diff --git a/sys/contrib/openzfs/cmd/zed/agents/zfs_retire.c b/sys/contrib/openzfs/cmd/zed/agents/zfs_retire.c index 28714ec295bb..f83ae09259ab 100644 --- a/sys/contrib/openzfs/cmd/zed/agents/zfs_retire.c +++ b/sys/contrib/openzfs/cmd/zed/agents/zfs_retire.c @@ -1,661 +1,663 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2006, 2010, Oracle and/or its affiliates. All rights reserved. * * Copyright (c) 2016, Intel Corporation. * Copyright (c) 2018, loli10K */ /* * The ZFS retire agent is responsible for managing hot spares across all pools. * When we see a device fault or a device removal, we try to open the associated * pool and look for any hot spares. We iterate over any available hot spares * and attempt a 'zpool replace' for each one. * * For vdevs diagnosed as faulty, the agent is also responsible for proactively * marking the vdev FAULTY (for I/O errors) or DEGRADED (for checksum errors). */ #include #include #include #include #include #include #include #include "zfs_agents.h" #include "fmd_api.h" typedef struct zfs_retire_repaired { struct zfs_retire_repaired *zrr_next; uint64_t zrr_pool; uint64_t zrr_vdev; } zfs_retire_repaired_t; typedef struct zfs_retire_data { libzfs_handle_t *zrd_hdl; zfs_retire_repaired_t *zrd_repaired; } zfs_retire_data_t; static void zfs_retire_clear_data(fmd_hdl_t *hdl, zfs_retire_data_t *zdp) { zfs_retire_repaired_t *zrp; while ((zrp = zdp->zrd_repaired) != NULL) { zdp->zrd_repaired = zrp->zrr_next; fmd_hdl_free(hdl, zrp, sizeof (zfs_retire_repaired_t)); } } /* * Find a pool with a matching GUID. */ typedef struct find_cbdata { uint64_t cb_guid; zpool_handle_t *cb_zhp; nvlist_t *cb_vdev; uint64_t cb_vdev_guid; uint64_t cb_num_spares; } find_cbdata_t; static int find_pool(zpool_handle_t *zhp, void *data) { find_cbdata_t *cbp = data; if (cbp->cb_guid == zpool_get_prop_int(zhp, ZPOOL_PROP_GUID, NULL)) { cbp->cb_zhp = zhp; return (1); } zpool_close(zhp); return (0); } /* * Find a vdev within a tree with a matching GUID. */ static nvlist_t * find_vdev(libzfs_handle_t *zhdl, nvlist_t *nv, uint64_t search_guid) { uint64_t guid; nvlist_t **child; uint_t c, children; nvlist_t *ret; if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) == 0 && guid == search_guid) { fmd_hdl_debug(fmd_module_hdl("zfs-retire"), "matched vdev %llu", guid); return (nv); } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) return (NULL); for (c = 0; c < children; c++) { if ((ret = find_vdev(zhdl, child[c], search_guid)) != NULL) return (ret); } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE, &child, &children) != 0) return (NULL); for (c = 0; c < children; c++) { if ((ret = find_vdev(zhdl, child[c], search_guid)) != NULL) return (ret); } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES, &child, &children) != 0) return (NULL); for (c = 0; c < children; c++) { if ((ret = find_vdev(zhdl, child[c], search_guid)) != NULL) return (ret); } return (NULL); } static int remove_spares(zpool_handle_t *zhp, void *data) { nvlist_t *config, *nvroot; nvlist_t **spares; uint_t nspares; char *devname; find_cbdata_t *cbp = data; uint64_t spareguid = 0; vdev_stat_t *vs; unsigned int c; config = zpool_get_config(zhp, NULL); if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) != 0) { zpool_close(zhp); return (0); } if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, &spares, &nspares) != 0) { zpool_close(zhp); return (0); } for (int i = 0; i < nspares; i++) { if (nvlist_lookup_uint64(spares[i], ZPOOL_CONFIG_GUID, &spareguid) == 0 && spareguid == cbp->cb_vdev_guid) { devname = zpool_vdev_name(NULL, zhp, spares[i], B_FALSE); nvlist_lookup_uint64_array(spares[i], ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &c); if (vs->vs_state != VDEV_STATE_REMOVED && zpool_vdev_remove_wanted(zhp, devname) == 0) cbp->cb_num_spares++; break; } } zpool_close(zhp); return (0); } /* * Given a vdev guid, find and remove all spares associated with it. */ static int find_and_remove_spares(libzfs_handle_t *zhdl, uint64_t vdev_guid) { find_cbdata_t cb; cb.cb_num_spares = 0; cb.cb_vdev_guid = vdev_guid; zpool_iter(zhdl, remove_spares, &cb); return (cb.cb_num_spares); } /* * Given a (pool, vdev) GUID pair, find the matching pool and vdev. */ static zpool_handle_t * find_by_guid(libzfs_handle_t *zhdl, uint64_t pool_guid, uint64_t vdev_guid, nvlist_t **vdevp) { find_cbdata_t cb; zpool_handle_t *zhp; nvlist_t *config, *nvroot; /* * Find the corresponding pool and make sure the vdev still exists. */ cb.cb_guid = pool_guid; if (zpool_iter(zhdl, find_pool, &cb) != 1) return (NULL); zhp = cb.cb_zhp; config = zpool_get_config(zhp, NULL); if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) != 0) { zpool_close(zhp); return (NULL); } if (vdev_guid != 0) { if ((*vdevp = find_vdev(zhdl, nvroot, vdev_guid)) == NULL) { zpool_close(zhp); return (NULL); } } return (zhp); } /* * Given a vdev, attempt to replace it with every known spare until one * succeeds or we run out of devices to try. * Return whether we were successful or not in replacing the device. */ static boolean_t replace_with_spare(fmd_hdl_t *hdl, zpool_handle_t *zhp, nvlist_t *vdev) { nvlist_t *config, *nvroot, *replacement; nvlist_t **spares; uint_t s, nspares; char *dev_name; zprop_source_t source; int ashift; config = zpool_get_config(zhp, NULL); if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) != 0) return (B_FALSE); /* * Find out if there are any hot spares available in the pool. */ if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, &spares, &nspares) != 0) return (B_FALSE); /* * lookup "ashift" pool property, we may need it for the replacement */ ashift = zpool_get_prop_int(zhp, ZPOOL_PROP_ASHIFT, &source); replacement = fmd_nvl_alloc(hdl, FMD_SLEEP); (void) nvlist_add_string(replacement, ZPOOL_CONFIG_TYPE, VDEV_TYPE_ROOT); dev_name = zpool_vdev_name(NULL, zhp, vdev, B_FALSE); /* * Try to replace each spare, ending when we successfully * replace it. */ for (s = 0; s < nspares; s++) { boolean_t rebuild = B_FALSE; const char *spare_name, *type; if (nvlist_lookup_string(spares[s], ZPOOL_CONFIG_PATH, &spare_name) != 0) continue; /* prefer sequential resilvering for distributed spares */ if ((nvlist_lookup_string(spares[s], ZPOOL_CONFIG_TYPE, &type) == 0) && strcmp(type, VDEV_TYPE_DRAID_SPARE) == 0) rebuild = B_TRUE; /* if set, add the "ashift" pool property to the spare nvlist */ if (source != ZPROP_SRC_DEFAULT) (void) nvlist_add_uint64(spares[s], ZPOOL_CONFIG_ASHIFT, ashift); (void) nvlist_add_nvlist_array(replacement, ZPOOL_CONFIG_CHILDREN, (const nvlist_t **)&spares[s], 1); fmd_hdl_debug(hdl, "zpool_vdev_replace '%s' with spare '%s'", dev_name, zfs_basename(spare_name)); if (zpool_vdev_attach(zhp, dev_name, spare_name, replacement, B_TRUE, rebuild) == 0) { free(dev_name); nvlist_free(replacement); return (B_TRUE); } } free(dev_name); nvlist_free(replacement); return (B_FALSE); } /* * Repair this vdev if we had diagnosed a 'fault.fs.zfs.device' and * ASRU is now usable. ZFS has found the device to be present and * functioning. */ static void zfs_vdev_repair(fmd_hdl_t *hdl, nvlist_t *nvl) { zfs_retire_data_t *zdp = fmd_hdl_getspecific(hdl); zfs_retire_repaired_t *zrp; uint64_t pool_guid, vdev_guid; if (nvlist_lookup_uint64(nvl, FM_EREPORT_PAYLOAD_ZFS_POOL_GUID, &pool_guid) != 0 || nvlist_lookup_uint64(nvl, FM_EREPORT_PAYLOAD_ZFS_VDEV_GUID, &vdev_guid) != 0) return; /* * Before checking the state of the ASRU, go through and see if we've * already made an attempt to repair this ASRU. This list is cleared * whenever we receive any kind of list event, and is designed to * prevent us from generating a feedback loop when we attempt repairs * against a faulted pool. The problem is that checking the unusable * state of the ASRU can involve opening the pool, which can post * statechange events but otherwise leave the pool in the faulted * state. This list allows us to detect when a statechange event is * due to our own request. */ for (zrp = zdp->zrd_repaired; zrp != NULL; zrp = zrp->zrr_next) { if (zrp->zrr_pool == pool_guid && zrp->zrr_vdev == vdev_guid) return; } zrp = fmd_hdl_alloc(hdl, sizeof (zfs_retire_repaired_t), FMD_SLEEP); zrp->zrr_next = zdp->zrd_repaired; zrp->zrr_pool = pool_guid; zrp->zrr_vdev = vdev_guid; zdp->zrd_repaired = zrp; fmd_hdl_debug(hdl, "marking repaired vdev %llu on pool %llu", vdev_guid, pool_guid); } static void zfs_retire_recv(fmd_hdl_t *hdl, fmd_event_t *ep, nvlist_t *nvl, const char *class) { (void) ep; uint64_t pool_guid, vdev_guid; zpool_handle_t *zhp; nvlist_t *resource, *fault; nvlist_t **faults; uint_t f, nfaults; zfs_retire_data_t *zdp = fmd_hdl_getspecific(hdl); libzfs_handle_t *zhdl = zdp->zrd_hdl; boolean_t fault_device, degrade_device; boolean_t is_repair; boolean_t l2arc = B_FALSE; boolean_t spare = B_FALSE; const char *scheme; nvlist_t *vdev = NULL; const char *uuid; int repair_done = 0; boolean_t retire; boolean_t is_disk; vdev_aux_t aux; uint64_t state = 0; vdev_stat_t *vs; unsigned int c; fmd_hdl_debug(hdl, "zfs_retire_recv: '%s'", class); (void) nvlist_lookup_uint64(nvl, FM_EREPORT_PAYLOAD_ZFS_VDEV_STATE, &state); /* * If this is a resource notifying us of device removal then simply * check for an available spare and continue unless the device is a * l2arc vdev, in which case we just offline it. */ if (strcmp(class, "resource.fs.zfs.removed") == 0 || (strcmp(class, "resource.fs.zfs.statechange") == 0 && (state == VDEV_STATE_REMOVED || state == VDEV_STATE_FAULTED))) { const char *devtype; char *devname; if (nvlist_lookup_string(nvl, FM_EREPORT_PAYLOAD_ZFS_VDEV_TYPE, &devtype) == 0) { if (strcmp(devtype, VDEV_TYPE_SPARE) == 0) spare = B_TRUE; else if (strcmp(devtype, VDEV_TYPE_L2CACHE) == 0) l2arc = B_TRUE; } if (nvlist_lookup_uint64(nvl, FM_EREPORT_PAYLOAD_ZFS_VDEV_GUID, &vdev_guid) != 0) return; if (spare) { int nspares = find_and_remove_spares(zhdl, vdev_guid); fmd_hdl_debug(hdl, "%d spares removed", nspares); return; } if (nvlist_lookup_uint64(nvl, FM_EREPORT_PAYLOAD_ZFS_POOL_GUID, &pool_guid) != 0) return; if ((zhp = find_by_guid(zhdl, pool_guid, vdev_guid, &vdev)) == NULL) return; devname = zpool_vdev_name(NULL, zhp, vdev, B_FALSE); nvlist_lookup_uint64_array(vdev, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &c); /* * If state removed is requested for already removed vdev, * its a loopback event from spa_async_remove(). Just * ignore it. */ if (vs->vs_state == VDEV_STATE_REMOVED && state == VDEV_STATE_REMOVED) return; /* Remove the vdev since device is unplugged */ + int remove_status = 0; if (l2arc || (strcmp(class, "resource.fs.zfs.removed") == 0)) { - int status = zpool_vdev_remove_wanted(zhp, devname); + remove_status = zpool_vdev_remove_wanted(zhp, devname); fmd_hdl_debug(hdl, "zpool_vdev_remove_wanted '%s'" - ", ret:%d", devname, status); + ", err:%d", devname, libzfs_errno(zhdl)); } /* Replace the vdev with a spare if its not a l2arc */ - if (!l2arc && (!fmd_prop_get_int32(hdl, "spare_on_remove") || + if (!l2arc && !remove_status && + (!fmd_prop_get_int32(hdl, "spare_on_remove") || replace_with_spare(hdl, zhp, vdev) == B_FALSE)) { /* Could not handle with spare */ fmd_hdl_debug(hdl, "no spare for '%s'", devname); } free(devname); zpool_close(zhp); return; } if (strcmp(class, FM_LIST_RESOLVED_CLASS) == 0) return; /* * Note: on Linux statechange events are more than just * healthy ones so we need to confirm the actual state value. */ if (strcmp(class, "resource.fs.zfs.statechange") == 0 && state == VDEV_STATE_HEALTHY) { zfs_vdev_repair(hdl, nvl); return; } if (strcmp(class, "sysevent.fs.zfs.vdev_remove") == 0) { zfs_vdev_repair(hdl, nvl); return; } zfs_retire_clear_data(hdl, zdp); if (strcmp(class, FM_LIST_REPAIRED_CLASS) == 0) is_repair = B_TRUE; else is_repair = B_FALSE; /* * We subscribe to zfs faults as well as all repair events. */ if (nvlist_lookup_nvlist_array(nvl, FM_SUSPECT_FAULT_LIST, &faults, &nfaults) != 0) return; for (f = 0; f < nfaults; f++) { fault = faults[f]; fault_device = B_FALSE; degrade_device = B_FALSE; is_disk = B_FALSE; if (nvlist_lookup_boolean_value(fault, FM_SUSPECT_RETIRE, &retire) == 0 && retire == 0) continue; /* * While we subscribe to fault.fs.zfs.*, we only take action * for faults targeting a specific vdev (open failure or SERD * failure). We also subscribe to fault.io.* events, so that * faulty disks will be faulted in the ZFS configuration. */ if (fmd_nvl_class_match(hdl, fault, "fault.fs.zfs.vdev.io")) { fault_device = B_TRUE; } else if (fmd_nvl_class_match(hdl, fault, "fault.fs.zfs.vdev.checksum")) { degrade_device = B_TRUE; } else if (fmd_nvl_class_match(hdl, fault, "fault.fs.zfs.device")) { fault_device = B_FALSE; } else if (fmd_nvl_class_match(hdl, fault, "fault.io.*")) { is_disk = B_TRUE; fault_device = B_TRUE; } else { continue; } if (is_disk) { continue; } else { /* * This is a ZFS fault. Lookup the resource, and * attempt to find the matching vdev. */ if (nvlist_lookup_nvlist(fault, FM_FAULT_RESOURCE, &resource) != 0 || nvlist_lookup_string(resource, FM_FMRI_SCHEME, &scheme) != 0) continue; if (strcmp(scheme, FM_FMRI_SCHEME_ZFS) != 0) continue; if (nvlist_lookup_uint64(resource, FM_FMRI_ZFS_POOL, &pool_guid) != 0) continue; if (nvlist_lookup_uint64(resource, FM_FMRI_ZFS_VDEV, &vdev_guid) != 0) { if (is_repair) vdev_guid = 0; else continue; } if ((zhp = find_by_guid(zhdl, pool_guid, vdev_guid, &vdev)) == NULL) continue; aux = VDEV_AUX_ERR_EXCEEDED; } if (vdev_guid == 0) { /* * For pool-level repair events, clear the entire pool. */ fmd_hdl_debug(hdl, "zpool_clear of pool '%s'", zpool_get_name(zhp)); (void) zpool_clear(zhp, NULL, NULL); zpool_close(zhp); continue; } /* * If this is a repair event, then mark the vdev as repaired and * continue. */ if (is_repair) { repair_done = 1; fmd_hdl_debug(hdl, "zpool_clear of pool '%s' vdev %llu", zpool_get_name(zhp), vdev_guid); (void) zpool_vdev_clear(zhp, vdev_guid); zpool_close(zhp); continue; } /* * Actively fault the device if needed. */ if (fault_device) (void) zpool_vdev_fault(zhp, vdev_guid, aux); if (degrade_device) (void) zpool_vdev_degrade(zhp, vdev_guid, aux); if (fault_device || degrade_device) fmd_hdl_debug(hdl, "zpool_vdev_%s: vdev %llu on '%s'", fault_device ? "fault" : "degrade", vdev_guid, zpool_get_name(zhp)); /* * Attempt to substitute a hot spare. */ (void) replace_with_spare(hdl, zhp, vdev); zpool_close(zhp); } if (strcmp(class, FM_LIST_REPAIRED_CLASS) == 0 && repair_done && nvlist_lookup_string(nvl, FM_SUSPECT_UUID, &uuid) == 0) fmd_case_uuresolved(hdl, uuid); } static const fmd_hdl_ops_t fmd_ops = { zfs_retire_recv, /* fmdo_recv */ NULL, /* fmdo_timeout */ NULL, /* fmdo_close */ NULL, /* fmdo_stats */ NULL, /* fmdo_gc */ }; static const fmd_prop_t fmd_props[] = { { "spare_on_remove", FMD_TYPE_BOOL, "true" }, { NULL, 0, NULL } }; static const fmd_hdl_info_t fmd_info = { "ZFS Retire Agent", "1.0", &fmd_ops, fmd_props }; void _zfs_retire_init(fmd_hdl_t *hdl) { zfs_retire_data_t *zdp; libzfs_handle_t *zhdl; if ((zhdl = libzfs_init()) == NULL) return; if (fmd_hdl_register(hdl, FMD_API_VERSION, &fmd_info) != 0) { libzfs_fini(zhdl); return; } zdp = fmd_hdl_zalloc(hdl, sizeof (zfs_retire_data_t), FMD_SLEEP); zdp->zrd_hdl = zhdl; fmd_hdl_setspecific(hdl, zdp); } void _zfs_retire_fini(fmd_hdl_t *hdl) { zfs_retire_data_t *zdp = fmd_hdl_getspecific(hdl); if (zdp != NULL) { zfs_retire_clear_data(hdl, zdp); libzfs_fini(zdp->zrd_hdl); fmd_hdl_free(hdl, zdp, sizeof (zfs_retire_data_t)); } } diff --git a/sys/contrib/openzfs/cmd/zpool/zpool_main.c b/sys/contrib/openzfs/cmd/zpool/zpool_main.c index 301c5f4bfc6f..013dd4a23380 100644 --- a/sys/contrib/openzfs/cmd/zpool/zpool_main.c +++ b/sys/contrib/openzfs/cmd/zpool/zpool_main.c @@ -1,11060 +1,11163 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright 2011 Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2011, 2020 by Delphix. All rights reserved. * Copyright (c) 2012 by Frederik Wessels. All rights reserved. * Copyright (c) 2012 by Cyril Plisko. All rights reserved. * Copyright (c) 2013 by Prasad Joshi (sTec). All rights reserved. * Copyright 2016 Igor Kozhukhov . * Copyright (c) 2017 Datto Inc. * Copyright (c) 2017 Open-E, Inc. All Rights Reserved. * Copyright (c) 2017, Intel Corporation. * Copyright (c) 2019, loli10K * Copyright (c) 2021, Colm Buckley * Copyright (c) 2021, Klara Inc. * Copyright [2021] Hewlett Packard Enterprise Development LP */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "zpool_util.h" #include "zfs_comutil.h" #include "zfeature_common.h" #include "statcommon.h" libzfs_handle_t *g_zfs; static int zpool_do_create(int, char **); static int zpool_do_destroy(int, char **); static int zpool_do_add(int, char **); static int zpool_do_remove(int, char **); static int zpool_do_labelclear(int, char **); static int zpool_do_checkpoint(int, char **); static int zpool_do_list(int, char **); static int zpool_do_iostat(int, char **); static int zpool_do_status(int, char **); static int zpool_do_online(int, char **); static int zpool_do_offline(int, char **); static int zpool_do_clear(int, char **); static int zpool_do_reopen(int, char **); static int zpool_do_reguid(int, char **); static int zpool_do_attach(int, char **); static int zpool_do_detach(int, char **); static int zpool_do_replace(int, char **); static int zpool_do_split(int, char **); static int zpool_do_initialize(int, char **); static int zpool_do_scrub(int, char **); static int zpool_do_resilver(int, char **); static int zpool_do_trim(int, char **); static int zpool_do_import(int, char **); static int zpool_do_export(int, char **); static int zpool_do_upgrade(int, char **); static int zpool_do_history(int, char **); static int zpool_do_events(int, char **); static int zpool_do_get(int, char **); static int zpool_do_set(int, char **); static int zpool_do_sync(int, char **); static int zpool_do_version(int, char **); static int zpool_do_wait(int, char **); static zpool_compat_status_t zpool_do_load_compat( const char *, boolean_t *); /* * These libumem hooks provide a reasonable set of defaults for the allocator's * debugging facilities. */ #ifdef DEBUG const char * _umem_debug_init(void) { return ("default,verbose"); /* $UMEM_DEBUG setting */ } const char * _umem_logging_init(void) { return ("fail,contents"); /* $UMEM_LOGGING setting */ } #endif typedef enum { HELP_ADD, HELP_ATTACH, HELP_CLEAR, HELP_CREATE, HELP_CHECKPOINT, HELP_DESTROY, HELP_DETACH, HELP_EXPORT, HELP_HISTORY, HELP_IMPORT, HELP_IOSTAT, HELP_LABELCLEAR, HELP_LIST, HELP_OFFLINE, HELP_ONLINE, HELP_REPLACE, HELP_REMOVE, HELP_INITIALIZE, HELP_SCRUB, HELP_RESILVER, HELP_TRIM, HELP_STATUS, HELP_UPGRADE, HELP_EVENTS, HELP_GET, HELP_SET, HELP_SPLIT, HELP_SYNC, HELP_REGUID, HELP_REOPEN, HELP_VERSION, HELP_WAIT } zpool_help_t; /* * Flags for stats to display with "zpool iostats" */ enum iostat_type { IOS_DEFAULT = 0, IOS_LATENCY = 1, IOS_QUEUES = 2, IOS_L_HISTO = 3, IOS_RQ_HISTO = 4, IOS_COUNT, /* always last element */ }; /* iostat_type entries as bitmasks */ #define IOS_DEFAULT_M (1ULL << IOS_DEFAULT) #define IOS_LATENCY_M (1ULL << IOS_LATENCY) #define IOS_QUEUES_M (1ULL << IOS_QUEUES) #define IOS_L_HISTO_M (1ULL << IOS_L_HISTO) #define IOS_RQ_HISTO_M (1ULL << IOS_RQ_HISTO) /* Mask of all the histo bits */ #define IOS_ANYHISTO_M (IOS_L_HISTO_M | IOS_RQ_HISTO_M) /* * Lookup table for iostat flags to nvlist names. Basically a list * of all the nvlists a flag requires. Also specifies the order in * which data gets printed in zpool iostat. */ static const char *vsx_type_to_nvlist[IOS_COUNT][15] = { [IOS_L_HISTO] = { ZPOOL_CONFIG_VDEV_TOT_R_LAT_HISTO, ZPOOL_CONFIG_VDEV_TOT_W_LAT_HISTO, ZPOOL_CONFIG_VDEV_DISK_R_LAT_HISTO, ZPOOL_CONFIG_VDEV_DISK_W_LAT_HISTO, ZPOOL_CONFIG_VDEV_SYNC_R_LAT_HISTO, ZPOOL_CONFIG_VDEV_SYNC_W_LAT_HISTO, ZPOOL_CONFIG_VDEV_ASYNC_R_LAT_HISTO, ZPOOL_CONFIG_VDEV_ASYNC_W_LAT_HISTO, ZPOOL_CONFIG_VDEV_SCRUB_LAT_HISTO, ZPOOL_CONFIG_VDEV_TRIM_LAT_HISTO, ZPOOL_CONFIG_VDEV_REBUILD_LAT_HISTO, NULL}, [IOS_LATENCY] = { ZPOOL_CONFIG_VDEV_TOT_R_LAT_HISTO, ZPOOL_CONFIG_VDEV_TOT_W_LAT_HISTO, ZPOOL_CONFIG_VDEV_DISK_R_LAT_HISTO, ZPOOL_CONFIG_VDEV_DISK_W_LAT_HISTO, ZPOOL_CONFIG_VDEV_TRIM_LAT_HISTO, ZPOOL_CONFIG_VDEV_REBUILD_LAT_HISTO, NULL}, [IOS_QUEUES] = { ZPOOL_CONFIG_VDEV_SYNC_R_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_SYNC_W_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_ASYNC_R_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_ASYNC_W_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_SCRUB_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_TRIM_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_REBUILD_ACTIVE_QUEUE, NULL}, [IOS_RQ_HISTO] = { ZPOOL_CONFIG_VDEV_SYNC_IND_R_HISTO, ZPOOL_CONFIG_VDEV_SYNC_AGG_R_HISTO, ZPOOL_CONFIG_VDEV_SYNC_IND_W_HISTO, ZPOOL_CONFIG_VDEV_SYNC_AGG_W_HISTO, ZPOOL_CONFIG_VDEV_ASYNC_IND_R_HISTO, ZPOOL_CONFIG_VDEV_ASYNC_AGG_R_HISTO, ZPOOL_CONFIG_VDEV_ASYNC_IND_W_HISTO, ZPOOL_CONFIG_VDEV_ASYNC_AGG_W_HISTO, ZPOOL_CONFIG_VDEV_IND_SCRUB_HISTO, ZPOOL_CONFIG_VDEV_AGG_SCRUB_HISTO, ZPOOL_CONFIG_VDEV_IND_TRIM_HISTO, ZPOOL_CONFIG_VDEV_AGG_TRIM_HISTO, ZPOOL_CONFIG_VDEV_IND_REBUILD_HISTO, ZPOOL_CONFIG_VDEV_AGG_REBUILD_HISTO, NULL}, }; /* * Given a cb->cb_flags with a histogram bit set, return the iostat_type. * Right now, only one histo bit is ever set at one time, so we can * just do a highbit64(a) */ #define IOS_HISTO_IDX(a) (highbit64(a & IOS_ANYHISTO_M) - 1) typedef struct zpool_command { const char *name; int (*func)(int, char **); zpool_help_t usage; } zpool_command_t; /* * Master command table. Each ZFS command has a name, associated function, and * usage message. The usage messages need to be internationalized, so we have * to have a function to return the usage message based on a command index. * * These commands are organized according to how they are displayed in the usage * message. An empty command (one with a NULL name) indicates an empty line in * the generic usage message. */ static zpool_command_t command_table[] = { { "version", zpool_do_version, HELP_VERSION }, { NULL }, { "create", zpool_do_create, HELP_CREATE }, { "destroy", zpool_do_destroy, HELP_DESTROY }, { NULL }, { "add", zpool_do_add, HELP_ADD }, { "remove", zpool_do_remove, HELP_REMOVE }, { NULL }, { "labelclear", zpool_do_labelclear, HELP_LABELCLEAR }, { NULL }, { "checkpoint", zpool_do_checkpoint, HELP_CHECKPOINT }, { NULL }, { "list", zpool_do_list, HELP_LIST }, { "iostat", zpool_do_iostat, HELP_IOSTAT }, { "status", zpool_do_status, HELP_STATUS }, { NULL }, { "online", zpool_do_online, HELP_ONLINE }, { "offline", zpool_do_offline, HELP_OFFLINE }, { "clear", zpool_do_clear, HELP_CLEAR }, { "reopen", zpool_do_reopen, HELP_REOPEN }, { NULL }, { "attach", zpool_do_attach, HELP_ATTACH }, { "detach", zpool_do_detach, HELP_DETACH }, { "replace", zpool_do_replace, HELP_REPLACE }, { "split", zpool_do_split, HELP_SPLIT }, { NULL }, { "initialize", zpool_do_initialize, HELP_INITIALIZE }, { "resilver", zpool_do_resilver, HELP_RESILVER }, { "scrub", zpool_do_scrub, HELP_SCRUB }, { "trim", zpool_do_trim, HELP_TRIM }, { NULL }, { "import", zpool_do_import, HELP_IMPORT }, { "export", zpool_do_export, HELP_EXPORT }, { "upgrade", zpool_do_upgrade, HELP_UPGRADE }, { "reguid", zpool_do_reguid, HELP_REGUID }, { NULL }, { "history", zpool_do_history, HELP_HISTORY }, { "events", zpool_do_events, HELP_EVENTS }, { NULL }, { "get", zpool_do_get, HELP_GET }, { "set", zpool_do_set, HELP_SET }, { "sync", zpool_do_sync, HELP_SYNC }, { NULL }, { "wait", zpool_do_wait, HELP_WAIT }, }; #define NCOMMAND (ARRAY_SIZE(command_table)) #define VDEV_ALLOC_CLASS_LOGS "logs" static zpool_command_t *current_command; static zfs_type_t current_prop_type = (ZFS_TYPE_POOL | ZFS_TYPE_VDEV); static char history_str[HIS_MAX_RECORD_LEN]; static boolean_t log_history = B_TRUE; static uint_t timestamp_fmt = NODATE; static const char * get_usage(zpool_help_t idx) { switch (idx) { case HELP_ADD: return (gettext("\tadd [-fgLnP] [-o property=value] " " ...\n")); case HELP_ATTACH: return (gettext("\tattach [-fsw] [-o property=value] " " \n")); case HELP_CLEAR: return (gettext("\tclear [-nF] [device]\n")); case HELP_CREATE: return (gettext("\tcreate [-fnd] [-o property=value] ... \n" "\t [-O file-system-property=value] ... \n" "\t [-m mountpoint] [-R root] ...\n")); case HELP_CHECKPOINT: return (gettext("\tcheckpoint [-d [-w]] ...\n")); case HELP_DESTROY: return (gettext("\tdestroy [-f] \n")); case HELP_DETACH: return (gettext("\tdetach \n")); case HELP_EXPORT: return (gettext("\texport [-af] ...\n")); case HELP_HISTORY: return (gettext("\thistory [-il] [] ...\n")); case HELP_IMPORT: return (gettext("\timport [-d dir] [-D]\n" "\timport [-o mntopts] [-o property=value] ... \n" "\t [-d dir | -c cachefile] [-D] [-l] [-f] [-m] [-N] " "[-R root] [-F [-n]] -a\n" "\timport [-o mntopts] [-o property=value] ... \n" "\t [-d dir | -c cachefile] [-D] [-l] [-f] [-m] [-N] " "[-R root] [-F [-n]]\n" "\t [--rewind-to-checkpoint] [newpool]\n")); case HELP_IOSTAT: return (gettext("\tiostat [[[-c [script1,script2,...]" "[-lq]]|[-rw]] [-T d | u] [-ghHLpPvy]\n" "\t [[pool ...]|[pool vdev ...]|[vdev ...]]" " [[-n] interval [count]]\n")); case HELP_LABELCLEAR: return (gettext("\tlabelclear [-f] \n")); case HELP_LIST: return (gettext("\tlist [-gHLpPv] [-o property[,...]] " "[-T d|u] [pool] ... \n" "\t [interval [count]]\n")); case HELP_OFFLINE: return (gettext("\toffline [-f] [-t] ...\n")); case HELP_ONLINE: return (gettext("\tonline [-e] ...\n")); case HELP_REPLACE: return (gettext("\treplace [-fsw] [-o property=value] " " [new-device]\n")); case HELP_REMOVE: return (gettext("\tremove [-npsw] ...\n")); case HELP_REOPEN: return (gettext("\treopen [-n] \n")); case HELP_INITIALIZE: - return (gettext("\tinitialize [-c | -s] [-w] " + return (gettext("\tinitialize [-c | -s | -u] [-w] " "[ ...]\n")); case HELP_SCRUB: - return (gettext("\tscrub [-s | -p] [-w] ...\n")); + return (gettext("\tscrub [-s | -p] [-w] [-e] ...\n")); case HELP_RESILVER: return (gettext("\tresilver ...\n")); case HELP_TRIM: return (gettext("\ttrim [-dw] [-r ] [-c | -s] " "[ ...]\n")); case HELP_STATUS: return (gettext("\tstatus [-c [script1,script2,...]] " "[-igLpPstvxD] [-T d|u] [pool] ... \n" "\t [interval [count]]\n")); case HELP_UPGRADE: return (gettext("\tupgrade\n" "\tupgrade -v\n" "\tupgrade [-V version] <-a | pool ...>\n")); case HELP_EVENTS: return (gettext("\tevents [-vHf [pool] | -c]\n")); case HELP_GET: return (gettext("\tget [-Hp] [-o \"all\" | field[,...]] " "<\"all\" | property[,...]> ...\n")); case HELP_SET: return (gettext("\tset \n" "\tset \n")); case HELP_SPLIT: return (gettext("\tsplit [-gLnPl] [-R altroot] [-o mntopts]\n" "\t [-o property=value] " "[ ...]\n")); case HELP_REGUID: return (gettext("\treguid \n")); case HELP_SYNC: return (gettext("\tsync [pool] ...\n")); case HELP_VERSION: return (gettext("\tversion\n")); case HELP_WAIT: return (gettext("\twait [-Hp] [-T d|u] [-t [,...]] " " [interval]\n")); default: __builtin_unreachable(); } } static void zpool_collect_leaves(zpool_handle_t *zhp, nvlist_t *nvroot, nvlist_t *res) { uint_t children = 0; nvlist_t **child; uint_t i; (void) nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN, &child, &children); if (children == 0) { char *path = zpool_vdev_name(g_zfs, zhp, nvroot, VDEV_NAME_PATH); if (strcmp(path, VDEV_TYPE_INDIRECT) != 0 && strcmp(path, VDEV_TYPE_HOLE) != 0) fnvlist_add_boolean(res, path); free(path); return; } for (i = 0; i < children; i++) { zpool_collect_leaves(zhp, child[i], res); } } /* * Callback routine that will print out a pool property value. */ static int print_pool_prop_cb(int prop, void *cb) { FILE *fp = cb; (void) fprintf(fp, "\t%-19s ", zpool_prop_to_name(prop)); if (zpool_prop_readonly(prop)) (void) fprintf(fp, " NO "); else (void) fprintf(fp, " YES "); if (zpool_prop_values(prop) == NULL) (void) fprintf(fp, "-\n"); else (void) fprintf(fp, "%s\n", zpool_prop_values(prop)); return (ZPROP_CONT); } /* * Callback routine that will print out a vdev property value. */ static int print_vdev_prop_cb(int prop, void *cb) { FILE *fp = cb; (void) fprintf(fp, "\t%-19s ", vdev_prop_to_name(prop)); if (vdev_prop_readonly(prop)) (void) fprintf(fp, " NO "); else (void) fprintf(fp, " YES "); if (vdev_prop_values(prop) == NULL) (void) fprintf(fp, "-\n"); else (void) fprintf(fp, "%s\n", vdev_prop_values(prop)); return (ZPROP_CONT); } /* * Display usage message. If we're inside a command, display only the usage for * that command. Otherwise, iterate over the entire command table and display * a complete usage message. */ static __attribute__((noreturn)) void usage(boolean_t requested) { FILE *fp = requested ? stdout : stderr; if (current_command == NULL) { int i; (void) fprintf(fp, gettext("usage: zpool command args ...\n")); (void) fprintf(fp, gettext("where 'command' is one of the following:\n\n")); for (i = 0; i < NCOMMAND; i++) { if (command_table[i].name == NULL) (void) fprintf(fp, "\n"); else (void) fprintf(fp, "%s", get_usage(command_table[i].usage)); } } else { (void) fprintf(fp, gettext("usage:\n")); (void) fprintf(fp, "%s", get_usage(current_command->usage)); } if (current_command != NULL && current_prop_type != (ZFS_TYPE_POOL | ZFS_TYPE_VDEV) && ((strcmp(current_command->name, "set") == 0) || (strcmp(current_command->name, "get") == 0) || (strcmp(current_command->name, "list") == 0))) { (void) fprintf(fp, "%s", gettext("\nthe following properties are supported:\n")); (void) fprintf(fp, "\n\t%-19s %s %s\n\n", "PROPERTY", "EDIT", "VALUES"); /* Iterate over all properties */ if (current_prop_type == ZFS_TYPE_POOL) { (void) zprop_iter(print_pool_prop_cb, fp, B_FALSE, B_TRUE, current_prop_type); (void) fprintf(fp, "\t%-19s ", "feature@..."); (void) fprintf(fp, "YES " "disabled | enabled | active\n"); (void) fprintf(fp, gettext("\nThe feature@ properties " "must be appended with a feature name.\n" "See zpool-features(7).\n")); } else if (current_prop_type == ZFS_TYPE_VDEV) { (void) zprop_iter(print_vdev_prop_cb, fp, B_FALSE, B_TRUE, current_prop_type); } } /* * See comments at end of main(). */ if (getenv("ZFS_ABORT") != NULL) { (void) printf("dumping core by request\n"); abort(); } exit(requested ? 0 : 2); } /* - * zpool initialize [-c | -s] [-w] [ ...] + * zpool initialize [-c | -s | -u] [-w] [ ...] * Initialize all unused blocks in the specified vdevs, or all vdevs in the pool * if none specified. * * -c Cancel. Ends active initializing. * -s Suspend. Initializing can then be restarted with no flags. + * -u Uninitialize. Clears initialization state. * -w Wait. Blocks until initializing has completed. */ int zpool_do_initialize(int argc, char **argv) { int c; char *poolname; zpool_handle_t *zhp; nvlist_t *vdevs; int err = 0; boolean_t wait = B_FALSE; struct option long_options[] = { {"cancel", no_argument, NULL, 'c'}, {"suspend", no_argument, NULL, 's'}, + {"uninit", no_argument, NULL, 'u'}, {"wait", no_argument, NULL, 'w'}, {0, 0, 0, 0} }; pool_initialize_func_t cmd_type = POOL_INITIALIZE_START; - while ((c = getopt_long(argc, argv, "csw", long_options, NULL)) != -1) { + while ((c = getopt_long(argc, argv, "csuw", long_options, + NULL)) != -1) { switch (c) { case 'c': if (cmd_type != POOL_INITIALIZE_START && cmd_type != POOL_INITIALIZE_CANCEL) { (void) fprintf(stderr, gettext("-c cannot be " "combined with other options\n")); usage(B_FALSE); } cmd_type = POOL_INITIALIZE_CANCEL; break; case 's': if (cmd_type != POOL_INITIALIZE_START && cmd_type != POOL_INITIALIZE_SUSPEND) { (void) fprintf(stderr, gettext("-s cannot be " "combined with other options\n")); usage(B_FALSE); } cmd_type = POOL_INITIALIZE_SUSPEND; break; + case 'u': + if (cmd_type != POOL_INITIALIZE_START && + cmd_type != POOL_INITIALIZE_UNINIT) { + (void) fprintf(stderr, gettext("-u cannot be " + "combined with other options\n")); + usage(B_FALSE); + } + cmd_type = POOL_INITIALIZE_UNINIT; + break; case 'w': wait = B_TRUE; break; case '?': if (optopt != 0) { (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); } else { (void) fprintf(stderr, gettext("invalid option '%s'\n"), argv[optind - 1]); } usage(B_FALSE); } } argc -= optind; argv += optind; if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name argument\n")); usage(B_FALSE); return (-1); } if (wait && (cmd_type != POOL_INITIALIZE_START)) { - (void) fprintf(stderr, gettext("-w cannot be used with -c or " - "-s\n")); + (void) fprintf(stderr, gettext("-w cannot be used with -c, -s" + "or -u\n")); usage(B_FALSE); } poolname = argv[0]; zhp = zpool_open(g_zfs, poolname); if (zhp == NULL) return (-1); vdevs = fnvlist_alloc(); if (argc == 1) { /* no individual leaf vdevs specified, so add them all */ nvlist_t *config = zpool_get_config(zhp, NULL); nvlist_t *nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE); zpool_collect_leaves(zhp, nvroot, vdevs); } else { for (int i = 1; i < argc; i++) { fnvlist_add_boolean(vdevs, argv[i]); } } if (wait) err = zpool_initialize_wait(zhp, cmd_type, vdevs); else err = zpool_initialize(zhp, cmd_type, vdevs); fnvlist_free(vdevs); zpool_close(zhp); return (err); } /* * print a pool vdev config for dry runs */ static void print_vdev_tree(zpool_handle_t *zhp, const char *name, nvlist_t *nv, int indent, const char *match, int name_flags) { nvlist_t **child; uint_t c, children; char *vname; boolean_t printed = B_FALSE; if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) { if (name != NULL) (void) printf("\t%*s%s\n", indent, "", name); return; } for (c = 0; c < children; c++) { uint64_t is_log = B_FALSE, is_hole = B_FALSE; const char *class = ""; (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_HOLE, &is_hole); if (is_hole == B_TRUE) { continue; } (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, &is_log); if (is_log) class = VDEV_ALLOC_BIAS_LOG; (void) nvlist_lookup_string(child[c], ZPOOL_CONFIG_ALLOCATION_BIAS, &class); if (strcmp(match, class) != 0) continue; if (!printed && name != NULL) { (void) printf("\t%*s%s\n", indent, "", name); printed = B_TRUE; } vname = zpool_vdev_name(g_zfs, zhp, child[c], name_flags); print_vdev_tree(zhp, vname, child[c], indent + 2, "", name_flags); free(vname); } } /* * Print the list of l2cache devices for dry runs. */ static void print_cache_list(nvlist_t *nv, int indent) { nvlist_t **child; uint_t c, children; if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE, &child, &children) == 0 && children > 0) { (void) printf("\t%*s%s\n", indent, "", "cache"); } else { return; } for (c = 0; c < children; c++) { char *vname; vname = zpool_vdev_name(g_zfs, NULL, child[c], 0); (void) printf("\t%*s%s\n", indent + 2, "", vname); free(vname); } } /* * Print the list of spares for dry runs. */ static void print_spare_list(nvlist_t *nv, int indent) { nvlist_t **child; uint_t c, children; if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES, &child, &children) == 0 && children > 0) { (void) printf("\t%*s%s\n", indent, "", "spares"); } else { return; } for (c = 0; c < children; c++) { char *vname; vname = zpool_vdev_name(g_zfs, NULL, child[c], 0); (void) printf("\t%*s%s\n", indent + 2, "", vname); free(vname); } } static boolean_t prop_list_contains_feature(nvlist_t *proplist) { nvpair_t *nvp; for (nvp = nvlist_next_nvpair(proplist, NULL); NULL != nvp; nvp = nvlist_next_nvpair(proplist, nvp)) { if (zpool_prop_feature(nvpair_name(nvp))) return (B_TRUE); } return (B_FALSE); } /* * Add a property pair (name, string-value) into a property nvlist. */ static int add_prop_list(const char *propname, const char *propval, nvlist_t **props, boolean_t poolprop) { zpool_prop_t prop = ZPOOL_PROP_INVAL; nvlist_t *proplist; const char *normnm; const char *strval; if (*props == NULL && nvlist_alloc(props, NV_UNIQUE_NAME, 0) != 0) { (void) fprintf(stderr, gettext("internal error: out of memory\n")); return (1); } proplist = *props; if (poolprop) { const char *vname = zpool_prop_to_name(ZPOOL_PROP_VERSION); const char *cname = zpool_prop_to_name(ZPOOL_PROP_COMPATIBILITY); if ((prop = zpool_name_to_prop(propname)) == ZPOOL_PROP_INVAL && (!zpool_prop_feature(propname) && !zpool_prop_vdev(propname))) { (void) fprintf(stderr, gettext("property '%s' is " "not a valid pool or vdev property\n"), propname); return (2); } /* * feature@ properties and version should not be specified * at the same time. */ if ((prop == ZPOOL_PROP_INVAL && zpool_prop_feature(propname) && nvlist_exists(proplist, vname)) || (prop == ZPOOL_PROP_VERSION && prop_list_contains_feature(proplist))) { (void) fprintf(stderr, gettext("'feature@' and " "'version' properties cannot be specified " "together\n")); return (2); } /* * if version is specified, only "legacy" compatibility * may be requested */ if ((prop == ZPOOL_PROP_COMPATIBILITY && strcmp(propval, ZPOOL_COMPAT_LEGACY) != 0 && nvlist_exists(proplist, vname)) || (prop == ZPOOL_PROP_VERSION && nvlist_exists(proplist, cname) && strcmp(fnvlist_lookup_string(proplist, cname), ZPOOL_COMPAT_LEGACY) != 0)) { (void) fprintf(stderr, gettext("when 'version' is " "specified, the 'compatibility' feature may only " "be set to '" ZPOOL_COMPAT_LEGACY "'\n")); return (2); } if (zpool_prop_feature(propname) || zpool_prop_vdev(propname)) normnm = propname; else normnm = zpool_prop_to_name(prop); } else { zfs_prop_t fsprop = zfs_name_to_prop(propname); if (zfs_prop_valid_for_type(fsprop, ZFS_TYPE_FILESYSTEM, B_FALSE)) { normnm = zfs_prop_to_name(fsprop); } else if (zfs_prop_user(propname) || zfs_prop_userquota(propname)) { normnm = propname; } else { (void) fprintf(stderr, gettext("property '%s' is " "not a valid filesystem property\n"), propname); return (2); } } if (nvlist_lookup_string(proplist, normnm, &strval) == 0 && prop != ZPOOL_PROP_CACHEFILE) { (void) fprintf(stderr, gettext("property '%s' " "specified multiple times\n"), propname); return (2); } if (nvlist_add_string(proplist, normnm, propval) != 0) { (void) fprintf(stderr, gettext("internal " "error: out of memory\n")); return (1); } return (0); } /* * Set a default property pair (name, string-value) in a property nvlist */ static int add_prop_list_default(const char *propname, const char *propval, nvlist_t **props) { const char *pval; if (nvlist_lookup_string(*props, propname, &pval) == 0) return (0); return (add_prop_list(propname, propval, props, B_TRUE)); } /* * zpool add [-fgLnP] [-o property=value] ... * * -f Force addition of devices, even if they appear in use * -g Display guid for individual vdev name. * -L Follow links when resolving vdev path name. * -n Do not add the devices, but display the resulting layout if * they were to be added. * -o Set property=value. * -P Display full path for vdev name. * * Adds the given vdevs to 'pool'. As with create, the bulk of this work is * handled by make_root_vdev(), which constructs the nvlist needed to pass to * libzfs. */ int zpool_do_add(int argc, char **argv) { boolean_t force = B_FALSE; boolean_t dryrun = B_FALSE; int name_flags = 0; int c; nvlist_t *nvroot; char *poolname; int ret; zpool_handle_t *zhp; nvlist_t *config; nvlist_t *props = NULL; char *propval; /* check options */ while ((c = getopt(argc, argv, "fgLno:P")) != -1) { switch (c) { case 'f': force = B_TRUE; break; case 'g': name_flags |= VDEV_NAME_GUID; break; case 'L': name_flags |= VDEV_NAME_FOLLOW_LINKS; break; case 'n': dryrun = B_TRUE; break; case 'o': if ((propval = strchr(optarg, '=')) == NULL) { (void) fprintf(stderr, gettext("missing " "'=' for -o option\n")); usage(B_FALSE); } *propval = '\0'; propval++; if ((strcmp(optarg, ZPOOL_CONFIG_ASHIFT) != 0) || (add_prop_list(optarg, propval, &props, B_TRUE))) usage(B_FALSE); break; case 'P': name_flags |= VDEV_NAME_PATH; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* get pool name and check number of arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name argument\n")); usage(B_FALSE); } if (argc < 2) { (void) fprintf(stderr, gettext("missing vdev specification\n")); usage(B_FALSE); } poolname = argv[0]; argc--; argv++; if ((zhp = zpool_open(g_zfs, poolname)) == NULL) return (1); if ((config = zpool_get_config(zhp, NULL)) == NULL) { (void) fprintf(stderr, gettext("pool '%s' is unavailable\n"), poolname); zpool_close(zhp); return (1); } /* unless manually specified use "ashift" pool property (if set) */ if (!nvlist_exists(props, ZPOOL_CONFIG_ASHIFT)) { int intval; zprop_source_t src; char strval[ZPOOL_MAXPROPLEN]; intval = zpool_get_prop_int(zhp, ZPOOL_PROP_ASHIFT, &src); if (src != ZPROP_SRC_DEFAULT) { (void) sprintf(strval, "%" PRId32, intval); verify(add_prop_list(ZPOOL_CONFIG_ASHIFT, strval, &props, B_TRUE) == 0); } } /* pass off to make_root_vdev for processing */ nvroot = make_root_vdev(zhp, props, force, !force, B_FALSE, dryrun, argc, argv); if (nvroot == NULL) { zpool_close(zhp); return (1); } if (dryrun) { nvlist_t *poolnvroot; nvlist_t **l2child, **sparechild; uint_t l2children, sparechildren, c; char *vname; boolean_t hadcache = B_FALSE, hadspare = B_FALSE; verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &poolnvroot) == 0); (void) printf(gettext("would update '%s' to the following " "configuration:\n\n"), zpool_get_name(zhp)); /* print original main pool and new tree */ print_vdev_tree(zhp, poolname, poolnvroot, 0, "", name_flags | VDEV_NAME_TYPE_ID); print_vdev_tree(zhp, NULL, nvroot, 0, "", name_flags); /* print other classes: 'dedup', 'special', and 'log' */ if (zfs_special_devs(poolnvroot, VDEV_ALLOC_BIAS_DEDUP)) { print_vdev_tree(zhp, "dedup", poolnvroot, 0, VDEV_ALLOC_BIAS_DEDUP, name_flags); print_vdev_tree(zhp, NULL, nvroot, 0, VDEV_ALLOC_BIAS_DEDUP, name_flags); } else if (zfs_special_devs(nvroot, VDEV_ALLOC_BIAS_DEDUP)) { print_vdev_tree(zhp, "dedup", nvroot, 0, VDEV_ALLOC_BIAS_DEDUP, name_flags); } if (zfs_special_devs(poolnvroot, VDEV_ALLOC_BIAS_SPECIAL)) { print_vdev_tree(zhp, "special", poolnvroot, 0, VDEV_ALLOC_BIAS_SPECIAL, name_flags); print_vdev_tree(zhp, NULL, nvroot, 0, VDEV_ALLOC_BIAS_SPECIAL, name_flags); } else if (zfs_special_devs(nvroot, VDEV_ALLOC_BIAS_SPECIAL)) { print_vdev_tree(zhp, "special", nvroot, 0, VDEV_ALLOC_BIAS_SPECIAL, name_flags); } if (num_logs(poolnvroot) > 0) { print_vdev_tree(zhp, "logs", poolnvroot, 0, VDEV_ALLOC_BIAS_LOG, name_flags); print_vdev_tree(zhp, NULL, nvroot, 0, VDEV_ALLOC_BIAS_LOG, name_flags); } else if (num_logs(nvroot) > 0) { print_vdev_tree(zhp, "logs", nvroot, 0, VDEV_ALLOC_BIAS_LOG, name_flags); } /* Do the same for the caches */ if (nvlist_lookup_nvlist_array(poolnvroot, ZPOOL_CONFIG_L2CACHE, &l2child, &l2children) == 0 && l2children) { hadcache = B_TRUE; (void) printf(gettext("\tcache\n")); for (c = 0; c < l2children; c++) { vname = zpool_vdev_name(g_zfs, NULL, l2child[c], name_flags); (void) printf("\t %s\n", vname); free(vname); } } if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE, &l2child, &l2children) == 0 && l2children) { if (!hadcache) (void) printf(gettext("\tcache\n")); for (c = 0; c < l2children; c++) { vname = zpool_vdev_name(g_zfs, NULL, l2child[c], name_flags); (void) printf("\t %s\n", vname); free(vname); } } /* And finally the spares */ if (nvlist_lookup_nvlist_array(poolnvroot, ZPOOL_CONFIG_SPARES, &sparechild, &sparechildren) == 0 && sparechildren > 0) { hadspare = B_TRUE; (void) printf(gettext("\tspares\n")); for (c = 0; c < sparechildren; c++) { vname = zpool_vdev_name(g_zfs, NULL, sparechild[c], name_flags); (void) printf("\t %s\n", vname); free(vname); } } if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, &sparechild, &sparechildren) == 0 && sparechildren > 0) { if (!hadspare) (void) printf(gettext("\tspares\n")); for (c = 0; c < sparechildren; c++) { vname = zpool_vdev_name(g_zfs, NULL, sparechild[c], name_flags); (void) printf("\t %s\n", vname); free(vname); } } ret = 0; } else { ret = (zpool_add(zhp, nvroot) != 0); } nvlist_free(props); nvlist_free(nvroot); zpool_close(zhp); return (ret); } /* * zpool remove [-npsw] ... * * Removes the given vdev from the pool. */ int zpool_do_remove(int argc, char **argv) { char *poolname; int i, ret = 0; zpool_handle_t *zhp = NULL; boolean_t stop = B_FALSE; int c; boolean_t noop = B_FALSE; boolean_t parsable = B_FALSE; boolean_t wait = B_FALSE; /* check options */ while ((c = getopt(argc, argv, "npsw")) != -1) { switch (c) { case 'n': noop = B_TRUE; break; case 'p': parsable = B_TRUE; break; case 's': stop = B_TRUE; break; case 'w': wait = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* get pool name and check number of arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name argument\n")); usage(B_FALSE); } poolname = argv[0]; if ((zhp = zpool_open(g_zfs, poolname)) == NULL) return (1); if (stop && noop) { zpool_close(zhp); (void) fprintf(stderr, gettext("stop request ignored\n")); return (0); } if (stop) { if (argc > 1) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } if (zpool_vdev_remove_cancel(zhp) != 0) ret = 1; if (wait) { (void) fprintf(stderr, gettext("invalid option " "combination: -w cannot be used with -s\n")); usage(B_FALSE); } } else { if (argc < 2) { (void) fprintf(stderr, gettext("missing device\n")); usage(B_FALSE); } for (i = 1; i < argc; i++) { if (noop) { uint64_t size; if (zpool_vdev_indirect_size(zhp, argv[i], &size) != 0) { ret = 1; break; } if (parsable) { (void) printf("%s %llu\n", argv[i], (unsigned long long)size); } else { char valstr[32]; zfs_nicenum(size, valstr, sizeof (valstr)); (void) printf("Memory that will be " "used after removing %s: %s\n", argv[i], valstr); } } else { if (zpool_vdev_remove(zhp, argv[i]) != 0) ret = 1; } } if (ret == 0 && wait) ret = zpool_wait(zhp, ZPOOL_WAIT_REMOVE); } zpool_close(zhp); return (ret); } /* * Return 1 if a vdev is active (being used in a pool) * Return 0 if a vdev is inactive (offlined or faulted, or not in active pool) * * This is useful for checking if a disk in an active pool is offlined or * faulted. */ static int vdev_is_active(char *vdev_path) { int fd; fd = open(vdev_path, O_EXCL); if (fd < 0) { return (1); /* cant open O_EXCL - disk is active */ } close(fd); return (0); /* disk is inactive in the pool */ } /* * zpool labelclear [-f] * * -f Force clearing the label for the vdevs which are members of * the exported or foreign pools. * * Verifies that the vdev is not active and zeros out the label information * on the device. */ int zpool_do_labelclear(int argc, char **argv) { char vdev[MAXPATHLEN]; char *name = NULL; int c, fd = -1, ret = 0; nvlist_t *config; pool_state_t state; boolean_t inuse = B_FALSE; boolean_t force = B_FALSE; /* check options */ while ((c = getopt(argc, argv, "f")) != -1) { switch (c) { case 'f': force = B_TRUE; break; default: (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* get vdev name */ if (argc < 1) { (void) fprintf(stderr, gettext("missing vdev name\n")); usage(B_FALSE); } if (argc > 1) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } (void) strlcpy(vdev, argv[0], sizeof (vdev)); /* * If we cannot open an absolute path, we quit. * Otherwise if the provided vdev name doesn't point to a file, * try prepending expected disk paths and partition numbers. */ if ((fd = open(vdev, O_RDWR)) < 0) { int error; if (vdev[0] == '/') { (void) fprintf(stderr, gettext("failed to open " "%s: %s\n"), vdev, strerror(errno)); return (1); } error = zfs_resolve_shortname(argv[0], vdev, MAXPATHLEN); if (error == 0 && zfs_dev_is_whole_disk(vdev)) { if (zfs_append_partition(vdev, MAXPATHLEN) == -1) error = ENOENT; } if (error || ((fd = open(vdev, O_RDWR)) < 0)) { if (errno == ENOENT) { (void) fprintf(stderr, gettext( "failed to find device %s, try " "specifying absolute path instead\n"), argv[0]); return (1); } (void) fprintf(stderr, gettext("failed to open %s:" " %s\n"), vdev, strerror(errno)); return (1); } } /* * Flush all dirty pages for the block device. This should not be * fatal when the device does not support BLKFLSBUF as would be the * case for a file vdev. */ if ((zfs_dev_flush(fd) != 0) && (errno != ENOTTY)) (void) fprintf(stderr, gettext("failed to invalidate " "cache for %s: %s\n"), vdev, strerror(errno)); if (zpool_read_label(fd, &config, NULL) != 0) { (void) fprintf(stderr, gettext("failed to read label from %s\n"), vdev); ret = 1; goto errout; } nvlist_free(config); ret = zpool_in_use(g_zfs, fd, &state, &name, &inuse); if (ret != 0) { (void) fprintf(stderr, gettext("failed to check state for %s\n"), vdev); ret = 1; goto errout; } if (!inuse) goto wipe_label; switch (state) { default: case POOL_STATE_ACTIVE: case POOL_STATE_SPARE: case POOL_STATE_L2CACHE: /* * We allow the user to call 'zpool offline -f' * on an offlined disk in an active pool. We can check if * the disk is online by calling vdev_is_active(). */ if (force && !vdev_is_active(vdev)) break; (void) fprintf(stderr, gettext( "%s is a member (%s) of pool \"%s\""), vdev, zpool_pool_state_to_name(state), name); if (force) { (void) fprintf(stderr, gettext( ". Offline the disk first to clear its label.")); } printf("\n"); ret = 1; goto errout; case POOL_STATE_EXPORTED: if (force) break; (void) fprintf(stderr, gettext( "use '-f' to override the following error:\n" "%s is a member of exported pool \"%s\"\n"), vdev, name); ret = 1; goto errout; case POOL_STATE_POTENTIALLY_ACTIVE: if (force) break; (void) fprintf(stderr, gettext( "use '-f' to override the following error:\n" "%s is a member of potentially active pool \"%s\"\n"), vdev, name); ret = 1; goto errout; case POOL_STATE_DESTROYED: /* inuse should never be set for a destroyed pool */ assert(0); break; } wipe_label: ret = zpool_clear_label(fd); if (ret != 0) { (void) fprintf(stderr, gettext("failed to clear label for %s\n"), vdev); } errout: free(name); (void) close(fd); return (ret); } /* * zpool create [-fnd] [-o property=value] ... * [-O file-system-property=value] ... * [-R root] [-m mountpoint] ... * * -f Force creation, even if devices appear in use * -n Do not create the pool, but display the resulting layout if it * were to be created. * -R Create a pool under an alternate root * -m Set default mountpoint for the root dataset. By default it's * '/' * -o Set property=value. * -o Set feature@feature=enabled|disabled. * -d Don't automatically enable all supported pool features * (individual features can be enabled with -o). * -O Set fsproperty=value in the pool's root file system * * Creates the named pool according to the given vdev specification. The * bulk of the vdev processing is done in make_root_vdev() in zpool_vdev.c. * Once we get the nvlist back from make_root_vdev(), we either print out the * contents (if '-n' was specified), or pass it to libzfs to do the creation. */ int zpool_do_create(int argc, char **argv) { boolean_t force = B_FALSE; boolean_t dryrun = B_FALSE; boolean_t enable_pool_features = B_TRUE; int c; nvlist_t *nvroot = NULL; char *poolname; char *tname = NULL; int ret = 1; char *altroot = NULL; char *compat = NULL; char *mountpoint = NULL; nvlist_t *fsprops = NULL; nvlist_t *props = NULL; char *propval; /* check options */ while ((c = getopt(argc, argv, ":fndR:m:o:O:t:")) != -1) { switch (c) { case 'f': force = B_TRUE; break; case 'n': dryrun = B_TRUE; break; case 'd': enable_pool_features = B_FALSE; break; case 'R': altroot = optarg; if (add_prop_list(zpool_prop_to_name( ZPOOL_PROP_ALTROOT), optarg, &props, B_TRUE)) goto errout; if (add_prop_list_default(zpool_prop_to_name( ZPOOL_PROP_CACHEFILE), "none", &props)) goto errout; break; case 'm': /* Equivalent to -O mountpoint=optarg */ mountpoint = optarg; break; case 'o': if ((propval = strchr(optarg, '=')) == NULL) { (void) fprintf(stderr, gettext("missing " "'=' for -o option\n")); goto errout; } *propval = '\0'; propval++; if (add_prop_list(optarg, propval, &props, B_TRUE)) goto errout; /* * If the user is creating a pool that doesn't support * feature flags, don't enable any features. */ if (zpool_name_to_prop(optarg) == ZPOOL_PROP_VERSION) { char *end; u_longlong_t ver; ver = strtoull(propval, &end, 10); if (*end == '\0' && ver < SPA_VERSION_FEATURES) { enable_pool_features = B_FALSE; } } if (zpool_name_to_prop(optarg) == ZPOOL_PROP_ALTROOT) altroot = propval; if (zpool_name_to_prop(optarg) == ZPOOL_PROP_COMPATIBILITY) compat = propval; break; case 'O': if ((propval = strchr(optarg, '=')) == NULL) { (void) fprintf(stderr, gettext("missing " "'=' for -O option\n")); goto errout; } *propval = '\0'; propval++; /* * Mountpoints are checked and then added later. * Uniquely among properties, they can be specified * more than once, to avoid conflict with -m. */ if (0 == strcmp(optarg, zfs_prop_to_name(ZFS_PROP_MOUNTPOINT))) { mountpoint = propval; } else if (add_prop_list(optarg, propval, &fsprops, B_FALSE)) { goto errout; } break; case 't': /* * Sanity check temporary pool name. */ if (strchr(optarg, '/') != NULL) { (void) fprintf(stderr, gettext("cannot create " "'%s': invalid character '/' in temporary " "name\n"), optarg); (void) fprintf(stderr, gettext("use 'zfs " "create' to create a dataset\n")); goto errout; } if (add_prop_list(zpool_prop_to_name( ZPOOL_PROP_TNAME), optarg, &props, B_TRUE)) goto errout; if (add_prop_list_default(zpool_prop_to_name( ZPOOL_PROP_CACHEFILE), "none", &props)) goto errout; tname = optarg; break; case ':': (void) fprintf(stderr, gettext("missing argument for " "'%c' option\n"), optopt); goto badusage; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); goto badusage; } } argc -= optind; argv += optind; /* get pool name and check number of arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name argument\n")); goto badusage; } if (argc < 2) { (void) fprintf(stderr, gettext("missing vdev specification\n")); goto badusage; } poolname = argv[0]; /* * As a special case, check for use of '/' in the name, and direct the * user to use 'zfs create' instead. */ if (strchr(poolname, '/') != NULL) { (void) fprintf(stderr, gettext("cannot create '%s': invalid " "character '/' in pool name\n"), poolname); (void) fprintf(stderr, gettext("use 'zfs create' to " "create a dataset\n")); goto errout; } /* pass off to make_root_vdev for bulk processing */ nvroot = make_root_vdev(NULL, props, force, !force, B_FALSE, dryrun, argc - 1, argv + 1); if (nvroot == NULL) goto errout; /* make_root_vdev() allows 0 toplevel children if there are spares */ if (!zfs_allocatable_devs(nvroot)) { (void) fprintf(stderr, gettext("invalid vdev " "specification: at least one toplevel vdev must be " "specified\n")); goto errout; } if (altroot != NULL && altroot[0] != '/') { (void) fprintf(stderr, gettext("invalid alternate root '%s': " "must be an absolute path\n"), altroot); goto errout; } /* * Check the validity of the mountpoint and direct the user to use the * '-m' mountpoint option if it looks like its in use. */ if (mountpoint == NULL || (strcmp(mountpoint, ZFS_MOUNTPOINT_LEGACY) != 0 && strcmp(mountpoint, ZFS_MOUNTPOINT_NONE) != 0)) { char buf[MAXPATHLEN]; DIR *dirp; if (mountpoint && mountpoint[0] != '/') { (void) fprintf(stderr, gettext("invalid mountpoint " "'%s': must be an absolute path, 'legacy', or " "'none'\n"), mountpoint); goto errout; } if (mountpoint == NULL) { if (altroot != NULL) (void) snprintf(buf, sizeof (buf), "%s/%s", altroot, poolname); else (void) snprintf(buf, sizeof (buf), "/%s", poolname); } else { if (altroot != NULL) (void) snprintf(buf, sizeof (buf), "%s%s", altroot, mountpoint); else (void) snprintf(buf, sizeof (buf), "%s", mountpoint); } if ((dirp = opendir(buf)) == NULL && errno != ENOENT) { (void) fprintf(stderr, gettext("mountpoint '%s' : " "%s\n"), buf, strerror(errno)); (void) fprintf(stderr, gettext("use '-m' " "option to provide a different default\n")); goto errout; } else if (dirp) { int count = 0; while (count < 3 && readdir(dirp) != NULL) count++; (void) closedir(dirp); if (count > 2) { (void) fprintf(stderr, gettext("mountpoint " "'%s' exists and is not empty\n"), buf); (void) fprintf(stderr, gettext("use '-m' " "option to provide a " "different default\n")); goto errout; } } } /* * Now that the mountpoint's validity has been checked, ensure that * the property is set appropriately prior to creating the pool. */ if (mountpoint != NULL) { ret = add_prop_list(zfs_prop_to_name(ZFS_PROP_MOUNTPOINT), mountpoint, &fsprops, B_FALSE); if (ret != 0) goto errout; } ret = 1; if (dryrun) { /* * For a dry run invocation, print out a basic message and run * through all the vdevs in the list and print out in an * appropriate hierarchy. */ (void) printf(gettext("would create '%s' with the " "following layout:\n\n"), poolname); print_vdev_tree(NULL, poolname, nvroot, 0, "", 0); print_vdev_tree(NULL, "dedup", nvroot, 0, VDEV_ALLOC_BIAS_DEDUP, 0); print_vdev_tree(NULL, "special", nvroot, 0, VDEV_ALLOC_BIAS_SPECIAL, 0); print_vdev_tree(NULL, "logs", nvroot, 0, VDEV_ALLOC_BIAS_LOG, 0); print_cache_list(nvroot, 0); print_spare_list(nvroot, 0); ret = 0; } else { /* * Load in feature set. * Note: if compatibility property not given, we'll have * NULL, which means 'all features'. */ boolean_t requested_features[SPA_FEATURES]; if (zpool_do_load_compat(compat, requested_features) != ZPOOL_COMPATIBILITY_OK) goto errout; /* * props contains list of features to enable. * For each feature: * - remove it if feature@name=disabled * - leave it there if feature@name=enabled * - add it if: * - enable_pool_features (ie: no '-d' or '-o version') * - it's supported by the kernel module * - it's in the requested feature set * - warn if it's enabled but not in compat */ for (spa_feature_t i = 0; i < SPA_FEATURES; i++) { char propname[MAXPATHLEN]; const char *propval; zfeature_info_t *feat = &spa_feature_table[i]; (void) snprintf(propname, sizeof (propname), "feature@%s", feat->fi_uname); if (!nvlist_lookup_string(props, propname, &propval)) { if (strcmp(propval, ZFS_FEATURE_DISABLED) == 0) { (void) nvlist_remove_all(props, propname); } else if (strcmp(propval, ZFS_FEATURE_ENABLED) == 0 && !requested_features[i]) { (void) fprintf(stderr, gettext( "Warning: feature \"%s\" enabled " "but is not in specified " "'compatibility' feature set.\n"), feat->fi_uname); } } else if ( enable_pool_features && feat->fi_zfs_mod_supported && requested_features[i]) { ret = add_prop_list(propname, ZFS_FEATURE_ENABLED, &props, B_TRUE); if (ret != 0) goto errout; } } ret = 1; if (zpool_create(g_zfs, poolname, nvroot, props, fsprops) == 0) { zfs_handle_t *pool = zfs_open(g_zfs, tname ? tname : poolname, ZFS_TYPE_FILESYSTEM); if (pool != NULL) { if (zfs_mount(pool, NULL, 0) == 0) { ret = zfs_share(pool, NULL); zfs_commit_shares(NULL); } zfs_close(pool); } } else if (libzfs_errno(g_zfs) == EZFS_INVALIDNAME) { (void) fprintf(stderr, gettext("pool name may have " "been omitted\n")); } } errout: nvlist_free(nvroot); nvlist_free(fsprops); nvlist_free(props); return (ret); badusage: nvlist_free(fsprops); nvlist_free(props); usage(B_FALSE); return (2); } /* * zpool destroy * * -f Forcefully unmount any datasets * * Destroy the given pool. Automatically unmounts any datasets in the pool. */ int zpool_do_destroy(int argc, char **argv) { boolean_t force = B_FALSE; int c; char *pool; zpool_handle_t *zhp; int ret; /* check options */ while ((c = getopt(argc, argv, "f")) != -1) { switch (c) { case 'f': force = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* check arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool argument\n")); usage(B_FALSE); } if (argc > 1) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } pool = argv[0]; if ((zhp = zpool_open_canfail(g_zfs, pool)) == NULL) { /* * As a special case, check for use of '/' in the name, and * direct the user to use 'zfs destroy' instead. */ if (strchr(pool, '/') != NULL) (void) fprintf(stderr, gettext("use 'zfs destroy' to " "destroy a dataset\n")); return (1); } if (zpool_disable_datasets(zhp, force) != 0) { (void) fprintf(stderr, gettext("could not destroy '%s': " "could not unmount datasets\n"), zpool_get_name(zhp)); zpool_close(zhp); return (1); } /* The history must be logged as part of the export */ log_history = B_FALSE; ret = (zpool_destroy(zhp, history_str) != 0); zpool_close(zhp); return (ret); } typedef struct export_cbdata { boolean_t force; boolean_t hardforce; } export_cbdata_t; /* * Export one pool */ static int zpool_export_one(zpool_handle_t *zhp, void *data) { export_cbdata_t *cb = data; if (zpool_disable_datasets(zhp, cb->force) != 0) return (1); /* The history must be logged as part of the export */ log_history = B_FALSE; if (cb->hardforce) { if (zpool_export_force(zhp, history_str) != 0) return (1); } else if (zpool_export(zhp, cb->force, history_str) != 0) { return (1); } return (0); } /* * zpool export [-f] ... * * -a Export all pools * -f Forcefully unmount datasets * * Export the given pools. By default, the command will attempt to cleanly * unmount any active datasets within the pool. If the '-f' flag is specified, * then the datasets will be forcefully unmounted. */ int zpool_do_export(int argc, char **argv) { export_cbdata_t cb; boolean_t do_all = B_FALSE; boolean_t force = B_FALSE; boolean_t hardforce = B_FALSE; int c, ret; /* check options */ while ((c = getopt(argc, argv, "afF")) != -1) { switch (c) { case 'a': do_all = B_TRUE; break; case 'f': force = B_TRUE; break; case 'F': hardforce = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } cb.force = force; cb.hardforce = hardforce; argc -= optind; argv += optind; if (do_all) { if (argc != 0) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } return (for_each_pool(argc, argv, B_TRUE, NULL, ZFS_TYPE_POOL, B_FALSE, zpool_export_one, &cb)); } /* check arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool argument\n")); usage(B_FALSE); } ret = for_each_pool(argc, argv, B_TRUE, NULL, ZFS_TYPE_POOL, B_FALSE, zpool_export_one, &cb); return (ret); } /* * Given a vdev configuration, determine the maximum width needed for the device * name column. */ static int max_width(zpool_handle_t *zhp, nvlist_t *nv, int depth, int max, int name_flags) { static const char *const subtypes[] = {ZPOOL_CONFIG_SPARES, ZPOOL_CONFIG_L2CACHE, ZPOOL_CONFIG_CHILDREN}; char *name = zpool_vdev_name(g_zfs, zhp, nv, name_flags); max = MAX(strlen(name) + depth, max); free(name); nvlist_t **child; uint_t children; for (size_t i = 0; i < ARRAY_SIZE(subtypes); ++i) if (nvlist_lookup_nvlist_array(nv, subtypes[i], &child, &children) == 0) for (uint_t c = 0; c < children; ++c) max = MAX(max_width(zhp, child[c], depth + 2, max, name_flags), max); return (max); } typedef struct spare_cbdata { uint64_t cb_guid; zpool_handle_t *cb_zhp; } spare_cbdata_t; static boolean_t find_vdev(nvlist_t *nv, uint64_t search) { uint64_t guid; nvlist_t **child; uint_t c, children; if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) == 0 && search == guid) return (B_TRUE); if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) == 0) { for (c = 0; c < children; c++) if (find_vdev(child[c], search)) return (B_TRUE); } return (B_FALSE); } static int find_spare(zpool_handle_t *zhp, void *data) { spare_cbdata_t *cbp = data; nvlist_t *config, *nvroot; config = zpool_get_config(zhp, NULL); verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); if (find_vdev(nvroot, cbp->cb_guid)) { cbp->cb_zhp = zhp; return (1); } zpool_close(zhp); return (0); } typedef struct status_cbdata { int cb_count; int cb_name_flags; int cb_namewidth; boolean_t cb_allpools; boolean_t cb_verbose; boolean_t cb_literal; boolean_t cb_explain; boolean_t cb_first; boolean_t cb_dedup_stats; boolean_t cb_print_status; boolean_t cb_print_slow_ios; boolean_t cb_print_vdev_init; boolean_t cb_print_vdev_trim; vdev_cmd_data_list_t *vcdl; } status_cbdata_t; /* Return 1 if string is NULL, empty, or whitespace; return 0 otherwise. */ static boolean_t is_blank_str(const char *str) { for (; str != NULL && *str != '\0'; ++str) if (!isblank(*str)) return (B_FALSE); return (B_TRUE); } /* Print command output lines for specific vdev in a specific pool */ static void zpool_print_cmd(vdev_cmd_data_list_t *vcdl, const char *pool, const char *path) { vdev_cmd_data_t *data; int i, j; const char *val; for (i = 0; i < vcdl->count; i++) { if ((strcmp(vcdl->data[i].path, path) != 0) || (strcmp(vcdl->data[i].pool, pool) != 0)) { /* Not the vdev we're looking for */ continue; } data = &vcdl->data[i]; /* Print out all the output values for this vdev */ for (j = 0; j < vcdl->uniq_cols_cnt; j++) { val = NULL; /* Does this vdev have values for this column? */ for (int k = 0; k < data->cols_cnt; k++) { if (strcmp(data->cols[k], vcdl->uniq_cols[j]) == 0) { /* yes it does, record the value */ val = data->lines[k]; break; } } /* * Mark empty values with dashes to make output * awk-able. */ if (val == NULL || is_blank_str(val)) val = "-"; printf("%*s", vcdl->uniq_cols_width[j], val); if (j < vcdl->uniq_cols_cnt - 1) fputs(" ", stdout); } /* Print out any values that aren't in a column at the end */ for (j = data->cols_cnt; j < data->lines_cnt; j++) { /* Did we have any columns? If so print a spacer. */ if (vcdl->uniq_cols_cnt > 0) fputs(" ", stdout); val = data->lines[j]; fputs(val ?: "", stdout); } break; } } /* * Print vdev initialization status for leaves */ static void print_status_initialize(vdev_stat_t *vs, boolean_t verbose) { if (verbose) { if ((vs->vs_initialize_state == VDEV_INITIALIZE_ACTIVE || vs->vs_initialize_state == VDEV_INITIALIZE_SUSPENDED || vs->vs_initialize_state == VDEV_INITIALIZE_COMPLETE) && !vs->vs_scan_removing) { char zbuf[1024]; char tbuf[256]; struct tm zaction_ts; time_t t = vs->vs_initialize_action_time; int initialize_pct = 100; if (vs->vs_initialize_state != VDEV_INITIALIZE_COMPLETE) { initialize_pct = (vs->vs_initialize_bytes_done * 100 / (vs->vs_initialize_bytes_est + 1)); } (void) localtime_r(&t, &zaction_ts); (void) strftime(tbuf, sizeof (tbuf), "%c", &zaction_ts); switch (vs->vs_initialize_state) { case VDEV_INITIALIZE_SUSPENDED: (void) snprintf(zbuf, sizeof (zbuf), ", %s %s", gettext("suspended, started at"), tbuf); break; case VDEV_INITIALIZE_ACTIVE: (void) snprintf(zbuf, sizeof (zbuf), ", %s %s", gettext("started at"), tbuf); break; case VDEV_INITIALIZE_COMPLETE: (void) snprintf(zbuf, sizeof (zbuf), ", %s %s", gettext("completed at"), tbuf); break; } (void) printf(gettext(" (%d%% initialized%s)"), initialize_pct, zbuf); } else { (void) printf(gettext(" (uninitialized)")); } } else if (vs->vs_initialize_state == VDEV_INITIALIZE_ACTIVE) { (void) printf(gettext(" (initializing)")); } } /* * Print vdev TRIM status for leaves */ static void print_status_trim(vdev_stat_t *vs, boolean_t verbose) { if (verbose) { if ((vs->vs_trim_state == VDEV_TRIM_ACTIVE || vs->vs_trim_state == VDEV_TRIM_SUSPENDED || vs->vs_trim_state == VDEV_TRIM_COMPLETE) && !vs->vs_scan_removing) { char zbuf[1024]; char tbuf[256]; struct tm zaction_ts; time_t t = vs->vs_trim_action_time; int trim_pct = 100; if (vs->vs_trim_state != VDEV_TRIM_COMPLETE) { trim_pct = (vs->vs_trim_bytes_done * 100 / (vs->vs_trim_bytes_est + 1)); } (void) localtime_r(&t, &zaction_ts); (void) strftime(tbuf, sizeof (tbuf), "%c", &zaction_ts); switch (vs->vs_trim_state) { case VDEV_TRIM_SUSPENDED: (void) snprintf(zbuf, sizeof (zbuf), ", %s %s", gettext("suspended, started at"), tbuf); break; case VDEV_TRIM_ACTIVE: (void) snprintf(zbuf, sizeof (zbuf), ", %s %s", gettext("started at"), tbuf); break; case VDEV_TRIM_COMPLETE: (void) snprintf(zbuf, sizeof (zbuf), ", %s %s", gettext("completed at"), tbuf); break; } (void) printf(gettext(" (%d%% trimmed%s)"), trim_pct, zbuf); } else if (vs->vs_trim_notsup) { (void) printf(gettext(" (trim unsupported)")); } else { (void) printf(gettext(" (untrimmed)")); } } else if (vs->vs_trim_state == VDEV_TRIM_ACTIVE) { (void) printf(gettext(" (trimming)")); } } /* * Return the color associated with a health string. This includes returning * NULL for no color change. */ static const char * health_str_to_color(const char *health) { if (strcmp(health, gettext("FAULTED")) == 0 || strcmp(health, gettext("SUSPENDED")) == 0 || strcmp(health, gettext("UNAVAIL")) == 0) { return (ANSI_RED); } if (strcmp(health, gettext("OFFLINE")) == 0 || strcmp(health, gettext("DEGRADED")) == 0 || strcmp(health, gettext("REMOVED")) == 0) { return (ANSI_YELLOW); } return (NULL); } /* * Print out configuration state as requested by status_callback. */ static void print_status_config(zpool_handle_t *zhp, status_cbdata_t *cb, const char *name, nvlist_t *nv, int depth, boolean_t isspare, vdev_rebuild_stat_t *vrs) { nvlist_t **child, *root; uint_t c, i, vsc, children; pool_scan_stat_t *ps = NULL; vdev_stat_t *vs; char rbuf[6], wbuf[6], cbuf[6]; char *vname; uint64_t notpresent; spare_cbdata_t spare_cb; const char *state; const char *type; const char *path = NULL; const char *rcolor = NULL, *wcolor = NULL, *ccolor = NULL; if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) children = 0; verify(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &vsc) == 0); verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &type) == 0); if (strcmp(type, VDEV_TYPE_INDIRECT) == 0) return; state = zpool_state_to_name(vs->vs_state, vs->vs_aux); if (isspare) { /* * For hot spares, we use the terms 'INUSE' and 'AVAILABLE' for * online drives. */ if (vs->vs_aux == VDEV_AUX_SPARED) state = gettext("INUSE"); else if (vs->vs_state == VDEV_STATE_HEALTHY) state = gettext("AVAIL"); } printf_color(health_str_to_color(state), "\t%*s%-*s %-8s", depth, "", cb->cb_namewidth - depth, name, state); if (!isspare) { if (vs->vs_read_errors) rcolor = ANSI_RED; if (vs->vs_write_errors) wcolor = ANSI_RED; if (vs->vs_checksum_errors) ccolor = ANSI_RED; if (cb->cb_literal) { fputc(' ', stdout); printf_color(rcolor, "%5llu", (u_longlong_t)vs->vs_read_errors); fputc(' ', stdout); printf_color(wcolor, "%5llu", (u_longlong_t)vs->vs_write_errors); fputc(' ', stdout); printf_color(ccolor, "%5llu", (u_longlong_t)vs->vs_checksum_errors); } else { zfs_nicenum(vs->vs_read_errors, rbuf, sizeof (rbuf)); zfs_nicenum(vs->vs_write_errors, wbuf, sizeof (wbuf)); zfs_nicenum(vs->vs_checksum_errors, cbuf, sizeof (cbuf)); fputc(' ', stdout); printf_color(rcolor, "%5s", rbuf); fputc(' ', stdout); printf_color(wcolor, "%5s", wbuf); fputc(' ', stdout); printf_color(ccolor, "%5s", cbuf); } if (cb->cb_print_slow_ios) { if (children == 0) { /* Only leafs vdevs have slow IOs */ zfs_nicenum(vs->vs_slow_ios, rbuf, sizeof (rbuf)); } else { snprintf(rbuf, sizeof (rbuf), "-"); } if (cb->cb_literal) printf(" %5llu", (u_longlong_t)vs->vs_slow_ios); else printf(" %5s", rbuf); } } if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT, ¬present) == 0) { verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) == 0); (void) printf(" %s %s", gettext("was"), path); } else if (vs->vs_aux != 0) { (void) printf(" "); color_start(ANSI_RED); switch (vs->vs_aux) { case VDEV_AUX_OPEN_FAILED: (void) printf(gettext("cannot open")); break; case VDEV_AUX_BAD_GUID_SUM: (void) printf(gettext("missing device")); break; case VDEV_AUX_NO_REPLICAS: (void) printf(gettext("insufficient replicas")); break; case VDEV_AUX_VERSION_NEWER: (void) printf(gettext("newer version")); break; case VDEV_AUX_UNSUP_FEAT: (void) printf(gettext("unsupported feature(s)")); break; case VDEV_AUX_ASHIFT_TOO_BIG: (void) printf(gettext("unsupported minimum blocksize")); break; case VDEV_AUX_SPARED: verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &spare_cb.cb_guid) == 0); if (zpool_iter(g_zfs, find_spare, &spare_cb) == 1) { if (strcmp(zpool_get_name(spare_cb.cb_zhp), zpool_get_name(zhp)) == 0) (void) printf(gettext("currently in " "use")); else (void) printf(gettext("in use by " "pool '%s'"), zpool_get_name(spare_cb.cb_zhp)); zpool_close(spare_cb.cb_zhp); } else { (void) printf(gettext("currently in use")); } break; case VDEV_AUX_ERR_EXCEEDED: (void) printf(gettext("too many errors")); break; case VDEV_AUX_IO_FAILURE: (void) printf(gettext("experienced I/O failures")); break; case VDEV_AUX_BAD_LOG: (void) printf(gettext("bad intent log")); break; case VDEV_AUX_EXTERNAL: (void) printf(gettext("external device fault")); break; case VDEV_AUX_SPLIT_POOL: (void) printf(gettext("split into new pool")); break; case VDEV_AUX_ACTIVE: (void) printf(gettext("currently in use")); break; case VDEV_AUX_CHILDREN_OFFLINE: (void) printf(gettext("all children offline")); break; case VDEV_AUX_BAD_LABEL: (void) printf(gettext("invalid label")); break; default: (void) printf(gettext("corrupted data")); break; } color_end(); } else if (children == 0 && !isspare && getenv("ZPOOL_STATUS_NON_NATIVE_ASHIFT_IGNORE") == NULL && VDEV_STAT_VALID(vs_physical_ashift, vsc) && vs->vs_configured_ashift < vs->vs_physical_ashift) { (void) printf( gettext(" block size: %dB configured, %dB native"), 1 << vs->vs_configured_ashift, 1 << vs->vs_physical_ashift); } if (vs->vs_scan_removing != 0) { (void) printf(gettext(" (removing)")); } else if (VDEV_STAT_VALID(vs_noalloc, vsc) && vs->vs_noalloc != 0) { (void) printf(gettext(" (non-allocating)")); } /* The root vdev has the scrub/resilver stats */ root = fnvlist_lookup_nvlist(zpool_get_config(zhp, NULL), ZPOOL_CONFIG_VDEV_TREE); (void) nvlist_lookup_uint64_array(root, ZPOOL_CONFIG_SCAN_STATS, (uint64_t **)&ps, &c); /* * If you force fault a drive that's resilvering, its scan stats can * get frozen in time, giving the false impression that it's * being resilvered. That's why we check the state to see if the vdev * is healthy before reporting "resilvering" or "repairing". */ if (ps != NULL && ps->pss_state == DSS_SCANNING && children == 0 && vs->vs_state == VDEV_STATE_HEALTHY) { if (vs->vs_scan_processed != 0) { (void) printf(gettext(" (%s)"), (ps->pss_func == POOL_SCAN_RESILVER) ? "resilvering" : "repairing"); } else if (vs->vs_resilver_deferred) { (void) printf(gettext(" (awaiting resilver)")); } } /* The top-level vdevs have the rebuild stats */ if (vrs != NULL && vrs->vrs_state == VDEV_REBUILD_ACTIVE && children == 0 && vs->vs_state == VDEV_STATE_HEALTHY) { if (vs->vs_rebuild_processed != 0) { (void) printf(gettext(" (resilvering)")); } } if (cb->vcdl != NULL) { if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) == 0) { printf(" "); zpool_print_cmd(cb->vcdl, zpool_get_name(zhp), path); } } /* Display vdev initialization and trim status for leaves. */ if (children == 0) { print_status_initialize(vs, cb->cb_print_vdev_init); print_status_trim(vs, cb->cb_print_vdev_trim); } (void) printf("\n"); for (c = 0; c < children; c++) { uint64_t islog = B_FALSE, ishole = B_FALSE; /* Don't print logs or holes here */ (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, &islog); (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_HOLE, &ishole); if (islog || ishole) continue; /* Only print normal classes here */ if (nvlist_exists(child[c], ZPOOL_CONFIG_ALLOCATION_BIAS)) continue; /* Provide vdev_rebuild_stats to children if available */ if (vrs == NULL) { (void) nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_REBUILD_STATS, (uint64_t **)&vrs, &i); } vname = zpool_vdev_name(g_zfs, zhp, child[c], cb->cb_name_flags | VDEV_NAME_TYPE_ID); print_status_config(zhp, cb, vname, child[c], depth + 2, isspare, vrs); free(vname); } } /* * Print the configuration of an exported pool. Iterate over all vdevs in the * pool, printing out the name and status for each one. */ static void print_import_config(status_cbdata_t *cb, const char *name, nvlist_t *nv, int depth) { nvlist_t **child; uint_t c, children; vdev_stat_t *vs; const char *type; char *vname; verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &type) == 0); if (strcmp(type, VDEV_TYPE_MISSING) == 0 || strcmp(type, VDEV_TYPE_HOLE) == 0) return; verify(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &c) == 0); (void) printf("\t%*s%-*s", depth, "", cb->cb_namewidth - depth, name); (void) printf(" %s", zpool_state_to_name(vs->vs_state, vs->vs_aux)); if (vs->vs_aux != 0) { (void) printf(" "); switch (vs->vs_aux) { case VDEV_AUX_OPEN_FAILED: (void) printf(gettext("cannot open")); break; case VDEV_AUX_BAD_GUID_SUM: (void) printf(gettext("missing device")); break; case VDEV_AUX_NO_REPLICAS: (void) printf(gettext("insufficient replicas")); break; case VDEV_AUX_VERSION_NEWER: (void) printf(gettext("newer version")); break; case VDEV_AUX_UNSUP_FEAT: (void) printf(gettext("unsupported feature(s)")); break; case VDEV_AUX_ERR_EXCEEDED: (void) printf(gettext("too many errors")); break; case VDEV_AUX_ACTIVE: (void) printf(gettext("currently in use")); break; case VDEV_AUX_CHILDREN_OFFLINE: (void) printf(gettext("all children offline")); break; case VDEV_AUX_BAD_LABEL: (void) printf(gettext("invalid label")); break; default: (void) printf(gettext("corrupted data")); break; } } (void) printf("\n"); if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) return; for (c = 0; c < children; c++) { uint64_t is_log = B_FALSE; (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, &is_log); if (is_log) continue; if (nvlist_exists(child[c], ZPOOL_CONFIG_ALLOCATION_BIAS)) continue; vname = zpool_vdev_name(g_zfs, NULL, child[c], cb->cb_name_flags | VDEV_NAME_TYPE_ID); print_import_config(cb, vname, child[c], depth + 2); free(vname); } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE, &child, &children) == 0) { (void) printf(gettext("\tcache\n")); for (c = 0; c < children; c++) { vname = zpool_vdev_name(g_zfs, NULL, child[c], cb->cb_name_flags); (void) printf("\t %s\n", vname); free(vname); } } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES, &child, &children) == 0) { (void) printf(gettext("\tspares\n")); for (c = 0; c < children; c++) { vname = zpool_vdev_name(g_zfs, NULL, child[c], cb->cb_name_flags); (void) printf("\t %s\n", vname); free(vname); } } } /* * Print specialized class vdevs. * * These are recorded as top level vdevs in the main pool child array * but with "is_log" set to 1 or an "alloc_bias" string. We use either * print_status_config() or print_import_config() to print the top level * class vdevs then any of their children (eg mirrored slogs) are printed * recursively - which works because only the top level vdev is marked. */ static void print_class_vdevs(zpool_handle_t *zhp, status_cbdata_t *cb, nvlist_t *nv, const char *class) { uint_t c, children; nvlist_t **child; boolean_t printed = B_FALSE; assert(zhp != NULL || !cb->cb_verbose); if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) return; for (c = 0; c < children; c++) { uint64_t is_log = B_FALSE; const char *bias = NULL; const char *type = NULL; (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, &is_log); if (is_log) { bias = (char *)VDEV_ALLOC_CLASS_LOGS; } else { (void) nvlist_lookup_string(child[c], ZPOOL_CONFIG_ALLOCATION_BIAS, &bias); (void) nvlist_lookup_string(child[c], ZPOOL_CONFIG_TYPE, &type); } if (bias == NULL || strcmp(bias, class) != 0) continue; if (!is_log && strcmp(type, VDEV_TYPE_INDIRECT) == 0) continue; if (!printed) { (void) printf("\t%s\t\n", gettext(class)); printed = B_TRUE; } char *name = zpool_vdev_name(g_zfs, zhp, child[c], cb->cb_name_flags | VDEV_NAME_TYPE_ID); if (cb->cb_print_status) print_status_config(zhp, cb, name, child[c], 2, B_FALSE, NULL); else print_import_config(cb, name, child[c], 2); free(name); } } /* * Display the status for the given pool. */ static int show_import(nvlist_t *config, boolean_t report_error) { uint64_t pool_state; vdev_stat_t *vs; const char *name; uint64_t guid; uint64_t hostid = 0; const char *msgid; const char *hostname = "unknown"; nvlist_t *nvroot, *nvinfo; zpool_status_t reason; zpool_errata_t errata; const char *health; uint_t vsc; const char *comment; status_cbdata_t cb = { 0 }; verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME, &name) == 0); verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID, &guid) == 0); verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE, &pool_state) == 0); verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &vsc) == 0); health = zpool_state_to_name(vs->vs_state, vs->vs_aux); reason = zpool_import_status(config, &msgid, &errata); /* * If we're importing using a cachefile, then we won't report any * errors unless we are in the scan phase of the import. */ if (reason != ZPOOL_STATUS_OK && !report_error) return (reason); (void) printf(gettext(" pool: %s\n"), name); (void) printf(gettext(" id: %llu\n"), (u_longlong_t)guid); (void) printf(gettext(" state: %s"), health); if (pool_state == POOL_STATE_DESTROYED) (void) printf(gettext(" (DESTROYED)")); (void) printf("\n"); switch (reason) { case ZPOOL_STATUS_MISSING_DEV_R: case ZPOOL_STATUS_MISSING_DEV_NR: case ZPOOL_STATUS_BAD_GUID_SUM: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices are " "missing from the system.\n")); break; case ZPOOL_STATUS_CORRUPT_LABEL_R: case ZPOOL_STATUS_CORRUPT_LABEL_NR: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices contains" " corrupted data.\n")); break; case ZPOOL_STATUS_CORRUPT_DATA: (void) printf( gettext(" status: The pool data is corrupted.\n")); break; case ZPOOL_STATUS_OFFLINE_DEV: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices " "are offlined.\n")); break; case ZPOOL_STATUS_CORRUPT_POOL: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool metadata is " "corrupted.\n")); break; case ZPOOL_STATUS_VERSION_OLDER: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool is formatted using " "a legacy on-disk version.\n")); break; case ZPOOL_STATUS_VERSION_NEWER: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool is formatted using " "an incompatible version.\n")); break; case ZPOOL_STATUS_FEAT_DISABLED: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("Some supported " "features are not enabled on the pool.\n\t" "(Note that they may be intentionally disabled " "if the\n\t'compatibility' property is set.)\n")); break; case ZPOOL_STATUS_COMPATIBILITY_ERR: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("Error reading or parsing " "the file(s) indicated by the 'compatibility'\n" "property.\n")); break; case ZPOOL_STATUS_INCOMPATIBLE_FEAT: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more features " "are enabled on the pool despite not being\n" "requested by the 'compatibility' property.\n")); break; case ZPOOL_STATUS_UNSUP_FEAT_READ: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool uses the following " "feature(s) not supported on this system:\n")); color_start(ANSI_YELLOW); zpool_print_unsup_feat(config); color_end(); break; case ZPOOL_STATUS_UNSUP_FEAT_WRITE: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool can only be " "accessed in read-only mode on this system. It\n\tcannot be" " accessed in read-write mode because it uses the " "following\n\tfeature(s) not supported on this system:\n")); color_start(ANSI_YELLOW); zpool_print_unsup_feat(config); color_end(); break; case ZPOOL_STATUS_HOSTID_ACTIVE: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool is currently " "imported by another system.\n")); break; case ZPOOL_STATUS_HOSTID_REQUIRED: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool has the " "multihost property on. It cannot\n\tbe safely imported " "when the system hostid is not set.\n")); break; case ZPOOL_STATUS_HOSTID_MISMATCH: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool was last accessed " "by another system.\n")); break; case ZPOOL_STATUS_FAULTED_DEV_R: case ZPOOL_STATUS_FAULTED_DEV_NR: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices are " "faulted.\n")); break; case ZPOOL_STATUS_BAD_LOG: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("An intent log record cannot " "be read.\n")); break; case ZPOOL_STATUS_RESILVERING: case ZPOOL_STATUS_REBUILDING: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices were " "being resilvered.\n")); break; case ZPOOL_STATUS_ERRATA: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("Errata #%d detected.\n"), errata); break; case ZPOOL_STATUS_NON_NATIVE_ASHIFT: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices are " "configured to use a non-native block size.\n" "\tExpect reduced performance.\n")); break; default: /* * No other status can be seen when importing pools. */ assert(reason == ZPOOL_STATUS_OK); } /* * Print out an action according to the overall state of the pool. */ if (vs->vs_state == VDEV_STATE_HEALTHY) { if (reason == ZPOOL_STATUS_VERSION_OLDER || reason == ZPOOL_STATUS_FEAT_DISABLED) { (void) printf(gettext(" action: The pool can be " "imported using its name or numeric identifier, " "though\n\tsome features will not be available " "without an explicit 'zpool upgrade'.\n")); } else if (reason == ZPOOL_STATUS_COMPATIBILITY_ERR) { (void) printf(gettext(" action: The pool can be " "imported using its name or numeric\n\tidentifier, " "though the file(s) indicated by its " "'compatibility'\n\tproperty cannot be parsed at " "this time.\n")); } else if (reason == ZPOOL_STATUS_HOSTID_MISMATCH) { (void) printf(gettext(" action: The pool can be " "imported using its name or numeric " "identifier and\n\tthe '-f' flag.\n")); } else if (reason == ZPOOL_STATUS_ERRATA) { switch (errata) { case ZPOOL_ERRATA_NONE: break; case ZPOOL_ERRATA_ZOL_2094_SCRUB: (void) printf(gettext(" action: The pool can " "be imported using its name or numeric " "identifier,\n\thowever there is a compat" "ibility issue which should be corrected" "\n\tby running 'zpool scrub'\n")); break; case ZPOOL_ERRATA_ZOL_2094_ASYNC_DESTROY: (void) printf(gettext(" action: The pool can" "not be imported with this version of ZFS " "due to\n\tan active asynchronous destroy. " "Revert to an earlier version\n\tand " "allow the destroy to complete before " "updating.\n")); break; case ZPOOL_ERRATA_ZOL_6845_ENCRYPTION: (void) printf(gettext(" action: Existing " "encrypted datasets contain an on-disk " "incompatibility, which\n\tneeds to be " "corrected. Backup these datasets to new " "encrypted datasets\n\tand destroy the " "old ones.\n")); break; case ZPOOL_ERRATA_ZOL_8308_ENCRYPTION: (void) printf(gettext(" action: Existing " "encrypted snapshots and bookmarks contain " "an on-disk\n\tincompatibility. This may " "cause on-disk corruption if they are used" "\n\twith 'zfs recv'. To correct the " "issue, enable the bookmark_v2 feature.\n\t" "No additional action is needed if there " "are no encrypted snapshots or\n\t" "bookmarks. If preserving the encrypted " "snapshots and bookmarks is\n\trequired, " "use a non-raw send to backup and restore " "them. Alternately,\n\tthey may be removed" " to resolve the incompatibility.\n")); break; default: /* * All errata must contain an action message. */ assert(0); } } else { (void) printf(gettext(" action: The pool can be " "imported using its name or numeric " "identifier.\n")); } } else if (vs->vs_state == VDEV_STATE_DEGRADED) { (void) printf(gettext(" action: The pool can be imported " "despite missing or damaged devices. The\n\tfault " "tolerance of the pool may be compromised if imported.\n")); } else { switch (reason) { case ZPOOL_STATUS_VERSION_NEWER: (void) printf(gettext(" action: The pool cannot be " "imported. Access the pool on a system running " "newer\n\tsoftware, or recreate the pool from " "backup.\n")); break; case ZPOOL_STATUS_UNSUP_FEAT_READ: printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("The pool cannot be " "imported. Access the pool on a system that " "supports\n\tthe required feature(s), or recreate " "the pool from backup.\n")); break; case ZPOOL_STATUS_UNSUP_FEAT_WRITE: printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("The pool cannot be " "imported in read-write mode. Import the pool " "with\n" "\t\"-o readonly=on\", access the pool on a system " "that supports the\n\trequired feature(s), or " "recreate the pool from backup.\n")); break; case ZPOOL_STATUS_MISSING_DEV_R: case ZPOOL_STATUS_MISSING_DEV_NR: case ZPOOL_STATUS_BAD_GUID_SUM: (void) printf(gettext(" action: The pool cannot be " "imported. Attach the missing\n\tdevices and try " "again.\n")); break; case ZPOOL_STATUS_HOSTID_ACTIVE: VERIFY0(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO, &nvinfo)); if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_HOSTNAME)) hostname = fnvlist_lookup_string(nvinfo, ZPOOL_CONFIG_MMP_HOSTNAME); if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_HOSTID)) hostid = fnvlist_lookup_uint64(nvinfo, ZPOOL_CONFIG_MMP_HOSTID); (void) printf(gettext(" action: The pool must be " "exported from %s (hostid=%"PRIx64")\n\tbefore it " "can be safely imported.\n"), hostname, hostid); break; case ZPOOL_STATUS_HOSTID_REQUIRED: (void) printf(gettext(" action: Set a unique system " "hostid with the zgenhostid(8) command.\n")); break; default: (void) printf(gettext(" action: The pool cannot be " "imported due to damaged devices or data.\n")); } } /* Print the comment attached to the pool. */ if (nvlist_lookup_string(config, ZPOOL_CONFIG_COMMENT, &comment) == 0) (void) printf(gettext("comment: %s\n"), comment); /* * If the state is "closed" or "can't open", and the aux state * is "corrupt data": */ if (((vs->vs_state == VDEV_STATE_CLOSED) || (vs->vs_state == VDEV_STATE_CANT_OPEN)) && (vs->vs_aux == VDEV_AUX_CORRUPT_DATA)) { if (pool_state == POOL_STATE_DESTROYED) (void) printf(gettext("\tThe pool was destroyed, " "but can be imported using the '-Df' flags.\n")); else if (pool_state != POOL_STATE_EXPORTED) (void) printf(gettext("\tThe pool may be active on " "another system, but can be imported using\n\t" "the '-f' flag.\n")); } if (msgid != NULL) { (void) printf(gettext( " see: https://openzfs.github.io/openzfs-docs/msg/%s\n"), msgid); } (void) printf(gettext(" config:\n\n")); cb.cb_namewidth = max_width(NULL, nvroot, 0, strlen(name), VDEV_NAME_TYPE_ID); if (cb.cb_namewidth < 10) cb.cb_namewidth = 10; print_import_config(&cb, name, nvroot, 0); print_class_vdevs(NULL, &cb, nvroot, VDEV_ALLOC_BIAS_DEDUP); print_class_vdevs(NULL, &cb, nvroot, VDEV_ALLOC_BIAS_SPECIAL); print_class_vdevs(NULL, &cb, nvroot, VDEV_ALLOC_CLASS_LOGS); if (reason == ZPOOL_STATUS_BAD_GUID_SUM) { (void) printf(gettext("\n\tAdditional devices are known to " "be part of this pool, though their\n\texact " "configuration cannot be determined.\n")); } return (0); } static boolean_t zfs_force_import_required(nvlist_t *config) { uint64_t state; uint64_t hostid = 0; nvlist_t *nvinfo; state = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE); (void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_HOSTID, &hostid); if (state != POOL_STATE_EXPORTED && hostid != get_system_hostid()) return (B_TRUE); nvinfo = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO); if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_STATE)) { mmp_state_t mmp_state = fnvlist_lookup_uint64(nvinfo, ZPOOL_CONFIG_MMP_STATE); if (mmp_state != MMP_STATE_INACTIVE) return (B_TRUE); } return (B_FALSE); } /* * Perform the import for the given configuration. This passes the heavy * lifting off to zpool_import_props(), and then mounts the datasets contained * within the pool. */ static int do_import(nvlist_t *config, const char *newname, const char *mntopts, nvlist_t *props, int flags) { int ret = 0; zpool_handle_t *zhp; const char *name; uint64_t version; name = fnvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME); version = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION); if (!SPA_VERSION_IS_SUPPORTED(version)) { (void) fprintf(stderr, gettext("cannot import '%s': pool " "is formatted using an unsupported ZFS version\n"), name); return (1); } else if (zfs_force_import_required(config) && !(flags & ZFS_IMPORT_ANY_HOST)) { mmp_state_t mmp_state = MMP_STATE_INACTIVE; nvlist_t *nvinfo; nvinfo = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO); if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_STATE)) mmp_state = fnvlist_lookup_uint64(nvinfo, ZPOOL_CONFIG_MMP_STATE); if (mmp_state == MMP_STATE_ACTIVE) { const char *hostname = ""; uint64_t hostid = 0; if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_HOSTNAME)) hostname = fnvlist_lookup_string(nvinfo, ZPOOL_CONFIG_MMP_HOSTNAME); if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_HOSTID)) hostid = fnvlist_lookup_uint64(nvinfo, ZPOOL_CONFIG_MMP_HOSTID); (void) fprintf(stderr, gettext("cannot import '%s': " "pool is imported on %s (hostid: " "0x%"PRIx64")\nExport the pool on the other " "system, then run 'zpool import'.\n"), name, hostname, hostid); } else if (mmp_state == MMP_STATE_NO_HOSTID) { (void) fprintf(stderr, gettext("Cannot import '%s': " "pool has the multihost property on and the\n" "system's hostid is not set. Set a unique hostid " "with the zgenhostid(8) command.\n"), name); } else { const char *hostname = ""; time_t timestamp = 0; uint64_t hostid = 0; if (nvlist_exists(config, ZPOOL_CONFIG_HOSTNAME)) hostname = fnvlist_lookup_string(config, ZPOOL_CONFIG_HOSTNAME); if (nvlist_exists(config, ZPOOL_CONFIG_TIMESTAMP)) timestamp = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_TIMESTAMP); if (nvlist_exists(config, ZPOOL_CONFIG_HOSTID)) hostid = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_HOSTID); (void) fprintf(stderr, gettext("cannot import '%s': " "pool was previously in use from another system.\n" "Last accessed by %s (hostid=%"PRIx64") at %s" "The pool can be imported, use 'zpool import -f' " "to import the pool.\n"), name, hostname, hostid, ctime(×tamp)); } return (1); } if (zpool_import_props(g_zfs, config, newname, props, flags) != 0) return (1); if (newname != NULL) name = newname; if ((zhp = zpool_open_canfail(g_zfs, name)) == NULL) return (1); /* * Loading keys is best effort. We don't want to return immediately * if it fails but we do want to give the error to the caller. */ if (flags & ZFS_IMPORT_LOAD_KEYS && zfs_crypto_attempt_load_keys(g_zfs, name) != 0) ret = 1; if (zpool_get_state(zhp) != POOL_STATE_UNAVAIL && !(flags & ZFS_IMPORT_ONLY) && zpool_enable_datasets(zhp, mntopts, 0) != 0) { zpool_close(zhp); return (1); } zpool_close(zhp); return (ret); } static int import_pools(nvlist_t *pools, nvlist_t *props, char *mntopts, int flags, char *orig_name, char *new_name, boolean_t do_destroyed, boolean_t pool_specified, boolean_t do_all, importargs_t *import) { nvlist_t *config = NULL; nvlist_t *found_config = NULL; uint64_t pool_state; /* * At this point we have a list of import candidate configs. Even if * we were searching by pool name or guid, we still need to * post-process the list to deal with pool state and possible * duplicate names. */ int err = 0; nvpair_t *elem = NULL; boolean_t first = B_TRUE; while ((elem = nvlist_next_nvpair(pools, elem)) != NULL) { verify(nvpair_value_nvlist(elem, &config) == 0); verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE, &pool_state) == 0); if (!do_destroyed && pool_state == POOL_STATE_DESTROYED) continue; if (do_destroyed && pool_state != POOL_STATE_DESTROYED) continue; verify(nvlist_add_nvlist(config, ZPOOL_LOAD_POLICY, import->policy) == 0); if (!pool_specified) { if (first) first = B_FALSE; else if (!do_all) (void) fputc('\n', stdout); if (do_all) { err |= do_import(config, NULL, mntopts, props, flags); } else { /* * If we're importing from cachefile, then * we don't want to report errors until we * are in the scan phase of the import. If * we get an error, then we return that error * to invoke the scan phase. */ if (import->cachefile && !import->scan) err = show_import(config, B_FALSE); else (void) show_import(config, B_TRUE); } } else if (import->poolname != NULL) { const char *name; /* * We are searching for a pool based on name. */ verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME, &name) == 0); if (strcmp(name, import->poolname) == 0) { if (found_config != NULL) { (void) fprintf(stderr, gettext( "cannot import '%s': more than " "one matching pool\n"), import->poolname); (void) fprintf(stderr, gettext( "import by numeric ID instead\n")); err = B_TRUE; } found_config = config; } } else { uint64_t guid; /* * Search for a pool by guid. */ verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID, &guid) == 0); if (guid == import->guid) found_config = config; } } /* * If we were searching for a specific pool, verify that we found a * pool, and then do the import. */ if (pool_specified && err == 0) { if (found_config == NULL) { (void) fprintf(stderr, gettext("cannot import '%s': " "no such pool available\n"), orig_name); err = B_TRUE; } else { err |= do_import(found_config, new_name, mntopts, props, flags); } } /* * If we were just looking for pools, report an error if none were * found. */ if (!pool_specified && first) (void) fprintf(stderr, gettext("no pools available to import\n")); return (err); } typedef struct target_exists_args { const char *poolname; uint64_t poolguid; } target_exists_args_t; static int name_or_guid_exists(zpool_handle_t *zhp, void *data) { target_exists_args_t *args = data; nvlist_t *config = zpool_get_config(zhp, NULL); int found = 0; if (config == NULL) return (0); if (args->poolname != NULL) { const char *pool_name; verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME, &pool_name) == 0); if (strcmp(pool_name, args->poolname) == 0) found = 1; } else { uint64_t pool_guid; verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID, &pool_guid) == 0); if (pool_guid == args->poolguid) found = 1; } zpool_close(zhp); return (found); } /* * zpool checkpoint * checkpoint --discard * * -d Discard the checkpoint from a checkpointed * --discard pool. * * -w Wait for discarding a checkpoint to complete. * --wait * * Checkpoints the specified pool, by taking a "snapshot" of its * current state. A pool can only have one checkpoint at a time. */ int zpool_do_checkpoint(int argc, char **argv) { boolean_t discard, wait; char *pool; zpool_handle_t *zhp; int c, err; struct option long_options[] = { {"discard", no_argument, NULL, 'd'}, {"wait", no_argument, NULL, 'w'}, {0, 0, 0, 0} }; discard = B_FALSE; wait = B_FALSE; while ((c = getopt_long(argc, argv, ":dw", long_options, NULL)) != -1) { switch (c) { case 'd': discard = B_TRUE; break; case 'w': wait = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } if (wait && !discard) { (void) fprintf(stderr, gettext("--wait only valid when " "--discard also specified\n")); usage(B_FALSE); } argc -= optind; argv += optind; if (argc < 1) { (void) fprintf(stderr, gettext("missing pool argument\n")); usage(B_FALSE); } if (argc > 1) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } pool = argv[0]; if ((zhp = zpool_open(g_zfs, pool)) == NULL) { /* As a special case, check for use of '/' in the name */ if (strchr(pool, '/') != NULL) (void) fprintf(stderr, gettext("'zpool checkpoint' " "doesn't work on datasets. To save the state " "of a dataset from a specific point in time " "please use 'zfs snapshot'\n")); return (1); } if (discard) { err = (zpool_discard_checkpoint(zhp) != 0); if (err == 0 && wait) err = zpool_wait(zhp, ZPOOL_WAIT_CKPT_DISCARD); } else { err = (zpool_checkpoint(zhp) != 0); } zpool_close(zhp); return (err); } #define CHECKPOINT_OPT 1024 /* * zpool import [-d dir] [-D] * import [-o mntopts] [-o prop=value] ... [-R root] [-D] [-l] * [-d dir | -c cachefile | -s] [-f] -a * import [-o mntopts] [-o prop=value] ... [-R root] [-D] [-l] * [-d dir | -c cachefile | -s] [-f] [-n] [-F] * [newpool] * * -c Read pool information from a cachefile instead of searching * devices. If importing from a cachefile config fails, then * fallback to searching for devices only in the directories that * exist in the cachefile. * * -d Scan in a specific directory, other than /dev/. More than * one directory can be specified using multiple '-d' options. * * -D Scan for previously destroyed pools or import all or only * specified destroyed pools. * * -R Temporarily import the pool, with all mountpoints relative to * the given root. The pool will remain exported when the machine * is rebooted. * * -V Import even in the presence of faulted vdevs. This is an * intentionally undocumented option for testing purposes, and * treats the pool configuration as complete, leaving any bad * vdevs in the FAULTED state. In other words, it does verbatim * import. * * -f Force import, even if it appears that the pool is active. * * -F Attempt rewind if necessary. * * -n See if rewind would work, but don't actually rewind. * * -N Import the pool but don't mount datasets. * * -T Specify a starting txg to use for import. This option is * intentionally undocumented option for testing purposes. * * -a Import all pools found. * * -l Load encryption keys while importing. * * -o Set property=value and/or temporary mount options (without '='). * * -s Scan using the default search path, the libblkid cache will * not be consulted. * * --rewind-to-checkpoint * Import the pool and revert back to the checkpoint. * * The import command scans for pools to import, and import pools based on pool * name and GUID. The pool can also be renamed as part of the import process. */ int zpool_do_import(int argc, char **argv) { char **searchdirs = NULL; char *env, *envdup = NULL; int nsearch = 0; int c; int err = 0; nvlist_t *pools = NULL; boolean_t do_all = B_FALSE; boolean_t do_destroyed = B_FALSE; char *mntopts = NULL; uint64_t searchguid = 0; char *searchname = NULL; char *propval; nvlist_t *policy = NULL; nvlist_t *props = NULL; int flags = ZFS_IMPORT_NORMAL; uint32_t rewind_policy = ZPOOL_NO_REWIND; boolean_t dryrun = B_FALSE; boolean_t do_rewind = B_FALSE; boolean_t xtreme_rewind = B_FALSE; boolean_t do_scan = B_FALSE; boolean_t pool_exists = B_FALSE; boolean_t pool_specified = B_FALSE; uint64_t txg = -1ULL; char *cachefile = NULL; importargs_t idata = { 0 }; char *endptr; struct option long_options[] = { {"rewind-to-checkpoint", no_argument, NULL, CHECKPOINT_OPT}, {0, 0, 0, 0} }; /* check options */ while ((c = getopt_long(argc, argv, ":aCc:d:DEfFlmnNo:R:stT:VX", long_options, NULL)) != -1) { switch (c) { case 'a': do_all = B_TRUE; break; case 'c': cachefile = optarg; break; case 'd': searchdirs = safe_realloc(searchdirs, (nsearch + 1) * sizeof (char *)); searchdirs[nsearch++] = optarg; break; case 'D': do_destroyed = B_TRUE; break; case 'f': flags |= ZFS_IMPORT_ANY_HOST; break; case 'F': do_rewind = B_TRUE; break; case 'l': flags |= ZFS_IMPORT_LOAD_KEYS; break; case 'm': flags |= ZFS_IMPORT_MISSING_LOG; break; case 'n': dryrun = B_TRUE; break; case 'N': flags |= ZFS_IMPORT_ONLY; break; case 'o': if ((propval = strchr(optarg, '=')) != NULL) { *propval = '\0'; propval++; if (add_prop_list(optarg, propval, &props, B_TRUE)) goto error; } else { mntopts = optarg; } break; case 'R': if (add_prop_list(zpool_prop_to_name( ZPOOL_PROP_ALTROOT), optarg, &props, B_TRUE)) goto error; if (add_prop_list_default(zpool_prop_to_name( ZPOOL_PROP_CACHEFILE), "none", &props)) goto error; break; case 's': do_scan = B_TRUE; break; case 't': flags |= ZFS_IMPORT_TEMP_NAME; if (add_prop_list_default(zpool_prop_to_name( ZPOOL_PROP_CACHEFILE), "none", &props)) goto error; break; case 'T': errno = 0; txg = strtoull(optarg, &endptr, 0); if (errno != 0 || *endptr != '\0') { (void) fprintf(stderr, gettext("invalid txg value\n")); usage(B_FALSE); } rewind_policy = ZPOOL_DO_REWIND | ZPOOL_EXTREME_REWIND; break; case 'V': flags |= ZFS_IMPORT_VERBATIM; break; case 'X': xtreme_rewind = B_TRUE; break; case CHECKPOINT_OPT: flags |= ZFS_IMPORT_CHECKPOINT; break; case ':': (void) fprintf(stderr, gettext("missing argument for " "'%c' option\n"), optopt); usage(B_FALSE); break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; if (cachefile && nsearch != 0) { (void) fprintf(stderr, gettext("-c is incompatible with -d\n")); usage(B_FALSE); } if (cachefile && do_scan) { (void) fprintf(stderr, gettext("-c is incompatible with -s\n")); usage(B_FALSE); } if ((flags & ZFS_IMPORT_LOAD_KEYS) && (flags & ZFS_IMPORT_ONLY)) { (void) fprintf(stderr, gettext("-l is incompatible with -N\n")); usage(B_FALSE); } if ((flags & ZFS_IMPORT_LOAD_KEYS) && !do_all && argc == 0) { (void) fprintf(stderr, gettext("-l is only meaningful during " "an import\n")); usage(B_FALSE); } if ((dryrun || xtreme_rewind) && !do_rewind) { (void) fprintf(stderr, gettext("-n or -X only meaningful with -F\n")); usage(B_FALSE); } if (dryrun) rewind_policy = ZPOOL_TRY_REWIND; else if (do_rewind) rewind_policy = ZPOOL_DO_REWIND; if (xtreme_rewind) rewind_policy |= ZPOOL_EXTREME_REWIND; /* In the future, we can capture further policy and include it here */ if (nvlist_alloc(&policy, NV_UNIQUE_NAME, 0) != 0 || nvlist_add_uint64(policy, ZPOOL_LOAD_REQUEST_TXG, txg) != 0 || nvlist_add_uint32(policy, ZPOOL_LOAD_REWIND_POLICY, rewind_policy) != 0) goto error; /* check argument count */ if (do_all) { if (argc != 0) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } } else { if (argc > 2) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } } /* * Check for the effective uid. We do this explicitly here because * otherwise any attempt to discover pools will silently fail. */ if (argc == 0 && geteuid() != 0) { (void) fprintf(stderr, gettext("cannot " "discover pools: permission denied\n")); free(searchdirs); nvlist_free(props); nvlist_free(policy); return (1); } /* * Depending on the arguments given, we do one of the following: * * Iterate through all pools and display information about * each one. * * -a Iterate through all pools and try to import each one. * * Find the pool that corresponds to the given GUID/pool * name and import that one. * * -D Above options applies only to destroyed pools. */ if (argc != 0) { char *endptr; errno = 0; searchguid = strtoull(argv[0], &endptr, 10); if (errno != 0 || *endptr != '\0') { searchname = argv[0]; searchguid = 0; } pool_specified = B_TRUE; /* * User specified a name or guid. Ensure it's unique. */ target_exists_args_t search = {searchname, searchguid}; pool_exists = zpool_iter(g_zfs, name_or_guid_exists, &search); } /* * Check the environment for the preferred search path. */ if ((searchdirs == NULL) && (env = getenv("ZPOOL_IMPORT_PATH"))) { char *dir, *tmp = NULL; envdup = strdup(env); for (dir = strtok_r(envdup, ":", &tmp); dir != NULL; dir = strtok_r(NULL, ":", &tmp)) { searchdirs = safe_realloc(searchdirs, (nsearch + 1) * sizeof (char *)); searchdirs[nsearch++] = dir; } } idata.path = searchdirs; idata.paths = nsearch; idata.poolname = searchname; idata.guid = searchguid; idata.cachefile = cachefile; idata.scan = do_scan; idata.policy = policy; libpc_handle_t lpch = { .lpc_lib_handle = g_zfs, .lpc_ops = &libzfs_config_ops, .lpc_printerr = B_TRUE }; pools = zpool_search_import(&lpch, &idata); if (pools != NULL && pool_exists && (argc == 1 || strcmp(argv[0], argv[1]) == 0)) { (void) fprintf(stderr, gettext("cannot import '%s': " "a pool with that name already exists\n"), argv[0]); (void) fprintf(stderr, gettext("use the form '%s " " ' to give it a new name\n"), "zpool import"); err = 1; } else if (pools == NULL && pool_exists) { (void) fprintf(stderr, gettext("cannot import '%s': " "a pool with that name is already created/imported,\n"), argv[0]); (void) fprintf(stderr, gettext("and no additional pools " "with that name were found\n")); err = 1; } else if (pools == NULL) { if (argc != 0) { (void) fprintf(stderr, gettext("cannot import '%s': " "no such pool available\n"), argv[0]); } err = 1; } if (err == 1) { free(searchdirs); free(envdup); nvlist_free(policy); nvlist_free(pools); nvlist_free(props); return (1); } err = import_pools(pools, props, mntopts, flags, argc >= 1 ? argv[0] : NULL, argc >= 2 ? argv[1] : NULL, do_destroyed, pool_specified, do_all, &idata); /* * If we're using the cachefile and we failed to import, then * fallback to scanning the directory for pools that match * those in the cachefile. */ if (err != 0 && cachefile != NULL) { (void) printf(gettext("cachefile import failed, retrying\n")); /* * We use the scan flag to gather the directories that exist * in the cachefile. If we need to fallback to searching for * the pool config, we will only search devices in these * directories. */ idata.scan = B_TRUE; nvlist_free(pools); pools = zpool_search_import(&lpch, &idata); err = import_pools(pools, props, mntopts, flags, argc >= 1 ? argv[0] : NULL, argc >= 2 ? argv[1] : NULL, do_destroyed, pool_specified, do_all, &idata); } error: nvlist_free(props); nvlist_free(pools); nvlist_free(policy); free(searchdirs); free(envdup); return (err ? 1 : 0); } /* * zpool sync [-f] [pool] ... * * -f (undocumented) force uberblock (and config including zpool cache file) * update. * * Sync the specified pool(s). * Without arguments "zpool sync" will sync all pools. * This command initiates TXG sync(s) and will return after the TXG(s) commit. * */ static int zpool_do_sync(int argc, char **argv) { int ret; boolean_t force = B_FALSE; /* check options */ while ((ret = getopt(argc, argv, "f")) != -1) { switch (ret) { case 'f': force = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* if argc == 0 we will execute zpool_sync_one on all pools */ ret = for_each_pool(argc, argv, B_FALSE, NULL, ZFS_TYPE_POOL, B_FALSE, zpool_sync_one, &force); return (ret); } typedef struct iostat_cbdata { uint64_t cb_flags; int cb_namewidth; int cb_iteration; boolean_t cb_verbose; boolean_t cb_literal; boolean_t cb_scripted; zpool_list_t *cb_list; vdev_cmd_data_list_t *vcdl; vdev_cbdata_t cb_vdevs; } iostat_cbdata_t; /* iostat labels */ typedef struct name_and_columns { const char *name; /* Column name */ unsigned int columns; /* Center name to this number of columns */ } name_and_columns_t; #define IOSTAT_MAX_LABELS 15 /* Max number of labels on one line */ static const name_and_columns_t iostat_top_labels[][IOSTAT_MAX_LABELS] = { [IOS_DEFAULT] = {{"capacity", 2}, {"operations", 2}, {"bandwidth", 2}, {NULL}}, [IOS_LATENCY] = {{"total_wait", 2}, {"disk_wait", 2}, {"syncq_wait", 2}, {"asyncq_wait", 2}, {"scrub", 1}, {"trim", 1}, {"rebuild", 1}, {NULL}}, [IOS_QUEUES] = {{"syncq_read", 2}, {"syncq_write", 2}, {"asyncq_read", 2}, {"asyncq_write", 2}, {"scrubq_read", 2}, {"trimq_write", 2}, {"rebuildq_write", 2}, {NULL}}, [IOS_L_HISTO] = {{"total_wait", 2}, {"disk_wait", 2}, {"syncq_wait", 2}, {"asyncq_wait", 2}, {NULL}}, [IOS_RQ_HISTO] = {{"sync_read", 2}, {"sync_write", 2}, {"async_read", 2}, {"async_write", 2}, {"scrub", 2}, {"trim", 2}, {"rebuild", 2}, {NULL}}, }; /* Shorthand - if "columns" field not set, default to 1 column */ static const name_and_columns_t iostat_bottom_labels[][IOSTAT_MAX_LABELS] = { [IOS_DEFAULT] = {{"alloc"}, {"free"}, {"read"}, {"write"}, {"read"}, {"write"}, {NULL}}, [IOS_LATENCY] = {{"read"}, {"write"}, {"read"}, {"write"}, {"read"}, {"write"}, {"read"}, {"write"}, {"wait"}, {"wait"}, {"wait"}, {NULL}}, [IOS_QUEUES] = {{"pend"}, {"activ"}, {"pend"}, {"activ"}, {"pend"}, {"activ"}, {"pend"}, {"activ"}, {"pend"}, {"activ"}, {"pend"}, {"activ"}, {"pend"}, {"activ"}, {NULL}}, [IOS_L_HISTO] = {{"read"}, {"write"}, {"read"}, {"write"}, {"read"}, {"write"}, {"read"}, {"write"}, {"scrub"}, {"trim"}, {"rebuild"}, {NULL}}, [IOS_RQ_HISTO] = {{"ind"}, {"agg"}, {"ind"}, {"agg"}, {"ind"}, {"agg"}, {"ind"}, {"agg"}, {"ind"}, {"agg"}, {"ind"}, {"agg"}, {"ind"}, {"agg"}, {NULL}}, }; static const char *histo_to_title[] = { [IOS_L_HISTO] = "latency", [IOS_RQ_HISTO] = "req_size", }; /* * Return the number of labels in a null-terminated name_and_columns_t * array. * */ static unsigned int label_array_len(const name_and_columns_t *labels) { int i = 0; while (labels[i].name) i++; return (i); } /* * Return the number of strings in a null-terminated string array. * For example: * * const char foo[] = {"bar", "baz", NULL} * * returns 2 */ static uint64_t str_array_len(const char *array[]) { uint64_t i = 0; while (array[i]) i++; return (i); } /* * Return a default column width for default/latency/queue columns. This does * not include histograms, which have their columns autosized. */ static unsigned int default_column_width(iostat_cbdata_t *cb, enum iostat_type type) { unsigned long column_width = 5; /* Normal niceprint */ static unsigned long widths[] = { /* * Choose some sane default column sizes for printing the * raw numbers. */ [IOS_DEFAULT] = 15, /* 1PB capacity */ [IOS_LATENCY] = 10, /* 1B ns = 10sec */ [IOS_QUEUES] = 6, /* 1M queue entries */ [IOS_L_HISTO] = 10, /* 1B ns = 10sec */ [IOS_RQ_HISTO] = 6, /* 1M queue entries */ }; if (cb->cb_literal) column_width = widths[type]; return (column_width); } /* * Print the column labels, i.e: * * capacity operations bandwidth * alloc free read write read write ... * * If force_column_width is set, use it for the column width. If not set, use * the default column width. */ static void print_iostat_labels(iostat_cbdata_t *cb, unsigned int force_column_width, const name_and_columns_t labels[][IOSTAT_MAX_LABELS]) { int i, idx, s; int text_start, rw_column_width, spaces_to_end; uint64_t flags = cb->cb_flags; uint64_t f; unsigned int column_width = force_column_width; /* For each bit set in flags */ for (f = flags; f; f &= ~(1ULL << idx)) { idx = lowbit64(f) - 1; if (!force_column_width) column_width = default_column_width(cb, idx); /* Print our top labels centered over "read write" label. */ for (i = 0; i < label_array_len(labels[idx]); i++) { const char *name = labels[idx][i].name; /* * We treat labels[][].columns == 0 as shorthand * for one column. It makes writing out the label * tables more concise. */ unsigned int columns = MAX(1, labels[idx][i].columns); unsigned int slen = strlen(name); rw_column_width = (column_width * columns) + (2 * (columns - 1)); text_start = (int)((rw_column_width) / columns - slen / columns); if (text_start < 0) text_start = 0; printf(" "); /* Two spaces between columns */ /* Space from beginning of column to label */ for (s = 0; s < text_start; s++) printf(" "); printf("%s", name); /* Print space after label to end of column */ spaces_to_end = rw_column_width - text_start - slen; if (spaces_to_end < 0) spaces_to_end = 0; for (s = 0; s < spaces_to_end; s++) printf(" "); } } } /* * print_cmd_columns - Print custom column titles from -c * * If the user specified the "zpool status|iostat -c" then print their custom * column titles in the header. For example, print_cmd_columns() would print * the " col1 col2" part of this: * * $ zpool iostat -vc 'echo col1=val1; echo col2=val2' * ... * capacity operations bandwidth * pool alloc free read write read write col1 col2 * ---------- ----- ----- ----- ----- ----- ----- ---- ---- * mypool 269K 1008M 0 0 107 946 * mirror 269K 1008M 0 0 107 946 * sdb - - 0 0 102 473 val1 val2 * sdc - - 0 0 5 473 val1 val2 * ---------- ----- ----- ----- ----- ----- ----- ---- ---- */ static void print_cmd_columns(vdev_cmd_data_list_t *vcdl, int use_dashes) { int i, j; vdev_cmd_data_t *data = &vcdl->data[0]; if (vcdl->count == 0 || data == NULL) return; /* * Each vdev cmd should have the same column names unless the user did * something weird with their cmd. Just take the column names from the * first vdev and assume it works for all of them. */ for (i = 0; i < vcdl->uniq_cols_cnt; i++) { printf(" "); if (use_dashes) { for (j = 0; j < vcdl->uniq_cols_width[i]; j++) printf("-"); } else { printf_color(ANSI_BOLD, "%*s", vcdl->uniq_cols_width[i], vcdl->uniq_cols[i]); } } } /* * Utility function to print out a line of dashes like: * * -------------------------------- ----- ----- ----- ----- ----- * * ...or a dashed named-row line like: * * logs - - - - - * * @cb: iostat data * * @force_column_width If non-zero, use the value as the column width. * Otherwise use the default column widths. * * @name: Print a dashed named-row line starting * with @name. Otherwise, print a regular * dashed line. */ static void print_iostat_dashes(iostat_cbdata_t *cb, unsigned int force_column_width, const char *name) { int i; unsigned int namewidth; uint64_t flags = cb->cb_flags; uint64_t f; int idx; const name_and_columns_t *labels; const char *title; if (cb->cb_flags & IOS_ANYHISTO_M) { title = histo_to_title[IOS_HISTO_IDX(cb->cb_flags)]; } else if (cb->cb_vdevs.cb_names_count) { title = "vdev"; } else { title = "pool"; } namewidth = MAX(MAX(strlen(title), cb->cb_namewidth), name ? strlen(name) : 0); if (name) { printf("%-*s", namewidth, name); } else { for (i = 0; i < namewidth; i++) (void) printf("-"); } /* For each bit in flags */ for (f = flags; f; f &= ~(1ULL << idx)) { unsigned int column_width; idx = lowbit64(f) - 1; if (force_column_width) column_width = force_column_width; else column_width = default_column_width(cb, idx); labels = iostat_bottom_labels[idx]; for (i = 0; i < label_array_len(labels); i++) { if (name) printf(" %*s-", column_width - 1, " "); else printf(" %.*s", column_width, "--------------------"); } } } static void print_iostat_separator_impl(iostat_cbdata_t *cb, unsigned int force_column_width) { print_iostat_dashes(cb, force_column_width, NULL); } static void print_iostat_separator(iostat_cbdata_t *cb) { print_iostat_separator_impl(cb, 0); } static void print_iostat_header_impl(iostat_cbdata_t *cb, unsigned int force_column_width, const char *histo_vdev_name) { unsigned int namewidth; const char *title; color_start(ANSI_BOLD); if (cb->cb_flags & IOS_ANYHISTO_M) { title = histo_to_title[IOS_HISTO_IDX(cb->cb_flags)]; } else if (cb->cb_vdevs.cb_names_count) { title = "vdev"; } else { title = "pool"; } namewidth = MAX(MAX(strlen(title), cb->cb_namewidth), histo_vdev_name ? strlen(histo_vdev_name) : 0); if (histo_vdev_name) printf("%-*s", namewidth, histo_vdev_name); else printf("%*s", namewidth, ""); print_iostat_labels(cb, force_column_width, iostat_top_labels); printf("\n"); printf("%-*s", namewidth, title); print_iostat_labels(cb, force_column_width, iostat_bottom_labels); if (cb->vcdl != NULL) print_cmd_columns(cb->vcdl, 0); printf("\n"); print_iostat_separator_impl(cb, force_column_width); if (cb->vcdl != NULL) print_cmd_columns(cb->vcdl, 1); color_end(); printf("\n"); } static void print_iostat_header(iostat_cbdata_t *cb) { print_iostat_header_impl(cb, 0, NULL); } /* * Prints a size string (i.e. 120M) with the suffix ("M") colored * by order of magnitude. Uses column_size to add padding. */ static void print_stat_color(const char *statbuf, unsigned int column_size) { fputs(" ", stdout); size_t len = strlen(statbuf); while (len < column_size) { fputc(' ', stdout); column_size--; } if (*statbuf == '0') { color_start(ANSI_GRAY); fputc('0', stdout); } else { for (; *statbuf; statbuf++) { if (*statbuf == 'K') color_start(ANSI_GREEN); else if (*statbuf == 'M') color_start(ANSI_YELLOW); else if (*statbuf == 'G') color_start(ANSI_RED); else if (*statbuf == 'T') color_start(ANSI_BOLD_BLUE); else if (*statbuf == 'P') color_start(ANSI_MAGENTA); else if (*statbuf == 'E') color_start(ANSI_CYAN); fputc(*statbuf, stdout); if (--column_size <= 0) break; } } color_end(); } /* * Display a single statistic. */ static void print_one_stat(uint64_t value, enum zfs_nicenum_format format, unsigned int column_size, boolean_t scripted) { char buf[64]; zfs_nicenum_format(value, buf, sizeof (buf), format); if (scripted) printf("\t%s", buf); else print_stat_color(buf, column_size); } /* * Calculate the default vdev stats * * Subtract oldvs from newvs, apply a scaling factor, and save the resulting * stats into calcvs. */ static void calc_default_iostats(vdev_stat_t *oldvs, vdev_stat_t *newvs, vdev_stat_t *calcvs) { int i; memcpy(calcvs, newvs, sizeof (*calcvs)); for (i = 0; i < ARRAY_SIZE(calcvs->vs_ops); i++) calcvs->vs_ops[i] = (newvs->vs_ops[i] - oldvs->vs_ops[i]); for (i = 0; i < ARRAY_SIZE(calcvs->vs_bytes); i++) calcvs->vs_bytes[i] = (newvs->vs_bytes[i] - oldvs->vs_bytes[i]); } /* * Internal representation of the extended iostats data. * * The extended iostat stats are exported in nvlists as either uint64_t arrays * or single uint64_t's. We make both look like arrays to make them easier * to process. In order to make single uint64_t's look like arrays, we set * __data to the stat data, and then set *data = &__data with count = 1. Then, * we can just use *data and count. */ struct stat_array { uint64_t *data; uint_t count; /* Number of entries in data[] */ uint64_t __data; /* Only used when data is a single uint64_t */ }; static uint64_t stat_histo_max(struct stat_array *nva, unsigned int len) { uint64_t max = 0; int i; for (i = 0; i < len; i++) max = MAX(max, array64_max(nva[i].data, nva[i].count)); return (max); } /* * Helper function to lookup a uint64_t array or uint64_t value and store its * data as a stat_array. If the nvpair is a single uint64_t value, then we make * it look like a one element array to make it easier to process. */ static int nvpair64_to_stat_array(nvlist_t *nvl, const char *name, struct stat_array *nva) { nvpair_t *tmp; int ret; verify(nvlist_lookup_nvpair(nvl, name, &tmp) == 0); switch (nvpair_type(tmp)) { case DATA_TYPE_UINT64_ARRAY: ret = nvpair_value_uint64_array(tmp, &nva->data, &nva->count); break; case DATA_TYPE_UINT64: ret = nvpair_value_uint64(tmp, &nva->__data); nva->data = &nva->__data; nva->count = 1; break; default: /* Not a uint64_t */ ret = EINVAL; break; } return (ret); } /* * Given a list of nvlist names, look up the extended stats in newnv and oldnv, * subtract them, and return the results in a newly allocated stat_array. * You must free the returned array after you are done with it with * free_calc_stats(). * * Additionally, you can set "oldnv" to NULL if you simply want the newnv * values. */ static struct stat_array * calc_and_alloc_stats_ex(const char **names, unsigned int len, nvlist_t *oldnv, nvlist_t *newnv) { nvlist_t *oldnvx = NULL, *newnvx; struct stat_array *oldnva, *newnva, *calcnva; int i, j; unsigned int alloc_size = (sizeof (struct stat_array)) * len; /* Extract our extended stats nvlist from the main list */ verify(nvlist_lookup_nvlist(newnv, ZPOOL_CONFIG_VDEV_STATS_EX, &newnvx) == 0); if (oldnv) { verify(nvlist_lookup_nvlist(oldnv, ZPOOL_CONFIG_VDEV_STATS_EX, &oldnvx) == 0); } newnva = safe_malloc(alloc_size); oldnva = safe_malloc(alloc_size); calcnva = safe_malloc(alloc_size); for (j = 0; j < len; j++) { verify(nvpair64_to_stat_array(newnvx, names[j], &newnva[j]) == 0); calcnva[j].count = newnva[j].count; alloc_size = calcnva[j].count * sizeof (calcnva[j].data[0]); calcnva[j].data = safe_malloc(alloc_size); memcpy(calcnva[j].data, newnva[j].data, alloc_size); if (oldnvx) { verify(nvpair64_to_stat_array(oldnvx, names[j], &oldnva[j]) == 0); for (i = 0; i < oldnva[j].count; i++) calcnva[j].data[i] -= oldnva[j].data[i]; } } free(newnva); free(oldnva); return (calcnva); } static void free_calc_stats(struct stat_array *nva, unsigned int len) { int i; for (i = 0; i < len; i++) free(nva[i].data); free(nva); } static void print_iostat_histo(struct stat_array *nva, unsigned int len, iostat_cbdata_t *cb, unsigned int column_width, unsigned int namewidth, double scale) { int i, j; char buf[6]; uint64_t val; enum zfs_nicenum_format format; unsigned int buckets; unsigned int start_bucket; if (cb->cb_literal) format = ZFS_NICENUM_RAW; else format = ZFS_NICENUM_1024; /* All these histos are the same size, so just use nva[0].count */ buckets = nva[0].count; if (cb->cb_flags & IOS_RQ_HISTO_M) { /* Start at 512 - req size should never be lower than this */ start_bucket = 9; } else { start_bucket = 0; } for (j = start_bucket; j < buckets; j++) { /* Print histogram bucket label */ if (cb->cb_flags & IOS_L_HISTO_M) { /* Ending range of this bucket */ val = (1UL << (j + 1)) - 1; zfs_nicetime(val, buf, sizeof (buf)); } else { /* Request size (starting range of bucket) */ val = (1UL << j); zfs_nicenum(val, buf, sizeof (buf)); } if (cb->cb_scripted) printf("%llu", (u_longlong_t)val); else printf("%-*s", namewidth, buf); /* Print the values on the line */ for (i = 0; i < len; i++) { print_one_stat(nva[i].data[j] * scale, format, column_width, cb->cb_scripted); } printf("\n"); } } static void print_solid_separator(unsigned int length) { while (length--) printf("-"); printf("\n"); } static void print_iostat_histos(iostat_cbdata_t *cb, nvlist_t *oldnv, nvlist_t *newnv, double scale, const char *name) { unsigned int column_width; unsigned int namewidth; unsigned int entire_width; enum iostat_type type; struct stat_array *nva; const char **names; unsigned int names_len; /* What type of histo are we? */ type = IOS_HISTO_IDX(cb->cb_flags); /* Get NULL-terminated array of nvlist names for our histo */ names = vsx_type_to_nvlist[type]; names_len = str_array_len(names); /* num of names */ nva = calc_and_alloc_stats_ex(names, names_len, oldnv, newnv); if (cb->cb_literal) { column_width = MAX(5, (unsigned int) log10(stat_histo_max(nva, names_len)) + 1); } else { column_width = 5; } namewidth = MAX(cb->cb_namewidth, strlen(histo_to_title[IOS_HISTO_IDX(cb->cb_flags)])); /* * Calculate the entire line width of what we're printing. The * +2 is for the two spaces between columns: */ /* read write */ /* ----- ----- */ /* |___| <---------- column_width */ /* */ /* |__________| <--- entire_width */ /* */ entire_width = namewidth + (column_width + 2) * label_array_len(iostat_bottom_labels[type]); if (cb->cb_scripted) printf("%s\n", name); else print_iostat_header_impl(cb, column_width, name); print_iostat_histo(nva, names_len, cb, column_width, namewidth, scale); free_calc_stats(nva, names_len); if (!cb->cb_scripted) print_solid_separator(entire_width); } /* * Calculate the average latency of a power-of-two latency histogram */ static uint64_t single_histo_average(uint64_t *histo, unsigned int buckets) { int i; uint64_t count = 0, total = 0; for (i = 0; i < buckets; i++) { /* * Our buckets are power-of-two latency ranges. Use the * midpoint latency of each bucket to calculate the average. * For example: * * Bucket Midpoint * 8ns-15ns: 12ns * 16ns-31ns: 24ns * ... */ if (histo[i] != 0) { total += histo[i] * (((1UL << i) + ((1UL << i)/2))); count += histo[i]; } } /* Prevent divide by zero */ return (count == 0 ? 0 : total / count); } static void print_iostat_queues(iostat_cbdata_t *cb, nvlist_t *newnv) { const char *names[] = { ZPOOL_CONFIG_VDEV_SYNC_R_PEND_QUEUE, ZPOOL_CONFIG_VDEV_SYNC_R_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_SYNC_W_PEND_QUEUE, ZPOOL_CONFIG_VDEV_SYNC_W_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_ASYNC_R_PEND_QUEUE, ZPOOL_CONFIG_VDEV_ASYNC_R_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_ASYNC_W_PEND_QUEUE, ZPOOL_CONFIG_VDEV_ASYNC_W_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_SCRUB_PEND_QUEUE, ZPOOL_CONFIG_VDEV_SCRUB_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_TRIM_PEND_QUEUE, ZPOOL_CONFIG_VDEV_TRIM_ACTIVE_QUEUE, ZPOOL_CONFIG_VDEV_REBUILD_PEND_QUEUE, ZPOOL_CONFIG_VDEV_REBUILD_ACTIVE_QUEUE, }; struct stat_array *nva; unsigned int column_width = default_column_width(cb, IOS_QUEUES); enum zfs_nicenum_format format; nva = calc_and_alloc_stats_ex(names, ARRAY_SIZE(names), NULL, newnv); if (cb->cb_literal) format = ZFS_NICENUM_RAW; else format = ZFS_NICENUM_1024; for (int i = 0; i < ARRAY_SIZE(names); i++) { uint64_t val = nva[i].data[0]; print_one_stat(val, format, column_width, cb->cb_scripted); } free_calc_stats(nva, ARRAY_SIZE(names)); } static void print_iostat_latency(iostat_cbdata_t *cb, nvlist_t *oldnv, nvlist_t *newnv) { int i; uint64_t val; const char *names[] = { ZPOOL_CONFIG_VDEV_TOT_R_LAT_HISTO, ZPOOL_CONFIG_VDEV_TOT_W_LAT_HISTO, ZPOOL_CONFIG_VDEV_DISK_R_LAT_HISTO, ZPOOL_CONFIG_VDEV_DISK_W_LAT_HISTO, ZPOOL_CONFIG_VDEV_SYNC_R_LAT_HISTO, ZPOOL_CONFIG_VDEV_SYNC_W_LAT_HISTO, ZPOOL_CONFIG_VDEV_ASYNC_R_LAT_HISTO, ZPOOL_CONFIG_VDEV_ASYNC_W_LAT_HISTO, ZPOOL_CONFIG_VDEV_SCRUB_LAT_HISTO, ZPOOL_CONFIG_VDEV_TRIM_LAT_HISTO, ZPOOL_CONFIG_VDEV_REBUILD_LAT_HISTO, }; struct stat_array *nva; unsigned int column_width = default_column_width(cb, IOS_LATENCY); enum zfs_nicenum_format format; nva = calc_and_alloc_stats_ex(names, ARRAY_SIZE(names), oldnv, newnv); if (cb->cb_literal) format = ZFS_NICENUM_RAWTIME; else format = ZFS_NICENUM_TIME; /* Print our avg latencies on the line */ for (i = 0; i < ARRAY_SIZE(names); i++) { /* Compute average latency for a latency histo */ val = single_histo_average(nva[i].data, nva[i].count); print_one_stat(val, format, column_width, cb->cb_scripted); } free_calc_stats(nva, ARRAY_SIZE(names)); } /* * Print default statistics (capacity/operations/bandwidth) */ static void print_iostat_default(vdev_stat_t *vs, iostat_cbdata_t *cb, double scale) { unsigned int column_width = default_column_width(cb, IOS_DEFAULT); enum zfs_nicenum_format format; char na; /* char to print for "not applicable" values */ if (cb->cb_literal) { format = ZFS_NICENUM_RAW; na = '0'; } else { format = ZFS_NICENUM_1024; na = '-'; } /* only toplevel vdevs have capacity stats */ if (vs->vs_space == 0) { if (cb->cb_scripted) printf("\t%c\t%c", na, na); else printf(" %*c %*c", column_width, na, column_width, na); } else { print_one_stat(vs->vs_alloc, format, column_width, cb->cb_scripted); print_one_stat(vs->vs_space - vs->vs_alloc, format, column_width, cb->cb_scripted); } print_one_stat((uint64_t)(vs->vs_ops[ZIO_TYPE_READ] * scale), format, column_width, cb->cb_scripted); print_one_stat((uint64_t)(vs->vs_ops[ZIO_TYPE_WRITE] * scale), format, column_width, cb->cb_scripted); print_one_stat((uint64_t)(vs->vs_bytes[ZIO_TYPE_READ] * scale), format, column_width, cb->cb_scripted); print_one_stat((uint64_t)(vs->vs_bytes[ZIO_TYPE_WRITE] * scale), format, column_width, cb->cb_scripted); } static const char *const class_name[] = { VDEV_ALLOC_BIAS_DEDUP, VDEV_ALLOC_BIAS_SPECIAL, VDEV_ALLOC_CLASS_LOGS }; /* * Print out all the statistics for the given vdev. This can either be the * toplevel configuration, or called recursively. If 'name' is NULL, then this * is a verbose output, and we don't want to display the toplevel pool stats. * * Returns the number of stat lines printed. */ static unsigned int print_vdev_stats(zpool_handle_t *zhp, const char *name, nvlist_t *oldnv, nvlist_t *newnv, iostat_cbdata_t *cb, int depth) { nvlist_t **oldchild, **newchild; uint_t c, children, oldchildren; vdev_stat_t *oldvs, *newvs, *calcvs; vdev_stat_t zerovs = { 0 }; char *vname; int i; int ret = 0; uint64_t tdelta; double scale; if (strcmp(name, VDEV_TYPE_INDIRECT) == 0) return (ret); calcvs = safe_malloc(sizeof (*calcvs)); if (oldnv != NULL) { verify(nvlist_lookup_uint64_array(oldnv, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&oldvs, &c) == 0); } else { oldvs = &zerovs; } /* Do we only want to see a specific vdev? */ for (i = 0; i < cb->cb_vdevs.cb_names_count; i++) { /* Yes we do. Is this the vdev? */ if (strcmp(name, cb->cb_vdevs.cb_names[i]) == 0) { /* * This is our vdev. Since it is the only vdev we * will be displaying, make depth = 0 so that it * doesn't get indented. */ depth = 0; break; } } if (cb->cb_vdevs.cb_names_count && (i == cb->cb_vdevs.cb_names_count)) { /* Couldn't match the name */ goto children; } verify(nvlist_lookup_uint64_array(newnv, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&newvs, &c) == 0); /* * Print the vdev name unless it's is a histogram. Histograms * display the vdev name in the header itself. */ if (!(cb->cb_flags & IOS_ANYHISTO_M)) { if (cb->cb_scripted) { printf("%s", name); } else { if (strlen(name) + depth > cb->cb_namewidth) (void) printf("%*s%s", depth, "", name); else (void) printf("%*s%s%*s", depth, "", name, (int)(cb->cb_namewidth - strlen(name) - depth), ""); } } /* Calculate our scaling factor */ tdelta = newvs->vs_timestamp - oldvs->vs_timestamp; if ((oldvs->vs_timestamp == 0) && (cb->cb_flags & IOS_ANYHISTO_M)) { /* * If we specify printing histograms with no time interval, then * print the histogram numbers over the entire lifetime of the * vdev. */ scale = 1; } else { if (tdelta == 0) scale = 1.0; else scale = (double)NANOSEC / tdelta; } if (cb->cb_flags & IOS_DEFAULT_M) { calc_default_iostats(oldvs, newvs, calcvs); print_iostat_default(calcvs, cb, scale); } if (cb->cb_flags & IOS_LATENCY_M) print_iostat_latency(cb, oldnv, newnv); if (cb->cb_flags & IOS_QUEUES_M) print_iostat_queues(cb, newnv); if (cb->cb_flags & IOS_ANYHISTO_M) { printf("\n"); print_iostat_histos(cb, oldnv, newnv, scale, name); } if (cb->vcdl != NULL) { const char *path; if (nvlist_lookup_string(newnv, ZPOOL_CONFIG_PATH, &path) == 0) { printf(" "); zpool_print_cmd(cb->vcdl, zpool_get_name(zhp), path); } } if (!(cb->cb_flags & IOS_ANYHISTO_M)) printf("\n"); ret++; children: free(calcvs); if (!cb->cb_verbose) return (ret); if (nvlist_lookup_nvlist_array(newnv, ZPOOL_CONFIG_CHILDREN, &newchild, &children) != 0) return (ret); if (oldnv) { if (nvlist_lookup_nvlist_array(oldnv, ZPOOL_CONFIG_CHILDREN, &oldchild, &oldchildren) != 0) return (ret); children = MIN(oldchildren, children); } /* * print normal top-level devices */ for (c = 0; c < children; c++) { uint64_t ishole = B_FALSE, islog = B_FALSE; (void) nvlist_lookup_uint64(newchild[c], ZPOOL_CONFIG_IS_HOLE, &ishole); (void) nvlist_lookup_uint64(newchild[c], ZPOOL_CONFIG_IS_LOG, &islog); if (ishole || islog) continue; if (nvlist_exists(newchild[c], ZPOOL_CONFIG_ALLOCATION_BIAS)) continue; vname = zpool_vdev_name(g_zfs, zhp, newchild[c], cb->cb_vdevs.cb_name_flags | VDEV_NAME_TYPE_ID); ret += print_vdev_stats(zhp, vname, oldnv ? oldchild[c] : NULL, newchild[c], cb, depth + 2); free(vname); } /* * print all other top-level devices */ for (uint_t n = 0; n < ARRAY_SIZE(class_name); n++) { boolean_t printed = B_FALSE; for (c = 0; c < children; c++) { uint64_t islog = B_FALSE; const char *bias = NULL; const char *type = NULL; (void) nvlist_lookup_uint64(newchild[c], ZPOOL_CONFIG_IS_LOG, &islog); if (islog) { bias = VDEV_ALLOC_CLASS_LOGS; } else { (void) nvlist_lookup_string(newchild[c], ZPOOL_CONFIG_ALLOCATION_BIAS, &bias); (void) nvlist_lookup_string(newchild[c], ZPOOL_CONFIG_TYPE, &type); } if (bias == NULL || strcmp(bias, class_name[n]) != 0) continue; if (!islog && strcmp(type, VDEV_TYPE_INDIRECT) == 0) continue; if (!printed) { if ((!(cb->cb_flags & IOS_ANYHISTO_M)) && !cb->cb_scripted && !cb->cb_vdevs.cb_names) { print_iostat_dashes(cb, 0, class_name[n]); } printf("\n"); printed = B_TRUE; } vname = zpool_vdev_name(g_zfs, zhp, newchild[c], cb->cb_vdevs.cb_name_flags | VDEV_NAME_TYPE_ID); ret += print_vdev_stats(zhp, vname, oldnv ? oldchild[c] : NULL, newchild[c], cb, depth + 2); free(vname); } } /* * Include level 2 ARC devices in iostat output */ if (nvlist_lookup_nvlist_array(newnv, ZPOOL_CONFIG_L2CACHE, &newchild, &children) != 0) return (ret); if (oldnv) { if (nvlist_lookup_nvlist_array(oldnv, ZPOOL_CONFIG_L2CACHE, &oldchild, &oldchildren) != 0) return (ret); children = MIN(oldchildren, children); } if (children > 0) { if ((!(cb->cb_flags & IOS_ANYHISTO_M)) && !cb->cb_scripted && !cb->cb_vdevs.cb_names) { print_iostat_dashes(cb, 0, "cache"); } printf("\n"); for (c = 0; c < children; c++) { vname = zpool_vdev_name(g_zfs, zhp, newchild[c], cb->cb_vdevs.cb_name_flags); ret += print_vdev_stats(zhp, vname, oldnv ? oldchild[c] : NULL, newchild[c], cb, depth + 2); free(vname); } } return (ret); } static int refresh_iostat(zpool_handle_t *zhp, void *data) { iostat_cbdata_t *cb = data; boolean_t missing; /* * If the pool has disappeared, remove it from the list and continue. */ if (zpool_refresh_stats(zhp, &missing) != 0) return (-1); if (missing) pool_list_remove(cb->cb_list, zhp); return (0); } /* * Callback to print out the iostats for the given pool. */ static int print_iostat(zpool_handle_t *zhp, void *data) { iostat_cbdata_t *cb = data; nvlist_t *oldconfig, *newconfig; nvlist_t *oldnvroot, *newnvroot; int ret; newconfig = zpool_get_config(zhp, &oldconfig); if (cb->cb_iteration == 1) oldconfig = NULL; verify(nvlist_lookup_nvlist(newconfig, ZPOOL_CONFIG_VDEV_TREE, &newnvroot) == 0); if (oldconfig == NULL) oldnvroot = NULL; else verify(nvlist_lookup_nvlist(oldconfig, ZPOOL_CONFIG_VDEV_TREE, &oldnvroot) == 0); ret = print_vdev_stats(zhp, zpool_get_name(zhp), oldnvroot, newnvroot, cb, 0); if ((ret != 0) && !(cb->cb_flags & IOS_ANYHISTO_M) && !cb->cb_scripted && cb->cb_verbose && !cb->cb_vdevs.cb_names_count) { print_iostat_separator(cb); if (cb->vcdl != NULL) { print_cmd_columns(cb->vcdl, 1); } printf("\n"); } return (ret); } static int get_columns(void) { struct winsize ws; int columns = 80; int error; if (isatty(STDOUT_FILENO)) { error = ioctl(STDOUT_FILENO, TIOCGWINSZ, &ws); if (error == 0) columns = ws.ws_col; } else { columns = 999; } return (columns); } /* * Return the required length of the pool/vdev name column. The minimum * allowed width and output formatting flags must be provided. */ static int get_namewidth(zpool_handle_t *zhp, int min_width, int flags, boolean_t verbose) { nvlist_t *config, *nvroot; int width = min_width; if ((config = zpool_get_config(zhp, NULL)) != NULL) { verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); size_t poolname_len = strlen(zpool_get_name(zhp)); if (verbose == B_FALSE) { width = MAX(poolname_len, min_width); } else { width = MAX(poolname_len, max_width(zhp, nvroot, 0, min_width, flags)); } } return (width); } /* * Parse the input string, get the 'interval' and 'count' value if there is one. */ static void get_interval_count(int *argcp, char **argv, float *iv, unsigned long *cnt) { float interval = 0; unsigned long count = 0; int argc = *argcp; /* * Determine if the last argument is an integer or a pool name */ if (argc > 0 && zfs_isnumber(argv[argc - 1])) { char *end; errno = 0; interval = strtof(argv[argc - 1], &end); if (*end == '\0' && errno == 0) { if (interval == 0) { (void) fprintf(stderr, gettext( "interval cannot be zero\n")); usage(B_FALSE); } /* * Ignore the last parameter */ argc--; } else { /* * If this is not a valid number, just plow on. The * user will get a more informative error message later * on. */ interval = 0; } } /* * If the last argument is also an integer, then we have both a count * and an interval. */ if (argc > 0 && zfs_isnumber(argv[argc - 1])) { char *end; errno = 0; count = interval; interval = strtof(argv[argc - 1], &end); if (*end == '\0' && errno == 0) { if (interval == 0) { (void) fprintf(stderr, gettext( "interval cannot be zero\n")); usage(B_FALSE); } /* * Ignore the last parameter */ argc--; } else { interval = 0; } } *iv = interval; *cnt = count; *argcp = argc; } static void get_timestamp_arg(char c) { if (c == 'u') timestamp_fmt = UDATE; else if (c == 'd') timestamp_fmt = DDATE; else usage(B_FALSE); } /* * Return stat flags that are supported by all pools by both the module and * zpool iostat. "*data" should be initialized to all 0xFFs before running. * It will get ANDed down until only the flags that are supported on all pools * remain. */ static int get_stat_flags_cb(zpool_handle_t *zhp, void *data) { uint64_t *mask = data; nvlist_t *config, *nvroot, *nvx; uint64_t flags = 0; int i, j; config = zpool_get_config(zhp, NULL); verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); /* Default stats are always supported, but for completeness.. */ if (nvlist_exists(nvroot, ZPOOL_CONFIG_VDEV_STATS)) flags |= IOS_DEFAULT_M; /* Get our extended stats nvlist from the main list */ if (nvlist_lookup_nvlist(nvroot, ZPOOL_CONFIG_VDEV_STATS_EX, &nvx) != 0) { /* * No extended stats; they're probably running an older * module. No big deal, we support that too. */ goto end; } /* For each extended stat, make sure all its nvpairs are supported */ for (j = 0; j < ARRAY_SIZE(vsx_type_to_nvlist); j++) { if (!vsx_type_to_nvlist[j][0]) continue; /* Start off by assuming the flag is supported, then check */ flags |= (1ULL << j); for (i = 0; vsx_type_to_nvlist[j][i]; i++) { if (!nvlist_exists(nvx, vsx_type_to_nvlist[j][i])) { /* flag isn't supported */ flags = flags & ~(1ULL << j); break; } } } end: *mask = *mask & flags; return (0); } /* * Return a bitmask of stats that are supported on all pools by both the module * and zpool iostat. */ static uint64_t get_stat_flags(zpool_list_t *list) { uint64_t mask = -1; /* * get_stat_flags_cb() will lop off bits from "mask" until only the * flags that are supported on all pools remain. */ pool_list_iter(list, B_FALSE, get_stat_flags_cb, &mask); return (mask); } /* * Return 1 if cb_data->cb_names[0] is this vdev's name, 0 otherwise. */ static int is_vdev_cb(void *zhp_data, nvlist_t *nv, void *cb_data) { uint64_t guid; vdev_cbdata_t *cb = cb_data; zpool_handle_t *zhp = zhp_data; if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0) return (0); return (guid == zpool_vdev_path_to_guid(zhp, cb->cb_names[0])); } /* * Returns 1 if cb_data->cb_names[0] is a vdev name, 0 otherwise. */ static int is_vdev(zpool_handle_t *zhp, void *cb_data) { return (for_each_vdev(zhp, is_vdev_cb, cb_data)); } /* * Check if vdevs are in a pool * * Return 1 if all argv[] strings are vdev names in pool "pool_name". Otherwise * return 0. If pool_name is NULL, then search all pools. */ static int are_vdevs_in_pool(int argc, char **argv, char *pool_name, vdev_cbdata_t *cb) { char **tmp_name; int ret = 0; int i; int pool_count = 0; if ((argc == 0) || !*argv) return (0); if (pool_name) pool_count = 1; /* Temporarily hijack cb_names for a second... */ tmp_name = cb->cb_names; /* Go though our list of prospective vdev names */ for (i = 0; i < argc; i++) { cb->cb_names = argv + i; /* Is this name a vdev in our pools? */ ret = for_each_pool(pool_count, &pool_name, B_TRUE, NULL, ZFS_TYPE_POOL, B_FALSE, is_vdev, cb); if (!ret) { /* No match */ break; } } cb->cb_names = tmp_name; return (ret); } static int is_pool_cb(zpool_handle_t *zhp, void *data) { char *name = data; if (strcmp(name, zpool_get_name(zhp)) == 0) return (1); return (0); } /* * Do we have a pool named *name? If so, return 1, otherwise 0. */ static int is_pool(char *name) { return (for_each_pool(0, NULL, B_TRUE, NULL, ZFS_TYPE_POOL, B_FALSE, is_pool_cb, name)); } /* Are all our argv[] strings pool names? If so return 1, 0 otherwise. */ static int are_all_pools(int argc, char **argv) { if ((argc == 0) || !*argv) return (0); while (--argc >= 0) if (!is_pool(argv[argc])) return (0); return (1); } /* * Helper function to print out vdev/pool names we can't resolve. Used for an * error message. */ static void error_list_unresolved_vdevs(int argc, char **argv, char *pool_name, vdev_cbdata_t *cb) { int i; char *name; char *str; for (i = 0; i < argc; i++) { name = argv[i]; if (is_pool(name)) str = gettext("pool"); else if (are_vdevs_in_pool(1, &name, pool_name, cb)) str = gettext("vdev in this pool"); else if (are_vdevs_in_pool(1, &name, NULL, cb)) str = gettext("vdev in another pool"); else str = gettext("unknown"); fprintf(stderr, "\t%s (%s)\n", name, str); } } /* * Same as get_interval_count(), but with additional checks to not misinterpret * guids as interval/count values. Assumes VDEV_NAME_GUID is set in * cb.cb_vdevs.cb_name_flags. */ static void get_interval_count_filter_guids(int *argc, char **argv, float *interval, unsigned long *count, iostat_cbdata_t *cb) { char **tmpargv = argv; int argc_for_interval = 0; /* Is the last arg an interval value? Or a guid? */ if (*argc >= 1 && !are_vdevs_in_pool(1, &argv[*argc - 1], NULL, &cb->cb_vdevs)) { /* * The last arg is not a guid, so it's probably an * interval value. */ argc_for_interval++; if (*argc >= 2 && !are_vdevs_in_pool(1, &argv[*argc - 2], NULL, &cb->cb_vdevs)) { /* * The 2nd to last arg is not a guid, so it's probably * an interval value. */ argc_for_interval++; } } /* Point to our list of possible intervals */ tmpargv = &argv[*argc - argc_for_interval]; *argc = *argc - argc_for_interval; get_interval_count(&argc_for_interval, tmpargv, interval, count); } /* * Floating point sleep(). Allows you to pass in a floating point value for * seconds. */ static void fsleep(float sec) { struct timespec req; req.tv_sec = floor(sec); req.tv_nsec = (sec - (float)req.tv_sec) * NANOSEC; nanosleep(&req, NULL); } /* * Terminal height, in rows. Returns -1 if stdout is not connected to a TTY or * if we were unable to determine its size. */ static int terminal_height(void) { struct winsize win; if (isatty(STDOUT_FILENO) == 0) return (-1); if (ioctl(STDOUT_FILENO, TIOCGWINSZ, &win) != -1 && win.ws_row > 0) return (win.ws_row); return (-1); } /* * Run one of the zpool status/iostat -c scripts with the help (-h) option and * print the result. * * name: Short name of the script ('iostat'). * path: Full path to the script ('/usr/local/etc/zfs/zpool.d/iostat'); */ static void print_zpool_script_help(char *name, char *path) { char *argv[] = {path, (char *)"-h", NULL}; char **lines = NULL; int lines_cnt = 0; int rc; rc = libzfs_run_process_get_stdout_nopath(path, argv, NULL, &lines, &lines_cnt); if (rc != 0 || lines == NULL || lines_cnt <= 0) { if (lines != NULL) libzfs_free_str_array(lines, lines_cnt); return; } for (int i = 0; i < lines_cnt; i++) if (!is_blank_str(lines[i])) printf(" %-14s %s\n", name, lines[i]); libzfs_free_str_array(lines, lines_cnt); } /* * Go though the zpool status/iostat -c scripts in the user's path, run their * help option (-h), and print out the results. */ static void print_zpool_dir_scripts(char *dirpath) { DIR *dir; struct dirent *ent; char fullpath[MAXPATHLEN]; struct stat dir_stat; if ((dir = opendir(dirpath)) != NULL) { /* print all the files and directories within directory */ while ((ent = readdir(dir)) != NULL) { if (snprintf(fullpath, sizeof (fullpath), "%s/%s", dirpath, ent->d_name) >= sizeof (fullpath)) { (void) fprintf(stderr, gettext("internal error: " "ZPOOL_SCRIPTS_PATH too large.\n")); exit(1); } /* Print the scripts */ if (stat(fullpath, &dir_stat) == 0) if (dir_stat.st_mode & S_IXUSR && S_ISREG(dir_stat.st_mode)) print_zpool_script_help(ent->d_name, fullpath); } closedir(dir); } } /* * Print out help text for all zpool status/iostat -c scripts. */ static void print_zpool_script_list(const char *subcommand) { char *dir, *sp, *tmp; printf(gettext("Available 'zpool %s -c' commands:\n"), subcommand); sp = zpool_get_cmd_search_path(); if (sp == NULL) return; for (dir = strtok_r(sp, ":", &tmp); dir != NULL; dir = strtok_r(NULL, ":", &tmp)) print_zpool_dir_scripts(dir); free(sp); } /* * Set the minimum pool/vdev name column width. The width must be at least 10, * but may be as large as the column width - 42 so it still fits on one line. * NOTE: 42 is the width of the default capacity/operations/bandwidth output */ static int get_namewidth_iostat(zpool_handle_t *zhp, void *data) { iostat_cbdata_t *cb = data; int width, available_width; /* * get_namewidth() returns the maximum width of any name in that column * for any pool/vdev/device line that will be output. */ width = get_namewidth(zhp, cb->cb_namewidth, cb->cb_vdevs.cb_name_flags | VDEV_NAME_TYPE_ID, cb->cb_verbose); /* * The width we are calculating is the width of the header and also the * padding width for names that are less than maximum width. The stats * take up 42 characters, so the width available for names is: */ available_width = get_columns() - 42; /* * If the maximum width fits on a screen, then great! Make everything * line up by justifying all lines to the same width. If that max * width is larger than what's available, the name plus stats won't fit * on one line, and justifying to that width would cause every line to * wrap on the screen. We only want lines with long names to wrap. * Limit the padding to what won't wrap. */ if (width > available_width) width = available_width; /* * And regardless of whatever the screen width is (get_columns can * return 0 if the width is not known or less than 42 for a narrow * terminal) have the width be a minimum of 10. */ if (width < 10) width = 10; /* Save the calculated width */ cb->cb_namewidth = width; return (0); } /* * zpool iostat [[-c [script1,script2,...]] [-lq]|[-rw]] [-ghHLpPvy] [-n name] * [-T d|u] [[ pool ...]|[pool vdev ...]|[vdev ...]] * [interval [count]] * * -c CMD For each vdev, run command CMD * -g Display guid for individual vdev name. * -L Follow links when resolving vdev path name. * -P Display full path for vdev name. * -v Display statistics for individual vdevs * -h Display help * -p Display values in parsable (exact) format. * -H Scripted mode. Don't display headers, and separate properties * by a single tab. * -l Display average latency * -q Display queue depths * -w Display latency histograms * -r Display request size histogram * -T Display a timestamp in date(1) or Unix format * -n Only print headers once * * This command can be tricky because we want to be able to deal with pool * creation/destruction as well as vdev configuration changes. The bulk of this * processing is handled by the pool_list_* routines in zpool_iter.c. We rely * on pool_list_update() to detect the addition of new pools. Configuration * changes are all handled within libzfs. */ int zpool_do_iostat(int argc, char **argv) { int c; int ret; int npools; float interval = 0; unsigned long count = 0; int winheight = 24; zpool_list_t *list; boolean_t verbose = B_FALSE; boolean_t latency = B_FALSE, l_histo = B_FALSE, rq_histo = B_FALSE; boolean_t queues = B_FALSE, parsable = B_FALSE, scripted = B_FALSE; boolean_t omit_since_boot = B_FALSE; boolean_t guid = B_FALSE; boolean_t follow_links = B_FALSE; boolean_t full_name = B_FALSE; boolean_t headers_once = B_FALSE; iostat_cbdata_t cb = { 0 }; char *cmd = NULL; /* Used for printing error message */ const char flag_to_arg[] = {[IOS_LATENCY] = 'l', [IOS_QUEUES] = 'q', [IOS_L_HISTO] = 'w', [IOS_RQ_HISTO] = 'r'}; uint64_t unsupported_flags; /* check options */ while ((c = getopt(argc, argv, "c:gLPT:vyhplqrwnH")) != -1) { switch (c) { case 'c': if (cmd != NULL) { fprintf(stderr, gettext("Can't set -c flag twice\n")); exit(1); } if (getenv("ZPOOL_SCRIPTS_ENABLED") != NULL && !libzfs_envvar_is_set("ZPOOL_SCRIPTS_ENABLED")) { fprintf(stderr, gettext( "Can't run -c, disabled by " "ZPOOL_SCRIPTS_ENABLED.\n")); exit(1); } if ((getuid() <= 0 || geteuid() <= 0) && !libzfs_envvar_is_set("ZPOOL_SCRIPTS_AS_ROOT")) { fprintf(stderr, gettext( "Can't run -c with root privileges " "unless ZPOOL_SCRIPTS_AS_ROOT is set.\n")); exit(1); } cmd = optarg; verbose = B_TRUE; break; case 'g': guid = B_TRUE; break; case 'L': follow_links = B_TRUE; break; case 'P': full_name = B_TRUE; break; case 'T': get_timestamp_arg(*optarg); break; case 'v': verbose = B_TRUE; break; case 'p': parsable = B_TRUE; break; case 'l': latency = B_TRUE; break; case 'q': queues = B_TRUE; break; case 'H': scripted = B_TRUE; break; case 'w': l_histo = B_TRUE; break; case 'r': rq_histo = B_TRUE; break; case 'y': omit_since_boot = B_TRUE; break; case 'n': headers_once = B_TRUE; break; case 'h': usage(B_FALSE); break; case '?': if (optopt == 'c') { print_zpool_script_list("iostat"); exit(0); } else { fprintf(stderr, gettext("invalid option '%c'\n"), optopt); } usage(B_FALSE); } } argc -= optind; argv += optind; cb.cb_literal = parsable; cb.cb_scripted = scripted; if (guid) cb.cb_vdevs.cb_name_flags |= VDEV_NAME_GUID; if (follow_links) cb.cb_vdevs.cb_name_flags |= VDEV_NAME_FOLLOW_LINKS; if (full_name) cb.cb_vdevs.cb_name_flags |= VDEV_NAME_PATH; cb.cb_iteration = 0; cb.cb_namewidth = 0; cb.cb_verbose = verbose; /* Get our interval and count values (if any) */ if (guid) { get_interval_count_filter_guids(&argc, argv, &interval, &count, &cb); } else { get_interval_count(&argc, argv, &interval, &count); } if (argc == 0) { /* No args, so just print the defaults. */ } else if (are_all_pools(argc, argv)) { /* All the args are pool names */ } else if (are_vdevs_in_pool(argc, argv, NULL, &cb.cb_vdevs)) { /* All the args are vdevs */ cb.cb_vdevs.cb_names = argv; cb.cb_vdevs.cb_names_count = argc; argc = 0; /* No pools to process */ } else if (are_all_pools(1, argv)) { /* The first arg is a pool name */ if (are_vdevs_in_pool(argc - 1, argv + 1, argv[0], &cb.cb_vdevs)) { /* ...and the rest are vdev names */ cb.cb_vdevs.cb_names = argv + 1; cb.cb_vdevs.cb_names_count = argc - 1; argc = 1; /* One pool to process */ } else { fprintf(stderr, gettext("Expected either a list of ")); fprintf(stderr, gettext("pools, or list of vdevs in")); fprintf(stderr, " \"%s\", ", argv[0]); fprintf(stderr, gettext("but got:\n")); error_list_unresolved_vdevs(argc - 1, argv + 1, argv[0], &cb.cb_vdevs); fprintf(stderr, "\n"); usage(B_FALSE); return (1); } } else { /* * The args don't make sense. The first arg isn't a pool name, * nor are all the args vdevs. */ fprintf(stderr, gettext("Unable to parse pools/vdevs list.\n")); fprintf(stderr, "\n"); return (1); } if (cb.cb_vdevs.cb_names_count != 0) { /* * If user specified vdevs, it implies verbose. */ cb.cb_verbose = B_TRUE; } /* * Construct the list of all interesting pools. */ ret = 0; if ((list = pool_list_get(argc, argv, NULL, ZFS_TYPE_POOL, parsable, &ret)) == NULL) return (1); if (pool_list_count(list) == 0 && argc != 0) { pool_list_free(list); return (1); } if (pool_list_count(list) == 0 && interval == 0) { pool_list_free(list); (void) fprintf(stderr, gettext("no pools available\n")); return (1); } if ((l_histo || rq_histo) && (cmd != NULL || latency || queues)) { pool_list_free(list); (void) fprintf(stderr, gettext("[-r|-w] isn't allowed with [-c|-l|-q]\n")); usage(B_FALSE); return (1); } if (l_histo && rq_histo) { pool_list_free(list); (void) fprintf(stderr, gettext("Only one of [-r|-w] can be passed at a time\n")); usage(B_FALSE); return (1); } /* * Enter the main iostat loop. */ cb.cb_list = list; if (l_histo) { /* * Histograms tables look out of place when you try to display * them with the other stats, so make a rule that you can only * print histograms by themselves. */ cb.cb_flags = IOS_L_HISTO_M; } else if (rq_histo) { cb.cb_flags = IOS_RQ_HISTO_M; } else { cb.cb_flags = IOS_DEFAULT_M; if (latency) cb.cb_flags |= IOS_LATENCY_M; if (queues) cb.cb_flags |= IOS_QUEUES_M; } /* * See if the module supports all the stats we want to display. */ unsupported_flags = cb.cb_flags & ~get_stat_flags(list); if (unsupported_flags) { uint64_t f; int idx; fprintf(stderr, gettext("The loaded zfs module doesn't support:")); /* for each bit set in unsupported_flags */ for (f = unsupported_flags; f; f &= ~(1ULL << idx)) { idx = lowbit64(f) - 1; fprintf(stderr, " -%c", flag_to_arg[idx]); } fprintf(stderr, ". Try running a newer module.\n"); pool_list_free(list); return (1); } for (;;) { if ((npools = pool_list_count(list)) == 0) (void) fprintf(stderr, gettext("no pools available\n")); else { /* * If this is the first iteration and -y was supplied * we skip any printing. */ boolean_t skip = (omit_since_boot && cb.cb_iteration == 0); /* * Refresh all statistics. This is done as an * explicit step before calculating the maximum name * width, so that any * configuration changes are * properly accounted for. */ (void) pool_list_iter(list, B_FALSE, refresh_iostat, &cb); /* * Iterate over all pools to determine the maximum width * for the pool / device name column across all pools. */ cb.cb_namewidth = 0; (void) pool_list_iter(list, B_FALSE, get_namewidth_iostat, &cb); if (timestamp_fmt != NODATE) print_timestamp(timestamp_fmt); if (cmd != NULL && cb.cb_verbose && !(cb.cb_flags & IOS_ANYHISTO_M)) { cb.vcdl = all_pools_for_each_vdev_run(argc, argv, cmd, g_zfs, cb.cb_vdevs.cb_names, cb.cb_vdevs.cb_names_count, cb.cb_vdevs.cb_name_flags); } else { cb.vcdl = NULL; } /* * Check terminal size so we can print headers * even when terminal window has its height * changed. */ winheight = terminal_height(); /* * Are we connected to TTY? If not, headers_once * should be true, to avoid breaking scripts. */ if (winheight < 0) headers_once = B_TRUE; /* * If it's the first time and we're not skipping it, * or either skip or verbose mode, print the header. * * The histogram code explicitly prints its header on * every vdev, so skip this for histograms. */ if (((++cb.cb_iteration == 1 && !skip) || (skip != verbose) || (!headers_once && (cb.cb_iteration % winheight) == 0)) && (!(cb.cb_flags & IOS_ANYHISTO_M)) && !cb.cb_scripted) print_iostat_header(&cb); if (skip) { (void) fsleep(interval); continue; } pool_list_iter(list, B_FALSE, print_iostat, &cb); /* * If there's more than one pool, and we're not in * verbose mode (which prints a separator for us), * then print a separator. * * In addition, if we're printing specific vdevs then * we also want an ending separator. */ if (((npools > 1 && !verbose && !(cb.cb_flags & IOS_ANYHISTO_M)) || (!(cb.cb_flags & IOS_ANYHISTO_M) && cb.cb_vdevs.cb_names_count)) && !cb.cb_scripted) { print_iostat_separator(&cb); if (cb.vcdl != NULL) print_cmd_columns(cb.vcdl, 1); printf("\n"); } if (cb.vcdl != NULL) free_vdev_cmd_data_list(cb.vcdl); } /* * Flush the output so that redirection to a file isn't buffered * indefinitely. */ (void) fflush(stdout); if (interval == 0) break; if (count != 0 && --count == 0) break; (void) fsleep(interval); } pool_list_free(list); return (ret); } typedef struct list_cbdata { boolean_t cb_verbose; int cb_name_flags; int cb_namewidth; boolean_t cb_scripted; zprop_list_t *cb_proplist; boolean_t cb_literal; } list_cbdata_t; /* * Given a list of columns to display, output appropriate headers for each one. */ static void print_header(list_cbdata_t *cb) { zprop_list_t *pl = cb->cb_proplist; char headerbuf[ZPOOL_MAXPROPLEN]; const char *header; boolean_t first = B_TRUE; boolean_t right_justify; size_t width = 0; for (; pl != NULL; pl = pl->pl_next) { width = pl->pl_width; if (first && cb->cb_verbose) { /* * Reset the width to accommodate the verbose listing * of devices. */ width = cb->cb_namewidth; } if (!first) (void) fputs(" ", stdout); else first = B_FALSE; right_justify = B_FALSE; if (pl->pl_prop != ZPROP_USERPROP) { header = zpool_prop_column_name(pl->pl_prop); right_justify = zpool_prop_align_right(pl->pl_prop); } else { int i; for (i = 0; pl->pl_user_prop[i] != '\0'; i++) headerbuf[i] = toupper(pl->pl_user_prop[i]); headerbuf[i] = '\0'; header = headerbuf; } if (pl->pl_next == NULL && !right_justify) (void) fputs(header, stdout); else if (right_justify) (void) printf("%*s", (int)width, header); else (void) printf("%-*s", (int)width, header); } (void) fputc('\n', stdout); } /* * Given a pool and a list of properties, print out all the properties according * to the described layout. Used by zpool_do_list(). */ static void print_pool(zpool_handle_t *zhp, list_cbdata_t *cb) { zprop_list_t *pl = cb->cb_proplist; boolean_t first = B_TRUE; char property[ZPOOL_MAXPROPLEN]; const char *propstr; boolean_t right_justify; size_t width; for (; pl != NULL; pl = pl->pl_next) { width = pl->pl_width; if (first && cb->cb_verbose) { /* * Reset the width to accommodate the verbose listing * of devices. */ width = cb->cb_namewidth; } if (!first) { if (cb->cb_scripted) (void) fputc('\t', stdout); else (void) fputs(" ", stdout); } else { first = B_FALSE; } right_justify = B_FALSE; if (pl->pl_prop != ZPROP_USERPROP) { if (zpool_get_prop(zhp, pl->pl_prop, property, sizeof (property), NULL, cb->cb_literal) != 0) propstr = "-"; else propstr = property; right_justify = zpool_prop_align_right(pl->pl_prop); } else if ((zpool_prop_feature(pl->pl_user_prop) || zpool_prop_unsupported(pl->pl_user_prop)) && zpool_prop_get_feature(zhp, pl->pl_user_prop, property, sizeof (property)) == 0) { propstr = property; } else if (zfs_prop_user(pl->pl_user_prop) && zpool_get_userprop(zhp, pl->pl_user_prop, property, sizeof (property), NULL) == 0) { propstr = property; } else { propstr = "-"; } /* * If this is being called in scripted mode, or if this is the * last column and it is left-justified, don't include a width * format specifier. */ if (cb->cb_scripted || (pl->pl_next == NULL && !right_justify)) (void) fputs(propstr, stdout); else if (right_justify) (void) printf("%*s", (int)width, propstr); else (void) printf("%-*s", (int)width, propstr); } (void) fputc('\n', stdout); } static void print_one_column(zpool_prop_t prop, uint64_t value, const char *str, boolean_t scripted, boolean_t valid, enum zfs_nicenum_format format) { char propval[64]; boolean_t fixed; size_t width = zprop_width(prop, &fixed, ZFS_TYPE_POOL); switch (prop) { case ZPOOL_PROP_SIZE: case ZPOOL_PROP_EXPANDSZ: case ZPOOL_PROP_CHECKPOINT: case ZPOOL_PROP_DEDUPRATIO: if (value == 0) (void) strlcpy(propval, "-", sizeof (propval)); else zfs_nicenum_format(value, propval, sizeof (propval), format); break; case ZPOOL_PROP_FRAGMENTATION: if (value == ZFS_FRAG_INVALID) { (void) strlcpy(propval, "-", sizeof (propval)); } else if (format == ZFS_NICENUM_RAW) { (void) snprintf(propval, sizeof (propval), "%llu", (unsigned long long)value); } else { (void) snprintf(propval, sizeof (propval), "%llu%%", (unsigned long long)value); } break; case ZPOOL_PROP_CAPACITY: /* capacity value is in parts-per-10,000 (aka permyriad) */ if (format == ZFS_NICENUM_RAW) (void) snprintf(propval, sizeof (propval), "%llu", (unsigned long long)value / 100); else (void) snprintf(propval, sizeof (propval), value < 1000 ? "%1.2f%%" : value < 10000 ? "%2.1f%%" : "%3.0f%%", value / 100.0); break; case ZPOOL_PROP_HEALTH: width = 8; (void) strlcpy(propval, str, sizeof (propval)); break; default: zfs_nicenum_format(value, propval, sizeof (propval), format); } if (!valid) (void) strlcpy(propval, "-", sizeof (propval)); if (scripted) (void) printf("\t%s", propval); else (void) printf(" %*s", (int)width, propval); } /* * print static default line per vdev * not compatible with '-o' option */ static void print_list_stats(zpool_handle_t *zhp, const char *name, nvlist_t *nv, list_cbdata_t *cb, int depth, boolean_t isspare) { nvlist_t **child; vdev_stat_t *vs; uint_t c, children; char *vname; boolean_t scripted = cb->cb_scripted; uint64_t islog = B_FALSE; const char *dashes = "%-*s - - - - " "- - - - -\n"; verify(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &c) == 0); if (name != NULL) { boolean_t toplevel = (vs->vs_space != 0); uint64_t cap; enum zfs_nicenum_format format; const char *state; if (cb->cb_literal) format = ZFS_NICENUM_RAW; else format = ZFS_NICENUM_1024; if (strcmp(name, VDEV_TYPE_INDIRECT) == 0) return; if (scripted) (void) printf("\t%s", name); else if (strlen(name) + depth > cb->cb_namewidth) (void) printf("%*s%s", depth, "", name); else (void) printf("%*s%s%*s", depth, "", name, (int)(cb->cb_namewidth - strlen(name) - depth), ""); /* * Print the properties for the individual vdevs. Some * properties are only applicable to toplevel vdevs. The * 'toplevel' boolean value is passed to the print_one_column() * to indicate that the value is valid. */ if (VDEV_STAT_VALID(vs_pspace, c) && vs->vs_pspace) print_one_column(ZPOOL_PROP_SIZE, vs->vs_pspace, NULL, scripted, B_TRUE, format); else print_one_column(ZPOOL_PROP_SIZE, vs->vs_space, NULL, scripted, toplevel, format); print_one_column(ZPOOL_PROP_ALLOCATED, vs->vs_alloc, NULL, scripted, toplevel, format); print_one_column(ZPOOL_PROP_FREE, vs->vs_space - vs->vs_alloc, NULL, scripted, toplevel, format); print_one_column(ZPOOL_PROP_CHECKPOINT, vs->vs_checkpoint_space, NULL, scripted, toplevel, format); print_one_column(ZPOOL_PROP_EXPANDSZ, vs->vs_esize, NULL, scripted, B_TRUE, format); print_one_column(ZPOOL_PROP_FRAGMENTATION, vs->vs_fragmentation, NULL, scripted, (vs->vs_fragmentation != ZFS_FRAG_INVALID && toplevel), format); cap = (vs->vs_space == 0) ? 0 : (vs->vs_alloc * 10000 / vs->vs_space); print_one_column(ZPOOL_PROP_CAPACITY, cap, NULL, scripted, toplevel, format); print_one_column(ZPOOL_PROP_DEDUPRATIO, 0, NULL, scripted, toplevel, format); state = zpool_state_to_name(vs->vs_state, vs->vs_aux); if (isspare) { if (vs->vs_aux == VDEV_AUX_SPARED) state = "INUSE"; else if (vs->vs_state == VDEV_STATE_HEALTHY) state = "AVAIL"; } print_one_column(ZPOOL_PROP_HEALTH, 0, state, scripted, B_TRUE, format); (void) fputc('\n', stdout); } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) return; /* list the normal vdevs first */ for (c = 0; c < children; c++) { uint64_t ishole = B_FALSE; if (nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_HOLE, &ishole) == 0 && ishole) continue; if (nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, &islog) == 0 && islog) continue; if (nvlist_exists(child[c], ZPOOL_CONFIG_ALLOCATION_BIAS)) continue; vname = zpool_vdev_name(g_zfs, zhp, child[c], cb->cb_name_flags | VDEV_NAME_TYPE_ID); print_list_stats(zhp, vname, child[c], cb, depth + 2, B_FALSE); free(vname); } /* list the classes: 'logs', 'dedup', and 'special' */ for (uint_t n = 0; n < ARRAY_SIZE(class_name); n++) { boolean_t printed = B_FALSE; for (c = 0; c < children; c++) { const char *bias = NULL; const char *type = NULL; if (nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, &islog) == 0 && islog) { bias = VDEV_ALLOC_CLASS_LOGS; } else { (void) nvlist_lookup_string(child[c], ZPOOL_CONFIG_ALLOCATION_BIAS, &bias); (void) nvlist_lookup_string(child[c], ZPOOL_CONFIG_TYPE, &type); } if (bias == NULL || strcmp(bias, class_name[n]) != 0) continue; if (!islog && strcmp(type, VDEV_TYPE_INDIRECT) == 0) continue; if (!printed) { /* LINTED E_SEC_PRINTF_VAR_FMT */ (void) printf(dashes, cb->cb_namewidth, class_name[n]); printed = B_TRUE; } vname = zpool_vdev_name(g_zfs, zhp, child[c], cb->cb_name_flags | VDEV_NAME_TYPE_ID); print_list_stats(zhp, vname, child[c], cb, depth + 2, B_FALSE); free(vname); } } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE, &child, &children) == 0 && children > 0) { /* LINTED E_SEC_PRINTF_VAR_FMT */ (void) printf(dashes, cb->cb_namewidth, "cache"); for (c = 0; c < children; c++) { vname = zpool_vdev_name(g_zfs, zhp, child[c], cb->cb_name_flags); print_list_stats(zhp, vname, child[c], cb, depth + 2, B_FALSE); free(vname); } } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES, &child, &children) == 0 && children > 0) { /* LINTED E_SEC_PRINTF_VAR_FMT */ (void) printf(dashes, cb->cb_namewidth, "spare"); for (c = 0; c < children; c++) { vname = zpool_vdev_name(g_zfs, zhp, child[c], cb->cb_name_flags); print_list_stats(zhp, vname, child[c], cb, depth + 2, B_TRUE); free(vname); } } } /* * Generic callback function to list a pool. */ static int list_callback(zpool_handle_t *zhp, void *data) { list_cbdata_t *cbp = data; print_pool(zhp, cbp); if (cbp->cb_verbose) { nvlist_t *config, *nvroot; config = zpool_get_config(zhp, NULL); verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); print_list_stats(zhp, NULL, nvroot, cbp, 0, B_FALSE); } return (0); } /* * Set the minimum pool/vdev name column width. The width must be at least 9, * but may be as large as needed. */ static int get_namewidth_list(zpool_handle_t *zhp, void *data) { list_cbdata_t *cb = data; int width; width = get_namewidth(zhp, cb->cb_namewidth, cb->cb_name_flags | VDEV_NAME_TYPE_ID, cb->cb_verbose); if (width < 9) width = 9; cb->cb_namewidth = width; return (0); } /* * zpool list [-gHLpP] [-o prop[,prop]*] [-T d|u] [pool] ... [interval [count]] * * -g Display guid for individual vdev name. * -H Scripted mode. Don't display headers, and separate properties * by a single tab. * -L Follow links when resolving vdev path name. * -o List of properties to display. Defaults to * "name,size,allocated,free,expandsize,fragmentation,capacity," * "dedupratio,health,altroot" * -p Display values in parsable (exact) format. * -P Display full path for vdev name. * -T Display a timestamp in date(1) or Unix format * * List all pools in the system, whether or not they're healthy. Output space * statistics for each one, as well as health status summary. */ int zpool_do_list(int argc, char **argv) { int c; int ret = 0; list_cbdata_t cb = { 0 }; static char default_props[] = "name,size,allocated,free,checkpoint,expandsize,fragmentation," "capacity,dedupratio,health,altroot"; char *props = default_props; float interval = 0; unsigned long count = 0; zpool_list_t *list; boolean_t first = B_TRUE; current_prop_type = ZFS_TYPE_POOL; /* check options */ while ((c = getopt(argc, argv, ":gHLo:pPT:v")) != -1) { switch (c) { case 'g': cb.cb_name_flags |= VDEV_NAME_GUID; break; case 'H': cb.cb_scripted = B_TRUE; break; case 'L': cb.cb_name_flags |= VDEV_NAME_FOLLOW_LINKS; break; case 'o': props = optarg; break; case 'P': cb.cb_name_flags |= VDEV_NAME_PATH; break; case 'p': cb.cb_literal = B_TRUE; break; case 'T': get_timestamp_arg(*optarg); break; case 'v': cb.cb_verbose = B_TRUE; cb.cb_namewidth = 8; /* 8 until precalc is avail */ break; case ':': (void) fprintf(stderr, gettext("missing argument for " "'%c' option\n"), optopt); usage(B_FALSE); break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; get_interval_count(&argc, argv, &interval, &count); if (zprop_get_list(g_zfs, props, &cb.cb_proplist, ZFS_TYPE_POOL) != 0) usage(B_FALSE); for (;;) { if ((list = pool_list_get(argc, argv, &cb.cb_proplist, ZFS_TYPE_POOL, cb.cb_literal, &ret)) == NULL) return (1); if (pool_list_count(list) == 0) break; cb.cb_namewidth = 0; (void) pool_list_iter(list, B_FALSE, get_namewidth_list, &cb); if (timestamp_fmt != NODATE) print_timestamp(timestamp_fmt); if (!cb.cb_scripted && (first || cb.cb_verbose)) { print_header(&cb); first = B_FALSE; } ret = pool_list_iter(list, B_TRUE, list_callback, &cb); if (interval == 0) break; if (count != 0 && --count == 0) break; pool_list_free(list); (void) fsleep(interval); } if (argc == 0 && !cb.cb_scripted && pool_list_count(list) == 0) { (void) printf(gettext("no pools available\n")); ret = 0; } pool_list_free(list); zprop_free_list(cb.cb_proplist); return (ret); } static int zpool_do_attach_or_replace(int argc, char **argv, int replacing) { boolean_t force = B_FALSE; boolean_t rebuild = B_FALSE; boolean_t wait = B_FALSE; int c; nvlist_t *nvroot; char *poolname, *old_disk, *new_disk; zpool_handle_t *zhp; nvlist_t *props = NULL; char *propval; int ret; /* check options */ while ((c = getopt(argc, argv, "fo:sw")) != -1) { switch (c) { case 'f': force = B_TRUE; break; case 'o': if ((propval = strchr(optarg, '=')) == NULL) { (void) fprintf(stderr, gettext("missing " "'=' for -o option\n")); usage(B_FALSE); } *propval = '\0'; propval++; if ((strcmp(optarg, ZPOOL_CONFIG_ASHIFT) != 0) || (add_prop_list(optarg, propval, &props, B_TRUE))) usage(B_FALSE); break; case 's': rebuild = B_TRUE; break; case 'w': wait = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* get pool name and check number of arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name argument\n")); usage(B_FALSE); } poolname = argv[0]; if (argc < 2) { (void) fprintf(stderr, gettext("missing specification\n")); usage(B_FALSE); } old_disk = argv[1]; if (argc < 3) { if (!replacing) { (void) fprintf(stderr, gettext("missing specification\n")); usage(B_FALSE); } new_disk = old_disk; argc -= 1; argv += 1; } else { new_disk = argv[2]; argc -= 2; argv += 2; } if (argc > 1) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } if ((zhp = zpool_open(g_zfs, poolname)) == NULL) { nvlist_free(props); return (1); } if (zpool_get_config(zhp, NULL) == NULL) { (void) fprintf(stderr, gettext("pool '%s' is unavailable\n"), poolname); zpool_close(zhp); nvlist_free(props); return (1); } /* unless manually specified use "ashift" pool property (if set) */ if (!nvlist_exists(props, ZPOOL_CONFIG_ASHIFT)) { int intval; zprop_source_t src; char strval[ZPOOL_MAXPROPLEN]; intval = zpool_get_prop_int(zhp, ZPOOL_PROP_ASHIFT, &src); if (src != ZPROP_SRC_DEFAULT) { (void) sprintf(strval, "%" PRId32, intval); verify(add_prop_list(ZPOOL_CONFIG_ASHIFT, strval, &props, B_TRUE) == 0); } } nvroot = make_root_vdev(zhp, props, force, B_FALSE, replacing, B_FALSE, argc, argv); if (nvroot == NULL) { zpool_close(zhp); nvlist_free(props); return (1); } ret = zpool_vdev_attach(zhp, old_disk, new_disk, nvroot, replacing, rebuild); if (ret == 0 && wait) ret = zpool_wait(zhp, replacing ? ZPOOL_WAIT_REPLACE : ZPOOL_WAIT_RESILVER); nvlist_free(props); nvlist_free(nvroot); zpool_close(zhp); return (ret); } /* * zpool replace [-fsw] [-o property=value] * * -f Force attach, even if appears to be in use. * -s Use sequential instead of healing reconstruction for resilver. * -o Set property=value. * -w Wait for replacing to complete before returning * * Replace with . */ int zpool_do_replace(int argc, char **argv) { return (zpool_do_attach_or_replace(argc, argv, B_TRUE)); } /* * zpool attach [-fsw] [-o property=value] * * -f Force attach, even if appears to be in use. * -s Use sequential instead of healing reconstruction for resilver. * -o Set property=value. * -w Wait for resilvering to complete before returning * * Attach to the mirror containing . If is not * part of a mirror, then will be transformed into a mirror of * and . In either case, will begin life * with a DTL of [0, now], and will immediately begin to resilver itself. */ int zpool_do_attach(int argc, char **argv) { return (zpool_do_attach_or_replace(argc, argv, B_FALSE)); } /* * zpool detach [-f] * * -f Force detach of , even if DTLs argue against it * (not supported yet) * * Detach a device from a mirror. The operation will be refused if * is the last device in the mirror, or if the DTLs indicate that this device * has the only valid copy of some data. */ int zpool_do_detach(int argc, char **argv) { int c; char *poolname, *path; zpool_handle_t *zhp; int ret; /* check options */ while ((c = getopt(argc, argv, "")) != -1) { switch (c) { case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* get pool name and check number of arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name argument\n")); usage(B_FALSE); } if (argc < 2) { (void) fprintf(stderr, gettext("missing specification\n")); usage(B_FALSE); } poolname = argv[0]; path = argv[1]; if ((zhp = zpool_open(g_zfs, poolname)) == NULL) return (1); ret = zpool_vdev_detach(zhp, path); zpool_close(zhp); return (ret); } /* * zpool split [-gLnP] [-o prop=val] ... * [-o mntopt] ... * [-R altroot] [ ...] * * -g Display guid for individual vdev name. * -L Follow links when resolving vdev path name. * -n Do not split the pool, but display the resulting layout if * it were to be split. * -o Set property=value, or set mount options. * -P Display full path for vdev name. * -R Mount the split-off pool under an alternate root. * -l Load encryption keys while importing. * * Splits the named pool and gives it the new pool name. Devices to be split * off may be listed, provided that no more than one device is specified * per top-level vdev mirror. The newly split pool is left in an exported * state unless -R is specified. * * Restrictions: the top-level of the pool pool must only be made up of * mirrors; all devices in the pool must be healthy; no device may be * undergoing a resilvering operation. */ int zpool_do_split(int argc, char **argv) { char *srcpool, *newpool, *propval; char *mntopts = NULL; splitflags_t flags; int c, ret = 0; boolean_t loadkeys = B_FALSE; zpool_handle_t *zhp; nvlist_t *config, *props = NULL; flags.dryrun = B_FALSE; flags.import = B_FALSE; flags.name_flags = 0; /* check options */ while ((c = getopt(argc, argv, ":gLR:lno:P")) != -1) { switch (c) { case 'g': flags.name_flags |= VDEV_NAME_GUID; break; case 'L': flags.name_flags |= VDEV_NAME_FOLLOW_LINKS; break; case 'R': flags.import = B_TRUE; if (add_prop_list( zpool_prop_to_name(ZPOOL_PROP_ALTROOT), optarg, &props, B_TRUE) != 0) { nvlist_free(props); usage(B_FALSE); } break; case 'l': loadkeys = B_TRUE; break; case 'n': flags.dryrun = B_TRUE; break; case 'o': if ((propval = strchr(optarg, '=')) != NULL) { *propval = '\0'; propval++; if (add_prop_list(optarg, propval, &props, B_TRUE) != 0) { nvlist_free(props); usage(B_FALSE); } } else { mntopts = optarg; } break; case 'P': flags.name_flags |= VDEV_NAME_PATH; break; case ':': (void) fprintf(stderr, gettext("missing argument for " "'%c' option\n"), optopt); usage(B_FALSE); break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); break; } } if (!flags.import && mntopts != NULL) { (void) fprintf(stderr, gettext("setting mntopts is only " "valid when importing the pool\n")); usage(B_FALSE); } if (!flags.import && loadkeys) { (void) fprintf(stderr, gettext("loading keys is only " "valid when importing the pool\n")); usage(B_FALSE); } argc -= optind; argv += optind; if (argc < 1) { (void) fprintf(stderr, gettext("Missing pool name\n")); usage(B_FALSE); } if (argc < 2) { (void) fprintf(stderr, gettext("Missing new pool name\n")); usage(B_FALSE); } srcpool = argv[0]; newpool = argv[1]; argc -= 2; argv += 2; if ((zhp = zpool_open(g_zfs, srcpool)) == NULL) { nvlist_free(props); return (1); } config = split_mirror_vdev(zhp, newpool, props, flags, argc, argv); if (config == NULL) { ret = 1; } else { if (flags.dryrun) { (void) printf(gettext("would create '%s' with the " "following layout:\n\n"), newpool); print_vdev_tree(NULL, newpool, config, 0, "", flags.name_flags); print_vdev_tree(NULL, "dedup", config, 0, VDEV_ALLOC_BIAS_DEDUP, 0); print_vdev_tree(NULL, "special", config, 0, VDEV_ALLOC_BIAS_SPECIAL, 0); } } zpool_close(zhp); if (ret != 0 || flags.dryrun || !flags.import) { nvlist_free(config); nvlist_free(props); return (ret); } /* * The split was successful. Now we need to open the new * pool and import it. */ if ((zhp = zpool_open_canfail(g_zfs, newpool)) == NULL) { nvlist_free(config); nvlist_free(props); return (1); } if (loadkeys) { ret = zfs_crypto_attempt_load_keys(g_zfs, newpool); if (ret != 0) ret = 1; } if (zpool_get_state(zhp) != POOL_STATE_UNAVAIL && zpool_enable_datasets(zhp, mntopts, 0) != 0) { ret = 1; (void) fprintf(stderr, gettext("Split was successful, but " "the datasets could not all be mounted\n")); (void) fprintf(stderr, gettext("Try doing '%s' with a " "different altroot\n"), "zpool import"); } zpool_close(zhp); nvlist_free(config); nvlist_free(props); return (ret); } /* * zpool online ... */ int zpool_do_online(int argc, char **argv) { int c, i; char *poolname; zpool_handle_t *zhp; int ret = 0; vdev_state_t newstate; int flags = 0; /* check options */ while ((c = getopt(argc, argv, "e")) != -1) { switch (c) { case 'e': flags |= ZFS_ONLINE_EXPAND; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* get pool name and check number of arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name\n")); usage(B_FALSE); } if (argc < 2) { (void) fprintf(stderr, gettext("missing device name\n")); usage(B_FALSE); } poolname = argv[0]; if ((zhp = zpool_open(g_zfs, poolname)) == NULL) return (1); for (i = 1; i < argc; i++) { vdev_state_t oldstate; boolean_t avail_spare, l2cache; nvlist_t *tgt = zpool_find_vdev(zhp, argv[i], &avail_spare, &l2cache, NULL); if (tgt == NULL) { ret = 1; continue; } uint_t vsc; oldstate = ((vdev_stat_t *)fnvlist_lookup_uint64_array(tgt, ZPOOL_CONFIG_VDEV_STATS, &vsc))->vs_state; if (zpool_vdev_online(zhp, argv[i], flags, &newstate) == 0) { if (newstate != VDEV_STATE_HEALTHY) { (void) printf(gettext("warning: device '%s' " "onlined, but remains in faulted state\n"), argv[i]); if (newstate == VDEV_STATE_FAULTED) (void) printf(gettext("use 'zpool " "clear' to restore a faulted " "device\n")); else (void) printf(gettext("use 'zpool " "replace' to replace devices " "that are no longer present\n")); if ((flags & ZFS_ONLINE_EXPAND)) { (void) printf(gettext("%s: failed " "to expand usable space on " "unhealthy device '%s'\n"), (oldstate >= VDEV_STATE_DEGRADED ? "error" : "warning"), argv[i]); if (oldstate >= VDEV_STATE_DEGRADED) { ret = 1; break; } } } } else { ret = 1; } } zpool_close(zhp); return (ret); } /* * zpool offline [-ft] ... * * -f Force the device into a faulted state. * * -t Only take the device off-line temporarily. The offline/faulted * state will not be persistent across reboots. */ int zpool_do_offline(int argc, char **argv) { int c, i; char *poolname; zpool_handle_t *zhp; int ret = 0; boolean_t istmp = B_FALSE; boolean_t fault = B_FALSE; /* check options */ while ((c = getopt(argc, argv, "ft")) != -1) { switch (c) { case 'f': fault = B_TRUE; break; case 't': istmp = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* get pool name and check number of arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name\n")); usage(B_FALSE); } if (argc < 2) { (void) fprintf(stderr, gettext("missing device name\n")); usage(B_FALSE); } poolname = argv[0]; if ((zhp = zpool_open(g_zfs, poolname)) == NULL) return (1); for (i = 1; i < argc; i++) { if (fault) { uint64_t guid = zpool_vdev_path_to_guid(zhp, argv[i]); vdev_aux_t aux; if (istmp == B_FALSE) { /* Force the fault to persist across imports */ aux = VDEV_AUX_EXTERNAL_PERSIST; } else { aux = VDEV_AUX_EXTERNAL; } if (guid == 0 || zpool_vdev_fault(zhp, guid, aux) != 0) ret = 1; } else { if (zpool_vdev_offline(zhp, argv[i], istmp) != 0) ret = 1; } } zpool_close(zhp); return (ret); } /* * zpool clear [device] * * Clear all errors associated with a pool or a particular device. */ int zpool_do_clear(int argc, char **argv) { int c; int ret = 0; boolean_t dryrun = B_FALSE; boolean_t do_rewind = B_FALSE; boolean_t xtreme_rewind = B_FALSE; uint32_t rewind_policy = ZPOOL_NO_REWIND; nvlist_t *policy = NULL; zpool_handle_t *zhp; char *pool, *device; /* check options */ while ((c = getopt(argc, argv, "FnX")) != -1) { switch (c) { case 'F': do_rewind = B_TRUE; break; case 'n': dryrun = B_TRUE; break; case 'X': xtreme_rewind = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name\n")); usage(B_FALSE); } if (argc > 2) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } if ((dryrun || xtreme_rewind) && !do_rewind) { (void) fprintf(stderr, gettext("-n or -X only meaningful with -F\n")); usage(B_FALSE); } if (dryrun) rewind_policy = ZPOOL_TRY_REWIND; else if (do_rewind) rewind_policy = ZPOOL_DO_REWIND; if (xtreme_rewind) rewind_policy |= ZPOOL_EXTREME_REWIND; /* In future, further rewind policy choices can be passed along here */ if (nvlist_alloc(&policy, NV_UNIQUE_NAME, 0) != 0 || nvlist_add_uint32(policy, ZPOOL_LOAD_REWIND_POLICY, rewind_policy) != 0) { return (1); } pool = argv[0]; device = argc == 2 ? argv[1] : NULL; if ((zhp = zpool_open_canfail(g_zfs, pool)) == NULL) { nvlist_free(policy); return (1); } if (zpool_clear(zhp, device, policy) != 0) ret = 1; zpool_close(zhp); nvlist_free(policy); return (ret); } /* * zpool reguid */ int zpool_do_reguid(int argc, char **argv) { int c; char *poolname; zpool_handle_t *zhp; int ret = 0; /* check options */ while ((c = getopt(argc, argv, "")) != -1) { switch (c) { case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* get pool name and check number of arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name\n")); usage(B_FALSE); } if (argc > 1) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } poolname = argv[0]; if ((zhp = zpool_open(g_zfs, poolname)) == NULL) return (1); ret = zpool_reguid(zhp); zpool_close(zhp); return (ret); } /* * zpool reopen * * Reopen the pool so that the kernel can update the sizes of all vdevs. */ int zpool_do_reopen(int argc, char **argv) { int c; int ret = 0; boolean_t scrub_restart = B_TRUE; /* check options */ while ((c = getopt(argc, argv, "n")) != -1) { switch (c) { case 'n': scrub_restart = B_FALSE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; /* if argc == 0 we will execute zpool_reopen_one on all pools */ ret = for_each_pool(argc, argv, B_TRUE, NULL, ZFS_TYPE_POOL, B_FALSE, zpool_reopen_one, &scrub_restart); return (ret); } typedef struct scrub_cbdata { int cb_type; pool_scrub_cmd_t cb_scrub_cmd; } scrub_cbdata_t; static boolean_t zpool_has_checkpoint(zpool_handle_t *zhp) { nvlist_t *config, *nvroot; config = zpool_get_config(zhp, NULL); if (config != NULL) { pool_checkpoint_stat_t *pcs = NULL; uint_t c; nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE); (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_CHECKPOINT_STATS, (uint64_t **)&pcs, &c); if (pcs == NULL || pcs->pcs_state == CS_NONE) return (B_FALSE); assert(pcs->pcs_state == CS_CHECKPOINT_EXISTS || pcs->pcs_state == CS_CHECKPOINT_DISCARDING); return (B_TRUE); } return (B_FALSE); } static int scrub_callback(zpool_handle_t *zhp, void *data) { scrub_cbdata_t *cb = data; int err; /* * Ignore faulted pools. */ if (zpool_get_state(zhp) == POOL_STATE_UNAVAIL) { (void) fprintf(stderr, gettext("cannot scan '%s': pool is " "currently unavailable\n"), zpool_get_name(zhp)); return (1); } err = zpool_scan(zhp, cb->cb_type, cb->cb_scrub_cmd); if (err == 0 && zpool_has_checkpoint(zhp) && cb->cb_type == POOL_SCAN_SCRUB) { (void) printf(gettext("warning: will not scrub state that " "belongs to the checkpoint of pool '%s'\n"), zpool_get_name(zhp)); } return (err != 0); } static int wait_callback(zpool_handle_t *zhp, void *data) { zpool_wait_activity_t *act = data; return (zpool_wait(zhp, *act)); } /* - * zpool scrub [-s | -p] [-w] ... + * zpool scrub [-s | -p] [-w] [-e] ... * + * -e Only scrub blocks in the error log. * -s Stop. Stops any in-progress scrub. * -p Pause. Pause in-progress scrub. * -w Wait. Blocks until scrub has completed. */ int zpool_do_scrub(int argc, char **argv) { int c; scrub_cbdata_t cb; boolean_t wait = B_FALSE; int error; cb.cb_type = POOL_SCAN_SCRUB; cb.cb_scrub_cmd = POOL_SCRUB_NORMAL; + boolean_t is_error_scrub = B_FALSE; + boolean_t is_pause = B_FALSE; + boolean_t is_stop = B_FALSE; + /* check options */ - while ((c = getopt(argc, argv, "spw")) != -1) { + while ((c = getopt(argc, argv, "spwe")) != -1) { switch (c) { + case 'e': + is_error_scrub = B_TRUE; + break; case 's': - cb.cb_type = POOL_SCAN_NONE; + is_stop = B_TRUE; break; case 'p': - cb.cb_scrub_cmd = POOL_SCRUB_PAUSE; + is_pause = B_TRUE; break; case 'w': wait = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } - if (cb.cb_type == POOL_SCAN_NONE && - cb.cb_scrub_cmd == POOL_SCRUB_PAUSE) { - (void) fprintf(stderr, gettext("invalid option combination: " - "-s and -p are mutually exclusive\n")); + if (is_pause && is_stop) { + (void) fprintf(stderr, gettext("invalid option " + "combination :-s and -p are mutually exclusive\n")); usage(B_FALSE); + } else { + if (is_error_scrub) + cb.cb_type = POOL_SCAN_ERRORSCRUB; + + if (is_pause) { + cb.cb_scrub_cmd = POOL_SCRUB_PAUSE; + } else if (is_stop) { + cb.cb_type = POOL_SCAN_NONE; + } else { + cb.cb_scrub_cmd = POOL_SCRUB_NORMAL; + } } if (wait && (cb.cb_type == POOL_SCAN_NONE || cb.cb_scrub_cmd == POOL_SCRUB_PAUSE)) { (void) fprintf(stderr, gettext("invalid option combination: " "-w cannot be used with -p or -s\n")); usage(B_FALSE); } argc -= optind; argv += optind; if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name argument\n")); usage(B_FALSE); } error = for_each_pool(argc, argv, B_TRUE, NULL, ZFS_TYPE_POOL, B_FALSE, scrub_callback, &cb); if (wait && !error) { zpool_wait_activity_t act = ZPOOL_WAIT_SCRUB; error = for_each_pool(argc, argv, B_TRUE, NULL, ZFS_TYPE_POOL, B_FALSE, wait_callback, &act); } return (error); } /* * zpool resilver ... * * Restarts any in-progress resilver */ int zpool_do_resilver(int argc, char **argv) { int c; scrub_cbdata_t cb; cb.cb_type = POOL_SCAN_RESILVER; cb.cb_scrub_cmd = POOL_SCRUB_NORMAL; /* check options */ while ((c = getopt(argc, argv, "")) != -1) { switch (c) { case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name argument\n")); usage(B_FALSE); } return (for_each_pool(argc, argv, B_TRUE, NULL, ZFS_TYPE_POOL, B_FALSE, scrub_callback, &cb)); } /* * zpool trim [-d] [-r ] [-c | -s] [ ...] * * -c Cancel. Ends any in-progress trim. * -d Secure trim. Requires kernel and device support. * -r Sets the TRIM rate in bytes (per second). Supports * adding a multiplier suffix such as 'k' or 'm'. * -s Suspend. TRIM can then be restarted with no flags. * -w Wait. Blocks until trimming has completed. */ int zpool_do_trim(int argc, char **argv) { struct option long_options[] = { {"cancel", no_argument, NULL, 'c'}, {"secure", no_argument, NULL, 'd'}, {"rate", required_argument, NULL, 'r'}, {"suspend", no_argument, NULL, 's'}, {"wait", no_argument, NULL, 'w'}, {0, 0, 0, 0} }; pool_trim_func_t cmd_type = POOL_TRIM_START; uint64_t rate = 0; boolean_t secure = B_FALSE; boolean_t wait = B_FALSE; int c; while ((c = getopt_long(argc, argv, "cdr:sw", long_options, NULL)) != -1) { switch (c) { case 'c': if (cmd_type != POOL_TRIM_START && cmd_type != POOL_TRIM_CANCEL) { (void) fprintf(stderr, gettext("-c cannot be " "combined with other options\n")); usage(B_FALSE); } cmd_type = POOL_TRIM_CANCEL; break; case 'd': if (cmd_type != POOL_TRIM_START) { (void) fprintf(stderr, gettext("-d cannot be " "combined with the -c or -s options\n")); usage(B_FALSE); } secure = B_TRUE; break; case 'r': if (cmd_type != POOL_TRIM_START) { (void) fprintf(stderr, gettext("-r cannot be " "combined with the -c or -s options\n")); usage(B_FALSE); } if (zfs_nicestrtonum(g_zfs, optarg, &rate) == -1) { (void) fprintf(stderr, "%s: %s\n", gettext("invalid value for rate"), libzfs_error_description(g_zfs)); usage(B_FALSE); } break; case 's': if (cmd_type != POOL_TRIM_START && cmd_type != POOL_TRIM_SUSPEND) { (void) fprintf(stderr, gettext("-s cannot be " "combined with other options\n")); usage(B_FALSE); } cmd_type = POOL_TRIM_SUSPEND; break; case 'w': wait = B_TRUE; break; case '?': if (optopt != 0) { (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); } else { (void) fprintf(stderr, gettext("invalid option '%s'\n"), argv[optind - 1]); } usage(B_FALSE); } } argc -= optind; argv += optind; if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name argument\n")); usage(B_FALSE); return (-1); } if (wait && (cmd_type != POOL_TRIM_START)) { (void) fprintf(stderr, gettext("-w cannot be used with -c or " "-s\n")); usage(B_FALSE); } char *poolname = argv[0]; zpool_handle_t *zhp = zpool_open(g_zfs, poolname); if (zhp == NULL) return (-1); trimflags_t trim_flags = { .secure = secure, .rate = rate, .wait = wait, }; nvlist_t *vdevs = fnvlist_alloc(); if (argc == 1) { /* no individual leaf vdevs specified, so add them all */ nvlist_t *config = zpool_get_config(zhp, NULL); nvlist_t *nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE); zpool_collect_leaves(zhp, nvroot, vdevs); trim_flags.fullpool = B_TRUE; } else { trim_flags.fullpool = B_FALSE; for (int i = 1; i < argc; i++) { fnvlist_add_boolean(vdevs, argv[i]); } } int error = zpool_trim(zhp, cmd_type, vdevs, &trim_flags); fnvlist_free(vdevs); zpool_close(zhp); return (error); } /* * Converts a total number of seconds to a human readable string broken * down in to days/hours/minutes/seconds. */ static void secs_to_dhms(uint64_t total, char *buf) { uint64_t days = total / 60 / 60 / 24; uint64_t hours = (total / 60 / 60) % 24; uint64_t mins = (total / 60) % 60; uint64_t secs = (total % 60); if (days > 0) { (void) sprintf(buf, "%llu days %02llu:%02llu:%02llu", (u_longlong_t)days, (u_longlong_t)hours, (u_longlong_t)mins, (u_longlong_t)secs); } else { (void) sprintf(buf, "%02llu:%02llu:%02llu", (u_longlong_t)hours, (u_longlong_t)mins, (u_longlong_t)secs); } } +/* + * Print out detailed error scrub status. + */ +static void +print_err_scrub_status(pool_scan_stat_t *ps) +{ + time_t start, end, pause; + uint64_t total_secs_left; + uint64_t secs_left, mins_left, hours_left, days_left; + uint64_t examined, to_be_examined; + + if (ps == NULL || ps->pss_error_scrub_func != POOL_SCAN_ERRORSCRUB) { + return; + } + + (void) printf(gettext(" scrub: ")); + + start = ps->pss_error_scrub_start; + end = ps->pss_error_scrub_end; + pause = ps->pss_pass_error_scrub_pause; + examined = ps->pss_error_scrub_examined; + to_be_examined = ps->pss_error_scrub_to_be_examined; + + assert(ps->pss_error_scrub_func == POOL_SCAN_ERRORSCRUB); + + if (ps->pss_error_scrub_state == DSS_FINISHED) { + total_secs_left = end - start; + days_left = total_secs_left / 60 / 60 / 24; + hours_left = (total_secs_left / 60 / 60) % 24; + mins_left = (total_secs_left / 60) % 60; + secs_left = (total_secs_left % 60); + + (void) printf(gettext("scrubbed %llu error blocks in %llu days " + "%02llu:%02llu:%02llu on %s"), (u_longlong_t)examined, + (u_longlong_t)days_left, (u_longlong_t)hours_left, + (u_longlong_t)mins_left, (u_longlong_t)secs_left, + ctime(&end)); + + return; + } else if (ps->pss_error_scrub_state == DSS_CANCELED) { + (void) printf(gettext("error scrub canceled on %s"), + ctime(&end)); + return; + } + assert(ps->pss_error_scrub_state == DSS_ERRORSCRUBBING); + + /* Error scrub is in progress. */ + if (pause == 0) { + (void) printf(gettext("error scrub in progress since %s"), + ctime(&start)); + } else { + (void) printf(gettext("error scrub paused since %s"), + ctime(&pause)); + (void) printf(gettext("\terror scrub started on %s"), + ctime(&start)); + } + + double fraction_done = (double)examined / (to_be_examined + examined); + (void) printf(gettext("\t%.2f%% done, issued I/O for %llu error" + " blocks"), 100 * fraction_done, (u_longlong_t)examined); + + (void) printf("\n"); +} + /* * Print out detailed scrub status. */ static void print_scan_scrub_resilver_status(pool_scan_stat_t *ps) { time_t start, end, pause; uint64_t pass_scanned, scanned, pass_issued, issued, total; uint64_t elapsed, scan_rate, issue_rate; double fraction_done; char processed_buf[7], scanned_buf[7], issued_buf[7], total_buf[7]; char srate_buf[7], irate_buf[7], time_buf[32]; printf(" "); printf_color(ANSI_BOLD, gettext("scan:")); printf(" "); /* If there's never been a scan, there's not much to say. */ if (ps == NULL || ps->pss_func == POOL_SCAN_NONE || ps->pss_func >= POOL_SCAN_FUNCS) { (void) printf(gettext("none requested\n")); return; } start = ps->pss_start_time; end = ps->pss_end_time; pause = ps->pss_pass_scrub_pause; zfs_nicebytes(ps->pss_processed, processed_buf, sizeof (processed_buf)); int is_resilver = ps->pss_func == POOL_SCAN_RESILVER; int is_scrub = ps->pss_func == POOL_SCAN_SCRUB; assert(is_resilver || is_scrub); /* Scan is finished or canceled. */ if (ps->pss_state == DSS_FINISHED) { secs_to_dhms(end - start, time_buf); if (is_scrub) { (void) printf(gettext("scrub repaired %s " "in %s with %llu errors on %s"), processed_buf, time_buf, (u_longlong_t)ps->pss_errors, ctime(&end)); } else if (is_resilver) { (void) printf(gettext("resilvered %s " "in %s with %llu errors on %s"), processed_buf, time_buf, (u_longlong_t)ps->pss_errors, ctime(&end)); } return; } else if (ps->pss_state == DSS_CANCELED) { if (is_scrub) { (void) printf(gettext("scrub canceled on %s"), ctime(&end)); } else if (is_resilver) { (void) printf(gettext("resilver canceled on %s"), ctime(&end)); } return; } assert(ps->pss_state == DSS_SCANNING); /* Scan is in progress. Resilvers can't be paused. */ if (is_scrub) { if (pause == 0) { (void) printf(gettext("scrub in progress since %s"), ctime(&start)); } else { (void) printf(gettext("scrub paused since %s"), ctime(&pause)); (void) printf(gettext("\tscrub started on %s"), ctime(&start)); } } else if (is_resilver) { (void) printf(gettext("resilver in progress since %s"), ctime(&start)); } scanned = ps->pss_examined; pass_scanned = ps->pss_pass_exam; issued = ps->pss_issued; pass_issued = ps->pss_pass_issued; total = ps->pss_to_examine; /* we are only done with a block once we have issued the IO for it */ fraction_done = (double)issued / total; /* elapsed time for this pass, rounding up to 1 if it's 0 */ elapsed = time(NULL) - ps->pss_pass_start; elapsed -= ps->pss_pass_scrub_spent_paused; elapsed = (elapsed != 0) ? elapsed : 1; scan_rate = pass_scanned / elapsed; issue_rate = pass_issued / elapsed; uint64_t total_secs_left = (issue_rate != 0 && total >= issued) ? ((total - issued) / issue_rate) : UINT64_MAX; secs_to_dhms(total_secs_left, time_buf); /* format all of the numbers we will be reporting */ zfs_nicebytes(scanned, scanned_buf, sizeof (scanned_buf)); zfs_nicebytes(issued, issued_buf, sizeof (issued_buf)); zfs_nicebytes(total, total_buf, sizeof (total_buf)); zfs_nicebytes(scan_rate, srate_buf, sizeof (srate_buf)); zfs_nicebytes(issue_rate, irate_buf, sizeof (irate_buf)); /* do not print estimated time if we have a paused scrub */ if (pause == 0) { (void) printf(gettext("\t%s scanned at %s/s, " "%s issued at %s/s, %s total\n"), scanned_buf, srate_buf, issued_buf, irate_buf, total_buf); } else { (void) printf(gettext("\t%s scanned, %s issued, %s total\n"), scanned_buf, issued_buf, total_buf); } if (is_resilver) { (void) printf(gettext("\t%s resilvered, %.2f%% done"), processed_buf, 100 * fraction_done); } else if (is_scrub) { (void) printf(gettext("\t%s repaired, %.2f%% done"), processed_buf, 100 * fraction_done); } if (pause == 0) { /* * Only provide an estimate iff: * 1) the time remaining is valid, and * 2) the issue rate exceeds 10 MB/s, and * 3) it's either: * a) a resilver which has started repairs, or * b) a scrub which has entered the issue phase. */ if (total_secs_left != UINT64_MAX && issue_rate >= 10 * 1024 * 1024 && ((is_resilver && ps->pss_processed > 0) || (is_scrub && issued > 0))) { (void) printf(gettext(", %s to go\n"), time_buf); } else { (void) printf(gettext(", no estimated " "completion time\n")); } } else { (void) printf(gettext("\n")); } } static void print_rebuild_status_impl(vdev_rebuild_stat_t *vrs, char *vdev_name) { if (vrs == NULL || vrs->vrs_state == VDEV_REBUILD_NONE) return; printf(" "); printf_color(ANSI_BOLD, gettext("scan:")); printf(" "); uint64_t bytes_scanned = vrs->vrs_bytes_scanned; uint64_t bytes_issued = vrs->vrs_bytes_issued; uint64_t bytes_rebuilt = vrs->vrs_bytes_rebuilt; uint64_t bytes_est = vrs->vrs_bytes_est; uint64_t scan_rate = (vrs->vrs_pass_bytes_scanned / (vrs->vrs_pass_time_ms + 1)) * 1000; uint64_t issue_rate = (vrs->vrs_pass_bytes_issued / (vrs->vrs_pass_time_ms + 1)) * 1000; double scan_pct = MIN((double)bytes_scanned * 100 / (bytes_est + 1), 100); /* Format all of the numbers we will be reporting */ char bytes_scanned_buf[7], bytes_issued_buf[7]; char bytes_rebuilt_buf[7], bytes_est_buf[7]; char scan_rate_buf[7], issue_rate_buf[7], time_buf[32]; zfs_nicebytes(bytes_scanned, bytes_scanned_buf, sizeof (bytes_scanned_buf)); zfs_nicebytes(bytes_issued, bytes_issued_buf, sizeof (bytes_issued_buf)); zfs_nicebytes(bytes_rebuilt, bytes_rebuilt_buf, sizeof (bytes_rebuilt_buf)); zfs_nicebytes(bytes_est, bytes_est_buf, sizeof (bytes_est_buf)); zfs_nicebytes(scan_rate, scan_rate_buf, sizeof (scan_rate_buf)); zfs_nicebytes(issue_rate, issue_rate_buf, sizeof (issue_rate_buf)); time_t start = vrs->vrs_start_time; time_t end = vrs->vrs_end_time; /* Rebuild is finished or canceled. */ if (vrs->vrs_state == VDEV_REBUILD_COMPLETE) { secs_to_dhms(vrs->vrs_scan_time_ms / 1000, time_buf); (void) printf(gettext("resilvered (%s) %s in %s " "with %llu errors on %s"), vdev_name, bytes_rebuilt_buf, time_buf, (u_longlong_t)vrs->vrs_errors, ctime(&end)); return; } else if (vrs->vrs_state == VDEV_REBUILD_CANCELED) { (void) printf(gettext("resilver (%s) canceled on %s"), vdev_name, ctime(&end)); return; } else if (vrs->vrs_state == VDEV_REBUILD_ACTIVE) { (void) printf(gettext("resilver (%s) in progress since %s"), vdev_name, ctime(&start)); } assert(vrs->vrs_state == VDEV_REBUILD_ACTIVE); secs_to_dhms(MAX((int64_t)bytes_est - (int64_t)bytes_scanned, 0) / MAX(scan_rate, 1), time_buf); (void) printf(gettext("\t%s scanned at %s/s, %s issued %s/s, " "%s total\n"), bytes_scanned_buf, scan_rate_buf, bytes_issued_buf, issue_rate_buf, bytes_est_buf); (void) printf(gettext("\t%s resilvered, %.2f%% done"), bytes_rebuilt_buf, scan_pct); if (vrs->vrs_state == VDEV_REBUILD_ACTIVE) { if (scan_rate >= 10 * 1024 * 1024) { (void) printf(gettext(", %s to go\n"), time_buf); } else { (void) printf(gettext(", no estimated " "completion time\n")); } } else { (void) printf(gettext("\n")); } } /* * Print rebuild status for top-level vdevs. */ static void print_rebuild_status(zpool_handle_t *zhp, nvlist_t *nvroot) { nvlist_t **child; uint_t children; if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) children = 0; for (uint_t c = 0; c < children; c++) { vdev_rebuild_stat_t *vrs; uint_t i; if (nvlist_lookup_uint64_array(child[c], ZPOOL_CONFIG_REBUILD_STATS, (uint64_t **)&vrs, &i) == 0) { char *name = zpool_vdev_name(g_zfs, zhp, child[c], VDEV_NAME_TYPE_ID); print_rebuild_status_impl(vrs, name); free(name); } } } /* * As we don't scrub checkpointed blocks, we want to warn the user that we * skipped scanning some blocks if a checkpoint exists or existed at any * time during the scan. If a sequential instead of healing reconstruction * was performed then the blocks were reconstructed. However, their checksums * have not been verified so we still print the warning. */ static void print_checkpoint_scan_warning(pool_scan_stat_t *ps, pool_checkpoint_stat_t *pcs) { if (ps == NULL || pcs == NULL) return; if (pcs->pcs_state == CS_NONE || pcs->pcs_state == CS_CHECKPOINT_DISCARDING) return; assert(pcs->pcs_state == CS_CHECKPOINT_EXISTS); if (ps->pss_state == DSS_NONE) return; if ((ps->pss_state == DSS_FINISHED || ps->pss_state == DSS_CANCELED) && ps->pss_end_time < pcs->pcs_start_time) return; if (ps->pss_state == DSS_FINISHED || ps->pss_state == DSS_CANCELED) { (void) printf(gettext(" scan warning: skipped blocks " "that are only referenced by the checkpoint.\n")); } else { assert(ps->pss_state == DSS_SCANNING); (void) printf(gettext(" scan warning: skipping blocks " "that are only referenced by the checkpoint.\n")); } } /* * Returns B_TRUE if there is an active rebuild in progress. Otherwise, * B_FALSE is returned and 'rebuild_end_time' is set to the end time for * the last completed (or cancelled) rebuild. */ static boolean_t check_rebuilding(nvlist_t *nvroot, uint64_t *rebuild_end_time) { nvlist_t **child; uint_t children; boolean_t rebuilding = B_FALSE; uint64_t end_time = 0; if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) children = 0; for (uint_t c = 0; c < children; c++) { vdev_rebuild_stat_t *vrs; uint_t i; if (nvlist_lookup_uint64_array(child[c], ZPOOL_CONFIG_REBUILD_STATS, (uint64_t **)&vrs, &i) == 0) { if (vrs->vrs_end_time > end_time) end_time = vrs->vrs_end_time; if (vrs->vrs_state == VDEV_REBUILD_ACTIVE) { rebuilding = B_TRUE; end_time = 0; break; } } } if (rebuild_end_time != NULL) *rebuild_end_time = end_time; return (rebuilding); } /* * Print the scan status. */ static void print_scan_status(zpool_handle_t *zhp, nvlist_t *nvroot) { uint64_t rebuild_end_time = 0, resilver_end_time = 0; boolean_t have_resilver = B_FALSE, have_scrub = B_FALSE; + boolean_t have_errorscrub = B_FALSE; boolean_t active_resilver = B_FALSE; pool_checkpoint_stat_t *pcs = NULL; pool_scan_stat_t *ps = NULL; uint_t c; + time_t scrub_start = 0, errorscrub_start = 0; if (nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_SCAN_STATS, (uint64_t **)&ps, &c) == 0) { if (ps->pss_func == POOL_SCAN_RESILVER) { resilver_end_time = ps->pss_end_time; active_resilver = (ps->pss_state == DSS_SCANNING); } + have_resilver = (ps->pss_func == POOL_SCAN_RESILVER); have_scrub = (ps->pss_func == POOL_SCAN_SCRUB); + scrub_start = ps->pss_start_time; + have_errorscrub = (ps->pss_error_scrub_func == + POOL_SCAN_ERRORSCRUB); + errorscrub_start = ps->pss_error_scrub_start; } boolean_t active_rebuild = check_rebuilding(nvroot, &rebuild_end_time); boolean_t have_rebuild = (active_rebuild || (rebuild_end_time > 0)); /* Always print the scrub status when available. */ - if (have_scrub) + if (have_scrub && scrub_start > errorscrub_start) print_scan_scrub_resilver_status(ps); + else if (have_errorscrub && errorscrub_start >= scrub_start) + print_err_scrub_status(ps); /* * When there is an active resilver or rebuild print its status. * Otherwise print the status of the last resilver or rebuild. */ if (active_resilver || (!active_rebuild && have_resilver && resilver_end_time && resilver_end_time > rebuild_end_time)) { print_scan_scrub_resilver_status(ps); } else if (active_rebuild || (!active_resilver && have_rebuild && rebuild_end_time && rebuild_end_time > resilver_end_time)) { print_rebuild_status(zhp, nvroot); } (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_CHECKPOINT_STATS, (uint64_t **)&pcs, &c); print_checkpoint_scan_warning(ps, pcs); } /* * Print out detailed removal status. */ static void print_removal_status(zpool_handle_t *zhp, pool_removal_stat_t *prs) { char copied_buf[7], examined_buf[7], total_buf[7], rate_buf[7]; time_t start, end; nvlist_t *config, *nvroot; nvlist_t **child; uint_t children; char *vdev_name; if (prs == NULL || prs->prs_state == DSS_NONE) return; /* * Determine name of vdev. */ config = zpool_get_config(zhp, NULL); nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE); verify(nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN, &child, &children) == 0); assert(prs->prs_removing_vdev < children); vdev_name = zpool_vdev_name(g_zfs, zhp, child[prs->prs_removing_vdev], B_TRUE); printf_color(ANSI_BOLD, gettext("remove: ")); start = prs->prs_start_time; end = prs->prs_end_time; zfs_nicenum(prs->prs_copied, copied_buf, sizeof (copied_buf)); /* * Removal is finished or canceled. */ if (prs->prs_state == DSS_FINISHED) { uint64_t minutes_taken = (end - start) / 60; (void) printf(gettext("Removal of vdev %llu copied %s " "in %lluh%um, completed on %s"), (longlong_t)prs->prs_removing_vdev, copied_buf, (u_longlong_t)(minutes_taken / 60), (uint_t)(minutes_taken % 60), ctime((time_t *)&end)); } else if (prs->prs_state == DSS_CANCELED) { (void) printf(gettext("Removal of %s canceled on %s"), vdev_name, ctime(&end)); } else { uint64_t copied, total, elapsed, mins_left, hours_left; double fraction_done; uint_t rate; assert(prs->prs_state == DSS_SCANNING); /* * Removal is in progress. */ (void) printf(gettext( "Evacuation of %s in progress since %s"), vdev_name, ctime(&start)); copied = prs->prs_copied > 0 ? prs->prs_copied : 1; total = prs->prs_to_copy; fraction_done = (double)copied / total; /* elapsed time for this pass */ elapsed = time(NULL) - prs->prs_start_time; elapsed = elapsed > 0 ? elapsed : 1; rate = copied / elapsed; rate = rate > 0 ? rate : 1; mins_left = ((total - copied) / rate) / 60; hours_left = mins_left / 60; zfs_nicenum(copied, examined_buf, sizeof (examined_buf)); zfs_nicenum(total, total_buf, sizeof (total_buf)); zfs_nicenum(rate, rate_buf, sizeof (rate_buf)); /* * do not print estimated time if hours_left is more than * 30 days */ (void) printf(gettext( "\t%s copied out of %s at %s/s, %.2f%% done"), examined_buf, total_buf, rate_buf, 100 * fraction_done); if (hours_left < (30 * 24)) { (void) printf(gettext(", %lluh%um to go\n"), (u_longlong_t)hours_left, (uint_t)(mins_left % 60)); } else { (void) printf(gettext( ", (copy is slow, no estimated time)\n")); } } free(vdev_name); if (prs->prs_mapping_memory > 0) { char mem_buf[7]; zfs_nicenum(prs->prs_mapping_memory, mem_buf, sizeof (mem_buf)); (void) printf(gettext( "\t%s memory used for removed device mappings\n"), mem_buf); } } static void print_checkpoint_status(pool_checkpoint_stat_t *pcs) { time_t start; char space_buf[7]; if (pcs == NULL || pcs->pcs_state == CS_NONE) return; (void) printf(gettext("checkpoint: ")); start = pcs->pcs_start_time; zfs_nicenum(pcs->pcs_space, space_buf, sizeof (space_buf)); if (pcs->pcs_state == CS_CHECKPOINT_EXISTS) { char *date = ctime(&start); /* * ctime() adds a newline at the end of the generated * string, thus the weird format specifier and the * strlen() call used to chop it off from the output. */ (void) printf(gettext("created %.*s, consumes %s\n"), (int)(strlen(date) - 1), date, space_buf); return; } assert(pcs->pcs_state == CS_CHECKPOINT_DISCARDING); (void) printf(gettext("discarding, %s remaining.\n"), space_buf); } static void print_error_log(zpool_handle_t *zhp) { nvlist_t *nverrlist = NULL; nvpair_t *elem; char *pathname; size_t len = MAXPATHLEN * 2; if (zpool_get_errlog(zhp, &nverrlist) != 0) return; (void) printf("errors: Permanent errors have been " "detected in the following files:\n\n"); pathname = safe_malloc(len); elem = NULL; while ((elem = nvlist_next_nvpair(nverrlist, elem)) != NULL) { nvlist_t *nv; uint64_t dsobj, obj; verify(nvpair_value_nvlist(elem, &nv) == 0); verify(nvlist_lookup_uint64(nv, ZPOOL_ERR_DATASET, &dsobj) == 0); verify(nvlist_lookup_uint64(nv, ZPOOL_ERR_OBJECT, &obj) == 0); zpool_obj_to_path(zhp, dsobj, obj, pathname, len); (void) printf("%7s %s\n", "", pathname); } free(pathname); nvlist_free(nverrlist); } static void print_spares(zpool_handle_t *zhp, status_cbdata_t *cb, nvlist_t **spares, uint_t nspares) { uint_t i; char *name; if (nspares == 0) return; (void) printf(gettext("\tspares\n")); for (i = 0; i < nspares; i++) { name = zpool_vdev_name(g_zfs, zhp, spares[i], cb->cb_name_flags); print_status_config(zhp, cb, name, spares[i], 2, B_TRUE, NULL); free(name); } } static void print_l2cache(zpool_handle_t *zhp, status_cbdata_t *cb, nvlist_t **l2cache, uint_t nl2cache) { uint_t i; char *name; if (nl2cache == 0) return; (void) printf(gettext("\tcache\n")); for (i = 0; i < nl2cache; i++) { name = zpool_vdev_name(g_zfs, zhp, l2cache[i], cb->cb_name_flags); print_status_config(zhp, cb, name, l2cache[i], 2, B_FALSE, NULL); free(name); } } static void print_dedup_stats(nvlist_t *config) { ddt_histogram_t *ddh; ddt_stat_t *dds; ddt_object_t *ddo; uint_t c; char dspace[6], mspace[6]; /* * If the pool was faulted then we may not have been able to * obtain the config. Otherwise, if we have anything in the dedup * table continue processing the stats. */ if (nvlist_lookup_uint64_array(config, ZPOOL_CONFIG_DDT_OBJ_STATS, (uint64_t **)&ddo, &c) != 0) return; (void) printf("\n"); (void) printf(gettext(" dedup: ")); if (ddo->ddo_count == 0) { (void) printf(gettext("no DDT entries\n")); return; } zfs_nicebytes(ddo->ddo_dspace, dspace, sizeof (dspace)); zfs_nicebytes(ddo->ddo_mspace, mspace, sizeof (mspace)); (void) printf("DDT entries %llu, size %s on disk, %s in core\n", (u_longlong_t)ddo->ddo_count, dspace, mspace); verify(nvlist_lookup_uint64_array(config, ZPOOL_CONFIG_DDT_STATS, (uint64_t **)&dds, &c) == 0); verify(nvlist_lookup_uint64_array(config, ZPOOL_CONFIG_DDT_HISTOGRAM, (uint64_t **)&ddh, &c) == 0); zpool_dump_ddt(dds, ddh); } /* * Display a summary of pool status. Displays a summary such as: * * pool: tank * status: DEGRADED * reason: One or more devices ... * see: https://openzfs.github.io/openzfs-docs/msg/ZFS-xxxx-01 * config: * mirror DEGRADED * c1t0d0 OK * c2t0d0 UNAVAIL * * When given the '-v' option, we print out the complete config. If the '-e' * option is specified, then we print out error rate information as well. */ static int status_callback(zpool_handle_t *zhp, void *data) { status_cbdata_t *cbp = data; nvlist_t *config, *nvroot; const char *msgid; zpool_status_t reason; zpool_errata_t errata; const char *health; uint_t c; vdev_stat_t *vs; config = zpool_get_config(zhp, NULL); reason = zpool_get_status(zhp, &msgid, &errata); cbp->cb_count++; /* * If we were given 'zpool status -x', only report those pools with * problems. */ if (cbp->cb_explain && (reason == ZPOOL_STATUS_OK || reason == ZPOOL_STATUS_VERSION_OLDER || reason == ZPOOL_STATUS_FEAT_DISABLED || reason == ZPOOL_STATUS_COMPATIBILITY_ERR || reason == ZPOOL_STATUS_INCOMPATIBLE_FEAT)) { if (!cbp->cb_allpools) { (void) printf(gettext("pool '%s' is healthy\n"), zpool_get_name(zhp)); if (cbp->cb_first) cbp->cb_first = B_FALSE; } return (0); } if (cbp->cb_first) cbp->cb_first = B_FALSE; else (void) printf("\n"); nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE); verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &c) == 0); health = zpool_get_state_str(zhp); printf(" "); printf_color(ANSI_BOLD, gettext("pool:")); printf(" %s\n", zpool_get_name(zhp)); fputc(' ', stdout); printf_color(ANSI_BOLD, gettext("state: ")); printf_color(health_str_to_color(health), "%s", health); fputc('\n', stdout); switch (reason) { case ZPOOL_STATUS_MISSING_DEV_R: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices could " "not be opened. Sufficient replicas exist for\n\tthe pool " "to continue functioning in a degraded state.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Attach the missing device " "and online it using 'zpool online'.\n")); break; case ZPOOL_STATUS_MISSING_DEV_NR: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices could " "not be opened. There are insufficient\n\treplicas for the" " pool to continue functioning.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Attach the missing device " "and online it using 'zpool online'.\n")); break; case ZPOOL_STATUS_CORRUPT_LABEL_R: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices could " "not be used because the label is missing or\n\tinvalid. " "Sufficient replicas exist for the pool to continue\n\t" "functioning in a degraded state.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Replace the device using " "'zpool replace'.\n")); break; case ZPOOL_STATUS_CORRUPT_LABEL_NR: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices could " "not be used because the label is missing \n\tor invalid. " "There are insufficient replicas for the pool to " "continue\n\tfunctioning.\n")); zpool_explain_recover(zpool_get_handle(zhp), zpool_get_name(zhp), reason, config); break; case ZPOOL_STATUS_FAILING_DEV: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices has " "experienced an unrecoverable error. An\n\tattempt was " "made to correct the error. Applications are " "unaffected.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Determine if the " "device needs to be replaced, and clear the errors\n\tusing" " 'zpool clear' or replace the device with 'zpool " "replace'.\n")); break; case ZPOOL_STATUS_OFFLINE_DEV: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices has " "been taken offline by the administrator.\n\tSufficient " "replicas exist for the pool to continue functioning in " "a\n\tdegraded state.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Online the device " "using 'zpool online' or replace the device with\n\t'zpool " "replace'.\n")); break; case ZPOOL_STATUS_REMOVED_DEV: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices has " "been removed by the administrator.\n\tSufficient " "replicas exist for the pool to continue functioning in " "a\n\tdegraded state.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Online the device " "using zpool online' or replace the device with\n\t'zpool " "replace'.\n")); break; case ZPOOL_STATUS_RESILVERING: case ZPOOL_STATUS_REBUILDING: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices is " "currently being resilvered. The pool will\n\tcontinue " "to function, possibly in a degraded state.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Wait for the resilver to " "complete.\n")); break; case ZPOOL_STATUS_REBUILD_SCRUB: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices have " "been sequentially resilvered, scrubbing\n\tthe pool " "is recommended.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Use 'zpool scrub' to " "verify all data checksums.\n")); break; case ZPOOL_STATUS_CORRUPT_DATA: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices has " "experienced an error resulting in data\n\tcorruption. " "Applications may be affected.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Restore the file in question" " if possible. Otherwise restore the\n\tentire pool from " "backup.\n")); break; case ZPOOL_STATUS_CORRUPT_POOL: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool metadata is " "corrupted and the pool cannot be opened.\n")); zpool_explain_recover(zpool_get_handle(zhp), zpool_get_name(zhp), reason, config); break; case ZPOOL_STATUS_VERSION_OLDER: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool is formatted using " "a legacy on-disk format. The pool can\n\tstill be used, " "but some features are unavailable.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Upgrade the pool using " "'zpool upgrade'. Once this is done, the\n\tpool will no " "longer be accessible on software that does not support\n\t" "feature flags.\n")); break; case ZPOOL_STATUS_VERSION_NEWER: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool has been upgraded " "to a newer, incompatible on-disk version.\n\tThe pool " "cannot be accessed on this system.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Access the pool from a " "system running more recent software, or\n\trestore the " "pool from backup.\n")); break; case ZPOOL_STATUS_FEAT_DISABLED: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("Some supported and " "requested features are not enabled on the pool.\n\t" "The pool can still be used, but some features are " "unavailable.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Enable all features using " "'zpool upgrade'. Once this is done,\n\tthe pool may no " "longer be accessible by software that does not support\n\t" "the features. See zpool-features(7) for details.\n")); break; case ZPOOL_STATUS_COMPATIBILITY_ERR: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("This pool has a " "compatibility list specified, but it could not be\n\t" "read/parsed at this time. The pool can still be used, " "but this\n\tshould be investigated.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Check the value of the " "'compatibility' property against the\n\t" "appropriate file in " ZPOOL_SYSCONF_COMPAT_D " or " ZPOOL_DATA_COMPAT_D ".\n")); break; case ZPOOL_STATUS_INCOMPATIBLE_FEAT: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more features " "are enabled on the pool despite not being\n\t" "requested by the 'compatibility' property.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Consider setting " "'compatibility' to an appropriate value, or\n\t" "adding needed features to the relevant file in\n\t" ZPOOL_SYSCONF_COMPAT_D " or " ZPOOL_DATA_COMPAT_D ".\n")); break; case ZPOOL_STATUS_UNSUP_FEAT_READ: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool cannot be accessed " "on this system because it uses the\n\tfollowing feature(s)" " not supported on this system:\n")); zpool_print_unsup_feat(config); (void) printf("\n"); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Access the pool from a " "system that supports the required feature(s),\n\tor " "restore the pool from backup.\n")); break; case ZPOOL_STATUS_UNSUP_FEAT_WRITE: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool can only be " "accessed in read-only mode on this system. It\n\tcannot be" " accessed in read-write mode because it uses the " "following\n\tfeature(s) not supported on this system:\n")); zpool_print_unsup_feat(config); (void) printf("\n"); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("The pool cannot be accessed " "in read-write mode. Import the pool with\n" "\t\"-o readonly=on\", access the pool from a system that " "supports the\n\trequired feature(s), or restore the " "pool from backup.\n")); break; case ZPOOL_STATUS_FAULTED_DEV_R: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices are " "faulted in response to persistent errors.\n\tSufficient " "replicas exist for the pool to continue functioning " "in a\n\tdegraded state.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Replace the faulted device, " "or use 'zpool clear' to mark the device\n\trepaired.\n")); break; case ZPOOL_STATUS_FAULTED_DEV_NR: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices are " "faulted in response to persistent errors. There are " "insufficient replicas for the pool to\n\tcontinue " "functioning.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Destroy and re-create the " "pool from a backup source. Manually marking the device\n" "\trepaired using 'zpool clear' may allow some data " "to be recovered.\n")); break; case ZPOOL_STATUS_IO_FAILURE_MMP: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("The pool is suspended " "because multihost writes failed or were delayed;\n\t" "another system could import the pool undetected.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Make sure the pool's devices" " are connected, then reboot your system and\n\timport the " "pool.\n")); break; case ZPOOL_STATUS_IO_FAILURE_WAIT: case ZPOOL_STATUS_IO_FAILURE_CONTINUE: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("One or more devices are " "faulted in response to IO failures.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Make sure the affected " "devices are connected, then run 'zpool clear'.\n")); break; case ZPOOL_STATUS_BAD_LOG: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("An intent log record " "could not be read.\n" "\tWaiting for administrator intervention to fix the " "faulted pool.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Either restore the affected " "device(s) and run 'zpool online',\n" "\tor ignore the intent log records by running " "'zpool clear'.\n")); break; case ZPOOL_STATUS_NON_NATIVE_ASHIFT: (void) printf(gettext("status: One or more devices are " "configured to use a non-native block size.\n" "\tExpect reduced performance.\n")); (void) printf(gettext("action: Replace affected devices with " "devices that support the\n\tconfigured block size, or " "migrate data to a properly configured\n\tpool.\n")); break; case ZPOOL_STATUS_HOSTID_MISMATCH: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("Mismatch between pool hostid" " and system hostid on imported pool.\n\tThis pool was " "previously imported into a system with a different " "hostid,\n\tand then was verbatim imported into this " "system.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("Export this pool on all " "systems on which it is imported.\n" "\tThen import it to correct the mismatch.\n")); break; case ZPOOL_STATUS_ERRATA: printf_color(ANSI_BOLD, gettext("status: ")); printf_color(ANSI_YELLOW, gettext("Errata #%d detected.\n"), errata); switch (errata) { case ZPOOL_ERRATA_NONE: break; case ZPOOL_ERRATA_ZOL_2094_SCRUB: printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("To correct the issue" " run 'zpool scrub'.\n")); break; case ZPOOL_ERRATA_ZOL_6845_ENCRYPTION: (void) printf(gettext("\tExisting encrypted datasets " "contain an on-disk incompatibility\n\twhich " "needs to be corrected.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("To correct the issue" " backup existing encrypted datasets to new\n\t" "encrypted datasets and destroy the old ones. " "'zfs mount -o ro' can\n\tbe used to temporarily " "mount existing encrypted datasets readonly.\n")); break; case ZPOOL_ERRATA_ZOL_8308_ENCRYPTION: (void) printf(gettext("\tExisting encrypted snapshots " "and bookmarks contain an on-disk\n\tincompat" "ibility. This may cause on-disk corruption if " "they are used\n\twith 'zfs recv'.\n")); printf_color(ANSI_BOLD, gettext("action: ")); printf_color(ANSI_YELLOW, gettext("To correct the" "issue, enable the bookmark_v2 feature. No " "additional\n\taction is needed if there are no " "encrypted snapshots or bookmarks.\n\tIf preserving" "the encrypted snapshots and bookmarks is required," " use\n\ta non-raw send to backup and restore them." " Alternately, they may be\n\tremoved to resolve " "the incompatibility.\n")); break; default: /* * All errata which allow the pool to be imported * must contain an action message. */ assert(0); } break; default: /* * The remaining errors can't actually be generated, yet. */ assert(reason == ZPOOL_STATUS_OK); } if (msgid != NULL) { printf(" "); printf_color(ANSI_BOLD, gettext("see:")); printf(gettext( " https://openzfs.github.io/openzfs-docs/msg/%s\n"), msgid); } if (config != NULL) { uint64_t nerr; nvlist_t **spares, **l2cache; uint_t nspares, nl2cache; pool_checkpoint_stat_t *pcs = NULL; pool_removal_stat_t *prs = NULL; print_scan_status(zhp, nvroot); (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_REMOVAL_STATS, (uint64_t **)&prs, &c); print_removal_status(zhp, prs); (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_CHECKPOINT_STATS, (uint64_t **)&pcs, &c); print_checkpoint_status(pcs); cbp->cb_namewidth = max_width(zhp, nvroot, 0, 0, cbp->cb_name_flags | VDEV_NAME_TYPE_ID); if (cbp->cb_namewidth < 10) cbp->cb_namewidth = 10; color_start(ANSI_BOLD); (void) printf(gettext("config:\n\n")); (void) printf(gettext("\t%-*s %-8s %5s %5s %5s"), cbp->cb_namewidth, "NAME", "STATE", "READ", "WRITE", "CKSUM"); color_end(); if (cbp->cb_print_slow_ios) { printf_color(ANSI_BOLD, " %5s", gettext("SLOW")); } if (cbp->vcdl != NULL) print_cmd_columns(cbp->vcdl, 0); printf("\n"); print_status_config(zhp, cbp, zpool_get_name(zhp), nvroot, 0, B_FALSE, NULL); print_class_vdevs(zhp, cbp, nvroot, VDEV_ALLOC_BIAS_DEDUP); print_class_vdevs(zhp, cbp, nvroot, VDEV_ALLOC_BIAS_SPECIAL); print_class_vdevs(zhp, cbp, nvroot, VDEV_ALLOC_CLASS_LOGS); if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache) == 0) print_l2cache(zhp, cbp, l2cache, nl2cache); if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, &spares, &nspares) == 0) print_spares(zhp, cbp, spares, nspares); if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_ERRCOUNT, &nerr) == 0) { (void) printf("\n"); if (nerr == 0) { (void) printf(gettext( "errors: No known data errors\n")); } else if (!cbp->cb_verbose) { (void) printf(gettext("errors: %llu data " "errors, use '-v' for a list\n"), (u_longlong_t)nerr); } else { print_error_log(zhp); } } if (cbp->cb_dedup_stats) print_dedup_stats(config); } else { (void) printf(gettext("config: The configuration cannot be " "determined.\n")); } return (0); } /* * zpool status [-c [script1,script2,...]] [-igLpPstvx] [-T d|u] [pool] ... * [interval [count]] * * -c CMD For each vdev, run command CMD * -i Display vdev initialization status. * -g Display guid for individual vdev name. * -L Follow links when resolving vdev path name. * -p Display values in parsable (exact) format. * -P Display full path for vdev name. * -s Display slow IOs column. * -v Display complete error logs * -x Display only pools with potential problems * -D Display dedup status (undocumented) * -t Display vdev TRIM status. * -T Display a timestamp in date(1) or Unix format * * Describes the health status of all pools or some subset. */ int zpool_do_status(int argc, char **argv) { int c; int ret; float interval = 0; unsigned long count = 0; status_cbdata_t cb = { 0 }; char *cmd = NULL; /* check options */ while ((c = getopt(argc, argv, "c:igLpPsvxDtT:")) != -1) { switch (c) { case 'c': if (cmd != NULL) { fprintf(stderr, gettext("Can't set -c flag twice\n")); exit(1); } if (getenv("ZPOOL_SCRIPTS_ENABLED") != NULL && !libzfs_envvar_is_set("ZPOOL_SCRIPTS_ENABLED")) { fprintf(stderr, gettext( "Can't run -c, disabled by " "ZPOOL_SCRIPTS_ENABLED.\n")); exit(1); } if ((getuid() <= 0 || geteuid() <= 0) && !libzfs_envvar_is_set("ZPOOL_SCRIPTS_AS_ROOT")) { fprintf(stderr, gettext( "Can't run -c with root privileges " "unless ZPOOL_SCRIPTS_AS_ROOT is set.\n")); exit(1); } cmd = optarg; break; case 'i': cb.cb_print_vdev_init = B_TRUE; break; case 'g': cb.cb_name_flags |= VDEV_NAME_GUID; break; case 'L': cb.cb_name_flags |= VDEV_NAME_FOLLOW_LINKS; break; case 'p': cb.cb_literal = B_TRUE; break; case 'P': cb.cb_name_flags |= VDEV_NAME_PATH; break; case 's': cb.cb_print_slow_ios = B_TRUE; break; case 'v': cb.cb_verbose = B_TRUE; break; case 'x': cb.cb_explain = B_TRUE; break; case 'D': cb.cb_dedup_stats = B_TRUE; break; case 't': cb.cb_print_vdev_trim = B_TRUE; break; case 'T': get_timestamp_arg(*optarg); break; case '?': if (optopt == 'c') { print_zpool_script_list("status"); exit(0); } else { fprintf(stderr, gettext("invalid option '%c'\n"), optopt); } usage(B_FALSE); } } argc -= optind; argv += optind; get_interval_count(&argc, argv, &interval, &count); if (argc == 0) cb.cb_allpools = B_TRUE; cb.cb_first = B_TRUE; cb.cb_print_status = B_TRUE; for (;;) { if (timestamp_fmt != NODATE) print_timestamp(timestamp_fmt); if (cmd != NULL) cb.vcdl = all_pools_for_each_vdev_run(argc, argv, cmd, NULL, NULL, 0, 0); ret = for_each_pool(argc, argv, B_TRUE, NULL, ZFS_TYPE_POOL, cb.cb_literal, status_callback, &cb); if (cb.vcdl != NULL) free_vdev_cmd_data_list(cb.vcdl); if (argc == 0 && cb.cb_count == 0) (void) fprintf(stderr, gettext("no pools available\n")); else if (cb.cb_explain && cb.cb_first && cb.cb_allpools) (void) printf(gettext("all pools are healthy\n")); if (ret != 0) return (ret); if (interval == 0) break; if (count != 0 && --count == 0) break; (void) fsleep(interval); } return (0); } typedef struct upgrade_cbdata { int cb_first; int cb_argc; uint64_t cb_version; char **cb_argv; } upgrade_cbdata_t; static int check_unsupp_fs(zfs_handle_t *zhp, void *unsupp_fs) { int zfs_version = (int)zfs_prop_get_int(zhp, ZFS_PROP_VERSION); int *count = (int *)unsupp_fs; if (zfs_version > ZPL_VERSION) { (void) printf(gettext("%s (v%d) is not supported by this " "implementation of ZFS.\n"), zfs_get_name(zhp), zfs_version); (*count)++; } zfs_iter_filesystems_v2(zhp, 0, check_unsupp_fs, unsupp_fs); zfs_close(zhp); return (0); } static int upgrade_version(zpool_handle_t *zhp, uint64_t version) { int ret; nvlist_t *config; uint64_t oldversion; int unsupp_fs = 0; config = zpool_get_config(zhp, NULL); verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION, &oldversion) == 0); char compat[ZFS_MAXPROPLEN]; if (zpool_get_prop(zhp, ZPOOL_PROP_COMPATIBILITY, compat, ZFS_MAXPROPLEN, NULL, B_FALSE) != 0) compat[0] = '\0'; assert(SPA_VERSION_IS_SUPPORTED(oldversion)); assert(oldversion < version); ret = zfs_iter_root(zpool_get_handle(zhp), check_unsupp_fs, &unsupp_fs); if (ret != 0) return (ret); if (unsupp_fs) { (void) fprintf(stderr, gettext("Upgrade not performed due " "to %d unsupported filesystems (max v%d).\n"), unsupp_fs, (int)ZPL_VERSION); return (1); } if (strcmp(compat, ZPOOL_COMPAT_LEGACY) == 0) { (void) fprintf(stderr, gettext("Upgrade not performed because " "'compatibility' property set to '" ZPOOL_COMPAT_LEGACY "'.\n")); return (1); } ret = zpool_upgrade(zhp, version); if (ret != 0) return (ret); if (version >= SPA_VERSION_FEATURES) { (void) printf(gettext("Successfully upgraded " "'%s' from version %llu to feature flags.\n"), zpool_get_name(zhp), (u_longlong_t)oldversion); } else { (void) printf(gettext("Successfully upgraded " "'%s' from version %llu to version %llu.\n"), zpool_get_name(zhp), (u_longlong_t)oldversion, (u_longlong_t)version); } return (0); } static int upgrade_enable_all(zpool_handle_t *zhp, int *countp) { int i, ret, count; boolean_t firstff = B_TRUE; nvlist_t *enabled = zpool_get_features(zhp); char compat[ZFS_MAXPROPLEN]; if (zpool_get_prop(zhp, ZPOOL_PROP_COMPATIBILITY, compat, ZFS_MAXPROPLEN, NULL, B_FALSE) != 0) compat[0] = '\0'; boolean_t requested_features[SPA_FEATURES]; if (zpool_do_load_compat(compat, requested_features) != ZPOOL_COMPATIBILITY_OK) return (-1); count = 0; for (i = 0; i < SPA_FEATURES; i++) { const char *fname = spa_feature_table[i].fi_uname; const char *fguid = spa_feature_table[i].fi_guid; if (!spa_feature_table[i].fi_zfs_mod_supported) continue; if (!nvlist_exists(enabled, fguid) && requested_features[i]) { char *propname; verify(-1 != asprintf(&propname, "feature@%s", fname)); ret = zpool_set_prop(zhp, propname, ZFS_FEATURE_ENABLED); if (ret != 0) { free(propname); return (ret); } count++; if (firstff) { (void) printf(gettext("Enabled the " "following features on '%s':\n"), zpool_get_name(zhp)); firstff = B_FALSE; } (void) printf(gettext(" %s\n"), fname); free(propname); } } if (countp != NULL) *countp = count; return (0); } static int upgrade_cb(zpool_handle_t *zhp, void *arg) { upgrade_cbdata_t *cbp = arg; nvlist_t *config; uint64_t version; boolean_t modified_pool = B_FALSE; int ret; config = zpool_get_config(zhp, NULL); verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION, &version) == 0); assert(SPA_VERSION_IS_SUPPORTED(version)); if (version < cbp->cb_version) { cbp->cb_first = B_FALSE; ret = upgrade_version(zhp, cbp->cb_version); if (ret != 0) return (ret); modified_pool = B_TRUE; /* * If they did "zpool upgrade -a", then we could * be doing ioctls to different pools. We need * to log this history once to each pool, and bypass * the normal history logging that happens in main(). */ (void) zpool_log_history(g_zfs, history_str); log_history = B_FALSE; } if (cbp->cb_version >= SPA_VERSION_FEATURES) { int count; ret = upgrade_enable_all(zhp, &count); if (ret != 0) return (ret); if (count > 0) { cbp->cb_first = B_FALSE; modified_pool = B_TRUE; } } if (modified_pool) { (void) printf("\n"); (void) after_zpool_upgrade(zhp); } return (0); } static int upgrade_list_older_cb(zpool_handle_t *zhp, void *arg) { upgrade_cbdata_t *cbp = arg; nvlist_t *config; uint64_t version; config = zpool_get_config(zhp, NULL); verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION, &version) == 0); assert(SPA_VERSION_IS_SUPPORTED(version)); if (version < SPA_VERSION_FEATURES) { if (cbp->cb_first) { (void) printf(gettext("The following pools are " "formatted with legacy version numbers and can\n" "be upgraded to use feature flags. After " "being upgraded, these pools\nwill no " "longer be accessible by software that does not " "support feature\nflags.\n\n" "Note that setting a pool's 'compatibility' " "feature to '" ZPOOL_COMPAT_LEGACY "' will\n" "inhibit upgrades.\n\n")); (void) printf(gettext("VER POOL\n")); (void) printf(gettext("--- ------------\n")); cbp->cb_first = B_FALSE; } (void) printf("%2llu %s\n", (u_longlong_t)version, zpool_get_name(zhp)); } return (0); } static int upgrade_list_disabled_cb(zpool_handle_t *zhp, void *arg) { upgrade_cbdata_t *cbp = arg; nvlist_t *config; uint64_t version; config = zpool_get_config(zhp, NULL); verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION, &version) == 0); if (version >= SPA_VERSION_FEATURES) { int i; boolean_t poolfirst = B_TRUE; nvlist_t *enabled = zpool_get_features(zhp); for (i = 0; i < SPA_FEATURES; i++) { const char *fguid = spa_feature_table[i].fi_guid; const char *fname = spa_feature_table[i].fi_uname; if (!spa_feature_table[i].fi_zfs_mod_supported) continue; if (!nvlist_exists(enabled, fguid)) { if (cbp->cb_first) { (void) printf(gettext("\nSome " "supported features are not " "enabled on the following pools. " "Once a\nfeature is enabled the " "pool may become incompatible with " "software\nthat does not support " "the feature. See " "zpool-features(7) for " "details.\n\n" "Note that the pool " "'compatibility' feature can be " "used to inhibit\nfeature " "upgrades.\n\n")); (void) printf(gettext("POOL " "FEATURE\n")); (void) printf(gettext("------" "---------\n")); cbp->cb_first = B_FALSE; } if (poolfirst) { (void) printf(gettext("%s\n"), zpool_get_name(zhp)); poolfirst = B_FALSE; } (void) printf(gettext(" %s\n"), fname); } /* * If they did "zpool upgrade -a", then we could * be doing ioctls to different pools. We need * to log this history once to each pool, and bypass * the normal history logging that happens in main(). */ (void) zpool_log_history(g_zfs, history_str); log_history = B_FALSE; } } return (0); } static int upgrade_one(zpool_handle_t *zhp, void *data) { boolean_t modified_pool = B_FALSE; upgrade_cbdata_t *cbp = data; uint64_t cur_version; int ret; if (strcmp("log", zpool_get_name(zhp)) == 0) { (void) fprintf(stderr, gettext("'log' is now a reserved word\n" "Pool 'log' must be renamed using export and import" " to upgrade.\n")); return (1); } cur_version = zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL); if (cur_version > cbp->cb_version) { (void) printf(gettext("Pool '%s' is already formatted " "using more current version '%llu'.\n\n"), zpool_get_name(zhp), (u_longlong_t)cur_version); return (0); } if (cbp->cb_version != SPA_VERSION && cur_version == cbp->cb_version) { (void) printf(gettext("Pool '%s' is already formatted " "using version %llu.\n\n"), zpool_get_name(zhp), (u_longlong_t)cbp->cb_version); return (0); } if (cur_version != cbp->cb_version) { modified_pool = B_TRUE; ret = upgrade_version(zhp, cbp->cb_version); if (ret != 0) return (ret); } if (cbp->cb_version >= SPA_VERSION_FEATURES) { int count = 0; ret = upgrade_enable_all(zhp, &count); if (ret != 0) return (ret); if (count != 0) { modified_pool = B_TRUE; } else if (cur_version == SPA_VERSION) { (void) printf(gettext("Pool '%s' already has all " "supported and requested features enabled.\n"), zpool_get_name(zhp)); } } if (modified_pool) { (void) printf("\n"); (void) after_zpool_upgrade(zhp); } return (0); } /* * zpool upgrade * zpool upgrade -v * zpool upgrade [-V version] <-a | pool ...> * * With no arguments, display downrev'd ZFS pool available for upgrade. * Individual pools can be upgraded by specifying the pool, and '-a' will * upgrade all pools. */ int zpool_do_upgrade(int argc, char **argv) { int c; upgrade_cbdata_t cb = { 0 }; int ret = 0; boolean_t showversions = B_FALSE; boolean_t upgradeall = B_FALSE; char *end; /* check options */ while ((c = getopt(argc, argv, ":avV:")) != -1) { switch (c) { case 'a': upgradeall = B_TRUE; break; case 'v': showversions = B_TRUE; break; case 'V': cb.cb_version = strtoll(optarg, &end, 10); if (*end != '\0' || !SPA_VERSION_IS_SUPPORTED(cb.cb_version)) { (void) fprintf(stderr, gettext("invalid version '%s'\n"), optarg); usage(B_FALSE); } break; case ':': (void) fprintf(stderr, gettext("missing argument for " "'%c' option\n"), optopt); usage(B_FALSE); break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } cb.cb_argc = argc; cb.cb_argv = argv; argc -= optind; argv += optind; if (cb.cb_version == 0) { cb.cb_version = SPA_VERSION; } else if (!upgradeall && argc == 0) { (void) fprintf(stderr, gettext("-V option is " "incompatible with other arguments\n")); usage(B_FALSE); } if (showversions) { if (upgradeall || argc != 0) { (void) fprintf(stderr, gettext("-v option is " "incompatible with other arguments\n")); usage(B_FALSE); } } else if (upgradeall) { if (argc != 0) { (void) fprintf(stderr, gettext("-a option should not " "be used along with a pool name\n")); usage(B_FALSE); } } (void) printf("%s", gettext("This system supports ZFS pool feature " "flags.\n\n")); if (showversions) { int i; (void) printf(gettext("The following features are " "supported:\n\n")); (void) printf(gettext("FEAT DESCRIPTION\n")); (void) printf("----------------------------------------------" "---------------\n"); for (i = 0; i < SPA_FEATURES; i++) { zfeature_info_t *fi = &spa_feature_table[i]; if (!fi->fi_zfs_mod_supported) continue; const char *ro = (fi->fi_flags & ZFEATURE_FLAG_READONLY_COMPAT) ? " (read-only compatible)" : ""; (void) printf("%-37s%s\n", fi->fi_uname, ro); (void) printf(" %s\n", fi->fi_desc); } (void) printf("\n"); (void) printf(gettext("The following legacy versions are also " "supported:\n\n")); (void) printf(gettext("VER DESCRIPTION\n")); (void) printf("--- -----------------------------------------" "---------------\n"); (void) printf(gettext(" 1 Initial ZFS version\n")); (void) printf(gettext(" 2 Ditto blocks " "(replicated metadata)\n")); (void) printf(gettext(" 3 Hot spares and double parity " "RAID-Z\n")); (void) printf(gettext(" 4 zpool history\n")); (void) printf(gettext(" 5 Compression using the gzip " "algorithm\n")); (void) printf(gettext(" 6 bootfs pool property\n")); (void) printf(gettext(" 7 Separate intent log devices\n")); (void) printf(gettext(" 8 Delegated administration\n")); (void) printf(gettext(" 9 refquota and refreservation " "properties\n")); (void) printf(gettext(" 10 Cache devices\n")); (void) printf(gettext(" 11 Improved scrub performance\n")); (void) printf(gettext(" 12 Snapshot properties\n")); (void) printf(gettext(" 13 snapused property\n")); (void) printf(gettext(" 14 passthrough-x aclinherit\n")); (void) printf(gettext(" 15 user/group space accounting\n")); (void) printf(gettext(" 16 stmf property support\n")); (void) printf(gettext(" 17 Triple-parity RAID-Z\n")); (void) printf(gettext(" 18 Snapshot user holds\n")); (void) printf(gettext(" 19 Log device removal\n")); (void) printf(gettext(" 20 Compression using zle " "(zero-length encoding)\n")); (void) printf(gettext(" 21 Deduplication\n")); (void) printf(gettext(" 22 Received properties\n")); (void) printf(gettext(" 23 Slim ZIL\n")); (void) printf(gettext(" 24 System attributes\n")); (void) printf(gettext(" 25 Improved scrub stats\n")); (void) printf(gettext(" 26 Improved snapshot deletion " "performance\n")); (void) printf(gettext(" 27 Improved snapshot creation " "performance\n")); (void) printf(gettext(" 28 Multiple vdev replacements\n")); (void) printf(gettext("\nFor more information on a particular " "version, including supported releases,\n")); (void) printf(gettext("see the ZFS Administration Guide.\n\n")); } else if (argc == 0 && upgradeall) { cb.cb_first = B_TRUE; ret = zpool_iter(g_zfs, upgrade_cb, &cb); if (ret == 0 && cb.cb_first) { if (cb.cb_version == SPA_VERSION) { (void) printf(gettext("All pools are already " "formatted using feature flags.\n\n")); (void) printf(gettext("Every feature flags " "pool already has all supported and " "requested features enabled.\n")); } else { (void) printf(gettext("All pools are already " "formatted with version %llu or higher.\n"), (u_longlong_t)cb.cb_version); } } } else if (argc == 0) { cb.cb_first = B_TRUE; ret = zpool_iter(g_zfs, upgrade_list_older_cb, &cb); assert(ret == 0); if (cb.cb_first) { (void) printf(gettext("All pools are formatted " "using feature flags.\n\n")); } else { (void) printf(gettext("\nUse 'zpool upgrade -v' " "for a list of available legacy versions.\n")); } cb.cb_first = B_TRUE; ret = zpool_iter(g_zfs, upgrade_list_disabled_cb, &cb); assert(ret == 0); if (cb.cb_first) { (void) printf(gettext("Every feature flags pool has " "all supported and requested features enabled.\n")); } else { (void) printf(gettext("\n")); } } else { ret = for_each_pool(argc, argv, B_FALSE, NULL, ZFS_TYPE_POOL, B_FALSE, upgrade_one, &cb); } return (ret); } typedef struct hist_cbdata { boolean_t first; boolean_t longfmt; boolean_t internal; } hist_cbdata_t; static void print_history_records(nvlist_t *nvhis, hist_cbdata_t *cb) { nvlist_t **records; uint_t numrecords; int i; verify(nvlist_lookup_nvlist_array(nvhis, ZPOOL_HIST_RECORD, &records, &numrecords) == 0); for (i = 0; i < numrecords; i++) { nvlist_t *rec = records[i]; char tbuf[64] = ""; if (nvlist_exists(rec, ZPOOL_HIST_TIME)) { time_t tsec; struct tm t; tsec = fnvlist_lookup_uint64(records[i], ZPOOL_HIST_TIME); (void) localtime_r(&tsec, &t); (void) strftime(tbuf, sizeof (tbuf), "%F.%T", &t); } if (nvlist_exists(rec, ZPOOL_HIST_ELAPSED_NS)) { uint64_t elapsed_ns = fnvlist_lookup_int64(records[i], ZPOOL_HIST_ELAPSED_NS); (void) snprintf(tbuf + strlen(tbuf), sizeof (tbuf) - strlen(tbuf), " (%lldms)", (long long)elapsed_ns / 1000 / 1000); } if (nvlist_exists(rec, ZPOOL_HIST_CMD)) { (void) printf("%s %s", tbuf, fnvlist_lookup_string(rec, ZPOOL_HIST_CMD)); } else if (nvlist_exists(rec, ZPOOL_HIST_INT_EVENT)) { int ievent = fnvlist_lookup_uint64(rec, ZPOOL_HIST_INT_EVENT); if (!cb->internal) continue; if (ievent >= ZFS_NUM_LEGACY_HISTORY_EVENTS) { (void) printf("%s unrecognized record:\n", tbuf); dump_nvlist(rec, 4); continue; } (void) printf("%s [internal %s txg:%lld] %s", tbuf, zfs_history_event_names[ievent], (longlong_t)fnvlist_lookup_uint64( rec, ZPOOL_HIST_TXG), fnvlist_lookup_string(rec, ZPOOL_HIST_INT_STR)); } else if (nvlist_exists(rec, ZPOOL_HIST_INT_NAME)) { if (!cb->internal) continue; (void) printf("%s [txg:%lld] %s", tbuf, (longlong_t)fnvlist_lookup_uint64( rec, ZPOOL_HIST_TXG), fnvlist_lookup_string(rec, ZPOOL_HIST_INT_NAME)); if (nvlist_exists(rec, ZPOOL_HIST_DSNAME)) { (void) printf(" %s (%llu)", fnvlist_lookup_string(rec, ZPOOL_HIST_DSNAME), (u_longlong_t)fnvlist_lookup_uint64(rec, ZPOOL_HIST_DSID)); } (void) printf(" %s", fnvlist_lookup_string(rec, ZPOOL_HIST_INT_STR)); } else if (nvlist_exists(rec, ZPOOL_HIST_IOCTL)) { if (!cb->internal) continue; (void) printf("%s ioctl %s\n", tbuf, fnvlist_lookup_string(rec, ZPOOL_HIST_IOCTL)); if (nvlist_exists(rec, ZPOOL_HIST_INPUT_NVL)) { (void) printf(" input:\n"); dump_nvlist(fnvlist_lookup_nvlist(rec, ZPOOL_HIST_INPUT_NVL), 8); } if (nvlist_exists(rec, ZPOOL_HIST_OUTPUT_NVL)) { (void) printf(" output:\n"); dump_nvlist(fnvlist_lookup_nvlist(rec, ZPOOL_HIST_OUTPUT_NVL), 8); } if (nvlist_exists(rec, ZPOOL_HIST_OUTPUT_SIZE)) { (void) printf(" output nvlist omitted; " "original size: %lldKB\n", (longlong_t)fnvlist_lookup_int64(rec, ZPOOL_HIST_OUTPUT_SIZE) / 1024); } if (nvlist_exists(rec, ZPOOL_HIST_ERRNO)) { (void) printf(" errno: %lld\n", (longlong_t)fnvlist_lookup_int64(rec, ZPOOL_HIST_ERRNO)); } } else { if (!cb->internal) continue; (void) printf("%s unrecognized record:\n", tbuf); dump_nvlist(rec, 4); } if (!cb->longfmt) { (void) printf("\n"); continue; } (void) printf(" ["); if (nvlist_exists(rec, ZPOOL_HIST_WHO)) { uid_t who = fnvlist_lookup_uint64(rec, ZPOOL_HIST_WHO); struct passwd *pwd = getpwuid(who); (void) printf("user %d ", (int)who); if (pwd != NULL) (void) printf("(%s) ", pwd->pw_name); } if (nvlist_exists(rec, ZPOOL_HIST_HOST)) { (void) printf("on %s", fnvlist_lookup_string(rec, ZPOOL_HIST_HOST)); } if (nvlist_exists(rec, ZPOOL_HIST_ZONE)) { (void) printf(":%s", fnvlist_lookup_string(rec, ZPOOL_HIST_ZONE)); } (void) printf("]"); (void) printf("\n"); } } /* * Print out the command history for a specific pool. */ static int get_history_one(zpool_handle_t *zhp, void *data) { nvlist_t *nvhis; int ret; hist_cbdata_t *cb = (hist_cbdata_t *)data; uint64_t off = 0; boolean_t eof = B_FALSE; cb->first = B_FALSE; (void) printf(gettext("History for '%s':\n"), zpool_get_name(zhp)); while (!eof) { if ((ret = zpool_get_history(zhp, &nvhis, &off, &eof)) != 0) return (ret); print_history_records(nvhis, cb); nvlist_free(nvhis); } (void) printf("\n"); return (ret); } /* * zpool history * * Displays the history of commands that modified pools. */ int zpool_do_history(int argc, char **argv) { hist_cbdata_t cbdata = { 0 }; int ret; int c; cbdata.first = B_TRUE; /* check options */ while ((c = getopt(argc, argv, "li")) != -1) { switch (c) { case 'l': cbdata.longfmt = B_TRUE; break; case 'i': cbdata.internal = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; ret = for_each_pool(argc, argv, B_FALSE, NULL, ZFS_TYPE_POOL, B_FALSE, get_history_one, &cbdata); if (argc == 0 && cbdata.first == B_TRUE) { (void) fprintf(stderr, gettext("no pools available\n")); return (0); } return (ret); } typedef struct ev_opts { int verbose; int scripted; int follow; int clear; char poolname[ZFS_MAX_DATASET_NAME_LEN]; } ev_opts_t; static void zpool_do_events_short(nvlist_t *nvl, ev_opts_t *opts) { char ctime_str[26], str[32]; const char *ptr; int64_t *tv; uint_t n; verify(nvlist_lookup_int64_array(nvl, FM_EREPORT_TIME, &tv, &n) == 0); memset(str, ' ', 32); (void) ctime_r((const time_t *)&tv[0], ctime_str); (void) memcpy(str, ctime_str+4, 6); /* 'Jun 30' */ (void) memcpy(str+7, ctime_str+20, 4); /* '1993' */ (void) memcpy(str+12, ctime_str+11, 8); /* '21:49:08' */ (void) sprintf(str+20, ".%09lld", (longlong_t)tv[1]); /* '.123456789' */ if (opts->scripted) (void) printf(gettext("%s\t"), str); else (void) printf(gettext("%s "), str); verify(nvlist_lookup_string(nvl, FM_CLASS, &ptr) == 0); (void) printf(gettext("%s\n"), ptr); } static void zpool_do_events_nvprint(nvlist_t *nvl, int depth) { nvpair_t *nvp; for (nvp = nvlist_next_nvpair(nvl, NULL); nvp != NULL; nvp = nvlist_next_nvpair(nvl, nvp)) { data_type_t type = nvpair_type(nvp); const char *name = nvpair_name(nvp); boolean_t b; uint8_t i8; uint16_t i16; uint32_t i32; uint64_t i64; const char *str; nvlist_t *cnv; printf(gettext("%*s%s = "), depth, "", name); switch (type) { case DATA_TYPE_BOOLEAN: printf(gettext("%s"), "1"); break; case DATA_TYPE_BOOLEAN_VALUE: (void) nvpair_value_boolean_value(nvp, &b); printf(gettext("%s"), b ? "1" : "0"); break; case DATA_TYPE_BYTE: (void) nvpair_value_byte(nvp, &i8); printf(gettext("0x%x"), i8); break; case DATA_TYPE_INT8: (void) nvpair_value_int8(nvp, (void *)&i8); printf(gettext("0x%x"), i8); break; case DATA_TYPE_UINT8: (void) nvpair_value_uint8(nvp, &i8); printf(gettext("0x%x"), i8); break; case DATA_TYPE_INT16: (void) nvpair_value_int16(nvp, (void *)&i16); printf(gettext("0x%x"), i16); break; case DATA_TYPE_UINT16: (void) nvpair_value_uint16(nvp, &i16); printf(gettext("0x%x"), i16); break; case DATA_TYPE_INT32: (void) nvpair_value_int32(nvp, (void *)&i32); printf(gettext("0x%x"), i32); break; case DATA_TYPE_UINT32: (void) nvpair_value_uint32(nvp, &i32); printf(gettext("0x%x"), i32); break; case DATA_TYPE_INT64: (void) nvpair_value_int64(nvp, (void *)&i64); printf(gettext("0x%llx"), (u_longlong_t)i64); break; case DATA_TYPE_UINT64: (void) nvpair_value_uint64(nvp, &i64); /* * translate vdev state values to readable * strings to aide zpool events consumers */ if (strcmp(name, FM_EREPORT_PAYLOAD_ZFS_VDEV_STATE) == 0 || strcmp(name, FM_EREPORT_PAYLOAD_ZFS_VDEV_LASTSTATE) == 0) { printf(gettext("\"%s\" (0x%llx)"), zpool_state_to_name(i64, VDEV_AUX_NONE), (u_longlong_t)i64); } else { printf(gettext("0x%llx"), (u_longlong_t)i64); } break; case DATA_TYPE_HRTIME: (void) nvpair_value_hrtime(nvp, (void *)&i64); printf(gettext("0x%llx"), (u_longlong_t)i64); break; case DATA_TYPE_STRING: (void) nvpair_value_string(nvp, &str); printf(gettext("\"%s\""), str ? str : ""); break; case DATA_TYPE_NVLIST: printf(gettext("(embedded nvlist)\n")); (void) nvpair_value_nvlist(nvp, &cnv); zpool_do_events_nvprint(cnv, depth + 8); printf(gettext("%*s(end %s)"), depth, "", name); break; case DATA_TYPE_NVLIST_ARRAY: { nvlist_t **val; uint_t i, nelem; (void) nvpair_value_nvlist_array(nvp, &val, &nelem); printf(gettext("(%d embedded nvlists)\n"), nelem); for (i = 0; i < nelem; i++) { printf(gettext("%*s%s[%d] = %s\n"), depth, "", name, i, "(embedded nvlist)"); zpool_do_events_nvprint(val[i], depth + 8); printf(gettext("%*s(end %s[%i])\n"), depth, "", name, i); } printf(gettext("%*s(end %s)\n"), depth, "", name); } break; case DATA_TYPE_INT8_ARRAY: { int8_t *val; uint_t i, nelem; (void) nvpair_value_int8_array(nvp, &val, &nelem); for (i = 0; i < nelem; i++) printf(gettext("0x%x "), val[i]); break; } case DATA_TYPE_UINT8_ARRAY: { uint8_t *val; uint_t i, nelem; (void) nvpair_value_uint8_array(nvp, &val, &nelem); for (i = 0; i < nelem; i++) printf(gettext("0x%x "), val[i]); break; } case DATA_TYPE_INT16_ARRAY: { int16_t *val; uint_t i, nelem; (void) nvpair_value_int16_array(nvp, &val, &nelem); for (i = 0; i < nelem; i++) printf(gettext("0x%x "), val[i]); break; } case DATA_TYPE_UINT16_ARRAY: { uint16_t *val; uint_t i, nelem; (void) nvpair_value_uint16_array(nvp, &val, &nelem); for (i = 0; i < nelem; i++) printf(gettext("0x%x "), val[i]); break; } case DATA_TYPE_INT32_ARRAY: { int32_t *val; uint_t i, nelem; (void) nvpair_value_int32_array(nvp, &val, &nelem); for (i = 0; i < nelem; i++) printf(gettext("0x%x "), val[i]); break; } case DATA_TYPE_UINT32_ARRAY: { uint32_t *val; uint_t i, nelem; (void) nvpair_value_uint32_array(nvp, &val, &nelem); for (i = 0; i < nelem; i++) printf(gettext("0x%x "), val[i]); break; } case DATA_TYPE_INT64_ARRAY: { int64_t *val; uint_t i, nelem; (void) nvpair_value_int64_array(nvp, &val, &nelem); for (i = 0; i < nelem; i++) printf(gettext("0x%llx "), (u_longlong_t)val[i]); break; } case DATA_TYPE_UINT64_ARRAY: { uint64_t *val; uint_t i, nelem; (void) nvpair_value_uint64_array(nvp, &val, &nelem); for (i = 0; i < nelem; i++) printf(gettext("0x%llx "), (u_longlong_t)val[i]); break; } case DATA_TYPE_STRING_ARRAY: { const char **str; uint_t i, nelem; (void) nvpair_value_string_array(nvp, &str, &nelem); for (i = 0; i < nelem; i++) printf(gettext("\"%s\" "), str[i] ? str[i] : ""); break; } case DATA_TYPE_BOOLEAN_ARRAY: case DATA_TYPE_BYTE_ARRAY: case DATA_TYPE_DOUBLE: case DATA_TYPE_DONTCARE: case DATA_TYPE_UNKNOWN: printf(gettext("")); break; } printf(gettext("\n")); } } static int zpool_do_events_next(ev_opts_t *opts) { nvlist_t *nvl; int zevent_fd, ret, dropped; const char *pool; zevent_fd = open(ZFS_DEV, O_RDWR); VERIFY(zevent_fd >= 0); if (!opts->scripted) (void) printf(gettext("%-30s %s\n"), "TIME", "CLASS"); while (1) { ret = zpool_events_next(g_zfs, &nvl, &dropped, (opts->follow ? ZEVENT_NONE : ZEVENT_NONBLOCK), zevent_fd); if (ret || nvl == NULL) break; if (dropped > 0) (void) printf(gettext("dropped %d events\n"), dropped); if (strlen(opts->poolname) > 0 && nvlist_lookup_string(nvl, FM_FMRI_ZFS_POOL, &pool) == 0 && strcmp(opts->poolname, pool) != 0) continue; zpool_do_events_short(nvl, opts); if (opts->verbose) { zpool_do_events_nvprint(nvl, 8); printf(gettext("\n")); } (void) fflush(stdout); nvlist_free(nvl); } VERIFY(0 == close(zevent_fd)); return (ret); } static int zpool_do_events_clear(void) { int count, ret; ret = zpool_events_clear(g_zfs, &count); if (!ret) (void) printf(gettext("cleared %d events\n"), count); return (ret); } /* * zpool events [-vHf [pool] | -c] * * Displays events logs by ZFS. */ int zpool_do_events(int argc, char **argv) { ev_opts_t opts = { 0 }; int ret; int c; /* check options */ while ((c = getopt(argc, argv, "vHfc")) != -1) { switch (c) { case 'v': opts.verbose = 1; break; case 'H': opts.scripted = 1; break; case 'f': opts.follow = 1; break; case 'c': opts.clear = 1; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; if (argc > 1) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } else if (argc == 1) { (void) strlcpy(opts.poolname, argv[0], sizeof (opts.poolname)); if (!zfs_name_valid(opts.poolname, ZFS_TYPE_POOL)) { (void) fprintf(stderr, gettext("invalid pool name '%s'\n"), opts.poolname); usage(B_FALSE); } } if ((argc == 1 || opts.verbose || opts.scripted || opts.follow) && opts.clear) { (void) fprintf(stderr, gettext("invalid options combined with -c\n")); usage(B_FALSE); } if (opts.clear) ret = zpool_do_events_clear(); else ret = zpool_do_events_next(&opts); return (ret); } static int get_callback_vdev(zpool_handle_t *zhp, char *vdevname, void *data) { zprop_get_cbdata_t *cbp = (zprop_get_cbdata_t *)data; char value[ZFS_MAXPROPLEN]; zprop_source_t srctype; for (zprop_list_t *pl = cbp->cb_proplist; pl != NULL; pl = pl->pl_next) { char *prop_name; /* * If the first property is pool name, it is a special * placeholder that we can skip. This will also skip * over the name property when 'all' is specified. */ if (pl->pl_prop == ZPOOL_PROP_NAME && pl == cbp->cb_proplist) continue; if (pl->pl_prop == ZPROP_INVAL) { prop_name = pl->pl_user_prop; } else { prop_name = (char *)vdev_prop_to_name(pl->pl_prop); } if (zpool_get_vdev_prop(zhp, vdevname, pl->pl_prop, prop_name, value, sizeof (value), &srctype, cbp->cb_literal) == 0) { zprop_print_one_property(vdevname, cbp, prop_name, value, srctype, NULL, NULL); } } return (0); } static int get_callback_vdev_cb(void *zhp_data, nvlist_t *nv, void *data) { zpool_handle_t *zhp = zhp_data; zprop_get_cbdata_t *cbp = (zprop_get_cbdata_t *)data; char *vdevname; const char *type; int ret; /* * zpool_vdev_name() transforms the root vdev name (i.e., root-0) to the * pool name for display purposes, which is not desired. Fallback to * zpool_vdev_name() when not dealing with the root vdev. */ type = fnvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE); if (zhp != NULL && strcmp(type, "root") == 0) vdevname = strdup("root-0"); else vdevname = zpool_vdev_name(g_zfs, zhp, nv, cbp->cb_vdevs.cb_name_flags); (void) vdev_expand_proplist(zhp, vdevname, &cbp->cb_proplist); ret = get_callback_vdev(zhp, vdevname, data); free(vdevname); return (ret); } static int get_callback(zpool_handle_t *zhp, void *data) { zprop_get_cbdata_t *cbp = (zprop_get_cbdata_t *)data; char value[ZFS_MAXPROPLEN]; zprop_source_t srctype; zprop_list_t *pl; int vid; if (cbp->cb_type == ZFS_TYPE_VDEV) { if (strcmp(cbp->cb_vdevs.cb_names[0], "all-vdevs") == 0) { for_each_vdev(zhp, get_callback_vdev_cb, data); } else { /* Adjust column widths for vdev properties */ for (vid = 0; vid < cbp->cb_vdevs.cb_names_count; vid++) { vdev_expand_proplist(zhp, cbp->cb_vdevs.cb_names[vid], &cbp->cb_proplist); } /* Display the properties */ for (vid = 0; vid < cbp->cb_vdevs.cb_names_count; vid++) { get_callback_vdev(zhp, cbp->cb_vdevs.cb_names[vid], data); } } } else { assert(cbp->cb_type == ZFS_TYPE_POOL); for (pl = cbp->cb_proplist; pl != NULL; pl = pl->pl_next) { /* * Skip the special fake placeholder. This will also * skip over the name property when 'all' is specified. */ if (pl->pl_prop == ZPOOL_PROP_NAME && pl == cbp->cb_proplist) continue; if (pl->pl_prop == ZPROP_INVAL && zfs_prop_user(pl->pl_user_prop)) { srctype = ZPROP_SRC_LOCAL; if (zpool_get_userprop(zhp, pl->pl_user_prop, value, sizeof (value), &srctype) != 0) continue; zprop_print_one_property(zpool_get_name(zhp), cbp, pl->pl_user_prop, value, srctype, NULL, NULL); } else if (pl->pl_prop == ZPROP_INVAL && (zpool_prop_feature(pl->pl_user_prop) || zpool_prop_unsupported(pl->pl_user_prop))) { srctype = ZPROP_SRC_LOCAL; if (zpool_prop_get_feature(zhp, pl->pl_user_prop, value, sizeof (value)) == 0) { zprop_print_one_property( zpool_get_name(zhp), cbp, pl->pl_user_prop, value, srctype, NULL, NULL); } } else { if (zpool_get_prop(zhp, pl->pl_prop, value, sizeof (value), &srctype, cbp->cb_literal) != 0) continue; zprop_print_one_property(zpool_get_name(zhp), cbp, zpool_prop_to_name(pl->pl_prop), value, srctype, NULL, NULL); } } } return (0); } /* * zpool get [-Hp] [-o "all" | field[,...]] <"all" | property[,...]> ... * * -H Scripted mode. Don't display headers, and separate properties * by a single tab. * -o List of columns to display. Defaults to * "name,property,value,source". * -p Display values in parsable (exact) format. * * Get properties of pools in the system. Output space statistics * for each one as well as other attributes. */ int zpool_do_get(int argc, char **argv) { zprop_get_cbdata_t cb = { 0 }; zprop_list_t fake_name = { 0 }; int ret; int c, i; char *propstr = NULL; char *vdev = NULL; cb.cb_first = B_TRUE; /* * Set up default columns and sources. */ cb.cb_sources = ZPROP_SRC_ALL; cb.cb_columns[0] = GET_COL_NAME; cb.cb_columns[1] = GET_COL_PROPERTY; cb.cb_columns[2] = GET_COL_VALUE; cb.cb_columns[3] = GET_COL_SOURCE; cb.cb_type = ZFS_TYPE_POOL; cb.cb_vdevs.cb_name_flags |= VDEV_NAME_TYPE_ID; current_prop_type = cb.cb_type; /* check options */ while ((c = getopt(argc, argv, ":Hpo:")) != -1) { switch (c) { case 'p': cb.cb_literal = B_TRUE; break; case 'H': cb.cb_scripted = B_TRUE; break; case 'o': memset(&cb.cb_columns, 0, sizeof (cb.cb_columns)); i = 0; for (char *tok; (tok = strsep(&optarg, ",")); ) { static const char *const col_opts[] = { "name", "property", "value", "source", "all" }; static const zfs_get_column_t col_cols[] = { GET_COL_NAME, GET_COL_PROPERTY, GET_COL_VALUE, GET_COL_SOURCE }; if (i == ZFS_GET_NCOLS - 1) { (void) fprintf(stderr, gettext("too " "many fields given to -o " "option\n")); usage(B_FALSE); } for (c = 0; c < ARRAY_SIZE(col_opts); ++c) if (strcmp(tok, col_opts[c]) == 0) goto found; (void) fprintf(stderr, gettext("invalid column name '%s'\n"), tok); usage(B_FALSE); found: if (c >= 4) { if (i > 0) { (void) fprintf(stderr, gettext("\"all\" conflicts " "with specific fields " "given to -o option\n")); usage(B_FALSE); } memcpy(cb.cb_columns, col_cols, sizeof (col_cols)); i = ZFS_GET_NCOLS - 1; } else cb.cb_columns[i++] = col_cols[c]; } break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; if (argc < 1) { (void) fprintf(stderr, gettext("missing property " "argument\n")); usage(B_FALSE); } /* Properties list is needed later by zprop_get_list() */ propstr = argv[0]; argc--; argv++; if (argc == 0) { /* No args, so just print the defaults. */ } else if (are_all_pools(argc, argv)) { /* All the args are pool names */ } else if (are_all_pools(1, argv)) { /* The first arg is a pool name */ if ((argc == 2 && strcmp(argv[1], "all-vdevs") == 0) || (argc == 2 && strcmp(argv[1], "root") == 0) || are_vdevs_in_pool(argc - 1, argv + 1, argv[0], &cb.cb_vdevs)) { if (strcmp(argv[1], "root") == 0) vdev = strdup("root-0"); else vdev = strdup(argv[1]); /* ... and the rest are vdev names */ cb.cb_vdevs.cb_names = &vdev; cb.cb_vdevs.cb_names_count = argc - 1; cb.cb_type = ZFS_TYPE_VDEV; argc = 1; /* One pool to process */ } else { fprintf(stderr, gettext("Expected a list of vdevs in" " \"%s\", but got:\n"), argv[0]); error_list_unresolved_vdevs(argc - 1, argv + 1, argv[0], &cb.cb_vdevs); fprintf(stderr, "\n"); usage(B_FALSE); return (1); } } else { /* * The first arg isn't a pool name, */ fprintf(stderr, gettext("missing pool name.\n")); fprintf(stderr, "\n"); usage(B_FALSE); return (1); } if (zprop_get_list(g_zfs, propstr, &cb.cb_proplist, cb.cb_type) != 0) { /* Use correct list of valid properties (pool or vdev) */ current_prop_type = cb.cb_type; usage(B_FALSE); } if (cb.cb_proplist != NULL) { fake_name.pl_prop = ZPOOL_PROP_NAME; fake_name.pl_width = strlen(gettext("NAME")); fake_name.pl_next = cb.cb_proplist; cb.cb_proplist = &fake_name; } ret = for_each_pool(argc, argv, B_TRUE, &cb.cb_proplist, cb.cb_type, cb.cb_literal, get_callback, &cb); if (cb.cb_proplist == &fake_name) zprop_free_list(fake_name.pl_next); else zprop_free_list(cb.cb_proplist); if (vdev != NULL) free(vdev); return (ret); } typedef struct set_cbdata { char *cb_propname; char *cb_value; zfs_type_t cb_type; vdev_cbdata_t cb_vdevs; boolean_t cb_any_successful; } set_cbdata_t; static int set_pool_callback(zpool_handle_t *zhp, set_cbdata_t *cb) { int error; /* Check if we have out-of-bounds features */ if (strcmp(cb->cb_propname, ZPOOL_CONFIG_COMPATIBILITY) == 0) { boolean_t features[SPA_FEATURES]; if (zpool_do_load_compat(cb->cb_value, features) != ZPOOL_COMPATIBILITY_OK) return (-1); nvlist_t *enabled = zpool_get_features(zhp); spa_feature_t i; for (i = 0; i < SPA_FEATURES; i++) { const char *fguid = spa_feature_table[i].fi_guid; if (nvlist_exists(enabled, fguid) && !features[i]) break; } if (i < SPA_FEATURES) (void) fprintf(stderr, gettext("Warning: one or " "more features already enabled on pool '%s'\n" "are not present in this compatibility set.\n"), zpool_get_name(zhp)); } /* if we're setting a feature, check it's in compatibility set */ if (zpool_prop_feature(cb->cb_propname) && strcmp(cb->cb_value, ZFS_FEATURE_ENABLED) == 0) { char *fname = strchr(cb->cb_propname, '@') + 1; spa_feature_t f; if (zfeature_lookup_name(fname, &f) == 0) { char compat[ZFS_MAXPROPLEN]; if (zpool_get_prop(zhp, ZPOOL_PROP_COMPATIBILITY, compat, ZFS_MAXPROPLEN, NULL, B_FALSE) != 0) compat[0] = '\0'; boolean_t features[SPA_FEATURES]; if (zpool_do_load_compat(compat, features) != ZPOOL_COMPATIBILITY_OK) { (void) fprintf(stderr, gettext("Error: " "cannot enable feature '%s' on pool '%s'\n" "because the pool's 'compatibility' " "property cannot be parsed.\n"), fname, zpool_get_name(zhp)); return (-1); } if (!features[f]) { (void) fprintf(stderr, gettext("Error: " "cannot enable feature '%s' on pool '%s'\n" "as it is not specified in this pool's " "current compatibility set.\n" "Consider setting 'compatibility' to a " "less restrictive set, or to 'off'.\n"), fname, zpool_get_name(zhp)); return (-1); } } } error = zpool_set_prop(zhp, cb->cb_propname, cb->cb_value); return (error); } static int set_callback(zpool_handle_t *zhp, void *data) { int error; set_cbdata_t *cb = (set_cbdata_t *)data; if (cb->cb_type == ZFS_TYPE_VDEV) { error = zpool_set_vdev_prop(zhp, *cb->cb_vdevs.cb_names, cb->cb_propname, cb->cb_value); } else { assert(cb->cb_type == ZFS_TYPE_POOL); error = set_pool_callback(zhp, cb); } cb->cb_any_successful = !error; return (error); } int zpool_do_set(int argc, char **argv) { set_cbdata_t cb = { 0 }; int error; char *vdev = NULL; current_prop_type = ZFS_TYPE_POOL; if (argc > 1 && argv[1][0] == '-') { (void) fprintf(stderr, gettext("invalid option '%c'\n"), argv[1][1]); usage(B_FALSE); } if (argc < 2) { (void) fprintf(stderr, gettext("missing property=value " "argument\n")); usage(B_FALSE); } if (argc < 3) { (void) fprintf(stderr, gettext("missing pool name\n")); usage(B_FALSE); } if (argc > 4) { (void) fprintf(stderr, gettext("too many pool names\n")); usage(B_FALSE); } cb.cb_propname = argv[1]; cb.cb_type = ZFS_TYPE_POOL; cb.cb_vdevs.cb_name_flags |= VDEV_NAME_TYPE_ID; cb.cb_value = strchr(cb.cb_propname, '='); if (cb.cb_value == NULL) { (void) fprintf(stderr, gettext("missing value in " "property=value argument\n")); usage(B_FALSE); } *(cb.cb_value) = '\0'; cb.cb_value++; argc -= 2; argv += 2; /* argv[0] is pool name */ if (!is_pool(argv[0])) { (void) fprintf(stderr, gettext("cannot open '%s': is not a pool\n"), argv[0]); return (EINVAL); } /* argv[1], when supplied, is vdev name */ if (argc == 2) { if (strcmp(argv[1], "root") == 0) vdev = strdup("root-0"); else vdev = strdup(argv[1]); if (!are_vdevs_in_pool(1, &vdev, argv[0], &cb.cb_vdevs)) { (void) fprintf(stderr, gettext( "cannot find '%s' in '%s': device not in pool\n"), vdev, argv[0]); free(vdev); return (EINVAL); } cb.cb_vdevs.cb_names = &vdev; cb.cb_vdevs.cb_names_count = 1; cb.cb_type = ZFS_TYPE_VDEV; } error = for_each_pool(1, argv, B_TRUE, NULL, ZFS_TYPE_POOL, B_FALSE, set_callback, &cb); if (vdev != NULL) free(vdev); return (error); } /* Add up the total number of bytes left to initialize/trim across all vdevs */ static uint64_t vdev_activity_remaining(nvlist_t *nv, zpool_wait_activity_t activity) { uint64_t bytes_remaining; nvlist_t **child; uint_t c, children; vdev_stat_t *vs; assert(activity == ZPOOL_WAIT_INITIALIZE || activity == ZPOOL_WAIT_TRIM); verify(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &c) == 0); if (activity == ZPOOL_WAIT_INITIALIZE && vs->vs_initialize_state == VDEV_INITIALIZE_ACTIVE) bytes_remaining = vs->vs_initialize_bytes_est - vs->vs_initialize_bytes_done; else if (activity == ZPOOL_WAIT_TRIM && vs->vs_trim_state == VDEV_TRIM_ACTIVE) bytes_remaining = vs->vs_trim_bytes_est - vs->vs_trim_bytes_done; else bytes_remaining = 0; if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) children = 0; for (c = 0; c < children; c++) bytes_remaining += vdev_activity_remaining(child[c], activity); return (bytes_remaining); } /* Add up the total number of bytes left to rebuild across top-level vdevs */ static uint64_t vdev_activity_top_remaining(nvlist_t *nv) { uint64_t bytes_remaining = 0; nvlist_t **child; uint_t children; int error; if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) children = 0; for (uint_t c = 0; c < children; c++) { vdev_rebuild_stat_t *vrs; uint_t i; error = nvlist_lookup_uint64_array(child[c], ZPOOL_CONFIG_REBUILD_STATS, (uint64_t **)&vrs, &i); if (error == 0) { if (vrs->vrs_state == VDEV_REBUILD_ACTIVE) { bytes_remaining += (vrs->vrs_bytes_est - vrs->vrs_bytes_rebuilt); } } } return (bytes_remaining); } /* Whether any vdevs are 'spare' or 'replacing' vdevs */ static boolean_t vdev_any_spare_replacing(nvlist_t *nv) { nvlist_t **child; uint_t c, children; const char *vdev_type; (void) nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &vdev_type); if (strcmp(vdev_type, VDEV_TYPE_REPLACING) == 0 || strcmp(vdev_type, VDEV_TYPE_SPARE) == 0 || strcmp(vdev_type, VDEV_TYPE_DRAID_SPARE) == 0) { return (B_TRUE); } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) children = 0; for (c = 0; c < children; c++) { if (vdev_any_spare_replacing(child[c])) return (B_TRUE); } return (B_FALSE); } typedef struct wait_data { char *wd_poolname; boolean_t wd_scripted; boolean_t wd_exact; boolean_t wd_headers_once; boolean_t wd_should_exit; /* Which activities to wait for */ boolean_t wd_enabled[ZPOOL_WAIT_NUM_ACTIVITIES]; float wd_interval; pthread_cond_t wd_cv; pthread_mutex_t wd_mutex; } wait_data_t; /* * Print to stdout a single line, containing one column for each activity that * we are waiting for specifying how many bytes of work are left for that * activity. */ static void print_wait_status_row(wait_data_t *wd, zpool_handle_t *zhp, int row) { nvlist_t *config, *nvroot; uint_t c; int i; pool_checkpoint_stat_t *pcs = NULL; pool_scan_stat_t *pss = NULL; pool_removal_stat_t *prs = NULL; const char *const headers[] = {"DISCARD", "FREE", "INITIALIZE", "REPLACE", "REMOVE", "RESILVER", "SCRUB", "TRIM"}; int col_widths[ZPOOL_WAIT_NUM_ACTIVITIES]; /* Calculate the width of each column */ for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++) { /* * Make sure we have enough space in the col for pretty-printed * numbers and for the column header, and then leave a couple * spaces between cols for readability. */ col_widths[i] = MAX(strlen(headers[i]), 6) + 2; } /* Print header if appropriate */ int term_height = terminal_height(); boolean_t reprint_header = (!wd->wd_headers_once && term_height > 0 && row % (term_height-1) == 0); if (!wd->wd_scripted && (row == 0 || reprint_header)) { for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++) { if (wd->wd_enabled[i]) (void) printf("%*s", col_widths[i], headers[i]); } (void) fputc('\n', stdout); } /* Bytes of work remaining in each activity */ int64_t bytes_rem[ZPOOL_WAIT_NUM_ACTIVITIES] = {0}; bytes_rem[ZPOOL_WAIT_FREE] = zpool_get_prop_int(zhp, ZPOOL_PROP_FREEING, NULL); config = zpool_get_config(zhp, NULL); nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE); (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_CHECKPOINT_STATS, (uint64_t **)&pcs, &c); if (pcs != NULL && pcs->pcs_state == CS_CHECKPOINT_DISCARDING) bytes_rem[ZPOOL_WAIT_CKPT_DISCARD] = pcs->pcs_space; (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_REMOVAL_STATS, (uint64_t **)&prs, &c); if (prs != NULL && prs->prs_state == DSS_SCANNING) bytes_rem[ZPOOL_WAIT_REMOVE] = prs->prs_to_copy - prs->prs_copied; (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_SCAN_STATS, (uint64_t **)&pss, &c); if (pss != NULL && pss->pss_state == DSS_SCANNING && pss->pss_pass_scrub_pause == 0) { int64_t rem = pss->pss_to_examine - pss->pss_issued; if (pss->pss_func == POOL_SCAN_SCRUB) bytes_rem[ZPOOL_WAIT_SCRUB] = rem; else bytes_rem[ZPOOL_WAIT_RESILVER] = rem; } else if (check_rebuilding(nvroot, NULL)) { bytes_rem[ZPOOL_WAIT_RESILVER] = vdev_activity_top_remaining(nvroot); } bytes_rem[ZPOOL_WAIT_INITIALIZE] = vdev_activity_remaining(nvroot, ZPOOL_WAIT_INITIALIZE); bytes_rem[ZPOOL_WAIT_TRIM] = vdev_activity_remaining(nvroot, ZPOOL_WAIT_TRIM); /* * A replace finishes after resilvering finishes, so the amount of work * left for a replace is the same as for resilvering. * * It isn't quite correct to say that if we have any 'spare' or * 'replacing' vdevs and a resilver is happening, then a replace is in * progress, like we do here. When a hot spare is used, the faulted vdev * is not removed after the hot spare is resilvered, so parent 'spare' * vdev is not removed either. So we could have a 'spare' vdev, but be * resilvering for a different reason. However, we use it as a heuristic * because we don't have access to the DTLs, which could tell us whether * or not we have really finished resilvering a hot spare. */ if (vdev_any_spare_replacing(nvroot)) bytes_rem[ZPOOL_WAIT_REPLACE] = bytes_rem[ZPOOL_WAIT_RESILVER]; if (timestamp_fmt != NODATE) print_timestamp(timestamp_fmt); for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++) { char buf[64]; if (!wd->wd_enabled[i]) continue; if (wd->wd_exact) (void) snprintf(buf, sizeof (buf), "%" PRIi64, bytes_rem[i]); else zfs_nicenum(bytes_rem[i], buf, sizeof (buf)); if (wd->wd_scripted) (void) printf(i == 0 ? "%s" : "\t%s", buf); else (void) printf(" %*s", col_widths[i] - 1, buf); } (void) printf("\n"); (void) fflush(stdout); } static void * wait_status_thread(void *arg) { wait_data_t *wd = (wait_data_t *)arg; zpool_handle_t *zhp; if ((zhp = zpool_open(g_zfs, wd->wd_poolname)) == NULL) return (void *)(1); for (int row = 0; ; row++) { boolean_t missing; struct timespec timeout; int ret = 0; (void) clock_gettime(CLOCK_REALTIME, &timeout); if (zpool_refresh_stats(zhp, &missing) != 0 || missing || zpool_props_refresh(zhp) != 0) { zpool_close(zhp); return (void *)(uintptr_t)(missing ? 0 : 1); } print_wait_status_row(wd, zhp, row); timeout.tv_sec += floor(wd->wd_interval); long nanos = timeout.tv_nsec + (wd->wd_interval - floor(wd->wd_interval)) * NANOSEC; if (nanos >= NANOSEC) { timeout.tv_sec++; timeout.tv_nsec = nanos - NANOSEC; } else { timeout.tv_nsec = nanos; } pthread_mutex_lock(&wd->wd_mutex); if (!wd->wd_should_exit) ret = pthread_cond_timedwait(&wd->wd_cv, &wd->wd_mutex, &timeout); pthread_mutex_unlock(&wd->wd_mutex); if (ret == 0) { break; /* signaled by main thread */ } else if (ret != ETIMEDOUT) { (void) fprintf(stderr, gettext("pthread_cond_timedwait " "failed: %s\n"), strerror(ret)); zpool_close(zhp); return (void *)(uintptr_t)(1); } } zpool_close(zhp); return (void *)(0); } int zpool_do_wait(int argc, char **argv) { boolean_t verbose = B_FALSE; int c, i; unsigned long count; pthread_t status_thr; int error = 0; zpool_handle_t *zhp; wait_data_t wd; wd.wd_scripted = B_FALSE; wd.wd_exact = B_FALSE; wd.wd_headers_once = B_FALSE; wd.wd_should_exit = B_FALSE; pthread_mutex_init(&wd.wd_mutex, NULL); pthread_cond_init(&wd.wd_cv, NULL); /* By default, wait for all types of activity. */ for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++) wd.wd_enabled[i] = B_TRUE; while ((c = getopt(argc, argv, "HpT:t:")) != -1) { switch (c) { case 'H': wd.wd_scripted = B_TRUE; break; case 'n': wd.wd_headers_once = B_TRUE; break; case 'p': wd.wd_exact = B_TRUE; break; case 'T': get_timestamp_arg(*optarg); break; case 't': /* Reset activities array */ memset(&wd.wd_enabled, 0, sizeof (wd.wd_enabled)); for (char *tok; (tok = strsep(&optarg, ",")); ) { static const char *const col_opts[] = { "discard", "free", "initialize", "replace", "remove", "resilver", "scrub", "trim" }; for (i = 0; i < ARRAY_SIZE(col_opts); ++i) if (strcmp(tok, col_opts[i]) == 0) { wd.wd_enabled[i] = B_TRUE; goto found; } (void) fprintf(stderr, gettext("invalid activity '%s'\n"), tok); usage(B_FALSE); found:; } break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); usage(B_FALSE); } } argc -= optind; argv += optind; get_interval_count(&argc, argv, &wd.wd_interval, &count); if (count != 0) { /* This subcmd only accepts an interval, not a count */ (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } if (wd.wd_interval != 0) verbose = B_TRUE; if (argc < 1) { (void) fprintf(stderr, gettext("missing 'pool' argument\n")); usage(B_FALSE); } if (argc > 1) { (void) fprintf(stderr, gettext("too many arguments\n")); usage(B_FALSE); } wd.wd_poolname = argv[0]; if ((zhp = zpool_open(g_zfs, wd.wd_poolname)) == NULL) return (1); if (verbose) { /* * We use a separate thread for printing status updates because * the main thread will call lzc_wait(), which blocks as long * as an activity is in progress, which can be a long time. */ if (pthread_create(&status_thr, NULL, wait_status_thread, &wd) != 0) { (void) fprintf(stderr, gettext("failed to create status" "thread: %s\n"), strerror(errno)); zpool_close(zhp); return (1); } } /* * Loop over all activities that we are supposed to wait for until none * of them are in progress. Note that this means we can end up waiting * for more activities to complete than just those that were in progress * when we began waiting; if an activity we are interested in begins * while we are waiting for another activity, we will wait for both to * complete before exiting. */ for (;;) { boolean_t missing = B_FALSE; boolean_t any_waited = B_FALSE; for (i = 0; i < ZPOOL_WAIT_NUM_ACTIVITIES; i++) { boolean_t waited; if (!wd.wd_enabled[i]) continue; error = zpool_wait_status(zhp, i, &missing, &waited); if (error != 0 || missing) break; any_waited = (any_waited || waited); } if (error != 0 || missing || !any_waited) break; } zpool_close(zhp); if (verbose) { uintptr_t status; pthread_mutex_lock(&wd.wd_mutex); wd.wd_should_exit = B_TRUE; pthread_cond_signal(&wd.wd_cv); pthread_mutex_unlock(&wd.wd_mutex); (void) pthread_join(status_thr, (void *)&status); if (status != 0) error = status; } pthread_mutex_destroy(&wd.wd_mutex); pthread_cond_destroy(&wd.wd_cv); return (error); } static int find_command_idx(const char *command, int *idx) { for (int i = 0; i < NCOMMAND; ++i) { if (command_table[i].name == NULL) continue; if (strcmp(command, command_table[i].name) == 0) { *idx = i; return (0); } } return (1); } /* * Display version message */ static int zpool_do_version(int argc, char **argv) { (void) argc, (void) argv; return (zfs_version_print() != 0); } /* * Do zpool_load_compat() and print error message on failure */ static zpool_compat_status_t zpool_do_load_compat(const char *compat, boolean_t *list) { char report[1024]; zpool_compat_status_t ret; ret = zpool_load_compat(compat, list, report, 1024); switch (ret) { case ZPOOL_COMPATIBILITY_OK: break; case ZPOOL_COMPATIBILITY_NOFILES: case ZPOOL_COMPATIBILITY_BADFILE: case ZPOOL_COMPATIBILITY_BADTOKEN: (void) fprintf(stderr, "Error: %s\n", report); break; case ZPOOL_COMPATIBILITY_WARNTOKEN: (void) fprintf(stderr, "Warning: %s\n", report); ret = ZPOOL_COMPATIBILITY_OK; break; } return (ret); } int main(int argc, char **argv) { int ret = 0; int i = 0; char *cmdname; char **newargv; (void) setlocale(LC_ALL, ""); (void) setlocale(LC_NUMERIC, "C"); (void) textdomain(TEXT_DOMAIN); srand(time(NULL)); opterr = 0; /* * Make sure the user has specified some command. */ if (argc < 2) { (void) fprintf(stderr, gettext("missing command\n")); usage(B_FALSE); } cmdname = argv[1]; /* * Special case '-?' */ if ((strcmp(cmdname, "-?") == 0) || strcmp(cmdname, "--help") == 0) usage(B_TRUE); /* * Special case '-V|--version' */ if ((strcmp(cmdname, "-V") == 0) || (strcmp(cmdname, "--version") == 0)) return (zpool_do_version(argc, argv)); if ((g_zfs = libzfs_init()) == NULL) { (void) fprintf(stderr, "%s\n", libzfs_error_init(errno)); return (1); } libzfs_print_on_error(g_zfs, B_TRUE); zfs_save_arguments(argc, argv, history_str, sizeof (history_str)); /* * Many commands modify input strings for string parsing reasons. * We create a copy to protect the original argv. */ newargv = safe_malloc((argc + 1) * sizeof (newargv[0])); for (i = 0; i < argc; i++) newargv[i] = strdup(argv[i]); newargv[argc] = NULL; /* * Run the appropriate command. */ if (find_command_idx(cmdname, &i) == 0) { current_command = &command_table[i]; ret = command_table[i].func(argc - 1, newargv + 1); } else if (strchr(cmdname, '=')) { verify(find_command_idx("set", &i) == 0); current_command = &command_table[i]; ret = command_table[i].func(argc, newargv); } else if (strcmp(cmdname, "freeze") == 0 && argc == 3) { /* * 'freeze' is a vile debugging abomination, so we treat * it as such. */ zfs_cmd_t zc = {"\0"}; (void) strlcpy(zc.zc_name, argv[2], sizeof (zc.zc_name)); ret = zfs_ioctl(g_zfs, ZFS_IOC_POOL_FREEZE, &zc); if (ret != 0) { (void) fprintf(stderr, gettext("failed to freeze pool: %d\n"), errno); ret = 1; } log_history = 0; } else { (void) fprintf(stderr, gettext("unrecognized " "command '%s'\n"), cmdname); usage(B_FALSE); ret = 1; } for (i = 0; i < argc; i++) free(newargv[i]); free(newargv); if (ret == 0 && log_history) (void) zpool_log_history(g_zfs, history_str); libzfs_fini(g_zfs); /* * The 'ZFS_ABORT' environment variable causes us to dump core on exit * for the purposes of running ::findleaks. */ if (getenv("ZFS_ABORT") != NULL) { (void) printf("dumping core by request\n"); abort(); } return (ret); } diff --git a/sys/contrib/openzfs/include/libzfs.h b/sys/contrib/openzfs/include/libzfs.h index 87d1ed738f2b..a7037e3e6266 100644 --- a/sys/contrib/openzfs/include/libzfs.h +++ b/sys/contrib/openzfs/include/libzfs.h @@ -1,1028 +1,1031 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2011, 2022 by Delphix. All rights reserved. * Copyright Joyent, Inc. * Copyright (c) 2013 Steven Hartland. All rights reserved. * Copyright (c) 2016, Intel Corporation. * Copyright 2016 Nexenta Systems, Inc. * Copyright (c) 2017 Open-E, Inc. All Rights Reserved. * Copyright (c) 2019 Datto Inc. * Copyright (c) 2021, Colm Buckley */ #ifndef _LIBZFS_H #define _LIBZFS_H extern __attribute__((visibility("default"))) #include #include #include #include #include #include #include #include #include #ifdef __cplusplus extern "C" { #endif /* * Miscellaneous ZFS constants */ #define ZFS_MAXPROPLEN MAXPATHLEN #define ZPOOL_MAXPROPLEN MAXPATHLEN /* * libzfs errors */ typedef enum zfs_error { EZFS_SUCCESS = 0, /* no error -- success */ EZFS_NOMEM = 2000, /* out of memory */ EZFS_BADPROP, /* invalid property value */ EZFS_PROPREADONLY, /* cannot set readonly property */ EZFS_PROPTYPE, /* property does not apply to dataset type */ EZFS_PROPNONINHERIT, /* property is not inheritable */ EZFS_PROPSPACE, /* bad quota or reservation */ EZFS_BADTYPE, /* dataset is not of appropriate type */ EZFS_BUSY, /* pool or dataset is busy */ EZFS_EXISTS, /* pool or dataset already exists */ EZFS_NOENT, /* no such pool or dataset */ EZFS_BADSTREAM, /* bad backup stream */ EZFS_DSREADONLY, /* dataset is readonly */ EZFS_VOLTOOBIG, /* volume is too large for 32-bit system */ EZFS_INVALIDNAME, /* invalid dataset name */ EZFS_BADRESTORE, /* unable to restore to destination */ EZFS_BADBACKUP, /* backup failed */ EZFS_BADTARGET, /* bad attach/detach/replace target */ EZFS_NODEVICE, /* no such device in pool */ EZFS_BADDEV, /* invalid device to add */ EZFS_NOREPLICAS, /* no valid replicas */ EZFS_RESILVERING, /* resilvering (healing reconstruction) */ EZFS_BADVERSION, /* unsupported version */ EZFS_POOLUNAVAIL, /* pool is currently unavailable */ EZFS_DEVOVERFLOW, /* too many devices in one vdev */ EZFS_BADPATH, /* must be an absolute path */ EZFS_CROSSTARGET, /* rename or clone across pool or dataset */ EZFS_ZONED, /* used improperly in local zone */ EZFS_MOUNTFAILED, /* failed to mount dataset */ EZFS_UMOUNTFAILED, /* failed to unmount dataset */ EZFS_UNSHARENFSFAILED, /* failed to unshare over nfs */ EZFS_SHARENFSFAILED, /* failed to share over nfs */ EZFS_PERM, /* permission denied */ EZFS_NOSPC, /* out of space */ EZFS_FAULT, /* bad address */ EZFS_IO, /* I/O error */ EZFS_INTR, /* signal received */ EZFS_ISSPARE, /* device is a hot spare */ EZFS_INVALCONFIG, /* invalid vdev configuration */ EZFS_RECURSIVE, /* recursive dependency */ EZFS_NOHISTORY, /* no history object */ EZFS_POOLPROPS, /* couldn't retrieve pool props */ EZFS_POOL_NOTSUP, /* ops not supported for this type of pool */ EZFS_POOL_INVALARG, /* invalid argument for this pool operation */ EZFS_NAMETOOLONG, /* dataset name is too long */ EZFS_OPENFAILED, /* open of device failed */ EZFS_NOCAP, /* couldn't get capacity */ EZFS_LABELFAILED, /* write of label failed */ EZFS_BADWHO, /* invalid permission who */ EZFS_BADPERM, /* invalid permission */ EZFS_BADPERMSET, /* invalid permission set name */ EZFS_NODELEGATION, /* delegated administration is disabled */ EZFS_UNSHARESMBFAILED, /* failed to unshare over smb */ EZFS_SHARESMBFAILED, /* failed to share over smb */ EZFS_BADCACHE, /* bad cache file */ EZFS_ISL2CACHE, /* device is for the level 2 ARC */ EZFS_VDEVNOTSUP, /* unsupported vdev type */ EZFS_NOTSUP, /* ops not supported on this dataset */ EZFS_ACTIVE_SPARE, /* pool has active shared spare devices */ EZFS_UNPLAYED_LOGS, /* log device has unplayed logs */ EZFS_REFTAG_RELE, /* snapshot release: tag not found */ EZFS_REFTAG_HOLD, /* snapshot hold: tag already exists */ EZFS_TAGTOOLONG, /* snapshot hold/rele: tag too long */ EZFS_PIPEFAILED, /* pipe create failed */ EZFS_THREADCREATEFAILED, /* thread create failed */ EZFS_POSTSPLIT_ONLINE, /* onlining a disk after splitting it */ EZFS_SCRUBBING, /* currently scrubbing */ + EZFS_ERRORSCRUBBING, /* currently error scrubbing */ + EZFS_ERRORSCRUB_PAUSED, /* error scrub currently paused */ EZFS_NO_SCRUB, /* no active scrub */ EZFS_DIFF, /* general failure of zfs diff */ EZFS_DIFFDATA, /* bad zfs diff data */ EZFS_POOLREADONLY, /* pool is in read-only mode */ EZFS_SCRUB_PAUSED, /* scrub currently paused */ + EZFS_SCRUB_PAUSED_TO_CANCEL, /* scrub currently paused */ EZFS_ACTIVE_POOL, /* pool is imported on a different system */ EZFS_CRYPTOFAILED, /* failed to setup encryption */ EZFS_NO_PENDING, /* cannot cancel, no operation is pending */ EZFS_CHECKPOINT_EXISTS, /* checkpoint exists */ EZFS_DISCARDING_CHECKPOINT, /* currently discarding a checkpoint */ EZFS_NO_CHECKPOINT, /* pool has no checkpoint */ EZFS_DEVRM_IN_PROGRESS, /* a device is currently being removed */ EZFS_VDEV_TOO_BIG, /* a device is too big to be used */ EZFS_IOC_NOTSUPPORTED, /* operation not supported by zfs module */ EZFS_TOOMANY, /* argument list too long */ EZFS_INITIALIZING, /* currently initializing */ EZFS_NO_INITIALIZE, /* no active initialize */ EZFS_WRONG_PARENT, /* invalid parent dataset (e.g ZVOL) */ EZFS_TRIMMING, /* currently trimming */ EZFS_NO_TRIM, /* no active trim */ EZFS_TRIM_NOTSUP, /* device does not support trim */ EZFS_NO_RESILVER_DEFER, /* pool doesn't support resilver_defer */ EZFS_EXPORT_IN_PROGRESS, /* currently exporting the pool */ EZFS_REBUILDING, /* resilvering (sequential reconstrution) */ EZFS_VDEV_NOTSUP, /* ops not supported for this type of vdev */ EZFS_NOT_USER_NAMESPACE, /* a file is not a user namespace */ EZFS_CKSUM, /* insufficient replicas */ EZFS_RESUME_EXISTS, /* Resume on existing dataset without force */ EZFS_UNKNOWN } zfs_error_t; /* * The following data structures are all part * of the zfs_allow_t data structure which is * used for printing 'allow' permissions. * It is a linked list of zfs_allow_t's which * then contain avl tree's for user/group/sets/... * and each one of the entries in those trees have * avl tree's for the permissions they belong to and * whether they are local,descendent or local+descendent * permissions. The AVL trees are used primarily for * sorting purposes, but also so that we can quickly find * a given user and or permission. */ typedef struct zfs_perm_node { avl_node_t z_node; char z_pname[MAXPATHLEN]; } zfs_perm_node_t; typedef struct zfs_allow_node { avl_node_t z_node; char z_key[MAXPATHLEN]; /* name, such as joe */ avl_tree_t z_localdescend; /* local+descendent perms */ avl_tree_t z_local; /* local permissions */ avl_tree_t z_descend; /* descendent permissions */ } zfs_allow_node_t; typedef struct zfs_allow { struct zfs_allow *z_next; char z_setpoint[MAXPATHLEN]; avl_tree_t z_sets; avl_tree_t z_crperms; avl_tree_t z_user; avl_tree_t z_group; avl_tree_t z_everyone; } zfs_allow_t; /* * Basic handle types */ typedef struct zfs_handle zfs_handle_t; typedef struct zpool_handle zpool_handle_t; typedef struct libzfs_handle libzfs_handle_t; _LIBZFS_H int zpool_wait(zpool_handle_t *, zpool_wait_activity_t); _LIBZFS_H int zpool_wait_status(zpool_handle_t *, zpool_wait_activity_t, boolean_t *, boolean_t *); /* * Library initialization */ _LIBZFS_H libzfs_handle_t *libzfs_init(void); _LIBZFS_H void libzfs_fini(libzfs_handle_t *); _LIBZFS_H libzfs_handle_t *zpool_get_handle(zpool_handle_t *); _LIBZFS_H libzfs_handle_t *zfs_get_handle(zfs_handle_t *); _LIBZFS_H void libzfs_print_on_error(libzfs_handle_t *, boolean_t); _LIBZFS_H void zfs_save_arguments(int argc, char **, char *, int); _LIBZFS_H int zpool_log_history(libzfs_handle_t *, const char *); _LIBZFS_H int libzfs_errno(libzfs_handle_t *); _LIBZFS_H const char *libzfs_error_init(int); _LIBZFS_H const char *libzfs_error_action(libzfs_handle_t *); _LIBZFS_H const char *libzfs_error_description(libzfs_handle_t *); _LIBZFS_H int zfs_standard_error(libzfs_handle_t *, int, const char *); _LIBZFS_H void libzfs_mnttab_init(libzfs_handle_t *); _LIBZFS_H void libzfs_mnttab_fini(libzfs_handle_t *); _LIBZFS_H void libzfs_mnttab_cache(libzfs_handle_t *, boolean_t); _LIBZFS_H int libzfs_mnttab_find(libzfs_handle_t *, const char *, struct mnttab *); _LIBZFS_H void libzfs_mnttab_add(libzfs_handle_t *, const char *, const char *, const char *); _LIBZFS_H void libzfs_mnttab_remove(libzfs_handle_t *, const char *); /* * Basic handle functions */ _LIBZFS_H zpool_handle_t *zpool_open(libzfs_handle_t *, const char *); _LIBZFS_H zpool_handle_t *zpool_open_canfail(libzfs_handle_t *, const char *); _LIBZFS_H void zpool_close(zpool_handle_t *); _LIBZFS_H const char *zpool_get_name(zpool_handle_t *); _LIBZFS_H int zpool_get_state(zpool_handle_t *); _LIBZFS_H const char *zpool_state_to_name(vdev_state_t, vdev_aux_t); _LIBZFS_H const char *zpool_pool_state_to_name(pool_state_t); _LIBZFS_H void zpool_free_handles(libzfs_handle_t *); /* * Iterate over all active pools in the system. */ typedef int (*zpool_iter_f)(zpool_handle_t *, void *); _LIBZFS_H int zpool_iter(libzfs_handle_t *, zpool_iter_f, void *); _LIBZFS_H boolean_t zpool_skip_pool(const char *); /* * Functions to create and destroy pools */ _LIBZFS_H int zpool_create(libzfs_handle_t *, const char *, nvlist_t *, nvlist_t *, nvlist_t *); _LIBZFS_H int zpool_destroy(zpool_handle_t *, const char *); _LIBZFS_H int zpool_add(zpool_handle_t *, nvlist_t *); typedef struct splitflags { /* do not split, but return the config that would be split off */ unsigned int dryrun : 1; /* after splitting, import the pool */ unsigned int import : 1; int name_flags; } splitflags_t; typedef struct trimflags { /* requested vdevs are for the entire pool */ boolean_t fullpool; /* request a secure trim, requires support from device */ boolean_t secure; /* after starting trim, block until trim completes */ boolean_t wait; /* trim at the requested rate in bytes/second */ uint64_t rate; } trimflags_t; /* * Functions to manipulate pool and vdev state */ _LIBZFS_H int zpool_scan(zpool_handle_t *, pool_scan_func_t, pool_scrub_cmd_t); _LIBZFS_H int zpool_initialize(zpool_handle_t *, pool_initialize_func_t, nvlist_t *); _LIBZFS_H int zpool_initialize_wait(zpool_handle_t *, pool_initialize_func_t, nvlist_t *); _LIBZFS_H int zpool_trim(zpool_handle_t *, pool_trim_func_t, nvlist_t *, trimflags_t *); _LIBZFS_H int zpool_clear(zpool_handle_t *, const char *, nvlist_t *); _LIBZFS_H int zpool_reguid(zpool_handle_t *); _LIBZFS_H int zpool_reopen_one(zpool_handle_t *, void *); _LIBZFS_H int zpool_sync_one(zpool_handle_t *, void *); _LIBZFS_H int zpool_vdev_online(zpool_handle_t *, const char *, int, vdev_state_t *); _LIBZFS_H int zpool_vdev_offline(zpool_handle_t *, const char *, boolean_t); _LIBZFS_H int zpool_vdev_attach(zpool_handle_t *, const char *, const char *, nvlist_t *, int, boolean_t); _LIBZFS_H int zpool_vdev_detach(zpool_handle_t *, const char *); _LIBZFS_H int zpool_vdev_remove(zpool_handle_t *, const char *); _LIBZFS_H int zpool_vdev_remove_cancel(zpool_handle_t *); _LIBZFS_H int zpool_vdev_indirect_size(zpool_handle_t *, const char *, uint64_t *); _LIBZFS_H int zpool_vdev_split(zpool_handle_t *, char *, nvlist_t **, nvlist_t *, splitflags_t); _LIBZFS_H int zpool_vdev_remove_wanted(zpool_handle_t *, const char *); _LIBZFS_H int zpool_vdev_fault(zpool_handle_t *, uint64_t, vdev_aux_t); _LIBZFS_H int zpool_vdev_degrade(zpool_handle_t *, uint64_t, vdev_aux_t); _LIBZFS_H int zpool_vdev_clear(zpool_handle_t *, uint64_t); _LIBZFS_H nvlist_t *zpool_find_vdev(zpool_handle_t *, const char *, boolean_t *, boolean_t *, boolean_t *); _LIBZFS_H nvlist_t *zpool_find_vdev_by_physpath(zpool_handle_t *, const char *, boolean_t *, boolean_t *, boolean_t *); _LIBZFS_H int zpool_label_disk(libzfs_handle_t *, zpool_handle_t *, const char *); _LIBZFS_H uint64_t zpool_vdev_path_to_guid(zpool_handle_t *zhp, const char *path); _LIBZFS_H const char *zpool_get_state_str(zpool_handle_t *); /* * Functions to manage pool properties */ _LIBZFS_H int zpool_set_prop(zpool_handle_t *, const char *, const char *); _LIBZFS_H int zpool_get_prop(zpool_handle_t *, zpool_prop_t, char *, size_t proplen, zprop_source_t *, boolean_t literal); _LIBZFS_H int zpool_get_userprop(zpool_handle_t *, const char *, char *, size_t proplen, zprop_source_t *); _LIBZFS_H uint64_t zpool_get_prop_int(zpool_handle_t *, zpool_prop_t, zprop_source_t *); _LIBZFS_H int zpool_props_refresh(zpool_handle_t *); _LIBZFS_H const char *zpool_prop_to_name(zpool_prop_t); _LIBZFS_H const char *zpool_prop_values(zpool_prop_t); /* * Functions to manage vdev properties */ _LIBZFS_H int zpool_get_vdev_prop_value(nvlist_t *, vdev_prop_t, char *, char *, size_t, zprop_source_t *, boolean_t); _LIBZFS_H int zpool_get_vdev_prop(zpool_handle_t *, const char *, vdev_prop_t, char *, char *, size_t, zprop_source_t *, boolean_t); _LIBZFS_H int zpool_get_all_vdev_props(zpool_handle_t *, const char *, nvlist_t **); _LIBZFS_H int zpool_set_vdev_prop(zpool_handle_t *, const char *, const char *, const char *); _LIBZFS_H const char *vdev_prop_to_name(vdev_prop_t); _LIBZFS_H const char *vdev_prop_values(vdev_prop_t); _LIBZFS_H boolean_t vdev_prop_user(const char *name); _LIBZFS_H const char *vdev_prop_column_name(vdev_prop_t); _LIBZFS_H boolean_t vdev_prop_align_right(vdev_prop_t); /* * Pool health statistics. */ typedef enum { /* * The following correspond to faults as defined in the (fault.fs.zfs.*) * event namespace. Each is associated with a corresponding message ID. * This must be kept in sync with the zfs_msgid_table in * lib/libzfs/libzfs_status.c. */ ZPOOL_STATUS_CORRUPT_CACHE, /* corrupt /kernel/drv/zpool.cache */ ZPOOL_STATUS_MISSING_DEV_R, /* missing device with replicas */ ZPOOL_STATUS_MISSING_DEV_NR, /* missing device with no replicas */ ZPOOL_STATUS_CORRUPT_LABEL_R, /* bad device label with replicas */ ZPOOL_STATUS_CORRUPT_LABEL_NR, /* bad device label with no replicas */ ZPOOL_STATUS_BAD_GUID_SUM, /* sum of device guids didn't match */ ZPOOL_STATUS_CORRUPT_POOL, /* pool metadata is corrupted */ ZPOOL_STATUS_CORRUPT_DATA, /* data errors in user (meta)data */ ZPOOL_STATUS_FAILING_DEV, /* device experiencing errors */ ZPOOL_STATUS_VERSION_NEWER, /* newer on-disk version */ ZPOOL_STATUS_HOSTID_MISMATCH, /* last accessed by another system */ ZPOOL_STATUS_HOSTID_ACTIVE, /* currently active on another system */ ZPOOL_STATUS_HOSTID_REQUIRED, /* multihost=on and hostid=0 */ ZPOOL_STATUS_IO_FAILURE_WAIT, /* failed I/O, failmode 'wait' */ ZPOOL_STATUS_IO_FAILURE_CONTINUE, /* failed I/O, failmode 'continue' */ ZPOOL_STATUS_IO_FAILURE_MMP, /* failed MMP, failmode not 'panic' */ ZPOOL_STATUS_BAD_LOG, /* cannot read log chain(s) */ ZPOOL_STATUS_ERRATA, /* informational errata available */ /* * If the pool has unsupported features but can still be opened in * read-only mode, its status is ZPOOL_STATUS_UNSUP_FEAT_WRITE. If the * pool has unsupported features but cannot be opened at all, its * status is ZPOOL_STATUS_UNSUP_FEAT_READ. */ ZPOOL_STATUS_UNSUP_FEAT_READ, /* unsupported features for read */ ZPOOL_STATUS_UNSUP_FEAT_WRITE, /* unsupported features for write */ /* * These faults have no corresponding message ID. At the time we are * checking the status, the original reason for the FMA fault (I/O or * checksum errors) has been lost. */ ZPOOL_STATUS_FAULTED_DEV_R, /* faulted device with replicas */ ZPOOL_STATUS_FAULTED_DEV_NR, /* faulted device with no replicas */ /* * The following are not faults per se, but still an error possibly * requiring administrative attention. There is no corresponding * message ID. */ ZPOOL_STATUS_VERSION_OLDER, /* older legacy on-disk version */ ZPOOL_STATUS_FEAT_DISABLED, /* supported features are disabled */ ZPOOL_STATUS_RESILVERING, /* device being resilvered */ ZPOOL_STATUS_OFFLINE_DEV, /* device offline */ ZPOOL_STATUS_REMOVED_DEV, /* removed device */ ZPOOL_STATUS_REBUILDING, /* device being rebuilt */ ZPOOL_STATUS_REBUILD_SCRUB, /* recommend scrubbing the pool */ ZPOOL_STATUS_NON_NATIVE_ASHIFT, /* (e.g. 512e dev with ashift of 9) */ ZPOOL_STATUS_COMPATIBILITY_ERR, /* bad 'compatibility' property */ ZPOOL_STATUS_INCOMPATIBLE_FEAT, /* feature set outside compatibility */ /* * Finally, the following indicates a healthy pool. */ ZPOOL_STATUS_OK } zpool_status_t; _LIBZFS_H zpool_status_t zpool_get_status(zpool_handle_t *, const char **, zpool_errata_t *); _LIBZFS_H zpool_status_t zpool_import_status(nvlist_t *, const char **, zpool_errata_t *); /* * Statistics and configuration functions. */ _LIBZFS_H nvlist_t *zpool_get_config(zpool_handle_t *, nvlist_t **); _LIBZFS_H nvlist_t *zpool_get_features(zpool_handle_t *); _LIBZFS_H int zpool_refresh_stats(zpool_handle_t *, boolean_t *); _LIBZFS_H int zpool_get_errlog(zpool_handle_t *, nvlist_t **); /* * Import and export functions */ _LIBZFS_H int zpool_export(zpool_handle_t *, boolean_t, const char *); _LIBZFS_H int zpool_export_force(zpool_handle_t *, const char *); _LIBZFS_H int zpool_import(libzfs_handle_t *, nvlist_t *, const char *, char *altroot); _LIBZFS_H int zpool_import_props(libzfs_handle_t *, nvlist_t *, const char *, nvlist_t *, int); _LIBZFS_H void zpool_print_unsup_feat(nvlist_t *config); /* * Miscellaneous pool functions */ struct zfs_cmd; _LIBZFS_H const char *const zfs_history_event_names[]; typedef enum { VDEV_NAME_PATH = 1 << 0, VDEV_NAME_GUID = 1 << 1, VDEV_NAME_FOLLOW_LINKS = 1 << 2, VDEV_NAME_TYPE_ID = 1 << 3, } vdev_name_t; _LIBZFS_H char *zpool_vdev_name(libzfs_handle_t *, zpool_handle_t *, nvlist_t *, int name_flags); _LIBZFS_H int zpool_upgrade(zpool_handle_t *, uint64_t); _LIBZFS_H int zpool_get_history(zpool_handle_t *, nvlist_t **, uint64_t *, boolean_t *); _LIBZFS_H int zpool_events_next(libzfs_handle_t *, nvlist_t **, int *, unsigned, int); _LIBZFS_H int zpool_events_clear(libzfs_handle_t *, int *); _LIBZFS_H int zpool_events_seek(libzfs_handle_t *, uint64_t, int); _LIBZFS_H void zpool_obj_to_path_ds(zpool_handle_t *, uint64_t, uint64_t, char *, size_t); _LIBZFS_H void zpool_obj_to_path(zpool_handle_t *, uint64_t, uint64_t, char *, size_t); _LIBZFS_H int zfs_ioctl(libzfs_handle_t *, int, struct zfs_cmd *); _LIBZFS_H void zpool_explain_recover(libzfs_handle_t *, const char *, int, nvlist_t *); _LIBZFS_H int zpool_checkpoint(zpool_handle_t *); _LIBZFS_H int zpool_discard_checkpoint(zpool_handle_t *); _LIBZFS_H boolean_t zpool_is_draid_spare(const char *); /* * Basic handle manipulations. These functions do not create or destroy the * underlying datasets, only the references to them. */ _LIBZFS_H zfs_handle_t *zfs_open(libzfs_handle_t *, const char *, int); _LIBZFS_H zfs_handle_t *zfs_handle_dup(zfs_handle_t *); _LIBZFS_H void zfs_close(zfs_handle_t *); _LIBZFS_H zfs_type_t zfs_get_type(const zfs_handle_t *); _LIBZFS_H zfs_type_t zfs_get_underlying_type(const zfs_handle_t *); _LIBZFS_H const char *zfs_get_name(const zfs_handle_t *); _LIBZFS_H zpool_handle_t *zfs_get_pool_handle(const zfs_handle_t *); _LIBZFS_H const char *zfs_get_pool_name(const zfs_handle_t *); /* * Property management functions. Some functions are shared with the kernel, * and are found in sys/fs/zfs.h. */ /* * zfs dataset property management */ _LIBZFS_H const char *zfs_prop_default_string(zfs_prop_t); _LIBZFS_H uint64_t zfs_prop_default_numeric(zfs_prop_t); _LIBZFS_H const char *zfs_prop_column_name(zfs_prop_t); _LIBZFS_H boolean_t zfs_prop_align_right(zfs_prop_t); _LIBZFS_H nvlist_t *zfs_valid_proplist(libzfs_handle_t *, zfs_type_t, nvlist_t *, uint64_t, zfs_handle_t *, zpool_handle_t *, boolean_t, const char *); _LIBZFS_H const char *zfs_prop_to_name(zfs_prop_t); _LIBZFS_H int zfs_prop_set(zfs_handle_t *, const char *, const char *); _LIBZFS_H int zfs_prop_set_list(zfs_handle_t *, nvlist_t *); _LIBZFS_H int zfs_prop_get(zfs_handle_t *, zfs_prop_t, char *, size_t, zprop_source_t *, char *, size_t, boolean_t); _LIBZFS_H int zfs_prop_get_recvd(zfs_handle_t *, const char *, char *, size_t, boolean_t); _LIBZFS_H int zfs_prop_get_numeric(zfs_handle_t *, zfs_prop_t, uint64_t *, zprop_source_t *, char *, size_t); _LIBZFS_H int zfs_prop_get_userquota_int(zfs_handle_t *zhp, const char *propname, uint64_t *propvalue); _LIBZFS_H int zfs_prop_get_userquota(zfs_handle_t *zhp, const char *propname, char *propbuf, int proplen, boolean_t literal); _LIBZFS_H int zfs_prop_get_written_int(zfs_handle_t *zhp, const char *propname, uint64_t *propvalue); _LIBZFS_H int zfs_prop_get_written(zfs_handle_t *zhp, const char *propname, char *propbuf, int proplen, boolean_t literal); _LIBZFS_H int zfs_prop_get_feature(zfs_handle_t *zhp, const char *propname, char *buf, size_t len); _LIBZFS_H uint64_t getprop_uint64(zfs_handle_t *, zfs_prop_t, const char **); _LIBZFS_H uint64_t zfs_prop_get_int(zfs_handle_t *, zfs_prop_t); _LIBZFS_H int zfs_prop_inherit(zfs_handle_t *, const char *, boolean_t); _LIBZFS_H const char *zfs_prop_values(zfs_prop_t); _LIBZFS_H int zfs_prop_is_string(zfs_prop_t prop); _LIBZFS_H nvlist_t *zfs_get_all_props(zfs_handle_t *); _LIBZFS_H nvlist_t *zfs_get_user_props(zfs_handle_t *); _LIBZFS_H nvlist_t *zfs_get_recvd_props(zfs_handle_t *); _LIBZFS_H nvlist_t *zfs_get_clones_nvl(zfs_handle_t *); _LIBZFS_H int zfs_wait_status(zfs_handle_t *, zfs_wait_activity_t, boolean_t *, boolean_t *); /* * zfs encryption management */ _LIBZFS_H int zfs_crypto_get_encryption_root(zfs_handle_t *, boolean_t *, char *); _LIBZFS_H int zfs_crypto_create(libzfs_handle_t *, char *, nvlist_t *, nvlist_t *, boolean_t stdin_available, uint8_t **, uint_t *); _LIBZFS_H int zfs_crypto_clone_check(libzfs_handle_t *, zfs_handle_t *, char *, nvlist_t *); _LIBZFS_H int zfs_crypto_attempt_load_keys(libzfs_handle_t *, const char *); _LIBZFS_H int zfs_crypto_load_key(zfs_handle_t *, boolean_t, const char *); _LIBZFS_H int zfs_crypto_unload_key(zfs_handle_t *); _LIBZFS_H int zfs_crypto_rewrap(zfs_handle_t *, nvlist_t *, boolean_t); typedef struct zprop_list { int pl_prop; char *pl_user_prop; struct zprop_list *pl_next; boolean_t pl_all; size_t pl_width; size_t pl_recvd_width; boolean_t pl_fixed; } zprop_list_t; _LIBZFS_H int zfs_expand_proplist(zfs_handle_t *, zprop_list_t **, boolean_t, boolean_t); _LIBZFS_H void zfs_prune_proplist(zfs_handle_t *, uint8_t *); _LIBZFS_H int vdev_expand_proplist(zpool_handle_t *, const char *, zprop_list_t **); #define ZFS_MOUNTPOINT_NONE "none" #define ZFS_MOUNTPOINT_LEGACY "legacy" #define ZFS_FEATURE_DISABLED "disabled" #define ZFS_FEATURE_ENABLED "enabled" #define ZFS_FEATURE_ACTIVE "active" #define ZFS_UNSUPPORTED_INACTIVE "inactive" #define ZFS_UNSUPPORTED_READONLY "readonly" /* * zpool property management */ _LIBZFS_H int zpool_expand_proplist(zpool_handle_t *, zprop_list_t **, zfs_type_t, boolean_t); _LIBZFS_H int zpool_prop_get_feature(zpool_handle_t *, const char *, char *, size_t); _LIBZFS_H const char *zpool_prop_default_string(zpool_prop_t); _LIBZFS_H uint64_t zpool_prop_default_numeric(zpool_prop_t); _LIBZFS_H const char *zpool_prop_column_name(zpool_prop_t); _LIBZFS_H boolean_t zpool_prop_align_right(zpool_prop_t); /* * Functions shared by zfs and zpool property management. */ _LIBZFS_H int zprop_iter(zprop_func func, void *cb, boolean_t show_all, boolean_t ordered, zfs_type_t type); _LIBZFS_H int zprop_get_list(libzfs_handle_t *, char *, zprop_list_t **, zfs_type_t); _LIBZFS_H void zprop_free_list(zprop_list_t *); #define ZFS_GET_NCOLS 5 typedef enum { GET_COL_NONE, GET_COL_NAME, GET_COL_PROPERTY, GET_COL_VALUE, GET_COL_RECVD, GET_COL_SOURCE } zfs_get_column_t; /* * Functions for printing zfs or zpool properties */ typedef struct vdev_cbdata { int cb_name_flags; char **cb_names; unsigned int cb_names_count; } vdev_cbdata_t; typedef struct zprop_get_cbdata { int cb_sources; zfs_get_column_t cb_columns[ZFS_GET_NCOLS]; int cb_colwidths[ZFS_GET_NCOLS + 1]; boolean_t cb_scripted; boolean_t cb_literal; boolean_t cb_first; zprop_list_t *cb_proplist; zfs_type_t cb_type; vdev_cbdata_t cb_vdevs; } zprop_get_cbdata_t; _LIBZFS_H void zprop_print_one_property(const char *, zprop_get_cbdata_t *, const char *, const char *, zprop_source_t, const char *, const char *); /* * Iterator functions. */ #define ZFS_ITER_RECURSE (1 << 0) #define ZFS_ITER_ARGS_CAN_BE_PATHS (1 << 1) #define ZFS_ITER_PROP_LISTSNAPS (1 << 2) #define ZFS_ITER_DEPTH_LIMIT (1 << 3) #define ZFS_ITER_RECVD_PROPS (1 << 4) #define ZFS_ITER_LITERAL_PROPS (1 << 5) #define ZFS_ITER_SIMPLE (1 << 6) typedef int (*zfs_iter_f)(zfs_handle_t *, void *); _LIBZFS_H int zfs_iter_root(libzfs_handle_t *, zfs_iter_f, void *); _LIBZFS_H int zfs_iter_children(zfs_handle_t *, zfs_iter_f, void *); _LIBZFS_H int zfs_iter_dependents(zfs_handle_t *, boolean_t, zfs_iter_f, void *); _LIBZFS_H int zfs_iter_filesystems(zfs_handle_t *, zfs_iter_f, void *); _LIBZFS_H int zfs_iter_snapshots(zfs_handle_t *, boolean_t, zfs_iter_f, void *, uint64_t, uint64_t); _LIBZFS_H int zfs_iter_snapshots_sorted(zfs_handle_t *, zfs_iter_f, void *, uint64_t, uint64_t); _LIBZFS_H int zfs_iter_snapspec(zfs_handle_t *, const char *, zfs_iter_f, void *); _LIBZFS_H int zfs_iter_bookmarks(zfs_handle_t *, zfs_iter_f, void *); _LIBZFS_H int zfs_iter_children_v2(zfs_handle_t *, int, zfs_iter_f, void *); _LIBZFS_H int zfs_iter_dependents_v2(zfs_handle_t *, int, boolean_t, zfs_iter_f, void *); _LIBZFS_H int zfs_iter_filesystems_v2(zfs_handle_t *, int, zfs_iter_f, void *); _LIBZFS_H int zfs_iter_snapshots_v2(zfs_handle_t *, int, zfs_iter_f, void *, uint64_t, uint64_t); _LIBZFS_H int zfs_iter_snapshots_sorted_v2(zfs_handle_t *, int, zfs_iter_f, void *, uint64_t, uint64_t); _LIBZFS_H int zfs_iter_snapspec_v2(zfs_handle_t *, int, const char *, zfs_iter_f, void *); _LIBZFS_H int zfs_iter_bookmarks_v2(zfs_handle_t *, int, zfs_iter_f, void *); _LIBZFS_H int zfs_iter_mounted(zfs_handle_t *, zfs_iter_f, void *); typedef struct get_all_cb { zfs_handle_t **cb_handles; size_t cb_alloc; size_t cb_used; } get_all_cb_t; _LIBZFS_H void zfs_foreach_mountpoint(libzfs_handle_t *, zfs_handle_t **, size_t, zfs_iter_f, void *, boolean_t); _LIBZFS_H void libzfs_add_handle(get_all_cb_t *, zfs_handle_t *); /* * Functions to create and destroy datasets. */ _LIBZFS_H int zfs_create(libzfs_handle_t *, const char *, zfs_type_t, nvlist_t *); _LIBZFS_H int zfs_create_ancestors(libzfs_handle_t *, const char *); _LIBZFS_H int zfs_destroy(zfs_handle_t *, boolean_t); _LIBZFS_H int zfs_destroy_snaps(zfs_handle_t *, char *, boolean_t); _LIBZFS_H int zfs_destroy_snaps_nvl(libzfs_handle_t *, nvlist_t *, boolean_t); _LIBZFS_H int zfs_destroy_snaps_nvl_os(libzfs_handle_t *, nvlist_t *); _LIBZFS_H int zfs_clone(zfs_handle_t *, const char *, nvlist_t *); _LIBZFS_H int zfs_snapshot(libzfs_handle_t *, const char *, boolean_t, nvlist_t *); _LIBZFS_H int zfs_snapshot_nvl(libzfs_handle_t *hdl, nvlist_t *snaps, nvlist_t *props); _LIBZFS_H int zfs_rollback(zfs_handle_t *, zfs_handle_t *, boolean_t); typedef struct renameflags { /* recursive rename */ unsigned int recursive : 1; /* don't unmount file systems */ unsigned int nounmount : 1; /* force unmount file systems */ unsigned int forceunmount : 1; } renameflags_t; _LIBZFS_H int zfs_rename(zfs_handle_t *, const char *, renameflags_t); typedef struct sendflags { /* Amount of extra information to print. */ int verbosity; /* recursive send (ie, -R) */ boolean_t replicate; /* for recursive send, skip sending missing snapshots */ boolean_t skipmissing; /* for incrementals, do all intermediate snapshots */ boolean_t doall; /* if dataset is a clone, do incremental from its origin */ boolean_t fromorigin; /* field no longer used, maintained for backwards compatibility */ boolean_t pad; /* send properties (ie, -p) */ boolean_t props; /* do not send (no-op, ie. -n) */ boolean_t dryrun; /* parsable verbose output (ie. -P) */ boolean_t parsable; /* show progress (ie. -v) */ boolean_t progress; /* show progress as process title (ie. -V) */ boolean_t progressastitle; /* large blocks (>128K) are permitted */ boolean_t largeblock; /* WRITE_EMBEDDED records of type DATA are permitted */ boolean_t embed_data; /* compressed WRITE records are permitted */ boolean_t compress; /* raw encrypted records are permitted */ boolean_t raw; /* only send received properties (ie. -b) */ boolean_t backup; /* include snapshot holds in send stream */ boolean_t holds; /* stream represents a partially received dataset */ boolean_t saved; } sendflags_t; typedef boolean_t (snapfilter_cb_t)(zfs_handle_t *, void *); _LIBZFS_H int zfs_send(zfs_handle_t *, const char *, const char *, sendflags_t *, int, snapfilter_cb_t, void *, nvlist_t **); _LIBZFS_H int zfs_send_one(zfs_handle_t *, const char *, int, sendflags_t *, const char *); _LIBZFS_H int zfs_send_progress(zfs_handle_t *, int, uint64_t *, uint64_t *); _LIBZFS_H int zfs_send_resume(libzfs_handle_t *, sendflags_t *, int outfd, const char *); _LIBZFS_H int zfs_send_saved(zfs_handle_t *, sendflags_t *, int, const char *); _LIBZFS_H nvlist_t *zfs_send_resume_token_to_nvlist(libzfs_handle_t *hdl, const char *token); _LIBZFS_H int zfs_promote(zfs_handle_t *); _LIBZFS_H int zfs_hold(zfs_handle_t *, const char *, const char *, boolean_t, int); _LIBZFS_H int zfs_hold_nvl(zfs_handle_t *, int, nvlist_t *); _LIBZFS_H int zfs_release(zfs_handle_t *, const char *, const char *, boolean_t); _LIBZFS_H int zfs_get_holds(zfs_handle_t *, nvlist_t **); _LIBZFS_H uint64_t zvol_volsize_to_reservation(zpool_handle_t *, uint64_t, nvlist_t *); typedef int (*zfs_userspace_cb_t)(void *arg, const char *domain, uid_t rid, uint64_t space); _LIBZFS_H int zfs_userspace(zfs_handle_t *, zfs_userquota_prop_t, zfs_userspace_cb_t, void *); _LIBZFS_H int zfs_get_fsacl(zfs_handle_t *, nvlist_t **); _LIBZFS_H int zfs_set_fsacl(zfs_handle_t *, boolean_t, nvlist_t *); typedef struct recvflags { /* print informational messages (ie, -v was specified) */ boolean_t verbose; /* the destination is a prefix, not the exact fs (ie, -d) */ boolean_t isprefix; /* * Only the tail of the sent snapshot path is appended to the * destination to determine the received snapshot name (ie, -e). */ boolean_t istail; /* do not actually do the recv, just check if it would work (ie, -n) */ boolean_t dryrun; /* rollback/destroy filesystems as necessary (eg, -F) */ boolean_t force; /* set "canmount=off" on all modified filesystems */ boolean_t canmountoff; /* * Mark the file systems as "resumable" and do not destroy them if the * receive is interrupted */ boolean_t resumable; /* byteswap flag is used internally; callers need not specify */ boolean_t byteswap; /* do not mount file systems as they are extracted (private) */ boolean_t nomount; /* Was holds flag set in the compound header? */ boolean_t holds; /* skip receive of snapshot holds */ boolean_t skipholds; /* mount the filesystem unless nomount is specified */ boolean_t domount; /* force unmount while recv snapshot (private) */ boolean_t forceunmount; /* use this recv to check (and heal if needed) an existing snapshot */ boolean_t heal; } recvflags_t; _LIBZFS_H int zfs_receive(libzfs_handle_t *, const char *, nvlist_t *, recvflags_t *, int, avl_tree_t *); typedef enum diff_flags { ZFS_DIFF_PARSEABLE = 1 << 0, ZFS_DIFF_TIMESTAMP = 1 << 1, ZFS_DIFF_CLASSIFY = 1 << 2, ZFS_DIFF_NO_MANGLE = 1 << 3 } diff_flags_t; _LIBZFS_H int zfs_show_diffs(zfs_handle_t *, int, const char *, const char *, int); /* * Miscellaneous functions. */ _LIBZFS_H const char *zfs_type_to_name(zfs_type_t); _LIBZFS_H void zfs_refresh_properties(zfs_handle_t *); _LIBZFS_H int zfs_name_valid(const char *, zfs_type_t); _LIBZFS_H zfs_handle_t *zfs_path_to_zhandle(libzfs_handle_t *, const char *, zfs_type_t); _LIBZFS_H int zfs_parent_name(zfs_handle_t *, char *, size_t); _LIBZFS_H boolean_t zfs_dataset_exists(libzfs_handle_t *, const char *, zfs_type_t); _LIBZFS_H int zfs_spa_version(zfs_handle_t *, int *); _LIBZFS_H boolean_t zfs_bookmark_exists(const char *path); /* * Mount support functions. */ _LIBZFS_H boolean_t is_mounted(libzfs_handle_t *, const char *special, char **); _LIBZFS_H boolean_t zfs_is_mounted(zfs_handle_t *, char **); _LIBZFS_H int zfs_mount(zfs_handle_t *, const char *, int); _LIBZFS_H int zfs_mount_at(zfs_handle_t *, const char *, int, const char *); _LIBZFS_H int zfs_unmount(zfs_handle_t *, const char *, int); _LIBZFS_H int zfs_unmountall(zfs_handle_t *, int); _LIBZFS_H int zfs_mount_delegation_check(void); #if defined(__linux__) || defined(__APPLE__) _LIBZFS_H int zfs_parse_mount_options(const char *mntopts, unsigned long *mntflags, unsigned long *zfsflags, int sloppy, char *badopt, char *mtabopt); _LIBZFS_H void zfs_adjust_mount_options(zfs_handle_t *zhp, const char *mntpoint, char *mntopts, char *mtabopt); #endif /* * Share support functions. * * enum sa_protocol * lists are terminated with SA_NO_PROTOCOL, * NULL means "all/any known to this libzfs". */ #define SA_NO_PROTOCOL -1 _LIBZFS_H boolean_t zfs_is_shared(zfs_handle_t *zhp, char **where, const enum sa_protocol *proto); _LIBZFS_H int zfs_share(zfs_handle_t *zhp, const enum sa_protocol *proto); _LIBZFS_H int zfs_unshare(zfs_handle_t *zhp, const char *mountpoint, const enum sa_protocol *proto); _LIBZFS_H int zfs_unshareall(zfs_handle_t *zhp, const enum sa_protocol *proto); _LIBZFS_H void zfs_commit_shares(const enum sa_protocol *proto); _LIBZFS_H void zfs_truncate_shares(const enum sa_protocol *proto); _LIBZFS_H int zfs_nicestrtonum(libzfs_handle_t *, const char *, uint64_t *); /* * Utility functions to run an external process. */ #define STDOUT_VERBOSE 0x01 #define STDERR_VERBOSE 0x02 #define NO_DEFAULT_PATH 0x04 /* Don't use $PATH to lookup the command */ _LIBZFS_H int libzfs_run_process(const char *, char **, int); _LIBZFS_H int libzfs_run_process_get_stdout(const char *, char *[], char *[], char **[], int *); _LIBZFS_H int libzfs_run_process_get_stdout_nopath(const char *, char *[], char *[], char **[], int *); _LIBZFS_H void libzfs_free_str_array(char **, int); _LIBZFS_H boolean_t libzfs_envvar_is_set(const char *); /* * Utility functions for zfs version */ _LIBZFS_H const char *zfs_version_userland(void); _LIBZFS_H char *zfs_version_kernel(void); _LIBZFS_H int zfs_version_print(void); /* * Given a device or file, determine if it is part of a pool. */ _LIBZFS_H int zpool_in_use(libzfs_handle_t *, int, pool_state_t *, char **, boolean_t *); /* * Label manipulation. */ _LIBZFS_H int zpool_clear_label(int); _LIBZFS_H int zpool_set_bootenv(zpool_handle_t *, const nvlist_t *); _LIBZFS_H int zpool_get_bootenv(zpool_handle_t *, nvlist_t **); /* * Management interfaces for SMB ACL files */ _LIBZFS_H int zfs_smb_acl_add(libzfs_handle_t *, char *, char *, char *); _LIBZFS_H int zfs_smb_acl_remove(libzfs_handle_t *, char *, char *, char *); _LIBZFS_H int zfs_smb_acl_purge(libzfs_handle_t *, char *, char *); _LIBZFS_H int zfs_smb_acl_rename(libzfs_handle_t *, char *, char *, char *, char *); /* * Enable and disable datasets within a pool by mounting/unmounting and * sharing/unsharing them. */ _LIBZFS_H int zpool_enable_datasets(zpool_handle_t *, const char *, int); _LIBZFS_H int zpool_disable_datasets(zpool_handle_t *, boolean_t); _LIBZFS_H void zpool_disable_datasets_os(zpool_handle_t *, boolean_t); _LIBZFS_H void zpool_disable_volume_os(const char *); /* * Parse a features file for -o compatibility */ typedef enum { ZPOOL_COMPATIBILITY_OK, ZPOOL_COMPATIBILITY_WARNTOKEN, ZPOOL_COMPATIBILITY_BADTOKEN, ZPOOL_COMPATIBILITY_BADFILE, ZPOOL_COMPATIBILITY_NOFILES } zpool_compat_status_t; _LIBZFS_H zpool_compat_status_t zpool_load_compat(const char *, boolean_t *, char *, size_t); #ifdef __FreeBSD__ /* * Attach/detach the given filesystem to/from the given jail. */ _LIBZFS_H int zfs_jail(zfs_handle_t *zhp, int jailid, int attach); /* * Set loader options for next boot. */ _LIBZFS_H int zpool_nextboot(libzfs_handle_t *, uint64_t, uint64_t, const char *); #endif /* __FreeBSD__ */ #ifdef __linux__ /* * Add or delete the given filesystem to/from the given user namespace. */ _LIBZFS_H int zfs_userns(zfs_handle_t *zhp, const char *nspath, int attach); #endif #ifdef __cplusplus } #endif #endif /* _LIBZFS_H */ diff --git a/sys/contrib/openzfs/include/libzfs_core.h b/sys/contrib/openzfs/include/libzfs_core.h index 14a4857c35da..867c18b9c226 100644 --- a/sys/contrib/openzfs/include/libzfs_core.h +++ b/sys/contrib/openzfs/include/libzfs_core.h @@ -1,162 +1,164 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2012, 2020 by Delphix. All rights reserved. * Copyright 2017 RackTop Systems. * Copyright (c) 2017 Open-E, Inc. All Rights Reserved. * Copyright (c) 2019 Datto Inc. */ #ifndef _LIBZFS_CORE_H #define _LIBZFS_CORE_H extern __attribute__((visibility("default"))) #include #include #include #include #ifdef __cplusplus extern "C" { #endif _LIBZFS_CORE_H int libzfs_core_init(void); _LIBZFS_CORE_H void libzfs_core_fini(void); struct zfs_cmd; _LIBZFS_CORE_H int lzc_ioctl_fd(int, unsigned long, struct zfs_cmd *); /* * NB: this type should be kept binary-compatible with dmu_objset_type_t. */ enum lzc_dataset_type { LZC_DATSET_TYPE_ZFS = 2, LZC_DATSET_TYPE_ZVOL }; _LIBZFS_CORE_H int lzc_snapshot(nvlist_t *, nvlist_t *, nvlist_t **); _LIBZFS_CORE_H int lzc_create(const char *, enum lzc_dataset_type, nvlist_t *, uint8_t *, uint_t); _LIBZFS_CORE_H int lzc_clone(const char *, const char *, nvlist_t *); _LIBZFS_CORE_H int lzc_promote(const char *, char *, int); _LIBZFS_CORE_H int lzc_destroy_snaps(nvlist_t *, boolean_t, nvlist_t **); _LIBZFS_CORE_H int lzc_bookmark(nvlist_t *, nvlist_t **); _LIBZFS_CORE_H int lzc_get_bookmarks(const char *, nvlist_t *, nvlist_t **); _LIBZFS_CORE_H int lzc_get_bookmark_props(const char *, nvlist_t **); _LIBZFS_CORE_H int lzc_destroy_bookmarks(nvlist_t *, nvlist_t **); _LIBZFS_CORE_H int lzc_load_key(const char *, boolean_t, uint8_t *, uint_t); _LIBZFS_CORE_H int lzc_unload_key(const char *); _LIBZFS_CORE_H int lzc_change_key(const char *, uint64_t, nvlist_t *, uint8_t *, uint_t); _LIBZFS_CORE_H int lzc_initialize(const char *, pool_initialize_func_t, nvlist_t *, nvlist_t **); _LIBZFS_CORE_H int lzc_trim(const char *, pool_trim_func_t, uint64_t, boolean_t, nvlist_t *, nvlist_t **); _LIBZFS_CORE_H int lzc_redact(const char *, const char *, nvlist_t *); _LIBZFS_CORE_H int lzc_snaprange_space(const char *, const char *, uint64_t *); _LIBZFS_CORE_H int lzc_hold(nvlist_t *, int, nvlist_t **); _LIBZFS_CORE_H int lzc_release(nvlist_t *, nvlist_t **); _LIBZFS_CORE_H int lzc_get_holds(const char *, nvlist_t **); enum lzc_send_flags { LZC_SEND_FLAG_EMBED_DATA = 1 << 0, LZC_SEND_FLAG_LARGE_BLOCK = 1 << 1, LZC_SEND_FLAG_COMPRESS = 1 << 2, LZC_SEND_FLAG_RAW = 1 << 3, LZC_SEND_FLAG_SAVED = 1 << 4, }; _LIBZFS_CORE_H int lzc_send_wrapper(int (*)(int, void *), int, void *); _LIBZFS_CORE_H int lzc_send(const char *, const char *, int, enum lzc_send_flags); _LIBZFS_CORE_H int lzc_send_resume(const char *, const char *, int, enum lzc_send_flags, uint64_t, uint64_t); _LIBZFS_CORE_H int lzc_send_space(const char *, const char *, enum lzc_send_flags, uint64_t *); struct dmu_replay_record; _LIBZFS_CORE_H int lzc_send_redacted(const char *, const char *, int, enum lzc_send_flags, const char *); _LIBZFS_CORE_H int lzc_send_resume_redacted(const char *, const char *, int, enum lzc_send_flags, uint64_t, uint64_t, const char *); _LIBZFS_CORE_H int lzc_receive(const char *, nvlist_t *, const char *, boolean_t, boolean_t, int); _LIBZFS_CORE_H int lzc_receive_resumable(const char *, nvlist_t *, const char *, boolean_t, boolean_t, int); _LIBZFS_CORE_H int lzc_receive_with_header(const char *, nvlist_t *, const char *, boolean_t, boolean_t, boolean_t, int, const struct dmu_replay_record *); _LIBZFS_CORE_H int lzc_receive_one(const char *, nvlist_t *, const char *, boolean_t, boolean_t, boolean_t, int, const struct dmu_replay_record *, int, uint64_t *, uint64_t *, uint64_t *, nvlist_t **); _LIBZFS_CORE_H int lzc_receive_with_cmdprops(const char *, nvlist_t *, nvlist_t *, uint8_t *, uint_t, const char *, boolean_t, boolean_t, boolean_t, int, const struct dmu_replay_record *, int, uint64_t *, uint64_t *, uint64_t *, nvlist_t **); _LIBZFS_CORE_H int lzc_receive_with_heal(const char *, nvlist_t *, nvlist_t *, uint8_t *, uint_t, const char *, boolean_t, boolean_t, boolean_t, boolean_t, int, const struct dmu_replay_record *, int, uint64_t *, uint64_t *, uint64_t *, nvlist_t **); _LIBZFS_CORE_H int lzc_send_space(const char *, const char *, enum lzc_send_flags, uint64_t *); _LIBZFS_CORE_H int lzc_send_space_resume_redacted(const char *, const char *, enum lzc_send_flags, uint64_t, uint64_t, uint64_t, const char *, int, uint64_t *); _LIBZFS_CORE_H uint64_t lzc_send_progress(int); _LIBZFS_CORE_H boolean_t lzc_exists(const char *); _LIBZFS_CORE_H int lzc_rollback(const char *, char *, int); _LIBZFS_CORE_H int lzc_rollback_to(const char *, const char *); _LIBZFS_CORE_H int lzc_rename(const char *, const char *); _LIBZFS_CORE_H int lzc_destroy(const char *); _LIBZFS_CORE_H int lzc_channel_program(const char *, const char *, uint64_t, uint64_t, nvlist_t *, nvlist_t **); _LIBZFS_CORE_H int lzc_channel_program_nosync(const char *, const char *, uint64_t, uint64_t, nvlist_t *, nvlist_t **); _LIBZFS_CORE_H int lzc_sync(const char *, nvlist_t *, nvlist_t **); _LIBZFS_CORE_H int lzc_reopen(const char *, boolean_t); _LIBZFS_CORE_H int lzc_pool_checkpoint(const char *); _LIBZFS_CORE_H int lzc_pool_checkpoint_discard(const char *); _LIBZFS_CORE_H int lzc_wait(const char *, zpool_wait_activity_t, boolean_t *); _LIBZFS_CORE_H int lzc_wait_tag(const char *, zpool_wait_activity_t, uint64_t, boolean_t *); _LIBZFS_CORE_H int lzc_wait_fs(const char *, zfs_wait_activity_t, boolean_t *); _LIBZFS_CORE_H int lzc_set_bootenv(const char *, const nvlist_t *); _LIBZFS_CORE_H int lzc_get_bootenv(const char *, nvlist_t **); _LIBZFS_CORE_H int lzc_get_vdev_prop(const char *, nvlist_t *, nvlist_t **); _LIBZFS_CORE_H int lzc_set_vdev_prop(const char *, nvlist_t *, nvlist_t **); +_LIBZFS_CORE_H int lzc_scrub(zfs_ioc_t, const char *, nvlist_t *, nvlist_t **); + #ifdef __cplusplus } #endif #endif /* _LIBZFS_CORE_H */ diff --git a/sys/contrib/openzfs/include/sys/dmu.h b/sys/contrib/openzfs/include/sys/dmu.h index 5ee6704668a4..7e57d133c2ec 100644 --- a/sys/contrib/openzfs/include/sys/dmu.h +++ b/sys/contrib/openzfs/include/sys/dmu.h @@ -1,1097 +1,1098 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2011, 2020 by Delphix. All rights reserved. * Copyright 2011 Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2012, Joyent, Inc. All rights reserved. * Copyright 2014 HybridCluster. All rights reserved. * Copyright (c) 2014 Spectra Logic Corporation, All rights reserved. * Copyright 2013 Saso Kiselkov. All rights reserved. * Copyright (c) 2017, Intel Corporation. * Copyright (c) 2022 Hewlett Packard Enterprise Development LP. */ /* Portions Copyright 2010 Robert Milkowski */ #ifndef _SYS_DMU_H #define _SYS_DMU_H /* * This file describes the interface that the DMU provides for its * consumers. * * The DMU also interacts with the SPA. That interface is described in * dmu_spa.h. */ #include #include #include #include #include #include #include #include #ifdef __cplusplus extern "C" { #endif struct page; struct vnode; struct spa; struct zilog; struct zio; struct blkptr; struct zap_cursor; struct dsl_dataset; struct dsl_pool; struct dnode; struct drr_begin; struct drr_end; struct zbookmark_phys; struct spa; struct nvlist; struct arc_buf; struct zio_prop; struct sa_handle; struct dsl_crypto_params; struct locked_range; typedef struct objset objset_t; typedef struct dmu_tx dmu_tx_t; typedef struct dsl_dir dsl_dir_t; typedef struct dnode dnode_t; typedef enum dmu_object_byteswap { DMU_BSWAP_UINT8, DMU_BSWAP_UINT16, DMU_BSWAP_UINT32, DMU_BSWAP_UINT64, DMU_BSWAP_ZAP, DMU_BSWAP_DNODE, DMU_BSWAP_OBJSET, DMU_BSWAP_ZNODE, DMU_BSWAP_OLDACL, DMU_BSWAP_ACL, /* * Allocating a new byteswap type number makes the on-disk format * incompatible with any other format that uses the same number. * * Data can usually be structured to work with one of the * DMU_BSWAP_UINT* or DMU_BSWAP_ZAP types. */ DMU_BSWAP_NUMFUNCS } dmu_object_byteswap_t; #define DMU_OT_NEWTYPE 0x80 #define DMU_OT_METADATA 0x40 #define DMU_OT_ENCRYPTED 0x20 #define DMU_OT_BYTESWAP_MASK 0x1f /* * Defines a uint8_t object type. Object types specify if the data * in the object is metadata (boolean) and how to byteswap the data * (dmu_object_byteswap_t). All of the types created by this method * are cached in the dbuf metadata cache. */ #define DMU_OT(byteswap, metadata, encrypted) \ (DMU_OT_NEWTYPE | \ ((metadata) ? DMU_OT_METADATA : 0) | \ ((encrypted) ? DMU_OT_ENCRYPTED : 0) | \ ((byteswap) & DMU_OT_BYTESWAP_MASK)) #define DMU_OT_IS_VALID(ot) (((ot) & DMU_OT_NEWTYPE) ? \ ((ot) & DMU_OT_BYTESWAP_MASK) < DMU_BSWAP_NUMFUNCS : \ (ot) < DMU_OT_NUMTYPES) #define DMU_OT_IS_METADATA_CACHED(ot) (((ot) & DMU_OT_NEWTYPE) ? \ B_TRUE : dmu_ot[(ot)].ot_dbuf_metadata_cache) /* * MDB doesn't have dmu_ot; it defines these macros itself. */ #ifndef ZFS_MDB #define DMU_OT_IS_METADATA_IMPL(ot) (dmu_ot[ot].ot_metadata) #define DMU_OT_IS_ENCRYPTED_IMPL(ot) (dmu_ot[ot].ot_encrypt) #define DMU_OT_BYTESWAP_IMPL(ot) (dmu_ot[ot].ot_byteswap) #endif #define DMU_OT_IS_METADATA(ot) (((ot) & DMU_OT_NEWTYPE) ? \ (((ot) & DMU_OT_METADATA) != 0) : \ DMU_OT_IS_METADATA_IMPL(ot)) #define DMU_OT_IS_DDT(ot) \ ((ot) == DMU_OT_DDT_ZAP) #define DMU_OT_IS_CRITICAL(ot) \ (DMU_OT_IS_METADATA(ot) && \ (ot) != DMU_OT_DNODE && \ (ot) != DMU_OT_DIRECTORY_CONTENTS && \ (ot) != DMU_OT_SA) /* Note: ztest uses DMU_OT_UINT64_OTHER as a proxy for file blocks */ #define DMU_OT_IS_FILE(ot) \ ((ot) == DMU_OT_PLAIN_FILE_CONTENTS || (ot) == DMU_OT_UINT64_OTHER) #define DMU_OT_IS_ENCRYPTED(ot) (((ot) & DMU_OT_NEWTYPE) ? \ (((ot) & DMU_OT_ENCRYPTED) != 0) : \ DMU_OT_IS_ENCRYPTED_IMPL(ot)) /* * These object types use bp_fill != 1 for their L0 bp's. Therefore they can't * have their data embedded (i.e. use a BP_IS_EMBEDDED() bp), because bp_fill * is repurposed for embedded BPs. */ #define DMU_OT_HAS_FILL(ot) \ ((ot) == DMU_OT_DNODE || (ot) == DMU_OT_OBJSET) #define DMU_OT_BYTESWAP(ot) (((ot) & DMU_OT_NEWTYPE) ? \ ((ot) & DMU_OT_BYTESWAP_MASK) : \ DMU_OT_BYTESWAP_IMPL(ot)) typedef enum dmu_object_type { DMU_OT_NONE, /* general: */ DMU_OT_OBJECT_DIRECTORY, /* ZAP */ DMU_OT_OBJECT_ARRAY, /* UINT64 */ DMU_OT_PACKED_NVLIST, /* UINT8 (XDR by nvlist_pack/unpack) */ DMU_OT_PACKED_NVLIST_SIZE, /* UINT64 */ DMU_OT_BPOBJ, /* UINT64 */ DMU_OT_BPOBJ_HDR, /* UINT64 */ /* spa: */ DMU_OT_SPACE_MAP_HEADER, /* UINT64 */ DMU_OT_SPACE_MAP, /* UINT64 */ /* zil: */ DMU_OT_INTENT_LOG, /* UINT64 */ /* dmu: */ DMU_OT_DNODE, /* DNODE */ DMU_OT_OBJSET, /* OBJSET */ /* dsl: */ DMU_OT_DSL_DIR, /* UINT64 */ DMU_OT_DSL_DIR_CHILD_MAP, /* ZAP */ DMU_OT_DSL_DS_SNAP_MAP, /* ZAP */ DMU_OT_DSL_PROPS, /* ZAP */ DMU_OT_DSL_DATASET, /* UINT64 */ /* zpl: */ DMU_OT_ZNODE, /* ZNODE */ DMU_OT_OLDACL, /* Old ACL */ DMU_OT_PLAIN_FILE_CONTENTS, /* UINT8 */ DMU_OT_DIRECTORY_CONTENTS, /* ZAP */ DMU_OT_MASTER_NODE, /* ZAP */ DMU_OT_UNLINKED_SET, /* ZAP */ /* zvol: */ DMU_OT_ZVOL, /* UINT8 */ DMU_OT_ZVOL_PROP, /* ZAP */ /* other; for testing only! */ DMU_OT_PLAIN_OTHER, /* UINT8 */ DMU_OT_UINT64_OTHER, /* UINT64 */ DMU_OT_ZAP_OTHER, /* ZAP */ /* new object types: */ DMU_OT_ERROR_LOG, /* ZAP */ DMU_OT_SPA_HISTORY, /* UINT8 */ DMU_OT_SPA_HISTORY_OFFSETS, /* spa_his_phys_t */ DMU_OT_POOL_PROPS, /* ZAP */ DMU_OT_DSL_PERMS, /* ZAP */ DMU_OT_ACL, /* ACL */ DMU_OT_SYSACL, /* SYSACL */ DMU_OT_FUID, /* FUID table (Packed NVLIST UINT8) */ DMU_OT_FUID_SIZE, /* FUID table size UINT64 */ DMU_OT_NEXT_CLONES, /* ZAP */ DMU_OT_SCAN_QUEUE, /* ZAP */ DMU_OT_USERGROUP_USED, /* ZAP */ DMU_OT_USERGROUP_QUOTA, /* ZAP */ DMU_OT_USERREFS, /* ZAP */ DMU_OT_DDT_ZAP, /* ZAP */ DMU_OT_DDT_STATS, /* ZAP */ DMU_OT_SA, /* System attr */ DMU_OT_SA_MASTER_NODE, /* ZAP */ DMU_OT_SA_ATTR_REGISTRATION, /* ZAP */ DMU_OT_SA_ATTR_LAYOUTS, /* ZAP */ DMU_OT_SCAN_XLATE, /* ZAP */ DMU_OT_DEDUP, /* fake dedup BP from ddt_bp_create() */ DMU_OT_DEADLIST, /* ZAP */ DMU_OT_DEADLIST_HDR, /* UINT64 */ DMU_OT_DSL_CLONES, /* ZAP */ DMU_OT_BPOBJ_SUBOBJ, /* UINT64 */ /* * Do not allocate new object types here. Doing so makes the on-disk * format incompatible with any other format that uses the same object * type number. * * When creating an object which does not have one of the above types * use the DMU_OTN_* type with the correct byteswap and metadata * values. * * The DMU_OTN_* types do not have entries in the dmu_ot table, * use the DMU_OT_IS_METADATA() and DMU_OT_BYTESWAP() macros instead * of indexing into dmu_ot directly (this works for both DMU_OT_* types * and DMU_OTN_* types). */ DMU_OT_NUMTYPES, /* * Names for valid types declared with DMU_OT(). */ DMU_OTN_UINT8_DATA = DMU_OT(DMU_BSWAP_UINT8, B_FALSE, B_FALSE), DMU_OTN_UINT8_METADATA = DMU_OT(DMU_BSWAP_UINT8, B_TRUE, B_FALSE), DMU_OTN_UINT16_DATA = DMU_OT(DMU_BSWAP_UINT16, B_FALSE, B_FALSE), DMU_OTN_UINT16_METADATA = DMU_OT(DMU_BSWAP_UINT16, B_TRUE, B_FALSE), DMU_OTN_UINT32_DATA = DMU_OT(DMU_BSWAP_UINT32, B_FALSE, B_FALSE), DMU_OTN_UINT32_METADATA = DMU_OT(DMU_BSWAP_UINT32, B_TRUE, B_FALSE), DMU_OTN_UINT64_DATA = DMU_OT(DMU_BSWAP_UINT64, B_FALSE, B_FALSE), DMU_OTN_UINT64_METADATA = DMU_OT(DMU_BSWAP_UINT64, B_TRUE, B_FALSE), DMU_OTN_ZAP_DATA = DMU_OT(DMU_BSWAP_ZAP, B_FALSE, B_FALSE), DMU_OTN_ZAP_METADATA = DMU_OT(DMU_BSWAP_ZAP, B_TRUE, B_FALSE), DMU_OTN_UINT8_ENC_DATA = DMU_OT(DMU_BSWAP_UINT8, B_FALSE, B_TRUE), DMU_OTN_UINT8_ENC_METADATA = DMU_OT(DMU_BSWAP_UINT8, B_TRUE, B_TRUE), DMU_OTN_UINT16_ENC_DATA = DMU_OT(DMU_BSWAP_UINT16, B_FALSE, B_TRUE), DMU_OTN_UINT16_ENC_METADATA = DMU_OT(DMU_BSWAP_UINT16, B_TRUE, B_TRUE), DMU_OTN_UINT32_ENC_DATA = DMU_OT(DMU_BSWAP_UINT32, B_FALSE, B_TRUE), DMU_OTN_UINT32_ENC_METADATA = DMU_OT(DMU_BSWAP_UINT32, B_TRUE, B_TRUE), DMU_OTN_UINT64_ENC_DATA = DMU_OT(DMU_BSWAP_UINT64, B_FALSE, B_TRUE), DMU_OTN_UINT64_ENC_METADATA = DMU_OT(DMU_BSWAP_UINT64, B_TRUE, B_TRUE), DMU_OTN_ZAP_ENC_DATA = DMU_OT(DMU_BSWAP_ZAP, B_FALSE, B_TRUE), DMU_OTN_ZAP_ENC_METADATA = DMU_OT(DMU_BSWAP_ZAP, B_TRUE, B_TRUE), } dmu_object_type_t; /* * These flags are intended to be used to specify the "txg_how" * parameter when calling the dmu_tx_assign() function. See the comment * above dmu_tx_assign() for more details on the meaning of these flags. */ #define TXG_NOWAIT (0ULL) #define TXG_WAIT (1ULL<<0) #define TXG_NOTHROTTLE (1ULL<<1) void byteswap_uint64_array(void *buf, size_t size); void byteswap_uint32_array(void *buf, size_t size); void byteswap_uint16_array(void *buf, size_t size); void byteswap_uint8_array(void *buf, size_t size); void zap_byteswap(void *buf, size_t size); void zfs_oldacl_byteswap(void *buf, size_t size); void zfs_acl_byteswap(void *buf, size_t size); void zfs_znode_byteswap(void *buf, size_t size); #define DS_FIND_SNAPSHOTS (1<<0) #define DS_FIND_CHILDREN (1<<1) #define DS_FIND_SERIALIZE (1<<2) /* * The maximum number of bytes that can be accessed as part of one * operation, including metadata. */ #define DMU_MAX_ACCESS (64 * 1024 * 1024) /* 64MB */ #define DMU_MAX_DELETEBLKCNT (20480) /* ~5MB of indirect blocks */ #define DMU_USERUSED_OBJECT (-1ULL) #define DMU_GROUPUSED_OBJECT (-2ULL) #define DMU_PROJECTUSED_OBJECT (-3ULL) /* * Zap prefix for object accounting in DMU_{USER,GROUP,PROJECT}USED_OBJECT. */ #define DMU_OBJACCT_PREFIX "obj-" #define DMU_OBJACCT_PREFIX_LEN 4 /* * artificial blkids for bonus buffer and spill blocks */ #define DMU_BONUS_BLKID (-1ULL) #define DMU_SPILL_BLKID (-2ULL) /* * Public routines to create, destroy, open, and close objsets. */ typedef void dmu_objset_create_sync_func_t(objset_t *os, void *arg, cred_t *cr, dmu_tx_t *tx); int dmu_objset_hold(const char *name, const void *tag, objset_t **osp); int dmu_objset_own(const char *name, dmu_objset_type_t type, boolean_t readonly, boolean_t key_required, const void *tag, objset_t **osp); void dmu_objset_rele(objset_t *os, const void *tag); void dmu_objset_disown(objset_t *os, boolean_t key_required, const void *tag); int dmu_objset_open_ds(struct dsl_dataset *ds, objset_t **osp); void dmu_objset_evict_dbufs(objset_t *os); int dmu_objset_create(const char *name, dmu_objset_type_t type, uint64_t flags, struct dsl_crypto_params *dcp, dmu_objset_create_sync_func_t func, void *arg); int dmu_objset_clone(const char *name, const char *origin); int dsl_destroy_snapshots_nvl(struct nvlist *snaps, boolean_t defer, struct nvlist *errlist); int dmu_objset_snapshot_one(const char *fsname, const char *snapname); int dmu_objset_find(const char *name, int func(const char *, void *), void *arg, int flags); void dmu_objset_byteswap(void *buf, size_t size); int dsl_dataset_rename_snapshot(const char *fsname, const char *oldsnapname, const char *newsnapname, boolean_t recursive); typedef struct dmu_buf { uint64_t db_object; /* object that this buffer is part of */ uint64_t db_offset; /* byte offset in this object */ uint64_t db_size; /* size of buffer in bytes */ void *db_data; /* data in buffer */ } dmu_buf_t; /* * The names of zap entries in the DIRECTORY_OBJECT of the MOS. */ #define DMU_POOL_DIRECTORY_OBJECT 1 #define DMU_POOL_CONFIG "config" #define DMU_POOL_FEATURES_FOR_WRITE "features_for_write" #define DMU_POOL_FEATURES_FOR_READ "features_for_read" #define DMU_POOL_FEATURE_DESCRIPTIONS "feature_descriptions" #define DMU_POOL_FEATURE_ENABLED_TXG "feature_enabled_txg" #define DMU_POOL_ROOT_DATASET "root_dataset" #define DMU_POOL_SYNC_BPOBJ "sync_bplist" #define DMU_POOL_ERRLOG_SCRUB "errlog_scrub" #define DMU_POOL_ERRLOG_LAST "errlog_last" #define DMU_POOL_SPARES "spares" #define DMU_POOL_DEFLATE "deflate" #define DMU_POOL_HISTORY "history" #define DMU_POOL_PROPS "pool_props" #define DMU_POOL_L2CACHE "l2cache" #define DMU_POOL_TMP_USERREFS "tmp_userrefs" #define DMU_POOL_DDT "DDT-%s-%s-%s" #define DMU_POOL_DDT_STATS "DDT-statistics" #define DMU_POOL_CREATION_VERSION "creation_version" #define DMU_POOL_SCAN "scan" +#define DMU_POOL_ERRORSCRUB "error_scrub" #define DMU_POOL_FREE_BPOBJ "free_bpobj" #define DMU_POOL_BPTREE_OBJ "bptree_obj" #define DMU_POOL_EMPTY_BPOBJ "empty_bpobj" #define DMU_POOL_CHECKSUM_SALT "org.illumos:checksum_salt" #define DMU_POOL_VDEV_ZAP_MAP "com.delphix:vdev_zap_map" #define DMU_POOL_REMOVING "com.delphix:removing" #define DMU_POOL_OBSOLETE_BPOBJ "com.delphix:obsolete_bpobj" #define DMU_POOL_CONDENSING_INDIRECT "com.delphix:condensing_indirect" #define DMU_POOL_ZPOOL_CHECKPOINT "com.delphix:zpool_checkpoint" #define DMU_POOL_LOG_SPACEMAP_ZAP "com.delphix:log_spacemap_zap" #define DMU_POOL_DELETED_CLONES "com.delphix:deleted_clones" /* * Allocate an object from this objset. The range of object numbers * available is (0, DN_MAX_OBJECT). Object 0 is the meta-dnode. * * The transaction must be assigned to a txg. The newly allocated * object will be "held" in the transaction (ie. you can modify the * newly allocated object in this transaction). * * dmu_object_alloc() chooses an object and returns it in *objectp. * * dmu_object_claim() allocates a specific object number. If that * number is already allocated, it fails and returns EEXIST. * * Return 0 on success, or ENOSPC or EEXIST as specified above. */ uint64_t dmu_object_alloc(objset_t *os, dmu_object_type_t ot, int blocksize, dmu_object_type_t bonus_type, int bonus_len, dmu_tx_t *tx); uint64_t dmu_object_alloc_ibs(objset_t *os, dmu_object_type_t ot, int blocksize, int indirect_blockshift, dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx); uint64_t dmu_object_alloc_dnsize(objset_t *os, dmu_object_type_t ot, int blocksize, dmu_object_type_t bonus_type, int bonus_len, int dnodesize, dmu_tx_t *tx); uint64_t dmu_object_alloc_hold(objset_t *os, dmu_object_type_t ot, int blocksize, int indirect_blockshift, dmu_object_type_t bonustype, int bonuslen, int dnodesize, dnode_t **allocated_dnode, const void *tag, dmu_tx_t *tx); int dmu_object_claim(objset_t *os, uint64_t object, dmu_object_type_t ot, int blocksize, dmu_object_type_t bonus_type, int bonus_len, dmu_tx_t *tx); int dmu_object_claim_dnsize(objset_t *os, uint64_t object, dmu_object_type_t ot, int blocksize, dmu_object_type_t bonus_type, int bonus_len, int dnodesize, dmu_tx_t *tx); int dmu_object_reclaim(objset_t *os, uint64_t object, dmu_object_type_t ot, int blocksize, dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *txp); int dmu_object_reclaim_dnsize(objset_t *os, uint64_t object, dmu_object_type_t ot, int blocksize, dmu_object_type_t bonustype, int bonuslen, int dnodesize, boolean_t keep_spill, dmu_tx_t *tx); int dmu_object_rm_spill(objset_t *os, uint64_t object, dmu_tx_t *tx); /* * Free an object from this objset. * * The object's data will be freed as well (ie. you don't need to call * dmu_free(object, 0, -1, tx)). * * The object need not be held in the transaction. * * If there are any holds on this object's buffers (via dmu_buf_hold()), * or tx holds on the object (via dmu_tx_hold_object()), you can not * free it; it fails and returns EBUSY. * * If the object is not allocated, it fails and returns ENOENT. * * Return 0 on success, or EBUSY or ENOENT as specified above. */ int dmu_object_free(objset_t *os, uint64_t object, dmu_tx_t *tx); /* * Find the next allocated or free object. * * The objectp parameter is in-out. It will be updated to be the next * object which is allocated. Ignore objects which have not been * modified since txg. * * XXX Can only be called on a objset with no dirty data. * * Returns 0 on success, or ENOENT if there are no more objects. */ int dmu_object_next(objset_t *os, uint64_t *objectp, boolean_t hole, uint64_t txg); /* * Set the number of levels on a dnode. nlevels must be greater than the * current number of levels or an EINVAL will be returned. */ int dmu_object_set_nlevels(objset_t *os, uint64_t object, int nlevels, dmu_tx_t *tx); /* * Set the data blocksize for an object. * * The object cannot have any blocks allocated beyond the first. If * the first block is allocated already, the new size must be greater * than the current block size. If these conditions are not met, * ENOTSUP will be returned. * * Returns 0 on success, or EBUSY if there are any holds on the object * contents, or ENOTSUP as described above. */ int dmu_object_set_blocksize(objset_t *os, uint64_t object, uint64_t size, int ibs, dmu_tx_t *tx); /* * Manually set the maxblkid on a dnode. This will adjust nlevels accordingly * to accommodate the change. When calling this function, the caller must * ensure that the object's nlevels can sufficiently support the new maxblkid. */ int dmu_object_set_maxblkid(objset_t *os, uint64_t object, uint64_t maxblkid, dmu_tx_t *tx); /* * Set the checksum property on a dnode. The new checksum algorithm will * apply to all newly written blocks; existing blocks will not be affected. */ void dmu_object_set_checksum(objset_t *os, uint64_t object, uint8_t checksum, dmu_tx_t *tx); /* * Set the compress property on a dnode. The new compression algorithm will * apply to all newly written blocks; existing blocks will not be affected. */ void dmu_object_set_compress(objset_t *os, uint64_t object, uint8_t compress, dmu_tx_t *tx); void dmu_write_embedded(objset_t *os, uint64_t object, uint64_t offset, void *data, uint8_t etype, uint8_t comp, int uncompressed_size, int compressed_size, int byteorder, dmu_tx_t *tx); void dmu_redact(objset_t *os, uint64_t object, uint64_t offset, uint64_t size, dmu_tx_t *tx); /* * Decide how to write a block: checksum, compression, number of copies, etc. */ #define WP_NOFILL 0x1 #define WP_DMU_SYNC 0x2 #define WP_SPILL 0x4 void dmu_write_policy(objset_t *os, dnode_t *dn, int level, int wp, struct zio_prop *zp); /* * The bonus data is accessed more or less like a regular buffer. * You must dmu_bonus_hold() to get the buffer, which will give you a * dmu_buf_t with db_offset==-1ULL, and db_size = the size of the bonus * data. As with any normal buffer, you must call dmu_buf_will_dirty() * before modifying it, and the * object must be held in an assigned transaction before calling * dmu_buf_will_dirty. You may use dmu_buf_set_user() on the bonus * buffer as well. You must release what you hold with dmu_buf_rele(). * * Returns ENOENT, EIO, or 0. */ int dmu_bonus_hold(objset_t *os, uint64_t object, const void *tag, dmu_buf_t **dbp); int dmu_bonus_hold_by_dnode(dnode_t *dn, const void *tag, dmu_buf_t **dbp, uint32_t flags); int dmu_bonus_max(void); int dmu_set_bonus(dmu_buf_t *, int, dmu_tx_t *); int dmu_set_bonustype(dmu_buf_t *, dmu_object_type_t, dmu_tx_t *); dmu_object_type_t dmu_get_bonustype(dmu_buf_t *); int dmu_rm_spill(objset_t *, uint64_t, dmu_tx_t *); /* * Special spill buffer support used by "SA" framework */ int dmu_spill_hold_by_bonus(dmu_buf_t *bonus, uint32_t flags, const void *tag, dmu_buf_t **dbp); int dmu_spill_hold_by_dnode(dnode_t *dn, uint32_t flags, const void *tag, dmu_buf_t **dbp); int dmu_spill_hold_existing(dmu_buf_t *bonus, const void *tag, dmu_buf_t **dbp); /* * Obtain the DMU buffer from the specified object which contains the * specified offset. dmu_buf_hold() puts a "hold" on the buffer, so * that it will remain in memory. You must release the hold with * dmu_buf_rele(). You must not access the dmu_buf_t after releasing * what you hold. You must have a hold on any dmu_buf_t* you pass to the DMU. * * You must call dmu_buf_read, dmu_buf_will_dirty, or dmu_buf_will_fill * on the returned buffer before reading or writing the buffer's * db_data. The comments for those routines describe what particular * operations are valid after calling them. * * The object number must be a valid, allocated object number. */ int dmu_buf_hold(objset_t *os, uint64_t object, uint64_t offset, const void *tag, dmu_buf_t **, int flags); int dmu_buf_hold_array(objset_t *os, uint64_t object, uint64_t offset, uint64_t length, int read, const void *tag, int *numbufsp, dmu_buf_t ***dbpp); int dmu_buf_hold_by_dnode(dnode_t *dn, uint64_t offset, const void *tag, dmu_buf_t **dbp, int flags); int dmu_buf_hold_array_by_dnode(dnode_t *dn, uint64_t offset, uint64_t length, boolean_t read, const void *tag, int *numbufsp, dmu_buf_t ***dbpp, uint32_t flags); /* * Add a reference to a dmu buffer that has already been held via * dmu_buf_hold() in the current context. */ void dmu_buf_add_ref(dmu_buf_t *db, const void *tag); /* * Attempt to add a reference to a dmu buffer that is in an unknown state, * using a pointer that may have been invalidated by eviction processing. * The request will succeed if the passed in dbuf still represents the * same os/object/blkid, is ineligible for eviction, and has at least * one hold by a user other than the syncer. */ boolean_t dmu_buf_try_add_ref(dmu_buf_t *, objset_t *os, uint64_t object, uint64_t blkid, const void *tag); void dmu_buf_rele(dmu_buf_t *db, const void *tag); uint64_t dmu_buf_refcount(dmu_buf_t *db); uint64_t dmu_buf_user_refcount(dmu_buf_t *db); /* * dmu_buf_hold_array holds the DMU buffers which contain all bytes in a * range of an object. A pointer to an array of dmu_buf_t*'s is * returned (in *dbpp). * * dmu_buf_rele_array releases the hold on an array of dmu_buf_t*'s, and * frees the array. The hold on the array of buffers MUST be released * with dmu_buf_rele_array. You can NOT release the hold on each buffer * individually with dmu_buf_rele. */ int dmu_buf_hold_array_by_bonus(dmu_buf_t *db, uint64_t offset, uint64_t length, boolean_t read, const void *tag, int *numbufsp, dmu_buf_t ***dbpp); void dmu_buf_rele_array(dmu_buf_t **, int numbufs, const void *tag); typedef void dmu_buf_evict_func_t(void *user_ptr); /* * A DMU buffer user object may be associated with a dbuf for the * duration of its lifetime. This allows the user of a dbuf (client) * to attach private data to a dbuf (e.g. in-core only data such as a * dnode_children_t, zap_t, or zap_leaf_t) and be optionally notified * when that dbuf has been evicted. Clients typically respond to the * eviction notification by freeing their private data, thus ensuring * the same lifetime for both dbuf and private data. * * The mapping from a dmu_buf_user_t to any client private data is the * client's responsibility. All current consumers of the API with private * data embed a dmu_buf_user_t as the first member of the structure for * their private data. This allows conversions between the two types * with a simple cast. Since the DMU buf user API never needs access * to the private data, other strategies can be employed if necessary * or convenient for the client (e.g. using container_of() to do the * conversion for private data that cannot have the dmu_buf_user_t as * its first member). * * Eviction callbacks are executed without the dbuf mutex held or any * other type of mechanism to guarantee that the dbuf is still available. * For this reason, users must assume the dbuf has already been freed * and not reference the dbuf from the callback context. * * Users requesting "immediate eviction" are notified as soon as the dbuf * is only referenced by dirty records (dirties == holds). Otherwise the * notification occurs after eviction processing for the dbuf begins. */ typedef struct dmu_buf_user { /* * Asynchronous user eviction callback state. */ taskq_ent_t dbu_tqent; /* * This instance's eviction function pointers. * * dbu_evict_func_sync is called synchronously and then * dbu_evict_func_async is executed asynchronously on a taskq. */ dmu_buf_evict_func_t *dbu_evict_func_sync; dmu_buf_evict_func_t *dbu_evict_func_async; #ifdef ZFS_DEBUG /* * Pointer to user's dbuf pointer. NULL for clients that do * not associate a dbuf with their user data. * * The dbuf pointer is cleared upon eviction so as to catch * use-after-evict bugs in clients. */ dmu_buf_t **dbu_clear_on_evict_dbufp; #endif } dmu_buf_user_t; /* * Initialize the given dmu_buf_user_t instance with the eviction function * evict_func, to be called when the user is evicted. * * NOTE: This function should only be called once on a given dmu_buf_user_t. * To allow enforcement of this, dbu must already be zeroed on entry. */ static inline void dmu_buf_init_user(dmu_buf_user_t *dbu, dmu_buf_evict_func_t *evict_func_sync, dmu_buf_evict_func_t *evict_func_async, dmu_buf_t **clear_on_evict_dbufp __maybe_unused) { ASSERT(dbu->dbu_evict_func_sync == NULL); ASSERT(dbu->dbu_evict_func_async == NULL); /* must have at least one evict func */ IMPLY(evict_func_sync == NULL, evict_func_async != NULL); dbu->dbu_evict_func_sync = evict_func_sync; dbu->dbu_evict_func_async = evict_func_async; taskq_init_ent(&dbu->dbu_tqent); #ifdef ZFS_DEBUG dbu->dbu_clear_on_evict_dbufp = clear_on_evict_dbufp; #endif } /* * Attach user data to a dbuf and mark it for normal (when the dbuf's * data is cleared or its reference count goes to zero) eviction processing. * * Returns NULL on success, or the existing user if another user currently * owns the buffer. */ void *dmu_buf_set_user(dmu_buf_t *db, dmu_buf_user_t *user); /* * Attach user data to a dbuf and mark it for immediate (its dirty and * reference counts are equal) eviction processing. * * Returns NULL on success, or the existing user if another user currently * owns the buffer. */ void *dmu_buf_set_user_ie(dmu_buf_t *db, dmu_buf_user_t *user); /* * Replace the current user of a dbuf. * * If given the current user of a dbuf, replaces the dbuf's user with * "new_user" and returns the user data pointer that was replaced. * Otherwise returns the current, and unmodified, dbuf user pointer. */ void *dmu_buf_replace_user(dmu_buf_t *db, dmu_buf_user_t *old_user, dmu_buf_user_t *new_user); /* * Remove the specified user data for a DMU buffer. * * Returns the user that was removed on success, or the current user if * another user currently owns the buffer. */ void *dmu_buf_remove_user(dmu_buf_t *db, dmu_buf_user_t *user); /* * Returns the user data (dmu_buf_user_t *) associated with this dbuf. */ void *dmu_buf_get_user(dmu_buf_t *db); objset_t *dmu_buf_get_objset(dmu_buf_t *db); dnode_t *dmu_buf_dnode_enter(dmu_buf_t *db); void dmu_buf_dnode_exit(dmu_buf_t *db); /* Block until any in-progress dmu buf user evictions complete. */ void dmu_buf_user_evict_wait(void); /* * Returns the blkptr associated with this dbuf, or NULL if not set. */ struct blkptr *dmu_buf_get_blkptr(dmu_buf_t *db); /* * Indicate that you are going to modify the buffer's data (db_data). * * The transaction (tx) must be assigned to a txg (ie. you've called * dmu_tx_assign()). The buffer's object must be held in the tx * (ie. you've called dmu_tx_hold_object(tx, db->db_object)). */ void dmu_buf_will_dirty(dmu_buf_t *db, dmu_tx_t *tx); boolean_t dmu_buf_is_dirty(dmu_buf_t *db, dmu_tx_t *tx); void dmu_buf_set_crypt_params(dmu_buf_t *db_fake, boolean_t byteorder, const uint8_t *salt, const uint8_t *iv, const uint8_t *mac, dmu_tx_t *tx); /* * You must create a transaction, then hold the objects which you will * (or might) modify as part of this transaction. Then you must assign * the transaction to a transaction group. Once the transaction has * been assigned, you can modify buffers which belong to held objects as * part of this transaction. You can't modify buffers before the * transaction has been assigned; you can't modify buffers which don't * belong to objects which this transaction holds; you can't hold * objects once the transaction has been assigned. You may hold an * object which you are going to free (with dmu_object_free()), but you * don't have to. * * You can abort the transaction before it has been assigned. * * Note that you may hold buffers (with dmu_buf_hold) at any time, * regardless of transaction state. */ #define DMU_NEW_OBJECT (-1ULL) #define DMU_OBJECT_END (-1ULL) dmu_tx_t *dmu_tx_create(objset_t *os); void dmu_tx_hold_write(dmu_tx_t *tx, uint64_t object, uint64_t off, int len); void dmu_tx_hold_write_by_dnode(dmu_tx_t *tx, dnode_t *dn, uint64_t off, int len); void dmu_tx_hold_append(dmu_tx_t *tx, uint64_t object, uint64_t off, int len); void dmu_tx_hold_append_by_dnode(dmu_tx_t *tx, dnode_t *dn, uint64_t off, int len); void dmu_tx_hold_clone_by_dnode(dmu_tx_t *tx, dnode_t *dn, uint64_t off, int len); void dmu_tx_hold_free(dmu_tx_t *tx, uint64_t object, uint64_t off, uint64_t len); void dmu_tx_hold_free_by_dnode(dmu_tx_t *tx, dnode_t *dn, uint64_t off, uint64_t len); void dmu_tx_hold_zap(dmu_tx_t *tx, uint64_t object, int add, const char *name); void dmu_tx_hold_zap_by_dnode(dmu_tx_t *tx, dnode_t *dn, int add, const char *name); void dmu_tx_hold_bonus(dmu_tx_t *tx, uint64_t object); void dmu_tx_hold_bonus_by_dnode(dmu_tx_t *tx, dnode_t *dn); void dmu_tx_hold_spill(dmu_tx_t *tx, uint64_t object); void dmu_tx_hold_sa(dmu_tx_t *tx, struct sa_handle *hdl, boolean_t may_grow); void dmu_tx_hold_sa_create(dmu_tx_t *tx, int total_size); void dmu_tx_abort(dmu_tx_t *tx); int dmu_tx_assign(dmu_tx_t *tx, uint64_t txg_how); void dmu_tx_wait(dmu_tx_t *tx); void dmu_tx_commit(dmu_tx_t *tx); void dmu_tx_mark_netfree(dmu_tx_t *tx); /* * To register a commit callback, dmu_tx_callback_register() must be called. * * dcb_data is a pointer to caller private data that is passed on as a * callback parameter. The caller is responsible for properly allocating and * freeing it. * * When registering a callback, the transaction must be already created, but * it cannot be committed or aborted. It can be assigned to a txg or not. * * The callback will be called after the transaction has been safely written * to stable storage and will also be called if the dmu_tx is aborted. * If there is any error which prevents the transaction from being committed to * disk, the callback will be called with a value of error != 0. * * When multiple callbacks are registered to the transaction, the callbacks * will be called in reverse order to let Lustre, the only user of commit * callback currently, take the fast path of its commit callback handling. */ typedef void dmu_tx_callback_func_t(void *dcb_data, int error); void dmu_tx_callback_register(dmu_tx_t *tx, dmu_tx_callback_func_t *dcb_func, void *dcb_data); void dmu_tx_do_callbacks(list_t *cb_list, int error); /* * Free up the data blocks for a defined range of a file. If size is * -1, the range from offset to end-of-file is freed. */ int dmu_free_range(objset_t *os, uint64_t object, uint64_t offset, uint64_t size, dmu_tx_t *tx); int dmu_free_long_range(objset_t *os, uint64_t object, uint64_t offset, uint64_t size); int dmu_free_long_object(objset_t *os, uint64_t object); /* * Convenience functions. * * Canfail routines will return 0 on success, or an errno if there is a * nonrecoverable I/O error. */ #define DMU_READ_PREFETCH 0 /* prefetch */ #define DMU_READ_NO_PREFETCH 1 /* don't prefetch */ #define DMU_READ_NO_DECRYPT 2 /* don't decrypt */ int dmu_read(objset_t *os, uint64_t object, uint64_t offset, uint64_t size, void *buf, uint32_t flags); int dmu_read_by_dnode(dnode_t *dn, uint64_t offset, uint64_t size, void *buf, uint32_t flags); void dmu_write(objset_t *os, uint64_t object, uint64_t offset, uint64_t size, const void *buf, dmu_tx_t *tx); void dmu_write_by_dnode(dnode_t *dn, uint64_t offset, uint64_t size, const void *buf, dmu_tx_t *tx); void dmu_prealloc(objset_t *os, uint64_t object, uint64_t offset, uint64_t size, dmu_tx_t *tx); #ifdef _KERNEL int dmu_read_uio(objset_t *os, uint64_t object, zfs_uio_t *uio, uint64_t size); int dmu_read_uio_dbuf(dmu_buf_t *zdb, zfs_uio_t *uio, uint64_t size); int dmu_read_uio_dnode(dnode_t *dn, zfs_uio_t *uio, uint64_t size); int dmu_write_uio(objset_t *os, uint64_t object, zfs_uio_t *uio, uint64_t size, dmu_tx_t *tx); int dmu_write_uio_dbuf(dmu_buf_t *zdb, zfs_uio_t *uio, uint64_t size, dmu_tx_t *tx); int dmu_write_uio_dnode(dnode_t *dn, zfs_uio_t *uio, uint64_t size, dmu_tx_t *tx); #endif struct arc_buf *dmu_request_arcbuf(dmu_buf_t *handle, int size); void dmu_return_arcbuf(struct arc_buf *buf); int dmu_assign_arcbuf_by_dnode(dnode_t *dn, uint64_t offset, struct arc_buf *buf, dmu_tx_t *tx); int dmu_assign_arcbuf_by_dbuf(dmu_buf_t *handle, uint64_t offset, struct arc_buf *buf, dmu_tx_t *tx); #define dmu_assign_arcbuf dmu_assign_arcbuf_by_dbuf extern uint_t zfs_max_recordsize; /* * Asynchronously try to read in the data. */ void dmu_prefetch(objset_t *os, uint64_t object, int64_t level, uint64_t offset, uint64_t len, enum zio_priority pri); typedef struct dmu_object_info { /* All sizes are in bytes unless otherwise indicated. */ uint32_t doi_data_block_size; uint32_t doi_metadata_block_size; dmu_object_type_t doi_type; dmu_object_type_t doi_bonus_type; uint64_t doi_bonus_size; uint8_t doi_indirection; /* 2 = dnode->indirect->data */ uint8_t doi_checksum; uint8_t doi_compress; uint8_t doi_nblkptr; uint8_t doi_pad[4]; uint64_t doi_dnodesize; uint64_t doi_physical_blocks_512; /* data + metadata, 512b blks */ uint64_t doi_max_offset; uint64_t doi_fill_count; /* number of non-empty blocks */ } dmu_object_info_t; typedef void (*const arc_byteswap_func_t)(void *buf, size_t size); typedef struct dmu_object_type_info { dmu_object_byteswap_t ot_byteswap; boolean_t ot_metadata; boolean_t ot_dbuf_metadata_cache; boolean_t ot_encrypt; const char *ot_name; } dmu_object_type_info_t; typedef const struct dmu_object_byteswap_info { arc_byteswap_func_t ob_func; const char *ob_name; } dmu_object_byteswap_info_t; extern const dmu_object_type_info_t dmu_ot[DMU_OT_NUMTYPES]; extern dmu_object_byteswap_info_t dmu_ot_byteswap[DMU_BSWAP_NUMFUNCS]; /* * Get information on a DMU object. * * Return 0 on success or ENOENT if object is not allocated. * * If doi is NULL, just indicates whether the object exists. */ int dmu_object_info(objset_t *os, uint64_t object, dmu_object_info_t *doi); void __dmu_object_info_from_dnode(struct dnode *dn, dmu_object_info_t *doi); /* Like dmu_object_info, but faster if you have a held dnode in hand. */ void dmu_object_info_from_dnode(dnode_t *dn, dmu_object_info_t *doi); /* Like dmu_object_info, but faster if you have a held dbuf in hand. */ void dmu_object_info_from_db(dmu_buf_t *db, dmu_object_info_t *doi); /* * Like dmu_object_info_from_db, but faster still when you only care about * the size. */ void dmu_object_size_from_db(dmu_buf_t *db, uint32_t *blksize, u_longlong_t *nblk512); void dmu_object_dnsize_from_db(dmu_buf_t *db, int *dnsize); typedef struct dmu_objset_stats { uint64_t dds_num_clones; /* number of clones of this */ uint64_t dds_creation_txg; uint64_t dds_guid; dmu_objset_type_t dds_type; uint8_t dds_is_snapshot; uint8_t dds_inconsistent; uint8_t dds_redacted; char dds_origin[ZFS_MAX_DATASET_NAME_LEN]; } dmu_objset_stats_t; /* * Get stats on a dataset. */ void dmu_objset_fast_stat(objset_t *os, dmu_objset_stats_t *stat); /* * Add entries to the nvlist for all the objset's properties. See * zfs_prop_table[] and zfs(1m) for details on the properties. */ void dmu_objset_stats(objset_t *os, struct nvlist *nv); /* * Get the space usage statistics for statvfs(). * * refdbytes is the amount of space "referenced" by this objset. * availbytes is the amount of space available to this objset, taking * into account quotas & reservations, assuming that no other objsets * use the space first. These values correspond to the 'referenced' and * 'available' properties, described in the zfs(1m) manpage. * * usedobjs and availobjs are the number of objects currently allocated, * and available. */ void dmu_objset_space(objset_t *os, uint64_t *refdbytesp, uint64_t *availbytesp, uint64_t *usedobjsp, uint64_t *availobjsp); /* * The fsid_guid is a 56-bit ID that can change to avoid collisions. * (Contrast with the ds_guid which is a 64-bit ID that will never * change, so there is a small probability that it will collide.) */ uint64_t dmu_objset_fsid_guid(objset_t *os); /* * Get the [cm]time for an objset's snapshot dir */ inode_timespec_t dmu_objset_snap_cmtime(objset_t *os); int dmu_objset_is_snapshot(objset_t *os); extern struct spa *dmu_objset_spa(objset_t *os); extern struct zilog *dmu_objset_zil(objset_t *os); extern struct dsl_pool *dmu_objset_pool(objset_t *os); extern struct dsl_dataset *dmu_objset_ds(objset_t *os); extern void dmu_objset_name(objset_t *os, char *buf); extern dmu_objset_type_t dmu_objset_type(objset_t *os); extern uint64_t dmu_objset_id(objset_t *os); extern uint64_t dmu_objset_dnodesize(objset_t *os); extern zfs_sync_type_t dmu_objset_syncprop(objset_t *os); extern zfs_logbias_op_t dmu_objset_logbias(objset_t *os); extern int dmu_objset_blksize(objset_t *os); extern int dmu_snapshot_list_next(objset_t *os, int namelen, char *name, uint64_t *id, uint64_t *offp, boolean_t *case_conflict); extern int dmu_snapshot_lookup(objset_t *os, const char *name, uint64_t *val); extern int dmu_snapshot_realname(objset_t *os, const char *name, char *real, int maxlen, boolean_t *conflict); extern int dmu_dir_list_next(objset_t *os, int namelen, char *name, uint64_t *idp, uint64_t *offp); typedef struct zfs_file_info { uint64_t zfi_user; uint64_t zfi_group; uint64_t zfi_project; uint64_t zfi_generation; } zfs_file_info_t; typedef int file_info_cb_t(dmu_object_type_t bonustype, const void *data, struct zfs_file_info *zoi); extern void dmu_objset_register_type(dmu_objset_type_t ost, file_info_cb_t *cb); extern void dmu_objset_set_user(objset_t *os, void *user_ptr); extern void *dmu_objset_get_user(objset_t *os); /* * Return the txg number for the given assigned transaction. */ uint64_t dmu_tx_get_txg(dmu_tx_t *tx); /* * Synchronous write. * If a parent zio is provided this function initiates a write on the * provided buffer as a child of the parent zio. * In the absence of a parent zio, the write is completed synchronously. * At write completion, blk is filled with the bp of the written block. * Note that while the data covered by this function will be on stable * storage when the write completes this new data does not become a * permanent part of the file until the associated transaction commits. */ /* * {zfs,zvol,ztest}_get_done() args */ typedef struct zgd { struct lwb *zgd_lwb; struct blkptr *zgd_bp; dmu_buf_t *zgd_db; struct zfs_locked_range *zgd_lr; void *zgd_private; } zgd_t; typedef void dmu_sync_cb_t(zgd_t *arg, int error); int dmu_sync(struct zio *zio, uint64_t txg, dmu_sync_cb_t *done, zgd_t *zgd); /* * Find the next hole or data block in file starting at *off * Return found offset in *off. Return ESRCH for end of file. */ int dmu_offset_next(objset_t *os, uint64_t object, boolean_t hole, uint64_t *off); int dmu_read_l0_bps(objset_t *os, uint64_t object, uint64_t offset, uint64_t length, struct blkptr *bps, size_t *nbpsp); int dmu_brt_clone(objset_t *os, uint64_t object, uint64_t offset, uint64_t length, dmu_tx_t *tx, const struct blkptr *bps, size_t nbps, boolean_t replay); /* * Initial setup and final teardown. */ extern void dmu_init(void); extern void dmu_fini(void); typedef void (*dmu_traverse_cb_t)(objset_t *os, void *arg, struct blkptr *bp, uint64_t object, uint64_t offset, int len); void dmu_traverse_objset(objset_t *os, uint64_t txg_start, dmu_traverse_cb_t cb, void *arg); int dmu_diff(const char *tosnap_name, const char *fromsnap_name, zfs_file_t *fp, offset_t *offp); /* CRC64 table */ #define ZFS_CRC64_POLY 0xC96C5795D7870F42ULL /* ECMA-182, reflected form */ extern uint64_t zfs_crc64_table[256]; extern uint_t dmu_prefetch_max; #ifdef __cplusplus } #endif #endif /* _SYS_DMU_H */ diff --git a/sys/contrib/openzfs/include/sys/dsl_scan.h b/sys/contrib/openzfs/include/sys/dsl_scan.h index 8925b5815a37..6753b4a8f359 100644 --- a/sys/contrib/openzfs/include/sys/dsl_scan.h +++ b/sys/contrib/openzfs/include/sys/dsl_scan.h @@ -1,195 +1,220 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2012, 2017 by Delphix. All rights reserved. * Copyright (c) 2017, 2019, Datto Inc. All rights reserved. */ #ifndef _SYS_DSL_SCAN_H #define _SYS_DSL_SCAN_H #include #include +#include #include #include #ifdef __cplusplus extern "C" { #endif struct objset; struct dsl_dir; struct dsl_dataset; struct dsl_pool; struct dmu_tx; extern int zfs_scan_suspend_progress; /* * All members of this structure must be uint64_t, for byteswap * purposes. */ typedef struct dsl_scan_phys { uint64_t scn_func; /* pool_scan_func_t */ uint64_t scn_state; /* dsl_scan_state_t */ uint64_t scn_queue_obj; uint64_t scn_min_txg; uint64_t scn_max_txg; uint64_t scn_cur_min_txg; uint64_t scn_cur_max_txg; uint64_t scn_start_time; uint64_t scn_end_time; uint64_t scn_to_examine; /* total bytes to be scanned */ uint64_t scn_examined; /* bytes scanned so far */ uint64_t scn_to_process; uint64_t scn_processed; uint64_t scn_errors; /* scan I/O error count */ uint64_t scn_ddt_class_max; ddt_bookmark_t scn_ddt_bookmark; zbookmark_phys_t scn_bookmark; uint64_t scn_flags; /* dsl_scan_flags_t */ } dsl_scan_phys_t; #define SCAN_PHYS_NUMINTS (sizeof (dsl_scan_phys_t) / sizeof (uint64_t)) typedef enum dsl_scan_flags { DSF_VISIT_DS_AGAIN = 1<<0, DSF_SCRUB_PAUSED = 1<<1, } dsl_scan_flags_t; #define DSL_SCAN_FLAGS_MASK (DSF_VISIT_DS_AGAIN) +typedef struct dsl_errorscrub_phys { + uint64_t dep_func; /* pool_scan_func_t */ + uint64_t dep_state; /* dsl_scan_state_t */ + uint64_t dep_cursor; /* serialized zap cursor for tracing progress */ + uint64_t dep_start_time; /* error scrub start time, unix timestamp */ + uint64_t dep_end_time; /* error scrub end time, unix timestamp */ + uint64_t dep_to_examine; /* total error blocks to be scrubbed */ + uint64_t dep_examined; /* blocks scrubbed so far */ + uint64_t dep_errors; /* error scrub I/O error count */ + uint64_t dep_paused_flags; /* flag for paused */ +} dsl_errorscrub_phys_t; + +#define ERRORSCRUB_PHYS_NUMINTS (sizeof (dsl_errorscrub_phys_t) \ + / sizeof (uint64_t)) + /* * Every pool will have one dsl_scan_t and this structure will contain * in-memory information about the scan and a pointer to the on-disk * representation (i.e. dsl_scan_phys_t). Most of the state of the scan * is contained on-disk to allow the scan to resume in the event of a reboot * or panic. This structure maintains information about the behavior of a * running scan, some caching information, and how it should traverse the pool. * * The following members of this structure direct the behavior of the scan: * * scn_suspending - a scan that cannot be completed in a single txg or * has exceeded its allotted time will need to suspend. * When this flag is set the scanner will stop traversing * the pool and write out the current state to disk. * * scn_restart_txg - directs the scanner to either restart or start a * a scan at the specified txg value. * * scn_done_txg - when a scan completes its traversal it will set * the completion txg to the next txg. This is necessary * to ensure that any blocks that were freed during * the scan but have not yet been processed (i.e deferred * frees) are accounted for. * * This structure also maintains information about deferred frees which are * a special kind of traversal. Deferred free can exist in either a bptree or * a bpobj structure. The scn_is_bptree flag will indicate the type of * deferred free that is in progress. If the deferred free is part of an * asynchronous destroy then the scn_async_destroying flag will be set. */ typedef struct dsl_scan { struct dsl_pool *scn_dp; uint64_t scn_restart_txg; uint64_t scn_done_txg; uint64_t scn_sync_start_time; uint64_t scn_issued_before_pass; /* for freeing blocks */ boolean_t scn_is_bptree; boolean_t scn_async_destroying; boolean_t scn_async_stalled; uint64_t scn_async_block_min_time_ms; /* flags and stats for controlling scan state */ boolean_t scn_is_sorted; /* doing sequential scan */ boolean_t scn_clearing; /* scan is issuing sequential extents */ boolean_t scn_checkpointing; /* scan is issuing all queued extents */ boolean_t scn_suspending; /* scan is suspending until next txg */ uint64_t scn_last_checkpoint; /* time of last checkpoint */ /* members for thread synchronization */ zio_t *scn_zio_root; /* root zio for waiting on IO */ taskq_t *scn_taskq; /* task queue for issuing extents */ /* for controlling scan prefetch, protected by spa_scrub_lock */ boolean_t scn_prefetch_stop; /* prefetch should stop */ zbookmark_phys_t scn_prefetch_bookmark; /* prefetch start bookmark */ avl_tree_t scn_prefetch_queue; /* priority queue of prefetch IOs */ uint64_t scn_maxinflight_bytes; /* max bytes in flight for pool */ /* per txg statistics */ uint64_t scn_visited_this_txg; /* total bps visited this txg */ uint64_t scn_dedup_frees_this_txg; /* dedup bps freed this txg */ uint64_t scn_holes_this_txg; uint64_t scn_lt_min_this_txg; uint64_t scn_gt_max_this_txg; uint64_t scn_ddt_contained_this_txg; uint64_t scn_objsets_visited_this_txg; uint64_t scn_avg_seg_size_this_txg; uint64_t scn_segs_this_txg; uint64_t scn_avg_zio_size_this_txg; uint64_t scn_zios_this_txg; + /* zap cursor for tracing error scrub progress */ + zap_cursor_t errorscrub_cursor; /* members needed for syncing scan status to disk */ dsl_scan_phys_t scn_phys; /* on disk representation of scan */ dsl_scan_phys_t scn_phys_cached; avl_tree_t scn_queue; /* queue of datasets to scan */ uint64_t scn_queues_pending; /* outstanding data to issue */ + /* members needed for syncing error scrub status to disk */ + dsl_errorscrub_phys_t errorscrub_phys; } dsl_scan_t; typedef struct dsl_scan_io_queue dsl_scan_io_queue_t; void scan_init(void); void scan_fini(void); int dsl_scan_init(struct dsl_pool *dp, uint64_t txg); int dsl_scan_setup_check(void *, dmu_tx_t *); void dsl_scan_setup_sync(void *, dmu_tx_t *); void dsl_scan_fini(struct dsl_pool *dp); void dsl_scan_sync(struct dsl_pool *, dmu_tx_t *); int dsl_scan_cancel(struct dsl_pool *); int dsl_scan(struct dsl_pool *, pool_scan_func_t); void dsl_scan_assess_vdev(struct dsl_pool *dp, vdev_t *vd); boolean_t dsl_scan_scrubbing(const struct dsl_pool *dp); -int dsl_scrub_set_pause_resume(const struct dsl_pool *dp, pool_scrub_cmd_t cmd); +boolean_t dsl_errorscrubbing(const struct dsl_pool *dp); +boolean_t dsl_errorscrub_active(dsl_scan_t *scn); void dsl_scan_restart_resilver(struct dsl_pool *, uint64_t txg); +int dsl_scrub_set_pause_resume(const struct dsl_pool *dp, + pool_scrub_cmd_t cmd); +void dsl_errorscrub_sync(struct dsl_pool *, dmu_tx_t *); boolean_t dsl_scan_resilvering(struct dsl_pool *dp); boolean_t dsl_scan_resilver_scheduled(struct dsl_pool *dp); boolean_t dsl_dataset_unstable(struct dsl_dataset *ds); void dsl_scan_ddt_entry(dsl_scan_t *scn, enum zio_checksum checksum, ddt_entry_t *dde, dmu_tx_t *tx); void dsl_scan_ds_destroyed(struct dsl_dataset *ds, struct dmu_tx *tx); void dsl_scan_ds_snapshotted(struct dsl_dataset *ds, struct dmu_tx *tx); void dsl_scan_ds_clone_swapped(struct dsl_dataset *ds1, struct dsl_dataset *ds2, struct dmu_tx *tx); boolean_t dsl_scan_active(dsl_scan_t *scn); boolean_t dsl_scan_is_paused_scrub(const dsl_scan_t *scn); +boolean_t dsl_errorscrub_is_paused(const dsl_scan_t *scn); void dsl_scan_freed(spa_t *spa, const blkptr_t *bp); void dsl_scan_io_queue_destroy(dsl_scan_io_queue_t *queue); void dsl_scan_io_queue_vdev_xfer(vdev_t *svd, vdev_t *tvd); #ifdef __cplusplus } #endif #endif /* _SYS_DSL_SCAN_H */ diff --git a/sys/contrib/openzfs/include/sys/fs/zfs.h b/sys/contrib/openzfs/include/sys/fs/zfs.h index 0734ff12280e..93193fa142da 100644 --- a/sys/contrib/openzfs/include/sys/fs/zfs.h +++ b/sys/contrib/openzfs/include/sys/fs/zfs.h @@ -1,1841 +1,1859 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2011, 2020 by Delphix. All rights reserved. * Copyright 2011 Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2013, 2017 Joyent, Inc. All rights reserved. * Copyright (c) 2014 Integros [integros.com] * Copyright (c) 2017, Intel Corporation. * Copyright (c) 2019 Datto Inc. * Portions Copyright 2010 Robert Milkowski * Copyright (c) 2021, Colm Buckley * Copyright (c) 2022 Hewlett Packard Enterprise Development LP. */ #ifndef _SYS_FS_ZFS_H #define _SYS_FS_ZFS_H extern __attribute__((visibility("default"))) #include #include #ifdef __cplusplus extern "C" { #endif /* * Types and constants shared between userland and the kernel. */ /* * Each dataset can be one of the following types. These constants can be * combined into masks that can be passed to various functions. */ typedef enum { ZFS_TYPE_INVALID = 0, ZFS_TYPE_FILESYSTEM = (1 << 0), ZFS_TYPE_SNAPSHOT = (1 << 1), ZFS_TYPE_VOLUME = (1 << 2), ZFS_TYPE_POOL = (1 << 3), ZFS_TYPE_BOOKMARK = (1 << 4), ZFS_TYPE_VDEV = (1 << 5), } zfs_type_t; /* * NB: lzc_dataset_type should be updated whenever a new objset type is added, * if it represents a real type of a dataset that can be created from userland. */ typedef enum dmu_objset_type { DMU_OST_NONE, DMU_OST_META, DMU_OST_ZFS, DMU_OST_ZVOL, DMU_OST_OTHER, /* For testing only! */ DMU_OST_ANY, /* Be careful! */ DMU_OST_NUMTYPES } dmu_objset_type_t; #define ZFS_TYPE_DATASET \ (ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME | ZFS_TYPE_SNAPSHOT) /* * All of these include the terminating NUL byte. */ #define ZAP_MAXNAMELEN 256 #define ZAP_MAXVALUELEN (1024 * 8) #define ZAP_OLDMAXVALUELEN 1024 #define ZFS_MAX_DATASET_NAME_LEN 256 /* * Dataset properties are identified by these constants and must be added to * the end of this list to ensure that external consumers are not affected * by the change. If you make any changes to this list, be sure to update * the property table in module/zcommon/zfs_prop.c. */ typedef enum { ZPROP_CONT = -2, ZPROP_INVAL = -1, ZPROP_USERPROP = ZPROP_INVAL, ZFS_PROP_TYPE = 0, ZFS_PROP_CREATION, ZFS_PROP_USED, ZFS_PROP_AVAILABLE, ZFS_PROP_REFERENCED, ZFS_PROP_COMPRESSRATIO, ZFS_PROP_MOUNTED, ZFS_PROP_ORIGIN, ZFS_PROP_QUOTA, ZFS_PROP_RESERVATION, ZFS_PROP_VOLSIZE, ZFS_PROP_VOLBLOCKSIZE, ZFS_PROP_RECORDSIZE, ZFS_PROP_MOUNTPOINT, ZFS_PROP_SHARENFS, ZFS_PROP_CHECKSUM, ZFS_PROP_COMPRESSION, ZFS_PROP_ATIME, ZFS_PROP_DEVICES, ZFS_PROP_EXEC, ZFS_PROP_SETUID, ZFS_PROP_READONLY, ZFS_PROP_ZONED, ZFS_PROP_SNAPDIR, ZFS_PROP_ACLMODE, ZFS_PROP_ACLINHERIT, ZFS_PROP_CREATETXG, ZFS_PROP_NAME, /* not exposed to the user */ ZFS_PROP_CANMOUNT, ZFS_PROP_ISCSIOPTIONS, /* not exposed to the user */ ZFS_PROP_XATTR, ZFS_PROP_NUMCLONES, /* not exposed to the user */ ZFS_PROP_COPIES, ZFS_PROP_VERSION, ZFS_PROP_UTF8ONLY, ZFS_PROP_NORMALIZE, ZFS_PROP_CASE, ZFS_PROP_VSCAN, ZFS_PROP_NBMAND, ZFS_PROP_SHARESMB, ZFS_PROP_REFQUOTA, ZFS_PROP_REFRESERVATION, ZFS_PROP_GUID, ZFS_PROP_PRIMARYCACHE, ZFS_PROP_SECONDARYCACHE, ZFS_PROP_USEDSNAP, ZFS_PROP_USEDDS, ZFS_PROP_USEDCHILD, ZFS_PROP_USEDREFRESERV, ZFS_PROP_USERACCOUNTING, /* not exposed to the user */ ZFS_PROP_STMF_SHAREINFO, /* not exposed to the user */ ZFS_PROP_DEFER_DESTROY, ZFS_PROP_USERREFS, ZFS_PROP_LOGBIAS, ZFS_PROP_UNIQUE, /* not exposed to the user */ ZFS_PROP_OBJSETID, ZFS_PROP_DEDUP, ZFS_PROP_MLSLABEL, ZFS_PROP_SYNC, ZFS_PROP_DNODESIZE, ZFS_PROP_REFRATIO, ZFS_PROP_WRITTEN, ZFS_PROP_CLONES, ZFS_PROP_LOGICALUSED, ZFS_PROP_LOGICALREFERENCED, ZFS_PROP_INCONSISTENT, /* not exposed to the user */ ZFS_PROP_VOLMODE, ZFS_PROP_FILESYSTEM_LIMIT, ZFS_PROP_SNAPSHOT_LIMIT, ZFS_PROP_FILESYSTEM_COUNT, ZFS_PROP_SNAPSHOT_COUNT, ZFS_PROP_SNAPDEV, ZFS_PROP_ACLTYPE, ZFS_PROP_SELINUX_CONTEXT, ZFS_PROP_SELINUX_FSCONTEXT, ZFS_PROP_SELINUX_DEFCONTEXT, ZFS_PROP_SELINUX_ROOTCONTEXT, ZFS_PROP_RELATIME, ZFS_PROP_REDUNDANT_METADATA, ZFS_PROP_OVERLAY, ZFS_PROP_PREV_SNAP, ZFS_PROP_RECEIVE_RESUME_TOKEN, ZFS_PROP_ENCRYPTION, ZFS_PROP_KEYLOCATION, ZFS_PROP_KEYFORMAT, ZFS_PROP_PBKDF2_SALT, ZFS_PROP_PBKDF2_ITERS, ZFS_PROP_ENCRYPTION_ROOT, ZFS_PROP_KEY_GUID, ZFS_PROP_KEYSTATUS, ZFS_PROP_REMAPTXG, /* obsolete - no longer used */ ZFS_PROP_SPECIAL_SMALL_BLOCKS, ZFS_PROP_IVSET_GUID, /* not exposed to the user */ ZFS_PROP_REDACTED, ZFS_PROP_REDACT_SNAPS, ZFS_PROP_SNAPSHOTS_CHANGED, ZFS_NUM_PROPS } zfs_prop_t; typedef enum { ZFS_PROP_USERUSED, ZFS_PROP_USERQUOTA, ZFS_PROP_GROUPUSED, ZFS_PROP_GROUPQUOTA, ZFS_PROP_USEROBJUSED, ZFS_PROP_USEROBJQUOTA, ZFS_PROP_GROUPOBJUSED, ZFS_PROP_GROUPOBJQUOTA, ZFS_PROP_PROJECTUSED, ZFS_PROP_PROJECTQUOTA, ZFS_PROP_PROJECTOBJUSED, ZFS_PROP_PROJECTOBJQUOTA, ZFS_NUM_USERQUOTA_PROPS } zfs_userquota_prop_t; _SYS_FS_ZFS_H const char *const zfs_userquota_prop_prefixes[ ZFS_NUM_USERQUOTA_PROPS]; /* * Pool properties are identified by these constants and must be added to the * end of this list to ensure that external consumers are not affected * by the change. Properties must be registered in zfs_prop_init(). */ typedef enum { ZPOOL_PROP_INVAL = -1, ZPOOL_PROP_NAME, ZPOOL_PROP_SIZE, ZPOOL_PROP_CAPACITY, ZPOOL_PROP_ALTROOT, ZPOOL_PROP_HEALTH, ZPOOL_PROP_GUID, ZPOOL_PROP_VERSION, ZPOOL_PROP_BOOTFS, ZPOOL_PROP_DELEGATION, ZPOOL_PROP_AUTOREPLACE, ZPOOL_PROP_CACHEFILE, ZPOOL_PROP_FAILUREMODE, ZPOOL_PROP_LISTSNAPS, ZPOOL_PROP_AUTOEXPAND, ZPOOL_PROP_DEDUPDITTO, ZPOOL_PROP_DEDUPRATIO, ZPOOL_PROP_FREE, ZPOOL_PROP_ALLOCATED, ZPOOL_PROP_READONLY, ZPOOL_PROP_ASHIFT, ZPOOL_PROP_COMMENT, ZPOOL_PROP_EXPANDSZ, ZPOOL_PROP_FREEING, ZPOOL_PROP_FRAGMENTATION, ZPOOL_PROP_LEAKED, ZPOOL_PROP_MAXBLOCKSIZE, ZPOOL_PROP_TNAME, ZPOOL_PROP_MAXDNODESIZE, ZPOOL_PROP_MULTIHOST, ZPOOL_PROP_CHECKPOINT, ZPOOL_PROP_LOAD_GUID, ZPOOL_PROP_AUTOTRIM, ZPOOL_PROP_COMPATIBILITY, ZPOOL_PROP_BCLONEUSED, ZPOOL_PROP_BCLONESAVED, ZPOOL_PROP_BCLONERATIO, ZPOOL_NUM_PROPS } zpool_prop_t; /* Small enough to not hog a whole line of printout in zpool(8). */ #define ZPROP_MAX_COMMENT 32 #define ZPROP_BOOLEAN_NA 2 #define ZPROP_VALUE "value" #define ZPROP_SOURCE "source" typedef enum { ZPROP_SRC_NONE = 0x1, ZPROP_SRC_DEFAULT = 0x2, ZPROP_SRC_TEMPORARY = 0x4, ZPROP_SRC_LOCAL = 0x8, ZPROP_SRC_INHERITED = 0x10, ZPROP_SRC_RECEIVED = 0x20 } zprop_source_t; #define ZPROP_SRC_ALL 0x3f #define ZPROP_SOURCE_VAL_RECVD "$recvd" #define ZPROP_N_MORE_ERRORS "N_MORE_ERRORS" /* * Dataset flag implemented as a special entry in the props zap object * indicating that the dataset has received properties on or after * SPA_VERSION_RECVD_PROPS. The first such receive blows away local properties * just as it did in earlier versions, and thereafter, local properties are * preserved. */ #define ZPROP_HAS_RECVD "$hasrecvd" typedef enum { ZPROP_ERR_NOCLEAR = 0x1, /* failure to clear existing props */ ZPROP_ERR_NORESTORE = 0x2 /* failure to restore props on error */ } zprop_errflags_t; typedef int (*zprop_func)(int, void *); /* * Properties to be set on the root file system of a new pool * are stuffed into their own nvlist, which is then included in * the properties nvlist with the pool properties. */ #define ZPOOL_ROOTFS_PROPS "root-props-nvl" /* * Length of 'written@' and 'written#' */ #define ZFS_WRITTEN_PROP_PREFIX_LEN 8 /* * VDEV properties are identified by these constants and must be added to the * end of this list to ensure that external consumers are not affected * by the change. If you make any changes to this list, be sure to update * the property table in usr/src/common/zfs/zpool_prop.c. */ typedef enum { VDEV_PROP_INVAL = -1, VDEV_PROP_USERPROP = VDEV_PROP_INVAL, VDEV_PROP_NAME, VDEV_PROP_CAPACITY, VDEV_PROP_STATE, VDEV_PROP_GUID, VDEV_PROP_ASIZE, VDEV_PROP_PSIZE, VDEV_PROP_ASHIFT, VDEV_PROP_SIZE, VDEV_PROP_FREE, VDEV_PROP_ALLOCATED, VDEV_PROP_COMMENT, VDEV_PROP_EXPANDSZ, VDEV_PROP_FRAGMENTATION, VDEV_PROP_BOOTSIZE, VDEV_PROP_PARITY, VDEV_PROP_PATH, VDEV_PROP_DEVID, VDEV_PROP_PHYS_PATH, VDEV_PROP_ENC_PATH, VDEV_PROP_FRU, VDEV_PROP_PARENT, VDEV_PROP_CHILDREN, VDEV_PROP_NUMCHILDREN, VDEV_PROP_READ_ERRORS, VDEV_PROP_WRITE_ERRORS, VDEV_PROP_CHECKSUM_ERRORS, VDEV_PROP_INITIALIZE_ERRORS, VDEV_PROP_OPS_NULL, VDEV_PROP_OPS_READ, VDEV_PROP_OPS_WRITE, VDEV_PROP_OPS_FREE, VDEV_PROP_OPS_CLAIM, VDEV_PROP_OPS_TRIM, VDEV_PROP_BYTES_NULL, VDEV_PROP_BYTES_READ, VDEV_PROP_BYTES_WRITE, VDEV_PROP_BYTES_FREE, VDEV_PROP_BYTES_CLAIM, VDEV_PROP_BYTES_TRIM, VDEV_PROP_REMOVING, VDEV_PROP_ALLOCATING, VDEV_PROP_FAILFAST, VDEV_PROP_CHECKSUM_N, VDEV_PROP_CHECKSUM_T, VDEV_PROP_IO_N, VDEV_PROP_IO_T, VDEV_NUM_PROPS } vdev_prop_t; /* * Dataset property functions shared between libzfs and kernel. */ _SYS_FS_ZFS_H const char *zfs_prop_default_string(zfs_prop_t); _SYS_FS_ZFS_H uint64_t zfs_prop_default_numeric(zfs_prop_t); _SYS_FS_ZFS_H boolean_t zfs_prop_readonly(zfs_prop_t); _SYS_FS_ZFS_H boolean_t zfs_prop_visible(zfs_prop_t prop); _SYS_FS_ZFS_H boolean_t zfs_prop_inheritable(zfs_prop_t); _SYS_FS_ZFS_H boolean_t zfs_prop_setonce(zfs_prop_t); _SYS_FS_ZFS_H boolean_t zfs_prop_encryption_key_param(zfs_prop_t); _SYS_FS_ZFS_H boolean_t zfs_prop_valid_keylocation(const char *, boolean_t); _SYS_FS_ZFS_H const char *zfs_prop_to_name(zfs_prop_t); _SYS_FS_ZFS_H zfs_prop_t zfs_name_to_prop(const char *); _SYS_FS_ZFS_H boolean_t zfs_prop_user(const char *); _SYS_FS_ZFS_H boolean_t zfs_prop_userquota(const char *); _SYS_FS_ZFS_H boolean_t zfs_prop_written(const char *); _SYS_FS_ZFS_H int zfs_prop_index_to_string(zfs_prop_t, uint64_t, const char **); _SYS_FS_ZFS_H int zfs_prop_string_to_index(zfs_prop_t, const char *, uint64_t *); _SYS_FS_ZFS_H uint64_t zfs_prop_random_value(zfs_prop_t, uint64_t seed); _SYS_FS_ZFS_H boolean_t zfs_prop_valid_for_type(int, zfs_type_t, boolean_t); /* * Pool property functions shared between libzfs and kernel. */ _SYS_FS_ZFS_H zpool_prop_t zpool_name_to_prop(const char *); _SYS_FS_ZFS_H const char *zpool_prop_to_name(zpool_prop_t); _SYS_FS_ZFS_H const char *zpool_prop_default_string(zpool_prop_t); _SYS_FS_ZFS_H uint64_t zpool_prop_default_numeric(zpool_prop_t); _SYS_FS_ZFS_H boolean_t zpool_prop_readonly(zpool_prop_t); _SYS_FS_ZFS_H boolean_t zpool_prop_setonce(zpool_prop_t); _SYS_FS_ZFS_H boolean_t zpool_prop_feature(const char *); _SYS_FS_ZFS_H boolean_t zpool_prop_unsupported(const char *); _SYS_FS_ZFS_H int zpool_prop_index_to_string(zpool_prop_t, uint64_t, const char **); _SYS_FS_ZFS_H int zpool_prop_string_to_index(zpool_prop_t, const char *, uint64_t *); _SYS_FS_ZFS_H uint64_t zpool_prop_random_value(zpool_prop_t, uint64_t seed); /* * VDEV property functions shared between libzfs and kernel. */ _SYS_FS_ZFS_H vdev_prop_t vdev_name_to_prop(const char *); _SYS_FS_ZFS_H boolean_t vdev_prop_user(const char *name); _SYS_FS_ZFS_H const char *vdev_prop_to_name(vdev_prop_t); _SYS_FS_ZFS_H const char *vdev_prop_default_string(vdev_prop_t); _SYS_FS_ZFS_H uint64_t vdev_prop_default_numeric(vdev_prop_t); _SYS_FS_ZFS_H boolean_t vdev_prop_readonly(vdev_prop_t prop); _SYS_FS_ZFS_H int vdev_prop_index_to_string(vdev_prop_t, uint64_t, const char **); _SYS_FS_ZFS_H int vdev_prop_string_to_index(vdev_prop_t, const char *, uint64_t *); _SYS_FS_ZFS_H boolean_t zpool_prop_vdev(const char *name); _SYS_FS_ZFS_H uint64_t vdev_prop_random_value(vdev_prop_t prop, uint64_t seed); /* * Definitions for the Delegation. */ typedef enum { ZFS_DELEG_WHO_UNKNOWN = 0, ZFS_DELEG_USER = 'u', ZFS_DELEG_USER_SETS = 'U', ZFS_DELEG_GROUP = 'g', ZFS_DELEG_GROUP_SETS = 'G', ZFS_DELEG_EVERYONE = 'e', ZFS_DELEG_EVERYONE_SETS = 'E', ZFS_DELEG_CREATE = 'c', ZFS_DELEG_CREATE_SETS = 'C', ZFS_DELEG_NAMED_SET = 's', ZFS_DELEG_NAMED_SET_SETS = 'S' } zfs_deleg_who_type_t; typedef enum { ZFS_DELEG_NONE = 0, ZFS_DELEG_PERM_LOCAL = 1, ZFS_DELEG_PERM_DESCENDENT = 2, ZFS_DELEG_PERM_LOCALDESCENDENT = 3, ZFS_DELEG_PERM_CREATE = 4 } zfs_deleg_inherit_t; #define ZFS_DELEG_PERM_UID "uid" #define ZFS_DELEG_PERM_GID "gid" #define ZFS_DELEG_PERM_GROUPS "groups" #define ZFS_MLSLABEL_DEFAULT "none" #define ZFS_SMB_ACL_SRC "src" #define ZFS_SMB_ACL_TARGET "target" typedef enum { ZFS_CANMOUNT_OFF = 0, ZFS_CANMOUNT_ON = 1, ZFS_CANMOUNT_NOAUTO = 2 } zfs_canmount_type_t; typedef enum { ZFS_LOGBIAS_LATENCY = 0, ZFS_LOGBIAS_THROUGHPUT = 1 } zfs_logbias_op_t; typedef enum zfs_share_op { ZFS_SHARE_NFS = 0, ZFS_UNSHARE_NFS = 1, ZFS_SHARE_SMB = 2, ZFS_UNSHARE_SMB = 3 } zfs_share_op_t; typedef enum zfs_smb_acl_op { ZFS_SMB_ACL_ADD, ZFS_SMB_ACL_REMOVE, ZFS_SMB_ACL_RENAME, ZFS_SMB_ACL_PURGE } zfs_smb_acl_op_t; typedef enum zfs_cache_type { ZFS_CACHE_NONE = 0, ZFS_CACHE_METADATA = 1, ZFS_CACHE_ALL = 2 } zfs_cache_type_t; typedef enum { ZFS_SYNC_STANDARD = 0, ZFS_SYNC_ALWAYS = 1, ZFS_SYNC_DISABLED = 2 } zfs_sync_type_t; typedef enum { ZFS_XATTR_OFF = 0, ZFS_XATTR_DIR = 1, ZFS_XATTR_SA = 2 } zfs_xattr_type_t; typedef enum { ZFS_DNSIZE_LEGACY = 0, ZFS_DNSIZE_AUTO = 1, ZFS_DNSIZE_1K = 1024, ZFS_DNSIZE_2K = 2048, ZFS_DNSIZE_4K = 4096, ZFS_DNSIZE_8K = 8192, ZFS_DNSIZE_16K = 16384 } zfs_dnsize_type_t; typedef enum { ZFS_REDUNDANT_METADATA_ALL, ZFS_REDUNDANT_METADATA_MOST, ZFS_REDUNDANT_METADATA_SOME, ZFS_REDUNDANT_METADATA_NONE } zfs_redundant_metadata_type_t; typedef enum { ZFS_VOLMODE_DEFAULT = 0, ZFS_VOLMODE_GEOM = 1, ZFS_VOLMODE_DEV = 2, ZFS_VOLMODE_NONE = 3 } zfs_volmode_t; typedef enum zfs_keystatus { ZFS_KEYSTATUS_NONE = 0, ZFS_KEYSTATUS_UNAVAILABLE, ZFS_KEYSTATUS_AVAILABLE, } zfs_keystatus_t; typedef enum zfs_keyformat { ZFS_KEYFORMAT_NONE = 0, ZFS_KEYFORMAT_RAW, ZFS_KEYFORMAT_HEX, ZFS_KEYFORMAT_PASSPHRASE, ZFS_KEYFORMAT_FORMATS } zfs_keyformat_t; typedef enum zfs_key_location { ZFS_KEYLOCATION_NONE = 0, ZFS_KEYLOCATION_PROMPT, ZFS_KEYLOCATION_URI, ZFS_KEYLOCATION_LOCATIONS } zfs_keylocation_t; #define DEFAULT_PBKDF2_ITERATIONS 350000 #define MIN_PBKDF2_ITERATIONS 100000 /* * On-disk version number. */ #define SPA_VERSION_1 1ULL #define SPA_VERSION_2 2ULL #define SPA_VERSION_3 3ULL #define SPA_VERSION_4 4ULL #define SPA_VERSION_5 5ULL #define SPA_VERSION_6 6ULL #define SPA_VERSION_7 7ULL #define SPA_VERSION_8 8ULL #define SPA_VERSION_9 9ULL #define SPA_VERSION_10 10ULL #define SPA_VERSION_11 11ULL #define SPA_VERSION_12 12ULL #define SPA_VERSION_13 13ULL #define SPA_VERSION_14 14ULL #define SPA_VERSION_15 15ULL #define SPA_VERSION_16 16ULL #define SPA_VERSION_17 17ULL #define SPA_VERSION_18 18ULL #define SPA_VERSION_19 19ULL #define SPA_VERSION_20 20ULL #define SPA_VERSION_21 21ULL #define SPA_VERSION_22 22ULL #define SPA_VERSION_23 23ULL #define SPA_VERSION_24 24ULL #define SPA_VERSION_25 25ULL #define SPA_VERSION_26 26ULL #define SPA_VERSION_27 27ULL #define SPA_VERSION_28 28ULL #define SPA_VERSION_5000 5000ULL /* * The incrementing pool version number has been replaced by pool feature * flags. For more details, see zfeature.c. */ #define SPA_VERSION SPA_VERSION_5000 #define SPA_VERSION_STRING "5000" /* * Symbolic names for the changes that caused a SPA_VERSION switch. * Used in the code when checking for presence or absence of a feature. * Feel free to define multiple symbolic names for each version if there * were multiple changes to on-disk structures during that version. * * NOTE: When checking the current SPA_VERSION in your code, be sure * to use spa_version() since it reports the version of the * last synced uberblock. Checking the in-flight version can * be dangerous in some cases. */ #define SPA_VERSION_INITIAL SPA_VERSION_1 #define SPA_VERSION_DITTO_BLOCKS SPA_VERSION_2 #define SPA_VERSION_SPARES SPA_VERSION_3 #define SPA_VERSION_RAIDZ2 SPA_VERSION_3 #define SPA_VERSION_BPOBJ_ACCOUNT SPA_VERSION_3 #define SPA_VERSION_RAIDZ_DEFLATE SPA_VERSION_3 #define SPA_VERSION_DNODE_BYTES SPA_VERSION_3 #define SPA_VERSION_ZPOOL_HISTORY SPA_VERSION_4 #define SPA_VERSION_GZIP_COMPRESSION SPA_VERSION_5 #define SPA_VERSION_BOOTFS SPA_VERSION_6 #define SPA_VERSION_SLOGS SPA_VERSION_7 #define SPA_VERSION_DELEGATED_PERMS SPA_VERSION_8 #define SPA_VERSION_FUID SPA_VERSION_9 #define SPA_VERSION_REFRESERVATION SPA_VERSION_9 #define SPA_VERSION_REFQUOTA SPA_VERSION_9 #define SPA_VERSION_UNIQUE_ACCURATE SPA_VERSION_9 #define SPA_VERSION_L2CACHE SPA_VERSION_10 #define SPA_VERSION_NEXT_CLONES SPA_VERSION_11 #define SPA_VERSION_ORIGIN SPA_VERSION_11 #define SPA_VERSION_DSL_SCRUB SPA_VERSION_11 #define SPA_VERSION_SNAP_PROPS SPA_VERSION_12 #define SPA_VERSION_USED_BREAKDOWN SPA_VERSION_13 #define SPA_VERSION_PASSTHROUGH_X SPA_VERSION_14 #define SPA_VERSION_USERSPACE SPA_VERSION_15 #define SPA_VERSION_STMF_PROP SPA_VERSION_16 #define SPA_VERSION_RAIDZ3 SPA_VERSION_17 #define SPA_VERSION_USERREFS SPA_VERSION_18 #define SPA_VERSION_HOLES SPA_VERSION_19 #define SPA_VERSION_ZLE_COMPRESSION SPA_VERSION_20 #define SPA_VERSION_DEDUP SPA_VERSION_21 #define SPA_VERSION_RECVD_PROPS SPA_VERSION_22 #define SPA_VERSION_SLIM_ZIL SPA_VERSION_23 #define SPA_VERSION_SA SPA_VERSION_24 #define SPA_VERSION_SCAN SPA_VERSION_25 #define SPA_VERSION_DIR_CLONES SPA_VERSION_26 #define SPA_VERSION_DEADLISTS SPA_VERSION_26 #define SPA_VERSION_FAST_SNAP SPA_VERSION_27 #define SPA_VERSION_MULTI_REPLACE SPA_VERSION_28 #define SPA_VERSION_BEFORE_FEATURES SPA_VERSION_28 #define SPA_VERSION_FEATURES SPA_VERSION_5000 #define SPA_VERSION_IS_SUPPORTED(v) \ (((v) >= SPA_VERSION_INITIAL && (v) <= SPA_VERSION_BEFORE_FEATURES) || \ ((v) >= SPA_VERSION_FEATURES && (v) <= SPA_VERSION)) /* * ZPL version - rev'd whenever an incompatible on-disk format change * occurs. This is independent of SPA/DMU/ZAP versioning. You must * also update the version_table[] and help message in zfs_prop.c. */ #define ZPL_VERSION_1 1ULL #define ZPL_VERSION_2 2ULL #define ZPL_VERSION_3 3ULL #define ZPL_VERSION_4 4ULL #define ZPL_VERSION_5 5ULL #define ZPL_VERSION ZPL_VERSION_5 #define ZPL_VERSION_STRING "5" #define ZPL_VERSION_INITIAL ZPL_VERSION_1 #define ZPL_VERSION_DIRENT_TYPE ZPL_VERSION_2 #define ZPL_VERSION_FUID ZPL_VERSION_3 #define ZPL_VERSION_NORMALIZATION ZPL_VERSION_3 #define ZPL_VERSION_SYSATTR ZPL_VERSION_3 #define ZPL_VERSION_USERSPACE ZPL_VERSION_4 #define ZPL_VERSION_SA ZPL_VERSION_5 /* Persistent L2ARC version */ #define L2ARC_PERSISTENT_VERSION_1 1ULL #define L2ARC_PERSISTENT_VERSION L2ARC_PERSISTENT_VERSION_1 #define L2ARC_PERSISTENT_VERSION_STRING "1" /* Rewind policy information */ #define ZPOOL_NO_REWIND 1 /* No policy - default behavior */ #define ZPOOL_NEVER_REWIND 2 /* Do not search for best txg or rewind */ #define ZPOOL_TRY_REWIND 4 /* Search for best txg, but do not rewind */ #define ZPOOL_DO_REWIND 8 /* Rewind to best txg w/in deferred frees */ #define ZPOOL_EXTREME_REWIND 16 /* Allow extreme measures to find best txg */ #define ZPOOL_REWIND_MASK 28 /* All the possible rewind bits */ #define ZPOOL_REWIND_POLICIES 31 /* All the possible policy bits */ typedef struct zpool_load_policy { uint32_t zlp_rewind; /* rewind policy requested */ uint64_t zlp_maxmeta; /* max acceptable meta-data errors */ uint64_t zlp_maxdata; /* max acceptable data errors */ uint64_t zlp_txg; /* specific txg to load */ } zpool_load_policy_t; /* * The following are configuration names used in the nvlist describing a pool's * configuration. New on-disk names should be prefixed with ":" * (e.g. "org.openzfs:") to avoid conflicting names being developed * independently. */ #define ZPOOL_CONFIG_VERSION "version" #define ZPOOL_CONFIG_POOL_NAME "name" #define ZPOOL_CONFIG_POOL_STATE "state" #define ZPOOL_CONFIG_POOL_TXG "txg" #define ZPOOL_CONFIG_POOL_GUID "pool_guid" #define ZPOOL_CONFIG_CREATE_TXG "create_txg" #define ZPOOL_CONFIG_TOP_GUID "top_guid" #define ZPOOL_CONFIG_VDEV_TREE "vdev_tree" #define ZPOOL_CONFIG_TYPE "type" #define ZPOOL_CONFIG_CHILDREN "children" #define ZPOOL_CONFIG_ID "id" #define ZPOOL_CONFIG_GUID "guid" #define ZPOOL_CONFIG_INDIRECT_OBJECT "com.delphix:indirect_object" #define ZPOOL_CONFIG_INDIRECT_BIRTHS "com.delphix:indirect_births" #define ZPOOL_CONFIG_PREV_INDIRECT_VDEV "com.delphix:prev_indirect_vdev" #define ZPOOL_CONFIG_PATH "path" #define ZPOOL_CONFIG_DEVID "devid" #define ZPOOL_CONFIG_SPARE_ID "spareid" #define ZPOOL_CONFIG_METASLAB_ARRAY "metaslab_array" #define ZPOOL_CONFIG_METASLAB_SHIFT "metaslab_shift" #define ZPOOL_CONFIG_ASHIFT "ashift" #define ZPOOL_CONFIG_ASIZE "asize" #define ZPOOL_CONFIG_DTL "DTL" #define ZPOOL_CONFIG_SCAN_STATS "scan_stats" /* not stored on disk */ #define ZPOOL_CONFIG_REMOVAL_STATS "removal_stats" /* not stored on disk */ #define ZPOOL_CONFIG_CHECKPOINT_STATS "checkpoint_stats" /* not on disk */ #define ZPOOL_CONFIG_VDEV_STATS "vdev_stats" /* not stored on disk */ #define ZPOOL_CONFIG_INDIRECT_SIZE "indirect_size" /* not stored on disk */ /* container nvlist of extended stats */ #define ZPOOL_CONFIG_VDEV_STATS_EX "vdev_stats_ex" /* Active queue read/write stats */ #define ZPOOL_CONFIG_VDEV_SYNC_R_ACTIVE_QUEUE "vdev_sync_r_active_queue" #define ZPOOL_CONFIG_VDEV_SYNC_W_ACTIVE_QUEUE "vdev_sync_w_active_queue" #define ZPOOL_CONFIG_VDEV_ASYNC_R_ACTIVE_QUEUE "vdev_async_r_active_queue" #define ZPOOL_CONFIG_VDEV_ASYNC_W_ACTIVE_QUEUE "vdev_async_w_active_queue" #define ZPOOL_CONFIG_VDEV_SCRUB_ACTIVE_QUEUE "vdev_async_scrub_active_queue" #define ZPOOL_CONFIG_VDEV_TRIM_ACTIVE_QUEUE "vdev_async_trim_active_queue" #define ZPOOL_CONFIG_VDEV_REBUILD_ACTIVE_QUEUE "vdev_rebuild_active_queue" /* Queue sizes */ #define ZPOOL_CONFIG_VDEV_SYNC_R_PEND_QUEUE "vdev_sync_r_pend_queue" #define ZPOOL_CONFIG_VDEV_SYNC_W_PEND_QUEUE "vdev_sync_w_pend_queue" #define ZPOOL_CONFIG_VDEV_ASYNC_R_PEND_QUEUE "vdev_async_r_pend_queue" #define ZPOOL_CONFIG_VDEV_ASYNC_W_PEND_QUEUE "vdev_async_w_pend_queue" #define ZPOOL_CONFIG_VDEV_SCRUB_PEND_QUEUE "vdev_async_scrub_pend_queue" #define ZPOOL_CONFIG_VDEV_TRIM_PEND_QUEUE "vdev_async_trim_pend_queue" #define ZPOOL_CONFIG_VDEV_REBUILD_PEND_QUEUE "vdev_rebuild_pend_queue" /* Latency read/write histogram stats */ #define ZPOOL_CONFIG_VDEV_TOT_R_LAT_HISTO "vdev_tot_r_lat_histo" #define ZPOOL_CONFIG_VDEV_TOT_W_LAT_HISTO "vdev_tot_w_lat_histo" #define ZPOOL_CONFIG_VDEV_DISK_R_LAT_HISTO "vdev_disk_r_lat_histo" #define ZPOOL_CONFIG_VDEV_DISK_W_LAT_HISTO "vdev_disk_w_lat_histo" #define ZPOOL_CONFIG_VDEV_SYNC_R_LAT_HISTO "vdev_sync_r_lat_histo" #define ZPOOL_CONFIG_VDEV_SYNC_W_LAT_HISTO "vdev_sync_w_lat_histo" #define ZPOOL_CONFIG_VDEV_ASYNC_R_LAT_HISTO "vdev_async_r_lat_histo" #define ZPOOL_CONFIG_VDEV_ASYNC_W_LAT_HISTO "vdev_async_w_lat_histo" #define ZPOOL_CONFIG_VDEV_SCRUB_LAT_HISTO "vdev_scrub_histo" #define ZPOOL_CONFIG_VDEV_TRIM_LAT_HISTO "vdev_trim_histo" #define ZPOOL_CONFIG_VDEV_REBUILD_LAT_HISTO "vdev_rebuild_histo" /* Request size histograms */ #define ZPOOL_CONFIG_VDEV_SYNC_IND_R_HISTO "vdev_sync_ind_r_histo" #define ZPOOL_CONFIG_VDEV_SYNC_IND_W_HISTO "vdev_sync_ind_w_histo" #define ZPOOL_CONFIG_VDEV_ASYNC_IND_R_HISTO "vdev_async_ind_r_histo" #define ZPOOL_CONFIG_VDEV_ASYNC_IND_W_HISTO "vdev_async_ind_w_histo" #define ZPOOL_CONFIG_VDEV_IND_SCRUB_HISTO "vdev_ind_scrub_histo" #define ZPOOL_CONFIG_VDEV_IND_TRIM_HISTO "vdev_ind_trim_histo" #define ZPOOL_CONFIG_VDEV_IND_REBUILD_HISTO "vdev_ind_rebuild_histo" #define ZPOOL_CONFIG_VDEV_SYNC_AGG_R_HISTO "vdev_sync_agg_r_histo" #define ZPOOL_CONFIG_VDEV_SYNC_AGG_W_HISTO "vdev_sync_agg_w_histo" #define ZPOOL_CONFIG_VDEV_ASYNC_AGG_R_HISTO "vdev_async_agg_r_histo" #define ZPOOL_CONFIG_VDEV_ASYNC_AGG_W_HISTO "vdev_async_agg_w_histo" #define ZPOOL_CONFIG_VDEV_AGG_SCRUB_HISTO "vdev_agg_scrub_histo" #define ZPOOL_CONFIG_VDEV_AGG_TRIM_HISTO "vdev_agg_trim_histo" #define ZPOOL_CONFIG_VDEV_AGG_REBUILD_HISTO "vdev_agg_rebuild_histo" /* Number of slow IOs */ #define ZPOOL_CONFIG_VDEV_SLOW_IOS "vdev_slow_ios" /* vdev enclosure sysfs path */ #define ZPOOL_CONFIG_VDEV_ENC_SYSFS_PATH "vdev_enc_sysfs_path" #define ZPOOL_CONFIG_WHOLE_DISK "whole_disk" #define ZPOOL_CONFIG_ERRCOUNT "error_count" #define ZPOOL_CONFIG_NOT_PRESENT "not_present" #define ZPOOL_CONFIG_SPARES "spares" #define ZPOOL_CONFIG_IS_SPARE "is_spare" #define ZPOOL_CONFIG_NPARITY "nparity" #define ZPOOL_CONFIG_HOSTID "hostid" #define ZPOOL_CONFIG_HOSTNAME "hostname" #define ZPOOL_CONFIG_LOADED_TIME "initial_load_time" #define ZPOOL_CONFIG_UNSPARE "unspare" #define ZPOOL_CONFIG_PHYS_PATH "phys_path" #define ZPOOL_CONFIG_IS_LOG "is_log" #define ZPOOL_CONFIG_L2CACHE "l2cache" #define ZPOOL_CONFIG_HOLE_ARRAY "hole_array" #define ZPOOL_CONFIG_VDEV_CHILDREN "vdev_children" #define ZPOOL_CONFIG_IS_HOLE "is_hole" #define ZPOOL_CONFIG_DDT_HISTOGRAM "ddt_histogram" #define ZPOOL_CONFIG_DDT_OBJ_STATS "ddt_object_stats" #define ZPOOL_CONFIG_DDT_STATS "ddt_stats" #define ZPOOL_CONFIG_SPLIT "splitcfg" #define ZPOOL_CONFIG_ORIG_GUID "orig_guid" #define ZPOOL_CONFIG_SPLIT_GUID "split_guid" #define ZPOOL_CONFIG_SPLIT_LIST "guid_list" #define ZPOOL_CONFIG_NONALLOCATING "non_allocating" #define ZPOOL_CONFIG_REMOVING "removing" #define ZPOOL_CONFIG_RESILVER_TXG "resilver_txg" #define ZPOOL_CONFIG_REBUILD_TXG "rebuild_txg" #define ZPOOL_CONFIG_COMMENT "comment" #define ZPOOL_CONFIG_SUSPENDED "suspended" /* not stored on disk */ #define ZPOOL_CONFIG_SUSPENDED_REASON "suspended_reason" /* not stored */ #define ZPOOL_CONFIG_TIMESTAMP "timestamp" /* not stored on disk */ #define ZPOOL_CONFIG_BOOTFS "bootfs" /* not stored on disk */ #define ZPOOL_CONFIG_MISSING_DEVICES "missing_vdevs" /* not stored on disk */ #define ZPOOL_CONFIG_LOAD_INFO "load_info" /* not stored on disk */ #define ZPOOL_CONFIG_REWIND_INFO "rewind_info" /* not stored on disk */ #define ZPOOL_CONFIG_UNSUP_FEAT "unsup_feat" /* not stored on disk */ #define ZPOOL_CONFIG_ENABLED_FEAT "enabled_feat" /* not stored on disk */ #define ZPOOL_CONFIG_CAN_RDONLY "can_rdonly" /* not stored on disk */ #define ZPOOL_CONFIG_FEATURES_FOR_READ "features_for_read" #define ZPOOL_CONFIG_FEATURE_STATS "feature_stats" /* not stored on disk */ #define ZPOOL_CONFIG_ERRATA "errata" /* not stored on disk */ #define ZPOOL_CONFIG_VDEV_ROOT_ZAP "com.klarasystems:vdev_zap_root" #define ZPOOL_CONFIG_VDEV_TOP_ZAP "com.delphix:vdev_zap_top" #define ZPOOL_CONFIG_VDEV_LEAF_ZAP "com.delphix:vdev_zap_leaf" #define ZPOOL_CONFIG_HAS_PER_VDEV_ZAPS "com.delphix:has_per_vdev_zaps" #define ZPOOL_CONFIG_RESILVER_DEFER "com.datto:resilver_defer" #define ZPOOL_CONFIG_CACHEFILE "cachefile" /* not stored on disk */ #define ZPOOL_CONFIG_MMP_STATE "mmp_state" /* not stored on disk */ #define ZPOOL_CONFIG_MMP_TXG "mmp_txg" /* not stored on disk */ #define ZPOOL_CONFIG_MMP_SEQ "mmp_seq" /* not stored on disk */ #define ZPOOL_CONFIG_MMP_HOSTNAME "mmp_hostname" /* not stored on disk */ #define ZPOOL_CONFIG_MMP_HOSTID "mmp_hostid" /* not stored on disk */ #define ZPOOL_CONFIG_ALLOCATION_BIAS "alloc_bias" /* not stored on disk */ #define ZPOOL_CONFIG_EXPANSION_TIME "expansion_time" /* not stored */ #define ZPOOL_CONFIG_REBUILD_STATS "org.openzfs:rebuild_stats" #define ZPOOL_CONFIG_COMPATIBILITY "compatibility" /* * The persistent vdev state is stored as separate values rather than a single * 'vdev_state' entry. This is because a device can be in multiple states, such * as offline and degraded. */ #define ZPOOL_CONFIG_OFFLINE "offline" #define ZPOOL_CONFIG_FAULTED "faulted" #define ZPOOL_CONFIG_DEGRADED "degraded" #define ZPOOL_CONFIG_REMOVED "removed" #define ZPOOL_CONFIG_FRU "fru" #define ZPOOL_CONFIG_AUX_STATE "aux_state" /* Pool load policy parameters */ #define ZPOOL_LOAD_POLICY "load-policy" #define ZPOOL_LOAD_REWIND_POLICY "load-rewind-policy" #define ZPOOL_LOAD_REQUEST_TXG "load-request-txg" #define ZPOOL_LOAD_META_THRESH "load-meta-thresh" #define ZPOOL_LOAD_DATA_THRESH "load-data-thresh" /* Rewind data discovered */ #define ZPOOL_CONFIG_LOAD_TIME "rewind_txg_ts" #define ZPOOL_CONFIG_LOAD_META_ERRORS "verify_meta_errors" #define ZPOOL_CONFIG_LOAD_DATA_ERRORS "verify_data_errors" #define ZPOOL_CONFIG_REWIND_TIME "seconds_of_rewind" /* dRAID configuration */ #define ZPOOL_CONFIG_DRAID_NDATA "draid_ndata" #define ZPOOL_CONFIG_DRAID_NSPARES "draid_nspares" #define ZPOOL_CONFIG_DRAID_NGROUPS "draid_ngroups" #define VDEV_TYPE_ROOT "root" #define VDEV_TYPE_MIRROR "mirror" #define VDEV_TYPE_REPLACING "replacing" #define VDEV_TYPE_RAIDZ "raidz" #define VDEV_TYPE_DRAID "draid" #define VDEV_TYPE_DRAID_SPARE "dspare" #define VDEV_TYPE_DISK "disk" #define VDEV_TYPE_FILE "file" #define VDEV_TYPE_MISSING "missing" #define VDEV_TYPE_HOLE "hole" #define VDEV_TYPE_SPARE "spare" #define VDEV_TYPE_LOG "log" #define VDEV_TYPE_L2CACHE "l2cache" #define VDEV_TYPE_INDIRECT "indirect" #define VDEV_RAIDZ_MAXPARITY 3 #define VDEV_DRAID_MAXPARITY 3 #define VDEV_DRAID_MIN_CHILDREN 2 #define VDEV_DRAID_MAX_CHILDREN UINT8_MAX /* VDEV_TOP_ZAP_* are used in top-level vdev ZAP objects. */ #define VDEV_TOP_ZAP_INDIRECT_OBSOLETE_SM \ "com.delphix:indirect_obsolete_sm" #define VDEV_TOP_ZAP_OBSOLETE_COUNTS_ARE_PRECISE \ "com.delphix:obsolete_counts_are_precise" #define VDEV_TOP_ZAP_POOL_CHECKPOINT_SM \ "com.delphix:pool_checkpoint_sm" #define VDEV_TOP_ZAP_MS_UNFLUSHED_PHYS_TXGS \ "com.delphix:ms_unflushed_phys_txgs" #define VDEV_TOP_ZAP_VDEV_REBUILD_PHYS \ "org.openzfs:vdev_rebuild" #define VDEV_TOP_ZAP_ALLOCATION_BIAS \ "org.zfsonlinux:allocation_bias" /* vdev metaslab allocation bias */ #define VDEV_ALLOC_BIAS_LOG "log" #define VDEV_ALLOC_BIAS_SPECIAL "special" #define VDEV_ALLOC_BIAS_DEDUP "dedup" /* vdev initialize state */ #define VDEV_LEAF_ZAP_INITIALIZE_LAST_OFFSET \ "com.delphix:next_offset_to_initialize" #define VDEV_LEAF_ZAP_INITIALIZE_STATE \ "com.delphix:vdev_initialize_state" #define VDEV_LEAF_ZAP_INITIALIZE_ACTION_TIME \ "com.delphix:vdev_initialize_action_time" /* vdev TRIM state */ #define VDEV_LEAF_ZAP_TRIM_LAST_OFFSET \ "org.zfsonlinux:next_offset_to_trim" #define VDEV_LEAF_ZAP_TRIM_STATE \ "org.zfsonlinux:vdev_trim_state" #define VDEV_LEAF_ZAP_TRIM_ACTION_TIME \ "org.zfsonlinux:vdev_trim_action_time" #define VDEV_LEAF_ZAP_TRIM_RATE \ "org.zfsonlinux:vdev_trim_rate" #define VDEV_LEAF_ZAP_TRIM_PARTIAL \ "org.zfsonlinux:vdev_trim_partial" #define VDEV_LEAF_ZAP_TRIM_SECURE \ "org.zfsonlinux:vdev_trim_secure" /* * This is needed in userland to report the minimum necessary device size. */ #define SPA_MINDEVSIZE (64ULL << 20) /* * Set if the fragmentation has not yet been calculated. This can happen * because the space maps have not been upgraded or the histogram feature * is not enabled. */ #define ZFS_FRAG_INVALID UINT64_MAX /* * The location of the pool configuration repository, shared between kernel and * userland. */ #define ZPOOL_CACHE_BOOT "/boot/zfs/zpool.cache" #define ZPOOL_CACHE "/etc/zfs/zpool.cache" /* * Settings for zpool compatibility features files */ #define ZPOOL_SYSCONF_COMPAT_D SYSCONFDIR "/zfs/compatibility.d" #define ZPOOL_DATA_COMPAT_D PKGDATADIR "/compatibility.d" #define ZPOOL_COMPAT_MAXSIZE 16384 /* * Hard-wired compatibility settings */ #define ZPOOL_COMPAT_LEGACY "legacy" #define ZPOOL_COMPAT_OFF "off" /* * vdev states are ordered from least to most healthy. * A vdev that's CANT_OPEN or below is considered unusable. */ typedef enum vdev_state { VDEV_STATE_UNKNOWN = 0, /* Uninitialized vdev */ VDEV_STATE_CLOSED, /* Not currently open */ VDEV_STATE_OFFLINE, /* Not allowed to open */ VDEV_STATE_REMOVED, /* Explicitly removed from system */ VDEV_STATE_CANT_OPEN, /* Tried to open, but failed */ VDEV_STATE_FAULTED, /* External request to fault device */ VDEV_STATE_DEGRADED, /* Replicated vdev with unhealthy kids */ VDEV_STATE_HEALTHY /* Presumed good */ } vdev_state_t; #define VDEV_STATE_ONLINE VDEV_STATE_HEALTHY /* * vdev aux states. When a vdev is in the CANT_OPEN state, the aux field * of the vdev stats structure uses these constants to distinguish why. */ typedef enum vdev_aux { VDEV_AUX_NONE, /* no error */ VDEV_AUX_OPEN_FAILED, /* ldi_open_*() or vn_open() failed */ VDEV_AUX_CORRUPT_DATA, /* bad label or disk contents */ VDEV_AUX_NO_REPLICAS, /* insufficient number of replicas */ VDEV_AUX_BAD_GUID_SUM, /* vdev guid sum doesn't match */ VDEV_AUX_TOO_SMALL, /* vdev size is too small */ VDEV_AUX_BAD_LABEL, /* the label is OK but invalid */ VDEV_AUX_VERSION_NEWER, /* on-disk version is too new */ VDEV_AUX_VERSION_OLDER, /* on-disk version is too old */ VDEV_AUX_UNSUP_FEAT, /* unsupported features */ VDEV_AUX_SPARED, /* hot spare used in another pool */ VDEV_AUX_ERR_EXCEEDED, /* too many errors */ VDEV_AUX_IO_FAILURE, /* experienced I/O failure */ VDEV_AUX_BAD_LOG, /* cannot read log chain(s) */ VDEV_AUX_EXTERNAL, /* external diagnosis or forced fault */ VDEV_AUX_SPLIT_POOL, /* vdev was split off into another pool */ VDEV_AUX_BAD_ASHIFT, /* vdev ashift is invalid */ VDEV_AUX_EXTERNAL_PERSIST, /* persistent forced fault */ VDEV_AUX_ACTIVE, /* vdev active on a different host */ VDEV_AUX_CHILDREN_OFFLINE, /* all children are offline */ VDEV_AUX_ASHIFT_TOO_BIG, /* vdev's min block size is too large */ } vdev_aux_t; /* * pool state. The following states are written to disk as part of the normal * SPA lifecycle: ACTIVE, EXPORTED, DESTROYED, SPARE, L2CACHE. The remaining * states are software abstractions used at various levels to communicate * pool state. */ typedef enum pool_state { POOL_STATE_ACTIVE = 0, /* In active use */ POOL_STATE_EXPORTED, /* Explicitly exported */ POOL_STATE_DESTROYED, /* Explicitly destroyed */ POOL_STATE_SPARE, /* Reserved for hot spare use */ POOL_STATE_L2CACHE, /* Level 2 ARC device */ POOL_STATE_UNINITIALIZED, /* Internal spa_t state */ POOL_STATE_UNAVAIL, /* Internal libzfs state */ POOL_STATE_POTENTIALLY_ACTIVE /* Internal libzfs state */ } pool_state_t; /* * mmp state. The following states provide additional detail describing * why a pool couldn't be safely imported. */ typedef enum mmp_state { MMP_STATE_ACTIVE = 0, /* In active use */ MMP_STATE_INACTIVE, /* Inactive and safe to import */ MMP_STATE_NO_HOSTID /* System hostid is not set */ } mmp_state_t; /* * Scan Functions. */ typedef enum pool_scan_func { POOL_SCAN_NONE, POOL_SCAN_SCRUB, POOL_SCAN_RESILVER, + POOL_SCAN_ERRORSCRUB, POOL_SCAN_FUNCS } pool_scan_func_t; /* * Used to control scrub pause and resume. */ typedef enum pool_scrub_cmd { POOL_SCRUB_NORMAL = 0, POOL_SCRUB_PAUSE, POOL_SCRUB_FLAGS_END } pool_scrub_cmd_t; typedef enum { CS_NONE, CS_CHECKPOINT_EXISTS, CS_CHECKPOINT_DISCARDING, CS_NUM_STATES } checkpoint_state_t; typedef struct pool_checkpoint_stat { uint64_t pcs_state; /* checkpoint_state_t */ uint64_t pcs_start_time; /* time checkpoint/discard started */ uint64_t pcs_space; /* checkpointed space */ } pool_checkpoint_stat_t; /* * ZIO types. Needed to interpret vdev statistics below. */ typedef enum zio_type { ZIO_TYPE_NULL = 0, ZIO_TYPE_READ, ZIO_TYPE_WRITE, ZIO_TYPE_FREE, ZIO_TYPE_CLAIM, ZIO_TYPE_IOCTL, ZIO_TYPE_TRIM, ZIO_TYPES } zio_type_t; /* * Pool statistics. Note: all fields should be 64-bit because this * is passed between kernel and userland as an nvlist uint64 array. */ typedef struct pool_scan_stat { /* values stored on disk */ uint64_t pss_func; /* pool_scan_func_t */ uint64_t pss_state; /* dsl_scan_state_t */ uint64_t pss_start_time; /* scan start time */ uint64_t pss_end_time; /* scan end time */ uint64_t pss_to_examine; /* total bytes to scan */ uint64_t pss_examined; /* total bytes located by scanner */ uint64_t pss_to_process; /* total bytes to process */ uint64_t pss_processed; /* total processed bytes */ uint64_t pss_errors; /* scan errors */ /* values not stored on disk */ uint64_t pss_pass_exam; /* examined bytes per scan pass */ uint64_t pss_pass_start; /* start time of a scan pass */ uint64_t pss_pass_scrub_pause; /* pause time of a scrub pass */ /* cumulative time scrub spent paused, needed for rate calculation */ uint64_t pss_pass_scrub_spent_paused; uint64_t pss_pass_issued; /* issued bytes per scan pass */ uint64_t pss_issued; /* total bytes checked by scanner */ + + /* error scrub values stored on disk */ + uint64_t pss_error_scrub_func; /* pool_scan_func_t */ + uint64_t pss_error_scrub_state; /* dsl_scan_state_t */ + uint64_t pss_error_scrub_start; /* error scrub start time */ + uint64_t pss_error_scrub_end; /* error scrub end time */ + uint64_t pss_error_scrub_examined; /* error blocks issued I/O */ + /* error blocks to be issued I/O */ + uint64_t pss_error_scrub_to_be_examined; + + /* error scrub values not stored on disk */ + /* error scrub pause time in milliseconds */ + uint64_t pss_pass_error_scrub_pause; + } pool_scan_stat_t; typedef struct pool_removal_stat { uint64_t prs_state; /* dsl_scan_state_t */ uint64_t prs_removing_vdev; uint64_t prs_start_time; uint64_t prs_end_time; uint64_t prs_to_copy; /* bytes that need to be copied */ uint64_t prs_copied; /* bytes copied so far */ /* * bytes of memory used for indirect mappings. * This includes all removed vdevs. */ uint64_t prs_mapping_memory; } pool_removal_stat_t; typedef enum dsl_scan_state { DSS_NONE, DSS_SCANNING, DSS_FINISHED, DSS_CANCELED, + DSS_ERRORSCRUBBING, DSS_NUM_STATES } dsl_scan_state_t; typedef struct vdev_rebuild_stat { uint64_t vrs_state; /* vdev_rebuild_state_t */ uint64_t vrs_start_time; /* time_t */ uint64_t vrs_end_time; /* time_t */ uint64_t vrs_scan_time_ms; /* total run time (millisecs) */ uint64_t vrs_bytes_scanned; /* allocated bytes scanned */ uint64_t vrs_bytes_issued; /* read bytes issued */ uint64_t vrs_bytes_rebuilt; /* rebuilt bytes */ uint64_t vrs_bytes_est; /* total bytes to scan */ uint64_t vrs_errors; /* scanning errors */ uint64_t vrs_pass_time_ms; /* pass run time (millisecs) */ uint64_t vrs_pass_bytes_scanned; /* bytes scanned since start/resume */ uint64_t vrs_pass_bytes_issued; /* bytes rebuilt since start/resume */ } vdev_rebuild_stat_t; /* * Errata described by https://openzfs.github.io/openzfs-docs/msg/ZFS-8000-ER. * The ordering of this enum must be maintained to ensure the errata identifiers * map to the correct documentation. New errata may only be appended to the * list and must contain corresponding documentation at the above link. */ typedef enum zpool_errata { ZPOOL_ERRATA_NONE, ZPOOL_ERRATA_ZOL_2094_SCRUB, ZPOOL_ERRATA_ZOL_2094_ASYNC_DESTROY, ZPOOL_ERRATA_ZOL_6845_ENCRYPTION, ZPOOL_ERRATA_ZOL_8308_ENCRYPTION, } zpool_errata_t; /* * Vdev statistics. Note: all fields should be 64-bit because this * is passed between kernel and user land as an nvlist uint64 array. * * The vs_ops[] and vs_bytes[] arrays must always be an array size of 6 in * order to keep subsequent members at their known fixed offsets. When * adding a new field it must be added to the end the structure. */ #define VS_ZIO_TYPES 6 typedef struct vdev_stat { hrtime_t vs_timestamp; /* time since vdev load */ uint64_t vs_state; /* vdev state */ uint64_t vs_aux; /* see vdev_aux_t */ uint64_t vs_alloc; /* space allocated */ uint64_t vs_space; /* total capacity */ uint64_t vs_dspace; /* deflated capacity */ uint64_t vs_rsize; /* replaceable dev size */ uint64_t vs_esize; /* expandable dev size */ uint64_t vs_ops[VS_ZIO_TYPES]; /* operation count */ uint64_t vs_bytes[VS_ZIO_TYPES]; /* bytes read/written */ uint64_t vs_read_errors; /* read errors */ uint64_t vs_write_errors; /* write errors */ uint64_t vs_checksum_errors; /* checksum errors */ uint64_t vs_initialize_errors; /* initializing errors */ uint64_t vs_self_healed; /* self-healed bytes */ uint64_t vs_scan_removing; /* removing? */ uint64_t vs_scan_processed; /* scan processed bytes */ uint64_t vs_fragmentation; /* device fragmentation */ uint64_t vs_initialize_bytes_done; /* bytes initialized */ uint64_t vs_initialize_bytes_est; /* total bytes to initialize */ uint64_t vs_initialize_state; /* vdev_initializing_state_t */ uint64_t vs_initialize_action_time; /* time_t */ uint64_t vs_checkpoint_space; /* checkpoint-consumed space */ uint64_t vs_resilver_deferred; /* resilver deferred */ uint64_t vs_slow_ios; /* slow IOs */ uint64_t vs_trim_errors; /* trimming errors */ uint64_t vs_trim_notsup; /* supported by device */ uint64_t vs_trim_bytes_done; /* bytes trimmed */ uint64_t vs_trim_bytes_est; /* total bytes to trim */ uint64_t vs_trim_state; /* vdev_trim_state_t */ uint64_t vs_trim_action_time; /* time_t */ uint64_t vs_rebuild_processed; /* bytes rebuilt */ uint64_t vs_configured_ashift; /* TLV vdev_ashift */ uint64_t vs_logical_ashift; /* vdev_logical_ashift */ uint64_t vs_physical_ashift; /* vdev_physical_ashift */ uint64_t vs_noalloc; /* allocations halted? */ uint64_t vs_pspace; /* physical capacity */ } vdev_stat_t; #define VDEV_STAT_VALID(field, uint64_t_field_count) \ ((uint64_t_field_count * sizeof (uint64_t)) >= \ (offsetof(vdev_stat_t, field) + sizeof (((vdev_stat_t *)NULL)->field))) /* * Extended stats * * These are stats which aren't included in the original iostat output. For * convenience, they are grouped together in vdev_stat_ex, although each stat * is individually exported as an nvlist. */ typedef struct vdev_stat_ex { /* Number of ZIOs issued to disk and waiting to finish */ uint64_t vsx_active_queue[ZIO_PRIORITY_NUM_QUEUEABLE]; /* Number of ZIOs pending to be issued to disk */ uint64_t vsx_pend_queue[ZIO_PRIORITY_NUM_QUEUEABLE]; /* * Below are the histograms for various latencies. Buckets are in * units of nanoseconds. */ /* * 2^37 nanoseconds = 134s. Timeouts will probably start kicking in * before this. */ #define VDEV_L_HISTO_BUCKETS 37 /* Latency histo buckets */ #define VDEV_RQ_HISTO_BUCKETS 25 /* Request size histo buckets */ /* Amount of time in ZIO queue (ns) */ uint64_t vsx_queue_histo[ZIO_PRIORITY_NUM_QUEUEABLE] [VDEV_L_HISTO_BUCKETS]; /* Total ZIO latency (ns). Includes queuing and disk access time */ uint64_t vsx_total_histo[ZIO_TYPES][VDEV_L_HISTO_BUCKETS]; /* Amount of time to read/write the disk (ns) */ uint64_t vsx_disk_histo[ZIO_TYPES][VDEV_L_HISTO_BUCKETS]; /* "lookup the bucket for a value" histogram macros */ #define HISTO(val, buckets) (val != 0 ? MIN(highbit64(val) - 1, \ buckets - 1) : 0) #define L_HISTO(a) HISTO(a, VDEV_L_HISTO_BUCKETS) #define RQ_HISTO(a) HISTO(a, VDEV_RQ_HISTO_BUCKETS) /* Physical IO histogram */ uint64_t vsx_ind_histo[ZIO_PRIORITY_NUM_QUEUEABLE] [VDEV_RQ_HISTO_BUCKETS]; /* Delegated (aggregated) physical IO histogram */ uint64_t vsx_agg_histo[ZIO_PRIORITY_NUM_QUEUEABLE] [VDEV_RQ_HISTO_BUCKETS]; } vdev_stat_ex_t; /* * Initialize functions. */ typedef enum pool_initialize_func { POOL_INITIALIZE_START, POOL_INITIALIZE_CANCEL, POOL_INITIALIZE_SUSPEND, + POOL_INITIALIZE_UNINIT, POOL_INITIALIZE_FUNCS } pool_initialize_func_t; /* * TRIM functions. */ typedef enum pool_trim_func { POOL_TRIM_START, POOL_TRIM_CANCEL, POOL_TRIM_SUSPEND, POOL_TRIM_FUNCS } pool_trim_func_t; /* * DDT statistics. Note: all fields should be 64-bit because this * is passed between kernel and userland as an nvlist uint64 array. */ typedef struct ddt_object { uint64_t ddo_count; /* number of elements in ddt */ uint64_t ddo_dspace; /* size of ddt on disk */ uint64_t ddo_mspace; /* size of ddt in-core */ } ddt_object_t; typedef struct ddt_stat { uint64_t dds_blocks; /* blocks */ uint64_t dds_lsize; /* logical size */ uint64_t dds_psize; /* physical size */ uint64_t dds_dsize; /* deflated allocated size */ uint64_t dds_ref_blocks; /* referenced blocks */ uint64_t dds_ref_lsize; /* referenced lsize * refcnt */ uint64_t dds_ref_psize; /* referenced psize * refcnt */ uint64_t dds_ref_dsize; /* referenced dsize * refcnt */ } ddt_stat_t; typedef struct ddt_histogram { ddt_stat_t ddh_stat[64]; /* power-of-two histogram buckets */ } ddt_histogram_t; #define ZVOL_DRIVER "zvol" #define ZFS_DRIVER "zfs" #define ZFS_DEV "/dev/zfs" #define ZFS_DEVDIR "/dev" #define ZFS_SUPER_MAGIC 0x2fc12fc1 /* general zvol path */ #define ZVOL_DIR "/dev/zvol/" #define ZVOL_MAJOR 230 #define ZVOL_MINOR_BITS 4 #define ZVOL_MINOR_MASK ((1U << ZVOL_MINOR_BITS) - 1) #define ZVOL_MINORS (1 << 4) #define ZVOL_DEV_NAME "zd" #define ZVOL_PROP_NAME "name" #define ZVOL_DEFAULT_BLOCKSIZE 16384 typedef enum { VDEV_INITIALIZE_NONE, VDEV_INITIALIZE_ACTIVE, VDEV_INITIALIZE_CANCELED, VDEV_INITIALIZE_SUSPENDED, VDEV_INITIALIZE_COMPLETE } vdev_initializing_state_t; typedef enum { VDEV_TRIM_NONE, VDEV_TRIM_ACTIVE, VDEV_TRIM_CANCELED, VDEV_TRIM_SUSPENDED, VDEV_TRIM_COMPLETE, } vdev_trim_state_t; typedef enum { VDEV_REBUILD_NONE, VDEV_REBUILD_ACTIVE, VDEV_REBUILD_CANCELED, VDEV_REBUILD_COMPLETE, } vdev_rebuild_state_t; /* * nvlist name constants. Facilitate restricting snapshot iteration range for * the "list next snapshot" ioctl */ #define SNAP_ITER_MIN_TXG "snap_iter_min_txg" #define SNAP_ITER_MAX_TXG "snap_iter_max_txg" /* * /dev/zfs ioctl numbers. * * These numbers cannot change over time. New ioctl numbers must be appended. */ typedef enum zfs_ioc { /* - * Core features - 81/128 numbers reserved. + * Core features - 88/128 numbers reserved. */ #ifdef __FreeBSD__ ZFS_IOC_FIRST = 0, #else ZFS_IOC_FIRST = ('Z' << 8), #endif ZFS_IOC = ZFS_IOC_FIRST, ZFS_IOC_POOL_CREATE = ZFS_IOC_FIRST, /* 0x5a00 */ ZFS_IOC_POOL_DESTROY, /* 0x5a01 */ ZFS_IOC_POOL_IMPORT, /* 0x5a02 */ ZFS_IOC_POOL_EXPORT, /* 0x5a03 */ ZFS_IOC_POOL_CONFIGS, /* 0x5a04 */ ZFS_IOC_POOL_STATS, /* 0x5a05 */ ZFS_IOC_POOL_TRYIMPORT, /* 0x5a06 */ ZFS_IOC_POOL_SCAN, /* 0x5a07 */ ZFS_IOC_POOL_FREEZE, /* 0x5a08 */ ZFS_IOC_POOL_UPGRADE, /* 0x5a09 */ ZFS_IOC_POOL_GET_HISTORY, /* 0x5a0a */ ZFS_IOC_VDEV_ADD, /* 0x5a0b */ ZFS_IOC_VDEV_REMOVE, /* 0x5a0c */ ZFS_IOC_VDEV_SET_STATE, /* 0x5a0d */ ZFS_IOC_VDEV_ATTACH, /* 0x5a0e */ ZFS_IOC_VDEV_DETACH, /* 0x5a0f */ ZFS_IOC_VDEV_SETPATH, /* 0x5a10 */ ZFS_IOC_VDEV_SETFRU, /* 0x5a11 */ ZFS_IOC_OBJSET_STATS, /* 0x5a12 */ ZFS_IOC_OBJSET_ZPLPROPS, /* 0x5a13 */ ZFS_IOC_DATASET_LIST_NEXT, /* 0x5a14 */ ZFS_IOC_SNAPSHOT_LIST_NEXT, /* 0x5a15 */ ZFS_IOC_SET_PROP, /* 0x5a16 */ ZFS_IOC_CREATE, /* 0x5a17 */ ZFS_IOC_DESTROY, /* 0x5a18 */ ZFS_IOC_ROLLBACK, /* 0x5a19 */ ZFS_IOC_RENAME, /* 0x5a1a */ ZFS_IOC_RECV, /* 0x5a1b */ ZFS_IOC_SEND, /* 0x5a1c */ ZFS_IOC_INJECT_FAULT, /* 0x5a1d */ ZFS_IOC_CLEAR_FAULT, /* 0x5a1e */ ZFS_IOC_INJECT_LIST_NEXT, /* 0x5a1f */ ZFS_IOC_ERROR_LOG, /* 0x5a20 */ ZFS_IOC_CLEAR, /* 0x5a21 */ ZFS_IOC_PROMOTE, /* 0x5a22 */ ZFS_IOC_SNAPSHOT, /* 0x5a23 */ ZFS_IOC_DSOBJ_TO_DSNAME, /* 0x5a24 */ ZFS_IOC_OBJ_TO_PATH, /* 0x5a25 */ ZFS_IOC_POOL_SET_PROPS, /* 0x5a26 */ ZFS_IOC_POOL_GET_PROPS, /* 0x5a27 */ ZFS_IOC_SET_FSACL, /* 0x5a28 */ ZFS_IOC_GET_FSACL, /* 0x5a29 */ ZFS_IOC_SHARE, /* 0x5a2a */ ZFS_IOC_INHERIT_PROP, /* 0x5a2b */ ZFS_IOC_SMB_ACL, /* 0x5a2c */ ZFS_IOC_USERSPACE_ONE, /* 0x5a2d */ ZFS_IOC_USERSPACE_MANY, /* 0x5a2e */ ZFS_IOC_USERSPACE_UPGRADE, /* 0x5a2f */ ZFS_IOC_HOLD, /* 0x5a30 */ ZFS_IOC_RELEASE, /* 0x5a31 */ ZFS_IOC_GET_HOLDS, /* 0x5a32 */ ZFS_IOC_OBJSET_RECVD_PROPS, /* 0x5a33 */ ZFS_IOC_VDEV_SPLIT, /* 0x5a34 */ ZFS_IOC_NEXT_OBJ, /* 0x5a35 */ ZFS_IOC_DIFF, /* 0x5a36 */ ZFS_IOC_TMP_SNAPSHOT, /* 0x5a37 */ ZFS_IOC_OBJ_TO_STATS, /* 0x5a38 */ ZFS_IOC_SPACE_WRITTEN, /* 0x5a39 */ ZFS_IOC_SPACE_SNAPS, /* 0x5a3a */ ZFS_IOC_DESTROY_SNAPS, /* 0x5a3b */ ZFS_IOC_POOL_REGUID, /* 0x5a3c */ ZFS_IOC_POOL_REOPEN, /* 0x5a3d */ ZFS_IOC_SEND_PROGRESS, /* 0x5a3e */ ZFS_IOC_LOG_HISTORY, /* 0x5a3f */ ZFS_IOC_SEND_NEW, /* 0x5a40 */ ZFS_IOC_SEND_SPACE, /* 0x5a41 */ ZFS_IOC_CLONE, /* 0x5a42 */ ZFS_IOC_BOOKMARK, /* 0x5a43 */ ZFS_IOC_GET_BOOKMARKS, /* 0x5a44 */ ZFS_IOC_DESTROY_BOOKMARKS, /* 0x5a45 */ ZFS_IOC_RECV_NEW, /* 0x5a46 */ ZFS_IOC_POOL_SYNC, /* 0x5a47 */ ZFS_IOC_CHANNEL_PROGRAM, /* 0x5a48 */ ZFS_IOC_LOAD_KEY, /* 0x5a49 */ ZFS_IOC_UNLOAD_KEY, /* 0x5a4a */ ZFS_IOC_CHANGE_KEY, /* 0x5a4b */ ZFS_IOC_REMAP, /* 0x5a4c */ ZFS_IOC_POOL_CHECKPOINT, /* 0x5a4d */ ZFS_IOC_POOL_DISCARD_CHECKPOINT, /* 0x5a4e */ ZFS_IOC_POOL_INITIALIZE, /* 0x5a4f */ ZFS_IOC_POOL_TRIM, /* 0x5a50 */ ZFS_IOC_REDACT, /* 0x5a51 */ ZFS_IOC_GET_BOOKMARK_PROPS, /* 0x5a52 */ ZFS_IOC_WAIT, /* 0x5a53 */ ZFS_IOC_WAIT_FS, /* 0x5a54 */ ZFS_IOC_VDEV_GET_PROPS, /* 0x5a55 */ ZFS_IOC_VDEV_SET_PROPS, /* 0x5a56 */ + ZFS_IOC_POOL_SCRUB, /* 0x5a57 */ /* * Per-platform (Optional) - 8/128 numbers reserved. */ ZFS_IOC_PLATFORM = ZFS_IOC_FIRST + 0x80, ZFS_IOC_EVENTS_NEXT, /* 0x81 (Linux) */ ZFS_IOC_EVENTS_CLEAR, /* 0x82 (Linux) */ ZFS_IOC_EVENTS_SEEK, /* 0x83 (Linux) */ ZFS_IOC_NEXTBOOT, /* 0x84 (FreeBSD) */ ZFS_IOC_JAIL, /* 0x85 (FreeBSD) */ ZFS_IOC_USERNS_ATTACH = ZFS_IOC_JAIL, /* 0x85 (Linux) */ ZFS_IOC_UNJAIL, /* 0x86 (FreeBSD) */ ZFS_IOC_USERNS_DETACH = ZFS_IOC_UNJAIL, /* 0x86 (Linux) */ ZFS_IOC_SET_BOOTENV, /* 0x87 */ ZFS_IOC_GET_BOOTENV, /* 0x88 */ ZFS_IOC_LAST } zfs_ioc_t; /* * zvol ioctl to get dataset name */ #define BLKZNAME _IOR(0x12, 125, char[ZFS_MAX_DATASET_NAME_LEN]) #ifdef __linux__ /* * IOCTLs to update and retrieve additional file level attributes on * Linux. */ #define ZFS_IOC_GETDOSFLAGS _IOR(0x83, 1, uint64_t) #define ZFS_IOC_SETDOSFLAGS _IOW(0x83, 2, uint64_t) /* * Additional file level attributes, that are stored * in the upper half of z_pflags */ #define ZFS_READONLY 0x0000000100000000ull #define ZFS_HIDDEN 0x0000000200000000ull #define ZFS_SYSTEM 0x0000000400000000ull #define ZFS_ARCHIVE 0x0000000800000000ull #define ZFS_IMMUTABLE 0x0000001000000000ull #define ZFS_NOUNLINK 0x0000002000000000ull #define ZFS_APPENDONLY 0x0000004000000000ull #define ZFS_NODUMP 0x0000008000000000ull #define ZFS_OPAQUE 0x0000010000000000ull #define ZFS_AV_QUARANTINED 0x0000020000000000ull #define ZFS_AV_MODIFIED 0x0000040000000000ull #define ZFS_REPARSE 0x0000080000000000ull #define ZFS_OFFLINE 0x0000100000000000ull #define ZFS_SPARSE 0x0000200000000000ull #define ZFS_DOS_FL_USER_VISIBLE (ZFS_IMMUTABLE | ZFS_APPENDONLY | \ ZFS_NOUNLINK | ZFS_ARCHIVE | ZFS_NODUMP | ZFS_SYSTEM | \ ZFS_HIDDEN | ZFS_READONLY | ZFS_REPARSE | ZFS_OFFLINE | \ ZFS_SPARSE) #endif /* * ZFS-specific error codes used for returning descriptive errors * to the userland through zfs ioctls. * * The enum implicitly includes all the error codes from errno.h. * New code should use and extend this enum for errors that are * not described precisely by generic errno codes. * * These numbers should not change over time. New entries should be appended. * * (Keep in sync with contrib/pyzfs/libzfs_core/_constants.py) */ typedef enum { ZFS_ERR_CHECKPOINT_EXISTS = 1024, ZFS_ERR_DISCARDING_CHECKPOINT, ZFS_ERR_NO_CHECKPOINT, ZFS_ERR_DEVRM_IN_PROGRESS, ZFS_ERR_VDEV_TOO_BIG, ZFS_ERR_IOC_CMD_UNAVAIL, ZFS_ERR_IOC_ARG_UNAVAIL, ZFS_ERR_IOC_ARG_REQUIRED, ZFS_ERR_IOC_ARG_BADTYPE, ZFS_ERR_WRONG_PARENT, ZFS_ERR_FROM_IVSET_GUID_MISSING, ZFS_ERR_FROM_IVSET_GUID_MISMATCH, ZFS_ERR_SPILL_BLOCK_FLAG_MISSING, ZFS_ERR_UNKNOWN_SEND_STREAM_FEATURE, ZFS_ERR_EXPORT_IN_PROGRESS, ZFS_ERR_BOOKMARK_SOURCE_NOT_ANCESTOR, ZFS_ERR_STREAM_TRUNCATED, ZFS_ERR_STREAM_LARGE_BLOCK_MISMATCH, ZFS_ERR_RESILVER_IN_PROGRESS, ZFS_ERR_REBUILD_IN_PROGRESS, ZFS_ERR_BADPROP, ZFS_ERR_VDEV_NOTSUP, ZFS_ERR_NOT_USER_NAMESPACE, ZFS_ERR_RESUME_EXISTS, ZFS_ERR_CRYPTO_NOTSUP, } zfs_errno_t; /* * Internal SPA load state. Used by FMA diagnosis engine. */ typedef enum { SPA_LOAD_NONE, /* no load in progress */ SPA_LOAD_OPEN, /* normal open */ SPA_LOAD_IMPORT, /* import in progress */ SPA_LOAD_TRYIMPORT, /* tryimport in progress */ SPA_LOAD_RECOVER, /* recovery requested */ SPA_LOAD_ERROR, /* load failed */ SPA_LOAD_CREATE /* creation in progress */ } spa_load_state_t; typedef enum { ZPOOL_WAIT_CKPT_DISCARD, ZPOOL_WAIT_FREE, ZPOOL_WAIT_INITIALIZE, ZPOOL_WAIT_REPLACE, ZPOOL_WAIT_REMOVE, ZPOOL_WAIT_RESILVER, ZPOOL_WAIT_SCRUB, ZPOOL_WAIT_TRIM, ZPOOL_WAIT_NUM_ACTIVITIES } zpool_wait_activity_t; typedef enum { ZFS_WAIT_DELETEQ, ZFS_WAIT_NUM_ACTIVITIES } zfs_wait_activity_t; /* * Bookmark name values. */ #define ZPOOL_ERR_LIST "error list" #define ZPOOL_ERR_DATASET "dataset" #define ZPOOL_ERR_OBJECT "object" #define HIS_MAX_RECORD_LEN (MAXPATHLEN + MAXPATHLEN + 1) /* * The following are names used in the nvlist describing * the pool's history log. */ #define ZPOOL_HIST_RECORD "history record" #define ZPOOL_HIST_TIME "history time" #define ZPOOL_HIST_CMD "history command" #define ZPOOL_HIST_WHO "history who" #define ZPOOL_HIST_ZONE "history zone" #define ZPOOL_HIST_HOST "history hostname" #define ZPOOL_HIST_TXG "history txg" #define ZPOOL_HIST_INT_EVENT "history internal event" #define ZPOOL_HIST_INT_STR "history internal str" #define ZPOOL_HIST_INT_NAME "internal_name" #define ZPOOL_HIST_IOCTL "ioctl" #define ZPOOL_HIST_INPUT_NVL "in_nvl" #define ZPOOL_HIST_OUTPUT_NVL "out_nvl" #define ZPOOL_HIST_OUTPUT_SIZE "out_size" #define ZPOOL_HIST_DSNAME "dsname" #define ZPOOL_HIST_DSID "dsid" #define ZPOOL_HIST_ERRNO "errno" #define ZPOOL_HIST_ELAPSED_NS "elapsed_ns" /* * Special nvlist name that will not have its args recorded in the pool's * history log. */ #define ZPOOL_HIDDEN_ARGS "hidden_args" /* * The following are names used when invoking ZFS_IOC_POOL_INITIALIZE. */ #define ZPOOL_INITIALIZE_COMMAND "initialize_command" #define ZPOOL_INITIALIZE_VDEVS "initialize_vdevs" /* * The following are names used when invoking ZFS_IOC_POOL_TRIM. */ #define ZPOOL_TRIM_COMMAND "trim_command" #define ZPOOL_TRIM_VDEVS "trim_vdevs" #define ZPOOL_TRIM_RATE "trim_rate" #define ZPOOL_TRIM_SECURE "trim_secure" /* * The following are names used when invoking ZFS_IOC_POOL_WAIT. */ #define ZPOOL_WAIT_ACTIVITY "wait_activity" #define ZPOOL_WAIT_TAG "wait_tag" #define ZPOOL_WAIT_WAITED "wait_waited" /* * The following are names used when invoking ZFS_IOC_VDEV_GET_PROP. */ #define ZPOOL_VDEV_PROPS_GET_VDEV "vdevprops_get_vdev" #define ZPOOL_VDEV_PROPS_GET_PROPS "vdevprops_get_props" /* * The following are names used when invoking ZFS_IOC_VDEV_SET_PROP. */ #define ZPOOL_VDEV_PROPS_SET_VDEV "vdevprops_set_vdev" #define ZPOOL_VDEV_PROPS_SET_PROPS "vdevprops_set_props" /* * The following are names used when invoking ZFS_IOC_WAIT_FS. */ #define ZFS_WAIT_ACTIVITY "wait_activity" #define ZFS_WAIT_WAITED "wait_waited" /* * Flags for ZFS_IOC_VDEV_SET_STATE */ #define ZFS_ONLINE_CHECKREMOVE 0x1 #define ZFS_ONLINE_UNSPARE 0x2 #define ZFS_ONLINE_FORCEFAULT 0x4 #define ZFS_ONLINE_EXPAND 0x8 #define ZFS_ONLINE_SPARE 0x10 #define ZFS_OFFLINE_TEMPORARY 0x1 /* * Flags for ZFS_IOC_POOL_IMPORT */ #define ZFS_IMPORT_NORMAL 0x0 #define ZFS_IMPORT_VERBATIM 0x1 #define ZFS_IMPORT_ANY_HOST 0x2 #define ZFS_IMPORT_MISSING_LOG 0x4 #define ZFS_IMPORT_ONLY 0x8 #define ZFS_IMPORT_TEMP_NAME 0x10 #define ZFS_IMPORT_SKIP_MMP 0x20 #define ZFS_IMPORT_LOAD_KEYS 0x40 #define ZFS_IMPORT_CHECKPOINT 0x80 /* * Channel program argument/return nvlist keys and defaults. */ #define ZCP_ARG_PROGRAM "program" #define ZCP_ARG_ARGLIST "arg" #define ZCP_ARG_SYNC "sync" #define ZCP_ARG_INSTRLIMIT "instrlimit" #define ZCP_ARG_MEMLIMIT "memlimit" #define ZCP_ARG_CLIARGV "argv" #define ZCP_RET_ERROR "error" #define ZCP_RET_RETURN "return" #define ZCP_DEFAULT_INSTRLIMIT (10 * 1000 * 1000) #define ZCP_MAX_INSTRLIMIT (10 * ZCP_DEFAULT_INSTRLIMIT) #define ZCP_DEFAULT_MEMLIMIT (10 * 1024 * 1024) #define ZCP_MAX_MEMLIMIT (10 * ZCP_DEFAULT_MEMLIMIT) /* * Sysevent payload members. ZFS will generate the following sysevents with the * given payloads: * * ESC_ZFS_RESILVER_START * ESC_ZFS_RESILVER_FINISH * * ZFS_EV_POOL_NAME DATA_TYPE_STRING * ZFS_EV_POOL_GUID DATA_TYPE_UINT64 * ZFS_EV_RESILVER_TYPE DATA_TYPE_STRING * * ESC_ZFS_POOL_DESTROY * ESC_ZFS_POOL_REGUID * * ZFS_EV_POOL_NAME DATA_TYPE_STRING * ZFS_EV_POOL_GUID DATA_TYPE_UINT64 * * ESC_ZFS_VDEV_REMOVE * ESC_ZFS_VDEV_CLEAR * ESC_ZFS_VDEV_CHECK * * ZFS_EV_POOL_NAME DATA_TYPE_STRING * ZFS_EV_POOL_GUID DATA_TYPE_UINT64 * ZFS_EV_VDEV_PATH DATA_TYPE_STRING (optional) * ZFS_EV_VDEV_GUID DATA_TYPE_UINT64 * * ESC_ZFS_HISTORY_EVENT * * ZFS_EV_POOL_NAME DATA_TYPE_STRING * ZFS_EV_POOL_GUID DATA_TYPE_UINT64 * ZFS_EV_HIST_TIME DATA_TYPE_UINT64 (optional) * ZFS_EV_HIST_CMD DATA_TYPE_STRING (optional) * ZFS_EV_HIST_WHO DATA_TYPE_UINT64 (optional) * ZFS_EV_HIST_ZONE DATA_TYPE_STRING (optional) * ZFS_EV_HIST_HOST DATA_TYPE_STRING (optional) * ZFS_EV_HIST_TXG DATA_TYPE_UINT64 (optional) * ZFS_EV_HIST_INT_EVENT DATA_TYPE_UINT64 (optional) * ZFS_EV_HIST_INT_STR DATA_TYPE_STRING (optional) * ZFS_EV_HIST_INT_NAME DATA_TYPE_STRING (optional) * ZFS_EV_HIST_IOCTL DATA_TYPE_STRING (optional) * ZFS_EV_HIST_DSNAME DATA_TYPE_STRING (optional) * ZFS_EV_HIST_DSID DATA_TYPE_UINT64 (optional) * * The ZFS_EV_HIST_* members will correspond to the ZPOOL_HIST_* members in the * history log nvlist. The keynames will be free of any spaces or other * characters that could be potentially unexpected to consumers of the * sysevents. */ #define ZFS_EV_POOL_NAME "pool_name" #define ZFS_EV_POOL_GUID "pool_guid" #define ZFS_EV_VDEV_PATH "vdev_path" #define ZFS_EV_VDEV_GUID "vdev_guid" #define ZFS_EV_HIST_TIME "history_time" #define ZFS_EV_HIST_CMD "history_command" #define ZFS_EV_HIST_WHO "history_who" #define ZFS_EV_HIST_ZONE "history_zone" #define ZFS_EV_HIST_HOST "history_hostname" #define ZFS_EV_HIST_TXG "history_txg" #define ZFS_EV_HIST_INT_EVENT "history_internal_event" #define ZFS_EV_HIST_INT_STR "history_internal_str" #define ZFS_EV_HIST_INT_NAME "history_internal_name" #define ZFS_EV_HIST_IOCTL "history_ioctl" #define ZFS_EV_HIST_DSNAME "history_dsname" #define ZFS_EV_HIST_DSID "history_dsid" #define ZFS_EV_RESILVER_TYPE "resilver_type" /* * We currently support block sizes from 512 bytes to 16MB. * The benefits of larger blocks, and thus larger IO, need to be weighed * against the cost of COWing a giant block to modify one byte, and the * large latency of reading or writing a large block. * * The recordsize property can not be set larger than zfs_max_recordsize * (default 16MB on 64-bit and 1MB on 32-bit). See the comment near * zfs_max_recordsize in dsl_dataset.c for details. * * Note that although the LSIZE field of the blkptr_t can store sizes up * to 32MB, the dnode's dn_datablkszsec can only store sizes up to * 32MB - 512 bytes. Therefore, we limit SPA_MAXBLOCKSIZE to 16MB. */ #define SPA_MINBLOCKSHIFT 9 #define SPA_OLD_MAXBLOCKSHIFT 17 #define SPA_MAXBLOCKSHIFT 24 #define SPA_MINBLOCKSIZE (1ULL << SPA_MINBLOCKSHIFT) #define SPA_OLD_MAXBLOCKSIZE (1ULL << SPA_OLD_MAXBLOCKSHIFT) #define SPA_MAXBLOCKSIZE (1ULL << SPA_MAXBLOCKSHIFT) /* supported encryption algorithms */ enum zio_encrypt { ZIO_CRYPT_INHERIT = 0, ZIO_CRYPT_ON, ZIO_CRYPT_OFF, ZIO_CRYPT_AES_128_CCM, ZIO_CRYPT_AES_192_CCM, ZIO_CRYPT_AES_256_CCM, ZIO_CRYPT_AES_128_GCM, ZIO_CRYPT_AES_192_GCM, ZIO_CRYPT_AES_256_GCM, ZIO_CRYPT_FUNCTIONS }; #define ZIO_CRYPT_ON_VALUE ZIO_CRYPT_AES_256_GCM #define ZIO_CRYPT_DEFAULT ZIO_CRYPT_OFF /* * xattr namespace prefixes. These are forbidden in xattr names. * * For cross-platform compatibility, xattrs in the user namespace should not be * prefixed with the namespace name, but for backwards compatibility with older * ZFS on Linux versions we do prefix the namespace. */ #define ZFS_XA_NS_FREEBSD_PREFIX "freebsd:" #define ZFS_XA_NS_FREEBSD_PREFIX_LEN strlen("freebsd:") #define ZFS_XA_NS_LINUX_SECURITY_PREFIX "security." #define ZFS_XA_NS_LINUX_SECURITY_PREFIX_LEN strlen("security.") #define ZFS_XA_NS_LINUX_SYSTEM_PREFIX "system." #define ZFS_XA_NS_LINUX_SYSTEM_PREFIX_LEN strlen("system.") #define ZFS_XA_NS_LINUX_TRUSTED_PREFIX "trusted." #define ZFS_XA_NS_LINUX_TRUSTED_PREFIX_LEN strlen("trusted.") #define ZFS_XA_NS_LINUX_USER_PREFIX "user." #define ZFS_XA_NS_LINUX_USER_PREFIX_LEN strlen("user.") #define ZFS_XA_NS_PREFIX_MATCH(ns, name) \ (strncmp(name, ZFS_XA_NS_##ns##_PREFIX, \ ZFS_XA_NS_##ns##_PREFIX_LEN) == 0) #define ZFS_XA_NS_PREFIX_FORBIDDEN(name) \ (ZFS_XA_NS_PREFIX_MATCH(FREEBSD, name) || \ ZFS_XA_NS_PREFIX_MATCH(LINUX_SECURITY, name) || \ ZFS_XA_NS_PREFIX_MATCH(LINUX_SYSTEM, name) || \ ZFS_XA_NS_PREFIX_MATCH(LINUX_TRUSTED, name) || \ ZFS_XA_NS_PREFIX_MATCH(LINUX_USER, name)) #ifdef __cplusplus } #endif #endif /* _SYS_FS_ZFS_H */ diff --git a/sys/contrib/openzfs/include/sys/spa.h b/sys/contrib/openzfs/include/sys/spa.h index 460ea2bfee4e..ed752967cca6 100644 --- a/sys/contrib/openzfs/include/sys/spa.h +++ b/sys/contrib/openzfs/include/sys/spa.h @@ -1,1239 +1,1247 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2011, 2021 by Delphix. All rights reserved. * Copyright 2011 Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2014 Spectra Logic Corporation, All rights reserved. * Copyright 2013 Saso Kiselkov. All rights reserved. * Copyright (c) 2014 Integros [integros.com] * Copyright 2017 Joyent, Inc. * Copyright (c) 2017, Intel Corporation. * Copyright (c) 2019, Allan Jude * Copyright (c) 2019, Klara Inc. * Copyright (c) 2019, Datto Inc. */ #ifndef _SYS_SPA_H #define _SYS_SPA_H #include #include #include #include #include #include #include #include #include #include #include #ifdef __cplusplus extern "C" { #endif /* * Forward references that lots of things need. */ typedef struct spa spa_t; typedef struct vdev vdev_t; typedef struct metaslab metaslab_t; typedef struct metaslab_group metaslab_group_t; typedef struct metaslab_class metaslab_class_t; typedef struct zio zio_t; typedef struct zilog zilog_t; typedef struct spa_aux_vdev spa_aux_vdev_t; typedef struct ddt ddt_t; typedef struct ddt_entry ddt_entry_t; typedef struct zbookmark_phys zbookmark_phys_t; typedef struct zbookmark_err_phys zbookmark_err_phys_t; struct bpobj; struct bplist; struct dsl_pool; struct dsl_dataset; struct dsl_crypto_params; /* * Alignment Shift (ashift) is an immutable, internal top-level vdev property * which can only be set at vdev creation time. Physical writes are always done * according to it, which makes 2^ashift the smallest possible IO on a vdev. * * We currently allow values ranging from 512 bytes (2^9 = 512) to 64 KiB * (2^16 = 65,536). */ #define ASHIFT_MIN 9 #define ASHIFT_MAX 16 /* * Size of block to hold the configuration data (a packed nvlist) */ #define SPA_CONFIG_BLOCKSIZE (1ULL << 14) /* * The DVA size encodings for LSIZE and PSIZE support blocks up to 32MB. * The ASIZE encoding should be at least 64 times larger (6 more bits) * to support up to 4-way RAID-Z mirror mode with worst-case gang block * overhead, three DVAs per bp, plus one more bit in case we do anything * else that expands the ASIZE. */ #define SPA_LSIZEBITS 16 /* LSIZE up to 32M (2^16 * 512) */ #define SPA_PSIZEBITS 16 /* PSIZE up to 32M (2^16 * 512) */ #define SPA_ASIZEBITS 24 /* ASIZE up to 64 times larger */ #define SPA_COMPRESSBITS 7 #define SPA_VDEVBITS 24 #define SPA_COMPRESSMASK ((1U << SPA_COMPRESSBITS) - 1) /* * All SPA data is represented by 128-bit data virtual addresses (DVAs). * The members of the dva_t should be considered opaque outside the SPA. */ typedef struct dva { uint64_t dva_word[2]; } dva_t; /* * Some checksums/hashes need a 256-bit initialization salt. This salt is kept * secret and is suitable for use in MAC algorithms as the key. */ typedef struct zio_cksum_salt { uint8_t zcs_bytes[32]; } zio_cksum_salt_t; /* * Each block is described by its DVAs, time of birth, checksum, etc. * The word-by-word, bit-by-bit layout of the blkptr is as follows: * * 64 56 48 40 32 24 16 8 0 * +-------+-------+-------+-------+-------+-------+-------+-------+ * 0 | pad | vdev1 | GRID | ASIZE | * +-------+-------+-------+-------+-------+-------+-------+-------+ * 1 |G| offset1 | * +-------+-------+-------+-------+-------+-------+-------+-------+ * 2 | pad | vdev2 | GRID | ASIZE | * +-------+-------+-------+-------+-------+-------+-------+-------+ * 3 |G| offset2 | * +-------+-------+-------+-------+-------+-------+-------+-------+ * 4 | pad | vdev3 | GRID | ASIZE | * +-------+-------+-------+-------+-------+-------+-------+-------+ * 5 |G| offset3 | * +-------+-------+-------+-------+-------+-------+-------+-------+ * 6 |BDX|lvl| type | cksum |E| comp| PSIZE | LSIZE | * +-------+-------+-------+-------+-------+-------+-------+-------+ * 7 | padding | * +-------+-------+-------+-------+-------+-------+-------+-------+ * 8 | padding | * +-------+-------+-------+-------+-------+-------+-------+-------+ * 9 | physical birth txg | * +-------+-------+-------+-------+-------+-------+-------+-------+ * a | logical birth txg | * +-------+-------+-------+-------+-------+-------+-------+-------+ * b | fill count | * +-------+-------+-------+-------+-------+-------+-------+-------+ * c | checksum[0] | * +-------+-------+-------+-------+-------+-------+-------+-------+ * d | checksum[1] | * +-------+-------+-------+-------+-------+-------+-------+-------+ * e | checksum[2] | * +-------+-------+-------+-------+-------+-------+-------+-------+ * f | checksum[3] | * +-------+-------+-------+-------+-------+-------+-------+-------+ * * Legend: * * vdev virtual device ID * offset offset into virtual device * LSIZE logical size * PSIZE physical size (after compression) * ASIZE allocated size (including RAID-Z parity and gang block headers) * GRID RAID-Z layout information (reserved for future use) * cksum checksum function * comp compression function * G gang block indicator * B byteorder (endianness) * D dedup * X encryption * E blkptr_t contains embedded data (see below) * lvl level of indirection * type DMU object type * phys birth txg when dva[0] was written; zero if same as logical birth txg * note that typically all the dva's would be written in this * txg, but they could be different if they were moved by * device removal. * log. birth transaction group in which the block was logically born * fill count number of non-zero blocks under this bp * checksum[4] 256-bit checksum of the data this bp describes */ /* * The blkptr_t's of encrypted blocks also need to store the encryption * parameters so that the block can be decrypted. This layout is as follows: * * 64 56 48 40 32 24 16 8 0 * +-------+-------+-------+-------+-------+-------+-------+-------+ * 0 | vdev1 | GRID | ASIZE | * +-------+-------+-------+-------+-------+-------+-------+-------+ * 1 |G| offset1 | * +-------+-------+-------+-------+-------+-------+-------+-------+ * 2 | vdev2 | GRID | ASIZE | * +-------+-------+-------+-------+-------+-------+-------+-------+ * 3 |G| offset2 | * +-------+-------+-------+-------+-------+-------+-------+-------+ * 4 | salt | * +-------+-------+-------+-------+-------+-------+-------+-------+ * 5 | IV1 | * +-------+-------+-------+-------+-------+-------+-------+-------+ * 6 |BDX|lvl| type | cksum |E| comp| PSIZE | LSIZE | * +-------+-------+-------+-------+-------+-------+-------+-------+ * 7 | padding | * +-------+-------+-------+-------+-------+-------+-------+-------+ * 8 | padding | * +-------+-------+-------+-------+-------+-------+-------+-------+ * 9 | physical birth txg | * +-------+-------+-------+-------+-------+-------+-------+-------+ * a | logical birth txg | * +-------+-------+-------+-------+-------+-------+-------+-------+ * b | IV2 | fill count | * +-------+-------+-------+-------+-------+-------+-------+-------+ * c | checksum[0] | * +-------+-------+-------+-------+-------+-------+-------+-------+ * d | checksum[1] | * +-------+-------+-------+-------+-------+-------+-------+-------+ * e | MAC[0] | * +-------+-------+-------+-------+-------+-------+-------+-------+ * f | MAC[1] | * +-------+-------+-------+-------+-------+-------+-------+-------+ * * Legend: * * salt Salt for generating encryption keys * IV1 First 64 bits of encryption IV * X Block requires encryption handling (set to 1) * E blkptr_t contains embedded data (set to 0, see below) * fill count number of non-zero blocks under this bp (truncated to 32 bits) * IV2 Last 32 bits of encryption IV * checksum[2] 128-bit checksum of the data this bp describes * MAC[2] 128-bit message authentication code for this data * * The X bit being set indicates that this block is one of 3 types. If this is * a level 0 block with an encrypted object type, the block is encrypted * (see BP_IS_ENCRYPTED()). If this is a level 0 block with an unencrypted * object type, this block is authenticated with an HMAC (see * BP_IS_AUTHENTICATED()). Otherwise (if level > 0), this bp will use the MAC * words to store a checksum-of-MACs from the level below (see * BP_HAS_INDIRECT_MAC_CKSUM()). For convenience in the code, BP_IS_PROTECTED() * refers to both encrypted and authenticated blocks and BP_USES_CRYPT() * refers to any of these 3 kinds of blocks. * * The additional encryption parameters are the salt, IV, and MAC which are * explained in greater detail in the block comment at the top of zio_crypt.c. * The MAC occupies half of the checksum space since it serves a very similar * purpose: to prevent data corruption on disk. The only functional difference * is that the checksum is used to detect on-disk corruption whether or not the * encryption key is loaded and the MAC provides additional protection against * malicious disk tampering. We use the 3rd DVA to store the salt and first * 64 bits of the IV. As a result encrypted blocks can only have 2 copies * maximum instead of the normal 3. The last 32 bits of the IV are stored in * the upper bits of what is usually the fill count. Note that only blocks at * level 0 or -2 are ever encrypted, which allows us to guarantee that these * 32 bits are not trampled over by other code (see zio_crypt.c for details). * The salt and IV are not used for authenticated bps or bps with an indirect * MAC checksum, so these blocks can utilize all 3 DVAs and the full 64 bits * for the fill count. */ /* * "Embedded" blkptr_t's don't actually point to a block, instead they * have a data payload embedded in the blkptr_t itself. See the comment * in blkptr.c for more details. * * The blkptr_t is laid out as follows: * * 64 56 48 40 32 24 16 8 0 * +-------+-------+-------+-------+-------+-------+-------+-------+ * 0 | payload | * 1 | payload | * 2 | payload | * 3 | payload | * 4 | payload | * 5 | payload | * +-------+-------+-------+-------+-------+-------+-------+-------+ * 6 |BDX|lvl| type | etype |E| comp| PSIZE| LSIZE | * +-------+-------+-------+-------+-------+-------+-------+-------+ * 7 | payload | * 8 | payload | * 9 | payload | * +-------+-------+-------+-------+-------+-------+-------+-------+ * a | logical birth txg | * +-------+-------+-------+-------+-------+-------+-------+-------+ * b | payload | * c | payload | * d | payload | * e | payload | * f | payload | * +-------+-------+-------+-------+-------+-------+-------+-------+ * * Legend: * * payload contains the embedded data * B (byteorder) byteorder (endianness) * D (dedup) padding (set to zero) * X encryption (set to zero) * E (embedded) set to one * lvl indirection level * type DMU object type * etype how to interpret embedded data (BP_EMBEDDED_TYPE_*) * comp compression function of payload * PSIZE size of payload after compression, in bytes * LSIZE logical size of payload, in bytes * note that 25 bits is enough to store the largest * "normal" BP's LSIZE (2^16 * 2^9) in bytes * log. birth transaction group in which the block was logically born * * Note that LSIZE and PSIZE are stored in bytes, whereas for non-embedded * bp's they are stored in units of SPA_MINBLOCKSHIFT. * Generally, the generic BP_GET_*() macros can be used on embedded BP's. * The B, D, X, lvl, type, and comp fields are stored the same as with normal * BP's so the BP_SET_* macros can be used with them. etype, PSIZE, LSIZE must * be set with the BPE_SET_* macros. BP_SET_EMBEDDED() should be called before * other macros, as they assert that they are only used on BP's of the correct * "embedded-ness". Encrypted blkptr_t's cannot be embedded because they use * the payload space for encryption parameters (see the comment above on * how encryption parameters are stored). */ #define BPE_GET_ETYPE(bp) \ (ASSERT(BP_IS_EMBEDDED(bp)), \ BF64_GET((bp)->blk_prop, 40, 8)) #define BPE_SET_ETYPE(bp, t) do { \ ASSERT(BP_IS_EMBEDDED(bp)); \ BF64_SET((bp)->blk_prop, 40, 8, t); \ } while (0) #define BPE_GET_LSIZE(bp) \ (ASSERT(BP_IS_EMBEDDED(bp)), \ BF64_GET_SB((bp)->blk_prop, 0, 25, 0, 1)) #define BPE_SET_LSIZE(bp, x) do { \ ASSERT(BP_IS_EMBEDDED(bp)); \ BF64_SET_SB((bp)->blk_prop, 0, 25, 0, 1, x); \ } while (0) #define BPE_GET_PSIZE(bp) \ (ASSERT(BP_IS_EMBEDDED(bp)), \ BF64_GET_SB((bp)->blk_prop, 25, 7, 0, 1)) #define BPE_SET_PSIZE(bp, x) do { \ ASSERT(BP_IS_EMBEDDED(bp)); \ BF64_SET_SB((bp)->blk_prop, 25, 7, 0, 1, x); \ } while (0) typedef enum bp_embedded_type { BP_EMBEDDED_TYPE_DATA, BP_EMBEDDED_TYPE_RESERVED, /* Reserved for Delphix byteswap feature. */ BP_EMBEDDED_TYPE_REDACTED, NUM_BP_EMBEDDED_TYPES } bp_embedded_type_t; #define BPE_NUM_WORDS 14 #define BPE_PAYLOAD_SIZE (BPE_NUM_WORDS * sizeof (uint64_t)) #define BPE_IS_PAYLOADWORD(bp, wp) \ ((wp) != &(bp)->blk_prop && (wp) != &(bp)->blk_birth) #define SPA_BLKPTRSHIFT 7 /* blkptr_t is 128 bytes */ #define SPA_DVAS_PER_BP 3 /* Number of DVAs in a bp */ #define SPA_SYNC_MIN_VDEVS 3 /* min vdevs to update during sync */ /* * A block is a hole when it has either 1) never been written to, or * 2) is zero-filled. In both cases, ZFS can return all zeroes for all reads * without physically allocating disk space. Holes are represented in the * blkptr_t structure by zeroed blk_dva. Correct checking for holes is * done through the BP_IS_HOLE macro. For holes, the logical size, level, * DMU object type, and birth times are all also stored for holes that * were written to at some point (i.e. were punched after having been filled). */ typedef struct blkptr { dva_t blk_dva[SPA_DVAS_PER_BP]; /* Data Virtual Addresses */ uint64_t blk_prop; /* size, compression, type, etc */ uint64_t blk_pad[2]; /* Extra space for the future */ uint64_t blk_phys_birth; /* txg when block was allocated */ uint64_t blk_birth; /* transaction group at birth */ uint64_t blk_fill; /* fill count */ zio_cksum_t blk_cksum; /* 256-bit checksum */ } blkptr_t; /* * Macros to get and set fields in a bp or DVA. */ /* * Note, for gang blocks, DVA_GET_ASIZE() is the total space allocated for * this gang DVA including its children BP's. The space allocated at this * DVA's vdev/offset is vdev_gang_header_asize(vdev). */ #define DVA_GET_ASIZE(dva) \ BF64_GET_SB((dva)->dva_word[0], 0, SPA_ASIZEBITS, SPA_MINBLOCKSHIFT, 0) #define DVA_SET_ASIZE(dva, x) \ BF64_SET_SB((dva)->dva_word[0], 0, SPA_ASIZEBITS, \ SPA_MINBLOCKSHIFT, 0, x) #define DVA_GET_GRID(dva) BF64_GET((dva)->dva_word[0], 24, 8) #define DVA_SET_GRID(dva, x) BF64_SET((dva)->dva_word[0], 24, 8, x) #define DVA_GET_VDEV(dva) BF64_GET((dva)->dva_word[0], 32, SPA_VDEVBITS) #define DVA_SET_VDEV(dva, x) \ BF64_SET((dva)->dva_word[0], 32, SPA_VDEVBITS, x) #define DVA_GET_OFFSET(dva) \ BF64_GET_SB((dva)->dva_word[1], 0, 63, SPA_MINBLOCKSHIFT, 0) #define DVA_SET_OFFSET(dva, x) \ BF64_SET_SB((dva)->dva_word[1], 0, 63, SPA_MINBLOCKSHIFT, 0, x) #define DVA_GET_GANG(dva) BF64_GET((dva)->dva_word[1], 63, 1) #define DVA_SET_GANG(dva, x) BF64_SET((dva)->dva_word[1], 63, 1, x) #define BP_GET_LSIZE(bp) \ (BP_IS_EMBEDDED(bp) ? \ (BPE_GET_ETYPE(bp) == BP_EMBEDDED_TYPE_DATA ? BPE_GET_LSIZE(bp) : 0): \ BF64_GET_SB((bp)->blk_prop, 0, SPA_LSIZEBITS, SPA_MINBLOCKSHIFT, 1)) #define BP_SET_LSIZE(bp, x) do { \ ASSERT(!BP_IS_EMBEDDED(bp)); \ BF64_SET_SB((bp)->blk_prop, \ 0, SPA_LSIZEBITS, SPA_MINBLOCKSHIFT, 1, x); \ } while (0) #define BP_GET_PSIZE(bp) \ (BP_IS_EMBEDDED(bp) ? 0 : \ BF64_GET_SB((bp)->blk_prop, 16, SPA_PSIZEBITS, SPA_MINBLOCKSHIFT, 1)) #define BP_SET_PSIZE(bp, x) do { \ ASSERT(!BP_IS_EMBEDDED(bp)); \ BF64_SET_SB((bp)->blk_prop, \ 16, SPA_PSIZEBITS, SPA_MINBLOCKSHIFT, 1, x); \ } while (0) #define BP_GET_COMPRESS(bp) \ BF64_GET((bp)->blk_prop, 32, SPA_COMPRESSBITS) #define BP_SET_COMPRESS(bp, x) \ BF64_SET((bp)->blk_prop, 32, SPA_COMPRESSBITS, x) #define BP_IS_EMBEDDED(bp) BF64_GET((bp)->blk_prop, 39, 1) #define BP_SET_EMBEDDED(bp, x) BF64_SET((bp)->blk_prop, 39, 1, x) #define BP_GET_CHECKSUM(bp) \ (BP_IS_EMBEDDED(bp) ? ZIO_CHECKSUM_OFF : \ BF64_GET((bp)->blk_prop, 40, 8)) #define BP_SET_CHECKSUM(bp, x) do { \ ASSERT(!BP_IS_EMBEDDED(bp)); \ BF64_SET((bp)->blk_prop, 40, 8, x); \ } while (0) #define BP_GET_TYPE(bp) BF64_GET((bp)->blk_prop, 48, 8) #define BP_SET_TYPE(bp, x) BF64_SET((bp)->blk_prop, 48, 8, x) #define BP_GET_LEVEL(bp) BF64_GET((bp)->blk_prop, 56, 5) #define BP_SET_LEVEL(bp, x) BF64_SET((bp)->blk_prop, 56, 5, x) /* encrypted, authenticated, and MAC cksum bps use the same bit */ #define BP_USES_CRYPT(bp) BF64_GET((bp)->blk_prop, 61, 1) #define BP_SET_CRYPT(bp, x) BF64_SET((bp)->blk_prop, 61, 1, x) #define BP_IS_ENCRYPTED(bp) \ (BP_USES_CRYPT(bp) && \ BP_GET_LEVEL(bp) <= 0 && \ DMU_OT_IS_ENCRYPTED(BP_GET_TYPE(bp))) #define BP_IS_AUTHENTICATED(bp) \ (BP_USES_CRYPT(bp) && \ BP_GET_LEVEL(bp) <= 0 && \ !DMU_OT_IS_ENCRYPTED(BP_GET_TYPE(bp))) #define BP_HAS_INDIRECT_MAC_CKSUM(bp) \ (BP_USES_CRYPT(bp) && BP_GET_LEVEL(bp) > 0) #define BP_IS_PROTECTED(bp) \ (BP_IS_ENCRYPTED(bp) || BP_IS_AUTHENTICATED(bp)) #define BP_GET_DEDUP(bp) BF64_GET((bp)->blk_prop, 62, 1) #define BP_SET_DEDUP(bp, x) BF64_SET((bp)->blk_prop, 62, 1, x) #define BP_GET_BYTEORDER(bp) BF64_GET((bp)->blk_prop, 63, 1) #define BP_SET_BYTEORDER(bp, x) BF64_SET((bp)->blk_prop, 63, 1, x) #define BP_GET_FREE(bp) BF64_GET((bp)->blk_fill, 0, 1) #define BP_SET_FREE(bp, x) BF64_SET((bp)->blk_fill, 0, 1, x) #define BP_PHYSICAL_BIRTH(bp) \ (BP_IS_EMBEDDED(bp) ? 0 : \ (bp)->blk_phys_birth ? (bp)->blk_phys_birth : (bp)->blk_birth) #define BP_SET_BIRTH(bp, logical, physical) \ { \ ASSERT(!BP_IS_EMBEDDED(bp)); \ (bp)->blk_birth = (logical); \ (bp)->blk_phys_birth = ((logical) == (physical) ? 0 : (physical)); \ } #define BP_GET_FILL(bp) \ ((BP_IS_ENCRYPTED(bp)) ? BF64_GET((bp)->blk_fill, 0, 32) : \ ((BP_IS_EMBEDDED(bp)) ? 1 : (bp)->blk_fill)) #define BP_SET_FILL(bp, fill) \ { \ if (BP_IS_ENCRYPTED(bp)) \ BF64_SET((bp)->blk_fill, 0, 32, fill); \ else \ (bp)->blk_fill = fill; \ } #define BP_GET_IV2(bp) \ (ASSERT(BP_IS_ENCRYPTED(bp)), \ BF64_GET((bp)->blk_fill, 32, 32)) #define BP_SET_IV2(bp, iv2) \ { \ ASSERT(BP_IS_ENCRYPTED(bp)); \ BF64_SET((bp)->blk_fill, 32, 32, iv2); \ } #define BP_IS_METADATA(bp) \ (BP_GET_LEVEL(bp) > 0 || DMU_OT_IS_METADATA(BP_GET_TYPE(bp))) #define BP_GET_ASIZE(bp) \ (BP_IS_EMBEDDED(bp) ? 0 : \ DVA_GET_ASIZE(&(bp)->blk_dva[0]) + \ DVA_GET_ASIZE(&(bp)->blk_dva[1]) + \ (DVA_GET_ASIZE(&(bp)->blk_dva[2]) * !BP_IS_ENCRYPTED(bp))) #define BP_GET_UCSIZE(bp) \ (BP_IS_METADATA(bp) ? BP_GET_PSIZE(bp) : BP_GET_LSIZE(bp)) #define BP_GET_NDVAS(bp) \ (BP_IS_EMBEDDED(bp) ? 0 : \ !!DVA_GET_ASIZE(&(bp)->blk_dva[0]) + \ !!DVA_GET_ASIZE(&(bp)->blk_dva[1]) + \ (!!DVA_GET_ASIZE(&(bp)->blk_dva[2]) * !BP_IS_ENCRYPTED(bp))) #define BP_COUNT_GANG(bp) \ (BP_IS_EMBEDDED(bp) ? 0 : \ (DVA_GET_GANG(&(bp)->blk_dva[0]) + \ DVA_GET_GANG(&(bp)->blk_dva[1]) + \ (DVA_GET_GANG(&(bp)->blk_dva[2]) * !BP_IS_ENCRYPTED(bp)))) #define DVA_EQUAL(dva1, dva2) \ ((dva1)->dva_word[1] == (dva2)->dva_word[1] && \ (dva1)->dva_word[0] == (dva2)->dva_word[0]) #define BP_EQUAL(bp1, bp2) \ (BP_PHYSICAL_BIRTH(bp1) == BP_PHYSICAL_BIRTH(bp2) && \ (bp1)->blk_birth == (bp2)->blk_birth && \ DVA_EQUAL(&(bp1)->blk_dva[0], &(bp2)->blk_dva[0]) && \ DVA_EQUAL(&(bp1)->blk_dva[1], &(bp2)->blk_dva[1]) && \ DVA_EQUAL(&(bp1)->blk_dva[2], &(bp2)->blk_dva[2])) #define DVA_IS_VALID(dva) (DVA_GET_ASIZE(dva) != 0) #define BP_IDENTITY(bp) (ASSERT(!BP_IS_EMBEDDED(bp)), &(bp)->blk_dva[0]) #define BP_IS_GANG(bp) \ (BP_IS_EMBEDDED(bp) ? B_FALSE : DVA_GET_GANG(BP_IDENTITY(bp))) #define DVA_IS_EMPTY(dva) ((dva)->dva_word[0] == 0ULL && \ (dva)->dva_word[1] == 0ULL) #define BP_IS_HOLE(bp) \ (!BP_IS_EMBEDDED(bp) && DVA_IS_EMPTY(BP_IDENTITY(bp))) #define BP_SET_REDACTED(bp) \ { \ BP_SET_EMBEDDED(bp, B_TRUE); \ BPE_SET_ETYPE(bp, BP_EMBEDDED_TYPE_REDACTED); \ } #define BP_IS_REDACTED(bp) \ (BP_IS_EMBEDDED(bp) && BPE_GET_ETYPE(bp) == BP_EMBEDDED_TYPE_REDACTED) /* BP_IS_RAIDZ(bp) assumes no block compression */ #define BP_IS_RAIDZ(bp) (DVA_GET_ASIZE(&(bp)->blk_dva[0]) > \ BP_GET_PSIZE(bp)) #define BP_ZERO(bp) \ { \ (bp)->blk_dva[0].dva_word[0] = 0; \ (bp)->blk_dva[0].dva_word[1] = 0; \ (bp)->blk_dva[1].dva_word[0] = 0; \ (bp)->blk_dva[1].dva_word[1] = 0; \ (bp)->blk_dva[2].dva_word[0] = 0; \ (bp)->blk_dva[2].dva_word[1] = 0; \ (bp)->blk_prop = 0; \ (bp)->blk_pad[0] = 0; \ (bp)->blk_pad[1] = 0; \ (bp)->blk_phys_birth = 0; \ (bp)->blk_birth = 0; \ (bp)->blk_fill = 0; \ ZIO_SET_CHECKSUM(&(bp)->blk_cksum, 0, 0, 0, 0); \ } #ifdef _ZFS_BIG_ENDIAN #define ZFS_HOST_BYTEORDER (0ULL) #else #define ZFS_HOST_BYTEORDER (1ULL) #endif #define BP_SHOULD_BYTESWAP(bp) (BP_GET_BYTEORDER(bp) != ZFS_HOST_BYTEORDER) #define BP_SPRINTF_LEN 400 /* * This macro allows code sharing between zfs, libzpool, and mdb. * 'func' is either kmem_scnprintf() or mdb_snprintf(). * 'ws' (whitespace) can be ' ' for single-line format, '\n' for multi-line. */ #define SNPRINTF_BLKPTR(func, ws, buf, size, bp, type, checksum, compress) \ { \ static const char *const copyname[] = \ { "zero", "single", "double", "triple" }; \ int len = 0; \ int copies = 0; \ const char *crypt_type; \ if (bp != NULL) { \ if (BP_IS_ENCRYPTED(bp)) { \ crypt_type = "encrypted"; \ /* LINTED E_SUSPICIOUS_COMPARISON */ \ } else if (BP_IS_AUTHENTICATED(bp)) { \ crypt_type = "authenticated"; \ } else if (BP_HAS_INDIRECT_MAC_CKSUM(bp)) { \ crypt_type = "indirect-MAC"; \ } else { \ crypt_type = "unencrypted"; \ } \ } \ if (bp == NULL) { \ len += func(buf + len, size - len, ""); \ } else if (BP_IS_HOLE(bp)) { \ len += func(buf + len, size - len, \ "HOLE [L%llu %s] " \ "size=%llxL birth=%lluL", \ (u_longlong_t)BP_GET_LEVEL(bp), \ type, \ (u_longlong_t)BP_GET_LSIZE(bp), \ (u_longlong_t)bp->blk_birth); \ } else if (BP_IS_EMBEDDED(bp)) { \ len = func(buf + len, size - len, \ "EMBEDDED [L%llu %s] et=%u %s " \ "size=%llxL/%llxP birth=%lluL", \ (u_longlong_t)BP_GET_LEVEL(bp), \ type, \ (int)BPE_GET_ETYPE(bp), \ compress, \ (u_longlong_t)BPE_GET_LSIZE(bp), \ (u_longlong_t)BPE_GET_PSIZE(bp), \ (u_longlong_t)bp->blk_birth); \ } else if (BP_IS_REDACTED(bp)) { \ len += func(buf + len, size - len, \ "REDACTED [L%llu %s] size=%llxL birth=%lluL", \ (u_longlong_t)BP_GET_LEVEL(bp), \ type, \ (u_longlong_t)BP_GET_LSIZE(bp), \ (u_longlong_t)bp->blk_birth); \ } else { \ for (int d = 0; d < BP_GET_NDVAS(bp); d++) { \ const dva_t *dva = &bp->blk_dva[d]; \ if (DVA_IS_VALID(dva)) \ copies++; \ len += func(buf + len, size - len, \ "DVA[%d]=<%llu:%llx:%llx>%c", d, \ (u_longlong_t)DVA_GET_VDEV(dva), \ (u_longlong_t)DVA_GET_OFFSET(dva), \ (u_longlong_t)DVA_GET_ASIZE(dva), \ ws); \ } \ ASSERT3S(copies, >, 0); \ if (BP_IS_ENCRYPTED(bp)) { \ len += func(buf + len, size - len, \ "salt=%llx iv=%llx:%llx%c", \ (u_longlong_t)bp->blk_dva[2].dva_word[0], \ (u_longlong_t)bp->blk_dva[2].dva_word[1], \ (u_longlong_t)BP_GET_IV2(bp), \ ws); \ } \ if (BP_IS_GANG(bp) && \ DVA_GET_ASIZE(&bp->blk_dva[2]) <= \ DVA_GET_ASIZE(&bp->blk_dva[1]) / 2) \ copies--; \ len += func(buf + len, size - len, \ "[L%llu %s] %s %s %s %s %s %s %s%c" \ "size=%llxL/%llxP birth=%lluL/%lluP fill=%llu%c" \ "cksum=%016llx:%016llx:%016llx:%016llx", \ (u_longlong_t)BP_GET_LEVEL(bp), \ type, \ checksum, \ compress, \ crypt_type, \ BP_GET_BYTEORDER(bp) == 0 ? "BE" : "LE", \ BP_IS_GANG(bp) ? "gang" : "contiguous", \ BP_GET_DEDUP(bp) ? "dedup" : "unique", \ copyname[copies], \ ws, \ (u_longlong_t)BP_GET_LSIZE(bp), \ (u_longlong_t)BP_GET_PSIZE(bp), \ (u_longlong_t)bp->blk_birth, \ (u_longlong_t)BP_PHYSICAL_BIRTH(bp), \ (u_longlong_t)BP_GET_FILL(bp), \ ws, \ (u_longlong_t)bp->blk_cksum.zc_word[0], \ (u_longlong_t)bp->blk_cksum.zc_word[1], \ (u_longlong_t)bp->blk_cksum.zc_word[2], \ (u_longlong_t)bp->blk_cksum.zc_word[3]); \ } \ ASSERT(len < size); \ } #define BP_GET_BUFC_TYPE(bp) \ (BP_IS_METADATA(bp) ? ARC_BUFC_METADATA : ARC_BUFC_DATA) typedef enum spa_import_type { SPA_IMPORT_EXISTING, SPA_IMPORT_ASSEMBLE } spa_import_type_t; typedef enum spa_mode { SPA_MODE_UNINIT = 0, SPA_MODE_READ = 1, SPA_MODE_WRITE = 2, } spa_mode_t; /* * Send TRIM commands in-line during normal pool operation while deleting. * OFF: no * ON: yes * NB: IN_FREEBSD_BASE is defined within the FreeBSD sources. */ typedef enum { SPA_AUTOTRIM_OFF = 0, /* default */ SPA_AUTOTRIM_ON, #ifdef IN_FREEBSD_BASE SPA_AUTOTRIM_DEFAULT = SPA_AUTOTRIM_ON, #else SPA_AUTOTRIM_DEFAULT = SPA_AUTOTRIM_OFF, #endif } spa_autotrim_t; /* * Reason TRIM command was issued, used internally for accounting purposes. */ typedef enum trim_type { TRIM_TYPE_MANUAL = 0, TRIM_TYPE_AUTO = 1, TRIM_TYPE_SIMPLE = 2 } trim_type_t; /* state manipulation functions */ extern int spa_open(const char *pool, spa_t **, const void *tag); extern int spa_open_rewind(const char *pool, spa_t **, const void *tag, nvlist_t *policy, nvlist_t **config); extern int spa_get_stats(const char *pool, nvlist_t **config, char *altroot, size_t buflen); extern int spa_create(const char *pool, nvlist_t *nvroot, nvlist_t *props, nvlist_t *zplprops, struct dsl_crypto_params *dcp); extern int spa_import(char *pool, nvlist_t *config, nvlist_t *props, uint64_t flags); extern nvlist_t *spa_tryimport(nvlist_t *tryconfig); extern int spa_destroy(const char *pool); extern int spa_checkpoint(const char *pool); extern int spa_checkpoint_discard(const char *pool); extern int spa_export(const char *pool, nvlist_t **oldconfig, boolean_t force, boolean_t hardforce); extern int spa_reset(const char *pool); extern void spa_async_request(spa_t *spa, int flag); extern void spa_async_unrequest(spa_t *spa, int flag); extern void spa_async_suspend(spa_t *spa); extern void spa_async_resume(spa_t *spa); extern int spa_async_tasks(spa_t *spa); extern spa_t *spa_inject_addref(char *pool); extern void spa_inject_delref(spa_t *spa); extern void spa_scan_stat_init(spa_t *spa); extern int spa_scan_get_stats(spa_t *spa, pool_scan_stat_t *ps); extern int bpobj_enqueue_alloc_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx); extern int bpobj_enqueue_free_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx); #define SPA_ASYNC_CONFIG_UPDATE 0x01 #define SPA_ASYNC_REMOVE 0x02 #define SPA_ASYNC_PROBE 0x04 #define SPA_ASYNC_RESILVER_DONE 0x08 #define SPA_ASYNC_RESILVER 0x10 #define SPA_ASYNC_AUTOEXPAND 0x20 #define SPA_ASYNC_REMOVE_DONE 0x40 #define SPA_ASYNC_REMOVE_STOP 0x80 #define SPA_ASYNC_INITIALIZE_RESTART 0x100 #define SPA_ASYNC_TRIM_RESTART 0x200 #define SPA_ASYNC_AUTOTRIM_RESTART 0x400 #define SPA_ASYNC_L2CACHE_REBUILD 0x800 #define SPA_ASYNC_L2CACHE_TRIM 0x1000 #define SPA_ASYNC_REBUILD_DONE 0x2000 #define SPA_ASYNC_DETACH_SPARE 0x4000 /* device manipulation */ extern int spa_vdev_add(spa_t *spa, nvlist_t *nvroot); extern int spa_vdev_attach(spa_t *spa, uint64_t guid, nvlist_t *nvroot, int replacing, int rebuild); extern int spa_vdev_detach(spa_t *spa, uint64_t guid, uint64_t pguid, int replace_done); extern int spa_vdev_alloc(spa_t *spa, uint64_t guid); extern int spa_vdev_noalloc(spa_t *spa, uint64_t guid); extern boolean_t spa_vdev_remove_active(spa_t *spa); extern int spa_vdev_initialize(spa_t *spa, nvlist_t *nv, uint64_t cmd_type, nvlist_t *vdev_errlist); extern int spa_vdev_trim(spa_t *spa, nvlist_t *nv, uint64_t cmd_type, uint64_t rate, boolean_t partial, boolean_t secure, nvlist_t *vdev_errlist); extern int spa_vdev_setpath(spa_t *spa, uint64_t guid, const char *newpath); extern int spa_vdev_setfru(spa_t *spa, uint64_t guid, const char *newfru); extern int spa_vdev_split_mirror(spa_t *spa, const char *newname, nvlist_t *config, nvlist_t *props, boolean_t exp); /* spare state (which is global across all pools) */ extern void spa_spare_add(vdev_t *vd); extern void spa_spare_remove(vdev_t *vd); extern boolean_t spa_spare_exists(uint64_t guid, uint64_t *pool, int *refcnt); extern void spa_spare_activate(vdev_t *vd); /* L2ARC state (which is global across all pools) */ extern void spa_l2cache_add(vdev_t *vd); extern void spa_l2cache_remove(vdev_t *vd); extern boolean_t spa_l2cache_exists(uint64_t guid, uint64_t *pool); extern void spa_l2cache_activate(vdev_t *vd); extern void spa_l2cache_drop(spa_t *spa); /* scanning */ extern int spa_scan(spa_t *spa, pool_scan_func_t func); extern int spa_scan_stop(spa_t *spa); extern int spa_scrub_pause_resume(spa_t *spa, pool_scrub_cmd_t flag); /* spa syncing */ extern void spa_sync(spa_t *spa, uint64_t txg); /* only for DMU use */ extern void spa_sync_allpools(void); extern uint_t zfs_sync_pass_deferred_free; /* spa namespace global mutex */ extern kmutex_t spa_namespace_lock; /* * SPA configuration functions in spa_config.c */ #define SPA_CONFIG_UPDATE_POOL 0 #define SPA_CONFIG_UPDATE_VDEVS 1 extern void spa_write_cachefile(spa_t *, boolean_t, boolean_t, boolean_t); extern void spa_config_load(void); extern nvlist_t *spa_all_configs(uint64_t *); extern void spa_config_set(spa_t *spa, nvlist_t *config); extern nvlist_t *spa_config_generate(spa_t *spa, vdev_t *vd, uint64_t txg, int getstats); extern void spa_config_update(spa_t *spa, int what); extern int spa_config_parse(spa_t *spa, vdev_t **vdp, nvlist_t *nv, vdev_t *parent, uint_t id, int atype); /* * Miscellaneous SPA routines in spa_misc.c */ /* Namespace manipulation */ extern spa_t *spa_lookup(const char *name); extern spa_t *spa_add(const char *name, nvlist_t *config, const char *altroot); extern void spa_remove(spa_t *spa); extern spa_t *spa_next(spa_t *prev); /* Refcount functions */ extern void spa_open_ref(spa_t *spa, const void *tag); extern void spa_close(spa_t *spa, const void *tag); extern void spa_async_close(spa_t *spa, const void *tag); extern boolean_t spa_refcount_zero(spa_t *spa); #define SCL_NONE 0x00 #define SCL_CONFIG 0x01 #define SCL_STATE 0x02 #define SCL_L2ARC 0x04 /* hack until L2ARC 2.0 */ #define SCL_ALLOC 0x08 #define SCL_ZIO 0x10 #define SCL_FREE 0x20 #define SCL_VDEV 0x40 #define SCL_LOCKS 7 #define SCL_ALL ((1 << SCL_LOCKS) - 1) #define SCL_STATE_ALL (SCL_STATE | SCL_L2ARC | SCL_ZIO) /* Historical pool statistics */ typedef struct spa_history_kstat { kmutex_t lock; uint64_t count; uint64_t size; kstat_t *kstat; void *priv; list_t list; } spa_history_kstat_t; typedef struct spa_history_list { uint64_t size; procfs_list_t procfs_list; } spa_history_list_t; typedef struct spa_stats { spa_history_list_t read_history; spa_history_list_t txg_history; spa_history_kstat_t tx_assign_histogram; spa_history_list_t mmp_history; spa_history_kstat_t state; /* pool state */ spa_history_kstat_t guid; /* pool guid */ spa_history_kstat_t iostats; } spa_stats_t; typedef enum txg_state { TXG_STATE_BIRTH = 0, TXG_STATE_OPEN = 1, TXG_STATE_QUIESCED = 2, TXG_STATE_WAIT_FOR_SYNC = 3, TXG_STATE_SYNCED = 4, TXG_STATE_COMMITTED = 5, } txg_state_t; typedef struct txg_stat { vdev_stat_t vs1; vdev_stat_t vs2; uint64_t txg; uint64_t ndirty; } txg_stat_t; /* Assorted pool IO kstats */ typedef struct spa_iostats { kstat_named_t trim_extents_written; kstat_named_t trim_bytes_written; kstat_named_t trim_extents_skipped; kstat_named_t trim_bytes_skipped; kstat_named_t trim_extents_failed; kstat_named_t trim_bytes_failed; kstat_named_t autotrim_extents_written; kstat_named_t autotrim_bytes_written; kstat_named_t autotrim_extents_skipped; kstat_named_t autotrim_bytes_skipped; kstat_named_t autotrim_extents_failed; kstat_named_t autotrim_bytes_failed; kstat_named_t simple_trim_extents_written; kstat_named_t simple_trim_bytes_written; kstat_named_t simple_trim_extents_skipped; kstat_named_t simple_trim_bytes_skipped; kstat_named_t simple_trim_extents_failed; kstat_named_t simple_trim_bytes_failed; } spa_iostats_t; extern void spa_stats_init(spa_t *spa); extern void spa_stats_destroy(spa_t *spa); extern void spa_read_history_add(spa_t *spa, const zbookmark_phys_t *zb, uint32_t aflags); extern void spa_txg_history_add(spa_t *spa, uint64_t txg, hrtime_t birth_time); extern int spa_txg_history_set(spa_t *spa, uint64_t txg, txg_state_t completed_state, hrtime_t completed_time); extern txg_stat_t *spa_txg_history_init_io(spa_t *, uint64_t, struct dsl_pool *); extern void spa_txg_history_fini_io(spa_t *, txg_stat_t *); extern void spa_tx_assign_add_nsecs(spa_t *spa, uint64_t nsecs); extern int spa_mmp_history_set_skip(spa_t *spa, uint64_t mmp_kstat_id); extern int spa_mmp_history_set(spa_t *spa, uint64_t mmp_kstat_id, int io_error, hrtime_t duration); extern void spa_mmp_history_add(spa_t *spa, uint64_t txg, uint64_t timestamp, uint64_t mmp_delay, vdev_t *vd, int label, uint64_t mmp_kstat_id, int error); extern void spa_iostats_trim_add(spa_t *spa, trim_type_t type, uint64_t extents_written, uint64_t bytes_written, uint64_t extents_skipped, uint64_t bytes_skipped, uint64_t extents_failed, uint64_t bytes_failed); extern void spa_import_progress_add(spa_t *spa); extern void spa_import_progress_remove(uint64_t spa_guid); extern int spa_import_progress_set_mmp_check(uint64_t pool_guid, uint64_t mmp_sec_remaining); extern int spa_import_progress_set_max_txg(uint64_t pool_guid, uint64_t max_txg); extern int spa_import_progress_set_state(uint64_t pool_guid, spa_load_state_t spa_load_state); /* Pool configuration locks */ extern int spa_config_tryenter(spa_t *spa, int locks, const void *tag, krw_t rw); extern void spa_config_enter(spa_t *spa, int locks, const void *tag, krw_t rw); extern void spa_config_enter_mmp(spa_t *spa, int locks, const void *tag, krw_t rw); extern void spa_config_exit(spa_t *spa, int locks, const void *tag); extern int spa_config_held(spa_t *spa, int locks, krw_t rw); /* Pool vdev add/remove lock */ extern uint64_t spa_vdev_enter(spa_t *spa); extern uint64_t spa_vdev_detach_enter(spa_t *spa, uint64_t guid); extern uint64_t spa_vdev_config_enter(spa_t *spa); extern void spa_vdev_config_exit(spa_t *spa, vdev_t *vd, uint64_t txg, int error, const char *tag); extern int spa_vdev_exit(spa_t *spa, vdev_t *vd, uint64_t txg, int error); /* Pool vdev state change lock */ extern void spa_vdev_state_enter(spa_t *spa, int oplock); extern int spa_vdev_state_exit(spa_t *spa, vdev_t *vd, int error); /* Log state */ typedef enum spa_log_state { SPA_LOG_UNKNOWN = 0, /* unknown log state */ SPA_LOG_MISSING, /* missing log(s) */ SPA_LOG_CLEAR, /* clear the log(s) */ SPA_LOG_GOOD, /* log(s) are good */ } spa_log_state_t; extern spa_log_state_t spa_get_log_state(spa_t *spa); extern void spa_set_log_state(spa_t *spa, spa_log_state_t state); extern int spa_reset_logs(spa_t *spa); /* Log claim callback */ extern void spa_claim_notify(zio_t *zio); extern void spa_deadman(void *); /* Accessor functions */ extern boolean_t spa_shutting_down(spa_t *spa); extern struct dsl_pool *spa_get_dsl(spa_t *spa); extern boolean_t spa_is_initializing(spa_t *spa); extern boolean_t spa_indirect_vdevs_loaded(spa_t *spa); extern blkptr_t *spa_get_rootblkptr(spa_t *spa); extern void spa_set_rootblkptr(spa_t *spa, const blkptr_t *bp); extern void spa_altroot(spa_t *, char *, size_t); extern uint32_t spa_sync_pass(spa_t *spa); extern char *spa_name(spa_t *spa); extern uint64_t spa_guid(spa_t *spa); extern uint64_t spa_load_guid(spa_t *spa); extern uint64_t spa_last_synced_txg(spa_t *spa); extern uint64_t spa_first_txg(spa_t *spa); extern uint64_t spa_syncing_txg(spa_t *spa); extern uint64_t spa_final_dirty_txg(spa_t *spa); extern uint64_t spa_version(spa_t *spa); extern pool_state_t spa_state(spa_t *spa); extern spa_load_state_t spa_load_state(spa_t *spa); extern uint64_t spa_freeze_txg(spa_t *spa); extern uint64_t spa_get_worst_case_asize(spa_t *spa, uint64_t lsize); extern uint64_t spa_get_dspace(spa_t *spa); extern uint64_t spa_get_checkpoint_space(spa_t *spa); extern uint64_t spa_get_slop_space(spa_t *spa); extern void spa_update_dspace(spa_t *spa); extern uint64_t spa_version(spa_t *spa); extern boolean_t spa_deflate(spa_t *spa); extern metaslab_class_t *spa_normal_class(spa_t *spa); extern metaslab_class_t *spa_log_class(spa_t *spa); extern metaslab_class_t *spa_embedded_log_class(spa_t *spa); extern metaslab_class_t *spa_special_class(spa_t *spa); extern metaslab_class_t *spa_dedup_class(spa_t *spa); extern metaslab_class_t *spa_preferred_class(spa_t *spa, uint64_t size, dmu_object_type_t objtype, uint_t level, uint_t special_smallblk); extern void spa_evicting_os_register(spa_t *, objset_t *os); extern void spa_evicting_os_deregister(spa_t *, objset_t *os); extern void spa_evicting_os_wait(spa_t *spa); extern int spa_max_replication(spa_t *spa); extern int spa_prev_software_version(spa_t *spa); extern uint64_t spa_get_failmode(spa_t *spa); extern uint64_t spa_get_deadman_failmode(spa_t *spa); extern void spa_set_deadman_failmode(spa_t *spa, const char *failmode); extern boolean_t spa_suspended(spa_t *spa); extern uint64_t spa_bootfs(spa_t *spa); extern uint64_t spa_delegation(spa_t *spa); extern objset_t *spa_meta_objset(spa_t *spa); extern space_map_t *spa_syncing_log_sm(spa_t *spa); extern uint64_t spa_deadman_synctime(spa_t *spa); extern uint64_t spa_deadman_ziotime(spa_t *spa); extern uint64_t spa_dirty_data(spa_t *spa); extern spa_autotrim_t spa_get_autotrim(spa_t *spa); /* Miscellaneous support routines */ extern void spa_load_failed(spa_t *spa, const char *fmt, ...) __attribute__((format(printf, 2, 3))); extern void spa_load_note(spa_t *spa, const char *fmt, ...) __attribute__((format(printf, 2, 3))); extern void spa_activate_mos_feature(spa_t *spa, const char *feature, dmu_tx_t *tx); extern void spa_deactivate_mos_feature(spa_t *spa, const char *feature); extern spa_t *spa_by_guid(uint64_t pool_guid, uint64_t device_guid); extern boolean_t spa_guid_exists(uint64_t pool_guid, uint64_t device_guid); extern char *spa_strdup(const char *); extern void spa_strfree(char *); extern uint64_t spa_generate_guid(spa_t *spa); extern void snprintf_blkptr(char *buf, size_t buflen, const blkptr_t *bp); extern void spa_freeze(spa_t *spa); extern int spa_change_guid(spa_t *spa); extern void spa_upgrade(spa_t *spa, uint64_t version); extern void spa_evict_all(void); extern vdev_t *spa_lookup_by_guid(spa_t *spa, uint64_t guid, boolean_t l2cache); extern boolean_t spa_has_l2cache(spa_t *, uint64_t guid); extern boolean_t spa_has_spare(spa_t *, uint64_t guid); extern uint64_t dva_get_dsize_sync(spa_t *spa, const dva_t *dva); extern uint64_t bp_get_dsize_sync(spa_t *spa, const blkptr_t *bp); extern uint64_t bp_get_dsize(spa_t *spa, const blkptr_t *bp); extern boolean_t spa_has_slogs(spa_t *spa); extern boolean_t spa_is_root(spa_t *spa); extern boolean_t spa_writeable(spa_t *spa); extern boolean_t spa_has_pending_synctask(spa_t *spa); extern int spa_maxblocksize(spa_t *spa); extern int spa_maxdnodesize(spa_t *spa); extern boolean_t spa_has_checkpoint(spa_t *spa); extern boolean_t spa_importing_readonly_checkpoint(spa_t *spa); extern boolean_t spa_suspend_async_destroy(spa_t *spa); extern uint64_t spa_min_claim_txg(spa_t *spa); extern boolean_t zfs_dva_valid(spa_t *spa, const dva_t *dva, const blkptr_t *bp); typedef void (*spa_remap_cb_t)(uint64_t vdev, uint64_t offset, uint64_t size, void *arg); extern boolean_t spa_remap_blkptr(spa_t *spa, blkptr_t *bp, spa_remap_cb_t callback, void *arg); extern uint64_t spa_get_last_removal_txg(spa_t *spa); extern boolean_t spa_trust_config(spa_t *spa); extern uint64_t spa_missing_tvds_allowed(spa_t *spa); extern void spa_set_missing_tvds(spa_t *spa, uint64_t missing); extern boolean_t spa_top_vdevs_spacemap_addressable(spa_t *spa); extern uint64_t spa_total_metaslabs(spa_t *spa); extern boolean_t spa_multihost(spa_t *spa); extern uint32_t spa_get_hostid(spa_t *spa); extern void spa_activate_allocation_classes(spa_t *, dmu_tx_t *); extern boolean_t spa_livelist_delete_check(spa_t *spa); extern spa_mode_t spa_mode(spa_t *spa); extern uint64_t zfs_strtonum(const char *str, char **nptr); extern char *spa_his_ievent_table[]; extern void spa_history_create_obj(spa_t *spa, dmu_tx_t *tx); extern int spa_history_get(spa_t *spa, uint64_t *offset, uint64_t *len_read, char *his_buf); extern int spa_history_log(spa_t *spa, const char *his_buf); extern int spa_history_log_nvl(spa_t *spa, nvlist_t *nvl); extern void spa_history_log_version(spa_t *spa, const char *operation, dmu_tx_t *tx); extern void spa_history_log_internal(spa_t *spa, const char *operation, dmu_tx_t *tx, const char *fmt, ...) __printflike(4, 5); extern void spa_history_log_internal_ds(struct dsl_dataset *ds, const char *op, dmu_tx_t *tx, const char *fmt, ...) __printflike(4, 5); extern void spa_history_log_internal_dd(dsl_dir_t *dd, const char *operation, dmu_tx_t *tx, const char *fmt, ...) __printflike(4, 5); extern const char *spa_state_to_name(spa_t *spa); /* error handling */ struct zbookmark_phys; extern void spa_log_error(spa_t *spa, const zbookmark_phys_t *zb, const uint64_t *birth); extern void spa_remove_error(spa_t *spa, zbookmark_phys_t *zb, const uint64_t *birth); extern int zfs_ereport_post(const char *clazz, spa_t *spa, vdev_t *vd, const zbookmark_phys_t *zb, zio_t *zio, uint64_t state); extern boolean_t zfs_ereport_is_valid(const char *clazz, spa_t *spa, vdev_t *vd, zio_t *zio); extern void zfs_ereport_taskq_fini(void); extern void zfs_ereport_clear(spa_t *spa, vdev_t *vd); extern nvlist_t *zfs_event_create(spa_t *spa, vdev_t *vd, const char *type, const char *name, nvlist_t *aux); extern void zfs_post_remove(spa_t *spa, vdev_t *vd); extern void zfs_post_state_change(spa_t *spa, vdev_t *vd, uint64_t laststate); extern void zfs_post_autoreplace(spa_t *spa, vdev_t *vd); extern uint64_t spa_approx_errlog_size(spa_t *spa); extern int spa_get_errlog(spa_t *spa, void *uaddr, uint64_t *count); +extern uint64_t spa_get_last_errlog_size(spa_t *spa); extern void spa_errlog_rotate(spa_t *spa); extern void spa_errlog_drain(spa_t *spa); extern void spa_errlog_sync(spa_t *spa, uint64_t txg); extern void spa_get_errlists(spa_t *spa, avl_tree_t *last, avl_tree_t *scrub); extern void spa_delete_dataset_errlog(spa_t *spa, uint64_t ds, dmu_tx_t *tx); extern void spa_swap_errlog(spa_t *spa, uint64_t new_head_ds, uint64_t old_head_ds, dmu_tx_t *tx); extern void sync_error_list(spa_t *spa, avl_tree_t *t, uint64_t *obj, dmu_tx_t *tx); extern void spa_upgrade_errlog(spa_t *spa, dmu_tx_t *tx); +extern int find_top_affected_fs(spa_t *spa, uint64_t head_ds, + zbookmark_err_phys_t *zep, uint64_t *top_affected_fs); +extern int find_birth_txg(struct dsl_dataset *ds, zbookmark_err_phys_t *zep, + uint64_t *birth_txg); +extern void zep_to_zb(uint64_t dataset, zbookmark_err_phys_t *zep, + zbookmark_phys_t *zb); +extern void name_to_errphys(char *buf, zbookmark_err_phys_t *zep); /* vdev cache */ extern void vdev_cache_stat_init(void); extern void vdev_cache_stat_fini(void); /* vdev mirror */ extern void vdev_mirror_stat_init(void); extern void vdev_mirror_stat_fini(void); /* Initialization and termination */ extern void spa_init(spa_mode_t mode); extern void spa_fini(void); extern void spa_boot_init(void); /* properties */ extern int spa_prop_set(spa_t *spa, nvlist_t *nvp); extern int spa_prop_get(spa_t *spa, nvlist_t **nvp); extern void spa_prop_clear_bootfs(spa_t *spa, uint64_t obj, dmu_tx_t *tx); extern void spa_configfile_set(spa_t *, nvlist_t *, boolean_t); /* asynchronous event notification */ extern void spa_event_notify(spa_t *spa, vdev_t *vdev, nvlist_t *hist_nvl, const char *name); extern void zfs_ereport_zvol_post(const char *subclass, const char *name, const char *device_name, const char *raw_name); /* waiting for pool activities to complete */ extern int spa_wait(const char *pool, zpool_wait_activity_t activity, boolean_t *waited); extern int spa_wait_tag(const char *name, zpool_wait_activity_t activity, uint64_t tag, boolean_t *waited); extern void spa_notify_waiters(spa_t *spa); extern void spa_wake_waiters(spa_t *spa); extern void spa_import_os(spa_t *spa); extern void spa_export_os(spa_t *spa); extern void spa_activate_os(spa_t *spa); extern void spa_deactivate_os(spa_t *spa); /* module param call functions */ int param_set_deadman_ziotime(ZFS_MODULE_PARAM_ARGS); int param_set_deadman_synctime(ZFS_MODULE_PARAM_ARGS); int param_set_slop_shift(ZFS_MODULE_PARAM_ARGS); int param_set_deadman_failmode(ZFS_MODULE_PARAM_ARGS); #ifdef ZFS_DEBUG #define dprintf_bp(bp, fmt, ...) do { \ if (zfs_flags & ZFS_DEBUG_DPRINTF) { \ char *__blkbuf = kmem_alloc(BP_SPRINTF_LEN, KM_SLEEP); \ snprintf_blkptr(__blkbuf, BP_SPRINTF_LEN, (bp)); \ dprintf(fmt " %s\n", __VA_ARGS__, __blkbuf); \ kmem_free(__blkbuf, BP_SPRINTF_LEN); \ } \ } while (0) #else #define dprintf_bp(bp, fmt, ...) #endif extern spa_mode_t spa_mode_global; extern int zfs_deadman_enabled; extern uint64_t zfs_deadman_synctime_ms; extern uint64_t zfs_deadman_ziotime_ms; extern uint64_t zfs_deadman_checktime_ms; extern kmem_cache_t *zio_buf_cache[]; extern kmem_cache_t *zio_data_buf_cache[]; #ifdef __cplusplus } #endif #endif /* _SYS_SPA_H */ diff --git a/sys/contrib/openzfs/include/sys/spa_impl.h b/sys/contrib/openzfs/include/sys/spa_impl.h index 5782c54bd78f..44afa763283a 100644 --- a/sys/contrib/openzfs/include/sys/spa_impl.h +++ b/sys/contrib/openzfs/include/sys/spa_impl.h @@ -1,470 +1,474 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2011, 2019 by Delphix. All rights reserved. * Copyright 2011 Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2014 Spectra Logic Corporation, All rights reserved. * Copyright 2013 Saso Kiselkov. All rights reserved. * Copyright (c) 2016 Actifio, Inc. All rights reserved. * Copyright (c) 2017, Intel Corporation. * Copyright (c) 2019 Datto Inc. */ #ifndef _SYS_SPA_IMPL_H #define _SYS_SPA_IMPL_H #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef __cplusplus extern "C" { #endif typedef struct spa_alloc { kmutex_t spaa_lock; avl_tree_t spaa_tree; } ____cacheline_aligned spa_alloc_t; typedef struct spa_error_entry { zbookmark_phys_t se_bookmark; char *se_name; avl_node_t se_avl; zbookmark_err_phys_t se_zep; /* not accounted in avl_find */ } spa_error_entry_t; typedef struct spa_history_phys { uint64_t sh_pool_create_len; /* ending offset of zpool create */ uint64_t sh_phys_max_off; /* physical EOF */ uint64_t sh_bof; /* logical BOF */ uint64_t sh_eof; /* logical EOF */ uint64_t sh_records_lost; /* num of records overwritten */ } spa_history_phys_t; /* * All members must be uint64_t, for byteswap purposes. */ typedef struct spa_removing_phys { uint64_t sr_state; /* dsl_scan_state_t */ /* * The vdev ID that we most recently attempted to remove, * or -1 if no removal has been attempted. */ uint64_t sr_removing_vdev; /* * The vdev ID that we most recently successfully removed, * or -1 if no devices have been removed. */ uint64_t sr_prev_indirect_vdev; uint64_t sr_start_time; uint64_t sr_end_time; /* * Note that we can not use the space map's or indirect mapping's * accounting as a substitute for these values, because we need to * count frees of not-yet-copied data as though it did the copy. * Otherwise, we could get into a situation where copied > to_copy, * or we complete before copied == to_copy. */ uint64_t sr_to_copy; /* bytes that need to be copied */ uint64_t sr_copied; /* bytes that have been copied or freed */ } spa_removing_phys_t; /* * This struct is stored as an entry in the DMU_POOL_DIRECTORY_OBJECT * (with key DMU_POOL_CONDENSING_INDIRECT). It is present if a condense * of an indirect vdev's mapping object is in progress. */ typedef struct spa_condensing_indirect_phys { /* * The vdev ID of the indirect vdev whose indirect mapping is * being condensed. */ uint64_t scip_vdev; /* * The vdev's old obsolete spacemap. This spacemap's contents are * being integrated into the new mapping. */ uint64_t scip_prev_obsolete_sm_object; /* * The new mapping object that is being created. */ uint64_t scip_next_mapping_object; } spa_condensing_indirect_phys_t; struct spa_aux_vdev { uint64_t sav_object; /* MOS object for device list */ nvlist_t *sav_config; /* cached device config */ vdev_t **sav_vdevs; /* devices */ int sav_count; /* number devices */ boolean_t sav_sync; /* sync the device list */ nvlist_t **sav_pending; /* pending device additions */ uint_t sav_npending; /* # pending devices */ }; typedef struct spa_config_lock { kmutex_t scl_lock; kthread_t *scl_writer; int scl_write_wanted; int scl_count; kcondvar_t scl_cv; } ____cacheline_aligned spa_config_lock_t; typedef struct spa_config_dirent { list_node_t scd_link; char *scd_path; } spa_config_dirent_t; typedef enum zio_taskq_type { ZIO_TASKQ_ISSUE = 0, ZIO_TASKQ_ISSUE_HIGH, ZIO_TASKQ_INTERRUPT, ZIO_TASKQ_INTERRUPT_HIGH, ZIO_TASKQ_TYPES } zio_taskq_type_t; /* * State machine for the zpool-poolname process. The states transitions * are done as follows: * * From To Routine * PROC_NONE -> PROC_CREATED spa_activate() * PROC_CREATED -> PROC_ACTIVE spa_thread() * PROC_ACTIVE -> PROC_DEACTIVATE spa_deactivate() * PROC_DEACTIVATE -> PROC_GONE spa_thread() * PROC_GONE -> PROC_NONE spa_deactivate() */ typedef enum spa_proc_state { SPA_PROC_NONE, /* spa_proc = &p0, no process created */ SPA_PROC_CREATED, /* spa_activate() has proc, is waiting */ SPA_PROC_ACTIVE, /* taskqs created, spa_proc set */ SPA_PROC_DEACTIVATE, /* spa_deactivate() requests process exit */ SPA_PROC_GONE /* spa_thread() is exiting, spa_proc = &p0 */ } spa_proc_state_t; typedef struct spa_taskqs { uint_t stqs_count; taskq_t **stqs_taskq; } spa_taskqs_t; typedef enum spa_all_vdev_zap_action { AVZ_ACTION_NONE = 0, AVZ_ACTION_DESTROY, /* Destroy all per-vdev ZAPs and the AVZ. */ AVZ_ACTION_REBUILD, /* Populate the new AVZ, see spa_avz_rebuild */ AVZ_ACTION_INITIALIZE } spa_avz_action_t; typedef enum spa_config_source { SPA_CONFIG_SRC_NONE = 0, SPA_CONFIG_SRC_SCAN, /* scan of path (default: /dev/dsk) */ SPA_CONFIG_SRC_CACHEFILE, /* any cachefile */ SPA_CONFIG_SRC_TRYIMPORT, /* returned from call to tryimport */ SPA_CONFIG_SRC_SPLIT, /* new pool in a pool split */ SPA_CONFIG_SRC_MOS /* MOS, but not always from right txg */ } spa_config_source_t; struct spa { /* * Fields protected by spa_namespace_lock. */ char spa_name[ZFS_MAX_DATASET_NAME_LEN]; /* pool name */ char *spa_comment; /* comment */ avl_node_t spa_avl; /* node in spa_namespace_avl */ nvlist_t *spa_config; /* last synced config */ nvlist_t *spa_config_syncing; /* currently syncing config */ nvlist_t *spa_config_splitting; /* config for splitting */ nvlist_t *spa_load_info; /* info and errors from load */ uint64_t spa_config_txg; /* txg of last config change */ uint32_t spa_sync_pass; /* iterate-to-convergence */ pool_state_t spa_state; /* pool state */ int spa_inject_ref; /* injection references */ uint8_t spa_sync_on; /* sync threads are running */ spa_load_state_t spa_load_state; /* current load operation */ boolean_t spa_indirect_vdevs_loaded; /* mappings loaded? */ boolean_t spa_trust_config; /* do we trust vdev tree? */ boolean_t spa_is_splitting; /* in the middle of a split? */ spa_config_source_t spa_config_source; /* where config comes from? */ uint64_t spa_import_flags; /* import specific flags */ spa_taskqs_t spa_zio_taskq[ZIO_TYPES][ZIO_TASKQ_TYPES]; dsl_pool_t *spa_dsl_pool; boolean_t spa_is_initializing; /* true while opening pool */ boolean_t spa_is_exporting; /* true while exporting pool */ metaslab_class_t *spa_normal_class; /* normal data class */ metaslab_class_t *spa_log_class; /* intent log data class */ metaslab_class_t *spa_embedded_log_class; /* log on normal vdevs */ metaslab_class_t *spa_special_class; /* special allocation class */ metaslab_class_t *spa_dedup_class; /* dedup allocation class */ uint64_t spa_first_txg; /* first txg after spa_open() */ uint64_t spa_final_txg; /* txg of export/destroy */ uint64_t spa_freeze_txg; /* freeze pool at this txg */ uint64_t spa_load_max_txg; /* best initial ub_txg */ uint64_t spa_claim_max_txg; /* highest claimed birth txg */ inode_timespec_t spa_loaded_ts; /* 1st successful open time */ objset_t *spa_meta_objset; /* copy of dp->dp_meta_objset */ kmutex_t spa_evicting_os_lock; /* Evicting objset list lock */ list_t spa_evicting_os_list; /* Objsets being evicted. */ kcondvar_t spa_evicting_os_cv; /* Objset Eviction Completion */ txg_list_t spa_vdev_txg_list; /* per-txg dirty vdev list */ vdev_t *spa_root_vdev; /* top-level vdev container */ uint64_t spa_min_ashift; /* of vdevs in normal class */ uint64_t spa_max_ashift; /* of vdevs in normal class */ uint64_t spa_min_alloc; /* of vdevs in normal class */ uint64_t spa_config_guid; /* config pool guid */ uint64_t spa_load_guid; /* spa_load initialized guid */ uint64_t spa_last_synced_guid; /* last synced guid */ list_t spa_config_dirty_list; /* vdevs with dirty config */ list_t spa_state_dirty_list; /* vdevs with dirty state */ /* * spa_allocs is an array, whose lengths is stored in spa_alloc_count. * There is one tree and one lock for each allocator, to help improve * allocation performance in write-heavy workloads. */ spa_alloc_t *spa_allocs; int spa_alloc_count; spa_aux_vdev_t spa_spares; /* hot spares */ spa_aux_vdev_t spa_l2cache; /* L2ARC cache devices */ nvlist_t *spa_label_features; /* Features for reading MOS */ uint64_t spa_config_object; /* MOS object for pool config */ uint64_t spa_config_generation; /* config generation number */ uint64_t spa_syncing_txg; /* txg currently syncing */ bpobj_t spa_deferred_bpobj; /* deferred-free bplist */ bplist_t spa_free_bplist[TXG_SIZE]; /* bplist of stuff to free */ zio_cksum_salt_t spa_cksum_salt; /* secret salt for cksum */ /* checksum context templates */ kmutex_t spa_cksum_tmpls_lock; void *spa_cksum_tmpls[ZIO_CHECKSUM_FUNCTIONS]; uberblock_t spa_ubsync; /* last synced uberblock */ uberblock_t spa_uberblock; /* current uberblock */ boolean_t spa_extreme_rewind; /* rewind past deferred frees */ kmutex_t spa_scrub_lock; /* resilver/scrub lock */ uint64_t spa_scrub_inflight; /* in-flight scrub bytes */ /* in-flight verification bytes */ uint64_t spa_load_verify_bytes; kcondvar_t spa_scrub_io_cv; /* scrub I/O completion */ uint8_t spa_scrub_active; /* active or suspended? */ uint8_t spa_scrub_type; /* type of scrub we're doing */ uint8_t spa_scrub_finished; /* indicator to rotate logs */ uint8_t spa_scrub_started; /* started since last boot */ uint8_t spa_scrub_reopen; /* scrub doing vdev_reopen */ uint64_t spa_scan_pass_start; /* start time per pass/reboot */ uint64_t spa_scan_pass_scrub_pause; /* scrub pause time */ uint64_t spa_scan_pass_scrub_spent_paused; /* total paused */ uint64_t spa_scan_pass_exam; /* examined bytes per pass */ uint64_t spa_scan_pass_issued; /* issued bytes per pass */ + /* error scrub pause time in milliseconds */ + uint64_t spa_scan_pass_errorscrub_pause; + /* total error scrub paused time in milliseconds */ + uint64_t spa_scan_pass_errorscrub_spent_paused; /* * We are in the middle of a resilver, and another resilver * is needed once this one completes. This is set iff any * vdev_resilver_deferred is set. */ boolean_t spa_resilver_deferred; kmutex_t spa_async_lock; /* protect async state */ kthread_t *spa_async_thread; /* thread doing async task */ int spa_async_suspended; /* async tasks suspended */ kcondvar_t spa_async_cv; /* wait for thread_exit() */ uint16_t spa_async_tasks; /* async task mask */ uint64_t spa_missing_tvds; /* unopenable tvds on load */ uint64_t spa_missing_tvds_allowed; /* allow loading spa? */ uint64_t spa_nonallocating_dspace; spa_removing_phys_t spa_removing_phys; spa_vdev_removal_t *spa_vdev_removal; spa_condensing_indirect_phys_t spa_condensing_indirect_phys; spa_condensing_indirect_t *spa_condensing_indirect; zthr_t *spa_condense_zthr; /* zthr doing condense. */ uint64_t spa_checkpoint_txg; /* the txg of the checkpoint */ spa_checkpoint_info_t spa_checkpoint_info; /* checkpoint accounting */ zthr_t *spa_checkpoint_discard_zthr; space_map_t *spa_syncing_log_sm; /* current log space map */ avl_tree_t spa_sm_logs_by_txg; kmutex_t spa_flushed_ms_lock; /* for metaslabs_by_flushed */ avl_tree_t spa_metaslabs_by_flushed; spa_unflushed_stats_t spa_unflushed_stats; list_t spa_log_summary; uint64_t spa_log_flushall_txg; zthr_t *spa_livelist_delete_zthr; /* deleting livelists */ zthr_t *spa_livelist_condense_zthr; /* condensing livelists */ uint64_t spa_livelists_to_delete; /* set of livelists to free */ livelist_condense_entry_t spa_to_condense; /* next to condense */ char *spa_root; /* alternate root directory */ uint64_t spa_ena; /* spa-wide ereport ENA */ int spa_last_open_failed; /* error if last open failed */ uint64_t spa_last_ubsync_txg; /* "best" uberblock txg */ uint64_t spa_last_ubsync_txg_ts; /* timestamp from that ub */ uint64_t spa_load_txg; /* ub txg that loaded */ uint64_t spa_load_txg_ts; /* timestamp from that ub */ uint64_t spa_load_meta_errors; /* verify metadata err count */ uint64_t spa_load_data_errors; /* verify data err count */ uint64_t spa_verify_min_txg; /* start txg of verify scrub */ kmutex_t spa_errlog_lock; /* error log lock */ uint64_t spa_errlog_last; /* last error log object */ uint64_t spa_errlog_scrub; /* scrub error log object */ kmutex_t spa_errlist_lock; /* error list/ereport lock */ avl_tree_t spa_errlist_last; /* last error list */ avl_tree_t spa_errlist_scrub; /* scrub error list */ avl_tree_t spa_errlist_healed; /* list of healed blocks */ uint64_t spa_deflate; /* should we deflate? */ uint64_t spa_history; /* history object */ kmutex_t spa_history_lock; /* history lock */ vdev_t *spa_pending_vdev; /* pending vdev additions */ kmutex_t spa_props_lock; /* property lock */ uint64_t spa_pool_props_object; /* object for properties */ uint64_t spa_bootfs; /* default boot filesystem */ uint64_t spa_failmode; /* failure mode for the pool */ uint64_t spa_deadman_failmode; /* failure mode for deadman */ uint64_t spa_delegation; /* delegation on/off */ list_t spa_config_list; /* previous cache file(s) */ /* per-CPU array of root of async I/O: */ zio_t **spa_async_zio_root; zio_t *spa_suspend_zio_root; /* root of all suspended I/O */ zio_t *spa_txg_zio[TXG_SIZE]; /* spa_sync() waits for this */ kmutex_t spa_suspend_lock; /* protects suspend_zio_root */ kcondvar_t spa_suspend_cv; /* notification of resume */ zio_suspend_reason_t spa_suspended; /* pool is suspended */ uint8_t spa_claiming; /* pool is doing zil_claim() */ boolean_t spa_is_root; /* pool is root */ int spa_minref; /* num refs when first opened */ spa_mode_t spa_mode; /* SPA_MODE_{READ|WRITE} */ boolean_t spa_read_spacemaps; /* spacemaps available if ro */ spa_log_state_t spa_log_state; /* log state */ uint64_t spa_autoexpand; /* lun expansion on/off */ ddt_t *spa_ddt[ZIO_CHECKSUM_FUNCTIONS]; /* in-core DDTs */ uint64_t spa_ddt_stat_object; /* DDT statistics */ uint64_t spa_dedup_dspace; /* Cache get_dedup_dspace() */ uint64_t spa_dedup_checksum; /* default dedup checksum */ uint64_t spa_dspace; /* dspace in normal class */ struct brt *spa_brt; /* in-core BRT */ kmutex_t spa_vdev_top_lock; /* dueling offline/remove */ kmutex_t spa_proc_lock; /* protects spa_proc* */ kcondvar_t spa_proc_cv; /* spa_proc_state transitions */ spa_proc_state_t spa_proc_state; /* see definition */ proc_t *spa_proc; /* "zpool-poolname" process */ uintptr_t spa_did; /* if procp != p0, did of t1 */ boolean_t spa_autoreplace; /* autoreplace set in open */ int spa_vdev_locks; /* locks grabbed */ uint64_t spa_creation_version; /* version at pool creation */ uint64_t spa_prev_software_version; /* See ub_software_version */ uint64_t spa_feat_for_write_obj; /* required to write to pool */ uint64_t spa_feat_for_read_obj; /* required to read from pool */ uint64_t spa_feat_desc_obj; /* Feature descriptions */ uint64_t spa_feat_enabled_txg_obj; /* Feature enabled txg */ kmutex_t spa_feat_stats_lock; /* protects spa_feat_stats */ nvlist_t *spa_feat_stats; /* Cache of enabled features */ /* cache feature refcounts */ uint64_t spa_feat_refcount_cache[SPA_FEATURES]; taskqid_t spa_deadman_tqid; /* Task id */ uint64_t spa_deadman_calls; /* number of deadman calls */ hrtime_t spa_sync_starttime; /* starting time of spa_sync */ uint64_t spa_deadman_synctime; /* deadman sync expiration */ uint64_t spa_deadman_ziotime; /* deadman zio expiration */ uint64_t spa_all_vdev_zaps; /* ZAP of per-vd ZAP obj #s */ spa_avz_action_t spa_avz_action; /* destroy/rebuild AVZ? */ uint64_t spa_autotrim; /* automatic background trim? */ uint64_t spa_errata; /* errata issues detected */ spa_stats_t spa_stats; /* assorted spa statistics */ spa_keystore_t spa_keystore; /* loaded crypto keys */ /* arc_memory_throttle() parameters during low memory condition */ uint64_t spa_lowmem_page_load; /* memory load during txg */ uint64_t spa_lowmem_last_txg; /* txg window start */ hrtime_t spa_ccw_fail_time; /* Conf cache write fail time */ taskq_t *spa_zvol_taskq; /* Taskq for minor management */ taskq_t *spa_prefetch_taskq; /* Taskq for prefetch threads */ uint64_t spa_multihost; /* multihost aware (mmp) */ mmp_thread_t spa_mmp; /* multihost mmp thread */ list_t spa_leaf_list; /* list of leaf vdevs */ uint64_t spa_leaf_list_gen; /* track leaf_list changes */ uint32_t spa_hostid; /* cached system hostid */ /* synchronization for threads in spa_wait */ kmutex_t spa_activities_lock; kcondvar_t spa_activities_cv; kcondvar_t spa_waiters_cv; int spa_waiters; /* number of waiting threads */ boolean_t spa_waiters_cancel; /* waiters should return */ char *spa_compatibility; /* compatibility file(s) */ /* * spa_refcount & spa_config_lock must be the last elements * because zfs_refcount_t changes size based on compilation options. * In order for the MDB module to function correctly, the other * fields must remain in the same location. */ spa_config_lock_t spa_config_lock[SCL_LOCKS]; /* config changes */ zfs_refcount_t spa_refcount; /* number of opens */ taskq_t *spa_upgrade_taskq; /* taskq for upgrade jobs */ }; extern char *spa_config_path; extern const char *zfs_deadman_failmode; extern uint_t spa_slop_shift; extern void spa_taskq_dispatch_ent(spa_t *spa, zio_type_t t, zio_taskq_type_t q, task_func_t *func, void *arg, uint_t flags, taskq_ent_t *ent); extern void spa_taskq_dispatch_sync(spa_t *, zio_type_t t, zio_taskq_type_t q, task_func_t *func, void *arg, uint_t flags); extern void spa_load_spares(spa_t *spa); extern void spa_load_l2cache(spa_t *spa); extern sysevent_t *spa_event_create(spa_t *spa, vdev_t *vd, nvlist_t *hist_nvl, const char *name); extern void spa_event_post(sysevent_t *ev); extern int param_set_deadman_failmode_common(const char *val); extern void spa_set_deadman_synctime(hrtime_t ns); extern void spa_set_deadman_ziotime(hrtime_t ns); extern const char *spa_history_zone(void); #ifdef __cplusplus } #endif #endif /* _SYS_SPA_IMPL_H */ diff --git a/sys/contrib/openzfs/include/sys/sysevent/eventdefs.h b/sys/contrib/openzfs/include/sys/sysevent/eventdefs.h index eb1dfd16c0fd..a21085257967 100644 --- a/sys/contrib/openzfs/include/sys/sysevent/eventdefs.h +++ b/sys/contrib/openzfs/include/sys/sysevent/eventdefs.h @@ -1,136 +1,141 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright 2015 Nexenta Systems, Inc. All rights reserved. * Copyright 2017 Joyent, Inc. */ #ifndef _SYS_SYSEVENT_EVENTDEFS_H #define _SYS_SYSEVENT_EVENTDEFS_H #ifdef __cplusplus extern "C" { #endif /* * eventdefs.h contains public definitions for sysevent types (classes * and subclasses). All additions/removal/changes are subject * to PSARC approval. */ /* Sysevent Class definitions */ #define EC_NONE "EC_none" #define EC_PRIV "EC_priv" #define EC_PLATFORM "EC_platform" /* events private to platform */ #define EC_DR "EC_dr" /* Dynamic reconfiguration event class */ #define EC_ENV "EC_env" /* Environmental monitor event class */ #define EC_DOMAIN "EC_domain" /* Domain event class */ #define EC_AP_DRIVER "EC_ap_driver" /* Alternate Pathing event class */ #define EC_IPMP "EC_ipmp" /* IP Multipathing event class */ #define EC_DEV_ADD "EC_dev_add" /* device add event class */ #define EC_DEV_REMOVE "EC_dev_remove" /* device remove event class */ #define EC_DEV_BRANCH "EC_dev_branch" /* device tree branch event class */ #define EC_DEV_STATUS "EC_dev_status" /* device status event class */ #define EC_FM "EC_fm" /* FMA error report event */ #define EC_ZFS "EC_zfs" /* ZFS event */ #define EC_DATALINK "EC_datalink" /* datalink event */ #define EC_VRRP "EC_vrrp" /* VRRP event */ /* * EC_DEV_ADD and EC_DEV_REMOVE subclass definitions - supporting attributes * (name/value pairs) are found in sys/sysevent/dev.h */ #define ESC_DISK "disk" /* disk device */ #define ESC_NETWORK "network" /* network interface */ #define ESC_PRINTER "printer" /* printer device */ #define ESC_LOFI "lofi" /* lofi device */ /* * EC_DEV_BRANCH subclass definitions - supporting attributes (name/value pairs) * are found in sys/sysevent/dev.h */ /* device tree branch added */ #define ESC_DEV_BRANCH_ADD "dev_branch_add" /* device tree branch removed */ #define ESC_DEV_BRANCH_REMOVE "dev_branch_remove" /* * EC_DEV_STATUS subclass definitions * * device capacity dynamically changed */ #define ESC_DEV_DLE "dev_dle" /* LUN has received an eject request from the user */ #define ESC_DEV_EJECT_REQUEST "dev_eject_request" /* FMA Fault and Error event protocol subclass */ #define ESC_FM_ERROR "error" #define ESC_FM_ERROR_REPLAY "error_replay" /* * ZFS subclass definitions. supporting attributes (name/value paris) are found * in sys/fs/zfs.h */ #define ESC_ZFS_RESILVER_START "resilver_start" #define ESC_ZFS_RESILVER_FINISH "resilver_finish" #define ESC_ZFS_VDEV_REMOVE "vdev_remove" #define ESC_ZFS_VDEV_REMOVE_AUX "vdev_remove_aux" #define ESC_ZFS_VDEV_REMOVE_DEV "vdev_remove_dev" #define ESC_ZFS_POOL_CREATE "pool_create" #define ESC_ZFS_POOL_DESTROY "pool_destroy" #define ESC_ZFS_POOL_IMPORT "pool_import" #define ESC_ZFS_POOL_EXPORT "pool_export" #define ESC_ZFS_VDEV_ADD "vdev_add" #define ESC_ZFS_VDEV_ATTACH "vdev_attach" #define ESC_ZFS_VDEV_CLEAR "vdev_clear" #define ESC_ZFS_VDEV_CHECK "vdev_check" #define ESC_ZFS_VDEV_ONLINE "vdev_online" #define ESC_ZFS_CONFIG_SYNC "config_sync" #define ESC_ZFS_SCRUB_START "scrub_start" #define ESC_ZFS_SCRUB_FINISH "scrub_finish" #define ESC_ZFS_SCRUB_ABORT "scrub_abort" #define ESC_ZFS_SCRUB_RESUME "scrub_resume" #define ESC_ZFS_SCRUB_PAUSED "scrub_paused" #define ESC_ZFS_VDEV_SPARE "vdev_spare" #define ESC_ZFS_VDEV_AUTOEXPAND "vdev_autoexpand" #define ESC_ZFS_BOOTFS_VDEV_ATTACH "bootfs_vdev_attach" #define ESC_ZFS_POOL_REGUID "pool_reguid" #define ESC_ZFS_HISTORY_EVENT "history_event" #define ESC_ZFS_TRIM_START "trim_start" #define ESC_ZFS_TRIM_FINISH "trim_finish" #define ESC_ZFS_TRIM_CANCEL "trim_cancel" #define ESC_ZFS_TRIM_RESUME "trim_resume" #define ESC_ZFS_TRIM_SUSPEND "trim_suspend" +#define ESC_ZFS_ERRORSCRUB_START "errorscrub_start" +#define ESC_ZFS_ERRORSCRUB_FINISH "errorscrub_finish" +#define ESC_ZFS_ERRORSCRUB_ABORT "errorscrub_abort" +#define ESC_ZFS_ERRORSCRUB_RESUME "errorscrub_resume" +#define ESC_ZFS_ERRORSCRUB_PAUSED "errorscrub_paused" /* * datalink subclass definitions. */ #define ESC_DATALINK_PHYS_ADD "datalink_phys_add" /* new physical link */ #ifdef __cplusplus } #endif #endif /* _SYS_SYSEVENT_EVENTDEFS_H */ diff --git a/sys/contrib/openzfs/include/sys/vdev_initialize.h b/sys/contrib/openzfs/include/sys/vdev_initialize.h index 4e63f063cb66..78702b7325a0 100644 --- a/sys/contrib/openzfs/include/sys/vdev_initialize.h +++ b/sys/contrib/openzfs/include/sys/vdev_initialize.h @@ -1,47 +1,48 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2016 by Delphix. All rights reserved. */ #ifndef _SYS_VDEV_INITIALIZE_H #define _SYS_VDEV_INITIALIZE_H #include #ifdef __cplusplus extern "C" { #endif extern void vdev_initialize(vdev_t *vd); +extern void vdev_uninitialize(vdev_t *vd); extern void vdev_initialize_stop(vdev_t *vd, vdev_initializing_state_t tgt_state, list_t *vd_list); extern void vdev_initialize_stop_all(vdev_t *vd, vdev_initializing_state_t tgt_state); extern void vdev_initialize_stop_wait(spa_t *spa, list_t *vd_list); extern void vdev_initialize_restart(vdev_t *vd); #ifdef __cplusplus } #endif #endif /* _SYS_VDEV_INITIALIZE_H */ diff --git a/sys/contrib/openzfs/lib/libzfs/libzfs.abi b/sys/contrib/openzfs/lib/libzfs/libzfs.abi index 732863dcffc7..6e53bcb41a87 100644 --- a/sys/contrib/openzfs/lib/libzfs/libzfs.abi +++ b/sys/contrib/openzfs/lib/libzfs/libzfs.abi @@ -1,9414 +1,9416 @@ - + + - + + diff --git a/sys/contrib/openzfs/lib/libzfs/libzfs_dataset.c b/sys/contrib/openzfs/lib/libzfs/libzfs_dataset.c index 138eca19acc3..fe9f3268d338 100644 --- a/sys/contrib/openzfs/lib/libzfs/libzfs_dataset.c +++ b/sys/contrib/openzfs/lib/libzfs/libzfs_dataset.c @@ -1,5634 +1,5640 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright 2019 Joyent, Inc. * Copyright (c) 2011, 2020 by Delphix. All rights reserved. * Copyright (c) 2012 DEY Storage Systems, Inc. All rights reserved. * Copyright (c) 2012 Pawel Jakub Dawidek . * Copyright (c) 2013 Martin Matuska. All rights reserved. * Copyright (c) 2013 Steven Hartland. All rights reserved. * Copyright 2017 Nexenta Systems, Inc. * Copyright 2016 Igor Kozhukhov * Copyright 2017-2018 RackTop Systems. * Copyright (c) 2019 Datto Inc. * Copyright (c) 2019, loli10K * Copyright (c) 2021 Matt Fiddaman */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef HAVE_IDMAP #include #include #include #endif /* HAVE_IDMAP */ #include #include #include #include #include #include #include "zfs_namecheck.h" #include "zfs_prop.h" #include "libzfs_impl.h" #include "zfs_deleg.h" static int userquota_propname_decode(const char *propname, boolean_t zoned, zfs_userquota_prop_t *typep, char *domain, int domainlen, uint64_t *ridp); /* * Given a single type (not a mask of types), return the type in a human * readable form. */ const char * zfs_type_to_name(zfs_type_t type) { switch (type) { case ZFS_TYPE_FILESYSTEM: return (dgettext(TEXT_DOMAIN, "filesystem")); case ZFS_TYPE_SNAPSHOT: return (dgettext(TEXT_DOMAIN, "snapshot")); case ZFS_TYPE_VOLUME: return (dgettext(TEXT_DOMAIN, "volume")); case ZFS_TYPE_POOL: return (dgettext(TEXT_DOMAIN, "pool")); case ZFS_TYPE_BOOKMARK: return (dgettext(TEXT_DOMAIN, "bookmark")); default: assert(!"unhandled zfs_type_t"); } return (NULL); } /* * Validate a ZFS path. This is used even before trying to open the dataset, to * provide a more meaningful error message. We call zfs_error_aux() to * explain exactly why the name was not valid. */ int zfs_validate_name(libzfs_handle_t *hdl, const char *path, int type, boolean_t modifying) { namecheck_err_t why; char what; if (!(type & ZFS_TYPE_SNAPSHOT) && strchr(path, '@') != NULL) { if (hdl != NULL) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "snapshot delimiter '@' is not expected here")); return (0); } if (type == ZFS_TYPE_SNAPSHOT && strchr(path, '@') == NULL) { if (hdl != NULL) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "missing '@' delimiter in snapshot name")); return (0); } if (!(type & ZFS_TYPE_BOOKMARK) && strchr(path, '#') != NULL) { if (hdl != NULL) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "bookmark delimiter '#' is not expected here")); return (0); } if (type == ZFS_TYPE_BOOKMARK && strchr(path, '#') == NULL) { if (hdl != NULL) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "missing '#' delimiter in bookmark name")); return (0); } if (modifying && strchr(path, '%') != NULL) { if (hdl != NULL) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid character %c in name"), '%'); return (0); } if (entity_namecheck(path, &why, &what) != 0) { if (hdl != NULL) { switch (why) { case NAME_ERR_TOOLONG: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "name is too long")); break; case NAME_ERR_LEADING_SLASH: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "leading slash in name")); break; case NAME_ERR_EMPTY_COMPONENT: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "empty component or misplaced '@'" " or '#' delimiter in name")); break; case NAME_ERR_TRAILING_SLASH: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "trailing slash in name")); break; case NAME_ERR_INVALCHAR: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid character " "'%c' in name"), what); break; case NAME_ERR_MULTIPLE_DELIMITERS: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "multiple '@' and/or '#' delimiters in " "name")); break; case NAME_ERR_NOLETTER: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool doesn't begin with a letter")); break; case NAME_ERR_RESERVED: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "name is reserved")); break; case NAME_ERR_DISKLIKE: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "reserved disk name")); break; case NAME_ERR_SELF_REF: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "self reference, '.' is found in name")); break; case NAME_ERR_PARENT_REF: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "parent reference, '..' is found in name")); break; default: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "(%d) not defined"), why); break; } } return (0); } return (-1); } int zfs_name_valid(const char *name, zfs_type_t type) { if (type == ZFS_TYPE_POOL) return (zpool_name_valid(NULL, B_FALSE, name)); return (zfs_validate_name(NULL, name, type, B_FALSE)); } /* * This function takes the raw DSL properties, and filters out the user-defined * properties into a separate nvlist. */ static nvlist_t * process_user_props(zfs_handle_t *zhp, nvlist_t *props) { libzfs_handle_t *hdl = zhp->zfs_hdl; nvpair_t *elem; nvlist_t *nvl; if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0) { (void) no_memory(hdl); return (NULL); } elem = NULL; while ((elem = nvlist_next_nvpair(props, elem)) != NULL) { if (!zfs_prop_user(nvpair_name(elem))) continue; nvlist_t *propval = fnvpair_value_nvlist(elem); if (nvlist_add_nvlist(nvl, nvpair_name(elem), propval) != 0) { nvlist_free(nvl); (void) no_memory(hdl); return (NULL); } } return (nvl); } static zpool_handle_t * zpool_add_handle(zfs_handle_t *zhp, const char *pool_name) { libzfs_handle_t *hdl = zhp->zfs_hdl; zpool_handle_t *zph; if ((zph = zpool_open_canfail(hdl, pool_name)) != NULL) { if (hdl->libzfs_pool_handles != NULL) zph->zpool_next = hdl->libzfs_pool_handles; hdl->libzfs_pool_handles = zph; } return (zph); } static zpool_handle_t * zpool_find_handle(zfs_handle_t *zhp, const char *pool_name, int len) { libzfs_handle_t *hdl = zhp->zfs_hdl; zpool_handle_t *zph = hdl->libzfs_pool_handles; while ((zph != NULL) && (strncmp(pool_name, zpool_get_name(zph), len) != 0)) zph = zph->zpool_next; return (zph); } /* * Returns a handle to the pool that contains the provided dataset. * If a handle to that pool already exists then that handle is returned. * Otherwise, a new handle is created and added to the list of handles. */ static zpool_handle_t * zpool_handle(zfs_handle_t *zhp) { char *pool_name; int len; zpool_handle_t *zph; len = strcspn(zhp->zfs_name, "/@#") + 1; pool_name = zfs_alloc(zhp->zfs_hdl, len); (void) strlcpy(pool_name, zhp->zfs_name, len); zph = zpool_find_handle(zhp, pool_name, len); if (zph == NULL) zph = zpool_add_handle(zhp, pool_name); free(pool_name); return (zph); } void zpool_free_handles(libzfs_handle_t *hdl) { zpool_handle_t *next, *zph = hdl->libzfs_pool_handles; while (zph != NULL) { next = zph->zpool_next; zpool_close(zph); zph = next; } hdl->libzfs_pool_handles = NULL; } /* * Utility function to gather stats (objset and zpl) for the given object. */ static int get_stats_ioctl(zfs_handle_t *zhp, zfs_cmd_t *zc) { libzfs_handle_t *hdl = zhp->zfs_hdl; (void) strlcpy(zc->zc_name, zhp->zfs_name, sizeof (zc->zc_name)); while (zfs_ioctl(hdl, ZFS_IOC_OBJSET_STATS, zc) != 0) { if (errno == ENOMEM) zcmd_expand_dst_nvlist(hdl, zc); else return (-1); } return (0); } /* * Utility function to get the received properties of the given object. */ static int get_recvd_props_ioctl(zfs_handle_t *zhp) { libzfs_handle_t *hdl = zhp->zfs_hdl; nvlist_t *recvdprops; zfs_cmd_t zc = {"\0"}; int err; zcmd_alloc_dst_nvlist(hdl, &zc, 0); (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); while (zfs_ioctl(hdl, ZFS_IOC_OBJSET_RECVD_PROPS, &zc) != 0) { if (errno == ENOMEM) zcmd_expand_dst_nvlist(hdl, &zc); else { zcmd_free_nvlists(&zc); return (-1); } } err = zcmd_read_dst_nvlist(zhp->zfs_hdl, &zc, &recvdprops); zcmd_free_nvlists(&zc); if (err != 0) return (-1); nvlist_free(zhp->zfs_recvd_props); zhp->zfs_recvd_props = recvdprops; return (0); } static int put_stats_zhdl(zfs_handle_t *zhp, zfs_cmd_t *zc) { nvlist_t *allprops, *userprops; zhp->zfs_dmustats = zc->zc_objset_stats; /* structure assignment */ if (zcmd_read_dst_nvlist(zhp->zfs_hdl, zc, &allprops) != 0) { return (-1); } /* * XXX Why do we store the user props separately, in addition to * storing them in zfs_props? */ if ((userprops = process_user_props(zhp, allprops)) == NULL) { nvlist_free(allprops); return (-1); } nvlist_free(zhp->zfs_props); nvlist_free(zhp->zfs_user_props); zhp->zfs_props = allprops; zhp->zfs_user_props = userprops; return (0); } static int get_stats(zfs_handle_t *zhp) { int rc = 0; zfs_cmd_t zc = {"\0"}; zcmd_alloc_dst_nvlist(zhp->zfs_hdl, &zc, 0); if (get_stats_ioctl(zhp, &zc) != 0) rc = -1; else if (put_stats_zhdl(zhp, &zc) != 0) rc = -1; zcmd_free_nvlists(&zc); return (rc); } /* * Refresh the properties currently stored in the handle. */ void zfs_refresh_properties(zfs_handle_t *zhp) { (void) get_stats(zhp); } /* * Makes a handle from the given dataset name. Used by zfs_open() and * zfs_iter_* to create child handles on the fly. */ static int make_dataset_handle_common(zfs_handle_t *zhp, zfs_cmd_t *zc) { if (put_stats_zhdl(zhp, zc) != 0) return (-1); /* * We've managed to open the dataset and gather statistics. Determine * the high-level type. */ if (zhp->zfs_dmustats.dds_type == DMU_OST_ZVOL) { zhp->zfs_head_type = ZFS_TYPE_VOLUME; } else if (zhp->zfs_dmustats.dds_type == DMU_OST_ZFS) { zhp->zfs_head_type = ZFS_TYPE_FILESYSTEM; } else if (zhp->zfs_dmustats.dds_type == DMU_OST_OTHER) { errno = EINVAL; return (-1); } else if (zhp->zfs_dmustats.dds_inconsistent) { errno = EBUSY; return (-1); } else { abort(); } if (zhp->zfs_dmustats.dds_is_snapshot) zhp->zfs_type = ZFS_TYPE_SNAPSHOT; else if (zhp->zfs_dmustats.dds_type == DMU_OST_ZVOL) zhp->zfs_type = ZFS_TYPE_VOLUME; else if (zhp->zfs_dmustats.dds_type == DMU_OST_ZFS) zhp->zfs_type = ZFS_TYPE_FILESYSTEM; else abort(); /* we should never see any other types */ if ((zhp->zpool_hdl = zpool_handle(zhp)) == NULL) return (-1); return (0); } zfs_handle_t * make_dataset_handle(libzfs_handle_t *hdl, const char *path) { zfs_cmd_t zc = {"\0"}; zfs_handle_t *zhp = calloc(1, sizeof (zfs_handle_t)); if (zhp == NULL) return (NULL); zhp->zfs_hdl = hdl; (void) strlcpy(zhp->zfs_name, path, sizeof (zhp->zfs_name)); zcmd_alloc_dst_nvlist(hdl, &zc, 0); if (get_stats_ioctl(zhp, &zc) == -1) { zcmd_free_nvlists(&zc); free(zhp); return (NULL); } if (make_dataset_handle_common(zhp, &zc) == -1) { free(zhp); zhp = NULL; } zcmd_free_nvlists(&zc); return (zhp); } zfs_handle_t * make_dataset_handle_zc(libzfs_handle_t *hdl, zfs_cmd_t *zc) { zfs_handle_t *zhp = calloc(1, sizeof (zfs_handle_t)); if (zhp == NULL) return (NULL); zhp->zfs_hdl = hdl; (void) strlcpy(zhp->zfs_name, zc->zc_name, sizeof (zhp->zfs_name)); if (make_dataset_handle_common(zhp, zc) == -1) { free(zhp); return (NULL); } return (zhp); } zfs_handle_t * make_dataset_simple_handle_zc(zfs_handle_t *pzhp, zfs_cmd_t *zc) { zfs_handle_t *zhp = calloc(1, sizeof (zfs_handle_t)); if (zhp == NULL) return (NULL); zhp->zfs_hdl = pzhp->zfs_hdl; (void) strlcpy(zhp->zfs_name, zc->zc_name, sizeof (zhp->zfs_name)); zhp->zfs_head_type = pzhp->zfs_type; zhp->zfs_type = ZFS_TYPE_SNAPSHOT; zhp->zpool_hdl = zpool_handle(zhp); if (zc->zc_objset_stats.dds_creation_txg != 0) { /* structure assignment */ zhp->zfs_dmustats = zc->zc_objset_stats; } else { if (get_stats_ioctl(zhp, zc) == -1) { zcmd_free_nvlists(zc); free(zhp); return (NULL); } if (make_dataset_handle_common(zhp, zc) == -1) { zcmd_free_nvlists(zc); free(zhp); return (NULL); } } if (zhp->zfs_dmustats.dds_is_snapshot || strchr(zc->zc_name, '@') != NULL) zhp->zfs_type = ZFS_TYPE_SNAPSHOT; else if (zhp->zfs_dmustats.dds_type == DMU_OST_ZVOL) zhp->zfs_type = ZFS_TYPE_VOLUME; else if (zhp->zfs_dmustats.dds_type == DMU_OST_ZFS) zhp->zfs_type = ZFS_TYPE_FILESYSTEM; return (zhp); } zfs_handle_t * zfs_handle_dup(zfs_handle_t *zhp_orig) { zfs_handle_t *zhp = calloc(1, sizeof (zfs_handle_t)); if (zhp == NULL) return (NULL); zhp->zfs_hdl = zhp_orig->zfs_hdl; zhp->zpool_hdl = zhp_orig->zpool_hdl; (void) strlcpy(zhp->zfs_name, zhp_orig->zfs_name, sizeof (zhp->zfs_name)); zhp->zfs_type = zhp_orig->zfs_type; zhp->zfs_head_type = zhp_orig->zfs_head_type; zhp->zfs_dmustats = zhp_orig->zfs_dmustats; if (zhp_orig->zfs_props != NULL) { if (nvlist_dup(zhp_orig->zfs_props, &zhp->zfs_props, 0) != 0) { (void) no_memory(zhp->zfs_hdl); zfs_close(zhp); return (NULL); } } if (zhp_orig->zfs_user_props != NULL) { if (nvlist_dup(zhp_orig->zfs_user_props, &zhp->zfs_user_props, 0) != 0) { (void) no_memory(zhp->zfs_hdl); zfs_close(zhp); return (NULL); } } if (zhp_orig->zfs_recvd_props != NULL) { if (nvlist_dup(zhp_orig->zfs_recvd_props, &zhp->zfs_recvd_props, 0)) { (void) no_memory(zhp->zfs_hdl); zfs_close(zhp); return (NULL); } } zhp->zfs_mntcheck = zhp_orig->zfs_mntcheck; if (zhp_orig->zfs_mntopts != NULL) { zhp->zfs_mntopts = zfs_strdup(zhp_orig->zfs_hdl, zhp_orig->zfs_mntopts); } zhp->zfs_props_table = zhp_orig->zfs_props_table; return (zhp); } boolean_t zfs_bookmark_exists(const char *path) { nvlist_t *bmarks; nvlist_t *props; char fsname[ZFS_MAX_DATASET_NAME_LEN]; char *bmark_name; char *pound; int err; boolean_t rv; (void) strlcpy(fsname, path, sizeof (fsname)); pound = strchr(fsname, '#'); if (pound == NULL) return (B_FALSE); *pound = '\0'; bmark_name = pound + 1; props = fnvlist_alloc(); err = lzc_get_bookmarks(fsname, props, &bmarks); nvlist_free(props); if (err != 0) { nvlist_free(bmarks); return (B_FALSE); } rv = nvlist_exists(bmarks, bmark_name); nvlist_free(bmarks); return (rv); } zfs_handle_t * make_bookmark_handle(zfs_handle_t *parent, const char *path, nvlist_t *bmark_props) { zfs_handle_t *zhp = calloc(1, sizeof (zfs_handle_t)); if (zhp == NULL) return (NULL); /* Fill in the name. */ zhp->zfs_hdl = parent->zfs_hdl; (void) strlcpy(zhp->zfs_name, path, sizeof (zhp->zfs_name)); /* Set the property lists. */ if (nvlist_dup(bmark_props, &zhp->zfs_props, 0) != 0) { free(zhp); return (NULL); } /* Set the types. */ zhp->zfs_head_type = parent->zfs_head_type; zhp->zfs_type = ZFS_TYPE_BOOKMARK; if ((zhp->zpool_hdl = zpool_handle(zhp)) == NULL) { nvlist_free(zhp->zfs_props); free(zhp); return (NULL); } return (zhp); } struct zfs_open_bookmarks_cb_data { const char *path; zfs_handle_t *zhp; }; static int zfs_open_bookmarks_cb(zfs_handle_t *zhp, void *data) { struct zfs_open_bookmarks_cb_data *dp = data; /* * Is it the one we are looking for? */ if (strcmp(dp->path, zfs_get_name(zhp)) == 0) { /* * We found it. Save it and let the caller know we are done. */ dp->zhp = zhp; return (EEXIST); } /* * Not found. Close the handle and ask for another one. */ zfs_close(zhp); return (0); } /* * Opens the given snapshot, bookmark, filesystem, or volume. The 'types' * argument is a mask of acceptable types. The function will print an * appropriate error message and return NULL if it can't be opened. */ zfs_handle_t * zfs_open(libzfs_handle_t *hdl, const char *path, int types) { zfs_handle_t *zhp; char errbuf[ERRBUFLEN]; char *bookp; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot open '%s'"), path); /* * Validate the name before we even try to open it. */ if (!zfs_validate_name(hdl, path, types, B_FALSE)) { (void) zfs_error(hdl, EZFS_INVALIDNAME, errbuf); errno = EINVAL; return (NULL); } /* * Bookmarks needs to be handled separately. */ bookp = strchr(path, '#'); if (bookp == NULL) { /* * Try to get stats for the dataset, which will tell us if it * exists. */ errno = 0; if ((zhp = make_dataset_handle(hdl, path)) == NULL) { (void) zfs_standard_error(hdl, errno, errbuf); return (NULL); } } else { char dsname[ZFS_MAX_DATASET_NAME_LEN]; zfs_handle_t *pzhp; struct zfs_open_bookmarks_cb_data cb_data = {path, NULL}; /* * We need to cut out '#' and everything after '#' * to get the parent dataset name only. */ assert(bookp - path < sizeof (dsname)); (void) strlcpy(dsname, path, MIN(sizeof (dsname), bookp - path + 1)); /* * Create handle for the parent dataset. */ errno = 0; if ((pzhp = make_dataset_handle(hdl, dsname)) == NULL) { (void) zfs_standard_error(hdl, errno, errbuf); return (NULL); } /* * Iterate bookmarks to find the right one. */ errno = 0; if ((zfs_iter_bookmarks_v2(pzhp, 0, zfs_open_bookmarks_cb, &cb_data) == 0) && (cb_data.zhp == NULL)) { (void) zfs_error(hdl, EZFS_NOENT, errbuf); zfs_close(pzhp); errno = ENOENT; return (NULL); } if (cb_data.zhp == NULL) { (void) zfs_standard_error(hdl, errno, errbuf); zfs_close(pzhp); return (NULL); } zhp = cb_data.zhp; /* * Cleanup. */ zfs_close(pzhp); } if (!(types & zhp->zfs_type)) { (void) zfs_error(hdl, EZFS_BADTYPE, errbuf); zfs_close(zhp); errno = EINVAL; return (NULL); } return (zhp); } /* * Release a ZFS handle. Nothing to do but free the associated memory. */ void zfs_close(zfs_handle_t *zhp) { if (zhp->zfs_mntopts) free(zhp->zfs_mntopts); nvlist_free(zhp->zfs_props); nvlist_free(zhp->zfs_user_props); nvlist_free(zhp->zfs_recvd_props); free(zhp); } typedef struct mnttab_node { struct mnttab mtn_mt; avl_node_t mtn_node; } mnttab_node_t; static int libzfs_mnttab_cache_compare(const void *arg1, const void *arg2) { const mnttab_node_t *mtn1 = (const mnttab_node_t *)arg1; const mnttab_node_t *mtn2 = (const mnttab_node_t *)arg2; int rv; rv = strcmp(mtn1->mtn_mt.mnt_special, mtn2->mtn_mt.mnt_special); return (TREE_ISIGN(rv)); } void libzfs_mnttab_init(libzfs_handle_t *hdl) { pthread_mutex_init(&hdl->libzfs_mnttab_cache_lock, NULL); assert(avl_numnodes(&hdl->libzfs_mnttab_cache) == 0); avl_create(&hdl->libzfs_mnttab_cache, libzfs_mnttab_cache_compare, sizeof (mnttab_node_t), offsetof(mnttab_node_t, mtn_node)); } static int libzfs_mnttab_update(libzfs_handle_t *hdl) { FILE *mnttab; struct mnttab entry; if ((mnttab = fopen(MNTTAB, "re")) == NULL) return (ENOENT); while (getmntent(mnttab, &entry) == 0) { mnttab_node_t *mtn; avl_index_t where; if (strcmp(entry.mnt_fstype, MNTTYPE_ZFS) != 0) continue; mtn = zfs_alloc(hdl, sizeof (mnttab_node_t)); mtn->mtn_mt.mnt_special = zfs_strdup(hdl, entry.mnt_special); mtn->mtn_mt.mnt_mountp = zfs_strdup(hdl, entry.mnt_mountp); mtn->mtn_mt.mnt_fstype = zfs_strdup(hdl, entry.mnt_fstype); mtn->mtn_mt.mnt_mntopts = zfs_strdup(hdl, entry.mnt_mntopts); /* Exclude duplicate mounts */ if (avl_find(&hdl->libzfs_mnttab_cache, mtn, &where) != NULL) { free(mtn->mtn_mt.mnt_special); free(mtn->mtn_mt.mnt_mountp); free(mtn->mtn_mt.mnt_fstype); free(mtn->mtn_mt.mnt_mntopts); free(mtn); continue; } avl_add(&hdl->libzfs_mnttab_cache, mtn); } (void) fclose(mnttab); return (0); } void libzfs_mnttab_fini(libzfs_handle_t *hdl) { void *cookie = NULL; mnttab_node_t *mtn; while ((mtn = avl_destroy_nodes(&hdl->libzfs_mnttab_cache, &cookie)) != NULL) { free(mtn->mtn_mt.mnt_special); free(mtn->mtn_mt.mnt_mountp); free(mtn->mtn_mt.mnt_fstype); free(mtn->mtn_mt.mnt_mntopts); free(mtn); } avl_destroy(&hdl->libzfs_mnttab_cache); (void) pthread_mutex_destroy(&hdl->libzfs_mnttab_cache_lock); } void libzfs_mnttab_cache(libzfs_handle_t *hdl, boolean_t enable) { hdl->libzfs_mnttab_enable = enable; } int libzfs_mnttab_find(libzfs_handle_t *hdl, const char *fsname, struct mnttab *entry) { FILE *mnttab; mnttab_node_t find; mnttab_node_t *mtn; int ret = ENOENT; if (!hdl->libzfs_mnttab_enable) { struct mnttab srch = { 0 }; if (avl_numnodes(&hdl->libzfs_mnttab_cache)) libzfs_mnttab_fini(hdl); if ((mnttab = fopen(MNTTAB, "re")) == NULL) return (ENOENT); srch.mnt_special = (char *)fsname; srch.mnt_fstype = (char *)MNTTYPE_ZFS; ret = getmntany(mnttab, entry, &srch) ? ENOENT : 0; (void) fclose(mnttab); return (ret); } pthread_mutex_lock(&hdl->libzfs_mnttab_cache_lock); if (avl_numnodes(&hdl->libzfs_mnttab_cache) == 0) { int error; if ((error = libzfs_mnttab_update(hdl)) != 0) { pthread_mutex_unlock(&hdl->libzfs_mnttab_cache_lock); return (error); } } find.mtn_mt.mnt_special = (char *)fsname; mtn = avl_find(&hdl->libzfs_mnttab_cache, &find, NULL); if (mtn) { *entry = mtn->mtn_mt; ret = 0; } pthread_mutex_unlock(&hdl->libzfs_mnttab_cache_lock); return (ret); } void libzfs_mnttab_add(libzfs_handle_t *hdl, const char *special, const char *mountp, const char *mntopts) { mnttab_node_t *mtn; pthread_mutex_lock(&hdl->libzfs_mnttab_cache_lock); if (avl_numnodes(&hdl->libzfs_mnttab_cache) != 0) { mtn = zfs_alloc(hdl, sizeof (mnttab_node_t)); mtn->mtn_mt.mnt_special = zfs_strdup(hdl, special); mtn->mtn_mt.mnt_mountp = zfs_strdup(hdl, mountp); mtn->mtn_mt.mnt_fstype = zfs_strdup(hdl, MNTTYPE_ZFS); mtn->mtn_mt.mnt_mntopts = zfs_strdup(hdl, mntopts); /* * Another thread may have already added this entry * via libzfs_mnttab_update. If so we should skip it. */ if (avl_find(&hdl->libzfs_mnttab_cache, mtn, NULL) != NULL) { free(mtn->mtn_mt.mnt_special); free(mtn->mtn_mt.mnt_mountp); free(mtn->mtn_mt.mnt_fstype); free(mtn->mtn_mt.mnt_mntopts); free(mtn); } else { avl_add(&hdl->libzfs_mnttab_cache, mtn); } } pthread_mutex_unlock(&hdl->libzfs_mnttab_cache_lock); } void libzfs_mnttab_remove(libzfs_handle_t *hdl, const char *fsname) { mnttab_node_t find; mnttab_node_t *ret; pthread_mutex_lock(&hdl->libzfs_mnttab_cache_lock); find.mtn_mt.mnt_special = (char *)fsname; if ((ret = avl_find(&hdl->libzfs_mnttab_cache, (void *)&find, NULL)) != NULL) { avl_remove(&hdl->libzfs_mnttab_cache, ret); free(ret->mtn_mt.mnt_special); free(ret->mtn_mt.mnt_mountp); free(ret->mtn_mt.mnt_fstype); free(ret->mtn_mt.mnt_mntopts); free(ret); } pthread_mutex_unlock(&hdl->libzfs_mnttab_cache_lock); } int zfs_spa_version(zfs_handle_t *zhp, int *spa_version) { zpool_handle_t *zpool_handle = zhp->zpool_hdl; if (zpool_handle == NULL) return (-1); *spa_version = zpool_get_prop_int(zpool_handle, ZPOOL_PROP_VERSION, NULL); return (0); } /* * The choice of reservation property depends on the SPA version. */ static int zfs_which_resv_prop(zfs_handle_t *zhp, zfs_prop_t *resv_prop) { int spa_version; if (zfs_spa_version(zhp, &spa_version) < 0) return (-1); if (spa_version >= SPA_VERSION_REFRESERVATION) *resv_prop = ZFS_PROP_REFRESERVATION; else *resv_prop = ZFS_PROP_RESERVATION; return (0); } /* * Given an nvlist of properties to set, validates that they are correct, and * parses any numeric properties (index, boolean, etc) if they are specified as * strings. */ nvlist_t * zfs_valid_proplist(libzfs_handle_t *hdl, zfs_type_t type, nvlist_t *nvl, uint64_t zoned, zfs_handle_t *zhp, zpool_handle_t *zpool_hdl, boolean_t key_params_ok, const char *errbuf) { nvpair_t *elem; uint64_t intval; const char *strval; zfs_prop_t prop; nvlist_t *ret; int chosen_normal = -1; int chosen_utf = -1; + int set_maxbs = 0; if (nvlist_alloc(&ret, NV_UNIQUE_NAME, 0) != 0) { (void) no_memory(hdl); return (NULL); } /* * Make sure this property is valid and applies to this type. */ elem = NULL; while ((elem = nvlist_next_nvpair(nvl, elem)) != NULL) { const char *propname = nvpair_name(elem); prop = zfs_name_to_prop(propname); if (prop == ZPROP_USERPROP && zfs_prop_user(propname)) { /* * This is a user property: make sure it's a * string, and that it's less than ZAP_MAXNAMELEN. */ if (nvpair_type(elem) != DATA_TYPE_STRING) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' must be a string"), propname); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } if (strlen(nvpair_name(elem)) >= ZAP_MAXNAMELEN) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "property name '%s' is too long"), propname); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } (void) nvpair_value_string(elem, &strval); if (nvlist_add_string(ret, propname, strval) != 0) { (void) no_memory(hdl); goto error; } continue; } /* * Currently, only user properties can be modified on * snapshots. */ if (type == ZFS_TYPE_SNAPSHOT) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "this property can not be modified for snapshots")); (void) zfs_error(hdl, EZFS_PROPTYPE, errbuf); goto error; } if (prop == ZPROP_USERPROP && zfs_prop_userquota(propname)) { zfs_userquota_prop_t uqtype; char *newpropname = NULL; char domain[128]; uint64_t rid; uint64_t valary[3]; int rc; if (userquota_propname_decode(propname, zoned, &uqtype, domain, sizeof (domain), &rid) != 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' has an invalid user/group name"), propname); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } if (uqtype != ZFS_PROP_USERQUOTA && uqtype != ZFS_PROP_GROUPQUOTA && uqtype != ZFS_PROP_USEROBJQUOTA && uqtype != ZFS_PROP_GROUPOBJQUOTA && uqtype != ZFS_PROP_PROJECTQUOTA && uqtype != ZFS_PROP_PROJECTOBJQUOTA) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' is readonly"), propname); (void) zfs_error(hdl, EZFS_PROPREADONLY, errbuf); goto error; } if (nvpair_type(elem) == DATA_TYPE_STRING) { (void) nvpair_value_string(elem, &strval); if (strcmp(strval, "none") == 0) { intval = 0; } else if (zfs_nicestrtonum(hdl, strval, &intval) != 0) { (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } } else if (nvpair_type(elem) == DATA_TYPE_UINT64) { (void) nvpair_value_uint64(elem, &intval); if (intval == 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "use 'none' to disable " "{user|group|project}quota")); goto error; } } else { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' must be a number"), propname); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } /* * Encode the prop name as * userquota@-domain, to make it easy * for the kernel to decode. */ rc = asprintf(&newpropname, "%s%llx-%s", zfs_userquota_prop_prefixes[uqtype], (longlong_t)rid, domain); if (rc == -1 || newpropname == NULL) { (void) no_memory(hdl); goto error; } valary[0] = uqtype; valary[1] = rid; valary[2] = intval; if (nvlist_add_uint64_array(ret, newpropname, valary, 3) != 0) { free(newpropname); (void) no_memory(hdl); goto error; } free(newpropname); continue; } else if (prop == ZPROP_USERPROP && zfs_prop_written(propname)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' is readonly"), propname); (void) zfs_error(hdl, EZFS_PROPREADONLY, errbuf); goto error; } if (prop == ZPROP_INVAL) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid property '%s'"), propname); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } if (!zfs_prop_valid_for_type(prop, type, B_FALSE)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' does not " "apply to datasets of this type"), propname); (void) zfs_error(hdl, EZFS_PROPTYPE, errbuf); goto error; } if (zfs_prop_readonly(prop) && !(zfs_prop_setonce(prop) && zhp == NULL) && !(zfs_prop_encryption_key_param(prop) && key_params_ok)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' is readonly"), propname); (void) zfs_error(hdl, EZFS_PROPREADONLY, errbuf); goto error; } if (zprop_parse_value(hdl, elem, prop, type, ret, &strval, &intval, errbuf) != 0) goto error; /* * Perform some additional checks for specific properties. */ switch (prop) { case ZFS_PROP_VERSION: { int version; if (zhp == NULL) break; version = zfs_prop_get_int(zhp, ZFS_PROP_VERSION); if (intval < version) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "Can not downgrade; already at version %u"), version); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } break; } case ZFS_PROP_VOLBLOCKSIZE: case ZFS_PROP_RECORDSIZE: { int maxbs = SPA_MAXBLOCKSIZE; char buf[64]; if (zpool_hdl != NULL) { maxbs = zpool_get_prop_int(zpool_hdl, ZPOOL_PROP_MAXBLOCKSIZE, NULL); } /* * The value must be a power of two between * SPA_MINBLOCKSIZE and maxbs. */ if (intval < SPA_MINBLOCKSIZE || intval > maxbs || !ISP2(intval)) { zfs_nicebytes(maxbs, buf, sizeof (buf)); zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' must be power of 2 from 512B " "to %s"), propname, buf); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } + /* save the ZFS_PROP_RECORDSIZE during create op */ + if (zpool_hdl == NULL && prop == ZFS_PROP_RECORDSIZE) { + set_maxbs = intval; + } break; } case ZFS_PROP_SPECIAL_SMALL_BLOCKS: { - int maxbs = SPA_OLD_MAXBLOCKSIZE; + int maxbs = + set_maxbs == 0 ? SPA_OLD_MAXBLOCKSIZE : set_maxbs; char buf[64]; if (zpool_hdl != NULL) { char state[64] = ""; maxbs = zpool_get_prop_int(zpool_hdl, ZPOOL_PROP_MAXBLOCKSIZE, NULL); /* * Issue a warning but do not fail so that * tests for settable properties succeed. */ if (zpool_prop_get_feature(zpool_hdl, "feature@allocation_classes", state, sizeof (state)) != 0 || strcmp(state, ZFS_FEATURE_ACTIVE) != 0) { (void) fprintf(stderr, gettext( "%s: property requires a special " "device in the pool\n"), propname); } } if (intval != 0 && (intval < SPA_MINBLOCKSIZE || intval > maxbs || !ISP2(intval))) { zfs_nicebytes(maxbs, buf, sizeof (buf)); zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid '%s=%llu' property: must be zero " "or a power of 2 from 512B to %s"), propname, (unsigned long long)intval, buf); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } break; } case ZFS_PROP_MLSLABEL: { #ifdef HAVE_MLSLABEL /* * Verify the mlslabel string and convert to * internal hex label string. */ m_label_t *new_sl; char *hex = NULL; /* internal label string */ /* Default value is already OK. */ if (strcasecmp(strval, ZFS_MLSLABEL_DEFAULT) == 0) break; /* Verify the label can be converted to binary form */ if (((new_sl = m_label_alloc(MAC_LABEL)) == NULL) || (str_to_label(strval, &new_sl, MAC_LABEL, L_NO_CORRECTION, NULL) == -1)) { goto badlabel; } /* Now translate to hex internal label string */ if (label_to_str(new_sl, &hex, M_INTERNAL, DEF_NAMES) != 0) { if (hex) free(hex); goto badlabel; } m_label_free(new_sl); /* If string is already in internal form, we're done. */ if (strcmp(strval, hex) == 0) { free(hex); break; } /* Replace the label string with the internal form. */ (void) nvlist_remove(ret, zfs_prop_to_name(prop), DATA_TYPE_STRING); fnvlist_add_string(ret, zfs_prop_to_name(prop), hex); free(hex); break; badlabel: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid mlslabel '%s'"), strval); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); m_label_free(new_sl); /* OK if null */ goto error; #else zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "mlslabels are unsupported")); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; #endif /* HAVE_MLSLABEL */ } case ZFS_PROP_MOUNTPOINT: { namecheck_err_t why; if (strcmp(strval, ZFS_MOUNTPOINT_NONE) == 0 || strcmp(strval, ZFS_MOUNTPOINT_LEGACY) == 0) break; if (mountpoint_namecheck(strval, &why)) { switch (why) { case NAME_ERR_LEADING_SLASH: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' must be an absolute path, " "'none', or 'legacy'"), propname); break; case NAME_ERR_TOOLONG: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "component of '%s' is too long"), propname); break; default: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "(%d) not defined"), why); break; } (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } zfs_fallthrough; } case ZFS_PROP_SHARESMB: case ZFS_PROP_SHARENFS: /* * For the mountpoint and sharenfs or sharesmb * properties, check if it can be set in a * global/non-global zone based on * the zoned property value: * * global zone non-global zone * -------------------------------------------------- * zoned=on mountpoint (no) mountpoint (yes) * sharenfs (no) sharenfs (no) * sharesmb (no) sharesmb (no) * * zoned=off mountpoint (yes) N/A * sharenfs (yes) * sharesmb (yes) */ if (zoned) { if (getzoneid() == GLOBAL_ZONEID) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' cannot be set on " "dataset in a non-global zone"), propname); (void) zfs_error(hdl, EZFS_ZONED, errbuf); goto error; } else if (prop == ZFS_PROP_SHARENFS || prop == ZFS_PROP_SHARESMB) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' cannot be set in " "a non-global zone"), propname); (void) zfs_error(hdl, EZFS_ZONED, errbuf); goto error; } } else if (getzoneid() != GLOBAL_ZONEID) { /* * If zoned property is 'off', this must be in * a global zone. If not, something is wrong. */ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' cannot be set while dataset " "'zoned' property is set"), propname); (void) zfs_error(hdl, EZFS_ZONED, errbuf); goto error; } /* * At this point, it is legitimate to set the * property. Now we want to make sure that the * property value is valid if it is sharenfs. */ if ((prop == ZFS_PROP_SHARENFS || prop == ZFS_PROP_SHARESMB) && strcmp(strval, "on") != 0 && strcmp(strval, "off") != 0) { enum sa_protocol proto; if (prop == ZFS_PROP_SHARESMB) proto = SA_PROTOCOL_SMB; else proto = SA_PROTOCOL_NFS; if (sa_validate_shareopts(strval, proto) != SA_OK) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' cannot be set to invalid " "options"), propname); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } } break; case ZFS_PROP_KEYLOCATION: if (!zfs_prop_valid_keylocation(strval, B_FALSE)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid keylocation")); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } if (zhp != NULL) { uint64_t crypt = zfs_prop_get_int(zhp, ZFS_PROP_ENCRYPTION); if (crypt == ZIO_CRYPT_OFF && strcmp(strval, "none") != 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "keylocation must be 'none' " "for unencrypted datasets")); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } else if (crypt != ZIO_CRYPT_OFF && strcmp(strval, "none") == 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "keylocation must not be 'none' " "for encrypted datasets")); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } } break; case ZFS_PROP_PBKDF2_ITERS: if (intval < MIN_PBKDF2_ITERATIONS) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "minimum pbkdf2 iterations is %u"), MIN_PBKDF2_ITERATIONS); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } break; case ZFS_PROP_UTF8ONLY: chosen_utf = (int)intval; break; case ZFS_PROP_NORMALIZE: chosen_normal = (int)intval; break; default: break; } /* * For changes to existing volumes, we have some additional * checks to enforce. */ if (type == ZFS_TYPE_VOLUME && zhp != NULL) { uint64_t blocksize = zfs_prop_get_int(zhp, ZFS_PROP_VOLBLOCKSIZE); char buf[64]; switch (prop) { case ZFS_PROP_VOLSIZE: if (intval % blocksize != 0) { zfs_nicebytes(blocksize, buf, sizeof (buf)); zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' must be a multiple of " "volume block size (%s)"), propname, buf); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } if (intval == 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' cannot be zero"), propname); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } break; default: break; } } /* check encryption properties */ if (zhp != NULL) { int64_t crypt = zfs_prop_get_int(zhp, ZFS_PROP_ENCRYPTION); switch (prop) { case ZFS_PROP_COPIES: if (crypt != ZIO_CRYPT_OFF && intval > 2) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "encrypted datasets cannot have " "3 copies")); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } break; default: break; } } } /* * If normalization was chosen, but no UTF8 choice was made, * enforce rejection of non-UTF8 names. * * If normalization was chosen, but rejecting non-UTF8 names * was explicitly not chosen, it is an error. * * If utf8only was turned off, but the parent has normalization, * turn off normalization. */ if (chosen_normal > 0 && chosen_utf < 0) { if (nvlist_add_uint64(ret, zfs_prop_to_name(ZFS_PROP_UTF8ONLY), 1) != 0) { (void) no_memory(hdl); goto error; } } else if (chosen_normal > 0 && chosen_utf == 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' must be set 'on' if normalization chosen"), zfs_prop_to_name(ZFS_PROP_UTF8ONLY)); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } else if (chosen_normal < 0 && chosen_utf == 0) { if (nvlist_add_uint64(ret, zfs_prop_to_name(ZFS_PROP_NORMALIZE), 0) != 0) { (void) no_memory(hdl); goto error; } } return (ret); error: nvlist_free(ret); return (NULL); } static int zfs_add_synthetic_resv(zfs_handle_t *zhp, nvlist_t *nvl) { uint64_t old_volsize; uint64_t new_volsize; uint64_t old_reservation; uint64_t new_reservation; zfs_prop_t resv_prop; nvlist_t *props; zpool_handle_t *zph = zpool_handle(zhp); /* * If this is an existing volume, and someone is setting the volsize, * make sure that it matches the reservation, or add it if necessary. */ old_volsize = zfs_prop_get_int(zhp, ZFS_PROP_VOLSIZE); if (zfs_which_resv_prop(zhp, &resv_prop) < 0) return (-1); old_reservation = zfs_prop_get_int(zhp, resv_prop); props = fnvlist_alloc(); fnvlist_add_uint64(props, zfs_prop_to_name(ZFS_PROP_VOLBLOCKSIZE), zfs_prop_get_int(zhp, ZFS_PROP_VOLBLOCKSIZE)); if ((zvol_volsize_to_reservation(zph, old_volsize, props) != old_reservation) || nvlist_exists(nvl, zfs_prop_to_name(resv_prop))) { fnvlist_free(props); return (0); } if (nvlist_lookup_uint64(nvl, zfs_prop_to_name(ZFS_PROP_VOLSIZE), &new_volsize) != 0) { fnvlist_free(props); return (-1); } new_reservation = zvol_volsize_to_reservation(zph, new_volsize, props); fnvlist_free(props); if (nvlist_add_uint64(nvl, zfs_prop_to_name(resv_prop), new_reservation) != 0) { (void) no_memory(zhp->zfs_hdl); return (-1); } return (1); } /* * Helper for 'zfs {set|clone} refreservation=auto'. Must be called after * zfs_valid_proplist(), as it is what sets the UINT64_MAX sentinel value. * Return codes must match zfs_add_synthetic_resv(). */ static int zfs_fix_auto_resv(zfs_handle_t *zhp, nvlist_t *nvl) { uint64_t volsize; uint64_t resvsize; zfs_prop_t prop; nvlist_t *props; if (!ZFS_IS_VOLUME(zhp)) { return (0); } if (zfs_which_resv_prop(zhp, &prop) != 0) { return (-1); } if (prop != ZFS_PROP_REFRESERVATION) { return (0); } if (nvlist_lookup_uint64(nvl, zfs_prop_to_name(prop), &resvsize) != 0) { /* No value being set, so it can't be "auto" */ return (0); } if (resvsize != UINT64_MAX) { /* Being set to a value other than "auto" */ return (0); } props = fnvlist_alloc(); fnvlist_add_uint64(props, zfs_prop_to_name(ZFS_PROP_VOLBLOCKSIZE), zfs_prop_get_int(zhp, ZFS_PROP_VOLBLOCKSIZE)); if (nvlist_lookup_uint64(nvl, zfs_prop_to_name(ZFS_PROP_VOLSIZE), &volsize) != 0) { volsize = zfs_prop_get_int(zhp, ZFS_PROP_VOLSIZE); } resvsize = zvol_volsize_to_reservation(zpool_handle(zhp), volsize, props); fnvlist_free(props); (void) nvlist_remove_all(nvl, zfs_prop_to_name(prop)); if (nvlist_add_uint64(nvl, zfs_prop_to_name(prop), resvsize) != 0) { (void) no_memory(zhp->zfs_hdl); return (-1); } return (1); } static boolean_t zfs_is_namespace_prop(zfs_prop_t prop) { switch (prop) { case ZFS_PROP_ATIME: case ZFS_PROP_RELATIME: case ZFS_PROP_DEVICES: case ZFS_PROP_EXEC: case ZFS_PROP_SETUID: case ZFS_PROP_READONLY: case ZFS_PROP_XATTR: case ZFS_PROP_NBMAND: return (B_TRUE); default: return (B_FALSE); } } /* * Given a property name and value, set the property for the given dataset. */ int zfs_prop_set(zfs_handle_t *zhp, const char *propname, const char *propval) { int ret = -1; char errbuf[ERRBUFLEN]; libzfs_handle_t *hdl = zhp->zfs_hdl; nvlist_t *nvl = NULL; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot set property for '%s'"), zhp->zfs_name); if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0 || nvlist_add_string(nvl, propname, propval) != 0) { (void) no_memory(hdl); goto error; } ret = zfs_prop_set_list(zhp, nvl); error: nvlist_free(nvl); return (ret); } /* * Given an nvlist of property names and values, set the properties for the * given dataset. */ int zfs_prop_set_list(zfs_handle_t *zhp, nvlist_t *props) { zfs_cmd_t zc = {"\0"}; int ret = -1; prop_changelist_t **cls = NULL; int cl_idx; char errbuf[ERRBUFLEN]; libzfs_handle_t *hdl = zhp->zfs_hdl; nvlist_t *nvl; int nvl_len = 0; int added_resv = 0; zfs_prop_t prop = 0; nvpair_t *elem; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot set property for '%s'"), zhp->zfs_name); if ((nvl = zfs_valid_proplist(hdl, zhp->zfs_type, props, zfs_prop_get_int(zhp, ZFS_PROP_ZONED), zhp, zhp->zpool_hdl, B_FALSE, errbuf)) == NULL) goto error; /* * We have to check for any extra properties which need to be added * before computing the length of the nvlist. */ for (elem = nvlist_next_nvpair(nvl, NULL); elem != NULL; elem = nvlist_next_nvpair(nvl, elem)) { if (zfs_name_to_prop(nvpair_name(elem)) == ZFS_PROP_VOLSIZE && (added_resv = zfs_add_synthetic_resv(zhp, nvl)) == -1) { goto error; } } if (added_resv != 1 && (added_resv = zfs_fix_auto_resv(zhp, nvl)) == -1) { goto error; } /* * Check how many properties we're setting and allocate an array to * store changelist pointers for postfix(). */ for (elem = nvlist_next_nvpair(nvl, NULL); elem != NULL; elem = nvlist_next_nvpair(nvl, elem)) nvl_len++; if ((cls = calloc(nvl_len, sizeof (prop_changelist_t *))) == NULL) goto error; cl_idx = 0; for (elem = nvlist_next_nvpair(nvl, NULL); elem != NULL; elem = nvlist_next_nvpair(nvl, elem)) { prop = zfs_name_to_prop(nvpair_name(elem)); assert(cl_idx < nvl_len); /* * We don't want to unmount & remount the dataset when changing * its canmount property to 'on' or 'noauto'. We only use * the changelist logic to unmount when setting canmount=off. */ if (prop != ZFS_PROP_CANMOUNT || (fnvpair_value_uint64(elem) == ZFS_CANMOUNT_OFF && zfs_is_mounted(zhp, NULL))) { cls[cl_idx] = changelist_gather(zhp, prop, 0, 0); if (cls[cl_idx] == NULL) goto error; } if (prop == ZFS_PROP_MOUNTPOINT && changelist_haszonedchild(cls[cl_idx])) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "child dataset with inherited mountpoint is used " "in a non-global zone")); ret = zfs_error(hdl, EZFS_ZONED, errbuf); goto error; } if (cls[cl_idx] != NULL && (ret = changelist_prefix(cls[cl_idx])) != 0) goto error; cl_idx++; } assert(cl_idx == nvl_len); /* * Execute the corresponding ioctl() to set this list of properties. */ (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); zcmd_write_src_nvlist(hdl, &zc, nvl); zcmd_alloc_dst_nvlist(hdl, &zc, 0); ret = zfs_ioctl(hdl, ZFS_IOC_SET_PROP, &zc); if (ret != 0) { if (zc.zc_nvlist_dst_filled == B_FALSE) { (void) zfs_standard_error(hdl, errno, errbuf); goto error; } /* Get the list of unset properties back and report them. */ nvlist_t *errorprops = NULL; if (zcmd_read_dst_nvlist(hdl, &zc, &errorprops) != 0) goto error; for (nvpair_t *elem = nvlist_next_nvpair(errorprops, NULL); elem != NULL; elem = nvlist_next_nvpair(errorprops, elem)) { prop = zfs_name_to_prop(nvpair_name(elem)); zfs_setprop_error(hdl, prop, errno, errbuf); } nvlist_free(errorprops); if (added_resv && errno == ENOSPC) { /* clean up the volsize property we tried to set */ uint64_t old_volsize = zfs_prop_get_int(zhp, ZFS_PROP_VOLSIZE); nvlist_free(nvl); nvl = NULL; zcmd_free_nvlists(&zc); if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0) goto error; if (nvlist_add_uint64(nvl, zfs_prop_to_name(ZFS_PROP_VOLSIZE), old_volsize) != 0) goto error; zcmd_write_src_nvlist(hdl, &zc, nvl); (void) zfs_ioctl(hdl, ZFS_IOC_SET_PROP, &zc); } } else { for (cl_idx = 0; cl_idx < nvl_len; cl_idx++) { if (cls[cl_idx] != NULL) { int clp_err = changelist_postfix(cls[cl_idx]); if (clp_err != 0) ret = clp_err; } } if (ret == 0) { /* * Refresh the statistics so the new property * value is reflected. */ (void) get_stats(zhp); /* * Remount the filesystem to propagate the change * if one of the options handled by the generic * Linux namespace layer has been modified. */ if (zfs_is_namespace_prop(prop) && zfs_is_mounted(zhp, NULL)) ret = zfs_mount(zhp, MNTOPT_REMOUNT, 0); } } error: nvlist_free(nvl); zcmd_free_nvlists(&zc); if (cls != NULL) { for (cl_idx = 0; cl_idx < nvl_len; cl_idx++) { if (cls[cl_idx] != NULL) changelist_free(cls[cl_idx]); } free(cls); } return (ret); } /* * Given a property, inherit the value from the parent dataset, or if received * is TRUE, revert to the received value, if any. */ int zfs_prop_inherit(zfs_handle_t *zhp, const char *propname, boolean_t received) { zfs_cmd_t zc = {"\0"}; int ret; prop_changelist_t *cl; libzfs_handle_t *hdl = zhp->zfs_hdl; char errbuf[ERRBUFLEN]; zfs_prop_t prop; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot inherit %s for '%s'"), propname, zhp->zfs_name); zc.zc_cookie = received; if ((prop = zfs_name_to_prop(propname)) == ZPROP_USERPROP) { /* * For user properties, the amount of work we have to do is very * small, so just do it here. */ if (!zfs_prop_user(propname)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid property")); return (zfs_error(hdl, EZFS_BADPROP, errbuf)); } (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); (void) strlcpy(zc.zc_value, propname, sizeof (zc.zc_value)); if (zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_INHERIT_PROP, &zc) != 0) return (zfs_standard_error(hdl, errno, errbuf)); (void) get_stats(zhp); return (0); } /* * Verify that this property is inheritable. */ if (zfs_prop_readonly(prop)) return (zfs_error(hdl, EZFS_PROPREADONLY, errbuf)); if (!zfs_prop_inheritable(prop) && !received) return (zfs_error(hdl, EZFS_PROPNONINHERIT, errbuf)); /* * Check to see if the value applies to this type */ if (!zfs_prop_valid_for_type(prop, zhp->zfs_type, B_FALSE)) return (zfs_error(hdl, EZFS_PROPTYPE, errbuf)); /* * Normalize the name, to get rid of shorthand abbreviations. */ propname = zfs_prop_to_name(prop); (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); (void) strlcpy(zc.zc_value, propname, sizeof (zc.zc_value)); if (prop == ZFS_PROP_MOUNTPOINT && getzoneid() == GLOBAL_ZONEID && zfs_prop_get_int(zhp, ZFS_PROP_ZONED)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "dataset is used in a non-global zone")); return (zfs_error(hdl, EZFS_ZONED, errbuf)); } /* * Determine datasets which will be affected by this change, if any. */ if ((cl = changelist_gather(zhp, prop, 0, 0)) == NULL) return (-1); if (prop == ZFS_PROP_MOUNTPOINT && changelist_haszonedchild(cl)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "child dataset with inherited mountpoint is used " "in a non-global zone")); ret = zfs_error(hdl, EZFS_ZONED, errbuf); goto error; } if ((ret = changelist_prefix(cl)) != 0) goto error; if (zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_INHERIT_PROP, &zc) != 0) { changelist_free(cl); return (zfs_standard_error(hdl, errno, errbuf)); } else { if ((ret = changelist_postfix(cl)) != 0) goto error; /* * Refresh the statistics so the new property is reflected. */ (void) get_stats(zhp); /* * Remount the filesystem to propagate the change * if one of the options handled by the generic * Linux namespace layer has been modified. */ if (zfs_is_namespace_prop(prop) && zfs_is_mounted(zhp, NULL)) ret = zfs_mount(zhp, MNTOPT_REMOUNT, 0); } error: changelist_free(cl); return (ret); } /* * True DSL properties are stored in an nvlist. The following two functions * extract them appropriately. */ uint64_t getprop_uint64(zfs_handle_t *zhp, zfs_prop_t prop, const char **source) { nvlist_t *nv; uint64_t value; *source = NULL; if (nvlist_lookup_nvlist(zhp->zfs_props, zfs_prop_to_name(prop), &nv) == 0) { value = fnvlist_lookup_uint64(nv, ZPROP_VALUE); (void) nvlist_lookup_string(nv, ZPROP_SOURCE, source); } else { verify(!zhp->zfs_props_table || zhp->zfs_props_table[prop] == B_TRUE); value = zfs_prop_default_numeric(prop); *source = ""; } return (value); } static const char * getprop_string(zfs_handle_t *zhp, zfs_prop_t prop, const char **source) { nvlist_t *nv; const char *value; *source = NULL; if (nvlist_lookup_nvlist(zhp->zfs_props, zfs_prop_to_name(prop), &nv) == 0) { value = fnvlist_lookup_string(nv, ZPROP_VALUE); (void) nvlist_lookup_string(nv, ZPROP_SOURCE, source); } else { verify(!zhp->zfs_props_table || zhp->zfs_props_table[prop] == B_TRUE); value = zfs_prop_default_string(prop); *source = ""; } return (value); } static boolean_t zfs_is_recvd_props_mode(zfs_handle_t *zhp) { return (zhp->zfs_props != NULL && zhp->zfs_props == zhp->zfs_recvd_props); } static void zfs_set_recvd_props_mode(zfs_handle_t *zhp, uintptr_t *cookie) { *cookie = (uintptr_t)zhp->zfs_props; zhp->zfs_props = zhp->zfs_recvd_props; } static void zfs_unset_recvd_props_mode(zfs_handle_t *zhp, uintptr_t *cookie) { zhp->zfs_props = (nvlist_t *)*cookie; *cookie = 0; } /* * Internal function for getting a numeric property. Both zfs_prop_get() and * zfs_prop_get_int() are built using this interface. * * Certain properties can be overridden using 'mount -o'. In this case, scan * the contents of the /proc/self/mounts entry, searching for the * appropriate options. If they differ from the on-disk values, report the * current values and mark the source "temporary". */ static int get_numeric_property(zfs_handle_t *zhp, zfs_prop_t prop, zprop_source_t *src, const char **source, uint64_t *val) { zfs_cmd_t zc = {"\0"}; nvlist_t *zplprops = NULL; struct mnttab mnt; const char *mntopt_on = NULL; const char *mntopt_off = NULL; boolean_t received = zfs_is_recvd_props_mode(zhp); *source = NULL; /* * If the property is being fetched for a snapshot, check whether * the property is valid for the snapshot's head dataset type. */ if (zhp->zfs_type == ZFS_TYPE_SNAPSHOT && !zfs_prop_valid_for_type(prop, zhp->zfs_head_type, B_TRUE)) { *val = zfs_prop_default_numeric(prop); return (-1); } switch (prop) { case ZFS_PROP_ATIME: mntopt_on = MNTOPT_ATIME; mntopt_off = MNTOPT_NOATIME; break; case ZFS_PROP_RELATIME: mntopt_on = MNTOPT_RELATIME; mntopt_off = MNTOPT_NORELATIME; break; case ZFS_PROP_DEVICES: mntopt_on = MNTOPT_DEVICES; mntopt_off = MNTOPT_NODEVICES; break; case ZFS_PROP_EXEC: mntopt_on = MNTOPT_EXEC; mntopt_off = MNTOPT_NOEXEC; break; case ZFS_PROP_READONLY: mntopt_on = MNTOPT_RO; mntopt_off = MNTOPT_RW; break; case ZFS_PROP_SETUID: mntopt_on = MNTOPT_SETUID; mntopt_off = MNTOPT_NOSETUID; break; case ZFS_PROP_XATTR: mntopt_on = MNTOPT_XATTR; mntopt_off = MNTOPT_NOXATTR; break; case ZFS_PROP_NBMAND: mntopt_on = MNTOPT_NBMAND; mntopt_off = MNTOPT_NONBMAND; break; default: break; } /* * Because looking up the mount options is potentially expensive * (iterating over all of /proc/self/mounts), we defer its * calculation until we're looking up a property which requires * its presence. */ if (!zhp->zfs_mntcheck && (mntopt_on != NULL || prop == ZFS_PROP_MOUNTED)) { libzfs_handle_t *hdl = zhp->zfs_hdl; struct mnttab entry; if (libzfs_mnttab_find(hdl, zhp->zfs_name, &entry) == 0) zhp->zfs_mntopts = zfs_strdup(hdl, entry.mnt_mntopts); zhp->zfs_mntcheck = B_TRUE; } if (zhp->zfs_mntopts == NULL) mnt.mnt_mntopts = (char *)""; else mnt.mnt_mntopts = zhp->zfs_mntopts; switch (prop) { case ZFS_PROP_ATIME: case ZFS_PROP_RELATIME: case ZFS_PROP_DEVICES: case ZFS_PROP_EXEC: case ZFS_PROP_READONLY: case ZFS_PROP_SETUID: #ifndef __FreeBSD__ case ZFS_PROP_XATTR: #endif case ZFS_PROP_NBMAND: *val = getprop_uint64(zhp, prop, source); if (received) break; if (hasmntopt(&mnt, mntopt_on) && !*val) { *val = B_TRUE; if (src) *src = ZPROP_SRC_TEMPORARY; } else if (hasmntopt(&mnt, mntopt_off) && *val) { *val = B_FALSE; if (src) *src = ZPROP_SRC_TEMPORARY; } break; case ZFS_PROP_CANMOUNT: case ZFS_PROP_VOLSIZE: case ZFS_PROP_QUOTA: case ZFS_PROP_REFQUOTA: case ZFS_PROP_RESERVATION: case ZFS_PROP_REFRESERVATION: case ZFS_PROP_FILESYSTEM_LIMIT: case ZFS_PROP_SNAPSHOT_LIMIT: case ZFS_PROP_FILESYSTEM_COUNT: case ZFS_PROP_SNAPSHOT_COUNT: *val = getprop_uint64(zhp, prop, source); if (*source == NULL) { /* not default, must be local */ *source = zhp->zfs_name; } break; case ZFS_PROP_MOUNTED: *val = (zhp->zfs_mntopts != NULL); break; case ZFS_PROP_NUMCLONES: *val = zhp->zfs_dmustats.dds_num_clones; break; case ZFS_PROP_VERSION: case ZFS_PROP_NORMALIZE: case ZFS_PROP_UTF8ONLY: case ZFS_PROP_CASE: zcmd_alloc_dst_nvlist(zhp->zfs_hdl, &zc, 0); (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); if (zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_OBJSET_ZPLPROPS, &zc)) { zcmd_free_nvlists(&zc); if (prop == ZFS_PROP_VERSION && zhp->zfs_type == ZFS_TYPE_VOLUME) *val = zfs_prop_default_numeric(prop); return (-1); } if (zcmd_read_dst_nvlist(zhp->zfs_hdl, &zc, &zplprops) != 0 || nvlist_lookup_uint64(zplprops, zfs_prop_to_name(prop), val) != 0) { zcmd_free_nvlists(&zc); return (-1); } nvlist_free(zplprops); zcmd_free_nvlists(&zc); break; case ZFS_PROP_INCONSISTENT: *val = zhp->zfs_dmustats.dds_inconsistent; break; case ZFS_PROP_REDACTED: *val = zhp->zfs_dmustats.dds_redacted; break; case ZFS_PROP_GUID: if (zhp->zfs_dmustats.dds_guid != 0) *val = zhp->zfs_dmustats.dds_guid; else *val = getprop_uint64(zhp, prop, source); break; case ZFS_PROP_CREATETXG: /* * We can directly read createtxg property from zfs * handle for Filesystem, Snapshot and ZVOL types. */ if (((zhp->zfs_type == ZFS_TYPE_FILESYSTEM) || (zhp->zfs_type == ZFS_TYPE_SNAPSHOT) || (zhp->zfs_type == ZFS_TYPE_VOLUME)) && (zhp->zfs_dmustats.dds_creation_txg != 0)) { *val = zhp->zfs_dmustats.dds_creation_txg; break; } else { *val = getprop_uint64(zhp, prop, source); } zfs_fallthrough; default: switch (zfs_prop_get_type(prop)) { case PROP_TYPE_NUMBER: case PROP_TYPE_INDEX: *val = getprop_uint64(zhp, prop, source); /* * If we tried to use a default value for a * readonly property, it means that it was not * present. Note this only applies to "truly" * readonly properties, not set-once properties * like volblocksize. */ if (zfs_prop_readonly(prop) && !zfs_prop_setonce(prop) && *source != NULL && (*source)[0] == '\0') { *source = NULL; return (-1); } break; case PROP_TYPE_STRING: default: zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN, "cannot get non-numeric property")); return (zfs_error(zhp->zfs_hdl, EZFS_BADPROP, dgettext(TEXT_DOMAIN, "internal error"))); } } return (0); } /* * Calculate the source type, given the raw source string. */ static void get_source(zfs_handle_t *zhp, zprop_source_t *srctype, const char *source, char *statbuf, size_t statlen) { if (statbuf == NULL || srctype == NULL || *srctype == ZPROP_SRC_TEMPORARY) { return; } if (source == NULL) { *srctype = ZPROP_SRC_NONE; } else if (source[0] == '\0') { *srctype = ZPROP_SRC_DEFAULT; } else if (strstr(source, ZPROP_SOURCE_VAL_RECVD) != NULL) { *srctype = ZPROP_SRC_RECEIVED; } else { if (strcmp(source, zhp->zfs_name) == 0) { *srctype = ZPROP_SRC_LOCAL; } else { (void) strlcpy(statbuf, source, statlen); *srctype = ZPROP_SRC_INHERITED; } } } int zfs_prop_get_recvd(zfs_handle_t *zhp, const char *propname, char *propbuf, size_t proplen, boolean_t literal) { zfs_prop_t prop; int err = 0; if (zhp->zfs_recvd_props == NULL) if (get_recvd_props_ioctl(zhp) != 0) return (-1); prop = zfs_name_to_prop(propname); if (prop != ZPROP_USERPROP) { uintptr_t cookie; if (!nvlist_exists(zhp->zfs_recvd_props, propname)) return (-1); zfs_set_recvd_props_mode(zhp, &cookie); err = zfs_prop_get(zhp, prop, propbuf, proplen, NULL, NULL, 0, literal); zfs_unset_recvd_props_mode(zhp, &cookie); } else { nvlist_t *propval; const char *recvdval; if (nvlist_lookup_nvlist(zhp->zfs_recvd_props, propname, &propval) != 0) return (-1); recvdval = fnvlist_lookup_string(propval, ZPROP_VALUE); (void) strlcpy(propbuf, recvdval, proplen); } return (err == 0 ? 0 : -1); } static int get_clones_string(zfs_handle_t *zhp, char *propbuf, size_t proplen) { nvlist_t *value; nvpair_t *pair; value = zfs_get_clones_nvl(zhp); if (value == NULL || nvlist_empty(value)) return (-1); propbuf[0] = '\0'; for (pair = nvlist_next_nvpair(value, NULL); pair != NULL; pair = nvlist_next_nvpair(value, pair)) { if (propbuf[0] != '\0') (void) strlcat(propbuf, ",", proplen); (void) strlcat(propbuf, nvpair_name(pair), proplen); } return (0); } struct get_clones_arg { uint64_t numclones; nvlist_t *value; const char *origin; char buf[ZFS_MAX_DATASET_NAME_LEN]; }; static int get_clones_cb(zfs_handle_t *zhp, void *arg) { struct get_clones_arg *gca = arg; if (gca->numclones == 0) { zfs_close(zhp); return (0); } if (zfs_prop_get(zhp, ZFS_PROP_ORIGIN, gca->buf, sizeof (gca->buf), NULL, NULL, 0, B_TRUE) != 0) goto out; if (strcmp(gca->buf, gca->origin) == 0) { fnvlist_add_boolean(gca->value, zfs_get_name(zhp)); gca->numclones--; } out: (void) zfs_iter_children_v2(zhp, 0, get_clones_cb, gca); zfs_close(zhp); return (0); } nvlist_t * zfs_get_clones_nvl(zfs_handle_t *zhp) { nvlist_t *nv, *value; if (nvlist_lookup_nvlist(zhp->zfs_props, zfs_prop_to_name(ZFS_PROP_CLONES), &nv) != 0) { struct get_clones_arg gca; /* * if this is a snapshot, then the kernel wasn't able * to get the clones. Do it by slowly iterating. */ if (zhp->zfs_type != ZFS_TYPE_SNAPSHOT) return (NULL); if (nvlist_alloc(&nv, NV_UNIQUE_NAME, 0) != 0) return (NULL); if (nvlist_alloc(&value, NV_UNIQUE_NAME, 0) != 0) { nvlist_free(nv); return (NULL); } gca.numclones = zfs_prop_get_int(zhp, ZFS_PROP_NUMCLONES); gca.value = value; gca.origin = zhp->zfs_name; if (gca.numclones != 0) { zfs_handle_t *root; char pool[ZFS_MAX_DATASET_NAME_LEN]; char *cp = pool; /* get the pool name */ (void) strlcpy(pool, zhp->zfs_name, sizeof (pool)); (void) strsep(&cp, "/@"); root = zfs_open(zhp->zfs_hdl, pool, ZFS_TYPE_FILESYSTEM); if (root == NULL) { nvlist_free(nv); nvlist_free(value); return (NULL); } (void) get_clones_cb(root, &gca); } if (gca.numclones != 0 || nvlist_add_nvlist(nv, ZPROP_VALUE, value) != 0 || nvlist_add_nvlist(zhp->zfs_props, zfs_prop_to_name(ZFS_PROP_CLONES), nv) != 0) { nvlist_free(nv); nvlist_free(value); return (NULL); } nvlist_free(nv); nvlist_free(value); nv = fnvlist_lookup_nvlist(zhp->zfs_props, zfs_prop_to_name(ZFS_PROP_CLONES)); } return (fnvlist_lookup_nvlist(nv, ZPROP_VALUE)); } static int get_rsnaps_string(zfs_handle_t *zhp, char *propbuf, size_t proplen) { nvlist_t *value; uint64_t *snaps; uint_t nsnaps; if (nvlist_lookup_nvlist(zhp->zfs_props, zfs_prop_to_name(ZFS_PROP_REDACT_SNAPS), &value) != 0) return (-1); if (nvlist_lookup_uint64_array(value, ZPROP_VALUE, &snaps, &nsnaps) != 0) return (-1); if (nsnaps == 0) { /* There's no redaction snapshots; pass a special value back */ (void) snprintf(propbuf, proplen, "none"); return (0); } propbuf[0] = '\0'; for (int i = 0; i < nsnaps; i++) { char buf[128]; if (propbuf[0] != '\0') (void) strlcat(propbuf, ",", proplen); (void) snprintf(buf, sizeof (buf), "%llu", (u_longlong_t)snaps[i]); (void) strlcat(propbuf, buf, proplen); } return (0); } /* * Accepts a property and value and checks that the value * matches the one found by the channel program. If they are * not equal, print both of them. */ static void zcp_check(zfs_handle_t *zhp, zfs_prop_t prop, uint64_t intval, const char *strval) { if (!zhp->zfs_hdl->libzfs_prop_debug) return; int error; char *poolname = zhp->zpool_hdl->zpool_name; const char *prop_name = zfs_prop_to_name(prop); const char *program = "args = ...\n" "ds = args['dataset']\n" "prop = args['property']\n" "value, setpoint = zfs.get_prop(ds, prop)\n" "return {value=value, setpoint=setpoint}\n"; nvlist_t *outnvl; nvlist_t *retnvl; nvlist_t *argnvl = fnvlist_alloc(); fnvlist_add_string(argnvl, "dataset", zhp->zfs_name); fnvlist_add_string(argnvl, "property", zfs_prop_to_name(prop)); error = lzc_channel_program_nosync(poolname, program, 10 * 1000 * 1000, 10 * 1024 * 1024, argnvl, &outnvl); if (error == 0) { retnvl = fnvlist_lookup_nvlist(outnvl, "return"); if (zfs_prop_get_type(prop) == PROP_TYPE_NUMBER) { int64_t ans; error = nvlist_lookup_int64(retnvl, "value", &ans); if (error != 0) { (void) fprintf(stderr, "%s: zcp check error: " "%u\n", prop_name, error); return; } if (ans != intval) { (void) fprintf(stderr, "%s: zfs found %llu, " "but zcp found %llu\n", prop_name, (u_longlong_t)intval, (u_longlong_t)ans); } } else { const char *str_ans; error = nvlist_lookup_string(retnvl, "value", &str_ans); if (error != 0) { (void) fprintf(stderr, "%s: zcp check error: " "%u\n", prop_name, error); return; } if (strcmp(strval, str_ans) != 0) { (void) fprintf(stderr, "%s: zfs found '%s', but zcp found '%s'\n", prop_name, strval, str_ans); } } } else { (void) fprintf(stderr, "%s: zcp check failed, channel program " "error: %u\n", prop_name, error); } nvlist_free(argnvl); nvlist_free(outnvl); } /* * Retrieve a property from the given object. If 'literal' is specified, then * numbers are left as exact values. Otherwise, numbers are converted to a * human-readable form. * * Returns 0 on success, or -1 on error. */ int zfs_prop_get(zfs_handle_t *zhp, zfs_prop_t prop, char *propbuf, size_t proplen, zprop_source_t *src, char *statbuf, size_t statlen, boolean_t literal) { const char *source = NULL; uint64_t val; const char *str; const char *strval; boolean_t received = zfs_is_recvd_props_mode(zhp); /* * Check to see if this property applies to our object */ if (!zfs_prop_valid_for_type(prop, zhp->zfs_type, B_FALSE)) return (-1); if (received && zfs_prop_readonly(prop)) return (-1); if (src) *src = ZPROP_SRC_NONE; switch (prop) { case ZFS_PROP_CREATION: /* * 'creation' is a time_t stored in the statistics. We convert * this into a string unless 'literal' is specified. */ { val = getprop_uint64(zhp, prop, &source); time_t time = (time_t)val; struct tm t; if (literal || localtime_r(&time, &t) == NULL || strftime(propbuf, proplen, "%a %b %e %k:%M %Y", &t) == 0) (void) snprintf(propbuf, proplen, "%llu", (u_longlong_t)val); } zcp_check(zhp, prop, val, NULL); break; case ZFS_PROP_MOUNTPOINT: /* * Getting the precise mountpoint can be tricky. * * - for 'none' or 'legacy', return those values. * - for inherited mountpoints, we want to take everything * after our ancestor and append it to the inherited value. * * If the pool has an alternate root, we want to prepend that * root to any values we return. */ str = getprop_string(zhp, prop, &source); if (str[0] == '/') { char buf[MAXPATHLEN]; char *root = buf; const char *relpath; /* * If we inherit the mountpoint, even from a dataset * with a received value, the source will be the path of * the dataset we inherit from. If source is * ZPROP_SOURCE_VAL_RECVD, the received value is not * inherited. */ if (strcmp(source, ZPROP_SOURCE_VAL_RECVD) == 0) { relpath = ""; } else { relpath = zhp->zfs_name + strlen(source); if (relpath[0] == '/') relpath++; } if ((zpool_get_prop(zhp->zpool_hdl, ZPOOL_PROP_ALTROOT, buf, MAXPATHLEN, NULL, B_FALSE)) || (strcmp(root, "-") == 0)) root[0] = '\0'; /* * Special case an alternate root of '/'. This will * avoid having multiple leading slashes in the * mountpoint path. */ if (strcmp(root, "/") == 0) root++; /* * If the mountpoint is '/' then skip over this * if we are obtaining either an alternate root or * an inherited mountpoint. */ if (str[1] == '\0' && (root[0] != '\0' || relpath[0] != '\0')) str++; if (relpath[0] == '\0') (void) snprintf(propbuf, proplen, "%s%s", root, str); else (void) snprintf(propbuf, proplen, "%s%s%s%s", root, str, relpath[0] == '@' ? "" : "/", relpath); } else { /* 'legacy' or 'none' */ (void) strlcpy(propbuf, str, proplen); } zcp_check(zhp, prop, 0, propbuf); break; case ZFS_PROP_ORIGIN: if (*zhp->zfs_dmustats.dds_origin != '\0') { str = (char *)&zhp->zfs_dmustats.dds_origin; } else { str = getprop_string(zhp, prop, &source); } if (str == NULL || *str == '\0') str = zfs_prop_default_string(prop); if (str == NULL) return (-1); (void) strlcpy(propbuf, str, proplen); zcp_check(zhp, prop, 0, str); break; case ZFS_PROP_REDACT_SNAPS: if (get_rsnaps_string(zhp, propbuf, proplen) != 0) return (-1); break; case ZFS_PROP_CLONES: if (get_clones_string(zhp, propbuf, proplen) != 0) return (-1); break; case ZFS_PROP_QUOTA: case ZFS_PROP_REFQUOTA: case ZFS_PROP_RESERVATION: case ZFS_PROP_REFRESERVATION: if (get_numeric_property(zhp, prop, src, &source, &val) != 0) return (-1); /* * If quota or reservation is 0, we translate this into 'none' * (unless literal is set), and indicate that it's the default * value. Otherwise, we print the number nicely and indicate * that its set locally. */ if (val == 0) { if (literal) (void) strlcpy(propbuf, "0", proplen); else (void) strlcpy(propbuf, "none", proplen); } else { if (literal) (void) snprintf(propbuf, proplen, "%llu", (u_longlong_t)val); else zfs_nicebytes(val, propbuf, proplen); } zcp_check(zhp, prop, val, NULL); break; case ZFS_PROP_FILESYSTEM_LIMIT: case ZFS_PROP_SNAPSHOT_LIMIT: case ZFS_PROP_FILESYSTEM_COUNT: case ZFS_PROP_SNAPSHOT_COUNT: if (get_numeric_property(zhp, prop, src, &source, &val) != 0) return (-1); /* * If limit is UINT64_MAX, we translate this into 'none', and * indicate that it's the default value. Otherwise, we print * the number nicely and indicate that it's set locally. */ if (val == UINT64_MAX) { (void) strlcpy(propbuf, "none", proplen); } else if (literal) { (void) snprintf(propbuf, proplen, "%llu", (u_longlong_t)val); } else { zfs_nicenum(val, propbuf, proplen); } zcp_check(zhp, prop, val, NULL); break; case ZFS_PROP_REFRATIO: case ZFS_PROP_COMPRESSRATIO: if (get_numeric_property(zhp, prop, src, &source, &val) != 0) return (-1); if (literal) (void) snprintf(propbuf, proplen, "%llu.%02llu", (u_longlong_t)(val / 100), (u_longlong_t)(val % 100)); else (void) snprintf(propbuf, proplen, "%llu.%02llux", (u_longlong_t)(val / 100), (u_longlong_t)(val % 100)); zcp_check(zhp, prop, val, NULL); break; case ZFS_PROP_TYPE: switch (zhp->zfs_type) { case ZFS_TYPE_FILESYSTEM: str = "filesystem"; break; case ZFS_TYPE_VOLUME: str = "volume"; break; case ZFS_TYPE_SNAPSHOT: str = "snapshot"; break; case ZFS_TYPE_BOOKMARK: str = "bookmark"; break; default: abort(); } (void) snprintf(propbuf, proplen, "%s", str); zcp_check(zhp, prop, 0, propbuf); break; case ZFS_PROP_MOUNTED: /* * The 'mounted' property is a pseudo-property that described * whether the filesystem is currently mounted. Even though * it's a boolean value, the typical values of "on" and "off" * don't make sense, so we translate to "yes" and "no". */ if (get_numeric_property(zhp, ZFS_PROP_MOUNTED, src, &source, &val) != 0) return (-1); if (val) (void) strlcpy(propbuf, "yes", proplen); else (void) strlcpy(propbuf, "no", proplen); break; case ZFS_PROP_NAME: /* * The 'name' property is a pseudo-property derived from the * dataset name. It is presented as a real property to simplify * consumers. */ (void) strlcpy(propbuf, zhp->zfs_name, proplen); zcp_check(zhp, prop, 0, propbuf); break; case ZFS_PROP_MLSLABEL: { #ifdef HAVE_MLSLABEL m_label_t *new_sl = NULL; char *ascii = NULL; /* human readable label */ (void) strlcpy(propbuf, getprop_string(zhp, prop, &source), proplen); if (literal || (strcasecmp(propbuf, ZFS_MLSLABEL_DEFAULT) == 0)) break; /* * Try to translate the internal hex string to * human-readable output. If there are any * problems just use the hex string. */ if (str_to_label(propbuf, &new_sl, MAC_LABEL, L_NO_CORRECTION, NULL) == -1) { m_label_free(new_sl); break; } if (label_to_str(new_sl, &ascii, M_LABEL, DEF_NAMES) != 0) { if (ascii) free(ascii); m_label_free(new_sl); break; } m_label_free(new_sl); (void) strlcpy(propbuf, ascii, proplen); free(ascii); #else (void) strlcpy(propbuf, getprop_string(zhp, prop, &source), proplen); #endif /* HAVE_MLSLABEL */ } break; case ZFS_PROP_GUID: case ZFS_PROP_KEY_GUID: case ZFS_PROP_IVSET_GUID: case ZFS_PROP_CREATETXG: case ZFS_PROP_OBJSETID: case ZFS_PROP_PBKDF2_ITERS: /* * These properties are stored as numbers, but they are * identifiers or counters. * We don't want them to be pretty printed, because pretty * printing truncates their values making them useless. */ if (get_numeric_property(zhp, prop, src, &source, &val) != 0) return (-1); (void) snprintf(propbuf, proplen, "%llu", (u_longlong_t)val); zcp_check(zhp, prop, val, NULL); break; case ZFS_PROP_REFERENCED: case ZFS_PROP_AVAILABLE: case ZFS_PROP_USED: case ZFS_PROP_USEDSNAP: case ZFS_PROP_USEDDS: case ZFS_PROP_USEDREFRESERV: case ZFS_PROP_USEDCHILD: if (get_numeric_property(zhp, prop, src, &source, &val) != 0) return (-1); if (literal) { (void) snprintf(propbuf, proplen, "%llu", (u_longlong_t)val); } else { zfs_nicebytes(val, propbuf, proplen); } zcp_check(zhp, prop, val, NULL); break; case ZFS_PROP_SNAPSHOTS_CHANGED: { if ((get_numeric_property(zhp, prop, src, &source, &val) != 0) || val == 0) { return (-1); } time_t time = (time_t)val; struct tm t; if (literal || localtime_r(&time, &t) == NULL || strftime(propbuf, proplen, "%a %b %e %k:%M:%S %Y", &t) == 0) (void) snprintf(propbuf, proplen, "%llu", (u_longlong_t)val); } zcp_check(zhp, prop, val, NULL); break; default: switch (zfs_prop_get_type(prop)) { case PROP_TYPE_NUMBER: if (get_numeric_property(zhp, prop, src, &source, &val) != 0) { return (-1); } if (literal) { (void) snprintf(propbuf, proplen, "%llu", (u_longlong_t)val); } else { zfs_nicenum(val, propbuf, proplen); } zcp_check(zhp, prop, val, NULL); break; case PROP_TYPE_STRING: str = getprop_string(zhp, prop, &source); if (str == NULL) return (-1); (void) strlcpy(propbuf, str, proplen); zcp_check(zhp, prop, 0, str); break; case PROP_TYPE_INDEX: if (get_numeric_property(zhp, prop, src, &source, &val) != 0) return (-1); if (zfs_prop_index_to_string(prop, val, &strval) != 0) return (-1); (void) strlcpy(propbuf, strval, proplen); zcp_check(zhp, prop, 0, strval); break; default: abort(); } } get_source(zhp, src, source, statbuf, statlen); return (0); } /* * Utility function to get the given numeric property. Does no validation that * the given property is the appropriate type; should only be used with * hard-coded property types. */ uint64_t zfs_prop_get_int(zfs_handle_t *zhp, zfs_prop_t prop) { const char *source; uint64_t val = 0; (void) get_numeric_property(zhp, prop, NULL, &source, &val); return (val); } static int zfs_prop_set_int(zfs_handle_t *zhp, zfs_prop_t prop, uint64_t val) { char buf[64]; (void) snprintf(buf, sizeof (buf), "%llu", (longlong_t)val); return (zfs_prop_set(zhp, zfs_prop_to_name(prop), buf)); } /* * Similar to zfs_prop_get(), but returns the value as an integer. */ int zfs_prop_get_numeric(zfs_handle_t *zhp, zfs_prop_t prop, uint64_t *value, zprop_source_t *src, char *statbuf, size_t statlen) { const char *source; /* * Check to see if this property applies to our object */ if (!zfs_prop_valid_for_type(prop, zhp->zfs_type, B_FALSE)) { return (zfs_error_fmt(zhp->zfs_hdl, EZFS_PROPTYPE, dgettext(TEXT_DOMAIN, "cannot get property '%s'"), zfs_prop_to_name(prop))); } if (src) *src = ZPROP_SRC_NONE; if (get_numeric_property(zhp, prop, src, &source, value) != 0) return (-1); get_source(zhp, src, source, statbuf, statlen); return (0); } #ifdef HAVE_IDMAP static int idmap_id_to_numeric_domain_rid(uid_t id, boolean_t isuser, char **domainp, idmap_rid_t *ridp) { idmap_get_handle_t *get_hdl = NULL; idmap_stat status; int err = EINVAL; if (idmap_get_create(&get_hdl) != IDMAP_SUCCESS) goto out; if (isuser) { err = idmap_get_sidbyuid(get_hdl, id, IDMAP_REQ_FLG_USE_CACHE, domainp, ridp, &status); } else { err = idmap_get_sidbygid(get_hdl, id, IDMAP_REQ_FLG_USE_CACHE, domainp, ridp, &status); } if (err == IDMAP_SUCCESS && idmap_get_mappings(get_hdl) == IDMAP_SUCCESS && status == IDMAP_SUCCESS) err = 0; else err = EINVAL; out: if (get_hdl) idmap_get_destroy(get_hdl); return (err); } #endif /* HAVE_IDMAP */ /* * convert the propname into parameters needed by kernel * Eg: userquota@ahrens -> ZFS_PROP_USERQUOTA, "", 126829 * Eg: userused@matt@domain -> ZFS_PROP_USERUSED, "S-1-123-456", 789 * Eg: groupquota@staff -> ZFS_PROP_GROUPQUOTA, "", 1234 * Eg: groupused@staff -> ZFS_PROP_GROUPUSED, "", 1234 * Eg: projectquota@123 -> ZFS_PROP_PROJECTQUOTA, "", 123 * Eg: projectused@789 -> ZFS_PROP_PROJECTUSED, "", 789 */ static int userquota_propname_decode(const char *propname, boolean_t zoned, zfs_userquota_prop_t *typep, char *domain, int domainlen, uint64_t *ridp) { zfs_userquota_prop_t type; char *cp; boolean_t isuser; boolean_t isgroup; boolean_t isproject; struct passwd *pw; struct group *gr; domain[0] = '\0'; /* Figure out the property type ({user|group|project}{quota|space}) */ for (type = 0; type < ZFS_NUM_USERQUOTA_PROPS; type++) { if (strncmp(propname, zfs_userquota_prop_prefixes[type], strlen(zfs_userquota_prop_prefixes[type])) == 0) break; } if (type == ZFS_NUM_USERQUOTA_PROPS) return (EINVAL); *typep = type; isuser = (type == ZFS_PROP_USERQUOTA || type == ZFS_PROP_USERUSED || type == ZFS_PROP_USEROBJQUOTA || type == ZFS_PROP_USEROBJUSED); isgroup = (type == ZFS_PROP_GROUPQUOTA || type == ZFS_PROP_GROUPUSED || type == ZFS_PROP_GROUPOBJQUOTA || type == ZFS_PROP_GROUPOBJUSED); isproject = (type == ZFS_PROP_PROJECTQUOTA || type == ZFS_PROP_PROJECTUSED || type == ZFS_PROP_PROJECTOBJQUOTA || type == ZFS_PROP_PROJECTOBJUSED); cp = strchr(propname, '@') + 1; if (isuser && (pw = getpwnam(cp)) != NULL) { if (zoned && getzoneid() == GLOBAL_ZONEID) return (ENOENT); *ridp = pw->pw_uid; } else if (isgroup && (gr = getgrnam(cp)) != NULL) { if (zoned && getzoneid() == GLOBAL_ZONEID) return (ENOENT); *ridp = gr->gr_gid; } else if (!isproject && strchr(cp, '@')) { #ifdef HAVE_IDMAP /* * It's a SID name (eg "user@domain") that needs to be * turned into S-1-domainID-RID. */ directory_error_t e; char *numericsid = NULL; char *end; if (zoned && getzoneid() == GLOBAL_ZONEID) return (ENOENT); if (isuser) { e = directory_sid_from_user_name(NULL, cp, &numericsid); } else { e = directory_sid_from_group_name(NULL, cp, &numericsid); } if (e != NULL) { directory_error_free(e); return (ENOENT); } if (numericsid == NULL) return (ENOENT); cp = numericsid; (void) strlcpy(domain, cp, domainlen); cp = strrchr(domain, '-'); *cp = '\0'; cp++; errno = 0; *ridp = strtoull(cp, &end, 10); free(numericsid); if (errno != 0 || *end != '\0') return (EINVAL); #else (void) domainlen; return (ENOSYS); #endif /* HAVE_IDMAP */ } else { /* It's a user/group/project ID (eg "12345"). */ uid_t id; char *end; id = strtoul(cp, &end, 10); if (*end != '\0') return (EINVAL); if (id > MAXUID && !isproject) { #ifdef HAVE_IDMAP /* It's an ephemeral ID. */ idmap_rid_t rid; char *mapdomain; if (idmap_id_to_numeric_domain_rid(id, isuser, &mapdomain, &rid) != 0) return (ENOENT); (void) strlcpy(domain, mapdomain, domainlen); *ridp = rid; #else return (ENOSYS); #endif /* HAVE_IDMAP */ } else { *ridp = id; } } return (0); } static int zfs_prop_get_userquota_common(zfs_handle_t *zhp, const char *propname, uint64_t *propvalue, zfs_userquota_prop_t *typep) { int err; zfs_cmd_t zc = {"\0"}; (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); err = userquota_propname_decode(propname, zfs_prop_get_int(zhp, ZFS_PROP_ZONED), typep, zc.zc_value, sizeof (zc.zc_value), &zc.zc_guid); zc.zc_objset_type = *typep; if (err) return (err); err = zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_USERSPACE_ONE, &zc); if (err) return (err); *propvalue = zc.zc_cookie; return (0); } int zfs_prop_get_userquota_int(zfs_handle_t *zhp, const char *propname, uint64_t *propvalue) { zfs_userquota_prop_t type; return (zfs_prop_get_userquota_common(zhp, propname, propvalue, &type)); } int zfs_prop_get_userquota(zfs_handle_t *zhp, const char *propname, char *propbuf, int proplen, boolean_t literal) { int err; uint64_t propvalue; zfs_userquota_prop_t type; err = zfs_prop_get_userquota_common(zhp, propname, &propvalue, &type); if (err) return (err); if (literal) { (void) snprintf(propbuf, proplen, "%llu", (u_longlong_t)propvalue); } else if (propvalue == 0 && (type == ZFS_PROP_USERQUOTA || type == ZFS_PROP_GROUPQUOTA || type == ZFS_PROP_USEROBJQUOTA || type == ZFS_PROP_GROUPOBJQUOTA || type == ZFS_PROP_PROJECTQUOTA || type == ZFS_PROP_PROJECTOBJQUOTA)) { (void) strlcpy(propbuf, "none", proplen); } else if (type == ZFS_PROP_USERQUOTA || type == ZFS_PROP_GROUPQUOTA || type == ZFS_PROP_USERUSED || type == ZFS_PROP_GROUPUSED || type == ZFS_PROP_PROJECTUSED || type == ZFS_PROP_PROJECTQUOTA) { zfs_nicebytes(propvalue, propbuf, proplen); } else { zfs_nicenum(propvalue, propbuf, proplen); } return (0); } /* * propname must start with "written@" or "written#". */ int zfs_prop_get_written_int(zfs_handle_t *zhp, const char *propname, uint64_t *propvalue) { int err; zfs_cmd_t zc = {"\0"}; const char *snapname; (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); assert(zfs_prop_written(propname)); snapname = propname + strlen("written@"); if (strchr(snapname, '@') != NULL || strchr(snapname, '#') != NULL) { /* full snapshot or bookmark name specified */ (void) strlcpy(zc.zc_value, snapname, sizeof (zc.zc_value)); } else { /* snapname is the short name, append it to zhp's fsname */ char *cp; (void) strlcpy(zc.zc_value, zhp->zfs_name, sizeof (zc.zc_value)); cp = strchr(zc.zc_value, '@'); if (cp != NULL) *cp = '\0'; (void) strlcat(zc.zc_value, snapname - 1, sizeof (zc.zc_value)); } err = zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_SPACE_WRITTEN, &zc); if (err) return (err); *propvalue = zc.zc_cookie; return (0); } int zfs_prop_get_written(zfs_handle_t *zhp, const char *propname, char *propbuf, int proplen, boolean_t literal) { int err; uint64_t propvalue; err = zfs_prop_get_written_int(zhp, propname, &propvalue); if (err) return (err); if (literal) { (void) snprintf(propbuf, proplen, "%llu", (u_longlong_t)propvalue); } else { zfs_nicebytes(propvalue, propbuf, proplen); } return (0); } /* * Returns the name of the given zfs handle. */ const char * zfs_get_name(const zfs_handle_t *zhp) { return (zhp->zfs_name); } /* * Returns the name of the parent pool for the given zfs handle. */ const char * zfs_get_pool_name(const zfs_handle_t *zhp) { return (zhp->zpool_hdl->zpool_name); } /* * Returns the type of the given zfs handle. */ zfs_type_t zfs_get_type(const zfs_handle_t *zhp) { return (zhp->zfs_type); } /* * Returns the type of the given zfs handle, * or, if a snapshot, the type of the snapshotted dataset. */ zfs_type_t zfs_get_underlying_type(const zfs_handle_t *zhp) { return (zhp->zfs_head_type); } /* * Is one dataset name a child dataset of another? * * Needs to handle these cases: * Dataset 1 "a/foo" "a/foo" "a/foo" "a/foo" * Dataset 2 "a/fo" "a/foobar" "a/bar/baz" "a/foo/bar" * Descendant? No. No. No. Yes. */ static boolean_t is_descendant(const char *ds1, const char *ds2) { size_t d1len = strlen(ds1); /* ds2 can't be a descendant if it's smaller */ if (strlen(ds2) < d1len) return (B_FALSE); /* otherwise, compare strings and verify that there's a '/' char */ return (ds2[d1len] == '/' && (strncmp(ds1, ds2, d1len) == 0)); } /* * Given a complete name, return just the portion that refers to the parent. * Will return -1 if there is no parent (path is just the name of the * pool). */ static int parent_name(const char *path, char *buf, size_t buflen) { char *slashp; (void) strlcpy(buf, path, buflen); if ((slashp = strrchr(buf, '/')) == NULL) return (-1); *slashp = '\0'; return (0); } int zfs_parent_name(zfs_handle_t *zhp, char *buf, size_t buflen) { return (parent_name(zfs_get_name(zhp), buf, buflen)); } /* * If accept_ancestor is false, then check to make sure that the given path has * a parent, and that it exists. If accept_ancestor is true, then find the * closest existing ancestor for the given path. In prefixlen return the * length of already existing prefix of the given path. We also fetch the * 'zoned' property, which is used to validate property settings when creating * new datasets. */ static int check_parents(libzfs_handle_t *hdl, const char *path, uint64_t *zoned, boolean_t accept_ancestor, int *prefixlen) { zfs_cmd_t zc = {"\0"}; char parent[ZFS_MAX_DATASET_NAME_LEN]; char *slash; zfs_handle_t *zhp; char errbuf[ERRBUFLEN]; uint64_t is_zoned; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot create '%s'"), path); /* get parent, and check to see if this is just a pool */ if (parent_name(path, parent, sizeof (parent)) != 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "missing dataset name")); return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); } /* check to see if the pool exists */ if ((slash = strchr(parent, '/')) == NULL) slash = parent + strlen(parent); (void) strlcpy(zc.zc_name, parent, MIN(sizeof (zc.zc_name), slash - parent + 1)); if (zfs_ioctl(hdl, ZFS_IOC_OBJSET_STATS, &zc) != 0 && errno == ENOENT) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "no such pool '%s'"), zc.zc_name); return (zfs_error(hdl, EZFS_NOENT, errbuf)); } /* check to see if the parent dataset exists */ while ((zhp = make_dataset_handle(hdl, parent)) == NULL) { if (errno == ENOENT && accept_ancestor) { /* * Go deeper to find an ancestor, give up on top level. */ if (parent_name(parent, parent, sizeof (parent)) != 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "no such pool '%s'"), zc.zc_name); return (zfs_error(hdl, EZFS_NOENT, errbuf)); } } else if (errno == ENOENT) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "parent does not exist")); return (zfs_error(hdl, EZFS_NOENT, errbuf)); } else return (zfs_standard_error(hdl, errno, errbuf)); } is_zoned = zfs_prop_get_int(zhp, ZFS_PROP_ZONED); if (zoned != NULL) *zoned = is_zoned; /* we are in a non-global zone, but parent is in the global zone */ if (getzoneid() != GLOBAL_ZONEID && !is_zoned) { (void) zfs_standard_error(hdl, EPERM, errbuf); zfs_close(zhp); return (-1); } /* make sure parent is a filesystem */ if (zfs_get_type(zhp) != ZFS_TYPE_FILESYSTEM) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "parent is not a filesystem")); (void) zfs_error(hdl, EZFS_BADTYPE, errbuf); zfs_close(zhp); return (-1); } zfs_close(zhp); if (prefixlen != NULL) *prefixlen = strlen(parent); return (0); } /* * Finds whether the dataset of the given type(s) exists. */ boolean_t zfs_dataset_exists(libzfs_handle_t *hdl, const char *path, zfs_type_t types) { zfs_handle_t *zhp; if (!zfs_validate_name(hdl, path, types, B_FALSE)) return (B_FALSE); /* * Try to get stats for the dataset, which will tell us if it exists. */ if ((zhp = make_dataset_handle(hdl, path)) != NULL) { int ds_type = zhp->zfs_type; zfs_close(zhp); if (types & ds_type) return (B_TRUE); } return (B_FALSE); } /* * Given a path to 'target', create all the ancestors between * the prefixlen portion of the path, and the target itself. * Fail if the initial prefixlen-ancestor does not already exist. */ int create_parents(libzfs_handle_t *hdl, char *target, int prefixlen) { zfs_handle_t *h; char *cp; const char *opname; /* make sure prefix exists */ cp = target + prefixlen; if (*cp != '/') { assert(strchr(cp, '/') == NULL); h = zfs_open(hdl, target, ZFS_TYPE_FILESYSTEM); } else { *cp = '\0'; h = zfs_open(hdl, target, ZFS_TYPE_FILESYSTEM); *cp = '/'; } if (h == NULL) return (-1); zfs_close(h); /* * Attempt to create, mount, and share any ancestor filesystems, * up to the prefixlen-long one. */ for (cp = target + prefixlen + 1; (cp = strchr(cp, '/')) != NULL; *cp = '/', cp++) { *cp = '\0'; h = make_dataset_handle(hdl, target); if (h) { /* it already exists, nothing to do here */ zfs_close(h); continue; } if (zfs_create(hdl, target, ZFS_TYPE_FILESYSTEM, NULL) != 0) { opname = dgettext(TEXT_DOMAIN, "create"); goto ancestorerr; } h = zfs_open(hdl, target, ZFS_TYPE_FILESYSTEM); if (h == NULL) { opname = dgettext(TEXT_DOMAIN, "open"); goto ancestorerr; } if (zfs_mount(h, NULL, 0) != 0) { opname = dgettext(TEXT_DOMAIN, "mount"); goto ancestorerr; } if (zfs_share(h, NULL) != 0) { opname = dgettext(TEXT_DOMAIN, "share"); goto ancestorerr; } zfs_close(h); } zfs_commit_shares(NULL); return (0); ancestorerr: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "failed to %s ancestor '%s'"), opname, target); return (-1); } /* * Creates non-existing ancestors of the given path. */ int zfs_create_ancestors(libzfs_handle_t *hdl, const char *path) { int prefix; char *path_copy; char errbuf[ERRBUFLEN]; int rc = 0; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot create '%s'"), path); /* * Check that we are not passing the nesting limit * before we start creating any ancestors. */ if (dataset_nestcheck(path) != 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "maximum name nesting depth exceeded")); return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); } if (check_parents(hdl, path, NULL, B_TRUE, &prefix) != 0) return (-1); if ((path_copy = strdup(path)) != NULL) { rc = create_parents(hdl, path_copy, prefix); free(path_copy); } if (path_copy == NULL || rc != 0) return (-1); return (0); } /* * Create a new filesystem or volume. */ int zfs_create(libzfs_handle_t *hdl, const char *path, zfs_type_t type, nvlist_t *props) { int ret; uint64_t size = 0; uint64_t blocksize = zfs_prop_default_numeric(ZFS_PROP_VOLBLOCKSIZE); uint64_t zoned; enum lzc_dataset_type ost; zpool_handle_t *zpool_handle; uint8_t *wkeydata = NULL; uint_t wkeylen = 0; char errbuf[ERRBUFLEN]; char parent[ZFS_MAX_DATASET_NAME_LEN]; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot create '%s'"), path); /* validate the path, taking care to note the extended error message */ if (!zfs_validate_name(hdl, path, type, B_TRUE)) return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); if (dataset_nestcheck(path) != 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "maximum name nesting depth exceeded")); return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); } /* validate parents exist */ if (check_parents(hdl, path, &zoned, B_FALSE, NULL) != 0) return (-1); /* * The failure modes when creating a dataset of a different type over * one that already exists is a little strange. In particular, if you * try to create a dataset on top of an existing dataset, the ioctl() * will return ENOENT, not EEXIST. To prevent this from happening, we * first try to see if the dataset exists. */ if (zfs_dataset_exists(hdl, path, ZFS_TYPE_DATASET)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "dataset already exists")); return (zfs_error(hdl, EZFS_EXISTS, errbuf)); } if (type == ZFS_TYPE_VOLUME) ost = LZC_DATSET_TYPE_ZVOL; else ost = LZC_DATSET_TYPE_ZFS; /* open zpool handle for prop validation */ char pool_path[ZFS_MAX_DATASET_NAME_LEN]; (void) strlcpy(pool_path, path, sizeof (pool_path)); /* truncate pool_path at first slash */ char *p = strchr(pool_path, '/'); if (p != NULL) *p = '\0'; if ((zpool_handle = zpool_open(hdl, pool_path)) == NULL) return (-1); if (props && (props = zfs_valid_proplist(hdl, type, props, zoned, NULL, zpool_handle, B_TRUE, errbuf)) == 0) { zpool_close(zpool_handle); return (-1); } zpool_close(zpool_handle); if (type == ZFS_TYPE_VOLUME) { /* * If we are creating a volume, the size and block size must * satisfy a few restraints. First, the blocksize must be a * valid block size between SPA_{MIN,MAX}BLOCKSIZE. Second, the * volsize must be a multiple of the block size, and cannot be * zero. */ if (props == NULL || nvlist_lookup_uint64(props, zfs_prop_to_name(ZFS_PROP_VOLSIZE), &size) != 0) { nvlist_free(props); zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "missing volume size")); return (zfs_error(hdl, EZFS_BADPROP, errbuf)); } if ((ret = nvlist_lookup_uint64(props, zfs_prop_to_name(ZFS_PROP_VOLBLOCKSIZE), &blocksize)) != 0) { if (ret == ENOENT) { blocksize = zfs_prop_default_numeric( ZFS_PROP_VOLBLOCKSIZE); } else { nvlist_free(props); zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "missing volume block size")); return (zfs_error(hdl, EZFS_BADPROP, errbuf)); } } if (size == 0) { nvlist_free(props); zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "volume size cannot be zero")); return (zfs_error(hdl, EZFS_BADPROP, errbuf)); } if (size % blocksize != 0) { nvlist_free(props); zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "volume size must be a multiple of volume block " "size")); return (zfs_error(hdl, EZFS_BADPROP, errbuf)); } } (void) parent_name(path, parent, sizeof (parent)); if (zfs_crypto_create(hdl, parent, props, NULL, B_TRUE, &wkeydata, &wkeylen) != 0) { nvlist_free(props); return (zfs_error(hdl, EZFS_CRYPTOFAILED, errbuf)); } /* create the dataset */ ret = lzc_create(path, ost, props, wkeydata, wkeylen); nvlist_free(props); if (wkeydata != NULL) free(wkeydata); /* check for failure */ if (ret != 0) { switch (errno) { case ENOENT: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "no such parent '%s'"), parent); return (zfs_error(hdl, EZFS_NOENT, errbuf)); case ENOTSUP: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool must be upgraded to set this " "property or value")); return (zfs_error(hdl, EZFS_BADVERSION, errbuf)); case EACCES: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "encryption root's key is not loaded " "or provided")); return (zfs_error(hdl, EZFS_CRYPTOFAILED, errbuf)); case ERANGE: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid property value(s) specified")); return (zfs_error(hdl, EZFS_BADPROP, errbuf)); #ifdef _ILP32 case EOVERFLOW: /* * This platform can't address a volume this big. */ if (type == ZFS_TYPE_VOLUME) return (zfs_error(hdl, EZFS_VOLTOOBIG, errbuf)); zfs_fallthrough; #endif default: return (zfs_standard_error(hdl, errno, errbuf)); } } return (0); } /* * Destroys the given dataset. The caller must make sure that the filesystem * isn't mounted, and that there are no active dependents. If the file system * does not exist this function does nothing. */ int zfs_destroy(zfs_handle_t *zhp, boolean_t defer) { int error; if (zhp->zfs_type != ZFS_TYPE_SNAPSHOT && defer) return (EINVAL); if (zhp->zfs_type == ZFS_TYPE_BOOKMARK) { nvlist_t *nv = fnvlist_alloc(); fnvlist_add_boolean(nv, zhp->zfs_name); error = lzc_destroy_bookmarks(nv, NULL); fnvlist_free(nv); if (error != 0) { return (zfs_standard_error_fmt(zhp->zfs_hdl, error, dgettext(TEXT_DOMAIN, "cannot destroy '%s'"), zhp->zfs_name)); } return (0); } if (zhp->zfs_type == ZFS_TYPE_SNAPSHOT) { nvlist_t *nv = fnvlist_alloc(); fnvlist_add_boolean(nv, zhp->zfs_name); error = lzc_destroy_snaps(nv, defer, NULL); fnvlist_free(nv); } else { error = lzc_destroy(zhp->zfs_name); } if (error != 0 && error != ENOENT) { return (zfs_standard_error_fmt(zhp->zfs_hdl, errno, dgettext(TEXT_DOMAIN, "cannot destroy '%s'"), zhp->zfs_name)); } remove_mountpoint(zhp); return (0); } struct destroydata { nvlist_t *nvl; const char *snapname; }; static int zfs_check_snap_cb(zfs_handle_t *zhp, void *arg) { struct destroydata *dd = arg; char name[ZFS_MAX_DATASET_NAME_LEN]; int rv = 0; if (snprintf(name, sizeof (name), "%s@%s", zhp->zfs_name, dd->snapname) >= sizeof (name)) return (EINVAL); if (lzc_exists(name)) fnvlist_add_boolean(dd->nvl, name); rv = zfs_iter_filesystems_v2(zhp, 0, zfs_check_snap_cb, dd); zfs_close(zhp); return (rv); } /* * Destroys all snapshots with the given name in zhp & descendants. */ int zfs_destroy_snaps(zfs_handle_t *zhp, char *snapname, boolean_t defer) { int ret; struct destroydata dd = { 0 }; dd.snapname = snapname; dd.nvl = fnvlist_alloc(); (void) zfs_check_snap_cb(zfs_handle_dup(zhp), &dd); if (nvlist_empty(dd.nvl)) { ret = zfs_standard_error_fmt(zhp->zfs_hdl, ENOENT, dgettext(TEXT_DOMAIN, "cannot destroy '%s@%s'"), zhp->zfs_name, snapname); } else { ret = zfs_destroy_snaps_nvl(zhp->zfs_hdl, dd.nvl, defer); } fnvlist_free(dd.nvl); return (ret); } /* * Destroys all the snapshots named in the nvlist. */ int zfs_destroy_snaps_nvl(libzfs_handle_t *hdl, nvlist_t *snaps, boolean_t defer) { nvlist_t *errlist = NULL; nvpair_t *pair; int ret = zfs_destroy_snaps_nvl_os(hdl, snaps); if (ret != 0) return (ret); ret = lzc_destroy_snaps(snaps, defer, &errlist); if (ret == 0) { nvlist_free(errlist); return (0); } if (nvlist_empty(errlist)) { char errbuf[ERRBUFLEN]; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot destroy snapshots")); ret = zfs_standard_error(hdl, ret, errbuf); } for (pair = nvlist_next_nvpair(errlist, NULL); pair != NULL; pair = nvlist_next_nvpair(errlist, pair)) { char errbuf[ERRBUFLEN]; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot destroy snapshot %s"), nvpair_name(pair)); switch (fnvpair_value_int32(pair)) { case EEXIST: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "snapshot is cloned")); ret = zfs_error(hdl, EZFS_EXISTS, errbuf); break; default: ret = zfs_standard_error(hdl, errno, errbuf); break; } } nvlist_free(errlist); return (ret); } /* * Clones the given dataset. The target must be of the same type as the source. */ int zfs_clone(zfs_handle_t *zhp, const char *target, nvlist_t *props) { char parent[ZFS_MAX_DATASET_NAME_LEN]; int ret; char errbuf[ERRBUFLEN]; libzfs_handle_t *hdl = zhp->zfs_hdl; uint64_t zoned; assert(zhp->zfs_type == ZFS_TYPE_SNAPSHOT); (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot create '%s'"), target); /* validate the target/clone name */ if (!zfs_validate_name(hdl, target, ZFS_TYPE_FILESYSTEM, B_TRUE)) return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); /* validate parents exist */ if (check_parents(hdl, target, &zoned, B_FALSE, NULL) != 0) return (-1); (void) parent_name(target, parent, sizeof (parent)); /* do the clone */ if (props) { zfs_type_t type = ZFS_TYPE_FILESYSTEM; if (ZFS_IS_VOLUME(zhp)) type = ZFS_TYPE_VOLUME; if ((props = zfs_valid_proplist(hdl, type, props, zoned, zhp, zhp->zpool_hdl, B_TRUE, errbuf)) == NULL) return (-1); if (zfs_fix_auto_resv(zhp, props) == -1) { nvlist_free(props); return (-1); } } if (zfs_crypto_clone_check(hdl, zhp, parent, props) != 0) { nvlist_free(props); return (zfs_error(hdl, EZFS_CRYPTOFAILED, errbuf)); } ret = lzc_clone(target, zhp->zfs_name, props); nvlist_free(props); if (ret != 0) { switch (errno) { case ENOENT: /* * The parent doesn't exist. We should have caught this * above, but there may a race condition that has since * destroyed the parent. * * At this point, we don't know whether it's the source * that doesn't exist anymore, or whether the target * dataset doesn't exist. */ zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN, "no such parent '%s'"), parent); return (zfs_error(zhp->zfs_hdl, EZFS_NOENT, errbuf)); case EXDEV: zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN, "source and target pools differ")); return (zfs_error(zhp->zfs_hdl, EZFS_CROSSTARGET, errbuf)); default: return (zfs_standard_error(zhp->zfs_hdl, errno, errbuf)); } } return (ret); } /* * Promotes the given clone fs to be the clone parent. */ int zfs_promote(zfs_handle_t *zhp) { libzfs_handle_t *hdl = zhp->zfs_hdl; char snapname[ZFS_MAX_DATASET_NAME_LEN]; int ret; char errbuf[ERRBUFLEN]; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot promote '%s'"), zhp->zfs_name); if (zhp->zfs_type == ZFS_TYPE_SNAPSHOT) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "snapshots can not be promoted")); return (zfs_error(hdl, EZFS_BADTYPE, errbuf)); } if (zhp->zfs_dmustats.dds_origin[0] == '\0') { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "not a cloned filesystem")); return (zfs_error(hdl, EZFS_BADTYPE, errbuf)); } if (!zfs_validate_name(hdl, zhp->zfs_name, zhp->zfs_type, B_TRUE)) return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); ret = lzc_promote(zhp->zfs_name, snapname, sizeof (snapname)); if (ret != 0) { switch (ret) { case EACCES: /* * Promoting encrypted dataset outside its * encryption root. */ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "cannot promote dataset outside its " "encryption root")); return (zfs_error(hdl, EZFS_EXISTS, errbuf)); case EEXIST: /* There is a conflicting snapshot name. */ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "conflicting snapshot '%s' from parent '%s'"), snapname, zhp->zfs_dmustats.dds_origin); return (zfs_error(hdl, EZFS_EXISTS, errbuf)); default: return (zfs_standard_error(hdl, ret, errbuf)); } } return (ret); } typedef struct snapdata { nvlist_t *sd_nvl; const char *sd_snapname; } snapdata_t; static int zfs_snapshot_cb(zfs_handle_t *zhp, void *arg) { snapdata_t *sd = arg; char name[ZFS_MAX_DATASET_NAME_LEN]; int rv = 0; if (zfs_prop_get_int(zhp, ZFS_PROP_INCONSISTENT) == 0) { if (snprintf(name, sizeof (name), "%s@%s", zfs_get_name(zhp), sd->sd_snapname) >= sizeof (name)) return (EINVAL); fnvlist_add_boolean(sd->sd_nvl, name); rv = zfs_iter_filesystems_v2(zhp, 0, zfs_snapshot_cb, sd); } zfs_close(zhp); return (rv); } /* * Creates snapshots. The keys in the snaps nvlist are the snapshots to be * created. */ int zfs_snapshot_nvl(libzfs_handle_t *hdl, nvlist_t *snaps, nvlist_t *props) { int ret; char errbuf[ERRBUFLEN]; nvpair_t *elem; nvlist_t *errors; zpool_handle_t *zpool_hdl; char pool[ZFS_MAX_DATASET_NAME_LEN]; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot create snapshots ")); elem = NULL; while ((elem = nvlist_next_nvpair(snaps, elem)) != NULL) { const char *snapname = nvpair_name(elem); /* validate the target name */ if (!zfs_validate_name(hdl, snapname, ZFS_TYPE_SNAPSHOT, B_TRUE)) { (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot create snapshot '%s'"), snapname); return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); } } /* * get pool handle for prop validation. assumes all snaps are in the * same pool, as does lzc_snapshot (below). */ elem = nvlist_next_nvpair(snaps, NULL); if (elem == NULL) return (-1); (void) strlcpy(pool, nvpair_name(elem), sizeof (pool)); pool[strcspn(pool, "/@")] = '\0'; zpool_hdl = zpool_open(hdl, pool); if (zpool_hdl == NULL) return (-1); if (props != NULL && (props = zfs_valid_proplist(hdl, ZFS_TYPE_SNAPSHOT, props, B_FALSE, NULL, zpool_hdl, B_FALSE, errbuf)) == NULL) { zpool_close(zpool_hdl); return (-1); } zpool_close(zpool_hdl); ret = lzc_snapshot(snaps, props, &errors); if (ret != 0) { boolean_t printed = B_FALSE; for (elem = nvlist_next_nvpair(errors, NULL); elem != NULL; elem = nvlist_next_nvpair(errors, elem)) { (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot create snapshot '%s'"), nvpair_name(elem)); (void) zfs_standard_error(hdl, fnvpair_value_int32(elem), errbuf); printed = B_TRUE; } if (!printed) { switch (ret) { case EXDEV: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "multiple snapshots of same " "fs not allowed")); (void) zfs_error(hdl, EZFS_EXISTS, errbuf); break; default: (void) zfs_standard_error(hdl, ret, errbuf); } } } nvlist_free(props); nvlist_free(errors); return (ret); } int zfs_snapshot(libzfs_handle_t *hdl, const char *path, boolean_t recursive, nvlist_t *props) { int ret; snapdata_t sd = { 0 }; char fsname[ZFS_MAX_DATASET_NAME_LEN]; char *cp; zfs_handle_t *zhp; char errbuf[ERRBUFLEN]; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot snapshot %s"), path); if (!zfs_validate_name(hdl, path, ZFS_TYPE_SNAPSHOT, B_TRUE)) return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); (void) strlcpy(fsname, path, sizeof (fsname)); cp = strchr(fsname, '@'); *cp = '\0'; sd.sd_snapname = cp + 1; if ((zhp = zfs_open(hdl, fsname, ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME)) == NULL) { return (-1); } sd.sd_nvl = fnvlist_alloc(); if (recursive) { (void) zfs_snapshot_cb(zfs_handle_dup(zhp), &sd); } else { fnvlist_add_boolean(sd.sd_nvl, path); } ret = zfs_snapshot_nvl(hdl, sd.sd_nvl, props); fnvlist_free(sd.sd_nvl); zfs_close(zhp); return (ret); } /* * Destroy any more recent snapshots. We invoke this callback on any dependents * of the snapshot first. If the 'cb_dependent' member is non-zero, then this * is a dependent and we should just destroy it without checking the transaction * group. */ typedef struct rollback_data { const char *cb_target; /* the snapshot */ uint64_t cb_create; /* creation time reference */ boolean_t cb_error; boolean_t cb_force; } rollback_data_t; static int rollback_destroy_dependent(zfs_handle_t *zhp, void *data) { rollback_data_t *cbp = data; prop_changelist_t *clp; /* We must destroy this clone; first unmount it */ clp = changelist_gather(zhp, ZFS_PROP_NAME, 0, cbp->cb_force ? MS_FORCE: 0); if (clp == NULL || changelist_prefix(clp) != 0) { cbp->cb_error = B_TRUE; zfs_close(zhp); return (0); } if (zfs_destroy(zhp, B_FALSE) != 0) cbp->cb_error = B_TRUE; else changelist_remove(clp, zhp->zfs_name); (void) changelist_postfix(clp); changelist_free(clp); zfs_close(zhp); return (0); } static int rollback_destroy(zfs_handle_t *zhp, void *data) { rollback_data_t *cbp = data; if (zfs_prop_get_int(zhp, ZFS_PROP_CREATETXG) > cbp->cb_create) { cbp->cb_error |= zfs_iter_dependents_v2(zhp, 0, B_FALSE, rollback_destroy_dependent, cbp); cbp->cb_error |= zfs_destroy(zhp, B_FALSE); } zfs_close(zhp); return (0); } /* * Given a dataset, rollback to a specific snapshot, discarding any * data changes since then and making it the active dataset. * * Any snapshots and bookmarks more recent than the target are * destroyed, along with their dependents (i.e. clones). */ int zfs_rollback(zfs_handle_t *zhp, zfs_handle_t *snap, boolean_t force) { rollback_data_t cb = { 0 }; int err; boolean_t restore_resv = 0; uint64_t old_volsize = 0, new_volsize; zfs_prop_t resv_prop = { 0 }; uint64_t min_txg = 0; assert(zhp->zfs_type == ZFS_TYPE_FILESYSTEM || zhp->zfs_type == ZFS_TYPE_VOLUME); /* * Destroy all recent snapshots and their dependents. */ cb.cb_force = force; cb.cb_target = snap->zfs_name; cb.cb_create = zfs_prop_get_int(snap, ZFS_PROP_CREATETXG); if (cb.cb_create > 0) min_txg = cb.cb_create; (void) zfs_iter_snapshots_v2(zhp, 0, rollback_destroy, &cb, min_txg, 0); (void) zfs_iter_bookmarks_v2(zhp, 0, rollback_destroy, &cb); if (cb.cb_error) return (-1); /* * Now that we have verified that the snapshot is the latest, * rollback to the given snapshot. */ if (zhp->zfs_type == ZFS_TYPE_VOLUME) { if (zfs_which_resv_prop(zhp, &resv_prop) < 0) return (-1); old_volsize = zfs_prop_get_int(zhp, ZFS_PROP_VOLSIZE); restore_resv = (old_volsize == zfs_prop_get_int(zhp, resv_prop)); } /* * Pass both the filesystem and the wanted snapshot names, * we would get an error back if the snapshot is destroyed or * a new snapshot is created before this request is processed. */ err = lzc_rollback_to(zhp->zfs_name, snap->zfs_name); if (err != 0) { char errbuf[ERRBUFLEN]; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot rollback '%s'"), zhp->zfs_name); switch (err) { case EEXIST: zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN, "there is a snapshot or bookmark more recent " "than '%s'"), snap->zfs_name); (void) zfs_error(zhp->zfs_hdl, EZFS_EXISTS, errbuf); break; case ESRCH: zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN, "'%s' is not found among snapshots of '%s'"), snap->zfs_name, zhp->zfs_name); (void) zfs_error(zhp->zfs_hdl, EZFS_NOENT, errbuf); break; case EINVAL: (void) zfs_error(zhp->zfs_hdl, EZFS_BADTYPE, errbuf); break; default: (void) zfs_standard_error(zhp->zfs_hdl, err, errbuf); } return (err); } /* * For volumes, if the pre-rollback volsize matched the pre- * rollback reservation and the volsize has changed then set * the reservation property to the post-rollback volsize. * Make a new handle since the rollback closed the dataset. */ if ((zhp->zfs_type == ZFS_TYPE_VOLUME) && (zhp = make_dataset_handle(zhp->zfs_hdl, zhp->zfs_name))) { if (restore_resv) { new_volsize = zfs_prop_get_int(zhp, ZFS_PROP_VOLSIZE); if (old_volsize != new_volsize) err = zfs_prop_set_int(zhp, resv_prop, new_volsize); } zfs_close(zhp); } return (err); } /* * Renames the given dataset. */ int zfs_rename(zfs_handle_t *zhp, const char *target, renameflags_t flags) { int ret = 0; zfs_cmd_t zc = {"\0"}; char *delim; prop_changelist_t *cl = NULL; char parent[ZFS_MAX_DATASET_NAME_LEN]; char property[ZFS_MAXPROPLEN]; libzfs_handle_t *hdl = zhp->zfs_hdl; char errbuf[ERRBUFLEN]; /* if we have the same exact name, just return success */ if (strcmp(zhp->zfs_name, target) == 0) return (0); (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot rename to '%s'"), target); /* make sure source name is valid */ if (!zfs_validate_name(hdl, zhp->zfs_name, zhp->zfs_type, B_TRUE)) return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); /* * Make sure the target name is valid */ if (zhp->zfs_type == ZFS_TYPE_SNAPSHOT) { if ((strchr(target, '@') == NULL) || *target == '@') { /* * Snapshot target name is abbreviated, * reconstruct full dataset name */ (void) strlcpy(parent, zhp->zfs_name, sizeof (parent)); delim = strchr(parent, '@'); if (strchr(target, '@') == NULL) *(++delim) = '\0'; else *delim = '\0'; (void) strlcat(parent, target, sizeof (parent)); target = parent; } else { /* * Make sure we're renaming within the same dataset. */ delim = strchr(target, '@'); if (strncmp(zhp->zfs_name, target, delim - target) != 0 || zhp->zfs_name[delim - target] != '@') { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "snapshots must be part of same " "dataset")); return (zfs_error(hdl, EZFS_CROSSTARGET, errbuf)); } } if (!zfs_validate_name(hdl, target, zhp->zfs_type, B_TRUE)) return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); } else { if (flags.recursive) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "recursive rename must be a snapshot")); return (zfs_error(hdl, EZFS_BADTYPE, errbuf)); } if (!zfs_validate_name(hdl, target, zhp->zfs_type, B_TRUE)) return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); /* validate parents */ if (check_parents(hdl, target, NULL, B_FALSE, NULL) != 0) return (-1); /* make sure we're in the same pool */ verify((delim = strchr(target, '/')) != NULL); if (strncmp(zhp->zfs_name, target, delim - target) != 0 || zhp->zfs_name[delim - target] != '/') { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "datasets must be within same pool")); return (zfs_error(hdl, EZFS_CROSSTARGET, errbuf)); } /* new name cannot be a child of the current dataset name */ if (is_descendant(zhp->zfs_name, target)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "New dataset name cannot be a descendant of " "current dataset name")); return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); } } (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot rename '%s'"), zhp->zfs_name); if (getzoneid() == GLOBAL_ZONEID && zfs_prop_get_int(zhp, ZFS_PROP_ZONED)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "dataset is used in a non-global zone")); return (zfs_error(hdl, EZFS_ZONED, errbuf)); } /* * Avoid unmounting file systems with mountpoint property set to * 'legacy' or 'none' even if -u option is not given. */ if (zhp->zfs_type == ZFS_TYPE_FILESYSTEM && !flags.recursive && !flags.nounmount && zfs_prop_get(zhp, ZFS_PROP_MOUNTPOINT, property, sizeof (property), NULL, NULL, 0, B_FALSE) == 0 && (strcmp(property, "legacy") == 0 || strcmp(property, "none") == 0)) { flags.nounmount = B_TRUE; } if (flags.recursive) { char *parentname = zfs_strdup(zhp->zfs_hdl, zhp->zfs_name); delim = strchr(parentname, '@'); *delim = '\0'; zfs_handle_t *zhrp = zfs_open(zhp->zfs_hdl, parentname, ZFS_TYPE_DATASET); free(parentname); if (zhrp == NULL) { ret = -1; goto error; } zfs_close(zhrp); } else if (zhp->zfs_type != ZFS_TYPE_SNAPSHOT) { if ((cl = changelist_gather(zhp, ZFS_PROP_NAME, flags.nounmount ? CL_GATHER_DONT_UNMOUNT : CL_GATHER_ITER_MOUNTED, flags.forceunmount ? MS_FORCE : 0)) == NULL) return (-1); if (changelist_haszonedchild(cl)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "child dataset with inherited mountpoint is used " "in a non-global zone")); (void) zfs_error(hdl, EZFS_ZONED, errbuf); ret = -1; goto error; } if ((ret = changelist_prefix(cl)) != 0) goto error; } if (ZFS_IS_VOLUME(zhp)) zc.zc_objset_type = DMU_OST_ZVOL; else zc.zc_objset_type = DMU_OST_ZFS; (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); (void) strlcpy(zc.zc_value, target, sizeof (zc.zc_value)); zc.zc_cookie = !!flags.recursive; zc.zc_cookie |= (!!flags.nounmount) << 1; if ((ret = zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_RENAME, &zc)) != 0) { /* * if it was recursive, the one that actually failed will * be in zc.zc_name */ (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot rename '%s'"), zc.zc_name); if (flags.recursive && errno == EEXIST) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "a child dataset already has a snapshot " "with the new name")); (void) zfs_error(hdl, EZFS_EXISTS, errbuf); } else if (errno == EACCES) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "cannot move encrypted child outside of " "its encryption root")); (void) zfs_error(hdl, EZFS_CRYPTOFAILED, errbuf); } else { (void) zfs_standard_error(zhp->zfs_hdl, errno, errbuf); } /* * On failure, we still want to remount any filesystems that * were previously mounted, so we don't alter the system state. */ if (cl != NULL) (void) changelist_postfix(cl); } else { if (cl != NULL) { changelist_rename(cl, zfs_get_name(zhp), target); ret = changelist_postfix(cl); } } error: if (cl != NULL) { changelist_free(cl); } return (ret); } nvlist_t * zfs_get_all_props(zfs_handle_t *zhp) { return (zhp->zfs_props); } nvlist_t * zfs_get_recvd_props(zfs_handle_t *zhp) { if (zhp->zfs_recvd_props == NULL) if (get_recvd_props_ioctl(zhp) != 0) return (NULL); return (zhp->zfs_recvd_props); } nvlist_t * zfs_get_user_props(zfs_handle_t *zhp) { return (zhp->zfs_user_props); } /* * This function is used by 'zfs list' to determine the exact set of columns to * display, and their maximum widths. This does two main things: * * - If this is a list of all properties, then expand the list to include * all native properties, and set a flag so that for each dataset we look * for new unique user properties and add them to the list. * * - For non fixed-width properties, keep track of the maximum width seen * so that we can size the column appropriately. If the user has * requested received property values, we also need to compute the width * of the RECEIVED column. */ int zfs_expand_proplist(zfs_handle_t *zhp, zprop_list_t **plp, boolean_t received, boolean_t literal) { libzfs_handle_t *hdl = zhp->zfs_hdl; zprop_list_t *entry; zprop_list_t **last, **start; nvlist_t *userprops, *propval; nvpair_t *elem; const char *strval; char buf[ZFS_MAXPROPLEN]; if (zprop_expand_list(hdl, plp, ZFS_TYPE_DATASET) != 0) return (-1); userprops = zfs_get_user_props(zhp); entry = *plp; if (entry->pl_all && nvlist_next_nvpair(userprops, NULL) != NULL) { /* * Go through and add any user properties as necessary. We * start by incrementing our list pointer to the first * non-native property. */ start = plp; while (*start != NULL) { if ((*start)->pl_prop == ZPROP_USERPROP) break; start = &(*start)->pl_next; } elem = NULL; while ((elem = nvlist_next_nvpair(userprops, elem)) != NULL) { /* * See if we've already found this property in our list. */ for (last = start; *last != NULL; last = &(*last)->pl_next) { if (strcmp((*last)->pl_user_prop, nvpair_name(elem)) == 0) break; } if (*last == NULL) { entry = zfs_alloc(hdl, sizeof (zprop_list_t)); entry->pl_user_prop = zfs_strdup(hdl, nvpair_name(elem)); entry->pl_prop = ZPROP_USERPROP; entry->pl_width = strlen(nvpair_name(elem)); entry->pl_all = B_TRUE; *last = entry; } } } /* * Now go through and check the width of any non-fixed columns */ for (entry = *plp; entry != NULL; entry = entry->pl_next) { if (entry->pl_fixed && !literal) continue; if (entry->pl_prop != ZPROP_USERPROP) { if (zfs_prop_get(zhp, entry->pl_prop, buf, sizeof (buf), NULL, NULL, 0, literal) == 0) { if (strlen(buf) > entry->pl_width) entry->pl_width = strlen(buf); } if (received && zfs_prop_get_recvd(zhp, zfs_prop_to_name(entry->pl_prop), buf, sizeof (buf), literal) == 0) if (strlen(buf) > entry->pl_recvd_width) entry->pl_recvd_width = strlen(buf); } else { if (nvlist_lookup_nvlist(userprops, entry->pl_user_prop, &propval) == 0) { strval = fnvlist_lookup_string(propval, ZPROP_VALUE); if (strlen(strval) > entry->pl_width) entry->pl_width = strlen(strval); } if (received && zfs_prop_get_recvd(zhp, entry->pl_user_prop, buf, sizeof (buf), literal) == 0) if (strlen(buf) > entry->pl_recvd_width) entry->pl_recvd_width = strlen(buf); } } return (0); } void zfs_prune_proplist(zfs_handle_t *zhp, uint8_t *props) { nvpair_t *curr; nvpair_t *next; /* * Keep a reference to the props-table against which we prune the * properties. */ zhp->zfs_props_table = props; curr = nvlist_next_nvpair(zhp->zfs_props, NULL); while (curr) { zfs_prop_t zfs_prop = zfs_name_to_prop(nvpair_name(curr)); next = nvlist_next_nvpair(zhp->zfs_props, curr); /* * User properties will result in ZPROP_USERPROP (an alias * for ZPROP_INVAL), and since we * only know how to prune standard ZFS properties, we always * leave these in the list. This can also happen if we * encounter an unknown DSL property (when running older * software, for example). */ if (zfs_prop != ZPROP_USERPROP && props[zfs_prop] == B_FALSE) (void) nvlist_remove(zhp->zfs_props, nvpair_name(curr), nvpair_type(curr)); curr = next; } } static int zfs_smb_acl_mgmt(libzfs_handle_t *hdl, char *dataset, char *path, zfs_smb_acl_op_t cmd, char *resource1, char *resource2) { zfs_cmd_t zc = {"\0"}; nvlist_t *nvlist = NULL; int error; (void) strlcpy(zc.zc_name, dataset, sizeof (zc.zc_name)); (void) strlcpy(zc.zc_value, path, sizeof (zc.zc_value)); zc.zc_cookie = (uint64_t)cmd; if (cmd == ZFS_SMB_ACL_RENAME) { if (nvlist_alloc(&nvlist, NV_UNIQUE_NAME, 0) != 0) { (void) no_memory(hdl); return (0); } } switch (cmd) { case ZFS_SMB_ACL_ADD: case ZFS_SMB_ACL_REMOVE: (void) strlcpy(zc.zc_string, resource1, sizeof (zc.zc_string)); break; case ZFS_SMB_ACL_RENAME: if (nvlist_add_string(nvlist, ZFS_SMB_ACL_SRC, resource1) != 0) { (void) no_memory(hdl); return (-1); } if (nvlist_add_string(nvlist, ZFS_SMB_ACL_TARGET, resource2) != 0) { (void) no_memory(hdl); return (-1); } zcmd_write_src_nvlist(hdl, &zc, nvlist); break; case ZFS_SMB_ACL_PURGE: break; default: return (-1); } error = ioctl(hdl->libzfs_fd, ZFS_IOC_SMB_ACL, &zc); nvlist_free(nvlist); return (error); } int zfs_smb_acl_add(libzfs_handle_t *hdl, char *dataset, char *path, char *resource) { return (zfs_smb_acl_mgmt(hdl, dataset, path, ZFS_SMB_ACL_ADD, resource, NULL)); } int zfs_smb_acl_remove(libzfs_handle_t *hdl, char *dataset, char *path, char *resource) { return (zfs_smb_acl_mgmt(hdl, dataset, path, ZFS_SMB_ACL_REMOVE, resource, NULL)); } int zfs_smb_acl_purge(libzfs_handle_t *hdl, char *dataset, char *path) { return (zfs_smb_acl_mgmt(hdl, dataset, path, ZFS_SMB_ACL_PURGE, NULL, NULL)); } int zfs_smb_acl_rename(libzfs_handle_t *hdl, char *dataset, char *path, char *oldname, char *newname) { return (zfs_smb_acl_mgmt(hdl, dataset, path, ZFS_SMB_ACL_RENAME, oldname, newname)); } int zfs_userspace(zfs_handle_t *zhp, zfs_userquota_prop_t type, zfs_userspace_cb_t func, void *arg) { zfs_cmd_t zc = {"\0"}; zfs_useracct_t buf[100]; libzfs_handle_t *hdl = zhp->zfs_hdl; int ret; (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); zc.zc_objset_type = type; zc.zc_nvlist_dst = (uintptr_t)buf; for (;;) { zfs_useracct_t *zua = buf; zc.zc_nvlist_dst_size = sizeof (buf); if (zfs_ioctl(hdl, ZFS_IOC_USERSPACE_MANY, &zc) != 0) { if ((errno == ENOTSUP && (type == ZFS_PROP_USEROBJUSED || type == ZFS_PROP_GROUPOBJUSED || type == ZFS_PROP_USEROBJQUOTA || type == ZFS_PROP_GROUPOBJQUOTA || type == ZFS_PROP_PROJECTOBJUSED || type == ZFS_PROP_PROJECTOBJQUOTA || type == ZFS_PROP_PROJECTUSED || type == ZFS_PROP_PROJECTQUOTA))) break; return (zfs_standard_error_fmt(hdl, errno, dgettext(TEXT_DOMAIN, "cannot get used/quota for %s"), zc.zc_name)); } if (zc.zc_nvlist_dst_size == 0) break; while (zc.zc_nvlist_dst_size > 0) { if ((ret = func(arg, zua->zu_domain, zua->zu_rid, zua->zu_space)) != 0) return (ret); zua++; zc.zc_nvlist_dst_size -= sizeof (zfs_useracct_t); } } return (0); } struct holdarg { nvlist_t *nvl; const char *snapname; const char *tag; boolean_t recursive; int error; }; static int zfs_hold_one(zfs_handle_t *zhp, void *arg) { struct holdarg *ha = arg; char name[ZFS_MAX_DATASET_NAME_LEN]; int rv = 0; if (snprintf(name, sizeof (name), "%s@%s", zhp->zfs_name, ha->snapname) >= sizeof (name)) return (EINVAL); if (lzc_exists(name)) fnvlist_add_string(ha->nvl, name, ha->tag); if (ha->recursive) rv = zfs_iter_filesystems_v2(zhp, 0, zfs_hold_one, ha); zfs_close(zhp); return (rv); } int zfs_hold(zfs_handle_t *zhp, const char *snapname, const char *tag, boolean_t recursive, int cleanup_fd) { int ret; struct holdarg ha; ha.nvl = fnvlist_alloc(); ha.snapname = snapname; ha.tag = tag; ha.recursive = recursive; (void) zfs_hold_one(zfs_handle_dup(zhp), &ha); if (nvlist_empty(ha.nvl)) { char errbuf[ERRBUFLEN]; fnvlist_free(ha.nvl); ret = ENOENT; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot hold snapshot '%s@%s'"), zhp->zfs_name, snapname); (void) zfs_standard_error(zhp->zfs_hdl, ret, errbuf); return (ret); } ret = zfs_hold_nvl(zhp, cleanup_fd, ha.nvl); fnvlist_free(ha.nvl); return (ret); } int zfs_hold_nvl(zfs_handle_t *zhp, int cleanup_fd, nvlist_t *holds) { int ret; nvlist_t *errors; libzfs_handle_t *hdl = zhp->zfs_hdl; char errbuf[ERRBUFLEN]; nvpair_t *elem; errors = NULL; ret = lzc_hold(holds, cleanup_fd, &errors); if (ret == 0) { /* There may be errors even in the success case. */ fnvlist_free(errors); return (0); } if (nvlist_empty(errors)) { /* no hold-specific errors */ (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot hold")); switch (ret) { case ENOTSUP: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool must be upgraded")); (void) zfs_error(hdl, EZFS_BADVERSION, errbuf); break; case EINVAL: (void) zfs_error(hdl, EZFS_BADTYPE, errbuf); break; default: (void) zfs_standard_error(hdl, ret, errbuf); } } for (elem = nvlist_next_nvpair(errors, NULL); elem != NULL; elem = nvlist_next_nvpair(errors, elem)) { (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot hold snapshot '%s'"), nvpair_name(elem)); switch (fnvpair_value_int32(elem)) { case E2BIG: /* * Temporary tags wind up having the ds object id * prepended. So even if we passed the length check * above, it's still possible for the tag to wind * up being slightly too long. */ (void) zfs_error(hdl, EZFS_TAGTOOLONG, errbuf); break; case EINVAL: (void) zfs_error(hdl, EZFS_BADTYPE, errbuf); break; case EEXIST: (void) zfs_error(hdl, EZFS_REFTAG_HOLD, errbuf); break; default: (void) zfs_standard_error(hdl, fnvpair_value_int32(elem), errbuf); } } fnvlist_free(errors); return (ret); } static int zfs_release_one(zfs_handle_t *zhp, void *arg) { struct holdarg *ha = arg; char name[ZFS_MAX_DATASET_NAME_LEN]; int rv = 0; nvlist_t *existing_holds; if (snprintf(name, sizeof (name), "%s@%s", zhp->zfs_name, ha->snapname) >= sizeof (name)) { ha->error = EINVAL; rv = EINVAL; } if (lzc_get_holds(name, &existing_holds) != 0) { ha->error = ENOENT; } else if (!nvlist_exists(existing_holds, ha->tag)) { ha->error = ESRCH; } else { nvlist_t *torelease = fnvlist_alloc(); fnvlist_add_boolean(torelease, ha->tag); fnvlist_add_nvlist(ha->nvl, name, torelease); fnvlist_free(torelease); } if (ha->recursive) rv = zfs_iter_filesystems_v2(zhp, 0, zfs_release_one, ha); zfs_close(zhp); return (rv); } int zfs_release(zfs_handle_t *zhp, const char *snapname, const char *tag, boolean_t recursive) { int ret; struct holdarg ha; nvlist_t *errors = NULL; nvpair_t *elem; libzfs_handle_t *hdl = zhp->zfs_hdl; char errbuf[ERRBUFLEN]; ha.nvl = fnvlist_alloc(); ha.snapname = snapname; ha.tag = tag; ha.recursive = recursive; ha.error = 0; (void) zfs_release_one(zfs_handle_dup(zhp), &ha); if (nvlist_empty(ha.nvl)) { fnvlist_free(ha.nvl); ret = ha.error; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot release hold from snapshot '%s@%s'"), zhp->zfs_name, snapname); if (ret == ESRCH) { (void) zfs_error(hdl, EZFS_REFTAG_RELE, errbuf); } else { (void) zfs_standard_error(hdl, ret, errbuf); } return (ret); } ret = lzc_release(ha.nvl, &errors); fnvlist_free(ha.nvl); if (ret == 0) { /* There may be errors even in the success case. */ fnvlist_free(errors); return (0); } if (nvlist_empty(errors)) { /* no hold-specific errors */ (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot release")); switch (errno) { case ENOTSUP: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool must be upgraded")); (void) zfs_error(hdl, EZFS_BADVERSION, errbuf); break; default: (void) zfs_standard_error(hdl, errno, errbuf); } } for (elem = nvlist_next_nvpair(errors, NULL); elem != NULL; elem = nvlist_next_nvpair(errors, elem)) { (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot release hold from snapshot '%s'"), nvpair_name(elem)); switch (fnvpair_value_int32(elem)) { case ESRCH: (void) zfs_error(hdl, EZFS_REFTAG_RELE, errbuf); break; case EINVAL: (void) zfs_error(hdl, EZFS_BADTYPE, errbuf); break; default: (void) zfs_standard_error(hdl, fnvpair_value_int32(elem), errbuf); } } fnvlist_free(errors); return (ret); } int zfs_get_fsacl(zfs_handle_t *zhp, nvlist_t **nvl) { zfs_cmd_t zc = {"\0"}; libzfs_handle_t *hdl = zhp->zfs_hdl; int nvsz = 2048; void *nvbuf; int err = 0; char errbuf[ERRBUFLEN]; assert(zhp->zfs_type == ZFS_TYPE_VOLUME || zhp->zfs_type == ZFS_TYPE_FILESYSTEM); tryagain: nvbuf = malloc(nvsz); if (nvbuf == NULL) { err = (zfs_error(hdl, EZFS_NOMEM, strerror(errno))); goto out; } zc.zc_nvlist_dst_size = nvsz; zc.zc_nvlist_dst = (uintptr_t)nvbuf; (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); if (zfs_ioctl(hdl, ZFS_IOC_GET_FSACL, &zc) != 0) { (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot get permissions on '%s'"), zc.zc_name); switch (errno) { case ENOMEM: free(nvbuf); nvsz = zc.zc_nvlist_dst_size; goto tryagain; case ENOTSUP: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool must be upgraded")); err = zfs_error(hdl, EZFS_BADVERSION, errbuf); break; case EINVAL: err = zfs_error(hdl, EZFS_BADTYPE, errbuf); break; case ENOENT: err = zfs_error(hdl, EZFS_NOENT, errbuf); break; default: err = zfs_standard_error(hdl, errno, errbuf); break; } } else { /* success */ int rc = nvlist_unpack(nvbuf, zc.zc_nvlist_dst_size, nvl, 0); if (rc) { err = zfs_standard_error_fmt(hdl, rc, dgettext( TEXT_DOMAIN, "cannot get permissions on '%s'"), zc.zc_name); } } free(nvbuf); out: return (err); } int zfs_set_fsacl(zfs_handle_t *zhp, boolean_t un, nvlist_t *nvl) { zfs_cmd_t zc = {"\0"}; libzfs_handle_t *hdl = zhp->zfs_hdl; char *nvbuf; char errbuf[ERRBUFLEN]; size_t nvsz; int err; assert(zhp->zfs_type == ZFS_TYPE_VOLUME || zhp->zfs_type == ZFS_TYPE_FILESYSTEM); err = nvlist_size(nvl, &nvsz, NV_ENCODE_NATIVE); assert(err == 0); nvbuf = malloc(nvsz); err = nvlist_pack(nvl, &nvbuf, &nvsz, NV_ENCODE_NATIVE, 0); assert(err == 0); zc.zc_nvlist_src_size = nvsz; zc.zc_nvlist_src = (uintptr_t)nvbuf; zc.zc_perm_action = un; (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); if (zfs_ioctl(hdl, ZFS_IOC_SET_FSACL, &zc) != 0) { (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot set permissions on '%s'"), zc.zc_name); switch (errno) { case ENOTSUP: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool must be upgraded")); err = zfs_error(hdl, EZFS_BADVERSION, errbuf); break; case EINVAL: err = zfs_error(hdl, EZFS_BADTYPE, errbuf); break; case ENOENT: err = zfs_error(hdl, EZFS_NOENT, errbuf); break; default: err = zfs_standard_error(hdl, errno, errbuf); break; } } free(nvbuf); return (err); } int zfs_get_holds(zfs_handle_t *zhp, nvlist_t **nvl) { int err; char errbuf[ERRBUFLEN]; err = lzc_get_holds(zhp->zfs_name, nvl); if (err != 0) { libzfs_handle_t *hdl = zhp->zfs_hdl; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot get holds for '%s'"), zhp->zfs_name); switch (err) { case ENOTSUP: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool must be upgraded")); err = zfs_error(hdl, EZFS_BADVERSION, errbuf); break; case EINVAL: err = zfs_error(hdl, EZFS_BADTYPE, errbuf); break; case ENOENT: err = zfs_error(hdl, EZFS_NOENT, errbuf); break; default: err = zfs_standard_error(hdl, errno, errbuf); break; } } return (err); } /* * The theory of raidz space accounting * * The "referenced" property of RAIDZ vdevs is scaled such that a 128KB block * will "reference" 128KB, even though it allocates more than that, to store the * parity information (and perhaps skip sectors). This concept of the * "referenced" (and other DMU space accounting) being lower than the allocated * space by a constant factor is called "raidz deflation." * * As mentioned above, the constant factor for raidz deflation assumes a 128KB * block size. However, zvols typically have a much smaller block size (default * 8KB). These smaller blocks may require proportionally much more parity * information (and perhaps skip sectors). In this case, the change to the * "referenced" property may be much more than the logical block size. * * Suppose a raidz vdev has 5 disks with ashift=12. A 128k block may be written * as follows. * * +-------+-------+-------+-------+-------+ * | disk1 | disk2 | disk3 | disk4 | disk5 | * +-------+-------+-------+-------+-------+ * | P0 | D0 | D8 | D16 | D24 | * | P1 | D1 | D9 | D17 | D25 | * | P2 | D2 | D10 | D18 | D26 | * | P3 | D3 | D11 | D19 | D27 | * | P4 | D4 | D12 | D20 | D28 | * | P5 | D5 | D13 | D21 | D29 | * | P6 | D6 | D14 | D22 | D30 | * | P7 | D7 | D15 | D23 | D31 | * +-------+-------+-------+-------+-------+ * * Above, notice that 160k was allocated: 8 x 4k parity sectors + 32 x 4k data * sectors. The dataset's referenced will increase by 128k and the pool's * allocated and free properties will be adjusted by 160k. * * A 4k block written to the same raidz vdev will require two 4k sectors. The * blank cells represent unallocated space. * * +-------+-------+-------+-------+-------+ * | disk1 | disk2 | disk3 | disk4 | disk5 | * +-------+-------+-------+-------+-------+ * | P0 | D0 | | | | * +-------+-------+-------+-------+-------+ * * Above, notice that the 4k block required one sector for parity and another * for data. vdev_raidz_asize() will return 8k and as such the pool's allocated * and free properties will be adjusted by 8k. The dataset will not be charged * 8k. Rather, it will be charged a value that is scaled according to the * overhead of the 128k block on the same vdev. This 8k allocation will be * charged 8k * 128k / 160k. 128k is from SPA_OLD_MAXBLOCKSIZE and 160k is as * calculated in the 128k block example above. * * Every raidz allocation is sized to be a multiple of nparity+1 sectors. That * is, every raidz1 allocation will be a multiple of 2 sectors, raidz2 * allocations are a multiple of 3 sectors, and raidz3 allocations are a * multiple of of 4 sectors. When a block does not fill the required number of * sectors, skip blocks (sectors) are used. * * An 8k block being written to a raidz vdev may be written as follows: * * +-------+-------+-------+-------+-------+ * | disk1 | disk2 | disk3 | disk4 | disk5 | * +-------+-------+-------+-------+-------+ * | P0 | D0 | D1 | S0 | | * +-------+-------+-------+-------+-------+ * * In order to maintain the nparity+1 allocation size, a skip block (S0) was * added. For this 8k block, the pool's allocated and free properties are * adjusted by 16k and the dataset's referenced is increased by 16k * 128k / * 160k. Again, 128k is from SPA_OLD_MAXBLOCKSIZE and 160k is as calculated in * the 128k block example above. * * The situation is slightly different for dRAID since the minimum allocation * size is the full group width. The same 8K block above would be written as * follows in a dRAID group: * * +-------+-------+-------+-------+-------+ * | disk1 | disk2 | disk3 | disk4 | disk5 | * +-------+-------+-------+-------+-------+ * | P0 | D0 | D1 | S0 | S1 | * +-------+-------+-------+-------+-------+ * * Compression may lead to a variety of block sizes being written for the same * volume or file. There is no clear way to reserve just the amount of space * that will be required, so the worst case (no compression) is assumed. * Note that metadata blocks will typically be compressed, so the reservation * size returned by zvol_volsize_to_reservation() will generally be slightly * larger than the maximum that the volume can reference. */ /* * Derived from function of same name in module/zfs/vdev_raidz.c. Returns the * amount of space (in bytes) that will be allocated for the specified block * size. Note that the "referenced" space accounted will be less than this, but * not necessarily equal to "blksize", due to RAIDZ deflation. */ static uint64_t vdev_raidz_asize(uint64_t ndisks, uint64_t nparity, uint64_t ashift, uint64_t blksize) { uint64_t asize, ndata; ASSERT3U(ndisks, >, nparity); ndata = ndisks - nparity; asize = ((blksize - 1) >> ashift) + 1; asize += nparity * ((asize + ndata - 1) / ndata); asize = roundup(asize, nparity + 1) << ashift; return (asize); } /* * Derived from function of same name in module/zfs/vdev_draid.c. Returns the * amount of space (in bytes) that will be allocated for the specified block * size. */ static uint64_t vdev_draid_asize(uint64_t ndisks, uint64_t nparity, uint64_t ashift, uint64_t blksize) { ASSERT3U(ndisks, >, nparity); uint64_t ndata = ndisks - nparity; uint64_t rows = ((blksize - 1) / (ndata << ashift)) + 1; uint64_t asize = (rows * ndisks) << ashift; return (asize); } /* * Determine how much space will be allocated if it lands on the most space- * inefficient top-level vdev. Returns the size in bytes required to store one * copy of the volume data. See theory comment above. */ static uint64_t volsize_from_vdevs(zpool_handle_t *zhp, uint64_t nblocks, uint64_t blksize) { nvlist_t *config, *tree, **vdevs; uint_t nvdevs; uint64_t ret = 0; config = zpool_get_config(zhp, NULL); if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &tree) != 0 || nvlist_lookup_nvlist_array(tree, ZPOOL_CONFIG_CHILDREN, &vdevs, &nvdevs) != 0) { return (nblocks * blksize); } for (int v = 0; v < nvdevs; v++) { const char *type; uint64_t nparity, ashift, asize, tsize; uint64_t volsize; if (nvlist_lookup_string(vdevs[v], ZPOOL_CONFIG_TYPE, &type) != 0) continue; if (strcmp(type, VDEV_TYPE_RAIDZ) != 0 && strcmp(type, VDEV_TYPE_DRAID) != 0) continue; if (nvlist_lookup_uint64(vdevs[v], ZPOOL_CONFIG_NPARITY, &nparity) != 0) continue; if (nvlist_lookup_uint64(vdevs[v], ZPOOL_CONFIG_ASHIFT, &ashift) != 0) continue; if (strcmp(type, VDEV_TYPE_RAIDZ) == 0) { nvlist_t **disks; uint_t ndisks; if (nvlist_lookup_nvlist_array(vdevs[v], ZPOOL_CONFIG_CHILDREN, &disks, &ndisks) != 0) continue; /* allocation size for the "typical" 128k block */ tsize = vdev_raidz_asize(ndisks, nparity, ashift, SPA_OLD_MAXBLOCKSIZE); /* allocation size for the blksize block */ asize = vdev_raidz_asize(ndisks, nparity, ashift, blksize); } else { uint64_t ndata; if (nvlist_lookup_uint64(vdevs[v], ZPOOL_CONFIG_DRAID_NDATA, &ndata) != 0) continue; /* allocation size for the "typical" 128k block */ tsize = vdev_draid_asize(ndata + nparity, nparity, ashift, SPA_OLD_MAXBLOCKSIZE); /* allocation size for the blksize block */ asize = vdev_draid_asize(ndata + nparity, nparity, ashift, blksize); } /* * Scale this size down as a ratio of 128k / tsize. * See theory statement above. */ volsize = nblocks * asize * SPA_OLD_MAXBLOCKSIZE / tsize; if (volsize > ret) { ret = volsize; } } if (ret == 0) { ret = nblocks * blksize; } return (ret); } /* * Convert the zvol's volume size to an appropriate reservation. See theory * comment above. * * Note: If this routine is updated, it is necessary to update the ZFS test * suite's shell version in reservation.shlib. */ uint64_t zvol_volsize_to_reservation(zpool_handle_t *zph, uint64_t volsize, nvlist_t *props) { uint64_t numdb; uint64_t nblocks, volblocksize; int ncopies; const char *strval; if (nvlist_lookup_string(props, zfs_prop_to_name(ZFS_PROP_COPIES), &strval) == 0) ncopies = atoi(strval); else ncopies = 1; if (nvlist_lookup_uint64(props, zfs_prop_to_name(ZFS_PROP_VOLBLOCKSIZE), &volblocksize) != 0) volblocksize = ZVOL_DEFAULT_BLOCKSIZE; nblocks = volsize / volblocksize; /* * Metadata defaults to using 128k blocks, not volblocksize blocks. For * this reason, only the data blocks are scaled based on vdev config. */ volsize = volsize_from_vdevs(zph, nblocks, volblocksize); /* start with metadnode L0-L6 */ numdb = 7; /* calculate number of indirects */ while (nblocks > 1) { nblocks += DNODES_PER_LEVEL - 1; nblocks /= DNODES_PER_LEVEL; numdb += nblocks; } numdb *= MIN(SPA_DVAS_PER_BP, ncopies + 1); volsize *= ncopies; /* * this is exactly DN_MAX_INDBLKSHIFT when metadata isn't * compressed, but in practice they compress down to about * 1100 bytes */ numdb *= 1ULL << DN_MAX_INDBLKSHIFT; volsize += numdb; return (volsize); } /* * Wait for the given activity and return the status of the wait (whether or not * any waiting was done) in the 'waited' parameter. Non-existent fses are * reported via the 'missing' parameter, rather than by printing an error * message. This is convenient when this function is called in a loop over a * long period of time (as it is, for example, by zfs's wait cmd). In that * scenario, a fs being exported or destroyed should be considered a normal * event, so we don't want to print an error when we find that the fs doesn't * exist. */ int zfs_wait_status(zfs_handle_t *zhp, zfs_wait_activity_t activity, boolean_t *missing, boolean_t *waited) { int error = lzc_wait_fs(zhp->zfs_name, activity, waited); *missing = (error == ENOENT); if (*missing) return (0); if (error != 0) { (void) zfs_standard_error_fmt(zhp->zfs_hdl, error, dgettext(TEXT_DOMAIN, "error waiting in fs '%s'"), zhp->zfs_name); } return (error); } diff --git a/sys/contrib/openzfs/lib/libzfs/libzfs_pool.c b/sys/contrib/openzfs/lib/libzfs/libzfs_pool.c index 4fb71b4e0dc8..d4af31c50cf8 100644 --- a/sys/contrib/openzfs/lib/libzfs/libzfs_pool.c +++ b/sys/contrib/openzfs/lib/libzfs/libzfs_pool.c @@ -1,5321 +1,5391 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright 2015 Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2011, 2020 by Delphix. All rights reserved. * Copyright 2016 Igor Kozhukhov * Copyright (c) 2018 Datto Inc. * Copyright (c) 2017 Open-E, Inc. All Rights Reserved. * Copyright (c) 2017, Intel Corporation. * Copyright (c) 2018, loli10K * Copyright (c) 2021, Colm Buckley * Copyright (c) 2021, Klara Inc. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "zfs_namecheck.h" #include "zfs_prop.h" #include "libzfs_impl.h" #include "zfs_comutil.h" #include "zfeature_common.h" static boolean_t zpool_vdev_is_interior(const char *name); typedef struct prop_flags { unsigned int create:1; /* Validate property on creation */ unsigned int import:1; /* Validate property on import */ unsigned int vdevprop:1; /* Validate property as a VDEV property */ } prop_flags_t; /* * ==================================================================== * zpool property functions * ==================================================================== */ static int zpool_get_all_props(zpool_handle_t *zhp) { zfs_cmd_t zc = {"\0"}; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); zcmd_alloc_dst_nvlist(hdl, &zc, 0); while (zfs_ioctl(hdl, ZFS_IOC_POOL_GET_PROPS, &zc) != 0) { if (errno == ENOMEM) zcmd_expand_dst_nvlist(hdl, &zc); else { zcmd_free_nvlists(&zc); return (-1); } } if (zcmd_read_dst_nvlist(hdl, &zc, &zhp->zpool_props) != 0) { zcmd_free_nvlists(&zc); return (-1); } zcmd_free_nvlists(&zc); return (0); } int zpool_props_refresh(zpool_handle_t *zhp) { nvlist_t *old_props; old_props = zhp->zpool_props; if (zpool_get_all_props(zhp) != 0) return (-1); nvlist_free(old_props); return (0); } static const char * zpool_get_prop_string(zpool_handle_t *zhp, zpool_prop_t prop, zprop_source_t *src) { nvlist_t *nv, *nvl; const char *value; zprop_source_t source; nvl = zhp->zpool_props; if (nvlist_lookup_nvlist(nvl, zpool_prop_to_name(prop), &nv) == 0) { source = fnvlist_lookup_uint64(nv, ZPROP_SOURCE); value = fnvlist_lookup_string(nv, ZPROP_VALUE); } else { source = ZPROP_SRC_DEFAULT; if ((value = zpool_prop_default_string(prop)) == NULL) value = "-"; } if (src) *src = source; return (value); } uint64_t zpool_get_prop_int(zpool_handle_t *zhp, zpool_prop_t prop, zprop_source_t *src) { nvlist_t *nv, *nvl; uint64_t value; zprop_source_t source; if (zhp->zpool_props == NULL && zpool_get_all_props(zhp)) { /* * zpool_get_all_props() has most likely failed because * the pool is faulted, but if all we need is the top level * vdev's guid then get it from the zhp config nvlist. */ if ((prop == ZPOOL_PROP_GUID) && (nvlist_lookup_nvlist(zhp->zpool_config, ZPOOL_CONFIG_VDEV_TREE, &nv) == 0) && (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &value) == 0)) { return (value); } return (zpool_prop_default_numeric(prop)); } nvl = zhp->zpool_props; if (nvlist_lookup_nvlist(nvl, zpool_prop_to_name(prop), &nv) == 0) { source = fnvlist_lookup_uint64(nv, ZPROP_SOURCE); value = fnvlist_lookup_uint64(nv, ZPROP_VALUE); } else { source = ZPROP_SRC_DEFAULT; value = zpool_prop_default_numeric(prop); } if (src) *src = source; return (value); } /* * Map VDEV STATE to printed strings. */ const char * zpool_state_to_name(vdev_state_t state, vdev_aux_t aux) { switch (state) { case VDEV_STATE_CLOSED: case VDEV_STATE_OFFLINE: return (gettext("OFFLINE")); case VDEV_STATE_REMOVED: return (gettext("REMOVED")); case VDEV_STATE_CANT_OPEN: if (aux == VDEV_AUX_CORRUPT_DATA || aux == VDEV_AUX_BAD_LOG) return (gettext("FAULTED")); else if (aux == VDEV_AUX_SPLIT_POOL) return (gettext("SPLIT")); else return (gettext("UNAVAIL")); case VDEV_STATE_FAULTED: return (gettext("FAULTED")); case VDEV_STATE_DEGRADED: return (gettext("DEGRADED")); case VDEV_STATE_HEALTHY: return (gettext("ONLINE")); default: break; } return (gettext("UNKNOWN")); } /* * Map POOL STATE to printed strings. */ const char * zpool_pool_state_to_name(pool_state_t state) { switch (state) { default: break; case POOL_STATE_ACTIVE: return (gettext("ACTIVE")); case POOL_STATE_EXPORTED: return (gettext("EXPORTED")); case POOL_STATE_DESTROYED: return (gettext("DESTROYED")); case POOL_STATE_SPARE: return (gettext("SPARE")); case POOL_STATE_L2CACHE: return (gettext("L2CACHE")); case POOL_STATE_UNINITIALIZED: return (gettext("UNINITIALIZED")); case POOL_STATE_UNAVAIL: return (gettext("UNAVAIL")); case POOL_STATE_POTENTIALLY_ACTIVE: return (gettext("POTENTIALLY_ACTIVE")); } return (gettext("UNKNOWN")); } /* * Given a pool handle, return the pool health string ("ONLINE", "DEGRADED", * "SUSPENDED", etc). */ const char * zpool_get_state_str(zpool_handle_t *zhp) { zpool_errata_t errata; zpool_status_t status; const char *str; status = zpool_get_status(zhp, NULL, &errata); if (zpool_get_state(zhp) == POOL_STATE_UNAVAIL) { str = gettext("FAULTED"); } else if (status == ZPOOL_STATUS_IO_FAILURE_WAIT || status == ZPOOL_STATUS_IO_FAILURE_MMP) { str = gettext("SUSPENDED"); } else { nvlist_t *nvroot = fnvlist_lookup_nvlist( zpool_get_config(zhp, NULL), ZPOOL_CONFIG_VDEV_TREE); uint_t vsc; vdev_stat_t *vs = (vdev_stat_t *)fnvlist_lookup_uint64_array( nvroot, ZPOOL_CONFIG_VDEV_STATS, &vsc); str = zpool_state_to_name(vs->vs_state, vs->vs_aux); } return (str); } /* * Get a zpool property value for 'prop' and return the value in * a pre-allocated buffer. */ int zpool_get_prop(zpool_handle_t *zhp, zpool_prop_t prop, char *buf, size_t len, zprop_source_t *srctype, boolean_t literal) { uint64_t intval; const char *strval; zprop_source_t src = ZPROP_SRC_NONE; if (zpool_get_state(zhp) == POOL_STATE_UNAVAIL) { switch (prop) { case ZPOOL_PROP_NAME: (void) strlcpy(buf, zpool_get_name(zhp), len); break; case ZPOOL_PROP_HEALTH: (void) strlcpy(buf, zpool_get_state_str(zhp), len); break; case ZPOOL_PROP_GUID: intval = zpool_get_prop_int(zhp, prop, &src); (void) snprintf(buf, len, "%llu", (u_longlong_t)intval); break; case ZPOOL_PROP_ALTROOT: case ZPOOL_PROP_CACHEFILE: case ZPOOL_PROP_COMMENT: case ZPOOL_PROP_COMPATIBILITY: if (zhp->zpool_props != NULL || zpool_get_all_props(zhp) == 0) { (void) strlcpy(buf, zpool_get_prop_string(zhp, prop, &src), len); break; } zfs_fallthrough; default: (void) strlcpy(buf, "-", len); break; } if (srctype != NULL) *srctype = src; return (0); } if (zhp->zpool_props == NULL && zpool_get_all_props(zhp) && prop != ZPOOL_PROP_NAME) return (-1); switch (zpool_prop_get_type(prop)) { case PROP_TYPE_STRING: (void) strlcpy(buf, zpool_get_prop_string(zhp, prop, &src), len); break; case PROP_TYPE_NUMBER: intval = zpool_get_prop_int(zhp, prop, &src); switch (prop) { case ZPOOL_PROP_SIZE: case ZPOOL_PROP_ALLOCATED: case ZPOOL_PROP_FREE: case ZPOOL_PROP_FREEING: case ZPOOL_PROP_LEAKED: case ZPOOL_PROP_ASHIFT: case ZPOOL_PROP_MAXBLOCKSIZE: case ZPOOL_PROP_MAXDNODESIZE: case ZPOOL_PROP_BCLONESAVED: case ZPOOL_PROP_BCLONEUSED: if (literal) (void) snprintf(buf, len, "%llu", (u_longlong_t)intval); else (void) zfs_nicenum(intval, buf, len); break; case ZPOOL_PROP_EXPANDSZ: case ZPOOL_PROP_CHECKPOINT: if (intval == 0) { (void) strlcpy(buf, "-", len); } else if (literal) { (void) snprintf(buf, len, "%llu", (u_longlong_t)intval); } else { (void) zfs_nicebytes(intval, buf, len); } break; case ZPOOL_PROP_CAPACITY: if (literal) { (void) snprintf(buf, len, "%llu", (u_longlong_t)intval); } else { (void) snprintf(buf, len, "%llu%%", (u_longlong_t)intval); } break; case ZPOOL_PROP_FRAGMENTATION: if (intval == UINT64_MAX) { (void) strlcpy(buf, "-", len); } else if (literal) { (void) snprintf(buf, len, "%llu", (u_longlong_t)intval); } else { (void) snprintf(buf, len, "%llu%%", (u_longlong_t)intval); } break; case ZPOOL_PROP_BCLONERATIO: case ZPOOL_PROP_DEDUPRATIO: if (literal) (void) snprintf(buf, len, "%llu.%02llu", (u_longlong_t)(intval / 100), (u_longlong_t)(intval % 100)); else (void) snprintf(buf, len, "%llu.%02llux", (u_longlong_t)(intval / 100), (u_longlong_t)(intval % 100)); break; case ZPOOL_PROP_HEALTH: (void) strlcpy(buf, zpool_get_state_str(zhp), len); break; case ZPOOL_PROP_VERSION: if (intval >= SPA_VERSION_FEATURES) { (void) snprintf(buf, len, "-"); break; } zfs_fallthrough; default: (void) snprintf(buf, len, "%llu", (u_longlong_t)intval); } break; case PROP_TYPE_INDEX: intval = zpool_get_prop_int(zhp, prop, &src); if (zpool_prop_index_to_string(prop, intval, &strval) != 0) return (-1); (void) strlcpy(buf, strval, len); break; default: abort(); } if (srctype) *srctype = src; return (0); } /* * Get a zpool property value for 'propname' and return the value in * a pre-allocated buffer. */ int zpool_get_userprop(zpool_handle_t *zhp, const char *propname, char *buf, size_t len, zprop_source_t *srctype) { nvlist_t *nv, *nvl; uint64_t ival; const char *value; zprop_source_t source = ZPROP_SRC_LOCAL; nvl = zhp->zpool_props; if (nvlist_lookup_nvlist(nvl, propname, &nv) == 0) { if (nvlist_lookup_uint64(nv, ZPROP_SOURCE, &ival) == 0) source = ival; verify(nvlist_lookup_string(nv, ZPROP_VALUE, &value) == 0); } else { source = ZPROP_SRC_DEFAULT; value = "-"; } if (srctype) *srctype = source; (void) strlcpy(buf, value, len); return (0); } /* * Check if the bootfs name has the same pool name as it is set to. * Assuming bootfs is a valid dataset name. */ static boolean_t bootfs_name_valid(const char *pool, const char *bootfs) { int len = strlen(pool); if (bootfs[0] == '\0') return (B_TRUE); if (!zfs_name_valid(bootfs, ZFS_TYPE_FILESYSTEM|ZFS_TYPE_SNAPSHOT)) return (B_FALSE); if (strncmp(pool, bootfs, len) == 0 && (bootfs[len] == '/' || bootfs[len] == '\0')) return (B_TRUE); return (B_FALSE); } /* * Given an nvlist of zpool properties to be set, validate that they are * correct, and parse any numeric properties (index, boolean, etc) if they are * specified as strings. */ static nvlist_t * zpool_valid_proplist(libzfs_handle_t *hdl, const char *poolname, nvlist_t *props, uint64_t version, prop_flags_t flags, char *errbuf) { nvpair_t *elem; nvlist_t *retprops; zpool_prop_t prop; const char *strval; uint64_t intval; const char *slash, *check; struct stat64 statbuf; zpool_handle_t *zhp; char report[1024]; if (nvlist_alloc(&retprops, NV_UNIQUE_NAME, 0) != 0) { (void) no_memory(hdl); return (NULL); } elem = NULL; while ((elem = nvlist_next_nvpair(props, elem)) != NULL) { const char *propname = nvpair_name(elem); if (flags.vdevprop && zpool_prop_vdev(propname)) { vdev_prop_t vprop = vdev_name_to_prop(propname); if (vdev_prop_readonly(vprop)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' " "is readonly"), propname); (void) zfs_error(hdl, EZFS_PROPREADONLY, errbuf); goto error; } if (zprop_parse_value(hdl, elem, vprop, ZFS_TYPE_VDEV, retprops, &strval, &intval, errbuf) != 0) goto error; continue; } else if (flags.vdevprop && vdev_prop_user(propname)) { if (nvlist_add_nvpair(retprops, elem) != 0) { (void) no_memory(hdl); goto error; } continue; } else if (flags.vdevprop) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid property: '%s'"), propname); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } prop = zpool_name_to_prop(propname); if (prop == ZPOOL_PROP_INVAL && zpool_prop_feature(propname)) { int err; char *fname = strchr(propname, '@') + 1; err = zfeature_lookup_name(fname, NULL); if (err != 0) { ASSERT3U(err, ==, ENOENT); zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "feature '%s' unsupported by kernel"), fname); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } if (nvpair_type(elem) != DATA_TYPE_STRING) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' must be a string"), propname); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } (void) nvpair_value_string(elem, &strval); if (strcmp(strval, ZFS_FEATURE_ENABLED) != 0 && strcmp(strval, ZFS_FEATURE_DISABLED) != 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "property '%s' can only be set to " "'enabled' or 'disabled'"), propname); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } if (!flags.create && strcmp(strval, ZFS_FEATURE_DISABLED) == 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "property '%s' can only be set to " "'disabled' at creation time"), propname); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } if (nvlist_add_uint64(retprops, propname, 0) != 0) { (void) no_memory(hdl); goto error; } continue; } else if (prop == ZPOOL_PROP_INVAL && zfs_prop_user(propname)) { /* * This is a user property: make sure it's a * string, and that it's less than ZAP_MAXNAMELEN. */ if (nvpair_type(elem) != DATA_TYPE_STRING) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' must be a string"), propname); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } if (strlen(nvpair_name(elem)) >= ZAP_MAXNAMELEN) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "property name '%s' is too long"), propname); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } (void) nvpair_value_string(elem, &strval); if (strlen(strval) >= ZFS_MAXPROPLEN) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "property value '%s' is too long"), strval); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } if (nvlist_add_string(retprops, propname, strval) != 0) { (void) no_memory(hdl); goto error; } continue; } /* * Make sure this property is valid and applies to this type. */ if (prop == ZPOOL_PROP_INVAL) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid property '%s'"), propname); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } if (zpool_prop_readonly(prop)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' " "is readonly"), propname); (void) zfs_error(hdl, EZFS_PROPREADONLY, errbuf); goto error; } if (!flags.create && zpool_prop_setonce(prop)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "property '%s' can only be set at " "creation time"), propname); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } if (zprop_parse_value(hdl, elem, prop, ZFS_TYPE_POOL, retprops, &strval, &intval, errbuf) != 0) goto error; /* * Perform additional checking for specific properties. */ switch (prop) { case ZPOOL_PROP_VERSION: if (intval < version || !SPA_VERSION_IS_SUPPORTED(intval)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "property '%s' number %llu is invalid."), propname, (unsigned long long)intval); (void) zfs_error(hdl, EZFS_BADVERSION, errbuf); goto error; } break; case ZPOOL_PROP_ASHIFT: if (intval != 0 && (intval < ASHIFT_MIN || intval > ASHIFT_MAX)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "property '%s' number %llu is invalid, " "only values between %" PRId32 " and %" PRId32 " are allowed."), propname, (unsigned long long)intval, ASHIFT_MIN, ASHIFT_MAX); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } break; case ZPOOL_PROP_BOOTFS: if (flags.create || flags.import) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "property '%s' cannot be set at creation " "or import time"), propname); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } if (version < SPA_VERSION_BOOTFS) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool must be upgraded to support " "'%s' property"), propname); (void) zfs_error(hdl, EZFS_BADVERSION, errbuf); goto error; } /* * bootfs property value has to be a dataset name and * the dataset has to be in the same pool as it sets to. */ if (!bootfs_name_valid(poolname, strval)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' " "is an invalid name"), strval); (void) zfs_error(hdl, EZFS_INVALIDNAME, errbuf); goto error; } if ((zhp = zpool_open_canfail(hdl, poolname)) == NULL) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "could not open pool '%s'"), poolname); (void) zfs_error(hdl, EZFS_OPENFAILED, errbuf); goto error; } zpool_close(zhp); break; case ZPOOL_PROP_ALTROOT: if (!flags.create && !flags.import) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "property '%s' can only be set during pool " "creation or import"), propname); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } if (strval[0] != '/') { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "bad alternate root '%s'"), strval); (void) zfs_error(hdl, EZFS_BADPATH, errbuf); goto error; } break; case ZPOOL_PROP_CACHEFILE: if (strval[0] == '\0') break; if (strcmp(strval, "none") == 0) break; if (strval[0] != '/') { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "property '%s' must be empty, an " "absolute path, or 'none'"), propname); (void) zfs_error(hdl, EZFS_BADPATH, errbuf); goto error; } slash = strrchr(strval, '/'); if (slash[1] == '\0' || strcmp(slash, "/.") == 0 || strcmp(slash, "/..") == 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' is not a valid file"), strval); (void) zfs_error(hdl, EZFS_BADPATH, errbuf); goto error; } *(char *)slash = '\0'; if (strval[0] != '\0' && (stat64(strval, &statbuf) != 0 || !S_ISDIR(statbuf.st_mode))) { *(char *)slash = '/'; zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' is not a valid directory"), strval); (void) zfs_error(hdl, EZFS_BADPATH, errbuf); goto error; } *(char *)slash = '/'; break; case ZPOOL_PROP_COMPATIBILITY: switch (zpool_load_compat(strval, NULL, report, 1024)) { case ZPOOL_COMPATIBILITY_OK: case ZPOOL_COMPATIBILITY_WARNTOKEN: break; case ZPOOL_COMPATIBILITY_BADFILE: case ZPOOL_COMPATIBILITY_BADTOKEN: case ZPOOL_COMPATIBILITY_NOFILES: zfs_error_aux(hdl, "%s", report); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } break; case ZPOOL_PROP_COMMENT: for (check = strval; *check != '\0'; check++) { if (!isprint(*check)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "comment may only have printable " "characters")); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } } if (strlen(strval) > ZPROP_MAX_COMMENT) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "comment must not exceed %d characters"), ZPROP_MAX_COMMENT); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } break; case ZPOOL_PROP_READONLY: if (!flags.import) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "property '%s' can only be set at " "import time"), propname); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } break; case ZPOOL_PROP_MULTIHOST: if (get_system_hostid() == 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "requires a non-zero system hostid")); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } break; case ZPOOL_PROP_DEDUPDITTO: printf("Note: property '%s' no longer has " "any effect\n", propname); break; default: break; } } return (retprops); error: nvlist_free(retprops); return (NULL); } /* * Set zpool property : propname=propval. */ int zpool_set_prop(zpool_handle_t *zhp, const char *propname, const char *propval) { zfs_cmd_t zc = {"\0"}; int ret = -1; char errbuf[ERRBUFLEN]; nvlist_t *nvl = NULL; nvlist_t *realprops; uint64_t version; prop_flags_t flags = { 0 }; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot set property for '%s'"), zhp->zpool_name); if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0) return (no_memory(zhp->zpool_hdl)); if (nvlist_add_string(nvl, propname, propval) != 0) { nvlist_free(nvl); return (no_memory(zhp->zpool_hdl)); } version = zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL); if ((realprops = zpool_valid_proplist(zhp->zpool_hdl, zhp->zpool_name, nvl, version, flags, errbuf)) == NULL) { nvlist_free(nvl); return (-1); } nvlist_free(nvl); nvl = realprops; /* * Execute the corresponding ioctl() to set this property. */ (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); zcmd_write_src_nvlist(zhp->zpool_hdl, &zc, nvl); ret = zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_POOL_SET_PROPS, &zc); zcmd_free_nvlists(&zc); nvlist_free(nvl); if (ret) (void) zpool_standard_error(zhp->zpool_hdl, errno, errbuf); else (void) zpool_props_refresh(zhp); return (ret); } int zpool_expand_proplist(zpool_handle_t *zhp, zprop_list_t **plp, zfs_type_t type, boolean_t literal) { libzfs_handle_t *hdl = zhp->zpool_hdl; zprop_list_t *entry; char buf[ZFS_MAXPROPLEN]; nvlist_t *features = NULL; nvpair_t *nvp; zprop_list_t **last; boolean_t firstexpand = (NULL == *plp); int i; if (zprop_expand_list(hdl, plp, type) != 0) return (-1); if (type == ZFS_TYPE_VDEV) return (0); last = plp; while (*last != NULL) last = &(*last)->pl_next; if ((*plp)->pl_all) features = zpool_get_features(zhp); if ((*plp)->pl_all && firstexpand) { /* Handle userprops in the all properties case */ if (zhp->zpool_props == NULL && zpool_props_refresh(zhp)) return (-1); nvp = NULL; while ((nvp = nvlist_next_nvpair(zhp->zpool_props, nvp)) != NULL) { const char *propname = nvpair_name(nvp); if (!zfs_prop_user(propname)) continue; entry = zfs_alloc(hdl, sizeof (zprop_list_t)); entry->pl_prop = ZPROP_USERPROP; entry->pl_user_prop = zfs_strdup(hdl, propname); entry->pl_width = strlen(entry->pl_user_prop); entry->pl_all = B_TRUE; *last = entry; last = &entry->pl_next; } for (i = 0; i < SPA_FEATURES; i++) { entry = zfs_alloc(hdl, sizeof (zprop_list_t)); entry->pl_prop = ZPROP_USERPROP; entry->pl_user_prop = zfs_asprintf(hdl, "feature@%s", spa_feature_table[i].fi_uname); entry->pl_width = strlen(entry->pl_user_prop); entry->pl_all = B_TRUE; *last = entry; last = &entry->pl_next; } } /* add any unsupported features */ for (nvp = nvlist_next_nvpair(features, NULL); nvp != NULL; nvp = nvlist_next_nvpair(features, nvp)) { char *propname; boolean_t found; if (zfeature_is_supported(nvpair_name(nvp))) continue; propname = zfs_asprintf(hdl, "unsupported@%s", nvpair_name(nvp)); /* * Before adding the property to the list make sure that no * other pool already added the same property. */ found = B_FALSE; entry = *plp; while (entry != NULL) { if (entry->pl_user_prop != NULL && strcmp(propname, entry->pl_user_prop) == 0) { found = B_TRUE; break; } entry = entry->pl_next; } if (found) { free(propname); continue; } entry = zfs_alloc(hdl, sizeof (zprop_list_t)); entry->pl_prop = ZPROP_USERPROP; entry->pl_user_prop = propname; entry->pl_width = strlen(entry->pl_user_prop); entry->pl_all = B_TRUE; *last = entry; last = &entry->pl_next; } for (entry = *plp; entry != NULL; entry = entry->pl_next) { if (entry->pl_fixed && !literal) continue; if (entry->pl_prop != ZPROP_USERPROP && zpool_get_prop(zhp, entry->pl_prop, buf, sizeof (buf), NULL, literal) == 0) { if (strlen(buf) > entry->pl_width) entry->pl_width = strlen(buf); } else if (entry->pl_prop == ZPROP_INVAL && zfs_prop_user(entry->pl_user_prop) && zpool_get_userprop(zhp, entry->pl_user_prop, buf, sizeof (buf), NULL) == 0) { if (strlen(buf) > entry->pl_width) entry->pl_width = strlen(buf); } } return (0); } int vdev_expand_proplist(zpool_handle_t *zhp, const char *vdevname, zprop_list_t **plp) { zprop_list_t *entry; char buf[ZFS_MAXPROPLEN]; const char *strval = NULL; int err = 0; nvpair_t *elem = NULL; nvlist_t *vprops = NULL; nvlist_t *propval = NULL; const char *propname; vdev_prop_t prop; zprop_list_t **last; for (entry = *plp; entry != NULL; entry = entry->pl_next) { if (entry->pl_fixed) continue; if (zpool_get_vdev_prop(zhp, vdevname, entry->pl_prop, entry->pl_user_prop, buf, sizeof (buf), NULL, B_FALSE) == 0) { if (strlen(buf) > entry->pl_width) entry->pl_width = strlen(buf); } if (entry->pl_prop == VDEV_PROP_NAME && strlen(vdevname) > entry->pl_width) entry->pl_width = strlen(vdevname); } /* Handle the all properties case */ last = plp; if (*last != NULL && (*last)->pl_all == B_TRUE) { while (*last != NULL) last = &(*last)->pl_next; err = zpool_get_all_vdev_props(zhp, vdevname, &vprops); if (err != 0) return (err); while ((elem = nvlist_next_nvpair(vprops, elem)) != NULL) { propname = nvpair_name(elem); /* Skip properties that are not user defined */ if ((prop = vdev_name_to_prop(propname)) != VDEV_PROP_USERPROP) continue; if (nvpair_value_nvlist(elem, &propval) != 0) continue; strval = fnvlist_lookup_string(propval, ZPROP_VALUE); entry = zfs_alloc(zhp->zpool_hdl, sizeof (zprop_list_t)); entry->pl_prop = prop; entry->pl_user_prop = zfs_strdup(zhp->zpool_hdl, propname); entry->pl_width = strlen(strval); entry->pl_all = B_TRUE; *last = entry; last = &entry->pl_next; } } return (0); } /* * Get the state for the given feature on the given ZFS pool. */ int zpool_prop_get_feature(zpool_handle_t *zhp, const char *propname, char *buf, size_t len) { uint64_t refcount; boolean_t found = B_FALSE; nvlist_t *features = zpool_get_features(zhp); boolean_t supported; const char *feature = strchr(propname, '@') + 1; supported = zpool_prop_feature(propname); ASSERT(supported || zpool_prop_unsupported(propname)); /* * Convert from feature name to feature guid. This conversion is * unnecessary for unsupported@... properties because they already * use guids. */ if (supported) { int ret; spa_feature_t fid; ret = zfeature_lookup_name(feature, &fid); if (ret != 0) { (void) strlcpy(buf, "-", len); return (ENOTSUP); } feature = spa_feature_table[fid].fi_guid; } if (nvlist_lookup_uint64(features, feature, &refcount) == 0) found = B_TRUE; if (supported) { if (!found) { (void) strlcpy(buf, ZFS_FEATURE_DISABLED, len); } else { if (refcount == 0) (void) strlcpy(buf, ZFS_FEATURE_ENABLED, len); else (void) strlcpy(buf, ZFS_FEATURE_ACTIVE, len); } } else { if (found) { if (refcount == 0) { (void) strcpy(buf, ZFS_UNSUPPORTED_INACTIVE); } else { (void) strcpy(buf, ZFS_UNSUPPORTED_READONLY); } } else { (void) strlcpy(buf, "-", len); return (ENOTSUP); } } return (0); } /* * Validate the given pool name, optionally putting an extended error message in * 'buf'. */ boolean_t zpool_name_valid(libzfs_handle_t *hdl, boolean_t isopen, const char *pool) { namecheck_err_t why; char what; int ret; ret = pool_namecheck(pool, &why, &what); /* * The rules for reserved pool names were extended at a later point. * But we need to support users with existing pools that may now be * invalid. So we only check for this expanded set of names during a * create (or import), and only in userland. */ if (ret == 0 && !isopen && (strncmp(pool, "mirror", 6) == 0 || strncmp(pool, "raidz", 5) == 0 || strncmp(pool, "draid", 5) == 0 || strncmp(pool, "spare", 5) == 0 || strcmp(pool, "log") == 0)) { if (hdl != NULL) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "name is reserved")); return (B_FALSE); } if (ret != 0) { if (hdl != NULL) { switch (why) { case NAME_ERR_TOOLONG: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "name is too long")); break; case NAME_ERR_INVALCHAR: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid character " "'%c' in pool name"), what); break; case NAME_ERR_NOLETTER: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "name must begin with a letter")); break; case NAME_ERR_RESERVED: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "name is reserved")); break; case NAME_ERR_DISKLIKE: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool name is reserved")); break; case NAME_ERR_LEADING_SLASH: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "leading slash in name")); break; case NAME_ERR_EMPTY_COMPONENT: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "empty component in name")); break; case NAME_ERR_TRAILING_SLASH: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "trailing slash in name")); break; case NAME_ERR_MULTIPLE_DELIMITERS: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "multiple '@' and/or '#' delimiters in " "name")); break; case NAME_ERR_NO_AT: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "permission set is missing '@'")); break; default: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "(%d) not defined"), why); break; } } return (B_FALSE); } return (B_TRUE); } /* * Open a handle to the given pool, even if the pool is currently in the FAULTED * state. */ zpool_handle_t * zpool_open_canfail(libzfs_handle_t *hdl, const char *pool) { zpool_handle_t *zhp; boolean_t missing; /* * Make sure the pool name is valid. */ if (!zpool_name_valid(hdl, B_TRUE, pool)) { (void) zfs_error_fmt(hdl, EZFS_INVALIDNAME, dgettext(TEXT_DOMAIN, "cannot open '%s'"), pool); return (NULL); } zhp = zfs_alloc(hdl, sizeof (zpool_handle_t)); zhp->zpool_hdl = hdl; (void) strlcpy(zhp->zpool_name, pool, sizeof (zhp->zpool_name)); if (zpool_refresh_stats(zhp, &missing) != 0) { zpool_close(zhp); return (NULL); } if (missing) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "no such pool")); (void) zfs_error_fmt(hdl, EZFS_NOENT, dgettext(TEXT_DOMAIN, "cannot open '%s'"), pool); zpool_close(zhp); return (NULL); } return (zhp); } /* * Like the above, but silent on error. Used when iterating over pools (because * the configuration cache may be out of date). */ int zpool_open_silent(libzfs_handle_t *hdl, const char *pool, zpool_handle_t **ret) { zpool_handle_t *zhp; boolean_t missing; zhp = zfs_alloc(hdl, sizeof (zpool_handle_t)); zhp->zpool_hdl = hdl; (void) strlcpy(zhp->zpool_name, pool, sizeof (zhp->zpool_name)); if (zpool_refresh_stats(zhp, &missing) != 0) { zpool_close(zhp); return (-1); } if (missing) { zpool_close(zhp); *ret = NULL; return (0); } *ret = zhp; return (0); } /* * Similar to zpool_open_canfail(), but refuses to open pools in the faulted * state. */ zpool_handle_t * zpool_open(libzfs_handle_t *hdl, const char *pool) { zpool_handle_t *zhp; if ((zhp = zpool_open_canfail(hdl, pool)) == NULL) return (NULL); if (zhp->zpool_state == POOL_STATE_UNAVAIL) { (void) zfs_error_fmt(hdl, EZFS_POOLUNAVAIL, dgettext(TEXT_DOMAIN, "cannot open '%s'"), zhp->zpool_name); zpool_close(zhp); return (NULL); } return (zhp); } /* * Close the handle. Simply frees the memory associated with the handle. */ void zpool_close(zpool_handle_t *zhp) { nvlist_free(zhp->zpool_config); nvlist_free(zhp->zpool_old_config); nvlist_free(zhp->zpool_props); free(zhp); } /* * Return the name of the pool. */ const char * zpool_get_name(zpool_handle_t *zhp) { return (zhp->zpool_name); } /* * Return the state of the pool (ACTIVE or UNAVAILABLE) */ int zpool_get_state(zpool_handle_t *zhp) { return (zhp->zpool_state); } /* * Check if vdev list contains a special vdev */ static boolean_t zpool_has_special_vdev(nvlist_t *nvroot) { nvlist_t **child; uint_t children; if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN, &child, &children) == 0) { for (uint_t c = 0; c < children; c++) { const char *bias; if (nvlist_lookup_string(child[c], ZPOOL_CONFIG_ALLOCATION_BIAS, &bias) == 0 && strcmp(bias, VDEV_ALLOC_BIAS_SPECIAL) == 0) { return (B_TRUE); } } } return (B_FALSE); } /* * Check if vdev list contains a dRAID vdev */ static boolean_t zpool_has_draid_vdev(nvlist_t *nvroot) { nvlist_t **child; uint_t children; if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN, &child, &children) == 0) { for (uint_t c = 0; c < children; c++) { const char *type; if (nvlist_lookup_string(child[c], ZPOOL_CONFIG_TYPE, &type) == 0 && strcmp(type, VDEV_TYPE_DRAID) == 0) { return (B_TRUE); } } } return (B_FALSE); } /* * Output a dRAID top-level vdev name in to the provided buffer. */ static char * zpool_draid_name(char *name, int len, uint64_t data, uint64_t parity, uint64_t spares, uint64_t children) { snprintf(name, len, "%s%llu:%llud:%lluc:%llus", VDEV_TYPE_DRAID, (u_longlong_t)parity, (u_longlong_t)data, (u_longlong_t)children, (u_longlong_t)spares); return (name); } /* * Return B_TRUE if the provided name is a dRAID spare name. */ boolean_t zpool_is_draid_spare(const char *name) { uint64_t spare_id, parity, vdev_id; if (sscanf(name, VDEV_TYPE_DRAID "%llu-%llu-%llu", (u_longlong_t *)&parity, (u_longlong_t *)&vdev_id, (u_longlong_t *)&spare_id) == 3) { return (B_TRUE); } return (B_FALSE); } /* * Create the named pool, using the provided vdev list. It is assumed * that the consumer has already validated the contents of the nvlist, so we * don't have to worry about error semantics. */ int zpool_create(libzfs_handle_t *hdl, const char *pool, nvlist_t *nvroot, nvlist_t *props, nvlist_t *fsprops) { zfs_cmd_t zc = {"\0"}; nvlist_t *zc_fsprops = NULL; nvlist_t *zc_props = NULL; nvlist_t *hidden_args = NULL; uint8_t *wkeydata = NULL; uint_t wkeylen = 0; char errbuf[ERRBUFLEN]; int ret = -1; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot create '%s'"), pool); if (!zpool_name_valid(hdl, B_FALSE, pool)) return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); zcmd_write_conf_nvlist(hdl, &zc, nvroot); if (props) { prop_flags_t flags = { .create = B_TRUE, .import = B_FALSE }; if ((zc_props = zpool_valid_proplist(hdl, pool, props, SPA_VERSION_1, flags, errbuf)) == NULL) { goto create_failed; } } if (fsprops) { uint64_t zoned; const char *zonestr; zoned = ((nvlist_lookup_string(fsprops, zfs_prop_to_name(ZFS_PROP_ZONED), &zonestr) == 0) && strcmp(zonestr, "on") == 0); if ((zc_fsprops = zfs_valid_proplist(hdl, ZFS_TYPE_FILESYSTEM, fsprops, zoned, NULL, NULL, B_TRUE, errbuf)) == NULL) { goto create_failed; } if (nvlist_exists(zc_fsprops, zfs_prop_to_name(ZFS_PROP_SPECIAL_SMALL_BLOCKS)) && !zpool_has_special_vdev(nvroot)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "%s property requires a special vdev"), zfs_prop_to_name(ZFS_PROP_SPECIAL_SMALL_BLOCKS)); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto create_failed; } if (!zc_props && (nvlist_alloc(&zc_props, NV_UNIQUE_NAME, 0) != 0)) { goto create_failed; } if (zfs_crypto_create(hdl, NULL, zc_fsprops, props, B_TRUE, &wkeydata, &wkeylen) != 0) { zfs_error(hdl, EZFS_CRYPTOFAILED, errbuf); goto create_failed; } if (nvlist_add_nvlist(zc_props, ZPOOL_ROOTFS_PROPS, zc_fsprops) != 0) { goto create_failed; } if (wkeydata != NULL) { if (nvlist_alloc(&hidden_args, NV_UNIQUE_NAME, 0) != 0) goto create_failed; if (nvlist_add_uint8_array(hidden_args, "wkeydata", wkeydata, wkeylen) != 0) goto create_failed; if (nvlist_add_nvlist(zc_props, ZPOOL_HIDDEN_ARGS, hidden_args) != 0) goto create_failed; } } if (zc_props) zcmd_write_src_nvlist(hdl, &zc, zc_props); (void) strlcpy(zc.zc_name, pool, sizeof (zc.zc_name)); if ((ret = zfs_ioctl(hdl, ZFS_IOC_POOL_CREATE, &zc)) != 0) { zcmd_free_nvlists(&zc); nvlist_free(zc_props); nvlist_free(zc_fsprops); nvlist_free(hidden_args); if (wkeydata != NULL) free(wkeydata); switch (errno) { case EBUSY: /* * This can happen if the user has specified the same * device multiple times. We can't reliably detect this * until we try to add it and see we already have a * label. This can also happen under if the device is * part of an active md or lvm device. */ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "one or more vdevs refer to the same device, or " "one of\nthe devices is part of an active md or " "lvm device")); return (zfs_error(hdl, EZFS_BADDEV, errbuf)); case ERANGE: /* * This happens if the record size is smaller or larger * than the allowed size range, or not a power of 2. * * NOTE: although zfs_valid_proplist is called earlier, * this case may have slipped through since the * pool does not exist yet and it is therefore * impossible to read properties e.g. max blocksize * from the pool. */ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "record size invalid")); return (zfs_error(hdl, EZFS_BADPROP, errbuf)); case EOVERFLOW: /* * This occurs when one of the devices is below * SPA_MINDEVSIZE. Unfortunately, we can't detect which * device was the problem device since there's no * reliable way to determine device size from userland. */ { char buf[64]; zfs_nicebytes(SPA_MINDEVSIZE, buf, sizeof (buf)); zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "one or more devices is less than the " "minimum size (%s)"), buf); } return (zfs_error(hdl, EZFS_BADDEV, errbuf)); case ENOSPC: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "one or more devices is out of space")); return (zfs_error(hdl, EZFS_BADDEV, errbuf)); case EINVAL: if (zpool_has_draid_vdev(nvroot) && zfeature_lookup_name("draid", NULL) != 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "dRAID vdevs are unsupported by the " "kernel")); return (zfs_error(hdl, EZFS_BADDEV, errbuf)); } else { return (zpool_standard_error(hdl, errno, errbuf)); } default: return (zpool_standard_error(hdl, errno, errbuf)); } } create_failed: zcmd_free_nvlists(&zc); nvlist_free(zc_props); nvlist_free(zc_fsprops); nvlist_free(hidden_args); if (wkeydata != NULL) free(wkeydata); return (ret); } /* * Destroy the given pool. It is up to the caller to ensure that there are no * datasets left in the pool. */ int zpool_destroy(zpool_handle_t *zhp, const char *log_str) { zfs_cmd_t zc = {"\0"}; zfs_handle_t *zfp = NULL; libzfs_handle_t *hdl = zhp->zpool_hdl; char errbuf[ERRBUFLEN]; if (zhp->zpool_state == POOL_STATE_ACTIVE && (zfp = zfs_open(hdl, zhp->zpool_name, ZFS_TYPE_FILESYSTEM)) == NULL) return (-1); (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); zc.zc_history = (uint64_t)(uintptr_t)log_str; if (zfs_ioctl(hdl, ZFS_IOC_POOL_DESTROY, &zc) != 0) { (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot destroy '%s'"), zhp->zpool_name); if (errno == EROFS) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "one or more devices is read only")); (void) zfs_error(hdl, EZFS_BADDEV, errbuf); } else { (void) zpool_standard_error(hdl, errno, errbuf); } if (zfp) zfs_close(zfp); return (-1); } if (zfp) { remove_mountpoint(zfp); zfs_close(zfp); } return (0); } /* * Create a checkpoint in the given pool. */ int zpool_checkpoint(zpool_handle_t *zhp) { libzfs_handle_t *hdl = zhp->zpool_hdl; char errbuf[ERRBUFLEN]; int error; error = lzc_pool_checkpoint(zhp->zpool_name); if (error != 0) { (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot checkpoint '%s'"), zhp->zpool_name); (void) zpool_standard_error(hdl, error, errbuf); return (-1); } return (0); } /* * Discard the checkpoint from the given pool. */ int zpool_discard_checkpoint(zpool_handle_t *zhp) { libzfs_handle_t *hdl = zhp->zpool_hdl; char errbuf[ERRBUFLEN]; int error; error = lzc_pool_checkpoint_discard(zhp->zpool_name); if (error != 0) { (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot discard checkpoint in '%s'"), zhp->zpool_name); (void) zpool_standard_error(hdl, error, errbuf); return (-1); } return (0); } /* * Add the given vdevs to the pool. The caller must have already performed the * necessary verification to ensure that the vdev specification is well-formed. */ int zpool_add(zpool_handle_t *zhp, nvlist_t *nvroot) { zfs_cmd_t zc = {"\0"}; int ret; libzfs_handle_t *hdl = zhp->zpool_hdl; char errbuf[ERRBUFLEN]; nvlist_t **spares, **l2cache; uint_t nspares, nl2cache; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot add to '%s'"), zhp->zpool_name); if (zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL) < SPA_VERSION_SPARES && nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, &spares, &nspares) == 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool must be " "upgraded to add hot spares")); return (zfs_error(hdl, EZFS_BADVERSION, errbuf)); } if (zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL) < SPA_VERSION_L2CACHE && nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache) == 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool must be " "upgraded to add cache devices")); return (zfs_error(hdl, EZFS_BADVERSION, errbuf)); } zcmd_write_conf_nvlist(hdl, &zc, nvroot); (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); if (zfs_ioctl(hdl, ZFS_IOC_VDEV_ADD, &zc) != 0) { switch (errno) { case EBUSY: /* * This can happen if the user has specified the same * device multiple times. We can't reliably detect this * until we try to add it and see we already have a * label. */ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "one or more vdevs refer to the same device")); (void) zfs_error(hdl, EZFS_BADDEV, errbuf); break; case EINVAL: if (zpool_has_draid_vdev(nvroot) && zfeature_lookup_name("draid", NULL) != 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "dRAID vdevs are unsupported by the " "kernel")); } else { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid config; a pool with removing/" "removed vdevs does not support adding " "raidz or dRAID vdevs")); } (void) zfs_error(hdl, EZFS_BADDEV, errbuf); break; case EOVERFLOW: /* * This occurs when one of the devices is below * SPA_MINDEVSIZE. Unfortunately, we can't detect which * device was the problem device since there's no * reliable way to determine device size from userland. */ { char buf[64]; zfs_nicebytes(SPA_MINDEVSIZE, buf, sizeof (buf)); zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "device is less than the minimum " "size (%s)"), buf); } (void) zfs_error(hdl, EZFS_BADDEV, errbuf); break; case ENOTSUP: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool must be upgraded to add these vdevs")); (void) zfs_error(hdl, EZFS_BADVERSION, errbuf); break; default: (void) zpool_standard_error(hdl, errno, errbuf); } ret = -1; } else { ret = 0; } zcmd_free_nvlists(&zc); return (ret); } /* * Exports the pool from the system. The caller must ensure that there are no * mounted datasets in the pool. */ static int zpool_export_common(zpool_handle_t *zhp, boolean_t force, boolean_t hardforce, const char *log_str) { zfs_cmd_t zc = {"\0"}; (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); zc.zc_cookie = force; zc.zc_guid = hardforce; zc.zc_history = (uint64_t)(uintptr_t)log_str; if (zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_POOL_EXPORT, &zc) != 0) { switch (errno) { case EXDEV: zfs_error_aux(zhp->zpool_hdl, dgettext(TEXT_DOMAIN, "use '-f' to override the following errors:\n" "'%s' has an active shared spare which could be" " used by other pools once '%s' is exported."), zhp->zpool_name, zhp->zpool_name); return (zfs_error_fmt(zhp->zpool_hdl, EZFS_ACTIVE_SPARE, dgettext(TEXT_DOMAIN, "cannot export '%s'"), zhp->zpool_name)); default: return (zpool_standard_error_fmt(zhp->zpool_hdl, errno, dgettext(TEXT_DOMAIN, "cannot export '%s'"), zhp->zpool_name)); } } return (0); } int zpool_export(zpool_handle_t *zhp, boolean_t force, const char *log_str) { return (zpool_export_common(zhp, force, B_FALSE, log_str)); } int zpool_export_force(zpool_handle_t *zhp, const char *log_str) { return (zpool_export_common(zhp, B_TRUE, B_TRUE, log_str)); } static void zpool_rewind_exclaim(libzfs_handle_t *hdl, const char *name, boolean_t dryrun, nvlist_t *config) { nvlist_t *nv = NULL; uint64_t rewindto; int64_t loss = -1; struct tm t; char timestr[128]; if (!hdl->libzfs_printerr || config == NULL) return; if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO, &nv) != 0 || nvlist_lookup_nvlist(nv, ZPOOL_CONFIG_REWIND_INFO, &nv) != 0) { return; } if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_LOAD_TIME, &rewindto) != 0) return; (void) nvlist_lookup_int64(nv, ZPOOL_CONFIG_REWIND_TIME, &loss); if (localtime_r((time_t *)&rewindto, &t) != NULL && strftime(timestr, 128, "%c", &t) != 0) { if (dryrun) { (void) printf(dgettext(TEXT_DOMAIN, "Would be able to return %s " "to its state as of %s.\n"), name, timestr); } else { (void) printf(dgettext(TEXT_DOMAIN, "Pool %s returned to its state as of %s.\n"), name, timestr); } if (loss > 120) { (void) printf(dgettext(TEXT_DOMAIN, "%s approximately %lld "), dryrun ? "Would discard" : "Discarded", ((longlong_t)loss + 30) / 60); (void) printf(dgettext(TEXT_DOMAIN, "minutes of transactions.\n")); } else if (loss > 0) { (void) printf(dgettext(TEXT_DOMAIN, "%s approximately %lld "), dryrun ? "Would discard" : "Discarded", (longlong_t)loss); (void) printf(dgettext(TEXT_DOMAIN, "seconds of transactions.\n")); } } } void zpool_explain_recover(libzfs_handle_t *hdl, const char *name, int reason, nvlist_t *config) { nvlist_t *nv = NULL; int64_t loss = -1; uint64_t edata = UINT64_MAX; uint64_t rewindto; struct tm t; char timestr[128]; if (!hdl->libzfs_printerr) return; if (reason >= 0) (void) printf(dgettext(TEXT_DOMAIN, "action: ")); else (void) printf(dgettext(TEXT_DOMAIN, "\t")); /* All attempted rewinds failed if ZPOOL_CONFIG_LOAD_TIME missing */ if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO, &nv) != 0 || nvlist_lookup_nvlist(nv, ZPOOL_CONFIG_REWIND_INFO, &nv) != 0 || nvlist_lookup_uint64(nv, ZPOOL_CONFIG_LOAD_TIME, &rewindto) != 0) goto no_info; (void) nvlist_lookup_int64(nv, ZPOOL_CONFIG_REWIND_TIME, &loss); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_LOAD_DATA_ERRORS, &edata); (void) printf(dgettext(TEXT_DOMAIN, "Recovery is possible, but will result in some data loss.\n")); if (localtime_r((time_t *)&rewindto, &t) != NULL && strftime(timestr, 128, "%c", &t) != 0) { (void) printf(dgettext(TEXT_DOMAIN, "\tReturning the pool to its state as of %s\n" "\tshould correct the problem. "), timestr); } else { (void) printf(dgettext(TEXT_DOMAIN, "\tReverting the pool to an earlier state " "should correct the problem.\n\t")); } if (loss > 120) { (void) printf(dgettext(TEXT_DOMAIN, "Approximately %lld minutes of data\n" "\tmust be discarded, irreversibly. "), ((longlong_t)loss + 30) / 60); } else if (loss > 0) { (void) printf(dgettext(TEXT_DOMAIN, "Approximately %lld seconds of data\n" "\tmust be discarded, irreversibly. "), (longlong_t)loss); } if (edata != 0 && edata != UINT64_MAX) { if (edata == 1) { (void) printf(dgettext(TEXT_DOMAIN, "After rewind, at least\n" "\tone persistent user-data error will remain. ")); } else { (void) printf(dgettext(TEXT_DOMAIN, "After rewind, several\n" "\tpersistent user-data errors will remain. ")); } } (void) printf(dgettext(TEXT_DOMAIN, "Recovery can be attempted\n\tby executing 'zpool %s -F %s'. "), reason >= 0 ? "clear" : "import", name); (void) printf(dgettext(TEXT_DOMAIN, "A scrub of the pool\n" "\tis strongly recommended after recovery.\n")); return; no_info: (void) printf(dgettext(TEXT_DOMAIN, "Destroy and re-create the pool from\n\ta backup source.\n")); } /* * zpool_import() is a contracted interface. Should be kept the same * if possible. * * Applications should use zpool_import_props() to import a pool with * new properties value to be set. */ int zpool_import(libzfs_handle_t *hdl, nvlist_t *config, const char *newname, char *altroot) { nvlist_t *props = NULL; int ret; if (altroot != NULL) { if (nvlist_alloc(&props, NV_UNIQUE_NAME, 0) != 0) { return (zfs_error_fmt(hdl, EZFS_NOMEM, dgettext(TEXT_DOMAIN, "cannot import '%s'"), newname)); } if (nvlist_add_string(props, zpool_prop_to_name(ZPOOL_PROP_ALTROOT), altroot) != 0 || nvlist_add_string(props, zpool_prop_to_name(ZPOOL_PROP_CACHEFILE), "none") != 0) { nvlist_free(props); return (zfs_error_fmt(hdl, EZFS_NOMEM, dgettext(TEXT_DOMAIN, "cannot import '%s'"), newname)); } } ret = zpool_import_props(hdl, config, newname, props, ZFS_IMPORT_NORMAL); nvlist_free(props); return (ret); } static void print_vdev_tree(libzfs_handle_t *hdl, const char *name, nvlist_t *nv, int indent) { nvlist_t **child; uint_t c, children; char *vname; uint64_t is_log = 0; (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_IS_LOG, &is_log); if (name != NULL) (void) printf("\t%*s%s%s\n", indent, "", name, is_log ? " [log]" : ""); if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) return; for (c = 0; c < children; c++) { vname = zpool_vdev_name(hdl, NULL, child[c], VDEV_NAME_TYPE_ID); print_vdev_tree(hdl, vname, child[c], indent + 2); free(vname); } } void zpool_print_unsup_feat(nvlist_t *config) { nvlist_t *nvinfo, *unsup_feat; nvinfo = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO); unsup_feat = fnvlist_lookup_nvlist(nvinfo, ZPOOL_CONFIG_UNSUP_FEAT); for (nvpair_t *nvp = nvlist_next_nvpair(unsup_feat, NULL); nvp != NULL; nvp = nvlist_next_nvpair(unsup_feat, nvp)) { const char *desc = fnvpair_value_string(nvp); if (strlen(desc) > 0) (void) printf("\t%s (%s)\n", nvpair_name(nvp), desc); else (void) printf("\t%s\n", nvpair_name(nvp)); } } /* * Import the given pool using the known configuration and a list of * properties to be set. The configuration should have come from * zpool_find_import(). The 'newname' parameters control whether the pool * is imported with a different name. */ int zpool_import_props(libzfs_handle_t *hdl, nvlist_t *config, const char *newname, nvlist_t *props, int flags) { zfs_cmd_t zc = {"\0"}; zpool_load_policy_t policy; nvlist_t *nv = NULL; nvlist_t *nvinfo = NULL; nvlist_t *missing = NULL; const char *thename; const char *origname; int ret; int error = 0; char errbuf[ERRBUFLEN]; origname = fnvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME); (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot import pool '%s'"), origname); if (newname != NULL) { if (!zpool_name_valid(hdl, B_FALSE, newname)) return (zfs_error_fmt(hdl, EZFS_INVALIDNAME, dgettext(TEXT_DOMAIN, "cannot import '%s'"), newname)); thename = newname; } else { thename = origname; } if (props != NULL) { uint64_t version; prop_flags_t flags = { .create = B_FALSE, .import = B_TRUE }; version = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION); if ((props = zpool_valid_proplist(hdl, origname, props, version, flags, errbuf)) == NULL) return (-1); zcmd_write_src_nvlist(hdl, &zc, props); nvlist_free(props); } (void) strlcpy(zc.zc_name, thename, sizeof (zc.zc_name)); zc.zc_guid = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID); zcmd_write_conf_nvlist(hdl, &zc, config); zcmd_alloc_dst_nvlist(hdl, &zc, zc.zc_nvlist_conf_size * 2); zc.zc_cookie = flags; while ((ret = zfs_ioctl(hdl, ZFS_IOC_POOL_IMPORT, &zc)) != 0 && errno == ENOMEM) zcmd_expand_dst_nvlist(hdl, &zc); if (ret != 0) error = errno; (void) zcmd_read_dst_nvlist(hdl, &zc, &nv); zcmd_free_nvlists(&zc); zpool_get_load_policy(config, &policy); if (error) { char desc[1024]; char aux[256]; /* * Dry-run failed, but we print out what success * looks like if we found a best txg */ if (policy.zlp_rewind & ZPOOL_TRY_REWIND) { zpool_rewind_exclaim(hdl, newname ? origname : thename, B_TRUE, nv); nvlist_free(nv); return (-1); } if (newname == NULL) (void) snprintf(desc, sizeof (desc), dgettext(TEXT_DOMAIN, "cannot import '%s'"), thename); else (void) snprintf(desc, sizeof (desc), dgettext(TEXT_DOMAIN, "cannot import '%s' as '%s'"), origname, thename); switch (error) { case ENOTSUP: if (nv != NULL && nvlist_lookup_nvlist(nv, ZPOOL_CONFIG_LOAD_INFO, &nvinfo) == 0 && nvlist_exists(nvinfo, ZPOOL_CONFIG_UNSUP_FEAT)) { (void) printf(dgettext(TEXT_DOMAIN, "This " "pool uses the following feature(s) not " "supported by this system:\n")); zpool_print_unsup_feat(nv); if (nvlist_exists(nvinfo, ZPOOL_CONFIG_CAN_RDONLY)) { (void) printf(dgettext(TEXT_DOMAIN, "All unsupported features are only " "required for writing to the pool." "\nThe pool can be imported using " "'-o readonly=on'.\n")); } } /* * Unsupported version. */ (void) zfs_error(hdl, EZFS_BADVERSION, desc); break; case EREMOTEIO: if (nv != NULL && nvlist_lookup_nvlist(nv, ZPOOL_CONFIG_LOAD_INFO, &nvinfo) == 0) { const char *hostname = ""; uint64_t hostid = 0; mmp_state_t mmp_state; mmp_state = fnvlist_lookup_uint64(nvinfo, ZPOOL_CONFIG_MMP_STATE); if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_HOSTNAME)) hostname = fnvlist_lookup_string(nvinfo, ZPOOL_CONFIG_MMP_HOSTNAME); if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_HOSTID)) hostid = fnvlist_lookup_uint64(nvinfo, ZPOOL_CONFIG_MMP_HOSTID); if (mmp_state == MMP_STATE_ACTIVE) { (void) snprintf(aux, sizeof (aux), dgettext(TEXT_DOMAIN, "pool is imp" "orted on host '%s' (hostid=%lx).\n" "Export the pool on the other " "system, then run 'zpool import'."), hostname, (unsigned long) hostid); } else if (mmp_state == MMP_STATE_NO_HOSTID) { (void) snprintf(aux, sizeof (aux), dgettext(TEXT_DOMAIN, "pool has " "the multihost property on and " "the\nsystem's hostid is not set. " "Set a unique system hostid with " "the zgenhostid(8) command.\n")); } (void) zfs_error_aux(hdl, "%s", aux); } (void) zfs_error(hdl, EZFS_ACTIVE_POOL, desc); break; case EINVAL: (void) zfs_error(hdl, EZFS_INVALCONFIG, desc); break; case EROFS: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "one or more devices is read only")); (void) zfs_error(hdl, EZFS_BADDEV, desc); break; case ENXIO: if (nv && nvlist_lookup_nvlist(nv, ZPOOL_CONFIG_LOAD_INFO, &nvinfo) == 0 && nvlist_lookup_nvlist(nvinfo, ZPOOL_CONFIG_MISSING_DEVICES, &missing) == 0) { (void) printf(dgettext(TEXT_DOMAIN, "The devices below are missing or " "corrupted, use '-m' to import the pool " "anyway:\n")); print_vdev_tree(hdl, NULL, missing, 2); (void) printf("\n"); } (void) zpool_standard_error(hdl, error, desc); break; case EEXIST: (void) zpool_standard_error(hdl, error, desc); break; case EBUSY: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "one or more devices are already in use\n")); (void) zfs_error(hdl, EZFS_BADDEV, desc); break; case ENAMETOOLONG: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "new name of at least one dataset is longer than " "the maximum allowable length")); (void) zfs_error(hdl, EZFS_NAMETOOLONG, desc); break; default: (void) zpool_standard_error(hdl, error, desc); zpool_explain_recover(hdl, newname ? origname : thename, -error, nv); break; } nvlist_free(nv); ret = -1; } else { zpool_handle_t *zhp; /* * This should never fail, but play it safe anyway. */ if (zpool_open_silent(hdl, thename, &zhp) != 0) ret = -1; else if (zhp != NULL) zpool_close(zhp); if (policy.zlp_rewind & (ZPOOL_DO_REWIND | ZPOOL_TRY_REWIND)) { zpool_rewind_exclaim(hdl, newname ? origname : thename, ((policy.zlp_rewind & ZPOOL_TRY_REWIND) != 0), nv); } nvlist_free(nv); } return (ret); } /* * Translate vdev names to guids. If a vdev_path is determined to be * unsuitable then a vd_errlist is allocated and the vdev path and errno * are added to it. */ static int zpool_translate_vdev_guids(zpool_handle_t *zhp, nvlist_t *vds, nvlist_t *vdev_guids, nvlist_t *guids_to_paths, nvlist_t **vd_errlist) { nvlist_t *errlist = NULL; int error = 0; for (nvpair_t *elem = nvlist_next_nvpair(vds, NULL); elem != NULL; elem = nvlist_next_nvpair(vds, elem)) { boolean_t spare, cache; const char *vd_path = nvpair_name(elem); nvlist_t *tgt = zpool_find_vdev(zhp, vd_path, &spare, &cache, NULL); if ((tgt == NULL) || cache || spare) { if (errlist == NULL) { errlist = fnvlist_alloc(); error = EINVAL; } uint64_t err = (tgt == NULL) ? EZFS_NODEVICE : (spare ? EZFS_ISSPARE : EZFS_ISL2CACHE); fnvlist_add_int64(errlist, vd_path, err); continue; } uint64_t guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID); fnvlist_add_uint64(vdev_guids, vd_path, guid); char msg[MAXNAMELEN]; (void) snprintf(msg, sizeof (msg), "%llu", (u_longlong_t)guid); fnvlist_add_string(guids_to_paths, msg, vd_path); } if (error != 0) { verify(errlist != NULL); if (vd_errlist != NULL) *vd_errlist = errlist; else fnvlist_free(errlist); } return (error); } static int xlate_init_err(int err) { switch (err) { case ENODEV: return (EZFS_NODEVICE); case EINVAL: case EROFS: return (EZFS_BADDEV); case EBUSY: return (EZFS_INITIALIZING); case ESRCH: return (EZFS_NO_INITIALIZE); } return (err); } /* - * Begin, suspend, or cancel the initialization (initializing of all free - * blocks) for the given vdevs in the given pool. + * Begin, suspend, cancel, or uninit (clear) the initialization (initializing + * of all free blocks) for the given vdevs in the given pool. */ static int zpool_initialize_impl(zpool_handle_t *zhp, pool_initialize_func_t cmd_type, nvlist_t *vds, boolean_t wait) { int err; nvlist_t *vdev_guids = fnvlist_alloc(); nvlist_t *guids_to_paths = fnvlist_alloc(); nvlist_t *vd_errlist = NULL; nvlist_t *errlist; nvpair_t *elem; err = zpool_translate_vdev_guids(zhp, vds, vdev_guids, guids_to_paths, &vd_errlist); if (err != 0) { verify(vd_errlist != NULL); goto list_errors; } err = lzc_initialize(zhp->zpool_name, cmd_type, vdev_guids, &errlist); if (err != 0) { - if (errlist != NULL) { - vd_errlist = fnvlist_lookup_nvlist(errlist, - ZPOOL_INITIALIZE_VDEVS); + if (errlist != NULL && nvlist_lookup_nvlist(errlist, + ZPOOL_INITIALIZE_VDEVS, &vd_errlist) == 0) { goto list_errors; } + + if (err == EINVAL && cmd_type == POOL_INITIALIZE_UNINIT) { + zfs_error_aux(zhp->zpool_hdl, dgettext(TEXT_DOMAIN, + "uninitialize is not supported by kernel")); + } + (void) zpool_standard_error(zhp->zpool_hdl, err, dgettext(TEXT_DOMAIN, "operation failed")); goto out; } if (wait) { for (elem = nvlist_next_nvpair(vdev_guids, NULL); elem != NULL; elem = nvlist_next_nvpair(vdev_guids, elem)) { uint64_t guid = fnvpair_value_uint64(elem); err = lzc_wait_tag(zhp->zpool_name, ZPOOL_WAIT_INITIALIZE, guid, NULL); if (err != 0) { (void) zpool_standard_error_fmt(zhp->zpool_hdl, err, dgettext(TEXT_DOMAIN, "error " "waiting for '%s' to initialize"), nvpair_name(elem)); goto out; } } } goto out; list_errors: for (elem = nvlist_next_nvpair(vd_errlist, NULL); elem != NULL; elem = nvlist_next_nvpair(vd_errlist, elem)) { int64_t vd_error = xlate_init_err(fnvpair_value_int64(elem)); const char *path; if (nvlist_lookup_string(guids_to_paths, nvpair_name(elem), &path) != 0) path = nvpair_name(elem); (void) zfs_error_fmt(zhp->zpool_hdl, vd_error, "cannot initialize '%s'", path); } out: fnvlist_free(vdev_guids); fnvlist_free(guids_to_paths); if (vd_errlist != NULL) fnvlist_free(vd_errlist); return (err == 0 ? 0 : -1); } int zpool_initialize(zpool_handle_t *zhp, pool_initialize_func_t cmd_type, nvlist_t *vds) { return (zpool_initialize_impl(zhp, cmd_type, vds, B_FALSE)); } int zpool_initialize_wait(zpool_handle_t *zhp, pool_initialize_func_t cmd_type, nvlist_t *vds) { return (zpool_initialize_impl(zhp, cmd_type, vds, B_TRUE)); } static int xlate_trim_err(int err) { switch (err) { case ENODEV: return (EZFS_NODEVICE); case EINVAL: case EROFS: return (EZFS_BADDEV); case EBUSY: return (EZFS_TRIMMING); case ESRCH: return (EZFS_NO_TRIM); case EOPNOTSUPP: return (EZFS_TRIM_NOTSUP); } return (err); } static int zpool_trim_wait(zpool_handle_t *zhp, nvlist_t *vdev_guids) { int err; nvpair_t *elem; for (elem = nvlist_next_nvpair(vdev_guids, NULL); elem != NULL; elem = nvlist_next_nvpair(vdev_guids, elem)) { uint64_t guid = fnvpair_value_uint64(elem); err = lzc_wait_tag(zhp->zpool_name, ZPOOL_WAIT_TRIM, guid, NULL); if (err != 0) { (void) zpool_standard_error_fmt(zhp->zpool_hdl, err, dgettext(TEXT_DOMAIN, "error " "waiting to trim '%s'"), nvpair_name(elem)); return (err); } } return (0); } /* * Check errlist and report any errors, omitting ones which should be * suppressed. Returns B_TRUE if any errors were reported. */ static boolean_t check_trim_errs(zpool_handle_t *zhp, trimflags_t *trim_flags, nvlist_t *guids_to_paths, nvlist_t *vds, nvlist_t *errlist) { nvpair_t *elem; boolean_t reported_errs = B_FALSE; int num_vds = 0; int num_suppressed_errs = 0; for (elem = nvlist_next_nvpair(vds, NULL); elem != NULL; elem = nvlist_next_nvpair(vds, elem)) { num_vds++; } for (elem = nvlist_next_nvpair(errlist, NULL); elem != NULL; elem = nvlist_next_nvpair(errlist, elem)) { int64_t vd_error = xlate_trim_err(fnvpair_value_int64(elem)); const char *path; /* * If only the pool was specified, and it was not a secure * trim then suppress warnings for individual vdevs which * do not support trimming. */ if (vd_error == EZFS_TRIM_NOTSUP && trim_flags->fullpool && !trim_flags->secure) { num_suppressed_errs++; continue; } reported_errs = B_TRUE; if (nvlist_lookup_string(guids_to_paths, nvpair_name(elem), &path) != 0) path = nvpair_name(elem); (void) zfs_error_fmt(zhp->zpool_hdl, vd_error, "cannot trim '%s'", path); } if (num_suppressed_errs == num_vds) { (void) zfs_error_aux(zhp->zpool_hdl, dgettext(TEXT_DOMAIN, "no devices in pool support trim operations")); (void) (zfs_error(zhp->zpool_hdl, EZFS_TRIM_NOTSUP, dgettext(TEXT_DOMAIN, "cannot trim"))); reported_errs = B_TRUE; } return (reported_errs); } /* * Begin, suspend, or cancel the TRIM (discarding of all free blocks) for * the given vdevs in the given pool. */ int zpool_trim(zpool_handle_t *zhp, pool_trim_func_t cmd_type, nvlist_t *vds, trimflags_t *trim_flags) { int err; int retval = 0; nvlist_t *vdev_guids = fnvlist_alloc(); nvlist_t *guids_to_paths = fnvlist_alloc(); nvlist_t *errlist = NULL; err = zpool_translate_vdev_guids(zhp, vds, vdev_guids, guids_to_paths, &errlist); if (err != 0) { check_trim_errs(zhp, trim_flags, guids_to_paths, vds, errlist); retval = -1; goto out; } err = lzc_trim(zhp->zpool_name, cmd_type, trim_flags->rate, trim_flags->secure, vdev_guids, &errlist); if (err != 0) { nvlist_t *vd_errlist; if (errlist != NULL && nvlist_lookup_nvlist(errlist, ZPOOL_TRIM_VDEVS, &vd_errlist) == 0) { if (check_trim_errs(zhp, trim_flags, guids_to_paths, vds, vd_errlist)) { retval = -1; goto out; } } else { char errbuf[ERRBUFLEN]; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "operation failed")); zpool_standard_error(zhp->zpool_hdl, err, errbuf); retval = -1; goto out; } } if (trim_flags->wait) retval = zpool_trim_wait(zhp, vdev_guids); out: if (errlist != NULL) fnvlist_free(errlist); fnvlist_free(vdev_guids); fnvlist_free(guids_to_paths); return (retval); } /* * Scan the pool. */ int zpool_scan(zpool_handle_t *zhp, pool_scan_func_t func, pool_scrub_cmd_t cmd) { - zfs_cmd_t zc = {"\0"}; char errbuf[ERRBUFLEN]; int err; libzfs_handle_t *hdl = zhp->zpool_hdl; - (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); - zc.zc_cookie = func; - zc.zc_flags = cmd; + nvlist_t *args = fnvlist_alloc(); + fnvlist_add_uint64(args, "scan_type", (uint64_t)func); + fnvlist_add_uint64(args, "scan_command", (uint64_t)cmd); - if (zfs_ioctl(hdl, ZFS_IOC_POOL_SCAN, &zc) == 0) + err = lzc_scrub(ZFS_IOC_POOL_SCRUB, zhp->zpool_name, args, NULL); + fnvlist_free(args); + + if (err == 0) { return (0); + } else if (err == ZFS_ERR_IOC_CMD_UNAVAIL) { + zfs_cmd_t zc = {"\0"}; + (void) strlcpy(zc.zc_name, zhp->zpool_name, + sizeof (zc.zc_name)); + zc.zc_cookie = func; + zc.zc_flags = cmd; - err = errno; + if (zfs_ioctl(hdl, ZFS_IOC_POOL_SCAN, &zc) == 0) + return (0); + } - /* ECANCELED on a scrub means we resumed a paused scrub */ - if (err == ECANCELED && func == POOL_SCAN_SCRUB && - cmd == POOL_SCRUB_NORMAL) + /* + * An ECANCELED on a scrub means one of the following: + * 1. we resumed a paused scrub. + * 2. we resumed a paused error scrub. + * 3. Error scrub is not run because of no error log. + */ + if (err == ECANCELED && (func == POOL_SCAN_SCRUB || + func == POOL_SCAN_ERRORSCRUB) && cmd == POOL_SCRUB_NORMAL) return (0); - - if (err == ENOENT && func != POOL_SCAN_NONE && cmd == POOL_SCRUB_NORMAL) + /* + * The following cases have been handled here: + * 1. Paused a scrub/error scrub if there is none in progress. + */ + if (err == ENOENT && func != POOL_SCAN_NONE && cmd == + POOL_SCRUB_PAUSE) { return (0); + } + + ASSERT3U(func, >=, POOL_SCAN_NONE); + ASSERT3U(func, <, POOL_SCAN_FUNCS); - if (func == POOL_SCAN_SCRUB) { + if (func == POOL_SCAN_SCRUB || func == POOL_SCAN_ERRORSCRUB) { if (cmd == POOL_SCRUB_PAUSE) { (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot pause scrubbing %s"), - zc.zc_name); + zhp->zpool_name); } else { assert(cmd == POOL_SCRUB_NORMAL); (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot scrub %s"), - zc.zc_name); + zhp->zpool_name); } } else if (func == POOL_SCAN_RESILVER) { assert(cmd == POOL_SCRUB_NORMAL); (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, - "cannot restart resilver on %s"), zc.zc_name); + "cannot restart resilver on %s"), zhp->zpool_name); } else if (func == POOL_SCAN_NONE) { (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, - "cannot cancel scrubbing %s"), zc.zc_name); + "cannot cancel scrubbing %s"), zhp->zpool_name); } else { assert(!"unexpected result"); } + /* + * With EBUSY, five cases are possible: + * + * Current state Requested + * 1. Normal Scrub Running Normal Scrub or Error Scrub + * 2. Normal Scrub Paused Error Scrub + * 3. Normal Scrub Paused Pause Normal Scrub + * 4. Error Scrub Running Normal Scrub or Error Scrub + * 5. Error Scrub Paused Pause Error Scrub + * 6. Resilvering Anything else + */ if (err == EBUSY) { nvlist_t *nvroot; pool_scan_stat_t *ps = NULL; uint_t psc; nvroot = fnvlist_lookup_nvlist(zhp->zpool_config, ZPOOL_CONFIG_VDEV_TREE); (void) nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_SCAN_STATS, (uint64_t **)&ps, &psc); if (ps && ps->pss_func == POOL_SCAN_SCRUB && ps->pss_state == DSS_SCANNING) { - if (cmd == POOL_SCRUB_PAUSE) - return (zfs_error(hdl, EZFS_SCRUB_PAUSED, + if (ps->pss_pass_scrub_pause == 0) { + /* handles case 1 */ + assert(cmd == POOL_SCRUB_NORMAL); + return (zfs_error(hdl, EZFS_SCRUBBING, errbuf)); - else - return (zfs_error(hdl, EZFS_SCRUBBING, errbuf)); + } else { + if (func == POOL_SCAN_ERRORSCRUB) { + /* handles case 2 */ + ASSERT3U(cmd, ==, POOL_SCRUB_NORMAL); + return (zfs_error(hdl, + EZFS_SCRUB_PAUSED_TO_CANCEL, + errbuf)); + } else { + /* handles case 3 */ + ASSERT3U(func, ==, POOL_SCAN_SCRUB); + ASSERT3U(cmd, ==, POOL_SCRUB_PAUSE); + return (zfs_error(hdl, + EZFS_SCRUB_PAUSED, errbuf)); + } + } + } else if (ps && + ps->pss_error_scrub_func == POOL_SCAN_ERRORSCRUB && + ps->pss_error_scrub_state == DSS_ERRORSCRUBBING) { + if (ps->pss_pass_error_scrub_pause == 0) { + /* handles case 4 */ + ASSERT3U(cmd, ==, POOL_SCRUB_NORMAL); + return (zfs_error(hdl, EZFS_ERRORSCRUBBING, + errbuf)); + } else { + /* handles case 5 */ + ASSERT3U(func, ==, POOL_SCAN_ERRORSCRUB); + ASSERT3U(cmd, ==, POOL_SCRUB_PAUSE); + return (zfs_error(hdl, EZFS_ERRORSCRUB_PAUSED, + errbuf)); + } } else { + /* handles case 6 */ return (zfs_error(hdl, EZFS_RESILVERING, errbuf)); } } else if (err == ENOENT) { return (zfs_error(hdl, EZFS_NO_SCRUB, errbuf)); } else if (err == ENOTSUP && func == POOL_SCAN_RESILVER) { return (zfs_error(hdl, EZFS_NO_RESILVER_DEFER, errbuf)); } else { return (zpool_standard_error(hdl, err, errbuf)); } } /* * Find a vdev that matches the search criteria specified. We use the * the nvpair name to determine how we should look for the device. * 'avail_spare' is set to TRUE if the provided guid refers to an AVAIL * spare; but FALSE if its an INUSE spare. */ static nvlist_t * vdev_to_nvlist_iter(nvlist_t *nv, nvlist_t *search, boolean_t *avail_spare, boolean_t *l2cache, boolean_t *log) { uint_t c, children; nvlist_t **child; nvlist_t *ret; uint64_t is_log; const char *srchkey; nvpair_t *pair = nvlist_next_nvpair(search, NULL); /* Nothing to look for */ if (search == NULL || pair == NULL) return (NULL); /* Obtain the key we will use to search */ srchkey = nvpair_name(pair); switch (nvpair_type(pair)) { case DATA_TYPE_UINT64: if (strcmp(srchkey, ZPOOL_CONFIG_GUID) == 0) { uint64_t srchval = fnvpair_value_uint64(pair); uint64_t theguid = fnvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID); if (theguid == srchval) return (nv); } break; case DATA_TYPE_STRING: { const char *srchval, *val; srchval = fnvpair_value_string(pair); if (nvlist_lookup_string(nv, srchkey, &val) != 0) break; /* * Search for the requested value. Special cases: * * - ZPOOL_CONFIG_PATH for whole disk entries. These end in * "-part1", or "p1". The suffix is hidden from the user, * but included in the string, so this matches around it. * - ZPOOL_CONFIG_PATH for short names zfs_strcmp_shortname() * is used to check all possible expanded paths. * - looking for a top-level vdev name (i.e. ZPOOL_CONFIG_TYPE). * * Otherwise, all other searches are simple string compares. */ if (strcmp(srchkey, ZPOOL_CONFIG_PATH) == 0) { uint64_t wholedisk = 0; (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK, &wholedisk); if (zfs_strcmp_pathname(srchval, val, wholedisk) == 0) return (nv); } else if (strcmp(srchkey, ZPOOL_CONFIG_TYPE) == 0) { char *type, *idx, *end, *p; uint64_t id, vdev_id; /* * Determine our vdev type, keeping in mind * that the srchval is composed of a type and * vdev id pair (i.e. mirror-4). */ if ((type = strdup(srchval)) == NULL) return (NULL); if ((p = strrchr(type, '-')) == NULL) { free(type); break; } idx = p + 1; *p = '\0'; /* * If the types don't match then keep looking. */ if (strncmp(val, type, strlen(val)) != 0) { free(type); break; } verify(zpool_vdev_is_interior(type)); id = fnvlist_lookup_uint64(nv, ZPOOL_CONFIG_ID); errno = 0; vdev_id = strtoull(idx, &end, 10); /* * If we are looking for a raidz and a parity is * specified, make sure it matches. */ int rzlen = strlen(VDEV_TYPE_RAIDZ); assert(rzlen == strlen(VDEV_TYPE_DRAID)); int typlen = strlen(type); if ((strncmp(type, VDEV_TYPE_RAIDZ, rzlen) == 0 || strncmp(type, VDEV_TYPE_DRAID, rzlen) == 0) && typlen != rzlen) { uint64_t vdev_parity; int parity = *(type + rzlen) - '0'; if (parity <= 0 || parity > 3 || (typlen - rzlen) != 1) { /* * Nonsense parity specified, can * never match */ free(type); return (NULL); } vdev_parity = fnvlist_lookup_uint64(nv, ZPOOL_CONFIG_NPARITY); if ((int)vdev_parity != parity) { free(type); break; } } free(type); if (errno != 0) return (NULL); /* * Now verify that we have the correct vdev id. */ if (vdev_id == id) return (nv); } /* * Common case */ if (strcmp(srchval, val) == 0) return (nv); break; } default: break; } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) return (NULL); for (c = 0; c < children; c++) { if ((ret = vdev_to_nvlist_iter(child[c], search, avail_spare, l2cache, NULL)) != NULL) { /* * The 'is_log' value is only set for the toplevel * vdev, not the leaf vdevs. So we always lookup the * log device from the root of the vdev tree (where * 'log' is non-NULL). */ if (log != NULL && nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, &is_log) == 0 && is_log) { *log = B_TRUE; } return (ret); } } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES, &child, &children) == 0) { for (c = 0; c < children; c++) { if ((ret = vdev_to_nvlist_iter(child[c], search, avail_spare, l2cache, NULL)) != NULL) { *avail_spare = B_TRUE; return (ret); } } } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE, &child, &children) == 0) { for (c = 0; c < children; c++) { if ((ret = vdev_to_nvlist_iter(child[c], search, avail_spare, l2cache, NULL)) != NULL) { *l2cache = B_TRUE; return (ret); } } } return (NULL); } /* * Given a physical path or guid, find the associated vdev. */ nvlist_t * zpool_find_vdev_by_physpath(zpool_handle_t *zhp, const char *ppath, boolean_t *avail_spare, boolean_t *l2cache, boolean_t *log) { nvlist_t *search, *nvroot, *ret; uint64_t guid; char *end; search = fnvlist_alloc(); guid = strtoull(ppath, &end, 0); if (guid != 0 && *end == '\0') { fnvlist_add_uint64(search, ZPOOL_CONFIG_GUID, guid); } else { fnvlist_add_string(search, ZPOOL_CONFIG_PHYS_PATH, ppath); } nvroot = fnvlist_lookup_nvlist(zhp->zpool_config, ZPOOL_CONFIG_VDEV_TREE); *avail_spare = B_FALSE; *l2cache = B_FALSE; if (log != NULL) *log = B_FALSE; ret = vdev_to_nvlist_iter(nvroot, search, avail_spare, l2cache, log); fnvlist_free(search); return (ret); } /* * Determine if we have an "interior" top-level vdev (i.e mirror/raidz). */ static boolean_t zpool_vdev_is_interior(const char *name) { if (strncmp(name, VDEV_TYPE_RAIDZ, strlen(VDEV_TYPE_RAIDZ)) == 0 || strncmp(name, VDEV_TYPE_SPARE, strlen(VDEV_TYPE_SPARE)) == 0 || strncmp(name, VDEV_TYPE_REPLACING, strlen(VDEV_TYPE_REPLACING)) == 0 || strncmp(name, VDEV_TYPE_ROOT, strlen(VDEV_TYPE_ROOT)) == 0 || strncmp(name, VDEV_TYPE_MIRROR, strlen(VDEV_TYPE_MIRROR)) == 0) return (B_TRUE); if (strncmp(name, VDEV_TYPE_DRAID, strlen(VDEV_TYPE_DRAID)) == 0 && !zpool_is_draid_spare(name)) return (B_TRUE); return (B_FALSE); } nvlist_t * zpool_find_vdev(zpool_handle_t *zhp, const char *path, boolean_t *avail_spare, boolean_t *l2cache, boolean_t *log) { char *end; nvlist_t *nvroot, *search, *ret; uint64_t guid; search = fnvlist_alloc(); guid = strtoull(path, &end, 0); if (guid != 0 && *end == '\0') { fnvlist_add_uint64(search, ZPOOL_CONFIG_GUID, guid); } else if (zpool_vdev_is_interior(path)) { fnvlist_add_string(search, ZPOOL_CONFIG_TYPE, path); } else { fnvlist_add_string(search, ZPOOL_CONFIG_PATH, path); } nvroot = fnvlist_lookup_nvlist(zhp->zpool_config, ZPOOL_CONFIG_VDEV_TREE); *avail_spare = B_FALSE; *l2cache = B_FALSE; if (log != NULL) *log = B_FALSE; ret = vdev_to_nvlist_iter(nvroot, search, avail_spare, l2cache, log); fnvlist_free(search); return (ret); } /* * Convert a vdev path to a GUID. Returns GUID or 0 on error. * * If is_spare, is_l2cache, or is_log is non-NULL, then store within it * if the VDEV is a spare, l2cache, or log device. If they're NULL then * ignore them. */ static uint64_t zpool_vdev_path_to_guid_impl(zpool_handle_t *zhp, const char *path, boolean_t *is_spare, boolean_t *is_l2cache, boolean_t *is_log) { boolean_t spare = B_FALSE, l2cache = B_FALSE, log = B_FALSE; nvlist_t *tgt; if ((tgt = zpool_find_vdev(zhp, path, &spare, &l2cache, &log)) == NULL) return (0); if (is_spare != NULL) *is_spare = spare; if (is_l2cache != NULL) *is_l2cache = l2cache; if (is_log != NULL) *is_log = log; return (fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID)); } /* Convert a vdev path to a GUID. Returns GUID or 0 on error. */ uint64_t zpool_vdev_path_to_guid(zpool_handle_t *zhp, const char *path) { return (zpool_vdev_path_to_guid_impl(zhp, path, NULL, NULL, NULL)); } /* * Bring the specified vdev online. The 'flags' parameter is a set of the * ZFS_ONLINE_* flags. */ int zpool_vdev_online(zpool_handle_t *zhp, const char *path, int flags, vdev_state_t *newstate) { zfs_cmd_t zc = {"\0"}; char errbuf[ERRBUFLEN]; nvlist_t *tgt; boolean_t avail_spare, l2cache, islog; libzfs_handle_t *hdl = zhp->zpool_hdl; if (flags & ZFS_ONLINE_EXPAND) { (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot expand %s"), path); } else { (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot online %s"), path); } (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache, &islog)) == NULL) return (zfs_error(hdl, EZFS_NODEVICE, errbuf)); zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID); if (!(flags & ZFS_ONLINE_SPARE) && avail_spare) return (zfs_error(hdl, EZFS_ISSPARE, errbuf)); #ifndef __FreeBSD__ const char *pathname; if ((flags & ZFS_ONLINE_EXPAND || zpool_get_prop_int(zhp, ZPOOL_PROP_AUTOEXPAND, NULL)) && nvlist_lookup_string(tgt, ZPOOL_CONFIG_PATH, &pathname) == 0) { uint64_t wholedisk = 0; (void) nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_WHOLE_DISK, &wholedisk); /* * XXX - L2ARC 1.0 devices can't support expansion. */ if (l2cache) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "cannot expand cache devices")); return (zfs_error(hdl, EZFS_VDEVNOTSUP, errbuf)); } if (wholedisk) { const char *fullpath = path; char buf[MAXPATHLEN]; int error; if (path[0] != '/') { error = zfs_resolve_shortname(path, buf, sizeof (buf)); if (error != 0) return (zfs_error(hdl, EZFS_NODEVICE, errbuf)); fullpath = buf; } error = zpool_relabel_disk(hdl, fullpath, errbuf); if (error != 0) return (error); } } #endif zc.zc_cookie = VDEV_STATE_ONLINE; zc.zc_obj = flags; if (zfs_ioctl(hdl, ZFS_IOC_VDEV_SET_STATE, &zc) != 0) { if (errno == EINVAL) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "was split " "from this pool into a new one. Use '%s' " "instead"), "zpool detach"); return (zfs_error(hdl, EZFS_POSTSPLIT_ONLINE, errbuf)); } return (zpool_standard_error(hdl, errno, errbuf)); } *newstate = zc.zc_cookie; return (0); } /* * Take the specified vdev offline */ int zpool_vdev_offline(zpool_handle_t *zhp, const char *path, boolean_t istmp) { zfs_cmd_t zc = {"\0"}; char errbuf[ERRBUFLEN]; nvlist_t *tgt; boolean_t avail_spare, l2cache; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot offline %s"), path); (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache, NULL)) == NULL) return (zfs_error(hdl, EZFS_NODEVICE, errbuf)); zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID); if (avail_spare) return (zfs_error(hdl, EZFS_ISSPARE, errbuf)); zc.zc_cookie = VDEV_STATE_OFFLINE; zc.zc_obj = istmp ? ZFS_OFFLINE_TEMPORARY : 0; if (zfs_ioctl(hdl, ZFS_IOC_VDEV_SET_STATE, &zc) == 0) return (0); switch (errno) { case EBUSY: /* * There are no other replicas of this device. */ return (zfs_error(hdl, EZFS_NOREPLICAS, errbuf)); case EEXIST: /* * The log device has unplayed logs */ return (zfs_error(hdl, EZFS_UNPLAYED_LOGS, errbuf)); default: return (zpool_standard_error(hdl, errno, errbuf)); } } /* * Remove the specified vdev asynchronously from the configuration, so * that it may come ONLINE if reinserted. This is called from zed on * Udev remove event. * Note: We also have a similar function zpool_vdev_remove() that * removes the vdev from the pool. */ int zpool_vdev_remove_wanted(zpool_handle_t *zhp, const char *path) { zfs_cmd_t zc = {"\0"}; char errbuf[ERRBUFLEN]; nvlist_t *tgt; boolean_t avail_spare, l2cache; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot remove %s"), path); (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache, NULL)) == NULL) return (zfs_error(hdl, EZFS_NODEVICE, errbuf)); zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID); zc.zc_cookie = VDEV_STATE_REMOVED; if (zfs_ioctl(hdl, ZFS_IOC_VDEV_SET_STATE, &zc) == 0) return (0); return (zpool_standard_error(hdl, errno, errbuf)); } /* * Mark the given vdev faulted. */ int zpool_vdev_fault(zpool_handle_t *zhp, uint64_t guid, vdev_aux_t aux) { zfs_cmd_t zc = {"\0"}; char errbuf[ERRBUFLEN]; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot fault %llu"), (u_longlong_t)guid); (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); zc.zc_guid = guid; zc.zc_cookie = VDEV_STATE_FAULTED; zc.zc_obj = aux; if (zfs_ioctl(hdl, ZFS_IOC_VDEV_SET_STATE, &zc) == 0) return (0); switch (errno) { case EBUSY: /* * There are no other replicas of this device. */ return (zfs_error(hdl, EZFS_NOREPLICAS, errbuf)); default: return (zpool_standard_error(hdl, errno, errbuf)); } } /* * Mark the given vdev degraded. */ int zpool_vdev_degrade(zpool_handle_t *zhp, uint64_t guid, vdev_aux_t aux) { zfs_cmd_t zc = {"\0"}; char errbuf[ERRBUFLEN]; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot degrade %llu"), (u_longlong_t)guid); (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); zc.zc_guid = guid; zc.zc_cookie = VDEV_STATE_DEGRADED; zc.zc_obj = aux; if (zfs_ioctl(hdl, ZFS_IOC_VDEV_SET_STATE, &zc) == 0) return (0); return (zpool_standard_error(hdl, errno, errbuf)); } /* * Returns TRUE if the given nvlist is a vdev that was originally swapped in as * a hot spare. */ static boolean_t is_replacing_spare(nvlist_t *search, nvlist_t *tgt, int which) { nvlist_t **child; uint_t c, children; if (nvlist_lookup_nvlist_array(search, ZPOOL_CONFIG_CHILDREN, &child, &children) == 0) { const char *type = fnvlist_lookup_string(search, ZPOOL_CONFIG_TYPE); if ((strcmp(type, VDEV_TYPE_SPARE) == 0 || strcmp(type, VDEV_TYPE_DRAID_SPARE) == 0) && children == 2 && child[which] == tgt) return (B_TRUE); for (c = 0; c < children; c++) if (is_replacing_spare(child[c], tgt, which)) return (B_TRUE); } return (B_FALSE); } /* * Attach new_disk (fully described by nvroot) to old_disk. * If 'replacing' is specified, the new disk will replace the old one. */ int zpool_vdev_attach(zpool_handle_t *zhp, const char *old_disk, const char *new_disk, nvlist_t *nvroot, int replacing, boolean_t rebuild) { zfs_cmd_t zc = {"\0"}; char errbuf[ERRBUFLEN]; int ret; nvlist_t *tgt; boolean_t avail_spare, l2cache, islog; uint64_t val; char *newname; nvlist_t **child; uint_t children; nvlist_t *config_root; libzfs_handle_t *hdl = zhp->zpool_hdl; if (replacing) (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot replace %s with %s"), old_disk, new_disk); else (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot attach %s to %s"), new_disk, old_disk); (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); if ((tgt = zpool_find_vdev(zhp, old_disk, &avail_spare, &l2cache, &islog)) == NULL) return (zfs_error(hdl, EZFS_NODEVICE, errbuf)); if (avail_spare) return (zfs_error(hdl, EZFS_ISSPARE, errbuf)); if (l2cache) return (zfs_error(hdl, EZFS_ISL2CACHE, errbuf)); zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID); zc.zc_cookie = replacing; zc.zc_simple = rebuild; if (rebuild && zfeature_lookup_guid("org.openzfs:device_rebuild", NULL) != 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "the loaded zfs module doesn't support device rebuilds")); return (zfs_error(hdl, EZFS_POOL_NOTSUP, errbuf)); } if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0 || children != 1) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "new device must be a single disk")); return (zfs_error(hdl, EZFS_INVALCONFIG, errbuf)); } config_root = fnvlist_lookup_nvlist(zpool_get_config(zhp, NULL), ZPOOL_CONFIG_VDEV_TREE); if ((newname = zpool_vdev_name(NULL, NULL, child[0], 0)) == NULL) return (-1); /* * If the target is a hot spare that has been swapped in, we can only * replace it with another hot spare. */ if (replacing && nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_IS_SPARE, &val) == 0 && (zpool_find_vdev(zhp, newname, &avail_spare, &l2cache, NULL) == NULL || !avail_spare) && is_replacing_spare(config_root, tgt, 1)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "can only be replaced by another hot spare")); free(newname); return (zfs_error(hdl, EZFS_BADTARGET, errbuf)); } free(newname); zcmd_write_conf_nvlist(hdl, &zc, nvroot); ret = zfs_ioctl(hdl, ZFS_IOC_VDEV_ATTACH, &zc); zcmd_free_nvlists(&zc); if (ret == 0) return (0); switch (errno) { case ENOTSUP: /* * Can't attach to or replace this type of vdev. */ if (replacing) { uint64_t version = zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL); if (islog) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "cannot replace a log with a spare")); } else if (rebuild) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "only mirror and dRAID vdevs support " "sequential reconstruction")); } else if (zpool_is_draid_spare(new_disk)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "dRAID spares can only replace child " "devices in their parent's dRAID vdev")); } else if (version >= SPA_VERSION_MULTI_REPLACE) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "already in replacing/spare config; wait " "for completion or use 'zpool detach'")); } else { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "cannot replace a replacing device")); } } else { char status[64] = {0}; zpool_prop_get_feature(zhp, "feature@device_rebuild", status, 63); if (rebuild && strncmp(status, ZFS_FEATURE_DISABLED, 64) == 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "device_rebuild feature must be enabled " "in order to use sequential " "reconstruction")); } else { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "can only attach to mirrors and top-level " "disks")); } } (void) zfs_error(hdl, EZFS_BADTARGET, errbuf); break; case EINVAL: /* * The new device must be a single disk. */ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "new device must be a single disk")); (void) zfs_error(hdl, EZFS_INVALCONFIG, errbuf); break; case EBUSY: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "%s is busy, " "or device removal is in progress"), new_disk); (void) zfs_error(hdl, EZFS_BADDEV, errbuf); break; case EOVERFLOW: /* * The new device is too small. */ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "device is too small")); (void) zfs_error(hdl, EZFS_BADDEV, errbuf); break; case EDOM: /* * The new device has a different optimal sector size. */ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "new device has a different optimal sector size; use the " "option '-o ashift=N' to override the optimal size")); (void) zfs_error(hdl, EZFS_BADDEV, errbuf); break; case ENAMETOOLONG: /* * The resulting top-level vdev spec won't fit in the label. */ (void) zfs_error(hdl, EZFS_DEVOVERFLOW, errbuf); break; default: (void) zpool_standard_error(hdl, errno, errbuf); } return (-1); } /* * Detach the specified device. */ int zpool_vdev_detach(zpool_handle_t *zhp, const char *path) { zfs_cmd_t zc = {"\0"}; char errbuf[ERRBUFLEN]; nvlist_t *tgt; boolean_t avail_spare, l2cache; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot detach %s"), path); (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache, NULL)) == NULL) return (zfs_error(hdl, EZFS_NODEVICE, errbuf)); if (avail_spare) return (zfs_error(hdl, EZFS_ISSPARE, errbuf)); if (l2cache) return (zfs_error(hdl, EZFS_ISL2CACHE, errbuf)); zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID); if (zfs_ioctl(hdl, ZFS_IOC_VDEV_DETACH, &zc) == 0) return (0); switch (errno) { case ENOTSUP: /* * Can't detach from this type of vdev. */ zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "only " "applicable to mirror and replacing vdevs")); (void) zfs_error(hdl, EZFS_BADTARGET, errbuf); break; case EBUSY: /* * There are no other replicas of this device. */ (void) zfs_error(hdl, EZFS_NOREPLICAS, errbuf); break; default: (void) zpool_standard_error(hdl, errno, errbuf); } return (-1); } /* * Find a mirror vdev in the source nvlist. * * The mchild array contains a list of disks in one of the top-level mirrors * of the source pool. The schild array contains a list of disks that the * user specified on the command line. We loop over the mchild array to * see if any entry in the schild array matches. * * If a disk in the mchild array is found in the schild array, we return * the index of that entry. Otherwise we return -1. */ static int find_vdev_entry(zpool_handle_t *zhp, nvlist_t **mchild, uint_t mchildren, nvlist_t **schild, uint_t schildren) { uint_t mc; for (mc = 0; mc < mchildren; mc++) { uint_t sc; char *mpath = zpool_vdev_name(zhp->zpool_hdl, zhp, mchild[mc], 0); for (sc = 0; sc < schildren; sc++) { char *spath = zpool_vdev_name(zhp->zpool_hdl, zhp, schild[sc], 0); boolean_t result = (strcmp(mpath, spath) == 0); free(spath); if (result) { free(mpath); return (mc); } } free(mpath); } return (-1); } /* * Split a mirror pool. If newroot points to null, then a new nvlist * is generated and it is the responsibility of the caller to free it. */ int zpool_vdev_split(zpool_handle_t *zhp, char *newname, nvlist_t **newroot, nvlist_t *props, splitflags_t flags) { zfs_cmd_t zc = {"\0"}; char errbuf[ERRBUFLEN]; const char *bias; nvlist_t *tree, *config, **child, **newchild, *newconfig = NULL; nvlist_t **varray = NULL, *zc_props = NULL; uint_t c, children, newchildren, lastlog = 0, vcount, found = 0; libzfs_handle_t *hdl = zhp->zpool_hdl; uint64_t vers, readonly = B_FALSE; boolean_t freelist = B_FALSE, memory_err = B_TRUE; int retval = 0; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "Unable to split %s"), zhp->zpool_name); if (!zpool_name_valid(hdl, B_FALSE, newname)) return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); if ((config = zpool_get_config(zhp, NULL)) == NULL) { (void) fprintf(stderr, gettext("Internal error: unable to " "retrieve pool configuration\n")); return (-1); } tree = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE); vers = fnvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION); if (props) { prop_flags_t flags = { .create = B_FALSE, .import = B_TRUE }; if ((zc_props = zpool_valid_proplist(hdl, zhp->zpool_name, props, vers, flags, errbuf)) == NULL) return (-1); (void) nvlist_lookup_uint64(zc_props, zpool_prop_to_name(ZPOOL_PROP_READONLY), &readonly); if (readonly) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "property %s can only be set at import time"), zpool_prop_to_name(ZPOOL_PROP_READONLY)); return (-1); } } if (nvlist_lookup_nvlist_array(tree, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "Source pool is missing vdev tree")); nvlist_free(zc_props); return (-1); } varray = zfs_alloc(hdl, children * sizeof (nvlist_t *)); vcount = 0; if (*newroot == NULL || nvlist_lookup_nvlist_array(*newroot, ZPOOL_CONFIG_CHILDREN, &newchild, &newchildren) != 0) newchildren = 0; for (c = 0; c < children; c++) { uint64_t is_log = B_FALSE, is_hole = B_FALSE; boolean_t is_special = B_FALSE, is_dedup = B_FALSE; const char *type; nvlist_t **mchild, *vdev; uint_t mchildren; int entry; /* * Unlike cache & spares, slogs are stored in the * ZPOOL_CONFIG_CHILDREN array. We filter them out here. */ (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, &is_log); (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_HOLE, &is_hole); if (is_log || is_hole) { /* * Create a hole vdev and put it in the config. */ if (nvlist_alloc(&vdev, NV_UNIQUE_NAME, 0) != 0) goto out; if (nvlist_add_string(vdev, ZPOOL_CONFIG_TYPE, VDEV_TYPE_HOLE) != 0) goto out; if (nvlist_add_uint64(vdev, ZPOOL_CONFIG_IS_HOLE, 1) != 0) goto out; if (lastlog == 0) lastlog = vcount; varray[vcount++] = vdev; continue; } lastlog = 0; type = fnvlist_lookup_string(child[c], ZPOOL_CONFIG_TYPE); if (strcmp(type, VDEV_TYPE_INDIRECT) == 0) { vdev = child[c]; if (nvlist_dup(vdev, &varray[vcount++], 0) != 0) goto out; continue; } else if (strcmp(type, VDEV_TYPE_MIRROR) != 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "Source pool must be composed only of mirrors\n")); retval = zfs_error(hdl, EZFS_INVALCONFIG, errbuf); goto out; } if (nvlist_lookup_string(child[c], ZPOOL_CONFIG_ALLOCATION_BIAS, &bias) == 0) { if (strcmp(bias, VDEV_ALLOC_BIAS_SPECIAL) == 0) is_special = B_TRUE; else if (strcmp(bias, VDEV_ALLOC_BIAS_DEDUP) == 0) is_dedup = B_TRUE; } verify(nvlist_lookup_nvlist_array(child[c], ZPOOL_CONFIG_CHILDREN, &mchild, &mchildren) == 0); /* find or add an entry for this top-level vdev */ if (newchildren > 0 && (entry = find_vdev_entry(zhp, mchild, mchildren, newchild, newchildren)) >= 0) { /* We found a disk that the user specified. */ vdev = mchild[entry]; ++found; } else { /* User didn't specify a disk for this vdev. */ vdev = mchild[mchildren - 1]; } if (nvlist_dup(vdev, &varray[vcount++], 0) != 0) goto out; if (flags.dryrun != 0) { if (is_dedup == B_TRUE) { if (nvlist_add_string(varray[vcount - 1], ZPOOL_CONFIG_ALLOCATION_BIAS, VDEV_ALLOC_BIAS_DEDUP) != 0) goto out; } else if (is_special == B_TRUE) { if (nvlist_add_string(varray[vcount - 1], ZPOOL_CONFIG_ALLOCATION_BIAS, VDEV_ALLOC_BIAS_SPECIAL) != 0) goto out; } } } /* did we find every disk the user specified? */ if (found != newchildren) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "Device list must " "include at most one disk from each mirror")); retval = zfs_error(hdl, EZFS_INVALCONFIG, errbuf); goto out; } /* Prepare the nvlist for populating. */ if (*newroot == NULL) { if (nvlist_alloc(newroot, NV_UNIQUE_NAME, 0) != 0) goto out; freelist = B_TRUE; if (nvlist_add_string(*newroot, ZPOOL_CONFIG_TYPE, VDEV_TYPE_ROOT) != 0) goto out; } else { verify(nvlist_remove_all(*newroot, ZPOOL_CONFIG_CHILDREN) == 0); } /* Add all the children we found */ if (nvlist_add_nvlist_array(*newroot, ZPOOL_CONFIG_CHILDREN, (const nvlist_t **)varray, lastlog == 0 ? vcount : lastlog) != 0) goto out; /* * If we're just doing a dry run, exit now with success. */ if (flags.dryrun) { memory_err = B_FALSE; freelist = B_FALSE; goto out; } /* now build up the config list & call the ioctl */ if (nvlist_alloc(&newconfig, NV_UNIQUE_NAME, 0) != 0) goto out; if (nvlist_add_nvlist(newconfig, ZPOOL_CONFIG_VDEV_TREE, *newroot) != 0 || nvlist_add_string(newconfig, ZPOOL_CONFIG_POOL_NAME, newname) != 0 || nvlist_add_uint64(newconfig, ZPOOL_CONFIG_VERSION, vers) != 0) goto out; /* * The new pool is automatically part of the namespace unless we * explicitly export it. */ if (!flags.import) zc.zc_cookie = ZPOOL_EXPORT_AFTER_SPLIT; (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); (void) strlcpy(zc.zc_string, newname, sizeof (zc.zc_string)); zcmd_write_conf_nvlist(hdl, &zc, newconfig); if (zc_props != NULL) zcmd_write_src_nvlist(hdl, &zc, zc_props); if (zfs_ioctl(hdl, ZFS_IOC_VDEV_SPLIT, &zc) != 0) { retval = zpool_standard_error(hdl, errno, errbuf); goto out; } freelist = B_FALSE; memory_err = B_FALSE; out: if (varray != NULL) { int v; for (v = 0; v < vcount; v++) nvlist_free(varray[v]); free(varray); } zcmd_free_nvlists(&zc); nvlist_free(zc_props); nvlist_free(newconfig); if (freelist) { nvlist_free(*newroot); *newroot = NULL; } if (retval != 0) return (retval); if (memory_err) return (no_memory(hdl)); return (0); } /* * Remove the given device. */ int zpool_vdev_remove(zpool_handle_t *zhp, const char *path) { zfs_cmd_t zc = {"\0"}; char errbuf[ERRBUFLEN]; nvlist_t *tgt; boolean_t avail_spare, l2cache, islog; libzfs_handle_t *hdl = zhp->zpool_hdl; uint64_t version; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot remove %s"), path); if (zpool_is_draid_spare(path)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "dRAID spares cannot be removed")); return (zfs_error(hdl, EZFS_NODEVICE, errbuf)); } (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache, &islog)) == NULL) return (zfs_error(hdl, EZFS_NODEVICE, errbuf)); version = zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL); if (islog && version < SPA_VERSION_HOLES) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool must be upgraded to support log removal")); return (zfs_error(hdl, EZFS_BADVERSION, errbuf)); } zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID); if (zfs_ioctl(hdl, ZFS_IOC_VDEV_REMOVE, &zc) == 0) return (0); switch (errno) { case EINVAL: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid config; all top-level vdevs must " "have the same sector size and not be raidz.")); (void) zfs_error(hdl, EZFS_INVALCONFIG, errbuf); break; case EBUSY: if (islog) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "Mount encrypted datasets to replay logs.")); } else { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "Pool busy; removal may already be in progress")); } (void) zfs_error(hdl, EZFS_BUSY, errbuf); break; case EACCES: if (islog) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "Mount encrypted datasets to replay logs.")); (void) zfs_error(hdl, EZFS_BUSY, errbuf); } else { (void) zpool_standard_error(hdl, errno, errbuf); } break; default: (void) zpool_standard_error(hdl, errno, errbuf); } return (-1); } int zpool_vdev_remove_cancel(zpool_handle_t *zhp) { zfs_cmd_t zc = {{0}}; char errbuf[ERRBUFLEN]; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot cancel removal")); (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); zc.zc_cookie = 1; if (zfs_ioctl(hdl, ZFS_IOC_VDEV_REMOVE, &zc) == 0) return (0); return (zpool_standard_error(hdl, errno, errbuf)); } int zpool_vdev_indirect_size(zpool_handle_t *zhp, const char *path, uint64_t *sizep) { char errbuf[ERRBUFLEN]; nvlist_t *tgt; boolean_t avail_spare, l2cache, islog; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot determine indirect size of %s"), path); if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache, &islog)) == NULL) return (zfs_error(hdl, EZFS_NODEVICE, errbuf)); if (avail_spare || l2cache || islog) { *sizep = 0; return (0); } if (nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_INDIRECT_SIZE, sizep) != 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "indirect size not available")); return (zfs_error(hdl, EINVAL, errbuf)); } return (0); } /* * Clear the errors for the pool, or the particular device if specified. */ int zpool_clear(zpool_handle_t *zhp, const char *path, nvlist_t *rewindnvl) { zfs_cmd_t zc = {"\0"}; char errbuf[ERRBUFLEN]; nvlist_t *tgt; zpool_load_policy_t policy; boolean_t avail_spare, l2cache; libzfs_handle_t *hdl = zhp->zpool_hdl; nvlist_t *nvi = NULL; int error; if (path) (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot clear errors for %s"), path); else (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot clear errors for %s"), zhp->zpool_name); (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); if (path) { if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache, NULL)) == NULL) return (zfs_error(hdl, EZFS_NODEVICE, errbuf)); /* * Don't allow error clearing for hot spares. Do allow * error clearing for l2cache devices. */ if (avail_spare) return (zfs_error(hdl, EZFS_ISSPARE, errbuf)); zc.zc_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID); } zpool_get_load_policy(rewindnvl, &policy); zc.zc_cookie = policy.zlp_rewind; zcmd_alloc_dst_nvlist(hdl, &zc, zhp->zpool_config_size * 2); zcmd_write_src_nvlist(hdl, &zc, rewindnvl); while ((error = zfs_ioctl(hdl, ZFS_IOC_CLEAR, &zc)) != 0 && errno == ENOMEM) zcmd_expand_dst_nvlist(hdl, &zc); if (!error || ((policy.zlp_rewind & ZPOOL_TRY_REWIND) && errno != EPERM && errno != EACCES)) { if (policy.zlp_rewind & (ZPOOL_DO_REWIND | ZPOOL_TRY_REWIND)) { (void) zcmd_read_dst_nvlist(hdl, &zc, &nvi); zpool_rewind_exclaim(hdl, zc.zc_name, ((policy.zlp_rewind & ZPOOL_TRY_REWIND) != 0), nvi); nvlist_free(nvi); } zcmd_free_nvlists(&zc); return (0); } zcmd_free_nvlists(&zc); return (zpool_standard_error(hdl, errno, errbuf)); } /* * Similar to zpool_clear(), but takes a GUID (used by fmd). */ int zpool_vdev_clear(zpool_handle_t *zhp, uint64_t guid) { zfs_cmd_t zc = {"\0"}; char errbuf[ERRBUFLEN]; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot clear errors for %llx"), (u_longlong_t)guid); (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); zc.zc_guid = guid; zc.zc_cookie = ZPOOL_NO_REWIND; if (zfs_ioctl(hdl, ZFS_IOC_CLEAR, &zc) == 0) return (0); return (zpool_standard_error(hdl, errno, errbuf)); } /* * Change the GUID for a pool. */ int zpool_reguid(zpool_handle_t *zhp) { char errbuf[ERRBUFLEN]; libzfs_handle_t *hdl = zhp->zpool_hdl; zfs_cmd_t zc = {"\0"}; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot reguid '%s'"), zhp->zpool_name); (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); if (zfs_ioctl(hdl, ZFS_IOC_POOL_REGUID, &zc) == 0) return (0); return (zpool_standard_error(hdl, errno, errbuf)); } /* * Reopen the pool. */ int zpool_reopen_one(zpool_handle_t *zhp, void *data) { libzfs_handle_t *hdl = zpool_get_handle(zhp); const char *pool_name = zpool_get_name(zhp); boolean_t *scrub_restart = data; int error; error = lzc_reopen(pool_name, *scrub_restart); if (error) { return (zpool_standard_error_fmt(hdl, error, dgettext(TEXT_DOMAIN, "cannot reopen '%s'"), pool_name)); } return (0); } /* call into libzfs_core to execute the sync IOCTL per pool */ int zpool_sync_one(zpool_handle_t *zhp, void *data) { int ret; libzfs_handle_t *hdl = zpool_get_handle(zhp); const char *pool_name = zpool_get_name(zhp); boolean_t *force = data; nvlist_t *innvl = fnvlist_alloc(); fnvlist_add_boolean_value(innvl, "force", *force); if ((ret = lzc_sync(pool_name, innvl, NULL)) != 0) { nvlist_free(innvl); return (zpool_standard_error_fmt(hdl, ret, dgettext(TEXT_DOMAIN, "sync '%s' failed"), pool_name)); } nvlist_free(innvl); return (0); } #define PATH_BUF_LEN 64 /* * Given a vdev, return the name to display in iostat. If the vdev has a path, * we use that, stripping off any leading "/dev/dsk/"; if not, we use the type. * We also check if this is a whole disk, in which case we strip off the * trailing 's0' slice name. * * This routine is also responsible for identifying when disks have been * reconfigured in a new location. The kernel will have opened the device by * devid, but the path will still refer to the old location. To catch this, we * first do a path -> devid translation (which is fast for the common case). If * the devid matches, we're done. If not, we do a reverse devid -> path * translation and issue the appropriate ioctl() to update the path of the vdev. * If 'zhp' is NULL, then this is an exported pool, and we don't need to do any * of these checks. */ char * zpool_vdev_name(libzfs_handle_t *hdl, zpool_handle_t *zhp, nvlist_t *nv, int name_flags) { const char *type, *tpath; const char *path; uint64_t value; char buf[PATH_BUF_LEN]; char tmpbuf[PATH_BUF_LEN * 2]; /* * vdev_name will be "root"/"root-0" for the root vdev, but it is the * zpool name that will be displayed to the user. */ type = fnvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE); if (zhp != NULL && strcmp(type, "root") == 0) return (zfs_strdup(hdl, zpool_get_name(zhp))); if (libzfs_envvar_is_set("ZPOOL_VDEV_NAME_PATH")) name_flags |= VDEV_NAME_PATH; if (libzfs_envvar_is_set("ZPOOL_VDEV_NAME_GUID")) name_flags |= VDEV_NAME_GUID; if (libzfs_envvar_is_set("ZPOOL_VDEV_NAME_FOLLOW_LINKS")) name_flags |= VDEV_NAME_FOLLOW_LINKS; if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT, &value) == 0 || name_flags & VDEV_NAME_GUID) { (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &value); (void) snprintf(buf, sizeof (buf), "%llu", (u_longlong_t)value); path = buf; } else if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &tpath) == 0) { path = tpath; if (name_flags & VDEV_NAME_FOLLOW_LINKS) { char *rp = realpath(path, NULL); if (rp) { strlcpy(buf, rp, sizeof (buf)); path = buf; free(rp); } } /* * For a block device only use the name. */ if ((strcmp(type, VDEV_TYPE_DISK) == 0) && !(name_flags & VDEV_NAME_PATH)) { path = zfs_strip_path(path); } /* * Remove the partition from the path if this is a whole disk. */ if (strcmp(type, VDEV_TYPE_DRAID_SPARE) != 0 && nvlist_lookup_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK, &value) == 0 && value && !(name_flags & VDEV_NAME_PATH)) { return (zfs_strip_partition(path)); } } else { path = type; /* * If it's a raidz device, we need to stick in the parity level. */ if (strcmp(path, VDEV_TYPE_RAIDZ) == 0) { value = fnvlist_lookup_uint64(nv, ZPOOL_CONFIG_NPARITY); (void) snprintf(buf, sizeof (buf), "%s%llu", path, (u_longlong_t)value); path = buf; } /* * If it's a dRAID device, we add parity, groups, and spares. */ if (strcmp(path, VDEV_TYPE_DRAID) == 0) { uint64_t ndata, nparity, nspares; nvlist_t **child; uint_t children; verify(nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) == 0); nparity = fnvlist_lookup_uint64(nv, ZPOOL_CONFIG_NPARITY); ndata = fnvlist_lookup_uint64(nv, ZPOOL_CONFIG_DRAID_NDATA); nspares = fnvlist_lookup_uint64(nv, ZPOOL_CONFIG_DRAID_NSPARES); path = zpool_draid_name(buf, sizeof (buf), ndata, nparity, nspares, children); } /* * We identify each top-level vdev by using a * naming convention. */ if (name_flags & VDEV_NAME_TYPE_ID) { uint64_t id = fnvlist_lookup_uint64(nv, ZPOOL_CONFIG_ID); (void) snprintf(tmpbuf, sizeof (tmpbuf), "%s-%llu", path, (u_longlong_t)id); path = tmpbuf; } } return (zfs_strdup(hdl, path)); } static int zbookmark_mem_compare(const void *a, const void *b) { return (memcmp(a, b, sizeof (zbookmark_phys_t))); } /* * Retrieve the persistent error log, uniquify the members, and return to the * caller. */ int zpool_get_errlog(zpool_handle_t *zhp, nvlist_t **nverrlistp) { zfs_cmd_t zc = {"\0"}; libzfs_handle_t *hdl = zhp->zpool_hdl; zbookmark_phys_t *buf; uint64_t buflen = 10000; /* approx. 1MB of RAM */ if (fnvlist_lookup_uint64(zhp->zpool_config, ZPOOL_CONFIG_ERRCOUNT) == 0) return (0); /* * Retrieve the raw error list from the kernel. If it doesn't fit, * allocate a larger buffer and retry. */ (void) strcpy(zc.zc_name, zhp->zpool_name); for (;;) { buf = zfs_alloc(zhp->zpool_hdl, buflen * sizeof (zbookmark_phys_t)); zc.zc_nvlist_dst = (uintptr_t)buf; zc.zc_nvlist_dst_size = buflen; if (zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_ERROR_LOG, &zc) != 0) { free(buf); if (errno == ENOMEM) { buflen *= 2; } else { return (zpool_standard_error_fmt(hdl, errno, dgettext(TEXT_DOMAIN, "errors: List of " "errors unavailable"))); } } else { break; } } /* * Sort the resulting bookmarks. This is a little confusing due to the * implementation of ZFS_IOC_ERROR_LOG. The bookmarks are copied last * to first, and 'zc_nvlist_dst_size' indicates the number of bookmarks * _not_ copied as part of the process. So we point the start of our * array appropriate and decrement the total number of elements. */ zbookmark_phys_t *zb = buf + zc.zc_nvlist_dst_size; uint64_t zblen = buflen - zc.zc_nvlist_dst_size; qsort(zb, zblen, sizeof (zbookmark_phys_t), zbookmark_mem_compare); verify(nvlist_alloc(nverrlistp, 0, KM_SLEEP) == 0); /* * Fill in the nverrlistp with nvlist's of dataset and object numbers. */ for (uint64_t i = 0; i < zblen; i++) { nvlist_t *nv; /* ignoring zb_blkid and zb_level for now */ if (i > 0 && zb[i-1].zb_objset == zb[i].zb_objset && zb[i-1].zb_object == zb[i].zb_object) continue; if (nvlist_alloc(&nv, NV_UNIQUE_NAME, KM_SLEEP) != 0) goto nomem; if (nvlist_add_uint64(nv, ZPOOL_ERR_DATASET, zb[i].zb_objset) != 0) { nvlist_free(nv); goto nomem; } if (nvlist_add_uint64(nv, ZPOOL_ERR_OBJECT, zb[i].zb_object) != 0) { nvlist_free(nv); goto nomem; } if (nvlist_add_nvlist(*nverrlistp, "ejk", nv) != 0) { nvlist_free(nv); goto nomem; } nvlist_free(nv); } free(buf); return (0); nomem: free(buf); return (no_memory(zhp->zpool_hdl)); } /* * Upgrade a ZFS pool to the latest on-disk version. */ int zpool_upgrade(zpool_handle_t *zhp, uint64_t new_version) { zfs_cmd_t zc = {"\0"}; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) strcpy(zc.zc_name, zhp->zpool_name); zc.zc_cookie = new_version; if (zfs_ioctl(hdl, ZFS_IOC_POOL_UPGRADE, &zc) != 0) return (zpool_standard_error_fmt(hdl, errno, dgettext(TEXT_DOMAIN, "cannot upgrade '%s'"), zhp->zpool_name)); return (0); } void zfs_save_arguments(int argc, char **argv, char *string, int len) { int i; (void) strlcpy(string, zfs_basename(argv[0]), len); for (i = 1; i < argc; i++) { (void) strlcat(string, " ", len); (void) strlcat(string, argv[i], len); } } int zpool_log_history(libzfs_handle_t *hdl, const char *message) { zfs_cmd_t zc = {"\0"}; nvlist_t *args; args = fnvlist_alloc(); fnvlist_add_string(args, "message", message); zcmd_write_src_nvlist(hdl, &zc, args); int err = zfs_ioctl(hdl, ZFS_IOC_LOG_HISTORY, &zc); nvlist_free(args); zcmd_free_nvlists(&zc); return (err); } /* * Perform ioctl to get some command history of a pool. * * 'buf' is the buffer to fill up to 'len' bytes. 'off' is the * logical offset of the history buffer to start reading from. * * Upon return, 'off' is the next logical offset to read from and * 'len' is the actual amount of bytes read into 'buf'. */ static int get_history(zpool_handle_t *zhp, char *buf, uint64_t *off, uint64_t *len) { zfs_cmd_t zc = {"\0"}; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); zc.zc_history = (uint64_t)(uintptr_t)buf; zc.zc_history_len = *len; zc.zc_history_offset = *off; if (zfs_ioctl(hdl, ZFS_IOC_POOL_GET_HISTORY, &zc) != 0) { switch (errno) { case EPERM: return (zfs_error_fmt(hdl, EZFS_PERM, dgettext(TEXT_DOMAIN, "cannot show history for pool '%s'"), zhp->zpool_name)); case ENOENT: return (zfs_error_fmt(hdl, EZFS_NOHISTORY, dgettext(TEXT_DOMAIN, "cannot get history for pool " "'%s'"), zhp->zpool_name)); case ENOTSUP: return (zfs_error_fmt(hdl, EZFS_BADVERSION, dgettext(TEXT_DOMAIN, "cannot get history for pool " "'%s', pool must be upgraded"), zhp->zpool_name)); default: return (zpool_standard_error_fmt(hdl, errno, dgettext(TEXT_DOMAIN, "cannot get history for '%s'"), zhp->zpool_name)); } } *len = zc.zc_history_len; *off = zc.zc_history_offset; return (0); } /* * Retrieve the command history of a pool. */ int zpool_get_history(zpool_handle_t *zhp, nvlist_t **nvhisp, uint64_t *off, boolean_t *eof) { libzfs_handle_t *hdl = zhp->zpool_hdl; char *buf; int buflen = 128 * 1024; nvlist_t **records = NULL; uint_t numrecords = 0; int err = 0, i; uint64_t start = *off; buf = zfs_alloc(hdl, buflen); /* process about 1MiB a time */ while (*off - start < 1024 * 1024) { uint64_t bytes_read = buflen; uint64_t leftover; if ((err = get_history(zhp, buf, off, &bytes_read)) != 0) break; /* if nothing else was read in, we're at EOF, just return */ if (!bytes_read) { *eof = B_TRUE; break; } if ((err = zpool_history_unpack(buf, bytes_read, &leftover, &records, &numrecords)) != 0) { zpool_standard_error_fmt(hdl, err, dgettext(TEXT_DOMAIN, "cannot get history for '%s'"), zhp->zpool_name); break; } *off -= leftover; if (leftover == bytes_read) { /* * no progress made, because buffer is not big enough * to hold this record; resize and retry. */ buflen *= 2; free(buf); buf = zfs_alloc(hdl, buflen); } } free(buf); if (!err) { *nvhisp = fnvlist_alloc(); fnvlist_add_nvlist_array(*nvhisp, ZPOOL_HIST_RECORD, (const nvlist_t **)records, numrecords); } for (i = 0; i < numrecords; i++) nvlist_free(records[i]); free(records); return (err); } /* * Retrieve the next event given the passed 'zevent_fd' file descriptor. * If there is a new event available 'nvp' will contain a newly allocated * nvlist and 'dropped' will be set to the number of missed events since * the last call to this function. When 'nvp' is set to NULL it indicates * no new events are available. In either case the function returns 0 and * it is up to the caller to free 'nvp'. In the case of a fatal error the * function will return a non-zero value. When the function is called in * blocking mode (the default, unless the ZEVENT_NONBLOCK flag is passed), * it will not return until a new event is available. */ int zpool_events_next(libzfs_handle_t *hdl, nvlist_t **nvp, int *dropped, unsigned flags, int zevent_fd) { zfs_cmd_t zc = {"\0"}; int error = 0; *nvp = NULL; *dropped = 0; zc.zc_cleanup_fd = zevent_fd; if (flags & ZEVENT_NONBLOCK) zc.zc_guid = ZEVENT_NONBLOCK; zcmd_alloc_dst_nvlist(hdl, &zc, ZEVENT_SIZE); retry: if (zfs_ioctl(hdl, ZFS_IOC_EVENTS_NEXT, &zc) != 0) { switch (errno) { case ESHUTDOWN: error = zfs_error_fmt(hdl, EZFS_POOLUNAVAIL, dgettext(TEXT_DOMAIN, "zfs shutdown")); goto out; case ENOENT: /* Blocking error case should not occur */ if (!(flags & ZEVENT_NONBLOCK)) error = zpool_standard_error_fmt(hdl, errno, dgettext(TEXT_DOMAIN, "cannot get event")); goto out; case ENOMEM: zcmd_expand_dst_nvlist(hdl, &zc); goto retry; default: error = zpool_standard_error_fmt(hdl, errno, dgettext(TEXT_DOMAIN, "cannot get event")); goto out; } } error = zcmd_read_dst_nvlist(hdl, &zc, nvp); if (error != 0) goto out; *dropped = (int)zc.zc_cookie; out: zcmd_free_nvlists(&zc); return (error); } /* * Clear all events. */ int zpool_events_clear(libzfs_handle_t *hdl, int *count) { zfs_cmd_t zc = {"\0"}; if (zfs_ioctl(hdl, ZFS_IOC_EVENTS_CLEAR, &zc) != 0) return (zpool_standard_error(hdl, errno, dgettext(TEXT_DOMAIN, "cannot clear events"))); if (count != NULL) *count = (int)zc.zc_cookie; /* # of events cleared */ return (0); } /* * Seek to a specific EID, ZEVENT_SEEK_START, or ZEVENT_SEEK_END for * the passed zevent_fd file handle. On success zero is returned, * otherwise -1 is returned and hdl->libzfs_error is set to the errno. */ int zpool_events_seek(libzfs_handle_t *hdl, uint64_t eid, int zevent_fd) { zfs_cmd_t zc = {"\0"}; int error = 0; zc.zc_guid = eid; zc.zc_cleanup_fd = zevent_fd; if (zfs_ioctl(hdl, ZFS_IOC_EVENTS_SEEK, &zc) != 0) { switch (errno) { case ENOENT: error = zfs_error_fmt(hdl, EZFS_NOENT, dgettext(TEXT_DOMAIN, "cannot get event")); break; case ENOMEM: error = zfs_error_fmt(hdl, EZFS_NOMEM, dgettext(TEXT_DOMAIN, "cannot get event")); break; default: error = zpool_standard_error_fmt(hdl, errno, dgettext(TEXT_DOMAIN, "cannot get event")); break; } } return (error); } static void zpool_obj_to_path_impl(zpool_handle_t *zhp, uint64_t dsobj, uint64_t obj, char *pathname, size_t len, boolean_t always_unmounted) { zfs_cmd_t zc = {"\0"}; boolean_t mounted = B_FALSE; char *mntpnt = NULL; char dsname[ZFS_MAX_DATASET_NAME_LEN]; if (dsobj == 0) { /* special case for the MOS */ (void) snprintf(pathname, len, ":<0x%llx>", (longlong_t)obj); return; } /* get the dataset's name */ (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); zc.zc_obj = dsobj; if (zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_DSOBJ_TO_DSNAME, &zc) != 0) { /* just write out a path of two object numbers */ (void) snprintf(pathname, len, "<0x%llx>:<0x%llx>", (longlong_t)dsobj, (longlong_t)obj); return; } (void) strlcpy(dsname, zc.zc_value, sizeof (dsname)); /* find out if the dataset is mounted */ mounted = !always_unmounted && is_mounted(zhp->zpool_hdl, dsname, &mntpnt); /* get the corrupted object's path */ (void) strlcpy(zc.zc_name, dsname, sizeof (zc.zc_name)); zc.zc_obj = obj; if (zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_OBJ_TO_PATH, &zc) == 0) { if (mounted) { (void) snprintf(pathname, len, "%s%s", mntpnt, zc.zc_value); } else { (void) snprintf(pathname, len, "%s:%s", dsname, zc.zc_value); } } else { (void) snprintf(pathname, len, "%s:<0x%llx>", dsname, (longlong_t)obj); } free(mntpnt); } void zpool_obj_to_path(zpool_handle_t *zhp, uint64_t dsobj, uint64_t obj, char *pathname, size_t len) { zpool_obj_to_path_impl(zhp, dsobj, obj, pathname, len, B_FALSE); } void zpool_obj_to_path_ds(zpool_handle_t *zhp, uint64_t dsobj, uint64_t obj, char *pathname, size_t len) { zpool_obj_to_path_impl(zhp, dsobj, obj, pathname, len, B_TRUE); } /* * Wait while the specified activity is in progress in the pool. */ int zpool_wait(zpool_handle_t *zhp, zpool_wait_activity_t activity) { boolean_t missing; int error = zpool_wait_status(zhp, activity, &missing, NULL); if (missing) { (void) zpool_standard_error_fmt(zhp->zpool_hdl, ENOENT, dgettext(TEXT_DOMAIN, "error waiting in pool '%s'"), zhp->zpool_name); return (ENOENT); } else { return (error); } } /* * Wait for the given activity and return the status of the wait (whether or not * any waiting was done) in the 'waited' parameter. Non-existent pools are * reported via the 'missing' parameter, rather than by printing an error * message. This is convenient when this function is called in a loop over a * long period of time (as it is, for example, by zpool's wait cmd). In that * scenario, a pool being exported or destroyed should be considered a normal * event, so we don't want to print an error when we find that the pool doesn't * exist. */ int zpool_wait_status(zpool_handle_t *zhp, zpool_wait_activity_t activity, boolean_t *missing, boolean_t *waited) { int error = lzc_wait(zhp->zpool_name, activity, waited); *missing = (error == ENOENT); if (*missing) return (0); if (error != 0) { (void) zpool_standard_error_fmt(zhp->zpool_hdl, error, dgettext(TEXT_DOMAIN, "error waiting in pool '%s'"), zhp->zpool_name); } return (error); } int zpool_set_bootenv(zpool_handle_t *zhp, const nvlist_t *envmap) { int error = lzc_set_bootenv(zhp->zpool_name, envmap); if (error != 0) { (void) zpool_standard_error_fmt(zhp->zpool_hdl, error, dgettext(TEXT_DOMAIN, "error setting bootenv in pool '%s'"), zhp->zpool_name); } return (error); } int zpool_get_bootenv(zpool_handle_t *zhp, nvlist_t **nvlp) { nvlist_t *nvl; int error; nvl = NULL; error = lzc_get_bootenv(zhp->zpool_name, &nvl); if (error != 0) { (void) zpool_standard_error_fmt(zhp->zpool_hdl, error, dgettext(TEXT_DOMAIN, "error getting bootenv in pool '%s'"), zhp->zpool_name); } else { *nvlp = nvl; } return (error); } /* * Attempt to read and parse feature file(s) (from "compatibility" property). * Files contain zpool feature names, comma or whitespace-separated. * Comments (# character to next newline) are discarded. * * Arguments: * compatibility : string containing feature filenames * features : either NULL or pointer to array of boolean * report : either NULL or pointer to string buffer * rlen : length of "report" buffer * * compatibility is NULL (unset), "", "off", "legacy", or list of * comma-separated filenames. filenames should either be absolute, * or relative to: * 1) ZPOOL_SYSCONF_COMPAT_D (eg: /etc/zfs/compatibility.d) or * 2) ZPOOL_DATA_COMPAT_D (eg: /usr/share/zfs/compatibility.d). * (Unset), "" or "off" => enable all features * "legacy" => disable all features * * Any feature names read from files which match unames in spa_feature_table * will have the corresponding boolean set in the features array (if non-NULL). * If more than one feature set specified, only features present in *all* of * them will be set. * * "report" if not NULL will be populated with a suitable status message. * * Return values: * ZPOOL_COMPATIBILITY_OK : files read and parsed ok * ZPOOL_COMPATIBILITY_BADFILE : file too big or not a text file * ZPOOL_COMPATIBILITY_BADTOKEN : SYSCONF file contains invalid feature name * ZPOOL_COMPATIBILITY_WARNTOKEN : DATA file contains invalid feature name * ZPOOL_COMPATIBILITY_NOFILES : no feature files found */ zpool_compat_status_t zpool_load_compat(const char *compat, boolean_t *features, char *report, size_t rlen) { int sdirfd, ddirfd, featfd; struct stat fs; char *fc; char *ps, *ls, *ws; char *file, *line, *word; char l_compat[ZFS_MAXPROPLEN]; boolean_t ret_nofiles = B_TRUE; boolean_t ret_badfile = B_FALSE; boolean_t ret_badtoken = B_FALSE; boolean_t ret_warntoken = B_FALSE; /* special cases (unset), "" and "off" => enable all features */ if (compat == NULL || compat[0] == '\0' || strcmp(compat, ZPOOL_COMPAT_OFF) == 0) { if (features != NULL) for (uint_t i = 0; i < SPA_FEATURES; i++) features[i] = B_TRUE; if (report != NULL) strlcpy(report, gettext("all features enabled"), rlen); return (ZPOOL_COMPATIBILITY_OK); } /* Final special case "legacy" => disable all features */ if (strcmp(compat, ZPOOL_COMPAT_LEGACY) == 0) { if (features != NULL) for (uint_t i = 0; i < SPA_FEATURES; i++) features[i] = B_FALSE; if (report != NULL) strlcpy(report, gettext("all features disabled"), rlen); return (ZPOOL_COMPATIBILITY_OK); } /* * Start with all true; will be ANDed with results from each file */ if (features != NULL) for (uint_t i = 0; i < SPA_FEATURES; i++) features[i] = B_TRUE; char err_badfile[ZFS_MAXPROPLEN] = ""; char err_badtoken[ZFS_MAXPROPLEN] = ""; /* * We ignore errors from the directory open() * as they're only needed if the filename is relative * which will be checked during the openat(). */ /* O_PATH safer than O_RDONLY if system allows it */ #if defined(O_PATH) #define ZC_DIR_FLAGS (O_DIRECTORY | O_CLOEXEC | O_PATH) #else #define ZC_DIR_FLAGS (O_DIRECTORY | O_CLOEXEC | O_RDONLY) #endif sdirfd = open(ZPOOL_SYSCONF_COMPAT_D, ZC_DIR_FLAGS); ddirfd = open(ZPOOL_DATA_COMPAT_D, ZC_DIR_FLAGS); (void) strlcpy(l_compat, compat, ZFS_MAXPROPLEN); for (file = strtok_r(l_compat, ",", &ps); file != NULL; file = strtok_r(NULL, ",", &ps)) { boolean_t l_features[SPA_FEATURES]; enum { Z_SYSCONF, Z_DATA } source; /* try sysconfdir first, then datadir */ source = Z_SYSCONF; if ((featfd = openat(sdirfd, file, O_RDONLY | O_CLOEXEC)) < 0) { featfd = openat(ddirfd, file, O_RDONLY | O_CLOEXEC); source = Z_DATA; } /* File readable and correct size? */ if (featfd < 0 || fstat(featfd, &fs) < 0 || fs.st_size < 1 || fs.st_size > ZPOOL_COMPAT_MAXSIZE) { (void) close(featfd); strlcat(err_badfile, file, ZFS_MAXPROPLEN); strlcat(err_badfile, " ", ZFS_MAXPROPLEN); ret_badfile = B_TRUE; continue; } /* Prefault the file if system allows */ #if defined(MAP_POPULATE) #define ZC_MMAP_FLAGS (MAP_PRIVATE | MAP_POPULATE) #elif defined(MAP_PREFAULT_READ) #define ZC_MMAP_FLAGS (MAP_PRIVATE | MAP_PREFAULT_READ) #else #define ZC_MMAP_FLAGS (MAP_PRIVATE) #endif /* private mmap() so we can strtok safely */ fc = (char *)mmap(NULL, fs.st_size, PROT_READ | PROT_WRITE, ZC_MMAP_FLAGS, featfd, 0); (void) close(featfd); /* map ok, and last character == newline? */ if (fc == MAP_FAILED || fc[fs.st_size - 1] != '\n') { (void) munmap((void *) fc, fs.st_size); strlcat(err_badfile, file, ZFS_MAXPROPLEN); strlcat(err_badfile, " ", ZFS_MAXPROPLEN); ret_badfile = B_TRUE; continue; } ret_nofiles = B_FALSE; for (uint_t i = 0; i < SPA_FEATURES; i++) l_features[i] = B_FALSE; /* replace final newline with NULL to ensure string ends */ fc[fs.st_size - 1] = '\0'; for (line = strtok_r(fc, "\n", &ls); line != NULL; line = strtok_r(NULL, "\n", &ls)) { /* discard comments */ char *r = strchr(line, '#'); if (r != NULL) *r = '\0'; for (word = strtok_r(line, ", \t", &ws); word != NULL; word = strtok_r(NULL, ", \t", &ws)) { /* Find matching feature name */ uint_t f; for (f = 0; f < SPA_FEATURES; f++) { zfeature_info_t *fi = &spa_feature_table[f]; if (strcmp(word, fi->fi_uname) == 0) { l_features[f] = B_TRUE; break; } } if (f < SPA_FEATURES) continue; /* found an unrecognized word */ /* lightly sanitize it */ if (strlen(word) > 32) word[32] = '\0'; for (char *c = word; *c != '\0'; c++) if (!isprint(*c)) *c = '?'; strlcat(err_badtoken, word, ZFS_MAXPROPLEN); strlcat(err_badtoken, " ", ZFS_MAXPROPLEN); if (source == Z_SYSCONF) ret_badtoken = B_TRUE; else ret_warntoken = B_TRUE; } } (void) munmap((void *) fc, fs.st_size); if (features != NULL) for (uint_t i = 0; i < SPA_FEATURES; i++) features[i] &= l_features[i]; } (void) close(sdirfd); (void) close(ddirfd); /* Return the most serious error */ if (ret_badfile) { if (report != NULL) snprintf(report, rlen, gettext("could not read/" "parse feature file(s): %s"), err_badfile); return (ZPOOL_COMPATIBILITY_BADFILE); } if (ret_nofiles) { if (report != NULL) strlcpy(report, gettext("no valid compatibility files specified"), rlen); return (ZPOOL_COMPATIBILITY_NOFILES); } if (ret_badtoken) { if (report != NULL) snprintf(report, rlen, gettext("invalid feature " "name(s) in local compatibility files: %s"), err_badtoken); return (ZPOOL_COMPATIBILITY_BADTOKEN); } if (ret_warntoken) { if (report != NULL) snprintf(report, rlen, gettext("unrecognized feature " "name(s) in distribution compatibility files: %s"), err_badtoken); return (ZPOOL_COMPATIBILITY_WARNTOKEN); } if (report != NULL) strlcpy(report, gettext("compatibility set ok"), rlen); return (ZPOOL_COMPATIBILITY_OK); } static int zpool_vdev_guid(zpool_handle_t *zhp, const char *vdevname, uint64_t *vdev_guid) { nvlist_t *tgt; boolean_t avail_spare, l2cache; verify(zhp != NULL); if (zpool_get_state(zhp) == POOL_STATE_UNAVAIL) { char errbuf[ERRBUFLEN]; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "pool is in an unavailable state")); return (zfs_error(zhp->zpool_hdl, EZFS_POOLUNAVAIL, errbuf)); } if ((tgt = zpool_find_vdev(zhp, vdevname, &avail_spare, &l2cache, NULL)) == NULL) { char errbuf[ERRBUFLEN]; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "can not find %s in %s"), vdevname, zhp->zpool_name); return (zfs_error(zhp->zpool_hdl, EZFS_NODEVICE, errbuf)); } *vdev_guid = fnvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID); return (0); } /* * Get a vdev property value for 'prop' and return the value in * a pre-allocated buffer. */ int zpool_get_vdev_prop_value(nvlist_t *nvprop, vdev_prop_t prop, char *prop_name, char *buf, size_t len, zprop_source_t *srctype, boolean_t literal) { nvlist_t *nv; const char *strval; uint64_t intval; zprop_source_t src = ZPROP_SRC_NONE; if (prop == VDEV_PROP_USERPROP) { /* user property, prop_name must contain the property name */ assert(prop_name != NULL); if (nvlist_lookup_nvlist(nvprop, prop_name, &nv) == 0) { src = fnvlist_lookup_uint64(nv, ZPROP_SOURCE); strval = fnvlist_lookup_string(nv, ZPROP_VALUE); } else { /* user prop not found */ return (-1); } (void) strlcpy(buf, strval, len); if (srctype) *srctype = src; return (0); } if (prop_name == NULL) prop_name = (char *)vdev_prop_to_name(prop); switch (vdev_prop_get_type(prop)) { case PROP_TYPE_STRING: if (nvlist_lookup_nvlist(nvprop, prop_name, &nv) == 0) { src = fnvlist_lookup_uint64(nv, ZPROP_SOURCE); strval = fnvlist_lookup_string(nv, ZPROP_VALUE); } else { src = ZPROP_SRC_DEFAULT; if ((strval = vdev_prop_default_string(prop)) == NULL) strval = "-"; } (void) strlcpy(buf, strval, len); break; case PROP_TYPE_NUMBER: if (nvlist_lookup_nvlist(nvprop, prop_name, &nv) == 0) { src = fnvlist_lookup_uint64(nv, ZPROP_SOURCE); intval = fnvlist_lookup_uint64(nv, ZPROP_VALUE); } else { src = ZPROP_SRC_DEFAULT; intval = vdev_prop_default_numeric(prop); } switch (prop) { case VDEV_PROP_ASIZE: case VDEV_PROP_PSIZE: case VDEV_PROP_SIZE: case VDEV_PROP_BOOTSIZE: case VDEV_PROP_ALLOCATED: case VDEV_PROP_FREE: case VDEV_PROP_READ_ERRORS: case VDEV_PROP_WRITE_ERRORS: case VDEV_PROP_CHECKSUM_ERRORS: case VDEV_PROP_INITIALIZE_ERRORS: case VDEV_PROP_OPS_NULL: case VDEV_PROP_OPS_READ: case VDEV_PROP_OPS_WRITE: case VDEV_PROP_OPS_FREE: case VDEV_PROP_OPS_CLAIM: case VDEV_PROP_OPS_TRIM: case VDEV_PROP_BYTES_NULL: case VDEV_PROP_BYTES_READ: case VDEV_PROP_BYTES_WRITE: case VDEV_PROP_BYTES_FREE: case VDEV_PROP_BYTES_CLAIM: case VDEV_PROP_BYTES_TRIM: if (literal) { (void) snprintf(buf, len, "%llu", (u_longlong_t)intval); } else { (void) zfs_nicenum(intval, buf, len); } break; case VDEV_PROP_EXPANDSZ: if (intval == 0) { (void) strlcpy(buf, "-", len); } else if (literal) { (void) snprintf(buf, len, "%llu", (u_longlong_t)intval); } else { (void) zfs_nicenum(intval, buf, len); } break; case VDEV_PROP_CAPACITY: if (literal) { (void) snprintf(buf, len, "%llu", (u_longlong_t)intval); } else { (void) snprintf(buf, len, "%llu%%", (u_longlong_t)intval); } break; case VDEV_PROP_CHECKSUM_N: case VDEV_PROP_CHECKSUM_T: case VDEV_PROP_IO_N: case VDEV_PROP_IO_T: if (intval == UINT64_MAX) { (void) strlcpy(buf, "-", len); } else { (void) snprintf(buf, len, "%llu", (u_longlong_t)intval); } break; case VDEV_PROP_FRAGMENTATION: if (intval == UINT64_MAX) { (void) strlcpy(buf, "-", len); } else { (void) snprintf(buf, len, "%llu%%", (u_longlong_t)intval); } break; case VDEV_PROP_STATE: if (literal) { (void) snprintf(buf, len, "%llu", (u_longlong_t)intval); } else { (void) strlcpy(buf, zpool_state_to_name(intval, VDEV_AUX_NONE), len); } break; default: (void) snprintf(buf, len, "%llu", (u_longlong_t)intval); } break; case PROP_TYPE_INDEX: if (nvlist_lookup_nvlist(nvprop, prop_name, &nv) == 0) { src = fnvlist_lookup_uint64(nv, ZPROP_SOURCE); intval = fnvlist_lookup_uint64(nv, ZPROP_VALUE); } else { src = ZPROP_SRC_DEFAULT; intval = vdev_prop_default_numeric(prop); } if (vdev_prop_index_to_string(prop, intval, (const char **)&strval) != 0) return (-1); (void) strlcpy(buf, strval, len); break; default: abort(); } if (srctype) *srctype = src; return (0); } /* * Get a vdev property value for 'prop_name' and return the value in * a pre-allocated buffer. */ int zpool_get_vdev_prop(zpool_handle_t *zhp, const char *vdevname, vdev_prop_t prop, char *prop_name, char *buf, size_t len, zprop_source_t *srctype, boolean_t literal) { nvlist_t *reqnvl, *reqprops; nvlist_t *retprops = NULL; uint64_t vdev_guid = 0; int ret; if ((ret = zpool_vdev_guid(zhp, vdevname, &vdev_guid)) != 0) return (ret); if (nvlist_alloc(&reqnvl, NV_UNIQUE_NAME, 0) != 0) return (no_memory(zhp->zpool_hdl)); if (nvlist_alloc(&reqprops, NV_UNIQUE_NAME, 0) != 0) return (no_memory(zhp->zpool_hdl)); fnvlist_add_uint64(reqnvl, ZPOOL_VDEV_PROPS_GET_VDEV, vdev_guid); if (prop != VDEV_PROP_USERPROP) { /* prop_name overrides prop value */ if (prop_name != NULL) prop = vdev_name_to_prop(prop_name); else prop_name = (char *)vdev_prop_to_name(prop); assert(prop < VDEV_NUM_PROPS); } assert(prop_name != NULL); if (nvlist_add_uint64(reqprops, prop_name, prop) != 0) { nvlist_free(reqnvl); nvlist_free(reqprops); return (no_memory(zhp->zpool_hdl)); } fnvlist_add_nvlist(reqnvl, ZPOOL_VDEV_PROPS_GET_PROPS, reqprops); ret = lzc_get_vdev_prop(zhp->zpool_name, reqnvl, &retprops); if (ret == 0) { ret = zpool_get_vdev_prop_value(retprops, prop, prop_name, buf, len, srctype, literal); } else { char errbuf[ERRBUFLEN]; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot get vdev property %s from" " %s in %s"), prop_name, vdevname, zhp->zpool_name); (void) zpool_standard_error(zhp->zpool_hdl, ret, errbuf); } nvlist_free(reqnvl); nvlist_free(reqprops); nvlist_free(retprops); return (ret); } /* * Get all vdev properties */ int zpool_get_all_vdev_props(zpool_handle_t *zhp, const char *vdevname, nvlist_t **outnvl) { nvlist_t *nvl = NULL; uint64_t vdev_guid = 0; int ret; if ((ret = zpool_vdev_guid(zhp, vdevname, &vdev_guid)) != 0) return (ret); if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0) return (no_memory(zhp->zpool_hdl)); fnvlist_add_uint64(nvl, ZPOOL_VDEV_PROPS_GET_VDEV, vdev_guid); ret = lzc_get_vdev_prop(zhp->zpool_name, nvl, outnvl); nvlist_free(nvl); if (ret) { char errbuf[ERRBUFLEN]; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot get vdev properties for" " %s in %s"), vdevname, zhp->zpool_name); (void) zpool_standard_error(zhp->zpool_hdl, errno, errbuf); } return (ret); } /* * Set vdev property */ int zpool_set_vdev_prop(zpool_handle_t *zhp, const char *vdevname, const char *propname, const char *propval) { int ret; nvlist_t *nvl = NULL; nvlist_t *outnvl = NULL; nvlist_t *props; nvlist_t *realprops; prop_flags_t flags = { 0 }; uint64_t version; uint64_t vdev_guid; if ((ret = zpool_vdev_guid(zhp, vdevname, &vdev_guid)) != 0) return (ret); if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0) return (no_memory(zhp->zpool_hdl)); if (nvlist_alloc(&props, NV_UNIQUE_NAME, 0) != 0) return (no_memory(zhp->zpool_hdl)); fnvlist_add_uint64(nvl, ZPOOL_VDEV_PROPS_SET_VDEV, vdev_guid); if (nvlist_add_string(props, propname, propval) != 0) { nvlist_free(props); return (no_memory(zhp->zpool_hdl)); } char errbuf[ERRBUFLEN]; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot set property %s for %s on %s"), propname, vdevname, zhp->zpool_name); flags.vdevprop = 1; version = zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL); if ((realprops = zpool_valid_proplist(zhp->zpool_hdl, zhp->zpool_name, props, version, flags, errbuf)) == NULL) { nvlist_free(props); nvlist_free(nvl); return (-1); } nvlist_free(props); props = realprops; fnvlist_add_nvlist(nvl, ZPOOL_VDEV_PROPS_SET_PROPS, props); ret = lzc_set_vdev_prop(zhp->zpool_name, nvl, &outnvl); nvlist_free(props); nvlist_free(nvl); nvlist_free(outnvl); if (ret) (void) zpool_standard_error(zhp->zpool_hdl, errno, errbuf); return (ret); } diff --git a/sys/contrib/openzfs/lib/libzfs/libzfs_util.c b/sys/contrib/openzfs/lib/libzfs/libzfs_util.c index 4b8a20160e02..b94abea3d581 100644 --- a/sys/contrib/openzfs/lib/libzfs/libzfs_util.c +++ b/sys/contrib/openzfs/lib/libzfs/libzfs_util.c @@ -1,2063 +1,2073 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright 2020 Joyent, Inc. All rights reserved. * Copyright (c) 2011, 2020 by Delphix. All rights reserved. * Copyright 2016 Igor Kozhukhov * Copyright (c) 2017 Datto Inc. * Copyright (c) 2020 The FreeBSD Foundation * * Portions of this software were developed by Allan Jude * under sponsorship from the FreeBSD Foundation. */ /* * Internal utility routines for the ZFS library. */ #include #include #include #include #include #include #include #include #include #if LIBFETCH_DYNAMIC #include #endif #include #include #include #include #include #include #include #include "libzfs_impl.h" #include "zfs_prop.h" #include "zfeature_common.h" #include #include /* * We only care about the scheme in order to match the scheme * with the handler. Each handler should validate the full URI * as necessary. */ #define URI_REGEX "^\\([A-Za-z][A-Za-z0-9+.\\-]*\\):" int libzfs_errno(libzfs_handle_t *hdl) { return (hdl->libzfs_error); } const char * libzfs_error_action(libzfs_handle_t *hdl) { return (hdl->libzfs_action); } const char * libzfs_error_description(libzfs_handle_t *hdl) { if (hdl->libzfs_desc[0] != '\0') return (hdl->libzfs_desc); switch (hdl->libzfs_error) { case EZFS_NOMEM: return (dgettext(TEXT_DOMAIN, "out of memory")); case EZFS_BADPROP: return (dgettext(TEXT_DOMAIN, "invalid property value")); case EZFS_PROPREADONLY: return (dgettext(TEXT_DOMAIN, "read-only property")); case EZFS_PROPTYPE: return (dgettext(TEXT_DOMAIN, "property doesn't apply to " "datasets of this type")); case EZFS_PROPNONINHERIT: return (dgettext(TEXT_DOMAIN, "property cannot be inherited")); case EZFS_PROPSPACE: return (dgettext(TEXT_DOMAIN, "invalid quota or reservation")); case EZFS_BADTYPE: return (dgettext(TEXT_DOMAIN, "operation not applicable to " "datasets of this type")); case EZFS_BUSY: return (dgettext(TEXT_DOMAIN, "pool or dataset is busy")); case EZFS_EXISTS: return (dgettext(TEXT_DOMAIN, "pool or dataset exists")); case EZFS_NOENT: return (dgettext(TEXT_DOMAIN, "no such pool or dataset")); case EZFS_BADSTREAM: return (dgettext(TEXT_DOMAIN, "invalid backup stream")); case EZFS_DSREADONLY: return (dgettext(TEXT_DOMAIN, "dataset is read-only")); case EZFS_VOLTOOBIG: return (dgettext(TEXT_DOMAIN, "volume size exceeds limit for " "this system")); case EZFS_INVALIDNAME: return (dgettext(TEXT_DOMAIN, "invalid name")); case EZFS_BADRESTORE: return (dgettext(TEXT_DOMAIN, "unable to restore to " "destination")); case EZFS_BADBACKUP: return (dgettext(TEXT_DOMAIN, "backup failed")); case EZFS_BADTARGET: return (dgettext(TEXT_DOMAIN, "invalid target vdev")); case EZFS_NODEVICE: return (dgettext(TEXT_DOMAIN, "no such device in pool")); case EZFS_BADDEV: return (dgettext(TEXT_DOMAIN, "invalid device")); case EZFS_NOREPLICAS: return (dgettext(TEXT_DOMAIN, "no valid replicas")); case EZFS_RESILVERING: return (dgettext(TEXT_DOMAIN, "currently resilvering")); case EZFS_BADVERSION: return (dgettext(TEXT_DOMAIN, "unsupported version or " "feature")); case EZFS_POOLUNAVAIL: return (dgettext(TEXT_DOMAIN, "pool is unavailable")); case EZFS_DEVOVERFLOW: return (dgettext(TEXT_DOMAIN, "too many devices in one vdev")); case EZFS_BADPATH: return (dgettext(TEXT_DOMAIN, "must be an absolute path")); case EZFS_CROSSTARGET: return (dgettext(TEXT_DOMAIN, "operation crosses datasets or " "pools")); case EZFS_ZONED: return (dgettext(TEXT_DOMAIN, "dataset in use by local zone")); case EZFS_MOUNTFAILED: return (dgettext(TEXT_DOMAIN, "mount failed")); case EZFS_UMOUNTFAILED: return (dgettext(TEXT_DOMAIN, "unmount failed")); case EZFS_UNSHARENFSFAILED: return (dgettext(TEXT_DOMAIN, "NFS share removal failed")); case EZFS_SHARENFSFAILED: return (dgettext(TEXT_DOMAIN, "NFS share creation failed")); case EZFS_UNSHARESMBFAILED: return (dgettext(TEXT_DOMAIN, "SMB share removal failed")); case EZFS_SHARESMBFAILED: return (dgettext(TEXT_DOMAIN, "SMB share creation failed")); case EZFS_PERM: return (dgettext(TEXT_DOMAIN, "permission denied")); case EZFS_NOSPC: return (dgettext(TEXT_DOMAIN, "out of space")); case EZFS_FAULT: return (dgettext(TEXT_DOMAIN, "bad address")); case EZFS_IO: return (dgettext(TEXT_DOMAIN, "I/O error")); case EZFS_INTR: return (dgettext(TEXT_DOMAIN, "signal received")); case EZFS_CKSUM: return (dgettext(TEXT_DOMAIN, "insufficient replicas")); case EZFS_ISSPARE: return (dgettext(TEXT_DOMAIN, "device is reserved as a hot " "spare")); case EZFS_INVALCONFIG: return (dgettext(TEXT_DOMAIN, "invalid vdev configuration")); case EZFS_RECURSIVE: return (dgettext(TEXT_DOMAIN, "recursive dataset dependency")); case EZFS_NOHISTORY: return (dgettext(TEXT_DOMAIN, "no history available")); case EZFS_POOLPROPS: return (dgettext(TEXT_DOMAIN, "failed to retrieve " "pool properties")); case EZFS_POOL_NOTSUP: return (dgettext(TEXT_DOMAIN, "operation not supported " "on this type of pool")); case EZFS_POOL_INVALARG: return (dgettext(TEXT_DOMAIN, "invalid argument for " "this pool operation")); case EZFS_NAMETOOLONG: return (dgettext(TEXT_DOMAIN, "dataset name is too long")); case EZFS_OPENFAILED: return (dgettext(TEXT_DOMAIN, "open failed")); case EZFS_NOCAP: return (dgettext(TEXT_DOMAIN, "disk capacity information could not be retrieved")); case EZFS_LABELFAILED: return (dgettext(TEXT_DOMAIN, "write of label failed")); case EZFS_BADWHO: return (dgettext(TEXT_DOMAIN, "invalid user/group")); case EZFS_BADPERM: return (dgettext(TEXT_DOMAIN, "invalid permission")); case EZFS_BADPERMSET: return (dgettext(TEXT_DOMAIN, "invalid permission set name")); case EZFS_NODELEGATION: return (dgettext(TEXT_DOMAIN, "delegated administration is " "disabled on pool")); case EZFS_BADCACHE: return (dgettext(TEXT_DOMAIN, "invalid or missing cache file")); case EZFS_ISL2CACHE: return (dgettext(TEXT_DOMAIN, "device is in use as a cache")); case EZFS_VDEVNOTSUP: return (dgettext(TEXT_DOMAIN, "vdev specification is not " "supported")); case EZFS_NOTSUP: return (dgettext(TEXT_DOMAIN, "operation not supported " "on this dataset")); case EZFS_IOC_NOTSUPPORTED: return (dgettext(TEXT_DOMAIN, "operation not supported by " "zfs kernel module")); case EZFS_ACTIVE_SPARE: return (dgettext(TEXT_DOMAIN, "pool has active shared spare " "device")); case EZFS_UNPLAYED_LOGS: return (dgettext(TEXT_DOMAIN, "log device has unplayed intent " "logs")); case EZFS_REFTAG_RELE: return (dgettext(TEXT_DOMAIN, "no such tag on this dataset")); case EZFS_REFTAG_HOLD: return (dgettext(TEXT_DOMAIN, "tag already exists on this " "dataset")); case EZFS_TAGTOOLONG: return (dgettext(TEXT_DOMAIN, "tag too long")); case EZFS_PIPEFAILED: return (dgettext(TEXT_DOMAIN, "pipe create failed")); case EZFS_THREADCREATEFAILED: return (dgettext(TEXT_DOMAIN, "thread create failed")); case EZFS_POSTSPLIT_ONLINE: return (dgettext(TEXT_DOMAIN, "disk was split from this pool " "into a new one")); case EZFS_SCRUB_PAUSED: return (dgettext(TEXT_DOMAIN, "scrub is paused; " - "use 'zpool scrub' to resume")); + "use 'zpool scrub' to resume scrub")); + case EZFS_SCRUB_PAUSED_TO_CANCEL: + return (dgettext(TEXT_DOMAIN, "scrub is paused; " + "use 'zpool scrub' to resume or 'zpool scrub -s' to " + "cancel scrub")); case EZFS_SCRUBBING: return (dgettext(TEXT_DOMAIN, "currently scrubbing; " - "use 'zpool scrub -s' to cancel current scrub")); + "use 'zpool scrub -s' to cancel scrub")); + case EZFS_ERRORSCRUBBING: + return (dgettext(TEXT_DOMAIN, "currently error scrubbing; " + "use 'zpool scrub -s' to cancel error scrub")); + case EZFS_ERRORSCRUB_PAUSED: + return (dgettext(TEXT_DOMAIN, "error scrub is paused; " + "use 'zpool scrub -e' to resume error scrub")); case EZFS_NO_SCRUB: return (dgettext(TEXT_DOMAIN, "there is no active scrub")); case EZFS_DIFF: return (dgettext(TEXT_DOMAIN, "unable to generate diffs")); case EZFS_DIFFDATA: return (dgettext(TEXT_DOMAIN, "invalid diff data")); case EZFS_POOLREADONLY: return (dgettext(TEXT_DOMAIN, "pool is read-only")); case EZFS_NO_PENDING: return (dgettext(TEXT_DOMAIN, "operation is not " "in progress")); case EZFS_CHECKPOINT_EXISTS: return (dgettext(TEXT_DOMAIN, "checkpoint exists")); case EZFS_DISCARDING_CHECKPOINT: return (dgettext(TEXT_DOMAIN, "currently discarding " "checkpoint")); case EZFS_NO_CHECKPOINT: return (dgettext(TEXT_DOMAIN, "checkpoint does not exist")); case EZFS_DEVRM_IN_PROGRESS: return (dgettext(TEXT_DOMAIN, "device removal in progress")); case EZFS_VDEV_TOO_BIG: return (dgettext(TEXT_DOMAIN, "device exceeds supported size")); case EZFS_ACTIVE_POOL: return (dgettext(TEXT_DOMAIN, "pool is imported on a " "different host")); case EZFS_CRYPTOFAILED: return (dgettext(TEXT_DOMAIN, "encryption failure")); case EZFS_TOOMANY: return (dgettext(TEXT_DOMAIN, "argument list too long")); case EZFS_INITIALIZING: return (dgettext(TEXT_DOMAIN, "currently initializing")); case EZFS_NO_INITIALIZE: return (dgettext(TEXT_DOMAIN, "there is no active " "initialization")); case EZFS_WRONG_PARENT: return (dgettext(TEXT_DOMAIN, "invalid parent dataset")); case EZFS_TRIMMING: return (dgettext(TEXT_DOMAIN, "currently trimming")); case EZFS_NO_TRIM: return (dgettext(TEXT_DOMAIN, "there is no active trim")); case EZFS_TRIM_NOTSUP: return (dgettext(TEXT_DOMAIN, "trim operations are not " "supported by this device")); case EZFS_NO_RESILVER_DEFER: return (dgettext(TEXT_DOMAIN, "this action requires the " "resilver_defer feature")); case EZFS_EXPORT_IN_PROGRESS: return (dgettext(TEXT_DOMAIN, "pool export in progress")); case EZFS_REBUILDING: return (dgettext(TEXT_DOMAIN, "currently sequentially " "resilvering")); case EZFS_VDEV_NOTSUP: return (dgettext(TEXT_DOMAIN, "operation not supported " "on this type of vdev")); case EZFS_NOT_USER_NAMESPACE: return (dgettext(TEXT_DOMAIN, "the provided file " "was not a user namespace file")); case EZFS_RESUME_EXISTS: return (dgettext(TEXT_DOMAIN, "Resuming recv on existing " "dataset without force")); case EZFS_UNKNOWN: return (dgettext(TEXT_DOMAIN, "unknown error")); default: assert(hdl->libzfs_error == 0); return (dgettext(TEXT_DOMAIN, "no error")); } } void zfs_error_aux(libzfs_handle_t *hdl, const char *fmt, ...) { va_list ap; va_start(ap, fmt); (void) vsnprintf(hdl->libzfs_desc, sizeof (hdl->libzfs_desc), fmt, ap); hdl->libzfs_desc_active = 1; va_end(ap); } static void zfs_verror(libzfs_handle_t *hdl, int error, const char *fmt, va_list ap) { (void) vsnprintf(hdl->libzfs_action, sizeof (hdl->libzfs_action), fmt, ap); hdl->libzfs_error = error; if (hdl->libzfs_desc_active) hdl->libzfs_desc_active = 0; else hdl->libzfs_desc[0] = '\0'; if (hdl->libzfs_printerr) { if (error == EZFS_UNKNOWN) { (void) fprintf(stderr, dgettext(TEXT_DOMAIN, "internal " "error: %s: %s\n"), hdl->libzfs_action, libzfs_error_description(hdl)); abort(); } (void) fprintf(stderr, "%s: %s\n", hdl->libzfs_action, libzfs_error_description(hdl)); if (error == EZFS_NOMEM) exit(1); } } int zfs_error(libzfs_handle_t *hdl, int error, const char *msg) { return (zfs_error_fmt(hdl, error, "%s", msg)); } int zfs_error_fmt(libzfs_handle_t *hdl, int error, const char *fmt, ...) { va_list ap; va_start(ap, fmt); zfs_verror(hdl, error, fmt, ap); va_end(ap); return (-1); } static int zfs_common_error(libzfs_handle_t *hdl, int error, const char *fmt, va_list ap) { switch (error) { case EPERM: case EACCES: zfs_verror(hdl, EZFS_PERM, fmt, ap); return (-1); case ECANCELED: zfs_verror(hdl, EZFS_NODELEGATION, fmt, ap); return (-1); case EIO: zfs_verror(hdl, EZFS_IO, fmt, ap); return (-1); case EFAULT: zfs_verror(hdl, EZFS_FAULT, fmt, ap); return (-1); case EINTR: zfs_verror(hdl, EZFS_INTR, fmt, ap); return (-1); case ECKSUM: zfs_verror(hdl, EZFS_CKSUM, fmt, ap); return (-1); } return (0); } int zfs_standard_error(libzfs_handle_t *hdl, int error, const char *msg) { return (zfs_standard_error_fmt(hdl, error, "%s", msg)); } int zfs_standard_error_fmt(libzfs_handle_t *hdl, int error, const char *fmt, ...) { va_list ap; va_start(ap, fmt); if (zfs_common_error(hdl, error, fmt, ap) != 0) { va_end(ap); return (-1); } switch (error) { case ENXIO: case ENODEV: case EPIPE: zfs_verror(hdl, EZFS_IO, fmt, ap); break; case ENOENT: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "dataset does not exist")); zfs_verror(hdl, EZFS_NOENT, fmt, ap); break; case ENOSPC: case EDQUOT: zfs_verror(hdl, EZFS_NOSPC, fmt, ap); break; case EEXIST: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "dataset already exists")); zfs_verror(hdl, EZFS_EXISTS, fmt, ap); break; case EBUSY: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "dataset is busy")); zfs_verror(hdl, EZFS_BUSY, fmt, ap); break; case EROFS: zfs_verror(hdl, EZFS_POOLREADONLY, fmt, ap); break; case ENAMETOOLONG: zfs_verror(hdl, EZFS_NAMETOOLONG, fmt, ap); break; case ENOTSUP: zfs_verror(hdl, EZFS_BADVERSION, fmt, ap); break; case EAGAIN: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool I/O is currently suspended")); zfs_verror(hdl, EZFS_POOLUNAVAIL, fmt, ap); break; case EREMOTEIO: zfs_verror(hdl, EZFS_ACTIVE_POOL, fmt, ap); break; case ZFS_ERR_UNKNOWN_SEND_STREAM_FEATURE: case ZFS_ERR_IOC_CMD_UNAVAIL: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "the loaded zfs " "module does not support this operation. A reboot may " "be required to enable this operation.")); zfs_verror(hdl, EZFS_IOC_NOTSUPPORTED, fmt, ap); break; case ZFS_ERR_IOC_ARG_UNAVAIL: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "the loaded zfs " "module does not support an option for this operation. " "A reboot may be required to enable this option.")); zfs_verror(hdl, EZFS_IOC_NOTSUPPORTED, fmt, ap); break; case ZFS_ERR_IOC_ARG_REQUIRED: case ZFS_ERR_IOC_ARG_BADTYPE: zfs_verror(hdl, EZFS_IOC_NOTSUPPORTED, fmt, ap); break; case ZFS_ERR_WRONG_PARENT: zfs_verror(hdl, EZFS_WRONG_PARENT, fmt, ap); break; case ZFS_ERR_BADPROP: zfs_verror(hdl, EZFS_BADPROP, fmt, ap); break; case ZFS_ERR_NOT_USER_NAMESPACE: zfs_verror(hdl, EZFS_NOT_USER_NAMESPACE, fmt, ap); break; default: zfs_error_aux(hdl, "%s", strerror(error)); zfs_verror(hdl, EZFS_UNKNOWN, fmt, ap); break; } va_end(ap); return (-1); } void zfs_setprop_error(libzfs_handle_t *hdl, zfs_prop_t prop, int err, char *errbuf) { switch (err) { case ENOSPC: /* * For quotas and reservations, ENOSPC indicates * something different; setting a quota or reservation * doesn't use any disk space. */ switch (prop) { case ZFS_PROP_QUOTA: case ZFS_PROP_REFQUOTA: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "size is less than current used or " "reserved space")); (void) zfs_error(hdl, EZFS_PROPSPACE, errbuf); break; case ZFS_PROP_RESERVATION: case ZFS_PROP_REFRESERVATION: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "size is greater than available space")); (void) zfs_error(hdl, EZFS_PROPSPACE, errbuf); break; default: (void) zfs_standard_error(hdl, err, errbuf); break; } break; case EBUSY: (void) zfs_standard_error(hdl, EBUSY, errbuf); break; case EROFS: (void) zfs_error(hdl, EZFS_DSREADONLY, errbuf); break; case E2BIG: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "property value too long")); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); break; case ENOTSUP: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool and or dataset must be upgraded to set this " "property or value")); (void) zfs_error(hdl, EZFS_BADVERSION, errbuf); break; case ERANGE: if (prop == ZFS_PROP_COMPRESSION || prop == ZFS_PROP_DNODESIZE || prop == ZFS_PROP_RECORDSIZE) { (void) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "property setting is not allowed on " "bootable datasets")); (void) zfs_error(hdl, EZFS_NOTSUP, errbuf); } else if (prop == ZFS_PROP_CHECKSUM || prop == ZFS_PROP_DEDUP) { (void) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "property setting is not allowed on " "root pools")); (void) zfs_error(hdl, EZFS_NOTSUP, errbuf); } else { (void) zfs_standard_error(hdl, err, errbuf); } break; case EINVAL: if (prop == ZPROP_INVAL) { (void) zfs_error(hdl, EZFS_BADPROP, errbuf); } else { (void) zfs_standard_error(hdl, err, errbuf); } break; case ZFS_ERR_BADPROP: (void) zfs_error(hdl, EZFS_BADPROP, errbuf); break; case EACCES: if (prop == ZFS_PROP_KEYLOCATION) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "keylocation may only be set on encryption roots")); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); } else { (void) zfs_standard_error(hdl, err, errbuf); } break; case EOVERFLOW: /* * This platform can't address a volume this big. */ #ifdef _ILP32 if (prop == ZFS_PROP_VOLSIZE) { (void) zfs_error(hdl, EZFS_VOLTOOBIG, errbuf); break; } zfs_fallthrough; #endif default: (void) zfs_standard_error(hdl, err, errbuf); } } int zpool_standard_error(libzfs_handle_t *hdl, int error, const char *msg) { return (zpool_standard_error_fmt(hdl, error, "%s", msg)); } int zpool_standard_error_fmt(libzfs_handle_t *hdl, int error, const char *fmt, ...) { va_list ap; va_start(ap, fmt); if (zfs_common_error(hdl, error, fmt, ap) != 0) { va_end(ap); return (-1); } switch (error) { case ENODEV: zfs_verror(hdl, EZFS_NODEVICE, fmt, ap); break; case ENOENT: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "no such pool or dataset")); zfs_verror(hdl, EZFS_NOENT, fmt, ap); break; case EEXIST: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool already exists")); zfs_verror(hdl, EZFS_EXISTS, fmt, ap); break; case EBUSY: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool is busy")); zfs_verror(hdl, EZFS_BUSY, fmt, ap); break; /* There is no pending operation to cancel */ case ENOTACTIVE: zfs_verror(hdl, EZFS_NO_PENDING, fmt, ap); break; case ENXIO: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "one or more devices is currently unavailable")); zfs_verror(hdl, EZFS_BADDEV, fmt, ap); break; case ENAMETOOLONG: zfs_verror(hdl, EZFS_DEVOVERFLOW, fmt, ap); break; case ENOTSUP: zfs_verror(hdl, EZFS_POOL_NOTSUP, fmt, ap); break; case EINVAL: zfs_verror(hdl, EZFS_POOL_INVALARG, fmt, ap); break; case ENOSPC: case EDQUOT: zfs_verror(hdl, EZFS_NOSPC, fmt, ap); break; case EAGAIN: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool I/O is currently suspended")); zfs_verror(hdl, EZFS_POOLUNAVAIL, fmt, ap); break; case EROFS: zfs_verror(hdl, EZFS_POOLREADONLY, fmt, ap); break; case EDOM: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "block size out of range or does not match")); zfs_verror(hdl, EZFS_BADPROP, fmt, ap); break; case EREMOTEIO: zfs_verror(hdl, EZFS_ACTIVE_POOL, fmt, ap); break; case ZFS_ERR_CHECKPOINT_EXISTS: zfs_verror(hdl, EZFS_CHECKPOINT_EXISTS, fmt, ap); break; case ZFS_ERR_DISCARDING_CHECKPOINT: zfs_verror(hdl, EZFS_DISCARDING_CHECKPOINT, fmt, ap); break; case ZFS_ERR_NO_CHECKPOINT: zfs_verror(hdl, EZFS_NO_CHECKPOINT, fmt, ap); break; case ZFS_ERR_DEVRM_IN_PROGRESS: zfs_verror(hdl, EZFS_DEVRM_IN_PROGRESS, fmt, ap); break; case ZFS_ERR_VDEV_TOO_BIG: zfs_verror(hdl, EZFS_VDEV_TOO_BIG, fmt, ap); break; case ZFS_ERR_EXPORT_IN_PROGRESS: zfs_verror(hdl, EZFS_EXPORT_IN_PROGRESS, fmt, ap); break; case ZFS_ERR_RESILVER_IN_PROGRESS: zfs_verror(hdl, EZFS_RESILVERING, fmt, ap); break; case ZFS_ERR_REBUILD_IN_PROGRESS: zfs_verror(hdl, EZFS_REBUILDING, fmt, ap); break; case ZFS_ERR_BADPROP: zfs_verror(hdl, EZFS_BADPROP, fmt, ap); break; case ZFS_ERR_VDEV_NOTSUP: zfs_verror(hdl, EZFS_VDEV_NOTSUP, fmt, ap); break; case ZFS_ERR_IOC_CMD_UNAVAIL: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "the loaded zfs " "module does not support this operation. A reboot may " "be required to enable this operation.")); zfs_verror(hdl, EZFS_IOC_NOTSUPPORTED, fmt, ap); break; case ZFS_ERR_IOC_ARG_UNAVAIL: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "the loaded zfs " "module does not support an option for this operation. " "A reboot may be required to enable this option.")); zfs_verror(hdl, EZFS_IOC_NOTSUPPORTED, fmt, ap); break; case ZFS_ERR_IOC_ARG_REQUIRED: case ZFS_ERR_IOC_ARG_BADTYPE: zfs_verror(hdl, EZFS_IOC_NOTSUPPORTED, fmt, ap); break; default: zfs_error_aux(hdl, "%s", strerror(error)); zfs_verror(hdl, EZFS_UNKNOWN, fmt, ap); } va_end(ap); return (-1); } /* * Display an out of memory error message and abort the current program. */ int no_memory(libzfs_handle_t *hdl) { return (zfs_error(hdl, EZFS_NOMEM, "internal error")); } /* * A safe form of malloc() which will die if the allocation fails. */ void * zfs_alloc(libzfs_handle_t *hdl, size_t size) { void *data; if ((data = calloc(1, size)) == NULL) (void) no_memory(hdl); return (data); } /* * A safe form of asprintf() which will die if the allocation fails. */ char * zfs_asprintf(libzfs_handle_t *hdl, const char *fmt, ...) { va_list ap; char *ret; int err; va_start(ap, fmt); err = vasprintf(&ret, fmt, ap); va_end(ap); if (err < 0) { (void) no_memory(hdl); ret = NULL; } return (ret); } /* * A safe form of realloc(), which also zeroes newly allocated space. */ void * zfs_realloc(libzfs_handle_t *hdl, void *ptr, size_t oldsize, size_t newsize) { void *ret; if ((ret = realloc(ptr, newsize)) == NULL) { (void) no_memory(hdl); return (NULL); } memset((char *)ret + oldsize, 0, newsize - oldsize); return (ret); } /* * A safe form of strdup() which will die if the allocation fails. */ char * zfs_strdup(libzfs_handle_t *hdl, const char *str) { char *ret; if ((ret = strdup(str)) == NULL) (void) no_memory(hdl); return (ret); } void libzfs_print_on_error(libzfs_handle_t *hdl, boolean_t printerr) { hdl->libzfs_printerr = printerr; } /* * Read lines from an open file descriptor and store them in an array of * strings until EOF. lines[] will be allocated and populated with all the * lines read. All newlines are replaced with NULL terminators for * convenience. lines[] must be freed after use with libzfs_free_str_array(). * * Returns the number of lines read. */ static int libzfs_read_stdout_from_fd(int fd, char **lines[]) { FILE *fp; int lines_cnt = 0; size_t len = 0; char *line = NULL; char **tmp_lines = NULL, **tmp; fp = fdopen(fd, "r"); if (fp == NULL) { close(fd); return (0); } while (getline(&line, &len, fp) != -1) { tmp = realloc(tmp_lines, sizeof (*tmp_lines) * (lines_cnt + 1)); if (tmp == NULL) { /* Return the lines we were able to process */ break; } tmp_lines = tmp; /* Remove newline if not EOF */ if (line[strlen(line) - 1] == '\n') line[strlen(line) - 1] = '\0'; tmp_lines[lines_cnt] = strdup(line); if (tmp_lines[lines_cnt] == NULL) break; ++lines_cnt; } free(line); fclose(fp); *lines = tmp_lines; return (lines_cnt); } static int libzfs_run_process_impl(const char *path, char *argv[], char *env[], int flags, char **lines[], int *lines_cnt) { pid_t pid; int error, devnull_fd; int link[2]; /* * Setup a pipe between our child and parent process if we're * reading stdout. */ if (lines != NULL && pipe2(link, O_NONBLOCK | O_CLOEXEC) == -1) return (-EPIPE); pid = fork(); if (pid == 0) { /* Child process */ devnull_fd = open("/dev/null", O_WRONLY | O_CLOEXEC); if (devnull_fd < 0) _exit(-1); if (!(flags & STDOUT_VERBOSE) && (lines == NULL)) (void) dup2(devnull_fd, STDOUT_FILENO); else if (lines != NULL) { /* Save the output to lines[] */ dup2(link[1], STDOUT_FILENO); } if (!(flags & STDERR_VERBOSE)) (void) dup2(devnull_fd, STDERR_FILENO); if (flags & NO_DEFAULT_PATH) { if (env == NULL) execv(path, argv); else execve(path, argv, env); } else { if (env == NULL) execvp(path, argv); else execvpe(path, argv, env); } _exit(-1); } else if (pid > 0) { /* Parent process */ int status; while ((error = waitpid(pid, &status, 0)) == -1 && errno == EINTR) ; if (error < 0 || !WIFEXITED(status)) return (-1); if (lines != NULL) { close(link[1]); *lines_cnt = libzfs_read_stdout_from_fd(link[0], lines); } return (WEXITSTATUS(status)); } return (-1); } int libzfs_run_process(const char *path, char *argv[], int flags) { return (libzfs_run_process_impl(path, argv, NULL, flags, NULL, NULL)); } /* * Run a command and store its stdout lines in an array of strings (lines[]). * lines[] is allocated and populated for you, and the number of lines is set in * lines_cnt. lines[] must be freed after use with libzfs_free_str_array(). * All newlines (\n) in lines[] are terminated for convenience. */ int libzfs_run_process_get_stdout(const char *path, char *argv[], char *env[], char **lines[], int *lines_cnt) { return (libzfs_run_process_impl(path, argv, env, 0, lines, lines_cnt)); } /* * Same as libzfs_run_process_get_stdout(), but run without $PATH set. This * means that *path needs to be the full path to the executable. */ int libzfs_run_process_get_stdout_nopath(const char *path, char *argv[], char *env[], char **lines[], int *lines_cnt) { return (libzfs_run_process_impl(path, argv, env, NO_DEFAULT_PATH, lines, lines_cnt)); } /* * Free an array of strings. Free both the strings contained in the array and * the array itself. */ void libzfs_free_str_array(char **strs, int count) { while (--count >= 0) free(strs[count]); free(strs); } /* * Returns 1 if environment variable is set to "YES", "yes", "ON", "on", or * a non-zero number. * * Returns 0 otherwise. */ boolean_t libzfs_envvar_is_set(const char *envvar) { char *env = getenv(envvar); return (env && (strtoul(env, NULL, 0) > 0 || (!strncasecmp(env, "YES", 3) && strnlen(env, 4) == 3) || (!strncasecmp(env, "ON", 2) && strnlen(env, 3) == 2))); } libzfs_handle_t * libzfs_init(void) { libzfs_handle_t *hdl; int error; char *env; if ((error = libzfs_load_module()) != 0) { errno = error; return (NULL); } if ((hdl = calloc(1, sizeof (libzfs_handle_t))) == NULL) { return (NULL); } if (regcomp(&hdl->libzfs_urire, URI_REGEX, 0) != 0) { free(hdl); return (NULL); } if ((hdl->libzfs_fd = open(ZFS_DEV, O_RDWR|O_EXCL|O_CLOEXEC)) < 0) { free(hdl); return (NULL); } if (libzfs_core_init() != 0) { (void) close(hdl->libzfs_fd); free(hdl); return (NULL); } zfs_prop_init(); zpool_prop_init(); zpool_feature_init(); vdev_prop_init(); libzfs_mnttab_init(hdl); fletcher_4_init(); if (getenv("ZFS_PROP_DEBUG") != NULL) { hdl->libzfs_prop_debug = B_TRUE; } if ((env = getenv("ZFS_SENDRECV_MAX_NVLIST")) != NULL) { if ((error = zfs_nicestrtonum(hdl, env, &hdl->libzfs_max_nvlist))) { errno = error; (void) close(hdl->libzfs_fd); free(hdl); return (NULL); } } else { hdl->libzfs_max_nvlist = (SPA_MAXBLOCKSIZE * 4); } /* * For testing, remove some settable properties and features */ if (libzfs_envvar_is_set("ZFS_SYSFS_PROP_SUPPORT_TEST")) { zprop_desc_t *proptbl; proptbl = zpool_prop_get_table(); proptbl[ZPOOL_PROP_COMMENT].pd_zfs_mod_supported = B_FALSE; proptbl = zfs_prop_get_table(); proptbl[ZFS_PROP_DNODESIZE].pd_zfs_mod_supported = B_FALSE; zfeature_info_t *ftbl = spa_feature_table; ftbl[SPA_FEATURE_LARGE_BLOCKS].fi_zfs_mod_supported = B_FALSE; } return (hdl); } void libzfs_fini(libzfs_handle_t *hdl) { (void) close(hdl->libzfs_fd); zpool_free_handles(hdl); namespace_clear(hdl); libzfs_mnttab_fini(hdl); libzfs_core_fini(); regfree(&hdl->libzfs_urire); fletcher_4_fini(); #if LIBFETCH_DYNAMIC if (hdl->libfetch != (void *)-1 && hdl->libfetch != NULL) (void) dlclose(hdl->libfetch); free(hdl->libfetch_load_error); #endif free(hdl); } libzfs_handle_t * zpool_get_handle(zpool_handle_t *zhp) { return (zhp->zpool_hdl); } libzfs_handle_t * zfs_get_handle(zfs_handle_t *zhp) { return (zhp->zfs_hdl); } zpool_handle_t * zfs_get_pool_handle(const zfs_handle_t *zhp) { return (zhp->zpool_hdl); } /* * Given a name, determine whether or not it's a valid path * (starts with '/' or "./"). If so, walk the mnttab trying * to match the device number. If not, treat the path as an * fs/vol/snap/bkmark name. */ zfs_handle_t * zfs_path_to_zhandle(libzfs_handle_t *hdl, const char *path, zfs_type_t argtype) { struct stat64 statbuf; struct extmnttab entry; if (path[0] != '/' && strncmp(path, "./", strlen("./")) != 0) { /* * It's not a valid path, assume it's a name of type 'argtype'. */ return (zfs_open(hdl, path, argtype)); } if (getextmntent(path, &entry, &statbuf) != 0) return (NULL); if (strcmp(entry.mnt_fstype, MNTTYPE_ZFS) != 0) { (void) fprintf(stderr, gettext("'%s': not a ZFS filesystem\n"), path); return (NULL); } return (zfs_open(hdl, entry.mnt_special, ZFS_TYPE_FILESYSTEM)); } /* * Initialize the zc_nvlist_dst member to prepare for receiving an nvlist from * an ioctl(). */ void zcmd_alloc_dst_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, size_t len) { if (len == 0) len = 256 * 1024; zc->zc_nvlist_dst_size = len; zc->zc_nvlist_dst = (uint64_t)(uintptr_t)zfs_alloc(hdl, zc->zc_nvlist_dst_size); } /* * Called when an ioctl() which returns an nvlist fails with ENOMEM. This will * expand the nvlist to the size specified in 'zc_nvlist_dst_size', which was * filled in by the kernel to indicate the actual required size. */ void zcmd_expand_dst_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc) { free((void *)(uintptr_t)zc->zc_nvlist_dst); zc->zc_nvlist_dst = (uint64_t)(uintptr_t)zfs_alloc(hdl, zc->zc_nvlist_dst_size); } /* * Called to free the src and dst nvlists stored in the command structure. */ void zcmd_free_nvlists(zfs_cmd_t *zc) { free((void *)(uintptr_t)zc->zc_nvlist_conf); free((void *)(uintptr_t)zc->zc_nvlist_src); free((void *)(uintptr_t)zc->zc_nvlist_dst); zc->zc_nvlist_conf = 0; zc->zc_nvlist_src = 0; zc->zc_nvlist_dst = 0; } static void zcmd_write_nvlist_com(libzfs_handle_t *hdl, uint64_t *outnv, uint64_t *outlen, nvlist_t *nvl) { char *packed; size_t len = fnvlist_size(nvl); packed = zfs_alloc(hdl, len); verify(nvlist_pack(nvl, &packed, &len, NV_ENCODE_NATIVE, 0) == 0); *outnv = (uint64_t)(uintptr_t)packed; *outlen = len; } void zcmd_write_conf_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, nvlist_t *nvl) { zcmd_write_nvlist_com(hdl, &zc->zc_nvlist_conf, &zc->zc_nvlist_conf_size, nvl); } void zcmd_write_src_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, nvlist_t *nvl) { zcmd_write_nvlist_com(hdl, &zc->zc_nvlist_src, &zc->zc_nvlist_src_size, nvl); } /* * Unpacks an nvlist from the ZFS ioctl command structure. */ int zcmd_read_dst_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, nvlist_t **nvlp) { if (nvlist_unpack((void *)(uintptr_t)zc->zc_nvlist_dst, zc->zc_nvlist_dst_size, nvlp, 0) != 0) return (no_memory(hdl)); return (0); } /* * ================================================================ * API shared by zfs and zpool property management * ================================================================ */ static void zprop_print_headers(zprop_get_cbdata_t *cbp, zfs_type_t type) { zprop_list_t *pl; int i; char *title; size_t len; cbp->cb_first = B_FALSE; if (cbp->cb_scripted) return; /* * Start with the length of the column headers. */ cbp->cb_colwidths[GET_COL_NAME] = strlen(dgettext(TEXT_DOMAIN, "NAME")); cbp->cb_colwidths[GET_COL_PROPERTY] = strlen(dgettext(TEXT_DOMAIN, "PROPERTY")); cbp->cb_colwidths[GET_COL_VALUE] = strlen(dgettext(TEXT_DOMAIN, "VALUE")); cbp->cb_colwidths[GET_COL_RECVD] = strlen(dgettext(TEXT_DOMAIN, "RECEIVED")); cbp->cb_colwidths[GET_COL_SOURCE] = strlen(dgettext(TEXT_DOMAIN, "SOURCE")); /* first property is always NAME */ assert(cbp->cb_proplist->pl_prop == ((type == ZFS_TYPE_POOL) ? ZPOOL_PROP_NAME : ((type == ZFS_TYPE_VDEV) ? VDEV_PROP_NAME : ZFS_PROP_NAME))); /* * Go through and calculate the widths for each column. For the * 'source' column, we kludge it up by taking the worst-case scenario of * inheriting from the longest name. This is acceptable because in the * majority of cases 'SOURCE' is the last column displayed, and we don't * use the width anyway. Note that the 'VALUE' column can be oversized, * if the name of the property is much longer than any values we find. */ for (pl = cbp->cb_proplist; pl != NULL; pl = pl->pl_next) { /* * 'PROPERTY' column */ if (pl->pl_prop != ZPROP_USERPROP) { const char *propname = (type == ZFS_TYPE_POOL) ? zpool_prop_to_name(pl->pl_prop) : ((type == ZFS_TYPE_VDEV) ? vdev_prop_to_name(pl->pl_prop) : zfs_prop_to_name(pl->pl_prop)); assert(propname != NULL); len = strlen(propname); if (len > cbp->cb_colwidths[GET_COL_PROPERTY]) cbp->cb_colwidths[GET_COL_PROPERTY] = len; } else { assert(pl->pl_user_prop != NULL); len = strlen(pl->pl_user_prop); if (len > cbp->cb_colwidths[GET_COL_PROPERTY]) cbp->cb_colwidths[GET_COL_PROPERTY] = len; } /* * 'VALUE' column. The first property is always the 'name' * property that was tacked on either by /sbin/zfs's * zfs_do_get() or when calling zprop_expand_list(), so we * ignore its width. If the user specified the name property * to display, then it will be later in the list in any case. */ if (pl != cbp->cb_proplist && pl->pl_width > cbp->cb_colwidths[GET_COL_VALUE]) cbp->cb_colwidths[GET_COL_VALUE] = pl->pl_width; /* 'RECEIVED' column. */ if (pl != cbp->cb_proplist && pl->pl_recvd_width > cbp->cb_colwidths[GET_COL_RECVD]) cbp->cb_colwidths[GET_COL_RECVD] = pl->pl_recvd_width; /* * 'NAME' and 'SOURCE' columns */ if (pl->pl_prop == ((type == ZFS_TYPE_POOL) ? ZPOOL_PROP_NAME : ((type == ZFS_TYPE_VDEV) ? VDEV_PROP_NAME : ZFS_PROP_NAME)) && pl->pl_width > cbp->cb_colwidths[GET_COL_NAME]) { cbp->cb_colwidths[GET_COL_NAME] = pl->pl_width; cbp->cb_colwidths[GET_COL_SOURCE] = pl->pl_width + strlen(dgettext(TEXT_DOMAIN, "inherited from")); } } /* * Now go through and print the headers. */ for (i = 0; i < ZFS_GET_NCOLS; i++) { switch (cbp->cb_columns[i]) { case GET_COL_NAME: title = dgettext(TEXT_DOMAIN, "NAME"); break; case GET_COL_PROPERTY: title = dgettext(TEXT_DOMAIN, "PROPERTY"); break; case GET_COL_VALUE: title = dgettext(TEXT_DOMAIN, "VALUE"); break; case GET_COL_RECVD: title = dgettext(TEXT_DOMAIN, "RECEIVED"); break; case GET_COL_SOURCE: title = dgettext(TEXT_DOMAIN, "SOURCE"); break; default: title = NULL; } if (title != NULL) { if (i == (ZFS_GET_NCOLS - 1) || cbp->cb_columns[i + 1] == GET_COL_NONE) (void) printf("%s", title); else (void) printf("%-*s ", cbp->cb_colwidths[cbp->cb_columns[i]], title); } } (void) printf("\n"); } /* * Display a single line of output, according to the settings in the callback * structure. */ void zprop_print_one_property(const char *name, zprop_get_cbdata_t *cbp, const char *propname, const char *value, zprop_source_t sourcetype, const char *source, const char *recvd_value) { int i; const char *str = NULL; char buf[128]; /* * Ignore those source types that the user has chosen to ignore. */ if ((sourcetype & cbp->cb_sources) == 0) return; if (cbp->cb_first) zprop_print_headers(cbp, cbp->cb_type); for (i = 0; i < ZFS_GET_NCOLS; i++) { switch (cbp->cb_columns[i]) { case GET_COL_NAME: str = name; break; case GET_COL_PROPERTY: str = propname; break; case GET_COL_VALUE: str = value; break; case GET_COL_SOURCE: switch (sourcetype) { case ZPROP_SRC_NONE: str = "-"; break; case ZPROP_SRC_DEFAULT: str = "default"; break; case ZPROP_SRC_LOCAL: str = "local"; break; case ZPROP_SRC_TEMPORARY: str = "temporary"; break; case ZPROP_SRC_INHERITED: (void) snprintf(buf, sizeof (buf), "inherited from %s", source); str = buf; break; case ZPROP_SRC_RECEIVED: str = "received"; break; default: str = NULL; assert(!"unhandled zprop_source_t"); } break; case GET_COL_RECVD: str = (recvd_value == NULL ? "-" : recvd_value); break; default: continue; } if (i == (ZFS_GET_NCOLS - 1) || cbp->cb_columns[i + 1] == GET_COL_NONE) (void) printf("%s", str); else if (cbp->cb_scripted) (void) printf("%s\t", str); else (void) printf("%-*s ", cbp->cb_colwidths[cbp->cb_columns[i]], str); } (void) printf("\n"); } /* * Given a numeric suffix, convert the value into a number of bits that the * resulting value must be shifted. */ static int str2shift(libzfs_handle_t *hdl, const char *buf) { const char *ends = "BKMGTPEZ"; int i; if (buf[0] == '\0') return (0); for (i = 0; i < strlen(ends); i++) { if (toupper(buf[0]) == ends[i]) break; } if (i == strlen(ends)) { if (hdl) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid numeric suffix '%s'"), buf); return (-1); } /* * Allow 'G' = 'GB' = 'GiB', case-insensitively. * However, 'BB' and 'BiB' are disallowed. */ if (buf[1] == '\0' || (toupper(buf[0]) != 'B' && ((toupper(buf[1]) == 'B' && buf[2] == '\0') || (toupper(buf[1]) == 'I' && toupper(buf[2]) == 'B' && buf[3] == '\0')))) return (10 * i); if (hdl) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid numeric suffix '%s'"), buf); return (-1); } /* * Convert a string of the form '100G' into a real number. Used when setting * properties or creating a volume. 'buf' is used to place an extended error * message for the caller to use. */ int zfs_nicestrtonum(libzfs_handle_t *hdl, const char *value, uint64_t *num) { char *end; int shift; *num = 0; /* Check to see if this looks like a number. */ if ((value[0] < '0' || value[0] > '9') && value[0] != '.') { if (hdl) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "bad numeric value '%s'"), value); return (-1); } /* Rely on strtoull() to process the numeric portion. */ errno = 0; *num = strtoull(value, &end, 10); /* * Check for ERANGE, which indicates that the value is too large to fit * in a 64-bit value. */ if (errno == ERANGE) { if (hdl) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "numeric value is too large")); return (-1); } /* * If we have a decimal value, then do the computation with floating * point arithmetic. Otherwise, use standard arithmetic. */ if (*end == '.') { double fval = strtod(value, &end); if ((shift = str2shift(hdl, end)) == -1) return (-1); fval *= pow(2, shift); /* * UINT64_MAX is not exactly representable as a double. * The closest representation is UINT64_MAX + 1, so we * use a >= comparison instead of > for the bounds check. */ if (fval >= (double)UINT64_MAX) { if (hdl) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "numeric value is too large")); return (-1); } *num = (uint64_t)fval; } else { if ((shift = str2shift(hdl, end)) == -1) return (-1); /* Check for overflow */ if (shift >= 64 || (*num << shift) >> shift != *num) { if (hdl) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "numeric value is too large")); return (-1); } *num <<= shift; } return (0); } /* * Given a propname=value nvpair to set, parse any numeric properties * (index, boolean, etc) if they are specified as strings and add the * resulting nvpair to the returned nvlist. * * At the DSL layer, all properties are either 64-bit numbers or strings. * We want the user to be able to ignore this fact and specify properties * as native values (numbers, for example) or as strings (to simplify * command line utilities). This also handles converting index types * (compression, checksum, etc) from strings to their on-disk index. */ int zprop_parse_value(libzfs_handle_t *hdl, nvpair_t *elem, int prop, zfs_type_t type, nvlist_t *ret, const char **svalp, uint64_t *ivalp, const char *errbuf) { data_type_t datatype = nvpair_type(elem); zprop_type_t proptype; const char *propname; const char *value; boolean_t isnone = B_FALSE; boolean_t isauto = B_FALSE; int err = 0; if (type == ZFS_TYPE_POOL) { proptype = zpool_prop_get_type(prop); propname = zpool_prop_to_name(prop); } else if (type == ZFS_TYPE_VDEV) { proptype = vdev_prop_get_type(prop); propname = vdev_prop_to_name(prop); } else { proptype = zfs_prop_get_type(prop); propname = zfs_prop_to_name(prop); } /* * Convert any properties to the internal DSL value types. */ *svalp = NULL; *ivalp = 0; switch (proptype) { case PROP_TYPE_STRING: if (datatype != DATA_TYPE_STRING) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' must be a string"), nvpair_name(elem)); goto error; } err = nvpair_value_string(elem, svalp); if (err != 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' is invalid"), nvpair_name(elem)); goto error; } if (strlen(*svalp) >= ZFS_MAXPROPLEN) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' is too long"), nvpair_name(elem)); goto error; } break; case PROP_TYPE_NUMBER: if (datatype == DATA_TYPE_STRING) { (void) nvpair_value_string(elem, &value); if (strcmp(value, "none") == 0) { isnone = B_TRUE; } else if (strcmp(value, "auto") == 0) { isauto = B_TRUE; } else if (zfs_nicestrtonum(hdl, value, ivalp) != 0) { goto error; } } else if (datatype == DATA_TYPE_UINT64) { (void) nvpair_value_uint64(elem, ivalp); } else { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' must be a number"), nvpair_name(elem)); goto error; } /* * Quota special: force 'none' and don't allow 0. */ if ((type & ZFS_TYPE_DATASET) && *ivalp == 0 && !isnone && (prop == ZFS_PROP_QUOTA || prop == ZFS_PROP_REFQUOTA)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "use 'none' to disable quota/refquota")); goto error; } /* * Special handling for "*_limit=none". In this case it's not * 0 but UINT64_MAX. */ if ((type & ZFS_TYPE_DATASET) && isnone && (prop == ZFS_PROP_FILESYSTEM_LIMIT || prop == ZFS_PROP_SNAPSHOT_LIMIT)) { *ivalp = UINT64_MAX; } /* * Special handling for "checksum_*=none". In this case it's not * 0 but UINT64_MAX. */ if ((type & ZFS_TYPE_VDEV) && isnone && (prop == VDEV_PROP_CHECKSUM_N || prop == VDEV_PROP_CHECKSUM_T || prop == VDEV_PROP_IO_N || prop == VDEV_PROP_IO_T)) { *ivalp = UINT64_MAX; } /* * Special handling for setting 'refreservation' to 'auto'. Use * UINT64_MAX to tell the caller to use zfs_fix_auto_resv(). * 'auto' is only allowed on volumes. */ if (isauto) { switch (prop) { case ZFS_PROP_REFRESERVATION: if ((type & ZFS_TYPE_VOLUME) == 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s=auto' only allowed on " "volumes"), nvpair_name(elem)); goto error; } *ivalp = UINT64_MAX; break; default: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'auto' is invalid value for '%s'"), nvpair_name(elem)); goto error; } } break; case PROP_TYPE_INDEX: if (datatype != DATA_TYPE_STRING) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' must be a string"), nvpair_name(elem)); goto error; } (void) nvpair_value_string(elem, &value); if (zprop_string_to_index(prop, value, ivalp, type) != 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' must be one of '%s'"), propname, zprop_values(prop, type)); goto error; } break; default: abort(); } /* * Add the result to our return set of properties. */ if (*svalp != NULL) { if (nvlist_add_string(ret, propname, *svalp) != 0) { (void) no_memory(hdl); return (-1); } } else { if (nvlist_add_uint64(ret, propname, *ivalp) != 0) { (void) no_memory(hdl); return (-1); } } return (0); error: (void) zfs_error(hdl, EZFS_BADPROP, errbuf); return (-1); } static int addlist(libzfs_handle_t *hdl, const char *propname, zprop_list_t **listp, zfs_type_t type) { int prop = zprop_name_to_prop(propname, type); if (prop != ZPROP_INVAL && !zprop_valid_for_type(prop, type, B_FALSE)) prop = ZPROP_INVAL; /* * Return failure if no property table entry was found and this isn't * a user-defined property. */ if (prop == ZPROP_USERPROP && ((type == ZFS_TYPE_POOL && !zfs_prop_user(propname) && !zpool_prop_feature(propname) && !zpool_prop_unsupported(propname)) || ((type == ZFS_TYPE_DATASET) && !zfs_prop_user(propname) && !zfs_prop_userquota(propname) && !zfs_prop_written(propname)) || ((type == ZFS_TYPE_VDEV) && !vdev_prop_user(propname)))) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid property '%s'"), propname); return (zfs_error(hdl, EZFS_BADPROP, dgettext(TEXT_DOMAIN, "bad property list"))); } zprop_list_t *entry = zfs_alloc(hdl, sizeof (*entry)); entry->pl_prop = prop; if (prop == ZPROP_USERPROP) { entry->pl_user_prop = zfs_strdup(hdl, propname); entry->pl_width = strlen(propname); } else { entry->pl_width = zprop_width(prop, &entry->pl_fixed, type); } *listp = entry; return (0); } /* * Given a comma-separated list of properties, construct a property list * containing both user-defined and native properties. This function will * return a NULL list if 'all' is specified, which can later be expanded * by zprop_expand_list(). */ int zprop_get_list(libzfs_handle_t *hdl, char *props, zprop_list_t **listp, zfs_type_t type) { *listp = NULL; /* * If 'all' is specified, return a NULL list. */ if (strcmp(props, "all") == 0) return (0); /* * If no props were specified, return an error. */ if (props[0] == '\0') { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "no properties specified")); return (zfs_error(hdl, EZFS_BADPROP, dgettext(TEXT_DOMAIN, "bad property list"))); } for (char *p; (p = strsep(&props, ",")); ) if (strcmp(p, "space") == 0) { static const char *const spaceprops[] = { "name", "avail", "used", "usedbysnapshots", "usedbydataset", "usedbyrefreservation", "usedbychildren" }; for (int i = 0; i < ARRAY_SIZE(spaceprops); i++) { if (addlist(hdl, spaceprops[i], listp, type)) return (-1); listp = &(*listp)->pl_next; } } else { if (addlist(hdl, p, listp, type)) return (-1); listp = &(*listp)->pl_next; } return (0); } void zprop_free_list(zprop_list_t *pl) { zprop_list_t *next; while (pl != NULL) { next = pl->pl_next; free(pl->pl_user_prop); free(pl); pl = next; } } typedef struct expand_data { zprop_list_t **last; libzfs_handle_t *hdl; zfs_type_t type; } expand_data_t; static int zprop_expand_list_cb(int prop, void *cb) { zprop_list_t *entry; expand_data_t *edp = cb; entry = zfs_alloc(edp->hdl, sizeof (zprop_list_t)); entry->pl_prop = prop; entry->pl_width = zprop_width(prop, &entry->pl_fixed, edp->type); entry->pl_all = B_TRUE; *(edp->last) = entry; edp->last = &entry->pl_next; return (ZPROP_CONT); } int zprop_expand_list(libzfs_handle_t *hdl, zprop_list_t **plp, zfs_type_t type) { zprop_list_t *entry; zprop_list_t **last; expand_data_t exp; if (*plp == NULL) { /* * If this is the very first time we've been called for an 'all' * specification, expand the list to include all native * properties. */ last = plp; exp.last = last; exp.hdl = hdl; exp.type = type; if (zprop_iter_common(zprop_expand_list_cb, &exp, B_FALSE, B_FALSE, type) == ZPROP_INVAL) return (-1); /* * Add 'name' to the beginning of the list, which is handled * specially. */ entry = zfs_alloc(hdl, sizeof (zprop_list_t)); entry->pl_prop = ((type == ZFS_TYPE_POOL) ? ZPOOL_PROP_NAME : ((type == ZFS_TYPE_VDEV) ? VDEV_PROP_NAME : ZFS_PROP_NAME)); entry->pl_width = zprop_width(entry->pl_prop, &entry->pl_fixed, type); entry->pl_all = B_TRUE; entry->pl_next = *plp; *plp = entry; } return (0); } int zprop_iter(zprop_func func, void *cb, boolean_t show_all, boolean_t ordered, zfs_type_t type) { return (zprop_iter_common(func, cb, show_all, ordered, type)); } const char * zfs_version_userland(void) { return (ZFS_META_ALIAS); } /* * Prints both zfs userland and kernel versions * Returns 0 on success, and -1 on error */ int zfs_version_print(void) { (void) puts(ZFS_META_ALIAS); char *kver = zfs_version_kernel(); if (kver == NULL) { fprintf(stderr, "zfs_version_kernel() failed: %s\n", strerror(errno)); return (-1); } (void) printf("zfs-kmod-%s\n", kver); free(kver); return (0); } /* * Return 1 if the user requested ANSI color output, and our terminal supports * it. Return 0 for no color. */ int use_color(void) { static int use_color = -1; char *term; /* * Optimization: * * For each zpool invocation, we do a single check to see if we should * be using color or not, and cache that value for the lifetime of the * the zpool command. That makes it cheap to call use_color() when * we're printing with color. We assume that the settings are not going * to change during the invocation of a zpool command (the user isn't * going to change the ZFS_COLOR value while zpool is running, for * example). */ if (use_color != -1) { /* * We've already figured out if we should be using color or * not. Return the cached value. */ return (use_color); } term = getenv("TERM"); /* * The user sets the ZFS_COLOR env var set to enable zpool ANSI color * output. However if NO_COLOR is set (https://no-color.org/) then * don't use it. Also, don't use color if terminal doesn't support * it. */ if (libzfs_envvar_is_set("ZFS_COLOR") && !libzfs_envvar_is_set("NO_COLOR") && isatty(STDOUT_FILENO) && term && strcmp("dumb", term) != 0 && strcmp("unknown", term) != 0) { /* Color supported */ use_color = 1; } else { use_color = 0; } return (use_color); } /* * The functions color_start() and color_end() are used for when you want * to colorize a block of text. * * For example: * color_start(ANSI_RED) * printf("hello"); * printf("world"); * color_end(); */ void color_start(const char *color) { if (color && use_color()) { fputs(color, stdout); fflush(stdout); } } void color_end(void) { if (use_color()) { fputs(ANSI_RESET, stdout); fflush(stdout); } } /* * printf() with a color. If color is NULL, then do a normal printf. */ int printf_color(const char *color, const char *format, ...) { va_list aptr; int rc; if (color) color_start(color); va_start(aptr, format); rc = vprintf(format, aptr); va_end(aptr); if (color) color_end(); return (rc); } diff --git a/sys/contrib/openzfs/lib/libzfs_core/libzfs_core.abi b/sys/contrib/openzfs/lib/libzfs_core/libzfs_core.abi index ec94a4650553..f2087186aa44 100644 --- a/sys/contrib/openzfs/lib/libzfs_core/libzfs_core.abi +++ b/sys/contrib/openzfs/lib/libzfs_core/libzfs_core.abi @@ -1,2814 +1,2920 @@ + - + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + diff --git a/sys/contrib/openzfs/lib/libzfs_core/libzfs_core.c b/sys/contrib/openzfs/lib/libzfs_core/libzfs_core.c index 254f14e04321..c63a16de5ab6 100644 --- a/sys/contrib/openzfs/lib/libzfs_core/libzfs_core.c +++ b/sys/contrib/openzfs/lib/libzfs_core/libzfs_core.c @@ -1,1894 +1,1901 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2012, 2020 by Delphix. All rights reserved. * Copyright (c) 2013 Steven Hartland. All rights reserved. * Copyright 2017 RackTop Systems. * Copyright (c) 2017 Open-E, Inc. All Rights Reserved. * Copyright (c) 2019, 2020 by Christian Schwarz. All rights reserved. * Copyright (c) 2019 Datto Inc. */ /* * LibZFS_Core (lzc) is intended to replace most functionality in libzfs. * It has the following characteristics: * * - Thread Safe. libzfs_core is accessible concurrently from multiple * threads. This is accomplished primarily by avoiding global data * (e.g. caching). Since it's thread-safe, there is no reason for a * process to have multiple libzfs "instances". Therefore, we store * our few pieces of data (e.g. the file descriptor) in global * variables. The fd is reference-counted so that the libzfs_core * library can be "initialized" multiple times (e.g. by different * consumers within the same process). * * - Committed Interface. The libzfs_core interface will be committed, * therefore consumers can compile against it and be confident that * their code will continue to work on future releases of this code. * Currently, the interface is Evolving (not Committed), but we intend * to commit to it once it is more complete and we determine that it * meets the needs of all consumers. * * - Programmatic Error Handling. libzfs_core communicates errors with * defined error numbers, and doesn't print anything to stdout/stderr. * * - Thin Layer. libzfs_core is a thin layer, marshaling arguments * to/from the kernel ioctls. There is generally a 1:1 correspondence * between libzfs_core functions and ioctls to ZFS_DEV. * * - Clear Atomicity. Because libzfs_core functions are generally 1:1 * with kernel ioctls, and kernel ioctls are general atomic, each * libzfs_core function is atomic. For example, creating multiple * snapshots with a single call to lzc_snapshot() is atomic -- it * can't fail with only some of the requested snapshots created, even * in the event of power loss or system crash. * * - Continued libzfs Support. Some higher-level operations (e.g. * support for "zfs send -R") are too complicated to fit the scope of * libzfs_core. This functionality will continue to live in libzfs. * Where appropriate, libzfs will use the underlying atomic operations * of libzfs_core. For example, libzfs may implement "zfs send -R | * zfs receive" by using individual "send one snapshot", rename, * destroy, and "receive one snapshot" operations in libzfs_core. * /sbin/zfs and /sbin/zpool will link with both libzfs and * libzfs_core. Other consumers should aim to use only libzfs_core, * since that will be the supported, stable interface going forwards. */ #include #include #include #include #include #ifdef ZFS_DEBUG #include #endif #include #include #include #include #include #include #include #include #include #if __FreeBSD__ #define BIG_PIPE_SIZE (64 * 1024) /* From sys/pipe.h */ #endif static int g_fd = -1; static pthread_mutex_t g_lock = PTHREAD_MUTEX_INITIALIZER; static int g_refcount; #ifdef ZFS_DEBUG static zfs_ioc_t fail_ioc_cmd = ZFS_IOC_LAST; static zfs_errno_t fail_ioc_err; static void libzfs_core_debug_ioc(void) { /* * To test running newer user space binaries with kernel's * that don't yet support an ioctl or a new ioctl arg we * provide an override to intentionally fail an ioctl. * * USAGE: * The override variable, ZFS_IOC_TEST, is of the form "cmd:err" * * For example, to fail a ZFS_IOC_POOL_CHECKPOINT with a * ZFS_ERR_IOC_CMD_UNAVAIL, the string would be "0x5a4d:1029" * * $ sudo sh -c "ZFS_IOC_TEST=0x5a4d:1029 zpool checkpoint tank" * cannot checkpoint 'tank': the loaded zfs module does not support * this operation. A reboot may be required to enable this operation. */ if (fail_ioc_cmd == ZFS_IOC_LAST) { char *ioc_test = getenv("ZFS_IOC_TEST"); unsigned int ioc_num = 0, ioc_err = 0; if (ioc_test != NULL && sscanf(ioc_test, "%i:%i", &ioc_num, &ioc_err) == 2 && ioc_num < ZFS_IOC_LAST) { fail_ioc_cmd = ioc_num; fail_ioc_err = ioc_err; } } } #endif int libzfs_core_init(void) { (void) pthread_mutex_lock(&g_lock); if (g_refcount == 0) { g_fd = open(ZFS_DEV, O_RDWR|O_CLOEXEC); if (g_fd < 0) { (void) pthread_mutex_unlock(&g_lock); return (errno); } } g_refcount++; #ifdef ZFS_DEBUG libzfs_core_debug_ioc(); #endif (void) pthread_mutex_unlock(&g_lock); return (0); } void libzfs_core_fini(void) { (void) pthread_mutex_lock(&g_lock); ASSERT3S(g_refcount, >, 0); g_refcount--; if (g_refcount == 0 && g_fd != -1) { (void) close(g_fd); g_fd = -1; } (void) pthread_mutex_unlock(&g_lock); } static int lzc_ioctl(zfs_ioc_t ioc, const char *name, nvlist_t *source, nvlist_t **resultp) { zfs_cmd_t zc = {"\0"}; int error = 0; char *packed = NULL; size_t size = 0; ASSERT3S(g_refcount, >, 0); VERIFY3S(g_fd, !=, -1); #ifdef ZFS_DEBUG if (ioc == fail_ioc_cmd) return (fail_ioc_err); #endif if (name != NULL) (void) strlcpy(zc.zc_name, name, sizeof (zc.zc_name)); if (source != NULL) { packed = fnvlist_pack(source, &size); zc.zc_nvlist_src = (uint64_t)(uintptr_t)packed; zc.zc_nvlist_src_size = size; } if (resultp != NULL) { *resultp = NULL; if (ioc == ZFS_IOC_CHANNEL_PROGRAM) { zc.zc_nvlist_dst_size = fnvlist_lookup_uint64(source, ZCP_ARG_MEMLIMIT); } else { zc.zc_nvlist_dst_size = MAX(size * 2, 128 * 1024); } zc.zc_nvlist_dst = (uint64_t)(uintptr_t) malloc(zc.zc_nvlist_dst_size); if (zc.zc_nvlist_dst == (uint64_t)0) { error = ENOMEM; goto out; } } while (lzc_ioctl_fd(g_fd, ioc, &zc) != 0) { /* * If ioctl exited with ENOMEM, we retry the ioctl after * increasing the size of the destination nvlist. * * Channel programs that exit with ENOMEM ran over the * lua memory sandbox; they should not be retried. */ if (errno == ENOMEM && resultp != NULL && ioc != ZFS_IOC_CHANNEL_PROGRAM) { free((void *)(uintptr_t)zc.zc_nvlist_dst); zc.zc_nvlist_dst_size *= 2; zc.zc_nvlist_dst = (uint64_t)(uintptr_t) malloc(zc.zc_nvlist_dst_size); if (zc.zc_nvlist_dst == (uint64_t)0) { error = ENOMEM; goto out; } } else { error = errno; break; } } if (zc.zc_nvlist_dst_filled && resultp != NULL) { *resultp = fnvlist_unpack((void *)(uintptr_t)zc.zc_nvlist_dst, zc.zc_nvlist_dst_size); } out: if (packed != NULL) fnvlist_pack_free(packed, size); free((void *)(uintptr_t)zc.zc_nvlist_dst); return (error); } +int +lzc_scrub(zfs_ioc_t ioc, const char *name, + nvlist_t *source, nvlist_t **resultp) +{ + return (lzc_ioctl(ioc, name, source, resultp)); +} + int lzc_create(const char *fsname, enum lzc_dataset_type type, nvlist_t *props, uint8_t *wkeydata, uint_t wkeylen) { int error; nvlist_t *hidden_args = NULL; nvlist_t *args = fnvlist_alloc(); fnvlist_add_int32(args, "type", (dmu_objset_type_t)type); if (props != NULL) fnvlist_add_nvlist(args, "props", props); if (wkeydata != NULL) { hidden_args = fnvlist_alloc(); fnvlist_add_uint8_array(hidden_args, "wkeydata", wkeydata, wkeylen); fnvlist_add_nvlist(args, ZPOOL_HIDDEN_ARGS, hidden_args); } error = lzc_ioctl(ZFS_IOC_CREATE, fsname, args, NULL); nvlist_free(hidden_args); nvlist_free(args); return (error); } int lzc_clone(const char *fsname, const char *origin, nvlist_t *props) { int error; nvlist_t *hidden_args = NULL; nvlist_t *args = fnvlist_alloc(); fnvlist_add_string(args, "origin", origin); if (props != NULL) fnvlist_add_nvlist(args, "props", props); error = lzc_ioctl(ZFS_IOC_CLONE, fsname, args, NULL); nvlist_free(hidden_args); nvlist_free(args); return (error); } int lzc_promote(const char *fsname, char *snapnamebuf, int snapnamelen) { /* * The promote ioctl is still legacy, so we need to construct our * own zfs_cmd_t rather than using lzc_ioctl(). */ zfs_cmd_t zc = {"\0"}; ASSERT3S(g_refcount, >, 0); VERIFY3S(g_fd, !=, -1); (void) strlcpy(zc.zc_name, fsname, sizeof (zc.zc_name)); if (lzc_ioctl_fd(g_fd, ZFS_IOC_PROMOTE, &zc) != 0) { int error = errno; if (error == EEXIST && snapnamebuf != NULL) (void) strlcpy(snapnamebuf, zc.zc_string, snapnamelen); return (error); } return (0); } int lzc_rename(const char *source, const char *target) { zfs_cmd_t zc = {"\0"}; int error; ASSERT3S(g_refcount, >, 0); VERIFY3S(g_fd, !=, -1); (void) strlcpy(zc.zc_name, source, sizeof (zc.zc_name)); (void) strlcpy(zc.zc_value, target, sizeof (zc.zc_value)); error = lzc_ioctl_fd(g_fd, ZFS_IOC_RENAME, &zc); if (error != 0) error = errno; return (error); } int lzc_destroy(const char *fsname) { int error; nvlist_t *args = fnvlist_alloc(); error = lzc_ioctl(ZFS_IOC_DESTROY, fsname, args, NULL); nvlist_free(args); return (error); } /* * Creates snapshots. * * The keys in the snaps nvlist are the snapshots to be created. * They must all be in the same pool. * * The props nvlist is properties to set. Currently only user properties * are supported. { user:prop_name -> string value } * * The returned results nvlist will have an entry for each snapshot that failed. * The value will be the (int32) error code. * * The return value will be 0 if all snapshots were created, otherwise it will * be the errno of a (unspecified) snapshot that failed. */ int lzc_snapshot(nvlist_t *snaps, nvlist_t *props, nvlist_t **errlist) { nvpair_t *elem; nvlist_t *args; int error; char pool[ZFS_MAX_DATASET_NAME_LEN]; *errlist = NULL; /* determine the pool name */ elem = nvlist_next_nvpair(snaps, NULL); if (elem == NULL) return (0); (void) strlcpy(pool, nvpair_name(elem), sizeof (pool)); pool[strcspn(pool, "/@")] = '\0'; args = fnvlist_alloc(); fnvlist_add_nvlist(args, "snaps", snaps); if (props != NULL) fnvlist_add_nvlist(args, "props", props); error = lzc_ioctl(ZFS_IOC_SNAPSHOT, pool, args, errlist); nvlist_free(args); return (error); } /* * Destroys snapshots. * * The keys in the snaps nvlist are the snapshots to be destroyed. * They must all be in the same pool. * * Snapshots that do not exist will be silently ignored. * * If 'defer' is not set, and a snapshot has user holds or clones, the * destroy operation will fail and none of the snapshots will be * destroyed. * * If 'defer' is set, and a snapshot has user holds or clones, it will be * marked for deferred destruction, and will be destroyed when the last hold * or clone is removed/destroyed. * * The return value will be 0 if all snapshots were destroyed (or marked for * later destruction if 'defer' is set) or didn't exist to begin with. * * Otherwise the return value will be the errno of a (unspecified) snapshot * that failed, no snapshots will be destroyed, and the errlist will have an * entry for each snapshot that failed. The value in the errlist will be * the (int32) error code. */ int lzc_destroy_snaps(nvlist_t *snaps, boolean_t defer, nvlist_t **errlist) { nvpair_t *elem; nvlist_t *args; int error; char pool[ZFS_MAX_DATASET_NAME_LEN]; /* determine the pool name */ elem = nvlist_next_nvpair(snaps, NULL); if (elem == NULL) return (0); (void) strlcpy(pool, nvpair_name(elem), sizeof (pool)); pool[strcspn(pool, "/@")] = '\0'; args = fnvlist_alloc(); fnvlist_add_nvlist(args, "snaps", snaps); if (defer) fnvlist_add_boolean(args, "defer"); error = lzc_ioctl(ZFS_IOC_DESTROY_SNAPS, pool, args, errlist); nvlist_free(args); return (error); } int lzc_snaprange_space(const char *firstsnap, const char *lastsnap, uint64_t *usedp) { nvlist_t *args; nvlist_t *result; int err; char fs[ZFS_MAX_DATASET_NAME_LEN]; char *atp; /* determine the fs name */ (void) strlcpy(fs, firstsnap, sizeof (fs)); atp = strchr(fs, '@'); if (atp == NULL) return (EINVAL); *atp = '\0'; args = fnvlist_alloc(); fnvlist_add_string(args, "firstsnap", firstsnap); err = lzc_ioctl(ZFS_IOC_SPACE_SNAPS, lastsnap, args, &result); nvlist_free(args); if (err == 0) *usedp = fnvlist_lookup_uint64(result, "used"); fnvlist_free(result); return (err); } boolean_t lzc_exists(const char *dataset) { /* * The objset_stats ioctl is still legacy, so we need to construct our * own zfs_cmd_t rather than using lzc_ioctl(). */ zfs_cmd_t zc = {"\0"}; ASSERT3S(g_refcount, >, 0); VERIFY3S(g_fd, !=, -1); (void) strlcpy(zc.zc_name, dataset, sizeof (zc.zc_name)); return (lzc_ioctl_fd(g_fd, ZFS_IOC_OBJSET_STATS, &zc) == 0); } /* * outnvl is unused. * It was added to preserve the function signature in case it is * needed in the future. */ int lzc_sync(const char *pool_name, nvlist_t *innvl, nvlist_t **outnvl) { (void) outnvl; return (lzc_ioctl(ZFS_IOC_POOL_SYNC, pool_name, innvl, NULL)); } /* * Create "user holds" on snapshots. If there is a hold on a snapshot, * the snapshot can not be destroyed. (However, it can be marked for deletion * by lzc_destroy_snaps(defer=B_TRUE).) * * The keys in the nvlist are snapshot names. * The snapshots must all be in the same pool. * The value is the name of the hold (string type). * * If cleanup_fd is not -1, it must be the result of open(ZFS_DEV, O_EXCL). * In this case, when the cleanup_fd is closed (including on process * termination), the holds will be released. If the system is shut down * uncleanly, the holds will be released when the pool is next opened * or imported. * * Holds for snapshots which don't exist will be skipped and have an entry * added to errlist, but will not cause an overall failure. * * The return value will be 0 if all holds, for snapshots that existed, * were successfully created. * * Otherwise the return value will be the errno of a (unspecified) hold that * failed and no holds will be created. * * In all cases the errlist will have an entry for each hold that failed * (name = snapshot), with its value being the error code (int32). */ int lzc_hold(nvlist_t *holds, int cleanup_fd, nvlist_t **errlist) { char pool[ZFS_MAX_DATASET_NAME_LEN]; nvlist_t *args; nvpair_t *elem; int error; /* determine the pool name */ elem = nvlist_next_nvpair(holds, NULL); if (elem == NULL) return (0); (void) strlcpy(pool, nvpair_name(elem), sizeof (pool)); pool[strcspn(pool, "/@")] = '\0'; args = fnvlist_alloc(); fnvlist_add_nvlist(args, "holds", holds); if (cleanup_fd != -1) fnvlist_add_int32(args, "cleanup_fd", cleanup_fd); error = lzc_ioctl(ZFS_IOC_HOLD, pool, args, errlist); nvlist_free(args); return (error); } /* * Release "user holds" on snapshots. If the snapshot has been marked for * deferred destroy (by lzc_destroy_snaps(defer=B_TRUE)), it does not have * any clones, and all the user holds are removed, then the snapshot will be * destroyed. * * The keys in the nvlist are snapshot names. * The snapshots must all be in the same pool. * The value is an nvlist whose keys are the holds to remove. * * Holds which failed to release because they didn't exist will have an entry * added to errlist, but will not cause an overall failure. * * The return value will be 0 if the nvl holds was empty or all holds that * existed, were successfully removed. * * Otherwise the return value will be the errno of a (unspecified) hold that * failed to release and no holds will be released. * * In all cases the errlist will have an entry for each hold that failed to * to release. */ int lzc_release(nvlist_t *holds, nvlist_t **errlist) { char pool[ZFS_MAX_DATASET_NAME_LEN]; nvpair_t *elem; /* determine the pool name */ elem = nvlist_next_nvpair(holds, NULL); if (elem == NULL) return (0); (void) strlcpy(pool, nvpair_name(elem), sizeof (pool)); pool[strcspn(pool, "/@")] = '\0'; return (lzc_ioctl(ZFS_IOC_RELEASE, pool, holds, errlist)); } /* * Retrieve list of user holds on the specified snapshot. * * On success, *holdsp will be set to an nvlist which the caller must free. * The keys are the names of the holds, and the value is the creation time * of the hold (uint64) in seconds since the epoch. */ int lzc_get_holds(const char *snapname, nvlist_t **holdsp) { return (lzc_ioctl(ZFS_IOC_GET_HOLDS, snapname, NULL, holdsp)); } static unsigned int max_pipe_buffer(int infd) { #if __linux__ static unsigned int max; if (max == 0) { max = 1048576; /* fs/pipe.c default */ FILE *procf = fopen("/proc/sys/fs/pipe-max-size", "re"); if (procf != NULL) { if (fscanf(procf, "%u", &max) <= 0) { /* ignore error: max untouched if parse fails */ } fclose(procf); } } unsigned int cur = fcntl(infd, F_GETPIPE_SZ); /* * Sadly, Linux has an unfixed deadlock if you do SETPIPE_SZ on a pipe * with data in it. * cf. #13232, https://bugzilla.kernel.org/show_bug.cgi?id=212295 * * And since the problem is in waking up the writer, there's nothing * we can do about it from here. * * So if people want to, they can set this, but they * may regret it... */ if (getenv("ZFS_SET_PIPE_MAX") == NULL) return (cur); if (cur < max && fcntl(infd, F_SETPIPE_SZ, max) != -1) cur = max; return (cur); #else /* FreeBSD automatically resizes */ (void) infd; return (BIG_PIPE_SIZE); #endif } #if __linux__ struct send_worker_ctx { int from; /* read end of pipe, with send data; closed on exit */ int to; /* original arbitrary output fd; mustn't be a pipe */ }; static void * send_worker(void *arg) { struct send_worker_ctx *ctx = arg; unsigned int bufsiz = max_pipe_buffer(ctx->from); ssize_t rd; while ((rd = splice(ctx->from, NULL, ctx->to, NULL, bufsiz, SPLICE_F_MOVE | SPLICE_F_MORE)) > 0) ; int err = (rd == -1) ? errno : 0; close(ctx->from); return ((void *)(uintptr_t)err); } #endif /* * Since Linux 5.10, 4d03e3cc59828c82ee89ea6e27a2f3cdf95aaadf * ("fs: don't allow kernel reads and writes without iter ops"), * ZFS_IOC_SEND* will EINVAL when writing to /dev/null, /dev/zero, &c. * * This wrapper transparently executes func() with a pipe * by spawning a thread to copy from that pipe to the original output * in the background. * * Returns the error from func(), if nonzero, * otherwise the error from the thread. * * No-op if orig_fd is -1, already a pipe (but the buffer size is bumped), * and on not-Linux; as such, it is safe to wrap/call wrapped functions * in a wrapped context. */ int lzc_send_wrapper(int (*func)(int, void *), int orig_fd, void *data) { #if __linux__ struct stat sb; if (orig_fd != -1 && fstat(orig_fd, &sb) == -1) return (errno); if (orig_fd == -1 || S_ISFIFO(sb.st_mode)) { if (orig_fd != -1) (void) max_pipe_buffer(orig_fd); return (func(orig_fd, data)); } if ((fcntl(orig_fd, F_GETFL) & O_ACCMODE) == O_RDONLY) return (errno = EBADF); int rw[2]; if (pipe2(rw, O_CLOEXEC) == -1) return (errno); int err; pthread_t send_thread; struct send_worker_ctx ctx = {.from = rw[0], .to = orig_fd}; if ((err = pthread_create(&send_thread, NULL, send_worker, &ctx)) != 0) { close(rw[0]); close(rw[1]); return (errno = err); } err = func(rw[1], data); void *send_err; close(rw[1]); pthread_join(send_thread, &send_err); if (err == 0 && send_err != 0) errno = err = (uintptr_t)send_err; return (err); #else return (func(orig_fd, data)); #endif } /* * Generate a zfs send stream for the specified snapshot and write it to * the specified file descriptor. * * "snapname" is the full name of the snapshot to send (e.g. "pool/fs@snap") * * If "from" is NULL, a full (non-incremental) stream will be sent. * If "from" is non-NULL, it must be the full name of a snapshot or * bookmark to send an incremental from (e.g. "pool/fs@earlier_snap" or * "pool/fs#earlier_bmark"). If non-NULL, the specified snapshot or * bookmark must represent an earlier point in the history of "snapname"). * It can be an earlier snapshot in the same filesystem or zvol as "snapname", * or it can be the origin of "snapname"'s filesystem, or an earlier * snapshot in the origin, etc. * * "fd" is the file descriptor to write the send stream to. * * If "flags" contains LZC_SEND_FLAG_LARGE_BLOCK, the stream is permitted * to contain DRR_WRITE records with drr_length > 128K, and DRR_OBJECT * records with drr_blksz > 128K. * * If "flags" contains LZC_SEND_FLAG_EMBED_DATA, the stream is permitted * to contain DRR_WRITE_EMBEDDED records with drr_etype==BP_EMBEDDED_TYPE_DATA, * which the receiving system must support (as indicated by support * for the "embedded_data" feature). * * If "flags" contains LZC_SEND_FLAG_COMPRESS, the stream is generated by using * compressed WRITE records for blocks which are compressed on disk and in * memory. If the lz4_compress feature is active on the sending system, then * the receiving system must have that feature enabled as well. * * If "flags" contains LZC_SEND_FLAG_RAW, the stream is generated, for encrypted * datasets, by sending data exactly as it exists on disk. This allows backups * to be taken even if encryption keys are not currently loaded. */ int lzc_send(const char *snapname, const char *from, int fd, enum lzc_send_flags flags) { return (lzc_send_resume_redacted(snapname, from, fd, flags, 0, 0, NULL)); } int lzc_send_redacted(const char *snapname, const char *from, int fd, enum lzc_send_flags flags, const char *redactbook) { return (lzc_send_resume_redacted(snapname, from, fd, flags, 0, 0, redactbook)); } int lzc_send_resume(const char *snapname, const char *from, int fd, enum lzc_send_flags flags, uint64_t resumeobj, uint64_t resumeoff) { return (lzc_send_resume_redacted(snapname, from, fd, flags, resumeobj, resumeoff, NULL)); } /* * snapname: The name of the "tosnap", or the snapshot whose contents we are * sending. * from: The name of the "fromsnap", or the incremental source. * fd: File descriptor to write the stream to. * flags: flags that determine features to be used by the stream. * resumeobj: Object to resume from, for resuming send * resumeoff: Offset to resume from, for resuming send. * redactnv: nvlist of string -> boolean(ignored) containing the names of all * the snapshots that we should redact with respect to. * redactbook: Name of the redaction bookmark to create. * * Pre-wrapped. */ static int lzc_send_resume_redacted_cb_impl(const char *snapname, const char *from, int fd, enum lzc_send_flags flags, uint64_t resumeobj, uint64_t resumeoff, const char *redactbook) { nvlist_t *args; int err; args = fnvlist_alloc(); fnvlist_add_int32(args, "fd", fd); if (from != NULL) fnvlist_add_string(args, "fromsnap", from); if (flags & LZC_SEND_FLAG_LARGE_BLOCK) fnvlist_add_boolean(args, "largeblockok"); if (flags & LZC_SEND_FLAG_EMBED_DATA) fnvlist_add_boolean(args, "embedok"); if (flags & LZC_SEND_FLAG_COMPRESS) fnvlist_add_boolean(args, "compressok"); if (flags & LZC_SEND_FLAG_RAW) fnvlist_add_boolean(args, "rawok"); if (flags & LZC_SEND_FLAG_SAVED) fnvlist_add_boolean(args, "savedok"); if (resumeobj != 0 || resumeoff != 0) { fnvlist_add_uint64(args, "resume_object", resumeobj); fnvlist_add_uint64(args, "resume_offset", resumeoff); } if (redactbook != NULL) fnvlist_add_string(args, "redactbook", redactbook); err = lzc_ioctl(ZFS_IOC_SEND_NEW, snapname, args, NULL); nvlist_free(args); return (err); } struct lzc_send_resume_redacted { const char *snapname; const char *from; enum lzc_send_flags flags; uint64_t resumeobj; uint64_t resumeoff; const char *redactbook; }; static int lzc_send_resume_redacted_cb(int fd, void *arg) { struct lzc_send_resume_redacted *zsrr = arg; return (lzc_send_resume_redacted_cb_impl(zsrr->snapname, zsrr->from, fd, zsrr->flags, zsrr->resumeobj, zsrr->resumeoff, zsrr->redactbook)); } int lzc_send_resume_redacted(const char *snapname, const char *from, int fd, enum lzc_send_flags flags, uint64_t resumeobj, uint64_t resumeoff, const char *redactbook) { struct lzc_send_resume_redacted zsrr = { .snapname = snapname, .from = from, .flags = flags, .resumeobj = resumeobj, .resumeoff = resumeoff, .redactbook = redactbook, }; return (lzc_send_wrapper(lzc_send_resume_redacted_cb, fd, &zsrr)); } /* * "from" can be NULL, a snapshot, or a bookmark. * * If from is NULL, a full (non-incremental) stream will be estimated. This * is calculated very efficiently. * * If from is a snapshot, lzc_send_space uses the deadlists attached to * each snapshot to efficiently estimate the stream size. * * If from is a bookmark, the indirect blocks in the destination snapshot * are traversed, looking for blocks with a birth time since the creation TXG of * the snapshot this bookmark was created from. This will result in * significantly more I/O and be less efficient than a send space estimation on * an equivalent snapshot. This process is also used if redact_snaps is * non-null. * * Pre-wrapped. */ static int lzc_send_space_resume_redacted_cb_impl(const char *snapname, const char *from, enum lzc_send_flags flags, uint64_t resumeobj, uint64_t resumeoff, uint64_t resume_bytes, const char *redactbook, int fd, uint64_t *spacep) { nvlist_t *args; nvlist_t *result; int err; args = fnvlist_alloc(); if (from != NULL) fnvlist_add_string(args, "from", from); if (flags & LZC_SEND_FLAG_LARGE_BLOCK) fnvlist_add_boolean(args, "largeblockok"); if (flags & LZC_SEND_FLAG_EMBED_DATA) fnvlist_add_boolean(args, "embedok"); if (flags & LZC_SEND_FLAG_COMPRESS) fnvlist_add_boolean(args, "compressok"); if (flags & LZC_SEND_FLAG_RAW) fnvlist_add_boolean(args, "rawok"); if (resumeobj != 0 || resumeoff != 0) { fnvlist_add_uint64(args, "resume_object", resumeobj); fnvlist_add_uint64(args, "resume_offset", resumeoff); fnvlist_add_uint64(args, "bytes", resume_bytes); } if (redactbook != NULL) fnvlist_add_string(args, "redactbook", redactbook); if (fd != -1) fnvlist_add_int32(args, "fd", fd); err = lzc_ioctl(ZFS_IOC_SEND_SPACE, snapname, args, &result); nvlist_free(args); if (err == 0) *spacep = fnvlist_lookup_uint64(result, "space"); nvlist_free(result); return (err); } struct lzc_send_space_resume_redacted { const char *snapname; const char *from; enum lzc_send_flags flags; uint64_t resumeobj; uint64_t resumeoff; uint64_t resume_bytes; const char *redactbook; uint64_t *spacep; }; static int lzc_send_space_resume_redacted_cb(int fd, void *arg) { struct lzc_send_space_resume_redacted *zssrr = arg; return (lzc_send_space_resume_redacted_cb_impl(zssrr->snapname, zssrr->from, zssrr->flags, zssrr->resumeobj, zssrr->resumeoff, zssrr->resume_bytes, zssrr->redactbook, fd, zssrr->spacep)); } int lzc_send_space_resume_redacted(const char *snapname, const char *from, enum lzc_send_flags flags, uint64_t resumeobj, uint64_t resumeoff, uint64_t resume_bytes, const char *redactbook, int fd, uint64_t *spacep) { struct lzc_send_space_resume_redacted zssrr = { .snapname = snapname, .from = from, .flags = flags, .resumeobj = resumeobj, .resumeoff = resumeoff, .resume_bytes = resume_bytes, .redactbook = redactbook, .spacep = spacep, }; return (lzc_send_wrapper(lzc_send_space_resume_redacted_cb, fd, &zssrr)); } int lzc_send_space(const char *snapname, const char *from, enum lzc_send_flags flags, uint64_t *spacep) { return (lzc_send_space_resume_redacted(snapname, from, flags, 0, 0, 0, NULL, -1, spacep)); } static int recv_read(int fd, void *buf, int ilen) { char *cp = buf; int rv; int len = ilen; do { rv = read(fd, cp, len); cp += rv; len -= rv; } while (rv > 0); if (rv < 0 || len != 0) return (EIO); return (0); } /* * Linux adds ZFS_IOC_RECV_NEW for resumable and raw streams and preserves the * legacy ZFS_IOC_RECV user/kernel interface. The new interface supports all * stream options but is currently only used for resumable streams. This way * updated user space utilities will interoperate with older kernel modules. * * Non-Linux OpenZFS platforms have opted to modify the legacy interface. */ static int recv_impl(const char *snapname, nvlist_t *recvdprops, nvlist_t *localprops, uint8_t *wkeydata, uint_t wkeylen, const char *origin, boolean_t force, boolean_t heal, boolean_t resumable, boolean_t raw, int input_fd, const dmu_replay_record_t *begin_record, uint64_t *read_bytes, uint64_t *errflags, nvlist_t **errors) { dmu_replay_record_t drr; char fsname[MAXPATHLEN]; char *atp; int error; boolean_t payload = B_FALSE; ASSERT3S(g_refcount, >, 0); VERIFY3S(g_fd, !=, -1); /* Set 'fsname' to the name of containing filesystem */ (void) strlcpy(fsname, snapname, sizeof (fsname)); atp = strchr(fsname, '@'); if (atp == NULL) return (EINVAL); *atp = '\0'; /* If the fs does not exist, try its parent. */ if (!lzc_exists(fsname)) { char *slashp = strrchr(fsname, '/'); if (slashp == NULL) return (ENOENT); *slashp = '\0'; } /* * It is not uncommon for gigabytes to be processed by zfs receive. * Speculatively increase the buffer size if supported by the platform. */ struct stat sb; if (fstat(input_fd, &sb) == -1) return (errno); if (S_ISFIFO(sb.st_mode)) (void) max_pipe_buffer(input_fd); /* * The begin_record is normally a non-byteswapped BEGIN record. * For resumable streams it may be set to any non-byteswapped * dmu_replay_record_t. */ if (begin_record == NULL) { error = recv_read(input_fd, &drr, sizeof (drr)); if (error != 0) return (error); } else { drr = *begin_record; payload = (begin_record->drr_payloadlen != 0); } /* * All receives with a payload should use the new interface. */ if (resumable || heal || raw || wkeydata != NULL || payload) { nvlist_t *outnvl = NULL; nvlist_t *innvl = fnvlist_alloc(); fnvlist_add_string(innvl, "snapname", snapname); if (recvdprops != NULL) fnvlist_add_nvlist(innvl, "props", recvdprops); if (localprops != NULL) fnvlist_add_nvlist(innvl, "localprops", localprops); if (wkeydata != NULL) { /* * wkeydata must be placed in the special * ZPOOL_HIDDEN_ARGS nvlist so that it * will not be printed to the zpool history. */ nvlist_t *hidden_args = fnvlist_alloc(); fnvlist_add_uint8_array(hidden_args, "wkeydata", wkeydata, wkeylen); fnvlist_add_nvlist(innvl, ZPOOL_HIDDEN_ARGS, hidden_args); nvlist_free(hidden_args); } if (origin != NULL && strlen(origin)) fnvlist_add_string(innvl, "origin", origin); fnvlist_add_byte_array(innvl, "begin_record", (uchar_t *)&drr, sizeof (drr)); fnvlist_add_int32(innvl, "input_fd", input_fd); if (force) fnvlist_add_boolean(innvl, "force"); if (resumable) fnvlist_add_boolean(innvl, "resumable"); if (heal) fnvlist_add_boolean(innvl, "heal"); error = lzc_ioctl(ZFS_IOC_RECV_NEW, fsname, innvl, &outnvl); if (error == 0 && read_bytes != NULL) error = nvlist_lookup_uint64(outnvl, "read_bytes", read_bytes); if (error == 0 && errflags != NULL) error = nvlist_lookup_uint64(outnvl, "error_flags", errflags); if (error == 0 && errors != NULL) { nvlist_t *nvl; error = nvlist_lookup_nvlist(outnvl, "errors", &nvl); if (error == 0) *errors = fnvlist_dup(nvl); } fnvlist_free(innvl); fnvlist_free(outnvl); } else { zfs_cmd_t zc = {"\0"}; char *rp_packed = NULL; char *lp_packed = NULL; size_t size; ASSERT3S(g_refcount, >, 0); (void) strlcpy(zc.zc_name, fsname, sizeof (zc.zc_name)); (void) strlcpy(zc.zc_value, snapname, sizeof (zc.zc_value)); if (recvdprops != NULL) { rp_packed = fnvlist_pack(recvdprops, &size); zc.zc_nvlist_src = (uint64_t)(uintptr_t)rp_packed; zc.zc_nvlist_src_size = size; } if (localprops != NULL) { lp_packed = fnvlist_pack(localprops, &size); zc.zc_nvlist_conf = (uint64_t)(uintptr_t)lp_packed; zc.zc_nvlist_conf_size = size; } if (origin != NULL) (void) strlcpy(zc.zc_string, origin, sizeof (zc.zc_string)); ASSERT3S(drr.drr_type, ==, DRR_BEGIN); zc.zc_begin_record = drr.drr_u.drr_begin; zc.zc_guid = force; zc.zc_cookie = input_fd; zc.zc_cleanup_fd = -1; zc.zc_action_handle = 0; zc.zc_nvlist_dst_size = 128 * 1024; zc.zc_nvlist_dst = (uint64_t)(uintptr_t) malloc(zc.zc_nvlist_dst_size); error = lzc_ioctl_fd(g_fd, ZFS_IOC_RECV, &zc); if (error != 0) { error = errno; } else { if (read_bytes != NULL) *read_bytes = zc.zc_cookie; if (errflags != NULL) *errflags = zc.zc_obj; if (errors != NULL) VERIFY0(nvlist_unpack( (void *)(uintptr_t)zc.zc_nvlist_dst, zc.zc_nvlist_dst_size, errors, KM_SLEEP)); } if (rp_packed != NULL) fnvlist_pack_free(rp_packed, size); if (lp_packed != NULL) fnvlist_pack_free(lp_packed, size); free((void *)(uintptr_t)zc.zc_nvlist_dst); } return (error); } /* * The simplest receive case: receive from the specified fd, creating the * specified snapshot. Apply the specified properties as "received" properties * (which can be overridden by locally-set properties). If the stream is a * clone, its origin snapshot must be specified by 'origin'. The 'force' * flag will cause the target filesystem to be rolled back or destroyed if * necessary to receive. * * Return 0 on success or an errno on failure. * * Note: this interface does not work on dedup'd streams * (those with DMU_BACKUP_FEATURE_DEDUP). */ int lzc_receive(const char *snapname, nvlist_t *props, const char *origin, boolean_t force, boolean_t raw, int fd) { return (recv_impl(snapname, props, NULL, NULL, 0, origin, force, B_FALSE, B_FALSE, raw, fd, NULL, NULL, NULL, NULL)); } /* * Like lzc_receive, but if the receive fails due to premature stream * termination, the intermediate state will be preserved on disk. In this * case, ECKSUM will be returned. The receive may subsequently be resumed * with a resuming send stream generated by lzc_send_resume(). */ int lzc_receive_resumable(const char *snapname, nvlist_t *props, const char *origin, boolean_t force, boolean_t raw, int fd) { return (recv_impl(snapname, props, NULL, NULL, 0, origin, force, B_FALSE, B_TRUE, raw, fd, NULL, NULL, NULL, NULL)); } /* * Like lzc_receive, but allows the caller to read the begin record and then to * pass it in. That could be useful if the caller wants to derive, for example, * the snapname or the origin parameters based on the information contained in * the begin record. * The begin record must be in its original form as read from the stream, * in other words, it should not be byteswapped. * * The 'resumable' parameter allows to obtain the same behavior as with * lzc_receive_resumable. */ int lzc_receive_with_header(const char *snapname, nvlist_t *props, const char *origin, boolean_t force, boolean_t resumable, boolean_t raw, int fd, const dmu_replay_record_t *begin_record) { if (begin_record == NULL) return (EINVAL); return (recv_impl(snapname, props, NULL, NULL, 0, origin, force, B_FALSE, resumable, raw, fd, begin_record, NULL, NULL, NULL)); } /* * Like lzc_receive, but allows the caller to pass all supported arguments * and retrieve all values returned. The only additional input parameter * is 'cleanup_fd' which is used to set a cleanup-on-exit file descriptor. * * The following parameters all provide return values. Several may be set * in the failure case and will contain additional information. * * The 'read_bytes' value will be set to the total number of bytes read. * * The 'errflags' value will contain zprop_errflags_t flags which are * used to describe any failures. * * The 'action_handle' and 'cleanup_fd' are no longer used, and are ignored. * * The 'errors' nvlist contains an entry for each unapplied received * property. Callers are responsible for freeing this nvlist. */ int lzc_receive_one(const char *snapname, nvlist_t *props, const char *origin, boolean_t force, boolean_t resumable, boolean_t raw, int input_fd, const dmu_replay_record_t *begin_record, int cleanup_fd, uint64_t *read_bytes, uint64_t *errflags, uint64_t *action_handle, nvlist_t **errors) { (void) action_handle, (void) cleanup_fd; return (recv_impl(snapname, props, NULL, NULL, 0, origin, force, B_FALSE, resumable, raw, input_fd, begin_record, read_bytes, errflags, errors)); } /* * Like lzc_receive_one, but allows the caller to pass an additional 'cmdprops' * argument. * * The 'cmdprops' nvlist contains both override ('zfs receive -o') and * exclude ('zfs receive -x') properties. Callers are responsible for freeing * this nvlist */ int lzc_receive_with_cmdprops(const char *snapname, nvlist_t *props, nvlist_t *cmdprops, uint8_t *wkeydata, uint_t wkeylen, const char *origin, boolean_t force, boolean_t resumable, boolean_t raw, int input_fd, const dmu_replay_record_t *begin_record, int cleanup_fd, uint64_t *read_bytes, uint64_t *errflags, uint64_t *action_handle, nvlist_t **errors) { (void) action_handle, (void) cleanup_fd; return (recv_impl(snapname, props, cmdprops, wkeydata, wkeylen, origin, force, B_FALSE, resumable, raw, input_fd, begin_record, read_bytes, errflags, errors)); } /* * Like lzc_receive_with_cmdprops, but allows the caller to pass an additional * 'heal' argument. * * The heal arguments tells us to heal the provided snapshot using the provided * send stream */ int lzc_receive_with_heal(const char *snapname, nvlist_t *props, nvlist_t *cmdprops, uint8_t *wkeydata, uint_t wkeylen, const char *origin, boolean_t force, boolean_t heal, boolean_t resumable, boolean_t raw, int input_fd, const dmu_replay_record_t *begin_record, int cleanup_fd, uint64_t *read_bytes, uint64_t *errflags, uint64_t *action_handle, nvlist_t **errors) { (void) action_handle, (void) cleanup_fd; return (recv_impl(snapname, props, cmdprops, wkeydata, wkeylen, origin, force, heal, resumable, raw, input_fd, begin_record, read_bytes, errflags, errors)); } /* * Roll back this filesystem or volume to its most recent snapshot. * If snapnamebuf is not NULL, it will be filled in with the name * of the most recent snapshot. * Note that the latest snapshot may change if a new one is concurrently * created or the current one is destroyed. lzc_rollback_to can be used * to roll back to a specific latest snapshot. * * Return 0 on success or an errno on failure. */ int lzc_rollback(const char *fsname, char *snapnamebuf, int snapnamelen) { nvlist_t *args; nvlist_t *result; int err; args = fnvlist_alloc(); err = lzc_ioctl(ZFS_IOC_ROLLBACK, fsname, args, &result); nvlist_free(args); if (err == 0 && snapnamebuf != NULL) { const char *snapname = fnvlist_lookup_string(result, "target"); (void) strlcpy(snapnamebuf, snapname, snapnamelen); } nvlist_free(result); return (err); } /* * Roll back this filesystem or volume to the specified snapshot, * if possible. * * Return 0 on success or an errno on failure. */ int lzc_rollback_to(const char *fsname, const char *snapname) { nvlist_t *args; nvlist_t *result; int err; args = fnvlist_alloc(); fnvlist_add_string(args, "target", snapname); err = lzc_ioctl(ZFS_IOC_ROLLBACK, fsname, args, &result); nvlist_free(args); nvlist_free(result); return (err); } /* * Creates new bookmarks from existing snapshot or bookmark. * * The bookmarks nvlist maps from the full name of the new bookmark to * the full name of the source snapshot or bookmark. * All the bookmarks and snapshots must be in the same pool. * The new bookmarks names must be unique. * => see function dsl_bookmark_create_nvl_validate * * The returned results nvlist will have an entry for each bookmark that failed. * The value will be the (int32) error code. * * The return value will be 0 if all bookmarks were created, otherwise it will * be the errno of a (undetermined) bookmarks that failed. */ int lzc_bookmark(nvlist_t *bookmarks, nvlist_t **errlist) { nvpair_t *elem; int error; char pool[ZFS_MAX_DATASET_NAME_LEN]; /* determine pool name from first bookmark */ elem = nvlist_next_nvpair(bookmarks, NULL); if (elem == NULL) return (0); (void) strlcpy(pool, nvpair_name(elem), sizeof (pool)); pool[strcspn(pool, "/#")] = '\0'; error = lzc_ioctl(ZFS_IOC_BOOKMARK, pool, bookmarks, errlist); return (error); } /* * Retrieve bookmarks. * * Retrieve the list of bookmarks for the given file system. The props * parameter is an nvlist of property names (with no values) that will be * returned for each bookmark. * * The following are valid properties on bookmarks, most of which are numbers * (represented as uint64 in the nvlist), except redact_snaps, which is a * uint64 array, and redact_complete, which is a boolean * * "guid" - globally unique identifier of the snapshot it refers to * "createtxg" - txg when the snapshot it refers to was created * "creation" - timestamp when the snapshot it refers to was created * "ivsetguid" - IVset guid for identifying encrypted snapshots * "redact_snaps" - list of guids of the redaction snapshots for the specified * bookmark. If the bookmark is not a redaction bookmark, the nvlist will * not contain an entry for this value. If it is redacted with respect to * no snapshots, it will contain value -> NULL uint64 array * "redact_complete" - boolean value; true if the redaction bookmark is * complete, false otherwise. * * The format of the returned nvlist as follows: * -> { * -> { * "value" -> uint64 * } * ... * "redact_snaps" -> { * "value" -> uint64 array * } * "redact_complete" -> { * "value" -> boolean value * } * } */ int lzc_get_bookmarks(const char *fsname, nvlist_t *props, nvlist_t **bmarks) { return (lzc_ioctl(ZFS_IOC_GET_BOOKMARKS, fsname, props, bmarks)); } /* * Get bookmark properties. * * Given a bookmark's full name, retrieve all properties for the bookmark. * * The format of the returned property list is as follows: * { * -> { * "value" -> uint64 * } * ... * "redact_snaps" -> { * "value" -> uint64 array * } */ int lzc_get_bookmark_props(const char *bookmark, nvlist_t **props) { int error; nvlist_t *innvl = fnvlist_alloc(); error = lzc_ioctl(ZFS_IOC_GET_BOOKMARK_PROPS, bookmark, innvl, props); fnvlist_free(innvl); return (error); } /* * Destroys bookmarks. * * The keys in the bmarks nvlist are the bookmarks to be destroyed. * They must all be in the same pool. Bookmarks are specified as * #. * * Bookmarks that do not exist will be silently ignored. * * The return value will be 0 if all bookmarks that existed were destroyed. * * Otherwise the return value will be the errno of a (undetermined) bookmark * that failed, no bookmarks will be destroyed, and the errlist will have an * entry for each bookmarks that failed. The value in the errlist will be * the (int32) error code. */ int lzc_destroy_bookmarks(nvlist_t *bmarks, nvlist_t **errlist) { nvpair_t *elem; int error; char pool[ZFS_MAX_DATASET_NAME_LEN]; /* determine the pool name */ elem = nvlist_next_nvpair(bmarks, NULL); if (elem == NULL) return (0); (void) strlcpy(pool, nvpair_name(elem), sizeof (pool)); pool[strcspn(pool, "/#")] = '\0'; error = lzc_ioctl(ZFS_IOC_DESTROY_BOOKMARKS, pool, bmarks, errlist); return (error); } static int lzc_channel_program_impl(const char *pool, const char *program, boolean_t sync, uint64_t instrlimit, uint64_t memlimit, nvlist_t *argnvl, nvlist_t **outnvl) { int error; nvlist_t *args; args = fnvlist_alloc(); fnvlist_add_string(args, ZCP_ARG_PROGRAM, program); fnvlist_add_nvlist(args, ZCP_ARG_ARGLIST, argnvl); fnvlist_add_boolean_value(args, ZCP_ARG_SYNC, sync); fnvlist_add_uint64(args, ZCP_ARG_INSTRLIMIT, instrlimit); fnvlist_add_uint64(args, ZCP_ARG_MEMLIMIT, memlimit); error = lzc_ioctl(ZFS_IOC_CHANNEL_PROGRAM, pool, args, outnvl); fnvlist_free(args); return (error); } /* * Executes a channel program. * * If this function returns 0 the channel program was successfully loaded and * ran without failing. Note that individual commands the channel program ran * may have failed and the channel program is responsible for reporting such * errors through outnvl if they are important. * * This method may also return: * * EINVAL The program contains syntax errors, or an invalid memory or time * limit was given. No part of the channel program was executed. * If caused by syntax errors, 'outnvl' contains information about the * errors. * * ECHRNG The program was executed, but encountered a runtime error, such as * calling a function with incorrect arguments, invoking the error() * function directly, failing an assert() command, etc. Some portion * of the channel program may have executed and committed changes. * Information about the failure can be found in 'outnvl'. * * ENOMEM The program fully executed, but the output buffer was not large * enough to store the returned value. No output is returned through * 'outnvl'. * * ENOSPC The program was terminated because it exceeded its memory usage * limit. Some portion of the channel program may have executed and * committed changes to disk. No output is returned through 'outnvl'. * * ETIME The program was terminated because it exceeded its Lua instruction * limit. Some portion of the channel program may have executed and * committed changes to disk. No output is returned through 'outnvl'. */ int lzc_channel_program(const char *pool, const char *program, uint64_t instrlimit, uint64_t memlimit, nvlist_t *argnvl, nvlist_t **outnvl) { return (lzc_channel_program_impl(pool, program, B_TRUE, instrlimit, memlimit, argnvl, outnvl)); } /* * Creates a checkpoint for the specified pool. * * If this function returns 0 the pool was successfully checkpointed. * * This method may also return: * * ZFS_ERR_CHECKPOINT_EXISTS * The pool already has a checkpoint. A pools can only have one * checkpoint at most, at any given time. * * ZFS_ERR_DISCARDING_CHECKPOINT * ZFS is in the middle of discarding a checkpoint for this pool. * The pool can be checkpointed again once the discard is done. * * ZFS_DEVRM_IN_PROGRESS * A vdev is currently being removed. The pool cannot be * checkpointed until the device removal is done. * * ZFS_VDEV_TOO_BIG * One or more top-level vdevs exceed the maximum vdev size * supported for this feature. */ int lzc_pool_checkpoint(const char *pool) { int error; nvlist_t *result = NULL; nvlist_t *args = fnvlist_alloc(); error = lzc_ioctl(ZFS_IOC_POOL_CHECKPOINT, pool, args, &result); fnvlist_free(args); fnvlist_free(result); return (error); } /* * Discard the checkpoint from the specified pool. * * If this function returns 0 the checkpoint was successfully discarded. * * This method may also return: * * ZFS_ERR_NO_CHECKPOINT * The pool does not have a checkpoint. * * ZFS_ERR_DISCARDING_CHECKPOINT * ZFS is already in the middle of discarding the checkpoint. */ int lzc_pool_checkpoint_discard(const char *pool) { int error; nvlist_t *result = NULL; nvlist_t *args = fnvlist_alloc(); error = lzc_ioctl(ZFS_IOC_POOL_DISCARD_CHECKPOINT, pool, args, &result); fnvlist_free(args); fnvlist_free(result); return (error); } /* * Executes a read-only channel program. * * A read-only channel program works programmatically the same way as a * normal channel program executed with lzc_channel_program(). The only * difference is it runs exclusively in open-context and therefore can * return faster. The downside to that, is that the program cannot change * on-disk state by calling functions from the zfs.sync submodule. * * The return values of this function (and their meaning) are exactly the * same as the ones described in lzc_channel_program(). */ int lzc_channel_program_nosync(const char *pool, const char *program, uint64_t timeout, uint64_t memlimit, nvlist_t *argnvl, nvlist_t **outnvl) { return (lzc_channel_program_impl(pool, program, B_FALSE, timeout, memlimit, argnvl, outnvl)); } int lzc_get_vdev_prop(const char *poolname, nvlist_t *innvl, nvlist_t **outnvl) { return (lzc_ioctl(ZFS_IOC_VDEV_GET_PROPS, poolname, innvl, outnvl)); } int lzc_set_vdev_prop(const char *poolname, nvlist_t *innvl, nvlist_t **outnvl) { return (lzc_ioctl(ZFS_IOC_VDEV_SET_PROPS, poolname, innvl, outnvl)); } /* * Performs key management functions * * crypto_cmd should be a value from dcp_cmd_t. If the command specifies to * load or change a wrapping key, the key should be specified in the * hidden_args nvlist so that it is not logged. */ int lzc_load_key(const char *fsname, boolean_t noop, uint8_t *wkeydata, uint_t wkeylen) { int error; nvlist_t *ioc_args; nvlist_t *hidden_args; if (wkeydata == NULL) return (EINVAL); ioc_args = fnvlist_alloc(); hidden_args = fnvlist_alloc(); fnvlist_add_uint8_array(hidden_args, "wkeydata", wkeydata, wkeylen); fnvlist_add_nvlist(ioc_args, ZPOOL_HIDDEN_ARGS, hidden_args); if (noop) fnvlist_add_boolean(ioc_args, "noop"); error = lzc_ioctl(ZFS_IOC_LOAD_KEY, fsname, ioc_args, NULL); nvlist_free(hidden_args); nvlist_free(ioc_args); return (error); } int lzc_unload_key(const char *fsname) { return (lzc_ioctl(ZFS_IOC_UNLOAD_KEY, fsname, NULL, NULL)); } int lzc_change_key(const char *fsname, uint64_t crypt_cmd, nvlist_t *props, uint8_t *wkeydata, uint_t wkeylen) { int error; nvlist_t *ioc_args = fnvlist_alloc(); nvlist_t *hidden_args = NULL; fnvlist_add_uint64(ioc_args, "crypt_cmd", crypt_cmd); if (wkeydata != NULL) { hidden_args = fnvlist_alloc(); fnvlist_add_uint8_array(hidden_args, "wkeydata", wkeydata, wkeylen); fnvlist_add_nvlist(ioc_args, ZPOOL_HIDDEN_ARGS, hidden_args); } if (props != NULL) fnvlist_add_nvlist(ioc_args, "props", props); error = lzc_ioctl(ZFS_IOC_CHANGE_KEY, fsname, ioc_args, NULL); nvlist_free(hidden_args); nvlist_free(ioc_args); return (error); } int lzc_reopen(const char *pool_name, boolean_t scrub_restart) { nvlist_t *args = fnvlist_alloc(); int error; fnvlist_add_boolean_value(args, "scrub_restart", scrub_restart); error = lzc_ioctl(ZFS_IOC_POOL_REOPEN, pool_name, args, NULL); nvlist_free(args); return (error); } /* * Changes initializing state. * * vdevs should be a list of (, guid) where guid is a uint64 vdev GUID. * The key is ignored. * * If there are errors related to vdev arguments, per-vdev errors are returned * in an nvlist with the key "vdevs". Each error is a (guid, errno) pair where * guid is stringified with PRIu64, and errno is one of the following as * an int64_t: * - ENODEV if the device was not found * - EINVAL if the devices is not a leaf or is not concrete (e.g. missing) * - EROFS if the device is not writeable * - EBUSY start requested but the device is already being either * initialized or trimmed * - ESRCH cancel/suspend requested but device is not being initialized * * If the errlist is empty, then return value will be: * - EINVAL if one or more arguments was invalid * - Other spa_open failures * - 0 if the operation succeeded */ int lzc_initialize(const char *poolname, pool_initialize_func_t cmd_type, nvlist_t *vdevs, nvlist_t **errlist) { int error; nvlist_t *args = fnvlist_alloc(); fnvlist_add_uint64(args, ZPOOL_INITIALIZE_COMMAND, (uint64_t)cmd_type); fnvlist_add_nvlist(args, ZPOOL_INITIALIZE_VDEVS, vdevs); error = lzc_ioctl(ZFS_IOC_POOL_INITIALIZE, poolname, args, errlist); fnvlist_free(args); return (error); } /* * Changes TRIM state. * * vdevs should be a list of (, guid) where guid is a uint64 vdev GUID. * The key is ignored. * * If there are errors related to vdev arguments, per-vdev errors are returned * in an nvlist with the key "vdevs". Each error is a (guid, errno) pair where * guid is stringified with PRIu64, and errno is one of the following as * an int64_t: * - ENODEV if the device was not found * - EINVAL if the devices is not a leaf or is not concrete (e.g. missing) * - EROFS if the device is not writeable * - EBUSY start requested but the device is already being either trimmed * or initialized * - ESRCH cancel/suspend requested but device is not being initialized * - EOPNOTSUPP if the device does not support TRIM (or secure TRIM) * * If the errlist is empty, then return value will be: * - EINVAL if one or more arguments was invalid * - Other spa_open failures * - 0 if the operation succeeded */ int lzc_trim(const char *poolname, pool_trim_func_t cmd_type, uint64_t rate, boolean_t secure, nvlist_t *vdevs, nvlist_t **errlist) { int error; nvlist_t *args = fnvlist_alloc(); fnvlist_add_uint64(args, ZPOOL_TRIM_COMMAND, (uint64_t)cmd_type); fnvlist_add_nvlist(args, ZPOOL_TRIM_VDEVS, vdevs); fnvlist_add_uint64(args, ZPOOL_TRIM_RATE, rate); fnvlist_add_boolean_value(args, ZPOOL_TRIM_SECURE, secure); error = lzc_ioctl(ZFS_IOC_POOL_TRIM, poolname, args, errlist); fnvlist_free(args); return (error); } /* * Create a redaction bookmark named bookname by redacting snapshot with respect * to all the snapshots in snapnv. */ int lzc_redact(const char *snapshot, const char *bookname, nvlist_t *snapnv) { nvlist_t *args = fnvlist_alloc(); fnvlist_add_string(args, "bookname", bookname); fnvlist_add_nvlist(args, "snapnv", snapnv); int error = lzc_ioctl(ZFS_IOC_REDACT, snapshot, args, NULL); fnvlist_free(args); return (error); } static int wait_common(const char *pool, zpool_wait_activity_t activity, boolean_t use_tag, uint64_t tag, boolean_t *waited) { nvlist_t *args = fnvlist_alloc(); nvlist_t *result = NULL; fnvlist_add_int32(args, ZPOOL_WAIT_ACTIVITY, activity); if (use_tag) fnvlist_add_uint64(args, ZPOOL_WAIT_TAG, tag); int error = lzc_ioctl(ZFS_IOC_WAIT, pool, args, &result); if (error == 0 && waited != NULL) *waited = fnvlist_lookup_boolean_value(result, ZPOOL_WAIT_WAITED); fnvlist_free(args); fnvlist_free(result); return (error); } int lzc_wait(const char *pool, zpool_wait_activity_t activity, boolean_t *waited) { return (wait_common(pool, activity, B_FALSE, 0, waited)); } int lzc_wait_tag(const char *pool, zpool_wait_activity_t activity, uint64_t tag, boolean_t *waited) { return (wait_common(pool, activity, B_TRUE, tag, waited)); } int lzc_wait_fs(const char *fs, zfs_wait_activity_t activity, boolean_t *waited) { nvlist_t *args = fnvlist_alloc(); nvlist_t *result = NULL; fnvlist_add_int32(args, ZFS_WAIT_ACTIVITY, activity); int error = lzc_ioctl(ZFS_IOC_WAIT_FS, fs, args, &result); if (error == 0 && waited != NULL) *waited = fnvlist_lookup_boolean_value(result, ZFS_WAIT_WAITED); fnvlist_free(args); fnvlist_free(result); return (error); } /* * Set the bootenv contents for the given pool. */ int lzc_set_bootenv(const char *pool, const nvlist_t *env) { return (lzc_ioctl(ZFS_IOC_SET_BOOTENV, pool, (nvlist_t *)env, NULL)); } /* * Get the contents of the bootenv of the given pool. */ int lzc_get_bootenv(const char *pool, nvlist_t **outnvl) { return (lzc_ioctl(ZFS_IOC_GET_BOOTENV, pool, NULL, outnvl)); } diff --git a/sys/contrib/openzfs/man/man4/zfs.4 b/sys/contrib/openzfs/man/man4/zfs.4 index d529147464fe..9ec940a94488 100644 --- a/sys/contrib/openzfs/man/man4/zfs.4 +++ b/sys/contrib/openzfs/man/man4/zfs.4 @@ -1,2593 +1,2596 @@ .\" .\" Copyright (c) 2013 by Turbo Fredriksson . All rights reserved. .\" Copyright (c) 2019, 2021 by Delphix. All rights reserved. .\" Copyright (c) 2019 Datto Inc. .\" The contents of this file are subject to the terms of the Common Development .\" and Distribution License (the "License"). You may not use this file except .\" in compliance with the License. You can obtain a copy of the license at .\" usr/src/OPENSOLARIS.LICENSE or https://opensource.org/licenses/CDDL-1.0. .\" .\" See the License for the specific language governing permissions and .\" limitations under the License. When distributing Covered Code, include this .\" CDDL HEADER in each file and include the License file at .\" usr/src/OPENSOLARIS.LICENSE. If applicable, add the following below this .\" CDDL HEADER, with the fields enclosed by brackets "[]" replaced with your .\" own identifying information: .\" Portions Copyright [yyyy] [name of copyright owner] .\" .Dd January 10, 2023 .Dt ZFS 4 .Os . .Sh NAME .Nm zfs .Nd tuning of the ZFS kernel module . .Sh DESCRIPTION The ZFS module supports these parameters: .Bl -tag -width Ds .It Sy dbuf_cache_max_bytes Ns = Ns Sy UINT64_MAX Ns B Pq u64 Maximum size in bytes of the dbuf cache. The target size is determined by the MIN versus .No 1/2^ Ns Sy dbuf_cache_shift Pq 1/32nd of the target ARC size. The behavior of the dbuf cache and its associated settings can be observed via the .Pa /proc/spl/kstat/zfs/dbufstats kstat. . .It Sy dbuf_metadata_cache_max_bytes Ns = Ns Sy UINT64_MAX Ns B Pq u64 Maximum size in bytes of the metadata dbuf cache. The target size is determined by the MIN versus .No 1/2^ Ns Sy dbuf_metadata_cache_shift Pq 1/64th of the target ARC size. The behavior of the metadata dbuf cache and its associated settings can be observed via the .Pa /proc/spl/kstat/zfs/dbufstats kstat. . .It Sy dbuf_cache_hiwater_pct Ns = Ns Sy 10 Ns % Pq uint The percentage over .Sy dbuf_cache_max_bytes when dbufs must be evicted directly. . .It Sy dbuf_cache_lowater_pct Ns = Ns Sy 10 Ns % Pq uint The percentage below .Sy dbuf_cache_max_bytes when the evict thread stops evicting dbufs. . .It Sy dbuf_cache_shift Ns = Ns Sy 5 Pq uint Set the size of the dbuf cache .Pq Sy dbuf_cache_max_bytes to a log2 fraction of the target ARC size. . .It Sy dbuf_metadata_cache_shift Ns = Ns Sy 6 Pq uint Set the size of the dbuf metadata cache .Pq Sy dbuf_metadata_cache_max_bytes to a log2 fraction of the target ARC size. . .It Sy dbuf_mutex_cache_shift Ns = Ns Sy 0 Pq uint Set the size of the mutex array for the dbuf cache. When set to .Sy 0 the array is dynamically sized based on total system memory. . .It Sy dmu_object_alloc_chunk_shift Ns = Ns Sy 7 Po 128 Pc Pq uint dnode slots allocated in a single operation as a power of 2. The default value minimizes lock contention for the bulk operation performed. . .It Sy dmu_prefetch_max Ns = Ns Sy 134217728 Ns B Po 128 MiB Pc Pq uint Limit the amount we can prefetch with one call to this amount in bytes. This helps to limit the amount of memory that can be used by prefetching. . .It Sy ignore_hole_birth Pq int Alias for .Sy send_holes_without_birth_time . . .It Sy l2arc_feed_again Ns = Ns Sy 1 Ns | Ns 0 Pq int Turbo L2ARC warm-up. When the L2ARC is cold the fill interval will be set as fast as possible. . .It Sy l2arc_feed_min_ms Ns = Ns Sy 200 Pq u64 Min feed interval in milliseconds. Requires .Sy l2arc_feed_again Ns = Ns Ar 1 and only applicable in related situations. . .It Sy l2arc_feed_secs Ns = Ns Sy 1 Pq u64 Seconds between L2ARC writing. . .It Sy l2arc_headroom Ns = Ns Sy 2 Pq u64 How far through the ARC lists to search for L2ARC cacheable content, expressed as a multiplier of .Sy l2arc_write_max . ARC persistence across reboots can be achieved with persistent L2ARC by setting this parameter to .Sy 0 , allowing the full length of ARC lists to be searched for cacheable content. . .It Sy l2arc_headroom_boost Ns = Ns Sy 200 Ns % Pq u64 Scales .Sy l2arc_headroom by this percentage when L2ARC contents are being successfully compressed before writing. A value of .Sy 100 disables this feature. . .It Sy l2arc_exclude_special Ns = Ns Sy 0 Ns | Ns 1 Pq int Controls whether buffers present on special vdevs are eligible for caching into L2ARC. If set to 1, exclude dbufs on special vdevs from being cached to L2ARC. . .It Sy l2arc_mfuonly Ns = Ns Sy 0 Ns | Ns 1 Pq int Controls whether only MFU metadata and data are cached from ARC into L2ARC. This may be desired to avoid wasting space on L2ARC when reading/writing large amounts of data that are not expected to be accessed more than once. .Pp The default is off, meaning both MRU and MFU data and metadata are cached. When turning off this feature, some MRU buffers will still be present in ARC and eventually cached on L2ARC. .No If Sy l2arc_noprefetch Ns = Ns Sy 0 , some prefetched buffers will be cached to L2ARC, and those might later transition to MRU, in which case the .Sy l2arc_mru_asize No arcstat will not be Sy 0 . .Pp Regardless of .Sy l2arc_noprefetch , some MFU buffers might be evicted from ARC, accessed later on as prefetches and transition to MRU as prefetches. If accessed again they are counted as MRU and the .Sy l2arc_mru_asize No arcstat will not be Sy 0 . .Pp The ARC status of L2ARC buffers when they were first cached in L2ARC can be seen in the .Sy l2arc_mru_asize , Sy l2arc_mfu_asize , No and Sy l2arc_prefetch_asize arcstats when importing the pool or onlining a cache device if persistent L2ARC is enabled. .Pp The .Sy evict_l2_eligible_mru arcstat does not take into account if this option is enabled as the information provided by the .Sy evict_l2_eligible_m[rf]u arcstats can be used to decide if toggling this option is appropriate for the current workload. . .It Sy l2arc_meta_percent Ns = Ns Sy 33 Ns % Pq uint Percent of ARC size allowed for L2ARC-only headers. Since L2ARC buffers are not evicted on memory pressure, too many headers on a system with an irrationally large L2ARC can render it slow or unusable. This parameter limits L2ARC writes and rebuilds to achieve the target. . .It Sy l2arc_trim_ahead Ns = Ns Sy 0 Ns % Pq u64 Trims ahead of the current write size .Pq Sy l2arc_write_max on L2ARC devices by this percentage of write size if we have filled the device. If set to .Sy 100 we TRIM twice the space required to accommodate upcoming writes. A minimum of .Sy 64 MiB will be trimmed. It also enables TRIM of the whole L2ARC device upon creation or addition to an existing pool or if the header of the device is invalid upon importing a pool or onlining a cache device. A value of .Sy 0 disables TRIM on L2ARC altogether and is the default as it can put significant stress on the underlying storage devices. This will vary depending of how well the specific device handles these commands. . .It Sy l2arc_noprefetch Ns = Ns Sy 1 Ns | Ns 0 Pq int Do not write buffers to L2ARC if they were prefetched but not used by applications. In case there are prefetched buffers in L2ARC and this option is later set, we do not read the prefetched buffers from L2ARC. Unsetting this option is useful for caching sequential reads from the disks to L2ARC and serve those reads from L2ARC later on. This may be beneficial in case the L2ARC device is significantly faster in sequential reads than the disks of the pool. .Pp Use .Sy 1 to disable and .Sy 0 to enable caching/reading prefetches to/from L2ARC. . .It Sy l2arc_norw Ns = Ns Sy 0 Ns | Ns 1 Pq int No reads during writes. . .It Sy l2arc_write_boost Ns = Ns Sy 8388608 Ns B Po 8 MiB Pc Pq u64 Cold L2ARC devices will have .Sy l2arc_write_max increased by this amount while they remain cold. . .It Sy l2arc_write_max Ns = Ns Sy 8388608 Ns B Po 8 MiB Pc Pq u64 Max write bytes per interval. . .It Sy l2arc_rebuild_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Rebuild the L2ARC when importing a pool (persistent L2ARC). This can be disabled if there are problems importing a pool or attaching an L2ARC device (e.g. the L2ARC device is slow in reading stored log metadata, or the metadata has become somehow fragmented/unusable). . .It Sy l2arc_rebuild_blocks_min_l2size Ns = Ns Sy 1073741824 Ns B Po 1 GiB Pc Pq u64 Mininum size of an L2ARC device required in order to write log blocks in it. The log blocks are used upon importing the pool to rebuild the persistent L2ARC. .Pp For L2ARC devices less than 1 GiB, the amount of data .Fn l2arc_evict evicts is significant compared to the amount of restored L2ARC data. In this case, do not write log blocks in L2ARC in order not to waste space. . .It Sy metaslab_aliquot Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64 Metaslab granularity, in bytes. This is roughly similar to what would be referred to as the "stripe size" in traditional RAID arrays. In normal operation, ZFS will try to write this amount of data to each disk before moving on to the next top-level vdev. . .It Sy metaslab_bias_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable metaslab group biasing based on their vdevs' over- or under-utilization relative to the pool. . .It Sy metaslab_force_ganging Ns = Ns Sy 16777217 Ns B Po 16 MiB + 1 B Pc Pq u64 Make some blocks above a certain size be gang blocks. This option is used by the test suite to facilitate testing. . .It Sy zfs_default_bs Ns = Ns Sy 9 Po 512 B Pc Pq int Default dnode block size as a power of 2. . .It Sy zfs_default_ibs Ns = Ns Sy 17 Po 128 KiB Pc Pq int Default dnode indirect block size as a power of 2. . .It Sy zfs_history_output_max Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64 When attempting to log an output nvlist of an ioctl in the on-disk history, the output will not be stored if it is larger than this size (in bytes). This must be less than .Sy DMU_MAX_ACCESS Pq 64 MiB . This applies primarily to .Fn zfs_ioc_channel_program Pq cf. Xr zfs-program 8 . . .It Sy zfs_keep_log_spacemaps_at_export Ns = Ns Sy 0 Ns | Ns 1 Pq int Prevent log spacemaps from being destroyed during pool exports and destroys. . .It Sy zfs_metaslab_segment_weight_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable/disable segment-based metaslab selection. . .It Sy zfs_metaslab_switch_threshold Ns = Ns Sy 2 Pq int When using segment-based metaslab selection, continue allocating from the active metaslab until this option's worth of buckets have been exhausted. . .It Sy metaslab_debug_load Ns = Ns Sy 0 Ns | Ns 1 Pq int Load all metaslabs during pool import. . .It Sy metaslab_debug_unload Ns = Ns Sy 0 Ns | Ns 1 Pq int Prevent metaslabs from being unloaded. . .It Sy metaslab_fragmentation_factor_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable use of the fragmentation metric in computing metaslab weights. . .It Sy metaslab_df_max_search Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq uint Maximum distance to search forward from the last offset. Without this limit, fragmented pools can see .Em >100`000 iterations and .Fn metaslab_block_picker becomes the performance limiting factor on high-performance storage. .Pp With the default setting of .Sy 16 MiB , we typically see less than .Em 500 iterations, even with very fragmented .Sy ashift Ns = Ns Sy 9 pools. The maximum number of iterations possible is .Sy metaslab_df_max_search / 2^(ashift+1) . With the default setting of .Sy 16 MiB this is .Em 16*1024 Pq with Sy ashift Ns = Ns Sy 9 or .Em 2*1024 Pq with Sy ashift Ns = Ns Sy 12 . . .It Sy metaslab_df_use_largest_segment Ns = Ns Sy 0 Ns | Ns 1 Pq int If not searching forward (due to .Sy metaslab_df_max_search , metaslab_df_free_pct , .No or Sy metaslab_df_alloc_threshold ) , this tunable controls which segment is used. If set, we will use the largest free segment. If unset, we will use a segment of at least the requested size. . .It Sy zfs_metaslab_max_size_cache_sec Ns = Ns Sy 3600 Ns s Po 1 hour Pc Pq u64 When we unload a metaslab, we cache the size of the largest free chunk. We use that cached size to determine whether or not to load a metaslab for a given allocation. As more frees accumulate in that metaslab while it's unloaded, the cached max size becomes less and less accurate. After a number of seconds controlled by this tunable, we stop considering the cached max size and start considering only the histogram instead. . .It Sy zfs_metaslab_mem_limit Ns = Ns Sy 25 Ns % Pq uint When we are loading a new metaslab, we check the amount of memory being used to store metaslab range trees. If it is over a threshold, we attempt to unload the least recently used metaslab to prevent the system from clogging all of its memory with range trees. This tunable sets the percentage of total system memory that is the threshold. . .It Sy zfs_metaslab_try_hard_before_gang Ns = Ns Sy 0 Ns | Ns 1 Pq int .Bl -item -compact .It If unset, we will first try normal allocation. .It If that fails then we will do a gang allocation. .It If that fails then we will do a "try hard" gang allocation. .It If that fails then we will have a multi-layer gang block. .El .Pp .Bl -item -compact .It If set, we will first try normal allocation. .It If that fails then we will do a "try hard" allocation. .It If that fails we will do a gang allocation. .It If that fails we will do a "try hard" gang allocation. .It If that fails then we will have a multi-layer gang block. .El . .It Sy zfs_metaslab_find_max_tries Ns = Ns Sy 100 Pq uint When not trying hard, we only consider this number of the best metaslabs. This improves performance, especially when there are many metaslabs per vdev and the allocation can't actually be satisfied (so we would otherwise iterate all metaslabs). . .It Sy zfs_vdev_default_ms_count Ns = Ns Sy 200 Pq uint When a vdev is added, target this number of metaslabs per top-level vdev. . .It Sy zfs_vdev_default_ms_shift Ns = Ns Sy 29 Po 512 MiB Pc Pq uint Default lower limit for metaslab size. . .It Sy zfs_vdev_max_ms_shift Ns = Ns Sy 34 Po 16 GiB Pc Pq uint Default upper limit for metaslab size. . .It Sy zfs_vdev_max_auto_ashift Ns = Ns Sy 14 Pq uint Maximum ashift used when optimizing for logical \[->] physical sector size on new top-level vdevs. May be increased up to .Sy ASHIFT_MAX Po 16 Pc , but this may negatively impact pool space efficiency. . .It Sy zfs_vdev_min_auto_ashift Ns = Ns Sy ASHIFT_MIN Po 9 Pc Pq uint Minimum ashift used when creating new top-level vdevs. . .It Sy zfs_vdev_min_ms_count Ns = Ns Sy 16 Pq uint Minimum number of metaslabs to create in a top-level vdev. . .It Sy vdev_validate_skip Ns = Ns Sy 0 Ns | Ns 1 Pq int Skip label validation steps during pool import. Changing is not recommended unless you know what you're doing and are recovering a damaged label. . .It Sy zfs_vdev_ms_count_limit Ns = Ns Sy 131072 Po 128k Pc Pq uint Practical upper limit of total metaslabs per top-level vdev. . .It Sy metaslab_preload_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable metaslab group preloading. . .It Sy metaslab_lba_weighting_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Give more weight to metaslabs with lower LBAs, assuming they have greater bandwidth, as is typically the case on a modern constant angular velocity disk drive. . .It Sy metaslab_unload_delay Ns = Ns Sy 32 Pq uint After a metaslab is used, we keep it loaded for this many TXGs, to attempt to reduce unnecessary reloading. Note that both this many TXGs and .Sy metaslab_unload_delay_ms milliseconds must pass before unloading will occur. . .It Sy metaslab_unload_delay_ms Ns = Ns Sy 600000 Ns ms Po 10 min Pc Pq uint After a metaslab is used, we keep it loaded for this many milliseconds, to attempt to reduce unnecessary reloading. Note, that both this many milliseconds and .Sy metaslab_unload_delay TXGs must pass before unloading will occur. . .It Sy reference_history Ns = Ns Sy 3 Pq uint Maximum reference holders being tracked when reference_tracking_enable is active. . .It Sy reference_tracking_enable Ns = Ns Sy 0 Ns | Ns 1 Pq int Track reference holders to .Sy refcount_t objects (debug builds only). . .It Sy send_holes_without_birth_time Ns = Ns Sy 1 Ns | Ns 0 Pq int When set, the .Sy hole_birth optimization will not be used, and all holes will always be sent during a .Nm zfs Cm send . This is useful if you suspect your datasets are affected by a bug in .Sy hole_birth . . .It Sy spa_config_path Ns = Ns Pa /etc/zfs/zpool.cache Pq charp SPA config file. . .It Sy spa_asize_inflation Ns = Ns Sy 24 Pq uint Multiplication factor used to estimate actual disk consumption from the size of data being written. The default value is a worst case estimate, but lower values may be valid for a given pool depending on its configuration. Pool administrators who understand the factors involved may wish to specify a more realistic inflation factor, particularly if they operate close to quota or capacity limits. . .It Sy spa_load_print_vdev_tree Ns = Ns Sy 0 Ns | Ns 1 Pq int Whether to print the vdev tree in the debugging message buffer during pool import. . .It Sy spa_load_verify_data Ns = Ns Sy 1 Ns | Ns 0 Pq int Whether to traverse data blocks during an "extreme rewind" .Pq Fl X import. .Pp An extreme rewind import normally performs a full traversal of all blocks in the pool for verification. If this parameter is unset, the traversal skips non-metadata blocks. It can be toggled once the import has started to stop or start the traversal of non-metadata blocks. . .It Sy spa_load_verify_metadata Ns = Ns Sy 1 Ns | Ns 0 Pq int Whether to traverse blocks during an "extreme rewind" .Pq Fl X pool import. .Pp An extreme rewind import normally performs a full traversal of all blocks in the pool for verification. If this parameter is unset, the traversal is not performed. It can be toggled once the import has started to stop or start the traversal. . .It Sy spa_load_verify_shift Ns = Ns Sy 4 Po 1/16th Pc Pq uint Sets the maximum number of bytes to consume during pool import to the log2 fraction of the target ARC size. . .It Sy spa_slop_shift Ns = Ns Sy 5 Po 1/32nd Pc Pq int Normally, we don't allow the last .Sy 3.2% Pq Sy 1/2^spa_slop_shift of space in the pool to be consumed. This ensures that we don't run the pool completely out of space, due to unaccounted changes (e.g. to the MOS). It also limits the worst-case time to allocate space. If we have less than this amount of free space, most ZPL operations (e.g. write, create) will return .Sy ENOSPC . . .It Sy spa_upgrade_errlog_limit Ns = Ns Sy 0 Pq uint Limits the number of on-disk error log entries that will be converted to the new format when enabling the .Sy head_errlog feature. The default is to convert all log entries. . .It Sy vdev_removal_max_span Ns = Ns Sy 32768 Ns B Po 32 KiB Pc Pq uint During top-level vdev removal, chunks of data are copied from the vdev which may include free space in order to trade bandwidth for IOPS. This parameter determines the maximum span of free space, in bytes, which will be included as "unnecessary" data in a chunk of copied data. .Pp The default value here was chosen to align with .Sy zfs_vdev_read_gap_limit , which is a similar concept when doing regular reads (but there's no reason it has to be the same). . .It Sy vdev_file_logical_ashift Ns = Ns Sy 9 Po 512 B Pc Pq u64 Logical ashift for file-based devices. . .It Sy vdev_file_physical_ashift Ns = Ns Sy 9 Po 512 B Pc Pq u64 Physical ashift for file-based devices. . .It Sy zap_iterate_prefetch Ns = Ns Sy 1 Ns | Ns 0 Pq int If set, when we start iterating over a ZAP object, prefetch the entire object (all leaf blocks). However, this is limited by .Sy dmu_prefetch_max . . .It Sy zap_micro_max_size Ns = Ns Sy 131072 Ns B Po 128 KiB Pc Pq int Maximum micro ZAP size. A micro ZAP is upgraded to a fat ZAP, once it grows beyond the specified size. . .It Sy zfetch_array_rd_sz Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64 If prefetching is enabled, disable prefetching for reads larger than this size. . .It Sy zfetch_min_distance Ns = Ns Sy 4194304 Ns B Po 4 MiB Pc Pq uint Min bytes to prefetch per stream. Prefetch distance starts from the demand access size and quickly grows to this value, doubling on each hit. After that it may grow further by 1/8 per hit, but only if some prefetch since last time haven't completed in time to satisfy demand request, i.e. prefetch depth didn't cover the read latency or the pool got saturated. . .It Sy zfetch_max_distance Ns = Ns Sy 67108864 Ns B Po 64 MiB Pc Pq uint Max bytes to prefetch per stream. . .It Sy zfetch_max_idistance Ns = Ns Sy 67108864 Ns B Po 64 MiB Pc Pq uint Max bytes to prefetch indirects for per stream. . .It Sy zfetch_max_streams Ns = Ns Sy 8 Pq uint Max number of streams per zfetch (prefetch streams per file). . .It Sy zfetch_min_sec_reap Ns = Ns Sy 1 Pq uint Min time before inactive prefetch stream can be reclaimed . .It Sy zfetch_max_sec_reap Ns = Ns Sy 2 Pq uint Max time before inactive prefetch stream can be deleted . .It Sy zfs_abd_scatter_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Enables ARC from using scatter/gather lists and forces all allocations to be linear in kernel memory. Disabling can improve performance in some code paths at the expense of fragmented kernel memory. . .It Sy zfs_abd_scatter_max_order Ns = Ns Sy MAX_ORDER\-1 Pq uint Maximum number of consecutive memory pages allocated in a single block for scatter/gather lists. .Pp The value of .Sy MAX_ORDER depends on kernel configuration. . .It Sy zfs_abd_scatter_min_size Ns = Ns Sy 1536 Ns B Po 1.5 KiB Pc Pq uint This is the minimum allocation size that will use scatter (page-based) ABDs. Smaller allocations will use linear ABDs. . .It Sy zfs_arc_dnode_limit Ns = Ns Sy 0 Ns B Pq u64 When the number of bytes consumed by dnodes in the ARC exceeds this number of bytes, try to unpin some of it in response to demand for non-metadata. This value acts as a ceiling to the amount of dnode metadata, and defaults to .Sy 0 , which indicates that a percent which is based on .Sy zfs_arc_dnode_limit_percent of the ARC meta buffers that may be used for dnodes. .It Sy zfs_arc_dnode_limit_percent Ns = Ns Sy 10 Ns % Pq u64 Percentage that can be consumed by dnodes of ARC meta buffers. .Pp See also .Sy zfs_arc_dnode_limit , which serves a similar purpose but has a higher priority if nonzero. . .It Sy zfs_arc_dnode_reduce_percent Ns = Ns Sy 10 Ns % Pq u64 Percentage of ARC dnodes to try to scan in response to demand for non-metadata when the number of bytes consumed by dnodes exceeds .Sy zfs_arc_dnode_limit . . .It Sy zfs_arc_average_blocksize Ns = Ns Sy 8192 Ns B Po 8 KiB Pc Pq uint The ARC's buffer hash table is sized based on the assumption of an average block size of this value. This works out to roughly 1 MiB of hash table per 1 GiB of physical memory with 8-byte pointers. For configurations with a known larger average block size, this value can be increased to reduce the memory footprint. . .It Sy zfs_arc_eviction_pct Ns = Ns Sy 200 Ns % Pq uint When .Fn arc_is_overflowing , .Fn arc_get_data_impl waits for this percent of the requested amount of data to be evicted. For example, by default, for every .Em 2 KiB that's evicted, .Em 1 KiB of it may be "reused" by a new allocation. Since this is above .Sy 100 Ns % , it ensures that progress is made towards getting .Sy arc_size No under Sy arc_c . Since this is finite, it ensures that allocations can still happen, even during the potentially long time that .Sy arc_size No is more than Sy arc_c . . .It Sy zfs_arc_evict_batch_limit Ns = Ns Sy 10 Pq uint Number ARC headers to evict per sub-list before proceeding to another sub-list. This batch-style operation prevents entire sub-lists from being evicted at once but comes at a cost of additional unlocking and locking. . .It Sy zfs_arc_grow_retry Ns = Ns Sy 0 Ns s Pq uint If set to a non zero value, it will replace the .Sy arc_grow_retry value with this value. The .Sy arc_grow_retry .No value Pq default Sy 5 Ns s is the number of seconds the ARC will wait before trying to resume growth after a memory pressure event. . .It Sy zfs_arc_lotsfree_percent Ns = Ns Sy 10 Ns % Pq int Throttle I/O when free system memory drops below this percentage of total system memory. Setting this value to .Sy 0 will disable the throttle. . .It Sy zfs_arc_max Ns = Ns Sy 0 Ns B Pq u64 Max size of ARC in bytes. If .Sy 0 , then the max size of ARC is determined by the amount of system memory installed. Under Linux, half of system memory will be used as the limit. Under .Fx , the larger of .Sy all_system_memory No \- Sy 1 GiB and .Sy 5/8 No \(mu Sy all_system_memory will be used as the limit. This value must be at least .Sy 67108864 Ns B Pq 64 MiB . .Pp This value can be changed dynamically, with some caveats. It cannot be set back to .Sy 0 while running, and reducing it below the current ARC size will not cause the ARC to shrink without memory pressure to induce shrinking. . .It Sy zfs_arc_meta_balance Ns = Ns Sy 500 Pq uint Balance between metadata and data on ghost hits. Values above 100 increase metadata caching by proportionally reducing effect of ghost data hits on target data/metadata rate. . .It Sy zfs_arc_min Ns = Ns Sy 0 Ns B Pq u64 Min size of ARC in bytes. .No If set to Sy 0 , arc_c_min will default to consuming the larger of .Sy 32 MiB and .Sy all_system_memory No / Sy 32 . . .It Sy zfs_arc_min_prefetch_ms Ns = Ns Sy 0 Ns ms Ns Po Ns ≡ Ns 1s Pc Pq uint Minimum time prefetched blocks are locked in the ARC. . .It Sy zfs_arc_min_prescient_prefetch_ms Ns = Ns Sy 0 Ns ms Ns Po Ns ≡ Ns 6s Pc Pq uint Minimum time "prescient prefetched" blocks are locked in the ARC. These blocks are meant to be prefetched fairly aggressively ahead of the code that may use them. . .It Sy zfs_arc_prune_task_threads Ns = Ns Sy 1 Pq int Number of arc_prune threads. .Fx does not need more than one. Linux may theoretically use one per mount point up to number of CPUs, but that was not proven to be useful. . .It Sy zfs_max_missing_tvds Ns = Ns Sy 0 Pq int Number of missing top-level vdevs which will be allowed during pool import (only in read-only mode). . .It Sy zfs_max_nvlist_src_size Ns = Sy 0 Pq u64 Maximum size in bytes allowed to be passed as .Sy zc_nvlist_src_size for ioctls on .Pa /dev/zfs . This prevents a user from causing the kernel to allocate an excessive amount of memory. When the limit is exceeded, the ioctl fails with .Sy EINVAL and a description of the error is sent to the .Pa zfs-dbgmsg log. This parameter should not need to be touched under normal circumstances. If .Sy 0 , equivalent to a quarter of the user-wired memory limit under .Fx and to .Sy 134217728 Ns B Pq 128 MiB under Linux. . .It Sy zfs_multilist_num_sublists Ns = Ns Sy 0 Pq uint To allow more fine-grained locking, each ARC state contains a series of lists for both data and metadata objects. Locking is performed at the level of these "sub-lists". This parameters controls the number of sub-lists per ARC state, and also applies to other uses of the multilist data structure. .Pp If .Sy 0 , equivalent to the greater of the number of online CPUs and .Sy 4 . . .It Sy zfs_arc_overflow_shift Ns = Ns Sy 8 Pq int The ARC size is considered to be overflowing if it exceeds the current ARC target size .Pq Sy arc_c by thresholds determined by this parameter. Exceeding by .Sy ( arc_c No >> Sy zfs_arc_overflow_shift ) No / Sy 2 starts ARC reclamation process. If that appears insufficient, exceeding by .Sy ( arc_c No >> Sy zfs_arc_overflow_shift ) No \(mu Sy 1.5 blocks new buffer allocation until the reclaim thread catches up. Started reclamation process continues till ARC size returns below the target size. .Pp The default value of .Sy 8 causes the ARC to start reclamation if it exceeds the target size by .Em 0.2% of the target size, and block allocations by .Em 0.6% . . .It Sy zfs_arc_shrink_shift Ns = Ns Sy 0 Pq uint If nonzero, this will update .Sy arc_shrink_shift Pq default Sy 7 with the new value. . .It Sy zfs_arc_pc_percent Ns = Ns Sy 0 Ns % Po off Pc Pq uint Percent of pagecache to reclaim ARC to. .Pp This tunable allows the ZFS ARC to play more nicely with the kernel's LRU pagecache. It can guarantee that the ARC size won't collapse under scanning pressure on the pagecache, yet still allows the ARC to be reclaimed down to .Sy zfs_arc_min if necessary. This value is specified as percent of pagecache size (as measured by .Sy NR_FILE_PAGES ) , where that percent may exceed .Sy 100 . This only operates during memory pressure/reclaim. . .It Sy zfs_arc_shrinker_limit Ns = Ns Sy 10000 Pq int This is a limit on how many pages the ARC shrinker makes available for eviction in response to one page allocation attempt. Note that in practice, the kernel's shrinker can ask us to evict up to about four times this for one allocation attempt. .Pp The default limit of .Sy 10000 Pq in practice, Em 160 MiB No per allocation attempt with 4 KiB pages limits the amount of time spent attempting to reclaim ARC memory to less than 100 ms per allocation attempt, even with a small average compressed block size of ~8 KiB. .Pp The parameter can be set to 0 (zero) to disable the limit, and only applies on Linux. . .It Sy zfs_arc_sys_free Ns = Ns Sy 0 Ns B Pq u64 The target number of bytes the ARC should leave as free memory on the system. If zero, equivalent to the bigger of .Sy 512 KiB No and Sy all_system_memory/64 . . .It Sy zfs_autoimport_disable Ns = Ns Sy 1 Ns | Ns 0 Pq int Disable pool import at module load by ignoring the cache file .Pq Sy spa_config_path . . .It Sy zfs_checksum_events_per_second Ns = Ns Sy 20 Ns /s Pq uint Rate limit checksum events to this many per second. Note that this should not be set below the ZED thresholds (currently 10 checksums over 10 seconds) or else the daemon may not trigger any action. . .It Sy zfs_commit_timeout_pct Ns = Ns Sy 5 Ns % Pq uint This controls the amount of time that a ZIL block (lwb) will remain "open" when it isn't "full", and it has a thread waiting for it to be committed to stable storage. The timeout is scaled based on a percentage of the last lwb latency to avoid significantly impacting the latency of each individual transaction record (itx). . .It Sy zfs_condense_indirect_commit_entry_delay_ms Ns = Ns Sy 0 Ns ms Pq int Vdev indirection layer (used for device removal) sleeps for this many milliseconds during mapping generation. Intended for use with the test suite to throttle vdev removal speed. . .It Sy zfs_condense_indirect_obsolete_pct Ns = Ns Sy 25 Ns % Pq uint Minimum percent of obsolete bytes in vdev mapping required to attempt to condense .Pq see Sy zfs_condense_indirect_vdevs_enable . Intended for use with the test suite to facilitate triggering condensing as needed. . .It Sy zfs_condense_indirect_vdevs_enable Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable condensing indirect vdev mappings. When set, attempt to condense indirect vdev mappings if the mapping uses more than .Sy zfs_condense_min_mapping_bytes bytes of memory and if the obsolete space map object uses more than .Sy zfs_condense_max_obsolete_bytes bytes on-disk. The condensing process is an attempt to save memory by removing obsolete mappings. . .It Sy zfs_condense_max_obsolete_bytes Ns = Ns Sy 1073741824 Ns B Po 1 GiB Pc Pq u64 Only attempt to condense indirect vdev mappings if the on-disk size of the obsolete space map object is greater than this number of bytes .Pq see Sy zfs_condense_indirect_vdevs_enable . . .It Sy zfs_condense_min_mapping_bytes Ns = Ns Sy 131072 Ns B Po 128 KiB Pc Pq u64 Minimum size vdev mapping to attempt to condense .Pq see Sy zfs_condense_indirect_vdevs_enable . . .It Sy zfs_dbgmsg_enable Ns = Ns Sy 1 Ns | Ns 0 Pq int Internally ZFS keeps a small log to facilitate debugging. The log is enabled by default, and can be disabled by unsetting this option. The contents of the log can be accessed by reading .Pa /proc/spl/kstat/zfs/dbgmsg . Writing .Sy 0 to the file clears the log. .Pp This setting does not influence debug prints due to .Sy zfs_flags . . .It Sy zfs_dbgmsg_maxsize Ns = Ns Sy 4194304 Ns B Po 4 MiB Pc Pq uint Maximum size of the internal ZFS debug log. . .It Sy zfs_dbuf_state_index Ns = Ns Sy 0 Pq int Historically used for controlling what reporting was available under .Pa /proc/spl/kstat/zfs . No effect. . .It Sy zfs_deadman_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int When a pool sync operation takes longer than .Sy zfs_deadman_synctime_ms , or when an individual I/O operation takes longer than .Sy zfs_deadman_ziotime_ms , then the operation is considered to be "hung". If .Sy zfs_deadman_enabled is set, then the deadman behavior is invoked as described by .Sy zfs_deadman_failmode . By default, the deadman is enabled and set to .Sy wait which results in "hung" I/O operations only being logged. The deadman is automatically disabled when a pool gets suspended. . .It Sy zfs_deadman_failmode Ns = Ns Sy wait Pq charp Controls the failure behavior when the deadman detects a "hung" I/O operation. Valid values are: .Bl -tag -compact -offset 4n -width "continue" .It Sy wait Wait for a "hung" operation to complete. For each "hung" operation a "deadman" event will be posted describing that operation. .It Sy continue Attempt to recover from a "hung" operation by re-dispatching it to the I/O pipeline if possible. .It Sy panic Panic the system. This can be used to facilitate automatic fail-over to a properly configured fail-over partner. .El . .It Sy zfs_deadman_checktime_ms Ns = Ns Sy 60000 Ns ms Po 1 min Pc Pq u64 Check time in milliseconds. This defines the frequency at which we check for hung I/O requests and potentially invoke the .Sy zfs_deadman_failmode behavior. . .It Sy zfs_deadman_synctime_ms Ns = Ns Sy 600000 Ns ms Po 10 min Pc Pq u64 Interval in milliseconds after which the deadman is triggered and also the interval after which a pool sync operation is considered to be "hung". Once this limit is exceeded the deadman will be invoked every .Sy zfs_deadman_checktime_ms milliseconds until the pool sync completes. . .It Sy zfs_deadman_ziotime_ms Ns = Ns Sy 300000 Ns ms Po 5 min Pc Pq u64 Interval in milliseconds after which the deadman is triggered and an individual I/O operation is considered to be "hung". As long as the operation remains "hung", the deadman will be invoked every .Sy zfs_deadman_checktime_ms milliseconds until the operation completes. . .It Sy zfs_dedup_prefetch Ns = Ns Sy 0 Ns | Ns 1 Pq int Enable prefetching dedup-ed blocks which are going to be freed. . .It Sy zfs_delay_min_dirty_percent Ns = Ns Sy 60 Ns % Pq uint Start to delay each transaction once there is this amount of dirty data, expressed as a percentage of .Sy zfs_dirty_data_max . This value should be at least .Sy zfs_vdev_async_write_active_max_dirty_percent . .No See Sx ZFS TRANSACTION DELAY . . .It Sy zfs_delay_scale Ns = Ns Sy 500000 Pq int This controls how quickly the transaction delay approaches infinity. Larger values cause longer delays for a given amount of dirty data. .Pp For the smoothest delay, this value should be about 1 billion divided by the maximum number of operations per second. This will smoothly handle between ten times and a tenth of this number. .No See Sx ZFS TRANSACTION DELAY . .Pp .Sy zfs_delay_scale No \(mu Sy zfs_dirty_data_max Em must No be smaller than Sy 2^64 . . .It Sy zfs_disable_ivset_guid_check Ns = Ns Sy 0 Ns | Ns 1 Pq int Disables requirement for IVset GUIDs to be present and match when doing a raw receive of encrypted datasets. Intended for users whose pools were created with OpenZFS pre-release versions and now have compatibility issues. . .It Sy zfs_key_max_salt_uses Ns = Ns Sy 400000000 Po 4*10^8 Pc Pq ulong Maximum number of uses of a single salt value before generating a new one for encrypted datasets. The default value is also the maximum. . .It Sy zfs_object_mutex_size Ns = Ns Sy 64 Pq uint Size of the znode hashtable used for holds. .Pp Due to the need to hold locks on objects that may not exist yet, kernel mutexes are not created per-object and instead a hashtable is used where collisions will result in objects waiting when there is not actually contention on the same object. . .It Sy zfs_slow_io_events_per_second Ns = Ns Sy 20 Ns /s Pq int Rate limit delay and deadman zevents (which report slow I/O operations) to this many per second. . .It Sy zfs_unflushed_max_mem_amt Ns = Ns Sy 1073741824 Ns B Po 1 GiB Pc Pq u64 Upper-bound limit for unflushed metadata changes to be held by the log spacemap in memory, in bytes. . .It Sy zfs_unflushed_max_mem_ppm Ns = Ns Sy 1000 Ns ppm Po 0.1% Pc Pq u64 Part of overall system memory that ZFS allows to be used for unflushed metadata changes by the log spacemap, in millionths. . .It Sy zfs_unflushed_log_block_max Ns = Ns Sy 131072 Po 128k Pc Pq u64 Describes the maximum number of log spacemap blocks allowed for each pool. The default value means that the space in all the log spacemaps can add up to no more than .Sy 131072 blocks (which means .Em 16 GiB of logical space before compression and ditto blocks, assuming that blocksize is .Em 128 KiB ) . .Pp This tunable is important because it involves a trade-off between import time after an unclean export and the frequency of flushing metaslabs. The higher this number is, the more log blocks we allow when the pool is active which means that we flush metaslabs less often and thus decrease the number of I/O operations for spacemap updates per TXG. At the same time though, that means that in the event of an unclean export, there will be more log spacemap blocks for us to read, inducing overhead in the import time of the pool. The lower the number, the amount of flushing increases, destroying log blocks quicker as they become obsolete faster, which leaves less blocks to be read during import time after a crash. .Pp Each log spacemap block existing during pool import leads to approximately one extra logical I/O issued. This is the reason why this tunable is exposed in terms of blocks rather than space used. . .It Sy zfs_unflushed_log_block_min Ns = Ns Sy 1000 Pq u64 If the number of metaslabs is small and our incoming rate is high, we could get into a situation that we are flushing all our metaslabs every TXG. Thus we always allow at least this many log blocks. . .It Sy zfs_unflushed_log_block_pct Ns = Ns Sy 400 Ns % Pq u64 Tunable used to determine the number of blocks that can be used for the spacemap log, expressed as a percentage of the total number of unflushed metaslabs in the pool. . .It Sy zfs_unflushed_log_txg_max Ns = Ns Sy 1000 Pq u64 Tunable limiting maximum time in TXGs any metaslab may remain unflushed. It effectively limits maximum number of unflushed per-TXG spacemap logs that need to be read after unclean pool export. . .It Sy zfs_unlink_suspend_progress Ns = Ns Sy 0 Ns | Ns 1 Pq uint When enabled, files will not be asynchronously removed from the list of pending unlinks and the space they consume will be leaked. Once this option has been disabled and the dataset is remounted, the pending unlinks will be processed and the freed space returned to the pool. This option is used by the test suite. . .It Sy zfs_delete_blocks Ns = Ns Sy 20480 Pq ulong This is the used to define a large file for the purposes of deletion. Files containing more than .Sy zfs_delete_blocks will be deleted asynchronously, while smaller files are deleted synchronously. Decreasing this value will reduce the time spent in an .Xr unlink 2 system call, at the expense of a longer delay before the freed space is available. This only applies on Linux. . .It Sy zfs_dirty_data_max Ns = Pq int Determines the dirty space limit in bytes. Once this limit is exceeded, new writes are halted until space frees up. This parameter takes precedence over .Sy zfs_dirty_data_max_percent . .No See Sx ZFS TRANSACTION DELAY . .Pp Defaults to .Sy physical_ram/10 , capped at .Sy zfs_dirty_data_max_max . . .It Sy zfs_dirty_data_max_max Ns = Pq int Maximum allowable value of .Sy zfs_dirty_data_max , expressed in bytes. This limit is only enforced at module load time, and will be ignored if .Sy zfs_dirty_data_max is later changed. This parameter takes precedence over .Sy zfs_dirty_data_max_max_percent . .No See Sx ZFS TRANSACTION DELAY . .Pp Defaults to .Sy min(physical_ram/4, 4GiB) , or .Sy min(physical_ram/4, 1GiB) for 32-bit systems. . .It Sy zfs_dirty_data_max_max_percent Ns = Ns Sy 25 Ns % Pq uint Maximum allowable value of .Sy zfs_dirty_data_max , expressed as a percentage of physical RAM. This limit is only enforced at module load time, and will be ignored if .Sy zfs_dirty_data_max is later changed. The parameter .Sy zfs_dirty_data_max_max takes precedence over this one. .No See Sx ZFS TRANSACTION DELAY . . .It Sy zfs_dirty_data_max_percent Ns = Ns Sy 10 Ns % Pq uint Determines the dirty space limit, expressed as a percentage of all memory. Once this limit is exceeded, new writes are halted until space frees up. The parameter .Sy zfs_dirty_data_max takes precedence over this one. .No See Sx ZFS TRANSACTION DELAY . .Pp Subject to .Sy zfs_dirty_data_max_max . . .It Sy zfs_dirty_data_sync_percent Ns = Ns Sy 20 Ns % Pq uint Start syncing out a transaction group if there's at least this much dirty data .Pq as a percentage of Sy zfs_dirty_data_max . This should be less than .Sy zfs_vdev_async_write_active_min_dirty_percent . . .It Sy zfs_wrlog_data_max Ns = Pq int The upper limit of write-transaction zil log data size in bytes. Write operations are throttled when approaching the limit until log data is cleared out after transaction group sync. Because of some overhead, it should be set at least 2 times the size of .Sy zfs_dirty_data_max .No to prevent harming normal write throughput . It also should be smaller than the size of the slog device if slog is present. .Pp Defaults to .Sy zfs_dirty_data_max*2 . .It Sy zfs_fallocate_reserve_percent Ns = Ns Sy 110 Ns % Pq uint Since ZFS is a copy-on-write filesystem with snapshots, blocks cannot be preallocated for a file in order to guarantee that later writes will not run out of space. Instead, .Xr fallocate 2 space preallocation only checks that sufficient space is currently available in the pool or the user's project quota allocation, and then creates a sparse file of the requested size. The requested space is multiplied by .Sy zfs_fallocate_reserve_percent to allow additional space for indirect blocks and other internal metadata. Setting this to .Sy 0 disables support for .Xr fallocate 2 and causes it to return .Sy EOPNOTSUPP . . .It Sy zfs_fletcher_4_impl Ns = Ns Sy fastest Pq string Select a fletcher 4 implementation. .Pp Supported selectors are: .Sy fastest , scalar , sse2 , ssse3 , avx2 , avx512f , avx512bw , .No and Sy aarch64_neon . All except .Sy fastest No and Sy scalar require instruction set extensions to be available, and will only appear if ZFS detects that they are present at runtime. If multiple implementations of fletcher 4 are available, the .Sy fastest will be chosen using a micro benchmark. Selecting .Sy scalar results in the original CPU-based calculation being used. Selecting any option other than .Sy fastest No or Sy scalar results in vector instructions from the respective CPU instruction set being used. . .It Sy zfs_blake3_impl Ns = Ns Sy fastest Pq string Select a BLAKE3 implementation. .Pp Supported selectors are: .Sy cycle , fastest , generic , sse2 , sse41 , avx2 , avx512 . All except .Sy cycle , fastest No and Sy generic require instruction set extensions to be available, and will only appear if ZFS detects that they are present at runtime. If multiple implementations of BLAKE3 are available, the .Sy fastest will be chosen using a micro benchmark. You can see the benchmark results by reading this kstat file: .Pa /proc/spl/kstat/zfs/chksum_bench . . .It Sy zfs_free_bpobj_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable/disable the processing of the free_bpobj object. . .It Sy zfs_async_block_max_blocks Ns = Ns Sy UINT64_MAX Po unlimited Pc Pq u64 Maximum number of blocks freed in a single TXG. . .It Sy zfs_max_async_dedup_frees Ns = Ns Sy 100000 Po 10^5 Pc Pq u64 Maximum number of dedup blocks freed in a single TXG. . .It Sy zfs_vdev_async_read_max_active Ns = Ns Sy 3 Pq uint Maximum asynchronous read I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_async_read_min_active Ns = Ns Sy 1 Pq uint Minimum asynchronous read I/O operation active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_async_write_active_max_dirty_percent Ns = Ns Sy 60 Ns % Pq uint When the pool has more than this much dirty data, use .Sy zfs_vdev_async_write_max_active to limit active async writes. If the dirty data is between the minimum and maximum, the active I/O limit is linearly interpolated. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_async_write_active_min_dirty_percent Ns = Ns Sy 30 Ns % Pq uint When the pool has less than this much dirty data, use .Sy zfs_vdev_async_write_min_active to limit active async writes. If the dirty data is between the minimum and maximum, the active I/O limit is linearly interpolated. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_async_write_max_active Ns = Ns Sy 10 Pq uint Maximum asynchronous write I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_async_write_min_active Ns = Ns Sy 2 Pq uint Minimum asynchronous write I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . .Pp Lower values are associated with better latency on rotational media but poorer resilver performance. The default value of .Sy 2 was chosen as a compromise. A value of .Sy 3 has been shown to improve resilver performance further at a cost of further increasing latency. . .It Sy zfs_vdev_initializing_max_active Ns = Ns Sy 1 Pq uint Maximum initializing I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_initializing_min_active Ns = Ns Sy 1 Pq uint Minimum initializing I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_max_active Ns = Ns Sy 1000 Pq uint The maximum number of I/O operations active to each device. Ideally, this will be at least the sum of each queue's .Sy max_active . .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_open_timeout_ms Ns = Ns Sy 1000 Pq uint Timeout value to wait before determining a device is missing during import. This is helpful for transient missing paths due to links being briefly removed and recreated in response to udev events. . .It Sy zfs_vdev_rebuild_max_active Ns = Ns Sy 3 Pq uint Maximum sequential resilver I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_rebuild_min_active Ns = Ns Sy 1 Pq uint Minimum sequential resilver I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_removal_max_active Ns = Ns Sy 2 Pq uint Maximum removal I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_removal_min_active Ns = Ns Sy 1 Pq uint Minimum removal I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_scrub_max_active Ns = Ns Sy 2 Pq uint Maximum scrub I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_scrub_min_active Ns = Ns Sy 1 Pq uint Minimum scrub I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_sync_read_max_active Ns = Ns Sy 10 Pq uint Maximum synchronous read I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_sync_read_min_active Ns = Ns Sy 10 Pq uint Minimum synchronous read I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_sync_write_max_active Ns = Ns Sy 10 Pq uint Maximum synchronous write I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_sync_write_min_active Ns = Ns Sy 10 Pq uint Minimum synchronous write I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_trim_max_active Ns = Ns Sy 2 Pq uint Maximum trim/discard I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_trim_min_active Ns = Ns Sy 1 Pq uint Minimum trim/discard I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_nia_delay Ns = Ns Sy 5 Pq uint For non-interactive I/O (scrub, resilver, removal, initialize and rebuild), the number of concurrently-active I/O operations is limited to .Sy zfs_*_min_active , unless the vdev is "idle". When there are no interactive I/O operations active (synchronous or otherwise), and .Sy zfs_vdev_nia_delay operations have completed since the last interactive operation, then the vdev is considered to be "idle", and the number of concurrently-active non-interactive operations is increased to .Sy zfs_*_max_active . .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_nia_credit Ns = Ns Sy 5 Pq uint Some HDDs tend to prioritize sequential I/O so strongly, that concurrent random I/O latency reaches several seconds. On some HDDs this happens even if sequential I/O operations are submitted one at a time, and so setting .Sy zfs_*_max_active Ns = Sy 1 does not help. To prevent non-interactive I/O, like scrub, from monopolizing the device, no more than .Sy zfs_vdev_nia_credit operations can be sent while there are outstanding incomplete interactive operations. This enforced wait ensures the HDD services the interactive I/O within a reasonable amount of time. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_queue_depth_pct Ns = Ns Sy 1000 Ns % Pq uint Maximum number of queued allocations per top-level vdev expressed as a percentage of .Sy zfs_vdev_async_write_max_active , which allows the system to detect devices that are more capable of handling allocations and to allocate more blocks to those devices. This allows for dynamic allocation distribution when devices are imbalanced, as fuller devices will tend to be slower than empty devices. .Pp Also see .Sy zio_dva_throttle_enabled . . .It Sy zfs_vdev_def_queue_depth Ns = Ns Sy 32 Pq uint Default queue depth for each vdev IO allocator. Higher values allow for better coalescing of sequential writes before sending them to the disk, but can increase transaction commit times. . .It Sy zfs_vdev_failfast_mask Ns = Ns Sy 1 Pq uint Defines if the driver should retire on a given error type. The following options may be bitwise-ored together: .TS box; lbz r l l . Value Name Description _ 1 Device No driver retries on device errors 2 Transport No driver retries on transport errors. 4 Driver No driver retries on driver errors. .TE . .It Sy zfs_expire_snapshot Ns = Ns Sy 300 Ns s Pq int Time before expiring .Pa .zfs/snapshot . . .It Sy zfs_admin_snapshot Ns = Ns Sy 0 Ns | Ns 1 Pq int Allow the creation, removal, or renaming of entries in the .Sy .zfs/snapshot directory to cause the creation, destruction, or renaming of snapshots. When enabled, this functionality works both locally and over NFS exports which have the .Em no_root_squash option set. . .It Sy zfs_flags Ns = Ns Sy 0 Pq int Set additional debugging flags. The following flags may be bitwise-ored together: .TS box; lbz r l l . Value Name Description _ 1 ZFS_DEBUG_DPRINTF Enable dprintf entries in the debug log. * 2 ZFS_DEBUG_DBUF_VERIFY Enable extra dbuf verifications. * 4 ZFS_DEBUG_DNODE_VERIFY Enable extra dnode verifications. 8 ZFS_DEBUG_SNAPNAMES Enable snapshot name verification. * 16 ZFS_DEBUG_MODIFY Check for illegally modified ARC buffers. 64 ZFS_DEBUG_ZIO_FREE Enable verification of block frees. 128 ZFS_DEBUG_HISTOGRAM_VERIFY Enable extra spacemap histogram verifications. 256 ZFS_DEBUG_METASLAB_VERIFY Verify space accounting on disk matches in-memory \fBrange_trees\fP. 512 ZFS_DEBUG_SET_ERROR Enable \fBSET_ERROR\fP and dprintf entries in the debug log. 1024 ZFS_DEBUG_INDIRECT_REMAP Verify split blocks created by device removal. 2048 ZFS_DEBUG_TRIM Verify TRIM ranges are always within the allocatable range tree. 4096 ZFS_DEBUG_LOG_SPACEMAP Verify that the log summary is consistent with the spacemap log and enable \fBzfs_dbgmsgs\fP for metaslab loading and flushing. .TE .Sy \& * No Requires debug build . . .It Sy zfs_btree_verify_intensity Ns = Ns Sy 0 Pq uint Enables btree verification. The following settings are culminative: .TS box; lbz r l l . Value Description 1 Verify height. 2 Verify pointers from children to parent. 3 Verify element counts. 4 Verify element order. (expensive) * 5 Verify unused memory is poisoned. (expensive) .TE .Sy \& * No Requires debug build . . .It Sy zfs_free_leak_on_eio Ns = Ns Sy 0 Ns | Ns 1 Pq int If destroy encounters an .Sy EIO while reading metadata (e.g. indirect blocks), space referenced by the missing metadata can not be freed. Normally this causes the background destroy to become "stalled", as it is unable to make forward progress. While in this stalled state, all remaining space to free from the error-encountering filesystem is "temporarily leaked". Set this flag to cause it to ignore the .Sy EIO , permanently leak the space from indirect blocks that can not be read, and continue to free everything else that it can. .Pp The default "stalling" behavior is useful if the storage partially fails (i.e. some but not all I/O operations fail), and then later recovers. In this case, we will be able to continue pool operations while it is partially failed, and when it recovers, we can continue to free the space, with no leaks. Note, however, that this case is actually fairly rare. .Pp Typically pools either .Bl -enum -compact -offset 4n -width "1." .It fail completely (but perhaps temporarily, e.g. due to a top-level vdev going offline), or .It have localized, permanent errors (e.g. disk returns the wrong data due to bit flip or firmware bug). .El In the former case, this setting does not matter because the pool will be suspended and the sync thread will not be able to make forward progress regardless. In the latter, because the error is permanent, the best we can do is leak the minimum amount of space, which is what setting this flag will do. It is therefore reasonable for this flag to normally be set, but we chose the more conservative approach of not setting it, so that there is no possibility of leaking space in the "partial temporary" failure case. . .It Sy zfs_free_min_time_ms Ns = Ns Sy 1000 Ns ms Po 1s Pc Pq uint During a .Nm zfs Cm destroy operation using the .Sy async_destroy feature, a minimum of this much time will be spent working on freeing blocks per TXG. . .It Sy zfs_obsolete_min_time_ms Ns = Ns Sy 500 Ns ms Pq uint Similar to .Sy zfs_free_min_time_ms , but for cleanup of old indirection records for removed vdevs. . .It Sy zfs_immediate_write_sz Ns = Ns Sy 32768 Ns B Po 32 KiB Pc Pq s64 Largest data block to write to the ZIL. Larger blocks will be treated as if the dataset being written to had the .Sy logbias Ns = Ns Sy throughput property set. . .It Sy zfs_initialize_value Ns = Ns Sy 16045690984833335022 Po 0xDEADBEEFDEADBEEE Pc Pq u64 Pattern written to vdev free space by .Xr zpool-initialize 8 . . .It Sy zfs_initialize_chunk_size Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64 Size of writes used by .Xr zpool-initialize 8 . This option is used by the test suite. . .It Sy zfs_livelist_max_entries Ns = Ns Sy 500000 Po 5*10^5 Pc Pq u64 The threshold size (in block pointers) at which we create a new sub-livelist. Larger sublists are more costly from a memory perspective but the fewer sublists there are, the lower the cost of insertion. . .It Sy zfs_livelist_min_percent_shared Ns = Ns Sy 75 Ns % Pq int If the amount of shared space between a snapshot and its clone drops below this threshold, the clone turns off the livelist and reverts to the old deletion method. This is in place because livelists no long give us a benefit once a clone has been overwritten enough. . .It Sy zfs_livelist_condense_new_alloc Ns = Ns Sy 0 Pq int Incremented each time an extra ALLOC blkptr is added to a livelist entry while it is being condensed. This option is used by the test suite to track race conditions. . .It Sy zfs_livelist_condense_sync_cancel Ns = Ns Sy 0 Pq int Incremented each time livelist condensing is canceled while in .Fn spa_livelist_condense_sync . This option is used by the test suite to track race conditions. . .It Sy zfs_livelist_condense_sync_pause Ns = Ns Sy 0 Ns | Ns 1 Pq int When set, the livelist condense process pauses indefinitely before executing the synctask \(em .Fn spa_livelist_condense_sync . This option is used by the test suite to trigger race conditions. . .It Sy zfs_livelist_condense_zthr_cancel Ns = Ns Sy 0 Pq int Incremented each time livelist condensing is canceled while in .Fn spa_livelist_condense_cb . This option is used by the test suite to track race conditions. . .It Sy zfs_livelist_condense_zthr_pause Ns = Ns Sy 0 Ns | Ns 1 Pq int When set, the livelist condense process pauses indefinitely before executing the open context condensing work in .Fn spa_livelist_condense_cb . This option is used by the test suite to trigger race conditions. . .It Sy zfs_lua_max_instrlimit Ns = Ns Sy 100000000 Po 10^8 Pc Pq u64 The maximum execution time limit that can be set for a ZFS channel program, specified as a number of Lua instructions. . .It Sy zfs_lua_max_memlimit Ns = Ns Sy 104857600 Po 100 MiB Pc Pq u64 The maximum memory limit that can be set for a ZFS channel program, specified in bytes. . .It Sy zfs_max_dataset_nesting Ns = Ns Sy 50 Pq int The maximum depth of nested datasets. This value can be tuned temporarily to fix existing datasets that exceed the predefined limit. . .It Sy zfs_max_log_walking Ns = Ns Sy 5 Pq u64 The number of past TXGs that the flushing algorithm of the log spacemap feature uses to estimate incoming log blocks. . .It Sy zfs_max_logsm_summary_length Ns = Ns Sy 10 Pq u64 Maximum number of rows allowed in the summary of the spacemap log. . .It Sy zfs_max_recordsize Ns = Ns Sy 16777216 Po 16 MiB Pc Pq uint We currently support block sizes from .Em 512 Po 512 B Pc No to Em 16777216 Po 16 MiB Pc . The benefits of larger blocks, and thus larger I/O, need to be weighed against the cost of COWing a giant block to modify one byte. Additionally, very large blocks can have an impact on I/O latency, and also potentially on the memory allocator. Therefore, we formerly forbade creating blocks larger than 1M. Larger blocks could be created by changing it, and pools with larger blocks can always be imported and used, regardless of this setting. . .It Sy zfs_allow_redacted_dataset_mount Ns = Ns Sy 0 Ns | Ns 1 Pq int Allow datasets received with redacted send/receive to be mounted. Normally disabled because these datasets may be missing key data. . .It Sy zfs_min_metaslabs_to_flush Ns = Ns Sy 1 Pq u64 Minimum number of metaslabs to flush per dirty TXG. . .It Sy zfs_metaslab_fragmentation_threshold Ns = Ns Sy 70 Ns % Pq uint Allow metaslabs to keep their active state as long as their fragmentation percentage is no more than this value. An active metaslab that exceeds this threshold will no longer keep its active status allowing better metaslabs to be selected. . .It Sy zfs_mg_fragmentation_threshold Ns = Ns Sy 95 Ns % Pq uint Metaslab groups are considered eligible for allocations if their fragmentation metric (measured as a percentage) is less than or equal to this value. If a metaslab group exceeds this threshold then it will be skipped unless all metaslab groups within the metaslab class have also crossed this threshold. . .It Sy zfs_mg_noalloc_threshold Ns = Ns Sy 0 Ns % Pq uint Defines a threshold at which metaslab groups should be eligible for allocations. The value is expressed as a percentage of free space beyond which a metaslab group is always eligible for allocations. If a metaslab group's free space is less than or equal to the threshold, the allocator will avoid allocating to that group unless all groups in the pool have reached the threshold. Once all groups have reached the threshold, all groups are allowed to accept allocations. The default value of .Sy 0 disables the feature and causes all metaslab groups to be eligible for allocations. .Pp This parameter allows one to deal with pools having heavily imbalanced vdevs such as would be the case when a new vdev has been added. Setting the threshold to a non-zero percentage will stop allocations from being made to vdevs that aren't filled to the specified percentage and allow lesser filled vdevs to acquire more allocations than they otherwise would under the old .Sy zfs_mg_alloc_failures facility. . .It Sy zfs_ddt_data_is_special Ns = Ns Sy 1 Ns | Ns 0 Pq int If enabled, ZFS will place DDT data into the special allocation class. . .It Sy zfs_user_indirect_is_special Ns = Ns Sy 1 Ns | Ns 0 Pq int If enabled, ZFS will place user data indirect blocks into the special allocation class. . .It Sy zfs_multihost_history Ns = Ns Sy 0 Pq uint Historical statistics for this many latest multihost updates will be available in .Pa /proc/spl/kstat/zfs/ Ns Ao Ar pool Ac Ns Pa /multihost . . .It Sy zfs_multihost_interval Ns = Ns Sy 1000 Ns ms Po 1 s Pc Pq u64 Used to control the frequency of multihost writes which are performed when the .Sy multihost pool property is on. This is one of the factors used to determine the length of the activity check during import. .Pp The multihost write period is .Sy zfs_multihost_interval No / Sy leaf-vdevs . On average a multihost write will be issued for each leaf vdev every .Sy zfs_multihost_interval milliseconds. In practice, the observed period can vary with the I/O load and this observed value is the delay which is stored in the uberblock. . .It Sy zfs_multihost_import_intervals Ns = Ns Sy 20 Pq uint Used to control the duration of the activity test on import. Smaller values of .Sy zfs_multihost_import_intervals will reduce the import time but increase the risk of failing to detect an active pool. The total activity check time is never allowed to drop below one second. .Pp On import the activity check waits a minimum amount of time determined by .Sy zfs_multihost_interval No \(mu Sy zfs_multihost_import_intervals , or the same product computed on the host which last had the pool imported, whichever is greater. The activity check time may be further extended if the value of MMP delay found in the best uberblock indicates actual multihost updates happened at longer intervals than .Sy zfs_multihost_interval . A minimum of .Em 100 ms is enforced. .Pp .Sy 0 No is equivalent to Sy 1 . . .It Sy zfs_multihost_fail_intervals Ns = Ns Sy 10 Pq uint Controls the behavior of the pool when multihost write failures or delays are detected. .Pp When .Sy 0 , multihost write failures or delays are ignored. The failures will still be reported to the ZED which depending on its configuration may take action such as suspending the pool or offlining a device. .Pp Otherwise, the pool will be suspended if .Sy zfs_multihost_fail_intervals No \(mu Sy zfs_multihost_interval milliseconds pass without a successful MMP write. This guarantees the activity test will see MMP writes if the pool is imported. .Sy 1 No is equivalent to Sy 2 ; this is necessary to prevent the pool from being suspended due to normal, small I/O latency variations. . .It Sy zfs_no_scrub_io Ns = Ns Sy 0 Ns | Ns 1 Pq int Set to disable scrub I/O. This results in scrubs not actually scrubbing data and simply doing a metadata crawl of the pool instead. . .It Sy zfs_no_scrub_prefetch Ns = Ns Sy 0 Ns | Ns 1 Pq int Set to disable block prefetching for scrubs. . .It Sy zfs_nocacheflush Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable cache flush operations on disks when writing. Setting this will cause pool corruption on power loss if a volatile out-of-order write cache is enabled. . .It Sy zfs_nopwrite_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Allow no-operation writes. The occurrence of nopwrites will further depend on other pool properties .Pq i.a. the checksumming and compression algorithms . . .It Sy zfs_dmu_offset_next_sync Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable forcing TXG sync to find holes. When enabled forces ZFS to sync data when .Sy SEEK_HOLE No or Sy SEEK_DATA flags are used allowing holes in a file to be accurately reported. When disabled holes will not be reported in recently dirtied files. . .It Sy zfs_pd_bytes_max Ns = Ns Sy 52428800 Ns B Po 50 MiB Pc Pq int The number of bytes which should be prefetched during a pool traversal, like .Nm zfs Cm send or other data crawling operations. . .It Sy zfs_traverse_indirect_prefetch_limit Ns = Ns Sy 32 Pq uint The number of blocks pointed by indirect (non-L0) block which should be prefetched during a pool traversal, like .Nm zfs Cm send or other data crawling operations. . .It Sy zfs_per_txg_dirty_frees_percent Ns = Ns Sy 30 Ns % Pq u64 Control percentage of dirtied indirect blocks from frees allowed into one TXG. After this threshold is crossed, additional frees will wait until the next TXG. .Sy 0 No disables this throttle . . .It Sy zfs_prefetch_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable predictive prefetch. Note that it leaves "prescient" prefetch .Pq for, e.g., Nm zfs Cm send intact. Unlike predictive prefetch, prescient prefetch never issues I/O that ends up not being needed, so it can't hurt performance. . .It Sy zfs_qat_checksum_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable QAT hardware acceleration for SHA256 checksums. May be unset after the ZFS modules have been loaded to initialize the QAT hardware as long as support is compiled in and the QAT driver is present. . .It Sy zfs_qat_compress_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable QAT hardware acceleration for gzip compression. May be unset after the ZFS modules have been loaded to initialize the QAT hardware as long as support is compiled in and the QAT driver is present. . .It Sy zfs_qat_encrypt_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable QAT hardware acceleration for AES-GCM encryption. May be unset after the ZFS modules have been loaded to initialize the QAT hardware as long as support is compiled in and the QAT driver is present. . .It Sy zfs_vnops_read_chunk_size Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64 Bytes to read per chunk. . .It Sy zfs_read_history Ns = Ns Sy 0 Pq uint Historical statistics for this many latest reads will be available in .Pa /proc/spl/kstat/zfs/ Ns Ao Ar pool Ac Ns Pa /reads . . .It Sy zfs_read_history_hits Ns = Ns Sy 0 Ns | Ns 1 Pq int Include cache hits in read history . .It Sy zfs_rebuild_max_segment Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64 Maximum read segment size to issue when sequentially resilvering a top-level vdev. . .It Sy zfs_rebuild_scrub_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Automatically start a pool scrub when the last active sequential resilver completes in order to verify the checksums of all blocks which have been resilvered. This is enabled by default and strongly recommended. . .It Sy zfs_rebuild_vdev_limit Ns = Ns Sy 67108864 Ns B Po 64 MiB Pc Pq u64 Maximum amount of I/O that can be concurrently issued for a sequential resilver per leaf device, given in bytes. . .It Sy zfs_reconstruct_indirect_combinations_max Ns = Ns Sy 4096 Pq int If an indirect split block contains more than this many possible unique combinations when being reconstructed, consider it too computationally expensive to check them all. Instead, try at most this many randomly selected combinations each time the block is accessed. This allows all segment copies to participate fairly in the reconstruction when all combinations cannot be checked and prevents repeated use of one bad copy. . .It Sy zfs_recover Ns = Ns Sy 0 Ns | Ns 1 Pq int Set to attempt to recover from fatal errors. This should only be used as a last resort, as it typically results in leaked space, or worse. . .It Sy zfs_removal_ignore_errors Ns = Ns Sy 0 Ns | Ns 1 Pq int Ignore hard I/O errors during device removal. When set, if a device encounters a hard I/O error during the removal process the removal will not be cancelled. This can result in a normally recoverable block becoming permanently damaged and is hence not recommended. This should only be used as a last resort when the pool cannot be returned to a healthy state prior to removing the device. . .It Sy zfs_removal_suspend_progress Ns = Ns Sy 0 Ns | Ns 1 Pq uint This is used by the test suite so that it can ensure that certain actions happen while in the middle of a removal. . .It Sy zfs_remove_max_segment Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq uint The largest contiguous segment that we will attempt to allocate when removing a device. If there is a performance problem with attempting to allocate large blocks, consider decreasing this. The default value is also the maximum. . .It Sy zfs_resilver_disable_defer Ns = Ns Sy 0 Ns | Ns 1 Pq int Ignore the .Sy resilver_defer feature, causing an operation that would start a resilver to immediately restart the one in progress. . .It Sy zfs_resilver_min_time_ms Ns = Ns Sy 3000 Ns ms Po 3 s Pc Pq uint Resilvers are processed by the sync thread. While resilvering, it will spend at least this much time working on a resilver between TXG flushes. . .It Sy zfs_scan_ignore_errors Ns = Ns Sy 0 Ns | Ns 1 Pq int If set, remove the DTL (dirty time list) upon completion of a pool scan (scrub), even if there were unrepairable errors. Intended to be used during pool repair or recovery to stop resilvering when the pool is next imported. . .It Sy zfs_scrub_min_time_ms Ns = Ns Sy 1000 Ns ms Po 1 s Pc Pq uint Scrubs are processed by the sync thread. While scrubbing, it will spend at least this much time working on a scrub between TXG flushes. . +.It Sy zfs_scrub_error_blocks_per_txg Ns = Ns Sy 4096 Pq uint +Error blocks to be scrubbed in one txg. +. .It Sy zfs_scan_checkpoint_intval Ns = Ns Sy 7200 Ns s Po 2 hour Pc Pq uint To preserve progress across reboots, the sequential scan algorithm periodically needs to stop metadata scanning and issue all the verification I/O to disk. The frequency of this flushing is determined by this tunable. . .It Sy zfs_scan_fill_weight Ns = Ns Sy 3 Pq uint This tunable affects how scrub and resilver I/O segments are ordered. A higher number indicates that we care more about how filled in a segment is, while a lower number indicates we care more about the size of the extent without considering the gaps within a segment. This value is only tunable upon module insertion. Changing the value afterwards will have no effect on scrub or resilver performance. . .It Sy zfs_scan_issue_strategy Ns = Ns Sy 0 Pq uint Determines the order that data will be verified while scrubbing or resilvering: .Bl -tag -compact -offset 4n -width "a" .It Sy 1 Data will be verified as sequentially as possible, given the amount of memory reserved for scrubbing .Pq see Sy zfs_scan_mem_lim_fact . This may improve scrub performance if the pool's data is very fragmented. .It Sy 2 The largest mostly-contiguous chunk of found data will be verified first. By deferring scrubbing of small segments, we may later find adjacent data to coalesce and increase the segment size. .It Sy 0 .No Use strategy Sy 1 No during normal verification .No and strategy Sy 2 No while taking a checkpoint . .El . .It Sy zfs_scan_legacy Ns = Ns Sy 0 Ns | Ns 1 Pq int If unset, indicates that scrubs and resilvers will gather metadata in memory before issuing sequential I/O. Otherwise indicates that the legacy algorithm will be used, where I/O is initiated as soon as it is discovered. Unsetting will not affect scrubs or resilvers that are already in progress. . .It Sy zfs_scan_max_ext_gap Ns = Ns Sy 2097152 Ns B Po 2 MiB Pc Pq int Sets the largest gap in bytes between scrub/resilver I/O operations that will still be considered sequential for sorting purposes. Changing this value will not affect scrubs or resilvers that are already in progress. . .It Sy zfs_scan_mem_lim_fact Ns = Ns Sy 20 Ns ^-1 Pq uint Maximum fraction of RAM used for I/O sorting by sequential scan algorithm. This tunable determines the hard limit for I/O sorting memory usage. When the hard limit is reached we stop scanning metadata and start issuing data verification I/O. This is done until we get below the soft limit. . .It Sy zfs_scan_mem_lim_soft_fact Ns = Ns Sy 20 Ns ^-1 Pq uint The fraction of the hard limit used to determined the soft limit for I/O sorting by the sequential scan algorithm. When we cross this limit from below no action is taken. When we cross this limit from above it is because we are issuing verification I/O. In this case (unless the metadata scan is done) we stop issuing verification I/O and start scanning metadata again until we get to the hard limit. . .It Sy zfs_scan_report_txgs Ns = Ns Sy 0 Ns | Ns 1 Pq uint When reporting resilver throughput and estimated completion time use the performance observed over roughly the last .Sy zfs_scan_report_txgs TXGs. When set to zero performance is calculated over the time between checkpoints. . .It Sy zfs_scan_strict_mem_lim Ns = Ns Sy 0 Ns | Ns 1 Pq int Enforce tight memory limits on pool scans when a sequential scan is in progress. When disabled, the memory limit may be exceeded by fast disks. . .It Sy zfs_scan_suspend_progress Ns = Ns Sy 0 Ns | Ns 1 Pq int Freezes a scrub/resilver in progress without actually pausing it. Intended for testing/debugging. . .It Sy zfs_scan_vdev_limit Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq int Maximum amount of data that can be concurrently issued at once for scrubs and resilvers per leaf device, given in bytes. . .It Sy zfs_send_corrupt_data Ns = Ns Sy 0 Ns | Ns 1 Pq int Allow sending of corrupt data (ignore read/checksum errors when sending). . .It Sy zfs_send_unmodified_spill_blocks Ns = Ns Sy 1 Ns | Ns 0 Pq int Include unmodified spill blocks in the send stream. Under certain circumstances, previous versions of ZFS could incorrectly remove the spill block from an existing object. Including unmodified copies of the spill blocks creates a backwards-compatible stream which will recreate a spill block if it was incorrectly removed. . .It Sy zfs_send_no_prefetch_queue_ff Ns = Ns Sy 20 Ns ^\-1 Pq uint The fill fraction of the .Nm zfs Cm send internal queues. The fill fraction controls the timing with which internal threads are woken up. . .It Sy zfs_send_no_prefetch_queue_length Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq uint The maximum number of bytes allowed in .Nm zfs Cm send Ns 's internal queues. . .It Sy zfs_send_queue_ff Ns = Ns Sy 20 Ns ^\-1 Pq uint The fill fraction of the .Nm zfs Cm send prefetch queue. The fill fraction controls the timing with which internal threads are woken up. . .It Sy zfs_send_queue_length Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq uint The maximum number of bytes allowed that will be prefetched by .Nm zfs Cm send . This value must be at least twice the maximum block size in use. . .It Sy zfs_recv_queue_ff Ns = Ns Sy 20 Ns ^\-1 Pq uint The fill fraction of the .Nm zfs Cm receive queue. The fill fraction controls the timing with which internal threads are woken up. . .It Sy zfs_recv_queue_length Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq uint The maximum number of bytes allowed in the .Nm zfs Cm receive queue. This value must be at least twice the maximum block size in use. . .It Sy zfs_recv_write_batch_size Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq uint The maximum amount of data, in bytes, that .Nm zfs Cm receive will write in one DMU transaction. This is the uncompressed size, even when receiving a compressed send stream. This setting will not reduce the write size below a single block. Capped at a maximum of .Sy 32 MiB . . .It Sy zfs_recv_best_effort_corrective Ns = Ns Sy 0 Pq int When this variable is set to non-zero a corrective receive: .Bl -enum -compact -offset 4n -width "1." .It Does not enforce the restriction of source & destination snapshot GUIDs matching. .It If there is an error during healing, the healing receive is not terminated instead it moves on to the next record. .El . .It Sy zfs_override_estimate_recordsize Ns = Ns Sy 0 Ns | Ns 1 Pq uint Setting this variable overrides the default logic for estimating block sizes when doing a .Nm zfs Cm send . The default heuristic is that the average block size will be the current recordsize. Override this value if most data in your dataset is not of that size and you require accurate zfs send size estimates. . .It Sy zfs_sync_pass_deferred_free Ns = Ns Sy 2 Pq uint Flushing of data to disk is done in passes. Defer frees starting in this pass. . .It Sy zfs_spa_discard_memory_limit Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq int Maximum memory used for prefetching a checkpoint's space map on each vdev while discarding the checkpoint. . .It Sy zfs_special_class_metadata_reserve_pct Ns = Ns Sy 25 Ns % Pq uint Only allow small data blocks to be allocated on the special and dedup vdev types when the available free space percentage on these vdevs exceeds this value. This ensures reserved space is available for pool metadata as the special vdevs approach capacity. . .It Sy zfs_sync_pass_dont_compress Ns = Ns Sy 8 Pq uint Starting in this sync pass, disable compression (including of metadata). With the default setting, in practice, we don't have this many sync passes, so this has no effect. .Pp The original intent was that disabling compression would help the sync passes to converge. However, in practice, disabling compression increases the average number of sync passes; because when we turn compression off, many blocks' size will change, and thus we have to re-allocate (not overwrite) them. It also increases the number of .Em 128 KiB allocations (e.g. for indirect blocks and spacemaps) because these will not be compressed. The .Em 128 KiB allocations are especially detrimental to performance on highly fragmented systems, which may have very few free segments of this size, and may need to load new metaslabs to satisfy these allocations. . .It Sy zfs_sync_pass_rewrite Ns = Ns Sy 2 Pq uint Rewrite new block pointers starting in this pass. . .It Sy zfs_sync_taskq_batch_pct Ns = Ns Sy 75 Ns % Pq int This controls the number of threads used by .Sy dp_sync_taskq . The default value of .Sy 75% will create a maximum of one thread per CPU. . .It Sy zfs_trim_extent_bytes_max Ns = Ns Sy 134217728 Ns B Po 128 MiB Pc Pq uint Maximum size of TRIM command. Larger ranges will be split into chunks no larger than this value before issuing. . .It Sy zfs_trim_extent_bytes_min Ns = Ns Sy 32768 Ns B Po 32 KiB Pc Pq uint Minimum size of TRIM commands. TRIM ranges smaller than this will be skipped, unless they're part of a larger range which was chunked. This is done because it's common for these small TRIMs to negatively impact overall performance. . .It Sy zfs_trim_metaslab_skip Ns = Ns Sy 0 Ns | Ns 1 Pq uint Skip uninitialized metaslabs during the TRIM process. This option is useful for pools constructed from large thinly-provisioned devices where TRIM operations are slow. As a pool ages, an increasing fraction of the pool's metaslabs will be initialized, progressively degrading the usefulness of this option. This setting is stored when starting a manual TRIM and will persist for the duration of the requested TRIM. . .It Sy zfs_trim_queue_limit Ns = Ns Sy 10 Pq uint Maximum number of queued TRIMs outstanding per leaf vdev. The number of concurrent TRIM commands issued to the device is controlled by .Sy zfs_vdev_trim_min_active No and Sy zfs_vdev_trim_max_active . . .It Sy zfs_trim_txg_batch Ns = Ns Sy 32 Pq uint The number of transaction groups' worth of frees which should be aggregated before TRIM operations are issued to the device. This setting represents a trade-off between issuing larger, more efficient TRIM operations and the delay before the recently trimmed space is available for use by the device. .Pp Increasing this value will allow frees to be aggregated for a longer time. This will result is larger TRIM operations and potentially increased memory usage. Decreasing this value will have the opposite effect. The default of .Sy 32 was determined to be a reasonable compromise. . .It Sy zfs_txg_history Ns = Ns Sy 0 Pq uint Historical statistics for this many latest TXGs will be available in .Pa /proc/spl/kstat/zfs/ Ns Ao Ar pool Ac Ns Pa /TXGs . . .It Sy zfs_txg_timeout Ns = Ns Sy 5 Ns s Pq uint Flush dirty data to disk at least every this many seconds (maximum TXG duration). . .It Sy zfs_vdev_aggregate_trim Ns = Ns Sy 0 Ns | Ns 1 Pq uint Allow TRIM I/O operations to be aggregated. This is normally not helpful because the extents to be trimmed will have been already been aggregated by the metaslab. This option is provided for debugging and performance analysis. . .It Sy zfs_vdev_aggregation_limit Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq uint Max vdev I/O aggregation size. . .It Sy zfs_vdev_aggregation_limit_non_rotating Ns = Ns Sy 131072 Ns B Po 128 KiB Pc Pq uint Max vdev I/O aggregation size for non-rotating media. . .It Sy zfs_vdev_cache_bshift Ns = Ns Sy 16 Po 64 KiB Pc Pq uint Shift size to inflate reads to. . .It Sy zfs_vdev_cache_max Ns = Ns Sy 16384 Ns B Po 16 KiB Pc Pq uint Inflate reads smaller than this value to meet the .Sy zfs_vdev_cache_bshift size .Pq default Sy 64 KiB . . .It Sy zfs_vdev_cache_size Ns = Ns Sy 0 Pq uint Total size of the per-disk cache in bytes. .Pp Currently this feature is disabled, as it has been found to not be helpful for performance and in some cases harmful. . .It Sy zfs_vdev_mirror_rotating_inc Ns = Ns Sy 0 Pq int A number by which the balancing algorithm increments the load calculation for the purpose of selecting the least busy mirror member when an I/O operation immediately follows its predecessor on rotational vdevs for the purpose of making decisions based on load. . .It Sy zfs_vdev_mirror_rotating_seek_inc Ns = Ns Sy 5 Pq int A number by which the balancing algorithm increments the load calculation for the purpose of selecting the least busy mirror member when an I/O operation lacks locality as defined by .Sy zfs_vdev_mirror_rotating_seek_offset . Operations within this that are not immediately following the previous operation are incremented by half. . .It Sy zfs_vdev_mirror_rotating_seek_offset Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq int The maximum distance for the last queued I/O operation in which the balancing algorithm considers an operation to have locality. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_mirror_non_rotating_inc Ns = Ns Sy 0 Pq int A number by which the balancing algorithm increments the load calculation for the purpose of selecting the least busy mirror member on non-rotational vdevs when I/O operations do not immediately follow one another. . .It Sy zfs_vdev_mirror_non_rotating_seek_inc Ns = Ns Sy 1 Pq int A number by which the balancing algorithm increments the load calculation for the purpose of selecting the least busy mirror member when an I/O operation lacks locality as defined by the .Sy zfs_vdev_mirror_rotating_seek_offset . Operations within this that are not immediately following the previous operation are incremented by half. . .It Sy zfs_vdev_read_gap_limit Ns = Ns Sy 32768 Ns B Po 32 KiB Pc Pq uint Aggregate read I/O operations if the on-disk gap between them is within this threshold. . .It Sy zfs_vdev_write_gap_limit Ns = Ns Sy 4096 Ns B Po 4 KiB Pc Pq uint Aggregate write I/O operations if the on-disk gap between them is within this threshold. . .It Sy zfs_vdev_raidz_impl Ns = Ns Sy fastest Pq string Select the raidz parity implementation to use. .Pp Variants that don't depend on CPU-specific features may be selected on module load, as they are supported on all systems. The remaining options may only be set after the module is loaded, as they are available only if the implementations are compiled in and supported on the running system. .Pp Once the module is loaded, .Pa /sys/module/zfs/parameters/zfs_vdev_raidz_impl will show the available options, with the currently selected one enclosed in square brackets. .Pp .TS lb l l . fastest selected by built-in benchmark original original implementation scalar scalar implementation sse2 SSE2 instruction set 64-bit x86 ssse3 SSSE3 instruction set 64-bit x86 avx2 AVX2 instruction set 64-bit x86 avx512f AVX512F instruction set 64-bit x86 avx512bw AVX512F & AVX512BW instruction sets 64-bit x86 aarch64_neon NEON Aarch64/64-bit ARMv8 aarch64_neonx2 NEON with more unrolling Aarch64/64-bit ARMv8 powerpc_altivec Altivec PowerPC .TE . .It Sy zfs_vdev_scheduler Pq charp .Sy DEPRECATED . Prints warning to kernel log for compatibility. . .It Sy zfs_zevent_len_max Ns = Ns Sy 512 Pq uint Max event queue length. Events in the queue can be viewed with .Xr zpool-events 8 . . .It Sy zfs_zevent_retain_max Ns = Ns Sy 2000 Pq int Maximum recent zevent records to retain for duplicate checking. Setting this to .Sy 0 disables duplicate detection. . .It Sy zfs_zevent_retain_expire_secs Ns = Ns Sy 900 Ns s Po 15 min Pc Pq int Lifespan for a recent ereport that was retained for duplicate checking. . .It Sy zfs_zil_clean_taskq_maxalloc Ns = Ns Sy 1048576 Pq int The maximum number of taskq entries that are allowed to be cached. When this limit is exceeded transaction records (itxs) will be cleaned synchronously. . .It Sy zfs_zil_clean_taskq_minalloc Ns = Ns Sy 1024 Pq int The number of taskq entries that are pre-populated when the taskq is first created and are immediately available for use. . .It Sy zfs_zil_clean_taskq_nthr_pct Ns = Ns Sy 100 Ns % Pq int This controls the number of threads used by .Sy dp_zil_clean_taskq . The default value of .Sy 100% will create a maximum of one thread per cpu. . .It Sy zil_maxblocksize Ns = Ns Sy 131072 Ns B Po 128 KiB Pc Pq uint This sets the maximum block size used by the ZIL. On very fragmented pools, lowering this .Pq typically to Sy 36 KiB can improve performance. . .It Sy zil_min_commit_timeout Ns = Ns Sy 5000 Pq u64 This sets the minimum delay in nanoseconds ZIL care to delay block commit, waiting for more records. If ZIL writes are too fast, kernel may not be able sleep for so short interval, increasing log latency above allowed by .Sy zfs_commit_timeout_pct . . .It Sy zil_nocacheflush Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable the cache flush commands that are normally sent to disk by the ZIL after an LWB write has completed. Setting this will cause ZIL corruption on power loss if a volatile out-of-order write cache is enabled. . .It Sy zil_replay_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable intent logging replay. Can be disabled for recovery from corrupted ZIL. . .It Sy zil_slog_bulk Ns = Ns Sy 786432 Ns B Po 768 KiB Pc Pq u64 Limit SLOG write size per commit executed with synchronous priority. Any writes above that will be executed with lower (asynchronous) priority to limit potential SLOG device abuse by single active ZIL writer. . .It Sy zfs_zil_saxattr Ns = Ns Sy 1 Ns | Ns 0 Pq int Setting this tunable to zero disables ZIL logging of new .Sy xattr Ns = Ns Sy sa records if the .Sy org.openzfs:zilsaxattr feature is enabled on the pool. This would only be necessary to work around bugs in the ZIL logging or replay code for this record type. The tunable has no effect if the feature is disabled. . .It Sy zfs_embedded_slog_min_ms Ns = Ns Sy 64 Pq uint Usually, one metaslab from each normal-class vdev is dedicated for use by the ZIL to log synchronous writes. However, if there are fewer than .Sy zfs_embedded_slog_min_ms metaslabs in the vdev, this functionality is disabled. This ensures that we don't set aside an unreasonable amount of space for the ZIL. . .It Sy zstd_earlyabort_pass Ns = Ns Sy 1 Pq uint Whether heuristic for detection of incompressible data with zstd levels >= 3 using LZ4 and zstd-1 passes is enabled. . .It Sy zstd_abort_size Ns = Ns Sy 131072 Pq uint Minimal uncompressed size (inclusive) of a record before the early abort heuristic will be attempted. . .It Sy zio_deadman_log_all Ns = Ns Sy 0 Ns | Ns 1 Pq int If non-zero, the zio deadman will produce debugging messages .Pq see Sy zfs_dbgmsg_enable for all zios, rather than only for leaf zios possessing a vdev. This is meant to be used by developers to gain diagnostic information for hang conditions which don't involve a mutex or other locking primitive: typically conditions in which a thread in the zio pipeline is looping indefinitely. . .It Sy zio_slow_io_ms Ns = Ns Sy 30000 Ns ms Po 30 s Pc Pq int When an I/O operation takes more than this much time to complete, it's marked as slow. Each slow operation causes a delay zevent. Slow I/O counters can be seen with .Nm zpool Cm status Fl s . . .It Sy zio_dva_throttle_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Throttle block allocations in the I/O pipeline. This allows for dynamic allocation distribution when devices are imbalanced. When enabled, the maximum number of pending allocations per top-level vdev is limited by .Sy zfs_vdev_queue_depth_pct . . .It Sy zfs_xattr_compat Ns = Ns 0 Ns | Ns 1 Pq int Control the naming scheme used when setting new xattrs in the user namespace. If .Sy 0 .Pq the default on Linux , user namespace xattr names are prefixed with the namespace, to be backwards compatible with previous versions of ZFS on Linux. If .Sy 1 .Pq the default on Fx , user namespace xattr names are not prefixed, to be backwards compatible with previous versions of ZFS on illumos and .Fx . .Pp Either naming scheme can be read on this and future versions of ZFS, regardless of this tunable, but legacy ZFS on illumos or .Fx are unable to read user namespace xattrs written in the Linux format, and legacy versions of ZFS on Linux are unable to read user namespace xattrs written in the legacy ZFS format. .Pp An existing xattr with the alternate naming scheme is removed when overwriting the xattr so as to not accumulate duplicates. . .It Sy zio_requeue_io_start_cut_in_line Ns = Ns Sy 0 Ns | Ns 1 Pq int Prioritize requeued I/O. . .It Sy zio_taskq_batch_pct Ns = Ns Sy 80 Ns % Pq uint Percentage of online CPUs which will run a worker thread for I/O. These workers are responsible for I/O work such as compression and checksum calculations. Fractional number of CPUs will be rounded down. .Pp The default value of .Sy 80% was chosen to avoid using all CPUs which can result in latency issues and inconsistent application performance, especially when slower compression and/or checksumming is enabled. . .It Sy zio_taskq_batch_tpq Ns = Ns Sy 0 Pq uint Number of worker threads per taskq. Lower values improve I/O ordering and CPU utilization, while higher reduces lock contention. .Pp If .Sy 0 , generate a system-dependent value close to 6 threads per taskq. . .It Sy zvol_inhibit_dev Ns = Ns Sy 0 Ns | Ns 1 Pq uint Do not create zvol device nodes. This may slightly improve startup time on systems with a very large number of zvols. . .It Sy zvol_major Ns = Ns Sy 230 Pq uint Major number for zvol block devices. . .It Sy zvol_max_discard_blocks Ns = Ns Sy 16384 Pq long Discard (TRIM) operations done on zvols will be done in batches of this many blocks, where block size is determined by the .Sy volblocksize property of a zvol. . .It Sy zvol_prefetch_bytes Ns = Ns Sy 131072 Ns B Po 128 KiB Pc Pq uint When adding a zvol to the system, prefetch this many bytes from the start and end of the volume. Prefetching these regions of the volume is desirable, because they are likely to be accessed immediately by .Xr blkid 8 or the kernel partitioner. . .It Sy zvol_request_sync Ns = Ns Sy 0 Ns | Ns 1 Pq uint When processing I/O requests for a zvol, submit them synchronously. This effectively limits the queue depth to .Em 1 for each I/O submitter. When unset, requests are handled asynchronously by a thread pool. The number of requests which can be handled concurrently is controlled by .Sy zvol_threads . .Sy zvol_request_sync is ignored when running on a kernel that supports block multiqueue .Pq Li blk-mq . . .It Sy zvol_threads Ns = Ns Sy 0 Pq uint The number of system wide threads to use for processing zvol block IOs. If .Sy 0 (the default) then internally set .Sy zvol_threads to the number of CPUs present or 32 (whichever is greater). . .It Sy zvol_blk_mq_threads Ns = Ns Sy 0 Pq uint The number of threads per zvol to use for queuing IO requests. This parameter will only appear if your kernel supports .Li blk-mq and is only read and assigned to a zvol at zvol load time. If .Sy 0 (the default) then internally set .Sy zvol_blk_mq_threads to the number of CPUs present. . .It Sy zvol_use_blk_mq Ns = Ns Sy 0 Ns | Ns 1 Pq uint Set to .Sy 1 to use the .Li blk-mq API for zvols. Set to .Sy 0 (the default) to use the legacy zvol APIs. This setting can give better or worse zvol performance depending on the workload. This parameter will only appear if your kernel supports .Li blk-mq and is only read and assigned to a zvol at zvol load time. . .It Sy zvol_blk_mq_blocks_per_thread Ns = Ns Sy 8 Pq uint If .Sy zvol_use_blk_mq is enabled, then process this number of .Sy volblocksize Ns -sized blocks per zvol thread. This tunable can be use to favor better performance for zvol reads (lower values) or writes (higher values). If set to .Sy 0 , then the zvol layer will process the maximum number of blocks per thread that it can. This parameter will only appear if your kernel supports .Li blk-mq and is only applied at each zvol's load time. . .It Sy zvol_blk_mq_queue_depth Ns = Ns Sy 0 Pq uint The queue_depth value for the zvol .Li blk-mq interface. This parameter will only appear if your kernel supports .Li blk-mq and is only applied at each zvol's load time. If .Sy 0 (the default) then use the kernel's default queue depth. Values are clamped to the kernel's .Dv BLKDEV_MIN_RQ and .Dv BLKDEV_MAX_RQ Ns / Ns Dv BLKDEV_DEFAULT_RQ limits. . .It Sy zvol_volmode Ns = Ns Sy 1 Pq uint Defines zvol block devices behaviour when .Sy volmode Ns = Ns Sy default : .Bl -tag -compact -offset 4n -width "a" .It Sy 1 .No equivalent to Sy full .It Sy 2 .No equivalent to Sy dev .It Sy 3 .No equivalent to Sy none .El . .It Sy zvol_enforce_quotas Ns = Ns Sy 0 Ns | Ns 1 Pq uint Enable strict ZVOL quota enforcement. The strict quota enforcement may have a performance impact. .El . .Sh ZFS I/O SCHEDULER ZFS issues I/O operations to leaf vdevs to satisfy and complete I/O operations. The scheduler determines when and in what order those operations are issued. The scheduler divides operations into five I/O classes, prioritized in the following order: sync read, sync write, async read, async write, and scrub/resilver. Each queue defines the minimum and maximum number of concurrent operations that may be issued to the device. In addition, the device has an aggregate maximum, .Sy zfs_vdev_max_active . Note that the sum of the per-queue minima must not exceed the aggregate maximum. If the sum of the per-queue maxima exceeds the aggregate maximum, then the number of active operations may reach .Sy zfs_vdev_max_active , in which case no further operations will be issued, regardless of whether all per-queue minima have been met. .Pp For many physical devices, throughput increases with the number of concurrent operations, but latency typically suffers. Furthermore, physical devices typically have a limit at which more concurrent operations have no effect on throughput or can actually cause it to decrease. .Pp The scheduler selects the next operation to issue by first looking for an I/O class whose minimum has not been satisfied. Once all are satisfied and the aggregate maximum has not been hit, the scheduler looks for classes whose maximum has not been satisfied. Iteration through the I/O classes is done in the order specified above. No further operations are issued if the aggregate maximum number of concurrent operations has been hit, or if there are no operations queued for an I/O class that has not hit its maximum. Every time an I/O operation is queued or an operation completes, the scheduler looks for new operations to issue. .Pp In general, smaller .Sy max_active Ns s will lead to lower latency of synchronous operations. Larger .Sy max_active Ns s may lead to higher overall throughput, depending on underlying storage. .Pp The ratio of the queues' .Sy max_active Ns s determines the balance of performance between reads, writes, and scrubs. For example, increasing .Sy zfs_vdev_scrub_max_active will cause the scrub or resilver to complete more quickly, but reads and writes to have higher latency and lower throughput. .Pp All I/O classes have a fixed maximum number of outstanding operations, except for the async write class. Asynchronous writes represent the data that is committed to stable storage during the syncing stage for transaction groups. Transaction groups enter the syncing state periodically, so the number of queued async writes will quickly burst up and then bleed down to zero. Rather than servicing them as quickly as possible, the I/O scheduler changes the maximum number of active async write operations according to the amount of dirty data in the pool. Since both throughput and latency typically increase with the number of concurrent operations issued to physical devices, reducing the burstiness in the number of simultaneous operations also stabilizes the response time of operations from other queues, in particular synchronous ones. In broad strokes, the I/O scheduler will issue more concurrent operations from the async write queue as there is more dirty data in the pool. . .Ss Async Writes The number of concurrent operations issued for the async write I/O class follows a piece-wise linear function defined by a few adjustable points: .Bd -literal | o---------| <-- \fBzfs_vdev_async_write_max_active\fP ^ | /^ | | | / | | active | / | | I/O | / | | count | / | | | / | | |-------o | | <-- \fBzfs_vdev_async_write_min_active\fP 0|_______^______|_________| 0% | | 100% of \fBzfs_dirty_data_max\fP | | | `-- \fBzfs_vdev_async_write_active_max_dirty_percent\fP `--------- \fBzfs_vdev_async_write_active_min_dirty_percent\fP .Ed .Pp Until the amount of dirty data exceeds a minimum percentage of the dirty data allowed in the pool, the I/O scheduler will limit the number of concurrent operations to the minimum. As that threshold is crossed, the number of concurrent operations issued increases linearly to the maximum at the specified maximum percentage of the dirty data allowed in the pool. .Pp Ideally, the amount of dirty data on a busy pool will stay in the sloped part of the function between .Sy zfs_vdev_async_write_active_min_dirty_percent and .Sy zfs_vdev_async_write_active_max_dirty_percent . If it exceeds the maximum percentage, this indicates that the rate of incoming data is greater than the rate that the backend storage can handle. In this case, we must further throttle incoming writes, as described in the next section. . .Sh ZFS TRANSACTION DELAY We delay transactions when we've determined that the backend storage isn't able to accommodate the rate of incoming writes. .Pp If there is already a transaction waiting, we delay relative to when that transaction will finish waiting. This way the calculated delay time is independent of the number of threads concurrently executing transactions. .Pp If we are the only waiter, wait relative to when the transaction started, rather than the current time. This credits the transaction for "time already served", e.g. reading indirect blocks. .Pp The minimum time for a transaction to take is calculated as .D1 min_time = min( Ns Sy zfs_delay_scale No \(mu Po Sy dirty No \- Sy min Pc / Po Sy max No \- Sy dirty Pc , 100ms) .Pp The delay has two degrees of freedom that can be adjusted via tunables. The percentage of dirty data at which we start to delay is defined by .Sy zfs_delay_min_dirty_percent . This should typically be at or above .Sy zfs_vdev_async_write_active_max_dirty_percent , so that we only start to delay after writing at full speed has failed to keep up with the incoming write rate. The scale of the curve is defined by .Sy zfs_delay_scale . Roughly speaking, this variable determines the amount of delay at the midpoint of the curve. .Bd -literal delay 10ms +-------------------------------------------------------------*+ | *| 9ms + *+ | *| 8ms + *+ | * | 7ms + * + | * | 6ms + * + | * | 5ms + * + | * | 4ms + * + | * | 3ms + * + | * | 2ms + (midpoint) * + | | ** | 1ms + v *** + | \fBzfs_delay_scale\fP ----------> ******** | 0 +-------------------------------------*********----------------+ 0% <- \fBzfs_dirty_data_max\fP -> 100% .Ed .Pp Note, that since the delay is added to the outstanding time remaining on the most recent transaction it's effectively the inverse of IOPS. Here, the midpoint of .Em 500 us translates to .Em 2000 IOPS . The shape of the curve was chosen such that small changes in the amount of accumulated dirty data in the first three quarters of the curve yield relatively small differences in the amount of delay. .Pp The effects can be easier to understand when the amount of delay is represented on a logarithmic scale: .Bd -literal delay 100ms +-------------------------------------------------------------++ + + | | + *+ 10ms + *+ + ** + | (midpoint) ** | + | ** + 1ms + v **** + + \fBzfs_delay_scale\fP ----------> ***** + | **** | + **** + 100us + ** + + * + | * | + * + 10us + * + + + | | + + +--------------------------------------------------------------+ 0% <- \fBzfs_dirty_data_max\fP -> 100% .Ed .Pp Note here that only as the amount of dirty data approaches its limit does the delay start to increase rapidly. The goal of a properly tuned system should be to keep the amount of dirty data out of that range by first ensuring that the appropriate limits are set for the I/O scheduler to reach optimal throughput on the back-end storage, and then by changing the value of .Sy zfs_delay_scale to increase the steepness of the curve. diff --git a/sys/contrib/openzfs/man/man8/zpool-initialize.8 b/sys/contrib/openzfs/man/man8/zpool-initialize.8 index eae711bff429..a9c8fd35aec9 100644 --- a/sys/contrib/openzfs/man/man8/zpool-initialize.8 +++ b/sys/contrib/openzfs/man/man8/zpool-initialize.8 @@ -1,73 +1,81 @@ .\" .\" CDDL HEADER START .\" .\" The contents of this file are subject to the terms of the .\" Common Development and Distribution License (the "License"). .\" You may not use this file except in compliance with the License. .\" .\" You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE .\" or https://opensource.org/licenses/CDDL-1.0. .\" See the License for the specific language governing permissions .\" and limitations under the License. .\" .\" When distributing Covered Code, include this CDDL HEADER in each .\" file and include the License file at usr/src/OPENSOLARIS.LICENSE. .\" If applicable, add the following below this CDDL HEADER, with the .\" fields enclosed by brackets "[]" replaced with your own identifying .\" information: Portions Copyright [yyyy] [name of copyright owner] .\" .\" CDDL HEADER END .\" .\" Copyright (c) 2007, Sun Microsystems, Inc. All Rights Reserved. .\" Copyright (c) 2012, 2018 by Delphix. All rights reserved. .\" Copyright (c) 2012 Cyril Plisko. All Rights Reserved. .\" Copyright (c) 2017 Datto Inc. .\" Copyright (c) 2018 George Melikov. All Rights Reserved. .\" Copyright 2017 Nexenta Systems, Inc. .\" Copyright (c) 2017 Open-E, Inc. All Rights Reserved. .\" .Dd May 27, 2021 .Dt ZPOOL-INITIALIZE 8 .Os . .Sh NAME .Nm zpool-initialize .Nd write to unallocated regions of ZFS storage pool .Sh SYNOPSIS .Nm zpool .Cm initialize -.Op Fl c Ns | Ns Fl s +.Op Fl c Ns | Ns Fl s | Ns Fl u .Op Fl w .Ar pool .Oo Ar device Oc Ns … . .Sh DESCRIPTION Begins initializing by writing to all unallocated regions on the specified devices, or all eligible devices in the pool if no individual devices are specified. Only leaf data or log devices may be initialized. .Bl -tag -width Ds .It Fl c , -cancel Cancel initializing on the specified devices, or all eligible devices if none are specified. If one or more target devices are invalid or are not currently being initialized, the command will fail and no cancellation will occur on any device. .It Fl s , -suspend Suspend initializing on the specified devices, or all eligible devices if none are specified. If one or more target devices are invalid or are not currently being initialized, the command will fail and no suspension will occur on any device. Initializing can then be resumed by running .Nm zpool Cm initialize with no flags on the relevant target devices. +.It Fl u , -uninit +Clears the initialization state on the specified devices, or all eligible +devices if none are specified. +If the devices are being actively initialized the command will fail. +After being cleared +.Nm zpool Cm initialize +with no flags can be used to re-initialize all unallocoated regions on +the relevant target devices. .It Fl w , -wait Wait until the devices have finished initializing before returning. .El . .Sh SEE ALSO .Xr zpool-add 8 , .Xr zpool-attach 8 , .Xr zpool-create 8 , .Xr zpool-online 8 , .Xr zpool-replace 8 , .Xr zpool-trim 8 diff --git a/sys/contrib/openzfs/man/man8/zpool-scrub.8 b/sys/contrib/openzfs/man/man8/zpool-scrub.8 index 1fdbb8a5d56d..138226e4562c 100644 --- a/sys/contrib/openzfs/man/man8/zpool-scrub.8 +++ b/sys/contrib/openzfs/man/man8/zpool-scrub.8 @@ -1,140 +1,159 @@ .\" .\" CDDL HEADER START .\" .\" The contents of this file are subject to the terms of the .\" Common Development and Distribution License (the "License"). .\" You may not use this file except in compliance with the License. .\" .\" You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE .\" or https://opensource.org/licenses/CDDL-1.0. .\" See the License for the specific language governing permissions .\" and limitations under the License. .\" .\" When distributing Covered Code, include this CDDL HEADER in each .\" file and include the License file at usr/src/OPENSOLARIS.LICENSE. .\" If applicable, add the following below this CDDL HEADER, with the .\" fields enclosed by brackets "[]" replaced with your own identifying .\" information: Portions Copyright [yyyy] [name of copyright owner] .\" .\" CDDL HEADER END .\" .\" Copyright (c) 2007, Sun Microsystems, Inc. All Rights Reserved. .\" Copyright (c) 2012, 2018 by Delphix. All rights reserved. .\" Copyright (c) 2012 Cyril Plisko. All Rights Reserved. .\" Copyright (c) 2017 Datto Inc. .\" Copyright (c) 2018, 2021 George Melikov. All Rights Reserved. .\" Copyright 2017 Nexenta Systems, Inc. .\" Copyright (c) 2017 Open-E, Inc. All Rights Reserved. .\" .Dd July 25, 2021 .Dt ZPOOL-SCRUB 8 .Os . .Sh NAME .Nm zpool-scrub .Nd begin or resume scrub of ZFS storage pools .Sh SYNOPSIS .Nm zpool .Cm scrub .Op Fl s Ns | Ns Fl p .Op Fl w +.Op Fl e .Ar pool Ns … . .Sh DESCRIPTION Begins a scrub or resumes a paused scrub. The scrub examines all data in the specified pools to verify that it checksums correctly. For replicated .Pq mirror, raidz, or draid devices, ZFS automatically repairs any damage discovered during the scrub. The .Nm zpool Cm status command reports the progress of the scrub and summarizes the results of the scrub upon completion. .Pp Scrubbing and resilvering are very similar operations. The difference is that resilvering only examines data that ZFS knows to be out of date .Po for example, when attaching a new device to a mirror or replacing an existing device .Pc , whereas scrubbing examines all data to discover silent errors due to hardware faults or disk failure. .Pp +When scrubbing a pool with encrypted filesystems the keys do not need to be +loaded. +However, if the keys are not loaded and an unrepairable checksum error is +detected the file name cannot be included in the +.Nm zpool Cm status Fl v +verbose error report. +.Pp Because scrubbing and resilvering are I/O-intensive operations, ZFS only allows one at a time. .Pp A scrub is split into two parts: metadata scanning and block scrubbing. The metadata scanning sorts blocks into large sequential ranges which can then be read much more efficiently from disk when issuing the scrub I/O. .Pp If a scrub is paused, the .Nm zpool Cm scrub resumes it. If a resilver is in progress, ZFS does not allow a scrub to be started until the resilver completes. .Pp Note that, due to changes in pool data on a live system, it is possible for scrubs to progress slightly beyond 100% completion. During this period, no completion time estimate will be provided. . .Sh OPTIONS .Bl -tag -width "-s" .It Fl s Stop scrubbing. .It Fl p Pause scrubbing. Scrub pause state and progress are periodically synced to disk. If the system is restarted or pool is exported during a paused scrub, even after import, scrub will remain paused until it is resumed. Once resumed the scrub will pick up from the place where it was last checkpointed to disk. To resume a paused scrub issue .Nm zpool Cm scrub +or +.Nm zpool Cm scrub +.Fl e again. .It Fl w Wait until scrub has completed before returning. +.It Fl e +Only scrub files with known data errors as reported by +.Nm zpool Cm status Fl v . +The pool must have been scrubbed at least once with the +.Sy head_errlog +feature enabled to use this option. +Error scrubbing cannot be run simultaneously with regular scrubbing or +resilvering, nor can it be run when a regular scrub is paused. .El .Sh EXAMPLES .Ss Example 1 Status of pool with ongoing scrub: .sp .Bd -literal -compact .No # Nm zpool Cm status ... scan: scrub in progress since Sun Jul 25 16:07:49 2021 403M scanned at 100M/s, 68.4M issued at 10.0M/s, 405M total 0B repaired, 16.91% done, 00:00:04 to go ... .Ed .Pp Where metadata which references 403M of file data has been scanned at 100M/s, and 68.4M of that file data has been scrubbed sequentially at 10.0M/s. .Sh PERIODIC SCRUB On machines using systemd, scrub timers can be enabled on per-pool basis. .Nm weekly and .Nm monthly timer units are provided. .Bl -tag -width Ds .It Xo .Xc .Nm systemctl .Cm enable .Cm zfs-scrub-\fIweekly\fB@\fIrpool\fB.timer .Cm --now .It Xo .Xc .Nm systemctl .Cm enable .Cm zfs-scrub-\fImonthly\fB@\fIotherpool\fB.timer .Cm --now .El . .Sh SEE ALSO .Xr systemd.timer 5 , .Xr zpool-iostat 8 , .Xr zpool-resilver 8 , .Xr zpool-status 8 diff --git a/sys/contrib/openzfs/module/zfs/dbuf.c b/sys/contrib/openzfs/module/zfs/dbuf.c index 049a62c1c171..272e712586fa 100644 --- a/sys/contrib/openzfs/module/zfs/dbuf.c +++ b/sys/contrib/openzfs/module/zfs/dbuf.c @@ -1,5258 +1,5260 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright 2011 Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2012, 2020 by Delphix. All rights reserved. * Copyright (c) 2013 by Saso Kiselkov. All rights reserved. * Copyright (c) 2014 Spectra Logic Corporation, All rights reserved. * Copyright (c) 2019, Klara Inc. * Copyright (c) 2019, Allan Jude * Copyright (c) 2021, 2022 by Pawel Jakub Dawidek */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include static kstat_t *dbuf_ksp; typedef struct dbuf_stats { /* * Various statistics about the size of the dbuf cache. */ kstat_named_t cache_count; kstat_named_t cache_size_bytes; kstat_named_t cache_size_bytes_max; /* * Statistics regarding the bounds on the dbuf cache size. */ kstat_named_t cache_target_bytes; kstat_named_t cache_lowater_bytes; kstat_named_t cache_hiwater_bytes; /* * Total number of dbuf cache evictions that have occurred. */ kstat_named_t cache_total_evicts; /* * The distribution of dbuf levels in the dbuf cache and * the total size of all dbufs at each level. */ kstat_named_t cache_levels[DN_MAX_LEVELS]; kstat_named_t cache_levels_bytes[DN_MAX_LEVELS]; /* * Statistics about the dbuf hash table. */ kstat_named_t hash_hits; kstat_named_t hash_misses; kstat_named_t hash_collisions; kstat_named_t hash_elements; kstat_named_t hash_elements_max; /* * Number of sublists containing more than one dbuf in the dbuf * hash table. Keep track of the longest hash chain. */ kstat_named_t hash_chains; kstat_named_t hash_chain_max; /* * Number of times a dbuf_create() discovers that a dbuf was * already created and in the dbuf hash table. */ kstat_named_t hash_insert_race; /* * Number of entries in the hash table dbuf and mutex arrays. */ kstat_named_t hash_table_count; kstat_named_t hash_mutex_count; /* * Statistics about the size of the metadata dbuf cache. */ kstat_named_t metadata_cache_count; kstat_named_t metadata_cache_size_bytes; kstat_named_t metadata_cache_size_bytes_max; /* * For diagnostic purposes, this is incremented whenever we can't add * something to the metadata cache because it's full, and instead put * the data in the regular dbuf cache. */ kstat_named_t metadata_cache_overflow; } dbuf_stats_t; dbuf_stats_t dbuf_stats = { { "cache_count", KSTAT_DATA_UINT64 }, { "cache_size_bytes", KSTAT_DATA_UINT64 }, { "cache_size_bytes_max", KSTAT_DATA_UINT64 }, { "cache_target_bytes", KSTAT_DATA_UINT64 }, { "cache_lowater_bytes", KSTAT_DATA_UINT64 }, { "cache_hiwater_bytes", KSTAT_DATA_UINT64 }, { "cache_total_evicts", KSTAT_DATA_UINT64 }, { { "cache_levels_N", KSTAT_DATA_UINT64 } }, { { "cache_levels_bytes_N", KSTAT_DATA_UINT64 } }, { "hash_hits", KSTAT_DATA_UINT64 }, { "hash_misses", KSTAT_DATA_UINT64 }, { "hash_collisions", KSTAT_DATA_UINT64 }, { "hash_elements", KSTAT_DATA_UINT64 }, { "hash_elements_max", KSTAT_DATA_UINT64 }, { "hash_chains", KSTAT_DATA_UINT64 }, { "hash_chain_max", KSTAT_DATA_UINT64 }, { "hash_insert_race", KSTAT_DATA_UINT64 }, { "hash_table_count", KSTAT_DATA_UINT64 }, { "hash_mutex_count", KSTAT_DATA_UINT64 }, { "metadata_cache_count", KSTAT_DATA_UINT64 }, { "metadata_cache_size_bytes", KSTAT_DATA_UINT64 }, { "metadata_cache_size_bytes_max", KSTAT_DATA_UINT64 }, { "metadata_cache_overflow", KSTAT_DATA_UINT64 } }; struct { wmsum_t cache_count; wmsum_t cache_total_evicts; wmsum_t cache_levels[DN_MAX_LEVELS]; wmsum_t cache_levels_bytes[DN_MAX_LEVELS]; wmsum_t hash_hits; wmsum_t hash_misses; wmsum_t hash_collisions; wmsum_t hash_chains; wmsum_t hash_insert_race; wmsum_t metadata_cache_count; wmsum_t metadata_cache_overflow; } dbuf_sums; #define DBUF_STAT_INCR(stat, val) \ wmsum_add(&dbuf_sums.stat, val); #define DBUF_STAT_DECR(stat, val) \ DBUF_STAT_INCR(stat, -(val)); #define DBUF_STAT_BUMP(stat) \ DBUF_STAT_INCR(stat, 1); #define DBUF_STAT_BUMPDOWN(stat) \ DBUF_STAT_INCR(stat, -1); #define DBUF_STAT_MAX(stat, v) { \ uint64_t _m; \ while ((v) > (_m = dbuf_stats.stat.value.ui64) && \ (_m != atomic_cas_64(&dbuf_stats.stat.value.ui64, _m, (v))))\ continue; \ } static void dbuf_write(dbuf_dirty_record_t *dr, arc_buf_t *data, dmu_tx_t *tx); static void dbuf_sync_leaf_verify_bonus_dnode(dbuf_dirty_record_t *dr); static int dbuf_read_verify_dnode_crypt(dmu_buf_impl_t *db, uint32_t flags); /* * Global data structures and functions for the dbuf cache. */ static kmem_cache_t *dbuf_kmem_cache; static taskq_t *dbu_evict_taskq; static kthread_t *dbuf_cache_evict_thread; static kmutex_t dbuf_evict_lock; static kcondvar_t dbuf_evict_cv; static boolean_t dbuf_evict_thread_exit; /* * There are two dbuf caches; each dbuf can only be in one of them at a time. * * 1. Cache of metadata dbufs, to help make read-heavy administrative commands * from /sbin/zfs run faster. The "metadata cache" specifically stores dbufs * that represent the metadata that describes filesystems/snapshots/ * bookmarks/properties/etc. We only evict from this cache when we export a * pool, to short-circuit as much I/O as possible for all administrative * commands that need the metadata. There is no eviction policy for this * cache, because we try to only include types in it which would occupy a * very small amount of space per object but create a large impact on the * performance of these commands. Instead, after it reaches a maximum size * (which should only happen on very small memory systems with a very large * number of filesystem objects), we stop taking new dbufs into the * metadata cache, instead putting them in the normal dbuf cache. * * 2. LRU cache of dbufs. The dbuf cache maintains a list of dbufs that * are not currently held but have been recently released. These dbufs * are not eligible for arc eviction until they are aged out of the cache. * Dbufs that are aged out of the cache will be immediately destroyed and * become eligible for arc eviction. * * Dbufs are added to these caches once the last hold is released. If a dbuf is * later accessed and still exists in the dbuf cache, then it will be removed * from the cache and later re-added to the head of the cache. * * If a given dbuf meets the requirements for the metadata cache, it will go * there, otherwise it will be considered for the generic LRU dbuf cache. The * caches and the refcounts tracking their sizes are stored in an array indexed * by those caches' matching enum values (from dbuf_cached_state_t). */ typedef struct dbuf_cache { multilist_t cache; zfs_refcount_t size ____cacheline_aligned; } dbuf_cache_t; dbuf_cache_t dbuf_caches[DB_CACHE_MAX]; /* Size limits for the caches */ static uint64_t dbuf_cache_max_bytes = UINT64_MAX; static uint64_t dbuf_metadata_cache_max_bytes = UINT64_MAX; /* Set the default sizes of the caches to log2 fraction of arc size */ static uint_t dbuf_cache_shift = 5; static uint_t dbuf_metadata_cache_shift = 6; /* Set the dbuf hash mutex count as log2 shift (dynamic by default) */ static uint_t dbuf_mutex_cache_shift = 0; static unsigned long dbuf_cache_target_bytes(void); static unsigned long dbuf_metadata_cache_target_bytes(void); /* * The LRU dbuf cache uses a three-stage eviction policy: * - A low water marker designates when the dbuf eviction thread * should stop evicting from the dbuf cache. * - When we reach the maximum size (aka mid water mark), we * signal the eviction thread to run. * - The high water mark indicates when the eviction thread * is unable to keep up with the incoming load and eviction must * happen in the context of the calling thread. * * The dbuf cache: * (max size) * low water mid water hi water * +----------------------------------------+----------+----------+ * | | | | * | | | | * | | | | * | | | | * +----------------------------------------+----------+----------+ * stop signal evict * evicting eviction directly * thread * * The high and low water marks indicate the operating range for the eviction * thread. The low water mark is, by default, 90% of the total size of the * cache and the high water mark is at 110% (both of these percentages can be * changed by setting dbuf_cache_lowater_pct and dbuf_cache_hiwater_pct, * respectively). The eviction thread will try to ensure that the cache remains * within this range by waking up every second and checking if the cache is * above the low water mark. The thread can also be woken up by callers adding * elements into the cache if the cache is larger than the mid water (i.e max * cache size). Once the eviction thread is woken up and eviction is required, * it will continue evicting buffers until it's able to reduce the cache size * to the low water mark. If the cache size continues to grow and hits the high * water mark, then callers adding elements to the cache will begin to evict * directly from the cache until the cache is no longer above the high water * mark. */ /* * The percentage above and below the maximum cache size. */ static uint_t dbuf_cache_hiwater_pct = 10; static uint_t dbuf_cache_lowater_pct = 10; static int dbuf_cons(void *vdb, void *unused, int kmflag) { (void) unused, (void) kmflag; dmu_buf_impl_t *db = vdb; memset(db, 0, sizeof (dmu_buf_impl_t)); mutex_init(&db->db_mtx, NULL, MUTEX_DEFAULT, NULL); rw_init(&db->db_rwlock, NULL, RW_DEFAULT, NULL); cv_init(&db->db_changed, NULL, CV_DEFAULT, NULL); multilist_link_init(&db->db_cache_link); zfs_refcount_create(&db->db_holds); return (0); } static void dbuf_dest(void *vdb, void *unused) { (void) unused; dmu_buf_impl_t *db = vdb; mutex_destroy(&db->db_mtx); rw_destroy(&db->db_rwlock); cv_destroy(&db->db_changed); ASSERT(!multilist_link_active(&db->db_cache_link)); zfs_refcount_destroy(&db->db_holds); } /* * dbuf hash table routines */ static dbuf_hash_table_t dbuf_hash_table; /* * We use Cityhash for this. It's fast, and has good hash properties without * requiring any large static buffers. */ static uint64_t dbuf_hash(void *os, uint64_t obj, uint8_t lvl, uint64_t blkid) { return (cityhash4((uintptr_t)os, obj, (uint64_t)lvl, blkid)); } #define DTRACE_SET_STATE(db, why) \ DTRACE_PROBE2(dbuf__state_change, dmu_buf_impl_t *, db, \ const char *, why) #define DBUF_EQUAL(dbuf, os, obj, level, blkid) \ ((dbuf)->db.db_object == (obj) && \ (dbuf)->db_objset == (os) && \ (dbuf)->db_level == (level) && \ (dbuf)->db_blkid == (blkid)) dmu_buf_impl_t * dbuf_find(objset_t *os, uint64_t obj, uint8_t level, uint64_t blkid, uint64_t *hash_out) { dbuf_hash_table_t *h = &dbuf_hash_table; uint64_t hv; uint64_t idx; dmu_buf_impl_t *db; hv = dbuf_hash(os, obj, level, blkid); idx = hv & h->hash_table_mask; mutex_enter(DBUF_HASH_MUTEX(h, idx)); for (db = h->hash_table[idx]; db != NULL; db = db->db_hash_next) { if (DBUF_EQUAL(db, os, obj, level, blkid)) { mutex_enter(&db->db_mtx); if (db->db_state != DB_EVICTING) { mutex_exit(DBUF_HASH_MUTEX(h, idx)); return (db); } mutex_exit(&db->db_mtx); } } mutex_exit(DBUF_HASH_MUTEX(h, idx)); if (hash_out != NULL) *hash_out = hv; return (NULL); } static dmu_buf_impl_t * dbuf_find_bonus(objset_t *os, uint64_t object) { dnode_t *dn; dmu_buf_impl_t *db = NULL; if (dnode_hold(os, object, FTAG, &dn) == 0) { rw_enter(&dn->dn_struct_rwlock, RW_READER); if (dn->dn_bonus != NULL) { db = dn->dn_bonus; mutex_enter(&db->db_mtx); } rw_exit(&dn->dn_struct_rwlock); dnode_rele(dn, FTAG); } return (db); } /* * Insert an entry into the hash table. If there is already an element * equal to elem in the hash table, then the already existing element * will be returned and the new element will not be inserted. * Otherwise returns NULL. */ static dmu_buf_impl_t * dbuf_hash_insert(dmu_buf_impl_t *db) { dbuf_hash_table_t *h = &dbuf_hash_table; objset_t *os = db->db_objset; uint64_t obj = db->db.db_object; int level = db->db_level; uint64_t blkid, idx; dmu_buf_impl_t *dbf; uint32_t i; blkid = db->db_blkid; ASSERT3U(dbuf_hash(os, obj, level, blkid), ==, db->db_hash); idx = db->db_hash & h->hash_table_mask; mutex_enter(DBUF_HASH_MUTEX(h, idx)); for (dbf = h->hash_table[idx], i = 0; dbf != NULL; dbf = dbf->db_hash_next, i++) { if (DBUF_EQUAL(dbf, os, obj, level, blkid)) { mutex_enter(&dbf->db_mtx); if (dbf->db_state != DB_EVICTING) { mutex_exit(DBUF_HASH_MUTEX(h, idx)); return (dbf); } mutex_exit(&dbf->db_mtx); } } if (i > 0) { DBUF_STAT_BUMP(hash_collisions); if (i == 1) DBUF_STAT_BUMP(hash_chains); DBUF_STAT_MAX(hash_chain_max, i); } mutex_enter(&db->db_mtx); db->db_hash_next = h->hash_table[idx]; h->hash_table[idx] = db; mutex_exit(DBUF_HASH_MUTEX(h, idx)); uint64_t he = atomic_inc_64_nv(&dbuf_stats.hash_elements.value.ui64); DBUF_STAT_MAX(hash_elements_max, he); return (NULL); } /* * This returns whether this dbuf should be stored in the metadata cache, which * is based on whether it's from one of the dnode types that store data related * to traversing dataset hierarchies. */ static boolean_t dbuf_include_in_metadata_cache(dmu_buf_impl_t *db) { DB_DNODE_ENTER(db); dmu_object_type_t type = DB_DNODE(db)->dn_type; DB_DNODE_EXIT(db); /* Check if this dbuf is one of the types we care about */ if (DMU_OT_IS_METADATA_CACHED(type)) { /* If we hit this, then we set something up wrong in dmu_ot */ ASSERT(DMU_OT_IS_METADATA(type)); /* * Sanity check for small-memory systems: don't allocate too * much memory for this purpose. */ if (zfs_refcount_count( &dbuf_caches[DB_DBUF_METADATA_CACHE].size) > dbuf_metadata_cache_target_bytes()) { DBUF_STAT_BUMP(metadata_cache_overflow); return (B_FALSE); } return (B_TRUE); } return (B_FALSE); } /* * Remove an entry from the hash table. It must be in the EVICTING state. */ static void dbuf_hash_remove(dmu_buf_impl_t *db) { dbuf_hash_table_t *h = &dbuf_hash_table; uint64_t idx; dmu_buf_impl_t *dbf, **dbp; ASSERT3U(dbuf_hash(db->db_objset, db->db.db_object, db->db_level, db->db_blkid), ==, db->db_hash); idx = db->db_hash & h->hash_table_mask; /* * We mustn't hold db_mtx to maintain lock ordering: * DBUF_HASH_MUTEX > db_mtx. */ ASSERT(zfs_refcount_is_zero(&db->db_holds)); ASSERT(db->db_state == DB_EVICTING); ASSERT(!MUTEX_HELD(&db->db_mtx)); mutex_enter(DBUF_HASH_MUTEX(h, idx)); dbp = &h->hash_table[idx]; while ((dbf = *dbp) != db) { dbp = &dbf->db_hash_next; ASSERT(dbf != NULL); } *dbp = db->db_hash_next; db->db_hash_next = NULL; if (h->hash_table[idx] && h->hash_table[idx]->db_hash_next == NULL) DBUF_STAT_BUMPDOWN(hash_chains); mutex_exit(DBUF_HASH_MUTEX(h, idx)); atomic_dec_64(&dbuf_stats.hash_elements.value.ui64); } typedef enum { DBVU_EVICTING, DBVU_NOT_EVICTING } dbvu_verify_type_t; static void dbuf_verify_user(dmu_buf_impl_t *db, dbvu_verify_type_t verify_type) { #ifdef ZFS_DEBUG int64_t holds; if (db->db_user == NULL) return; /* Only data blocks support the attachment of user data. */ ASSERT(db->db_level == 0); /* Clients must resolve a dbuf before attaching user data. */ ASSERT(db->db.db_data != NULL); ASSERT3U(db->db_state, ==, DB_CACHED); holds = zfs_refcount_count(&db->db_holds); if (verify_type == DBVU_EVICTING) { /* * Immediate eviction occurs when holds == dirtycnt. * For normal eviction buffers, holds is zero on * eviction, except when dbuf_fix_old_data() calls * dbuf_clear_data(). However, the hold count can grow * during eviction even though db_mtx is held (see * dmu_bonus_hold() for an example), so we can only * test the generic invariant that holds >= dirtycnt. */ ASSERT3U(holds, >=, db->db_dirtycnt); } else { if (db->db_user_immediate_evict == TRUE) ASSERT3U(holds, >=, db->db_dirtycnt); else ASSERT3U(holds, >, 0); } #endif } static void dbuf_evict_user(dmu_buf_impl_t *db) { dmu_buf_user_t *dbu = db->db_user; ASSERT(MUTEX_HELD(&db->db_mtx)); if (dbu == NULL) return; dbuf_verify_user(db, DBVU_EVICTING); db->db_user = NULL; #ifdef ZFS_DEBUG if (dbu->dbu_clear_on_evict_dbufp != NULL) *dbu->dbu_clear_on_evict_dbufp = NULL; #endif /* * There are two eviction callbacks - one that we call synchronously * and one that we invoke via a taskq. The async one is useful for * avoiding lock order reversals and limiting stack depth. * * Note that if we have a sync callback but no async callback, * it's likely that the sync callback will free the structure * containing the dbu. In that case we need to take care to not * dereference dbu after calling the sync evict func. */ boolean_t has_async = (dbu->dbu_evict_func_async != NULL); if (dbu->dbu_evict_func_sync != NULL) dbu->dbu_evict_func_sync(dbu); if (has_async) { taskq_dispatch_ent(dbu_evict_taskq, dbu->dbu_evict_func_async, dbu, 0, &dbu->dbu_tqent); } } boolean_t dbuf_is_metadata(dmu_buf_impl_t *db) { /* * Consider indirect blocks and spill blocks to be meta data. */ if (db->db_level > 0 || db->db_blkid == DMU_SPILL_BLKID) { return (B_TRUE); } else { boolean_t is_metadata; DB_DNODE_ENTER(db); is_metadata = DMU_OT_IS_METADATA(DB_DNODE(db)->dn_type); DB_DNODE_EXIT(db); return (is_metadata); } } /* * We want to exclude buffers that are on a special allocation class from * L2ARC. */ boolean_t dbuf_is_l2cacheable(dmu_buf_impl_t *db) { if (db->db_objset->os_secondary_cache == ZFS_CACHE_ALL || (db->db_objset->os_secondary_cache == ZFS_CACHE_METADATA && dbuf_is_metadata(db))) { if (l2arc_exclude_special == 0) return (B_TRUE); blkptr_t *bp = db->db_blkptr; if (bp == NULL || BP_IS_HOLE(bp)) return (B_FALSE); uint64_t vdev = DVA_GET_VDEV(bp->blk_dva); vdev_t *rvd = db->db_objset->os_spa->spa_root_vdev; vdev_t *vd = NULL; if (vdev < rvd->vdev_children) vd = rvd->vdev_child[vdev]; if (vd == NULL) return (B_TRUE); if (vd->vdev_alloc_bias != VDEV_BIAS_SPECIAL && vd->vdev_alloc_bias != VDEV_BIAS_DEDUP) return (B_TRUE); } return (B_FALSE); } static inline boolean_t dnode_level_is_l2cacheable(blkptr_t *bp, dnode_t *dn, int64_t level) { if (dn->dn_objset->os_secondary_cache == ZFS_CACHE_ALL || (dn->dn_objset->os_secondary_cache == ZFS_CACHE_METADATA && (level > 0 || DMU_OT_IS_METADATA(dn->dn_handle->dnh_dnode->dn_type)))) { if (l2arc_exclude_special == 0) return (B_TRUE); if (bp == NULL || BP_IS_HOLE(bp)) return (B_FALSE); uint64_t vdev = DVA_GET_VDEV(bp->blk_dva); vdev_t *rvd = dn->dn_objset->os_spa->spa_root_vdev; vdev_t *vd = NULL; if (vdev < rvd->vdev_children) vd = rvd->vdev_child[vdev]; if (vd == NULL) return (B_TRUE); if (vd->vdev_alloc_bias != VDEV_BIAS_SPECIAL && vd->vdev_alloc_bias != VDEV_BIAS_DEDUP) return (B_TRUE); } return (B_FALSE); } /* * This function *must* return indices evenly distributed between all * sublists of the multilist. This is needed due to how the dbuf eviction * code is laid out; dbuf_evict_thread() assumes dbufs are evenly * distributed between all sublists and uses this assumption when * deciding which sublist to evict from and how much to evict from it. */ static unsigned int dbuf_cache_multilist_index_func(multilist_t *ml, void *obj) { dmu_buf_impl_t *db = obj; /* * The assumption here, is the hash value for a given * dmu_buf_impl_t will remain constant throughout it's lifetime * (i.e. it's objset, object, level and blkid fields don't change). * Thus, we don't need to store the dbuf's sublist index * on insertion, as this index can be recalculated on removal. * * Also, the low order bits of the hash value are thought to be * distributed evenly. Otherwise, in the case that the multilist * has a power of two number of sublists, each sublists' usage * would not be evenly distributed. In this context full 64bit * division would be a waste of time, so limit it to 32 bits. */ return ((unsigned int)dbuf_hash(db->db_objset, db->db.db_object, db->db_level, db->db_blkid) % multilist_get_num_sublists(ml)); } /* * The target size of the dbuf cache can grow with the ARC target, * unless limited by the tunable dbuf_cache_max_bytes. */ static inline unsigned long dbuf_cache_target_bytes(void) { return (MIN(dbuf_cache_max_bytes, arc_target_bytes() >> dbuf_cache_shift)); } /* * The target size of the dbuf metadata cache can grow with the ARC target, * unless limited by the tunable dbuf_metadata_cache_max_bytes. */ static inline unsigned long dbuf_metadata_cache_target_bytes(void) { return (MIN(dbuf_metadata_cache_max_bytes, arc_target_bytes() >> dbuf_metadata_cache_shift)); } static inline uint64_t dbuf_cache_hiwater_bytes(void) { uint64_t dbuf_cache_target = dbuf_cache_target_bytes(); return (dbuf_cache_target + (dbuf_cache_target * dbuf_cache_hiwater_pct) / 100); } static inline uint64_t dbuf_cache_lowater_bytes(void) { uint64_t dbuf_cache_target = dbuf_cache_target_bytes(); return (dbuf_cache_target - (dbuf_cache_target * dbuf_cache_lowater_pct) / 100); } static inline boolean_t dbuf_cache_above_lowater(void) { return (zfs_refcount_count(&dbuf_caches[DB_DBUF_CACHE].size) > dbuf_cache_lowater_bytes()); } /* * Evict the oldest eligible dbuf from the dbuf cache. */ static void dbuf_evict_one(void) { int idx = multilist_get_random_index(&dbuf_caches[DB_DBUF_CACHE].cache); multilist_sublist_t *mls = multilist_sublist_lock( &dbuf_caches[DB_DBUF_CACHE].cache, idx); ASSERT(!MUTEX_HELD(&dbuf_evict_lock)); dmu_buf_impl_t *db = multilist_sublist_tail(mls); while (db != NULL && mutex_tryenter(&db->db_mtx) == 0) { db = multilist_sublist_prev(mls, db); } DTRACE_PROBE2(dbuf__evict__one, dmu_buf_impl_t *, db, multilist_sublist_t *, mls); if (db != NULL) { multilist_sublist_remove(mls, db); multilist_sublist_unlock(mls); (void) zfs_refcount_remove_many( &dbuf_caches[DB_DBUF_CACHE].size, db->db.db_size, db); DBUF_STAT_BUMPDOWN(cache_levels[db->db_level]); DBUF_STAT_BUMPDOWN(cache_count); DBUF_STAT_DECR(cache_levels_bytes[db->db_level], db->db.db_size); ASSERT3U(db->db_caching_status, ==, DB_DBUF_CACHE); db->db_caching_status = DB_NO_CACHE; dbuf_destroy(db); DBUF_STAT_BUMP(cache_total_evicts); } else { multilist_sublist_unlock(mls); } } /* * The dbuf evict thread is responsible for aging out dbufs from the * cache. Once the cache has reached it's maximum size, dbufs are removed * and destroyed. The eviction thread will continue running until the size * of the dbuf cache is at or below the maximum size. Once the dbuf is aged * out of the cache it is destroyed and becomes eligible for arc eviction. */ static __attribute__((noreturn)) void dbuf_evict_thread(void *unused) { (void) unused; callb_cpr_t cpr; CALLB_CPR_INIT(&cpr, &dbuf_evict_lock, callb_generic_cpr, FTAG); mutex_enter(&dbuf_evict_lock); while (!dbuf_evict_thread_exit) { while (!dbuf_cache_above_lowater() && !dbuf_evict_thread_exit) { CALLB_CPR_SAFE_BEGIN(&cpr); (void) cv_timedwait_idle_hires(&dbuf_evict_cv, &dbuf_evict_lock, SEC2NSEC(1), MSEC2NSEC(1), 0); CALLB_CPR_SAFE_END(&cpr, &dbuf_evict_lock); } mutex_exit(&dbuf_evict_lock); /* * Keep evicting as long as we're above the low water mark * for the cache. We do this without holding the locks to * minimize lock contention. */ while (dbuf_cache_above_lowater() && !dbuf_evict_thread_exit) { dbuf_evict_one(); } mutex_enter(&dbuf_evict_lock); } dbuf_evict_thread_exit = B_FALSE; cv_broadcast(&dbuf_evict_cv); CALLB_CPR_EXIT(&cpr); /* drops dbuf_evict_lock */ thread_exit(); } /* * Wake up the dbuf eviction thread if the dbuf cache is at its max size. * If the dbuf cache is at its high water mark, then evict a dbuf from the * dbuf cache using the caller's context. */ static void dbuf_evict_notify(uint64_t size) { /* * We check if we should evict without holding the dbuf_evict_lock, * because it's OK to occasionally make the wrong decision here, * and grabbing the lock results in massive lock contention. */ if (size > dbuf_cache_target_bytes()) { if (size > dbuf_cache_hiwater_bytes()) dbuf_evict_one(); cv_signal(&dbuf_evict_cv); } } static int dbuf_kstat_update(kstat_t *ksp, int rw) { dbuf_stats_t *ds = ksp->ks_data; dbuf_hash_table_t *h = &dbuf_hash_table; if (rw == KSTAT_WRITE) return (SET_ERROR(EACCES)); ds->cache_count.value.ui64 = wmsum_value(&dbuf_sums.cache_count); ds->cache_size_bytes.value.ui64 = zfs_refcount_count(&dbuf_caches[DB_DBUF_CACHE].size); ds->cache_target_bytes.value.ui64 = dbuf_cache_target_bytes(); ds->cache_hiwater_bytes.value.ui64 = dbuf_cache_hiwater_bytes(); ds->cache_lowater_bytes.value.ui64 = dbuf_cache_lowater_bytes(); ds->cache_total_evicts.value.ui64 = wmsum_value(&dbuf_sums.cache_total_evicts); for (int i = 0; i < DN_MAX_LEVELS; i++) { ds->cache_levels[i].value.ui64 = wmsum_value(&dbuf_sums.cache_levels[i]); ds->cache_levels_bytes[i].value.ui64 = wmsum_value(&dbuf_sums.cache_levels_bytes[i]); } ds->hash_hits.value.ui64 = wmsum_value(&dbuf_sums.hash_hits); ds->hash_misses.value.ui64 = wmsum_value(&dbuf_sums.hash_misses); ds->hash_collisions.value.ui64 = wmsum_value(&dbuf_sums.hash_collisions); ds->hash_chains.value.ui64 = wmsum_value(&dbuf_sums.hash_chains); ds->hash_insert_race.value.ui64 = wmsum_value(&dbuf_sums.hash_insert_race); ds->hash_table_count.value.ui64 = h->hash_table_mask + 1; ds->hash_mutex_count.value.ui64 = h->hash_mutex_mask + 1; ds->metadata_cache_count.value.ui64 = wmsum_value(&dbuf_sums.metadata_cache_count); ds->metadata_cache_size_bytes.value.ui64 = zfs_refcount_count( &dbuf_caches[DB_DBUF_METADATA_CACHE].size); ds->metadata_cache_overflow.value.ui64 = wmsum_value(&dbuf_sums.metadata_cache_overflow); return (0); } void dbuf_init(void) { uint64_t hmsize, hsize = 1ULL << 16; dbuf_hash_table_t *h = &dbuf_hash_table; /* * The hash table is big enough to fill one eighth of physical memory * with an average block size of zfs_arc_average_blocksize (default 8K). * By default, the table will take up * totalmem * sizeof(void*) / 8K (1MB per GB with 8-byte pointers). */ while (hsize * zfs_arc_average_blocksize < arc_all_memory() / 8) hsize <<= 1; h->hash_table = NULL; while (h->hash_table == NULL) { h->hash_table_mask = hsize - 1; h->hash_table = vmem_zalloc(hsize * sizeof (void *), KM_SLEEP); if (h->hash_table == NULL) hsize >>= 1; ASSERT3U(hsize, >=, 1ULL << 10); } /* * The hash table buckets are protected by an array of mutexes where * each mutex is reponsible for protecting 128 buckets. A minimum * array size of 8192 is targeted to avoid contention. */ if (dbuf_mutex_cache_shift == 0) hmsize = MAX(hsize >> 7, 1ULL << 13); else hmsize = 1ULL << MIN(dbuf_mutex_cache_shift, 24); h->hash_mutexes = NULL; while (h->hash_mutexes == NULL) { h->hash_mutex_mask = hmsize - 1; h->hash_mutexes = vmem_zalloc(hmsize * sizeof (kmutex_t), KM_SLEEP); if (h->hash_mutexes == NULL) hmsize >>= 1; } dbuf_kmem_cache = kmem_cache_create("dmu_buf_impl_t", sizeof (dmu_buf_impl_t), 0, dbuf_cons, dbuf_dest, NULL, NULL, NULL, 0); for (int i = 0; i < hmsize; i++) mutex_init(&h->hash_mutexes[i], NULL, MUTEX_DEFAULT, NULL); dbuf_stats_init(h); /* * All entries are queued via taskq_dispatch_ent(), so min/maxalloc * configuration is not required. */ dbu_evict_taskq = taskq_create("dbu_evict", 1, defclsyspri, 0, 0, 0); for (dbuf_cached_state_t dcs = 0; dcs < DB_CACHE_MAX; dcs++) { multilist_create(&dbuf_caches[dcs].cache, sizeof (dmu_buf_impl_t), offsetof(dmu_buf_impl_t, db_cache_link), dbuf_cache_multilist_index_func); zfs_refcount_create(&dbuf_caches[dcs].size); } dbuf_evict_thread_exit = B_FALSE; mutex_init(&dbuf_evict_lock, NULL, MUTEX_DEFAULT, NULL); cv_init(&dbuf_evict_cv, NULL, CV_DEFAULT, NULL); dbuf_cache_evict_thread = thread_create(NULL, 0, dbuf_evict_thread, NULL, 0, &p0, TS_RUN, minclsyspri); wmsum_init(&dbuf_sums.cache_count, 0); wmsum_init(&dbuf_sums.cache_total_evicts, 0); for (int i = 0; i < DN_MAX_LEVELS; i++) { wmsum_init(&dbuf_sums.cache_levels[i], 0); wmsum_init(&dbuf_sums.cache_levels_bytes[i], 0); } wmsum_init(&dbuf_sums.hash_hits, 0); wmsum_init(&dbuf_sums.hash_misses, 0); wmsum_init(&dbuf_sums.hash_collisions, 0); wmsum_init(&dbuf_sums.hash_chains, 0); wmsum_init(&dbuf_sums.hash_insert_race, 0); wmsum_init(&dbuf_sums.metadata_cache_count, 0); wmsum_init(&dbuf_sums.metadata_cache_overflow, 0); dbuf_ksp = kstat_create("zfs", 0, "dbufstats", "misc", KSTAT_TYPE_NAMED, sizeof (dbuf_stats) / sizeof (kstat_named_t), KSTAT_FLAG_VIRTUAL); if (dbuf_ksp != NULL) { for (int i = 0; i < DN_MAX_LEVELS; i++) { snprintf(dbuf_stats.cache_levels[i].name, KSTAT_STRLEN, "cache_level_%d", i); dbuf_stats.cache_levels[i].data_type = KSTAT_DATA_UINT64; snprintf(dbuf_stats.cache_levels_bytes[i].name, KSTAT_STRLEN, "cache_level_%d_bytes", i); dbuf_stats.cache_levels_bytes[i].data_type = KSTAT_DATA_UINT64; } dbuf_ksp->ks_data = &dbuf_stats; dbuf_ksp->ks_update = dbuf_kstat_update; kstat_install(dbuf_ksp); } } void dbuf_fini(void) { dbuf_hash_table_t *h = &dbuf_hash_table; dbuf_stats_destroy(); for (int i = 0; i < (h->hash_mutex_mask + 1); i++) mutex_destroy(&h->hash_mutexes[i]); vmem_free(h->hash_table, (h->hash_table_mask + 1) * sizeof (void *)); vmem_free(h->hash_mutexes, (h->hash_mutex_mask + 1) * sizeof (kmutex_t)); kmem_cache_destroy(dbuf_kmem_cache); taskq_destroy(dbu_evict_taskq); mutex_enter(&dbuf_evict_lock); dbuf_evict_thread_exit = B_TRUE; while (dbuf_evict_thread_exit) { cv_signal(&dbuf_evict_cv); cv_wait(&dbuf_evict_cv, &dbuf_evict_lock); } mutex_exit(&dbuf_evict_lock); mutex_destroy(&dbuf_evict_lock); cv_destroy(&dbuf_evict_cv); for (dbuf_cached_state_t dcs = 0; dcs < DB_CACHE_MAX; dcs++) { zfs_refcount_destroy(&dbuf_caches[dcs].size); multilist_destroy(&dbuf_caches[dcs].cache); } if (dbuf_ksp != NULL) { kstat_delete(dbuf_ksp); dbuf_ksp = NULL; } wmsum_fini(&dbuf_sums.cache_count); wmsum_fini(&dbuf_sums.cache_total_evicts); for (int i = 0; i < DN_MAX_LEVELS; i++) { wmsum_fini(&dbuf_sums.cache_levels[i]); wmsum_fini(&dbuf_sums.cache_levels_bytes[i]); } wmsum_fini(&dbuf_sums.hash_hits); wmsum_fini(&dbuf_sums.hash_misses); wmsum_fini(&dbuf_sums.hash_collisions); wmsum_fini(&dbuf_sums.hash_chains); wmsum_fini(&dbuf_sums.hash_insert_race); wmsum_fini(&dbuf_sums.metadata_cache_count); wmsum_fini(&dbuf_sums.metadata_cache_overflow); } /* * Other stuff. */ #ifdef ZFS_DEBUG static void dbuf_verify(dmu_buf_impl_t *db) { dnode_t *dn; dbuf_dirty_record_t *dr; uint32_t txg_prev; ASSERT(MUTEX_HELD(&db->db_mtx)); if (!(zfs_flags & ZFS_DEBUG_DBUF_VERIFY)) return; ASSERT(db->db_objset != NULL); DB_DNODE_ENTER(db); dn = DB_DNODE(db); if (dn == NULL) { ASSERT(db->db_parent == NULL); ASSERT(db->db_blkptr == NULL); } else { ASSERT3U(db->db.db_object, ==, dn->dn_object); ASSERT3P(db->db_objset, ==, dn->dn_objset); ASSERT3U(db->db_level, <, dn->dn_nlevels); ASSERT(db->db_blkid == DMU_BONUS_BLKID || db->db_blkid == DMU_SPILL_BLKID || !avl_is_empty(&dn->dn_dbufs)); } if (db->db_blkid == DMU_BONUS_BLKID) { ASSERT(dn != NULL); ASSERT3U(db->db.db_size, >=, dn->dn_bonuslen); ASSERT3U(db->db.db_offset, ==, DMU_BONUS_BLKID); } else if (db->db_blkid == DMU_SPILL_BLKID) { ASSERT(dn != NULL); ASSERT0(db->db.db_offset); } else { ASSERT3U(db->db.db_offset, ==, db->db_blkid * db->db.db_size); } if ((dr = list_head(&db->db_dirty_records)) != NULL) { ASSERT(dr->dr_dbuf == db); txg_prev = dr->dr_txg; for (dr = list_next(&db->db_dirty_records, dr); dr != NULL; dr = list_next(&db->db_dirty_records, dr)) { ASSERT(dr->dr_dbuf == db); ASSERT(txg_prev > dr->dr_txg); txg_prev = dr->dr_txg; } } /* * We can't assert that db_size matches dn_datablksz because it * can be momentarily different when another thread is doing * dnode_set_blksz(). */ if (db->db_level == 0 && db->db.db_object == DMU_META_DNODE_OBJECT) { dr = db->db_data_pending; /* * It should only be modified in syncing context, so * make sure we only have one copy of the data. */ ASSERT(dr == NULL || dr->dt.dl.dr_data == db->db_buf); } /* verify db->db_blkptr */ if (db->db_blkptr) { if (db->db_parent == dn->dn_dbuf) { /* db is pointed to by the dnode */ /* ASSERT3U(db->db_blkid, <, dn->dn_nblkptr); */ if (DMU_OBJECT_IS_SPECIAL(db->db.db_object)) ASSERT(db->db_parent == NULL); else ASSERT(db->db_parent != NULL); if (db->db_blkid != DMU_SPILL_BLKID) ASSERT3P(db->db_blkptr, ==, &dn->dn_phys->dn_blkptr[db->db_blkid]); } else { /* db is pointed to by an indirect block */ int epb __maybe_unused = db->db_parent->db.db_size >> SPA_BLKPTRSHIFT; ASSERT3U(db->db_parent->db_level, ==, db->db_level+1); ASSERT3U(db->db_parent->db.db_object, ==, db->db.db_object); /* * dnode_grow_indblksz() can make this fail if we don't * have the parent's rwlock. XXX indblksz no longer * grows. safe to do this now? */ if (RW_LOCK_HELD(&db->db_parent->db_rwlock)) { ASSERT3P(db->db_blkptr, ==, ((blkptr_t *)db->db_parent->db.db_data + db->db_blkid % epb)); } } } if ((db->db_blkptr == NULL || BP_IS_HOLE(db->db_blkptr)) && (db->db_buf == NULL || db->db_buf->b_data) && db->db.db_data && db->db_blkid != DMU_BONUS_BLKID && db->db_state != DB_FILL && (dn == NULL || !dn->dn_free_txg)) { /* * If the blkptr isn't set but they have nonzero data, * it had better be dirty, otherwise we'll lose that * data when we evict this buffer. * * There is an exception to this rule for indirect blocks; in * this case, if the indirect block is a hole, we fill in a few * fields on each of the child blocks (importantly, birth time) * to prevent hole birth times from being lost when you * partially fill in a hole. */ if (db->db_dirtycnt == 0) { if (db->db_level == 0) { uint64_t *buf = db->db.db_data; int i; for (i = 0; i < db->db.db_size >> 3; i++) { ASSERT(buf[i] == 0); } } else { blkptr_t *bps = db->db.db_data; ASSERT3U(1 << DB_DNODE(db)->dn_indblkshift, ==, db->db.db_size); /* * We want to verify that all the blkptrs in the * indirect block are holes, but we may have * automatically set up a few fields for them. * We iterate through each blkptr and verify * they only have those fields set. */ for (int i = 0; i < db->db.db_size / sizeof (blkptr_t); i++) { blkptr_t *bp = &bps[i]; ASSERT(ZIO_CHECKSUM_IS_ZERO( &bp->blk_cksum)); ASSERT( DVA_IS_EMPTY(&bp->blk_dva[0]) && DVA_IS_EMPTY(&bp->blk_dva[1]) && DVA_IS_EMPTY(&bp->blk_dva[2])); ASSERT0(bp->blk_fill); ASSERT0(bp->blk_pad[0]); ASSERT0(bp->blk_pad[1]); ASSERT(!BP_IS_EMBEDDED(bp)); ASSERT(BP_IS_HOLE(bp)); ASSERT0(bp->blk_phys_birth); } } } } DB_DNODE_EXIT(db); } #endif static void dbuf_clear_data(dmu_buf_impl_t *db) { ASSERT(MUTEX_HELD(&db->db_mtx)); dbuf_evict_user(db); ASSERT3P(db->db_buf, ==, NULL); db->db.db_data = NULL; if (db->db_state != DB_NOFILL) { db->db_state = DB_UNCACHED; DTRACE_SET_STATE(db, "clear data"); } } static void dbuf_set_data(dmu_buf_impl_t *db, arc_buf_t *buf) { ASSERT(MUTEX_HELD(&db->db_mtx)); ASSERT(buf != NULL); db->db_buf = buf; ASSERT(buf->b_data != NULL); db->db.db_data = buf->b_data; } static arc_buf_t * dbuf_alloc_arcbuf(dmu_buf_impl_t *db) { spa_t *spa = db->db_objset->os_spa; return (arc_alloc_buf(spa, db, DBUF_GET_BUFC_TYPE(db), db->db.db_size)); } /* * Loan out an arc_buf for read. Return the loaned arc_buf. */ arc_buf_t * dbuf_loan_arcbuf(dmu_buf_impl_t *db) { arc_buf_t *abuf; ASSERT(db->db_blkid != DMU_BONUS_BLKID); mutex_enter(&db->db_mtx); if (arc_released(db->db_buf) || zfs_refcount_count(&db->db_holds) > 1) { int blksz = db->db.db_size; spa_t *spa = db->db_objset->os_spa; mutex_exit(&db->db_mtx); abuf = arc_loan_buf(spa, B_FALSE, blksz); memcpy(abuf->b_data, db->db.db_data, blksz); } else { abuf = db->db_buf; arc_loan_inuse_buf(abuf, db); db->db_buf = NULL; dbuf_clear_data(db); mutex_exit(&db->db_mtx); } return (abuf); } /* * Calculate which level n block references the data at the level 0 offset * provided. */ uint64_t dbuf_whichblock(const dnode_t *dn, const int64_t level, const uint64_t offset) { if (dn->dn_datablkshift != 0 && dn->dn_indblkshift != 0) { /* * The level n blkid is equal to the level 0 blkid divided by * the number of level 0s in a level n block. * * The level 0 blkid is offset >> datablkshift = * offset / 2^datablkshift. * * The number of level 0s in a level n is the number of block * pointers in an indirect block, raised to the power of level. * This is 2^(indblkshift - SPA_BLKPTRSHIFT)^level = * 2^(level*(indblkshift - SPA_BLKPTRSHIFT)). * * Thus, the level n blkid is: offset / * ((2^datablkshift)*(2^(level*(indblkshift-SPA_BLKPTRSHIFT)))) * = offset / 2^(datablkshift + level * * (indblkshift - SPA_BLKPTRSHIFT)) * = offset >> (datablkshift + level * * (indblkshift - SPA_BLKPTRSHIFT)) */ const unsigned exp = dn->dn_datablkshift + level * (dn->dn_indblkshift - SPA_BLKPTRSHIFT); if (exp >= 8 * sizeof (offset)) { /* This only happens on the highest indirection level */ ASSERT3U(level, ==, dn->dn_nlevels - 1); return (0); } ASSERT3U(exp, <, 8 * sizeof (offset)); return (offset >> exp); } else { ASSERT3U(offset, <, dn->dn_datablksz); return (0); } } /* * This function is used to lock the parent of the provided dbuf. This should be * used when modifying or reading db_blkptr. */ db_lock_type_t dmu_buf_lock_parent(dmu_buf_impl_t *db, krw_t rw, const void *tag) { enum db_lock_type ret = DLT_NONE; if (db->db_parent != NULL) { rw_enter(&db->db_parent->db_rwlock, rw); ret = DLT_PARENT; } else if (dmu_objset_ds(db->db_objset) != NULL) { rrw_enter(&dmu_objset_ds(db->db_objset)->ds_bp_rwlock, rw, tag); ret = DLT_OBJSET; } /* * We only return a DLT_NONE lock when it's the top-most indirect block * of the meta-dnode of the MOS. */ return (ret); } /* * We need to pass the lock type in because it's possible that the block will * move from being the topmost indirect block in a dnode (and thus, have no * parent) to not the top-most via an indirection increase. This would cause a * panic if we didn't pass the lock type in. */ void dmu_buf_unlock_parent(dmu_buf_impl_t *db, db_lock_type_t type, const void *tag) { if (type == DLT_PARENT) rw_exit(&db->db_parent->db_rwlock); else if (type == DLT_OBJSET) rrw_exit(&dmu_objset_ds(db->db_objset)->ds_bp_rwlock, tag); } static void dbuf_read_done(zio_t *zio, const zbookmark_phys_t *zb, const blkptr_t *bp, arc_buf_t *buf, void *vdb) { (void) zb, (void) bp; dmu_buf_impl_t *db = vdb; mutex_enter(&db->db_mtx); ASSERT3U(db->db_state, ==, DB_READ); /* * All reads are synchronous, so we must have a hold on the dbuf */ ASSERT(zfs_refcount_count(&db->db_holds) > 0); ASSERT(db->db_buf == NULL); ASSERT(db->db.db_data == NULL); if (buf == NULL) { /* i/o error */ ASSERT(zio == NULL || zio->io_error != 0); ASSERT(db->db_blkid != DMU_BONUS_BLKID); ASSERT3P(db->db_buf, ==, NULL); db->db_state = DB_UNCACHED; DTRACE_SET_STATE(db, "i/o error"); } else if (db->db_level == 0 && db->db_freed_in_flight) { /* freed in flight */ ASSERT(zio == NULL || zio->io_error == 0); arc_release(buf, db); memset(buf->b_data, 0, db->db.db_size); arc_buf_freeze(buf); db->db_freed_in_flight = FALSE; dbuf_set_data(db, buf); db->db_state = DB_CACHED; DTRACE_SET_STATE(db, "freed in flight"); } else { /* success */ ASSERT(zio == NULL || zio->io_error == 0); dbuf_set_data(db, buf); db->db_state = DB_CACHED; DTRACE_SET_STATE(db, "successful read"); } cv_broadcast(&db->db_changed); dbuf_rele_and_unlock(db, NULL, B_FALSE); } /* * Shortcut for performing reads on bonus dbufs. Returns * an error if we fail to verify the dnode associated with * a decrypted block. Otherwise success. */ static int dbuf_read_bonus(dmu_buf_impl_t *db, dnode_t *dn, uint32_t flags) { int bonuslen, max_bonuslen, err; err = dbuf_read_verify_dnode_crypt(db, flags); if (err) return (err); bonuslen = MIN(dn->dn_bonuslen, dn->dn_phys->dn_bonuslen); max_bonuslen = DN_SLOTS_TO_BONUSLEN(dn->dn_num_slots); ASSERT(MUTEX_HELD(&db->db_mtx)); ASSERT(DB_DNODE_HELD(db)); ASSERT3U(bonuslen, <=, db->db.db_size); db->db.db_data = kmem_alloc(max_bonuslen, KM_SLEEP); arc_space_consume(max_bonuslen, ARC_SPACE_BONUS); if (bonuslen < max_bonuslen) memset(db->db.db_data, 0, max_bonuslen); if (bonuslen) memcpy(db->db.db_data, DN_BONUS(dn->dn_phys), bonuslen); db->db_state = DB_CACHED; DTRACE_SET_STATE(db, "bonus buffer filled"); return (0); } static void dbuf_handle_indirect_hole(dmu_buf_impl_t *db, dnode_t *dn, blkptr_t *dbbp) { blkptr_t *bps = db->db.db_data; uint32_t indbs = 1ULL << dn->dn_indblkshift; int n_bps = indbs >> SPA_BLKPTRSHIFT; for (int i = 0; i < n_bps; i++) { blkptr_t *bp = &bps[i]; ASSERT3U(BP_GET_LSIZE(dbbp), ==, indbs); BP_SET_LSIZE(bp, BP_GET_LEVEL(dbbp) == 1 ? dn->dn_datablksz : BP_GET_LSIZE(dbbp)); BP_SET_TYPE(bp, BP_GET_TYPE(dbbp)); BP_SET_LEVEL(bp, BP_GET_LEVEL(dbbp) - 1); BP_SET_BIRTH(bp, dbbp->blk_birth, 0); } } /* * Handle reads on dbufs that are holes, if necessary. This function * requires that the dbuf's mutex is held. Returns success (0) if action * was taken, ENOENT if no action was taken. */ static int dbuf_read_hole(dmu_buf_impl_t *db, dnode_t *dn, blkptr_t *bp) { ASSERT(MUTEX_HELD(&db->db_mtx)); int is_hole = bp == NULL || BP_IS_HOLE(bp); /* * For level 0 blocks only, if the above check fails: * Recheck BP_IS_HOLE() after dnode_block_freed() in case dnode_sync() * processes the delete record and clears the bp while we are waiting * for the dn_mtx (resulting in a "no" from block_freed). */ if (!is_hole && db->db_level == 0) is_hole = dnode_block_freed(dn, db->db_blkid) || BP_IS_HOLE(bp); if (is_hole) { dbuf_set_data(db, dbuf_alloc_arcbuf(db)); memset(db->db.db_data, 0, db->db.db_size); if (bp != NULL && db->db_level > 0 && BP_IS_HOLE(bp) && bp->blk_birth != 0) { dbuf_handle_indirect_hole(db, dn, bp); } db->db_state = DB_CACHED; DTRACE_SET_STATE(db, "hole read satisfied"); return (0); } return (ENOENT); } /* * This function ensures that, when doing a decrypting read of a block, * we make sure we have decrypted the dnode associated with it. We must do * this so that we ensure we are fully authenticating the checksum-of-MACs * tree from the root of the objset down to this block. Indirect blocks are * always verified against their secure checksum-of-MACs assuming that the * dnode containing them is correct. Now that we are doing a decrypting read, * we can be sure that the key is loaded and verify that assumption. This is * especially important considering that we always read encrypted dnode * blocks as raw data (without verifying their MACs) to start, and * decrypt / authenticate them when we need to read an encrypted bonus buffer. */ static int dbuf_read_verify_dnode_crypt(dmu_buf_impl_t *db, uint32_t flags) { int err = 0; objset_t *os = db->db_objset; arc_buf_t *dnode_abuf; dnode_t *dn; zbookmark_phys_t zb; ASSERT(MUTEX_HELD(&db->db_mtx)); if ((flags & DB_RF_NO_DECRYPT) != 0 || !os->os_encrypted || os->os_raw_receive) return (0); DB_DNODE_ENTER(db); dn = DB_DNODE(db); dnode_abuf = (dn->dn_dbuf != NULL) ? dn->dn_dbuf->db_buf : NULL; if (dnode_abuf == NULL || !arc_is_encrypted(dnode_abuf)) { DB_DNODE_EXIT(db); return (0); } SET_BOOKMARK(&zb, dmu_objset_id(os), DMU_META_DNODE_OBJECT, 0, dn->dn_dbuf->db_blkid); err = arc_untransform(dnode_abuf, os->os_spa, &zb, B_TRUE); /* * An error code of EACCES tells us that the key is still not * available. This is ok if we are only reading authenticated * (and therefore non-encrypted) blocks. */ if (err == EACCES && ((db->db_blkid != DMU_BONUS_BLKID && !DMU_OT_IS_ENCRYPTED(dn->dn_type)) || (db->db_blkid == DMU_BONUS_BLKID && !DMU_OT_IS_ENCRYPTED(dn->dn_bonustype)))) err = 0; DB_DNODE_EXIT(db); return (err); } /* * Drops db_mtx and the parent lock specified by dblt and tag before * returning. */ static int dbuf_read_impl(dmu_buf_impl_t *db, zio_t *zio, uint32_t flags, db_lock_type_t dblt, const void *tag) { dnode_t *dn; zbookmark_phys_t zb; uint32_t aflags = ARC_FLAG_NOWAIT; int err, zio_flags; blkptr_t bp, *bpp; DB_DNODE_ENTER(db); dn = DB_DNODE(db); ASSERT(!zfs_refcount_is_zero(&db->db_holds)); ASSERT(MUTEX_HELD(&db->db_mtx)); ASSERT(db->db_state == DB_UNCACHED || db->db_state == DB_NOFILL); ASSERT(db->db_buf == NULL); ASSERT(db->db_parent == NULL || RW_LOCK_HELD(&db->db_parent->db_rwlock)); if (db->db_blkid == DMU_BONUS_BLKID) { err = dbuf_read_bonus(db, dn, flags); goto early_unlock; } if (db->db_state == DB_UNCACHED) { if (db->db_blkptr == NULL) { bpp = NULL; } else { bp = *db->db_blkptr; bpp = &bp; } } else { dbuf_dirty_record_t *dr; ASSERT3S(db->db_state, ==, DB_NOFILL); /* * Block cloning: If we have a pending block clone, * we don't want to read the underlying block, but the content * of the block being cloned, so we have the most recent data. */ dr = list_head(&db->db_dirty_records); if (dr == NULL || !dr->dt.dl.dr_brtwrite) { err = EIO; goto early_unlock; } bp = dr->dt.dl.dr_overridden_by; bpp = &bp; } err = dbuf_read_hole(db, dn, bpp); if (err == 0) goto early_unlock; ASSERT(bpp != NULL); /* * Any attempt to read a redacted block should result in an error. This * will never happen under normal conditions, but can be useful for * debugging purposes. */ if (BP_IS_REDACTED(bpp)) { ASSERT(dsl_dataset_feature_is_active( db->db_objset->os_dsl_dataset, SPA_FEATURE_REDACTED_DATASETS)); err = SET_ERROR(EIO); goto early_unlock; } SET_BOOKMARK(&zb, dmu_objset_id(db->db_objset), db->db.db_object, db->db_level, db->db_blkid); /* * All bps of an encrypted os should have the encryption bit set. * If this is not true it indicates tampering and we report an error. */ if (db->db_objset->os_encrypted && !BP_USES_CRYPT(bpp)) { spa_log_error(db->db_objset->os_spa, &zb, &bpp->blk_birth); zfs_panic_recover("unencrypted block in encrypted " "object set %llu", dmu_objset_id(db->db_objset)); err = SET_ERROR(EIO); goto early_unlock; } err = dbuf_read_verify_dnode_crypt(db, flags); if (err != 0) goto early_unlock; DB_DNODE_EXIT(db); db->db_state = DB_READ; DTRACE_SET_STATE(db, "read issued"); mutex_exit(&db->db_mtx); if (!DBUF_IS_CACHEABLE(db)) aflags |= ARC_FLAG_UNCACHED; else if (dbuf_is_l2cacheable(db)) aflags |= ARC_FLAG_L2CACHE; dbuf_add_ref(db, NULL); zio_flags = (flags & DB_RF_CANFAIL) ? ZIO_FLAG_CANFAIL : ZIO_FLAG_MUSTSUCCEED; if ((flags & DB_RF_NO_DECRYPT) && BP_IS_PROTECTED(db->db_blkptr)) zio_flags |= ZIO_FLAG_RAW; /* * The zio layer will copy the provided blkptr later, but we have our * own copy so that we can release the parent's rwlock. We have to * do that so that if dbuf_read_done is called synchronously (on * an l1 cache hit) we don't acquire the db_mtx while holding the * parent's rwlock, which would be a lock ordering violation. */ dmu_buf_unlock_parent(db, dblt, tag); (void) arc_read(zio, db->db_objset->os_spa, bpp, dbuf_read_done, db, ZIO_PRIORITY_SYNC_READ, zio_flags, &aflags, &zb); return (err); early_unlock: DB_DNODE_EXIT(db); mutex_exit(&db->db_mtx); dmu_buf_unlock_parent(db, dblt, tag); return (err); } /* * This is our just-in-time copy function. It makes a copy of buffers that * have been modified in a previous transaction group before we access them in * the current active group. * * This function is used in three places: when we are dirtying a buffer for the * first time in a txg, when we are freeing a range in a dnode that includes * this buffer, and when we are accessing a buffer which was received compressed * and later referenced in a WRITE_BYREF record. * * Note that when we are called from dbuf_free_range() we do not put a hold on * the buffer, we just traverse the active dbuf list for the dnode. */ static void dbuf_fix_old_data(dmu_buf_impl_t *db, uint64_t txg) { dbuf_dirty_record_t *dr = list_head(&db->db_dirty_records); ASSERT(MUTEX_HELD(&db->db_mtx)); ASSERT(db->db.db_data != NULL); ASSERT(db->db_level == 0); ASSERT(db->db.db_object != DMU_META_DNODE_OBJECT); if (dr == NULL || (dr->dt.dl.dr_data != ((db->db_blkid == DMU_BONUS_BLKID) ? db->db.db_data : db->db_buf))) return; /* * If the last dirty record for this dbuf has not yet synced * and its referencing the dbuf data, either: * reset the reference to point to a new copy, * or (if there a no active holders) * just null out the current db_data pointer. */ ASSERT3U(dr->dr_txg, >=, txg - 2); if (db->db_blkid == DMU_BONUS_BLKID) { dnode_t *dn = DB_DNODE(db); int bonuslen = DN_SLOTS_TO_BONUSLEN(dn->dn_num_slots); dr->dt.dl.dr_data = kmem_alloc(bonuslen, KM_SLEEP); arc_space_consume(bonuslen, ARC_SPACE_BONUS); memcpy(dr->dt.dl.dr_data, db->db.db_data, bonuslen); } else if (zfs_refcount_count(&db->db_holds) > db->db_dirtycnt) { dnode_t *dn = DB_DNODE(db); int size = arc_buf_size(db->db_buf); arc_buf_contents_t type = DBUF_GET_BUFC_TYPE(db); spa_t *spa = db->db_objset->os_spa; enum zio_compress compress_type = arc_get_compression(db->db_buf); uint8_t complevel = arc_get_complevel(db->db_buf); if (arc_is_encrypted(db->db_buf)) { boolean_t byteorder; uint8_t salt[ZIO_DATA_SALT_LEN]; uint8_t iv[ZIO_DATA_IV_LEN]; uint8_t mac[ZIO_DATA_MAC_LEN]; arc_get_raw_params(db->db_buf, &byteorder, salt, iv, mac); dr->dt.dl.dr_data = arc_alloc_raw_buf(spa, db, dmu_objset_id(dn->dn_objset), byteorder, salt, iv, mac, dn->dn_type, size, arc_buf_lsize(db->db_buf), compress_type, complevel); } else if (compress_type != ZIO_COMPRESS_OFF) { ASSERT3U(type, ==, ARC_BUFC_DATA); dr->dt.dl.dr_data = arc_alloc_compressed_buf(spa, db, size, arc_buf_lsize(db->db_buf), compress_type, complevel); } else { dr->dt.dl.dr_data = arc_alloc_buf(spa, db, type, size); } memcpy(dr->dt.dl.dr_data->b_data, db->db.db_data, size); } else { db->db_buf = NULL; dbuf_clear_data(db); } } int dbuf_read(dmu_buf_impl_t *db, zio_t *zio, uint32_t flags) { int err = 0; boolean_t prefetch; dnode_t *dn; /* * We don't have to hold the mutex to check db_state because it * can't be freed while we have a hold on the buffer. */ ASSERT(!zfs_refcount_is_zero(&db->db_holds)); DB_DNODE_ENTER(db); dn = DB_DNODE(db); prefetch = db->db_level == 0 && db->db_blkid != DMU_BONUS_BLKID && (flags & DB_RF_NOPREFETCH) == 0 && dn != NULL; mutex_enter(&db->db_mtx); if (flags & DB_RF_PARTIAL_FIRST) db->db_partial_read = B_TRUE; else if (!(flags & DB_RF_PARTIAL_MORE)) db->db_partial_read = B_FALSE; if (db->db_state == DB_CACHED) { /* * Ensure that this block's dnode has been decrypted if * the caller has requested decrypted data. */ err = dbuf_read_verify_dnode_crypt(db, flags); /* * If the arc buf is compressed or encrypted and the caller * requested uncompressed data, we need to untransform it * before returning. We also call arc_untransform() on any * unauthenticated blocks, which will verify their MAC if * the key is now available. */ if (err == 0 && db->db_buf != NULL && (flags & DB_RF_NO_DECRYPT) == 0 && (arc_is_encrypted(db->db_buf) || arc_is_unauthenticated(db->db_buf) || arc_get_compression(db->db_buf) != ZIO_COMPRESS_OFF)) { spa_t *spa = dn->dn_objset->os_spa; zbookmark_phys_t zb; SET_BOOKMARK(&zb, dmu_objset_id(db->db_objset), db->db.db_object, db->db_level, db->db_blkid); dbuf_fix_old_data(db, spa_syncing_txg(spa)); err = arc_untransform(db->db_buf, spa, &zb, B_FALSE); dbuf_set_data(db, db->db_buf); } mutex_exit(&db->db_mtx); if (err == 0 && prefetch) { dmu_zfetch(&dn->dn_zfetch, db->db_blkid, 1, B_TRUE, B_FALSE, flags & DB_RF_HAVESTRUCT); } DB_DNODE_EXIT(db); DBUF_STAT_BUMP(hash_hits); } else if (db->db_state == DB_UNCACHED || db->db_state == DB_NOFILL) { boolean_t need_wait = B_FALSE; db_lock_type_t dblt = dmu_buf_lock_parent(db, RW_READER, FTAG); if (zio == NULL && (db->db_state == DB_NOFILL || (db->db_blkptr != NULL && !BP_IS_HOLE(db->db_blkptr)))) { spa_t *spa = dn->dn_objset->os_spa; zio = zio_root(spa, NULL, NULL, ZIO_FLAG_CANFAIL); need_wait = B_TRUE; } err = dbuf_read_impl(db, zio, flags, dblt, FTAG); /* * dbuf_read_impl has dropped db_mtx and our parent's rwlock * for us */ if (!err && prefetch) { dmu_zfetch(&dn->dn_zfetch, db->db_blkid, 1, B_TRUE, db->db_state != DB_CACHED, flags & DB_RF_HAVESTRUCT); } DB_DNODE_EXIT(db); DBUF_STAT_BUMP(hash_misses); /* * If we created a zio_root we must execute it to avoid * leaking it, even if it isn't attached to any work due * to an error in dbuf_read_impl(). */ if (need_wait) { if (err == 0) err = zio_wait(zio); else VERIFY0(zio_wait(zio)); } } else { /* * Another reader came in while the dbuf was in flight * between UNCACHED and CACHED. Either a writer will finish * writing the buffer (sending the dbuf to CACHED) or the * first reader's request will reach the read_done callback * and send the dbuf to CACHED. Otherwise, a failure * occurred and the dbuf went to UNCACHED. */ mutex_exit(&db->db_mtx); if (prefetch) { dmu_zfetch(&dn->dn_zfetch, db->db_blkid, 1, B_TRUE, B_TRUE, flags & DB_RF_HAVESTRUCT); } DB_DNODE_EXIT(db); DBUF_STAT_BUMP(hash_misses); /* Skip the wait per the caller's request. */ if ((flags & DB_RF_NEVERWAIT) == 0) { mutex_enter(&db->db_mtx); while (db->db_state == DB_READ || db->db_state == DB_FILL) { ASSERT(db->db_state == DB_READ || (flags & DB_RF_HAVESTRUCT) == 0); DTRACE_PROBE2(blocked__read, dmu_buf_impl_t *, db, zio_t *, zio); cv_wait(&db->db_changed, &db->db_mtx); } if (db->db_state == DB_UNCACHED) err = SET_ERROR(EIO); mutex_exit(&db->db_mtx); } } return (err); } static void dbuf_noread(dmu_buf_impl_t *db) { ASSERT(!zfs_refcount_is_zero(&db->db_holds)); ASSERT(db->db_blkid != DMU_BONUS_BLKID); mutex_enter(&db->db_mtx); while (db->db_state == DB_READ || db->db_state == DB_FILL) cv_wait(&db->db_changed, &db->db_mtx); if (db->db_state == DB_UNCACHED) { ASSERT(db->db_buf == NULL); ASSERT(db->db.db_data == NULL); dbuf_set_data(db, dbuf_alloc_arcbuf(db)); db->db_state = DB_FILL; DTRACE_SET_STATE(db, "assigning filled buffer"); } else if (db->db_state == DB_NOFILL) { dbuf_clear_data(db); } else { ASSERT3U(db->db_state, ==, DB_CACHED); } mutex_exit(&db->db_mtx); } void dbuf_unoverride(dbuf_dirty_record_t *dr) { dmu_buf_impl_t *db = dr->dr_dbuf; blkptr_t *bp = &dr->dt.dl.dr_overridden_by; uint64_t txg = dr->dr_txg; boolean_t release; ASSERT(MUTEX_HELD(&db->db_mtx)); /* * This assert is valid because dmu_sync() expects to be called by * a zilog's get_data while holding a range lock. This call only * comes from dbuf_dirty() callers who must also hold a range lock. */ ASSERT(dr->dt.dl.dr_override_state != DR_IN_DMU_SYNC); ASSERT(db->db_level == 0); if (db->db_blkid == DMU_BONUS_BLKID || dr->dt.dl.dr_override_state == DR_NOT_OVERRIDDEN) return; ASSERT(db->db_data_pending != dr); /* free this block */ if (!BP_IS_HOLE(bp) && !dr->dt.dl.dr_nopwrite) zio_free(db->db_objset->os_spa, txg, bp); release = !dr->dt.dl.dr_brtwrite; dr->dt.dl.dr_override_state = DR_NOT_OVERRIDDEN; dr->dt.dl.dr_nopwrite = B_FALSE; dr->dt.dl.dr_brtwrite = B_FALSE; dr->dt.dl.dr_has_raw_params = B_FALSE; /* * Release the already-written buffer, so we leave it in * a consistent dirty state. Note that all callers are * modifying the buffer, so they will immediately do * another (redundant) arc_release(). Therefore, leave * the buf thawed to save the effort of freezing & * immediately re-thawing it. */ if (release) arc_release(dr->dt.dl.dr_data, db); } /* * Evict (if its unreferenced) or clear (if its referenced) any level-0 * data blocks in the free range, so that any future readers will find * empty blocks. */ void dbuf_free_range(dnode_t *dn, uint64_t start_blkid, uint64_t end_blkid, dmu_tx_t *tx) { dmu_buf_impl_t *db_search; dmu_buf_impl_t *db, *db_next; uint64_t txg = tx->tx_txg; avl_index_t where; dbuf_dirty_record_t *dr; if (end_blkid > dn->dn_maxblkid && !(start_blkid == DMU_SPILL_BLKID || end_blkid == DMU_SPILL_BLKID)) end_blkid = dn->dn_maxblkid; dprintf_dnode(dn, "start=%llu end=%llu\n", (u_longlong_t)start_blkid, (u_longlong_t)end_blkid); db_search = kmem_alloc(sizeof (dmu_buf_impl_t), KM_SLEEP); db_search->db_level = 0; db_search->db_blkid = start_blkid; db_search->db_state = DB_SEARCH; mutex_enter(&dn->dn_dbufs_mtx); db = avl_find(&dn->dn_dbufs, db_search, &where); ASSERT3P(db, ==, NULL); db = avl_nearest(&dn->dn_dbufs, where, AVL_AFTER); for (; db != NULL; db = db_next) { db_next = AVL_NEXT(&dn->dn_dbufs, db); ASSERT(db->db_blkid != DMU_BONUS_BLKID); if (db->db_level != 0 || db->db_blkid > end_blkid) { break; } ASSERT3U(db->db_blkid, >=, start_blkid); /* found a level 0 buffer in the range */ mutex_enter(&db->db_mtx); if (dbuf_undirty(db, tx)) { /* mutex has been dropped and dbuf destroyed */ continue; } if (db->db_state == DB_UNCACHED || db->db_state == DB_NOFILL || db->db_state == DB_EVICTING) { ASSERT(db->db.db_data == NULL); mutex_exit(&db->db_mtx); continue; } if (db->db_state == DB_READ || db->db_state == DB_FILL) { /* will be handled in dbuf_read_done or dbuf_rele */ db->db_freed_in_flight = TRUE; mutex_exit(&db->db_mtx); continue; } if (zfs_refcount_count(&db->db_holds) == 0) { ASSERT(db->db_buf); dbuf_destroy(db); continue; } /* The dbuf is referenced */ dr = list_head(&db->db_dirty_records); if (dr != NULL) { if (dr->dr_txg == txg) { /* * This buffer is "in-use", re-adjust the file * size to reflect that this buffer may * contain new data when we sync. */ if (db->db_blkid != DMU_SPILL_BLKID && db->db_blkid > dn->dn_maxblkid) dn->dn_maxblkid = db->db_blkid; dbuf_unoverride(dr); } else { /* * This dbuf is not dirty in the open context. * Either uncache it (if its not referenced in * the open context) or reset its contents to * empty. */ dbuf_fix_old_data(db, txg); } } /* clear the contents if its cached */ if (db->db_state == DB_CACHED) { ASSERT(db->db.db_data != NULL); arc_release(db->db_buf, db); rw_enter(&db->db_rwlock, RW_WRITER); memset(db->db.db_data, 0, db->db.db_size); rw_exit(&db->db_rwlock); arc_buf_freeze(db->db_buf); } mutex_exit(&db->db_mtx); } mutex_exit(&dn->dn_dbufs_mtx); kmem_free(db_search, sizeof (dmu_buf_impl_t)); } void dbuf_new_size(dmu_buf_impl_t *db, int size, dmu_tx_t *tx) { arc_buf_t *buf, *old_buf; dbuf_dirty_record_t *dr; int osize = db->db.db_size; arc_buf_contents_t type = DBUF_GET_BUFC_TYPE(db); dnode_t *dn; ASSERT(db->db_blkid != DMU_BONUS_BLKID); DB_DNODE_ENTER(db); dn = DB_DNODE(db); /* * XXX we should be doing a dbuf_read, checking the return * value and returning that up to our callers */ dmu_buf_will_dirty(&db->db, tx); /* create the data buffer for the new block */ buf = arc_alloc_buf(dn->dn_objset->os_spa, db, type, size); /* copy old block data to the new block */ old_buf = db->db_buf; memcpy(buf->b_data, old_buf->b_data, MIN(osize, size)); /* zero the remainder */ if (size > osize) memset((uint8_t *)buf->b_data + osize, 0, size - osize); mutex_enter(&db->db_mtx); dbuf_set_data(db, buf); arc_buf_destroy(old_buf, db); db->db.db_size = size; dr = list_head(&db->db_dirty_records); /* dirty record added by dmu_buf_will_dirty() */ VERIFY(dr != NULL); if (db->db_level == 0) dr->dt.dl.dr_data = buf; ASSERT3U(dr->dr_txg, ==, tx->tx_txg); ASSERT3U(dr->dr_accounted, ==, osize); dr->dr_accounted = size; mutex_exit(&db->db_mtx); dmu_objset_willuse_space(dn->dn_objset, size - osize, tx); DB_DNODE_EXIT(db); } void dbuf_release_bp(dmu_buf_impl_t *db) { objset_t *os __maybe_unused = db->db_objset; ASSERT(dsl_pool_sync_context(dmu_objset_pool(os))); ASSERT(arc_released(os->os_phys_buf) || list_link_active(&os->os_dsl_dataset->ds_synced_link)); ASSERT(db->db_parent == NULL || arc_released(db->db_parent->db_buf)); (void) arc_release(db->db_buf, db); } /* * We already have a dirty record for this TXG, and we are being * dirtied again. */ static void dbuf_redirty(dbuf_dirty_record_t *dr) { dmu_buf_impl_t *db = dr->dr_dbuf; ASSERT(MUTEX_HELD(&db->db_mtx)); if (db->db_level == 0 && db->db_blkid != DMU_BONUS_BLKID) { /* * If this buffer has already been written out, * we now need to reset its state. */ dbuf_unoverride(dr); if (db->db.db_object != DMU_META_DNODE_OBJECT && db->db_state != DB_NOFILL) { /* Already released on initial dirty, so just thaw. */ ASSERT(arc_released(db->db_buf)); arc_buf_thaw(db->db_buf); } } } dbuf_dirty_record_t * dbuf_dirty_lightweight(dnode_t *dn, uint64_t blkid, dmu_tx_t *tx) { rw_enter(&dn->dn_struct_rwlock, RW_READER); IMPLY(dn->dn_objset->os_raw_receive, dn->dn_maxblkid >= blkid); dnode_new_blkid(dn, blkid, tx, B_TRUE, B_FALSE); ASSERT(dn->dn_maxblkid >= blkid); dbuf_dirty_record_t *dr = kmem_zalloc(sizeof (*dr), KM_SLEEP); list_link_init(&dr->dr_dirty_node); list_link_init(&dr->dr_dbuf_node); dr->dr_dnode = dn; dr->dr_txg = tx->tx_txg; dr->dt.dll.dr_blkid = blkid; dr->dr_accounted = dn->dn_datablksz; /* * There should not be any dbuf for the block that we're dirtying. * Otherwise the buffer contents could be inconsistent between the * dbuf and the lightweight dirty record. */ ASSERT3P(NULL, ==, dbuf_find(dn->dn_objset, dn->dn_object, 0, blkid, NULL)); mutex_enter(&dn->dn_mtx); int txgoff = tx->tx_txg & TXG_MASK; if (dn->dn_free_ranges[txgoff] != NULL) { range_tree_clear(dn->dn_free_ranges[txgoff], blkid, 1); } if (dn->dn_nlevels == 1) { ASSERT3U(blkid, <, dn->dn_nblkptr); list_insert_tail(&dn->dn_dirty_records[txgoff], dr); mutex_exit(&dn->dn_mtx); rw_exit(&dn->dn_struct_rwlock); dnode_setdirty(dn, tx); } else { mutex_exit(&dn->dn_mtx); int epbs = dn->dn_indblkshift - SPA_BLKPTRSHIFT; dmu_buf_impl_t *parent_db = dbuf_hold_level(dn, 1, blkid >> epbs, FTAG); rw_exit(&dn->dn_struct_rwlock); if (parent_db == NULL) { kmem_free(dr, sizeof (*dr)); return (NULL); } int err = dbuf_read(parent_db, NULL, (DB_RF_NOPREFETCH | DB_RF_CANFAIL)); if (err != 0) { dbuf_rele(parent_db, FTAG); kmem_free(dr, sizeof (*dr)); return (NULL); } dbuf_dirty_record_t *parent_dr = dbuf_dirty(parent_db, tx); dbuf_rele(parent_db, FTAG); mutex_enter(&parent_dr->dt.di.dr_mtx); ASSERT3U(parent_dr->dr_txg, ==, tx->tx_txg); list_insert_tail(&parent_dr->dt.di.dr_children, dr); mutex_exit(&parent_dr->dt.di.dr_mtx); dr->dr_parent = parent_dr; } dmu_objset_willuse_space(dn->dn_objset, dr->dr_accounted, tx); return (dr); } dbuf_dirty_record_t * dbuf_dirty(dmu_buf_impl_t *db, dmu_tx_t *tx) { dnode_t *dn; objset_t *os; dbuf_dirty_record_t *dr, *dr_next, *dr_head; int txgoff = tx->tx_txg & TXG_MASK; boolean_t drop_struct_rwlock = B_FALSE; ASSERT(tx->tx_txg != 0); ASSERT(!zfs_refcount_is_zero(&db->db_holds)); DMU_TX_DIRTY_BUF(tx, db); DB_DNODE_ENTER(db); dn = DB_DNODE(db); /* * Shouldn't dirty a regular buffer in syncing context. Private * objects may be dirtied in syncing context, but only if they * were already pre-dirtied in open context. */ #ifdef ZFS_DEBUG if (dn->dn_objset->os_dsl_dataset != NULL) { rrw_enter(&dn->dn_objset->os_dsl_dataset->ds_bp_rwlock, RW_READER, FTAG); } ASSERT(!dmu_tx_is_syncing(tx) || BP_IS_HOLE(dn->dn_objset->os_rootbp) || DMU_OBJECT_IS_SPECIAL(dn->dn_object) || dn->dn_objset->os_dsl_dataset == NULL); if (dn->dn_objset->os_dsl_dataset != NULL) rrw_exit(&dn->dn_objset->os_dsl_dataset->ds_bp_rwlock, FTAG); #endif /* * We make this assert for private objects as well, but after we * check if we're already dirty. They are allowed to re-dirty * in syncing context. */ ASSERT(dn->dn_object == DMU_META_DNODE_OBJECT || dn->dn_dirtyctx == DN_UNDIRTIED || dn->dn_dirtyctx == (dmu_tx_is_syncing(tx) ? DN_DIRTY_SYNC : DN_DIRTY_OPEN)); mutex_enter(&db->db_mtx); /* * XXX make this true for indirects too? The problem is that * transactions created with dmu_tx_create_assigned() from * syncing context don't bother holding ahead. */ ASSERT(db->db_level != 0 || db->db_state == DB_CACHED || db->db_state == DB_FILL || db->db_state == DB_NOFILL); mutex_enter(&dn->dn_mtx); dnode_set_dirtyctx(dn, tx, db); if (tx->tx_txg > dn->dn_dirty_txg) dn->dn_dirty_txg = tx->tx_txg; mutex_exit(&dn->dn_mtx); if (db->db_blkid == DMU_SPILL_BLKID) dn->dn_have_spill = B_TRUE; /* * If this buffer is already dirty, we're done. */ dr_head = list_head(&db->db_dirty_records); ASSERT(dr_head == NULL || dr_head->dr_txg <= tx->tx_txg || db->db.db_object == DMU_META_DNODE_OBJECT); dr_next = dbuf_find_dirty_lte(db, tx->tx_txg); if (dr_next && dr_next->dr_txg == tx->tx_txg) { DB_DNODE_EXIT(db); dbuf_redirty(dr_next); mutex_exit(&db->db_mtx); return (dr_next); } /* * Only valid if not already dirty. */ ASSERT(dn->dn_object == 0 || dn->dn_dirtyctx == DN_UNDIRTIED || dn->dn_dirtyctx == (dmu_tx_is_syncing(tx) ? DN_DIRTY_SYNC : DN_DIRTY_OPEN)); ASSERT3U(dn->dn_nlevels, >, db->db_level); /* * We should only be dirtying in syncing context if it's the * mos or we're initializing the os or it's a special object. * However, we are allowed to dirty in syncing context provided * we already dirtied it in open context. Hence we must make * this assertion only if we're not already dirty. */ os = dn->dn_objset; VERIFY3U(tx->tx_txg, <=, spa_final_dirty_txg(os->os_spa)); #ifdef ZFS_DEBUG if (dn->dn_objset->os_dsl_dataset != NULL) rrw_enter(&os->os_dsl_dataset->ds_bp_rwlock, RW_READER, FTAG); ASSERT(!dmu_tx_is_syncing(tx) || DMU_OBJECT_IS_SPECIAL(dn->dn_object) || os->os_dsl_dataset == NULL || BP_IS_HOLE(os->os_rootbp)); if (dn->dn_objset->os_dsl_dataset != NULL) rrw_exit(&os->os_dsl_dataset->ds_bp_rwlock, FTAG); #endif ASSERT(db->db.db_size != 0); dprintf_dbuf(db, "size=%llx\n", (u_longlong_t)db->db.db_size); if (db->db_blkid != DMU_BONUS_BLKID && db->db_state != DB_NOFILL) { dmu_objset_willuse_space(os, db->db.db_size, tx); } /* * If this buffer is dirty in an old transaction group we need * to make a copy of it so that the changes we make in this * transaction group won't leak out when we sync the older txg. */ dr = kmem_zalloc(sizeof (dbuf_dirty_record_t), KM_SLEEP); list_link_init(&dr->dr_dirty_node); list_link_init(&dr->dr_dbuf_node); dr->dr_dnode = dn; if (db->db_level == 0) { void *data_old = db->db_buf; if (db->db_state != DB_NOFILL) { if (db->db_blkid == DMU_BONUS_BLKID) { dbuf_fix_old_data(db, tx->tx_txg); data_old = db->db.db_data; } else if (db->db.db_object != DMU_META_DNODE_OBJECT) { /* * Release the data buffer from the cache so * that we can modify it without impacting * possible other users of this cached data * block. Note that indirect blocks and * private objects are not released until the * syncing state (since they are only modified * then). */ arc_release(db->db_buf, db); dbuf_fix_old_data(db, tx->tx_txg); data_old = db->db_buf; } ASSERT(data_old != NULL); } dr->dt.dl.dr_data = data_old; } else { mutex_init(&dr->dt.di.dr_mtx, NULL, MUTEX_NOLOCKDEP, NULL); list_create(&dr->dt.di.dr_children, sizeof (dbuf_dirty_record_t), offsetof(dbuf_dirty_record_t, dr_dirty_node)); } if (db->db_blkid != DMU_BONUS_BLKID && db->db_state != DB_NOFILL) { dr->dr_accounted = db->db.db_size; } dr->dr_dbuf = db; dr->dr_txg = tx->tx_txg; list_insert_before(&db->db_dirty_records, dr_next, dr); /* * We could have been freed_in_flight between the dbuf_noread * and dbuf_dirty. We win, as though the dbuf_noread() had * happened after the free. */ if (db->db_level == 0 && db->db_blkid != DMU_BONUS_BLKID && db->db_blkid != DMU_SPILL_BLKID) { mutex_enter(&dn->dn_mtx); if (dn->dn_free_ranges[txgoff] != NULL) { range_tree_clear(dn->dn_free_ranges[txgoff], db->db_blkid, 1); } mutex_exit(&dn->dn_mtx); db->db_freed_in_flight = FALSE; } /* * This buffer is now part of this txg */ dbuf_add_ref(db, (void *)(uintptr_t)tx->tx_txg); db->db_dirtycnt += 1; ASSERT3U(db->db_dirtycnt, <=, 3); mutex_exit(&db->db_mtx); if (db->db_blkid == DMU_BONUS_BLKID || db->db_blkid == DMU_SPILL_BLKID) { mutex_enter(&dn->dn_mtx); ASSERT(!list_link_active(&dr->dr_dirty_node)); list_insert_tail(&dn->dn_dirty_records[txgoff], dr); mutex_exit(&dn->dn_mtx); dnode_setdirty(dn, tx); DB_DNODE_EXIT(db); return (dr); } if (!RW_WRITE_HELD(&dn->dn_struct_rwlock)) { rw_enter(&dn->dn_struct_rwlock, RW_READER); drop_struct_rwlock = B_TRUE; } /* * If we are overwriting a dedup BP, then unless it is snapshotted, * when we get to syncing context we will need to decrement its * refcount in the DDT. Prefetch the relevant DDT block so that * syncing context won't have to wait for the i/o. */ if (db->db_blkptr != NULL) { db_lock_type_t dblt = dmu_buf_lock_parent(db, RW_READER, FTAG); ddt_prefetch(os->os_spa, db->db_blkptr); dmu_buf_unlock_parent(db, dblt, FTAG); } /* * We need to hold the dn_struct_rwlock to make this assertion, * because it protects dn_phys / dn_next_nlevels from changing. */ ASSERT((dn->dn_phys->dn_nlevels == 0 && db->db_level == 0) || dn->dn_phys->dn_nlevels > db->db_level || dn->dn_next_nlevels[txgoff] > db->db_level || dn->dn_next_nlevels[(tx->tx_txg-1) & TXG_MASK] > db->db_level || dn->dn_next_nlevels[(tx->tx_txg-2) & TXG_MASK] > db->db_level); if (db->db_level == 0) { ASSERT(!db->db_objset->os_raw_receive || dn->dn_maxblkid >= db->db_blkid); dnode_new_blkid(dn, db->db_blkid, tx, drop_struct_rwlock, B_FALSE); ASSERT(dn->dn_maxblkid >= db->db_blkid); } if (db->db_level+1 < dn->dn_nlevels) { dmu_buf_impl_t *parent = db->db_parent; dbuf_dirty_record_t *di; int parent_held = FALSE; if (db->db_parent == NULL || db->db_parent == dn->dn_dbuf) { int epbs = dn->dn_indblkshift - SPA_BLKPTRSHIFT; parent = dbuf_hold_level(dn, db->db_level + 1, db->db_blkid >> epbs, FTAG); ASSERT(parent != NULL); parent_held = TRUE; } if (drop_struct_rwlock) rw_exit(&dn->dn_struct_rwlock); ASSERT3U(db->db_level + 1, ==, parent->db_level); di = dbuf_dirty(parent, tx); if (parent_held) dbuf_rele(parent, FTAG); mutex_enter(&db->db_mtx); /* * Since we've dropped the mutex, it's possible that * dbuf_undirty() might have changed this out from under us. */ if (list_head(&db->db_dirty_records) == dr || dn->dn_object == DMU_META_DNODE_OBJECT) { mutex_enter(&di->dt.di.dr_mtx); ASSERT3U(di->dr_txg, ==, tx->tx_txg); ASSERT(!list_link_active(&dr->dr_dirty_node)); list_insert_tail(&di->dt.di.dr_children, dr); mutex_exit(&di->dt.di.dr_mtx); dr->dr_parent = di; } mutex_exit(&db->db_mtx); } else { ASSERT(db->db_level + 1 == dn->dn_nlevels); ASSERT(db->db_blkid < dn->dn_nblkptr); ASSERT(db->db_parent == NULL || db->db_parent == dn->dn_dbuf); mutex_enter(&dn->dn_mtx); ASSERT(!list_link_active(&dr->dr_dirty_node)); list_insert_tail(&dn->dn_dirty_records[txgoff], dr); mutex_exit(&dn->dn_mtx); if (drop_struct_rwlock) rw_exit(&dn->dn_struct_rwlock); } dnode_setdirty(dn, tx); DB_DNODE_EXIT(db); return (dr); } static void dbuf_undirty_bonus(dbuf_dirty_record_t *dr) { dmu_buf_impl_t *db = dr->dr_dbuf; if (dr->dt.dl.dr_data != db->db.db_data) { struct dnode *dn = dr->dr_dnode; int max_bonuslen = DN_SLOTS_TO_BONUSLEN(dn->dn_num_slots); kmem_free(dr->dt.dl.dr_data, max_bonuslen); arc_space_return(max_bonuslen, ARC_SPACE_BONUS); } db->db_data_pending = NULL; ASSERT(list_next(&db->db_dirty_records, dr) == NULL); list_remove(&db->db_dirty_records, dr); if (dr->dr_dbuf->db_level != 0) { mutex_destroy(&dr->dt.di.dr_mtx); list_destroy(&dr->dt.di.dr_children); } kmem_free(dr, sizeof (dbuf_dirty_record_t)); ASSERT3U(db->db_dirtycnt, >, 0); db->db_dirtycnt -= 1; } /* * Undirty a buffer in the transaction group referenced by the given * transaction. Return whether this evicted the dbuf. */ boolean_t dbuf_undirty(dmu_buf_impl_t *db, dmu_tx_t *tx) { uint64_t txg = tx->tx_txg; boolean_t brtwrite; ASSERT(txg != 0); /* * Due to our use of dn_nlevels below, this can only be called * in open context, unless we are operating on the MOS. * From syncing context, dn_nlevels may be different from the * dn_nlevels used when dbuf was dirtied. */ ASSERT(db->db_objset == dmu_objset_pool(db->db_objset)->dp_meta_objset || txg != spa_syncing_txg(dmu_objset_spa(db->db_objset))); ASSERT(db->db_blkid != DMU_BONUS_BLKID); ASSERT0(db->db_level); ASSERT(MUTEX_HELD(&db->db_mtx)); /* * If this buffer is not dirty, we're done. */ dbuf_dirty_record_t *dr = dbuf_find_dirty_eq(db, txg); if (dr == NULL) return (B_FALSE); ASSERT(dr->dr_dbuf == db); brtwrite = dr->dt.dl.dr_brtwrite; if (brtwrite) { /* * We are freeing a block that we cloned in the same * transaction group. */ brt_pending_remove(dmu_objset_spa(db->db_objset), &dr->dt.dl.dr_overridden_by, tx); } dnode_t *dn = dr->dr_dnode; dprintf_dbuf(db, "size=%llx\n", (u_longlong_t)db->db.db_size); ASSERT(db->db.db_size != 0); dsl_pool_undirty_space(dmu_objset_pool(dn->dn_objset), dr->dr_accounted, txg); list_remove(&db->db_dirty_records, dr); /* * Note that there are three places in dbuf_dirty() * where this dirty record may be put on a list. * Make sure to do a list_remove corresponding to * every one of those list_insert calls. */ if (dr->dr_parent) { mutex_enter(&dr->dr_parent->dt.di.dr_mtx); list_remove(&dr->dr_parent->dt.di.dr_children, dr); mutex_exit(&dr->dr_parent->dt.di.dr_mtx); } else if (db->db_blkid == DMU_SPILL_BLKID || db->db_level + 1 == dn->dn_nlevels) { ASSERT(db->db_blkptr == NULL || db->db_parent == dn->dn_dbuf); mutex_enter(&dn->dn_mtx); list_remove(&dn->dn_dirty_records[txg & TXG_MASK], dr); mutex_exit(&dn->dn_mtx); } if (db->db_state != DB_NOFILL && !brtwrite) { dbuf_unoverride(dr); ASSERT(db->db_buf != NULL); ASSERT(dr->dt.dl.dr_data != NULL); if (dr->dt.dl.dr_data != db->db_buf) arc_buf_destroy(dr->dt.dl.dr_data, db); } kmem_free(dr, sizeof (dbuf_dirty_record_t)); ASSERT(db->db_dirtycnt > 0); db->db_dirtycnt -= 1; if (zfs_refcount_remove(&db->db_holds, (void *)(uintptr_t)txg) == 0) { ASSERT(db->db_state == DB_NOFILL || brtwrite || arc_released(db->db_buf)); dbuf_destroy(db); return (B_TRUE); } return (B_FALSE); } static void dmu_buf_will_dirty_impl(dmu_buf_t *db_fake, int flags, dmu_tx_t *tx) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake; boolean_t undirty = B_FALSE; ASSERT(tx->tx_txg != 0); ASSERT(!zfs_refcount_is_zero(&db->db_holds)); /* * Quick check for dirtiness. For already dirty blocks, this * reduces runtime of this function by >90%, and overall performance * by 50% for some workloads (e.g. file deletion with indirect blocks * cached). */ mutex_enter(&db->db_mtx); if (db->db_state == DB_CACHED || db->db_state == DB_NOFILL) { dbuf_dirty_record_t *dr = dbuf_find_dirty_eq(db, tx->tx_txg); /* * It's possible that it is already dirty but not cached, * because there are some calls to dbuf_dirty() that don't * go through dmu_buf_will_dirty(). */ if (dr != NULL) { if (dr->dt.dl.dr_brtwrite) { /* * Block cloning: If we are dirtying a cloned * block, we cannot simply redirty it, because * this dr has no data associated with it. * We will go through a full undirtying below, * before dirtying it again. */ undirty = B_TRUE; } else { /* This dbuf is already dirty and cached. */ dbuf_redirty(dr); mutex_exit(&db->db_mtx); return; } } } mutex_exit(&db->db_mtx); DB_DNODE_ENTER(db); if (RW_WRITE_HELD(&DB_DNODE(db)->dn_struct_rwlock)) flags |= DB_RF_HAVESTRUCT; DB_DNODE_EXIT(db); /* * Block cloning: Do the dbuf_read() before undirtying the dbuf, as we * want to make sure dbuf_read() will read the pending cloned block and * not the uderlying block that is being replaced. dbuf_undirty() will * do dbuf_unoverride(), so we will end up with cloned block content, * without overridden BP. */ (void) dbuf_read(db, NULL, flags); if (undirty) { mutex_enter(&db->db_mtx); VERIFY(!dbuf_undirty(db, tx)); mutex_exit(&db->db_mtx); } (void) dbuf_dirty(db, tx); } void dmu_buf_will_dirty(dmu_buf_t *db_fake, dmu_tx_t *tx) { dmu_buf_will_dirty_impl(db_fake, DB_RF_MUST_SUCCEED | DB_RF_NOPREFETCH, tx); } boolean_t dmu_buf_is_dirty(dmu_buf_t *db_fake, dmu_tx_t *tx) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake; dbuf_dirty_record_t *dr; mutex_enter(&db->db_mtx); dr = dbuf_find_dirty_eq(db, tx->tx_txg); mutex_exit(&db->db_mtx); return (dr != NULL); } void dmu_buf_will_clone(dmu_buf_t *db_fake, dmu_tx_t *tx) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake; /* * Block cloning: We are going to clone into this block, so undirty * modifications done to this block so far in this txg. This includes * writes and clones into this block. */ mutex_enter(&db->db_mtx); VERIFY(!dbuf_undirty(db, tx)); ASSERT(list_head(&db->db_dirty_records) == NULL); if (db->db_buf != NULL) { arc_buf_destroy(db->db_buf, db); db->db_buf = NULL; } mutex_exit(&db->db_mtx); dmu_buf_will_not_fill(db_fake, tx); } void dmu_buf_will_not_fill(dmu_buf_t *db_fake, dmu_tx_t *tx) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake; + mutex_enter(&db->db_mtx); db->db_state = DB_NOFILL; DTRACE_SET_STATE(db, "allocating NOFILL buffer"); + mutex_exit(&db->db_mtx); dbuf_noread(db); (void) dbuf_dirty(db, tx); } void dmu_buf_will_fill(dmu_buf_t *db_fake, dmu_tx_t *tx) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake; ASSERT(db->db_blkid != DMU_BONUS_BLKID); ASSERT(tx->tx_txg != 0); ASSERT(db->db_level == 0); ASSERT(!zfs_refcount_is_zero(&db->db_holds)); ASSERT(db->db.db_object != DMU_META_DNODE_OBJECT || dmu_tx_private_ok(tx)); + mutex_enter(&db->db_mtx); if (db->db_state == DB_NOFILL) { /* * Block cloning: We will be completely overwriting a block * cloned in this transaction group, so let's undirty the * pending clone and mark the block as uncached. This will be * as if the clone was never done. */ - mutex_enter(&db->db_mtx); VERIFY(!dbuf_undirty(db, tx)); - mutex_exit(&db->db_mtx); db->db_state = DB_UNCACHED; } + mutex_exit(&db->db_mtx); dbuf_noread(db); (void) dbuf_dirty(db, tx); } /* * This function is effectively the same as dmu_buf_will_dirty(), but * indicates the caller expects raw encrypted data in the db, and provides * the crypt params (byteorder, salt, iv, mac) which should be stored in the * blkptr_t when this dbuf is written. This is only used for blocks of * dnodes, during raw receive. */ void dmu_buf_set_crypt_params(dmu_buf_t *db_fake, boolean_t byteorder, const uint8_t *salt, const uint8_t *iv, const uint8_t *mac, dmu_tx_t *tx) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake; dbuf_dirty_record_t *dr; /* * dr_has_raw_params is only processed for blocks of dnodes * (see dbuf_sync_dnode_leaf_crypt()). */ ASSERT3U(db->db.db_object, ==, DMU_META_DNODE_OBJECT); ASSERT3U(db->db_level, ==, 0); ASSERT(db->db_objset->os_raw_receive); dmu_buf_will_dirty_impl(db_fake, DB_RF_MUST_SUCCEED | DB_RF_NOPREFETCH | DB_RF_NO_DECRYPT, tx); dr = dbuf_find_dirty_eq(db, tx->tx_txg); ASSERT3P(dr, !=, NULL); dr->dt.dl.dr_has_raw_params = B_TRUE; dr->dt.dl.dr_byteorder = byteorder; memcpy(dr->dt.dl.dr_salt, salt, ZIO_DATA_SALT_LEN); memcpy(dr->dt.dl.dr_iv, iv, ZIO_DATA_IV_LEN); memcpy(dr->dt.dl.dr_mac, mac, ZIO_DATA_MAC_LEN); } static void dbuf_override_impl(dmu_buf_impl_t *db, const blkptr_t *bp, dmu_tx_t *tx) { struct dirty_leaf *dl; dbuf_dirty_record_t *dr; dr = list_head(&db->db_dirty_records); ASSERT3P(dr, !=, NULL); ASSERT3U(dr->dr_txg, ==, tx->tx_txg); dl = &dr->dt.dl; dl->dr_overridden_by = *bp; dl->dr_override_state = DR_OVERRIDDEN; dl->dr_overridden_by.blk_birth = dr->dr_txg; } void dmu_buf_fill_done(dmu_buf_t *dbuf, dmu_tx_t *tx) { (void) tx; dmu_buf_impl_t *db = (dmu_buf_impl_t *)dbuf; dbuf_states_t old_state; mutex_enter(&db->db_mtx); DBUF_VERIFY(db); old_state = db->db_state; db->db_state = DB_CACHED; if (old_state == DB_FILL) { if (db->db_level == 0 && db->db_freed_in_flight) { ASSERT(db->db_blkid != DMU_BONUS_BLKID); /* we were freed while filling */ /* XXX dbuf_undirty? */ memset(db->db.db_data, 0, db->db.db_size); db->db_freed_in_flight = FALSE; DTRACE_SET_STATE(db, "fill done handling freed in flight"); } else { DTRACE_SET_STATE(db, "fill done"); } cv_broadcast(&db->db_changed); } mutex_exit(&db->db_mtx); } void dmu_buf_write_embedded(dmu_buf_t *dbuf, void *data, bp_embedded_type_t etype, enum zio_compress comp, int uncompressed_size, int compressed_size, int byteorder, dmu_tx_t *tx) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)dbuf; struct dirty_leaf *dl; dmu_object_type_t type; dbuf_dirty_record_t *dr; if (etype == BP_EMBEDDED_TYPE_DATA) { ASSERT(spa_feature_is_active(dmu_objset_spa(db->db_objset), SPA_FEATURE_EMBEDDED_DATA)); } DB_DNODE_ENTER(db); type = DB_DNODE(db)->dn_type; DB_DNODE_EXIT(db); ASSERT0(db->db_level); ASSERT(db->db_blkid != DMU_BONUS_BLKID); dmu_buf_will_not_fill(dbuf, tx); dr = list_head(&db->db_dirty_records); ASSERT3P(dr, !=, NULL); ASSERT3U(dr->dr_txg, ==, tx->tx_txg); dl = &dr->dt.dl; encode_embedded_bp_compressed(&dl->dr_overridden_by, data, comp, uncompressed_size, compressed_size); BPE_SET_ETYPE(&dl->dr_overridden_by, etype); BP_SET_TYPE(&dl->dr_overridden_by, type); BP_SET_LEVEL(&dl->dr_overridden_by, 0); BP_SET_BYTEORDER(&dl->dr_overridden_by, byteorder); dl->dr_override_state = DR_OVERRIDDEN; dl->dr_overridden_by.blk_birth = dr->dr_txg; } void dmu_buf_redact(dmu_buf_t *dbuf, dmu_tx_t *tx) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)dbuf; dmu_object_type_t type; ASSERT(dsl_dataset_feature_is_active(db->db_objset->os_dsl_dataset, SPA_FEATURE_REDACTED_DATASETS)); DB_DNODE_ENTER(db); type = DB_DNODE(db)->dn_type; DB_DNODE_EXIT(db); ASSERT0(db->db_level); dmu_buf_will_not_fill(dbuf, tx); blkptr_t bp = { { { {0} } } }; BP_SET_TYPE(&bp, type); BP_SET_LEVEL(&bp, 0); BP_SET_BIRTH(&bp, tx->tx_txg, 0); BP_SET_REDACTED(&bp); BPE_SET_LSIZE(&bp, dbuf->db_size); dbuf_override_impl(db, &bp, tx); } /* * Directly assign a provided arc buf to a given dbuf if it's not referenced * by anybody except our caller. Otherwise copy arcbuf's contents to dbuf. */ void dbuf_assign_arcbuf(dmu_buf_impl_t *db, arc_buf_t *buf, dmu_tx_t *tx) { ASSERT(!zfs_refcount_is_zero(&db->db_holds)); ASSERT(db->db_blkid != DMU_BONUS_BLKID); ASSERT(db->db_level == 0); ASSERT3U(dbuf_is_metadata(db), ==, arc_is_metadata(buf)); ASSERT(buf != NULL); ASSERT3U(arc_buf_lsize(buf), ==, db->db.db_size); ASSERT(tx->tx_txg != 0); arc_return_buf(buf, db); ASSERT(arc_released(buf)); mutex_enter(&db->db_mtx); while (db->db_state == DB_READ || db->db_state == DB_FILL) cv_wait(&db->db_changed, &db->db_mtx); ASSERT(db->db_state == DB_CACHED || db->db_state == DB_UNCACHED); if (db->db_state == DB_CACHED && zfs_refcount_count(&db->db_holds) - 1 > db->db_dirtycnt) { /* * In practice, we will never have a case where we have an * encrypted arc buffer while additional holds exist on the * dbuf. We don't handle this here so we simply assert that * fact instead. */ ASSERT(!arc_is_encrypted(buf)); mutex_exit(&db->db_mtx); (void) dbuf_dirty(db, tx); memcpy(db->db.db_data, buf->b_data, db->db.db_size); arc_buf_destroy(buf, db); return; } if (db->db_state == DB_CACHED) { dbuf_dirty_record_t *dr = list_head(&db->db_dirty_records); ASSERT(db->db_buf != NULL); if (dr != NULL && dr->dr_txg == tx->tx_txg) { ASSERT(dr->dt.dl.dr_data == db->db_buf); if (!arc_released(db->db_buf)) { ASSERT(dr->dt.dl.dr_override_state == DR_OVERRIDDEN); arc_release(db->db_buf, db); } dr->dt.dl.dr_data = buf; arc_buf_destroy(db->db_buf, db); } else if (dr == NULL || dr->dt.dl.dr_data != db->db_buf) { arc_release(db->db_buf, db); arc_buf_destroy(db->db_buf, db); } db->db_buf = NULL; } ASSERT(db->db_buf == NULL); dbuf_set_data(db, buf); db->db_state = DB_FILL; DTRACE_SET_STATE(db, "filling assigned arcbuf"); mutex_exit(&db->db_mtx); (void) dbuf_dirty(db, tx); dmu_buf_fill_done(&db->db, tx); } void dbuf_destroy(dmu_buf_impl_t *db) { dnode_t *dn; dmu_buf_impl_t *parent = db->db_parent; dmu_buf_impl_t *dndb; ASSERT(MUTEX_HELD(&db->db_mtx)); ASSERT(zfs_refcount_is_zero(&db->db_holds)); if (db->db_buf != NULL) { arc_buf_destroy(db->db_buf, db); db->db_buf = NULL; } if (db->db_blkid == DMU_BONUS_BLKID) { int slots = DB_DNODE(db)->dn_num_slots; int bonuslen = DN_SLOTS_TO_BONUSLEN(slots); if (db->db.db_data != NULL) { kmem_free(db->db.db_data, bonuslen); arc_space_return(bonuslen, ARC_SPACE_BONUS); db->db_state = DB_UNCACHED; DTRACE_SET_STATE(db, "buffer cleared"); } } dbuf_clear_data(db); if (multilist_link_active(&db->db_cache_link)) { ASSERT(db->db_caching_status == DB_DBUF_CACHE || db->db_caching_status == DB_DBUF_METADATA_CACHE); multilist_remove(&dbuf_caches[db->db_caching_status].cache, db); (void) zfs_refcount_remove_many( &dbuf_caches[db->db_caching_status].size, db->db.db_size, db); if (db->db_caching_status == DB_DBUF_METADATA_CACHE) { DBUF_STAT_BUMPDOWN(metadata_cache_count); } else { DBUF_STAT_BUMPDOWN(cache_levels[db->db_level]); DBUF_STAT_BUMPDOWN(cache_count); DBUF_STAT_DECR(cache_levels_bytes[db->db_level], db->db.db_size); } db->db_caching_status = DB_NO_CACHE; } ASSERT(db->db_state == DB_UNCACHED || db->db_state == DB_NOFILL); ASSERT(db->db_data_pending == NULL); ASSERT(list_is_empty(&db->db_dirty_records)); db->db_state = DB_EVICTING; DTRACE_SET_STATE(db, "buffer eviction started"); db->db_blkptr = NULL; /* * Now that db_state is DB_EVICTING, nobody else can find this via * the hash table. We can now drop db_mtx, which allows us to * acquire the dn_dbufs_mtx. */ mutex_exit(&db->db_mtx); DB_DNODE_ENTER(db); dn = DB_DNODE(db); dndb = dn->dn_dbuf; if (db->db_blkid != DMU_BONUS_BLKID) { boolean_t needlock = !MUTEX_HELD(&dn->dn_dbufs_mtx); if (needlock) mutex_enter_nested(&dn->dn_dbufs_mtx, NESTED_SINGLE); avl_remove(&dn->dn_dbufs, db); membar_producer(); DB_DNODE_EXIT(db); if (needlock) mutex_exit(&dn->dn_dbufs_mtx); /* * Decrementing the dbuf count means that the hold corresponding * to the removed dbuf is no longer discounted in dnode_move(), * so the dnode cannot be moved until after we release the hold. * The membar_producer() ensures visibility of the decremented * value in dnode_move(), since DB_DNODE_EXIT doesn't actually * release any lock. */ mutex_enter(&dn->dn_mtx); dnode_rele_and_unlock(dn, db, B_TRUE); db->db_dnode_handle = NULL; dbuf_hash_remove(db); } else { DB_DNODE_EXIT(db); } ASSERT(zfs_refcount_is_zero(&db->db_holds)); db->db_parent = NULL; ASSERT(db->db_buf == NULL); ASSERT(db->db.db_data == NULL); ASSERT(db->db_hash_next == NULL); ASSERT(db->db_blkptr == NULL); ASSERT(db->db_data_pending == NULL); ASSERT3U(db->db_caching_status, ==, DB_NO_CACHE); ASSERT(!multilist_link_active(&db->db_cache_link)); /* * If this dbuf is referenced from an indirect dbuf, * decrement the ref count on the indirect dbuf. */ if (parent && parent != dndb) { mutex_enter(&parent->db_mtx); dbuf_rele_and_unlock(parent, db, B_TRUE); } kmem_cache_free(dbuf_kmem_cache, db); arc_space_return(sizeof (dmu_buf_impl_t), ARC_SPACE_DBUF); } /* * Note: While bpp will always be updated if the function returns success, * parentp will not be updated if the dnode does not have dn_dbuf filled in; * this happens when the dnode is the meta-dnode, or {user|group|project}used * object. */ __attribute__((always_inline)) static inline int dbuf_findbp(dnode_t *dn, int level, uint64_t blkid, int fail_sparse, dmu_buf_impl_t **parentp, blkptr_t **bpp) { *parentp = NULL; *bpp = NULL; ASSERT(blkid != DMU_BONUS_BLKID); if (blkid == DMU_SPILL_BLKID) { mutex_enter(&dn->dn_mtx); if (dn->dn_have_spill && (dn->dn_phys->dn_flags & DNODE_FLAG_SPILL_BLKPTR)) *bpp = DN_SPILL_BLKPTR(dn->dn_phys); else *bpp = NULL; dbuf_add_ref(dn->dn_dbuf, NULL); *parentp = dn->dn_dbuf; mutex_exit(&dn->dn_mtx); return (0); } int nlevels = (dn->dn_phys->dn_nlevels == 0) ? 1 : dn->dn_phys->dn_nlevels; int epbs = dn->dn_indblkshift - SPA_BLKPTRSHIFT; ASSERT3U(level * epbs, <, 64); ASSERT(RW_LOCK_HELD(&dn->dn_struct_rwlock)); /* * This assertion shouldn't trip as long as the max indirect block size * is less than 1M. The reason for this is that up to that point, * the number of levels required to address an entire object with blocks * of size SPA_MINBLOCKSIZE satisfies nlevels * epbs + 1 <= 64. In * other words, if N * epbs + 1 > 64, then if (N-1) * epbs + 1 > 55 * (i.e. we can address the entire object), objects will all use at most * N-1 levels and the assertion won't overflow. However, once epbs is * 13, 4 * 13 + 1 = 53, but 5 * 13 + 1 = 66. Then, 4 levels will not be * enough to address an entire object, so objects will have 5 levels, * but then this assertion will overflow. * * All this is to say that if we ever increase DN_MAX_INDBLKSHIFT, we * need to redo this logic to handle overflows. */ ASSERT(level >= nlevels || ((nlevels - level - 1) * epbs) + highbit64(dn->dn_phys->dn_nblkptr) <= 64); if (level >= nlevels || blkid >= ((uint64_t)dn->dn_phys->dn_nblkptr << ((nlevels - level - 1) * epbs)) || (fail_sparse && blkid > (dn->dn_phys->dn_maxblkid >> (level * epbs)))) { /* the buffer has no parent yet */ return (SET_ERROR(ENOENT)); } else if (level < nlevels-1) { /* this block is referenced from an indirect block */ int err; err = dbuf_hold_impl(dn, level + 1, blkid >> epbs, fail_sparse, FALSE, NULL, parentp); if (err) return (err); err = dbuf_read(*parentp, NULL, (DB_RF_HAVESTRUCT | DB_RF_NOPREFETCH | DB_RF_CANFAIL)); if (err) { dbuf_rele(*parentp, NULL); *parentp = NULL; return (err); } rw_enter(&(*parentp)->db_rwlock, RW_READER); *bpp = ((blkptr_t *)(*parentp)->db.db_data) + (blkid & ((1ULL << epbs) - 1)); if (blkid > (dn->dn_phys->dn_maxblkid >> (level * epbs))) ASSERT(BP_IS_HOLE(*bpp)); rw_exit(&(*parentp)->db_rwlock); return (0); } else { /* the block is referenced from the dnode */ ASSERT3U(level, ==, nlevels-1); ASSERT(dn->dn_phys->dn_nblkptr == 0 || blkid < dn->dn_phys->dn_nblkptr); if (dn->dn_dbuf) { dbuf_add_ref(dn->dn_dbuf, NULL); *parentp = dn->dn_dbuf; } *bpp = &dn->dn_phys->dn_blkptr[blkid]; return (0); } } static dmu_buf_impl_t * dbuf_create(dnode_t *dn, uint8_t level, uint64_t blkid, dmu_buf_impl_t *parent, blkptr_t *blkptr, uint64_t hash) { objset_t *os = dn->dn_objset; dmu_buf_impl_t *db, *odb; ASSERT(RW_LOCK_HELD(&dn->dn_struct_rwlock)); ASSERT(dn->dn_type != DMU_OT_NONE); db = kmem_cache_alloc(dbuf_kmem_cache, KM_SLEEP); list_create(&db->db_dirty_records, sizeof (dbuf_dirty_record_t), offsetof(dbuf_dirty_record_t, dr_dbuf_node)); db->db_objset = os; db->db.db_object = dn->dn_object; db->db_level = level; db->db_blkid = blkid; db->db_dirtycnt = 0; db->db_dnode_handle = dn->dn_handle; db->db_parent = parent; db->db_blkptr = blkptr; db->db_hash = hash; db->db_user = NULL; db->db_user_immediate_evict = FALSE; db->db_freed_in_flight = FALSE; db->db_pending_evict = FALSE; if (blkid == DMU_BONUS_BLKID) { ASSERT3P(parent, ==, dn->dn_dbuf); db->db.db_size = DN_SLOTS_TO_BONUSLEN(dn->dn_num_slots) - (dn->dn_nblkptr-1) * sizeof (blkptr_t); ASSERT3U(db->db.db_size, >=, dn->dn_bonuslen); db->db.db_offset = DMU_BONUS_BLKID; db->db_state = DB_UNCACHED; DTRACE_SET_STATE(db, "bonus buffer created"); db->db_caching_status = DB_NO_CACHE; /* the bonus dbuf is not placed in the hash table */ arc_space_consume(sizeof (dmu_buf_impl_t), ARC_SPACE_DBUF); return (db); } else if (blkid == DMU_SPILL_BLKID) { db->db.db_size = (blkptr != NULL) ? BP_GET_LSIZE(blkptr) : SPA_MINBLOCKSIZE; db->db.db_offset = 0; } else { int blocksize = db->db_level ? 1 << dn->dn_indblkshift : dn->dn_datablksz; db->db.db_size = blocksize; db->db.db_offset = db->db_blkid * blocksize; } /* * Hold the dn_dbufs_mtx while we get the new dbuf * in the hash table *and* added to the dbufs list. * This prevents a possible deadlock with someone * trying to look up this dbuf before it's added to the * dn_dbufs list. */ mutex_enter(&dn->dn_dbufs_mtx); db->db_state = DB_EVICTING; /* not worth logging this state change */ if ((odb = dbuf_hash_insert(db)) != NULL) { /* someone else inserted it first */ mutex_exit(&dn->dn_dbufs_mtx); kmem_cache_free(dbuf_kmem_cache, db); DBUF_STAT_BUMP(hash_insert_race); return (odb); } avl_add(&dn->dn_dbufs, db); db->db_state = DB_UNCACHED; DTRACE_SET_STATE(db, "regular buffer created"); db->db_caching_status = DB_NO_CACHE; mutex_exit(&dn->dn_dbufs_mtx); arc_space_consume(sizeof (dmu_buf_impl_t), ARC_SPACE_DBUF); if (parent && parent != dn->dn_dbuf) dbuf_add_ref(parent, db); ASSERT(dn->dn_object == DMU_META_DNODE_OBJECT || zfs_refcount_count(&dn->dn_holds) > 0); (void) zfs_refcount_add(&dn->dn_holds, db); dprintf_dbuf(db, "db=%p\n", db); return (db); } /* * This function returns a block pointer and information about the object, * given a dnode and a block. This is a publicly accessible version of * dbuf_findbp that only returns some information, rather than the * dbuf. Note that the dnode passed in must be held, and the dn_struct_rwlock * should be locked as (at least) a reader. */ int dbuf_dnode_findbp(dnode_t *dn, uint64_t level, uint64_t blkid, blkptr_t *bp, uint16_t *datablkszsec, uint8_t *indblkshift) { dmu_buf_impl_t *dbp = NULL; blkptr_t *bp2; int err = 0; ASSERT(RW_LOCK_HELD(&dn->dn_struct_rwlock)); err = dbuf_findbp(dn, level, blkid, B_FALSE, &dbp, &bp2); if (err == 0) { ASSERT3P(bp2, !=, NULL); *bp = *bp2; if (dbp != NULL) dbuf_rele(dbp, NULL); if (datablkszsec != NULL) *datablkszsec = dn->dn_phys->dn_datablkszsec; if (indblkshift != NULL) *indblkshift = dn->dn_phys->dn_indblkshift; } return (err); } typedef struct dbuf_prefetch_arg { spa_t *dpa_spa; /* The spa to issue the prefetch in. */ zbookmark_phys_t dpa_zb; /* The target block to prefetch. */ int dpa_epbs; /* Entries (blkptr_t's) Per Block Shift. */ int dpa_curlevel; /* The current level that we're reading */ dnode_t *dpa_dnode; /* The dnode associated with the prefetch */ zio_priority_t dpa_prio; /* The priority I/Os should be issued at. */ zio_t *dpa_zio; /* The parent zio_t for all prefetches. */ arc_flags_t dpa_aflags; /* Flags to pass to the final prefetch. */ dbuf_prefetch_fn dpa_cb; /* prefetch completion callback */ void *dpa_arg; /* prefetch completion arg */ } dbuf_prefetch_arg_t; static void dbuf_prefetch_fini(dbuf_prefetch_arg_t *dpa, boolean_t io_done) { if (dpa->dpa_cb != NULL) { dpa->dpa_cb(dpa->dpa_arg, dpa->dpa_zb.zb_level, dpa->dpa_zb.zb_blkid, io_done); } kmem_free(dpa, sizeof (*dpa)); } static void dbuf_issue_final_prefetch_done(zio_t *zio, const zbookmark_phys_t *zb, const blkptr_t *iobp, arc_buf_t *abuf, void *private) { (void) zio, (void) zb, (void) iobp; dbuf_prefetch_arg_t *dpa = private; if (abuf != NULL) arc_buf_destroy(abuf, private); dbuf_prefetch_fini(dpa, B_TRUE); } /* * Actually issue the prefetch read for the block given. */ static void dbuf_issue_final_prefetch(dbuf_prefetch_arg_t *dpa, blkptr_t *bp) { ASSERT(!BP_IS_REDACTED(bp) || dsl_dataset_feature_is_active( dpa->dpa_dnode->dn_objset->os_dsl_dataset, SPA_FEATURE_REDACTED_DATASETS)); if (BP_IS_HOLE(bp) || BP_IS_EMBEDDED(bp) || BP_IS_REDACTED(bp)) return (dbuf_prefetch_fini(dpa, B_FALSE)); int zio_flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE; arc_flags_t aflags = dpa->dpa_aflags | ARC_FLAG_NOWAIT | ARC_FLAG_PREFETCH | ARC_FLAG_NO_BUF; /* dnodes are always read as raw and then converted later */ if (BP_GET_TYPE(bp) == DMU_OT_DNODE && BP_IS_PROTECTED(bp) && dpa->dpa_curlevel == 0) zio_flags |= ZIO_FLAG_RAW; ASSERT3U(dpa->dpa_curlevel, ==, BP_GET_LEVEL(bp)); ASSERT3U(dpa->dpa_curlevel, ==, dpa->dpa_zb.zb_level); ASSERT(dpa->dpa_zio != NULL); (void) arc_read(dpa->dpa_zio, dpa->dpa_spa, bp, dbuf_issue_final_prefetch_done, dpa, dpa->dpa_prio, zio_flags, &aflags, &dpa->dpa_zb); } /* * Called when an indirect block above our prefetch target is read in. This * will either read in the next indirect block down the tree or issue the actual * prefetch if the next block down is our target. */ static void dbuf_prefetch_indirect_done(zio_t *zio, const zbookmark_phys_t *zb, const blkptr_t *iobp, arc_buf_t *abuf, void *private) { (void) zb, (void) iobp; dbuf_prefetch_arg_t *dpa = private; ASSERT3S(dpa->dpa_zb.zb_level, <, dpa->dpa_curlevel); ASSERT3S(dpa->dpa_curlevel, >, 0); if (abuf == NULL) { ASSERT(zio == NULL || zio->io_error != 0); dbuf_prefetch_fini(dpa, B_TRUE); return; } ASSERT(zio == NULL || zio->io_error == 0); /* * The dpa_dnode is only valid if we are called with a NULL * zio. This indicates that the arc_read() returned without * first calling zio_read() to issue a physical read. Once * a physical read is made the dpa_dnode must be invalidated * as the locks guarding it may have been dropped. If the * dpa_dnode is still valid, then we want to add it to the dbuf * cache. To do so, we must hold the dbuf associated with the block * we just prefetched, read its contents so that we associate it * with an arc_buf_t, and then release it. */ if (zio != NULL) { ASSERT3S(BP_GET_LEVEL(zio->io_bp), ==, dpa->dpa_curlevel); if (zio->io_flags & ZIO_FLAG_RAW_COMPRESS) { ASSERT3U(BP_GET_PSIZE(zio->io_bp), ==, zio->io_size); } else { ASSERT3U(BP_GET_LSIZE(zio->io_bp), ==, zio->io_size); } ASSERT3P(zio->io_spa, ==, dpa->dpa_spa); dpa->dpa_dnode = NULL; } else if (dpa->dpa_dnode != NULL) { uint64_t curblkid = dpa->dpa_zb.zb_blkid >> (dpa->dpa_epbs * (dpa->dpa_curlevel - dpa->dpa_zb.zb_level)); dmu_buf_impl_t *db = dbuf_hold_level(dpa->dpa_dnode, dpa->dpa_curlevel, curblkid, FTAG); if (db == NULL) { arc_buf_destroy(abuf, private); dbuf_prefetch_fini(dpa, B_TRUE); return; } (void) dbuf_read(db, NULL, DB_RF_MUST_SUCCEED | DB_RF_NOPREFETCH | DB_RF_HAVESTRUCT); dbuf_rele(db, FTAG); } dpa->dpa_curlevel--; uint64_t nextblkid = dpa->dpa_zb.zb_blkid >> (dpa->dpa_epbs * (dpa->dpa_curlevel - dpa->dpa_zb.zb_level)); blkptr_t *bp = ((blkptr_t *)abuf->b_data) + P2PHASE(nextblkid, 1ULL << dpa->dpa_epbs); ASSERT(!BP_IS_REDACTED(bp) || (dpa->dpa_dnode && dsl_dataset_feature_is_active( dpa->dpa_dnode->dn_objset->os_dsl_dataset, SPA_FEATURE_REDACTED_DATASETS))); if (BP_IS_HOLE(bp) || BP_IS_REDACTED(bp)) { arc_buf_destroy(abuf, private); dbuf_prefetch_fini(dpa, B_TRUE); return; } else if (dpa->dpa_curlevel == dpa->dpa_zb.zb_level) { ASSERT3U(nextblkid, ==, dpa->dpa_zb.zb_blkid); dbuf_issue_final_prefetch(dpa, bp); } else { arc_flags_t iter_aflags = ARC_FLAG_NOWAIT; zbookmark_phys_t zb; /* flag if L2ARC eligible, l2arc_noprefetch then decides */ if (dpa->dpa_aflags & ARC_FLAG_L2CACHE) iter_aflags |= ARC_FLAG_L2CACHE; ASSERT3U(dpa->dpa_curlevel, ==, BP_GET_LEVEL(bp)); SET_BOOKMARK(&zb, dpa->dpa_zb.zb_objset, dpa->dpa_zb.zb_object, dpa->dpa_curlevel, nextblkid); (void) arc_read(dpa->dpa_zio, dpa->dpa_spa, bp, dbuf_prefetch_indirect_done, dpa, ZIO_PRIORITY_SYNC_READ, ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE, &iter_aflags, &zb); } arc_buf_destroy(abuf, private); } /* * Issue prefetch reads for the given block on the given level. If the indirect * blocks above that block are not in memory, we will read them in * asynchronously. As a result, this call never blocks waiting for a read to * complete. Note that the prefetch might fail if the dataset is encrypted and * the encryption key is unmapped before the IO completes. */ int dbuf_prefetch_impl(dnode_t *dn, int64_t level, uint64_t blkid, zio_priority_t prio, arc_flags_t aflags, dbuf_prefetch_fn cb, void *arg) { blkptr_t bp; int epbs, nlevels, curlevel; uint64_t curblkid; ASSERT(blkid != DMU_BONUS_BLKID); ASSERT(RW_LOCK_HELD(&dn->dn_struct_rwlock)); if (blkid > dn->dn_maxblkid) goto no_issue; if (level == 0 && dnode_block_freed(dn, blkid)) goto no_issue; /* * This dnode hasn't been written to disk yet, so there's nothing to * prefetch. */ nlevels = dn->dn_phys->dn_nlevels; if (level >= nlevels || dn->dn_phys->dn_nblkptr == 0) goto no_issue; epbs = dn->dn_phys->dn_indblkshift - SPA_BLKPTRSHIFT; if (dn->dn_phys->dn_maxblkid < blkid << (epbs * level)) goto no_issue; dmu_buf_impl_t *db = dbuf_find(dn->dn_objset, dn->dn_object, level, blkid, NULL); if (db != NULL) { mutex_exit(&db->db_mtx); /* * This dbuf already exists. It is either CACHED, or * (we assume) about to be read or filled. */ goto no_issue; } /* * Find the closest ancestor (indirect block) of the target block * that is present in the cache. In this indirect block, we will * find the bp that is at curlevel, curblkid. */ curlevel = level; curblkid = blkid; while (curlevel < nlevels - 1) { int parent_level = curlevel + 1; uint64_t parent_blkid = curblkid >> epbs; dmu_buf_impl_t *db; if (dbuf_hold_impl(dn, parent_level, parent_blkid, FALSE, TRUE, FTAG, &db) == 0) { blkptr_t *bpp = db->db_buf->b_data; bp = bpp[P2PHASE(curblkid, 1 << epbs)]; dbuf_rele(db, FTAG); break; } curlevel = parent_level; curblkid = parent_blkid; } if (curlevel == nlevels - 1) { /* No cached indirect blocks found. */ ASSERT3U(curblkid, <, dn->dn_phys->dn_nblkptr); bp = dn->dn_phys->dn_blkptr[curblkid]; } ASSERT(!BP_IS_REDACTED(&bp) || dsl_dataset_feature_is_active(dn->dn_objset->os_dsl_dataset, SPA_FEATURE_REDACTED_DATASETS)); if (BP_IS_HOLE(&bp) || BP_IS_REDACTED(&bp)) goto no_issue; ASSERT3U(curlevel, ==, BP_GET_LEVEL(&bp)); zio_t *pio = zio_root(dmu_objset_spa(dn->dn_objset), NULL, NULL, ZIO_FLAG_CANFAIL); dbuf_prefetch_arg_t *dpa = kmem_zalloc(sizeof (*dpa), KM_SLEEP); dsl_dataset_t *ds = dn->dn_objset->os_dsl_dataset; SET_BOOKMARK(&dpa->dpa_zb, ds != NULL ? ds->ds_object : DMU_META_OBJSET, dn->dn_object, level, blkid); dpa->dpa_curlevel = curlevel; dpa->dpa_prio = prio; dpa->dpa_aflags = aflags; dpa->dpa_spa = dn->dn_objset->os_spa; dpa->dpa_dnode = dn; dpa->dpa_epbs = epbs; dpa->dpa_zio = pio; dpa->dpa_cb = cb; dpa->dpa_arg = arg; if (!DNODE_LEVEL_IS_CACHEABLE(dn, level)) dpa->dpa_aflags |= ARC_FLAG_UNCACHED; else if (dnode_level_is_l2cacheable(&bp, dn, level)) dpa->dpa_aflags |= ARC_FLAG_L2CACHE; /* * If we have the indirect just above us, no need to do the asynchronous * prefetch chain; we'll just run the last step ourselves. If we're at * a higher level, though, we want to issue the prefetches for all the * indirect blocks asynchronously, so we can go on with whatever we were * doing. */ if (curlevel == level) { ASSERT3U(curblkid, ==, blkid); dbuf_issue_final_prefetch(dpa, &bp); } else { arc_flags_t iter_aflags = ARC_FLAG_NOWAIT; zbookmark_phys_t zb; /* flag if L2ARC eligible, l2arc_noprefetch then decides */ if (dnode_level_is_l2cacheable(&bp, dn, level)) iter_aflags |= ARC_FLAG_L2CACHE; SET_BOOKMARK(&zb, ds != NULL ? ds->ds_object : DMU_META_OBJSET, dn->dn_object, curlevel, curblkid); (void) arc_read(dpa->dpa_zio, dpa->dpa_spa, &bp, dbuf_prefetch_indirect_done, dpa, ZIO_PRIORITY_SYNC_READ, ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE, &iter_aflags, &zb); } /* * We use pio here instead of dpa_zio since it's possible that * dpa may have already been freed. */ zio_nowait(pio); return (1); no_issue: if (cb != NULL) cb(arg, level, blkid, B_FALSE); return (0); } int dbuf_prefetch(dnode_t *dn, int64_t level, uint64_t blkid, zio_priority_t prio, arc_flags_t aflags) { return (dbuf_prefetch_impl(dn, level, blkid, prio, aflags, NULL, NULL)); } /* * Helper function for dbuf_hold_impl() to copy a buffer. Handles * the case of encrypted, compressed and uncompressed buffers by * allocating the new buffer, respectively, with arc_alloc_raw_buf(), * arc_alloc_compressed_buf() or arc_alloc_buf().* * * NOTE: Declared noinline to avoid stack bloat in dbuf_hold_impl(). */ noinline static void dbuf_hold_copy(dnode_t *dn, dmu_buf_impl_t *db) { dbuf_dirty_record_t *dr = db->db_data_pending; arc_buf_t *data = dr->dt.dl.dr_data; enum zio_compress compress_type = arc_get_compression(data); uint8_t complevel = arc_get_complevel(data); if (arc_is_encrypted(data)) { boolean_t byteorder; uint8_t salt[ZIO_DATA_SALT_LEN]; uint8_t iv[ZIO_DATA_IV_LEN]; uint8_t mac[ZIO_DATA_MAC_LEN]; arc_get_raw_params(data, &byteorder, salt, iv, mac); dbuf_set_data(db, arc_alloc_raw_buf(dn->dn_objset->os_spa, db, dmu_objset_id(dn->dn_objset), byteorder, salt, iv, mac, dn->dn_type, arc_buf_size(data), arc_buf_lsize(data), compress_type, complevel)); } else if (compress_type != ZIO_COMPRESS_OFF) { dbuf_set_data(db, arc_alloc_compressed_buf( dn->dn_objset->os_spa, db, arc_buf_size(data), arc_buf_lsize(data), compress_type, complevel)); } else { dbuf_set_data(db, arc_alloc_buf(dn->dn_objset->os_spa, db, DBUF_GET_BUFC_TYPE(db), db->db.db_size)); } rw_enter(&db->db_rwlock, RW_WRITER); memcpy(db->db.db_data, data->b_data, arc_buf_size(data)); rw_exit(&db->db_rwlock); } /* * Returns with db_holds incremented, and db_mtx not held. * Note: dn_struct_rwlock must be held. */ int dbuf_hold_impl(dnode_t *dn, uint8_t level, uint64_t blkid, boolean_t fail_sparse, boolean_t fail_uncached, const void *tag, dmu_buf_impl_t **dbp) { dmu_buf_impl_t *db, *parent = NULL; uint64_t hv; /* If the pool has been created, verify the tx_sync_lock is not held */ spa_t *spa = dn->dn_objset->os_spa; dsl_pool_t *dp = spa->spa_dsl_pool; if (dp != NULL) { ASSERT(!MUTEX_HELD(&dp->dp_tx.tx_sync_lock)); } ASSERT(blkid != DMU_BONUS_BLKID); ASSERT(RW_LOCK_HELD(&dn->dn_struct_rwlock)); ASSERT3U(dn->dn_nlevels, >, level); *dbp = NULL; /* dbuf_find() returns with db_mtx held */ db = dbuf_find(dn->dn_objset, dn->dn_object, level, blkid, &hv); if (db == NULL) { blkptr_t *bp = NULL; int err; if (fail_uncached) return (SET_ERROR(ENOENT)); ASSERT3P(parent, ==, NULL); err = dbuf_findbp(dn, level, blkid, fail_sparse, &parent, &bp); if (fail_sparse) { if (err == 0 && bp && BP_IS_HOLE(bp)) err = SET_ERROR(ENOENT); if (err) { if (parent) dbuf_rele(parent, NULL); return (err); } } if (err && err != ENOENT) return (err); db = dbuf_create(dn, level, blkid, parent, bp, hv); } if (fail_uncached && db->db_state != DB_CACHED) { mutex_exit(&db->db_mtx); return (SET_ERROR(ENOENT)); } if (db->db_buf != NULL) { arc_buf_access(db->db_buf); ASSERT3P(db->db.db_data, ==, db->db_buf->b_data); } ASSERT(db->db_buf == NULL || arc_referenced(db->db_buf)); /* * If this buffer is currently syncing out, and we are * still referencing it from db_data, we need to make a copy * of it in case we decide we want to dirty it again in this txg. */ if (db->db_level == 0 && db->db_blkid != DMU_BONUS_BLKID && dn->dn_object != DMU_META_DNODE_OBJECT && db->db_state == DB_CACHED && db->db_data_pending) { dbuf_dirty_record_t *dr = db->db_data_pending; if (dr->dt.dl.dr_data == db->db_buf) { ASSERT3P(db->db_buf, !=, NULL); dbuf_hold_copy(dn, db); } } if (multilist_link_active(&db->db_cache_link)) { ASSERT(zfs_refcount_is_zero(&db->db_holds)); ASSERT(db->db_caching_status == DB_DBUF_CACHE || db->db_caching_status == DB_DBUF_METADATA_CACHE); multilist_remove(&dbuf_caches[db->db_caching_status].cache, db); (void) zfs_refcount_remove_many( &dbuf_caches[db->db_caching_status].size, db->db.db_size, db); if (db->db_caching_status == DB_DBUF_METADATA_CACHE) { DBUF_STAT_BUMPDOWN(metadata_cache_count); } else { DBUF_STAT_BUMPDOWN(cache_levels[db->db_level]); DBUF_STAT_BUMPDOWN(cache_count); DBUF_STAT_DECR(cache_levels_bytes[db->db_level], db->db.db_size); } db->db_caching_status = DB_NO_CACHE; } (void) zfs_refcount_add(&db->db_holds, tag); DBUF_VERIFY(db); mutex_exit(&db->db_mtx); /* NOTE: we can't rele the parent until after we drop the db_mtx */ if (parent) dbuf_rele(parent, NULL); ASSERT3P(DB_DNODE(db), ==, dn); ASSERT3U(db->db_blkid, ==, blkid); ASSERT3U(db->db_level, ==, level); *dbp = db; return (0); } dmu_buf_impl_t * dbuf_hold(dnode_t *dn, uint64_t blkid, const void *tag) { return (dbuf_hold_level(dn, 0, blkid, tag)); } dmu_buf_impl_t * dbuf_hold_level(dnode_t *dn, int level, uint64_t blkid, const void *tag) { dmu_buf_impl_t *db; int err = dbuf_hold_impl(dn, level, blkid, FALSE, FALSE, tag, &db); return (err ? NULL : db); } void dbuf_create_bonus(dnode_t *dn) { ASSERT(RW_WRITE_HELD(&dn->dn_struct_rwlock)); ASSERT(dn->dn_bonus == NULL); dn->dn_bonus = dbuf_create(dn, 0, DMU_BONUS_BLKID, dn->dn_dbuf, NULL, dbuf_hash(dn->dn_objset, dn->dn_object, 0, DMU_BONUS_BLKID)); } int dbuf_spill_set_blksz(dmu_buf_t *db_fake, uint64_t blksz, dmu_tx_t *tx) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake; if (db->db_blkid != DMU_SPILL_BLKID) return (SET_ERROR(ENOTSUP)); if (blksz == 0) blksz = SPA_MINBLOCKSIZE; ASSERT3U(blksz, <=, spa_maxblocksize(dmu_objset_spa(db->db_objset))); blksz = P2ROUNDUP(blksz, SPA_MINBLOCKSIZE); dbuf_new_size(db, blksz, tx); return (0); } void dbuf_rm_spill(dnode_t *dn, dmu_tx_t *tx) { dbuf_free_range(dn, DMU_SPILL_BLKID, DMU_SPILL_BLKID, tx); } #pragma weak dmu_buf_add_ref = dbuf_add_ref void dbuf_add_ref(dmu_buf_impl_t *db, const void *tag) { int64_t holds = zfs_refcount_add(&db->db_holds, tag); VERIFY3S(holds, >, 1); } #pragma weak dmu_buf_try_add_ref = dbuf_try_add_ref boolean_t dbuf_try_add_ref(dmu_buf_t *db_fake, objset_t *os, uint64_t obj, uint64_t blkid, const void *tag) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake; dmu_buf_impl_t *found_db; boolean_t result = B_FALSE; if (blkid == DMU_BONUS_BLKID) found_db = dbuf_find_bonus(os, obj); else found_db = dbuf_find(os, obj, 0, blkid, NULL); if (found_db != NULL) { if (db == found_db && dbuf_refcount(db) > db->db_dirtycnt) { (void) zfs_refcount_add(&db->db_holds, tag); result = B_TRUE; } mutex_exit(&found_db->db_mtx); } return (result); } /* * If you call dbuf_rele() you had better not be referencing the dnode handle * unless you have some other direct or indirect hold on the dnode. (An indirect * hold is a hold on one of the dnode's dbufs, including the bonus buffer.) * Without that, the dbuf_rele() could lead to a dnode_rele() followed by the * dnode's parent dbuf evicting its dnode handles. */ void dbuf_rele(dmu_buf_impl_t *db, const void *tag) { mutex_enter(&db->db_mtx); dbuf_rele_and_unlock(db, tag, B_FALSE); } void dmu_buf_rele(dmu_buf_t *db, const void *tag) { dbuf_rele((dmu_buf_impl_t *)db, tag); } /* * dbuf_rele() for an already-locked dbuf. This is necessary to allow * db_dirtycnt and db_holds to be updated atomically. The 'evicting' * argument should be set if we are already in the dbuf-evicting code * path, in which case we don't want to recursively evict. This allows us to * avoid deeply nested stacks that would have a call flow similar to this: * * dbuf_rele()-->dbuf_rele_and_unlock()-->dbuf_evict_notify() * ^ | * | | * +-----dbuf_destroy()<--dbuf_evict_one()<--------+ * */ void dbuf_rele_and_unlock(dmu_buf_impl_t *db, const void *tag, boolean_t evicting) { int64_t holds; uint64_t size; ASSERT(MUTEX_HELD(&db->db_mtx)); DBUF_VERIFY(db); /* * Remove the reference to the dbuf before removing its hold on the * dnode so we can guarantee in dnode_move() that a referenced bonus * buffer has a corresponding dnode hold. */ holds = zfs_refcount_remove(&db->db_holds, tag); ASSERT(holds >= 0); /* * We can't freeze indirects if there is a possibility that they * may be modified in the current syncing context. */ if (db->db_buf != NULL && holds == (db->db_level == 0 ? db->db_dirtycnt : 0)) { arc_buf_freeze(db->db_buf); } if (holds == db->db_dirtycnt && db->db_level == 0 && db->db_user_immediate_evict) dbuf_evict_user(db); if (holds == 0) { if (db->db_blkid == DMU_BONUS_BLKID) { dnode_t *dn; boolean_t evict_dbuf = db->db_pending_evict; /* * If the dnode moves here, we cannot cross this * barrier until the move completes. */ DB_DNODE_ENTER(db); dn = DB_DNODE(db); atomic_dec_32(&dn->dn_dbufs_count); /* * Decrementing the dbuf count means that the bonus * buffer's dnode hold is no longer discounted in * dnode_move(). The dnode cannot move until after * the dnode_rele() below. */ DB_DNODE_EXIT(db); /* * Do not reference db after its lock is dropped. * Another thread may evict it. */ mutex_exit(&db->db_mtx); if (evict_dbuf) dnode_evict_bonus(dn); dnode_rele(dn, db); } else if (db->db_buf == NULL) { /* * This is a special case: we never associated this * dbuf with any data allocated from the ARC. */ ASSERT(db->db_state == DB_UNCACHED || db->db_state == DB_NOFILL); dbuf_destroy(db); } else if (arc_released(db->db_buf)) { /* * This dbuf has anonymous data associated with it. */ dbuf_destroy(db); } else if (!(DBUF_IS_CACHEABLE(db) || db->db_partial_read) || db->db_pending_evict) { dbuf_destroy(db); } else if (!multilist_link_active(&db->db_cache_link)) { ASSERT3U(db->db_caching_status, ==, DB_NO_CACHE); dbuf_cached_state_t dcs = dbuf_include_in_metadata_cache(db) ? DB_DBUF_METADATA_CACHE : DB_DBUF_CACHE; db->db_caching_status = dcs; multilist_insert(&dbuf_caches[dcs].cache, db); uint64_t db_size = db->db.db_size; size = zfs_refcount_add_many( &dbuf_caches[dcs].size, db_size, db); uint8_t db_level = db->db_level; mutex_exit(&db->db_mtx); if (dcs == DB_DBUF_METADATA_CACHE) { DBUF_STAT_BUMP(metadata_cache_count); DBUF_STAT_MAX(metadata_cache_size_bytes_max, size); } else { DBUF_STAT_BUMP(cache_count); DBUF_STAT_MAX(cache_size_bytes_max, size); DBUF_STAT_BUMP(cache_levels[db_level]); DBUF_STAT_INCR(cache_levels_bytes[db_level], db_size); } if (dcs == DB_DBUF_CACHE && !evicting) dbuf_evict_notify(size); } } else { mutex_exit(&db->db_mtx); } } #pragma weak dmu_buf_refcount = dbuf_refcount uint64_t dbuf_refcount(dmu_buf_impl_t *db) { return (zfs_refcount_count(&db->db_holds)); } uint64_t dmu_buf_user_refcount(dmu_buf_t *db_fake) { uint64_t holds; dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake; mutex_enter(&db->db_mtx); ASSERT3U(zfs_refcount_count(&db->db_holds), >=, db->db_dirtycnt); holds = zfs_refcount_count(&db->db_holds) - db->db_dirtycnt; mutex_exit(&db->db_mtx); return (holds); } void * dmu_buf_replace_user(dmu_buf_t *db_fake, dmu_buf_user_t *old_user, dmu_buf_user_t *new_user) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake; mutex_enter(&db->db_mtx); dbuf_verify_user(db, DBVU_NOT_EVICTING); if (db->db_user == old_user) db->db_user = new_user; else old_user = db->db_user; dbuf_verify_user(db, DBVU_NOT_EVICTING); mutex_exit(&db->db_mtx); return (old_user); } void * dmu_buf_set_user(dmu_buf_t *db_fake, dmu_buf_user_t *user) { return (dmu_buf_replace_user(db_fake, NULL, user)); } void * dmu_buf_set_user_ie(dmu_buf_t *db_fake, dmu_buf_user_t *user) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake; db->db_user_immediate_evict = TRUE; return (dmu_buf_set_user(db_fake, user)); } void * dmu_buf_remove_user(dmu_buf_t *db_fake, dmu_buf_user_t *user) { return (dmu_buf_replace_user(db_fake, user, NULL)); } void * dmu_buf_get_user(dmu_buf_t *db_fake) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake; dbuf_verify_user(db, DBVU_NOT_EVICTING); return (db->db_user); } void dmu_buf_user_evict_wait(void) { taskq_wait(dbu_evict_taskq); } blkptr_t * dmu_buf_get_blkptr(dmu_buf_t *db) { dmu_buf_impl_t *dbi = (dmu_buf_impl_t *)db; return (dbi->db_blkptr); } objset_t * dmu_buf_get_objset(dmu_buf_t *db) { dmu_buf_impl_t *dbi = (dmu_buf_impl_t *)db; return (dbi->db_objset); } dnode_t * dmu_buf_dnode_enter(dmu_buf_t *db) { dmu_buf_impl_t *dbi = (dmu_buf_impl_t *)db; DB_DNODE_ENTER(dbi); return (DB_DNODE(dbi)); } void dmu_buf_dnode_exit(dmu_buf_t *db) { dmu_buf_impl_t *dbi = (dmu_buf_impl_t *)db; DB_DNODE_EXIT(dbi); } static void dbuf_check_blkptr(dnode_t *dn, dmu_buf_impl_t *db) { /* ASSERT(dmu_tx_is_syncing(tx) */ ASSERT(MUTEX_HELD(&db->db_mtx)); if (db->db_blkptr != NULL) return; if (db->db_blkid == DMU_SPILL_BLKID) { db->db_blkptr = DN_SPILL_BLKPTR(dn->dn_phys); BP_ZERO(db->db_blkptr); return; } if (db->db_level == dn->dn_phys->dn_nlevels-1) { /* * This buffer was allocated at a time when there was * no available blkptrs from the dnode, or it was * inappropriate to hook it in (i.e., nlevels mismatch). */ ASSERT(db->db_blkid < dn->dn_phys->dn_nblkptr); ASSERT(db->db_parent == NULL); db->db_parent = dn->dn_dbuf; db->db_blkptr = &dn->dn_phys->dn_blkptr[db->db_blkid]; DBUF_VERIFY(db); } else { dmu_buf_impl_t *parent = db->db_parent; int epbs = dn->dn_phys->dn_indblkshift - SPA_BLKPTRSHIFT; ASSERT(dn->dn_phys->dn_nlevels > 1); if (parent == NULL) { mutex_exit(&db->db_mtx); rw_enter(&dn->dn_struct_rwlock, RW_READER); parent = dbuf_hold_level(dn, db->db_level + 1, db->db_blkid >> epbs, db); rw_exit(&dn->dn_struct_rwlock); mutex_enter(&db->db_mtx); db->db_parent = parent; } db->db_blkptr = (blkptr_t *)parent->db.db_data + (db->db_blkid & ((1ULL << epbs) - 1)); DBUF_VERIFY(db); } } static void dbuf_sync_bonus(dbuf_dirty_record_t *dr, dmu_tx_t *tx) { dmu_buf_impl_t *db = dr->dr_dbuf; void *data = dr->dt.dl.dr_data; ASSERT0(db->db_level); ASSERT(MUTEX_HELD(&db->db_mtx)); ASSERT(db->db_blkid == DMU_BONUS_BLKID); ASSERT(data != NULL); dnode_t *dn = dr->dr_dnode; ASSERT3U(DN_MAX_BONUS_LEN(dn->dn_phys), <=, DN_SLOTS_TO_BONUSLEN(dn->dn_phys->dn_extra_slots + 1)); memcpy(DN_BONUS(dn->dn_phys), data, DN_MAX_BONUS_LEN(dn->dn_phys)); dbuf_sync_leaf_verify_bonus_dnode(dr); dbuf_undirty_bonus(dr); dbuf_rele_and_unlock(db, (void *)(uintptr_t)tx->tx_txg, B_FALSE); } /* * When syncing out a blocks of dnodes, adjust the block to deal with * encryption. Normally, we make sure the block is decrypted before writing * it. If we have crypt params, then we are writing a raw (encrypted) block, * from a raw receive. In this case, set the ARC buf's crypt params so * that the BP will be filled with the correct byteorder, salt, iv, and mac. */ static void dbuf_prepare_encrypted_dnode_leaf(dbuf_dirty_record_t *dr) { int err; dmu_buf_impl_t *db = dr->dr_dbuf; ASSERT(MUTEX_HELD(&db->db_mtx)); ASSERT3U(db->db.db_object, ==, DMU_META_DNODE_OBJECT); ASSERT3U(db->db_level, ==, 0); if (!db->db_objset->os_raw_receive && arc_is_encrypted(db->db_buf)) { zbookmark_phys_t zb; /* * Unfortunately, there is currently no mechanism for * syncing context to handle decryption errors. An error * here is only possible if an attacker maliciously * changed a dnode block and updated the associated * checksums going up the block tree. */ SET_BOOKMARK(&zb, dmu_objset_id(db->db_objset), db->db.db_object, db->db_level, db->db_blkid); err = arc_untransform(db->db_buf, db->db_objset->os_spa, &zb, B_TRUE); if (err) panic("Invalid dnode block MAC"); } else if (dr->dt.dl.dr_has_raw_params) { (void) arc_release(dr->dt.dl.dr_data, db); arc_convert_to_raw(dr->dt.dl.dr_data, dmu_objset_id(db->db_objset), dr->dt.dl.dr_byteorder, DMU_OT_DNODE, dr->dt.dl.dr_salt, dr->dt.dl.dr_iv, dr->dt.dl.dr_mac); } } /* * dbuf_sync_indirect() is called recursively from dbuf_sync_list() so it * is critical the we not allow the compiler to inline this function in to * dbuf_sync_list() thereby drastically bloating the stack usage. */ noinline static void dbuf_sync_indirect(dbuf_dirty_record_t *dr, dmu_tx_t *tx) { dmu_buf_impl_t *db = dr->dr_dbuf; dnode_t *dn = dr->dr_dnode; ASSERT(dmu_tx_is_syncing(tx)); dprintf_dbuf_bp(db, db->db_blkptr, "blkptr=%p", db->db_blkptr); mutex_enter(&db->db_mtx); ASSERT(db->db_level > 0); DBUF_VERIFY(db); /* Read the block if it hasn't been read yet. */ if (db->db_buf == NULL) { mutex_exit(&db->db_mtx); (void) dbuf_read(db, NULL, DB_RF_MUST_SUCCEED); mutex_enter(&db->db_mtx); } ASSERT3U(db->db_state, ==, DB_CACHED); ASSERT(db->db_buf != NULL); /* Indirect block size must match what the dnode thinks it is. */ ASSERT3U(db->db.db_size, ==, 1<dn_phys->dn_indblkshift); dbuf_check_blkptr(dn, db); /* Provide the pending dirty record to child dbufs */ db->db_data_pending = dr; mutex_exit(&db->db_mtx); dbuf_write(dr, db->db_buf, tx); zio_t *zio = dr->dr_zio; mutex_enter(&dr->dt.di.dr_mtx); dbuf_sync_list(&dr->dt.di.dr_children, db->db_level - 1, tx); ASSERT(list_head(&dr->dt.di.dr_children) == NULL); mutex_exit(&dr->dt.di.dr_mtx); zio_nowait(zio); } /* * Verify that the size of the data in our bonus buffer does not exceed * its recorded size. * * The purpose of this verification is to catch any cases in development * where the size of a phys structure (i.e space_map_phys_t) grows and, * due to incorrect feature management, older pools expect to read more * data even though they didn't actually write it to begin with. * * For a example, this would catch an error in the feature logic where we * open an older pool and we expect to write the space map histogram of * a space map with size SPACE_MAP_SIZE_V0. */ static void dbuf_sync_leaf_verify_bonus_dnode(dbuf_dirty_record_t *dr) { #ifdef ZFS_DEBUG dnode_t *dn = dr->dr_dnode; /* * Encrypted bonus buffers can have data past their bonuslen. * Skip the verification of these blocks. */ if (DMU_OT_IS_ENCRYPTED(dn->dn_bonustype)) return; uint16_t bonuslen = dn->dn_phys->dn_bonuslen; uint16_t maxbonuslen = DN_SLOTS_TO_BONUSLEN(dn->dn_num_slots); ASSERT3U(bonuslen, <=, maxbonuslen); arc_buf_t *datap = dr->dt.dl.dr_data; char *datap_end = ((char *)datap) + bonuslen; char *datap_max = ((char *)datap) + maxbonuslen; /* ensure that everything is zero after our data */ for (; datap_end < datap_max; datap_end++) ASSERT(*datap_end == 0); #endif } static blkptr_t * dbuf_lightweight_bp(dbuf_dirty_record_t *dr) { /* This must be a lightweight dirty record. */ ASSERT3P(dr->dr_dbuf, ==, NULL); dnode_t *dn = dr->dr_dnode; if (dn->dn_phys->dn_nlevels == 1) { VERIFY3U(dr->dt.dll.dr_blkid, <, dn->dn_phys->dn_nblkptr); return (&dn->dn_phys->dn_blkptr[dr->dt.dll.dr_blkid]); } else { dmu_buf_impl_t *parent_db = dr->dr_parent->dr_dbuf; int epbs = dn->dn_indblkshift - SPA_BLKPTRSHIFT; VERIFY3U(parent_db->db_level, ==, 1); VERIFY3P(parent_db->db_dnode_handle->dnh_dnode, ==, dn); VERIFY3U(dr->dt.dll.dr_blkid >> epbs, ==, parent_db->db_blkid); blkptr_t *bp = parent_db->db.db_data; return (&bp[dr->dt.dll.dr_blkid & ((1 << epbs) - 1)]); } } static void dbuf_lightweight_ready(zio_t *zio) { dbuf_dirty_record_t *dr = zio->io_private; blkptr_t *bp = zio->io_bp; if (zio->io_error != 0) return; dnode_t *dn = dr->dr_dnode; blkptr_t *bp_orig = dbuf_lightweight_bp(dr); spa_t *spa = dmu_objset_spa(dn->dn_objset); int64_t delta = bp_get_dsize_sync(spa, bp) - bp_get_dsize_sync(spa, bp_orig); dnode_diduse_space(dn, delta); uint64_t blkid = dr->dt.dll.dr_blkid; mutex_enter(&dn->dn_mtx); if (blkid > dn->dn_phys->dn_maxblkid) { ASSERT0(dn->dn_objset->os_raw_receive); dn->dn_phys->dn_maxblkid = blkid; } mutex_exit(&dn->dn_mtx); if (!BP_IS_EMBEDDED(bp)) { uint64_t fill = BP_IS_HOLE(bp) ? 0 : 1; BP_SET_FILL(bp, fill); } dmu_buf_impl_t *parent_db; EQUIV(dr->dr_parent == NULL, dn->dn_phys->dn_nlevels == 1); if (dr->dr_parent == NULL) { parent_db = dn->dn_dbuf; } else { parent_db = dr->dr_parent->dr_dbuf; } rw_enter(&parent_db->db_rwlock, RW_WRITER); *bp_orig = *bp; rw_exit(&parent_db->db_rwlock); } static void dbuf_lightweight_physdone(zio_t *zio) { dbuf_dirty_record_t *dr = zio->io_private; dsl_pool_t *dp = spa_get_dsl(zio->io_spa); ASSERT3U(dr->dr_txg, ==, zio->io_txg); /* * The callback will be called io_phys_children times. Retire one * portion of our dirty space each time we are called. Any rounding * error will be cleaned up by dbuf_lightweight_done(). */ int delta = dr->dr_accounted / zio->io_phys_children; dsl_pool_undirty_space(dp, delta, zio->io_txg); } static void dbuf_lightweight_done(zio_t *zio) { dbuf_dirty_record_t *dr = zio->io_private; VERIFY0(zio->io_error); objset_t *os = dr->dr_dnode->dn_objset; dmu_tx_t *tx = os->os_synctx; if (zio->io_flags & (ZIO_FLAG_IO_REWRITE | ZIO_FLAG_NOPWRITE)) { ASSERT(BP_EQUAL(zio->io_bp, &zio->io_bp_orig)); } else { dsl_dataset_t *ds = os->os_dsl_dataset; (void) dsl_dataset_block_kill(ds, &zio->io_bp_orig, tx, B_TRUE); dsl_dataset_block_born(ds, zio->io_bp, tx); } /* * See comment in dbuf_write_done(). */ if (zio->io_phys_children == 0) { dsl_pool_undirty_space(dmu_objset_pool(os), dr->dr_accounted, zio->io_txg); } else { dsl_pool_undirty_space(dmu_objset_pool(os), dr->dr_accounted % zio->io_phys_children, zio->io_txg); } abd_free(dr->dt.dll.dr_abd); kmem_free(dr, sizeof (*dr)); } noinline static void dbuf_sync_lightweight(dbuf_dirty_record_t *dr, dmu_tx_t *tx) { dnode_t *dn = dr->dr_dnode; zio_t *pio; if (dn->dn_phys->dn_nlevels == 1) { pio = dn->dn_zio; } else { pio = dr->dr_parent->dr_zio; } zbookmark_phys_t zb = { .zb_objset = dmu_objset_id(dn->dn_objset), .zb_object = dn->dn_object, .zb_level = 0, .zb_blkid = dr->dt.dll.dr_blkid, }; /* * See comment in dbuf_write(). This is so that zio->io_bp_orig * will have the old BP in dbuf_lightweight_done(). */ dr->dr_bp_copy = *dbuf_lightweight_bp(dr); dr->dr_zio = zio_write(pio, dmu_objset_spa(dn->dn_objset), dmu_tx_get_txg(tx), &dr->dr_bp_copy, dr->dt.dll.dr_abd, dn->dn_datablksz, abd_get_size(dr->dt.dll.dr_abd), &dr->dt.dll.dr_props, dbuf_lightweight_ready, NULL, dbuf_lightweight_physdone, dbuf_lightweight_done, dr, ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_MUSTSUCCEED | dr->dt.dll.dr_flags, &zb); zio_nowait(dr->dr_zio); } /* * dbuf_sync_leaf() is called recursively from dbuf_sync_list() so it is * critical the we not allow the compiler to inline this function in to * dbuf_sync_list() thereby drastically bloating the stack usage. */ noinline static void dbuf_sync_leaf(dbuf_dirty_record_t *dr, dmu_tx_t *tx) { arc_buf_t **datap = &dr->dt.dl.dr_data; dmu_buf_impl_t *db = dr->dr_dbuf; dnode_t *dn = dr->dr_dnode; objset_t *os; uint64_t txg = tx->tx_txg; ASSERT(dmu_tx_is_syncing(tx)); dprintf_dbuf_bp(db, db->db_blkptr, "blkptr=%p", db->db_blkptr); mutex_enter(&db->db_mtx); /* * To be synced, we must be dirtied. But we * might have been freed after the dirty. */ if (db->db_state == DB_UNCACHED) { /* This buffer has been freed since it was dirtied */ ASSERT(db->db.db_data == NULL); } else if (db->db_state == DB_FILL) { /* This buffer was freed and is now being re-filled */ ASSERT(db->db.db_data != dr->dt.dl.dr_data); } else { ASSERT(db->db_state == DB_CACHED || db->db_state == DB_NOFILL); } DBUF_VERIFY(db); if (db->db_blkid == DMU_SPILL_BLKID) { mutex_enter(&dn->dn_mtx); if (!(dn->dn_phys->dn_flags & DNODE_FLAG_SPILL_BLKPTR)) { /* * In the previous transaction group, the bonus buffer * was entirely used to store the attributes for the * dnode which overrode the dn_spill field. However, * when adding more attributes to the file a spill * block was required to hold the extra attributes. * * Make sure to clear the garbage left in the dn_spill * field from the previous attributes in the bonus * buffer. Otherwise, after writing out the spill * block to the new allocated dva, it will free * the old block pointed to by the invalid dn_spill. */ db->db_blkptr = NULL; } dn->dn_phys->dn_flags |= DNODE_FLAG_SPILL_BLKPTR; mutex_exit(&dn->dn_mtx); } /* * If this is a bonus buffer, simply copy the bonus data into the * dnode. It will be written out when the dnode is synced (and it * will be synced, since it must have been dirty for dbuf_sync to * be called). */ if (db->db_blkid == DMU_BONUS_BLKID) { ASSERT(dr->dr_dbuf == db); dbuf_sync_bonus(dr, tx); return; } os = dn->dn_objset; /* * This function may have dropped the db_mtx lock allowing a dmu_sync * operation to sneak in. As a result, we need to ensure that we * don't check the dr_override_state until we have returned from * dbuf_check_blkptr. */ dbuf_check_blkptr(dn, db); /* * If this buffer is in the middle of an immediate write, * wait for the synchronous IO to complete. */ while (dr->dt.dl.dr_override_state == DR_IN_DMU_SYNC) { ASSERT(dn->dn_object != DMU_META_DNODE_OBJECT); cv_wait(&db->db_changed, &db->db_mtx); } /* * If this is a dnode block, ensure it is appropriately encrypted * or decrypted, depending on what we are writing to it this txg. */ if (os->os_encrypted && dn->dn_object == DMU_META_DNODE_OBJECT) dbuf_prepare_encrypted_dnode_leaf(dr); if (db->db_state != DB_NOFILL && dn->dn_object != DMU_META_DNODE_OBJECT && zfs_refcount_count(&db->db_holds) > 1 && dr->dt.dl.dr_override_state != DR_OVERRIDDEN && *datap == db->db_buf) { /* * If this buffer is currently "in use" (i.e., there * are active holds and db_data still references it), * then make a copy before we start the write so that * any modifications from the open txg will not leak * into this write. * * NOTE: this copy does not need to be made for * objects only modified in the syncing context (e.g. * DNONE_DNODE blocks). */ int psize = arc_buf_size(*datap); int lsize = arc_buf_lsize(*datap); arc_buf_contents_t type = DBUF_GET_BUFC_TYPE(db); enum zio_compress compress_type = arc_get_compression(*datap); uint8_t complevel = arc_get_complevel(*datap); if (arc_is_encrypted(*datap)) { boolean_t byteorder; uint8_t salt[ZIO_DATA_SALT_LEN]; uint8_t iv[ZIO_DATA_IV_LEN]; uint8_t mac[ZIO_DATA_MAC_LEN]; arc_get_raw_params(*datap, &byteorder, salt, iv, mac); *datap = arc_alloc_raw_buf(os->os_spa, db, dmu_objset_id(os), byteorder, salt, iv, mac, dn->dn_type, psize, lsize, compress_type, complevel); } else if (compress_type != ZIO_COMPRESS_OFF) { ASSERT3U(type, ==, ARC_BUFC_DATA); *datap = arc_alloc_compressed_buf(os->os_spa, db, psize, lsize, compress_type, complevel); } else { *datap = arc_alloc_buf(os->os_spa, db, type, psize); } memcpy((*datap)->b_data, db->db.db_data, psize); } db->db_data_pending = dr; mutex_exit(&db->db_mtx); dbuf_write(dr, *datap, tx); ASSERT(!list_link_active(&dr->dr_dirty_node)); if (dn->dn_object == DMU_META_DNODE_OBJECT) { list_insert_tail(&dn->dn_dirty_records[txg & TXG_MASK], dr); } else { zio_nowait(dr->dr_zio); } } void dbuf_sync_list(list_t *list, int level, dmu_tx_t *tx) { dbuf_dirty_record_t *dr; while ((dr = list_head(list))) { if (dr->dr_zio != NULL) { /* * If we find an already initialized zio then we * are processing the meta-dnode, and we have finished. * The dbufs for all dnodes are put back on the list * during processing, so that we can zio_wait() * these IOs after initiating all child IOs. */ ASSERT3U(dr->dr_dbuf->db.db_object, ==, DMU_META_DNODE_OBJECT); break; } list_remove(list, dr); if (dr->dr_dbuf == NULL) { dbuf_sync_lightweight(dr, tx); } else { if (dr->dr_dbuf->db_blkid != DMU_BONUS_BLKID && dr->dr_dbuf->db_blkid != DMU_SPILL_BLKID) { VERIFY3U(dr->dr_dbuf->db_level, ==, level); } if (dr->dr_dbuf->db_level > 0) dbuf_sync_indirect(dr, tx); else dbuf_sync_leaf(dr, tx); } } } static void dbuf_write_ready(zio_t *zio, arc_buf_t *buf, void *vdb) { (void) buf; dmu_buf_impl_t *db = vdb; dnode_t *dn; blkptr_t *bp = zio->io_bp; blkptr_t *bp_orig = &zio->io_bp_orig; spa_t *spa = zio->io_spa; int64_t delta; uint64_t fill = 0; int i; ASSERT3P(db->db_blkptr, !=, NULL); ASSERT3P(&db->db_data_pending->dr_bp_copy, ==, bp); DB_DNODE_ENTER(db); dn = DB_DNODE(db); delta = bp_get_dsize_sync(spa, bp) - bp_get_dsize_sync(spa, bp_orig); dnode_diduse_space(dn, delta - zio->io_prev_space_delta); zio->io_prev_space_delta = delta; if (bp->blk_birth != 0) { ASSERT((db->db_blkid != DMU_SPILL_BLKID && BP_GET_TYPE(bp) == dn->dn_type) || (db->db_blkid == DMU_SPILL_BLKID && BP_GET_TYPE(bp) == dn->dn_bonustype) || BP_IS_EMBEDDED(bp)); ASSERT(BP_GET_LEVEL(bp) == db->db_level); } mutex_enter(&db->db_mtx); #ifdef ZFS_DEBUG if (db->db_blkid == DMU_SPILL_BLKID) { ASSERT(dn->dn_phys->dn_flags & DNODE_FLAG_SPILL_BLKPTR); ASSERT(!(BP_IS_HOLE(bp)) && db->db_blkptr == DN_SPILL_BLKPTR(dn->dn_phys)); } #endif if (db->db_level == 0) { mutex_enter(&dn->dn_mtx); if (db->db_blkid > dn->dn_phys->dn_maxblkid && db->db_blkid != DMU_SPILL_BLKID) { ASSERT0(db->db_objset->os_raw_receive); dn->dn_phys->dn_maxblkid = db->db_blkid; } mutex_exit(&dn->dn_mtx); if (dn->dn_type == DMU_OT_DNODE) { i = 0; while (i < db->db.db_size) { dnode_phys_t *dnp = (void *)(((char *)db->db.db_data) + i); i += DNODE_MIN_SIZE; if (dnp->dn_type != DMU_OT_NONE) { fill++; for (int j = 0; j < dnp->dn_nblkptr; j++) { (void) zfs_blkptr_verify(spa, &dnp->dn_blkptr[j], BLK_CONFIG_SKIP, BLK_VERIFY_HALT); } if (dnp->dn_flags & DNODE_FLAG_SPILL_BLKPTR) { (void) zfs_blkptr_verify(spa, DN_SPILL_BLKPTR(dnp), BLK_CONFIG_SKIP, BLK_VERIFY_HALT); } i += dnp->dn_extra_slots * DNODE_MIN_SIZE; } } } else { if (BP_IS_HOLE(bp)) { fill = 0; } else { fill = 1; } } } else { blkptr_t *ibp = db->db.db_data; ASSERT3U(db->db.db_size, ==, 1<dn_phys->dn_indblkshift); for (i = db->db.db_size >> SPA_BLKPTRSHIFT; i > 0; i--, ibp++) { if (BP_IS_HOLE(ibp)) continue; (void) zfs_blkptr_verify(spa, ibp, BLK_CONFIG_SKIP, BLK_VERIFY_HALT); fill += BP_GET_FILL(ibp); } } DB_DNODE_EXIT(db); if (!BP_IS_EMBEDDED(bp)) BP_SET_FILL(bp, fill); mutex_exit(&db->db_mtx); db_lock_type_t dblt = dmu_buf_lock_parent(db, RW_WRITER, FTAG); *db->db_blkptr = *bp; dmu_buf_unlock_parent(db, dblt, FTAG); } /* * This function gets called just prior to running through the compression * stage of the zio pipeline. If we're an indirect block comprised of only * holes, then we want this indirect to be compressed away to a hole. In * order to do that we must zero out any information about the holes that * this indirect points to prior to before we try to compress it. */ static void dbuf_write_children_ready(zio_t *zio, arc_buf_t *buf, void *vdb) { (void) zio, (void) buf; dmu_buf_impl_t *db = vdb; dnode_t *dn; blkptr_t *bp; unsigned int epbs, i; ASSERT3U(db->db_level, >, 0); DB_DNODE_ENTER(db); dn = DB_DNODE(db); epbs = dn->dn_phys->dn_indblkshift - SPA_BLKPTRSHIFT; ASSERT3U(epbs, <, 31); /* Determine if all our children are holes */ for (i = 0, bp = db->db.db_data; i < 1ULL << epbs; i++, bp++) { if (!BP_IS_HOLE(bp)) break; } /* * If all the children are holes, then zero them all out so that * we may get compressed away. */ if (i == 1ULL << epbs) { /* * We only found holes. Grab the rwlock to prevent * anybody from reading the blocks we're about to * zero out. */ rw_enter(&db->db_rwlock, RW_WRITER); memset(db->db.db_data, 0, db->db.db_size); rw_exit(&db->db_rwlock); } DB_DNODE_EXIT(db); } /* * The SPA will call this callback several times for each zio - once * for every physical child i/o (zio->io_phys_children times). This * allows the DMU to monitor the progress of each logical i/o. For example, * there may be 2 copies of an indirect block, or many fragments of a RAID-Z * block. There may be a long delay before all copies/fragments are completed, * so this callback allows us to retire dirty space gradually, as the physical * i/os complete. */ static void dbuf_write_physdone(zio_t *zio, arc_buf_t *buf, void *arg) { (void) buf; dmu_buf_impl_t *db = arg; objset_t *os = db->db_objset; dsl_pool_t *dp = dmu_objset_pool(os); dbuf_dirty_record_t *dr; int delta = 0; dr = db->db_data_pending; ASSERT3U(dr->dr_txg, ==, zio->io_txg); /* * The callback will be called io_phys_children times. Retire one * portion of our dirty space each time we are called. Any rounding * error will be cleaned up by dbuf_write_done(). */ delta = dr->dr_accounted / zio->io_phys_children; dsl_pool_undirty_space(dp, delta, zio->io_txg); } static void dbuf_write_done(zio_t *zio, arc_buf_t *buf, void *vdb) { (void) buf; dmu_buf_impl_t *db = vdb; blkptr_t *bp_orig = &zio->io_bp_orig; blkptr_t *bp = db->db_blkptr; objset_t *os = db->db_objset; dmu_tx_t *tx = os->os_synctx; ASSERT0(zio->io_error); ASSERT(db->db_blkptr == bp); /* * For nopwrites and rewrites we ensure that the bp matches our * original and bypass all the accounting. */ if (zio->io_flags & (ZIO_FLAG_IO_REWRITE | ZIO_FLAG_NOPWRITE)) { ASSERT(BP_EQUAL(bp, bp_orig)); } else { dsl_dataset_t *ds = os->os_dsl_dataset; (void) dsl_dataset_block_kill(ds, bp_orig, tx, B_TRUE); dsl_dataset_block_born(ds, bp, tx); } mutex_enter(&db->db_mtx); DBUF_VERIFY(db); dbuf_dirty_record_t *dr = db->db_data_pending; dnode_t *dn = dr->dr_dnode; ASSERT(!list_link_active(&dr->dr_dirty_node)); ASSERT(dr->dr_dbuf == db); ASSERT(list_next(&db->db_dirty_records, dr) == NULL); list_remove(&db->db_dirty_records, dr); #ifdef ZFS_DEBUG if (db->db_blkid == DMU_SPILL_BLKID) { ASSERT(dn->dn_phys->dn_flags & DNODE_FLAG_SPILL_BLKPTR); ASSERT(!(BP_IS_HOLE(db->db_blkptr)) && db->db_blkptr == DN_SPILL_BLKPTR(dn->dn_phys)); } #endif if (db->db_level == 0) { ASSERT(db->db_blkid != DMU_BONUS_BLKID); ASSERT(dr->dt.dl.dr_override_state == DR_NOT_OVERRIDDEN); if (db->db_state != DB_NOFILL) { if (dr->dt.dl.dr_data != NULL && dr->dt.dl.dr_data != db->db_buf) { arc_buf_destroy(dr->dt.dl.dr_data, db); } } } else { ASSERT(list_head(&dr->dt.di.dr_children) == NULL); ASSERT3U(db->db.db_size, ==, 1 << dn->dn_phys->dn_indblkshift); if (!BP_IS_HOLE(db->db_blkptr)) { int epbs __maybe_unused = dn->dn_phys->dn_indblkshift - SPA_BLKPTRSHIFT; ASSERT3U(db->db_blkid, <=, dn->dn_phys->dn_maxblkid >> (db->db_level * epbs)); ASSERT3U(BP_GET_LSIZE(db->db_blkptr), ==, db->db.db_size); } mutex_destroy(&dr->dt.di.dr_mtx); list_destroy(&dr->dt.di.dr_children); } cv_broadcast(&db->db_changed); ASSERT(db->db_dirtycnt > 0); db->db_dirtycnt -= 1; db->db_data_pending = NULL; dbuf_rele_and_unlock(db, (void *)(uintptr_t)tx->tx_txg, B_FALSE); /* * If we didn't do a physical write in this ZIO and we * still ended up here, it means that the space of the * dbuf that we just released (and undirtied) above hasn't * been marked as undirtied in the pool's accounting. * * Thus, we undirty that space in the pool's view of the * world here. For physical writes this type of update * happens in dbuf_write_physdone(). * * If we did a physical write, cleanup any rounding errors * that came up due to writing multiple copies of a block * on disk [see dbuf_write_physdone()]. */ if (zio->io_phys_children == 0) { dsl_pool_undirty_space(dmu_objset_pool(os), dr->dr_accounted, zio->io_txg); } else { dsl_pool_undirty_space(dmu_objset_pool(os), dr->dr_accounted % zio->io_phys_children, zio->io_txg); } kmem_free(dr, sizeof (dbuf_dirty_record_t)); } static void dbuf_write_nofill_ready(zio_t *zio) { dbuf_write_ready(zio, NULL, zio->io_private); } static void dbuf_write_nofill_done(zio_t *zio) { dbuf_write_done(zio, NULL, zio->io_private); } static void dbuf_write_override_ready(zio_t *zio) { dbuf_dirty_record_t *dr = zio->io_private; dmu_buf_impl_t *db = dr->dr_dbuf; dbuf_write_ready(zio, NULL, db); } static void dbuf_write_override_done(zio_t *zio) { dbuf_dirty_record_t *dr = zio->io_private; dmu_buf_impl_t *db = dr->dr_dbuf; blkptr_t *obp = &dr->dt.dl.dr_overridden_by; mutex_enter(&db->db_mtx); if (!BP_EQUAL(zio->io_bp, obp)) { if (!BP_IS_HOLE(obp)) dsl_free(spa_get_dsl(zio->io_spa), zio->io_txg, obp); arc_release(dr->dt.dl.dr_data, db); } mutex_exit(&db->db_mtx); dbuf_write_done(zio, NULL, db); if (zio->io_abd != NULL) abd_free(zio->io_abd); } typedef struct dbuf_remap_impl_callback_arg { objset_t *drica_os; uint64_t drica_blk_birth; dmu_tx_t *drica_tx; } dbuf_remap_impl_callback_arg_t; static void dbuf_remap_impl_callback(uint64_t vdev, uint64_t offset, uint64_t size, void *arg) { dbuf_remap_impl_callback_arg_t *drica = arg; objset_t *os = drica->drica_os; spa_t *spa = dmu_objset_spa(os); dmu_tx_t *tx = drica->drica_tx; ASSERT(dsl_pool_sync_context(spa_get_dsl(spa))); if (os == spa_meta_objset(spa)) { spa_vdev_indirect_mark_obsolete(spa, vdev, offset, size, tx); } else { dsl_dataset_block_remapped(dmu_objset_ds(os), vdev, offset, size, drica->drica_blk_birth, tx); } } static void dbuf_remap_impl(dnode_t *dn, blkptr_t *bp, krwlock_t *rw, dmu_tx_t *tx) { blkptr_t bp_copy = *bp; spa_t *spa = dmu_objset_spa(dn->dn_objset); dbuf_remap_impl_callback_arg_t drica; ASSERT(dsl_pool_sync_context(spa_get_dsl(spa))); drica.drica_os = dn->dn_objset; drica.drica_blk_birth = bp->blk_birth; drica.drica_tx = tx; if (spa_remap_blkptr(spa, &bp_copy, dbuf_remap_impl_callback, &drica)) { /* * If the blkptr being remapped is tracked by a livelist, * then we need to make sure the livelist reflects the update. * First, cancel out the old blkptr by appending a 'FREE' * entry. Next, add an 'ALLOC' to track the new version. This * way we avoid trying to free an inaccurate blkptr at delete. * Note that embedded blkptrs are not tracked in livelists. */ if (dn->dn_objset != spa_meta_objset(spa)) { dsl_dataset_t *ds = dmu_objset_ds(dn->dn_objset); if (dsl_deadlist_is_open(&ds->ds_dir->dd_livelist) && bp->blk_birth > ds->ds_dir->dd_origin_txg) { ASSERT(!BP_IS_EMBEDDED(bp)); ASSERT(dsl_dir_is_clone(ds->ds_dir)); ASSERT(spa_feature_is_enabled(spa, SPA_FEATURE_LIVELIST)); bplist_append(&ds->ds_dir->dd_pending_frees, bp); bplist_append(&ds->ds_dir->dd_pending_allocs, &bp_copy); } } /* * The db_rwlock prevents dbuf_read_impl() from * dereferencing the BP while we are changing it. To * avoid lock contention, only grab it when we are actually * changing the BP. */ if (rw != NULL) rw_enter(rw, RW_WRITER); *bp = bp_copy; if (rw != NULL) rw_exit(rw); } } /* * Remap any existing BP's to concrete vdevs, if possible. */ static void dbuf_remap(dnode_t *dn, dmu_buf_impl_t *db, dmu_tx_t *tx) { spa_t *spa = dmu_objset_spa(db->db_objset); ASSERT(dsl_pool_sync_context(spa_get_dsl(spa))); if (!spa_feature_is_active(spa, SPA_FEATURE_DEVICE_REMOVAL)) return; if (db->db_level > 0) { blkptr_t *bp = db->db.db_data; for (int i = 0; i < db->db.db_size >> SPA_BLKPTRSHIFT; i++) { dbuf_remap_impl(dn, &bp[i], &db->db_rwlock, tx); } } else if (db->db.db_object == DMU_META_DNODE_OBJECT) { dnode_phys_t *dnp = db->db.db_data; ASSERT3U(db->db_dnode_handle->dnh_dnode->dn_type, ==, DMU_OT_DNODE); for (int i = 0; i < db->db.db_size >> DNODE_SHIFT; i += dnp[i].dn_extra_slots + 1) { for (int j = 0; j < dnp[i].dn_nblkptr; j++) { krwlock_t *lock = (dn->dn_dbuf == NULL ? NULL : &dn->dn_dbuf->db_rwlock); dbuf_remap_impl(dn, &dnp[i].dn_blkptr[j], lock, tx); } } } } /* Issue I/O to commit a dirty buffer to disk. */ static void dbuf_write(dbuf_dirty_record_t *dr, arc_buf_t *data, dmu_tx_t *tx) { dmu_buf_impl_t *db = dr->dr_dbuf; dnode_t *dn = dr->dr_dnode; objset_t *os; dmu_buf_impl_t *parent = db->db_parent; uint64_t txg = tx->tx_txg; zbookmark_phys_t zb; zio_prop_t zp; zio_t *pio; /* parent I/O */ int wp_flag = 0; ASSERT(dmu_tx_is_syncing(tx)); os = dn->dn_objset; if (db->db_state != DB_NOFILL) { if (db->db_level > 0 || dn->dn_type == DMU_OT_DNODE) { /* * Private object buffers are released here rather * than in dbuf_dirty() since they are only modified * in the syncing context and we don't want the * overhead of making multiple copies of the data. */ if (BP_IS_HOLE(db->db_blkptr)) { arc_buf_thaw(data); } else { dbuf_release_bp(db); } dbuf_remap(dn, db, tx); } } if (parent != dn->dn_dbuf) { /* Our parent is an indirect block. */ /* We have a dirty parent that has been scheduled for write. */ ASSERT(parent && parent->db_data_pending); /* Our parent's buffer is one level closer to the dnode. */ ASSERT(db->db_level == parent->db_level-1); /* * We're about to modify our parent's db_data by modifying * our block pointer, so the parent must be released. */ ASSERT(arc_released(parent->db_buf)); pio = parent->db_data_pending->dr_zio; } else { /* Our parent is the dnode itself. */ ASSERT((db->db_level == dn->dn_phys->dn_nlevels-1 && db->db_blkid != DMU_SPILL_BLKID) || (db->db_blkid == DMU_SPILL_BLKID && db->db_level == 0)); if (db->db_blkid != DMU_SPILL_BLKID) ASSERT3P(db->db_blkptr, ==, &dn->dn_phys->dn_blkptr[db->db_blkid]); pio = dn->dn_zio; } ASSERT(db->db_level == 0 || data == db->db_buf); ASSERT3U(db->db_blkptr->blk_birth, <=, txg); ASSERT(pio); SET_BOOKMARK(&zb, os->os_dsl_dataset ? os->os_dsl_dataset->ds_object : DMU_META_OBJSET, db->db.db_object, db->db_level, db->db_blkid); if (db->db_blkid == DMU_SPILL_BLKID) wp_flag = WP_SPILL; wp_flag |= (db->db_state == DB_NOFILL) ? WP_NOFILL : 0; dmu_write_policy(os, dn, db->db_level, wp_flag, &zp); /* * We copy the blkptr now (rather than when we instantiate the dirty * record), because its value can change between open context and * syncing context. We do not need to hold dn_struct_rwlock to read * db_blkptr because we are in syncing context. */ dr->dr_bp_copy = *db->db_blkptr; if (db->db_level == 0 && dr->dt.dl.dr_override_state == DR_OVERRIDDEN) { /* * The BP for this block has been provided by open context * (by dmu_sync() or dmu_buf_write_embedded()). */ abd_t *contents = (data != NULL) ? abd_get_from_buf(data->b_data, arc_buf_size(data)) : NULL; dr->dr_zio = zio_write(pio, os->os_spa, txg, &dr->dr_bp_copy, contents, db->db.db_size, db->db.db_size, &zp, dbuf_write_override_ready, NULL, NULL, dbuf_write_override_done, dr, ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_MUSTSUCCEED, &zb); mutex_enter(&db->db_mtx); dr->dt.dl.dr_override_state = DR_NOT_OVERRIDDEN; zio_write_override(dr->dr_zio, &dr->dt.dl.dr_overridden_by, dr->dt.dl.dr_copies, dr->dt.dl.dr_nopwrite, dr->dt.dl.dr_brtwrite); mutex_exit(&db->db_mtx); } else if (db->db_state == DB_NOFILL) { ASSERT(zp.zp_checksum == ZIO_CHECKSUM_OFF || zp.zp_checksum == ZIO_CHECKSUM_NOPARITY); dr->dr_zio = zio_write(pio, os->os_spa, txg, &dr->dr_bp_copy, NULL, db->db.db_size, db->db.db_size, &zp, dbuf_write_nofill_ready, NULL, NULL, dbuf_write_nofill_done, db, ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_MUSTSUCCEED | ZIO_FLAG_NODATA, &zb); } else { ASSERT(arc_released(data)); /* * For indirect blocks, we want to setup the children * ready callback so that we can properly handle an indirect * block that only contains holes. */ arc_write_done_func_t *children_ready_cb = NULL; if (db->db_level != 0) children_ready_cb = dbuf_write_children_ready; dr->dr_zio = arc_write(pio, os->os_spa, txg, &dr->dr_bp_copy, data, !DBUF_IS_CACHEABLE(db), dbuf_is_l2cacheable(db), &zp, dbuf_write_ready, children_ready_cb, dbuf_write_physdone, dbuf_write_done, db, ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_MUSTSUCCEED, &zb); } } EXPORT_SYMBOL(dbuf_find); EXPORT_SYMBOL(dbuf_is_metadata); EXPORT_SYMBOL(dbuf_destroy); EXPORT_SYMBOL(dbuf_loan_arcbuf); EXPORT_SYMBOL(dbuf_whichblock); EXPORT_SYMBOL(dbuf_read); EXPORT_SYMBOL(dbuf_unoverride); EXPORT_SYMBOL(dbuf_free_range); EXPORT_SYMBOL(dbuf_new_size); EXPORT_SYMBOL(dbuf_release_bp); EXPORT_SYMBOL(dbuf_dirty); EXPORT_SYMBOL(dmu_buf_set_crypt_params); EXPORT_SYMBOL(dmu_buf_will_dirty); EXPORT_SYMBOL(dmu_buf_is_dirty); EXPORT_SYMBOL(dmu_buf_will_clone); EXPORT_SYMBOL(dmu_buf_will_not_fill); EXPORT_SYMBOL(dmu_buf_will_fill); EXPORT_SYMBOL(dmu_buf_fill_done); EXPORT_SYMBOL(dmu_buf_rele); EXPORT_SYMBOL(dbuf_assign_arcbuf); EXPORT_SYMBOL(dbuf_prefetch); EXPORT_SYMBOL(dbuf_hold_impl); EXPORT_SYMBOL(dbuf_hold); EXPORT_SYMBOL(dbuf_hold_level); EXPORT_SYMBOL(dbuf_create_bonus); EXPORT_SYMBOL(dbuf_spill_set_blksz); EXPORT_SYMBOL(dbuf_rm_spill); EXPORT_SYMBOL(dbuf_add_ref); EXPORT_SYMBOL(dbuf_rele); EXPORT_SYMBOL(dbuf_rele_and_unlock); EXPORT_SYMBOL(dbuf_refcount); EXPORT_SYMBOL(dbuf_sync_list); EXPORT_SYMBOL(dmu_buf_set_user); EXPORT_SYMBOL(dmu_buf_set_user_ie); EXPORT_SYMBOL(dmu_buf_get_user); EXPORT_SYMBOL(dmu_buf_get_blkptr); ZFS_MODULE_PARAM(zfs_dbuf_cache, dbuf_cache_, max_bytes, U64, ZMOD_RW, "Maximum size in bytes of the dbuf cache."); ZFS_MODULE_PARAM(zfs_dbuf_cache, dbuf_cache_, hiwater_pct, UINT, ZMOD_RW, "Percentage over dbuf_cache_max_bytes for direct dbuf eviction."); ZFS_MODULE_PARAM(zfs_dbuf_cache, dbuf_cache_, lowater_pct, UINT, ZMOD_RW, "Percentage below dbuf_cache_max_bytes when dbuf eviction stops."); ZFS_MODULE_PARAM(zfs_dbuf, dbuf_, metadata_cache_max_bytes, U64, ZMOD_RW, "Maximum size in bytes of dbuf metadata cache."); ZFS_MODULE_PARAM(zfs_dbuf, dbuf_, cache_shift, UINT, ZMOD_RW, "Set size of dbuf cache to log2 fraction of arc size."); ZFS_MODULE_PARAM(zfs_dbuf, dbuf_, metadata_cache_shift, UINT, ZMOD_RW, "Set size of dbuf metadata cache to log2 fraction of arc size."); ZFS_MODULE_PARAM(zfs_dbuf, dbuf_, mutex_cache_shift, UINT, ZMOD_RD, "Set size of dbuf cache mutex array as log2 shift."); diff --git a/sys/contrib/openzfs/module/zfs/dsl_scan.c b/sys/contrib/openzfs/module/zfs/dsl_scan.c index d398b6705551..5e3559b251e3 100644 --- a/sys/contrib/openzfs/module/zfs/dsl_scan.c +++ b/sys/contrib/openzfs/module/zfs/dsl_scan.c @@ -1,4561 +1,5247 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2008, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2011, 2021 by Delphix. All rights reserved. * Copyright 2016 Gary Mills * Copyright (c) 2017, 2019, Datto Inc. All rights reserved. * Copyright (c) 2015, Nexenta Systems, Inc. All rights reserved. * Copyright 2019 Joyent, Inc. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include +#include #ifdef _KERNEL #include #endif /* * Grand theory statement on scan queue sorting * * Scanning is implemented by recursively traversing all indirection levels * in an object and reading all blocks referenced from said objects. This * results in us approximately traversing the object from lowest logical * offset to the highest. For best performance, we would want the logical * blocks to be physically contiguous. However, this is frequently not the * case with pools given the allocation patterns of copy-on-write filesystems. * So instead, we put the I/Os into a reordering queue and issue them in a * way that will most benefit physical disks (LBA-order). * * Queue management: * * Ideally, we would want to scan all metadata and queue up all block I/O * prior to starting to issue it, because that allows us to do an optimal * sorting job. This can however consume large amounts of memory. Therefore * we continuously monitor the size of the queues and constrain them to 5% * (zfs_scan_mem_lim_fact) of physmem. If the queues grow larger than this * limit, we clear out a few of the largest extents at the head of the queues * to make room for more scanning. Hopefully, these extents will be fairly * large and contiguous, allowing us to approach sequential I/O throughput * even without a fully sorted tree. * * Metadata scanning takes place in dsl_scan_visit(), which is called from * dsl_scan_sync() every spa_sync(). If we have either fully scanned all * metadata on the pool, or we need to make room in memory because our * queues are too large, dsl_scan_visit() is postponed and * scan_io_queues_run() is called from dsl_scan_sync() instead. This implies * that metadata scanning and queued I/O issuing are mutually exclusive. This * allows us to provide maximum sequential I/O throughput for the majority of * I/O's issued since sequential I/O performance is significantly negatively * impacted if it is interleaved with random I/O. * * Implementation Notes * * One side effect of the queued scanning algorithm is that the scanning code * needs to be notified whenever a block is freed. This is needed to allow * the scanning code to remove these I/Os from the issuing queue. Additionally, * we do not attempt to queue gang blocks to be issued sequentially since this * is very hard to do and would have an extremely limited performance benefit. * Instead, we simply issue gang I/Os as soon as we find them using the legacy * algorithm. * * Backwards compatibility * * This new algorithm is backwards compatible with the legacy on-disk data * structures (and therefore does not require a new feature flag). * Periodically during scanning (see zfs_scan_checkpoint_intval), the scan * will stop scanning metadata (in logical order) and wait for all outstanding * sorted I/O to complete. Once this is done, we write out a checkpoint * bookmark, indicating that we have scanned everything logically before it. * If the pool is imported on a machine without the new sorting algorithm, * the scan simply resumes from the last checkpoint using the legacy algorithm. */ typedef int (scan_cb_t)(dsl_pool_t *, const blkptr_t *, const zbookmark_phys_t *); static scan_cb_t dsl_scan_scrub_cb; static int scan_ds_queue_compare(const void *a, const void *b); static int scan_prefetch_queue_compare(const void *a, const void *b); static void scan_ds_queue_clear(dsl_scan_t *scn); static void scan_ds_prefetch_queue_clear(dsl_scan_t *scn); static boolean_t scan_ds_queue_contains(dsl_scan_t *scn, uint64_t dsobj, uint64_t *txg); static void scan_ds_queue_insert(dsl_scan_t *scn, uint64_t dsobj, uint64_t txg); static void scan_ds_queue_remove(dsl_scan_t *scn, uint64_t dsobj); static void scan_ds_queue_sync(dsl_scan_t *scn, dmu_tx_t *tx); static uint64_t dsl_scan_count_data_disks(spa_t *spa); +static void read_by_block_level(dsl_scan_t *scn, zbookmark_phys_t zb); extern uint_t zfs_vdev_async_write_active_min_dirty_percent; static int zfs_scan_blkstats = 0; /* * 'zpool status' uses bytes processed per pass to report throughput and * estimate time remaining. We define a pass to start when the scanning * phase completes for a sequential resilver. Optionally, this value * may be used to reset the pass statistics every N txgs to provide an * estimated completion time based on currently observed performance. */ static uint_t zfs_scan_report_txgs = 0; /* * By default zfs will check to ensure it is not over the hard memory * limit before each txg. If finer-grained control of this is needed * this value can be set to 1 to enable checking before scanning each * block. */ static int zfs_scan_strict_mem_lim = B_FALSE; /* * Maximum number of parallelly executed bytes per leaf vdev. We attempt * to strike a balance here between keeping the vdev queues full of I/Os * at all times and not overflowing the queues to cause long latency, * which would cause long txg sync times. No matter what, we will not * overload the drives with I/O, since that is protected by * zfs_vdev_scrub_max_active. */ static uint64_t zfs_scan_vdev_limit = 16 << 20; static uint_t zfs_scan_issue_strategy = 0; /* don't queue & sort zios, go direct */ static int zfs_scan_legacy = B_FALSE; static uint64_t zfs_scan_max_ext_gap = 2 << 20; /* in bytes */ /* * fill_weight is non-tunable at runtime, so we copy it at module init from * zfs_scan_fill_weight. Runtime adjustments to zfs_scan_fill_weight would * break queue sorting. */ static uint_t zfs_scan_fill_weight = 3; static uint64_t fill_weight; /* See dsl_scan_should_clear() for details on the memory limit tunables */ static const uint64_t zfs_scan_mem_lim_min = 16 << 20; /* bytes */ static const uint64_t zfs_scan_mem_lim_soft_max = 128 << 20; /* bytes */ /* fraction of physmem */ static uint_t zfs_scan_mem_lim_fact = 20; /* fraction of mem lim above */ static uint_t zfs_scan_mem_lim_soft_fact = 20; /* minimum milliseconds to scrub per txg */ static uint_t zfs_scrub_min_time_ms = 1000; /* minimum milliseconds to obsolete per txg */ static uint_t zfs_obsolete_min_time_ms = 500; /* minimum milliseconds to free per txg */ static uint_t zfs_free_min_time_ms = 1000; /* minimum milliseconds to resilver per txg */ static uint_t zfs_resilver_min_time_ms = 3000; static uint_t zfs_scan_checkpoint_intval = 7200; /* in seconds */ int zfs_scan_suspend_progress = 0; /* set to prevent scans from progressing */ static int zfs_no_scrub_io = B_FALSE; /* set to disable scrub i/o */ static int zfs_no_scrub_prefetch = B_FALSE; /* set to disable scrub prefetch */ static const enum ddt_class zfs_scrub_ddt_class_max = DDT_CLASS_DUPLICATE; /* max number of blocks to free in a single TXG */ static uint64_t zfs_async_block_max_blocks = UINT64_MAX; /* max number of dedup blocks to free in a single TXG */ static uint64_t zfs_max_async_dedup_frees = 100000; /* set to disable resilver deferring */ static int zfs_resilver_disable_defer = B_FALSE; /* * We wait a few txgs after importing a pool to begin scanning so that * the import / mounting code isn't held up by scrub / resilver IO. * Unfortunately, it is a bit difficult to determine exactly how long * this will take since userspace will trigger fs mounts asynchronously * and the kernel will create zvol minors asynchronously. As a result, * the value provided here is a bit arbitrary, but represents a * reasonable estimate of how many txgs it will take to finish fully * importing a pool */ #define SCAN_IMPORT_WAIT_TXGS 5 #define DSL_SCAN_IS_SCRUB_RESILVER(scn) \ ((scn)->scn_phys.scn_func == POOL_SCAN_SCRUB || \ (scn)->scn_phys.scn_func == POOL_SCAN_RESILVER) /* * Enable/disable the processing of the free_bpobj object. */ static int zfs_free_bpobj_enabled = 1; +/* Error blocks to be scrubbed in one txg. */ +unsigned long zfs_scrub_error_blocks_per_txg = 1 << 12; + /* the order has to match pool_scan_type */ static scan_cb_t *scan_funcs[POOL_SCAN_FUNCS] = { NULL, dsl_scan_scrub_cb, /* POOL_SCAN_SCRUB */ dsl_scan_scrub_cb, /* POOL_SCAN_RESILVER */ }; /* In core node for the scn->scn_queue. Represents a dataset to be scanned */ typedef struct { uint64_t sds_dsobj; uint64_t sds_txg; avl_node_t sds_node; } scan_ds_t; /* * This controls what conditions are placed on dsl_scan_sync_state(): * SYNC_OPTIONAL) write out scn_phys iff scn_queues_pending == 0 * SYNC_MANDATORY) write out scn_phys always. scn_queues_pending must be 0. * SYNC_CACHED) if scn_queues_pending == 0, write out scn_phys. Otherwise * write out the scn_phys_cached version. * See dsl_scan_sync_state for details. */ typedef enum { SYNC_OPTIONAL, SYNC_MANDATORY, SYNC_CACHED } state_sync_type_t; /* * This struct represents the minimum information needed to reconstruct a * zio for sequential scanning. This is useful because many of these will * accumulate in the sequential IO queues before being issued, so saving * memory matters here. */ typedef struct scan_io { /* fields from blkptr_t */ uint64_t sio_blk_prop; uint64_t sio_phys_birth; uint64_t sio_birth; zio_cksum_t sio_cksum; uint32_t sio_nr_dvas; /* fields from zio_t */ uint32_t sio_flags; zbookmark_phys_t sio_zb; /* members for queue sorting */ union { avl_node_t sio_addr_node; /* link into issuing queue */ list_node_t sio_list_node; /* link for issuing to disk */ } sio_nodes; /* * There may be up to SPA_DVAS_PER_BP DVAs here from the bp, * depending on how many were in the original bp. Only the * first DVA is really used for sorting and issuing purposes. * The other DVAs (if provided) simply exist so that the zio * layer can find additional copies to repair from in the * event of an error. This array must go at the end of the * struct to allow this for the variable number of elements. */ dva_t sio_dva[]; } scan_io_t; #define SIO_SET_OFFSET(sio, x) DVA_SET_OFFSET(&(sio)->sio_dva[0], x) #define SIO_SET_ASIZE(sio, x) DVA_SET_ASIZE(&(sio)->sio_dva[0], x) #define SIO_GET_OFFSET(sio) DVA_GET_OFFSET(&(sio)->sio_dva[0]) #define SIO_GET_ASIZE(sio) DVA_GET_ASIZE(&(sio)->sio_dva[0]) #define SIO_GET_END_OFFSET(sio) \ (SIO_GET_OFFSET(sio) + SIO_GET_ASIZE(sio)) #define SIO_GET_MUSED(sio) \ (sizeof (scan_io_t) + ((sio)->sio_nr_dvas * sizeof (dva_t))) struct dsl_scan_io_queue { dsl_scan_t *q_scn; /* associated dsl_scan_t */ vdev_t *q_vd; /* top-level vdev that this queue represents */ zio_t *q_zio; /* scn_zio_root child for waiting on IO */ /* trees used for sorting I/Os and extents of I/Os */ range_tree_t *q_exts_by_addr; zfs_btree_t q_exts_by_size; avl_tree_t q_sios_by_addr; uint64_t q_sio_memused; uint64_t q_last_ext_addr; /* members for zio rate limiting */ uint64_t q_maxinflight_bytes; uint64_t q_inflight_bytes; kcondvar_t q_zio_cv; /* used under vd->vdev_scan_io_queue_lock */ /* per txg statistics */ uint64_t q_total_seg_size_this_txg; uint64_t q_segs_this_txg; uint64_t q_total_zio_size_this_txg; uint64_t q_zios_this_txg; }; /* private data for dsl_scan_prefetch_cb() */ typedef struct scan_prefetch_ctx { zfs_refcount_t spc_refcnt; /* refcount for memory management */ dsl_scan_t *spc_scn; /* dsl_scan_t for the pool */ boolean_t spc_root; /* is this prefetch for an objset? */ uint8_t spc_indblkshift; /* dn_indblkshift of current dnode */ uint16_t spc_datablkszsec; /* dn_idatablkszsec of current dnode */ } scan_prefetch_ctx_t; /* private data for dsl_scan_prefetch() */ typedef struct scan_prefetch_issue_ctx { avl_node_t spic_avl_node; /* link into scn->scn_prefetch_queue */ scan_prefetch_ctx_t *spic_spc; /* spc for the callback */ blkptr_t spic_bp; /* bp to prefetch */ zbookmark_phys_t spic_zb; /* bookmark to prefetch */ } scan_prefetch_issue_ctx_t; static void scan_exec_io(dsl_pool_t *dp, const blkptr_t *bp, int zio_flags, const zbookmark_phys_t *zb, dsl_scan_io_queue_t *queue); static void scan_io_queue_insert_impl(dsl_scan_io_queue_t *queue, scan_io_t *sio); static dsl_scan_io_queue_t *scan_io_queue_create(vdev_t *vd); static void scan_io_queues_destroy(dsl_scan_t *scn); static kmem_cache_t *sio_cache[SPA_DVAS_PER_BP]; /* sio->sio_nr_dvas must be set so we know which cache to free from */ static void sio_free(scan_io_t *sio) { ASSERT3U(sio->sio_nr_dvas, >, 0); ASSERT3U(sio->sio_nr_dvas, <=, SPA_DVAS_PER_BP); kmem_cache_free(sio_cache[sio->sio_nr_dvas - 1], sio); } /* It is up to the caller to set sio->sio_nr_dvas for freeing */ static scan_io_t * sio_alloc(unsigned short nr_dvas) { ASSERT3U(nr_dvas, >, 0); ASSERT3U(nr_dvas, <=, SPA_DVAS_PER_BP); return (kmem_cache_alloc(sio_cache[nr_dvas - 1], KM_SLEEP)); } void scan_init(void) { /* * This is used in ext_size_compare() to weight segments * based on how sparse they are. This cannot be changed * mid-scan and the tree comparison functions don't currently * have a mechanism for passing additional context to the * compare functions. Thus we store this value globally and * we only allow it to be set at module initialization time */ fill_weight = zfs_scan_fill_weight; for (int i = 0; i < SPA_DVAS_PER_BP; i++) { char name[36]; (void) snprintf(name, sizeof (name), "sio_cache_%d", i); sio_cache[i] = kmem_cache_create(name, (sizeof (scan_io_t) + ((i + 1) * sizeof (dva_t))), 0, NULL, NULL, NULL, NULL, NULL, 0); } } void scan_fini(void) { for (int i = 0; i < SPA_DVAS_PER_BP; i++) { kmem_cache_destroy(sio_cache[i]); } } static inline boolean_t dsl_scan_is_running(const dsl_scan_t *scn) { return (scn->scn_phys.scn_state == DSS_SCANNING); } boolean_t dsl_scan_resilvering(dsl_pool_t *dp) { return (dsl_scan_is_running(dp->dp_scan) && dp->dp_scan->scn_phys.scn_func == POOL_SCAN_RESILVER); } static inline void sio2bp(const scan_io_t *sio, blkptr_t *bp) { memset(bp, 0, sizeof (*bp)); bp->blk_prop = sio->sio_blk_prop; bp->blk_phys_birth = sio->sio_phys_birth; bp->blk_birth = sio->sio_birth; bp->blk_fill = 1; /* we always only work with data pointers */ bp->blk_cksum = sio->sio_cksum; ASSERT3U(sio->sio_nr_dvas, >, 0); ASSERT3U(sio->sio_nr_dvas, <=, SPA_DVAS_PER_BP); memcpy(bp->blk_dva, sio->sio_dva, sio->sio_nr_dvas * sizeof (dva_t)); } static inline void bp2sio(const blkptr_t *bp, scan_io_t *sio, int dva_i) { sio->sio_blk_prop = bp->blk_prop; sio->sio_phys_birth = bp->blk_phys_birth; sio->sio_birth = bp->blk_birth; sio->sio_cksum = bp->blk_cksum; sio->sio_nr_dvas = BP_GET_NDVAS(bp); /* * Copy the DVAs to the sio. We need all copies of the block so * that the self healing code can use the alternate copies if the * first is corrupted. We want the DVA at index dva_i to be first * in the sio since this is the primary one that we want to issue. */ for (int i = 0, j = dva_i; i < sio->sio_nr_dvas; i++, j++) { sio->sio_dva[i] = bp->blk_dva[j % sio->sio_nr_dvas]; } } int dsl_scan_init(dsl_pool_t *dp, uint64_t txg) { int err; dsl_scan_t *scn; spa_t *spa = dp->dp_spa; uint64_t f; scn = dp->dp_scan = kmem_zalloc(sizeof (dsl_scan_t), KM_SLEEP); scn->scn_dp = dp; /* * It's possible that we're resuming a scan after a reboot so * make sure that the scan_async_destroying flag is initialized * appropriately. */ ASSERT(!scn->scn_async_destroying); scn->scn_async_destroying = spa_feature_is_active(dp->dp_spa, SPA_FEATURE_ASYNC_DESTROY); /* * Calculate the max number of in-flight bytes for pool-wide * scanning operations (minimum 1MB, maximum 1/4 of arc_c_max). * Limits for the issuing phase are done per top-level vdev and * are handled separately. */ scn->scn_maxinflight_bytes = MIN(arc_c_max / 4, MAX(1ULL << 20, zfs_scan_vdev_limit * dsl_scan_count_data_disks(spa))); avl_create(&scn->scn_queue, scan_ds_queue_compare, sizeof (scan_ds_t), offsetof(scan_ds_t, sds_node)); avl_create(&scn->scn_prefetch_queue, scan_prefetch_queue_compare, sizeof (scan_prefetch_issue_ctx_t), offsetof(scan_prefetch_issue_ctx_t, spic_avl_node)); err = zap_lookup(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, "scrub_func", sizeof (uint64_t), 1, &f); if (err == 0) { /* * There was an old-style scrub in progress. Restart a * new-style scrub from the beginning. */ scn->scn_restart_txg = txg; zfs_dbgmsg("old-style scrub was in progress for %s; " "restarting new-style scrub in txg %llu", spa->spa_name, (longlong_t)scn->scn_restart_txg); /* * Load the queue obj from the old location so that it * can be freed by dsl_scan_done(). */ (void) zap_lookup(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, "scrub_queue", sizeof (uint64_t), 1, &scn->scn_phys.scn_queue_obj); } else { + err = zap_lookup(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, + DMU_POOL_ERRORSCRUB, sizeof (uint64_t), + ERRORSCRUB_PHYS_NUMINTS, &scn->errorscrub_phys); + + if (err != 0 && err != ENOENT) + return (err); + err = zap_lookup(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_SCAN, sizeof (uint64_t), SCAN_PHYS_NUMINTS, &scn->scn_phys); + /* * Detect if the pool contains the signature of #2094. If it * does properly update the scn->scn_phys structure and notify * the administrator by setting an errata for the pool. */ if (err == EOVERFLOW) { uint64_t zaptmp[SCAN_PHYS_NUMINTS + 1]; VERIFY3S(SCAN_PHYS_NUMINTS, ==, 24); VERIFY3S(offsetof(dsl_scan_phys_t, scn_flags), ==, (23 * sizeof (uint64_t))); err = zap_lookup(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_SCAN, sizeof (uint64_t), SCAN_PHYS_NUMINTS + 1, &zaptmp); if (err == 0) { uint64_t overflow = zaptmp[SCAN_PHYS_NUMINTS]; if (overflow & ~DSL_SCAN_FLAGS_MASK || scn->scn_async_destroying) { spa->spa_errata = ZPOOL_ERRATA_ZOL_2094_ASYNC_DESTROY; return (EOVERFLOW); } memcpy(&scn->scn_phys, zaptmp, SCAN_PHYS_NUMINTS * sizeof (uint64_t)); scn->scn_phys.scn_flags = overflow; /* Required scrub already in progress. */ if (scn->scn_phys.scn_state == DSS_FINISHED || scn->scn_phys.scn_state == DSS_CANCELED) spa->spa_errata = ZPOOL_ERRATA_ZOL_2094_SCRUB; } } if (err == ENOENT) return (0); else if (err) return (err); /* * We might be restarting after a reboot, so jump the issued * counter to how far we've scanned. We know we're consistent * up to here. */ scn->scn_issued_before_pass = scn->scn_phys.scn_examined; if (dsl_scan_is_running(scn) && spa_prev_software_version(dp->dp_spa) < SPA_VERSION_SCAN) { /* * A new-type scrub was in progress on an old * pool, and the pool was accessed by old * software. Restart from the beginning, since * the old software may have changed the pool in * the meantime. */ scn->scn_restart_txg = txg; zfs_dbgmsg("new-style scrub for %s was modified " "by old software; restarting in txg %llu", spa->spa_name, (longlong_t)scn->scn_restart_txg); } else if (dsl_scan_resilvering(dp)) { /* * If a resilver is in progress and there are already * errors, restart it instead of finishing this scan and * then restarting it. If there haven't been any errors * then remember that the incore DTL is valid. */ if (scn->scn_phys.scn_errors > 0) { scn->scn_restart_txg = txg; zfs_dbgmsg("resilver can't excise DTL_MISSING " "when finished; restarting on %s in txg " "%llu", spa->spa_name, (u_longlong_t)scn->scn_restart_txg); } else { /* it's safe to excise DTL when finished */ spa->spa_scrub_started = B_TRUE; } } } memcpy(&scn->scn_phys_cached, &scn->scn_phys, sizeof (scn->scn_phys)); /* reload the queue into the in-core state */ if (scn->scn_phys.scn_queue_obj != 0) { zap_cursor_t zc; zap_attribute_t za; for (zap_cursor_init(&zc, dp->dp_meta_objset, scn->scn_phys.scn_queue_obj); zap_cursor_retrieve(&zc, &za) == 0; (void) zap_cursor_advance(&zc)) { scan_ds_queue_insert(scn, zfs_strtonum(za.za_name, NULL), za.za_first_integer); } zap_cursor_fini(&zc); } spa_scan_stat_init(spa); vdev_scan_stat_init(spa->spa_root_vdev); return (0); } void dsl_scan_fini(dsl_pool_t *dp) { if (dp->dp_scan != NULL) { dsl_scan_t *scn = dp->dp_scan; if (scn->scn_taskq != NULL) taskq_destroy(scn->scn_taskq); scan_ds_queue_clear(scn); avl_destroy(&scn->scn_queue); scan_ds_prefetch_queue_clear(scn); avl_destroy(&scn->scn_prefetch_queue); kmem_free(dp->dp_scan, sizeof (dsl_scan_t)); dp->dp_scan = NULL; } } static boolean_t dsl_scan_restarting(dsl_scan_t *scn, dmu_tx_t *tx) { return (scn->scn_restart_txg != 0 && scn->scn_restart_txg <= tx->tx_txg); } boolean_t dsl_scan_resilver_scheduled(dsl_pool_t *dp) { return ((dp->dp_scan && dp->dp_scan->scn_restart_txg != 0) || (spa_async_tasks(dp->dp_spa) & SPA_ASYNC_RESILVER)); } boolean_t dsl_scan_scrubbing(const dsl_pool_t *dp) { dsl_scan_phys_t *scn_phys = &dp->dp_scan->scn_phys; return (scn_phys->scn_state == DSS_SCANNING && scn_phys->scn_func == POOL_SCAN_SCRUB); } +boolean_t +dsl_errorscrubbing(const dsl_pool_t *dp) +{ + dsl_errorscrub_phys_t *errorscrub_phys = &dp->dp_scan->errorscrub_phys; + + return (errorscrub_phys->dep_state == DSS_ERRORSCRUBBING && + errorscrub_phys->dep_func == POOL_SCAN_ERRORSCRUB); +} + +boolean_t +dsl_errorscrub_is_paused(const dsl_scan_t *scn) +{ + return (dsl_errorscrubbing(scn->scn_dp) && + scn->errorscrub_phys.dep_paused_flags); +} + boolean_t dsl_scan_is_paused_scrub(const dsl_scan_t *scn) { return (dsl_scan_scrubbing(scn->scn_dp) && scn->scn_phys.scn_flags & DSF_SCRUB_PAUSED); } +static void +dsl_errorscrub_sync_state(dsl_scan_t *scn, dmu_tx_t *tx) +{ + scn->errorscrub_phys.dep_cursor = + zap_cursor_serialize(&scn->errorscrub_cursor); + + VERIFY0(zap_update(scn->scn_dp->dp_meta_objset, + DMU_POOL_DIRECTORY_OBJECT, + DMU_POOL_ERRORSCRUB, sizeof (uint64_t), ERRORSCRUB_PHYS_NUMINTS, + &scn->errorscrub_phys, tx)); +} + +static void +dsl_errorscrub_setup_sync(void *arg, dmu_tx_t *tx) +{ + dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan; + pool_scan_func_t *funcp = arg; + dsl_pool_t *dp = scn->scn_dp; + spa_t *spa = dp->dp_spa; + + ASSERT(!dsl_scan_is_running(scn)); + ASSERT(!dsl_errorscrubbing(scn->scn_dp)); + ASSERT(*funcp > POOL_SCAN_NONE && *funcp < POOL_SCAN_FUNCS); + + memset(&scn->errorscrub_phys, 0, sizeof (scn->errorscrub_phys)); + scn->errorscrub_phys.dep_func = *funcp; + scn->errorscrub_phys.dep_state = DSS_ERRORSCRUBBING; + scn->errorscrub_phys.dep_start_time = gethrestime_sec(); + scn->errorscrub_phys.dep_to_examine = spa_get_last_errlog_size(spa); + scn->errorscrub_phys.dep_examined = 0; + scn->errorscrub_phys.dep_errors = 0; + scn->errorscrub_phys.dep_cursor = 0; + zap_cursor_init_serialized(&scn->errorscrub_cursor, + spa->spa_meta_objset, spa->spa_errlog_last, + scn->errorscrub_phys.dep_cursor); + + vdev_config_dirty(spa->spa_root_vdev); + spa_event_notify(spa, NULL, NULL, ESC_ZFS_ERRORSCRUB_START); + + dsl_errorscrub_sync_state(scn, tx); + + spa_history_log_internal(spa, "error scrub setup", tx, + "func=%u mintxg=%u maxtxg=%llu", + *funcp, 0, (u_longlong_t)tx->tx_txg); +} + +static int +dsl_errorscrub_setup_check(void *arg, dmu_tx_t *tx) +{ + (void) arg; + dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan; + + if (dsl_scan_is_running(scn) || (dsl_errorscrubbing(scn->scn_dp))) { + return (SET_ERROR(EBUSY)); + } + + if (spa_get_last_errlog_size(scn->scn_dp->dp_spa) == 0) { + return (ECANCELED); + } + return (0); +} + /* * Writes out a persistent dsl_scan_phys_t record to the pool directory. * Because we can be running in the block sorting algorithm, we do not always * want to write out the record, only when it is "safe" to do so. This safety * condition is achieved by making sure that the sorting queues are empty * (scn_queues_pending == 0). When this condition is not true, the sync'd state * is inconsistent with how much actual scanning progress has been made. The * kind of sync to be performed is specified by the sync_type argument. If the * sync is optional, we only sync if the queues are empty. If the sync is * mandatory, we do a hard ASSERT to make sure that the queues are empty. The * third possible state is a "cached" sync. This is done in response to: * 1) The dataset that was in the last sync'd dsl_scan_phys_t having been * destroyed, so we wouldn't be able to restart scanning from it. * 2) The snapshot that was in the last sync'd dsl_scan_phys_t having been * superseded by a newer snapshot. * 3) The dataset that was in the last sync'd dsl_scan_phys_t having been * swapped with its clone. * In all cases, a cached sync simply rewrites the last record we've written, * just slightly modified. For the modifications that are performed to the * last written dsl_scan_phys_t, see dsl_scan_ds_destroyed, * dsl_scan_ds_snapshotted and dsl_scan_ds_clone_swapped. */ static void dsl_scan_sync_state(dsl_scan_t *scn, dmu_tx_t *tx, state_sync_type_t sync_type) { int i; spa_t *spa = scn->scn_dp->dp_spa; ASSERT(sync_type != SYNC_MANDATORY || scn->scn_queues_pending == 0); if (scn->scn_queues_pending == 0) { for (i = 0; i < spa->spa_root_vdev->vdev_children; i++) { vdev_t *vd = spa->spa_root_vdev->vdev_child[i]; dsl_scan_io_queue_t *q = vd->vdev_scan_io_queue; if (q == NULL) continue; mutex_enter(&vd->vdev_scan_io_queue_lock); ASSERT3P(avl_first(&q->q_sios_by_addr), ==, NULL); ASSERT3P(zfs_btree_first(&q->q_exts_by_size, NULL), ==, NULL); ASSERT3P(range_tree_first(q->q_exts_by_addr), ==, NULL); mutex_exit(&vd->vdev_scan_io_queue_lock); } if (scn->scn_phys.scn_queue_obj != 0) scan_ds_queue_sync(scn, tx); VERIFY0(zap_update(scn->scn_dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_SCAN, sizeof (uint64_t), SCAN_PHYS_NUMINTS, &scn->scn_phys, tx)); memcpy(&scn->scn_phys_cached, &scn->scn_phys, sizeof (scn->scn_phys)); if (scn->scn_checkpointing) zfs_dbgmsg("finish scan checkpoint for %s", spa->spa_name); scn->scn_checkpointing = B_FALSE; scn->scn_last_checkpoint = ddi_get_lbolt(); } else if (sync_type == SYNC_CACHED) { VERIFY0(zap_update(scn->scn_dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_SCAN, sizeof (uint64_t), SCAN_PHYS_NUMINTS, &scn->scn_phys_cached, tx)); } } int dsl_scan_setup_check(void *arg, dmu_tx_t *tx) { (void) arg; dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan; vdev_t *rvd = scn->scn_dp->dp_spa->spa_root_vdev; - if (dsl_scan_is_running(scn) || vdev_rebuild_active(rvd)) + if (dsl_scan_is_running(scn) || vdev_rebuild_active(rvd) || + dsl_errorscrubbing(scn->scn_dp)) return (SET_ERROR(EBUSY)); return (0); } void dsl_scan_setup_sync(void *arg, dmu_tx_t *tx) { + (void) arg; dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan; pool_scan_func_t *funcp = arg; dmu_object_type_t ot = 0; dsl_pool_t *dp = scn->scn_dp; spa_t *spa = dp->dp_spa; ASSERT(!dsl_scan_is_running(scn)); ASSERT(*funcp > POOL_SCAN_NONE && *funcp < POOL_SCAN_FUNCS); memset(&scn->scn_phys, 0, sizeof (scn->scn_phys)); + + /* + * If we are starting a fresh scrub, we erase the error scrub + * information from disk. + */ + memset(&scn->errorscrub_phys, 0, sizeof (scn->errorscrub_phys)); + dsl_errorscrub_sync_state(scn, tx); + scn->scn_phys.scn_func = *funcp; scn->scn_phys.scn_state = DSS_SCANNING; scn->scn_phys.scn_min_txg = 0; scn->scn_phys.scn_max_txg = tx->tx_txg; scn->scn_phys.scn_ddt_class_max = DDT_CLASSES - 1; /* the entire DDT */ scn->scn_phys.scn_start_time = gethrestime_sec(); scn->scn_phys.scn_errors = 0; scn->scn_phys.scn_to_examine = spa->spa_root_vdev->vdev_stat.vs_alloc; scn->scn_issued_before_pass = 0; scn->scn_restart_txg = 0; scn->scn_done_txg = 0; scn->scn_last_checkpoint = 0; scn->scn_checkpointing = B_FALSE; spa_scan_stat_init(spa); vdev_scan_stat_init(spa->spa_root_vdev); if (DSL_SCAN_IS_SCRUB_RESILVER(scn)) { scn->scn_phys.scn_ddt_class_max = zfs_scrub_ddt_class_max; /* rewrite all disk labels */ vdev_config_dirty(spa->spa_root_vdev); if (vdev_resilver_needed(spa->spa_root_vdev, &scn->scn_phys.scn_min_txg, &scn->scn_phys.scn_max_txg)) { nvlist_t *aux = fnvlist_alloc(); fnvlist_add_string(aux, ZFS_EV_RESILVER_TYPE, "healing"); spa_event_notify(spa, NULL, aux, ESC_ZFS_RESILVER_START); nvlist_free(aux); } else { spa_event_notify(spa, NULL, NULL, ESC_ZFS_SCRUB_START); } spa->spa_scrub_started = B_TRUE; /* * If this is an incremental scrub, limit the DDT scrub phase * to just the auto-ditto class (for correctness); the rest * of the scrub should go faster using top-down pruning. */ if (scn->scn_phys.scn_min_txg > TXG_INITIAL) scn->scn_phys.scn_ddt_class_max = DDT_CLASS_DITTO; /* * When starting a resilver clear any existing rebuild state. * This is required to prevent stale rebuild status from * being reported when a rebuild is run, then a resilver and * finally a scrub. In which case only the scrub status * should be reported by 'zpool status'. */ if (scn->scn_phys.scn_func == POOL_SCAN_RESILVER) { vdev_t *rvd = spa->spa_root_vdev; for (uint64_t i = 0; i < rvd->vdev_children; i++) { vdev_t *vd = rvd->vdev_child[i]; vdev_rebuild_clear_sync( (void *)(uintptr_t)vd->vdev_id, tx); } } } /* back to the generic stuff */ if (zfs_scan_blkstats) { if (dp->dp_blkstats == NULL) { dp->dp_blkstats = vmem_alloc(sizeof (zfs_all_blkstats_t), KM_SLEEP); } memset(&dp->dp_blkstats->zab_type, 0, sizeof (dp->dp_blkstats->zab_type)); } else { if (dp->dp_blkstats) { vmem_free(dp->dp_blkstats, sizeof (zfs_all_blkstats_t)); dp->dp_blkstats = NULL; } } if (spa_version(spa) < SPA_VERSION_DSL_SCRUB) ot = DMU_OT_ZAP_OTHER; scn->scn_phys.scn_queue_obj = zap_create(dp->dp_meta_objset, ot ? ot : DMU_OT_SCAN_QUEUE, DMU_OT_NONE, 0, tx); memcpy(&scn->scn_phys_cached, &scn->scn_phys, sizeof (scn->scn_phys)); dsl_scan_sync_state(scn, tx, SYNC_MANDATORY); spa_history_log_internal(spa, "scan setup", tx, "func=%u mintxg=%llu maxtxg=%llu", *funcp, (u_longlong_t)scn->scn_phys.scn_min_txg, (u_longlong_t)scn->scn_phys.scn_max_txg); } /* - * Called by the ZFS_IOC_POOL_SCAN ioctl to start a scrub or resilver. - * Can also be called to resume a paused scrub. + * Called by ZFS_IOC_POOL_SCRUB and ZFS_IOC_POOL_SCAN ioctl to start a scrub, + * error scrub or resilver. Can also be called to resume a paused scrub or + * error scrub. */ int dsl_scan(dsl_pool_t *dp, pool_scan_func_t func) { spa_t *spa = dp->dp_spa; dsl_scan_t *scn = dp->dp_scan; /* * Purge all vdev caches and probe all devices. We do this here * rather than in sync context because this requires a writer lock * on the spa_config lock, which we can't do from sync context. The * spa_scrub_reopen flag indicates that vdev_open() should not * attempt to start another scrub. */ spa_vdev_state_enter(spa, SCL_NONE); spa->spa_scrub_reopen = B_TRUE; vdev_reopen(spa->spa_root_vdev); spa->spa_scrub_reopen = B_FALSE; (void) spa_vdev_state_exit(spa, NULL, 0); if (func == POOL_SCAN_RESILVER) { dsl_scan_restart_resilver(spa->spa_dsl_pool, 0); return (0); } + if (func == POOL_SCAN_ERRORSCRUB) { + if (dsl_errorscrub_is_paused(dp->dp_scan)) { + /* + * got error scrub start cmd, resume paused error scrub. + */ + int err = dsl_scrub_set_pause_resume(scn->scn_dp, + POOL_SCRUB_NORMAL); + if (err == 0) { + spa_event_notify(spa, NULL, NULL, + ESC_ZFS_ERRORSCRUB_RESUME); + return (ECANCELED); + } + return (SET_ERROR(err)); + } + + return (dsl_sync_task(spa_name(dp->dp_spa), + dsl_errorscrub_setup_check, dsl_errorscrub_setup_sync, + &func, 0, ZFS_SPACE_CHECK_RESERVED)); + } + if (func == POOL_SCAN_SCRUB && dsl_scan_is_paused_scrub(scn)) { /* got scrub start cmd, resume paused scrub */ int err = dsl_scrub_set_pause_resume(scn->scn_dp, POOL_SCRUB_NORMAL); if (err == 0) { spa_event_notify(spa, NULL, NULL, ESC_ZFS_SCRUB_RESUME); return (SET_ERROR(ECANCELED)); } - return (SET_ERROR(err)); } return (dsl_sync_task(spa_name(spa), dsl_scan_setup_check, dsl_scan_setup_sync, &func, 0, ZFS_SPACE_CHECK_EXTRA_RESERVED)); } +static void +dsl_errorscrub_done(dsl_scan_t *scn, boolean_t complete, dmu_tx_t *tx) +{ + dsl_pool_t *dp = scn->scn_dp; + spa_t *spa = dp->dp_spa; + + if (complete) { + spa_event_notify(spa, NULL, NULL, ESC_ZFS_ERRORSCRUB_FINISH); + spa_history_log_internal(spa, "error scrub done", tx, + "errors=%llu", (u_longlong_t)spa_approx_errlog_size(spa)); + } else { + spa_history_log_internal(spa, "error scrub canceled", tx, + "errors=%llu", (u_longlong_t)spa_approx_errlog_size(spa)); + } + + scn->errorscrub_phys.dep_state = complete ? DSS_FINISHED : DSS_CANCELED; + spa->spa_scrub_active = B_FALSE; + spa_errlog_rotate(spa); + scn->errorscrub_phys.dep_end_time = gethrestime_sec(); + zap_cursor_fini(&scn->errorscrub_cursor); + + if (spa->spa_errata == ZPOOL_ERRATA_ZOL_2094_SCRUB) + spa->spa_errata = 0; + + ASSERT(!dsl_errorscrubbing(scn->scn_dp)); +} + static void dsl_scan_done(dsl_scan_t *scn, boolean_t complete, dmu_tx_t *tx) { static const char *old_names[] = { "scrub_bookmark", "scrub_ddt_bookmark", "scrub_ddt_class_max", "scrub_queue", "scrub_min_txg", "scrub_max_txg", "scrub_func", "scrub_errors", NULL }; dsl_pool_t *dp = scn->scn_dp; spa_t *spa = dp->dp_spa; int i; /* Remove any remnants of an old-style scrub. */ for (i = 0; old_names[i]; i++) { (void) zap_remove(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, old_names[i], tx); } if (scn->scn_phys.scn_queue_obj != 0) { VERIFY0(dmu_object_free(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, tx)); scn->scn_phys.scn_queue_obj = 0; } scan_ds_queue_clear(scn); scan_ds_prefetch_queue_clear(scn); scn->scn_phys.scn_flags &= ~DSF_SCRUB_PAUSED; /* * If we were "restarted" from a stopped state, don't bother * with anything else. */ if (!dsl_scan_is_running(scn)) { ASSERT(!scn->scn_is_sorted); return; } if (scn->scn_is_sorted) { scan_io_queues_destroy(scn); scn->scn_is_sorted = B_FALSE; if (scn->scn_taskq != NULL) { taskq_destroy(scn->scn_taskq); scn->scn_taskq = NULL; } } scn->scn_phys.scn_state = complete ? DSS_FINISHED : DSS_CANCELED; spa_notify_waiters(spa); if (dsl_scan_restarting(scn, tx)) spa_history_log_internal(spa, "scan aborted, restarting", tx, "errors=%llu", (u_longlong_t)spa_approx_errlog_size(spa)); else if (!complete) spa_history_log_internal(spa, "scan cancelled", tx, "errors=%llu", (u_longlong_t)spa_approx_errlog_size(spa)); else spa_history_log_internal(spa, "scan done", tx, "errors=%llu", (u_longlong_t)spa_approx_errlog_size(spa)); if (DSL_SCAN_IS_SCRUB_RESILVER(scn)) { spa->spa_scrub_active = B_FALSE; /* * If the scrub/resilver completed, update all DTLs to * reflect this. Whether it succeeded or not, vacate * all temporary scrub DTLs. * * As the scrub does not currently support traversing * data that have been freed but are part of a checkpoint, * we don't mark the scrub as done in the DTLs as faults * may still exist in those vdevs. */ if (complete && !spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT)) { vdev_dtl_reassess(spa->spa_root_vdev, tx->tx_txg, scn->scn_phys.scn_max_txg, B_TRUE, B_FALSE); if (scn->scn_phys.scn_min_txg) { nvlist_t *aux = fnvlist_alloc(); fnvlist_add_string(aux, ZFS_EV_RESILVER_TYPE, "healing"); spa_event_notify(spa, NULL, aux, ESC_ZFS_RESILVER_FINISH); nvlist_free(aux); } else { spa_event_notify(spa, NULL, NULL, ESC_ZFS_SCRUB_FINISH); } } else { vdev_dtl_reassess(spa->spa_root_vdev, tx->tx_txg, 0, B_TRUE, B_FALSE); } spa_errlog_rotate(spa); /* * Don't clear flag until after vdev_dtl_reassess to ensure that * DTL_MISSING will get updated when possible. */ spa->spa_scrub_started = B_FALSE; /* * We may have finished replacing a device. * Let the async thread assess this and handle the detach. */ spa_async_request(spa, SPA_ASYNC_RESILVER_DONE); /* * Clear any resilver_deferred flags in the config. * If there are drives that need resilvering, kick * off an asynchronous request to start resilver. * vdev_clear_resilver_deferred() may update the config * before the resilver can restart. In the event of * a crash during this period, the spa loading code * will find the drives that need to be resilvered * and start the resilver then. */ if (spa_feature_is_enabled(spa, SPA_FEATURE_RESILVER_DEFER) && vdev_clear_resilver_deferred(spa->spa_root_vdev, tx)) { spa_history_log_internal(spa, "starting deferred resilver", tx, "errors=%llu", (u_longlong_t)spa_approx_errlog_size(spa)); spa_async_request(spa, SPA_ASYNC_RESILVER); } /* Clear recent error events (i.e. duplicate events tracking) */ if (complete) zfs_ereport_clear(spa, NULL); } scn->scn_phys.scn_end_time = gethrestime_sec(); if (spa->spa_errata == ZPOOL_ERRATA_ZOL_2094_SCRUB) spa->spa_errata = 0; ASSERT(!dsl_scan_is_running(scn)); } +static int +dsl_errorscrub_pause_resume_check(void *arg, dmu_tx_t *tx) +{ + pool_scrub_cmd_t *cmd = arg; + dsl_pool_t *dp = dmu_tx_pool(tx); + dsl_scan_t *scn = dp->dp_scan; + + if (*cmd == POOL_SCRUB_PAUSE) { + /* + * can't pause a error scrub when there is no in-progress + * error scrub. + */ + if (!dsl_errorscrubbing(dp)) + return (SET_ERROR(ENOENT)); + + /* can't pause a paused error scrub */ + if (dsl_errorscrub_is_paused(scn)) + return (SET_ERROR(EBUSY)); + } else if (*cmd != POOL_SCRUB_NORMAL) { + return (SET_ERROR(ENOTSUP)); + } + + return (0); +} + +static void +dsl_errorscrub_pause_resume_sync(void *arg, dmu_tx_t *tx) +{ + pool_scrub_cmd_t *cmd = arg; + dsl_pool_t *dp = dmu_tx_pool(tx); + spa_t *spa = dp->dp_spa; + dsl_scan_t *scn = dp->dp_scan; + + if (*cmd == POOL_SCRUB_PAUSE) { + spa->spa_scan_pass_errorscrub_pause = gethrestime_sec(); + scn->errorscrub_phys.dep_paused_flags = B_TRUE; + dsl_errorscrub_sync_state(scn, tx); + spa_event_notify(spa, NULL, NULL, ESC_ZFS_ERRORSCRUB_PAUSED); + } else { + ASSERT3U(*cmd, ==, POOL_SCRUB_NORMAL); + if (dsl_errorscrub_is_paused(scn)) { + /* + * We need to keep track of how much time we spend + * paused per pass so that we can adjust the error scrub + * rate shown in the output of 'zpool status'. + */ + spa->spa_scan_pass_errorscrub_spent_paused += + gethrestime_sec() - + spa->spa_scan_pass_errorscrub_pause; + + spa->spa_scan_pass_errorscrub_pause = 0; + scn->errorscrub_phys.dep_paused_flags = B_FALSE; + + zap_cursor_init_serialized( + &scn->errorscrub_cursor, + spa->spa_meta_objset, spa->spa_errlog_last, + scn->errorscrub_phys.dep_cursor); + + dsl_errorscrub_sync_state(scn, tx); + } + } +} + +static int +dsl_errorscrub_cancel_check(void *arg, dmu_tx_t *tx) +{ + (void) arg; + dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan; + /* can't cancel a error scrub when there is no one in-progress */ + if (!dsl_errorscrubbing(scn->scn_dp)) + return (SET_ERROR(ENOENT)); + return (0); +} + +static void +dsl_errorscrub_cancel_sync(void *arg, dmu_tx_t *tx) +{ + (void) arg; + dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan; + + dsl_errorscrub_done(scn, B_FALSE, tx); + dsl_errorscrub_sync_state(scn, tx); + spa_event_notify(scn->scn_dp->dp_spa, NULL, NULL, + ESC_ZFS_ERRORSCRUB_ABORT); +} + static int dsl_scan_cancel_check(void *arg, dmu_tx_t *tx) { (void) arg; dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan; if (!dsl_scan_is_running(scn)) return (SET_ERROR(ENOENT)); return (0); } static void dsl_scan_cancel_sync(void *arg, dmu_tx_t *tx) { (void) arg; dsl_scan_t *scn = dmu_tx_pool(tx)->dp_scan; dsl_scan_done(scn, B_FALSE, tx); dsl_scan_sync_state(scn, tx, SYNC_MANDATORY); spa_event_notify(scn->scn_dp->dp_spa, NULL, NULL, ESC_ZFS_SCRUB_ABORT); } int dsl_scan_cancel(dsl_pool_t *dp) { + if (dsl_errorscrubbing(dp)) { + return (dsl_sync_task(spa_name(dp->dp_spa), + dsl_errorscrub_cancel_check, dsl_errorscrub_cancel_sync, + NULL, 3, ZFS_SPACE_CHECK_RESERVED)); + } return (dsl_sync_task(spa_name(dp->dp_spa), dsl_scan_cancel_check, dsl_scan_cancel_sync, NULL, 3, ZFS_SPACE_CHECK_RESERVED)); } static int dsl_scrub_pause_resume_check(void *arg, dmu_tx_t *tx) { pool_scrub_cmd_t *cmd = arg; dsl_pool_t *dp = dmu_tx_pool(tx); dsl_scan_t *scn = dp->dp_scan; if (*cmd == POOL_SCRUB_PAUSE) { /* can't pause a scrub when there is no in-progress scrub */ if (!dsl_scan_scrubbing(dp)) return (SET_ERROR(ENOENT)); /* can't pause a paused scrub */ if (dsl_scan_is_paused_scrub(scn)) return (SET_ERROR(EBUSY)); } else if (*cmd != POOL_SCRUB_NORMAL) { return (SET_ERROR(ENOTSUP)); } return (0); } static void dsl_scrub_pause_resume_sync(void *arg, dmu_tx_t *tx) { pool_scrub_cmd_t *cmd = arg; dsl_pool_t *dp = dmu_tx_pool(tx); spa_t *spa = dp->dp_spa; dsl_scan_t *scn = dp->dp_scan; if (*cmd == POOL_SCRUB_PAUSE) { /* can't pause a scrub when there is no in-progress scrub */ spa->spa_scan_pass_scrub_pause = gethrestime_sec(); scn->scn_phys.scn_flags |= DSF_SCRUB_PAUSED; scn->scn_phys_cached.scn_flags |= DSF_SCRUB_PAUSED; dsl_scan_sync_state(scn, tx, SYNC_CACHED); spa_event_notify(spa, NULL, NULL, ESC_ZFS_SCRUB_PAUSED); spa_notify_waiters(spa); } else { ASSERT3U(*cmd, ==, POOL_SCRUB_NORMAL); if (dsl_scan_is_paused_scrub(scn)) { /* * We need to keep track of how much time we spend * paused per pass so that we can adjust the scrub rate * shown in the output of 'zpool status' */ spa->spa_scan_pass_scrub_spent_paused += gethrestime_sec() - spa->spa_scan_pass_scrub_pause; spa->spa_scan_pass_scrub_pause = 0; scn->scn_phys.scn_flags &= ~DSF_SCRUB_PAUSED; scn->scn_phys_cached.scn_flags &= ~DSF_SCRUB_PAUSED; dsl_scan_sync_state(scn, tx, SYNC_CACHED); } } } /* * Set scrub pause/resume state if it makes sense to do so */ int dsl_scrub_set_pause_resume(const dsl_pool_t *dp, pool_scrub_cmd_t cmd) { + if (dsl_errorscrubbing(dp)) { + return (dsl_sync_task(spa_name(dp->dp_spa), + dsl_errorscrub_pause_resume_check, + dsl_errorscrub_pause_resume_sync, &cmd, 3, + ZFS_SPACE_CHECK_RESERVED)); + } return (dsl_sync_task(spa_name(dp->dp_spa), dsl_scrub_pause_resume_check, dsl_scrub_pause_resume_sync, &cmd, 3, ZFS_SPACE_CHECK_RESERVED)); } /* start a new scan, or restart an existing one. */ void dsl_scan_restart_resilver(dsl_pool_t *dp, uint64_t txg) { if (txg == 0) { dmu_tx_t *tx; tx = dmu_tx_create_dd(dp->dp_mos_dir); VERIFY(0 == dmu_tx_assign(tx, TXG_WAIT)); txg = dmu_tx_get_txg(tx); dp->dp_scan->scn_restart_txg = txg; dmu_tx_commit(tx); } else { dp->dp_scan->scn_restart_txg = txg; } zfs_dbgmsg("restarting resilver for %s at txg=%llu", dp->dp_spa->spa_name, (longlong_t)txg); } void dsl_free(dsl_pool_t *dp, uint64_t txg, const blkptr_t *bp) { zio_free(dp->dp_spa, txg, bp); } void dsl_free_sync(zio_t *pio, dsl_pool_t *dp, uint64_t txg, const blkptr_t *bpp) { ASSERT(dsl_pool_sync_context(dp)); zio_nowait(zio_free_sync(pio, dp->dp_spa, txg, bpp, pio->io_flags)); } static int scan_ds_queue_compare(const void *a, const void *b) { const scan_ds_t *sds_a = a, *sds_b = b; if (sds_a->sds_dsobj < sds_b->sds_dsobj) return (-1); if (sds_a->sds_dsobj == sds_b->sds_dsobj) return (0); return (1); } static void scan_ds_queue_clear(dsl_scan_t *scn) { void *cookie = NULL; scan_ds_t *sds; while ((sds = avl_destroy_nodes(&scn->scn_queue, &cookie)) != NULL) { kmem_free(sds, sizeof (*sds)); } } static boolean_t scan_ds_queue_contains(dsl_scan_t *scn, uint64_t dsobj, uint64_t *txg) { scan_ds_t srch, *sds; srch.sds_dsobj = dsobj; sds = avl_find(&scn->scn_queue, &srch, NULL); if (sds != NULL && txg != NULL) *txg = sds->sds_txg; return (sds != NULL); } static void scan_ds_queue_insert(dsl_scan_t *scn, uint64_t dsobj, uint64_t txg) { scan_ds_t *sds; avl_index_t where; sds = kmem_zalloc(sizeof (*sds), KM_SLEEP); sds->sds_dsobj = dsobj; sds->sds_txg = txg; VERIFY3P(avl_find(&scn->scn_queue, sds, &where), ==, NULL); avl_insert(&scn->scn_queue, sds, where); } static void scan_ds_queue_remove(dsl_scan_t *scn, uint64_t dsobj) { scan_ds_t srch, *sds; srch.sds_dsobj = dsobj; sds = avl_find(&scn->scn_queue, &srch, NULL); VERIFY(sds != NULL); avl_remove(&scn->scn_queue, sds); kmem_free(sds, sizeof (*sds)); } static void scan_ds_queue_sync(dsl_scan_t *scn, dmu_tx_t *tx) { dsl_pool_t *dp = scn->scn_dp; spa_t *spa = dp->dp_spa; dmu_object_type_t ot = (spa_version(spa) >= SPA_VERSION_DSL_SCRUB) ? DMU_OT_SCAN_QUEUE : DMU_OT_ZAP_OTHER; ASSERT0(scn->scn_queues_pending); ASSERT(scn->scn_phys.scn_queue_obj != 0); VERIFY0(dmu_object_free(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, tx)); scn->scn_phys.scn_queue_obj = zap_create(dp->dp_meta_objset, ot, DMU_OT_NONE, 0, tx); for (scan_ds_t *sds = avl_first(&scn->scn_queue); sds != NULL; sds = AVL_NEXT(&scn->scn_queue, sds)) { VERIFY0(zap_add_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, sds->sds_dsobj, sds->sds_txg, tx)); } } /* * Computes the memory limit state that we're currently in. A sorted scan * needs quite a bit of memory to hold the sorting queue, so we need to * reasonably constrain the size so it doesn't impact overall system * performance. We compute two limits: * 1) Hard memory limit: if the amount of memory used by the sorting * queues on a pool gets above this value, we stop the metadata * scanning portion and start issuing the queued up and sorted * I/Os to reduce memory usage. * This limit is calculated as a fraction of physmem (by default 5%). * We constrain the lower bound of the hard limit to an absolute * minimum of zfs_scan_mem_lim_min (default: 16 MiB). We also constrain * the upper bound to 5% of the total pool size - no chance we'll * ever need that much memory, but just to keep the value in check. * 2) Soft memory limit: once we hit the hard memory limit, we start * issuing I/O to reduce queue memory usage, but we don't want to * completely empty out the queues, since we might be able to find I/Os * that will fill in the gaps of our non-sequential IOs at some point * in the future. So we stop the issuing of I/Os once the amount of * memory used drops below the soft limit (at which point we stop issuing * I/O and start scanning metadata again). * * This limit is calculated by subtracting a fraction of the hard * limit from the hard limit. By default this fraction is 5%, so * the soft limit is 95% of the hard limit. We cap the size of the * difference between the hard and soft limits at an absolute * maximum of zfs_scan_mem_lim_soft_max (default: 128 MiB) - this is * sufficient to not cause too frequent switching between the * metadata scan and I/O issue (even at 2k recordsize, 128 MiB's * worth of queues is about 1.2 GiB of on-pool data, so scanning * that should take at least a decent fraction of a second). */ static boolean_t dsl_scan_should_clear(dsl_scan_t *scn) { spa_t *spa = scn->scn_dp->dp_spa; vdev_t *rvd = scn->scn_dp->dp_spa->spa_root_vdev; uint64_t alloc, mlim_hard, mlim_soft, mused; alloc = metaslab_class_get_alloc(spa_normal_class(spa)); alloc += metaslab_class_get_alloc(spa_special_class(spa)); alloc += metaslab_class_get_alloc(spa_dedup_class(spa)); mlim_hard = MAX((physmem / zfs_scan_mem_lim_fact) * PAGESIZE, zfs_scan_mem_lim_min); mlim_hard = MIN(mlim_hard, alloc / 20); mlim_soft = mlim_hard - MIN(mlim_hard / zfs_scan_mem_lim_soft_fact, zfs_scan_mem_lim_soft_max); mused = 0; for (uint64_t i = 0; i < rvd->vdev_children; i++) { vdev_t *tvd = rvd->vdev_child[i]; dsl_scan_io_queue_t *queue; mutex_enter(&tvd->vdev_scan_io_queue_lock); queue = tvd->vdev_scan_io_queue; if (queue != NULL) { /* * # of extents in exts_by_addr = # in exts_by_size. * B-tree efficiency is ~75%, but can be as low as 50%. */ mused += zfs_btree_numnodes(&queue->q_exts_by_size) * ((sizeof (range_seg_gap_t) + sizeof (uint64_t)) * 3 / 2) + queue->q_sio_memused; } mutex_exit(&tvd->vdev_scan_io_queue_lock); } dprintf("current scan memory usage: %llu bytes\n", (longlong_t)mused); if (mused == 0) ASSERT0(scn->scn_queues_pending); /* * If we are above our hard limit, we need to clear out memory. * If we are below our soft limit, we need to accumulate sequential IOs. * Otherwise, we should keep doing whatever we are currently doing. */ if (mused >= mlim_hard) return (B_TRUE); else if (mused < mlim_soft) return (B_FALSE); else return (scn->scn_clearing); } static boolean_t dsl_scan_check_suspend(dsl_scan_t *scn, const zbookmark_phys_t *zb) { /* we never skip user/group accounting objects */ if (zb && (int64_t)zb->zb_object < 0) return (B_FALSE); if (scn->scn_suspending) return (B_TRUE); /* we're already suspending */ if (!ZB_IS_ZERO(&scn->scn_phys.scn_bookmark)) return (B_FALSE); /* we're resuming */ /* We only know how to resume from level-0 and objset blocks. */ if (zb && (zb->zb_level != 0 && zb->zb_level != ZB_ROOT_LEVEL)) return (B_FALSE); /* * We suspend if: * - we have scanned for at least the minimum time (default 1 sec * for scrub, 3 sec for resilver), and either we have sufficient * dirty data that we are starting to write more quickly * (default 30%), someone is explicitly waiting for this txg * to complete, or we have used up all of the time in the txg * timeout (default 5 sec). * or * - the spa is shutting down because this pool is being exported * or the machine is rebooting. * or * - the scan queue has reached its memory use limit */ uint64_t curr_time_ns = gethrtime(); uint64_t scan_time_ns = curr_time_ns - scn->scn_sync_start_time; uint64_t sync_time_ns = curr_time_ns - scn->scn_dp->dp_spa->spa_sync_starttime; uint64_t dirty_min_bytes = zfs_dirty_data_max * zfs_vdev_async_write_active_min_dirty_percent / 100; uint_t mintime = (scn->scn_phys.scn_func == POOL_SCAN_RESILVER) ? zfs_resilver_min_time_ms : zfs_scrub_min_time_ms; if ((NSEC2MSEC(scan_time_ns) > mintime && (scn->scn_dp->dp_dirty_total >= dirty_min_bytes || txg_sync_waiting(scn->scn_dp) || NSEC2SEC(sync_time_ns) >= zfs_txg_timeout)) || spa_shutting_down(scn->scn_dp->dp_spa) || (zfs_scan_strict_mem_lim && dsl_scan_should_clear(scn))) { if (zb && zb->zb_level == ZB_ROOT_LEVEL) { dprintf("suspending at first available bookmark " "%llx/%llx/%llx/%llx\n", (longlong_t)zb->zb_objset, (longlong_t)zb->zb_object, (longlong_t)zb->zb_level, (longlong_t)zb->zb_blkid); SET_BOOKMARK(&scn->scn_phys.scn_bookmark, zb->zb_objset, 0, 0, 0); } else if (zb != NULL) { dprintf("suspending at bookmark %llx/%llx/%llx/%llx\n", (longlong_t)zb->zb_objset, (longlong_t)zb->zb_object, (longlong_t)zb->zb_level, (longlong_t)zb->zb_blkid); scn->scn_phys.scn_bookmark = *zb; } else { #ifdef ZFS_DEBUG dsl_scan_phys_t *scnp = &scn->scn_phys; dprintf("suspending at at DDT bookmark " "%llx/%llx/%llx/%llx\n", (longlong_t)scnp->scn_ddt_bookmark.ddb_class, (longlong_t)scnp->scn_ddt_bookmark.ddb_type, (longlong_t)scnp->scn_ddt_bookmark.ddb_checksum, (longlong_t)scnp->scn_ddt_bookmark.ddb_cursor); #endif } scn->scn_suspending = B_TRUE; return (B_TRUE); } return (B_FALSE); } +static boolean_t +dsl_error_scrub_check_suspend(dsl_scan_t *scn, const zbookmark_phys_t *zb) +{ + /* + * We suspend if: + * - we have scrubbed for at least the minimum time (default 1 sec + * for error scrub), someone is explicitly waiting for this txg + * to complete, or we have used up all of the time in the txg + * timeout (default 5 sec). + * or + * - the spa is shutting down because this pool is being exported + * or the machine is rebooting. + */ + uint64_t curr_time_ns = gethrtime(); + uint64_t error_scrub_time_ns = curr_time_ns - scn->scn_sync_start_time; + uint64_t sync_time_ns = curr_time_ns - + scn->scn_dp->dp_spa->spa_sync_starttime; + int mintime = zfs_scrub_min_time_ms; + + if ((NSEC2MSEC(error_scrub_time_ns) > mintime && + (txg_sync_waiting(scn->scn_dp) || + NSEC2SEC(sync_time_ns) >= zfs_txg_timeout)) || + spa_shutting_down(scn->scn_dp->dp_spa)) { + if (zb) { + dprintf("error scrub suspending at bookmark " + "%llx/%llx/%llx/%llx\n", + (longlong_t)zb->zb_objset, + (longlong_t)zb->zb_object, + (longlong_t)zb->zb_level, + (longlong_t)zb->zb_blkid); + } + return (B_TRUE); + } + return (B_FALSE); +} + typedef struct zil_scan_arg { dsl_pool_t *zsa_dp; zil_header_t *zsa_zh; } zil_scan_arg_t; static int dsl_scan_zil_block(zilog_t *zilog, const blkptr_t *bp, void *arg, uint64_t claim_txg) { (void) zilog; zil_scan_arg_t *zsa = arg; dsl_pool_t *dp = zsa->zsa_dp; dsl_scan_t *scn = dp->dp_scan; zil_header_t *zh = zsa->zsa_zh; zbookmark_phys_t zb; ASSERT(!BP_IS_REDACTED(bp)); if (BP_IS_HOLE(bp) || bp->blk_birth <= scn->scn_phys.scn_cur_min_txg) return (0); /* * One block ("stubby") can be allocated a long time ago; we * want to visit that one because it has been allocated * (on-disk) even if it hasn't been claimed (even though for * scrub there's nothing to do to it). */ if (claim_txg == 0 && bp->blk_birth >= spa_min_claim_txg(dp->dp_spa)) return (0); SET_BOOKMARK(&zb, zh->zh_log.blk_cksum.zc_word[ZIL_ZC_OBJSET], ZB_ZIL_OBJECT, ZB_ZIL_LEVEL, bp->blk_cksum.zc_word[ZIL_ZC_SEQ]); VERIFY(0 == scan_funcs[scn->scn_phys.scn_func](dp, bp, &zb)); return (0); } static int dsl_scan_zil_record(zilog_t *zilog, const lr_t *lrc, void *arg, uint64_t claim_txg) { (void) zilog; if (lrc->lrc_txtype == TX_WRITE) { zil_scan_arg_t *zsa = arg; dsl_pool_t *dp = zsa->zsa_dp; dsl_scan_t *scn = dp->dp_scan; zil_header_t *zh = zsa->zsa_zh; const lr_write_t *lr = (const lr_write_t *)lrc; const blkptr_t *bp = &lr->lr_blkptr; zbookmark_phys_t zb; ASSERT(!BP_IS_REDACTED(bp)); if (BP_IS_HOLE(bp) || bp->blk_birth <= scn->scn_phys.scn_cur_min_txg) return (0); /* * birth can be < claim_txg if this record's txg is * already txg sync'ed (but this log block contains * other records that are not synced) */ if (claim_txg == 0 || bp->blk_birth < claim_txg) return (0); ASSERT3U(BP_GET_LSIZE(bp), !=, 0); SET_BOOKMARK(&zb, zh->zh_log.blk_cksum.zc_word[ZIL_ZC_OBJSET], lr->lr_foid, ZB_ZIL_LEVEL, lr->lr_offset / BP_GET_LSIZE(bp)); VERIFY(0 == scan_funcs[scn->scn_phys.scn_func](dp, bp, &zb)); } return (0); } static void dsl_scan_zil(dsl_pool_t *dp, zil_header_t *zh) { uint64_t claim_txg = zh->zh_claim_txg; zil_scan_arg_t zsa = { dp, zh }; zilog_t *zilog; ASSERT(spa_writeable(dp->dp_spa)); /* * We only want to visit blocks that have been claimed but not yet * replayed (or, in read-only mode, blocks that *would* be claimed). */ if (claim_txg == 0) return; zilog = zil_alloc(dp->dp_meta_objset, zh); (void) zil_parse(zilog, dsl_scan_zil_block, dsl_scan_zil_record, &zsa, claim_txg, B_FALSE); zil_free(zilog); } /* * We compare scan_prefetch_issue_ctx_t's based on their bookmarks. The idea * here is to sort the AVL tree by the order each block will be needed. */ static int scan_prefetch_queue_compare(const void *a, const void *b) { const scan_prefetch_issue_ctx_t *spic_a = a, *spic_b = b; const scan_prefetch_ctx_t *spc_a = spic_a->spic_spc; const scan_prefetch_ctx_t *spc_b = spic_b->spic_spc; return (zbookmark_compare(spc_a->spc_datablkszsec, spc_a->spc_indblkshift, spc_b->spc_datablkszsec, spc_b->spc_indblkshift, &spic_a->spic_zb, &spic_b->spic_zb)); } static void scan_prefetch_ctx_rele(scan_prefetch_ctx_t *spc, const void *tag) { if (zfs_refcount_remove(&spc->spc_refcnt, tag) == 0) { zfs_refcount_destroy(&spc->spc_refcnt); kmem_free(spc, sizeof (scan_prefetch_ctx_t)); } } static scan_prefetch_ctx_t * scan_prefetch_ctx_create(dsl_scan_t *scn, dnode_phys_t *dnp, const void *tag) { scan_prefetch_ctx_t *spc; spc = kmem_alloc(sizeof (scan_prefetch_ctx_t), KM_SLEEP); zfs_refcount_create(&spc->spc_refcnt); zfs_refcount_add(&spc->spc_refcnt, tag); spc->spc_scn = scn; if (dnp != NULL) { spc->spc_datablkszsec = dnp->dn_datablkszsec; spc->spc_indblkshift = dnp->dn_indblkshift; spc->spc_root = B_FALSE; } else { spc->spc_datablkszsec = 0; spc->spc_indblkshift = 0; spc->spc_root = B_TRUE; } return (spc); } static void scan_prefetch_ctx_add_ref(scan_prefetch_ctx_t *spc, const void *tag) { zfs_refcount_add(&spc->spc_refcnt, tag); } static void scan_ds_prefetch_queue_clear(dsl_scan_t *scn) { spa_t *spa = scn->scn_dp->dp_spa; void *cookie = NULL; scan_prefetch_issue_ctx_t *spic = NULL; mutex_enter(&spa->spa_scrub_lock); while ((spic = avl_destroy_nodes(&scn->scn_prefetch_queue, &cookie)) != NULL) { scan_prefetch_ctx_rele(spic->spic_spc, scn); kmem_free(spic, sizeof (scan_prefetch_issue_ctx_t)); } mutex_exit(&spa->spa_scrub_lock); } static boolean_t dsl_scan_check_prefetch_resume(scan_prefetch_ctx_t *spc, const zbookmark_phys_t *zb) { zbookmark_phys_t *last_zb = &spc->spc_scn->scn_prefetch_bookmark; dnode_phys_t tmp_dnp; dnode_phys_t *dnp = (spc->spc_root) ? NULL : &tmp_dnp; if (zb->zb_objset != last_zb->zb_objset) return (B_TRUE); if ((int64_t)zb->zb_object < 0) return (B_FALSE); tmp_dnp.dn_datablkszsec = spc->spc_datablkszsec; tmp_dnp.dn_indblkshift = spc->spc_indblkshift; if (zbookmark_subtree_completed(dnp, zb, last_zb)) return (B_TRUE); return (B_FALSE); } static void dsl_scan_prefetch(scan_prefetch_ctx_t *spc, blkptr_t *bp, zbookmark_phys_t *zb) { avl_index_t idx; dsl_scan_t *scn = spc->spc_scn; spa_t *spa = scn->scn_dp->dp_spa; scan_prefetch_issue_ctx_t *spic; if (zfs_no_scrub_prefetch || BP_IS_REDACTED(bp)) return; if (BP_IS_HOLE(bp) || bp->blk_birth <= scn->scn_phys.scn_cur_min_txg || (BP_GET_LEVEL(bp) == 0 && BP_GET_TYPE(bp) != DMU_OT_DNODE && BP_GET_TYPE(bp) != DMU_OT_OBJSET)) return; if (dsl_scan_check_prefetch_resume(spc, zb)) return; scan_prefetch_ctx_add_ref(spc, scn); spic = kmem_alloc(sizeof (scan_prefetch_issue_ctx_t), KM_SLEEP); spic->spic_spc = spc; spic->spic_bp = *bp; spic->spic_zb = *zb; /* * Add the IO to the queue of blocks to prefetch. This allows us to * prioritize blocks that we will need first for the main traversal * thread. */ mutex_enter(&spa->spa_scrub_lock); if (avl_find(&scn->scn_prefetch_queue, spic, &idx) != NULL) { /* this block is already queued for prefetch */ kmem_free(spic, sizeof (scan_prefetch_issue_ctx_t)); scan_prefetch_ctx_rele(spc, scn); mutex_exit(&spa->spa_scrub_lock); return; } avl_insert(&scn->scn_prefetch_queue, spic, idx); cv_broadcast(&spa->spa_scrub_io_cv); mutex_exit(&spa->spa_scrub_lock); } static void dsl_scan_prefetch_dnode(dsl_scan_t *scn, dnode_phys_t *dnp, uint64_t objset, uint64_t object) { int i; zbookmark_phys_t zb; scan_prefetch_ctx_t *spc; if (dnp->dn_nblkptr == 0 && !(dnp->dn_flags & DNODE_FLAG_SPILL_BLKPTR)) return; SET_BOOKMARK(&zb, objset, object, 0, 0); spc = scan_prefetch_ctx_create(scn, dnp, FTAG); for (i = 0; i < dnp->dn_nblkptr; i++) { zb.zb_level = BP_GET_LEVEL(&dnp->dn_blkptr[i]); zb.zb_blkid = i; dsl_scan_prefetch(spc, &dnp->dn_blkptr[i], &zb); } if (dnp->dn_flags & DNODE_FLAG_SPILL_BLKPTR) { zb.zb_level = 0; zb.zb_blkid = DMU_SPILL_BLKID; dsl_scan_prefetch(spc, DN_SPILL_BLKPTR(dnp), &zb); } scan_prefetch_ctx_rele(spc, FTAG); } static void dsl_scan_prefetch_cb(zio_t *zio, const zbookmark_phys_t *zb, const blkptr_t *bp, arc_buf_t *buf, void *private) { (void) zio; scan_prefetch_ctx_t *spc = private; dsl_scan_t *scn = spc->spc_scn; spa_t *spa = scn->scn_dp->dp_spa; /* broadcast that the IO has completed for rate limiting purposes */ mutex_enter(&spa->spa_scrub_lock); ASSERT3U(spa->spa_scrub_inflight, >=, BP_GET_PSIZE(bp)); spa->spa_scrub_inflight -= BP_GET_PSIZE(bp); cv_broadcast(&spa->spa_scrub_io_cv); mutex_exit(&spa->spa_scrub_lock); /* if there was an error or we are done prefetching, just cleanup */ if (buf == NULL || scn->scn_prefetch_stop) goto out; if (BP_GET_LEVEL(bp) > 0) { int i; blkptr_t *cbp; int epb = BP_GET_LSIZE(bp) >> SPA_BLKPTRSHIFT; zbookmark_phys_t czb; for (i = 0, cbp = buf->b_data; i < epb; i++, cbp++) { SET_BOOKMARK(&czb, zb->zb_objset, zb->zb_object, zb->zb_level - 1, zb->zb_blkid * epb + i); dsl_scan_prefetch(spc, cbp, &czb); } } else if (BP_GET_TYPE(bp) == DMU_OT_DNODE) { dnode_phys_t *cdnp; int i; int epb = BP_GET_LSIZE(bp) >> DNODE_SHIFT; for (i = 0, cdnp = buf->b_data; i < epb; i += cdnp->dn_extra_slots + 1, cdnp += cdnp->dn_extra_slots + 1) { dsl_scan_prefetch_dnode(scn, cdnp, zb->zb_objset, zb->zb_blkid * epb + i); } } else if (BP_GET_TYPE(bp) == DMU_OT_OBJSET) { objset_phys_t *osp = buf->b_data; dsl_scan_prefetch_dnode(scn, &osp->os_meta_dnode, zb->zb_objset, DMU_META_DNODE_OBJECT); if (OBJSET_BUF_HAS_USERUSED(buf)) { dsl_scan_prefetch_dnode(scn, &osp->os_groupused_dnode, zb->zb_objset, DMU_GROUPUSED_OBJECT); dsl_scan_prefetch_dnode(scn, &osp->os_userused_dnode, zb->zb_objset, DMU_USERUSED_OBJECT); } } out: if (buf != NULL) arc_buf_destroy(buf, private); scan_prefetch_ctx_rele(spc, scn); } static void dsl_scan_prefetch_thread(void *arg) { dsl_scan_t *scn = arg; spa_t *spa = scn->scn_dp->dp_spa; scan_prefetch_issue_ctx_t *spic; /* loop until we are told to stop */ while (!scn->scn_prefetch_stop) { arc_flags_t flags = ARC_FLAG_NOWAIT | ARC_FLAG_PRESCIENT_PREFETCH | ARC_FLAG_PREFETCH; int zio_flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_SCAN_THREAD; mutex_enter(&spa->spa_scrub_lock); /* * Wait until we have an IO to issue and are not above our * maximum in flight limit. */ while (!scn->scn_prefetch_stop && (avl_numnodes(&scn->scn_prefetch_queue) == 0 || spa->spa_scrub_inflight >= scn->scn_maxinflight_bytes)) { cv_wait(&spa->spa_scrub_io_cv, &spa->spa_scrub_lock); } /* recheck if we should stop since we waited for the cv */ if (scn->scn_prefetch_stop) { mutex_exit(&spa->spa_scrub_lock); break; } /* remove the prefetch IO from the tree */ spic = avl_first(&scn->scn_prefetch_queue); spa->spa_scrub_inflight += BP_GET_PSIZE(&spic->spic_bp); avl_remove(&scn->scn_prefetch_queue, spic); mutex_exit(&spa->spa_scrub_lock); if (BP_IS_PROTECTED(&spic->spic_bp)) { ASSERT(BP_GET_TYPE(&spic->spic_bp) == DMU_OT_DNODE || BP_GET_TYPE(&spic->spic_bp) == DMU_OT_OBJSET); ASSERT3U(BP_GET_LEVEL(&spic->spic_bp), ==, 0); zio_flags |= ZIO_FLAG_RAW; } /* issue the prefetch asynchronously */ (void) arc_read(scn->scn_zio_root, scn->scn_dp->dp_spa, &spic->spic_bp, dsl_scan_prefetch_cb, spic->spic_spc, ZIO_PRIORITY_SCRUB, zio_flags, &flags, &spic->spic_zb); kmem_free(spic, sizeof (scan_prefetch_issue_ctx_t)); } ASSERT(scn->scn_prefetch_stop); /* free any prefetches we didn't get to complete */ mutex_enter(&spa->spa_scrub_lock); while ((spic = avl_first(&scn->scn_prefetch_queue)) != NULL) { avl_remove(&scn->scn_prefetch_queue, spic); scan_prefetch_ctx_rele(spic->spic_spc, scn); kmem_free(spic, sizeof (scan_prefetch_issue_ctx_t)); } ASSERT0(avl_numnodes(&scn->scn_prefetch_queue)); mutex_exit(&spa->spa_scrub_lock); } static boolean_t dsl_scan_check_resume(dsl_scan_t *scn, const dnode_phys_t *dnp, const zbookmark_phys_t *zb) { /* * We never skip over user/group accounting objects (obj<0) */ if (!ZB_IS_ZERO(&scn->scn_phys.scn_bookmark) && (int64_t)zb->zb_object >= 0) { /* * If we already visited this bp & everything below (in * a prior txg sync), don't bother doing it again. */ if (zbookmark_subtree_completed(dnp, zb, &scn->scn_phys.scn_bookmark)) return (B_TRUE); /* * If we found the block we're trying to resume from, or * we went past it, zero it out to indicate that it's OK * to start checking for suspending again. */ if (zbookmark_subtree_tbd(dnp, zb, &scn->scn_phys.scn_bookmark)) { dprintf("resuming at %llx/%llx/%llx/%llx\n", (longlong_t)zb->zb_objset, (longlong_t)zb->zb_object, (longlong_t)zb->zb_level, (longlong_t)zb->zb_blkid); memset(&scn->scn_phys.scn_bookmark, 0, sizeof (*zb)); } } return (B_FALSE); } static void dsl_scan_visitbp(blkptr_t *bp, const zbookmark_phys_t *zb, dnode_phys_t *dnp, dsl_dataset_t *ds, dsl_scan_t *scn, dmu_objset_type_t ostype, dmu_tx_t *tx); inline __attribute__((always_inline)) static void dsl_scan_visitdnode( dsl_scan_t *, dsl_dataset_t *ds, dmu_objset_type_t ostype, dnode_phys_t *dnp, uint64_t object, dmu_tx_t *tx); /* * Return nonzero on i/o error. * Return new buf to write out in *bufp. */ inline __attribute__((always_inline)) static int dsl_scan_recurse(dsl_scan_t *scn, dsl_dataset_t *ds, dmu_objset_type_t ostype, dnode_phys_t *dnp, const blkptr_t *bp, const zbookmark_phys_t *zb, dmu_tx_t *tx) { dsl_pool_t *dp = scn->scn_dp; spa_t *spa = dp->dp_spa; int zio_flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_SCAN_THREAD; int err; ASSERT(!BP_IS_REDACTED(bp)); /* * There is an unlikely case of encountering dnodes with contradicting * dn_bonuslen and DNODE_FLAG_SPILL_BLKPTR flag before in files created * or modified before commit 4254acb was merged. As it is not possible * to know which of the two is correct, report an error. */ if (dnp != NULL && dnp->dn_bonuslen > DN_MAX_BONUS_LEN(dnp)) { scn->scn_phys.scn_errors++; spa_log_error(spa, zb, &bp->blk_birth); return (SET_ERROR(EINVAL)); } if (BP_GET_LEVEL(bp) > 0) { arc_flags_t flags = ARC_FLAG_WAIT; int i; blkptr_t *cbp; int epb = BP_GET_LSIZE(bp) >> SPA_BLKPTRSHIFT; arc_buf_t *buf; err = arc_read(NULL, spa, bp, arc_getbuf_func, &buf, ZIO_PRIORITY_SCRUB, zio_flags, &flags, zb); if (err) { scn->scn_phys.scn_errors++; return (err); } for (i = 0, cbp = buf->b_data; i < epb; i++, cbp++) { zbookmark_phys_t czb; SET_BOOKMARK(&czb, zb->zb_objset, zb->zb_object, zb->zb_level - 1, zb->zb_blkid * epb + i); dsl_scan_visitbp(cbp, &czb, dnp, ds, scn, ostype, tx); } arc_buf_destroy(buf, &buf); } else if (BP_GET_TYPE(bp) == DMU_OT_DNODE) { arc_flags_t flags = ARC_FLAG_WAIT; dnode_phys_t *cdnp; int i; int epb = BP_GET_LSIZE(bp) >> DNODE_SHIFT; arc_buf_t *buf; if (BP_IS_PROTECTED(bp)) { ASSERT3U(BP_GET_COMPRESS(bp), ==, ZIO_COMPRESS_OFF); zio_flags |= ZIO_FLAG_RAW; } err = arc_read(NULL, spa, bp, arc_getbuf_func, &buf, ZIO_PRIORITY_SCRUB, zio_flags, &flags, zb); if (err) { scn->scn_phys.scn_errors++; return (err); } for (i = 0, cdnp = buf->b_data; i < epb; i += cdnp->dn_extra_slots + 1, cdnp += cdnp->dn_extra_slots + 1) { dsl_scan_visitdnode(scn, ds, ostype, cdnp, zb->zb_blkid * epb + i, tx); } arc_buf_destroy(buf, &buf); } else if (BP_GET_TYPE(bp) == DMU_OT_OBJSET) { arc_flags_t flags = ARC_FLAG_WAIT; objset_phys_t *osp; arc_buf_t *buf; err = arc_read(NULL, spa, bp, arc_getbuf_func, &buf, ZIO_PRIORITY_SCRUB, zio_flags, &flags, zb); if (err) { scn->scn_phys.scn_errors++; return (err); } osp = buf->b_data; dsl_scan_visitdnode(scn, ds, osp->os_type, &osp->os_meta_dnode, DMU_META_DNODE_OBJECT, tx); if (OBJSET_BUF_HAS_USERUSED(buf)) { /* * We also always visit user/group/project accounting * objects, and never skip them, even if we are * suspending. This is necessary so that the * space deltas from this txg get integrated. */ if (OBJSET_BUF_HAS_PROJECTUSED(buf)) dsl_scan_visitdnode(scn, ds, osp->os_type, &osp->os_projectused_dnode, DMU_PROJECTUSED_OBJECT, tx); dsl_scan_visitdnode(scn, ds, osp->os_type, &osp->os_groupused_dnode, DMU_GROUPUSED_OBJECT, tx); dsl_scan_visitdnode(scn, ds, osp->os_type, &osp->os_userused_dnode, DMU_USERUSED_OBJECT, tx); } arc_buf_destroy(buf, &buf); } else if (!zfs_blkptr_verify(spa, bp, BLK_CONFIG_NEEDED, BLK_VERIFY_LOG)) { /* * Sanity check the block pointer contents, this is handled * by arc_read() for the cases above. */ scn->scn_phys.scn_errors++; spa_log_error(spa, zb, &bp->blk_birth); return (SET_ERROR(EINVAL)); } return (0); } inline __attribute__((always_inline)) static void dsl_scan_visitdnode(dsl_scan_t *scn, dsl_dataset_t *ds, dmu_objset_type_t ostype, dnode_phys_t *dnp, uint64_t object, dmu_tx_t *tx) { int j; for (j = 0; j < dnp->dn_nblkptr; j++) { zbookmark_phys_t czb; SET_BOOKMARK(&czb, ds ? ds->ds_object : 0, object, dnp->dn_nlevels - 1, j); dsl_scan_visitbp(&dnp->dn_blkptr[j], &czb, dnp, ds, scn, ostype, tx); } if (dnp->dn_flags & DNODE_FLAG_SPILL_BLKPTR) { zbookmark_phys_t czb; SET_BOOKMARK(&czb, ds ? ds->ds_object : 0, object, 0, DMU_SPILL_BLKID); dsl_scan_visitbp(DN_SPILL_BLKPTR(dnp), &czb, dnp, ds, scn, ostype, tx); } } /* * The arguments are in this order because mdb can only print the * first 5; we want them to be useful. */ static void dsl_scan_visitbp(blkptr_t *bp, const zbookmark_phys_t *zb, dnode_phys_t *dnp, dsl_dataset_t *ds, dsl_scan_t *scn, dmu_objset_type_t ostype, dmu_tx_t *tx) { dsl_pool_t *dp = scn->scn_dp; blkptr_t *bp_toread = NULL; if (dsl_scan_check_suspend(scn, zb)) return; if (dsl_scan_check_resume(scn, dnp, zb)) return; scn->scn_visited_this_txg++; if (BP_IS_HOLE(bp)) { scn->scn_holes_this_txg++; return; } if (BP_IS_REDACTED(bp)) { ASSERT(dsl_dataset_feature_is_active(ds, SPA_FEATURE_REDACTED_DATASETS)); return; } /* * Check if this block contradicts any filesystem flags. */ spa_feature_t f = SPA_FEATURE_LARGE_BLOCKS; if (BP_GET_LSIZE(bp) > SPA_OLD_MAXBLOCKSIZE) ASSERT(dsl_dataset_feature_is_active(ds, f)); f = zio_checksum_to_feature(BP_GET_CHECKSUM(bp)); if (f != SPA_FEATURE_NONE) ASSERT(dsl_dataset_feature_is_active(ds, f)); f = zio_compress_to_feature(BP_GET_COMPRESS(bp)); if (f != SPA_FEATURE_NONE) ASSERT(dsl_dataset_feature_is_active(ds, f)); if (bp->blk_birth <= scn->scn_phys.scn_cur_min_txg) { scn->scn_lt_min_this_txg++; return; } bp_toread = kmem_alloc(sizeof (blkptr_t), KM_SLEEP); *bp_toread = *bp; if (dsl_scan_recurse(scn, ds, ostype, dnp, bp_toread, zb, tx) != 0) goto out; /* * If dsl_scan_ddt() has already visited this block, it will have * already done any translations or scrubbing, so don't call the * callback again. */ if (ddt_class_contains(dp->dp_spa, scn->scn_phys.scn_ddt_class_max, bp)) { scn->scn_ddt_contained_this_txg++; goto out; } /* * If this block is from the future (after cur_max_txg), then we * are doing this on behalf of a deleted snapshot, and we will * revisit the future block on the next pass of this dataset. * Don't scan it now unless we need to because something * under it was modified. */ if (BP_PHYSICAL_BIRTH(bp) > scn->scn_phys.scn_cur_max_txg) { scn->scn_gt_max_this_txg++; goto out; } scan_funcs[scn->scn_phys.scn_func](dp, bp, zb); out: kmem_free(bp_toread, sizeof (blkptr_t)); } static void dsl_scan_visit_rootbp(dsl_scan_t *scn, dsl_dataset_t *ds, blkptr_t *bp, dmu_tx_t *tx) { zbookmark_phys_t zb; scan_prefetch_ctx_t *spc; SET_BOOKMARK(&zb, ds ? ds->ds_object : DMU_META_OBJSET, ZB_ROOT_OBJECT, ZB_ROOT_LEVEL, ZB_ROOT_BLKID); if (ZB_IS_ZERO(&scn->scn_phys.scn_bookmark)) { SET_BOOKMARK(&scn->scn_prefetch_bookmark, zb.zb_objset, 0, 0, 0); } else { scn->scn_prefetch_bookmark = scn->scn_phys.scn_bookmark; } scn->scn_objsets_visited_this_txg++; spc = scan_prefetch_ctx_create(scn, NULL, FTAG); dsl_scan_prefetch(spc, bp, &zb); scan_prefetch_ctx_rele(spc, FTAG); dsl_scan_visitbp(bp, &zb, NULL, ds, scn, DMU_OST_NONE, tx); dprintf_ds(ds, "finished scan%s", ""); } static void ds_destroyed_scn_phys(dsl_dataset_t *ds, dsl_scan_phys_t *scn_phys) { if (scn_phys->scn_bookmark.zb_objset == ds->ds_object) { if (ds->ds_is_snapshot) { /* * Note: * - scn_cur_{min,max}_txg stays the same. * - Setting the flag is not really necessary if * scn_cur_max_txg == scn_max_txg, because there * is nothing after this snapshot that we care * about. However, we set it anyway and then * ignore it when we retraverse it in * dsl_scan_visitds(). */ scn_phys->scn_bookmark.zb_objset = dsl_dataset_phys(ds)->ds_next_snap_obj; zfs_dbgmsg("destroying ds %llu on %s; currently " "traversing; reset zb_objset to %llu", (u_longlong_t)ds->ds_object, ds->ds_dir->dd_pool->dp_spa->spa_name, (u_longlong_t)dsl_dataset_phys(ds)-> ds_next_snap_obj); scn_phys->scn_flags |= DSF_VISIT_DS_AGAIN; } else { SET_BOOKMARK(&scn_phys->scn_bookmark, ZB_DESTROYED_OBJSET, 0, 0, 0); zfs_dbgmsg("destroying ds %llu on %s; currently " "traversing; reset bookmark to -1,0,0,0", (u_longlong_t)ds->ds_object, ds->ds_dir->dd_pool->dp_spa->spa_name); } } } /* * Invoked when a dataset is destroyed. We need to make sure that: * * 1) If it is the dataset that was currently being scanned, we write * a new dsl_scan_phys_t and marking the objset reference in it * as destroyed. * 2) Remove it from the work queue, if it was present. * * If the dataset was actually a snapshot, instead of marking the dataset * as destroyed, we instead substitute the next snapshot in line. */ void dsl_scan_ds_destroyed(dsl_dataset_t *ds, dmu_tx_t *tx) { dsl_pool_t *dp = ds->ds_dir->dd_pool; dsl_scan_t *scn = dp->dp_scan; uint64_t mintxg; if (!dsl_scan_is_running(scn)) return; ds_destroyed_scn_phys(ds, &scn->scn_phys); ds_destroyed_scn_phys(ds, &scn->scn_phys_cached); if (scan_ds_queue_contains(scn, ds->ds_object, &mintxg)) { scan_ds_queue_remove(scn, ds->ds_object); if (ds->ds_is_snapshot) scan_ds_queue_insert(scn, dsl_dataset_phys(ds)->ds_next_snap_obj, mintxg); } if (zap_lookup_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds->ds_object, &mintxg) == 0) { ASSERT3U(dsl_dataset_phys(ds)->ds_num_children, <=, 1); VERIFY3U(0, ==, zap_remove_int(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds->ds_object, tx)); if (ds->ds_is_snapshot) { /* * We keep the same mintxg; it could be > * ds_creation_txg if the previous snapshot was * deleted too. */ VERIFY(zap_add_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, dsl_dataset_phys(ds)->ds_next_snap_obj, mintxg, tx) == 0); zfs_dbgmsg("destroying ds %llu on %s; in queue; " "replacing with %llu", (u_longlong_t)ds->ds_object, dp->dp_spa->spa_name, (u_longlong_t)dsl_dataset_phys(ds)-> ds_next_snap_obj); } else { zfs_dbgmsg("destroying ds %llu on %s; in queue; " "removing", (u_longlong_t)ds->ds_object, dp->dp_spa->spa_name); } } /* * dsl_scan_sync() should be called after this, and should sync * out our changed state, but just to be safe, do it here. */ dsl_scan_sync_state(scn, tx, SYNC_CACHED); } static void ds_snapshotted_bookmark(dsl_dataset_t *ds, zbookmark_phys_t *scn_bookmark) { if (scn_bookmark->zb_objset == ds->ds_object) { scn_bookmark->zb_objset = dsl_dataset_phys(ds)->ds_prev_snap_obj; zfs_dbgmsg("snapshotting ds %llu on %s; currently traversing; " "reset zb_objset to %llu", (u_longlong_t)ds->ds_object, ds->ds_dir->dd_pool->dp_spa->spa_name, (u_longlong_t)dsl_dataset_phys(ds)->ds_prev_snap_obj); } } /* * Called when a dataset is snapshotted. If we were currently traversing * this snapshot, we reset our bookmark to point at the newly created * snapshot. We also modify our work queue to remove the old snapshot and * replace with the new one. */ void dsl_scan_ds_snapshotted(dsl_dataset_t *ds, dmu_tx_t *tx) { dsl_pool_t *dp = ds->ds_dir->dd_pool; dsl_scan_t *scn = dp->dp_scan; uint64_t mintxg; if (!dsl_scan_is_running(scn)) return; ASSERT(dsl_dataset_phys(ds)->ds_prev_snap_obj != 0); ds_snapshotted_bookmark(ds, &scn->scn_phys.scn_bookmark); ds_snapshotted_bookmark(ds, &scn->scn_phys_cached.scn_bookmark); if (scan_ds_queue_contains(scn, ds->ds_object, &mintxg)) { scan_ds_queue_remove(scn, ds->ds_object); scan_ds_queue_insert(scn, dsl_dataset_phys(ds)->ds_prev_snap_obj, mintxg); } if (zap_lookup_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds->ds_object, &mintxg) == 0) { VERIFY3U(0, ==, zap_remove_int(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds->ds_object, tx)); VERIFY(zap_add_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, dsl_dataset_phys(ds)->ds_prev_snap_obj, mintxg, tx) == 0); zfs_dbgmsg("snapshotting ds %llu on %s; in queue; " "replacing with %llu", (u_longlong_t)ds->ds_object, dp->dp_spa->spa_name, (u_longlong_t)dsl_dataset_phys(ds)->ds_prev_snap_obj); } dsl_scan_sync_state(scn, tx, SYNC_CACHED); } static void ds_clone_swapped_bookmark(dsl_dataset_t *ds1, dsl_dataset_t *ds2, zbookmark_phys_t *scn_bookmark) { if (scn_bookmark->zb_objset == ds1->ds_object) { scn_bookmark->zb_objset = ds2->ds_object; zfs_dbgmsg("clone_swap ds %llu on %s; currently traversing; " "reset zb_objset to %llu", (u_longlong_t)ds1->ds_object, ds1->ds_dir->dd_pool->dp_spa->spa_name, (u_longlong_t)ds2->ds_object); } else if (scn_bookmark->zb_objset == ds2->ds_object) { scn_bookmark->zb_objset = ds1->ds_object; zfs_dbgmsg("clone_swap ds %llu on %s; currently traversing; " "reset zb_objset to %llu", (u_longlong_t)ds2->ds_object, ds2->ds_dir->dd_pool->dp_spa->spa_name, (u_longlong_t)ds1->ds_object); } } /* * Called when an origin dataset and its clone are swapped. If we were * currently traversing the dataset, we need to switch to traversing the * newly promoted clone. */ void dsl_scan_ds_clone_swapped(dsl_dataset_t *ds1, dsl_dataset_t *ds2, dmu_tx_t *tx) { dsl_pool_t *dp = ds1->ds_dir->dd_pool; dsl_scan_t *scn = dp->dp_scan; uint64_t mintxg1, mintxg2; boolean_t ds1_queued, ds2_queued; if (!dsl_scan_is_running(scn)) return; ds_clone_swapped_bookmark(ds1, ds2, &scn->scn_phys.scn_bookmark); ds_clone_swapped_bookmark(ds1, ds2, &scn->scn_phys_cached.scn_bookmark); /* * Handle the in-memory scan queue. */ ds1_queued = scan_ds_queue_contains(scn, ds1->ds_object, &mintxg1); ds2_queued = scan_ds_queue_contains(scn, ds2->ds_object, &mintxg2); /* Sanity checking. */ if (ds1_queued) { ASSERT3U(mintxg1, ==, dsl_dataset_phys(ds1)->ds_prev_snap_txg); ASSERT3U(mintxg1, ==, dsl_dataset_phys(ds2)->ds_prev_snap_txg); } if (ds2_queued) { ASSERT3U(mintxg2, ==, dsl_dataset_phys(ds1)->ds_prev_snap_txg); ASSERT3U(mintxg2, ==, dsl_dataset_phys(ds2)->ds_prev_snap_txg); } if (ds1_queued && ds2_queued) { /* * If both are queued, we don't need to do anything. * The swapping code below would not handle this case correctly, * since we can't insert ds2 if it is already there. That's * because scan_ds_queue_insert() prohibits a duplicate insert * and panics. */ } else if (ds1_queued) { scan_ds_queue_remove(scn, ds1->ds_object); scan_ds_queue_insert(scn, ds2->ds_object, mintxg1); } else if (ds2_queued) { scan_ds_queue_remove(scn, ds2->ds_object); scan_ds_queue_insert(scn, ds1->ds_object, mintxg2); } /* * Handle the on-disk scan queue. * The on-disk state is an out-of-date version of the in-memory state, * so the in-memory and on-disk values for ds1_queued and ds2_queued may * be different. Therefore we need to apply the swap logic to the * on-disk state independently of the in-memory state. */ ds1_queued = zap_lookup_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds1->ds_object, &mintxg1) == 0; ds2_queued = zap_lookup_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds2->ds_object, &mintxg2) == 0; /* Sanity checking. */ if (ds1_queued) { ASSERT3U(mintxg1, ==, dsl_dataset_phys(ds1)->ds_prev_snap_txg); ASSERT3U(mintxg1, ==, dsl_dataset_phys(ds2)->ds_prev_snap_txg); } if (ds2_queued) { ASSERT3U(mintxg2, ==, dsl_dataset_phys(ds1)->ds_prev_snap_txg); ASSERT3U(mintxg2, ==, dsl_dataset_phys(ds2)->ds_prev_snap_txg); } if (ds1_queued && ds2_queued) { /* * If both are queued, we don't need to do anything. * Alternatively, we could check for EEXIST from * zap_add_int_key() and back out to the original state, but * that would be more work than checking for this case upfront. */ } else if (ds1_queued) { VERIFY3S(0, ==, zap_remove_int(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds1->ds_object, tx)); VERIFY3S(0, ==, zap_add_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds2->ds_object, mintxg1, tx)); zfs_dbgmsg("clone_swap ds %llu on %s; in queue; " "replacing with %llu", (u_longlong_t)ds1->ds_object, dp->dp_spa->spa_name, (u_longlong_t)ds2->ds_object); } else if (ds2_queued) { VERIFY3S(0, ==, zap_remove_int(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds2->ds_object, tx)); VERIFY3S(0, ==, zap_add_int_key(dp->dp_meta_objset, scn->scn_phys.scn_queue_obj, ds1->ds_object, mintxg2, tx)); zfs_dbgmsg("clone_swap ds %llu on %s; in queue; " "replacing with %llu", (u_longlong_t)ds2->ds_object, dp->dp_spa->spa_name, (u_longlong_t)ds1->ds_object); } dsl_scan_sync_state(scn, tx, SYNC_CACHED); } static int enqueue_clones_cb(dsl_pool_t *dp, dsl_dataset_t *hds, void *arg) { uint64_t originobj = *(uint64_t *)arg; dsl_dataset_t *ds; int err; dsl_scan_t *scn = dp->dp_scan; if (dsl_dir_phys(hds->ds_dir)->dd_origin_obj != originobj) return (0); err = dsl_dataset_hold_obj(dp, hds->ds_object, FTAG, &ds); if (err) return (err); while (dsl_dataset_phys(ds)->ds_prev_snap_obj != originobj) { dsl_dataset_t *prev; err = dsl_dataset_hold_obj(dp, dsl_dataset_phys(ds)->ds_prev_snap_obj, FTAG, &prev); dsl_dataset_rele(ds, FTAG); if (err) return (err); ds = prev; } scan_ds_queue_insert(scn, ds->ds_object, dsl_dataset_phys(ds)->ds_prev_snap_txg); dsl_dataset_rele(ds, FTAG); return (0); } static void dsl_scan_visitds(dsl_scan_t *scn, uint64_t dsobj, dmu_tx_t *tx) { dsl_pool_t *dp = scn->scn_dp; dsl_dataset_t *ds; VERIFY3U(0, ==, dsl_dataset_hold_obj(dp, dsobj, FTAG, &ds)); if (scn->scn_phys.scn_cur_min_txg >= scn->scn_phys.scn_max_txg) { /* * This can happen if this snapshot was created after the * scan started, and we already completed a previous snapshot * that was created after the scan started. This snapshot * only references blocks with: * * birth < our ds_creation_txg * cur_min_txg is no less than ds_creation_txg. * We have already visited these blocks. * or * birth > scn_max_txg * The scan requested not to visit these blocks. * * Subsequent snapshots (and clones) can reference our * blocks, or blocks with even higher birth times. * Therefore we do not need to visit them either, * so we do not add them to the work queue. * * Note that checking for cur_min_txg >= cur_max_txg * is not sufficient, because in that case we may need to * visit subsequent snapshots. This happens when min_txg > 0, * which raises cur_min_txg. In this case we will visit * this dataset but skip all of its blocks, because the * rootbp's birth time is < cur_min_txg. Then we will * add the next snapshots/clones to the work queue. */ char *dsname = kmem_alloc(ZFS_MAX_DATASET_NAME_LEN, KM_SLEEP); dsl_dataset_name(ds, dsname); zfs_dbgmsg("scanning dataset %llu (%s) is unnecessary because " "cur_min_txg (%llu) >= max_txg (%llu)", (longlong_t)dsobj, dsname, (longlong_t)scn->scn_phys.scn_cur_min_txg, (longlong_t)scn->scn_phys.scn_max_txg); kmem_free(dsname, MAXNAMELEN); goto out; } /* * Only the ZIL in the head (non-snapshot) is valid. Even though * snapshots can have ZIL block pointers (which may be the same * BP as in the head), they must be ignored. In addition, $ORIGIN * doesn't have a objset (i.e. its ds_bp is a hole) so we don't * need to look for a ZIL in it either. So we traverse the ZIL here, * rather than in scan_recurse(), because the regular snapshot * block-sharing rules don't apply to it. */ if (!dsl_dataset_is_snapshot(ds) && (dp->dp_origin_snap == NULL || ds->ds_dir != dp->dp_origin_snap->ds_dir)) { objset_t *os; if (dmu_objset_from_ds(ds, &os) != 0) { goto out; } dsl_scan_zil(dp, &os->os_zil_header); } /* * Iterate over the bps in this ds. */ dmu_buf_will_dirty(ds->ds_dbuf, tx); rrw_enter(&ds->ds_bp_rwlock, RW_READER, FTAG); dsl_scan_visit_rootbp(scn, ds, &dsl_dataset_phys(ds)->ds_bp, tx); rrw_exit(&ds->ds_bp_rwlock, FTAG); char *dsname = kmem_alloc(ZFS_MAX_DATASET_NAME_LEN, KM_SLEEP); dsl_dataset_name(ds, dsname); zfs_dbgmsg("scanned dataset %llu (%s) with min=%llu max=%llu; " "suspending=%u", (longlong_t)dsobj, dsname, (longlong_t)scn->scn_phys.scn_cur_min_txg, (longlong_t)scn->scn_phys.scn_cur_max_txg, (int)scn->scn_suspending); kmem_free(dsname, ZFS_MAX_DATASET_NAME_LEN); if (scn->scn_suspending) goto out; /* * We've finished this pass over this dataset. */ /* * If we did not completely visit this dataset, do another pass. */ if (scn->scn_phys.scn_flags & DSF_VISIT_DS_AGAIN) { zfs_dbgmsg("incomplete pass on %s; visiting again", dp->dp_spa->spa_name); scn->scn_phys.scn_flags &= ~DSF_VISIT_DS_AGAIN; scan_ds_queue_insert(scn, ds->ds_object, scn->scn_phys.scn_cur_max_txg); goto out; } /* * Add descendant datasets to work queue. */ if (dsl_dataset_phys(ds)->ds_next_snap_obj != 0) { scan_ds_queue_insert(scn, dsl_dataset_phys(ds)->ds_next_snap_obj, dsl_dataset_phys(ds)->ds_creation_txg); } if (dsl_dataset_phys(ds)->ds_num_children > 1) { boolean_t usenext = B_FALSE; if (dsl_dataset_phys(ds)->ds_next_clones_obj != 0) { uint64_t count; /* * A bug in a previous version of the code could * cause upgrade_clones_cb() to not set * ds_next_snap_obj when it should, leading to a * missing entry. Therefore we can only use the * next_clones_obj when its count is correct. */ int err = zap_count(dp->dp_meta_objset, dsl_dataset_phys(ds)->ds_next_clones_obj, &count); if (err == 0 && count == dsl_dataset_phys(ds)->ds_num_children - 1) usenext = B_TRUE; } if (usenext) { zap_cursor_t zc; zap_attribute_t za; for (zap_cursor_init(&zc, dp->dp_meta_objset, dsl_dataset_phys(ds)->ds_next_clones_obj); zap_cursor_retrieve(&zc, &za) == 0; (void) zap_cursor_advance(&zc)) { scan_ds_queue_insert(scn, zfs_strtonum(za.za_name, NULL), dsl_dataset_phys(ds)->ds_creation_txg); } zap_cursor_fini(&zc); } else { VERIFY0(dmu_objset_find_dp(dp, dp->dp_root_dir_obj, enqueue_clones_cb, &ds->ds_object, DS_FIND_CHILDREN)); } } out: dsl_dataset_rele(ds, FTAG); } static int enqueue_cb(dsl_pool_t *dp, dsl_dataset_t *hds, void *arg) { (void) arg; dsl_dataset_t *ds; int err; dsl_scan_t *scn = dp->dp_scan; err = dsl_dataset_hold_obj(dp, hds->ds_object, FTAG, &ds); if (err) return (err); while (dsl_dataset_phys(ds)->ds_prev_snap_obj != 0) { dsl_dataset_t *prev; err = dsl_dataset_hold_obj(dp, dsl_dataset_phys(ds)->ds_prev_snap_obj, FTAG, &prev); if (err) { dsl_dataset_rele(ds, FTAG); return (err); } /* * If this is a clone, we don't need to worry about it for now. */ if (dsl_dataset_phys(prev)->ds_next_snap_obj != ds->ds_object) { dsl_dataset_rele(ds, FTAG); dsl_dataset_rele(prev, FTAG); return (0); } dsl_dataset_rele(ds, FTAG); ds = prev; } scan_ds_queue_insert(scn, ds->ds_object, dsl_dataset_phys(ds)->ds_prev_snap_txg); dsl_dataset_rele(ds, FTAG); return (0); } void dsl_scan_ddt_entry(dsl_scan_t *scn, enum zio_checksum checksum, ddt_entry_t *dde, dmu_tx_t *tx) { (void) tx; const ddt_key_t *ddk = &dde->dde_key; ddt_phys_t *ddp = dde->dde_phys; blkptr_t bp; zbookmark_phys_t zb = { 0 }; if (!dsl_scan_is_running(scn)) return; /* * This function is special because it is the only thing * that can add scan_io_t's to the vdev scan queues from * outside dsl_scan_sync(). For the most part this is ok * as long as it is called from within syncing context. * However, dsl_scan_sync() expects that no new sio's will * be added between when all the work for a scan is done * and the next txg when the scan is actually marked as * completed. This check ensures we do not issue new sio's * during this period. */ if (scn->scn_done_txg != 0) return; for (int p = 0; p < DDT_PHYS_TYPES; p++, ddp++) { if (ddp->ddp_phys_birth == 0 || ddp->ddp_phys_birth > scn->scn_phys.scn_max_txg) continue; ddt_bp_create(checksum, ddk, ddp, &bp); scn->scn_visited_this_txg++; scan_funcs[scn->scn_phys.scn_func](scn->scn_dp, &bp, &zb); } } /* * Scrub/dedup interaction. * * If there are N references to a deduped block, we don't want to scrub it * N times -- ideally, we should scrub it exactly once. * * We leverage the fact that the dde's replication class (enum ddt_class) * is ordered from highest replication class (DDT_CLASS_DITTO) to lowest * (DDT_CLASS_UNIQUE) so that we may walk the DDT in that order. * * To prevent excess scrubbing, the scrub begins by walking the DDT * to find all blocks with refcnt > 1, and scrubs each of these once. * Since there are two replication classes which contain blocks with * refcnt > 1, we scrub the highest replication class (DDT_CLASS_DITTO) first. * Finally the top-down scrub begins, only visiting blocks with refcnt == 1. * * There would be nothing more to say if a block's refcnt couldn't change * during a scrub, but of course it can so we must account for changes * in a block's replication class. * * Here's an example of what can occur: * * If a block has refcnt > 1 during the DDT scrub phase, but has refcnt == 1 * when visited during the top-down scrub phase, it will be scrubbed twice. * This negates our scrub optimization, but is otherwise harmless. * * If a block has refcnt == 1 during the DDT scrub phase, but has refcnt > 1 * on each visit during the top-down scrub phase, it will never be scrubbed. * To catch this, ddt_sync_entry() notifies the scrub code whenever a block's * reference class transitions to a higher level (i.e DDT_CLASS_UNIQUE to * DDT_CLASS_DUPLICATE); if it transitions from refcnt == 1 to refcnt > 1 * while a scrub is in progress, it scrubs the block right then. */ static void dsl_scan_ddt(dsl_scan_t *scn, dmu_tx_t *tx) { ddt_bookmark_t *ddb = &scn->scn_phys.scn_ddt_bookmark; ddt_entry_t dde = {{{{0}}}}; int error; uint64_t n = 0; while ((error = ddt_walk(scn->scn_dp->dp_spa, ddb, &dde)) == 0) { ddt_t *ddt; if (ddb->ddb_class > scn->scn_phys.scn_ddt_class_max) break; dprintf("visiting ddb=%llu/%llu/%llu/%llx\n", (longlong_t)ddb->ddb_class, (longlong_t)ddb->ddb_type, (longlong_t)ddb->ddb_checksum, (longlong_t)ddb->ddb_cursor); /* There should be no pending changes to the dedup table */ ddt = scn->scn_dp->dp_spa->spa_ddt[ddb->ddb_checksum]; ASSERT(avl_first(&ddt->ddt_tree) == NULL); dsl_scan_ddt_entry(scn, ddb->ddb_checksum, &dde, tx); n++; if (dsl_scan_check_suspend(scn, NULL)) break; } zfs_dbgmsg("scanned %llu ddt entries on %s with class_max = %u; " "suspending=%u", (longlong_t)n, scn->scn_dp->dp_spa->spa_name, (int)scn->scn_phys.scn_ddt_class_max, (int)scn->scn_suspending); ASSERT(error == 0 || error == ENOENT); ASSERT(error != ENOENT || ddb->ddb_class > scn->scn_phys.scn_ddt_class_max); } static uint64_t dsl_scan_ds_maxtxg(dsl_dataset_t *ds) { uint64_t smt = ds->ds_dir->dd_pool->dp_scan->scn_phys.scn_max_txg; if (ds->ds_is_snapshot) return (MIN(smt, dsl_dataset_phys(ds)->ds_creation_txg)); return (smt); } static void dsl_scan_visit(dsl_scan_t *scn, dmu_tx_t *tx) { scan_ds_t *sds; dsl_pool_t *dp = scn->scn_dp; if (scn->scn_phys.scn_ddt_bookmark.ddb_class <= scn->scn_phys.scn_ddt_class_max) { scn->scn_phys.scn_cur_min_txg = scn->scn_phys.scn_min_txg; scn->scn_phys.scn_cur_max_txg = scn->scn_phys.scn_max_txg; dsl_scan_ddt(scn, tx); if (scn->scn_suspending) return; } if (scn->scn_phys.scn_bookmark.zb_objset == DMU_META_OBJSET) { /* First do the MOS & ORIGIN */ scn->scn_phys.scn_cur_min_txg = scn->scn_phys.scn_min_txg; scn->scn_phys.scn_cur_max_txg = scn->scn_phys.scn_max_txg; dsl_scan_visit_rootbp(scn, NULL, &dp->dp_meta_rootbp, tx); spa_set_rootblkptr(dp->dp_spa, &dp->dp_meta_rootbp); if (scn->scn_suspending) return; if (spa_version(dp->dp_spa) < SPA_VERSION_DSL_SCRUB) { VERIFY0(dmu_objset_find_dp(dp, dp->dp_root_dir_obj, enqueue_cb, NULL, DS_FIND_CHILDREN)); } else { dsl_scan_visitds(scn, dp->dp_origin_snap->ds_object, tx); } ASSERT(!scn->scn_suspending); } else if (scn->scn_phys.scn_bookmark.zb_objset != ZB_DESTROYED_OBJSET) { uint64_t dsobj = scn->scn_phys.scn_bookmark.zb_objset; /* * If we were suspended, continue from here. Note if the * ds we were suspended on was deleted, the zb_objset may * be -1, so we will skip this and find a new objset * below. */ dsl_scan_visitds(scn, dsobj, tx); if (scn->scn_suspending) return; } /* * In case we suspended right at the end of the ds, zero the * bookmark so we don't think that we're still trying to resume. */ memset(&scn->scn_phys.scn_bookmark, 0, sizeof (zbookmark_phys_t)); /* * Keep pulling things out of the dataset avl queue. Updates to the * persistent zap-object-as-queue happen only at checkpoints. */ while ((sds = avl_first(&scn->scn_queue)) != NULL) { dsl_dataset_t *ds; uint64_t dsobj = sds->sds_dsobj; uint64_t txg = sds->sds_txg; /* dequeue and free the ds from the queue */ scan_ds_queue_remove(scn, dsobj); sds = NULL; /* set up min / max txg */ VERIFY3U(0, ==, dsl_dataset_hold_obj(dp, dsobj, FTAG, &ds)); if (txg != 0) { scn->scn_phys.scn_cur_min_txg = MAX(scn->scn_phys.scn_min_txg, txg); } else { scn->scn_phys.scn_cur_min_txg = MAX(scn->scn_phys.scn_min_txg, dsl_dataset_phys(ds)->ds_prev_snap_txg); } scn->scn_phys.scn_cur_max_txg = dsl_scan_ds_maxtxg(ds); dsl_dataset_rele(ds, FTAG); dsl_scan_visitds(scn, dsobj, tx); if (scn->scn_suspending) return; } /* No more objsets to fetch, we're done */ scn->scn_phys.scn_bookmark.zb_objset = ZB_DESTROYED_OBJSET; ASSERT0(scn->scn_suspending); } static uint64_t dsl_scan_count_data_disks(spa_t *spa) { vdev_t *rvd = spa->spa_root_vdev; uint64_t i, leaves = 0; for (i = 0; i < rvd->vdev_children; i++) { vdev_t *vd = rvd->vdev_child[i]; if (vd->vdev_islog || vd->vdev_isspare || vd->vdev_isl2cache) continue; leaves += vdev_get_ndisks(vd) - vdev_get_nparity(vd); } return (leaves); } static void scan_io_queues_update_zio_stats(dsl_scan_io_queue_t *q, const blkptr_t *bp) { int i; uint64_t cur_size = 0; for (i = 0; i < BP_GET_NDVAS(bp); i++) { cur_size += DVA_GET_ASIZE(&bp->blk_dva[i]); } q->q_total_zio_size_this_txg += cur_size; q->q_zios_this_txg++; } static void scan_io_queues_update_seg_stats(dsl_scan_io_queue_t *q, uint64_t start, uint64_t end) { q->q_total_seg_size_this_txg += end - start; q->q_segs_this_txg++; } static boolean_t scan_io_queue_check_suspend(dsl_scan_t *scn) { /* See comment in dsl_scan_check_suspend() */ uint64_t curr_time_ns = gethrtime(); uint64_t scan_time_ns = curr_time_ns - scn->scn_sync_start_time; uint64_t sync_time_ns = curr_time_ns - scn->scn_dp->dp_spa->spa_sync_starttime; uint64_t dirty_min_bytes = zfs_dirty_data_max * zfs_vdev_async_write_active_min_dirty_percent / 100; uint_t mintime = (scn->scn_phys.scn_func == POOL_SCAN_RESILVER) ? zfs_resilver_min_time_ms : zfs_scrub_min_time_ms; return ((NSEC2MSEC(scan_time_ns) > mintime && (scn->scn_dp->dp_dirty_total >= dirty_min_bytes || txg_sync_waiting(scn->scn_dp) || NSEC2SEC(sync_time_ns) >= zfs_txg_timeout)) || spa_shutting_down(scn->scn_dp->dp_spa)); } /* * Given a list of scan_io_t's in io_list, this issues the I/Os out to * disk. This consumes the io_list and frees the scan_io_t's. This is * called when emptying queues, either when we're up against the memory * limit or when we have finished scanning. Returns B_TRUE if we stopped * processing the list before we finished. Any sios that were not issued * will remain in the io_list. */ static boolean_t scan_io_queue_issue(dsl_scan_io_queue_t *queue, list_t *io_list) { dsl_scan_t *scn = queue->q_scn; scan_io_t *sio; boolean_t suspended = B_FALSE; while ((sio = list_head(io_list)) != NULL) { blkptr_t bp; if (scan_io_queue_check_suspend(scn)) { suspended = B_TRUE; break; } sio2bp(sio, &bp); scan_exec_io(scn->scn_dp, &bp, sio->sio_flags, &sio->sio_zb, queue); (void) list_remove_head(io_list); scan_io_queues_update_zio_stats(queue, &bp); sio_free(sio); } return (suspended); } /* * This function removes sios from an IO queue which reside within a given * range_seg_t and inserts them (in offset order) into a list. Note that * we only ever return a maximum of 32 sios at once. If there are more sios * to process within this segment that did not make it onto the list we * return B_TRUE and otherwise B_FALSE. */ static boolean_t scan_io_queue_gather(dsl_scan_io_queue_t *queue, range_seg_t *rs, list_t *list) { scan_io_t *srch_sio, *sio, *next_sio; avl_index_t idx; uint_t num_sios = 0; int64_t bytes_issued = 0; ASSERT(rs != NULL); ASSERT(MUTEX_HELD(&queue->q_vd->vdev_scan_io_queue_lock)); srch_sio = sio_alloc(1); srch_sio->sio_nr_dvas = 1; SIO_SET_OFFSET(srch_sio, rs_get_start(rs, queue->q_exts_by_addr)); /* * The exact start of the extent might not contain any matching zios, * so if that's the case, examine the next one in the tree. */ sio = avl_find(&queue->q_sios_by_addr, srch_sio, &idx); sio_free(srch_sio); if (sio == NULL) sio = avl_nearest(&queue->q_sios_by_addr, idx, AVL_AFTER); while (sio != NULL && SIO_GET_OFFSET(sio) < rs_get_end(rs, queue->q_exts_by_addr) && num_sios <= 32) { ASSERT3U(SIO_GET_OFFSET(sio), >=, rs_get_start(rs, queue->q_exts_by_addr)); ASSERT3U(SIO_GET_END_OFFSET(sio), <=, rs_get_end(rs, queue->q_exts_by_addr)); next_sio = AVL_NEXT(&queue->q_sios_by_addr, sio); avl_remove(&queue->q_sios_by_addr, sio); if (avl_is_empty(&queue->q_sios_by_addr)) atomic_add_64(&queue->q_scn->scn_queues_pending, -1); queue->q_sio_memused -= SIO_GET_MUSED(sio); bytes_issued += SIO_GET_ASIZE(sio); num_sios++; list_insert_tail(list, sio); sio = next_sio; } /* * We limit the number of sios we process at once to 32 to avoid * biting off more than we can chew. If we didn't take everything * in the segment we update it to reflect the work we were able to * complete. Otherwise, we remove it from the range tree entirely. */ if (sio != NULL && SIO_GET_OFFSET(sio) < rs_get_end(rs, queue->q_exts_by_addr)) { range_tree_adjust_fill(queue->q_exts_by_addr, rs, -bytes_issued); range_tree_resize_segment(queue->q_exts_by_addr, rs, SIO_GET_OFFSET(sio), rs_get_end(rs, queue->q_exts_by_addr) - SIO_GET_OFFSET(sio)); queue->q_last_ext_addr = SIO_GET_OFFSET(sio); return (B_TRUE); } else { uint64_t rstart = rs_get_start(rs, queue->q_exts_by_addr); uint64_t rend = rs_get_end(rs, queue->q_exts_by_addr); range_tree_remove(queue->q_exts_by_addr, rstart, rend - rstart); queue->q_last_ext_addr = -1; return (B_FALSE); } } /* * This is called from the queue emptying thread and selects the next * extent from which we are to issue I/Os. The behavior of this function * depends on the state of the scan, the current memory consumption and * whether or not we are performing a scan shutdown. * 1) We select extents in an elevator algorithm (LBA-order) if the scan * needs to perform a checkpoint * 2) We select the largest available extent if we are up against the * memory limit. * 3) Otherwise we don't select any extents. */ static range_seg_t * scan_io_queue_fetch_ext(dsl_scan_io_queue_t *queue) { dsl_scan_t *scn = queue->q_scn; range_tree_t *rt = queue->q_exts_by_addr; ASSERT(MUTEX_HELD(&queue->q_vd->vdev_scan_io_queue_lock)); ASSERT(scn->scn_is_sorted); if (!scn->scn_checkpointing && !scn->scn_clearing) return (NULL); /* * During normal clearing, we want to issue our largest segments * first, keeping IO as sequential as possible, and leaving the * smaller extents for later with the hope that they might eventually * grow to larger sequential segments. However, when the scan is * checkpointing, no new extents will be added to the sorting queue, * so the way we are sorted now is as good as it will ever get. * In this case, we instead switch to issuing extents in LBA order. */ if ((zfs_scan_issue_strategy < 1 && scn->scn_checkpointing) || zfs_scan_issue_strategy == 1) return (range_tree_first(rt)); /* * Try to continue previous extent if it is not completed yet. After * shrink in scan_io_queue_gather() it may no longer be the best, but * otherwise we leave shorter remnant every txg. */ uint64_t start; uint64_t size = 1ULL << rt->rt_shift; range_seg_t *addr_rs; if (queue->q_last_ext_addr != -1) { start = queue->q_last_ext_addr; addr_rs = range_tree_find(rt, start, size); if (addr_rs != NULL) return (addr_rs); } /* * Nothing to continue, so find new best extent. */ uint64_t *v = zfs_btree_first(&queue->q_exts_by_size, NULL); if (v == NULL) return (NULL); queue->q_last_ext_addr = start = *v << rt->rt_shift; /* * We need to get the original entry in the by_addr tree so we can * modify it. */ addr_rs = range_tree_find(rt, start, size); ASSERT3P(addr_rs, !=, NULL); ASSERT3U(rs_get_start(addr_rs, rt), ==, start); ASSERT3U(rs_get_end(addr_rs, rt), >, start); return (addr_rs); } static void scan_io_queues_run_one(void *arg) { dsl_scan_io_queue_t *queue = arg; kmutex_t *q_lock = &queue->q_vd->vdev_scan_io_queue_lock; boolean_t suspended = B_FALSE; range_seg_t *rs; scan_io_t *sio; zio_t *zio; list_t sio_list; ASSERT(queue->q_scn->scn_is_sorted); list_create(&sio_list, sizeof (scan_io_t), offsetof(scan_io_t, sio_nodes.sio_list_node)); zio = zio_null(queue->q_scn->scn_zio_root, queue->q_scn->scn_dp->dp_spa, NULL, NULL, NULL, ZIO_FLAG_CANFAIL); mutex_enter(q_lock); queue->q_zio = zio; /* Calculate maximum in-flight bytes for this vdev. */ queue->q_maxinflight_bytes = MAX(1, zfs_scan_vdev_limit * (vdev_get_ndisks(queue->q_vd) - vdev_get_nparity(queue->q_vd))); /* reset per-queue scan statistics for this txg */ queue->q_total_seg_size_this_txg = 0; queue->q_segs_this_txg = 0; queue->q_total_zio_size_this_txg = 0; queue->q_zios_this_txg = 0; /* loop until we run out of time or sios */ while ((rs = scan_io_queue_fetch_ext(queue)) != NULL) { uint64_t seg_start = 0, seg_end = 0; boolean_t more_left; ASSERT(list_is_empty(&sio_list)); /* loop while we still have sios left to process in this rs */ do { scan_io_t *first_sio, *last_sio; /* * We have selected which extent needs to be * processed next. Gather up the corresponding sios. */ more_left = scan_io_queue_gather(queue, rs, &sio_list); ASSERT(!list_is_empty(&sio_list)); first_sio = list_head(&sio_list); last_sio = list_tail(&sio_list); seg_end = SIO_GET_END_OFFSET(last_sio); if (seg_start == 0) seg_start = SIO_GET_OFFSET(first_sio); /* * Issuing sios can take a long time so drop the * queue lock. The sio queue won't be updated by * other threads since we're in syncing context so * we can be sure that our trees will remain exactly * as we left them. */ mutex_exit(q_lock); suspended = scan_io_queue_issue(queue, &sio_list); mutex_enter(q_lock); if (suspended) break; } while (more_left); /* update statistics for debugging purposes */ scan_io_queues_update_seg_stats(queue, seg_start, seg_end); if (suspended) break; } /* * If we were suspended in the middle of processing, * requeue any unfinished sios and exit. */ while ((sio = list_head(&sio_list)) != NULL) { list_remove(&sio_list, sio); scan_io_queue_insert_impl(queue, sio); } queue->q_zio = NULL; mutex_exit(q_lock); zio_nowait(zio); list_destroy(&sio_list); } /* * Performs an emptying run on all scan queues in the pool. This just * punches out one thread per top-level vdev, each of which processes * only that vdev's scan queue. We can parallelize the I/O here because * we know that each queue's I/Os only affect its own top-level vdev. * * This function waits for the queue runs to complete, and must be * called from dsl_scan_sync (or in general, syncing context). */ static void scan_io_queues_run(dsl_scan_t *scn) { spa_t *spa = scn->scn_dp->dp_spa; ASSERT(scn->scn_is_sorted); ASSERT(spa_config_held(spa, SCL_CONFIG, RW_READER)); if (scn->scn_queues_pending == 0) return; if (scn->scn_taskq == NULL) { int nthreads = spa->spa_root_vdev->vdev_children; /* * We need to make this taskq *always* execute as many * threads in parallel as we have top-level vdevs and no * less, otherwise strange serialization of the calls to * scan_io_queues_run_one can occur during spa_sync runs * and that significantly impacts performance. */ scn->scn_taskq = taskq_create("dsl_scan_iss", nthreads, minclsyspri, nthreads, nthreads, TASKQ_PREPOPULATE); } for (uint64_t i = 0; i < spa->spa_root_vdev->vdev_children; i++) { vdev_t *vd = spa->spa_root_vdev->vdev_child[i]; mutex_enter(&vd->vdev_scan_io_queue_lock); if (vd->vdev_scan_io_queue != NULL) { VERIFY(taskq_dispatch(scn->scn_taskq, scan_io_queues_run_one, vd->vdev_scan_io_queue, TQ_SLEEP) != TASKQID_INVALID); } mutex_exit(&vd->vdev_scan_io_queue_lock); } /* * Wait for the queues to finish issuing their IOs for this run * before we return. There may still be IOs in flight at this * point. */ taskq_wait(scn->scn_taskq); } static boolean_t dsl_scan_async_block_should_pause(dsl_scan_t *scn) { uint64_t elapsed_nanosecs; if (zfs_recover) return (B_FALSE); if (zfs_async_block_max_blocks != 0 && scn->scn_visited_this_txg >= zfs_async_block_max_blocks) { return (B_TRUE); } if (zfs_max_async_dedup_frees != 0 && scn->scn_dedup_frees_this_txg >= zfs_max_async_dedup_frees) { return (B_TRUE); } elapsed_nanosecs = gethrtime() - scn->scn_sync_start_time; return (elapsed_nanosecs / NANOSEC > zfs_txg_timeout || (NSEC2MSEC(elapsed_nanosecs) > scn->scn_async_block_min_time_ms && txg_sync_waiting(scn->scn_dp)) || spa_shutting_down(scn->scn_dp->dp_spa)); } static int dsl_scan_free_block_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx) { dsl_scan_t *scn = arg; if (!scn->scn_is_bptree || (BP_GET_LEVEL(bp) == 0 && BP_GET_TYPE(bp) != DMU_OT_OBJSET)) { if (dsl_scan_async_block_should_pause(scn)) return (SET_ERROR(ERESTART)); } zio_nowait(zio_free_sync(scn->scn_zio_root, scn->scn_dp->dp_spa, dmu_tx_get_txg(tx), bp, 0)); dsl_dir_diduse_space(tx->tx_pool->dp_free_dir, DD_USED_HEAD, -bp_get_dsize_sync(scn->scn_dp->dp_spa, bp), -BP_GET_PSIZE(bp), -BP_GET_UCSIZE(bp), tx); scn->scn_visited_this_txg++; if (BP_GET_DEDUP(bp)) scn->scn_dedup_frees_this_txg++; return (0); } static void dsl_scan_update_stats(dsl_scan_t *scn) { spa_t *spa = scn->scn_dp->dp_spa; uint64_t i; uint64_t seg_size_total = 0, zio_size_total = 0; uint64_t seg_count_total = 0, zio_count_total = 0; for (i = 0; i < spa->spa_root_vdev->vdev_children; i++) { vdev_t *vd = spa->spa_root_vdev->vdev_child[i]; dsl_scan_io_queue_t *queue = vd->vdev_scan_io_queue; if (queue == NULL) continue; seg_size_total += queue->q_total_seg_size_this_txg; zio_size_total += queue->q_total_zio_size_this_txg; seg_count_total += queue->q_segs_this_txg; zio_count_total += queue->q_zios_this_txg; } if (seg_count_total == 0 || zio_count_total == 0) { scn->scn_avg_seg_size_this_txg = 0; scn->scn_avg_zio_size_this_txg = 0; scn->scn_segs_this_txg = 0; scn->scn_zios_this_txg = 0; return; } scn->scn_avg_seg_size_this_txg = seg_size_total / seg_count_total; scn->scn_avg_zio_size_this_txg = zio_size_total / zio_count_total; scn->scn_segs_this_txg = seg_count_total; scn->scn_zios_this_txg = zio_count_total; } static int bpobj_dsl_scan_free_block_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed, dmu_tx_t *tx) { ASSERT(!bp_freed); return (dsl_scan_free_block_cb(arg, bp, tx)); } static int dsl_scan_obsolete_block_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed, dmu_tx_t *tx) { ASSERT(!bp_freed); dsl_scan_t *scn = arg; const dva_t *dva = &bp->blk_dva[0]; if (dsl_scan_async_block_should_pause(scn)) return (SET_ERROR(ERESTART)); spa_vdev_indirect_mark_obsolete(scn->scn_dp->dp_spa, DVA_GET_VDEV(dva), DVA_GET_OFFSET(dva), DVA_GET_ASIZE(dva), tx); scn->scn_visited_this_txg++; return (0); } boolean_t dsl_scan_active(dsl_scan_t *scn) { spa_t *spa = scn->scn_dp->dp_spa; uint64_t used = 0, comp, uncomp; boolean_t clones_left; if (spa->spa_load_state != SPA_LOAD_NONE) return (B_FALSE); if (spa_shutting_down(spa)) return (B_FALSE); if ((dsl_scan_is_running(scn) && !dsl_scan_is_paused_scrub(scn)) || (scn->scn_async_destroying && !scn->scn_async_stalled)) return (B_TRUE); if (spa_version(scn->scn_dp->dp_spa) >= SPA_VERSION_DEADLISTS) { (void) bpobj_space(&scn->scn_dp->dp_free_bpobj, &used, &comp, &uncomp); } clones_left = spa_livelist_delete_check(spa); return ((used != 0) || (clones_left)); } +boolean_t +dsl_errorscrub_active(dsl_scan_t *scn) +{ + spa_t *spa = scn->scn_dp->dp_spa; + if (spa->spa_load_state != SPA_LOAD_NONE) + return (B_FALSE); + if (spa_shutting_down(spa)) + return (B_FALSE); + if (dsl_errorscrubbing(scn->scn_dp)) + return (B_TRUE); + return (B_FALSE); +} + static boolean_t dsl_scan_check_deferred(vdev_t *vd) { boolean_t need_resilver = B_FALSE; for (int c = 0; c < vd->vdev_children; c++) { need_resilver |= dsl_scan_check_deferred(vd->vdev_child[c]); } if (!vdev_is_concrete(vd) || vd->vdev_aux || !vd->vdev_ops->vdev_op_leaf) return (need_resilver); if (!vd->vdev_resilver_deferred) need_resilver = B_TRUE; return (need_resilver); } static boolean_t dsl_scan_need_resilver(spa_t *spa, const dva_t *dva, size_t psize, uint64_t phys_birth) { vdev_t *vd; vd = vdev_lookup_top(spa, DVA_GET_VDEV(dva)); if (vd->vdev_ops == &vdev_indirect_ops) { /* * The indirect vdev can point to multiple * vdevs. For simplicity, always create * the resilver zio_t. zio_vdev_io_start() * will bypass the child resilver i/o's if * they are on vdevs that don't have DTL's. */ return (B_TRUE); } if (DVA_GET_GANG(dva)) { /* * Gang members may be spread across multiple * vdevs, so the best estimate we have is the * scrub range, which has already been checked. * XXX -- it would be better to change our * allocation policy to ensure that all * gang members reside on the same vdev. */ return (B_TRUE); } /* * Check if the top-level vdev must resilver this offset. * When the offset does not intersect with a dirty leaf DTL * then it may be possible to skip the resilver IO. The psize * is provided instead of asize to simplify the check for RAIDZ. */ if (!vdev_dtl_need_resilver(vd, dva, psize, phys_birth)) return (B_FALSE); /* * Check that this top-level vdev has a device under it which * is resilvering and is not deferred. */ if (!dsl_scan_check_deferred(vd)) return (B_FALSE); return (B_TRUE); } static int dsl_process_async_destroys(dsl_pool_t *dp, dmu_tx_t *tx) { dsl_scan_t *scn = dp->dp_scan; spa_t *spa = dp->dp_spa; int err = 0; if (spa_suspend_async_destroy(spa)) return (0); if (zfs_free_bpobj_enabled && spa_version(spa) >= SPA_VERSION_DEADLISTS) { scn->scn_is_bptree = B_FALSE; scn->scn_async_block_min_time_ms = zfs_free_min_time_ms; scn->scn_zio_root = zio_root(spa, NULL, NULL, ZIO_FLAG_MUSTSUCCEED); err = bpobj_iterate(&dp->dp_free_bpobj, bpobj_dsl_scan_free_block_cb, scn, tx); VERIFY0(zio_wait(scn->scn_zio_root)); scn->scn_zio_root = NULL; if (err != 0 && err != ERESTART) zfs_panic_recover("error %u from bpobj_iterate()", err); } if (err == 0 && spa_feature_is_active(spa, SPA_FEATURE_ASYNC_DESTROY)) { ASSERT(scn->scn_async_destroying); scn->scn_is_bptree = B_TRUE; scn->scn_zio_root = zio_root(spa, NULL, NULL, ZIO_FLAG_MUSTSUCCEED); err = bptree_iterate(dp->dp_meta_objset, dp->dp_bptree_obj, B_TRUE, dsl_scan_free_block_cb, scn, tx); VERIFY0(zio_wait(scn->scn_zio_root)); scn->scn_zio_root = NULL; if (err == EIO || err == ECKSUM) { err = 0; } else if (err != 0 && err != ERESTART) { zfs_panic_recover("error %u from " "traverse_dataset_destroyed()", err); } if (bptree_is_empty(dp->dp_meta_objset, dp->dp_bptree_obj)) { /* finished; deactivate async destroy feature */ spa_feature_decr(spa, SPA_FEATURE_ASYNC_DESTROY, tx); ASSERT(!spa_feature_is_active(spa, SPA_FEATURE_ASYNC_DESTROY)); VERIFY0(zap_remove(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_BPTREE_OBJ, tx)); VERIFY0(bptree_free(dp->dp_meta_objset, dp->dp_bptree_obj, tx)); dp->dp_bptree_obj = 0; scn->scn_async_destroying = B_FALSE; scn->scn_async_stalled = B_FALSE; } else { /* * If we didn't make progress, mark the async * destroy as stalled, so that we will not initiate * a spa_sync() on its behalf. Note that we only * check this if we are not finished, because if the * bptree had no blocks for us to visit, we can * finish without "making progress". */ scn->scn_async_stalled = (scn->scn_visited_this_txg == 0); } } if (scn->scn_visited_this_txg) { zfs_dbgmsg("freed %llu blocks in %llums from " "free_bpobj/bptree on %s in txg %llu; err=%u", (longlong_t)scn->scn_visited_this_txg, (longlong_t) NSEC2MSEC(gethrtime() - scn->scn_sync_start_time), spa->spa_name, (longlong_t)tx->tx_txg, err); scn->scn_visited_this_txg = 0; scn->scn_dedup_frees_this_txg = 0; /* * Write out changes to the DDT and the BRT that may be required * as a result of the blocks freed. This ensures that the DDT * and the BRT are clean when a scrub/resilver runs. */ ddt_sync(spa, tx->tx_txg); brt_sync(spa, tx->tx_txg); } if (err != 0) return (err); if (dp->dp_free_dir != NULL && !scn->scn_async_destroying && zfs_free_leak_on_eio && (dsl_dir_phys(dp->dp_free_dir)->dd_used_bytes != 0 || dsl_dir_phys(dp->dp_free_dir)->dd_compressed_bytes != 0 || dsl_dir_phys(dp->dp_free_dir)->dd_uncompressed_bytes != 0)) { /* * We have finished background destroying, but there is still * some space left in the dp_free_dir. Transfer this leaked * space to the dp_leak_dir. */ if (dp->dp_leak_dir == NULL) { rrw_enter(&dp->dp_config_rwlock, RW_WRITER, FTAG); (void) dsl_dir_create_sync(dp, dp->dp_root_dir, LEAK_DIR_NAME, tx); VERIFY0(dsl_pool_open_special_dir(dp, LEAK_DIR_NAME, &dp->dp_leak_dir)); rrw_exit(&dp->dp_config_rwlock, FTAG); } dsl_dir_diduse_space(dp->dp_leak_dir, DD_USED_HEAD, dsl_dir_phys(dp->dp_free_dir)->dd_used_bytes, dsl_dir_phys(dp->dp_free_dir)->dd_compressed_bytes, dsl_dir_phys(dp->dp_free_dir)->dd_uncompressed_bytes, tx); dsl_dir_diduse_space(dp->dp_free_dir, DD_USED_HEAD, -dsl_dir_phys(dp->dp_free_dir)->dd_used_bytes, -dsl_dir_phys(dp->dp_free_dir)->dd_compressed_bytes, -dsl_dir_phys(dp->dp_free_dir)->dd_uncompressed_bytes, tx); } if (dp->dp_free_dir != NULL && !scn->scn_async_destroying && !spa_livelist_delete_check(spa)) { /* finished; verify that space accounting went to zero */ ASSERT0(dsl_dir_phys(dp->dp_free_dir)->dd_used_bytes); ASSERT0(dsl_dir_phys(dp->dp_free_dir)->dd_compressed_bytes); ASSERT0(dsl_dir_phys(dp->dp_free_dir)->dd_uncompressed_bytes); } spa_notify_waiters(spa); EQUIV(bpobj_is_open(&dp->dp_obsolete_bpobj), 0 == zap_contains(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_OBSOLETE_BPOBJ)); if (err == 0 && bpobj_is_open(&dp->dp_obsolete_bpobj)) { ASSERT(spa_feature_is_active(dp->dp_spa, SPA_FEATURE_OBSOLETE_COUNTS)); scn->scn_is_bptree = B_FALSE; scn->scn_async_block_min_time_ms = zfs_obsolete_min_time_ms; err = bpobj_iterate(&dp->dp_obsolete_bpobj, dsl_scan_obsolete_block_cb, scn, tx); if (err != 0 && err != ERESTART) zfs_panic_recover("error %u from bpobj_iterate()", err); if (bpobj_is_empty(&dp->dp_obsolete_bpobj)) dsl_pool_destroy_obsolete_bpobj(dp, tx); } return (0); } +static void +name_to_bookmark(char *buf, zbookmark_phys_t *zb) +{ + zb->zb_objset = zfs_strtonum(buf, &buf); + ASSERT(*buf == ':'); + zb->zb_object = zfs_strtonum(buf + 1, &buf); + ASSERT(*buf == ':'); + zb->zb_level = (int)zfs_strtonum(buf + 1, &buf); + ASSERT(*buf == ':'); + zb->zb_blkid = zfs_strtonum(buf + 1, &buf); + ASSERT(*buf == '\0'); +} + +static void +name_to_object(char *buf, uint64_t *obj) +{ + *obj = zfs_strtonum(buf, &buf); + ASSERT(*buf == '\0'); +} + +static void +read_by_block_level(dsl_scan_t *scn, zbookmark_phys_t zb) +{ + dsl_pool_t *dp = scn->scn_dp; + dsl_dataset_t *ds; + objset_t *os; + if (dsl_dataset_hold_obj(dp, zb.zb_objset, FTAG, &ds) != 0) + return; + + if (dmu_objset_from_ds(ds, &os) != 0) { + dsl_dataset_rele(ds, FTAG); + return; + } + + /* + * If the key is not loaded dbuf_dnode_findbp() will error out with + * EACCES. However in that case dnode_hold() will eventually call + * dbuf_read()->zio_wait() which may call spa_log_error(). This will + * lead to a deadlock due to us holding the mutex spa_errlist_lock. + * Avoid this by checking here if the keys are loaded, if not return. + * If the keys are not loaded the head_errlog feature is meaningless + * as we cannot figure out the birth txg of the block pointer. + */ + if (dsl_dataset_get_keystatus(ds->ds_dir) == + ZFS_KEYSTATUS_UNAVAILABLE) { + dsl_dataset_rele(ds, FTAG); + return; + } + + dnode_t *dn; + blkptr_t bp; + + if (dnode_hold(os, zb.zb_object, FTAG, &dn) != 0) { + dsl_dataset_rele(ds, FTAG); + return; + } + + rw_enter(&dn->dn_struct_rwlock, RW_READER); + int error = dbuf_dnode_findbp(dn, zb.zb_level, zb.zb_blkid, &bp, NULL, + NULL); + + if (error) { + rw_exit(&dn->dn_struct_rwlock); + dnode_rele(dn, FTAG); + dsl_dataset_rele(ds, FTAG); + return; + } + + if (!error && BP_IS_HOLE(&bp)) { + rw_exit(&dn->dn_struct_rwlock); + dnode_rele(dn, FTAG); + dsl_dataset_rele(ds, FTAG); + return; + } + + int zio_flags = ZIO_FLAG_SCAN_THREAD | ZIO_FLAG_RAW | + ZIO_FLAG_CANFAIL | ZIO_FLAG_SCRUB; + + /* If it's an intent log block, failure is expected. */ + if (zb.zb_level == ZB_ZIL_LEVEL) + zio_flags |= ZIO_FLAG_SPECULATIVE; + + ASSERT(!BP_IS_EMBEDDED(&bp)); + scan_exec_io(dp, &bp, zio_flags, &zb, NULL); + rw_exit(&dn->dn_struct_rwlock); + dnode_rele(dn, FTAG); + dsl_dataset_rele(ds, FTAG); +} + +/* + * We keep track of the scrubbed error blocks in "count". This will be used + * when deciding whether we exceeded zfs_scrub_error_blocks_per_txg. This + * function is modelled after check_filesystem(). + */ +static int +scrub_filesystem(spa_t *spa, uint64_t fs, zbookmark_err_phys_t *zep, + int *count) +{ + dsl_dataset_t *ds; + dsl_pool_t *dp = spa->spa_dsl_pool; + dsl_scan_t *scn = dp->dp_scan; + + int error = dsl_dataset_hold_obj(dp, fs, FTAG, &ds); + if (error != 0) + return (error); + + uint64_t latest_txg; + uint64_t txg_to_consider = spa->spa_syncing_txg; + boolean_t check_snapshot = B_TRUE; + + error = find_birth_txg(ds, zep, &latest_txg); + + /* + * If find_birth_txg() errors out, then err on the side of caution and + * proceed. In worst case scenario scrub all objects. If zep->zb_birth + * is 0 (e.g. in case of encryption with unloaded keys) also proceed to + * scrub all objects. + */ + if (error == 0 && zep->zb_birth == latest_txg) { + /* Block neither free nor re written. */ + zbookmark_phys_t zb; + zep_to_zb(fs, zep, &zb); + scn->scn_zio_root = zio_root(spa, NULL, NULL, + ZIO_FLAG_CANFAIL); + /* We have already acquired the config lock for spa */ + read_by_block_level(scn, zb); + + (void) zio_wait(scn->scn_zio_root); + scn->scn_zio_root = NULL; + + scn->errorscrub_phys.dep_examined++; + scn->errorscrub_phys.dep_to_examine--; + (*count)++; + if ((*count) == zfs_scrub_error_blocks_per_txg || + dsl_error_scrub_check_suspend(scn, &zb)) { + dsl_dataset_rele(ds, FTAG); + return (SET_ERROR(EFAULT)); + } + + check_snapshot = B_FALSE; + } else if (error == 0) { + txg_to_consider = latest_txg; + } + + /* + * Retrieve the number of snapshots if the dataset is not a snapshot. + */ + uint64_t snap_count = 0; + if (dsl_dataset_phys(ds)->ds_snapnames_zapobj != 0) { + + error = zap_count(spa->spa_meta_objset, + dsl_dataset_phys(ds)->ds_snapnames_zapobj, &snap_count); + + if (error != 0) { + dsl_dataset_rele(ds, FTAG); + return (error); + } + } + + if (snap_count == 0) { + /* Filesystem without snapshots. */ + dsl_dataset_rele(ds, FTAG); + return (0); + } + + uint64_t snap_obj = dsl_dataset_phys(ds)->ds_prev_snap_obj; + uint64_t snap_obj_txg = dsl_dataset_phys(ds)->ds_prev_snap_txg; + + dsl_dataset_rele(ds, FTAG); + + /* Check only snapshots created from this file system. */ + while (snap_obj != 0 && zep->zb_birth < snap_obj_txg && + snap_obj_txg <= txg_to_consider) { + + error = dsl_dataset_hold_obj(dp, snap_obj, FTAG, &ds); + if (error != 0) + return (error); + + if (dsl_dir_phys(ds->ds_dir)->dd_head_dataset_obj != fs) { + snap_obj = dsl_dataset_phys(ds)->ds_prev_snap_obj; + snap_obj_txg = dsl_dataset_phys(ds)->ds_prev_snap_txg; + dsl_dataset_rele(ds, FTAG); + continue; + } + + boolean_t affected = B_TRUE; + if (check_snapshot) { + uint64_t blk_txg; + error = find_birth_txg(ds, zep, &blk_txg); + + /* + * Scrub the snapshot also when zb_birth == 0 or when + * find_birth_txg() returns an error. + */ + affected = (error == 0 && zep->zb_birth == blk_txg) || + (error != 0) || (zep->zb_birth == 0); + } + + /* Scrub snapshots. */ + if (affected) { + zbookmark_phys_t zb; + zep_to_zb(snap_obj, zep, &zb); + scn->scn_zio_root = zio_root(spa, NULL, NULL, + ZIO_FLAG_CANFAIL); + /* We have already acquired the config lock for spa */ + read_by_block_level(scn, zb); + + (void) zio_wait(scn->scn_zio_root); + scn->scn_zio_root = NULL; + + scn->errorscrub_phys.dep_examined++; + scn->errorscrub_phys.dep_to_examine--; + (*count)++; + if ((*count) == zfs_scrub_error_blocks_per_txg || + dsl_error_scrub_check_suspend(scn, &zb)) { + dsl_dataset_rele(ds, FTAG); + return (EFAULT); + } + } + snap_obj_txg = dsl_dataset_phys(ds)->ds_prev_snap_txg; + snap_obj = dsl_dataset_phys(ds)->ds_prev_snap_obj; + dsl_dataset_rele(ds, FTAG); + } + return (0); +} + +void +dsl_errorscrub_sync(dsl_pool_t *dp, dmu_tx_t *tx) +{ + spa_t *spa = dp->dp_spa; + dsl_scan_t *scn = dp->dp_scan; + + /* + * Only process scans in sync pass 1. + */ + + if (spa_sync_pass(spa) > 1) + return; + + /* + * If the spa is shutting down, then stop scanning. This will + * ensure that the scan does not dirty any new data during the + * shutdown phase. + */ + if (spa_shutting_down(spa)) + return; + + if (!dsl_errorscrub_active(scn) || dsl_errorscrub_is_paused(scn)) { + return; + } + + if (dsl_scan_resilvering(scn->scn_dp)) { + /* cancel the error scrub if resilver started */ + dsl_scan_cancel(scn->scn_dp); + return; + } + + spa->spa_scrub_active = B_TRUE; + scn->scn_sync_start_time = gethrtime(); + + /* + * zfs_scan_suspend_progress can be set to disable scrub progress. + * See more detailed comment in dsl_scan_sync(). + */ + if (zfs_scan_suspend_progress) { + uint64_t scan_time_ns = gethrtime() - scn->scn_sync_start_time; + int mintime = zfs_scrub_min_time_ms; + + while (zfs_scan_suspend_progress && + !txg_sync_waiting(scn->scn_dp) && + !spa_shutting_down(scn->scn_dp->dp_spa) && + NSEC2MSEC(scan_time_ns) < mintime) { + delay(hz); + scan_time_ns = gethrtime() - scn->scn_sync_start_time; + } + return; + } + + int i = 0; + zap_attribute_t *za; + zbookmark_phys_t *zb; + boolean_t limit_exceeded = B_FALSE; + + za = kmem_zalloc(sizeof (zap_attribute_t), KM_SLEEP); + zb = kmem_zalloc(sizeof (zbookmark_phys_t), KM_SLEEP); + + if (!spa_feature_is_enabled(spa, SPA_FEATURE_HEAD_ERRLOG)) { + for (; zap_cursor_retrieve(&scn->errorscrub_cursor, za) == 0; + zap_cursor_advance(&scn->errorscrub_cursor)) { + name_to_bookmark(za->za_name, zb); + + scn->scn_zio_root = zio_root(dp->dp_spa, NULL, + NULL, ZIO_FLAG_CANFAIL); + dsl_pool_config_enter(dp, FTAG); + read_by_block_level(scn, *zb); + dsl_pool_config_exit(dp, FTAG); + + (void) zio_wait(scn->scn_zio_root); + scn->scn_zio_root = NULL; + + scn->errorscrub_phys.dep_examined += 1; + scn->errorscrub_phys.dep_to_examine -= 1; + i++; + if (i == zfs_scrub_error_blocks_per_txg || + dsl_error_scrub_check_suspend(scn, zb)) { + limit_exceeded = B_TRUE; + break; + } + } + + if (!limit_exceeded) + dsl_errorscrub_done(scn, B_TRUE, tx); + + dsl_errorscrub_sync_state(scn, tx); + kmem_free(za, sizeof (*za)); + kmem_free(zb, sizeof (*zb)); + return; + } + + int error = 0; + for (; zap_cursor_retrieve(&scn->errorscrub_cursor, za) == 0; + zap_cursor_advance(&scn->errorscrub_cursor)) { + + zap_cursor_t *head_ds_cursor; + zap_attribute_t *head_ds_attr; + zbookmark_err_phys_t head_ds_block; + + head_ds_cursor = kmem_zalloc(sizeof (zap_cursor_t), KM_SLEEP); + head_ds_attr = kmem_zalloc(sizeof (zap_attribute_t), KM_SLEEP); + + uint64_t head_ds_err_obj = za->za_first_integer; + uint64_t head_ds; + name_to_object(za->za_name, &head_ds); + boolean_t config_held = B_FALSE; + uint64_t top_affected_fs; + + for (zap_cursor_init(head_ds_cursor, spa->spa_meta_objset, + head_ds_err_obj); zap_cursor_retrieve(head_ds_cursor, + head_ds_attr) == 0; zap_cursor_advance(head_ds_cursor)) { + + name_to_errphys(head_ds_attr->za_name, &head_ds_block); + + /* + * In case we are called from spa_sync the pool + * config is already held. + */ + if (!dsl_pool_config_held(dp)) { + dsl_pool_config_enter(dp, FTAG); + config_held = B_TRUE; + } + + error = find_top_affected_fs(spa, + head_ds, &head_ds_block, &top_affected_fs); + if (error) + break; + + error = scrub_filesystem(spa, top_affected_fs, + &head_ds_block, &i); + + if (error == SET_ERROR(EFAULT)) { + limit_exceeded = B_TRUE; + break; + } + } + + zap_cursor_fini(head_ds_cursor); + kmem_free(head_ds_cursor, sizeof (*head_ds_cursor)); + kmem_free(head_ds_attr, sizeof (*head_ds_attr)); + + if (config_held) + dsl_pool_config_exit(dp, FTAG); + } + + kmem_free(za, sizeof (*za)); + kmem_free(zb, sizeof (*zb)); + if (!limit_exceeded) + dsl_errorscrub_done(scn, B_TRUE, tx); + + dsl_errorscrub_sync_state(scn, tx); +} + /* * This is the primary entry point for scans that is called from syncing * context. Scans must happen entirely during syncing context so that we * can guarantee that blocks we are currently scanning will not change out * from under us. While a scan is active, this function controls how quickly * transaction groups proceed, instead of the normal handling provided by * txg_sync_thread(). */ void dsl_scan_sync(dsl_pool_t *dp, dmu_tx_t *tx) { int err = 0; dsl_scan_t *scn = dp->dp_scan; spa_t *spa = dp->dp_spa; state_sync_type_t sync_type = SYNC_OPTIONAL; if (spa->spa_resilver_deferred && !spa_feature_is_active(dp->dp_spa, SPA_FEATURE_RESILVER_DEFER)) spa_feature_incr(spa, SPA_FEATURE_RESILVER_DEFER, tx); /* * Check for scn_restart_txg before checking spa_load_state, so * that we can restart an old-style scan while the pool is being * imported (see dsl_scan_init). We also restart scans if there * is a deferred resilver and the user has manually disabled * deferred resilvers via the tunable. */ if (dsl_scan_restarting(scn, tx) || (spa->spa_resilver_deferred && zfs_resilver_disable_defer)) { pool_scan_func_t func = POOL_SCAN_SCRUB; dsl_scan_done(scn, B_FALSE, tx); if (vdev_resilver_needed(spa->spa_root_vdev, NULL, NULL)) func = POOL_SCAN_RESILVER; zfs_dbgmsg("restarting scan func=%u on %s txg=%llu", func, dp->dp_spa->spa_name, (longlong_t)tx->tx_txg); dsl_scan_setup_sync(&func, tx); } /* * Only process scans in sync pass 1. */ if (spa_sync_pass(spa) > 1) return; /* * If the spa is shutting down, then stop scanning. This will * ensure that the scan does not dirty any new data during the * shutdown phase. */ if (spa_shutting_down(spa)) return; /* * If the scan is inactive due to a stalled async destroy, try again. */ if (!scn->scn_async_stalled && !dsl_scan_active(scn)) return; /* reset scan statistics */ scn->scn_visited_this_txg = 0; scn->scn_dedup_frees_this_txg = 0; scn->scn_holes_this_txg = 0; scn->scn_lt_min_this_txg = 0; scn->scn_gt_max_this_txg = 0; scn->scn_ddt_contained_this_txg = 0; scn->scn_objsets_visited_this_txg = 0; scn->scn_avg_seg_size_this_txg = 0; scn->scn_segs_this_txg = 0; scn->scn_avg_zio_size_this_txg = 0; scn->scn_zios_this_txg = 0; scn->scn_suspending = B_FALSE; scn->scn_sync_start_time = gethrtime(); spa->spa_scrub_active = B_TRUE; /* * First process the async destroys. If we suspend, don't do * any scrubbing or resilvering. This ensures that there are no * async destroys while we are scanning, so the scan code doesn't * have to worry about traversing it. It is also faster to free the * blocks than to scrub them. */ err = dsl_process_async_destroys(dp, tx); if (err != 0) return; if (!dsl_scan_is_running(scn) || dsl_scan_is_paused_scrub(scn)) return; /* * Wait a few txgs after importing to begin scanning so that * we can get the pool imported quickly. */ if (spa->spa_syncing_txg < spa->spa_first_txg + SCAN_IMPORT_WAIT_TXGS) return; /* * zfs_scan_suspend_progress can be set to disable scan progress. * We don't want to spin the txg_sync thread, so we add a delay * here to simulate the time spent doing a scan. This is mostly * useful for testing and debugging. */ if (zfs_scan_suspend_progress) { uint64_t scan_time_ns = gethrtime() - scn->scn_sync_start_time; uint_t mintime = (scn->scn_phys.scn_func == POOL_SCAN_RESILVER) ? zfs_resilver_min_time_ms : zfs_scrub_min_time_ms; while (zfs_scan_suspend_progress && !txg_sync_waiting(scn->scn_dp) && !spa_shutting_down(scn->scn_dp->dp_spa) && NSEC2MSEC(scan_time_ns) < mintime) { delay(hz); scan_time_ns = gethrtime() - scn->scn_sync_start_time; } return; } /* * Disabled by default, set zfs_scan_report_txgs to report * average performance over the last zfs_scan_report_txgs TXGs. */ if (!dsl_scan_is_paused_scrub(scn) && zfs_scan_report_txgs != 0 && tx->tx_txg % zfs_scan_report_txgs == 0) { scn->scn_issued_before_pass += spa->spa_scan_pass_issued; spa_scan_stat_init(spa); } /* * It is possible to switch from unsorted to sorted at any time, * but afterwards the scan will remain sorted unless reloaded from * a checkpoint after a reboot. */ if (!zfs_scan_legacy) { scn->scn_is_sorted = B_TRUE; if (scn->scn_last_checkpoint == 0) scn->scn_last_checkpoint = ddi_get_lbolt(); } /* * For sorted scans, determine what kind of work we will be doing * this txg based on our memory limitations and whether or not we * need to perform a checkpoint. */ if (scn->scn_is_sorted) { /* * If we are over our checkpoint interval, set scn_clearing * so that we can begin checkpointing immediately. The * checkpoint allows us to save a consistent bookmark * representing how much data we have scrubbed so far. * Otherwise, use the memory limit to determine if we should * scan for metadata or start issue scrub IOs. We accumulate * metadata until we hit our hard memory limit at which point * we issue scrub IOs until we are at our soft memory limit. */ if (scn->scn_checkpointing || ddi_get_lbolt() - scn->scn_last_checkpoint > SEC_TO_TICK(zfs_scan_checkpoint_intval)) { if (!scn->scn_checkpointing) zfs_dbgmsg("begin scan checkpoint for %s", spa->spa_name); scn->scn_checkpointing = B_TRUE; scn->scn_clearing = B_TRUE; } else { boolean_t should_clear = dsl_scan_should_clear(scn); if (should_clear && !scn->scn_clearing) { zfs_dbgmsg("begin scan clearing for %s", spa->spa_name); scn->scn_clearing = B_TRUE; } else if (!should_clear && scn->scn_clearing) { zfs_dbgmsg("finish scan clearing for %s", spa->spa_name); scn->scn_clearing = B_FALSE; } } } else { ASSERT0(scn->scn_checkpointing); ASSERT0(scn->scn_clearing); } if (!scn->scn_clearing && scn->scn_done_txg == 0) { /* Need to scan metadata for more blocks to scrub */ dsl_scan_phys_t *scnp = &scn->scn_phys; taskqid_t prefetch_tqid; /* * Calculate the max number of in-flight bytes for pool-wide * scanning operations (minimum 1MB, maximum 1/4 of arc_c_max). * Limits for the issuing phase are done per top-level vdev and * are handled separately. */ scn->scn_maxinflight_bytes = MIN(arc_c_max / 4, MAX(1ULL << 20, zfs_scan_vdev_limit * dsl_scan_count_data_disks(spa))); if (scnp->scn_ddt_bookmark.ddb_class <= scnp->scn_ddt_class_max) { ASSERT(ZB_IS_ZERO(&scnp->scn_bookmark)); zfs_dbgmsg("doing scan sync for %s txg %llu; " "ddt bm=%llu/%llu/%llu/%llx", spa->spa_name, (longlong_t)tx->tx_txg, (longlong_t)scnp->scn_ddt_bookmark.ddb_class, (longlong_t)scnp->scn_ddt_bookmark.ddb_type, (longlong_t)scnp->scn_ddt_bookmark.ddb_checksum, (longlong_t)scnp->scn_ddt_bookmark.ddb_cursor); } else { zfs_dbgmsg("doing scan sync for %s txg %llu; " "bm=%llu/%llu/%llu/%llu", spa->spa_name, (longlong_t)tx->tx_txg, (longlong_t)scnp->scn_bookmark.zb_objset, (longlong_t)scnp->scn_bookmark.zb_object, (longlong_t)scnp->scn_bookmark.zb_level, (longlong_t)scnp->scn_bookmark.zb_blkid); } scn->scn_zio_root = zio_root(dp->dp_spa, NULL, NULL, ZIO_FLAG_CANFAIL); scn->scn_prefetch_stop = B_FALSE; prefetch_tqid = taskq_dispatch(dp->dp_sync_taskq, dsl_scan_prefetch_thread, scn, TQ_SLEEP); ASSERT(prefetch_tqid != TASKQID_INVALID); dsl_pool_config_enter(dp, FTAG); dsl_scan_visit(scn, tx); dsl_pool_config_exit(dp, FTAG); mutex_enter(&dp->dp_spa->spa_scrub_lock); scn->scn_prefetch_stop = B_TRUE; cv_broadcast(&spa->spa_scrub_io_cv); mutex_exit(&dp->dp_spa->spa_scrub_lock); taskq_wait_id(dp->dp_sync_taskq, prefetch_tqid); (void) zio_wait(scn->scn_zio_root); scn->scn_zio_root = NULL; zfs_dbgmsg("scan visited %llu blocks of %s in %llums " "(%llu os's, %llu holes, %llu < mintxg, " "%llu in ddt, %llu > maxtxg)", (longlong_t)scn->scn_visited_this_txg, spa->spa_name, (longlong_t)NSEC2MSEC(gethrtime() - scn->scn_sync_start_time), (longlong_t)scn->scn_objsets_visited_this_txg, (longlong_t)scn->scn_holes_this_txg, (longlong_t)scn->scn_lt_min_this_txg, (longlong_t)scn->scn_ddt_contained_this_txg, (longlong_t)scn->scn_gt_max_this_txg); if (!scn->scn_suspending) { ASSERT0(avl_numnodes(&scn->scn_queue)); scn->scn_done_txg = tx->tx_txg + 1; if (scn->scn_is_sorted) { scn->scn_checkpointing = B_TRUE; scn->scn_clearing = B_TRUE; scn->scn_issued_before_pass += spa->spa_scan_pass_issued; spa_scan_stat_init(spa); } zfs_dbgmsg("scan complete for %s txg %llu", spa->spa_name, (longlong_t)tx->tx_txg); } } else if (scn->scn_is_sorted && scn->scn_queues_pending != 0) { ASSERT(scn->scn_clearing); /* need to issue scrubbing IOs from per-vdev queues */ scn->scn_zio_root = zio_root(dp->dp_spa, NULL, NULL, ZIO_FLAG_CANFAIL); scan_io_queues_run(scn); (void) zio_wait(scn->scn_zio_root); scn->scn_zio_root = NULL; /* calculate and dprintf the current memory usage */ (void) dsl_scan_should_clear(scn); dsl_scan_update_stats(scn); zfs_dbgmsg("scan issued %llu blocks for %s (%llu segs) " "in %llums (avg_block_size = %llu, avg_seg_size = %llu)", (longlong_t)scn->scn_zios_this_txg, spa->spa_name, (longlong_t)scn->scn_segs_this_txg, (longlong_t)NSEC2MSEC(gethrtime() - scn->scn_sync_start_time), (longlong_t)scn->scn_avg_zio_size_this_txg, (longlong_t)scn->scn_avg_seg_size_this_txg); } else if (scn->scn_done_txg != 0 && scn->scn_done_txg <= tx->tx_txg) { /* Finished with everything. Mark the scrub as complete */ zfs_dbgmsg("scan issuing complete txg %llu for %s", (longlong_t)tx->tx_txg, spa->spa_name); ASSERT3U(scn->scn_done_txg, !=, 0); ASSERT0(spa->spa_scrub_inflight); ASSERT0(scn->scn_queues_pending); dsl_scan_done(scn, B_TRUE, tx); sync_type = SYNC_MANDATORY; } dsl_scan_sync_state(scn, tx, sync_type); } static void count_block_issued(spa_t *spa, const blkptr_t *bp, boolean_t all) { /* * Don't count embedded bp's, since we already did the work of * scanning these when we scanned the containing block. */ if (BP_IS_EMBEDDED(bp)) return; /* * Update the spa's stats on how many bytes we have issued. * Sequential scrubs create a zio for each DVA of the bp. Each * of these will include all DVAs for repair purposes, but the * zio code will only try the first one unless there is an issue. * Therefore, we should only count the first DVA for these IOs. */ atomic_add_64(&spa->spa_scan_pass_issued, all ? BP_GET_ASIZE(bp) : DVA_GET_ASIZE(&bp->blk_dva[0])); } static void count_block(zfs_all_blkstats_t *zab, const blkptr_t *bp) { /* * If we resume after a reboot, zab will be NULL; don't record * incomplete stats in that case. */ if (zab == NULL) return; for (int i = 0; i < 4; i++) { int l = (i < 2) ? BP_GET_LEVEL(bp) : DN_MAX_LEVELS; int t = (i & 1) ? BP_GET_TYPE(bp) : DMU_OT_TOTAL; if (t & DMU_OT_NEWTYPE) t = DMU_OT_OTHER; zfs_blkstat_t *zb = &zab->zab_type[l][t]; int equal; zb->zb_count++; zb->zb_asize += BP_GET_ASIZE(bp); zb->zb_lsize += BP_GET_LSIZE(bp); zb->zb_psize += BP_GET_PSIZE(bp); zb->zb_gangs += BP_COUNT_GANG(bp); switch (BP_GET_NDVAS(bp)) { case 2: if (DVA_GET_VDEV(&bp->blk_dva[0]) == DVA_GET_VDEV(&bp->blk_dva[1])) zb->zb_ditto_2_of_2_samevdev++; break; case 3: equal = (DVA_GET_VDEV(&bp->blk_dva[0]) == DVA_GET_VDEV(&bp->blk_dva[1])) + (DVA_GET_VDEV(&bp->blk_dva[0]) == DVA_GET_VDEV(&bp->blk_dva[2])) + (DVA_GET_VDEV(&bp->blk_dva[1]) == DVA_GET_VDEV(&bp->blk_dva[2])); if (equal == 1) zb->zb_ditto_2_of_3_samevdev++; else if (equal == 3) zb->zb_ditto_3_of_3_samevdev++; break; } } } static void scan_io_queue_insert_impl(dsl_scan_io_queue_t *queue, scan_io_t *sio) { avl_index_t idx; dsl_scan_t *scn = queue->q_scn; ASSERT(MUTEX_HELD(&queue->q_vd->vdev_scan_io_queue_lock)); if (unlikely(avl_is_empty(&queue->q_sios_by_addr))) atomic_add_64(&scn->scn_queues_pending, 1); if (avl_find(&queue->q_sios_by_addr, sio, &idx) != NULL) { /* block is already scheduled for reading */ sio_free(sio); return; } avl_insert(&queue->q_sios_by_addr, sio, idx); queue->q_sio_memused += SIO_GET_MUSED(sio); range_tree_add(queue->q_exts_by_addr, SIO_GET_OFFSET(sio), SIO_GET_ASIZE(sio)); } /* * Given all the info we got from our metadata scanning process, we * construct a scan_io_t and insert it into the scan sorting queue. The * I/O must already be suitable for us to process. This is controlled * by dsl_scan_enqueue(). */ static void scan_io_queue_insert(dsl_scan_io_queue_t *queue, const blkptr_t *bp, int dva_i, int zio_flags, const zbookmark_phys_t *zb) { scan_io_t *sio = sio_alloc(BP_GET_NDVAS(bp)); ASSERT0(BP_IS_GANG(bp)); ASSERT(MUTEX_HELD(&queue->q_vd->vdev_scan_io_queue_lock)); bp2sio(bp, sio, dva_i); sio->sio_flags = zio_flags; sio->sio_zb = *zb; queue->q_last_ext_addr = -1; scan_io_queue_insert_impl(queue, sio); } /* * Given a set of I/O parameters as discovered by the metadata traversal * process, attempts to place the I/O into the sorted queues (if allowed), * or immediately executes the I/O. */ static void dsl_scan_enqueue(dsl_pool_t *dp, const blkptr_t *bp, int zio_flags, const zbookmark_phys_t *zb) { spa_t *spa = dp->dp_spa; ASSERT(!BP_IS_EMBEDDED(bp)); /* * Gang blocks are hard to issue sequentially, so we just issue them * here immediately instead of queuing them. */ if (!dp->dp_scan->scn_is_sorted || BP_IS_GANG(bp)) { scan_exec_io(dp, bp, zio_flags, zb, NULL); return; } for (int i = 0; i < BP_GET_NDVAS(bp); i++) { dva_t dva; vdev_t *vdev; dva = bp->blk_dva[i]; vdev = vdev_lookup_top(spa, DVA_GET_VDEV(&dva)); ASSERT(vdev != NULL); mutex_enter(&vdev->vdev_scan_io_queue_lock); if (vdev->vdev_scan_io_queue == NULL) vdev->vdev_scan_io_queue = scan_io_queue_create(vdev); ASSERT(dp->dp_scan != NULL); scan_io_queue_insert(vdev->vdev_scan_io_queue, bp, i, zio_flags, zb); mutex_exit(&vdev->vdev_scan_io_queue_lock); } } static int dsl_scan_scrub_cb(dsl_pool_t *dp, const blkptr_t *bp, const zbookmark_phys_t *zb) { dsl_scan_t *scn = dp->dp_scan; spa_t *spa = dp->dp_spa; uint64_t phys_birth = BP_PHYSICAL_BIRTH(bp); size_t psize = BP_GET_PSIZE(bp); boolean_t needs_io = B_FALSE; int zio_flags = ZIO_FLAG_SCAN_THREAD | ZIO_FLAG_RAW | ZIO_FLAG_CANFAIL; count_block(dp->dp_blkstats, bp); if (phys_birth <= scn->scn_phys.scn_min_txg || phys_birth >= scn->scn_phys.scn_max_txg) { count_block_issued(spa, bp, B_TRUE); return (0); } /* Embedded BP's have phys_birth==0, so we reject them above. */ ASSERT(!BP_IS_EMBEDDED(bp)); ASSERT(DSL_SCAN_IS_SCRUB_RESILVER(scn)); if (scn->scn_phys.scn_func == POOL_SCAN_SCRUB) { zio_flags |= ZIO_FLAG_SCRUB; needs_io = B_TRUE; } else { ASSERT3U(scn->scn_phys.scn_func, ==, POOL_SCAN_RESILVER); zio_flags |= ZIO_FLAG_RESILVER; needs_io = B_FALSE; } /* If it's an intent log block, failure is expected. */ if (zb->zb_level == ZB_ZIL_LEVEL) zio_flags |= ZIO_FLAG_SPECULATIVE; for (int d = 0; d < BP_GET_NDVAS(bp); d++) { const dva_t *dva = &bp->blk_dva[d]; /* * Keep track of how much data we've examined so that * zpool(8) status can make useful progress reports. */ uint64_t asize = DVA_GET_ASIZE(dva); scn->scn_phys.scn_examined += asize; spa->spa_scan_pass_exam += asize; /* if it's a resilver, this may not be in the target range */ if (!needs_io) needs_io = dsl_scan_need_resilver(spa, dva, psize, phys_birth); } if (needs_io && !zfs_no_scrub_io) { dsl_scan_enqueue(dp, bp, zio_flags, zb); } else { count_block_issued(spa, bp, B_TRUE); } /* do not relocate this block */ return (0); } static void dsl_scan_scrub_done(zio_t *zio) { spa_t *spa = zio->io_spa; blkptr_t *bp = zio->io_bp; dsl_scan_io_queue_t *queue = zio->io_private; abd_free(zio->io_abd); if (queue == NULL) { mutex_enter(&spa->spa_scrub_lock); ASSERT3U(spa->spa_scrub_inflight, >=, BP_GET_PSIZE(bp)); spa->spa_scrub_inflight -= BP_GET_PSIZE(bp); cv_broadcast(&spa->spa_scrub_io_cv); mutex_exit(&spa->spa_scrub_lock); } else { mutex_enter(&queue->q_vd->vdev_scan_io_queue_lock); ASSERT3U(queue->q_inflight_bytes, >=, BP_GET_PSIZE(bp)); queue->q_inflight_bytes -= BP_GET_PSIZE(bp); cv_broadcast(&queue->q_zio_cv); mutex_exit(&queue->q_vd->vdev_scan_io_queue_lock); } if (zio->io_error && (zio->io_error != ECKSUM || !(zio->io_flags & ZIO_FLAG_SPECULATIVE))) { - atomic_inc_64(&spa->spa_dsl_pool->dp_scan->scn_phys.scn_errors); + if (dsl_errorscrubbing(spa->spa_dsl_pool) && + !dsl_errorscrub_is_paused(spa->spa_dsl_pool->dp_scan)) { + atomic_inc_64(&spa->spa_dsl_pool->dp_scan + ->errorscrub_phys.dep_errors); + } else { + atomic_inc_64(&spa->spa_dsl_pool->dp_scan->scn_phys + .scn_errors); + } } } /* * Given a scanning zio's information, executes the zio. The zio need * not necessarily be only sortable, this function simply executes the * zio, no matter what it is. The optional queue argument allows the * caller to specify that they want per top level vdev IO rate limiting * instead of the legacy global limiting. */ static void scan_exec_io(dsl_pool_t *dp, const blkptr_t *bp, int zio_flags, const zbookmark_phys_t *zb, dsl_scan_io_queue_t *queue) { spa_t *spa = dp->dp_spa; dsl_scan_t *scn = dp->dp_scan; size_t size = BP_GET_PSIZE(bp); abd_t *data = abd_alloc_for_io(size, B_FALSE); zio_t *pio; if (queue == NULL) { ASSERT3U(scn->scn_maxinflight_bytes, >, 0); mutex_enter(&spa->spa_scrub_lock); while (spa->spa_scrub_inflight >= scn->scn_maxinflight_bytes) cv_wait(&spa->spa_scrub_io_cv, &spa->spa_scrub_lock); spa->spa_scrub_inflight += BP_GET_PSIZE(bp); mutex_exit(&spa->spa_scrub_lock); pio = scn->scn_zio_root; } else { kmutex_t *q_lock = &queue->q_vd->vdev_scan_io_queue_lock; ASSERT3U(queue->q_maxinflight_bytes, >, 0); mutex_enter(q_lock); while (queue->q_inflight_bytes >= queue->q_maxinflight_bytes) cv_wait(&queue->q_zio_cv, q_lock); queue->q_inflight_bytes += BP_GET_PSIZE(bp); pio = queue->q_zio; mutex_exit(q_lock); } ASSERT(pio != NULL); count_block_issued(spa, bp, queue == NULL); zio_nowait(zio_read(pio, spa, bp, data, size, dsl_scan_scrub_done, queue, ZIO_PRIORITY_SCRUB, zio_flags, zb)); } /* * This is the primary extent sorting algorithm. We balance two parameters: * 1) how many bytes of I/O are in an extent * 2) how well the extent is filled with I/O (as a fraction of its total size) * Since we allow extents to have gaps between their constituent I/Os, it's * possible to have a fairly large extent that contains the same amount of * I/O bytes than a much smaller extent, which just packs the I/O more tightly. * The algorithm sorts based on a score calculated from the extent's size, * the relative fill volume (in %) and a "fill weight" parameter that controls * the split between whether we prefer larger extents or more well populated * extents: * * SCORE = FILL_IN_BYTES + (FILL_IN_PERCENT * FILL_IN_BYTES * FILL_WEIGHT) * * Example: * 1) assume extsz = 64 MiB * 2) assume fill = 32 MiB (extent is half full) * 3) assume fill_weight = 3 * 4) SCORE = 32M + (((32M * 100) / 64M) * 3 * 32M) / 100 * SCORE = 32M + (50 * 3 * 32M) / 100 * SCORE = 32M + (4800M / 100) * SCORE = 32M + 48M * ^ ^ * | +--- final total relative fill-based score * +--------- final total fill-based score * SCORE = 80M * * As can be seen, at fill_ratio=3, the algorithm is slightly biased towards * extents that are more completely filled (in a 3:2 ratio) vs just larger. * Note that as an optimization, we replace multiplication and division by * 100 with bitshifting by 7 (which effectively multiplies and divides by 128). * * Since we do not care if one extent is only few percent better than another, * compress the score into 6 bits via binary logarithm AKA highbit64() and * put into otherwise unused due to ashift high bits of offset. This allows * to reduce q_exts_by_size B-tree elements to only 64 bits and compare them * with single operation. Plus it makes scrubs more sequential and reduces * chances that minor extent change move it within the B-tree. */ static int ext_size_compare(const void *x, const void *y) { const uint64_t *a = x, *b = y; return (TREE_CMP(*a, *b)); } static void ext_size_create(range_tree_t *rt, void *arg) { (void) rt; zfs_btree_t *size_tree = arg; zfs_btree_create(size_tree, ext_size_compare, sizeof (uint64_t)); } static void ext_size_destroy(range_tree_t *rt, void *arg) { (void) rt; zfs_btree_t *size_tree = arg; ASSERT0(zfs_btree_numnodes(size_tree)); zfs_btree_destroy(size_tree); } static uint64_t ext_size_value(range_tree_t *rt, range_seg_gap_t *rsg) { (void) rt; uint64_t size = rsg->rs_end - rsg->rs_start; uint64_t score = rsg->rs_fill + ((((rsg->rs_fill << 7) / size) * fill_weight * rsg->rs_fill) >> 7); ASSERT3U(rt->rt_shift, >=, 8); return (((uint64_t)(64 - highbit64(score)) << 56) | rsg->rs_start); } static void ext_size_add(range_tree_t *rt, range_seg_t *rs, void *arg) { zfs_btree_t *size_tree = arg; ASSERT3U(rt->rt_type, ==, RANGE_SEG_GAP); uint64_t v = ext_size_value(rt, (range_seg_gap_t *)rs); zfs_btree_add(size_tree, &v); } static void ext_size_remove(range_tree_t *rt, range_seg_t *rs, void *arg) { zfs_btree_t *size_tree = arg; ASSERT3U(rt->rt_type, ==, RANGE_SEG_GAP); uint64_t v = ext_size_value(rt, (range_seg_gap_t *)rs); zfs_btree_remove(size_tree, &v); } static void ext_size_vacate(range_tree_t *rt, void *arg) { zfs_btree_t *size_tree = arg; zfs_btree_clear(size_tree); zfs_btree_destroy(size_tree); ext_size_create(rt, arg); } static const range_tree_ops_t ext_size_ops = { .rtop_create = ext_size_create, .rtop_destroy = ext_size_destroy, .rtop_add = ext_size_add, .rtop_remove = ext_size_remove, .rtop_vacate = ext_size_vacate }; /* * Comparator for the q_sios_by_addr tree. Sorting is simply performed * based on LBA-order (from lowest to highest). */ static int sio_addr_compare(const void *x, const void *y) { const scan_io_t *a = x, *b = y; return (TREE_CMP(SIO_GET_OFFSET(a), SIO_GET_OFFSET(b))); } /* IO queues are created on demand when they are needed. */ static dsl_scan_io_queue_t * scan_io_queue_create(vdev_t *vd) { dsl_scan_t *scn = vd->vdev_spa->spa_dsl_pool->dp_scan; dsl_scan_io_queue_t *q = kmem_zalloc(sizeof (*q), KM_SLEEP); q->q_scn = scn; q->q_vd = vd; q->q_sio_memused = 0; q->q_last_ext_addr = -1; cv_init(&q->q_zio_cv, NULL, CV_DEFAULT, NULL); q->q_exts_by_addr = range_tree_create_gap(&ext_size_ops, RANGE_SEG_GAP, &q->q_exts_by_size, 0, vd->vdev_ashift, zfs_scan_max_ext_gap); avl_create(&q->q_sios_by_addr, sio_addr_compare, sizeof (scan_io_t), offsetof(scan_io_t, sio_nodes.sio_addr_node)); return (q); } /* * Destroys a scan queue and all segments and scan_io_t's contained in it. * No further execution of I/O occurs, anything pending in the queue is * simply freed without being executed. */ void dsl_scan_io_queue_destroy(dsl_scan_io_queue_t *queue) { dsl_scan_t *scn = queue->q_scn; scan_io_t *sio; void *cookie = NULL; ASSERT(MUTEX_HELD(&queue->q_vd->vdev_scan_io_queue_lock)); if (!avl_is_empty(&queue->q_sios_by_addr)) atomic_add_64(&scn->scn_queues_pending, -1); while ((sio = avl_destroy_nodes(&queue->q_sios_by_addr, &cookie)) != NULL) { ASSERT(range_tree_contains(queue->q_exts_by_addr, SIO_GET_OFFSET(sio), SIO_GET_ASIZE(sio))); queue->q_sio_memused -= SIO_GET_MUSED(sio); sio_free(sio); } ASSERT0(queue->q_sio_memused); range_tree_vacate(queue->q_exts_by_addr, NULL, queue); range_tree_destroy(queue->q_exts_by_addr); avl_destroy(&queue->q_sios_by_addr); cv_destroy(&queue->q_zio_cv); kmem_free(queue, sizeof (*queue)); } /* * Properly transfers a dsl_scan_queue_t from `svd' to `tvd'. This is * called on behalf of vdev_top_transfer when creating or destroying * a mirror vdev due to zpool attach/detach. */ void dsl_scan_io_queue_vdev_xfer(vdev_t *svd, vdev_t *tvd) { mutex_enter(&svd->vdev_scan_io_queue_lock); mutex_enter(&tvd->vdev_scan_io_queue_lock); VERIFY3P(tvd->vdev_scan_io_queue, ==, NULL); tvd->vdev_scan_io_queue = svd->vdev_scan_io_queue; svd->vdev_scan_io_queue = NULL; if (tvd->vdev_scan_io_queue != NULL) tvd->vdev_scan_io_queue->q_vd = tvd; mutex_exit(&tvd->vdev_scan_io_queue_lock); mutex_exit(&svd->vdev_scan_io_queue_lock); } static void scan_io_queues_destroy(dsl_scan_t *scn) { vdev_t *rvd = scn->scn_dp->dp_spa->spa_root_vdev; for (uint64_t i = 0; i < rvd->vdev_children; i++) { vdev_t *tvd = rvd->vdev_child[i]; mutex_enter(&tvd->vdev_scan_io_queue_lock); if (tvd->vdev_scan_io_queue != NULL) dsl_scan_io_queue_destroy(tvd->vdev_scan_io_queue); tvd->vdev_scan_io_queue = NULL; mutex_exit(&tvd->vdev_scan_io_queue_lock); } } static void dsl_scan_freed_dva(spa_t *spa, const blkptr_t *bp, int dva_i) { dsl_pool_t *dp = spa->spa_dsl_pool; dsl_scan_t *scn = dp->dp_scan; vdev_t *vdev; kmutex_t *q_lock; dsl_scan_io_queue_t *queue; scan_io_t *srch_sio, *sio; avl_index_t idx; uint64_t start, size; vdev = vdev_lookup_top(spa, DVA_GET_VDEV(&bp->blk_dva[dva_i])); ASSERT(vdev != NULL); q_lock = &vdev->vdev_scan_io_queue_lock; queue = vdev->vdev_scan_io_queue; mutex_enter(q_lock); if (queue == NULL) { mutex_exit(q_lock); return; } srch_sio = sio_alloc(BP_GET_NDVAS(bp)); bp2sio(bp, srch_sio, dva_i); start = SIO_GET_OFFSET(srch_sio); size = SIO_GET_ASIZE(srch_sio); /* * We can find the zio in two states: * 1) Cold, just sitting in the queue of zio's to be issued at * some point in the future. In this case, all we do is * remove the zio from the q_sios_by_addr tree, decrement * its data volume from the containing range_seg_t and * resort the q_exts_by_size tree to reflect that the * range_seg_t has lost some of its 'fill'. We don't shorten * the range_seg_t - this is usually rare enough not to be * worth the extra hassle of trying keep track of precise * extent boundaries. * 2) Hot, where the zio is currently in-flight in * dsl_scan_issue_ios. In this case, we can't simply * reach in and stop the in-flight zio's, so we instead * block the caller. Eventually, dsl_scan_issue_ios will * be done with issuing the zio's it gathered and will * signal us. */ sio = avl_find(&queue->q_sios_by_addr, srch_sio, &idx); sio_free(srch_sio); if (sio != NULL) { blkptr_t tmpbp; /* Got it while it was cold in the queue */ ASSERT3U(start, ==, SIO_GET_OFFSET(sio)); ASSERT3U(size, ==, SIO_GET_ASIZE(sio)); avl_remove(&queue->q_sios_by_addr, sio); if (avl_is_empty(&queue->q_sios_by_addr)) atomic_add_64(&scn->scn_queues_pending, -1); queue->q_sio_memused -= SIO_GET_MUSED(sio); ASSERT(range_tree_contains(queue->q_exts_by_addr, start, size)); range_tree_remove_fill(queue->q_exts_by_addr, start, size); /* count the block as though we issued it */ sio2bp(sio, &tmpbp); count_block_issued(spa, &tmpbp, B_FALSE); sio_free(sio); } mutex_exit(q_lock); } /* * Callback invoked when a zio_free() zio is executing. This needs to be * intercepted to prevent the zio from deallocating a particular portion * of disk space and it then getting reallocated and written to, while we * still have it queued up for processing. */ void dsl_scan_freed(spa_t *spa, const blkptr_t *bp) { dsl_pool_t *dp = spa->spa_dsl_pool; dsl_scan_t *scn = dp->dp_scan; ASSERT(!BP_IS_EMBEDDED(bp)); ASSERT(scn != NULL); if (!dsl_scan_is_running(scn)) return; for (int i = 0; i < BP_GET_NDVAS(bp); i++) dsl_scan_freed_dva(spa, bp, i); } /* * Check if a vdev needs resilvering (non-empty DTL), if so, and resilver has * not started, start it. Otherwise, only restart if max txg in DTL range is * greater than the max txg in the current scan. If the DTL max is less than * the scan max, then the vdev has not missed any new data since the resilver * started, so a restart is not needed. */ void dsl_scan_assess_vdev(dsl_pool_t *dp, vdev_t *vd) { uint64_t min, max; if (!vdev_resilver_needed(vd, &min, &max)) return; if (!dsl_scan_resilvering(dp)) { spa_async_request(dp->dp_spa, SPA_ASYNC_RESILVER); return; } if (max <= dp->dp_scan->scn_phys.scn_max_txg) return; /* restart is needed, check if it can be deferred */ if (spa_feature_is_enabled(dp->dp_spa, SPA_FEATURE_RESILVER_DEFER)) vdev_defer_resilver(vd); else spa_async_request(dp->dp_spa, SPA_ASYNC_RESILVER); } ZFS_MODULE_PARAM(zfs, zfs_, scan_vdev_limit, U64, ZMOD_RW, "Max bytes in flight per leaf vdev for scrubs and resilvers"); ZFS_MODULE_PARAM(zfs, zfs_, scrub_min_time_ms, UINT, ZMOD_RW, "Min millisecs to scrub per txg"); ZFS_MODULE_PARAM(zfs, zfs_, obsolete_min_time_ms, UINT, ZMOD_RW, "Min millisecs to obsolete per txg"); ZFS_MODULE_PARAM(zfs, zfs_, free_min_time_ms, UINT, ZMOD_RW, "Min millisecs to free per txg"); ZFS_MODULE_PARAM(zfs, zfs_, resilver_min_time_ms, UINT, ZMOD_RW, "Min millisecs to resilver per txg"); ZFS_MODULE_PARAM(zfs, zfs_, scan_suspend_progress, INT, ZMOD_RW, "Set to prevent scans from progressing"); ZFS_MODULE_PARAM(zfs, zfs_, no_scrub_io, INT, ZMOD_RW, "Set to disable scrub I/O"); ZFS_MODULE_PARAM(zfs, zfs_, no_scrub_prefetch, INT, ZMOD_RW, "Set to disable scrub prefetching"); ZFS_MODULE_PARAM(zfs, zfs_, async_block_max_blocks, U64, ZMOD_RW, "Max number of blocks freed in one txg"); ZFS_MODULE_PARAM(zfs, zfs_, max_async_dedup_frees, U64, ZMOD_RW, "Max number of dedup blocks freed in one txg"); ZFS_MODULE_PARAM(zfs, zfs_, free_bpobj_enabled, INT, ZMOD_RW, "Enable processing of the free_bpobj"); ZFS_MODULE_PARAM(zfs, zfs_, scan_blkstats, INT, ZMOD_RW, "Enable block statistics calculation during scrub"); ZFS_MODULE_PARAM(zfs, zfs_, scan_mem_lim_fact, UINT, ZMOD_RW, "Fraction of RAM for scan hard limit"); ZFS_MODULE_PARAM(zfs, zfs_, scan_issue_strategy, UINT, ZMOD_RW, "IO issuing strategy during scrubbing. 0 = default, 1 = LBA, 2 = size"); ZFS_MODULE_PARAM(zfs, zfs_, scan_legacy, INT, ZMOD_RW, "Scrub using legacy non-sequential method"); ZFS_MODULE_PARAM(zfs, zfs_, scan_checkpoint_intval, UINT, ZMOD_RW, "Scan progress on-disk checkpointing interval"); ZFS_MODULE_PARAM(zfs, zfs_, scan_max_ext_gap, U64, ZMOD_RW, "Max gap in bytes between sequential scrub / resilver I/Os"); ZFS_MODULE_PARAM(zfs, zfs_, scan_mem_lim_soft_fact, UINT, ZMOD_RW, "Fraction of hard limit used as soft limit"); ZFS_MODULE_PARAM(zfs, zfs_, scan_strict_mem_lim, INT, ZMOD_RW, "Tunable to attempt to reduce lock contention"); ZFS_MODULE_PARAM(zfs, zfs_, scan_fill_weight, UINT, ZMOD_RW, "Tunable to adjust bias towards more filled segments during scans"); ZFS_MODULE_PARAM(zfs, zfs_, scan_report_txgs, UINT, ZMOD_RW, "Tunable to report resilver performance over the last N txgs"); ZFS_MODULE_PARAM(zfs, zfs_, resilver_disable_defer, INT, ZMOD_RW, "Process all resilvers immediately"); + +ZFS_MODULE_PARAM(zfs, zfs_, scrub_error_blocks_per_txg, U64, ZMOD_RW, + "Error blocks to be scrubbed in one txg"); +/* END CSTYLED */ diff --git a/sys/contrib/openzfs/module/zfs/spa.c b/sys/contrib/openzfs/module/zfs/spa.c index 16396170273c..1fc2c5e8c55d 100644 --- a/sys/contrib/openzfs/module/zfs/spa.c +++ b/sys/contrib/openzfs/module/zfs/spa.c @@ -1,10165 +1,10178 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2011, 2020 by Delphix. All rights reserved. * Copyright (c) 2018, Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2014 Spectra Logic Corporation, All rights reserved. * Copyright 2013 Saso Kiselkov. All rights reserved. * Copyright (c) 2014 Integros [integros.com] * Copyright 2016 Toomas Soome * Copyright (c) 2016 Actifio, Inc. All rights reserved. * Copyright 2018 Joyent, Inc. * Copyright (c) 2017, 2019, Datto Inc. All rights reserved. * Copyright 2017 Joyent, Inc. * Copyright (c) 2017, Intel Corporation. * Copyright (c) 2021, Colm Buckley */ /* * SPA: Storage Pool Allocator * * This file contains all the routines used when modifying on-disk SPA state. * This includes opening, importing, destroying, exporting a pool, and syncing a * pool. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef _KERNEL #include #include #include #include #include #endif /* _KERNEL */ #include "zfs_prop.h" #include "zfs_comutil.h" /* * The interval, in seconds, at which failed configuration cache file writes * should be retried. */ int zfs_ccw_retry_interval = 300; typedef enum zti_modes { ZTI_MODE_FIXED, /* value is # of threads (min 1) */ ZTI_MODE_BATCH, /* cpu-intensive; value is ignored */ ZTI_MODE_SCALE, /* Taskqs scale with CPUs. */ ZTI_MODE_NULL, /* don't create a taskq */ ZTI_NMODES } zti_modes_t; #define ZTI_P(n, q) { ZTI_MODE_FIXED, (n), (q) } #define ZTI_PCT(n) { ZTI_MODE_ONLINE_PERCENT, (n), 1 } #define ZTI_BATCH { ZTI_MODE_BATCH, 0, 1 } #define ZTI_SCALE { ZTI_MODE_SCALE, 0, 1 } #define ZTI_NULL { ZTI_MODE_NULL, 0, 0 } #define ZTI_N(n) ZTI_P(n, 1) #define ZTI_ONE ZTI_N(1) typedef struct zio_taskq_info { zti_modes_t zti_mode; uint_t zti_value; uint_t zti_count; } zio_taskq_info_t; static const char *const zio_taskq_types[ZIO_TASKQ_TYPES] = { "iss", "iss_h", "int", "int_h" }; /* * This table defines the taskq settings for each ZFS I/O type. When * initializing a pool, we use this table to create an appropriately sized * taskq. Some operations are low volume and therefore have a small, static * number of threads assigned to their taskqs using the ZTI_N(#) or ZTI_ONE * macros. Other operations process a large amount of data; the ZTI_BATCH * macro causes us to create a taskq oriented for throughput. Some operations * are so high frequency and short-lived that the taskq itself can become a * point of lock contention. The ZTI_P(#, #) macro indicates that we need an * additional degree of parallelism specified by the number of threads per- * taskq and the number of taskqs; when dispatching an event in this case, the * particular taskq is chosen at random. ZTI_SCALE is similar to ZTI_BATCH, * but with number of taskqs also scaling with number of CPUs. * * The different taskq priorities are to handle the different contexts (issue * and interrupt) and then to reserve threads for ZIO_PRIORITY_NOW I/Os that * need to be handled with minimum delay. */ static const zio_taskq_info_t zio_taskqs[ZIO_TYPES][ZIO_TASKQ_TYPES] = { /* ISSUE ISSUE_HIGH INTR INTR_HIGH */ { ZTI_ONE, ZTI_NULL, ZTI_ONE, ZTI_NULL }, /* NULL */ { ZTI_N(8), ZTI_NULL, ZTI_SCALE, ZTI_NULL }, /* READ */ { ZTI_BATCH, ZTI_N(5), ZTI_SCALE, ZTI_N(5) }, /* WRITE */ { ZTI_SCALE, ZTI_NULL, ZTI_ONE, ZTI_NULL }, /* FREE */ { ZTI_ONE, ZTI_NULL, ZTI_ONE, ZTI_NULL }, /* CLAIM */ { ZTI_ONE, ZTI_NULL, ZTI_ONE, ZTI_NULL }, /* IOCTL */ { ZTI_N(4), ZTI_NULL, ZTI_ONE, ZTI_NULL }, /* TRIM */ }; static void spa_sync_version(void *arg, dmu_tx_t *tx); static void spa_sync_props(void *arg, dmu_tx_t *tx); static boolean_t spa_has_active_shared_spare(spa_t *spa); static int spa_load_impl(spa_t *spa, spa_import_type_t type, const char **ereport); static void spa_vdev_resilver_done(spa_t *spa); static uint_t zio_taskq_batch_pct = 80; /* 1 thread per cpu in pset */ static uint_t zio_taskq_batch_tpq; /* threads per taskq */ static const boolean_t zio_taskq_sysdc = B_TRUE; /* use SDC scheduling class */ static const uint_t zio_taskq_basedc = 80; /* base duty cycle */ static const boolean_t spa_create_process = B_TRUE; /* no process => no sysdc */ /* * Report any spa_load_verify errors found, but do not fail spa_load. * This is used by zdb to analyze non-idle pools. */ boolean_t spa_load_verify_dryrun = B_FALSE; /* * Allow read spacemaps in case of readonly import (spa_mode == SPA_MODE_READ). * This is used by zdb for spacemaps verification. */ boolean_t spa_mode_readable_spacemaps = B_FALSE; /* * This (illegal) pool name is used when temporarily importing a spa_t in order * to get the vdev stats associated with the imported devices. */ #define TRYIMPORT_NAME "$import" /* * For debugging purposes: print out vdev tree during pool import. */ static int spa_load_print_vdev_tree = B_FALSE; /* * A non-zero value for zfs_max_missing_tvds means that we allow importing * pools with missing top-level vdevs. This is strictly intended for advanced * pool recovery cases since missing data is almost inevitable. Pools with * missing devices can only be imported read-only for safety reasons, and their * fail-mode will be automatically set to "continue". * * With 1 missing vdev we should be able to import the pool and mount all * datasets. User data that was not modified after the missing device has been * added should be recoverable. This means that snapshots created prior to the * addition of that device should be completely intact. * * With 2 missing vdevs, some datasets may fail to mount since there are * dataset statistics that are stored as regular metadata. Some data might be * recoverable if those vdevs were added recently. * * With 3 or more missing vdevs, the pool is severely damaged and MOS entries * may be missing entirely. Chances of data recovery are very low. Note that * there are also risks of performing an inadvertent rewind as we might be * missing all the vdevs with the latest uberblocks. */ uint64_t zfs_max_missing_tvds = 0; /* * The parameters below are similar to zfs_max_missing_tvds but are only * intended for a preliminary open of the pool with an untrusted config which * might be incomplete or out-dated. * * We are more tolerant for pools opened from a cachefile since we could have * an out-dated cachefile where a device removal was not registered. * We could have set the limit arbitrarily high but in the case where devices * are really missing we would want to return the proper error codes; we chose * SPA_DVAS_PER_BP - 1 so that some copies of the MOS would still be available * and we get a chance to retrieve the trusted config. */ uint64_t zfs_max_missing_tvds_cachefile = SPA_DVAS_PER_BP - 1; /* * In the case where config was assembled by scanning device paths (/dev/dsks * by default) we are less tolerant since all the existing devices should have * been detected and we want spa_load to return the right error codes. */ uint64_t zfs_max_missing_tvds_scan = 0; /* * Debugging aid that pauses spa_sync() towards the end. */ static const boolean_t zfs_pause_spa_sync = B_FALSE; /* * Variables to indicate the livelist condense zthr func should wait at certain * points for the livelist to be removed - used to test condense/destroy races */ static int zfs_livelist_condense_zthr_pause = 0; static int zfs_livelist_condense_sync_pause = 0; /* * Variables to track whether or not condense cancellation has been * triggered in testing. */ static int zfs_livelist_condense_sync_cancel = 0; static int zfs_livelist_condense_zthr_cancel = 0; /* * Variable to track whether or not extra ALLOC blkptrs were added to a * livelist entry while it was being condensed (caused by the way we track * remapped blkptrs in dbuf_remap_impl) */ static int zfs_livelist_condense_new_alloc = 0; /* * ========================================================================== * SPA properties routines * ========================================================================== */ /* * Add a (source=src, propname=propval) list to an nvlist. */ static void spa_prop_add_list(nvlist_t *nvl, zpool_prop_t prop, const char *strval, uint64_t intval, zprop_source_t src) { const char *propname = zpool_prop_to_name(prop); nvlist_t *propval; propval = fnvlist_alloc(); fnvlist_add_uint64(propval, ZPROP_SOURCE, src); if (strval != NULL) fnvlist_add_string(propval, ZPROP_VALUE, strval); else fnvlist_add_uint64(propval, ZPROP_VALUE, intval); fnvlist_add_nvlist(nvl, propname, propval); nvlist_free(propval); } /* * Add a user property (source=src, propname=propval) to an nvlist. */ static void spa_prop_add_user(nvlist_t *nvl, const char *propname, char *strval, zprop_source_t src) { nvlist_t *propval; VERIFY(nvlist_alloc(&propval, NV_UNIQUE_NAME, KM_SLEEP) == 0); VERIFY(nvlist_add_uint64(propval, ZPROP_SOURCE, src) == 0); VERIFY(nvlist_add_string(propval, ZPROP_VALUE, strval) == 0); VERIFY(nvlist_add_nvlist(nvl, propname, propval) == 0); nvlist_free(propval); } /* * Get property values from the spa configuration. */ static void spa_prop_get_config(spa_t *spa, nvlist_t **nvp) { vdev_t *rvd = spa->spa_root_vdev; dsl_pool_t *pool = spa->spa_dsl_pool; uint64_t size, alloc, cap, version; const zprop_source_t src = ZPROP_SRC_NONE; spa_config_dirent_t *dp; metaslab_class_t *mc = spa_normal_class(spa); ASSERT(MUTEX_HELD(&spa->spa_props_lock)); if (rvd != NULL) { alloc = metaslab_class_get_alloc(mc); alloc += metaslab_class_get_alloc(spa_special_class(spa)); alloc += metaslab_class_get_alloc(spa_dedup_class(spa)); alloc += metaslab_class_get_alloc(spa_embedded_log_class(spa)); size = metaslab_class_get_space(mc); size += metaslab_class_get_space(spa_special_class(spa)); size += metaslab_class_get_space(spa_dedup_class(spa)); size += metaslab_class_get_space(spa_embedded_log_class(spa)); spa_prop_add_list(*nvp, ZPOOL_PROP_NAME, spa_name(spa), 0, src); spa_prop_add_list(*nvp, ZPOOL_PROP_SIZE, NULL, size, src); spa_prop_add_list(*nvp, ZPOOL_PROP_ALLOCATED, NULL, alloc, src); spa_prop_add_list(*nvp, ZPOOL_PROP_FREE, NULL, size - alloc, src); spa_prop_add_list(*nvp, ZPOOL_PROP_CHECKPOINT, NULL, spa->spa_checkpoint_info.sci_dspace, src); spa_prop_add_list(*nvp, ZPOOL_PROP_FRAGMENTATION, NULL, metaslab_class_fragmentation(mc), src); spa_prop_add_list(*nvp, ZPOOL_PROP_EXPANDSZ, NULL, metaslab_class_expandable_space(mc), src); spa_prop_add_list(*nvp, ZPOOL_PROP_READONLY, NULL, (spa_mode(spa) == SPA_MODE_READ), src); cap = (size == 0) ? 0 : (alloc * 100 / size); spa_prop_add_list(*nvp, ZPOOL_PROP_CAPACITY, NULL, cap, src); spa_prop_add_list(*nvp, ZPOOL_PROP_DEDUPRATIO, NULL, ddt_get_pool_dedup_ratio(spa), src); spa_prop_add_list(*nvp, ZPOOL_PROP_BCLONEUSED, NULL, brt_get_used(spa), src); spa_prop_add_list(*nvp, ZPOOL_PROP_BCLONESAVED, NULL, brt_get_saved(spa), src); spa_prop_add_list(*nvp, ZPOOL_PROP_BCLONERATIO, NULL, brt_get_ratio(spa), src); spa_prop_add_list(*nvp, ZPOOL_PROP_HEALTH, NULL, rvd->vdev_state, src); version = spa_version(spa); if (version == zpool_prop_default_numeric(ZPOOL_PROP_VERSION)) { spa_prop_add_list(*nvp, ZPOOL_PROP_VERSION, NULL, version, ZPROP_SRC_DEFAULT); } else { spa_prop_add_list(*nvp, ZPOOL_PROP_VERSION, NULL, version, ZPROP_SRC_LOCAL); } spa_prop_add_list(*nvp, ZPOOL_PROP_LOAD_GUID, NULL, spa_load_guid(spa), src); } if (pool != NULL) { /* * The $FREE directory was introduced in SPA_VERSION_DEADLISTS, * when opening pools before this version freedir will be NULL. */ if (pool->dp_free_dir != NULL) { spa_prop_add_list(*nvp, ZPOOL_PROP_FREEING, NULL, dsl_dir_phys(pool->dp_free_dir)->dd_used_bytes, src); } else { spa_prop_add_list(*nvp, ZPOOL_PROP_FREEING, NULL, 0, src); } if (pool->dp_leak_dir != NULL) { spa_prop_add_list(*nvp, ZPOOL_PROP_LEAKED, NULL, dsl_dir_phys(pool->dp_leak_dir)->dd_used_bytes, src); } else { spa_prop_add_list(*nvp, ZPOOL_PROP_LEAKED, NULL, 0, src); } } spa_prop_add_list(*nvp, ZPOOL_PROP_GUID, NULL, spa_guid(spa), src); if (spa->spa_comment != NULL) { spa_prop_add_list(*nvp, ZPOOL_PROP_COMMENT, spa->spa_comment, 0, ZPROP_SRC_LOCAL); } if (spa->spa_compatibility != NULL) { spa_prop_add_list(*nvp, ZPOOL_PROP_COMPATIBILITY, spa->spa_compatibility, 0, ZPROP_SRC_LOCAL); } if (spa->spa_root != NULL) spa_prop_add_list(*nvp, ZPOOL_PROP_ALTROOT, spa->spa_root, 0, ZPROP_SRC_LOCAL); if (spa_feature_is_enabled(spa, SPA_FEATURE_LARGE_BLOCKS)) { spa_prop_add_list(*nvp, ZPOOL_PROP_MAXBLOCKSIZE, NULL, MIN(zfs_max_recordsize, SPA_MAXBLOCKSIZE), ZPROP_SRC_NONE); } else { spa_prop_add_list(*nvp, ZPOOL_PROP_MAXBLOCKSIZE, NULL, SPA_OLD_MAXBLOCKSIZE, ZPROP_SRC_NONE); } if (spa_feature_is_enabled(spa, SPA_FEATURE_LARGE_DNODE)) { spa_prop_add_list(*nvp, ZPOOL_PROP_MAXDNODESIZE, NULL, DNODE_MAX_SIZE, ZPROP_SRC_NONE); } else { spa_prop_add_list(*nvp, ZPOOL_PROP_MAXDNODESIZE, NULL, DNODE_MIN_SIZE, ZPROP_SRC_NONE); } if ((dp = list_head(&spa->spa_config_list)) != NULL) { if (dp->scd_path == NULL) { spa_prop_add_list(*nvp, ZPOOL_PROP_CACHEFILE, "none", 0, ZPROP_SRC_LOCAL); } else if (strcmp(dp->scd_path, spa_config_path) != 0) { spa_prop_add_list(*nvp, ZPOOL_PROP_CACHEFILE, dp->scd_path, 0, ZPROP_SRC_LOCAL); } } } /* * Get zpool property values. */ int spa_prop_get(spa_t *spa, nvlist_t **nvp) { objset_t *mos = spa->spa_meta_objset; zap_cursor_t zc; zap_attribute_t za; dsl_pool_t *dp; int err; err = nvlist_alloc(nvp, NV_UNIQUE_NAME, KM_SLEEP); if (err) return (err); dp = spa_get_dsl(spa); dsl_pool_config_enter(dp, FTAG); mutex_enter(&spa->spa_props_lock); /* * Get properties from the spa config. */ spa_prop_get_config(spa, nvp); /* If no pool property object, no more prop to get. */ if (mos == NULL || spa->spa_pool_props_object == 0) goto out; /* * Get properties from the MOS pool property object. */ for (zap_cursor_init(&zc, mos, spa->spa_pool_props_object); (err = zap_cursor_retrieve(&zc, &za)) == 0; zap_cursor_advance(&zc)) { uint64_t intval = 0; char *strval = NULL; zprop_source_t src = ZPROP_SRC_DEFAULT; zpool_prop_t prop; if ((prop = zpool_name_to_prop(za.za_name)) == ZPOOL_PROP_INVAL && !zfs_prop_user(za.za_name)) continue; switch (za.za_integer_length) { case 8: /* integer property */ if (za.za_first_integer != zpool_prop_default_numeric(prop)) src = ZPROP_SRC_LOCAL; if (prop == ZPOOL_PROP_BOOTFS) { dsl_dataset_t *ds = NULL; err = dsl_dataset_hold_obj(dp, za.za_first_integer, FTAG, &ds); if (err != 0) break; strval = kmem_alloc(ZFS_MAX_DATASET_NAME_LEN, KM_SLEEP); dsl_dataset_name(ds, strval); dsl_dataset_rele(ds, FTAG); } else { strval = NULL; intval = za.za_first_integer; } spa_prop_add_list(*nvp, prop, strval, intval, src); if (strval != NULL) kmem_free(strval, ZFS_MAX_DATASET_NAME_LEN); break; case 1: /* string property */ strval = kmem_alloc(za.za_num_integers, KM_SLEEP); err = zap_lookup(mos, spa->spa_pool_props_object, za.za_name, 1, za.za_num_integers, strval); if (err) { kmem_free(strval, za.za_num_integers); break; } if (prop != ZPOOL_PROP_INVAL) { spa_prop_add_list(*nvp, prop, strval, 0, src); } else { src = ZPROP_SRC_LOCAL; spa_prop_add_user(*nvp, za.za_name, strval, src); } kmem_free(strval, za.za_num_integers); break; default: break; } } zap_cursor_fini(&zc); out: mutex_exit(&spa->spa_props_lock); dsl_pool_config_exit(dp, FTAG); if (err && err != ENOENT) { nvlist_free(*nvp); *nvp = NULL; return (err); } return (0); } /* * Validate the given pool properties nvlist and modify the list * for the property values to be set. */ static int spa_prop_validate(spa_t *spa, nvlist_t *props) { nvpair_t *elem; int error = 0, reset_bootfs = 0; uint64_t objnum = 0; boolean_t has_feature = B_FALSE; elem = NULL; while ((elem = nvlist_next_nvpair(props, elem)) != NULL) { uint64_t intval; const char *strval, *slash, *check, *fname; const char *propname = nvpair_name(elem); zpool_prop_t prop = zpool_name_to_prop(propname); switch (prop) { case ZPOOL_PROP_INVAL: /* * Sanitize the input. */ if (zfs_prop_user(propname)) { if (strlen(propname) >= ZAP_MAXNAMELEN) { error = SET_ERROR(ENAMETOOLONG); break; } if (strlen(fnvpair_value_string(elem)) >= ZAP_MAXVALUELEN) { error = SET_ERROR(E2BIG); break; } } else if (zpool_prop_feature(propname)) { if (nvpair_type(elem) != DATA_TYPE_UINT64) { error = SET_ERROR(EINVAL); break; } if (nvpair_value_uint64(elem, &intval) != 0) { error = SET_ERROR(EINVAL); break; } if (intval != 0) { error = SET_ERROR(EINVAL); break; } fname = strchr(propname, '@') + 1; if (zfeature_lookup_name(fname, NULL) != 0) { error = SET_ERROR(EINVAL); break; } has_feature = B_TRUE; } else { error = SET_ERROR(EINVAL); break; } break; case ZPOOL_PROP_VERSION: error = nvpair_value_uint64(elem, &intval); if (!error && (intval < spa_version(spa) || intval > SPA_VERSION_BEFORE_FEATURES || has_feature)) error = SET_ERROR(EINVAL); break; case ZPOOL_PROP_DELEGATION: case ZPOOL_PROP_AUTOREPLACE: case ZPOOL_PROP_LISTSNAPS: case ZPOOL_PROP_AUTOEXPAND: case ZPOOL_PROP_AUTOTRIM: error = nvpair_value_uint64(elem, &intval); if (!error && intval > 1) error = SET_ERROR(EINVAL); break; case ZPOOL_PROP_MULTIHOST: error = nvpair_value_uint64(elem, &intval); if (!error && intval > 1) error = SET_ERROR(EINVAL); if (!error) { uint32_t hostid = zone_get_hostid(NULL); if (hostid) spa->spa_hostid = hostid; else error = SET_ERROR(ENOTSUP); } break; case ZPOOL_PROP_BOOTFS: /* * If the pool version is less than SPA_VERSION_BOOTFS, * or the pool is still being created (version == 0), * the bootfs property cannot be set. */ if (spa_version(spa) < SPA_VERSION_BOOTFS) { error = SET_ERROR(ENOTSUP); break; } /* * Make sure the vdev config is bootable */ if (!vdev_is_bootable(spa->spa_root_vdev)) { error = SET_ERROR(ENOTSUP); break; } reset_bootfs = 1; error = nvpair_value_string(elem, &strval); if (!error) { objset_t *os; if (strval == NULL || strval[0] == '\0') { objnum = zpool_prop_default_numeric( ZPOOL_PROP_BOOTFS); break; } error = dmu_objset_hold(strval, FTAG, &os); if (error != 0) break; /* Must be ZPL. */ if (dmu_objset_type(os) != DMU_OST_ZFS) { error = SET_ERROR(ENOTSUP); } else { objnum = dmu_objset_id(os); } dmu_objset_rele(os, FTAG); } break; case ZPOOL_PROP_FAILUREMODE: error = nvpair_value_uint64(elem, &intval); if (!error && intval > ZIO_FAILURE_MODE_PANIC) error = SET_ERROR(EINVAL); /* * This is a special case which only occurs when * the pool has completely failed. This allows * the user to change the in-core failmode property * without syncing it out to disk (I/Os might * currently be blocked). We do this by returning * EIO to the caller (spa_prop_set) to trick it * into thinking we encountered a property validation * error. */ if (!error && spa_suspended(spa)) { spa->spa_failmode = intval; error = SET_ERROR(EIO); } break; case ZPOOL_PROP_CACHEFILE: if ((error = nvpair_value_string(elem, &strval)) != 0) break; if (strval[0] == '\0') break; if (strcmp(strval, "none") == 0) break; if (strval[0] != '/') { error = SET_ERROR(EINVAL); break; } slash = strrchr(strval, '/'); ASSERT(slash != NULL); if (slash[1] == '\0' || strcmp(slash, "/.") == 0 || strcmp(slash, "/..") == 0) error = SET_ERROR(EINVAL); break; case ZPOOL_PROP_COMMENT: if ((error = nvpair_value_string(elem, &strval)) != 0) break; for (check = strval; *check != '\0'; check++) { if (!isprint(*check)) { error = SET_ERROR(EINVAL); break; } } if (strlen(strval) > ZPROP_MAX_COMMENT) error = SET_ERROR(E2BIG); break; default: break; } if (error) break; } (void) nvlist_remove_all(props, zpool_prop_to_name(ZPOOL_PROP_DEDUPDITTO)); if (!error && reset_bootfs) { error = nvlist_remove(props, zpool_prop_to_name(ZPOOL_PROP_BOOTFS), DATA_TYPE_STRING); if (!error) { error = nvlist_add_uint64(props, zpool_prop_to_name(ZPOOL_PROP_BOOTFS), objnum); } } return (error); } void spa_configfile_set(spa_t *spa, nvlist_t *nvp, boolean_t need_sync) { const char *cachefile; spa_config_dirent_t *dp; if (nvlist_lookup_string(nvp, zpool_prop_to_name(ZPOOL_PROP_CACHEFILE), &cachefile) != 0) return; dp = kmem_alloc(sizeof (spa_config_dirent_t), KM_SLEEP); if (cachefile[0] == '\0') dp->scd_path = spa_strdup(spa_config_path); else if (strcmp(cachefile, "none") == 0) dp->scd_path = NULL; else dp->scd_path = spa_strdup(cachefile); list_insert_head(&spa->spa_config_list, dp); if (need_sync) spa_async_request(spa, SPA_ASYNC_CONFIG_UPDATE); } int spa_prop_set(spa_t *spa, nvlist_t *nvp) { int error; nvpair_t *elem = NULL; boolean_t need_sync = B_FALSE; if ((error = spa_prop_validate(spa, nvp)) != 0) return (error); while ((elem = nvlist_next_nvpair(nvp, elem)) != NULL) { zpool_prop_t prop = zpool_name_to_prop(nvpair_name(elem)); if (prop == ZPOOL_PROP_CACHEFILE || prop == ZPOOL_PROP_ALTROOT || prop == ZPOOL_PROP_READONLY) continue; if (prop == ZPOOL_PROP_INVAL && zfs_prop_user(nvpair_name(elem))) { need_sync = B_TRUE; break; } if (prop == ZPOOL_PROP_VERSION || prop == ZPOOL_PROP_INVAL) { uint64_t ver = 0; if (prop == ZPOOL_PROP_VERSION) { VERIFY(nvpair_value_uint64(elem, &ver) == 0); } else { ASSERT(zpool_prop_feature(nvpair_name(elem))); ver = SPA_VERSION_FEATURES; need_sync = B_TRUE; } /* Save time if the version is already set. */ if (ver == spa_version(spa)) continue; /* * In addition to the pool directory object, we might * create the pool properties object, the features for * read object, the features for write object, or the * feature descriptions object. */ error = dsl_sync_task(spa->spa_name, NULL, spa_sync_version, &ver, 6, ZFS_SPACE_CHECK_RESERVED); if (error) return (error); continue; } need_sync = B_TRUE; break; } if (need_sync) { return (dsl_sync_task(spa->spa_name, NULL, spa_sync_props, nvp, 6, ZFS_SPACE_CHECK_RESERVED)); } return (0); } /* * If the bootfs property value is dsobj, clear it. */ void spa_prop_clear_bootfs(spa_t *spa, uint64_t dsobj, dmu_tx_t *tx) { if (spa->spa_bootfs == dsobj && spa->spa_pool_props_object != 0) { VERIFY(zap_remove(spa->spa_meta_objset, spa->spa_pool_props_object, zpool_prop_to_name(ZPOOL_PROP_BOOTFS), tx) == 0); spa->spa_bootfs = 0; } } static int spa_change_guid_check(void *arg, dmu_tx_t *tx) { uint64_t *newguid __maybe_unused = arg; spa_t *spa = dmu_tx_pool(tx)->dp_spa; vdev_t *rvd = spa->spa_root_vdev; uint64_t vdev_state; if (spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT)) { int error = (spa_has_checkpoint(spa)) ? ZFS_ERR_CHECKPOINT_EXISTS : ZFS_ERR_DISCARDING_CHECKPOINT; return (SET_ERROR(error)); } spa_config_enter(spa, SCL_STATE, FTAG, RW_READER); vdev_state = rvd->vdev_state; spa_config_exit(spa, SCL_STATE, FTAG); if (vdev_state != VDEV_STATE_HEALTHY) return (SET_ERROR(ENXIO)); ASSERT3U(spa_guid(spa), !=, *newguid); return (0); } static void spa_change_guid_sync(void *arg, dmu_tx_t *tx) { uint64_t *newguid = arg; spa_t *spa = dmu_tx_pool(tx)->dp_spa; uint64_t oldguid; vdev_t *rvd = spa->spa_root_vdev; oldguid = spa_guid(spa); spa_config_enter(spa, SCL_STATE, FTAG, RW_READER); rvd->vdev_guid = *newguid; rvd->vdev_guid_sum += (*newguid - oldguid); vdev_config_dirty(rvd); spa_config_exit(spa, SCL_STATE, FTAG); spa_history_log_internal(spa, "guid change", tx, "old=%llu new=%llu", (u_longlong_t)oldguid, (u_longlong_t)*newguid); } /* * Change the GUID for the pool. This is done so that we can later * re-import a pool built from a clone of our own vdevs. We will modify * the root vdev's guid, our own pool guid, and then mark all of our * vdevs dirty. Note that we must make sure that all our vdevs are * online when we do this, or else any vdevs that weren't present * would be orphaned from our pool. We are also going to issue a * sysevent to update any watchers. */ int spa_change_guid(spa_t *spa) { int error; uint64_t guid; mutex_enter(&spa->spa_vdev_top_lock); mutex_enter(&spa_namespace_lock); guid = spa_generate_guid(NULL); error = dsl_sync_task(spa->spa_name, spa_change_guid_check, spa_change_guid_sync, &guid, 5, ZFS_SPACE_CHECK_RESERVED); if (error == 0) { /* * Clear the kobj flag from all the vdevs to allow * vdev_cache_process_kobj_evt() to post events to all the * vdevs since GUID is updated. */ vdev_clear_kobj_evt(spa->spa_root_vdev); for (int i = 0; i < spa->spa_l2cache.sav_count; i++) vdev_clear_kobj_evt(spa->spa_l2cache.sav_vdevs[i]); spa_write_cachefile(spa, B_FALSE, B_TRUE, B_TRUE); spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_REGUID); } mutex_exit(&spa_namespace_lock); mutex_exit(&spa->spa_vdev_top_lock); return (error); } /* * ========================================================================== * SPA state manipulation (open/create/destroy/import/export) * ========================================================================== */ static int spa_error_entry_compare(const void *a, const void *b) { const spa_error_entry_t *sa = (const spa_error_entry_t *)a; const spa_error_entry_t *sb = (const spa_error_entry_t *)b; int ret; ret = memcmp(&sa->se_bookmark, &sb->se_bookmark, sizeof (zbookmark_phys_t)); return (TREE_ISIGN(ret)); } /* * Utility function which retrieves copies of the current logs and * re-initializes them in the process. */ void spa_get_errlists(spa_t *spa, avl_tree_t *last, avl_tree_t *scrub) { ASSERT(MUTEX_HELD(&spa->spa_errlist_lock)); memcpy(last, &spa->spa_errlist_last, sizeof (avl_tree_t)); memcpy(scrub, &spa->spa_errlist_scrub, sizeof (avl_tree_t)); avl_create(&spa->spa_errlist_scrub, spa_error_entry_compare, sizeof (spa_error_entry_t), offsetof(spa_error_entry_t, se_avl)); avl_create(&spa->spa_errlist_last, spa_error_entry_compare, sizeof (spa_error_entry_t), offsetof(spa_error_entry_t, se_avl)); } static void spa_taskqs_init(spa_t *spa, zio_type_t t, zio_taskq_type_t q) { const zio_taskq_info_t *ztip = &zio_taskqs[t][q]; enum zti_modes mode = ztip->zti_mode; uint_t value = ztip->zti_value; uint_t count = ztip->zti_count; spa_taskqs_t *tqs = &spa->spa_zio_taskq[t][q]; uint_t cpus, flags = TASKQ_DYNAMIC; boolean_t batch = B_FALSE; switch (mode) { case ZTI_MODE_FIXED: ASSERT3U(value, >, 0); break; case ZTI_MODE_BATCH: batch = B_TRUE; flags |= TASKQ_THREADS_CPU_PCT; value = MIN(zio_taskq_batch_pct, 100); break; case ZTI_MODE_SCALE: flags |= TASKQ_THREADS_CPU_PCT; /* * We want more taskqs to reduce lock contention, but we want * less for better request ordering and CPU utilization. */ cpus = MAX(1, boot_ncpus * zio_taskq_batch_pct / 100); if (zio_taskq_batch_tpq > 0) { count = MAX(1, (cpus + zio_taskq_batch_tpq / 2) / zio_taskq_batch_tpq); } else { /* * Prefer 6 threads per taskq, but no more taskqs * than threads in them on large systems. For 80%: * * taskq taskq total * cpus taskqs percent threads threads * ------- ------- ------- ------- ------- * 1 1 80% 1 1 * 2 1 80% 1 1 * 4 1 80% 3 3 * 8 2 40% 3 6 * 16 3 27% 4 12 * 32 5 16% 5 25 * 64 7 11% 7 49 * 128 10 8% 10 100 * 256 14 6% 15 210 */ count = 1 + cpus / 6; while (count * count > cpus) count--; } /* Limit each taskq within 100% to not trigger assertion. */ count = MAX(count, (zio_taskq_batch_pct + 99) / 100); value = (zio_taskq_batch_pct + count / 2) / count; break; case ZTI_MODE_NULL: tqs->stqs_count = 0; tqs->stqs_taskq = NULL; return; default: panic("unrecognized mode for %s_%s taskq (%u:%u) in " "spa_activate()", zio_type_name[t], zio_taskq_types[q], mode, value); break; } ASSERT3U(count, >, 0); tqs->stqs_count = count; tqs->stqs_taskq = kmem_alloc(count * sizeof (taskq_t *), KM_SLEEP); for (uint_t i = 0; i < count; i++) { taskq_t *tq; char name[32]; if (count > 1) (void) snprintf(name, sizeof (name), "%s_%s_%u", zio_type_name[t], zio_taskq_types[q], i); else (void) snprintf(name, sizeof (name), "%s_%s", zio_type_name[t], zio_taskq_types[q]); if (zio_taskq_sysdc && spa->spa_proc != &p0) { if (batch) flags |= TASKQ_DC_BATCH; (void) zio_taskq_basedc; tq = taskq_create_sysdc(name, value, 50, INT_MAX, spa->spa_proc, zio_taskq_basedc, flags); } else { pri_t pri = maxclsyspri; /* * The write issue taskq can be extremely CPU * intensive. Run it at slightly less important * priority than the other taskqs. * * Under Linux and FreeBSD this means incrementing * the priority value as opposed to platforms like * illumos where it should be decremented. * * On FreeBSD, if priorities divided by four (RQ_PPQ) * are equal then a difference between them is * insignificant. */ if (t == ZIO_TYPE_WRITE && q == ZIO_TASKQ_ISSUE) { #if defined(__linux__) pri++; #elif defined(__FreeBSD__) pri += 4; #else #error "unknown OS" #endif } tq = taskq_create_proc(name, value, pri, 50, INT_MAX, spa->spa_proc, flags); } tqs->stqs_taskq[i] = tq; } } static void spa_taskqs_fini(spa_t *spa, zio_type_t t, zio_taskq_type_t q) { spa_taskqs_t *tqs = &spa->spa_zio_taskq[t][q]; if (tqs->stqs_taskq == NULL) { ASSERT3U(tqs->stqs_count, ==, 0); return; } for (uint_t i = 0; i < tqs->stqs_count; i++) { ASSERT3P(tqs->stqs_taskq[i], !=, NULL); taskq_destroy(tqs->stqs_taskq[i]); } kmem_free(tqs->stqs_taskq, tqs->stqs_count * sizeof (taskq_t *)); tqs->stqs_taskq = NULL; } /* * Dispatch a task to the appropriate taskq for the ZFS I/O type and priority. * Note that a type may have multiple discrete taskqs to avoid lock contention * on the taskq itself. In that case we choose which taskq at random by using * the low bits of gethrtime(). */ void spa_taskq_dispatch_ent(spa_t *spa, zio_type_t t, zio_taskq_type_t q, task_func_t *func, void *arg, uint_t flags, taskq_ent_t *ent) { spa_taskqs_t *tqs = &spa->spa_zio_taskq[t][q]; taskq_t *tq; ASSERT3P(tqs->stqs_taskq, !=, NULL); ASSERT3U(tqs->stqs_count, !=, 0); if (tqs->stqs_count == 1) { tq = tqs->stqs_taskq[0]; } else { tq = tqs->stqs_taskq[((uint64_t)gethrtime()) % tqs->stqs_count]; } taskq_dispatch_ent(tq, func, arg, flags, ent); } /* * Same as spa_taskq_dispatch_ent() but block on the task until completion. */ void spa_taskq_dispatch_sync(spa_t *spa, zio_type_t t, zio_taskq_type_t q, task_func_t *func, void *arg, uint_t flags) { spa_taskqs_t *tqs = &spa->spa_zio_taskq[t][q]; taskq_t *tq; taskqid_t id; ASSERT3P(tqs->stqs_taskq, !=, NULL); ASSERT3U(tqs->stqs_count, !=, 0); if (tqs->stqs_count == 1) { tq = tqs->stqs_taskq[0]; } else { tq = tqs->stqs_taskq[((uint64_t)gethrtime()) % tqs->stqs_count]; } id = taskq_dispatch(tq, func, arg, flags); if (id) taskq_wait_id(tq, id); } static void spa_create_zio_taskqs(spa_t *spa) { for (int t = 0; t < ZIO_TYPES; t++) { for (int q = 0; q < ZIO_TASKQ_TYPES; q++) { spa_taskqs_init(spa, t, q); } } } /* * Disabled until spa_thread() can be adapted for Linux. */ #undef HAVE_SPA_THREAD #if defined(_KERNEL) && defined(HAVE_SPA_THREAD) static void spa_thread(void *arg) { psetid_t zio_taskq_psrset_bind = PS_NONE; callb_cpr_t cprinfo; spa_t *spa = arg; user_t *pu = PTOU(curproc); CALLB_CPR_INIT(&cprinfo, &spa->spa_proc_lock, callb_generic_cpr, spa->spa_name); ASSERT(curproc != &p0); (void) snprintf(pu->u_psargs, sizeof (pu->u_psargs), "zpool-%s", spa->spa_name); (void) strlcpy(pu->u_comm, pu->u_psargs, sizeof (pu->u_comm)); /* bind this thread to the requested psrset */ if (zio_taskq_psrset_bind != PS_NONE) { pool_lock(); mutex_enter(&cpu_lock); mutex_enter(&pidlock); mutex_enter(&curproc->p_lock); if (cpupart_bind_thread(curthread, zio_taskq_psrset_bind, 0, NULL, NULL) == 0) { curthread->t_bind_pset = zio_taskq_psrset_bind; } else { cmn_err(CE_WARN, "Couldn't bind process for zfs pool \"%s\" to " "pset %d\n", spa->spa_name, zio_taskq_psrset_bind); } mutex_exit(&curproc->p_lock); mutex_exit(&pidlock); mutex_exit(&cpu_lock); pool_unlock(); } if (zio_taskq_sysdc) { sysdc_thread_enter(curthread, 100, 0); } spa->spa_proc = curproc; spa->spa_did = curthread->t_did; spa_create_zio_taskqs(spa); mutex_enter(&spa->spa_proc_lock); ASSERT(spa->spa_proc_state == SPA_PROC_CREATED); spa->spa_proc_state = SPA_PROC_ACTIVE; cv_broadcast(&spa->spa_proc_cv); CALLB_CPR_SAFE_BEGIN(&cprinfo); while (spa->spa_proc_state == SPA_PROC_ACTIVE) cv_wait(&spa->spa_proc_cv, &spa->spa_proc_lock); CALLB_CPR_SAFE_END(&cprinfo, &spa->spa_proc_lock); ASSERT(spa->spa_proc_state == SPA_PROC_DEACTIVATE); spa->spa_proc_state = SPA_PROC_GONE; spa->spa_proc = &p0; cv_broadcast(&spa->spa_proc_cv); CALLB_CPR_EXIT(&cprinfo); /* drops spa_proc_lock */ mutex_enter(&curproc->p_lock); lwp_exit(); } #endif /* * Activate an uninitialized pool. */ static void spa_activate(spa_t *spa, spa_mode_t mode) { ASSERT(spa->spa_state == POOL_STATE_UNINITIALIZED); spa->spa_state = POOL_STATE_ACTIVE; spa->spa_mode = mode; spa->spa_read_spacemaps = spa_mode_readable_spacemaps; spa->spa_normal_class = metaslab_class_create(spa, &zfs_metaslab_ops); spa->spa_log_class = metaslab_class_create(spa, &zfs_metaslab_ops); spa->spa_embedded_log_class = metaslab_class_create(spa, &zfs_metaslab_ops); spa->spa_special_class = metaslab_class_create(spa, &zfs_metaslab_ops); spa->spa_dedup_class = metaslab_class_create(spa, &zfs_metaslab_ops); /* Try to create a covering process */ mutex_enter(&spa->spa_proc_lock); ASSERT(spa->spa_proc_state == SPA_PROC_NONE); ASSERT(spa->spa_proc == &p0); spa->spa_did = 0; (void) spa_create_process; #ifdef HAVE_SPA_THREAD /* Only create a process if we're going to be around a while. */ if (spa_create_process && strcmp(spa->spa_name, TRYIMPORT_NAME) != 0) { if (newproc(spa_thread, (caddr_t)spa, syscid, maxclsyspri, NULL, 0) == 0) { spa->spa_proc_state = SPA_PROC_CREATED; while (spa->spa_proc_state == SPA_PROC_CREATED) { cv_wait(&spa->spa_proc_cv, &spa->spa_proc_lock); } ASSERT(spa->spa_proc_state == SPA_PROC_ACTIVE); ASSERT(spa->spa_proc != &p0); ASSERT(spa->spa_did != 0); } else { #ifdef _KERNEL cmn_err(CE_WARN, "Couldn't create process for zfs pool \"%s\"\n", spa->spa_name); #endif } } #endif /* HAVE_SPA_THREAD */ mutex_exit(&spa->spa_proc_lock); /* If we didn't create a process, we need to create our taskqs. */ if (spa->spa_proc == &p0) { spa_create_zio_taskqs(spa); } for (size_t i = 0; i < TXG_SIZE; i++) { spa->spa_txg_zio[i] = zio_root(spa, NULL, NULL, ZIO_FLAG_CANFAIL); } list_create(&spa->spa_config_dirty_list, sizeof (vdev_t), offsetof(vdev_t, vdev_config_dirty_node)); list_create(&spa->spa_evicting_os_list, sizeof (objset_t), offsetof(objset_t, os_evicting_node)); list_create(&spa->spa_state_dirty_list, sizeof (vdev_t), offsetof(vdev_t, vdev_state_dirty_node)); txg_list_create(&spa->spa_vdev_txg_list, spa, offsetof(struct vdev, vdev_txg_node)); avl_create(&spa->spa_errlist_scrub, spa_error_entry_compare, sizeof (spa_error_entry_t), offsetof(spa_error_entry_t, se_avl)); avl_create(&spa->spa_errlist_last, spa_error_entry_compare, sizeof (spa_error_entry_t), offsetof(spa_error_entry_t, se_avl)); avl_create(&spa->spa_errlist_healed, spa_error_entry_compare, sizeof (spa_error_entry_t), offsetof(spa_error_entry_t, se_avl)); spa_activate_os(spa); spa_keystore_init(&spa->spa_keystore); /* * This taskq is used to perform zvol-minor-related tasks * asynchronously. This has several advantages, including easy * resolution of various deadlocks. * * The taskq must be single threaded to ensure tasks are always * processed in the order in which they were dispatched. * * A taskq per pool allows one to keep the pools independent. * This way if one pool is suspended, it will not impact another. * * The preferred location to dispatch a zvol minor task is a sync * task. In this context, there is easy access to the spa_t and minimal * error handling is required because the sync task must succeed. */ spa->spa_zvol_taskq = taskq_create("z_zvol", 1, defclsyspri, 1, INT_MAX, 0); /* * Taskq dedicated to prefetcher threads: this is used to prevent the * pool traverse code from monopolizing the global (and limited) * system_taskq by inappropriately scheduling long running tasks on it. */ spa->spa_prefetch_taskq = taskq_create("z_prefetch", 100, defclsyspri, 1, INT_MAX, TASKQ_DYNAMIC | TASKQ_THREADS_CPU_PCT); /* * The taskq to upgrade datasets in this pool. Currently used by * feature SPA_FEATURE_USEROBJ_ACCOUNTING/SPA_FEATURE_PROJECT_QUOTA. */ spa->spa_upgrade_taskq = taskq_create("z_upgrade", 100, defclsyspri, 1, INT_MAX, TASKQ_DYNAMIC | TASKQ_THREADS_CPU_PCT); } /* * Opposite of spa_activate(). */ static void spa_deactivate(spa_t *spa) { ASSERT(spa->spa_sync_on == B_FALSE); ASSERT(spa->spa_dsl_pool == NULL); ASSERT(spa->spa_root_vdev == NULL); ASSERT(spa->spa_async_zio_root == NULL); ASSERT(spa->spa_state != POOL_STATE_UNINITIALIZED); spa_evicting_os_wait(spa); if (spa->spa_zvol_taskq) { taskq_destroy(spa->spa_zvol_taskq); spa->spa_zvol_taskq = NULL; } if (spa->spa_prefetch_taskq) { taskq_destroy(spa->spa_prefetch_taskq); spa->spa_prefetch_taskq = NULL; } if (spa->spa_upgrade_taskq) { taskq_destroy(spa->spa_upgrade_taskq); spa->spa_upgrade_taskq = NULL; } txg_list_destroy(&spa->spa_vdev_txg_list); list_destroy(&spa->spa_config_dirty_list); list_destroy(&spa->spa_evicting_os_list); list_destroy(&spa->spa_state_dirty_list); taskq_cancel_id(system_delay_taskq, spa->spa_deadman_tqid); for (int t = 0; t < ZIO_TYPES; t++) { for (int q = 0; q < ZIO_TASKQ_TYPES; q++) { spa_taskqs_fini(spa, t, q); } } for (size_t i = 0; i < TXG_SIZE; i++) { ASSERT3P(spa->spa_txg_zio[i], !=, NULL); VERIFY0(zio_wait(spa->spa_txg_zio[i])); spa->spa_txg_zio[i] = NULL; } metaslab_class_destroy(spa->spa_normal_class); spa->spa_normal_class = NULL; metaslab_class_destroy(spa->spa_log_class); spa->spa_log_class = NULL; metaslab_class_destroy(spa->spa_embedded_log_class); spa->spa_embedded_log_class = NULL; metaslab_class_destroy(spa->spa_special_class); spa->spa_special_class = NULL; metaslab_class_destroy(spa->spa_dedup_class); spa->spa_dedup_class = NULL; /* * If this was part of an import or the open otherwise failed, we may * still have errors left in the queues. Empty them just in case. */ spa_errlog_drain(spa); avl_destroy(&spa->spa_errlist_scrub); avl_destroy(&spa->spa_errlist_last); avl_destroy(&spa->spa_errlist_healed); spa_keystore_fini(&spa->spa_keystore); spa->spa_state = POOL_STATE_UNINITIALIZED; mutex_enter(&spa->spa_proc_lock); if (spa->spa_proc_state != SPA_PROC_NONE) { ASSERT(spa->spa_proc_state == SPA_PROC_ACTIVE); spa->spa_proc_state = SPA_PROC_DEACTIVATE; cv_broadcast(&spa->spa_proc_cv); while (spa->spa_proc_state == SPA_PROC_DEACTIVATE) { ASSERT(spa->spa_proc != &p0); cv_wait(&spa->spa_proc_cv, &spa->spa_proc_lock); } ASSERT(spa->spa_proc_state == SPA_PROC_GONE); spa->spa_proc_state = SPA_PROC_NONE; } ASSERT(spa->spa_proc == &p0); mutex_exit(&spa->spa_proc_lock); /* * We want to make sure spa_thread() has actually exited the ZFS * module, so that the module can't be unloaded out from underneath * it. */ if (spa->spa_did != 0) { thread_join(spa->spa_did); spa->spa_did = 0; } spa_deactivate_os(spa); } /* * Verify a pool configuration, and construct the vdev tree appropriately. This * will create all the necessary vdevs in the appropriate layout, with each vdev * in the CLOSED state. This will prep the pool before open/creation/import. * All vdev validation is done by the vdev_alloc() routine. */ int spa_config_parse(spa_t *spa, vdev_t **vdp, nvlist_t *nv, vdev_t *parent, uint_t id, int atype) { nvlist_t **child; uint_t children; int error; if ((error = vdev_alloc(spa, vdp, nv, parent, id, atype)) != 0) return (error); if ((*vdp)->vdev_ops->vdev_op_leaf) return (0); error = nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children); if (error == ENOENT) return (0); if (error) { vdev_free(*vdp); *vdp = NULL; return (SET_ERROR(EINVAL)); } for (int c = 0; c < children; c++) { vdev_t *vd; if ((error = spa_config_parse(spa, &vd, child[c], *vdp, c, atype)) != 0) { vdev_free(*vdp); *vdp = NULL; return (error); } } ASSERT(*vdp != NULL); return (0); } static boolean_t spa_should_flush_logs_on_unload(spa_t *spa) { if (!spa_feature_is_active(spa, SPA_FEATURE_LOG_SPACEMAP)) return (B_FALSE); if (!spa_writeable(spa)) return (B_FALSE); if (!spa->spa_sync_on) return (B_FALSE); if (spa_state(spa) != POOL_STATE_EXPORTED) return (B_FALSE); if (zfs_keep_log_spacemaps_at_export) return (B_FALSE); return (B_TRUE); } /* * Opens a transaction that will set the flag that will instruct * spa_sync to attempt to flush all the metaslabs for that txg. */ static void spa_unload_log_sm_flush_all(spa_t *spa) { dmu_tx_t *tx = dmu_tx_create_dd(spa_get_dsl(spa)->dp_mos_dir); VERIFY0(dmu_tx_assign(tx, TXG_WAIT)); ASSERT3U(spa->spa_log_flushall_txg, ==, 0); spa->spa_log_flushall_txg = dmu_tx_get_txg(tx); dmu_tx_commit(tx); txg_wait_synced(spa_get_dsl(spa), spa->spa_log_flushall_txg); } static void spa_unload_log_sm_metadata(spa_t *spa) { void *cookie = NULL; spa_log_sm_t *sls; while ((sls = avl_destroy_nodes(&spa->spa_sm_logs_by_txg, &cookie)) != NULL) { VERIFY0(sls->sls_mscount); kmem_free(sls, sizeof (spa_log_sm_t)); } for (log_summary_entry_t *e = list_head(&spa->spa_log_summary); e != NULL; e = list_head(&spa->spa_log_summary)) { VERIFY0(e->lse_mscount); list_remove(&spa->spa_log_summary, e); kmem_free(e, sizeof (log_summary_entry_t)); } spa->spa_unflushed_stats.sus_nblocks = 0; spa->spa_unflushed_stats.sus_memused = 0; spa->spa_unflushed_stats.sus_blocklimit = 0; } static void spa_destroy_aux_threads(spa_t *spa) { if (spa->spa_condense_zthr != NULL) { zthr_destroy(spa->spa_condense_zthr); spa->spa_condense_zthr = NULL; } if (spa->spa_checkpoint_discard_zthr != NULL) { zthr_destroy(spa->spa_checkpoint_discard_zthr); spa->spa_checkpoint_discard_zthr = NULL; } if (spa->spa_livelist_delete_zthr != NULL) { zthr_destroy(spa->spa_livelist_delete_zthr); spa->spa_livelist_delete_zthr = NULL; } if (spa->spa_livelist_condense_zthr != NULL) { zthr_destroy(spa->spa_livelist_condense_zthr); spa->spa_livelist_condense_zthr = NULL; } } /* * Opposite of spa_load(). */ static void spa_unload(spa_t *spa) { ASSERT(MUTEX_HELD(&spa_namespace_lock)); ASSERT(spa_state(spa) != POOL_STATE_UNINITIALIZED); spa_import_progress_remove(spa_guid(spa)); spa_load_note(spa, "UNLOADING"); spa_wake_waiters(spa); /* * If we have set the spa_final_txg, we have already performed the * tasks below in spa_export_common(). We should not redo it here since * we delay the final TXGs beyond what spa_final_txg is set at. */ if (spa->spa_final_txg == UINT64_MAX) { /* * If the log space map feature is enabled and the pool is * getting exported (but not destroyed), we want to spend some * time flushing as many metaslabs as we can in an attempt to * destroy log space maps and save import time. */ if (spa_should_flush_logs_on_unload(spa)) spa_unload_log_sm_flush_all(spa); /* * Stop async tasks. */ spa_async_suspend(spa); if (spa->spa_root_vdev) { vdev_t *root_vdev = spa->spa_root_vdev; vdev_initialize_stop_all(root_vdev, VDEV_INITIALIZE_ACTIVE); vdev_trim_stop_all(root_vdev, VDEV_TRIM_ACTIVE); vdev_autotrim_stop_all(spa); vdev_rebuild_stop_all(spa); } } /* * Stop syncing. */ if (spa->spa_sync_on) { txg_sync_stop(spa->spa_dsl_pool); spa->spa_sync_on = B_FALSE; } /* * This ensures that there is no async metaslab prefetching * while we attempt to unload the spa. */ if (spa->spa_root_vdev != NULL) { for (int c = 0; c < spa->spa_root_vdev->vdev_children; c++) { vdev_t *vc = spa->spa_root_vdev->vdev_child[c]; if (vc->vdev_mg != NULL) taskq_wait(vc->vdev_mg->mg_taskq); } } if (spa->spa_mmp.mmp_thread) mmp_thread_stop(spa); /* * Wait for any outstanding async I/O to complete. */ if (spa->spa_async_zio_root != NULL) { for (int i = 0; i < max_ncpus; i++) (void) zio_wait(spa->spa_async_zio_root[i]); kmem_free(spa->spa_async_zio_root, max_ncpus * sizeof (void *)); spa->spa_async_zio_root = NULL; } if (spa->spa_vdev_removal != NULL) { spa_vdev_removal_destroy(spa->spa_vdev_removal); spa->spa_vdev_removal = NULL; } spa_destroy_aux_threads(spa); spa_condense_fini(spa); bpobj_close(&spa->spa_deferred_bpobj); spa_config_enter(spa, SCL_ALL, spa, RW_WRITER); /* * Close all vdevs. */ if (spa->spa_root_vdev) vdev_free(spa->spa_root_vdev); ASSERT(spa->spa_root_vdev == NULL); /* * Close the dsl pool. */ if (spa->spa_dsl_pool) { dsl_pool_close(spa->spa_dsl_pool); spa->spa_dsl_pool = NULL; spa->spa_meta_objset = NULL; } ddt_unload(spa); brt_unload(spa); spa_unload_log_sm_metadata(spa); /* * Drop and purge level 2 cache */ spa_l2cache_drop(spa); if (spa->spa_spares.sav_vdevs) { for (int i = 0; i < spa->spa_spares.sav_count; i++) vdev_free(spa->spa_spares.sav_vdevs[i]); kmem_free(spa->spa_spares.sav_vdevs, spa->spa_spares.sav_count * sizeof (void *)); spa->spa_spares.sav_vdevs = NULL; } if (spa->spa_spares.sav_config) { nvlist_free(spa->spa_spares.sav_config); spa->spa_spares.sav_config = NULL; } spa->spa_spares.sav_count = 0; if (spa->spa_l2cache.sav_vdevs) { for (int i = 0; i < spa->spa_l2cache.sav_count; i++) { vdev_clear_stats(spa->spa_l2cache.sav_vdevs[i]); vdev_free(spa->spa_l2cache.sav_vdevs[i]); } kmem_free(spa->spa_l2cache.sav_vdevs, spa->spa_l2cache.sav_count * sizeof (void *)); spa->spa_l2cache.sav_vdevs = NULL; } if (spa->spa_l2cache.sav_config) { nvlist_free(spa->spa_l2cache.sav_config); spa->spa_l2cache.sav_config = NULL; } spa->spa_l2cache.sav_count = 0; spa->spa_async_suspended = 0; spa->spa_indirect_vdevs_loaded = B_FALSE; if (spa->spa_comment != NULL) { spa_strfree(spa->spa_comment); spa->spa_comment = NULL; } if (spa->spa_compatibility != NULL) { spa_strfree(spa->spa_compatibility); spa->spa_compatibility = NULL; } spa_config_exit(spa, SCL_ALL, spa); } /* * Load (or re-load) the current list of vdevs describing the active spares for * this pool. When this is called, we have some form of basic information in * 'spa_spares.sav_config'. We parse this into vdevs, try to open them, and * then re-generate a more complete list including status information. */ void spa_load_spares(spa_t *spa) { nvlist_t **spares; uint_t nspares; int i; vdev_t *vd, *tvd; #ifndef _KERNEL /* * zdb opens both the current state of the pool and the * checkpointed state (if present), with a different spa_t. * * As spare vdevs are shared among open pools, we skip loading * them when we load the checkpointed state of the pool. */ if (!spa_writeable(spa)) return; #endif ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL); /* * First, close and free any existing spare vdevs. */ if (spa->spa_spares.sav_vdevs) { for (i = 0; i < spa->spa_spares.sav_count; i++) { vd = spa->spa_spares.sav_vdevs[i]; /* Undo the call to spa_activate() below */ if ((tvd = spa_lookup_by_guid(spa, vd->vdev_guid, B_FALSE)) != NULL && tvd->vdev_isspare) spa_spare_remove(tvd); vdev_close(vd); vdev_free(vd); } kmem_free(spa->spa_spares.sav_vdevs, spa->spa_spares.sav_count * sizeof (void *)); } if (spa->spa_spares.sav_config == NULL) nspares = 0; else VERIFY0(nvlist_lookup_nvlist_array(spa->spa_spares.sav_config, ZPOOL_CONFIG_SPARES, &spares, &nspares)); spa->spa_spares.sav_count = (int)nspares; spa->spa_spares.sav_vdevs = NULL; if (nspares == 0) return; /* * Construct the array of vdevs, opening them to get status in the * process. For each spare, there is potentially two different vdev_t * structures associated with it: one in the list of spares (used only * for basic validation purposes) and one in the active vdev * configuration (if it's spared in). During this phase we open and * validate each vdev on the spare list. If the vdev also exists in the * active configuration, then we also mark this vdev as an active spare. */ spa->spa_spares.sav_vdevs = kmem_zalloc(nspares * sizeof (void *), KM_SLEEP); for (i = 0; i < spa->spa_spares.sav_count; i++) { VERIFY(spa_config_parse(spa, &vd, spares[i], NULL, 0, VDEV_ALLOC_SPARE) == 0); ASSERT(vd != NULL); spa->spa_spares.sav_vdevs[i] = vd; if ((tvd = spa_lookup_by_guid(spa, vd->vdev_guid, B_FALSE)) != NULL) { if (!tvd->vdev_isspare) spa_spare_add(tvd); /* * We only mark the spare active if we were successfully * able to load the vdev. Otherwise, importing a pool * with a bad active spare would result in strange * behavior, because multiple pool would think the spare * is actively in use. * * There is a vulnerability here to an equally bizarre * circumstance, where a dead active spare is later * brought back to life (onlined or otherwise). Given * the rarity of this scenario, and the extra complexity * it adds, we ignore the possibility. */ if (!vdev_is_dead(tvd)) spa_spare_activate(tvd); } vd->vdev_top = vd; vd->vdev_aux = &spa->spa_spares; if (vdev_open(vd) != 0) continue; if (vdev_validate_aux(vd) == 0) spa_spare_add(vd); } /* * Recompute the stashed list of spares, with status information * this time. */ fnvlist_remove(spa->spa_spares.sav_config, ZPOOL_CONFIG_SPARES); spares = kmem_alloc(spa->spa_spares.sav_count * sizeof (void *), KM_SLEEP); for (i = 0; i < spa->spa_spares.sav_count; i++) spares[i] = vdev_config_generate(spa, spa->spa_spares.sav_vdevs[i], B_TRUE, VDEV_CONFIG_SPARE); fnvlist_add_nvlist_array(spa->spa_spares.sav_config, ZPOOL_CONFIG_SPARES, (const nvlist_t * const *)spares, spa->spa_spares.sav_count); for (i = 0; i < spa->spa_spares.sav_count; i++) nvlist_free(spares[i]); kmem_free(spares, spa->spa_spares.sav_count * sizeof (void *)); } /* * Load (or re-load) the current list of vdevs describing the active l2cache for * this pool. When this is called, we have some form of basic information in * 'spa_l2cache.sav_config'. We parse this into vdevs, try to open them, and * then re-generate a more complete list including status information. * Devices which are already active have their details maintained, and are * not re-opened. */ void spa_load_l2cache(spa_t *spa) { nvlist_t **l2cache = NULL; uint_t nl2cache; int i, j, oldnvdevs; uint64_t guid; vdev_t *vd, **oldvdevs, **newvdevs; spa_aux_vdev_t *sav = &spa->spa_l2cache; #ifndef _KERNEL /* * zdb opens both the current state of the pool and the * checkpointed state (if present), with a different spa_t. * * As L2 caches are part of the ARC which is shared among open * pools, we skip loading them when we load the checkpointed * state of the pool. */ if (!spa_writeable(spa)) return; #endif ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL); oldvdevs = sav->sav_vdevs; oldnvdevs = sav->sav_count; sav->sav_vdevs = NULL; sav->sav_count = 0; if (sav->sav_config == NULL) { nl2cache = 0; newvdevs = NULL; goto out; } VERIFY0(nvlist_lookup_nvlist_array(sav->sav_config, ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache)); newvdevs = kmem_alloc(nl2cache * sizeof (void *), KM_SLEEP); /* * Process new nvlist of vdevs. */ for (i = 0; i < nl2cache; i++) { guid = fnvlist_lookup_uint64(l2cache[i], ZPOOL_CONFIG_GUID); newvdevs[i] = NULL; for (j = 0; j < oldnvdevs; j++) { vd = oldvdevs[j]; if (vd != NULL && guid == vd->vdev_guid) { /* * Retain previous vdev for add/remove ops. */ newvdevs[i] = vd; oldvdevs[j] = NULL; break; } } if (newvdevs[i] == NULL) { /* * Create new vdev */ VERIFY(spa_config_parse(spa, &vd, l2cache[i], NULL, 0, VDEV_ALLOC_L2CACHE) == 0); ASSERT(vd != NULL); newvdevs[i] = vd; /* * Commit this vdev as an l2cache device, * even if it fails to open. */ spa_l2cache_add(vd); vd->vdev_top = vd; vd->vdev_aux = sav; spa_l2cache_activate(vd); if (vdev_open(vd) != 0) continue; (void) vdev_validate_aux(vd); if (!vdev_is_dead(vd)) l2arc_add_vdev(spa, vd); /* * Upon cache device addition to a pool or pool * creation with a cache device or if the header * of the device is invalid we issue an async * TRIM command for the whole device which will * execute if l2arc_trim_ahead > 0. */ spa_async_request(spa, SPA_ASYNC_L2CACHE_TRIM); } } sav->sav_vdevs = newvdevs; sav->sav_count = (int)nl2cache; /* * Recompute the stashed list of l2cache devices, with status * information this time. */ fnvlist_remove(sav->sav_config, ZPOOL_CONFIG_L2CACHE); if (sav->sav_count > 0) l2cache = kmem_alloc(sav->sav_count * sizeof (void *), KM_SLEEP); for (i = 0; i < sav->sav_count; i++) l2cache[i] = vdev_config_generate(spa, sav->sav_vdevs[i], B_TRUE, VDEV_CONFIG_L2CACHE); fnvlist_add_nvlist_array(sav->sav_config, ZPOOL_CONFIG_L2CACHE, (const nvlist_t * const *)l2cache, sav->sav_count); out: /* * Purge vdevs that were dropped */ if (oldvdevs) { for (i = 0; i < oldnvdevs; i++) { uint64_t pool; vd = oldvdevs[i]; if (vd != NULL) { ASSERT(vd->vdev_isl2cache); if (spa_l2cache_exists(vd->vdev_guid, &pool) && pool != 0ULL && l2arc_vdev_present(vd)) l2arc_remove_vdev(vd); vdev_clear_stats(vd); vdev_free(vd); } } kmem_free(oldvdevs, oldnvdevs * sizeof (void *)); } for (i = 0; i < sav->sav_count; i++) nvlist_free(l2cache[i]); if (sav->sav_count) kmem_free(l2cache, sav->sav_count * sizeof (void *)); } static int load_nvlist(spa_t *spa, uint64_t obj, nvlist_t **value) { dmu_buf_t *db; char *packed = NULL; size_t nvsize = 0; int error; *value = NULL; error = dmu_bonus_hold(spa->spa_meta_objset, obj, FTAG, &db); if (error) return (error); nvsize = *(uint64_t *)db->db_data; dmu_buf_rele(db, FTAG); packed = vmem_alloc(nvsize, KM_SLEEP); error = dmu_read(spa->spa_meta_objset, obj, 0, nvsize, packed, DMU_READ_PREFETCH); if (error == 0) error = nvlist_unpack(packed, nvsize, value, 0); vmem_free(packed, nvsize); return (error); } /* * Concrete top-level vdevs that are not missing and are not logs. At every * spa_sync we write new uberblocks to at least SPA_SYNC_MIN_VDEVS core tvds. */ static uint64_t spa_healthy_core_tvds(spa_t *spa) { vdev_t *rvd = spa->spa_root_vdev; uint64_t tvds = 0; for (uint64_t i = 0; i < rvd->vdev_children; i++) { vdev_t *vd = rvd->vdev_child[i]; if (vd->vdev_islog) continue; if (vdev_is_concrete(vd) && !vdev_is_dead(vd)) tvds++; } return (tvds); } /* * Checks to see if the given vdev could not be opened, in which case we post a * sysevent to notify the autoreplace code that the device has been removed. */ static void spa_check_removed(vdev_t *vd) { for (uint64_t c = 0; c < vd->vdev_children; c++) spa_check_removed(vd->vdev_child[c]); if (vd->vdev_ops->vdev_op_leaf && vdev_is_dead(vd) && vdev_is_concrete(vd)) { zfs_post_autoreplace(vd->vdev_spa, vd); spa_event_notify(vd->vdev_spa, vd, NULL, ESC_ZFS_VDEV_CHECK); } } static int spa_check_for_missing_logs(spa_t *spa) { vdev_t *rvd = spa->spa_root_vdev; /* * If we're doing a normal import, then build up any additional * diagnostic information about missing log devices. * We'll pass this up to the user for further processing. */ if (!(spa->spa_import_flags & ZFS_IMPORT_MISSING_LOG)) { nvlist_t **child, *nv; uint64_t idx = 0; child = kmem_alloc(rvd->vdev_children * sizeof (nvlist_t *), KM_SLEEP); nv = fnvlist_alloc(); for (uint64_t c = 0; c < rvd->vdev_children; c++) { vdev_t *tvd = rvd->vdev_child[c]; /* * We consider a device as missing only if it failed * to open (i.e. offline or faulted is not considered * as missing). */ if (tvd->vdev_islog && tvd->vdev_state == VDEV_STATE_CANT_OPEN) { child[idx++] = vdev_config_generate(spa, tvd, B_FALSE, VDEV_CONFIG_MISSING); } } if (idx > 0) { fnvlist_add_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, (const nvlist_t * const *)child, idx); fnvlist_add_nvlist(spa->spa_load_info, ZPOOL_CONFIG_MISSING_DEVICES, nv); for (uint64_t i = 0; i < idx; i++) nvlist_free(child[i]); } nvlist_free(nv); kmem_free(child, rvd->vdev_children * sizeof (char **)); if (idx > 0) { spa_load_failed(spa, "some log devices are missing"); vdev_dbgmsg_print_tree(rvd, 2); return (SET_ERROR(ENXIO)); } } else { for (uint64_t c = 0; c < rvd->vdev_children; c++) { vdev_t *tvd = rvd->vdev_child[c]; if (tvd->vdev_islog && tvd->vdev_state == VDEV_STATE_CANT_OPEN) { spa_set_log_state(spa, SPA_LOG_CLEAR); spa_load_note(spa, "some log devices are " "missing, ZIL is dropped."); vdev_dbgmsg_print_tree(rvd, 2); break; } } } return (0); } /* * Check for missing log devices */ static boolean_t spa_check_logs(spa_t *spa) { boolean_t rv = B_FALSE; dsl_pool_t *dp = spa_get_dsl(spa); switch (spa->spa_log_state) { default: break; case SPA_LOG_MISSING: /* need to recheck in case slog has been restored */ case SPA_LOG_UNKNOWN: rv = (dmu_objset_find_dp(dp, dp->dp_root_dir_obj, zil_check_log_chain, NULL, DS_FIND_CHILDREN) != 0); if (rv) spa_set_log_state(spa, SPA_LOG_MISSING); break; } return (rv); } /* * Passivate any log vdevs (note, does not apply to embedded log metaslabs). */ static boolean_t spa_passivate_log(spa_t *spa) { vdev_t *rvd = spa->spa_root_vdev; boolean_t slog_found = B_FALSE; ASSERT(spa_config_held(spa, SCL_ALLOC, RW_WRITER)); for (int c = 0; c < rvd->vdev_children; c++) { vdev_t *tvd = rvd->vdev_child[c]; if (tvd->vdev_islog) { ASSERT3P(tvd->vdev_log_mg, ==, NULL); metaslab_group_passivate(tvd->vdev_mg); slog_found = B_TRUE; } } return (slog_found); } /* * Activate any log vdevs (note, does not apply to embedded log metaslabs). */ static void spa_activate_log(spa_t *spa) { vdev_t *rvd = spa->spa_root_vdev; ASSERT(spa_config_held(spa, SCL_ALLOC, RW_WRITER)); for (int c = 0; c < rvd->vdev_children; c++) { vdev_t *tvd = rvd->vdev_child[c]; if (tvd->vdev_islog) { ASSERT3P(tvd->vdev_log_mg, ==, NULL); metaslab_group_activate(tvd->vdev_mg); } } } int spa_reset_logs(spa_t *spa) { int error; error = dmu_objset_find(spa_name(spa), zil_reset, NULL, DS_FIND_CHILDREN); if (error == 0) { /* * We successfully offlined the log device, sync out the * current txg so that the "stubby" block can be removed * by zil_sync(). */ txg_wait_synced(spa->spa_dsl_pool, 0); } return (error); } static void spa_aux_check_removed(spa_aux_vdev_t *sav) { for (int i = 0; i < sav->sav_count; i++) spa_check_removed(sav->sav_vdevs[i]); } void spa_claim_notify(zio_t *zio) { spa_t *spa = zio->io_spa; if (zio->io_error) return; mutex_enter(&spa->spa_props_lock); /* any mutex will do */ if (spa->spa_claim_max_txg < zio->io_bp->blk_birth) spa->spa_claim_max_txg = zio->io_bp->blk_birth; mutex_exit(&spa->spa_props_lock); } typedef struct spa_load_error { boolean_t sle_verify_data; uint64_t sle_meta_count; uint64_t sle_data_count; } spa_load_error_t; static void spa_load_verify_done(zio_t *zio) { blkptr_t *bp = zio->io_bp; spa_load_error_t *sle = zio->io_private; dmu_object_type_t type = BP_GET_TYPE(bp); int error = zio->io_error; spa_t *spa = zio->io_spa; abd_free(zio->io_abd); if (error) { if ((BP_GET_LEVEL(bp) != 0 || DMU_OT_IS_METADATA(type)) && type != DMU_OT_INTENT_LOG) atomic_inc_64(&sle->sle_meta_count); else atomic_inc_64(&sle->sle_data_count); } mutex_enter(&spa->spa_scrub_lock); spa->spa_load_verify_bytes -= BP_GET_PSIZE(bp); cv_broadcast(&spa->spa_scrub_io_cv); mutex_exit(&spa->spa_scrub_lock); } /* * Maximum number of inflight bytes is the log2 fraction of the arc size. * By default, we set it to 1/16th of the arc. */ static uint_t spa_load_verify_shift = 4; static int spa_load_verify_metadata = B_TRUE; static int spa_load_verify_data = B_TRUE; static int spa_load_verify_cb(spa_t *spa, zilog_t *zilog, const blkptr_t *bp, const zbookmark_phys_t *zb, const dnode_phys_t *dnp, void *arg) { zio_t *rio = arg; spa_load_error_t *sle = rio->io_private; (void) zilog, (void) dnp; /* * Note: normally this routine will not be called if * spa_load_verify_metadata is not set. However, it may be useful * to manually set the flag after the traversal has begun. */ if (!spa_load_verify_metadata) return (0); /* * Sanity check the block pointer in order to detect obvious damage * before using the contents in subsequent checks or in zio_read(). * When damaged consider it to be a metadata error since we cannot * trust the BP_GET_TYPE and BP_GET_LEVEL values. */ if (!zfs_blkptr_verify(spa, bp, BLK_CONFIG_NEEDED, BLK_VERIFY_LOG)) { atomic_inc_64(&sle->sle_meta_count); return (0); } if (zb->zb_level == ZB_DNODE_LEVEL || BP_IS_HOLE(bp) || BP_IS_EMBEDDED(bp) || BP_IS_REDACTED(bp)) return (0); if (!BP_IS_METADATA(bp) && (!spa_load_verify_data || !sle->sle_verify_data)) return (0); uint64_t maxinflight_bytes = arc_target_bytes() >> spa_load_verify_shift; size_t size = BP_GET_PSIZE(bp); mutex_enter(&spa->spa_scrub_lock); while (spa->spa_load_verify_bytes >= maxinflight_bytes) cv_wait(&spa->spa_scrub_io_cv, &spa->spa_scrub_lock); spa->spa_load_verify_bytes += size; mutex_exit(&spa->spa_scrub_lock); zio_nowait(zio_read(rio, spa, bp, abd_alloc_for_io(size, B_FALSE), size, spa_load_verify_done, rio->io_private, ZIO_PRIORITY_SCRUB, ZIO_FLAG_SPECULATIVE | ZIO_FLAG_CANFAIL | ZIO_FLAG_SCRUB | ZIO_FLAG_RAW, zb)); return (0); } static int verify_dataset_name_len(dsl_pool_t *dp, dsl_dataset_t *ds, void *arg) { (void) dp, (void) arg; if (dsl_dataset_namelen(ds) >= ZFS_MAX_DATASET_NAME_LEN) return (SET_ERROR(ENAMETOOLONG)); return (0); } static int spa_load_verify(spa_t *spa) { zio_t *rio; spa_load_error_t sle = { 0 }; zpool_load_policy_t policy; boolean_t verify_ok = B_FALSE; int error = 0; zpool_get_load_policy(spa->spa_config, &policy); if (policy.zlp_rewind & ZPOOL_NEVER_REWIND || policy.zlp_maxmeta == UINT64_MAX) return (0); dsl_pool_config_enter(spa->spa_dsl_pool, FTAG); error = dmu_objset_find_dp(spa->spa_dsl_pool, spa->spa_dsl_pool->dp_root_dir_obj, verify_dataset_name_len, NULL, DS_FIND_CHILDREN); dsl_pool_config_exit(spa->spa_dsl_pool, FTAG); if (error != 0) return (error); /* * Verify data only if we are rewinding or error limit was set. * Otherwise nothing except dbgmsg care about it to waste time. */ sle.sle_verify_data = (policy.zlp_rewind & ZPOOL_REWIND_MASK) || (policy.zlp_maxdata < UINT64_MAX); rio = zio_root(spa, NULL, &sle, ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE); if (spa_load_verify_metadata) { if (spa->spa_extreme_rewind) { spa_load_note(spa, "performing a complete scan of the " "pool since extreme rewind is on. This may take " "a very long time.\n (spa_load_verify_data=%u, " "spa_load_verify_metadata=%u)", spa_load_verify_data, spa_load_verify_metadata); } error = traverse_pool(spa, spa->spa_verify_min_txg, TRAVERSE_PRE | TRAVERSE_PREFETCH_METADATA | TRAVERSE_NO_DECRYPT, spa_load_verify_cb, rio); } (void) zio_wait(rio); ASSERT0(spa->spa_load_verify_bytes); spa->spa_load_meta_errors = sle.sle_meta_count; spa->spa_load_data_errors = sle.sle_data_count; if (sle.sle_meta_count != 0 || sle.sle_data_count != 0) { spa_load_note(spa, "spa_load_verify found %llu metadata errors " "and %llu data errors", (u_longlong_t)sle.sle_meta_count, (u_longlong_t)sle.sle_data_count); } if (spa_load_verify_dryrun || (!error && sle.sle_meta_count <= policy.zlp_maxmeta && sle.sle_data_count <= policy.zlp_maxdata)) { int64_t loss = 0; verify_ok = B_TRUE; spa->spa_load_txg = spa->spa_uberblock.ub_txg; spa->spa_load_txg_ts = spa->spa_uberblock.ub_timestamp; loss = spa->spa_last_ubsync_txg_ts - spa->spa_load_txg_ts; fnvlist_add_uint64(spa->spa_load_info, ZPOOL_CONFIG_LOAD_TIME, spa->spa_load_txg_ts); fnvlist_add_int64(spa->spa_load_info, ZPOOL_CONFIG_REWIND_TIME, loss); fnvlist_add_uint64(spa->spa_load_info, ZPOOL_CONFIG_LOAD_META_ERRORS, sle.sle_meta_count); fnvlist_add_uint64(spa->spa_load_info, ZPOOL_CONFIG_LOAD_DATA_ERRORS, sle.sle_data_count); } else { spa->spa_load_max_txg = spa->spa_uberblock.ub_txg; } if (spa_load_verify_dryrun) return (0); if (error) { if (error != ENXIO && error != EIO) error = SET_ERROR(EIO); return (error); } return (verify_ok ? 0 : EIO); } /* * Find a value in the pool props object. */ static void spa_prop_find(spa_t *spa, zpool_prop_t prop, uint64_t *val) { (void) zap_lookup(spa->spa_meta_objset, spa->spa_pool_props_object, zpool_prop_to_name(prop), sizeof (uint64_t), 1, val); } /* * Find a value in the pool directory object. */ static int spa_dir_prop(spa_t *spa, const char *name, uint64_t *val, boolean_t log_enoent) { int error = zap_lookup(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT, name, sizeof (uint64_t), 1, val); if (error != 0 && (error != ENOENT || log_enoent)) { spa_load_failed(spa, "couldn't get '%s' value in MOS directory " "[error=%d]", name, error); } return (error); } static int spa_vdev_err(vdev_t *vdev, vdev_aux_t aux, int err) { vdev_set_state(vdev, B_TRUE, VDEV_STATE_CANT_OPEN, aux); return (SET_ERROR(err)); } boolean_t spa_livelist_delete_check(spa_t *spa) { return (spa->spa_livelists_to_delete != 0); } static boolean_t spa_livelist_delete_cb_check(void *arg, zthr_t *z) { (void) z; spa_t *spa = arg; return (spa_livelist_delete_check(spa)); } static int delete_blkptr_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx) { spa_t *spa = arg; zio_free(spa, tx->tx_txg, bp); dsl_dir_diduse_space(tx->tx_pool->dp_free_dir, DD_USED_HEAD, -bp_get_dsize_sync(spa, bp), -BP_GET_PSIZE(bp), -BP_GET_UCSIZE(bp), tx); return (0); } static int dsl_get_next_livelist_obj(objset_t *os, uint64_t zap_obj, uint64_t *llp) { int err; zap_cursor_t zc; zap_attribute_t za; zap_cursor_init(&zc, os, zap_obj); err = zap_cursor_retrieve(&zc, &za); zap_cursor_fini(&zc); if (err == 0) *llp = za.za_first_integer; return (err); } /* * Components of livelist deletion that must be performed in syncing * context: freeing block pointers and updating the pool-wide data * structures to indicate how much work is left to do */ typedef struct sublist_delete_arg { spa_t *spa; dsl_deadlist_t *ll; uint64_t key; bplist_t *to_free; } sublist_delete_arg_t; static void sublist_delete_sync(void *arg, dmu_tx_t *tx) { sublist_delete_arg_t *sda = arg; spa_t *spa = sda->spa; dsl_deadlist_t *ll = sda->ll; uint64_t key = sda->key; bplist_t *to_free = sda->to_free; bplist_iterate(to_free, delete_blkptr_cb, spa, tx); dsl_deadlist_remove_entry(ll, key, tx); } typedef struct livelist_delete_arg { spa_t *spa; uint64_t ll_obj; uint64_t zap_obj; } livelist_delete_arg_t; static void livelist_delete_sync(void *arg, dmu_tx_t *tx) { livelist_delete_arg_t *lda = arg; spa_t *spa = lda->spa; uint64_t ll_obj = lda->ll_obj; uint64_t zap_obj = lda->zap_obj; objset_t *mos = spa->spa_meta_objset; uint64_t count; /* free the livelist and decrement the feature count */ VERIFY0(zap_remove_int(mos, zap_obj, ll_obj, tx)); dsl_deadlist_free(mos, ll_obj, tx); spa_feature_decr(spa, SPA_FEATURE_LIVELIST, tx); VERIFY0(zap_count(mos, zap_obj, &count)); if (count == 0) { /* no more livelists to delete */ VERIFY0(zap_remove(mos, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_DELETED_CLONES, tx)); VERIFY0(zap_destroy(mos, zap_obj, tx)); spa->spa_livelists_to_delete = 0; spa_notify_waiters(spa); } } /* * Load in the value for the livelist to be removed and open it. Then, * load its first sublist and determine which block pointers should actually * be freed. Then, call a synctask which performs the actual frees and updates * the pool-wide livelist data. */ static void spa_livelist_delete_cb(void *arg, zthr_t *z) { spa_t *spa = arg; uint64_t ll_obj = 0, count; objset_t *mos = spa->spa_meta_objset; uint64_t zap_obj = spa->spa_livelists_to_delete; /* * Determine the next livelist to delete. This function should only * be called if there is at least one deleted clone. */ VERIFY0(dsl_get_next_livelist_obj(mos, zap_obj, &ll_obj)); VERIFY0(zap_count(mos, ll_obj, &count)); if (count > 0) { dsl_deadlist_t *ll; dsl_deadlist_entry_t *dle; bplist_t to_free; ll = kmem_zalloc(sizeof (dsl_deadlist_t), KM_SLEEP); dsl_deadlist_open(ll, mos, ll_obj); dle = dsl_deadlist_first(ll); ASSERT3P(dle, !=, NULL); bplist_create(&to_free); int err = dsl_process_sub_livelist(&dle->dle_bpobj, &to_free, z, NULL); if (err == 0) { sublist_delete_arg_t sync_arg = { .spa = spa, .ll = ll, .key = dle->dle_mintxg, .to_free = &to_free }; zfs_dbgmsg("deleting sublist (id %llu) from" " livelist %llu, %lld remaining", (u_longlong_t)dle->dle_bpobj.bpo_object, (u_longlong_t)ll_obj, (longlong_t)count - 1); VERIFY0(dsl_sync_task(spa_name(spa), NULL, sublist_delete_sync, &sync_arg, 0, ZFS_SPACE_CHECK_DESTROY)); } else { VERIFY3U(err, ==, EINTR); } bplist_clear(&to_free); bplist_destroy(&to_free); dsl_deadlist_close(ll); kmem_free(ll, sizeof (dsl_deadlist_t)); } else { livelist_delete_arg_t sync_arg = { .spa = spa, .ll_obj = ll_obj, .zap_obj = zap_obj }; zfs_dbgmsg("deletion of livelist %llu completed", (u_longlong_t)ll_obj); VERIFY0(dsl_sync_task(spa_name(spa), NULL, livelist_delete_sync, &sync_arg, 0, ZFS_SPACE_CHECK_DESTROY)); } } static void spa_start_livelist_destroy_thread(spa_t *spa) { ASSERT3P(spa->spa_livelist_delete_zthr, ==, NULL); spa->spa_livelist_delete_zthr = zthr_create("z_livelist_destroy", spa_livelist_delete_cb_check, spa_livelist_delete_cb, spa, minclsyspri); } typedef struct livelist_new_arg { bplist_t *allocs; bplist_t *frees; } livelist_new_arg_t; static int livelist_track_new_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed, dmu_tx_t *tx) { ASSERT(tx == NULL); livelist_new_arg_t *lna = arg; if (bp_freed) { bplist_append(lna->frees, bp); } else { bplist_append(lna->allocs, bp); zfs_livelist_condense_new_alloc++; } return (0); } typedef struct livelist_condense_arg { spa_t *spa; bplist_t to_keep; uint64_t first_size; uint64_t next_size; } livelist_condense_arg_t; static void spa_livelist_condense_sync(void *arg, dmu_tx_t *tx) { livelist_condense_arg_t *lca = arg; spa_t *spa = lca->spa; bplist_t new_frees; dsl_dataset_t *ds = spa->spa_to_condense.ds; /* Have we been cancelled? */ if (spa->spa_to_condense.cancelled) { zfs_livelist_condense_sync_cancel++; goto out; } dsl_deadlist_entry_t *first = spa->spa_to_condense.first; dsl_deadlist_entry_t *next = spa->spa_to_condense.next; dsl_deadlist_t *ll = &ds->ds_dir->dd_livelist; /* * It's possible that the livelist was changed while the zthr was * running. Therefore, we need to check for new blkptrs in the two * entries being condensed and continue to track them in the livelist. * Because of the way we handle remapped blkptrs (see dbuf_remap_impl), * it's possible that the newly added blkptrs are FREEs or ALLOCs so * we need to sort them into two different bplists. */ uint64_t first_obj = first->dle_bpobj.bpo_object; uint64_t next_obj = next->dle_bpobj.bpo_object; uint64_t cur_first_size = first->dle_bpobj.bpo_phys->bpo_num_blkptrs; uint64_t cur_next_size = next->dle_bpobj.bpo_phys->bpo_num_blkptrs; bplist_create(&new_frees); livelist_new_arg_t new_bps = { .allocs = &lca->to_keep, .frees = &new_frees, }; if (cur_first_size > lca->first_size) { VERIFY0(livelist_bpobj_iterate_from_nofree(&first->dle_bpobj, livelist_track_new_cb, &new_bps, lca->first_size)); } if (cur_next_size > lca->next_size) { VERIFY0(livelist_bpobj_iterate_from_nofree(&next->dle_bpobj, livelist_track_new_cb, &new_bps, lca->next_size)); } dsl_deadlist_clear_entry(first, ll, tx); ASSERT(bpobj_is_empty(&first->dle_bpobj)); dsl_deadlist_remove_entry(ll, next->dle_mintxg, tx); bplist_iterate(&lca->to_keep, dsl_deadlist_insert_alloc_cb, ll, tx); bplist_iterate(&new_frees, dsl_deadlist_insert_free_cb, ll, tx); bplist_destroy(&new_frees); char dsname[ZFS_MAX_DATASET_NAME_LEN]; dsl_dataset_name(ds, dsname); zfs_dbgmsg("txg %llu condensing livelist of %s (id %llu), bpobj %llu " "(%llu blkptrs) and bpobj %llu (%llu blkptrs) -> bpobj %llu " "(%llu blkptrs)", (u_longlong_t)tx->tx_txg, dsname, (u_longlong_t)ds->ds_object, (u_longlong_t)first_obj, (u_longlong_t)cur_first_size, (u_longlong_t)next_obj, (u_longlong_t)cur_next_size, (u_longlong_t)first->dle_bpobj.bpo_object, (u_longlong_t)first->dle_bpobj.bpo_phys->bpo_num_blkptrs); out: dmu_buf_rele(ds->ds_dbuf, spa); spa->spa_to_condense.ds = NULL; bplist_clear(&lca->to_keep); bplist_destroy(&lca->to_keep); kmem_free(lca, sizeof (livelist_condense_arg_t)); spa->spa_to_condense.syncing = B_FALSE; } static void spa_livelist_condense_cb(void *arg, zthr_t *t) { while (zfs_livelist_condense_zthr_pause && !(zthr_has_waiters(t) || zthr_iscancelled(t))) delay(1); spa_t *spa = arg; dsl_deadlist_entry_t *first = spa->spa_to_condense.first; dsl_deadlist_entry_t *next = spa->spa_to_condense.next; uint64_t first_size, next_size; livelist_condense_arg_t *lca = kmem_alloc(sizeof (livelist_condense_arg_t), KM_SLEEP); bplist_create(&lca->to_keep); /* * Process the livelists (matching FREEs and ALLOCs) in open context * so we have minimal work in syncing context to condense. * * We save bpobj sizes (first_size and next_size) to use later in * syncing context to determine if entries were added to these sublists * while in open context. This is possible because the clone is still * active and open for normal writes and we want to make sure the new, * unprocessed blockpointers are inserted into the livelist normally. * * Note that dsl_process_sub_livelist() both stores the size number of * blockpointers and iterates over them while the bpobj's lock held, so * the sizes returned to us are consistent which what was actually * processed. */ int err = dsl_process_sub_livelist(&first->dle_bpobj, &lca->to_keep, t, &first_size); if (err == 0) err = dsl_process_sub_livelist(&next->dle_bpobj, &lca->to_keep, t, &next_size); if (err == 0) { while (zfs_livelist_condense_sync_pause && !(zthr_has_waiters(t) || zthr_iscancelled(t))) delay(1); dmu_tx_t *tx = dmu_tx_create_dd(spa_get_dsl(spa)->dp_mos_dir); dmu_tx_mark_netfree(tx); dmu_tx_hold_space(tx, 1); err = dmu_tx_assign(tx, TXG_NOWAIT | TXG_NOTHROTTLE); if (err == 0) { /* * Prevent the condense zthr restarting before * the synctask completes. */ spa->spa_to_condense.syncing = B_TRUE; lca->spa = spa; lca->first_size = first_size; lca->next_size = next_size; dsl_sync_task_nowait(spa_get_dsl(spa), spa_livelist_condense_sync, lca, tx); dmu_tx_commit(tx); return; } } /* * Condensing can not continue: either it was externally stopped or * we were unable to assign to a tx because the pool has run out of * space. In the second case, we'll just end up trying to condense * again in a later txg. */ ASSERT(err != 0); bplist_clear(&lca->to_keep); bplist_destroy(&lca->to_keep); kmem_free(lca, sizeof (livelist_condense_arg_t)); dmu_buf_rele(spa->spa_to_condense.ds->ds_dbuf, spa); spa->spa_to_condense.ds = NULL; if (err == EINTR) zfs_livelist_condense_zthr_cancel++; } /* * Check that there is something to condense but that a condense is not * already in progress and that condensing has not been cancelled. */ static boolean_t spa_livelist_condense_cb_check(void *arg, zthr_t *z) { (void) z; spa_t *spa = arg; if ((spa->spa_to_condense.ds != NULL) && (spa->spa_to_condense.syncing == B_FALSE) && (spa->spa_to_condense.cancelled == B_FALSE)) { return (B_TRUE); } return (B_FALSE); } static void spa_start_livelist_condensing_thread(spa_t *spa) { spa->spa_to_condense.ds = NULL; spa->spa_to_condense.first = NULL; spa->spa_to_condense.next = NULL; spa->spa_to_condense.syncing = B_FALSE; spa->spa_to_condense.cancelled = B_FALSE; ASSERT3P(spa->spa_livelist_condense_zthr, ==, NULL); spa->spa_livelist_condense_zthr = zthr_create("z_livelist_condense", spa_livelist_condense_cb_check, spa_livelist_condense_cb, spa, minclsyspri); } static void spa_spawn_aux_threads(spa_t *spa) { ASSERT(spa_writeable(spa)); ASSERT(MUTEX_HELD(&spa_namespace_lock)); spa_start_indirect_condensing_thread(spa); spa_start_livelist_destroy_thread(spa); spa_start_livelist_condensing_thread(spa); ASSERT3P(spa->spa_checkpoint_discard_zthr, ==, NULL); spa->spa_checkpoint_discard_zthr = zthr_create("z_checkpoint_discard", spa_checkpoint_discard_thread_check, spa_checkpoint_discard_thread, spa, minclsyspri); } /* * Fix up config after a partly-completed split. This is done with the * ZPOOL_CONFIG_SPLIT nvlist. Both the splitting pool and the split-off * pool have that entry in their config, but only the splitting one contains * a list of all the guids of the vdevs that are being split off. * * This function determines what to do with that list: either rejoin * all the disks to the pool, or complete the splitting process. To attempt * the rejoin, each disk that is offlined is marked online again, and * we do a reopen() call. If the vdev label for every disk that was * marked online indicates it was successfully split off (VDEV_AUX_SPLIT_POOL) * then we call vdev_split() on each disk, and complete the split. * * Otherwise we leave the config alone, with all the vdevs in place in * the original pool. */ static void spa_try_repair(spa_t *spa, nvlist_t *config) { uint_t extracted; uint64_t *glist; uint_t i, gcount; nvlist_t *nvl; vdev_t **vd; boolean_t attempt_reopen; if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_SPLIT, &nvl) != 0) return; /* check that the config is complete */ if (nvlist_lookup_uint64_array(nvl, ZPOOL_CONFIG_SPLIT_LIST, &glist, &gcount) != 0) return; vd = kmem_zalloc(gcount * sizeof (vdev_t *), KM_SLEEP); /* attempt to online all the vdevs & validate */ attempt_reopen = B_TRUE; for (i = 0; i < gcount; i++) { if (glist[i] == 0) /* vdev is hole */ continue; vd[i] = spa_lookup_by_guid(spa, glist[i], B_FALSE); if (vd[i] == NULL) { /* * Don't bother attempting to reopen the disks; * just do the split. */ attempt_reopen = B_FALSE; } else { /* attempt to re-online it */ vd[i]->vdev_offline = B_FALSE; } } if (attempt_reopen) { vdev_reopen(spa->spa_root_vdev); /* check each device to see what state it's in */ for (extracted = 0, i = 0; i < gcount; i++) { if (vd[i] != NULL && vd[i]->vdev_stat.vs_aux != VDEV_AUX_SPLIT_POOL) break; ++extracted; } } /* * If every disk has been moved to the new pool, or if we never * even attempted to look at them, then we split them off for * good. */ if (!attempt_reopen || gcount == extracted) { for (i = 0; i < gcount; i++) if (vd[i] != NULL) vdev_split(vd[i]); vdev_reopen(spa->spa_root_vdev); } kmem_free(vd, gcount * sizeof (vdev_t *)); } static int spa_load(spa_t *spa, spa_load_state_t state, spa_import_type_t type) { const char *ereport = FM_EREPORT_ZFS_POOL; int error; spa->spa_load_state = state; (void) spa_import_progress_set_state(spa_guid(spa), spa_load_state(spa)); gethrestime(&spa->spa_loaded_ts); error = spa_load_impl(spa, type, &ereport); /* * Don't count references from objsets that are already closed * and are making their way through the eviction process. */ spa_evicting_os_wait(spa); spa->spa_minref = zfs_refcount_count(&spa->spa_refcount); if (error) { if (error != EEXIST) { spa->spa_loaded_ts.tv_sec = 0; spa->spa_loaded_ts.tv_nsec = 0; } if (error != EBADF) { (void) zfs_ereport_post(ereport, spa, NULL, NULL, NULL, 0); } } spa->spa_load_state = error ? SPA_LOAD_ERROR : SPA_LOAD_NONE; spa->spa_ena = 0; (void) spa_import_progress_set_state(spa_guid(spa), spa_load_state(spa)); return (error); } #ifdef ZFS_DEBUG /* * Count the number of per-vdev ZAPs associated with all of the vdevs in the * vdev tree rooted in the given vd, and ensure that each ZAP is present in the * spa's per-vdev ZAP list. */ static uint64_t vdev_count_verify_zaps(vdev_t *vd) { spa_t *spa = vd->vdev_spa; uint64_t total = 0; if (spa_feature_is_active(vd->vdev_spa, SPA_FEATURE_AVZ_V2) && vd->vdev_root_zap != 0) { total++; ASSERT0(zap_lookup_int(spa->spa_meta_objset, spa->spa_all_vdev_zaps, vd->vdev_root_zap)); } if (vd->vdev_top_zap != 0) { total++; ASSERT0(zap_lookup_int(spa->spa_meta_objset, spa->spa_all_vdev_zaps, vd->vdev_top_zap)); } if (vd->vdev_leaf_zap != 0) { total++; ASSERT0(zap_lookup_int(spa->spa_meta_objset, spa->spa_all_vdev_zaps, vd->vdev_leaf_zap)); } for (uint64_t i = 0; i < vd->vdev_children; i++) { total += vdev_count_verify_zaps(vd->vdev_child[i]); } return (total); } #else #define vdev_count_verify_zaps(vd) ((void) sizeof (vd), 0) #endif /* * Determine whether the activity check is required. */ static boolean_t spa_activity_check_required(spa_t *spa, uberblock_t *ub, nvlist_t *label, nvlist_t *config) { uint64_t state = 0; uint64_t hostid = 0; uint64_t tryconfig_txg = 0; uint64_t tryconfig_timestamp = 0; uint16_t tryconfig_mmp_seq = 0; nvlist_t *nvinfo; if (nvlist_exists(config, ZPOOL_CONFIG_LOAD_INFO)) { nvinfo = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO); (void) nvlist_lookup_uint64(nvinfo, ZPOOL_CONFIG_MMP_TXG, &tryconfig_txg); (void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_TIMESTAMP, &tryconfig_timestamp); (void) nvlist_lookup_uint16(nvinfo, ZPOOL_CONFIG_MMP_SEQ, &tryconfig_mmp_seq); } (void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE, &state); /* * Disable the MMP activity check - This is used by zdb which * is intended to be used on potentially active pools. */ if (spa->spa_import_flags & ZFS_IMPORT_SKIP_MMP) return (B_FALSE); /* * Skip the activity check when the MMP feature is disabled. */ if (ub->ub_mmp_magic == MMP_MAGIC && ub->ub_mmp_delay == 0) return (B_FALSE); /* * If the tryconfig_ values are nonzero, they are the results of an * earlier tryimport. If they all match the uberblock we just found, * then the pool has not changed and we return false so we do not test * a second time. */ if (tryconfig_txg && tryconfig_txg == ub->ub_txg && tryconfig_timestamp && tryconfig_timestamp == ub->ub_timestamp && tryconfig_mmp_seq && tryconfig_mmp_seq == (MMP_SEQ_VALID(ub) ? MMP_SEQ(ub) : 0)) return (B_FALSE); /* * Allow the activity check to be skipped when importing the pool * on the same host which last imported it. Since the hostid from * configuration may be stale use the one read from the label. */ if (nvlist_exists(label, ZPOOL_CONFIG_HOSTID)) hostid = fnvlist_lookup_uint64(label, ZPOOL_CONFIG_HOSTID); if (hostid == spa_get_hostid(spa)) return (B_FALSE); /* * Skip the activity test when the pool was cleanly exported. */ if (state != POOL_STATE_ACTIVE) return (B_FALSE); return (B_TRUE); } /* * Nanoseconds the activity check must watch for changes on-disk. */ static uint64_t spa_activity_check_duration(spa_t *spa, uberblock_t *ub) { uint64_t import_intervals = MAX(zfs_multihost_import_intervals, 1); uint64_t multihost_interval = MSEC2NSEC( MMP_INTERVAL_OK(zfs_multihost_interval)); uint64_t import_delay = MAX(NANOSEC, import_intervals * multihost_interval); /* * Local tunables determine a minimum duration except for the case * where we know when the remote host will suspend the pool if MMP * writes do not land. * * See Big Theory comment at the top of mmp.c for the reasoning behind * these cases and times. */ ASSERT(MMP_IMPORT_SAFETY_FACTOR >= 100); if (MMP_INTERVAL_VALID(ub) && MMP_FAIL_INT_VALID(ub) && MMP_FAIL_INT(ub) > 0) { /* MMP on remote host will suspend pool after failed writes */ import_delay = MMP_FAIL_INT(ub) * MSEC2NSEC(MMP_INTERVAL(ub)) * MMP_IMPORT_SAFETY_FACTOR / 100; zfs_dbgmsg("fail_intvals>0 import_delay=%llu ub_mmp " "mmp_fails=%llu ub_mmp mmp_interval=%llu " "import_intervals=%llu", (u_longlong_t)import_delay, (u_longlong_t)MMP_FAIL_INT(ub), (u_longlong_t)MMP_INTERVAL(ub), (u_longlong_t)import_intervals); } else if (MMP_INTERVAL_VALID(ub) && MMP_FAIL_INT_VALID(ub) && MMP_FAIL_INT(ub) == 0) { /* MMP on remote host will never suspend pool */ import_delay = MAX(import_delay, (MSEC2NSEC(MMP_INTERVAL(ub)) + ub->ub_mmp_delay) * import_intervals); zfs_dbgmsg("fail_intvals=0 import_delay=%llu ub_mmp " "mmp_interval=%llu ub_mmp_delay=%llu " "import_intervals=%llu", (u_longlong_t)import_delay, (u_longlong_t)MMP_INTERVAL(ub), (u_longlong_t)ub->ub_mmp_delay, (u_longlong_t)import_intervals); } else if (MMP_VALID(ub)) { /* * zfs-0.7 compatibility case */ import_delay = MAX(import_delay, (multihost_interval + ub->ub_mmp_delay) * import_intervals); zfs_dbgmsg("import_delay=%llu ub_mmp_delay=%llu " "import_intervals=%llu leaves=%u", (u_longlong_t)import_delay, (u_longlong_t)ub->ub_mmp_delay, (u_longlong_t)import_intervals, vdev_count_leaves(spa)); } else { /* Using local tunings is the only reasonable option */ zfs_dbgmsg("pool last imported on non-MMP aware " "host using import_delay=%llu multihost_interval=%llu " "import_intervals=%llu", (u_longlong_t)import_delay, (u_longlong_t)multihost_interval, (u_longlong_t)import_intervals); } return (import_delay); } /* * Perform the import activity check. If the user canceled the import or * we detected activity then fail. */ static int spa_activity_check(spa_t *spa, uberblock_t *ub, nvlist_t *config) { uint64_t txg = ub->ub_txg; uint64_t timestamp = ub->ub_timestamp; uint64_t mmp_config = ub->ub_mmp_config; uint16_t mmp_seq = MMP_SEQ_VALID(ub) ? MMP_SEQ(ub) : 0; uint64_t import_delay; hrtime_t import_expire; nvlist_t *mmp_label = NULL; vdev_t *rvd = spa->spa_root_vdev; kcondvar_t cv; kmutex_t mtx; int error = 0; cv_init(&cv, NULL, CV_DEFAULT, NULL); mutex_init(&mtx, NULL, MUTEX_DEFAULT, NULL); mutex_enter(&mtx); /* * If ZPOOL_CONFIG_MMP_TXG is present an activity check was performed * during the earlier tryimport. If the txg recorded there is 0 then * the pool is known to be active on another host. * * Otherwise, the pool might be in use on another host. Check for * changes in the uberblocks on disk if necessary. */ if (nvlist_exists(config, ZPOOL_CONFIG_LOAD_INFO)) { nvlist_t *nvinfo = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_LOAD_INFO); if (nvlist_exists(nvinfo, ZPOOL_CONFIG_MMP_TXG) && fnvlist_lookup_uint64(nvinfo, ZPOOL_CONFIG_MMP_TXG) == 0) { vdev_uberblock_load(rvd, ub, &mmp_label); error = SET_ERROR(EREMOTEIO); goto out; } } import_delay = spa_activity_check_duration(spa, ub); /* Add a small random factor in case of simultaneous imports (0-25%) */ import_delay += import_delay * random_in_range(250) / 1000; import_expire = gethrtime() + import_delay; while (gethrtime() < import_expire) { (void) spa_import_progress_set_mmp_check(spa_guid(spa), NSEC2SEC(import_expire - gethrtime())); vdev_uberblock_load(rvd, ub, &mmp_label); if (txg != ub->ub_txg || timestamp != ub->ub_timestamp || mmp_seq != (MMP_SEQ_VALID(ub) ? MMP_SEQ(ub) : 0)) { zfs_dbgmsg("multihost activity detected " "txg %llu ub_txg %llu " "timestamp %llu ub_timestamp %llu " "mmp_config %#llx ub_mmp_config %#llx", (u_longlong_t)txg, (u_longlong_t)ub->ub_txg, (u_longlong_t)timestamp, (u_longlong_t)ub->ub_timestamp, (u_longlong_t)mmp_config, (u_longlong_t)ub->ub_mmp_config); error = SET_ERROR(EREMOTEIO); break; } if (mmp_label) { nvlist_free(mmp_label); mmp_label = NULL; } error = cv_timedwait_sig(&cv, &mtx, ddi_get_lbolt() + hz); if (error != -1) { error = SET_ERROR(EINTR); break; } error = 0; } out: mutex_exit(&mtx); mutex_destroy(&mtx); cv_destroy(&cv); /* * If the pool is determined to be active store the status in the * spa->spa_load_info nvlist. If the remote hostname or hostid are * available from configuration read from disk store them as well. * This allows 'zpool import' to generate a more useful message. * * ZPOOL_CONFIG_MMP_STATE - observed pool status (mandatory) * ZPOOL_CONFIG_MMP_HOSTNAME - hostname from the active pool * ZPOOL_CONFIG_MMP_HOSTID - hostid from the active pool */ if (error == EREMOTEIO) { const char *hostname = ""; uint64_t hostid = 0; if (mmp_label) { if (nvlist_exists(mmp_label, ZPOOL_CONFIG_HOSTNAME)) { hostname = fnvlist_lookup_string(mmp_label, ZPOOL_CONFIG_HOSTNAME); fnvlist_add_string(spa->spa_load_info, ZPOOL_CONFIG_MMP_HOSTNAME, hostname); } if (nvlist_exists(mmp_label, ZPOOL_CONFIG_HOSTID)) { hostid = fnvlist_lookup_uint64(mmp_label, ZPOOL_CONFIG_HOSTID); fnvlist_add_uint64(spa->spa_load_info, ZPOOL_CONFIG_MMP_HOSTID, hostid); } } fnvlist_add_uint64(spa->spa_load_info, ZPOOL_CONFIG_MMP_STATE, MMP_STATE_ACTIVE); fnvlist_add_uint64(spa->spa_load_info, ZPOOL_CONFIG_MMP_TXG, 0); error = spa_vdev_err(rvd, VDEV_AUX_ACTIVE, EREMOTEIO); } if (mmp_label) nvlist_free(mmp_label); return (error); } static int spa_verify_host(spa_t *spa, nvlist_t *mos_config) { uint64_t hostid; const char *hostname; uint64_t myhostid = 0; if (!spa_is_root(spa) && nvlist_lookup_uint64(mos_config, ZPOOL_CONFIG_HOSTID, &hostid) == 0) { hostname = fnvlist_lookup_string(mos_config, ZPOOL_CONFIG_HOSTNAME); myhostid = zone_get_hostid(NULL); if (hostid != 0 && myhostid != 0 && hostid != myhostid) { cmn_err(CE_WARN, "pool '%s' could not be " "loaded as it was last accessed by " "another system (host: %s hostid: 0x%llx). " "See: https://openzfs.github.io/openzfs-docs/msg/" "ZFS-8000-EY", spa_name(spa), hostname, (u_longlong_t)hostid); spa_load_failed(spa, "hostid verification failed: pool " "last accessed by host: %s (hostid: 0x%llx)", hostname, (u_longlong_t)hostid); return (SET_ERROR(EBADF)); } } return (0); } static int spa_ld_parse_config(spa_t *spa, spa_import_type_t type) { int error = 0; nvlist_t *nvtree, *nvl, *config = spa->spa_config; int parse; vdev_t *rvd; uint64_t pool_guid; const char *comment; const char *compatibility; /* * Versioning wasn't explicitly added to the label until later, so if * it's not present treat it as the initial version. */ if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION, &spa->spa_ubsync.ub_version) != 0) spa->spa_ubsync.ub_version = SPA_VERSION_INITIAL; if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID, &pool_guid)) { spa_load_failed(spa, "invalid config provided: '%s' missing", ZPOOL_CONFIG_POOL_GUID); return (SET_ERROR(EINVAL)); } /* * If we are doing an import, ensure that the pool is not already * imported by checking if its pool guid already exists in the * spa namespace. * * The only case that we allow an already imported pool to be * imported again, is when the pool is checkpointed and we want to * look at its checkpointed state from userland tools like zdb. */ #ifdef _KERNEL if ((spa->spa_load_state == SPA_LOAD_IMPORT || spa->spa_load_state == SPA_LOAD_TRYIMPORT) && spa_guid_exists(pool_guid, 0)) { #else if ((spa->spa_load_state == SPA_LOAD_IMPORT || spa->spa_load_state == SPA_LOAD_TRYIMPORT) && spa_guid_exists(pool_guid, 0) && !spa_importing_readonly_checkpoint(spa)) { #endif spa_load_failed(spa, "a pool with guid %llu is already open", (u_longlong_t)pool_guid); return (SET_ERROR(EEXIST)); } spa->spa_config_guid = pool_guid; nvlist_free(spa->spa_load_info); spa->spa_load_info = fnvlist_alloc(); ASSERT(spa->spa_comment == NULL); if (nvlist_lookup_string(config, ZPOOL_CONFIG_COMMENT, &comment) == 0) spa->spa_comment = spa_strdup(comment); ASSERT(spa->spa_compatibility == NULL); if (nvlist_lookup_string(config, ZPOOL_CONFIG_COMPATIBILITY, &compatibility) == 0) spa->spa_compatibility = spa_strdup(compatibility); (void) nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_TXG, &spa->spa_config_txg); if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_SPLIT, &nvl) == 0) spa->spa_config_splitting = fnvlist_dup(nvl); if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvtree)) { spa_load_failed(spa, "invalid config provided: '%s' missing", ZPOOL_CONFIG_VDEV_TREE); return (SET_ERROR(EINVAL)); } /* * Create "The Godfather" zio to hold all async IOs */ spa->spa_async_zio_root = kmem_alloc(max_ncpus * sizeof (void *), KM_SLEEP); for (int i = 0; i < max_ncpus; i++) { spa->spa_async_zio_root[i] = zio_root(spa, NULL, NULL, ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE | ZIO_FLAG_GODFATHER); } /* * Parse the configuration into a vdev tree. We explicitly set the * value that will be returned by spa_version() since parsing the * configuration requires knowing the version number. */ spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); parse = (type == SPA_IMPORT_EXISTING ? VDEV_ALLOC_LOAD : VDEV_ALLOC_SPLIT); error = spa_config_parse(spa, &rvd, nvtree, NULL, 0, parse); spa_config_exit(spa, SCL_ALL, FTAG); if (error != 0) { spa_load_failed(spa, "unable to parse config [error=%d]", error); return (error); } ASSERT(spa->spa_root_vdev == rvd); ASSERT3U(spa->spa_min_ashift, >=, SPA_MINBLOCKSHIFT); ASSERT3U(spa->spa_max_ashift, <=, SPA_MAXBLOCKSHIFT); if (type != SPA_IMPORT_ASSEMBLE) { ASSERT(spa_guid(spa) == pool_guid); } return (0); } /* * Recursively open all vdevs in the vdev tree. This function is called twice: * first with the untrusted config, then with the trusted config. */ static int spa_ld_open_vdevs(spa_t *spa) { int error = 0; /* * spa_missing_tvds_allowed defines how many top-level vdevs can be * missing/unopenable for the root vdev to be still considered openable. */ if (spa->spa_trust_config) { spa->spa_missing_tvds_allowed = zfs_max_missing_tvds; } else if (spa->spa_config_source == SPA_CONFIG_SRC_CACHEFILE) { spa->spa_missing_tvds_allowed = zfs_max_missing_tvds_cachefile; } else if (spa->spa_config_source == SPA_CONFIG_SRC_SCAN) { spa->spa_missing_tvds_allowed = zfs_max_missing_tvds_scan; } else { spa->spa_missing_tvds_allowed = 0; } spa->spa_missing_tvds_allowed = MAX(zfs_max_missing_tvds, spa->spa_missing_tvds_allowed); spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); error = vdev_open(spa->spa_root_vdev); spa_config_exit(spa, SCL_ALL, FTAG); if (spa->spa_missing_tvds != 0) { spa_load_note(spa, "vdev tree has %lld missing top-level " "vdevs.", (u_longlong_t)spa->spa_missing_tvds); if (spa->spa_trust_config && (spa->spa_mode & SPA_MODE_WRITE)) { /* * Although theoretically we could allow users to open * incomplete pools in RW mode, we'd need to add a lot * of extra logic (e.g. adjust pool space to account * for missing vdevs). * This limitation also prevents users from accidentally * opening the pool in RW mode during data recovery and * damaging it further. */ spa_load_note(spa, "pools with missing top-level " "vdevs can only be opened in read-only mode."); error = SET_ERROR(ENXIO); } else { spa_load_note(spa, "current settings allow for maximum " "%lld missing top-level vdevs at this stage.", (u_longlong_t)spa->spa_missing_tvds_allowed); } } if (error != 0) { spa_load_failed(spa, "unable to open vdev tree [error=%d]", error); } if (spa->spa_missing_tvds != 0 || error != 0) vdev_dbgmsg_print_tree(spa->spa_root_vdev, 2); return (error); } /* * We need to validate the vdev labels against the configuration that * we have in hand. This function is called twice: first with an untrusted * config, then with a trusted config. The validation is more strict when the * config is trusted. */ static int spa_ld_validate_vdevs(spa_t *spa) { int error = 0; vdev_t *rvd = spa->spa_root_vdev; spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); error = vdev_validate(rvd); spa_config_exit(spa, SCL_ALL, FTAG); if (error != 0) { spa_load_failed(spa, "vdev_validate failed [error=%d]", error); return (error); } if (rvd->vdev_state <= VDEV_STATE_CANT_OPEN) { spa_load_failed(spa, "cannot open vdev tree after invalidating " "some vdevs"); vdev_dbgmsg_print_tree(rvd, 2); return (SET_ERROR(ENXIO)); } return (0); } static void spa_ld_select_uberblock_done(spa_t *spa, uberblock_t *ub) { spa->spa_state = POOL_STATE_ACTIVE; spa->spa_ubsync = spa->spa_uberblock; spa->spa_verify_min_txg = spa->spa_extreme_rewind ? TXG_INITIAL - 1 : spa_last_synced_txg(spa) - TXG_DEFER_SIZE - 1; spa->spa_first_txg = spa->spa_last_ubsync_txg ? spa->spa_last_ubsync_txg : spa_last_synced_txg(spa) + 1; spa->spa_claim_max_txg = spa->spa_first_txg; spa->spa_prev_software_version = ub->ub_software_version; } static int spa_ld_select_uberblock(spa_t *spa, spa_import_type_t type) { vdev_t *rvd = spa->spa_root_vdev; nvlist_t *label; uberblock_t *ub = &spa->spa_uberblock; boolean_t activity_check = B_FALSE; /* * If we are opening the checkpointed state of the pool by * rewinding to it, at this point we will have written the * checkpointed uberblock to the vdev labels, so searching * the labels will find the right uberblock. However, if * we are opening the checkpointed state read-only, we have * not modified the labels. Therefore, we must ignore the * labels and continue using the spa_uberblock that was set * by spa_ld_checkpoint_rewind. * * Note that it would be fine to ignore the labels when * rewinding (opening writeable) as well. However, if we * crash just after writing the labels, we will end up * searching the labels. Doing so in the common case means * that this code path gets exercised normally, rather than * just in the edge case. */ if (ub->ub_checkpoint_txg != 0 && spa_importing_readonly_checkpoint(spa)) { spa_ld_select_uberblock_done(spa, ub); return (0); } /* * Find the best uberblock. */ vdev_uberblock_load(rvd, ub, &label); /* * If we weren't able to find a single valid uberblock, return failure. */ if (ub->ub_txg == 0) { nvlist_free(label); spa_load_failed(spa, "no valid uberblock found"); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, ENXIO)); } if (spa->spa_load_max_txg != UINT64_MAX) { (void) spa_import_progress_set_max_txg(spa_guid(spa), (u_longlong_t)spa->spa_load_max_txg); } spa_load_note(spa, "using uberblock with txg=%llu", (u_longlong_t)ub->ub_txg); /* * For pools which have the multihost property on determine if the * pool is truly inactive and can be safely imported. Prevent * hosts which don't have a hostid set from importing the pool. */ activity_check = spa_activity_check_required(spa, ub, label, spa->spa_config); if (activity_check) { if (ub->ub_mmp_magic == MMP_MAGIC && ub->ub_mmp_delay && spa_get_hostid(spa) == 0) { nvlist_free(label); fnvlist_add_uint64(spa->spa_load_info, ZPOOL_CONFIG_MMP_STATE, MMP_STATE_NO_HOSTID); return (spa_vdev_err(rvd, VDEV_AUX_ACTIVE, EREMOTEIO)); } int error = spa_activity_check(spa, ub, spa->spa_config); if (error) { nvlist_free(label); return (error); } fnvlist_add_uint64(spa->spa_load_info, ZPOOL_CONFIG_MMP_STATE, MMP_STATE_INACTIVE); fnvlist_add_uint64(spa->spa_load_info, ZPOOL_CONFIG_MMP_TXG, ub->ub_txg); fnvlist_add_uint16(spa->spa_load_info, ZPOOL_CONFIG_MMP_SEQ, (MMP_SEQ_VALID(ub) ? MMP_SEQ(ub) : 0)); } /* * If the pool has an unsupported version we can't open it. */ if (!SPA_VERSION_IS_SUPPORTED(ub->ub_version)) { nvlist_free(label); spa_load_failed(spa, "version %llu is not supported", (u_longlong_t)ub->ub_version); return (spa_vdev_err(rvd, VDEV_AUX_VERSION_NEWER, ENOTSUP)); } if (ub->ub_version >= SPA_VERSION_FEATURES) { nvlist_t *features; /* * If we weren't able to find what's necessary for reading the * MOS in the label, return failure. */ if (label == NULL) { spa_load_failed(spa, "label config unavailable"); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, ENXIO)); } if (nvlist_lookup_nvlist(label, ZPOOL_CONFIG_FEATURES_FOR_READ, &features) != 0) { nvlist_free(label); spa_load_failed(spa, "invalid label: '%s' missing", ZPOOL_CONFIG_FEATURES_FOR_READ); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, ENXIO)); } /* * Update our in-core representation with the definitive values * from the label. */ nvlist_free(spa->spa_label_features); spa->spa_label_features = fnvlist_dup(features); } nvlist_free(label); /* * Look through entries in the label nvlist's features_for_read. If * there is a feature listed there which we don't understand then we * cannot open a pool. */ if (ub->ub_version >= SPA_VERSION_FEATURES) { nvlist_t *unsup_feat; unsup_feat = fnvlist_alloc(); for (nvpair_t *nvp = nvlist_next_nvpair(spa->spa_label_features, NULL); nvp != NULL; nvp = nvlist_next_nvpair(spa->spa_label_features, nvp)) { if (!zfeature_is_supported(nvpair_name(nvp))) { fnvlist_add_string(unsup_feat, nvpair_name(nvp), ""); } } if (!nvlist_empty(unsup_feat)) { fnvlist_add_nvlist(spa->spa_load_info, ZPOOL_CONFIG_UNSUP_FEAT, unsup_feat); nvlist_free(unsup_feat); spa_load_failed(spa, "some features are unsupported"); return (spa_vdev_err(rvd, VDEV_AUX_UNSUP_FEAT, ENOTSUP)); } nvlist_free(unsup_feat); } if (type != SPA_IMPORT_ASSEMBLE && spa->spa_config_splitting) { spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); spa_try_repair(spa, spa->spa_config); spa_config_exit(spa, SCL_ALL, FTAG); nvlist_free(spa->spa_config_splitting); spa->spa_config_splitting = NULL; } /* * Initialize internal SPA structures. */ spa_ld_select_uberblock_done(spa, ub); return (0); } static int spa_ld_open_rootbp(spa_t *spa) { int error = 0; vdev_t *rvd = spa->spa_root_vdev; error = dsl_pool_init(spa, spa->spa_first_txg, &spa->spa_dsl_pool); if (error != 0) { spa_load_failed(spa, "unable to open rootbp in dsl_pool_init " "[error=%d]", error); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); } spa->spa_meta_objset = spa->spa_dsl_pool->dp_meta_objset; return (0); } static int spa_ld_trusted_config(spa_t *spa, spa_import_type_t type, boolean_t reloading) { vdev_t *mrvd, *rvd = spa->spa_root_vdev; nvlist_t *nv, *mos_config, *policy; int error = 0, copy_error; uint64_t healthy_tvds, healthy_tvds_mos; uint64_t mos_config_txg; if (spa_dir_prop(spa, DMU_POOL_CONFIG, &spa->spa_config_object, B_TRUE) != 0) return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); /* * If we're assembling a pool from a split, the config provided is * already trusted so there is nothing to do. */ if (type == SPA_IMPORT_ASSEMBLE) return (0); healthy_tvds = spa_healthy_core_tvds(spa); if (load_nvlist(spa, spa->spa_config_object, &mos_config) != 0) { spa_load_failed(spa, "unable to retrieve MOS config"); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); } /* * If we are doing an open, pool owner wasn't verified yet, thus do * the verification here. */ if (spa->spa_load_state == SPA_LOAD_OPEN) { error = spa_verify_host(spa, mos_config); if (error != 0) { nvlist_free(mos_config); return (error); } } nv = fnvlist_lookup_nvlist(mos_config, ZPOOL_CONFIG_VDEV_TREE); spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); /* * Build a new vdev tree from the trusted config */ error = spa_config_parse(spa, &mrvd, nv, NULL, 0, VDEV_ALLOC_LOAD); if (error != 0) { nvlist_free(mos_config); spa_config_exit(spa, SCL_ALL, FTAG); spa_load_failed(spa, "spa_config_parse failed [error=%d]", error); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, error)); } /* * Vdev paths in the MOS may be obsolete. If the untrusted config was * obtained by scanning /dev/dsk, then it will have the right vdev * paths. We update the trusted MOS config with this information. * We first try to copy the paths with vdev_copy_path_strict, which * succeeds only when both configs have exactly the same vdev tree. * If that fails, we fall back to a more flexible method that has a * best effort policy. */ copy_error = vdev_copy_path_strict(rvd, mrvd); if (copy_error != 0 || spa_load_print_vdev_tree) { spa_load_note(spa, "provided vdev tree:"); vdev_dbgmsg_print_tree(rvd, 2); spa_load_note(spa, "MOS vdev tree:"); vdev_dbgmsg_print_tree(mrvd, 2); } if (copy_error != 0) { spa_load_note(spa, "vdev_copy_path_strict failed, falling " "back to vdev_copy_path_relaxed"); vdev_copy_path_relaxed(rvd, mrvd); } vdev_close(rvd); vdev_free(rvd); spa->spa_root_vdev = mrvd; rvd = mrvd; spa_config_exit(spa, SCL_ALL, FTAG); /* * We will use spa_config if we decide to reload the spa or if spa_load * fails and we rewind. We must thus regenerate the config using the * MOS information with the updated paths. ZPOOL_LOAD_POLICY is used to * pass settings on how to load the pool and is not stored in the MOS. * We copy it over to our new, trusted config. */ mos_config_txg = fnvlist_lookup_uint64(mos_config, ZPOOL_CONFIG_POOL_TXG); nvlist_free(mos_config); mos_config = spa_config_generate(spa, NULL, mos_config_txg, B_FALSE); if (nvlist_lookup_nvlist(spa->spa_config, ZPOOL_LOAD_POLICY, &policy) == 0) fnvlist_add_nvlist(mos_config, ZPOOL_LOAD_POLICY, policy); spa_config_set(spa, mos_config); spa->spa_config_source = SPA_CONFIG_SRC_MOS; /* * Now that we got the config from the MOS, we should be more strict * in checking blkptrs and can make assumptions about the consistency * of the vdev tree. spa_trust_config must be set to true before opening * vdevs in order for them to be writeable. */ spa->spa_trust_config = B_TRUE; /* * Open and validate the new vdev tree */ error = spa_ld_open_vdevs(spa); if (error != 0) return (error); error = spa_ld_validate_vdevs(spa); if (error != 0) return (error); if (copy_error != 0 || spa_load_print_vdev_tree) { spa_load_note(spa, "final vdev tree:"); vdev_dbgmsg_print_tree(rvd, 2); } if (spa->spa_load_state != SPA_LOAD_TRYIMPORT && !spa->spa_extreme_rewind && zfs_max_missing_tvds == 0) { /* * Sanity check to make sure that we are indeed loading the * latest uberblock. If we missed SPA_SYNC_MIN_VDEVS tvds * in the config provided and they happened to be the only ones * to have the latest uberblock, we could involuntarily perform * an extreme rewind. */ healthy_tvds_mos = spa_healthy_core_tvds(spa); if (healthy_tvds_mos - healthy_tvds >= SPA_SYNC_MIN_VDEVS) { spa_load_note(spa, "config provided misses too many " "top-level vdevs compared to MOS (%lld vs %lld). ", (u_longlong_t)healthy_tvds, (u_longlong_t)healthy_tvds_mos); spa_load_note(spa, "vdev tree:"); vdev_dbgmsg_print_tree(rvd, 2); if (reloading) { spa_load_failed(spa, "config was already " "provided from MOS. Aborting."); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); } spa_load_note(spa, "spa must be reloaded using MOS " "config"); return (SET_ERROR(EAGAIN)); } } error = spa_check_for_missing_logs(spa); if (error != 0) return (spa_vdev_err(rvd, VDEV_AUX_BAD_GUID_SUM, ENXIO)); if (rvd->vdev_guid_sum != spa->spa_uberblock.ub_guid_sum) { spa_load_failed(spa, "uberblock guid sum doesn't match MOS " "guid sum (%llu != %llu)", (u_longlong_t)spa->spa_uberblock.ub_guid_sum, (u_longlong_t)rvd->vdev_guid_sum); return (spa_vdev_err(rvd, VDEV_AUX_BAD_GUID_SUM, ENXIO)); } return (0); } static int spa_ld_open_indirect_vdev_metadata(spa_t *spa) { int error = 0; vdev_t *rvd = spa->spa_root_vdev; /* * Everything that we read before spa_remove_init() must be stored * on concreted vdevs. Therefore we do this as early as possible. */ error = spa_remove_init(spa); if (error != 0) { spa_load_failed(spa, "spa_remove_init failed [error=%d]", error); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); } /* * Retrieve information needed to condense indirect vdev mappings. */ error = spa_condense_init(spa); if (error != 0) { spa_load_failed(spa, "spa_condense_init failed [error=%d]", error); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, error)); } return (0); } static int spa_ld_check_features(spa_t *spa, boolean_t *missing_feat_writep) { int error = 0; vdev_t *rvd = spa->spa_root_vdev; if (spa_version(spa) >= SPA_VERSION_FEATURES) { boolean_t missing_feat_read = B_FALSE; nvlist_t *unsup_feat, *enabled_feat; if (spa_dir_prop(spa, DMU_POOL_FEATURES_FOR_READ, &spa->spa_feat_for_read_obj, B_TRUE) != 0) { return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); } if (spa_dir_prop(spa, DMU_POOL_FEATURES_FOR_WRITE, &spa->spa_feat_for_write_obj, B_TRUE) != 0) { return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); } if (spa_dir_prop(spa, DMU_POOL_FEATURE_DESCRIPTIONS, &spa->spa_feat_desc_obj, B_TRUE) != 0) { return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); } enabled_feat = fnvlist_alloc(); unsup_feat = fnvlist_alloc(); if (!spa_features_check(spa, B_FALSE, unsup_feat, enabled_feat)) missing_feat_read = B_TRUE; if (spa_writeable(spa) || spa->spa_load_state == SPA_LOAD_TRYIMPORT) { if (!spa_features_check(spa, B_TRUE, unsup_feat, enabled_feat)) { *missing_feat_writep = B_TRUE; } } fnvlist_add_nvlist(spa->spa_load_info, ZPOOL_CONFIG_ENABLED_FEAT, enabled_feat); if (!nvlist_empty(unsup_feat)) { fnvlist_add_nvlist(spa->spa_load_info, ZPOOL_CONFIG_UNSUP_FEAT, unsup_feat); } fnvlist_free(enabled_feat); fnvlist_free(unsup_feat); if (!missing_feat_read) { fnvlist_add_boolean(spa->spa_load_info, ZPOOL_CONFIG_CAN_RDONLY); } /* * If the state is SPA_LOAD_TRYIMPORT, our objective is * twofold: to determine whether the pool is available for * import in read-write mode and (if it is not) whether the * pool is available for import in read-only mode. If the pool * is available for import in read-write mode, it is displayed * as available in userland; if it is not available for import * in read-only mode, it is displayed as unavailable in * userland. If the pool is available for import in read-only * mode but not read-write mode, it is displayed as unavailable * in userland with a special note that the pool is actually * available for open in read-only mode. * * As a result, if the state is SPA_LOAD_TRYIMPORT and we are * missing a feature for write, we must first determine whether * the pool can be opened read-only before returning to * userland in order to know whether to display the * abovementioned note. */ if (missing_feat_read || (*missing_feat_writep && spa_writeable(spa))) { spa_load_failed(spa, "pool uses unsupported features"); return (spa_vdev_err(rvd, VDEV_AUX_UNSUP_FEAT, ENOTSUP)); } /* * Load refcounts for ZFS features from disk into an in-memory * cache during SPA initialization. */ for (spa_feature_t i = 0; i < SPA_FEATURES; i++) { uint64_t refcount; error = feature_get_refcount_from_disk(spa, &spa_feature_table[i], &refcount); if (error == 0) { spa->spa_feat_refcount_cache[i] = refcount; } else if (error == ENOTSUP) { spa->spa_feat_refcount_cache[i] = SPA_FEATURE_DISABLED; } else { spa_load_failed(spa, "error getting refcount " "for feature %s [error=%d]", spa_feature_table[i].fi_guid, error); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); } } } if (spa_feature_is_active(spa, SPA_FEATURE_ENABLED_TXG)) { if (spa_dir_prop(spa, DMU_POOL_FEATURE_ENABLED_TXG, &spa->spa_feat_enabled_txg_obj, B_TRUE) != 0) return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); } /* * Encryption was added before bookmark_v2, even though bookmark_v2 * is now a dependency. If this pool has encryption enabled without * bookmark_v2, trigger an errata message. */ if (spa_feature_is_enabled(spa, SPA_FEATURE_ENCRYPTION) && !spa_feature_is_enabled(spa, SPA_FEATURE_BOOKMARK_V2)) { spa->spa_errata = ZPOOL_ERRATA_ZOL_8308_ENCRYPTION; } return (0); } static int spa_ld_load_special_directories(spa_t *spa) { int error = 0; vdev_t *rvd = spa->spa_root_vdev; spa->spa_is_initializing = B_TRUE; error = dsl_pool_open(spa->spa_dsl_pool); spa->spa_is_initializing = B_FALSE; if (error != 0) { spa_load_failed(spa, "dsl_pool_open failed [error=%d]", error); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); } return (0); } static int spa_ld_get_props(spa_t *spa) { int error = 0; uint64_t obj; vdev_t *rvd = spa->spa_root_vdev; /* Grab the checksum salt from the MOS. */ error = zap_lookup(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_CHECKSUM_SALT, 1, sizeof (spa->spa_cksum_salt.zcs_bytes), spa->spa_cksum_salt.zcs_bytes); if (error == ENOENT) { /* Generate a new salt for subsequent use */ (void) random_get_pseudo_bytes(spa->spa_cksum_salt.zcs_bytes, sizeof (spa->spa_cksum_salt.zcs_bytes)); } else if (error != 0) { spa_load_failed(spa, "unable to retrieve checksum salt from " "MOS [error=%d]", error); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); } if (spa_dir_prop(spa, DMU_POOL_SYNC_BPOBJ, &obj, B_TRUE) != 0) return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); error = bpobj_open(&spa->spa_deferred_bpobj, spa->spa_meta_objset, obj); if (error != 0) { spa_load_failed(spa, "error opening deferred-frees bpobj " "[error=%d]", error); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); } /* * Load the bit that tells us to use the new accounting function * (raid-z deflation). If we have an older pool, this will not * be present. */ error = spa_dir_prop(spa, DMU_POOL_DEFLATE, &spa->spa_deflate, B_FALSE); if (error != 0 && error != ENOENT) return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); error = spa_dir_prop(spa, DMU_POOL_CREATION_VERSION, &spa->spa_creation_version, B_FALSE); if (error != 0 && error != ENOENT) return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); /* * Load the persistent error log. If we have an older pool, this will * not be present. */ error = spa_dir_prop(spa, DMU_POOL_ERRLOG_LAST, &spa->spa_errlog_last, B_FALSE); if (error != 0 && error != ENOENT) return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); error = spa_dir_prop(spa, DMU_POOL_ERRLOG_SCRUB, &spa->spa_errlog_scrub, B_FALSE); if (error != 0 && error != ENOENT) return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); /* * Load the livelist deletion field. If a livelist is queued for * deletion, indicate that in the spa */ error = spa_dir_prop(spa, DMU_POOL_DELETED_CLONES, &spa->spa_livelists_to_delete, B_FALSE); if (error != 0 && error != ENOENT) return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); /* * Load the history object. If we have an older pool, this * will not be present. */ error = spa_dir_prop(spa, DMU_POOL_HISTORY, &spa->spa_history, B_FALSE); if (error != 0 && error != ENOENT) return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); /* * Load the per-vdev ZAP map. If we have an older pool, this will not * be present; in this case, defer its creation to a later time to * avoid dirtying the MOS this early / out of sync context. See * spa_sync_config_object. */ /* The sentinel is only available in the MOS config. */ nvlist_t *mos_config; if (load_nvlist(spa, spa->spa_config_object, &mos_config) != 0) { spa_load_failed(spa, "unable to retrieve MOS config"); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); } error = spa_dir_prop(spa, DMU_POOL_VDEV_ZAP_MAP, &spa->spa_all_vdev_zaps, B_FALSE); if (error == ENOENT) { VERIFY(!nvlist_exists(mos_config, ZPOOL_CONFIG_HAS_PER_VDEV_ZAPS)); spa->spa_avz_action = AVZ_ACTION_INITIALIZE; ASSERT0(vdev_count_verify_zaps(spa->spa_root_vdev)); } else if (error != 0) { nvlist_free(mos_config); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); } else if (!nvlist_exists(mos_config, ZPOOL_CONFIG_HAS_PER_VDEV_ZAPS)) { /* * An older version of ZFS overwrote the sentinel value, so * we have orphaned per-vdev ZAPs in the MOS. Defer their * destruction to later; see spa_sync_config_object. */ spa->spa_avz_action = AVZ_ACTION_DESTROY; /* * We're assuming that no vdevs have had their ZAPs created * before this. Better be sure of it. */ ASSERT0(vdev_count_verify_zaps(spa->spa_root_vdev)); } nvlist_free(mos_config); spa->spa_delegation = zpool_prop_default_numeric(ZPOOL_PROP_DELEGATION); error = spa_dir_prop(spa, DMU_POOL_PROPS, &spa->spa_pool_props_object, B_FALSE); if (error && error != ENOENT) return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); if (error == 0) { uint64_t autoreplace = 0; spa_prop_find(spa, ZPOOL_PROP_BOOTFS, &spa->spa_bootfs); spa_prop_find(spa, ZPOOL_PROP_AUTOREPLACE, &autoreplace); spa_prop_find(spa, ZPOOL_PROP_DELEGATION, &spa->spa_delegation); spa_prop_find(spa, ZPOOL_PROP_FAILUREMODE, &spa->spa_failmode); spa_prop_find(spa, ZPOOL_PROP_AUTOEXPAND, &spa->spa_autoexpand); spa_prop_find(spa, ZPOOL_PROP_MULTIHOST, &spa->spa_multihost); spa_prop_find(spa, ZPOOL_PROP_AUTOTRIM, &spa->spa_autotrim); spa->spa_autoreplace = (autoreplace != 0); } /* * If we are importing a pool with missing top-level vdevs, * we enforce that the pool doesn't panic or get suspended on * error since the likelihood of missing data is extremely high. */ if (spa->spa_missing_tvds > 0 && spa->spa_failmode != ZIO_FAILURE_MODE_CONTINUE && spa->spa_load_state != SPA_LOAD_TRYIMPORT) { spa_load_note(spa, "forcing failmode to 'continue' " "as some top level vdevs are missing"); spa->spa_failmode = ZIO_FAILURE_MODE_CONTINUE; } return (0); } static int spa_ld_open_aux_vdevs(spa_t *spa, spa_import_type_t type) { int error = 0; vdev_t *rvd = spa->spa_root_vdev; /* * If we're assembling the pool from the split-off vdevs of * an existing pool, we don't want to attach the spares & cache * devices. */ /* * Load any hot spares for this pool. */ error = spa_dir_prop(spa, DMU_POOL_SPARES, &spa->spa_spares.sav_object, B_FALSE); if (error != 0 && error != ENOENT) return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); if (error == 0 && type != SPA_IMPORT_ASSEMBLE) { ASSERT(spa_version(spa) >= SPA_VERSION_SPARES); if (load_nvlist(spa, spa->spa_spares.sav_object, &spa->spa_spares.sav_config) != 0) { spa_load_failed(spa, "error loading spares nvlist"); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); } spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); spa_load_spares(spa); spa_config_exit(spa, SCL_ALL, FTAG); } else if (error == 0) { spa->spa_spares.sav_sync = B_TRUE; } /* * Load any level 2 ARC devices for this pool. */ error = spa_dir_prop(spa, DMU_POOL_L2CACHE, &spa->spa_l2cache.sav_object, B_FALSE); if (error != 0 && error != ENOENT) return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); if (error == 0 && type != SPA_IMPORT_ASSEMBLE) { ASSERT(spa_version(spa) >= SPA_VERSION_L2CACHE); if (load_nvlist(spa, spa->spa_l2cache.sav_object, &spa->spa_l2cache.sav_config) != 0) { spa_load_failed(spa, "error loading l2cache nvlist"); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); } spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); spa_load_l2cache(spa); spa_config_exit(spa, SCL_ALL, FTAG); } else if (error == 0) { spa->spa_l2cache.sav_sync = B_TRUE; } return (0); } static int spa_ld_load_vdev_metadata(spa_t *spa) { int error = 0; vdev_t *rvd = spa->spa_root_vdev; /* * If the 'multihost' property is set, then never allow a pool to * be imported when the system hostid is zero. The exception to * this rule is zdb which is always allowed to access pools. */ if (spa_multihost(spa) && spa_get_hostid(spa) == 0 && (spa->spa_import_flags & ZFS_IMPORT_SKIP_MMP) == 0) { fnvlist_add_uint64(spa->spa_load_info, ZPOOL_CONFIG_MMP_STATE, MMP_STATE_NO_HOSTID); return (spa_vdev_err(rvd, VDEV_AUX_ACTIVE, EREMOTEIO)); } /* * If the 'autoreplace' property is set, then post a resource notifying * the ZFS DE that it should not issue any faults for unopenable * devices. We also iterate over the vdevs, and post a sysevent for any * unopenable vdevs so that the normal autoreplace handler can take * over. */ if (spa->spa_autoreplace && spa->spa_load_state != SPA_LOAD_TRYIMPORT) { spa_check_removed(spa->spa_root_vdev); /* * For the import case, this is done in spa_import(), because * at this point we're using the spare definitions from * the MOS config, not necessarily from the userland config. */ if (spa->spa_load_state != SPA_LOAD_IMPORT) { spa_aux_check_removed(&spa->spa_spares); spa_aux_check_removed(&spa->spa_l2cache); } } /* * Load the vdev metadata such as metaslabs, DTLs, spacemap object, etc. */ error = vdev_load(rvd); if (error != 0) { spa_load_failed(spa, "vdev_load failed [error=%d]", error); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, error)); } error = spa_ld_log_spacemaps(spa); if (error != 0) { spa_load_failed(spa, "spa_ld_log_spacemaps failed [error=%d]", error); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, error)); } /* * Propagate the leaf DTLs we just loaded all the way up the vdev tree. */ spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); vdev_dtl_reassess(rvd, 0, 0, B_FALSE, B_FALSE); spa_config_exit(spa, SCL_ALL, FTAG); return (0); } static int spa_ld_load_dedup_tables(spa_t *spa) { int error = 0; vdev_t *rvd = spa->spa_root_vdev; error = ddt_load(spa); if (error != 0) { spa_load_failed(spa, "ddt_load failed [error=%d]", error); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); } return (0); } static int spa_ld_load_brt(spa_t *spa) { int error = 0; vdev_t *rvd = spa->spa_root_vdev; error = brt_load(spa); if (error != 0) { spa_load_failed(spa, "brt_load failed [error=%d]", error); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, EIO)); } return (0); } static int spa_ld_verify_logs(spa_t *spa, spa_import_type_t type, const char **ereport) { vdev_t *rvd = spa->spa_root_vdev; if (type != SPA_IMPORT_ASSEMBLE && spa_writeable(spa)) { boolean_t missing = spa_check_logs(spa); if (missing) { if (spa->spa_missing_tvds != 0) { spa_load_note(spa, "spa_check_logs failed " "so dropping the logs"); } else { *ereport = FM_EREPORT_ZFS_LOG_REPLAY; spa_load_failed(spa, "spa_check_logs failed"); return (spa_vdev_err(rvd, VDEV_AUX_BAD_LOG, ENXIO)); } } } return (0); } static int spa_ld_verify_pool_data(spa_t *spa) { int error = 0; vdev_t *rvd = spa->spa_root_vdev; /* * We've successfully opened the pool, verify that we're ready * to start pushing transactions. */ if (spa->spa_load_state != SPA_LOAD_TRYIMPORT) { error = spa_load_verify(spa); if (error != 0) { spa_load_failed(spa, "spa_load_verify failed " "[error=%d]", error); return (spa_vdev_err(rvd, VDEV_AUX_CORRUPT_DATA, error)); } } return (0); } static void spa_ld_claim_log_blocks(spa_t *spa) { dmu_tx_t *tx; dsl_pool_t *dp = spa_get_dsl(spa); /* * Claim log blocks that haven't been committed yet. * This must all happen in a single txg. * Note: spa_claim_max_txg is updated by spa_claim_notify(), * invoked from zil_claim_log_block()'s i/o done callback. * Price of rollback is that we abandon the log. */ spa->spa_claiming = B_TRUE; tx = dmu_tx_create_assigned(dp, spa_first_txg(spa)); (void) dmu_objset_find_dp(dp, dp->dp_root_dir_obj, zil_claim, tx, DS_FIND_CHILDREN); dmu_tx_commit(tx); spa->spa_claiming = B_FALSE; spa_set_log_state(spa, SPA_LOG_GOOD); } static void spa_ld_check_for_config_update(spa_t *spa, uint64_t config_cache_txg, boolean_t update_config_cache) { vdev_t *rvd = spa->spa_root_vdev; int need_update = B_FALSE; /* * If the config cache is stale, or we have uninitialized * metaslabs (see spa_vdev_add()), then update the config. * * If this is a verbatim import, trust the current * in-core spa_config and update the disk labels. */ if (update_config_cache || config_cache_txg != spa->spa_config_txg || spa->spa_load_state == SPA_LOAD_IMPORT || spa->spa_load_state == SPA_LOAD_RECOVER || (spa->spa_import_flags & ZFS_IMPORT_VERBATIM)) need_update = B_TRUE; for (int c = 0; c < rvd->vdev_children; c++) if (rvd->vdev_child[c]->vdev_ms_array == 0) need_update = B_TRUE; /* * Update the config cache asynchronously in case we're the * root pool, in which case the config cache isn't writable yet. */ if (need_update) spa_async_request(spa, SPA_ASYNC_CONFIG_UPDATE); } static void spa_ld_prepare_for_reload(spa_t *spa) { spa_mode_t mode = spa->spa_mode; int async_suspended = spa->spa_async_suspended; spa_unload(spa); spa_deactivate(spa); spa_activate(spa, mode); /* * We save the value of spa_async_suspended as it gets reset to 0 by * spa_unload(). We want to restore it back to the original value before * returning as we might be calling spa_async_resume() later. */ spa->spa_async_suspended = async_suspended; } static int spa_ld_read_checkpoint_txg(spa_t *spa) { uberblock_t checkpoint; int error = 0; ASSERT0(spa->spa_checkpoint_txg); ASSERT(MUTEX_HELD(&spa_namespace_lock)); error = zap_lookup(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_ZPOOL_CHECKPOINT, sizeof (uint64_t), sizeof (uberblock_t) / sizeof (uint64_t), &checkpoint); if (error == ENOENT) return (0); if (error != 0) return (error); ASSERT3U(checkpoint.ub_txg, !=, 0); ASSERT3U(checkpoint.ub_checkpoint_txg, !=, 0); ASSERT3U(checkpoint.ub_timestamp, !=, 0); spa->spa_checkpoint_txg = checkpoint.ub_txg; spa->spa_checkpoint_info.sci_timestamp = checkpoint.ub_timestamp; return (0); } static int spa_ld_mos_init(spa_t *spa, spa_import_type_t type) { int error = 0; ASSERT(MUTEX_HELD(&spa_namespace_lock)); ASSERT(spa->spa_config_source != SPA_CONFIG_SRC_NONE); /* * Never trust the config that is provided unless we are assembling * a pool following a split. * This means don't trust blkptrs and the vdev tree in general. This * also effectively puts the spa in read-only mode since * spa_writeable() checks for spa_trust_config to be true. * We will later load a trusted config from the MOS. */ if (type != SPA_IMPORT_ASSEMBLE) spa->spa_trust_config = B_FALSE; /* * Parse the config provided to create a vdev tree. */ error = spa_ld_parse_config(spa, type); if (error != 0) return (error); spa_import_progress_add(spa); /* * Now that we have the vdev tree, try to open each vdev. This involves * opening the underlying physical device, retrieving its geometry and * probing the vdev with a dummy I/O. The state of each vdev will be set * based on the success of those operations. After this we'll be ready * to read from the vdevs. */ error = spa_ld_open_vdevs(spa); if (error != 0) return (error); /* * Read the label of each vdev and make sure that the GUIDs stored * there match the GUIDs in the config provided. * If we're assembling a new pool that's been split off from an * existing pool, the labels haven't yet been updated so we skip * validation for now. */ if (type != SPA_IMPORT_ASSEMBLE) { error = spa_ld_validate_vdevs(spa); if (error != 0) return (error); } /* * Read all vdev labels to find the best uberblock (i.e. latest, * unless spa_load_max_txg is set) and store it in spa_uberblock. We * get the list of features required to read blkptrs in the MOS from * the vdev label with the best uberblock and verify that our version * of zfs supports them all. */ error = spa_ld_select_uberblock(spa, type); if (error != 0) return (error); /* * Pass that uberblock to the dsl_pool layer which will open the root * blkptr. This blkptr points to the latest version of the MOS and will * allow us to read its contents. */ error = spa_ld_open_rootbp(spa); if (error != 0) return (error); return (0); } static int spa_ld_checkpoint_rewind(spa_t *spa) { uberblock_t checkpoint; int error = 0; ASSERT(MUTEX_HELD(&spa_namespace_lock)); ASSERT(spa->spa_import_flags & ZFS_IMPORT_CHECKPOINT); error = zap_lookup(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_ZPOOL_CHECKPOINT, sizeof (uint64_t), sizeof (uberblock_t) / sizeof (uint64_t), &checkpoint); if (error != 0) { spa_load_failed(spa, "unable to retrieve checkpointed " "uberblock from the MOS config [error=%d]", error); if (error == ENOENT) error = ZFS_ERR_NO_CHECKPOINT; return (error); } ASSERT3U(checkpoint.ub_txg, <, spa->spa_uberblock.ub_txg); ASSERT3U(checkpoint.ub_txg, ==, checkpoint.ub_checkpoint_txg); /* * We need to update the txg and timestamp of the checkpointed * uberblock to be higher than the latest one. This ensures that * the checkpointed uberblock is selected if we were to close and * reopen the pool right after we've written it in the vdev labels. * (also see block comment in vdev_uberblock_compare) */ checkpoint.ub_txg = spa->spa_uberblock.ub_txg + 1; checkpoint.ub_timestamp = gethrestime_sec(); /* * Set current uberblock to be the checkpointed uberblock. */ spa->spa_uberblock = checkpoint; /* * If we are doing a normal rewind, then the pool is open for * writing and we sync the "updated" checkpointed uberblock to * disk. Once this is done, we've basically rewound the whole * pool and there is no way back. * * There are cases when we don't want to attempt and sync the * checkpointed uberblock to disk because we are opening a * pool as read-only. Specifically, verifying the checkpointed * state with zdb, and importing the checkpointed state to get * a "preview" of its content. */ if (spa_writeable(spa)) { vdev_t *rvd = spa->spa_root_vdev; spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); vdev_t *svd[SPA_SYNC_MIN_VDEVS] = { NULL }; int svdcount = 0; int children = rvd->vdev_children; int c0 = random_in_range(children); for (int c = 0; c < children; c++) { vdev_t *vd = rvd->vdev_child[(c0 + c) % children]; /* Stop when revisiting the first vdev */ if (c > 0 && svd[0] == vd) break; if (vd->vdev_ms_array == 0 || vd->vdev_islog || !vdev_is_concrete(vd)) continue; svd[svdcount++] = vd; if (svdcount == SPA_SYNC_MIN_VDEVS) break; } error = vdev_config_sync(svd, svdcount, spa->spa_first_txg); if (error == 0) spa->spa_last_synced_guid = rvd->vdev_guid; spa_config_exit(spa, SCL_ALL, FTAG); if (error != 0) { spa_load_failed(spa, "failed to write checkpointed " "uberblock to the vdev labels [error=%d]", error); return (error); } } return (0); } static int spa_ld_mos_with_trusted_config(spa_t *spa, spa_import_type_t type, boolean_t *update_config_cache) { int error; /* * Parse the config for pool, open and validate vdevs, * select an uberblock, and use that uberblock to open * the MOS. */ error = spa_ld_mos_init(spa, type); if (error != 0) return (error); /* * Retrieve the trusted config stored in the MOS and use it to create * a new, exact version of the vdev tree, then reopen all vdevs. */ error = spa_ld_trusted_config(spa, type, B_FALSE); if (error == EAGAIN) { if (update_config_cache != NULL) *update_config_cache = B_TRUE; /* * Redo the loading process with the trusted config if it is * too different from the untrusted config. */ spa_ld_prepare_for_reload(spa); spa_load_note(spa, "RELOADING"); error = spa_ld_mos_init(spa, type); if (error != 0) return (error); error = spa_ld_trusted_config(spa, type, B_TRUE); if (error != 0) return (error); } else if (error != 0) { return (error); } return (0); } /* * Load an existing storage pool, using the config provided. This config * describes which vdevs are part of the pool and is later validated against * partial configs present in each vdev's label and an entire copy of the * config stored in the MOS. */ static int spa_load_impl(spa_t *spa, spa_import_type_t type, const char **ereport) { int error = 0; boolean_t missing_feat_write = B_FALSE; boolean_t checkpoint_rewind = (spa->spa_import_flags & ZFS_IMPORT_CHECKPOINT); boolean_t update_config_cache = B_FALSE; ASSERT(MUTEX_HELD(&spa_namespace_lock)); ASSERT(spa->spa_config_source != SPA_CONFIG_SRC_NONE); spa_load_note(spa, "LOADING"); error = spa_ld_mos_with_trusted_config(spa, type, &update_config_cache); if (error != 0) return (error); /* * If we are rewinding to the checkpoint then we need to repeat * everything we've done so far in this function but this time * selecting the checkpointed uberblock and using that to open * the MOS. */ if (checkpoint_rewind) { /* * If we are rewinding to the checkpoint update config cache * anyway. */ update_config_cache = B_TRUE; /* * Extract the checkpointed uberblock from the current MOS * and use this as the pool's uberblock from now on. If the * pool is imported as writeable we also write the checkpoint * uberblock to the labels, making the rewind permanent. */ error = spa_ld_checkpoint_rewind(spa); if (error != 0) return (error); /* * Redo the loading process again with the * checkpointed uberblock. */ spa_ld_prepare_for_reload(spa); spa_load_note(spa, "LOADING checkpointed uberblock"); error = spa_ld_mos_with_trusted_config(spa, type, NULL); if (error != 0) return (error); } /* * Retrieve the checkpoint txg if the pool has a checkpoint. */ error = spa_ld_read_checkpoint_txg(spa); if (error != 0) return (error); /* * Retrieve the mapping of indirect vdevs. Those vdevs were removed * from the pool and their contents were re-mapped to other vdevs. Note * that everything that we read before this step must have been * rewritten on concrete vdevs after the last device removal was * initiated. Otherwise we could be reading from indirect vdevs before * we have loaded their mappings. */ error = spa_ld_open_indirect_vdev_metadata(spa); if (error != 0) return (error); /* * Retrieve the full list of active features from the MOS and check if * they are all supported. */ error = spa_ld_check_features(spa, &missing_feat_write); if (error != 0) return (error); /* * Load several special directories from the MOS needed by the dsl_pool * layer. */ error = spa_ld_load_special_directories(spa); if (error != 0) return (error); /* * Retrieve pool properties from the MOS. */ error = spa_ld_get_props(spa); if (error != 0) return (error); /* * Retrieve the list of auxiliary devices - cache devices and spares - * and open them. */ error = spa_ld_open_aux_vdevs(spa, type); if (error != 0) return (error); /* * Load the metadata for all vdevs. Also check if unopenable devices * should be autoreplaced. */ error = spa_ld_load_vdev_metadata(spa); if (error != 0) return (error); error = spa_ld_load_dedup_tables(spa); if (error != 0) return (error); error = spa_ld_load_brt(spa); if (error != 0) return (error); /* * Verify the logs now to make sure we don't have any unexpected errors * when we claim log blocks later. */ error = spa_ld_verify_logs(spa, type, ereport); if (error != 0) return (error); if (missing_feat_write) { ASSERT(spa->spa_load_state == SPA_LOAD_TRYIMPORT); /* * At this point, we know that we can open the pool in * read-only mode but not read-write mode. We now have enough * information and can return to userland. */ return (spa_vdev_err(spa->spa_root_vdev, VDEV_AUX_UNSUP_FEAT, ENOTSUP)); } /* * Traverse the last txgs to make sure the pool was left off in a safe * state. When performing an extreme rewind, we verify the whole pool, * which can take a very long time. */ error = spa_ld_verify_pool_data(spa); if (error != 0) return (error); /* * Calculate the deflated space for the pool. This must be done before * we write anything to the pool because we'd need to update the space * accounting using the deflated sizes. */ spa_update_dspace(spa); /* * We have now retrieved all the information we needed to open the * pool. If we are importing the pool in read-write mode, a few * additional steps must be performed to finish the import. */ if (spa_writeable(spa) && (spa->spa_load_state == SPA_LOAD_RECOVER || spa->spa_load_max_txg == UINT64_MAX)) { uint64_t config_cache_txg = spa->spa_config_txg; ASSERT(spa->spa_load_state != SPA_LOAD_TRYIMPORT); /* * In case of a checkpoint rewind, log the original txg * of the checkpointed uberblock. */ if (checkpoint_rewind) { spa_history_log_internal(spa, "checkpoint rewind", NULL, "rewound state to txg=%llu", (u_longlong_t)spa->spa_uberblock.ub_checkpoint_txg); } /* * Traverse the ZIL and claim all blocks. */ spa_ld_claim_log_blocks(spa); /* * Kick-off the syncing thread. */ spa->spa_sync_on = B_TRUE; txg_sync_start(spa->spa_dsl_pool); mmp_thread_start(spa); /* * Wait for all claims to sync. We sync up to the highest * claimed log block birth time so that claimed log blocks * don't appear to be from the future. spa_claim_max_txg * will have been set for us by ZIL traversal operations * performed above. */ txg_wait_synced(spa->spa_dsl_pool, spa->spa_claim_max_txg); /* * Check if we need to request an update of the config. On the * next sync, we would update the config stored in vdev labels * and the cachefile (by default /etc/zfs/zpool.cache). */ spa_ld_check_for_config_update(spa, config_cache_txg, update_config_cache); /* * Check if a rebuild was in progress and if so resume it. * Then check all DTLs to see if anything needs resilvering. * The resilver will be deferred if a rebuild was started. */ if (vdev_rebuild_active(spa->spa_root_vdev)) { vdev_rebuild_restart(spa); } else if (!dsl_scan_resilvering(spa->spa_dsl_pool) && vdev_resilver_needed(spa->spa_root_vdev, NULL, NULL)) { spa_async_request(spa, SPA_ASYNC_RESILVER); } /* * Log the fact that we booted up (so that we can detect if * we rebooted in the middle of an operation). */ spa_history_log_version(spa, "open", NULL); spa_restart_removal(spa); spa_spawn_aux_threads(spa); /* * Delete any inconsistent datasets. * * Note: * Since we may be issuing deletes for clones here, * we make sure to do so after we've spawned all the * auxiliary threads above (from which the livelist * deletion zthr is part of). */ (void) dmu_objset_find(spa_name(spa), dsl_destroy_inconsistent, NULL, DS_FIND_CHILDREN); /* * Clean up any stale temporary dataset userrefs. */ dsl_pool_clean_tmp_userrefs(spa->spa_dsl_pool); spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER); vdev_initialize_restart(spa->spa_root_vdev); vdev_trim_restart(spa->spa_root_vdev); vdev_autotrim_restart(spa); spa_config_exit(spa, SCL_CONFIG, FTAG); } spa_import_progress_remove(spa_guid(spa)); spa_async_request(spa, SPA_ASYNC_L2CACHE_REBUILD); spa_load_note(spa, "LOADED"); return (0); } static int spa_load_retry(spa_t *spa, spa_load_state_t state) { spa_mode_t mode = spa->spa_mode; spa_unload(spa); spa_deactivate(spa); spa->spa_load_max_txg = spa->spa_uberblock.ub_txg - 1; spa_activate(spa, mode); spa_async_suspend(spa); spa_load_note(spa, "spa_load_retry: rewind, max txg: %llu", (u_longlong_t)spa->spa_load_max_txg); return (spa_load(spa, state, SPA_IMPORT_EXISTING)); } /* * If spa_load() fails this function will try loading prior txg's. If * 'state' is SPA_LOAD_RECOVER and one of these loads succeeds the pool * will be rewound to that txg. If 'state' is not SPA_LOAD_RECOVER this * function will not rewind the pool and will return the same error as * spa_load(). */ static int spa_load_best(spa_t *spa, spa_load_state_t state, uint64_t max_request, int rewind_flags) { nvlist_t *loadinfo = NULL; nvlist_t *config = NULL; int load_error, rewind_error; uint64_t safe_rewind_txg; uint64_t min_txg; if (spa->spa_load_txg && state == SPA_LOAD_RECOVER) { spa->spa_load_max_txg = spa->spa_load_txg; spa_set_log_state(spa, SPA_LOG_CLEAR); } else { spa->spa_load_max_txg = max_request; if (max_request != UINT64_MAX) spa->spa_extreme_rewind = B_TRUE; } load_error = rewind_error = spa_load(spa, state, SPA_IMPORT_EXISTING); if (load_error == 0) return (0); if (load_error == ZFS_ERR_NO_CHECKPOINT) { /* * When attempting checkpoint-rewind on a pool with no * checkpoint, we should not attempt to load uberblocks * from previous txgs when spa_load fails. */ ASSERT(spa->spa_import_flags & ZFS_IMPORT_CHECKPOINT); spa_import_progress_remove(spa_guid(spa)); return (load_error); } if (spa->spa_root_vdev != NULL) config = spa_config_generate(spa, NULL, -1ULL, B_TRUE); spa->spa_last_ubsync_txg = spa->spa_uberblock.ub_txg; spa->spa_last_ubsync_txg_ts = spa->spa_uberblock.ub_timestamp; if (rewind_flags & ZPOOL_NEVER_REWIND) { nvlist_free(config); spa_import_progress_remove(spa_guid(spa)); return (load_error); } if (state == SPA_LOAD_RECOVER) { /* Price of rolling back is discarding txgs, including log */ spa_set_log_state(spa, SPA_LOG_CLEAR); } else { /* * If we aren't rolling back save the load info from our first * import attempt so that we can restore it after attempting * to rewind. */ loadinfo = spa->spa_load_info; spa->spa_load_info = fnvlist_alloc(); } spa->spa_load_max_txg = spa->spa_last_ubsync_txg; safe_rewind_txg = spa->spa_last_ubsync_txg - TXG_DEFER_SIZE; min_txg = (rewind_flags & ZPOOL_EXTREME_REWIND) ? TXG_INITIAL : safe_rewind_txg; /* * Continue as long as we're finding errors, we're still within * the acceptable rewind range, and we're still finding uberblocks */ while (rewind_error && spa->spa_uberblock.ub_txg >= min_txg && spa->spa_uberblock.ub_txg <= spa->spa_load_max_txg) { if (spa->spa_load_max_txg < safe_rewind_txg) spa->spa_extreme_rewind = B_TRUE; rewind_error = spa_load_retry(spa, state); } spa->spa_extreme_rewind = B_FALSE; spa->spa_load_max_txg = UINT64_MAX; if (config && (rewind_error || state != SPA_LOAD_RECOVER)) spa_config_set(spa, config); else nvlist_free(config); if (state == SPA_LOAD_RECOVER) { ASSERT3P(loadinfo, ==, NULL); spa_import_progress_remove(spa_guid(spa)); return (rewind_error); } else { /* Store the rewind info as part of the initial load info */ fnvlist_add_nvlist(loadinfo, ZPOOL_CONFIG_REWIND_INFO, spa->spa_load_info); /* Restore the initial load info */ fnvlist_free(spa->spa_load_info); spa->spa_load_info = loadinfo; spa_import_progress_remove(spa_guid(spa)); return (load_error); } } /* * Pool Open/Import * * The import case is identical to an open except that the configuration is sent * down from userland, instead of grabbed from the configuration cache. For the * case of an open, the pool configuration will exist in the * POOL_STATE_UNINITIALIZED state. * * The stats information (gen/count/ustats) is used to gather vdev statistics at * the same time open the pool, without having to keep around the spa_t in some * ambiguous state. */ static int spa_open_common(const char *pool, spa_t **spapp, const void *tag, nvlist_t *nvpolicy, nvlist_t **config) { spa_t *spa; spa_load_state_t state = SPA_LOAD_OPEN; int error; int locked = B_FALSE; int firstopen = B_FALSE; *spapp = NULL; /* * As disgusting as this is, we need to support recursive calls to this * function because dsl_dir_open() is called during spa_load(), and ends * up calling spa_open() again. The real fix is to figure out how to * avoid dsl_dir_open() calling this in the first place. */ if (MUTEX_NOT_HELD(&spa_namespace_lock)) { mutex_enter(&spa_namespace_lock); locked = B_TRUE; } if ((spa = spa_lookup(pool)) == NULL) { if (locked) mutex_exit(&spa_namespace_lock); return (SET_ERROR(ENOENT)); } if (spa->spa_state == POOL_STATE_UNINITIALIZED) { zpool_load_policy_t policy; firstopen = B_TRUE; zpool_get_load_policy(nvpolicy ? nvpolicy : spa->spa_config, &policy); if (policy.zlp_rewind & ZPOOL_DO_REWIND) state = SPA_LOAD_RECOVER; spa_activate(spa, spa_mode_global); if (state != SPA_LOAD_RECOVER) spa->spa_last_ubsync_txg = spa->spa_load_txg = 0; spa->spa_config_source = SPA_CONFIG_SRC_CACHEFILE; zfs_dbgmsg("spa_open_common: opening %s", pool); error = spa_load_best(spa, state, policy.zlp_txg, policy.zlp_rewind); if (error == EBADF) { /* * If vdev_validate() returns failure (indicated by * EBADF), it indicates that one of the vdevs indicates * that the pool has been exported or destroyed. If * this is the case, the config cache is out of sync and * we should remove the pool from the namespace. */ spa_unload(spa); spa_deactivate(spa); spa_write_cachefile(spa, B_TRUE, B_TRUE, B_FALSE); spa_remove(spa); if (locked) mutex_exit(&spa_namespace_lock); return (SET_ERROR(ENOENT)); } if (error) { /* * We can't open the pool, but we still have useful * information: the state of each vdev after the * attempted vdev_open(). Return this to the user. */ if (config != NULL && spa->spa_config) { *config = fnvlist_dup(spa->spa_config); fnvlist_add_nvlist(*config, ZPOOL_CONFIG_LOAD_INFO, spa->spa_load_info); } spa_unload(spa); spa_deactivate(spa); spa->spa_last_open_failed = error; if (locked) mutex_exit(&spa_namespace_lock); *spapp = NULL; return (error); } } spa_open_ref(spa, tag); if (config != NULL) *config = spa_config_generate(spa, NULL, -1ULL, B_TRUE); /* * If we've recovered the pool, pass back any information we * gathered while doing the load. */ if (state == SPA_LOAD_RECOVER && config != NULL) { fnvlist_add_nvlist(*config, ZPOOL_CONFIG_LOAD_INFO, spa->spa_load_info); } if (locked) { spa->spa_last_open_failed = 0; spa->spa_last_ubsync_txg = 0; spa->spa_load_txg = 0; mutex_exit(&spa_namespace_lock); } if (firstopen) zvol_create_minors_recursive(spa_name(spa)); *spapp = spa; return (0); } int spa_open_rewind(const char *name, spa_t **spapp, const void *tag, nvlist_t *policy, nvlist_t **config) { return (spa_open_common(name, spapp, tag, policy, config)); } int spa_open(const char *name, spa_t **spapp, const void *tag) { return (spa_open_common(name, spapp, tag, NULL, NULL)); } /* * Lookup the given spa_t, incrementing the inject count in the process, * preventing it from being exported or destroyed. */ spa_t * spa_inject_addref(char *name) { spa_t *spa; mutex_enter(&spa_namespace_lock); if ((spa = spa_lookup(name)) == NULL) { mutex_exit(&spa_namespace_lock); return (NULL); } spa->spa_inject_ref++; mutex_exit(&spa_namespace_lock); return (spa); } void spa_inject_delref(spa_t *spa) { mutex_enter(&spa_namespace_lock); spa->spa_inject_ref--; mutex_exit(&spa_namespace_lock); } /* * Add spares device information to the nvlist. */ static void spa_add_spares(spa_t *spa, nvlist_t *config) { nvlist_t **spares; uint_t i, nspares; nvlist_t *nvroot; uint64_t guid; vdev_stat_t *vs; uint_t vsc; uint64_t pool; ASSERT(spa_config_held(spa, SCL_CONFIG, RW_READER)); if (spa->spa_spares.sav_count == 0) return; nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE); VERIFY0(nvlist_lookup_nvlist_array(spa->spa_spares.sav_config, ZPOOL_CONFIG_SPARES, &spares, &nspares)); if (nspares != 0) { fnvlist_add_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, (const nvlist_t * const *)spares, nspares); VERIFY0(nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, &spares, &nspares)); /* * Go through and find any spares which have since been * repurposed as an active spare. If this is the case, update * their status appropriately. */ for (i = 0; i < nspares; i++) { guid = fnvlist_lookup_uint64(spares[i], ZPOOL_CONFIG_GUID); VERIFY0(nvlist_lookup_uint64_array(spares[i], ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &vsc)); if (spa_spare_exists(guid, &pool, NULL) && pool != 0ULL) { vs->vs_state = VDEV_STATE_CANT_OPEN; vs->vs_aux = VDEV_AUX_SPARED; } else { vs->vs_state = spa->spa_spares.sav_vdevs[i]->vdev_state; } } } } /* * Add l2cache device information to the nvlist, including vdev stats. */ static void spa_add_l2cache(spa_t *spa, nvlist_t *config) { nvlist_t **l2cache; uint_t i, j, nl2cache; nvlist_t *nvroot; uint64_t guid; vdev_t *vd; vdev_stat_t *vs; uint_t vsc; ASSERT(spa_config_held(spa, SCL_CONFIG, RW_READER)); if (spa->spa_l2cache.sav_count == 0) return; nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE); VERIFY0(nvlist_lookup_nvlist_array(spa->spa_l2cache.sav_config, ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache)); if (nl2cache != 0) { fnvlist_add_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE, (const nvlist_t * const *)l2cache, nl2cache); VERIFY0(nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache)); /* * Update level 2 cache device stats. */ for (i = 0; i < nl2cache; i++) { guid = fnvlist_lookup_uint64(l2cache[i], ZPOOL_CONFIG_GUID); vd = NULL; for (j = 0; j < spa->spa_l2cache.sav_count; j++) { if (guid == spa->spa_l2cache.sav_vdevs[j]->vdev_guid) { vd = spa->spa_l2cache.sav_vdevs[j]; break; } } ASSERT(vd != NULL); VERIFY0(nvlist_lookup_uint64_array(l2cache[i], ZPOOL_CONFIG_VDEV_STATS, (uint64_t **)&vs, &vsc)); vdev_get_stats(vd, vs); vdev_config_generate_stats(vd, l2cache[i]); } } } static void spa_feature_stats_from_disk(spa_t *spa, nvlist_t *features) { zap_cursor_t zc; zap_attribute_t za; if (spa->spa_feat_for_read_obj != 0) { for (zap_cursor_init(&zc, spa->spa_meta_objset, spa->spa_feat_for_read_obj); zap_cursor_retrieve(&zc, &za) == 0; zap_cursor_advance(&zc)) { ASSERT(za.za_integer_length == sizeof (uint64_t) && za.za_num_integers == 1); VERIFY0(nvlist_add_uint64(features, za.za_name, za.za_first_integer)); } zap_cursor_fini(&zc); } if (spa->spa_feat_for_write_obj != 0) { for (zap_cursor_init(&zc, spa->spa_meta_objset, spa->spa_feat_for_write_obj); zap_cursor_retrieve(&zc, &za) == 0; zap_cursor_advance(&zc)) { ASSERT(za.za_integer_length == sizeof (uint64_t) && za.za_num_integers == 1); VERIFY0(nvlist_add_uint64(features, za.za_name, za.za_first_integer)); } zap_cursor_fini(&zc); } } static void spa_feature_stats_from_cache(spa_t *spa, nvlist_t *features) { int i; for (i = 0; i < SPA_FEATURES; i++) { zfeature_info_t feature = spa_feature_table[i]; uint64_t refcount; if (feature_get_refcount(spa, &feature, &refcount) != 0) continue; VERIFY0(nvlist_add_uint64(features, feature.fi_guid, refcount)); } } /* * Store a list of pool features and their reference counts in the * config. * * The first time this is called on a spa, allocate a new nvlist, fetch * the pool features and reference counts from disk, then save the list * in the spa. In subsequent calls on the same spa use the saved nvlist * and refresh its values from the cached reference counts. This * ensures we don't block here on I/O on a suspended pool so 'zpool * clear' can resume the pool. */ static void spa_add_feature_stats(spa_t *spa, nvlist_t *config) { nvlist_t *features; ASSERT(spa_config_held(spa, SCL_CONFIG, RW_READER)); mutex_enter(&spa->spa_feat_stats_lock); features = spa->spa_feat_stats; if (features != NULL) { spa_feature_stats_from_cache(spa, features); } else { VERIFY0(nvlist_alloc(&features, NV_UNIQUE_NAME, KM_SLEEP)); spa->spa_feat_stats = features; spa_feature_stats_from_disk(spa, features); } VERIFY0(nvlist_add_nvlist(config, ZPOOL_CONFIG_FEATURE_STATS, features)); mutex_exit(&spa->spa_feat_stats_lock); } int spa_get_stats(const char *name, nvlist_t **config, char *altroot, size_t buflen) { int error; spa_t *spa; *config = NULL; error = spa_open_common(name, &spa, FTAG, NULL, config); if (spa != NULL) { /* * This still leaves a window of inconsistency where the spares * or l2cache devices could change and the config would be * self-inconsistent. */ spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER); if (*config != NULL) { uint64_t loadtimes[2]; loadtimes[0] = spa->spa_loaded_ts.tv_sec; loadtimes[1] = spa->spa_loaded_ts.tv_nsec; fnvlist_add_uint64_array(*config, ZPOOL_CONFIG_LOADED_TIME, loadtimes, 2); fnvlist_add_uint64(*config, ZPOOL_CONFIG_ERRCOUNT, spa_approx_errlog_size(spa)); if (spa_suspended(spa)) { fnvlist_add_uint64(*config, ZPOOL_CONFIG_SUSPENDED, spa->spa_failmode); fnvlist_add_uint64(*config, ZPOOL_CONFIG_SUSPENDED_REASON, spa->spa_suspended); } spa_add_spares(spa, *config); spa_add_l2cache(spa, *config); spa_add_feature_stats(spa, *config); } } /* * We want to get the alternate root even for faulted pools, so we cheat * and call spa_lookup() directly. */ if (altroot) { if (spa == NULL) { mutex_enter(&spa_namespace_lock); spa = spa_lookup(name); if (spa) spa_altroot(spa, altroot, buflen); else altroot[0] = '\0'; spa = NULL; mutex_exit(&spa_namespace_lock); } else { spa_altroot(spa, altroot, buflen); } } if (spa != NULL) { spa_config_exit(spa, SCL_CONFIG, FTAG); spa_close(spa, FTAG); } return (error); } /* * Validate that the auxiliary device array is well formed. We must have an * array of nvlists, each which describes a valid leaf vdev. If this is an * import (mode is VDEV_ALLOC_SPARE), then we allow corrupted spares to be * specified, as long as they are well-formed. */ static int spa_validate_aux_devs(spa_t *spa, nvlist_t *nvroot, uint64_t crtxg, int mode, spa_aux_vdev_t *sav, const char *config, uint64_t version, vdev_labeltype_t label) { nvlist_t **dev; uint_t i, ndev; vdev_t *vd; int error; ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL); /* * It's acceptable to have no devs specified. */ if (nvlist_lookup_nvlist_array(nvroot, config, &dev, &ndev) != 0) return (0); if (ndev == 0) return (SET_ERROR(EINVAL)); /* * Make sure the pool is formatted with a version that supports this * device type. */ if (spa_version(spa) < version) return (SET_ERROR(ENOTSUP)); /* * Set the pending device list so we correctly handle device in-use * checking. */ sav->sav_pending = dev; sav->sav_npending = ndev; for (i = 0; i < ndev; i++) { if ((error = spa_config_parse(spa, &vd, dev[i], NULL, 0, mode)) != 0) goto out; if (!vd->vdev_ops->vdev_op_leaf) { vdev_free(vd); error = SET_ERROR(EINVAL); goto out; } vd->vdev_top = vd; if ((error = vdev_open(vd)) == 0 && (error = vdev_label_init(vd, crtxg, label)) == 0) { fnvlist_add_uint64(dev[i], ZPOOL_CONFIG_GUID, vd->vdev_guid); } vdev_free(vd); if (error && (mode != VDEV_ALLOC_SPARE && mode != VDEV_ALLOC_L2CACHE)) goto out; else error = 0; } out: sav->sav_pending = NULL; sav->sav_npending = 0; return (error); } static int spa_validate_aux(spa_t *spa, nvlist_t *nvroot, uint64_t crtxg, int mode) { int error; ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL); if ((error = spa_validate_aux_devs(spa, nvroot, crtxg, mode, &spa->spa_spares, ZPOOL_CONFIG_SPARES, SPA_VERSION_SPARES, VDEV_LABEL_SPARE)) != 0) { return (error); } return (spa_validate_aux_devs(spa, nvroot, crtxg, mode, &spa->spa_l2cache, ZPOOL_CONFIG_L2CACHE, SPA_VERSION_L2CACHE, VDEV_LABEL_L2CACHE)); } static void spa_set_aux_vdevs(spa_aux_vdev_t *sav, nvlist_t **devs, int ndevs, const char *config) { int i; if (sav->sav_config != NULL) { nvlist_t **olddevs; uint_t oldndevs; nvlist_t **newdevs; /* * Generate new dev list by concatenating with the * current dev list. */ VERIFY0(nvlist_lookup_nvlist_array(sav->sav_config, config, &olddevs, &oldndevs)); newdevs = kmem_alloc(sizeof (void *) * (ndevs + oldndevs), KM_SLEEP); for (i = 0; i < oldndevs; i++) newdevs[i] = fnvlist_dup(olddevs[i]); for (i = 0; i < ndevs; i++) newdevs[i + oldndevs] = fnvlist_dup(devs[i]); fnvlist_remove(sav->sav_config, config); fnvlist_add_nvlist_array(sav->sav_config, config, (const nvlist_t * const *)newdevs, ndevs + oldndevs); for (i = 0; i < oldndevs + ndevs; i++) nvlist_free(newdevs[i]); kmem_free(newdevs, (oldndevs + ndevs) * sizeof (void *)); } else { /* * Generate a new dev list. */ sav->sav_config = fnvlist_alloc(); fnvlist_add_nvlist_array(sav->sav_config, config, (const nvlist_t * const *)devs, ndevs); } } /* * Stop and drop level 2 ARC devices */ void spa_l2cache_drop(spa_t *spa) { vdev_t *vd; int i; spa_aux_vdev_t *sav = &spa->spa_l2cache; for (i = 0; i < sav->sav_count; i++) { uint64_t pool; vd = sav->sav_vdevs[i]; ASSERT(vd != NULL); if (spa_l2cache_exists(vd->vdev_guid, &pool) && pool != 0ULL && l2arc_vdev_present(vd)) l2arc_remove_vdev(vd); } } /* * Verify encryption parameters for spa creation. If we are encrypting, we must * have the encryption feature flag enabled. */ static int spa_create_check_encryption_params(dsl_crypto_params_t *dcp, boolean_t has_encryption) { if (dcp->cp_crypt != ZIO_CRYPT_OFF && dcp->cp_crypt != ZIO_CRYPT_INHERIT && !has_encryption) return (SET_ERROR(ENOTSUP)); return (dmu_objset_create_crypt_check(NULL, dcp, NULL)); } /* * Pool Creation */ int spa_create(const char *pool, nvlist_t *nvroot, nvlist_t *props, nvlist_t *zplprops, dsl_crypto_params_t *dcp) { spa_t *spa; const char *altroot = NULL; vdev_t *rvd; dsl_pool_t *dp; dmu_tx_t *tx; int error = 0; uint64_t txg = TXG_INITIAL; nvlist_t **spares, **l2cache; uint_t nspares, nl2cache; uint64_t version, obj, ndraid = 0; boolean_t has_features; boolean_t has_encryption; boolean_t has_allocclass; spa_feature_t feat; const char *feat_name; const char *poolname; nvlist_t *nvl; if (props == NULL || nvlist_lookup_string(props, "tname", &poolname) != 0) poolname = (char *)pool; /* * If this pool already exists, return failure. */ mutex_enter(&spa_namespace_lock); if (spa_lookup(poolname) != NULL) { mutex_exit(&spa_namespace_lock); return (SET_ERROR(EEXIST)); } /* * Allocate a new spa_t structure. */ nvl = fnvlist_alloc(); fnvlist_add_string(nvl, ZPOOL_CONFIG_POOL_NAME, pool); (void) nvlist_lookup_string(props, zpool_prop_to_name(ZPOOL_PROP_ALTROOT), &altroot); spa = spa_add(poolname, nvl, altroot); fnvlist_free(nvl); spa_activate(spa, spa_mode_global); if (props && (error = spa_prop_validate(spa, props))) { spa_deactivate(spa); spa_remove(spa); mutex_exit(&spa_namespace_lock); return (error); } /* * Temporary pool names should never be written to disk. */ if (poolname != pool) spa->spa_import_flags |= ZFS_IMPORT_TEMP_NAME; has_features = B_FALSE; has_encryption = B_FALSE; has_allocclass = B_FALSE; for (nvpair_t *elem = nvlist_next_nvpair(props, NULL); elem != NULL; elem = nvlist_next_nvpair(props, elem)) { if (zpool_prop_feature(nvpair_name(elem))) { has_features = B_TRUE; feat_name = strchr(nvpair_name(elem), '@') + 1; VERIFY0(zfeature_lookup_name(feat_name, &feat)); if (feat == SPA_FEATURE_ENCRYPTION) has_encryption = B_TRUE; if (feat == SPA_FEATURE_ALLOCATION_CLASSES) has_allocclass = B_TRUE; } } /* verify encryption params, if they were provided */ if (dcp != NULL) { error = spa_create_check_encryption_params(dcp, has_encryption); if (error != 0) { spa_deactivate(spa); spa_remove(spa); mutex_exit(&spa_namespace_lock); return (error); } } if (!has_allocclass && zfs_special_devs(nvroot, NULL)) { spa_deactivate(spa); spa_remove(spa); mutex_exit(&spa_namespace_lock); return (ENOTSUP); } if (has_features || nvlist_lookup_uint64(props, zpool_prop_to_name(ZPOOL_PROP_VERSION), &version) != 0) { version = SPA_VERSION; } ASSERT(SPA_VERSION_IS_SUPPORTED(version)); spa->spa_first_txg = txg; spa->spa_uberblock.ub_txg = txg - 1; spa->spa_uberblock.ub_version = version; spa->spa_ubsync = spa->spa_uberblock; spa->spa_load_state = SPA_LOAD_CREATE; spa->spa_removing_phys.sr_state = DSS_NONE; spa->spa_removing_phys.sr_removing_vdev = -1; spa->spa_removing_phys.sr_prev_indirect_vdev = -1; spa->spa_indirect_vdevs_loaded = B_TRUE; /* * Create "The Godfather" zio to hold all async IOs */ spa->spa_async_zio_root = kmem_alloc(max_ncpus * sizeof (void *), KM_SLEEP); for (int i = 0; i < max_ncpus; i++) { spa->spa_async_zio_root[i] = zio_root(spa, NULL, NULL, ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE | ZIO_FLAG_GODFATHER); } /* * Create the root vdev. */ spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); error = spa_config_parse(spa, &rvd, nvroot, NULL, 0, VDEV_ALLOC_ADD); ASSERT(error != 0 || rvd != NULL); ASSERT(error != 0 || spa->spa_root_vdev == rvd); if (error == 0 && !zfs_allocatable_devs(nvroot)) error = SET_ERROR(EINVAL); if (error == 0 && (error = vdev_create(rvd, txg, B_FALSE)) == 0 && (error = vdev_draid_spare_create(nvroot, rvd, &ndraid, 0)) == 0 && (error = spa_validate_aux(spa, nvroot, txg, VDEV_ALLOC_ADD)) == 0) { /* * instantiate the metaslab groups (this will dirty the vdevs) * we can no longer error exit past this point */ for (int c = 0; error == 0 && c < rvd->vdev_children; c++) { vdev_t *vd = rvd->vdev_child[c]; vdev_metaslab_set_size(vd); vdev_expand(vd, txg); } } spa_config_exit(spa, SCL_ALL, FTAG); if (error != 0) { spa_unload(spa); spa_deactivate(spa); spa_remove(spa); mutex_exit(&spa_namespace_lock); return (error); } /* * Get the list of spares, if specified. */ if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, &spares, &nspares) == 0) { spa->spa_spares.sav_config = fnvlist_alloc(); fnvlist_add_nvlist_array(spa->spa_spares.sav_config, ZPOOL_CONFIG_SPARES, (const nvlist_t * const *)spares, nspares); spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); spa_load_spares(spa); spa_config_exit(spa, SCL_ALL, FTAG); spa->spa_spares.sav_sync = B_TRUE; } /* * Get the list of level 2 cache devices, if specified. */ if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache) == 0) { VERIFY0(nvlist_alloc(&spa->spa_l2cache.sav_config, NV_UNIQUE_NAME, KM_SLEEP)); fnvlist_add_nvlist_array(spa->spa_l2cache.sav_config, ZPOOL_CONFIG_L2CACHE, (const nvlist_t * const *)l2cache, nl2cache); spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); spa_load_l2cache(spa); spa_config_exit(spa, SCL_ALL, FTAG); spa->spa_l2cache.sav_sync = B_TRUE; } spa->spa_is_initializing = B_TRUE; spa->spa_dsl_pool = dp = dsl_pool_create(spa, zplprops, dcp, txg); spa->spa_is_initializing = B_FALSE; /* * Create DDTs (dedup tables). */ ddt_create(spa); /* * Create BRT table and BRT table object. */ brt_create(spa); spa_update_dspace(spa); tx = dmu_tx_create_assigned(dp, txg); /* * Create the pool's history object. */ if (version >= SPA_VERSION_ZPOOL_HISTORY && !spa->spa_history) spa_history_create_obj(spa, tx); spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_CREATE); spa_history_log_version(spa, "create", tx); /* * Create the pool config object. */ spa->spa_config_object = dmu_object_alloc(spa->spa_meta_objset, DMU_OT_PACKED_NVLIST, SPA_CONFIG_BLOCKSIZE, DMU_OT_PACKED_NVLIST_SIZE, sizeof (uint64_t), tx); if (zap_add(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_CONFIG, sizeof (uint64_t), 1, &spa->spa_config_object, tx) != 0) { cmn_err(CE_PANIC, "failed to add pool config"); } if (zap_add(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_CREATION_VERSION, sizeof (uint64_t), 1, &version, tx) != 0) { cmn_err(CE_PANIC, "failed to add pool version"); } /* Newly created pools with the right version are always deflated. */ if (version >= SPA_VERSION_RAIDZ_DEFLATE) { spa->spa_deflate = TRUE; if (zap_add(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_DEFLATE, sizeof (uint64_t), 1, &spa->spa_deflate, tx) != 0) { cmn_err(CE_PANIC, "failed to add deflate"); } } /* * Create the deferred-free bpobj. Turn off compression * because sync-to-convergence takes longer if the blocksize * keeps changing. */ obj = bpobj_alloc(spa->spa_meta_objset, 1 << 14, tx); dmu_object_set_compress(spa->spa_meta_objset, obj, ZIO_COMPRESS_OFF, tx); if (zap_add(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_SYNC_BPOBJ, sizeof (uint64_t), 1, &obj, tx) != 0) { cmn_err(CE_PANIC, "failed to add bpobj"); } VERIFY3U(0, ==, bpobj_open(&spa->spa_deferred_bpobj, spa->spa_meta_objset, obj)); /* * Generate some random noise for salted checksums to operate on. */ (void) random_get_pseudo_bytes(spa->spa_cksum_salt.zcs_bytes, sizeof (spa->spa_cksum_salt.zcs_bytes)); /* * Set pool properties. */ spa->spa_bootfs = zpool_prop_default_numeric(ZPOOL_PROP_BOOTFS); spa->spa_delegation = zpool_prop_default_numeric(ZPOOL_PROP_DELEGATION); spa->spa_failmode = zpool_prop_default_numeric(ZPOOL_PROP_FAILUREMODE); spa->spa_autoexpand = zpool_prop_default_numeric(ZPOOL_PROP_AUTOEXPAND); spa->spa_multihost = zpool_prop_default_numeric(ZPOOL_PROP_MULTIHOST); spa->spa_autotrim = zpool_prop_default_numeric(ZPOOL_PROP_AUTOTRIM); if (props != NULL) { spa_configfile_set(spa, props, B_FALSE); spa_sync_props(props, tx); } for (int i = 0; i < ndraid; i++) spa_feature_incr(spa, SPA_FEATURE_DRAID, tx); dmu_tx_commit(tx); spa->spa_sync_on = B_TRUE; txg_sync_start(dp); mmp_thread_start(spa); txg_wait_synced(dp, txg); spa_spawn_aux_threads(spa); spa_write_cachefile(spa, B_FALSE, B_TRUE, B_TRUE); /* * Don't count references from objsets that are already closed * and are making their way through the eviction process. */ spa_evicting_os_wait(spa); spa->spa_minref = zfs_refcount_count(&spa->spa_refcount); spa->spa_load_state = SPA_LOAD_NONE; spa_import_os(spa); mutex_exit(&spa_namespace_lock); return (0); } /* * Import a non-root pool into the system. */ int spa_import(char *pool, nvlist_t *config, nvlist_t *props, uint64_t flags) { spa_t *spa; const char *altroot = NULL; spa_load_state_t state = SPA_LOAD_IMPORT; zpool_load_policy_t policy; spa_mode_t mode = spa_mode_global; uint64_t readonly = B_FALSE; int error; nvlist_t *nvroot; nvlist_t **spares, **l2cache; uint_t nspares, nl2cache; /* * If a pool with this name exists, return failure. */ mutex_enter(&spa_namespace_lock); if (spa_lookup(pool) != NULL) { mutex_exit(&spa_namespace_lock); return (SET_ERROR(EEXIST)); } /* * Create and initialize the spa structure. */ (void) nvlist_lookup_string(props, zpool_prop_to_name(ZPOOL_PROP_ALTROOT), &altroot); (void) nvlist_lookup_uint64(props, zpool_prop_to_name(ZPOOL_PROP_READONLY), &readonly); if (readonly) mode = SPA_MODE_READ; spa = spa_add(pool, config, altroot); spa->spa_import_flags = flags; /* * Verbatim import - Take a pool and insert it into the namespace * as if it had been loaded at boot. */ if (spa->spa_import_flags & ZFS_IMPORT_VERBATIM) { if (props != NULL) spa_configfile_set(spa, props, B_FALSE); spa_write_cachefile(spa, B_FALSE, B_TRUE, B_FALSE); spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_IMPORT); zfs_dbgmsg("spa_import: verbatim import of %s", pool); mutex_exit(&spa_namespace_lock); return (0); } spa_activate(spa, mode); /* * Don't start async tasks until we know everything is healthy. */ spa_async_suspend(spa); zpool_get_load_policy(config, &policy); if (policy.zlp_rewind & ZPOOL_DO_REWIND) state = SPA_LOAD_RECOVER; spa->spa_config_source = SPA_CONFIG_SRC_TRYIMPORT; if (state != SPA_LOAD_RECOVER) { spa->spa_last_ubsync_txg = spa->spa_load_txg = 0; zfs_dbgmsg("spa_import: importing %s", pool); } else { zfs_dbgmsg("spa_import: importing %s, max_txg=%lld " "(RECOVERY MODE)", pool, (longlong_t)policy.zlp_txg); } error = spa_load_best(spa, state, policy.zlp_txg, policy.zlp_rewind); /* * Propagate anything learned while loading the pool and pass it * back to caller (i.e. rewind info, missing devices, etc). */ fnvlist_add_nvlist(config, ZPOOL_CONFIG_LOAD_INFO, spa->spa_load_info); spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); /* * Toss any existing sparelist, as it doesn't have any validity * anymore, and conflicts with spa_has_spare(). */ if (spa->spa_spares.sav_config) { nvlist_free(spa->spa_spares.sav_config); spa->spa_spares.sav_config = NULL; spa_load_spares(spa); } if (spa->spa_l2cache.sav_config) { nvlist_free(spa->spa_l2cache.sav_config); spa->spa_l2cache.sav_config = NULL; spa_load_l2cache(spa); } nvroot = fnvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE); spa_config_exit(spa, SCL_ALL, FTAG); if (props != NULL) spa_configfile_set(spa, props, B_FALSE); if (error != 0 || (props && spa_writeable(spa) && (error = spa_prop_set(spa, props)))) { spa_unload(spa); spa_deactivate(spa); spa_remove(spa); mutex_exit(&spa_namespace_lock); return (error); } spa_async_resume(spa); /* * Override any spares and level 2 cache devices as specified by * the user, as these may have correct device names/devids, etc. */ if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, &spares, &nspares) == 0) { if (spa->spa_spares.sav_config) fnvlist_remove(spa->spa_spares.sav_config, ZPOOL_CONFIG_SPARES); else spa->spa_spares.sav_config = fnvlist_alloc(); fnvlist_add_nvlist_array(spa->spa_spares.sav_config, ZPOOL_CONFIG_SPARES, (const nvlist_t * const *)spares, nspares); spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); spa_load_spares(spa); spa_config_exit(spa, SCL_ALL, FTAG); spa->spa_spares.sav_sync = B_TRUE; } if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache) == 0) { if (spa->spa_l2cache.sav_config) fnvlist_remove(spa->spa_l2cache.sav_config, ZPOOL_CONFIG_L2CACHE); else spa->spa_l2cache.sav_config = fnvlist_alloc(); fnvlist_add_nvlist_array(spa->spa_l2cache.sav_config, ZPOOL_CONFIG_L2CACHE, (const nvlist_t * const *)l2cache, nl2cache); spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); spa_load_l2cache(spa); spa_config_exit(spa, SCL_ALL, FTAG); spa->spa_l2cache.sav_sync = B_TRUE; } /* * Check for any removed devices. */ if (spa->spa_autoreplace) { spa_aux_check_removed(&spa->spa_spares); spa_aux_check_removed(&spa->spa_l2cache); } if (spa_writeable(spa)) { /* * Update the config cache to include the newly-imported pool. */ spa_config_update(spa, SPA_CONFIG_UPDATE_POOL); } /* * It's possible that the pool was expanded while it was exported. * We kick off an async task to handle this for us. */ spa_async_request(spa, SPA_ASYNC_AUTOEXPAND); spa_history_log_version(spa, "import", NULL); spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_IMPORT); mutex_exit(&spa_namespace_lock); zvol_create_minors_recursive(pool); spa_import_os(spa); return (0); } nvlist_t * spa_tryimport(nvlist_t *tryconfig) { nvlist_t *config = NULL; const char *poolname, *cachefile; spa_t *spa; uint64_t state; int error; zpool_load_policy_t policy; if (nvlist_lookup_string(tryconfig, ZPOOL_CONFIG_POOL_NAME, &poolname)) return (NULL); if (nvlist_lookup_uint64(tryconfig, ZPOOL_CONFIG_POOL_STATE, &state)) return (NULL); /* * Create and initialize the spa structure. */ mutex_enter(&spa_namespace_lock); spa = spa_add(TRYIMPORT_NAME, tryconfig, NULL); spa_activate(spa, SPA_MODE_READ); /* * Rewind pool if a max txg was provided. */ zpool_get_load_policy(spa->spa_config, &policy); if (policy.zlp_txg != UINT64_MAX) { spa->spa_load_max_txg = policy.zlp_txg; spa->spa_extreme_rewind = B_TRUE; zfs_dbgmsg("spa_tryimport: importing %s, max_txg=%lld", poolname, (longlong_t)policy.zlp_txg); } else { zfs_dbgmsg("spa_tryimport: importing %s", poolname); } if (nvlist_lookup_string(tryconfig, ZPOOL_CONFIG_CACHEFILE, &cachefile) == 0) { zfs_dbgmsg("spa_tryimport: using cachefile '%s'", cachefile); spa->spa_config_source = SPA_CONFIG_SRC_CACHEFILE; } else { spa->spa_config_source = SPA_CONFIG_SRC_SCAN; } /* * spa_import() relies on a pool config fetched by spa_try_import() * for spare/cache devices. Import flags are not passed to * spa_tryimport(), which makes it return early due to a missing log * device and missing retrieving the cache device and spare eventually. * Passing ZFS_IMPORT_MISSING_LOG to spa_tryimport() makes it fetch * the correct configuration regardless of the missing log device. */ spa->spa_import_flags |= ZFS_IMPORT_MISSING_LOG; error = spa_load(spa, SPA_LOAD_TRYIMPORT, SPA_IMPORT_EXISTING); /* * If 'tryconfig' was at least parsable, return the current config. */ if (spa->spa_root_vdev != NULL) { config = spa_config_generate(spa, NULL, -1ULL, B_TRUE); fnvlist_add_string(config, ZPOOL_CONFIG_POOL_NAME, poolname); fnvlist_add_uint64(config, ZPOOL_CONFIG_POOL_STATE, state); fnvlist_add_uint64(config, ZPOOL_CONFIG_TIMESTAMP, spa->spa_uberblock.ub_timestamp); fnvlist_add_nvlist(config, ZPOOL_CONFIG_LOAD_INFO, spa->spa_load_info); fnvlist_add_uint64(config, ZPOOL_CONFIG_ERRATA, spa->spa_errata); /* * If the bootfs property exists on this pool then we * copy it out so that external consumers can tell which * pools are bootable. */ if ((!error || error == EEXIST) && spa->spa_bootfs) { char *tmpname = kmem_alloc(MAXPATHLEN, KM_SLEEP); /* * We have to play games with the name since the * pool was opened as TRYIMPORT_NAME. */ if (dsl_dsobj_to_dsname(spa_name(spa), spa->spa_bootfs, tmpname) == 0) { char *cp; char *dsname; dsname = kmem_alloc(MAXPATHLEN, KM_SLEEP); cp = strchr(tmpname, '/'); if (cp == NULL) { (void) strlcpy(dsname, tmpname, MAXPATHLEN); } else { (void) snprintf(dsname, MAXPATHLEN, "%s/%s", poolname, ++cp); } fnvlist_add_string(config, ZPOOL_CONFIG_BOOTFS, dsname); kmem_free(dsname, MAXPATHLEN); } kmem_free(tmpname, MAXPATHLEN); } /* * Add the list of hot spares and level 2 cache devices. */ spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER); spa_add_spares(spa, config); spa_add_l2cache(spa, config); spa_config_exit(spa, SCL_CONFIG, FTAG); } spa_unload(spa); spa_deactivate(spa); spa_remove(spa); mutex_exit(&spa_namespace_lock); return (config); } /* * Pool export/destroy * * The act of destroying or exporting a pool is very simple. We make sure there * is no more pending I/O and any references to the pool are gone. Then, we * update the pool state and sync all the labels to disk, removing the * configuration from the cache afterwards. If the 'hardforce' flag is set, then * we don't sync the labels or remove the configuration cache. */ static int spa_export_common(const char *pool, int new_state, nvlist_t **oldconfig, boolean_t force, boolean_t hardforce) { int error; spa_t *spa; if (oldconfig) *oldconfig = NULL; if (!(spa_mode_global & SPA_MODE_WRITE)) return (SET_ERROR(EROFS)); mutex_enter(&spa_namespace_lock); if ((spa = spa_lookup(pool)) == NULL) { mutex_exit(&spa_namespace_lock); return (SET_ERROR(ENOENT)); } if (spa->spa_is_exporting) { /* the pool is being exported by another thread */ mutex_exit(&spa_namespace_lock); return (SET_ERROR(ZFS_ERR_EXPORT_IN_PROGRESS)); } spa->spa_is_exporting = B_TRUE; /* * Put a hold on the pool, drop the namespace lock, stop async tasks, * reacquire the namespace lock, and see if we can export. */ spa_open_ref(spa, FTAG); mutex_exit(&spa_namespace_lock); spa_async_suspend(spa); if (spa->spa_zvol_taskq) { zvol_remove_minors(spa, spa_name(spa), B_TRUE); taskq_wait(spa->spa_zvol_taskq); } mutex_enter(&spa_namespace_lock); spa_close(spa, FTAG); if (spa->spa_state == POOL_STATE_UNINITIALIZED) goto export_spa; /* * The pool will be in core if it's openable, in which case we can * modify its state. Objsets may be open only because they're dirty, * so we have to force it to sync before checking spa_refcnt. */ if (spa->spa_sync_on) { txg_wait_synced(spa->spa_dsl_pool, 0); spa_evicting_os_wait(spa); } /* * A pool cannot be exported or destroyed if there are active * references. If we are resetting a pool, allow references by * fault injection handlers. */ if (!spa_refcount_zero(spa) || (spa->spa_inject_ref != 0)) { error = SET_ERROR(EBUSY); goto fail; } if (spa->spa_sync_on) { vdev_t *rvd = spa->spa_root_vdev; /* * A pool cannot be exported if it has an active shared spare. * This is to prevent other pools stealing the active spare * from an exported pool. At user's own will, such pool can * be forcedly exported. */ if (!force && new_state == POOL_STATE_EXPORTED && spa_has_active_shared_spare(spa)) { error = SET_ERROR(EXDEV); goto fail; } /* * We're about to export or destroy this pool. Make sure * we stop all initialization and trim activity here before * we set the spa_final_txg. This will ensure that all * dirty data resulting from the initialization is * committed to disk before we unload the pool. */ vdev_initialize_stop_all(rvd, VDEV_INITIALIZE_ACTIVE); vdev_trim_stop_all(rvd, VDEV_TRIM_ACTIVE); vdev_autotrim_stop_all(spa); vdev_rebuild_stop_all(spa); /* * We want this to be reflected on every label, * so mark them all dirty. spa_unload() will do the * final sync that pushes these changes out. */ if (new_state != POOL_STATE_UNINITIALIZED && !hardforce) { spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); spa->spa_state = new_state; vdev_config_dirty(rvd); spa_config_exit(spa, SCL_ALL, FTAG); } /* * If the log space map feature is enabled and the pool is * getting exported (but not destroyed), we want to spend some * time flushing as many metaslabs as we can in an attempt to * destroy log space maps and save import time. This has to be * done before we set the spa_final_txg, otherwise * spa_sync() -> spa_flush_metaslabs() may dirty the final TXGs. * spa_should_flush_logs_on_unload() should be called after * spa_state has been set to the new_state. */ if (spa_should_flush_logs_on_unload(spa)) spa_unload_log_sm_flush_all(spa); if (new_state != POOL_STATE_UNINITIALIZED && !hardforce) { spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); spa->spa_final_txg = spa_last_synced_txg(spa) + TXG_DEFER_SIZE + 1; spa_config_exit(spa, SCL_ALL, FTAG); } } export_spa: spa_export_os(spa); if (new_state == POOL_STATE_DESTROYED) spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_DESTROY); else if (new_state == POOL_STATE_EXPORTED) spa_event_notify(spa, NULL, NULL, ESC_ZFS_POOL_EXPORT); if (spa->spa_state != POOL_STATE_UNINITIALIZED) { spa_unload(spa); spa_deactivate(spa); } if (oldconfig && spa->spa_config) *oldconfig = fnvlist_dup(spa->spa_config); if (new_state != POOL_STATE_UNINITIALIZED) { if (!hardforce) spa_write_cachefile(spa, B_TRUE, B_TRUE, B_FALSE); spa_remove(spa); } else { /* * If spa_remove() is not called for this spa_t and * there is any possibility that it can be reused, * we make sure to reset the exporting flag. */ spa->spa_is_exporting = B_FALSE; } mutex_exit(&spa_namespace_lock); return (0); fail: spa->spa_is_exporting = B_FALSE; spa_async_resume(spa); mutex_exit(&spa_namespace_lock); return (error); } /* * Destroy a storage pool. */ int spa_destroy(const char *pool) { return (spa_export_common(pool, POOL_STATE_DESTROYED, NULL, B_FALSE, B_FALSE)); } /* * Export a storage pool. */ int spa_export(const char *pool, nvlist_t **oldconfig, boolean_t force, boolean_t hardforce) { return (spa_export_common(pool, POOL_STATE_EXPORTED, oldconfig, force, hardforce)); } /* * Similar to spa_export(), this unloads the spa_t without actually removing it * from the namespace in any way. */ int spa_reset(const char *pool) { return (spa_export_common(pool, POOL_STATE_UNINITIALIZED, NULL, B_FALSE, B_FALSE)); } /* * ========================================================================== * Device manipulation * ========================================================================== */ /* * This is called as a synctask to increment the draid feature flag */ static void spa_draid_feature_incr(void *arg, dmu_tx_t *tx) { spa_t *spa = dmu_tx_pool(tx)->dp_spa; int draid = (int)(uintptr_t)arg; for (int c = 0; c < draid; c++) spa_feature_incr(spa, SPA_FEATURE_DRAID, tx); } /* * Add a device to a storage pool. */ int spa_vdev_add(spa_t *spa, nvlist_t *nvroot) { uint64_t txg, ndraid = 0; int error; vdev_t *rvd = spa->spa_root_vdev; vdev_t *vd, *tvd; nvlist_t **spares, **l2cache; uint_t nspares, nl2cache; ASSERT(spa_writeable(spa)); txg = spa_vdev_enter(spa); if ((error = spa_config_parse(spa, &vd, nvroot, NULL, 0, VDEV_ALLOC_ADD)) != 0) return (spa_vdev_exit(spa, NULL, txg, error)); spa->spa_pending_vdev = vd; /* spa_vdev_exit() will clear this */ if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, &spares, &nspares) != 0) nspares = 0; if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE, &l2cache, &nl2cache) != 0) nl2cache = 0; if (vd->vdev_children == 0 && nspares == 0 && nl2cache == 0) return (spa_vdev_exit(spa, vd, txg, EINVAL)); if (vd->vdev_children != 0 && (error = vdev_create(vd, txg, B_FALSE)) != 0) { return (spa_vdev_exit(spa, vd, txg, error)); } /* * The virtual dRAID spares must be added after vdev tree is created * and the vdev guids are generated. The guid of their associated * dRAID is stored in the config and used when opening the spare. */ if ((error = vdev_draid_spare_create(nvroot, vd, &ndraid, rvd->vdev_children)) == 0) { if (ndraid > 0 && nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, &spares, &nspares) != 0) nspares = 0; } else { return (spa_vdev_exit(spa, vd, txg, error)); } /* * We must validate the spares and l2cache devices after checking the * children. Otherwise, vdev_inuse() will blindly overwrite the spare. */ if ((error = spa_validate_aux(spa, nvroot, txg, VDEV_ALLOC_ADD)) != 0) return (spa_vdev_exit(spa, vd, txg, error)); /* * If we are in the middle of a device removal, we can only add * devices which match the existing devices in the pool. * If we are in the middle of a removal, or have some indirect * vdevs, we can not add raidz or dRAID top levels. */ if (spa->spa_vdev_removal != NULL || spa->spa_removing_phys.sr_prev_indirect_vdev != -1) { for (int c = 0; c < vd->vdev_children; c++) { tvd = vd->vdev_child[c]; if (spa->spa_vdev_removal != NULL && tvd->vdev_ashift != spa->spa_max_ashift) { return (spa_vdev_exit(spa, vd, txg, EINVAL)); } /* Fail if top level vdev is raidz or a dRAID */ if (vdev_get_nparity(tvd) != 0) return (spa_vdev_exit(spa, vd, txg, EINVAL)); /* * Need the top level mirror to be * a mirror of leaf vdevs only */ if (tvd->vdev_ops == &vdev_mirror_ops) { for (uint64_t cid = 0; cid < tvd->vdev_children; cid++) { vdev_t *cvd = tvd->vdev_child[cid]; if (!cvd->vdev_ops->vdev_op_leaf) { return (spa_vdev_exit(spa, vd, txg, EINVAL)); } } } } } for (int c = 0; c < vd->vdev_children; c++) { tvd = vd->vdev_child[c]; vdev_remove_child(vd, tvd); tvd->vdev_id = rvd->vdev_children; vdev_add_child(rvd, tvd); vdev_config_dirty(tvd); } if (nspares != 0) { spa_set_aux_vdevs(&spa->spa_spares, spares, nspares, ZPOOL_CONFIG_SPARES); spa_load_spares(spa); spa->spa_spares.sav_sync = B_TRUE; } if (nl2cache != 0) { spa_set_aux_vdevs(&spa->spa_l2cache, l2cache, nl2cache, ZPOOL_CONFIG_L2CACHE); spa_load_l2cache(spa); spa->spa_l2cache.sav_sync = B_TRUE; } /* * We can't increment a feature while holding spa_vdev so we * have to do it in a synctask. */ if (ndraid != 0) { dmu_tx_t *tx; tx = dmu_tx_create_assigned(spa->spa_dsl_pool, txg); dsl_sync_task_nowait(spa->spa_dsl_pool, spa_draid_feature_incr, (void *)(uintptr_t)ndraid, tx); dmu_tx_commit(tx); } /* * We have to be careful when adding new vdevs to an existing pool. * If other threads start allocating from these vdevs before we * sync the config cache, and we lose power, then upon reboot we may * fail to open the pool because there are DVAs that the config cache * can't translate. Therefore, we first add the vdevs without * initializing metaslabs; sync the config cache (via spa_vdev_exit()); * and then let spa_config_update() initialize the new metaslabs. * * spa_load() checks for added-but-not-initialized vdevs, so that * if we lose power at any point in this sequence, the remaining * steps will be completed the next time we load the pool. */ (void) spa_vdev_exit(spa, vd, txg, 0); mutex_enter(&spa_namespace_lock); spa_config_update(spa, SPA_CONFIG_UPDATE_POOL); spa_event_notify(spa, NULL, NULL, ESC_ZFS_VDEV_ADD); mutex_exit(&spa_namespace_lock); return (0); } /* * Attach a device to a mirror. The arguments are the path to any device * in the mirror, and the nvroot for the new device. If the path specifies * a device that is not mirrored, we automatically insert the mirror vdev. * * If 'replacing' is specified, the new device is intended to replace the * existing device; in this case the two devices are made into their own * mirror using the 'replacing' vdev, which is functionally identical to * the mirror vdev (it actually reuses all the same ops) but has a few * extra rules: you can't attach to it after it's been created, and upon * completion of resilvering, the first disk (the one being replaced) * is automatically detached. * * If 'rebuild' is specified, then sequential reconstruction (a.ka. rebuild) * should be performed instead of traditional healing reconstruction. From * an administrators perspective these are both resilver operations. */ int spa_vdev_attach(spa_t *spa, uint64_t guid, nvlist_t *nvroot, int replacing, int rebuild) { uint64_t txg, dtl_max_txg; vdev_t *rvd = spa->spa_root_vdev; vdev_t *oldvd, *newvd, *newrootvd, *pvd, *tvd; vdev_ops_t *pvops; char *oldvdpath, *newvdpath; int newvd_isspare; int error; ASSERT(spa_writeable(spa)); txg = spa_vdev_enter(spa); oldvd = spa_lookup_by_guid(spa, guid, B_FALSE); ASSERT(MUTEX_HELD(&spa_namespace_lock)); if (spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT)) { error = (spa_has_checkpoint(spa)) ? ZFS_ERR_CHECKPOINT_EXISTS : ZFS_ERR_DISCARDING_CHECKPOINT; return (spa_vdev_exit(spa, NULL, txg, error)); } if (rebuild) { if (!spa_feature_is_enabled(spa, SPA_FEATURE_DEVICE_REBUILD)) return (spa_vdev_exit(spa, NULL, txg, ENOTSUP)); if (dsl_scan_resilvering(spa_get_dsl(spa))) return (spa_vdev_exit(spa, NULL, txg, ZFS_ERR_RESILVER_IN_PROGRESS)); } else { if (vdev_rebuild_active(rvd)) return (spa_vdev_exit(spa, NULL, txg, ZFS_ERR_REBUILD_IN_PROGRESS)); } if (spa->spa_vdev_removal != NULL) return (spa_vdev_exit(spa, NULL, txg, EBUSY)); if (oldvd == NULL) return (spa_vdev_exit(spa, NULL, txg, ENODEV)); if (!oldvd->vdev_ops->vdev_op_leaf) return (spa_vdev_exit(spa, NULL, txg, ENOTSUP)); pvd = oldvd->vdev_parent; if (spa_config_parse(spa, &newrootvd, nvroot, NULL, 0, VDEV_ALLOC_ATTACH) != 0) return (spa_vdev_exit(spa, NULL, txg, EINVAL)); if (newrootvd->vdev_children != 1) return (spa_vdev_exit(spa, newrootvd, txg, EINVAL)); newvd = newrootvd->vdev_child[0]; if (!newvd->vdev_ops->vdev_op_leaf) return (spa_vdev_exit(spa, newrootvd, txg, EINVAL)); if ((error = vdev_create(newrootvd, txg, replacing)) != 0) return (spa_vdev_exit(spa, newrootvd, txg, error)); /* * log, dedup and special vdevs should not be replaced by spares. */ if ((oldvd->vdev_top->vdev_alloc_bias != VDEV_BIAS_NONE || oldvd->vdev_top->vdev_islog) && newvd->vdev_isspare) { return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP)); } /* * A dRAID spare can only replace a child of its parent dRAID vdev. */ if (newvd->vdev_ops == &vdev_draid_spare_ops && oldvd->vdev_top != vdev_draid_spare_get_parent(newvd)) { return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP)); } if (rebuild) { /* * For rebuilds, the top vdev must support reconstruction * using only space maps. This means the only allowable * vdevs types are the root vdev, a mirror, or dRAID. */ tvd = pvd; if (pvd->vdev_top != NULL) tvd = pvd->vdev_top; if (tvd->vdev_ops != &vdev_mirror_ops && tvd->vdev_ops != &vdev_root_ops && tvd->vdev_ops != &vdev_draid_ops) { return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP)); } } if (!replacing) { /* * For attach, the only allowable parent is a mirror or the root * vdev. */ if (pvd->vdev_ops != &vdev_mirror_ops && pvd->vdev_ops != &vdev_root_ops) return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP)); pvops = &vdev_mirror_ops; } else { /* * Active hot spares can only be replaced by inactive hot * spares. */ if (pvd->vdev_ops == &vdev_spare_ops && oldvd->vdev_isspare && !spa_has_spare(spa, newvd->vdev_guid)) return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP)); /* * If the source is a hot spare, and the parent isn't already a * spare, then we want to create a new hot spare. Otherwise, we * want to create a replacing vdev. The user is not allowed to * attach to a spared vdev child unless the 'isspare' state is * the same (spare replaces spare, non-spare replaces * non-spare). */ if (pvd->vdev_ops == &vdev_replacing_ops && spa_version(spa) < SPA_VERSION_MULTI_REPLACE) { return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP)); } else if (pvd->vdev_ops == &vdev_spare_ops && newvd->vdev_isspare != oldvd->vdev_isspare) { return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP)); } if (newvd->vdev_isspare) pvops = &vdev_spare_ops; else pvops = &vdev_replacing_ops; } /* * Make sure the new device is big enough. */ if (newvd->vdev_asize < vdev_get_min_asize(oldvd)) return (spa_vdev_exit(spa, newrootvd, txg, EOVERFLOW)); /* * The new device cannot have a higher alignment requirement * than the top-level vdev. */ if (newvd->vdev_ashift > oldvd->vdev_top->vdev_ashift) return (spa_vdev_exit(spa, newrootvd, txg, ENOTSUP)); /* * If this is an in-place replacement, update oldvd's path and devid * to make it distinguishable from newvd, and unopenable from now on. */ if (strcmp(oldvd->vdev_path, newvd->vdev_path) == 0) { spa_strfree(oldvd->vdev_path); oldvd->vdev_path = kmem_alloc(strlen(newvd->vdev_path) + 5, KM_SLEEP); (void) snprintf(oldvd->vdev_path, strlen(newvd->vdev_path) + 5, "%s/%s", newvd->vdev_path, "old"); if (oldvd->vdev_devid != NULL) { spa_strfree(oldvd->vdev_devid); oldvd->vdev_devid = NULL; } } /* * If the parent is not a mirror, or if we're replacing, insert the new * mirror/replacing/spare vdev above oldvd. */ if (pvd->vdev_ops != pvops) pvd = vdev_add_parent(oldvd, pvops); ASSERT(pvd->vdev_top->vdev_parent == rvd); ASSERT(pvd->vdev_ops == pvops); ASSERT(oldvd->vdev_parent == pvd); /* * Extract the new device from its root and add it to pvd. */ vdev_remove_child(newrootvd, newvd); newvd->vdev_id = pvd->vdev_children; newvd->vdev_crtxg = oldvd->vdev_crtxg; vdev_add_child(pvd, newvd); /* * Reevaluate the parent vdev state. */ vdev_propagate_state(pvd); tvd = newvd->vdev_top; ASSERT(pvd->vdev_top == tvd); ASSERT(tvd->vdev_parent == rvd); vdev_config_dirty(tvd); /* * Set newvd's DTL to [TXG_INITIAL, dtl_max_txg) so that we account * for any dmu_sync-ed blocks. It will propagate upward when * spa_vdev_exit() calls vdev_dtl_reassess(). */ dtl_max_txg = txg + TXG_CONCURRENT_STATES; vdev_dtl_dirty(newvd, DTL_MISSING, TXG_INITIAL, dtl_max_txg - TXG_INITIAL); if (newvd->vdev_isspare) { spa_spare_activate(newvd); spa_event_notify(spa, newvd, NULL, ESC_ZFS_VDEV_SPARE); } oldvdpath = spa_strdup(oldvd->vdev_path); newvdpath = spa_strdup(newvd->vdev_path); newvd_isspare = newvd->vdev_isspare; /* * Mark newvd's DTL dirty in this txg. */ vdev_dirty(tvd, VDD_DTL, newvd, txg); /* * Schedule the resilver or rebuild to restart in the future. We do * this to ensure that dmu_sync-ed blocks have been stitched into the * respective datasets. */ if (rebuild) { newvd->vdev_rebuild_txg = txg; vdev_rebuild(tvd); } else { newvd->vdev_resilver_txg = txg; if (dsl_scan_resilvering(spa_get_dsl(spa)) && spa_feature_is_enabled(spa, SPA_FEATURE_RESILVER_DEFER)) { vdev_defer_resilver(newvd); } else { dsl_scan_restart_resilver(spa->spa_dsl_pool, dtl_max_txg); } } if (spa->spa_bootfs) spa_event_notify(spa, newvd, NULL, ESC_ZFS_BOOTFS_VDEV_ATTACH); spa_event_notify(spa, newvd, NULL, ESC_ZFS_VDEV_ATTACH); /* * Commit the config */ (void) spa_vdev_exit(spa, newrootvd, dtl_max_txg, 0); spa_history_log_internal(spa, "vdev attach", NULL, "%s vdev=%s %s vdev=%s", replacing && newvd_isspare ? "spare in" : replacing ? "replace" : "attach", newvdpath, replacing ? "for" : "to", oldvdpath); spa_strfree(oldvdpath); spa_strfree(newvdpath); return (0); } /* * Detach a device from a mirror or replacing vdev. * * If 'replace_done' is specified, only detach if the parent * is a replacing or a spare vdev. */ int spa_vdev_detach(spa_t *spa, uint64_t guid, uint64_t pguid, int replace_done) { uint64_t txg; int error; vdev_t *rvd __maybe_unused = spa->spa_root_vdev; vdev_t *vd, *pvd, *cvd, *tvd; boolean_t unspare = B_FALSE; uint64_t unspare_guid = 0; char *vdpath; ASSERT(spa_writeable(spa)); txg = spa_vdev_detach_enter(spa, guid); vd = spa_lookup_by_guid(spa, guid, B_FALSE); /* * Besides being called directly from the userland through the * ioctl interface, spa_vdev_detach() can be potentially called * at the end of spa_vdev_resilver_done(). * * In the regular case, when we have a checkpoint this shouldn't * happen as we never empty the DTLs of a vdev during the scrub * [see comment in dsl_scan_done()]. Thus spa_vdev_resilvering_done() * should never get here when we have a checkpoint. * * That said, even in a case when we checkpoint the pool exactly * as spa_vdev_resilver_done() calls this function everything * should be fine as the resilver will return right away. */ ASSERT(MUTEX_HELD(&spa_namespace_lock)); if (spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT)) { error = (spa_has_checkpoint(spa)) ? ZFS_ERR_CHECKPOINT_EXISTS : ZFS_ERR_DISCARDING_CHECKPOINT; return (spa_vdev_exit(spa, NULL, txg, error)); } if (vd == NULL) return (spa_vdev_exit(spa, NULL, txg, ENODEV)); if (!vd->vdev_ops->vdev_op_leaf) return (spa_vdev_exit(spa, NULL, txg, ENOTSUP)); pvd = vd->vdev_parent; /* * If the parent/child relationship is not as expected, don't do it. * Consider M(A,R(B,C)) -- that is, a mirror of A with a replacing * vdev that's replacing B with C. The user's intent in replacing * is to go from M(A,B) to M(A,C). If the user decides to cancel * the replace by detaching C, the expected behavior is to end up * M(A,B). But suppose that right after deciding to detach C, * the replacement of B completes. We would have M(A,C), and then * ask to detach C, which would leave us with just A -- not what * the user wanted. To prevent this, we make sure that the * parent/child relationship hasn't changed -- in this example, * that C's parent is still the replacing vdev R. */ if (pvd->vdev_guid != pguid && pguid != 0) return (spa_vdev_exit(spa, NULL, txg, EBUSY)); /* * Only 'replacing' or 'spare' vdevs can be replaced. */ if (replace_done && pvd->vdev_ops != &vdev_replacing_ops && pvd->vdev_ops != &vdev_spare_ops) return (spa_vdev_exit(spa, NULL, txg, ENOTSUP)); ASSERT(pvd->vdev_ops != &vdev_spare_ops || spa_version(spa) >= SPA_VERSION_SPARES); /* * Only mirror, replacing, and spare vdevs support detach. */ if (pvd->vdev_ops != &vdev_replacing_ops && pvd->vdev_ops != &vdev_mirror_ops && pvd->vdev_ops != &vdev_spare_ops) return (spa_vdev_exit(spa, NULL, txg, ENOTSUP)); /* * If this device has the only valid copy of some data, * we cannot safely detach it. */ if (vdev_dtl_required(vd)) return (spa_vdev_exit(spa, NULL, txg, EBUSY)); ASSERT(pvd->vdev_children >= 2); /* * If we are detaching the second disk from a replacing vdev, then * check to see if we changed the original vdev's path to have "/old" * at the end in spa_vdev_attach(). If so, undo that change now. */ if (pvd->vdev_ops == &vdev_replacing_ops && vd->vdev_id > 0 && vd->vdev_path != NULL) { size_t len = strlen(vd->vdev_path); for (int c = 0; c < pvd->vdev_children; c++) { cvd = pvd->vdev_child[c]; if (cvd == vd || cvd->vdev_path == NULL) continue; if (strncmp(cvd->vdev_path, vd->vdev_path, len) == 0 && strcmp(cvd->vdev_path + len, "/old") == 0) { spa_strfree(cvd->vdev_path); cvd->vdev_path = spa_strdup(vd->vdev_path); break; } } } /* * If we are detaching the original disk from a normal spare, then it * implies that the spare should become a real disk, and be removed * from the active spare list for the pool. dRAID spares on the * other hand are coupled to the pool and thus should never be removed * from the spares list. */ if (pvd->vdev_ops == &vdev_spare_ops && vd->vdev_id == 0) { vdev_t *last_cvd = pvd->vdev_child[pvd->vdev_children - 1]; if (last_cvd->vdev_isspare && last_cvd->vdev_ops != &vdev_draid_spare_ops) { unspare = B_TRUE; } } /* * Erase the disk labels so the disk can be used for other things. * This must be done after all other error cases are handled, * but before we disembowel vd (so we can still do I/O to it). * But if we can't do it, don't treat the error as fatal -- * it may be that the unwritability of the disk is the reason * it's being detached! */ (void) vdev_label_init(vd, 0, VDEV_LABEL_REMOVE); /* * Remove vd from its parent and compact the parent's children. */ vdev_remove_child(pvd, vd); vdev_compact_children(pvd); /* * Remember one of the remaining children so we can get tvd below. */ cvd = pvd->vdev_child[pvd->vdev_children - 1]; /* * If we need to remove the remaining child from the list of hot spares, * do it now, marking the vdev as no longer a spare in the process. * We must do this before vdev_remove_parent(), because that can * change the GUID if it creates a new toplevel GUID. For a similar * reason, we must remove the spare now, in the same txg as the detach; * otherwise someone could attach a new sibling, change the GUID, and * the subsequent attempt to spa_vdev_remove(unspare_guid) would fail. */ if (unspare) { ASSERT(cvd->vdev_isspare); spa_spare_remove(cvd); unspare_guid = cvd->vdev_guid; (void) spa_vdev_remove(spa, unspare_guid, B_TRUE); cvd->vdev_unspare = B_TRUE; } /* * If the parent mirror/replacing vdev only has one child, * the parent is no longer needed. Remove it from the tree. */ if (pvd->vdev_children == 1) { if (pvd->vdev_ops == &vdev_spare_ops) cvd->vdev_unspare = B_FALSE; vdev_remove_parent(cvd); } /* * We don't set tvd until now because the parent we just removed * may have been the previous top-level vdev. */ tvd = cvd->vdev_top; ASSERT(tvd->vdev_parent == rvd); /* * Reevaluate the parent vdev state. */ vdev_propagate_state(cvd); /* * If the 'autoexpand' property is set on the pool then automatically * try to expand the size of the pool. For example if the device we * just detached was smaller than the others, it may be possible to * add metaslabs (i.e. grow the pool). We need to reopen the vdev * first so that we can obtain the updated sizes of the leaf vdevs. */ if (spa->spa_autoexpand) { vdev_reopen(tvd); vdev_expand(tvd, txg); } vdev_config_dirty(tvd); /* * Mark vd's DTL as dirty in this txg. vdev_dtl_sync() will see that * vd->vdev_detached is set and free vd's DTL object in syncing context. * But first make sure we're not on any *other* txg's DTL list, to * prevent vd from being accessed after it's freed. */ vdpath = spa_strdup(vd->vdev_path ? vd->vdev_path : "none"); for (int t = 0; t < TXG_SIZE; t++) (void) txg_list_remove_this(&tvd->vdev_dtl_list, vd, t); vd->vdev_detached = B_TRUE; vdev_dirty(tvd, VDD_DTL, vd, txg); spa_event_notify(spa, vd, NULL, ESC_ZFS_VDEV_REMOVE); spa_notify_waiters(spa); /* hang on to the spa before we release the lock */ spa_open_ref(spa, FTAG); error = spa_vdev_exit(spa, vd, txg, 0); spa_history_log_internal(spa, "detach", NULL, "vdev=%s", vdpath); spa_strfree(vdpath); /* * If this was the removal of the original device in a hot spare vdev, * then we want to go through and remove the device from the hot spare * list of every other pool. */ if (unspare) { spa_t *altspa = NULL; mutex_enter(&spa_namespace_lock); while ((altspa = spa_next(altspa)) != NULL) { if (altspa->spa_state != POOL_STATE_ACTIVE || altspa == spa) continue; spa_open_ref(altspa, FTAG); mutex_exit(&spa_namespace_lock); (void) spa_vdev_remove(altspa, unspare_guid, B_TRUE); mutex_enter(&spa_namespace_lock); spa_close(altspa, FTAG); } mutex_exit(&spa_namespace_lock); /* search the rest of the vdevs for spares to remove */ spa_vdev_resilver_done(spa); } /* all done with the spa; OK to release */ mutex_enter(&spa_namespace_lock); spa_close(spa, FTAG); mutex_exit(&spa_namespace_lock); return (error); } static int spa_vdev_initialize_impl(spa_t *spa, uint64_t guid, uint64_t cmd_type, list_t *vd_list) { ASSERT(MUTEX_HELD(&spa_namespace_lock)); spa_config_enter(spa, SCL_CONFIG | SCL_STATE, FTAG, RW_READER); /* Look up vdev and ensure it's a leaf. */ vdev_t *vd = spa_lookup_by_guid(spa, guid, B_FALSE); if (vd == NULL || vd->vdev_detached) { spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG); return (SET_ERROR(ENODEV)); } else if (!vd->vdev_ops->vdev_op_leaf || !vdev_is_concrete(vd)) { spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG); return (SET_ERROR(EINVAL)); } else if (!vdev_writeable(vd)) { spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG); return (SET_ERROR(EROFS)); } mutex_enter(&vd->vdev_initialize_lock); spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG); /* * When we activate an initialize action we check to see * if the vdev_initialize_thread is NULL. We do this instead * of using the vdev_initialize_state since there might be * a previous initialization process which has completed but * the thread is not exited. */ if (cmd_type == POOL_INITIALIZE_START && (vd->vdev_initialize_thread != NULL || vd->vdev_top->vdev_removing)) { mutex_exit(&vd->vdev_initialize_lock); return (SET_ERROR(EBUSY)); } else if (cmd_type == POOL_INITIALIZE_CANCEL && (vd->vdev_initialize_state != VDEV_INITIALIZE_ACTIVE && vd->vdev_initialize_state != VDEV_INITIALIZE_SUSPENDED)) { mutex_exit(&vd->vdev_initialize_lock); return (SET_ERROR(ESRCH)); } else if (cmd_type == POOL_INITIALIZE_SUSPEND && vd->vdev_initialize_state != VDEV_INITIALIZE_ACTIVE) { mutex_exit(&vd->vdev_initialize_lock); return (SET_ERROR(ESRCH)); + } else if (cmd_type == POOL_INITIALIZE_UNINIT && + vd->vdev_initialize_thread != NULL) { + mutex_exit(&vd->vdev_initialize_lock); + return (SET_ERROR(EBUSY)); } switch (cmd_type) { case POOL_INITIALIZE_START: vdev_initialize(vd); break; case POOL_INITIALIZE_CANCEL: vdev_initialize_stop(vd, VDEV_INITIALIZE_CANCELED, vd_list); break; case POOL_INITIALIZE_SUSPEND: vdev_initialize_stop(vd, VDEV_INITIALIZE_SUSPENDED, vd_list); break; + case POOL_INITIALIZE_UNINIT: + vdev_uninitialize(vd); + break; default: panic("invalid cmd_type %llu", (unsigned long long)cmd_type); } mutex_exit(&vd->vdev_initialize_lock); return (0); } int spa_vdev_initialize(spa_t *spa, nvlist_t *nv, uint64_t cmd_type, nvlist_t *vdev_errlist) { int total_errors = 0; list_t vd_list; list_create(&vd_list, sizeof (vdev_t), offsetof(vdev_t, vdev_initialize_node)); /* * We hold the namespace lock through the whole function * to prevent any changes to the pool while we're starting or * stopping initialization. The config and state locks are held so that * we can properly assess the vdev state before we commit to * the initializing operation. */ mutex_enter(&spa_namespace_lock); for (nvpair_t *pair = nvlist_next_nvpair(nv, NULL); pair != NULL; pair = nvlist_next_nvpair(nv, pair)) { uint64_t vdev_guid = fnvpair_value_uint64(pair); int error = spa_vdev_initialize_impl(spa, vdev_guid, cmd_type, &vd_list); if (error != 0) { char guid_as_str[MAXNAMELEN]; (void) snprintf(guid_as_str, sizeof (guid_as_str), "%llu", (unsigned long long)vdev_guid); fnvlist_add_int64(vdev_errlist, guid_as_str, error); total_errors++; } } /* Wait for all initialize threads to stop. */ vdev_initialize_stop_wait(spa, &vd_list); /* Sync out the initializing state */ txg_wait_synced(spa->spa_dsl_pool, 0); mutex_exit(&spa_namespace_lock); list_destroy(&vd_list); return (total_errors); } static int spa_vdev_trim_impl(spa_t *spa, uint64_t guid, uint64_t cmd_type, uint64_t rate, boolean_t partial, boolean_t secure, list_t *vd_list) { ASSERT(MUTEX_HELD(&spa_namespace_lock)); spa_config_enter(spa, SCL_CONFIG | SCL_STATE, FTAG, RW_READER); /* Look up vdev and ensure it's a leaf. */ vdev_t *vd = spa_lookup_by_guid(spa, guid, B_FALSE); if (vd == NULL || vd->vdev_detached) { spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG); return (SET_ERROR(ENODEV)); } else if (!vd->vdev_ops->vdev_op_leaf || !vdev_is_concrete(vd)) { spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG); return (SET_ERROR(EINVAL)); } else if (!vdev_writeable(vd)) { spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG); return (SET_ERROR(EROFS)); } else if (!vd->vdev_has_trim) { spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG); return (SET_ERROR(EOPNOTSUPP)); } else if (secure && !vd->vdev_has_securetrim) { spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG); return (SET_ERROR(EOPNOTSUPP)); } mutex_enter(&vd->vdev_trim_lock); spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG); /* * When we activate a TRIM action we check to see if the * vdev_trim_thread is NULL. We do this instead of using the * vdev_trim_state since there might be a previous TRIM process * which has completed but the thread is not exited. */ if (cmd_type == POOL_TRIM_START && (vd->vdev_trim_thread != NULL || vd->vdev_top->vdev_removing)) { mutex_exit(&vd->vdev_trim_lock); return (SET_ERROR(EBUSY)); } else if (cmd_type == POOL_TRIM_CANCEL && (vd->vdev_trim_state != VDEV_TRIM_ACTIVE && vd->vdev_trim_state != VDEV_TRIM_SUSPENDED)) { mutex_exit(&vd->vdev_trim_lock); return (SET_ERROR(ESRCH)); } else if (cmd_type == POOL_TRIM_SUSPEND && vd->vdev_trim_state != VDEV_TRIM_ACTIVE) { mutex_exit(&vd->vdev_trim_lock); return (SET_ERROR(ESRCH)); } switch (cmd_type) { case POOL_TRIM_START: vdev_trim(vd, rate, partial, secure); break; case POOL_TRIM_CANCEL: vdev_trim_stop(vd, VDEV_TRIM_CANCELED, vd_list); break; case POOL_TRIM_SUSPEND: vdev_trim_stop(vd, VDEV_TRIM_SUSPENDED, vd_list); break; default: panic("invalid cmd_type %llu", (unsigned long long)cmd_type); } mutex_exit(&vd->vdev_trim_lock); return (0); } /* * Initiates a manual TRIM for the requested vdevs. This kicks off individual * TRIM threads for each child vdev. These threads pass over all of the free * space in the vdev's metaslabs and issues TRIM commands for that space. */ int spa_vdev_trim(spa_t *spa, nvlist_t *nv, uint64_t cmd_type, uint64_t rate, boolean_t partial, boolean_t secure, nvlist_t *vdev_errlist) { int total_errors = 0; list_t vd_list; list_create(&vd_list, sizeof (vdev_t), offsetof(vdev_t, vdev_trim_node)); /* * We hold the namespace lock through the whole function * to prevent any changes to the pool while we're starting or * stopping TRIM. The config and state locks are held so that * we can properly assess the vdev state before we commit to * the TRIM operation. */ mutex_enter(&spa_namespace_lock); for (nvpair_t *pair = nvlist_next_nvpair(nv, NULL); pair != NULL; pair = nvlist_next_nvpair(nv, pair)) { uint64_t vdev_guid = fnvpair_value_uint64(pair); int error = spa_vdev_trim_impl(spa, vdev_guid, cmd_type, rate, partial, secure, &vd_list); if (error != 0) { char guid_as_str[MAXNAMELEN]; (void) snprintf(guid_as_str, sizeof (guid_as_str), "%llu", (unsigned long long)vdev_guid); fnvlist_add_int64(vdev_errlist, guid_as_str, error); total_errors++; } } /* Wait for all TRIM threads to stop. */ vdev_trim_stop_wait(spa, &vd_list); /* Sync out the TRIM state */ txg_wait_synced(spa->spa_dsl_pool, 0); mutex_exit(&spa_namespace_lock); list_destroy(&vd_list); return (total_errors); } /* * Split a set of devices from their mirrors, and create a new pool from them. */ int spa_vdev_split_mirror(spa_t *spa, const char *newname, nvlist_t *config, nvlist_t *props, boolean_t exp) { int error = 0; uint64_t txg, *glist; spa_t *newspa; uint_t c, children, lastlog; nvlist_t **child, *nvl, *tmp; dmu_tx_t *tx; const char *altroot = NULL; vdev_t *rvd, **vml = NULL; /* vdev modify list */ boolean_t activate_slog; ASSERT(spa_writeable(spa)); txg = spa_vdev_enter(spa); ASSERT(MUTEX_HELD(&spa_namespace_lock)); if (spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT)) { error = (spa_has_checkpoint(spa)) ? ZFS_ERR_CHECKPOINT_EXISTS : ZFS_ERR_DISCARDING_CHECKPOINT; return (spa_vdev_exit(spa, NULL, txg, error)); } /* clear the log and flush everything up to now */ activate_slog = spa_passivate_log(spa); (void) spa_vdev_config_exit(spa, NULL, txg, 0, FTAG); error = spa_reset_logs(spa); txg = spa_vdev_config_enter(spa); if (activate_slog) spa_activate_log(spa); if (error != 0) return (spa_vdev_exit(spa, NULL, txg, error)); /* check new spa name before going any further */ if (spa_lookup(newname) != NULL) return (spa_vdev_exit(spa, NULL, txg, EEXIST)); /* * scan through all the children to ensure they're all mirrors */ if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvl) != 0 || nvlist_lookup_nvlist_array(nvl, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) return (spa_vdev_exit(spa, NULL, txg, EINVAL)); /* first, check to ensure we've got the right child count */ rvd = spa->spa_root_vdev; lastlog = 0; for (c = 0; c < rvd->vdev_children; c++) { vdev_t *vd = rvd->vdev_child[c]; /* don't count the holes & logs as children */ if (vd->vdev_islog || (vd->vdev_ops != &vdev_indirect_ops && !vdev_is_concrete(vd))) { if (lastlog == 0) lastlog = c; continue; } lastlog = 0; } if (children != (lastlog != 0 ? lastlog : rvd->vdev_children)) return (spa_vdev_exit(spa, NULL, txg, EINVAL)); /* next, ensure no spare or cache devices are part of the split */ if (nvlist_lookup_nvlist(nvl, ZPOOL_CONFIG_SPARES, &tmp) == 0 || nvlist_lookup_nvlist(nvl, ZPOOL_CONFIG_L2CACHE, &tmp) == 0) return (spa_vdev_exit(spa, NULL, txg, EINVAL)); vml = kmem_zalloc(children * sizeof (vdev_t *), KM_SLEEP); glist = kmem_zalloc(children * sizeof (uint64_t), KM_SLEEP); /* then, loop over each vdev and validate it */ for (c = 0; c < children; c++) { uint64_t is_hole = 0; (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_HOLE, &is_hole); if (is_hole != 0) { if (spa->spa_root_vdev->vdev_child[c]->vdev_ishole || spa->spa_root_vdev->vdev_child[c]->vdev_islog) { continue; } else { error = SET_ERROR(EINVAL); break; } } /* deal with indirect vdevs */ if (spa->spa_root_vdev->vdev_child[c]->vdev_ops == &vdev_indirect_ops) continue; /* which disk is going to be split? */ if (nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_GUID, &glist[c]) != 0) { error = SET_ERROR(EINVAL); break; } /* look it up in the spa */ vml[c] = spa_lookup_by_guid(spa, glist[c], B_FALSE); if (vml[c] == NULL) { error = SET_ERROR(ENODEV); break; } /* make sure there's nothing stopping the split */ if (vml[c]->vdev_parent->vdev_ops != &vdev_mirror_ops || vml[c]->vdev_islog || !vdev_is_concrete(vml[c]) || vml[c]->vdev_isspare || vml[c]->vdev_isl2cache || !vdev_writeable(vml[c]) || vml[c]->vdev_children != 0 || vml[c]->vdev_state != VDEV_STATE_HEALTHY || c != spa->spa_root_vdev->vdev_child[c]->vdev_id) { error = SET_ERROR(EINVAL); break; } if (vdev_dtl_required(vml[c]) || vdev_resilver_needed(vml[c], NULL, NULL)) { error = SET_ERROR(EBUSY); break; } /* we need certain info from the top level */ fnvlist_add_uint64(child[c], ZPOOL_CONFIG_METASLAB_ARRAY, vml[c]->vdev_top->vdev_ms_array); fnvlist_add_uint64(child[c], ZPOOL_CONFIG_METASLAB_SHIFT, vml[c]->vdev_top->vdev_ms_shift); fnvlist_add_uint64(child[c], ZPOOL_CONFIG_ASIZE, vml[c]->vdev_top->vdev_asize); fnvlist_add_uint64(child[c], ZPOOL_CONFIG_ASHIFT, vml[c]->vdev_top->vdev_ashift); /* transfer per-vdev ZAPs */ ASSERT3U(vml[c]->vdev_leaf_zap, !=, 0); VERIFY0(nvlist_add_uint64(child[c], ZPOOL_CONFIG_VDEV_LEAF_ZAP, vml[c]->vdev_leaf_zap)); ASSERT3U(vml[c]->vdev_top->vdev_top_zap, !=, 0); VERIFY0(nvlist_add_uint64(child[c], ZPOOL_CONFIG_VDEV_TOP_ZAP, vml[c]->vdev_parent->vdev_top_zap)); } if (error != 0) { kmem_free(vml, children * sizeof (vdev_t *)); kmem_free(glist, children * sizeof (uint64_t)); return (spa_vdev_exit(spa, NULL, txg, error)); } /* stop writers from using the disks */ for (c = 0; c < children; c++) { if (vml[c] != NULL) vml[c]->vdev_offline = B_TRUE; } vdev_reopen(spa->spa_root_vdev); /* * Temporarily record the splitting vdevs in the spa config. This * will disappear once the config is regenerated. */ nvl = fnvlist_alloc(); fnvlist_add_uint64_array(nvl, ZPOOL_CONFIG_SPLIT_LIST, glist, children); kmem_free(glist, children * sizeof (uint64_t)); mutex_enter(&spa->spa_props_lock); fnvlist_add_nvlist(spa->spa_config, ZPOOL_CONFIG_SPLIT, nvl); mutex_exit(&spa->spa_props_lock); spa->spa_config_splitting = nvl; vdev_config_dirty(spa->spa_root_vdev); /* configure and create the new pool */ fnvlist_add_string(config, ZPOOL_CONFIG_POOL_NAME, newname); fnvlist_add_uint64(config, ZPOOL_CONFIG_POOL_STATE, exp ? POOL_STATE_EXPORTED : POOL_STATE_ACTIVE); fnvlist_add_uint64(config, ZPOOL_CONFIG_VERSION, spa_version(spa)); fnvlist_add_uint64(config, ZPOOL_CONFIG_POOL_TXG, spa->spa_config_txg); fnvlist_add_uint64(config, ZPOOL_CONFIG_POOL_GUID, spa_generate_guid(NULL)); VERIFY0(nvlist_add_boolean(config, ZPOOL_CONFIG_HAS_PER_VDEV_ZAPS)); (void) nvlist_lookup_string(props, zpool_prop_to_name(ZPOOL_PROP_ALTROOT), &altroot); /* add the new pool to the namespace */ newspa = spa_add(newname, config, altroot); newspa->spa_avz_action = AVZ_ACTION_REBUILD; newspa->spa_config_txg = spa->spa_config_txg; spa_set_log_state(newspa, SPA_LOG_CLEAR); /* release the spa config lock, retaining the namespace lock */ spa_vdev_config_exit(spa, NULL, txg, 0, FTAG); if (zio_injection_enabled) zio_handle_panic_injection(spa, FTAG, 1); spa_activate(newspa, spa_mode_global); spa_async_suspend(newspa); /* * Temporarily stop the initializing and TRIM activity. We set the * state to ACTIVE so that we know to resume initializing or TRIM * once the split has completed. */ list_t vd_initialize_list; list_create(&vd_initialize_list, sizeof (vdev_t), offsetof(vdev_t, vdev_initialize_node)); list_t vd_trim_list; list_create(&vd_trim_list, sizeof (vdev_t), offsetof(vdev_t, vdev_trim_node)); for (c = 0; c < children; c++) { if (vml[c] != NULL && vml[c]->vdev_ops != &vdev_indirect_ops) { mutex_enter(&vml[c]->vdev_initialize_lock); vdev_initialize_stop(vml[c], VDEV_INITIALIZE_ACTIVE, &vd_initialize_list); mutex_exit(&vml[c]->vdev_initialize_lock); mutex_enter(&vml[c]->vdev_trim_lock); vdev_trim_stop(vml[c], VDEV_TRIM_ACTIVE, &vd_trim_list); mutex_exit(&vml[c]->vdev_trim_lock); } } vdev_initialize_stop_wait(spa, &vd_initialize_list); vdev_trim_stop_wait(spa, &vd_trim_list); list_destroy(&vd_initialize_list); list_destroy(&vd_trim_list); newspa->spa_config_source = SPA_CONFIG_SRC_SPLIT; newspa->spa_is_splitting = B_TRUE; /* create the new pool from the disks of the original pool */ error = spa_load(newspa, SPA_LOAD_IMPORT, SPA_IMPORT_ASSEMBLE); if (error) goto out; /* if that worked, generate a real config for the new pool */ if (newspa->spa_root_vdev != NULL) { newspa->spa_config_splitting = fnvlist_alloc(); fnvlist_add_uint64(newspa->spa_config_splitting, ZPOOL_CONFIG_SPLIT_GUID, spa_guid(spa)); spa_config_set(newspa, spa_config_generate(newspa, NULL, -1ULL, B_TRUE)); } /* set the props */ if (props != NULL) { spa_configfile_set(newspa, props, B_FALSE); error = spa_prop_set(newspa, props); if (error) goto out; } /* flush everything */ txg = spa_vdev_config_enter(newspa); vdev_config_dirty(newspa->spa_root_vdev); (void) spa_vdev_config_exit(newspa, NULL, txg, 0, FTAG); if (zio_injection_enabled) zio_handle_panic_injection(spa, FTAG, 2); spa_async_resume(newspa); /* finally, update the original pool's config */ txg = spa_vdev_config_enter(spa); tx = dmu_tx_create_dd(spa_get_dsl(spa)->dp_mos_dir); error = dmu_tx_assign(tx, TXG_WAIT); if (error != 0) dmu_tx_abort(tx); for (c = 0; c < children; c++) { if (vml[c] != NULL && vml[c]->vdev_ops != &vdev_indirect_ops) { vdev_t *tvd = vml[c]->vdev_top; /* * Need to be sure the detachable VDEV is not * on any *other* txg's DTL list to prevent it * from being accessed after it's freed. */ for (int t = 0; t < TXG_SIZE; t++) { (void) txg_list_remove_this( &tvd->vdev_dtl_list, vml[c], t); } vdev_split(vml[c]); if (error == 0) spa_history_log_internal(spa, "detach", tx, "vdev=%s", vml[c]->vdev_path); vdev_free(vml[c]); } } spa->spa_avz_action = AVZ_ACTION_REBUILD; vdev_config_dirty(spa->spa_root_vdev); spa->spa_config_splitting = NULL; nvlist_free(nvl); if (error == 0) dmu_tx_commit(tx); (void) spa_vdev_exit(spa, NULL, txg, 0); if (zio_injection_enabled) zio_handle_panic_injection(spa, FTAG, 3); /* split is complete; log a history record */ spa_history_log_internal(newspa, "split", NULL, "from pool %s", spa_name(spa)); newspa->spa_is_splitting = B_FALSE; kmem_free(vml, children * sizeof (vdev_t *)); /* if we're not going to mount the filesystems in userland, export */ if (exp) error = spa_export_common(newname, POOL_STATE_EXPORTED, NULL, B_FALSE, B_FALSE); return (error); out: spa_unload(newspa); spa_deactivate(newspa); spa_remove(newspa); txg = spa_vdev_config_enter(spa); /* re-online all offlined disks */ for (c = 0; c < children; c++) { if (vml[c] != NULL) vml[c]->vdev_offline = B_FALSE; } /* restart initializing or trimming disks as necessary */ spa_async_request(spa, SPA_ASYNC_INITIALIZE_RESTART); spa_async_request(spa, SPA_ASYNC_TRIM_RESTART); spa_async_request(spa, SPA_ASYNC_AUTOTRIM_RESTART); vdev_reopen(spa->spa_root_vdev); nvlist_free(spa->spa_config_splitting); spa->spa_config_splitting = NULL; (void) spa_vdev_exit(spa, NULL, txg, error); kmem_free(vml, children * sizeof (vdev_t *)); return (error); } /* * Find any device that's done replacing, or a vdev marked 'unspare' that's * currently spared, so we can detach it. */ static vdev_t * spa_vdev_resilver_done_hunt(vdev_t *vd) { vdev_t *newvd, *oldvd; for (int c = 0; c < vd->vdev_children; c++) { oldvd = spa_vdev_resilver_done_hunt(vd->vdev_child[c]); if (oldvd != NULL) return (oldvd); } /* * Check for a completed replacement. We always consider the first * vdev in the list to be the oldest vdev, and the last one to be * the newest (see spa_vdev_attach() for how that works). In * the case where the newest vdev is faulted, we will not automatically * remove it after a resilver completes. This is OK as it will require * user intervention to determine which disk the admin wishes to keep. */ if (vd->vdev_ops == &vdev_replacing_ops) { ASSERT(vd->vdev_children > 1); newvd = vd->vdev_child[vd->vdev_children - 1]; oldvd = vd->vdev_child[0]; if (vdev_dtl_empty(newvd, DTL_MISSING) && vdev_dtl_empty(newvd, DTL_OUTAGE) && !vdev_dtl_required(oldvd)) return (oldvd); } /* * Check for a completed resilver with the 'unspare' flag set. * Also potentially update faulted state. */ if (vd->vdev_ops == &vdev_spare_ops) { vdev_t *first = vd->vdev_child[0]; vdev_t *last = vd->vdev_child[vd->vdev_children - 1]; if (last->vdev_unspare) { oldvd = first; newvd = last; } else if (first->vdev_unspare) { oldvd = last; newvd = first; } else { oldvd = NULL; } if (oldvd != NULL && vdev_dtl_empty(newvd, DTL_MISSING) && vdev_dtl_empty(newvd, DTL_OUTAGE) && !vdev_dtl_required(oldvd)) return (oldvd); vdev_propagate_state(vd); /* * If there are more than two spares attached to a disk, * and those spares are not required, then we want to * attempt to free them up now so that they can be used * by other pools. Once we're back down to a single * disk+spare, we stop removing them. */ if (vd->vdev_children > 2) { newvd = vd->vdev_child[1]; if (newvd->vdev_isspare && last->vdev_isspare && vdev_dtl_empty(last, DTL_MISSING) && vdev_dtl_empty(last, DTL_OUTAGE) && !vdev_dtl_required(newvd)) return (newvd); } } return (NULL); } static void spa_vdev_resilver_done(spa_t *spa) { vdev_t *vd, *pvd, *ppvd; uint64_t guid, sguid, pguid, ppguid; spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); while ((vd = spa_vdev_resilver_done_hunt(spa->spa_root_vdev)) != NULL) { pvd = vd->vdev_parent; ppvd = pvd->vdev_parent; guid = vd->vdev_guid; pguid = pvd->vdev_guid; ppguid = ppvd->vdev_guid; sguid = 0; /* * If we have just finished replacing a hot spared device, then * we need to detach the parent's first child (the original hot * spare) as well. */ if (ppvd->vdev_ops == &vdev_spare_ops && pvd->vdev_id == 0 && ppvd->vdev_children == 2) { ASSERT(pvd->vdev_ops == &vdev_replacing_ops); sguid = ppvd->vdev_child[1]->vdev_guid; } ASSERT(vd->vdev_resilver_txg == 0 || !vdev_dtl_required(vd)); spa_config_exit(spa, SCL_ALL, FTAG); if (spa_vdev_detach(spa, guid, pguid, B_TRUE) != 0) return; if (sguid && spa_vdev_detach(spa, sguid, ppguid, B_TRUE) != 0) return; spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); } spa_config_exit(spa, SCL_ALL, FTAG); /* * If a detach was not performed above replace waiters will not have * been notified. In which case we must do so now. */ spa_notify_waiters(spa); } /* * Update the stored path or FRU for this vdev. */ static int spa_vdev_set_common(spa_t *spa, uint64_t guid, const char *value, boolean_t ispath) { vdev_t *vd; boolean_t sync = B_FALSE; ASSERT(spa_writeable(spa)); spa_vdev_state_enter(spa, SCL_ALL); if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL) return (spa_vdev_state_exit(spa, NULL, ENOENT)); if (!vd->vdev_ops->vdev_op_leaf) return (spa_vdev_state_exit(spa, NULL, ENOTSUP)); if (ispath) { if (strcmp(value, vd->vdev_path) != 0) { spa_strfree(vd->vdev_path); vd->vdev_path = spa_strdup(value); sync = B_TRUE; } } else { if (vd->vdev_fru == NULL) { vd->vdev_fru = spa_strdup(value); sync = B_TRUE; } else if (strcmp(value, vd->vdev_fru) != 0) { spa_strfree(vd->vdev_fru); vd->vdev_fru = spa_strdup(value); sync = B_TRUE; } } return (spa_vdev_state_exit(spa, sync ? vd : NULL, 0)); } int spa_vdev_setpath(spa_t *spa, uint64_t guid, const char *newpath) { return (spa_vdev_set_common(spa, guid, newpath, B_TRUE)); } int spa_vdev_setfru(spa_t *spa, uint64_t guid, const char *newfru) { return (spa_vdev_set_common(spa, guid, newfru, B_FALSE)); } /* * ========================================================================== * SPA Scanning * ========================================================================== */ int spa_scrub_pause_resume(spa_t *spa, pool_scrub_cmd_t cmd) { ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == 0); if (dsl_scan_resilvering(spa->spa_dsl_pool)) return (SET_ERROR(EBUSY)); return (dsl_scrub_set_pause_resume(spa->spa_dsl_pool, cmd)); } int spa_scan_stop(spa_t *spa) { ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == 0); if (dsl_scan_resilvering(spa->spa_dsl_pool)) return (SET_ERROR(EBUSY)); + return (dsl_scan_cancel(spa->spa_dsl_pool)); } int spa_scan(spa_t *spa, pool_scan_func_t func) { ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == 0); if (func >= POOL_SCAN_FUNCS || func == POOL_SCAN_NONE) return (SET_ERROR(ENOTSUP)); if (func == POOL_SCAN_RESILVER && !spa_feature_is_enabled(spa, SPA_FEATURE_RESILVER_DEFER)) return (SET_ERROR(ENOTSUP)); /* * If a resilver was requested, but there is no DTL on a * writeable leaf device, we have nothing to do. */ if (func == POOL_SCAN_RESILVER && !vdev_resilver_needed(spa->spa_root_vdev, NULL, NULL)) { spa_async_request(spa, SPA_ASYNC_RESILVER_DONE); return (0); } + if (func == POOL_SCAN_ERRORSCRUB && + !spa_feature_is_enabled(spa, SPA_FEATURE_HEAD_ERRLOG)) + return (SET_ERROR(ENOTSUP)); + return (dsl_scan(spa->spa_dsl_pool, func)); } /* * ========================================================================== * SPA async task processing * ========================================================================== */ static void spa_async_remove(spa_t *spa, vdev_t *vd) { if (vd->vdev_remove_wanted) { vd->vdev_remove_wanted = B_FALSE; vd->vdev_delayed_close = B_FALSE; vdev_set_state(vd, B_FALSE, VDEV_STATE_REMOVED, VDEV_AUX_NONE); /* * We want to clear the stats, but we don't want to do a full * vdev_clear() as that will cause us to throw away * degraded/faulted state as well as attempt to reopen the * device, all of which is a waste. */ vd->vdev_stat.vs_read_errors = 0; vd->vdev_stat.vs_write_errors = 0; vd->vdev_stat.vs_checksum_errors = 0; vdev_state_dirty(vd->vdev_top); /* Tell userspace that the vdev is gone. */ zfs_post_remove(spa, vd); } for (int c = 0; c < vd->vdev_children; c++) spa_async_remove(spa, vd->vdev_child[c]); } static void spa_async_probe(spa_t *spa, vdev_t *vd) { if (vd->vdev_probe_wanted) { vd->vdev_probe_wanted = B_FALSE; vdev_reopen(vd); /* vdev_open() does the actual probe */ } for (int c = 0; c < vd->vdev_children; c++) spa_async_probe(spa, vd->vdev_child[c]); } static void spa_async_autoexpand(spa_t *spa, vdev_t *vd) { if (!spa->spa_autoexpand) return; for (int c = 0; c < vd->vdev_children; c++) { vdev_t *cvd = vd->vdev_child[c]; spa_async_autoexpand(spa, cvd); } if (!vd->vdev_ops->vdev_op_leaf || vd->vdev_physpath == NULL) return; spa_event_notify(vd->vdev_spa, vd, NULL, ESC_ZFS_VDEV_AUTOEXPAND); } static __attribute__((noreturn)) void spa_async_thread(void *arg) { spa_t *spa = (spa_t *)arg; dsl_pool_t *dp = spa->spa_dsl_pool; int tasks; ASSERT(spa->spa_sync_on); mutex_enter(&spa->spa_async_lock); tasks = spa->spa_async_tasks; spa->spa_async_tasks = 0; mutex_exit(&spa->spa_async_lock); /* * See if the config needs to be updated. */ if (tasks & SPA_ASYNC_CONFIG_UPDATE) { uint64_t old_space, new_space; mutex_enter(&spa_namespace_lock); old_space = metaslab_class_get_space(spa_normal_class(spa)); old_space += metaslab_class_get_space(spa_special_class(spa)); old_space += metaslab_class_get_space(spa_dedup_class(spa)); old_space += metaslab_class_get_space( spa_embedded_log_class(spa)); spa_config_update(spa, SPA_CONFIG_UPDATE_POOL); new_space = metaslab_class_get_space(spa_normal_class(spa)); new_space += metaslab_class_get_space(spa_special_class(spa)); new_space += metaslab_class_get_space(spa_dedup_class(spa)); new_space += metaslab_class_get_space( spa_embedded_log_class(spa)); mutex_exit(&spa_namespace_lock); /* * If the pool grew as a result of the config update, * then log an internal history event. */ if (new_space != old_space) { spa_history_log_internal(spa, "vdev online", NULL, "pool '%s' size: %llu(+%llu)", spa_name(spa), (u_longlong_t)new_space, (u_longlong_t)(new_space - old_space)); } } /* * See if any devices need to be marked REMOVED. */ if (tasks & SPA_ASYNC_REMOVE) { spa_vdev_state_enter(spa, SCL_NONE); spa_async_remove(spa, spa->spa_root_vdev); for (int i = 0; i < spa->spa_l2cache.sav_count; i++) spa_async_remove(spa, spa->spa_l2cache.sav_vdevs[i]); for (int i = 0; i < spa->spa_spares.sav_count; i++) spa_async_remove(spa, spa->spa_spares.sav_vdevs[i]); (void) spa_vdev_state_exit(spa, NULL, 0); } if ((tasks & SPA_ASYNC_AUTOEXPAND) && !spa_suspended(spa)) { spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER); spa_async_autoexpand(spa, spa->spa_root_vdev); spa_config_exit(spa, SCL_CONFIG, FTAG); } /* * See if any devices need to be probed. */ if (tasks & SPA_ASYNC_PROBE) { spa_vdev_state_enter(spa, SCL_NONE); spa_async_probe(spa, spa->spa_root_vdev); (void) spa_vdev_state_exit(spa, NULL, 0); } /* * If any devices are done replacing, detach them. */ if (tasks & SPA_ASYNC_RESILVER_DONE || tasks & SPA_ASYNC_REBUILD_DONE || tasks & SPA_ASYNC_DETACH_SPARE) { spa_vdev_resilver_done(spa); } /* * Kick off a resilver. */ if (tasks & SPA_ASYNC_RESILVER && !vdev_rebuild_active(spa->spa_root_vdev) && (!dsl_scan_resilvering(dp) || !spa_feature_is_enabled(dp->dp_spa, SPA_FEATURE_RESILVER_DEFER))) dsl_scan_restart_resilver(dp, 0); if (tasks & SPA_ASYNC_INITIALIZE_RESTART) { mutex_enter(&spa_namespace_lock); spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER); vdev_initialize_restart(spa->spa_root_vdev); spa_config_exit(spa, SCL_CONFIG, FTAG); mutex_exit(&spa_namespace_lock); } if (tasks & SPA_ASYNC_TRIM_RESTART) { mutex_enter(&spa_namespace_lock); spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER); vdev_trim_restart(spa->spa_root_vdev); spa_config_exit(spa, SCL_CONFIG, FTAG); mutex_exit(&spa_namespace_lock); } if (tasks & SPA_ASYNC_AUTOTRIM_RESTART) { mutex_enter(&spa_namespace_lock); spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER); vdev_autotrim_restart(spa); spa_config_exit(spa, SCL_CONFIG, FTAG); mutex_exit(&spa_namespace_lock); } /* * Kick off L2 cache whole device TRIM. */ if (tasks & SPA_ASYNC_L2CACHE_TRIM) { mutex_enter(&spa_namespace_lock); spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER); vdev_trim_l2arc(spa); spa_config_exit(spa, SCL_CONFIG, FTAG); mutex_exit(&spa_namespace_lock); } /* * Kick off L2 cache rebuilding. */ if (tasks & SPA_ASYNC_L2CACHE_REBUILD) { mutex_enter(&spa_namespace_lock); spa_config_enter(spa, SCL_L2ARC, FTAG, RW_READER); l2arc_spa_rebuild_start(spa); spa_config_exit(spa, SCL_L2ARC, FTAG); mutex_exit(&spa_namespace_lock); } /* * Let the world know that we're done. */ mutex_enter(&spa->spa_async_lock); spa->spa_async_thread = NULL; cv_broadcast(&spa->spa_async_cv); mutex_exit(&spa->spa_async_lock); thread_exit(); } void spa_async_suspend(spa_t *spa) { mutex_enter(&spa->spa_async_lock); spa->spa_async_suspended++; while (spa->spa_async_thread != NULL) cv_wait(&spa->spa_async_cv, &spa->spa_async_lock); mutex_exit(&spa->spa_async_lock); spa_vdev_remove_suspend(spa); zthr_t *condense_thread = spa->spa_condense_zthr; if (condense_thread != NULL) zthr_cancel(condense_thread); zthr_t *discard_thread = spa->spa_checkpoint_discard_zthr; if (discard_thread != NULL) zthr_cancel(discard_thread); zthr_t *ll_delete_thread = spa->spa_livelist_delete_zthr; if (ll_delete_thread != NULL) zthr_cancel(ll_delete_thread); zthr_t *ll_condense_thread = spa->spa_livelist_condense_zthr; if (ll_condense_thread != NULL) zthr_cancel(ll_condense_thread); } void spa_async_resume(spa_t *spa) { mutex_enter(&spa->spa_async_lock); ASSERT(spa->spa_async_suspended != 0); spa->spa_async_suspended--; mutex_exit(&spa->spa_async_lock); spa_restart_removal(spa); zthr_t *condense_thread = spa->spa_condense_zthr; if (condense_thread != NULL) zthr_resume(condense_thread); zthr_t *discard_thread = spa->spa_checkpoint_discard_zthr; if (discard_thread != NULL) zthr_resume(discard_thread); zthr_t *ll_delete_thread = spa->spa_livelist_delete_zthr; if (ll_delete_thread != NULL) zthr_resume(ll_delete_thread); zthr_t *ll_condense_thread = spa->spa_livelist_condense_zthr; if (ll_condense_thread != NULL) zthr_resume(ll_condense_thread); } static boolean_t spa_async_tasks_pending(spa_t *spa) { uint_t non_config_tasks; uint_t config_task; boolean_t config_task_suspended; non_config_tasks = spa->spa_async_tasks & ~SPA_ASYNC_CONFIG_UPDATE; config_task = spa->spa_async_tasks & SPA_ASYNC_CONFIG_UPDATE; if (spa->spa_ccw_fail_time == 0) { config_task_suspended = B_FALSE; } else { config_task_suspended = (gethrtime() - spa->spa_ccw_fail_time) < ((hrtime_t)zfs_ccw_retry_interval * NANOSEC); } return (non_config_tasks || (config_task && !config_task_suspended)); } static void spa_async_dispatch(spa_t *spa) { mutex_enter(&spa->spa_async_lock); if (spa_async_tasks_pending(spa) && !spa->spa_async_suspended && spa->spa_async_thread == NULL) spa->spa_async_thread = thread_create(NULL, 0, spa_async_thread, spa, 0, &p0, TS_RUN, maxclsyspri); mutex_exit(&spa->spa_async_lock); } void spa_async_request(spa_t *spa, int task) { zfs_dbgmsg("spa=%s async request task=%u", spa->spa_name, task); mutex_enter(&spa->spa_async_lock); spa->spa_async_tasks |= task; mutex_exit(&spa->spa_async_lock); } int spa_async_tasks(spa_t *spa) { return (spa->spa_async_tasks); } /* * ========================================================================== * SPA syncing routines * ========================================================================== */ static int bpobj_enqueue_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed, dmu_tx_t *tx) { bpobj_t *bpo = arg; bpobj_enqueue(bpo, bp, bp_freed, tx); return (0); } int bpobj_enqueue_alloc_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx) { return (bpobj_enqueue_cb(arg, bp, B_FALSE, tx)); } int bpobj_enqueue_free_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx) { return (bpobj_enqueue_cb(arg, bp, B_TRUE, tx)); } static int spa_free_sync_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx) { zio_t *pio = arg; zio_nowait(zio_free_sync(pio, pio->io_spa, dmu_tx_get_txg(tx), bp, pio->io_flags)); return (0); } static int bpobj_spa_free_sync_cb(void *arg, const blkptr_t *bp, boolean_t bp_freed, dmu_tx_t *tx) { ASSERT(!bp_freed); return (spa_free_sync_cb(arg, bp, tx)); } /* * Note: this simple function is not inlined to make it easier to dtrace the * amount of time spent syncing frees. */ static void spa_sync_frees(spa_t *spa, bplist_t *bpl, dmu_tx_t *tx) { zio_t *zio = zio_root(spa, NULL, NULL, 0); bplist_iterate(bpl, spa_free_sync_cb, zio, tx); VERIFY(zio_wait(zio) == 0); } /* * Note: this simple function is not inlined to make it easier to dtrace the * amount of time spent syncing deferred frees. */ static void spa_sync_deferred_frees(spa_t *spa, dmu_tx_t *tx) { if (spa_sync_pass(spa) != 1) return; /* * Note: * If the log space map feature is active, we stop deferring * frees to the next TXG and therefore running this function * would be considered a no-op as spa_deferred_bpobj should * not have any entries. * * That said we run this function anyway (instead of returning * immediately) for the edge-case scenario where we just * activated the log space map feature in this TXG but we have * deferred frees from the previous TXG. */ zio_t *zio = zio_root(spa, NULL, NULL, 0); VERIFY3U(bpobj_iterate(&spa->spa_deferred_bpobj, bpobj_spa_free_sync_cb, zio, tx), ==, 0); VERIFY0(zio_wait(zio)); } static void spa_sync_nvlist(spa_t *spa, uint64_t obj, nvlist_t *nv, dmu_tx_t *tx) { char *packed = NULL; size_t bufsize; size_t nvsize = 0; dmu_buf_t *db; VERIFY(nvlist_size(nv, &nvsize, NV_ENCODE_XDR) == 0); /* * Write full (SPA_CONFIG_BLOCKSIZE) blocks of configuration * information. This avoids the dmu_buf_will_dirty() path and * saves us a pre-read to get data we don't actually care about. */ bufsize = P2ROUNDUP((uint64_t)nvsize, SPA_CONFIG_BLOCKSIZE); packed = vmem_alloc(bufsize, KM_SLEEP); VERIFY(nvlist_pack(nv, &packed, &nvsize, NV_ENCODE_XDR, KM_SLEEP) == 0); memset(packed + nvsize, 0, bufsize - nvsize); dmu_write(spa->spa_meta_objset, obj, 0, bufsize, packed, tx); vmem_free(packed, bufsize); VERIFY(0 == dmu_bonus_hold(spa->spa_meta_objset, obj, FTAG, &db)); dmu_buf_will_dirty(db, tx); *(uint64_t *)db->db_data = nvsize; dmu_buf_rele(db, FTAG); } static void spa_sync_aux_dev(spa_t *spa, spa_aux_vdev_t *sav, dmu_tx_t *tx, const char *config, const char *entry) { nvlist_t *nvroot; nvlist_t **list; int i; if (!sav->sav_sync) return; /* * Update the MOS nvlist describing the list of available devices. * spa_validate_aux() will have already made sure this nvlist is * valid and the vdevs are labeled appropriately. */ if (sav->sav_object == 0) { sav->sav_object = dmu_object_alloc(spa->spa_meta_objset, DMU_OT_PACKED_NVLIST, 1 << 14, DMU_OT_PACKED_NVLIST_SIZE, sizeof (uint64_t), tx); VERIFY(zap_update(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT, entry, sizeof (uint64_t), 1, &sav->sav_object, tx) == 0); } nvroot = fnvlist_alloc(); if (sav->sav_count == 0) { fnvlist_add_nvlist_array(nvroot, config, (const nvlist_t * const *)NULL, 0); } else { list = kmem_alloc(sav->sav_count*sizeof (void *), KM_SLEEP); for (i = 0; i < sav->sav_count; i++) list[i] = vdev_config_generate(spa, sav->sav_vdevs[i], B_FALSE, VDEV_CONFIG_L2CACHE); fnvlist_add_nvlist_array(nvroot, config, (const nvlist_t * const *)list, sav->sav_count); for (i = 0; i < sav->sav_count; i++) nvlist_free(list[i]); kmem_free(list, sav->sav_count * sizeof (void *)); } spa_sync_nvlist(spa, sav->sav_object, nvroot, tx); nvlist_free(nvroot); sav->sav_sync = B_FALSE; } /* * Rebuild spa's all-vdev ZAP from the vdev ZAPs indicated in each vdev_t. * The all-vdev ZAP must be empty. */ static void spa_avz_build(vdev_t *vd, uint64_t avz, dmu_tx_t *tx) { spa_t *spa = vd->vdev_spa; if (vd->vdev_root_zap != 0 && spa_feature_is_active(spa, SPA_FEATURE_AVZ_V2)) { VERIFY0(zap_add_int(spa->spa_meta_objset, avz, vd->vdev_root_zap, tx)); } if (vd->vdev_top_zap != 0) { VERIFY0(zap_add_int(spa->spa_meta_objset, avz, vd->vdev_top_zap, tx)); } if (vd->vdev_leaf_zap != 0) { VERIFY0(zap_add_int(spa->spa_meta_objset, avz, vd->vdev_leaf_zap, tx)); } for (uint64_t i = 0; i < vd->vdev_children; i++) { spa_avz_build(vd->vdev_child[i], avz, tx); } } static void spa_sync_config_object(spa_t *spa, dmu_tx_t *tx) { nvlist_t *config; /* * If the pool is being imported from a pre-per-vdev-ZAP version of ZFS, * its config may not be dirty but we still need to build per-vdev ZAPs. * Similarly, if the pool is being assembled (e.g. after a split), we * need to rebuild the AVZ although the config may not be dirty. */ if (list_is_empty(&spa->spa_config_dirty_list) && spa->spa_avz_action == AVZ_ACTION_NONE) return; spa_config_enter(spa, SCL_STATE, FTAG, RW_READER); ASSERT(spa->spa_avz_action == AVZ_ACTION_NONE || spa->spa_avz_action == AVZ_ACTION_INITIALIZE || spa->spa_all_vdev_zaps != 0); if (spa->spa_avz_action == AVZ_ACTION_REBUILD) { /* Make and build the new AVZ */ uint64_t new_avz = zap_create(spa->spa_meta_objset, DMU_OTN_ZAP_METADATA, DMU_OT_NONE, 0, tx); spa_avz_build(spa->spa_root_vdev, new_avz, tx); /* Diff old AVZ with new one */ zap_cursor_t zc; zap_attribute_t za; for (zap_cursor_init(&zc, spa->spa_meta_objset, spa->spa_all_vdev_zaps); zap_cursor_retrieve(&zc, &za) == 0; zap_cursor_advance(&zc)) { uint64_t vdzap = za.za_first_integer; if (zap_lookup_int(spa->spa_meta_objset, new_avz, vdzap) == ENOENT) { /* * ZAP is listed in old AVZ but not in new one; * destroy it */ VERIFY0(zap_destroy(spa->spa_meta_objset, vdzap, tx)); } } zap_cursor_fini(&zc); /* Destroy the old AVZ */ VERIFY0(zap_destroy(spa->spa_meta_objset, spa->spa_all_vdev_zaps, tx)); /* Replace the old AVZ in the dir obj with the new one */ VERIFY0(zap_update(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_VDEV_ZAP_MAP, sizeof (new_avz), 1, &new_avz, tx)); spa->spa_all_vdev_zaps = new_avz; } else if (spa->spa_avz_action == AVZ_ACTION_DESTROY) { zap_cursor_t zc; zap_attribute_t za; /* Walk through the AVZ and destroy all listed ZAPs */ for (zap_cursor_init(&zc, spa->spa_meta_objset, spa->spa_all_vdev_zaps); zap_cursor_retrieve(&zc, &za) == 0; zap_cursor_advance(&zc)) { uint64_t zap = za.za_first_integer; VERIFY0(zap_destroy(spa->spa_meta_objset, zap, tx)); } zap_cursor_fini(&zc); /* Destroy and unlink the AVZ itself */ VERIFY0(zap_destroy(spa->spa_meta_objset, spa->spa_all_vdev_zaps, tx)); VERIFY0(zap_remove(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_VDEV_ZAP_MAP, tx)); spa->spa_all_vdev_zaps = 0; } if (spa->spa_all_vdev_zaps == 0) { spa->spa_all_vdev_zaps = zap_create_link(spa->spa_meta_objset, DMU_OTN_ZAP_METADATA, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_VDEV_ZAP_MAP, tx); } spa->spa_avz_action = AVZ_ACTION_NONE; /* Create ZAPs for vdevs that don't have them. */ vdev_construct_zaps(spa->spa_root_vdev, tx); config = spa_config_generate(spa, spa->spa_root_vdev, dmu_tx_get_txg(tx), B_FALSE); /* * If we're upgrading the spa version then make sure that * the config object gets updated with the correct version. */ if (spa->spa_ubsync.ub_version < spa->spa_uberblock.ub_version) fnvlist_add_uint64(config, ZPOOL_CONFIG_VERSION, spa->spa_uberblock.ub_version); spa_config_exit(spa, SCL_STATE, FTAG); nvlist_free(spa->spa_config_syncing); spa->spa_config_syncing = config; spa_sync_nvlist(spa, spa->spa_config_object, config, tx); } static void spa_sync_version(void *arg, dmu_tx_t *tx) { uint64_t *versionp = arg; uint64_t version = *versionp; spa_t *spa = dmu_tx_pool(tx)->dp_spa; /* * Setting the version is special cased when first creating the pool. */ ASSERT(tx->tx_txg != TXG_INITIAL); ASSERT(SPA_VERSION_IS_SUPPORTED(version)); ASSERT(version >= spa_version(spa)); spa->spa_uberblock.ub_version = version; vdev_config_dirty(spa->spa_root_vdev); spa_history_log_internal(spa, "set", tx, "version=%lld", (longlong_t)version); } /* * Set zpool properties. */ static void spa_sync_props(void *arg, dmu_tx_t *tx) { nvlist_t *nvp = arg; spa_t *spa = dmu_tx_pool(tx)->dp_spa; objset_t *mos = spa->spa_meta_objset; nvpair_t *elem = NULL; mutex_enter(&spa->spa_props_lock); while ((elem = nvlist_next_nvpair(nvp, elem))) { uint64_t intval; const char *strval, *fname; zpool_prop_t prop; const char *propname; const char *elemname = nvpair_name(elem); zprop_type_t proptype; spa_feature_t fid; switch (prop = zpool_name_to_prop(elemname)) { case ZPOOL_PROP_VERSION: intval = fnvpair_value_uint64(elem); /* * The version is synced separately before other * properties and should be correct by now. */ ASSERT3U(spa_version(spa), >=, intval); break; case ZPOOL_PROP_ALTROOT: /* * 'altroot' is a non-persistent property. It should * have been set temporarily at creation or import time. */ ASSERT(spa->spa_root != NULL); break; case ZPOOL_PROP_READONLY: case ZPOOL_PROP_CACHEFILE: /* * 'readonly' and 'cachefile' are also non-persistent * properties. */ break; case ZPOOL_PROP_COMMENT: strval = fnvpair_value_string(elem); if (spa->spa_comment != NULL) spa_strfree(spa->spa_comment); spa->spa_comment = spa_strdup(strval); /* * We need to dirty the configuration on all the vdevs * so that their labels get updated. We also need to * update the cache file to keep it in sync with the * MOS version. It's unnecessary to do this for pool * creation since the vdev's configuration has already * been dirtied. */ if (tx->tx_txg != TXG_INITIAL) { vdev_config_dirty(spa->spa_root_vdev); spa_async_request(spa, SPA_ASYNC_CONFIG_UPDATE); } spa_history_log_internal(spa, "set", tx, "%s=%s", elemname, strval); break; case ZPOOL_PROP_COMPATIBILITY: strval = fnvpair_value_string(elem); if (spa->spa_compatibility != NULL) spa_strfree(spa->spa_compatibility); spa->spa_compatibility = spa_strdup(strval); /* * Dirty the configuration on vdevs as above. */ if (tx->tx_txg != TXG_INITIAL) { vdev_config_dirty(spa->spa_root_vdev); spa_async_request(spa, SPA_ASYNC_CONFIG_UPDATE); } spa_history_log_internal(spa, "set", tx, "%s=%s", nvpair_name(elem), strval); break; case ZPOOL_PROP_INVAL: if (zpool_prop_feature(elemname)) { fname = strchr(elemname, '@') + 1; VERIFY0(zfeature_lookup_name(fname, &fid)); spa_feature_enable(spa, fid, tx); spa_history_log_internal(spa, "set", tx, "%s=enabled", elemname); break; } else if (!zfs_prop_user(elemname)) { ASSERT(zpool_prop_feature(elemname)); break; } zfs_fallthrough; default: /* * Set pool property values in the poolprops mos object. */ if (spa->spa_pool_props_object == 0) { spa->spa_pool_props_object = zap_create_link(mos, DMU_OT_POOL_PROPS, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_PROPS, tx); } /* normalize the property name */ - propname = zpool_prop_to_name(prop); - proptype = zpool_prop_get_type(prop); - if (prop == ZPOOL_PROP_INVAL && - zfs_prop_user(elemname)) { + if (prop == ZPOOL_PROP_INVAL) { propname = elemname; proptype = PROP_TYPE_STRING; + } else { + propname = zpool_prop_to_name(prop); + proptype = zpool_prop_get_type(prop); } if (nvpair_type(elem) == DATA_TYPE_STRING) { ASSERT(proptype == PROP_TYPE_STRING); strval = fnvpair_value_string(elem); VERIFY0(zap_update(mos, spa->spa_pool_props_object, propname, 1, strlen(strval) + 1, strval, tx)); spa_history_log_internal(spa, "set", tx, "%s=%s", elemname, strval); } else if (nvpair_type(elem) == DATA_TYPE_UINT64) { intval = fnvpair_value_uint64(elem); if (proptype == PROP_TYPE_INDEX) { const char *unused; VERIFY0(zpool_prop_index_to_string( prop, intval, &unused)); } VERIFY0(zap_update(mos, spa->spa_pool_props_object, propname, 8, 1, &intval, tx)); spa_history_log_internal(spa, "set", tx, "%s=%lld", elemname, (longlong_t)intval); switch (prop) { case ZPOOL_PROP_DELEGATION: spa->spa_delegation = intval; break; case ZPOOL_PROP_BOOTFS: spa->spa_bootfs = intval; break; case ZPOOL_PROP_FAILUREMODE: spa->spa_failmode = intval; break; case ZPOOL_PROP_AUTOTRIM: spa->spa_autotrim = intval; spa_async_request(spa, SPA_ASYNC_AUTOTRIM_RESTART); break; case ZPOOL_PROP_AUTOEXPAND: spa->spa_autoexpand = intval; if (tx->tx_txg != TXG_INITIAL) spa_async_request(spa, SPA_ASYNC_AUTOEXPAND); break; case ZPOOL_PROP_MULTIHOST: spa->spa_multihost = intval; break; default: break; } } else { ASSERT(0); /* not allowed */ } } } mutex_exit(&spa->spa_props_lock); } /* * Perform one-time upgrade on-disk changes. spa_version() does not * reflect the new version this txg, so there must be no changes this * txg to anything that the upgrade code depends on after it executes. * Therefore this must be called after dsl_pool_sync() does the sync * tasks. */ static void spa_sync_upgrades(spa_t *spa, dmu_tx_t *tx) { if (spa_sync_pass(spa) != 1) return; dsl_pool_t *dp = spa->spa_dsl_pool; rrw_enter(&dp->dp_config_rwlock, RW_WRITER, FTAG); if (spa->spa_ubsync.ub_version < SPA_VERSION_ORIGIN && spa->spa_uberblock.ub_version >= SPA_VERSION_ORIGIN) { dsl_pool_create_origin(dp, tx); /* Keeping the origin open increases spa_minref */ spa->spa_minref += 3; } if (spa->spa_ubsync.ub_version < SPA_VERSION_NEXT_CLONES && spa->spa_uberblock.ub_version >= SPA_VERSION_NEXT_CLONES) { dsl_pool_upgrade_clones(dp, tx); } if (spa->spa_ubsync.ub_version < SPA_VERSION_DIR_CLONES && spa->spa_uberblock.ub_version >= SPA_VERSION_DIR_CLONES) { dsl_pool_upgrade_dir_clones(dp, tx); /* Keeping the freedir open increases spa_minref */ spa->spa_minref += 3; } if (spa->spa_ubsync.ub_version < SPA_VERSION_FEATURES && spa->spa_uberblock.ub_version >= SPA_VERSION_FEATURES) { spa_feature_create_zap_objects(spa, tx); } /* * LZ4_COMPRESS feature's behaviour was changed to activate_on_enable * when possibility to use lz4 compression for metadata was added * Old pools that have this feature enabled must be upgraded to have * this feature active */ if (spa->spa_uberblock.ub_version >= SPA_VERSION_FEATURES) { boolean_t lz4_en = spa_feature_is_enabled(spa, SPA_FEATURE_LZ4_COMPRESS); boolean_t lz4_ac = spa_feature_is_active(spa, SPA_FEATURE_LZ4_COMPRESS); if (lz4_en && !lz4_ac) spa_feature_incr(spa, SPA_FEATURE_LZ4_COMPRESS, tx); } /* * If we haven't written the salt, do so now. Note that the * feature may not be activated yet, but that's fine since * the presence of this ZAP entry is backwards compatible. */ if (zap_contains(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_CHECKSUM_SALT) == ENOENT) { VERIFY0(zap_add(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_CHECKSUM_SALT, 1, sizeof (spa->spa_cksum_salt.zcs_bytes), spa->spa_cksum_salt.zcs_bytes, tx)); } rrw_exit(&dp->dp_config_rwlock, FTAG); } static void vdev_indirect_state_sync_verify(vdev_t *vd) { vdev_indirect_mapping_t *vim __maybe_unused = vd->vdev_indirect_mapping; vdev_indirect_births_t *vib __maybe_unused = vd->vdev_indirect_births; if (vd->vdev_ops == &vdev_indirect_ops) { ASSERT(vim != NULL); ASSERT(vib != NULL); } uint64_t obsolete_sm_object = 0; ASSERT0(vdev_obsolete_sm_object(vd, &obsolete_sm_object)); if (obsolete_sm_object != 0) { ASSERT(vd->vdev_obsolete_sm != NULL); ASSERT(vd->vdev_removing || vd->vdev_ops == &vdev_indirect_ops); ASSERT(vdev_indirect_mapping_num_entries(vim) > 0); ASSERT(vdev_indirect_mapping_bytes_mapped(vim) > 0); ASSERT3U(obsolete_sm_object, ==, space_map_object(vd->vdev_obsolete_sm)); ASSERT3U(vdev_indirect_mapping_bytes_mapped(vim), >=, space_map_allocated(vd->vdev_obsolete_sm)); } ASSERT(vd->vdev_obsolete_segments != NULL); /* * Since frees / remaps to an indirect vdev can only * happen in syncing context, the obsolete segments * tree must be empty when we start syncing. */ ASSERT0(range_tree_space(vd->vdev_obsolete_segments)); } /* * Set the top-level vdev's max queue depth. Evaluate each top-level's * async write queue depth in case it changed. The max queue depth will * not change in the middle of syncing out this txg. */ static void spa_sync_adjust_vdev_max_queue_depth(spa_t *spa) { ASSERT(spa_writeable(spa)); vdev_t *rvd = spa->spa_root_vdev; uint32_t max_queue_depth = zfs_vdev_async_write_max_active * zfs_vdev_queue_depth_pct / 100; metaslab_class_t *normal = spa_normal_class(spa); metaslab_class_t *special = spa_special_class(spa); metaslab_class_t *dedup = spa_dedup_class(spa); uint64_t slots_per_allocator = 0; for (int c = 0; c < rvd->vdev_children; c++) { vdev_t *tvd = rvd->vdev_child[c]; metaslab_group_t *mg = tvd->vdev_mg; if (mg == NULL || !metaslab_group_initialized(mg)) continue; metaslab_class_t *mc = mg->mg_class; if (mc != normal && mc != special && mc != dedup) continue; /* * It is safe to do a lock-free check here because only async * allocations look at mg_max_alloc_queue_depth, and async * allocations all happen from spa_sync(). */ for (int i = 0; i < mg->mg_allocators; i++) { ASSERT0(zfs_refcount_count( &(mg->mg_allocator[i].mga_alloc_queue_depth))); } mg->mg_max_alloc_queue_depth = max_queue_depth; for (int i = 0; i < mg->mg_allocators; i++) { mg->mg_allocator[i].mga_cur_max_alloc_queue_depth = zfs_vdev_def_queue_depth; } slots_per_allocator += zfs_vdev_def_queue_depth; } for (int i = 0; i < spa->spa_alloc_count; i++) { ASSERT0(zfs_refcount_count(&normal->mc_allocator[i]. mca_alloc_slots)); ASSERT0(zfs_refcount_count(&special->mc_allocator[i]. mca_alloc_slots)); ASSERT0(zfs_refcount_count(&dedup->mc_allocator[i]. mca_alloc_slots)); normal->mc_allocator[i].mca_alloc_max_slots = slots_per_allocator; special->mc_allocator[i].mca_alloc_max_slots = slots_per_allocator; dedup->mc_allocator[i].mca_alloc_max_slots = slots_per_allocator; } normal->mc_alloc_throttle_enabled = zio_dva_throttle_enabled; special->mc_alloc_throttle_enabled = zio_dva_throttle_enabled; dedup->mc_alloc_throttle_enabled = zio_dva_throttle_enabled; } static void spa_sync_condense_indirect(spa_t *spa, dmu_tx_t *tx) { ASSERT(spa_writeable(spa)); vdev_t *rvd = spa->spa_root_vdev; for (int c = 0; c < rvd->vdev_children; c++) { vdev_t *vd = rvd->vdev_child[c]; vdev_indirect_state_sync_verify(vd); if (vdev_indirect_should_condense(vd)) { spa_condense_indirect_start_sync(vd, tx); break; } } } static void spa_sync_iterate_to_convergence(spa_t *spa, dmu_tx_t *tx) { objset_t *mos = spa->spa_meta_objset; dsl_pool_t *dp = spa->spa_dsl_pool; uint64_t txg = tx->tx_txg; bplist_t *free_bpl = &spa->spa_free_bplist[txg & TXG_MASK]; do { int pass = ++spa->spa_sync_pass; spa_sync_config_object(spa, tx); spa_sync_aux_dev(spa, &spa->spa_spares, tx, ZPOOL_CONFIG_SPARES, DMU_POOL_SPARES); spa_sync_aux_dev(spa, &spa->spa_l2cache, tx, ZPOOL_CONFIG_L2CACHE, DMU_POOL_L2CACHE); spa_errlog_sync(spa, txg); dsl_pool_sync(dp, txg); if (pass < zfs_sync_pass_deferred_free || spa_feature_is_active(spa, SPA_FEATURE_LOG_SPACEMAP)) { /* * If the log space map feature is active we don't * care about deferred frees and the deferred bpobj * as the log space map should effectively have the * same results (i.e. appending only to one object). */ spa_sync_frees(spa, free_bpl, tx); } else { /* * We can not defer frees in pass 1, because * we sync the deferred frees later in pass 1. */ ASSERT3U(pass, >, 1); bplist_iterate(free_bpl, bpobj_enqueue_alloc_cb, &spa->spa_deferred_bpobj, tx); } brt_sync(spa, txg); ddt_sync(spa, txg); dsl_scan_sync(dp, tx); + dsl_errorscrub_sync(dp, tx); svr_sync(spa, tx); spa_sync_upgrades(spa, tx); spa_flush_metaslabs(spa, tx); vdev_t *vd = NULL; while ((vd = txg_list_remove(&spa->spa_vdev_txg_list, txg)) != NULL) vdev_sync(vd, txg); /* * Note: We need to check if the MOS is dirty because we could * have marked the MOS dirty without updating the uberblock * (e.g. if we have sync tasks but no dirty user data). We need * to check the uberblock's rootbp because it is updated if we * have synced out dirty data (though in this case the MOS will * most likely also be dirty due to second order effects, we * don't want to rely on that here). */ if (pass == 1 && spa->spa_uberblock.ub_rootbp.blk_birth < txg && !dmu_objset_is_dirty(mos, txg)) { /* * Nothing changed on the first pass, therefore this * TXG is a no-op. Avoid syncing deferred frees, so * that we can keep this TXG as a no-op. */ ASSERT(txg_list_empty(&dp->dp_dirty_datasets, txg)); ASSERT(txg_list_empty(&dp->dp_dirty_dirs, txg)); ASSERT(txg_list_empty(&dp->dp_sync_tasks, txg)); ASSERT(txg_list_empty(&dp->dp_early_sync_tasks, txg)); break; } spa_sync_deferred_frees(spa, tx); } while (dmu_objset_is_dirty(mos, txg)); } /* * Rewrite the vdev configuration (which includes the uberblock) to * commit the transaction group. * * If there are no dirty vdevs, we sync the uberblock to a few random * top-level vdevs that are known to be visible in the config cache * (see spa_vdev_add() for a complete description). If there *are* dirty * vdevs, sync the uberblock to all vdevs. */ static void spa_sync_rewrite_vdev_config(spa_t *spa, dmu_tx_t *tx) { vdev_t *rvd = spa->spa_root_vdev; uint64_t txg = tx->tx_txg; for (;;) { int error = 0; /* * We hold SCL_STATE to prevent vdev open/close/etc. * while we're attempting to write the vdev labels. */ spa_config_enter(spa, SCL_STATE, FTAG, RW_READER); if (list_is_empty(&spa->spa_config_dirty_list)) { vdev_t *svd[SPA_SYNC_MIN_VDEVS] = { NULL }; int svdcount = 0; int children = rvd->vdev_children; int c0 = random_in_range(children); for (int c = 0; c < children; c++) { vdev_t *vd = rvd->vdev_child[(c0 + c) % children]; /* Stop when revisiting the first vdev */ if (c > 0 && svd[0] == vd) break; if (vd->vdev_ms_array == 0 || vd->vdev_islog || !vdev_is_concrete(vd)) continue; svd[svdcount++] = vd; if (svdcount == SPA_SYNC_MIN_VDEVS) break; } error = vdev_config_sync(svd, svdcount, txg); } else { error = vdev_config_sync(rvd->vdev_child, rvd->vdev_children, txg); } if (error == 0) spa->spa_last_synced_guid = rvd->vdev_guid; spa_config_exit(spa, SCL_STATE, FTAG); if (error == 0) break; zio_suspend(spa, NULL, ZIO_SUSPEND_IOERR); zio_resume_wait(spa); } } /* * Sync the specified transaction group. New blocks may be dirtied as * part of the process, so we iterate until it converges. */ void spa_sync(spa_t *spa, uint64_t txg) { vdev_t *vd = NULL; VERIFY(spa_writeable(spa)); /* * Wait for i/os issued in open context that need to complete * before this txg syncs. */ (void) zio_wait(spa->spa_txg_zio[txg & TXG_MASK]); spa->spa_txg_zio[txg & TXG_MASK] = zio_root(spa, NULL, NULL, ZIO_FLAG_CANFAIL); /* * Now that there can be no more cloning in this transaction group, * but we are still before issuing frees, we can process pending BRT * updates. */ brt_pending_apply(spa, txg); /* * Lock out configuration changes. */ spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER); spa->spa_syncing_txg = txg; spa->spa_sync_pass = 0; for (int i = 0; i < spa->spa_alloc_count; i++) { mutex_enter(&spa->spa_allocs[i].spaa_lock); VERIFY0(avl_numnodes(&spa->spa_allocs[i].spaa_tree)); mutex_exit(&spa->spa_allocs[i].spaa_lock); } /* * If there are any pending vdev state changes, convert them * into config changes that go out with this transaction group. */ spa_config_enter(spa, SCL_STATE, FTAG, RW_READER); while ((vd = list_head(&spa->spa_state_dirty_list)) != NULL) { /* Avoid holding the write lock unless actually necessary */ if (vd->vdev_aux == NULL) { vdev_state_clean(vd); vdev_config_dirty(vd); continue; } /* * We need the write lock here because, for aux vdevs, * calling vdev_config_dirty() modifies sav_config. * This is ugly and will become unnecessary when we * eliminate the aux vdev wart by integrating all vdevs * into the root vdev tree. */ spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG); spa_config_enter(spa, SCL_CONFIG | SCL_STATE, FTAG, RW_WRITER); while ((vd = list_head(&spa->spa_state_dirty_list)) != NULL) { vdev_state_clean(vd); vdev_config_dirty(vd); } spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG); spa_config_enter(spa, SCL_CONFIG | SCL_STATE, FTAG, RW_READER); } spa_config_exit(spa, SCL_STATE, FTAG); dsl_pool_t *dp = spa->spa_dsl_pool; dmu_tx_t *tx = dmu_tx_create_assigned(dp, txg); spa->spa_sync_starttime = gethrtime(); taskq_cancel_id(system_delay_taskq, spa->spa_deadman_tqid); spa->spa_deadman_tqid = taskq_dispatch_delay(system_delay_taskq, spa_deadman, spa, TQ_SLEEP, ddi_get_lbolt() + NSEC_TO_TICK(spa->spa_deadman_synctime)); /* * If we are upgrading to SPA_VERSION_RAIDZ_DEFLATE this txg, * set spa_deflate if we have no raid-z vdevs. */ if (spa->spa_ubsync.ub_version < SPA_VERSION_RAIDZ_DEFLATE && spa->spa_uberblock.ub_version >= SPA_VERSION_RAIDZ_DEFLATE) { vdev_t *rvd = spa->spa_root_vdev; int i; for (i = 0; i < rvd->vdev_children; i++) { vd = rvd->vdev_child[i]; if (vd->vdev_deflate_ratio != SPA_MINBLOCKSIZE) break; } if (i == rvd->vdev_children) { spa->spa_deflate = TRUE; VERIFY0(zap_add(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_DEFLATE, sizeof (uint64_t), 1, &spa->spa_deflate, tx)); } } spa_sync_adjust_vdev_max_queue_depth(spa); spa_sync_condense_indirect(spa, tx); spa_sync_iterate_to_convergence(spa, tx); #ifdef ZFS_DEBUG if (!list_is_empty(&spa->spa_config_dirty_list)) { /* * Make sure that the number of ZAPs for all the vdevs matches * the number of ZAPs in the per-vdev ZAP list. This only gets * called if the config is dirty; otherwise there may be * outstanding AVZ operations that weren't completed in * spa_sync_config_object. */ uint64_t all_vdev_zap_entry_count; ASSERT0(zap_count(spa->spa_meta_objset, spa->spa_all_vdev_zaps, &all_vdev_zap_entry_count)); ASSERT3U(vdev_count_verify_zaps(spa->spa_root_vdev), ==, all_vdev_zap_entry_count); } #endif if (spa->spa_vdev_removal != NULL) { ASSERT0(spa->spa_vdev_removal->svr_bytes_done[txg & TXG_MASK]); } spa_sync_rewrite_vdev_config(spa, tx); dmu_tx_commit(tx); taskq_cancel_id(system_delay_taskq, spa->spa_deadman_tqid); spa->spa_deadman_tqid = 0; /* * Clear the dirty config list. */ while ((vd = list_head(&spa->spa_config_dirty_list)) != NULL) vdev_config_clean(vd); /* * Now that the new config has synced transactionally, * let it become visible to the config cache. */ if (spa->spa_config_syncing != NULL) { spa_config_set(spa, spa->spa_config_syncing); spa->spa_config_txg = txg; spa->spa_config_syncing = NULL; } dsl_pool_sync_done(dp, txg); for (int i = 0; i < spa->spa_alloc_count; i++) { mutex_enter(&spa->spa_allocs[i].spaa_lock); VERIFY0(avl_numnodes(&spa->spa_allocs[i].spaa_tree)); mutex_exit(&spa->spa_allocs[i].spaa_lock); } /* * Update usable space statistics. */ while ((vd = txg_list_remove(&spa->spa_vdev_txg_list, TXG_CLEAN(txg))) != NULL) vdev_sync_done(vd, txg); metaslab_class_evict_old(spa->spa_normal_class, txg); metaslab_class_evict_old(spa->spa_log_class, txg); spa_sync_close_syncing_log_sm(spa); spa_update_dspace(spa); if (spa_get_autotrim(spa) == SPA_AUTOTRIM_ON) vdev_autotrim_kick(spa); /* * It had better be the case that we didn't dirty anything * since vdev_config_sync(). */ ASSERT(txg_list_empty(&dp->dp_dirty_datasets, txg)); ASSERT(txg_list_empty(&dp->dp_dirty_dirs, txg)); ASSERT(txg_list_empty(&spa->spa_vdev_txg_list, txg)); while (zfs_pause_spa_sync) delay(1); spa->spa_sync_pass = 0; /* * Update the last synced uberblock here. We want to do this at * the end of spa_sync() so that consumers of spa_last_synced_txg() * will be guaranteed that all the processing associated with * that txg has been completed. */ spa->spa_ubsync = spa->spa_uberblock; spa_config_exit(spa, SCL_CONFIG, FTAG); spa_handle_ignored_writes(spa); /* * If any async tasks have been requested, kick them off. */ spa_async_dispatch(spa); } /* * Sync all pools. We don't want to hold the namespace lock across these * operations, so we take a reference on the spa_t and drop the lock during the * sync. */ void spa_sync_allpools(void) { spa_t *spa = NULL; mutex_enter(&spa_namespace_lock); while ((spa = spa_next(spa)) != NULL) { if (spa_state(spa) != POOL_STATE_ACTIVE || !spa_writeable(spa) || spa_suspended(spa)) continue; spa_open_ref(spa, FTAG); mutex_exit(&spa_namespace_lock); txg_wait_synced(spa_get_dsl(spa), 0); mutex_enter(&spa_namespace_lock); spa_close(spa, FTAG); } mutex_exit(&spa_namespace_lock); } /* * ========================================================================== * Miscellaneous routines * ========================================================================== */ /* * Remove all pools in the system. */ void spa_evict_all(void) { spa_t *spa; /* * Remove all cached state. All pools should be closed now, * so every spa in the AVL tree should be unreferenced. */ mutex_enter(&spa_namespace_lock); while ((spa = spa_next(NULL)) != NULL) { /* * Stop async tasks. The async thread may need to detach * a device that's been replaced, which requires grabbing * spa_namespace_lock, so we must drop it here. */ spa_open_ref(spa, FTAG); mutex_exit(&spa_namespace_lock); spa_async_suspend(spa); mutex_enter(&spa_namespace_lock); spa_close(spa, FTAG); if (spa->spa_state != POOL_STATE_UNINITIALIZED) { spa_unload(spa); spa_deactivate(spa); } spa_remove(spa); } mutex_exit(&spa_namespace_lock); } vdev_t * spa_lookup_by_guid(spa_t *spa, uint64_t guid, boolean_t aux) { vdev_t *vd; int i; if ((vd = vdev_lookup_by_guid(spa->spa_root_vdev, guid)) != NULL) return (vd); if (aux) { for (i = 0; i < spa->spa_l2cache.sav_count; i++) { vd = spa->spa_l2cache.sav_vdevs[i]; if (vd->vdev_guid == guid) return (vd); } for (i = 0; i < spa->spa_spares.sav_count; i++) { vd = spa->spa_spares.sav_vdevs[i]; if (vd->vdev_guid == guid) return (vd); } } return (NULL); } void spa_upgrade(spa_t *spa, uint64_t version) { ASSERT(spa_writeable(spa)); spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); /* * This should only be called for a non-faulted pool, and since a * future version would result in an unopenable pool, this shouldn't be * possible. */ ASSERT(SPA_VERSION_IS_SUPPORTED(spa->spa_uberblock.ub_version)); ASSERT3U(version, >=, spa->spa_uberblock.ub_version); spa->spa_uberblock.ub_version = version; vdev_config_dirty(spa->spa_root_vdev); spa_config_exit(spa, SCL_ALL, FTAG); txg_wait_synced(spa_get_dsl(spa), 0); } static boolean_t spa_has_aux_vdev(spa_t *spa, uint64_t guid, spa_aux_vdev_t *sav) { (void) spa; int i; uint64_t vdev_guid; for (i = 0; i < sav->sav_count; i++) if (sav->sav_vdevs[i]->vdev_guid == guid) return (B_TRUE); for (i = 0; i < sav->sav_npending; i++) { if (nvlist_lookup_uint64(sav->sav_pending[i], ZPOOL_CONFIG_GUID, &vdev_guid) == 0 && vdev_guid == guid) return (B_TRUE); } return (B_FALSE); } boolean_t spa_has_l2cache(spa_t *spa, uint64_t guid) { return (spa_has_aux_vdev(spa, guid, &spa->spa_l2cache)); } boolean_t spa_has_spare(spa_t *spa, uint64_t guid) { return (spa_has_aux_vdev(spa, guid, &spa->spa_spares)); } /* * Check if a pool has an active shared spare device. * Note: reference count of an active spare is 2, as a spare and as a replace */ static boolean_t spa_has_active_shared_spare(spa_t *spa) { int i, refcnt; uint64_t pool; spa_aux_vdev_t *sav = &spa->spa_spares; for (i = 0; i < sav->sav_count; i++) { if (spa_spare_exists(sav->sav_vdevs[i]->vdev_guid, &pool, &refcnt) && pool != 0ULL && pool == spa_guid(spa) && refcnt > 2) return (B_TRUE); } return (B_FALSE); } uint64_t spa_total_metaslabs(spa_t *spa) { vdev_t *rvd = spa->spa_root_vdev; uint64_t m = 0; for (uint64_t c = 0; c < rvd->vdev_children; c++) { vdev_t *vd = rvd->vdev_child[c]; if (!vdev_is_concrete(vd)) continue; m += vd->vdev_ms_count; } return (m); } /* * Notify any waiting threads that some activity has switched from being in- * progress to not-in-progress so that the thread can wake up and determine * whether it is finished waiting. */ void spa_notify_waiters(spa_t *spa) { /* * Acquiring spa_activities_lock here prevents the cv_broadcast from * happening between the waiting thread's check and cv_wait. */ mutex_enter(&spa->spa_activities_lock); cv_broadcast(&spa->spa_activities_cv); mutex_exit(&spa->spa_activities_lock); } /* * Notify any waiting threads that the pool is exporting, and then block until * they are finished using the spa_t. */ void spa_wake_waiters(spa_t *spa) { mutex_enter(&spa->spa_activities_lock); spa->spa_waiters_cancel = B_TRUE; cv_broadcast(&spa->spa_activities_cv); while (spa->spa_waiters != 0) cv_wait(&spa->spa_waiters_cv, &spa->spa_activities_lock); spa->spa_waiters_cancel = B_FALSE; mutex_exit(&spa->spa_activities_lock); } /* Whether the vdev or any of its descendants are being initialized/trimmed. */ static boolean_t spa_vdev_activity_in_progress_impl(vdev_t *vd, zpool_wait_activity_t activity) { spa_t *spa = vd->vdev_spa; ASSERT(spa_config_held(spa, SCL_CONFIG | SCL_STATE, RW_READER)); ASSERT(MUTEX_HELD(&spa->spa_activities_lock)); ASSERT(activity == ZPOOL_WAIT_INITIALIZE || activity == ZPOOL_WAIT_TRIM); kmutex_t *lock = activity == ZPOOL_WAIT_INITIALIZE ? &vd->vdev_initialize_lock : &vd->vdev_trim_lock; mutex_exit(&spa->spa_activities_lock); mutex_enter(lock); mutex_enter(&spa->spa_activities_lock); boolean_t in_progress = (activity == ZPOOL_WAIT_INITIALIZE) ? (vd->vdev_initialize_state == VDEV_INITIALIZE_ACTIVE) : (vd->vdev_trim_state == VDEV_TRIM_ACTIVE); mutex_exit(lock); if (in_progress) return (B_TRUE); for (int i = 0; i < vd->vdev_children; i++) { if (spa_vdev_activity_in_progress_impl(vd->vdev_child[i], activity)) return (B_TRUE); } return (B_FALSE); } /* * If use_guid is true, this checks whether the vdev specified by guid is * being initialized/trimmed. Otherwise, it checks whether any vdev in the pool * is being initialized/trimmed. The caller must hold the config lock and * spa_activities_lock. */ static int spa_vdev_activity_in_progress(spa_t *spa, boolean_t use_guid, uint64_t guid, zpool_wait_activity_t activity, boolean_t *in_progress) { mutex_exit(&spa->spa_activities_lock); spa_config_enter(spa, SCL_CONFIG | SCL_STATE, FTAG, RW_READER); mutex_enter(&spa->spa_activities_lock); vdev_t *vd; if (use_guid) { vd = spa_lookup_by_guid(spa, guid, B_FALSE); if (vd == NULL || !vd->vdev_ops->vdev_op_leaf) { spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG); return (EINVAL); } } else { vd = spa->spa_root_vdev; } *in_progress = spa_vdev_activity_in_progress_impl(vd, activity); spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG); return (0); } /* * Locking for waiting threads * --------------------------- * * Waiting threads need a way to check whether a given activity is in progress, * and then, if it is, wait for it to complete. Each activity will have some * in-memory representation of the relevant on-disk state which can be used to * determine whether or not the activity is in progress. The in-memory state and * the locking used to protect it will be different for each activity, and may * not be suitable for use with a cvar (e.g., some state is protected by the * config lock). To allow waiting threads to wait without any races, another * lock, spa_activities_lock, is used. * * When the state is checked, both the activity-specific lock (if there is one) * and spa_activities_lock are held. In some cases, the activity-specific lock * is acquired explicitly (e.g. the config lock). In others, the locking is * internal to some check (e.g. bpobj_is_empty). After checking, the waiting * thread releases the activity-specific lock and, if the activity is in * progress, then cv_waits using spa_activities_lock. * * The waiting thread is woken when another thread, one completing some * activity, updates the state of the activity and then calls * spa_notify_waiters, which will cv_broadcast. This 'completing' thread only * needs to hold its activity-specific lock when updating the state, and this * lock can (but doesn't have to) be dropped before calling spa_notify_waiters. * * Because spa_notify_waiters acquires spa_activities_lock before broadcasting, * and because it is held when the waiting thread checks the state of the * activity, it can never be the case that the completing thread both updates * the activity state and cv_broadcasts in between the waiting thread's check * and cv_wait. Thus, a waiting thread can never miss a wakeup. * * In order to prevent deadlock, when the waiting thread does its check, in some * cases it will temporarily drop spa_activities_lock in order to acquire the * activity-specific lock. The order in which spa_activities_lock and the * activity specific lock are acquired in the waiting thread is determined by * the order in which they are acquired in the completing thread; if the * completing thread calls spa_notify_waiters with the activity-specific lock * held, then the waiting thread must also acquire the activity-specific lock * first. */ static int spa_activity_in_progress(spa_t *spa, zpool_wait_activity_t activity, boolean_t use_tag, uint64_t tag, boolean_t *in_progress) { int error = 0; ASSERT(MUTEX_HELD(&spa->spa_activities_lock)); switch (activity) { case ZPOOL_WAIT_CKPT_DISCARD: *in_progress = (spa_feature_is_active(spa, SPA_FEATURE_POOL_CHECKPOINT) && zap_contains(spa_meta_objset(spa), DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_ZPOOL_CHECKPOINT) == ENOENT); break; case ZPOOL_WAIT_FREE: *in_progress = ((spa_version(spa) >= SPA_VERSION_DEADLISTS && !bpobj_is_empty(&spa->spa_dsl_pool->dp_free_bpobj)) || spa_feature_is_active(spa, SPA_FEATURE_ASYNC_DESTROY) || spa_livelist_delete_check(spa)); break; case ZPOOL_WAIT_INITIALIZE: case ZPOOL_WAIT_TRIM: error = spa_vdev_activity_in_progress(spa, use_tag, tag, activity, in_progress); break; case ZPOOL_WAIT_REPLACE: mutex_exit(&spa->spa_activities_lock); spa_config_enter(spa, SCL_CONFIG | SCL_STATE, FTAG, RW_READER); mutex_enter(&spa->spa_activities_lock); *in_progress = vdev_replace_in_progress(spa->spa_root_vdev); spa_config_exit(spa, SCL_CONFIG | SCL_STATE, FTAG); break; case ZPOOL_WAIT_REMOVE: *in_progress = (spa->spa_removing_phys.sr_state == DSS_SCANNING); break; case ZPOOL_WAIT_RESILVER: if ((*in_progress = vdev_rebuild_active(spa->spa_root_vdev))) break; zfs_fallthrough; case ZPOOL_WAIT_SCRUB: { boolean_t scanning, paused, is_scrub; dsl_scan_t *scn = spa->spa_dsl_pool->dp_scan; is_scrub = (scn->scn_phys.scn_func == POOL_SCAN_SCRUB); scanning = (scn->scn_phys.scn_state == DSS_SCANNING); paused = dsl_scan_is_paused_scrub(scn); *in_progress = (scanning && !paused && is_scrub == (activity == ZPOOL_WAIT_SCRUB)); break; } default: panic("unrecognized value for activity %d", activity); } return (error); } static int spa_wait_common(const char *pool, zpool_wait_activity_t activity, boolean_t use_tag, uint64_t tag, boolean_t *waited) { /* * The tag is used to distinguish between instances of an activity. * 'initialize' and 'trim' are the only activities that we use this for. * The other activities can only have a single instance in progress in a * pool at one time, making the tag unnecessary. * * There can be multiple devices being replaced at once, but since they * all finish once resilvering finishes, we don't bother keeping track * of them individually, we just wait for them all to finish. */ if (use_tag && activity != ZPOOL_WAIT_INITIALIZE && activity != ZPOOL_WAIT_TRIM) return (EINVAL); if (activity < 0 || activity >= ZPOOL_WAIT_NUM_ACTIVITIES) return (EINVAL); spa_t *spa; int error = spa_open(pool, &spa, FTAG); if (error != 0) return (error); /* * Increment the spa's waiter count so that we can call spa_close and * still ensure that the spa_t doesn't get freed before this thread is * finished with it when the pool is exported. We want to call spa_close * before we start waiting because otherwise the additional ref would * prevent the pool from being exported or destroyed throughout the * potentially long wait. */ mutex_enter(&spa->spa_activities_lock); spa->spa_waiters++; spa_close(spa, FTAG); *waited = B_FALSE; for (;;) { boolean_t in_progress; error = spa_activity_in_progress(spa, activity, use_tag, tag, &in_progress); if (error || !in_progress || spa->spa_waiters_cancel) break; *waited = B_TRUE; if (cv_wait_sig(&spa->spa_activities_cv, &spa->spa_activities_lock) == 0) { error = EINTR; break; } } spa->spa_waiters--; cv_signal(&spa->spa_waiters_cv); mutex_exit(&spa->spa_activities_lock); return (error); } /* * Wait for a particular instance of the specified activity to complete, where * the instance is identified by 'tag' */ int spa_wait_tag(const char *pool, zpool_wait_activity_t activity, uint64_t tag, boolean_t *waited) { return (spa_wait_common(pool, activity, B_TRUE, tag, waited)); } /* * Wait for all instances of the specified activity complete */ int spa_wait(const char *pool, zpool_wait_activity_t activity, boolean_t *waited) { return (spa_wait_common(pool, activity, B_FALSE, 0, waited)); } sysevent_t * spa_event_create(spa_t *spa, vdev_t *vd, nvlist_t *hist_nvl, const char *name) { sysevent_t *ev = NULL; #ifdef _KERNEL nvlist_t *resource; resource = zfs_event_create(spa, vd, FM_SYSEVENT_CLASS, name, hist_nvl); if (resource) { ev = kmem_alloc(sizeof (sysevent_t), KM_SLEEP); ev->resource = resource; } #else (void) spa, (void) vd, (void) hist_nvl, (void) name; #endif return (ev); } void spa_event_post(sysevent_t *ev) { #ifdef _KERNEL if (ev) { zfs_zevent_post(ev->resource, NULL, zfs_zevent_post_cb); kmem_free(ev, sizeof (*ev)); } #else (void) ev; #endif } /* * Post a zevent corresponding to the given sysevent. The 'name' must be one * of the event definitions in sys/sysevent/eventdefs.h. The payload will be * filled in from the spa and (optionally) the vdev. This doesn't do anything * in the userland libzpool, as we don't want consumers to misinterpret ztest * or zdb as real changes. */ void spa_event_notify(spa_t *spa, vdev_t *vd, nvlist_t *hist_nvl, const char *name) { spa_event_post(spa_event_create(spa, vd, hist_nvl, name)); } /* state manipulation functions */ EXPORT_SYMBOL(spa_open); EXPORT_SYMBOL(spa_open_rewind); EXPORT_SYMBOL(spa_get_stats); EXPORT_SYMBOL(spa_create); EXPORT_SYMBOL(spa_import); EXPORT_SYMBOL(spa_tryimport); EXPORT_SYMBOL(spa_destroy); EXPORT_SYMBOL(spa_export); EXPORT_SYMBOL(spa_reset); EXPORT_SYMBOL(spa_async_request); EXPORT_SYMBOL(spa_async_suspend); EXPORT_SYMBOL(spa_async_resume); EXPORT_SYMBOL(spa_inject_addref); EXPORT_SYMBOL(spa_inject_delref); EXPORT_SYMBOL(spa_scan_stat_init); EXPORT_SYMBOL(spa_scan_get_stats); /* device manipulation */ EXPORT_SYMBOL(spa_vdev_add); EXPORT_SYMBOL(spa_vdev_attach); EXPORT_SYMBOL(spa_vdev_detach); EXPORT_SYMBOL(spa_vdev_setpath); EXPORT_SYMBOL(spa_vdev_setfru); EXPORT_SYMBOL(spa_vdev_split_mirror); /* spare statech is global across all pools) */ EXPORT_SYMBOL(spa_spare_add); EXPORT_SYMBOL(spa_spare_remove); EXPORT_SYMBOL(spa_spare_exists); EXPORT_SYMBOL(spa_spare_activate); /* L2ARC statech is global across all pools) */ EXPORT_SYMBOL(spa_l2cache_add); EXPORT_SYMBOL(spa_l2cache_remove); EXPORT_SYMBOL(spa_l2cache_exists); EXPORT_SYMBOL(spa_l2cache_activate); EXPORT_SYMBOL(spa_l2cache_drop); /* scanning */ EXPORT_SYMBOL(spa_scan); EXPORT_SYMBOL(spa_scan_stop); /* spa syncing */ EXPORT_SYMBOL(spa_sync); /* only for DMU use */ EXPORT_SYMBOL(spa_sync_allpools); /* properties */ EXPORT_SYMBOL(spa_prop_set); EXPORT_SYMBOL(spa_prop_get); EXPORT_SYMBOL(spa_prop_clear_bootfs); /* asynchronous event notification */ EXPORT_SYMBOL(spa_event_notify); /* BEGIN CSTYLED */ ZFS_MODULE_PARAM(zfs_spa, spa_, load_verify_shift, UINT, ZMOD_RW, "log2 fraction of arc that can be used by inflight I/Os when " "verifying pool during import"); /* END CSTYLED */ ZFS_MODULE_PARAM(zfs_spa, spa_, load_verify_metadata, INT, ZMOD_RW, "Set to traverse metadata on pool import"); ZFS_MODULE_PARAM(zfs_spa, spa_, load_verify_data, INT, ZMOD_RW, "Set to traverse data on pool import"); ZFS_MODULE_PARAM(zfs_spa, spa_, load_print_vdev_tree, INT, ZMOD_RW, "Print vdev tree to zfs_dbgmsg during pool import"); ZFS_MODULE_PARAM(zfs_zio, zio_, taskq_batch_pct, UINT, ZMOD_RD, "Percentage of CPUs to run an IO worker thread"); ZFS_MODULE_PARAM(zfs_zio, zio_, taskq_batch_tpq, UINT, ZMOD_RD, "Number of threads per IO worker taskqueue"); /* BEGIN CSTYLED */ ZFS_MODULE_PARAM(zfs, zfs_, max_missing_tvds, U64, ZMOD_RW, "Allow importing pool with up to this number of missing top-level " "vdevs (in read-only mode)"); /* END CSTYLED */ ZFS_MODULE_PARAM(zfs_livelist_condense, zfs_livelist_condense_, zthr_pause, INT, ZMOD_RW, "Set the livelist condense zthr to pause"); ZFS_MODULE_PARAM(zfs_livelist_condense, zfs_livelist_condense_, sync_pause, INT, ZMOD_RW, "Set the livelist condense synctask to pause"); /* BEGIN CSTYLED */ ZFS_MODULE_PARAM(zfs_livelist_condense, zfs_livelist_condense_, sync_cancel, INT, ZMOD_RW, "Whether livelist condensing was canceled in the synctask"); ZFS_MODULE_PARAM(zfs_livelist_condense, zfs_livelist_condense_, zthr_cancel, INT, ZMOD_RW, "Whether livelist condensing was canceled in the zthr function"); ZFS_MODULE_PARAM(zfs_livelist_condense, zfs_livelist_condense_, new_alloc, INT, ZMOD_RW, "Whether extra ALLOC blkptrs were added to a livelist entry while it " "was being condensed"); /* END CSTYLED */ diff --git a/sys/contrib/openzfs/module/zfs/spa_errlog.c b/sys/contrib/openzfs/module/zfs/spa_errlog.c index 31719063a227..2e5c22c11490 100644 --- a/sys/contrib/openzfs/module/zfs/spa_errlog.c +++ b/sys/contrib/openzfs/module/zfs/spa_errlog.c @@ -1,1465 +1,1489 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2006, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2013, 2014, Delphix. All rights reserved. * Copyright (c) 2019 Datto Inc. * Copyright (c) 2021, 2022, George Amanakis. All rights reserved. */ /* * Routines to manage the on-disk persistent error log. * * Each pool stores a log of all logical data errors seen during normal * operation. This is actually the union of two distinct logs: the last log, * and the current log. All errors seen are logged to the current log. When a * scrub completes, the current log becomes the last log, the last log is thrown * out, and the current log is reinitialized. This way, if an error is somehow * corrected, a new scrub will show that it no longer exists, and will be * deleted from the log when the scrub completes. * * The log is stored using a ZAP object whose key is a string form of the * zbookmark_phys tuple (objset, object, level, blkid), and whose contents is an * optional 'objset:object' human-readable string describing the data. When an * error is first logged, this string will be empty, indicating that no name is * known. This prevents us from having to issue a potentially large amount of * I/O to discover the object name during an error path. Instead, we do the * calculation when the data is requested, storing the result so future queries * will be faster. * * If the head_errlog feature is enabled, a different on-disk format is used. * The error log of each head dataset is stored separately in the zap object * and keyed by the head id. This enables listing every dataset affected in * userland. In order to be able to track whether an error block has been * modified or added to snapshots since it was marked as an error, a new tuple * is introduced: zbookmark_err_phys_t. It allows the storage of the birth * transaction group of an error block on-disk. The birth transaction group is * used by check_filesystem() to assess whether this block was freed, * re-written or added to a snapshot since its marking as an error. * * This log is then shipped into an nvlist where the key is the dataset name and * the value is the object name. Userland is then responsible for uniquifying * this list and displaying it to the user. */ #include #include #include #include #include #include #include #include #include #define NAME_MAX_LEN 64 typedef struct clones { uint64_t clone_ds; list_node_t node; } clones_t; /* * spa_upgrade_errlog_limit : A zfs module parameter that controls the number * of on-disk error log entries that will be converted to the new * format when enabling head_errlog. Defaults to 0 which converts * all log entries. */ static uint_t spa_upgrade_errlog_limit = 0; /* * Convert a bookmark to a string. */ static void bookmark_to_name(zbookmark_phys_t *zb, char *buf, size_t len) { (void) snprintf(buf, len, "%llx:%llx:%llx:%llx", (u_longlong_t)zb->zb_objset, (u_longlong_t)zb->zb_object, (u_longlong_t)zb->zb_level, (u_longlong_t)zb->zb_blkid); } /* * Convert an err_phys to a string. */ static void errphys_to_name(zbookmark_err_phys_t *zep, char *buf, size_t len) { (void) snprintf(buf, len, "%llx:%llx:%llx:%llx", (u_longlong_t)zep->zb_object, (u_longlong_t)zep->zb_level, (u_longlong_t)zep->zb_blkid, (u_longlong_t)zep->zb_birth); } /* * Convert a string to a err_phys. */ -static void +void name_to_errphys(char *buf, zbookmark_err_phys_t *zep) { zep->zb_object = zfs_strtonum(buf, &buf); ASSERT(*buf == ':'); zep->zb_level = (int)zfs_strtonum(buf + 1, &buf); ASSERT(*buf == ':'); zep->zb_blkid = zfs_strtonum(buf + 1, &buf); ASSERT(*buf == ':'); zep->zb_birth = zfs_strtonum(buf + 1, &buf); ASSERT(*buf == '\0'); } /* * Convert a string to a bookmark. */ static void name_to_bookmark(char *buf, zbookmark_phys_t *zb) { zb->zb_objset = zfs_strtonum(buf, &buf); ASSERT(*buf == ':'); zb->zb_object = zfs_strtonum(buf + 1, &buf); ASSERT(*buf == ':'); zb->zb_level = (int)zfs_strtonum(buf + 1, &buf); ASSERT(*buf == ':'); zb->zb_blkid = zfs_strtonum(buf + 1, &buf); ASSERT(*buf == '\0'); } -#ifdef _KERNEL -static void +void zep_to_zb(uint64_t dataset, zbookmark_err_phys_t *zep, zbookmark_phys_t *zb) { zb->zb_objset = dataset; zb->zb_object = zep->zb_object; zb->zb_level = zep->zb_level; zb->zb_blkid = zep->zb_blkid; } -#endif static void name_to_object(char *buf, uint64_t *obj) { *obj = zfs_strtonum(buf, &buf); ASSERT(*buf == '\0'); } /* * Retrieve the head filesystem. */ static int get_head_ds(spa_t *spa, uint64_t dsobj, uint64_t *head_ds) { dsl_dataset_t *ds; int error = dsl_dataset_hold_obj_flags(spa->spa_dsl_pool, dsobj, DS_HOLD_FLAG_DECRYPT, FTAG, &ds); if (error != 0) return (error); ASSERT(head_ds); *head_ds = dsl_dir_phys(ds->ds_dir)->dd_head_dataset_obj; dsl_dataset_rele_flags(ds, DS_HOLD_FLAG_DECRYPT, FTAG); return (error); } /* * Log an uncorrectable error to the persistent error log. We add it to the * spa's list of pending errors. The changes are actually synced out to disk * during spa_errlog_sync(). */ void spa_log_error(spa_t *spa, const zbookmark_phys_t *zb, const uint64_t *birth) { spa_error_entry_t search; spa_error_entry_t *new; avl_tree_t *tree; avl_index_t where; /* * If we are trying to import a pool, ignore any errors, as we won't be * writing to the pool any time soon. */ if (spa_load_state(spa) == SPA_LOAD_TRYIMPORT) return; mutex_enter(&spa->spa_errlist_lock); /* * If we have had a request to rotate the log, log it to the next list * instead of the current one. */ if (spa->spa_scrub_active || spa->spa_scrub_finished) tree = &spa->spa_errlist_scrub; else tree = &spa->spa_errlist_last; search.se_bookmark = *zb; if (avl_find(tree, &search, &where) != NULL) { mutex_exit(&spa->spa_errlist_lock); return; } new = kmem_zalloc(sizeof (spa_error_entry_t), KM_SLEEP); new->se_bookmark = *zb; /* * If the head_errlog feature is enabled, store the birth txg now. In * case the file is deleted before spa_errlog_sync() runs, we will not * be able to retrieve the birth txg. */ if (spa_feature_is_enabled(spa, SPA_FEATURE_HEAD_ERRLOG)) { new->se_zep.zb_object = zb->zb_object; new->se_zep.zb_level = zb->zb_level; new->se_zep.zb_blkid = zb->zb_blkid; /* * birth may end up being NULL, e.g. in zio_done(). We * will handle this in process_error_block(). */ if (birth != NULL) new->se_zep.zb_birth = *birth; } avl_insert(tree, new, where); mutex_exit(&spa->spa_errlist_lock); } -#ifdef _KERNEL -static int +int find_birth_txg(dsl_dataset_t *ds, zbookmark_err_phys_t *zep, uint64_t *birth_txg) { objset_t *os; int error = dmu_objset_from_ds(ds, &os); if (error != 0) return (error); dnode_t *dn; blkptr_t bp; error = dnode_hold(os, zep->zb_object, FTAG, &dn); if (error != 0) return (error); rw_enter(&dn->dn_struct_rwlock, RW_READER); error = dbuf_dnode_findbp(dn, zep->zb_level, zep->zb_blkid, &bp, NULL, NULL); if (error == 0 && BP_IS_HOLE(&bp)) error = SET_ERROR(ENOENT); *birth_txg = bp.blk_birth; rw_exit(&dn->dn_struct_rwlock); dnode_rele(dn, FTAG); return (error); } +/* + * This function finds the oldest affected filesystem containing an error + * block. + */ +int +find_top_affected_fs(spa_t *spa, uint64_t head_ds, zbookmark_err_phys_t *zep, + uint64_t *top_affected_fs) +{ + uint64_t oldest_dsobj; + int error = dsl_dataset_oldest_snapshot(spa, head_ds, zep->zb_birth, + &oldest_dsobj); + if (error != 0) + return (error); + + dsl_dataset_t *ds; + error = dsl_dataset_hold_obj_flags(spa->spa_dsl_pool, oldest_dsobj, + DS_HOLD_FLAG_DECRYPT, FTAG, &ds); + if (error != 0) + return (error); + + *top_affected_fs = + dsl_dir_phys(ds->ds_dir)->dd_head_dataset_obj; + dsl_dataset_rele_flags(ds, DS_HOLD_FLAG_DECRYPT, FTAG); + return (0); +} + + +#ifdef _KERNEL /* * Copy the bookmark to the end of the user-space buffer which starts at * uaddr and has *count unused entries, and decrement *count by 1. */ static int copyout_entry(const zbookmark_phys_t *zb, void *uaddr, uint64_t *count) { if (*count == 0) return (SET_ERROR(ENOMEM)); *count -= 1; if (copyout(zb, (char *)uaddr + (*count) * sizeof (zbookmark_phys_t), sizeof (zbookmark_phys_t)) != 0) return (SET_ERROR(EFAULT)); return (0); } /* * Each time the error block is referenced by a snapshot or clone, add a * zbookmark_phys_t entry to the userspace array at uaddr. The array is * filled from the back and the in-out parameter *count is modified to be the - * number of unused entries at the beginning of the array. + * number of unused entries at the beginning of the array. The function + * scrub_filesystem() is modelled after this one. */ static int check_filesystem(spa_t *spa, uint64_t head_ds, zbookmark_err_phys_t *zep, void *uaddr, uint64_t *count, list_t *clones_list) { dsl_dataset_t *ds; dsl_pool_t *dp = spa->spa_dsl_pool; int error = dsl_dataset_hold_obj_flags(dp, head_ds, DS_HOLD_FLAG_DECRYPT, FTAG, &ds); if (error != 0) return (error); uint64_t latest_txg; uint64_t txg_to_consider = spa->spa_syncing_txg; boolean_t check_snapshot = B_TRUE; error = find_birth_txg(ds, zep, &latest_txg); /* * If find_birth_txg() errors out otherwise, let txg_to_consider be * equal to the spa's syncing txg: if check_filesystem() errors out * then affected snapshots or clones will not be checked. */ if (error == 0 && zep->zb_birth == latest_txg) { /* Block neither free nor rewritten. */ zbookmark_phys_t zb; zep_to_zb(head_ds, zep, &zb); error = copyout_entry(&zb, uaddr, count); if (error != 0) { dsl_dataset_rele_flags(ds, DS_HOLD_FLAG_DECRYPT, FTAG); return (error); } check_snapshot = B_FALSE; } else if (error == 0) { txg_to_consider = latest_txg; } /* * Retrieve the number of snapshots if the dataset is not a snapshot. */ uint64_t snap_count = 0; if (dsl_dataset_phys(ds)->ds_snapnames_zapobj != 0) { error = zap_count(spa->spa_meta_objset, dsl_dataset_phys(ds)->ds_snapnames_zapobj, &snap_count); if (error != 0) { dsl_dataset_rele_flags(ds, DS_HOLD_FLAG_DECRYPT, FTAG); return (error); } } if (snap_count == 0) { /* Filesystem without snapshots. */ dsl_dataset_rele_flags(ds, DS_HOLD_FLAG_DECRYPT, FTAG); return (0); } uint64_t *snap_obj_array = kmem_zalloc(snap_count * sizeof (uint64_t), KM_SLEEP); int aff_snap_count = 0; uint64_t snap_obj = dsl_dataset_phys(ds)->ds_prev_snap_obj; uint64_t snap_obj_txg = dsl_dataset_phys(ds)->ds_prev_snap_txg; uint64_t zap_clone = dsl_dir_phys(ds->ds_dir)->dd_clones; dsl_dataset_rele_flags(ds, DS_HOLD_FLAG_DECRYPT, FTAG); /* Check only snapshots created from this file system. */ while (snap_obj != 0 && zep->zb_birth < snap_obj_txg && snap_obj_txg <= txg_to_consider) { error = dsl_dataset_hold_obj_flags(dp, snap_obj, DS_HOLD_FLAG_DECRYPT, FTAG, &ds); if (error != 0) goto out; if (dsl_dir_phys(ds->ds_dir)->dd_head_dataset_obj != head_ds) { snap_obj = dsl_dataset_phys(ds)->ds_prev_snap_obj; snap_obj_txg = dsl_dataset_phys(ds)->ds_prev_snap_txg; dsl_dataset_rele_flags(ds, DS_HOLD_FLAG_DECRYPT, FTAG); continue; } boolean_t affected = B_TRUE; if (check_snapshot) { uint64_t blk_txg; error = find_birth_txg(ds, zep, &blk_txg); affected = (error == 0 && zep->zb_birth == blk_txg); } /* Report errors in snapshots. */ if (affected) { snap_obj_array[aff_snap_count] = snap_obj; aff_snap_count++; zbookmark_phys_t zb; zep_to_zb(snap_obj, zep, &zb); error = copyout_entry(&zb, uaddr, count); if (error != 0) { dsl_dataset_rele_flags(ds, DS_HOLD_FLAG_DECRYPT, FTAG); goto out; } } snap_obj = dsl_dataset_phys(ds)->ds_prev_snap_obj; snap_obj_txg = dsl_dataset_phys(ds)->ds_prev_snap_txg; dsl_dataset_rele_flags(ds, DS_HOLD_FLAG_DECRYPT, FTAG); } if (zap_clone == 0 || aff_snap_count == 0) return (0); /* Check clones. */ zap_cursor_t *zc; zap_attribute_t *za; zc = kmem_zalloc(sizeof (zap_cursor_t), KM_SLEEP); za = kmem_zalloc(sizeof (zap_attribute_t), KM_SLEEP); for (zap_cursor_init(zc, spa->spa_meta_objset, zap_clone); zap_cursor_retrieve(zc, za) == 0; zap_cursor_advance(zc)) { dsl_dataset_t *clone; error = dsl_dataset_hold_obj_flags(dp, za->za_first_integer, DS_HOLD_FLAG_DECRYPT, FTAG, &clone); if (error != 0) break; /* * Only clones whose origins were affected could also * have affected snapshots. */ boolean_t found = B_FALSE; for (int i = 0; i < snap_count; i++) { if (dsl_dir_phys(clone->ds_dir)->dd_origin_obj == snap_obj_array[i]) found = B_TRUE; } dsl_dataset_rele_flags(clone, DS_HOLD_FLAG_DECRYPT, FTAG); if (!found) continue; clones_t *ct = kmem_zalloc(sizeof (*ct), KM_SLEEP); ct->clone_ds = za->za_first_integer; list_insert_tail(clones_list, ct); } zap_cursor_fini(zc); kmem_free(za, sizeof (*za)); kmem_free(zc, sizeof (*zc)); out: kmem_free(snap_obj_array, sizeof (*snap_obj_array)); return (error); } -static int -find_top_affected_fs(spa_t *spa, uint64_t head_ds, zbookmark_err_phys_t *zep, - uint64_t *top_affected_fs) -{ - uint64_t oldest_dsobj; - int error = dsl_dataset_oldest_snapshot(spa, head_ds, zep->zb_birth, - &oldest_dsobj); - if (error != 0) - return (error); - - dsl_dataset_t *ds; - error = dsl_dataset_hold_obj_flags(spa->spa_dsl_pool, oldest_dsobj, - DS_HOLD_FLAG_DECRYPT, FTAG, &ds); - if (error != 0) - return (error); - - *top_affected_fs = - dsl_dir_phys(ds->ds_dir)->dd_head_dataset_obj; - dsl_dataset_rele_flags(ds, DS_HOLD_FLAG_DECRYPT, FTAG); - return (0); -} - static int process_error_block(spa_t *spa, uint64_t head_ds, zbookmark_err_phys_t *zep, void *uaddr, uint64_t *count) { /* * If zb_birth == 0 or head_ds == 0 it means we failed to retrieve the * birth txg or the head filesystem of the block pointer. This may * happen e.g. when an encrypted filesystem is not mounted or when * the key is not loaded. In this case do not proceed to * check_filesystem(), instead do the accounting here. */ if (zep->zb_birth == 0 || head_ds == 0) { zbookmark_phys_t zb; zep_to_zb(head_ds, zep, &zb); int error = copyout_entry(&zb, uaddr, count); if (error != 0) { return (error); } return (0); } uint64_t top_affected_fs; uint64_t init_count = *count; int error = find_top_affected_fs(spa, head_ds, zep, &top_affected_fs); if (error == 0) { clones_t *ct; list_t clones_list; list_create(&clones_list, sizeof (clones_t), offsetof(clones_t, node)); error = check_filesystem(spa, top_affected_fs, zep, uaddr, count, &clones_list); while ((ct = list_remove_head(&clones_list)) != NULL) { error = check_filesystem(spa, ct->clone_ds, zep, uaddr, count, &clones_list); kmem_free(ct, sizeof (*ct)); if (error) { while (!list_is_empty(&clones_list)) { ct = list_remove_head(&clones_list); kmem_free(ct, sizeof (*ct)); } break; } } list_destroy(&clones_list); } if (error == 0 && init_count == *count) { /* * If we reach this point, no errors have been detected * in the checked filesystems/snapshots. Before returning mark * the error block to be removed from the error lists and logs. */ zbookmark_phys_t zb; zep_to_zb(head_ds, zep, &zb); spa_remove_error(spa, &zb, &zep->zb_birth); } return (error); } #endif +/* Return the number of errors in the error log */ +uint64_t +spa_get_last_errlog_size(spa_t *spa) +{ + uint64_t total = 0, count; + mutex_enter(&spa->spa_errlog_lock); + + if (spa->spa_errlog_last != 0 && + zap_count(spa->spa_meta_objset, spa->spa_errlog_last, + &count) == 0) + total += count; + mutex_exit(&spa->spa_errlog_lock); + return (total); +} + /* * If a healed bookmark matches an entry in the error log we stash it in a tree * so that we can later remove the related log entries in sync context. */ static void spa_add_healed_error(spa_t *spa, uint64_t obj, zbookmark_phys_t *healed_zb, const uint64_t *birth) { char name[NAME_MAX_LEN]; if (obj == 0) return; boolean_t held_list = B_FALSE; boolean_t held_log = B_FALSE; if (!spa_feature_is_enabled(spa, SPA_FEATURE_HEAD_ERRLOG)) { bookmark_to_name(healed_zb, name, sizeof (name)); if (zap_contains(spa->spa_meta_objset, healed_zb->zb_objset, name) == 0) { if (!MUTEX_HELD(&spa->spa_errlog_lock)) { mutex_enter(&spa->spa_errlog_lock); held_log = B_TRUE; } /* * Found an error matching healed zb, add zb to our * tree of healed errors */ avl_tree_t *tree = &spa->spa_errlist_healed; spa_error_entry_t search; spa_error_entry_t *new; avl_index_t where; search.se_bookmark = *healed_zb; if (!MUTEX_HELD(&spa->spa_errlist_lock)) { mutex_enter(&spa->spa_errlist_lock); held_list = B_TRUE; } if (avl_find(tree, &search, &where) != NULL) { if (held_list) mutex_exit(&spa->spa_errlist_lock); if (held_log) mutex_exit(&spa->spa_errlog_lock); return; } new = kmem_zalloc(sizeof (spa_error_entry_t), KM_SLEEP); new->se_bookmark = *healed_zb; avl_insert(tree, new, where); if (held_list) mutex_exit(&spa->spa_errlist_lock); if (held_log) mutex_exit(&spa->spa_errlog_lock); } return; } zbookmark_err_phys_t healed_zep; healed_zep.zb_object = healed_zb->zb_object; healed_zep.zb_level = healed_zb->zb_level; healed_zep.zb_blkid = healed_zb->zb_blkid; if (birth != NULL) healed_zep.zb_birth = *birth; else healed_zep.zb_birth = 0; errphys_to_name(&healed_zep, name, sizeof (name)); zap_cursor_t zc; zap_attribute_t za; for (zap_cursor_init(&zc, spa->spa_meta_objset, spa->spa_errlog_last); zap_cursor_retrieve(&zc, &za) == 0; zap_cursor_advance(&zc)) { if (zap_contains(spa->spa_meta_objset, za.za_first_integer, name) == 0) { if (!MUTEX_HELD(&spa->spa_errlog_lock)) { mutex_enter(&spa->spa_errlog_lock); held_log = B_TRUE; } avl_tree_t *tree = &spa->spa_errlist_healed; spa_error_entry_t search; spa_error_entry_t *new; avl_index_t where; search.se_bookmark = *healed_zb; if (!MUTEX_HELD(&spa->spa_errlist_lock)) { mutex_enter(&spa->spa_errlist_lock); held_list = B_TRUE; } if (avl_find(tree, &search, &where) != NULL) { if (held_list) mutex_exit(&spa->spa_errlist_lock); if (held_log) mutex_exit(&spa->spa_errlog_lock); continue; } new = kmem_zalloc(sizeof (spa_error_entry_t), KM_SLEEP); new->se_bookmark = *healed_zb; new->se_zep = healed_zep; avl_insert(tree, new, where); if (held_list) mutex_exit(&spa->spa_errlist_lock); if (held_log) mutex_exit(&spa->spa_errlog_lock); } } zap_cursor_fini(&zc); } /* * If this error exists in the given tree remove it. */ static void remove_error_from_list(spa_t *spa, avl_tree_t *t, const zbookmark_phys_t *zb) { spa_error_entry_t search, *found; avl_index_t where; mutex_enter(&spa->spa_errlist_lock); search.se_bookmark = *zb; if ((found = avl_find(t, &search, &where)) != NULL) { avl_remove(t, found); kmem_free(found, sizeof (spa_error_entry_t)); } mutex_exit(&spa->spa_errlist_lock); } /* * Removes all of the recv healed errors from both on-disk error logs */ static void spa_remove_healed_errors(spa_t *spa, avl_tree_t *s, avl_tree_t *l, dmu_tx_t *tx) { char name[NAME_MAX_LEN]; spa_error_entry_t *se; void *cookie = NULL; ASSERT(MUTEX_HELD(&spa->spa_errlog_lock)); while ((se = avl_destroy_nodes(&spa->spa_errlist_healed, &cookie)) != NULL) { remove_error_from_list(spa, s, &se->se_bookmark); remove_error_from_list(spa, l, &se->se_bookmark); - kmem_free(se, sizeof (spa_error_entry_t)); if (!spa_feature_is_enabled(spa, SPA_FEATURE_HEAD_ERRLOG)) { bookmark_to_name(&se->se_bookmark, name, sizeof (name)); (void) zap_remove(spa->spa_meta_objset, spa->spa_errlog_last, name, tx); (void) zap_remove(spa->spa_meta_objset, spa->spa_errlog_scrub, name, tx); } else { errphys_to_name(&se->se_zep, name, sizeof (name)); zap_cursor_t zc; zap_attribute_t za; for (zap_cursor_init(&zc, spa->spa_meta_objset, spa->spa_errlog_last); zap_cursor_retrieve(&zc, &za) == 0; zap_cursor_advance(&zc)) { zap_remove(spa->spa_meta_objset, za.za_first_integer, name, tx); } zap_cursor_fini(&zc); for (zap_cursor_init(&zc, spa->spa_meta_objset, spa->spa_errlog_scrub); zap_cursor_retrieve(&zc, &za) == 0; zap_cursor_advance(&zc)) { zap_remove(spa->spa_meta_objset, za.za_first_integer, name, tx); } zap_cursor_fini(&zc); } + kmem_free(se, sizeof (spa_error_entry_t)); } } /* * Stash away healed bookmarks to remove them from the on-disk error logs * later in spa_remove_healed_errors(). */ void spa_remove_error(spa_t *spa, zbookmark_phys_t *zb, const uint64_t *birth) { spa_add_healed_error(spa, spa->spa_errlog_last, zb, birth); spa_add_healed_error(spa, spa->spa_errlog_scrub, zb, birth); } static uint64_t approx_errlog_size_impl(spa_t *spa, uint64_t spa_err_obj) { if (spa_err_obj == 0) return (0); uint64_t total = 0; zap_cursor_t zc; zap_attribute_t za; for (zap_cursor_init(&zc, spa->spa_meta_objset, spa_err_obj); zap_cursor_retrieve(&zc, &za) == 0; zap_cursor_advance(&zc)) { uint64_t count; if (zap_count(spa->spa_meta_objset, za.za_first_integer, &count) == 0) total += count; } zap_cursor_fini(&zc); return (total); } /* * Return the approximate number of errors currently in the error log. This * will be nonzero if there are some errors, but otherwise it may be more * or less than the number of entries returned by spa_get_errlog(). */ uint64_t spa_approx_errlog_size(spa_t *spa) { uint64_t total = 0; if (!spa_feature_is_enabled(spa, SPA_FEATURE_HEAD_ERRLOG)) { mutex_enter(&spa->spa_errlog_lock); uint64_t count; if (spa->spa_errlog_scrub != 0 && zap_count(spa->spa_meta_objset, spa->spa_errlog_scrub, &count) == 0) total += count; if (spa->spa_errlog_last != 0 && !spa->spa_scrub_finished && zap_count(spa->spa_meta_objset, spa->spa_errlog_last, &count) == 0) total += count; mutex_exit(&spa->spa_errlog_lock); } else { mutex_enter(&spa->spa_errlog_lock); total += approx_errlog_size_impl(spa, spa->spa_errlog_last); total += approx_errlog_size_impl(spa, spa->spa_errlog_scrub); mutex_exit(&spa->spa_errlog_lock); } mutex_enter(&spa->spa_errlist_lock); total += avl_numnodes(&spa->spa_errlist_last); total += avl_numnodes(&spa->spa_errlist_scrub); mutex_exit(&spa->spa_errlist_lock); return (total); } /* * This function sweeps through an on-disk error log and stores all bookmarks * as error bookmarks in a new ZAP object. At the end we discard the old one, * and spa_update_errlog() will set the spa's on-disk error log to new ZAP * object. */ static void sync_upgrade_errlog(spa_t *spa, uint64_t spa_err_obj, uint64_t *newobj, dmu_tx_t *tx) { zap_cursor_t zc; zap_attribute_t za; zbookmark_phys_t zb; uint64_t count; *newobj = zap_create(spa->spa_meta_objset, DMU_OT_ERROR_LOG, DMU_OT_NONE, 0, tx); /* * If we cannnot perform the upgrade we should clear the old on-disk * error logs. */ if (zap_count(spa->spa_meta_objset, spa_err_obj, &count) != 0) { VERIFY0(dmu_object_free(spa->spa_meta_objset, spa_err_obj, tx)); return; } for (zap_cursor_init(&zc, spa->spa_meta_objset, spa_err_obj); zap_cursor_retrieve(&zc, &za) == 0; zap_cursor_advance(&zc)) { if (spa_upgrade_errlog_limit != 0 && zc.zc_cd == spa_upgrade_errlog_limit) break; name_to_bookmark(za.za_name, &zb); zbookmark_err_phys_t zep; zep.zb_object = zb.zb_object; zep.zb_level = zb.zb_level; zep.zb_blkid = zb.zb_blkid; zep.zb_birth = 0; /* * In case of an error we should simply continue instead of * returning prematurely. See the next comment. */ uint64_t head_ds; dsl_pool_t *dp = spa->spa_dsl_pool; dsl_dataset_t *ds; objset_t *os; int error = dsl_dataset_hold_obj_flags(dp, zb.zb_objset, DS_HOLD_FLAG_DECRYPT, FTAG, &ds); if (error != 0) continue; head_ds = dsl_dir_phys(ds->ds_dir)->dd_head_dataset_obj; /* * The objset and the dnode are required for getting the block * pointer, which is used to determine if BP_IS_HOLE(). If * getting the objset or the dnode fails, do not create a * zap entry (presuming we know the dataset) as this may create * spurious errors that we cannot ever resolve. If an error is * truly persistent, it should re-appear after a scan. */ if (dmu_objset_from_ds(ds, &os) != 0) { dsl_dataset_rele_flags(ds, DS_HOLD_FLAG_DECRYPT, FTAG); continue; } dnode_t *dn; blkptr_t bp; if (dnode_hold(os, zep.zb_object, FTAG, &dn) != 0) { dsl_dataset_rele_flags(ds, DS_HOLD_FLAG_DECRYPT, FTAG); continue; } rw_enter(&dn->dn_struct_rwlock, RW_READER); error = dbuf_dnode_findbp(dn, zep.zb_level, zep.zb_blkid, &bp, NULL, NULL); if (error == EACCES) error = 0; else if (!error) zep.zb_birth = bp.blk_birth; rw_exit(&dn->dn_struct_rwlock); dnode_rele(dn, FTAG); dsl_dataset_rele_flags(ds, DS_HOLD_FLAG_DECRYPT, FTAG); if (error != 0 || BP_IS_HOLE(&bp)) continue; uint64_t err_obj; error = zap_lookup_int_key(spa->spa_meta_objset, *newobj, head_ds, &err_obj); if (error == ENOENT) { err_obj = zap_create(spa->spa_meta_objset, DMU_OT_ERROR_LOG, DMU_OT_NONE, 0, tx); (void) zap_update_int_key(spa->spa_meta_objset, *newobj, head_ds, err_obj, tx); } char buf[64]; errphys_to_name(&zep, buf, sizeof (buf)); const char *name = ""; (void) zap_update(spa->spa_meta_objset, err_obj, buf, 1, strlen(name) + 1, name, tx); } zap_cursor_fini(&zc); VERIFY0(dmu_object_free(spa->spa_meta_objset, spa_err_obj, tx)); } void spa_upgrade_errlog(spa_t *spa, dmu_tx_t *tx) { uint64_t newobj = 0; mutex_enter(&spa->spa_errlog_lock); if (spa->spa_errlog_last != 0) { sync_upgrade_errlog(spa, spa->spa_errlog_last, &newobj, tx); spa->spa_errlog_last = newobj; } if (spa->spa_errlog_scrub != 0) { sync_upgrade_errlog(spa, spa->spa_errlog_scrub, &newobj, tx); spa->spa_errlog_scrub = newobj; } mutex_exit(&spa->spa_errlog_lock); } #ifdef _KERNEL /* * If an error block is shared by two datasets it will be counted twice. */ static int process_error_log(spa_t *spa, uint64_t obj, void *uaddr, uint64_t *count) { if (obj == 0) return (0); zap_cursor_t *zc; zap_attribute_t *za; zc = kmem_zalloc(sizeof (zap_cursor_t), KM_SLEEP); za = kmem_zalloc(sizeof (zap_attribute_t), KM_SLEEP); if (!spa_feature_is_enabled(spa, SPA_FEATURE_HEAD_ERRLOG)) { for (zap_cursor_init(zc, spa->spa_meta_objset, obj); zap_cursor_retrieve(zc, za) == 0; zap_cursor_advance(zc)) { if (*count == 0) { zap_cursor_fini(zc); kmem_free(zc, sizeof (*zc)); kmem_free(za, sizeof (*za)); return (SET_ERROR(ENOMEM)); } zbookmark_phys_t zb; name_to_bookmark(za->za_name, &zb); int error = copyout_entry(&zb, uaddr, count); if (error != 0) { zap_cursor_fini(zc); kmem_free(zc, sizeof (*zc)); kmem_free(za, sizeof (*za)); return (error); } } zap_cursor_fini(zc); kmem_free(zc, sizeof (*zc)); kmem_free(za, sizeof (*za)); return (0); } for (zap_cursor_init(zc, spa->spa_meta_objset, obj); zap_cursor_retrieve(zc, za) == 0; zap_cursor_advance(zc)) { zap_cursor_t *head_ds_cursor; zap_attribute_t *head_ds_attr; head_ds_cursor = kmem_zalloc(sizeof (zap_cursor_t), KM_SLEEP); head_ds_attr = kmem_zalloc(sizeof (zap_attribute_t), KM_SLEEP); uint64_t head_ds_err_obj = za->za_first_integer; uint64_t head_ds; name_to_object(za->za_name, &head_ds); for (zap_cursor_init(head_ds_cursor, spa->spa_meta_objset, head_ds_err_obj); zap_cursor_retrieve(head_ds_cursor, head_ds_attr) == 0; zap_cursor_advance(head_ds_cursor)) { zbookmark_err_phys_t head_ds_block; name_to_errphys(head_ds_attr->za_name, &head_ds_block); int error = process_error_block(spa, head_ds, &head_ds_block, uaddr, count); if (error != 0) { zap_cursor_fini(head_ds_cursor); kmem_free(head_ds_cursor, sizeof (*head_ds_cursor)); kmem_free(head_ds_attr, sizeof (*head_ds_attr)); zap_cursor_fini(zc); kmem_free(za, sizeof (*za)); kmem_free(zc, sizeof (*zc)); return (error); } } zap_cursor_fini(head_ds_cursor); kmem_free(head_ds_cursor, sizeof (*head_ds_cursor)); kmem_free(head_ds_attr, sizeof (*head_ds_attr)); } zap_cursor_fini(zc); kmem_free(za, sizeof (*za)); kmem_free(zc, sizeof (*zc)); return (0); } static int process_error_list(spa_t *spa, avl_tree_t *list, void *uaddr, uint64_t *count) { spa_error_entry_t *se; if (!spa_feature_is_enabled(spa, SPA_FEATURE_HEAD_ERRLOG)) { for (se = avl_first(list); se != NULL; se = AVL_NEXT(list, se)) { int error = copyout_entry(&se->se_bookmark, uaddr, count); if (error != 0) { return (error); } } return (0); } for (se = avl_first(list); se != NULL; se = AVL_NEXT(list, se)) { uint64_t head_ds = 0; int error = get_head_ds(spa, se->se_bookmark.zb_objset, &head_ds); /* * If get_head_ds() errors out, set the head filesystem * to the filesystem stored in the bookmark of the * error block. */ if (error != 0) head_ds = se->se_bookmark.zb_objset; error = process_error_block(spa, head_ds, &se->se_zep, uaddr, count); if (error != 0) return (error); } return (0); } #endif /* * Copy all known errors to userland as an array of bookmarks. This is * actually a union of the on-disk last log and current log, as well as any * pending error requests. * * Because the act of reading the on-disk log could cause errors to be * generated, we have two separate locks: one for the error log and one for the * in-core error lists. We only need the error list lock to log and error, so * we grab the error log lock while we read the on-disk logs, and only pick up * the error list lock when we are finished. */ int spa_get_errlog(spa_t *spa, void *uaddr, uint64_t *count) { int ret = 0; #ifdef _KERNEL /* * The pool config lock is needed to hold a dataset_t via (among other * places) process_error_list() -> process_error_block()-> * find_top_affected_fs(), and lock ordering requires that we get it * before the spa_errlog_lock. */ dsl_pool_config_enter(spa->spa_dsl_pool, FTAG); mutex_enter(&spa->spa_errlog_lock); ret = process_error_log(spa, spa->spa_errlog_scrub, uaddr, count); if (!ret && !spa->spa_scrub_finished) ret = process_error_log(spa, spa->spa_errlog_last, uaddr, count); mutex_enter(&spa->spa_errlist_lock); if (!ret) ret = process_error_list(spa, &spa->spa_errlist_scrub, uaddr, count); if (!ret) ret = process_error_list(spa, &spa->spa_errlist_last, uaddr, count); mutex_exit(&spa->spa_errlist_lock); mutex_exit(&spa->spa_errlog_lock); dsl_pool_config_exit(spa->spa_dsl_pool, FTAG); #else (void) spa, (void) uaddr, (void) count; #endif return (ret); } /* * Called when a scrub completes. This simply set a bit which tells which AVL * tree to add new errors. spa_errlog_sync() is responsible for actually * syncing the changes to the underlying objects. */ void spa_errlog_rotate(spa_t *spa) { mutex_enter(&spa->spa_errlist_lock); spa->spa_scrub_finished = B_TRUE; mutex_exit(&spa->spa_errlist_lock); } /* * Discard any pending errors from the spa_t. Called when unloading a faulted * pool, as the errors encountered during the open cannot be synced to disk. */ void spa_errlog_drain(spa_t *spa) { spa_error_entry_t *se; void *cookie; mutex_enter(&spa->spa_errlist_lock); cookie = NULL; while ((se = avl_destroy_nodes(&spa->spa_errlist_last, &cookie)) != NULL) kmem_free(se, sizeof (spa_error_entry_t)); cookie = NULL; while ((se = avl_destroy_nodes(&spa->spa_errlist_scrub, &cookie)) != NULL) kmem_free(se, sizeof (spa_error_entry_t)); mutex_exit(&spa->spa_errlist_lock); } /* * Process a list of errors into the current on-disk log. */ void sync_error_list(spa_t *spa, avl_tree_t *t, uint64_t *obj, dmu_tx_t *tx) { spa_error_entry_t *se; char buf[NAME_MAX_LEN]; void *cookie; if (avl_numnodes(t) == 0) return; /* create log if necessary */ if (*obj == 0) *obj = zap_create(spa->spa_meta_objset, DMU_OT_ERROR_LOG, DMU_OT_NONE, 0, tx); /* add errors to the current log */ if (!spa_feature_is_enabled(spa, SPA_FEATURE_HEAD_ERRLOG)) { for (se = avl_first(t); se != NULL; se = AVL_NEXT(t, se)) { bookmark_to_name(&se->se_bookmark, buf, sizeof (buf)); const char *name = se->se_name ? se->se_name : ""; (void) zap_update(spa->spa_meta_objset, *obj, buf, 1, strlen(name) + 1, name, tx); } } else { for (se = avl_first(t); se != NULL; se = AVL_NEXT(t, se)) { zbookmark_err_phys_t zep; zep.zb_object = se->se_zep.zb_object; zep.zb_level = se->se_zep.zb_level; zep.zb_blkid = se->se_zep.zb_blkid; zep.zb_birth = se->se_zep.zb_birth; uint64_t head_ds = 0; int error = get_head_ds(spa, se->se_bookmark.zb_objset, &head_ds); /* * If get_head_ds() errors out, set the head filesystem * to the filesystem stored in the bookmark of the * error block. */ if (error != 0) head_ds = se->se_bookmark.zb_objset; uint64_t err_obj; error = zap_lookup_int_key(spa->spa_meta_objset, *obj, head_ds, &err_obj); if (error == ENOENT) { err_obj = zap_create(spa->spa_meta_objset, DMU_OT_ERROR_LOG, DMU_OT_NONE, 0, tx); (void) zap_update_int_key(spa->spa_meta_objset, *obj, head_ds, err_obj, tx); } errphys_to_name(&zep, buf, sizeof (buf)); const char *name = se->se_name ? se->se_name : ""; (void) zap_update(spa->spa_meta_objset, err_obj, buf, 1, strlen(name) + 1, name, tx); } } /* purge the error list */ cookie = NULL; while ((se = avl_destroy_nodes(t, &cookie)) != NULL) kmem_free(se, sizeof (spa_error_entry_t)); } static void delete_errlog(spa_t *spa, uint64_t spa_err_obj, dmu_tx_t *tx) { if (spa_feature_is_enabled(spa, SPA_FEATURE_HEAD_ERRLOG)) { zap_cursor_t zc; zap_attribute_t za; for (zap_cursor_init(&zc, spa->spa_meta_objset, spa_err_obj); zap_cursor_retrieve(&zc, &za) == 0; zap_cursor_advance(&zc)) { VERIFY0(dmu_object_free(spa->spa_meta_objset, za.za_first_integer, tx)); } zap_cursor_fini(&zc); } VERIFY0(dmu_object_free(spa->spa_meta_objset, spa_err_obj, tx)); } /* * Sync the error log out to disk. This is a little tricky because the act of * writing the error log requires the spa_errlist_lock. So, we need to lock the * error lists, take a copy of the lists, and then reinitialize them. Then, we * drop the error list lock and take the error log lock, at which point we * do the errlog processing. Then, if we encounter an I/O error during this * process, we can successfully add the error to the list. Note that this will * result in the perpetual recycling of errors, but it is an unlikely situation * and not a performance critical operation. */ void spa_errlog_sync(spa_t *spa, uint64_t txg) { dmu_tx_t *tx; avl_tree_t scrub, last; int scrub_finished; mutex_enter(&spa->spa_errlist_lock); /* * Bail out early under normal circumstances. */ if (avl_numnodes(&spa->spa_errlist_scrub) == 0 && avl_numnodes(&spa->spa_errlist_last) == 0 && avl_numnodes(&spa->spa_errlist_healed) == 0 && !spa->spa_scrub_finished) { mutex_exit(&spa->spa_errlist_lock); return; } spa_get_errlists(spa, &last, &scrub); scrub_finished = spa->spa_scrub_finished; spa->spa_scrub_finished = B_FALSE; mutex_exit(&spa->spa_errlist_lock); /* * The pool config lock is needed to hold a dataset_t via * sync_error_list() -> get_head_ds(), and lock ordering * requires that we get it before the spa_errlog_lock. */ dsl_pool_config_enter(spa->spa_dsl_pool, FTAG); mutex_enter(&spa->spa_errlog_lock); tx = dmu_tx_create_assigned(spa->spa_dsl_pool, txg); /* * Remove healed errors from errors. */ spa_remove_healed_errors(spa, &last, &scrub, tx); /* * Sync out the current list of errors. */ sync_error_list(spa, &last, &spa->spa_errlog_last, tx); /* * Rotate the log if necessary. */ if (scrub_finished) { if (spa->spa_errlog_last != 0) delete_errlog(spa, spa->spa_errlog_last, tx); spa->spa_errlog_last = spa->spa_errlog_scrub; spa->spa_errlog_scrub = 0; sync_error_list(spa, &scrub, &spa->spa_errlog_last, tx); } /* * Sync out any pending scrub errors. */ sync_error_list(spa, &scrub, &spa->spa_errlog_scrub, tx); /* * Update the MOS to reflect the new values. */ (void) zap_update(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_ERRLOG_LAST, sizeof (uint64_t), 1, &spa->spa_errlog_last, tx); (void) zap_update(spa->spa_meta_objset, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_ERRLOG_SCRUB, sizeof (uint64_t), 1, &spa->spa_errlog_scrub, tx); dmu_tx_commit(tx); mutex_exit(&spa->spa_errlog_lock); dsl_pool_config_exit(spa->spa_dsl_pool, FTAG); } static void delete_dataset_errlog(spa_t *spa, uint64_t spa_err_obj, uint64_t ds, dmu_tx_t *tx) { if (spa_err_obj == 0) return; zap_cursor_t zc; zap_attribute_t za; for (zap_cursor_init(&zc, spa->spa_meta_objset, spa_err_obj); zap_cursor_retrieve(&zc, &za) == 0; zap_cursor_advance(&zc)) { uint64_t head_ds; name_to_object(za.za_name, &head_ds); if (head_ds == ds) { (void) zap_remove(spa->spa_meta_objset, spa_err_obj, za.za_name, tx); VERIFY0(dmu_object_free(spa->spa_meta_objset, za.za_first_integer, tx)); break; } } zap_cursor_fini(&zc); } void spa_delete_dataset_errlog(spa_t *spa, uint64_t ds, dmu_tx_t *tx) { mutex_enter(&spa->spa_errlog_lock); delete_dataset_errlog(spa, spa->spa_errlog_scrub, ds, tx); delete_dataset_errlog(spa, spa->spa_errlog_last, ds, tx); mutex_exit(&spa->spa_errlog_lock); } static int find_txg_ancestor_snapshot(spa_t *spa, uint64_t new_head, uint64_t old_head, uint64_t *txg) { dsl_dataset_t *ds; dsl_pool_t *dp = spa->spa_dsl_pool; int error = dsl_dataset_hold_obj_flags(dp, old_head, DS_HOLD_FLAG_DECRYPT, FTAG, &ds); if (error != 0) return (error); uint64_t prev_obj = dsl_dataset_phys(ds)->ds_prev_snap_obj; uint64_t prev_obj_txg = dsl_dataset_phys(ds)->ds_prev_snap_txg; while (prev_obj != 0) { dsl_dataset_rele_flags(ds, DS_HOLD_FLAG_DECRYPT, FTAG); if ((error = dsl_dataset_hold_obj_flags(dp, prev_obj, DS_HOLD_FLAG_DECRYPT, FTAG, &ds)) == 0 && dsl_dir_phys(ds->ds_dir)->dd_head_dataset_obj == new_head) break; if (error != 0) return (error); prev_obj_txg = dsl_dataset_phys(ds)->ds_prev_snap_txg; prev_obj = dsl_dataset_phys(ds)->ds_prev_snap_obj; } dsl_dataset_rele_flags(ds, DS_HOLD_FLAG_DECRYPT, FTAG); ASSERT(prev_obj != 0); *txg = prev_obj_txg; return (0); } static void swap_errlog(spa_t *spa, uint64_t spa_err_obj, uint64_t new_head, uint64_t old_head, dmu_tx_t *tx) { if (spa_err_obj == 0) return; uint64_t old_head_errlog; int error = zap_lookup_int_key(spa->spa_meta_objset, spa_err_obj, old_head, &old_head_errlog); /* If no error log, then there is nothing to do. */ if (error != 0) return; uint64_t txg; error = find_txg_ancestor_snapshot(spa, new_head, old_head, &txg); if (error != 0) return; /* * Create an error log if the file system being promoted does not * already have one. */ uint64_t new_head_errlog; error = zap_lookup_int_key(spa->spa_meta_objset, spa_err_obj, new_head, &new_head_errlog); if (error != 0) { new_head_errlog = zap_create(spa->spa_meta_objset, DMU_OT_ERROR_LOG, DMU_OT_NONE, 0, tx); (void) zap_update_int_key(spa->spa_meta_objset, spa_err_obj, new_head, new_head_errlog, tx); } zap_cursor_t zc; zap_attribute_t za; zbookmark_err_phys_t err_block; for (zap_cursor_init(&zc, spa->spa_meta_objset, old_head_errlog); zap_cursor_retrieve(&zc, &za) == 0; zap_cursor_advance(&zc)) { const char *name = ""; name_to_errphys(za.za_name, &err_block); if (err_block.zb_birth < txg) { (void) zap_update(spa->spa_meta_objset, new_head_errlog, za.za_name, 1, strlen(name) + 1, name, tx); (void) zap_remove(spa->spa_meta_objset, old_head_errlog, za.za_name, tx); } } zap_cursor_fini(&zc); } void spa_swap_errlog(spa_t *spa, uint64_t new_head_ds, uint64_t old_head_ds, dmu_tx_t *tx) { mutex_enter(&spa->spa_errlog_lock); swap_errlog(spa, spa->spa_errlog_scrub, new_head_ds, old_head_ds, tx); swap_errlog(spa, spa->spa_errlog_last, new_head_ds, old_head_ds, tx); mutex_exit(&spa->spa_errlog_lock); } #if defined(_KERNEL) /* error handling */ EXPORT_SYMBOL(spa_log_error); EXPORT_SYMBOL(spa_approx_errlog_size); +EXPORT_SYMBOL(spa_get_last_errlog_size); EXPORT_SYMBOL(spa_get_errlog); EXPORT_SYMBOL(spa_errlog_rotate); EXPORT_SYMBOL(spa_errlog_drain); EXPORT_SYMBOL(spa_errlog_sync); EXPORT_SYMBOL(spa_get_errlists); EXPORT_SYMBOL(spa_delete_dataset_errlog); EXPORT_SYMBOL(spa_swap_errlog); EXPORT_SYMBOL(sync_error_list); EXPORT_SYMBOL(spa_upgrade_errlog); +EXPORT_SYMBOL(find_top_affected_fs); +EXPORT_SYMBOL(find_birth_txg); +EXPORT_SYMBOL(zep_to_zb); +EXPORT_SYMBOL(name_to_errphys); #endif /* BEGIN CSTYLED */ ZFS_MODULE_PARAM(zfs_spa, spa_, upgrade_errlog_limit, UINT, ZMOD_RW, "Limit the number of errors which will be upgraded to the new " "on-disk error log when enabling head_errlog"); /* END CSTYLED */ diff --git a/sys/contrib/openzfs/module/zfs/spa_misc.c b/sys/contrib/openzfs/module/zfs/spa_misc.c index 54a0eeccf27b..89e1ce7165db 100644 --- a/sys/contrib/openzfs/module/zfs/spa_misc.c +++ b/sys/contrib/openzfs/module/zfs/spa_misc.c @@ -1,2984 +1,3007 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2011, 2019 by Delphix. All rights reserved. * Copyright 2015 Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2014 Spectra Logic Corporation, All rights reserved. * Copyright 2013 Saso Kiselkov. All rights reserved. * Copyright (c) 2017 Datto Inc. * Copyright (c) 2017, Intel Corporation. * Copyright (c) 2019, loli10K . All rights reserved. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "zfs_prop.h" #include #include #include #include /* * SPA locking * * There are three basic locks for managing spa_t structures: * * spa_namespace_lock (global mutex) * * This lock must be acquired to do any of the following: * * - Lookup a spa_t by name * - Add or remove a spa_t from the namespace * - Increase spa_refcount from non-zero * - Check if spa_refcount is zero * - Rename a spa_t * - add/remove/attach/detach devices * - Held for the duration of create/destroy/import/export * * It does not need to handle recursion. A create or destroy may * reference objects (files or zvols) in other pools, but by * definition they must have an existing reference, and will never need * to lookup a spa_t by name. * * spa_refcount (per-spa zfs_refcount_t protected by mutex) * * This reference count keep track of any active users of the spa_t. The * spa_t cannot be destroyed or freed while this is non-zero. Internally, * the refcount is never really 'zero' - opening a pool implicitly keeps * some references in the DMU. Internally we check against spa_minref, but * present the image of a zero/non-zero value to consumers. * * spa_config_lock[] (per-spa array of rwlocks) * * This protects the spa_t from config changes, and must be held in * the following circumstances: * * - RW_READER to perform I/O to the spa * - RW_WRITER to change the vdev config * * The locking order is fairly straightforward: * * spa_namespace_lock -> spa_refcount * * The namespace lock must be acquired to increase the refcount from 0 * or to check if it is zero. * * spa_refcount -> spa_config_lock[] * * There must be at least one valid reference on the spa_t to acquire * the config lock. * * spa_namespace_lock -> spa_config_lock[] * * The namespace lock must always be taken before the config lock. * * * The spa_namespace_lock can be acquired directly and is globally visible. * * The namespace is manipulated using the following functions, all of which * require the spa_namespace_lock to be held. * * spa_lookup() Lookup a spa_t by name. * * spa_add() Create a new spa_t in the namespace. * * spa_remove() Remove a spa_t from the namespace. This also * frees up any memory associated with the spa_t. * * spa_next() Returns the next spa_t in the system, or the * first if NULL is passed. * * spa_evict_all() Shutdown and remove all spa_t structures in * the system. * * spa_guid_exists() Determine whether a pool/device guid exists. * * The spa_refcount is manipulated using the following functions: * * spa_open_ref() Adds a reference to the given spa_t. Must be * called with spa_namespace_lock held if the * refcount is currently zero. * * spa_close() Remove a reference from the spa_t. This will * not free the spa_t or remove it from the * namespace. No locking is required. * * spa_refcount_zero() Returns true if the refcount is currently * zero. Must be called with spa_namespace_lock * held. * * The spa_config_lock[] is an array of rwlocks, ordered as follows: * SCL_CONFIG > SCL_STATE > SCL_ALLOC > SCL_ZIO > SCL_FREE > SCL_VDEV. * spa_config_lock[] is manipulated with spa_config_{enter,exit,held}(). * * To read the configuration, it suffices to hold one of these locks as reader. * To modify the configuration, you must hold all locks as writer. To modify * vdev state without altering the vdev tree's topology (e.g. online/offline), * you must hold SCL_STATE and SCL_ZIO as writer. * * We use these distinct config locks to avoid recursive lock entry. * For example, spa_sync() (which holds SCL_CONFIG as reader) induces * block allocations (SCL_ALLOC), which may require reading space maps * from disk (dmu_read() -> zio_read() -> SCL_ZIO). * * The spa config locks cannot be normal rwlocks because we need the * ability to hand off ownership. For example, SCL_ZIO is acquired * by the issuing thread and later released by an interrupt thread. * They do, however, obey the usual write-wanted semantics to prevent * writer (i.e. system administrator) starvation. * * The lock acquisition rules are as follows: * * SCL_CONFIG * Protects changes to the vdev tree topology, such as vdev * add/remove/attach/detach. Protects the dirty config list * (spa_config_dirty_list) and the set of spares and l2arc devices. * * SCL_STATE * Protects changes to pool state and vdev state, such as vdev * online/offline/fault/degrade/clear. Protects the dirty state list * (spa_state_dirty_list) and global pool state (spa_state). * * SCL_ALLOC * Protects changes to metaslab groups and classes. * Held as reader by metaslab_alloc() and metaslab_claim(). * * SCL_ZIO * Held by bp-level zios (those which have no io_vd upon entry) * to prevent changes to the vdev tree. The bp-level zio implicitly * protects all of its vdev child zios, which do not hold SCL_ZIO. * * SCL_FREE * Protects changes to metaslab groups and classes. * Held as reader by metaslab_free(). SCL_FREE is distinct from * SCL_ALLOC, and lower than SCL_ZIO, so that we can safely free * blocks in zio_done() while another i/o that holds either * SCL_ALLOC or SCL_ZIO is waiting for this i/o to complete. * * SCL_VDEV * Held as reader to prevent changes to the vdev tree during trivial * inquiries such as bp_get_dsize(). SCL_VDEV is distinct from the * other locks, and lower than all of them, to ensure that it's safe * to acquire regardless of caller context. * * In addition, the following rules apply: * * (a) spa_props_lock protects pool properties, spa_config and spa_config_list. * The lock ordering is SCL_CONFIG > spa_props_lock. * * (b) I/O operations on leaf vdevs. For any zio operation that takes * an explicit vdev_t argument -- such as zio_ioctl(), zio_read_phys(), * or zio_write_phys() -- the caller must ensure that the config cannot * cannot change in the interim, and that the vdev cannot be reopened. * SCL_STATE as reader suffices for both. * * The vdev configuration is protected by spa_vdev_enter() / spa_vdev_exit(). * * spa_vdev_enter() Acquire the namespace lock and the config lock * for writing. * * spa_vdev_exit() Release the config lock, wait for all I/O * to complete, sync the updated configs to the * cache, and release the namespace lock. * * vdev state is protected by spa_vdev_state_enter() / spa_vdev_state_exit(). * Like spa_vdev_enter/exit, these are convenience wrappers -- the actual * locking is, always, based on spa_namespace_lock and spa_config_lock[]. */ static avl_tree_t spa_namespace_avl; kmutex_t spa_namespace_lock; static kcondvar_t spa_namespace_cv; static const int spa_max_replication_override = SPA_DVAS_PER_BP; static kmutex_t spa_spare_lock; static avl_tree_t spa_spare_avl; static kmutex_t spa_l2cache_lock; static avl_tree_t spa_l2cache_avl; spa_mode_t spa_mode_global = SPA_MODE_UNINIT; #ifdef ZFS_DEBUG /* * Everything except dprintf, set_error, spa, and indirect_remap is on * by default in debug builds. */ int zfs_flags = ~(ZFS_DEBUG_DPRINTF | ZFS_DEBUG_SET_ERROR | ZFS_DEBUG_INDIRECT_REMAP); #else int zfs_flags = 0; #endif /* * zfs_recover can be set to nonzero to attempt to recover from * otherwise-fatal errors, typically caused by on-disk corruption. When * set, calls to zfs_panic_recover() will turn into warning messages. * This should only be used as a last resort, as it typically results * in leaked space, or worse. */ int zfs_recover = B_FALSE; /* * If destroy encounters an EIO while reading metadata (e.g. indirect * blocks), space referenced by the missing metadata can not be freed. * Normally this causes the background destroy to become "stalled", as * it is unable to make forward progress. While in this stalled state, * all remaining space to free from the error-encountering filesystem is * "temporarily leaked". Set this flag to cause it to ignore the EIO, * permanently leak the space from indirect blocks that can not be read, * and continue to free everything else that it can. * * The default, "stalling" behavior is useful if the storage partially * fails (i.e. some but not all i/os fail), and then later recovers. In * this case, we will be able to continue pool operations while it is * partially failed, and when it recovers, we can continue to free the * space, with no leaks. However, note that this case is actually * fairly rare. * * Typically pools either (a) fail completely (but perhaps temporarily, * e.g. a top-level vdev going offline), or (b) have localized, * permanent errors (e.g. disk returns the wrong data due to bit flip or * firmware bug). In case (a), this setting does not matter because the * pool will be suspended and the sync thread will not be able to make * forward progress regardless. In case (b), because the error is * permanent, the best we can do is leak the minimum amount of space, * which is what setting this flag will do. Therefore, it is reasonable * for this flag to normally be set, but we chose the more conservative * approach of not setting it, so that there is no possibility of * leaking space in the "partial temporary" failure case. */ int zfs_free_leak_on_eio = B_FALSE; /* * Expiration time in milliseconds. This value has two meanings. First it is * used to determine when the spa_deadman() logic should fire. By default the * spa_deadman() will fire if spa_sync() has not completed in 600 seconds. * Secondly, the value determines if an I/O is considered "hung". Any I/O that * has not completed in zfs_deadman_synctime_ms is considered "hung" resulting * in one of three behaviors controlled by zfs_deadman_failmode. */ uint64_t zfs_deadman_synctime_ms = 600000UL; /* 10 min. */ /* * This value controls the maximum amount of time zio_wait() will block for an * outstanding IO. By default this is 300 seconds at which point the "hung" * behavior will be applied as described for zfs_deadman_synctime_ms. */ uint64_t zfs_deadman_ziotime_ms = 300000UL; /* 5 min. */ /* * Check time in milliseconds. This defines the frequency at which we check * for hung I/O. */ uint64_t zfs_deadman_checktime_ms = 60000UL; /* 1 min. */ /* * By default the deadman is enabled. */ int zfs_deadman_enabled = B_TRUE; /* * Controls the behavior of the deadman when it detects a "hung" I/O. * Valid values are zfs_deadman_failmode=. * * wait - Wait for the "hung" I/O (default) * continue - Attempt to recover from a "hung" I/O * panic - Panic the system */ const char *zfs_deadman_failmode = "wait"; /* * The worst case is single-sector max-parity RAID-Z blocks, in which * case the space requirement is exactly (VDEV_RAIDZ_MAXPARITY + 1) * times the size; so just assume that. Add to this the fact that * we can have up to 3 DVAs per bp, and one more factor of 2 because * the block may be dittoed with up to 3 DVAs by ddt_sync(). All together, * the worst case is: * (VDEV_RAIDZ_MAXPARITY + 1) * SPA_DVAS_PER_BP * 2 == 24 */ uint_t spa_asize_inflation = 24; /* * Normally, we don't allow the last 3.2% (1/(2^spa_slop_shift)) of space in * the pool to be consumed (bounded by spa_max_slop). This ensures that we * don't run the pool completely out of space, due to unaccounted changes (e.g. * to the MOS). It also limits the worst-case time to allocate space. If we * have less than this amount of free space, most ZPL operations (e.g. write, * create) will return ENOSPC. The ZIL metaslabs (spa_embedded_log_class) are * also part of this 3.2% of space which can't be consumed by normal writes; * the slop space "proper" (spa_get_slop_space()) is decreased by the embedded * log space. * * Certain operations (e.g. file removal, most administrative actions) can * use half the slop space. They will only return ENOSPC if less than half * the slop space is free. Typically, once the pool has less than the slop * space free, the user will use these operations to free up space in the pool. * These are the operations that call dsl_pool_adjustedsize() with the netfree * argument set to TRUE. * * Operations that are almost guaranteed to free up space in the absence of * a pool checkpoint can use up to three quarters of the slop space * (e.g zfs destroy). * * A very restricted set of operations are always permitted, regardless of * the amount of free space. These are the operations that call * dsl_sync_task(ZFS_SPACE_CHECK_NONE). If these operations result in a net * increase in the amount of space used, it is possible to run the pool * completely out of space, causing it to be permanently read-only. * * Note that on very small pools, the slop space will be larger than * 3.2%, in an effort to have it be at least spa_min_slop (128MB), * but we never allow it to be more than half the pool size. * * Further, on very large pools, the slop space will be smaller than * 3.2%, to avoid reserving much more space than we actually need; bounded * by spa_max_slop (128GB). * * See also the comments in zfs_space_check_t. */ uint_t spa_slop_shift = 5; static const uint64_t spa_min_slop = 128ULL * 1024 * 1024; static const uint64_t spa_max_slop = 128ULL * 1024 * 1024 * 1024; static const int spa_allocators = 4; void spa_load_failed(spa_t *spa, const char *fmt, ...) { va_list adx; char buf[256]; va_start(adx, fmt); (void) vsnprintf(buf, sizeof (buf), fmt, adx); va_end(adx); zfs_dbgmsg("spa_load(%s, config %s): FAILED: %s", spa->spa_name, spa->spa_trust_config ? "trusted" : "untrusted", buf); } void spa_load_note(spa_t *spa, const char *fmt, ...) { va_list adx; char buf[256]; va_start(adx, fmt); (void) vsnprintf(buf, sizeof (buf), fmt, adx); va_end(adx); zfs_dbgmsg("spa_load(%s, config %s): %s", spa->spa_name, spa->spa_trust_config ? "trusted" : "untrusted", buf); } /* * By default dedup and user data indirects land in the special class */ static int zfs_ddt_data_is_special = B_TRUE; static int zfs_user_indirect_is_special = B_TRUE; /* * The percentage of special class final space reserved for metadata only. * Once we allocate 100 - zfs_special_class_metadata_reserve_pct we only * let metadata into the class. */ static uint_t zfs_special_class_metadata_reserve_pct = 25; /* * ========================================================================== * SPA config locking * ========================================================================== */ static void spa_config_lock_init(spa_t *spa) { for (int i = 0; i < SCL_LOCKS; i++) { spa_config_lock_t *scl = &spa->spa_config_lock[i]; mutex_init(&scl->scl_lock, NULL, MUTEX_DEFAULT, NULL); cv_init(&scl->scl_cv, NULL, CV_DEFAULT, NULL); scl->scl_writer = NULL; scl->scl_write_wanted = 0; scl->scl_count = 0; } } static void spa_config_lock_destroy(spa_t *spa) { for (int i = 0; i < SCL_LOCKS; i++) { spa_config_lock_t *scl = &spa->spa_config_lock[i]; mutex_destroy(&scl->scl_lock); cv_destroy(&scl->scl_cv); ASSERT(scl->scl_writer == NULL); ASSERT(scl->scl_write_wanted == 0); ASSERT(scl->scl_count == 0); } } int spa_config_tryenter(spa_t *spa, int locks, const void *tag, krw_t rw) { for (int i = 0; i < SCL_LOCKS; i++) { spa_config_lock_t *scl = &spa->spa_config_lock[i]; if (!(locks & (1 << i))) continue; mutex_enter(&scl->scl_lock); if (rw == RW_READER) { if (scl->scl_writer || scl->scl_write_wanted) { mutex_exit(&scl->scl_lock); spa_config_exit(spa, locks & ((1 << i) - 1), tag); return (0); } } else { ASSERT(scl->scl_writer != curthread); if (scl->scl_count != 0) { mutex_exit(&scl->scl_lock); spa_config_exit(spa, locks & ((1 << i) - 1), tag); return (0); } scl->scl_writer = curthread; } scl->scl_count++; mutex_exit(&scl->scl_lock); } return (1); } static void spa_config_enter_impl(spa_t *spa, int locks, const void *tag, krw_t rw, int mmp_flag) { (void) tag; int wlocks_held = 0; ASSERT3U(SCL_LOCKS, <, sizeof (wlocks_held) * NBBY); for (int i = 0; i < SCL_LOCKS; i++) { spa_config_lock_t *scl = &spa->spa_config_lock[i]; if (scl->scl_writer == curthread) wlocks_held |= (1 << i); if (!(locks & (1 << i))) continue; mutex_enter(&scl->scl_lock); if (rw == RW_READER) { while (scl->scl_writer || (!mmp_flag && scl->scl_write_wanted)) { cv_wait(&scl->scl_cv, &scl->scl_lock); } } else { ASSERT(scl->scl_writer != curthread); while (scl->scl_count != 0) { scl->scl_write_wanted++; cv_wait(&scl->scl_cv, &scl->scl_lock); scl->scl_write_wanted--; } scl->scl_writer = curthread; } scl->scl_count++; mutex_exit(&scl->scl_lock); } ASSERT3U(wlocks_held, <=, locks); } void spa_config_enter(spa_t *spa, int locks, const void *tag, krw_t rw) { spa_config_enter_impl(spa, locks, tag, rw, 0); } /* * The spa_config_enter_mmp() allows the mmp thread to cut in front of * outstanding write lock requests. This is needed since the mmp updates are * time sensitive and failure to service them promptly will result in a * suspended pool. This pool suspension has been seen in practice when there is * a single disk in a pool that is responding slowly and presumably about to * fail. */ void spa_config_enter_mmp(spa_t *spa, int locks, const void *tag, krw_t rw) { spa_config_enter_impl(spa, locks, tag, rw, 1); } void spa_config_exit(spa_t *spa, int locks, const void *tag) { (void) tag; for (int i = SCL_LOCKS - 1; i >= 0; i--) { spa_config_lock_t *scl = &spa->spa_config_lock[i]; if (!(locks & (1 << i))) continue; mutex_enter(&scl->scl_lock); ASSERT(scl->scl_count > 0); if (--scl->scl_count == 0) { ASSERT(scl->scl_writer == NULL || scl->scl_writer == curthread); scl->scl_writer = NULL; /* OK in either case */ cv_broadcast(&scl->scl_cv); } mutex_exit(&scl->scl_lock); } } int spa_config_held(spa_t *spa, int locks, krw_t rw) { int locks_held = 0; for (int i = 0; i < SCL_LOCKS; i++) { spa_config_lock_t *scl = &spa->spa_config_lock[i]; if (!(locks & (1 << i))) continue; if ((rw == RW_READER && scl->scl_count != 0) || (rw == RW_WRITER && scl->scl_writer == curthread)) locks_held |= 1 << i; } return (locks_held); } /* * ========================================================================== * SPA namespace functions * ========================================================================== */ /* * Lookup the named spa_t in the AVL tree. The spa_namespace_lock must be held. * Returns NULL if no matching spa_t is found. */ spa_t * spa_lookup(const char *name) { static spa_t search; /* spa_t is large; don't allocate on stack */ spa_t *spa; avl_index_t where; char *cp; ASSERT(MUTEX_HELD(&spa_namespace_lock)); (void) strlcpy(search.spa_name, name, sizeof (search.spa_name)); /* * If it's a full dataset name, figure out the pool name and * just use that. */ cp = strpbrk(search.spa_name, "/@#"); if (cp != NULL) *cp = '\0'; spa = avl_find(&spa_namespace_avl, &search, &where); return (spa); } /* * Fires when spa_sync has not completed within zfs_deadman_synctime_ms. * If the zfs_deadman_enabled flag is set then it inspects all vdev queues * looking for potentially hung I/Os. */ void spa_deadman(void *arg) { spa_t *spa = arg; /* Disable the deadman if the pool is suspended. */ if (spa_suspended(spa)) return; zfs_dbgmsg("slow spa_sync: started %llu seconds ago, calls %llu", (gethrtime() - spa->spa_sync_starttime) / NANOSEC, (u_longlong_t)++spa->spa_deadman_calls); if (zfs_deadman_enabled) vdev_deadman(spa->spa_root_vdev, FTAG); spa->spa_deadman_tqid = taskq_dispatch_delay(system_delay_taskq, spa_deadman, spa, TQ_SLEEP, ddi_get_lbolt() + MSEC_TO_TICK(zfs_deadman_checktime_ms)); } static int spa_log_sm_sort_by_txg(const void *va, const void *vb) { const spa_log_sm_t *a = va; const spa_log_sm_t *b = vb; return (TREE_CMP(a->sls_txg, b->sls_txg)); } /* * Create an uninitialized spa_t with the given name. Requires * spa_namespace_lock. The caller must ensure that the spa_t doesn't already * exist by calling spa_lookup() first. */ spa_t * spa_add(const char *name, nvlist_t *config, const char *altroot) { spa_t *spa; spa_config_dirent_t *dp; ASSERT(MUTEX_HELD(&spa_namespace_lock)); spa = kmem_zalloc(sizeof (spa_t), KM_SLEEP); mutex_init(&spa->spa_async_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_errlist_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_errlog_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_evicting_os_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_history_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_proc_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_props_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_cksum_tmpls_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_scrub_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_suspend_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_vdev_top_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_feat_stats_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_flushed_ms_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa->spa_activities_lock, NULL, MUTEX_DEFAULT, NULL); cv_init(&spa->spa_async_cv, NULL, CV_DEFAULT, NULL); cv_init(&spa->spa_evicting_os_cv, NULL, CV_DEFAULT, NULL); cv_init(&spa->spa_proc_cv, NULL, CV_DEFAULT, NULL); cv_init(&spa->spa_scrub_io_cv, NULL, CV_DEFAULT, NULL); cv_init(&spa->spa_suspend_cv, NULL, CV_DEFAULT, NULL); cv_init(&spa->spa_activities_cv, NULL, CV_DEFAULT, NULL); cv_init(&spa->spa_waiters_cv, NULL, CV_DEFAULT, NULL); for (int t = 0; t < TXG_SIZE; t++) bplist_create(&spa->spa_free_bplist[t]); (void) strlcpy(spa->spa_name, name, sizeof (spa->spa_name)); spa->spa_state = POOL_STATE_UNINITIALIZED; spa->spa_freeze_txg = UINT64_MAX; spa->spa_final_txg = UINT64_MAX; spa->spa_load_max_txg = UINT64_MAX; spa->spa_proc = &p0; spa->spa_proc_state = SPA_PROC_NONE; spa->spa_trust_config = B_TRUE; spa->spa_hostid = zone_get_hostid(NULL); spa->spa_deadman_synctime = MSEC2NSEC(zfs_deadman_synctime_ms); spa->spa_deadman_ziotime = MSEC2NSEC(zfs_deadman_ziotime_ms); spa_set_deadman_failmode(spa, zfs_deadman_failmode); zfs_refcount_create(&spa->spa_refcount); spa_config_lock_init(spa); spa_stats_init(spa); avl_add(&spa_namespace_avl, spa); /* * Set the alternate root, if there is one. */ if (altroot) spa->spa_root = spa_strdup(altroot); spa->spa_alloc_count = spa_allocators; spa->spa_allocs = kmem_zalloc(spa->spa_alloc_count * sizeof (spa_alloc_t), KM_SLEEP); for (int i = 0; i < spa->spa_alloc_count; i++) { mutex_init(&spa->spa_allocs[i].spaa_lock, NULL, MUTEX_DEFAULT, NULL); avl_create(&spa->spa_allocs[i].spaa_tree, zio_bookmark_compare, sizeof (zio_t), offsetof(zio_t, io_alloc_node)); } avl_create(&spa->spa_metaslabs_by_flushed, metaslab_sort_by_flushed, sizeof (metaslab_t), offsetof(metaslab_t, ms_spa_txg_node)); avl_create(&spa->spa_sm_logs_by_txg, spa_log_sm_sort_by_txg, sizeof (spa_log_sm_t), offsetof(spa_log_sm_t, sls_node)); list_create(&spa->spa_log_summary, sizeof (log_summary_entry_t), offsetof(log_summary_entry_t, lse_node)); /* * Every pool starts with the default cachefile */ list_create(&spa->spa_config_list, sizeof (spa_config_dirent_t), offsetof(spa_config_dirent_t, scd_link)); dp = kmem_zalloc(sizeof (spa_config_dirent_t), KM_SLEEP); dp->scd_path = altroot ? NULL : spa_strdup(spa_config_path); list_insert_head(&spa->spa_config_list, dp); VERIFY(nvlist_alloc(&spa->spa_load_info, NV_UNIQUE_NAME, KM_SLEEP) == 0); if (config != NULL) { nvlist_t *features; if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_FEATURES_FOR_READ, &features) == 0) { VERIFY(nvlist_dup(features, &spa->spa_label_features, 0) == 0); } VERIFY(nvlist_dup(config, &spa->spa_config, 0) == 0); } if (spa->spa_label_features == NULL) { VERIFY(nvlist_alloc(&spa->spa_label_features, NV_UNIQUE_NAME, KM_SLEEP) == 0); } spa->spa_min_ashift = INT_MAX; spa->spa_max_ashift = 0; spa->spa_min_alloc = INT_MAX; /* Reset cached value */ spa->spa_dedup_dspace = ~0ULL; /* * As a pool is being created, treat all features as disabled by * setting SPA_FEATURE_DISABLED for all entries in the feature * refcount cache. */ for (int i = 0; i < SPA_FEATURES; i++) { spa->spa_feat_refcount_cache[i] = SPA_FEATURE_DISABLED; } list_create(&spa->spa_leaf_list, sizeof (vdev_t), offsetof(vdev_t, vdev_leaf_node)); return (spa); } /* * Removes a spa_t from the namespace, freeing up any memory used. Requires * spa_namespace_lock. This is called only after the spa_t has been closed and * deactivated. */ void spa_remove(spa_t *spa) { spa_config_dirent_t *dp; ASSERT(MUTEX_HELD(&spa_namespace_lock)); ASSERT(spa_state(spa) == POOL_STATE_UNINITIALIZED); ASSERT3U(zfs_refcount_count(&spa->spa_refcount), ==, 0); ASSERT0(spa->spa_waiters); nvlist_free(spa->spa_config_splitting); avl_remove(&spa_namespace_avl, spa); cv_broadcast(&spa_namespace_cv); if (spa->spa_root) spa_strfree(spa->spa_root); while ((dp = list_head(&spa->spa_config_list)) != NULL) { list_remove(&spa->spa_config_list, dp); if (dp->scd_path != NULL) spa_strfree(dp->scd_path); kmem_free(dp, sizeof (spa_config_dirent_t)); } for (int i = 0; i < spa->spa_alloc_count; i++) { avl_destroy(&spa->spa_allocs[i].spaa_tree); mutex_destroy(&spa->spa_allocs[i].spaa_lock); } kmem_free(spa->spa_allocs, spa->spa_alloc_count * sizeof (spa_alloc_t)); avl_destroy(&spa->spa_metaslabs_by_flushed); avl_destroy(&spa->spa_sm_logs_by_txg); list_destroy(&spa->spa_log_summary); list_destroy(&spa->spa_config_list); list_destroy(&spa->spa_leaf_list); nvlist_free(spa->spa_label_features); nvlist_free(spa->spa_load_info); nvlist_free(spa->spa_feat_stats); spa_config_set(spa, NULL); zfs_refcount_destroy(&spa->spa_refcount); spa_stats_destroy(spa); spa_config_lock_destroy(spa); for (int t = 0; t < TXG_SIZE; t++) bplist_destroy(&spa->spa_free_bplist[t]); zio_checksum_templates_free(spa); cv_destroy(&spa->spa_async_cv); cv_destroy(&spa->spa_evicting_os_cv); cv_destroy(&spa->spa_proc_cv); cv_destroy(&spa->spa_scrub_io_cv); cv_destroy(&spa->spa_suspend_cv); cv_destroy(&spa->spa_activities_cv); cv_destroy(&spa->spa_waiters_cv); mutex_destroy(&spa->spa_flushed_ms_lock); mutex_destroy(&spa->spa_async_lock); mutex_destroy(&spa->spa_errlist_lock); mutex_destroy(&spa->spa_errlog_lock); mutex_destroy(&spa->spa_evicting_os_lock); mutex_destroy(&spa->spa_history_lock); mutex_destroy(&spa->spa_proc_lock); mutex_destroy(&spa->spa_props_lock); mutex_destroy(&spa->spa_cksum_tmpls_lock); mutex_destroy(&spa->spa_scrub_lock); mutex_destroy(&spa->spa_suspend_lock); mutex_destroy(&spa->spa_vdev_top_lock); mutex_destroy(&spa->spa_feat_stats_lock); mutex_destroy(&spa->spa_activities_lock); kmem_free(spa, sizeof (spa_t)); } /* * Given a pool, return the next pool in the namespace, or NULL if there is * none. If 'prev' is NULL, return the first pool. */ spa_t * spa_next(spa_t *prev) { ASSERT(MUTEX_HELD(&spa_namespace_lock)); if (prev) return (AVL_NEXT(&spa_namespace_avl, prev)); else return (avl_first(&spa_namespace_avl)); } /* * ========================================================================== * SPA refcount functions * ========================================================================== */ /* * Add a reference to the given spa_t. Must have at least one reference, or * have the namespace lock held. */ void spa_open_ref(spa_t *spa, const void *tag) { ASSERT(zfs_refcount_count(&spa->spa_refcount) >= spa->spa_minref || MUTEX_HELD(&spa_namespace_lock)); (void) zfs_refcount_add(&spa->spa_refcount, tag); } /* * Remove a reference to the given spa_t. Must have at least one reference, or * have the namespace lock held. */ void spa_close(spa_t *spa, const void *tag) { ASSERT(zfs_refcount_count(&spa->spa_refcount) > spa->spa_minref || MUTEX_HELD(&spa_namespace_lock)); (void) zfs_refcount_remove(&spa->spa_refcount, tag); } /* * Remove a reference to the given spa_t held by a dsl dir that is * being asynchronously released. Async releases occur from a taskq * performing eviction of dsl datasets and dirs. The namespace lock * isn't held and the hold by the object being evicted may contribute to * spa_minref (e.g. dataset or directory released during pool export), * so the asserts in spa_close() do not apply. */ void spa_async_close(spa_t *spa, const void *tag) { (void) zfs_refcount_remove(&spa->spa_refcount, tag); } /* * Check to see if the spa refcount is zero. Must be called with * spa_namespace_lock held. We really compare against spa_minref, which is the * number of references acquired when opening a pool */ boolean_t spa_refcount_zero(spa_t *spa) { ASSERT(MUTEX_HELD(&spa_namespace_lock)); return (zfs_refcount_count(&spa->spa_refcount) == spa->spa_minref); } /* * ========================================================================== * SPA spare and l2cache tracking * ========================================================================== */ /* * Hot spares and cache devices are tracked using the same code below, * for 'auxiliary' devices. */ typedef struct spa_aux { uint64_t aux_guid; uint64_t aux_pool; avl_node_t aux_avl; int aux_count; } spa_aux_t; static inline int spa_aux_compare(const void *a, const void *b) { const spa_aux_t *sa = (const spa_aux_t *)a; const spa_aux_t *sb = (const spa_aux_t *)b; return (TREE_CMP(sa->aux_guid, sb->aux_guid)); } static void spa_aux_add(vdev_t *vd, avl_tree_t *avl) { avl_index_t where; spa_aux_t search; spa_aux_t *aux; search.aux_guid = vd->vdev_guid; if ((aux = avl_find(avl, &search, &where)) != NULL) { aux->aux_count++; } else { aux = kmem_zalloc(sizeof (spa_aux_t), KM_SLEEP); aux->aux_guid = vd->vdev_guid; aux->aux_count = 1; avl_insert(avl, aux, where); } } static void spa_aux_remove(vdev_t *vd, avl_tree_t *avl) { spa_aux_t search; spa_aux_t *aux; avl_index_t where; search.aux_guid = vd->vdev_guid; aux = avl_find(avl, &search, &where); ASSERT(aux != NULL); if (--aux->aux_count == 0) { avl_remove(avl, aux); kmem_free(aux, sizeof (spa_aux_t)); } else if (aux->aux_pool == spa_guid(vd->vdev_spa)) { aux->aux_pool = 0ULL; } } static boolean_t spa_aux_exists(uint64_t guid, uint64_t *pool, int *refcnt, avl_tree_t *avl) { spa_aux_t search, *found; search.aux_guid = guid; found = avl_find(avl, &search, NULL); if (pool) { if (found) *pool = found->aux_pool; else *pool = 0ULL; } if (refcnt) { if (found) *refcnt = found->aux_count; else *refcnt = 0; } return (found != NULL); } static void spa_aux_activate(vdev_t *vd, avl_tree_t *avl) { spa_aux_t search, *found; avl_index_t where; search.aux_guid = vd->vdev_guid; found = avl_find(avl, &search, &where); ASSERT(found != NULL); ASSERT(found->aux_pool == 0ULL); found->aux_pool = spa_guid(vd->vdev_spa); } /* * Spares are tracked globally due to the following constraints: * * - A spare may be part of multiple pools. * - A spare may be added to a pool even if it's actively in use within * another pool. * - A spare in use in any pool can only be the source of a replacement if * the target is a spare in the same pool. * * We keep track of all spares on the system through the use of a reference * counted AVL tree. When a vdev is added as a spare, or used as a replacement * spare, then we bump the reference count in the AVL tree. In addition, we set * the 'vdev_isspare' member to indicate that the device is a spare (active or * inactive). When a spare is made active (used to replace a device in the * pool), we also keep track of which pool its been made a part of. * * The 'spa_spare_lock' protects the AVL tree. These functions are normally * called under the spa_namespace lock as part of vdev reconfiguration. The * separate spare lock exists for the status query path, which does not need to * be completely consistent with respect to other vdev configuration changes. */ static int spa_spare_compare(const void *a, const void *b) { return (spa_aux_compare(a, b)); } void spa_spare_add(vdev_t *vd) { mutex_enter(&spa_spare_lock); ASSERT(!vd->vdev_isspare); spa_aux_add(vd, &spa_spare_avl); vd->vdev_isspare = B_TRUE; mutex_exit(&spa_spare_lock); } void spa_spare_remove(vdev_t *vd) { mutex_enter(&spa_spare_lock); ASSERT(vd->vdev_isspare); spa_aux_remove(vd, &spa_spare_avl); vd->vdev_isspare = B_FALSE; mutex_exit(&spa_spare_lock); } boolean_t spa_spare_exists(uint64_t guid, uint64_t *pool, int *refcnt) { boolean_t found; mutex_enter(&spa_spare_lock); found = spa_aux_exists(guid, pool, refcnt, &spa_spare_avl); mutex_exit(&spa_spare_lock); return (found); } void spa_spare_activate(vdev_t *vd) { mutex_enter(&spa_spare_lock); ASSERT(vd->vdev_isspare); spa_aux_activate(vd, &spa_spare_avl); mutex_exit(&spa_spare_lock); } /* * Level 2 ARC devices are tracked globally for the same reasons as spares. * Cache devices currently only support one pool per cache device, and so * for these devices the aux reference count is currently unused beyond 1. */ static int spa_l2cache_compare(const void *a, const void *b) { return (spa_aux_compare(a, b)); } void spa_l2cache_add(vdev_t *vd) { mutex_enter(&spa_l2cache_lock); ASSERT(!vd->vdev_isl2cache); spa_aux_add(vd, &spa_l2cache_avl); vd->vdev_isl2cache = B_TRUE; mutex_exit(&spa_l2cache_lock); } void spa_l2cache_remove(vdev_t *vd) { mutex_enter(&spa_l2cache_lock); ASSERT(vd->vdev_isl2cache); spa_aux_remove(vd, &spa_l2cache_avl); vd->vdev_isl2cache = B_FALSE; mutex_exit(&spa_l2cache_lock); } boolean_t spa_l2cache_exists(uint64_t guid, uint64_t *pool) { boolean_t found; mutex_enter(&spa_l2cache_lock); found = spa_aux_exists(guid, pool, NULL, &spa_l2cache_avl); mutex_exit(&spa_l2cache_lock); return (found); } void spa_l2cache_activate(vdev_t *vd) { mutex_enter(&spa_l2cache_lock); ASSERT(vd->vdev_isl2cache); spa_aux_activate(vd, &spa_l2cache_avl); mutex_exit(&spa_l2cache_lock); } /* * ========================================================================== * SPA vdev locking * ========================================================================== */ /* * Lock the given spa_t for the purpose of adding or removing a vdev. * Grabs the global spa_namespace_lock plus the spa config lock for writing. * It returns the next transaction group for the spa_t. */ uint64_t spa_vdev_enter(spa_t *spa) { mutex_enter(&spa->spa_vdev_top_lock); mutex_enter(&spa_namespace_lock); vdev_autotrim_stop_all(spa); return (spa_vdev_config_enter(spa)); } /* * The same as spa_vdev_enter() above but additionally takes the guid of * the vdev being detached. When there is a rebuild in process it will be * suspended while the vdev tree is modified then resumed by spa_vdev_exit(). * The rebuild is canceled if only a single child remains after the detach. */ uint64_t spa_vdev_detach_enter(spa_t *spa, uint64_t guid) { mutex_enter(&spa->spa_vdev_top_lock); mutex_enter(&spa_namespace_lock); vdev_autotrim_stop_all(spa); if (guid != 0) { vdev_t *vd = spa_lookup_by_guid(spa, guid, B_FALSE); if (vd) { vdev_rebuild_stop_wait(vd->vdev_top); } } return (spa_vdev_config_enter(spa)); } /* * Internal implementation for spa_vdev_enter(). Used when a vdev * operation requires multiple syncs (i.e. removing a device) while * keeping the spa_namespace_lock held. */ uint64_t spa_vdev_config_enter(spa_t *spa) { ASSERT(MUTEX_HELD(&spa_namespace_lock)); spa_config_enter(spa, SCL_ALL, spa, RW_WRITER); return (spa_last_synced_txg(spa) + 1); } /* * Used in combination with spa_vdev_config_enter() to allow the syncing * of multiple transactions without releasing the spa_namespace_lock. */ void spa_vdev_config_exit(spa_t *spa, vdev_t *vd, uint64_t txg, int error, const char *tag) { ASSERT(MUTEX_HELD(&spa_namespace_lock)); int config_changed = B_FALSE; ASSERT(txg > spa_last_synced_txg(spa)); spa->spa_pending_vdev = NULL; /* * Reassess the DTLs. */ vdev_dtl_reassess(spa->spa_root_vdev, 0, 0, B_FALSE, B_FALSE); if (error == 0 && !list_is_empty(&spa->spa_config_dirty_list)) { config_changed = B_TRUE; spa->spa_config_generation++; } /* * Verify the metaslab classes. */ ASSERT(metaslab_class_validate(spa_normal_class(spa)) == 0); ASSERT(metaslab_class_validate(spa_log_class(spa)) == 0); ASSERT(metaslab_class_validate(spa_embedded_log_class(spa)) == 0); ASSERT(metaslab_class_validate(spa_special_class(spa)) == 0); ASSERT(metaslab_class_validate(spa_dedup_class(spa)) == 0); spa_config_exit(spa, SCL_ALL, spa); /* * Panic the system if the specified tag requires it. This * is useful for ensuring that configurations are updated * transactionally. */ if (zio_injection_enabled) zio_handle_panic_injection(spa, tag, 0); /* * Note: this txg_wait_synced() is important because it ensures * that there won't be more than one config change per txg. * This allows us to use the txg as the generation number. */ if (error == 0) txg_wait_synced(spa->spa_dsl_pool, txg); if (vd != NULL) { ASSERT(!vd->vdev_detached || vd->vdev_dtl_sm == NULL); if (vd->vdev_ops->vdev_op_leaf) { mutex_enter(&vd->vdev_initialize_lock); vdev_initialize_stop(vd, VDEV_INITIALIZE_CANCELED, NULL); mutex_exit(&vd->vdev_initialize_lock); mutex_enter(&vd->vdev_trim_lock); vdev_trim_stop(vd, VDEV_TRIM_CANCELED, NULL); mutex_exit(&vd->vdev_trim_lock); } /* * The vdev may be both a leaf and top-level device. */ vdev_autotrim_stop_wait(vd); spa_config_enter(spa, SCL_STATE_ALL, spa, RW_WRITER); vdev_free(vd); spa_config_exit(spa, SCL_STATE_ALL, spa); } /* * If the config changed, update the config cache. */ if (config_changed) spa_write_cachefile(spa, B_FALSE, B_TRUE, B_TRUE); } /* * Unlock the spa_t after adding or removing a vdev. Besides undoing the * locking of spa_vdev_enter(), we also want make sure the transactions have * synced to disk, and then update the global configuration cache with the new * information. */ int spa_vdev_exit(spa_t *spa, vdev_t *vd, uint64_t txg, int error) { vdev_autotrim_restart(spa); vdev_rebuild_restart(spa); spa_vdev_config_exit(spa, vd, txg, error, FTAG); mutex_exit(&spa_namespace_lock); mutex_exit(&spa->spa_vdev_top_lock); return (error); } /* * Lock the given spa_t for the purpose of changing vdev state. */ void spa_vdev_state_enter(spa_t *spa, int oplocks) { int locks = SCL_STATE_ALL | oplocks; /* * Root pools may need to read of the underlying devfs filesystem * when opening up a vdev. Unfortunately if we're holding the * SCL_ZIO lock it will result in a deadlock when we try to issue * the read from the root filesystem. Instead we "prefetch" * the associated vnodes that we need prior to opening the * underlying devices and cache them so that we can prevent * any I/O when we are doing the actual open. */ if (spa_is_root(spa)) { int low = locks & ~(SCL_ZIO - 1); int high = locks & ~low; spa_config_enter(spa, high, spa, RW_WRITER); vdev_hold(spa->spa_root_vdev); spa_config_enter(spa, low, spa, RW_WRITER); } else { spa_config_enter(spa, locks, spa, RW_WRITER); } spa->spa_vdev_locks = locks; } int spa_vdev_state_exit(spa_t *spa, vdev_t *vd, int error) { boolean_t config_changed = B_FALSE; vdev_t *vdev_top; if (vd == NULL || vd == spa->spa_root_vdev) { vdev_top = spa->spa_root_vdev; } else { vdev_top = vd->vdev_top; } if (vd != NULL || error == 0) vdev_dtl_reassess(vdev_top, 0, 0, B_FALSE, B_FALSE); if (vd != NULL) { if (vd != spa->spa_root_vdev) vdev_state_dirty(vdev_top); config_changed = B_TRUE; spa->spa_config_generation++; } if (spa_is_root(spa)) vdev_rele(spa->spa_root_vdev); ASSERT3U(spa->spa_vdev_locks, >=, SCL_STATE_ALL); spa_config_exit(spa, spa->spa_vdev_locks, spa); /* * If anything changed, wait for it to sync. This ensures that, * from the system administrator's perspective, zpool(8) commands * are synchronous. This is important for things like zpool offline: * when the command completes, you expect no further I/O from ZFS. */ if (vd != NULL) txg_wait_synced(spa->spa_dsl_pool, 0); /* * If the config changed, update the config cache. */ if (config_changed) { mutex_enter(&spa_namespace_lock); spa_write_cachefile(spa, B_FALSE, B_TRUE, B_FALSE); mutex_exit(&spa_namespace_lock); } return (error); } /* * ========================================================================== * Miscellaneous functions * ========================================================================== */ void spa_activate_mos_feature(spa_t *spa, const char *feature, dmu_tx_t *tx) { if (!nvlist_exists(spa->spa_label_features, feature)) { fnvlist_add_boolean(spa->spa_label_features, feature); /* * When we are creating the pool (tx_txg==TXG_INITIAL), we can't * dirty the vdev config because lock SCL_CONFIG is not held. * Thankfully, in this case we don't need to dirty the config * because it will be written out anyway when we finish * creating the pool. */ if (tx->tx_txg != TXG_INITIAL) vdev_config_dirty(spa->spa_root_vdev); } } void spa_deactivate_mos_feature(spa_t *spa, const char *feature) { if (nvlist_remove_all(spa->spa_label_features, feature) == 0) vdev_config_dirty(spa->spa_root_vdev); } /* * Return the spa_t associated with given pool_guid, if it exists. If * device_guid is non-zero, determine whether the pool exists *and* contains * a device with the specified device_guid. */ spa_t * spa_by_guid(uint64_t pool_guid, uint64_t device_guid) { spa_t *spa; avl_tree_t *t = &spa_namespace_avl; ASSERT(MUTEX_HELD(&spa_namespace_lock)); for (spa = avl_first(t); spa != NULL; spa = AVL_NEXT(t, spa)) { if (spa->spa_state == POOL_STATE_UNINITIALIZED) continue; if (spa->spa_root_vdev == NULL) continue; if (spa_guid(spa) == pool_guid) { if (device_guid == 0) break; if (vdev_lookup_by_guid(spa->spa_root_vdev, device_guid) != NULL) break; /* * Check any devices we may be in the process of adding. */ if (spa->spa_pending_vdev) { if (vdev_lookup_by_guid(spa->spa_pending_vdev, device_guid) != NULL) break; } } } return (spa); } /* * Determine whether a pool with the given pool_guid exists. */ boolean_t spa_guid_exists(uint64_t pool_guid, uint64_t device_guid) { return (spa_by_guid(pool_guid, device_guid) != NULL); } char * spa_strdup(const char *s) { size_t len; char *new; len = strlen(s); new = kmem_alloc(len + 1, KM_SLEEP); memcpy(new, s, len + 1); return (new); } void spa_strfree(char *s) { kmem_free(s, strlen(s) + 1); } uint64_t spa_generate_guid(spa_t *spa) { uint64_t guid; if (spa != NULL) { do { (void) random_get_pseudo_bytes((void *)&guid, sizeof (guid)); } while (guid == 0 || spa_guid_exists(spa_guid(spa), guid)); } else { do { (void) random_get_pseudo_bytes((void *)&guid, sizeof (guid)); } while (guid == 0 || spa_guid_exists(guid, 0)); } return (guid); } void snprintf_blkptr(char *buf, size_t buflen, const blkptr_t *bp) { char type[256]; const char *checksum = NULL; const char *compress = NULL; if (bp != NULL) { if (BP_GET_TYPE(bp) & DMU_OT_NEWTYPE) { dmu_object_byteswap_t bswap = DMU_OT_BYTESWAP(BP_GET_TYPE(bp)); (void) snprintf(type, sizeof (type), "bswap %s %s", DMU_OT_IS_METADATA(BP_GET_TYPE(bp)) ? "metadata" : "data", dmu_ot_byteswap[bswap].ob_name); } else { (void) strlcpy(type, dmu_ot[BP_GET_TYPE(bp)].ot_name, sizeof (type)); } if (!BP_IS_EMBEDDED(bp)) { checksum = zio_checksum_table[BP_GET_CHECKSUM(bp)].ci_name; } compress = zio_compress_table[BP_GET_COMPRESS(bp)].ci_name; } SNPRINTF_BLKPTR(kmem_scnprintf, ' ', buf, buflen, bp, type, checksum, compress); } void spa_freeze(spa_t *spa) { uint64_t freeze_txg = 0; spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER); if (spa->spa_freeze_txg == UINT64_MAX) { freeze_txg = spa_last_synced_txg(spa) + TXG_SIZE; spa->spa_freeze_txg = freeze_txg; } spa_config_exit(spa, SCL_ALL, FTAG); if (freeze_txg != 0) txg_wait_synced(spa_get_dsl(spa), freeze_txg); } void zfs_panic_recover(const char *fmt, ...) { va_list adx; va_start(adx, fmt); vcmn_err(zfs_recover ? CE_WARN : CE_PANIC, fmt, adx); va_end(adx); } /* * This is a stripped-down version of strtoull, suitable only for converting * lowercase hexadecimal numbers that don't overflow. */ uint64_t zfs_strtonum(const char *str, char **nptr) { uint64_t val = 0; char c; int digit; while ((c = *str) != '\0') { if (c >= '0' && c <= '9') digit = c - '0'; else if (c >= 'a' && c <= 'f') digit = 10 + c - 'a'; else break; val *= 16; val += digit; str++; } if (nptr) *nptr = (char *)str; return (val); } void spa_activate_allocation_classes(spa_t *spa, dmu_tx_t *tx) { /* * We bump the feature refcount for each special vdev added to the pool */ ASSERT(spa_feature_is_enabled(spa, SPA_FEATURE_ALLOCATION_CLASSES)); spa_feature_incr(spa, SPA_FEATURE_ALLOCATION_CLASSES, tx); } /* * ========================================================================== * Accessor functions * ========================================================================== */ boolean_t spa_shutting_down(spa_t *spa) { return (spa->spa_async_suspended); } dsl_pool_t * spa_get_dsl(spa_t *spa) { return (spa->spa_dsl_pool); } boolean_t spa_is_initializing(spa_t *spa) { return (spa->spa_is_initializing); } boolean_t spa_indirect_vdevs_loaded(spa_t *spa) { return (spa->spa_indirect_vdevs_loaded); } blkptr_t * spa_get_rootblkptr(spa_t *spa) { return (&spa->spa_ubsync.ub_rootbp); } void spa_set_rootblkptr(spa_t *spa, const blkptr_t *bp) { spa->spa_uberblock.ub_rootbp = *bp; } void spa_altroot(spa_t *spa, char *buf, size_t buflen) { if (spa->spa_root == NULL) buf[0] = '\0'; else (void) strlcpy(buf, spa->spa_root, buflen); } uint32_t spa_sync_pass(spa_t *spa) { return (spa->spa_sync_pass); } char * spa_name(spa_t *spa) { return (spa->spa_name); } uint64_t spa_guid(spa_t *spa) { dsl_pool_t *dp = spa_get_dsl(spa); uint64_t guid; /* * If we fail to parse the config during spa_load(), we can go through * the error path (which posts an ereport) and end up here with no root * vdev. We stash the original pool guid in 'spa_config_guid' to handle * this case. */ if (spa->spa_root_vdev == NULL) return (spa->spa_config_guid); guid = spa->spa_last_synced_guid != 0 ? spa->spa_last_synced_guid : spa->spa_root_vdev->vdev_guid; /* * Return the most recently synced out guid unless we're * in syncing context. */ if (dp && dsl_pool_sync_context(dp)) return (spa->spa_root_vdev->vdev_guid); else return (guid); } uint64_t spa_load_guid(spa_t *spa) { /* * This is a GUID that exists solely as a reference for the * purposes of the arc. It is generated at load time, and * is never written to persistent storage. */ return (spa->spa_load_guid); } uint64_t spa_last_synced_txg(spa_t *spa) { return (spa->spa_ubsync.ub_txg); } uint64_t spa_first_txg(spa_t *spa) { return (spa->spa_first_txg); } uint64_t spa_syncing_txg(spa_t *spa) { return (spa->spa_syncing_txg); } /* * Return the last txg where data can be dirtied. The final txgs * will be used to just clear out any deferred frees that remain. */ uint64_t spa_final_dirty_txg(spa_t *spa) { return (spa->spa_final_txg - TXG_DEFER_SIZE); } pool_state_t spa_state(spa_t *spa) { return (spa->spa_state); } spa_load_state_t spa_load_state(spa_t *spa) { return (spa->spa_load_state); } uint64_t spa_freeze_txg(spa_t *spa) { return (spa->spa_freeze_txg); } /* * Return the inflated asize for a logical write in bytes. This is used by the * DMU to calculate the space a logical write will require on disk. * If lsize is smaller than the largest physical block size allocatable on this * pool we use its value instead, since the write will end up using the whole * block anyway. */ uint64_t spa_get_worst_case_asize(spa_t *spa, uint64_t lsize) { if (lsize == 0) return (0); /* No inflation needed */ return (MAX(lsize, 1 << spa->spa_max_ashift) * spa_asize_inflation); } /* * Return the amount of slop space in bytes. It is typically 1/32 of the pool * (3.2%), minus the embedded log space. On very small pools, it may be * slightly larger than this. On very large pools, it will be capped to * the value of spa_max_slop. The embedded log space is not included in * spa_dspace. By subtracting it, the usable space (per "zfs list") is a * constant 97% of the total space, regardless of metaslab size (assuming the * default spa_slop_shift=5 and a non-tiny pool). * * See the comment above spa_slop_shift for more details. */ uint64_t spa_get_slop_space(spa_t *spa) { uint64_t space = 0; uint64_t slop = 0; /* * Make sure spa_dedup_dspace has been set. */ if (spa->spa_dedup_dspace == ~0ULL) spa_update_dspace(spa); /* * spa_get_dspace() includes the space only logically "used" by * deduplicated data, so since it's not useful to reserve more * space with more deduplicated data, we subtract that out here. */ space = spa_get_dspace(spa) - spa->spa_dedup_dspace; slop = MIN(space >> spa_slop_shift, spa_max_slop); /* * Subtract the embedded log space, but no more than half the (3.2%) * unusable space. Note, the "no more than half" is only relevant if * zfs_embedded_slog_min_ms >> spa_slop_shift < 2, which is not true by * default. */ uint64_t embedded_log = metaslab_class_get_dspace(spa_embedded_log_class(spa)); slop -= MIN(embedded_log, slop >> 1); /* * Slop space should be at least spa_min_slop, but no more than half * the entire pool. */ slop = MAX(slop, MIN(space >> 1, spa_min_slop)); return (slop); } uint64_t spa_get_dspace(spa_t *spa) { return (spa->spa_dspace); } uint64_t spa_get_checkpoint_space(spa_t *spa) { return (spa->spa_checkpoint_info.sci_dspace); } void spa_update_dspace(spa_t *spa) { spa->spa_dspace = metaslab_class_get_dspace(spa_normal_class(spa)) + ddt_get_dedup_dspace(spa) + brt_get_dspace(spa); if (spa->spa_nonallocating_dspace > 0) { /* * Subtract the space provided by all non-allocating vdevs that * contribute to dspace. If a file is overwritten, its old * blocks are freed and new blocks are allocated. If there are * no snapshots of the file, the available space should remain * the same. The old blocks could be freed from the * non-allocating vdev, but the new blocks must be allocated on * other (allocating) vdevs. By reserving the entire size of * the non-allocating vdevs (including allocated space), we * ensure that there will be enough space on the allocating * vdevs for this file overwrite to succeed. * * Note that the DMU/DSL doesn't actually know or care * how much space is allocated (it does its own tracking * of how much space has been logically used). So it * doesn't matter that the data we are moving may be * allocated twice (on the old device and the new device). */ ASSERT3U(spa->spa_dspace, >=, spa->spa_nonallocating_dspace); spa->spa_dspace -= spa->spa_nonallocating_dspace; } } /* * Return the failure mode that has been set to this pool. The default * behavior will be to block all I/Os when a complete failure occurs. */ uint64_t spa_get_failmode(spa_t *spa) { return (spa->spa_failmode); } boolean_t spa_suspended(spa_t *spa) { return (spa->spa_suspended != ZIO_SUSPEND_NONE); } uint64_t spa_version(spa_t *spa) { return (spa->spa_ubsync.ub_version); } boolean_t spa_deflate(spa_t *spa) { return (spa->spa_deflate); } metaslab_class_t * spa_normal_class(spa_t *spa) { return (spa->spa_normal_class); } metaslab_class_t * spa_log_class(spa_t *spa) { return (spa->spa_log_class); } metaslab_class_t * spa_embedded_log_class(spa_t *spa) { return (spa->spa_embedded_log_class); } metaslab_class_t * spa_special_class(spa_t *spa) { return (spa->spa_special_class); } metaslab_class_t * spa_dedup_class(spa_t *spa) { return (spa->spa_dedup_class); } /* * Locate an appropriate allocation class */ metaslab_class_t * spa_preferred_class(spa_t *spa, uint64_t size, dmu_object_type_t objtype, uint_t level, uint_t special_smallblk) { /* * ZIL allocations determine their class in zio_alloc_zil(). */ ASSERT(objtype != DMU_OT_INTENT_LOG); boolean_t has_special_class = spa->spa_special_class->mc_groups != 0; if (DMU_OT_IS_DDT(objtype)) { if (spa->spa_dedup_class->mc_groups != 0) return (spa_dedup_class(spa)); else if (has_special_class && zfs_ddt_data_is_special) return (spa_special_class(spa)); else return (spa_normal_class(spa)); } /* Indirect blocks for user data can land in special if allowed */ if (level > 0 && (DMU_OT_IS_FILE(objtype) || objtype == DMU_OT_ZVOL)) { if (has_special_class && zfs_user_indirect_is_special) return (spa_special_class(spa)); else return (spa_normal_class(spa)); } if (DMU_OT_IS_METADATA(objtype) || level > 0) { if (has_special_class) return (spa_special_class(spa)); else return (spa_normal_class(spa)); } /* * Allow small file blocks in special class in some cases (like * for the dRAID vdev feature). But always leave a reserve of * zfs_special_class_metadata_reserve_pct exclusively for metadata. */ if (DMU_OT_IS_FILE(objtype) && has_special_class && size <= special_smallblk) { metaslab_class_t *special = spa_special_class(spa); uint64_t alloc = metaslab_class_get_alloc(special); uint64_t space = metaslab_class_get_space(special); uint64_t limit = (space * (100 - zfs_special_class_metadata_reserve_pct)) / 100; if (alloc < limit) return (special); } return (spa_normal_class(spa)); } void spa_evicting_os_register(spa_t *spa, objset_t *os) { mutex_enter(&spa->spa_evicting_os_lock); list_insert_head(&spa->spa_evicting_os_list, os); mutex_exit(&spa->spa_evicting_os_lock); } void spa_evicting_os_deregister(spa_t *spa, objset_t *os) { mutex_enter(&spa->spa_evicting_os_lock); list_remove(&spa->spa_evicting_os_list, os); cv_broadcast(&spa->spa_evicting_os_cv); mutex_exit(&spa->spa_evicting_os_lock); } void spa_evicting_os_wait(spa_t *spa) { mutex_enter(&spa->spa_evicting_os_lock); while (!list_is_empty(&spa->spa_evicting_os_list)) cv_wait(&spa->spa_evicting_os_cv, &spa->spa_evicting_os_lock); mutex_exit(&spa->spa_evicting_os_lock); dmu_buf_user_evict_wait(); } int spa_max_replication(spa_t *spa) { /* * As of SPA_VERSION == SPA_VERSION_DITTO_BLOCKS, we are able to * handle BPs with more than one DVA allocated. Set our max * replication level accordingly. */ if (spa_version(spa) < SPA_VERSION_DITTO_BLOCKS) return (1); return (MIN(SPA_DVAS_PER_BP, spa_max_replication_override)); } int spa_prev_software_version(spa_t *spa) { return (spa->spa_prev_software_version); } uint64_t spa_deadman_synctime(spa_t *spa) { return (spa->spa_deadman_synctime); } spa_autotrim_t spa_get_autotrim(spa_t *spa) { return (spa->spa_autotrim); } uint64_t spa_deadman_ziotime(spa_t *spa) { return (spa->spa_deadman_ziotime); } uint64_t spa_get_deadman_failmode(spa_t *spa) { return (spa->spa_deadman_failmode); } void spa_set_deadman_failmode(spa_t *spa, const char *failmode) { if (strcmp(failmode, "wait") == 0) spa->spa_deadman_failmode = ZIO_FAILURE_MODE_WAIT; else if (strcmp(failmode, "continue") == 0) spa->spa_deadman_failmode = ZIO_FAILURE_MODE_CONTINUE; else if (strcmp(failmode, "panic") == 0) spa->spa_deadman_failmode = ZIO_FAILURE_MODE_PANIC; else spa->spa_deadman_failmode = ZIO_FAILURE_MODE_WAIT; } void spa_set_deadman_ziotime(hrtime_t ns) { spa_t *spa = NULL; if (spa_mode_global != SPA_MODE_UNINIT) { mutex_enter(&spa_namespace_lock); while ((spa = spa_next(spa)) != NULL) spa->spa_deadman_ziotime = ns; mutex_exit(&spa_namespace_lock); } } void spa_set_deadman_synctime(hrtime_t ns) { spa_t *spa = NULL; if (spa_mode_global != SPA_MODE_UNINIT) { mutex_enter(&spa_namespace_lock); while ((spa = spa_next(spa)) != NULL) spa->spa_deadman_synctime = ns; mutex_exit(&spa_namespace_lock); } } uint64_t dva_get_dsize_sync(spa_t *spa, const dva_t *dva) { uint64_t asize = DVA_GET_ASIZE(dva); uint64_t dsize = asize; ASSERT(spa_config_held(spa, SCL_ALL, RW_READER) != 0); if (asize != 0 && spa->spa_deflate) { vdev_t *vd = vdev_lookup_top(spa, DVA_GET_VDEV(dva)); if (vd != NULL) dsize = (asize >> SPA_MINBLOCKSHIFT) * vd->vdev_deflate_ratio; } return (dsize); } uint64_t bp_get_dsize_sync(spa_t *spa, const blkptr_t *bp) { uint64_t dsize = 0; for (int d = 0; d < BP_GET_NDVAS(bp); d++) dsize += dva_get_dsize_sync(spa, &bp->blk_dva[d]); return (dsize); } uint64_t bp_get_dsize(spa_t *spa, const blkptr_t *bp) { uint64_t dsize = 0; spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER); for (int d = 0; d < BP_GET_NDVAS(bp); d++) dsize += dva_get_dsize_sync(spa, &bp->blk_dva[d]); spa_config_exit(spa, SCL_VDEV, FTAG); return (dsize); } uint64_t spa_dirty_data(spa_t *spa) { return (spa->spa_dsl_pool->dp_dirty_total); } /* * ========================================================================== * SPA Import Progress Routines * ========================================================================== */ typedef struct spa_import_progress { uint64_t pool_guid; /* unique id for updates */ char *pool_name; spa_load_state_t spa_load_state; uint64_t mmp_sec_remaining; /* MMP activity check */ uint64_t spa_load_max_txg; /* rewind txg */ procfs_list_node_t smh_node; } spa_import_progress_t; spa_history_list_t *spa_import_progress_list = NULL; static int spa_import_progress_show_header(struct seq_file *f) { seq_printf(f, "%-20s %-14s %-14s %-12s %s\n", "pool_guid", "load_state", "multihost_secs", "max_txg", "pool_name"); return (0); } static int spa_import_progress_show(struct seq_file *f, void *data) { spa_import_progress_t *sip = (spa_import_progress_t *)data; seq_printf(f, "%-20llu %-14llu %-14llu %-12llu %s\n", (u_longlong_t)sip->pool_guid, (u_longlong_t)sip->spa_load_state, (u_longlong_t)sip->mmp_sec_remaining, (u_longlong_t)sip->spa_load_max_txg, (sip->pool_name ? sip->pool_name : "-")); return (0); } /* Remove oldest elements from list until there are no more than 'size' left */ static void spa_import_progress_truncate(spa_history_list_t *shl, unsigned int size) { spa_import_progress_t *sip; while (shl->size > size) { sip = list_remove_head(&shl->procfs_list.pl_list); if (sip->pool_name) spa_strfree(sip->pool_name); kmem_free(sip, sizeof (spa_import_progress_t)); shl->size--; } IMPLY(size == 0, list_is_empty(&shl->procfs_list.pl_list)); } static void spa_import_progress_init(void) { spa_import_progress_list = kmem_zalloc(sizeof (spa_history_list_t), KM_SLEEP); spa_import_progress_list->size = 0; spa_import_progress_list->procfs_list.pl_private = spa_import_progress_list; procfs_list_install("zfs", NULL, "import_progress", 0644, &spa_import_progress_list->procfs_list, spa_import_progress_show, spa_import_progress_show_header, NULL, offsetof(spa_import_progress_t, smh_node)); } static void spa_import_progress_destroy(void) { spa_history_list_t *shl = spa_import_progress_list; procfs_list_uninstall(&shl->procfs_list); spa_import_progress_truncate(shl, 0); procfs_list_destroy(&shl->procfs_list); kmem_free(shl, sizeof (spa_history_list_t)); } int spa_import_progress_set_state(uint64_t pool_guid, spa_load_state_t load_state) { spa_history_list_t *shl = spa_import_progress_list; spa_import_progress_t *sip; int error = ENOENT; if (shl->size == 0) return (0); mutex_enter(&shl->procfs_list.pl_lock); for (sip = list_tail(&shl->procfs_list.pl_list); sip != NULL; sip = list_prev(&shl->procfs_list.pl_list, sip)) { if (sip->pool_guid == pool_guid) { sip->spa_load_state = load_state; error = 0; break; } } mutex_exit(&shl->procfs_list.pl_lock); return (error); } int spa_import_progress_set_max_txg(uint64_t pool_guid, uint64_t load_max_txg) { spa_history_list_t *shl = spa_import_progress_list; spa_import_progress_t *sip; int error = ENOENT; if (shl->size == 0) return (0); mutex_enter(&shl->procfs_list.pl_lock); for (sip = list_tail(&shl->procfs_list.pl_list); sip != NULL; sip = list_prev(&shl->procfs_list.pl_list, sip)) { if (sip->pool_guid == pool_guid) { sip->spa_load_max_txg = load_max_txg; error = 0; break; } } mutex_exit(&shl->procfs_list.pl_lock); return (error); } int spa_import_progress_set_mmp_check(uint64_t pool_guid, uint64_t mmp_sec_remaining) { spa_history_list_t *shl = spa_import_progress_list; spa_import_progress_t *sip; int error = ENOENT; if (shl->size == 0) return (0); mutex_enter(&shl->procfs_list.pl_lock); for (sip = list_tail(&shl->procfs_list.pl_list); sip != NULL; sip = list_prev(&shl->procfs_list.pl_list, sip)) { if (sip->pool_guid == pool_guid) { sip->mmp_sec_remaining = mmp_sec_remaining; error = 0; break; } } mutex_exit(&shl->procfs_list.pl_lock); return (error); } /* * A new import is in progress, add an entry. */ void spa_import_progress_add(spa_t *spa) { spa_history_list_t *shl = spa_import_progress_list; spa_import_progress_t *sip; const char *poolname = NULL; sip = kmem_zalloc(sizeof (spa_import_progress_t), KM_SLEEP); sip->pool_guid = spa_guid(spa); (void) nvlist_lookup_string(spa->spa_config, ZPOOL_CONFIG_POOL_NAME, &poolname); if (poolname == NULL) poolname = spa_name(spa); sip->pool_name = spa_strdup(poolname); sip->spa_load_state = spa_load_state(spa); mutex_enter(&shl->procfs_list.pl_lock); procfs_list_add(&shl->procfs_list, sip); shl->size++; mutex_exit(&shl->procfs_list.pl_lock); } void spa_import_progress_remove(uint64_t pool_guid) { spa_history_list_t *shl = spa_import_progress_list; spa_import_progress_t *sip; mutex_enter(&shl->procfs_list.pl_lock); for (sip = list_tail(&shl->procfs_list.pl_list); sip != NULL; sip = list_prev(&shl->procfs_list.pl_list, sip)) { if (sip->pool_guid == pool_guid) { if (sip->pool_name) spa_strfree(sip->pool_name); list_remove(&shl->procfs_list.pl_list, sip); shl->size--; kmem_free(sip, sizeof (spa_import_progress_t)); break; } } mutex_exit(&shl->procfs_list.pl_lock); } /* * ========================================================================== * Initialization and Termination * ========================================================================== */ static int spa_name_compare(const void *a1, const void *a2) { const spa_t *s1 = a1; const spa_t *s2 = a2; int s; s = strcmp(s1->spa_name, s2->spa_name); return (TREE_ISIGN(s)); } void spa_boot_init(void) { spa_config_load(); } void spa_init(spa_mode_t mode) { mutex_init(&spa_namespace_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa_spare_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&spa_l2cache_lock, NULL, MUTEX_DEFAULT, NULL); cv_init(&spa_namespace_cv, NULL, CV_DEFAULT, NULL); avl_create(&spa_namespace_avl, spa_name_compare, sizeof (spa_t), offsetof(spa_t, spa_avl)); avl_create(&spa_spare_avl, spa_spare_compare, sizeof (spa_aux_t), offsetof(spa_aux_t, aux_avl)); avl_create(&spa_l2cache_avl, spa_l2cache_compare, sizeof (spa_aux_t), offsetof(spa_aux_t, aux_avl)); spa_mode_global = mode; #ifndef _KERNEL if (spa_mode_global != SPA_MODE_READ && dprintf_find_string("watch")) { struct sigaction sa; sa.sa_flags = SA_SIGINFO; sigemptyset(&sa.sa_mask); sa.sa_sigaction = arc_buf_sigsegv; if (sigaction(SIGSEGV, &sa, NULL) == -1) { perror("could not enable watchpoints: " "sigaction(SIGSEGV, ...) = "); } else { arc_watch = B_TRUE; } } #endif fm_init(); zfs_refcount_init(); unique_init(); zfs_btree_init(); metaslab_stat_init(); brt_init(); ddt_init(); zio_init(); dmu_init(); zil_init(); vdev_cache_stat_init(); vdev_mirror_stat_init(); vdev_raidz_math_init(); vdev_file_init(); zfs_prop_init(); chksum_init(); zpool_prop_init(); zpool_feature_init(); spa_config_load(); vdev_prop_init(); l2arc_start(); scan_init(); qat_init(); spa_import_progress_init(); } void spa_fini(void) { l2arc_stop(); spa_evict_all(); vdev_file_fini(); vdev_cache_stat_fini(); vdev_mirror_stat_fini(); vdev_raidz_math_fini(); chksum_fini(); zil_fini(); dmu_fini(); zio_fini(); ddt_fini(); brt_fini(); metaslab_stat_fini(); zfs_btree_fini(); unique_fini(); zfs_refcount_fini(); fm_fini(); scan_fini(); qat_fini(); spa_import_progress_destroy(); avl_destroy(&spa_namespace_avl); avl_destroy(&spa_spare_avl); avl_destroy(&spa_l2cache_avl); cv_destroy(&spa_namespace_cv); mutex_destroy(&spa_namespace_lock); mutex_destroy(&spa_spare_lock); mutex_destroy(&spa_l2cache_lock); } /* * Return whether this pool has a dedicated slog device. No locking needed. * It's not a problem if the wrong answer is returned as it's only for * performance and not correctness. */ boolean_t spa_has_slogs(spa_t *spa) { return (spa->spa_log_class->mc_groups != 0); } spa_log_state_t spa_get_log_state(spa_t *spa) { return (spa->spa_log_state); } void spa_set_log_state(spa_t *spa, spa_log_state_t state) { spa->spa_log_state = state; } boolean_t spa_is_root(spa_t *spa) { return (spa->spa_is_root); } boolean_t spa_writeable(spa_t *spa) { return (!!(spa->spa_mode & SPA_MODE_WRITE) && spa->spa_trust_config); } /* * Returns true if there is a pending sync task in any of the current * syncing txg, the current quiescing txg, or the current open txg. */ boolean_t spa_has_pending_synctask(spa_t *spa) { return (!txg_all_lists_empty(&spa->spa_dsl_pool->dp_sync_tasks) || !txg_all_lists_empty(&spa->spa_dsl_pool->dp_early_sync_tasks)); } spa_mode_t spa_mode(spa_t *spa) { return (spa->spa_mode); } uint64_t spa_bootfs(spa_t *spa) { return (spa->spa_bootfs); } uint64_t spa_delegation(spa_t *spa) { return (spa->spa_delegation); } objset_t * spa_meta_objset(spa_t *spa) { return (spa->spa_meta_objset); } enum zio_checksum spa_dedup_checksum(spa_t *spa) { return (spa->spa_dedup_checksum); } /* * Reset pool scan stat per scan pass (or reboot). */ void spa_scan_stat_init(spa_t *spa) { /* data not stored on disk */ spa->spa_scan_pass_start = gethrestime_sec(); if (dsl_scan_is_paused_scrub(spa->spa_dsl_pool->dp_scan)) spa->spa_scan_pass_scrub_pause = spa->spa_scan_pass_start; else spa->spa_scan_pass_scrub_pause = 0; + + if (dsl_errorscrub_is_paused(spa->spa_dsl_pool->dp_scan)) + spa->spa_scan_pass_errorscrub_pause = spa->spa_scan_pass_start; + else + spa->spa_scan_pass_errorscrub_pause = 0; + spa->spa_scan_pass_scrub_spent_paused = 0; spa->spa_scan_pass_exam = 0; spa->spa_scan_pass_issued = 0; + + // error scrub stats + spa->spa_scan_pass_errorscrub_spent_paused = 0; } /* * Get scan stats for zpool status reports */ int spa_scan_get_stats(spa_t *spa, pool_scan_stat_t *ps) { dsl_scan_t *scn = spa->spa_dsl_pool ? spa->spa_dsl_pool->dp_scan : NULL; - if (scn == NULL || scn->scn_phys.scn_func == POOL_SCAN_NONE) + if (scn == NULL || (scn->scn_phys.scn_func == POOL_SCAN_NONE && + scn->errorscrub_phys.dep_func == POOL_SCAN_NONE)) return (SET_ERROR(ENOENT)); + memset(ps, 0, sizeof (pool_scan_stat_t)); /* data stored on disk */ ps->pss_func = scn->scn_phys.scn_func; ps->pss_state = scn->scn_phys.scn_state; ps->pss_start_time = scn->scn_phys.scn_start_time; ps->pss_end_time = scn->scn_phys.scn_end_time; ps->pss_to_examine = scn->scn_phys.scn_to_examine; ps->pss_examined = scn->scn_phys.scn_examined; ps->pss_to_process = scn->scn_phys.scn_to_process; ps->pss_processed = scn->scn_phys.scn_processed; ps->pss_errors = scn->scn_phys.scn_errors; /* data not stored on disk */ ps->pss_pass_exam = spa->spa_scan_pass_exam; ps->pss_pass_start = spa->spa_scan_pass_start; ps->pss_pass_scrub_pause = spa->spa_scan_pass_scrub_pause; ps->pss_pass_scrub_spent_paused = spa->spa_scan_pass_scrub_spent_paused; ps->pss_pass_issued = spa->spa_scan_pass_issued; ps->pss_issued = scn->scn_issued_before_pass + spa->spa_scan_pass_issued; + /* error scrub data stored on disk */ + ps->pss_error_scrub_func = scn->errorscrub_phys.dep_func; + ps->pss_error_scrub_state = scn->errorscrub_phys.dep_state; + ps->pss_error_scrub_start = scn->errorscrub_phys.dep_start_time; + ps->pss_error_scrub_end = scn->errorscrub_phys.dep_end_time; + ps->pss_error_scrub_examined = scn->errorscrub_phys.dep_examined; + ps->pss_error_scrub_to_be_examined = + scn->errorscrub_phys.dep_to_examine; + + /* error scrub data not stored on disk */ + ps->pss_pass_error_scrub_pause = spa->spa_scan_pass_errorscrub_pause; + return (0); } int spa_maxblocksize(spa_t *spa) { if (spa_feature_is_enabled(spa, SPA_FEATURE_LARGE_BLOCKS)) return (SPA_MAXBLOCKSIZE); else return (SPA_OLD_MAXBLOCKSIZE); } /* * Returns the txg that the last device removal completed. No indirect mappings * have been added since this txg. */ uint64_t spa_get_last_removal_txg(spa_t *spa) { uint64_t vdevid; uint64_t ret = -1ULL; spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER); /* * sr_prev_indirect_vdev is only modified while holding all the * config locks, so it is sufficient to hold SCL_VDEV as reader when * examining it. */ vdevid = spa->spa_removing_phys.sr_prev_indirect_vdev; while (vdevid != -1ULL) { vdev_t *vd = vdev_lookup_top(spa, vdevid); vdev_indirect_births_t *vib = vd->vdev_indirect_births; ASSERT3P(vd->vdev_ops, ==, &vdev_indirect_ops); /* * If the removal did not remap any data, we don't care. */ if (vdev_indirect_births_count(vib) != 0) { ret = vdev_indirect_births_last_entry_txg(vib); break; } vdevid = vd->vdev_indirect_config.vic_prev_indirect_vdev; } spa_config_exit(spa, SCL_VDEV, FTAG); IMPLY(ret != -1ULL, spa_feature_is_active(spa, SPA_FEATURE_DEVICE_REMOVAL)); return (ret); } int spa_maxdnodesize(spa_t *spa) { if (spa_feature_is_enabled(spa, SPA_FEATURE_LARGE_DNODE)) return (DNODE_MAX_SIZE); else return (DNODE_MIN_SIZE); } boolean_t spa_multihost(spa_t *spa) { return (spa->spa_multihost ? B_TRUE : B_FALSE); } uint32_t spa_get_hostid(spa_t *spa) { return (spa->spa_hostid); } boolean_t spa_trust_config(spa_t *spa) { return (spa->spa_trust_config); } uint64_t spa_missing_tvds_allowed(spa_t *spa) { return (spa->spa_missing_tvds_allowed); } space_map_t * spa_syncing_log_sm(spa_t *spa) { return (spa->spa_syncing_log_sm); } void spa_set_missing_tvds(spa_t *spa, uint64_t missing) { spa->spa_missing_tvds = missing; } /* * Return the pool state string ("ONLINE", "DEGRADED", "SUSPENDED", etc). */ const char * spa_state_to_name(spa_t *spa) { ASSERT3P(spa, !=, NULL); /* * it is possible for the spa to exist, without root vdev * as the spa transitions during import/export */ vdev_t *rvd = spa->spa_root_vdev; if (rvd == NULL) { return ("TRANSITIONING"); } vdev_state_t state = rvd->vdev_state; vdev_aux_t aux = rvd->vdev_stat.vs_aux; if (spa_suspended(spa) && (spa_get_failmode(spa) != ZIO_FAILURE_MODE_CONTINUE)) return ("SUSPENDED"); switch (state) { case VDEV_STATE_CLOSED: case VDEV_STATE_OFFLINE: return ("OFFLINE"); case VDEV_STATE_REMOVED: return ("REMOVED"); case VDEV_STATE_CANT_OPEN: if (aux == VDEV_AUX_CORRUPT_DATA || aux == VDEV_AUX_BAD_LOG) return ("FAULTED"); else if (aux == VDEV_AUX_SPLIT_POOL) return ("SPLIT"); else return ("UNAVAIL"); case VDEV_STATE_FAULTED: return ("FAULTED"); case VDEV_STATE_DEGRADED: return ("DEGRADED"); case VDEV_STATE_HEALTHY: return ("ONLINE"); default: break; } return ("UNKNOWN"); } boolean_t spa_top_vdevs_spacemap_addressable(spa_t *spa) { vdev_t *rvd = spa->spa_root_vdev; for (uint64_t c = 0; c < rvd->vdev_children; c++) { if (!vdev_is_spacemap_addressable(rvd->vdev_child[c])) return (B_FALSE); } return (B_TRUE); } boolean_t spa_has_checkpoint(spa_t *spa) { return (spa->spa_checkpoint_txg != 0); } boolean_t spa_importing_readonly_checkpoint(spa_t *spa) { return ((spa->spa_import_flags & ZFS_IMPORT_CHECKPOINT) && spa->spa_mode == SPA_MODE_READ); } uint64_t spa_min_claim_txg(spa_t *spa) { uint64_t checkpoint_txg = spa->spa_uberblock.ub_checkpoint_txg; if (checkpoint_txg != 0) return (checkpoint_txg + 1); return (spa->spa_first_txg); } /* * If there is a checkpoint, async destroys may consume more space from * the pool instead of freeing it. In an attempt to save the pool from * getting suspended when it is about to run out of space, we stop * processing async destroys. */ boolean_t spa_suspend_async_destroy(spa_t *spa) { dsl_pool_t *dp = spa_get_dsl(spa); uint64_t unreserved = dsl_pool_unreserved_space(dp, ZFS_SPACE_CHECK_EXTRA_RESERVED); uint64_t used = dsl_dir_phys(dp->dp_root_dir)->dd_used_bytes; uint64_t avail = (unreserved > used) ? (unreserved - used) : 0; if (spa_has_checkpoint(spa) && avail == 0) return (B_TRUE); return (B_FALSE); } #if defined(_KERNEL) int param_set_deadman_failmode_common(const char *val) { spa_t *spa = NULL; char *p; if (val == NULL) return (SET_ERROR(EINVAL)); if ((p = strchr(val, '\n')) != NULL) *p = '\0'; if (strcmp(val, "wait") != 0 && strcmp(val, "continue") != 0 && strcmp(val, "panic")) return (SET_ERROR(EINVAL)); if (spa_mode_global != SPA_MODE_UNINIT) { mutex_enter(&spa_namespace_lock); while ((spa = spa_next(spa)) != NULL) spa_set_deadman_failmode(spa, val); mutex_exit(&spa_namespace_lock); } return (0); } #endif /* Namespace manipulation */ EXPORT_SYMBOL(spa_lookup); EXPORT_SYMBOL(spa_add); EXPORT_SYMBOL(spa_remove); EXPORT_SYMBOL(spa_next); /* Refcount functions */ EXPORT_SYMBOL(spa_open_ref); EXPORT_SYMBOL(spa_close); EXPORT_SYMBOL(spa_refcount_zero); /* Pool configuration lock */ EXPORT_SYMBOL(spa_config_tryenter); EXPORT_SYMBOL(spa_config_enter); EXPORT_SYMBOL(spa_config_exit); EXPORT_SYMBOL(spa_config_held); /* Pool vdev add/remove lock */ EXPORT_SYMBOL(spa_vdev_enter); EXPORT_SYMBOL(spa_vdev_exit); /* Pool vdev state change lock */ EXPORT_SYMBOL(spa_vdev_state_enter); EXPORT_SYMBOL(spa_vdev_state_exit); /* Accessor functions */ EXPORT_SYMBOL(spa_shutting_down); EXPORT_SYMBOL(spa_get_dsl); EXPORT_SYMBOL(spa_get_rootblkptr); EXPORT_SYMBOL(spa_set_rootblkptr); EXPORT_SYMBOL(spa_altroot); EXPORT_SYMBOL(spa_sync_pass); EXPORT_SYMBOL(spa_name); EXPORT_SYMBOL(spa_guid); EXPORT_SYMBOL(spa_last_synced_txg); EXPORT_SYMBOL(spa_first_txg); EXPORT_SYMBOL(spa_syncing_txg); EXPORT_SYMBOL(spa_version); EXPORT_SYMBOL(spa_state); EXPORT_SYMBOL(spa_load_state); EXPORT_SYMBOL(spa_freeze_txg); EXPORT_SYMBOL(spa_get_dspace); EXPORT_SYMBOL(spa_update_dspace); EXPORT_SYMBOL(spa_deflate); EXPORT_SYMBOL(spa_normal_class); EXPORT_SYMBOL(spa_log_class); EXPORT_SYMBOL(spa_special_class); EXPORT_SYMBOL(spa_preferred_class); EXPORT_SYMBOL(spa_max_replication); EXPORT_SYMBOL(spa_prev_software_version); EXPORT_SYMBOL(spa_get_failmode); EXPORT_SYMBOL(spa_suspended); EXPORT_SYMBOL(spa_bootfs); EXPORT_SYMBOL(spa_delegation); EXPORT_SYMBOL(spa_meta_objset); EXPORT_SYMBOL(spa_maxblocksize); EXPORT_SYMBOL(spa_maxdnodesize); /* Miscellaneous support routines */ EXPORT_SYMBOL(spa_guid_exists); EXPORT_SYMBOL(spa_strdup); EXPORT_SYMBOL(spa_strfree); EXPORT_SYMBOL(spa_generate_guid); EXPORT_SYMBOL(snprintf_blkptr); EXPORT_SYMBOL(spa_freeze); EXPORT_SYMBOL(spa_upgrade); EXPORT_SYMBOL(spa_evict_all); EXPORT_SYMBOL(spa_lookup_by_guid); EXPORT_SYMBOL(spa_has_spare); EXPORT_SYMBOL(dva_get_dsize_sync); EXPORT_SYMBOL(bp_get_dsize_sync); EXPORT_SYMBOL(bp_get_dsize); EXPORT_SYMBOL(spa_has_slogs); EXPORT_SYMBOL(spa_is_root); EXPORT_SYMBOL(spa_writeable); EXPORT_SYMBOL(spa_mode); EXPORT_SYMBOL(spa_namespace_lock); EXPORT_SYMBOL(spa_trust_config); EXPORT_SYMBOL(spa_missing_tvds_allowed); EXPORT_SYMBOL(spa_set_missing_tvds); EXPORT_SYMBOL(spa_state_to_name); EXPORT_SYMBOL(spa_importing_readonly_checkpoint); EXPORT_SYMBOL(spa_min_claim_txg); EXPORT_SYMBOL(spa_suspend_async_destroy); EXPORT_SYMBOL(spa_has_checkpoint); EXPORT_SYMBOL(spa_top_vdevs_spacemap_addressable); ZFS_MODULE_PARAM(zfs, zfs_, flags, UINT, ZMOD_RW, "Set additional debugging flags"); ZFS_MODULE_PARAM(zfs, zfs_, recover, INT, ZMOD_RW, "Set to attempt to recover from fatal errors"); ZFS_MODULE_PARAM(zfs, zfs_, free_leak_on_eio, INT, ZMOD_RW, "Set to ignore IO errors during free and permanently leak the space"); ZFS_MODULE_PARAM(zfs_deadman, zfs_deadman_, checktime_ms, U64, ZMOD_RW, "Dead I/O check interval in milliseconds"); ZFS_MODULE_PARAM(zfs_deadman, zfs_deadman_, enabled, INT, ZMOD_RW, "Enable deadman timer"); ZFS_MODULE_PARAM(zfs_spa, spa_, asize_inflation, UINT, ZMOD_RW, "SPA size estimate multiplication factor"); ZFS_MODULE_PARAM(zfs, zfs_, ddt_data_is_special, INT, ZMOD_RW, "Place DDT data into the special class"); ZFS_MODULE_PARAM(zfs, zfs_, user_indirect_is_special, INT, ZMOD_RW, "Place user data indirect blocks into the special class"); /* BEGIN CSTYLED */ ZFS_MODULE_PARAM_CALL(zfs_deadman, zfs_deadman_, failmode, param_set_deadman_failmode, param_get_charp, ZMOD_RW, "Failmode for deadman timer"); ZFS_MODULE_PARAM_CALL(zfs_deadman, zfs_deadman_, synctime_ms, param_set_deadman_synctime, spl_param_get_u64, ZMOD_RW, "Pool sync expiration time in milliseconds"); ZFS_MODULE_PARAM_CALL(zfs_deadman, zfs_deadman_, ziotime_ms, param_set_deadman_ziotime, spl_param_get_u64, ZMOD_RW, "IO expiration time in milliseconds"); ZFS_MODULE_PARAM(zfs, zfs_, special_class_metadata_reserve_pct, UINT, ZMOD_RW, "Small file blocks in special vdevs depends on this much " "free space available"); /* END CSTYLED */ ZFS_MODULE_PARAM_CALL(zfs_spa, spa_, slop_shift, param_set_slop_shift, param_get_uint, ZMOD_RW, "Reserved free space in pool"); diff --git a/sys/contrib/openzfs/module/zfs/vdev.c b/sys/contrib/openzfs/module/zfs/vdev.c index 4bfd95861e02..c243dddb7e61 100644 --- a/sys/contrib/openzfs/module/zfs/vdev.c +++ b/sys/contrib/openzfs/module/zfs/vdev.c @@ -1,6365 +1,6372 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2011, 2021 by Delphix. All rights reserved. * Copyright 2017 Nexenta Systems, Inc. * Copyright (c) 2014 Integros [integros.com] * Copyright 2016 Toomas Soome * Copyright 2017 Joyent, Inc. * Copyright (c) 2017, Intel Corporation. * Copyright (c) 2019, Datto Inc. All rights reserved. * Copyright (c) 2021, Klara Inc. * Copyright [2021] Hewlett Packard Enterprise Development LP */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "zfs_prop.h" /* * One metaslab from each (normal-class) vdev is used by the ZIL. These are * called "embedded slog metaslabs", are referenced by vdev_log_mg, and are * part of the spa_embedded_log_class. The metaslab with the most free space * in each vdev is selected for this purpose when the pool is opened (or a * vdev is added). See vdev_metaslab_init(). * * Log blocks can be allocated from the following locations. Each one is tried * in order until the allocation succeeds: * 1. dedicated log vdevs, aka "slog" (spa_log_class) * 2. embedded slog metaslabs (spa_embedded_log_class) * 3. other metaslabs in normal vdevs (spa_normal_class) * * zfs_embedded_slog_min_ms disables the embedded slog if there are fewer * than this number of metaslabs in the vdev. This ensures that we don't set * aside an unreasonable amount of space for the ZIL. If set to less than * 1 << (spa_slop_shift + 1), on small pools the usable space may be reduced * (by more than 1<vdev_path != NULL) { zfs_dbgmsg("%s vdev '%s': %s", vd->vdev_ops->vdev_op_type, vd->vdev_path, buf); } else { zfs_dbgmsg("%s-%llu vdev (guid %llu): %s", vd->vdev_ops->vdev_op_type, (u_longlong_t)vd->vdev_id, (u_longlong_t)vd->vdev_guid, buf); } } void vdev_dbgmsg_print_tree(vdev_t *vd, int indent) { char state[20]; if (vd->vdev_ishole || vd->vdev_ops == &vdev_missing_ops) { zfs_dbgmsg("%*svdev %llu: %s", indent, "", (u_longlong_t)vd->vdev_id, vd->vdev_ops->vdev_op_type); return; } switch (vd->vdev_state) { case VDEV_STATE_UNKNOWN: (void) snprintf(state, sizeof (state), "unknown"); break; case VDEV_STATE_CLOSED: (void) snprintf(state, sizeof (state), "closed"); break; case VDEV_STATE_OFFLINE: (void) snprintf(state, sizeof (state), "offline"); break; case VDEV_STATE_REMOVED: (void) snprintf(state, sizeof (state), "removed"); break; case VDEV_STATE_CANT_OPEN: (void) snprintf(state, sizeof (state), "can't open"); break; case VDEV_STATE_FAULTED: (void) snprintf(state, sizeof (state), "faulted"); break; case VDEV_STATE_DEGRADED: (void) snprintf(state, sizeof (state), "degraded"); break; case VDEV_STATE_HEALTHY: (void) snprintf(state, sizeof (state), "healthy"); break; default: (void) snprintf(state, sizeof (state), "", (uint_t)vd->vdev_state); } zfs_dbgmsg("%*svdev %u: %s%s, guid: %llu, path: %s, %s", indent, "", (int)vd->vdev_id, vd->vdev_ops->vdev_op_type, vd->vdev_islog ? " (log)" : "", (u_longlong_t)vd->vdev_guid, vd->vdev_path ? vd->vdev_path : "N/A", state); for (uint64_t i = 0; i < vd->vdev_children; i++) vdev_dbgmsg_print_tree(vd->vdev_child[i], indent + 2); } /* * Virtual device management. */ static vdev_ops_t *const vdev_ops_table[] = { &vdev_root_ops, &vdev_raidz_ops, &vdev_draid_ops, &vdev_draid_spare_ops, &vdev_mirror_ops, &vdev_replacing_ops, &vdev_spare_ops, &vdev_disk_ops, &vdev_file_ops, &vdev_missing_ops, &vdev_hole_ops, &vdev_indirect_ops, NULL }; /* * Given a vdev type, return the appropriate ops vector. */ static vdev_ops_t * vdev_getops(const char *type) { vdev_ops_t *ops, *const *opspp; for (opspp = vdev_ops_table; (ops = *opspp) != NULL; opspp++) if (strcmp(ops->vdev_op_type, type) == 0) break; return (ops); } /* * Given a vdev and a metaslab class, find which metaslab group we're * interested in. All vdevs may belong to two different metaslab classes. * Dedicated slog devices use only the primary metaslab group, rather than a * separate log group. For embedded slogs, the vdev_log_mg will be non-NULL. */ metaslab_group_t * vdev_get_mg(vdev_t *vd, metaslab_class_t *mc) { if (mc == spa_embedded_log_class(vd->vdev_spa) && vd->vdev_log_mg != NULL) return (vd->vdev_log_mg); else return (vd->vdev_mg); } void vdev_default_xlate(vdev_t *vd, const range_seg64_t *logical_rs, range_seg64_t *physical_rs, range_seg64_t *remain_rs) { (void) vd, (void) remain_rs; physical_rs->rs_start = logical_rs->rs_start; physical_rs->rs_end = logical_rs->rs_end; } /* * Derive the enumerated allocation bias from string input. * String origin is either the per-vdev zap or zpool(8). */ static vdev_alloc_bias_t vdev_derive_alloc_bias(const char *bias) { vdev_alloc_bias_t alloc_bias = VDEV_BIAS_NONE; if (strcmp(bias, VDEV_ALLOC_BIAS_LOG) == 0) alloc_bias = VDEV_BIAS_LOG; else if (strcmp(bias, VDEV_ALLOC_BIAS_SPECIAL) == 0) alloc_bias = VDEV_BIAS_SPECIAL; else if (strcmp(bias, VDEV_ALLOC_BIAS_DEDUP) == 0) alloc_bias = VDEV_BIAS_DEDUP; return (alloc_bias); } /* * Default asize function: return the MAX of psize with the asize of * all children. This is what's used by anything other than RAID-Z. */ uint64_t vdev_default_asize(vdev_t *vd, uint64_t psize) { uint64_t asize = P2ROUNDUP(psize, 1ULL << vd->vdev_top->vdev_ashift); uint64_t csize; for (int c = 0; c < vd->vdev_children; c++) { csize = vdev_psize_to_asize(vd->vdev_child[c], psize); asize = MAX(asize, csize); } return (asize); } uint64_t vdev_default_min_asize(vdev_t *vd) { return (vd->vdev_min_asize); } /* * Get the minimum allocatable size. We define the allocatable size as * the vdev's asize rounded to the nearest metaslab. This allows us to * replace or attach devices which don't have the same physical size but * can still satisfy the same number of allocations. */ uint64_t vdev_get_min_asize(vdev_t *vd) { vdev_t *pvd = vd->vdev_parent; /* * If our parent is NULL (inactive spare or cache) or is the root, * just return our own asize. */ if (pvd == NULL) return (vd->vdev_asize); /* * The top-level vdev just returns the allocatable size rounded * to the nearest metaslab. */ if (vd == vd->vdev_top) return (P2ALIGN(vd->vdev_asize, 1ULL << vd->vdev_ms_shift)); return (pvd->vdev_ops->vdev_op_min_asize(pvd)); } void vdev_set_min_asize(vdev_t *vd) { vd->vdev_min_asize = vdev_get_min_asize(vd); for (int c = 0; c < vd->vdev_children; c++) vdev_set_min_asize(vd->vdev_child[c]); } /* * Get the minimal allocation size for the top-level vdev. */ uint64_t vdev_get_min_alloc(vdev_t *vd) { uint64_t min_alloc = 1ULL << vd->vdev_ashift; if (vd->vdev_ops->vdev_op_min_alloc != NULL) min_alloc = vd->vdev_ops->vdev_op_min_alloc(vd); return (min_alloc); } /* * Get the parity level for a top-level vdev. */ uint64_t vdev_get_nparity(vdev_t *vd) { uint64_t nparity = 0; if (vd->vdev_ops->vdev_op_nparity != NULL) nparity = vd->vdev_ops->vdev_op_nparity(vd); return (nparity); } static int vdev_prop_get_int(vdev_t *vd, vdev_prop_t prop, uint64_t *value) { spa_t *spa = vd->vdev_spa; objset_t *mos = spa->spa_meta_objset; uint64_t objid; int err; if (vd->vdev_root_zap != 0) { objid = vd->vdev_root_zap; } else if (vd->vdev_top_zap != 0) { objid = vd->vdev_top_zap; } else if (vd->vdev_leaf_zap != 0) { objid = vd->vdev_leaf_zap; } else { return (EINVAL); } err = zap_lookup(mos, objid, vdev_prop_to_name(prop), sizeof (uint64_t), 1, value); if (err == ENOENT) *value = vdev_prop_default_numeric(prop); return (err); } /* * Get the number of data disks for a top-level vdev. */ uint64_t vdev_get_ndisks(vdev_t *vd) { uint64_t ndisks = 1; if (vd->vdev_ops->vdev_op_ndisks != NULL) ndisks = vd->vdev_ops->vdev_op_ndisks(vd); return (ndisks); } vdev_t * vdev_lookup_top(spa_t *spa, uint64_t vdev) { vdev_t *rvd = spa->spa_root_vdev; ASSERT(spa_config_held(spa, SCL_ALL, RW_READER) != 0); if (vdev < rvd->vdev_children) { ASSERT(rvd->vdev_child[vdev] != NULL); return (rvd->vdev_child[vdev]); } return (NULL); } vdev_t * vdev_lookup_by_guid(vdev_t *vd, uint64_t guid) { vdev_t *mvd; if (vd->vdev_guid == guid) return (vd); for (int c = 0; c < vd->vdev_children; c++) if ((mvd = vdev_lookup_by_guid(vd->vdev_child[c], guid)) != NULL) return (mvd); return (NULL); } static int vdev_count_leaves_impl(vdev_t *vd) { int n = 0; if (vd->vdev_ops->vdev_op_leaf) return (1); for (int c = 0; c < vd->vdev_children; c++) n += vdev_count_leaves_impl(vd->vdev_child[c]); return (n); } int vdev_count_leaves(spa_t *spa) { int rc; spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER); rc = vdev_count_leaves_impl(spa->spa_root_vdev); spa_config_exit(spa, SCL_VDEV, FTAG); return (rc); } void vdev_add_child(vdev_t *pvd, vdev_t *cvd) { size_t oldsize, newsize; uint64_t id = cvd->vdev_id; vdev_t **newchild; ASSERT(spa_config_held(cvd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL); ASSERT(cvd->vdev_parent == NULL); cvd->vdev_parent = pvd; if (pvd == NULL) return; ASSERT(id >= pvd->vdev_children || pvd->vdev_child[id] == NULL); oldsize = pvd->vdev_children * sizeof (vdev_t *); pvd->vdev_children = MAX(pvd->vdev_children, id + 1); newsize = pvd->vdev_children * sizeof (vdev_t *); newchild = kmem_alloc(newsize, KM_SLEEP); if (pvd->vdev_child != NULL) { memcpy(newchild, pvd->vdev_child, oldsize); kmem_free(pvd->vdev_child, oldsize); } pvd->vdev_child = newchild; pvd->vdev_child[id] = cvd; cvd->vdev_top = (pvd->vdev_top ? pvd->vdev_top: cvd); ASSERT(cvd->vdev_top->vdev_parent->vdev_parent == NULL); /* * Walk up all ancestors to update guid sum. */ for (; pvd != NULL; pvd = pvd->vdev_parent) pvd->vdev_guid_sum += cvd->vdev_guid_sum; if (cvd->vdev_ops->vdev_op_leaf) { list_insert_head(&cvd->vdev_spa->spa_leaf_list, cvd); cvd->vdev_spa->spa_leaf_list_gen++; } } void vdev_remove_child(vdev_t *pvd, vdev_t *cvd) { int c; uint_t id = cvd->vdev_id; ASSERT(cvd->vdev_parent == pvd); if (pvd == NULL) return; ASSERT(id < pvd->vdev_children); ASSERT(pvd->vdev_child[id] == cvd); pvd->vdev_child[id] = NULL; cvd->vdev_parent = NULL; for (c = 0; c < pvd->vdev_children; c++) if (pvd->vdev_child[c]) break; if (c == pvd->vdev_children) { kmem_free(pvd->vdev_child, c * sizeof (vdev_t *)); pvd->vdev_child = NULL; pvd->vdev_children = 0; } if (cvd->vdev_ops->vdev_op_leaf) { spa_t *spa = cvd->vdev_spa; list_remove(&spa->spa_leaf_list, cvd); spa->spa_leaf_list_gen++; } /* * Walk up all ancestors to update guid sum. */ for (; pvd != NULL; pvd = pvd->vdev_parent) pvd->vdev_guid_sum -= cvd->vdev_guid_sum; } /* * Remove any holes in the child array. */ void vdev_compact_children(vdev_t *pvd) { vdev_t **newchild, *cvd; int oldc = pvd->vdev_children; int newc; ASSERT(spa_config_held(pvd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL); if (oldc == 0) return; for (int c = newc = 0; c < oldc; c++) if (pvd->vdev_child[c]) newc++; if (newc > 0) { newchild = kmem_zalloc(newc * sizeof (vdev_t *), KM_SLEEP); for (int c = newc = 0; c < oldc; c++) { if ((cvd = pvd->vdev_child[c]) != NULL) { newchild[newc] = cvd; cvd->vdev_id = newc++; } } } else { newchild = NULL; } kmem_free(pvd->vdev_child, oldc * sizeof (vdev_t *)); pvd->vdev_child = newchild; pvd->vdev_children = newc; } /* * Allocate and minimally initialize a vdev_t. */ vdev_t * vdev_alloc_common(spa_t *spa, uint_t id, uint64_t guid, vdev_ops_t *ops) { vdev_t *vd; vdev_indirect_config_t *vic; vd = kmem_zalloc(sizeof (vdev_t), KM_SLEEP); vic = &vd->vdev_indirect_config; if (spa->spa_root_vdev == NULL) { ASSERT(ops == &vdev_root_ops); spa->spa_root_vdev = vd; spa->spa_load_guid = spa_generate_guid(NULL); } if (guid == 0 && ops != &vdev_hole_ops) { if (spa->spa_root_vdev == vd) { /* * The root vdev's guid will also be the pool guid, * which must be unique among all pools. */ guid = spa_generate_guid(NULL); } else { /* * Any other vdev's guid must be unique within the pool. */ guid = spa_generate_guid(spa); } ASSERT(!spa_guid_exists(spa_guid(spa), guid)); } vd->vdev_spa = spa; vd->vdev_id = id; vd->vdev_guid = guid; vd->vdev_guid_sum = guid; vd->vdev_ops = ops; vd->vdev_state = VDEV_STATE_CLOSED; vd->vdev_ishole = (ops == &vdev_hole_ops); vic->vic_prev_indirect_vdev = UINT64_MAX; rw_init(&vd->vdev_indirect_rwlock, NULL, RW_DEFAULT, NULL); mutex_init(&vd->vdev_obsolete_lock, NULL, MUTEX_DEFAULT, NULL); vd->vdev_obsolete_segments = range_tree_create(NULL, RANGE_SEG64, NULL, 0, 0); /* * Initialize rate limit structs for events. We rate limit ZIO delay * and checksum events so that we don't overwhelm ZED with thousands * of events when a disk is acting up. */ zfs_ratelimit_init(&vd->vdev_delay_rl, &zfs_slow_io_events_per_second, 1); zfs_ratelimit_init(&vd->vdev_deadman_rl, &zfs_slow_io_events_per_second, 1); zfs_ratelimit_init(&vd->vdev_checksum_rl, &zfs_checksum_events_per_second, 1); /* * Default Thresholds for tuning ZED */ vd->vdev_checksum_n = vdev_prop_default_numeric(VDEV_PROP_CHECKSUM_N); vd->vdev_checksum_t = vdev_prop_default_numeric(VDEV_PROP_CHECKSUM_T); vd->vdev_io_n = vdev_prop_default_numeric(VDEV_PROP_IO_N); vd->vdev_io_t = vdev_prop_default_numeric(VDEV_PROP_IO_T); list_link_init(&vd->vdev_config_dirty_node); list_link_init(&vd->vdev_state_dirty_node); list_link_init(&vd->vdev_initialize_node); list_link_init(&vd->vdev_leaf_node); list_link_init(&vd->vdev_trim_node); mutex_init(&vd->vdev_dtl_lock, NULL, MUTEX_NOLOCKDEP, NULL); mutex_init(&vd->vdev_stat_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&vd->vdev_probe_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&vd->vdev_scan_io_queue_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&vd->vdev_initialize_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&vd->vdev_initialize_io_lock, NULL, MUTEX_DEFAULT, NULL); cv_init(&vd->vdev_initialize_cv, NULL, CV_DEFAULT, NULL); cv_init(&vd->vdev_initialize_io_cv, NULL, CV_DEFAULT, NULL); mutex_init(&vd->vdev_trim_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&vd->vdev_autotrim_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&vd->vdev_trim_io_lock, NULL, MUTEX_DEFAULT, NULL); cv_init(&vd->vdev_trim_cv, NULL, CV_DEFAULT, NULL); cv_init(&vd->vdev_autotrim_cv, NULL, CV_DEFAULT, NULL); cv_init(&vd->vdev_autotrim_kick_cv, NULL, CV_DEFAULT, NULL); cv_init(&vd->vdev_trim_io_cv, NULL, CV_DEFAULT, NULL); mutex_init(&vd->vdev_rebuild_lock, NULL, MUTEX_DEFAULT, NULL); cv_init(&vd->vdev_rebuild_cv, NULL, CV_DEFAULT, NULL); for (int t = 0; t < DTL_TYPES; t++) { vd->vdev_dtl[t] = range_tree_create(NULL, RANGE_SEG64, NULL, 0, 0); } txg_list_create(&vd->vdev_ms_list, spa, offsetof(struct metaslab, ms_txg_node)); txg_list_create(&vd->vdev_dtl_list, spa, offsetof(struct vdev, vdev_dtl_node)); vd->vdev_stat.vs_timestamp = gethrtime(); vdev_queue_init(vd); vdev_cache_init(vd); return (vd); } /* * Allocate a new vdev. The 'alloctype' is used to control whether we are * creating a new vdev or loading an existing one - the behavior is slightly * different for each case. */ int vdev_alloc(spa_t *spa, vdev_t **vdp, nvlist_t *nv, vdev_t *parent, uint_t id, int alloctype) { vdev_ops_t *ops; const char *type; uint64_t guid = 0, islog; vdev_t *vd; vdev_indirect_config_t *vic; const char *tmp = NULL; int rc; vdev_alloc_bias_t alloc_bias = VDEV_BIAS_NONE; boolean_t top_level = (parent && !parent->vdev_parent); ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL); if (nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &type) != 0) return (SET_ERROR(EINVAL)); if ((ops = vdev_getops(type)) == NULL) return (SET_ERROR(EINVAL)); /* * If this is a load, get the vdev guid from the nvlist. * Otherwise, vdev_alloc_common() will generate one for us. */ if (alloctype == VDEV_ALLOC_LOAD) { uint64_t label_id; if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_ID, &label_id) || label_id != id) return (SET_ERROR(EINVAL)); if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0) return (SET_ERROR(EINVAL)); } else if (alloctype == VDEV_ALLOC_SPARE) { if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0) return (SET_ERROR(EINVAL)); } else if (alloctype == VDEV_ALLOC_L2CACHE) { if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0) return (SET_ERROR(EINVAL)); } else if (alloctype == VDEV_ALLOC_ROOTPOOL) { if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &guid) != 0) return (SET_ERROR(EINVAL)); } /* * The first allocated vdev must be of type 'root'. */ if (ops != &vdev_root_ops && spa->spa_root_vdev == NULL) return (SET_ERROR(EINVAL)); /* * Determine whether we're a log vdev. */ islog = 0; (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_IS_LOG, &islog); if (islog && spa_version(spa) < SPA_VERSION_SLOGS) return (SET_ERROR(ENOTSUP)); if (ops == &vdev_hole_ops && spa_version(spa) < SPA_VERSION_HOLES) return (SET_ERROR(ENOTSUP)); if (top_level && alloctype == VDEV_ALLOC_ADD) { const char *bias; /* * If creating a top-level vdev, check for allocation * classes input. */ if (nvlist_lookup_string(nv, ZPOOL_CONFIG_ALLOCATION_BIAS, &bias) == 0) { alloc_bias = vdev_derive_alloc_bias(bias); /* spa_vdev_add() expects feature to be enabled */ if (spa->spa_load_state != SPA_LOAD_CREATE && !spa_feature_is_enabled(spa, SPA_FEATURE_ALLOCATION_CLASSES)) { return (SET_ERROR(ENOTSUP)); } } /* spa_vdev_add() expects feature to be enabled */ if (ops == &vdev_draid_ops && spa->spa_load_state != SPA_LOAD_CREATE && !spa_feature_is_enabled(spa, SPA_FEATURE_DRAID)) { return (SET_ERROR(ENOTSUP)); } } /* * Initialize the vdev specific data. This is done before calling * vdev_alloc_common() since it may fail and this simplifies the * error reporting and cleanup code paths. */ void *tsd = NULL; if (ops->vdev_op_init != NULL) { rc = ops->vdev_op_init(spa, nv, &tsd); if (rc != 0) { return (rc); } } vd = vdev_alloc_common(spa, id, guid, ops); vd->vdev_tsd = tsd; vd->vdev_islog = islog; if (top_level && alloc_bias != VDEV_BIAS_NONE) vd->vdev_alloc_bias = alloc_bias; if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &tmp) == 0) vd->vdev_path = spa_strdup(tmp); /* * ZPOOL_CONFIG_AUX_STATE = "external" means we previously forced a * fault on a vdev and want it to persist across imports (like with * zpool offline -f). */ rc = nvlist_lookup_string(nv, ZPOOL_CONFIG_AUX_STATE, &tmp); if (rc == 0 && tmp != NULL && strcmp(tmp, "external") == 0) { vd->vdev_stat.vs_aux = VDEV_AUX_EXTERNAL; vd->vdev_faulted = 1; vd->vdev_label_aux = VDEV_AUX_EXTERNAL; } if (nvlist_lookup_string(nv, ZPOOL_CONFIG_DEVID, &tmp) == 0) vd->vdev_devid = spa_strdup(tmp); if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PHYS_PATH, &tmp) == 0) vd->vdev_physpath = spa_strdup(tmp); if (nvlist_lookup_string(nv, ZPOOL_CONFIG_VDEV_ENC_SYSFS_PATH, &tmp) == 0) vd->vdev_enc_sysfs_path = spa_strdup(tmp); if (nvlist_lookup_string(nv, ZPOOL_CONFIG_FRU, &tmp) == 0) vd->vdev_fru = spa_strdup(tmp); /* * Set the whole_disk property. If it's not specified, leave the value * as -1. */ if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_WHOLE_DISK, &vd->vdev_wholedisk) != 0) vd->vdev_wholedisk = -1ULL; vic = &vd->vdev_indirect_config; ASSERT0(vic->vic_mapping_object); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_INDIRECT_OBJECT, &vic->vic_mapping_object); ASSERT0(vic->vic_births_object); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_INDIRECT_BIRTHS, &vic->vic_births_object); ASSERT3U(vic->vic_prev_indirect_vdev, ==, UINT64_MAX); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_PREV_INDIRECT_VDEV, &vic->vic_prev_indirect_vdev); /* * Look for the 'not present' flag. This will only be set if the device * was not present at the time of import. */ (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT, &vd->vdev_not_present); /* * Get the alignment requirement. */ (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_ASHIFT, &vd->vdev_ashift); /* * Retrieve the vdev creation time. */ (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_CREATE_TXG, &vd->vdev_crtxg); if (vd->vdev_ops == &vdev_root_ops && (alloctype == VDEV_ALLOC_LOAD || alloctype == VDEV_ALLOC_SPLIT || alloctype == VDEV_ALLOC_ROOTPOOL)) { (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_VDEV_ROOT_ZAP, &vd->vdev_root_zap); } /* * If we're a top-level vdev, try to load the allocation parameters. */ if (top_level && (alloctype == VDEV_ALLOC_LOAD || alloctype == VDEV_ALLOC_SPLIT)) { (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_METASLAB_ARRAY, &vd->vdev_ms_array); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_METASLAB_SHIFT, &vd->vdev_ms_shift); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_ASIZE, &vd->vdev_asize); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NONALLOCATING, &vd->vdev_noalloc); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_REMOVING, &vd->vdev_removing); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_VDEV_TOP_ZAP, &vd->vdev_top_zap); } else { ASSERT0(vd->vdev_top_zap); } if (top_level && alloctype != VDEV_ALLOC_ATTACH) { ASSERT(alloctype == VDEV_ALLOC_LOAD || alloctype == VDEV_ALLOC_ADD || alloctype == VDEV_ALLOC_SPLIT || alloctype == VDEV_ALLOC_ROOTPOOL); /* Note: metaslab_group_create() is now deferred */ } if (vd->vdev_ops->vdev_op_leaf && (alloctype == VDEV_ALLOC_LOAD || alloctype == VDEV_ALLOC_SPLIT)) { (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_VDEV_LEAF_ZAP, &vd->vdev_leaf_zap); } else { ASSERT0(vd->vdev_leaf_zap); } /* * If we're a leaf vdev, try to load the DTL object and other state. */ if (vd->vdev_ops->vdev_op_leaf && (alloctype == VDEV_ALLOC_LOAD || alloctype == VDEV_ALLOC_L2CACHE || alloctype == VDEV_ALLOC_ROOTPOOL)) { if (alloctype == VDEV_ALLOC_LOAD) { (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_DTL, &vd->vdev_dtl_object); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_UNSPARE, &vd->vdev_unspare); } if (alloctype == VDEV_ALLOC_ROOTPOOL) { uint64_t spare = 0; if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_IS_SPARE, &spare) == 0 && spare) spa_spare_add(vd); } (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_OFFLINE, &vd->vdev_offline); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_RESILVER_TXG, &vd->vdev_resilver_txg); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_REBUILD_TXG, &vd->vdev_rebuild_txg); if (nvlist_exists(nv, ZPOOL_CONFIG_RESILVER_DEFER)) vdev_defer_resilver(vd); /* * In general, when importing a pool we want to ignore the * persistent fault state, as the diagnosis made on another * system may not be valid in the current context. The only * exception is if we forced a vdev to a persistently faulted * state with 'zpool offline -f'. The persistent fault will * remain across imports until cleared. * * Local vdevs will remain in the faulted state. */ if (spa_load_state(spa) == SPA_LOAD_OPEN || spa_load_state(spa) == SPA_LOAD_IMPORT) { (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_FAULTED, &vd->vdev_faulted); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_DEGRADED, &vd->vdev_degraded); (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_REMOVED, &vd->vdev_removed); if (vd->vdev_faulted || vd->vdev_degraded) { const char *aux; vd->vdev_label_aux = VDEV_AUX_ERR_EXCEEDED; if (nvlist_lookup_string(nv, ZPOOL_CONFIG_AUX_STATE, &aux) == 0 && strcmp(aux, "external") == 0) vd->vdev_label_aux = VDEV_AUX_EXTERNAL; else vd->vdev_faulted = 0ULL; } } } /* * Add ourselves to the parent's list of children. */ vdev_add_child(parent, vd); *vdp = vd; return (0); } void vdev_free(vdev_t *vd) { spa_t *spa = vd->vdev_spa; ASSERT3P(vd->vdev_initialize_thread, ==, NULL); ASSERT3P(vd->vdev_trim_thread, ==, NULL); ASSERT3P(vd->vdev_autotrim_thread, ==, NULL); ASSERT3P(vd->vdev_rebuild_thread, ==, NULL); /* * Scan queues are normally destroyed at the end of a scan. If the * queue exists here, that implies the vdev is being removed while * the scan is still running. */ if (vd->vdev_scan_io_queue != NULL) { mutex_enter(&vd->vdev_scan_io_queue_lock); dsl_scan_io_queue_destroy(vd->vdev_scan_io_queue); vd->vdev_scan_io_queue = NULL; mutex_exit(&vd->vdev_scan_io_queue_lock); } /* * vdev_free() implies closing the vdev first. This is simpler than * trying to ensure complicated semantics for all callers. */ vdev_close(vd); ASSERT(!list_link_active(&vd->vdev_config_dirty_node)); ASSERT(!list_link_active(&vd->vdev_state_dirty_node)); /* * Free all children. */ for (int c = 0; c < vd->vdev_children; c++) vdev_free(vd->vdev_child[c]); ASSERT(vd->vdev_child == NULL); ASSERT(vd->vdev_guid_sum == vd->vdev_guid); if (vd->vdev_ops->vdev_op_fini != NULL) vd->vdev_ops->vdev_op_fini(vd); /* * Discard allocation state. */ if (vd->vdev_mg != NULL) { vdev_metaslab_fini(vd); metaslab_group_destroy(vd->vdev_mg); vd->vdev_mg = NULL; } if (vd->vdev_log_mg != NULL) { ASSERT0(vd->vdev_ms_count); metaslab_group_destroy(vd->vdev_log_mg); vd->vdev_log_mg = NULL; } ASSERT0(vd->vdev_stat.vs_space); ASSERT0(vd->vdev_stat.vs_dspace); ASSERT0(vd->vdev_stat.vs_alloc); /* * Remove this vdev from its parent's child list. */ vdev_remove_child(vd->vdev_parent, vd); ASSERT(vd->vdev_parent == NULL); ASSERT(!list_link_active(&vd->vdev_leaf_node)); /* * Clean up vdev structure. */ vdev_queue_fini(vd); vdev_cache_fini(vd); if (vd->vdev_path) spa_strfree(vd->vdev_path); if (vd->vdev_devid) spa_strfree(vd->vdev_devid); if (vd->vdev_physpath) spa_strfree(vd->vdev_physpath); if (vd->vdev_enc_sysfs_path) spa_strfree(vd->vdev_enc_sysfs_path); if (vd->vdev_fru) spa_strfree(vd->vdev_fru); if (vd->vdev_isspare) spa_spare_remove(vd); if (vd->vdev_isl2cache) spa_l2cache_remove(vd); txg_list_destroy(&vd->vdev_ms_list); txg_list_destroy(&vd->vdev_dtl_list); mutex_enter(&vd->vdev_dtl_lock); space_map_close(vd->vdev_dtl_sm); for (int t = 0; t < DTL_TYPES; t++) { range_tree_vacate(vd->vdev_dtl[t], NULL, NULL); range_tree_destroy(vd->vdev_dtl[t]); } mutex_exit(&vd->vdev_dtl_lock); EQUIV(vd->vdev_indirect_births != NULL, vd->vdev_indirect_mapping != NULL); if (vd->vdev_indirect_births != NULL) { vdev_indirect_mapping_close(vd->vdev_indirect_mapping); vdev_indirect_births_close(vd->vdev_indirect_births); } if (vd->vdev_obsolete_sm != NULL) { ASSERT(vd->vdev_removing || vd->vdev_ops == &vdev_indirect_ops); space_map_close(vd->vdev_obsolete_sm); vd->vdev_obsolete_sm = NULL; } range_tree_destroy(vd->vdev_obsolete_segments); rw_destroy(&vd->vdev_indirect_rwlock); mutex_destroy(&vd->vdev_obsolete_lock); mutex_destroy(&vd->vdev_dtl_lock); mutex_destroy(&vd->vdev_stat_lock); mutex_destroy(&vd->vdev_probe_lock); mutex_destroy(&vd->vdev_scan_io_queue_lock); mutex_destroy(&vd->vdev_initialize_lock); mutex_destroy(&vd->vdev_initialize_io_lock); cv_destroy(&vd->vdev_initialize_io_cv); cv_destroy(&vd->vdev_initialize_cv); mutex_destroy(&vd->vdev_trim_lock); mutex_destroy(&vd->vdev_autotrim_lock); mutex_destroy(&vd->vdev_trim_io_lock); cv_destroy(&vd->vdev_trim_cv); cv_destroy(&vd->vdev_autotrim_cv); cv_destroy(&vd->vdev_autotrim_kick_cv); cv_destroy(&vd->vdev_trim_io_cv); mutex_destroy(&vd->vdev_rebuild_lock); cv_destroy(&vd->vdev_rebuild_cv); zfs_ratelimit_fini(&vd->vdev_delay_rl); zfs_ratelimit_fini(&vd->vdev_deadman_rl); zfs_ratelimit_fini(&vd->vdev_checksum_rl); if (vd == spa->spa_root_vdev) spa->spa_root_vdev = NULL; kmem_free(vd, sizeof (vdev_t)); } /* * Transfer top-level vdev state from svd to tvd. */ static void vdev_top_transfer(vdev_t *svd, vdev_t *tvd) { spa_t *spa = svd->vdev_spa; metaslab_t *msp; vdev_t *vd; int t; ASSERT(tvd == tvd->vdev_top); tvd->vdev_pending_fastwrite = svd->vdev_pending_fastwrite; tvd->vdev_ms_array = svd->vdev_ms_array; tvd->vdev_ms_shift = svd->vdev_ms_shift; tvd->vdev_ms_count = svd->vdev_ms_count; tvd->vdev_top_zap = svd->vdev_top_zap; svd->vdev_ms_array = 0; svd->vdev_ms_shift = 0; svd->vdev_ms_count = 0; svd->vdev_top_zap = 0; if (tvd->vdev_mg) ASSERT3P(tvd->vdev_mg, ==, svd->vdev_mg); if (tvd->vdev_log_mg) ASSERT3P(tvd->vdev_log_mg, ==, svd->vdev_log_mg); tvd->vdev_mg = svd->vdev_mg; tvd->vdev_log_mg = svd->vdev_log_mg; tvd->vdev_ms = svd->vdev_ms; svd->vdev_mg = NULL; svd->vdev_log_mg = NULL; svd->vdev_ms = NULL; if (tvd->vdev_mg != NULL) tvd->vdev_mg->mg_vd = tvd; if (tvd->vdev_log_mg != NULL) tvd->vdev_log_mg->mg_vd = tvd; tvd->vdev_checkpoint_sm = svd->vdev_checkpoint_sm; svd->vdev_checkpoint_sm = NULL; tvd->vdev_alloc_bias = svd->vdev_alloc_bias; svd->vdev_alloc_bias = VDEV_BIAS_NONE; tvd->vdev_stat.vs_alloc = svd->vdev_stat.vs_alloc; tvd->vdev_stat.vs_space = svd->vdev_stat.vs_space; tvd->vdev_stat.vs_dspace = svd->vdev_stat.vs_dspace; svd->vdev_stat.vs_alloc = 0; svd->vdev_stat.vs_space = 0; svd->vdev_stat.vs_dspace = 0; /* * State which may be set on a top-level vdev that's in the * process of being removed. */ ASSERT0(tvd->vdev_indirect_config.vic_births_object); ASSERT0(tvd->vdev_indirect_config.vic_mapping_object); ASSERT3U(tvd->vdev_indirect_config.vic_prev_indirect_vdev, ==, -1ULL); ASSERT3P(tvd->vdev_indirect_mapping, ==, NULL); ASSERT3P(tvd->vdev_indirect_births, ==, NULL); ASSERT3P(tvd->vdev_obsolete_sm, ==, NULL); ASSERT0(tvd->vdev_noalloc); ASSERT0(tvd->vdev_removing); ASSERT0(tvd->vdev_rebuilding); tvd->vdev_noalloc = svd->vdev_noalloc; tvd->vdev_removing = svd->vdev_removing; tvd->vdev_rebuilding = svd->vdev_rebuilding; tvd->vdev_rebuild_config = svd->vdev_rebuild_config; tvd->vdev_indirect_config = svd->vdev_indirect_config; tvd->vdev_indirect_mapping = svd->vdev_indirect_mapping; tvd->vdev_indirect_births = svd->vdev_indirect_births; range_tree_swap(&svd->vdev_obsolete_segments, &tvd->vdev_obsolete_segments); tvd->vdev_obsolete_sm = svd->vdev_obsolete_sm; svd->vdev_indirect_config.vic_mapping_object = 0; svd->vdev_indirect_config.vic_births_object = 0; svd->vdev_indirect_config.vic_prev_indirect_vdev = -1ULL; svd->vdev_indirect_mapping = NULL; svd->vdev_indirect_births = NULL; svd->vdev_obsolete_sm = NULL; svd->vdev_noalloc = 0; svd->vdev_removing = 0; svd->vdev_rebuilding = 0; for (t = 0; t < TXG_SIZE; t++) { while ((msp = txg_list_remove(&svd->vdev_ms_list, t)) != NULL) (void) txg_list_add(&tvd->vdev_ms_list, msp, t); while ((vd = txg_list_remove(&svd->vdev_dtl_list, t)) != NULL) (void) txg_list_add(&tvd->vdev_dtl_list, vd, t); if (txg_list_remove_this(&spa->spa_vdev_txg_list, svd, t)) (void) txg_list_add(&spa->spa_vdev_txg_list, tvd, t); } if (list_link_active(&svd->vdev_config_dirty_node)) { vdev_config_clean(svd); vdev_config_dirty(tvd); } if (list_link_active(&svd->vdev_state_dirty_node)) { vdev_state_clean(svd); vdev_state_dirty(tvd); } tvd->vdev_deflate_ratio = svd->vdev_deflate_ratio; svd->vdev_deflate_ratio = 0; tvd->vdev_islog = svd->vdev_islog; svd->vdev_islog = 0; dsl_scan_io_queue_vdev_xfer(svd, tvd); } static void vdev_top_update(vdev_t *tvd, vdev_t *vd) { if (vd == NULL) return; vd->vdev_top = tvd; for (int c = 0; c < vd->vdev_children; c++) vdev_top_update(tvd, vd->vdev_child[c]); } /* * Add a mirror/replacing vdev above an existing vdev. There is no need to * call .vdev_op_init() since mirror/replacing vdevs do not have private state. */ vdev_t * vdev_add_parent(vdev_t *cvd, vdev_ops_t *ops) { spa_t *spa = cvd->vdev_spa; vdev_t *pvd = cvd->vdev_parent; vdev_t *mvd; ASSERT(spa_config_held(spa, SCL_ALL, RW_WRITER) == SCL_ALL); mvd = vdev_alloc_common(spa, cvd->vdev_id, 0, ops); mvd->vdev_asize = cvd->vdev_asize; mvd->vdev_min_asize = cvd->vdev_min_asize; mvd->vdev_max_asize = cvd->vdev_max_asize; mvd->vdev_psize = cvd->vdev_psize; mvd->vdev_ashift = cvd->vdev_ashift; mvd->vdev_logical_ashift = cvd->vdev_logical_ashift; mvd->vdev_physical_ashift = cvd->vdev_physical_ashift; mvd->vdev_state = cvd->vdev_state; mvd->vdev_crtxg = cvd->vdev_crtxg; vdev_remove_child(pvd, cvd); vdev_add_child(pvd, mvd); cvd->vdev_id = mvd->vdev_children; vdev_add_child(mvd, cvd); vdev_top_update(cvd->vdev_top, cvd->vdev_top); if (mvd == mvd->vdev_top) vdev_top_transfer(cvd, mvd); return (mvd); } /* * Remove a 1-way mirror/replacing vdev from the tree. */ void vdev_remove_parent(vdev_t *cvd) { vdev_t *mvd = cvd->vdev_parent; vdev_t *pvd = mvd->vdev_parent; ASSERT(spa_config_held(cvd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL); ASSERT(mvd->vdev_children == 1); ASSERT(mvd->vdev_ops == &vdev_mirror_ops || mvd->vdev_ops == &vdev_replacing_ops || mvd->vdev_ops == &vdev_spare_ops); cvd->vdev_ashift = mvd->vdev_ashift; cvd->vdev_logical_ashift = mvd->vdev_logical_ashift; cvd->vdev_physical_ashift = mvd->vdev_physical_ashift; vdev_remove_child(mvd, cvd); vdev_remove_child(pvd, mvd); /* * If cvd will replace mvd as a top-level vdev, preserve mvd's guid. * Otherwise, we could have detached an offline device, and when we * go to import the pool we'll think we have two top-level vdevs, * instead of a different version of the same top-level vdev. */ if (mvd->vdev_top == mvd) { uint64_t guid_delta = mvd->vdev_guid - cvd->vdev_guid; cvd->vdev_orig_guid = cvd->vdev_guid; cvd->vdev_guid += guid_delta; cvd->vdev_guid_sum += guid_delta; /* * If pool not set for autoexpand, we need to also preserve * mvd's asize to prevent automatic expansion of cvd. * Otherwise if we are adjusting the mirror by attaching and * detaching children of non-uniform sizes, the mirror could * autoexpand, unexpectedly requiring larger devices to * re-establish the mirror. */ if (!cvd->vdev_spa->spa_autoexpand) cvd->vdev_asize = mvd->vdev_asize; } cvd->vdev_id = mvd->vdev_id; vdev_add_child(pvd, cvd); vdev_top_update(cvd->vdev_top, cvd->vdev_top); if (cvd == cvd->vdev_top) vdev_top_transfer(mvd, cvd); ASSERT(mvd->vdev_children == 0); vdev_free(mvd); } void vdev_metaslab_group_create(vdev_t *vd) { spa_t *spa = vd->vdev_spa; /* * metaslab_group_create was delayed until allocation bias was available */ if (vd->vdev_mg == NULL) { metaslab_class_t *mc; if (vd->vdev_islog && vd->vdev_alloc_bias == VDEV_BIAS_NONE) vd->vdev_alloc_bias = VDEV_BIAS_LOG; ASSERT3U(vd->vdev_islog, ==, (vd->vdev_alloc_bias == VDEV_BIAS_LOG)); switch (vd->vdev_alloc_bias) { case VDEV_BIAS_LOG: mc = spa_log_class(spa); break; case VDEV_BIAS_SPECIAL: mc = spa_special_class(spa); break; case VDEV_BIAS_DEDUP: mc = spa_dedup_class(spa); break; default: mc = spa_normal_class(spa); } vd->vdev_mg = metaslab_group_create(mc, vd, spa->spa_alloc_count); if (!vd->vdev_islog) { vd->vdev_log_mg = metaslab_group_create( spa_embedded_log_class(spa), vd, 1); } /* * The spa ashift min/max only apply for the normal metaslab * class. Class destination is late binding so ashift boundary * setting had to wait until now. */ if (vd->vdev_top == vd && vd->vdev_ashift != 0 && mc == spa_normal_class(spa) && vd->vdev_aux == NULL) { if (vd->vdev_ashift > spa->spa_max_ashift) spa->spa_max_ashift = vd->vdev_ashift; if (vd->vdev_ashift < spa->spa_min_ashift) spa->spa_min_ashift = vd->vdev_ashift; uint64_t min_alloc = vdev_get_min_alloc(vd); if (min_alloc < spa->spa_min_alloc) spa->spa_min_alloc = min_alloc; } } } int vdev_metaslab_init(vdev_t *vd, uint64_t txg) { spa_t *spa = vd->vdev_spa; uint64_t oldc = vd->vdev_ms_count; uint64_t newc = vd->vdev_asize >> vd->vdev_ms_shift; metaslab_t **mspp; int error; boolean_t expanding = (oldc != 0); ASSERT(txg == 0 || spa_config_held(spa, SCL_ALLOC, RW_WRITER)); /* * This vdev is not being allocated from yet or is a hole. */ if (vd->vdev_ms_shift == 0) return (0); ASSERT(!vd->vdev_ishole); ASSERT(oldc <= newc); mspp = vmem_zalloc(newc * sizeof (*mspp), KM_SLEEP); if (expanding) { memcpy(mspp, vd->vdev_ms, oldc * sizeof (*mspp)); vmem_free(vd->vdev_ms, oldc * sizeof (*mspp)); } vd->vdev_ms = mspp; vd->vdev_ms_count = newc; for (uint64_t m = oldc; m < newc; m++) { uint64_t object = 0; /* * vdev_ms_array may be 0 if we are creating the "fake" * metaslabs for an indirect vdev for zdb's leak detection. * See zdb_leak_init(). */ if (txg == 0 && vd->vdev_ms_array != 0) { error = dmu_read(spa->spa_meta_objset, vd->vdev_ms_array, m * sizeof (uint64_t), sizeof (uint64_t), &object, DMU_READ_PREFETCH); if (error != 0) { vdev_dbgmsg(vd, "unable to read the metaslab " "array [error=%d]", error); return (error); } } error = metaslab_init(vd->vdev_mg, m, object, txg, &(vd->vdev_ms[m])); if (error != 0) { vdev_dbgmsg(vd, "metaslab_init failed [error=%d]", error); return (error); } } /* * Find the emptiest metaslab on the vdev and mark it for use for * embedded slog by moving it from the regular to the log metaslab * group. */ if (vd->vdev_mg->mg_class == spa_normal_class(spa) && vd->vdev_ms_count > zfs_embedded_slog_min_ms && avl_is_empty(&vd->vdev_log_mg->mg_metaslab_tree)) { uint64_t slog_msid = 0; uint64_t smallest = UINT64_MAX; /* * Note, we only search the new metaslabs, because the old * (pre-existing) ones may be active (e.g. have non-empty * range_tree's), and we don't move them to the new * metaslab_t. */ for (uint64_t m = oldc; m < newc; m++) { uint64_t alloc = space_map_allocated(vd->vdev_ms[m]->ms_sm); if (alloc < smallest) { slog_msid = m; smallest = alloc; } } metaslab_t *slog_ms = vd->vdev_ms[slog_msid]; /* * The metaslab was marked as dirty at the end of * metaslab_init(). Remove it from the dirty list so that we * can uninitialize and reinitialize it to the new class. */ if (txg != 0) { (void) txg_list_remove_this(&vd->vdev_ms_list, slog_ms, txg); } uint64_t sm_obj = space_map_object(slog_ms->ms_sm); metaslab_fini(slog_ms); VERIFY0(metaslab_init(vd->vdev_log_mg, slog_msid, sm_obj, txg, &vd->vdev_ms[slog_msid])); } if (txg == 0) spa_config_enter(spa, SCL_ALLOC, FTAG, RW_WRITER); /* * If the vdev is marked as non-allocating then don't * activate the metaslabs since we want to ensure that * no allocations are performed on this device. */ if (vd->vdev_noalloc) { /* track non-allocating vdev space */ spa->spa_nonallocating_dspace += spa_deflate(spa) ? vd->vdev_stat.vs_dspace : vd->vdev_stat.vs_space; } else if (!expanding) { metaslab_group_activate(vd->vdev_mg); if (vd->vdev_log_mg != NULL) metaslab_group_activate(vd->vdev_log_mg); } if (txg == 0) spa_config_exit(spa, SCL_ALLOC, FTAG); return (0); } void vdev_metaslab_fini(vdev_t *vd) { if (vd->vdev_checkpoint_sm != NULL) { ASSERT(spa_feature_is_active(vd->vdev_spa, SPA_FEATURE_POOL_CHECKPOINT)); space_map_close(vd->vdev_checkpoint_sm); /* * Even though we close the space map, we need to set its * pointer to NULL. The reason is that vdev_metaslab_fini() * may be called multiple times for certain operations * (i.e. when destroying a pool) so we need to ensure that * this clause never executes twice. This logic is similar * to the one used for the vdev_ms clause below. */ vd->vdev_checkpoint_sm = NULL; } if (vd->vdev_ms != NULL) { metaslab_group_t *mg = vd->vdev_mg; metaslab_group_passivate(mg); if (vd->vdev_log_mg != NULL) { ASSERT(!vd->vdev_islog); metaslab_group_passivate(vd->vdev_log_mg); } uint64_t count = vd->vdev_ms_count; for (uint64_t m = 0; m < count; m++) { metaslab_t *msp = vd->vdev_ms[m]; if (msp != NULL) metaslab_fini(msp); } vmem_free(vd->vdev_ms, count * sizeof (metaslab_t *)); vd->vdev_ms = NULL; vd->vdev_ms_count = 0; for (int i = 0; i < RANGE_TREE_HISTOGRAM_SIZE; i++) { ASSERT0(mg->mg_histogram[i]); if (vd->vdev_log_mg != NULL) ASSERT0(vd->vdev_log_mg->mg_histogram[i]); } } ASSERT0(vd->vdev_ms_count); ASSERT3U(vd->vdev_pending_fastwrite, ==, 0); } typedef struct vdev_probe_stats { boolean_t vps_readable; boolean_t vps_writeable; int vps_flags; } vdev_probe_stats_t; static void vdev_probe_done(zio_t *zio) { spa_t *spa = zio->io_spa; vdev_t *vd = zio->io_vd; vdev_probe_stats_t *vps = zio->io_private; ASSERT(vd->vdev_probe_zio != NULL); if (zio->io_type == ZIO_TYPE_READ) { if (zio->io_error == 0) vps->vps_readable = 1; if (zio->io_error == 0 && spa_writeable(spa)) { zio_nowait(zio_write_phys(vd->vdev_probe_zio, vd, zio->io_offset, zio->io_size, zio->io_abd, ZIO_CHECKSUM_OFF, vdev_probe_done, vps, ZIO_PRIORITY_SYNC_WRITE, vps->vps_flags, B_TRUE)); } else { abd_free(zio->io_abd); } } else if (zio->io_type == ZIO_TYPE_WRITE) { if (zio->io_error == 0) vps->vps_writeable = 1; abd_free(zio->io_abd); } else if (zio->io_type == ZIO_TYPE_NULL) { zio_t *pio; zio_link_t *zl; vd->vdev_cant_read |= !vps->vps_readable; vd->vdev_cant_write |= !vps->vps_writeable; if (vdev_readable(vd) && (vdev_writeable(vd) || !spa_writeable(spa))) { zio->io_error = 0; } else { ASSERT(zio->io_error != 0); vdev_dbgmsg(vd, "failed probe"); (void) zfs_ereport_post(FM_EREPORT_ZFS_PROBE_FAILURE, spa, vd, NULL, NULL, 0); zio->io_error = SET_ERROR(ENXIO); } mutex_enter(&vd->vdev_probe_lock); ASSERT(vd->vdev_probe_zio == zio); vd->vdev_probe_zio = NULL; mutex_exit(&vd->vdev_probe_lock); zl = NULL; while ((pio = zio_walk_parents(zio, &zl)) != NULL) if (!vdev_accessible(vd, pio)) pio->io_error = SET_ERROR(ENXIO); kmem_free(vps, sizeof (*vps)); } } /* * Determine whether this device is accessible. * * Read and write to several known locations: the pad regions of each * vdev label but the first, which we leave alone in case it contains * a VTOC. */ zio_t * vdev_probe(vdev_t *vd, zio_t *zio) { spa_t *spa = vd->vdev_spa; vdev_probe_stats_t *vps = NULL; zio_t *pio; ASSERT(vd->vdev_ops->vdev_op_leaf); /* * Don't probe the probe. */ if (zio && (zio->io_flags & ZIO_FLAG_PROBE)) return (NULL); /* * To prevent 'probe storms' when a device fails, we create * just one probe i/o at a time. All zios that want to probe * this vdev will become parents of the probe io. */ mutex_enter(&vd->vdev_probe_lock); if ((pio = vd->vdev_probe_zio) == NULL) { vps = kmem_zalloc(sizeof (*vps), KM_SLEEP); vps->vps_flags = ZIO_FLAG_CANFAIL | ZIO_FLAG_PROBE | ZIO_FLAG_DONT_CACHE | ZIO_FLAG_DONT_AGGREGATE | ZIO_FLAG_TRYHARD; if (spa_config_held(spa, SCL_ZIO, RW_WRITER)) { /* * vdev_cant_read and vdev_cant_write can only * transition from TRUE to FALSE when we have the * SCL_ZIO lock as writer; otherwise they can only * transition from FALSE to TRUE. This ensures that * any zio looking at these values can assume that * failures persist for the life of the I/O. That's * important because when a device has intermittent * connectivity problems, we want to ensure that * they're ascribed to the device (ENXIO) and not * the zio (EIO). * * Since we hold SCL_ZIO as writer here, clear both * values so the probe can reevaluate from first * principles. */ vps->vps_flags |= ZIO_FLAG_CONFIG_WRITER; vd->vdev_cant_read = B_FALSE; vd->vdev_cant_write = B_FALSE; } vd->vdev_probe_zio = pio = zio_null(NULL, spa, vd, vdev_probe_done, vps, vps->vps_flags | ZIO_FLAG_DONT_PROPAGATE); /* * We can't change the vdev state in this context, so we * kick off an async task to do it on our behalf. */ if (zio != NULL) { vd->vdev_probe_wanted = B_TRUE; spa_async_request(spa, SPA_ASYNC_PROBE); } } if (zio != NULL) zio_add_child(zio, pio); mutex_exit(&vd->vdev_probe_lock); if (vps == NULL) { ASSERT(zio != NULL); return (NULL); } for (int l = 1; l < VDEV_LABELS; l++) { zio_nowait(zio_read_phys(pio, vd, vdev_label_offset(vd->vdev_psize, l, offsetof(vdev_label_t, vl_be)), VDEV_PAD_SIZE, abd_alloc_for_io(VDEV_PAD_SIZE, B_TRUE), ZIO_CHECKSUM_OFF, vdev_probe_done, vps, ZIO_PRIORITY_SYNC_READ, vps->vps_flags, B_TRUE)); } if (zio == NULL) return (pio); zio_nowait(pio); return (NULL); } static void vdev_load_child(void *arg) { vdev_t *vd = arg; vd->vdev_load_error = vdev_load(vd); } static void vdev_open_child(void *arg) { vdev_t *vd = arg; vd->vdev_open_thread = curthread; vd->vdev_open_error = vdev_open(vd); vd->vdev_open_thread = NULL; } static boolean_t vdev_uses_zvols(vdev_t *vd) { #ifdef _KERNEL if (zvol_is_zvol(vd->vdev_path)) return (B_TRUE); #endif for (int c = 0; c < vd->vdev_children; c++) if (vdev_uses_zvols(vd->vdev_child[c])) return (B_TRUE); return (B_FALSE); } /* * Returns B_TRUE if the passed child should be opened. */ static boolean_t vdev_default_open_children_func(vdev_t *vd) { (void) vd; return (B_TRUE); } /* * Open the requested child vdevs. If any of the leaf vdevs are using * a ZFS volume then do the opens in a single thread. This avoids a * deadlock when the current thread is holding the spa_namespace_lock. */ static void vdev_open_children_impl(vdev_t *vd, vdev_open_children_func_t *open_func) { int children = vd->vdev_children; taskq_t *tq = taskq_create("vdev_open", children, minclsyspri, children, children, TASKQ_PREPOPULATE); vd->vdev_nonrot = B_TRUE; for (int c = 0; c < children; c++) { vdev_t *cvd = vd->vdev_child[c]; if (open_func(cvd) == B_FALSE) continue; if (tq == NULL || vdev_uses_zvols(vd)) { cvd->vdev_open_error = vdev_open(cvd); } else { VERIFY(taskq_dispatch(tq, vdev_open_child, cvd, TQ_SLEEP) != TASKQID_INVALID); } vd->vdev_nonrot &= cvd->vdev_nonrot; } if (tq != NULL) { taskq_wait(tq); taskq_destroy(tq); } } /* * Open all child vdevs. */ void vdev_open_children(vdev_t *vd) { vdev_open_children_impl(vd, vdev_default_open_children_func); } /* * Conditionally open a subset of child vdevs. */ void vdev_open_children_subset(vdev_t *vd, vdev_open_children_func_t *open_func) { vdev_open_children_impl(vd, open_func); } /* * Compute the raidz-deflation ratio. Note, we hard-code * in 128k (1 << 17) because it is the "typical" blocksize. * Even though SPA_MAXBLOCKSIZE changed, this algorithm can not change, * otherwise it would inconsistently account for existing bp's. */ static void vdev_set_deflate_ratio(vdev_t *vd) { if (vd == vd->vdev_top && !vd->vdev_ishole && vd->vdev_ashift != 0) { vd->vdev_deflate_ratio = (1 << 17) / (vdev_psize_to_asize(vd, 1 << 17) >> SPA_MINBLOCKSHIFT); } } /* * Choose the best of two ashifts, preferring one between logical ashift * (absolute minimum) and administrator defined maximum, otherwise take * the biggest of the two. */ uint64_t vdev_best_ashift(uint64_t logical, uint64_t a, uint64_t b) { if (a > logical && a <= zfs_vdev_max_auto_ashift) { if (b <= logical || b > zfs_vdev_max_auto_ashift) return (a); else return (MAX(a, b)); } else if (b <= logical || b > zfs_vdev_max_auto_ashift) return (MAX(a, b)); return (b); } /* * Maximize performance by inflating the configured ashift for top level * vdevs to be as close to the physical ashift as possible while maintaining * administrator defined limits and ensuring it doesn't go below the * logical ashift. */ static void vdev_ashift_optimize(vdev_t *vd) { ASSERT(vd == vd->vdev_top); if (vd->vdev_ashift < vd->vdev_physical_ashift && vd->vdev_physical_ashift <= zfs_vdev_max_auto_ashift) { vd->vdev_ashift = MIN( MAX(zfs_vdev_max_auto_ashift, vd->vdev_ashift), MAX(zfs_vdev_min_auto_ashift, vd->vdev_physical_ashift)); } else { /* * If the logical and physical ashifts are the same, then * we ensure that the top-level vdev's ashift is not smaller * than our minimum ashift value. For the unusual case * where logical ashift > physical ashift, we can't cap * the calculated ashift based on max ashift as that * would cause failures. * We still check if we need to increase it to match * the min ashift. */ vd->vdev_ashift = MAX(zfs_vdev_min_auto_ashift, vd->vdev_ashift); } } /* * Prepare a virtual device for access. */ int vdev_open(vdev_t *vd) { spa_t *spa = vd->vdev_spa; int error; uint64_t osize = 0; uint64_t max_osize = 0; uint64_t asize, max_asize, psize; uint64_t logical_ashift = 0; uint64_t physical_ashift = 0; ASSERT(vd->vdev_open_thread == curthread || spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL); ASSERT(vd->vdev_state == VDEV_STATE_CLOSED || vd->vdev_state == VDEV_STATE_CANT_OPEN || vd->vdev_state == VDEV_STATE_OFFLINE); vd->vdev_stat.vs_aux = VDEV_AUX_NONE; vd->vdev_cant_read = B_FALSE; vd->vdev_cant_write = B_FALSE; vd->vdev_min_asize = vdev_get_min_asize(vd); /* * If this vdev is not removed, check its fault status. If it's * faulted, bail out of the open. */ if (!vd->vdev_removed && vd->vdev_faulted) { ASSERT(vd->vdev_children == 0); ASSERT(vd->vdev_label_aux == VDEV_AUX_ERR_EXCEEDED || vd->vdev_label_aux == VDEV_AUX_EXTERNAL); vdev_set_state(vd, B_TRUE, VDEV_STATE_FAULTED, vd->vdev_label_aux); return (SET_ERROR(ENXIO)); } else if (vd->vdev_offline) { ASSERT(vd->vdev_children == 0); vdev_set_state(vd, B_TRUE, VDEV_STATE_OFFLINE, VDEV_AUX_NONE); return (SET_ERROR(ENXIO)); } error = vd->vdev_ops->vdev_op_open(vd, &osize, &max_osize, &logical_ashift, &physical_ashift); /* Keep the device in removed state if unplugged */ if (error == ENOENT && vd->vdev_removed) { vdev_set_state(vd, B_TRUE, VDEV_STATE_REMOVED, VDEV_AUX_NONE); return (error); } /* * Physical volume size should never be larger than its max size, unless * the disk has shrunk while we were reading it or the device is buggy * or damaged: either way it's not safe for use, bail out of the open. */ if (osize > max_osize) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_OPEN_FAILED); return (SET_ERROR(ENXIO)); } /* * Reset the vdev_reopening flag so that we actually close * the vdev on error. */ vd->vdev_reopening = B_FALSE; if (zio_injection_enabled && error == 0) error = zio_handle_device_injection(vd, NULL, SET_ERROR(ENXIO)); if (error) { if (vd->vdev_removed && vd->vdev_stat.vs_aux != VDEV_AUX_OPEN_FAILED) vd->vdev_removed = B_FALSE; if (vd->vdev_stat.vs_aux == VDEV_AUX_CHILDREN_OFFLINE) { vdev_set_state(vd, B_TRUE, VDEV_STATE_OFFLINE, vd->vdev_stat.vs_aux); } else { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, vd->vdev_stat.vs_aux); } return (error); } vd->vdev_removed = B_FALSE; /* * Recheck the faulted flag now that we have confirmed that * the vdev is accessible. If we're faulted, bail. */ if (vd->vdev_faulted) { ASSERT(vd->vdev_children == 0); ASSERT(vd->vdev_label_aux == VDEV_AUX_ERR_EXCEEDED || vd->vdev_label_aux == VDEV_AUX_EXTERNAL); vdev_set_state(vd, B_TRUE, VDEV_STATE_FAULTED, vd->vdev_label_aux); return (SET_ERROR(ENXIO)); } if (vd->vdev_degraded) { ASSERT(vd->vdev_children == 0); vdev_set_state(vd, B_TRUE, VDEV_STATE_DEGRADED, VDEV_AUX_ERR_EXCEEDED); } else { vdev_set_state(vd, B_TRUE, VDEV_STATE_HEALTHY, 0); } /* * For hole or missing vdevs we just return success. */ if (vd->vdev_ishole || vd->vdev_ops == &vdev_missing_ops) return (0); for (int c = 0; c < vd->vdev_children; c++) { if (vd->vdev_child[c]->vdev_state != VDEV_STATE_HEALTHY) { vdev_set_state(vd, B_TRUE, VDEV_STATE_DEGRADED, VDEV_AUX_NONE); break; } } osize = P2ALIGN(osize, (uint64_t)sizeof (vdev_label_t)); max_osize = P2ALIGN(max_osize, (uint64_t)sizeof (vdev_label_t)); if (vd->vdev_children == 0) { if (osize < SPA_MINDEVSIZE) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_TOO_SMALL); return (SET_ERROR(EOVERFLOW)); } psize = osize; asize = osize - (VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE); max_asize = max_osize - (VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE); } else { if (vd->vdev_parent != NULL && osize < SPA_MINDEVSIZE - (VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE)) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_TOO_SMALL); return (SET_ERROR(EOVERFLOW)); } psize = 0; asize = osize; max_asize = max_osize; } /* * If the vdev was expanded, record this so that we can re-create the * uberblock rings in labels {2,3}, during the next sync. */ if ((psize > vd->vdev_psize) && (vd->vdev_psize != 0)) vd->vdev_copy_uberblocks = B_TRUE; vd->vdev_psize = psize; /* * Make sure the allocatable size hasn't shrunk too much. */ if (asize < vd->vdev_min_asize) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_BAD_LABEL); return (SET_ERROR(EINVAL)); } /* * We can always set the logical/physical ashift members since * their values are only used to calculate the vdev_ashift when * the device is first added to the config. These values should * not be used for anything else since they may change whenever * the device is reopened and we don't store them in the label. */ vd->vdev_physical_ashift = MAX(physical_ashift, vd->vdev_physical_ashift); vd->vdev_logical_ashift = MAX(logical_ashift, vd->vdev_logical_ashift); if (vd->vdev_asize == 0) { /* * This is the first-ever open, so use the computed values. * For compatibility, a different ashift can be requested. */ vd->vdev_asize = asize; vd->vdev_max_asize = max_asize; /* * If the vdev_ashift was not overridden at creation time, * then set it the logical ashift and optimize the ashift. */ if (vd->vdev_ashift == 0) { vd->vdev_ashift = vd->vdev_logical_ashift; if (vd->vdev_logical_ashift > ASHIFT_MAX) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_ASHIFT_TOO_BIG); return (SET_ERROR(EDOM)); } if (vd->vdev_top == vd) { vdev_ashift_optimize(vd); } } if (vd->vdev_ashift != 0 && (vd->vdev_ashift < ASHIFT_MIN || vd->vdev_ashift > ASHIFT_MAX)) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_BAD_ASHIFT); return (SET_ERROR(EDOM)); } } else { /* * Make sure the alignment required hasn't increased. */ if (vd->vdev_ashift > vd->vdev_top->vdev_ashift && vd->vdev_ops->vdev_op_leaf) { (void) zfs_ereport_post( FM_EREPORT_ZFS_DEVICE_BAD_ASHIFT, spa, vd, NULL, NULL, 0); vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_BAD_LABEL); return (SET_ERROR(EDOM)); } vd->vdev_max_asize = max_asize; } /* * If all children are healthy we update asize if either: * The asize has increased, due to a device expansion caused by dynamic * LUN growth or vdev replacement, and automatic expansion is enabled; * making the additional space available. * * The asize has decreased, due to a device shrink usually caused by a * vdev replace with a smaller device. This ensures that calculations * based of max_asize and asize e.g. esize are always valid. It's safe * to do this as we've already validated that asize is greater than * vdev_min_asize. */ if (vd->vdev_state == VDEV_STATE_HEALTHY && ((asize > vd->vdev_asize && (vd->vdev_expanding || spa->spa_autoexpand)) || (asize < vd->vdev_asize))) vd->vdev_asize = asize; vdev_set_min_asize(vd); /* * Ensure we can issue some IO before declaring the * vdev open for business. */ if (vd->vdev_ops->vdev_op_leaf && (error = zio_wait(vdev_probe(vd, NULL))) != 0) { vdev_set_state(vd, B_TRUE, VDEV_STATE_FAULTED, VDEV_AUX_ERR_EXCEEDED); return (error); } /* * Track the minimum allocation size. */ if (vd->vdev_top == vd && vd->vdev_ashift != 0 && vd->vdev_islog == 0 && vd->vdev_aux == NULL) { uint64_t min_alloc = vdev_get_min_alloc(vd); if (min_alloc < spa->spa_min_alloc) spa->spa_min_alloc = min_alloc; } /* * If this is a leaf vdev, assess whether a resilver is needed. * But don't do this if we are doing a reopen for a scrub, since * this would just restart the scrub we are already doing. */ if (vd->vdev_ops->vdev_op_leaf && !spa->spa_scrub_reopen) dsl_scan_assess_vdev(spa->spa_dsl_pool, vd); return (0); } static void vdev_validate_child(void *arg) { vdev_t *vd = arg; vd->vdev_validate_thread = curthread; vd->vdev_validate_error = vdev_validate(vd); vd->vdev_validate_thread = NULL; } /* * Called once the vdevs are all opened, this routine validates the label * contents. This needs to be done before vdev_load() so that we don't * inadvertently do repair I/Os to the wrong device. * * This function will only return failure if one of the vdevs indicates that it * has since been destroyed or exported. This is only possible if * /etc/zfs/zpool.cache was readonly at the time. Otherwise, the vdev state * will be updated but the function will return 0. */ int vdev_validate(vdev_t *vd) { spa_t *spa = vd->vdev_spa; taskq_t *tq = NULL; nvlist_t *label; uint64_t guid = 0, aux_guid = 0, top_guid; uint64_t state; nvlist_t *nvl; uint64_t txg; int children = vd->vdev_children; if (vdev_validate_skip) return (0); if (children > 0) { tq = taskq_create("vdev_validate", children, minclsyspri, children, children, TASKQ_PREPOPULATE); } for (uint64_t c = 0; c < children; c++) { vdev_t *cvd = vd->vdev_child[c]; if (tq == NULL || vdev_uses_zvols(cvd)) { vdev_validate_child(cvd); } else { VERIFY(taskq_dispatch(tq, vdev_validate_child, cvd, TQ_SLEEP) != TASKQID_INVALID); } } if (tq != NULL) { taskq_wait(tq); taskq_destroy(tq); } for (int c = 0; c < children; c++) { int error = vd->vdev_child[c]->vdev_validate_error; if (error != 0) return (SET_ERROR(EBADF)); } /* * If the device has already failed, or was marked offline, don't do * any further validation. Otherwise, label I/O will fail and we will * overwrite the previous state. */ if (!vd->vdev_ops->vdev_op_leaf || !vdev_readable(vd)) return (0); /* * If we are performing an extreme rewind, we allow for a label that * was modified at a point after the current txg. * If config lock is not held do not check for the txg. spa_sync could * be updating the vdev's label before updating spa_last_synced_txg. */ if (spa->spa_extreme_rewind || spa_last_synced_txg(spa) == 0 || spa_config_held(spa, SCL_CONFIG, RW_WRITER) != SCL_CONFIG) txg = UINT64_MAX; else txg = spa_last_synced_txg(spa); if ((label = vdev_label_read_config(vd, txg)) == NULL) { vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_BAD_LABEL); vdev_dbgmsg(vd, "vdev_validate: failed reading config for " "txg %llu", (u_longlong_t)txg); return (0); } /* * Determine if this vdev has been split off into another * pool. If so, then refuse to open it. */ if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_SPLIT_GUID, &aux_guid) == 0 && aux_guid == spa_guid(spa)) { vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_SPLIT_POOL); nvlist_free(label); vdev_dbgmsg(vd, "vdev_validate: vdev split into other pool"); return (0); } if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_GUID, &guid) != 0) { vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); nvlist_free(label); vdev_dbgmsg(vd, "vdev_validate: '%s' missing from label", ZPOOL_CONFIG_POOL_GUID); return (0); } /* * If config is not trusted then ignore the spa guid check. This is * necessary because if the machine crashed during a re-guid the new * guid might have been written to all of the vdev labels, but not the * cached config. The check will be performed again once we have the * trusted config from the MOS. */ if (spa->spa_trust_config && guid != spa_guid(spa)) { vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); nvlist_free(label); vdev_dbgmsg(vd, "vdev_validate: vdev label pool_guid doesn't " "match config (%llu != %llu)", (u_longlong_t)guid, (u_longlong_t)spa_guid(spa)); return (0); } if (nvlist_lookup_nvlist(label, ZPOOL_CONFIG_VDEV_TREE, &nvl) != 0 || nvlist_lookup_uint64(nvl, ZPOOL_CONFIG_ORIG_GUID, &aux_guid) != 0) aux_guid = 0; if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_GUID, &guid) != 0) { vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); nvlist_free(label); vdev_dbgmsg(vd, "vdev_validate: '%s' missing from label", ZPOOL_CONFIG_GUID); return (0); } if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_TOP_GUID, &top_guid) != 0) { vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); nvlist_free(label); vdev_dbgmsg(vd, "vdev_validate: '%s' missing from label", ZPOOL_CONFIG_TOP_GUID); return (0); } /* * If this vdev just became a top-level vdev because its sibling was * detached, it will have adopted the parent's vdev guid -- but the * label may or may not be on disk yet. Fortunately, either version * of the label will have the same top guid, so if we're a top-level * vdev, we can safely compare to that instead. * However, if the config comes from a cachefile that failed to update * after the detach, a top-level vdev will appear as a non top-level * vdev in the config. Also relax the constraints if we perform an * extreme rewind. * * If we split this vdev off instead, then we also check the * original pool's guid. We don't want to consider the vdev * corrupt if it is partway through a split operation. */ if (vd->vdev_guid != guid && vd->vdev_guid != aux_guid) { boolean_t mismatch = B_FALSE; if (spa->spa_trust_config && !spa->spa_extreme_rewind) { if (vd != vd->vdev_top || vd->vdev_guid != top_guid) mismatch = B_TRUE; } else { if (vd->vdev_guid != top_guid && vd->vdev_top->vdev_guid != guid) mismatch = B_TRUE; } if (mismatch) { vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); nvlist_free(label); vdev_dbgmsg(vd, "vdev_validate: config guid " "doesn't match label guid"); vdev_dbgmsg(vd, "CONFIG: guid %llu, top_guid %llu", (u_longlong_t)vd->vdev_guid, (u_longlong_t)vd->vdev_top->vdev_guid); vdev_dbgmsg(vd, "LABEL: guid %llu, top_guid %llu, " "aux_guid %llu", (u_longlong_t)guid, (u_longlong_t)top_guid, (u_longlong_t)aux_guid); return (0); } } if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_STATE, &state) != 0) { vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); nvlist_free(label); vdev_dbgmsg(vd, "vdev_validate: '%s' missing from label", ZPOOL_CONFIG_POOL_STATE); return (0); } nvlist_free(label); /* * If this is a verbatim import, no need to check the * state of the pool. */ if (!(spa->spa_import_flags & ZFS_IMPORT_VERBATIM) && spa_load_state(spa) == SPA_LOAD_OPEN && state != POOL_STATE_ACTIVE) { vdev_dbgmsg(vd, "vdev_validate: invalid pool state (%llu) " "for spa %s", (u_longlong_t)state, spa->spa_name); return (SET_ERROR(EBADF)); } /* * If we were able to open and validate a vdev that was * previously marked permanently unavailable, clear that state * now. */ if (vd->vdev_not_present) vd->vdev_not_present = 0; return (0); } static void vdev_copy_path_impl(vdev_t *svd, vdev_t *dvd) { char *old, *new; if (svd->vdev_path != NULL && dvd->vdev_path != NULL) { if (strcmp(svd->vdev_path, dvd->vdev_path) != 0) { zfs_dbgmsg("vdev_copy_path: vdev %llu: path changed " "from '%s' to '%s'", (u_longlong_t)dvd->vdev_guid, dvd->vdev_path, svd->vdev_path); spa_strfree(dvd->vdev_path); dvd->vdev_path = spa_strdup(svd->vdev_path); } } else if (svd->vdev_path != NULL) { dvd->vdev_path = spa_strdup(svd->vdev_path); zfs_dbgmsg("vdev_copy_path: vdev %llu: path set to '%s'", (u_longlong_t)dvd->vdev_guid, dvd->vdev_path); } /* * Our enclosure sysfs path may have changed between imports */ old = dvd->vdev_enc_sysfs_path; new = svd->vdev_enc_sysfs_path; if ((old != NULL && new == NULL) || (old == NULL && new != NULL) || ((old != NULL && new != NULL) && strcmp(new, old) != 0)) { zfs_dbgmsg("vdev_copy_path: vdev %llu: vdev_enc_sysfs_path " "changed from '%s' to '%s'", (u_longlong_t)dvd->vdev_guid, old, new); if (dvd->vdev_enc_sysfs_path) spa_strfree(dvd->vdev_enc_sysfs_path); if (svd->vdev_enc_sysfs_path) { dvd->vdev_enc_sysfs_path = spa_strdup( svd->vdev_enc_sysfs_path); } else { dvd->vdev_enc_sysfs_path = NULL; } } } /* * Recursively copy vdev paths from one vdev to another. Source and destination * vdev trees must have same geometry otherwise return error. Intended to copy * paths from userland config into MOS config. */ int vdev_copy_path_strict(vdev_t *svd, vdev_t *dvd) { if ((svd->vdev_ops == &vdev_missing_ops) || (svd->vdev_ishole && dvd->vdev_ishole) || (dvd->vdev_ops == &vdev_indirect_ops)) return (0); if (svd->vdev_ops != dvd->vdev_ops) { vdev_dbgmsg(svd, "vdev_copy_path: vdev type mismatch: %s != %s", svd->vdev_ops->vdev_op_type, dvd->vdev_ops->vdev_op_type); return (SET_ERROR(EINVAL)); } if (svd->vdev_guid != dvd->vdev_guid) { vdev_dbgmsg(svd, "vdev_copy_path: guids mismatch (%llu != " "%llu)", (u_longlong_t)svd->vdev_guid, (u_longlong_t)dvd->vdev_guid); return (SET_ERROR(EINVAL)); } if (svd->vdev_children != dvd->vdev_children) { vdev_dbgmsg(svd, "vdev_copy_path: children count mismatch: " "%llu != %llu", (u_longlong_t)svd->vdev_children, (u_longlong_t)dvd->vdev_children); return (SET_ERROR(EINVAL)); } for (uint64_t i = 0; i < svd->vdev_children; i++) { int error = vdev_copy_path_strict(svd->vdev_child[i], dvd->vdev_child[i]); if (error != 0) return (error); } if (svd->vdev_ops->vdev_op_leaf) vdev_copy_path_impl(svd, dvd); return (0); } static void vdev_copy_path_search(vdev_t *stvd, vdev_t *dvd) { ASSERT(stvd->vdev_top == stvd); ASSERT3U(stvd->vdev_id, ==, dvd->vdev_top->vdev_id); for (uint64_t i = 0; i < dvd->vdev_children; i++) { vdev_copy_path_search(stvd, dvd->vdev_child[i]); } if (!dvd->vdev_ops->vdev_op_leaf || !vdev_is_concrete(dvd)) return; /* * The idea here is that while a vdev can shift positions within * a top vdev (when replacing, attaching mirror, etc.) it cannot * step outside of it. */ vdev_t *vd = vdev_lookup_by_guid(stvd, dvd->vdev_guid); if (vd == NULL || vd->vdev_ops != dvd->vdev_ops) return; ASSERT(vd->vdev_ops->vdev_op_leaf); vdev_copy_path_impl(vd, dvd); } /* * Recursively copy vdev paths from one root vdev to another. Source and * destination vdev trees may differ in geometry. For each destination leaf * vdev, search a vdev with the same guid and top vdev id in the source. * Intended to copy paths from userland config into MOS config. */ void vdev_copy_path_relaxed(vdev_t *srvd, vdev_t *drvd) { uint64_t children = MIN(srvd->vdev_children, drvd->vdev_children); ASSERT(srvd->vdev_ops == &vdev_root_ops); ASSERT(drvd->vdev_ops == &vdev_root_ops); for (uint64_t i = 0; i < children; i++) { vdev_copy_path_search(srvd->vdev_child[i], drvd->vdev_child[i]); } } /* * Close a virtual device. */ void vdev_close(vdev_t *vd) { vdev_t *pvd = vd->vdev_parent; spa_t *spa __maybe_unused = vd->vdev_spa; ASSERT(vd != NULL); ASSERT(vd->vdev_open_thread == curthread || spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL); /* * If our parent is reopening, then we are as well, unless we are * going offline. */ if (pvd != NULL && pvd->vdev_reopening) vd->vdev_reopening = (pvd->vdev_reopening && !vd->vdev_offline); vd->vdev_ops->vdev_op_close(vd); vdev_cache_purge(vd); /* * We record the previous state before we close it, so that if we are * doing a reopen(), we don't generate FMA ereports if we notice that * it's still faulted. */ vd->vdev_prevstate = vd->vdev_state; if (vd->vdev_offline) vd->vdev_state = VDEV_STATE_OFFLINE; else vd->vdev_state = VDEV_STATE_CLOSED; vd->vdev_stat.vs_aux = VDEV_AUX_NONE; } void vdev_hold(vdev_t *vd) { spa_t *spa = vd->vdev_spa; ASSERT(spa_is_root(spa)); if (spa->spa_state == POOL_STATE_UNINITIALIZED) return; for (int c = 0; c < vd->vdev_children; c++) vdev_hold(vd->vdev_child[c]); if (vd->vdev_ops->vdev_op_leaf && vd->vdev_ops->vdev_op_hold != NULL) vd->vdev_ops->vdev_op_hold(vd); } void vdev_rele(vdev_t *vd) { ASSERT(spa_is_root(vd->vdev_spa)); for (int c = 0; c < vd->vdev_children; c++) vdev_rele(vd->vdev_child[c]); if (vd->vdev_ops->vdev_op_leaf && vd->vdev_ops->vdev_op_rele != NULL) vd->vdev_ops->vdev_op_rele(vd); } /* * Reopen all interior vdevs and any unopened leaves. We don't actually * reopen leaf vdevs which had previously been opened as they might deadlock * on the spa_config_lock. Instead we only obtain the leaf's physical size. * If the leaf has never been opened then open it, as usual. */ void vdev_reopen(vdev_t *vd) { spa_t *spa = vd->vdev_spa; ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL); /* set the reopening flag unless we're taking the vdev offline */ vd->vdev_reopening = !vd->vdev_offline; vdev_close(vd); (void) vdev_open(vd); /* * Call vdev_validate() here to make sure we have the same device. * Otherwise, a device with an invalid label could be successfully * opened in response to vdev_reopen(). */ if (vd->vdev_aux) { (void) vdev_validate_aux(vd); if (vdev_readable(vd) && vdev_writeable(vd) && vd->vdev_aux == &spa->spa_l2cache) { /* * In case the vdev is present we should evict all ARC * buffers and pointers to log blocks and reclaim their * space before restoring its contents to L2ARC. */ if (l2arc_vdev_present(vd)) { l2arc_rebuild_vdev(vd, B_TRUE); } else { l2arc_add_vdev(spa, vd); } spa_async_request(spa, SPA_ASYNC_L2CACHE_REBUILD); spa_async_request(spa, SPA_ASYNC_L2CACHE_TRIM); } } else { (void) vdev_validate(vd); } /* * Reassess parent vdev's health. */ vdev_propagate_state(vd); } int vdev_create(vdev_t *vd, uint64_t txg, boolean_t isreplacing) { int error; /* * Normally, partial opens (e.g. of a mirror) are allowed. * For a create, however, we want to fail the request if * there are any components we can't open. */ error = vdev_open(vd); if (error || vd->vdev_state != VDEV_STATE_HEALTHY) { vdev_close(vd); return (error ? error : SET_ERROR(ENXIO)); } /* * Recursively load DTLs and initialize all labels. */ if ((error = vdev_dtl_load(vd)) != 0 || (error = vdev_label_init(vd, txg, isreplacing ? VDEV_LABEL_REPLACE : VDEV_LABEL_CREATE)) != 0) { vdev_close(vd); return (error); } return (0); } void vdev_metaslab_set_size(vdev_t *vd) { uint64_t asize = vd->vdev_asize; uint64_t ms_count = asize >> zfs_vdev_default_ms_shift; uint64_t ms_shift; /* * There are two dimensions to the metaslab sizing calculation: * the size of the metaslab and the count of metaslabs per vdev. * * The default values used below are a good balance between memory * usage (larger metaslab size means more memory needed for loaded * metaslabs; more metaslabs means more memory needed for the * metaslab_t structs), metaslab load time (larger metaslabs take * longer to load), and metaslab sync time (more metaslabs means * more time spent syncing all of them). * * In general, we aim for zfs_vdev_default_ms_count (200) metaslabs. * The range of the dimensions are as follows: * * 2^29 <= ms_size <= 2^34 * 16 <= ms_count <= 131,072 * * On the lower end of vdev sizes, we aim for metaslabs sizes of * at least 512MB (2^29) to minimize fragmentation effects when * testing with smaller devices. However, the count constraint * of at least 16 metaslabs will override this minimum size goal. * * On the upper end of vdev sizes, we aim for a maximum metaslab * size of 16GB. However, we will cap the total count to 2^17 * metaslabs to keep our memory footprint in check and let the * metaslab size grow from there if that limit is hit. * * The net effect of applying above constrains is summarized below. * * vdev size metaslab count * --------------|----------------- * < 8GB ~16 * 8GB - 100GB one per 512MB * 100GB - 3TB ~200 * 3TB - 2PB one per 16GB * > 2PB ~131,072 * -------------------------------- * * Finally, note that all of the above calculate the initial * number of metaslabs. Expanding a top-level vdev will result * in additional metaslabs being allocated making it possible * to exceed the zfs_vdev_ms_count_limit. */ if (ms_count < zfs_vdev_min_ms_count) ms_shift = highbit64(asize / zfs_vdev_min_ms_count); else if (ms_count > zfs_vdev_default_ms_count) ms_shift = highbit64(asize / zfs_vdev_default_ms_count); else ms_shift = zfs_vdev_default_ms_shift; if (ms_shift < SPA_MAXBLOCKSHIFT) { ms_shift = SPA_MAXBLOCKSHIFT; } else if (ms_shift > zfs_vdev_max_ms_shift) { ms_shift = zfs_vdev_max_ms_shift; /* cap the total count to constrain memory footprint */ if ((asize >> ms_shift) > zfs_vdev_ms_count_limit) ms_shift = highbit64(asize / zfs_vdev_ms_count_limit); } vd->vdev_ms_shift = ms_shift; ASSERT3U(vd->vdev_ms_shift, >=, SPA_MAXBLOCKSHIFT); } void vdev_dirty(vdev_t *vd, int flags, void *arg, uint64_t txg) { ASSERT(vd == vd->vdev_top); /* indirect vdevs don't have metaslabs or dtls */ ASSERT(vdev_is_concrete(vd) || flags == 0); ASSERT(ISP2(flags)); ASSERT(spa_writeable(vd->vdev_spa)); if (flags & VDD_METASLAB) (void) txg_list_add(&vd->vdev_ms_list, arg, txg); if (flags & VDD_DTL) (void) txg_list_add(&vd->vdev_dtl_list, arg, txg); (void) txg_list_add(&vd->vdev_spa->spa_vdev_txg_list, vd, txg); } void vdev_dirty_leaves(vdev_t *vd, int flags, uint64_t txg) { for (int c = 0; c < vd->vdev_children; c++) vdev_dirty_leaves(vd->vdev_child[c], flags, txg); if (vd->vdev_ops->vdev_op_leaf) vdev_dirty(vd->vdev_top, flags, vd, txg); } /* * DTLs. * * A vdev's DTL (dirty time log) is the set of transaction groups for which * the vdev has less than perfect replication. There are four kinds of DTL: * * DTL_MISSING: txgs for which the vdev has no valid copies of the data * * DTL_PARTIAL: txgs for which data is available, but not fully replicated * * DTL_SCRUB: the txgs that could not be repaired by the last scrub; upon * scrub completion, DTL_SCRUB replaces DTL_MISSING in the range of * txgs that was scrubbed. * * DTL_OUTAGE: txgs which cannot currently be read, whether due to * persistent errors or just some device being offline. * Unlike the other three, the DTL_OUTAGE map is not generally * maintained; it's only computed when needed, typically to * determine whether a device can be detached. * * For leaf vdevs, DTL_MISSING and DTL_PARTIAL are identical: the device * either has the data or it doesn't. * * For interior vdevs such as mirror and RAID-Z the picture is more complex. * A vdev's DTL_PARTIAL is the union of its children's DTL_PARTIALs, because * if any child is less than fully replicated, then so is its parent. * A vdev's DTL_MISSING is a modified union of its children's DTL_MISSINGs, * comprising only those txgs which appear in 'maxfaults' or more children; * those are the txgs we don't have enough replication to read. For example, * double-parity RAID-Z can tolerate up to two missing devices (maxfaults == 2); * thus, its DTL_MISSING consists of the set of txgs that appear in more than * two child DTL_MISSING maps. * * It should be clear from the above that to compute the DTLs and outage maps * for all vdevs, it suffices to know just the leaf vdevs' DTL_MISSING maps. * Therefore, that is all we keep on disk. When loading the pool, or after * a configuration change, we generate all other DTLs from first principles. */ void vdev_dtl_dirty(vdev_t *vd, vdev_dtl_type_t t, uint64_t txg, uint64_t size) { range_tree_t *rt = vd->vdev_dtl[t]; ASSERT(t < DTL_TYPES); ASSERT(vd != vd->vdev_spa->spa_root_vdev); ASSERT(spa_writeable(vd->vdev_spa)); mutex_enter(&vd->vdev_dtl_lock); if (!range_tree_contains(rt, txg, size)) range_tree_add(rt, txg, size); mutex_exit(&vd->vdev_dtl_lock); } boolean_t vdev_dtl_contains(vdev_t *vd, vdev_dtl_type_t t, uint64_t txg, uint64_t size) { range_tree_t *rt = vd->vdev_dtl[t]; boolean_t dirty = B_FALSE; ASSERT(t < DTL_TYPES); ASSERT(vd != vd->vdev_spa->spa_root_vdev); /* * While we are loading the pool, the DTLs have not been loaded yet. * This isn't a problem but it can result in devices being tried * which are known to not have the data. In which case, the import * is relying on the checksum to ensure that we get the right data. * Note that while importing we are only reading the MOS, which is * always checksummed. */ mutex_enter(&vd->vdev_dtl_lock); if (!range_tree_is_empty(rt)) dirty = range_tree_contains(rt, txg, size); mutex_exit(&vd->vdev_dtl_lock); return (dirty); } boolean_t vdev_dtl_empty(vdev_t *vd, vdev_dtl_type_t t) { range_tree_t *rt = vd->vdev_dtl[t]; boolean_t empty; mutex_enter(&vd->vdev_dtl_lock); empty = range_tree_is_empty(rt); mutex_exit(&vd->vdev_dtl_lock); return (empty); } /* * Check if the txg falls within the range which must be * resilvered. DVAs outside this range can always be skipped. */ boolean_t vdev_default_need_resilver(vdev_t *vd, const dva_t *dva, size_t psize, uint64_t phys_birth) { (void) dva, (void) psize; /* Set by sequential resilver. */ if (phys_birth == TXG_UNKNOWN) return (B_TRUE); return (vdev_dtl_contains(vd, DTL_PARTIAL, phys_birth, 1)); } /* * Returns B_TRUE if the vdev determines the DVA needs to be resilvered. */ boolean_t vdev_dtl_need_resilver(vdev_t *vd, const dva_t *dva, size_t psize, uint64_t phys_birth) { ASSERT(vd != vd->vdev_spa->spa_root_vdev); if (vd->vdev_ops->vdev_op_need_resilver == NULL || vd->vdev_ops->vdev_op_leaf) return (B_TRUE); return (vd->vdev_ops->vdev_op_need_resilver(vd, dva, psize, phys_birth)); } /* * Returns the lowest txg in the DTL range. */ static uint64_t vdev_dtl_min(vdev_t *vd) { ASSERT(MUTEX_HELD(&vd->vdev_dtl_lock)); ASSERT3U(range_tree_space(vd->vdev_dtl[DTL_MISSING]), !=, 0); ASSERT0(vd->vdev_children); return (range_tree_min(vd->vdev_dtl[DTL_MISSING]) - 1); } /* * Returns the highest txg in the DTL. */ static uint64_t vdev_dtl_max(vdev_t *vd) { ASSERT(MUTEX_HELD(&vd->vdev_dtl_lock)); ASSERT3U(range_tree_space(vd->vdev_dtl[DTL_MISSING]), !=, 0); ASSERT0(vd->vdev_children); return (range_tree_max(vd->vdev_dtl[DTL_MISSING])); } /* * Determine if a resilvering vdev should remove any DTL entries from * its range. If the vdev was resilvering for the entire duration of the * scan then it should excise that range from its DTLs. Otherwise, this * vdev is considered partially resilvered and should leave its DTL * entries intact. The comment in vdev_dtl_reassess() describes how we * excise the DTLs. */ static boolean_t vdev_dtl_should_excise(vdev_t *vd, boolean_t rebuild_done) { ASSERT0(vd->vdev_children); if (vd->vdev_state < VDEV_STATE_DEGRADED) return (B_FALSE); if (vd->vdev_resilver_deferred) return (B_FALSE); if (range_tree_is_empty(vd->vdev_dtl[DTL_MISSING])) return (B_TRUE); if (rebuild_done) { vdev_rebuild_t *vr = &vd->vdev_top->vdev_rebuild_config; vdev_rebuild_phys_t *vrp = &vr->vr_rebuild_phys; /* Rebuild not initiated by attach */ if (vd->vdev_rebuild_txg == 0) return (B_TRUE); /* * When a rebuild completes without error then all missing data * up to the rebuild max txg has been reconstructed and the DTL * is eligible for excision. */ if (vrp->vrp_rebuild_state == VDEV_REBUILD_COMPLETE && vdev_dtl_max(vd) <= vrp->vrp_max_txg) { ASSERT3U(vrp->vrp_min_txg, <=, vdev_dtl_min(vd)); ASSERT3U(vrp->vrp_min_txg, <, vd->vdev_rebuild_txg); ASSERT3U(vd->vdev_rebuild_txg, <=, vrp->vrp_max_txg); return (B_TRUE); } } else { dsl_scan_t *scn = vd->vdev_spa->spa_dsl_pool->dp_scan; dsl_scan_phys_t *scnp __maybe_unused = &scn->scn_phys; /* Resilver not initiated by attach */ if (vd->vdev_resilver_txg == 0) return (B_TRUE); /* * When a resilver is initiated the scan will assign the * scn_max_txg value to the highest txg value that exists * in all DTLs. If this device's max DTL is not part of this * scan (i.e. it is not in the range (scn_min_txg, scn_max_txg] * then it is not eligible for excision. */ if (vdev_dtl_max(vd) <= scn->scn_phys.scn_max_txg) { ASSERT3U(scnp->scn_min_txg, <=, vdev_dtl_min(vd)); ASSERT3U(scnp->scn_min_txg, <, vd->vdev_resilver_txg); ASSERT3U(vd->vdev_resilver_txg, <=, scnp->scn_max_txg); return (B_TRUE); } } return (B_FALSE); } /* * Reassess DTLs after a config change or scrub completion. If txg == 0 no * write operations will be issued to the pool. */ void vdev_dtl_reassess(vdev_t *vd, uint64_t txg, uint64_t scrub_txg, boolean_t scrub_done, boolean_t rebuild_done) { spa_t *spa = vd->vdev_spa; avl_tree_t reftree; int minref; ASSERT(spa_config_held(spa, SCL_ALL, RW_READER) != 0); for (int c = 0; c < vd->vdev_children; c++) vdev_dtl_reassess(vd->vdev_child[c], txg, scrub_txg, scrub_done, rebuild_done); if (vd == spa->spa_root_vdev || !vdev_is_concrete(vd) || vd->vdev_aux) return; if (vd->vdev_ops->vdev_op_leaf) { dsl_scan_t *scn = spa->spa_dsl_pool->dp_scan; vdev_rebuild_t *vr = &vd->vdev_top->vdev_rebuild_config; boolean_t check_excise = B_FALSE; boolean_t wasempty = B_TRUE; mutex_enter(&vd->vdev_dtl_lock); /* * If requested, pretend the scan or rebuild completed cleanly. */ if (zfs_scan_ignore_errors) { if (scn != NULL) scn->scn_phys.scn_errors = 0; if (vr != NULL) vr->vr_rebuild_phys.vrp_errors = 0; } if (scrub_txg != 0 && !range_tree_is_empty(vd->vdev_dtl[DTL_MISSING])) { wasempty = B_FALSE; zfs_dbgmsg("guid:%llu txg:%llu scrub:%llu started:%d " "dtl:%llu/%llu errors:%llu", (u_longlong_t)vd->vdev_guid, (u_longlong_t)txg, (u_longlong_t)scrub_txg, spa->spa_scrub_started, (u_longlong_t)vdev_dtl_min(vd), (u_longlong_t)vdev_dtl_max(vd), (u_longlong_t)(scn ? scn->scn_phys.scn_errors : 0)); } /* * If we've completed a scrub/resilver or a rebuild cleanly * then determine if this vdev should remove any DTLs. We * only want to excise regions on vdevs that were available * during the entire duration of this scan. */ if (rebuild_done && vr != NULL && vr->vr_rebuild_phys.vrp_errors == 0) { check_excise = B_TRUE; } else { if (spa->spa_scrub_started || (scn != NULL && scn->scn_phys.scn_errors == 0)) { check_excise = B_TRUE; } } if (scrub_txg && check_excise && vdev_dtl_should_excise(vd, rebuild_done)) { /* * We completed a scrub, resilver or rebuild up to * scrub_txg. If we did it without rebooting, then * the scrub dtl will be valid, so excise the old * region and fold in the scrub dtl. Otherwise, * leave the dtl as-is if there was an error. * * There's little trick here: to excise the beginning * of the DTL_MISSING map, we put it into a reference * tree and then add a segment with refcnt -1 that * covers the range [0, scrub_txg). This means * that each txg in that range has refcnt -1 or 0. * We then add DTL_SCRUB with a refcnt of 2, so that * entries in the range [0, scrub_txg) will have a * positive refcnt -- either 1 or 2. We then convert * the reference tree into the new DTL_MISSING map. */ space_reftree_create(&reftree); space_reftree_add_map(&reftree, vd->vdev_dtl[DTL_MISSING], 1); space_reftree_add_seg(&reftree, 0, scrub_txg, -1); space_reftree_add_map(&reftree, vd->vdev_dtl[DTL_SCRUB], 2); space_reftree_generate_map(&reftree, vd->vdev_dtl[DTL_MISSING], 1); space_reftree_destroy(&reftree); if (!range_tree_is_empty(vd->vdev_dtl[DTL_MISSING])) { zfs_dbgmsg("update DTL_MISSING:%llu/%llu", (u_longlong_t)vdev_dtl_min(vd), (u_longlong_t)vdev_dtl_max(vd)); } else if (!wasempty) { zfs_dbgmsg("DTL_MISSING is now empty"); } } range_tree_vacate(vd->vdev_dtl[DTL_PARTIAL], NULL, NULL); range_tree_walk(vd->vdev_dtl[DTL_MISSING], range_tree_add, vd->vdev_dtl[DTL_PARTIAL]); if (scrub_done) range_tree_vacate(vd->vdev_dtl[DTL_SCRUB], NULL, NULL); range_tree_vacate(vd->vdev_dtl[DTL_OUTAGE], NULL, NULL); if (!vdev_readable(vd)) range_tree_add(vd->vdev_dtl[DTL_OUTAGE], 0, -1ULL); else range_tree_walk(vd->vdev_dtl[DTL_MISSING], range_tree_add, vd->vdev_dtl[DTL_OUTAGE]); /* * If the vdev was resilvering or rebuilding and no longer * has any DTLs then reset the appropriate flag and dirty * the top level so that we persist the change. */ if (txg != 0 && range_tree_is_empty(vd->vdev_dtl[DTL_MISSING]) && range_tree_is_empty(vd->vdev_dtl[DTL_OUTAGE])) { if (vd->vdev_rebuild_txg != 0) { vd->vdev_rebuild_txg = 0; vdev_config_dirty(vd->vdev_top); } else if (vd->vdev_resilver_txg != 0) { vd->vdev_resilver_txg = 0; vdev_config_dirty(vd->vdev_top); } } mutex_exit(&vd->vdev_dtl_lock); if (txg != 0) vdev_dirty(vd->vdev_top, VDD_DTL, vd, txg); return; } mutex_enter(&vd->vdev_dtl_lock); for (int t = 0; t < DTL_TYPES; t++) { /* account for child's outage in parent's missing map */ int s = (t == DTL_MISSING) ? DTL_OUTAGE: t; if (t == DTL_SCRUB) continue; /* leaf vdevs only */ if (t == DTL_PARTIAL) minref = 1; /* i.e. non-zero */ else if (vdev_get_nparity(vd) != 0) minref = vdev_get_nparity(vd) + 1; /* RAID-Z, dRAID */ else minref = vd->vdev_children; /* any kind of mirror */ space_reftree_create(&reftree); for (int c = 0; c < vd->vdev_children; c++) { vdev_t *cvd = vd->vdev_child[c]; mutex_enter(&cvd->vdev_dtl_lock); space_reftree_add_map(&reftree, cvd->vdev_dtl[s], 1); mutex_exit(&cvd->vdev_dtl_lock); } space_reftree_generate_map(&reftree, vd->vdev_dtl[t], minref); space_reftree_destroy(&reftree); } mutex_exit(&vd->vdev_dtl_lock); } /* * Iterate over all the vdevs except spare, and post kobj events */ void vdev_post_kobj_evt(vdev_t *vd) { if (vd->vdev_ops->vdev_op_kobj_evt_post && vd->vdev_kobj_flag == B_FALSE) { vd->vdev_kobj_flag = B_TRUE; vd->vdev_ops->vdev_op_kobj_evt_post(vd); } for (int c = 0; c < vd->vdev_children; c++) vdev_post_kobj_evt(vd->vdev_child[c]); } /* * Iterate over all the vdevs except spare, and clear kobj events */ void vdev_clear_kobj_evt(vdev_t *vd) { vd->vdev_kobj_flag = B_FALSE; for (int c = 0; c < vd->vdev_children; c++) vdev_clear_kobj_evt(vd->vdev_child[c]); } int vdev_dtl_load(vdev_t *vd) { spa_t *spa = vd->vdev_spa; objset_t *mos = spa->spa_meta_objset; range_tree_t *rt; int error = 0; if (vd->vdev_ops->vdev_op_leaf && vd->vdev_dtl_object != 0) { ASSERT(vdev_is_concrete(vd)); /* * If the dtl cannot be sync'd there is no need to open it. */ if (spa->spa_mode == SPA_MODE_READ && !spa->spa_read_spacemaps) return (0); error = space_map_open(&vd->vdev_dtl_sm, mos, vd->vdev_dtl_object, 0, -1ULL, 0); if (error) return (error); ASSERT(vd->vdev_dtl_sm != NULL); rt = range_tree_create(NULL, RANGE_SEG64, NULL, 0, 0); error = space_map_load(vd->vdev_dtl_sm, rt, SM_ALLOC); if (error == 0) { mutex_enter(&vd->vdev_dtl_lock); range_tree_walk(rt, range_tree_add, vd->vdev_dtl[DTL_MISSING]); mutex_exit(&vd->vdev_dtl_lock); } range_tree_vacate(rt, NULL, NULL); range_tree_destroy(rt); return (error); } for (int c = 0; c < vd->vdev_children; c++) { error = vdev_dtl_load(vd->vdev_child[c]); if (error != 0) break; } return (error); } static void vdev_zap_allocation_data(vdev_t *vd, dmu_tx_t *tx) { spa_t *spa = vd->vdev_spa; objset_t *mos = spa->spa_meta_objset; vdev_alloc_bias_t alloc_bias = vd->vdev_alloc_bias; const char *string; ASSERT(alloc_bias != VDEV_BIAS_NONE); string = (alloc_bias == VDEV_BIAS_LOG) ? VDEV_ALLOC_BIAS_LOG : (alloc_bias == VDEV_BIAS_SPECIAL) ? VDEV_ALLOC_BIAS_SPECIAL : (alloc_bias == VDEV_BIAS_DEDUP) ? VDEV_ALLOC_BIAS_DEDUP : NULL; ASSERT(string != NULL); VERIFY0(zap_add(mos, vd->vdev_top_zap, VDEV_TOP_ZAP_ALLOCATION_BIAS, 1, strlen(string) + 1, string, tx)); if (alloc_bias == VDEV_BIAS_SPECIAL || alloc_bias == VDEV_BIAS_DEDUP) { spa_activate_allocation_classes(spa, tx); } } void vdev_destroy_unlink_zap(vdev_t *vd, uint64_t zapobj, dmu_tx_t *tx) { spa_t *spa = vd->vdev_spa; VERIFY0(zap_destroy(spa->spa_meta_objset, zapobj, tx)); VERIFY0(zap_remove_int(spa->spa_meta_objset, spa->spa_all_vdev_zaps, zapobj, tx)); } uint64_t vdev_create_link_zap(vdev_t *vd, dmu_tx_t *tx) { spa_t *spa = vd->vdev_spa; uint64_t zap = zap_create(spa->spa_meta_objset, DMU_OTN_ZAP_METADATA, DMU_OT_NONE, 0, tx); ASSERT(zap != 0); VERIFY0(zap_add_int(spa->spa_meta_objset, spa->spa_all_vdev_zaps, zap, tx)); return (zap); } void vdev_construct_zaps(vdev_t *vd, dmu_tx_t *tx) { if (vd->vdev_ops != &vdev_hole_ops && vd->vdev_ops != &vdev_missing_ops && vd->vdev_ops != &vdev_root_ops && !vd->vdev_top->vdev_removing) { if (vd->vdev_ops->vdev_op_leaf && vd->vdev_leaf_zap == 0) { vd->vdev_leaf_zap = vdev_create_link_zap(vd, tx); } if (vd == vd->vdev_top && vd->vdev_top_zap == 0) { vd->vdev_top_zap = vdev_create_link_zap(vd, tx); if (vd->vdev_alloc_bias != VDEV_BIAS_NONE) vdev_zap_allocation_data(vd, tx); } } if (vd->vdev_ops == &vdev_root_ops && vd->vdev_root_zap == 0 && spa_feature_is_enabled(vd->vdev_spa, SPA_FEATURE_AVZ_V2)) { if (!spa_feature_is_active(vd->vdev_spa, SPA_FEATURE_AVZ_V2)) spa_feature_incr(vd->vdev_spa, SPA_FEATURE_AVZ_V2, tx); vd->vdev_root_zap = vdev_create_link_zap(vd, tx); } for (uint64_t i = 0; i < vd->vdev_children; i++) { vdev_construct_zaps(vd->vdev_child[i], tx); } } static void vdev_dtl_sync(vdev_t *vd, uint64_t txg) { spa_t *spa = vd->vdev_spa; range_tree_t *rt = vd->vdev_dtl[DTL_MISSING]; objset_t *mos = spa->spa_meta_objset; range_tree_t *rtsync; dmu_tx_t *tx; uint64_t object = space_map_object(vd->vdev_dtl_sm); ASSERT(vdev_is_concrete(vd)); ASSERT(vd->vdev_ops->vdev_op_leaf); tx = dmu_tx_create_assigned(spa->spa_dsl_pool, txg); if (vd->vdev_detached || vd->vdev_top->vdev_removing) { mutex_enter(&vd->vdev_dtl_lock); space_map_free(vd->vdev_dtl_sm, tx); space_map_close(vd->vdev_dtl_sm); vd->vdev_dtl_sm = NULL; mutex_exit(&vd->vdev_dtl_lock); /* * We only destroy the leaf ZAP for detached leaves or for * removed log devices. Removed data devices handle leaf ZAP * cleanup later, once cancellation is no longer possible. */ if (vd->vdev_leaf_zap != 0 && (vd->vdev_detached || vd->vdev_top->vdev_islog)) { vdev_destroy_unlink_zap(vd, vd->vdev_leaf_zap, tx); vd->vdev_leaf_zap = 0; } dmu_tx_commit(tx); return; } if (vd->vdev_dtl_sm == NULL) { uint64_t new_object; new_object = space_map_alloc(mos, zfs_vdev_dtl_sm_blksz, tx); VERIFY3U(new_object, !=, 0); VERIFY0(space_map_open(&vd->vdev_dtl_sm, mos, new_object, 0, -1ULL, 0)); ASSERT(vd->vdev_dtl_sm != NULL); } rtsync = range_tree_create(NULL, RANGE_SEG64, NULL, 0, 0); mutex_enter(&vd->vdev_dtl_lock); range_tree_walk(rt, range_tree_add, rtsync); mutex_exit(&vd->vdev_dtl_lock); space_map_truncate(vd->vdev_dtl_sm, zfs_vdev_dtl_sm_blksz, tx); space_map_write(vd->vdev_dtl_sm, rtsync, SM_ALLOC, SM_NO_VDEVID, tx); range_tree_vacate(rtsync, NULL, NULL); range_tree_destroy(rtsync); /* * If the object for the space map has changed then dirty * the top level so that we update the config. */ if (object != space_map_object(vd->vdev_dtl_sm)) { vdev_dbgmsg(vd, "txg %llu, spa %s, DTL old object %llu, " "new object %llu", (u_longlong_t)txg, spa_name(spa), (u_longlong_t)object, (u_longlong_t)space_map_object(vd->vdev_dtl_sm)); vdev_config_dirty(vd->vdev_top); } dmu_tx_commit(tx); } /* * Determine whether the specified vdev can be offlined/detached/removed * without losing data. */ boolean_t vdev_dtl_required(vdev_t *vd) { spa_t *spa = vd->vdev_spa; vdev_t *tvd = vd->vdev_top; uint8_t cant_read = vd->vdev_cant_read; boolean_t required; ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL); if (vd == spa->spa_root_vdev || vd == tvd) return (B_TRUE); /* * Temporarily mark the device as unreadable, and then determine * whether this results in any DTL outages in the top-level vdev. * If not, we can safely offline/detach/remove the device. */ vd->vdev_cant_read = B_TRUE; vdev_dtl_reassess(tvd, 0, 0, B_FALSE, B_FALSE); required = !vdev_dtl_empty(tvd, DTL_OUTAGE); vd->vdev_cant_read = cant_read; vdev_dtl_reassess(tvd, 0, 0, B_FALSE, B_FALSE); if (!required && zio_injection_enabled) { required = !!zio_handle_device_injection(vd, NULL, SET_ERROR(ECHILD)); } return (required); } /* * Determine if resilver is needed, and if so the txg range. */ boolean_t vdev_resilver_needed(vdev_t *vd, uint64_t *minp, uint64_t *maxp) { boolean_t needed = B_FALSE; uint64_t thismin = UINT64_MAX; uint64_t thismax = 0; if (vd->vdev_children == 0) { mutex_enter(&vd->vdev_dtl_lock); if (!range_tree_is_empty(vd->vdev_dtl[DTL_MISSING]) && vdev_writeable(vd)) { thismin = vdev_dtl_min(vd); thismax = vdev_dtl_max(vd); needed = B_TRUE; } mutex_exit(&vd->vdev_dtl_lock); } else { for (int c = 0; c < vd->vdev_children; c++) { vdev_t *cvd = vd->vdev_child[c]; uint64_t cmin, cmax; if (vdev_resilver_needed(cvd, &cmin, &cmax)) { thismin = MIN(thismin, cmin); thismax = MAX(thismax, cmax); needed = B_TRUE; } } } if (needed && minp) { *minp = thismin; *maxp = thismax; } return (needed); } /* * Gets the checkpoint space map object from the vdev's ZAP. On success sm_obj * will contain either the checkpoint spacemap object or zero if none exists. * All other errors are returned to the caller. */ int vdev_checkpoint_sm_object(vdev_t *vd, uint64_t *sm_obj) { ASSERT0(spa_config_held(vd->vdev_spa, SCL_ALL, RW_WRITER)); if (vd->vdev_top_zap == 0) { *sm_obj = 0; return (0); } int error = zap_lookup(spa_meta_objset(vd->vdev_spa), vd->vdev_top_zap, VDEV_TOP_ZAP_POOL_CHECKPOINT_SM, sizeof (uint64_t), 1, sm_obj); if (error == ENOENT) { *sm_obj = 0; error = 0; } return (error); } int vdev_load(vdev_t *vd) { int children = vd->vdev_children; int error = 0; taskq_t *tq = NULL; /* * It's only worthwhile to use the taskq for the root vdev, because the * slow part is metaslab_init, and that only happens for top-level * vdevs. */ if (vd->vdev_ops == &vdev_root_ops && vd->vdev_children > 0) { tq = taskq_create("vdev_load", children, minclsyspri, children, children, TASKQ_PREPOPULATE); } /* * Recursively load all children. */ for (int c = 0; c < vd->vdev_children; c++) { vdev_t *cvd = vd->vdev_child[c]; if (tq == NULL || vdev_uses_zvols(cvd)) { cvd->vdev_load_error = vdev_load(cvd); } else { VERIFY(taskq_dispatch(tq, vdev_load_child, cvd, TQ_SLEEP) != TASKQID_INVALID); } } if (tq != NULL) { taskq_wait(tq); taskq_destroy(tq); } for (int c = 0; c < vd->vdev_children; c++) { int error = vd->vdev_child[c]->vdev_load_error; if (error != 0) return (error); } vdev_set_deflate_ratio(vd); /* * On spa_load path, grab the allocation bias from our zap */ if (vd == vd->vdev_top && vd->vdev_top_zap != 0) { spa_t *spa = vd->vdev_spa; char bias_str[64]; error = zap_lookup(spa->spa_meta_objset, vd->vdev_top_zap, VDEV_TOP_ZAP_ALLOCATION_BIAS, 1, sizeof (bias_str), bias_str); if (error == 0) { ASSERT(vd->vdev_alloc_bias == VDEV_BIAS_NONE); vd->vdev_alloc_bias = vdev_derive_alloc_bias(bias_str); } else if (error != ENOENT) { vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); vdev_dbgmsg(vd, "vdev_load: zap_lookup(top_zap=%llu) " "failed [error=%d]", (u_longlong_t)vd->vdev_top_zap, error); return (error); } } if (vd == vd->vdev_top && vd->vdev_top_zap != 0) { spa_t *spa = vd->vdev_spa; uint64_t failfast; error = zap_lookup(spa->spa_meta_objset, vd->vdev_top_zap, vdev_prop_to_name(VDEV_PROP_FAILFAST), sizeof (failfast), 1, &failfast); if (error == 0) { vd->vdev_failfast = failfast & 1; } else if (error == ENOENT) { vd->vdev_failfast = vdev_prop_default_numeric( VDEV_PROP_FAILFAST); } else { vdev_dbgmsg(vd, "vdev_load: zap_lookup(top_zap=%llu) " "failed [error=%d]", (u_longlong_t)vd->vdev_top_zap, error); } } /* * Load any rebuild state from the top-level vdev zap. */ if (vd == vd->vdev_top && vd->vdev_top_zap != 0) { error = vdev_rebuild_load(vd); if (error && error != ENOTSUP) { vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); vdev_dbgmsg(vd, "vdev_load: vdev_rebuild_load " "failed [error=%d]", error); return (error); } } if (vd->vdev_top_zap != 0 || vd->vdev_leaf_zap != 0) { uint64_t zapobj; if (vd->vdev_top_zap != 0) zapobj = vd->vdev_top_zap; else zapobj = vd->vdev_leaf_zap; error = vdev_prop_get_int(vd, VDEV_PROP_CHECKSUM_N, &vd->vdev_checksum_n); if (error && error != ENOENT) vdev_dbgmsg(vd, "vdev_load: zap_lookup(zap=%llu) " "failed [error=%d]", (u_longlong_t)zapobj, error); error = vdev_prop_get_int(vd, VDEV_PROP_CHECKSUM_T, &vd->vdev_checksum_t); if (error && error != ENOENT) vdev_dbgmsg(vd, "vdev_load: zap_lookup(zap=%llu) " "failed [error=%d]", (u_longlong_t)zapobj, error); error = vdev_prop_get_int(vd, VDEV_PROP_IO_N, &vd->vdev_io_n); if (error && error != ENOENT) vdev_dbgmsg(vd, "vdev_load: zap_lookup(zap=%llu) " "failed [error=%d]", (u_longlong_t)zapobj, error); error = vdev_prop_get_int(vd, VDEV_PROP_IO_T, &vd->vdev_io_t); if (error && error != ENOENT) vdev_dbgmsg(vd, "vdev_load: zap_lookup(zap=%llu) " "failed [error=%d]", (u_longlong_t)zapobj, error); } /* * If this is a top-level vdev, initialize its metaslabs. */ if (vd == vd->vdev_top && vdev_is_concrete(vd)) { vdev_metaslab_group_create(vd); if (vd->vdev_ashift == 0 || vd->vdev_asize == 0) { vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); vdev_dbgmsg(vd, "vdev_load: invalid size. ashift=%llu, " "asize=%llu", (u_longlong_t)vd->vdev_ashift, (u_longlong_t)vd->vdev_asize); return (SET_ERROR(ENXIO)); } error = vdev_metaslab_init(vd, 0); if (error != 0) { vdev_dbgmsg(vd, "vdev_load: metaslab_init failed " "[error=%d]", error); vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); return (error); } uint64_t checkpoint_sm_obj; error = vdev_checkpoint_sm_object(vd, &checkpoint_sm_obj); if (error == 0 && checkpoint_sm_obj != 0) { objset_t *mos = spa_meta_objset(vd->vdev_spa); ASSERT(vd->vdev_asize != 0); ASSERT3P(vd->vdev_checkpoint_sm, ==, NULL); error = space_map_open(&vd->vdev_checkpoint_sm, mos, checkpoint_sm_obj, 0, vd->vdev_asize, vd->vdev_ashift); if (error != 0) { vdev_dbgmsg(vd, "vdev_load: space_map_open " "failed for checkpoint spacemap (obj %llu) " "[error=%d]", (u_longlong_t)checkpoint_sm_obj, error); return (error); } ASSERT3P(vd->vdev_checkpoint_sm, !=, NULL); /* * Since the checkpoint_sm contains free entries * exclusively we can use space_map_allocated() to * indicate the cumulative checkpointed space that * has been freed. */ vd->vdev_stat.vs_checkpoint_space = -space_map_allocated(vd->vdev_checkpoint_sm); vd->vdev_spa->spa_checkpoint_info.sci_dspace += vd->vdev_stat.vs_checkpoint_space; } else if (error != 0) { vdev_dbgmsg(vd, "vdev_load: failed to retrieve " "checkpoint space map object from vdev ZAP " "[error=%d]", error); return (error); } } /* * If this is a leaf vdev, load its DTL. */ if (vd->vdev_ops->vdev_op_leaf && (error = vdev_dtl_load(vd)) != 0) { vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); vdev_dbgmsg(vd, "vdev_load: vdev_dtl_load failed " "[error=%d]", error); return (error); } uint64_t obsolete_sm_object; error = vdev_obsolete_sm_object(vd, &obsolete_sm_object); if (error == 0 && obsolete_sm_object != 0) { objset_t *mos = vd->vdev_spa->spa_meta_objset; ASSERT(vd->vdev_asize != 0); ASSERT3P(vd->vdev_obsolete_sm, ==, NULL); if ((error = space_map_open(&vd->vdev_obsolete_sm, mos, obsolete_sm_object, 0, vd->vdev_asize, 0))) { vdev_set_state(vd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); vdev_dbgmsg(vd, "vdev_load: space_map_open failed for " "obsolete spacemap (obj %llu) [error=%d]", (u_longlong_t)obsolete_sm_object, error); return (error); } } else if (error != 0) { vdev_dbgmsg(vd, "vdev_load: failed to retrieve obsolete " "space map object from vdev ZAP [error=%d]", error); return (error); } return (0); } /* * The special vdev case is used for hot spares and l2cache devices. Its * sole purpose it to set the vdev state for the associated vdev. To do this, * we make sure that we can open the underlying device, then try to read the * label, and make sure that the label is sane and that it hasn't been * repurposed to another pool. */ int vdev_validate_aux(vdev_t *vd) { nvlist_t *label; uint64_t guid, version; uint64_t state; if (!vdev_readable(vd)) return (0); if ((label = vdev_label_read_config(vd, -1ULL)) == NULL) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); return (-1); } if (nvlist_lookup_uint64(label, ZPOOL_CONFIG_VERSION, &version) != 0 || !SPA_VERSION_IS_SUPPORTED(version) || nvlist_lookup_uint64(label, ZPOOL_CONFIG_GUID, &guid) != 0 || guid != vd->vdev_guid || nvlist_lookup_uint64(label, ZPOOL_CONFIG_POOL_STATE, &state) != 0) { vdev_set_state(vd, B_TRUE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); nvlist_free(label); return (-1); } /* * We don't actually check the pool state here. If it's in fact in * use by another pool, we update this fact on the fly when requested. */ nvlist_free(label); return (0); } static void vdev_destroy_ms_flush_data(vdev_t *vd, dmu_tx_t *tx) { objset_t *mos = spa_meta_objset(vd->vdev_spa); if (vd->vdev_top_zap == 0) return; uint64_t object = 0; int err = zap_lookup(mos, vd->vdev_top_zap, VDEV_TOP_ZAP_MS_UNFLUSHED_PHYS_TXGS, sizeof (uint64_t), 1, &object); if (err == ENOENT) return; VERIFY0(err); VERIFY0(dmu_object_free(mos, object, tx)); VERIFY0(zap_remove(mos, vd->vdev_top_zap, VDEV_TOP_ZAP_MS_UNFLUSHED_PHYS_TXGS, tx)); } /* * Free the objects used to store this vdev's spacemaps, and the array * that points to them. */ void vdev_destroy_spacemaps(vdev_t *vd, dmu_tx_t *tx) { if (vd->vdev_ms_array == 0) return; objset_t *mos = vd->vdev_spa->spa_meta_objset; uint64_t array_count = vd->vdev_asize >> vd->vdev_ms_shift; size_t array_bytes = array_count * sizeof (uint64_t); uint64_t *smobj_array = kmem_alloc(array_bytes, KM_SLEEP); VERIFY0(dmu_read(mos, vd->vdev_ms_array, 0, array_bytes, smobj_array, 0)); for (uint64_t i = 0; i < array_count; i++) { uint64_t smobj = smobj_array[i]; if (smobj == 0) continue; space_map_free_obj(mos, smobj, tx); } kmem_free(smobj_array, array_bytes); VERIFY0(dmu_object_free(mos, vd->vdev_ms_array, tx)); vdev_destroy_ms_flush_data(vd, tx); vd->vdev_ms_array = 0; } static void vdev_remove_empty_log(vdev_t *vd, uint64_t txg) { spa_t *spa = vd->vdev_spa; ASSERT(vd->vdev_islog); ASSERT(vd == vd->vdev_top); ASSERT3U(txg, ==, spa_syncing_txg(spa)); dmu_tx_t *tx = dmu_tx_create_assigned(spa_get_dsl(spa), txg); vdev_destroy_spacemaps(vd, tx); if (vd->vdev_top_zap != 0) { vdev_destroy_unlink_zap(vd, vd->vdev_top_zap, tx); vd->vdev_top_zap = 0; } dmu_tx_commit(tx); } void vdev_sync_done(vdev_t *vd, uint64_t txg) { metaslab_t *msp; boolean_t reassess = !txg_list_empty(&vd->vdev_ms_list, TXG_CLEAN(txg)); ASSERT(vdev_is_concrete(vd)); while ((msp = txg_list_remove(&vd->vdev_ms_list, TXG_CLEAN(txg))) != NULL) metaslab_sync_done(msp, txg); if (reassess) { metaslab_sync_reassess(vd->vdev_mg); if (vd->vdev_log_mg != NULL) metaslab_sync_reassess(vd->vdev_log_mg); } } void vdev_sync(vdev_t *vd, uint64_t txg) { spa_t *spa = vd->vdev_spa; vdev_t *lvd; metaslab_t *msp; ASSERT3U(txg, ==, spa->spa_syncing_txg); dmu_tx_t *tx = dmu_tx_create_assigned(spa->spa_dsl_pool, txg); if (range_tree_space(vd->vdev_obsolete_segments) > 0) { ASSERT(vd->vdev_removing || vd->vdev_ops == &vdev_indirect_ops); vdev_indirect_sync_obsolete(vd, tx); /* * If the vdev is indirect, it can't have dirty * metaslabs or DTLs. */ if (vd->vdev_ops == &vdev_indirect_ops) { ASSERT(txg_list_empty(&vd->vdev_ms_list, txg)); ASSERT(txg_list_empty(&vd->vdev_dtl_list, txg)); dmu_tx_commit(tx); return; } } ASSERT(vdev_is_concrete(vd)); if (vd->vdev_ms_array == 0 && vd->vdev_ms_shift != 0 && !vd->vdev_removing) { ASSERT(vd == vd->vdev_top); ASSERT0(vd->vdev_indirect_config.vic_mapping_object); vd->vdev_ms_array = dmu_object_alloc(spa->spa_meta_objset, DMU_OT_OBJECT_ARRAY, 0, DMU_OT_NONE, 0, tx); ASSERT(vd->vdev_ms_array != 0); vdev_config_dirty(vd); } while ((msp = txg_list_remove(&vd->vdev_ms_list, txg)) != NULL) { metaslab_sync(msp, txg); (void) txg_list_add(&vd->vdev_ms_list, msp, TXG_CLEAN(txg)); } while ((lvd = txg_list_remove(&vd->vdev_dtl_list, txg)) != NULL) vdev_dtl_sync(lvd, txg); /* * If this is an empty log device being removed, destroy the * metadata associated with it. */ if (vd->vdev_islog && vd->vdev_stat.vs_alloc == 0 && vd->vdev_removing) vdev_remove_empty_log(vd, txg); (void) txg_list_add(&spa->spa_vdev_txg_list, vd, TXG_CLEAN(txg)); dmu_tx_commit(tx); } uint64_t vdev_psize_to_asize(vdev_t *vd, uint64_t psize) { return (vd->vdev_ops->vdev_op_asize(vd, psize)); } /* * Mark the given vdev faulted. A faulted vdev behaves as if the device could * not be opened, and no I/O is attempted. */ int vdev_fault(spa_t *spa, uint64_t guid, vdev_aux_t aux) { vdev_t *vd, *tvd; spa_vdev_state_enter(spa, SCL_NONE); if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL) return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENODEV))); if (!vd->vdev_ops->vdev_op_leaf) return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENOTSUP))); tvd = vd->vdev_top; /* * If user did a 'zpool offline -f' then make the fault persist across * reboots. */ if (aux == VDEV_AUX_EXTERNAL_PERSIST) { /* * There are two kinds of forced faults: temporary and * persistent. Temporary faults go away at pool import, while * persistent faults stay set. Both types of faults can be * cleared with a zpool clear. * * We tell if a vdev is persistently faulted by looking at the * ZPOOL_CONFIG_AUX_STATE nvpair. If it's set to "external" at * import then it's a persistent fault. Otherwise, it's * temporary. We get ZPOOL_CONFIG_AUX_STATE set to "external" * by setting vd.vdev_stat.vs_aux to VDEV_AUX_EXTERNAL. This * tells vdev_config_generate() (which gets run later) to set * ZPOOL_CONFIG_AUX_STATE to "external" in the nvlist. */ vd->vdev_stat.vs_aux = VDEV_AUX_EXTERNAL; vd->vdev_tmpoffline = B_FALSE; aux = VDEV_AUX_EXTERNAL; } else { vd->vdev_tmpoffline = B_TRUE; } /* * We don't directly use the aux state here, but if we do a * vdev_reopen(), we need this value to be present to remember why we * were faulted. */ vd->vdev_label_aux = aux; /* * Faulted state takes precedence over degraded. */ vd->vdev_delayed_close = B_FALSE; vd->vdev_faulted = 1ULL; vd->vdev_degraded = 0ULL; vdev_set_state(vd, B_FALSE, VDEV_STATE_FAULTED, aux); /* * If this device has the only valid copy of the data, then * back off and simply mark the vdev as degraded instead. */ if (!tvd->vdev_islog && vd->vdev_aux == NULL && vdev_dtl_required(vd)) { vd->vdev_degraded = 1ULL; vd->vdev_faulted = 0ULL; /* * If we reopen the device and it's not dead, only then do we * mark it degraded. */ vdev_reopen(tvd); if (vdev_readable(vd)) vdev_set_state(vd, B_FALSE, VDEV_STATE_DEGRADED, aux); } return (spa_vdev_state_exit(spa, vd, 0)); } /* * Mark the given vdev degraded. A degraded vdev is purely an indication to the * user that something is wrong. The vdev continues to operate as normal as far * as I/O is concerned. */ int vdev_degrade(spa_t *spa, uint64_t guid, vdev_aux_t aux) { vdev_t *vd; spa_vdev_state_enter(spa, SCL_NONE); if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL) return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENODEV))); if (!vd->vdev_ops->vdev_op_leaf) return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENOTSUP))); /* * If the vdev is already faulted, then don't do anything. */ if (vd->vdev_faulted || vd->vdev_degraded) return (spa_vdev_state_exit(spa, NULL, 0)); vd->vdev_degraded = 1ULL; if (!vdev_is_dead(vd)) vdev_set_state(vd, B_FALSE, VDEV_STATE_DEGRADED, aux); return (spa_vdev_state_exit(spa, vd, 0)); } int vdev_remove_wanted(spa_t *spa, uint64_t guid) { vdev_t *vd; spa_vdev_state_enter(spa, SCL_NONE); if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL) return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENODEV))); /* - * If the vdev is already removed, then don't do anything. + * If the vdev is already removed, or expanding which can trigger + * repartition add/remove events, then don't do anything. */ - if (vd->vdev_removed) + if (vd->vdev_removed || vd->vdev_expanding) return (spa_vdev_state_exit(spa, NULL, 0)); + /* + * Confirm the vdev has been removed, otherwise don't do anything. + */ + if (vd->vdev_ops->vdev_op_leaf && !zio_wait(vdev_probe(vd, NULL))) + return (spa_vdev_state_exit(spa, NULL, SET_ERROR(EEXIST))); + vd->vdev_remove_wanted = B_TRUE; spa_async_request(spa, SPA_ASYNC_REMOVE); return (spa_vdev_state_exit(spa, vd, 0)); } /* * Online the given vdev. * * If 'ZFS_ONLINE_UNSPARE' is set, it implies two things. First, any attached * spare device should be detached when the device finishes resilvering. * Second, the online should be treated like a 'test' online case, so no FMA * events are generated if the device fails to open. */ int vdev_online(spa_t *spa, uint64_t guid, uint64_t flags, vdev_state_t *newstate) { vdev_t *vd, *tvd, *pvd, *rvd = spa->spa_root_vdev; boolean_t wasoffline; vdev_state_t oldstate; spa_vdev_state_enter(spa, SCL_NONE); if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL) return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENODEV))); if (!vd->vdev_ops->vdev_op_leaf) return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENOTSUP))); wasoffline = (vd->vdev_offline || vd->vdev_tmpoffline); oldstate = vd->vdev_state; tvd = vd->vdev_top; vd->vdev_offline = B_FALSE; vd->vdev_tmpoffline = B_FALSE; vd->vdev_checkremove = !!(flags & ZFS_ONLINE_CHECKREMOVE); vd->vdev_forcefault = !!(flags & ZFS_ONLINE_FORCEFAULT); /* XXX - L2ARC 1.0 does not support expansion */ if (!vd->vdev_aux) { for (pvd = vd; pvd != rvd; pvd = pvd->vdev_parent) pvd->vdev_expanding = !!((flags & ZFS_ONLINE_EXPAND) || spa->spa_autoexpand); vd->vdev_expansion_time = gethrestime_sec(); } vdev_reopen(tvd); vd->vdev_checkremove = vd->vdev_forcefault = B_FALSE; if (!vd->vdev_aux) { for (pvd = vd; pvd != rvd; pvd = pvd->vdev_parent) pvd->vdev_expanding = B_FALSE; } if (newstate) *newstate = vd->vdev_state; if ((flags & ZFS_ONLINE_UNSPARE) && !vdev_is_dead(vd) && vd->vdev_parent && vd->vdev_parent->vdev_ops == &vdev_spare_ops && vd->vdev_parent->vdev_child[0] == vd) vd->vdev_unspare = B_TRUE; if ((flags & ZFS_ONLINE_EXPAND) || spa->spa_autoexpand) { /* XXX - L2ARC 1.0 does not support expansion */ if (vd->vdev_aux) return (spa_vdev_state_exit(spa, vd, ENOTSUP)); spa_async_request(spa, SPA_ASYNC_CONFIG_UPDATE); } /* Restart initializing if necessary */ mutex_enter(&vd->vdev_initialize_lock); if (vdev_writeable(vd) && vd->vdev_initialize_thread == NULL && vd->vdev_initialize_state == VDEV_INITIALIZE_ACTIVE) { (void) vdev_initialize(vd); } mutex_exit(&vd->vdev_initialize_lock); /* * Restart trimming if necessary. We do not restart trimming for cache * devices here. This is triggered by l2arc_rebuild_vdev() * asynchronously for the whole device or in l2arc_evict() as it evicts * space for upcoming writes. */ mutex_enter(&vd->vdev_trim_lock); if (vdev_writeable(vd) && !vd->vdev_isl2cache && vd->vdev_trim_thread == NULL && vd->vdev_trim_state == VDEV_TRIM_ACTIVE) { (void) vdev_trim(vd, vd->vdev_trim_rate, vd->vdev_trim_partial, vd->vdev_trim_secure); } mutex_exit(&vd->vdev_trim_lock); if (wasoffline || (oldstate < VDEV_STATE_DEGRADED && vd->vdev_state >= VDEV_STATE_DEGRADED)) { spa_event_notify(spa, vd, NULL, ESC_ZFS_VDEV_ONLINE); /* * Asynchronously detach spare vdev if resilver or * rebuild is not required */ if (vd->vdev_unspare && !dsl_scan_resilvering(spa->spa_dsl_pool) && !dsl_scan_resilver_scheduled(spa->spa_dsl_pool) && !vdev_rebuild_active(tvd)) spa_async_request(spa, SPA_ASYNC_DETACH_SPARE); } return (spa_vdev_state_exit(spa, vd, 0)); } static int vdev_offline_locked(spa_t *spa, uint64_t guid, uint64_t flags) { vdev_t *vd, *tvd; int error = 0; uint64_t generation; metaslab_group_t *mg; top: spa_vdev_state_enter(spa, SCL_ALLOC); if ((vd = spa_lookup_by_guid(spa, guid, B_TRUE)) == NULL) return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENODEV))); if (!vd->vdev_ops->vdev_op_leaf) return (spa_vdev_state_exit(spa, NULL, SET_ERROR(ENOTSUP))); if (vd->vdev_ops == &vdev_draid_spare_ops) return (spa_vdev_state_exit(spa, NULL, ENOTSUP)); tvd = vd->vdev_top; mg = tvd->vdev_mg; generation = spa->spa_config_generation + 1; /* * If the device isn't already offline, try to offline it. */ if (!vd->vdev_offline) { /* * If this device has the only valid copy of some data, * don't allow it to be offlined. Log devices are always * expendable. */ if (!tvd->vdev_islog && vd->vdev_aux == NULL && vdev_dtl_required(vd)) return (spa_vdev_state_exit(spa, NULL, SET_ERROR(EBUSY))); /* * If the top-level is a slog and it has had allocations * then proceed. We check that the vdev's metaslab group * is not NULL since it's possible that we may have just * added this vdev but not yet initialized its metaslabs. */ if (tvd->vdev_islog && mg != NULL) { /* * Prevent any future allocations. */ ASSERT3P(tvd->vdev_log_mg, ==, NULL); metaslab_group_passivate(mg); (void) spa_vdev_state_exit(spa, vd, 0); error = spa_reset_logs(spa); /* * If the log device was successfully reset but has * checkpointed data, do not offline it. */ if (error == 0 && tvd->vdev_checkpoint_sm != NULL) { ASSERT3U(space_map_allocated( tvd->vdev_checkpoint_sm), !=, 0); error = ZFS_ERR_CHECKPOINT_EXISTS; } spa_vdev_state_enter(spa, SCL_ALLOC); /* * Check to see if the config has changed. */ if (error || generation != spa->spa_config_generation) { metaslab_group_activate(mg); if (error) return (spa_vdev_state_exit(spa, vd, error)); (void) spa_vdev_state_exit(spa, vd, 0); goto top; } ASSERT0(tvd->vdev_stat.vs_alloc); } /* * Offline this device and reopen its top-level vdev. * If the top-level vdev is a log device then just offline * it. Otherwise, if this action results in the top-level * vdev becoming unusable, undo it and fail the request. */ vd->vdev_offline = B_TRUE; vdev_reopen(tvd); if (!tvd->vdev_islog && vd->vdev_aux == NULL && vdev_is_dead(tvd)) { vd->vdev_offline = B_FALSE; vdev_reopen(tvd); return (spa_vdev_state_exit(spa, NULL, SET_ERROR(EBUSY))); } /* * Add the device back into the metaslab rotor so that * once we online the device it's open for business. */ if (tvd->vdev_islog && mg != NULL) metaslab_group_activate(mg); } vd->vdev_tmpoffline = !!(flags & ZFS_OFFLINE_TEMPORARY); return (spa_vdev_state_exit(spa, vd, 0)); } int vdev_offline(spa_t *spa, uint64_t guid, uint64_t flags) { int error; mutex_enter(&spa->spa_vdev_top_lock); error = vdev_offline_locked(spa, guid, flags); mutex_exit(&spa->spa_vdev_top_lock); return (error); } /* * Clear the error counts associated with this vdev. Unlike vdev_online() and * vdev_offline(), we assume the spa config is locked. We also clear all * children. If 'vd' is NULL, then the user wants to clear all vdevs. */ void vdev_clear(spa_t *spa, vdev_t *vd) { vdev_t *rvd = spa->spa_root_vdev; ASSERT(spa_config_held(spa, SCL_STATE_ALL, RW_WRITER) == SCL_STATE_ALL); if (vd == NULL) vd = rvd; vd->vdev_stat.vs_read_errors = 0; vd->vdev_stat.vs_write_errors = 0; vd->vdev_stat.vs_checksum_errors = 0; vd->vdev_stat.vs_slow_ios = 0; for (int c = 0; c < vd->vdev_children; c++) vdev_clear(spa, vd->vdev_child[c]); /* * It makes no sense to "clear" an indirect or removed vdev. */ if (!vdev_is_concrete(vd) || vd->vdev_removed) return; /* * If we're in the FAULTED state or have experienced failed I/O, then * clear the persistent state and attempt to reopen the device. We * also mark the vdev config dirty, so that the new faulted state is * written out to disk. */ if (vd->vdev_faulted || vd->vdev_degraded || !vdev_readable(vd) || !vdev_writeable(vd)) { /* * When reopening in response to a clear event, it may be due to * a fmadm repair request. In this case, if the device is * still broken, we want to still post the ereport again. */ vd->vdev_forcefault = B_TRUE; vd->vdev_faulted = vd->vdev_degraded = 0ULL; vd->vdev_cant_read = B_FALSE; vd->vdev_cant_write = B_FALSE; vd->vdev_stat.vs_aux = 0; vdev_reopen(vd == rvd ? rvd : vd->vdev_top); vd->vdev_forcefault = B_FALSE; if (vd != rvd && vdev_writeable(vd->vdev_top)) vdev_state_dirty(vd->vdev_top); /* If a resilver isn't required, check if vdevs can be culled */ if (vd->vdev_aux == NULL && !vdev_is_dead(vd) && !dsl_scan_resilvering(spa->spa_dsl_pool) && !dsl_scan_resilver_scheduled(spa->spa_dsl_pool)) spa_async_request(spa, SPA_ASYNC_RESILVER_DONE); spa_event_notify(spa, vd, NULL, ESC_ZFS_VDEV_CLEAR); } /* * When clearing a FMA-diagnosed fault, we always want to * unspare the device, as we assume that the original spare was * done in response to the FMA fault. */ if (!vdev_is_dead(vd) && vd->vdev_parent != NULL && vd->vdev_parent->vdev_ops == &vdev_spare_ops && vd->vdev_parent->vdev_child[0] == vd) vd->vdev_unspare = B_TRUE; /* Clear recent error events cache (i.e. duplicate events tracking) */ zfs_ereport_clear(spa, vd); } boolean_t vdev_is_dead(vdev_t *vd) { /* * Holes and missing devices are always considered "dead". * This simplifies the code since we don't have to check for * these types of devices in the various code paths. * Instead we rely on the fact that we skip over dead devices * before issuing I/O to them. */ return (vd->vdev_state < VDEV_STATE_DEGRADED || vd->vdev_ops == &vdev_hole_ops || vd->vdev_ops == &vdev_missing_ops); } boolean_t vdev_readable(vdev_t *vd) { return (!vdev_is_dead(vd) && !vd->vdev_cant_read); } boolean_t vdev_writeable(vdev_t *vd) { return (!vdev_is_dead(vd) && !vd->vdev_cant_write && vdev_is_concrete(vd)); } boolean_t vdev_allocatable(vdev_t *vd) { uint64_t state = vd->vdev_state; /* * We currently allow allocations from vdevs which may be in the * process of reopening (i.e. VDEV_STATE_CLOSED). If the device * fails to reopen then we'll catch it later when we're holding * the proper locks. Note that we have to get the vdev state * in a local variable because although it changes atomically, * we're asking two separate questions about it. */ return (!(state < VDEV_STATE_DEGRADED && state != VDEV_STATE_CLOSED) && !vd->vdev_cant_write && vdev_is_concrete(vd) && vd->vdev_mg->mg_initialized); } boolean_t vdev_accessible(vdev_t *vd, zio_t *zio) { ASSERT(zio->io_vd == vd); if (vdev_is_dead(vd) || vd->vdev_remove_wanted) return (B_FALSE); if (zio->io_type == ZIO_TYPE_READ) return (!vd->vdev_cant_read); if (zio->io_type == ZIO_TYPE_WRITE) return (!vd->vdev_cant_write); return (B_TRUE); } static void vdev_get_child_stat(vdev_t *cvd, vdev_stat_t *vs, vdev_stat_t *cvs) { /* * Exclude the dRAID spare when aggregating to avoid double counting * the ops and bytes. These IOs are counted by the physical leaves. */ if (cvd->vdev_ops == &vdev_draid_spare_ops) return; for (int t = 0; t < VS_ZIO_TYPES; t++) { vs->vs_ops[t] += cvs->vs_ops[t]; vs->vs_bytes[t] += cvs->vs_bytes[t]; } cvs->vs_scan_removing = cvd->vdev_removing; } /* * Get extended stats */ static void vdev_get_child_stat_ex(vdev_t *cvd, vdev_stat_ex_t *vsx, vdev_stat_ex_t *cvsx) { (void) cvd; int t, b; for (t = 0; t < ZIO_TYPES; t++) { for (b = 0; b < ARRAY_SIZE(vsx->vsx_disk_histo[0]); b++) vsx->vsx_disk_histo[t][b] += cvsx->vsx_disk_histo[t][b]; for (b = 0; b < ARRAY_SIZE(vsx->vsx_total_histo[0]); b++) { vsx->vsx_total_histo[t][b] += cvsx->vsx_total_histo[t][b]; } } for (t = 0; t < ZIO_PRIORITY_NUM_QUEUEABLE; t++) { for (b = 0; b < ARRAY_SIZE(vsx->vsx_queue_histo[0]); b++) { vsx->vsx_queue_histo[t][b] += cvsx->vsx_queue_histo[t][b]; } vsx->vsx_active_queue[t] += cvsx->vsx_active_queue[t]; vsx->vsx_pend_queue[t] += cvsx->vsx_pend_queue[t]; for (b = 0; b < ARRAY_SIZE(vsx->vsx_ind_histo[0]); b++) vsx->vsx_ind_histo[t][b] += cvsx->vsx_ind_histo[t][b]; for (b = 0; b < ARRAY_SIZE(vsx->vsx_agg_histo[0]); b++) vsx->vsx_agg_histo[t][b] += cvsx->vsx_agg_histo[t][b]; } } boolean_t vdev_is_spacemap_addressable(vdev_t *vd) { if (spa_feature_is_active(vd->vdev_spa, SPA_FEATURE_SPACEMAP_V2)) return (B_TRUE); /* * If double-word space map entries are not enabled we assume * 47 bits of the space map entry are dedicated to the entry's * offset (see SM_OFFSET_BITS in space_map.h). We then use that * to calculate the maximum address that can be described by a * space map entry for the given device. */ uint64_t shift = vd->vdev_ashift + SM_OFFSET_BITS; if (shift >= 63) /* detect potential overflow */ return (B_TRUE); return (vd->vdev_asize < (1ULL << shift)); } /* * Get statistics for the given vdev. */ static void vdev_get_stats_ex_impl(vdev_t *vd, vdev_stat_t *vs, vdev_stat_ex_t *vsx) { int t; /* * If we're getting stats on the root vdev, aggregate the I/O counts * over all top-level vdevs (i.e. the direct children of the root). */ if (!vd->vdev_ops->vdev_op_leaf) { if (vs) { memset(vs->vs_ops, 0, sizeof (vs->vs_ops)); memset(vs->vs_bytes, 0, sizeof (vs->vs_bytes)); } if (vsx) memset(vsx, 0, sizeof (*vsx)); for (int c = 0; c < vd->vdev_children; c++) { vdev_t *cvd = vd->vdev_child[c]; vdev_stat_t *cvs = &cvd->vdev_stat; vdev_stat_ex_t *cvsx = &cvd->vdev_stat_ex; vdev_get_stats_ex_impl(cvd, cvs, cvsx); if (vs) vdev_get_child_stat(cvd, vs, cvs); if (vsx) vdev_get_child_stat_ex(cvd, vsx, cvsx); } } else { /* * We're a leaf. Just copy our ZIO active queue stats in. The * other leaf stats are updated in vdev_stat_update(). */ if (!vsx) return; memcpy(vsx, &vd->vdev_stat_ex, sizeof (vd->vdev_stat_ex)); for (t = 0; t < ARRAY_SIZE(vd->vdev_queue.vq_class); t++) { vsx->vsx_active_queue[t] = vd->vdev_queue.vq_class[t].vqc_active; vsx->vsx_pend_queue[t] = avl_numnodes( &vd->vdev_queue.vq_class[t].vqc_queued_tree); } } } void vdev_get_stats_ex(vdev_t *vd, vdev_stat_t *vs, vdev_stat_ex_t *vsx) { vdev_t *tvd = vd->vdev_top; mutex_enter(&vd->vdev_stat_lock); if (vs) { memcpy(vs, &vd->vdev_stat, sizeof (*vs)); vs->vs_timestamp = gethrtime() - vs->vs_timestamp; vs->vs_state = vd->vdev_state; vs->vs_rsize = vdev_get_min_asize(vd); if (vd->vdev_ops->vdev_op_leaf) { vs->vs_pspace = vd->vdev_psize; vs->vs_rsize += VDEV_LABEL_START_SIZE + VDEV_LABEL_END_SIZE; /* * Report initializing progress. Since we don't * have the initializing locks held, this is only * an estimate (although a fairly accurate one). */ vs->vs_initialize_bytes_done = vd->vdev_initialize_bytes_done; vs->vs_initialize_bytes_est = vd->vdev_initialize_bytes_est; vs->vs_initialize_state = vd->vdev_initialize_state; vs->vs_initialize_action_time = vd->vdev_initialize_action_time; /* * Report manual TRIM progress. Since we don't have * the manual TRIM locks held, this is only an * estimate (although fairly accurate one). */ vs->vs_trim_notsup = !vd->vdev_has_trim; vs->vs_trim_bytes_done = vd->vdev_trim_bytes_done; vs->vs_trim_bytes_est = vd->vdev_trim_bytes_est; vs->vs_trim_state = vd->vdev_trim_state; vs->vs_trim_action_time = vd->vdev_trim_action_time; /* Set when there is a deferred resilver. */ vs->vs_resilver_deferred = vd->vdev_resilver_deferred; } /* * Report expandable space on top-level, non-auxiliary devices * only. The expandable space is reported in terms of metaslab * sized units since that determines how much space the pool * can expand. */ if (vd->vdev_aux == NULL && tvd != NULL) { vs->vs_esize = P2ALIGN( vd->vdev_max_asize - vd->vdev_asize, 1ULL << tvd->vdev_ms_shift); } vs->vs_configured_ashift = vd->vdev_top != NULL ? vd->vdev_top->vdev_ashift : vd->vdev_ashift; vs->vs_logical_ashift = vd->vdev_logical_ashift; if (vd->vdev_physical_ashift <= ASHIFT_MAX) vs->vs_physical_ashift = vd->vdev_physical_ashift; else vs->vs_physical_ashift = 0; /* * Report fragmentation and rebuild progress for top-level, * non-auxiliary, concrete devices. */ if (vd->vdev_aux == NULL && vd == vd->vdev_top && vdev_is_concrete(vd)) { /* * The vdev fragmentation rating doesn't take into * account the embedded slog metaslab (vdev_log_mg). * Since it's only one metaslab, it would have a tiny * impact on the overall fragmentation. */ vs->vs_fragmentation = (vd->vdev_mg != NULL) ? vd->vdev_mg->mg_fragmentation : 0; } vs->vs_noalloc = MAX(vd->vdev_noalloc, tvd ? tvd->vdev_noalloc : 0); } vdev_get_stats_ex_impl(vd, vs, vsx); mutex_exit(&vd->vdev_stat_lock); } void vdev_get_stats(vdev_t *vd, vdev_stat_t *vs) { return (vdev_get_stats_ex(vd, vs, NULL)); } void vdev_clear_stats(vdev_t *vd) { mutex_enter(&vd->vdev_stat_lock); vd->vdev_stat.vs_space = 0; vd->vdev_stat.vs_dspace = 0; vd->vdev_stat.vs_alloc = 0; mutex_exit(&vd->vdev_stat_lock); } void vdev_scan_stat_init(vdev_t *vd) { vdev_stat_t *vs = &vd->vdev_stat; for (int c = 0; c < vd->vdev_children; c++) vdev_scan_stat_init(vd->vdev_child[c]); mutex_enter(&vd->vdev_stat_lock); vs->vs_scan_processed = 0; mutex_exit(&vd->vdev_stat_lock); } void vdev_stat_update(zio_t *zio, uint64_t psize) { spa_t *spa = zio->io_spa; vdev_t *rvd = spa->spa_root_vdev; vdev_t *vd = zio->io_vd ? zio->io_vd : rvd; vdev_t *pvd; uint64_t txg = zio->io_txg; /* Suppress ASAN false positive */ #ifdef __SANITIZE_ADDRESS__ vdev_stat_t *vs = vd ? &vd->vdev_stat : NULL; vdev_stat_ex_t *vsx = vd ? &vd->vdev_stat_ex : NULL; #else vdev_stat_t *vs = &vd->vdev_stat; vdev_stat_ex_t *vsx = &vd->vdev_stat_ex; #endif zio_type_t type = zio->io_type; int flags = zio->io_flags; /* * If this i/o is a gang leader, it didn't do any actual work. */ if (zio->io_gang_tree) return; if (zio->io_error == 0) { /* * If this is a root i/o, don't count it -- we've already * counted the top-level vdevs, and vdev_get_stats() will * aggregate them when asked. This reduces contention on * the root vdev_stat_lock and implicitly handles blocks * that compress away to holes, for which there is no i/o. * (Holes never create vdev children, so all the counters * remain zero, which is what we want.) * * Note: this only applies to successful i/o (io_error == 0) * because unlike i/o counts, errors are not additive. * When reading a ditto block, for example, failure of * one top-level vdev does not imply a root-level error. */ if (vd == rvd) return; ASSERT(vd == zio->io_vd); if (flags & ZIO_FLAG_IO_BYPASS) return; mutex_enter(&vd->vdev_stat_lock); if (flags & ZIO_FLAG_IO_REPAIR) { /* * Repair is the result of a resilver issued by the * scan thread (spa_sync). */ if (flags & ZIO_FLAG_SCAN_THREAD) { dsl_scan_t *scn = spa->spa_dsl_pool->dp_scan; dsl_scan_phys_t *scn_phys = &scn->scn_phys; uint64_t *processed = &scn_phys->scn_processed; if (vd->vdev_ops->vdev_op_leaf) atomic_add_64(processed, psize); vs->vs_scan_processed += psize; } /* * Repair is the result of a rebuild issued by the * rebuild thread (vdev_rebuild_thread). To avoid * double counting repaired bytes the virtual dRAID * spare vdev is excluded from the processed bytes. */ if (zio->io_priority == ZIO_PRIORITY_REBUILD) { vdev_t *tvd = vd->vdev_top; vdev_rebuild_t *vr = &tvd->vdev_rebuild_config; vdev_rebuild_phys_t *vrp = &vr->vr_rebuild_phys; uint64_t *rebuilt = &vrp->vrp_bytes_rebuilt; if (vd->vdev_ops->vdev_op_leaf && vd->vdev_ops != &vdev_draid_spare_ops) { atomic_add_64(rebuilt, psize); } vs->vs_rebuild_processed += psize; } if (flags & ZIO_FLAG_SELF_HEAL) vs->vs_self_healed += psize; } /* * The bytes/ops/histograms are recorded at the leaf level and * aggregated into the higher level vdevs in vdev_get_stats(). */ if (vd->vdev_ops->vdev_op_leaf && (zio->io_priority < ZIO_PRIORITY_NUM_QUEUEABLE)) { zio_type_t vs_type = type; zio_priority_t priority = zio->io_priority; /* * TRIM ops and bytes are reported to user space as * ZIO_TYPE_IOCTL. This is done to preserve the * vdev_stat_t structure layout for user space. */ if (type == ZIO_TYPE_TRIM) vs_type = ZIO_TYPE_IOCTL; /* * Solely for the purposes of 'zpool iostat -lqrw' * reporting use the priority to categorize the IO. * Only the following are reported to user space: * * ZIO_PRIORITY_SYNC_READ, * ZIO_PRIORITY_SYNC_WRITE, * ZIO_PRIORITY_ASYNC_READ, * ZIO_PRIORITY_ASYNC_WRITE, * ZIO_PRIORITY_SCRUB, * ZIO_PRIORITY_TRIM, * ZIO_PRIORITY_REBUILD. */ if (priority == ZIO_PRIORITY_INITIALIZING) { ASSERT3U(type, ==, ZIO_TYPE_WRITE); priority = ZIO_PRIORITY_ASYNC_WRITE; } else if (priority == ZIO_PRIORITY_REMOVAL) { priority = ((type == ZIO_TYPE_WRITE) ? ZIO_PRIORITY_ASYNC_WRITE : ZIO_PRIORITY_ASYNC_READ); } vs->vs_ops[vs_type]++; vs->vs_bytes[vs_type] += psize; if (flags & ZIO_FLAG_DELEGATED) { vsx->vsx_agg_histo[priority] [RQ_HISTO(zio->io_size)]++; } else { vsx->vsx_ind_histo[priority] [RQ_HISTO(zio->io_size)]++; } if (zio->io_delta && zio->io_delay) { vsx->vsx_queue_histo[priority] [L_HISTO(zio->io_delta - zio->io_delay)]++; vsx->vsx_disk_histo[type] [L_HISTO(zio->io_delay)]++; vsx->vsx_total_histo[type] [L_HISTO(zio->io_delta)]++; } } mutex_exit(&vd->vdev_stat_lock); return; } if (flags & ZIO_FLAG_SPECULATIVE) return; /* * If this is an I/O error that is going to be retried, then ignore the * error. Otherwise, the user may interpret B_FAILFAST I/O errors as * hard errors, when in reality they can happen for any number of * innocuous reasons (bus resets, MPxIO link failure, etc). */ if (zio->io_error == EIO && !(zio->io_flags & ZIO_FLAG_IO_RETRY)) return; /* * Intent logs writes won't propagate their error to the root * I/O so don't mark these types of failures as pool-level * errors. */ if (zio->io_vd == NULL && (zio->io_flags & ZIO_FLAG_DONT_PROPAGATE)) return; if (type == ZIO_TYPE_WRITE && txg != 0 && (!(flags & ZIO_FLAG_IO_REPAIR) || (flags & ZIO_FLAG_SCAN_THREAD) || spa->spa_claiming)) { /* * This is either a normal write (not a repair), or it's * a repair induced by the scrub thread, or it's a repair * made by zil_claim() during spa_load() in the first txg. * In the normal case, we commit the DTL change in the same * txg as the block was born. In the scrub-induced repair * case, we know that scrubs run in first-pass syncing context, * so we commit the DTL change in spa_syncing_txg(spa). * In the zil_claim() case, we commit in spa_first_txg(spa). * * We currently do not make DTL entries for failed spontaneous * self-healing writes triggered by normal (non-scrubbing) * reads, because we have no transactional context in which to * do so -- and it's not clear that it'd be desirable anyway. */ if (vd->vdev_ops->vdev_op_leaf) { uint64_t commit_txg = txg; if (flags & ZIO_FLAG_SCAN_THREAD) { ASSERT(flags & ZIO_FLAG_IO_REPAIR); ASSERT(spa_sync_pass(spa) == 1); vdev_dtl_dirty(vd, DTL_SCRUB, txg, 1); commit_txg = spa_syncing_txg(spa); } else if (spa->spa_claiming) { ASSERT(flags & ZIO_FLAG_IO_REPAIR); commit_txg = spa_first_txg(spa); } ASSERT(commit_txg >= spa_syncing_txg(spa)); if (vdev_dtl_contains(vd, DTL_MISSING, txg, 1)) return; for (pvd = vd; pvd != rvd; pvd = pvd->vdev_parent) vdev_dtl_dirty(pvd, DTL_PARTIAL, txg, 1); vdev_dirty(vd->vdev_top, VDD_DTL, vd, commit_txg); } if (vd != rvd) vdev_dtl_dirty(vd, DTL_MISSING, txg, 1); } } int64_t vdev_deflated_space(vdev_t *vd, int64_t space) { ASSERT((space & (SPA_MINBLOCKSIZE-1)) == 0); ASSERT(vd->vdev_deflate_ratio != 0 || vd->vdev_isl2cache); return ((space >> SPA_MINBLOCKSHIFT) * vd->vdev_deflate_ratio); } /* * Update the in-core space usage stats for this vdev, its metaslab class, * and the root vdev. */ void vdev_space_update(vdev_t *vd, int64_t alloc_delta, int64_t defer_delta, int64_t space_delta) { (void) defer_delta; int64_t dspace_delta; spa_t *spa = vd->vdev_spa; vdev_t *rvd = spa->spa_root_vdev; ASSERT(vd == vd->vdev_top); /* * Apply the inverse of the psize-to-asize (ie. RAID-Z) space-expansion * factor. We must calculate this here and not at the root vdev * because the root vdev's psize-to-asize is simply the max of its * children's, thus not accurate enough for us. */ dspace_delta = vdev_deflated_space(vd, space_delta); mutex_enter(&vd->vdev_stat_lock); /* ensure we won't underflow */ if (alloc_delta < 0) { ASSERT3U(vd->vdev_stat.vs_alloc, >=, -alloc_delta); } vd->vdev_stat.vs_alloc += alloc_delta; vd->vdev_stat.vs_space += space_delta; vd->vdev_stat.vs_dspace += dspace_delta; mutex_exit(&vd->vdev_stat_lock); /* every class but log contributes to root space stats */ if (vd->vdev_mg != NULL && !vd->vdev_islog) { ASSERT(!vd->vdev_isl2cache); mutex_enter(&rvd->vdev_stat_lock); rvd->vdev_stat.vs_alloc += alloc_delta; rvd->vdev_stat.vs_space += space_delta; rvd->vdev_stat.vs_dspace += dspace_delta; mutex_exit(&rvd->vdev_stat_lock); } /* Note: metaslab_class_space_update moved to metaslab_space_update */ } /* * Mark a top-level vdev's config as dirty, placing it on the dirty list * so that it will be written out next time the vdev configuration is synced. * If the root vdev is specified (vdev_top == NULL), dirty all top-level vdevs. */ void vdev_config_dirty(vdev_t *vd) { spa_t *spa = vd->vdev_spa; vdev_t *rvd = spa->spa_root_vdev; int c; ASSERT(spa_writeable(spa)); /* * If this is an aux vdev (as with l2cache and spare devices), then we * update the vdev config manually and set the sync flag. */ if (vd->vdev_aux != NULL) { spa_aux_vdev_t *sav = vd->vdev_aux; nvlist_t **aux; uint_t naux; for (c = 0; c < sav->sav_count; c++) { if (sav->sav_vdevs[c] == vd) break; } if (c == sav->sav_count) { /* * We're being removed. There's nothing more to do. */ ASSERT(sav->sav_sync == B_TRUE); return; } sav->sav_sync = B_TRUE; if (nvlist_lookup_nvlist_array(sav->sav_config, ZPOOL_CONFIG_L2CACHE, &aux, &naux) != 0) { VERIFY(nvlist_lookup_nvlist_array(sav->sav_config, ZPOOL_CONFIG_SPARES, &aux, &naux) == 0); } ASSERT(c < naux); /* * Setting the nvlist in the middle if the array is a little * sketchy, but it will work. */ nvlist_free(aux[c]); aux[c] = vdev_config_generate(spa, vd, B_TRUE, 0); return; } /* * The dirty list is protected by the SCL_CONFIG lock. The caller * must either hold SCL_CONFIG as writer, or must be the sync thread * (which holds SCL_CONFIG as reader). There's only one sync thread, * so this is sufficient to ensure mutual exclusion. */ ASSERT(spa_config_held(spa, SCL_CONFIG, RW_WRITER) || (dsl_pool_sync_context(spa_get_dsl(spa)) && spa_config_held(spa, SCL_CONFIG, RW_READER))); if (vd == rvd) { for (c = 0; c < rvd->vdev_children; c++) vdev_config_dirty(rvd->vdev_child[c]); } else { ASSERT(vd == vd->vdev_top); if (!list_link_active(&vd->vdev_config_dirty_node) && vdev_is_concrete(vd)) { list_insert_head(&spa->spa_config_dirty_list, vd); } } } void vdev_config_clean(vdev_t *vd) { spa_t *spa = vd->vdev_spa; ASSERT(spa_config_held(spa, SCL_CONFIG, RW_WRITER) || (dsl_pool_sync_context(spa_get_dsl(spa)) && spa_config_held(spa, SCL_CONFIG, RW_READER))); ASSERT(list_link_active(&vd->vdev_config_dirty_node)); list_remove(&spa->spa_config_dirty_list, vd); } /* * Mark a top-level vdev's state as dirty, so that the next pass of * spa_sync() can convert this into vdev_config_dirty(). We distinguish * the state changes from larger config changes because they require * much less locking, and are often needed for administrative actions. */ void vdev_state_dirty(vdev_t *vd) { spa_t *spa = vd->vdev_spa; ASSERT(spa_writeable(spa)); ASSERT(vd == vd->vdev_top); /* * The state list is protected by the SCL_STATE lock. The caller * must either hold SCL_STATE as writer, or must be the sync thread * (which holds SCL_STATE as reader). There's only one sync thread, * so this is sufficient to ensure mutual exclusion. */ ASSERT(spa_config_held(spa, SCL_STATE, RW_WRITER) || (dsl_pool_sync_context(spa_get_dsl(spa)) && spa_config_held(spa, SCL_STATE, RW_READER))); if (!list_link_active(&vd->vdev_state_dirty_node) && vdev_is_concrete(vd)) list_insert_head(&spa->spa_state_dirty_list, vd); } void vdev_state_clean(vdev_t *vd) { spa_t *spa = vd->vdev_spa; ASSERT(spa_config_held(spa, SCL_STATE, RW_WRITER) || (dsl_pool_sync_context(spa_get_dsl(spa)) && spa_config_held(spa, SCL_STATE, RW_READER))); ASSERT(list_link_active(&vd->vdev_state_dirty_node)); list_remove(&spa->spa_state_dirty_list, vd); } /* * Propagate vdev state up from children to parent. */ void vdev_propagate_state(vdev_t *vd) { spa_t *spa = vd->vdev_spa; vdev_t *rvd = spa->spa_root_vdev; int degraded = 0, faulted = 0; int corrupted = 0; vdev_t *child; if (vd->vdev_children > 0) { for (int c = 0; c < vd->vdev_children; c++) { child = vd->vdev_child[c]; /* * Don't factor holes or indirect vdevs into the * decision. */ if (!vdev_is_concrete(child)) continue; if (!vdev_readable(child) || (!vdev_writeable(child) && spa_writeable(spa))) { /* * Root special: if there is a top-level log * device, treat the root vdev as if it were * degraded. */ if (child->vdev_islog && vd == rvd) degraded++; else faulted++; } else if (child->vdev_state <= VDEV_STATE_DEGRADED) { degraded++; } if (child->vdev_stat.vs_aux == VDEV_AUX_CORRUPT_DATA) corrupted++; } vd->vdev_ops->vdev_op_state_change(vd, faulted, degraded); /* * Root special: if there is a top-level vdev that cannot be * opened due to corrupted metadata, then propagate the root * vdev's aux state as 'corrupt' rather than 'insufficient * replicas'. */ if (corrupted && vd == rvd && rvd->vdev_state == VDEV_STATE_CANT_OPEN) vdev_set_state(rvd, B_FALSE, VDEV_STATE_CANT_OPEN, VDEV_AUX_CORRUPT_DATA); } if (vd->vdev_parent) vdev_propagate_state(vd->vdev_parent); } /* * Set a vdev's state. If this is during an open, we don't update the parent * state, because we're in the process of opening children depth-first. * Otherwise, we propagate the change to the parent. * * If this routine places a device in a faulted state, an appropriate ereport is * generated. */ void vdev_set_state(vdev_t *vd, boolean_t isopen, vdev_state_t state, vdev_aux_t aux) { uint64_t save_state; spa_t *spa = vd->vdev_spa; if (state == vd->vdev_state) { /* * Since vdev_offline() code path is already in an offline * state we can miss a statechange event to OFFLINE. Check * the previous state to catch this condition. */ if (vd->vdev_ops->vdev_op_leaf && (state == VDEV_STATE_OFFLINE) && (vd->vdev_prevstate >= VDEV_STATE_FAULTED)) { /* post an offline state change */ zfs_post_state_change(spa, vd, vd->vdev_prevstate); } vd->vdev_stat.vs_aux = aux; return; } save_state = vd->vdev_state; vd->vdev_state = state; vd->vdev_stat.vs_aux = aux; /* * If we are setting the vdev state to anything but an open state, then * always close the underlying device unless the device has requested * a delayed close (i.e. we're about to remove or fault the device). * Otherwise, we keep accessible but invalid devices open forever. * We don't call vdev_close() itself, because that implies some extra * checks (offline, etc) that we don't want here. This is limited to * leaf devices, because otherwise closing the device will affect other * children. */ if (!vd->vdev_delayed_close && vdev_is_dead(vd) && vd->vdev_ops->vdev_op_leaf) vd->vdev_ops->vdev_op_close(vd); if (vd->vdev_removed && state == VDEV_STATE_CANT_OPEN && (aux == VDEV_AUX_OPEN_FAILED || vd->vdev_checkremove)) { /* * If the previous state is set to VDEV_STATE_REMOVED, then this * device was previously marked removed and someone attempted to * reopen it. If this failed due to a nonexistent device, then * keep the device in the REMOVED state. We also let this be if * it is one of our special test online cases, which is only * attempting to online the device and shouldn't generate an FMA * fault. */ vd->vdev_state = VDEV_STATE_REMOVED; vd->vdev_stat.vs_aux = VDEV_AUX_NONE; } else if (state == VDEV_STATE_REMOVED) { vd->vdev_removed = B_TRUE; } else if (state == VDEV_STATE_CANT_OPEN) { /* * If we fail to open a vdev during an import or recovery, we * mark it as "not available", which signifies that it was * never there to begin with. Failure to open such a device * is not considered an error. */ if ((spa_load_state(spa) == SPA_LOAD_IMPORT || spa_load_state(spa) == SPA_LOAD_RECOVER) && vd->vdev_ops->vdev_op_leaf) vd->vdev_not_present = 1; /* * Post the appropriate ereport. If the 'prevstate' field is * set to something other than VDEV_STATE_UNKNOWN, it indicates * that this is part of a vdev_reopen(). In this case, we don't * want to post the ereport if the device was already in the * CANT_OPEN state beforehand. * * If the 'checkremove' flag is set, then this is an attempt to * online the device in response to an insertion event. If we * hit this case, then we have detected an insertion event for a * faulted or offline device that wasn't in the removed state. * In this scenario, we don't post an ereport because we are * about to replace the device, or attempt an online with * vdev_forcefault, which will generate the fault for us. */ if ((vd->vdev_prevstate != state || vd->vdev_forcefault) && !vd->vdev_not_present && !vd->vdev_checkremove && vd != spa->spa_root_vdev) { const char *class; switch (aux) { case VDEV_AUX_OPEN_FAILED: class = FM_EREPORT_ZFS_DEVICE_OPEN_FAILED; break; case VDEV_AUX_CORRUPT_DATA: class = FM_EREPORT_ZFS_DEVICE_CORRUPT_DATA; break; case VDEV_AUX_NO_REPLICAS: class = FM_EREPORT_ZFS_DEVICE_NO_REPLICAS; break; case VDEV_AUX_BAD_GUID_SUM: class = FM_EREPORT_ZFS_DEVICE_BAD_GUID_SUM; break; case VDEV_AUX_TOO_SMALL: class = FM_EREPORT_ZFS_DEVICE_TOO_SMALL; break; case VDEV_AUX_BAD_LABEL: class = FM_EREPORT_ZFS_DEVICE_BAD_LABEL; break; case VDEV_AUX_BAD_ASHIFT: class = FM_EREPORT_ZFS_DEVICE_BAD_ASHIFT; break; default: class = FM_EREPORT_ZFS_DEVICE_UNKNOWN; } (void) zfs_ereport_post(class, spa, vd, NULL, NULL, save_state); } /* Erase any notion of persistent removed state */ vd->vdev_removed = B_FALSE; } else { vd->vdev_removed = B_FALSE; } /* * Notify ZED of any significant state-change on a leaf vdev. * */ if (vd->vdev_ops->vdev_op_leaf) { /* preserve original state from a vdev_reopen() */ if ((vd->vdev_prevstate != VDEV_STATE_UNKNOWN) && (vd->vdev_prevstate != vd->vdev_state) && (save_state <= VDEV_STATE_CLOSED)) save_state = vd->vdev_prevstate; /* filter out state change due to initial vdev_open */ if (save_state > VDEV_STATE_CLOSED) zfs_post_state_change(spa, vd, save_state); } if (!isopen && vd->vdev_parent) vdev_propagate_state(vd->vdev_parent); } boolean_t vdev_children_are_offline(vdev_t *vd) { ASSERT(!vd->vdev_ops->vdev_op_leaf); for (uint64_t i = 0; i < vd->vdev_children; i++) { if (vd->vdev_child[i]->vdev_state != VDEV_STATE_OFFLINE) return (B_FALSE); } return (B_TRUE); } /* * Check the vdev configuration to ensure that it's capable of supporting * a root pool. We do not support partial configuration. */ boolean_t vdev_is_bootable(vdev_t *vd) { if (!vd->vdev_ops->vdev_op_leaf) { const char *vdev_type = vd->vdev_ops->vdev_op_type; if (strcmp(vdev_type, VDEV_TYPE_MISSING) == 0) return (B_FALSE); } for (int c = 0; c < vd->vdev_children; c++) { if (!vdev_is_bootable(vd->vdev_child[c])) return (B_FALSE); } return (B_TRUE); } boolean_t vdev_is_concrete(vdev_t *vd) { vdev_ops_t *ops = vd->vdev_ops; if (ops == &vdev_indirect_ops || ops == &vdev_hole_ops || ops == &vdev_missing_ops || ops == &vdev_root_ops) { return (B_FALSE); } else { return (B_TRUE); } } /* * Determine if a log device has valid content. If the vdev was * removed or faulted in the MOS config then we know that * the content on the log device has already been written to the pool. */ boolean_t vdev_log_state_valid(vdev_t *vd) { if (vd->vdev_ops->vdev_op_leaf && !vd->vdev_faulted && !vd->vdev_removed) return (B_TRUE); for (int c = 0; c < vd->vdev_children; c++) if (vdev_log_state_valid(vd->vdev_child[c])) return (B_TRUE); return (B_FALSE); } /* * Expand a vdev if possible. */ void vdev_expand(vdev_t *vd, uint64_t txg) { ASSERT(vd->vdev_top == vd); ASSERT(spa_config_held(vd->vdev_spa, SCL_ALL, RW_WRITER) == SCL_ALL); ASSERT(vdev_is_concrete(vd)); vdev_set_deflate_ratio(vd); if ((vd->vdev_asize >> vd->vdev_ms_shift) > vd->vdev_ms_count && vdev_is_concrete(vd)) { vdev_metaslab_group_create(vd); VERIFY(vdev_metaslab_init(vd, txg) == 0); vdev_config_dirty(vd); } } /* * Split a vdev. */ void vdev_split(vdev_t *vd) { vdev_t *cvd, *pvd = vd->vdev_parent; VERIFY3U(pvd->vdev_children, >, 1); vdev_remove_child(pvd, vd); vdev_compact_children(pvd); ASSERT3P(pvd->vdev_child, !=, NULL); cvd = pvd->vdev_child[0]; if (pvd->vdev_children == 1) { vdev_remove_parent(cvd); cvd->vdev_splitting = B_TRUE; } vdev_propagate_state(cvd); } void vdev_deadman(vdev_t *vd, const char *tag) { for (int c = 0; c < vd->vdev_children; c++) { vdev_t *cvd = vd->vdev_child[c]; vdev_deadman(cvd, tag); } if (vd->vdev_ops->vdev_op_leaf) { vdev_queue_t *vq = &vd->vdev_queue; mutex_enter(&vq->vq_lock); if (avl_numnodes(&vq->vq_active_tree) > 0) { spa_t *spa = vd->vdev_spa; zio_t *fio; uint64_t delta; zfs_dbgmsg("slow vdev: %s has %lu active IOs", vd->vdev_path, avl_numnodes(&vq->vq_active_tree)); /* * Look at the head of all the pending queues, * if any I/O has been outstanding for longer than * the spa_deadman_synctime invoke the deadman logic. */ fio = avl_first(&vq->vq_active_tree); delta = gethrtime() - fio->io_timestamp; if (delta > spa_deadman_synctime(spa)) zio_deadman(fio, tag); } mutex_exit(&vq->vq_lock); } } void vdev_defer_resilver(vdev_t *vd) { ASSERT(vd->vdev_ops->vdev_op_leaf); vd->vdev_resilver_deferred = B_TRUE; vd->vdev_spa->spa_resilver_deferred = B_TRUE; } /* * Clears the resilver deferred flag on all leaf devs under vd. Returns * B_TRUE if we have devices that need to be resilvered and are available to * accept resilver I/Os. */ boolean_t vdev_clear_resilver_deferred(vdev_t *vd, dmu_tx_t *tx) { boolean_t resilver_needed = B_FALSE; spa_t *spa = vd->vdev_spa; for (int c = 0; c < vd->vdev_children; c++) { vdev_t *cvd = vd->vdev_child[c]; resilver_needed |= vdev_clear_resilver_deferred(cvd, tx); } if (vd == spa->spa_root_vdev && spa_feature_is_active(spa, SPA_FEATURE_RESILVER_DEFER)) { spa_feature_decr(spa, SPA_FEATURE_RESILVER_DEFER, tx); vdev_config_dirty(vd); spa->spa_resilver_deferred = B_FALSE; return (resilver_needed); } if (!vdev_is_concrete(vd) || vd->vdev_aux || !vd->vdev_ops->vdev_op_leaf) return (resilver_needed); vd->vdev_resilver_deferred = B_FALSE; return (!vdev_is_dead(vd) && !vd->vdev_offline && vdev_resilver_needed(vd, NULL, NULL)); } boolean_t vdev_xlate_is_empty(range_seg64_t *rs) { return (rs->rs_start == rs->rs_end); } /* * Translate a logical range to the first contiguous physical range for the * specified vdev_t. This function is initially called with a leaf vdev and * will walk each parent vdev until it reaches a top-level vdev. Once the * top-level is reached the physical range is initialized and the recursive * function begins to unwind. As it unwinds it calls the parent's vdev * specific translation function to do the real conversion. */ void vdev_xlate(vdev_t *vd, const range_seg64_t *logical_rs, range_seg64_t *physical_rs, range_seg64_t *remain_rs) { /* * Walk up the vdev tree */ if (vd != vd->vdev_top) { vdev_xlate(vd->vdev_parent, logical_rs, physical_rs, remain_rs); } else { /* * We've reached the top-level vdev, initialize the physical * range to the logical range and set an empty remaining * range then start to unwind. */ physical_rs->rs_start = logical_rs->rs_start; physical_rs->rs_end = logical_rs->rs_end; remain_rs->rs_start = logical_rs->rs_start; remain_rs->rs_end = logical_rs->rs_start; return; } vdev_t *pvd = vd->vdev_parent; ASSERT3P(pvd, !=, NULL); ASSERT3P(pvd->vdev_ops->vdev_op_xlate, !=, NULL); /* * As this recursive function unwinds, translate the logical * range into its physical and any remaining components by calling * the vdev specific translate function. */ range_seg64_t intermediate = { 0 }; pvd->vdev_ops->vdev_op_xlate(vd, physical_rs, &intermediate, remain_rs); physical_rs->rs_start = intermediate.rs_start; physical_rs->rs_end = intermediate.rs_end; } void vdev_xlate_walk(vdev_t *vd, const range_seg64_t *logical_rs, vdev_xlate_func_t *func, void *arg) { range_seg64_t iter_rs = *logical_rs; range_seg64_t physical_rs; range_seg64_t remain_rs; while (!vdev_xlate_is_empty(&iter_rs)) { vdev_xlate(vd, &iter_rs, &physical_rs, &remain_rs); /* * With raidz and dRAID, it's possible that the logical range * does not live on this leaf vdev. Only when there is a non- * zero physical size call the provided function. */ if (!vdev_xlate_is_empty(&physical_rs)) func(arg, &physical_rs); iter_rs = remain_rs; } } static char * vdev_name(vdev_t *vd, char *buf, int buflen) { if (vd->vdev_path == NULL) { if (strcmp(vd->vdev_ops->vdev_op_type, "root") == 0) { strlcpy(buf, vd->vdev_spa->spa_name, buflen); } else if (!vd->vdev_ops->vdev_op_leaf) { snprintf(buf, buflen, "%s-%llu", vd->vdev_ops->vdev_op_type, (u_longlong_t)vd->vdev_id); } } else { strlcpy(buf, vd->vdev_path, buflen); } return (buf); } /* * Look at the vdev tree and determine whether any devices are currently being * replaced. */ boolean_t vdev_replace_in_progress(vdev_t *vdev) { ASSERT(spa_config_held(vdev->vdev_spa, SCL_ALL, RW_READER) != 0); if (vdev->vdev_ops == &vdev_replacing_ops) return (B_TRUE); /* * A 'spare' vdev indicates that we have a replace in progress, unless * it has exactly two children, and the second, the hot spare, has * finished being resilvered. */ if (vdev->vdev_ops == &vdev_spare_ops && (vdev->vdev_children > 2 || !vdev_dtl_empty(vdev->vdev_child[1], DTL_MISSING))) return (B_TRUE); for (int i = 0; i < vdev->vdev_children; i++) { if (vdev_replace_in_progress(vdev->vdev_child[i])) return (B_TRUE); } return (B_FALSE); } /* * Add a (source=src, propname=propval) list to an nvlist. */ static void vdev_prop_add_list(nvlist_t *nvl, const char *propname, const char *strval, uint64_t intval, zprop_source_t src) { nvlist_t *propval; propval = fnvlist_alloc(); fnvlist_add_uint64(propval, ZPROP_SOURCE, src); if (strval != NULL) fnvlist_add_string(propval, ZPROP_VALUE, strval); else fnvlist_add_uint64(propval, ZPROP_VALUE, intval); fnvlist_add_nvlist(nvl, propname, propval); nvlist_free(propval); } static void vdev_props_set_sync(void *arg, dmu_tx_t *tx) { vdev_t *vd; nvlist_t *nvp = arg; spa_t *spa = dmu_tx_pool(tx)->dp_spa; objset_t *mos = spa->spa_meta_objset; nvpair_t *elem = NULL; uint64_t vdev_guid; nvlist_t *nvprops; vdev_guid = fnvlist_lookup_uint64(nvp, ZPOOL_VDEV_PROPS_SET_VDEV); nvprops = fnvlist_lookup_nvlist(nvp, ZPOOL_VDEV_PROPS_SET_PROPS); vd = spa_lookup_by_guid(spa, vdev_guid, B_TRUE); /* this vdev could get removed while waiting for this sync task */ if (vd == NULL) return; mutex_enter(&spa->spa_props_lock); while ((elem = nvlist_next_nvpair(nvprops, elem)) != NULL) { uint64_t intval, objid = 0; const char *strval; vdev_prop_t prop; const char *propname = nvpair_name(elem); zprop_type_t proptype; /* * Set vdev property values in the vdev props mos object. */ if (vd->vdev_root_zap != 0) { objid = vd->vdev_root_zap; } else if (vd->vdev_top_zap != 0) { objid = vd->vdev_top_zap; } else if (vd->vdev_leaf_zap != 0) { objid = vd->vdev_leaf_zap; } else { /* * XXX: implement vdev_props_set_check() */ panic("vdev not root/top/leaf"); } switch (prop = vdev_name_to_prop(propname)) { case VDEV_PROP_USERPROP: if (vdev_prop_user(propname)) { strval = fnvpair_value_string(elem); if (strlen(strval) == 0) { /* remove the property if value == "" */ (void) zap_remove(mos, objid, propname, tx); } else { VERIFY0(zap_update(mos, objid, propname, 1, strlen(strval) + 1, strval, tx)); } spa_history_log_internal(spa, "vdev set", tx, "vdev_guid=%llu: %s=%s", (u_longlong_t)vdev_guid, nvpair_name(elem), strval); } break; default: /* normalize the property name */ propname = vdev_prop_to_name(prop); proptype = vdev_prop_get_type(prop); if (nvpair_type(elem) == DATA_TYPE_STRING) { ASSERT(proptype == PROP_TYPE_STRING); strval = fnvpair_value_string(elem); VERIFY0(zap_update(mos, objid, propname, 1, strlen(strval) + 1, strval, tx)); spa_history_log_internal(spa, "vdev set", tx, "vdev_guid=%llu: %s=%s", (u_longlong_t)vdev_guid, nvpair_name(elem), strval); } else if (nvpair_type(elem) == DATA_TYPE_UINT64) { intval = fnvpair_value_uint64(elem); if (proptype == PROP_TYPE_INDEX) { const char *unused; VERIFY0(vdev_prop_index_to_string( prop, intval, &unused)); } VERIFY0(zap_update(mos, objid, propname, sizeof (uint64_t), 1, &intval, tx)); spa_history_log_internal(spa, "vdev set", tx, "vdev_guid=%llu: %s=%lld", (u_longlong_t)vdev_guid, nvpair_name(elem), (longlong_t)intval); } else { panic("invalid vdev property type %u", nvpair_type(elem)); } } } mutex_exit(&spa->spa_props_lock); } int vdev_prop_set(vdev_t *vd, nvlist_t *innvl, nvlist_t *outnvl) { spa_t *spa = vd->vdev_spa; nvpair_t *elem = NULL; uint64_t vdev_guid; nvlist_t *nvprops; int error = 0; ASSERT(vd != NULL); if (nvlist_lookup_uint64(innvl, ZPOOL_VDEV_PROPS_SET_VDEV, &vdev_guid) != 0) return (SET_ERROR(EINVAL)); if (nvlist_lookup_nvlist(innvl, ZPOOL_VDEV_PROPS_SET_PROPS, &nvprops) != 0) return (SET_ERROR(EINVAL)); if ((vd = spa_lookup_by_guid(spa, vdev_guid, B_TRUE)) == NULL) return (SET_ERROR(EINVAL)); while ((elem = nvlist_next_nvpair(nvprops, elem)) != NULL) { const char *propname = nvpair_name(elem); vdev_prop_t prop = vdev_name_to_prop(propname); uint64_t intval = 0; const char *strval = NULL; if (prop == VDEV_PROP_USERPROP && !vdev_prop_user(propname)) { error = EINVAL; goto end; } if (vdev_prop_readonly(prop)) { error = EROFS; goto end; } /* Special Processing */ switch (prop) { case VDEV_PROP_PATH: if (vd->vdev_path == NULL) { error = EROFS; break; } if (nvpair_value_string(elem, &strval) != 0) { error = EINVAL; break; } /* New path must start with /dev/ */ if (strncmp(strval, "/dev/", 5)) { error = EINVAL; break; } error = spa_vdev_setpath(spa, vdev_guid, strval); break; case VDEV_PROP_ALLOCATING: if (nvpair_value_uint64(elem, &intval) != 0) { error = EINVAL; break; } if (intval != vd->vdev_noalloc) break; if (intval == 0) error = spa_vdev_noalloc(spa, vdev_guid); else error = spa_vdev_alloc(spa, vdev_guid); break; case VDEV_PROP_FAILFAST: if (nvpair_value_uint64(elem, &intval) != 0) { error = EINVAL; break; } vd->vdev_failfast = intval & 1; break; case VDEV_PROP_CHECKSUM_N: if (nvpair_value_uint64(elem, &intval) != 0) { error = EINVAL; break; } vd->vdev_checksum_n = intval; break; case VDEV_PROP_CHECKSUM_T: if (nvpair_value_uint64(elem, &intval) != 0) { error = EINVAL; break; } vd->vdev_checksum_t = intval; break; case VDEV_PROP_IO_N: if (nvpair_value_uint64(elem, &intval) != 0) { error = EINVAL; break; } vd->vdev_io_n = intval; break; case VDEV_PROP_IO_T: if (nvpair_value_uint64(elem, &intval) != 0) { error = EINVAL; break; } vd->vdev_io_t = intval; break; default: /* Most processing is done in vdev_props_set_sync */ break; } end: if (error != 0) { intval = error; vdev_prop_add_list(outnvl, propname, strval, intval, 0); return (error); } } return (dsl_sync_task(spa->spa_name, NULL, vdev_props_set_sync, innvl, 6, ZFS_SPACE_CHECK_EXTRA_RESERVED)); } int vdev_prop_get(vdev_t *vd, nvlist_t *innvl, nvlist_t *outnvl) { spa_t *spa = vd->vdev_spa; objset_t *mos = spa->spa_meta_objset; int err = 0; uint64_t objid; uint64_t vdev_guid; nvpair_t *elem = NULL; nvlist_t *nvprops = NULL; uint64_t intval = 0; char *strval = NULL; const char *propname = NULL; vdev_prop_t prop; ASSERT(vd != NULL); ASSERT(mos != NULL); if (nvlist_lookup_uint64(innvl, ZPOOL_VDEV_PROPS_GET_VDEV, &vdev_guid) != 0) return (SET_ERROR(EINVAL)); nvlist_lookup_nvlist(innvl, ZPOOL_VDEV_PROPS_GET_PROPS, &nvprops); if (vd->vdev_root_zap != 0) { objid = vd->vdev_root_zap; } else if (vd->vdev_top_zap != 0) { objid = vd->vdev_top_zap; } else if (vd->vdev_leaf_zap != 0) { objid = vd->vdev_leaf_zap; } else { return (SET_ERROR(EINVAL)); } ASSERT(objid != 0); mutex_enter(&spa->spa_props_lock); if (nvprops != NULL) { char namebuf[64] = { 0 }; while ((elem = nvlist_next_nvpair(nvprops, elem)) != NULL) { intval = 0; strval = NULL; propname = nvpair_name(elem); prop = vdev_name_to_prop(propname); zprop_source_t src = ZPROP_SRC_DEFAULT; uint64_t integer_size, num_integers; switch (prop) { /* Special Read-only Properties */ case VDEV_PROP_NAME: strval = vdev_name(vd, namebuf, sizeof (namebuf)); if (strval == NULL) continue; vdev_prop_add_list(outnvl, propname, strval, 0, ZPROP_SRC_NONE); continue; case VDEV_PROP_CAPACITY: /* percent used */ intval = (vd->vdev_stat.vs_dspace == 0) ? 0 : (vd->vdev_stat.vs_alloc * 100 / vd->vdev_stat.vs_dspace); vdev_prop_add_list(outnvl, propname, NULL, intval, ZPROP_SRC_NONE); continue; case VDEV_PROP_STATE: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_state, ZPROP_SRC_NONE); continue; case VDEV_PROP_GUID: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_guid, ZPROP_SRC_NONE); continue; case VDEV_PROP_ASIZE: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_asize, ZPROP_SRC_NONE); continue; case VDEV_PROP_PSIZE: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_psize, ZPROP_SRC_NONE); continue; case VDEV_PROP_ASHIFT: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_ashift, ZPROP_SRC_NONE); continue; case VDEV_PROP_SIZE: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_dspace, ZPROP_SRC_NONE); continue; case VDEV_PROP_FREE: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_dspace - vd->vdev_stat.vs_alloc, ZPROP_SRC_NONE); continue; case VDEV_PROP_ALLOCATED: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_alloc, ZPROP_SRC_NONE); continue; case VDEV_PROP_EXPANDSZ: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_esize, ZPROP_SRC_NONE); continue; case VDEV_PROP_FRAGMENTATION: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_fragmentation, ZPROP_SRC_NONE); continue; case VDEV_PROP_PARITY: vdev_prop_add_list(outnvl, propname, NULL, vdev_get_nparity(vd), ZPROP_SRC_NONE); continue; case VDEV_PROP_PATH: if (vd->vdev_path == NULL) continue; vdev_prop_add_list(outnvl, propname, vd->vdev_path, 0, ZPROP_SRC_NONE); continue; case VDEV_PROP_DEVID: if (vd->vdev_devid == NULL) continue; vdev_prop_add_list(outnvl, propname, vd->vdev_devid, 0, ZPROP_SRC_NONE); continue; case VDEV_PROP_PHYS_PATH: if (vd->vdev_physpath == NULL) continue; vdev_prop_add_list(outnvl, propname, vd->vdev_physpath, 0, ZPROP_SRC_NONE); continue; case VDEV_PROP_ENC_PATH: if (vd->vdev_enc_sysfs_path == NULL) continue; vdev_prop_add_list(outnvl, propname, vd->vdev_enc_sysfs_path, 0, ZPROP_SRC_NONE); continue; case VDEV_PROP_FRU: if (vd->vdev_fru == NULL) continue; vdev_prop_add_list(outnvl, propname, vd->vdev_fru, 0, ZPROP_SRC_NONE); continue; case VDEV_PROP_PARENT: if (vd->vdev_parent != NULL) { strval = vdev_name(vd->vdev_parent, namebuf, sizeof (namebuf)); vdev_prop_add_list(outnvl, propname, strval, 0, ZPROP_SRC_NONE); } continue; case VDEV_PROP_CHILDREN: if (vd->vdev_children > 0) strval = kmem_zalloc(ZAP_MAXVALUELEN, KM_SLEEP); for (uint64_t i = 0; i < vd->vdev_children; i++) { const char *vname; vname = vdev_name(vd->vdev_child[i], namebuf, sizeof (namebuf)); if (vname == NULL) vname = "(unknown)"; if (strlen(strval) > 0) strlcat(strval, ",", ZAP_MAXVALUELEN); strlcat(strval, vname, ZAP_MAXVALUELEN); } if (strval != NULL) { vdev_prop_add_list(outnvl, propname, strval, 0, ZPROP_SRC_NONE); kmem_free(strval, ZAP_MAXVALUELEN); } continue; case VDEV_PROP_NUMCHILDREN: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_children, ZPROP_SRC_NONE); continue; case VDEV_PROP_READ_ERRORS: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_read_errors, ZPROP_SRC_NONE); continue; case VDEV_PROP_WRITE_ERRORS: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_write_errors, ZPROP_SRC_NONE); continue; case VDEV_PROP_CHECKSUM_ERRORS: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_checksum_errors, ZPROP_SRC_NONE); continue; case VDEV_PROP_INITIALIZE_ERRORS: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_initialize_errors, ZPROP_SRC_NONE); continue; case VDEV_PROP_OPS_NULL: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_ops[ZIO_TYPE_NULL], ZPROP_SRC_NONE); continue; case VDEV_PROP_OPS_READ: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_ops[ZIO_TYPE_READ], ZPROP_SRC_NONE); continue; case VDEV_PROP_OPS_WRITE: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_ops[ZIO_TYPE_WRITE], ZPROP_SRC_NONE); continue; case VDEV_PROP_OPS_FREE: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_ops[ZIO_TYPE_FREE], ZPROP_SRC_NONE); continue; case VDEV_PROP_OPS_CLAIM: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_ops[ZIO_TYPE_CLAIM], ZPROP_SRC_NONE); continue; case VDEV_PROP_OPS_TRIM: /* * TRIM ops and bytes are reported to user * space as ZIO_TYPE_IOCTL. This is done to * preserve the vdev_stat_t structure layout * for user space. */ vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_ops[ZIO_TYPE_IOCTL], ZPROP_SRC_NONE); continue; case VDEV_PROP_BYTES_NULL: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_bytes[ZIO_TYPE_NULL], ZPROP_SRC_NONE); continue; case VDEV_PROP_BYTES_READ: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_bytes[ZIO_TYPE_READ], ZPROP_SRC_NONE); continue; case VDEV_PROP_BYTES_WRITE: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_bytes[ZIO_TYPE_WRITE], ZPROP_SRC_NONE); continue; case VDEV_PROP_BYTES_FREE: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_bytes[ZIO_TYPE_FREE], ZPROP_SRC_NONE); continue; case VDEV_PROP_BYTES_CLAIM: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_bytes[ZIO_TYPE_CLAIM], ZPROP_SRC_NONE); continue; case VDEV_PROP_BYTES_TRIM: /* * TRIM ops and bytes are reported to user * space as ZIO_TYPE_IOCTL. This is done to * preserve the vdev_stat_t structure layout * for user space. */ vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_stat.vs_bytes[ZIO_TYPE_IOCTL], ZPROP_SRC_NONE); continue; case VDEV_PROP_REMOVING: vdev_prop_add_list(outnvl, propname, NULL, vd->vdev_removing, ZPROP_SRC_NONE); continue; /* Numeric Properites */ case VDEV_PROP_ALLOCATING: /* Leaf vdevs cannot have this property */ if (vd->vdev_mg == NULL && vd->vdev_top != NULL) { src = ZPROP_SRC_NONE; intval = ZPROP_BOOLEAN_NA; } else { err = vdev_prop_get_int(vd, prop, &intval); if (err && err != ENOENT) break; if (intval == vdev_prop_default_numeric(prop)) src = ZPROP_SRC_DEFAULT; else src = ZPROP_SRC_LOCAL; } vdev_prop_add_list(outnvl, propname, NULL, intval, src); break; case VDEV_PROP_FAILFAST: src = ZPROP_SRC_LOCAL; strval = NULL; err = zap_lookup(mos, objid, nvpair_name(elem), sizeof (uint64_t), 1, &intval); if (err == ENOENT) { intval = vdev_prop_default_numeric( prop); err = 0; } else if (err) { break; } if (intval == vdev_prop_default_numeric(prop)) src = ZPROP_SRC_DEFAULT; vdev_prop_add_list(outnvl, propname, strval, intval, src); break; case VDEV_PROP_CHECKSUM_N: case VDEV_PROP_CHECKSUM_T: case VDEV_PROP_IO_N: case VDEV_PROP_IO_T: err = vdev_prop_get_int(vd, prop, &intval); if (err && err != ENOENT) break; if (intval == vdev_prop_default_numeric(prop)) src = ZPROP_SRC_DEFAULT; else src = ZPROP_SRC_LOCAL; vdev_prop_add_list(outnvl, propname, NULL, intval, src); break; /* Text Properties */ case VDEV_PROP_COMMENT: /* Exists in the ZAP below */ /* FALLTHRU */ case VDEV_PROP_USERPROP: /* User Properites */ src = ZPROP_SRC_LOCAL; err = zap_length(mos, objid, nvpair_name(elem), &integer_size, &num_integers); if (err) break; switch (integer_size) { case 8: /* User properties cannot be integers */ err = EINVAL; break; case 1: /* string property */ strval = kmem_alloc(num_integers, KM_SLEEP); err = zap_lookup(mos, objid, nvpair_name(elem), 1, num_integers, strval); if (err) { kmem_free(strval, num_integers); break; } vdev_prop_add_list(outnvl, propname, strval, 0, src); kmem_free(strval, num_integers); break; } break; default: err = ENOENT; break; } if (err) break; } } else { /* * Get all properties from the MOS vdev property object. */ zap_cursor_t zc; zap_attribute_t za; for (zap_cursor_init(&zc, mos, objid); (err = zap_cursor_retrieve(&zc, &za)) == 0; zap_cursor_advance(&zc)) { intval = 0; strval = NULL; zprop_source_t src = ZPROP_SRC_DEFAULT; propname = za.za_name; switch (za.za_integer_length) { case 8: /* We do not allow integer user properties */ /* This is likely an internal value */ break; case 1: /* string property */ strval = kmem_alloc(za.za_num_integers, KM_SLEEP); err = zap_lookup(mos, objid, za.za_name, 1, za.za_num_integers, strval); if (err) { kmem_free(strval, za.za_num_integers); break; } vdev_prop_add_list(outnvl, propname, strval, 0, src); kmem_free(strval, za.za_num_integers); break; default: break; } } zap_cursor_fini(&zc); } mutex_exit(&spa->spa_props_lock); if (err && err != ENOENT) { return (err); } return (0); } EXPORT_SYMBOL(vdev_fault); EXPORT_SYMBOL(vdev_degrade); EXPORT_SYMBOL(vdev_online); EXPORT_SYMBOL(vdev_offline); EXPORT_SYMBOL(vdev_clear); ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, default_ms_count, UINT, ZMOD_RW, "Target number of metaslabs per top-level vdev"); ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, default_ms_shift, UINT, ZMOD_RW, "Default lower limit for metaslab size"); ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, max_ms_shift, UINT, ZMOD_RW, "Default upper limit for metaslab size"); ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, min_ms_count, UINT, ZMOD_RW, "Minimum number of metaslabs per top-level vdev"); ZFS_MODULE_PARAM(zfs_vdev, zfs_vdev_, ms_count_limit, UINT, ZMOD_RW, "Practical upper limit of total metaslabs per top-level vdev"); ZFS_MODULE_PARAM(zfs, zfs_, slow_io_events_per_second, UINT, ZMOD_RW, "Rate limit slow IO (delay) events to this many per second"); /* BEGIN CSTYLED */ ZFS_MODULE_PARAM(zfs, zfs_, checksum_events_per_second, UINT, ZMOD_RW, "Rate limit checksum events to this many checksum errors per second " "(do not set below ZED threshold)."); /* END CSTYLED */ ZFS_MODULE_PARAM(zfs, zfs_, scan_ignore_errors, INT, ZMOD_RW, "Ignore errors during resilver/scrub"); ZFS_MODULE_PARAM(zfs_vdev, vdev_, validate_skip, INT, ZMOD_RW, "Bypass vdev_validate()"); ZFS_MODULE_PARAM(zfs, zfs_, nocacheflush, INT, ZMOD_RW, "Disable cache flushes"); ZFS_MODULE_PARAM(zfs, zfs_, embedded_slog_min_ms, UINT, ZMOD_RW, "Minimum number of metaslabs required to dedicate one for log blocks"); /* BEGIN CSTYLED */ ZFS_MODULE_PARAM_CALL(zfs_vdev, zfs_vdev_, min_auto_ashift, param_set_min_auto_ashift, param_get_uint, ZMOD_RW, "Minimum ashift used when creating new top-level vdevs"); ZFS_MODULE_PARAM_CALL(zfs_vdev, zfs_vdev_, max_auto_ashift, param_set_max_auto_ashift, param_get_uint, ZMOD_RW, "Maximum ashift used when optimizing for logical -> physical sector " "size on new top-level vdevs"); /* END CSTYLED */ diff --git a/sys/contrib/openzfs/module/zfs/vdev_initialize.c b/sys/contrib/openzfs/module/zfs/vdev_initialize.c index 75beb0cc3d12..ffdcef1972c3 100644 --- a/sys/contrib/openzfs/module/zfs/vdev_initialize.c +++ b/sys/contrib/openzfs/module/zfs/vdev_initialize.c @@ -1,762 +1,824 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2016, 2019 by Delphix. All rights reserved. */ #include #include #include #include #include #include #include #include #include /* * Value that is written to disk during initialization. */ static uint64_t zfs_initialize_value = 0xdeadbeefdeadbeeeULL; /* maximum number of I/Os outstanding per leaf vdev */ static const int zfs_initialize_limit = 1; /* size of initializing writes; default 1MiB, see zfs_remove_max_segment */ static uint64_t zfs_initialize_chunk_size = 1024 * 1024; static boolean_t vdev_initialize_should_stop(vdev_t *vd) { return (vd->vdev_initialize_exit_wanted || !vdev_writeable(vd) || vd->vdev_detached || vd->vdev_top->vdev_removing); } static void vdev_initialize_zap_update_sync(void *arg, dmu_tx_t *tx) { /* * We pass in the guid instead of the vdev_t since the vdev may * have been freed prior to the sync task being processed. This * happens when a vdev is detached as we call spa_config_vdev_exit(), * stop the initializing thread, schedule the sync task, and free * the vdev. Later when the scheduled sync task is invoked, it would * find that the vdev has been freed. */ uint64_t guid = *(uint64_t *)arg; uint64_t txg = dmu_tx_get_txg(tx); kmem_free(arg, sizeof (uint64_t)); vdev_t *vd = spa_lookup_by_guid(tx->tx_pool->dp_spa, guid, B_FALSE); if (vd == NULL || vd->vdev_top->vdev_removing || !vdev_is_concrete(vd)) return; uint64_t last_offset = vd->vdev_initialize_offset[txg & TXG_MASK]; vd->vdev_initialize_offset[txg & TXG_MASK] = 0; VERIFY(vd->vdev_leaf_zap != 0); objset_t *mos = vd->vdev_spa->spa_meta_objset; if (last_offset > 0) { vd->vdev_initialize_last_offset = last_offset; VERIFY0(zap_update(mos, vd->vdev_leaf_zap, VDEV_LEAF_ZAP_INITIALIZE_LAST_OFFSET, sizeof (last_offset), 1, &last_offset, tx)); } if (vd->vdev_initialize_action_time > 0) { uint64_t val = (uint64_t)vd->vdev_initialize_action_time; VERIFY0(zap_update(mos, vd->vdev_leaf_zap, VDEV_LEAF_ZAP_INITIALIZE_ACTION_TIME, sizeof (val), 1, &val, tx)); } uint64_t initialize_state = vd->vdev_initialize_state; VERIFY0(zap_update(mos, vd->vdev_leaf_zap, VDEV_LEAF_ZAP_INITIALIZE_STATE, sizeof (initialize_state), 1, &initialize_state, tx)); } +static void +vdev_initialize_zap_remove_sync(void *arg, dmu_tx_t *tx) +{ + uint64_t guid = *(uint64_t *)arg; + + kmem_free(arg, sizeof (uint64_t)); + + vdev_t *vd = spa_lookup_by_guid(tx->tx_pool->dp_spa, guid, B_FALSE); + if (vd == NULL || vd->vdev_top->vdev_removing || !vdev_is_concrete(vd)) + return; + + ASSERT3S(vd->vdev_initialize_state, ==, VDEV_INITIALIZE_NONE); + ASSERT3U(vd->vdev_leaf_zap, !=, 0); + + vd->vdev_initialize_last_offset = 0; + vd->vdev_initialize_action_time = 0; + + objset_t *mos = vd->vdev_spa->spa_meta_objset; + int error; + + error = zap_remove(mos, vd->vdev_leaf_zap, + VDEV_LEAF_ZAP_INITIALIZE_LAST_OFFSET, tx); + VERIFY(error == 0 || error == ENOENT); + + error = zap_remove(mos, vd->vdev_leaf_zap, + VDEV_LEAF_ZAP_INITIALIZE_STATE, tx); + VERIFY(error == 0 || error == ENOENT); + + error = zap_remove(mos, vd->vdev_leaf_zap, + VDEV_LEAF_ZAP_INITIALIZE_ACTION_TIME, tx); + VERIFY(error == 0 || error == ENOENT); +} + static void vdev_initialize_change_state(vdev_t *vd, vdev_initializing_state_t new_state) { ASSERT(MUTEX_HELD(&vd->vdev_initialize_lock)); spa_t *spa = vd->vdev_spa; if (new_state == vd->vdev_initialize_state) return; /* * Copy the vd's guid, this will be freed by the sync task. */ uint64_t *guid = kmem_zalloc(sizeof (uint64_t), KM_SLEEP); *guid = vd->vdev_guid; /* * If we're suspending, then preserving the original start time. */ if (vd->vdev_initialize_state != VDEV_INITIALIZE_SUSPENDED) { vd->vdev_initialize_action_time = gethrestime_sec(); } vdev_initializing_state_t old_state = vd->vdev_initialize_state; vd->vdev_initialize_state = new_state; dmu_tx_t *tx = dmu_tx_create_dd(spa_get_dsl(spa)->dp_mos_dir); VERIFY0(dmu_tx_assign(tx, TXG_WAIT)); - dsl_sync_task_nowait(spa_get_dsl(spa), vdev_initialize_zap_update_sync, - guid, tx); + + if (new_state != VDEV_INITIALIZE_NONE) { + dsl_sync_task_nowait(spa_get_dsl(spa), + vdev_initialize_zap_update_sync, guid, tx); + } else { + dsl_sync_task_nowait(spa_get_dsl(spa), + vdev_initialize_zap_remove_sync, guid, tx); + } switch (new_state) { case VDEV_INITIALIZE_ACTIVE: spa_history_log_internal(spa, "initialize", tx, "vdev=%s activated", vd->vdev_path); break; case VDEV_INITIALIZE_SUSPENDED: spa_history_log_internal(spa, "initialize", tx, "vdev=%s suspended", vd->vdev_path); break; case VDEV_INITIALIZE_CANCELED: if (old_state == VDEV_INITIALIZE_ACTIVE || old_state == VDEV_INITIALIZE_SUSPENDED) spa_history_log_internal(spa, "initialize", tx, "vdev=%s canceled", vd->vdev_path); break; case VDEV_INITIALIZE_COMPLETE: spa_history_log_internal(spa, "initialize", tx, "vdev=%s complete", vd->vdev_path); break; + case VDEV_INITIALIZE_NONE: + spa_history_log_internal(spa, "uninitialize", tx, + "vdev=%s", vd->vdev_path); + break; default: panic("invalid state %llu", (unsigned long long)new_state); } dmu_tx_commit(tx); if (new_state != VDEV_INITIALIZE_ACTIVE) spa_notify_waiters(spa); } static void vdev_initialize_cb(zio_t *zio) { vdev_t *vd = zio->io_vd; mutex_enter(&vd->vdev_initialize_io_lock); if (zio->io_error == ENXIO && !vdev_writeable(vd)) { /* * The I/O failed because the vdev was unavailable; roll the * last offset back. (This works because spa_sync waits on * spa_txg_zio before it runs sync tasks.) */ uint64_t *off = &vd->vdev_initialize_offset[zio->io_txg & TXG_MASK]; *off = MIN(*off, zio->io_offset); } else { /* * Since initializing is best-effort, we ignore I/O errors and * rely on vdev_probe to determine if the errors are more * critical. */ if (zio->io_error != 0) vd->vdev_stat.vs_initialize_errors++; vd->vdev_initialize_bytes_done += zio->io_orig_size; } ASSERT3U(vd->vdev_initialize_inflight, >, 0); vd->vdev_initialize_inflight--; cv_broadcast(&vd->vdev_initialize_io_cv); mutex_exit(&vd->vdev_initialize_io_lock); spa_config_exit(vd->vdev_spa, SCL_STATE_ALL, vd); } /* Takes care of physical writing and limiting # of concurrent ZIOs. */ static int vdev_initialize_write(vdev_t *vd, uint64_t start, uint64_t size, abd_t *data) { spa_t *spa = vd->vdev_spa; /* Limit inflight initializing I/Os */ mutex_enter(&vd->vdev_initialize_io_lock); while (vd->vdev_initialize_inflight >= zfs_initialize_limit) { cv_wait(&vd->vdev_initialize_io_cv, &vd->vdev_initialize_io_lock); } vd->vdev_initialize_inflight++; mutex_exit(&vd->vdev_initialize_io_lock); dmu_tx_t *tx = dmu_tx_create_dd(spa_get_dsl(spa)->dp_mos_dir); VERIFY0(dmu_tx_assign(tx, TXG_WAIT)); uint64_t txg = dmu_tx_get_txg(tx); spa_config_enter(spa, SCL_STATE_ALL, vd, RW_READER); mutex_enter(&vd->vdev_initialize_lock); if (vd->vdev_initialize_offset[txg & TXG_MASK] == 0) { uint64_t *guid = kmem_zalloc(sizeof (uint64_t), KM_SLEEP); *guid = vd->vdev_guid; /* This is the first write of this txg. */ dsl_sync_task_nowait(spa_get_dsl(spa), vdev_initialize_zap_update_sync, guid, tx); } /* * We know the vdev struct will still be around since all * consumers of vdev_free must stop the initialization first. */ if (vdev_initialize_should_stop(vd)) { mutex_enter(&vd->vdev_initialize_io_lock); ASSERT3U(vd->vdev_initialize_inflight, >, 0); vd->vdev_initialize_inflight--; mutex_exit(&vd->vdev_initialize_io_lock); spa_config_exit(vd->vdev_spa, SCL_STATE_ALL, vd); mutex_exit(&vd->vdev_initialize_lock); dmu_tx_commit(tx); return (SET_ERROR(EINTR)); } mutex_exit(&vd->vdev_initialize_lock); vd->vdev_initialize_offset[txg & TXG_MASK] = start + size; zio_nowait(zio_write_phys(spa->spa_txg_zio[txg & TXG_MASK], vd, start, size, data, ZIO_CHECKSUM_OFF, vdev_initialize_cb, NULL, ZIO_PRIORITY_INITIALIZING, ZIO_FLAG_CANFAIL, B_FALSE)); /* vdev_initialize_cb releases SCL_STATE_ALL */ dmu_tx_commit(tx); return (0); } /* * Callback to fill each ABD chunk with zfs_initialize_value. len must be * divisible by sizeof (uint64_t), and buf must be 8-byte aligned. The ABD * allocation will guarantee these for us. */ static int vdev_initialize_block_fill(void *buf, size_t len, void *unused) { (void) unused; ASSERT0(len % sizeof (uint64_t)); for (uint64_t i = 0; i < len; i += sizeof (uint64_t)) { *(uint64_t *)((char *)(buf) + i) = zfs_initialize_value; } return (0); } static abd_t * vdev_initialize_block_alloc(void) { /* Allocate ABD for filler data */ abd_t *data = abd_alloc_for_io(zfs_initialize_chunk_size, B_FALSE); ASSERT0(zfs_initialize_chunk_size % sizeof (uint64_t)); (void) abd_iterate_func(data, 0, zfs_initialize_chunk_size, vdev_initialize_block_fill, NULL); return (data); } static void vdev_initialize_block_free(abd_t *data) { abd_free(data); } static int vdev_initialize_ranges(vdev_t *vd, abd_t *data) { range_tree_t *rt = vd->vdev_initialize_tree; zfs_btree_t *bt = &rt->rt_root; zfs_btree_index_t where; for (range_seg_t *rs = zfs_btree_first(bt, &where); rs != NULL; rs = zfs_btree_next(bt, &where, &where)) { uint64_t size = rs_get_end(rs, rt) - rs_get_start(rs, rt); /* Split range into legally-sized physical chunks */ uint64_t writes_required = ((size - 1) / zfs_initialize_chunk_size) + 1; for (uint64_t w = 0; w < writes_required; w++) { int error; error = vdev_initialize_write(vd, VDEV_LABEL_START_SIZE + rs_get_start(rs, rt) + (w * zfs_initialize_chunk_size), MIN(size - (w * zfs_initialize_chunk_size), zfs_initialize_chunk_size), data); if (error != 0) return (error); } } return (0); } static void vdev_initialize_xlate_last_rs_end(void *arg, range_seg64_t *physical_rs) { uint64_t *last_rs_end = (uint64_t *)arg; if (physical_rs->rs_end > *last_rs_end) *last_rs_end = physical_rs->rs_end; } static void vdev_initialize_xlate_progress(void *arg, range_seg64_t *physical_rs) { vdev_t *vd = (vdev_t *)arg; uint64_t size = physical_rs->rs_end - physical_rs->rs_start; vd->vdev_initialize_bytes_est += size; if (vd->vdev_initialize_last_offset > physical_rs->rs_end) { vd->vdev_initialize_bytes_done += size; } else if (vd->vdev_initialize_last_offset > physical_rs->rs_start && vd->vdev_initialize_last_offset < physical_rs->rs_end) { vd->vdev_initialize_bytes_done += vd->vdev_initialize_last_offset - physical_rs->rs_start; } } static void vdev_initialize_calculate_progress(vdev_t *vd) { ASSERT(spa_config_held(vd->vdev_spa, SCL_CONFIG, RW_READER) || spa_config_held(vd->vdev_spa, SCL_CONFIG, RW_WRITER)); ASSERT(vd->vdev_leaf_zap != 0); vd->vdev_initialize_bytes_est = 0; vd->vdev_initialize_bytes_done = 0; for (uint64_t i = 0; i < vd->vdev_top->vdev_ms_count; i++) { metaslab_t *msp = vd->vdev_top->vdev_ms[i]; mutex_enter(&msp->ms_lock); uint64_t ms_free = (msp->ms_size - metaslab_allocated_space(msp)) / vdev_get_ndisks(vd->vdev_top); /* * Convert the metaslab range to a physical range * on our vdev. We use this to determine if we are * in the middle of this metaslab range. */ range_seg64_t logical_rs, physical_rs, remain_rs; logical_rs.rs_start = msp->ms_start; logical_rs.rs_end = msp->ms_start + msp->ms_size; /* Metaslab space after this offset has not been initialized */ vdev_xlate(vd, &logical_rs, &physical_rs, &remain_rs); if (vd->vdev_initialize_last_offset <= physical_rs.rs_start) { vd->vdev_initialize_bytes_est += ms_free; mutex_exit(&msp->ms_lock); continue; } /* Metaslab space before this offset has been initialized */ uint64_t last_rs_end = physical_rs.rs_end; if (!vdev_xlate_is_empty(&remain_rs)) { vdev_xlate_walk(vd, &remain_rs, vdev_initialize_xlate_last_rs_end, &last_rs_end); } if (vd->vdev_initialize_last_offset > last_rs_end) { vd->vdev_initialize_bytes_done += ms_free; vd->vdev_initialize_bytes_est += ms_free; mutex_exit(&msp->ms_lock); continue; } /* * If we get here, we're in the middle of initializing this * metaslab. Load it and walk the free tree for more accurate * progress estimation. */ VERIFY0(metaslab_load(msp)); zfs_btree_index_t where; range_tree_t *rt = msp->ms_allocatable; for (range_seg_t *rs = zfs_btree_first(&rt->rt_root, &where); rs; rs = zfs_btree_next(&rt->rt_root, &where, &where)) { logical_rs.rs_start = rs_get_start(rs, rt); logical_rs.rs_end = rs_get_end(rs, rt); vdev_xlate_walk(vd, &logical_rs, vdev_initialize_xlate_progress, vd); } mutex_exit(&msp->ms_lock); } } static int vdev_initialize_load(vdev_t *vd) { int err = 0; ASSERT(spa_config_held(vd->vdev_spa, SCL_CONFIG, RW_READER) || spa_config_held(vd->vdev_spa, SCL_CONFIG, RW_WRITER)); ASSERT(vd->vdev_leaf_zap != 0); if (vd->vdev_initialize_state == VDEV_INITIALIZE_ACTIVE || vd->vdev_initialize_state == VDEV_INITIALIZE_SUSPENDED) { err = zap_lookup(vd->vdev_spa->spa_meta_objset, vd->vdev_leaf_zap, VDEV_LEAF_ZAP_INITIALIZE_LAST_OFFSET, sizeof (vd->vdev_initialize_last_offset), 1, &vd->vdev_initialize_last_offset); if (err == ENOENT) { vd->vdev_initialize_last_offset = 0; err = 0; } } vdev_initialize_calculate_progress(vd); return (err); } static void vdev_initialize_xlate_range_add(void *arg, range_seg64_t *physical_rs) { vdev_t *vd = arg; /* Only add segments that we have not visited yet */ if (physical_rs->rs_end <= vd->vdev_initialize_last_offset) return; /* Pick up where we left off mid-range. */ if (vd->vdev_initialize_last_offset > physical_rs->rs_start) { zfs_dbgmsg("range write: vd %s changed (%llu, %llu) to " "(%llu, %llu)", vd->vdev_path, (u_longlong_t)physical_rs->rs_start, (u_longlong_t)physical_rs->rs_end, (u_longlong_t)vd->vdev_initialize_last_offset, (u_longlong_t)physical_rs->rs_end); ASSERT3U(physical_rs->rs_end, >, vd->vdev_initialize_last_offset); physical_rs->rs_start = vd->vdev_initialize_last_offset; } ASSERT3U(physical_rs->rs_end, >, physical_rs->rs_start); range_tree_add(vd->vdev_initialize_tree, physical_rs->rs_start, physical_rs->rs_end - physical_rs->rs_start); } /* * Convert the logical range into a physical range and add it to our * avl tree. */ static void vdev_initialize_range_add(void *arg, uint64_t start, uint64_t size) { vdev_t *vd = arg; range_seg64_t logical_rs; logical_rs.rs_start = start; logical_rs.rs_end = start + size; ASSERT(vd->vdev_ops->vdev_op_leaf); vdev_xlate_walk(vd, &logical_rs, vdev_initialize_xlate_range_add, arg); } static __attribute__((noreturn)) void vdev_initialize_thread(void *arg) { vdev_t *vd = arg; spa_t *spa = vd->vdev_spa; int error = 0; uint64_t ms_count = 0; ASSERT(vdev_is_concrete(vd)); spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER); vd->vdev_initialize_last_offset = 0; VERIFY0(vdev_initialize_load(vd)); abd_t *deadbeef = vdev_initialize_block_alloc(); vd->vdev_initialize_tree = range_tree_create(NULL, RANGE_SEG64, NULL, 0, 0); for (uint64_t i = 0; !vd->vdev_detached && i < vd->vdev_top->vdev_ms_count; i++) { metaslab_t *msp = vd->vdev_top->vdev_ms[i]; boolean_t unload_when_done = B_FALSE; /* * If we've expanded the top-level vdev or it's our * first pass, calculate our progress. */ if (vd->vdev_top->vdev_ms_count != ms_count) { vdev_initialize_calculate_progress(vd); ms_count = vd->vdev_top->vdev_ms_count; } spa_config_exit(spa, SCL_CONFIG, FTAG); metaslab_disable(msp); mutex_enter(&msp->ms_lock); if (!msp->ms_loaded && !msp->ms_loading) unload_when_done = B_TRUE; VERIFY0(metaslab_load(msp)); range_tree_walk(msp->ms_allocatable, vdev_initialize_range_add, vd); mutex_exit(&msp->ms_lock); error = vdev_initialize_ranges(vd, deadbeef); metaslab_enable(msp, B_TRUE, unload_when_done); spa_config_enter(spa, SCL_CONFIG, FTAG, RW_READER); range_tree_vacate(vd->vdev_initialize_tree, NULL, NULL); if (error != 0) break; } spa_config_exit(spa, SCL_CONFIG, FTAG); mutex_enter(&vd->vdev_initialize_io_lock); while (vd->vdev_initialize_inflight > 0) { cv_wait(&vd->vdev_initialize_io_cv, &vd->vdev_initialize_io_lock); } mutex_exit(&vd->vdev_initialize_io_lock); range_tree_destroy(vd->vdev_initialize_tree); vdev_initialize_block_free(deadbeef); vd->vdev_initialize_tree = NULL; mutex_enter(&vd->vdev_initialize_lock); if (!vd->vdev_initialize_exit_wanted) { if (vdev_writeable(vd)) { vdev_initialize_change_state(vd, VDEV_INITIALIZE_COMPLETE); } else if (vd->vdev_faulted) { vdev_initialize_change_state(vd, VDEV_INITIALIZE_CANCELED); } } ASSERT(vd->vdev_initialize_thread != NULL || vd->vdev_initialize_inflight == 0); /* * Drop the vdev_initialize_lock while we sync out the * txg since it's possible that a device might be trying to * come online and must check to see if it needs to restart an * initialization. That thread will be holding the spa_config_lock * which would prevent the txg_wait_synced from completing. */ mutex_exit(&vd->vdev_initialize_lock); txg_wait_synced(spa_get_dsl(spa), 0); mutex_enter(&vd->vdev_initialize_lock); vd->vdev_initialize_thread = NULL; cv_broadcast(&vd->vdev_initialize_cv); mutex_exit(&vd->vdev_initialize_lock); thread_exit(); } /* * Initiates a device. Caller must hold vdev_initialize_lock. * Device must be a leaf and not already be initializing. */ void vdev_initialize(vdev_t *vd) { ASSERT(MUTEX_HELD(&vd->vdev_initialize_lock)); ASSERT(vd->vdev_ops->vdev_op_leaf); ASSERT(vdev_is_concrete(vd)); ASSERT3P(vd->vdev_initialize_thread, ==, NULL); ASSERT(!vd->vdev_detached); ASSERT(!vd->vdev_initialize_exit_wanted); ASSERT(!vd->vdev_top->vdev_removing); vdev_initialize_change_state(vd, VDEV_INITIALIZE_ACTIVE); vd->vdev_initialize_thread = thread_create(NULL, 0, vdev_initialize_thread, vd, 0, &p0, TS_RUN, maxclsyspri); } +/* + * Uninitializes a device. Caller must hold vdev_initialize_lock. + * Device must be a leaf and not already be initializing. + */ +void +vdev_uninitialize(vdev_t *vd) +{ + ASSERT(MUTEX_HELD(&vd->vdev_initialize_lock)); + ASSERT(vd->vdev_ops->vdev_op_leaf); + ASSERT(vdev_is_concrete(vd)); + ASSERT3P(vd->vdev_initialize_thread, ==, NULL); + ASSERT(!vd->vdev_detached); + ASSERT(!vd->vdev_initialize_exit_wanted); + ASSERT(!vd->vdev_top->vdev_removing); + + vdev_initialize_change_state(vd, VDEV_INITIALIZE_NONE); +} + /* * Wait for the initialize thread to be terminated (cancelled or stopped). */ static void vdev_initialize_stop_wait_impl(vdev_t *vd) { ASSERT(MUTEX_HELD(&vd->vdev_initialize_lock)); while (vd->vdev_initialize_thread != NULL) cv_wait(&vd->vdev_initialize_cv, &vd->vdev_initialize_lock); ASSERT3P(vd->vdev_initialize_thread, ==, NULL); vd->vdev_initialize_exit_wanted = B_FALSE; } /* * Wait for vdev initialize threads which were either to cleanly exit. */ void vdev_initialize_stop_wait(spa_t *spa, list_t *vd_list) { (void) spa; vdev_t *vd; ASSERT(MUTEX_HELD(&spa_namespace_lock)); while ((vd = list_remove_head(vd_list)) != NULL) { mutex_enter(&vd->vdev_initialize_lock); vdev_initialize_stop_wait_impl(vd); mutex_exit(&vd->vdev_initialize_lock); } } /* * Stop initializing a device, with the resultant initializing state being * tgt_state. For blocking behavior pass NULL for vd_list. Otherwise, when * a list_t is provided the stopping vdev is inserted in to the list. Callers * are then required to call vdev_initialize_stop_wait() to block for all the * initialization threads to exit. The caller must hold vdev_initialize_lock * and must not be writing to the spa config, as the initializing thread may * try to enter the config as a reader before exiting. */ void vdev_initialize_stop(vdev_t *vd, vdev_initializing_state_t tgt_state, list_t *vd_list) { ASSERT(!spa_config_held(vd->vdev_spa, SCL_CONFIG|SCL_STATE, RW_WRITER)); ASSERT(MUTEX_HELD(&vd->vdev_initialize_lock)); ASSERT(vd->vdev_ops->vdev_op_leaf); ASSERT(vdev_is_concrete(vd)); /* * Allow cancel requests to proceed even if the initialize thread * has stopped. */ if (vd->vdev_initialize_thread == NULL && tgt_state != VDEV_INITIALIZE_CANCELED) { return; } vdev_initialize_change_state(vd, tgt_state); vd->vdev_initialize_exit_wanted = B_TRUE; if (vd_list == NULL) { vdev_initialize_stop_wait_impl(vd); } else { ASSERT(MUTEX_HELD(&spa_namespace_lock)); list_insert_tail(vd_list, vd); } } static void vdev_initialize_stop_all_impl(vdev_t *vd, vdev_initializing_state_t tgt_state, list_t *vd_list) { if (vd->vdev_ops->vdev_op_leaf && vdev_is_concrete(vd)) { mutex_enter(&vd->vdev_initialize_lock); vdev_initialize_stop(vd, tgt_state, vd_list); mutex_exit(&vd->vdev_initialize_lock); return; } for (uint64_t i = 0; i < vd->vdev_children; i++) { vdev_initialize_stop_all_impl(vd->vdev_child[i], tgt_state, vd_list); } } /* * Convenience function to stop initializing of a vdev tree and set all * initialize thread pointers to NULL. */ void vdev_initialize_stop_all(vdev_t *vd, vdev_initializing_state_t tgt_state) { spa_t *spa = vd->vdev_spa; list_t vd_list; ASSERT(MUTEX_HELD(&spa_namespace_lock)); list_create(&vd_list, sizeof (vdev_t), offsetof(vdev_t, vdev_initialize_node)); vdev_initialize_stop_all_impl(vd, tgt_state, &vd_list); vdev_initialize_stop_wait(spa, &vd_list); if (vd->vdev_spa->spa_sync_on) { /* Make sure that our state has been synced to disk */ txg_wait_synced(spa_get_dsl(vd->vdev_spa), 0); } list_destroy(&vd_list); } void vdev_initialize_restart(vdev_t *vd) { ASSERT(MUTEX_HELD(&spa_namespace_lock)); ASSERT(!spa_config_held(vd->vdev_spa, SCL_ALL, RW_WRITER)); if (vd->vdev_leaf_zap != 0) { mutex_enter(&vd->vdev_initialize_lock); uint64_t initialize_state = VDEV_INITIALIZE_NONE; int err = zap_lookup(vd->vdev_spa->spa_meta_objset, vd->vdev_leaf_zap, VDEV_LEAF_ZAP_INITIALIZE_STATE, sizeof (initialize_state), 1, &initialize_state); ASSERT(err == 0 || err == ENOENT); vd->vdev_initialize_state = initialize_state; uint64_t timestamp = 0; err = zap_lookup(vd->vdev_spa->spa_meta_objset, vd->vdev_leaf_zap, VDEV_LEAF_ZAP_INITIALIZE_ACTION_TIME, sizeof (timestamp), 1, ×tamp); ASSERT(err == 0 || err == ENOENT); vd->vdev_initialize_action_time = timestamp; if (vd->vdev_initialize_state == VDEV_INITIALIZE_SUSPENDED || vd->vdev_offline) { /* load progress for reporting, but don't resume */ VERIFY0(vdev_initialize_load(vd)); } else if (vd->vdev_initialize_state == VDEV_INITIALIZE_ACTIVE && vdev_writeable(vd) && !vd->vdev_top->vdev_removing && vd->vdev_initialize_thread == NULL) { vdev_initialize(vd); } mutex_exit(&vd->vdev_initialize_lock); } for (uint64_t i = 0; i < vd->vdev_children; i++) { vdev_initialize_restart(vd->vdev_child[i]); } } EXPORT_SYMBOL(vdev_initialize); +EXPORT_SYMBOL(vdev_uninitialize); EXPORT_SYMBOL(vdev_initialize_stop); EXPORT_SYMBOL(vdev_initialize_stop_all); EXPORT_SYMBOL(vdev_initialize_stop_wait); EXPORT_SYMBOL(vdev_initialize_restart); ZFS_MODULE_PARAM(zfs, zfs_, initialize_value, U64, ZMOD_RW, "Value written during zpool initialize"); ZFS_MODULE_PARAM(zfs, zfs_, initialize_chunk_size, U64, ZMOD_RW, "Size in bytes of writes by zpool initialize"); diff --git a/sys/contrib/openzfs/module/zfs/zfs_ioctl.c b/sys/contrib/openzfs/module/zfs/zfs_ioctl.c index 3b1e2ae5fb5d..f91a2f3bbca5 100644 --- a/sys/contrib/openzfs/module/zfs/zfs_ioctl.c +++ b/sys/contrib/openzfs/module/zfs/zfs_ioctl.c @@ -1,7883 +1,7930 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Portions Copyright 2011 Martin Matuska * Copyright 2015, OmniTI Computer Consulting, Inc. All rights reserved. * Copyright (c) 2012 Pawel Jakub Dawidek * Copyright (c) 2014, 2016 Joyent, Inc. All rights reserved. * Copyright 2016 Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2014, Joyent, Inc. All rights reserved. * Copyright (c) 2011, 2020 by Delphix. All rights reserved. * Copyright (c) 2013 by Saso Kiselkov. All rights reserved. * Copyright (c) 2013 Steven Hartland. All rights reserved. * Copyright (c) 2014 Integros [integros.com] * Copyright 2016 Toomas Soome * Copyright (c) 2016 Actifio, Inc. All rights reserved. * Copyright (c) 2018, loli10K . All rights reserved. * Copyright 2017 RackTop Systems. * Copyright (c) 2017 Open-E, Inc. All Rights Reserved. * Copyright (c) 2019 Datto Inc. * Copyright (c) 2019, 2020 by Christian Schwarz. All rights reserved. * Copyright (c) 2019, 2021, Klara Inc. * Copyright (c) 2019, Allan Jude */ /* * ZFS ioctls. * * This file handles the ioctls to /dev/zfs, used for configuring ZFS storage * pools and filesystems, e.g. with /sbin/zfs and /sbin/zpool. * * There are two ways that we handle ioctls: the legacy way where almost * all of the logic is in the ioctl callback, and the new way where most * of the marshalling is handled in the common entry point, zfsdev_ioctl(). * * Non-legacy ioctls should be registered by calling * zfs_ioctl_register() from zfs_ioctl_init(). The ioctl is invoked * from userland by lzc_ioctl(). * * The registration arguments are as follows: * * const char *name * The name of the ioctl. This is used for history logging. If the * ioctl returns successfully (the callback returns 0), and allow_log * is true, then a history log entry will be recorded with the input & * output nvlists. The log entry can be printed with "zpool history -i". * * zfs_ioc_t ioc * The ioctl request number, which userland will pass to ioctl(2). * We want newer versions of libzfs and libzfs_core to run against * existing zfs kernel modules (i.e. a deferred reboot after an update). * Therefore the ioctl numbers cannot change from release to release. * * zfs_secpolicy_func_t *secpolicy * This function will be called before the zfs_ioc_func_t, to * determine if this operation is permitted. It should return EPERM * on failure, and 0 on success. Checks include determining if the * dataset is visible in this zone, and if the user has either all * zfs privileges in the zone (SYS_MOUNT), or has been granted permission * to do this operation on this dataset with "zfs allow". * * zfs_ioc_namecheck_t namecheck * This specifies what to expect in the zfs_cmd_t:zc_name -- a pool * name, a dataset name, or nothing. If the name is not well-formed, * the ioctl will fail and the callback will not be called. * Therefore, the callback can assume that the name is well-formed * (e.g. is null-terminated, doesn't have more than one '@' character, * doesn't have invalid characters). * * zfs_ioc_poolcheck_t pool_check * This specifies requirements on the pool state. If the pool does * not meet them (is suspended or is readonly), the ioctl will fail * and the callback will not be called. If any checks are specified * (i.e. it is not POOL_CHECK_NONE), namecheck must not be NO_NAME. * Multiple checks can be or-ed together (e.g. POOL_CHECK_SUSPENDED | * POOL_CHECK_READONLY). * * zfs_ioc_key_t *nvl_keys * The list of expected/allowable innvl input keys. This list is used * to validate the nvlist input to the ioctl. * * boolean_t smush_outnvlist * If smush_outnvlist is true, then the output is presumed to be a * list of errors, and it will be "smushed" down to fit into the * caller's buffer, by removing some entries and replacing them with a * single "N_MORE_ERRORS" entry indicating how many were removed. See * nvlist_smush() for details. If smush_outnvlist is false, and the * outnvlist does not fit into the userland-provided buffer, then the * ioctl will fail with ENOMEM. * * zfs_ioc_func_t *func * The callback function that will perform the operation. * * The callback should return 0 on success, or an error number on * failure. If the function fails, the userland ioctl will return -1, * and errno will be set to the callback's return value. The callback * will be called with the following arguments: * * const char *name * The name of the pool or dataset to operate on, from * zfs_cmd_t:zc_name. The 'namecheck' argument specifies the * expected type (pool, dataset, or none). * * nvlist_t *innvl * The input nvlist, deserialized from zfs_cmd_t:zc_nvlist_src. Or * NULL if no input nvlist was provided. Changes to this nvlist are * ignored. If the input nvlist could not be deserialized, the * ioctl will fail and the callback will not be called. * * nvlist_t *outnvl * The output nvlist, initially empty. The callback can fill it in, * and it will be returned to userland by serializing it into * zfs_cmd_t:zc_nvlist_dst. If it is non-empty, and serialization * fails (e.g. because the caller didn't supply a large enough * buffer), then the overall ioctl will fail. See the * 'smush_nvlist' argument above for additional behaviors. * * There are two typical uses of the output nvlist: * - To return state, e.g. property values. In this case, * smush_outnvlist should be false. If the buffer was not large * enough, the caller will reallocate a larger buffer and try * the ioctl again. * * - To return multiple errors from an ioctl which makes on-disk * changes. In this case, smush_outnvlist should be true. * Ioctls which make on-disk modifications should generally not * use the outnvl if they succeed, because the caller can not * distinguish between the operation failing, and * deserialization failing. * * IOCTL Interface Errors * * The following ioctl input errors can be returned: * ZFS_ERR_IOC_CMD_UNAVAIL the ioctl number is not supported by kernel * ZFS_ERR_IOC_ARG_UNAVAIL an input argument is not supported by kernel * ZFS_ERR_IOC_ARG_REQUIRED a required input argument is missing * ZFS_ERR_IOC_ARG_BADTYPE an input argument has an invalid type */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "zfs_namecheck.h" #include "zfs_prop.h" #include "zfs_deleg.h" #include "zfs_comutil.h" #include #include #include kmutex_t zfsdev_state_lock; static zfsdev_state_t zfsdev_state_listhead; /* * Limit maximum nvlist size. We don't want users passing in insane values * for zc->zc_nvlist_src_size, since we will need to allocate that much memory. * Defaults to 0=auto which is handled by platform code. */ uint64_t zfs_max_nvlist_src_size = 0; /* * When logging the output nvlist of an ioctl in the on-disk history, limit * the logged size to this many bytes. This must be less than DMU_MAX_ACCESS. * This applies primarily to zfs_ioc_channel_program(). */ static uint64_t zfs_history_output_max = 1024 * 1024; uint_t zfs_fsyncer_key; uint_t zfs_allow_log_key; /* DATA_TYPE_ANY is used when zkey_type can vary. */ #define DATA_TYPE_ANY DATA_TYPE_UNKNOWN typedef struct zfs_ioc_vec { zfs_ioc_legacy_func_t *zvec_legacy_func; zfs_ioc_func_t *zvec_func; zfs_secpolicy_func_t *zvec_secpolicy; zfs_ioc_namecheck_t zvec_namecheck; boolean_t zvec_allow_log; zfs_ioc_poolcheck_t zvec_pool_check; boolean_t zvec_smush_outnvlist; const char *zvec_name; const zfs_ioc_key_t *zvec_nvl_keys; size_t zvec_nvl_key_count; } zfs_ioc_vec_t; /* This array is indexed by zfs_userquota_prop_t */ static const char *userquota_perms[] = { ZFS_DELEG_PERM_USERUSED, ZFS_DELEG_PERM_USERQUOTA, ZFS_DELEG_PERM_GROUPUSED, ZFS_DELEG_PERM_GROUPQUOTA, ZFS_DELEG_PERM_USEROBJUSED, ZFS_DELEG_PERM_USEROBJQUOTA, ZFS_DELEG_PERM_GROUPOBJUSED, ZFS_DELEG_PERM_GROUPOBJQUOTA, ZFS_DELEG_PERM_PROJECTUSED, ZFS_DELEG_PERM_PROJECTQUOTA, ZFS_DELEG_PERM_PROJECTOBJUSED, ZFS_DELEG_PERM_PROJECTOBJQUOTA, }; static int zfs_ioc_userspace_upgrade(zfs_cmd_t *zc); static int zfs_ioc_id_quota_upgrade(zfs_cmd_t *zc); static int zfs_check_settable(const char *name, nvpair_t *property, cred_t *cr); static int zfs_check_clearable(const char *dataset, nvlist_t *props, nvlist_t **errors); static int zfs_fill_zplprops_root(uint64_t, nvlist_t *, nvlist_t *, boolean_t *); int zfs_set_prop_nvlist(const char *, zprop_source_t, nvlist_t *, nvlist_t *); static int get_nvlist(uint64_t nvl, uint64_t size, int iflag, nvlist_t **nvp); static void history_str_free(char *buf) { kmem_free(buf, HIS_MAX_RECORD_LEN); } static char * history_str_get(zfs_cmd_t *zc) { char *buf; if (zc->zc_history == 0) return (NULL); buf = kmem_alloc(HIS_MAX_RECORD_LEN, KM_SLEEP); if (copyinstr((void *)(uintptr_t)zc->zc_history, buf, HIS_MAX_RECORD_LEN, NULL) != 0) { history_str_free(buf); return (NULL); } buf[HIS_MAX_RECORD_LEN -1] = '\0'; return (buf); } /* * Return non-zero if the spa version is less than requested version. */ static int zfs_earlier_version(const char *name, int version) { spa_t *spa; if (spa_open(name, &spa, FTAG) == 0) { if (spa_version(spa) < version) { spa_close(spa, FTAG); return (1); } spa_close(spa, FTAG); } return (0); } /* * Return TRUE if the ZPL version is less than requested version. */ static boolean_t zpl_earlier_version(const char *name, int version) { objset_t *os; boolean_t rc = B_TRUE; if (dmu_objset_hold(name, FTAG, &os) == 0) { uint64_t zplversion; if (dmu_objset_type(os) != DMU_OST_ZFS) { dmu_objset_rele(os, FTAG); return (B_TRUE); } /* XXX reading from non-owned objset */ if (zfs_get_zplprop(os, ZFS_PROP_VERSION, &zplversion) == 0) rc = zplversion < version; dmu_objset_rele(os, FTAG); } return (rc); } static void zfs_log_history(zfs_cmd_t *zc) { spa_t *spa; char *buf; if ((buf = history_str_get(zc)) == NULL) return; if (spa_open(zc->zc_name, &spa, FTAG) == 0) { if (spa_version(spa) >= SPA_VERSION_ZPOOL_HISTORY) (void) spa_history_log(spa, buf); spa_close(spa, FTAG); } history_str_free(buf); } /* * Policy for top-level read operations (list pools). Requires no privileges, * and can be used in the local zone, as there is no associated dataset. */ static int zfs_secpolicy_none(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) zc, (void) innvl, (void) cr; return (0); } /* * Policy for dataset read operations (list children, get statistics). Requires * no privileges, but must be visible in the local zone. */ static int zfs_secpolicy_read(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) innvl, (void) cr; if (INGLOBALZONE(curproc) || zone_dataset_visible(zc->zc_name, NULL)) return (0); return (SET_ERROR(ENOENT)); } static int zfs_dozonecheck_impl(const char *dataset, uint64_t zoned, cred_t *cr) { int writable = 1; /* * The dataset must be visible by this zone -- check this first * so they don't see EPERM on something they shouldn't know about. */ if (!INGLOBALZONE(curproc) && !zone_dataset_visible(dataset, &writable)) return (SET_ERROR(ENOENT)); if (INGLOBALZONE(curproc)) { /* * If the fs is zoned, only root can access it from the * global zone. */ if (secpolicy_zfs(cr) && zoned) return (SET_ERROR(EPERM)); } else { /* * If we are in a local zone, the 'zoned' property must be set. */ if (!zoned) return (SET_ERROR(EPERM)); /* must be writable by this zone */ if (!writable) return (SET_ERROR(EPERM)); } return (0); } static int zfs_dozonecheck(const char *dataset, cred_t *cr) { uint64_t zoned; if (dsl_prop_get_integer(dataset, zfs_prop_to_name(ZFS_PROP_ZONED), &zoned, NULL)) return (SET_ERROR(ENOENT)); return (zfs_dozonecheck_impl(dataset, zoned, cr)); } static int zfs_dozonecheck_ds(const char *dataset, dsl_dataset_t *ds, cred_t *cr) { uint64_t zoned; if (dsl_prop_get_int_ds(ds, zfs_prop_to_name(ZFS_PROP_ZONED), &zoned)) return (SET_ERROR(ENOENT)); return (zfs_dozonecheck_impl(dataset, zoned, cr)); } static int zfs_secpolicy_write_perms_ds(const char *name, dsl_dataset_t *ds, const char *perm, cred_t *cr) { int error; error = zfs_dozonecheck_ds(name, ds, cr); if (error == 0) { error = secpolicy_zfs(cr); if (error != 0) error = dsl_deleg_access_impl(ds, perm, cr); } return (error); } static int zfs_secpolicy_write_perms(const char *name, const char *perm, cred_t *cr) { int error; dsl_dataset_t *ds; dsl_pool_t *dp; /* * First do a quick check for root in the global zone, which * is allowed to do all write_perms. This ensures that zfs_ioc_* * will get to handle nonexistent datasets. */ if (INGLOBALZONE(curproc) && secpolicy_zfs(cr) == 0) return (0); error = dsl_pool_hold(name, FTAG, &dp); if (error != 0) return (error); error = dsl_dataset_hold(dp, name, FTAG, &ds); if (error != 0) { dsl_pool_rele(dp, FTAG); return (error); } error = zfs_secpolicy_write_perms_ds(name, ds, perm, cr); dsl_dataset_rele(ds, FTAG); dsl_pool_rele(dp, FTAG); return (error); } /* * Policy for setting the security label property. * * Returns 0 for success, non-zero for access and other errors. */ static int zfs_set_slabel_policy(const char *name, const char *strval, cred_t *cr) { #ifdef HAVE_MLSLABEL char ds_hexsl[MAXNAMELEN]; bslabel_t ds_sl, new_sl; boolean_t new_default = FALSE; uint64_t zoned; int needed_priv = -1; int error; /* First get the existing dataset label. */ error = dsl_prop_get(name, zfs_prop_to_name(ZFS_PROP_MLSLABEL), 1, sizeof (ds_hexsl), &ds_hexsl, NULL); if (error != 0) return (SET_ERROR(EPERM)); if (strcasecmp(strval, ZFS_MLSLABEL_DEFAULT) == 0) new_default = TRUE; /* The label must be translatable */ if (!new_default && (hexstr_to_label(strval, &new_sl) != 0)) return (SET_ERROR(EINVAL)); /* * In a non-global zone, disallow attempts to set a label that * doesn't match that of the zone; otherwise no other checks * are needed. */ if (!INGLOBALZONE(curproc)) { if (new_default || !blequal(&new_sl, CR_SL(CRED()))) return (SET_ERROR(EPERM)); return (0); } /* * For global-zone datasets (i.e., those whose zoned property is * "off", verify that the specified new label is valid for the * global zone. */ if (dsl_prop_get_integer(name, zfs_prop_to_name(ZFS_PROP_ZONED), &zoned, NULL)) return (SET_ERROR(EPERM)); if (!zoned) { if (zfs_check_global_label(name, strval) != 0) return (SET_ERROR(EPERM)); } /* * If the existing dataset label is nondefault, check if the * dataset is mounted (label cannot be changed while mounted). * Get the zfsvfs_t; if there isn't one, then the dataset isn't * mounted (or isn't a dataset, doesn't exist, ...). */ if (strcasecmp(ds_hexsl, ZFS_MLSLABEL_DEFAULT) != 0) { objset_t *os; static const char *setsl_tag = "setsl_tag"; /* * Try to own the dataset; abort if there is any error, * (e.g., already mounted, in use, or other error). */ error = dmu_objset_own(name, DMU_OST_ZFS, B_TRUE, B_TRUE, setsl_tag, &os); if (error != 0) return (SET_ERROR(EPERM)); dmu_objset_disown(os, B_TRUE, setsl_tag); if (new_default) { needed_priv = PRIV_FILE_DOWNGRADE_SL; goto out_check; } if (hexstr_to_label(strval, &new_sl) != 0) return (SET_ERROR(EPERM)); if (blstrictdom(&ds_sl, &new_sl)) needed_priv = PRIV_FILE_DOWNGRADE_SL; else if (blstrictdom(&new_sl, &ds_sl)) needed_priv = PRIV_FILE_UPGRADE_SL; } else { /* dataset currently has a default label */ if (!new_default) needed_priv = PRIV_FILE_UPGRADE_SL; } out_check: if (needed_priv != -1) return (PRIV_POLICY(cr, needed_priv, B_FALSE, EPERM, NULL)); return (0); #else return (SET_ERROR(ENOTSUP)); #endif /* HAVE_MLSLABEL */ } static int zfs_secpolicy_setprop(const char *dsname, zfs_prop_t prop, nvpair_t *propval, cred_t *cr) { const char *strval; /* * Check permissions for special properties. */ switch (prop) { default: break; case ZFS_PROP_ZONED: /* * Disallow setting of 'zoned' from within a local zone. */ if (!INGLOBALZONE(curproc)) return (SET_ERROR(EPERM)); break; case ZFS_PROP_QUOTA: case ZFS_PROP_FILESYSTEM_LIMIT: case ZFS_PROP_SNAPSHOT_LIMIT: if (!INGLOBALZONE(curproc)) { uint64_t zoned; char setpoint[ZFS_MAX_DATASET_NAME_LEN]; /* * Unprivileged users are allowed to modify the * limit on things *under* (ie. contained by) * the thing they own. */ if (dsl_prop_get_integer(dsname, zfs_prop_to_name(ZFS_PROP_ZONED), &zoned, setpoint)) return (SET_ERROR(EPERM)); if (!zoned || strlen(dsname) <= strlen(setpoint)) return (SET_ERROR(EPERM)); } break; case ZFS_PROP_MLSLABEL: if (!is_system_labeled()) return (SET_ERROR(EPERM)); if (nvpair_value_string(propval, &strval) == 0) { int err; err = zfs_set_slabel_policy(dsname, strval, CRED()); if (err != 0) return (err); } break; } return (zfs_secpolicy_write_perms(dsname, zfs_prop_to_name(prop), cr)); } static int zfs_secpolicy_set_fsacl(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { /* * permission to set permissions will be evaluated later in * dsl_deleg_can_allow() */ (void) innvl; return (zfs_dozonecheck(zc->zc_name, cr)); } static int zfs_secpolicy_rollback(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) innvl; return (zfs_secpolicy_write_perms(zc->zc_name, ZFS_DELEG_PERM_ROLLBACK, cr)); } static int zfs_secpolicy_send(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) innvl; dsl_pool_t *dp; dsl_dataset_t *ds; const char *cp; int error; /* * Generate the current snapshot name from the given objsetid, then * use that name for the secpolicy/zone checks. */ cp = strchr(zc->zc_name, '@'); if (cp == NULL) return (SET_ERROR(EINVAL)); error = dsl_pool_hold(zc->zc_name, FTAG, &dp); if (error != 0) return (error); error = dsl_dataset_hold_obj(dp, zc->zc_sendobj, FTAG, &ds); if (error != 0) { dsl_pool_rele(dp, FTAG); return (error); } dsl_dataset_name(ds, zc->zc_name); error = zfs_secpolicy_write_perms_ds(zc->zc_name, ds, ZFS_DELEG_PERM_SEND, cr); dsl_dataset_rele(ds, FTAG); dsl_pool_rele(dp, FTAG); return (error); } static int zfs_secpolicy_send_new(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) innvl; return (zfs_secpolicy_write_perms(zc->zc_name, ZFS_DELEG_PERM_SEND, cr)); } static int zfs_secpolicy_share(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) zc, (void) innvl, (void) cr; return (SET_ERROR(ENOTSUP)); } static int zfs_secpolicy_smb_acl(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) zc, (void) innvl, (void) cr; return (SET_ERROR(ENOTSUP)); } static int zfs_get_parent(const char *datasetname, char *parent, int parentsize) { char *cp; /* * Remove the @bla or /bla from the end of the name to get the parent. */ (void) strlcpy(parent, datasetname, parentsize); cp = strrchr(parent, '@'); if (cp != NULL) { cp[0] = '\0'; } else { cp = strrchr(parent, '/'); if (cp == NULL) return (SET_ERROR(ENOENT)); cp[0] = '\0'; } return (0); } int zfs_secpolicy_destroy_perms(const char *name, cred_t *cr) { int error; if ((error = zfs_secpolicy_write_perms(name, ZFS_DELEG_PERM_MOUNT, cr)) != 0) return (error); return (zfs_secpolicy_write_perms(name, ZFS_DELEG_PERM_DESTROY, cr)); } static int zfs_secpolicy_destroy(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) innvl; return (zfs_secpolicy_destroy_perms(zc->zc_name, cr)); } /* * Destroying snapshots with delegated permissions requires * descendant mount and destroy permissions. */ static int zfs_secpolicy_destroy_snaps(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) zc; nvlist_t *snaps; nvpair_t *pair, *nextpair; int error = 0; snaps = fnvlist_lookup_nvlist(innvl, "snaps"); for (pair = nvlist_next_nvpair(snaps, NULL); pair != NULL; pair = nextpair) { nextpair = nvlist_next_nvpair(snaps, pair); error = zfs_secpolicy_destroy_perms(nvpair_name(pair), cr); if (error == ENOENT) { /* * Ignore any snapshots that don't exist (we consider * them "already destroyed"). Remove the name from the * nvl here in case the snapshot is created between * now and when we try to destroy it (in which case * we don't want to destroy it since we haven't * checked for permission). */ fnvlist_remove_nvpair(snaps, pair); error = 0; } if (error != 0) break; } return (error); } int zfs_secpolicy_rename_perms(const char *from, const char *to, cred_t *cr) { char parentname[ZFS_MAX_DATASET_NAME_LEN]; int error; if ((error = zfs_secpolicy_write_perms(from, ZFS_DELEG_PERM_RENAME, cr)) != 0) return (error); if ((error = zfs_secpolicy_write_perms(from, ZFS_DELEG_PERM_MOUNT, cr)) != 0) return (error); if ((error = zfs_get_parent(to, parentname, sizeof (parentname))) != 0) return (error); if ((error = zfs_secpolicy_write_perms(parentname, ZFS_DELEG_PERM_CREATE, cr)) != 0) return (error); if ((error = zfs_secpolicy_write_perms(parentname, ZFS_DELEG_PERM_MOUNT, cr)) != 0) return (error); return (error); } static int zfs_secpolicy_rename(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) innvl; return (zfs_secpolicy_rename_perms(zc->zc_name, zc->zc_value, cr)); } static int zfs_secpolicy_promote(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) innvl; dsl_pool_t *dp; dsl_dataset_t *clone; int error; error = zfs_secpolicy_write_perms(zc->zc_name, ZFS_DELEG_PERM_PROMOTE, cr); if (error != 0) return (error); error = dsl_pool_hold(zc->zc_name, FTAG, &dp); if (error != 0) return (error); error = dsl_dataset_hold(dp, zc->zc_name, FTAG, &clone); if (error == 0) { char parentname[ZFS_MAX_DATASET_NAME_LEN]; dsl_dataset_t *origin = NULL; dsl_dir_t *dd; dd = clone->ds_dir; error = dsl_dataset_hold_obj(dd->dd_pool, dsl_dir_phys(dd)->dd_origin_obj, FTAG, &origin); if (error != 0) { dsl_dataset_rele(clone, FTAG); dsl_pool_rele(dp, FTAG); return (error); } error = zfs_secpolicy_write_perms_ds(zc->zc_name, clone, ZFS_DELEG_PERM_MOUNT, cr); dsl_dataset_name(origin, parentname); if (error == 0) { error = zfs_secpolicy_write_perms_ds(parentname, origin, ZFS_DELEG_PERM_PROMOTE, cr); } dsl_dataset_rele(clone, FTAG); dsl_dataset_rele(origin, FTAG); } dsl_pool_rele(dp, FTAG); return (error); } static int zfs_secpolicy_recv(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) innvl; int error; if ((error = zfs_secpolicy_write_perms(zc->zc_name, ZFS_DELEG_PERM_RECEIVE, cr)) != 0) return (error); if ((error = zfs_secpolicy_write_perms(zc->zc_name, ZFS_DELEG_PERM_MOUNT, cr)) != 0) return (error); return (zfs_secpolicy_write_perms(zc->zc_name, ZFS_DELEG_PERM_CREATE, cr)); } int zfs_secpolicy_snapshot_perms(const char *name, cred_t *cr) { return (zfs_secpolicy_write_perms(name, ZFS_DELEG_PERM_SNAPSHOT, cr)); } /* * Check for permission to create each snapshot in the nvlist. */ static int zfs_secpolicy_snapshot(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) zc; nvlist_t *snaps; int error = 0; nvpair_t *pair; snaps = fnvlist_lookup_nvlist(innvl, "snaps"); for (pair = nvlist_next_nvpair(snaps, NULL); pair != NULL; pair = nvlist_next_nvpair(snaps, pair)) { char *name = (char *)nvpair_name(pair); char *atp = strchr(name, '@'); if (atp == NULL) { error = SET_ERROR(EINVAL); break; } *atp = '\0'; error = zfs_secpolicy_snapshot_perms(name, cr); *atp = '@'; if (error != 0) break; } return (error); } /* * Check for permission to create each bookmark in the nvlist. */ static int zfs_secpolicy_bookmark(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) zc; int error = 0; for (nvpair_t *pair = nvlist_next_nvpair(innvl, NULL); pair != NULL; pair = nvlist_next_nvpair(innvl, pair)) { char *name = (char *)nvpair_name(pair); char *hashp = strchr(name, '#'); if (hashp == NULL) { error = SET_ERROR(EINVAL); break; } *hashp = '\0'; error = zfs_secpolicy_write_perms(name, ZFS_DELEG_PERM_BOOKMARK, cr); *hashp = '#'; if (error != 0) break; } return (error); } static int zfs_secpolicy_destroy_bookmarks(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) zc; nvpair_t *pair, *nextpair; int error = 0; for (pair = nvlist_next_nvpair(innvl, NULL); pair != NULL; pair = nextpair) { char *name = (char *)nvpair_name(pair); char *hashp = strchr(name, '#'); nextpair = nvlist_next_nvpair(innvl, pair); if (hashp == NULL) { error = SET_ERROR(EINVAL); break; } *hashp = '\0'; error = zfs_secpolicy_write_perms(name, ZFS_DELEG_PERM_DESTROY, cr); *hashp = '#'; if (error == ENOENT) { /* * Ignore any filesystems that don't exist (we consider * their bookmarks "already destroyed"). Remove * the name from the nvl here in case the filesystem * is created between now and when we try to destroy * the bookmark (in which case we don't want to * destroy it since we haven't checked for permission). */ fnvlist_remove_nvpair(innvl, pair); error = 0; } if (error != 0) break; } return (error); } static int zfs_secpolicy_log_history(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) zc, (void) innvl, (void) cr; /* * Even root must have a proper TSD so that we know what pool * to log to. */ if (tsd_get(zfs_allow_log_key) == NULL) return (SET_ERROR(EPERM)); return (0); } static int zfs_secpolicy_create_clone(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { char parentname[ZFS_MAX_DATASET_NAME_LEN]; int error; const char *origin; if ((error = zfs_get_parent(zc->zc_name, parentname, sizeof (parentname))) != 0) return (error); if (nvlist_lookup_string(innvl, "origin", &origin) == 0 && (error = zfs_secpolicy_write_perms(origin, ZFS_DELEG_PERM_CLONE, cr)) != 0) return (error); if ((error = zfs_secpolicy_write_perms(parentname, ZFS_DELEG_PERM_CREATE, cr)) != 0) return (error); return (zfs_secpolicy_write_perms(parentname, ZFS_DELEG_PERM_MOUNT, cr)); } /* * Policy for pool operations - create/destroy pools, add vdevs, etc. Requires * SYS_CONFIG privilege, which is not available in a local zone. */ int zfs_secpolicy_config(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) zc, (void) innvl; if (secpolicy_sys_config(cr, B_FALSE) != 0) return (SET_ERROR(EPERM)); return (0); } /* * Policy for object to name lookups. */ static int zfs_secpolicy_diff(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) innvl; int error; if (secpolicy_sys_config(cr, B_FALSE) == 0) return (0); error = zfs_secpolicy_write_perms(zc->zc_name, ZFS_DELEG_PERM_DIFF, cr); return (error); } /* * Policy for fault injection. Requires all privileges. */ static int zfs_secpolicy_inject(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) zc, (void) innvl; return (secpolicy_zinject(cr)); } static int zfs_secpolicy_inherit_prop(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) innvl; zfs_prop_t prop = zfs_name_to_prop(zc->zc_value); if (prop == ZPROP_USERPROP) { if (!zfs_prop_user(zc->zc_value)) return (SET_ERROR(EINVAL)); return (zfs_secpolicy_write_perms(zc->zc_name, ZFS_DELEG_PERM_USERPROP, cr)); } else { return (zfs_secpolicy_setprop(zc->zc_name, prop, NULL, cr)); } } static int zfs_secpolicy_userspace_one(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { int err = zfs_secpolicy_read(zc, innvl, cr); if (err) return (err); if (zc->zc_objset_type >= ZFS_NUM_USERQUOTA_PROPS) return (SET_ERROR(EINVAL)); if (zc->zc_value[0] == 0) { /* * They are asking about a posix uid/gid. If it's * themself, allow it. */ if (zc->zc_objset_type == ZFS_PROP_USERUSED || zc->zc_objset_type == ZFS_PROP_USERQUOTA || zc->zc_objset_type == ZFS_PROP_USEROBJUSED || zc->zc_objset_type == ZFS_PROP_USEROBJQUOTA) { if (zc->zc_guid == crgetuid(cr)) return (0); } else if (zc->zc_objset_type == ZFS_PROP_GROUPUSED || zc->zc_objset_type == ZFS_PROP_GROUPQUOTA || zc->zc_objset_type == ZFS_PROP_GROUPOBJUSED || zc->zc_objset_type == ZFS_PROP_GROUPOBJQUOTA) { if (groupmember(zc->zc_guid, cr)) return (0); } /* else is for project quota/used */ } return (zfs_secpolicy_write_perms(zc->zc_name, userquota_perms[zc->zc_objset_type], cr)); } static int zfs_secpolicy_userspace_many(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { int err = zfs_secpolicy_read(zc, innvl, cr); if (err) return (err); if (zc->zc_objset_type >= ZFS_NUM_USERQUOTA_PROPS) return (SET_ERROR(EINVAL)); return (zfs_secpolicy_write_perms(zc->zc_name, userquota_perms[zc->zc_objset_type], cr)); } static int zfs_secpolicy_userspace_upgrade(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) innvl; return (zfs_secpolicy_setprop(zc->zc_name, ZFS_PROP_VERSION, NULL, cr)); } static int zfs_secpolicy_hold(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) zc; nvpair_t *pair; nvlist_t *holds; int error; holds = fnvlist_lookup_nvlist(innvl, "holds"); for (pair = nvlist_next_nvpair(holds, NULL); pair != NULL; pair = nvlist_next_nvpair(holds, pair)) { char fsname[ZFS_MAX_DATASET_NAME_LEN]; error = dmu_fsname(nvpair_name(pair), fsname); if (error != 0) return (error); error = zfs_secpolicy_write_perms(fsname, ZFS_DELEG_PERM_HOLD, cr); if (error != 0) return (error); } return (0); } static int zfs_secpolicy_release(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { (void) zc; nvpair_t *pair; int error; for (pair = nvlist_next_nvpair(innvl, NULL); pair != NULL; pair = nvlist_next_nvpair(innvl, pair)) { char fsname[ZFS_MAX_DATASET_NAME_LEN]; error = dmu_fsname(nvpair_name(pair), fsname); if (error != 0) return (error); error = zfs_secpolicy_write_perms(fsname, ZFS_DELEG_PERM_RELEASE, cr); if (error != 0) return (error); } return (0); } /* * Policy for allowing temporary snapshots to be taken or released */ static int zfs_secpolicy_tmp_snapshot(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { /* * A temporary snapshot is the same as a snapshot, * hold, destroy and release all rolled into one. * Delegated diff alone is sufficient that we allow this. */ int error; if (zfs_secpolicy_write_perms(zc->zc_name, ZFS_DELEG_PERM_DIFF, cr) == 0) return (0); error = zfs_secpolicy_snapshot_perms(zc->zc_name, cr); if (innvl != NULL) { if (error == 0) error = zfs_secpolicy_hold(zc, innvl, cr); if (error == 0) error = zfs_secpolicy_release(zc, innvl, cr); if (error == 0) error = zfs_secpolicy_destroy(zc, innvl, cr); } return (error); } static int zfs_secpolicy_load_key(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { return (zfs_secpolicy_write_perms(zc->zc_name, ZFS_DELEG_PERM_LOAD_KEY, cr)); } static int zfs_secpolicy_change_key(zfs_cmd_t *zc, nvlist_t *innvl, cred_t *cr) { return (zfs_secpolicy_write_perms(zc->zc_name, ZFS_DELEG_PERM_CHANGE_KEY, cr)); } /* * Returns the nvlist as specified by the user in the zfs_cmd_t. */ static int get_nvlist(uint64_t nvl, uint64_t size, int iflag, nvlist_t **nvp) { char *packed; int error; nvlist_t *list = NULL; /* * Read in and unpack the user-supplied nvlist. */ if (size == 0) return (SET_ERROR(EINVAL)); packed = vmem_alloc(size, KM_SLEEP); if (ddi_copyin((void *)(uintptr_t)nvl, packed, size, iflag) != 0) { vmem_free(packed, size); return (SET_ERROR(EFAULT)); } if ((error = nvlist_unpack(packed, size, &list, 0)) != 0) { vmem_free(packed, size); return (error); } vmem_free(packed, size); *nvp = list; return (0); } /* * Reduce the size of this nvlist until it can be serialized in 'max' bytes. * Entries will be removed from the end of the nvlist, and one int32 entry * named "N_MORE_ERRORS" will be added indicating how many entries were * removed. */ static int nvlist_smush(nvlist_t *errors, size_t max) { size_t size; size = fnvlist_size(errors); if (size > max) { nvpair_t *more_errors; int n = 0; if (max < 1024) return (SET_ERROR(ENOMEM)); fnvlist_add_int32(errors, ZPROP_N_MORE_ERRORS, 0); more_errors = nvlist_prev_nvpair(errors, NULL); do { nvpair_t *pair = nvlist_prev_nvpair(errors, more_errors); fnvlist_remove_nvpair(errors, pair); n++; size = fnvlist_size(errors); } while (size > max); fnvlist_remove_nvpair(errors, more_errors); fnvlist_add_int32(errors, ZPROP_N_MORE_ERRORS, n); ASSERT3U(fnvlist_size(errors), <=, max); } return (0); } static int put_nvlist(zfs_cmd_t *zc, nvlist_t *nvl) { char *packed = NULL; int error = 0; size_t size; size = fnvlist_size(nvl); if (size > zc->zc_nvlist_dst_size) { error = SET_ERROR(ENOMEM); } else { packed = fnvlist_pack(nvl, &size); if (ddi_copyout(packed, (void *)(uintptr_t)zc->zc_nvlist_dst, size, zc->zc_iflags) != 0) error = SET_ERROR(EFAULT); fnvlist_pack_free(packed, size); } zc->zc_nvlist_dst_size = size; zc->zc_nvlist_dst_filled = B_TRUE; return (error); } int getzfsvfs_impl(objset_t *os, zfsvfs_t **zfvp) { int error = 0; if (dmu_objset_type(os) != DMU_OST_ZFS) { return (SET_ERROR(EINVAL)); } mutex_enter(&os->os_user_ptr_lock); *zfvp = dmu_objset_get_user(os); /* bump s_active only when non-zero to prevent umount race */ error = zfs_vfs_ref(zfvp); mutex_exit(&os->os_user_ptr_lock); return (error); } int getzfsvfs(const char *dsname, zfsvfs_t **zfvp) { objset_t *os; int error; error = dmu_objset_hold(dsname, FTAG, &os); if (error != 0) return (error); error = getzfsvfs_impl(os, zfvp); dmu_objset_rele(os, FTAG); return (error); } /* * Find a zfsvfs_t for a mounted filesystem, or create our own, in which * case its z_sb will be NULL, and it will be opened as the owner. * If 'writer' is set, the z_teardown_lock will be held for RW_WRITER, * which prevents all inode ops from running. */ static int zfsvfs_hold(const char *name, const void *tag, zfsvfs_t **zfvp, boolean_t writer) { int error = 0; if (getzfsvfs(name, zfvp) != 0) error = zfsvfs_create(name, B_FALSE, zfvp); if (error == 0) { if (writer) ZFS_TEARDOWN_ENTER_WRITE(*zfvp, tag); else ZFS_TEARDOWN_ENTER_READ(*zfvp, tag); if ((*zfvp)->z_unmounted) { /* * XXX we could probably try again, since the unmounting * thread should be just about to disassociate the * objset from the zfsvfs. */ ZFS_TEARDOWN_EXIT(*zfvp, tag); return (SET_ERROR(EBUSY)); } } return (error); } static void zfsvfs_rele(zfsvfs_t *zfsvfs, const void *tag) { ZFS_TEARDOWN_EXIT(zfsvfs, tag); if (zfs_vfs_held(zfsvfs)) { zfs_vfs_rele(zfsvfs); } else { dmu_objset_disown(zfsvfs->z_os, B_TRUE, zfsvfs); zfsvfs_free(zfsvfs); } } static int zfs_ioc_pool_create(zfs_cmd_t *zc) { int error; nvlist_t *config, *props = NULL; nvlist_t *rootprops = NULL; nvlist_t *zplprops = NULL; dsl_crypto_params_t *dcp = NULL; const char *spa_name = zc->zc_name; boolean_t unload_wkey = B_TRUE; if ((error = get_nvlist(zc->zc_nvlist_conf, zc->zc_nvlist_conf_size, zc->zc_iflags, &config))) return (error); if (zc->zc_nvlist_src_size != 0 && (error = get_nvlist(zc->zc_nvlist_src, zc->zc_nvlist_src_size, zc->zc_iflags, &props))) { nvlist_free(config); return (error); } if (props) { nvlist_t *nvl = NULL; nvlist_t *hidden_args = NULL; uint64_t version = SPA_VERSION; const char *tname; (void) nvlist_lookup_uint64(props, zpool_prop_to_name(ZPOOL_PROP_VERSION), &version); if (!SPA_VERSION_IS_SUPPORTED(version)) { error = SET_ERROR(EINVAL); goto pool_props_bad; } (void) nvlist_lookup_nvlist(props, ZPOOL_ROOTFS_PROPS, &nvl); if (nvl) { error = nvlist_dup(nvl, &rootprops, KM_SLEEP); if (error != 0) goto pool_props_bad; (void) nvlist_remove_all(props, ZPOOL_ROOTFS_PROPS); } (void) nvlist_lookup_nvlist(props, ZPOOL_HIDDEN_ARGS, &hidden_args); error = dsl_crypto_params_create_nvlist(DCP_CMD_NONE, rootprops, hidden_args, &dcp); if (error != 0) goto pool_props_bad; (void) nvlist_remove_all(props, ZPOOL_HIDDEN_ARGS); VERIFY(nvlist_alloc(&zplprops, NV_UNIQUE_NAME, KM_SLEEP) == 0); error = zfs_fill_zplprops_root(version, rootprops, zplprops, NULL); if (error != 0) goto pool_props_bad; if (nvlist_lookup_string(props, zpool_prop_to_name(ZPOOL_PROP_TNAME), &tname) == 0) spa_name = tname; } error = spa_create(zc->zc_name, config, props, zplprops, dcp); /* * Set the remaining root properties */ if (!error && (error = zfs_set_prop_nvlist(spa_name, ZPROP_SRC_LOCAL, rootprops, NULL)) != 0) { (void) spa_destroy(spa_name); unload_wkey = B_FALSE; /* spa_destroy() unloads wrapping keys */ } pool_props_bad: nvlist_free(rootprops); nvlist_free(zplprops); nvlist_free(config); nvlist_free(props); dsl_crypto_params_free(dcp, unload_wkey && !!error); return (error); } static int zfs_ioc_pool_destroy(zfs_cmd_t *zc) { int error; zfs_log_history(zc); error = spa_destroy(zc->zc_name); return (error); } static int zfs_ioc_pool_import(zfs_cmd_t *zc) { nvlist_t *config, *props = NULL; uint64_t guid; int error; if ((error = get_nvlist(zc->zc_nvlist_conf, zc->zc_nvlist_conf_size, zc->zc_iflags, &config)) != 0) return (error); if (zc->zc_nvlist_src_size != 0 && (error = get_nvlist(zc->zc_nvlist_src, zc->zc_nvlist_src_size, zc->zc_iflags, &props))) { nvlist_free(config); return (error); } if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID, &guid) != 0 || guid != zc->zc_guid) error = SET_ERROR(EINVAL); else error = spa_import(zc->zc_name, config, props, zc->zc_cookie); if (zc->zc_nvlist_dst != 0) { int err; if ((err = put_nvlist(zc, config)) != 0) error = err; } nvlist_free(config); nvlist_free(props); return (error); } static int zfs_ioc_pool_export(zfs_cmd_t *zc) { int error; boolean_t force = (boolean_t)zc->zc_cookie; boolean_t hardforce = (boolean_t)zc->zc_guid; zfs_log_history(zc); error = spa_export(zc->zc_name, NULL, force, hardforce); return (error); } static int zfs_ioc_pool_configs(zfs_cmd_t *zc) { nvlist_t *configs; int error; if ((configs = spa_all_configs(&zc->zc_cookie)) == NULL) return (SET_ERROR(EEXIST)); error = put_nvlist(zc, configs); nvlist_free(configs); return (error); } /* * inputs: * zc_name name of the pool * * outputs: * zc_cookie real errno * zc_nvlist_dst config nvlist * zc_nvlist_dst_size size of config nvlist */ static int zfs_ioc_pool_stats(zfs_cmd_t *zc) { nvlist_t *config; int error; int ret = 0; error = spa_get_stats(zc->zc_name, &config, zc->zc_value, sizeof (zc->zc_value)); if (config != NULL) { ret = put_nvlist(zc, config); nvlist_free(config); /* * The config may be present even if 'error' is non-zero. * In this case we return success, and preserve the real errno * in 'zc_cookie'. */ zc->zc_cookie = error; } else { ret = error; } return (ret); } /* * Try to import the given pool, returning pool stats as appropriate so that * user land knows which devices are available and overall pool health. */ static int zfs_ioc_pool_tryimport(zfs_cmd_t *zc) { nvlist_t *tryconfig, *config = NULL; int error; if ((error = get_nvlist(zc->zc_nvlist_conf, zc->zc_nvlist_conf_size, zc->zc_iflags, &tryconfig)) != 0) return (error); config = spa_tryimport(tryconfig); nvlist_free(tryconfig); if (config == NULL) return (SET_ERROR(EINVAL)); error = put_nvlist(zc, config); nvlist_free(config); return (error); } /* * inputs: * zc_name name of the pool * zc_cookie scan func (pool_scan_func_t) * zc_flags scrub pause/resume flag (pool_scrub_cmd_t) */ static int zfs_ioc_pool_scan(zfs_cmd_t *zc) { spa_t *spa; int error; if (zc->zc_flags >= POOL_SCRUB_FLAGS_END) return (SET_ERROR(EINVAL)); if ((error = spa_open(zc->zc_name, &spa, FTAG)) != 0) return (error); if (zc->zc_flags == POOL_SCRUB_PAUSE) error = spa_scrub_pause_resume(spa, POOL_SCRUB_PAUSE); else if (zc->zc_cookie == POOL_SCAN_NONE) error = spa_scan_stop(spa); else error = spa_scan(spa, zc->zc_cookie); spa_close(spa, FTAG); return (error); } +/* + * inputs: + * poolname name of the pool + * scan_type scan func (pool_scan_func_t) + * scan_command scrub pause/resume flag (pool_scrub_cmd_t) + */ +static const zfs_ioc_key_t zfs_keys_pool_scrub[] = { + {"scan_type", DATA_TYPE_UINT64, 0}, + {"scan_command", DATA_TYPE_UINT64, 0}, +}; + +static int +zfs_ioc_pool_scrub(const char *poolname, nvlist_t *innvl, nvlist_t *outnvl) +{ + spa_t *spa; + int error; + uint64_t scan_type, scan_cmd; + + if (nvlist_lookup_uint64(innvl, "scan_type", &scan_type) != 0) + return (SET_ERROR(EINVAL)); + if (nvlist_lookup_uint64(innvl, "scan_command", &scan_cmd) != 0) + return (SET_ERROR(EINVAL)); + + if (scan_cmd >= POOL_SCRUB_FLAGS_END) + return (SET_ERROR(EINVAL)); + + if ((error = spa_open(poolname, &spa, FTAG)) != 0) + return (error); + + if (scan_cmd == POOL_SCRUB_PAUSE) { + error = spa_scrub_pause_resume(spa, POOL_SCRUB_PAUSE); + } else if (scan_type == POOL_SCAN_NONE) { + error = spa_scan_stop(spa); + } else { + error = spa_scan(spa, scan_type); + } + + spa_close(spa, FTAG); + return (error); +} + static int zfs_ioc_pool_freeze(zfs_cmd_t *zc) { spa_t *spa; int error; error = spa_open(zc->zc_name, &spa, FTAG); if (error == 0) { spa_freeze(spa); spa_close(spa, FTAG); } return (error); } static int zfs_ioc_pool_upgrade(zfs_cmd_t *zc) { spa_t *spa; int error; if ((error = spa_open(zc->zc_name, &spa, FTAG)) != 0) return (error); if (zc->zc_cookie < spa_version(spa) || !SPA_VERSION_IS_SUPPORTED(zc->zc_cookie)) { spa_close(spa, FTAG); return (SET_ERROR(EINVAL)); } spa_upgrade(spa, zc->zc_cookie); spa_close(spa, FTAG); return (error); } static int zfs_ioc_pool_get_history(zfs_cmd_t *zc) { spa_t *spa; char *hist_buf; uint64_t size; int error; if ((size = zc->zc_history_len) == 0) return (SET_ERROR(EINVAL)); if ((error = spa_open(zc->zc_name, &spa, FTAG)) != 0) return (error); if (spa_version(spa) < SPA_VERSION_ZPOOL_HISTORY) { spa_close(spa, FTAG); return (SET_ERROR(ENOTSUP)); } hist_buf = vmem_alloc(size, KM_SLEEP); if ((error = spa_history_get(spa, &zc->zc_history_offset, &zc->zc_history_len, hist_buf)) == 0) { error = ddi_copyout(hist_buf, (void *)(uintptr_t)zc->zc_history, zc->zc_history_len, zc->zc_iflags); } spa_close(spa, FTAG); vmem_free(hist_buf, size); return (error); } static int zfs_ioc_pool_reguid(zfs_cmd_t *zc) { spa_t *spa; int error; error = spa_open(zc->zc_name, &spa, FTAG); if (error == 0) { error = spa_change_guid(spa); spa_close(spa, FTAG); } return (error); } static int zfs_ioc_dsobj_to_dsname(zfs_cmd_t *zc) { return (dsl_dsobj_to_dsname(zc->zc_name, zc->zc_obj, zc->zc_value)); } /* * inputs: * zc_name name of filesystem * zc_obj object to find * * outputs: * zc_value name of object */ static int zfs_ioc_obj_to_path(zfs_cmd_t *zc) { objset_t *os; int error; /* XXX reading from objset not owned */ if ((error = dmu_objset_hold_flags(zc->zc_name, B_TRUE, FTAG, &os)) != 0) return (error); if (dmu_objset_type(os) != DMU_OST_ZFS) { dmu_objset_rele_flags(os, B_TRUE, FTAG); return (SET_ERROR(EINVAL)); } error = zfs_obj_to_path(os, zc->zc_obj, zc->zc_value, sizeof (zc->zc_value)); dmu_objset_rele_flags(os, B_TRUE, FTAG); return (error); } /* * inputs: * zc_name name of filesystem * zc_obj object to find * * outputs: * zc_stat stats on object * zc_value path to object */ static int zfs_ioc_obj_to_stats(zfs_cmd_t *zc) { objset_t *os; int error; /* XXX reading from objset not owned */ if ((error = dmu_objset_hold_flags(zc->zc_name, B_TRUE, FTAG, &os)) != 0) return (error); if (dmu_objset_type(os) != DMU_OST_ZFS) { dmu_objset_rele_flags(os, B_TRUE, FTAG); return (SET_ERROR(EINVAL)); } error = zfs_obj_to_stats(os, zc->zc_obj, &zc->zc_stat, zc->zc_value, sizeof (zc->zc_value)); dmu_objset_rele_flags(os, B_TRUE, FTAG); return (error); } static int zfs_ioc_vdev_add(zfs_cmd_t *zc) { spa_t *spa; int error; nvlist_t *config; error = spa_open(zc->zc_name, &spa, FTAG); if (error != 0) return (error); error = get_nvlist(zc->zc_nvlist_conf, zc->zc_nvlist_conf_size, zc->zc_iflags, &config); if (error == 0) { error = spa_vdev_add(spa, config); nvlist_free(config); } spa_close(spa, FTAG); return (error); } /* * inputs: * zc_name name of the pool * zc_guid guid of vdev to remove * zc_cookie cancel removal */ static int zfs_ioc_vdev_remove(zfs_cmd_t *zc) { spa_t *spa; int error; error = spa_open(zc->zc_name, &spa, FTAG); if (error != 0) return (error); if (zc->zc_cookie != 0) { error = spa_vdev_remove_cancel(spa); } else { error = spa_vdev_remove(spa, zc->zc_guid, B_FALSE); } spa_close(spa, FTAG); return (error); } static int zfs_ioc_vdev_set_state(zfs_cmd_t *zc) { spa_t *spa; int error; vdev_state_t newstate = VDEV_STATE_UNKNOWN; if ((error = spa_open(zc->zc_name, &spa, FTAG)) != 0) return (error); switch (zc->zc_cookie) { case VDEV_STATE_ONLINE: error = vdev_online(spa, zc->zc_guid, zc->zc_obj, &newstate); break; case VDEV_STATE_OFFLINE: error = vdev_offline(spa, zc->zc_guid, zc->zc_obj); break; case VDEV_STATE_FAULTED: if (zc->zc_obj != VDEV_AUX_ERR_EXCEEDED && zc->zc_obj != VDEV_AUX_EXTERNAL && zc->zc_obj != VDEV_AUX_EXTERNAL_PERSIST) zc->zc_obj = VDEV_AUX_ERR_EXCEEDED; error = vdev_fault(spa, zc->zc_guid, zc->zc_obj); break; case VDEV_STATE_DEGRADED: if (zc->zc_obj != VDEV_AUX_ERR_EXCEEDED && zc->zc_obj != VDEV_AUX_EXTERNAL) zc->zc_obj = VDEV_AUX_ERR_EXCEEDED; error = vdev_degrade(spa, zc->zc_guid, zc->zc_obj); break; case VDEV_STATE_REMOVED: error = vdev_remove_wanted(spa, zc->zc_guid); break; default: error = SET_ERROR(EINVAL); } zc->zc_cookie = newstate; spa_close(spa, FTAG); return (error); } static int zfs_ioc_vdev_attach(zfs_cmd_t *zc) { spa_t *spa; nvlist_t *config; int replacing = zc->zc_cookie; int rebuild = zc->zc_simple; int error; if ((error = spa_open(zc->zc_name, &spa, FTAG)) != 0) return (error); if ((error = get_nvlist(zc->zc_nvlist_conf, zc->zc_nvlist_conf_size, zc->zc_iflags, &config)) == 0) { error = spa_vdev_attach(spa, zc->zc_guid, config, replacing, rebuild); nvlist_free(config); } spa_close(spa, FTAG); return (error); } static int zfs_ioc_vdev_detach(zfs_cmd_t *zc) { spa_t *spa; int error; if ((error = spa_open(zc->zc_name, &spa, FTAG)) != 0) return (error); error = spa_vdev_detach(spa, zc->zc_guid, 0, B_FALSE); spa_close(spa, FTAG); return (error); } static int zfs_ioc_vdev_split(zfs_cmd_t *zc) { spa_t *spa; nvlist_t *config, *props = NULL; int error; boolean_t exp = !!(zc->zc_cookie & ZPOOL_EXPORT_AFTER_SPLIT); if ((error = spa_open(zc->zc_name, &spa, FTAG)) != 0) return (error); if ((error = get_nvlist(zc->zc_nvlist_conf, zc->zc_nvlist_conf_size, zc->zc_iflags, &config))) { spa_close(spa, FTAG); return (error); } if (zc->zc_nvlist_src_size != 0 && (error = get_nvlist(zc->zc_nvlist_src, zc->zc_nvlist_src_size, zc->zc_iflags, &props))) { spa_close(spa, FTAG); nvlist_free(config); return (error); } error = spa_vdev_split_mirror(spa, zc->zc_string, config, props, exp); spa_close(spa, FTAG); nvlist_free(config); nvlist_free(props); return (error); } static int zfs_ioc_vdev_setpath(zfs_cmd_t *zc) { spa_t *spa; const char *path = zc->zc_value; uint64_t guid = zc->zc_guid; int error; error = spa_open(zc->zc_name, &spa, FTAG); if (error != 0) return (error); error = spa_vdev_setpath(spa, guid, path); spa_close(spa, FTAG); return (error); } static int zfs_ioc_vdev_setfru(zfs_cmd_t *zc) { spa_t *spa; const char *fru = zc->zc_value; uint64_t guid = zc->zc_guid; int error; error = spa_open(zc->zc_name, &spa, FTAG); if (error != 0) return (error); error = spa_vdev_setfru(spa, guid, fru); spa_close(spa, FTAG); return (error); } static int zfs_ioc_objset_stats_impl(zfs_cmd_t *zc, objset_t *os) { int error = 0; nvlist_t *nv; dmu_objset_fast_stat(os, &zc->zc_objset_stats); if (!zc->zc_simple && zc->zc_nvlist_dst != 0 && (error = dsl_prop_get_all(os, &nv)) == 0) { dmu_objset_stats(os, nv); /* * NB: zvol_get_stats() will read the objset contents, * which we aren't supposed to do with a * DS_MODE_USER hold, because it could be * inconsistent. So this is a bit of a workaround... * XXX reading without owning */ if (!zc->zc_objset_stats.dds_inconsistent && dmu_objset_type(os) == DMU_OST_ZVOL) { error = zvol_get_stats(os, nv); if (error == EIO) { nvlist_free(nv); return (error); } VERIFY0(error); } if (error == 0) error = put_nvlist(zc, nv); nvlist_free(nv); } return (error); } /* * inputs: * zc_name name of filesystem * zc_nvlist_dst_size size of buffer for property nvlist * * outputs: * zc_objset_stats stats * zc_nvlist_dst property nvlist * zc_nvlist_dst_size size of property nvlist */ static int zfs_ioc_objset_stats(zfs_cmd_t *zc) { objset_t *os; int error; error = dmu_objset_hold(zc->zc_name, FTAG, &os); if (error == 0) { error = zfs_ioc_objset_stats_impl(zc, os); dmu_objset_rele(os, FTAG); } return (error); } /* * inputs: * zc_name name of filesystem * zc_nvlist_dst_size size of buffer for property nvlist * * outputs: * zc_nvlist_dst received property nvlist * zc_nvlist_dst_size size of received property nvlist * * Gets received properties (distinct from local properties on or after * SPA_VERSION_RECVD_PROPS) for callers who want to differentiate received from * local property values. */ static int zfs_ioc_objset_recvd_props(zfs_cmd_t *zc) { int error = 0; nvlist_t *nv; /* * Without this check, we would return local property values if the * caller has not already received properties on or after * SPA_VERSION_RECVD_PROPS. */ if (!dsl_prop_get_hasrecvd(zc->zc_name)) return (SET_ERROR(ENOTSUP)); if (zc->zc_nvlist_dst != 0 && (error = dsl_prop_get_received(zc->zc_name, &nv)) == 0) { error = put_nvlist(zc, nv); nvlist_free(nv); } return (error); } static int nvl_add_zplprop(objset_t *os, nvlist_t *props, zfs_prop_t prop) { uint64_t value; int error; /* * zfs_get_zplprop() will either find a value or give us * the default value (if there is one). */ if ((error = zfs_get_zplprop(os, prop, &value)) != 0) return (error); VERIFY(nvlist_add_uint64(props, zfs_prop_to_name(prop), value) == 0); return (0); } /* * inputs: * zc_name name of filesystem * zc_nvlist_dst_size size of buffer for zpl property nvlist * * outputs: * zc_nvlist_dst zpl property nvlist * zc_nvlist_dst_size size of zpl property nvlist */ static int zfs_ioc_objset_zplprops(zfs_cmd_t *zc) { objset_t *os; int err; /* XXX reading without owning */ if ((err = dmu_objset_hold(zc->zc_name, FTAG, &os))) return (err); dmu_objset_fast_stat(os, &zc->zc_objset_stats); /* * NB: nvl_add_zplprop() will read the objset contents, * which we aren't supposed to do with a DS_MODE_USER * hold, because it could be inconsistent. */ if (zc->zc_nvlist_dst != 0 && !zc->zc_objset_stats.dds_inconsistent && dmu_objset_type(os) == DMU_OST_ZFS) { nvlist_t *nv; VERIFY(nvlist_alloc(&nv, NV_UNIQUE_NAME, KM_SLEEP) == 0); if ((err = nvl_add_zplprop(os, nv, ZFS_PROP_VERSION)) == 0 && (err = nvl_add_zplprop(os, nv, ZFS_PROP_NORMALIZE)) == 0 && (err = nvl_add_zplprop(os, nv, ZFS_PROP_UTF8ONLY)) == 0 && (err = nvl_add_zplprop(os, nv, ZFS_PROP_CASE)) == 0) err = put_nvlist(zc, nv); nvlist_free(nv); } else { err = SET_ERROR(ENOENT); } dmu_objset_rele(os, FTAG); return (err); } /* * inputs: * zc_name name of filesystem * zc_cookie zap cursor * zc_nvlist_dst_size size of buffer for property nvlist * * outputs: * zc_name name of next filesystem * zc_cookie zap cursor * zc_objset_stats stats * zc_nvlist_dst property nvlist * zc_nvlist_dst_size size of property nvlist */ static int zfs_ioc_dataset_list_next(zfs_cmd_t *zc) { objset_t *os; int error; char *p; size_t orig_len = strlen(zc->zc_name); top: if ((error = dmu_objset_hold(zc->zc_name, FTAG, &os))) { if (error == ENOENT) error = SET_ERROR(ESRCH); return (error); } p = strrchr(zc->zc_name, '/'); if (p == NULL || p[1] != '\0') (void) strlcat(zc->zc_name, "/", sizeof (zc->zc_name)); p = zc->zc_name + strlen(zc->zc_name); do { error = dmu_dir_list_next(os, sizeof (zc->zc_name) - (p - zc->zc_name), p, NULL, &zc->zc_cookie); if (error == ENOENT) error = SET_ERROR(ESRCH); } while (error == 0 && zfs_dataset_name_hidden(zc->zc_name)); dmu_objset_rele(os, FTAG); /* * If it's an internal dataset (ie. with a '$' in its name), * don't try to get stats for it, otherwise we'll return ENOENT. */ if (error == 0 && strchr(zc->zc_name, '$') == NULL) { error = zfs_ioc_objset_stats(zc); /* fill in the stats */ if (error == ENOENT) { /* We lost a race with destroy, get the next one. */ zc->zc_name[orig_len] = '\0'; goto top; } } return (error); } /* * inputs: * zc_name name of filesystem * zc_cookie zap cursor * zc_nvlist_src iteration range nvlist * zc_nvlist_src_size size of iteration range nvlist * * outputs: * zc_name name of next snapshot * zc_objset_stats stats * zc_nvlist_dst property nvlist * zc_nvlist_dst_size size of property nvlist */ static int zfs_ioc_snapshot_list_next(zfs_cmd_t *zc) { int error; objset_t *os, *ossnap; dsl_dataset_t *ds; uint64_t min_txg = 0, max_txg = 0; if (zc->zc_nvlist_src_size != 0) { nvlist_t *props = NULL; error = get_nvlist(zc->zc_nvlist_src, zc->zc_nvlist_src_size, zc->zc_iflags, &props); if (error != 0) return (error); (void) nvlist_lookup_uint64(props, SNAP_ITER_MIN_TXG, &min_txg); (void) nvlist_lookup_uint64(props, SNAP_ITER_MAX_TXG, &max_txg); nvlist_free(props); } error = dmu_objset_hold(zc->zc_name, FTAG, &os); if (error != 0) { return (error == ENOENT ? SET_ERROR(ESRCH) : error); } /* * A dataset name of maximum length cannot have any snapshots, * so exit immediately. */ if (strlcat(zc->zc_name, "@", sizeof (zc->zc_name)) >= ZFS_MAX_DATASET_NAME_LEN) { dmu_objset_rele(os, FTAG); return (SET_ERROR(ESRCH)); } while (error == 0) { if (issig(JUSTLOOKING) && issig(FORREAL)) { error = SET_ERROR(EINTR); break; } error = dmu_snapshot_list_next(os, sizeof (zc->zc_name) - strlen(zc->zc_name), zc->zc_name + strlen(zc->zc_name), &zc->zc_obj, &zc->zc_cookie, NULL); if (error == ENOENT) { error = SET_ERROR(ESRCH); break; } else if (error != 0) { break; } error = dsl_dataset_hold_obj(dmu_objset_pool(os), zc->zc_obj, FTAG, &ds); if (error != 0) break; if ((min_txg != 0 && dsl_get_creationtxg(ds) < min_txg) || (max_txg != 0 && dsl_get_creationtxg(ds) > max_txg)) { dsl_dataset_rele(ds, FTAG); /* undo snapshot name append */ *(strchr(zc->zc_name, '@') + 1) = '\0'; /* skip snapshot */ continue; } if (zc->zc_simple) { dsl_dataset_fast_stat(ds, &zc->zc_objset_stats); dsl_dataset_rele(ds, FTAG); break; } if ((error = dmu_objset_from_ds(ds, &ossnap)) != 0) { dsl_dataset_rele(ds, FTAG); break; } if ((error = zfs_ioc_objset_stats_impl(zc, ossnap)) != 0) { dsl_dataset_rele(ds, FTAG); break; } dsl_dataset_rele(ds, FTAG); break; } dmu_objset_rele(os, FTAG); /* if we failed, undo the @ that we tacked on to zc_name */ if (error != 0) *strchr(zc->zc_name, '@') = '\0'; return (error); } static int zfs_prop_set_userquota(const char *dsname, nvpair_t *pair) { const char *propname = nvpair_name(pair); uint64_t *valary; unsigned int vallen; const char *dash, *domain; zfs_userquota_prop_t type; uint64_t rid; uint64_t quota; zfsvfs_t *zfsvfs; int err; if (nvpair_type(pair) == DATA_TYPE_NVLIST) { nvlist_t *attrs; VERIFY(nvpair_value_nvlist(pair, &attrs) == 0); if (nvlist_lookup_nvpair(attrs, ZPROP_VALUE, &pair) != 0) return (SET_ERROR(EINVAL)); } /* * A correctly constructed propname is encoded as * userquota@-. */ if ((dash = strchr(propname, '-')) == NULL || nvpair_value_uint64_array(pair, &valary, &vallen) != 0 || vallen != 3) return (SET_ERROR(EINVAL)); domain = dash + 1; type = valary[0]; rid = valary[1]; quota = valary[2]; err = zfsvfs_hold(dsname, FTAG, &zfsvfs, B_FALSE); if (err == 0) { err = zfs_set_userquota(zfsvfs, type, domain, rid, quota); zfsvfs_rele(zfsvfs, FTAG); } return (err); } /* * If the named property is one that has a special function to set its value, * return 0 on success and a positive error code on failure; otherwise if it is * not one of the special properties handled by this function, return -1. * * XXX: It would be better for callers of the property interface if we handled * these special cases in dsl_prop.c (in the dsl layer). */ static int zfs_prop_set_special(const char *dsname, zprop_source_t source, nvpair_t *pair) { const char *propname = nvpair_name(pair); zfs_prop_t prop = zfs_name_to_prop(propname); uint64_t intval = 0; const char *strval = NULL; int err = -1; if (prop == ZPROP_USERPROP) { if (zfs_prop_userquota(propname)) return (zfs_prop_set_userquota(dsname, pair)); return (-1); } if (nvpair_type(pair) == DATA_TYPE_NVLIST) { nvlist_t *attrs; VERIFY(nvpair_value_nvlist(pair, &attrs) == 0); VERIFY(nvlist_lookup_nvpair(attrs, ZPROP_VALUE, &pair) == 0); } /* all special properties are numeric except for keylocation */ if (zfs_prop_get_type(prop) == PROP_TYPE_STRING) { strval = fnvpair_value_string(pair); } else { intval = fnvpair_value_uint64(pair); } switch (prop) { case ZFS_PROP_QUOTA: err = dsl_dir_set_quota(dsname, source, intval); break; case ZFS_PROP_REFQUOTA: err = dsl_dataset_set_refquota(dsname, source, intval); break; case ZFS_PROP_FILESYSTEM_LIMIT: case ZFS_PROP_SNAPSHOT_LIMIT: if (intval == UINT64_MAX) { /* clearing the limit, just do it */ err = 0; } else { err = dsl_dir_activate_fs_ss_limit(dsname); } /* * Set err to -1 to force the zfs_set_prop_nvlist code down the * default path to set the value in the nvlist. */ if (err == 0) err = -1; break; case ZFS_PROP_KEYLOCATION: err = dsl_crypto_can_set_keylocation(dsname, strval); /* * Set err to -1 to force the zfs_set_prop_nvlist code down the * default path to set the value in the nvlist. */ if (err == 0) err = -1; break; case ZFS_PROP_RESERVATION: err = dsl_dir_set_reservation(dsname, source, intval); break; case ZFS_PROP_REFRESERVATION: err = dsl_dataset_set_refreservation(dsname, source, intval); break; case ZFS_PROP_COMPRESSION: err = dsl_dataset_set_compression(dsname, source, intval); /* * Set err to -1 to force the zfs_set_prop_nvlist code down the * default path to set the value in the nvlist. */ if (err == 0) err = -1; break; case ZFS_PROP_VOLSIZE: err = zvol_set_volsize(dsname, intval); break; case ZFS_PROP_SNAPDEV: err = zvol_set_snapdev(dsname, source, intval); break; case ZFS_PROP_VOLMODE: err = zvol_set_volmode(dsname, source, intval); break; case ZFS_PROP_VERSION: { zfsvfs_t *zfsvfs; if ((err = zfsvfs_hold(dsname, FTAG, &zfsvfs, B_TRUE)) != 0) break; err = zfs_set_version(zfsvfs, intval); zfsvfs_rele(zfsvfs, FTAG); if (err == 0 && intval >= ZPL_VERSION_USERSPACE) { zfs_cmd_t *zc; zc = kmem_zalloc(sizeof (zfs_cmd_t), KM_SLEEP); (void) strlcpy(zc->zc_name, dsname, sizeof (zc->zc_name)); (void) zfs_ioc_userspace_upgrade(zc); (void) zfs_ioc_id_quota_upgrade(zc); kmem_free(zc, sizeof (zfs_cmd_t)); } break; } default: err = -1; } return (err); } static boolean_t zfs_is_namespace_prop(zfs_prop_t prop) { switch (prop) { case ZFS_PROP_ATIME: case ZFS_PROP_RELATIME: case ZFS_PROP_DEVICES: case ZFS_PROP_EXEC: case ZFS_PROP_SETUID: case ZFS_PROP_READONLY: case ZFS_PROP_XATTR: case ZFS_PROP_NBMAND: return (B_TRUE); default: return (B_FALSE); } } /* * This function is best effort. If it fails to set any of the given properties, * it continues to set as many as it can and returns the last error * encountered. If the caller provides a non-NULL errlist, it will be filled in * with the list of names of all the properties that failed along with the * corresponding error numbers. * * If every property is set successfully, zero is returned and errlist is not * modified. */ int zfs_set_prop_nvlist(const char *dsname, zprop_source_t source, nvlist_t *nvl, nvlist_t *errlist) { nvpair_t *pair; nvpair_t *propval; int rv = 0; int err; uint64_t intval; const char *strval; boolean_t should_update_mount_cache = B_FALSE; nvlist_t *genericnvl = fnvlist_alloc(); nvlist_t *retrynvl = fnvlist_alloc(); retry: pair = NULL; while ((pair = nvlist_next_nvpair(nvl, pair)) != NULL) { const char *propname = nvpair_name(pair); zfs_prop_t prop = zfs_name_to_prop(propname); err = 0; /* decode the property value */ propval = pair; if (nvpair_type(pair) == DATA_TYPE_NVLIST) { nvlist_t *attrs; attrs = fnvpair_value_nvlist(pair); if (nvlist_lookup_nvpair(attrs, ZPROP_VALUE, &propval) != 0) err = SET_ERROR(EINVAL); } /* Validate value type */ if (err == 0 && source == ZPROP_SRC_INHERITED) { /* inherited properties are expected to be booleans */ if (nvpair_type(propval) != DATA_TYPE_BOOLEAN) err = SET_ERROR(EINVAL); } else if (err == 0 && prop == ZPROP_USERPROP) { if (zfs_prop_user(propname)) { if (nvpair_type(propval) != DATA_TYPE_STRING) err = SET_ERROR(EINVAL); } else if (zfs_prop_userquota(propname)) { if (nvpair_type(propval) != DATA_TYPE_UINT64_ARRAY) err = SET_ERROR(EINVAL); } else { err = SET_ERROR(EINVAL); } } else if (err == 0) { if (nvpair_type(propval) == DATA_TYPE_STRING) { if (zfs_prop_get_type(prop) != PROP_TYPE_STRING) err = SET_ERROR(EINVAL); } else if (nvpair_type(propval) == DATA_TYPE_UINT64) { const char *unused; intval = fnvpair_value_uint64(propval); switch (zfs_prop_get_type(prop)) { case PROP_TYPE_NUMBER: break; case PROP_TYPE_STRING: err = SET_ERROR(EINVAL); break; case PROP_TYPE_INDEX: if (zfs_prop_index_to_string(prop, intval, &unused) != 0) err = SET_ERROR(ZFS_ERR_BADPROP); break; default: cmn_err(CE_PANIC, "unknown property type"); } } else { err = SET_ERROR(EINVAL); } } /* Validate permissions */ if (err == 0) err = zfs_check_settable(dsname, pair, CRED()); if (err == 0) { if (source == ZPROP_SRC_INHERITED) err = -1; /* does not need special handling */ else err = zfs_prop_set_special(dsname, source, pair); if (err == -1) { /* * For better performance we build up a list of * properties to set in a single transaction. */ err = nvlist_add_nvpair(genericnvl, pair); } else if (err != 0 && nvl != retrynvl) { /* * This may be a spurious error caused by * receiving quota and reservation out of order. * Try again in a second pass. */ err = nvlist_add_nvpair(retrynvl, pair); } } if (err != 0) { if (errlist != NULL) fnvlist_add_int32(errlist, propname, err); rv = err; } if (zfs_is_namespace_prop(prop)) should_update_mount_cache = B_TRUE; } if (nvl != retrynvl && !nvlist_empty(retrynvl)) { nvl = retrynvl; goto retry; } if (nvlist_empty(genericnvl)) goto out; /* * Try to set them all in one batch. */ err = dsl_props_set(dsname, source, genericnvl); if (err == 0) goto out; /* * If batching fails, we still want to set as many properties as we * can, so try setting them individually. */ pair = NULL; while ((pair = nvlist_next_nvpair(genericnvl, pair)) != NULL) { const char *propname = nvpair_name(pair); propval = pair; if (nvpair_type(pair) == DATA_TYPE_NVLIST) { nvlist_t *attrs; attrs = fnvpair_value_nvlist(pair); propval = fnvlist_lookup_nvpair(attrs, ZPROP_VALUE); } if (nvpair_type(propval) == DATA_TYPE_STRING) { strval = fnvpair_value_string(propval); err = dsl_prop_set_string(dsname, propname, source, strval); } else if (nvpair_type(propval) == DATA_TYPE_BOOLEAN) { err = dsl_prop_inherit(dsname, propname, source); } else { intval = fnvpair_value_uint64(propval); err = dsl_prop_set_int(dsname, propname, source, intval); } if (err != 0) { if (errlist != NULL) { fnvlist_add_int32(errlist, propname, err); } rv = err; } } out: if (should_update_mount_cache) zfs_ioctl_update_mount_cache(dsname); nvlist_free(genericnvl); nvlist_free(retrynvl); return (rv); } /* * Check that all the properties are valid user properties. */ static int zfs_check_userprops(nvlist_t *nvl) { nvpair_t *pair = NULL; while ((pair = nvlist_next_nvpair(nvl, pair)) != NULL) { const char *propname = nvpair_name(pair); if (!zfs_prop_user(propname) || nvpair_type(pair) != DATA_TYPE_STRING) return (SET_ERROR(EINVAL)); if (strlen(propname) >= ZAP_MAXNAMELEN) return (SET_ERROR(ENAMETOOLONG)); if (strlen(fnvpair_value_string(pair)) >= ZAP_MAXVALUELEN) return (SET_ERROR(E2BIG)); } return (0); } static void props_skip(nvlist_t *props, nvlist_t *skipped, nvlist_t **newprops) { nvpair_t *pair; VERIFY(nvlist_alloc(newprops, NV_UNIQUE_NAME, KM_SLEEP) == 0); pair = NULL; while ((pair = nvlist_next_nvpair(props, pair)) != NULL) { if (nvlist_exists(skipped, nvpair_name(pair))) continue; VERIFY(nvlist_add_nvpair(*newprops, pair) == 0); } } static int clear_received_props(const char *dsname, nvlist_t *props, nvlist_t *skipped) { int err = 0; nvlist_t *cleared_props = NULL; props_skip(props, skipped, &cleared_props); if (!nvlist_empty(cleared_props)) { /* * Acts on local properties until the dataset has received * properties at least once on or after SPA_VERSION_RECVD_PROPS. */ zprop_source_t flags = (ZPROP_SRC_NONE | (dsl_prop_get_hasrecvd(dsname) ? ZPROP_SRC_RECEIVED : 0)); err = zfs_set_prop_nvlist(dsname, flags, cleared_props, NULL); } nvlist_free(cleared_props); return (err); } /* * inputs: * zc_name name of filesystem * zc_value name of property to set * zc_nvlist_src{_size} nvlist of properties to apply * zc_cookie received properties flag * * outputs: * zc_nvlist_dst{_size} error for each unapplied received property */ static int zfs_ioc_set_prop(zfs_cmd_t *zc) { nvlist_t *nvl; boolean_t received = zc->zc_cookie; zprop_source_t source = (received ? ZPROP_SRC_RECEIVED : ZPROP_SRC_LOCAL); nvlist_t *errors; int error; if ((error = get_nvlist(zc->zc_nvlist_src, zc->zc_nvlist_src_size, zc->zc_iflags, &nvl)) != 0) return (error); if (received) { nvlist_t *origprops; if (dsl_prop_get_received(zc->zc_name, &origprops) == 0) { (void) clear_received_props(zc->zc_name, origprops, nvl); nvlist_free(origprops); } error = dsl_prop_set_hasrecvd(zc->zc_name); } errors = fnvlist_alloc(); if (error == 0) error = zfs_set_prop_nvlist(zc->zc_name, source, nvl, errors); if (zc->zc_nvlist_dst != 0 && errors != NULL) { (void) put_nvlist(zc, errors); } nvlist_free(errors); nvlist_free(nvl); return (error); } /* * inputs: * zc_name name of filesystem * zc_value name of property to inherit * zc_cookie revert to received value if TRUE * * outputs: none */ static int zfs_ioc_inherit_prop(zfs_cmd_t *zc) { const char *propname = zc->zc_value; zfs_prop_t prop = zfs_name_to_prop(propname); boolean_t received = zc->zc_cookie; zprop_source_t source = (received ? ZPROP_SRC_NONE /* revert to received value, if any */ : ZPROP_SRC_INHERITED); /* explicitly inherit */ nvlist_t *dummy; nvpair_t *pair; zprop_type_t type; int err; if (!received) { /* * Only check this in the non-received case. We want to allow * 'inherit -S' to revert non-inheritable properties like quota * and reservation to the received or default values even though * they are not considered inheritable. */ if (prop != ZPROP_USERPROP && !zfs_prop_inheritable(prop)) return (SET_ERROR(EINVAL)); } if (prop == ZPROP_USERPROP) { if (!zfs_prop_user(propname)) return (SET_ERROR(EINVAL)); type = PROP_TYPE_STRING; } else if (prop == ZFS_PROP_VOLSIZE || prop == ZFS_PROP_VERSION) { return (SET_ERROR(EINVAL)); } else { type = zfs_prop_get_type(prop); } /* * zfs_prop_set_special() expects properties in the form of an * nvpair with type info. */ dummy = fnvlist_alloc(); switch (type) { case PROP_TYPE_STRING: VERIFY(0 == nvlist_add_string(dummy, propname, "")); break; case PROP_TYPE_NUMBER: case PROP_TYPE_INDEX: VERIFY(0 == nvlist_add_uint64(dummy, propname, 0)); break; default: err = SET_ERROR(EINVAL); goto errout; } pair = nvlist_next_nvpair(dummy, NULL); if (pair == NULL) { err = SET_ERROR(EINVAL); } else { err = zfs_prop_set_special(zc->zc_name, source, pair); if (err == -1) /* property is not "special", needs handling */ err = dsl_prop_inherit(zc->zc_name, zc->zc_value, source); } errout: nvlist_free(dummy); return (err); } static int zfs_ioc_pool_set_props(zfs_cmd_t *zc) { nvlist_t *props; spa_t *spa; int error; nvpair_t *pair; if ((error = get_nvlist(zc->zc_nvlist_src, zc->zc_nvlist_src_size, zc->zc_iflags, &props))) return (error); /* * If the only property is the configfile, then just do a spa_lookup() * to handle the faulted case. */ pair = nvlist_next_nvpair(props, NULL); if (pair != NULL && strcmp(nvpair_name(pair), zpool_prop_to_name(ZPOOL_PROP_CACHEFILE)) == 0 && nvlist_next_nvpair(props, pair) == NULL) { mutex_enter(&spa_namespace_lock); if ((spa = spa_lookup(zc->zc_name)) != NULL) { spa_configfile_set(spa, props, B_FALSE); spa_write_cachefile(spa, B_FALSE, B_TRUE, B_FALSE); } mutex_exit(&spa_namespace_lock); if (spa != NULL) { nvlist_free(props); return (0); } } if ((error = spa_open(zc->zc_name, &spa, FTAG)) != 0) { nvlist_free(props); return (error); } error = spa_prop_set(spa, props); nvlist_free(props); spa_close(spa, FTAG); return (error); } static int zfs_ioc_pool_get_props(zfs_cmd_t *zc) { spa_t *spa; int error; nvlist_t *nvp = NULL; if ((error = spa_open(zc->zc_name, &spa, FTAG)) != 0) { /* * If the pool is faulted, there may be properties we can still * get (such as altroot and cachefile), so attempt to get them * anyway. */ mutex_enter(&spa_namespace_lock); if ((spa = spa_lookup(zc->zc_name)) != NULL) error = spa_prop_get(spa, &nvp); mutex_exit(&spa_namespace_lock); } else { error = spa_prop_get(spa, &nvp); spa_close(spa, FTAG); } if (error == 0 && zc->zc_nvlist_dst != 0) error = put_nvlist(zc, nvp); else error = SET_ERROR(EFAULT); nvlist_free(nvp); return (error); } /* * innvl: { * "vdevprops_set_vdev" -> guid * "vdevprops_set_props" -> { prop -> value } * } * * outnvl: propname -> error code (int32) */ static const zfs_ioc_key_t zfs_keys_vdev_set_props[] = { {ZPOOL_VDEV_PROPS_SET_VDEV, DATA_TYPE_UINT64, 0}, {ZPOOL_VDEV_PROPS_SET_PROPS, DATA_TYPE_NVLIST, 0} }; static int zfs_ioc_vdev_set_props(const char *poolname, nvlist_t *innvl, nvlist_t *outnvl) { spa_t *spa; int error; vdev_t *vd; uint64_t vdev_guid; /* Early validation */ if (nvlist_lookup_uint64(innvl, ZPOOL_VDEV_PROPS_SET_VDEV, &vdev_guid) != 0) return (SET_ERROR(EINVAL)); if (outnvl == NULL) return (SET_ERROR(EINVAL)); if ((error = spa_open(poolname, &spa, FTAG)) != 0) return (error); ASSERT(spa_writeable(spa)); if ((vd = spa_lookup_by_guid(spa, vdev_guid, B_TRUE)) == NULL) { spa_close(spa, FTAG); return (SET_ERROR(ENOENT)); } error = vdev_prop_set(vd, innvl, outnvl); spa_close(spa, FTAG); return (error); } /* * innvl: { * "vdevprops_get_vdev" -> guid * (optional) "vdevprops_get_props" -> { propname -> propid } * } * * outnvl: propname -> value */ static const zfs_ioc_key_t zfs_keys_vdev_get_props[] = { {ZPOOL_VDEV_PROPS_GET_VDEV, DATA_TYPE_UINT64, 0}, {ZPOOL_VDEV_PROPS_GET_PROPS, DATA_TYPE_NVLIST, ZK_OPTIONAL} }; static int zfs_ioc_vdev_get_props(const char *poolname, nvlist_t *innvl, nvlist_t *outnvl) { spa_t *spa; int error; vdev_t *vd; uint64_t vdev_guid; /* Early validation */ if (nvlist_lookup_uint64(innvl, ZPOOL_VDEV_PROPS_GET_VDEV, &vdev_guid) != 0) return (SET_ERROR(EINVAL)); if (outnvl == NULL) return (SET_ERROR(EINVAL)); if ((error = spa_open(poolname, &spa, FTAG)) != 0) return (error); if ((vd = spa_lookup_by_guid(spa, vdev_guid, B_TRUE)) == NULL) { spa_close(spa, FTAG); return (SET_ERROR(ENOENT)); } error = vdev_prop_get(vd, innvl, outnvl); spa_close(spa, FTAG); return (error); } /* * inputs: * zc_name name of filesystem * zc_nvlist_src{_size} nvlist of delegated permissions * zc_perm_action allow/unallow flag * * outputs: none */ static int zfs_ioc_set_fsacl(zfs_cmd_t *zc) { int error; nvlist_t *fsaclnv = NULL; if ((error = get_nvlist(zc->zc_nvlist_src, zc->zc_nvlist_src_size, zc->zc_iflags, &fsaclnv)) != 0) return (error); /* * Verify nvlist is constructed correctly */ if (zfs_deleg_verify_nvlist(fsaclnv) != 0) { nvlist_free(fsaclnv); return (SET_ERROR(EINVAL)); } /* * If we don't have PRIV_SYS_MOUNT, then validate * that user is allowed to hand out each permission in * the nvlist(s) */ error = secpolicy_zfs(CRED()); if (error != 0) { if (zc->zc_perm_action == B_FALSE) { error = dsl_deleg_can_allow(zc->zc_name, fsaclnv, CRED()); } else { error = dsl_deleg_can_unallow(zc->zc_name, fsaclnv, CRED()); } } if (error == 0) error = dsl_deleg_set(zc->zc_name, fsaclnv, zc->zc_perm_action); nvlist_free(fsaclnv); return (error); } /* * inputs: * zc_name name of filesystem * * outputs: * zc_nvlist_src{_size} nvlist of delegated permissions */ static int zfs_ioc_get_fsacl(zfs_cmd_t *zc) { nvlist_t *nvp; int error; if ((error = dsl_deleg_get(zc->zc_name, &nvp)) == 0) { error = put_nvlist(zc, nvp); nvlist_free(nvp); } return (error); } static void zfs_create_cb(objset_t *os, void *arg, cred_t *cr, dmu_tx_t *tx) { zfs_creat_t *zct = arg; zfs_create_fs(os, cr, zct->zct_zplprops, tx); } #define ZFS_PROP_UNDEFINED ((uint64_t)-1) /* * inputs: * os parent objset pointer (NULL if root fs) * fuids_ok fuids allowed in this version of the spa? * sa_ok SAs allowed in this version of the spa? * createprops list of properties requested by creator * * outputs: * zplprops values for the zplprops we attach to the master node object * is_ci true if requested file system will be purely case-insensitive * * Determine the settings for utf8only, normalization and * casesensitivity. Specific values may have been requested by the * creator and/or we can inherit values from the parent dataset. If * the file system is of too early a vintage, a creator can not * request settings for these properties, even if the requested * setting is the default value. We don't actually want to create dsl * properties for these, so remove them from the source nvlist after * processing. */ static int zfs_fill_zplprops_impl(objset_t *os, uint64_t zplver, boolean_t fuids_ok, boolean_t sa_ok, nvlist_t *createprops, nvlist_t *zplprops, boolean_t *is_ci) { uint64_t sense = ZFS_PROP_UNDEFINED; uint64_t norm = ZFS_PROP_UNDEFINED; uint64_t u8 = ZFS_PROP_UNDEFINED; int error; ASSERT(zplprops != NULL); /* parent dataset must be a filesystem */ if (os != NULL && os->os_phys->os_type != DMU_OST_ZFS) return (SET_ERROR(ZFS_ERR_WRONG_PARENT)); /* * Pull out creator prop choices, if any. */ if (createprops) { (void) nvlist_lookup_uint64(createprops, zfs_prop_to_name(ZFS_PROP_VERSION), &zplver); (void) nvlist_lookup_uint64(createprops, zfs_prop_to_name(ZFS_PROP_NORMALIZE), &norm); (void) nvlist_remove_all(createprops, zfs_prop_to_name(ZFS_PROP_NORMALIZE)); (void) nvlist_lookup_uint64(createprops, zfs_prop_to_name(ZFS_PROP_UTF8ONLY), &u8); (void) nvlist_remove_all(createprops, zfs_prop_to_name(ZFS_PROP_UTF8ONLY)); (void) nvlist_lookup_uint64(createprops, zfs_prop_to_name(ZFS_PROP_CASE), &sense); (void) nvlist_remove_all(createprops, zfs_prop_to_name(ZFS_PROP_CASE)); } /* * If the zpl version requested is whacky or the file system * or pool is version is too "young" to support normalization * and the creator tried to set a value for one of the props, * error out. */ if ((zplver < ZPL_VERSION_INITIAL || zplver > ZPL_VERSION) || (zplver >= ZPL_VERSION_FUID && !fuids_ok) || (zplver >= ZPL_VERSION_SA && !sa_ok) || (zplver < ZPL_VERSION_NORMALIZATION && (norm != ZFS_PROP_UNDEFINED || u8 != ZFS_PROP_UNDEFINED || sense != ZFS_PROP_UNDEFINED))) return (SET_ERROR(ENOTSUP)); /* * Put the version in the zplprops */ VERIFY(nvlist_add_uint64(zplprops, zfs_prop_to_name(ZFS_PROP_VERSION), zplver) == 0); if (norm == ZFS_PROP_UNDEFINED && (error = zfs_get_zplprop(os, ZFS_PROP_NORMALIZE, &norm)) != 0) return (error); VERIFY(nvlist_add_uint64(zplprops, zfs_prop_to_name(ZFS_PROP_NORMALIZE), norm) == 0); /* * If we're normalizing, names must always be valid UTF-8 strings. */ if (norm) u8 = 1; if (u8 == ZFS_PROP_UNDEFINED && (error = zfs_get_zplprop(os, ZFS_PROP_UTF8ONLY, &u8)) != 0) return (error); VERIFY(nvlist_add_uint64(zplprops, zfs_prop_to_name(ZFS_PROP_UTF8ONLY), u8) == 0); if (sense == ZFS_PROP_UNDEFINED && (error = zfs_get_zplprop(os, ZFS_PROP_CASE, &sense)) != 0) return (error); VERIFY(nvlist_add_uint64(zplprops, zfs_prop_to_name(ZFS_PROP_CASE), sense) == 0); if (is_ci) *is_ci = (sense == ZFS_CASE_INSENSITIVE); return (0); } static int zfs_fill_zplprops(const char *dataset, nvlist_t *createprops, nvlist_t *zplprops, boolean_t *is_ci) { boolean_t fuids_ok, sa_ok; uint64_t zplver = ZPL_VERSION; objset_t *os = NULL; char parentname[ZFS_MAX_DATASET_NAME_LEN]; spa_t *spa; uint64_t spa_vers; int error; zfs_get_parent(dataset, parentname, sizeof (parentname)); if ((error = spa_open(dataset, &spa, FTAG)) != 0) return (error); spa_vers = spa_version(spa); spa_close(spa, FTAG); zplver = zfs_zpl_version_map(spa_vers); fuids_ok = (zplver >= ZPL_VERSION_FUID); sa_ok = (zplver >= ZPL_VERSION_SA); /* * Open parent object set so we can inherit zplprop values. */ if ((error = dmu_objset_hold(parentname, FTAG, &os)) != 0) return (error); error = zfs_fill_zplprops_impl(os, zplver, fuids_ok, sa_ok, createprops, zplprops, is_ci); dmu_objset_rele(os, FTAG); return (error); } static int zfs_fill_zplprops_root(uint64_t spa_vers, nvlist_t *createprops, nvlist_t *zplprops, boolean_t *is_ci) { boolean_t fuids_ok; boolean_t sa_ok; uint64_t zplver = ZPL_VERSION; int error; zplver = zfs_zpl_version_map(spa_vers); fuids_ok = (zplver >= ZPL_VERSION_FUID); sa_ok = (zplver >= ZPL_VERSION_SA); error = zfs_fill_zplprops_impl(NULL, zplver, fuids_ok, sa_ok, createprops, zplprops, is_ci); return (error); } /* * innvl: { * "type" -> dmu_objset_type_t (int32) * (optional) "props" -> { prop -> value } * (optional) "hidden_args" -> { "wkeydata" -> value } * raw uint8_t array of encryption wrapping key data (32 bytes) * } * * outnvl: propname -> error code (int32) */ static const zfs_ioc_key_t zfs_keys_create[] = { {"type", DATA_TYPE_INT32, 0}, {"props", DATA_TYPE_NVLIST, ZK_OPTIONAL}, {"hidden_args", DATA_TYPE_NVLIST, ZK_OPTIONAL}, }; static int zfs_ioc_create(const char *fsname, nvlist_t *innvl, nvlist_t *outnvl) { int error = 0; zfs_creat_t zct = { 0 }; nvlist_t *nvprops = NULL; nvlist_t *hidden_args = NULL; void (*cbfunc)(objset_t *os, void *arg, cred_t *cr, dmu_tx_t *tx); dmu_objset_type_t type; boolean_t is_insensitive = B_FALSE; dsl_crypto_params_t *dcp = NULL; type = (dmu_objset_type_t)fnvlist_lookup_int32(innvl, "type"); (void) nvlist_lookup_nvlist(innvl, "props", &nvprops); (void) nvlist_lookup_nvlist(innvl, ZPOOL_HIDDEN_ARGS, &hidden_args); switch (type) { case DMU_OST_ZFS: cbfunc = zfs_create_cb; break; case DMU_OST_ZVOL: cbfunc = zvol_create_cb; break; default: cbfunc = NULL; break; } if (strchr(fsname, '@') || strchr(fsname, '%')) return (SET_ERROR(EINVAL)); zct.zct_props = nvprops; if (cbfunc == NULL) return (SET_ERROR(EINVAL)); if (type == DMU_OST_ZVOL) { uint64_t volsize, volblocksize; if (nvprops == NULL) return (SET_ERROR(EINVAL)); if (nvlist_lookup_uint64(nvprops, zfs_prop_to_name(ZFS_PROP_VOLSIZE), &volsize) != 0) return (SET_ERROR(EINVAL)); if ((error = nvlist_lookup_uint64(nvprops, zfs_prop_to_name(ZFS_PROP_VOLBLOCKSIZE), &volblocksize)) != 0 && error != ENOENT) return (SET_ERROR(EINVAL)); if (error != 0) volblocksize = zfs_prop_default_numeric( ZFS_PROP_VOLBLOCKSIZE); if ((error = zvol_check_volblocksize(fsname, volblocksize)) != 0 || (error = zvol_check_volsize(volsize, volblocksize)) != 0) return (error); } else if (type == DMU_OST_ZFS) { int error; /* * We have to have normalization and * case-folding flags correct when we do the * file system creation, so go figure them out * now. */ VERIFY(nvlist_alloc(&zct.zct_zplprops, NV_UNIQUE_NAME, KM_SLEEP) == 0); error = zfs_fill_zplprops(fsname, nvprops, zct.zct_zplprops, &is_insensitive); if (error != 0) { nvlist_free(zct.zct_zplprops); return (error); } } error = dsl_crypto_params_create_nvlist(DCP_CMD_NONE, nvprops, hidden_args, &dcp); if (error != 0) { nvlist_free(zct.zct_zplprops); return (error); } error = dmu_objset_create(fsname, type, is_insensitive ? DS_FLAG_CI_DATASET : 0, dcp, cbfunc, &zct); nvlist_free(zct.zct_zplprops); dsl_crypto_params_free(dcp, !!error); /* * It would be nice to do this atomically. */ if (error == 0) { error = zfs_set_prop_nvlist(fsname, ZPROP_SRC_LOCAL, nvprops, outnvl); if (error != 0) { spa_t *spa; int error2; /* * Volumes will return EBUSY and cannot be destroyed * until all asynchronous minor handling (e.g. from * setting the volmode property) has completed. Wait for * the spa_zvol_taskq to drain then retry. */ error2 = dsl_destroy_head(fsname); while ((error2 == EBUSY) && (type == DMU_OST_ZVOL)) { error2 = spa_open(fsname, &spa, FTAG); if (error2 == 0) { taskq_wait(spa->spa_zvol_taskq); spa_close(spa, FTAG); } error2 = dsl_destroy_head(fsname); } } } return (error); } /* * innvl: { * "origin" -> name of origin snapshot * (optional) "props" -> { prop -> value } * (optional) "hidden_args" -> { "wkeydata" -> value } * raw uint8_t array of encryption wrapping key data (32 bytes) * } * * outputs: * outnvl: propname -> error code (int32) */ static const zfs_ioc_key_t zfs_keys_clone[] = { {"origin", DATA_TYPE_STRING, 0}, {"props", DATA_TYPE_NVLIST, ZK_OPTIONAL}, {"hidden_args", DATA_TYPE_NVLIST, ZK_OPTIONAL}, }; static int zfs_ioc_clone(const char *fsname, nvlist_t *innvl, nvlist_t *outnvl) { int error = 0; nvlist_t *nvprops = NULL; const char *origin_name; origin_name = fnvlist_lookup_string(innvl, "origin"); (void) nvlist_lookup_nvlist(innvl, "props", &nvprops); if (strchr(fsname, '@') || strchr(fsname, '%')) return (SET_ERROR(EINVAL)); if (dataset_namecheck(origin_name, NULL, NULL) != 0) return (SET_ERROR(EINVAL)); error = dmu_objset_clone(fsname, origin_name); /* * It would be nice to do this atomically. */ if (error == 0) { error = zfs_set_prop_nvlist(fsname, ZPROP_SRC_LOCAL, nvprops, outnvl); if (error != 0) (void) dsl_destroy_head(fsname); } return (error); } static const zfs_ioc_key_t zfs_keys_remap[] = { /* no nvl keys */ }; static int zfs_ioc_remap(const char *fsname, nvlist_t *innvl, nvlist_t *outnvl) { /* This IOCTL is no longer supported. */ (void) fsname, (void) innvl, (void) outnvl; return (0); } /* * innvl: { * "snaps" -> { snapshot1, snapshot2 } * (optional) "props" -> { prop -> value (string) } * } * * outnvl: snapshot -> error code (int32) */ static const zfs_ioc_key_t zfs_keys_snapshot[] = { {"snaps", DATA_TYPE_NVLIST, 0}, {"props", DATA_TYPE_NVLIST, ZK_OPTIONAL}, }; static int zfs_ioc_snapshot(const char *poolname, nvlist_t *innvl, nvlist_t *outnvl) { nvlist_t *snaps; nvlist_t *props = NULL; int error, poollen; nvpair_t *pair; (void) nvlist_lookup_nvlist(innvl, "props", &props); if (!nvlist_empty(props) && zfs_earlier_version(poolname, SPA_VERSION_SNAP_PROPS)) return (SET_ERROR(ENOTSUP)); if ((error = zfs_check_userprops(props)) != 0) return (error); snaps = fnvlist_lookup_nvlist(innvl, "snaps"); poollen = strlen(poolname); for (pair = nvlist_next_nvpair(snaps, NULL); pair != NULL; pair = nvlist_next_nvpair(snaps, pair)) { const char *name = nvpair_name(pair); char *cp = strchr(name, '@'); /* * The snap name must contain an @, and the part after it must * contain only valid characters. */ if (cp == NULL || zfs_component_namecheck(cp + 1, NULL, NULL) != 0) return (SET_ERROR(EINVAL)); /* * The snap must be in the specified pool. */ if (strncmp(name, poolname, poollen) != 0 || (name[poollen] != '/' && name[poollen] != '@')) return (SET_ERROR(EXDEV)); /* * Check for permission to set the properties on the fs. */ if (!nvlist_empty(props)) { *cp = '\0'; error = zfs_secpolicy_write_perms(name, ZFS_DELEG_PERM_USERPROP, CRED()); *cp = '@'; if (error != 0) return (error); } /* This must be the only snap of this fs. */ for (nvpair_t *pair2 = nvlist_next_nvpair(snaps, pair); pair2 != NULL; pair2 = nvlist_next_nvpair(snaps, pair2)) { if (strncmp(name, nvpair_name(pair2), cp - name + 1) == 0) { return (SET_ERROR(EXDEV)); } } } error = dsl_dataset_snapshot(snaps, props, outnvl); return (error); } /* * innvl: "message" -> string */ static const zfs_ioc_key_t zfs_keys_log_history[] = { {"message", DATA_TYPE_STRING, 0}, }; static int zfs_ioc_log_history(const char *unused, nvlist_t *innvl, nvlist_t *outnvl) { (void) unused, (void) outnvl; const char *message; char *poolname; spa_t *spa; int error; /* * The poolname in the ioctl is not set, we get it from the TSD, * which was set at the end of the last successful ioctl that allows * logging. The secpolicy func already checked that it is set. * Only one log ioctl is allowed after each successful ioctl, so * we clear the TSD here. */ poolname = tsd_get(zfs_allow_log_key); if (poolname == NULL) return (SET_ERROR(EINVAL)); (void) tsd_set(zfs_allow_log_key, NULL); error = spa_open(poolname, &spa, FTAG); kmem_strfree(poolname); if (error != 0) return (error); message = fnvlist_lookup_string(innvl, "message"); if (spa_version(spa) < SPA_VERSION_ZPOOL_HISTORY) { spa_close(spa, FTAG); return (SET_ERROR(ENOTSUP)); } error = spa_history_log(spa, message); spa_close(spa, FTAG); return (error); } /* * This ioctl is used to set the bootenv configuration on the current * pool. This configuration is stored in the second padding area of the label, * and it is used by the bootloader(s) to store the bootloader and/or system * specific data. * The data is stored as nvlist data stream, and is protected by * an embedded checksum. * The version can have two possible values: * VB_RAW: nvlist should have key GRUB_ENVMAP, value DATA_TYPE_STRING. * VB_NVLIST: nvlist with arbitrary pairs. */ static const zfs_ioc_key_t zfs_keys_set_bootenv[] = { {"version", DATA_TYPE_UINT64, 0}, {"", DATA_TYPE_ANY, ZK_OPTIONAL | ZK_WILDCARDLIST}, }; static int zfs_ioc_set_bootenv(const char *name, nvlist_t *innvl, nvlist_t *outnvl) { int error; spa_t *spa; if ((error = spa_open(name, &spa, FTAG)) != 0) return (error); spa_vdev_state_enter(spa, SCL_ALL); error = vdev_label_write_bootenv(spa->spa_root_vdev, innvl); (void) spa_vdev_state_exit(spa, NULL, 0); spa_close(spa, FTAG); return (error); } static const zfs_ioc_key_t zfs_keys_get_bootenv[] = { /* no nvl keys */ }; static int zfs_ioc_get_bootenv(const char *name, nvlist_t *innvl, nvlist_t *outnvl) { spa_t *spa; int error; if ((error = spa_open(name, &spa, FTAG)) != 0) return (error); spa_vdev_state_enter(spa, SCL_ALL); error = vdev_label_read_bootenv(spa->spa_root_vdev, outnvl); (void) spa_vdev_state_exit(spa, NULL, 0); spa_close(spa, FTAG); return (error); } /* * The dp_config_rwlock must not be held when calling this, because the * unmount may need to write out data. * * This function is best-effort. Callers must deal gracefully if it * remains mounted (or is remounted after this call). * * Returns 0 if the argument is not a snapshot, or it is not currently a * filesystem, or we were able to unmount it. Returns error code otherwise. */ void zfs_unmount_snap(const char *snapname) { if (strchr(snapname, '@') == NULL) return; (void) zfsctl_snapshot_unmount(snapname, MNT_FORCE); } static int zfs_unmount_snap_cb(const char *snapname, void *arg) { (void) arg; zfs_unmount_snap(snapname); return (0); } /* * When a clone is destroyed, its origin may also need to be destroyed, * in which case it must be unmounted. This routine will do that unmount * if necessary. */ void zfs_destroy_unmount_origin(const char *fsname) { int error; objset_t *os; dsl_dataset_t *ds; error = dmu_objset_hold(fsname, FTAG, &os); if (error != 0) return; ds = dmu_objset_ds(os); if (dsl_dir_is_clone(ds->ds_dir) && DS_IS_DEFER_DESTROY(ds->ds_prev)) { char originname[ZFS_MAX_DATASET_NAME_LEN]; dsl_dataset_name(ds->ds_prev, originname); dmu_objset_rele(os, FTAG); zfs_unmount_snap(originname); } else { dmu_objset_rele(os, FTAG); } } /* * innvl: { * "snaps" -> { snapshot1, snapshot2 } * (optional boolean) "defer" * } * * outnvl: snapshot -> error code (int32) */ static const zfs_ioc_key_t zfs_keys_destroy_snaps[] = { {"snaps", DATA_TYPE_NVLIST, 0}, {"defer", DATA_TYPE_BOOLEAN, ZK_OPTIONAL}, }; static int zfs_ioc_destroy_snaps(const char *poolname, nvlist_t *innvl, nvlist_t *outnvl) { int poollen; nvlist_t *snaps; nvpair_t *pair; boolean_t defer; spa_t *spa; snaps = fnvlist_lookup_nvlist(innvl, "snaps"); defer = nvlist_exists(innvl, "defer"); poollen = strlen(poolname); for (pair = nvlist_next_nvpair(snaps, NULL); pair != NULL; pair = nvlist_next_nvpair(snaps, pair)) { const char *name = nvpair_name(pair); /* * The snap must be in the specified pool to prevent the * invalid removal of zvol minors below. */ if (strncmp(name, poolname, poollen) != 0 || (name[poollen] != '/' && name[poollen] != '@')) return (SET_ERROR(EXDEV)); zfs_unmount_snap(nvpair_name(pair)); if (spa_open(name, &spa, FTAG) == 0) { zvol_remove_minors(spa, name, B_TRUE); spa_close(spa, FTAG); } } return (dsl_destroy_snapshots_nvl(snaps, defer, outnvl)); } /* * Create bookmarks. The bookmark names are of the form #. * All bookmarks and snapshots must be in the same pool. * dsl_bookmark_create_nvl_validate describes the nvlist schema in more detail. * * innvl: { * new_bookmark1 -> existing_snapshot, * new_bookmark2 -> existing_bookmark, * } * * outnvl: bookmark -> error code (int32) * */ static const zfs_ioc_key_t zfs_keys_bookmark[] = { {"...", DATA_TYPE_STRING, ZK_WILDCARDLIST}, }; static int zfs_ioc_bookmark(const char *poolname, nvlist_t *innvl, nvlist_t *outnvl) { (void) poolname; return (dsl_bookmark_create(innvl, outnvl)); } /* * innvl: { * property 1, property 2, ... * } * * outnvl: { * bookmark name 1 -> { property 1, property 2, ... }, * bookmark name 2 -> { property 1, property 2, ... } * } * */ static const zfs_ioc_key_t zfs_keys_get_bookmarks[] = { {"...", DATA_TYPE_BOOLEAN, ZK_WILDCARDLIST | ZK_OPTIONAL}, }; static int zfs_ioc_get_bookmarks(const char *fsname, nvlist_t *innvl, nvlist_t *outnvl) { return (dsl_get_bookmarks(fsname, innvl, outnvl)); } /* * innvl is not used. * * outnvl: { * property 1, property 2, ... * } * */ static const zfs_ioc_key_t zfs_keys_get_bookmark_props[] = { /* no nvl keys */ }; static int zfs_ioc_get_bookmark_props(const char *bookmark, nvlist_t *innvl, nvlist_t *outnvl) { (void) innvl; char fsname[ZFS_MAX_DATASET_NAME_LEN]; char *bmname; bmname = strchr(bookmark, '#'); if (bmname == NULL) return (SET_ERROR(EINVAL)); bmname++; (void) strlcpy(fsname, bookmark, sizeof (fsname)); *(strchr(fsname, '#')) = '\0'; return (dsl_get_bookmark_props(fsname, bmname, outnvl)); } /* * innvl: { * bookmark name 1, bookmark name 2 * } * * outnvl: bookmark -> error code (int32) * */ static const zfs_ioc_key_t zfs_keys_destroy_bookmarks[] = { {"...", DATA_TYPE_BOOLEAN, ZK_WILDCARDLIST}, }; static int zfs_ioc_destroy_bookmarks(const char *poolname, nvlist_t *innvl, nvlist_t *outnvl) { int error, poollen; poollen = strlen(poolname); for (nvpair_t *pair = nvlist_next_nvpair(innvl, NULL); pair != NULL; pair = nvlist_next_nvpair(innvl, pair)) { const char *name = nvpair_name(pair); const char *cp = strchr(name, '#'); /* * The bookmark name must contain an #, and the part after it * must contain only valid characters. */ if (cp == NULL || zfs_component_namecheck(cp + 1, NULL, NULL) != 0) return (SET_ERROR(EINVAL)); /* * The bookmark must be in the specified pool. */ if (strncmp(name, poolname, poollen) != 0 || (name[poollen] != '/' && name[poollen] != '#')) return (SET_ERROR(EXDEV)); } error = dsl_bookmark_destroy(innvl, outnvl); return (error); } static const zfs_ioc_key_t zfs_keys_channel_program[] = { {"program", DATA_TYPE_STRING, 0}, {"arg", DATA_TYPE_ANY, 0}, {"sync", DATA_TYPE_BOOLEAN_VALUE, ZK_OPTIONAL}, {"instrlimit", DATA_TYPE_UINT64, ZK_OPTIONAL}, {"memlimit", DATA_TYPE_UINT64, ZK_OPTIONAL}, }; static int zfs_ioc_channel_program(const char *poolname, nvlist_t *innvl, nvlist_t *outnvl) { const char *program; uint64_t instrlimit, memlimit; boolean_t sync_flag; nvpair_t *nvarg = NULL; program = fnvlist_lookup_string(innvl, ZCP_ARG_PROGRAM); if (0 != nvlist_lookup_boolean_value(innvl, ZCP_ARG_SYNC, &sync_flag)) { sync_flag = B_TRUE; } if (0 != nvlist_lookup_uint64(innvl, ZCP_ARG_INSTRLIMIT, &instrlimit)) { instrlimit = ZCP_DEFAULT_INSTRLIMIT; } if (0 != nvlist_lookup_uint64(innvl, ZCP_ARG_MEMLIMIT, &memlimit)) { memlimit = ZCP_DEFAULT_MEMLIMIT; } nvarg = fnvlist_lookup_nvpair(innvl, ZCP_ARG_ARGLIST); if (instrlimit == 0 || instrlimit > zfs_lua_max_instrlimit) return (SET_ERROR(EINVAL)); if (memlimit == 0 || memlimit > zfs_lua_max_memlimit) return (SET_ERROR(EINVAL)); return (zcp_eval(poolname, program, sync_flag, instrlimit, memlimit, nvarg, outnvl)); } /* * innvl: unused * outnvl: empty */ static const zfs_ioc_key_t zfs_keys_pool_checkpoint[] = { /* no nvl keys */ }; static int zfs_ioc_pool_checkpoint(const char *poolname, nvlist_t *innvl, nvlist_t *outnvl) { (void) innvl, (void) outnvl; return (spa_checkpoint(poolname)); } /* * innvl: unused * outnvl: empty */ static const zfs_ioc_key_t zfs_keys_pool_discard_checkpoint[] = { /* no nvl keys */ }; static int zfs_ioc_pool_discard_checkpoint(const char *poolname, nvlist_t *innvl, nvlist_t *outnvl) { (void) innvl, (void) outnvl; return (spa_checkpoint_discard(poolname)); } /* * inputs: * zc_name name of dataset to destroy * zc_defer_destroy mark for deferred destroy * * outputs: none */ static int zfs_ioc_destroy(zfs_cmd_t *zc) { objset_t *os; dmu_objset_type_t ost; int err; err = dmu_objset_hold(zc->zc_name, FTAG, &os); if (err != 0) return (err); ost = dmu_objset_type(os); dmu_objset_rele(os, FTAG); if (ost == DMU_OST_ZFS) zfs_unmount_snap(zc->zc_name); if (strchr(zc->zc_name, '@')) { err = dsl_destroy_snapshot(zc->zc_name, zc->zc_defer_destroy); } else { err = dsl_destroy_head(zc->zc_name); if (err == EEXIST) { /* * It is possible that the given DS may have * hidden child (%recv) datasets - "leftovers" * resulting from the previously interrupted * 'zfs receive'. * * 6 extra bytes for /%recv */ char namebuf[ZFS_MAX_DATASET_NAME_LEN + 6]; if (snprintf(namebuf, sizeof (namebuf), "%s/%s", zc->zc_name, recv_clone_name) >= sizeof (namebuf)) return (SET_ERROR(EINVAL)); /* * Try to remove the hidden child (%recv) and after * that try to remove the target dataset. * If the hidden child (%recv) does not exist * the original error (EEXIST) will be returned */ err = dsl_destroy_head(namebuf); if (err == 0) err = dsl_destroy_head(zc->zc_name); else if (err == ENOENT) err = SET_ERROR(EEXIST); } } return (err); } /* * innvl: { * "initialize_command" -> POOL_INITIALIZE_{CANCEL|START|SUSPEND} (uint64) * "initialize_vdevs": { -> guids to initialize (nvlist) * "vdev_path_1": vdev_guid_1, (uint64), * "vdev_path_2": vdev_guid_2, (uint64), * ... * }, * } * * outnvl: { * "initialize_vdevs": { -> initialization errors (nvlist) * "vdev_path_1": errno, see function body for possible errnos (uint64) * "vdev_path_2": errno, ... (uint64) * ... * } * } * * EINVAL is returned for an unknown commands or if any of the provided vdev * guids have be specified with a type other than uint64. */ static const zfs_ioc_key_t zfs_keys_pool_initialize[] = { {ZPOOL_INITIALIZE_COMMAND, DATA_TYPE_UINT64, 0}, {ZPOOL_INITIALIZE_VDEVS, DATA_TYPE_NVLIST, 0} }; static int zfs_ioc_pool_initialize(const char *poolname, nvlist_t *innvl, nvlist_t *outnvl) { uint64_t cmd_type; if (nvlist_lookup_uint64(innvl, ZPOOL_INITIALIZE_COMMAND, &cmd_type) != 0) { return (SET_ERROR(EINVAL)); } if (!(cmd_type == POOL_INITIALIZE_CANCEL || cmd_type == POOL_INITIALIZE_START || - cmd_type == POOL_INITIALIZE_SUSPEND)) { + cmd_type == POOL_INITIALIZE_SUSPEND || + cmd_type == POOL_INITIALIZE_UNINIT)) { return (SET_ERROR(EINVAL)); } nvlist_t *vdev_guids; if (nvlist_lookup_nvlist(innvl, ZPOOL_INITIALIZE_VDEVS, &vdev_guids) != 0) { return (SET_ERROR(EINVAL)); } for (nvpair_t *pair = nvlist_next_nvpair(vdev_guids, NULL); pair != NULL; pair = nvlist_next_nvpair(vdev_guids, pair)) { uint64_t vdev_guid; if (nvpair_value_uint64(pair, &vdev_guid) != 0) { return (SET_ERROR(EINVAL)); } } spa_t *spa; int error = spa_open(poolname, &spa, FTAG); if (error != 0) return (error); nvlist_t *vdev_errlist = fnvlist_alloc(); int total_errors = spa_vdev_initialize(spa, vdev_guids, cmd_type, vdev_errlist); if (fnvlist_size(vdev_errlist) > 0) { fnvlist_add_nvlist(outnvl, ZPOOL_INITIALIZE_VDEVS, vdev_errlist); } fnvlist_free(vdev_errlist); spa_close(spa, FTAG); return (total_errors > 0 ? SET_ERROR(EINVAL) : 0); } /* * innvl: { * "trim_command" -> POOL_TRIM_{CANCEL|START|SUSPEND} (uint64) * "trim_vdevs": { -> guids to TRIM (nvlist) * "vdev_path_1": vdev_guid_1, (uint64), * "vdev_path_2": vdev_guid_2, (uint64), * ... * }, * "trim_rate" -> Target TRIM rate in bytes/sec. * "trim_secure" -> Set to request a secure TRIM. * } * * outnvl: { * "trim_vdevs": { -> TRIM errors (nvlist) * "vdev_path_1": errno, see function body for possible errnos (uint64) * "vdev_path_2": errno, ... (uint64) * ... * } * } * * EINVAL is returned for an unknown commands or if any of the provided vdev * guids have be specified with a type other than uint64. */ static const zfs_ioc_key_t zfs_keys_pool_trim[] = { {ZPOOL_TRIM_COMMAND, DATA_TYPE_UINT64, 0}, {ZPOOL_TRIM_VDEVS, DATA_TYPE_NVLIST, 0}, {ZPOOL_TRIM_RATE, DATA_TYPE_UINT64, ZK_OPTIONAL}, {ZPOOL_TRIM_SECURE, DATA_TYPE_BOOLEAN_VALUE, ZK_OPTIONAL}, }; static int zfs_ioc_pool_trim(const char *poolname, nvlist_t *innvl, nvlist_t *outnvl) { uint64_t cmd_type; if (nvlist_lookup_uint64(innvl, ZPOOL_TRIM_COMMAND, &cmd_type) != 0) return (SET_ERROR(EINVAL)); if (!(cmd_type == POOL_TRIM_CANCEL || cmd_type == POOL_TRIM_START || cmd_type == POOL_TRIM_SUSPEND)) { return (SET_ERROR(EINVAL)); } nvlist_t *vdev_guids; if (nvlist_lookup_nvlist(innvl, ZPOOL_TRIM_VDEVS, &vdev_guids) != 0) return (SET_ERROR(EINVAL)); for (nvpair_t *pair = nvlist_next_nvpair(vdev_guids, NULL); pair != NULL; pair = nvlist_next_nvpair(vdev_guids, pair)) { uint64_t vdev_guid; if (nvpair_value_uint64(pair, &vdev_guid) != 0) { return (SET_ERROR(EINVAL)); } } /* Optional, defaults to maximum rate when not provided */ uint64_t rate; if (nvlist_lookup_uint64(innvl, ZPOOL_TRIM_RATE, &rate) != 0) rate = 0; /* Optional, defaults to standard TRIM when not provided */ boolean_t secure; if (nvlist_lookup_boolean_value(innvl, ZPOOL_TRIM_SECURE, &secure) != 0) { secure = B_FALSE; } spa_t *spa; int error = spa_open(poolname, &spa, FTAG); if (error != 0) return (error); nvlist_t *vdev_errlist = fnvlist_alloc(); int total_errors = spa_vdev_trim(spa, vdev_guids, cmd_type, rate, !!zfs_trim_metaslab_skip, secure, vdev_errlist); if (fnvlist_size(vdev_errlist) > 0) fnvlist_add_nvlist(outnvl, ZPOOL_TRIM_VDEVS, vdev_errlist); fnvlist_free(vdev_errlist); spa_close(spa, FTAG); return (total_errors > 0 ? SET_ERROR(EINVAL) : 0); } /* * This ioctl waits for activity of a particular type to complete. If there is * no activity of that type in progress, it returns immediately, and the * returned value "waited" is false. If there is activity in progress, and no * tag is passed in, the ioctl blocks until all activity of that type is * complete, and then returns with "waited" set to true. * * If a tag is provided, it identifies a particular instance of an activity to * wait for. Currently, this is only valid for use with 'initialize', because * that is the only activity for which there can be multiple instances running * concurrently. In the case of 'initialize', the tag corresponds to the guid of * the vdev on which to wait. * * If a thread waiting in the ioctl receives a signal, the call will return * immediately, and the return value will be EINTR. * * innvl: { * "wait_activity" -> int32_t * (optional) "wait_tag" -> uint64_t * } * * outnvl: "waited" -> boolean_t */ static const zfs_ioc_key_t zfs_keys_pool_wait[] = { {ZPOOL_WAIT_ACTIVITY, DATA_TYPE_INT32, 0}, {ZPOOL_WAIT_TAG, DATA_TYPE_UINT64, ZK_OPTIONAL}, }; static int zfs_ioc_wait(const char *name, nvlist_t *innvl, nvlist_t *outnvl) { int32_t activity; uint64_t tag; boolean_t waited; int error; if (nvlist_lookup_int32(innvl, ZPOOL_WAIT_ACTIVITY, &activity) != 0) return (EINVAL); if (nvlist_lookup_uint64(innvl, ZPOOL_WAIT_TAG, &tag) == 0) error = spa_wait_tag(name, activity, tag, &waited); else error = spa_wait(name, activity, &waited); if (error == 0) fnvlist_add_boolean_value(outnvl, ZPOOL_WAIT_WAITED, waited); return (error); } /* * This ioctl waits for activity of a particular type to complete. If there is * no activity of that type in progress, it returns immediately, and the * returned value "waited" is false. If there is activity in progress, and no * tag is passed in, the ioctl blocks until all activity of that type is * complete, and then returns with "waited" set to true. * * If a thread waiting in the ioctl receives a signal, the call will return * immediately, and the return value will be EINTR. * * innvl: { * "wait_activity" -> int32_t * } * * outnvl: "waited" -> boolean_t */ static const zfs_ioc_key_t zfs_keys_fs_wait[] = { {ZFS_WAIT_ACTIVITY, DATA_TYPE_INT32, 0}, }; static int zfs_ioc_wait_fs(const char *name, nvlist_t *innvl, nvlist_t *outnvl) { int32_t activity; boolean_t waited = B_FALSE; int error; dsl_pool_t *dp; dsl_dir_t *dd; dsl_dataset_t *ds; if (nvlist_lookup_int32(innvl, ZFS_WAIT_ACTIVITY, &activity) != 0) return (SET_ERROR(EINVAL)); if (activity >= ZFS_WAIT_NUM_ACTIVITIES || activity < 0) return (SET_ERROR(EINVAL)); if ((error = dsl_pool_hold(name, FTAG, &dp)) != 0) return (error); if ((error = dsl_dataset_hold(dp, name, FTAG, &ds)) != 0) { dsl_pool_rele(dp, FTAG); return (error); } dd = ds->ds_dir; mutex_enter(&dd->dd_activity_lock); dd->dd_activity_waiters++; /* * We get a long-hold here so that the dsl_dataset_t and dsl_dir_t * aren't evicted while we're waiting. Normally this is prevented by * holding the pool, but we can't do that while we're waiting since * that would prevent TXGs from syncing out. Some of the functionality * of long-holds (e.g. preventing deletion) is unnecessary for this * case, since we would cancel the waiters before proceeding with a * deletion. An alternative mechanism for keeping the dataset around * could be developed but this is simpler. */ dsl_dataset_long_hold(ds, FTAG); dsl_pool_rele(dp, FTAG); error = dsl_dir_wait(dd, ds, activity, &waited); dsl_dataset_long_rele(ds, FTAG); dd->dd_activity_waiters--; if (dd->dd_activity_waiters == 0) cv_signal(&dd->dd_activity_cv); mutex_exit(&dd->dd_activity_lock); dsl_dataset_rele(ds, FTAG); if (error == 0) fnvlist_add_boolean_value(outnvl, ZFS_WAIT_WAITED, waited); return (error); } /* * fsname is name of dataset to rollback (to most recent snapshot) * * innvl may contain name of expected target snapshot * * outnvl: "target" -> name of most recent snapshot * } */ static const zfs_ioc_key_t zfs_keys_rollback[] = { {"target", DATA_TYPE_STRING, ZK_OPTIONAL}, }; static int zfs_ioc_rollback(const char *fsname, nvlist_t *innvl, nvlist_t *outnvl) { zfsvfs_t *zfsvfs; zvol_state_handle_t *zv; const char *target = NULL; int error; (void) nvlist_lookup_string(innvl, "target", &target); if (target != NULL) { const char *cp = strchr(target, '@'); /* * The snap name must contain an @, and the part after it must * contain only valid characters. */ if (cp == NULL || zfs_component_namecheck(cp + 1, NULL, NULL) != 0) return (SET_ERROR(EINVAL)); } if (getzfsvfs(fsname, &zfsvfs) == 0) { dsl_dataset_t *ds; ds = dmu_objset_ds(zfsvfs->z_os); error = zfs_suspend_fs(zfsvfs); if (error == 0) { int resume_err; error = dsl_dataset_rollback(fsname, target, zfsvfs, outnvl); resume_err = zfs_resume_fs(zfsvfs, ds); error = error ? error : resume_err; } zfs_vfs_rele(zfsvfs); } else if ((zv = zvol_suspend(fsname)) != NULL) { error = dsl_dataset_rollback(fsname, target, zvol_tag(zv), outnvl); zvol_resume(zv); } else { error = dsl_dataset_rollback(fsname, target, NULL, outnvl); } return (error); } static int recursive_unmount(const char *fsname, void *arg) { const char *snapname = arg; char *fullname; fullname = kmem_asprintf("%s@%s", fsname, snapname); zfs_unmount_snap(fullname); kmem_strfree(fullname); return (0); } /* * * snapname is the snapshot to redact. * innvl: { * "bookname" -> (string) * shortname of the redaction bookmark to generate * "snapnv" -> (nvlist, values ignored) * snapshots to redact snapname with respect to * } * * outnvl is unused */ static const zfs_ioc_key_t zfs_keys_redact[] = { {"bookname", DATA_TYPE_STRING, 0}, {"snapnv", DATA_TYPE_NVLIST, 0}, }; static int zfs_ioc_redact(const char *snapname, nvlist_t *innvl, nvlist_t *outnvl) { (void) outnvl; nvlist_t *redactnvl = NULL; const char *redactbook = NULL; if (nvlist_lookup_nvlist(innvl, "snapnv", &redactnvl) != 0) return (SET_ERROR(EINVAL)); if (fnvlist_num_pairs(redactnvl) == 0) return (SET_ERROR(ENXIO)); if (nvlist_lookup_string(innvl, "bookname", &redactbook) != 0) return (SET_ERROR(EINVAL)); return (dmu_redact_snap(snapname, redactnvl, redactbook)); } /* * inputs: * zc_name old name of dataset * zc_value new name of dataset * zc_cookie recursive flag (only valid for snapshots) * * outputs: none */ static int zfs_ioc_rename(zfs_cmd_t *zc) { objset_t *os; dmu_objset_type_t ost; boolean_t recursive = zc->zc_cookie & 1; boolean_t nounmount = !!(zc->zc_cookie & 2); char *at; int err; /* "zfs rename" from and to ...%recv datasets should both fail */ zc->zc_name[sizeof (zc->zc_name) - 1] = '\0'; zc->zc_value[sizeof (zc->zc_value) - 1] = '\0'; if (dataset_namecheck(zc->zc_name, NULL, NULL) != 0 || dataset_namecheck(zc->zc_value, NULL, NULL) != 0 || strchr(zc->zc_name, '%') || strchr(zc->zc_value, '%')) return (SET_ERROR(EINVAL)); err = dmu_objset_hold(zc->zc_name, FTAG, &os); if (err != 0) return (err); ost = dmu_objset_type(os); dmu_objset_rele(os, FTAG); at = strchr(zc->zc_name, '@'); if (at != NULL) { /* snaps must be in same fs */ int error; if (strncmp(zc->zc_name, zc->zc_value, at - zc->zc_name + 1)) return (SET_ERROR(EXDEV)); *at = '\0'; if (ost == DMU_OST_ZFS && !nounmount) { error = dmu_objset_find(zc->zc_name, recursive_unmount, at + 1, recursive ? DS_FIND_CHILDREN : 0); if (error != 0) { *at = '@'; return (error); } } error = dsl_dataset_rename_snapshot(zc->zc_name, at + 1, strchr(zc->zc_value, '@') + 1, recursive); *at = '@'; return (error); } else { return (dsl_dir_rename(zc->zc_name, zc->zc_value)); } } static int zfs_check_settable(const char *dsname, nvpair_t *pair, cred_t *cr) { const char *propname = nvpair_name(pair); boolean_t issnap = (strchr(dsname, '@') != NULL); zfs_prop_t prop = zfs_name_to_prop(propname); uint64_t intval, compval; int err; if (prop == ZPROP_USERPROP) { if (zfs_prop_user(propname)) { if ((err = zfs_secpolicy_write_perms(dsname, ZFS_DELEG_PERM_USERPROP, cr))) return (err); return (0); } if (!issnap && zfs_prop_userquota(propname)) { const char *perm = NULL; const char *uq_prefix = zfs_userquota_prop_prefixes[ZFS_PROP_USERQUOTA]; const char *gq_prefix = zfs_userquota_prop_prefixes[ZFS_PROP_GROUPQUOTA]; const char *uiq_prefix = zfs_userquota_prop_prefixes[ZFS_PROP_USEROBJQUOTA]; const char *giq_prefix = zfs_userquota_prop_prefixes[ZFS_PROP_GROUPOBJQUOTA]; const char *pq_prefix = zfs_userquota_prop_prefixes[ZFS_PROP_PROJECTQUOTA]; const char *piq_prefix = zfs_userquota_prop_prefixes[\ ZFS_PROP_PROJECTOBJQUOTA]; if (strncmp(propname, uq_prefix, strlen(uq_prefix)) == 0) { perm = ZFS_DELEG_PERM_USERQUOTA; } else if (strncmp(propname, uiq_prefix, strlen(uiq_prefix)) == 0) { perm = ZFS_DELEG_PERM_USEROBJQUOTA; } else if (strncmp(propname, gq_prefix, strlen(gq_prefix)) == 0) { perm = ZFS_DELEG_PERM_GROUPQUOTA; } else if (strncmp(propname, giq_prefix, strlen(giq_prefix)) == 0) { perm = ZFS_DELEG_PERM_GROUPOBJQUOTA; } else if (strncmp(propname, pq_prefix, strlen(pq_prefix)) == 0) { perm = ZFS_DELEG_PERM_PROJECTQUOTA; } else if (strncmp(propname, piq_prefix, strlen(piq_prefix)) == 0) { perm = ZFS_DELEG_PERM_PROJECTOBJQUOTA; } else { /* {USER|GROUP|PROJECT}USED are read-only */ return (SET_ERROR(EINVAL)); } if ((err = zfs_secpolicy_write_perms(dsname, perm, cr))) return (err); return (0); } return (SET_ERROR(EINVAL)); } if (issnap) return (SET_ERROR(EINVAL)); if (nvpair_type(pair) == DATA_TYPE_NVLIST) { /* * dsl_prop_get_all_impl() returns properties in this * format. */ nvlist_t *attrs; VERIFY(nvpair_value_nvlist(pair, &attrs) == 0); VERIFY(nvlist_lookup_nvpair(attrs, ZPROP_VALUE, &pair) == 0); } /* * Check that this value is valid for this pool version */ switch (prop) { case ZFS_PROP_COMPRESSION: /* * If the user specified gzip compression, make sure * the SPA supports it. We ignore any errors here since * we'll catch them later. */ if (nvpair_value_uint64(pair, &intval) == 0) { compval = ZIO_COMPRESS_ALGO(intval); if (compval >= ZIO_COMPRESS_GZIP_1 && compval <= ZIO_COMPRESS_GZIP_9 && zfs_earlier_version(dsname, SPA_VERSION_GZIP_COMPRESSION)) { return (SET_ERROR(ENOTSUP)); } if (compval == ZIO_COMPRESS_ZLE && zfs_earlier_version(dsname, SPA_VERSION_ZLE_COMPRESSION)) return (SET_ERROR(ENOTSUP)); if (compval == ZIO_COMPRESS_LZ4) { spa_t *spa; if ((err = spa_open(dsname, &spa, FTAG)) != 0) return (err); if (!spa_feature_is_enabled(spa, SPA_FEATURE_LZ4_COMPRESS)) { spa_close(spa, FTAG); return (SET_ERROR(ENOTSUP)); } spa_close(spa, FTAG); } if (compval == ZIO_COMPRESS_ZSTD) { spa_t *spa; if ((err = spa_open(dsname, &spa, FTAG)) != 0) return (err); if (!spa_feature_is_enabled(spa, SPA_FEATURE_ZSTD_COMPRESS)) { spa_close(spa, FTAG); return (SET_ERROR(ENOTSUP)); } spa_close(spa, FTAG); } } break; case ZFS_PROP_COPIES: if (zfs_earlier_version(dsname, SPA_VERSION_DITTO_BLOCKS)) return (SET_ERROR(ENOTSUP)); break; case ZFS_PROP_VOLBLOCKSIZE: case ZFS_PROP_RECORDSIZE: /* Record sizes above 128k need the feature to be enabled */ if (nvpair_value_uint64(pair, &intval) == 0 && intval > SPA_OLD_MAXBLOCKSIZE) { spa_t *spa; /* * We don't allow setting the property above 1MB, * unless the tunable has been changed. */ if (intval > zfs_max_recordsize || intval > SPA_MAXBLOCKSIZE) return (SET_ERROR(ERANGE)); if ((err = spa_open(dsname, &spa, FTAG)) != 0) return (err); if (!spa_feature_is_enabled(spa, SPA_FEATURE_LARGE_BLOCKS)) { spa_close(spa, FTAG); return (SET_ERROR(ENOTSUP)); } spa_close(spa, FTAG); } break; case ZFS_PROP_DNODESIZE: /* Dnode sizes above 512 need the feature to be enabled */ if (nvpair_value_uint64(pair, &intval) == 0 && intval != ZFS_DNSIZE_LEGACY) { spa_t *spa; if ((err = spa_open(dsname, &spa, FTAG)) != 0) return (err); if (!spa_feature_is_enabled(spa, SPA_FEATURE_LARGE_DNODE)) { spa_close(spa, FTAG); return (SET_ERROR(ENOTSUP)); } spa_close(spa, FTAG); } break; case ZFS_PROP_SPECIAL_SMALL_BLOCKS: /* * This property could require the allocation classes * feature to be active for setting, however we allow * it so that tests of settable properties succeed. * The CLI will issue a warning in this case. */ break; case ZFS_PROP_SHARESMB: if (zpl_earlier_version(dsname, ZPL_VERSION_FUID)) return (SET_ERROR(ENOTSUP)); break; case ZFS_PROP_ACLINHERIT: if (nvpair_type(pair) == DATA_TYPE_UINT64 && nvpair_value_uint64(pair, &intval) == 0) { if (intval == ZFS_ACL_PASSTHROUGH_X && zfs_earlier_version(dsname, SPA_VERSION_PASSTHROUGH_X)) return (SET_ERROR(ENOTSUP)); } break; case ZFS_PROP_CHECKSUM: case ZFS_PROP_DEDUP: { spa_feature_t feature; spa_t *spa; int err; /* dedup feature version checks */ if (prop == ZFS_PROP_DEDUP && zfs_earlier_version(dsname, SPA_VERSION_DEDUP)) return (SET_ERROR(ENOTSUP)); if (nvpair_type(pair) == DATA_TYPE_UINT64 && nvpair_value_uint64(pair, &intval) == 0) { /* check prop value is enabled in features */ feature = zio_checksum_to_feature( intval & ZIO_CHECKSUM_MASK); if (feature == SPA_FEATURE_NONE) break; if ((err = spa_open(dsname, &spa, FTAG)) != 0) return (err); if (!spa_feature_is_enabled(spa, feature)) { spa_close(spa, FTAG); return (SET_ERROR(ENOTSUP)); } spa_close(spa, FTAG); } break; } default: break; } return (zfs_secpolicy_setprop(dsname, prop, pair, CRED())); } /* * Removes properties from the given props list that fail permission checks * needed to clear them and to restore them in case of a receive error. For each * property, make sure we have both set and inherit permissions. * * Returns the first error encountered if any permission checks fail. If the * caller provides a non-NULL errlist, it also gives the complete list of names * of all the properties that failed a permission check along with the * corresponding error numbers. The caller is responsible for freeing the * returned errlist. * * If every property checks out successfully, zero is returned and the list * pointed at by errlist is NULL. */ static int zfs_check_clearable(const char *dataset, nvlist_t *props, nvlist_t **errlist) { zfs_cmd_t *zc; nvpair_t *pair, *next_pair; nvlist_t *errors; int err, rv = 0; if (props == NULL) return (0); VERIFY(nvlist_alloc(&errors, NV_UNIQUE_NAME, KM_SLEEP) == 0); zc = kmem_alloc(sizeof (zfs_cmd_t), KM_SLEEP); (void) strlcpy(zc->zc_name, dataset, sizeof (zc->zc_name)); pair = nvlist_next_nvpair(props, NULL); while (pair != NULL) { next_pair = nvlist_next_nvpair(props, pair); (void) strlcpy(zc->zc_value, nvpair_name(pair), sizeof (zc->zc_value)); if ((err = zfs_check_settable(dataset, pair, CRED())) != 0 || (err = zfs_secpolicy_inherit_prop(zc, NULL, CRED())) != 0) { VERIFY(nvlist_remove_nvpair(props, pair) == 0); VERIFY(nvlist_add_int32(errors, zc->zc_value, err) == 0); } pair = next_pair; } kmem_free(zc, sizeof (zfs_cmd_t)); if ((pair = nvlist_next_nvpair(errors, NULL)) == NULL) { nvlist_free(errors); errors = NULL; } else { VERIFY(nvpair_value_int32(pair, &rv) == 0); } if (errlist == NULL) nvlist_free(errors); else *errlist = errors; return (rv); } static boolean_t propval_equals(nvpair_t *p1, nvpair_t *p2) { if (nvpair_type(p1) == DATA_TYPE_NVLIST) { /* dsl_prop_get_all_impl() format */ nvlist_t *attrs; VERIFY(nvpair_value_nvlist(p1, &attrs) == 0); VERIFY(nvlist_lookup_nvpair(attrs, ZPROP_VALUE, &p1) == 0); } if (nvpair_type(p2) == DATA_TYPE_NVLIST) { nvlist_t *attrs; VERIFY(nvpair_value_nvlist(p2, &attrs) == 0); VERIFY(nvlist_lookup_nvpair(attrs, ZPROP_VALUE, &p2) == 0); } if (nvpair_type(p1) != nvpair_type(p2)) return (B_FALSE); if (nvpair_type(p1) == DATA_TYPE_STRING) { const char *valstr1, *valstr2; VERIFY(nvpair_value_string(p1, &valstr1) == 0); VERIFY(nvpair_value_string(p2, &valstr2) == 0); return (strcmp(valstr1, valstr2) == 0); } else { uint64_t intval1, intval2; VERIFY(nvpair_value_uint64(p1, &intval1) == 0); VERIFY(nvpair_value_uint64(p2, &intval2) == 0); return (intval1 == intval2); } } /* * Remove properties from props if they are not going to change (as determined * by comparison with origprops). Remove them from origprops as well, since we * do not need to clear or restore properties that won't change. */ static void props_reduce(nvlist_t *props, nvlist_t *origprops) { nvpair_t *pair, *next_pair; if (origprops == NULL) return; /* all props need to be received */ pair = nvlist_next_nvpair(props, NULL); while (pair != NULL) { const char *propname = nvpair_name(pair); nvpair_t *match; next_pair = nvlist_next_nvpair(props, pair); if ((nvlist_lookup_nvpair(origprops, propname, &match) != 0) || !propval_equals(pair, match)) goto next; /* need to set received value */ /* don't clear the existing received value */ (void) nvlist_remove_nvpair(origprops, match); /* don't bother receiving the property */ (void) nvlist_remove_nvpair(props, pair); next: pair = next_pair; } } /* * Extract properties that cannot be set PRIOR to the receipt of a dataset. * For example, refquota cannot be set until after the receipt of a dataset, * because in replication streams, an older/earlier snapshot may exceed the * refquota. We want to receive the older/earlier snapshot, but setting * refquota pre-receipt will set the dsl's ACTUAL quota, which will prevent * the older/earlier snapshot from being received (with EDQUOT). * * The ZFS test "zfs_receive_011_pos" demonstrates such a scenario. * * libzfs will need to be judicious handling errors encountered by props * extracted by this function. */ static nvlist_t * extract_delay_props(nvlist_t *props) { nvlist_t *delayprops; nvpair_t *nvp, *tmp; static const zfs_prop_t delayable[] = { ZFS_PROP_REFQUOTA, ZFS_PROP_KEYLOCATION, /* * Setting ZFS_PROP_SHARESMB requires the objset type to be * known, which is not possible prior to receipt of raw sends. */ ZFS_PROP_SHARESMB, 0 }; int i; VERIFY(nvlist_alloc(&delayprops, NV_UNIQUE_NAME, KM_SLEEP) == 0); for (nvp = nvlist_next_nvpair(props, NULL); nvp != NULL; nvp = nvlist_next_nvpair(props, nvp)) { /* * strcmp() is safe because zfs_prop_to_name() always returns * a bounded string. */ for (i = 0; delayable[i] != 0; i++) { if (strcmp(zfs_prop_to_name(delayable[i]), nvpair_name(nvp)) == 0) { break; } } if (delayable[i] != 0) { tmp = nvlist_prev_nvpair(props, nvp); VERIFY(nvlist_add_nvpair(delayprops, nvp) == 0); VERIFY(nvlist_remove_nvpair(props, nvp) == 0); nvp = tmp; } } if (nvlist_empty(delayprops)) { nvlist_free(delayprops); delayprops = NULL; } return (delayprops); } static void zfs_allow_log_destroy(void *arg) { char *poolname = arg; if (poolname != NULL) kmem_strfree(poolname); } #ifdef ZFS_DEBUG static boolean_t zfs_ioc_recv_inject_err; #endif /* * nvlist 'errors' is always allocated. It will contain descriptions of * encountered errors, if any. It's the callers responsibility to free. */ static int zfs_ioc_recv_impl(char *tofs, char *tosnap, const char *origin, nvlist_t *recvprops, nvlist_t *localprops, nvlist_t *hidden_args, boolean_t force, boolean_t heal, boolean_t resumable, int input_fd, dmu_replay_record_t *begin_record, uint64_t *read_bytes, uint64_t *errflags, nvlist_t **errors) { dmu_recv_cookie_t drc; int error = 0; int props_error = 0; offset_t off, noff; nvlist_t *local_delayprops = NULL; nvlist_t *recv_delayprops = NULL; nvlist_t *inherited_delayprops = NULL; nvlist_t *origprops = NULL; /* existing properties */ nvlist_t *origrecvd = NULL; /* existing received properties */ boolean_t first_recvd_props = B_FALSE; boolean_t tofs_was_redacted; zfs_file_t *input_fp; *read_bytes = 0; *errflags = 0; *errors = fnvlist_alloc(); off = 0; if ((input_fp = zfs_file_get(input_fd)) == NULL) return (SET_ERROR(EBADF)); noff = off = zfs_file_off(input_fp); error = dmu_recv_begin(tofs, tosnap, begin_record, force, heal, resumable, localprops, hidden_args, origin, &drc, input_fp, &off); if (error != 0) goto out; tofs_was_redacted = dsl_get_redacted(drc.drc_ds); /* * Set properties before we receive the stream so that they are applied * to the new data. Note that we must call dmu_recv_stream() if * dmu_recv_begin() succeeds. */ if (recvprops != NULL && !drc.drc_newfs) { if (spa_version(dsl_dataset_get_spa(drc.drc_ds)) >= SPA_VERSION_RECVD_PROPS && !dsl_prop_get_hasrecvd(tofs)) first_recvd_props = B_TRUE; /* * If new received properties are supplied, they are to * completely replace the existing received properties, * so stash away the existing ones. */ if (dsl_prop_get_received(tofs, &origrecvd) == 0) { nvlist_t *errlist = NULL; /* * Don't bother writing a property if its value won't * change (and avoid the unnecessary security checks). * * The first receive after SPA_VERSION_RECVD_PROPS is a * special case where we blow away all local properties * regardless. */ if (!first_recvd_props) props_reduce(recvprops, origrecvd); if (zfs_check_clearable(tofs, origrecvd, &errlist) != 0) (void) nvlist_merge(*errors, errlist, 0); nvlist_free(errlist); if (clear_received_props(tofs, origrecvd, first_recvd_props ? NULL : recvprops) != 0) *errflags |= ZPROP_ERR_NOCLEAR; } else { *errflags |= ZPROP_ERR_NOCLEAR; } } /* * Stash away existing properties so we can restore them on error unless * we're doing the first receive after SPA_VERSION_RECVD_PROPS, in which * case "origrecvd" will take care of that. */ if (localprops != NULL && !drc.drc_newfs && !first_recvd_props) { objset_t *os; if (dmu_objset_hold(tofs, FTAG, &os) == 0) { if (dsl_prop_get_all(os, &origprops) != 0) { *errflags |= ZPROP_ERR_NOCLEAR; } dmu_objset_rele(os, FTAG); } else { *errflags |= ZPROP_ERR_NOCLEAR; } } if (recvprops != NULL) { props_error = dsl_prop_set_hasrecvd(tofs); if (props_error == 0) { recv_delayprops = extract_delay_props(recvprops); (void) zfs_set_prop_nvlist(tofs, ZPROP_SRC_RECEIVED, recvprops, *errors); } } if (localprops != NULL) { nvlist_t *oprops = fnvlist_alloc(); nvlist_t *xprops = fnvlist_alloc(); nvpair_t *nvp = NULL; while ((nvp = nvlist_next_nvpair(localprops, nvp)) != NULL) { if (nvpair_type(nvp) == DATA_TYPE_BOOLEAN) { /* -x property */ const char *name = nvpair_name(nvp); zfs_prop_t prop = zfs_name_to_prop(name); if (prop != ZPROP_USERPROP) { if (!zfs_prop_inheritable(prop)) continue; } else if (!zfs_prop_user(name)) continue; fnvlist_add_boolean(xprops, name); } else { /* -o property=value */ fnvlist_add_nvpair(oprops, nvp); } } local_delayprops = extract_delay_props(oprops); (void) zfs_set_prop_nvlist(tofs, ZPROP_SRC_LOCAL, oprops, *errors); inherited_delayprops = extract_delay_props(xprops); (void) zfs_set_prop_nvlist(tofs, ZPROP_SRC_INHERITED, xprops, *errors); nvlist_free(oprops); nvlist_free(xprops); } error = dmu_recv_stream(&drc, &off); if (error == 0) { zfsvfs_t *zfsvfs = NULL; zvol_state_handle_t *zv = NULL; if (getzfsvfs(tofs, &zfsvfs) == 0) { /* online recv */ dsl_dataset_t *ds; int end_err; boolean_t stream_is_redacted = DMU_GET_FEATUREFLAGS( begin_record->drr_u.drr_begin. drr_versioninfo) & DMU_BACKUP_FEATURE_REDACTED; ds = dmu_objset_ds(zfsvfs->z_os); error = zfs_suspend_fs(zfsvfs); /* * If the suspend fails, then the recv_end will * likely also fail, and clean up after itself. */ end_err = dmu_recv_end(&drc, zfsvfs); /* * If the dataset was not redacted, but we received a * redacted stream onto it, we need to unmount the * dataset. Otherwise, resume the filesystem. */ if (error == 0 && !drc.drc_newfs && stream_is_redacted && !tofs_was_redacted) { error = zfs_end_fs(zfsvfs, ds); } else if (error == 0) { error = zfs_resume_fs(zfsvfs, ds); } error = error ? error : end_err; zfs_vfs_rele(zfsvfs); } else if ((zv = zvol_suspend(tofs)) != NULL) { error = dmu_recv_end(&drc, zvol_tag(zv)); zvol_resume(zv); } else { error = dmu_recv_end(&drc, NULL); } /* Set delayed properties now, after we're done receiving. */ if (recv_delayprops != NULL && error == 0) { (void) zfs_set_prop_nvlist(tofs, ZPROP_SRC_RECEIVED, recv_delayprops, *errors); } if (local_delayprops != NULL && error == 0) { (void) zfs_set_prop_nvlist(tofs, ZPROP_SRC_LOCAL, local_delayprops, *errors); } if (inherited_delayprops != NULL && error == 0) { (void) zfs_set_prop_nvlist(tofs, ZPROP_SRC_INHERITED, inherited_delayprops, *errors); } } /* * Merge delayed props back in with initial props, in case * we're DEBUG and zfs_ioc_recv_inject_err is set (which means * we have to make sure clear_received_props() includes * the delayed properties). * * Since zfs_ioc_recv_inject_err is only in DEBUG kernels, * using ASSERT() will be just like a VERIFY. */ if (recv_delayprops != NULL) { ASSERT(nvlist_merge(recvprops, recv_delayprops, 0) == 0); nvlist_free(recv_delayprops); } if (local_delayprops != NULL) { ASSERT(nvlist_merge(localprops, local_delayprops, 0) == 0); nvlist_free(local_delayprops); } if (inherited_delayprops != NULL) { ASSERT(nvlist_merge(localprops, inherited_delayprops, 0) == 0); nvlist_free(inherited_delayprops); } *read_bytes = off - noff; #ifdef ZFS_DEBUG if (zfs_ioc_recv_inject_err) { zfs_ioc_recv_inject_err = B_FALSE; error = 1; } #endif /* * On error, restore the original props. */ if (error != 0 && recvprops != NULL && !drc.drc_newfs) { if (clear_received_props(tofs, recvprops, NULL) != 0) { /* * We failed to clear the received properties. * Since we may have left a $recvd value on the * system, we can't clear the $hasrecvd flag. */ *errflags |= ZPROP_ERR_NORESTORE; } else if (first_recvd_props) { dsl_prop_unset_hasrecvd(tofs); } if (origrecvd == NULL && !drc.drc_newfs) { /* We failed to stash the original properties. */ *errflags |= ZPROP_ERR_NORESTORE; } /* * dsl_props_set() will not convert RECEIVED to LOCAL on or * after SPA_VERSION_RECVD_PROPS, so we need to specify LOCAL * explicitly if we're restoring local properties cleared in the * first new-style receive. */ if (origrecvd != NULL && zfs_set_prop_nvlist(tofs, (first_recvd_props ? ZPROP_SRC_LOCAL : ZPROP_SRC_RECEIVED), origrecvd, NULL) != 0) { /* * We stashed the original properties but failed to * restore them. */ *errflags |= ZPROP_ERR_NORESTORE; } } if (error != 0 && localprops != NULL && !drc.drc_newfs && !first_recvd_props) { nvlist_t *setprops; nvlist_t *inheritprops; nvpair_t *nvp; if (origprops == NULL) { /* We failed to stash the original properties. */ *errflags |= ZPROP_ERR_NORESTORE; goto out; } /* Restore original props */ setprops = fnvlist_alloc(); inheritprops = fnvlist_alloc(); nvp = NULL; while ((nvp = nvlist_next_nvpair(localprops, nvp)) != NULL) { const char *name = nvpair_name(nvp); const char *source; nvlist_t *attrs; if (!nvlist_exists(origprops, name)) { /* * Property was not present or was explicitly * inherited before the receive, restore this. */ fnvlist_add_boolean(inheritprops, name); continue; } attrs = fnvlist_lookup_nvlist(origprops, name); source = fnvlist_lookup_string(attrs, ZPROP_SOURCE); /* Skip received properties */ if (strcmp(source, ZPROP_SOURCE_VAL_RECVD) == 0) continue; if (strcmp(source, tofs) == 0) { /* Property was locally set */ fnvlist_add_nvlist(setprops, name, attrs); } else { /* Property was implicitly inherited */ fnvlist_add_boolean(inheritprops, name); } } if (zfs_set_prop_nvlist(tofs, ZPROP_SRC_LOCAL, setprops, NULL) != 0) *errflags |= ZPROP_ERR_NORESTORE; if (zfs_set_prop_nvlist(tofs, ZPROP_SRC_INHERITED, inheritprops, NULL) != 0) *errflags |= ZPROP_ERR_NORESTORE; nvlist_free(setprops); nvlist_free(inheritprops); } out: zfs_file_put(input_fp); nvlist_free(origrecvd); nvlist_free(origprops); if (error == 0) error = props_error; return (error); } /* * inputs: * zc_name name of containing filesystem (unused) * zc_nvlist_src{_size} nvlist of properties to apply * zc_nvlist_conf{_size} nvlist of properties to exclude * (DATA_TYPE_BOOLEAN) and override (everything else) * zc_value name of snapshot to create * zc_string name of clone origin (if DRR_FLAG_CLONE) * zc_cookie file descriptor to recv from * zc_begin_record the BEGIN record of the stream (not byteswapped) * zc_guid force flag * * outputs: * zc_cookie number of bytes read * zc_obj zprop_errflags_t * zc_nvlist_dst{_size} error for each unapplied received property */ static int zfs_ioc_recv(zfs_cmd_t *zc) { dmu_replay_record_t begin_record; nvlist_t *errors = NULL; nvlist_t *recvdprops = NULL; nvlist_t *localprops = NULL; const char *origin = NULL; char *tosnap; char tofs[ZFS_MAX_DATASET_NAME_LEN]; int error = 0; if (dataset_namecheck(zc->zc_value, NULL, NULL) != 0 || strchr(zc->zc_value, '@') == NULL || strchr(zc->zc_value, '%')) return (SET_ERROR(EINVAL)); (void) strlcpy(tofs, zc->zc_value, sizeof (tofs)); tosnap = strchr(tofs, '@'); *tosnap++ = '\0'; if (zc->zc_nvlist_src != 0 && (error = get_nvlist(zc->zc_nvlist_src, zc->zc_nvlist_src_size, zc->zc_iflags, &recvdprops)) != 0) return (error); if (zc->zc_nvlist_conf != 0 && (error = get_nvlist(zc->zc_nvlist_conf, zc->zc_nvlist_conf_size, zc->zc_iflags, &localprops)) != 0) return (error); if (zc->zc_string[0]) origin = zc->zc_string; begin_record.drr_type = DRR_BEGIN; begin_record.drr_payloadlen = 0; begin_record.drr_u.drr_begin = zc->zc_begin_record; error = zfs_ioc_recv_impl(tofs, tosnap, origin, recvdprops, localprops, NULL, zc->zc_guid, B_FALSE, B_FALSE, zc->zc_cookie, &begin_record, &zc->zc_cookie, &zc->zc_obj, &errors); nvlist_free(recvdprops); nvlist_free(localprops); /* * Now that all props, initial and delayed, are set, report the prop * errors to the caller. */ if (zc->zc_nvlist_dst_size != 0 && errors != NULL && (nvlist_smush(errors, zc->zc_nvlist_dst_size) != 0 || put_nvlist(zc, errors) != 0)) { /* * Caller made zc->zc_nvlist_dst less than the minimum expected * size or supplied an invalid address. */ error = SET_ERROR(EINVAL); } nvlist_free(errors); return (error); } /* * innvl: { * "snapname" -> full name of the snapshot to create * (optional) "props" -> received properties to set (nvlist) * (optional) "localprops" -> override and exclude properties (nvlist) * (optional) "origin" -> name of clone origin (DRR_FLAG_CLONE) * "begin_record" -> non-byteswapped dmu_replay_record_t * "input_fd" -> file descriptor to read stream from (int32) * (optional) "force" -> force flag (value ignored) * (optional) "heal" -> use send stream to heal data corruption * (optional) "resumable" -> resumable flag (value ignored) * (optional) "cleanup_fd" -> unused * (optional) "action_handle" -> unused * (optional) "hidden_args" -> { "wkeydata" -> value } * } * * outnvl: { * "read_bytes" -> number of bytes read * "error_flags" -> zprop_errflags_t * "errors" -> error for each unapplied received property (nvlist) * } */ static const zfs_ioc_key_t zfs_keys_recv_new[] = { {"snapname", DATA_TYPE_STRING, 0}, {"props", DATA_TYPE_NVLIST, ZK_OPTIONAL}, {"localprops", DATA_TYPE_NVLIST, ZK_OPTIONAL}, {"origin", DATA_TYPE_STRING, ZK_OPTIONAL}, {"begin_record", DATA_TYPE_BYTE_ARRAY, 0}, {"input_fd", DATA_TYPE_INT32, 0}, {"force", DATA_TYPE_BOOLEAN, ZK_OPTIONAL}, {"heal", DATA_TYPE_BOOLEAN, ZK_OPTIONAL}, {"resumable", DATA_TYPE_BOOLEAN, ZK_OPTIONAL}, {"cleanup_fd", DATA_TYPE_INT32, ZK_OPTIONAL}, {"action_handle", DATA_TYPE_UINT64, ZK_OPTIONAL}, {"hidden_args", DATA_TYPE_NVLIST, ZK_OPTIONAL}, }; static int zfs_ioc_recv_new(const char *fsname, nvlist_t *innvl, nvlist_t *outnvl) { dmu_replay_record_t *begin_record; uint_t begin_record_size; nvlist_t *errors = NULL; nvlist_t *recvprops = NULL; nvlist_t *localprops = NULL; nvlist_t *hidden_args = NULL; const char *snapname; const char *origin = NULL; char *tosnap; char tofs[ZFS_MAX_DATASET_NAME_LEN]; boolean_t force; boolean_t heal; boolean_t resumable; uint64_t read_bytes = 0; uint64_t errflags = 0; int input_fd = -1; int error; snapname = fnvlist_lookup_string(innvl, "snapname"); if (dataset_namecheck(snapname, NULL, NULL) != 0 || strchr(snapname, '@') == NULL || strchr(snapname, '%')) return (SET_ERROR(EINVAL)); (void) strlcpy(tofs, snapname, sizeof (tofs)); tosnap = strchr(tofs, '@'); *tosnap++ = '\0'; error = nvlist_lookup_string(innvl, "origin", &origin); if (error && error != ENOENT) return (error); error = nvlist_lookup_byte_array(innvl, "begin_record", (uchar_t **)&begin_record, &begin_record_size); if (error != 0 || begin_record_size != sizeof (*begin_record)) return (SET_ERROR(EINVAL)); input_fd = fnvlist_lookup_int32(innvl, "input_fd"); force = nvlist_exists(innvl, "force"); heal = nvlist_exists(innvl, "heal"); resumable = nvlist_exists(innvl, "resumable"); /* we still use "props" here for backwards compatibility */ error = nvlist_lookup_nvlist(innvl, "props", &recvprops); if (error && error != ENOENT) return (error); error = nvlist_lookup_nvlist(innvl, "localprops", &localprops); if (error && error != ENOENT) return (error); error = nvlist_lookup_nvlist(innvl, ZPOOL_HIDDEN_ARGS, &hidden_args); if (error && error != ENOENT) return (error); error = zfs_ioc_recv_impl(tofs, tosnap, origin, recvprops, localprops, hidden_args, force, heal, resumable, input_fd, begin_record, &read_bytes, &errflags, &errors); fnvlist_add_uint64(outnvl, "read_bytes", read_bytes); fnvlist_add_uint64(outnvl, "error_flags", errflags); fnvlist_add_nvlist(outnvl, "errors", errors); nvlist_free(errors); nvlist_free(recvprops); nvlist_free(localprops); return (error); } typedef struct dump_bytes_io { zfs_file_t *dbi_fp; caddr_t dbi_buf; int dbi_len; int dbi_err; } dump_bytes_io_t; static void dump_bytes_cb(void *arg) { dump_bytes_io_t *dbi = (dump_bytes_io_t *)arg; zfs_file_t *fp; caddr_t buf; fp = dbi->dbi_fp; buf = dbi->dbi_buf; dbi->dbi_err = zfs_file_write(fp, buf, dbi->dbi_len, NULL); } static int dump_bytes(objset_t *os, void *buf, int len, void *arg) { dump_bytes_io_t dbi; dbi.dbi_fp = arg; dbi.dbi_buf = buf; dbi.dbi_len = len; #if defined(HAVE_LARGE_STACKS) dump_bytes_cb(&dbi); #else /* * The vn_rdwr() call is performed in a taskq to ensure that there is * always enough stack space to write safely to the target filesystem. * The ZIO_TYPE_FREE threads are used because there can be a lot of * them and they are used in vdev_file.c for a similar purpose. */ spa_taskq_dispatch_sync(dmu_objset_spa(os), ZIO_TYPE_FREE, ZIO_TASKQ_ISSUE, dump_bytes_cb, &dbi, TQ_SLEEP); #endif /* HAVE_LARGE_STACKS */ return (dbi.dbi_err); } /* * inputs: * zc_name name of snapshot to send * zc_cookie file descriptor to send stream to * zc_obj fromorigin flag (mutually exclusive with zc_fromobj) * zc_sendobj objsetid of snapshot to send * zc_fromobj objsetid of incremental fromsnap (may be zero) * zc_guid if set, estimate size of stream only. zc_cookie is ignored. * output size in zc_objset_type. * zc_flags lzc_send_flags * * outputs: * zc_objset_type estimated size, if zc_guid is set * * NOTE: This is no longer the preferred interface, any new functionality * should be added to zfs_ioc_send_new() instead. */ static int zfs_ioc_send(zfs_cmd_t *zc) { int error; offset_t off; boolean_t estimate = (zc->zc_guid != 0); boolean_t embedok = (zc->zc_flags & 0x1); boolean_t large_block_ok = (zc->zc_flags & 0x2); boolean_t compressok = (zc->zc_flags & 0x4); boolean_t rawok = (zc->zc_flags & 0x8); boolean_t savedok = (zc->zc_flags & 0x10); if (zc->zc_obj != 0) { dsl_pool_t *dp; dsl_dataset_t *tosnap; error = dsl_pool_hold(zc->zc_name, FTAG, &dp); if (error != 0) return (error); error = dsl_dataset_hold_obj(dp, zc->zc_sendobj, FTAG, &tosnap); if (error != 0) { dsl_pool_rele(dp, FTAG); return (error); } if (dsl_dir_is_clone(tosnap->ds_dir)) zc->zc_fromobj = dsl_dir_phys(tosnap->ds_dir)->dd_origin_obj; dsl_dataset_rele(tosnap, FTAG); dsl_pool_rele(dp, FTAG); } if (estimate) { dsl_pool_t *dp; dsl_dataset_t *tosnap; dsl_dataset_t *fromsnap = NULL; error = dsl_pool_hold(zc->zc_name, FTAG, &dp); if (error != 0) return (error); error = dsl_dataset_hold_obj(dp, zc->zc_sendobj, FTAG, &tosnap); if (error != 0) { dsl_pool_rele(dp, FTAG); return (error); } if (zc->zc_fromobj != 0) { error = dsl_dataset_hold_obj(dp, zc->zc_fromobj, FTAG, &fromsnap); if (error != 0) { dsl_dataset_rele(tosnap, FTAG); dsl_pool_rele(dp, FTAG); return (error); } } error = dmu_send_estimate_fast(tosnap, fromsnap, NULL, compressok || rawok, savedok, &zc->zc_objset_type); if (fromsnap != NULL) dsl_dataset_rele(fromsnap, FTAG); dsl_dataset_rele(tosnap, FTAG); dsl_pool_rele(dp, FTAG); } else { zfs_file_t *fp; dmu_send_outparams_t out = {0}; if ((fp = zfs_file_get(zc->zc_cookie)) == NULL) return (SET_ERROR(EBADF)); off = zfs_file_off(fp); out.dso_outfunc = dump_bytes; out.dso_arg = fp; out.dso_dryrun = B_FALSE; error = dmu_send_obj(zc->zc_name, zc->zc_sendobj, zc->zc_fromobj, embedok, large_block_ok, compressok, rawok, savedok, zc->zc_cookie, &off, &out); zfs_file_put(fp); } return (error); } /* * inputs: * zc_name name of snapshot on which to report progress * zc_cookie file descriptor of send stream * * outputs: * zc_cookie number of bytes written in send stream thus far * zc_objset_type logical size of data traversed by send thus far */ static int zfs_ioc_send_progress(zfs_cmd_t *zc) { dsl_pool_t *dp; dsl_dataset_t *ds; dmu_sendstatus_t *dsp = NULL; int error; error = dsl_pool_hold(zc->zc_name, FTAG, &dp); if (error != 0) return (error); error = dsl_dataset_hold(dp, zc->zc_name, FTAG, &ds); if (error != 0) { dsl_pool_rele(dp, FTAG); return (error); } mutex_enter(&ds->ds_sendstream_lock); /* * Iterate over all the send streams currently active on this dataset. * If there's one which matches the specified file descriptor _and_ the * stream was started by the current process, return the progress of * that stream. */ for (dsp = list_head(&ds->ds_sendstreams); dsp != NULL; dsp = list_next(&ds->ds_sendstreams, dsp)) { if (dsp->dss_outfd == zc->zc_cookie && zfs_proc_is_caller(dsp->dss_proc)) break; } if (dsp != NULL) { zc->zc_cookie = atomic_cas_64((volatile uint64_t *)dsp->dss_off, 0, 0); /* This is the closest thing we have to atomic_read_64. */ zc->zc_objset_type = atomic_cas_64(&dsp->dss_blocks, 0, 0); } else { error = SET_ERROR(ENOENT); } mutex_exit(&ds->ds_sendstream_lock); dsl_dataset_rele(ds, FTAG); dsl_pool_rele(dp, FTAG); return (error); } static int zfs_ioc_inject_fault(zfs_cmd_t *zc) { int id, error; error = zio_inject_fault(zc->zc_name, (int)zc->zc_guid, &id, &zc->zc_inject_record); if (error == 0) zc->zc_guid = (uint64_t)id; return (error); } static int zfs_ioc_clear_fault(zfs_cmd_t *zc) { return (zio_clear_fault((int)zc->zc_guid)); } static int zfs_ioc_inject_list_next(zfs_cmd_t *zc) { int id = (int)zc->zc_guid; int error; error = zio_inject_list_next(&id, zc->zc_name, sizeof (zc->zc_name), &zc->zc_inject_record); zc->zc_guid = id; return (error); } static int zfs_ioc_error_log(zfs_cmd_t *zc) { spa_t *spa; int error; if ((error = spa_open(zc->zc_name, &spa, FTAG)) != 0) return (error); error = spa_get_errlog(spa, (void *)(uintptr_t)zc->zc_nvlist_dst, &zc->zc_nvlist_dst_size); spa_close(spa, FTAG); return (error); } static int zfs_ioc_clear(zfs_cmd_t *zc) { spa_t *spa; vdev_t *vd; int error; /* * On zpool clear we also fix up missing slogs */ mutex_enter(&spa_namespace_lock); spa = spa_lookup(zc->zc_name); if (spa == NULL) { mutex_exit(&spa_namespace_lock); return (SET_ERROR(EIO)); } if (spa_get_log_state(spa) == SPA_LOG_MISSING) { /* we need to let spa_open/spa_load clear the chains */ spa_set_log_state(spa, SPA_LOG_CLEAR); } spa->spa_last_open_failed = 0; mutex_exit(&spa_namespace_lock); if (zc->zc_cookie & ZPOOL_NO_REWIND) { error = spa_open(zc->zc_name, &spa, FTAG); } else { nvlist_t *policy; nvlist_t *config = NULL; if (zc->zc_nvlist_src == 0) return (SET_ERROR(EINVAL)); if ((error = get_nvlist(zc->zc_nvlist_src, zc->zc_nvlist_src_size, zc->zc_iflags, &policy)) == 0) { error = spa_open_rewind(zc->zc_name, &spa, FTAG, policy, &config); if (config != NULL) { int err; if ((err = put_nvlist(zc, config)) != 0) error = err; nvlist_free(config); } nvlist_free(policy); } } if (error != 0) return (error); /* * If multihost is enabled, resuming I/O is unsafe as another * host may have imported the pool. */ if (spa_multihost(spa) && spa_suspended(spa)) return (SET_ERROR(EINVAL)); spa_vdev_state_enter(spa, SCL_NONE); if (zc->zc_guid == 0) { vd = NULL; } else { vd = spa_lookup_by_guid(spa, zc->zc_guid, B_TRUE); if (vd == NULL) { error = SET_ERROR(ENODEV); (void) spa_vdev_state_exit(spa, NULL, error); spa_close(spa, FTAG); return (error); } } vdev_clear(spa, vd); (void) spa_vdev_state_exit(spa, spa_suspended(spa) ? NULL : spa->spa_root_vdev, 0); /* * Resume any suspended I/Os. */ if (zio_resume(spa) != 0) error = SET_ERROR(EIO); spa_close(spa, FTAG); return (error); } /* * Reopen all the vdevs associated with the pool. * * innvl: { * "scrub_restart" -> when true and scrub is running, allow to restart * scrub as the side effect of the reopen (boolean). * } * * outnvl is unused */ static const zfs_ioc_key_t zfs_keys_pool_reopen[] = { {"scrub_restart", DATA_TYPE_BOOLEAN_VALUE, ZK_OPTIONAL}, }; static int zfs_ioc_pool_reopen(const char *pool, nvlist_t *innvl, nvlist_t *outnvl) { (void) outnvl; spa_t *spa; int error; boolean_t rc, scrub_restart = B_TRUE; if (innvl) { error = nvlist_lookup_boolean_value(innvl, "scrub_restart", &rc); if (error == 0) scrub_restart = rc; } error = spa_open(pool, &spa, FTAG); if (error != 0) return (error); spa_vdev_state_enter(spa, SCL_NONE); /* * If the scrub_restart flag is B_FALSE and a scrub is already * in progress then set spa_scrub_reopen flag to B_TRUE so that * we don't restart the scrub as a side effect of the reopen. * Otherwise, let vdev_open() decided if a resilver is required. */ spa->spa_scrub_reopen = (!scrub_restart && dsl_scan_scrubbing(spa->spa_dsl_pool)); vdev_reopen(spa->spa_root_vdev); spa->spa_scrub_reopen = B_FALSE; (void) spa_vdev_state_exit(spa, NULL, 0); spa_close(spa, FTAG); return (0); } /* * inputs: * zc_name name of filesystem * * outputs: * zc_string name of conflicting snapshot, if there is one */ static int zfs_ioc_promote(zfs_cmd_t *zc) { dsl_pool_t *dp; dsl_dataset_t *ds, *ods; char origin[ZFS_MAX_DATASET_NAME_LEN]; char *cp; int error; zc->zc_name[sizeof (zc->zc_name) - 1] = '\0'; if (dataset_namecheck(zc->zc_name, NULL, NULL) != 0 || strchr(zc->zc_name, '%')) return (SET_ERROR(EINVAL)); error = dsl_pool_hold(zc->zc_name, FTAG, &dp); if (error != 0) return (error); error = dsl_dataset_hold(dp, zc->zc_name, FTAG, &ds); if (error != 0) { dsl_pool_rele(dp, FTAG); return (error); } if (!dsl_dir_is_clone(ds->ds_dir)) { dsl_dataset_rele(ds, FTAG); dsl_pool_rele(dp, FTAG); return (SET_ERROR(EINVAL)); } error = dsl_dataset_hold_obj(dp, dsl_dir_phys(ds->ds_dir)->dd_origin_obj, FTAG, &ods); if (error != 0) { dsl_dataset_rele(ds, FTAG); dsl_pool_rele(dp, FTAG); return (error); } dsl_dataset_name(ods, origin); dsl_dataset_rele(ods, FTAG); dsl_dataset_rele(ds, FTAG); dsl_pool_rele(dp, FTAG); /* * We don't need to unmount *all* the origin fs's snapshots, but * it's easier. */ cp = strchr(origin, '@'); if (cp) *cp = '\0'; (void) dmu_objset_find(origin, zfs_unmount_snap_cb, NULL, DS_FIND_SNAPSHOTS); return (dsl_dataset_promote(zc->zc_name, zc->zc_string)); } /* * Retrieve a single {user|group|project}{used|quota}@... property. * * inputs: * zc_name name of filesystem * zc_objset_type zfs_userquota_prop_t * zc_value domain name (eg. "S-1-234-567-89") * zc_guid RID/UID/GID * * outputs: * zc_cookie property value */ static int zfs_ioc_userspace_one(zfs_cmd_t *zc) { zfsvfs_t *zfsvfs; int error; if (zc->zc_objset_type >= ZFS_NUM_USERQUOTA_PROPS) return (SET_ERROR(EINVAL)); error = zfsvfs_hold(zc->zc_name, FTAG, &zfsvfs, B_FALSE); if (error != 0) return (error); error = zfs_userspace_one(zfsvfs, zc->zc_objset_type, zc->zc_value, zc->zc_guid, &zc->zc_cookie); zfsvfs_rele(zfsvfs, FTAG); return (error); } /* * inputs: * zc_name name of filesystem * zc_cookie zap cursor * zc_objset_type zfs_userquota_prop_t * zc_nvlist_dst[_size] buffer to fill (not really an nvlist) * * outputs: * zc_nvlist_dst[_size] data buffer (array of zfs_useracct_t) * zc_cookie zap cursor */ static int zfs_ioc_userspace_many(zfs_cmd_t *zc) { zfsvfs_t *zfsvfs; int bufsize = zc->zc_nvlist_dst_size; if (bufsize <= 0) return (SET_ERROR(ENOMEM)); int error = zfsvfs_hold(zc->zc_name, FTAG, &zfsvfs, B_FALSE); if (error != 0) return (error); void *buf = vmem_alloc(bufsize, KM_SLEEP); error = zfs_userspace_many(zfsvfs, zc->zc_objset_type, &zc->zc_cookie, buf, &zc->zc_nvlist_dst_size); if (error == 0) { error = xcopyout(buf, (void *)(uintptr_t)zc->zc_nvlist_dst, zc->zc_nvlist_dst_size); } vmem_free(buf, bufsize); zfsvfs_rele(zfsvfs, FTAG); return (error); } /* * inputs: * zc_name name of filesystem * * outputs: * none */ static int zfs_ioc_userspace_upgrade(zfs_cmd_t *zc) { int error = 0; zfsvfs_t *zfsvfs; if (getzfsvfs(zc->zc_name, &zfsvfs) == 0) { if (!dmu_objset_userused_enabled(zfsvfs->z_os)) { /* * If userused is not enabled, it may be because the * objset needs to be closed & reopened (to grow the * objset_phys_t). Suspend/resume the fs will do that. */ dsl_dataset_t *ds, *newds; ds = dmu_objset_ds(zfsvfs->z_os); error = zfs_suspend_fs(zfsvfs); if (error == 0) { dmu_objset_refresh_ownership(ds, &newds, B_TRUE, zfsvfs); error = zfs_resume_fs(zfsvfs, newds); } } if (error == 0) { mutex_enter(&zfsvfs->z_os->os_upgrade_lock); if (zfsvfs->z_os->os_upgrade_id == 0) { /* clear potential error code and retry */ zfsvfs->z_os->os_upgrade_status = 0; mutex_exit(&zfsvfs->z_os->os_upgrade_lock); dsl_pool_config_enter( dmu_objset_pool(zfsvfs->z_os), FTAG); dmu_objset_userspace_upgrade(zfsvfs->z_os); dsl_pool_config_exit( dmu_objset_pool(zfsvfs->z_os), FTAG); } else { mutex_exit(&zfsvfs->z_os->os_upgrade_lock); } taskq_wait_id(zfsvfs->z_os->os_spa->spa_upgrade_taskq, zfsvfs->z_os->os_upgrade_id); error = zfsvfs->z_os->os_upgrade_status; } zfs_vfs_rele(zfsvfs); } else { objset_t *os; /* XXX kind of reading contents without owning */ error = dmu_objset_hold_flags(zc->zc_name, B_TRUE, FTAG, &os); if (error != 0) return (error); mutex_enter(&os->os_upgrade_lock); if (os->os_upgrade_id == 0) { /* clear potential error code and retry */ os->os_upgrade_status = 0; mutex_exit(&os->os_upgrade_lock); dmu_objset_userspace_upgrade(os); } else { mutex_exit(&os->os_upgrade_lock); } dsl_pool_rele(dmu_objset_pool(os), FTAG); taskq_wait_id(os->os_spa->spa_upgrade_taskq, os->os_upgrade_id); error = os->os_upgrade_status; dsl_dataset_rele_flags(dmu_objset_ds(os), DS_HOLD_FLAG_DECRYPT, FTAG); } return (error); } /* * inputs: * zc_name name of filesystem * * outputs: * none */ static int zfs_ioc_id_quota_upgrade(zfs_cmd_t *zc) { objset_t *os; int error; error = dmu_objset_hold_flags(zc->zc_name, B_TRUE, FTAG, &os); if (error != 0) return (error); if (dmu_objset_userobjspace_upgradable(os) || dmu_objset_projectquota_upgradable(os)) { mutex_enter(&os->os_upgrade_lock); if (os->os_upgrade_id == 0) { /* clear potential error code and retry */ os->os_upgrade_status = 0; mutex_exit(&os->os_upgrade_lock); dmu_objset_id_quota_upgrade(os); } else { mutex_exit(&os->os_upgrade_lock); } dsl_pool_rele(dmu_objset_pool(os), FTAG); taskq_wait_id(os->os_spa->spa_upgrade_taskq, os->os_upgrade_id); error = os->os_upgrade_status; } else { dsl_pool_rele(dmu_objset_pool(os), FTAG); } dsl_dataset_rele_flags(dmu_objset_ds(os), DS_HOLD_FLAG_DECRYPT, FTAG); return (error); } static int zfs_ioc_share(zfs_cmd_t *zc) { return (SET_ERROR(ENOSYS)); } /* * inputs: * zc_name name of containing filesystem * zc_obj object # beyond which we want next in-use object # * * outputs: * zc_obj next in-use object # */ static int zfs_ioc_next_obj(zfs_cmd_t *zc) { objset_t *os = NULL; int error; error = dmu_objset_hold(zc->zc_name, FTAG, &os); if (error != 0) return (error); error = dmu_object_next(os, &zc->zc_obj, B_FALSE, 0); dmu_objset_rele(os, FTAG); return (error); } /* * inputs: * zc_name name of filesystem * zc_value prefix name for snapshot * zc_cleanup_fd cleanup-on-exit file descriptor for calling process * * outputs: * zc_value short name of new snapshot */ static int zfs_ioc_tmp_snapshot(zfs_cmd_t *zc) { char *snap_name; char *hold_name; minor_t minor; zfs_file_t *fp = zfs_onexit_fd_hold(zc->zc_cleanup_fd, &minor); if (fp == NULL) return (SET_ERROR(EBADF)); snap_name = kmem_asprintf("%s-%016llx", zc->zc_value, (u_longlong_t)ddi_get_lbolt64()); hold_name = kmem_asprintf("%%%s", zc->zc_value); int error = dsl_dataset_snapshot_tmp(zc->zc_name, snap_name, minor, hold_name); if (error == 0) (void) strlcpy(zc->zc_value, snap_name, sizeof (zc->zc_value)); kmem_strfree(snap_name); kmem_strfree(hold_name); zfs_onexit_fd_rele(fp); return (error); } /* * inputs: * zc_name name of "to" snapshot * zc_value name of "from" snapshot * zc_cookie file descriptor to write diff data on * * outputs: * dmu_diff_record_t's to the file descriptor */ static int zfs_ioc_diff(zfs_cmd_t *zc) { zfs_file_t *fp; offset_t off; int error; if ((fp = zfs_file_get(zc->zc_cookie)) == NULL) return (SET_ERROR(EBADF)); off = zfs_file_off(fp); error = dmu_diff(zc->zc_name, zc->zc_value, fp, &off); zfs_file_put(fp); return (error); } static int zfs_ioc_smb_acl(zfs_cmd_t *zc) { return (SET_ERROR(ENOTSUP)); } /* * innvl: { * "holds" -> { snapname -> holdname (string), ... } * (optional) "cleanup_fd" -> fd (int32) * } * * outnvl: { * snapname -> error value (int32) * ... * } */ static const zfs_ioc_key_t zfs_keys_hold[] = { {"holds", DATA_TYPE_NVLIST, 0}, {"cleanup_fd", DATA_TYPE_INT32, ZK_OPTIONAL}, }; static int zfs_ioc_hold(const char *pool, nvlist_t *args, nvlist_t *errlist) { (void) pool; nvpair_t *pair; nvlist_t *holds; int cleanup_fd = -1; int error; minor_t minor = 0; zfs_file_t *fp = NULL; holds = fnvlist_lookup_nvlist(args, "holds"); /* make sure the user didn't pass us any invalid (empty) tags */ for (pair = nvlist_next_nvpair(holds, NULL); pair != NULL; pair = nvlist_next_nvpair(holds, pair)) { const char *htag; error = nvpair_value_string(pair, &htag); if (error != 0) return (SET_ERROR(error)); if (strlen(htag) == 0) return (SET_ERROR(EINVAL)); } if (nvlist_lookup_int32(args, "cleanup_fd", &cleanup_fd) == 0) { fp = zfs_onexit_fd_hold(cleanup_fd, &minor); if (fp == NULL) return (SET_ERROR(EBADF)); } error = dsl_dataset_user_hold(holds, minor, errlist); if (fp != NULL) { ASSERT3U(minor, !=, 0); zfs_onexit_fd_rele(fp); } return (SET_ERROR(error)); } /* * innvl is not used. * * outnvl: { * holdname -> time added (uint64 seconds since epoch) * ... * } */ static const zfs_ioc_key_t zfs_keys_get_holds[] = { /* no nvl keys */ }; static int zfs_ioc_get_holds(const char *snapname, nvlist_t *args, nvlist_t *outnvl) { (void) args; return (dsl_dataset_get_holds(snapname, outnvl)); } /* * innvl: { * snapname -> { holdname, ... } * ... * } * * outnvl: { * snapname -> error value (int32) * ... * } */ static const zfs_ioc_key_t zfs_keys_release[] = { {"...", DATA_TYPE_NVLIST, ZK_WILDCARDLIST}, }; static int zfs_ioc_release(const char *pool, nvlist_t *holds, nvlist_t *errlist) { (void) pool; return (dsl_dataset_user_release(holds, errlist)); } /* * inputs: * zc_guid flags (ZEVENT_NONBLOCK) * zc_cleanup_fd zevent file descriptor * * outputs: * zc_nvlist_dst next nvlist event * zc_cookie dropped events since last get */ static int zfs_ioc_events_next(zfs_cmd_t *zc) { zfs_zevent_t *ze; nvlist_t *event = NULL; minor_t minor; uint64_t dropped = 0; int error; zfs_file_t *fp = zfs_zevent_fd_hold(zc->zc_cleanup_fd, &minor, &ze); if (fp == NULL) return (SET_ERROR(EBADF)); do { error = zfs_zevent_next(ze, &event, &zc->zc_nvlist_dst_size, &dropped); if (event != NULL) { zc->zc_cookie = dropped; error = put_nvlist(zc, event); nvlist_free(event); } if (zc->zc_guid & ZEVENT_NONBLOCK) break; if ((error == 0) || (error != ENOENT)) break; error = zfs_zevent_wait(ze); if (error != 0) break; } while (1); zfs_zevent_fd_rele(fp); return (error); } /* * outputs: * zc_cookie cleared events count */ static int zfs_ioc_events_clear(zfs_cmd_t *zc) { uint_t count; zfs_zevent_drain_all(&count); zc->zc_cookie = count; return (0); } /* * inputs: * zc_guid eid | ZEVENT_SEEK_START | ZEVENT_SEEK_END * zc_cleanup zevent file descriptor */ static int zfs_ioc_events_seek(zfs_cmd_t *zc) { zfs_zevent_t *ze; minor_t minor; int error; zfs_file_t *fp = zfs_zevent_fd_hold(zc->zc_cleanup_fd, &minor, &ze); if (fp == NULL) return (SET_ERROR(EBADF)); error = zfs_zevent_seek(ze, zc->zc_guid); zfs_zevent_fd_rele(fp); return (error); } /* * inputs: * zc_name name of later filesystem or snapshot * zc_value full name of old snapshot or bookmark * * outputs: * zc_cookie space in bytes * zc_objset_type compressed space in bytes * zc_perm_action uncompressed space in bytes */ static int zfs_ioc_space_written(zfs_cmd_t *zc) { int error; dsl_pool_t *dp; dsl_dataset_t *new; error = dsl_pool_hold(zc->zc_name, FTAG, &dp); if (error != 0) return (error); error = dsl_dataset_hold(dp, zc->zc_name, FTAG, &new); if (error != 0) { dsl_pool_rele(dp, FTAG); return (error); } if (strchr(zc->zc_value, '#') != NULL) { zfs_bookmark_phys_t bmp; error = dsl_bookmark_lookup(dp, zc->zc_value, new, &bmp); if (error == 0) { error = dsl_dataset_space_written_bookmark(&bmp, new, &zc->zc_cookie, &zc->zc_objset_type, &zc->zc_perm_action); } } else { dsl_dataset_t *old; error = dsl_dataset_hold(dp, zc->zc_value, FTAG, &old); if (error == 0) { error = dsl_dataset_space_written(old, new, &zc->zc_cookie, &zc->zc_objset_type, &zc->zc_perm_action); dsl_dataset_rele(old, FTAG); } } dsl_dataset_rele(new, FTAG); dsl_pool_rele(dp, FTAG); return (error); } /* * innvl: { * "firstsnap" -> snapshot name * } * * outnvl: { * "used" -> space in bytes * "compressed" -> compressed space in bytes * "uncompressed" -> uncompressed space in bytes * } */ static const zfs_ioc_key_t zfs_keys_space_snaps[] = { {"firstsnap", DATA_TYPE_STRING, 0}, }; static int zfs_ioc_space_snaps(const char *lastsnap, nvlist_t *innvl, nvlist_t *outnvl) { int error; dsl_pool_t *dp; dsl_dataset_t *new, *old; const char *firstsnap; uint64_t used, comp, uncomp; firstsnap = fnvlist_lookup_string(innvl, "firstsnap"); error = dsl_pool_hold(lastsnap, FTAG, &dp); if (error != 0) return (error); error = dsl_dataset_hold(dp, lastsnap, FTAG, &new); if (error == 0 && !new->ds_is_snapshot) { dsl_dataset_rele(new, FTAG); error = SET_ERROR(EINVAL); } if (error != 0) { dsl_pool_rele(dp, FTAG); return (error); } error = dsl_dataset_hold(dp, firstsnap, FTAG, &old); if (error == 0 && !old->ds_is_snapshot) { dsl_dataset_rele(old, FTAG); error = SET_ERROR(EINVAL); } if (error != 0) { dsl_dataset_rele(new, FTAG); dsl_pool_rele(dp, FTAG); return (error); } error = dsl_dataset_space_wouldfree(old, new, &used, &comp, &uncomp); dsl_dataset_rele(old, FTAG); dsl_dataset_rele(new, FTAG); dsl_pool_rele(dp, FTAG); fnvlist_add_uint64(outnvl, "used", used); fnvlist_add_uint64(outnvl, "compressed", comp); fnvlist_add_uint64(outnvl, "uncompressed", uncomp); return (error); } /* * innvl: { * "fd" -> file descriptor to write stream to (int32) * (optional) "fromsnap" -> full snap name to send an incremental from * (optional) "largeblockok" -> (value ignored) * indicates that blocks > 128KB are permitted * (optional) "embedok" -> (value ignored) * presence indicates DRR_WRITE_EMBEDDED records are permitted * (optional) "compressok" -> (value ignored) * presence indicates compressed DRR_WRITE records are permitted * (optional) "rawok" -> (value ignored) * presence indicates raw encrypted records should be used. * (optional) "savedok" -> (value ignored) * presence indicates we should send a partially received snapshot * (optional) "resume_object" and "resume_offset" -> (uint64) * if present, resume send stream from specified object and offset. * (optional) "redactbook" -> (string) * if present, use this bookmark's redaction list to generate a redacted * send stream * } * * outnvl is unused */ static const zfs_ioc_key_t zfs_keys_send_new[] = { {"fd", DATA_TYPE_INT32, 0}, {"fromsnap", DATA_TYPE_STRING, ZK_OPTIONAL}, {"largeblockok", DATA_TYPE_BOOLEAN, ZK_OPTIONAL}, {"embedok", DATA_TYPE_BOOLEAN, ZK_OPTIONAL}, {"compressok", DATA_TYPE_BOOLEAN, ZK_OPTIONAL}, {"rawok", DATA_TYPE_BOOLEAN, ZK_OPTIONAL}, {"savedok", DATA_TYPE_BOOLEAN, ZK_OPTIONAL}, {"resume_object", DATA_TYPE_UINT64, ZK_OPTIONAL}, {"resume_offset", DATA_TYPE_UINT64, ZK_OPTIONAL}, {"redactbook", DATA_TYPE_STRING, ZK_OPTIONAL}, }; static int zfs_ioc_send_new(const char *snapname, nvlist_t *innvl, nvlist_t *outnvl) { (void) outnvl; int error; offset_t off; const char *fromname = NULL; int fd; zfs_file_t *fp; boolean_t largeblockok; boolean_t embedok; boolean_t compressok; boolean_t rawok; boolean_t savedok; uint64_t resumeobj = 0; uint64_t resumeoff = 0; const char *redactbook = NULL; fd = fnvlist_lookup_int32(innvl, "fd"); (void) nvlist_lookup_string(innvl, "fromsnap", &fromname); largeblockok = nvlist_exists(innvl, "largeblockok"); embedok = nvlist_exists(innvl, "embedok"); compressok = nvlist_exists(innvl, "compressok"); rawok = nvlist_exists(innvl, "rawok"); savedok = nvlist_exists(innvl, "savedok"); (void) nvlist_lookup_uint64(innvl, "resume_object", &resumeobj); (void) nvlist_lookup_uint64(innvl, "resume_offset", &resumeoff); (void) nvlist_lookup_string(innvl, "redactbook", &redactbook); if ((fp = zfs_file_get(fd)) == NULL) return (SET_ERROR(EBADF)); off = zfs_file_off(fp); dmu_send_outparams_t out = {0}; out.dso_outfunc = dump_bytes; out.dso_arg = fp; out.dso_dryrun = B_FALSE; error = dmu_send(snapname, fromname, embedok, largeblockok, compressok, rawok, savedok, resumeobj, resumeoff, redactbook, fd, &off, &out); zfs_file_put(fp); return (error); } static int send_space_sum(objset_t *os, void *buf, int len, void *arg) { (void) os, (void) buf; uint64_t *size = arg; *size += len; return (0); } /* * Determine approximately how large a zfs send stream will be -- the number * of bytes that will be written to the fd supplied to zfs_ioc_send_new(). * * innvl: { * (optional) "from" -> full snap or bookmark name to send an incremental * from * (optional) "largeblockok" -> (value ignored) * indicates that blocks > 128KB are permitted * (optional) "embedok" -> (value ignored) * presence indicates DRR_WRITE_EMBEDDED records are permitted * (optional) "compressok" -> (value ignored) * presence indicates compressed DRR_WRITE records are permitted * (optional) "rawok" -> (value ignored) * presence indicates raw encrypted records should be used. * (optional) "resume_object" and "resume_offset" -> (uint64) * if present, resume send stream from specified object and offset. * (optional) "fd" -> file descriptor to use as a cookie for progress * tracking (int32) * } * * outnvl: { * "space" -> bytes of space (uint64) * } */ static const zfs_ioc_key_t zfs_keys_send_space[] = { {"from", DATA_TYPE_STRING, ZK_OPTIONAL}, {"fromsnap", DATA_TYPE_STRING, ZK_OPTIONAL}, {"largeblockok", DATA_TYPE_BOOLEAN, ZK_OPTIONAL}, {"embedok", DATA_TYPE_BOOLEAN, ZK_OPTIONAL}, {"compressok", DATA_TYPE_BOOLEAN, ZK_OPTIONAL}, {"rawok", DATA_TYPE_BOOLEAN, ZK_OPTIONAL}, {"fd", DATA_TYPE_INT32, ZK_OPTIONAL}, {"redactbook", DATA_TYPE_STRING, ZK_OPTIONAL}, {"resume_object", DATA_TYPE_UINT64, ZK_OPTIONAL}, {"resume_offset", DATA_TYPE_UINT64, ZK_OPTIONAL}, {"bytes", DATA_TYPE_UINT64, ZK_OPTIONAL}, }; static int zfs_ioc_send_space(const char *snapname, nvlist_t *innvl, nvlist_t *outnvl) { dsl_pool_t *dp; dsl_dataset_t *tosnap; dsl_dataset_t *fromsnap = NULL; int error; const char *fromname = NULL; const char *redactlist_book = NULL; boolean_t largeblockok; boolean_t embedok; boolean_t compressok; boolean_t rawok; boolean_t savedok; uint64_t space = 0; boolean_t full_estimate = B_FALSE; uint64_t resumeobj = 0; uint64_t resumeoff = 0; uint64_t resume_bytes = 0; int32_t fd = -1; zfs_bookmark_phys_t zbm = {0}; error = dsl_pool_hold(snapname, FTAG, &dp); if (error != 0) return (error); error = dsl_dataset_hold(dp, snapname, FTAG, &tosnap); if (error != 0) { dsl_pool_rele(dp, FTAG); return (error); } (void) nvlist_lookup_int32(innvl, "fd", &fd); largeblockok = nvlist_exists(innvl, "largeblockok"); embedok = nvlist_exists(innvl, "embedok"); compressok = nvlist_exists(innvl, "compressok"); rawok = nvlist_exists(innvl, "rawok"); savedok = nvlist_exists(innvl, "savedok"); boolean_t from = (nvlist_lookup_string(innvl, "from", &fromname) == 0); boolean_t altbook = (nvlist_lookup_string(innvl, "redactbook", &redactlist_book) == 0); (void) nvlist_lookup_uint64(innvl, "resume_object", &resumeobj); (void) nvlist_lookup_uint64(innvl, "resume_offset", &resumeoff); (void) nvlist_lookup_uint64(innvl, "bytes", &resume_bytes); if (altbook) { full_estimate = B_TRUE; } else if (from) { if (strchr(fromname, '#')) { error = dsl_bookmark_lookup(dp, fromname, tosnap, &zbm); /* * dsl_bookmark_lookup() will fail with EXDEV if * the from-bookmark and tosnap are at the same txg. * However, it's valid to do a send (and therefore, * a send estimate) from and to the same time point, * if the bookmark is redacted (the incremental send * can change what's redacted on the target). In * this case, dsl_bookmark_lookup() fills in zbm * but returns EXDEV. Ignore this error. */ if (error == EXDEV && zbm.zbm_redaction_obj != 0 && zbm.zbm_guid == dsl_dataset_phys(tosnap)->ds_guid) error = 0; if (error != 0) { dsl_dataset_rele(tosnap, FTAG); dsl_pool_rele(dp, FTAG); return (error); } if (zbm.zbm_redaction_obj != 0 || !(zbm.zbm_flags & ZBM_FLAG_HAS_FBN)) { full_estimate = B_TRUE; } } else if (strchr(fromname, '@')) { error = dsl_dataset_hold(dp, fromname, FTAG, &fromsnap); if (error != 0) { dsl_dataset_rele(tosnap, FTAG); dsl_pool_rele(dp, FTAG); return (error); } if (!dsl_dataset_is_before(tosnap, fromsnap, 0)) { full_estimate = B_TRUE; dsl_dataset_rele(fromsnap, FTAG); } } else { /* * from is not properly formatted as a snapshot or * bookmark */ dsl_dataset_rele(tosnap, FTAG); dsl_pool_rele(dp, FTAG); return (SET_ERROR(EINVAL)); } } if (full_estimate) { dmu_send_outparams_t out = {0}; offset_t off = 0; out.dso_outfunc = send_space_sum; out.dso_arg = &space; out.dso_dryrun = B_TRUE; /* * We have to release these holds so dmu_send can take them. It * will do all the error checking we need. */ dsl_dataset_rele(tosnap, FTAG); dsl_pool_rele(dp, FTAG); error = dmu_send(snapname, fromname, embedok, largeblockok, compressok, rawok, savedok, resumeobj, resumeoff, redactlist_book, fd, &off, &out); } else { error = dmu_send_estimate_fast(tosnap, fromsnap, (from && strchr(fromname, '#') != NULL ? &zbm : NULL), compressok || rawok, savedok, &space); space -= resume_bytes; if (fromsnap != NULL) dsl_dataset_rele(fromsnap, FTAG); dsl_dataset_rele(tosnap, FTAG); dsl_pool_rele(dp, FTAG); } fnvlist_add_uint64(outnvl, "space", space); return (error); } /* * Sync the currently open TXG to disk for the specified pool. * This is somewhat similar to 'zfs_sync()'. * For cases that do not result in error this ioctl will wait for * the currently open TXG to commit before returning back to the caller. * * innvl: { * "force" -> when true, force uberblock update even if there is no dirty data. * In addition this will cause the vdev configuration to be written * out including updating the zpool cache file. (boolean_t) * } * * onvl is unused */ static const zfs_ioc_key_t zfs_keys_pool_sync[] = { {"force", DATA_TYPE_BOOLEAN_VALUE, 0}, }; static int zfs_ioc_pool_sync(const char *pool, nvlist_t *innvl, nvlist_t *onvl) { (void) onvl; int err; boolean_t rc, force = B_FALSE; spa_t *spa; if ((err = spa_open(pool, &spa, FTAG)) != 0) return (err); if (innvl) { err = nvlist_lookup_boolean_value(innvl, "force", &rc); if (err == 0) force = rc; } if (force) { spa_config_enter(spa, SCL_CONFIG, FTAG, RW_WRITER); vdev_config_dirty(spa->spa_root_vdev); spa_config_exit(spa, SCL_CONFIG, FTAG); } txg_wait_synced(spa_get_dsl(spa), 0); spa_close(spa, FTAG); return (0); } /* * Load a user's wrapping key into the kernel. * innvl: { * "hidden_args" -> { "wkeydata" -> value } * raw uint8_t array of encryption wrapping key data (32 bytes) * (optional) "noop" -> (value ignored) * presence indicated key should only be verified, not loaded * } */ static const zfs_ioc_key_t zfs_keys_load_key[] = { {"hidden_args", DATA_TYPE_NVLIST, 0}, {"noop", DATA_TYPE_BOOLEAN, ZK_OPTIONAL}, }; static int zfs_ioc_load_key(const char *dsname, nvlist_t *innvl, nvlist_t *outnvl) { (void) outnvl; int ret; dsl_crypto_params_t *dcp = NULL; nvlist_t *hidden_args; boolean_t noop = nvlist_exists(innvl, "noop"); if (strchr(dsname, '@') != NULL || strchr(dsname, '%') != NULL) { ret = SET_ERROR(EINVAL); goto error; } hidden_args = fnvlist_lookup_nvlist(innvl, ZPOOL_HIDDEN_ARGS); ret = dsl_crypto_params_create_nvlist(DCP_CMD_NONE, NULL, hidden_args, &dcp); if (ret != 0) goto error; ret = spa_keystore_load_wkey(dsname, dcp, noop); if (ret != 0) goto error; dsl_crypto_params_free(dcp, noop); return (0); error: dsl_crypto_params_free(dcp, B_TRUE); return (ret); } /* * Unload a user's wrapping key from the kernel. * Both innvl and outnvl are unused. */ static const zfs_ioc_key_t zfs_keys_unload_key[] = { /* no nvl keys */ }; static int zfs_ioc_unload_key(const char *dsname, nvlist_t *innvl, nvlist_t *outnvl) { (void) innvl, (void) outnvl; int ret = 0; if (strchr(dsname, '@') != NULL || strchr(dsname, '%') != NULL) { ret = (SET_ERROR(EINVAL)); goto out; } ret = spa_keystore_unload_wkey(dsname); if (ret != 0) goto out; out: return (ret); } /* * Changes a user's wrapping key used to decrypt a dataset. The keyformat, * keylocation, pbkdf2salt, and pbkdf2iters properties can also be specified * here to change how the key is derived in userspace. * * innvl: { * "hidden_args" (optional) -> { "wkeydata" -> value } * raw uint8_t array of new encryption wrapping key data (32 bytes) * "props" (optional) -> { prop -> value } * } * * outnvl is unused */ static const zfs_ioc_key_t zfs_keys_change_key[] = { {"crypt_cmd", DATA_TYPE_UINT64, ZK_OPTIONAL}, {"hidden_args", DATA_TYPE_NVLIST, ZK_OPTIONAL}, {"props", DATA_TYPE_NVLIST, ZK_OPTIONAL}, }; static int zfs_ioc_change_key(const char *dsname, nvlist_t *innvl, nvlist_t *outnvl) { (void) outnvl; int ret; uint64_t cmd = DCP_CMD_NONE; dsl_crypto_params_t *dcp = NULL; nvlist_t *args = NULL, *hidden_args = NULL; if (strchr(dsname, '@') != NULL || strchr(dsname, '%') != NULL) { ret = (SET_ERROR(EINVAL)); goto error; } (void) nvlist_lookup_uint64(innvl, "crypt_cmd", &cmd); (void) nvlist_lookup_nvlist(innvl, "props", &args); (void) nvlist_lookup_nvlist(innvl, ZPOOL_HIDDEN_ARGS, &hidden_args); ret = dsl_crypto_params_create_nvlist(cmd, args, hidden_args, &dcp); if (ret != 0) goto error; ret = spa_keystore_change_key(dsname, dcp); if (ret != 0) goto error; dsl_crypto_params_free(dcp, B_FALSE); return (0); error: dsl_crypto_params_free(dcp, B_TRUE); return (ret); } static zfs_ioc_vec_t zfs_ioc_vec[ZFS_IOC_LAST - ZFS_IOC_FIRST]; static void zfs_ioctl_register_legacy(zfs_ioc_t ioc, zfs_ioc_legacy_func_t *func, zfs_secpolicy_func_t *secpolicy, zfs_ioc_namecheck_t namecheck, boolean_t log_history, zfs_ioc_poolcheck_t pool_check) { zfs_ioc_vec_t *vec = &zfs_ioc_vec[ioc - ZFS_IOC_FIRST]; ASSERT3U(ioc, >=, ZFS_IOC_FIRST); ASSERT3U(ioc, <, ZFS_IOC_LAST); ASSERT3P(vec->zvec_legacy_func, ==, NULL); ASSERT3P(vec->zvec_func, ==, NULL); vec->zvec_legacy_func = func; vec->zvec_secpolicy = secpolicy; vec->zvec_namecheck = namecheck; vec->zvec_allow_log = log_history; vec->zvec_pool_check = pool_check; } /* * See the block comment at the beginning of this file for details on * each argument to this function. */ void zfs_ioctl_register(const char *name, zfs_ioc_t ioc, zfs_ioc_func_t *func, zfs_secpolicy_func_t *secpolicy, zfs_ioc_namecheck_t namecheck, zfs_ioc_poolcheck_t pool_check, boolean_t smush_outnvlist, boolean_t allow_log, const zfs_ioc_key_t *nvl_keys, size_t num_keys) { zfs_ioc_vec_t *vec = &zfs_ioc_vec[ioc - ZFS_IOC_FIRST]; ASSERT3U(ioc, >=, ZFS_IOC_FIRST); ASSERT3U(ioc, <, ZFS_IOC_LAST); ASSERT3P(vec->zvec_legacy_func, ==, NULL); ASSERT3P(vec->zvec_func, ==, NULL); /* if we are logging, the name must be valid */ ASSERT(!allow_log || namecheck != NO_NAME); vec->zvec_name = name; vec->zvec_func = func; vec->zvec_secpolicy = secpolicy; vec->zvec_namecheck = namecheck; vec->zvec_pool_check = pool_check; vec->zvec_smush_outnvlist = smush_outnvlist; vec->zvec_allow_log = allow_log; vec->zvec_nvl_keys = nvl_keys; vec->zvec_nvl_key_count = num_keys; } static void zfs_ioctl_register_pool(zfs_ioc_t ioc, zfs_ioc_legacy_func_t *func, zfs_secpolicy_func_t *secpolicy, boolean_t log_history, zfs_ioc_poolcheck_t pool_check) { zfs_ioctl_register_legacy(ioc, func, secpolicy, POOL_NAME, log_history, pool_check); } void zfs_ioctl_register_dataset_nolog(zfs_ioc_t ioc, zfs_ioc_legacy_func_t *func, zfs_secpolicy_func_t *secpolicy, zfs_ioc_poolcheck_t pool_check) { zfs_ioctl_register_legacy(ioc, func, secpolicy, DATASET_NAME, B_FALSE, pool_check); } static void zfs_ioctl_register_pool_modify(zfs_ioc_t ioc, zfs_ioc_legacy_func_t *func) { zfs_ioctl_register_legacy(ioc, func, zfs_secpolicy_config, POOL_NAME, B_TRUE, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY); } static void zfs_ioctl_register_pool_meta(zfs_ioc_t ioc, zfs_ioc_legacy_func_t *func, zfs_secpolicy_func_t *secpolicy) { zfs_ioctl_register_legacy(ioc, func, secpolicy, NO_NAME, B_FALSE, POOL_CHECK_NONE); } static void zfs_ioctl_register_dataset_read_secpolicy(zfs_ioc_t ioc, zfs_ioc_legacy_func_t *func, zfs_secpolicy_func_t *secpolicy) { zfs_ioctl_register_legacy(ioc, func, secpolicy, DATASET_NAME, B_FALSE, POOL_CHECK_SUSPENDED); } static void zfs_ioctl_register_dataset_read(zfs_ioc_t ioc, zfs_ioc_legacy_func_t *func) { zfs_ioctl_register_dataset_read_secpolicy(ioc, func, zfs_secpolicy_read); } static void zfs_ioctl_register_dataset_modify(zfs_ioc_t ioc, zfs_ioc_legacy_func_t *func, zfs_secpolicy_func_t *secpolicy) { zfs_ioctl_register_legacy(ioc, func, secpolicy, DATASET_NAME, B_TRUE, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY); } static void zfs_ioctl_init(void) { zfs_ioctl_register("snapshot", ZFS_IOC_SNAPSHOT, zfs_ioc_snapshot, zfs_secpolicy_snapshot, POOL_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE, zfs_keys_snapshot, ARRAY_SIZE(zfs_keys_snapshot)); zfs_ioctl_register("log_history", ZFS_IOC_LOG_HISTORY, zfs_ioc_log_history, zfs_secpolicy_log_history, NO_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_FALSE, B_FALSE, zfs_keys_log_history, ARRAY_SIZE(zfs_keys_log_history)); zfs_ioctl_register("space_snaps", ZFS_IOC_SPACE_SNAPS, zfs_ioc_space_snaps, zfs_secpolicy_read, DATASET_NAME, POOL_CHECK_SUSPENDED, B_FALSE, B_FALSE, zfs_keys_space_snaps, ARRAY_SIZE(zfs_keys_space_snaps)); zfs_ioctl_register("send", ZFS_IOC_SEND_NEW, zfs_ioc_send_new, zfs_secpolicy_send_new, DATASET_NAME, POOL_CHECK_SUSPENDED, B_FALSE, B_FALSE, zfs_keys_send_new, ARRAY_SIZE(zfs_keys_send_new)); zfs_ioctl_register("send_space", ZFS_IOC_SEND_SPACE, zfs_ioc_send_space, zfs_secpolicy_read, DATASET_NAME, POOL_CHECK_SUSPENDED, B_FALSE, B_FALSE, zfs_keys_send_space, ARRAY_SIZE(zfs_keys_send_space)); zfs_ioctl_register("create", ZFS_IOC_CREATE, zfs_ioc_create, zfs_secpolicy_create_clone, DATASET_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE, zfs_keys_create, ARRAY_SIZE(zfs_keys_create)); zfs_ioctl_register("clone", ZFS_IOC_CLONE, zfs_ioc_clone, zfs_secpolicy_create_clone, DATASET_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE, zfs_keys_clone, ARRAY_SIZE(zfs_keys_clone)); zfs_ioctl_register("remap", ZFS_IOC_REMAP, zfs_ioc_remap, zfs_secpolicy_none, DATASET_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_FALSE, B_TRUE, zfs_keys_remap, ARRAY_SIZE(zfs_keys_remap)); zfs_ioctl_register("destroy_snaps", ZFS_IOC_DESTROY_SNAPS, zfs_ioc_destroy_snaps, zfs_secpolicy_destroy_snaps, POOL_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE, zfs_keys_destroy_snaps, ARRAY_SIZE(zfs_keys_destroy_snaps)); zfs_ioctl_register("hold", ZFS_IOC_HOLD, zfs_ioc_hold, zfs_secpolicy_hold, POOL_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE, zfs_keys_hold, ARRAY_SIZE(zfs_keys_hold)); zfs_ioctl_register("release", ZFS_IOC_RELEASE, zfs_ioc_release, zfs_secpolicy_release, POOL_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE, zfs_keys_release, ARRAY_SIZE(zfs_keys_release)); zfs_ioctl_register("get_holds", ZFS_IOC_GET_HOLDS, zfs_ioc_get_holds, zfs_secpolicy_read, DATASET_NAME, POOL_CHECK_SUSPENDED, B_FALSE, B_FALSE, zfs_keys_get_holds, ARRAY_SIZE(zfs_keys_get_holds)); zfs_ioctl_register("rollback", ZFS_IOC_ROLLBACK, zfs_ioc_rollback, zfs_secpolicy_rollback, DATASET_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_FALSE, B_TRUE, zfs_keys_rollback, ARRAY_SIZE(zfs_keys_rollback)); zfs_ioctl_register("bookmark", ZFS_IOC_BOOKMARK, zfs_ioc_bookmark, zfs_secpolicy_bookmark, POOL_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE, zfs_keys_bookmark, ARRAY_SIZE(zfs_keys_bookmark)); zfs_ioctl_register("get_bookmarks", ZFS_IOC_GET_BOOKMARKS, zfs_ioc_get_bookmarks, zfs_secpolicy_read, DATASET_NAME, POOL_CHECK_SUSPENDED, B_FALSE, B_FALSE, zfs_keys_get_bookmarks, ARRAY_SIZE(zfs_keys_get_bookmarks)); zfs_ioctl_register("get_bookmark_props", ZFS_IOC_GET_BOOKMARK_PROPS, zfs_ioc_get_bookmark_props, zfs_secpolicy_read, ENTITY_NAME, POOL_CHECK_SUSPENDED, B_FALSE, B_FALSE, zfs_keys_get_bookmark_props, ARRAY_SIZE(zfs_keys_get_bookmark_props)); zfs_ioctl_register("destroy_bookmarks", ZFS_IOC_DESTROY_BOOKMARKS, zfs_ioc_destroy_bookmarks, zfs_secpolicy_destroy_bookmarks, POOL_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE, zfs_keys_destroy_bookmarks, ARRAY_SIZE(zfs_keys_destroy_bookmarks)); zfs_ioctl_register("receive", ZFS_IOC_RECV_NEW, zfs_ioc_recv_new, zfs_secpolicy_recv, DATASET_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE, zfs_keys_recv_new, ARRAY_SIZE(zfs_keys_recv_new)); zfs_ioctl_register("load-key", ZFS_IOC_LOAD_KEY, zfs_ioc_load_key, zfs_secpolicy_load_key, DATASET_NAME, POOL_CHECK_SUSPENDED, B_TRUE, B_TRUE, zfs_keys_load_key, ARRAY_SIZE(zfs_keys_load_key)); zfs_ioctl_register("unload-key", ZFS_IOC_UNLOAD_KEY, zfs_ioc_unload_key, zfs_secpolicy_load_key, DATASET_NAME, POOL_CHECK_SUSPENDED, B_TRUE, B_TRUE, zfs_keys_unload_key, ARRAY_SIZE(zfs_keys_unload_key)); zfs_ioctl_register("change-key", ZFS_IOC_CHANGE_KEY, zfs_ioc_change_key, zfs_secpolicy_change_key, DATASET_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE, zfs_keys_change_key, ARRAY_SIZE(zfs_keys_change_key)); zfs_ioctl_register("sync", ZFS_IOC_POOL_SYNC, zfs_ioc_pool_sync, zfs_secpolicy_none, POOL_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_FALSE, B_FALSE, zfs_keys_pool_sync, ARRAY_SIZE(zfs_keys_pool_sync)); zfs_ioctl_register("reopen", ZFS_IOC_POOL_REOPEN, zfs_ioc_pool_reopen, zfs_secpolicy_config, POOL_NAME, POOL_CHECK_SUSPENDED, B_TRUE, B_TRUE, zfs_keys_pool_reopen, ARRAY_SIZE(zfs_keys_pool_reopen)); zfs_ioctl_register("channel_program", ZFS_IOC_CHANNEL_PROGRAM, zfs_ioc_channel_program, zfs_secpolicy_config, POOL_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE, zfs_keys_channel_program, ARRAY_SIZE(zfs_keys_channel_program)); zfs_ioctl_register("redact", ZFS_IOC_REDACT, zfs_ioc_redact, zfs_secpolicy_config, DATASET_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE, zfs_keys_redact, ARRAY_SIZE(zfs_keys_redact)); zfs_ioctl_register("zpool_checkpoint", ZFS_IOC_POOL_CHECKPOINT, zfs_ioc_pool_checkpoint, zfs_secpolicy_config, POOL_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE, zfs_keys_pool_checkpoint, ARRAY_SIZE(zfs_keys_pool_checkpoint)); zfs_ioctl_register("zpool_discard_checkpoint", ZFS_IOC_POOL_DISCARD_CHECKPOINT, zfs_ioc_pool_discard_checkpoint, zfs_secpolicy_config, POOL_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE, zfs_keys_pool_discard_checkpoint, ARRAY_SIZE(zfs_keys_pool_discard_checkpoint)); zfs_ioctl_register("initialize", ZFS_IOC_POOL_INITIALIZE, zfs_ioc_pool_initialize, zfs_secpolicy_config, POOL_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE, zfs_keys_pool_initialize, ARRAY_SIZE(zfs_keys_pool_initialize)); zfs_ioctl_register("trim", ZFS_IOC_POOL_TRIM, zfs_ioc_pool_trim, zfs_secpolicy_config, POOL_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_TRUE, B_TRUE, zfs_keys_pool_trim, ARRAY_SIZE(zfs_keys_pool_trim)); zfs_ioctl_register("wait", ZFS_IOC_WAIT, zfs_ioc_wait, zfs_secpolicy_none, POOL_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_FALSE, B_FALSE, zfs_keys_pool_wait, ARRAY_SIZE(zfs_keys_pool_wait)); zfs_ioctl_register("wait_fs", ZFS_IOC_WAIT_FS, zfs_ioc_wait_fs, zfs_secpolicy_none, DATASET_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_FALSE, B_FALSE, zfs_keys_fs_wait, ARRAY_SIZE(zfs_keys_fs_wait)); zfs_ioctl_register("set_bootenv", ZFS_IOC_SET_BOOTENV, zfs_ioc_set_bootenv, zfs_secpolicy_config, POOL_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_FALSE, B_TRUE, zfs_keys_set_bootenv, ARRAY_SIZE(zfs_keys_set_bootenv)); zfs_ioctl_register("get_bootenv", ZFS_IOC_GET_BOOTENV, zfs_ioc_get_bootenv, zfs_secpolicy_none, POOL_NAME, POOL_CHECK_SUSPENDED, B_FALSE, B_TRUE, zfs_keys_get_bootenv, ARRAY_SIZE(zfs_keys_get_bootenv)); zfs_ioctl_register("zpool_vdev_get_props", ZFS_IOC_VDEV_GET_PROPS, zfs_ioc_vdev_get_props, zfs_secpolicy_read, POOL_NAME, POOL_CHECK_NONE, B_FALSE, B_FALSE, zfs_keys_vdev_get_props, ARRAY_SIZE(zfs_keys_vdev_get_props)); zfs_ioctl_register("zpool_vdev_set_props", ZFS_IOC_VDEV_SET_PROPS, zfs_ioc_vdev_set_props, zfs_secpolicy_config, POOL_NAME, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY, B_FALSE, B_FALSE, zfs_keys_vdev_set_props, ARRAY_SIZE(zfs_keys_vdev_set_props)); + zfs_ioctl_register("scrub", ZFS_IOC_POOL_SCRUB, + zfs_ioc_pool_scrub, zfs_secpolicy_config, POOL_NAME, + POOL_CHECK_NONE, B_TRUE, B_TRUE, + zfs_keys_pool_scrub, ARRAY_SIZE(zfs_keys_pool_scrub)); + /* IOCTLS that use the legacy function signature */ zfs_ioctl_register_legacy(ZFS_IOC_POOL_FREEZE, zfs_ioc_pool_freeze, zfs_secpolicy_config, NO_NAME, B_FALSE, POOL_CHECK_READONLY); zfs_ioctl_register_pool(ZFS_IOC_POOL_CREATE, zfs_ioc_pool_create, zfs_secpolicy_config, B_TRUE, POOL_CHECK_NONE); zfs_ioctl_register_pool_modify(ZFS_IOC_POOL_SCAN, zfs_ioc_pool_scan); zfs_ioctl_register_pool_modify(ZFS_IOC_POOL_UPGRADE, zfs_ioc_pool_upgrade); zfs_ioctl_register_pool_modify(ZFS_IOC_VDEV_ADD, zfs_ioc_vdev_add); zfs_ioctl_register_pool_modify(ZFS_IOC_VDEV_REMOVE, zfs_ioc_vdev_remove); zfs_ioctl_register_pool_modify(ZFS_IOC_VDEV_SET_STATE, zfs_ioc_vdev_set_state); zfs_ioctl_register_pool_modify(ZFS_IOC_VDEV_ATTACH, zfs_ioc_vdev_attach); zfs_ioctl_register_pool_modify(ZFS_IOC_VDEV_DETACH, zfs_ioc_vdev_detach); zfs_ioctl_register_pool_modify(ZFS_IOC_VDEV_SETPATH, zfs_ioc_vdev_setpath); zfs_ioctl_register_pool_modify(ZFS_IOC_VDEV_SETFRU, zfs_ioc_vdev_setfru); zfs_ioctl_register_pool_modify(ZFS_IOC_POOL_SET_PROPS, zfs_ioc_pool_set_props); zfs_ioctl_register_pool_modify(ZFS_IOC_VDEV_SPLIT, zfs_ioc_vdev_split); zfs_ioctl_register_pool_modify(ZFS_IOC_POOL_REGUID, zfs_ioc_pool_reguid); zfs_ioctl_register_pool_meta(ZFS_IOC_POOL_CONFIGS, zfs_ioc_pool_configs, zfs_secpolicy_none); zfs_ioctl_register_pool_meta(ZFS_IOC_POOL_TRYIMPORT, zfs_ioc_pool_tryimport, zfs_secpolicy_config); zfs_ioctl_register_pool_meta(ZFS_IOC_INJECT_FAULT, zfs_ioc_inject_fault, zfs_secpolicy_inject); zfs_ioctl_register_pool_meta(ZFS_IOC_CLEAR_FAULT, zfs_ioc_clear_fault, zfs_secpolicy_inject); zfs_ioctl_register_pool_meta(ZFS_IOC_INJECT_LIST_NEXT, zfs_ioc_inject_list_next, zfs_secpolicy_inject); /* * pool destroy, and export don't log the history as part of * zfsdev_ioctl, but rather zfs_ioc_pool_export * does the logging of those commands. */ zfs_ioctl_register_pool(ZFS_IOC_POOL_DESTROY, zfs_ioc_pool_destroy, zfs_secpolicy_config, B_FALSE, POOL_CHECK_SUSPENDED); zfs_ioctl_register_pool(ZFS_IOC_POOL_EXPORT, zfs_ioc_pool_export, zfs_secpolicy_config, B_FALSE, POOL_CHECK_SUSPENDED); zfs_ioctl_register_pool(ZFS_IOC_POOL_STATS, zfs_ioc_pool_stats, zfs_secpolicy_read, B_FALSE, POOL_CHECK_NONE); zfs_ioctl_register_pool(ZFS_IOC_POOL_GET_PROPS, zfs_ioc_pool_get_props, zfs_secpolicy_read, B_FALSE, POOL_CHECK_NONE); zfs_ioctl_register_pool(ZFS_IOC_ERROR_LOG, zfs_ioc_error_log, zfs_secpolicy_inject, B_FALSE, POOL_CHECK_SUSPENDED); zfs_ioctl_register_pool(ZFS_IOC_DSOBJ_TO_DSNAME, zfs_ioc_dsobj_to_dsname, zfs_secpolicy_diff, B_FALSE, POOL_CHECK_SUSPENDED); zfs_ioctl_register_pool(ZFS_IOC_POOL_GET_HISTORY, zfs_ioc_pool_get_history, zfs_secpolicy_config, B_FALSE, POOL_CHECK_SUSPENDED); zfs_ioctl_register_pool(ZFS_IOC_POOL_IMPORT, zfs_ioc_pool_import, zfs_secpolicy_config, B_TRUE, POOL_CHECK_NONE); zfs_ioctl_register_pool(ZFS_IOC_CLEAR, zfs_ioc_clear, zfs_secpolicy_config, B_TRUE, POOL_CHECK_READONLY); zfs_ioctl_register_dataset_read(ZFS_IOC_SPACE_WRITTEN, zfs_ioc_space_written); zfs_ioctl_register_dataset_read(ZFS_IOC_OBJSET_RECVD_PROPS, zfs_ioc_objset_recvd_props); zfs_ioctl_register_dataset_read(ZFS_IOC_NEXT_OBJ, zfs_ioc_next_obj); zfs_ioctl_register_dataset_read(ZFS_IOC_GET_FSACL, zfs_ioc_get_fsacl); zfs_ioctl_register_dataset_read(ZFS_IOC_OBJSET_STATS, zfs_ioc_objset_stats); zfs_ioctl_register_dataset_read(ZFS_IOC_OBJSET_ZPLPROPS, zfs_ioc_objset_zplprops); zfs_ioctl_register_dataset_read(ZFS_IOC_DATASET_LIST_NEXT, zfs_ioc_dataset_list_next); zfs_ioctl_register_dataset_read(ZFS_IOC_SNAPSHOT_LIST_NEXT, zfs_ioc_snapshot_list_next); zfs_ioctl_register_dataset_read(ZFS_IOC_SEND_PROGRESS, zfs_ioc_send_progress); zfs_ioctl_register_dataset_read_secpolicy(ZFS_IOC_DIFF, zfs_ioc_diff, zfs_secpolicy_diff); zfs_ioctl_register_dataset_read_secpolicy(ZFS_IOC_OBJ_TO_STATS, zfs_ioc_obj_to_stats, zfs_secpolicy_diff); zfs_ioctl_register_dataset_read_secpolicy(ZFS_IOC_OBJ_TO_PATH, zfs_ioc_obj_to_path, zfs_secpolicy_diff); zfs_ioctl_register_dataset_read_secpolicy(ZFS_IOC_USERSPACE_ONE, zfs_ioc_userspace_one, zfs_secpolicy_userspace_one); zfs_ioctl_register_dataset_read_secpolicy(ZFS_IOC_USERSPACE_MANY, zfs_ioc_userspace_many, zfs_secpolicy_userspace_many); zfs_ioctl_register_dataset_read_secpolicy(ZFS_IOC_SEND, zfs_ioc_send, zfs_secpolicy_send); zfs_ioctl_register_dataset_modify(ZFS_IOC_SET_PROP, zfs_ioc_set_prop, zfs_secpolicy_none); zfs_ioctl_register_dataset_modify(ZFS_IOC_DESTROY, zfs_ioc_destroy, zfs_secpolicy_destroy); zfs_ioctl_register_dataset_modify(ZFS_IOC_RENAME, zfs_ioc_rename, zfs_secpolicy_rename); zfs_ioctl_register_dataset_modify(ZFS_IOC_RECV, zfs_ioc_recv, zfs_secpolicy_recv); zfs_ioctl_register_dataset_modify(ZFS_IOC_PROMOTE, zfs_ioc_promote, zfs_secpolicy_promote); zfs_ioctl_register_dataset_modify(ZFS_IOC_INHERIT_PROP, zfs_ioc_inherit_prop, zfs_secpolicy_inherit_prop); zfs_ioctl_register_dataset_modify(ZFS_IOC_SET_FSACL, zfs_ioc_set_fsacl, zfs_secpolicy_set_fsacl); zfs_ioctl_register_dataset_nolog(ZFS_IOC_SHARE, zfs_ioc_share, zfs_secpolicy_share, POOL_CHECK_NONE); zfs_ioctl_register_dataset_nolog(ZFS_IOC_SMB_ACL, zfs_ioc_smb_acl, zfs_secpolicy_smb_acl, POOL_CHECK_NONE); zfs_ioctl_register_dataset_nolog(ZFS_IOC_USERSPACE_UPGRADE, zfs_ioc_userspace_upgrade, zfs_secpolicy_userspace_upgrade, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY); zfs_ioctl_register_dataset_nolog(ZFS_IOC_TMP_SNAPSHOT, zfs_ioc_tmp_snapshot, zfs_secpolicy_tmp_snapshot, POOL_CHECK_SUSPENDED | POOL_CHECK_READONLY); zfs_ioctl_register_legacy(ZFS_IOC_EVENTS_NEXT, zfs_ioc_events_next, zfs_secpolicy_config, NO_NAME, B_FALSE, POOL_CHECK_NONE); zfs_ioctl_register_legacy(ZFS_IOC_EVENTS_CLEAR, zfs_ioc_events_clear, zfs_secpolicy_config, NO_NAME, B_FALSE, POOL_CHECK_NONE); zfs_ioctl_register_legacy(ZFS_IOC_EVENTS_SEEK, zfs_ioc_events_seek, zfs_secpolicy_config, NO_NAME, B_FALSE, POOL_CHECK_NONE); zfs_ioctl_init_os(); } /* * Verify that for non-legacy ioctls the input nvlist * pairs match against the expected input. * * Possible errors are: * ZFS_ERR_IOC_ARG_UNAVAIL An unrecognized nvpair was encountered * ZFS_ERR_IOC_ARG_REQUIRED A required nvpair is missing * ZFS_ERR_IOC_ARG_BADTYPE Invalid type for nvpair */ static int zfs_check_input_nvpairs(nvlist_t *innvl, const zfs_ioc_vec_t *vec) { const zfs_ioc_key_t *nvl_keys = vec->zvec_nvl_keys; boolean_t required_keys_found = B_FALSE; /* * examine each input pair */ for (nvpair_t *pair = nvlist_next_nvpair(innvl, NULL); pair != NULL; pair = nvlist_next_nvpair(innvl, pair)) { const char *name = nvpair_name(pair); data_type_t type = nvpair_type(pair); boolean_t identified = B_FALSE; /* * check pair against the documented names and type */ for (int k = 0; k < vec->zvec_nvl_key_count; k++) { /* if not a wild card name, check for an exact match */ if ((nvl_keys[k].zkey_flags & ZK_WILDCARDLIST) == 0 && strcmp(nvl_keys[k].zkey_name, name) != 0) continue; identified = B_TRUE; if (nvl_keys[k].zkey_type != DATA_TYPE_ANY && nvl_keys[k].zkey_type != type) { return (SET_ERROR(ZFS_ERR_IOC_ARG_BADTYPE)); } if (nvl_keys[k].zkey_flags & ZK_OPTIONAL) continue; required_keys_found = B_TRUE; break; } /* allow an 'optional' key, everything else is invalid */ if (!identified && (strcmp(name, "optional") != 0 || type != DATA_TYPE_NVLIST)) { return (SET_ERROR(ZFS_ERR_IOC_ARG_UNAVAIL)); } } /* verify that all required keys were found */ for (int k = 0; k < vec->zvec_nvl_key_count; k++) { if (nvl_keys[k].zkey_flags & ZK_OPTIONAL) continue; if (nvl_keys[k].zkey_flags & ZK_WILDCARDLIST) { /* at least one non-optional key is expected here */ if (!required_keys_found) return (SET_ERROR(ZFS_ERR_IOC_ARG_REQUIRED)); continue; } if (!nvlist_exists(innvl, nvl_keys[k].zkey_name)) return (SET_ERROR(ZFS_ERR_IOC_ARG_REQUIRED)); } return (0); } static int pool_status_check(const char *name, zfs_ioc_namecheck_t type, zfs_ioc_poolcheck_t check) { spa_t *spa; int error; ASSERT(type == POOL_NAME || type == DATASET_NAME || type == ENTITY_NAME); if (check & POOL_CHECK_NONE) return (0); error = spa_open(name, &spa, FTAG); if (error == 0) { if ((check & POOL_CHECK_SUSPENDED) && spa_suspended(spa)) error = SET_ERROR(EAGAIN); else if ((check & POOL_CHECK_READONLY) && !spa_writeable(spa)) error = SET_ERROR(EROFS); spa_close(spa, FTAG); } return (error); } int zfsdev_getminor(zfs_file_t *fp, minor_t *minorp) { zfsdev_state_t *zs, *fpd; ASSERT(!MUTEX_HELD(&zfsdev_state_lock)); fpd = zfs_file_private(fp); if (fpd == NULL) return (SET_ERROR(EBADF)); mutex_enter(&zfsdev_state_lock); for (zs = &zfsdev_state_listhead; zs != NULL; zs = zs->zs_next) { if (zs->zs_minor == -1) continue; if (fpd == zs) { *minorp = fpd->zs_minor; mutex_exit(&zfsdev_state_lock); return (0); } } mutex_exit(&zfsdev_state_lock); return (SET_ERROR(EBADF)); } void * zfsdev_get_state(minor_t minor, enum zfsdev_state_type which) { zfsdev_state_t *zs; for (zs = &zfsdev_state_listhead; zs != NULL; zs = zs->zs_next) { if (zs->zs_minor == minor) { membar_consumer(); switch (which) { case ZST_ONEXIT: return (zs->zs_onexit); case ZST_ZEVENT: return (zs->zs_zevent); case ZST_ALL: return (zs); } } } return (NULL); } /* * Find a free minor number. The zfsdev_state_list is expected to * be short since it is only a list of currently open file handles. */ static minor_t zfsdev_minor_alloc(void) { static minor_t last_minor = 0; minor_t m; ASSERT(MUTEX_HELD(&zfsdev_state_lock)); for (m = last_minor + 1; m != last_minor; m++) { if (m > ZFSDEV_MAX_MINOR) m = 1; if (zfsdev_get_state(m, ZST_ALL) == NULL) { last_minor = m; return (m); } } return (0); } int zfsdev_state_init(void *priv) { zfsdev_state_t *zs, *zsprev = NULL; minor_t minor; boolean_t newzs = B_FALSE; ASSERT(MUTEX_HELD(&zfsdev_state_lock)); minor = zfsdev_minor_alloc(); if (minor == 0) return (SET_ERROR(ENXIO)); for (zs = &zfsdev_state_listhead; zs != NULL; zs = zs->zs_next) { if (zs->zs_minor == -1) break; zsprev = zs; } if (!zs) { zs = kmem_zalloc(sizeof (zfsdev_state_t), KM_SLEEP); newzs = B_TRUE; } zfsdev_private_set_state(priv, zs); zfs_onexit_init((zfs_onexit_t **)&zs->zs_onexit); zfs_zevent_init((zfs_zevent_t **)&zs->zs_zevent); /* * In order to provide for lock-free concurrent read access * to the minor list in zfsdev_get_state(), new entries * must be completely written before linking them into the * list whereas existing entries are already linked; the last * operation must be updating zs_minor (from -1 to the new * value). */ if (newzs) { zs->zs_minor = minor; membar_producer(); zsprev->zs_next = zs; } else { membar_producer(); zs->zs_minor = minor; } return (0); } void zfsdev_state_destroy(void *priv) { zfsdev_state_t *zs = zfsdev_private_get_state(priv); ASSERT(zs != NULL); ASSERT3S(zs->zs_minor, >, 0); /* * The last reference to this zfsdev file descriptor is being dropped. * We don't have to worry about lookup grabbing this state object, and * zfsdev_state_init() will not try to reuse this object until it is * invalidated by setting zs_minor to -1. Invalidation must be done * last, with a memory barrier to ensure ordering. This lets us avoid * taking the global zfsdev state lock around destruction. */ zfs_onexit_destroy(zs->zs_onexit); zfs_zevent_destroy(zs->zs_zevent); zs->zs_onexit = NULL; zs->zs_zevent = NULL; membar_producer(); zs->zs_minor = -1; } long zfsdev_ioctl_common(uint_t vecnum, zfs_cmd_t *zc, int flag) { int error, cmd; const zfs_ioc_vec_t *vec; char *saved_poolname = NULL; uint64_t max_nvlist_src_size; size_t saved_poolname_len = 0; nvlist_t *innvl = NULL; fstrans_cookie_t cookie; hrtime_t start_time = gethrtime(); cmd = vecnum; error = 0; if (vecnum >= sizeof (zfs_ioc_vec) / sizeof (zfs_ioc_vec[0])) return (SET_ERROR(ZFS_ERR_IOC_CMD_UNAVAIL)); vec = &zfs_ioc_vec[vecnum]; /* * The registered ioctl list may be sparse, verify that either * a normal or legacy handler are registered. */ if (vec->zvec_func == NULL && vec->zvec_legacy_func == NULL) return (SET_ERROR(ZFS_ERR_IOC_CMD_UNAVAIL)); zc->zc_iflags = flag & FKIOCTL; max_nvlist_src_size = zfs_max_nvlist_src_size_os(); if (zc->zc_nvlist_src_size > max_nvlist_src_size) { /* * Make sure the user doesn't pass in an insane value for * zc_nvlist_src_size. We have to check, since we will end * up allocating that much memory inside of get_nvlist(). This * prevents a nefarious user from allocating tons of kernel * memory. * * Also, we return EINVAL instead of ENOMEM here. The reason * being that returning ENOMEM from an ioctl() has a special * connotation; that the user's size value is too small and * needs to be expanded to hold the nvlist. See * zcmd_expand_dst_nvlist() for details. */ error = SET_ERROR(EINVAL); /* User's size too big */ } else if (zc->zc_nvlist_src_size != 0) { error = get_nvlist(zc->zc_nvlist_src, zc->zc_nvlist_src_size, zc->zc_iflags, &innvl); if (error != 0) goto out; } /* * Ensure that all pool/dataset names are valid before we pass down to * the lower layers. */ zc->zc_name[sizeof (zc->zc_name) - 1] = '\0'; switch (vec->zvec_namecheck) { case POOL_NAME: if (pool_namecheck(zc->zc_name, NULL, NULL) != 0) error = SET_ERROR(EINVAL); else error = pool_status_check(zc->zc_name, vec->zvec_namecheck, vec->zvec_pool_check); break; case DATASET_NAME: if (dataset_namecheck(zc->zc_name, NULL, NULL) != 0) error = SET_ERROR(EINVAL); else error = pool_status_check(zc->zc_name, vec->zvec_namecheck, vec->zvec_pool_check); break; case ENTITY_NAME: if (entity_namecheck(zc->zc_name, NULL, NULL) != 0) { error = SET_ERROR(EINVAL); } else { error = pool_status_check(zc->zc_name, vec->zvec_namecheck, vec->zvec_pool_check); } break; case NO_NAME: break; } /* * Ensure that all input pairs are valid before we pass them down * to the lower layers. * * The vectored functions can use fnvlist_lookup_{type} for any * required pairs since zfs_check_input_nvpairs() confirmed that * they exist and are of the correct type. */ if (error == 0 && vec->zvec_func != NULL) { error = zfs_check_input_nvpairs(innvl, vec); if (error != 0) goto out; } if (error == 0) { cookie = spl_fstrans_mark(); error = vec->zvec_secpolicy(zc, innvl, CRED()); spl_fstrans_unmark(cookie); } if (error != 0) goto out; /* legacy ioctls can modify zc_name */ /* * Can't use kmem_strdup() as we might truncate the string and * kmem_strfree() would then free with incorrect size. */ saved_poolname_len = strlen(zc->zc_name) + 1; saved_poolname = kmem_alloc(saved_poolname_len, KM_SLEEP); strlcpy(saved_poolname, zc->zc_name, saved_poolname_len); saved_poolname[strcspn(saved_poolname, "/@#")] = '\0'; if (vec->zvec_func != NULL) { nvlist_t *outnvl; int puterror = 0; spa_t *spa; nvlist_t *lognv = NULL; ASSERT(vec->zvec_legacy_func == NULL); /* * Add the innvl to the lognv before calling the func, * in case the func changes the innvl. */ if (vec->zvec_allow_log) { lognv = fnvlist_alloc(); fnvlist_add_string(lognv, ZPOOL_HIST_IOCTL, vec->zvec_name); if (!nvlist_empty(innvl)) { fnvlist_add_nvlist(lognv, ZPOOL_HIST_INPUT_NVL, innvl); } } outnvl = fnvlist_alloc(); cookie = spl_fstrans_mark(); error = vec->zvec_func(zc->zc_name, innvl, outnvl); spl_fstrans_unmark(cookie); /* * Some commands can partially execute, modify state, and still * return an error. In these cases, attempt to record what * was modified. */ if ((error == 0 || (cmd == ZFS_IOC_CHANNEL_PROGRAM && error != EINVAL)) && vec->zvec_allow_log && spa_open(zc->zc_name, &spa, FTAG) == 0) { if (!nvlist_empty(outnvl)) { size_t out_size = fnvlist_size(outnvl); if (out_size > zfs_history_output_max) { fnvlist_add_int64(lognv, ZPOOL_HIST_OUTPUT_SIZE, out_size); } else { fnvlist_add_nvlist(lognv, ZPOOL_HIST_OUTPUT_NVL, outnvl); } } if (error != 0) { fnvlist_add_int64(lognv, ZPOOL_HIST_ERRNO, error); } fnvlist_add_int64(lognv, ZPOOL_HIST_ELAPSED_NS, gethrtime() - start_time); (void) spa_history_log_nvl(spa, lognv); spa_close(spa, FTAG); } fnvlist_free(lognv); if (!nvlist_empty(outnvl) || zc->zc_nvlist_dst_size != 0) { int smusherror = 0; if (vec->zvec_smush_outnvlist) { smusherror = nvlist_smush(outnvl, zc->zc_nvlist_dst_size); } if (smusherror == 0) puterror = put_nvlist(zc, outnvl); } if (puterror != 0) error = puterror; nvlist_free(outnvl); } else { cookie = spl_fstrans_mark(); error = vec->zvec_legacy_func(zc); spl_fstrans_unmark(cookie); } out: nvlist_free(innvl); if (error == 0 && vec->zvec_allow_log) { char *s = tsd_get(zfs_allow_log_key); if (s != NULL) kmem_strfree(s); (void) tsd_set(zfs_allow_log_key, kmem_strdup(saved_poolname)); } if (saved_poolname != NULL) kmem_free(saved_poolname, saved_poolname_len); return (error); } int zfs_kmod_init(void) { int error; if ((error = zvol_init()) != 0) return (error); spa_init(SPA_MODE_READ | SPA_MODE_WRITE); zfs_init(); zfs_ioctl_init(); mutex_init(&zfsdev_state_lock, NULL, MUTEX_DEFAULT, NULL); zfsdev_state_listhead.zs_minor = -1; if ((error = zfsdev_attach()) != 0) goto out; tsd_create(&zfs_fsyncer_key, NULL); tsd_create(&rrw_tsd_key, rrw_tsd_destroy); tsd_create(&zfs_allow_log_key, zfs_allow_log_destroy); return (0); out: zfs_fini(); spa_fini(); zvol_fini(); return (error); } void zfs_kmod_fini(void) { zfsdev_state_t *zs, *zsnext = NULL; zfsdev_detach(); mutex_destroy(&zfsdev_state_lock); for (zs = &zfsdev_state_listhead; zs != NULL; zs = zsnext) { zsnext = zs->zs_next; if (zs->zs_onexit) zfs_onexit_destroy(zs->zs_onexit); if (zs->zs_zevent) zfs_zevent_destroy(zs->zs_zevent); if (zs != &zfsdev_state_listhead) kmem_free(zs, sizeof (zfsdev_state_t)); } zfs_ereport_taskq_fini(); /* run before zfs_fini() on Linux */ zfs_fini(); spa_fini(); zvol_fini(); tsd_destroy(&zfs_fsyncer_key); tsd_destroy(&rrw_tsd_key); tsd_destroy(&zfs_allow_log_key); } ZFS_MODULE_PARAM(zfs, zfs_, max_nvlist_src_size, U64, ZMOD_RW, "Maximum size in bytes allowed for src nvlist passed with ZFS ioctls"); ZFS_MODULE_PARAM(zfs, zfs_, history_output_max, U64, ZMOD_RW, "Maximum size in bytes of ZFS ioctl output that will be logged"); diff --git a/sys/contrib/openzfs/module/zfs/zil.c b/sys/contrib/openzfs/module/zfs/zil.c index c37da89dd438..d887e4900d1d 100644 --- a/sys/contrib/openzfs/module/zfs/zil.c +++ b/sys/contrib/openzfs/module/zfs/zil.c @@ -1,4124 +1,4089 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2011, 2018 by Delphix. All rights reserved. * Copyright (c) 2014 Integros [integros.com] * Copyright (c) 2018 Datto Inc. */ /* Portions Copyright 2010 Robert Milkowski */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include /* * The ZFS Intent Log (ZIL) saves "transaction records" (itxs) of system * calls that change the file system. Each itx has enough information to * be able to replay them after a system crash, power loss, or * equivalent failure mode. These are stored in memory until either: * * 1. they are committed to the pool by the DMU transaction group * (txg), at which point they can be discarded; or * 2. they are committed to the on-disk ZIL for the dataset being * modified (e.g. due to an fsync, O_DSYNC, or other synchronous * requirement). * * In the event of a crash or power loss, the itxs contained by each * dataset's on-disk ZIL will be replayed when that dataset is first * instantiated (e.g. if the dataset is a normal filesystem, when it is * first mounted). * * As hinted at above, there is one ZIL per dataset (both the in-memory * representation, and the on-disk representation). The on-disk format * consists of 3 parts: * * - a single, per-dataset, ZIL header; which points to a chain of * - zero or more ZIL blocks; each of which contains * - zero or more ZIL records * * A ZIL record holds the information necessary to replay a single * system call transaction. A ZIL block can hold many ZIL records, and * the blocks are chained together, similarly to a singly linked list. * * Each ZIL block contains a block pointer (blkptr_t) to the next ZIL * block in the chain, and the ZIL header points to the first block in * the chain. * * Note, there is not a fixed place in the pool to hold these ZIL * blocks; they are dynamically allocated and freed as needed from the * blocks available on the pool, though they can be preferentially * allocated from a dedicated "log" vdev. */ /* * This controls the amount of time that a ZIL block (lwb) will remain * "open" when it isn't "full", and it has a thread waiting for it to be * committed to stable storage. Please refer to the zil_commit_waiter() * function (and the comments within it) for more details. */ static uint_t zfs_commit_timeout_pct = 5; /* * Minimal time we care to delay commit waiting for more ZIL records. * At least FreeBSD kernel can't sleep for less than 2us at its best. * So requests to sleep for less then 5us is a waste of CPU time with * a risk of significant log latency increase due to oversleep. */ static uint64_t zil_min_commit_timeout = 5000; /* * See zil.h for more information about these fields. */ static zil_kstat_values_t zil_stats = { { "zil_commit_count", KSTAT_DATA_UINT64 }, { "zil_commit_writer_count", KSTAT_DATA_UINT64 }, { "zil_itx_count", KSTAT_DATA_UINT64 }, { "zil_itx_indirect_count", KSTAT_DATA_UINT64 }, { "zil_itx_indirect_bytes", KSTAT_DATA_UINT64 }, { "zil_itx_copied_count", KSTAT_DATA_UINT64 }, { "zil_itx_copied_bytes", KSTAT_DATA_UINT64 }, { "zil_itx_needcopy_count", KSTAT_DATA_UINT64 }, { "zil_itx_needcopy_bytes", KSTAT_DATA_UINT64 }, { "zil_itx_metaslab_normal_count", KSTAT_DATA_UINT64 }, { "zil_itx_metaslab_normal_bytes", KSTAT_DATA_UINT64 }, { "zil_itx_metaslab_slog_count", KSTAT_DATA_UINT64 }, { "zil_itx_metaslab_slog_bytes", KSTAT_DATA_UINT64 }, }; static zil_sums_t zil_sums_global; static kstat_t *zil_kstats_global; /* * Disable intent logging replay. This global ZIL switch affects all pools. */ int zil_replay_disable = 0; /* * Disable the DKIOCFLUSHWRITECACHE commands that are normally sent to * the disk(s) by the ZIL after an LWB write has completed. Setting this * will cause ZIL corruption on power loss if a volatile out-of-order * write cache is enabled. */ static int zil_nocacheflush = 0; /* * Limit SLOG write size per commit executed with synchronous priority. * Any writes above that will be executed with lower (asynchronous) priority * to limit potential SLOG device abuse by single active ZIL writer. */ static uint64_t zil_slog_bulk = 768 * 1024; static kmem_cache_t *zil_lwb_cache; static kmem_cache_t *zil_zcw_cache; -#define LWB_EMPTY(lwb) ((BP_GET_LSIZE(&lwb->lwb_blk) - \ - sizeof (zil_chain_t)) == (lwb->lwb_sz - lwb->lwb_nused)) - static int zil_bp_compare(const void *x1, const void *x2) { const dva_t *dva1 = &((zil_bp_node_t *)x1)->zn_dva; const dva_t *dva2 = &((zil_bp_node_t *)x2)->zn_dva; int cmp = TREE_CMP(DVA_GET_VDEV(dva1), DVA_GET_VDEV(dva2)); if (likely(cmp)) return (cmp); return (TREE_CMP(DVA_GET_OFFSET(dva1), DVA_GET_OFFSET(dva2))); } static void zil_bp_tree_init(zilog_t *zilog) { avl_create(&zilog->zl_bp_tree, zil_bp_compare, sizeof (zil_bp_node_t), offsetof(zil_bp_node_t, zn_node)); } static void zil_bp_tree_fini(zilog_t *zilog) { avl_tree_t *t = &zilog->zl_bp_tree; zil_bp_node_t *zn; void *cookie = NULL; while ((zn = avl_destroy_nodes(t, &cookie)) != NULL) kmem_free(zn, sizeof (zil_bp_node_t)); avl_destroy(t); } int zil_bp_tree_add(zilog_t *zilog, const blkptr_t *bp) { avl_tree_t *t = &zilog->zl_bp_tree; const dva_t *dva; zil_bp_node_t *zn; avl_index_t where; if (BP_IS_EMBEDDED(bp)) return (0); dva = BP_IDENTITY(bp); if (avl_find(t, dva, &where) != NULL) return (SET_ERROR(EEXIST)); zn = kmem_alloc(sizeof (zil_bp_node_t), KM_SLEEP); zn->zn_dva = *dva; avl_insert(t, zn, where); return (0); } static zil_header_t * zil_header_in_syncing_context(zilog_t *zilog) { return ((zil_header_t *)zilog->zl_header); } static void zil_init_log_chain(zilog_t *zilog, blkptr_t *bp) { zio_cksum_t *zc = &bp->blk_cksum; (void) random_get_pseudo_bytes((void *)&zc->zc_word[ZIL_ZC_GUID_0], sizeof (zc->zc_word[ZIL_ZC_GUID_0])); (void) random_get_pseudo_bytes((void *)&zc->zc_word[ZIL_ZC_GUID_1], sizeof (zc->zc_word[ZIL_ZC_GUID_1])); zc->zc_word[ZIL_ZC_OBJSET] = dmu_objset_id(zilog->zl_os); zc->zc_word[ZIL_ZC_SEQ] = 1ULL; } static int zil_kstats_global_update(kstat_t *ksp, int rw) { zil_kstat_values_t *zs = ksp->ks_data; ASSERT3P(&zil_stats, ==, zs); if (rw == KSTAT_WRITE) { return (SET_ERROR(EACCES)); } zil_kstat_values_update(zs, &zil_sums_global); return (0); } /* * Read a log block and make sure it's valid. */ static int zil_read_log_block(zilog_t *zilog, boolean_t decrypt, const blkptr_t *bp, blkptr_t *nbp, void *dst, char **end) { zio_flag_t zio_flags = ZIO_FLAG_CANFAIL; arc_flags_t aflags = ARC_FLAG_WAIT; arc_buf_t *abuf = NULL; zbookmark_phys_t zb; int error; if (zilog->zl_header->zh_claim_txg == 0) zio_flags |= ZIO_FLAG_SPECULATIVE | ZIO_FLAG_SCRUB; if (!(zilog->zl_header->zh_flags & ZIL_CLAIM_LR_SEQ_VALID)) zio_flags |= ZIO_FLAG_SPECULATIVE; if (!decrypt) zio_flags |= ZIO_FLAG_RAW; SET_BOOKMARK(&zb, bp->blk_cksum.zc_word[ZIL_ZC_OBJSET], ZB_ZIL_OBJECT, ZB_ZIL_LEVEL, bp->blk_cksum.zc_word[ZIL_ZC_SEQ]); error = arc_read(NULL, zilog->zl_spa, bp, arc_getbuf_func, &abuf, ZIO_PRIORITY_SYNC_READ, zio_flags, &aflags, &zb); if (error == 0) { zio_cksum_t cksum = bp->blk_cksum; /* * Validate the checksummed log block. * * Sequence numbers should be... sequential. The checksum * verifier for the next block should be bp's checksum plus 1. * * Also check the log chain linkage and size used. */ cksum.zc_word[ZIL_ZC_SEQ]++; if (BP_GET_CHECKSUM(bp) == ZIO_CHECKSUM_ZILOG2) { zil_chain_t *zilc = abuf->b_data; char *lr = (char *)(zilc + 1); uint64_t len = zilc->zc_nused - sizeof (zil_chain_t); if (memcmp(&cksum, &zilc->zc_next_blk.blk_cksum, sizeof (cksum)) || BP_IS_HOLE(&zilc->zc_next_blk)) { error = SET_ERROR(ECKSUM); } else { ASSERT3U(len, <=, SPA_OLD_MAXBLOCKSIZE); memcpy(dst, lr, len); *end = (char *)dst + len; *nbp = zilc->zc_next_blk; } } else { char *lr = abuf->b_data; uint64_t size = BP_GET_LSIZE(bp); zil_chain_t *zilc = (zil_chain_t *)(lr + size) - 1; if (memcmp(&cksum, &zilc->zc_next_blk.blk_cksum, sizeof (cksum)) || BP_IS_HOLE(&zilc->zc_next_blk) || (zilc->zc_nused > (size - sizeof (*zilc)))) { error = SET_ERROR(ECKSUM); } else { ASSERT3U(zilc->zc_nused, <=, SPA_OLD_MAXBLOCKSIZE); memcpy(dst, lr, zilc->zc_nused); *end = (char *)dst + zilc->zc_nused; *nbp = zilc->zc_next_blk; } } arc_buf_destroy(abuf, &abuf); } return (error); } /* * Read a TX_WRITE log data block. */ static int zil_read_log_data(zilog_t *zilog, const lr_write_t *lr, void *wbuf) { zio_flag_t zio_flags = ZIO_FLAG_CANFAIL; const blkptr_t *bp = &lr->lr_blkptr; arc_flags_t aflags = ARC_FLAG_WAIT; arc_buf_t *abuf = NULL; zbookmark_phys_t zb; int error; if (BP_IS_HOLE(bp)) { if (wbuf != NULL) memset(wbuf, 0, MAX(BP_GET_LSIZE(bp), lr->lr_length)); return (0); } if (zilog->zl_header->zh_claim_txg == 0) zio_flags |= ZIO_FLAG_SPECULATIVE | ZIO_FLAG_SCRUB; /* * If we are not using the resulting data, we are just checking that * it hasn't been corrupted so we don't need to waste CPU time * decompressing and decrypting it. */ if (wbuf == NULL) zio_flags |= ZIO_FLAG_RAW; ASSERT3U(BP_GET_LSIZE(bp), !=, 0); SET_BOOKMARK(&zb, dmu_objset_id(zilog->zl_os), lr->lr_foid, ZB_ZIL_LEVEL, lr->lr_offset / BP_GET_LSIZE(bp)); error = arc_read(NULL, zilog->zl_spa, bp, arc_getbuf_func, &abuf, ZIO_PRIORITY_SYNC_READ, zio_flags, &aflags, &zb); if (error == 0) { if (wbuf != NULL) memcpy(wbuf, abuf->b_data, arc_buf_size(abuf)); arc_buf_destroy(abuf, &abuf); } return (error); } void zil_sums_init(zil_sums_t *zs) { wmsum_init(&zs->zil_commit_count, 0); wmsum_init(&zs->zil_commit_writer_count, 0); wmsum_init(&zs->zil_itx_count, 0); wmsum_init(&zs->zil_itx_indirect_count, 0); wmsum_init(&zs->zil_itx_indirect_bytes, 0); wmsum_init(&zs->zil_itx_copied_count, 0); wmsum_init(&zs->zil_itx_copied_bytes, 0); wmsum_init(&zs->zil_itx_needcopy_count, 0); wmsum_init(&zs->zil_itx_needcopy_bytes, 0); wmsum_init(&zs->zil_itx_metaslab_normal_count, 0); wmsum_init(&zs->zil_itx_metaslab_normal_bytes, 0); wmsum_init(&zs->zil_itx_metaslab_slog_count, 0); wmsum_init(&zs->zil_itx_metaslab_slog_bytes, 0); } void zil_sums_fini(zil_sums_t *zs) { wmsum_fini(&zs->zil_commit_count); wmsum_fini(&zs->zil_commit_writer_count); wmsum_fini(&zs->zil_itx_count); wmsum_fini(&zs->zil_itx_indirect_count); wmsum_fini(&zs->zil_itx_indirect_bytes); wmsum_fini(&zs->zil_itx_copied_count); wmsum_fini(&zs->zil_itx_copied_bytes); wmsum_fini(&zs->zil_itx_needcopy_count); wmsum_fini(&zs->zil_itx_needcopy_bytes); wmsum_fini(&zs->zil_itx_metaslab_normal_count); wmsum_fini(&zs->zil_itx_metaslab_normal_bytes); wmsum_fini(&zs->zil_itx_metaslab_slog_count); wmsum_fini(&zs->zil_itx_metaslab_slog_bytes); } void zil_kstat_values_update(zil_kstat_values_t *zs, zil_sums_t *zil_sums) { zs->zil_commit_count.value.ui64 = wmsum_value(&zil_sums->zil_commit_count); zs->zil_commit_writer_count.value.ui64 = wmsum_value(&zil_sums->zil_commit_writer_count); zs->zil_itx_count.value.ui64 = wmsum_value(&zil_sums->zil_itx_count); zs->zil_itx_indirect_count.value.ui64 = wmsum_value(&zil_sums->zil_itx_indirect_count); zs->zil_itx_indirect_bytes.value.ui64 = wmsum_value(&zil_sums->zil_itx_indirect_bytes); zs->zil_itx_copied_count.value.ui64 = wmsum_value(&zil_sums->zil_itx_copied_count); zs->zil_itx_copied_bytes.value.ui64 = wmsum_value(&zil_sums->zil_itx_copied_bytes); zs->zil_itx_needcopy_count.value.ui64 = wmsum_value(&zil_sums->zil_itx_needcopy_count); zs->zil_itx_needcopy_bytes.value.ui64 = wmsum_value(&zil_sums->zil_itx_needcopy_bytes); zs->zil_itx_metaslab_normal_count.value.ui64 = wmsum_value(&zil_sums->zil_itx_metaslab_normal_count); zs->zil_itx_metaslab_normal_bytes.value.ui64 = wmsum_value(&zil_sums->zil_itx_metaslab_normal_bytes); zs->zil_itx_metaslab_slog_count.value.ui64 = wmsum_value(&zil_sums->zil_itx_metaslab_slog_count); zs->zil_itx_metaslab_slog_bytes.value.ui64 = wmsum_value(&zil_sums->zil_itx_metaslab_slog_bytes); } /* * Parse the intent log, and call parse_func for each valid record within. */ int zil_parse(zilog_t *zilog, zil_parse_blk_func_t *parse_blk_func, zil_parse_lr_func_t *parse_lr_func, void *arg, uint64_t txg, boolean_t decrypt) { const zil_header_t *zh = zilog->zl_header; boolean_t claimed = !!zh->zh_claim_txg; uint64_t claim_blk_seq = claimed ? zh->zh_claim_blk_seq : UINT64_MAX; uint64_t claim_lr_seq = claimed ? zh->zh_claim_lr_seq : UINT64_MAX; uint64_t max_blk_seq = 0; uint64_t max_lr_seq = 0; uint64_t blk_count = 0; uint64_t lr_count = 0; blkptr_t blk, next_blk = {{{{0}}}}; char *lrbuf, *lrp; int error = 0; /* * Old logs didn't record the maximum zh_claim_lr_seq. */ if (!(zh->zh_flags & ZIL_CLAIM_LR_SEQ_VALID)) claim_lr_seq = UINT64_MAX; /* * Starting at the block pointed to by zh_log we read the log chain. * For each block in the chain we strongly check that block to * ensure its validity. We stop when an invalid block is found. * For each block pointer in the chain we call parse_blk_func(). * For each record in each valid block we call parse_lr_func(). * If the log has been claimed, stop if we encounter a sequence * number greater than the highest claimed sequence number. */ lrbuf = zio_buf_alloc(SPA_OLD_MAXBLOCKSIZE); zil_bp_tree_init(zilog); for (blk = zh->zh_log; !BP_IS_HOLE(&blk); blk = next_blk) { uint64_t blk_seq = blk.blk_cksum.zc_word[ZIL_ZC_SEQ]; int reclen; char *end = NULL; if (blk_seq > claim_blk_seq) break; error = parse_blk_func(zilog, &blk, arg, txg); if (error != 0) break; ASSERT3U(max_blk_seq, <, blk_seq); max_blk_seq = blk_seq; blk_count++; if (max_lr_seq == claim_lr_seq && max_blk_seq == claim_blk_seq) break; error = zil_read_log_block(zilog, decrypt, &blk, &next_blk, lrbuf, &end); if (error != 0) { if (claimed) { char name[ZFS_MAX_DATASET_NAME_LEN]; dmu_objset_name(zilog->zl_os, name); cmn_err(CE_WARN, "ZFS read log block error %d, " "dataset %s, seq 0x%llx\n", error, name, (u_longlong_t)blk_seq); } break; } for (lrp = lrbuf; lrp < end; lrp += reclen) { lr_t *lr = (lr_t *)lrp; reclen = lr->lrc_reclen; ASSERT3U(reclen, >=, sizeof (lr_t)); if (lr->lrc_seq > claim_lr_seq) goto done; error = parse_lr_func(zilog, lr, arg, txg); if (error != 0) goto done; ASSERT3U(max_lr_seq, <, lr->lrc_seq); max_lr_seq = lr->lrc_seq; lr_count++; } } done: zilog->zl_parse_error = error; zilog->zl_parse_blk_seq = max_blk_seq; zilog->zl_parse_lr_seq = max_lr_seq; zilog->zl_parse_blk_count = blk_count; zilog->zl_parse_lr_count = lr_count; zil_bp_tree_fini(zilog); zio_buf_free(lrbuf, SPA_OLD_MAXBLOCKSIZE); return (error); } static int zil_clear_log_block(zilog_t *zilog, const blkptr_t *bp, void *tx, uint64_t first_txg) { (void) tx; ASSERT(!BP_IS_HOLE(bp)); /* * As we call this function from the context of a rewind to a * checkpoint, each ZIL block whose txg is later than the txg * that we rewind to is invalid. Thus, we return -1 so * zil_parse() doesn't attempt to read it. */ if (bp->blk_birth >= first_txg) return (-1); if (zil_bp_tree_add(zilog, bp) != 0) return (0); zio_free(zilog->zl_spa, first_txg, bp); return (0); } static int zil_noop_log_record(zilog_t *zilog, const lr_t *lrc, void *tx, uint64_t first_txg) { (void) zilog, (void) lrc, (void) tx, (void) first_txg; return (0); } static int zil_claim_log_block(zilog_t *zilog, const blkptr_t *bp, void *tx, uint64_t first_txg) { /* * Claim log block if not already committed and not already claimed. * If tx == NULL, just verify that the block is claimable. */ if (BP_IS_HOLE(bp) || bp->blk_birth < first_txg || zil_bp_tree_add(zilog, bp) != 0) return (0); return (zio_wait(zio_claim(NULL, zilog->zl_spa, tx == NULL ? 0 : first_txg, bp, spa_claim_notify, NULL, ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE | ZIO_FLAG_SCRUB))); } static int zil_claim_write(zilog_t *zilog, const lr_t *lrc, void *tx, uint64_t first_txg) { lr_write_t *lr = (lr_write_t *)lrc; int error; ASSERT(lrc->lrc_txtype == TX_WRITE); /* * If the block is not readable, don't claim it. This can happen * in normal operation when a log block is written to disk before * some of the dmu_sync() blocks it points to. In this case, the * transaction cannot have been committed to anyone (we would have * waited for all writes to be stable first), so it is semantically * correct to declare this the end of the log. */ if (lr->lr_blkptr.blk_birth >= first_txg) { error = zil_read_log_data(zilog, lr, NULL); if (error != 0) return (error); } return (zil_claim_log_block(zilog, &lr->lr_blkptr, tx, first_txg)); } static int zil_claim_clone_range(zilog_t *zilog, const lr_t *lrc, void *tx) { const lr_clone_range_t *lr = (const lr_clone_range_t *)lrc; const blkptr_t *bp; spa_t *spa; uint_t ii; ASSERT(lrc->lrc_txtype == TX_CLONE_RANGE); if (tx == NULL) { return (0); } /* * XXX: Do we need to byteswap lr? */ spa = zilog->zl_spa; for (ii = 0; ii < lr->lr_nbps; ii++) { bp = &lr->lr_bps[ii]; /* * When data in embedded into BP there is no need to create * BRT entry as there is no data block. Just copy the BP as * it contains the data. */ if (!BP_IS_HOLE(bp) && !BP_IS_EMBEDDED(bp)) { brt_pending_add(spa, bp, tx); } } return (0); } static int zil_claim_log_record(zilog_t *zilog, const lr_t *lrc, void *tx, uint64_t first_txg) { switch (lrc->lrc_txtype) { case TX_WRITE: return (zil_claim_write(zilog, lrc, tx, first_txg)); case TX_CLONE_RANGE: return (zil_claim_clone_range(zilog, lrc, tx)); default: return (0); } } static int zil_free_log_block(zilog_t *zilog, const blkptr_t *bp, void *tx, uint64_t claim_txg) { (void) claim_txg; zio_free(zilog->zl_spa, dmu_tx_get_txg(tx), bp); return (0); } static int zil_free_write(zilog_t *zilog, const lr_t *lrc, void *tx, uint64_t claim_txg) { lr_write_t *lr = (lr_write_t *)lrc; blkptr_t *bp = &lr->lr_blkptr; ASSERT(lrc->lrc_txtype == TX_WRITE); /* * If we previously claimed it, we need to free it. */ if (bp->blk_birth >= claim_txg && zil_bp_tree_add(zilog, bp) == 0 && !BP_IS_HOLE(bp)) { zio_free(zilog->zl_spa, dmu_tx_get_txg(tx), bp); } return (0); } static int zil_free_clone_range(zilog_t *zilog, const lr_t *lrc, void *tx) { const lr_clone_range_t *lr = (const lr_clone_range_t *)lrc; const blkptr_t *bp; spa_t *spa; uint_t ii; ASSERT(lrc->lrc_txtype == TX_CLONE_RANGE); if (tx == NULL) { return (0); } spa = zilog->zl_spa; for (ii = 0; ii < lr->lr_nbps; ii++) { bp = &lr->lr_bps[ii]; if (!BP_IS_HOLE(bp)) { zio_free(spa, dmu_tx_get_txg(tx), bp); } } return (0); } static int zil_free_log_record(zilog_t *zilog, const lr_t *lrc, void *tx, uint64_t claim_txg) { if (claim_txg == 0) { return (0); } switch (lrc->lrc_txtype) { case TX_WRITE: return (zil_free_write(zilog, lrc, tx, claim_txg)); case TX_CLONE_RANGE: return (zil_free_clone_range(zilog, lrc, tx)); default: return (0); } } static int zil_lwb_vdev_compare(const void *x1, const void *x2) { const uint64_t v1 = ((zil_vdev_node_t *)x1)->zv_vdev; const uint64_t v2 = ((zil_vdev_node_t *)x2)->zv_vdev; return (TREE_CMP(v1, v2)); } static lwb_t * zil_alloc_lwb(zilog_t *zilog, blkptr_t *bp, boolean_t slog, uint64_t txg, boolean_t fastwrite) { lwb_t *lwb; lwb = kmem_cache_alloc(zil_lwb_cache, KM_SLEEP); lwb->lwb_zilog = zilog; lwb->lwb_blk = *bp; lwb->lwb_fastwrite = fastwrite; lwb->lwb_slog = slog; lwb->lwb_state = LWB_STATE_CLOSED; lwb->lwb_buf = zio_buf_alloc(BP_GET_LSIZE(bp)); lwb->lwb_max_txg = txg; lwb->lwb_write_zio = NULL; lwb->lwb_root_zio = NULL; lwb->lwb_issued_timestamp = 0; lwb->lwb_issued_txg = 0; if (BP_GET_CHECKSUM(bp) == ZIO_CHECKSUM_ZILOG2) { lwb->lwb_nused = sizeof (zil_chain_t); lwb->lwb_sz = BP_GET_LSIZE(bp); } else { lwb->lwb_nused = 0; lwb->lwb_sz = BP_GET_LSIZE(bp) - sizeof (zil_chain_t); } mutex_enter(&zilog->zl_lock); list_insert_tail(&zilog->zl_lwb_list, lwb); mutex_exit(&zilog->zl_lock); - ASSERT(!MUTEX_HELD(&lwb->lwb_vdev_lock)); - ASSERT(avl_is_empty(&lwb->lwb_vdev_tree)); - VERIFY(list_is_empty(&lwb->lwb_waiters)); - VERIFY(list_is_empty(&lwb->lwb_itxs)); - return (lwb); } static void zil_free_lwb(zilog_t *zilog, lwb_t *lwb) { ASSERT(MUTEX_HELD(&zilog->zl_lock)); ASSERT(!MUTEX_HELD(&lwb->lwb_vdev_lock)); - VERIFY(list_is_empty(&lwb->lwb_waiters)); - VERIFY(list_is_empty(&lwb->lwb_itxs)); + ASSERT(list_is_empty(&lwb->lwb_waiters)); + ASSERT(list_is_empty(&lwb->lwb_itxs)); ASSERT(avl_is_empty(&lwb->lwb_vdev_tree)); ASSERT3P(lwb->lwb_write_zio, ==, NULL); ASSERT3P(lwb->lwb_root_zio, ==, NULL); ASSERT3U(lwb->lwb_max_txg, <=, spa_syncing_txg(zilog->zl_spa)); ASSERT(lwb->lwb_state == LWB_STATE_CLOSED || lwb->lwb_state == LWB_STATE_FLUSH_DONE); /* * Clear the zilog's field to indicate this lwb is no longer * valid, and prevent use-after-free errors. */ if (zilog->zl_last_lwb_opened == lwb) zilog->zl_last_lwb_opened = NULL; kmem_cache_free(zil_lwb_cache, lwb); } /* * Called when we create in-memory log transactions so that we know * to cleanup the itxs at the end of spa_sync(). */ static void zilog_dirty(zilog_t *zilog, uint64_t txg) { dsl_pool_t *dp = zilog->zl_dmu_pool; dsl_dataset_t *ds = dmu_objset_ds(zilog->zl_os); ASSERT(spa_writeable(zilog->zl_spa)); if (ds->ds_is_snapshot) panic("dirtying snapshot!"); if (txg_list_add(&dp->dp_dirty_zilogs, zilog, txg)) { /* up the hold count until we can be written out */ dmu_buf_add_ref(ds->ds_dbuf, zilog); zilog->zl_dirty_max_txg = MAX(txg, zilog->zl_dirty_max_txg); } } /* * Determine if the zil is dirty in the specified txg. Callers wanting to * ensure that the dirty state does not change must hold the itxg_lock for * the specified txg. Holding the lock will ensure that the zil cannot be * dirtied (zil_itx_assign) or cleaned (zil_clean) while we check its current * state. */ static boolean_t __maybe_unused zilog_is_dirty_in_txg(zilog_t *zilog, uint64_t txg) { dsl_pool_t *dp = zilog->zl_dmu_pool; if (txg_list_member(&dp->dp_dirty_zilogs, zilog, txg & TXG_MASK)) return (B_TRUE); return (B_FALSE); } /* * Determine if the zil is dirty. The zil is considered dirty if it has * any pending itx records that have not been cleaned by zil_clean(). */ static boolean_t zilog_is_dirty(zilog_t *zilog) { dsl_pool_t *dp = zilog->zl_dmu_pool; for (int t = 0; t < TXG_SIZE; t++) { if (txg_list_member(&dp->dp_dirty_zilogs, zilog, t)) return (B_TRUE); } return (B_FALSE); } /* * Its called in zil_commit context (zil_process_commit_list()/zil_create()). * It activates SPA_FEATURE_ZILSAXATTR feature, if its enabled. * Check dsl_dataset_feature_is_active to avoid txg_wait_synced() on every * zil_commit. */ static void zil_commit_activate_saxattr_feature(zilog_t *zilog) { dsl_dataset_t *ds = dmu_objset_ds(zilog->zl_os); uint64_t txg = 0; dmu_tx_t *tx = NULL; if (spa_feature_is_enabled(zilog->zl_spa, SPA_FEATURE_ZILSAXATTR) && dmu_objset_type(zilog->zl_os) != DMU_OST_ZVOL && !dsl_dataset_feature_is_active(ds, SPA_FEATURE_ZILSAXATTR)) { tx = dmu_tx_create(zilog->zl_os); VERIFY0(dmu_tx_assign(tx, TXG_WAIT)); dsl_dataset_dirty(ds, tx); txg = dmu_tx_get_txg(tx); mutex_enter(&ds->ds_lock); ds->ds_feature_activation[SPA_FEATURE_ZILSAXATTR] = (void *)B_TRUE; mutex_exit(&ds->ds_lock); dmu_tx_commit(tx); txg_wait_synced(zilog->zl_dmu_pool, txg); } } /* * Create an on-disk intent log. */ static lwb_t * zil_create(zilog_t *zilog) { const zil_header_t *zh = zilog->zl_header; lwb_t *lwb = NULL; uint64_t txg = 0; dmu_tx_t *tx = NULL; blkptr_t blk; int error = 0; boolean_t fastwrite = FALSE; boolean_t slog = FALSE; dsl_dataset_t *ds = dmu_objset_ds(zilog->zl_os); /* * Wait for any previous destroy to complete. */ txg_wait_synced(zilog->zl_dmu_pool, zilog->zl_destroy_txg); ASSERT(zh->zh_claim_txg == 0); ASSERT(zh->zh_replay_seq == 0); blk = zh->zh_log; /* * Allocate an initial log block if: * - there isn't one already * - the existing block is the wrong endianness */ if (BP_IS_HOLE(&blk) || BP_SHOULD_BYTESWAP(&blk)) { tx = dmu_tx_create(zilog->zl_os); VERIFY0(dmu_tx_assign(tx, TXG_WAIT)); dsl_dataset_dirty(dmu_objset_ds(zilog->zl_os), tx); txg = dmu_tx_get_txg(tx); if (!BP_IS_HOLE(&blk)) { zio_free(zilog->zl_spa, txg, &blk); BP_ZERO(&blk); } error = zio_alloc_zil(zilog->zl_spa, zilog->zl_os, txg, &blk, ZIL_MIN_BLKSZ, &slog); fastwrite = TRUE; if (error == 0) zil_init_log_chain(zilog, &blk); } /* * Allocate a log write block (lwb) for the first log block. */ if (error == 0) lwb = zil_alloc_lwb(zilog, &blk, slog, txg, fastwrite); /* * If we just allocated the first log block, commit our transaction * and wait for zil_sync() to stuff the block pointer into zh_log. * (zh is part of the MOS, so we cannot modify it in open context.) */ if (tx != NULL) { /* * If "zilsaxattr" feature is enabled on zpool, then activate * it now when we're creating the ZIL chain. We can't wait with * this until we write the first xattr log record because we * need to wait for the feature activation to sync out. */ if (spa_feature_is_enabled(zilog->zl_spa, SPA_FEATURE_ZILSAXATTR) && dmu_objset_type(zilog->zl_os) != DMU_OST_ZVOL) { mutex_enter(&ds->ds_lock); ds->ds_feature_activation[SPA_FEATURE_ZILSAXATTR] = (void *)B_TRUE; mutex_exit(&ds->ds_lock); } dmu_tx_commit(tx); txg_wait_synced(zilog->zl_dmu_pool, txg); } else { /* * This branch covers the case where we enable the feature on a * zpool that has existing ZIL headers. */ zil_commit_activate_saxattr_feature(zilog); } IMPLY(spa_feature_is_enabled(zilog->zl_spa, SPA_FEATURE_ZILSAXATTR) && dmu_objset_type(zilog->zl_os) != DMU_OST_ZVOL, dsl_dataset_feature_is_active(ds, SPA_FEATURE_ZILSAXATTR)); ASSERT(error != 0 || memcmp(&blk, &zh->zh_log, sizeof (blk)) == 0); IMPLY(error == 0, lwb != NULL); return (lwb); } /* * In one tx, free all log blocks and clear the log header. If keep_first * is set, then we're replaying a log with no content. We want to keep the * first block, however, so that the first synchronous transaction doesn't * require a txg_wait_synced() in zil_create(). We don't need to * txg_wait_synced() here either when keep_first is set, because both * zil_create() and zil_destroy() will wait for any in-progress destroys * to complete. * Return B_TRUE if there were any entries to replay. */ boolean_t zil_destroy(zilog_t *zilog, boolean_t keep_first) { const zil_header_t *zh = zilog->zl_header; lwb_t *lwb; dmu_tx_t *tx; uint64_t txg; /* * Wait for any previous destroy to complete. */ txg_wait_synced(zilog->zl_dmu_pool, zilog->zl_destroy_txg); zilog->zl_old_header = *zh; /* debugging aid */ if (BP_IS_HOLE(&zh->zh_log)) return (B_FALSE); tx = dmu_tx_create(zilog->zl_os); VERIFY0(dmu_tx_assign(tx, TXG_WAIT)); dsl_dataset_dirty(dmu_objset_ds(zilog->zl_os), tx); txg = dmu_tx_get_txg(tx); mutex_enter(&zilog->zl_lock); ASSERT3U(zilog->zl_destroy_txg, <, txg); zilog->zl_destroy_txg = txg; zilog->zl_keep_first = keep_first; if (!list_is_empty(&zilog->zl_lwb_list)) { ASSERT(zh->zh_claim_txg == 0); VERIFY(!keep_first); - while ((lwb = list_head(&zilog->zl_lwb_list)) != NULL) { + while ((lwb = list_remove_head(&zilog->zl_lwb_list)) != NULL) { if (lwb->lwb_fastwrite) metaslab_fastwrite_unmark(zilog->zl_spa, &lwb->lwb_blk); - - list_remove(&zilog->zl_lwb_list, lwb); if (lwb->lwb_buf != NULL) zio_buf_free(lwb->lwb_buf, lwb->lwb_sz); zio_free(zilog->zl_spa, txg, &lwb->lwb_blk); zil_free_lwb(zilog, lwb); } } else if (!keep_first) { zil_destroy_sync(zilog, tx); } mutex_exit(&zilog->zl_lock); dmu_tx_commit(tx); return (B_TRUE); } void zil_destroy_sync(zilog_t *zilog, dmu_tx_t *tx) { ASSERT(list_is_empty(&zilog->zl_lwb_list)); (void) zil_parse(zilog, zil_free_log_block, zil_free_log_record, tx, zilog->zl_header->zh_claim_txg, B_FALSE); } int zil_claim(dsl_pool_t *dp, dsl_dataset_t *ds, void *txarg) { dmu_tx_t *tx = txarg; zilog_t *zilog; uint64_t first_txg; zil_header_t *zh; objset_t *os; int error; error = dmu_objset_own_obj(dp, ds->ds_object, DMU_OST_ANY, B_FALSE, B_FALSE, FTAG, &os); if (error != 0) { /* * EBUSY indicates that the objset is inconsistent, in which * case it can not have a ZIL. */ if (error != EBUSY) { cmn_err(CE_WARN, "can't open objset for %llu, error %u", (unsigned long long)ds->ds_object, error); } return (0); } zilog = dmu_objset_zil(os); zh = zil_header_in_syncing_context(zilog); ASSERT3U(tx->tx_txg, ==, spa_first_txg(zilog->zl_spa)); first_txg = spa_min_claim_txg(zilog->zl_spa); /* * If the spa_log_state is not set to be cleared, check whether * the current uberblock is a checkpoint one and if the current * header has been claimed before moving on. * * If the current uberblock is a checkpointed uberblock then * one of the following scenarios took place: * * 1] We are currently rewinding to the checkpoint of the pool. * 2] We crashed in the middle of a checkpoint rewind but we * did manage to write the checkpointed uberblock to the * vdev labels, so when we tried to import the pool again * the checkpointed uberblock was selected from the import * procedure. * * In both cases we want to zero out all the ZIL blocks, except * the ones that have been claimed at the time of the checkpoint * (their zh_claim_txg != 0). The reason is that these blocks * may be corrupted since we may have reused their locations on * disk after we took the checkpoint. * * We could try to set spa_log_state to SPA_LOG_CLEAR earlier * when we first figure out whether the current uberblock is * checkpointed or not. Unfortunately, that would discard all * the logs, including the ones that are claimed, and we would * leak space. */ if (spa_get_log_state(zilog->zl_spa) == SPA_LOG_CLEAR || (zilog->zl_spa->spa_uberblock.ub_checkpoint_txg != 0 && zh->zh_claim_txg == 0)) { if (!BP_IS_HOLE(&zh->zh_log)) { (void) zil_parse(zilog, zil_clear_log_block, zil_noop_log_record, tx, first_txg, B_FALSE); } BP_ZERO(&zh->zh_log); if (os->os_encrypted) os->os_next_write_raw[tx->tx_txg & TXG_MASK] = B_TRUE; dsl_dataset_dirty(dmu_objset_ds(os), tx); dmu_objset_disown(os, B_FALSE, FTAG); return (0); } /* * If we are not rewinding and opening the pool normally, then * the min_claim_txg should be equal to the first txg of the pool. */ ASSERT3U(first_txg, ==, spa_first_txg(zilog->zl_spa)); /* * Claim all log blocks if we haven't already done so, and remember * the highest claimed sequence number. This ensures that if we can * read only part of the log now (e.g. due to a missing device), * but we can read the entire log later, we will not try to replay * or destroy beyond the last block we successfully claimed. */ ASSERT3U(zh->zh_claim_txg, <=, first_txg); if (zh->zh_claim_txg == 0 && !BP_IS_HOLE(&zh->zh_log)) { (void) zil_parse(zilog, zil_claim_log_block, zil_claim_log_record, tx, first_txg, B_FALSE); zh->zh_claim_txg = first_txg; zh->zh_claim_blk_seq = zilog->zl_parse_blk_seq; zh->zh_claim_lr_seq = zilog->zl_parse_lr_seq; if (zilog->zl_parse_lr_count || zilog->zl_parse_blk_count > 1) zh->zh_flags |= ZIL_REPLAY_NEEDED; zh->zh_flags |= ZIL_CLAIM_LR_SEQ_VALID; if (os->os_encrypted) os->os_next_write_raw[tx->tx_txg & TXG_MASK] = B_TRUE; dsl_dataset_dirty(dmu_objset_ds(os), tx); } ASSERT3U(first_txg, ==, (spa_last_synced_txg(zilog->zl_spa) + 1)); dmu_objset_disown(os, B_FALSE, FTAG); return (0); } /* * Check the log by walking the log chain. * Checksum errors are ok as they indicate the end of the chain. * Any other error (no device or read failure) returns an error. */ int zil_check_log_chain(dsl_pool_t *dp, dsl_dataset_t *ds, void *tx) { (void) dp; zilog_t *zilog; objset_t *os; blkptr_t *bp; int error; ASSERT(tx == NULL); error = dmu_objset_from_ds(ds, &os); if (error != 0) { cmn_err(CE_WARN, "can't open objset %llu, error %d", (unsigned long long)ds->ds_object, error); return (0); } zilog = dmu_objset_zil(os); bp = (blkptr_t *)&zilog->zl_header->zh_log; if (!BP_IS_HOLE(bp)) { vdev_t *vd; boolean_t valid = B_TRUE; /* * Check the first block and determine if it's on a log device * which may have been removed or faulted prior to loading this * pool. If so, there's no point in checking the rest of the * log as its content should have already been synced to the * pool. */ spa_config_enter(os->os_spa, SCL_STATE, FTAG, RW_READER); vd = vdev_lookup_top(os->os_spa, DVA_GET_VDEV(&bp->blk_dva[0])); if (vd->vdev_islog && vdev_is_dead(vd)) valid = vdev_log_state_valid(vd); spa_config_exit(os->os_spa, SCL_STATE, FTAG); if (!valid) return (0); /* * Check whether the current uberblock is checkpointed (e.g. * we are rewinding) and whether the current header has been * claimed or not. If it hasn't then skip verifying it. We * do this because its ZIL blocks may be part of the pool's * state before the rewind, which is no longer valid. */ zil_header_t *zh = zil_header_in_syncing_context(zilog); if (zilog->zl_spa->spa_uberblock.ub_checkpoint_txg != 0 && zh->zh_claim_txg == 0) return (0); } /* * Because tx == NULL, zil_claim_log_block() will not actually claim * any blocks, but just determine whether it is possible to do so. * In addition to checking the log chain, zil_claim_log_block() * will invoke zio_claim() with a done func of spa_claim_notify(), * which will update spa_max_claim_txg. See spa_load() for details. */ error = zil_parse(zilog, zil_claim_log_block, zil_claim_log_record, tx, zilog->zl_header->zh_claim_txg ? -1ULL : spa_min_claim_txg(os->os_spa), B_FALSE); return ((error == ECKSUM || error == ENOENT) ? 0 : error); } /* * When an itx is "skipped", this function is used to properly mark the * waiter as "done, and signal any thread(s) waiting on it. An itx can * be skipped (and not committed to an lwb) for a variety of reasons, * one of them being that the itx was committed via spa_sync(), prior to * it being committed to an lwb; this can happen if a thread calling * zil_commit() is racing with spa_sync(). */ static void zil_commit_waiter_skip(zil_commit_waiter_t *zcw) { mutex_enter(&zcw->zcw_lock); ASSERT3B(zcw->zcw_done, ==, B_FALSE); zcw->zcw_done = B_TRUE; cv_broadcast(&zcw->zcw_cv); mutex_exit(&zcw->zcw_lock); } /* * This function is used when the given waiter is to be linked into an * lwb's "lwb_waiter" list; i.e. when the itx is committed to the lwb. * At this point, the waiter will no longer be referenced by the itx, * and instead, will be referenced by the lwb. */ static void zil_commit_waiter_link_lwb(zil_commit_waiter_t *zcw, lwb_t *lwb) { /* * The lwb_waiters field of the lwb is protected by the zilog's * zl_lock, thus it must be held when calling this function. */ ASSERT(MUTEX_HELD(&lwb->lwb_zilog->zl_lock)); mutex_enter(&zcw->zcw_lock); ASSERT(!list_link_active(&zcw->zcw_node)); ASSERT3P(zcw->zcw_lwb, ==, NULL); ASSERT3P(lwb, !=, NULL); ASSERT(lwb->lwb_state == LWB_STATE_OPENED || lwb->lwb_state == LWB_STATE_ISSUED || lwb->lwb_state == LWB_STATE_WRITE_DONE); list_insert_tail(&lwb->lwb_waiters, zcw); zcw->zcw_lwb = lwb; mutex_exit(&zcw->zcw_lock); } /* * This function is used when zio_alloc_zil() fails to allocate a ZIL * block, and the given waiter must be linked to the "nolwb waiters" * list inside of zil_process_commit_list(). */ static void zil_commit_waiter_link_nolwb(zil_commit_waiter_t *zcw, list_t *nolwb) { mutex_enter(&zcw->zcw_lock); ASSERT(!list_link_active(&zcw->zcw_node)); ASSERT3P(zcw->zcw_lwb, ==, NULL); list_insert_tail(nolwb, zcw); mutex_exit(&zcw->zcw_lock); } void zil_lwb_add_block(lwb_t *lwb, const blkptr_t *bp) { avl_tree_t *t = &lwb->lwb_vdev_tree; avl_index_t where; zil_vdev_node_t *zv, zvsearch; int ndvas = BP_GET_NDVAS(bp); int i; if (zil_nocacheflush) return; mutex_enter(&lwb->lwb_vdev_lock); for (i = 0; i < ndvas; i++) { zvsearch.zv_vdev = DVA_GET_VDEV(&bp->blk_dva[i]); if (avl_find(t, &zvsearch, &where) == NULL) { zv = kmem_alloc(sizeof (*zv), KM_SLEEP); zv->zv_vdev = zvsearch.zv_vdev; avl_insert(t, zv, where); } } mutex_exit(&lwb->lwb_vdev_lock); } static void zil_lwb_flush_defer(lwb_t *lwb, lwb_t *nlwb) { avl_tree_t *src = &lwb->lwb_vdev_tree; avl_tree_t *dst = &nlwb->lwb_vdev_tree; void *cookie = NULL; zil_vdev_node_t *zv; ASSERT3S(lwb->lwb_state, ==, LWB_STATE_WRITE_DONE); ASSERT3S(nlwb->lwb_state, !=, LWB_STATE_WRITE_DONE); ASSERT3S(nlwb->lwb_state, !=, LWB_STATE_FLUSH_DONE); /* * While 'lwb' is at a point in its lifetime where lwb_vdev_tree does * not need the protection of lwb_vdev_lock (it will only be modified * while holding zilog->zl_lock) as its writes and those of its * children have all completed. The younger 'nlwb' may be waiting on * future writes to additional vdevs. */ mutex_enter(&nlwb->lwb_vdev_lock); /* * Tear down the 'lwb' vdev tree, ensuring that entries which do not * exist in 'nlwb' are moved to it, freeing any would-be duplicates. */ while ((zv = avl_destroy_nodes(src, &cookie)) != NULL) { avl_index_t where; if (avl_find(dst, zv, &where) == NULL) { avl_insert(dst, zv, where); } else { kmem_free(zv, sizeof (*zv)); } } mutex_exit(&nlwb->lwb_vdev_lock); } void zil_lwb_add_txg(lwb_t *lwb, uint64_t txg) { lwb->lwb_max_txg = MAX(lwb->lwb_max_txg, txg); } /* * This function is a called after all vdevs associated with a given lwb * write have completed their DKIOCFLUSHWRITECACHE command; or as soon * as the lwb write completes, if "zil_nocacheflush" is set. Further, * all "previous" lwb's will have completed before this function is * called; i.e. this function is called for all previous lwbs before * it's called for "this" lwb (enforced via zio the dependencies * configured in zil_lwb_set_zio_dependency()). * * The intention is for this function to be called as soon as the * contents of an lwb are considered "stable" on disk, and will survive * any sudden loss of power. At this point, any threads waiting for the * lwb to reach this state are signalled, and the "waiter" structures * are marked "done". */ static void zil_lwb_flush_vdevs_done(zio_t *zio) { lwb_t *lwb = zio->io_private; zilog_t *zilog = lwb->lwb_zilog; zil_commit_waiter_t *zcw; itx_t *itx; uint64_t txg; spa_config_exit(zilog->zl_spa, SCL_STATE, lwb); - zio_buf_free(lwb->lwb_buf, lwb->lwb_sz); + hrtime_t t = gethrtime() - lwb->lwb_issued_timestamp; mutex_enter(&zilog->zl_lock); - /* - * If we have had an allocation failure and the txg is - * waiting to sync then we want zil_sync() to remove the lwb so - * that it's not picked up as the next new one in - * zil_process_commit_list(). zil_sync() will only remove the - * lwb if lwb_buf is null. - */ - lwb->lwb_buf = NULL; - - ASSERT3U(lwb->lwb_issued_timestamp, >, 0); - zilog->zl_last_lwb_latency = (zilog->zl_last_lwb_latency * 3 + - gethrtime() - lwb->lwb_issued_timestamp) / 4; + zilog->zl_last_lwb_latency = (zilog->zl_last_lwb_latency * 7 + t) / 8; lwb->lwb_root_zio = NULL; ASSERT3S(lwb->lwb_state, ==, LWB_STATE_WRITE_DONE); lwb->lwb_state = LWB_STATE_FLUSH_DONE; if (zilog->zl_last_lwb_opened == lwb) { /* * Remember the highest committed log sequence number * for ztest. We only update this value when all the log * writes succeeded, because ztest wants to ASSERT that * it got the whole log chain. */ zilog->zl_commit_lr_seq = zilog->zl_lr_seq; } - while ((itx = list_head(&lwb->lwb_itxs)) != NULL) { - list_remove(&lwb->lwb_itxs, itx); + while ((itx = list_remove_head(&lwb->lwb_itxs)) != NULL) zil_itx_destroy(itx); - } - while ((zcw = list_head(&lwb->lwb_waiters)) != NULL) { + while ((zcw = list_remove_head(&lwb->lwb_waiters)) != NULL) { mutex_enter(&zcw->zcw_lock); - ASSERT(list_link_active(&zcw->zcw_node)); - list_remove(&lwb->lwb_waiters, zcw); - ASSERT3P(zcw->zcw_lwb, ==, lwb); zcw->zcw_lwb = NULL; /* * We expect any ZIO errors from child ZIOs to have been * propagated "up" to this specific LWB's root ZIO, in * order for this error handling to work correctly. This * includes ZIO errors from either this LWB's write or * flush, as well as any errors from other dependent LWBs * (e.g. a root LWB ZIO that might be a child of this LWB). * * With that said, it's important to note that LWB flush * errors are not propagated up to the LWB root ZIO. * This is incorrect behavior, and results in VDEV flush * errors not being handled correctly here. See the * comment above the call to "zio_flush" for details. */ zcw->zcw_zio_error = zio->io_error; ASSERT3B(zcw->zcw_done, ==, B_FALSE); zcw->zcw_done = B_TRUE; cv_broadcast(&zcw->zcw_cv); mutex_exit(&zcw->zcw_lock); } mutex_exit(&zilog->zl_lock); mutex_enter(&zilog->zl_lwb_io_lock); txg = lwb->lwb_issued_txg; ASSERT3U(zilog->zl_lwb_inflight[txg & TXG_MASK], >, 0); zilog->zl_lwb_inflight[txg & TXG_MASK]--; if (zilog->zl_lwb_inflight[txg & TXG_MASK] == 0) cv_broadcast(&zilog->zl_lwb_io_cv); mutex_exit(&zilog->zl_lwb_io_lock); } /* * Wait for the completion of all issued write/flush of that txg provided. * It guarantees zil_lwb_flush_vdevs_done() is called and returned. */ static void zil_lwb_flush_wait_all(zilog_t *zilog, uint64_t txg) { ASSERT3U(txg, ==, spa_syncing_txg(zilog->zl_spa)); mutex_enter(&zilog->zl_lwb_io_lock); while (zilog->zl_lwb_inflight[txg & TXG_MASK] > 0) cv_wait(&zilog->zl_lwb_io_cv, &zilog->zl_lwb_io_lock); mutex_exit(&zilog->zl_lwb_io_lock); #ifdef ZFS_DEBUG mutex_enter(&zilog->zl_lock); mutex_enter(&zilog->zl_lwb_io_lock); lwb_t *lwb = list_head(&zilog->zl_lwb_list); while (lwb != NULL && lwb->lwb_max_txg <= txg) { if (lwb->lwb_issued_txg <= txg) { ASSERT(lwb->lwb_state != LWB_STATE_ISSUED); ASSERT(lwb->lwb_state != LWB_STATE_WRITE_DONE); IMPLY(lwb->lwb_issued_txg > 0, lwb->lwb_state == LWB_STATE_FLUSH_DONE); } - IMPLY(lwb->lwb_state == LWB_STATE_FLUSH_DONE, + IMPLY(lwb->lwb_state == LWB_STATE_WRITE_DONE || + lwb->lwb_state == LWB_STATE_FLUSH_DONE, lwb->lwb_buf == NULL); lwb = list_next(&zilog->zl_lwb_list, lwb); } mutex_exit(&zilog->zl_lwb_io_lock); mutex_exit(&zilog->zl_lock); #endif } /* * This is called when an lwb's write zio completes. The callback's * purpose is to issue the DKIOCFLUSHWRITECACHE commands for the vdevs * in the lwb's lwb_vdev_tree. The tree will contain the vdevs involved * in writing out this specific lwb's data, and in the case that cache * flushes have been deferred, vdevs involved in writing the data for * previous lwbs. The writes corresponding to all the vdevs in the * lwb_vdev_tree will have completed by the time this is called, due to * the zio dependencies configured in zil_lwb_set_zio_dependency(), * which takes deferred flushes into account. The lwb will be "done" * once zil_lwb_flush_vdevs_done() is called, which occurs in the zio * completion callback for the lwb's root zio. */ static void zil_lwb_write_done(zio_t *zio) { lwb_t *lwb = zio->io_private; spa_t *spa = zio->io_spa; zilog_t *zilog = lwb->lwb_zilog; avl_tree_t *t = &lwb->lwb_vdev_tree; void *cookie = NULL; zil_vdev_node_t *zv; lwb_t *nlwb; ASSERT3S(spa_config_held(spa, SCL_STATE, RW_READER), !=, 0); ASSERT(BP_GET_COMPRESS(zio->io_bp) == ZIO_COMPRESS_OFF); ASSERT(BP_GET_TYPE(zio->io_bp) == DMU_OT_INTENT_LOG); ASSERT(BP_GET_LEVEL(zio->io_bp) == 0); ASSERT(BP_GET_BYTEORDER(zio->io_bp) == ZFS_HOST_BYTEORDER); ASSERT(!BP_IS_GANG(zio->io_bp)); ASSERT(!BP_IS_HOLE(zio->io_bp)); ASSERT(BP_GET_FILL(zio->io_bp) == 0); abd_free(zio->io_abd); + zio_buf_free(lwb->lwb_buf, lwb->lwb_sz); + lwb->lwb_buf = NULL; mutex_enter(&zilog->zl_lock); ASSERT3S(lwb->lwb_state, ==, LWB_STATE_ISSUED); lwb->lwb_state = LWB_STATE_WRITE_DONE; lwb->lwb_write_zio = NULL; lwb->lwb_fastwrite = FALSE; nlwb = list_next(&zilog->zl_lwb_list, lwb); mutex_exit(&zilog->zl_lock); if (avl_numnodes(t) == 0) return; /* * If there was an IO error, we're not going to call zio_flush() * on these vdevs, so we simply empty the tree and free the * nodes. We avoid calling zio_flush() since there isn't any * good reason for doing so, after the lwb block failed to be * written out. * * Additionally, we don't perform any further error handling at * this point (e.g. setting "zcw_zio_error" appropriately), as * we expect that to occur in "zil_lwb_flush_vdevs_done" (thus, * we expect any error seen here, to have been propagated to * that function). */ if (zio->io_error != 0) { while ((zv = avl_destroy_nodes(t, &cookie)) != NULL) kmem_free(zv, sizeof (*zv)); return; } /* * If this lwb does not have any threads waiting for it to * complete, we want to defer issuing the DKIOCFLUSHWRITECACHE * command to the vdevs written to by "this" lwb, and instead * rely on the "next" lwb to handle the DKIOCFLUSHWRITECACHE * command for those vdevs. Thus, we merge the vdev tree of * "this" lwb with the vdev tree of the "next" lwb in the list, * and assume the "next" lwb will handle flushing the vdevs (or * deferring the flush(s) again). * * This is a useful performance optimization, especially for * workloads with lots of async write activity and few sync * write and/or fsync activity, as it has the potential to * coalesce multiple flush commands to a vdev into one. */ - if (list_head(&lwb->lwb_waiters) == NULL && nlwb != NULL) { + if (list_is_empty(&lwb->lwb_waiters) && nlwb != NULL) { zil_lwb_flush_defer(lwb, nlwb); ASSERT(avl_is_empty(&lwb->lwb_vdev_tree)); return; } while ((zv = avl_destroy_nodes(t, &cookie)) != NULL) { vdev_t *vd = vdev_lookup_top(spa, zv->zv_vdev); - if (vd != NULL) { + if (vd != NULL && !vd->vdev_nowritecache) { /* * The "ZIO_FLAG_DONT_PROPAGATE" is currently * always used within "zio_flush". This means, * any errors when flushing the vdev(s), will * (unfortunately) not be handled correctly, * since these "zio_flush" errors will not be * propagated up to "zil_lwb_flush_vdevs_done". */ zio_flush(lwb->lwb_root_zio, vd); } kmem_free(zv, sizeof (*zv)); } } static void zil_lwb_set_zio_dependency(zilog_t *zilog, lwb_t *lwb) { lwb_t *last_lwb_opened = zilog->zl_last_lwb_opened; ASSERT(MUTEX_HELD(&zilog->zl_issuer_lock)); ASSERT(MUTEX_HELD(&zilog->zl_lock)); /* * The zilog's "zl_last_lwb_opened" field is used to build the * lwb/zio dependency chain, which is used to preserve the * ordering of lwb completions that is required by the semantics * of the ZIL. Each new lwb zio becomes a parent of the * "previous" lwb zio, such that the new lwb's zio cannot * complete until the "previous" lwb's zio completes. * * This is required by the semantics of zil_commit(); the commit * waiters attached to the lwbs will be woken in the lwb zio's * completion callback, so this zio dependency graph ensures the * waiters are woken in the correct order (the same order the * lwbs were created). */ if (last_lwb_opened != NULL && last_lwb_opened->lwb_state != LWB_STATE_FLUSH_DONE) { ASSERT(last_lwb_opened->lwb_state == LWB_STATE_OPENED || last_lwb_opened->lwb_state == LWB_STATE_ISSUED || last_lwb_opened->lwb_state == LWB_STATE_WRITE_DONE); ASSERT3P(last_lwb_opened->lwb_root_zio, !=, NULL); zio_add_child(lwb->lwb_root_zio, last_lwb_opened->lwb_root_zio); /* * If the previous lwb's write hasn't already completed, * we also want to order the completion of the lwb write * zios (above, we only order the completion of the lwb * root zios). This is required because of how we can * defer the DKIOCFLUSHWRITECACHE commands for each lwb. * * When the DKIOCFLUSHWRITECACHE commands are deferred, * the previous lwb will rely on this lwb to flush the * vdevs written to by that previous lwb. Thus, we need * to ensure this lwb doesn't issue the flush until * after the previous lwb's write completes. We ensure * this ordering by setting the zio parent/child * relationship here. * * Without this relationship on the lwb's write zio, * it's possible for this lwb's write to complete prior * to the previous lwb's write completing; and thus, the * vdevs for the previous lwb would be flushed prior to * that lwb's data being written to those vdevs (the * vdevs are flushed in the lwb write zio's completion * handler, zil_lwb_write_done()). */ if (last_lwb_opened->lwb_state != LWB_STATE_WRITE_DONE) { ASSERT(last_lwb_opened->lwb_state == LWB_STATE_OPENED || last_lwb_opened->lwb_state == LWB_STATE_ISSUED); ASSERT3P(last_lwb_opened->lwb_write_zio, !=, NULL); zio_add_child(lwb->lwb_write_zio, last_lwb_opened->lwb_write_zio); } } } /* * This function's purpose is to "open" an lwb such that it is ready to * accept new itxs being committed to it. To do this, the lwb's zio * structures are created, and linked to the lwb. This function is * idempotent; if the passed in lwb has already been opened, this * function is essentially a no-op. */ static void zil_lwb_write_open(zilog_t *zilog, lwb_t *lwb) { zbookmark_phys_t zb; zio_priority_t prio; ASSERT(MUTEX_HELD(&zilog->zl_issuer_lock)); ASSERT3P(lwb, !=, NULL); EQUIV(lwb->lwb_root_zio == NULL, lwb->lwb_state == LWB_STATE_CLOSED); EQUIV(lwb->lwb_root_zio != NULL, lwb->lwb_state == LWB_STATE_OPENED); SET_BOOKMARK(&zb, lwb->lwb_blk.blk_cksum.zc_word[ZIL_ZC_OBJSET], ZB_ZIL_OBJECT, ZB_ZIL_LEVEL, lwb->lwb_blk.blk_cksum.zc_word[ZIL_ZC_SEQ]); /* Lock so zil_sync() doesn't fastwrite_unmark after zio is created */ mutex_enter(&zilog->zl_lock); if (lwb->lwb_root_zio == NULL) { abd_t *lwb_abd = abd_get_from_buf(lwb->lwb_buf, BP_GET_LSIZE(&lwb->lwb_blk)); if (!lwb->lwb_fastwrite) { metaslab_fastwrite_mark(zilog->zl_spa, &lwb->lwb_blk); lwb->lwb_fastwrite = 1; } if (!lwb->lwb_slog || zilog->zl_cur_used <= zil_slog_bulk) prio = ZIO_PRIORITY_SYNC_WRITE; else prio = ZIO_PRIORITY_ASYNC_WRITE; lwb->lwb_root_zio = zio_root(zilog->zl_spa, zil_lwb_flush_vdevs_done, lwb, ZIO_FLAG_CANFAIL); ASSERT3P(lwb->lwb_root_zio, !=, NULL); lwb->lwb_write_zio = zio_rewrite(lwb->lwb_root_zio, zilog->zl_spa, 0, &lwb->lwb_blk, lwb_abd, BP_GET_LSIZE(&lwb->lwb_blk), zil_lwb_write_done, lwb, prio, ZIO_FLAG_CANFAIL | ZIO_FLAG_FASTWRITE, &zb); ASSERT3P(lwb->lwb_write_zio, !=, NULL); lwb->lwb_state = LWB_STATE_OPENED; zil_lwb_set_zio_dependency(zilog, lwb); zilog->zl_last_lwb_opened = lwb; } mutex_exit(&zilog->zl_lock); ASSERT3P(lwb->lwb_root_zio, !=, NULL); ASSERT3P(lwb->lwb_write_zio, !=, NULL); ASSERT3S(lwb->lwb_state, ==, LWB_STATE_OPENED); } /* * Define a limited set of intent log block sizes. * * These must be a multiple of 4KB. Note only the amount used (again * aligned to 4KB) actually gets written. However, we can't always just * allocate SPA_OLD_MAXBLOCKSIZE as the slog space could be exhausted. */ static const struct { uint64_t limit; uint64_t blksz; } zil_block_buckets[] = { { 4096, 4096 }, /* non TX_WRITE */ { 8192 + 4096, 8192 + 4096 }, /* database */ { 32768 + 4096, 32768 + 4096 }, /* NFS writes */ { 65536 + 4096, 65536 + 4096 }, /* 64KB writes */ { 131072, 131072 }, /* < 128KB writes */ { 131072 +4096, 65536 + 4096 }, /* 128KB writes */ { UINT64_MAX, SPA_OLD_MAXBLOCKSIZE}, /* > 128KB writes */ }; /* * Maximum block size used by the ZIL. This is picked up when the ZIL is * initialized. Otherwise this should not be used directly; see * zl_max_block_size instead. */ static uint_t zil_maxblocksize = SPA_OLD_MAXBLOCKSIZE; /* * Start a log block write and advance to the next log block. * Calls are serialized. */ static lwb_t * zil_lwb_write_issue(zilog_t *zilog, lwb_t *lwb) { lwb_t *nlwb = NULL; zil_chain_t *zilc; spa_t *spa = zilog->zl_spa; blkptr_t *bp; dmu_tx_t *tx; uint64_t txg; uint64_t zil_blksz, wsz; int i, error; boolean_t slog; ASSERT(MUTEX_HELD(&zilog->zl_issuer_lock)); ASSERT3P(lwb->lwb_root_zio, !=, NULL); ASSERT3P(lwb->lwb_write_zio, !=, NULL); ASSERT3S(lwb->lwb_state, ==, LWB_STATE_OPENED); if (BP_GET_CHECKSUM(&lwb->lwb_blk) == ZIO_CHECKSUM_ZILOG2) { zilc = (zil_chain_t *)lwb->lwb_buf; bp = &zilc->zc_next_blk; } else { zilc = (zil_chain_t *)(lwb->lwb_buf + lwb->lwb_sz); bp = &zilc->zc_next_blk; } ASSERT(lwb->lwb_nused <= lwb->lwb_sz); /* * Allocate the next block and save its address in this block * before writing it in order to establish the log chain. */ tx = dmu_tx_create(zilog->zl_os); /* * Since we are not going to create any new dirty data, and we * can even help with clearing the existing dirty data, we * should not be subject to the dirty data based delays. We * use TXG_NOTHROTTLE to bypass the delay mechanism. */ VERIFY0(dmu_tx_assign(tx, TXG_WAIT | TXG_NOTHROTTLE)); dsl_dataset_dirty(dmu_objset_ds(zilog->zl_os), tx); txg = dmu_tx_get_txg(tx); mutex_enter(&zilog->zl_lwb_io_lock); lwb->lwb_issued_txg = txg; zilog->zl_lwb_inflight[txg & TXG_MASK]++; zilog->zl_lwb_max_issued_txg = MAX(txg, zilog->zl_lwb_max_issued_txg); mutex_exit(&zilog->zl_lwb_io_lock); /* * Log blocks are pre-allocated. Here we select the size of the next * block, based on size used in the last block. * - first find the smallest bucket that will fit the block from a * limited set of block sizes. This is because it's faster to write * blocks allocated from the same metaslab as they are adjacent or * close. * - next find the maximum from the new suggested size and an array of * previous sizes. This lessens a picket fence effect of wrongly * guessing the size if we have a stream of say 2k, 64k, 2k, 64k * requests. * * Note we only write what is used, but we can't just allocate * the maximum block size because we can exhaust the available * pool log space. */ zil_blksz = zilog->zl_cur_used + sizeof (zil_chain_t); for (i = 0; zil_blksz > zil_block_buckets[i].limit; i++) continue; zil_blksz = MIN(zil_block_buckets[i].blksz, zilog->zl_max_block_size); zilog->zl_prev_blks[zilog->zl_prev_rotor] = zil_blksz; for (i = 0; i < ZIL_PREV_BLKS; i++) zil_blksz = MAX(zil_blksz, zilog->zl_prev_blks[i]); zilog->zl_prev_rotor = (zilog->zl_prev_rotor + 1) & (ZIL_PREV_BLKS - 1); BP_ZERO(bp); error = zio_alloc_zil(spa, zilog->zl_os, txg, bp, zil_blksz, &slog); if (slog) { ZIL_STAT_BUMP(zilog, zil_itx_metaslab_slog_count); ZIL_STAT_INCR(zilog, zil_itx_metaslab_slog_bytes, lwb->lwb_nused); } else { ZIL_STAT_BUMP(zilog, zil_itx_metaslab_normal_count); ZIL_STAT_INCR(zilog, zil_itx_metaslab_normal_bytes, lwb->lwb_nused); } if (error == 0) { ASSERT3U(bp->blk_birth, ==, txg); bp->blk_cksum = lwb->lwb_blk.blk_cksum; bp->blk_cksum.zc_word[ZIL_ZC_SEQ]++; /* * Allocate a new log write block (lwb). */ nlwb = zil_alloc_lwb(zilog, bp, slog, txg, TRUE); } if (BP_GET_CHECKSUM(&lwb->lwb_blk) == ZIO_CHECKSUM_ZILOG2) { /* For Slim ZIL only write what is used. */ wsz = P2ROUNDUP_TYPED(lwb->lwb_nused, ZIL_MIN_BLKSZ, uint64_t); ASSERT3U(wsz, <=, lwb->lwb_sz); zio_shrink(lwb->lwb_write_zio, wsz); wsz = lwb->lwb_write_zio->io_size; } else { wsz = lwb->lwb_sz; } zilc->zc_pad = 0; zilc->zc_nused = lwb->lwb_nused; zilc->zc_eck.zec_cksum = lwb->lwb_blk.blk_cksum; /* * clear unused data for security */ memset(lwb->lwb_buf + lwb->lwb_nused, 0, wsz - lwb->lwb_nused); spa_config_enter(zilog->zl_spa, SCL_STATE, lwb, RW_READER); zil_lwb_add_block(lwb, &lwb->lwb_blk); lwb->lwb_issued_timestamp = gethrtime(); lwb->lwb_state = LWB_STATE_ISSUED; zio_nowait(lwb->lwb_root_zio); zio_nowait(lwb->lwb_write_zio); dmu_tx_commit(tx); /* * If there was an allocation failure then nlwb will be null which * forces a txg_wait_synced(). */ return (nlwb); } /* * Maximum amount of data that can be put into single log block. */ uint64_t zil_max_log_data(zilog_t *zilog, size_t hdrsize) { return (zilog->zl_max_block_size - sizeof (zil_chain_t) - hdrsize); } /* * Maximum amount of log space we agree to waste to reduce number of * WR_NEED_COPY chunks to reduce zl_get_data() overhead (~12%). */ static inline uint64_t zil_max_waste_space(zilog_t *zilog) { return (zil_max_log_data(zilog, sizeof (lr_write_t)) / 8); } /* * Maximum amount of write data for WR_COPIED. For correctness, consumers * must fall back to WR_NEED_COPY if we can't fit the entire record into one * maximum sized log block, because each WR_COPIED record must fit in a * single log block. For space efficiency, we want to fit two records into a * max-sized log block. */ uint64_t zil_max_copied_data(zilog_t *zilog) { return ((zilog->zl_max_block_size - sizeof (zil_chain_t)) / 2 - sizeof (lr_write_t)); } static lwb_t * zil_lwb_commit(zilog_t *zilog, itx_t *itx, lwb_t *lwb) { lr_t *lrcb, *lrc; lr_write_t *lrwb, *lrw; char *lr_buf; uint64_t dlen, dnow, dpad, lwb_sp, reclen, txg, max_log_data; ASSERT(MUTEX_HELD(&zilog->zl_issuer_lock)); ASSERT3P(lwb, !=, NULL); ASSERT3P(lwb->lwb_buf, !=, NULL); zil_lwb_write_open(zilog, lwb); lrc = &itx->itx_lr; lrw = (lr_write_t *)lrc; /* * A commit itx doesn't represent any on-disk state; instead * it's simply used as a place holder on the commit list, and * provides a mechanism for attaching a "commit waiter" onto the * correct lwb (such that the waiter can be signalled upon * completion of that lwb). Thus, we don't process this itx's * log record if it's a commit itx (these itx's don't have log * records), and instead link the itx's waiter onto the lwb's * list of waiters. * * For more details, see the comment above zil_commit(). */ if (lrc->lrc_txtype == TX_COMMIT) { mutex_enter(&zilog->zl_lock); zil_commit_waiter_link_lwb(itx->itx_private, lwb); itx->itx_private = NULL; mutex_exit(&zilog->zl_lock); return (lwb); } if (lrc->lrc_txtype == TX_WRITE && itx->itx_wr_state == WR_NEED_COPY) { dlen = P2ROUNDUP_TYPED( lrw->lr_length, sizeof (uint64_t), uint64_t); dpad = dlen - lrw->lr_length; } else { dlen = dpad = 0; } reclen = lrc->lrc_reclen; zilog->zl_cur_used += (reclen + dlen); txg = lrc->lrc_txg; - ASSERT3U(zilog->zl_cur_used, <, UINT64_MAX - (reclen + dlen)); - cont: /* * If this record won't fit in the current log block, start a new one. * For WR_NEED_COPY optimize layout for minimal number of chunks. */ lwb_sp = lwb->lwb_sz - lwb->lwb_nused; max_log_data = zil_max_log_data(zilog, sizeof (lr_write_t)); if (reclen > lwb_sp || (reclen + dlen > lwb_sp && lwb_sp < zil_max_waste_space(zilog) && (dlen % max_log_data == 0 || lwb_sp < reclen + dlen % max_log_data))) { lwb = zil_lwb_write_issue(zilog, lwb); if (lwb == NULL) return (NULL); zil_lwb_write_open(zilog, lwb); - ASSERT(LWB_EMPTY(lwb)); lwb_sp = lwb->lwb_sz - lwb->lwb_nused; /* * There must be enough space in the new, empty log block to * hold reclen. For WR_COPIED, we need to fit the whole * record in one block, and reclen is the header size + the * data size. For WR_NEED_COPY, we can create multiple * records, splitting the data into multiple blocks, so we * only need to fit one word of data per block; in this case * reclen is just the header size (no data). */ ASSERT3U(reclen + MIN(dlen, sizeof (uint64_t)), <=, lwb_sp); } dnow = MIN(dlen, lwb_sp - reclen); lr_buf = lwb->lwb_buf + lwb->lwb_nused; memcpy(lr_buf, lrc, reclen); lrcb = (lr_t *)lr_buf; /* Like lrc, but inside lwb. */ lrwb = (lr_write_t *)lrcb; /* Like lrw, but inside lwb. */ ZIL_STAT_BUMP(zilog, zil_itx_count); /* * If it's a write, fetch the data or get its blkptr as appropriate. */ if (lrc->lrc_txtype == TX_WRITE) { if (txg > spa_freeze_txg(zilog->zl_spa)) txg_wait_synced(zilog->zl_dmu_pool, txg); if (itx->itx_wr_state == WR_COPIED) { ZIL_STAT_BUMP(zilog, zil_itx_copied_count); ZIL_STAT_INCR(zilog, zil_itx_copied_bytes, lrw->lr_length); } else { char *dbuf; int error; if (itx->itx_wr_state == WR_NEED_COPY) { dbuf = lr_buf + reclen; lrcb->lrc_reclen += dnow; if (lrwb->lr_length > dnow) lrwb->lr_length = dnow; lrw->lr_offset += dnow; lrw->lr_length -= dnow; ZIL_STAT_BUMP(zilog, zil_itx_needcopy_count); ZIL_STAT_INCR(zilog, zil_itx_needcopy_bytes, dnow); } else { ASSERT3S(itx->itx_wr_state, ==, WR_INDIRECT); dbuf = NULL; ZIL_STAT_BUMP(zilog, zil_itx_indirect_count); ZIL_STAT_INCR(zilog, zil_itx_indirect_bytes, lrw->lr_length); } /* * We pass in the "lwb_write_zio" rather than * "lwb_root_zio" so that the "lwb_write_zio" * becomes the parent of any zio's created by * the "zl_get_data" callback. The vdevs are * flushed after the "lwb_write_zio" completes, * so we want to make sure that completion * callback waits for these additional zio's, * such that the vdevs used by those zio's will * be included in the lwb's vdev tree, and those * vdevs will be properly flushed. If we passed * in "lwb_root_zio" here, then these additional * vdevs may not be flushed; e.g. if these zio's * completed after "lwb_write_zio" completed. */ error = zilog->zl_get_data(itx->itx_private, itx->itx_gen, lrwb, dbuf, lwb, lwb->lwb_write_zio); if (dbuf != NULL && error == 0 && dnow == dlen) /* Zero any padding bytes in the last block. */ memset((char *)dbuf + lrwb->lr_length, 0, dpad); /* * Typically, the only return values we should see from * ->zl_get_data() are 0, EIO, ENOENT, EEXIST or * EALREADY. However, it is also possible to see other * error values such as ENOSPC or EINVAL from * dmu_read() -> dnode_hold() -> dnode_hold_impl() or * ENXIO as well as a multitude of others from the * block layer through dmu_buf_hold() -> dbuf_read() * -> zio_wait(), as well as through dmu_read() -> * dnode_hold() -> dnode_hold_impl() -> dbuf_read() -> * zio_wait(). When these errors happen, we can assume * that neither an immediate write nor an indirect * write occurred, so we need to fall back to * txg_wait_synced(). This is unusual, so we print to * dmesg whenever one of these errors occurs. */ switch (error) { case 0: break; default: cmn_err(CE_WARN, "zil_lwb_commit() received " "unexpected error %d from ->zl_get_data()" ". Falling back to txg_wait_synced().", error); zfs_fallthrough; case EIO: txg_wait_synced(zilog->zl_dmu_pool, txg); zfs_fallthrough; case ENOENT: zfs_fallthrough; case EEXIST: zfs_fallthrough; case EALREADY: return (lwb); } } } /* * We're actually making an entry, so update lrc_seq to be the * log record sequence number. Note that this is generally not * equal to the itx sequence number because not all transactions * are synchronous, and sometimes spa_sync() gets there first. */ lrcb->lrc_seq = ++zilog->zl_lr_seq; lwb->lwb_nused += reclen + dnow; zil_lwb_add_txg(lwb, txg); ASSERT3U(lwb->lwb_nused, <=, lwb->lwb_sz); ASSERT0(P2PHASE(lwb->lwb_nused, sizeof (uint64_t))); dlen -= dnow; if (dlen > 0) { zilog->zl_cur_used += reclen; goto cont; } return (lwb); } itx_t * zil_itx_create(uint64_t txtype, size_t olrsize) { size_t itxsize, lrsize; itx_t *itx; lrsize = P2ROUNDUP_TYPED(olrsize, sizeof (uint64_t), size_t); itxsize = offsetof(itx_t, itx_lr) + lrsize; itx = zio_data_buf_alloc(itxsize); itx->itx_lr.lrc_txtype = txtype; itx->itx_lr.lrc_reclen = lrsize; itx->itx_lr.lrc_seq = 0; /* defensive */ memset((char *)&itx->itx_lr + olrsize, 0, lrsize - olrsize); itx->itx_sync = B_TRUE; /* default is synchronous */ itx->itx_callback = NULL; itx->itx_callback_data = NULL; itx->itx_size = itxsize; return (itx); } void zil_itx_destroy(itx_t *itx) { IMPLY(itx->itx_lr.lrc_txtype == TX_COMMIT, itx->itx_callback == NULL); IMPLY(itx->itx_callback != NULL, itx->itx_lr.lrc_txtype != TX_COMMIT); if (itx->itx_callback != NULL) itx->itx_callback(itx->itx_callback_data); zio_data_buf_free(itx, itx->itx_size); } /* * Free up the sync and async itxs. The itxs_t has already been detached * so no locks are needed. */ static void zil_itxg_clean(void *arg) { itx_t *itx; list_t *list; avl_tree_t *t; void *cookie; itxs_t *itxs = arg; itx_async_node_t *ian; list = &itxs->i_sync_list; - while ((itx = list_head(list)) != NULL) { + while ((itx = list_remove_head(list)) != NULL) { /* * In the general case, commit itxs will not be found * here, as they'll be committed to an lwb via * zil_lwb_commit(), and free'd in that function. Having * said that, it is still possible for commit itxs to be * found here, due to the following race: * * - a thread calls zil_commit() which assigns the * commit itx to a per-txg i_sync_list * - zil_itxg_clean() is called (e.g. via spa_sync()) * while the waiter is still on the i_sync_list * * There's nothing to prevent syncing the txg while the * waiter is on the i_sync_list. This normally doesn't * happen because spa_sync() is slower than zil_commit(), * but if zil_commit() calls txg_wait_synced() (e.g. * because zil_create() or zil_commit_writer_stall() is * called) we will hit this case. */ if (itx->itx_lr.lrc_txtype == TX_COMMIT) zil_commit_waiter_skip(itx->itx_private); - list_remove(list, itx); zil_itx_destroy(itx); } cookie = NULL; t = &itxs->i_async_tree; while ((ian = avl_destroy_nodes(t, &cookie)) != NULL) { list = &ian->ia_list; - while ((itx = list_head(list)) != NULL) { - list_remove(list, itx); + while ((itx = list_remove_head(list)) != NULL) { /* commit itxs should never be on the async lists. */ ASSERT3U(itx->itx_lr.lrc_txtype, !=, TX_COMMIT); zil_itx_destroy(itx); } list_destroy(list); kmem_free(ian, sizeof (itx_async_node_t)); } avl_destroy(t); kmem_free(itxs, sizeof (itxs_t)); } static int zil_aitx_compare(const void *x1, const void *x2) { const uint64_t o1 = ((itx_async_node_t *)x1)->ia_foid; const uint64_t o2 = ((itx_async_node_t *)x2)->ia_foid; return (TREE_CMP(o1, o2)); } /* * Remove all async itx with the given oid. */ void zil_remove_async(zilog_t *zilog, uint64_t oid) { uint64_t otxg, txg; itx_async_node_t *ian; avl_tree_t *t; avl_index_t where; list_t clean_list; itx_t *itx; ASSERT(oid != 0); list_create(&clean_list, sizeof (itx_t), offsetof(itx_t, itx_node)); if (spa_freeze_txg(zilog->zl_spa) != UINT64_MAX) /* ziltest support */ otxg = ZILTEST_TXG; else otxg = spa_last_synced_txg(zilog->zl_spa) + 1; for (txg = otxg; txg < (otxg + TXG_CONCURRENT_STATES); txg++) { itxg_t *itxg = &zilog->zl_itxg[txg & TXG_MASK]; mutex_enter(&itxg->itxg_lock); if (itxg->itxg_txg != txg) { mutex_exit(&itxg->itxg_lock); continue; } /* * Locate the object node and append its list. */ t = &itxg->itxg_itxs->i_async_tree; ian = avl_find(t, &oid, &where); if (ian != NULL) list_move_tail(&clean_list, &ian->ia_list); mutex_exit(&itxg->itxg_lock); } - while ((itx = list_head(&clean_list)) != NULL) { - list_remove(&clean_list, itx); + while ((itx = list_remove_head(&clean_list)) != NULL) { /* commit itxs should never be on the async lists. */ ASSERT3U(itx->itx_lr.lrc_txtype, !=, TX_COMMIT); zil_itx_destroy(itx); } list_destroy(&clean_list); } void zil_itx_assign(zilog_t *zilog, itx_t *itx, dmu_tx_t *tx) { uint64_t txg; itxg_t *itxg; itxs_t *itxs, *clean = NULL; /* * Ensure the data of a renamed file is committed before the rename. */ if ((itx->itx_lr.lrc_txtype & ~TX_CI) == TX_RENAME) zil_async_to_sync(zilog, itx->itx_oid); if (spa_freeze_txg(zilog->zl_spa) != UINT64_MAX) txg = ZILTEST_TXG; else txg = dmu_tx_get_txg(tx); itxg = &zilog->zl_itxg[txg & TXG_MASK]; mutex_enter(&itxg->itxg_lock); itxs = itxg->itxg_itxs; if (itxg->itxg_txg != txg) { if (itxs != NULL) { /* * The zil_clean callback hasn't got around to cleaning * this itxg. Save the itxs for release below. * This should be rare. */ zfs_dbgmsg("zil_itx_assign: missed itx cleanup for " "txg %llu", (u_longlong_t)itxg->itxg_txg); clean = itxg->itxg_itxs; } itxg->itxg_txg = txg; itxs = itxg->itxg_itxs = kmem_zalloc(sizeof (itxs_t), KM_SLEEP); list_create(&itxs->i_sync_list, sizeof (itx_t), offsetof(itx_t, itx_node)); avl_create(&itxs->i_async_tree, zil_aitx_compare, sizeof (itx_async_node_t), offsetof(itx_async_node_t, ia_node)); } if (itx->itx_sync) { list_insert_tail(&itxs->i_sync_list, itx); } else { avl_tree_t *t = &itxs->i_async_tree; uint64_t foid = LR_FOID_GET_OBJ(((lr_ooo_t *)&itx->itx_lr)->lr_foid); itx_async_node_t *ian; avl_index_t where; ian = avl_find(t, &foid, &where); if (ian == NULL) { ian = kmem_alloc(sizeof (itx_async_node_t), KM_SLEEP); list_create(&ian->ia_list, sizeof (itx_t), offsetof(itx_t, itx_node)); ian->ia_foid = foid; avl_insert(t, ian, where); } list_insert_tail(&ian->ia_list, itx); } itx->itx_lr.lrc_txg = dmu_tx_get_txg(tx); /* * We don't want to dirty the ZIL using ZILTEST_TXG, because * zil_clean() will never be called using ZILTEST_TXG. Thus, we * need to be careful to always dirty the ZIL using the "real" * TXG (not itxg_txg) even when the SPA is frozen. */ zilog_dirty(zilog, dmu_tx_get_txg(tx)); mutex_exit(&itxg->itxg_lock); /* Release the old itxs now we've dropped the lock */ if (clean != NULL) zil_itxg_clean(clean); } /* * If there are any in-memory intent log transactions which have now been * synced then start up a taskq to free them. We should only do this after we * have written out the uberblocks (i.e. txg has been committed) so that * don't inadvertently clean out in-memory log records that would be required * by zil_commit(). */ void zil_clean(zilog_t *zilog, uint64_t synced_txg) { itxg_t *itxg = &zilog->zl_itxg[synced_txg & TXG_MASK]; itxs_t *clean_me; ASSERT3U(synced_txg, <, ZILTEST_TXG); mutex_enter(&itxg->itxg_lock); if (itxg->itxg_itxs == NULL || itxg->itxg_txg == ZILTEST_TXG) { mutex_exit(&itxg->itxg_lock); return; } ASSERT3U(itxg->itxg_txg, <=, synced_txg); ASSERT3U(itxg->itxg_txg, !=, 0); clean_me = itxg->itxg_itxs; itxg->itxg_itxs = NULL; itxg->itxg_txg = 0; mutex_exit(&itxg->itxg_lock); /* * Preferably start a task queue to free up the old itxs but * if taskq_dispatch can't allocate resources to do that then * free it in-line. This should be rare. Note, using TQ_SLEEP * created a bad performance problem. */ ASSERT3P(zilog->zl_dmu_pool, !=, NULL); ASSERT3P(zilog->zl_dmu_pool->dp_zil_clean_taskq, !=, NULL); taskqid_t id = taskq_dispatch(zilog->zl_dmu_pool->dp_zil_clean_taskq, zil_itxg_clean, clean_me, TQ_NOSLEEP); if (id == TASKQID_INVALID) zil_itxg_clean(clean_me); } /* * This function will traverse the queue of itxs that need to be * committed, and move them onto the ZIL's zl_itx_commit_list. */ static void zil_get_commit_list(zilog_t *zilog) { uint64_t otxg, txg; list_t *commit_list = &zilog->zl_itx_commit_list; ASSERT(MUTEX_HELD(&zilog->zl_issuer_lock)); if (spa_freeze_txg(zilog->zl_spa) != UINT64_MAX) /* ziltest support */ otxg = ZILTEST_TXG; else otxg = spa_last_synced_txg(zilog->zl_spa) + 1; /* * This is inherently racy, since there is nothing to prevent * the last synced txg from changing. That's okay since we'll * only commit things in the future. */ for (txg = otxg; txg < (otxg + TXG_CONCURRENT_STATES); txg++) { itxg_t *itxg = &zilog->zl_itxg[txg & TXG_MASK]; mutex_enter(&itxg->itxg_lock); if (itxg->itxg_txg != txg) { mutex_exit(&itxg->itxg_lock); continue; } /* * If we're adding itx records to the zl_itx_commit_list, * then the zil better be dirty in this "txg". We can assert * that here since we're holding the itxg_lock which will * prevent spa_sync from cleaning it. Once we add the itxs * to the zl_itx_commit_list we must commit it to disk even * if it's unnecessary (i.e. the txg was synced). */ ASSERT(zilog_is_dirty_in_txg(zilog, txg) || spa_freeze_txg(zilog->zl_spa) != UINT64_MAX); list_move_tail(commit_list, &itxg->itxg_itxs->i_sync_list); mutex_exit(&itxg->itxg_lock); } } /* * Move the async itxs for a specified object to commit into sync lists. */ void zil_async_to_sync(zilog_t *zilog, uint64_t foid) { uint64_t otxg, txg; itx_async_node_t *ian; avl_tree_t *t; avl_index_t where; if (spa_freeze_txg(zilog->zl_spa) != UINT64_MAX) /* ziltest support */ otxg = ZILTEST_TXG; else otxg = spa_last_synced_txg(zilog->zl_spa) + 1; /* * This is inherently racy, since there is nothing to prevent * the last synced txg from changing. */ for (txg = otxg; txg < (otxg + TXG_CONCURRENT_STATES); txg++) { itxg_t *itxg = &zilog->zl_itxg[txg & TXG_MASK]; mutex_enter(&itxg->itxg_lock); if (itxg->itxg_txg != txg) { mutex_exit(&itxg->itxg_lock); continue; } /* * If a foid is specified then find that node and append its * list. Otherwise walk the tree appending all the lists * to the sync list. We add to the end rather than the * beginning to ensure the create has happened. */ t = &itxg->itxg_itxs->i_async_tree; if (foid != 0) { ian = avl_find(t, &foid, &where); if (ian != NULL) { list_move_tail(&itxg->itxg_itxs->i_sync_list, &ian->ia_list); } } else { void *cookie = NULL; while ((ian = avl_destroy_nodes(t, &cookie)) != NULL) { list_move_tail(&itxg->itxg_itxs->i_sync_list, &ian->ia_list); list_destroy(&ian->ia_list); kmem_free(ian, sizeof (itx_async_node_t)); } } mutex_exit(&itxg->itxg_lock); } } /* * This function will prune commit itxs that are at the head of the * commit list (it won't prune past the first non-commit itx), and * either: a) attach them to the last lwb that's still pending * completion, or b) skip them altogether. * * This is used as a performance optimization to prevent commit itxs * from generating new lwbs when it's unnecessary to do so. */ static void zil_prune_commit_list(zilog_t *zilog) { itx_t *itx; ASSERT(MUTEX_HELD(&zilog->zl_issuer_lock)); while ((itx = list_head(&zilog->zl_itx_commit_list)) != NULL) { lr_t *lrc = &itx->itx_lr; if (lrc->lrc_txtype != TX_COMMIT) break; mutex_enter(&zilog->zl_lock); lwb_t *last_lwb = zilog->zl_last_lwb_opened; if (last_lwb == NULL || last_lwb->lwb_state == LWB_STATE_FLUSH_DONE) { /* * All of the itxs this waiter was waiting on * must have already completed (or there were * never any itx's for it to wait on), so it's * safe to skip this waiter and mark it done. */ zil_commit_waiter_skip(itx->itx_private); } else { zil_commit_waiter_link_lwb(itx->itx_private, last_lwb); itx->itx_private = NULL; } mutex_exit(&zilog->zl_lock); list_remove(&zilog->zl_itx_commit_list, itx); zil_itx_destroy(itx); } IMPLY(itx != NULL, itx->itx_lr.lrc_txtype != TX_COMMIT); } static void zil_commit_writer_stall(zilog_t *zilog) { /* * When zio_alloc_zil() fails to allocate the next lwb block on * disk, we must call txg_wait_synced() to ensure all of the * lwbs in the zilog's zl_lwb_list are synced and then freed (in * zil_sync()), such that any subsequent ZIL writer (i.e. a call * to zil_process_commit_list()) will have to call zil_create(), * and start a new ZIL chain. * * Since zil_alloc_zil() failed, the lwb that was previously * issued does not have a pointer to the "next" lwb on disk. * Thus, if another ZIL writer thread was to allocate the "next" * on-disk lwb, that block could be leaked in the event of a * crash (because the previous lwb on-disk would not point to * it). * * We must hold the zilog's zl_issuer_lock while we do this, to * ensure no new threads enter zil_process_commit_list() until * all lwb's in the zl_lwb_list have been synced and freed * (which is achieved via the txg_wait_synced() call). */ ASSERT(MUTEX_HELD(&zilog->zl_issuer_lock)); txg_wait_synced(zilog->zl_dmu_pool, 0); - ASSERT3P(list_tail(&zilog->zl_lwb_list), ==, NULL); + ASSERT(list_is_empty(&zilog->zl_lwb_list)); } /* * This function will traverse the commit list, creating new lwbs as * needed, and committing the itxs from the commit list to these newly * created lwbs. Additionally, as a new lwb is created, the previous * lwb will be issued to the zio layer to be written to disk. */ static void zil_process_commit_list(zilog_t *zilog) { spa_t *spa = zilog->zl_spa; list_t nolwb_itxs; list_t nolwb_waiters; lwb_t *lwb, *plwb; itx_t *itx; boolean_t first = B_TRUE; ASSERT(MUTEX_HELD(&zilog->zl_issuer_lock)); /* * Return if there's nothing to commit before we dirty the fs by * calling zil_create(). */ - if (list_head(&zilog->zl_itx_commit_list) == NULL) + if (list_is_empty(&zilog->zl_itx_commit_list)) return; list_create(&nolwb_itxs, sizeof (itx_t), offsetof(itx_t, itx_node)); list_create(&nolwb_waiters, sizeof (zil_commit_waiter_t), offsetof(zil_commit_waiter_t, zcw_node)); lwb = list_tail(&zilog->zl_lwb_list); if (lwb == NULL) { lwb = zil_create(zilog); } else { /* * Activate SPA_FEATURE_ZILSAXATTR for the cases where ZIL will * have already been created (zl_lwb_list not empty). */ zil_commit_activate_saxattr_feature(zilog); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_ISSUED); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_WRITE_DONE); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_FLUSH_DONE); first = (lwb->lwb_state != LWB_STATE_OPENED) && ((plwb = list_prev(&zilog->zl_lwb_list, lwb)) == NULL || plwb->lwb_state == LWB_STATE_FLUSH_DONE); } - while ((itx = list_head(&zilog->zl_itx_commit_list)) != NULL) { + while ((itx = list_remove_head(&zilog->zl_itx_commit_list)) != NULL) { lr_t *lrc = &itx->itx_lr; uint64_t txg = lrc->lrc_txg; ASSERT3U(txg, !=, 0); if (lrc->lrc_txtype == TX_COMMIT) { DTRACE_PROBE2(zil__process__commit__itx, zilog_t *, zilog, itx_t *, itx); } else { DTRACE_PROBE2(zil__process__normal__itx, zilog_t *, zilog, itx_t *, itx); } - list_remove(&zilog->zl_itx_commit_list, itx); - boolean_t synced = txg <= spa_last_synced_txg(spa); boolean_t frozen = txg > spa_freeze_txg(spa); /* * If the txg of this itx has already been synced out, then * we don't need to commit this itx to an lwb. This is * because the data of this itx will have already been * written to the main pool. This is inherently racy, and * it's still ok to commit an itx whose txg has already * been synced; this will result in a write that's * unnecessary, but will do no harm. * * With that said, we always want to commit TX_COMMIT itxs * to an lwb, regardless of whether or not that itx's txg * has been synced out. We do this to ensure any OPENED lwb * will always have at least one zil_commit_waiter_t linked * to the lwb. * * As a counter-example, if we skipped TX_COMMIT itx's * whose txg had already been synced, the following * situation could occur if we happened to be racing with * spa_sync: * * 1. We commit a non-TX_COMMIT itx to an lwb, where the * itx's txg is 10 and the last synced txg is 9. * 2. spa_sync finishes syncing out txg 10. * 3. We move to the next itx in the list, it's a TX_COMMIT * whose txg is 10, so we skip it rather than committing * it to the lwb used in (1). * * If the itx that is skipped in (3) is the last TX_COMMIT * itx in the commit list, than it's possible for the lwb * used in (1) to remain in the OPENED state indefinitely. * * To prevent the above scenario from occurring, ensuring * that once an lwb is OPENED it will transition to ISSUED * and eventually DONE, we always commit TX_COMMIT itx's to * an lwb here, even if that itx's txg has already been * synced. * * Finally, if the pool is frozen, we _always_ commit the * itx. The point of freezing the pool is to prevent data * from being written to the main pool via spa_sync, and * instead rely solely on the ZIL to persistently store the * data; i.e. when the pool is frozen, the last synced txg * value can't be trusted. */ if (frozen || !synced || lrc->lrc_txtype == TX_COMMIT) { if (lwb != NULL) { lwb = zil_lwb_commit(zilog, itx, lwb); if (lwb == NULL) list_insert_tail(&nolwb_itxs, itx); else list_insert_tail(&lwb->lwb_itxs, itx); } else { if (lrc->lrc_txtype == TX_COMMIT) { zil_commit_waiter_link_nolwb( itx->itx_private, &nolwb_waiters); } list_insert_tail(&nolwb_itxs, itx); } } else { ASSERT3S(lrc->lrc_txtype, !=, TX_COMMIT); zil_itx_destroy(itx); } } if (lwb == NULL) { /* * This indicates zio_alloc_zil() failed to allocate the * "next" lwb on-disk. When this happens, we must stall * the ZIL write pipeline; see the comment within * zil_commit_writer_stall() for more details. */ zil_commit_writer_stall(zilog); /* * Additionally, we have to signal and mark the "nolwb" * waiters as "done" here, since without an lwb, we * can't do this via zil_lwb_flush_vdevs_done() like * normal. */ zil_commit_waiter_t *zcw; - while ((zcw = list_head(&nolwb_waiters)) != NULL) { + while ((zcw = list_remove_head(&nolwb_waiters)) != NULL) zil_commit_waiter_skip(zcw); - list_remove(&nolwb_waiters, zcw); - } /* * And finally, we have to destroy the itx's that * couldn't be committed to an lwb; this will also call * the itx's callback if one exists for the itx. */ - while ((itx = list_head(&nolwb_itxs)) != NULL) { - list_remove(&nolwb_itxs, itx); + while ((itx = list_remove_head(&nolwb_itxs)) != NULL) zil_itx_destroy(itx); - } } else { ASSERT(list_is_empty(&nolwb_waiters)); ASSERT3P(lwb, !=, NULL); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_ISSUED); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_WRITE_DONE); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_FLUSH_DONE); /* * At this point, the ZIL block pointed at by the "lwb" * variable is in one of the following states: "closed" * or "open". * * If it's "closed", then no itxs have been committed to * it, so there's no point in issuing its zio (i.e. it's * "empty"). * * If it's "open", then it contains one or more itxs that * eventually need to be committed to stable storage. In * this case we intentionally do not issue the lwb's zio * to disk yet, and instead rely on one of the following * two mechanisms for issuing the zio: * * 1. Ideally, there will be more ZIL activity occurring * on the system, such that this function will be * immediately called again (not necessarily by the same * thread) and this lwb's zio will be issued via * zil_lwb_commit(). This way, the lwb is guaranteed to * be "full" when it is issued to disk, and we'll make * use of the lwb's size the best we can. * * 2. If there isn't sufficient ZIL activity occurring on * the system, such that this lwb's zio isn't issued via * zil_lwb_commit(), zil_commit_waiter() will issue the * lwb's zio. If this occurs, the lwb is not guaranteed * to be "full" by the time its zio is issued, and means * the size of the lwb was "too large" given the amount * of ZIL activity occurring on the system at that time. * * We do this for a couple of reasons: * * 1. To try and reduce the number of IOPs needed to * write the same number of itxs. If an lwb has space * available in its buffer for more itxs, and more itxs * will be committed relatively soon (relative to the * latency of performing a write), then it's beneficial * to wait for these "next" itxs. This way, more itxs * can be committed to stable storage with fewer writes. * * 2. To try and use the largest lwb block size that the * incoming rate of itxs can support. Again, this is to * try and pack as many itxs into as few lwbs as * possible, without significantly impacting the latency * of each individual itx. * * If we had no already running or open LWBs, it can be * the workload is single-threaded. And if the ZIL write * latency is very small or if the LWB is almost full, it * may be cheaper to bypass the delay. */ if (lwb->lwb_state == LWB_STATE_OPENED && first) { hrtime_t sleep = zilog->zl_last_lwb_latency * zfs_commit_timeout_pct / 100; if (sleep < zil_min_commit_timeout || lwb->lwb_sz - lwb->lwb_nused < lwb->lwb_sz / 8) { lwb = zil_lwb_write_issue(zilog, lwb); zilog->zl_cur_used = 0; if (lwb == NULL) zil_commit_writer_stall(zilog); } } } } /* * This function is responsible for ensuring the passed in commit waiter * (and associated commit itx) is committed to an lwb. If the waiter is * not already committed to an lwb, all itxs in the zilog's queue of * itxs will be processed. The assumption is the passed in waiter's * commit itx will found in the queue just like the other non-commit * itxs, such that when the entire queue is processed, the waiter will * have been committed to an lwb. * * The lwb associated with the passed in waiter is not guaranteed to * have been issued by the time this function completes. If the lwb is * not issued, we rely on future calls to zil_commit_writer() to issue * the lwb, or the timeout mechanism found in zil_commit_waiter(). */ static void zil_commit_writer(zilog_t *zilog, zil_commit_waiter_t *zcw) { ASSERT(!MUTEX_HELD(&zilog->zl_lock)); ASSERT(spa_writeable(zilog->zl_spa)); mutex_enter(&zilog->zl_issuer_lock); if (zcw->zcw_lwb != NULL || zcw->zcw_done) { /* * It's possible that, while we were waiting to acquire * the "zl_issuer_lock", another thread committed this * waiter to an lwb. If that occurs, we bail out early, * without processing any of the zilog's queue of itxs. * * On certain workloads and system configurations, the * "zl_issuer_lock" can become highly contended. In an * attempt to reduce this contention, we immediately drop * the lock if the waiter has already been processed. * * We've measured this optimization to reduce CPU spent * contending on this lock by up to 5%, using a system * with 32 CPUs, low latency storage (~50 usec writes), * and 1024 threads performing sync writes. */ goto out; } ZIL_STAT_BUMP(zilog, zil_commit_writer_count); zil_get_commit_list(zilog); zil_prune_commit_list(zilog); zil_process_commit_list(zilog); out: mutex_exit(&zilog->zl_issuer_lock); } static void zil_commit_waiter_timeout(zilog_t *zilog, zil_commit_waiter_t *zcw) { ASSERT(!MUTEX_HELD(&zilog->zl_issuer_lock)); ASSERT(MUTEX_HELD(&zcw->zcw_lock)); ASSERT3B(zcw->zcw_done, ==, B_FALSE); lwb_t *lwb = zcw->zcw_lwb; ASSERT3P(lwb, !=, NULL); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_CLOSED); /* * If the lwb has already been issued by another thread, we can * immediately return since there's no work to be done (the * point of this function is to issue the lwb). Additionally, we * do this prior to acquiring the zl_issuer_lock, to avoid * acquiring it when it's not necessary to do so. */ if (lwb->lwb_state == LWB_STATE_ISSUED || lwb->lwb_state == LWB_STATE_WRITE_DONE || lwb->lwb_state == LWB_STATE_FLUSH_DONE) return; /* * In order to call zil_lwb_write_issue() we must hold the * zilog's "zl_issuer_lock". We can't simply acquire that lock, * since we're already holding the commit waiter's "zcw_lock", * and those two locks are acquired in the opposite order * elsewhere. */ mutex_exit(&zcw->zcw_lock); mutex_enter(&zilog->zl_issuer_lock); mutex_enter(&zcw->zcw_lock); /* * Since we just dropped and re-acquired the commit waiter's * lock, we have to re-check to see if the waiter was marked * "done" during that process. If the waiter was marked "done", * the "lwb" pointer is no longer valid (it can be free'd after * the waiter is marked "done"), so without this check we could * wind up with a use-after-free error below. */ if (zcw->zcw_done) goto out; ASSERT3P(lwb, ==, zcw->zcw_lwb); /* * We've already checked this above, but since we hadn't acquired * the zilog's zl_issuer_lock, we have to perform this check a * second time while holding the lock. * * We don't need to hold the zl_lock since the lwb cannot transition * from OPENED to ISSUED while we hold the zl_issuer_lock. The lwb * _can_ transition from ISSUED to DONE, but it's OK to race with * that transition since we treat the lwb the same, whether it's in * the ISSUED or DONE states. * * The important thing, is we treat the lwb differently depending on * if it's ISSUED or OPENED, and block any other threads that might * attempt to issue this lwb. For that reason we hold the * zl_issuer_lock when checking the lwb_state; we must not call * zil_lwb_write_issue() if the lwb had already been issued. * * See the comment above the lwb_state_t structure definition for * more details on the lwb states, and locking requirements. */ if (lwb->lwb_state == LWB_STATE_ISSUED || lwb->lwb_state == LWB_STATE_WRITE_DONE || lwb->lwb_state == LWB_STATE_FLUSH_DONE) goto out; ASSERT3S(lwb->lwb_state, ==, LWB_STATE_OPENED); /* * As described in the comments above zil_commit_waiter() and * zil_process_commit_list(), we need to issue this lwb's zio * since we've reached the commit waiter's timeout and it still * hasn't been issued. */ lwb_t *nlwb = zil_lwb_write_issue(zilog, lwb); - IMPLY(nlwb != NULL, lwb->lwb_state != LWB_STATE_OPENED); + ASSERT3S(lwb->lwb_state, !=, LWB_STATE_OPENED); /* * Since the lwb's zio hadn't been issued by the time this thread * reached its timeout, we reset the zilog's "zl_cur_used" field * to influence the zil block size selection algorithm. * * By having to issue the lwb's zio here, it means the size of the * lwb was too large, given the incoming throughput of itxs. By * setting "zl_cur_used" to zero, we communicate this fact to the * block size selection algorithm, so it can take this information * into account, and potentially select a smaller size for the * next lwb block that is allocated. */ zilog->zl_cur_used = 0; if (nlwb == NULL) { /* * When zil_lwb_write_issue() returns NULL, this * indicates zio_alloc_zil() failed to allocate the * "next" lwb on-disk. When this occurs, the ZIL write * pipeline must be stalled; see the comment within the * zil_commit_writer_stall() function for more details. * * We must drop the commit waiter's lock prior to * calling zil_commit_writer_stall() or else we can wind * up with the following deadlock: * * - This thread is waiting for the txg to sync while * holding the waiter's lock; txg_wait_synced() is * used within txg_commit_writer_stall(). * * - The txg can't sync because it is waiting for this * lwb's zio callback to call dmu_tx_commit(). * * - The lwb's zio callback can't call dmu_tx_commit() * because it's blocked trying to acquire the waiter's * lock, which occurs prior to calling dmu_tx_commit() */ mutex_exit(&zcw->zcw_lock); zil_commit_writer_stall(zilog); mutex_enter(&zcw->zcw_lock); } out: mutex_exit(&zilog->zl_issuer_lock); ASSERT(MUTEX_HELD(&zcw->zcw_lock)); } /* * This function is responsible for performing the following two tasks: * * 1. its primary responsibility is to block until the given "commit * waiter" is considered "done". * * 2. its secondary responsibility is to issue the zio for the lwb that * the given "commit waiter" is waiting on, if this function has * waited "long enough" and the lwb is still in the "open" state. * * Given a sufficient amount of itxs being generated and written using * the ZIL, the lwb's zio will be issued via the zil_lwb_commit() * function. If this does not occur, this secondary responsibility will * ensure the lwb is issued even if there is not other synchronous * activity on the system. * * For more details, see zil_process_commit_list(); more specifically, * the comment at the bottom of that function. */ static void zil_commit_waiter(zilog_t *zilog, zil_commit_waiter_t *zcw) { ASSERT(!MUTEX_HELD(&zilog->zl_lock)); ASSERT(!MUTEX_HELD(&zilog->zl_issuer_lock)); ASSERT(spa_writeable(zilog->zl_spa)); mutex_enter(&zcw->zcw_lock); /* * The timeout is scaled based on the lwb latency to avoid * significantly impacting the latency of each individual itx. * For more details, see the comment at the bottom of the * zil_process_commit_list() function. */ int pct = MAX(zfs_commit_timeout_pct, 1); hrtime_t sleep = (zilog->zl_last_lwb_latency * pct) / 100; hrtime_t wakeup = gethrtime() + sleep; boolean_t timedout = B_FALSE; while (!zcw->zcw_done) { ASSERT(MUTEX_HELD(&zcw->zcw_lock)); lwb_t *lwb = zcw->zcw_lwb; /* * Usually, the waiter will have a non-NULL lwb field here, * but it's possible for it to be NULL as a result of * zil_commit() racing with spa_sync(). * * When zil_clean() is called, it's possible for the itxg * list (which may be cleaned via a taskq) to contain * commit itxs. When this occurs, the commit waiters linked * off of these commit itxs will not be committed to an * lwb. Additionally, these commit waiters will not be * marked done until zil_commit_waiter_skip() is called via * zil_itxg_clean(). * * Thus, it's possible for this commit waiter (i.e. the * "zcw" variable) to be found in this "in between" state; * where it's "zcw_lwb" field is NULL, and it hasn't yet * been skipped, so it's "zcw_done" field is still B_FALSE. */ IMPLY(lwb != NULL, lwb->lwb_state != LWB_STATE_CLOSED); if (lwb != NULL && lwb->lwb_state == LWB_STATE_OPENED) { ASSERT3B(timedout, ==, B_FALSE); /* * If the lwb hasn't been issued yet, then we * need to wait with a timeout, in case this * function needs to issue the lwb after the * timeout is reached; responsibility (2) from * the comment above this function. */ int rc = cv_timedwait_hires(&zcw->zcw_cv, &zcw->zcw_lock, wakeup, USEC2NSEC(1), CALLOUT_FLAG_ABSOLUTE); if (rc != -1 || zcw->zcw_done) continue; timedout = B_TRUE; zil_commit_waiter_timeout(zilog, zcw); if (!zcw->zcw_done) { /* * If the commit waiter has already been * marked "done", it's possible for the * waiter's lwb structure to have already * been freed. Thus, we can only reliably * make these assertions if the waiter * isn't done. */ ASSERT3P(lwb, ==, zcw->zcw_lwb); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_OPENED); } } else { /* * If the lwb isn't open, then it must have already * been issued. In that case, there's no need to * use a timeout when waiting for the lwb to * complete. * * Additionally, if the lwb is NULL, the waiter * will soon be signaled and marked done via * zil_clean() and zil_itxg_clean(), so no timeout * is required. */ IMPLY(lwb != NULL, lwb->lwb_state == LWB_STATE_ISSUED || lwb->lwb_state == LWB_STATE_WRITE_DONE || lwb->lwb_state == LWB_STATE_FLUSH_DONE); cv_wait(&zcw->zcw_cv, &zcw->zcw_lock); } } mutex_exit(&zcw->zcw_lock); } static zil_commit_waiter_t * zil_alloc_commit_waiter(void) { zil_commit_waiter_t *zcw = kmem_cache_alloc(zil_zcw_cache, KM_SLEEP); cv_init(&zcw->zcw_cv, NULL, CV_DEFAULT, NULL); mutex_init(&zcw->zcw_lock, NULL, MUTEX_DEFAULT, NULL); list_link_init(&zcw->zcw_node); zcw->zcw_lwb = NULL; zcw->zcw_done = B_FALSE; zcw->zcw_zio_error = 0; return (zcw); } static void zil_free_commit_waiter(zil_commit_waiter_t *zcw) { ASSERT(!list_link_active(&zcw->zcw_node)); ASSERT3P(zcw->zcw_lwb, ==, NULL); ASSERT3B(zcw->zcw_done, ==, B_TRUE); mutex_destroy(&zcw->zcw_lock); cv_destroy(&zcw->zcw_cv); kmem_cache_free(zil_zcw_cache, zcw); } /* * This function is used to create a TX_COMMIT itx and assign it. This * way, it will be linked into the ZIL's list of synchronous itxs, and * then later committed to an lwb (or skipped) when * zil_process_commit_list() is called. */ static void zil_commit_itx_assign(zilog_t *zilog, zil_commit_waiter_t *zcw) { dmu_tx_t *tx = dmu_tx_create(zilog->zl_os); /* * Since we are not going to create any new dirty data, and we * can even help with clearing the existing dirty data, we * should not be subject to the dirty data based delays. We * use TXG_NOTHROTTLE to bypass the delay mechanism. */ VERIFY0(dmu_tx_assign(tx, TXG_WAIT | TXG_NOTHROTTLE)); itx_t *itx = zil_itx_create(TX_COMMIT, sizeof (lr_t)); itx->itx_sync = B_TRUE; itx->itx_private = zcw; zil_itx_assign(zilog, itx, tx); dmu_tx_commit(tx); } /* * Commit ZFS Intent Log transactions (itxs) to stable storage. * * When writing ZIL transactions to the on-disk representation of the * ZIL, the itxs are committed to a Log Write Block (lwb). Multiple * itxs can be committed to a single lwb. Once a lwb is written and * committed to stable storage (i.e. the lwb is written, and vdevs have * been flushed), each itx that was committed to that lwb is also * considered to be committed to stable storage. * * When an itx is committed to an lwb, the log record (lr_t) contained * by the itx is copied into the lwb's zio buffer, and once this buffer * is written to disk, it becomes an on-disk ZIL block. * * As itxs are generated, they're inserted into the ZIL's queue of * uncommitted itxs. The semantics of zil_commit() are such that it will * block until all itxs that were in the queue when it was called, are * committed to stable storage. * * If "foid" is zero, this means all "synchronous" and "asynchronous" * itxs, for all objects in the dataset, will be committed to stable * storage prior to zil_commit() returning. If "foid" is non-zero, all * "synchronous" itxs for all objects, but only "asynchronous" itxs * that correspond to the foid passed in, will be committed to stable * storage prior to zil_commit() returning. * * Generally speaking, when zil_commit() is called, the consumer doesn't * actually care about _all_ of the uncommitted itxs. Instead, they're * simply trying to waiting for a specific itx to be committed to disk, * but the interface(s) for interacting with the ZIL don't allow such * fine-grained communication. A better interface would allow a consumer * to create and assign an itx, and then pass a reference to this itx to * zil_commit(); such that zil_commit() would return as soon as that * specific itx was committed to disk (instead of waiting for _all_ * itxs to be committed). * * When a thread calls zil_commit() a special "commit itx" will be * generated, along with a corresponding "waiter" for this commit itx. * zil_commit() will wait on this waiter's CV, such that when the waiter * is marked done, and signaled, zil_commit() will return. * * This commit itx is inserted into the queue of uncommitted itxs. This * provides an easy mechanism for determining which itxs were in the * queue prior to zil_commit() having been called, and which itxs were * added after zil_commit() was called. * * The commit itx is special; it doesn't have any on-disk representation. * When a commit itx is "committed" to an lwb, the waiter associated * with it is linked onto the lwb's list of waiters. Then, when that lwb * completes, each waiter on the lwb's list is marked done and signaled * -- allowing the thread waiting on the waiter to return from zil_commit(). * * It's important to point out a few critical factors that allow us * to make use of the commit itxs, commit waiters, per-lwb lists of * commit waiters, and zio completion callbacks like we're doing: * * 1. The list of waiters for each lwb is traversed, and each commit * waiter is marked "done" and signaled, in the zio completion * callback of the lwb's zio[*]. * * * Actually, the waiters are signaled in the zio completion * callback of the root zio for the DKIOCFLUSHWRITECACHE commands * that are sent to the vdevs upon completion of the lwb zio. * * 2. When the itxs are inserted into the ZIL's queue of uncommitted * itxs, the order in which they are inserted is preserved[*]; as * itxs are added to the queue, they are added to the tail of * in-memory linked lists. * * When committing the itxs to lwbs (to be written to disk), they * are committed in the same order in which the itxs were added to * the uncommitted queue's linked list(s); i.e. the linked list of * itxs to commit is traversed from head to tail, and each itx is * committed to an lwb in that order. * * * To clarify: * * - the order of "sync" itxs is preserved w.r.t. other * "sync" itxs, regardless of the corresponding objects. * - the order of "async" itxs is preserved w.r.t. other * "async" itxs corresponding to the same object. * - the order of "async" itxs is *not* preserved w.r.t. other * "async" itxs corresponding to different objects. * - the order of "sync" itxs w.r.t. "async" itxs (or vice * versa) is *not* preserved, even for itxs that correspond * to the same object. * * For more details, see: zil_itx_assign(), zil_async_to_sync(), * zil_get_commit_list(), and zil_process_commit_list(). * * 3. The lwbs represent a linked list of blocks on disk. Thus, any * lwb cannot be considered committed to stable storage, until its * "previous" lwb is also committed to stable storage. This fact, * coupled with the fact described above, means that itxs are * committed in (roughly) the order in which they were generated. * This is essential because itxs are dependent on prior itxs. * Thus, we *must not* deem an itx as being committed to stable * storage, until *all* prior itxs have also been committed to * stable storage. * * To enforce this ordering of lwb zio's, while still leveraging as * much of the underlying storage performance as possible, we rely * on two fundamental concepts: * * 1. The creation and issuance of lwb zio's is protected by * the zilog's "zl_issuer_lock", which ensures only a single * thread is creating and/or issuing lwb's at a time * 2. The "previous" lwb is a child of the "current" lwb * (leveraging the zio parent-child dependency graph) * * By relying on this parent-child zio relationship, we can have * many lwb zio's concurrently issued to the underlying storage, * but the order in which they complete will be the same order in * which they were created. */ void zil_commit(zilog_t *zilog, uint64_t foid) { /* * We should never attempt to call zil_commit on a snapshot for * a couple of reasons: * * 1. A snapshot may never be modified, thus it cannot have any * in-flight itxs that would have modified the dataset. * * 2. By design, when zil_commit() is called, a commit itx will * be assigned to this zilog; as a result, the zilog will be * dirtied. We must not dirty the zilog of a snapshot; there's * checks in the code that enforce this invariant, and will * cause a panic if it's not upheld. */ ASSERT3B(dmu_objset_is_snapshot(zilog->zl_os), ==, B_FALSE); if (zilog->zl_sync == ZFS_SYNC_DISABLED) return; if (!spa_writeable(zilog->zl_spa)) { /* * If the SPA is not writable, there should never be any * pending itxs waiting to be committed to disk. If that * weren't true, we'd skip writing those itxs out, and * would break the semantics of zil_commit(); thus, we're * verifying that truth before we return to the caller. */ ASSERT(list_is_empty(&zilog->zl_lwb_list)); ASSERT3P(zilog->zl_last_lwb_opened, ==, NULL); for (int i = 0; i < TXG_SIZE; i++) ASSERT3P(zilog->zl_itxg[i].itxg_itxs, ==, NULL); return; } /* * If the ZIL is suspended, we don't want to dirty it by calling * zil_commit_itx_assign() below, nor can we write out * lwbs like would be done in zil_commit_write(). Thus, we * simply rely on txg_wait_synced() to maintain the necessary * semantics, and avoid calling those functions altogether. */ if (zilog->zl_suspend > 0) { txg_wait_synced(zilog->zl_dmu_pool, 0); return; } zil_commit_impl(zilog, foid); } void zil_commit_impl(zilog_t *zilog, uint64_t foid) { ZIL_STAT_BUMP(zilog, zil_commit_count); /* * Move the "async" itxs for the specified foid to the "sync" * queues, such that they will be later committed (or skipped) * to an lwb when zil_process_commit_list() is called. * * Since these "async" itxs must be committed prior to this * call to zil_commit returning, we must perform this operation * before we call zil_commit_itx_assign(). */ zil_async_to_sync(zilog, foid); /* * We allocate a new "waiter" structure which will initially be * linked to the commit itx using the itx's "itx_private" field. * Since the commit itx doesn't represent any on-disk state, * when it's committed to an lwb, rather than copying the its * lr_t into the lwb's buffer, the commit itx's "waiter" will be * added to the lwb's list of waiters. Then, when the lwb is * committed to stable storage, each waiter in the lwb's list of * waiters will be marked "done", and signalled. * * We must create the waiter and assign the commit itx prior to * calling zil_commit_writer(), or else our specific commit itx * is not guaranteed to be committed to an lwb prior to calling * zil_commit_waiter(). */ zil_commit_waiter_t *zcw = zil_alloc_commit_waiter(); zil_commit_itx_assign(zilog, zcw); zil_commit_writer(zilog, zcw); zil_commit_waiter(zilog, zcw); if (zcw->zcw_zio_error != 0) { /* * If there was an error writing out the ZIL blocks that * this thread is waiting on, then we fallback to * relying on spa_sync() to write out the data this * thread is waiting on. Obviously this has performance * implications, but the expectation is for this to be * an exceptional case, and shouldn't occur often. */ DTRACE_PROBE2(zil__commit__io__error, zilog_t *, zilog, zil_commit_waiter_t *, zcw); txg_wait_synced(zilog->zl_dmu_pool, 0); } zil_free_commit_waiter(zcw); } /* * Called in syncing context to free committed log blocks and update log header. */ void zil_sync(zilog_t *zilog, dmu_tx_t *tx) { zil_header_t *zh = zil_header_in_syncing_context(zilog); uint64_t txg = dmu_tx_get_txg(tx); spa_t *spa = zilog->zl_spa; uint64_t *replayed_seq = &zilog->zl_replayed_seq[txg & TXG_MASK]; lwb_t *lwb; /* * We don't zero out zl_destroy_txg, so make sure we don't try * to destroy it twice. */ if (spa_sync_pass(spa) != 1) return; zil_lwb_flush_wait_all(zilog, txg); mutex_enter(&zilog->zl_lock); ASSERT(zilog->zl_stop_sync == 0); if (*replayed_seq != 0) { ASSERT(zh->zh_replay_seq < *replayed_seq); zh->zh_replay_seq = *replayed_seq; *replayed_seq = 0; } if (zilog->zl_destroy_txg == txg) { blkptr_t blk = zh->zh_log; dsl_dataset_t *ds = dmu_objset_ds(zilog->zl_os); - ASSERT(list_head(&zilog->zl_lwb_list) == NULL); + ASSERT(list_is_empty(&zilog->zl_lwb_list)); memset(zh, 0, sizeof (zil_header_t)); memset(zilog->zl_replayed_seq, 0, sizeof (zilog->zl_replayed_seq)); if (zilog->zl_keep_first) { /* * If this block was part of log chain that couldn't * be claimed because a device was missing during * zil_claim(), but that device later returns, * then this block could erroneously appear valid. * To guard against this, assign a new GUID to the new * log chain so it doesn't matter what blk points to. */ zil_init_log_chain(zilog, &blk); zh->zh_log = blk; } else { /* * A destroyed ZIL chain can't contain any TX_SETSAXATTR * records. So, deactivate the feature for this dataset. * We activate it again when we start a new ZIL chain. */ if (dsl_dataset_feature_is_active(ds, SPA_FEATURE_ZILSAXATTR)) dsl_dataset_deactivate_feature(ds, SPA_FEATURE_ZILSAXATTR, tx); } } while ((lwb = list_head(&zilog->zl_lwb_list)) != NULL) { zh->zh_log = lwb->lwb_blk; - if (lwb->lwb_buf != NULL || lwb->lwb_max_txg > txg) + if (lwb->lwb_state != LWB_STATE_FLUSH_DONE || + lwb->lwb_max_txg > txg) break; list_remove(&zilog->zl_lwb_list, lwb); zio_free(spa, txg, &lwb->lwb_blk); zil_free_lwb(zilog, lwb); /* * If we don't have anything left in the lwb list then * we've had an allocation failure and we need to zero * out the zil_header blkptr so that we don't end * up freeing the same block twice. */ - if (list_head(&zilog->zl_lwb_list) == NULL) + if (list_is_empty(&zilog->zl_lwb_list)) BP_ZERO(&zh->zh_log); } /* * Remove fastwrite on any blocks that have been pre-allocated for * the next commit. This prevents fastwrite counter pollution by * unused, long-lived LWBs. */ for (; lwb != NULL; lwb = list_next(&zilog->zl_lwb_list, lwb)) { if (lwb->lwb_fastwrite && !lwb->lwb_write_zio) { metaslab_fastwrite_unmark(zilog->zl_spa, &lwb->lwb_blk); lwb->lwb_fastwrite = 0; } } mutex_exit(&zilog->zl_lock); } static int zil_lwb_cons(void *vbuf, void *unused, int kmflag) { (void) unused, (void) kmflag; lwb_t *lwb = vbuf; list_create(&lwb->lwb_itxs, sizeof (itx_t), offsetof(itx_t, itx_node)); list_create(&lwb->lwb_waiters, sizeof (zil_commit_waiter_t), offsetof(zil_commit_waiter_t, zcw_node)); avl_create(&lwb->lwb_vdev_tree, zil_lwb_vdev_compare, sizeof (zil_vdev_node_t), offsetof(zil_vdev_node_t, zv_node)); mutex_init(&lwb->lwb_vdev_lock, NULL, MUTEX_DEFAULT, NULL); return (0); } static void zil_lwb_dest(void *vbuf, void *unused) { (void) unused; lwb_t *lwb = vbuf; mutex_destroy(&lwb->lwb_vdev_lock); avl_destroy(&lwb->lwb_vdev_tree); list_destroy(&lwb->lwb_waiters); list_destroy(&lwb->lwb_itxs); } void zil_init(void) { zil_lwb_cache = kmem_cache_create("zil_lwb_cache", sizeof (lwb_t), 0, zil_lwb_cons, zil_lwb_dest, NULL, NULL, NULL, 0); zil_zcw_cache = kmem_cache_create("zil_zcw_cache", sizeof (zil_commit_waiter_t), 0, NULL, NULL, NULL, NULL, NULL, 0); zil_sums_init(&zil_sums_global); zil_kstats_global = kstat_create("zfs", 0, "zil", "misc", KSTAT_TYPE_NAMED, sizeof (zil_stats) / sizeof (kstat_named_t), KSTAT_FLAG_VIRTUAL); if (zil_kstats_global != NULL) { zil_kstats_global->ks_data = &zil_stats; zil_kstats_global->ks_update = zil_kstats_global_update; zil_kstats_global->ks_private = NULL; kstat_install(zil_kstats_global); } } void zil_fini(void) { kmem_cache_destroy(zil_zcw_cache); kmem_cache_destroy(zil_lwb_cache); if (zil_kstats_global != NULL) { kstat_delete(zil_kstats_global); zil_kstats_global = NULL; } zil_sums_fini(&zil_sums_global); } void zil_set_sync(zilog_t *zilog, uint64_t sync) { zilog->zl_sync = sync; } void zil_set_logbias(zilog_t *zilog, uint64_t logbias) { zilog->zl_logbias = logbias; } zilog_t * zil_alloc(objset_t *os, zil_header_t *zh_phys) { zilog_t *zilog; zilog = kmem_zalloc(sizeof (zilog_t), KM_SLEEP); zilog->zl_header = zh_phys; zilog->zl_os = os; zilog->zl_spa = dmu_objset_spa(os); zilog->zl_dmu_pool = dmu_objset_pool(os); zilog->zl_destroy_txg = TXG_INITIAL - 1; zilog->zl_logbias = dmu_objset_logbias(os); zilog->zl_sync = dmu_objset_syncprop(os); zilog->zl_dirty_max_txg = 0; zilog->zl_last_lwb_opened = NULL; zilog->zl_last_lwb_latency = 0; zilog->zl_max_block_size = zil_maxblocksize; mutex_init(&zilog->zl_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&zilog->zl_issuer_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&zilog->zl_lwb_io_lock, NULL, MUTEX_DEFAULT, NULL); for (int i = 0; i < TXG_SIZE; i++) { mutex_init(&zilog->zl_itxg[i].itxg_lock, NULL, MUTEX_DEFAULT, NULL); } list_create(&zilog->zl_lwb_list, sizeof (lwb_t), offsetof(lwb_t, lwb_node)); list_create(&zilog->zl_itx_commit_list, sizeof (itx_t), offsetof(itx_t, itx_node)); cv_init(&zilog->zl_cv_suspend, NULL, CV_DEFAULT, NULL); cv_init(&zilog->zl_lwb_io_cv, NULL, CV_DEFAULT, NULL); return (zilog); } void zil_free(zilog_t *zilog) { int i; zilog->zl_stop_sync = 1; ASSERT0(zilog->zl_suspend); ASSERT0(zilog->zl_suspending); ASSERT(list_is_empty(&zilog->zl_lwb_list)); list_destroy(&zilog->zl_lwb_list); ASSERT(list_is_empty(&zilog->zl_itx_commit_list)); list_destroy(&zilog->zl_itx_commit_list); for (i = 0; i < TXG_SIZE; i++) { /* * It's possible for an itx to be generated that doesn't dirty * a txg (e.g. ztest TX_TRUNCATE). So there's no zil_clean() * callback to remove the entry. We remove those here. * * Also free up the ziltest itxs. */ if (zilog->zl_itxg[i].itxg_itxs) zil_itxg_clean(zilog->zl_itxg[i].itxg_itxs); mutex_destroy(&zilog->zl_itxg[i].itxg_lock); } mutex_destroy(&zilog->zl_issuer_lock); mutex_destroy(&zilog->zl_lock); mutex_destroy(&zilog->zl_lwb_io_lock); cv_destroy(&zilog->zl_cv_suspend); cv_destroy(&zilog->zl_lwb_io_cv); kmem_free(zilog, sizeof (zilog_t)); } /* * Open an intent log. */ zilog_t * zil_open(objset_t *os, zil_get_data_t *get_data, zil_sums_t *zil_sums) { zilog_t *zilog = dmu_objset_zil(os); ASSERT3P(zilog->zl_get_data, ==, NULL); ASSERT3P(zilog->zl_last_lwb_opened, ==, NULL); ASSERT(list_is_empty(&zilog->zl_lwb_list)); zilog->zl_get_data = get_data; zilog->zl_sums = zil_sums; return (zilog); } /* * Close an intent log. */ void zil_close(zilog_t *zilog) { lwb_t *lwb; uint64_t txg; if (!dmu_objset_is_snapshot(zilog->zl_os)) { zil_commit(zilog, 0); } else { - ASSERT3P(list_tail(&zilog->zl_lwb_list), ==, NULL); + ASSERT(list_is_empty(&zilog->zl_lwb_list)); ASSERT0(zilog->zl_dirty_max_txg); ASSERT3B(zilog_is_dirty(zilog), ==, B_FALSE); } mutex_enter(&zilog->zl_lock); lwb = list_tail(&zilog->zl_lwb_list); if (lwb == NULL) txg = zilog->zl_dirty_max_txg; else txg = MAX(zilog->zl_dirty_max_txg, lwb->lwb_max_txg); mutex_exit(&zilog->zl_lock); /* * zl_lwb_max_issued_txg may be larger than lwb_max_txg. It depends * on the time when the dmu_tx transaction is assigned in * zil_lwb_write_issue(). */ mutex_enter(&zilog->zl_lwb_io_lock); txg = MAX(zilog->zl_lwb_max_issued_txg, txg); mutex_exit(&zilog->zl_lwb_io_lock); /* * We need to use txg_wait_synced() to wait until that txg is synced. * zil_sync() will guarantee all lwbs up to that txg have been * written out, flushed, and cleaned. */ if (txg != 0) txg_wait_synced(zilog->zl_dmu_pool, txg); if (zilog_is_dirty(zilog)) zfs_dbgmsg("zil (%px) is dirty, txg %llu", zilog, (u_longlong_t)txg); if (txg < spa_freeze_txg(zilog->zl_spa)) VERIFY(!zilog_is_dirty(zilog)); zilog->zl_get_data = NULL; /* * We should have only one lwb left on the list; remove it now. */ mutex_enter(&zilog->zl_lock); - lwb = list_head(&zilog->zl_lwb_list); + lwb = list_remove_head(&zilog->zl_lwb_list); if (lwb != NULL) { - ASSERT3P(lwb, ==, list_tail(&zilog->zl_lwb_list)); + ASSERT(list_is_empty(&zilog->zl_lwb_list)); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_ISSUED); if (lwb->lwb_fastwrite) metaslab_fastwrite_unmark(zilog->zl_spa, &lwb->lwb_blk); - list_remove(&zilog->zl_lwb_list, lwb); zio_buf_free(lwb->lwb_buf, lwb->lwb_sz); zil_free_lwb(zilog, lwb); } mutex_exit(&zilog->zl_lock); } static const char *suspend_tag = "zil suspending"; /* * Suspend an intent log. While in suspended mode, we still honor * synchronous semantics, but we rely on txg_wait_synced() to do it. * On old version pools, we suspend the log briefly when taking a * snapshot so that it will have an empty intent log. * * Long holds are not really intended to be used the way we do here -- * held for such a short time. A concurrent caller of dsl_dataset_long_held() * could fail. Therefore we take pains to only put a long hold if it is * actually necessary. Fortunately, it will only be necessary if the * objset is currently mounted (or the ZVOL equivalent). In that case it * will already have a long hold, so we are not really making things any worse. * * Ideally, we would locate the existing long-holder (i.e. the zfsvfs_t or * zvol_state_t), and use their mechanism to prevent their hold from being * dropped (e.g. VFS_HOLD()). However, that would be even more pain for * very little gain. * * if cookiep == NULL, this does both the suspend & resume. * Otherwise, it returns with the dataset "long held", and the cookie * should be passed into zil_resume(). */ int zil_suspend(const char *osname, void **cookiep) { objset_t *os; zilog_t *zilog; const zil_header_t *zh; int error; error = dmu_objset_hold(osname, suspend_tag, &os); if (error != 0) return (error); zilog = dmu_objset_zil(os); mutex_enter(&zilog->zl_lock); zh = zilog->zl_header; if (zh->zh_flags & ZIL_REPLAY_NEEDED) { /* unplayed log */ mutex_exit(&zilog->zl_lock); dmu_objset_rele(os, suspend_tag); return (SET_ERROR(EBUSY)); } /* * Don't put a long hold in the cases where we can avoid it. This * is when there is no cookie so we are doing a suspend & resume * (i.e. called from zil_vdev_offline()), and there's nothing to do * for the suspend because it's already suspended, or there's no ZIL. */ if (cookiep == NULL && !zilog->zl_suspending && (zilog->zl_suspend > 0 || BP_IS_HOLE(&zh->zh_log))) { mutex_exit(&zilog->zl_lock); dmu_objset_rele(os, suspend_tag); return (0); } dsl_dataset_long_hold(dmu_objset_ds(os), suspend_tag); dsl_pool_rele(dmu_objset_pool(os), suspend_tag); zilog->zl_suspend++; if (zilog->zl_suspend > 1) { /* * Someone else is already suspending it. * Just wait for them to finish. */ while (zilog->zl_suspending) cv_wait(&zilog->zl_cv_suspend, &zilog->zl_lock); mutex_exit(&zilog->zl_lock); if (cookiep == NULL) zil_resume(os); else *cookiep = os; return (0); } /* * If there is no pointer to an on-disk block, this ZIL must not * be active (e.g. filesystem not mounted), so there's nothing * to clean up. */ if (BP_IS_HOLE(&zh->zh_log)) { ASSERT(cookiep != NULL); /* fast path already handled */ *cookiep = os; mutex_exit(&zilog->zl_lock); return (0); } /* * The ZIL has work to do. Ensure that the associated encryption * key will remain mapped while we are committing the log by * grabbing a reference to it. If the key isn't loaded we have no * choice but to return an error until the wrapping key is loaded. */ if (os->os_encrypted && dsl_dataset_create_key_mapping(dmu_objset_ds(os)) != 0) { zilog->zl_suspend--; mutex_exit(&zilog->zl_lock); dsl_dataset_long_rele(dmu_objset_ds(os), suspend_tag); dsl_dataset_rele(dmu_objset_ds(os), suspend_tag); return (SET_ERROR(EACCES)); } zilog->zl_suspending = B_TRUE; mutex_exit(&zilog->zl_lock); /* * We need to use zil_commit_impl to ensure we wait for all * LWB_STATE_OPENED and LWB_STATE_ISSUED lwbs to be committed * to disk before proceeding. If we used zil_commit instead, it * would just call txg_wait_synced(), because zl_suspend is set. * txg_wait_synced() doesn't wait for these lwb's to be * LWB_STATE_FLUSH_DONE before returning. */ zil_commit_impl(zilog, 0); /* * Now that we've ensured all lwb's are LWB_STATE_FLUSH_DONE, we * use txg_wait_synced() to ensure the data from the zilog has * migrated to the main pool before calling zil_destroy(). */ txg_wait_synced(zilog->zl_dmu_pool, 0); zil_destroy(zilog, B_FALSE); mutex_enter(&zilog->zl_lock); zilog->zl_suspending = B_FALSE; cv_broadcast(&zilog->zl_cv_suspend); mutex_exit(&zilog->zl_lock); if (os->os_encrypted) dsl_dataset_remove_key_mapping(dmu_objset_ds(os)); if (cookiep == NULL) zil_resume(os); else *cookiep = os; return (0); } void zil_resume(void *cookie) { objset_t *os = cookie; zilog_t *zilog = dmu_objset_zil(os); mutex_enter(&zilog->zl_lock); ASSERT(zilog->zl_suspend != 0); zilog->zl_suspend--; mutex_exit(&zilog->zl_lock); dsl_dataset_long_rele(dmu_objset_ds(os), suspend_tag); dsl_dataset_rele(dmu_objset_ds(os), suspend_tag); } typedef struct zil_replay_arg { zil_replay_func_t *const *zr_replay; void *zr_arg; boolean_t zr_byteswap; char *zr_lr; } zil_replay_arg_t; static int zil_replay_error(zilog_t *zilog, const lr_t *lr, int error) { char name[ZFS_MAX_DATASET_NAME_LEN]; zilog->zl_replaying_seq--; /* didn't actually replay this one */ dmu_objset_name(zilog->zl_os, name); cmn_err(CE_WARN, "ZFS replay transaction error %d, " "dataset %s, seq 0x%llx, txtype %llu %s\n", error, name, (u_longlong_t)lr->lrc_seq, (u_longlong_t)(lr->lrc_txtype & ~TX_CI), (lr->lrc_txtype & TX_CI) ? "CI" : ""); return (error); } static int zil_replay_log_record(zilog_t *zilog, const lr_t *lr, void *zra, uint64_t claim_txg) { zil_replay_arg_t *zr = zra; const zil_header_t *zh = zilog->zl_header; uint64_t reclen = lr->lrc_reclen; uint64_t txtype = lr->lrc_txtype; int error = 0; zilog->zl_replaying_seq = lr->lrc_seq; if (lr->lrc_seq <= zh->zh_replay_seq) /* already replayed */ return (0); if (lr->lrc_txg < claim_txg) /* already committed */ return (0); /* Strip case-insensitive bit, still present in log record */ txtype &= ~TX_CI; if (txtype == 0 || txtype >= TX_MAX_TYPE) return (zil_replay_error(zilog, lr, EINVAL)); /* * If this record type can be logged out of order, the object * (lr_foid) may no longer exist. That's legitimate, not an error. */ if (TX_OOO(txtype)) { error = dmu_object_info(zilog->zl_os, LR_FOID_GET_OBJ(((lr_ooo_t *)lr)->lr_foid), NULL); if (error == ENOENT || error == EEXIST) return (0); } /* * Make a copy of the data so we can revise and extend it. */ memcpy(zr->zr_lr, lr, reclen); /* * If this is a TX_WRITE with a blkptr, suck in the data. */ if (txtype == TX_WRITE && reclen == sizeof (lr_write_t)) { error = zil_read_log_data(zilog, (lr_write_t *)lr, zr->zr_lr + reclen); if (error != 0) return (zil_replay_error(zilog, lr, error)); } /* * The log block containing this lr may have been byteswapped * so that we can easily examine common fields like lrc_txtype. * However, the log is a mix of different record types, and only the * replay vectors know how to byteswap their records. Therefore, if * the lr was byteswapped, undo it before invoking the replay vector. */ if (zr->zr_byteswap) byteswap_uint64_array(zr->zr_lr, reclen); /* * We must now do two things atomically: replay this log record, * and update the log header sequence number to reflect the fact that * we did so. At the end of each replay function the sequence number * is updated if we are in replay mode. */ error = zr->zr_replay[txtype](zr->zr_arg, zr->zr_lr, zr->zr_byteswap); if (error != 0) { /* * The DMU's dnode layer doesn't see removes until the txg * commits, so a subsequent claim can spuriously fail with * EEXIST. So if we receive any error we try syncing out * any removes then retry the transaction. Note that we * specify B_FALSE for byteswap now, so we don't do it twice. */ txg_wait_synced(spa_get_dsl(zilog->zl_spa), 0); error = zr->zr_replay[txtype](zr->zr_arg, zr->zr_lr, B_FALSE); if (error != 0) return (zil_replay_error(zilog, lr, error)); } return (0); } static int zil_incr_blks(zilog_t *zilog, const blkptr_t *bp, void *arg, uint64_t claim_txg) { (void) bp, (void) arg, (void) claim_txg; zilog->zl_replay_blks++; return (0); } /* * If this dataset has a non-empty intent log, replay it and destroy it. * Return B_TRUE if there were any entries to replay. */ boolean_t zil_replay(objset_t *os, void *arg, zil_replay_func_t *const replay_func[TX_MAX_TYPE]) { zilog_t *zilog = dmu_objset_zil(os); const zil_header_t *zh = zilog->zl_header; zil_replay_arg_t zr; if ((zh->zh_flags & ZIL_REPLAY_NEEDED) == 0) { return (zil_destroy(zilog, B_TRUE)); } zr.zr_replay = replay_func; zr.zr_arg = arg; zr.zr_byteswap = BP_SHOULD_BYTESWAP(&zh->zh_log); zr.zr_lr = vmem_alloc(2 * SPA_MAXBLOCKSIZE, KM_SLEEP); /* * Wait for in-progress removes to sync before starting replay. */ txg_wait_synced(zilog->zl_dmu_pool, 0); zilog->zl_replay = B_TRUE; zilog->zl_replay_time = ddi_get_lbolt(); ASSERT(zilog->zl_replay_blks == 0); (void) zil_parse(zilog, zil_incr_blks, zil_replay_log_record, &zr, zh->zh_claim_txg, B_TRUE); vmem_free(zr.zr_lr, 2 * SPA_MAXBLOCKSIZE); zil_destroy(zilog, B_FALSE); txg_wait_synced(zilog->zl_dmu_pool, zilog->zl_destroy_txg); zilog->zl_replay = B_FALSE; return (B_TRUE); } boolean_t zil_replaying(zilog_t *zilog, dmu_tx_t *tx) { if (zilog->zl_sync == ZFS_SYNC_DISABLED) return (B_TRUE); if (zilog->zl_replay) { dsl_dataset_dirty(dmu_objset_ds(zilog->zl_os), tx); zilog->zl_replayed_seq[dmu_tx_get_txg(tx) & TXG_MASK] = zilog->zl_replaying_seq; return (B_TRUE); } return (B_FALSE); } int zil_reset(const char *osname, void *arg) { (void) arg; int error = zil_suspend(osname, NULL); /* EACCES means crypto key not loaded */ if ((error == EACCES) || (error == EBUSY)) return (SET_ERROR(error)); if (error != 0) return (SET_ERROR(EEXIST)); return (0); } EXPORT_SYMBOL(zil_alloc); EXPORT_SYMBOL(zil_free); EXPORT_SYMBOL(zil_open); EXPORT_SYMBOL(zil_close); EXPORT_SYMBOL(zil_replay); EXPORT_SYMBOL(zil_replaying); EXPORT_SYMBOL(zil_destroy); EXPORT_SYMBOL(zil_destroy_sync); EXPORT_SYMBOL(zil_itx_create); EXPORT_SYMBOL(zil_itx_destroy); EXPORT_SYMBOL(zil_itx_assign); EXPORT_SYMBOL(zil_commit); EXPORT_SYMBOL(zil_claim); EXPORT_SYMBOL(zil_check_log_chain); EXPORT_SYMBOL(zil_sync); EXPORT_SYMBOL(zil_clean); EXPORT_SYMBOL(zil_suspend); EXPORT_SYMBOL(zil_resume); EXPORT_SYMBOL(zil_lwb_add_block); EXPORT_SYMBOL(zil_bp_tree_add); EXPORT_SYMBOL(zil_set_sync); EXPORT_SYMBOL(zil_set_logbias); EXPORT_SYMBOL(zil_sums_init); EXPORT_SYMBOL(zil_sums_fini); EXPORT_SYMBOL(zil_kstat_values_update); ZFS_MODULE_PARAM(zfs, zfs_, commit_timeout_pct, UINT, ZMOD_RW, "ZIL block open timeout percentage"); ZFS_MODULE_PARAM(zfs_zil, zil_, min_commit_timeout, U64, ZMOD_RW, "Minimum delay we care for ZIL block commit"); ZFS_MODULE_PARAM(zfs_zil, zil_, replay_disable, INT, ZMOD_RW, "Disable intent logging replay"); ZFS_MODULE_PARAM(zfs_zil, zil_, nocacheflush, INT, ZMOD_RW, "Disable ZIL cache flushes"); ZFS_MODULE_PARAM(zfs_zil, zil_, slog_bulk, U64, ZMOD_RW, "Limit in bytes slog sync writes per commit"); ZFS_MODULE_PARAM(zfs_zil, zil_, maxblocksize, UINT, ZMOD_RW, "Limit in bytes of ZIL log block size"); diff --git a/sys/contrib/openzfs/tests/runfiles/common.run b/sys/contrib/openzfs/tests/runfiles/common.run index e2137ac596d9..9ed1a6d37a97 100644 --- a/sys/contrib/openzfs/tests/runfiles/common.run +++ b/sys/contrib/openzfs/tests/runfiles/common.run @@ -1,979 +1,982 @@ # # This file and its contents are supplied under the terms of the # Common Development and Distribution License ("CDDL"), version 1.0. # You may only use this file in accordance with the terms of version # 1.0 of the CDDL. # # A full copy of the text of the CDDL should have accompanied this # source. A copy of the CDDL is also available via the Internet at # http://www.illumos.org/license/CDDL. # # This run file contains all of the common functional tests. When # adding a new test consider also adding it to the sanity.run file # if the new test runs to completion in only a few seconds. # # Approximate run time: 4-5 hours # [DEFAULT] pre = setup quiet = False pre_user = root user = root timeout = 600 post_user = root post = cleanup failsafe_user = root failsafe = callbacks/zfs_failsafe outputdir = /var/tmp/test_results tags = ['functional'] [tests/functional/acl/off] tests = ['dosmode', 'posixmode'] tags = ['functional', 'acl'] [tests/functional/alloc_class] tests = ['alloc_class_001_pos', 'alloc_class_002_neg', 'alloc_class_003_pos', 'alloc_class_004_pos', 'alloc_class_005_pos', 'alloc_class_006_pos', 'alloc_class_007_pos', 'alloc_class_008_pos', 'alloc_class_009_pos', 'alloc_class_010_pos', 'alloc_class_011_neg', 'alloc_class_012_pos', - 'alloc_class_013_pos'] + 'alloc_class_013_pos', 'alloc_class_014_neg', 'alloc_class_015_pos'] tags = ['functional', 'alloc_class'] [tests/functional/append] tests = ['file_append', 'threadsappend_001_pos'] tags = ['functional', 'append'] [tests/functional/arc] tests = ['dbufstats_001_pos', 'dbufstats_002_pos', 'dbufstats_003_pos', 'arcstats_runtime_tuning'] tags = ['functional', 'arc'] [tests/functional/atime] tests = ['atime_001_pos', 'atime_002_neg', 'root_atime_off', 'root_atime_on'] tags = ['functional', 'atime'] [tests/functional/bootfs] tests = ['bootfs_001_pos', 'bootfs_002_neg', 'bootfs_003_pos', 'bootfs_004_neg', 'bootfs_005_neg', 'bootfs_006_pos', 'bootfs_007_pos', 'bootfs_008_pos'] tags = ['functional', 'bootfs'] [tests/functional/btree] tests = ['btree_positive', 'btree_negative'] tags = ['functional', 'btree'] pre = post = [tests/functional/cache] tests = ['cache_001_pos', 'cache_002_pos', 'cache_003_pos', 'cache_004_neg', 'cache_005_neg', 'cache_006_pos', 'cache_007_neg', 'cache_008_neg', 'cache_009_pos', 'cache_010_pos', 'cache_011_pos', 'cache_012_pos'] tags = ['functional', 'cache'] [tests/functional/cachefile] tests = ['cachefile_001_pos', 'cachefile_002_pos', 'cachefile_003_pos', 'cachefile_004_pos'] tags = ['functional', 'cachefile'] [tests/functional/casenorm] tests = ['case_all_values', 'norm_all_values', 'mixed_create_failure', 'sensitive_none_lookup', 'sensitive_none_delete', 'sensitive_formd_lookup', 'sensitive_formd_delete', 'insensitive_none_lookup', 'insensitive_none_delete', 'insensitive_formd_lookup', 'insensitive_formd_delete', 'mixed_none_lookup', 'mixed_none_lookup_ci', 'mixed_none_delete', 'mixed_formd_lookup', 'mixed_formd_lookup_ci', 'mixed_formd_delete'] tags = ['functional', 'casenorm'] [tests/functional/channel_program/lua_core] tests = ['tst.args_to_lua', 'tst.divide_by_zero', 'tst.exists', 'tst.integer_illegal', 'tst.integer_overflow', 'tst.language_functions_neg', 'tst.language_functions_pos', 'tst.large_prog', 'tst.libraries', 'tst.memory_limit', 'tst.nested_neg', 'tst.nested_pos', 'tst.nvlist_to_lua', 'tst.recursive_neg', 'tst.recursive_pos', 'tst.return_large', 'tst.return_nvlist_neg', 'tst.return_nvlist_pos', 'tst.return_recursive_table', 'tst.stack_gsub', 'tst.timeout'] tags = ['functional', 'channel_program', 'lua_core'] [tests/functional/channel_program/synctask_core] tests = ['tst.destroy_fs', 'tst.destroy_snap', 'tst.get_count_and_limit', 'tst.get_index_props', 'tst.get_mountpoint', 'tst.get_neg', 'tst.get_number_props', 'tst.get_string_props', 'tst.get_type', 'tst.get_userquota', 'tst.get_written', 'tst.inherit', 'tst.list_bookmarks', 'tst.list_children', 'tst.list_clones', 'tst.list_holds', 'tst.list_snapshots', 'tst.list_system_props', 'tst.list_user_props', 'tst.parse_args_neg','tst.promote_conflict', 'tst.promote_multiple', 'tst.promote_simple', 'tst.rollback_mult', 'tst.rollback_one', 'tst.set_props', 'tst.snapshot_destroy', 'tst.snapshot_neg', 'tst.snapshot_recursive', 'tst.snapshot_rename', 'tst.snapshot_simple', 'tst.bookmark.create', 'tst.bookmark.copy', 'tst.terminate_by_signal' ] tags = ['functional', 'channel_program', 'synctask_core'] [tests/functional/checksum] tests = ['run_edonr_test', 'run_sha2_test', 'run_skein_test', 'run_blake3_test', 'filetest_001_pos', 'filetest_002_pos'] tags = ['functional', 'checksum'] [tests/functional/clean_mirror] tests = [ 'clean_mirror_001_pos', 'clean_mirror_002_pos', 'clean_mirror_003_pos', 'clean_mirror_004_pos'] tags = ['functional', 'clean_mirror'] [tests/functional/cli_root/zdb] tests = ['zdb_002_pos', 'zdb_003_pos', 'zdb_004_pos', 'zdb_005_pos', 'zdb_006_pos', 'zdb_args_neg', 'zdb_args_pos', 'zdb_block_size_histogram', 'zdb_checksum', 'zdb_decompress', 'zdb_display_block', 'zdb_encrypted', 'zdb_label_checksum', 'zdb_object_range_neg', 'zdb_object_range_pos', 'zdb_objset_id', 'zdb_decompress_zstd', 'zdb_recover', 'zdb_recover_2'] pre = post = tags = ['functional', 'cli_root', 'zdb'] [tests/functional/cli_root/zfs] tests = ['zfs_001_neg', 'zfs_002_pos'] tags = ['functional', 'cli_root', 'zfs'] [tests/functional/cli_root/zfs_bookmark] tests = ['zfs_bookmark_cliargs'] tags = ['functional', 'cli_root', 'zfs_bookmark'] [tests/functional/cli_root/zfs_change-key] tests = ['zfs_change-key', 'zfs_change-key_child', 'zfs_change-key_format', 'zfs_change-key_inherit', 'zfs_change-key_load', 'zfs_change-key_location', 'zfs_change-key_pbkdf2iters', 'zfs_change-key_clones'] tags = ['functional', 'cli_root', 'zfs_change-key'] [tests/functional/cli_root/zfs_clone] tests = ['zfs_clone_001_neg', 'zfs_clone_002_pos', 'zfs_clone_003_pos', 'zfs_clone_004_pos', 'zfs_clone_005_pos', 'zfs_clone_006_pos', 'zfs_clone_007_pos', 'zfs_clone_008_neg', 'zfs_clone_009_neg', 'zfs_clone_010_pos', 'zfs_clone_encrypted', 'zfs_clone_deeply_nested', 'zfs_clone_rm_nested'] tags = ['functional', 'cli_root', 'zfs_clone'] [tests/functional/cli_root/zfs_copies] tests = ['zfs_copies_001_pos', 'zfs_copies_002_pos', 'zfs_copies_003_pos', 'zfs_copies_004_neg', 'zfs_copies_005_neg', 'zfs_copies_006_pos'] tags = ['functional', 'cli_root', 'zfs_copies'] [tests/functional/cli_root/zfs_create] tests = ['zfs_create_001_pos', 'zfs_create_002_pos', 'zfs_create_003_pos', 'zfs_create_004_pos', 'zfs_create_005_pos', 'zfs_create_006_pos', 'zfs_create_007_pos', 'zfs_create_008_neg', 'zfs_create_009_neg', 'zfs_create_010_neg', 'zfs_create_011_pos', 'zfs_create_012_pos', 'zfs_create_013_pos', 'zfs_create_014_pos', 'zfs_create_encrypted', 'zfs_create_crypt_combos', 'zfs_create_dryrun', 'zfs_create_nomount', 'zfs_create_verbose'] tags = ['functional', 'cli_root', 'zfs_create'] [tests/functional/cli_root/zfs_destroy] tests = ['zfs_clone_livelist_condense_and_disable', 'zfs_clone_livelist_condense_races', 'zfs_clone_livelist_dedup', 'zfs_destroy_001_pos', 'zfs_destroy_002_pos', 'zfs_destroy_003_pos', 'zfs_destroy_004_pos', 'zfs_destroy_005_neg', 'zfs_destroy_006_neg', 'zfs_destroy_007_neg', 'zfs_destroy_008_pos', 'zfs_destroy_009_pos', 'zfs_destroy_010_pos', 'zfs_destroy_011_pos', 'zfs_destroy_012_pos', 'zfs_destroy_013_neg', 'zfs_destroy_014_pos', 'zfs_destroy_015_pos', 'zfs_destroy_016_pos', 'zfs_destroy_clone_livelist', 'zfs_destroy_dev_removal', 'zfs_destroy_dev_removal_condense'] tags = ['functional', 'cli_root', 'zfs_destroy'] [tests/functional/cli_root/zfs_diff] tests = ['zfs_diff_changes', 'zfs_diff_cliargs', 'zfs_diff_timestamp', 'zfs_diff_types', 'zfs_diff_encrypted', 'zfs_diff_mangle'] tags = ['functional', 'cli_root', 'zfs_diff'] [tests/functional/cli_root/zfs_get] tests = ['zfs_get_001_pos', 'zfs_get_002_pos', 'zfs_get_003_pos', 'zfs_get_004_pos', 'zfs_get_005_neg', 'zfs_get_006_neg', 'zfs_get_007_neg', 'zfs_get_008_pos', 'zfs_get_009_pos', 'zfs_get_010_neg'] tags = ['functional', 'cli_root', 'zfs_get'] [tests/functional/cli_root/zfs_ids_to_path] tests = ['zfs_ids_to_path_001_pos'] tags = ['functional', 'cli_root', 'zfs_ids_to_path'] [tests/functional/cli_root/zfs_inherit] tests = ['zfs_inherit_001_neg', 'zfs_inherit_002_neg', 'zfs_inherit_003_pos', 'zfs_inherit_mountpoint'] tags = ['functional', 'cli_root', 'zfs_inherit'] [tests/functional/cli_root/zfs_load-key] tests = ['zfs_load-key', 'zfs_load-key_all', 'zfs_load-key_file', 'zfs_load-key_https', 'zfs_load-key_location', 'zfs_load-key_noop', 'zfs_load-key_recursive'] tags = ['functional', 'cli_root', 'zfs_load-key'] [tests/functional/cli_root/zfs_mount] tests = ['zfs_mount_001_pos', 'zfs_mount_002_pos', 'zfs_mount_003_pos', 'zfs_mount_004_pos', 'zfs_mount_005_pos', 'zfs_mount_007_pos', 'zfs_mount_009_neg', 'zfs_mount_010_neg', 'zfs_mount_011_neg', 'zfs_mount_012_pos', 'zfs_mount_all_001_pos', 'zfs_mount_encrypted', 'zfs_mount_remount', 'zfs_mount_all_fail', 'zfs_mount_all_mountpoints', 'zfs_mount_test_race'] tags = ['functional', 'cli_root', 'zfs_mount'] [tests/functional/cli_root/zfs_program] tests = ['zfs_program_json'] tags = ['functional', 'cli_root', 'zfs_program'] [tests/functional/cli_root/zfs_promote] tests = ['zfs_promote_001_pos', 'zfs_promote_002_pos', 'zfs_promote_003_pos', 'zfs_promote_004_pos', 'zfs_promote_005_pos', 'zfs_promote_006_neg', 'zfs_promote_007_neg', 'zfs_promote_008_pos', 'zfs_promote_encryptionroot'] tags = ['functional', 'cli_root', 'zfs_promote'] [tests/functional/cli_root/zfs_property] tests = ['zfs_written_property_001_pos'] tags = ['functional', 'cli_root', 'zfs_property'] [tests/functional/cli_root/zfs_receive] tests = ['zfs_receive_001_pos', 'zfs_receive_002_pos', 'zfs_receive_003_pos', 'zfs_receive_004_neg', 'zfs_receive_005_neg', 'zfs_receive_006_pos', 'zfs_receive_007_neg', 'zfs_receive_008_pos', 'zfs_receive_009_neg', 'zfs_receive_010_pos', 'zfs_receive_011_pos', 'zfs_receive_012_pos', 'zfs_receive_013_pos', 'zfs_receive_014_pos', 'zfs_receive_015_pos', 'zfs_receive_016_pos', 'receive-o-x_props_override', 'receive-o-x_props_aliases', 'zfs_receive_from_encrypted', 'zfs_receive_to_encrypted', 'zfs_receive_raw', 'zfs_receive_raw_incremental', 'zfs_receive_-e', 'zfs_receive_raw_-d', 'zfs_receive_from_zstd', 'zfs_receive_new_props', 'zfs_receive_-wR-encrypted-mix', 'zfs_receive_corrective', 'zfs_receive_compressed_corrective', 'zfs_receive_large_block_corrective'] tags = ['functional', 'cli_root', 'zfs_receive'] [tests/functional/cli_root/zfs_rename] tests = ['zfs_rename_001_pos', 'zfs_rename_002_pos', 'zfs_rename_003_pos', 'zfs_rename_004_neg', 'zfs_rename_005_neg', 'zfs_rename_006_pos', 'zfs_rename_007_pos', 'zfs_rename_008_pos', 'zfs_rename_009_neg', 'zfs_rename_010_neg', 'zfs_rename_011_pos', 'zfs_rename_012_neg', 'zfs_rename_013_pos', 'zfs_rename_014_neg', 'zfs_rename_encrypted_child', 'zfs_rename_to_encrypted', 'zfs_rename_mountpoint', 'zfs_rename_nounmount'] tags = ['functional', 'cli_root', 'zfs_rename'] [tests/functional/cli_root/zfs_reservation] tests = ['zfs_reservation_001_pos', 'zfs_reservation_002_pos'] tags = ['functional', 'cli_root', 'zfs_reservation'] [tests/functional/cli_root/zfs_rollback] tests = ['zfs_rollback_001_pos', 'zfs_rollback_002_pos', 'zfs_rollback_003_neg', 'zfs_rollback_004_neg'] tags = ['functional', 'cli_root', 'zfs_rollback'] [tests/functional/cli_root/zfs_send] tests = ['zfs_send_001_pos', 'zfs_send_002_pos', 'zfs_send_003_pos', 'zfs_send_004_neg', 'zfs_send_005_pos', 'zfs_send_006_pos', 'zfs_send_007_pos', 'zfs_send_encrypted', 'zfs_send_encrypted_unloaded', 'zfs_send_raw', 'zfs_send_sparse', 'zfs_send-b', 'zfs_send_skip_missing'] tags = ['functional', 'cli_root', 'zfs_send'] [tests/functional/cli_root/zfs_set] tests = ['cache_001_pos', 'cache_002_neg', 'canmount_001_pos', 'canmount_002_pos', 'canmount_003_pos', 'canmount_004_pos', 'checksum_001_pos', 'compression_001_pos', 'mountpoint_001_pos', 'mountpoint_002_pos', 'reservation_001_neg', 'user_property_002_pos', 'share_mount_001_neg', 'snapdir_001_pos', 'onoffs_001_pos', 'user_property_001_pos', 'user_property_003_neg', 'readonly_001_pos', 'user_property_004_pos', 'version_001_neg', 'zfs_set_001_neg', 'zfs_set_002_neg', 'zfs_set_003_neg', 'property_alias_001_pos', 'mountpoint_003_pos', 'ro_props_001_pos', 'zfs_set_keylocation', 'zfs_set_feature_activation'] tags = ['functional', 'cli_root', 'zfs_set'] [tests/functional/cli_root/zfs_share] tests = ['zfs_share_001_pos', 'zfs_share_002_pos', 'zfs_share_003_pos', 'zfs_share_004_pos', 'zfs_share_006_pos', 'zfs_share_008_neg', 'zfs_share_010_neg', 'zfs_share_011_pos', 'zfs_share_concurrent_shares'] tags = ['functional', 'cli_root', 'zfs_share'] [tests/functional/cli_root/zfs_snapshot] tests = ['zfs_snapshot_001_neg', 'zfs_snapshot_002_neg', 'zfs_snapshot_003_neg', 'zfs_snapshot_004_neg', 'zfs_snapshot_005_neg', 'zfs_snapshot_006_pos', 'zfs_snapshot_007_neg', 'zfs_snapshot_008_neg', 'zfs_snapshot_009_pos'] tags = ['functional', 'cli_root', 'zfs_snapshot'] [tests/functional/cli_root/zfs_unload-key] tests = ['zfs_unload-key', 'zfs_unload-key_all', 'zfs_unload-key_recursive'] tags = ['functional', 'cli_root', 'zfs_unload-key'] [tests/functional/cli_root/zfs_unmount] tests = ['zfs_unmount_001_pos', 'zfs_unmount_002_pos', 'zfs_unmount_003_pos', 'zfs_unmount_004_pos', 'zfs_unmount_005_pos', 'zfs_unmount_006_pos', 'zfs_unmount_007_neg', 'zfs_unmount_008_neg', 'zfs_unmount_009_pos', 'zfs_unmount_all_001_pos', 'zfs_unmount_nested', 'zfs_unmount_unload_keys'] tags = ['functional', 'cli_root', 'zfs_unmount'] [tests/functional/cli_root/zfs_unshare] tests = ['zfs_unshare_001_pos', 'zfs_unshare_002_pos', 'zfs_unshare_003_pos', 'zfs_unshare_004_neg', 'zfs_unshare_005_neg', 'zfs_unshare_006_pos', 'zfs_unshare_007_pos'] tags = ['functional', 'cli_root', 'zfs_unshare'] [tests/functional/cli_root/zfs_upgrade] tests = ['zfs_upgrade_001_pos', 'zfs_upgrade_002_pos', 'zfs_upgrade_003_pos', 'zfs_upgrade_004_pos', 'zfs_upgrade_005_pos', 'zfs_upgrade_006_neg', 'zfs_upgrade_007_neg'] tags = ['functional', 'cli_root', 'zfs_upgrade'] [tests/functional/cli_root/zfs_wait] tests = ['zfs_wait_deleteq', 'zfs_wait_getsubopt'] tags = ['functional', 'cli_root', 'zfs_wait'] [tests/functional/cli_root/zhack] tests = ['zhack_label_repair_001', 'zhack_label_repair_002', 'zhack_label_repair_003', 'zhack_label_repair_004'] pre = post = tags = ['functional', 'cli_root', 'zhack'] [tests/functional/cli_root/zpool] tests = ['zpool_001_neg', 'zpool_002_pos', 'zpool_003_pos', 'zpool_colors'] tags = ['functional', 'cli_root', 'zpool'] [tests/functional/cli_root/zpool_add] tests = ['zpool_add_001_pos', 'zpool_add_002_pos', 'zpool_add_003_pos', 'zpool_add_004_pos', 'zpool_add_006_pos', 'zpool_add_007_neg', 'zpool_add_008_neg', 'zpool_add_009_neg', 'zpool_add_010_pos', 'add-o_ashift', 'add_prop_ashift', 'zpool_add_dryrun_output'] tags = ['functional', 'cli_root', 'zpool_add'] [tests/functional/cli_root/zpool_attach] tests = ['zpool_attach_001_neg', 'attach-o_ashift'] tags = ['functional', 'cli_root', 'zpool_attach'] [tests/functional/cli_root/zpool_clear] tests = ['zpool_clear_001_pos', 'zpool_clear_002_neg', 'zpool_clear_003_neg', 'zpool_clear_readonly'] tags = ['functional', 'cli_root', 'zpool_clear'] [tests/functional/cli_root/zpool_create] tests = ['zpool_create_001_pos', 'zpool_create_002_pos', 'zpool_create_003_pos', 'zpool_create_004_pos', 'zpool_create_005_pos', 'zpool_create_006_pos', 'zpool_create_007_neg', 'zpool_create_008_pos', 'zpool_create_009_neg', 'zpool_create_010_neg', 'zpool_create_011_neg', 'zpool_create_012_neg', 'zpool_create_014_neg', 'zpool_create_015_neg', 'zpool_create_017_neg', 'zpool_create_018_pos', 'zpool_create_019_pos', 'zpool_create_020_pos', 'zpool_create_021_pos', 'zpool_create_022_pos', 'zpool_create_023_neg', 'zpool_create_024_pos', 'zpool_create_encrypted', 'zpool_create_crypt_combos', 'zpool_create_draid_001_pos', 'zpool_create_draid_002_pos', 'zpool_create_draid_003_pos', 'zpool_create_draid_004_pos', 'zpool_create_features_001_pos', 'zpool_create_features_002_pos', 'zpool_create_features_003_pos', 'zpool_create_features_004_neg', 'zpool_create_features_005_pos', 'zpool_create_features_006_pos', 'zpool_create_features_007_pos', 'zpool_create_features_008_pos', 'zpool_create_features_009_pos', 'create-o_ashift', 'zpool_create_tempname', 'zpool_create_dryrun_output'] tags = ['functional', 'cli_root', 'zpool_create'] [tests/functional/cli_root/zpool_destroy] tests = ['zpool_destroy_001_pos', 'zpool_destroy_002_pos', 'zpool_destroy_003_neg'] pre = post = tags = ['functional', 'cli_root', 'zpool_destroy'] [tests/functional/cli_root/zpool_detach] tests = ['zpool_detach_001_neg'] tags = ['functional', 'cli_root', 'zpool_detach'] [tests/functional/cli_root/zpool_events] tests = ['zpool_events_clear', 'zpool_events_cliargs', 'zpool_events_follow', 'zpool_events_poolname', 'zpool_events_errors', 'zpool_events_duplicates', 'zpool_events_clear_retained'] tags = ['functional', 'cli_root', 'zpool_events'] [tests/functional/cli_root/zpool_export] tests = ['zpool_export_001_pos', 'zpool_export_002_pos', 'zpool_export_003_neg', 'zpool_export_004_pos'] tags = ['functional', 'cli_root', 'zpool_export'] [tests/functional/cli_root/zpool_get] tests = ['zpool_get_001_pos', 'zpool_get_002_pos', 'zpool_get_003_pos', 'zpool_get_004_neg', 'zpool_get_005_pos', 'vdev_get_001_pos'] tags = ['functional', 'cli_root', 'zpool_get'] [tests/functional/cli_root/zpool_history] tests = ['zpool_history_001_neg', 'zpool_history_002_pos'] tags = ['functional', 'cli_root', 'zpool_history'] [tests/functional/cli_root/zpool_import] tests = ['zpool_import_001_pos', 'zpool_import_002_pos', 'zpool_import_003_pos', 'zpool_import_004_pos', 'zpool_import_005_pos', 'zpool_import_006_pos', 'zpool_import_007_pos', 'zpool_import_008_pos', 'zpool_import_009_neg', 'zpool_import_010_pos', 'zpool_import_011_neg', 'zpool_import_012_pos', 'zpool_import_013_neg', 'zpool_import_014_pos', 'zpool_import_015_pos', 'zpool_import_016_pos', 'zpool_import_017_pos', 'zpool_import_features_001_pos', 'zpool_import_features_002_neg', 'zpool_import_features_003_pos', 'zpool_import_missing_001_pos', 'zpool_import_missing_002_pos', 'zpool_import_missing_003_pos', 'zpool_import_rename_001_pos', 'zpool_import_all_001_pos', 'zpool_import_encrypted', 'zpool_import_encrypted_load', 'zpool_import_errata3', 'zpool_import_errata4', 'import_cachefile_device_added', 'import_cachefile_device_removed', 'import_cachefile_device_replaced', 'import_cachefile_mirror_attached', 'import_cachefile_mirror_detached', 'import_cachefile_paths_changed', 'import_cachefile_shared_device', 'import_devices_missing', 'import_log_missing', 'import_paths_changed', 'import_rewind_config_changed', 'import_rewind_device_replaced'] tags = ['functional', 'cli_root', 'zpool_import'] timeout = 1200 [tests/functional/cli_root/zpool_labelclear] tests = ['zpool_labelclear_active', 'zpool_labelclear_exported', 'zpool_labelclear_removed', 'zpool_labelclear_valid'] pre = post = tags = ['functional', 'cli_root', 'zpool_labelclear'] [tests/functional/cli_root/zpool_initialize] tests = ['zpool_initialize_attach_detach_add_remove', 'zpool_initialize_fault_export_import_online', 'zpool_initialize_import_export', 'zpool_initialize_offline_export_import_online', 'zpool_initialize_online_offline', 'zpool_initialize_split', 'zpool_initialize_start_and_cancel_neg', 'zpool_initialize_start_and_cancel_pos', 'zpool_initialize_suspend_resume', + 'zpool_initialize_uninit', 'zpool_initialize_unsupported_vdevs', 'zpool_initialize_verify_checksums', 'zpool_initialize_verify_initialized'] pre = tags = ['functional', 'cli_root', 'zpool_initialize'] [tests/functional/cli_root/zpool_offline] tests = ['zpool_offline_001_pos', 'zpool_offline_002_neg', 'zpool_offline_003_pos'] tags = ['functional', 'cli_root', 'zpool_offline'] [tests/functional/cli_root/zpool_online] tests = ['zpool_online_001_pos', 'zpool_online_002_neg'] tags = ['functional', 'cli_root', 'zpool_online'] [tests/functional/cli_root/zpool_remove] tests = ['zpool_remove_001_neg', 'zpool_remove_002_pos', 'zpool_remove_003_pos'] tags = ['functional', 'cli_root', 'zpool_remove'] [tests/functional/cli_root/zpool_replace] tests = ['zpool_replace_001_neg', 'replace-o_ashift', 'replace_prop_ashift'] tags = ['functional', 'cli_root', 'zpool_replace'] [tests/functional/cli_root/zpool_resilver] tests = ['zpool_resilver_bad_args', 'zpool_resilver_restart'] tags = ['functional', 'cli_root', 'zpool_resilver'] [tests/functional/cli_root/zpool_scrub] tests = ['zpool_scrub_001_neg', 'zpool_scrub_002_pos', 'zpool_scrub_003_pos', 'zpool_scrub_004_pos', 'zpool_scrub_005_pos', 'zpool_scrub_encrypted_unloaded', 'zpool_scrub_print_repairing', - 'zpool_scrub_offline_device', 'zpool_scrub_multiple_copies'] + 'zpool_scrub_offline_device', 'zpool_scrub_multiple_copies', + 'zpool_error_scrub_001_pos', 'zpool_error_scrub_002_pos', + 'zpool_error_scrub_003_pos', 'zpool_error_scrub_004_pos'] tags = ['functional', 'cli_root', 'zpool_scrub'] [tests/functional/cli_root/zpool_set] tests = ['zpool_set_001_pos', 'zpool_set_002_neg', 'zpool_set_003_neg', 'zpool_set_ashift', 'zpool_set_features', 'vdev_set_001_pos', 'user_property_001_pos', 'user_property_002_neg'] tags = ['functional', 'cli_root', 'zpool_set'] [tests/functional/cli_root/zpool_split] tests = ['zpool_split_cliargs', 'zpool_split_devices', 'zpool_split_encryption', 'zpool_split_props', 'zpool_split_vdevs', 'zpool_split_resilver', 'zpool_split_indirect', 'zpool_split_dryrun_output'] tags = ['functional', 'cli_root', 'zpool_split'] [tests/functional/cli_root/zpool_status] tests = ['zpool_status_001_pos', 'zpool_status_002_pos', 'zpool_status_003_pos', 'zpool_status_004_pos', 'zpool_status_005_pos', 'zpool_status_006_pos', 'zpool_status_007_pos', 'zpool_status_features_001_pos'] tags = ['functional', 'cli_root', 'zpool_status'] [tests/functional/cli_root/zpool_sync] tests = ['zpool_sync_001_pos', 'zpool_sync_002_neg'] tags = ['functional', 'cli_root', 'zpool_sync'] [tests/functional/cli_root/zpool_trim] tests = ['zpool_trim_attach_detach_add_remove', 'zpool_trim_fault_export_import_online', 'zpool_trim_import_export', 'zpool_trim_multiple', 'zpool_trim_neg', 'zpool_trim_offline_export_import_online', 'zpool_trim_online_offline', 'zpool_trim_partial', 'zpool_trim_rate', 'zpool_trim_rate_neg', 'zpool_trim_secure', 'zpool_trim_split', 'zpool_trim_start_and_cancel_neg', 'zpool_trim_start_and_cancel_pos', 'zpool_trim_suspend_resume', 'zpool_trim_unsupported_vdevs', 'zpool_trim_verify_checksums', 'zpool_trim_verify_trimmed'] tags = ['functional', 'zpool_trim'] [tests/functional/cli_root/zpool_upgrade] tests = ['zpool_upgrade_001_pos', 'zpool_upgrade_002_pos', 'zpool_upgrade_003_pos', 'zpool_upgrade_004_pos', 'zpool_upgrade_005_neg', 'zpool_upgrade_006_neg', 'zpool_upgrade_007_pos', 'zpool_upgrade_008_pos', 'zpool_upgrade_009_neg', 'zpool_upgrade_features_001_pos'] tags = ['functional', 'cli_root', 'zpool_upgrade'] [tests/functional/cli_root/zpool_wait] tests = ['zpool_wait_discard', 'zpool_wait_freeing', 'zpool_wait_initialize_basic', 'zpool_wait_initialize_cancel', 'zpool_wait_initialize_flag', 'zpool_wait_multiple', 'zpool_wait_no_activity', 'zpool_wait_remove', 'zpool_wait_remove_cancel', 'zpool_wait_trim_basic', 'zpool_wait_trim_cancel', 'zpool_wait_trim_flag', 'zpool_wait_usage'] tags = ['functional', 'cli_root', 'zpool_wait'] [tests/functional/cli_root/zpool_wait/scan] tests = ['zpool_wait_replace_cancel', 'zpool_wait_rebuild', 'zpool_wait_resilver', 'zpool_wait_scrub_cancel', 'zpool_wait_replace', 'zpool_wait_scrub_basic', 'zpool_wait_scrub_flag'] tags = ['functional', 'cli_root', 'zpool_wait'] [tests/functional/cli_user/misc] tests = ['zdb_001_neg', 'zfs_001_neg', 'zfs_allow_001_neg', 'zfs_clone_001_neg', 'zfs_create_001_neg', 'zfs_destroy_001_neg', 'zfs_get_001_neg', 'zfs_inherit_001_neg', 'zfs_mount_001_neg', 'zfs_promote_001_neg', 'zfs_receive_001_neg', 'zfs_rename_001_neg', 'zfs_rollback_001_neg', 'zfs_send_001_neg', 'zfs_set_001_neg', 'zfs_share_001_neg', 'zfs_snapshot_001_neg', 'zfs_unallow_001_neg', 'zfs_unmount_001_neg', 'zfs_unshare_001_neg', 'zfs_upgrade_001_neg', 'zpool_001_neg', 'zpool_add_001_neg', 'zpool_attach_001_neg', 'zpool_clear_001_neg', 'zpool_create_001_neg', 'zpool_destroy_001_neg', 'zpool_detach_001_neg', 'zpool_export_001_neg', 'zpool_get_001_neg', 'zpool_history_001_neg', 'zpool_import_001_neg', 'zpool_import_002_neg', 'zpool_offline_001_neg', 'zpool_online_001_neg', 'zpool_remove_001_neg', 'zpool_replace_001_neg', 'zpool_scrub_001_neg', 'zpool_set_001_neg', 'zpool_status_001_neg', 'zpool_upgrade_001_neg', 'arcstat_001_pos', 'arc_summary_001_pos', 'arc_summary_002_neg', 'zpool_wait_privilege', 'zilstat_001_pos'] user = tags = ['functional', 'cli_user', 'misc'] [tests/functional/cli_user/zfs_list] tests = ['zfs_list_001_pos', 'zfs_list_002_pos', 'zfs_list_003_pos', 'zfs_list_004_neg', 'zfs_list_005_neg', 'zfs_list_007_pos', 'zfs_list_008_neg'] user = tags = ['functional', 'cli_user', 'zfs_list'] [tests/functional/cli_user/zpool_iostat] tests = ['zpool_iostat_001_neg', 'zpool_iostat_002_pos', 'zpool_iostat_003_neg', 'zpool_iostat_004_pos', 'zpool_iostat_005_pos', 'zpool_iostat_-c_disable', 'zpool_iostat_-c_homedir', 'zpool_iostat_-c_searchpath'] user = tags = ['functional', 'cli_user', 'zpool_iostat'] [tests/functional/cli_user/zpool_list] tests = ['zpool_list_001_pos', 'zpool_list_002_neg'] user = tags = ['functional', 'cli_user', 'zpool_list'] [tests/functional/cli_user/zpool_status] tests = ['zpool_status_003_pos', 'zpool_status_-c_disable', 'zpool_status_-c_homedir', 'zpool_status_-c_searchpath'] user = tags = ['functional', 'cli_user', 'zpool_status'] [tests/functional/compression] tests = ['compress_001_pos', 'compress_002_pos', 'compress_003_pos', 'l2arc_compressed_arc', 'l2arc_compressed_arc_disabled', 'l2arc_encrypted', 'l2arc_encrypted_no_compressed_arc'] tags = ['functional', 'compression'] [tests/functional/cp_files] tests = ['cp_files_001_pos'] tags = ['functional', 'cp_files'] [tests/functional/crtime] tests = ['crtime_001_pos' ] tags = ['functional', 'crtime'] [tests/functional/ctime] tests = ['ctime_001_pos' ] tags = ['functional', 'ctime'] [tests/functional/deadman] tests = ['deadman_ratelimit', 'deadman_sync', 'deadman_zio'] pre = post = tags = ['functional', 'deadman'] [tests/functional/delegate] tests = ['zfs_allow_001_pos', 'zfs_allow_002_pos', 'zfs_allow_003_pos', 'zfs_allow_004_pos', 'zfs_allow_005_pos', 'zfs_allow_006_pos', 'zfs_allow_007_pos', 'zfs_allow_008_pos', 'zfs_allow_009_neg', 'zfs_allow_010_pos', 'zfs_allow_011_neg', 'zfs_allow_012_neg', 'zfs_unallow_001_pos', 'zfs_unallow_002_pos', 'zfs_unallow_003_pos', 'zfs_unallow_004_pos', 'zfs_unallow_005_pos', 'zfs_unallow_006_pos', 'zfs_unallow_007_neg', 'zfs_unallow_008_neg'] tags = ['functional', 'delegate'] [tests/functional/exec] tests = ['exec_001_pos', 'exec_002_neg'] tags = ['functional', 'exec'] [tests/functional/fallocate] tests = ['fallocate_punch-hole'] tags = ['functional', 'fallocate'] [tests/functional/features/async_destroy] tests = ['async_destroy_001_pos'] tags = ['functional', 'features', 'async_destroy'] [tests/functional/features/large_dnode] tests = ['large_dnode_001_pos', 'large_dnode_003_pos', 'large_dnode_004_neg', 'large_dnode_005_pos', 'large_dnode_007_neg', 'large_dnode_009_pos'] tags = ['functional', 'features', 'large_dnode'] [tests/functional/grow] pre = post = tests = ['grow_pool_001_pos', 'grow_replicas_001_pos'] tags = ['functional', 'grow'] [tests/functional/history] tests = ['history_001_pos', 'history_002_pos', 'history_003_pos', 'history_004_pos', 'history_005_neg', 'history_006_neg', 'history_007_pos', 'history_008_pos', 'history_009_pos', 'history_010_pos'] tags = ['functional', 'history'] [tests/functional/hkdf] pre = post = tests = ['hkdf_test'] tags = ['functional', 'hkdf'] [tests/functional/inheritance] tests = ['inherit_001_pos'] pre = tags = ['functional', 'inheritance'] [tests/functional/io] tests = ['sync', 'psync', 'posixaio', 'mmap'] tags = ['functional', 'io'] [tests/functional/inuse] tests = ['inuse_004_pos', 'inuse_005_pos', 'inuse_008_pos', 'inuse_009_pos'] post = tags = ['functional', 'inuse'] [tests/functional/large_files] tests = ['large_files_001_pos', 'large_files_002_pos'] tags = ['functional', 'large_files'] [tests/functional/limits] tests = ['filesystem_count', 'filesystem_limit', 'snapshot_count', 'snapshot_limit'] tags = ['functional', 'limits'] [tests/functional/link_count] tests = ['link_count_001', 'link_count_root_inode'] tags = ['functional', 'link_count'] [tests/functional/migration] tests = ['migration_001_pos', 'migration_002_pos', 'migration_003_pos', 'migration_004_pos', 'migration_005_pos', 'migration_006_pos', 'migration_007_pos', 'migration_008_pos', 'migration_009_pos', 'migration_010_pos', 'migration_011_pos', 'migration_012_pos'] tags = ['functional', 'migration'] [tests/functional/mmap] tests = ['mmap_mixed', 'mmap_read_001_pos', 'mmap_seek_001_pos', 'mmap_sync_001_pos', 'mmap_write_001_pos'] tags = ['functional', 'mmap'] [tests/functional/mount] tests = ['umount_001', 'umountall_001'] tags = ['functional', 'mount'] [tests/functional/mv_files] tests = ['mv_files_001_pos', 'mv_files_002_pos', 'random_creation'] tags = ['functional', 'mv_files'] [tests/functional/nestedfs] tests = ['nestedfs_001_pos'] tags = ['functional', 'nestedfs'] [tests/functional/no_space] tests = ['enospc_001_pos', 'enospc_002_pos', 'enospc_003_pos', 'enospc_df', 'enospc_ganging', 'enospc_rm'] tags = ['functional', 'no_space'] [tests/functional/nopwrite] tests = ['nopwrite_copies', 'nopwrite_mtime', 'nopwrite_negative', 'nopwrite_promoted_clone', 'nopwrite_recsize', 'nopwrite_sync', 'nopwrite_varying_compression', 'nopwrite_volume'] tags = ['functional', 'nopwrite'] [tests/functional/online_offline] tests = ['online_offline_001_pos', 'online_offline_002_neg', 'online_offline_003_neg'] tags = ['functional', 'online_offline'] [tests/functional/pool_checkpoint] tests = ['checkpoint_after_rewind', 'checkpoint_big_rewind', 'checkpoint_capacity', 'checkpoint_conf_change', 'checkpoint_discard', 'checkpoint_discard_busy', 'checkpoint_discard_many', 'checkpoint_indirect', 'checkpoint_invalid', 'checkpoint_lun_expsz', 'checkpoint_open', 'checkpoint_removal', 'checkpoint_rewind', 'checkpoint_ro_rewind', 'checkpoint_sm_scale', 'checkpoint_twice', 'checkpoint_vdev_add', 'checkpoint_zdb', 'checkpoint_zhack_feat'] tags = ['functional', 'pool_checkpoint'] timeout = 1800 [tests/functional/pool_names] tests = ['pool_names_001_pos', 'pool_names_002_neg'] pre = post = tags = ['functional', 'pool_names'] [tests/functional/poolversion] tests = ['poolversion_001_pos', 'poolversion_002_pos'] tags = ['functional', 'poolversion'] [tests/functional/pyzfs] tests = ['pyzfs_unittest'] pre = post = tags = ['functional', 'pyzfs'] [tests/functional/quota] tests = ['quota_001_pos', 'quota_002_pos', 'quota_003_pos', 'quota_004_pos', 'quota_005_pos', 'quota_006_neg'] tags = ['functional', 'quota'] [tests/functional/redacted_send] tests = ['redacted_compressed', 'redacted_contents', 'redacted_deleted', 'redacted_disabled_feature', 'redacted_embedded', 'redacted_holes', 'redacted_incrementals', 'redacted_largeblocks', 'redacted_many_clones', 'redacted_mixed_recsize', 'redacted_mounts', 'redacted_negative', 'redacted_origin', 'redacted_panic', 'redacted_props', 'redacted_resume', 'redacted_size', 'redacted_volume'] tags = ['functional', 'redacted_send'] [tests/functional/raidz] tests = ['raidz_001_neg', 'raidz_002_pos', 'raidz_003_pos', 'raidz_004_pos'] tags = ['functional', 'raidz'] [tests/functional/redundancy] tests = ['redundancy_draid', 'redundancy_draid1', 'redundancy_draid2', 'redundancy_draid3', 'redundancy_draid_damaged1', 'redundancy_draid_damaged2', 'redundancy_draid_spare1', 'redundancy_draid_spare2', 'redundancy_draid_spare3', 'redundancy_mirror', 'redundancy_raidz', 'redundancy_raidz1', 'redundancy_raidz2', 'redundancy_raidz3', 'redundancy_stripe'] tags = ['functional', 'redundancy'] timeout = 1200 [tests/functional/refquota] tests = ['refquota_001_pos', 'refquota_002_pos', 'refquota_003_pos', 'refquota_004_pos', 'refquota_005_pos', 'refquota_006_neg', 'refquota_007_neg', 'refquota_008_neg'] tags = ['functional', 'refquota'] [tests/functional/refreserv] tests = ['refreserv_001_pos', 'refreserv_002_pos', 'refreserv_003_pos', 'refreserv_004_pos', 'refreserv_005_pos', 'refreserv_multi_raidz', 'refreserv_raidz'] tags = ['functional', 'refreserv'] [tests/functional/removal] pre = tests = ['removal_all_vdev', 'removal_cancel', 'removal_check_space', 'removal_condense_export', 'removal_multiple_indirection', 'removal_nopwrite', 'removal_remap_deadlists', 'removal_resume_export', 'removal_sanity', 'removal_with_add', 'removal_with_create_fs', 'removal_with_dedup', 'removal_with_errors', 'removal_with_export', 'removal_with_indirect', 'removal_with_ganging', 'removal_with_faulted', 'removal_with_remove', 'removal_with_scrub', 'removal_with_send', 'removal_with_send_recv', 'removal_with_snapshot', 'removal_with_write', 'removal_with_zdb', 'remove_expanded', 'remove_mirror', 'remove_mirror_sanity', 'remove_raidz', 'remove_indirect', 'remove_attach_mirror', 'removal_reservation'] tags = ['functional', 'removal'] [tests/functional/rename_dirs] tests = ['rename_dirs_001_pos'] tags = ['functional', 'rename_dirs'] [tests/functional/replacement] tests = ['attach_import', 'attach_multiple', 'attach_rebuild', 'attach_resilver', 'detach', 'rebuild_disabled_feature', 'rebuild_multiple', 'rebuild_raidz', 'replace_import', 'replace_rebuild', 'replace_resilver', 'resilver_restart_001', 'resilver_restart_002', 'scrub_cancel'] tags = ['functional', 'replacement'] [tests/functional/reservation] tests = ['reservation_001_pos', 'reservation_002_pos', 'reservation_003_pos', 'reservation_004_pos', 'reservation_005_pos', 'reservation_006_pos', 'reservation_007_pos', 'reservation_008_pos', 'reservation_009_pos', 'reservation_010_pos', 'reservation_011_pos', 'reservation_012_pos', 'reservation_013_pos', 'reservation_014_pos', 'reservation_015_pos', 'reservation_016_pos', 'reservation_017_pos', 'reservation_018_pos', 'reservation_019_pos', 'reservation_020_pos', 'reservation_021_neg', 'reservation_022_pos'] tags = ['functional', 'reservation'] [tests/functional/rootpool] tests = ['rootpool_002_neg', 'rootpool_003_neg', 'rootpool_007_pos'] tags = ['functional', 'rootpool'] [tests/functional/rsend] tests = ['recv_dedup', 'recv_dedup_encrypted_zvol', 'rsend_001_pos', 'rsend_002_pos', 'rsend_003_pos', 'rsend_004_pos', 'rsend_005_pos', 'rsend_006_pos', 'rsend_007_pos', 'rsend_008_pos', 'rsend_009_pos', 'rsend_010_pos', 'rsend_011_pos', 'rsend_012_pos', 'rsend_013_pos', 'rsend_014_pos', 'rsend_016_neg', 'rsend_019_pos', 'rsend_020_pos', 'rsend_021_pos', 'rsend_022_pos', 'rsend_024_pos', 'rsend_025_pos', 'rsend_026_neg', 'rsend_027_pos', 'rsend_028_neg', 'rsend_029_neg', 'rsend_030_pos', 'rsend_031_pos', 'send-c_verify_ratio', 'send-c_verify_contents', 'send-c_props', 'send-c_incremental', 'send-c_volume', 'send-c_zstream_recompress', 'send-c_zstreamdump', 'send-c_lz4_disabled', 'send-c_recv_lz4_disabled', 'send-c_mixed_compression', 'send-c_stream_size_estimate', 'send-c_embedded_blocks', 'send-c_resume', 'send-cpL_varied_recsize', 'send-c_recv_dedup', 'send-L_toggle', 'send_encrypted_incremental', 'send_encrypted_freeobjects', 'send_encrypted_hierarchy', 'send_encrypted_props', 'send_encrypted_truncated_files', 'send_freeobjects', 'send_realloc_files', 'send_realloc_encrypted_files', 'send_spill_block', 'send_holds', 'send_hole_birth', 'send_mixed_raw', 'send-wR_encrypted_zvol', 'send_partial_dataset', 'send_invalid', 'send_doall', 'send_raw_spill_block', 'send_raw_ashift', 'send_raw_large_blocks'] tags = ['functional', 'rsend'] [tests/functional/scrub_mirror] tests = ['scrub_mirror_001_pos', 'scrub_mirror_002_pos', 'scrub_mirror_003_pos', 'scrub_mirror_004_pos'] tags = ['functional', 'scrub_mirror'] [tests/functional/slog] tests = ['slog_001_pos', 'slog_002_pos', 'slog_003_pos', 'slog_004_pos', 'slog_005_pos', 'slog_006_pos', 'slog_007_pos', 'slog_008_neg', 'slog_009_neg', 'slog_010_neg', 'slog_011_neg', 'slog_012_neg', 'slog_013_pos', 'slog_014_pos', 'slog_015_neg', 'slog_replay_fs_001', 'slog_replay_fs_002', 'slog_replay_volume', 'slog_016_pos'] tags = ['functional', 'slog'] [tests/functional/snapshot] tests = ['clone_001_pos', 'rollback_001_pos', 'rollback_002_pos', 'rollback_003_pos', 'snapshot_001_pos', 'snapshot_002_pos', 'snapshot_003_pos', 'snapshot_004_pos', 'snapshot_005_pos', 'snapshot_006_pos', 'snapshot_007_pos', 'snapshot_008_pos', 'snapshot_009_pos', 'snapshot_010_pos', 'snapshot_011_pos', 'snapshot_012_pos', 'snapshot_013_pos', 'snapshot_014_pos', 'snapshot_017_pos', 'snapshot_018_pos'] tags = ['functional', 'snapshot'] [tests/functional/snapused] tests = ['snapused_001_pos', 'snapused_002_pos', 'snapused_003_pos', 'snapused_004_pos', 'snapused_005_pos'] tags = ['functional', 'snapused'] [tests/functional/sparse] tests = ['sparse_001_pos'] tags = ['functional', 'sparse'] [tests/functional/stat] tests = ['stat_001_pos'] tags = ['functional', 'stat'] [tests/functional/suid] tests = ['suid_write_to_suid', 'suid_write_to_sgid', 'suid_write_to_suid_sgid', 'suid_write_to_none', 'suid_write_zil_replay'] tags = ['functional', 'suid'] [tests/functional/trim] tests = ['autotrim_integrity', 'autotrim_config', 'autotrim_trim_integrity', 'trim_integrity', 'trim_config', 'trim_l2arc'] tags = ['functional', 'trim'] [tests/functional/truncate] tests = ['truncate_001_pos', 'truncate_002_pos', 'truncate_timestamps'] tags = ['functional', 'truncate'] [tests/functional/upgrade] tests = ['upgrade_userobj_001_pos', 'upgrade_readonly_pool'] tags = ['functional', 'upgrade'] [tests/functional/userquota] tests = [ 'userquota_001_pos', 'userquota_002_pos', 'userquota_003_pos', 'userquota_004_pos', 'userquota_005_neg', 'userquota_006_pos', 'userquota_007_pos', 'userquota_008_pos', 'userquota_009_pos', 'userquota_010_pos', 'userquota_011_pos', 'userquota_012_neg', 'userspace_001_pos', 'userspace_002_pos', 'userspace_encrypted', 'userspace_send_encrypted', 'userspace_encrypted_13709'] tags = ['functional', 'userquota'] [tests/functional/vdev_zaps] tests = ['vdev_zaps_001_pos', 'vdev_zaps_002_pos', 'vdev_zaps_003_pos', 'vdev_zaps_004_pos', 'vdev_zaps_005_pos', 'vdev_zaps_006_pos', 'vdev_zaps_007_pos'] tags = ['functional', 'vdev_zaps'] [tests/functional/write_dirs] tests = ['write_dirs_001_pos', 'write_dirs_002_pos'] tags = ['functional', 'write_dirs'] [tests/functional/xattr] tests = ['xattr_001_pos', 'xattr_002_neg', 'xattr_003_neg', 'xattr_004_pos', 'xattr_005_pos', 'xattr_006_pos', 'xattr_007_neg', 'xattr_011_pos', 'xattr_012_pos', 'xattr_013_pos', 'xattr_compat'] tags = ['functional', 'xattr'] [tests/functional/zvol/zvol_ENOSPC] tests = ['zvol_ENOSPC_001_pos'] tags = ['functional', 'zvol', 'zvol_ENOSPC'] [tests/functional/zvol/zvol_cli] tests = ['zvol_cli_001_pos', 'zvol_cli_002_pos', 'zvol_cli_003_neg'] tags = ['functional', 'zvol', 'zvol_cli'] [tests/functional/zvol/zvol_misc] tests = ['zvol_misc_002_pos', 'zvol_misc_hierarchy', 'zvol_misc_rename_inuse', 'zvol_misc_snapdev', 'zvol_misc_trim', 'zvol_misc_volmode', 'zvol_misc_zil'] tags = ['functional', 'zvol', 'zvol_misc'] [tests/functional/zvol/zvol_stress] tests = ['zvol_stress'] tags = ['functional', 'zvol', 'zvol_stress'] [tests/functional/zvol/zvol_swap] tests = ['zvol_swap_001_pos', 'zvol_swap_002_pos', 'zvol_swap_004_pos'] tags = ['functional', 'zvol', 'zvol_swap'] [tests/functional/libzfs] tests = ['many_fds', 'libzfs_input'] tags = ['functional', 'libzfs'] [tests/functional/log_spacemap] tests = ['log_spacemap_import_logs'] pre = post = tags = ['functional', 'log_spacemap'] [tests/functional/l2arc] tests = ['l2arc_arcstats_pos', 'l2arc_mfuonly_pos', 'l2arc_l2miss_pos', 'persist_l2arc_001_pos', 'persist_l2arc_002_pos', 'persist_l2arc_003_neg', 'persist_l2arc_004_pos', 'persist_l2arc_005_pos'] tags = ['functional', 'l2arc'] [tests/functional/zpool_influxdb] tests = ['zpool_influxdb'] tags = ['functional', 'zpool_influxdb'] diff --git a/sys/contrib/openzfs/tests/test-runner/bin/test-runner.py.in b/sys/contrib/openzfs/tests/test-runner/bin/test-runner.py.in index c454bf8d7c6a..422ebd7bc8bf 100755 --- a/sys/contrib/openzfs/tests/test-runner/bin/test-runner.py.in +++ b/sys/contrib/openzfs/tests/test-runner/bin/test-runner.py.in @@ -1,1150 +1,1156 @@ #!/usr/bin/env @PYTHON_SHEBANG@ # # This file and its contents are supplied under the terms of the # Common Development and Distribution License ("CDDL"), version 1.0. # You may only use this file in accordance with the terms of version # 1.0 of the CDDL. # # A full copy of the text of the CDDL should have accompanied this # source. A copy of the CDDL is also available via the Internet at # http://www.illumos.org/license/CDDL. # # # Copyright (c) 2012, 2018 by Delphix. All rights reserved. # Copyright (c) 2019 Datto Inc. # # This script must remain compatible with Python 3.6+. # import os import sys import ctypes import re import configparser from datetime import datetime from optparse import OptionParser from pwd import getpwnam from pwd import getpwuid from select import select from subprocess import PIPE from subprocess import Popen from subprocess import check_output from threading import Timer from time import time, CLOCK_MONOTONIC from os.path import exists BASEDIR = '/var/tmp/test_results' TESTDIR = '/usr/share/zfs/' KMEMLEAK_FILE = '/sys/kernel/debug/kmemleak' KILL = 'kill' TRUE = 'true' SUDO = 'sudo' LOG_FILE = 'LOG_FILE' LOG_OUT = 'LOG_OUT' LOG_ERR = 'LOG_ERR' LOG_FILE_OBJ = None try: from time import monotonic as monotonic_time except ImportError: class timespec(ctypes.Structure): _fields_ = [ ('tv_sec', ctypes.c_long), ('tv_nsec', ctypes.c_long) ] librt = ctypes.CDLL('librt.so.1', use_errno=True) clock_gettime = librt.clock_gettime clock_gettime.argtypes = [ctypes.c_int, ctypes.POINTER(timespec)] def monotonic_time(): t = timespec() if clock_gettime(CLOCK_MONOTONIC, ctypes.pointer(t)) != 0: errno_ = ctypes.get_errno() raise OSError(errno_, os.strerror(errno_)) return t.tv_sec + t.tv_nsec * 1e-9 class Result(object): total = 0 runresults = {'PASS': 0, 'FAIL': 0, 'SKIP': 0, 'KILLED': 0, 'RERAN': 0} def __init__(self): self.starttime = None self.returncode = None self.runtime = '' self.stdout = [] self.stderr = [] self.kmemleak = '' self.result = '' def done(self, proc, killed, reran): """ Finalize the results of this Cmd. """ Result.total += 1 m, s = divmod(monotonic_time() - self.starttime, 60) self.runtime = '%02d:%02d' % (m, s) self.returncode = proc.returncode if reran is True: Result.runresults['RERAN'] += 1 if killed: self.result = 'KILLED' Result.runresults['KILLED'] += 1 elif len(self.kmemleak) > 0: self.result = 'FAIL' Result.runresults['FAIL'] += 1 elif self.returncode == 0: self.result = 'PASS' Result.runresults['PASS'] += 1 elif self.returncode == 4: self.result = 'SKIP' Result.runresults['SKIP'] += 1 elif self.returncode != 0: self.result = 'FAIL' Result.runresults['FAIL'] += 1 class Output(object): """ This class is a slightly modified version of the 'Stream' class found here: http://goo.gl/aSGfv """ def __init__(self, stream): self.stream = stream self._buf = b'' self.lines = [] def fileno(self): return self.stream.fileno() def read(self, drain=0): """ Read from the file descriptor. If 'drain' set, read until EOF. """ while self._read() is not None: if not drain: break def _read(self): """ Read up to 4k of data from this output stream. Collect the output up to the last newline, and append it to any leftover data from a previous call. The lines are stored as a (timestamp, data) tuple for easy sorting/merging later. """ fd = self.fileno() buf = os.read(fd, 4096) if not buf: return None if b'\n' not in buf: self._buf += buf return [] buf = self._buf + buf tmp, rest = buf.rsplit(b'\n', 1) self._buf = rest now = datetime.now() rows = tmp.split(b'\n') self.lines += [(now, r) for r in rows] class Cmd(object): verified_users = [] def __init__(self, pathname, identifier=None, outputdir=None, timeout=None, user=None, tags=None): self.pathname = pathname self.identifier = identifier self.outputdir = outputdir or 'BASEDIR' """ The timeout for tests is measured in wall-clock time """ self.timeout = timeout self.user = user or '' self.killed = False self.reran = None self.result = Result() if self.timeout is None: self.timeout = 60 def __str__(self): return '''\ Pathname: %s Identifier: %s Outputdir: %s Timeout: %d User: %s ''' % (self.pathname, self.identifier, self.outputdir, self.timeout, self.user) - def kill_cmd(self, proc, keyboard_interrupt=False): + def kill_cmd(self, proc, options, kmemleak, keyboard_interrupt=False): """ Kill a running command due to timeout, or ^C from the keyboard. If sudo is required, this user was verified previously. """ self.killed = True do_sudo = len(self.user) != 0 signal = '-TERM' cmd = [SUDO, KILL, signal, str(proc.pid)] if not do_sudo: del cmd[0] try: kp = Popen(cmd) kp.wait() except Exception: pass """ If this is not a user-initiated kill and the test has not been reran before we consider if the test needs to be reran: If the test has spent some time hibernating and didn't run the whole length of time before being timed out we will rerun the test. """ if keyboard_interrupt is False and self.reran is None: runtime = monotonic_time() - self.result.starttime if int(self.timeout) > runtime: self.killed = False self.reran = False - self.run(False) + self.run(options, dryrun=False, kmemleak=kmemleak) self.reran = True def update_cmd_privs(self, cmd, user): """ If a user has been specified to run this Cmd and we're not already running as that user, prepend the appropriate sudo command to run as that user. """ me = getpwuid(os.getuid()) if not user or user is me: if os.path.isfile(cmd+'.ksh') and os.access(cmd+'.ksh', os.X_OK): cmd += '.ksh' if os.path.isfile(cmd+'.sh') and os.access(cmd+'.sh', os.X_OK): cmd += '.sh' return cmd if not os.path.isfile(cmd): if os.path.isfile(cmd+'.ksh') and os.access(cmd+'.ksh', os.X_OK): cmd += '.ksh' if os.path.isfile(cmd+'.sh') and os.access(cmd+'.sh', os.X_OK): cmd += '.sh' ret = '%s -E -u %s %s' % (SUDO, user, cmd) return ret.split(' ') def collect_output(self, proc): """ Read from stdout/stderr as data becomes available, until the process is no longer running. Return the lines from the stdout and stderr Output objects. """ out = Output(proc.stdout) err = Output(proc.stderr) res = [] while proc.returncode is None: proc.poll() res = select([out, err], [], [], .1) for fd in res[0]: fd.read() for fd in res[0]: fd.read(drain=1) return out.lines, err.lines - def run(self, dryrun, kmemleak, kmsg): + def run(self, options, dryrun=None, kmemleak=None): """ This is the main function that runs each individual test. Determine whether or not the command requires sudo, and modify it if needed. Run the command, and update the result object. """ + if dryrun is None: + dryrun = options.dryrun if dryrun is True: print(self) return + if kmemleak is None: + kmemleak = options.kmemleak privcmd = self.update_cmd_privs(self.pathname, self.user) try: old = os.umask(0) if not os.path.isdir(self.outputdir): os.makedirs(self.outputdir, mode=0o777) os.umask(old) except OSError as e: fail('%s' % e) """ Log each test we run to /dev/kmsg (on Linux), so if there's a kernel warning we'll be able to match it up to a particular test. """ - if kmsg is True and exists("/dev/kmsg"): + if options.kmsg is True and exists("/dev/kmsg"): try: kp = Popen([SUDO, "sh", "-c", f"echo ZTS run {self.pathname} > /dev/kmsg"]) kp.wait() except Exception: pass self.result.starttime = monotonic_time() if kmemleak: cmd = f'{SUDO} sh -c "echo clear > {KMEMLEAK_FILE}"' check_output(cmd, shell=True) proc = Popen(privcmd, stdout=PIPE, stderr=PIPE) # Allow a special timeout value of 0 to mean infinity if int(self.timeout) == 0: self.timeout = sys.maxsize / (10 ** 9) - t = Timer(int(self.timeout), self.kill_cmd, [proc]) + t = Timer( + int(self.timeout), self.kill_cmd, [proc, options, kmemleak] + ) try: t.start() self.result.stdout, self.result.stderr = self.collect_output(proc) if kmemleak: cmd = f'{SUDO} sh -c "echo scan > {KMEMLEAK_FILE}"' check_output(cmd, shell=True) cmd = f'{SUDO} cat {KMEMLEAK_FILE}' self.result.kmemleak = check_output(cmd, shell=True) except KeyboardInterrupt: - self.kill_cmd(proc, True) + self.kill_cmd(proc, options, kmemleak, True) fail('\nRun terminated at user request.') finally: t.cancel() if self.reran is not False: self.result.done(proc, self.killed, self.reran) def skip(self): """ Initialize enough of the test result that we can log a skipped command. """ Result.total += 1 Result.runresults['SKIP'] += 1 self.result.stdout = self.result.stderr = [] self.result.starttime = monotonic_time() m, s = divmod(monotonic_time() - self.result.starttime, 60) self.result.runtime = '%02d:%02d' % (m, s) self.result.result = 'SKIP' def log(self, options, suppress_console=False): """ This function is responsible for writing all output. This includes the console output, the logfile of all results (with timestamped merged stdout and stderr), and for each test, the unmodified stdout/stderr/merged in its own file. """ logname = getpwuid(os.getuid()).pw_name rer = '' if self.reran is True: rer = ' (RERAN)' user = ' (run as %s)' % (self.user if len(self.user) else logname) if self.identifier: msga = 'Test (%s): %s%s ' % (self.identifier, self.pathname, user) else: msga = 'Test: %s%s ' % (self.pathname, user) msgb = '[%s] [%s]%s\n' % (self.result.runtime, self.result.result, rer) pad = ' ' * (80 - (len(msga) + len(msgb))) result_line = msga + pad + msgb # The result line is always written to the log file. If -q was # specified only failures are written to the console, otherwise # the result line is written to the console. The console output # may be suppressed by calling log() with suppress_console=True. write_log(bytearray(result_line, encoding='utf-8'), LOG_FILE) if not suppress_console: if not options.quiet: write_log(result_line, LOG_OUT) elif options.quiet and self.result.result != 'PASS': write_log(result_line, LOG_OUT) lines = sorted(self.result.stdout + self.result.stderr, key=lambda x: x[0]) # Write timestamped output (stdout and stderr) to the logfile for dt, line in lines: timestamp = bytearray(dt.strftime("%H:%M:%S.%f ")[:11], encoding='utf-8') write_log(b'%s %s\n' % (timestamp, line), LOG_FILE) # Write the separate stdout/stderr/merged files, if the data exists if len(self.result.stdout): with open(os.path.join(self.outputdir, 'stdout'), 'wb') as out: for _, line in self.result.stdout: os.write(out.fileno(), b'%s\n' % line) if len(self.result.stderr): with open(os.path.join(self.outputdir, 'stderr'), 'wb') as err: for _, line in self.result.stderr: os.write(err.fileno(), b'%s\n' % line) if len(self.result.stdout) and len(self.result.stderr): with open(os.path.join(self.outputdir, 'merged'), 'wb') as merged: for _, line in lines: os.write(merged.fileno(), b'%s\n' % line) if len(self.result.kmemleak): with open(os.path.join(self.outputdir, 'kmemleak'), 'wb') as kmem: kmem.write(self.result.kmemleak) class Test(Cmd): props = ['outputdir', 'timeout', 'user', 'pre', 'pre_user', 'post', 'post_user', 'failsafe', 'failsafe_user', 'tags'] def __init__(self, pathname, pre=None, pre_user=None, post=None, post_user=None, failsafe=None, failsafe_user=None, tags=None, **kwargs): super(Test, self).__init__(pathname, **kwargs) self.pre = pre or '' self.pre_user = pre_user or '' self.post = post or '' self.post_user = post_user or '' self.failsafe = failsafe or '' self.failsafe_user = failsafe_user or '' self.tags = tags or [] def __str__(self): post_user = pre_user = failsafe_user = '' if len(self.pre_user): pre_user = ' (as %s)' % (self.pre_user) if len(self.post_user): post_user = ' (as %s)' % (self.post_user) if len(self.failsafe_user): failsafe_user = ' (as %s)' % (self.failsafe_user) return '''\ Pathname: %s Identifier: %s Outputdir: %s Timeout: %d User: %s Pre: %s%s Post: %s%s Failsafe: %s%s Tags: %s ''' % (self.pathname, self.identifier, self.outputdir, self.timeout, self.user, self.pre, pre_user, self.post, post_user, self.failsafe, failsafe_user, self.tags) def verify(self): """ Check the pre/post/failsafe scripts, user and Test. Omit the Test from this run if there are any problems. """ files = [self.pre, self.pathname, self.post, self.failsafe] users = [self.pre_user, self.user, self.post_user, self.failsafe_user] for f in [f for f in files if len(f)]: if not verify_file(f): write_log("Warning: Test '%s' not added to this run because" " it failed verification.\n" % f, LOG_ERR) return False for user in [user for user in users if len(user)]: if not verify_user(user): write_log("Not adding Test '%s' to this run.\n" % self.pathname, LOG_ERR) return False return True - def run(self, options): + def run(self, options, dryrun=None, kmemleak=None): """ Create Cmd instances for the pre/post/failsafe scripts. If the pre script doesn't pass, skip this Test. Run the post script regardless. If the Test is killed, also run the failsafe script. """ odir = os.path.join(self.outputdir, os.path.basename(self.pre)) pretest = Cmd(self.pre, identifier=self.identifier, outputdir=odir, timeout=self.timeout, user=self.pre_user) test = Cmd(self.pathname, identifier=self.identifier, outputdir=self.outputdir, timeout=self.timeout, user=self.user) odir = os.path.join(self.outputdir, os.path.basename(self.failsafe)) failsafe = Cmd(self.failsafe, identifier=self.identifier, outputdir=odir, timeout=self.timeout, user=self.failsafe_user) odir = os.path.join(self.outputdir, os.path.basename(self.post)) posttest = Cmd(self.post, identifier=self.identifier, outputdir=odir, timeout=self.timeout, user=self.post_user) cont = True if len(pretest.pathname): - pretest.run(options.dryrun, False, options.kmsg) + pretest.run(options, kmemleak=False) cont = pretest.result.result == 'PASS' pretest.log(options) if cont: - test.run(options.dryrun, options.kmemleak, options.kmsg) + test.run(options, kmemleak=kmemleak) if test.result.result == 'KILLED' and len(failsafe.pathname): - failsafe.run(options.dryrun, False, options.kmsg) + failsafe.run(options, kmemleak=False) failsafe.log(options, suppress_console=True) else: test.skip() test.log(options) if len(posttest.pathname): - posttest.run(options.dryrun, False, options.kmsg) + posttest.run(options, kmemleak=False) posttest.log(options) class TestGroup(Test): props = Test.props + ['tests'] def __init__(self, pathname, tests=None, **kwargs): super(TestGroup, self).__init__(pathname, **kwargs) self.tests = tests or [] def __str__(self): post_user = pre_user = failsafe_user = '' if len(self.pre_user): pre_user = ' (as %s)' % (self.pre_user) if len(self.post_user): post_user = ' (as %s)' % (self.post_user) if len(self.failsafe_user): failsafe_user = ' (as %s)' % (self.failsafe_user) return '''\ Pathname: %s Identifier: %s Outputdir: %s Tests: %s Timeout: %s User: %s Pre: %s%s Post: %s%s Failsafe: %s%s Tags: %s ''' % (self.pathname, self.identifier, self.outputdir, self.tests, self.timeout, self.user, self.pre, pre_user, self.post, post_user, self.failsafe, failsafe_user, self.tags) def filter(self, keeplist): self.tests = [x for x in self.tests if x in keeplist] def verify(self): """ Check the pre/post/failsafe scripts, user and tests in this TestGroup. Omit the TestGroup entirely, or simply delete the relevant tests in the group, if that's all that's required. """ # If the pre/post/failsafe scripts are relative pathnames, convert to # absolute, so they stand a chance of passing verification. if len(self.pre) and not os.path.isabs(self.pre): self.pre = os.path.join(self.pathname, self.pre) if len(self.post) and not os.path.isabs(self.post): self.post = os.path.join(self.pathname, self.post) if len(self.failsafe) and not os.path.isabs(self.failsafe): self.post = os.path.join(self.pathname, self.post) auxfiles = [self.pre, self.post, self.failsafe] users = [self.pre_user, self.user, self.post_user, self.failsafe_user] for f in [f for f in auxfiles if len(f)]: if f != self.failsafe and self.pathname != os.path.dirname(f): write_log("Warning: TestGroup '%s' not added to this run. " "Auxiliary script '%s' exists in a different " "directory.\n" % (self.pathname, f), LOG_ERR) return False if not verify_file(f): write_log("Warning: TestGroup '%s' not added to this run. " "Auxiliary script '%s' failed verification.\n" % (self.pathname, f), LOG_ERR) return False for user in [user for user in users if len(user)]: if not verify_user(user): write_log("Not adding TestGroup '%s' to this run.\n" % self.pathname, LOG_ERR) return False # If one of the tests is invalid, delete it, log it, and drive on. for test in self.tests: if not verify_file(os.path.join(self.pathname, test)): del self.tests[self.tests.index(test)] write_log("Warning: Test '%s' removed from TestGroup '%s' " "because it failed verification.\n" % (test, self.pathname), LOG_ERR) return len(self.tests) != 0 - def run(self, options): + def run(self, options, dryrun=None, kmemleak=None): """ Create Cmd instances for the pre/post/failsafe scripts. If the pre script doesn't pass, skip all the tests in this TestGroup. Run the post script regardless. Run the failsafe script when a test is killed. """ # tags assigned to this test group also include the test names if options.tags and not set(self.tags).intersection(set(options.tags)): return odir = os.path.join(self.outputdir, os.path.basename(self.pre)) pretest = Cmd(self.pre, outputdir=odir, timeout=self.timeout, user=self.pre_user, identifier=self.identifier) odir = os.path.join(self.outputdir, os.path.basename(self.post)) posttest = Cmd(self.post, outputdir=odir, timeout=self.timeout, user=self.post_user, identifier=self.identifier) cont = True if len(pretest.pathname): - pretest.run(options.dryrun, False, options.kmsg) + pretest.run(options, dryrun=dryrun, kmemleak=False) cont = pretest.result.result == 'PASS' pretest.log(options) for fname in self.tests: odir = os.path.join(self.outputdir, fname) test = Cmd(os.path.join(self.pathname, fname), outputdir=odir, timeout=self.timeout, user=self.user, identifier=self.identifier) odir = os.path.join(odir, os.path.basename(self.failsafe)) failsafe = Cmd(self.failsafe, outputdir=odir, timeout=self.timeout, user=self.failsafe_user, identifier=self.identifier) if cont: - test.run(options.dryrun, options.kmemleak, options.kmsg) + test.run(options, dryrun=dryrun, kmemleak=kmemleak) if test.result.result == 'KILLED' and len(failsafe.pathname): - failsafe.run(options.dryrun, False, options.kmsg) + failsafe.run(options, dryrun=dryrun, kmemleak=False) failsafe.log(options, suppress_console=True) else: test.skip() test.log(options) if len(posttest.pathname): - posttest.run(options.dryrun, False, options.kmsg) + posttest.run(options, dryrun=dryrun, kmemleak=False) posttest.log(options) class TestRun(object): props = ['quiet', 'outputdir'] def __init__(self, options): self.tests = {} self.testgroups = {} self.starttime = time() self.timestamp = datetime.now().strftime('%Y%m%dT%H%M%S') self.outputdir = os.path.join(options.outputdir, self.timestamp) self.setup_logging(options) self.defaults = [ ('outputdir', BASEDIR), ('quiet', False), ('timeout', 60), ('user', ''), ('pre', ''), ('pre_user', ''), ('post', ''), ('post_user', ''), ('failsafe', ''), ('failsafe_user', ''), ('tags', []) ] def __str__(self): s = 'TestRun:\n outputdir: %s\n' % self.outputdir s += 'TESTS:\n' for key in sorted(self.tests.keys()): s += '%s%s' % (self.tests[key].__str__(), '\n') s += 'TESTGROUPS:\n' for key in sorted(self.testgroups.keys()): s += '%s%s' % (self.testgroups[key].__str__(), '\n') return s def addtest(self, pathname, options): """ Create a new Test, and apply any properties that were passed in from the command line. If it passes verification, add it to the TestRun. """ test = Test(pathname) for prop in Test.props: setattr(test, prop, getattr(options, prop)) if test.verify(): self.tests[pathname] = test def addtestgroup(self, dirname, filenames, options): """ Create a new TestGroup, and apply any properties that were passed in from the command line. If it passes verification, add it to the TestRun. """ if dirname not in self.testgroups: testgroup = TestGroup(dirname) for prop in Test.props: setattr(testgroup, prop, getattr(options, prop)) # Prevent pre/post/failsafe scripts from running as regular tests for f in [testgroup.pre, testgroup.post, testgroup.failsafe]: if f in filenames: del filenames[filenames.index(f)] self.testgroups[dirname] = testgroup self.testgroups[dirname].tests = sorted(filenames) testgroup.verify() def filter(self, keeplist): for group in list(self.testgroups.keys()): if group not in keeplist: del self.testgroups[group] continue g = self.testgroups[group] if g.pre and os.path.basename(g.pre) in keeplist[group]: continue g.filter(keeplist[group]) for test in list(self.tests.keys()): directory, base = os.path.split(test) if directory not in keeplist or base not in keeplist[directory]: del self.tests[test] def read(self, options): """ Read in the specified runfiles, and apply the TestRun properties listed in the 'DEFAULT' section to our TestRun. Then read each section, and apply the appropriate properties to the Test or TestGroup. Properties from individual sections override those set in the 'DEFAULT' section. If the Test or TestGroup passes verification, add it to the TestRun. """ config = configparser.RawConfigParser() parsed = config.read(options.runfiles) failed = options.runfiles - set(parsed) if len(failed): files = ' '.join(sorted(failed)) fail("Couldn't read config files: %s" % files) for opt in TestRun.props: if config.has_option('DEFAULT', opt): setattr(self, opt, config.get('DEFAULT', opt)) self.outputdir = os.path.join(self.outputdir, self.timestamp) testdir = options.testdir for section in config.sections(): if 'tests' in config.options(section): parts = section.split(':', 1) sectiondir = parts[0] identifier = parts[1] if len(parts) == 2 else None if os.path.isdir(sectiondir): pathname = sectiondir elif os.path.isdir(os.path.join(testdir, sectiondir)): pathname = os.path.join(testdir, sectiondir) else: pathname = sectiondir testgroup = TestGroup(os.path.abspath(pathname), identifier=identifier) for prop in TestGroup.props: for sect in ['DEFAULT', section]: if config.has_option(sect, prop): if prop == 'tags': setattr(testgroup, prop, eval(config.get(sect, prop))) elif prop == 'failsafe': failsafe = config.get(sect, prop) setattr(testgroup, prop, os.path.join(testdir, failsafe)) else: setattr(testgroup, prop, config.get(sect, prop)) # Repopulate tests using eval to convert the string to a list testgroup.tests = eval(config.get(section, 'tests')) if testgroup.verify(): self.testgroups[section] = testgroup else: test = Test(section) for prop in Test.props: for sect in ['DEFAULT', section]: if config.has_option(sect, prop): if prop == 'failsafe': failsafe = config.get(sect, prop) setattr(test, prop, os.path.join(testdir, failsafe)) else: setattr(test, prop, config.get(sect, prop)) if test.verify(): self.tests[section] = test def write(self, options): """ Create a configuration file for editing and later use. The 'DEFAULT' section of the config file is created from the properties that were specified on the command line. Tests are simply added as sections that inherit everything from the 'DEFAULT' section. TestGroups are the same, except they get an option including all the tests to run in that directory. """ defaults = dict([(prop, getattr(options, prop)) for prop, _ in self.defaults]) config = configparser.RawConfigParser(defaults) for test in sorted(self.tests.keys()): config.add_section(test) for prop in Test.props: if prop not in self.props: config.set(test, prop, getattr(self.tests[test], prop)) for testgroup in sorted(self.testgroups.keys()): config.add_section(testgroup) config.set(testgroup, 'tests', self.testgroups[testgroup].tests) for prop in TestGroup.props: if prop not in self.props: config.set(testgroup, prop, getattr(self.testgroups[testgroup], prop)) try: with open(options.template, 'w') as f: return config.write(f) except IOError: fail('Could not open \'%s\' for writing.' % options.template) def complete_outputdirs(self): """ Collect all the pathnames for Tests, and TestGroups. Work backwards one pathname component at a time, to create a unique directory name in which to deposit test output. Tests will be able to write output files directly in the newly modified outputdir. TestGroups will be able to create one subdirectory per test in the outputdir, and are guaranteed uniqueness because a group can only contain files in one directory. Pre and post tests will create a directory rooted at the outputdir of the Test or TestGroup in question for their output. Failsafe scripts will create a directory rooted at the outputdir of each Test for their output. """ done = False components = 0 tmp_dict = dict(list(self.tests.items()) + list(self.testgroups.items())) total = len(tmp_dict) base = self.outputdir while not done: paths = [] components -= 1 for testfile in list(tmp_dict.keys()): uniq = '/'.join(testfile.split('/')[components:]).lstrip('/') if uniq not in paths: paths.append(uniq) tmp_dict[testfile].outputdir = os.path.join(base, uniq) else: break done = total == len(paths) def setup_logging(self, options): """ This function creates the output directory and gets a file object for the logfile. This function must be called before write_log() can be used. """ if options.dryrun is True: return global LOG_FILE_OBJ if not options.template: try: old = os.umask(0) os.makedirs(self.outputdir, mode=0o777) os.umask(old) filename = os.path.join(self.outputdir, 'log') LOG_FILE_OBJ = open(filename, buffering=0, mode='wb') except OSError as e: fail('%s' % e) def run(self, options): """ Walk through all the Tests and TestGroups, calling run(). """ try: os.chdir(self.outputdir) except OSError: fail('Could not change to directory %s' % self.outputdir) # make a symlink to the output for the currently running test logsymlink = os.path.join(self.outputdir, '../current') if os.path.islink(logsymlink): os.unlink(logsymlink) if not os.path.exists(logsymlink): os.symlink(self.outputdir, logsymlink) else: write_log('Could not make a symlink to directory %s\n' % self.outputdir, LOG_ERR) if options.kmemleak: cmd = f'{SUDO} -c "echo scan=0 > {KMEMLEAK_FILE}"' check_output(cmd, shell=True) iteration = 0 while iteration < options.iterations: for test in sorted(self.tests.keys()): self.tests[test].run(options) for testgroup in sorted(self.testgroups.keys()): self.testgroups[testgroup].run(options) iteration += 1 def summary(self): if Result.total == 0: return 2 print('\nResults Summary') for key in list(Result.runresults.keys()): if Result.runresults[key] != 0: print('%s\t% 4d' % (key, Result.runresults[key])) m, s = divmod(time() - self.starttime, 60) h, m = divmod(m, 60) print('\nRunning Time:\t%02d:%02d:%02d' % (h, m, s)) print('Percent passed:\t%.1f%%' % ((float(Result.runresults['PASS']) / float(Result.total)) * 100)) print('Log directory:\t%s' % self.outputdir) if Result.runresults['FAIL'] > 0: return 1 if Result.runresults['KILLED'] > 0: return 1 if Result.runresults['RERAN'] > 0: return 3 return 0 def write_log(msg, target): """ Write the provided message to standard out, standard error or the logfile. If specifying LOG_FILE, then `msg` must be a bytes like object. This way we can still handle output from tests that may be in unexpected encodings. """ if target == LOG_OUT: os.write(sys.stdout.fileno(), bytearray(msg, encoding='utf-8')) elif target == LOG_ERR: os.write(sys.stderr.fileno(), bytearray(msg, encoding='utf-8')) elif target == LOG_FILE: os.write(LOG_FILE_OBJ.fileno(), msg) else: fail('log_msg called with unknown target "%s"' % target) def verify_file(pathname): """ Verify that the supplied pathname is an executable regular file. """ if os.path.isdir(pathname) or os.path.islink(pathname): return False for ext in '', '.ksh', '.sh': script_path = pathname + ext if os.path.isfile(script_path) and os.access(script_path, os.X_OK): return True return False def verify_user(user): """ Verify that the specified user exists on this system, and can execute sudo without being prompted for a password. """ testcmd = [SUDO, '-n', '-u', user, TRUE] if user in Cmd.verified_users: return True try: getpwnam(user) except KeyError: write_log("Warning: user '%s' does not exist.\n" % user, LOG_ERR) return False p = Popen(testcmd) p.wait() if p.returncode != 0: write_log("Warning: user '%s' cannot use passwordless sudo.\n" % user, LOG_ERR) return False else: Cmd.verified_users.append(user) return True def find_tests(testrun, options): """ For the given list of pathnames, add files as Tests. For directories, if do_groups is True, add the directory as a TestGroup. If False, recursively search for executable files. """ for p in sorted(options.pathnames): if os.path.isdir(p): for dirname, _, filenames in os.walk(p): if options.do_groups: testrun.addtestgroup(dirname, filenames, options) else: for f in sorted(filenames): testrun.addtest(os.path.join(dirname, f), options) else: testrun.addtest(p, options) def filter_tests(testrun, options): try: fh = open(options.logfile, "r") except Exception as e: fail('%s' % e) failed = {} while True: line = fh.readline() if not line: break m = re.match(r'Test: .*(tests/.*)/(\S+).*\[FAIL\]', line) if not m: continue group, test = m.group(1, 2) try: failed[group].append(test) except KeyError: failed[group] = [test] fh.close() testrun.filter(failed) def fail(retstr, ret=1): print('%s: %s' % (sys.argv[0], retstr)) exit(ret) def kmemleak_cb(option, opt_str, value, parser): if not os.path.exists(KMEMLEAK_FILE): fail(f"File '{KMEMLEAK_FILE}' doesn't exist. " + "Enable CONFIG_DEBUG_KMEMLEAK in kernel configuration.") setattr(parser.values, option.dest, True) def options_cb(option, opt_str, value, parser): path_options = ['outputdir', 'template', 'testdir', 'logfile'] if opt_str in parser.rargs: fail('%s may only be specified once.' % opt_str) if option.dest == 'runfiles': parser.values.cmd = 'rdconfig' value = set(os.path.abspath(p) for p in value.split(',')) if option.dest == 'tags': value = [x.strip() for x in value.split(',')] if option.dest in path_options: setattr(parser.values, option.dest, os.path.abspath(value)) else: setattr(parser.values, option.dest, value) def parse_args(): parser = OptionParser() parser.add_option('-c', action='callback', callback=options_cb, type='string', dest='runfiles', metavar='runfiles', help='Specify tests to run via config files.') parser.add_option('-d', action='store_true', default=False, dest='dryrun', help='Dry run. Print tests, but take no other action.') parser.add_option('-l', action='callback', callback=options_cb, default=None, dest='logfile', metavar='logfile', type='string', help='Read logfile and re-run tests which failed.') parser.add_option('-g', action='store_true', default=False, dest='do_groups', help='Make directories TestGroups.') parser.add_option('-o', action='callback', callback=options_cb, default=BASEDIR, dest='outputdir', type='string', metavar='outputdir', help='Specify an output directory.') parser.add_option('-i', action='callback', callback=options_cb, default=TESTDIR, dest='testdir', type='string', metavar='testdir', help='Specify a test directory.') parser.add_option('-K', action='store_true', default=False, dest='kmsg', help='Log tests names to /dev/kmsg') parser.add_option('-m', action='callback', callback=kmemleak_cb, default=False, dest='kmemleak', help='Enable kmemleak reporting (Linux only)') parser.add_option('-p', action='callback', callback=options_cb, default='', dest='pre', metavar='script', type='string', help='Specify a pre script.') parser.add_option('-P', action='callback', callback=options_cb, default='', dest='post', metavar='script', type='string', help='Specify a post script.') parser.add_option('-q', action='store_true', default=False, dest='quiet', help='Silence on the console during a test run.') parser.add_option('-s', action='callback', callback=options_cb, default='', dest='failsafe', metavar='script', type='string', help='Specify a failsafe script.') parser.add_option('-S', action='callback', callback=options_cb, default='', dest='failsafe_user', metavar='failsafe_user', type='string', help='Specify a user to execute the failsafe script.') parser.add_option('-t', action='callback', callback=options_cb, default=60, dest='timeout', metavar='seconds', type='int', help='Timeout (in seconds) for an individual test.') parser.add_option('-u', action='callback', callback=options_cb, default='', dest='user', metavar='user', type='string', help='Specify a different user name to run as.') parser.add_option('-w', action='callback', callback=options_cb, default=None, dest='template', metavar='template', type='string', help='Create a new config file.') parser.add_option('-x', action='callback', callback=options_cb, default='', dest='pre_user', metavar='pre_user', type='string', help='Specify a user to execute the pre script.') parser.add_option('-X', action='callback', callback=options_cb, default='', dest='post_user', metavar='post_user', type='string', help='Specify a user to execute the post script.') parser.add_option('-T', action='callback', callback=options_cb, default='', dest='tags', metavar='tags', type='string', help='Specify tags to execute specific test groups.') parser.add_option('-I', action='callback', callback=options_cb, default=1, dest='iterations', metavar='iterations', type='int', help='Number of times to run the test run.') (options, pathnames) = parser.parse_args() if options.runfiles and len(pathnames): fail('Extraneous arguments.') options.pathnames = [os.path.abspath(path) for path in pathnames] return options def main(): options = parse_args() testrun = TestRun(options) if options.runfiles: testrun.read(options) else: find_tests(testrun, options) if options.logfile: filter_tests(testrun, options) if options.template: testrun.write(options) exit(0) testrun.complete_outputdirs() testrun.run(options) exit(testrun.summary()) if __name__ == '__main__': main() diff --git a/sys/contrib/openzfs/tests/test-runner/bin/zts-report.py.in b/sys/contrib/openzfs/tests/test-runner/bin/zts-report.py.in index 63470bc041c6..3f7498f5c6bf 100755 --- a/sys/contrib/openzfs/tests/test-runner/bin/zts-report.py.in +++ b/sys/contrib/openzfs/tests/test-runner/bin/zts-report.py.in @@ -1,451 +1,452 @@ #!/usr/bin/env @PYTHON_SHEBANG@ # # This file and its contents are supplied under the terms of the # Common Development and Distribution License ("CDDL"), version 1.0. # You may only use this file in accordance with the terms of version # 1.0 of the CDDL. # # A full copy of the text of the CDDL should have accompanied this # source. A copy of the CDDL is also available via the Internet at # http://www.illumos.org/license/CDDL. # # # Copyright (c) 2017 by Delphix. All rights reserved. # Copyright (c) 2018 by Lawrence Livermore National Security, LLC. # # This script must remain compatible with Python 3.6+. # import os import re import sys import argparse # # This script parses the stdout of zfstest, which has this format: # # Test: /path/to/testa (run as root) [00:00] [PASS] # Test: /path/to/testb (run as jkennedy) [00:00] [PASS] # Test: /path/to/testc (run as root) [00:00] [FAIL] # [...many more results...] # # Results Summary # FAIL 22 # SKIP 32 # PASS 1156 # # Running Time: 02:50:31 # Percent passed: 95.5% # Log directory: /var/tmp/test_results/20180615T205926 # # # Common generic reasons for a test or test group to be skipped. # # Some test cases are known to fail in ways which are not harmful or dangerous. # In these cases simply mark the test as a known failure until it can be # updated and the issue resolved. Note that it's preferable to open a unique # issue on the GitHub issue tracker for each test case failure. # known_reason = 'Known issue' # # Some tests require that a test user be able to execute the zfs utilities. # This may not be possible when testing in-tree due to the default permissions # on the user's home directory. When testing this can be resolved by granting # group read access. # # chmod 0750 $HOME # exec_reason = 'Test user execute permissions required for utilities' # # Some tests require that the kernel supports renameat2 syscall. # renameat2_reason = 'Kernel renameat2 support required' # # Some tests require the O_TMPFILE flag which was first introduced in the # 3.11 kernel. # tmpfile_reason = 'Kernel O_TMPFILE support required' # # Some tests require the statx(2) system call on Linux which was first # introduced in the 4.11 kernel. # statx_reason = 'Kernel statx(2) system call required on Linux' # # Some tests require that the lsattr utility support the project id feature. # project_id_reason = 'lsattr with set/show project ID required' # # Some tests require that the kernel support user namespaces. # user_ns_reason = 'Kernel user namespace support required' # # Some rewind tests can fail since nothing guarantees that old MOS blocks # are not overwritten. Snapshots protect datasets and data files but not # the MOS. Reasonable efforts are made in the test case to increase the # odds that some txgs will have their MOS data left untouched, but it is # never a sure thing. # rewind_reason = 'Arbitrary pool rewind is not guaranteed' # # Some tests require a minimum version of the fio benchmark utility. # Older distributions such as CentOS 6.x only provide fio-2.0.13. # fio_reason = 'Fio v2.3 or newer required' # # Some tests require that the DISKS provided support the discard operation. # Normally this is not an issue because loop back devices are used for DISKS # and they support discard (TRIM/UNMAP). # trim_reason = 'DISKS must support discard (TRIM/UNMAP)' # # Some tests on FreeBSD require the fspacectl(2) system call and the # truncate(1) utility supporting the -d option. The system call was first # introduced in FreeBSD version 1400032. # fspacectl_reason = 'fspacectl(2) and truncate -d support required' # # Some tests are not applicable to a platform or need to be updated to operate # in the manor required by the platform. Any tests which are skipped for this # reason will be suppressed in the final analysis output. # na_reason = "Not applicable" # # Some test cases doesn't have all requirements to run on Github actions CI. # ci_reason = 'CI runner doesn\'t have all requirements' # # Idmapped mount is only supported in kernel version >= 5.12 # idmap_reason = 'Idmapped mount needs kernel 5.12+' # # These tests are known to fail, thus we use this list to prevent these # failures from failing the job as a whole; only unexpected failures # bubble up to cause this script to exit with a non-zero exit status. # # Format: { 'test-name': ['expected result', 'issue-number | reason'] } # # For each known failure it is recommended to link to a GitHub issue by # setting the reason to the issue number. Alternately, one of the generic # reasons listed above can be used. # known = { 'casenorm/mixed_none_lookup_ci': ['FAIL', 7633], 'casenorm/mixed_formd_lookup_ci': ['FAIL', 7633], 'cli_root/zpool_import/import_rewind_device_replaced': ['FAIL', rewind_reason], 'cli_user/misc/zfs_share_001_neg': ['SKIP', na_reason], 'cli_user/misc/zfs_unshare_001_neg': ['SKIP', na_reason], 'privilege/setup': ['SKIP', na_reason], 'refreserv/refreserv_004_pos': ['FAIL', known_reason], 'rootpool/setup': ['SKIP', na_reason], 'rsend/rsend_008_pos': ['SKIP', 6066], 'vdev_zaps/vdev_zaps_007_pos': ['FAIL', known_reason], } if sys.platform.startswith('freebsd'): known.update({ 'cli_root/zfs_receive/receive-o-x_props_override': ['FAIL', known_reason], 'cli_root/zpool_wait/zpool_wait_trim_basic': ['SKIP', trim_reason], 'cli_root/zpool_wait/zpool_wait_trim_cancel': ['SKIP', trim_reason], 'cli_root/zpool_wait/zpool_wait_trim_flag': ['SKIP', trim_reason], 'cli_root/zfs_unshare/zfs_unshare_008_pos': ['SKIP', na_reason], 'link_count/link_count_001': ['SKIP', na_reason], 'casenorm/mixed_create_failure': ['FAIL', 13215], 'mmap/mmap_sync_001_pos': ['SKIP', na_reason], }) elif sys.platform.startswith('linux'): known.update({ 'casenorm/mixed_formd_lookup': ['FAIL', 7633], 'casenorm/mixed_formd_delete': ['FAIL', 7633], 'casenorm/sensitive_formd_lookup': ['FAIL', 7633], 'casenorm/sensitive_formd_delete': ['FAIL', 7633], 'removal/removal_with_zdb': ['SKIP', known_reason], 'cli_root/zfs_unshare/zfs_unshare_002_pos': ['SKIP', na_reason], }) # # These tests may occasionally fail or be skipped. We want there failures # to be reported but only unexpected failures should bubble up to cause # this script to exit with a non-zero exit status. # # Format: { 'test-name': ['expected result', 'issue-number | reason'] } # # For each known failure it is recommended to link to a GitHub issue by # setting the reason to the issue number. Alternately, one of the generic # reasons listed above can be used. # maybe = { 'append/threadsappend_001_pos': ['FAIL', 6136], 'chattr/setup': ['SKIP', exec_reason], 'crtime/crtime_001_pos': ['SKIP', statx_reason], 'cli_root/zdb/zdb_006_pos': ['FAIL', known_reason], 'cli_root/zfs_destroy/zfs_destroy_dev_removal_condense': ['FAIL', known_reason], 'cli_root/zfs_get/zfs_get_004_pos': ['FAIL', known_reason], 'cli_root/zfs_get/zfs_get_009_pos': ['SKIP', 5479], 'cli_root/zfs_rollback/zfs_rollback_001_pos': ['FAIL', known_reason], 'cli_root/zfs_rollback/zfs_rollback_002_pos': ['FAIL', known_reason], 'cli_root/zfs_snapshot/zfs_snapshot_002_neg': ['FAIL', known_reason], 'cli_root/zfs_unshare/zfs_unshare_006_pos': ['SKIP', na_reason], 'cli_root/zpool_add/zpool_add_004_pos': ['FAIL', known_reason], 'cli_root/zpool_destroy/zpool_destroy_001_pos': ['SKIP', 6145], 'cli_root/zpool_import/zpool_import_missing_003_pos': ['SKIP', 6839], 'cli_root/zpool_initialize/zpool_initialize_import_export': ['FAIL', 11948], 'cli_root/zpool_labelclear/zpool_labelclear_removed': ['FAIL', known_reason], 'cli_root/zpool_trim/setup': ['SKIP', trim_reason], 'cli_root/zpool_upgrade/zpool_upgrade_004_pos': ['FAIL', 6141], 'delegate/setup': ['SKIP', exec_reason], 'fallocate/fallocate_punch-hole': ['SKIP', fspacectl_reason], 'history/history_004_pos': ['FAIL', 7026], 'history/history_005_neg': ['FAIL', 6680], 'history/history_006_neg': ['FAIL', 5657], 'history/history_008_pos': ['FAIL', known_reason], 'history/history_010_pos': ['SKIP', exec_reason], 'io/mmap': ['SKIP', fio_reason], 'largest_pool/largest_pool_001_pos': ['FAIL', known_reason], 'mmp/mmp_on_uberblocks': ['FAIL', known_reason], 'pam/setup': ['SKIP', "pamtester might be not available"], 'pool_checkpoint/checkpoint_discard_busy': ['FAIL', 11946], 'projectquota/setup': ['SKIP', exec_reason], 'removal/removal_condense_export': ['FAIL', known_reason], 'renameat2/setup': ['SKIP', renameat2_reason], 'reservation/reservation_008_pos': ['FAIL', 7741], 'reservation/reservation_018_pos': ['FAIL', 5642], 'snapshot/clone_001_pos': ['FAIL', known_reason], 'snapshot/snapshot_009_pos': ['FAIL', 7961], 'snapshot/snapshot_010_pos': ['FAIL', 7961], 'snapused/snapused_004_pos': ['FAIL', 5513], 'tmpfile/setup': ['SKIP', tmpfile_reason], 'trim/setup': ['SKIP', trim_reason], 'upgrade/upgrade_projectquota_001_pos': ['SKIP', project_id_reason], 'user_namespace/setup': ['SKIP', user_ns_reason], 'userquota/setup': ['SKIP', exec_reason], 'vdev_zaps/vdev_zaps_004_pos': ['FAIL', known_reason], 'zvol/zvol_ENOSPC/zvol_ENOSPC_001_pos': ['FAIL', 5848], } if sys.platform.startswith('freebsd'): maybe.update({ 'cli_root/zfs_copies/zfs_copies_002_pos': ['FAIL', known_reason], 'cli_root/zfs_inherit/zfs_inherit_001_neg': ['FAIL', known_reason], 'cli_root/zfs_share/zfs_share_concurrent_shares': ['FAIL', known_reason], 'cli_root/zpool_import/zpool_import_012_pos': ['FAIL', known_reason], 'delegate/zfs_allow_003_pos': ['FAIL', known_reason], 'inheritance/inherit_001_pos': ['FAIL', 11829], 'resilver/resilver_restart_001': ['FAIL', known_reason], 'pool_checkpoint/checkpoint_big_rewind': ['FAIL', 12622], 'pool_checkpoint/checkpoint_indirect': ['FAIL', 12623], 'snapshot/snapshot_002_pos': ['FAIL', '14831'], }) elif sys.platform.startswith('linux'): maybe.update({ 'cli_root/zfs_rename/zfs_rename_002_pos': ['FAIL', known_reason], 'cli_root/zpool_reopen/zpool_reopen_003_pos': ['FAIL', known_reason], 'fault/auto_online_002_pos': ['FAIL', 11889], + 'fault/auto_replace_001_pos': ['FAIL', 14851], 'fault/auto_spare_002_pos': ['FAIL', 11889], 'fault/auto_spare_multiple': ['FAIL', 11889], 'fault/auto_spare_shared': ['FAIL', 11889], 'fault/decompress_fault': ['FAIL', 11889], 'io/io_uring': ['SKIP', 'io_uring support required'], 'limits/filesystem_limit': ['SKIP', known_reason], 'limits/snapshot_limit': ['SKIP', known_reason], 'mmp/mmp_active_import': ['FAIL', known_reason], 'mmp/mmp_exported_import': ['FAIL', known_reason], 'mmp/mmp_inactive_import': ['FAIL', known_reason], 'zvol/zvol_misc/zvol_misc_snapdev': ['FAIL', 12621], 'zvol/zvol_misc/zvol_misc_volmode': ['FAIL', known_reason], 'idmap_mount/idmap_mount_001': ['SKIP', idmap_reason], 'idmap_mount/idmap_mount_002': ['SKIP', idmap_reason], 'idmap_mount/idmap_mount_003': ['SKIP', idmap_reason], 'idmap_mount/idmap_mount_004': ['SKIP', idmap_reason], 'idmap_mount/idmap_mount_005': ['SKIP', idmap_reason], }) # Not all Github actions runners have scsi_debug module, so we may skip # some tests which use it. if os.environ.get('CI') == 'true': known.update({ 'cli_root/zpool_expand/zpool_expand_001_pos': ['SKIP', ci_reason], 'cli_root/zpool_expand/zpool_expand_003_neg': ['SKIP', ci_reason], 'cli_root/zpool_expand/zpool_expand_005_pos': ['SKIP', ci_reason], 'cli_root/zpool_reopen/setup': ['SKIP', ci_reason], 'cli_root/zpool_reopen/zpool_reopen_001_pos': ['SKIP', ci_reason], 'cli_root/zpool_reopen/zpool_reopen_002_pos': ['SKIP', ci_reason], 'cli_root/zpool_reopen/zpool_reopen_003_pos': ['SKIP', ci_reason], 'cli_root/zpool_reopen/zpool_reopen_004_pos': ['SKIP', ci_reason], 'cli_root/zpool_reopen/zpool_reopen_005_pos': ['SKIP', ci_reason], 'cli_root/zpool_reopen/zpool_reopen_006_neg': ['SKIP', ci_reason], 'cli_root/zpool_reopen/zpool_reopen_007_pos': ['SKIP', ci_reason], 'cli_root/zpool_split/zpool_split_wholedisk': ['SKIP', ci_reason], 'fault/auto_offline_001_pos': ['SKIP', ci_reason], 'fault/auto_online_001_pos': ['SKIP', ci_reason], 'fault/auto_online_002_pos': ['SKIP', ci_reason], 'fault/auto_replace_001_pos': ['SKIP', ci_reason], 'fault/auto_spare_ashift': ['SKIP', ci_reason], 'fault/auto_spare_shared': ['SKIP', ci_reason], 'procfs/pool_state': ['SKIP', ci_reason], }) maybe.update({ 'events/events_002_pos': ['FAIL', 11546], }) def process_results(pathname): try: f = open(pathname) except IOError as e: print('Error opening file:', e) sys.exit(1) prefix = '/zfs-tests/tests/(?:functional|perf/regression)/' pattern = \ r'^Test(?:\s+\(\S+\))?:' + \ rf'\s*\S*{prefix}(\S+)' + \ r'\s*\(run as (\S+)\)\s*\[(\S+)\]\s*\[(\S+)\]' pattern_log = r'^\s*Log directory:\s*(\S*)' d = {} logdir = 'Could not determine log directory.' for line in f.readlines(): m = re.match(pattern, line) if m and len(m.groups()) == 4: d[m.group(1)] = m.group(4) continue m = re.match(pattern_log, line) if m: logdir = m.group(1) return d, logdir class ListMaybesAction(argparse.Action): def __init__(self, option_strings, dest="SUPPRESS", default="SUPPRESS", help="list flaky tests and exit"): super(ListMaybesAction, self).__init__( option_strings=option_strings, dest=dest, default=default, nargs=0, help=help) def __call__(self, parser, namespace, values, option_string=None): for test in maybe: print(test) sys.exit(0) if __name__ == "__main__": parser = argparse.ArgumentParser(description='Analyze ZTS logs') parser.add_argument('logfile') parser.add_argument('--list-maybes', action=ListMaybesAction) parser.add_argument('--no-maybes', action='store_false', dest='maybes') args = parser.parse_args() results, logdir = process_results(args.logfile) if not results: print("\n\nNo test results were found.") print("Log directory:", logdir) sys.exit(0) expected = [] unexpected = [] all_maybes = True for test in list(results.keys()): if results[test] == "PASS": continue setup = test.replace(os.path.basename(test), "setup") if results[test] == "SKIP" and test != setup: if setup in known and known[setup][0] == "SKIP": continue if setup in maybe and maybe[setup][0] == "SKIP": continue if (test in known and results[test] in known[test][0]): expected.append(test) elif test in maybe and results[test] in maybe[test][0]: if results[test] == 'SKIP' or args.maybes: expected.append(test) elif not args.maybes: unexpected.append(test) else: unexpected.append(test) all_maybes = False print("\nTests with results other than PASS that are expected:") for test in sorted(expected): issue_url = 'https://github.com/openzfs/zfs/issues/' # Include the reason why the result is expected, given the following: # 1. Suppress test results which set the "Not applicable" reason. # 2. Numerical reasons are assumed to be GitHub issue numbers. # 3. When an entire test group is skipped only report the setup reason. if test in known: if known[test][1] == na_reason: continue elif isinstance(known[test][1], int): expect = f"{issue_url}{known[test][1]}" else: expect = known[test][1] elif test in maybe: if isinstance(maybe[test][1], int): expect = f"{issue_url}{maybe[test][1]}" else: expect = maybe[test][1] elif setup in known and known[setup][0] == "SKIP" and setup != test: continue elif setup in maybe and maybe[setup][0] == "SKIP" and setup != test: continue else: expect = "UNKNOWN REASON" print(f" {results[test]} {test} ({expect})") print("\nTests with result of PASS that are unexpected:") for test in sorted(known.keys()): # We probably should not be silently ignoring the case # where "test" is not in "results". if test not in results or results[test] != "PASS": continue print(f" {results[test]} {test} (expected {known[test][0]})") print("\nTests with results other than PASS that are unexpected:") for test in sorted(unexpected): expect = "PASS" if test not in known else known[test][0] print(f" {results[test]} {test} (expected {expect})") if len(unexpected) == 0: sys.exit(0) elif not args.maybes and all_maybes: sys.exit(2) else: sys.exit(1) diff --git a/sys/contrib/openzfs/tests/zfs-tests/cmd/libzfs_input_check.c b/sys/contrib/openzfs/tests/zfs-tests/cmd/libzfs_input_check.c index a1dfaefd7105..c661718a296c 100644 --- a/sys/contrib/openzfs/tests/zfs-tests/cmd/libzfs_input_check.c +++ b/sys/contrib/openzfs/tests/zfs-tests/cmd/libzfs_input_check.c @@ -1,1067 +1,1082 @@ /* * CDDL HEADER START * * This file and its contents are supplied under the terms of the * Common Development and Distribution License ("CDDL"), version 1.0. * You may only use this file in accordance with the terms of version * 1.0 of the CDDL. * * A full copy of the text of the CDDL should have accompanied this * source. A copy of the CDDL is also available via the Internet at * http://www.illumos.org/license/CDDL. * * CDDL HEADER END */ /* * Copyright (c) 2018 by Delphix. All rights reserved. */ #include #include #include #include #include #include #include #include #include +#include /* * Test the nvpair inputs for the non-legacy zfs ioctl commands. */ static boolean_t unexpected_failures; static int zfs_fd; static const char *active_test; /* * Tracks which zfs_ioc_t commands were tested */ static boolean_t ioc_tested[ZFS_IOC_LAST - ZFS_IOC_FIRST]; /* * Legacy ioctls that are skipped (for now) */ static const zfs_ioc_t ioc_skip[] = { ZFS_IOC_POOL_CREATE, ZFS_IOC_POOL_DESTROY, ZFS_IOC_POOL_IMPORT, ZFS_IOC_POOL_EXPORT, ZFS_IOC_POOL_CONFIGS, ZFS_IOC_POOL_STATS, ZFS_IOC_POOL_TRYIMPORT, ZFS_IOC_POOL_SCAN, ZFS_IOC_POOL_FREEZE, ZFS_IOC_POOL_UPGRADE, ZFS_IOC_POOL_GET_HISTORY, ZFS_IOC_VDEV_ADD, ZFS_IOC_VDEV_REMOVE, ZFS_IOC_VDEV_SET_STATE, ZFS_IOC_VDEV_ATTACH, ZFS_IOC_VDEV_DETACH, ZFS_IOC_VDEV_SETPATH, ZFS_IOC_VDEV_SETFRU, ZFS_IOC_OBJSET_STATS, ZFS_IOC_OBJSET_ZPLPROPS, ZFS_IOC_DATASET_LIST_NEXT, ZFS_IOC_SNAPSHOT_LIST_NEXT, ZFS_IOC_SET_PROP, ZFS_IOC_DESTROY, ZFS_IOC_RENAME, ZFS_IOC_RECV, ZFS_IOC_SEND, ZFS_IOC_INJECT_FAULT, ZFS_IOC_CLEAR_FAULT, ZFS_IOC_INJECT_LIST_NEXT, ZFS_IOC_ERROR_LOG, ZFS_IOC_CLEAR, ZFS_IOC_PROMOTE, ZFS_IOC_DSOBJ_TO_DSNAME, ZFS_IOC_OBJ_TO_PATH, ZFS_IOC_POOL_SET_PROPS, ZFS_IOC_POOL_GET_PROPS, ZFS_IOC_SET_FSACL, ZFS_IOC_GET_FSACL, ZFS_IOC_SHARE, ZFS_IOC_INHERIT_PROP, ZFS_IOC_SMB_ACL, ZFS_IOC_USERSPACE_ONE, ZFS_IOC_USERSPACE_MANY, ZFS_IOC_USERSPACE_UPGRADE, ZFS_IOC_OBJSET_RECVD_PROPS, ZFS_IOC_VDEV_SPLIT, ZFS_IOC_NEXT_OBJ, ZFS_IOC_DIFF, ZFS_IOC_TMP_SNAPSHOT, ZFS_IOC_OBJ_TO_STATS, ZFS_IOC_SPACE_WRITTEN, ZFS_IOC_POOL_REGUID, ZFS_IOC_SEND_PROGRESS, ZFS_IOC_EVENTS_NEXT, ZFS_IOC_EVENTS_CLEAR, ZFS_IOC_EVENTS_SEEK, ZFS_IOC_NEXTBOOT, ZFS_IOC_JAIL, ZFS_IOC_UNJAIL, }; #define IOC_INPUT_TEST(ioc, name, req, opt, err) \ IOC_INPUT_TEST_IMPL(ioc, name, req, opt, err, B_FALSE) #define IOC_INPUT_TEST_WILD(ioc, name, req, opt, err) \ IOC_INPUT_TEST_IMPL(ioc, name, req, opt, err, B_TRUE) #define IOC_INPUT_TEST_IMPL(ioc, name, req, opt, err, wild) \ do { \ active_test = __func__ + 5; \ ioc_tested[ioc - ZFS_IOC_FIRST] = B_TRUE; \ lzc_ioctl_test(ioc, name, req, opt, err, wild); \ } while (0) /* * run a zfs ioctl command, verify expected results and log failures */ static void lzc_ioctl_run(zfs_ioc_t ioc, const char *name, nvlist_t *innvl, int expected) { zfs_cmd_t zc = {"\0"}; char *packed = NULL; const char *variant; size_t size = 0; int error = 0; switch (expected) { case ZFS_ERR_IOC_ARG_UNAVAIL: variant = "unsupported input"; break; case ZFS_ERR_IOC_ARG_REQUIRED: variant = "missing input"; break; case ZFS_ERR_IOC_ARG_BADTYPE: variant = "invalid input type"; break; default: variant = "valid input"; break; } packed = fnvlist_pack(innvl, &size); (void) strlcpy(zc.zc_name, name, sizeof (zc.zc_name)); zc.zc_name[sizeof (zc.zc_name) - 1] = '\0'; zc.zc_nvlist_src = (uint64_t)(uintptr_t)packed; zc.zc_nvlist_src_size = size; zc.zc_nvlist_dst_size = MAX(size * 2, 128 * 1024); zc.zc_nvlist_dst = (uint64_t)(uintptr_t)malloc(zc.zc_nvlist_dst_size); if (lzc_ioctl_fd(zfs_fd, ioc, &zc) != 0) error = errno; if (error != expected) { unexpected_failures = B_TRUE; (void) fprintf(stderr, "%s: Unexpected result with %s, " "error %d (expecting %d)\n", active_test, variant, error, expected); } fnvlist_pack_free(packed, size); free((void *)(uintptr_t)zc.zc_nvlist_dst); } /* * Test each ioc for the following ioctl input errors: * ZFS_ERR_IOC_ARG_UNAVAIL an input argument is not supported by kernel * ZFS_ERR_IOC_ARG_REQUIRED a required input argument is missing * ZFS_ERR_IOC_ARG_BADTYPE an input argument has an invalid type */ static int lzc_ioctl_test(zfs_ioc_t ioc, const char *name, nvlist_t *required, nvlist_t *optional, int expected_error, boolean_t wildcard) { nvlist_t *input = fnvlist_alloc(); nvlist_t *future = fnvlist_alloc(); int error = 0; if (required != NULL) { for (nvpair_t *pair = nvlist_next_nvpair(required, NULL); pair != NULL; pair = nvlist_next_nvpair(required, pair)) { fnvlist_add_nvpair(input, pair); } } if (optional != NULL) { for (nvpair_t *pair = nvlist_next_nvpair(optional, NULL); pair != NULL; pair = nvlist_next_nvpair(optional, pair)) { fnvlist_add_nvpair(input, pair); } } /* * Generic input run with 'optional' nvlist pair */ if (!wildcard) fnvlist_add_nvlist(input, "optional", future); lzc_ioctl_run(ioc, name, input, expected_error); if (!wildcard) fnvlist_remove(input, "optional"); /* * Bogus input value */ if (!wildcard) { fnvlist_add_string(input, "bogus_input", "bogus"); lzc_ioctl_run(ioc, name, input, ZFS_ERR_IOC_ARG_UNAVAIL); fnvlist_remove(input, "bogus_input"); } /* * Missing required inputs */ if (required != NULL) { nvlist_t *empty = fnvlist_alloc(); lzc_ioctl_run(ioc, name, empty, ZFS_ERR_IOC_ARG_REQUIRED); nvlist_free(empty); } /* * Wrong nvpair type */ if (required != NULL || optional != NULL) { /* * switch the type of one of the input pairs */ for (nvpair_t *pair = nvlist_next_nvpair(input, NULL); pair != NULL; pair = nvlist_next_nvpair(input, pair)) { char pname[MAXNAMELEN]; data_type_t ptype; strlcpy(pname, nvpair_name(pair), sizeof (pname)); pname[sizeof (pname) - 1] = '\0'; ptype = nvpair_type(pair); fnvlist_remove_nvpair(input, pair); switch (ptype) { case DATA_TYPE_STRING: fnvlist_add_uint64(input, pname, 42); break; default: fnvlist_add_string(input, pname, "bogus"); break; } } lzc_ioctl_run(ioc, name, input, ZFS_ERR_IOC_ARG_BADTYPE); } nvlist_free(future); nvlist_free(input); return (error); } static void test_pool_sync(const char *pool) { nvlist_t *required = fnvlist_alloc(); fnvlist_add_boolean_value(required, "force", B_TRUE); IOC_INPUT_TEST(ZFS_IOC_POOL_SYNC, pool, required, NULL, 0); nvlist_free(required); } static void test_pool_reopen(const char *pool) { nvlist_t *optional = fnvlist_alloc(); fnvlist_add_boolean_value(optional, "scrub_restart", B_FALSE); IOC_INPUT_TEST(ZFS_IOC_POOL_REOPEN, pool, NULL, optional, 0); nvlist_free(optional); } static void test_pool_checkpoint(const char *pool) { IOC_INPUT_TEST(ZFS_IOC_POOL_CHECKPOINT, pool, NULL, NULL, 0); } static void test_pool_discard_checkpoint(const char *pool) { int err = lzc_pool_checkpoint(pool); if (err == 0 || err == ZFS_ERR_CHECKPOINT_EXISTS) IOC_INPUT_TEST(ZFS_IOC_POOL_DISCARD_CHECKPOINT, pool, NULL, NULL, 0); } static void test_log_history(const char *pool) { nvlist_t *required = fnvlist_alloc(); fnvlist_add_string(required, "message", "input check"); IOC_INPUT_TEST(ZFS_IOC_LOG_HISTORY, pool, required, NULL, 0); nvlist_free(required); } static void test_create(const char *pool) { char dataset[MAXNAMELEN + 32]; (void) snprintf(dataset, sizeof (dataset), "%s/create-fs", pool); nvlist_t *required = fnvlist_alloc(); nvlist_t *optional = fnvlist_alloc(); nvlist_t *props = fnvlist_alloc(); fnvlist_add_int32(required, "type", DMU_OST_ZFS); fnvlist_add_uint64(props, "recordsize", 8192); fnvlist_add_nvlist(optional, "props", props); IOC_INPUT_TEST(ZFS_IOC_CREATE, dataset, required, optional, 0); nvlist_free(required); nvlist_free(optional); } static void test_snapshot(const char *pool, const char *snapshot) { nvlist_t *required = fnvlist_alloc(); nvlist_t *optional = fnvlist_alloc(); nvlist_t *snaps = fnvlist_alloc(); nvlist_t *props = fnvlist_alloc(); fnvlist_add_boolean(snaps, snapshot); fnvlist_add_nvlist(required, "snaps", snaps); fnvlist_add_string(props, "org.openzfs:launch", "September 17th, 2013"); fnvlist_add_nvlist(optional, "props", props); IOC_INPUT_TEST(ZFS_IOC_SNAPSHOT, pool, required, optional, 0); nvlist_free(props); nvlist_free(snaps); nvlist_free(optional); nvlist_free(required); } static void test_space_snaps(const char *snapshot) { nvlist_t *required = fnvlist_alloc(); fnvlist_add_string(required, "firstsnap", snapshot); IOC_INPUT_TEST(ZFS_IOC_SPACE_SNAPS, snapshot, required, NULL, 0); nvlist_free(required); } static void test_destroy_snaps(const char *pool, const char *snapshot) { nvlist_t *required = fnvlist_alloc(); nvlist_t *snaps = fnvlist_alloc(); fnvlist_add_boolean(snaps, snapshot); fnvlist_add_nvlist(required, "snaps", snaps); IOC_INPUT_TEST(ZFS_IOC_DESTROY_SNAPS, pool, required, NULL, 0); nvlist_free(snaps); nvlist_free(required); } static void test_bookmark(const char *pool, const char *snapshot, const char *bookmark) { nvlist_t *required = fnvlist_alloc(); fnvlist_add_string(required, bookmark, snapshot); IOC_INPUT_TEST_WILD(ZFS_IOC_BOOKMARK, pool, required, NULL, 0); nvlist_free(required); } static void test_get_bookmarks(const char *dataset) { nvlist_t *optional = fnvlist_alloc(); fnvlist_add_boolean(optional, "guid"); fnvlist_add_boolean(optional, "createtxg"); fnvlist_add_boolean(optional, "creation"); IOC_INPUT_TEST_WILD(ZFS_IOC_GET_BOOKMARKS, dataset, NULL, optional, 0); nvlist_free(optional); } static void test_destroy_bookmarks(const char *pool, const char *bookmark) { nvlist_t *required = fnvlist_alloc(); fnvlist_add_boolean(required, bookmark); IOC_INPUT_TEST_WILD(ZFS_IOC_DESTROY_BOOKMARKS, pool, required, NULL, 0); nvlist_free(required); } static void test_clone(const char *snapshot, const char *clone) { nvlist_t *required = fnvlist_alloc(); nvlist_t *optional = fnvlist_alloc(); nvlist_t *props = fnvlist_alloc(); fnvlist_add_string(required, "origin", snapshot); IOC_INPUT_TEST(ZFS_IOC_CLONE, clone, required, NULL, 0); nvlist_free(props); nvlist_free(optional); nvlist_free(required); } static void test_rollback(const char *dataset, const char *snapshot) { nvlist_t *optional = fnvlist_alloc(); fnvlist_add_string(optional, "target", snapshot); IOC_INPUT_TEST(ZFS_IOC_ROLLBACK, dataset, NULL, optional, B_FALSE); nvlist_free(optional); } static void test_hold(const char *pool, const char *snapshot) { nvlist_t *required = fnvlist_alloc(); nvlist_t *optional = fnvlist_alloc(); nvlist_t *holds = fnvlist_alloc(); fnvlist_add_string(holds, snapshot, "libzfs_check_hold"); fnvlist_add_nvlist(required, "holds", holds); fnvlist_add_int32(optional, "cleanup_fd", zfs_fd); IOC_INPUT_TEST(ZFS_IOC_HOLD, pool, required, optional, 0); nvlist_free(holds); nvlist_free(optional); nvlist_free(required); } static void test_get_holds(const char *snapshot) { IOC_INPUT_TEST(ZFS_IOC_GET_HOLDS, snapshot, NULL, NULL, 0); } static void test_release(const char *pool, const char *snapshot) { nvlist_t *required = fnvlist_alloc(); nvlist_t *release = fnvlist_alloc(); fnvlist_add_boolean(release, "libzfs_check_hold"); fnvlist_add_nvlist(required, snapshot, release); IOC_INPUT_TEST_WILD(ZFS_IOC_RELEASE, pool, required, NULL, 0); nvlist_free(release); nvlist_free(required); } static void test_send_new(const char *snapshot, int fd) { nvlist_t *required = fnvlist_alloc(); nvlist_t *optional = fnvlist_alloc(); fnvlist_add_int32(required, "fd", fd); fnvlist_add_boolean(optional, "largeblockok"); fnvlist_add_boolean(optional, "embedok"); fnvlist_add_boolean(optional, "compressok"); fnvlist_add_boolean(optional, "rawok"); /* * TODO - Resumable send is harder to set up. So we currently * ignore testing for that variant. */ #if 0 fnvlist_add_string(optional, "fromsnap", from); fnvlist_add_uint64(optional, "resume_object", resumeobj); fnvlist_add_uint64(optional, "resume_offset", offset); fnvlist_add_boolean(optional, "savedok"); #endif IOC_INPUT_TEST(ZFS_IOC_SEND_NEW, snapshot, required, optional, 0); nvlist_free(optional); nvlist_free(required); } static void test_recv_new(const char *dataset, int fd) { dmu_replay_record_t drr = { 0 }; nvlist_t *required = fnvlist_alloc(); nvlist_t *optional = fnvlist_alloc(); nvlist_t *props = fnvlist_alloc(); char snapshot[MAXNAMELEN + 32]; ssize_t count; int cleanup_fd = open(ZFS_DEV, O_RDWR); if (cleanup_fd == -1) { (void) fprintf(stderr, "open(%s) failed: %s\n", ZFS_DEV, strerror(errno)); exit(EXIT_FAILURE); } (void) snprintf(snapshot, sizeof (snapshot), "%s@replicant", dataset); count = pread(fd, &drr, sizeof (drr), 0); if (count != sizeof (drr)) { (void) fprintf(stderr, "could not read stream: %s\n", strerror(errno)); } fnvlist_add_string(required, "snapname", snapshot); fnvlist_add_byte_array(required, "begin_record", (uchar_t *)&drr, sizeof (drr)); fnvlist_add_int32(required, "input_fd", fd); fnvlist_add_string(props, "org.openzfs:launch", "September 17th, 2013"); fnvlist_add_nvlist(optional, "localprops", props); fnvlist_add_boolean(optional, "force"); fnvlist_add_boolean(optional, "heal"); fnvlist_add_int32(optional, "cleanup_fd", cleanup_fd); /* * TODO - Resumable receive is harder to set up. So we currently * ignore testing for one. */ #if 0 fnvlist_add_nvlist(optional, "props", recvdprops); fnvlist_add_string(optional, "origin", origin); fnvlist_add_boolean(optional, "resumable"); fnvlist_add_uint64(optional, "action_handle", *action_handle); #endif IOC_INPUT_TEST(ZFS_IOC_RECV_NEW, dataset, required, optional, ENOTSUP); nvlist_free(props); nvlist_free(optional); nvlist_free(required); (void) close(cleanup_fd); } static void test_send_space(const char *snapshot1, const char *snapshot2) { nvlist_t *optional = fnvlist_alloc(); fnvlist_add_string(optional, "from", snapshot1); fnvlist_add_boolean(optional, "largeblockok"); fnvlist_add_boolean(optional, "embedok"); fnvlist_add_boolean(optional, "compressok"); fnvlist_add_boolean(optional, "rawok"); IOC_INPUT_TEST(ZFS_IOC_SEND_SPACE, snapshot2, NULL, optional, 0); nvlist_free(optional); } static void test_remap(const char *dataset) { IOC_INPUT_TEST(ZFS_IOC_REMAP, dataset, NULL, NULL, 0); } static void test_channel_program(const char *pool) { const char *program = "arg = ...\n" "argv = arg[\"argv\"]\n" "return argv[1]"; const char *const argv[1] = { "Hello World!" }; nvlist_t *required = fnvlist_alloc(); nvlist_t *optional = fnvlist_alloc(); nvlist_t *args = fnvlist_alloc(); fnvlist_add_string(required, "program", program); fnvlist_add_string_array(args, "argv", argv, 1); fnvlist_add_nvlist(required, "arg", args); fnvlist_add_boolean_value(optional, "sync", B_TRUE); fnvlist_add_uint64(optional, "instrlimit", 1000 * 1000); fnvlist_add_uint64(optional, "memlimit", 8192 * 1024); IOC_INPUT_TEST(ZFS_IOC_CHANNEL_PROGRAM, pool, required, optional, 0); nvlist_free(args); nvlist_free(optional); nvlist_free(required); } #define WRAPPING_KEY_LEN 32 static void test_load_key(const char *dataset) { nvlist_t *required = fnvlist_alloc(); nvlist_t *optional = fnvlist_alloc(); nvlist_t *hidden = fnvlist_alloc(); uint8_t keydata[WRAPPING_KEY_LEN] = {0}; fnvlist_add_uint8_array(hidden, "wkeydata", keydata, sizeof (keydata)); fnvlist_add_nvlist(required, "hidden_args", hidden); fnvlist_add_boolean(optional, "noop"); IOC_INPUT_TEST(ZFS_IOC_LOAD_KEY, dataset, required, optional, EINVAL); nvlist_free(hidden); nvlist_free(optional); nvlist_free(required); } static void test_change_key(const char *dataset) { IOC_INPUT_TEST(ZFS_IOC_CHANGE_KEY, dataset, NULL, NULL, EINVAL); } static void test_unload_key(const char *dataset) { IOC_INPUT_TEST(ZFS_IOC_UNLOAD_KEY, dataset, NULL, NULL, EACCES); } static void test_vdev_initialize(const char *pool) { nvlist_t *required = fnvlist_alloc(); nvlist_t *vdev_guids = fnvlist_alloc(); fnvlist_add_uint64(vdev_guids, "path", 0xdeadbeefdeadbeef); fnvlist_add_uint64(required, ZPOOL_INITIALIZE_COMMAND, POOL_INITIALIZE_START); fnvlist_add_nvlist(required, ZPOOL_INITIALIZE_VDEVS, vdev_guids); IOC_INPUT_TEST(ZFS_IOC_POOL_INITIALIZE, pool, required, NULL, EINVAL); nvlist_free(vdev_guids); nvlist_free(required); } static void test_vdev_trim(const char *pool) { nvlist_t *required = fnvlist_alloc(); nvlist_t *optional = fnvlist_alloc(); nvlist_t *vdev_guids = fnvlist_alloc(); fnvlist_add_uint64(vdev_guids, "path", 0xdeadbeefdeadbeef); fnvlist_add_uint64(required, ZPOOL_TRIM_COMMAND, POOL_TRIM_START); fnvlist_add_nvlist(required, ZPOOL_TRIM_VDEVS, vdev_guids); fnvlist_add_uint64(optional, ZPOOL_TRIM_RATE, 1ULL << 30); fnvlist_add_boolean_value(optional, ZPOOL_TRIM_SECURE, B_TRUE); IOC_INPUT_TEST(ZFS_IOC_POOL_TRIM, pool, required, optional, EINVAL); nvlist_free(vdev_guids); nvlist_free(optional); nvlist_free(required); } +/* Test with invalid values */ +static void +test_scrub(const char *pool) +{ + nvlist_t *required = fnvlist_alloc(); + fnvlist_add_uint64(required, "scan_type", POOL_SCAN_FUNCS + 1); + fnvlist_add_uint64(required, "scan_command", POOL_SCRUB_FLAGS_END + 1); + IOC_INPUT_TEST(ZFS_IOC_POOL_SCRUB, pool, required, NULL, EINVAL); + nvlist_free(required); +} + static int zfs_destroy(const char *dataset) { zfs_cmd_t zc = {"\0"}; int err; (void) strlcpy(zc.zc_name, dataset, sizeof (zc.zc_name)); zc.zc_name[sizeof (zc.zc_name) - 1] = '\0'; err = lzc_ioctl_fd(zfs_fd, ZFS_IOC_DESTROY, &zc); return (err == 0 ? 0 : errno); } static void test_redact(const char *snapshot1, const char *snapshot2) { nvlist_t *required = fnvlist_alloc(); nvlist_t *snapnv = fnvlist_alloc(); char bookmark[MAXNAMELEN + 32]; fnvlist_add_string(required, "bookname", "testbookmark"); fnvlist_add_boolean(snapnv, snapshot2); fnvlist_add_nvlist(required, "snapnv", snapnv); IOC_INPUT_TEST(ZFS_IOC_REDACT, snapshot1, required, NULL, 0); nvlist_free(snapnv); nvlist_free(required); strlcpy(bookmark, snapshot1, sizeof (bookmark)); *strchr(bookmark, '@') = '\0'; strlcat(bookmark, "#testbookmark", sizeof (bookmark) - strlen(bookmark)); zfs_destroy(bookmark); } static void test_get_bookmark_props(const char *bookmark) { IOC_INPUT_TEST(ZFS_IOC_GET_BOOKMARK_PROPS, bookmark, NULL, NULL, 0); } static void test_wait(const char *pool) { nvlist_t *required = fnvlist_alloc(); nvlist_t *optional = fnvlist_alloc(); fnvlist_add_int32(required, "wait_activity", 2); fnvlist_add_uint64(optional, "wait_tag", 0xdeadbeefdeadbeef); IOC_INPUT_TEST(ZFS_IOC_WAIT, pool, required, optional, EINVAL); nvlist_free(required); nvlist_free(optional); } static void test_wait_fs(const char *dataset) { nvlist_t *required = fnvlist_alloc(); fnvlist_add_int32(required, "wait_activity", 2); IOC_INPUT_TEST(ZFS_IOC_WAIT_FS, dataset, required, NULL, EINVAL); nvlist_free(required); } static void test_get_bootenv(const char *pool) { IOC_INPUT_TEST(ZFS_IOC_GET_BOOTENV, pool, NULL, NULL, 0); } static void test_set_bootenv(const char *pool) { nvlist_t *required = fnvlist_alloc(); fnvlist_add_uint64(required, "version", VB_RAW); fnvlist_add_string(required, GRUB_ENVMAP, "test"); IOC_INPUT_TEST_WILD(ZFS_IOC_SET_BOOTENV, pool, required, NULL, 0); nvlist_free(required); } static void zfs_ioc_input_tests(const char *pool) { char filepath[] = "/tmp/ioc_test_file_XXXXXX"; char dataset[ZFS_MAX_DATASET_NAME_LEN]; char snapbase[ZFS_MAX_DATASET_NAME_LEN + 32]; char snapshot[ZFS_MAX_DATASET_NAME_LEN + 32]; char bookmark[ZFS_MAX_DATASET_NAME_LEN + 32]; char backup[ZFS_MAX_DATASET_NAME_LEN]; char clone[ZFS_MAX_DATASET_NAME_LEN]; char clonesnap[ZFS_MAX_DATASET_NAME_LEN + 32]; int tmpfd, err; /* * Setup names and create a working dataset */ (void) snprintf(dataset, sizeof (dataset), "%s/test-fs", pool); (void) snprintf(snapbase, sizeof (snapbase), "%s@snapbase", dataset); (void) snprintf(snapshot, sizeof (snapshot), "%s@snapshot", dataset); (void) snprintf(bookmark, sizeof (bookmark), "%s#bookmark", dataset); (void) snprintf(clone, sizeof (clone), "%s/test-fs-clone", pool); (void) snprintf(clonesnap, sizeof (clonesnap), "%s@snap", clone); (void) snprintf(backup, sizeof (backup), "%s/backup", pool); err = lzc_create(dataset, LZC_DATSET_TYPE_ZFS, NULL, NULL, -1); if (err) { (void) fprintf(stderr, "could not create '%s': %s\n", dataset, strerror(errno)); exit(2); } tmpfd = mkstemp(filepath); if (tmpfd < 0) { (void) fprintf(stderr, "could not create '%s': %s\n", filepath, strerror(errno)); exit(2); } /* * run a test for each ioctl * Note that some test build on previous test operations */ test_pool_sync(pool); test_pool_reopen(pool); test_pool_checkpoint(pool); test_pool_discard_checkpoint(pool); test_log_history(pool); test_create(dataset); test_snapshot(pool, snapbase); test_snapshot(pool, snapshot); test_space_snaps(snapshot); test_send_space(snapbase, snapshot); test_send_new(snapshot, tmpfd); test_recv_new(backup, tmpfd); test_bookmark(pool, snapshot, bookmark); test_get_bookmarks(dataset); test_get_bookmark_props(bookmark); test_destroy_bookmarks(pool, bookmark); test_hold(pool, snapshot); test_get_holds(snapshot); test_release(pool, snapshot); test_clone(snapshot, clone); test_snapshot(pool, clonesnap); test_redact(snapshot, clonesnap); zfs_destroy(clonesnap); zfs_destroy(clone); test_rollback(dataset, snapshot); test_destroy_snaps(pool, snapshot); test_destroy_snaps(pool, snapbase); test_remap(dataset); test_channel_program(pool); test_load_key(dataset); test_change_key(dataset); test_unload_key(dataset); test_vdev_initialize(pool); test_vdev_trim(pool); test_wait(pool); test_wait_fs(dataset); test_set_bootenv(pool); test_get_bootenv(pool); + test_scrub(pool); + /* * cleanup */ zfs_cmd_t zc = {"\0"}; nvlist_t *snaps = fnvlist_alloc(); fnvlist_add_boolean(snaps, snapshot); (void) lzc_destroy_snaps(snaps, B_FALSE, NULL); nvlist_free(snaps); (void) zfs_destroy(dataset); (void) zfs_destroy(backup); (void) close(tmpfd); (void) unlink(filepath); /* * All the unused slots should yield ZFS_ERR_IOC_CMD_UNAVAIL */ for (int i = 0; i < ARRAY_SIZE(ioc_skip); i++) { if (ioc_tested[ioc_skip[i] - ZFS_IOC_FIRST]) (void) fprintf(stderr, "cmd %d tested, not skipped!\n", (int)(ioc_skip[i] - ZFS_IOC_FIRST)); ioc_tested[ioc_skip[i] - ZFS_IOC_FIRST] = B_TRUE; } (void) strlcpy(zc.zc_name, pool, sizeof (zc.zc_name)); zc.zc_name[sizeof (zc.zc_name) - 1] = '\0'; for (unsigned ioc = ZFS_IOC_FIRST; ioc < ZFS_IOC_LAST; ioc++) { unsigned cmd = ioc - ZFS_IOC_FIRST; if (ioc_tested[cmd]) continue; if (lzc_ioctl_fd(zfs_fd, ioc, &zc) != 0 && errno != ZFS_ERR_IOC_CMD_UNAVAIL) { (void) fprintf(stderr, "cmd %d is missing a test case " "(%d)\n", cmd, errno); } } } enum zfs_ioc_ref { #ifdef __FreeBSD__ ZFS_IOC_BASE = 0, #else ZFS_IOC_BASE = ('Z' << 8), #endif ZFS_IOC_PLATFORM_BASE = ZFS_IOC_BASE + 0x80, }; /* * Canonical reference check of /dev/zfs ioctl numbers. * These cannot change and new ioctl numbers must be appended. */ static boolean_t validate_ioc_values(void) { boolean_t result = B_TRUE; #define CHECK(expr) do { \ if (!(expr)) { \ result = B_FALSE; \ fprintf(stderr, "(%s) === FALSE\n", #expr); \ } \ } while (0) CHECK(ZFS_IOC_BASE + 0 == ZFS_IOC_POOL_CREATE); CHECK(ZFS_IOC_BASE + 1 == ZFS_IOC_POOL_DESTROY); CHECK(ZFS_IOC_BASE + 2 == ZFS_IOC_POOL_IMPORT); CHECK(ZFS_IOC_BASE + 3 == ZFS_IOC_POOL_EXPORT); CHECK(ZFS_IOC_BASE + 4 == ZFS_IOC_POOL_CONFIGS); CHECK(ZFS_IOC_BASE + 5 == ZFS_IOC_POOL_STATS); CHECK(ZFS_IOC_BASE + 6 == ZFS_IOC_POOL_TRYIMPORT); CHECK(ZFS_IOC_BASE + 7 == ZFS_IOC_POOL_SCAN); CHECK(ZFS_IOC_BASE + 8 == ZFS_IOC_POOL_FREEZE); CHECK(ZFS_IOC_BASE + 9 == ZFS_IOC_POOL_UPGRADE); CHECK(ZFS_IOC_BASE + 10 == ZFS_IOC_POOL_GET_HISTORY); CHECK(ZFS_IOC_BASE + 11 == ZFS_IOC_VDEV_ADD); CHECK(ZFS_IOC_BASE + 12 == ZFS_IOC_VDEV_REMOVE); CHECK(ZFS_IOC_BASE + 13 == ZFS_IOC_VDEV_SET_STATE); CHECK(ZFS_IOC_BASE + 14 == ZFS_IOC_VDEV_ATTACH); CHECK(ZFS_IOC_BASE + 15 == ZFS_IOC_VDEV_DETACH); CHECK(ZFS_IOC_BASE + 16 == ZFS_IOC_VDEV_SETPATH); CHECK(ZFS_IOC_BASE + 17 == ZFS_IOC_VDEV_SETFRU); CHECK(ZFS_IOC_BASE + 18 == ZFS_IOC_OBJSET_STATS); CHECK(ZFS_IOC_BASE + 19 == ZFS_IOC_OBJSET_ZPLPROPS); CHECK(ZFS_IOC_BASE + 20 == ZFS_IOC_DATASET_LIST_NEXT); CHECK(ZFS_IOC_BASE + 21 == ZFS_IOC_SNAPSHOT_LIST_NEXT); CHECK(ZFS_IOC_BASE + 22 == ZFS_IOC_SET_PROP); CHECK(ZFS_IOC_BASE + 23 == ZFS_IOC_CREATE); CHECK(ZFS_IOC_BASE + 24 == ZFS_IOC_DESTROY); CHECK(ZFS_IOC_BASE + 25 == ZFS_IOC_ROLLBACK); CHECK(ZFS_IOC_BASE + 26 == ZFS_IOC_RENAME); CHECK(ZFS_IOC_BASE + 27 == ZFS_IOC_RECV); CHECK(ZFS_IOC_BASE + 28 == ZFS_IOC_SEND); CHECK(ZFS_IOC_BASE + 29 == ZFS_IOC_INJECT_FAULT); CHECK(ZFS_IOC_BASE + 30 == ZFS_IOC_CLEAR_FAULT); CHECK(ZFS_IOC_BASE + 31 == ZFS_IOC_INJECT_LIST_NEXT); CHECK(ZFS_IOC_BASE + 32 == ZFS_IOC_ERROR_LOG); CHECK(ZFS_IOC_BASE + 33 == ZFS_IOC_CLEAR); CHECK(ZFS_IOC_BASE + 34 == ZFS_IOC_PROMOTE); CHECK(ZFS_IOC_BASE + 35 == ZFS_IOC_SNAPSHOT); CHECK(ZFS_IOC_BASE + 36 == ZFS_IOC_DSOBJ_TO_DSNAME); CHECK(ZFS_IOC_BASE + 37 == ZFS_IOC_OBJ_TO_PATH); CHECK(ZFS_IOC_BASE + 38 == ZFS_IOC_POOL_SET_PROPS); CHECK(ZFS_IOC_BASE + 39 == ZFS_IOC_POOL_GET_PROPS); CHECK(ZFS_IOC_BASE + 40 == ZFS_IOC_SET_FSACL); CHECK(ZFS_IOC_BASE + 41 == ZFS_IOC_GET_FSACL); CHECK(ZFS_IOC_BASE + 42 == ZFS_IOC_SHARE); CHECK(ZFS_IOC_BASE + 43 == ZFS_IOC_INHERIT_PROP); CHECK(ZFS_IOC_BASE + 44 == ZFS_IOC_SMB_ACL); CHECK(ZFS_IOC_BASE + 45 == ZFS_IOC_USERSPACE_ONE); CHECK(ZFS_IOC_BASE + 46 == ZFS_IOC_USERSPACE_MANY); CHECK(ZFS_IOC_BASE + 47 == ZFS_IOC_USERSPACE_UPGRADE); CHECK(ZFS_IOC_BASE + 48 == ZFS_IOC_HOLD); CHECK(ZFS_IOC_BASE + 49 == ZFS_IOC_RELEASE); CHECK(ZFS_IOC_BASE + 50 == ZFS_IOC_GET_HOLDS); CHECK(ZFS_IOC_BASE + 51 == ZFS_IOC_OBJSET_RECVD_PROPS); CHECK(ZFS_IOC_BASE + 52 == ZFS_IOC_VDEV_SPLIT); CHECK(ZFS_IOC_BASE + 53 == ZFS_IOC_NEXT_OBJ); CHECK(ZFS_IOC_BASE + 54 == ZFS_IOC_DIFF); CHECK(ZFS_IOC_BASE + 55 == ZFS_IOC_TMP_SNAPSHOT); CHECK(ZFS_IOC_BASE + 56 == ZFS_IOC_OBJ_TO_STATS); CHECK(ZFS_IOC_BASE + 57 == ZFS_IOC_SPACE_WRITTEN); CHECK(ZFS_IOC_BASE + 58 == ZFS_IOC_SPACE_SNAPS); CHECK(ZFS_IOC_BASE + 59 == ZFS_IOC_DESTROY_SNAPS); CHECK(ZFS_IOC_BASE + 60 == ZFS_IOC_POOL_REGUID); CHECK(ZFS_IOC_BASE + 61 == ZFS_IOC_POOL_REOPEN); CHECK(ZFS_IOC_BASE + 62 == ZFS_IOC_SEND_PROGRESS); CHECK(ZFS_IOC_BASE + 63 == ZFS_IOC_LOG_HISTORY); CHECK(ZFS_IOC_BASE + 64 == ZFS_IOC_SEND_NEW); CHECK(ZFS_IOC_BASE + 65 == ZFS_IOC_SEND_SPACE); CHECK(ZFS_IOC_BASE + 66 == ZFS_IOC_CLONE); CHECK(ZFS_IOC_BASE + 67 == ZFS_IOC_BOOKMARK); CHECK(ZFS_IOC_BASE + 68 == ZFS_IOC_GET_BOOKMARKS); CHECK(ZFS_IOC_BASE + 69 == ZFS_IOC_DESTROY_BOOKMARKS); CHECK(ZFS_IOC_BASE + 70 == ZFS_IOC_RECV_NEW); CHECK(ZFS_IOC_BASE + 71 == ZFS_IOC_POOL_SYNC); CHECK(ZFS_IOC_BASE + 72 == ZFS_IOC_CHANNEL_PROGRAM); CHECK(ZFS_IOC_BASE + 73 == ZFS_IOC_LOAD_KEY); CHECK(ZFS_IOC_BASE + 74 == ZFS_IOC_UNLOAD_KEY); CHECK(ZFS_IOC_BASE + 75 == ZFS_IOC_CHANGE_KEY); CHECK(ZFS_IOC_BASE + 76 == ZFS_IOC_REMAP); CHECK(ZFS_IOC_BASE + 77 == ZFS_IOC_POOL_CHECKPOINT); CHECK(ZFS_IOC_BASE + 78 == ZFS_IOC_POOL_DISCARD_CHECKPOINT); CHECK(ZFS_IOC_BASE + 79 == ZFS_IOC_POOL_INITIALIZE); CHECK(ZFS_IOC_BASE + 80 == ZFS_IOC_POOL_TRIM); CHECK(ZFS_IOC_BASE + 81 == ZFS_IOC_REDACT); CHECK(ZFS_IOC_BASE + 82 == ZFS_IOC_GET_BOOKMARK_PROPS); CHECK(ZFS_IOC_BASE + 83 == ZFS_IOC_WAIT); CHECK(ZFS_IOC_BASE + 84 == ZFS_IOC_WAIT_FS); + CHECK(ZFS_IOC_BASE + 87 == ZFS_IOC_POOL_SCRUB); CHECK(ZFS_IOC_PLATFORM_BASE + 1 == ZFS_IOC_EVENTS_NEXT); CHECK(ZFS_IOC_PLATFORM_BASE + 2 == ZFS_IOC_EVENTS_CLEAR); CHECK(ZFS_IOC_PLATFORM_BASE + 3 == ZFS_IOC_EVENTS_SEEK); CHECK(ZFS_IOC_PLATFORM_BASE + 4 == ZFS_IOC_NEXTBOOT); CHECK(ZFS_IOC_PLATFORM_BASE + 5 == ZFS_IOC_JAIL); CHECK(ZFS_IOC_PLATFORM_BASE + 6 == ZFS_IOC_UNJAIL); CHECK(ZFS_IOC_PLATFORM_BASE + 7 == ZFS_IOC_SET_BOOTENV); CHECK(ZFS_IOC_PLATFORM_BASE + 8 == ZFS_IOC_GET_BOOTENV); #undef CHECK return (result); } int main(int argc, const char *argv[]) { if (argc != 2) { (void) fprintf(stderr, "usage: %s \n", argv[0]); exit(2); } if (!validate_ioc_values()) { (void) fprintf(stderr, "WARNING: zfs_ioc_t has binary " "incompatible command values\n"); exit(3); } (void) libzfs_core_init(); zfs_fd = open(ZFS_DEV, O_RDWR); if (zfs_fd < 0) { (void) fprintf(stderr, "open: %s\n", strerror(errno)); libzfs_core_fini(); exit(2); } zfs_ioc_input_tests(argv[1]); (void) close(zfs_fd); libzfs_core_fini(); return (unexpected_failures); } diff --git a/sys/contrib/openzfs/tests/zfs-tests/include/libtest.shlib b/sys/contrib/openzfs/tests/zfs-tests/include/libtest.shlib index 8521f271be54..133f8387ddaf 100644 --- a/sys/contrib/openzfs/tests/zfs-tests/include/libtest.shlib +++ b/sys/contrib/openzfs/tests/zfs-tests/include/libtest.shlib @@ -1,3857 +1,3875 @@ # # CDDL HEADER START # # The contents of this file are subject to the terms of the # Common Development and Distribution License (the "License"). # You may not use this file except in compliance with the License. # # You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE # or https://opensource.org/licenses/CDDL-1.0. # See the License for the specific language governing permissions # and limitations under the License. # # When distributing Covered Code, include this CDDL HEADER in each # file and include the License file at usr/src/OPENSOLARIS.LICENSE. # If applicable, add the following below this CDDL HEADER, with the # fields enclosed by brackets "[]" replaced with your own identifying # information: Portions Copyright [yyyy] [name of copyright owner] # # CDDL HEADER END # # # Copyright (c) 2009, Sun Microsystems Inc. All rights reserved. # Copyright (c) 2012, 2020, Delphix. All rights reserved. # Copyright (c) 2017, Tim Chase. All rights reserved. # Copyright (c) 2017, Nexenta Systems Inc. All rights reserved. # Copyright (c) 2017, Lawrence Livermore National Security LLC. # Copyright (c) 2017, Datto Inc. All rights reserved. # Copyright (c) 2017, Open-E Inc. All rights reserved. # Copyright (c) 2021, The FreeBSD Foundation. # Use is subject to license terms. # . ${STF_SUITE}/include/tunables.cfg . ${STF_TOOLS}/include/logapi.shlib . ${STF_SUITE}/include/math.shlib . ${STF_SUITE}/include/blkdev.shlib # # Apply constrained path when available. This is required since the # PATH may have been modified by sudo's secure_path behavior. # if [ -n "$STF_PATH" ]; then export PATH="$STF_PATH" fi # # Generic dot version comparison function # # Returns success when version $1 is greater than or equal to $2. # function compare_version_gte { [ "$(printf "$1\n$2" | sort -V | tail -n1)" = "$1" ] } # Linux kernel version comparison function # # $1 Linux version ("4.10", "2.6.32") or blank for installed Linux version # # Used for comparison: if [ $(linux_version) -ge $(linux_version "2.6.32") ] # function linux_version { typeset ver="$1" [ -z "$ver" ] && ver=$(uname -r | grep -Eo "^[0-9]+\.[0-9]+\.[0-9]+") typeset version major minor _ IFS='.' read -r version major minor _ <<<"$ver" [ -z "$version" ] && version=0 [ -z "$major" ] && major=0 [ -z "$minor" ] && minor=0 echo $((version * 100000 + major * 1000 + minor)) } # Determine if this is a Linux test system # # Return 0 if platform Linux, 1 if otherwise function is_linux { [ "$UNAME" = "Linux" ] } # Determine if this is an illumos test system # # Return 0 if platform illumos, 1 if otherwise function is_illumos { [ "$UNAME" = "illumos" ] } # Determine if this is a FreeBSD test system # # Return 0 if platform FreeBSD, 1 if otherwise function is_freebsd { [ "$UNAME" = "FreeBSD" ] } # Determine if this is a 32-bit system # # Return 0 if platform is 32-bit, 1 if otherwise function is_32bit { [ $(getconf LONG_BIT) = "32" ] } # Determine if kmemleak is enabled # # Return 0 if kmemleak is enabled, 1 if otherwise function is_kmemleak { is_linux && [ -e /sys/kernel/debug/kmemleak ] } # Determine whether a dataset is mounted # # $1 dataset name # $2 filesystem type; optional - defaulted to zfs # # Return 0 if dataset is mounted; 1 if unmounted; 2 on error function ismounted { typeset fstype=$2 [[ -z $fstype ]] && fstype=zfs typeset out dir name case $fstype in zfs) if [[ "$1" == "/"* ]] ; then ! zfs mount | awk -v fs="$1" '$2 == fs {exit 1}' else ! zfs mount | awk -v ds="$1" '$1 == ds {exit 1}' fi ;; ufs|nfs) if is_freebsd; then mount -pt $fstype | while read dev dir _t _flags; do [[ "$1" == "$dev" || "$1" == "$dir" ]] && return 0 done else out=$(df -F $fstype $1 2>/dev/null) || return dir=${out%%\(*} dir=${dir%% *} name=${out##*\(} name=${name%%\)*} name=${name%% *} [[ "$1" == "$dir" || "$1" == "$name" ]] && return 0 fi ;; ext*) df -t $fstype $1 > /dev/null 2>&1 ;; zvol) if [[ -L "$ZVOL_DEVDIR/$1" ]]; then link=$(readlink -f $ZVOL_DEVDIR/$1) [[ -n "$link" ]] && \ mount | grep -q "^$link" && \ return 0 fi ;; *) false ;; esac } # Return 0 if a dataset is mounted; 1 otherwise # # $1 dataset name # $2 filesystem type; optional - defaulted to zfs function mounted { ismounted $1 $2 } # Return 0 if a dataset is unmounted; 1 otherwise # # $1 dataset name # $2 filesystem type; optional - defaulted to zfs function unmounted { ! ismounted $1 $2 } function default_setup { default_setup_noexit "$@" log_pass } function default_setup_no_mountpoint { default_setup_noexit "$1" "$2" "$3" "yes" log_pass } # # Given a list of disks, setup storage pools and datasets. # function default_setup_noexit { typeset disklist=$1 typeset container=$2 typeset volume=$3 typeset no_mountpoint=$4 log_note begin default_setup_noexit if is_global_zone; then if poolexists $TESTPOOL ; then destroy_pool $TESTPOOL fi [[ -d /$TESTPOOL ]] && rm -rf /$TESTPOOL log_must zpool create -f $TESTPOOL $disklist else reexport_pool fi rm -rf $TESTDIR || log_unresolved Could not remove $TESTDIR mkdir -p $TESTDIR || log_unresolved Could not create $TESTDIR log_must zfs create $TESTPOOL/$TESTFS if [[ -z $no_mountpoint ]]; then log_must zfs set mountpoint=$TESTDIR $TESTPOOL/$TESTFS fi if [[ -n $container ]]; then rm -rf $TESTDIR1 || \ log_unresolved Could not remove $TESTDIR1 mkdir -p $TESTDIR1 || \ log_unresolved Could not create $TESTDIR1 log_must zfs create $TESTPOOL/$TESTCTR log_must zfs set canmount=off $TESTPOOL/$TESTCTR log_must zfs create $TESTPOOL/$TESTCTR/$TESTFS1 if [[ -z $no_mountpoint ]]; then log_must zfs set mountpoint=$TESTDIR1 \ $TESTPOOL/$TESTCTR/$TESTFS1 fi fi if [[ -n $volume ]]; then if is_global_zone ; then log_must zfs create -V $VOLSIZE $TESTPOOL/$TESTVOL block_device_wait else log_must zfs create $TESTPOOL/$TESTVOL fi fi } # # Given a list of disks, setup a storage pool, file system and # a container. # function default_container_setup { typeset disklist=$1 default_setup "$disklist" "true" } # # Given a list of disks, setup a storage pool,file system # and a volume. # function default_volume_setup { typeset disklist=$1 default_setup "$disklist" "" "true" } # # Given a list of disks, setup a storage pool,file system, # a container and a volume. # function default_container_volume_setup { typeset disklist=$1 default_setup "$disklist" "true" "true" } # # Create a snapshot on a filesystem or volume. Defaultly create a snapshot on # filesystem # # $1 Existing filesystem or volume name. Default, $TESTPOOL/$TESTFS # $2 snapshot name. Default, $TESTSNAP # function create_snapshot { typeset fs_vol=${1:-$TESTPOOL/$TESTFS} typeset snap=${2:-$TESTSNAP} [[ -z $fs_vol ]] && log_fail "Filesystem or volume's name is undefined." [[ -z $snap ]] && log_fail "Snapshot's name is undefined." if snapexists $fs_vol@$snap; then log_fail "$fs_vol@$snap already exists." fi datasetexists $fs_vol || \ log_fail "$fs_vol must exist." log_must zfs snapshot $fs_vol@$snap } # # Create a clone from a snapshot, default clone name is $TESTCLONE. # # $1 Existing snapshot, $TESTPOOL/$TESTFS@$TESTSNAP is default. # $2 Clone name, $TESTPOOL/$TESTCLONE is default. # function create_clone # snapshot clone { typeset snap=${1:-$TESTPOOL/$TESTFS@$TESTSNAP} typeset clone=${2:-$TESTPOOL/$TESTCLONE} [[ -z $snap ]] && \ log_fail "Snapshot name is undefined." [[ -z $clone ]] && \ log_fail "Clone name is undefined." log_must zfs clone $snap $clone } # # Create a bookmark of the given snapshot. Defaultly create a bookmark on # filesystem. # # $1 Existing filesystem or volume name. Default, $TESTFS # $2 Existing snapshot name. Default, $TESTSNAP # $3 bookmark name. Default, $TESTBKMARK # function create_bookmark { typeset fs_vol=${1:-$TESTFS} typeset snap=${2:-$TESTSNAP} typeset bkmark=${3:-$TESTBKMARK} [[ -z $fs_vol ]] && log_fail "Filesystem or volume's name is undefined." [[ -z $snap ]] && log_fail "Snapshot's name is undefined." [[ -z $bkmark ]] && log_fail "Bookmark's name is undefined." if bkmarkexists $fs_vol#$bkmark; then log_fail "$fs_vol#$bkmark already exists." fi datasetexists $fs_vol || \ log_fail "$fs_vol must exist." snapexists $fs_vol@$snap || \ log_fail "$fs_vol@$snap must exist." log_must zfs bookmark $fs_vol@$snap $fs_vol#$bkmark } # # Create a temporary clone result of an interrupted resumable 'zfs receive' # $1 Destination filesystem name. Must not exist, will be created as the result # of this function along with its %recv temporary clone # $2 Source filesystem name. Must not exist, will be created and destroyed # function create_recv_clone { typeset recvfs="$1" typeset sendfs="${2:-$TESTPOOL/create_recv_clone}" typeset snap="$sendfs@snap1" typeset incr="$sendfs@snap2" typeset mountpoint="$TESTDIR/create_recv_clone" typeset sendfile="$TESTDIR/create_recv_clone.zsnap" [[ -z $recvfs ]] && log_fail "Recv filesystem's name is undefined." datasetexists $recvfs && log_fail "Recv filesystem must not exist." datasetexists $sendfs && log_fail "Send filesystem must not exist." log_must zfs create -o compression=off -o mountpoint="$mountpoint" $sendfs log_must zfs snapshot $snap log_must eval "zfs send $snap | zfs recv -u $recvfs" log_must mkfile 1m "$mountpoint/data" log_must zfs snapshot $incr log_must eval "zfs send -i $snap $incr | dd bs=10K count=1 \ iflag=fullblock > $sendfile" log_mustnot eval "zfs recv -su $recvfs < $sendfile" destroy_dataset "$sendfs" "-r" log_must rm -f "$sendfile" if [[ $(get_prop 'inconsistent' "$recvfs/%recv") -ne 1 ]]; then log_fail "Error creating temporary $recvfs/%recv clone" fi } function default_mirror_setup { default_mirror_setup_noexit $1 $2 $3 log_pass } # # Given a pair of disks, set up a storage pool and dataset for the mirror # @parameters: $1 the primary side of the mirror # $2 the secondary side of the mirror # @uses: ZPOOL ZFS TESTPOOL TESTFS function default_mirror_setup_noexit { readonly func="default_mirror_setup_noexit" typeset primary=$1 typeset secondary=$2 [[ -z $primary ]] && \ log_fail "$func: No parameters passed" [[ -z $secondary ]] && \ log_fail "$func: No secondary partition passed" [[ -d /$TESTPOOL ]] && rm -rf /$TESTPOOL log_must zpool create -f $TESTPOOL mirror $@ log_must zfs create $TESTPOOL/$TESTFS log_must zfs set mountpoint=$TESTDIR $TESTPOOL/$TESTFS } # # Destroy the configured testpool mirrors. # the mirrors are of the form ${TESTPOOL}{number} # @uses: ZPOOL ZFS TESTPOOL function destroy_mirrors { default_cleanup_noexit log_pass } function default_raidz_setup { default_raidz_setup_noexit "$*" log_pass } # # Given a minimum of two disks, set up a storage pool and dataset for the raid-z # $1 the list of disks # function default_raidz_setup_noexit { typeset disklist="$*" disks=(${disklist[*]}) if [[ ${#disks[*]} -lt 2 ]]; then log_fail "A raid-z requires a minimum of two disks." fi [[ -d /$TESTPOOL ]] && rm -rf /$TESTPOOL log_must zpool create -f $TESTPOOL raidz $disklist log_must zfs create $TESTPOOL/$TESTFS log_must zfs set mountpoint=$TESTDIR $TESTPOOL/$TESTFS } # # Common function used to cleanup storage pools and datasets. # # Invoked at the start of the test suite to ensure the system # is in a known state, and also at the end of each set of # sub-tests to ensure errors from one set of tests doesn't # impact the execution of the next set. function default_cleanup { default_cleanup_noexit log_pass } # # Utility function used to list all available pool names. # # NOTE: $KEEP is a variable containing pool names, separated by a newline # character, that must be excluded from the returned list. # function get_all_pools { zpool list -H -o name | grep -Fvx "$KEEP" | grep -v "$NO_POOLS" } function default_cleanup_noexit { typeset pool="" # # Destroying the pool will also destroy any # filesystems it contains. # if is_global_zone; then zfs unmount -a > /dev/null 2>&1 ALL_POOLS=$(get_all_pools) # Here, we loop through the pools we're allowed to # destroy, only destroying them if it's safe to do # so. while [ ! -z ${ALL_POOLS} ] do for pool in ${ALL_POOLS} do if safe_to_destroy_pool $pool ; then destroy_pool $pool fi done ALL_POOLS=$(get_all_pools) done zfs mount -a else typeset fs="" for fs in $(zfs list -H -o name \ | grep "^$ZONE_POOL/$ZONE_CTR[01234]/"); do destroy_dataset "$fs" "-Rf" done # Need cleanup here to avoid garbage dir left. for fs in $(zfs list -H -o name); do [[ $fs == /$ZONE_POOL ]] && continue [[ -d $fs ]] && log_must rm -rf $fs/* done # # Reset the $ZONE_POOL/$ZONE_CTR[01234] file systems property to # the default value # for fs in $(zfs list -H -o name); do if [[ $fs == $ZONE_POOL/$ZONE_CTR[01234] ]]; then log_must zfs set reservation=none $fs log_must zfs set recordsize=128K $fs log_must zfs set mountpoint=/$fs $fs typeset enc=$(get_prop encryption $fs) if [ -z "$enc" ] || [ "$enc" = "off" ]; then log_must zfs set checksum=on $fs fi log_must zfs set compression=off $fs log_must zfs set atime=on $fs log_must zfs set devices=off $fs log_must zfs set exec=on $fs log_must zfs set setuid=on $fs log_must zfs set readonly=off $fs log_must zfs set snapdir=hidden $fs log_must zfs set aclmode=groupmask $fs log_must zfs set aclinherit=secure $fs fi done fi [[ -d $TESTDIR ]] && \ log_must rm -rf $TESTDIR disk1=${DISKS%% *} if is_mpath_device $disk1; then delete_partitions fi rm -f $TEST_BASE_DIR/{err,out} } # # Common function used to cleanup storage pools, file systems # and containers. # function default_container_cleanup { if ! is_global_zone; then reexport_pool fi ismounted $TESTPOOL/$TESTCTR/$TESTFS1 && log_must zfs unmount $TESTPOOL/$TESTCTR/$TESTFS1 destroy_dataset "$TESTPOOL/$TESTCTR/$TESTFS1" "-R" destroy_dataset "$TESTPOOL/$TESTCTR" "-Rf" [[ -e $TESTDIR1 ]] && \ log_must rm -rf $TESTDIR1 default_cleanup } # # Common function used to cleanup snapshot of file system or volume. Default to # delete the file system's snapshot # # $1 snapshot name # function destroy_snapshot { typeset snap=${1:-$TESTPOOL/$TESTFS@$TESTSNAP} if ! snapexists $snap; then log_fail "'$snap' does not exist." fi # # For the sake of the value which come from 'get_prop' is not equal # to the really mountpoint when the snapshot is unmounted. So, firstly # check and make sure this snapshot's been mounted in current system. # typeset mtpt="" if ismounted $snap; then mtpt=$(get_prop mountpoint $snap) fi destroy_dataset "$snap" [[ $mtpt != "" && -d $mtpt ]] && \ log_must rm -rf $mtpt } # # Common function used to cleanup clone. # # $1 clone name # function destroy_clone { typeset clone=${1:-$TESTPOOL/$TESTCLONE} if ! datasetexists $clone; then log_fail "'$clone' does not existed." fi # With the same reason in destroy_snapshot typeset mtpt="" if ismounted $clone; then mtpt=$(get_prop mountpoint $clone) fi destroy_dataset "$clone" [[ $mtpt != "" && -d $mtpt ]] && \ log_must rm -rf $mtpt } # # Common function used to cleanup bookmark of file system or volume. Default # to delete the file system's bookmark. # # $1 bookmark name # function destroy_bookmark { typeset bkmark=${1:-$TESTPOOL/$TESTFS#$TESTBKMARK} if ! bkmarkexists $bkmark; then log_fail "'$bkmarkp' does not existed." fi destroy_dataset "$bkmark" } # Return 0 if a snapshot exists; $? otherwise # # $1 - snapshot name function snapexists { zfs list -H -t snapshot "$1" > /dev/null 2>&1 } # # Return 0 if a bookmark exists; $? otherwise # # $1 - bookmark name # function bkmarkexists { zfs list -H -t bookmark "$1" > /dev/null 2>&1 } # # Return 0 if a hold exists; $? otherwise # # $1 - hold tag # $2 - snapshot name # function holdexists { ! zfs holds "$2" | awk -v t="$1" '$2 ~ t { exit 1 }' } # # Set a property to a certain value on a dataset. # Sets a property of the dataset to the value as passed in. # @param: # $1 dataset who's property is being set # $2 property to set # $3 value to set property to # @return: # 0 if the property could be set. # non-zero otherwise. # @use: ZFS # function dataset_setprop { typeset fn=dataset_setprop if (($# < 3)); then log_note "$fn: Insufficient parameters (need 3, had $#)" return 1 fi typeset output= output=$(zfs set $2=$3 $1 2>&1) typeset rv=$? if ((rv != 0)); then log_note "Setting property on $1 failed." log_note "property $2=$3" log_note "Return Code: $rv" log_note "Output: $output" return $rv fi return 0 } # # Check a numeric assertion # @parameter: $@ the assertion to check # @output: big loud notice if assertion failed # @use: log_fail # function assert { (($@)) || log_fail "$@" } # # Function to format partition size of a disk # Given a disk cxtxdx reduces all partitions # to 0 size # function zero_partitions # { typeset diskname=$1 typeset i if is_freebsd; then gpart destroy -F $diskname elif is_linux; then DSK=$DEV_DSKDIR/$diskname DSK=$(echo $DSK | sed -e "s|//|/|g") log_must parted $DSK -s -- mklabel gpt blockdev --rereadpt $DSK 2>/dev/null block_device_wait else for i in 0 1 3 4 5 6 7 do log_must set_partition $i "" 0mb $diskname done fi return 0 } # # Given a slice, size and disk, this function # formats the slice to the specified size. # Size should be specified with units as per # the `format` command requirements eg. 100mb 3gb # # NOTE: This entire interface is problematic for the Linux parted utility # which requires the end of the partition to be specified. It would be # best to retire this interface and replace it with something more flexible. # At the moment a best effort is made. # # arguments: function set_partition { typeset -i slicenum=$1 typeset start=$2 typeset size=$3 typeset disk=${4#$DEV_DSKDIR/} disk=${disk#$DEV_RDSKDIR/} case "$UNAME" in Linux) if [[ -z $size || -z $disk ]]; then log_fail "The size or disk name is unspecified." fi disk=$DEV_DSKDIR/$disk typeset size_mb=${size%%[mMgG]} size_mb=${size_mb%%[mMgG][bB]} if [[ ${size:1:1} == 'g' ]]; then ((size_mb = size_mb * 1024)) fi # Create GPT partition table when setting slice 0 or # when the device doesn't already contain a GPT label. parted $disk -s -- print 1 >/dev/null typeset ret_val=$? if [[ $slicenum -eq 0 || $ret_val -ne 0 ]]; then if ! parted $disk -s -- mklabel gpt; then log_note "Failed to create GPT partition table on $disk" return 1 fi fi # When no start is given align on the first cylinder. if [[ -z "$start" ]]; then start=1 fi # Determine the cylinder size for the device and using # that calculate the end offset in cylinders. typeset -i cly_size_kb=0 cly_size_kb=$(parted -m $disk -s -- unit cyl print | awk -F '[:k.]' 'NR == 3 {print $4}') ((end = (size_mb * 1024 / cly_size_kb) + start)) parted $disk -s -- \ mkpart part$slicenum ${start}cyl ${end}cyl typeset ret_val=$? if [[ $ret_val -ne 0 ]]; then log_note "Failed to create partition $slicenum on $disk" return 1 fi blockdev --rereadpt $disk 2>/dev/null block_device_wait $disk ;; FreeBSD) if [[ -z $size || -z $disk ]]; then log_fail "The size or disk name is unspecified." fi disk=$DEV_DSKDIR/$disk if [[ $slicenum -eq 0 ]] || ! gpart show $disk >/dev/null 2>&1; then gpart destroy -F $disk >/dev/null 2>&1 if ! gpart create -s GPT $disk; then log_note "Failed to create GPT partition table on $disk" return 1 fi fi typeset index=$((slicenum + 1)) if [[ -n $start ]]; then start="-b $start" fi gpart add -t freebsd-zfs $start -s $size -i $index $disk if [[ $ret_val -ne 0 ]]; then log_note "Failed to create partition $slicenum on $disk" return 1 fi block_device_wait $disk ;; *) if [[ -z $slicenum || -z $size || -z $disk ]]; then log_fail "The slice, size or disk name is unspecified." fi typeset format_file=/var/tmp/format_in.$$ echo "partition" >$format_file echo "$slicenum" >> $format_file echo "" >> $format_file echo "" >> $format_file echo "$start" >> $format_file echo "$size" >> $format_file echo "label" >> $format_file echo "" >> $format_file echo "q" >> $format_file echo "q" >> $format_file format -e -s -d $disk -f $format_file typeset ret_val=$? rm -f $format_file ;; esac if [[ $ret_val -ne 0 ]]; then log_note "Unable to format $disk slice $slicenum to $size" return 1 fi return 0 } # # Delete all partitions on all disks - this is specifically for the use of multipath # devices which currently can only be used in the test suite as raw/un-partitioned # devices (ie a zpool cannot be created on a whole mpath device that has partitions) # function delete_partitions { typeset disk if [[ -z $DISKSARRAY ]]; then DISKSARRAY=$DISKS fi if is_linux; then typeset -i part for disk in $DISKSARRAY; do for (( part = 1; part < MAX_PARTITIONS; part++ )); do typeset partition=${disk}${SLICE_PREFIX}${part} parted $DEV_DSKDIR/$disk -s rm $part > /dev/null 2>&1 if lsblk | grep -qF ${partition}; then log_fail "Partition ${partition} not deleted" else log_note "Partition ${partition} deleted" fi done done elif is_freebsd; then for disk in $DISKSARRAY; do if gpart destroy -F $disk; then log_note "Partitions for ${disk} deleted" else log_fail "Partitions for ${disk} not deleted" fi done fi } # # Get the end cyl of the given slice # function get_endslice # { typeset disk=$1 typeset slice=$2 if [[ -z $disk || -z $slice ]] ; then log_fail "The disk name or slice number is unspecified." fi case "$UNAME" in Linux) endcyl=$(parted -s $DEV_DSKDIR/$disk -- unit cyl print | \ awk "/part${slice}/"' {sub(/cyl/, "", $3); print $3}') ((endcyl = (endcyl + 1))) ;; FreeBSD) disk=${disk#/dev/zvol/} disk=${disk%p*} slice=$((slice + 1)) endcyl=$(gpart show $disk | \ awk -v slice=$slice '$3 == slice { print $1 + $2 }') ;; *) disk=${disk#/dev/dsk/} disk=${disk#/dev/rdsk/} disk=${disk%s*} typeset -i ratio=0 ratio=$(prtvtoc /dev/rdsk/${disk}s2 | \ awk '/sectors\/cylinder/ {print $2}') if ((ratio == 0)); then return fi typeset -i endcyl=$(prtvtoc -h /dev/rdsk/${disk}s2 | awk -v token="$slice" '$1 == token {print $6}') ((endcyl = (endcyl + 1) / ratio)) ;; esac echo $endcyl } # # Given a size,disk and total slice number, this function formats the # disk slices from 0 to the total slice number with the same specified # size. # function partition_disk # { typeset -i i=0 typeset slice_size=$1 typeset disk_name=$2 typeset total_slices=$3 typeset cyl zero_partitions $disk_name while ((i < $total_slices)); do if ! is_linux; then if ((i == 2)); then ((i = i + 1)) continue fi fi log_must set_partition $i "$cyl" $slice_size $disk_name cyl=$(get_endslice $disk_name $i) ((i = i+1)) done } # # This function continues to write to a filenum number of files into dirnum # number of directories until either file_write returns an error or the # maximum number of files per directory have been written. # # Usage: # fill_fs [destdir] [dirnum] [filenum] [bytes] [num_writes] [data] # # Return value: 0 on success # non 0 on error # # Where : # destdir: is the directory where everything is to be created under # dirnum: the maximum number of subdirectories to use, -1 no limit # filenum: the maximum number of files per subdirectory # bytes: number of bytes to write # num_writes: number of types to write out bytes # data: the data that will be written # # E.g. # fill_fs /testdir 20 25 1024 256 0 # # Note: bytes * num_writes equals the size of the testfile # function fill_fs # destdir dirnum filenum bytes num_writes data { typeset destdir=${1:-$TESTDIR} typeset -i dirnum=${2:-50} typeset -i filenum=${3:-50} typeset -i bytes=${4:-8192} typeset -i num_writes=${5:-10240} typeset data=${6:-0} mkdir -p $destdir/{1..$dirnum} for f in $destdir/{1..$dirnum}/$TESTFILE{1..$filenum}; do file_write -o create -f $f -b $bytes -c $num_writes -d $data \ || return done } # Get the specified dataset property in parsable format or fail function get_prop # property dataset { typeset prop=$1 typeset dataset=$2 zfs get -Hpo value "$prop" "$dataset" || log_fail "zfs get $prop $dataset" } # Get the specified pool property in parsable format or fail function get_pool_prop # property pool { typeset prop=$1 typeset pool=$2 zpool get -Hpo value "$prop" "$pool" || log_fail "zpool get $prop $pool" } # Return 0 if a pool exists; $? otherwise # # $1 - pool name function poolexists { typeset pool=$1 if [[ -z $pool ]]; then log_note "No pool name given." return 1 fi zpool get name "$pool" > /dev/null 2>&1 } # Return 0 if all the specified datasets exist; $? otherwise # # $1-n dataset name function datasetexists { if (($# == 0)); then log_note "No dataset name given." return 1 fi zfs get name "$@" > /dev/null 2>&1 } # return 0 if none of the specified datasets exists, otherwise return 1. # # $1-n dataset name function datasetnonexists { if (($# == 0)); then log_note "No dataset name given." return 1 fi while (($# > 0)); do zfs list -H -t filesystem,snapshot,volume $1 > /dev/null 2>&1 \ && return 1 shift done return 0 } # FreeBSD breaks exports(5) at whitespace and doesn't process escapes # Solaris just breaks # # cf. https://github.com/openzfs/zfs/pull/13165#issuecomment-1059845807 # # Linux can have spaces (which are \OOO-escaped), # but can't have backslashes because they're parsed recursively function shares_can_have_whitespace { is_linux } function is_shared_freebsd { typeset fs=$1 pgrep -q mountd && showmount -E | grep -qx "$fs" } function is_shared_illumos { typeset fs=$1 typeset mtpt for mtpt in `share | awk '{print $2}'` ; do if [[ $mtpt == $fs ]] ; then return 0 fi done typeset stat=$(svcs -H -o STA nfs/server:default) if [[ $stat != "ON" ]]; then log_note "Current nfs/server status: $stat" fi return 1 } function is_shared_linux { typeset fs=$1 ! exportfs -s | awk -v fs="${fs//\\/\\\\}" '/^\// && $1 == fs {exit 1}' } # # Given a mountpoint, or a dataset name, determine if it is shared via NFS. # # Returns 0 if shared, 1 otherwise. # function is_shared { typeset fs=$1 typeset mtpt if [[ $fs != "/"* ]] ; then if datasetnonexists "$fs" ; then return 1 else mtpt=$(get_prop mountpoint "$fs") case "$mtpt" in none|legacy|-) return 1 ;; *) fs=$mtpt ;; esac fi fi case "$UNAME" in FreeBSD) is_shared_freebsd "$fs" ;; Linux) is_shared_linux "$fs" ;; *) is_shared_illumos "$fs" ;; esac } function is_exported_illumos { typeset fs=$1 typeset mtpt _ while read -r mtpt _; do [ "$mtpt" = "$fs" ] && return done < /etc/dfs/sharetab return 1 } function is_exported_freebsd { typeset fs=$1 typeset mtpt _ while read -r mtpt _; do [ "$mtpt" = "$fs" ] && return done < /etc/zfs/exports return 1 } function is_exported_linux { typeset fs=$1 typeset mtpt _ while read -r mtpt _; do [ "$(printf "$mtpt")" = "$fs" ] && return done < /etc/exports.d/zfs.exports return 1 } # # Given a mountpoint, or a dataset name, determine if it is exported via # the os-specific NFS exports file. # # Returns 0 if exported, 1 otherwise. # function is_exported { typeset fs=$1 typeset mtpt if [[ $fs != "/"* ]] ; then if datasetnonexists "$fs" ; then return 1 else mtpt=$(get_prop mountpoint "$fs") case $mtpt in none|legacy|-) return 1 ;; *) fs=$mtpt ;; esac fi fi case "$UNAME" in FreeBSD) is_exported_freebsd "$fs" ;; Linux) is_exported_linux "$fs" ;; *) is_exported_illumos "$fs" ;; esac } # # Given a dataset name determine if it is shared via SMB. # # Returns 0 if shared, 1 otherwise. # function is_shared_smb { typeset fs=$1 datasetexists "$fs" || return if is_linux; then net usershare list | grep -xFq "${fs//[-\/]/_}" else log_note "SMB on $UNAME currently unsupported by the test framework" return 1 fi } # # Given a mountpoint, determine if it is not shared via NFS. # # Returns 0 if not shared, 1 otherwise. # function not_shared { ! is_shared $1 } # # Given a dataset determine if it is not shared via SMB. # # Returns 0 if not shared, 1 otherwise. # function not_shared_smb { ! is_shared_smb $1 } # # Helper function to unshare a mountpoint. # function unshare_fs #fs { typeset fs=$1 if is_shared $fs || is_shared_smb $fs; then log_must zfs unshare $fs fi } # # Helper function to share a NFS mountpoint. # function share_nfs #fs { typeset fs=$1 is_shared "$fs" && return case "$UNAME" in Linux) log_must exportfs "*:$fs" ;; FreeBSD) typeset mountd read -r mountd < /var/run/mountd.pid log_must eval "printf '%s\t\n' \"$fs\" >> /etc/zfs/exports" log_must kill -s HUP "$mountd" ;; *) log_must share -F nfs "$fs" ;; esac return 0 } # # Helper function to unshare a NFS mountpoint. # function unshare_nfs #fs { typeset fs=$1 ! is_shared "$fs" && return case "$UNAME" in Linux) log_must exportfs -u "*:$fs" ;; FreeBSD) typeset mountd read -r mountd < /var/run/mountd.pid awk -v fs="${fs//\\/\\\\}" '$1 != fs' /etc/zfs/exports > /etc/zfs/exports.$$ log_must mv /etc/zfs/exports.$$ /etc/zfs/exports log_must kill -s HUP "$mountd" ;; *) log_must unshare -F nfs $fs ;; esac return 0 } # # Helper function to show NFS shares. # function showshares_nfs { case "$UNAME" in Linux) exportfs -v ;; FreeBSD) showmount ;; *) share -F nfs ;; esac } function check_nfs { case "$UNAME" in Linux) exportfs -s ;; FreeBSD) showmount -e ;; *) log_unsupported "Unknown platform" ;; esac || log_unsupported "The NFS utilities are not installed" } # # Check NFS server status and trigger it online. # function setup_nfs_server { # Cannot share directory in non-global zone. # if ! is_global_zone; then log_note "Cannot trigger NFS server by sharing in LZ." return fi if is_linux; then # # Re-synchronize /var/lib/nfs/etab with /etc/exports and # /etc/exports.d./* to provide a clean test environment. # log_must exportfs -r log_note "NFS server must be started prior to running ZTS." return elif is_freebsd; then log_must kill -s HUP $(/dev/null) [ $cur_zone = "global" ] fi } # # Verify whether test is permitted to run from # global zone, local zone, or both # # $1 zone limit, could be "global", "local", or "both"(no limit) # # Return 0 if permitted, otherwise exit with log_unsupported # function verify_runnable # zone limit { typeset limit=$1 [[ -z $limit ]] && return 0 if is_global_zone ; then case $limit in global|both) ;; local) log_unsupported "Test is unable to run from "\ "global zone." ;; *) log_note "Warning: unknown limit $limit - " \ "use both." ;; esac else case $limit in local|both) ;; global) log_unsupported "Test is unable to run from "\ "local zone." ;; *) log_note "Warning: unknown limit $limit - " \ "use both." ;; esac reexport_pool fi return 0 } # Return 0 if create successfully or the pool exists; $? otherwise # Note: In local zones, this function should return 0 silently. # # $1 - pool name # $2-n - [keyword] devs_list function create_pool #pool devs_list { typeset pool=${1%%/*} shift if [[ -z $pool ]]; then log_note "Missing pool name." return 1 fi if poolexists $pool ; then destroy_pool $pool fi if is_global_zone ; then [[ -d /$pool ]] && rm -rf /$pool log_must zpool create -f $pool $@ fi return 0 } # Return 0 if destroy successfully or the pool exists; $? otherwise # Note: In local zones, this function should return 0 silently. # # $1 - pool name # Destroy pool with the given parameters. function destroy_pool #pool { typeset pool=${1%%/*} typeset mtpt if [[ -z $pool ]]; then log_note "No pool name given." return 1 fi if is_global_zone ; then if poolexists "$pool" ; then mtpt=$(get_prop mountpoint "$pool") # At times, syseventd/udev activity can cause attempts # to destroy a pool to fail with EBUSY. We retry a few # times allowing failures before requiring the destroy # to succeed. log_must_busy zpool destroy -f $pool [[ -d $mtpt ]] && \ log_must rm -rf $mtpt else log_note "Pool does not exist. ($pool)" return 1 fi fi return 0 } # Return 0 if created successfully; $? otherwise # # $1 - dataset name # $2-n - dataset options function create_dataset #dataset dataset_options { typeset dataset=$1 shift if [[ -z $dataset ]]; then log_note "Missing dataset name." return 1 fi if datasetexists $dataset ; then destroy_dataset $dataset fi log_must zfs create $@ $dataset return 0 } # Return 0 if destroy successfully or the dataset exists; $? otherwise # Note: In local zones, this function should return 0 silently. # # $1 - dataset name # $2 - custom arguments for zfs destroy # Destroy dataset with the given parameters. function destroy_dataset # dataset [args] { typeset dataset=$1 typeset mtpt typeset args=${2:-""} if [[ -z $dataset ]]; then log_note "No dataset name given." return 1 fi if is_global_zone ; then if datasetexists "$dataset" ; then mtpt=$(get_prop mountpoint "$dataset") log_must_busy zfs destroy $args $dataset [ -d $mtpt ] && log_must rm -rf $mtpt else log_note "Dataset does not exist. ($dataset)" return 1 fi fi return 0 } # # Reexport TESTPOOL & TESTPOOL(1-4) # function reexport_pool { typeset -i cntctr=5 typeset -i i=0 while ((i < cntctr)); do if ((i == 0)); then TESTPOOL=$ZONE_POOL/$ZONE_CTR$i if ! ismounted $TESTPOOL; then log_must zfs mount $TESTPOOL fi else eval TESTPOOL$i=$ZONE_POOL/$ZONE_CTR$i if eval ! ismounted \$TESTPOOL$i; then log_must eval zfs mount \$TESTPOOL$i fi fi ((i += 1)) done } # # Verify a given disk or pool state # # Return 0 is pool/disk matches expected state, 1 otherwise # function check_state # pool disk state{online,offline,degraded} { typeset pool=$1 typeset disk=${2#$DEV_DSKDIR/} typeset state=$3 [[ -z $pool ]] || [[ -z $state ]] \ && log_fail "Arguments invalid or missing" if [[ -z $disk ]]; then #check pool state only zpool get -H -o value health $pool | grep -qi "$state" else zpool status -v $pool | grep "$disk" | grep -qi "$state" fi } # # Get the mountpoint of snapshot # For the snapshot use /.zfs/snapshot/ # as its mountpoint # function snapshot_mountpoint { typeset dataset=${1:-$TESTPOOL/$TESTFS@$TESTSNAP} if [[ $dataset != *@* ]]; then log_fail "Error name of snapshot '$dataset'." fi typeset fs=${dataset%@*} typeset snap=${dataset#*@} if [[ -z $fs || -z $snap ]]; then log_fail "Error name of snapshot '$dataset'." fi echo $(get_prop mountpoint $fs)/.zfs/snapshot/$snap } # # Given a device and 'ashift' value verify it's correctly set on every label # function verify_ashift # device ashift { typeset device="$1" typeset ashift="$2" zdb -e -lll $device | awk -v ashift=$ashift ' /ashift: / { if (ashift != $2) exit 1; else count++; } END { exit (count != 4); }' } # # Given a pool and file system, this function will verify the file system # using the zdb internal tool. Note that the pool is exported and imported # to ensure it has consistent state. # function verify_filesys # pool filesystem dir { typeset pool="$1" typeset filesys="$2" typeset zdbout="/tmp/zdbout.$$" shift shift typeset dirs=$@ typeset search_path="" log_note "Calling zdb to verify filesystem '$filesys'" zfs unmount -a > /dev/null 2>&1 log_must zpool export $pool if [[ -n $dirs ]] ; then for dir in $dirs ; do search_path="$search_path -d $dir" done fi log_must zpool import $search_path $pool if ! zdb -cudi $filesys > $zdbout 2>&1; then log_note "Output: zdb -cudi $filesys" cat $zdbout rm -f $zdbout log_fail "zdb detected errors with: '$filesys'" fi log_must zfs mount -a log_must rm -rf $zdbout } # # Given a pool issue a scrub and verify that no checksum errors are reported. # function verify_pool { typeset pool=${1:-$TESTPOOL} log_must zpool scrub $pool log_must wait_scrubbed $pool typeset -i cksum=$(zpool status $pool | awk ' !NF { isvdev = 0 } isvdev { errors += $NF } /CKSUM$/ { isvdev = 1 } END { print errors } ') if [[ $cksum != 0 ]]; then log_must zpool status -v log_fail "Unexpected CKSUM errors found on $pool ($cksum)" fi } # # Given a pool, and this function list all disks in the pool # function get_disklist # pool { echo $(zpool iostat -v $1 | awk '(NR > 4) {print $1}' | \ grep -vEe '^-----' -e "^(mirror|raidz[1-3]|draid[1-3]|spare|log|cache|special|dedup)|\-[0-9]$") } # # Given a pool, and this function list all disks in the pool with their full # path (like "/dev/sda" instead of "sda"). # function get_disklist_fullpath # pool { get_disklist "-P $1" } # /** # This function kills a given list of processes after a time period. We use # this in the stress tests instead of STF_TIMEOUT so that we can have processes # run for a fixed amount of time, yet still pass. Tests that hit STF_TIMEOUT # would be listed as FAIL, which we don't want : we're happy with stress tests # running for a certain amount of time, then finishing. # # @param $1 the time in seconds after which we should terminate these processes # @param $2..$n the processes we wish to terminate. # */ function stress_timeout { typeset -i TIMEOUT=$1 shift typeset cpids="$@" log_note "Waiting for child processes($cpids). " \ "It could last dozens of minutes, please be patient ..." log_must sleep $TIMEOUT log_note "Killing child processes after ${TIMEOUT} stress timeout." typeset pid for pid in $cpids; do ps -p $pid > /dev/null 2>&1 && log_must kill -USR1 $pid done } # # Verify a given hotspare disk is inuse or avail # # Return 0 is pool/disk matches expected state, 1 otherwise # function check_hotspare_state # pool disk state{inuse,avail} { typeset pool=$1 typeset disk=${2#$DEV_DSKDIR/} typeset state=$3 cur_state=$(get_device_state $pool $disk "spares") [ $state = $cur_state ] } # # Wait until a hotspare transitions to a given state or times out. # # Return 0 when pool/disk matches expected state, 1 on timeout. # function wait_hotspare_state # pool disk state timeout { typeset pool=$1 typeset disk=${2#*$DEV_DSKDIR/} typeset state=$3 typeset timeout=${4:-60} typeset -i i=0 while [[ $i -lt $timeout ]]; do if check_hotspare_state $pool $disk $state; then return 0 fi i=$((i+1)) sleep 1 done return 1 } # # Verify a given vdev disk is inuse or avail # # Return 0 is pool/disk matches expected state, 1 otherwise # function check_vdev_state # pool disk state{online,offline,unavail,removed} { typeset pool=$1 typeset disk=${2#*$DEV_DSKDIR/} typeset state=$3 cur_state=$(get_device_state $pool $disk) [ $state = $cur_state ] } # # Wait until a vdev transitions to a given state or times out. # # Return 0 when pool/disk matches expected state, 1 on timeout. # function wait_vdev_state # pool disk state timeout { typeset pool=$1 typeset disk=${2#*$DEV_DSKDIR/} typeset state=$3 typeset timeout=${4:-60} typeset -i i=0 while [[ $i -lt $timeout ]]; do if check_vdev_state $pool $disk $state; then return 0 fi i=$((i+1)) sleep 1 done return 1 } # # Check the output of 'zpool status -v ', # and to see if the content of contain the specified. # # Return 0 is contain, 1 otherwise # function check_pool_status # pool token keyword { typeset pool=$1 typeset token=$2 typeset keyword=$3 typeset verbose=${4:-false} scan=$(zpool status -v "$pool" 2>/dev/null | awk -v token="$token:" '$1==token') if [[ $verbose == true ]]; then log_note $scan fi echo $scan | grep -qi "$keyword" } # # The following functions are instance of check_pool_status() # is_pool_resilvering - to check if the pool resilver is in progress # is_pool_resilvered - to check if the pool resilver is completed # is_pool_scrubbing - to check if the pool scrub is in progress # is_pool_scrubbed - to check if the pool scrub is completed # is_pool_scrub_stopped - to check if the pool scrub is stopped # is_pool_scrub_paused - to check if the pool scrub has paused # is_pool_removing - to check if the pool removing is a vdev # is_pool_removed - to check if the pool remove is completed # is_pool_discarding - to check if the pool checkpoint is being discarded # is_pool_replacing - to check if the pool is performing a replacement # function is_pool_resilvering #pool { check_pool_status "$1" "scan" \ "resilver[ ()0-9A-Za-z:_-]* in progress since" $2 } function is_pool_resilvered #pool { check_pool_status "$1" "scan" "resilvered " $2 } function is_pool_scrubbing #pool { check_pool_status "$1" "scan" "scrub in progress since " $2 } +function is_pool_error_scrubbing #pool +{ + check_pool_status "$1" "scrub" "error scrub in progress since " $2 + return $? +} + function is_pool_scrubbed #pool { check_pool_status "$1" "scan" "scrub repaired" $2 } function is_pool_scrub_stopped #pool { check_pool_status "$1" "scan" "scrub canceled" $2 } +function is_pool_error_scrub_stopped #pool +{ + check_pool_status "$1" "scrub" "error scrub canceled on " $2 + return $? +} + function is_pool_scrub_paused #pool { check_pool_status "$1" "scan" "scrub paused since " $2 } +function is_pool_error_scrub_paused #pool +{ + check_pool_status "$1" "scrub" "error scrub paused since " $2 + return $? +} + function is_pool_removing #pool { check_pool_status "$1" "remove" "in progress since " } function is_pool_removed #pool { check_pool_status "$1" "remove" "completed on" } function is_pool_discarding #pool { check_pool_status "$1" "checkpoint" "discarding" } function is_pool_replacing #pool { zpool status "$1" | grep -qE 'replacing-[0-9]+' } function wait_for_degraded { typeset pool=$1 typeset timeout=${2:-30} typeset t0=$SECONDS while :; do [[ $(get_pool_prop health $pool) == "DEGRADED" ]] && break log_note "$pool is not yet degraded." sleep 1 if ((SECONDS - t0 > $timeout)); then log_note "$pool not degraded after $timeout seconds." return 1 fi done return 0 } # # Use create_pool()/destroy_pool() to clean up the information in # in the given disk to avoid slice overlapping. # function cleanup_devices #vdevs { typeset pool="foopool$$" for vdev in $@; do zero_partitions $vdev done poolexists $pool && destroy_pool $pool create_pool $pool $@ destroy_pool $pool return 0 } #/** # A function to find and locate free disks on a system or from given # disks as the parameter. It works by locating disks that are in use # as swap devices and dump devices, and also disks listed in /etc/vfstab # # $@ given disks to find which are free, default is all disks in # the test system # # @return a string containing the list of available disks #*/ function find_disks { # Trust provided list, no attempt is made to locate unused devices. if is_linux || is_freebsd; then echo "$@" return fi sfi=/tmp/swaplist.$$ dmpi=/tmp/dumpdev.$$ max_finddisksnum=${MAX_FINDDISKSNUM:-6} swap -l > $sfi dumpadm > $dmpi 2>/dev/null disks=${@:-$(echo "" | format -e 2>/dev/null | awk ' BEGIN { FS="."; } /^Specify disk/{ searchdisks=0; } { if (searchdisks && $2 !~ "^$"){ split($2,arr," "); print arr[1]; } } /^AVAILABLE DISK SELECTIONS:/{ searchdisks=1; } ')} unused="" for disk in $disks; do # Check for mounted grep -q "${disk}[sp]" /etc/mnttab && continue # Check for swap grep -q "${disk}[sp]" $sfi && continue # check for dump device grep -q "${disk}[sp]" $dmpi && continue # check to see if this disk hasn't been explicitly excluded # by a user-set environment variable echo "${ZFS_HOST_DEVICES_IGNORE}" | grep -q "${disk}" && continue unused_candidates="$unused_candidates $disk" done rm $sfi $dmpi # now just check to see if those disks do actually exist # by looking for a device pointing to the first slice in # each case. limit the number to max_finddisksnum count=0 for disk in $unused_candidates; do if is_disk_device $DEV_DSKDIR/${disk}s0 && \ [ $count -lt $max_finddisksnum ]; then unused="$unused $disk" # do not impose limit if $@ is provided [[ -z $@ ]] && ((count = count + 1)) fi done # finally, return our disk list echo $unused } function add_user_freebsd # { typeset group=$1 typeset user=$2 typeset basedir=$3 # Check to see if the user exists. if id $user > /dev/null 2>&1; then return 0 fi # Assign 1000 as the base uid typeset -i uid=1000 while true; do pw useradd -u $uid -g $group -d $basedir/$user -m -n $user case $? in 0) break ;; # The uid is not unique 65) ((uid += 1)) ;; *) return 1 ;; esac if [[ $uid == 65000 ]]; then log_fail "No user id available under 65000 for $user" fi done # Silence MOTD touch $basedir/$user/.hushlogin return 0 } # # Delete the specified user. # # $1 login name # function del_user_freebsd # { typeset user=$1 if id $user > /dev/null 2>&1; then log_must pw userdel $user fi return 0 } # # Select valid gid and create specified group. # # $1 group name # function add_group_freebsd # { typeset group=$1 # See if the group already exists. if pw groupshow $group >/dev/null 2>&1; then return 0 fi # Assign 1000 as the base gid typeset -i gid=1000 while true; do pw groupadd -g $gid -n $group > /dev/null 2>&1 case $? in 0) return 0 ;; # The gid is not unique 65) ((gid += 1)) ;; *) return 1 ;; esac if [[ $gid == 65000 ]]; then log_fail "No user id available under 65000 for $group" fi done } # # Delete the specified group. # # $1 group name # function del_group_freebsd # { typeset group=$1 pw groupdel -n $group > /dev/null 2>&1 case $? in # Group does not exist, or was deleted successfully. 0|6|65) return 0 ;; # Name already exists as a group name 9) log_must pw groupdel $group ;; *) return 1 ;; esac return 0 } function add_user_illumos # { typeset group=$1 typeset user=$2 typeset basedir=$3 log_must useradd -g $group -d $basedir/$user -m $user return 0 } function del_user_illumos # { typeset user=$1 if id $user > /dev/null 2>&1; then log_must_retry "currently used" 6 userdel $user fi return 0 } function add_group_illumos # { typeset group=$1 typeset -i gid=100 while true; do groupadd -g $gid $group > /dev/null 2>&1 case $? in 0) return 0 ;; # The gid is not unique 4) ((gid += 1)) ;; *) return 1 ;; esac done } function del_group_illumos # { typeset group=$1 groupmod -n $grp $grp > /dev/null 2>&1 case $? in # Group does not exist. 6) return 0 ;; # Name already exists as a group name 9) log_must groupdel $grp ;; *) return 1 ;; esac } function add_user_linux # { typeset group=$1 typeset user=$2 typeset basedir=$3 log_must useradd -g $group -d $basedir/$user -m $user # Add new users to the same group and the command line utils. # This allows them to be run out of the original users home # directory as long as it permissioned to be group readable. cmd_group=$(stat --format="%G" $(command -v zfs)) log_must usermod -a -G $cmd_group $user return 0 } function del_user_linux # { typeset user=$1 if id $user > /dev/null 2>&1; then log_must_retry "currently used" 6 userdel $user fi } function add_group_linux # { typeset group=$1 # Assign 100 as the base gid, a larger value is selected for # Linux because for many distributions 1000 and under are reserved. while true; do groupadd $group > /dev/null 2>&1 case $? in 0) return 0 ;; *) return 1 ;; esac done } function del_group_linux # { typeset group=$1 getent group $group > /dev/null 2>&1 case $? in # Group does not exist. 2) return 0 ;; # Name already exists as a group name 0) log_must groupdel $group ;; *) return 1 ;; esac return 0 } # # Add specified user to specified group # # $1 group name # $2 user name # $3 base of the homedir (optional) # function add_user # { typeset group=$1 typeset user=$2 typeset basedir=${3:-"/var/tmp"} if ((${#group} == 0 || ${#user} == 0)); then log_fail "group name or user name are not defined." fi case "$UNAME" in FreeBSD) add_user_freebsd "$group" "$user" "$basedir" ;; Linux) add_user_linux "$group" "$user" "$basedir" ;; *) add_user_illumos "$group" "$user" "$basedir" ;; esac return 0 } # # Delete the specified user. # # $1 login name # $2 base of the homedir (optional) # function del_user # { typeset user=$1 typeset basedir=${2:-"/var/tmp"} if ((${#user} == 0)); then log_fail "login name is necessary." fi case "$UNAME" in FreeBSD) del_user_freebsd "$user" ;; Linux) del_user_linux "$user" ;; *) del_user_illumos "$user" ;; esac [[ -d $basedir/$user ]] && rm -fr $basedir/$user return 0 } # # Select valid gid and create specified group. # # $1 group name # function add_group # { typeset group=$1 if ((${#group} == 0)); then log_fail "group name is necessary." fi case "$UNAME" in FreeBSD) add_group_freebsd "$group" ;; Linux) add_group_linux "$group" ;; *) add_group_illumos "$group" ;; esac return 0 } # # Delete the specified group. # # $1 group name # function del_group # { typeset group=$1 if ((${#group} == 0)); then log_fail "group name is necessary." fi case "$UNAME" in FreeBSD) del_group_freebsd "$group" ;; Linux) del_group_linux "$group" ;; *) del_group_illumos "$group" ;; esac return 0 } # # This function will return true if it's safe to destroy the pool passed # as argument 1. It checks for pools based on zvols and files, and also # files contained in a pool that may have a different mountpoint. # function safe_to_destroy_pool { # $1 the pool name typeset pool="" typeset DONT_DESTROY="" # We check that by deleting the $1 pool, we're not # going to pull the rug out from other pools. Do this # by looking at all other pools, ensuring that they # aren't built from files or zvols contained in this pool. for pool in $(zpool list -H -o name) do ALTMOUNTPOOL="" # this is a list of the top-level directories in each of the # files that make up the path to the files the pool is based on FILEPOOL=$(zpool status -v $pool | awk -v pool="/$1/" '$0 ~ pool {print $1}') # this is a list of the zvols that make up the pool ZVOLPOOL=$(zpool status -v $pool | awk -v zvols="$ZVOL_DEVDIR/$1$" '$0 ~ zvols {print $1}') # also want to determine if it's a file-based pool using an # alternate mountpoint... POOL_FILE_DIRS=$(zpool status -v $pool | \ awk '/\// {print $1}' | \ awk -F/ '!/dev/ {print $2}') for pooldir in $POOL_FILE_DIRS do OUTPUT=$(zfs list -H -r -o mountpoint $1 | \ awk -v pd="${pooldir}$" '$0 ~ pd {print $1}') ALTMOUNTPOOL="${ALTMOUNTPOOL}${OUTPUT}" done if [ ! -z "$ZVOLPOOL" ] then DONT_DESTROY="true" log_note "Pool $pool is built from $ZVOLPOOL on $1" fi if [ ! -z "$FILEPOOL" ] then DONT_DESTROY="true" log_note "Pool $pool is built from $FILEPOOL on $1" fi if [ ! -z "$ALTMOUNTPOOL" ] then DONT_DESTROY="true" log_note "Pool $pool is built from $ALTMOUNTPOOL on $1" fi done if [ -z "${DONT_DESTROY}" ] then return 0 else log_note "Warning: it is not safe to destroy $1!" return 1 fi } # # Verify zfs operation with -p option work as expected # $1 operation, value could be create, clone or rename # $2 dataset type, value could be fs or vol # $3 dataset name # $4 new dataset name # function verify_opt_p_ops { typeset ops=$1 typeset datatype=$2 typeset dataset=$3 typeset newdataset=$4 if [[ $datatype != "fs" && $datatype != "vol" ]]; then log_fail "$datatype is not supported." fi # check parameters accordingly case $ops in create) newdataset=$dataset dataset="" if [[ $datatype == "vol" ]]; then ops="create -V $VOLSIZE" fi ;; clone) if [[ -z $newdataset ]]; then log_fail "newdataset should not be empty" \ "when ops is $ops." fi log_must datasetexists $dataset log_must snapexists $dataset ;; rename) if [[ -z $newdataset ]]; then log_fail "newdataset should not be empty" \ "when ops is $ops." fi log_must datasetexists $dataset ;; *) log_fail "$ops is not supported." ;; esac # make sure the upper level filesystem does not exist destroy_dataset "${newdataset%/*}" "-rRf" # without -p option, operation will fail log_mustnot zfs $ops $dataset $newdataset log_mustnot datasetexists $newdataset ${newdataset%/*} # with -p option, operation should succeed log_must zfs $ops -p $dataset $newdataset block_device_wait if ! datasetexists $newdataset ; then log_fail "-p option does not work for $ops" fi # when $ops is create or clone, redo the operation still return zero if [[ $ops != "rename" ]]; then log_must zfs $ops -p $dataset $newdataset fi return 0 } # # Get configuration of pool # $1 pool name # $2 config name # function get_config { typeset pool=$1 typeset config=$2 if ! poolexists "$pool" ; then return 1 fi if [ "$(get_pool_prop cachefile "$pool")" = "none" ]; then zdb -e $pool else zdb -C $pool fi | awk -F: -v cfg="$config:" '$0 ~ cfg {sub(/^'\''/, $2); sub(/'\''$/, $2); print $2}' } # # Privated function. Random select one of items from arguments. # # $1 count # $2-n string # function _random_get { typeset cnt=$1 shift typeset str="$@" typeset -i ind ((ind = RANDOM % cnt + 1)) echo "$str" | cut -f $ind -d ' ' } # # Random select one of item from arguments which include NONE string # function random_get_with_non { typeset -i cnt=$# ((cnt =+ 1)) _random_get "$cnt" "$@" } # # Random select one of item from arguments which doesn't include NONE string # function random_get { _random_get "$#" "$@" } # # The function will generate a dataset name with specific length # $1, the length of the name # $2, the base string to construct the name # function gen_dataset_name { typeset -i len=$1 typeset basestr="$2" typeset -i baselen=${#basestr} typeset -i iter=0 typeset l_name="" if ((len % baselen == 0)); then ((iter = len / baselen)) else ((iter = len / baselen + 1)) fi while ((iter > 0)); do l_name="${l_name}$basestr" ((iter -= 1)) done echo $l_name } # # Get cksum tuple of dataset # $1 dataset name # # sample zdb output: # Dataset data/test [ZPL], ID 355, cr_txg 2413856, 31.0K, 7 objects, rootbp # DVA[0]=<0:803046400:200> DVA[1]=<0:81199000:200> [L0 DMU objset] fletcher4 # lzjb LE contiguous unique double size=800L/200P birth=2413856L/2413856P # fill=7 cksum=11ce125712:643a9c18ee2:125e25238fca0:254a3f74b59744 function datasetcksum { typeset cksum sync sync_all_pools zdb -vvv $1 | awk -F= -v ds="^Dataset $1 "'\\[' '$0 ~ ds && /cksum/ {print $7}' } # # Get the given disk/slice state from the specific field of the pool # function get_device_state #pool disk field("", "spares","logs") { typeset pool=$1 typeset disk=${2#$DEV_DSKDIR/} typeset field=${3:-$pool} zpool status -v "$pool" 2>/dev/null | \ awk -v device=$disk -v pool=$pool -v field=$field \ 'BEGIN {startconfig=0; startfield=0; } /config:/ {startconfig=1} (startconfig==1) && ($1==field) {startfield=1; next;} (startfield==1) && ($1==device) {print $2; exit;} (startfield==1) && ($1==field || $1 ~ "^spares$" || $1 ~ "^logs$") {startfield=0}' } # # get the root filesystem name if it's zfsroot system. # # return: root filesystem name function get_rootfs { typeset rootfs="" if is_freebsd; then rootfs=$(mount -p | awk '$2 == "/" && $3 == "zfs" {print $1}') elif ! is_linux; then rootfs=$(awk '$2 == "/" && $3 == "zfs" {print $1}' \ /etc/mnttab) fi if [[ -z "$rootfs" ]]; then log_fail "Can not get rootfs" fi if datasetexists $rootfs; then echo $rootfs else log_fail "This is not a zfsroot system." fi } # # get the rootfs's pool name # return: # rootpool name # function get_rootpool { typeset rootfs=$(get_rootfs) echo ${rootfs%%/*} } # # To verify if the require numbers of disks is given # function verify_disk_count { typeset -i min=${2:-1} typeset -i count=$(echo "$1" | wc -w) if ((count < min)); then log_untested "A minimum of $min disks is required to run." \ " You specified $count disk(s)" fi } function ds_is_volume { typeset type=$(get_prop type $1) [ $type = "volume" ] } function ds_is_filesystem { typeset type=$(get_prop type $1) [ $type = "filesystem" ] } # # Check if Trusted Extensions are installed and enabled # function is_te_enabled { svcs -H -o state labeld 2>/dev/null | grep -q "enabled" } # Return the number of CPUs (cross-platform) function get_num_cpus { if is_linux ; then grep -c '^processor' /proc/cpuinfo elif is_freebsd; then sysctl -n kern.smp.cpus else psrinfo | wc -l fi } # Utility function to determine if a system has multiple cpus. function is_mp { [[ $(get_num_cpus) -gt 1 ]] } function get_cpu_freq { if is_linux; then lscpu | awk '/CPU MHz/ { print $3 }' elif is_freebsd; then sysctl -n hw.clockrate else psrinfo -v 0 | awk '/processor operates at/ {print $6}' fi } # Run the given command as the user provided. function user_run { typeset user=$1 shift log_note "user: $user" log_note "cmd: $*" typeset out=$TEST_BASE_DIR/out typeset err=$TEST_BASE_DIR/err sudo -Eu $user env PATH="$PATH" ksh <<<"$*" >$out 2>$err typeset res=$? log_note "out: $(<$out)" log_note "err: $(<$err)" return $res } # # Check if the pool contains the specified vdevs # # $1 pool # $2..n ... # # Return 0 if the vdevs are contained in the pool, 1 if any of the specified # vdevs is not in the pool, and 2 if pool name is missing. # function vdevs_in_pool { typeset pool=$1 typeset vdev if [[ -z $pool ]]; then log_note "Missing pool name." return 2 fi shift # We could use 'zpool list' to only get the vdevs of the pool but we # can't reference a mirror/raidz vdev using its ID (i.e mirror-0), # therefore we use the 'zpool status' output. typeset tmpfile=$(mktemp) zpool status -v "$pool" | grep -A 1000 "config:" >$tmpfile for vdev in "$@"; do grep -wq ${vdev##*/} $tmpfile || return 1 done rm -f $tmpfile return 0 } function get_max { typeset -l i max=$1 shift for i in "$@"; do max=$((max > i ? max : i)) done echo $max } # Write data that can be compressed into a directory function write_compressible { typeset dir=$1 typeset megs=$2 typeset nfiles=${3:-1} typeset bs=${4:-1024k} typeset fname=${5:-file} [[ -d $dir ]] || log_fail "No directory: $dir" # Under Linux fio is not currently used since its behavior can # differ significantly across versions. This includes missing # command line options and cases where the --buffer_compress_* # options fail to behave as expected. if is_linux; then typeset file_bytes=$(to_bytes $megs) typeset bs_bytes=4096 typeset blocks=$(($file_bytes / $bs_bytes)) for (( i = 0; i < $nfiles; i++ )); do truncate -s $file_bytes $dir/$fname.$i # Write every third block to get 66% compression. for (( j = 0; j < $blocks; j += 3 )); do dd if=/dev/urandom of=$dir/$fname.$i \ seek=$j bs=$bs_bytes count=1 \ conv=notrunc >/dev/null 2>&1 done done else command -v fio > /dev/null || log_unsupported "fio missing" log_must eval fio \ --name=job \ --fallocate=0 \ --minimal \ --randrepeat=0 \ --buffer_compress_percentage=66 \ --buffer_compress_chunk=4096 \ --directory="$dir" \ --numjobs="$nfiles" \ --nrfiles="$nfiles" \ --rw=write \ --bs="$bs" \ --filesize="$megs" \ "--filename_format='$fname.\$jobnum' >/dev/null" fi } function get_objnum { typeset pathname=$1 typeset objnum [[ -e $pathname ]] || log_fail "No such file or directory: $pathname" if is_freebsd; then objnum=$(stat -f "%i" $pathname) else objnum=$(stat -c %i $pathname) fi echo $objnum } # # Sync data to the pool # # $1 pool name # $2 boolean to force uberblock (and config including zpool cache file) update # function sync_pool #pool { typeset pool=${1:-$TESTPOOL} typeset force=${2:-false} if [[ $force == true ]]; then log_must zpool sync -f $pool else log_must zpool sync $pool fi return 0 } # # Sync all pools # # $1 boolean to force uberblock (and config including zpool cache file) update # function sync_all_pools # { typeset force=${1:-false} if [[ $force == true ]]; then log_must zpool sync -f else log_must zpool sync fi return 0 } # # Wait for zpool 'freeing' property drops to zero. # # $1 pool name # function wait_freeing #pool { typeset pool=${1:-$TESTPOOL} while true; do [[ "0" == "$(zpool list -Ho freeing $pool)" ]] && break log_must sleep 1 done } # # Wait for every device replace operation to complete # # $1 pool name # $2 timeout # function wait_replacing #pool timeout { typeset timeout=${2:-300} typeset pool=${1:-$TESTPOOL} for (( timer = 0; timer < $timeout; timer++ )); do is_pool_replacing $pool || break; sleep 1; done } # Wait for a pool to be scrubbed # # $1 pool name # $2 timeout # function wait_scrubbed #pool timeout { typeset timeout=${2:-300} typeset pool=${1:-$TESTPOOL} for (( timer = 0; timer < $timeout; timer++ )); do is_pool_scrubbed $pool && break; sleep 1; done } # Backup the zed.rc in our test directory so that we can edit it for our test. # # Returns: Backup file name. You will need to pass this to zed_rc_restore(). function zed_rc_backup { zedrc_backup="$(mktemp)" cp $ZEDLET_DIR/zed.rc $zedrc_backup echo $zedrc_backup } function zed_rc_restore { mv $1 $ZEDLET_DIR/zed.rc } # # Setup custom environment for the ZED. # # $@ Optional list of zedlets to run under zed. function zed_setup { if ! is_linux; then log_unsupported "No zed on $UNAME" fi if [[ ! -d $ZEDLET_DIR ]]; then log_must mkdir $ZEDLET_DIR fi if [[ ! -e $VDEVID_CONF ]]; then log_must touch $VDEVID_CONF fi if [[ -e $VDEVID_CONF_ETC ]]; then log_fail "Must not have $VDEVID_CONF_ETC file present on system" fi EXTRA_ZEDLETS=$@ # Create a symlink for /etc/zfs/vdev_id.conf file. log_must ln -s $VDEVID_CONF $VDEVID_CONF_ETC # Setup minimal ZED configuration. Individual test cases should # add additional ZEDLETs as needed for their specific test. log_must cp ${ZEDLET_ETC_DIR}/zed.rc $ZEDLET_DIR log_must cp ${ZEDLET_ETC_DIR}/zed-functions.sh $ZEDLET_DIR # Scripts must only be user writable. if [[ -n "$EXTRA_ZEDLETS" ]] ; then saved_umask=$(umask) log_must umask 0022 for i in $EXTRA_ZEDLETS ; do log_must cp ${ZEDLET_LIBEXEC_DIR}/$i $ZEDLET_DIR done log_must umask $saved_umask fi # Customize the zed.rc file to enable the full debug log. log_must sed -i '/\#ZED_DEBUG_LOG=.*/d' $ZEDLET_DIR/zed.rc echo "ZED_DEBUG_LOG=$ZED_DEBUG_LOG" >>$ZEDLET_DIR/zed.rc } # # Cleanup custom ZED environment. # # $@ Optional list of zedlets to remove from our test zed.d directory. function zed_cleanup { if ! is_linux; then return fi for extra_zedlet; do log_must rm -f ${ZEDLET_DIR}/$extra_zedlet done log_must rm -fd ${ZEDLET_DIR}/zed.rc ${ZEDLET_DIR}/zed-functions.sh ${ZEDLET_DIR}/all-syslog.sh ${ZEDLET_DIR}/all-debug.sh ${ZEDLET_DIR}/state \ $ZED_LOG $ZED_DEBUG_LOG $VDEVID_CONF_ETC $VDEVID_CONF \ $ZEDLET_DIR } # # Check if ZED is currently running; if so, returns PIDs # function zed_check { if ! is_linux; then return fi zedpids="$(pgrep -x zed)" zedpids2="$(pgrep -x lt-zed)" echo ${zedpids} ${zedpids2} } # # Check if ZED is currently running, if not start ZED. # function zed_start { if ! is_linux; then return fi # ZEDLET_DIR=/var/tmp/zed if [[ ! -d $ZEDLET_DIR ]]; then log_must mkdir $ZEDLET_DIR fi # Verify the ZED is not already running. zedpids=$(zed_check) if [ -n "$zedpids" ]; then # We never, ever, really want it to just keep going if zed # is already running - usually this implies our test cases # will break very strangely because whatever we wanted to # configure zed for won't be listening to our changes in the # tmpdir log_fail "ZED already running - ${zedpids}" else log_note "Starting ZED" # run ZED in the background and redirect foreground logging # output to $ZED_LOG. log_must truncate -s 0 $ZED_DEBUG_LOG log_must eval "zed -vF -d $ZEDLET_DIR -P $PATH" \ "-s $ZEDLET_DIR/state -j 1 2>$ZED_LOG &" fi return 0 } # # Kill ZED process # function zed_stop { if ! is_linux; then return "" fi log_note "Stopping ZED" while true; do zedpids=$(zed_check) [ ! -n "$zedpids" ] && break log_must kill $zedpids sleep 1 done return 0 } # # Drain all zevents # function zed_events_drain { while [ $(zpool events -H | wc -l) -ne 0 ]; do sleep 1 zpool events -c >/dev/null done } # Set a variable in zed.rc to something, un-commenting it in the process. # # $1 variable # $2 value function zed_rc_set { var="$1" val="$2" # Remove the line cmd="'/$var/d'" eval sed -i $cmd $ZEDLET_DIR/zed.rc # Add it at the end echo "$var=$val" >> $ZEDLET_DIR/zed.rc } # # Check is provided device is being active used as a swap device. # function is_swap_inuse { typeset device=$1 if [[ -z $device ]] ; then log_note "No device specified." return 1 fi case "$UNAME" in Linux) swapon -s | grep -wq $(readlink -f $device) ;; FreeBSD) swapctl -l | grep -wq $device ;; *) swap -l | grep -wq $device ;; esac } # # Setup a swap device using the provided device. # function swap_setup { typeset swapdev=$1 case "$UNAME" in Linux) log_must eval "mkswap $swapdev > /dev/null 2>&1" log_must swapon $swapdev ;; FreeBSD) log_must swapctl -a $swapdev ;; *) log_must swap -a $swapdev ;; esac return 0 } # # Cleanup a swap device on the provided device. # function swap_cleanup { typeset swapdev=$1 if is_swap_inuse $swapdev; then if is_linux; then log_must swapoff $swapdev elif is_freebsd; then log_must swapoff $swapdev else log_must swap -d $swapdev fi fi return 0 } # # Set a global system tunable (64-bit value) # # $1 tunable name (use a NAME defined in tunables.cfg) # $2 tunable values # function set_tunable64 { set_tunable_impl "$1" "$2" Z } # # Set a global system tunable (32-bit value) # # $1 tunable name (use a NAME defined in tunables.cfg) # $2 tunable values # function set_tunable32 { set_tunable_impl "$1" "$2" W } function set_tunable_impl { typeset name="$1" typeset value="$2" typeset mdb_cmd="$3" eval "typeset tunable=\$$name" case "$tunable" in UNSUPPORTED) log_unsupported "Tunable '$name' is unsupported on $UNAME" ;; "") log_fail "Tunable '$name' must be added to tunables.cfg" ;; *) ;; esac [[ -z "$value" ]] && return 1 [[ -z "$mdb_cmd" ]] && return 1 case "$UNAME" in Linux) typeset zfs_tunables="/sys/module/zfs/parameters" echo "$value" >"$zfs_tunables/$tunable" ;; FreeBSD) sysctl vfs.zfs.$tunable=$value ;; SunOS) echo "${tunable}/${mdb_cmd}0t${value}" | mdb -kw ;; esac } # # Get a global system tunable # # $1 tunable name (use a NAME defined in tunables.cfg) # function get_tunable { get_tunable_impl "$1" } function get_tunable_impl { typeset name="$1" typeset module="${2:-zfs}" typeset check_only="$3" eval "typeset tunable=\$$name" case "$tunable" in UNSUPPORTED) if [ -z "$check_only" ] ; then log_unsupported "Tunable '$name' is unsupported on $UNAME" else return 1 fi ;; "") if [ -z "$check_only" ] ; then log_fail "Tunable '$name' must be added to tunables.cfg" else return 1 fi ;; *) ;; esac case "$UNAME" in Linux) typeset zfs_tunables="/sys/module/$module/parameters" cat $zfs_tunables/$tunable ;; FreeBSD) sysctl -n vfs.zfs.$tunable ;; SunOS) [[ "$module" -eq "zfs" ]] || return 1 ;; esac } # Does a tunable exist? # # $1: Tunable name function tunable_exists { get_tunable_impl $1 "zfs" 1 } # # Compute MD5 digest for given file or stdin if no file given. # Note: file path must not contain spaces # function md5digest { typeset file=$1 case "$UNAME" in FreeBSD) md5 -q $file ;; *) typeset sum _ read -r sum _ < <(md5sum -b $file) echo $sum ;; esac } # # Compute SHA256 digest for given file or stdin if no file given. # Note: file path must not contain spaces # function sha256digest { typeset file=$1 case "$UNAME" in FreeBSD) sha256 -q $file ;; *) typeset sum _ read -r sum _ < <(sha256sum -b $file) echo $sum ;; esac } function new_fs # { case "$UNAME" in FreeBSD) newfs "$@" ;; *) echo y | newfs -v "$@" ;; esac } function stat_size # { typeset path=$1 case "$UNAME" in FreeBSD) stat -f %z "$path" ;; *) stat -c %s "$path" ;; esac } function stat_mtime # { typeset path=$1 case "$UNAME" in FreeBSD) stat -f %m "$path" ;; *) stat -c %Y "$path" ;; esac } function stat_ctime # { typeset path=$1 case "$UNAME" in FreeBSD) stat -f %c "$path" ;; *) stat -c %Z "$path" ;; esac } function stat_crtime # { typeset path=$1 case "$UNAME" in FreeBSD) stat -f %B "$path" ;; *) stat -c %W "$path" ;; esac } function stat_generation # { typeset path=$1 case "$UNAME" in Linux) getversion "${path}" ;; *) stat -f %v "${path}" ;; esac } # Run a command as if it was being run in a TTY. # # Usage: # # faketty command # function faketty { if is_freebsd; then script -q /dev/null env "$@" else script --return --quiet -c "$*" /dev/null fi } # # Produce a random permutation of the integers in a given range (inclusive). # function range_shuffle # begin end { typeset -i begin=$1 typeset -i end=$2 seq ${begin} ${end} | sort -R } # # Cross-platform xattr helpers # function get_xattr # name path { typeset name=$1 typeset path=$2 case "$UNAME" in FreeBSD) getextattr -qq user "${name}" "${path}" ;; *) attr -qg "${name}" "${path}" ;; esac } function set_xattr # name value path { typeset name=$1 typeset value=$2 typeset path=$3 case "$UNAME" in FreeBSD) setextattr user "${name}" "${value}" "${path}" ;; *) attr -qs "${name}" -V "${value}" "${path}" ;; esac } function set_xattr_stdin # name value { typeset name=$1 typeset path=$2 case "$UNAME" in FreeBSD) setextattr -i user "${name}" "${path}" ;; *) attr -qs "${name}" "${path}" ;; esac } function rm_xattr # name path { typeset name=$1 typeset path=$2 case "$UNAME" in FreeBSD) rmextattr -q user "${name}" "${path}" ;; *) attr -qr "${name}" "${path}" ;; esac } function ls_xattr # path { typeset path=$1 case "$UNAME" in FreeBSD) lsextattr -qq user "${path}" ;; *) attr -ql "${path}" ;; esac } function kstat # stat flags? { typeset stat=$1 typeset flags=${2-"-n"} case "$UNAME" in FreeBSD) sysctl $flags kstat.zfs.misc.$stat ;; Linux) cat "/proc/spl/kstat/zfs/$stat" 2>/dev/null ;; *) false ;; esac } function get_arcstat # stat { typeset stat=$1 case "$UNAME" in FreeBSD) kstat arcstats.$stat ;; Linux) kstat arcstats | awk "/$stat/"' { print $3 }' ;; *) false ;; esac } function punch_hole # offset length file { typeset offset=$1 typeset length=$2 typeset file=$3 case "$UNAME" in FreeBSD) truncate -d -o $offset -l $length "$file" ;; Linux) fallocate --punch-hole --offset $offset --length $length "$file" ;; *) false ;; esac } function zero_range # offset length file { typeset offset=$1 typeset length=$2 typeset file=$3 case "$UNAME" in Linux) fallocate --zero-range --offset $offset --length $length "$file" ;; *) false ;; esac } # # Wait for the specified arcstat to reach non-zero quiescence. # If echo is 1 echo the value after reaching quiescence, otherwise # if echo is 0 print the arcstat we are waiting on. # function arcstat_quiescence # stat echo { typeset stat=$1 typeset echo=$2 typeset do_once=true if [[ $echo -eq 0 ]]; then echo "Waiting for arcstat $1 quiescence." fi while $do_once || [ $stat1 -ne $stat2 ] || [ $stat2 -eq 0 ]; do typeset stat1=$(get_arcstat $stat) sleep 2 typeset stat2=$(get_arcstat $stat) do_once=false done if [[ $echo -eq 1 ]]; then echo $stat2 fi } function arcstat_quiescence_noecho # stat { typeset stat=$1 arcstat_quiescence $stat 0 } function arcstat_quiescence_echo # stat { typeset stat=$1 arcstat_quiescence $stat 1 } # # Given an array of pids, wait until all processes # have completed and check their return status. # function wait_for_children #children { rv=0 children=("$@") for child in "${children[@]}" do child_exit=0 wait ${child} || child_exit=$? if [ $child_exit -ne 0 ]; then echo "child ${child} failed with ${child_exit}" rv=1 fi done return $rv } # # Compare two directory trees recursively in a manner similar to diff(1), but # using rsync. If there are any discrepancies, a summary of the differences are # output and a non-zero error is returned. # # If you're comparing a directory after a ZIL replay, you should set # LIBTEST_DIFF_ZIL_REPLAY=1 or use replay_directory_diff which will cause # directory_diff to ignore mtime changes (the ZIL replay won't fix up mtime # information). # function directory_diff # dir_a dir_b { dir_a="$1" dir_b="$2" zil_replay="${LIBTEST_DIFF_ZIL_REPLAY:-0}" # If one of the directories doesn't exist, return 2. This is to match the # semantics of diff. if ! [ -d "$dir_a" -a -d "$dir_b" ]; then return 2 fi # Run rsync with --dry-run --itemize-changes to get something akin to diff # output, but rsync is far more thorough in detecting differences (diff # doesn't compare file metadata, and cannot handle special files). # # Also make sure to filter out non-user.* xattrs when comparing. On # SELinux-enabled systems the copied tree will probably have different # SELinux labels. args=("-nicaAHX" '--filter=-x! user.*' "--delete") # NOTE: Quite a few rsync builds do not support --crtimes which would be # necessary to verify that creation times are being maintained properly. # Unfortunately because of this we cannot use it unconditionally but we can # check if this rsync build supports it and use it then. This check is # based on the same check in the rsync test suite (testsuite/crtimes.test). # # We check ctimes even with zil_replay=1 because the ZIL does store # creation times and we should make sure they match (if the creation times # do not match there is a "c" entry in one of the columns). if rsync --version | grep -q "[, ] crtimes"; then args+=("--crtimes") else log_note "This rsync package does not support --crtimes (-N)." fi # If we are testing a ZIL replay, we need to ignore timestamp changes. # Unfortunately --no-times doesn't do what we want -- it will still tell # you if the timestamps don't match but rsync will set the timestamps to # the current time (leading to an itemised change entry). It's simpler to # just filter out those lines. if [ "$zil_replay" -eq 0 ]; then filter=("cat") else # Different rsync versions have different numbers of columns. So just # require that aside from the first two, all other columns must be # blank (literal ".") or a timestamp field ("[tT]"). filter=("grep" "-v" '^\..[.Tt]\+ ') fi diff="$(rsync "${args[@]}" "$dir_a/" "$dir_b/" | "${filter[@]}")" rv=0 if [ -n "$diff" ]; then echo "$diff" rv=1 fi return $rv } # # Compare two directory trees recursively, without checking whether the mtimes # match (creation times will be checked if the available rsync binary supports # it). This is necessary for ZIL replay checks (because the ZIL does not # contain mtimes and thus after a ZIL replay, mtimes won't match). # # This is shorthand for LIBTEST_DIFF_ZIL_REPLAY=1 directory_diff <...>. # function replay_directory_diff # dir_a dir_b { LIBTEST_DIFF_ZIL_REPLAY=1 directory_diff "$@" } # # Put coredumps into $1/core.{basename} # # Output must be saved and passed to pop_coredump_pattern on cleanup # function push_coredump_pattern # dir { ulimit -c unlimited case "$UNAME" in Linux) cat /proc/sys/kernel/core_pattern /proc/sys/kernel/core_uses_pid echo "$1/core.%e" >/proc/sys/kernel/core_pattern && echo 0 >/proc/sys/kernel/core_uses_pid ;; FreeBSD) sysctl -n kern.corefile sysctl kern.corefile="$1/core.%N" >/dev/null ;; *) # Nothing to output – set only for this shell coreadm -p "$1/core.%f" ;; esac } # # Put coredumps back into the default location # function pop_coredump_pattern { [ -s "$1" ] || return 0 case "$UNAME" in Linux) typeset pat pid { read -r pat; read -r pid; } < "$1" echo "$pat" >/proc/sys/kernel/core_pattern && echo "$pid" >/proc/sys/kernel/core_uses_pid ;; FreeBSD) sysctl kern.corefile="$(<"$1")" >/dev/null ;; esac } diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/Makefile.am b/sys/contrib/openzfs/tests/zfs-tests/tests/Makefile.am index 9299a4ca9b47..ad4aec543299 100644 --- a/sys/contrib/openzfs/tests/zfs-tests/tests/Makefile.am +++ b/sys/contrib/openzfs/tests/zfs-tests/tests/Makefile.am @@ -1,2043 +1,2050 @@ CLEANFILES = dist_noinst_DATA = include $(top_srcdir)/config/Substfiles.am datadir_zfs_tests_testsdir = $(datadir)/$(PACKAGE)/zfs-tests/tests nobase_dist_datadir_zfs_tests_tests_DATA = \ perf/nfs-sample.cfg \ perf/perf.shlib \ \ perf/fio/mkfiles.fio \ perf/fio/random_reads.fio \ perf/fio/random_readwrite.fio \ perf/fio/random_readwrite_fixed.fio \ perf/fio/random_writes.fio \ perf/fio/sequential_reads.fio \ perf/fio/sequential_readwrite.fio \ perf/fio/sequential_writes.fio nobase_dist_datadir_zfs_tests_tests_SCRIPTS = \ perf/regression/random_reads.ksh \ perf/regression/random_readwrite.ksh \ perf/regression/random_readwrite_fixed.ksh \ perf/regression/random_writes.ksh \ perf/regression/random_writes_zil.ksh \ perf/regression/sequential_reads_arc_cached_clone.ksh \ perf/regression/sequential_reads_arc_cached.ksh \ perf/regression/sequential_reads_dbuf_cached.ksh \ perf/regression/sequential_reads.ksh \ perf/regression/sequential_writes.ksh \ perf/regression/setup.ksh \ \ perf/scripts/prefetch_io.sh # These lists can be regenerated by running make regen-tests at the root, or, on a *clean* source: # find functional/ ! -type d ! -name .gitignore ! -name .dirstamp ! -name '*.Po' ! -executable -name '*.in' | sort | sed 's/\.in$//;s/^/\t/;$!s/$/ \\/' # find functional/ ! -type d ! -name .gitignore ! -name .dirstamp ! -name '*.Po' -executable -name '*.in' | sort | sed 's/\.in$//;s/^/\t/;$!s/$/ \\/' # find functional/ ! -type d ! -name .gitignore ! -name .dirstamp ! -name '*.Po' ! -name '*.in' ! -name '*.c' | grep -Fe /simd -e /tmpfile | sort | sed 's/^/\t/;$!s/$/ \\/' # find functional/ ! -type d ! -name .gitignore ! -name .dirstamp ! -name '*.Po' ! -executable ! -name '*.in' ! -name '*.c' | grep -vFe /simd -e /tmpfile | sort | sed 's/^/\t/;$!s/$/ \\/' # find functional/ ! -type d ! -name .gitignore ! -name .dirstamp ! -name '*.Po' -executable ! -name '*.in' ! -name '*.c' | grep -vFe /simd -e /tmpfile | sort | sed 's/^/\t/;$!s/$/ \\/' # # simd and tmpfile are Linux-only and not installed elsewhere # # C programs are specced in ../Makefile.am above as part of the main Makefile find_common := find functional/ ! -type d ! -name .gitignore ! -name .dirstamp ! -name '*.Po' regen: @$(MAKE) -C $(top_builddir) clean @$(MAKE) clean $(SED) $(ac_inplace) '/^# -- >8 --/q' Makefile.am echo >> Makefile.am echo 'nobase_nodist_datadir_zfs_tests_tests_DATA = \' >> Makefile.am $(find_common) ! -executable -name '*.in' | sort | sed 's/\.in$$//;s/^/\t/;$$!s/$$/ \\/' >> Makefile.am echo 'nobase_nodist_datadir_zfs_tests_tests_SCRIPTS = \' >> Makefile.am $(find_common) -executable -name '*.in' | sort | sed 's/\.in$$//;s/^/\t/;$$!s/$$/ \\/' >> Makefile.am echo >> Makefile.am echo 'SUBSTFILES += $$(nobase_nodist_datadir_zfs_tests_tests_DATA) $$(nobase_nodist_datadir_zfs_tests_tests_SCRIPTS)' >> Makefile.am echo >> Makefile.am echo 'if BUILD_LINUX' >> Makefile.am echo 'nobase_dist_datadir_zfs_tests_tests_SCRIPTS += \' >> Makefile.am $(find_common) ! -name '*.in' ! -name '*.c' | grep -Fe /simd -e /tmpfile | sort | sed 's/^/\t/;$$!s/$$/ \\/' >> Makefile.am echo 'endif' >> Makefile.am echo >> Makefile.am echo 'nobase_dist_datadir_zfs_tests_tests_DATA += \' >> Makefile.am $(find_common) ! -executable ! -name '*.in' ! -name '*.c' | grep -vFe /simd -e /tmpfile | sort | sed 's/^/\t/;$$!s/$$/ \\/' >> Makefile.am echo >> Makefile.am echo 'nobase_dist_datadir_zfs_tests_tests_SCRIPTS += \' >> Makefile.am $(find_common) -executable ! -name '*.in' ! -name '*.c' | grep -vFe /simd -e /tmpfile | sort | sed 's/^/\t/;$$!s/$$/ \\/' >> Makefile.am # -- >8 -- nobase_nodist_datadir_zfs_tests_tests_DATA = \ functional/pam/utilities.kshlib nobase_nodist_datadir_zfs_tests_tests_SCRIPTS = \ functional/pyzfs/pyzfs_unittest.ksh SUBSTFILES += $(nobase_nodist_datadir_zfs_tests_tests_DATA) $(nobase_nodist_datadir_zfs_tests_tests_SCRIPTS) if BUILD_LINUX nobase_dist_datadir_zfs_tests_tests_SCRIPTS += \ functional/simd/simd_supported.ksh \ functional/tmpfile/cleanup.ksh \ functional/tmpfile/setup.ksh endif nobase_dist_datadir_zfs_tests_tests_DATA += \ functional/acl/acl.cfg \ functional/acl/acl_common.kshlib \ functional/alloc_class/alloc_class.cfg \ functional/alloc_class/alloc_class.kshlib \ functional/atime/atime.cfg \ functional/atime/atime_common.kshlib \ functional/cache/cache.cfg \ functional/cache/cache.kshlib \ functional/cachefile/cachefile.cfg \ functional/cachefile/cachefile.kshlib \ functional/casenorm/casenorm.cfg \ functional/casenorm/casenorm.kshlib \ functional/channel_program/channel_common.kshlib \ functional/channel_program/lua_core/tst.args_to_lua.out \ functional/channel_program/lua_core/tst.args_to_lua.zcp \ functional/channel_program/lua_core/tst.divide_by_zero.err \ functional/channel_program/lua_core/tst.divide_by_zero.zcp \ functional/channel_program/lua_core/tst.exists.zcp \ functional/channel_program/lua_core/tst.large_prog.out \ functional/channel_program/lua_core/tst.large_prog.zcp \ functional/channel_program/lua_core/tst.lib_base.lua \ functional/channel_program/lua_core/tst.lib_coroutine.lua \ functional/channel_program/lua_core/tst.lib_strings.lua \ functional/channel_program/lua_core/tst.lib_table.lua \ functional/channel_program/lua_core/tst.nested_neg.zcp \ functional/channel_program/lua_core/tst.nested_pos.zcp \ functional/channel_program/lua_core/tst.recursive.zcp \ functional/channel_program/lua_core/tst.return_large.zcp \ functional/channel_program/lua_core/tst.return_recursive_table.zcp \ functional/channel_program/lua_core/tst.stack_gsub.err \ functional/channel_program/lua_core/tst.stack_gsub.zcp \ functional/channel_program/lua_core/tst.timeout.zcp \ functional/channel_program/synctask_core/tst.bookmark.copy.zcp \ functional/channel_program/synctask_core/tst.bookmark.create.zcp \ functional/channel_program/synctask_core/tst.get_index_props.out \ functional/channel_program/synctask_core/tst.get_index_props.zcp \ functional/channel_program/synctask_core/tst.get_number_props.out \ functional/channel_program/synctask_core/tst.get_number_props.zcp \ functional/channel_program/synctask_core/tst.get_string_props.out \ functional/channel_program/synctask_core/tst.get_string_props.zcp \ functional/channel_program/synctask_core/tst.promote_conflict.zcp \ functional/channel_program/synctask_core/tst.set_props.zcp \ functional/channel_program/synctask_core/tst.snapshot_destroy.zcp \ functional/channel_program/synctask_core/tst.snapshot_neg.zcp \ functional/channel_program/synctask_core/tst.snapshot_recursive.zcp \ functional/channel_program/synctask_core/tst.snapshot_rename.zcp \ functional/channel_program/synctask_core/tst.snapshot_simple.zcp \ functional/checksum/default.cfg \ functional/clean_mirror/clean_mirror_common.kshlib \ functional/clean_mirror/default.cfg \ functional/cli_root/cli_common.kshlib \ functional/cli_root/zfs_copies/zfs_copies.cfg \ functional/cli_root/zfs_copies/zfs_copies.kshlib \ functional/cli_root/zfs_create/properties.kshlib \ functional/cli_root/zfs_create/zfs_create.cfg \ functional/cli_root/zfs_create/zfs_create_common.kshlib \ functional/cli_root/zfs_destroy/zfs_destroy.cfg \ functional/cli_root/zfs_destroy/zfs_destroy_common.kshlib \ functional/cli_root/zfs_get/zfs_get_common.kshlib \ functional/cli_root/zfs_get/zfs_get_list_d.kshlib \ functional/cli_root/zfs_jail/jail.conf \ functional/cli_root/zfs_load-key/HEXKEY \ functional/cli_root/zfs_load-key/PASSPHRASE \ functional/cli_root/zfs_load-key/RAWKEY \ functional/cli_root/zfs_load-key/zfs_load-key.cfg \ functional/cli_root/zfs_load-key/zfs_load-key_common.kshlib \ functional/cli_root/zfs_mount/zfs_mount.cfg \ functional/cli_root/zfs_mount/zfs_mount.kshlib \ functional/cli_root/zfs_promote/zfs_promote.cfg \ functional/cli_root/zfs_receive/zstd_test_data.txt \ functional/cli_root/zfs_rename/zfs_rename.cfg \ functional/cli_root/zfs_rename/zfs_rename.kshlib \ functional/cli_root/zfs_rollback/zfs_rollback.cfg \ functional/cli_root/zfs_rollback/zfs_rollback_common.kshlib \ functional/cli_root/zfs_send/zfs_send.cfg \ functional/cli_root/zfs_set/zfs_set_common.kshlib \ functional/cli_root/zfs_share/zfs_share.cfg \ functional/cli_root/zfs_snapshot/zfs_snapshot.cfg \ functional/cli_root/zfs_unmount/zfs_unmount.cfg \ functional/cli_root/zfs_unmount/zfs_unmount.kshlib \ functional/cli_root/zfs_upgrade/zfs_upgrade.kshlib \ functional/cli_root/zfs_wait/zfs_wait.kshlib \ functional/cli_root/zpool_add/zpool_add.cfg \ functional/cli_root/zpool_add/zpool_add.kshlib \ functional/cli_root/zpool_clear/zpool_clear.cfg \ functional/cli_root/zpool_create/draidcfg.gz \ functional/cli_root/zpool_create/zpool_create.cfg \ functional/cli_root/zpool_create/zpool_create.shlib \ functional/cli_root/zpool_destroy/zpool_destroy.cfg \ functional/cli_root/zpool_events/zpool_events.cfg \ functional/cli_root/zpool_events/zpool_events.kshlib \ functional/cli_root/zpool_expand/zpool_expand.cfg \ functional/cli_root/zpool_export/zpool_export.cfg \ functional/cli_root/zpool_export/zpool_export.kshlib \ functional/cli_root/zpool_get/vdev_get.cfg \ functional/cli_root/zpool_get/zpool_get.cfg \ functional/cli_root/zpool_get/zpool_get_parsable.cfg \ functional/cli_root/zpool_import/blockfiles/cryptv0.dat.bz2 \ functional/cli_root/zpool_import/blockfiles/missing_ivset.dat.bz2 \ functional/cli_root/zpool_import/blockfiles/unclean_export.dat.bz2 \ functional/cli_root/zpool_import/zpool_import.cfg \ functional/cli_root/zpool_import/zpool_import.kshlib \ functional/cli_root/zpool_initialize/zpool_initialize.kshlib \ functional/cli_root/zpool_labelclear/labelclear.cfg \ functional/cli_root/zpool_remove/zpool_remove.cfg \ functional/cli_root/zpool_reopen/zpool_reopen.cfg \ functional/cli_root/zpool_reopen/zpool_reopen.shlib \ functional/cli_root/zpool_resilver/zpool_resilver.cfg \ functional/cli_root/zpool_scrub/zpool_scrub.cfg \ functional/cli_root/zpool_split/zpool_split.cfg \ functional/cli_root/zpool_trim/zpool_trim.kshlib \ functional/cli_root/zpool_upgrade/blockfiles/zfs-broken-mirror1.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-broken-mirror2.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v10.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v11.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v12.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v13.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v14.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v15.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1mirror1.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1mirror2.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1mirror3.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1raidz1.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1raidz2.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1raidz3.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1stripe1.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1stripe2.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v1stripe3.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2mirror1.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2mirror2.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2mirror3.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2raidz1.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2raidz2.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2raidz3.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2stripe1.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2stripe2.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v2stripe3.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3hotspare1.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3hotspare2.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3hotspare3.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3mirror1.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3mirror2.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3mirror3.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3raidz1.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3raidz21.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3raidz22.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3raidz23.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3raidz2.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3raidz3.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3stripe1.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3stripe2.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v3stripe3.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v4.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v5.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v6.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v7.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v8.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v999.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-v9.dat.bz2 \ functional/cli_root/zpool_upgrade/blockfiles/zfs-pool-vBROKEN.dat.bz2 \ functional/cli_root/zpool_upgrade/zpool_upgrade.cfg \ functional/cli_root/zpool_upgrade/zpool_upgrade.kshlib \ functional/cli_root/zpool_wait/zpool_wait.kshlib \ functional/cli_root/zhack/library.kshlib \ functional/cli_user/misc/misc.cfg \ functional/cli_user/zfs_list/zfs_list.cfg \ functional/cli_user/zfs_list/zfs_list.kshlib \ functional/compression/compress.cfg \ functional/compression/testpool_zstd.tar.gz \ functional/deadman/deadman.cfg \ functional/delegate/delegate.cfg \ functional/delegate/delegate_common.kshlib \ functional/devices/devices.cfg \ functional/devices/devices_common.kshlib \ functional/events/events.cfg \ functional/events/events_common.kshlib \ functional/fault/fault.cfg \ functional/grow/grow.cfg \ functional/history/history.cfg \ functional/history/history_common.kshlib \ functional/history/i386.migratedpool.DAT.Z \ functional/history/i386.orig_history.txt \ functional/history/sparc.migratedpool.DAT.Z \ functional/history/sparc.orig_history.txt \ functional/history/zfs-pool-v4.dat.Z \ functional/inheritance/config001.cfg \ functional/inheritance/config002.cfg \ functional/inheritance/config003.cfg \ functional/inheritance/config004.cfg \ functional/inheritance/config005.cfg \ functional/inheritance/config006.cfg \ functional/inheritance/config007.cfg \ functional/inheritance/config008.cfg \ functional/inheritance/config009.cfg \ functional/inheritance/config010.cfg \ functional/inheritance/config011.cfg \ functional/inheritance/config012.cfg \ functional/inheritance/config013.cfg \ functional/inheritance/config014.cfg \ functional/inheritance/config015.cfg \ functional/inheritance/config016.cfg \ functional/inheritance/config017.cfg \ functional/inheritance/config018.cfg \ functional/inheritance/config019.cfg \ functional/inheritance/config020.cfg \ functional/inheritance/config021.cfg \ functional/inheritance/config022.cfg \ functional/inheritance/config023.cfg \ functional/inheritance/config024.cfg \ functional/inheritance/inherit.kshlib \ functional/inheritance/README.config \ functional/inheritance/README.state \ functional/inheritance/state001.cfg \ functional/inheritance/state002.cfg \ functional/inheritance/state003.cfg \ functional/inheritance/state004.cfg \ functional/inheritance/state005.cfg \ functional/inheritance/state006.cfg \ functional/inheritance/state007.cfg \ functional/inheritance/state008.cfg \ functional/inheritance/state009.cfg \ functional/inheritance/state010.cfg \ functional/inheritance/state011.cfg \ functional/inheritance/state012.cfg \ functional/inheritance/state013.cfg \ functional/inheritance/state014.cfg \ functional/inheritance/state015.cfg \ functional/inheritance/state016.cfg \ functional/inheritance/state017.cfg \ functional/inheritance/state018.cfg \ functional/inheritance/state019.cfg \ functional/inheritance/state020.cfg \ functional/inheritance/state021.cfg \ functional/inheritance/state022.cfg \ functional/inheritance/state023.cfg \ functional/inheritance/state024.cfg \ functional/inuse/inuse.cfg \ functional/io/io.cfg \ functional/l2arc/l2arc.cfg \ functional/largest_pool/largest_pool.cfg \ functional/migration/migration.cfg \ functional/migration/migration.kshlib \ functional/mmap/mmap.cfg \ functional/mmp/mmp.cfg \ functional/mmp/mmp.kshlib \ functional/mv_files/mv_files.cfg \ functional/mv_files/mv_files_common.kshlib \ functional/nopwrite/nopwrite.shlib \ functional/no_space/enospc.cfg \ functional/online_offline/online_offline.cfg \ functional/pool_checkpoint/pool_checkpoint.kshlib \ functional/projectquota/projectquota.cfg \ functional/projectquota/projectquota_common.kshlib \ functional/quota/quota.cfg \ functional/quota/quota.kshlib \ functional/redacted_send/redacted.cfg \ functional/redacted_send/redacted.kshlib \ functional/redundancy/redundancy.cfg \ functional/redundancy/redundancy.kshlib \ functional/refreserv/refreserv.cfg \ functional/removal/removal.kshlib \ functional/replacement/replacement.cfg \ functional/reservation/reservation.cfg \ functional/reservation/reservation.shlib \ functional/rsend/dedup_encrypted_zvol.bz2 \ functional/rsend/dedup_encrypted_zvol.zsend.bz2 \ functional/rsend/dedup.zsend.bz2 \ functional/rsend/fs.tar.gz \ functional/rsend/rsend.cfg \ functional/rsend/rsend.kshlib \ functional/scrub_mirror/default.cfg \ functional/scrub_mirror/scrub_mirror_common.kshlib \ functional/slog/slog.cfg \ functional/slog/slog.kshlib \ functional/snapshot/snapshot.cfg \ functional/snapused/snapused.kshlib \ functional/sparse/sparse.cfg \ functional/trim/trim.cfg \ functional/trim/trim.kshlib \ functional/truncate/truncate.cfg \ functional/upgrade/upgrade_common.kshlib \ functional/user_namespace/user_namespace.cfg \ functional/user_namespace/user_namespace_common.kshlib \ functional/userquota/13709_reproducer.bz2 \ functional/userquota/userquota.cfg \ functional/userquota/userquota_common.kshlib \ functional/vdev_zaps/vdev_zaps.kshlib \ functional/xattr/xattr.cfg \ functional/xattr/xattr_common.kshlib \ functional/zvol/zvol.cfg \ functional/zvol/zvol_cli/zvol_cli.cfg \ functional/zvol/zvol_common.shlib \ functional/zvol/zvol_ENOSPC/zvol_ENOSPC.cfg \ functional/zvol/zvol_misc/zvol_misc_common.kshlib \ functional/zvol/zvol_swap/zvol_swap.cfg \ functional/idmap_mount/idmap_mount.cfg \ functional/idmap_mount/idmap_mount_common.kshlib nobase_dist_datadir_zfs_tests_tests_SCRIPTS += \ functional/acl/off/cleanup.ksh \ functional/acl/off/dosmode.ksh \ functional/acl/off/posixmode.ksh \ functional/acl/off/setup.ksh \ functional/acl/posix/cleanup.ksh \ functional/acl/posix/posix_001_pos.ksh \ functional/acl/posix/posix_002_pos.ksh \ functional/acl/posix/posix_003_pos.ksh \ functional/acl/posix/posix_004_pos.ksh \ functional/acl/posix-sa/cleanup.ksh \ functional/acl/posix-sa/posix_001_pos.ksh \ functional/acl/posix-sa/posix_002_pos.ksh \ functional/acl/posix-sa/posix_003_pos.ksh \ functional/acl/posix-sa/posix_004_pos.ksh \ functional/acl/posix-sa/setup.ksh \ functional/acl/posix/setup.ksh \ functional/alloc_class/alloc_class_001_pos.ksh \ functional/alloc_class/alloc_class_002_neg.ksh \ functional/alloc_class/alloc_class_003_pos.ksh \ functional/alloc_class/alloc_class_004_pos.ksh \ functional/alloc_class/alloc_class_005_pos.ksh \ functional/alloc_class/alloc_class_006_pos.ksh \ functional/alloc_class/alloc_class_007_pos.ksh \ functional/alloc_class/alloc_class_008_pos.ksh \ functional/alloc_class/alloc_class_009_pos.ksh \ functional/alloc_class/alloc_class_010_pos.ksh \ functional/alloc_class/alloc_class_011_neg.ksh \ functional/alloc_class/alloc_class_012_pos.ksh \ functional/alloc_class/alloc_class_013_pos.ksh \ + functional/alloc_class/alloc_class_014_neg.ksh \ + functional/alloc_class/alloc_class_015_pos.ksh \ functional/alloc_class/cleanup.ksh \ functional/alloc_class/setup.ksh \ functional/append/file_append.ksh \ functional/append/threadsappend_001_pos.ksh \ functional/append/cleanup.ksh \ functional/append/setup.ksh \ functional/arc/arcstats_runtime_tuning.ksh \ functional/arc/cleanup.ksh \ functional/arc/dbufstats_001_pos.ksh \ functional/arc/dbufstats_002_pos.ksh \ functional/arc/dbufstats_003_pos.ksh \ functional/arc/setup.ksh \ functional/atime/atime_001_pos.ksh \ functional/atime/atime_002_neg.ksh \ functional/atime/atime_003_pos.ksh \ functional/atime/cleanup.ksh \ functional/atime/root_atime_off.ksh \ functional/atime/root_atime_on.ksh \ functional/atime/root_relatime_on.ksh \ functional/atime/setup.ksh \ functional/bootfs/bootfs_001_pos.ksh \ functional/bootfs/bootfs_002_neg.ksh \ functional/bootfs/bootfs_003_pos.ksh \ functional/bootfs/bootfs_004_neg.ksh \ functional/bootfs/bootfs_005_neg.ksh \ functional/bootfs/bootfs_006_pos.ksh \ functional/bootfs/bootfs_007_pos.ksh \ functional/bootfs/bootfs_008_pos.ksh \ functional/bootfs/cleanup.ksh \ functional/bootfs/setup.ksh \ functional/btree/btree_negative.ksh \ functional/btree/btree_positive.ksh \ functional/cache/cache_001_pos.ksh \ functional/cache/cache_002_pos.ksh \ functional/cache/cache_003_pos.ksh \ functional/cache/cache_004_neg.ksh \ functional/cache/cache_005_neg.ksh \ functional/cache/cache_006_pos.ksh \ functional/cache/cache_007_neg.ksh \ functional/cache/cache_008_neg.ksh \ functional/cache/cache_009_pos.ksh \ functional/cache/cache_010_pos.ksh \ functional/cache/cache_011_pos.ksh \ functional/cache/cache_012_pos.ksh \ functional/cache/cleanup.ksh \ functional/cachefile/cachefile_001_pos.ksh \ functional/cachefile/cachefile_002_pos.ksh \ functional/cachefile/cachefile_003_pos.ksh \ functional/cachefile/cachefile_004_pos.ksh \ functional/cachefile/cleanup.ksh \ functional/cachefile/setup.ksh \ functional/cache/setup.ksh \ functional/casenorm/case_all_values.ksh \ functional/casenorm/cleanup.ksh \ functional/casenorm/insensitive_formd_delete.ksh \ functional/casenorm/insensitive_formd_lookup.ksh \ functional/casenorm/insensitive_none_delete.ksh \ functional/casenorm/insensitive_none_lookup.ksh \ functional/casenorm/mixed_create_failure.ksh \ functional/casenorm/mixed_formd_delete.ksh \ functional/casenorm/mixed_formd_lookup_ci.ksh \ functional/casenorm/mixed_formd_lookup.ksh \ functional/casenorm/mixed_none_delete.ksh \ functional/casenorm/mixed_none_lookup_ci.ksh \ functional/casenorm/mixed_none_lookup.ksh \ functional/casenorm/norm_all_values.ksh \ functional/casenorm/sensitive_formd_delete.ksh \ functional/casenorm/sensitive_formd_lookup.ksh \ functional/casenorm/sensitive_none_delete.ksh \ functional/casenorm/sensitive_none_lookup.ksh \ functional/casenorm/setup.ksh \ functional/channel_program/lua_core/cleanup.ksh \ functional/channel_program/lua_core/setup.ksh \ functional/channel_program/lua_core/tst.args_to_lua.ksh \ functional/channel_program/lua_core/tst.divide_by_zero.ksh \ functional/channel_program/lua_core/tst.exists.ksh \ functional/channel_program/lua_core/tst.integer_illegal.ksh \ functional/channel_program/lua_core/tst.integer_overflow.ksh \ functional/channel_program/lua_core/tst.language_functions_neg.ksh \ functional/channel_program/lua_core/tst.language_functions_pos.ksh \ functional/channel_program/lua_core/tst.large_prog.ksh \ functional/channel_program/lua_core/tst.libraries.ksh \ functional/channel_program/lua_core/tst.memory_limit.ksh \ functional/channel_program/lua_core/tst.nested_neg.ksh \ functional/channel_program/lua_core/tst.nested_pos.ksh \ functional/channel_program/lua_core/tst.nvlist_to_lua.ksh \ functional/channel_program/lua_core/tst.recursive_neg.ksh \ functional/channel_program/lua_core/tst.recursive_pos.ksh \ functional/channel_program/lua_core/tst.return_large.ksh \ functional/channel_program/lua_core/tst.return_nvlist_neg.ksh \ functional/channel_program/lua_core/tst.return_nvlist_pos.ksh \ functional/channel_program/lua_core/tst.return_recursive_table.ksh \ functional/channel_program/lua_core/tst.stack_gsub.ksh \ functional/channel_program/lua_core/tst.timeout.ksh \ functional/channel_program/synctask_core/cleanup.ksh \ functional/channel_program/synctask_core/setup.ksh \ functional/channel_program/synctask_core/tst.bookmark.copy.ksh \ functional/channel_program/synctask_core/tst.bookmark.create.ksh \ functional/channel_program/synctask_core/tst.destroy_fs.ksh \ functional/channel_program/synctask_core/tst.destroy_snap.ksh \ functional/channel_program/synctask_core/tst.get_count_and_limit.ksh \ functional/channel_program/synctask_core/tst.get_index_props.ksh \ functional/channel_program/synctask_core/tst.get_mountpoint.ksh \ functional/channel_program/synctask_core/tst.get_neg.ksh \ functional/channel_program/synctask_core/tst.get_number_props.ksh \ functional/channel_program/synctask_core/tst.get_string_props.ksh \ functional/channel_program/synctask_core/tst.get_type.ksh \ functional/channel_program/synctask_core/tst.get_userquota.ksh \ functional/channel_program/synctask_core/tst.get_written.ksh \ functional/channel_program/synctask_core/tst.inherit.ksh \ functional/channel_program/synctask_core/tst.list_bookmarks.ksh \ functional/channel_program/synctask_core/tst.list_children.ksh \ functional/channel_program/synctask_core/tst.list_clones.ksh \ functional/channel_program/synctask_core/tst.list_holds.ksh \ functional/channel_program/synctask_core/tst.list_snapshots.ksh \ functional/channel_program/synctask_core/tst.list_system_props.ksh \ functional/channel_program/synctask_core/tst.list_user_props.ksh \ functional/channel_program/synctask_core/tst.parse_args_neg.ksh \ functional/channel_program/synctask_core/tst.promote_conflict.ksh \ functional/channel_program/synctask_core/tst.promote_multiple.ksh \ functional/channel_program/synctask_core/tst.promote_simple.ksh \ functional/channel_program/synctask_core/tst.rollback_mult.ksh \ functional/channel_program/synctask_core/tst.rollback_one.ksh \ functional/channel_program/synctask_core/tst.set_props.ksh \ functional/channel_program/synctask_core/tst.snapshot_destroy.ksh \ functional/channel_program/synctask_core/tst.snapshot_neg.ksh \ functional/channel_program/synctask_core/tst.snapshot_recursive.ksh \ functional/channel_program/synctask_core/tst.snapshot_rename.ksh \ functional/channel_program/synctask_core/tst.snapshot_simple.ksh \ functional/channel_program/synctask_core/tst.terminate_by_signal.ksh \ functional/chattr/chattr_001_pos.ksh \ functional/chattr/chattr_002_neg.ksh \ functional/chattr/cleanup.ksh \ functional/chattr/setup.ksh \ functional/checksum/cleanup.ksh \ functional/checksum/filetest_001_pos.ksh \ functional/checksum/filetest_002_pos.ksh \ functional/checksum/run_blake3_test.ksh \ functional/checksum/run_edonr_test.ksh \ functional/checksum/run_sha2_test.ksh \ functional/checksum/run_skein_test.ksh \ functional/checksum/setup.ksh \ functional/clean_mirror/clean_mirror_001_pos.ksh \ functional/clean_mirror/clean_mirror_002_pos.ksh \ functional/clean_mirror/clean_mirror_003_pos.ksh \ functional/clean_mirror/clean_mirror_004_pos.ksh \ functional/clean_mirror/cleanup.ksh \ functional/clean_mirror/setup.ksh \ functional/cli_root/zdb/zdb_002_pos.ksh \ functional/cli_root/zdb/zdb_003_pos.ksh \ functional/cli_root/zdb/zdb_004_pos.ksh \ functional/cli_root/zdb/zdb_005_pos.ksh \ functional/cli_root/zdb/zdb_006_pos.ksh \ functional/cli_root/zdb/zdb_args_neg.ksh \ functional/cli_root/zdb/zdb_args_pos.ksh \ functional/cli_root/zdb/zdb_block_size_histogram.ksh \ functional/cli_root/zdb/zdb_checksum.ksh \ functional/cli_root/zdb/zdb_decompress.ksh \ functional/cli_root/zdb/zdb_decompress_zstd.ksh \ functional/cli_root/zdb/zdb_display_block.ksh \ functional/cli_root/zdb/zdb_encrypted.ksh \ functional/cli_root/zdb/zdb_label_checksum.ksh \ functional/cli_root/zdb/zdb_object_range_neg.ksh \ functional/cli_root/zdb/zdb_object_range_pos.ksh \ functional/cli_root/zdb/zdb_objset_id.ksh \ functional/cli_root/zdb/zdb_recover_2.ksh \ functional/cli_root/zdb/zdb_recover.ksh \ functional/cli_root/zfs_bookmark/cleanup.ksh \ functional/cli_root/zfs_bookmark/setup.ksh \ functional/cli_root/zfs_bookmark/zfs_bookmark_cliargs.ksh \ functional/cli_root/zfs_change-key/cleanup.ksh \ functional/cli_root/zfs_change-key/setup.ksh \ functional/cli_root/zfs_change-key/zfs_change-key_child.ksh \ functional/cli_root/zfs_change-key/zfs_change-key_clones.ksh \ functional/cli_root/zfs_change-key/zfs_change-key_format.ksh \ functional/cli_root/zfs_change-key/zfs_change-key_inherit.ksh \ functional/cli_root/zfs_change-key/zfs_change-key.ksh \ functional/cli_root/zfs_change-key/zfs_change-key_load.ksh \ functional/cli_root/zfs_change-key/zfs_change-key_location.ksh \ functional/cli_root/zfs_change-key/zfs_change-key_pbkdf2iters.ksh \ functional/cli_root/zfs/cleanup.ksh \ functional/cli_root/zfs_clone/cleanup.ksh \ functional/cli_root/zfs_clone/setup.ksh \ functional/cli_root/zfs_clone/zfs_clone_001_neg.ksh \ functional/cli_root/zfs_clone/zfs_clone_002_pos.ksh \ functional/cli_root/zfs_clone/zfs_clone_003_pos.ksh \ functional/cli_root/zfs_clone/zfs_clone_004_pos.ksh \ functional/cli_root/zfs_clone/zfs_clone_005_pos.ksh \ functional/cli_root/zfs_clone/zfs_clone_006_pos.ksh \ functional/cli_root/zfs_clone/zfs_clone_007_pos.ksh \ functional/cli_root/zfs_clone/zfs_clone_008_neg.ksh \ functional/cli_root/zfs_clone/zfs_clone_009_neg.ksh \ functional/cli_root/zfs_clone/zfs_clone_010_pos.ksh \ functional/cli_root/zfs_clone/zfs_clone_deeply_nested.ksh \ functional/cli_root/zfs_clone/zfs_clone_encrypted.ksh \ functional/cli_root/zfs_clone/zfs_clone_rm_nested.ksh \ functional/cli_root/zfs_copies/cleanup.ksh \ functional/cli_root/zfs_copies/setup.ksh \ functional/cli_root/zfs_copies/zfs_copies_001_pos.ksh \ functional/cli_root/zfs_copies/zfs_copies_002_pos.ksh \ functional/cli_root/zfs_copies/zfs_copies_003_pos.ksh \ functional/cli_root/zfs_copies/zfs_copies_004_neg.ksh \ functional/cli_root/zfs_copies/zfs_copies_005_neg.ksh \ functional/cli_root/zfs_copies/zfs_copies_006_pos.ksh \ functional/cli_root/zfs_create/cleanup.ksh \ functional/cli_root/zfs_create/setup.ksh \ functional/cli_root/zfs_create/zfs_create_001_pos.ksh \ functional/cli_root/zfs_create/zfs_create_002_pos.ksh \ functional/cli_root/zfs_create/zfs_create_003_pos.ksh \ functional/cli_root/zfs_create/zfs_create_004_pos.ksh \ functional/cli_root/zfs_create/zfs_create_005_pos.ksh \ functional/cli_root/zfs_create/zfs_create_006_pos.ksh \ functional/cli_root/zfs_create/zfs_create_007_pos.ksh \ functional/cli_root/zfs_create/zfs_create_008_neg.ksh \ functional/cli_root/zfs_create/zfs_create_009_neg.ksh \ functional/cli_root/zfs_create/zfs_create_010_neg.ksh \ functional/cli_root/zfs_create/zfs_create_011_pos.ksh \ functional/cli_root/zfs_create/zfs_create_012_pos.ksh \ functional/cli_root/zfs_create/zfs_create_013_pos.ksh \ functional/cli_root/zfs_create/zfs_create_014_pos.ksh \ functional/cli_root/zfs_create/zfs_create_crypt_combos.ksh \ functional/cli_root/zfs_create/zfs_create_dryrun.ksh \ functional/cli_root/zfs_create/zfs_create_encrypted.ksh \ functional/cli_root/zfs_create/zfs_create_nomount.ksh \ functional/cli_root/zfs_create/zfs_create_verbose.ksh \ functional/cli_root/zfs_destroy/cleanup.ksh \ functional/cli_root/zfs_destroy/setup.ksh \ functional/cli_root/zfs_destroy/zfs_clone_livelist_condense_and_disable.ksh \ functional/cli_root/zfs_destroy/zfs_clone_livelist_condense_races.ksh \ functional/cli_root/zfs_destroy/zfs_clone_livelist_dedup.ksh \ functional/cli_root/zfs_destroy/zfs_destroy_001_pos.ksh \ functional/cli_root/zfs_destroy/zfs_destroy_002_pos.ksh \ functional/cli_root/zfs_destroy/zfs_destroy_003_pos.ksh \ functional/cli_root/zfs_destroy/zfs_destroy_004_pos.ksh \ functional/cli_root/zfs_destroy/zfs_destroy_005_neg.ksh \ functional/cli_root/zfs_destroy/zfs_destroy_006_neg.ksh \ functional/cli_root/zfs_destroy/zfs_destroy_007_neg.ksh \ functional/cli_root/zfs_destroy/zfs_destroy_008_pos.ksh \ functional/cli_root/zfs_destroy/zfs_destroy_009_pos.ksh \ functional/cli_root/zfs_destroy/zfs_destroy_010_pos.ksh \ functional/cli_root/zfs_destroy/zfs_destroy_011_pos.ksh \ functional/cli_root/zfs_destroy/zfs_destroy_012_pos.ksh \ functional/cli_root/zfs_destroy/zfs_destroy_013_neg.ksh \ functional/cli_root/zfs_destroy/zfs_destroy_014_pos.ksh \ functional/cli_root/zfs_destroy/zfs_destroy_015_pos.ksh \ functional/cli_root/zfs_destroy/zfs_destroy_016_pos.ksh \ functional/cli_root/zfs_destroy/zfs_destroy_clone_livelist.ksh \ functional/cli_root/zfs_destroy/zfs_destroy_dev_removal_condense.ksh \ functional/cli_root/zfs_destroy/zfs_destroy_dev_removal.ksh \ functional/cli_root/zfs_diff/cleanup.ksh \ functional/cli_root/zfs_diff/setup.ksh \ functional/cli_root/zfs_diff/zfs_diff_changes.ksh \ functional/cli_root/zfs_diff/zfs_diff_cliargs.ksh \ functional/cli_root/zfs_diff/zfs_diff_encrypted.ksh \ functional/cli_root/zfs_diff/zfs_diff_mangle.ksh \ functional/cli_root/zfs_diff/zfs_diff_timestamp.ksh \ functional/cli_root/zfs_diff/zfs_diff_types.ksh \ functional/cli_root/zfs_get/cleanup.ksh \ functional/cli_root/zfs_get/setup.ksh \ functional/cli_root/zfs_get/zfs_get_001_pos.ksh \ functional/cli_root/zfs_get/zfs_get_002_pos.ksh \ functional/cli_root/zfs_get/zfs_get_003_pos.ksh \ functional/cli_root/zfs_get/zfs_get_004_pos.ksh \ functional/cli_root/zfs_get/zfs_get_005_neg.ksh \ functional/cli_root/zfs_get/zfs_get_006_neg.ksh \ functional/cli_root/zfs_get/zfs_get_007_neg.ksh \ functional/cli_root/zfs_get/zfs_get_008_pos.ksh \ functional/cli_root/zfs_get/zfs_get_009_pos.ksh \ functional/cli_root/zfs_get/zfs_get_010_neg.ksh \ functional/cli_root/zfs_ids_to_path/cleanup.ksh \ functional/cli_root/zfs_ids_to_path/setup.ksh \ functional/cli_root/zfs_ids_to_path/zfs_ids_to_path_001_pos.ksh \ functional/cli_root/zfs_inherit/cleanup.ksh \ functional/cli_root/zfs_inherit/setup.ksh \ functional/cli_root/zfs_inherit/zfs_inherit_001_neg.ksh \ functional/cli_root/zfs_inherit/zfs_inherit_002_neg.ksh \ functional/cli_root/zfs_inherit/zfs_inherit_003_pos.ksh \ functional/cli_root/zfs_inherit/zfs_inherit_mountpoint.ksh \ functional/cli_root/zfs_jail/cleanup.ksh \ functional/cli_root/zfs_jail/setup.ksh \ functional/cli_root/zfs_jail/zfs_jail_001_pos.ksh \ functional/cli_root/zfs_load-key/cleanup.ksh \ functional/cli_root/zfs_load-key/setup.ksh \ functional/cli_root/zfs_load-key/zfs_load-key_all.ksh \ functional/cli_root/zfs_load-key/zfs_load-key_file.ksh \ functional/cli_root/zfs_load-key/zfs_load-key_https.ksh \ functional/cli_root/zfs_load-key/zfs_load-key.ksh \ functional/cli_root/zfs_load-key/zfs_load-key_location.ksh \ functional/cli_root/zfs_load-key/zfs_load-key_noop.ksh \ functional/cli_root/zfs_load-key/zfs_load-key_recursive.ksh \ functional/cli_root/zfs_mount/cleanup.ksh \ functional/cli_root/zfs_mount/setup.ksh \ functional/cli_root/zfs_mount/zfs_mount_001_pos.ksh \ functional/cli_root/zfs_mount/zfs_mount_002_pos.ksh \ functional/cli_root/zfs_mount/zfs_mount_003_pos.ksh \ functional/cli_root/zfs_mount/zfs_mount_004_pos.ksh \ functional/cli_root/zfs_mount/zfs_mount_005_pos.ksh \ functional/cli_root/zfs_mount/zfs_mount_006_pos.ksh \ functional/cli_root/zfs_mount/zfs_mount_007_pos.ksh \ functional/cli_root/zfs_mount/zfs_mount_008_pos.ksh \ functional/cli_root/zfs_mount/zfs_mount_009_neg.ksh \ functional/cli_root/zfs_mount/zfs_mount_010_neg.ksh \ functional/cli_root/zfs_mount/zfs_mount_011_neg.ksh \ functional/cli_root/zfs_mount/zfs_mount_012_pos.ksh \ functional/cli_root/zfs_mount/zfs_mount_013_pos.ksh \ functional/cli_root/zfs_mount/zfs_mount_014_neg.ksh \ functional/cli_root/zfs_mount/zfs_mount_all_001_pos.ksh \ functional/cli_root/zfs_mount/zfs_mount_all_fail.ksh \ functional/cli_root/zfs_mount/zfs_mount_all_mountpoints.ksh \ functional/cli_root/zfs_mount/zfs_mount_encrypted.ksh \ functional/cli_root/zfs_mount/zfs_mount_remount.ksh \ functional/cli_root/zfs_mount/zfs_mount_test_race.ksh \ functional/cli_root/zfs_mount/zfs_multi_mount.ksh \ functional/cli_root/zfs_program/cleanup.ksh \ functional/cli_root/zfs_program/setup.ksh \ functional/cli_root/zfs_program/zfs_program_json.ksh \ functional/cli_root/zfs_promote/cleanup.ksh \ functional/cli_root/zfs_promote/setup.ksh \ functional/cli_root/zfs_promote/zfs_promote_001_pos.ksh \ functional/cli_root/zfs_promote/zfs_promote_002_pos.ksh \ functional/cli_root/zfs_promote/zfs_promote_003_pos.ksh \ functional/cli_root/zfs_promote/zfs_promote_004_pos.ksh \ functional/cli_root/zfs_promote/zfs_promote_005_pos.ksh \ functional/cli_root/zfs_promote/zfs_promote_006_neg.ksh \ functional/cli_root/zfs_promote/zfs_promote_007_neg.ksh \ functional/cli_root/zfs_promote/zfs_promote_008_pos.ksh \ functional/cli_root/zfs_promote/zfs_promote_encryptionroot.ksh \ functional/cli_root/zfs_property/cleanup.ksh \ functional/cli_root/zfs_property/setup.ksh \ functional/cli_root/zfs_property/zfs_written_property_001_pos.ksh \ functional/cli_root/zfs_receive/cleanup.ksh \ functional/cli_root/zfs_receive/receive-o-x_props_aliases.ksh \ functional/cli_root/zfs_receive/receive-o-x_props_override.ksh \ functional/cli_root/zfs_receive/setup.ksh \ functional/cli_root/zfs_receive/zfs_receive_001_pos.ksh \ functional/cli_root/zfs_receive/zfs_receive_002_pos.ksh \ functional/cli_root/zfs_receive/zfs_receive_003_pos.ksh \ functional/cli_root/zfs_receive/zfs_receive_004_neg.ksh \ functional/cli_root/zfs_receive/zfs_receive_005_neg.ksh \ functional/cli_root/zfs_receive/zfs_receive_006_pos.ksh \ functional/cli_root/zfs_receive/zfs_receive_007_neg.ksh \ functional/cli_root/zfs_receive/zfs_receive_008_pos.ksh \ functional/cli_root/zfs_receive/zfs_receive_009_neg.ksh \ functional/cli_root/zfs_receive/zfs_receive_010_pos.ksh \ functional/cli_root/zfs_receive/zfs_receive_011_pos.ksh \ functional/cli_root/zfs_receive/zfs_receive_012_pos.ksh \ functional/cli_root/zfs_receive/zfs_receive_013_pos.ksh \ functional/cli_root/zfs_receive/zfs_receive_014_pos.ksh \ functional/cli_root/zfs_receive/zfs_receive_015_pos.ksh \ functional/cli_root/zfs_receive/zfs_receive_016_pos.ksh \ functional/cli_root/zfs_receive/zfs_receive_-e.ksh \ functional/cli_root/zfs_receive/zfs_receive_from_encrypted.ksh \ functional/cli_root/zfs_receive/zfs_receive_from_zstd.ksh \ functional/cli_root/zfs_receive/zfs_receive_new_props.ksh \ functional/cli_root/zfs_receive/zfs_receive_raw_-d.ksh \ functional/cli_root/zfs_receive/zfs_receive_raw_incremental.ksh \ functional/cli_root/zfs_receive/zfs_receive_raw.ksh \ functional/cli_root/zfs_receive/zfs_receive_to_encrypted.ksh \ functional/cli_root/zfs_receive/zfs_receive_-wR-encrypted-mix.ksh \ functional/cli_root/zfs_receive/zfs_receive_corrective.ksh \ functional/cli_root/zfs_receive/zfs_receive_compressed_corrective.ksh \ functional/cli_root/zfs_receive/zfs_receive_large_block_corrective.ksh \ functional/cli_root/zfs_rename/cleanup.ksh \ functional/cli_root/zfs_rename/setup.ksh \ functional/cli_root/zfs_rename/zfs_rename_001_pos.ksh \ functional/cli_root/zfs_rename/zfs_rename_002_pos.ksh \ functional/cli_root/zfs_rename/zfs_rename_003_pos.ksh \ functional/cli_root/zfs_rename/zfs_rename_004_neg.ksh \ functional/cli_root/zfs_rename/zfs_rename_005_neg.ksh \ functional/cli_root/zfs_rename/zfs_rename_006_pos.ksh \ functional/cli_root/zfs_rename/zfs_rename_007_pos.ksh \ functional/cli_root/zfs_rename/zfs_rename_008_pos.ksh \ functional/cli_root/zfs_rename/zfs_rename_009_neg.ksh \ functional/cli_root/zfs_rename/zfs_rename_010_neg.ksh \ functional/cli_root/zfs_rename/zfs_rename_011_pos.ksh \ functional/cli_root/zfs_rename/zfs_rename_012_neg.ksh \ functional/cli_root/zfs_rename/zfs_rename_013_pos.ksh \ functional/cli_root/zfs_rename/zfs_rename_014_neg.ksh \ functional/cli_root/zfs_rename/zfs_rename_encrypted_child.ksh \ functional/cli_root/zfs_rename/zfs_rename_mountpoint.ksh \ functional/cli_root/zfs_rename/zfs_rename_nounmount.ksh \ functional/cli_root/zfs_rename/zfs_rename_to_encrypted.ksh \ functional/cli_root/zfs_reservation/cleanup.ksh \ functional/cli_root/zfs_reservation/setup.ksh \ functional/cli_root/zfs_reservation/zfs_reservation_001_pos.ksh \ functional/cli_root/zfs_reservation/zfs_reservation_002_pos.ksh \ functional/cli_root/zfs_rollback/cleanup.ksh \ functional/cli_root/zfs_rollback/setup.ksh \ functional/cli_root/zfs_rollback/zfs_rollback_001_pos.ksh \ functional/cli_root/zfs_rollback/zfs_rollback_002_pos.ksh \ functional/cli_root/zfs_rollback/zfs_rollback_003_neg.ksh \ functional/cli_root/zfs_rollback/zfs_rollback_004_neg.ksh \ functional/cli_root/zfs_send/cleanup.ksh \ functional/cli_root/zfs_send/setup.ksh \ functional/cli_root/zfs_send/zfs_send_001_pos.ksh \ functional/cli_root/zfs_send/zfs_send_002_pos.ksh \ functional/cli_root/zfs_send/zfs_send_003_pos.ksh \ functional/cli_root/zfs_send/zfs_send_004_neg.ksh \ functional/cli_root/zfs_send/zfs_send_005_pos.ksh \ functional/cli_root/zfs_send/zfs_send_006_pos.ksh \ functional/cli_root/zfs_send/zfs_send_007_pos.ksh \ functional/cli_root/zfs_send/zfs_send-b.ksh \ functional/cli_root/zfs_send/zfs_send_encrypted.ksh \ functional/cli_root/zfs_send/zfs_send_encrypted_unloaded.ksh \ functional/cli_root/zfs_send/zfs_send_raw.ksh \ functional/cli_root/zfs_send/zfs_send_skip_missing.ksh \ functional/cli_root/zfs_send/zfs_send_sparse.ksh \ functional/cli_root/zfs_set/cache_001_pos.ksh \ functional/cli_root/zfs_set/cache_002_neg.ksh \ functional/cli_root/zfs_set/canmount_001_pos.ksh \ functional/cli_root/zfs_set/canmount_002_pos.ksh \ functional/cli_root/zfs_set/canmount_003_pos.ksh \ functional/cli_root/zfs_set/canmount_004_pos.ksh \ functional/cli_root/zfs_set/checksum_001_pos.ksh \ functional/cli_root/zfs_set/cleanup.ksh \ functional/cli_root/zfs_set/compression_001_pos.ksh \ functional/cli_root/zfs_set/mountpoint_001_pos.ksh \ functional/cli_root/zfs_set/mountpoint_002_pos.ksh \ functional/cli_root/zfs_set/mountpoint_003_pos.ksh \ functional/cli_root/zfs_set/onoffs_001_pos.ksh \ functional/cli_root/zfs_set/property_alias_001_pos.ksh \ functional/cli_root/zfs_set/readonly_001_pos.ksh \ functional/cli_root/zfs_set/reservation_001_neg.ksh \ functional/cli_root/zfs_set/ro_props_001_pos.ksh \ functional/cli_root/zfs_set/setup.ksh \ functional/cli_root/zfs_set/share_mount_001_neg.ksh \ functional/cli_root/zfs_set/snapdir_001_pos.ksh \ functional/cli_root/zfs/setup.ksh \ functional/cli_root/zfs_set/user_property_001_pos.ksh \ functional/cli_root/zfs_set/user_property_002_pos.ksh \ functional/cli_root/zfs_set/user_property_003_neg.ksh \ functional/cli_root/zfs_set/user_property_004_pos.ksh \ functional/cli_root/zfs_set/version_001_neg.ksh \ functional/cli_root/zfs_set/zfs_set_001_neg.ksh \ functional/cli_root/zfs_set/zfs_set_002_neg.ksh \ functional/cli_root/zfs_set/zfs_set_003_neg.ksh \ functional/cli_root/zfs_set/zfs_set_feature_activation.ksh \ functional/cli_root/zfs_set/zfs_set_keylocation.ksh \ functional/cli_root/zfs_share/cleanup.ksh \ functional/cli_root/zfs_share/setup.ksh \ functional/cli_root/zfs_share/zfs_share_001_pos.ksh \ functional/cli_root/zfs_share/zfs_share_002_pos.ksh \ functional/cli_root/zfs_share/zfs_share_003_pos.ksh \ functional/cli_root/zfs_share/zfs_share_004_pos.ksh \ functional/cli_root/zfs_share/zfs_share_005_pos.ksh \ functional/cli_root/zfs_share/zfs_share_006_pos.ksh \ functional/cli_root/zfs_share/zfs_share_007_neg.ksh \ functional/cli_root/zfs_share/zfs_share_008_neg.ksh \ functional/cli_root/zfs_share/zfs_share_009_neg.ksh \ functional/cli_root/zfs_share/zfs_share_010_neg.ksh \ functional/cli_root/zfs_share/zfs_share_011_pos.ksh \ functional/cli_root/zfs_share/zfs_share_012_pos.ksh \ functional/cli_root/zfs_share/zfs_share_013_pos.ksh \ functional/cli_root/zfs_share/zfs_share_concurrent_shares.ksh \ functional/cli_root/zfs_snapshot/cleanup.ksh \ functional/cli_root/zfs_snapshot/setup.ksh \ functional/cli_root/zfs_snapshot/zfs_snapshot_001_neg.ksh \ functional/cli_root/zfs_snapshot/zfs_snapshot_002_neg.ksh \ functional/cli_root/zfs_snapshot/zfs_snapshot_003_neg.ksh \ functional/cli_root/zfs_snapshot/zfs_snapshot_004_neg.ksh \ functional/cli_root/zfs_snapshot/zfs_snapshot_005_neg.ksh \ functional/cli_root/zfs_snapshot/zfs_snapshot_006_pos.ksh \ functional/cli_root/zfs_snapshot/zfs_snapshot_007_neg.ksh \ functional/cli_root/zfs_snapshot/zfs_snapshot_008_neg.ksh \ functional/cli_root/zfs_snapshot/zfs_snapshot_009_pos.ksh \ functional/cli_root/zfs_sysfs/cleanup.ksh \ functional/cli_root/zfs_sysfs/setup.ksh \ functional/cli_root/zfs_sysfs/zfeature_set_unsupported.ksh \ functional/cli_root/zfs_sysfs/zfs_get_unsupported.ksh \ functional/cli_root/zfs_sysfs/zfs_set_unsupported.ksh \ functional/cli_root/zfs_sysfs/zfs_sysfs_live.ksh \ functional/cli_root/zfs_sysfs/zpool_get_unsupported.ksh \ functional/cli_root/zfs_sysfs/zpool_set_unsupported.ksh \ functional/cli_root/zfs_unload-key/cleanup.ksh \ functional/cli_root/zfs_unload-key/setup.ksh \ functional/cli_root/zfs_unload-key/zfs_unload-key_all.ksh \ functional/cli_root/zfs_unload-key/zfs_unload-key.ksh \ functional/cli_root/zfs_unload-key/zfs_unload-key_recursive.ksh \ functional/cli_root/zfs_unmount/cleanup.ksh \ functional/cli_root/zfs_unmount/setup.ksh \ functional/cli_root/zfs_unmount/zfs_unmount_001_pos.ksh \ functional/cli_root/zfs_unmount/zfs_unmount_002_pos.ksh \ functional/cli_root/zfs_unmount/zfs_unmount_003_pos.ksh \ functional/cli_root/zfs_unmount/zfs_unmount_004_pos.ksh \ functional/cli_root/zfs_unmount/zfs_unmount_005_pos.ksh \ functional/cli_root/zfs_unmount/zfs_unmount_006_pos.ksh \ functional/cli_root/zfs_unmount/zfs_unmount_007_neg.ksh \ functional/cli_root/zfs_unmount/zfs_unmount_008_neg.ksh \ functional/cli_root/zfs_unmount/zfs_unmount_009_pos.ksh \ functional/cli_root/zfs_unmount/zfs_unmount_all_001_pos.ksh \ functional/cli_root/zfs_unmount/zfs_unmount_nested.ksh \ functional/cli_root/zfs_unmount/zfs_unmount_unload_keys.ksh \ functional/cli_root/zfs_unshare/cleanup.ksh \ functional/cli_root/zfs_unshare/setup.ksh \ functional/cli_root/zfs_unshare/zfs_unshare_001_pos.ksh \ functional/cli_root/zfs_unshare/zfs_unshare_002_pos.ksh \ functional/cli_root/zfs_unshare/zfs_unshare_003_pos.ksh \ functional/cli_root/zfs_unshare/zfs_unshare_004_neg.ksh \ functional/cli_root/zfs_unshare/zfs_unshare_005_neg.ksh \ functional/cli_root/zfs_unshare/zfs_unshare_006_pos.ksh \ functional/cli_root/zfs_unshare/zfs_unshare_007_pos.ksh \ functional/cli_root/zfs_unshare/zfs_unshare_008_pos.ksh \ functional/cli_root/zfs_upgrade/cleanup.ksh \ functional/cli_root/zfs_upgrade/setup.ksh \ functional/cli_root/zfs_upgrade/zfs_upgrade_001_pos.ksh \ functional/cli_root/zfs_upgrade/zfs_upgrade_002_pos.ksh \ functional/cli_root/zfs_upgrade/zfs_upgrade_003_pos.ksh \ functional/cli_root/zfs_upgrade/zfs_upgrade_004_pos.ksh \ functional/cli_root/zfs_upgrade/zfs_upgrade_005_pos.ksh \ functional/cli_root/zfs_upgrade/zfs_upgrade_006_neg.ksh \ functional/cli_root/zfs_upgrade/zfs_upgrade_007_neg.ksh \ functional/cli_root/zfs_wait/cleanup.ksh \ functional/cli_root/zfs_wait/setup.ksh \ functional/cli_root/zfs_wait/zfs_wait_deleteq.ksh \ functional/cli_root/zfs_wait/zfs_wait_getsubopt.ksh \ functional/cli_root/zfs/zfs_001_neg.ksh \ functional/cli_root/zfs/zfs_002_pos.ksh \ functional/cli_root/zfs/zfs_003_neg.ksh \ functional/cli_root/zhack/zhack_label_repair_001.ksh \ functional/cli_root/zhack/zhack_label_repair_002.ksh \ functional/cli_root/zhack/zhack_label_repair_003.ksh \ functional/cli_root/zhack/zhack_label_repair_004.ksh \ functional/cli_root/zpool_add/add_nested_replacing_spare.ksh \ functional/cli_root/zpool_add/add-o_ashift.ksh \ functional/cli_root/zpool_add/add_prop_ashift.ksh \ functional/cli_root/zpool_add/cleanup.ksh \ functional/cli_root/zpool_add/setup.ksh \ functional/cli_root/zpool_add/zpool_add_001_pos.ksh \ functional/cli_root/zpool_add/zpool_add_002_pos.ksh \ functional/cli_root/zpool_add/zpool_add_003_pos.ksh \ functional/cli_root/zpool_add/zpool_add_004_pos.ksh \ functional/cli_root/zpool_add/zpool_add_005_pos.ksh \ functional/cli_root/zpool_add/zpool_add_006_pos.ksh \ functional/cli_root/zpool_add/zpool_add_007_neg.ksh \ functional/cli_root/zpool_add/zpool_add_008_neg.ksh \ functional/cli_root/zpool_add/zpool_add_009_neg.ksh \ functional/cli_root/zpool_add/zpool_add_010_pos.ksh \ functional/cli_root/zpool_add/zpool_add_dryrun_output.ksh \ functional/cli_root/zpool_attach/attach-o_ashift.ksh \ functional/cli_root/zpool_attach/cleanup.ksh \ functional/cli_root/zpool_attach/setup.ksh \ functional/cli_root/zpool_attach/zpool_attach_001_neg.ksh \ functional/cli_root/zpool/cleanup.ksh \ functional/cli_root/zpool_clear/cleanup.ksh \ functional/cli_root/zpool_clear/setup.ksh \ functional/cli_root/zpool_clear/zpool_clear_001_pos.ksh \ functional/cli_root/zpool_clear/zpool_clear_002_neg.ksh \ functional/cli_root/zpool_clear/zpool_clear_003_neg.ksh \ functional/cli_root/zpool_clear/zpool_clear_readonly.ksh \ functional/cli_root/zpool_create/cleanup.ksh \ functional/cli_root/zpool_create/create-o_ashift.ksh \ functional/cli_root/zpool_create/setup.ksh \ functional/cli_root/zpool_create/zpool_create_001_pos.ksh \ functional/cli_root/zpool_create/zpool_create_002_pos.ksh \ functional/cli_root/zpool_create/zpool_create_003_pos.ksh \ functional/cli_root/zpool_create/zpool_create_004_pos.ksh \ functional/cli_root/zpool_create/zpool_create_005_pos.ksh \ functional/cli_root/zpool_create/zpool_create_006_pos.ksh \ functional/cli_root/zpool_create/zpool_create_007_neg.ksh \ functional/cli_root/zpool_create/zpool_create_008_pos.ksh \ functional/cli_root/zpool_create/zpool_create_009_neg.ksh \ functional/cli_root/zpool_create/zpool_create_010_neg.ksh \ functional/cli_root/zpool_create/zpool_create_011_neg.ksh \ functional/cli_root/zpool_create/zpool_create_012_neg.ksh \ functional/cli_root/zpool_create/zpool_create_014_neg.ksh \ functional/cli_root/zpool_create/zpool_create_015_neg.ksh \ functional/cli_root/zpool_create/zpool_create_016_pos.ksh \ functional/cli_root/zpool_create/zpool_create_017_neg.ksh \ functional/cli_root/zpool_create/zpool_create_018_pos.ksh \ functional/cli_root/zpool_create/zpool_create_019_pos.ksh \ functional/cli_root/zpool_create/zpool_create_020_pos.ksh \ functional/cli_root/zpool_create/zpool_create_021_pos.ksh \ functional/cli_root/zpool_create/zpool_create_022_pos.ksh \ functional/cli_root/zpool_create/zpool_create_023_neg.ksh \ functional/cli_root/zpool_create/zpool_create_024_pos.ksh \ functional/cli_root/zpool_create/zpool_create_crypt_combos.ksh \ functional/cli_root/zpool_create/zpool_create_draid_001_pos.ksh \ functional/cli_root/zpool_create/zpool_create_draid_002_pos.ksh \ functional/cli_root/zpool_create/zpool_create_draid_003_pos.ksh \ functional/cli_root/zpool_create/zpool_create_draid_004_pos.ksh \ functional/cli_root/zpool_create/zpool_create_dryrun_output.ksh \ functional/cli_root/zpool_create/zpool_create_encrypted.ksh \ functional/cli_root/zpool_create/zpool_create_features_001_pos.ksh \ functional/cli_root/zpool_create/zpool_create_features_002_pos.ksh \ functional/cli_root/zpool_create/zpool_create_features_003_pos.ksh \ functional/cli_root/zpool_create/zpool_create_features_004_neg.ksh \ functional/cli_root/zpool_create/zpool_create_features_005_pos.ksh \ functional/cli_root/zpool_create/zpool_create_features_006_pos.ksh \ functional/cli_root/zpool_create/zpool_create_features_007_pos.ksh \ functional/cli_root/zpool_create/zpool_create_features_008_pos.ksh \ functional/cli_root/zpool_create/zpool_create_features_009_pos.ksh \ functional/cli_root/zpool_create/zpool_create_tempname.ksh \ functional/cli_root/zpool_destroy/zpool_destroy_001_pos.ksh \ functional/cli_root/zpool_destroy/zpool_destroy_002_pos.ksh \ functional/cli_root/zpool_destroy/zpool_destroy_003_neg.ksh \ functional/cli_root/zpool_detach/cleanup.ksh \ functional/cli_root/zpool_detach/setup.ksh \ functional/cli_root/zpool_detach/zpool_detach_001_neg.ksh \ functional/cli_root/zpool_events/cleanup.ksh \ functional/cli_root/zpool_events/setup.ksh \ functional/cli_root/zpool_events/zpool_events_clear.ksh \ functional/cli_root/zpool_events/zpool_events_clear_retained.ksh \ functional/cli_root/zpool_events/zpool_events_cliargs.ksh \ functional/cli_root/zpool_events/zpool_events_duplicates.ksh \ functional/cli_root/zpool_events/zpool_events_errors.ksh \ functional/cli_root/zpool_events/zpool_events_follow.ksh \ functional/cli_root/zpool_events/zpool_events_poolname.ksh \ functional/cli_root/zpool_expand/cleanup.ksh \ functional/cli_root/zpool_expand/setup.ksh \ functional/cli_root/zpool_expand/zpool_expand_001_pos.ksh \ functional/cli_root/zpool_expand/zpool_expand_002_pos.ksh \ functional/cli_root/zpool_expand/zpool_expand_003_neg.ksh \ functional/cli_root/zpool_expand/zpool_expand_004_pos.ksh \ functional/cli_root/zpool_expand/zpool_expand_005_pos.ksh \ functional/cli_root/zpool_export/cleanup.ksh \ functional/cli_root/zpool_export/setup.ksh \ functional/cli_root/zpool_export/zpool_export_001_pos.ksh \ functional/cli_root/zpool_export/zpool_export_002_pos.ksh \ functional/cli_root/zpool_export/zpool_export_003_neg.ksh \ functional/cli_root/zpool_export/zpool_export_004_pos.ksh \ functional/cli_root/zpool_get/cleanup.ksh \ functional/cli_root/zpool_get/setup.ksh \ functional/cli_root/zpool_get/vdev_get_001_pos.ksh \ functional/cli_root/zpool_get/zpool_get_001_pos.ksh \ functional/cli_root/zpool_get/zpool_get_002_pos.ksh \ functional/cli_root/zpool_get/zpool_get_003_pos.ksh \ functional/cli_root/zpool_get/zpool_get_004_neg.ksh \ functional/cli_root/zpool_get/zpool_get_005_pos.ksh \ functional/cli_root/zpool_history/cleanup.ksh \ functional/cli_root/zpool_history/setup.ksh \ functional/cli_root/zpool_history/zpool_history_001_neg.ksh \ functional/cli_root/zpool_history/zpool_history_002_pos.ksh \ functional/cli_root/zpool_import/cleanup.ksh \ functional/cli_root/zpool_import/import_cachefile_device_added.ksh \ functional/cli_root/zpool_import/import_cachefile_device_removed.ksh \ functional/cli_root/zpool_import/import_cachefile_device_replaced.ksh \ functional/cli_root/zpool_import/import_cachefile_mirror_attached.ksh \ functional/cli_root/zpool_import/import_cachefile_mirror_detached.ksh \ functional/cli_root/zpool_import/import_cachefile_paths_changed.ksh \ functional/cli_root/zpool_import/import_cachefile_shared_device.ksh \ functional/cli_root/zpool_import/import_devices_missing.ksh \ functional/cli_root/zpool_import/import_log_missing.ksh \ functional/cli_root/zpool_import/import_paths_changed.ksh \ functional/cli_root/zpool_import/import_rewind_config_changed.ksh \ functional/cli_root/zpool_import/import_rewind_device_replaced.ksh \ functional/cli_root/zpool_import/setup.ksh \ functional/cli_root/zpool_import/zpool_import_001_pos.ksh \ functional/cli_root/zpool_import/zpool_import_002_pos.ksh \ functional/cli_root/zpool_import/zpool_import_003_pos.ksh \ functional/cli_root/zpool_import/zpool_import_004_pos.ksh \ functional/cli_root/zpool_import/zpool_import_005_pos.ksh \ functional/cli_root/zpool_import/zpool_import_006_pos.ksh \ functional/cli_root/zpool_import/zpool_import_007_pos.ksh \ functional/cli_root/zpool_import/zpool_import_008_pos.ksh \ functional/cli_root/zpool_import/zpool_import_009_neg.ksh \ functional/cli_root/zpool_import/zpool_import_010_pos.ksh \ functional/cli_root/zpool_import/zpool_import_011_neg.ksh \ functional/cli_root/zpool_import/zpool_import_012_pos.ksh \ functional/cli_root/zpool_import/zpool_import_013_neg.ksh \ functional/cli_root/zpool_import/zpool_import_014_pos.ksh \ functional/cli_root/zpool_import/zpool_import_015_pos.ksh \ functional/cli_root/zpool_import/zpool_import_016_pos.ksh \ functional/cli_root/zpool_import/zpool_import_017_pos.ksh \ functional/cli_root/zpool_import/zpool_import_all_001_pos.ksh \ functional/cli_root/zpool_import/zpool_import_encrypted.ksh \ functional/cli_root/zpool_import/zpool_import_encrypted_load.ksh \ functional/cli_root/zpool_import/zpool_import_errata3.ksh \ functional/cli_root/zpool_import/zpool_import_errata4.ksh \ functional/cli_root/zpool_import/zpool_import_features_001_pos.ksh \ functional/cli_root/zpool_import/zpool_import_features_002_neg.ksh \ functional/cli_root/zpool_import/zpool_import_features_003_pos.ksh \ functional/cli_root/zpool_import/zpool_import_missing_001_pos.ksh \ functional/cli_root/zpool_import/zpool_import_missing_002_pos.ksh \ functional/cli_root/zpool_import/zpool_import_missing_003_pos.ksh \ functional/cli_root/zpool_import/zpool_import_rename_001_pos.ksh \ functional/cli_root/zpool_initialize/cleanup.ksh \ functional/cli_root/zpool_initialize/zpool_initialize_attach_detach_add_remove.ksh \ functional/cli_root/zpool_initialize/zpool_initialize_fault_export_import_online.ksh \ functional/cli_root/zpool_initialize/zpool_initialize_import_export.ksh \ functional/cli_root/zpool_initialize/zpool_initialize_offline_export_import_online.ksh \ functional/cli_root/zpool_initialize/zpool_initialize_online_offline.ksh \ functional/cli_root/zpool_initialize/zpool_initialize_split.ksh \ functional/cli_root/zpool_initialize/zpool_initialize_start_and_cancel_neg.ksh \ functional/cli_root/zpool_initialize/zpool_initialize_start_and_cancel_pos.ksh \ functional/cli_root/zpool_initialize/zpool_initialize_suspend_resume.ksh \ + functional/cli_root/zpool_initialize/zpool_initialize_uninit.ksh \ functional/cli_root/zpool_initialize/zpool_initialize_unsupported_vdevs.ksh \ functional/cli_root/zpool_initialize/zpool_initialize_verify_checksums.ksh \ functional/cli_root/zpool_initialize/zpool_initialize_verify_initialized.ksh \ functional/cli_root/zpool_labelclear/zpool_labelclear_active.ksh \ functional/cli_root/zpool_labelclear/zpool_labelclear_exported.ksh \ functional/cli_root/zpool_labelclear/zpool_labelclear_removed.ksh \ functional/cli_root/zpool_labelclear/zpool_labelclear_valid.ksh \ functional/cli_root/zpool_offline/cleanup.ksh \ functional/cli_root/zpool_offline/setup.ksh \ functional/cli_root/zpool_offline/zpool_offline_001_pos.ksh \ functional/cli_root/zpool_offline/zpool_offline_002_neg.ksh \ functional/cli_root/zpool_offline/zpool_offline_003_pos.ksh \ functional/cli_root/zpool_online/cleanup.ksh \ functional/cli_root/zpool_online/setup.ksh \ functional/cli_root/zpool_online/zpool_online_001_pos.ksh \ functional/cli_root/zpool_online/zpool_online_002_neg.ksh \ functional/cli_root/zpool_remove/cleanup.ksh \ functional/cli_root/zpool_remove/setup.ksh \ functional/cli_root/zpool_remove/zpool_remove_001_neg.ksh \ functional/cli_root/zpool_remove/zpool_remove_002_pos.ksh \ functional/cli_root/zpool_remove/zpool_remove_003_pos.ksh \ functional/cli_root/zpool_reopen/cleanup.ksh \ functional/cli_root/zpool_reopen/setup.ksh \ functional/cli_root/zpool_reopen/zpool_reopen_001_pos.ksh \ functional/cli_root/zpool_reopen/zpool_reopen_002_pos.ksh \ functional/cli_root/zpool_reopen/zpool_reopen_003_pos.ksh \ functional/cli_root/zpool_reopen/zpool_reopen_004_pos.ksh \ functional/cli_root/zpool_reopen/zpool_reopen_005_pos.ksh \ functional/cli_root/zpool_reopen/zpool_reopen_006_neg.ksh \ functional/cli_root/zpool_reopen/zpool_reopen_007_pos.ksh \ functional/cli_root/zpool_replace/cleanup.ksh \ functional/cli_root/zpool_replace/replace-o_ashift.ksh \ functional/cli_root/zpool_replace/replace_prop_ashift.ksh \ functional/cli_root/zpool_replace/setup.ksh \ functional/cli_root/zpool_replace/zpool_replace_001_neg.ksh \ functional/cli_root/zpool_resilver/cleanup.ksh \ functional/cli_root/zpool_resilver/setup.ksh \ functional/cli_root/zpool_resilver/zpool_resilver_bad_args.ksh \ functional/cli_root/zpool_resilver/zpool_resilver_restart.ksh \ functional/cli_root/zpool_scrub/cleanup.ksh \ functional/cli_root/zpool_scrub/setup.ksh \ functional/cli_root/zpool_scrub/zpool_scrub_001_neg.ksh \ functional/cli_root/zpool_scrub/zpool_scrub_002_pos.ksh \ functional/cli_root/zpool_scrub/zpool_scrub_003_pos.ksh \ functional/cli_root/zpool_scrub/zpool_scrub_004_pos.ksh \ functional/cli_root/zpool_scrub/zpool_scrub_005_pos.ksh \ functional/cli_root/zpool_scrub/zpool_scrub_encrypted_unloaded.ksh \ functional/cli_root/zpool_scrub/zpool_scrub_multiple_copies.ksh \ functional/cli_root/zpool_scrub/zpool_scrub_offline_device.ksh \ functional/cli_root/zpool_scrub/zpool_scrub_print_repairing.ksh \ + functional/cli_root/zpool_scrub/zpool_error_scrub_001_pos.ksh \ + functional/cli_root/zpool_scrub/zpool_error_scrub_002_pos.ksh \ + functional/cli_root/zpool_scrub/zpool_error_scrub_003_pos.ksh \ + functional/cli_root/zpool_scrub/zpool_error_scrub_004_pos.ksh \ functional/cli_root/zpool_set/cleanup.ksh \ functional/cli_root/zpool_set/setup.ksh \ functional/cli_root/zpool/setup.ksh \ functional/cli_root/zpool_set/vdev_set_001_pos.ksh \ functional/cli_root/zpool_set/zpool_set_common.kshlib \ functional/cli_root/zpool_set/zpool_set_001_pos.ksh \ functional/cli_root/zpool_set/zpool_set_002_neg.ksh \ functional/cli_root/zpool_set/zpool_set_003_neg.ksh \ functional/cli_root/zpool_set/zpool_set_ashift.ksh \ functional/cli_root/zpool_set/user_property_001_pos.ksh \ functional/cli_root/zpool_set/user_property_002_neg.ksh \ functional/cli_root/zpool_set/zpool_set_features.ksh \ functional/cli_root/zpool_split/cleanup.ksh \ functional/cli_root/zpool_split/setup.ksh \ functional/cli_root/zpool_split/zpool_split_cliargs.ksh \ functional/cli_root/zpool_split/zpool_split_devices.ksh \ functional/cli_root/zpool_split/zpool_split_dryrun_output.ksh \ functional/cli_root/zpool_split/zpool_split_encryption.ksh \ functional/cli_root/zpool_split/zpool_split_indirect.ksh \ functional/cli_root/zpool_split/zpool_split_props.ksh \ functional/cli_root/zpool_split/zpool_split_resilver.ksh \ functional/cli_root/zpool_split/zpool_split_vdevs.ksh \ functional/cli_root/zpool_split/zpool_split_wholedisk.ksh \ functional/cli_root/zpool_status/cleanup.ksh \ functional/cli_root/zpool_status/setup.ksh \ functional/cli_root/zpool_status/zpool_status_001_pos.ksh \ functional/cli_root/zpool_status/zpool_status_002_pos.ksh \ functional/cli_root/zpool_status/zpool_status_003_pos.ksh \ functional/cli_root/zpool_status/zpool_status_004_pos.ksh \ functional/cli_root/zpool_status/zpool_status_005_pos.ksh \ functional/cli_root/zpool_status/zpool_status_006_pos.ksh \ functional/cli_root/zpool_status/zpool_status_007_pos.ksh \ functional/cli_root/zpool_status/zpool_status_features_001_pos.ksh \ functional/cli_root/zpool_sync/cleanup.ksh \ functional/cli_root/zpool_sync/setup.ksh \ functional/cli_root/zpool_sync/zpool_sync_001_pos.ksh \ functional/cli_root/zpool_sync/zpool_sync_002_neg.ksh \ functional/cli_root/zpool_trim/cleanup.ksh \ functional/cli_root/zpool_trim/setup.ksh \ functional/cli_root/zpool_trim/zpool_trim_attach_detach_add_remove.ksh \ functional/cli_root/zpool_trim/zpool_trim_fault_export_import_online.ksh \ functional/cli_root/zpool_trim/zpool_trim_import_export.ksh \ functional/cli_root/zpool_trim/zpool_trim_multiple.ksh \ functional/cli_root/zpool_trim/zpool_trim_neg.ksh \ functional/cli_root/zpool_trim/zpool_trim_offline_export_import_online.ksh \ functional/cli_root/zpool_trim/zpool_trim_online_offline.ksh \ functional/cli_root/zpool_trim/zpool_trim_partial.ksh \ functional/cli_root/zpool_trim/zpool_trim_rate.ksh \ functional/cli_root/zpool_trim/zpool_trim_rate_neg.ksh \ functional/cli_root/zpool_trim/zpool_trim_secure.ksh \ functional/cli_root/zpool_trim/zpool_trim_split.ksh \ functional/cli_root/zpool_trim/zpool_trim_start_and_cancel_neg.ksh \ functional/cli_root/zpool_trim/zpool_trim_start_and_cancel_pos.ksh \ functional/cli_root/zpool_trim/zpool_trim_suspend_resume.ksh \ functional/cli_root/zpool_trim/zpool_trim_unsupported_vdevs.ksh \ functional/cli_root/zpool_trim/zpool_trim_verify_checksums.ksh \ functional/cli_root/zpool_trim/zpool_trim_verify_trimmed.ksh \ functional/cli_root/zpool_upgrade/cleanup.ksh \ functional/cli_root/zpool_upgrade/setup.ksh \ functional/cli_root/zpool_upgrade/zpool_upgrade_001_pos.ksh \ functional/cli_root/zpool_upgrade/zpool_upgrade_002_pos.ksh \ functional/cli_root/zpool_upgrade/zpool_upgrade_003_pos.ksh \ functional/cli_root/zpool_upgrade/zpool_upgrade_004_pos.ksh \ functional/cli_root/zpool_upgrade/zpool_upgrade_005_neg.ksh \ functional/cli_root/zpool_upgrade/zpool_upgrade_006_neg.ksh \ functional/cli_root/zpool_upgrade/zpool_upgrade_007_pos.ksh \ functional/cli_root/zpool_upgrade/zpool_upgrade_008_pos.ksh \ functional/cli_root/zpool_upgrade/zpool_upgrade_009_neg.ksh \ functional/cli_root/zpool_upgrade/zpool_upgrade_features_001_pos.ksh \ functional/cli_root/zpool_wait/cleanup.ksh \ functional/cli_root/zpool_wait/scan/cleanup.ksh \ functional/cli_root/zpool_wait/scan/setup.ksh \ functional/cli_root/zpool_wait/scan/zpool_wait_rebuild.ksh \ functional/cli_root/zpool_wait/scan/zpool_wait_replace_cancel.ksh \ functional/cli_root/zpool_wait/scan/zpool_wait_replace.ksh \ functional/cli_root/zpool_wait/scan/zpool_wait_resilver.ksh \ functional/cli_root/zpool_wait/scan/zpool_wait_scrub_basic.ksh \ functional/cli_root/zpool_wait/scan/zpool_wait_scrub_cancel.ksh \ functional/cli_root/zpool_wait/scan/zpool_wait_scrub_flag.ksh \ functional/cli_root/zpool_wait/setup.ksh \ functional/cli_root/zpool_wait/zpool_wait_discard.ksh \ functional/cli_root/zpool_wait/zpool_wait_freeing.ksh \ functional/cli_root/zpool_wait/zpool_wait_initialize_basic.ksh \ functional/cli_root/zpool_wait/zpool_wait_initialize_cancel.ksh \ functional/cli_root/zpool_wait/zpool_wait_initialize_flag.ksh \ functional/cli_root/zpool_wait/zpool_wait_multiple.ksh \ functional/cli_root/zpool_wait/zpool_wait_no_activity.ksh \ functional/cli_root/zpool_wait/zpool_wait_remove_cancel.ksh \ functional/cli_root/zpool_wait/zpool_wait_remove.ksh \ functional/cli_root/zpool_wait/zpool_wait_trim_basic.ksh \ functional/cli_root/zpool_wait/zpool_wait_trim_cancel.ksh \ functional/cli_root/zpool_wait/zpool_wait_trim_flag.ksh \ functional/cli_root/zpool_wait/zpool_wait_usage.ksh \ functional/cli_root/zpool/zpool_001_neg.ksh \ functional/cli_root/zpool/zpool_002_pos.ksh \ functional/cli_root/zpool/zpool_003_pos.ksh \ functional/cli_root/zpool/zpool_colors.ksh \ functional/cli_user/misc/arcstat_001_pos.ksh \ functional/cli_user/misc/arc_summary_001_pos.ksh \ functional/cli_user/misc/arc_summary_002_neg.ksh \ functional/cli_user/misc/zilstat_001_pos.ksh \ functional/cli_user/misc/cleanup.ksh \ functional/cli_user/misc/setup.ksh \ functional/cli_user/misc/zdb_001_neg.ksh \ functional/cli_user/misc/zfs_001_neg.ksh \ functional/cli_user/misc/zfs_allow_001_neg.ksh \ functional/cli_user/misc/zfs_clone_001_neg.ksh \ functional/cli_user/misc/zfs_create_001_neg.ksh \ functional/cli_user/misc/zfs_destroy_001_neg.ksh \ functional/cli_user/misc/zfs_get_001_neg.ksh \ functional/cli_user/misc/zfs_inherit_001_neg.ksh \ functional/cli_user/misc/zfs_mount_001_neg.ksh \ functional/cli_user/misc/zfs_promote_001_neg.ksh \ functional/cli_user/misc/zfs_receive_001_neg.ksh \ functional/cli_user/misc/zfs_rename_001_neg.ksh \ functional/cli_user/misc/zfs_rollback_001_neg.ksh \ functional/cli_user/misc/zfs_send_001_neg.ksh \ functional/cli_user/misc/zfs_set_001_neg.ksh \ functional/cli_user/misc/zfs_share_001_neg.ksh \ functional/cli_user/misc/zfs_snapshot_001_neg.ksh \ functional/cli_user/misc/zfs_unallow_001_neg.ksh \ functional/cli_user/misc/zfs_unmount_001_neg.ksh \ functional/cli_user/misc/zfs_unshare_001_neg.ksh \ functional/cli_user/misc/zfs_upgrade_001_neg.ksh \ functional/cli_user/misc/zpool_001_neg.ksh \ functional/cli_user/misc/zpool_add_001_neg.ksh \ functional/cli_user/misc/zpool_attach_001_neg.ksh \ functional/cli_user/misc/zpool_clear_001_neg.ksh \ functional/cli_user/misc/zpool_create_001_neg.ksh \ functional/cli_user/misc/zpool_destroy_001_neg.ksh \ functional/cli_user/misc/zpool_detach_001_neg.ksh \ functional/cli_user/misc/zpool_export_001_neg.ksh \ functional/cli_user/misc/zpool_get_001_neg.ksh \ functional/cli_user/misc/zpool_history_001_neg.ksh \ functional/cli_user/misc/zpool_import_001_neg.ksh \ functional/cli_user/misc/zpool_import_002_neg.ksh \ functional/cli_user/misc/zpool_offline_001_neg.ksh \ functional/cli_user/misc/zpool_online_001_neg.ksh \ functional/cli_user/misc/zpool_remove_001_neg.ksh \ functional/cli_user/misc/zpool_replace_001_neg.ksh \ functional/cli_user/misc/zpool_scrub_001_neg.ksh \ functional/cli_user/misc/zpool_set_001_neg.ksh \ functional/cli_user/misc/zpool_status_001_neg.ksh \ functional/cli_user/misc/zpool_upgrade_001_neg.ksh \ functional/cli_user/misc/zpool_wait_privilege.ksh \ functional/cli_user/zfs_list/cleanup.ksh \ functional/cli_user/zfs_list/setup.ksh \ functional/cli_user/zfs_list/zfs_list_001_pos.ksh \ functional/cli_user/zfs_list/zfs_list_002_pos.ksh \ functional/cli_user/zfs_list/zfs_list_003_pos.ksh \ functional/cli_user/zfs_list/zfs_list_004_neg.ksh \ functional/cli_user/zfs_list/zfs_list_005_neg.ksh \ functional/cli_user/zfs_list/zfs_list_007_pos.ksh \ functional/cli_user/zfs_list/zfs_list_008_neg.ksh \ functional/cli_user/zpool_iostat/cleanup.ksh \ functional/cli_user/zpool_iostat/setup.ksh \ functional/cli_user/zpool_iostat/zpool_iostat_001_neg.ksh \ functional/cli_user/zpool_iostat/zpool_iostat_002_pos.ksh \ functional/cli_user/zpool_iostat/zpool_iostat_003_neg.ksh \ functional/cli_user/zpool_iostat/zpool_iostat_004_pos.ksh \ functional/cli_user/zpool_iostat/zpool_iostat_005_pos.ksh \ functional/cli_user/zpool_iostat/zpool_iostat_-c_disable.ksh \ functional/cli_user/zpool_iostat/zpool_iostat_-c_homedir.ksh \ functional/cli_user/zpool_iostat/zpool_iostat_-c_searchpath.ksh \ functional/cli_user/zpool_list/cleanup.ksh \ functional/cli_user/zpool_list/setup.ksh \ functional/cli_user/zpool_list/zpool_list_001_pos.ksh \ functional/cli_user/zpool_list/zpool_list_002_neg.ksh \ functional/cli_user/zpool_status/cleanup.ksh \ functional/cli_user/zpool_status/setup.ksh \ functional/cli_user/zpool_status/zpool_status_003_pos.ksh \ functional/cli_user/zpool_status/zpool_status_-c_disable.ksh \ functional/cli_user/zpool_status/zpool_status_-c_homedir.ksh \ functional/cli_user/zpool_status/zpool_status_-c_searchpath.ksh \ functional/compression/cleanup.ksh \ functional/compression/compress_001_pos.ksh \ functional/compression/compress_002_pos.ksh \ functional/compression/compress_003_pos.ksh \ functional/compression/compress_004_pos.ksh \ functional/compression/compress_zstd_bswap.ksh \ functional/compression/l2arc_compressed_arc_disabled.ksh \ functional/compression/l2arc_compressed_arc.ksh \ functional/compression/l2arc_encrypted.ksh \ functional/compression/l2arc_encrypted_no_compressed_arc.ksh \ functional/compression/setup.ksh \ functional/cp_files/cleanup.ksh \ functional/cp_files/cp_files_001_pos.ksh \ functional/cp_files/setup.ksh \ functional/crtime/cleanup.ksh \ functional/crtime/crtime_001_pos.ksh \ functional/crtime/setup.ksh \ functional/ctime/cleanup.ksh \ functional/ctime/ctime_001_pos.ksh \ functional/ctime/setup.ksh \ functional/deadman/deadman_ratelimit.ksh \ functional/deadman/deadman_sync.ksh \ functional/deadman/deadman_zio.ksh \ functional/delegate/cleanup.ksh \ functional/delegate/setup.ksh \ functional/delegate/zfs_allow_001_pos.ksh \ functional/delegate/zfs_allow_002_pos.ksh \ functional/delegate/zfs_allow_003_pos.ksh \ functional/delegate/zfs_allow_004_pos.ksh \ functional/delegate/zfs_allow_005_pos.ksh \ functional/delegate/zfs_allow_006_pos.ksh \ functional/delegate/zfs_allow_007_pos.ksh \ functional/delegate/zfs_allow_008_pos.ksh \ functional/delegate/zfs_allow_009_neg.ksh \ functional/delegate/zfs_allow_010_pos.ksh \ functional/delegate/zfs_allow_011_neg.ksh \ functional/delegate/zfs_allow_012_neg.ksh \ functional/delegate/zfs_unallow_001_pos.ksh \ functional/delegate/zfs_unallow_002_pos.ksh \ functional/delegate/zfs_unallow_003_pos.ksh \ functional/delegate/zfs_unallow_004_pos.ksh \ functional/delegate/zfs_unallow_005_pos.ksh \ functional/delegate/zfs_unallow_006_pos.ksh \ functional/delegate/zfs_unallow_007_neg.ksh \ functional/delegate/zfs_unallow_008_neg.ksh \ functional/devices/cleanup.ksh \ functional/devices/devices_001_pos.ksh \ functional/devices/devices_002_neg.ksh \ functional/devices/devices_003_pos.ksh \ functional/devices/setup.ksh \ functional/dos_attributes/cleanup.ksh \ functional/dos_attributes/read_dos_attrs_001.ksh \ functional/dos_attributes/setup.ksh \ functional/dos_attributes/write_dos_attrs_001.ksh \ functional/events/cleanup.ksh \ functional/events/events_001_pos.ksh \ functional/events/events_002_pos.ksh \ functional/events/setup.ksh \ functional/events/zed_cksum_config.ksh \ functional/events/zed_cksum_reported.ksh \ functional/events/zed_fd_spill.ksh \ functional/events/zed_io_config.ksh \ functional/events/zed_rc_filter.ksh \ functional/exec/cleanup.ksh \ functional/exec/exec_001_pos.ksh \ functional/exec/exec_002_neg.ksh \ functional/exec/setup.ksh \ functional/fadvise/cleanup.ksh \ functional/fadvise/fadvise_sequential.ksh \ functional/fadvise/setup.ksh \ functional/fallocate/cleanup.ksh \ functional/fallocate/fallocate_prealloc.ksh \ functional/fallocate/fallocate_punch-hole.ksh \ functional/fallocate/fallocate_zero-range.ksh \ functional/fallocate/setup.ksh \ functional/fault/auto_offline_001_pos.ksh \ functional/fault/auto_online_001_pos.ksh \ functional/fault/auto_online_002_pos.ksh \ functional/fault/auto_replace_001_pos.ksh \ functional/fault/auto_spare_001_pos.ksh \ functional/fault/auto_spare_002_pos.ksh \ functional/fault/auto_spare_ashift.ksh \ functional/fault/auto_spare_multiple.ksh \ functional/fault/auto_spare_shared.ksh \ functional/fault/cleanup.ksh \ functional/fault/decompress_fault.ksh \ functional/fault/decrypt_fault.ksh \ functional/fault/scrub_after_resilver.ksh \ functional/fault/setup.ksh \ functional/fault/zpool_status_-s.ksh \ functional/features/async_destroy/async_destroy_001_pos.ksh \ functional/features/async_destroy/cleanup.ksh \ functional/features/async_destroy/setup.ksh \ functional/features/large_dnode/cleanup.ksh \ functional/features/large_dnode/large_dnode_001_pos.ksh \ functional/features/large_dnode/large_dnode_002_pos.ksh \ functional/features/large_dnode/large_dnode_003_pos.ksh \ functional/features/large_dnode/large_dnode_004_neg.ksh \ functional/features/large_dnode/large_dnode_005_pos.ksh \ functional/features/large_dnode/large_dnode_006_pos.ksh \ functional/features/large_dnode/large_dnode_007_neg.ksh \ functional/features/large_dnode/large_dnode_008_pos.ksh \ functional/features/large_dnode/large_dnode_009_pos.ksh \ functional/features/large_dnode/setup.ksh \ functional/grow/grow_pool_001_pos.ksh \ functional/grow/grow_replicas_001_pos.ksh \ functional/history/cleanup.ksh \ functional/history/history_001_pos.ksh \ functional/history/history_002_pos.ksh \ functional/history/history_003_pos.ksh \ functional/history/history_004_pos.ksh \ functional/history/history_005_neg.ksh \ functional/history/history_006_neg.ksh \ functional/history/history_007_pos.ksh \ functional/history/history_008_pos.ksh \ functional/history/history_009_pos.ksh \ functional/history/history_010_pos.ksh \ functional/history/setup.ksh \ functional/inheritance/cleanup.ksh \ functional/inheritance/inherit_001_pos.ksh \ functional/inuse/inuse_001_pos.ksh \ functional/inuse/inuse_003_pos.ksh \ functional/inuse/inuse_004_pos.ksh \ functional/inuse/inuse_005_pos.ksh \ functional/inuse/inuse_006_pos.ksh \ functional/inuse/inuse_007_pos.ksh \ functional/inuse/inuse_008_pos.ksh \ functional/inuse/inuse_009_pos.ksh \ functional/inuse/setup.ksh \ functional/io/cleanup.ksh \ functional/io/io_uring.ksh \ functional/io/libaio.ksh \ functional/io/mmap.ksh \ functional/io/posixaio.ksh \ functional/io/psync.ksh \ functional/io/setup.ksh \ functional/io/sync.ksh \ functional/l2arc/cleanup.ksh \ functional/l2arc/l2arc_arcstats_pos.ksh \ functional/l2arc/l2arc_l2miss_pos.ksh \ functional/l2arc/l2arc_mfuonly_pos.ksh \ functional/l2arc/persist_l2arc_001_pos.ksh \ functional/l2arc/persist_l2arc_002_pos.ksh \ functional/l2arc/persist_l2arc_003_neg.ksh \ functional/l2arc/persist_l2arc_004_pos.ksh \ functional/l2arc/persist_l2arc_005_pos.ksh \ functional/l2arc/setup.ksh \ functional/large_files/cleanup.ksh \ functional/large_files/large_files_001_pos.ksh \ functional/large_files/large_files_002_pos.ksh \ functional/large_files/setup.ksh \ functional/largest_pool/largest_pool_001_pos.ksh \ functional/libzfs/cleanup.ksh \ functional/libzfs/libzfs_input.ksh \ functional/libzfs/setup.ksh \ functional/limits/cleanup.ksh \ functional/limits/filesystem_count.ksh \ functional/limits/filesystem_limit.ksh \ functional/limits/setup.ksh \ functional/limits/snapshot_count.ksh \ functional/limits/snapshot_limit.ksh \ functional/link_count/cleanup.ksh \ functional/link_count/link_count_001.ksh \ functional/link_count/link_count_root_inode.ksh \ functional/link_count/setup.ksh \ functional/log_spacemap/log_spacemap_import_logs.ksh \ functional/migration/cleanup.ksh \ functional/migration/migration_001_pos.ksh \ functional/migration/migration_002_pos.ksh \ functional/migration/migration_003_pos.ksh \ functional/migration/migration_004_pos.ksh \ functional/migration/migration_005_pos.ksh \ functional/migration/migration_006_pos.ksh \ functional/migration/migration_007_pos.ksh \ functional/migration/migration_008_pos.ksh \ functional/migration/migration_009_pos.ksh \ functional/migration/migration_010_pos.ksh \ functional/migration/migration_011_pos.ksh \ functional/migration/migration_012_pos.ksh \ functional/migration/setup.ksh \ functional/mmap/cleanup.ksh \ functional/mmap/mmap_libaio_001_pos.ksh \ functional/mmap/mmap_mixed.ksh \ functional/mmap/mmap_read_001_pos.ksh \ functional/mmap/mmap_seek_001_pos.ksh \ functional/mmap/mmap_sync_001_pos.ksh \ functional/mmap/mmap_write_001_pos.ksh \ functional/mmap/setup.ksh \ functional/mmp/cleanup.ksh \ functional/mmp/mmp_active_import.ksh \ functional/mmp/mmp_exported_import.ksh \ functional/mmp/mmp_hostid.ksh \ functional/mmp/mmp_inactive_import.ksh \ functional/mmp/mmp_interval.ksh \ functional/mmp/mmp_on_off.ksh \ functional/mmp/mmp_on_thread.ksh \ functional/mmp/mmp_on_uberblocks.ksh \ functional/mmp/mmp_on_zdb.ksh \ functional/mmp/mmp_reset_interval.ksh \ functional/mmp/mmp_write_distribution.ksh \ functional/mmp/mmp_write_uberblocks.ksh \ functional/mmp/multihost_history.ksh \ functional/mmp/setup.ksh \ functional/mount/cleanup.ksh \ functional/mount/setup.ksh \ functional/mount/umount_001.ksh \ functional/mount/umountall_001.ksh \ functional/mount/umount_unlinked_drain.ksh \ functional/mv_files/cleanup.ksh \ functional/mv_files/mv_files_001_pos.ksh \ functional/mv_files/mv_files_002_pos.ksh \ functional/mv_files/random_creation.ksh \ functional/mv_files/setup.ksh \ functional/nestedfs/cleanup.ksh \ functional/nestedfs/nestedfs_001_pos.ksh \ functional/nestedfs/setup.ksh \ functional/nopwrite/cleanup.ksh \ functional/nopwrite/nopwrite_copies.ksh \ functional/nopwrite/nopwrite_mtime.ksh \ functional/nopwrite/nopwrite_negative.ksh \ functional/nopwrite/nopwrite_promoted_clone.ksh \ functional/nopwrite/nopwrite_recsize.ksh \ functional/nopwrite/nopwrite_sync.ksh \ functional/nopwrite/nopwrite_varying_compression.ksh \ functional/nopwrite/nopwrite_volume.ksh \ functional/nopwrite/setup.ksh \ functional/no_space/cleanup.ksh \ functional/no_space/enospc_001_pos.ksh \ functional/no_space/enospc_002_pos.ksh \ functional/no_space/enospc_003_pos.ksh \ functional/no_space/enospc_df.ksh \ functional/no_space/enospc_ganging.ksh \ functional/no_space/enospc_rm.ksh \ functional/no_space/setup.ksh \ functional/online_offline/cleanup.ksh \ functional/online_offline/online_offline_001_pos.ksh \ functional/online_offline/online_offline_002_neg.ksh \ functional/online_offline/online_offline_003_neg.ksh \ functional/online_offline/setup.ksh \ functional/pam/cleanup.ksh \ functional/pam/pam_basic.ksh \ functional/pam/pam_nounmount.ksh \ functional/pam/pam_short_password.ksh \ functional/pam/setup.ksh \ functional/pool_checkpoint/checkpoint_after_rewind.ksh \ functional/pool_checkpoint/checkpoint_big_rewind.ksh \ functional/pool_checkpoint/checkpoint_capacity.ksh \ functional/pool_checkpoint/checkpoint_conf_change.ksh \ functional/pool_checkpoint/checkpoint_discard_busy.ksh \ functional/pool_checkpoint/checkpoint_discard.ksh \ functional/pool_checkpoint/checkpoint_discard_many.ksh \ functional/pool_checkpoint/checkpoint_indirect.ksh \ functional/pool_checkpoint/checkpoint_invalid.ksh \ functional/pool_checkpoint/checkpoint_lun_expsz.ksh \ functional/pool_checkpoint/checkpoint_open.ksh \ functional/pool_checkpoint/checkpoint_removal.ksh \ functional/pool_checkpoint/checkpoint_rewind.ksh \ functional/pool_checkpoint/checkpoint_ro_rewind.ksh \ functional/pool_checkpoint/checkpoint_sm_scale.ksh \ functional/pool_checkpoint/checkpoint_twice.ksh \ functional/pool_checkpoint/checkpoint_vdev_add.ksh \ functional/pool_checkpoint/checkpoint_zdb.ksh \ functional/pool_checkpoint/checkpoint_zhack_feat.ksh \ functional/pool_checkpoint/cleanup.ksh \ functional/pool_checkpoint/setup.ksh \ functional/pool_names/pool_names_001_pos.ksh \ functional/pool_names/pool_names_002_neg.ksh \ functional/poolversion/cleanup.ksh \ functional/poolversion/poolversion_001_pos.ksh \ functional/poolversion/poolversion_002_pos.ksh \ functional/poolversion/setup.ksh \ functional/privilege/cleanup.ksh \ functional/privilege/privilege_001_pos.ksh \ functional/privilege/privilege_002_pos.ksh \ functional/privilege/setup.ksh \ functional/procfs/cleanup.ksh \ functional/procfs/pool_state.ksh \ functional/procfs/procfs_list_basic.ksh \ functional/procfs/procfs_list_concurrent_readers.ksh \ functional/procfs/procfs_list_stale_read.ksh \ functional/procfs/setup.ksh \ functional/projectquota/cleanup.ksh \ functional/projectquota/projectid_001_pos.ksh \ functional/projectquota/projectid_002_pos.ksh \ functional/projectquota/projectid_003_pos.ksh \ functional/projectquota/projectquota_001_pos.ksh \ functional/projectquota/projectquota_002_pos.ksh \ functional/projectquota/projectquota_003_pos.ksh \ functional/projectquota/projectquota_004_neg.ksh \ functional/projectquota/projectquota_005_pos.ksh \ functional/projectquota/projectquota_006_pos.ksh \ functional/projectquota/projectquota_007_pos.ksh \ functional/projectquota/projectquota_008_pos.ksh \ functional/projectquota/projectquota_009_pos.ksh \ functional/projectquota/projectspace_001_pos.ksh \ functional/projectquota/projectspace_002_pos.ksh \ functional/projectquota/projectspace_003_pos.ksh \ functional/projectquota/projectspace_004_pos.ksh \ functional/projectquota/projecttree_001_pos.ksh \ functional/projectquota/projecttree_002_pos.ksh \ functional/projectquota/projecttree_003_neg.ksh \ functional/projectquota/setup.ksh \ functional/quota/cleanup.ksh \ functional/quota/quota_001_pos.ksh \ functional/quota/quota_002_pos.ksh \ functional/quota/quota_003_pos.ksh \ functional/quota/quota_004_pos.ksh \ functional/quota/quota_005_pos.ksh \ functional/quota/quota_006_neg.ksh \ functional/quota/setup.ksh \ functional/raidz/cleanup.ksh \ functional/raidz/raidz_001_neg.ksh \ functional/raidz/raidz_002_pos.ksh \ functional/raidz/raidz_003_pos.ksh \ functional/raidz/raidz_004_pos.ksh \ functional/raidz/setup.ksh \ functional/redacted_send/cleanup.ksh \ functional/redacted_send/redacted_compressed.ksh \ functional/redacted_send/redacted_contents.ksh \ functional/redacted_send/redacted_deleted.ksh \ functional/redacted_send/redacted_disabled_feature.ksh \ functional/redacted_send/redacted_embedded.ksh \ functional/redacted_send/redacted_holes.ksh \ functional/redacted_send/redacted_incrementals.ksh \ functional/redacted_send/redacted_largeblocks.ksh \ functional/redacted_send/redacted_many_clones.ksh \ functional/redacted_send/redacted_mixed_recsize.ksh \ functional/redacted_send/redacted_mounts.ksh \ functional/redacted_send/redacted_negative.ksh \ functional/redacted_send/redacted_origin.ksh \ functional/redacted_send/redacted_panic.ksh \ functional/redacted_send/redacted_props.ksh \ functional/redacted_send/redacted_resume.ksh \ functional/redacted_send/redacted_size.ksh \ functional/redacted_send/redacted_volume.ksh \ functional/redacted_send/setup.ksh \ functional/redundancy/cleanup.ksh \ functional/redundancy/redundancy_draid1.ksh \ functional/redundancy/redundancy_draid2.ksh \ functional/redundancy/redundancy_draid3.ksh \ functional/redundancy/redundancy_draid_damaged1.ksh \ functional/redundancy/redundancy_draid_damaged2.ksh \ functional/redundancy/redundancy_draid.ksh \ functional/redundancy/redundancy_draid_spare1.ksh \ functional/redundancy/redundancy_draid_spare2.ksh \ functional/redundancy/redundancy_draid_spare3.ksh \ functional/redundancy/redundancy_mirror.ksh \ functional/redundancy/redundancy_raidz1.ksh \ functional/redundancy/redundancy_raidz2.ksh \ functional/redundancy/redundancy_raidz3.ksh \ functional/redundancy/redundancy_raidz.ksh \ functional/redundancy/redundancy_stripe.ksh \ functional/redundancy/setup.ksh \ functional/refquota/cleanup.ksh \ functional/refquota/refquota_001_pos.ksh \ functional/refquota/refquota_002_pos.ksh \ functional/refquota/refquota_003_pos.ksh \ functional/refquota/refquota_004_pos.ksh \ functional/refquota/refquota_005_pos.ksh \ functional/refquota/refquota_006_neg.ksh \ functional/refquota/refquota_007_neg.ksh \ functional/refquota/refquota_008_neg.ksh \ functional/refquota/setup.ksh \ functional/refreserv/cleanup.ksh \ functional/refreserv/refreserv_001_pos.ksh \ functional/refreserv/refreserv_002_pos.ksh \ functional/refreserv/refreserv_003_pos.ksh \ functional/refreserv/refreserv_004_pos.ksh \ functional/refreserv/refreserv_005_pos.ksh \ functional/refreserv/refreserv_multi_raidz.ksh \ functional/refreserv/refreserv_raidz.ksh \ functional/refreserv/setup.ksh \ functional/removal/cleanup.ksh \ functional/removal/removal_all_vdev.ksh \ functional/removal/removal_cancel.ksh \ functional/removal/removal_check_space.ksh \ functional/removal/removal_condense_export.ksh \ functional/removal/removal_multiple_indirection.ksh \ functional/removal/removal_nopwrite.ksh \ functional/removal/removal_remap_deadlists.ksh \ functional/removal/removal_reservation.ksh \ functional/removal/removal_resume_export.ksh \ functional/removal/removal_sanity.ksh \ functional/removal/removal_with_add.ksh \ functional/removal/removal_with_create_fs.ksh \ functional/removal/removal_with_dedup.ksh \ functional/removal/removal_with_errors.ksh \ functional/removal/removal_with_export.ksh \ functional/removal/removal_with_faulted.ksh \ functional/removal/removal_with_ganging.ksh \ functional/removal/removal_with_indirect.ksh \ functional/removal/removal_with_remove.ksh \ functional/removal/removal_with_scrub.ksh \ functional/removal/removal_with_send.ksh \ functional/removal/removal_with_send_recv.ksh \ functional/removal/removal_with_snapshot.ksh \ functional/removal/removal_with_write.ksh \ functional/removal/removal_with_zdb.ksh \ functional/removal/remove_attach_mirror.ksh \ functional/removal/remove_expanded.ksh \ functional/removal/remove_indirect.ksh \ functional/removal/remove_mirror.ksh \ functional/removal/remove_mirror_sanity.ksh \ functional/removal/remove_raidz.ksh \ functional/rename_dirs/cleanup.ksh \ functional/rename_dirs/rename_dirs_001_pos.ksh \ functional/rename_dirs/setup.ksh \ functional/renameat2/cleanup.ksh \ functional/renameat2/setup.ksh \ functional/renameat2/renameat2_exchange.ksh \ functional/renameat2/renameat2_noreplace.ksh \ functional/renameat2/renameat2_whiteout.ksh \ functional/replacement/attach_import.ksh \ functional/replacement/attach_multiple.ksh \ functional/replacement/attach_rebuild.ksh \ functional/replacement/attach_resilver.ksh \ functional/replacement/cleanup.ksh \ functional/replacement/detach.ksh \ functional/replacement/rebuild_disabled_feature.ksh \ functional/replacement/rebuild_multiple.ksh \ functional/replacement/rebuild_raidz.ksh \ functional/replacement/replace_import.ksh \ functional/replacement/replace_rebuild.ksh \ functional/replacement/replace_resilver.ksh \ functional/replacement/resilver_restart_001.ksh \ functional/replacement/resilver_restart_002.ksh \ functional/replacement/scrub_cancel.ksh \ functional/replacement/setup.ksh \ functional/reservation/cleanup.ksh \ functional/reservation/reservation_001_pos.ksh \ functional/reservation/reservation_002_pos.ksh \ functional/reservation/reservation_003_pos.ksh \ functional/reservation/reservation_004_pos.ksh \ functional/reservation/reservation_005_pos.ksh \ functional/reservation/reservation_006_pos.ksh \ functional/reservation/reservation_007_pos.ksh \ functional/reservation/reservation_008_pos.ksh \ functional/reservation/reservation_009_pos.ksh \ functional/reservation/reservation_010_pos.ksh \ functional/reservation/reservation_011_pos.ksh \ functional/reservation/reservation_012_pos.ksh \ functional/reservation/reservation_013_pos.ksh \ functional/reservation/reservation_014_pos.ksh \ functional/reservation/reservation_015_pos.ksh \ functional/reservation/reservation_016_pos.ksh \ functional/reservation/reservation_017_pos.ksh \ functional/reservation/reservation_018_pos.ksh \ functional/reservation/reservation_019_pos.ksh \ functional/reservation/reservation_020_pos.ksh \ functional/reservation/reservation_021_neg.ksh \ functional/reservation/reservation_022_pos.ksh \ functional/reservation/setup.ksh \ functional/rootpool/cleanup.ksh \ functional/rootpool/rootpool_002_neg.ksh \ functional/rootpool/rootpool_003_neg.ksh \ functional/rootpool/rootpool_007_pos.ksh \ functional/rootpool/setup.ksh \ functional/rsend/cleanup.ksh \ functional/rsend/recv_dedup_encrypted_zvol.ksh \ functional/rsend/recv_dedup.ksh \ functional/rsend/rsend_001_pos.ksh \ functional/rsend/rsend_002_pos.ksh \ functional/rsend/rsend_003_pos.ksh \ functional/rsend/rsend_004_pos.ksh \ functional/rsend/rsend_005_pos.ksh \ functional/rsend/rsend_006_pos.ksh \ functional/rsend/rsend_007_pos.ksh \ functional/rsend/rsend_008_pos.ksh \ functional/rsend/rsend_009_pos.ksh \ functional/rsend/rsend_010_pos.ksh \ functional/rsend/rsend_011_pos.ksh \ functional/rsend/rsend_012_pos.ksh \ functional/rsend/rsend_013_pos.ksh \ functional/rsend/rsend_014_pos.ksh \ functional/rsend/rsend_016_neg.ksh \ functional/rsend/rsend_019_pos.ksh \ functional/rsend/rsend_020_pos.ksh \ functional/rsend/rsend_021_pos.ksh \ functional/rsend/rsend_022_pos.ksh \ functional/rsend/rsend_024_pos.ksh \ functional/rsend/rsend_025_pos.ksh \ functional/rsend/rsend_026_neg.ksh \ functional/rsend/rsend_027_pos.ksh \ functional/rsend/rsend_028_neg.ksh \ functional/rsend/rsend_029_neg.ksh \ functional/rsend/rsend_030_pos.ksh \ functional/rsend/rsend_031_pos.ksh \ functional/rsend/send-c_embedded_blocks.ksh \ functional/rsend/send-c_incremental.ksh \ functional/rsend/send-c_lz4_disabled.ksh \ functional/rsend/send-c_mixed_compression.ksh \ functional/rsend/send-c_props.ksh \ functional/rsend/send-c_recv_dedup.ksh \ functional/rsend/send-c_recv_lz4_disabled.ksh \ functional/rsend/send-c_resume.ksh \ functional/rsend/send-c_stream_size_estimate.ksh \ functional/rsend/send-c_verify_contents.ksh \ functional/rsend/send-c_verify_ratio.ksh \ functional/rsend/send-c_volume.ksh \ functional/rsend/send-c_zstream_recompress.ksh \ functional/rsend/send-c_zstreamdump.ksh \ functional/rsend/send-cpL_varied_recsize.ksh \ functional/rsend/send_doall.ksh \ functional/rsend/send_encrypted_incremental.ksh \ functional/rsend/send_encrypted_files.ksh \ functional/rsend/send_encrypted_freeobjects.ksh \ functional/rsend/send_encrypted_hierarchy.ksh \ functional/rsend/send_encrypted_props.ksh \ functional/rsend/send_encrypted_truncated_files.ksh \ functional/rsend/send_freeobjects.ksh \ functional/rsend/send_holds.ksh \ functional/rsend/send_hole_birth.ksh \ functional/rsend/send_invalid.ksh \ functional/rsend/send-L_toggle.ksh \ functional/rsend/send_mixed_raw.ksh \ functional/rsend/send_partial_dataset.ksh \ functional/rsend/send_raw_ashift.ksh \ functional/rsend/send_raw_spill_block.ksh \ functional/rsend/send_raw_large_blocks.ksh \ functional/rsend/send_realloc_dnode_size.ksh \ functional/rsend/send_realloc_encrypted_files.ksh \ functional/rsend/send_realloc_files.ksh \ functional/rsend/send_spill_block.ksh \ functional/rsend/send-wR_encrypted_zvol.ksh \ functional/rsend/setup.ksh \ functional/scrub_mirror/cleanup.ksh \ functional/scrub_mirror/scrub_mirror_001_pos.ksh \ functional/scrub_mirror/scrub_mirror_002_pos.ksh \ functional/scrub_mirror/scrub_mirror_003_pos.ksh \ functional/scrub_mirror/scrub_mirror_004_pos.ksh \ functional/scrub_mirror/setup.ksh \ functional/slog/cleanup.ksh \ functional/slog/setup.ksh \ functional/slog/slog_001_pos.ksh \ functional/slog/slog_002_pos.ksh \ functional/slog/slog_003_pos.ksh \ functional/slog/slog_004_pos.ksh \ functional/slog/slog_005_pos.ksh \ functional/slog/slog_006_pos.ksh \ functional/slog/slog_007_pos.ksh \ functional/slog/slog_008_neg.ksh \ functional/slog/slog_009_neg.ksh \ functional/slog/slog_010_neg.ksh \ functional/slog/slog_011_neg.ksh \ functional/slog/slog_012_neg.ksh \ functional/slog/slog_013_pos.ksh \ functional/slog/slog_014_pos.ksh \ functional/slog/slog_015_neg.ksh \ functional/slog/slog_016_pos.ksh \ functional/slog/slog_replay_fs_001.ksh \ functional/slog/slog_replay_fs_002.ksh \ functional/slog/slog_replay_volume.ksh \ functional/snapshot/cleanup.ksh \ functional/snapshot/clone_001_pos.ksh \ functional/snapshot/rollback_001_pos.ksh \ functional/snapshot/rollback_002_pos.ksh \ functional/snapshot/rollback_003_pos.ksh \ functional/snapshot/setup.ksh \ functional/snapshot/snapshot_001_pos.ksh \ functional/snapshot/snapshot_002_pos.ksh \ functional/snapshot/snapshot_003_pos.ksh \ functional/snapshot/snapshot_004_pos.ksh \ functional/snapshot/snapshot_005_pos.ksh \ functional/snapshot/snapshot_006_pos.ksh \ functional/snapshot/snapshot_007_pos.ksh \ functional/snapshot/snapshot_008_pos.ksh \ functional/snapshot/snapshot_009_pos.ksh \ functional/snapshot/snapshot_010_pos.ksh \ functional/snapshot/snapshot_011_pos.ksh \ functional/snapshot/snapshot_012_pos.ksh \ functional/snapshot/snapshot_013_pos.ksh \ functional/snapshot/snapshot_014_pos.ksh \ functional/snapshot/snapshot_015_pos.ksh \ functional/snapshot/snapshot_016_pos.ksh \ functional/snapshot/snapshot_017_pos.ksh \ functional/snapshot/snapshot_018_pos.ksh \ functional/snapused/cleanup.ksh \ functional/snapused/setup.ksh \ functional/snapused/snapused_001_pos.ksh \ functional/snapused/snapused_002_pos.ksh \ functional/snapused/snapused_003_pos.ksh \ functional/snapused/snapused_004_pos.ksh \ functional/snapused/snapused_005_pos.ksh \ functional/sparse/cleanup.ksh \ functional/sparse/setup.ksh \ functional/sparse/sparse_001_pos.ksh \ functional/stat/cleanup.ksh \ functional/stat/setup.ksh \ functional/stat/stat_001_pos.ksh \ functional/suid/cleanup.ksh \ functional/suid/setup.ksh \ functional/suid/suid_write_to_none.ksh \ functional/suid/suid_write_to_sgid.ksh \ functional/suid/suid_write_to_suid.ksh \ functional/suid/suid_write_to_suid_sgid.ksh \ functional/suid/suid_write_zil_replay.ksh \ functional/trim/autotrim_config.ksh \ functional/trim/autotrim_integrity.ksh \ functional/trim/autotrim_trim_integrity.ksh \ functional/trim/cleanup.ksh \ functional/trim/setup.ksh \ functional/trim/trim_config.ksh \ functional/trim/trim_integrity.ksh \ functional/trim/trim_l2arc.ksh \ functional/truncate/cleanup.ksh \ functional/truncate/setup.ksh \ functional/truncate/truncate_001_pos.ksh \ functional/truncate/truncate_002_pos.ksh \ functional/truncate/truncate_timestamps.ksh \ functional/upgrade/cleanup.ksh \ functional/upgrade/setup.ksh \ functional/upgrade/upgrade_projectquota_001_pos.ksh \ functional/upgrade/upgrade_readonly_pool.ksh \ functional/upgrade/upgrade_userobj_001_pos.ksh \ functional/user_namespace/cleanup.ksh \ functional/user_namespace/setup.ksh \ functional/user_namespace/user_namespace_001.ksh \ functional/user_namespace/user_namespace_002.ksh \ functional/user_namespace/user_namespace_003.ksh \ functional/user_namespace/user_namespace_004.ksh \ functional/userquota/cleanup.ksh \ functional/userquota/groupspace_001_pos.ksh \ functional/userquota/groupspace_002_pos.ksh \ functional/userquota/groupspace_003_pos.ksh \ functional/userquota/setup.ksh \ functional/userquota/userquota_001_pos.ksh \ functional/userquota/userquota_002_pos.ksh \ functional/userquota/userquota_003_pos.ksh \ functional/userquota/userquota_004_pos.ksh \ functional/userquota/userquota_005_neg.ksh \ functional/userquota/userquota_006_pos.ksh \ functional/userquota/userquota_007_pos.ksh \ functional/userquota/userquota_008_pos.ksh \ functional/userquota/userquota_009_pos.ksh \ functional/userquota/userquota_010_pos.ksh \ functional/userquota/userquota_011_pos.ksh \ functional/userquota/userquota_012_neg.ksh \ functional/userquota/userquota_013_pos.ksh \ functional/userquota/userspace_001_pos.ksh \ functional/userquota/userspace_002_pos.ksh \ functional/userquota/userspace_003_pos.ksh \ functional/userquota/userspace_encrypted.ksh \ functional/userquota/userspace_send_encrypted.ksh \ functional/userquota/userspace_encrypted_13709.ksh \ functional/vdev_zaps/cleanup.ksh \ functional/vdev_zaps/setup.ksh \ functional/vdev_zaps/vdev_zaps_001_pos.ksh \ functional/vdev_zaps/vdev_zaps_002_pos.ksh \ functional/vdev_zaps/vdev_zaps_003_pos.ksh \ functional/vdev_zaps/vdev_zaps_004_pos.ksh \ functional/vdev_zaps/vdev_zaps_005_pos.ksh \ functional/vdev_zaps/vdev_zaps_006_pos.ksh \ functional/vdev_zaps/vdev_zaps_007_pos.ksh \ functional/write_dirs/cleanup.ksh \ functional/write_dirs/setup.ksh \ functional/write_dirs/write_dirs_001_pos.ksh \ functional/write_dirs/write_dirs_002_pos.ksh \ functional/xattr/cleanup.ksh \ functional/xattr/setup.ksh \ functional/xattr/xattr_001_pos.ksh \ functional/xattr/xattr_002_neg.ksh \ functional/xattr/xattr_003_neg.ksh \ functional/xattr/xattr_004_pos.ksh \ functional/xattr/xattr_005_pos.ksh \ functional/xattr/xattr_006_pos.ksh \ functional/xattr/xattr_007_neg.ksh \ functional/xattr/xattr_008_pos.ksh \ functional/xattr/xattr_009_neg.ksh \ functional/xattr/xattr_010_neg.ksh \ functional/xattr/xattr_011_pos.ksh \ functional/xattr/xattr_012_pos.ksh \ functional/xattr/xattr_013_pos.ksh \ functional/xattr/xattr_compat.ksh \ functional/zpool_influxdb/cleanup.ksh \ functional/zpool_influxdb/setup.ksh \ functional/zpool_influxdb/zpool_influxdb.ksh \ functional/zvol/zvol_cli/cleanup.ksh \ functional/zvol/zvol_cli/setup.ksh \ functional/zvol/zvol_cli/zvol_cli_001_pos.ksh \ functional/zvol/zvol_cli/zvol_cli_002_pos.ksh \ functional/zvol/zvol_cli/zvol_cli_003_neg.ksh \ functional/zvol/zvol_ENOSPC/cleanup.ksh \ functional/zvol/zvol_ENOSPC/setup.ksh \ functional/zvol/zvol_ENOSPC/zvol_ENOSPC_001_pos.ksh \ functional/zvol/zvol_misc/cleanup.ksh \ functional/zvol/zvol_misc/setup.ksh \ functional/zvol/zvol_misc/zvol_misc_001_neg.ksh \ functional/zvol/zvol_misc/zvol_misc_002_pos.ksh \ functional/zvol/zvol_misc/zvol_misc_003_neg.ksh \ functional/zvol/zvol_misc/zvol_misc_004_pos.ksh \ functional/zvol/zvol_misc/zvol_misc_005_neg.ksh \ functional/zvol/zvol_misc/zvol_misc_006_pos.ksh \ functional/zvol/zvol_misc/zvol_misc_fua.ksh \ functional/zvol/zvol_misc/zvol_misc_hierarchy.ksh \ functional/zvol/zvol_misc/zvol_misc_rename_inuse.ksh \ functional/zvol/zvol_misc/zvol_misc_snapdev.ksh \ functional/zvol/zvol_misc/zvol_misc_trim.ksh \ functional/zvol/zvol_misc/zvol_misc_volmode.ksh \ functional/zvol/zvol_misc/zvol_misc_zil.ksh \ functional/zvol/zvol_stress/cleanup.ksh \ functional/zvol/zvol_stress/setup.ksh \ functional/zvol/zvol_stress/zvol_stress.ksh \ functional/zvol/zvol_swap/cleanup.ksh \ functional/zvol/zvol_swap/setup.ksh \ functional/zvol/zvol_swap/zvol_swap_001_pos.ksh \ functional/zvol/zvol_swap/zvol_swap_002_pos.ksh \ functional/zvol/zvol_swap/zvol_swap_003_pos.ksh \ functional/zvol/zvol_swap/zvol_swap_004_pos.ksh \ functional/zvol/zvol_swap/zvol_swap_005_pos.ksh \ functional/zvol/zvol_swap/zvol_swap_006_pos.ksh \ functional/idmap_mount/cleanup.ksh \ functional/idmap_mount/setup.ksh \ functional/idmap_mount/idmap_mount_001.ksh \ functional/idmap_mount/idmap_mount_002.ksh \ functional/idmap_mount/idmap_mount_003.ksh \ functional/idmap_mount/idmap_mount_004.ksh \ functional/idmap_mount/idmap_mount_005.ksh diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/alloc_class/alloc_class_014_neg.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/alloc_class/alloc_class_014_neg.ksh new file mode 100755 index 000000000000..1b52014fd2d9 --- /dev/null +++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/alloc_class/alloc_class_014_neg.ksh @@ -0,0 +1,38 @@ +#!/bin/ksh -p + +# +# This file and its contents are supplied under the terms of the +# Common Development and Distribution License ("CDDL"), version 1.0. +# You may only use this file in accordance with the terms of version +# 1.0 of the CDDL. +# +# A full copy of the text of the CDDL should have accompanied this +# source. A copy of the CDDL is also available via the Internet at +# http://www.illumos.org/license/CDDL. +# + +. $STF_SUITE/tests/functional/alloc_class/alloc_class.kshlib + +# +# DESCRIPTION: +# Setting the special_small_blocks property greater than recordsize fails. +# + +verify_runnable "global" + +claim="Setting the special_small_blocks property greater than recordsize fails" + +log_assert $claim +log_onexit cleanup +log_must disk_setup + +for size in 512 4096 32768 131072 524288 1048576 +do + let bigger=$size*2 + log_mustnot zpool create -O recordsize=$size \ + -O special_small_blocks=$bigger \ + $TESTPOOL raidz $ZPOOL_DISKS special mirror \ + $CLASS_DISK0 $CLASS_DISK1 +done + +log_pass $claim diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/alloc_class/alloc_class_015_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/alloc_class/alloc_class_015_pos.ksh new file mode 100755 index 000000000000..49c468af6702 --- /dev/null +++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/alloc_class/alloc_class_015_pos.ksh @@ -0,0 +1,45 @@ +#!/bin/ksh -p + +# +# This file and its contents are supplied under the terms of the +# Common Development and Distribution License ("CDDL"), version 1.0. +# You may only use this file in accordance with the terms of version +# 1.0 of the CDDL. +# +# A full copy of the text of the CDDL should have accompanied this +# source. A copy of the CDDL is also available via the Internet at +# http://www.illumos.org/license/CDDL. +# + +. $STF_SUITE/tests/functional/alloc_class/alloc_class.kshlib + +# +# DESCRIPTION: +# Can set special_small_blocks property less than or equal to recordsize. +# + +verify_runnable "global" + +claim="Can set special_small_blocks property less than or equal to recordsize" + +log_assert $claim +log_onexit cleanup +log_must disk_setup + +for size in 8192 32768 131072 524288 1048576 +do + let smaller=$size/2 + log_must zpool create -O recordsize=$size \ + -O special_small_blocks=$smaller \ + $TESTPOOL raidz $ZPOOL_DISKS special mirror \ + $CLASS_DISK0 $CLASS_DISK1 + log_must zpool destroy -f "$TESTPOOL" + + log_must zpool create -O recordsize=$size \ + -O special_small_blocks=$size \ + $TESTPOOL raidz $ZPOOL_DISKS special mirror \ + $CLASS_DISK0 $CLASS_DISK1 + log_must zpool destroy -f "$TESTPOOL" +done + +log_pass $claim diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_initialize/zpool_initialize_uninit.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_initialize/zpool_initialize_uninit.ksh new file mode 100755 index 000000000000..17f776cfbc20 --- /dev/null +++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_initialize/zpool_initialize_uninit.ksh @@ -0,0 +1,141 @@ +#!/bin/ksh -p +# +# CDDL HEADER START +# +# The contents of this file are subject to the terms of the +# Common Development and Distribution License (the "License"). +# You may not use this file except in compliance with the License. +# +# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE +# or https://opensource.org/licenses/CDDL-1.0. +# See the License for the specific language governing permissions +# and limitations under the License. +# +# When distributing Covered Code, include this CDDL HEADER in each +# file and include the License file at usr/src/OPENSOLARIS.LICENSE. +# If applicable, add the following below this CDDL HEADER, with the +# fields enclosed by brackets "[]" replaced with your own identifying +# information: Portions Copyright [yyyy] [name of copyright owner] +# +# CDDL HEADER END +# + +# +# Copyright (c) 2016 by Delphix. All rights reserved. +# Copyright (C) 2023 Lawrence Livermore National Security, LLC. +# +. $STF_SUITE/include/libtest.shlib +. $STF_SUITE/tests/functional/cli_root/zpool_initialize/zpool_initialize.kshlib + +# +# DESCRIPTION: +# Starting, stopping, uninitializing, and restart an initialize works. +# +# STRATEGY: +# 1. Create a one-disk pool. +# 2. Verify uninitialize succeeds for uninitialized pool. +# 3. Verify pool wide cancel|suspend + uninit +# a. Start initializing and verify that initializing is active. +# b. Verify uninitialize fails when actively initializing. +# c. Cancel or suspend initializing and verify that initializing is not active. +# d. Verify uninitialize succeeds after being cancelled. +# 4. Verify per-disk cancel|suspend + uninit +# + +DISK1="$(echo $DISKS | cut -d' ' -f1)" +DISK2="$(echo $DISKS | cut -d' ' -f2)" +DISK3="$(echo $DISKS | cut -d' ' -f3)" + +function status_check # pool disk1-state disk2-state disk3-state +{ + typeset pool="$1" + typeset disk1_state="$2" + typeset disk2_state="$3" + typeset disk3_state="$4" + + state=$(zpool status -i "$pool" | grep "$DISK1" | grep "$disk1_state") + if [[ -z "$state" ]]; then + log_fail "DISK1 state; expected='$disk1_state' got '$state'" + fi + + state=$(zpool status -i "$pool" | grep "$DISK2" | grep "$disk2_state") + if [[ -z "$state" ]]; then + log_fail "DISK2 state; expected='$disk2_state' got '$state'" + fi + + state=$(zpool status -i "$pool" | grep "$DISK3" | grep "$disk3_state") + if [[ -z "$state" ]]; then + log_fail "DISK3 state; expected='$disk3_state' got '$state'" + fi +} + +function status_check_all # pool disk-state +{ + typeset pool="$1" + typeset disk_state="$2" + + status_check "$pool" "$disk_state" "$disk_state" "$disk_state" +} + +# 1. Create a one-disk pool. +log_must zpool create -f $TESTPOOL $DISK1 $DISK2 $DISK3 +status_check_all $TESTPOOL "uninitialized" + +# 2. Verify uninitialize succeeds for uninitialized pool. +log_must zpool initialize -u $TESTPOOL +status_check_all $TESTPOOL "uninitialized" + +# 3. Verify pool wide cancel + uninit +log_must zpool initialize $TESTPOOL +status_check_all $TESTPOOL "[[:digit:]]* initialized" + +log_mustnot zpool initialize -u $TESTPOOL +status_check_all $TESTPOOL "[[:digit:]]* initialized" + +log_must zpool initialize -c $TESTPOOL +status_check_all $TESTPOOL "uninitialized" + +log_must zpool initialize -u $TESTPOOL +status_check_all $TESTPOOL "uninitialized" + +# 3. Verify pool wide suspend + uninit +log_must zpool initialize $TESTPOOL +status_check_all $TESTPOOL "[[:digit:]]* initialized" + +log_mustnot zpool initialize -u $TESTPOOL +status_check_all $TESTPOOL "[[:digit:]]* initialized" + +log_must zpool initialize -s $TESTPOOL +status_check_all $TESTPOOL "suspended" + +log_must zpool initialize -u $TESTPOOL +status_check_all $TESTPOOL "uninitialized" + +# 4. Verify per-disk cancel|suspend + uninit +log_must zpool initialize $TESTPOOL +status_check_all $TESTPOOL "[[:digit:]]* initialized" + +log_must zpool initialize -c $TESTPOOL $DISK1 +log_must zpool initialize -s $TESTPOOL $DISK2 +log_mustnot zpool initialize -u $TESTPOOL $DISK3 +status_check $TESTPOOL "uninitialized" "suspended" "[[:digit:]]* initialized" + +log_must zpool initialize -u $TESTPOOL $DISK1 +status_check $TESTPOOL "uninitialized" "suspended" "[[:digit:]]* initialized" + +log_must zpool initialize -u $TESTPOOL $DISK2 +status_check $TESTPOOL "uninitialized" "uninitialized" "[[:digit:]]* initialized" + +log_must zpool initialize $TESTPOOL $DISK1 +status_check $TESTPOOL "[[:digit:]]* initialized" "uninitialized" "[[:digit:]]* initialized" + +log_must zpool initialize $TESTPOOL $DISK2 +status_check_all $TESTPOOL "[[:digit:]]* initialized" + +log_must zpool initialize -s $TESTPOOL +status_check_all $TESTPOOL "suspended" + +log_must zpool initialize -u $TESTPOOL $DISK1 $DISK2 $DISK3 +status_check_all $TESTPOOL "uninitialized" + +log_pass "Initialize start + cancel/suspend + uninit + start works" diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_scrub/zpool_error_scrub_001_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_scrub/zpool_error_scrub_001_pos.ksh new file mode 100755 index 000000000000..e414cd1beaad --- /dev/null +++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_scrub/zpool_error_scrub_001_pos.ksh @@ -0,0 +1,79 @@ +#!/bin/ksh -p +# +# CDDL HEADER START +# +# The contents of this file are subject to the terms of the +# Common Development and Distribution License (the "License"). +# You may not use this file except in compliance with the License. +# +# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE +# or http://www.opensolaris.org/os/licensing. +# See the License for the specific language governing permissions +# and limitations under the License. +# +# When distributing Covered Code, include this CDDL HEADER in each +# file and include the License file at usr/src/OPENSOLARIS.LICENSE. +# If applicable, add the following below this CDDL HEADER, with the +# fields enclosed by brackets "[]" replaced with your own identifying +# information: Portions Copyright [yyyy] [name of copyright owner] +# +# CDDL HEADER END +# +# Copyright (c) 2019, Delphix. All rights reserved. +# Copyright (c) 2023, George Amanakis. All rights reserved. +# + +. $STF_SUITE/include/libtest.shlib +. $STF_SUITE/tests/functional/cli_root/zpool_scrub/zpool_scrub.cfg + +# +# DESCRIPTION: +# Verify scrub -e, -p, and -s show the right status. +# +# STRATEGY: +# 1. Create a pool and create a 10MB file in it. +# 2. Start a error scrub (-e) and verify it's doing a scrub. +# 3. Pause error scrub (-p) and verify it's paused. +# 4. Try to pause a paused error scrub (-p) and make sure that fails. +# 5. Resume the paused error scrub and verify again it's doing a scrub. +# 6. Verify zpool scrub -s succeed when the system is error scrubbing. +# + +verify_runnable "global" + +function cleanup +{ + log_must set_tunable32 SCAN_SUSPEND_PROGRESS 0 + log_must zinject -c all + rm -f /$TESTPOOL/10m_file +} + +log_onexit cleanup + +log_assert "Verify scrub -e, -p, and -s show the right status." + +log_must fio --rw=write --name=job --size=10M --filename=/$TESTPOOL/10m_file + +log_must zpool export $TESTPOOL +log_must zpool import $TESTPOOL +log_must zinject -t data -e checksum -f 100 -am /$TESTPOOL/10m_file + +# create some error blocks +dd if=/$TESTPOOL/10m_file bs=1M count=1 || true + +# sync error blocks to disk +log_must sync_pool $TESTPOOL + +log_must set_tunable32 SCAN_SUSPEND_PROGRESS 1 +log_must zpool scrub -e $TESTPOOL +log_must is_pool_error_scrubbing $TESTPOOL true +log_must zpool scrub -p $TESTPOOL +log_must is_pool_error_scrub_paused $TESTPOOL true +log_mustnot zpool scrub -p $TESTPOOL +log_must is_pool_error_scrub_paused $TESTPOOL true +log_must zpool scrub -e $TESTPOOL +log_must is_pool_error_scrubbing $TESTPOOL true +log_must zpool scrub -s $TESTPOOL +log_must is_pool_error_scrub_stopped $TESTPOOL true + +log_pass "Verified scrub -e, -p, and -s show expected status." diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_scrub/zpool_error_scrub_002_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_scrub/zpool_error_scrub_002_pos.ksh new file mode 100755 index 000000000000..daa11c3949c6 --- /dev/null +++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_scrub/zpool_error_scrub_002_pos.ksh @@ -0,0 +1,99 @@ +#!/bin/ksh -p +# +# CDDL HEADER START +# +# The contents of this file are subject to the terms of the +# Common Development and Distribution License (the "License"). +# You may not use this file except in compliance with the License. +# +# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE +# or http://www.opensolaris.org/os/licensing. +# See the License for the specific language governing permissions +# and limitations under the License. +# +# When distributing Covered Code, include this CDDL HEADER in each +# file and include the License file at usr/src/OPENSOLARIS.LICENSE. +# If applicable, add the following below this CDDL HEADER, with the +# fields enclosed by brackets "[]" replaced with your own identifying +# information: Portions Copyright [yyyy] [name of copyright owner] +# +# CDDL HEADER END +# +# Copyright (c) 2019, Delphix. All rights reserved. +# Copyright (c) 2023, George Amanakis. All rights reserved. +# + +. $STF_SUITE/include/libtest.shlib +. $STF_SUITE/tests/functional/cli_root/zpool_scrub/zpool_scrub.cfg + +# +# DESCRIPTION: +# Verify regular scrub and error scrub can't run at the same time. +# +# STRATEGY: +# 1. Create a pool and create a 10MB file in it. +# 2. Start a scrub and verify it's doing a scrub. +# 3. Start a error scrub (-e) and verify it fails. +# 4. Pause scrub (-p) and verify it's paused. +# 5. Start a error scrub (-e) verify it fails again. +# 6. Resume the paused scrub, verify it and cancel it. +# 7. Start a error scrub (-e) and verify it's doing error scrub. +# 8. Start a scrub and verify it fails. +# 9. Cancel error scrub (-e) and verify it is canceled. +# 10. Start scrub, verify it, cancel it and verify it. +# + +verify_runnable "global" + +function cleanup +{ + log_must set_tunable32 SCAN_SUSPEND_PROGRESS 0 + log_must zinject -c all + rm -f /$TESTPOOL/10m_file +} + +log_onexit cleanup + +log_assert "Verify regular scrub and error scrub can't run at the same time." + +log_must fio --rw=write --name=job --size=10M --filename=/$TESTPOOL/10m_file + +log_must zpool export $TESTPOOL +log_must zpool import $TESTPOOL +log_must zinject -t data -e checksum -f 100 -am /$TESTPOOL/10m_file + +# create some error blocks before error scrub is requested. +dd if=/$TESTPOOL/10m_file bs=1M count=1 || true +# sync error blocks to disk +log_must sync_pool $TESTPOOL + +log_must set_tunable32 SCAN_SUSPEND_PROGRESS 1 + +log_must zpool scrub $TESTPOOL +log_must is_pool_scrubbing $TESTPOOL true +log_mustnot zpool scrub -e $TESTPOOL +log_must zpool scrub -p $TESTPOOL +log_must is_pool_scrub_paused $TESTPOOL true +log_mustnot zpool scrub -e $TESTPOOL +log_must zpool scrub $TESTPOOL +log_must is_pool_scrubbing $TESTPOOL true +log_must zpool scrub -s $TESTPOOL +log_must is_pool_scrub_stopped $TESTPOOL true + +# create some error blocks before error scrub is requested. +dd if=/$TESTPOOL/10m_file bs=1M count=1 || true +# sync error blocks to disk +log_must sync_pool $TESTPOOL + +log_must zpool scrub -e $TESTPOOL +log_must is_pool_error_scrubbing $TESTPOOL true +log_mustnot zpool scrub $TESTPOOL +log_must zpool scrub -s $TESTPOOL +log_must is_pool_error_scrub_stopped $TESTPOOL true + +log_must zpool scrub $TESTPOOL +log_must is_pool_scrubbing $TESTPOOL true +log_must zpool scrub -s $TESTPOOL +log_must is_pool_scrub_stopped $TESTPOOL true + +log_pass "Verified regular scrub and error scrub can't run at the same time." diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_scrub/zpool_error_scrub_003_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_scrub/zpool_error_scrub_003_pos.ksh new file mode 100755 index 000000000000..d0066fdbb4a3 --- /dev/null +++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_scrub/zpool_error_scrub_003_pos.ksh @@ -0,0 +1,109 @@ +#!/bin/ksh -p +# +# CDDL HEADER START +# +# The contents of this file are subject to the terms of the +# Common Development and Distribution License (the "License"). +# You may not use this file except in compliance with the License. +# +# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE +# or http://www.opensolaris.org/os/licensing. +# See the License for the specific language governing permissions +# and limitations under the License. +# +# When distributing Covered Code, include this CDDL HEADER in each +# file and include the License file at usr/src/OPENSOLARIS.LICENSE. +# If applicable, add the following below this CDDL HEADER, with the +# fields enclosed by brackets "[]" replaced with your own identifying +# information: Portions Copyright [yyyy] [name of copyright owner] +# +# CDDL HEADER END +# +# Copyright (c) 2019, Delphix. All rights reserved. +# Copyright (c) 2023, George Amanakis. All rights reserved. +# + +. $STF_SUITE/include/libtest.shlib +. $STF_SUITE/tests/functional/cli_root/zpool_scrub/zpool_scrub.cfg + +# +# DESCRIPTION: +# Verify error scrub clears the errorlog, if errors no longer exist. +# +# STRATEGY: +# 1. Create a pool and create file in it. +# 2. Zinject errors and read using dd to log errors to disk. +# 3. Make sure file name is mentioned in the list of error files. +# 4. Start error scrub and wait for it finish. +# 5. Check scrub ran and errors are still reported. +# 6. Clear corruption and error scrub again. +# 7. Check scrub ran and errors are cleared. +# + +verify_runnable "global" + +function cleanup +{ + zinject -c all + rm -f /$TESTPOOL2/$TESTFILE0 + destroy_pool $TESTPOOL2 +} + +log_onexit cleanup + +log_assert "Verify error scrub clears the errorlog, if errors no longer exist." + +truncate -s $MINVDEVSIZE $TESTDIR/vdev_a +log_must zpool create -f -O primarycache=none $TESTPOOL2 $TESTDIR/vdev_a +log_must zfs create $TESTPOOL2/$TESTFS1 +typeset file=/$TESTPOOL2/$TESTFS1/$TESTFILE0 +log_must dd if=/dev/urandom of=$file bs=2M count=10 + +lastfs="$(zfs list -r $TESTPOOL2 | tail -1 | awk '{print $1}')" +for i in {1..3}; do + log_must zfs snap $lastfs@snap$i + log_must zfs clone $lastfs@snap$i $TESTPOOL2/clone$i + lastfs="$(zfs list -r $TESTPOOL2/clone$i | tail -1 | awk '{print $1}')" +done + +log_must zinject -t data -e checksum -f 100 -a $file +dd if=$file of=/dev/null bs=2M count=10 + +# Important: sync error log to disk +log_must sync_pool $TESTPOOL2 + +# Check reported errors +log_must zpool status -v $TESTPOOL2 +log_must eval "zpool status -v $TESTPOOL2 | \ + grep \"Permanent errors have been detected\"" +log_must eval "zpool status -v | grep '$TESTPOOL2/$TESTFS1/$TESTFILE0'" +log_must eval "zpool status -v | grep '$TESTPOOL2/$TESTFS1@snap1:/$TESTFILE0'" +log_must eval "zpool status -v | grep '$TESTPOOL2/clone1/$TESTFILE0'" +log_must eval "zpool status -v | grep '$TESTPOOL2/clone1@snap2:/$TESTFILE0'" +log_must eval "zpool status -v | grep '$TESTPOOL2/clone2/$TESTFILE0'" +log_must eval "zpool status -v | grep '$TESTPOOL2/clone2@snap3:/$TESTFILE0'" +log_must eval "zpool status -v | grep '$TESTPOOL2/clone3/$TESTFILE0'" + +# Check errors are reported if corruption persists +log_must zpool scrub -e -w $TESTPOOL2 +log_must eval "zpool status -v | grep 'error blocks'" +log_must zpool status -v $TESTPOOL2 +log_must eval "zpool status -v $TESTPOOL2 | \ + grep \"Permanent errors have been detected\"" +log_must eval "zpool status -v | grep '$TESTPOOL2/$TESTFS1/$TESTFILE0'" +log_must eval "zpool status -v | grep '$TESTPOOL2/$TESTFS1@snap1:/$TESTFILE0'" +log_must eval "zpool status -v | grep '$TESTPOOL2/clone1/$TESTFILE0'" +log_must eval "zpool status -v | grep '$TESTPOOL2/clone1@snap2:/$TESTFILE0'" +log_must eval "zpool status -v | grep '$TESTPOOL2/clone2/$TESTFILE0'" +log_must eval "zpool status -v | grep '$TESTPOOL2/clone2@snap3:/$TESTFILE0'" +log_must eval "zpool status -v | grep '$TESTPOOL2/clone3/$TESTFILE0'" + +# Check errors are cleared +log_must zinject -c all +log_must zpool scrub -e -w $TESTPOOL2 +log_must zpool status -v $TESTPOOL2 +log_must eval "zpool status -v | grep 'error blocks'" +log_mustnot eval "zpool status -v | grep '$TESTFILE0'" + + +log_pass "Verify error scrub clears the errorlog, if errors no longer exist." diff --git a/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_scrub/zpool_error_scrub_004_pos.ksh b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_scrub/zpool_error_scrub_004_pos.ksh new file mode 100755 index 000000000000..c88b9b0c8d33 --- /dev/null +++ b/sys/contrib/openzfs/tests/zfs-tests/tests/functional/cli_root/zpool_scrub/zpool_error_scrub_004_pos.ksh @@ -0,0 +1,54 @@ +#!/bin/ksh -p +# +# CDDL HEADER START +# +# The contents of this file are subject to the terms of the +# Common Development and Distribution License (the "License"). +# You may not use this file except in compliance with the License. +# +# You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE +# or http://www.opensolaris.org/os/licensing. +# See the License for the specific language governing permissions +# and limitations under the License. +# +# When distributing Covered Code, include this CDDL HEADER in each +# file and include the License file at usr/src/OPENSOLARIS.LICENSE. +# If applicable, add the following below this CDDL HEADER, with the +# fields enclosed by brackets "[]" replaced with your own identifying +# information: Portions Copyright [yyyy] [name of copyright owner] +# +# CDDL HEADER END +# +# Copyright (c) 2023, George Amanakis. All rights reserved. +# + +. $STF_SUITE/include/libtest.shlib +. $STF_SUITE/tests/functional/cli_root/zpool_scrub/zpool_scrub.cfg + +# +# DESCRIPTION: +# Verify error scrub clears the errorlog, if errors no longer exist. +# +# STRATEGY: +# 1. Create a pool with head_errlog disabled. +# 2. Run an error scrub and verify it is not supported. +# + +verify_runnable "global" + +function cleanup +{ + rm -f /$TESTPOOL2/$TESTFILE0 + destroy_pool $TESTPOOL2 +} + +log_onexit cleanup + +log_assert "Verify error scrub cannot run without the head_errlog feature." + +truncate -s $MINVDEVSIZE $TESTDIR/vdev_a +log_must zpool create -f -o feature@head_errlog=disabled $TESTPOOL2 $TESTDIR/vdev_a +log_mustnot zpool scrub -ew $TESTPOOL2 + +log_pass "Verify error scrub cannot run without the head_errlog feature." + diff --git a/sys/modules/zfs/zfs_config.h b/sys/modules/zfs/zfs_config.h index ec45185797b4..d0ad587ed327 100644 --- a/sys/modules/zfs/zfs_config.h +++ b/sys/modules/zfs/zfs_config.h @@ -1,1087 +1,1087 @@ /* * $FreeBSD$ */ /* zfs_config.h. Generated from zfs_config.h.in by configure. */ /* zfs_config.h.in. Generated from configure.ac by autoheader. */ /* Define to 1 if translation of program messages to the user's native language is requested. */ /* #undef ENABLE_NLS */ /* bio_end_io_t wants 1 arg */ /* #undef HAVE_1ARG_BIO_END_IO_T */ /* lookup_bdev() wants 1 arg */ /* #undef HAVE_1ARG_LOOKUP_BDEV */ /* submit_bio() wants 1 arg */ /* #undef HAVE_1ARG_SUBMIT_BIO */ /* bdi_setup_and_register() wants 2 args */ /* #undef HAVE_2ARGS_BDI_SETUP_AND_REGISTER */ /* vfs_getattr wants 2 args */ /* #undef HAVE_2ARGS_VFS_GETATTR */ /* zlib_deflate_workspacesize() wants 2 args */ /* #undef HAVE_2ARGS_ZLIB_DEFLATE_WORKSPACESIZE */ /* bdi_setup_and_register() wants 3 args */ /* #undef HAVE_3ARGS_BDI_SETUP_AND_REGISTER */ /* vfs_getattr wants 3 args */ /* #undef HAVE_3ARGS_VFS_GETATTR */ /* vfs_getattr wants 4 args */ /* #undef HAVE_4ARGS_VFS_GETATTR */ /* kernel has access_ok with 'type' parameter */ /* #undef HAVE_ACCESS_OK_TYPE */ /* posix_acl has refcount_t */ /* #undef HAVE_ACL_REFCOUNT */ /* add_disk() returns int */ /* #undef HAVE_ADD_DISK_RET */ /* Define if host toolchain supports AES */ #define HAVE_AES 1 /* Define if you have [rt] */ #define HAVE_AIO_H 1 #ifdef __amd64__ #ifndef RESCUE /* Define if host toolchain supports AVX */ #define HAVE_AVX 1 #endif /* Define if host toolchain supports AVX2 */ #define HAVE_AVX2 1 /* Define if host toolchain supports AVX512BW */ #define HAVE_AVX512BW 1 /* Define if host toolchain supports AVX512CD */ #define HAVE_AVX512CD 1 /* Define if host toolchain supports AVX512DQ */ #define HAVE_AVX512DQ 1 /* Define if host toolchain supports AVX512ER */ #define HAVE_AVX512ER 1 /* Define if host toolchain supports AVX512F */ #define HAVE_AVX512F 1 /* Define if host toolchain supports AVX512IFMA */ #define HAVE_AVX512IFMA 1 /* Define if host toolchain supports AVX512PF */ #define HAVE_AVX512PF 1 /* Define if host toolchain supports AVX512VBMI */ #define HAVE_AVX512VBMI 1 /* Define if host toolchain supports AVX512VL */ #define HAVE_AVX512VL 1 #endif /* bdevname() is available */ /* #undef HAVE_BDEVNAME */ /* bdev_check_media_change() exists */ /* #undef HAVE_BDEV_CHECK_MEDIA_CHANGE */ /* bdev_*_io_acct() available */ /* #undef HAVE_BDEV_IO_ACCT_63 */ /* bdev_*_io_acct() available */ /* #undef HAVE_BDEV_IO_ACCT_OLD */ /* bdev_kobj() exists */ /* #undef HAVE_BDEV_KOBJ */ /* bdev_max_discard_sectors() is available */ /* #undef HAVE_BDEV_MAX_DISCARD_SECTORS */ /* bdev_max_secure_erase_sectors() is available */ /* #undef HAVE_BDEV_MAX_SECURE_ERASE_SECTORS */ /* block_device_operations->submit_bio() returns void */ /* #undef HAVE_BDEV_SUBMIT_BIO_RETURNS_VOID */ /* bdev_whole() is available */ /* #undef HAVE_BDEV_WHOLE */ /* bio_alloc() takes 4 arguments */ /* #undef HAVE_BIO_ALLOC_4ARG */ /* bio->bi_bdev->bd_disk exists */ /* #undef HAVE_BIO_BDEV_DISK */ /* bio->bi_opf is defined */ /* #undef HAVE_BIO_BI_OPF */ /* bio->bi_status exists */ /* #undef HAVE_BIO_BI_STATUS */ /* bio has bi_iter */ /* #undef HAVE_BIO_BVEC_ITER */ /* bio_*_io_acct() available */ /* #undef HAVE_BIO_IO_ACCT */ /* bio_max_segs() is implemented */ /* #undef HAVE_BIO_MAX_SEGS */ /* bio_set_dev() is available */ /* #undef HAVE_BIO_SET_DEV */ /* bio_set_dev() GPL-only */ /* #undef HAVE_BIO_SET_DEV_GPL_ONLY */ /* bio_set_dev() is a macro */ /* #undef HAVE_BIO_SET_DEV_MACRO */ /* bio_set_op_attrs is available */ /* #undef HAVE_BIO_SET_OP_ATTRS */ /* blkdev_get_by_path() handles ERESTARTSYS */ /* #undef HAVE_BLKDEV_GET_ERESTARTSYS */ /* blkdev_issue_discard() is available */ /* #undef HAVE_BLKDEV_ISSUE_DISCARD */ /* blkdev_issue_secure_erase() is available */ /* #undef HAVE_BLKDEV_ISSUE_SECURE_ERASE */ /* blkdev_reread_part() exists */ /* #undef HAVE_BLKDEV_REREAD_PART */ /* blkg_tryget() is available */ /* #undef HAVE_BLKG_TRYGET */ /* blkg_tryget() GPL-only */ /* #undef HAVE_BLKG_TRYGET_GPL_ONLY */ /* blk_alloc_disk() exists */ /* #undef HAVE_BLK_ALLOC_DISK */ /* blk_alloc_queue() expects request function */ /* #undef HAVE_BLK_ALLOC_QUEUE_REQUEST_FN */ /* blk_alloc_queue_rh() expects request function */ /* #undef HAVE_BLK_ALLOC_QUEUE_REQUEST_FN_RH */ /* blk_cleanup_disk() exists */ /* #undef HAVE_BLK_CLEANUP_DISK */ /* block multiqueue is available */ /* #undef HAVE_BLK_MQ */ /* blk queue backing_dev_info is dynamic */ /* #undef HAVE_BLK_QUEUE_BDI_DYNAMIC */ /* blk_queue_discard() is available */ /* #undef HAVE_BLK_QUEUE_DISCARD */ /* blk_queue_flag_clear() exists */ /* #undef HAVE_BLK_QUEUE_FLAG_CLEAR */ /* blk_queue_flag_set() exists */ /* #undef HAVE_BLK_QUEUE_FLAG_SET */ /* blk_queue_flush() is available */ /* #undef HAVE_BLK_QUEUE_FLUSH */ /* blk_queue_flush() is GPL-only */ /* #undef HAVE_BLK_QUEUE_FLUSH_GPL_ONLY */ /* blk_queue_secdiscard() is available */ /* #undef HAVE_BLK_QUEUE_SECDISCARD */ /* blk_queue_secure_erase() is available */ /* #undef HAVE_BLK_QUEUE_SECURE_ERASE */ /* blk_queue_update_readahead() exists */ /* #undef HAVE_BLK_QUEUE_UPDATE_READAHEAD */ /* blk_queue_write_cache() exists */ /* #undef HAVE_BLK_QUEUE_WRITE_CACHE */ /* blk_queue_write_cache() is GPL-only */ /* #undef HAVE_BLK_QUEUE_WRITE_CACHE_GPL_ONLY */ /* Define if revalidate_disk() in block_device_operations */ /* #undef HAVE_BLOCK_DEVICE_OPERATIONS_REVALIDATE_DISK */ /* Define to 1 if you have the Mac OS X function CFLocaleCopyCurrent in the CoreFoundation framework. */ /* #undef HAVE_CFLOCALECOPYCURRENT */ /* Define to 1 if you have the Mac OS X function CFLocaleCopyPreferredLanguages in the CoreFoundation framework. */ /* #undef HAVE_CFLOCALECOPYPREFERREDLANGUAGES */ /* Define to 1 if you have the Mac OS X function CFPreferencesCopyAppValue in the CoreFoundation framework. */ /* #undef HAVE_CFPREFERENCESCOPYAPPVALUE */ /* check_disk_change() exists */ /* #undef HAVE_CHECK_DISK_CHANGE */ /* clear_inode() is available */ /* #undef HAVE_CLEAR_INODE */ /* dentry uses const struct dentry_operations */ /* #undef HAVE_CONST_DENTRY_OPERATIONS */ /* copy_from_iter() is available */ /* #undef HAVE_COPY_FROM_ITER */ /* copy_to_iter() is available */ /* #undef HAVE_COPY_TO_ITER */ /* cpu_has_feature() is GPL-only */ /* #undef HAVE_CPU_HAS_FEATURE_GPL_ONLY */ /* yes */ /* #undef HAVE_CPU_HOTPLUG */ /* current_time() exists */ /* #undef HAVE_CURRENT_TIME */ /* Define if the GNU dcgettext() function is already present or preinstalled. */ /* #undef HAVE_DCGETTEXT */ /* DECLARE_EVENT_CLASS() is available */ /* #undef HAVE_DECLARE_EVENT_CLASS */ /* dentry aliases are in d_u member */ /* #undef HAVE_DENTRY_D_U_ALIASES */ /* dequeue_signal() takes 4 arguments */ /* #undef HAVE_DEQUEUE_SIGNAL_4ARG */ /* lookup_bdev() wants dev_t arg */ /* #undef HAVE_DEVT_LOOKUP_BDEV */ /* sops->dirty_inode() wants flags */ /* #undef HAVE_DIRTY_INODE_WITH_FLAGS */ /* disk_*_io_acct() available */ /* #undef HAVE_DISK_IO_ACCT */ /* disk_update_readahead() exists */ /* #undef HAVE_DISK_UPDATE_READAHEAD */ /* Define to 1 if you have the header file. */ #define HAVE_DLFCN_H 1 /* d_make_root() is available */ /* #undef HAVE_D_MAKE_ROOT */ /* d_prune_aliases() is available */ /* #undef HAVE_D_PRUNE_ALIASES */ /* dops->d_revalidate() operation takes nameidata */ /* #undef HAVE_D_REVALIDATE_NAMEIDATA */ /* eops->encode_fh() wants child and parent inodes */ /* #undef HAVE_ENCODE_FH_WITH_INODE */ /* sops->evict_inode() exists */ /* #undef HAVE_EVICT_INODE */ /* FALLOC_FL_ZERO_RANGE is defined */ /* #undef HAVE_FALLOC_FL_ZERO_RANGE */ /* fault_in_iov_iter_readable() is available */ /* #undef HAVE_FAULT_IN_IOV_ITER_READABLE */ /* filemap_range_has_page() is available */ /* #undef HAVE_FILEMAP_RANGE_HAS_PAGE */ /* fops->aio_fsync() exists */ /* #undef HAVE_FILE_AIO_FSYNC */ /* file_dentry() is available */ /* #undef HAVE_FILE_DENTRY */ /* fops->fadvise() exists */ /* #undef HAVE_FILE_FADVISE */ /* file_inode() is available */ /* #undef HAVE_FILE_INODE */ /* flush_dcache_page() is GPL-only */ /* #undef HAVE_FLUSH_DCACHE_PAGE_GPL_ONLY */ /* iops->follow_link() cookie */ /* #undef HAVE_FOLLOW_LINK_COOKIE */ /* iops->follow_link() nameidata */ /* #undef HAVE_FOLLOW_LINK_NAMEIDATA */ /* Define if compiler supports -Wformat-overflow */ /* #undef HAVE_FORMAT_OVERFLOW */ /* fops->fsync() with range */ /* #undef HAVE_FSYNC_RANGE */ /* fops->fsync() without dentry */ /* #undef HAVE_FSYNC_WITHOUT_DENTRY */ /* yes */ /* #undef HAVE_GENERIC_FADVISE */ /* generic_fillattr requires struct mnt_idmap* */ /* #undef HAVE_GENERIC_FILLATTR_IDMAP */ /* generic_fillattr requires struct user_namespace* */ /* #undef HAVE_GENERIC_FILLATTR_USERNS */ /* generic_*_io_acct() 3 arg available */ /* #undef HAVE_GENERIC_IO_ACCT_3ARG */ /* generic_*_io_acct() 4 arg available */ /* #undef HAVE_GENERIC_IO_ACCT_4ARG */ /* generic_readlink is global */ /* #undef HAVE_GENERIC_READLINK */ /* generic_setxattr() exists */ /* #undef HAVE_GENERIC_SETXATTR */ /* generic_write_checks() takes kiocb */ /* #undef HAVE_GENERIC_WRITE_CHECKS_KIOCB */ /* Define if the GNU gettext() function is already present or preinstalled. */ /* #undef HAVE_GETTEXT */ /* iops->get_acl() exists */ /* #undef HAVE_GET_ACL */ /* iops->get_acl() takes rcu */ /* #undef HAVE_GET_ACL_RCU */ /* has iops->get_inode_acl() */ /* #undef HAVE_GET_INODE_ACL */ /* iops->get_link() cookie */ /* #undef HAVE_GET_LINK_COOKIE */ /* iops->get_link() delayed */ /* #undef HAVE_GET_LINK_DELAYED */ /* group_info->gid exists */ /* #undef HAVE_GROUP_INFO_GID */ /* has_capability() is available */ /* #undef HAVE_HAS_CAPABILITY */ /* iattr->ia_vfsuid and iattr->ia_vfsgid exist */ /* #undef HAVE_IATTR_VFSID */ /* Define if you have the iconv() function and it works. */ #define HAVE_ICONV 1 /* iops->getattr() takes struct mnt_idmap* */ /* #undef HAVE_IDMAP_IOPS_GETATTR */ /* iops->setattr() takes struct mnt_idmap* */ /* #undef HAVE_IDMAP_IOPS_SETATTR */ /* APIs for idmapped mount are present */ /* #undef HAVE_IDMAP_MNT_API */ /* Define if compiler supports -Wimplicit-fallthrough */ /* #undef HAVE_IMPLICIT_FALLTHROUGH */ /* Define if compiler supports -Winfinite-recursion */ /* #undef HAVE_INFINITE_RECURSION */ /* yes */ /* #undef HAVE_INODE_LOCK_SHARED */ /* inode_owner_or_capable() exists */ /* #undef HAVE_INODE_OWNER_OR_CAPABLE */ /* inode_owner_or_capable() takes mnt_idmap */ /* #undef HAVE_INODE_OWNER_OR_CAPABLE_IDMAP */ /* inode_owner_or_capable() takes user_ns */ /* #undef HAVE_INODE_OWNER_OR_CAPABLE_USERNS */ /* inode_set_flags() exists */ /* #undef HAVE_INODE_SET_FLAGS */ /* inode_set_iversion() exists */ /* #undef HAVE_INODE_SET_IVERSION */ /* inode->i_*time's are timespec64 */ /* #undef HAVE_INODE_TIMESPEC64_TIMES */ /* timestamp_truncate() exists */ /* #undef HAVE_INODE_TIMESTAMP_TRUNCATE */ /* Define to 1 if you have the header file. */ #define HAVE_INTTYPES_H 1 /* in_compat_syscall() is available */ /* #undef HAVE_IN_COMPAT_SYSCALL */ /* iops->create() takes struct mnt_idmap* */ /* #undef HAVE_IOPS_CREATE_IDMAP */ /* iops->create() takes struct user_namespace* */ /* #undef HAVE_IOPS_CREATE_USERNS */ /* iops->mkdir() takes struct mnt_idmap* */ /* #undef HAVE_IOPS_MKDIR_IDMAP */ /* iops->mkdir() takes struct user_namespace* */ /* #undef HAVE_IOPS_MKDIR_USERNS */ /* iops->mknod() takes struct mnt_idmap* */ /* #undef HAVE_IOPS_MKNOD_IDMAP */ /* iops->mknod() takes struct user_namespace* */ /* #undef HAVE_IOPS_MKNOD_USERNS */ /* iops->permission() takes struct mnt_idmap* */ /* #undef HAVE_IOPS_PERMISSION_IDMAP */ /* iops->permission() takes struct user_namespace* */ /* #undef HAVE_IOPS_PERMISSION_USERNS */ /* iops->rename() takes struct mnt_idmap* */ /* #undef HAVE_IOPS_RENAME_IDMAP */ /* iops->rename() takes struct user_namespace* */ /* #undef HAVE_IOPS_RENAME_USERNS */ /* iops->setattr() exists */ /* #undef HAVE_IOPS_SETATTR */ /* iops->symlink() takes struct mnt_idmap* */ /* #undef HAVE_IOPS_SYMLINK_IDMAP */ /* iops->symlink() takes struct user_namespace* */ /* #undef HAVE_IOPS_SYMLINK_USERNS */ /* iov_iter_advance() is available */ /* #undef HAVE_IOV_ITER_ADVANCE */ /* iov_iter_count() is available */ /* #undef HAVE_IOV_ITER_COUNT */ /* iov_iter_fault_in_readable() is available */ /* #undef HAVE_IOV_ITER_FAULT_IN_READABLE */ /* iov_iter_revert() is available */ /* #undef HAVE_IOV_ITER_REVERT */ /* iov_iter_type() is available */ /* #undef HAVE_IOV_ITER_TYPE */ /* iov_iter types are available */ /* #undef HAVE_IOV_ITER_TYPES */ /* yes */ /* #undef HAVE_IO_SCHEDULE_TIMEOUT */ /* Define to 1 if you have the `issetugid' function. */ #define HAVE_ISSETUGID 1 /* kernel has kernel_fpu_* functions */ /* #undef HAVE_KERNEL_FPU */ /* kernel has asm/fpu/api.h */ /* #undef HAVE_KERNEL_FPU_API_HEADER */ /* kernel fpu internal */ /* #undef HAVE_KERNEL_FPU_INTERNAL */ /* kernel has asm/fpu/internal.h */ /* #undef HAVE_KERNEL_FPU_INTERNAL_HEADER */ /* uncached_acl_sentinel() exists */ /* #undef HAVE_KERNEL_GET_ACL_HANDLE_CACHE */ /* Define if compiler supports -Winfinite-recursion */ /* #undef HAVE_KERNEL_INFINITE_RECURSION */ /* kernel does stack verification */ /* #undef HAVE_KERNEL_OBJTOOL */ /* kernel has linux/objtool.h */ /* #undef HAVE_KERNEL_OBJTOOL_HEADER */ /* kernel_read() take loff_t pointer */ /* #undef HAVE_KERNEL_READ_PPOS */ /* timer_list.function gets a timer_list */ /* #undef HAVE_KERNEL_TIMER_FUNCTION_TIMER_LIST */ /* struct timer_list has a flags member */ /* #undef HAVE_KERNEL_TIMER_LIST_FLAGS */ /* timer_setup() is available */ /* #undef HAVE_KERNEL_TIMER_SETUP */ /* kernel_write() take loff_t pointer */ /* #undef HAVE_KERNEL_WRITE_PPOS */ /* kmem_cache_create_usercopy() exists */ /* #undef HAVE_KMEM_CACHE_CREATE_USERCOPY */ /* kstrtoul() exists */ /* #undef HAVE_KSTRTOUL */ /* ktime_get_coarse_real_ts64() exists */ /* #undef HAVE_KTIME_GET_COARSE_REAL_TS64 */ /* ktime_get_raw_ts64() exists */ /* #undef HAVE_KTIME_GET_RAW_TS64 */ /* kvmalloc exists */ /* #undef HAVE_KVMALLOC */ /* Define if you have [aio] */ /* #undef HAVE_LIBAIO */ /* Define if you have [blkid] */ /* #undef HAVE_LIBBLKID */ /* Define if you have [crypto] */ #define HAVE_LIBCRYPTO 1 /* Define if you have [tirpc] */ /* #undef HAVE_LIBTIRPC */ /* Define if you have [udev] */ /* #undef HAVE_LIBUDEV */ /* Define if you have [uuid] */ /* #undef HAVE_LIBUUID */ /* linux/blk-cgroup.h exists */ /* #undef HAVE_LINUX_BLK_CGROUP_HEADER */ /* lseek_execute() is available */ /* #undef HAVE_LSEEK_EXECUTE */ /* makedev() is declared in sys/mkdev.h */ /* #undef HAVE_MAKEDEV_IN_MKDEV */ /* makedev() is declared in sys/sysmacros.h */ /* #undef HAVE_MAKEDEV_IN_SYSMACROS */ /* Noting that make_request_fn() returns blk_qc_t */ /* #undef HAVE_MAKE_REQUEST_FN_RET_QC */ /* Noting that make_request_fn() returns void */ /* #undef HAVE_MAKE_REQUEST_FN_RET_VOID */ /* iops->mkdir() takes umode_t */ /* #undef HAVE_MKDIR_UMODE_T */ /* Define to 1 if you have the `mlockall' function. */ #define HAVE_MLOCKALL 1 /* lookup_bdev() wants mode arg */ /* #undef HAVE_MODE_LOOKUP_BDEV */ /* Define if host toolchain supports MOVBE */ #define HAVE_MOVBE 1 /* new_sync_read()/new_sync_write() are available */ /* #undef HAVE_NEW_SYNC_READ */ /* folio_wait_bit() exists */ /* #undef HAVE_PAGEMAP_FOLIO_WAIT_BIT */ /* part_to_dev() exists */ /* #undef HAVE_PART_TO_DEV */ /* iops->getattr() takes a path */ /* #undef HAVE_PATH_IOPS_GETATTR */ /* Define if host toolchain supports PCLMULQDQ */ #define HAVE_PCLMULQDQ 1 /* percpu_counter_add_batch() is defined */ /* #undef HAVE_PERCPU_COUNTER_ADD_BATCH */ /* percpu_counter_init() wants gfp_t */ /* #undef HAVE_PERCPU_COUNTER_INIT_WITH_GFP */ /* posix_acl_chmod() exists */ /* #undef HAVE_POSIX_ACL_CHMOD */ /* posix_acl_from_xattr() needs user_ns */ /* #undef HAVE_POSIX_ACL_FROM_XATTR_USERNS */ /* posix_acl_release() is available */ /* #undef HAVE_POSIX_ACL_RELEASE */ /* posix_acl_release() is GPL-only */ /* #undef HAVE_POSIX_ACL_RELEASE_GPL_ONLY */ /* posix_acl_valid() wants user namespace */ /* #undef HAVE_POSIX_ACL_VALID_WITH_NS */ /* proc_ops structure exists */ /* #undef HAVE_PROC_OPS_STRUCT */ /* iops->put_link() cookie */ /* #undef HAVE_PUT_LINK_COOKIE */ /* iops->put_link() delayed */ /* #undef HAVE_PUT_LINK_DELAYED */ /* iops->put_link() nameidata */ /* #undef HAVE_PUT_LINK_NAMEIDATA */ /* If available, contains the Python version number currently in use. */ #define HAVE_PYTHON "3.7" /* qat is enabled and existed */ /* #undef HAVE_QAT */ /* register_shrinker is vararg */ /* #undef HAVE_REGISTER_SHRINKER_VARARG */ /* iops->rename2() exists */ /* #undef HAVE_RENAME2 */ /* struct inode_operations_wrapper takes .rename2() */ /* #undef HAVE_RENAME2_OPERATIONS_WRAPPER */ /* iops->rename() wants flags */ /* #undef HAVE_RENAME_WANTS_FLAGS */ /* REQ_DISCARD is defined */ /* #undef HAVE_REQ_DISCARD */ /* REQ_FLUSH is defined */ /* #undef HAVE_REQ_FLUSH */ /* REQ_OP_DISCARD is defined */ /* #undef HAVE_REQ_OP_DISCARD */ /* REQ_OP_FLUSH is defined */ /* #undef HAVE_REQ_OP_FLUSH */ /* REQ_OP_SECURE_ERASE is defined */ /* #undef HAVE_REQ_OP_SECURE_ERASE */ /* REQ_PREFLUSH is defined */ /* #undef HAVE_REQ_PREFLUSH */ /* revalidate_disk() is available */ /* #undef HAVE_REVALIDATE_DISK */ /* revalidate_disk_size() is available */ /* #undef HAVE_REVALIDATE_DISK_SIZE */ /* struct rw_semaphore has member activity */ /* #undef HAVE_RWSEM_ACTIVITY */ /* struct rw_semaphore has atomic_long_t member count */ /* #undef HAVE_RWSEM_ATOMIC_LONG_COUNT */ /* linux/sched/signal.h exists */ /* #undef HAVE_SCHED_SIGNAL_HEADER */ /* Define to 1 if you have the header file. */ #define HAVE_SECURITY_PAM_MODULES_H 1 /* setattr_prepare() accepts mnt_idmap */ /* #undef HAVE_SETATTR_PREPARE_IDMAP */ /* setattr_prepare() is available, doesn't accept user_namespace */ /* #undef HAVE_SETATTR_PREPARE_NO_USERNS */ /* setattr_prepare() accepts user_namespace */ /* #undef HAVE_SETATTR_PREPARE_USERNS */ /* iops->set_acl() exists, takes 3 args */ /* #undef HAVE_SET_ACL */ /* iops->set_acl() takes 4 args, arg1 is struct mnt_idmap * */ /* #undef HAVE_SET_ACL_IDMAP_DENTRY */ /* iops->set_acl() takes 4 args */ /* #undef HAVE_SET_ACL_USERNS */ /* iops->set_acl() takes 4 args, arg2 is struct dentry * */ /* #undef HAVE_SET_ACL_USERNS_DENTRY_ARG2 */ /* set_cached_acl() is usable */ /* #undef HAVE_SET_CACHED_ACL_USABLE */ /* set_special_state() exists */ /* #undef HAVE_SET_SPECIAL_STATE */ /* struct shrink_control exists */ /* #undef HAVE_SHRINK_CONTROL_STRUCT */ /* kernel_siginfo_t exists */ /* #undef HAVE_SIGINFO */ /* signal_stop() exists */ /* #undef HAVE_SIGNAL_STOP */ /* new shrinker callback wants 2 args */ /* #undef HAVE_SINGLE_SHRINKER_CALLBACK */ /* cs->count_objects exists */ /* #undef HAVE_SPLIT_SHRINKER_CALLBACK */ #if defined(__amd64__) || defined(__i386__) /* Define if host toolchain supports SSE */ #define HAVE_SSE 1 /* Define if host toolchain supports SSE2 */ #define HAVE_SSE2 1 /* Define if host toolchain supports SSE3 */ #define HAVE_SSE3 1 /* Define if host toolchain supports SSE4.1 */ #define HAVE_SSE4_1 1 /* Define if host toolchain supports SSE4.2 */ #define HAVE_SSE4_2 1 /* Define if host toolchain supports SSSE3 */ #define HAVE_SSSE3 1 #endif /* STACK_FRAME_NON_STANDARD is defined */ /* #undef HAVE_STACK_FRAME_NON_STANDARD */ /* standalone exists */ /* #undef HAVE_STANDALONE_LINUX_STDARG */ /* Define to 1 if you have the header file. */ #define HAVE_STDINT_H 1 /* Define to 1 if you have the header file. */ #define HAVE_STDIO_H 1 /* Define to 1 if you have the header file. */ #define HAVE_STDLIB_H 1 /* Define to 1 if you have the header file. */ #define HAVE_STRINGS_H 1 /* Define to 1 if you have the header file. */ #define HAVE_STRING_H 1 /* Define to 1 if you have the `strlcat' function. */ #define HAVE_STRLCAT 1 /* Define to 1 if you have the `strlcpy' function. */ #define HAVE_STRLCPY 1 /* submit_bio is member of struct block_device_operations */ /* #undef HAVE_SUBMIT_BIO_IN_BLOCK_DEVICE_OPERATIONS */ /* super_setup_bdi_name() exits */ /* #undef HAVE_SUPER_SETUP_BDI_NAME */ /* super_block->s_user_ns exists */ /* #undef HAVE_SUPER_USER_NS */ /* struct kobj_type has default_groups */ /* #undef HAVE_SYSFS_DEFAULT_GROUPS */ /* Define to 1 if you have the header file. */ #define HAVE_SYS_STAT_H 1 /* Define to 1 if you have the header file. */ #define HAVE_SYS_TYPES_H 1 /* i_op->tmpfile() exists */ /* #undef HAVE_TMPFILE */ /* i_op->tmpfile() uses old dentry signature */ /* #undef HAVE_TMPFILE_DENTRY */ /* i_op->tmpfile() has mnt_idmap */ /* #undef HAVE_TMPFILE_IDMAP */ /* i_op->tmpfile() has userns */ /* #undef HAVE_TMPFILE_USERNS */ /* totalhigh_pages() exists */ /* #undef HAVE_TOTALHIGH_PAGES */ /* kernel has totalram_pages() */ /* #undef HAVE_TOTALRAM_PAGES_FUNC */ /* Define to 1 if you have the `udev_device_get_is_initialized' function. */ /* #undef HAVE_UDEV_DEVICE_GET_IS_INITIALIZED */ /* kernel has __kernel_fpu_* functions */ /* #undef HAVE_UNDERSCORE_KERNEL_FPU */ /* Define to 1 if you have the header file. */ #define HAVE_UNISTD_H 1 /* iops->getattr() takes struct user_namespace* */ /* #undef HAVE_USERNS_IOPS_GETATTR */ /* iops->setattr() takes struct user_namespace* */ /* #undef HAVE_USERNS_IOPS_SETATTR */ /* user_namespace->ns.inum exists */ /* #undef HAVE_USER_NS_COMMON_INUM */ /* iops->getattr() takes a vfsmount */ /* #undef HAVE_VFSMOUNT_IOPS_GETATTR */ /* aops->direct_IO() uses iovec */ /* #undef HAVE_VFS_DIRECT_IO_IOVEC */ /* aops->direct_IO() uses iov_iter without rw */ /* #undef HAVE_VFS_DIRECT_IO_ITER */ /* aops->direct_IO() uses iov_iter with offset */ /* #undef HAVE_VFS_DIRECT_IO_ITER_OFFSET */ /* aops->direct_IO() uses iov_iter with rw and offset */ /* #undef HAVE_VFS_DIRECT_IO_ITER_RW_OFFSET */ /* filemap_dirty_folio exists */ /* #undef HAVE_VFS_FILEMAP_DIRTY_FOLIO */ /* All required iov_iter interfaces are available */ /* #undef HAVE_VFS_IOV_ITER */ /* fops->iterate() is available */ /* #undef HAVE_VFS_ITERATE */ /* fops->iterate_shared() is available */ /* #undef HAVE_VFS_ITERATE_SHARED */ /* fops->readdir() is available */ /* #undef HAVE_VFS_READDIR */ /* address_space_operations->readpages exists */ /* #undef HAVE_VFS_READPAGES */ /* read_folio exists */ /* #undef HAVE_VFS_READ_FOLIO */ /* fops->read/write_iter() are available */ /* #undef HAVE_VFS_RW_ITERATE */ /* __set_page_dirty_nobuffers exists */ /* #undef HAVE_VFS_SET_PAGE_DIRTY_NOBUFFERS */ /* __vmalloc page flags exists */ /* #undef HAVE_VMALLOC_PAGE_KERNEL */ /* yes */ /* #undef HAVE_WAIT_ON_BIT_ACTION */ /* wait_queue_entry_t exists */ /* #undef HAVE_WAIT_QUEUE_ENTRY_T */ /* wq_head->head and wq_entry->entry exist */ /* #undef HAVE_WAIT_QUEUE_HEAD_ENTRY */ /* int (*writepage_t)() takes struct folio* */ /* #undef HAVE_WRITEPAGE_T_FOLIO */ /* xattr_handler->get() wants dentry */ /* #undef HAVE_XATTR_GET_DENTRY */ /* xattr_handler->get() wants both dentry and inode */ /* #undef HAVE_XATTR_GET_DENTRY_INODE */ /* xattr_handler->get() wants dentry and inode and flags */ /* #undef HAVE_XATTR_GET_DENTRY_INODE_FLAGS */ /* xattr_handler->get() wants xattr_handler */ /* #undef HAVE_XATTR_GET_HANDLER */ /* xattr_handler has name */ /* #undef HAVE_XATTR_HANDLER_NAME */ /* xattr_handler->list() wants dentry */ /* #undef HAVE_XATTR_LIST_DENTRY */ /* xattr_handler->list() wants xattr_handler */ /* #undef HAVE_XATTR_LIST_HANDLER */ /* xattr_handler->list() wants simple */ /* #undef HAVE_XATTR_LIST_SIMPLE */ /* xattr_handler->set() wants dentry */ /* #undef HAVE_XATTR_SET_DENTRY */ /* xattr_handler->set() wants both dentry and inode */ /* #undef HAVE_XATTR_SET_DENTRY_INODE */ /* xattr_handler->set() wants xattr_handler */ /* #undef HAVE_XATTR_SET_HANDLER */ /* xattr_handler->set() takes mnt_idmap */ /* #undef HAVE_XATTR_SET_IDMAP */ /* xattr_handler->set() takes user_namespace */ /* #undef HAVE_XATTR_SET_USERNS */ /* Define if host toolchain supports XSAVE */ #define HAVE_XSAVE 1 /* Define if host toolchain supports XSAVEOPT */ #define HAVE_XSAVEOPT 1 /* Define if host toolchain supports XSAVES */ #define HAVE_XSAVES 1 /* ZERO_PAGE() is GPL-only */ /* #undef HAVE_ZERO_PAGE_GPL_ONLY */ /* Define if you have [z] */ #define HAVE_ZLIB 1 /* __posix_acl_chmod() exists */ /* #undef HAVE___POSIX_ACL_CHMOD */ /* kernel exports FPU functions */ /* #undef KERNEL_EXPORTS_X86_FPU */ /* TBD: fetch(3) support */ #if 0 /* whether the chosen libfetch is to be loaded at run-time */ #define LIBFETCH_DYNAMIC 1 /* libfetch is fetch(3) */ #define LIBFETCH_IS_FETCH 1 /* libfetch is libcurl */ #define LIBFETCH_IS_LIBCURL 0 /* soname of chosen libfetch */ #define LIBFETCH_SONAME "libfetch.so.6" #endif /* Define to the sub-directory where libtool stores uninstalled libraries. */ #define LT_OBJDIR ".libs/" /* make_request_fn() return type */ /* #undef MAKE_REQUEST_FN_RET */ /* struct shrink_control has nid */ /* #undef SHRINK_CONTROL_HAS_NID */ /* using complete_and_exit() instead */ /* #undef SPL_KTHREAD_COMPLETE_AND_EXIT */ /* Defined for legacy compatibility. */ #define SPL_META_ALIAS ZFS_META_ALIAS /* Defined for legacy compatibility. */ #define SPL_META_RELEASE ZFS_META_RELEASE /* Defined for legacy compatibility. */ #define SPL_META_VERSION ZFS_META_VERSION /* pde_data() is PDE_DATA() */ /* #undef SPL_PDE_DATA */ /* Define to 1 if all of the C90 standard headers exist (not just the ones required in a freestanding environment). This macro is provided for backward compatibility; new code need not use it. */ #define SYSTEM_FREEBSD 1 /* True if ZFS is to be compiled for a Linux system */ /* #undef SYSTEM_LINUX */ /* Version number of package */ /* #undef ZFS_DEBUG */ /* /dev/zfs minor */ /* #undef ZFS_DEVICE_MINOR */ /* enum node_stat_item contains NR_FILE_PAGES */ /* #undef ZFS_ENUM_NODE_STAT_ITEM_NR_FILE_PAGES */ /* enum node_stat_item contains NR_INACTIVE_ANON */ /* #undef ZFS_ENUM_NODE_STAT_ITEM_NR_INACTIVE_ANON */ /* enum node_stat_item contains NR_INACTIVE_FILE */ /* #undef ZFS_ENUM_NODE_STAT_ITEM_NR_INACTIVE_FILE */ /* enum zone_stat_item contains NR_FILE_PAGES */ /* #undef ZFS_ENUM_ZONE_STAT_ITEM_NR_FILE_PAGES */ /* enum zone_stat_item contains NR_INACTIVE_ANON */ /* #undef ZFS_ENUM_ZONE_STAT_ITEM_NR_INACTIVE_ANON */ /* enum zone_stat_item contains NR_INACTIVE_FILE */ /* #undef ZFS_ENUM_ZONE_STAT_ITEM_NR_INACTIVE_FILE */ /* GENHD_FL_EXT_DEVT flag is not available */ /* #undef ZFS_GENHD_FL_EXT_DEVT */ /* GENHD_FL_NO_PART_SCAN flag is available */ /* #undef ZFS_GENHD_FL_NO_PART */ /* global_node_page_state() exists */ /* #undef ZFS_GLOBAL_NODE_PAGE_STATE */ /* global_zone_page_state() exists */ /* #undef ZFS_GLOBAL_ZONE_PAGE_STATE */ /* Define to 1 if GPL-only symbols can be used */ /* #undef ZFS_IS_GPL_COMPATIBLE */ /* Define the project alias string. */ -#define ZFS_META_ALIAS "zfs-2.1.99-FreeBSD_ge61076683" +#define ZFS_META_ALIAS "zfs-2.1.99-FreeBSD_gad0a55461" /* Define the project author. */ #define ZFS_META_AUTHOR "OpenZFS" /* Define the project release date. */ /* #undef ZFS_META_DATA */ /* Define the maximum compatible kernel version. */ #define ZFS_META_KVER_MAX "6.2" /* Define the minimum compatible kernel version. */ #define ZFS_META_KVER_MIN "3.10" /* Define the project license. */ #define ZFS_META_LICENSE "CDDL" /* Define the libtool library 'age' version information. */ /* #undef ZFS_META_LT_AGE */ /* Define the libtool library 'current' version information. */ /* #undef ZFS_META_LT_CURRENT */ /* Define the libtool library 'revision' version information. */ /* #undef ZFS_META_LT_REVISION */ /* Define the project name. */ #define ZFS_META_NAME "zfs" /* Define the project release. */ -#define ZFS_META_RELEASE "FreeBSD_ge61076683" +#define ZFS_META_RELEASE "FreeBSD_gad0a55461" /* Define the project version. */ #define ZFS_META_VERSION "2.1.99" /* count is located in percpu_ref.data */ /* #undef ZFS_PERCPU_REF_COUNT_IN_DATA */ diff --git a/sys/modules/zfs/zfs_gitrev.h b/sys/modules/zfs/zfs_gitrev.h index 29954b8d4719..5c7ebbec1dc5 100644 --- a/sys/modules/zfs/zfs_gitrev.h +++ b/sys/modules/zfs/zfs_gitrev.h @@ -1 +1 @@ -#define ZFS_META_GITREV "zfs-2.1.99-1944-ge61076683" +#define ZFS_META_GITREV "zfs-2.1.99-1955-gad0a55461"