diff --git a/include/sys/arc.h b/include/sys/arc.h index db7a64aa2e22..5780554b5c22 100644 --- a/include/sys/arc.h +++ b/include/sys/arc.h @@ -1,230 +1,234 @@ /* * 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) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2012, 2014 by Delphix. All rights reserved. * Copyright (c) 2013 by Saso Kiselkov. All rights reserved. */ #ifndef _SYS_ARC_H #define _SYS_ARC_H #include #ifdef __cplusplus extern "C" { #endif #include #include #include #include /* * Used by arc_flush() to inform arc_evict_state() that it should evict * all available buffers from the arc state being passed in. */ #define ARC_EVICT_ALL -1ULL typedef struct arc_buf_hdr arc_buf_hdr_t; typedef struct arc_buf arc_buf_t; typedef struct arc_prune arc_prune_t; typedef void arc_done_func_t(zio_t *zio, arc_buf_t *buf, void *private); typedef void arc_prune_func_t(int64_t bytes, void *private); typedef int arc_evict_func_t(void *private); /* Shared module parameters */ extern int zfs_arc_average_blocksize; /* generic arc_done_func_t's which you can use */ arc_done_func_t arc_bcopy_func; arc_done_func_t arc_getbuf_func; /* generic arc_prune_func_t wrapper for callbacks */ struct arc_prune { arc_prune_func_t *p_pfunc; void *p_private; uint64_t p_adjust; list_node_t p_node; refcount_t p_refcnt; }; typedef enum arc_strategy { ARC_STRATEGY_META_ONLY = 0, /* Evict only meta data buffers */ ARC_STRATEGY_META_BALANCED = 1, /* Evict data buffers if needed */ } arc_strategy_t; typedef enum arc_flags { /* * Public flags that can be passed into the ARC by external consumers. */ ARC_FLAG_NONE = 1 << 0, /* No flags set */ ARC_FLAG_WAIT = 1 << 1, /* perform sync I/O */ ARC_FLAG_NOWAIT = 1 << 2, /* perform async I/O */ ARC_FLAG_PREFETCH = 1 << 3, /* I/O is a prefetch */ ARC_FLAG_CACHED = 1 << 4, /* I/O was in cache */ ARC_FLAG_L2CACHE = 1 << 5, /* cache in L2ARC */ ARC_FLAG_L2COMPRESS = 1 << 6, /* compress in L2ARC */ + ARC_FLAG_PREDICTIVE_PREFETCH = 1 << 7, /* I/O from zfetch */ /* * Private ARC flags. These flags are private ARC only flags that * will show up in b_flags in the arc_hdr_buf_t. These flags should * only be set by ARC code. */ - ARC_FLAG_IN_HASH_TABLE = 1 << 7, /* buffer is hashed */ - ARC_FLAG_IO_IN_PROGRESS = 1 << 8, /* I/O in progress */ - ARC_FLAG_IO_ERROR = 1 << 9, /* I/O failed for buf */ - ARC_FLAG_FREED_IN_READ = 1 << 10, /* freed during read */ - ARC_FLAG_BUF_AVAILABLE = 1 << 11, /* block not in use */ - ARC_FLAG_INDIRECT = 1 << 12, /* indirect block */ - ARC_FLAG_L2_WRITING = 1 << 13, /* write in progress */ - ARC_FLAG_L2_EVICTED = 1 << 14, /* evicted during I/O */ - ARC_FLAG_L2_WRITE_HEAD = 1 << 15, /* head of write list */ + ARC_FLAG_IN_HASH_TABLE = 1 << 8, /* buffer is hashed */ + ARC_FLAG_IO_IN_PROGRESS = 1 << 9, /* I/O in progress */ + ARC_FLAG_IO_ERROR = 1 << 10, /* I/O failed for buf */ + ARC_FLAG_FREED_IN_READ = 1 << 11, /* freed during read */ + ARC_FLAG_BUF_AVAILABLE = 1 << 12, /* block not in use */ + ARC_FLAG_INDIRECT = 1 << 13, /* indirect block */ + /* Indicates that block was read with ASYNC priority. */ + ARC_FLAG_PRIO_ASYNC_READ = 1 << 14, + ARC_FLAG_L2_WRITING = 1 << 15, /* write in progress */ + ARC_FLAG_L2_EVICTED = 1 << 16, /* evicted during I/O */ + ARC_FLAG_L2_WRITE_HEAD = 1 << 17, /* head of write list */ /* indicates that the buffer contains metadata (otherwise, data) */ - ARC_FLAG_BUFC_METADATA = 1 << 16, + ARC_FLAG_BUFC_METADATA = 1 << 18, /* Flags specifying whether optional hdr struct fields are defined */ - ARC_FLAG_HAS_L1HDR = 1 << 17, - ARC_FLAG_HAS_L2HDR = 1 << 18, + ARC_FLAG_HAS_L1HDR = 1 << 19, + ARC_FLAG_HAS_L2HDR = 1 << 20, + } arc_flags_t; struct arc_buf { arc_buf_hdr_t *b_hdr; arc_buf_t *b_next; kmutex_t b_evict_lock; void *b_data; arc_evict_func_t *b_efunc; void *b_private; }; typedef enum arc_buf_contents { ARC_BUFC_DATA, /* buffer contains data */ ARC_BUFC_METADATA, /* buffer contains metadata */ ARC_BUFC_NUMTYPES } arc_buf_contents_t; /* * The following breakdows of arc_size exist for kstat only. */ typedef enum arc_space_type { ARC_SPACE_DATA, ARC_SPACE_META, ARC_SPACE_HDRS, ARC_SPACE_L2HDRS, ARC_SPACE_OTHER, ARC_SPACE_NUMTYPES } arc_space_type_t; typedef enum arc_state_type { ARC_STATE_ANON, ARC_STATE_MRU, ARC_STATE_MRU_GHOST, ARC_STATE_MFU, ARC_STATE_MFU_GHOST, ARC_STATE_L2C_ONLY, ARC_STATE_NUMTYPES } arc_state_type_t; typedef struct arc_buf_info { arc_state_type_t abi_state_type; arc_buf_contents_t abi_state_contents; uint32_t abi_flags; uint32_t abi_datacnt; uint64_t abi_size; uint64_t abi_spa; uint64_t abi_access; uint32_t abi_mru_hits; uint32_t abi_mru_ghost_hits; uint32_t abi_mfu_hits; uint32_t abi_mfu_ghost_hits; uint32_t abi_l2arc_hits; uint32_t abi_holds; uint64_t abi_l2arc_dattr; uint64_t abi_l2arc_asize; enum zio_compress abi_l2arc_compress; } arc_buf_info_t; void arc_space_consume(uint64_t space, arc_space_type_t type); void arc_space_return(uint64_t space, arc_space_type_t type); arc_buf_t *arc_buf_alloc(spa_t *spa, uint64_t size, void *tag, arc_buf_contents_t type); arc_buf_t *arc_loan_buf(spa_t *spa, uint64_t size); void arc_return_buf(arc_buf_t *buf, void *tag); void arc_loan_inuse_buf(arc_buf_t *buf, void *tag); void arc_buf_add_ref(arc_buf_t *buf, void *tag); boolean_t arc_buf_remove_ref(arc_buf_t *buf, void *tag); void arc_buf_info(arc_buf_t *buf, arc_buf_info_t *abi, int state_index); uint64_t arc_buf_size(arc_buf_t *buf); void arc_release(arc_buf_t *buf, void *tag); int arc_released(arc_buf_t *buf); void arc_buf_sigsegv(int sig, siginfo_t *si, void *unused); void arc_buf_freeze(arc_buf_t *buf); void arc_buf_thaw(arc_buf_t *buf); boolean_t arc_buf_eviction_needed(arc_buf_t *buf); #ifdef ZFS_DEBUG int arc_referenced(arc_buf_t *buf); #endif int arc_read(zio_t *pio, spa_t *spa, const blkptr_t *bp, arc_done_func_t *done, void *private, zio_priority_t priority, int flags, arc_flags_t *arc_flags, const zbookmark_phys_t *zb); zio_t *arc_write(zio_t *pio, spa_t *spa, uint64_t txg, blkptr_t *bp, arc_buf_t *buf, boolean_t l2arc, boolean_t l2arc_compress, const zio_prop_t *zp, arc_done_func_t *ready, arc_done_func_t *physdone, arc_done_func_t *done, void *private, zio_priority_t priority, int zio_flags, const zbookmark_phys_t *zb); arc_prune_t *arc_add_prune_callback(arc_prune_func_t *func, void *private); void arc_remove_prune_callback(arc_prune_t *p); void arc_freed(spa_t *spa, const blkptr_t *bp); void arc_set_callback(arc_buf_t *buf, arc_evict_func_t *func, void *private); boolean_t arc_clear_callback(arc_buf_t *buf); void arc_flush(spa_t *spa, boolean_t retry); void arc_tempreserve_clear(uint64_t reserve); int arc_tempreserve_space(uint64_t reserve, uint64_t txg); void arc_init(void); void arc_fini(void); /* * Level 2 ARC */ void l2arc_add_vdev(spa_t *spa, vdev_t *vd); void l2arc_remove_vdev(vdev_t *vd); boolean_t l2arc_vdev_present(vdev_t *vd); void l2arc_init(void); void l2arc_fini(void); void l2arc_start(void); void l2arc_stop(void); #ifndef _KERNEL extern boolean_t arc_watch; #endif #ifdef __cplusplus } #endif #endif /* _SYS_ARC_H */ diff --git a/include/sys/dmu.h b/include/sys/dmu.h index 75d85951a334..0a5b2809b05a 100644 --- a/include/sys/dmu.h +++ b/include/sys/dmu.h @@ -1,932 +1,933 @@ /* * 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) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2011, 2014 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. */ /* 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 #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; typedef struct objset objset_t; typedef struct dmu_tx dmu_tx_t; typedef struct dsl_dir dsl_dir_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_BYTESWAP_MASK 0x3f /* * 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). */ #define DMU_OT(byteswap, metadata) \ (DMU_OT_NEWTYPE | \ ((metadata) ? DMU_OT_METADATA : 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(ot) (((ot) & DMU_OT_NEWTYPE) ? \ ((ot) & DMU_OT_METADATA) : \ dmu_ot[(int)(ot)].ot_metadata) /* * 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[(int)(ot)].ot_byteswap) 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_METDATA() 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), DMU_OTN_UINT8_METADATA = DMU_OT(DMU_BSWAP_UINT8, B_TRUE), DMU_OTN_UINT16_DATA = DMU_OT(DMU_BSWAP_UINT16, B_FALSE), DMU_OTN_UINT16_METADATA = DMU_OT(DMU_BSWAP_UINT16, B_TRUE), DMU_OTN_UINT32_DATA = DMU_OT(DMU_BSWAP_UINT32, B_FALSE), DMU_OTN_UINT32_METADATA = DMU_OT(DMU_BSWAP_UINT32, B_TRUE), DMU_OTN_UINT64_DATA = DMU_OT(DMU_BSWAP_UINT64, B_FALSE), DMU_OTN_UINT64_METADATA = DMU_OT(DMU_BSWAP_UINT64, B_TRUE), DMU_OTN_ZAP_DATA = DMU_OT(DMU_BSWAP_ZAP, B_FALSE), DMU_OTN_ZAP_METADATA = DMU_OT(DMU_BSWAP_ZAP, B_TRUE), } dmu_object_type_t; typedef enum txg_how { TXG_WAIT = 1, TXG_NOWAIT, TXG_WAITED, } txg_how_t; 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) /* * 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. */ int dmu_objset_hold(const char *name, void *tag, objset_t **osp); int dmu_objset_own(const char *name, dmu_objset_type_t type, boolean_t readonly, void *tag, objset_t **osp); void dmu_objset_rele(objset_t *os, void *tag); void dmu_objset_disown(objset_t *os, 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, void (*func)(objset_t *os, void *arg, cred_t *cr, dmu_tx_t *tx), 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_snapshot_tmp(const char *, const char *, int); int dmu_objset_find(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_FREE_BPOBJ "free_bpobj" #define DMU_POOL_BPTREE_OBJ "bptree_obj" #define DMU_POOL_EMPTY_BPOBJ "empty_bpobj" /* * 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); 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_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); /* * 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 data blocksize for an object. * * The object cannot have any blocks allcated 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); /* * 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); /* * 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, struct dnode *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_read() to * read db_data, 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, void *tag, dmu_buf_t **); 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, void *tag, dmu_buf_t **dbp); int dmu_spill_hold_by_dnode(struct dnode *dn, uint32_t flags, void *tag, dmu_buf_t **dbp); int dmu_spill_hold_existing(dmu_buf_t *bonus, 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, void *tag, dmu_buf_t **, int 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, 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, void *tag); void dmu_buf_rele(dmu_buf_t *db, void *tag); uint64_t dmu_buf_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, int read, void *tag, int *numbufsp, dmu_buf_t ***dbpp); + uint64_t length, boolean_t read, void *tag, + int *numbufsp, dmu_buf_t ***dbpp); void dmu_buf_rele_array(dmu_buf_t **, int numbufs, 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 pointer. */ dmu_buf_evict_func_t *dbu_evict_func; #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. */ #ifdef __lint /* Very ugly, but it beats issuing suppression directives in many Makefiles. */ extern void dmu_buf_init_user(dmu_buf_user_t *dbu, dmu_buf_evict_func_t *evict_func, dmu_buf_t **clear_on_evict_dbufp); #else /* __lint */ static inline void dmu_buf_init_user(dmu_buf_user_t *dbu, dmu_buf_evict_func_t *evict_func, dmu_buf_t **clear_on_evict_dbufp) { ASSERT(dbu->dbu_evict_func == NULL); ASSERT(evict_func != NULL); dbu->dbu_evict_func = evict_func; taskq_init_ent(&dbu->dbu_tqent); #ifdef ZFS_DEBUG dbu->dbu_clear_on_evict_dbufp = clear_on_evict_dbufp; #endif } #endif /* __lint */ /* * 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); /* 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); /* * Tells if the given dbuf is freeable. */ boolean_t dmu_buf_freeable(dmu_buf_t *); /* * 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_free(dmu_tx_t *tx, uint64_t object, 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_bonus(dmu_tx_t *tx, uint64_t object); 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, enum txg_how txg_how); void dmu_tx_wait(dmu_tx_t *tx); void dmu_tx_commit(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. */ 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); /* * 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 */ int dmu_read(objset_t *os, uint64_t object, 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_prealloc(objset_t *os, uint64_t object, uint64_t offset, uint64_t size, dmu_tx_t *tx); #ifdef _KERNEL #include int dmu_read_uio(objset_t *os, uint64_t object, struct uio *uio, uint64_t size); int dmu_read_uio_dbuf(dmu_buf_t *zdb, struct uio *uio, uint64_t size); int dmu_write_uio(objset_t *os, uint64_t object, struct uio *uio, uint64_t size, dmu_tx_t *tx); int dmu_write_uio_dbuf(dmu_buf_t *zdb, struct uio *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); void dmu_assign_arcbuf(dmu_buf_t *handle, uint64_t offset, struct arc_buf *buf, dmu_tx_t *tx); int dmu_xuio_init(struct xuio *uio, int niov); void dmu_xuio_fini(struct xuio *uio); int dmu_xuio_add(struct xuio *uio, struct arc_buf *abuf, offset_t off, size_t n); int dmu_xuio_cnt(struct xuio *uio); struct arc_buf *dmu_xuio_arcbuf(struct xuio *uio, int i); void dmu_xuio_clear(struct xuio *uio, int i); void xuio_stat_wbuf_copied(void); void xuio_stat_wbuf_nocopy(void); extern int zfs_prefetch_disable; extern int 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_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; char *ot_name; } dmu_object_type_info_t; typedef const struct dmu_object_byteswap_info { arc_byteswap_func_t ob_func; char *ob_name; } dmu_object_byteswap_info_t; extern const dmu_object_type_info_t dmu_ot[DMU_OT_NUMTYPES]; extern const 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(struct dnode *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. This is specifically optimized for zfs_getattr(). */ void dmu_object_size_from_db(dmu_buf_t *db, uint32_t *blksize, u_longlong_t *nblk512); 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; char dds_origin[MAXNAMELEN]; } 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 */ timestruc_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 zfs_sync_type_t dmu_objset_syncprop(objset_t *os); extern zfs_logbias_op_t dmu_objset_logbias(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, 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 int objset_used_cb_t(dmu_object_type_t bonustype, void *bonus, uint64_t *userp, uint64_t *groupp); extern void dmu_objset_register_type(dmu_objset_type_t ost, objset_used_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 zilog *zgd_zilog; struct blkptr *zgd_bp; dmu_buf_t *zgd_db; struct rl *zgd_rl; 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); /* * 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, struct vnode *vp, offset_t *offp); /* CRC64 table */ #define ZFS_CRC64_POLY 0xC96C5795D7870F42ULL /* ECMA-182, reflected form */ extern uint64_t zfs_crc64_table[256]; extern int zfs_mdcomp_disable; #ifdef __cplusplus } #endif #endif /* _SYS_DMU_H */ diff --git a/include/sys/dmu_zfetch.h b/include/sys/dmu_zfetch.h index 38ed1d872129..df33f182bd4f 100644 --- a/include/sys/dmu_zfetch.h +++ b/include/sys/dmu_zfetch.h @@ -1,76 +1,69 @@ /* * 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 2009 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ -#ifndef _DFETCH_H -#define _DFETCH_H +/* + * Copyright (c) 2014 by Delphix. All rights reserved. + */ + +#ifndef _DMU_ZFETCH_H +#define _DMU_ZFETCH_H #include #ifdef __cplusplus extern "C" { #endif extern unsigned long zfetch_array_rd_sz; struct dnode; /* so we can reference dnode */ -typedef enum zfetch_dirn { - ZFETCH_FORWARD = 1, /* prefetch increasing block numbers */ - ZFETCH_BACKWARD = -1 /* prefetch decreasing block numbers */ -} zfetch_dirn_t; - typedef struct zstream { - uint64_t zst_offset; /* offset of starting block in range */ - uint64_t zst_len; /* length of range, in blocks */ - zfetch_dirn_t zst_direction; /* direction of prefetch */ - uint64_t zst_stride; /* length of stride, in blocks */ - uint64_t zst_ph_offset; /* prefetch offset, in blocks */ - uint64_t zst_cap; /* prefetch limit (cap), in blocks */ - kmutex_t zst_lock; /* protects stream */ - clock_t zst_last; /* lbolt of last prefetch */ - list_node_t zst_node; /* next zstream here */ + uint64_t zs_blkid; /* expect next access at this blkid */ + uint64_t zs_pf_blkid; /* next block to prefetch */ + kmutex_t zs_lock; /* protects stream */ + hrtime_t zs_atime; /* time last prefetch issued */ + list_node_t zs_node; /* link for zf_stream */ } zstream_t; typedef struct zfetch { krwlock_t zf_rwlock; /* protects zfetch structure */ - list_t zf_stream; /* AVL tree of zstream_t's */ + list_t zf_stream; /* list of zstream_t's */ struct dnode *zf_dnode; /* dnode that owns this zfetch */ - uint32_t zf_stream_cnt; /* # of active streams */ - uint64_t zf_alloc_fail; /* # of failed attempts to alloc strm */ } zfetch_t; void zfetch_init(void); void zfetch_fini(void); void dmu_zfetch_init(zfetch_t *, struct dnode *); -void dmu_zfetch_rele(zfetch_t *); -void dmu_zfetch(zfetch_t *, uint64_t, uint64_t, int); +void dmu_zfetch_fini(zfetch_t *); +void dmu_zfetch(zfetch_t *, uint64_t, uint64_t); #ifdef __cplusplus } #endif -#endif /* _DFETCH_H */ +#endif /* _DMU_ZFETCH_H */ diff --git a/include/sys/trace_arc.h b/include/sys/trace_arc.h index 31c3cdcb9b21..55dbdf19b055 100644 --- a/include/sys/trace_arc.h +++ b/include/sys/trace_arc.h @@ -1,422 +1,424 @@ /* * 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 */ #include #if defined(_KERNEL) && defined(HAVE_DECLARE_EVENT_CLASS) #undef TRACE_SYSTEM #define TRACE_SYSTEM zfs #undef TRACE_SYSTEM_VAR #define TRACE_SYSTEM_VAR zfs_arc #if !defined(_TRACE_ARC_H) || defined(TRACE_HEADER_MULTI_READ) #define _TRACE_ARC_H #include #include /* * Generic support for one argument tracepoints of the form: * * DTRACE_PROBE1(..., * arc_buf_hdr_t *, ...); */ DECLARE_EVENT_CLASS(zfs_arc_buf_hdr_class, TP_PROTO(arc_buf_hdr_t *ab), TP_ARGS(ab), TP_STRUCT__entry( __array(uint64_t, hdr_dva_word, 2) __field(uint64_t, hdr_birth) __field(uint32_t, hdr_flags) __field(uint32_t, hdr_datacnt) __field(arc_buf_contents_t, hdr_type) __field(uint64_t, hdr_size) __field(uint64_t, hdr_spa) __field(arc_state_type_t, hdr_state_type) __field(clock_t, hdr_access) __field(uint32_t, hdr_mru_hits) __field(uint32_t, hdr_mru_ghost_hits) __field(uint32_t, hdr_mfu_hits) __field(uint32_t, hdr_mfu_ghost_hits) __field(uint32_t, hdr_l2_hits) __field(int64_t, hdr_refcount) ), TP_fast_assign( __entry->hdr_dva_word[0] = ab->b_dva.dva_word[0]; __entry->hdr_dva_word[1] = ab->b_dva.dva_word[1]; __entry->hdr_birth = ab->b_birth; __entry->hdr_flags = ab->b_flags; __entry->hdr_datacnt = ab->b_l1hdr.b_datacnt; __entry->hdr_size = ab->b_size; __entry->hdr_spa = ab->b_spa; __entry->hdr_state_type = ab->b_l1hdr.b_state->arcs_state; __entry->hdr_access = ab->b_l1hdr.b_arc_access; __entry->hdr_mru_hits = ab->b_l1hdr.b_mru_hits; __entry->hdr_mru_ghost_hits = ab->b_l1hdr.b_mru_ghost_hits; __entry->hdr_mfu_hits = ab->b_l1hdr.b_mfu_hits; __entry->hdr_mfu_ghost_hits = ab->b_l1hdr.b_mfu_ghost_hits; __entry->hdr_l2_hits = ab->b_l1hdr.b_l2_hits; __entry->hdr_refcount = ab->b_l1hdr.b_refcnt.rc_count; ), TP_printk("hdr { dva 0x%llx:0x%llx birth %llu " "flags 0x%x datacnt %u type %u size %llu spa %llu " "state_type %u access %lu mru_hits %u mru_ghost_hits %u " "mfu_hits %u mfu_ghost_hits %u l2_hits %u refcount %lli }", __entry->hdr_dva_word[0], __entry->hdr_dva_word[1], __entry->hdr_birth, __entry->hdr_flags, __entry->hdr_datacnt, __entry->hdr_type, __entry->hdr_size, __entry->hdr_spa, __entry->hdr_state_type, __entry->hdr_access, __entry->hdr_mru_hits, __entry->hdr_mru_ghost_hits, __entry->hdr_mfu_hits, __entry->hdr_mfu_ghost_hits, __entry->hdr_l2_hits, __entry->hdr_refcount) ); #define DEFINE_ARC_BUF_HDR_EVENT(name) \ DEFINE_EVENT(zfs_arc_buf_hdr_class, name, \ TP_PROTO(arc_buf_hdr_t *ab), \ TP_ARGS(ab)) DEFINE_ARC_BUF_HDR_EVENT(zfs_arc__hit); DEFINE_ARC_BUF_HDR_EVENT(zfs_arc__evict); DEFINE_ARC_BUF_HDR_EVENT(zfs_arc__delete); DEFINE_ARC_BUF_HDR_EVENT(zfs_new_state__mru); DEFINE_ARC_BUF_HDR_EVENT(zfs_new_state__mfu); +DEFINE_ARC_BUF_HDR_EVENT(zfs_arc__sync__wait__for__async); +DEFINE_ARC_BUF_HDR_EVENT(zfs_arc__demand__hit__predictive__prefetch); DEFINE_ARC_BUF_HDR_EVENT(zfs_l2arc__hit); DEFINE_ARC_BUF_HDR_EVENT(zfs_l2arc__miss); /* * Generic support for two argument tracepoints of the form: * * DTRACE_PROBE2(..., * vdev_t *, ..., * zio_t *, ...); */ #define ZIO_TP_STRUCT_ENTRY \ __field(zio_type_t, zio_type) \ __field(int, zio_cmd) \ __field(zio_priority_t, zio_priority) \ __field(uint64_t, zio_size) \ __field(uint64_t, zio_orig_size) \ __field(uint64_t, zio_offset) \ __field(hrtime_t, zio_timestamp) \ __field(hrtime_t, zio_delta) \ __field(uint64_t, zio_delay) \ __field(enum zio_flag, zio_flags) \ __field(enum zio_stage, zio_stage) \ __field(enum zio_stage, zio_pipeline) \ __field(enum zio_flag, zio_orig_flags) \ __field(enum zio_stage, zio_orig_stage) \ __field(enum zio_stage, zio_orig_pipeline) \ __field(uint8_t, zio_reexecute) \ __field(uint64_t, zio_txg) \ __field(int, zio_error) \ __field(uint64_t, zio_ena) \ \ __field(enum zio_checksum, zp_checksum) \ __field(enum zio_compress, zp_compress) \ __field(dmu_object_type_t, zp_type) \ __field(uint8_t, zp_level) \ __field(uint8_t, zp_copies) \ __field(boolean_t, zp_dedup) \ __field(boolean_t, zp_dedup_verify) \ __field(boolean_t, zp_nopwrite) #define ZIO_TP_FAST_ASSIGN \ __entry->zio_type = zio->io_type; \ __entry->zio_cmd = zio->io_cmd; \ __entry->zio_priority = zio->io_priority; \ __entry->zio_size = zio->io_size; \ __entry->zio_orig_size = zio->io_orig_size; \ __entry->zio_offset = zio->io_offset; \ __entry->zio_timestamp = zio->io_timestamp; \ __entry->zio_delta = zio->io_delta; \ __entry->zio_delay = zio->io_delay; \ __entry->zio_flags = zio->io_flags; \ __entry->zio_stage = zio->io_stage; \ __entry->zio_pipeline = zio->io_pipeline; \ __entry->zio_orig_flags = zio->io_orig_flags; \ __entry->zio_orig_stage = zio->io_orig_stage; \ __entry->zio_orig_pipeline = zio->io_orig_pipeline; \ __entry->zio_reexecute = zio->io_reexecute; \ __entry->zio_txg = zio->io_txg; \ __entry->zio_error = zio->io_error; \ __entry->zio_ena = zio->io_ena; \ \ __entry->zp_checksum = zio->io_prop.zp_checksum; \ __entry->zp_compress = zio->io_prop.zp_compress; \ __entry->zp_type = zio->io_prop.zp_type; \ __entry->zp_level = zio->io_prop.zp_level; \ __entry->zp_copies = zio->io_prop.zp_copies; \ __entry->zp_dedup = zio->io_prop.zp_dedup; \ __entry->zp_nopwrite = zio->io_prop.zp_nopwrite; \ __entry->zp_dedup_verify = zio->io_prop.zp_dedup_verify; #define ZIO_TP_PRINTK_FMT \ "zio { type %u cmd %i prio %u size %llu orig_size %llu " \ "offset %llu timestamp %llu delta %llu delay %llu " \ "flags 0x%x stage 0x%x pipeline 0x%x orig_flags 0x%x " \ "orig_stage 0x%x orig_pipeline 0x%x reexecute %u " \ "txg %llu error %d ena %llu prop { checksum %u compress %u " \ "type %u level %u copies %u dedup %u dedup_verify %u nopwrite %u } }" #define ZIO_TP_PRINTK_ARGS \ __entry->zio_type, __entry->zio_cmd, __entry->zio_priority, \ __entry->zio_size, __entry->zio_orig_size, __entry->zio_offset, \ __entry->zio_timestamp, __entry->zio_delta, __entry->zio_delay, \ __entry->zio_flags, __entry->zio_stage, __entry->zio_pipeline, \ __entry->zio_orig_flags, __entry->zio_orig_stage, \ __entry->zio_orig_pipeline, __entry->zio_reexecute, \ __entry->zio_txg, __entry->zio_error, __entry->zio_ena, \ __entry->zp_checksum, __entry->zp_compress, __entry->zp_type, \ __entry->zp_level, __entry->zp_copies, __entry->zp_dedup, \ __entry->zp_dedup_verify, __entry->zp_nopwrite DECLARE_EVENT_CLASS(zfs_l2arc_rw_class, TP_PROTO(vdev_t *vd, zio_t *zio), TP_ARGS(vd, zio), TP_STRUCT__entry( __field(uint64_t, vdev_id) __field(uint64_t, vdev_guid) __field(uint64_t, vdev_state) ZIO_TP_STRUCT_ENTRY ), TP_fast_assign( __entry->vdev_id = vd->vdev_id; __entry->vdev_guid = vd->vdev_guid; __entry->vdev_state = vd->vdev_state; ZIO_TP_FAST_ASSIGN ), TP_printk("vdev { id %llu guid %llu state %llu } " ZIO_TP_PRINTK_FMT, __entry->vdev_id, __entry->vdev_guid, __entry->vdev_state, ZIO_TP_PRINTK_ARGS) ); #define DEFINE_L2ARC_RW_EVENT(name) \ DEFINE_EVENT(zfs_l2arc_rw_class, name, \ TP_PROTO(vdev_t *vd, zio_t *zio), \ TP_ARGS(vd, zio)) DEFINE_L2ARC_RW_EVENT(zfs_l2arc__read); DEFINE_L2ARC_RW_EVENT(zfs_l2arc__write); /* * Generic support for two argument tracepoints of the form: * * DTRACE_PROBE2(..., * zio_t *, ..., * l2arc_write_callback_t *, ...); */ DECLARE_EVENT_CLASS(zfs_l2arc_iodone_class, TP_PROTO(zio_t *zio, l2arc_write_callback_t *cb), TP_ARGS(zio, cb), TP_STRUCT__entry(ZIO_TP_STRUCT_ENTRY), TP_fast_assign(ZIO_TP_FAST_ASSIGN), TP_printk(ZIO_TP_PRINTK_FMT, ZIO_TP_PRINTK_ARGS) ); #define DEFINE_L2ARC_IODONE_EVENT(name) \ DEFINE_EVENT(zfs_l2arc_iodone_class, name, \ TP_PROTO(zio_t *zio, l2arc_write_callback_t *cb), \ TP_ARGS(zio, cb)) DEFINE_L2ARC_IODONE_EVENT(zfs_l2arc__iodone); /* * Generic support for four argument tracepoints of the form: * * DTRACE_PROBE4(..., * arc_buf_hdr_t *, ..., * const blkptr_t *, * uint64_t, * const zbookmark_phys_t *); */ DECLARE_EVENT_CLASS(zfs_arc_miss_class, TP_PROTO(arc_buf_hdr_t *hdr, const blkptr_t *bp, uint64_t size, const zbookmark_phys_t *zb), TP_ARGS(hdr, bp, size, zb), TP_STRUCT__entry( __array(uint64_t, hdr_dva_word, 2) __field(uint64_t, hdr_birth) __field(uint32_t, hdr_flags) __field(uint32_t, hdr_datacnt) __field(arc_buf_contents_t, hdr_type) __field(uint64_t, hdr_size) __field(uint64_t, hdr_spa) __field(arc_state_type_t, hdr_state_type) __field(clock_t, hdr_access) __field(uint32_t, hdr_mru_hits) __field(uint32_t, hdr_mru_ghost_hits) __field(uint32_t, hdr_mfu_hits) __field(uint32_t, hdr_mfu_ghost_hits) __field(uint32_t, hdr_l2_hits) __field(int64_t, hdr_refcount) __array(uint64_t, bp_dva0, 2) __array(uint64_t, bp_dva1, 2) __array(uint64_t, bp_dva2, 2) __array(uint64_t, bp_cksum, 4) __field(uint64_t, bp_lsize) __field(uint64_t, zb_objset) __field(uint64_t, zb_object) __field(int64_t, zb_level) __field(uint64_t, zb_blkid) ), TP_fast_assign( __entry->hdr_dva_word[0] = hdr->b_dva.dva_word[0]; __entry->hdr_dva_word[1] = hdr->b_dva.dva_word[1]; __entry->hdr_birth = hdr->b_birth; __entry->hdr_flags = hdr->b_flags; __entry->hdr_datacnt = hdr->b_l1hdr.b_datacnt; __entry->hdr_size = hdr->b_size; __entry->hdr_spa = hdr->b_spa; __entry->hdr_state_type = hdr->b_l1hdr.b_state->arcs_state; __entry->hdr_access = hdr->b_l1hdr.b_arc_access; __entry->hdr_mru_hits = hdr->b_l1hdr.b_mru_hits; __entry->hdr_mru_ghost_hits = hdr->b_l1hdr.b_mru_ghost_hits; __entry->hdr_mfu_hits = hdr->b_l1hdr.b_mfu_hits; __entry->hdr_mfu_ghost_hits = hdr->b_l1hdr.b_mfu_ghost_hits; __entry->hdr_l2_hits = hdr->b_l1hdr.b_l2_hits; __entry->hdr_refcount = hdr->b_l1hdr.b_refcnt.rc_count; __entry->bp_dva0[0] = bp->blk_dva[0].dva_word[0]; __entry->bp_dva0[1] = bp->blk_dva[0].dva_word[1]; __entry->bp_dva1[0] = bp->blk_dva[1].dva_word[0]; __entry->bp_dva1[1] = bp->blk_dva[1].dva_word[1]; __entry->bp_dva2[0] = bp->blk_dva[2].dva_word[0]; __entry->bp_dva2[1] = bp->blk_dva[2].dva_word[1]; __entry->bp_cksum[0] = bp->blk_cksum.zc_word[0]; __entry->bp_cksum[1] = bp->blk_cksum.zc_word[1]; __entry->bp_cksum[2] = bp->blk_cksum.zc_word[2]; __entry->bp_cksum[3] = bp->blk_cksum.zc_word[3]; __entry->bp_lsize = size; __entry->zb_objset = zb->zb_objset; __entry->zb_object = zb->zb_object; __entry->zb_level = zb->zb_level; __entry->zb_blkid = zb->zb_blkid; ), TP_printk("hdr { dva 0x%llx:0x%llx birth %llu " "flags 0x%x datacnt %u size %llu spa %llu state_type %u " "access %lu mru_hits %u mru_ghost_hits %u mfu_hits %u " "mfu_ghost_hits %u l2_hits %u refcount %lli } " "bp { dva0 0x%llx:0x%llx dva1 0x%llx:0x%llx dva2 " "0x%llx:0x%llx cksum 0x%llx:0x%llx:0x%llx:0x%llx " "lsize %llu } zb { objset %llu object %llu level %lli " "blkid %llu }", __entry->hdr_dva_word[0], __entry->hdr_dva_word[1], __entry->hdr_birth, __entry->hdr_flags, __entry->hdr_datacnt, __entry->hdr_size, __entry->hdr_spa, __entry->hdr_state_type, __entry->hdr_access, __entry->hdr_mru_hits, __entry->hdr_mru_ghost_hits, __entry->hdr_mfu_hits, __entry->hdr_mfu_ghost_hits, __entry->hdr_l2_hits, __entry->hdr_refcount, __entry->bp_dva0[0], __entry->bp_dva0[1], __entry->bp_dva1[0], __entry->bp_dva1[1], __entry->bp_dva2[0], __entry->bp_dva2[1], __entry->bp_cksum[0], __entry->bp_cksum[1], __entry->bp_cksum[2], __entry->bp_cksum[3], __entry->bp_lsize, __entry->zb_objset, __entry->zb_object, __entry->zb_level, __entry->zb_blkid) ); #define DEFINE_ARC_MISS_EVENT(name) \ DEFINE_EVENT(zfs_arc_miss_class, name, \ TP_PROTO(arc_buf_hdr_t *hdr, \ const blkptr_t *bp, uint64_t size, const zbookmark_phys_t *zb), \ TP_ARGS(hdr, bp, size, zb)) DEFINE_ARC_MISS_EVENT(zfs_arc__miss); /* * Generic support for four argument tracepoints of the form: * * DTRACE_PROBE4(..., * l2arc_dev_t *, ..., * list_t *, ..., * uint64_t, ..., * boolean_t, ...); */ DECLARE_EVENT_CLASS(zfs_l2arc_evict_class, TP_PROTO(l2arc_dev_t *dev, list_t *buflist, uint64_t taddr, boolean_t all), TP_ARGS(dev, buflist, taddr, all), TP_STRUCT__entry( __field(uint64_t, vdev_id) __field(uint64_t, vdev_guid) __field(uint64_t, vdev_state) __field(uint64_t, l2ad_hand) __field(uint64_t, l2ad_start) __field(uint64_t, l2ad_end) __field(boolean_t, l2ad_first) __field(boolean_t, l2ad_writing) __field(uint64_t, taddr) __field(boolean_t, all) ), TP_fast_assign( __entry->vdev_id = dev->l2ad_vdev->vdev_id; __entry->vdev_guid = dev->l2ad_vdev->vdev_guid; __entry->vdev_state = dev->l2ad_vdev->vdev_state; __entry->l2ad_hand = dev->l2ad_hand; __entry->l2ad_start = dev->l2ad_start; __entry->l2ad_end = dev->l2ad_end; __entry->l2ad_first = dev->l2ad_first; __entry->l2ad_writing = dev->l2ad_writing; __entry->taddr = taddr; __entry->all = all; ), TP_printk("l2ad { vdev { id %llu guid %llu state %llu } " "hand %llu start %llu end %llu " "first %d writing %d } taddr %llu all %d", __entry->vdev_id, __entry->vdev_guid, __entry->vdev_state, __entry->l2ad_hand, __entry->l2ad_start, __entry->l2ad_end, __entry->l2ad_first, __entry->l2ad_writing, __entry->taddr, __entry->all) ); #define DEFINE_L2ARC_EVICT_EVENT(name) \ DEFINE_EVENT(zfs_l2arc_evict_class, name, \ TP_PROTO(l2arc_dev_t *dev, \ list_t *buflist, uint64_t taddr, boolean_t all), \ TP_ARGS(dev, buflist, taddr, all)) DEFINE_L2ARC_EVICT_EVENT(zfs_l2arc__evict); #endif /* _TRACE_ARC_H */ #undef TRACE_INCLUDE_PATH #undef TRACE_INCLUDE_FILE #define TRACE_INCLUDE_PATH sys #define TRACE_INCLUDE_FILE trace_arc #include #endif /* _KERNEL && HAVE_DECLARE_EVENT_CLASS */ diff --git a/man/man5/zfs-module-parameters.5 b/man/man5/zfs-module-parameters.5 index 6be382eacafe..eabd8ebe9eeb 100644 --- a/man/man5/zfs-module-parameters.5 +++ b/man/man5/zfs-module-parameters.5 @@ -1,1848 +1,1851 @@ '\" te .\" Copyright (c) 2013 by Turbo Fredriksson . All rights reserved. .\" 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] .TH ZFS-MODULE-PARAMETERS 5 "Nov 16, 2013" .SH NAME zfs\-module\-parameters \- ZFS module parameters .SH DESCRIPTION .sp .LP Description of the different parameters to the ZFS module. .SS "Module parameters" .sp .LP .sp .ne 2 .na \fBl2arc_feed_again\fR (int) .ad .RS 12n Turbo L2ARC warmup .sp Use \fB1\fR for yes (default) and \fB0\fR to disable. .RE .sp .ne 2 .na \fBl2arc_feed_min_ms\fR (ulong) .ad .RS 12n Min feed interval in milliseconds .sp Default value: \fB200\fR. .RE .sp .ne 2 .na \fBl2arc_feed_secs\fR (ulong) .ad .RS 12n Seconds between L2ARC writing .sp Default value: \fB1\fR. .RE .sp .ne 2 .na \fBl2arc_headroom\fR (ulong) .ad .RS 12n Number of max device writes to precache .sp Default value: \fB2\fR. .RE .sp .ne 2 .na \fBl2arc_headroom_boost\fR (ulong) .ad .RS 12n Compressed l2arc_headroom multiplier .sp Default value: \fB200\fR. .RE .sp .ne 2 .na \fBl2arc_nocompress\fR (int) .ad .RS 12n Skip compressing L2ARC buffers .sp Use \fB1\fR for yes and \fB0\fR for no (default). .RE .sp .ne 2 .na \fBl2arc_noprefetch\fR (int) .ad .RS 12n Skip caching prefetched buffers .sp Use \fB1\fR for yes (default) and \fB0\fR to disable. .RE .sp .ne 2 .na \fBl2arc_norw\fR (int) .ad .RS 12n No reads during writes .sp Use \fB1\fR for yes and \fB0\fR for no (default). .RE .sp .ne 2 .na \fBl2arc_write_boost\fR (ulong) .ad .RS 12n Extra write bytes during device warmup .sp Default value: \fB8,388,608\fR. .RE .sp .ne 2 .na \fBl2arc_write_max\fR (ulong) .ad .RS 12n Max write bytes per interval .sp Default value: \fB8,388,608\fR. .RE .sp .ne 2 .na \fBmetaslab_aliquot\fR (ulong) .ad .RS 12n 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 a top-level vdev before moving on to the next one. .sp Default value: \fB524,288\fR. .RE .sp .ne 2 .na \fBmetaslab_bias_enabled\fR (int) .ad .RS 12n Enable metaslab group biasing based on its vdev's over- or under-utilization relative to the pool. .sp Use \fB1\fR for yes (default) and \fB0\fR for no. .RE .sp .ne 2 .na \fBmetaslab_debug_load\fR (int) .ad .RS 12n Load all metaslabs during pool import. .sp Use \fB1\fR for yes and \fB0\fR for no (default). .RE .sp .ne 2 .na \fBmetaslab_debug_unload\fR (int) .ad .RS 12n Prevent metaslabs from being unloaded. .sp Use \fB1\fR for yes and \fB0\fR for no (default). .RE .sp .ne 2 .na \fBmetaslab_fragmentation_factor_enabled\fR (int) .ad .RS 12n Enable use of the fragmentation metric in computing metaslab weights. .sp Use \fB1\fR for yes (default) and \fB0\fR for no. .RE .sp .ne 2 .na \fBmetaslabs_per_vdev\fR (int) .ad .RS 12n When a vdev is added, it will be divided into approximately (but no more than) this number of metaslabs. .sp Default value: \fB200\fR. .RE .sp .ne 2 .na \fBmetaslab_preload_enabled\fR (int) .ad .RS 12n Enable metaslab group preloading. .sp Use \fB1\fR for yes (default) and \fB0\fR for no. .RE .sp .ne 2 .na \fBmetaslab_lba_weighting_enabled\fR (int) .ad .RS 12n 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. .sp Use \fB1\fR for yes (default) and \fB0\fR for no. .RE .sp .ne 2 .na \fBspa_config_path\fR (charp) .ad .RS 12n SPA config file .sp Default value: \fB/etc/zfs/zpool.cache\fR. .RE .sp .ne 2 .na \fBspa_asize_inflation\fR (int) .ad .RS 12n 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. .sp Default value: 24 .RE .sp .ne 2 .na \fBspa_load_verify_data\fR (int) .ad .RS 12n Whether to traverse data blocks during an "extreme rewind" (\fB-X\fR) import. Use 0 to disable and 1 to enable. An extreme rewind import normally performs a full traversal of all blocks in the pool for verification. If this parameter is set to 0, 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. .sp Default value: 1 .RE .sp .ne 2 .na \fBspa_load_verify_metadata\fR (int) .ad .RS 12n Whether to traverse blocks during an "extreme rewind" (\fB-X\fR) pool import. Use 0 to disable and 1 to enable. An extreme rewind import normally performs a full traversal of all blocks in the pool for verification. If this parameter is set to 1, the traversal is not performed. It can be toggled once the import has started to stop or start the traversal. .sp Default value: 1 .RE .sp .ne 2 .na \fBspa_load_verify_maxinflight\fR (int) .ad .RS 12n Maximum concurrent I/Os during the traversal performed during an "extreme rewind" (\fB-X\fR) pool import. .sp Default value: 10000 .RE .sp .ne 2 .na \fBspa_slop_shift\fR (int) .ad .RS 12n Normally, we don't allow the last 3.2% (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 ENOSPC. .sp Default value: 5 .RE .sp .ne 2 .na \fBzfetch_array_rd_sz\fR (ulong) .ad .RS 12n If prefetching is enabled, disable prefetching for reads larger than this size. .sp Default value: \fB1,048,576\fR. .RE .sp .ne 2 .na -\fBzfetch_block_cap\fR (uint) +\fBzfetch_max_distance\fR (uint) .ad .RS 12n -Max number of blocks to prefetch at a time +Max bytes to prefetch per stream (default 8MB). .sp -Default value: \fB256\fR. +Default value: \fB8,388,608\fR. .RE .sp .ne 2 .na \fBzfetch_max_streams\fR (uint) .ad .RS 12n Max number of streams per zfetch (prefetch streams per file). .sp Default value: \fB8\fR. .RE .sp .ne 2 .na \fBzfetch_min_sec_reap\fR (uint) .ad .RS 12n Min time before an active prefetch stream can be reclaimed .sp Default value: \fB2\fR. .RE .sp .ne 2 .na \fBzfs_arc_average_blocksize\fR (int) .ad .RS 12n The ARC's buffer hash table is sized based on the assumption of an average block size of \fBzfs_arc_average_blocksize\fR (default 8K). This works out to roughly 1MB of hash table per 1GB 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. .sp Default value: \fB8192\fR. .RE .sp .ne 2 .na \fBzfs_arc_evict_batch_limit\fR (int) .ad .RS 12n 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. .sp Default value: \fB10\fR. .RE .sp .ne 2 .na \fBzfs_arc_grow_retry\fR (int) .ad .RS 12n Seconds before growing arc size .sp Default value: \fB5\fR. .RE .sp .ne 2 .na \fBzfs_arc_lotsfree_percent\fR (int) .ad .RS 12n Throttle I/O when free system memory drops below this percentage of total system memory. Setting this value to 0 will disable the throttle. .sp Default value: \fB10\fR. .RE .sp .ne 2 .na \fBzfs_arc_max\fR (ulong) .ad .RS 12n Max arc size .sp Default value: \fB0\fR. .RE .sp .ne 2 .na \fBzfs_arc_meta_limit\fR (ulong) .ad .RS 12n The maximum allowed size in bytes that meta data buffers are allowed to consume in the ARC. When this limit is reached meta data buffers will be reclaimed even if the overall arc_c_max has not been reached. This value defaults to 0 which indicates that 3/4 of the ARC may be used for meta data. .sp Default value: \fB0\fR. .RE .sp .ne 2 .na \fBzfs_arc_meta_min\fR (ulong) .ad .RS 12n The minimum allowed size in bytes that meta data buffers may consume in the ARC. This value defaults to 0 which disables a floor on the amount of the ARC devoted meta data. .sp Default value: \fB0\fR. .RE .sp .ne 2 .na \fBzfs_arc_meta_prune\fR (int) .ad .RS 12n The number of dentries and inodes to be scanned looking for entries which can be dropped. This may be required when the ARC reaches the \fBzfs_arc_meta_limit\fR because dentries and inodes can pin buffers in the ARC. Increasing this value will cause to dentry and inode caches to be pruned more aggressively. Setting this value to 0 will disable pruning the inode and dentry caches. .sp Default value: \fB10,000\fR. .RE .sp .ne 2 .na \fBzfs_arc_meta_adjust_restarts\fR (ulong) .ad .RS 12n The number of restart passes to make while scanning the ARC attempting the free buffers in order to stay below the \fBzfs_arc_meta_limit\fR. This value should not need to be tuned but is available to facilitate performance analysis. .sp Default value: \fB4096\fR. .RE .sp .ne 2 .na \fBzfs_arc_min\fR (ulong) .ad .RS 12n Min arc size .sp Default value: \fB100\fR. .RE .sp .ne 2 .na \fBzfs_arc_min_prefetch_lifespan\fR (int) .ad .RS 12n Min life of prefetch block .sp Default value: \fB100\fR. .RE .sp .ne 2 .na \fBzfs_arc_num_sublists_per_state\fR (int) .ad .RS 12n To allow more fine-grained locking, each ARC state contains a series of lists for both data and meta data objects. Locking is performed at the level of these "sub-lists". This parameters controls the number of sub-lists per ARC state. .sp Default value: 1 or the number of on-online CPUs, whichever is greater .RE .sp .ne 2 .na \fBzfs_arc_overflow_shift\fR (int) .ad .RS 12n The ARC size is considered to be overflowing if it exceeds the current ARC target size (arc_c) by a threshold determined by this parameter. The threshold is calculated as a fraction of arc_c using the formula "arc_c >> \fBzfs_arc_overflow_shift\fR". The default value of 8 causes the ARC to be considered to be overflowing if it exceeds the target size by 1/256th (0.3%) of the target size. When the ARC is overflowing, new buffer allocations are stalled until the reclaim thread catches up and the overflow condition no longer exists. .sp Default value: \fB8\fR. .RE .sp .ne 2 .na \fBzfs_arc_p_min_shift\fR (int) .ad .RS 12n arc_c shift to calc min/max arc_p .sp Default value: \fB4\fR. .RE .sp .ne 2 .na \fBzfs_arc_p_aggressive_disable\fR (int) .ad .RS 12n Disable aggressive arc_p growth .sp Use \fB1\fR for yes (default) and \fB0\fR to disable. .RE .sp .ne 2 .na \fBzfs_arc_p_dampener_disable\fR (int) .ad .RS 12n Disable arc_p adapt dampener .sp Use \fB1\fR for yes (default) and \fB0\fR to disable. .RE .sp .ne 2 .na \fBzfs_arc_shrink_shift\fR (int) .ad .RS 12n log2(fraction of arc to reclaim) .sp Default value: \fB5\fR. .RE .sp .ne 2 .na \fBzfs_arc_sys_free\fR (ulong) .ad .RS 12n The target number of bytes the ARC should leave as free memory on the system. Defaults to the larger of 1/64 of physical memory or 512K. Setting this option to a non-zero value will override the default. .sp Default value: \fB0\fR. .RE .sp .ne 2 .na \fBzfs_autoimport_disable\fR (int) .ad .RS 12n Disable pool import at module load by ignoring the cache file (typically \fB/etc/zfs/zpool.cache\fR). .sp Use \fB1\fR for yes (default) and \fB0\fR for no. .RE .sp .ne 2 .na \fBzfs_dbgmsg_enable\fR (int) .ad .RS 12n Internally ZFS keeps a small log to facilitate debugging. By default the log is disabled, to enable it set this option to 1. The contents of the log can be accessed by reading the /proc/spl/kstat/zfs/dbgmsg file. Writing 0 to this proc file clears the log. .sp Default value: \fB0\fR. .RE .sp .ne 2 .na \fBzfs_dbgmsg_maxsize\fR (int) .ad .RS 12n The maximum size in bytes of the internal ZFS debug log. .sp Default value: \fB4M\fR. .RE .sp .ne 2 .na \fBzfs_dbuf_state_index\fR (int) .ad .RS 12n Calculate arc header index .sp Default value: \fB0\fR. .RE .sp .ne 2 .na \fBzfs_deadman_enabled\fR (int) .ad .RS 12n Enable deadman timer .sp Use \fB1\fR for yes (default) and \fB0\fR to disable. .RE .sp .ne 2 .na \fBzfs_deadman_synctime_ms\fR (ulong) .ad .RS 12n 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 1000 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 a zevent being logged. .sp Default value: \fB1,000,000\fR. .RE .sp .ne 2 .na \fBzfs_dedup_prefetch\fR (int) .ad .RS 12n Enable prefetching dedup-ed blks .sp Use \fB1\fR for yes and \fB0\fR to disable (default). .RE .sp .ne 2 .na \fBzfs_delay_min_dirty_percent\fR (int) .ad .RS 12n Start to delay each transaction once there is this amount of dirty data, expressed as a percentage of \fBzfs_dirty_data_max\fR. This value should be >= zfs_vdev_async_write_active_max_dirty_percent. See the section "ZFS TRANSACTION DELAY". .sp Default value: \fB60\fR. .RE .sp .ne 2 .na \fBzfs_delay_scale\fR (int) .ad .RS 12n This controls how quickly the transaction delay approaches infinity. Larger values cause longer delays for a given amount of dirty data. .sp 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 10x and 1/10th this number. .sp See the section "ZFS TRANSACTION DELAY". .sp Note: \fBzfs_delay_scale\fR * \fBzfs_dirty_data_max\fR must be < 2^64. .sp Default value: \fB500,000\fR. .RE .sp .ne 2 .na \fBzfs_dirty_data_max\fR (int) .ad .RS 12n 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 \fBzfs_dirty_data_max_percent\fR. See the section "ZFS TRANSACTION DELAY". .sp Default value: 10 percent of all memory, capped at \fBzfs_dirty_data_max_max\fR. .RE .sp .ne 2 .na \fBzfs_dirty_data_max_max\fR (int) .ad .RS 12n Maximum allowable value of \fBzfs_dirty_data_max\fR, expressed in bytes. This limit is only enforced at module load time, and will be ignored if \fBzfs_dirty_data_max\fR is later changed. This parameter takes precedence over \fBzfs_dirty_data_max_max_percent\fR. See the section "ZFS TRANSACTION DELAY". .sp Default value: 25% of physical RAM. .RE .sp .ne 2 .na \fBzfs_dirty_data_max_max_percent\fR (int) .ad .RS 12n Maximum allowable value of \fBzfs_dirty_data_max\fR, expressed as a percentage of physical RAM. This limit is only enforced at module load time, and will be ignored if \fBzfs_dirty_data_max\fR is later changed. The parameter \fBzfs_dirty_data_max_max\fR takes precedence over this one. See the section "ZFS TRANSACTION DELAY". .sp Default value: 25 .RE .sp .ne 2 .na \fBzfs_dirty_data_max_percent\fR (int) .ad .RS 12n 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 \fBzfs_dirty_data_max\fR takes precedence over this one. See the section "ZFS TRANSACTION DELAY". .sp Default value: 10%, subject to \fBzfs_dirty_data_max_max\fR. .RE .sp .ne 2 .na \fBzfs_dirty_data_sync\fR (int) .ad .RS 12n Start syncing out a transaction group if there is at least this much dirty data. .sp Default value: \fB67,108,864\fR. .RE .sp .ne 2 .na \fBzfs_free_max_blocks\fR (ulong) .ad .RS 12n Maximum number of blocks freed in a single txg. .sp Default value: \fB100,000\fR. .RE .sp .ne 2 .na \fBzfs_vdev_async_read_max_active\fR (int) .ad .RS 12n Maxium asynchronous read I/Os active to each device. See the section "ZFS I/O SCHEDULER". .sp Default value: \fB3\fR. .RE .sp .ne 2 .na \fBzfs_vdev_async_read_min_active\fR (int) .ad .RS 12n Minimum asynchronous read I/Os active to each device. See the section "ZFS I/O SCHEDULER". .sp Default value: \fB1\fR. .RE .sp .ne 2 .na \fBzfs_vdev_async_write_active_max_dirty_percent\fR (int) .ad .RS 12n When the pool has more than \fBzfs_vdev_async_write_active_max_dirty_percent\fR dirty data, use \fBzfs_vdev_async_write_max_active\fR to limit active async writes. If the dirty data is between min and max, the active I/O limit is linearly interpolated. See the section "ZFS I/O SCHEDULER". .sp Default value: \fB60\fR. .RE .sp .ne 2 .na \fBzfs_vdev_async_write_active_min_dirty_percent\fR (int) .ad .RS 12n When the pool has less than \fBzfs_vdev_async_write_active_min_dirty_percent\fR dirty data, use \fBzfs_vdev_async_write_min_active\fR to limit active async writes. If the dirty data is between min and max, the active I/O limit is linearly interpolated. See the section "ZFS I/O SCHEDULER". .sp Default value: \fB30\fR. .RE .sp .ne 2 .na \fBzfs_vdev_async_write_max_active\fR (int) .ad .RS 12n Maxium asynchronous write I/Os active to each device. See the section "ZFS I/O SCHEDULER". .sp Default value: \fB10\fR. .RE .sp .ne 2 .na \fBzfs_vdev_async_write_min_active\fR (int) .ad .RS 12n Minimum asynchronous write I/Os active to each device. See the section "ZFS I/O SCHEDULER". .sp Default value: \fB1\fR. .RE .sp .ne 2 .na \fBzfs_vdev_max_active\fR (int) .ad .RS 12n The maximum number of I/Os active to each device. Ideally, this will be >= the sum of each queue's max_active. It must be at least the sum of each queue's min_active. See the section "ZFS I/O SCHEDULER". .sp Default value: \fB1,000\fR. .RE .sp .ne 2 .na \fBzfs_vdev_scrub_max_active\fR (int) .ad .RS 12n Maxium scrub I/Os active to each device. See the section "ZFS I/O SCHEDULER". .sp Default value: \fB2\fR. .RE .sp .ne 2 .na \fBzfs_vdev_scrub_min_active\fR (int) .ad .RS 12n Minimum scrub I/Os active to each device. See the section "ZFS I/O SCHEDULER". .sp Default value: \fB1\fR. .RE .sp .ne 2 .na \fBzfs_vdev_sync_read_max_active\fR (int) .ad .RS 12n Maxium synchronous read I/Os active to each device. See the section "ZFS I/O SCHEDULER". .sp Default value: \fB10\fR. .RE .sp .ne 2 .na \fBzfs_vdev_sync_read_min_active\fR (int) .ad .RS 12n Minimum synchronous read I/Os active to each device. See the section "ZFS I/O SCHEDULER". .sp Default value: \fB10\fR. .RE .sp .ne 2 .na \fBzfs_vdev_sync_write_max_active\fR (int) .ad .RS 12n Maxium synchronous write I/Os active to each device. See the section "ZFS I/O SCHEDULER". .sp Default value: \fB10\fR. .RE .sp .ne 2 .na \fBzfs_vdev_sync_write_min_active\fR (int) .ad .RS 12n Minimum synchronous write I/Os active to each device. See the section "ZFS I/O SCHEDULER". .sp Default value: \fB10\fR. .RE .sp .ne 2 .na \fBzfs_disable_dup_eviction\fR (int) .ad .RS 12n Disable duplicate buffer eviction .sp Use \fB1\fR for yes and \fB0\fR for no (default). .RE .sp .ne 2 .na \fBzfs_expire_snapshot\fR (int) .ad .RS 12n Seconds to expire .zfs/snapshot .sp Default value: \fB300\fR. .RE .sp .ne 2 .na \fBzfs_admin_snapshot\fR (int) .ad .RS 12n Allow the creation, removal, or renaming of entries in the .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 'no_root_squash' option set. This functionality is disabled by default. .sp Use \fB1\fR for yes and \fB0\fR for no (default). .RE .sp .ne 2 .na \fBzfs_flags\fR (int) .ad .RS 12n Set additional debugging flags. The following flags may be bitwise-or'd together. .sp .TS box; rB lB lB lB r l. Value Symbolic 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. _ 32 ZFS_DEBUG_SPA Enable spa_dbgmsg entries in the debug log. _ 64 ZFS_DEBUG_ZIO_FREE Enable verification of block frees. _ 128 ZFS_DEBUG_HISTOGRAM_VERIFY Enable extra spacemap histogram verifications. .TE .sp * Requires debug build. .sp Default value: \fB0\fR. .RE .sp .ne 2 .na \fBzfs_free_leak_on_eio\fR (int) .ad .RS 12n 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. .sp Default value: \fB0\fR. .RE .sp .ne 2 .na \fBzfs_free_min_time_ms\fR (int) .ad .RS 12n Min millisecs to free per txg .sp Default value: \fB1,000\fR. .RE .sp .ne 2 .na \fBzfs_immediate_write_sz\fR (long) .ad .RS 12n Largest data block to write to zil .sp Default value: \fB32,768\fR. .RE .sp .ne 2 .na \fBzfs_max_recordsize\fR (int) .ad .RS 12n 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. Additionally, very large blocks can have an impact on i/o latency, and also potentially on the memory allocator. Therefore, we do not allow the recordsize to be set larger than zfs_max_recordsize (default 1MB). Larger blocks can be created by changing this tunable, and pools with larger blocks can always be imported and used, regardless of this setting. .sp Default value: \fB1,048,576\fR. .RE .sp .ne 2 .na \fBzfs_mdcomp_disable\fR (int) .ad .RS 12n Disable meta data compression .sp Use \fB1\fR for yes and \fB0\fR for no (default). .RE .sp .ne 2 .na \fBzfs_metaslab_fragmentation_threshold\fR (int) .ad .RS 12n Allow metaslabs to keep their active state as long as their fragmentation percentage is less than or equal to this value. An active metaslab that exceeds this threshold will no longer keep its active status allowing better metaslabs to be selected. .sp Default value: \fB70\fR. .RE .sp .ne 2 .na \fBzfs_mg_fragmentation_threshold\fR (int) .ad .RS 12n Metaslab groups are considered eligible for allocations if their fragmenation 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. .sp Default value: \fB85\fR. .RE .sp .ne 2 .na \fBzfs_mg_noalloc_threshold\fR (int) .ad .RS 12n 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 0 disables the feature and causes all metaslab groups to be eligible for allocations. This parameter allows 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 \fBzfs_mg_alloc_failures\fR facility. .sp Default value: \fB0\fR. .RE .sp .ne 2 .na \fBzfs_no_scrub_io\fR (int) .ad .RS 12n Set for no scrub I/O .sp Use \fB1\fR for yes and \fB0\fR for no (default). .RE .sp .ne 2 .na \fBzfs_no_scrub_prefetch\fR (int) .ad .RS 12n Set for no scrub prefetching .sp Use \fB1\fR for yes and \fB0\fR for no (default). .RE .sp .ne 2 .na \fBzfs_nocacheflush\fR (int) .ad .RS 12n Disable cache flushes .sp Use \fB1\fR for yes and \fB0\fR for no (default). .RE .sp .ne 2 .na \fBzfs_nopwrite_enabled\fR (int) .ad .RS 12n Enable NOP writes .sp Use \fB1\fR for yes (default) and \fB0\fR to disable. .RE .sp .ne 2 .na \fBzfs_pd_bytes_max\fR (int) .ad .RS 12n The number of bytes which should be prefetched. .sp Default value: \fB52,428,800\fR. .RE .sp .ne 2 .na \fBzfs_prefetch_disable\fR (int) .ad .RS 12n -Disable all ZFS prefetching +This tunable disables predictive prefetch. Note that it leaves "prescient" +prefetch (e.g. prefetch for zfs send) intact. Unlike predictive prefetch, +prescient prefetch never issues i/os that end up not being needed, so it +can't hurt performance. .sp Use \fB1\fR for yes and \fB0\fR for no (default). .RE .sp .ne 2 .na \fBzfs_read_chunk_size\fR (long) .ad .RS 12n Bytes to read per chunk .sp Default value: \fB1,048,576\fR. .RE .sp .ne 2 .na \fBzfs_read_history\fR (int) .ad .RS 12n Historic statistics for the last N reads .sp Default value: \fB0\fR. .RE .sp .ne 2 .na \fBzfs_read_history_hits\fR (int) .ad .RS 12n Include cache hits in read history .sp Use \fB1\fR for yes and \fB0\fR for no (default). .RE .sp .ne 2 .na \fBzfs_recover\fR (int) .ad .RS 12n 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. .sp Use \fB1\fR for yes and \fB0\fR for no (default). .RE .sp .ne 2 .na \fBzfs_resilver_delay\fR (int) .ad .RS 12n Number of ticks to delay prior to issuing a resilver I/O operation when a non-resilver or non-scrub I/O operation has occurred within the past \fBzfs_scan_idle\fR ticks. .sp Default value: \fB2\fR. .RE .sp .ne 2 .na \fBzfs_resilver_min_time_ms\fR (int) .ad .RS 12n Min millisecs to resilver per txg .sp Default value: \fB3,000\fR. .RE .sp .ne 2 .na \fBzfs_scan_idle\fR (int) .ad .RS 12n Idle window in clock ticks. During a scrub or a resilver, if a non-scrub or non-resilver I/O operation has occurred during this window, the next scrub or resilver operation is delayed by, respectively \fBzfs_scrub_delay\fR or \fBzfs_resilver_delay\fR ticks. .sp Default value: \fB50\fR. .RE .sp .ne 2 .na \fBzfs_scan_min_time_ms\fR (int) .ad .RS 12n Min millisecs to scrub per txg .sp Default value: \fB1,000\fR. .RE .sp .ne 2 .na \fBzfs_scrub_delay\fR (int) .ad .RS 12n Number of ticks to delay prior to issuing a scrub I/O operation when a non-scrub or non-resilver I/O operation has occurred within the past \fBzfs_scan_idle\fR ticks. .sp Default value: \fB4\fR. .RE .sp .ne 2 .na \fBzfs_send_corrupt_data\fR (int) .ad .RS 12n Allow to send corrupt data (ignore read/checksum errors when sending data) .sp Use \fB1\fR for yes and \fB0\fR for no (default). .RE .sp .ne 2 .na \fBzfs_sync_pass_deferred_free\fR (int) .ad .RS 12n Defer frees starting in this pass .sp Default value: \fB2\fR. .RE .sp .ne 2 .na \fBzfs_sync_pass_dont_compress\fR (int) .ad .RS 12n Don't compress starting in this pass .sp Default value: \fB5\fR. .RE .sp .ne 2 .na \fBzfs_sync_pass_rewrite\fR (int) .ad .RS 12n Rewrite new bps starting in this pass .sp Default value: \fB2\fR. .RE .sp .ne 2 .na \fBzfs_top_maxinflight\fR (int) .ad .RS 12n Max I/Os per top-level vdev during scrub or resilver operations. .sp Default value: \fB32\fR. .RE .sp .ne 2 .na \fBzfs_txg_history\fR (int) .ad .RS 12n Historic statistics for the last N txgs .sp Default value: \fB0\fR. .RE .sp .ne 2 .na \fBzfs_txg_timeout\fR (int) .ad .RS 12n Max seconds worth of delta per txg .sp Default value: \fB5\fR. .RE .sp .ne 2 .na \fBzfs_vdev_aggregation_limit\fR (int) .ad .RS 12n Max vdev I/O aggregation size .sp Default value: \fB131,072\fR. .RE .sp .ne 2 .na \fBzfs_vdev_cache_bshift\fR (int) .ad .RS 12n Shift size to inflate reads too .sp Default value: \fB16\fR. .RE .sp .ne 2 .na \fBzfs_vdev_cache_max\fR (int) .ad .RS 12n Inflate reads small than max .RE .sp .ne 2 .na \fBzfs_vdev_cache_size\fR (int) .ad .RS 12n Total size of the per-disk cache .sp Default value: \fB0\fR. .RE .sp .ne 2 .na \fBzfs_vdev_mirror_switch_us\fR (int) .ad .RS 12n Switch mirrors every N usecs .sp Default value: \fB10,000\fR. .RE .sp .ne 2 .na \fBzfs_vdev_read_gap_limit\fR (int) .ad .RS 12n Aggregate read I/O over gap .sp Default value: \fB32,768\fR. .RE .sp .ne 2 .na \fBzfs_vdev_scheduler\fR (charp) .ad .RS 12n I/O scheduler .sp Default value: \fBnoop\fR. .RE .sp .ne 2 .na \fBzfs_vdev_write_gap_limit\fR (int) .ad .RS 12n Aggregate write I/O over gap .sp Default value: \fB4,096\fR. .RE .sp .ne 2 .na \fBzfs_zevent_cols\fR (int) .ad .RS 12n Max event column width .sp Default value: \fB80\fR. .RE .sp .ne 2 .na \fBzfs_zevent_console\fR (int) .ad .RS 12n Log events to the console .sp Use \fB1\fR for yes and \fB0\fR for no (default). .RE .sp .ne 2 .na \fBzfs_zevent_len_max\fR (int) .ad .RS 12n Max event queue length .sp Default value: \fB0\fR. .RE .sp .ne 2 .na \fBzil_replay_disable\fR (int) .ad .RS 12n Disable intent logging replay .sp Use \fB1\fR for yes and \fB0\fR for no (default). .RE .sp .ne 2 .na \fBzil_slog_limit\fR (ulong) .ad .RS 12n Max commit bytes to separate log device .sp Default value: \fB1,048,576\fR. .RE .sp .ne 2 .na \fBzio_delay_max\fR (int) .ad .RS 12n Max zio millisecond delay before posting event .sp Default value: \fB30,000\fR. .RE .sp .ne 2 .na \fBzio_requeue_io_start_cut_in_line\fR (int) .ad .RS 12n Prioritize requeued I/O .sp Default value: \fB0\fR. .RE .sp .ne 2 .na \fBzio_taskq_batch_pct\fR (uint) .ad .RS 12n Percentage of online CPUs (or CPU cores, etc) which will run a worker thread for IO. These workers are responsible for IO work such as compression and checksum calculations. Fractional number of CPUs will be rounded down. .sp The default value of 75 was chosen to avoid using all CPUs which can result in latency issues and inconsistent application performance, especially when high compression is enabled. .sp Default value: \fB75\fR. .RE .sp .ne 2 .na \fBzvol_inhibit_dev\fR (uint) .ad .RS 12n Do not create zvol device nodes .sp Use \fB1\fR for yes and \fB0\fR for no (default). .RE .sp .ne 2 .na \fBzvol_major\fR (uint) .ad .RS 12n Major number for zvol device .sp Default value: \fB230\fR. .RE .sp .ne 2 .na \fBzvol_max_discard_blocks\fR (ulong) .ad .RS 12n Max number of blocks to discard at once .sp Default value: \fB16,384\fR. .RE .sp .ne 2 .na \fBzvol_prefetch_bytes\fR (uint) .ad .RS 12n When adding a zvol to the system prefetch \fBzvol_prefetch_bytes\fR 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 \fBblkid(8)\fR or by the kernel scanning for a partition table. .sp Default value: \fB131,072\fR. .RE .SH ZFS I/O SCHEDULER ZFS issues I/O operations to leaf vdevs to satisfy and complete I/Os. The I/O scheduler determines when and in what order those operations are issued. The I/O 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, \fBzfs_vdev_max_active\fR. Note that the sum of the per-queue minimums must not exceed the aggregate maximum. If the sum of the per-queue maximums exceeds the aggregate maximum, then the number of active I/Os may reach \fBzfs_vdev_max_active\fR, in which case no further I/Os will be issued regardless of whether all per-queue minimums have been met. .sp For many physical devices, throughput increases with the number of concurrent operations, but latency typically suffers. Further, physical devices typically have a limit at which more concurrent operations have no effect on throughput or can actually cause it to decrease. .sp 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 is queued or an operation completes, the I/O scheduler looks for new operations to issue. .sp In general, smaller max_active's will lead to lower latency of synchronous operations. Larger max_active's may lead to higher overall throughput, depending on underlying storage. .sp The ratio of the queues' max_actives determines the balance of performance between reads, writes, and scrubs. E.g., increasing \fBzfs_vdev_scrub_max_active\fR will cause the scrub or resilver to complete more quickly, but reads and writes to have higher latency and lower throughput. .sp 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 I/Os 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 concurrent operations also stabilizes the response time of operations from other -- and in particular synchronous -- queues. In broad strokes, the I/O scheduler will issue more concurrent operations from the async write queue as there's more dirty data in the pool. .sp Async Writes .sp 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. .nf | o---------| <-- zfs_vdev_async_write_max_active ^ | /^ | | | / | | active | / | | I/O | / | | count | / | | | / | | |-------o | | <-- zfs_vdev_async_write_min_active 0|_______^______|_________| 0% | | 100% of zfs_dirty_data_max | | | `-- zfs_vdev_async_write_active_max_dirty_percent `--------- zfs_vdev_async_write_active_min_dirty_percent .fi 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. .sp Ideally, the amount of dirty data on a busy pool will stay in the sloped part of the function between \fBzfs_vdev_async_write_active_min_dirty_percent\fR and \fBzfs_vdev_async_write_active_max_dirty_percent\fR. 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. .sp 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. .sp 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. .sp The minimum time for a transaction to take is calculated as: .nf min_time = zfs_delay_scale * (dirty - min) / (max - dirty) min_time is then capped at 100 milliseconds. .fi .sp 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 \fBzfs_delay_min_dirty_percent\fR. This should typically be at or above \fBzfs_vdev_async_write_active_max_dirty_percent\fR 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 \fBzfs_delay_scale\fR. Roughly speaking, this variable determines the amount of delay at the midpoint of the curve. .sp .nf delay 10ms +-------------------------------------------------------------*+ | *| 9ms + *+ | *| 8ms + *+ | * | 7ms + * + | * | 6ms + * + | * | 5ms + * + | * | 4ms + * + | * | 3ms + * + | * | 2ms + (midpoint) * + | | ** | 1ms + v *** + | zfs_delay_scale ----------> ******** | 0 +-------------------------------------*********----------------+ 0% <- zfs_dirty_data_max -> 100% .fi .sp Note that since the delay is added to the outstanding time remaining on the most recent transaction, the delay is effectively the inverse of IOPS. Here the midpoint of 500us translates to 2000 IOPS. The shape of the curve was chosen such that small changes in the amount of accumulated dirty data in the first 3/4 of the curve yield relatively small differences in the amount of delay. .sp The effects can be easier to understand when the amount of delay is represented on a log scale: .sp .nf delay 100ms +-------------------------------------------------------------++ + + | | + *+ 10ms + *+ + ** + | (midpoint) ** | + | ** + 1ms + v **** + + zfs_delay_scale ----------> ***** + | **** | + **** + 100us + ** + + * + | * | + * + 10us + * + + + | | + + +--------------------------------------------------------------+ 0% <- zfs_dirty_data_max -> 100% .fi .sp 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 backend storage, and then by changing the value of \fBzfs_delay_scale\fR to increase the steepness of the curve. diff --git a/module/zfs/arc.c b/module/zfs/arc.c index c3d88679f86c..b8707474d523 100644 --- a/module/zfs/arc.c +++ b/module/zfs/arc.c @@ -1,7097 +1,7157 @@ /* * 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) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2012, Joyent, Inc. All rights reserved. * Copyright (c) 2011, 2015 by Delphix. All rights reserved. * Copyright (c) 2014 by Saso Kiselkov. All rights reserved. * Copyright 2014 Nexenta Systems, Inc. All rights reserved. */ /* * DVA-based Adjustable Replacement Cache * * While much of the theory of operation used here is * based on the self-tuning, low overhead replacement cache * presented by Megiddo and Modha at FAST 2003, there are some * significant differences: * * 1. The Megiddo and Modha model assumes any page is evictable. * Pages in its cache cannot be "locked" into memory. This makes * the eviction algorithm simple: evict the last page in the list. * This also make the performance characteristics easy to reason * about. Our cache is not so simple. At any given moment, some * subset of the blocks in the cache are un-evictable because we * have handed out a reference to them. Blocks are only evictable * when there are no external references active. This makes * eviction far more problematic: we choose to evict the evictable * blocks that are the "lowest" in the list. * * There are times when it is not possible to evict the requested * space. In these circumstances we are unable to adjust the cache * size. To prevent the cache growing unbounded at these times we * implement a "cache throttle" that slows the flow of new data * into the cache until we can make space available. * * 2. The Megiddo and Modha model assumes a fixed cache size. * Pages are evicted when the cache is full and there is a cache * miss. Our model has a variable sized cache. It grows with * high use, but also tries to react to memory pressure from the * operating system: decreasing its size when system memory is * tight. * * 3. The Megiddo and Modha model assumes a fixed page size. All * elements of the cache are therefore exactly the same size. So * when adjusting the cache size following a cache miss, its simply * a matter of choosing a single page to evict. In our model, we * have variable sized cache blocks (rangeing from 512 bytes to * 128K bytes). We therefore choose a set of blocks to evict to make * space for a cache miss that approximates as closely as possible * the space used by the new block. * * See also: "ARC: A Self-Tuning, Low Overhead Replacement Cache" * by N. Megiddo & D. Modha, FAST 2003 */ /* * The locking model: * * A new reference to a cache buffer can be obtained in two * ways: 1) via a hash table lookup using the DVA as a key, * or 2) via one of the ARC lists. The arc_read() interface * uses method 1, while the internal arc algorithms for * adjusting the cache use method 2. We therefore provide two * types of locks: 1) the hash table lock array, and 2) the * arc list locks. * * Buffers do not have their own mutexes, rather they rely on the * hash table mutexes for the bulk of their protection (i.e. most * fields in the arc_buf_hdr_t are protected by these mutexes). * * buf_hash_find() returns the appropriate mutex (held) when it * locates the requested buffer in the hash table. It returns * NULL for the mutex if the buffer was not in the table. * * buf_hash_remove() expects the appropriate hash mutex to be * already held before it is invoked. * * Each arc state also has a mutex which is used to protect the * buffer list associated with the state. When attempting to * obtain a hash table lock while holding an arc list lock you * must use: mutex_tryenter() to avoid deadlock. Also note that * the active state mutex must be held before the ghost state mutex. * * Arc buffers may have an associated eviction callback function. * This function will be invoked prior to removing the buffer (e.g. * in arc_do_user_evicts()). Note however that the data associated * with the buffer may be evicted prior to the callback. The callback * must be made with *no locks held* (to prevent deadlock). Additionally, * the users of callbacks must ensure that their private data is * protected from simultaneous callbacks from arc_clear_callback() * and arc_do_user_evicts(). * * It as also possible to register a callback which is run when the * arc_meta_limit is reached and no buffers can be safely evicted. In * this case the arc user should drop a reference on some arc buffers so * they can be reclaimed and the arc_meta_limit honored. For example, * when using the ZPL each dentry holds a references on a znode. These * dentries must be pruned before the arc buffer holding the znode can * be safely evicted. * * Note that the majority of the performance stats are manipulated * with atomic operations. * * The L2ARC uses the l2ad_mtx on each vdev for the following: * * - L2ARC buflist creation * - L2ARC buflist eviction * - L2ARC write completion, which walks L2ARC buflists * - ARC header destruction, as it removes from L2ARC buflists * - ARC header release, as it removes from L2ARC buflists */ #include #include #include #include #include #include #include #include #include #include #ifdef _KERNEL #include #include #include #include #include #endif #include #include #include #include #include #include #ifndef _KERNEL /* set with ZFS_DEBUG=watch, to enable watchpoints on frozen buffers */ boolean_t arc_watch = B_FALSE; #endif static kmutex_t arc_reclaim_lock; static kcondvar_t arc_reclaim_thread_cv; static boolean_t arc_reclaim_thread_exit; static kcondvar_t arc_reclaim_waiters_cv; static kmutex_t arc_user_evicts_lock; static kcondvar_t arc_user_evicts_cv; static boolean_t arc_user_evicts_thread_exit; /* * The number of headers to evict in arc_evict_state_impl() before * dropping the sublist lock and evicting from another sublist. A lower * value means we're more likely to evict the "correct" header (i.e. the * oldest header in the arc state), but comes with higher overhead * (i.e. more invocations of arc_evict_state_impl()). */ int zfs_arc_evict_batch_limit = 10; /* * The number of sublists used for each of the arc state lists. If this * is not set to a suitable value by the user, it will be configured to * the number of CPUs on the system in arc_init(). */ int zfs_arc_num_sublists_per_state = 0; /* number of seconds before growing cache again */ static int arc_grow_retry = 5; /* shift of arc_c for calculating overflow limit in arc_get_data_buf */ int zfs_arc_overflow_shift = 8; /* shift of arc_c for calculating both min and max arc_p */ static int arc_p_min_shift = 4; /* log2(fraction of arc to reclaim) */ static int arc_shrink_shift = 7; /* * log2(fraction of ARC which must be free to allow growing). * I.e. If there is less than arc_c >> arc_no_grow_shift free memory, * when reading a new block into the ARC, we will evict an equal-sized block * from the ARC. * * This must be less than arc_shrink_shift, so that when we shrink the ARC, * we will still not allow it to grow. */ int arc_no_grow_shift = 5; /* * minimum lifespan of a prefetch block in clock ticks * (initialized in arc_init()) */ static int arc_min_prefetch_lifespan; /* * If this percent of memory is free, don't throttle. */ int arc_lotsfree_percent = 10; static int arc_dead; /* * The arc has filled available memory and has now warmed up. */ static boolean_t arc_warm; /* * These tunables are for performance analysis. */ unsigned long zfs_arc_max = 0; unsigned long zfs_arc_min = 0; unsigned long zfs_arc_meta_limit = 0; unsigned long zfs_arc_meta_min = 0; int zfs_arc_grow_retry = 0; int zfs_arc_shrink_shift = 0; int zfs_arc_p_min_shift = 0; int zfs_disable_dup_eviction = 0; int zfs_arc_average_blocksize = 8 * 1024; /* 8KB */ /* * These tunables are Linux specific */ unsigned long zfs_arc_sys_free = 0; int zfs_arc_min_prefetch_lifespan = 0; int zfs_arc_p_aggressive_disable = 1; int zfs_arc_p_dampener_disable = 1; int zfs_arc_meta_prune = 10000; int zfs_arc_meta_strategy = ARC_STRATEGY_META_BALANCED; int zfs_arc_meta_adjust_restarts = 4096; int zfs_arc_lotsfree_percent = 10; /* The 6 states: */ static arc_state_t ARC_anon; static arc_state_t ARC_mru; static arc_state_t ARC_mru_ghost; static arc_state_t ARC_mfu; static arc_state_t ARC_mfu_ghost; static arc_state_t ARC_l2c_only; typedef struct arc_stats { kstat_named_t arcstat_hits; kstat_named_t arcstat_misses; kstat_named_t arcstat_demand_data_hits; kstat_named_t arcstat_demand_data_misses; kstat_named_t arcstat_demand_metadata_hits; kstat_named_t arcstat_demand_metadata_misses; kstat_named_t arcstat_prefetch_data_hits; kstat_named_t arcstat_prefetch_data_misses; kstat_named_t arcstat_prefetch_metadata_hits; kstat_named_t arcstat_prefetch_metadata_misses; kstat_named_t arcstat_mru_hits; kstat_named_t arcstat_mru_ghost_hits; kstat_named_t arcstat_mfu_hits; kstat_named_t arcstat_mfu_ghost_hits; kstat_named_t arcstat_deleted; /* * Number of buffers that could not be evicted because the hash lock * was held by another thread. The lock may not necessarily be held * by something using the same buffer, since hash locks are shared * by multiple buffers. */ kstat_named_t arcstat_mutex_miss; /* * Number of buffers skipped because they have I/O in progress, are * indrect prefetch buffers that have not lived long enough, or are * not from the spa we're trying to evict from. */ kstat_named_t arcstat_evict_skip; /* * Number of times arc_evict_state() was unable to evict enough * buffers to reach its target amount. */ kstat_named_t arcstat_evict_not_enough; kstat_named_t arcstat_evict_l2_cached; kstat_named_t arcstat_evict_l2_eligible; kstat_named_t arcstat_evict_l2_ineligible; kstat_named_t arcstat_evict_l2_skip; kstat_named_t arcstat_hash_elements; kstat_named_t arcstat_hash_elements_max; kstat_named_t arcstat_hash_collisions; kstat_named_t arcstat_hash_chains; kstat_named_t arcstat_hash_chain_max; kstat_named_t arcstat_p; kstat_named_t arcstat_c; kstat_named_t arcstat_c_min; kstat_named_t arcstat_c_max; kstat_named_t arcstat_size; /* * Number of bytes consumed by internal ARC structures necessary * for tracking purposes; these structures are not actually * backed by ARC buffers. This includes arc_buf_hdr_t structures * (allocated via arc_buf_hdr_t_full and arc_buf_hdr_t_l2only * caches), and arc_buf_t structures (allocated via arc_buf_t * cache). */ kstat_named_t arcstat_hdr_size; /* * Number of bytes consumed by ARC buffers of type equal to * ARC_BUFC_DATA. This is generally consumed by buffers backing * on disk user data (e.g. plain file contents). */ kstat_named_t arcstat_data_size; /* * Number of bytes consumed by ARC buffers of type equal to * ARC_BUFC_METADATA. This is generally consumed by buffers * backing on disk data that is used for internal ZFS * structures (e.g. ZAP, dnode, indirect blocks, etc). */ kstat_named_t arcstat_metadata_size; /* * Number of bytes consumed by various buffers and structures * not actually backed with ARC buffers. This includes bonus * buffers (allocated directly via zio_buf_* functions), * dmu_buf_impl_t structures (allocated via dmu_buf_impl_t * cache), and dnode_t structures (allocated via dnode_t cache). */ kstat_named_t arcstat_other_size; /* * Total number of bytes consumed by ARC buffers residing in the * arc_anon state. This includes *all* buffers in the arc_anon * state; e.g. data, metadata, evictable, and unevictable buffers * are all included in this value. */ kstat_named_t arcstat_anon_size; /* * Number of bytes consumed by ARC buffers that meet the * following criteria: backing buffers of type ARC_BUFC_DATA, * residing in the arc_anon state, and are eligible for eviction * (e.g. have no outstanding holds on the buffer). */ kstat_named_t arcstat_anon_evictable_data; /* * Number of bytes consumed by ARC buffers that meet the * following criteria: backing buffers of type ARC_BUFC_METADATA, * residing in the arc_anon state, and are eligible for eviction * (e.g. have no outstanding holds on the buffer). */ kstat_named_t arcstat_anon_evictable_metadata; /* * Total number of bytes consumed by ARC buffers residing in the * arc_mru state. This includes *all* buffers in the arc_mru * state; e.g. data, metadata, evictable, and unevictable buffers * are all included in this value. */ kstat_named_t arcstat_mru_size; /* * Number of bytes consumed by ARC buffers that meet the * following criteria: backing buffers of type ARC_BUFC_DATA, * residing in the arc_mru state, and are eligible for eviction * (e.g. have no outstanding holds on the buffer). */ kstat_named_t arcstat_mru_evictable_data; /* * Number of bytes consumed by ARC buffers that meet the * following criteria: backing buffers of type ARC_BUFC_METADATA, * residing in the arc_mru state, and are eligible for eviction * (e.g. have no outstanding holds on the buffer). */ kstat_named_t arcstat_mru_evictable_metadata; /* * Total number of bytes that *would have been* consumed by ARC * buffers in the arc_mru_ghost state. The key thing to note * here, is the fact that this size doesn't actually indicate * RAM consumption. The ghost lists only consist of headers and * don't actually have ARC buffers linked off of these headers. * Thus, *if* the headers had associated ARC buffers, these * buffers *would have* consumed this number of bytes. */ kstat_named_t arcstat_mru_ghost_size; /* * Number of bytes that *would have been* consumed by ARC * buffers that are eligible for eviction, of type * ARC_BUFC_DATA, and linked off the arc_mru_ghost state. */ kstat_named_t arcstat_mru_ghost_evictable_data; /* * Number of bytes that *would have been* consumed by ARC * buffers that are eligible for eviction, of type * ARC_BUFC_METADATA, and linked off the arc_mru_ghost state. */ kstat_named_t arcstat_mru_ghost_evictable_metadata; /* * Total number of bytes consumed by ARC buffers residing in the * arc_mfu state. This includes *all* buffers in the arc_mfu * state; e.g. data, metadata, evictable, and unevictable buffers * are all included in this value. */ kstat_named_t arcstat_mfu_size; /* * Number of bytes consumed by ARC buffers that are eligible for * eviction, of type ARC_BUFC_DATA, and reside in the arc_mfu * state. */ kstat_named_t arcstat_mfu_evictable_data; /* * Number of bytes consumed by ARC buffers that are eligible for * eviction, of type ARC_BUFC_METADATA, and reside in the * arc_mfu state. */ kstat_named_t arcstat_mfu_evictable_metadata; /* * Total number of bytes that *would have been* consumed by ARC * buffers in the arc_mfu_ghost state. See the comment above * arcstat_mru_ghost_size for more details. */ kstat_named_t arcstat_mfu_ghost_size; /* * Number of bytes that *would have been* consumed by ARC * buffers that are eligible for eviction, of type * ARC_BUFC_DATA, and linked off the arc_mfu_ghost state. */ kstat_named_t arcstat_mfu_ghost_evictable_data; /* * Number of bytes that *would have been* consumed by ARC * buffers that are eligible for eviction, of type * ARC_BUFC_METADATA, and linked off the arc_mru_ghost state. */ kstat_named_t arcstat_mfu_ghost_evictable_metadata; kstat_named_t arcstat_l2_hits; kstat_named_t arcstat_l2_misses; kstat_named_t arcstat_l2_feeds; kstat_named_t arcstat_l2_rw_clash; kstat_named_t arcstat_l2_read_bytes; kstat_named_t arcstat_l2_write_bytes; kstat_named_t arcstat_l2_writes_sent; kstat_named_t arcstat_l2_writes_done; kstat_named_t arcstat_l2_writes_error; kstat_named_t arcstat_l2_writes_lock_retry; kstat_named_t arcstat_l2_evict_lock_retry; kstat_named_t arcstat_l2_evict_reading; kstat_named_t arcstat_l2_evict_l1cached; kstat_named_t arcstat_l2_free_on_write; kstat_named_t arcstat_l2_cdata_free_on_write; kstat_named_t arcstat_l2_abort_lowmem; kstat_named_t arcstat_l2_cksum_bad; kstat_named_t arcstat_l2_io_error; kstat_named_t arcstat_l2_size; kstat_named_t arcstat_l2_asize; kstat_named_t arcstat_l2_hdr_size; kstat_named_t arcstat_l2_compress_successes; kstat_named_t arcstat_l2_compress_zeros; kstat_named_t arcstat_l2_compress_failures; kstat_named_t arcstat_memory_throttle_count; kstat_named_t arcstat_duplicate_buffers; kstat_named_t arcstat_duplicate_buffers_size; kstat_named_t arcstat_duplicate_reads; kstat_named_t arcstat_memory_direct_count; kstat_named_t arcstat_memory_indirect_count; kstat_named_t arcstat_no_grow; kstat_named_t arcstat_tempreserve; kstat_named_t arcstat_loaned_bytes; kstat_named_t arcstat_prune; kstat_named_t arcstat_meta_used; kstat_named_t arcstat_meta_limit; kstat_named_t arcstat_meta_max; kstat_named_t arcstat_meta_min; + kstat_named_t arcstat_sync_wait_for_async; + kstat_named_t arcstat_demand_hit_predictive_prefetch; kstat_named_t arcstat_need_free; kstat_named_t arcstat_sys_free; } arc_stats_t; static arc_stats_t arc_stats = { { "hits", KSTAT_DATA_UINT64 }, { "misses", KSTAT_DATA_UINT64 }, { "demand_data_hits", KSTAT_DATA_UINT64 }, { "demand_data_misses", KSTAT_DATA_UINT64 }, { "demand_metadata_hits", KSTAT_DATA_UINT64 }, { "demand_metadata_misses", KSTAT_DATA_UINT64 }, { "prefetch_data_hits", KSTAT_DATA_UINT64 }, { "prefetch_data_misses", KSTAT_DATA_UINT64 }, { "prefetch_metadata_hits", KSTAT_DATA_UINT64 }, { "prefetch_metadata_misses", KSTAT_DATA_UINT64 }, { "mru_hits", KSTAT_DATA_UINT64 }, { "mru_ghost_hits", KSTAT_DATA_UINT64 }, { "mfu_hits", KSTAT_DATA_UINT64 }, { "mfu_ghost_hits", KSTAT_DATA_UINT64 }, { "deleted", KSTAT_DATA_UINT64 }, { "mutex_miss", KSTAT_DATA_UINT64 }, { "evict_skip", KSTAT_DATA_UINT64 }, { "evict_not_enough", KSTAT_DATA_UINT64 }, { "evict_l2_cached", KSTAT_DATA_UINT64 }, { "evict_l2_eligible", KSTAT_DATA_UINT64 }, { "evict_l2_ineligible", KSTAT_DATA_UINT64 }, { "evict_l2_skip", KSTAT_DATA_UINT64 }, { "hash_elements", KSTAT_DATA_UINT64 }, { "hash_elements_max", KSTAT_DATA_UINT64 }, { "hash_collisions", KSTAT_DATA_UINT64 }, { "hash_chains", KSTAT_DATA_UINT64 }, { "hash_chain_max", KSTAT_DATA_UINT64 }, { "p", KSTAT_DATA_UINT64 }, { "c", KSTAT_DATA_UINT64 }, { "c_min", KSTAT_DATA_UINT64 }, { "c_max", KSTAT_DATA_UINT64 }, { "size", KSTAT_DATA_UINT64 }, { "hdr_size", KSTAT_DATA_UINT64 }, { "data_size", KSTAT_DATA_UINT64 }, { "metadata_size", KSTAT_DATA_UINT64 }, { "other_size", KSTAT_DATA_UINT64 }, { "anon_size", KSTAT_DATA_UINT64 }, { "anon_evictable_data", KSTAT_DATA_UINT64 }, { "anon_evictable_metadata", KSTAT_DATA_UINT64 }, { "mru_size", KSTAT_DATA_UINT64 }, { "mru_evictable_data", KSTAT_DATA_UINT64 }, { "mru_evictable_metadata", KSTAT_DATA_UINT64 }, { "mru_ghost_size", KSTAT_DATA_UINT64 }, { "mru_ghost_evictable_data", KSTAT_DATA_UINT64 }, { "mru_ghost_evictable_metadata", KSTAT_DATA_UINT64 }, { "mfu_size", KSTAT_DATA_UINT64 }, { "mfu_evictable_data", KSTAT_DATA_UINT64 }, { "mfu_evictable_metadata", KSTAT_DATA_UINT64 }, { "mfu_ghost_size", KSTAT_DATA_UINT64 }, { "mfu_ghost_evictable_data", KSTAT_DATA_UINT64 }, { "mfu_ghost_evictable_metadata", KSTAT_DATA_UINT64 }, { "l2_hits", KSTAT_DATA_UINT64 }, { "l2_misses", KSTAT_DATA_UINT64 }, { "l2_feeds", KSTAT_DATA_UINT64 }, { "l2_rw_clash", KSTAT_DATA_UINT64 }, { "l2_read_bytes", KSTAT_DATA_UINT64 }, { "l2_write_bytes", KSTAT_DATA_UINT64 }, { "l2_writes_sent", KSTAT_DATA_UINT64 }, { "l2_writes_done", KSTAT_DATA_UINT64 }, { "l2_writes_error", KSTAT_DATA_UINT64 }, { "l2_writes_lock_retry", KSTAT_DATA_UINT64 }, { "l2_evict_lock_retry", KSTAT_DATA_UINT64 }, { "l2_evict_reading", KSTAT_DATA_UINT64 }, { "l2_evict_l1cached", KSTAT_DATA_UINT64 }, { "l2_free_on_write", KSTAT_DATA_UINT64 }, { "l2_cdata_free_on_write", KSTAT_DATA_UINT64 }, { "l2_abort_lowmem", KSTAT_DATA_UINT64 }, { "l2_cksum_bad", KSTAT_DATA_UINT64 }, { "l2_io_error", KSTAT_DATA_UINT64 }, { "l2_size", KSTAT_DATA_UINT64 }, { "l2_asize", KSTAT_DATA_UINT64 }, { "l2_hdr_size", KSTAT_DATA_UINT64 }, { "l2_compress_successes", KSTAT_DATA_UINT64 }, { "l2_compress_zeros", KSTAT_DATA_UINT64 }, { "l2_compress_failures", KSTAT_DATA_UINT64 }, { "memory_throttle_count", KSTAT_DATA_UINT64 }, { "duplicate_buffers", KSTAT_DATA_UINT64 }, { "duplicate_buffers_size", KSTAT_DATA_UINT64 }, { "duplicate_reads", KSTAT_DATA_UINT64 }, { "memory_direct_count", KSTAT_DATA_UINT64 }, { "memory_indirect_count", KSTAT_DATA_UINT64 }, { "arc_no_grow", KSTAT_DATA_UINT64 }, { "arc_tempreserve", KSTAT_DATA_UINT64 }, { "arc_loaned_bytes", KSTAT_DATA_UINT64 }, { "arc_prune", KSTAT_DATA_UINT64 }, { "arc_meta_used", KSTAT_DATA_UINT64 }, { "arc_meta_limit", KSTAT_DATA_UINT64 }, { "arc_meta_max", KSTAT_DATA_UINT64 }, { "arc_meta_min", KSTAT_DATA_UINT64 }, + { "sync_wait_for_async", KSTAT_DATA_UINT64 }, + { "demand_hit_predictive_prefetch", KSTAT_DATA_UINT64 }, { "arc_need_free", KSTAT_DATA_UINT64 }, { "arc_sys_free", KSTAT_DATA_UINT64 } }; #define ARCSTAT(stat) (arc_stats.stat.value.ui64) #define ARCSTAT_INCR(stat, val) \ atomic_add_64(&arc_stats.stat.value.ui64, (val)) #define ARCSTAT_BUMP(stat) ARCSTAT_INCR(stat, 1) #define ARCSTAT_BUMPDOWN(stat) ARCSTAT_INCR(stat, -1) #define ARCSTAT_MAX(stat, val) { \ uint64_t m; \ while ((val) > (m = arc_stats.stat.value.ui64) && \ (m != atomic_cas_64(&arc_stats.stat.value.ui64, m, (val)))) \ continue; \ } #define ARCSTAT_MAXSTAT(stat) \ ARCSTAT_MAX(stat##_max, arc_stats.stat.value.ui64) /* * We define a macro to allow ARC hits/misses to be easily broken down by * two separate conditions, giving a total of four different subtypes for * each of hits and misses (so eight statistics total). */ #define ARCSTAT_CONDSTAT(cond1, stat1, notstat1, cond2, stat2, notstat2, stat) \ if (cond1) { \ if (cond2) { \ ARCSTAT_BUMP(arcstat_##stat1##_##stat2##_##stat); \ } else { \ ARCSTAT_BUMP(arcstat_##stat1##_##notstat2##_##stat); \ } \ } else { \ if (cond2) { \ ARCSTAT_BUMP(arcstat_##notstat1##_##stat2##_##stat); \ } else { \ ARCSTAT_BUMP(arcstat_##notstat1##_##notstat2##_##stat);\ } \ } kstat_t *arc_ksp; static arc_state_t *arc_anon; static arc_state_t *arc_mru; static arc_state_t *arc_mru_ghost; static arc_state_t *arc_mfu; static arc_state_t *arc_mfu_ghost; static arc_state_t *arc_l2c_only; /* * There are several ARC variables that are critical to export as kstats -- * but we don't want to have to grovel around in the kstat whenever we wish to * manipulate them. For these variables, we therefore define them to be in * terms of the statistic variable. This assures that we are not introducing * the possibility of inconsistency by having shadow copies of the variables, * while still allowing the code to be readable. */ #define arc_size ARCSTAT(arcstat_size) /* actual total arc size */ #define arc_p ARCSTAT(arcstat_p) /* target size of MRU */ #define arc_c ARCSTAT(arcstat_c) /* target size of cache */ #define arc_c_min ARCSTAT(arcstat_c_min) /* min target cache size */ #define arc_c_max ARCSTAT(arcstat_c_max) /* max target cache size */ #define arc_no_grow ARCSTAT(arcstat_no_grow) #define arc_tempreserve ARCSTAT(arcstat_tempreserve) #define arc_loaned_bytes ARCSTAT(arcstat_loaned_bytes) #define arc_meta_limit ARCSTAT(arcstat_meta_limit) /* max size for metadata */ #define arc_meta_min ARCSTAT(arcstat_meta_min) /* min size for metadata */ #define arc_meta_used ARCSTAT(arcstat_meta_used) /* size of metadata */ #define arc_meta_max ARCSTAT(arcstat_meta_max) /* max size of metadata */ #define arc_need_free ARCSTAT(arcstat_need_free) /* bytes to be freed */ #define arc_sys_free ARCSTAT(arcstat_sys_free) /* target system free bytes */ #define L2ARC_IS_VALID_COMPRESS(_c_) \ ((_c_) == ZIO_COMPRESS_LZ4 || (_c_) == ZIO_COMPRESS_EMPTY) static list_t arc_prune_list; static kmutex_t arc_prune_mtx; static taskq_t *arc_prune_taskq; static arc_buf_t *arc_eviction_list; static arc_buf_hdr_t arc_eviction_hdr; #define GHOST_STATE(state) \ ((state) == arc_mru_ghost || (state) == arc_mfu_ghost || \ (state) == arc_l2c_only) #define HDR_IN_HASH_TABLE(hdr) ((hdr)->b_flags & ARC_FLAG_IN_HASH_TABLE) #define HDR_IO_IN_PROGRESS(hdr) ((hdr)->b_flags & ARC_FLAG_IO_IN_PROGRESS) #define HDR_IO_ERROR(hdr) ((hdr)->b_flags & ARC_FLAG_IO_ERROR) #define HDR_PREFETCH(hdr) ((hdr)->b_flags & ARC_FLAG_PREFETCH) #define HDR_FREED_IN_READ(hdr) ((hdr)->b_flags & ARC_FLAG_FREED_IN_READ) #define HDR_BUF_AVAILABLE(hdr) ((hdr)->b_flags & ARC_FLAG_BUF_AVAILABLE) #define HDR_L2CACHE(hdr) ((hdr)->b_flags & ARC_FLAG_L2CACHE) #define HDR_L2COMPRESS(hdr) ((hdr)->b_flags & ARC_FLAG_L2COMPRESS) #define HDR_L2_READING(hdr) \ (((hdr)->b_flags & ARC_FLAG_IO_IN_PROGRESS) && \ ((hdr)->b_flags & ARC_FLAG_HAS_L2HDR)) #define HDR_L2_WRITING(hdr) ((hdr)->b_flags & ARC_FLAG_L2_WRITING) #define HDR_L2_EVICTED(hdr) ((hdr)->b_flags & ARC_FLAG_L2_EVICTED) #define HDR_L2_WRITE_HEAD(hdr) ((hdr)->b_flags & ARC_FLAG_L2_WRITE_HEAD) #define HDR_ISTYPE_METADATA(hdr) \ ((hdr)->b_flags & ARC_FLAG_BUFC_METADATA) #define HDR_ISTYPE_DATA(hdr) (!HDR_ISTYPE_METADATA(hdr)) #define HDR_HAS_L1HDR(hdr) ((hdr)->b_flags & ARC_FLAG_HAS_L1HDR) #define HDR_HAS_L2HDR(hdr) ((hdr)->b_flags & ARC_FLAG_HAS_L2HDR) /* * Other sizes */ #define HDR_FULL_SIZE ((int64_t)sizeof (arc_buf_hdr_t)) #define HDR_L2ONLY_SIZE ((int64_t)offsetof(arc_buf_hdr_t, b_l1hdr)) /* * Hash table routines */ #define HT_LOCK_ALIGN 64 #define HT_LOCK_PAD (P2NPHASE(sizeof (kmutex_t), (HT_LOCK_ALIGN))) struct ht_lock { kmutex_t ht_lock; #ifdef _KERNEL unsigned char pad[HT_LOCK_PAD]; #endif }; #define BUF_LOCKS 8192 typedef struct buf_hash_table { uint64_t ht_mask; arc_buf_hdr_t **ht_table; struct ht_lock ht_locks[BUF_LOCKS]; } buf_hash_table_t; static buf_hash_table_t buf_hash_table; #define BUF_HASH_INDEX(spa, dva, birth) \ (buf_hash(spa, dva, birth) & buf_hash_table.ht_mask) #define BUF_HASH_LOCK_NTRY(idx) (buf_hash_table.ht_locks[idx & (BUF_LOCKS-1)]) #define BUF_HASH_LOCK(idx) (&(BUF_HASH_LOCK_NTRY(idx).ht_lock)) #define HDR_LOCK(hdr) \ (BUF_HASH_LOCK(BUF_HASH_INDEX(hdr->b_spa, &hdr->b_dva, hdr->b_birth))) uint64_t zfs_crc64_table[256]; /* * Level 2 ARC */ #define L2ARC_WRITE_SIZE (8 * 1024 * 1024) /* initial write max */ #define L2ARC_HEADROOM 2 /* num of writes */ /* * If we discover during ARC scan any buffers to be compressed, we boost * our headroom for the next scanning cycle by this percentage multiple. */ #define L2ARC_HEADROOM_BOOST 200 #define L2ARC_FEED_SECS 1 /* caching interval secs */ #define L2ARC_FEED_MIN_MS 200 /* min caching interval ms */ /* * Used to distinguish headers that are being process by * l2arc_write_buffers(), but have yet to be assigned to a l2arc disk * address. This can happen when the header is added to the l2arc's list * of buffers to write in the first stage of l2arc_write_buffers(), but * has not yet been written out which happens in the second stage of * l2arc_write_buffers(). */ #define L2ARC_ADDR_UNSET ((uint64_t)(-1)) #define l2arc_writes_sent ARCSTAT(arcstat_l2_writes_sent) #define l2arc_writes_done ARCSTAT(arcstat_l2_writes_done) /* L2ARC Performance Tunables */ unsigned long l2arc_write_max = L2ARC_WRITE_SIZE; /* def max write size */ unsigned long l2arc_write_boost = L2ARC_WRITE_SIZE; /* extra warmup write */ unsigned long l2arc_headroom = L2ARC_HEADROOM; /* # of dev writes */ unsigned long l2arc_headroom_boost = L2ARC_HEADROOM_BOOST; unsigned long l2arc_feed_secs = L2ARC_FEED_SECS; /* interval seconds */ unsigned long l2arc_feed_min_ms = L2ARC_FEED_MIN_MS; /* min interval msecs */ int l2arc_noprefetch = B_TRUE; /* don't cache prefetch bufs */ int l2arc_nocompress = B_FALSE; /* don't compress bufs */ int l2arc_feed_again = B_TRUE; /* turbo warmup */ int l2arc_norw = B_FALSE; /* no reads during writes */ /* * L2ARC Internals */ static list_t L2ARC_dev_list; /* device list */ static list_t *l2arc_dev_list; /* device list pointer */ static kmutex_t l2arc_dev_mtx; /* device list mutex */ static l2arc_dev_t *l2arc_dev_last; /* last device used */ static list_t L2ARC_free_on_write; /* free after write buf list */ static list_t *l2arc_free_on_write; /* free after write list ptr */ static kmutex_t l2arc_free_on_write_mtx; /* mutex for list */ static uint64_t l2arc_ndev; /* number of devices */ typedef struct l2arc_read_callback { arc_buf_t *l2rcb_buf; /* read buffer */ spa_t *l2rcb_spa; /* spa */ blkptr_t l2rcb_bp; /* original blkptr */ zbookmark_phys_t l2rcb_zb; /* original bookmark */ int l2rcb_flags; /* original flags */ enum zio_compress l2rcb_compress; /* applied compress */ } l2arc_read_callback_t; typedef struct l2arc_data_free { /* protected by l2arc_free_on_write_mtx */ void *l2df_data; size_t l2df_size; void (*l2df_func)(void *, size_t); list_node_t l2df_list_node; } l2arc_data_free_t; static kmutex_t l2arc_feed_thr_lock; static kcondvar_t l2arc_feed_thr_cv; static uint8_t l2arc_thread_exit; static void arc_get_data_buf(arc_buf_t *); static void arc_access(arc_buf_hdr_t *, kmutex_t *); static boolean_t arc_is_overflowing(void); static void arc_buf_watch(arc_buf_t *); static void arc_tuning_update(void); static arc_buf_contents_t arc_buf_type(arc_buf_hdr_t *); static uint32_t arc_bufc_to_flags(arc_buf_contents_t); static boolean_t l2arc_write_eligible(uint64_t, arc_buf_hdr_t *); static void l2arc_read_done(zio_t *); static boolean_t l2arc_compress_buf(arc_buf_hdr_t *); static void l2arc_decompress_zio(zio_t *, arc_buf_hdr_t *, enum zio_compress); static void l2arc_release_cdata_buf(arc_buf_hdr_t *); static uint64_t buf_hash(uint64_t spa, const dva_t *dva, uint64_t birth) { uint8_t *vdva = (uint8_t *)dva; uint64_t crc = -1ULL; int i; ASSERT(zfs_crc64_table[128] == ZFS_CRC64_POLY); for (i = 0; i < sizeof (dva_t); i++) crc = (crc >> 8) ^ zfs_crc64_table[(crc ^ vdva[i]) & 0xFF]; crc ^= (spa>>8) ^ birth; return (crc); } #define BUF_EMPTY(buf) \ ((buf)->b_dva.dva_word[0] == 0 && \ (buf)->b_dva.dva_word[1] == 0) #define BUF_EQUAL(spa, dva, birth, buf) \ ((buf)->b_dva.dva_word[0] == (dva)->dva_word[0]) && \ ((buf)->b_dva.dva_word[1] == (dva)->dva_word[1]) && \ ((buf)->b_birth == birth) && ((buf)->b_spa == spa) static void buf_discard_identity(arc_buf_hdr_t *hdr) { hdr->b_dva.dva_word[0] = 0; hdr->b_dva.dva_word[1] = 0; hdr->b_birth = 0; } static arc_buf_hdr_t * buf_hash_find(uint64_t spa, const blkptr_t *bp, kmutex_t **lockp) { const dva_t *dva = BP_IDENTITY(bp); uint64_t birth = BP_PHYSICAL_BIRTH(bp); uint64_t idx = BUF_HASH_INDEX(spa, dva, birth); kmutex_t *hash_lock = BUF_HASH_LOCK(idx); arc_buf_hdr_t *hdr; mutex_enter(hash_lock); for (hdr = buf_hash_table.ht_table[idx]; hdr != NULL; hdr = hdr->b_hash_next) { if (BUF_EQUAL(spa, dva, birth, hdr)) { *lockp = hash_lock; return (hdr); } } mutex_exit(hash_lock); *lockp = NULL; return (NULL); } /* * 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. * If lockp == NULL, the caller is assumed to already hold the hash lock. */ static arc_buf_hdr_t * buf_hash_insert(arc_buf_hdr_t *hdr, kmutex_t **lockp) { uint64_t idx = BUF_HASH_INDEX(hdr->b_spa, &hdr->b_dva, hdr->b_birth); kmutex_t *hash_lock = BUF_HASH_LOCK(idx); arc_buf_hdr_t *fhdr; uint32_t i; ASSERT(!DVA_IS_EMPTY(&hdr->b_dva)); ASSERT(hdr->b_birth != 0); ASSERT(!HDR_IN_HASH_TABLE(hdr)); if (lockp != NULL) { *lockp = hash_lock; mutex_enter(hash_lock); } else { ASSERT(MUTEX_HELD(hash_lock)); } for (fhdr = buf_hash_table.ht_table[idx], i = 0; fhdr != NULL; fhdr = fhdr->b_hash_next, i++) { if (BUF_EQUAL(hdr->b_spa, &hdr->b_dva, hdr->b_birth, fhdr)) return (fhdr); } hdr->b_hash_next = buf_hash_table.ht_table[idx]; buf_hash_table.ht_table[idx] = hdr; hdr->b_flags |= ARC_FLAG_IN_HASH_TABLE; /* collect some hash table performance data */ if (i > 0) { ARCSTAT_BUMP(arcstat_hash_collisions); if (i == 1) ARCSTAT_BUMP(arcstat_hash_chains); ARCSTAT_MAX(arcstat_hash_chain_max, i); } ARCSTAT_BUMP(arcstat_hash_elements); ARCSTAT_MAXSTAT(arcstat_hash_elements); return (NULL); } static void buf_hash_remove(arc_buf_hdr_t *hdr) { arc_buf_hdr_t *fhdr, **hdrp; uint64_t idx = BUF_HASH_INDEX(hdr->b_spa, &hdr->b_dva, hdr->b_birth); ASSERT(MUTEX_HELD(BUF_HASH_LOCK(idx))); ASSERT(HDR_IN_HASH_TABLE(hdr)); hdrp = &buf_hash_table.ht_table[idx]; while ((fhdr = *hdrp) != hdr) { ASSERT(fhdr != NULL); hdrp = &fhdr->b_hash_next; } *hdrp = hdr->b_hash_next; hdr->b_hash_next = NULL; hdr->b_flags &= ~ARC_FLAG_IN_HASH_TABLE; /* collect some hash table performance data */ ARCSTAT_BUMPDOWN(arcstat_hash_elements); if (buf_hash_table.ht_table[idx] && buf_hash_table.ht_table[idx]->b_hash_next == NULL) ARCSTAT_BUMPDOWN(arcstat_hash_chains); } /* * Global data structures and functions for the buf kmem cache. */ static kmem_cache_t *hdr_full_cache; static kmem_cache_t *hdr_l2only_cache; static kmem_cache_t *buf_cache; static void buf_fini(void) { int i; #if defined(_KERNEL) && defined(HAVE_SPL) /* * Large allocations which do not require contiguous pages * should be using vmem_free() in the linux kernel\ */ vmem_free(buf_hash_table.ht_table, (buf_hash_table.ht_mask + 1) * sizeof (void *)); #else kmem_free(buf_hash_table.ht_table, (buf_hash_table.ht_mask + 1) * sizeof (void *)); #endif for (i = 0; i < BUF_LOCKS; i++) mutex_destroy(&buf_hash_table.ht_locks[i].ht_lock); kmem_cache_destroy(hdr_full_cache); kmem_cache_destroy(hdr_l2only_cache); kmem_cache_destroy(buf_cache); } /* * Constructor callback - called when the cache is empty * and a new buf is requested. */ /* ARGSUSED */ static int hdr_full_cons(void *vbuf, void *unused, int kmflag) { arc_buf_hdr_t *hdr = vbuf; bzero(hdr, HDR_FULL_SIZE); cv_init(&hdr->b_l1hdr.b_cv, NULL, CV_DEFAULT, NULL); refcount_create(&hdr->b_l1hdr.b_refcnt); mutex_init(&hdr->b_l1hdr.b_freeze_lock, NULL, MUTEX_DEFAULT, NULL); list_link_init(&hdr->b_l1hdr.b_arc_node); list_link_init(&hdr->b_l2hdr.b_l2node); multilist_link_init(&hdr->b_l1hdr.b_arc_node); arc_space_consume(HDR_FULL_SIZE, ARC_SPACE_HDRS); return (0); } /* ARGSUSED */ static int hdr_l2only_cons(void *vbuf, void *unused, int kmflag) { arc_buf_hdr_t *hdr = vbuf; bzero(hdr, HDR_L2ONLY_SIZE); arc_space_consume(HDR_L2ONLY_SIZE, ARC_SPACE_L2HDRS); return (0); } /* ARGSUSED */ static int buf_cons(void *vbuf, void *unused, int kmflag) { arc_buf_t *buf = vbuf; bzero(buf, sizeof (arc_buf_t)); mutex_init(&buf->b_evict_lock, NULL, MUTEX_DEFAULT, NULL); arc_space_consume(sizeof (arc_buf_t), ARC_SPACE_HDRS); return (0); } /* * Destructor callback - called when a cached buf is * no longer required. */ /* ARGSUSED */ static void hdr_full_dest(void *vbuf, void *unused) { arc_buf_hdr_t *hdr = vbuf; ASSERT(BUF_EMPTY(hdr)); cv_destroy(&hdr->b_l1hdr.b_cv); refcount_destroy(&hdr->b_l1hdr.b_refcnt); mutex_destroy(&hdr->b_l1hdr.b_freeze_lock); ASSERT(!multilist_link_active(&hdr->b_l1hdr.b_arc_node)); arc_space_return(HDR_FULL_SIZE, ARC_SPACE_HDRS); } /* ARGSUSED */ static void hdr_l2only_dest(void *vbuf, void *unused) { ASSERTV(arc_buf_hdr_t *hdr = vbuf); ASSERT(BUF_EMPTY(hdr)); arc_space_return(HDR_L2ONLY_SIZE, ARC_SPACE_L2HDRS); } /* ARGSUSED */ static void buf_dest(void *vbuf, void *unused) { arc_buf_t *buf = vbuf; mutex_destroy(&buf->b_evict_lock); arc_space_return(sizeof (arc_buf_t), ARC_SPACE_HDRS); } /* * Reclaim callback -- invoked when memory is low. */ /* ARGSUSED */ static void hdr_recl(void *unused) { dprintf("hdr_recl called\n"); /* * umem calls the reclaim func when we destroy the buf cache, * which is after we do arc_fini(). */ if (!arc_dead) cv_signal(&arc_reclaim_thread_cv); } static void buf_init(void) { uint64_t *ct; uint64_t hsize = 1ULL << 12; int i, j; /* * The hash table is big enough to fill all 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 < physmem * PAGESIZE) hsize <<= 1; retry: buf_hash_table.ht_mask = hsize - 1; #if defined(_KERNEL) && defined(HAVE_SPL) /* * Large allocations which do not require contiguous pages * should be using vmem_alloc() in the linux kernel */ buf_hash_table.ht_table = vmem_zalloc(hsize * sizeof (void*), KM_SLEEP); #else buf_hash_table.ht_table = kmem_zalloc(hsize * sizeof (void*), KM_NOSLEEP); #endif if (buf_hash_table.ht_table == NULL) { ASSERT(hsize > (1ULL << 8)); hsize >>= 1; goto retry; } hdr_full_cache = kmem_cache_create("arc_buf_hdr_t_full", HDR_FULL_SIZE, 0, hdr_full_cons, hdr_full_dest, hdr_recl, NULL, NULL, 0); hdr_l2only_cache = kmem_cache_create("arc_buf_hdr_t_l2only", HDR_L2ONLY_SIZE, 0, hdr_l2only_cons, hdr_l2only_dest, hdr_recl, NULL, NULL, 0); buf_cache = kmem_cache_create("arc_buf_t", sizeof (arc_buf_t), 0, buf_cons, buf_dest, NULL, NULL, NULL, 0); for (i = 0; i < 256; i++) for (ct = zfs_crc64_table + i, *ct = i, j = 8; j > 0; j--) *ct = (*ct >> 1) ^ (-(*ct & 1) & ZFS_CRC64_POLY); for (i = 0; i < BUF_LOCKS; i++) { mutex_init(&buf_hash_table.ht_locks[i].ht_lock, NULL, MUTEX_DEFAULT, NULL); } } /* * Transition between the two allocation states for the arc_buf_hdr struct. * The arc_buf_hdr struct can be allocated with (hdr_full_cache) or without * (hdr_l2only_cache) the fields necessary for the L1 cache - the smaller * version is used when a cache buffer is only in the L2ARC in order to reduce * memory usage. */ static arc_buf_hdr_t * arc_hdr_realloc(arc_buf_hdr_t *hdr, kmem_cache_t *old, kmem_cache_t *new) { arc_buf_hdr_t *nhdr; l2arc_dev_t *dev; ASSERT(HDR_HAS_L2HDR(hdr)); ASSERT((old == hdr_full_cache && new == hdr_l2only_cache) || (old == hdr_l2only_cache && new == hdr_full_cache)); dev = hdr->b_l2hdr.b_dev; nhdr = kmem_cache_alloc(new, KM_PUSHPAGE); ASSERT(MUTEX_HELD(HDR_LOCK(hdr))); buf_hash_remove(hdr); bcopy(hdr, nhdr, HDR_L2ONLY_SIZE); if (new == hdr_full_cache) { nhdr->b_flags |= ARC_FLAG_HAS_L1HDR; /* * arc_access and arc_change_state need to be aware that a * header has just come out of L2ARC, so we set its state to * l2c_only even though it's about to change. */ nhdr->b_l1hdr.b_state = arc_l2c_only; /* Verify previous threads set to NULL before freeing */ ASSERT3P(nhdr->b_l1hdr.b_tmp_cdata, ==, NULL); } else { ASSERT(hdr->b_l1hdr.b_buf == NULL); ASSERT0(hdr->b_l1hdr.b_datacnt); /* * If we've reached here, We must have been called from * arc_evict_hdr(), as such we should have already been * removed from any ghost list we were previously on * (which protects us from racing with arc_evict_state), * thus no locking is needed during this check. */ ASSERT(!multilist_link_active(&hdr->b_l1hdr.b_arc_node)); /* * A buffer must not be moved into the arc_l2c_only * state if it's not finished being written out to the * l2arc device. Otherwise, the b_l1hdr.b_tmp_cdata field * might try to be accessed, even though it was removed. */ VERIFY(!HDR_L2_WRITING(hdr)); VERIFY3P(hdr->b_l1hdr.b_tmp_cdata, ==, NULL); nhdr->b_flags &= ~ARC_FLAG_HAS_L1HDR; } /* * The header has been reallocated so we need to re-insert it into any * lists it was on. */ (void) buf_hash_insert(nhdr, NULL); ASSERT(list_link_active(&hdr->b_l2hdr.b_l2node)); mutex_enter(&dev->l2ad_mtx); /* * We must place the realloc'ed header back into the list at * the same spot. Otherwise, if it's placed earlier in the list, * l2arc_write_buffers() could find it during the function's * write phase, and try to write it out to the l2arc. */ list_insert_after(&dev->l2ad_buflist, hdr, nhdr); list_remove(&dev->l2ad_buflist, hdr); mutex_exit(&dev->l2ad_mtx); /* * Since we're using the pointer address as the tag when * incrementing and decrementing the l2ad_alloc refcount, we * must remove the old pointer (that we're about to destroy) and * add the new pointer to the refcount. Otherwise we'd remove * the wrong pointer address when calling arc_hdr_destroy() later. */ (void) refcount_remove_many(&dev->l2ad_alloc, hdr->b_l2hdr.b_asize, hdr); (void) refcount_add_many(&dev->l2ad_alloc, nhdr->b_l2hdr.b_asize, nhdr); buf_discard_identity(hdr); hdr->b_freeze_cksum = NULL; kmem_cache_free(old, hdr); return (nhdr); } #define ARC_MINTIME (hz>>4) /* 62 ms */ static void arc_cksum_verify(arc_buf_t *buf) { zio_cksum_t zc; if (!(zfs_flags & ZFS_DEBUG_MODIFY)) return; mutex_enter(&buf->b_hdr->b_l1hdr.b_freeze_lock); if (buf->b_hdr->b_freeze_cksum == NULL || HDR_IO_ERROR(buf->b_hdr)) { mutex_exit(&buf->b_hdr->b_l1hdr.b_freeze_lock); return; } fletcher_2_native(buf->b_data, buf->b_hdr->b_size, &zc); if (!ZIO_CHECKSUM_EQUAL(*buf->b_hdr->b_freeze_cksum, zc)) panic("buffer modified while frozen!"); mutex_exit(&buf->b_hdr->b_l1hdr.b_freeze_lock); } static int arc_cksum_equal(arc_buf_t *buf) { zio_cksum_t zc; int equal; mutex_enter(&buf->b_hdr->b_l1hdr.b_freeze_lock); fletcher_2_native(buf->b_data, buf->b_hdr->b_size, &zc); equal = ZIO_CHECKSUM_EQUAL(*buf->b_hdr->b_freeze_cksum, zc); mutex_exit(&buf->b_hdr->b_l1hdr.b_freeze_lock); return (equal); } static void arc_cksum_compute(arc_buf_t *buf, boolean_t force) { if (!force && !(zfs_flags & ZFS_DEBUG_MODIFY)) return; mutex_enter(&buf->b_hdr->b_l1hdr.b_freeze_lock); if (buf->b_hdr->b_freeze_cksum != NULL) { mutex_exit(&buf->b_hdr->b_l1hdr.b_freeze_lock); return; } buf->b_hdr->b_freeze_cksum = kmem_alloc(sizeof (zio_cksum_t), KM_SLEEP); fletcher_2_native(buf->b_data, buf->b_hdr->b_size, buf->b_hdr->b_freeze_cksum); mutex_exit(&buf->b_hdr->b_l1hdr.b_freeze_lock); arc_buf_watch(buf); } #ifndef _KERNEL void arc_buf_sigsegv(int sig, siginfo_t *si, void *unused) { panic("Got SIGSEGV at address: 0x%lx\n", (long) si->si_addr); } #endif /* ARGSUSED */ static void arc_buf_unwatch(arc_buf_t *buf) { #ifndef _KERNEL if (arc_watch) { ASSERT0(mprotect(buf->b_data, buf->b_hdr->b_size, PROT_READ | PROT_WRITE)); } #endif } /* ARGSUSED */ static void arc_buf_watch(arc_buf_t *buf) { #ifndef _KERNEL if (arc_watch) ASSERT0(mprotect(buf->b_data, buf->b_hdr->b_size, PROT_READ)); #endif } static arc_buf_contents_t arc_buf_type(arc_buf_hdr_t *hdr) { if (HDR_ISTYPE_METADATA(hdr)) { return (ARC_BUFC_METADATA); } else { return (ARC_BUFC_DATA); } } static uint32_t arc_bufc_to_flags(arc_buf_contents_t type) { switch (type) { case ARC_BUFC_DATA: /* metadata field is 0 if buffer contains normal data */ return (0); case ARC_BUFC_METADATA: return (ARC_FLAG_BUFC_METADATA); default: break; } panic("undefined ARC buffer type!"); return ((uint32_t)-1); } void arc_buf_thaw(arc_buf_t *buf) { if (zfs_flags & ZFS_DEBUG_MODIFY) { if (buf->b_hdr->b_l1hdr.b_state != arc_anon) panic("modifying non-anon buffer!"); if (HDR_IO_IN_PROGRESS(buf->b_hdr)) panic("modifying buffer while i/o in progress!"); arc_cksum_verify(buf); } mutex_enter(&buf->b_hdr->b_l1hdr.b_freeze_lock); if (buf->b_hdr->b_freeze_cksum != NULL) { kmem_free(buf->b_hdr->b_freeze_cksum, sizeof (zio_cksum_t)); buf->b_hdr->b_freeze_cksum = NULL; } mutex_exit(&buf->b_hdr->b_l1hdr.b_freeze_lock); arc_buf_unwatch(buf); } void arc_buf_freeze(arc_buf_t *buf) { kmutex_t *hash_lock; if (!(zfs_flags & ZFS_DEBUG_MODIFY)) return; hash_lock = HDR_LOCK(buf->b_hdr); mutex_enter(hash_lock); ASSERT(buf->b_hdr->b_freeze_cksum != NULL || buf->b_hdr->b_l1hdr.b_state == arc_anon); arc_cksum_compute(buf, B_FALSE); mutex_exit(hash_lock); } static void add_reference(arc_buf_hdr_t *hdr, kmutex_t *hash_lock, void *tag) { arc_state_t *state; ASSERT(HDR_HAS_L1HDR(hdr)); ASSERT(MUTEX_HELD(hash_lock)); state = hdr->b_l1hdr.b_state; if ((refcount_add(&hdr->b_l1hdr.b_refcnt, tag) == 1) && (state != arc_anon)) { /* We don't use the L2-only state list. */ if (state != arc_l2c_only) { arc_buf_contents_t type = arc_buf_type(hdr); uint64_t delta = hdr->b_size * hdr->b_l1hdr.b_datacnt; multilist_t *list = &state->arcs_list[type]; uint64_t *size = &state->arcs_lsize[type]; multilist_remove(list, hdr); if (GHOST_STATE(state)) { ASSERT0(hdr->b_l1hdr.b_datacnt); ASSERT3P(hdr->b_l1hdr.b_buf, ==, NULL); delta = hdr->b_size; } ASSERT(delta > 0); ASSERT3U(*size, >=, delta); atomic_add_64(size, -delta); } /* remove the prefetch flag if we get a reference */ hdr->b_flags &= ~ARC_FLAG_PREFETCH; } } static int remove_reference(arc_buf_hdr_t *hdr, kmutex_t *hash_lock, void *tag) { int cnt; arc_state_t *state = hdr->b_l1hdr.b_state; ASSERT(HDR_HAS_L1HDR(hdr)); ASSERT(state == arc_anon || MUTEX_HELD(hash_lock)); ASSERT(!GHOST_STATE(state)); /* * arc_l2c_only counts as a ghost state so we don't need to explicitly * check to prevent usage of the arc_l2c_only list. */ if (((cnt = refcount_remove(&hdr->b_l1hdr.b_refcnt, tag)) == 0) && (state != arc_anon)) { arc_buf_contents_t type = arc_buf_type(hdr); multilist_t *list = &state->arcs_list[type]; uint64_t *size = &state->arcs_lsize[type]; multilist_insert(list, hdr); ASSERT(hdr->b_l1hdr.b_datacnt > 0); atomic_add_64(size, hdr->b_size * hdr->b_l1hdr.b_datacnt); } return (cnt); } /* * Returns detailed information about a specific arc buffer. When the * state_index argument is set the function will calculate the arc header * list position for its arc state. Since this requires a linear traversal * callers are strongly encourage not to do this. However, it can be helpful * for targeted analysis so the functionality is provided. */ void arc_buf_info(arc_buf_t *ab, arc_buf_info_t *abi, int state_index) { arc_buf_hdr_t *hdr = ab->b_hdr; l1arc_buf_hdr_t *l1hdr = NULL; l2arc_buf_hdr_t *l2hdr = NULL; arc_state_t *state = NULL; if (HDR_HAS_L1HDR(hdr)) { l1hdr = &hdr->b_l1hdr; state = l1hdr->b_state; } if (HDR_HAS_L2HDR(hdr)) l2hdr = &hdr->b_l2hdr; memset(abi, 0, sizeof (arc_buf_info_t)); abi->abi_flags = hdr->b_flags; if (l1hdr) { abi->abi_datacnt = l1hdr->b_datacnt; abi->abi_access = l1hdr->b_arc_access; abi->abi_mru_hits = l1hdr->b_mru_hits; abi->abi_mru_ghost_hits = l1hdr->b_mru_ghost_hits; abi->abi_mfu_hits = l1hdr->b_mfu_hits; abi->abi_mfu_ghost_hits = l1hdr->b_mfu_ghost_hits; abi->abi_holds = refcount_count(&l1hdr->b_refcnt); } if (l2hdr) { abi->abi_l2arc_dattr = l2hdr->b_daddr; abi->abi_l2arc_asize = l2hdr->b_asize; abi->abi_l2arc_compress = l2hdr->b_compress; abi->abi_l2arc_hits = l2hdr->b_hits; } abi->abi_state_type = state ? state->arcs_state : ARC_STATE_ANON; abi->abi_state_contents = arc_buf_type(hdr); abi->abi_size = hdr->b_size; } /* * Move the supplied buffer to the indicated state. The hash lock * for the buffer must be held by the caller. */ static void arc_change_state(arc_state_t *new_state, arc_buf_hdr_t *hdr, kmutex_t *hash_lock) { arc_state_t *old_state; int64_t refcnt; uint32_t datacnt; uint64_t from_delta, to_delta; arc_buf_contents_t buftype = arc_buf_type(hdr); /* * We almost always have an L1 hdr here, since we call arc_hdr_realloc() * in arc_read() when bringing a buffer out of the L2ARC. However, the * L1 hdr doesn't always exist when we change state to arc_anon before * destroying a header, in which case reallocating to add the L1 hdr is * pointless. */ if (HDR_HAS_L1HDR(hdr)) { old_state = hdr->b_l1hdr.b_state; refcnt = refcount_count(&hdr->b_l1hdr.b_refcnt); datacnt = hdr->b_l1hdr.b_datacnt; } else { old_state = arc_l2c_only; refcnt = 0; datacnt = 0; } ASSERT(MUTEX_HELD(hash_lock)); ASSERT3P(new_state, !=, old_state); ASSERT(refcnt == 0 || datacnt > 0); ASSERT(!GHOST_STATE(new_state) || datacnt == 0); ASSERT(old_state != arc_anon || datacnt <= 1); from_delta = to_delta = datacnt * hdr->b_size; /* * If this buffer is evictable, transfer it from the * old state list to the new state list. */ if (refcnt == 0) { if (old_state != arc_anon && old_state != arc_l2c_only) { uint64_t *size = &old_state->arcs_lsize[buftype]; ASSERT(HDR_HAS_L1HDR(hdr)); multilist_remove(&old_state->arcs_list[buftype], hdr); /* * If prefetching out of the ghost cache, * we will have a non-zero datacnt. */ if (GHOST_STATE(old_state) && datacnt == 0) { /* ghost elements have a ghost size */ ASSERT(hdr->b_l1hdr.b_buf == NULL); from_delta = hdr->b_size; } ASSERT3U(*size, >=, from_delta); atomic_add_64(size, -from_delta); } if (new_state != arc_anon && new_state != arc_l2c_only) { uint64_t *size = &new_state->arcs_lsize[buftype]; /* * An L1 header always exists here, since if we're * moving to some L1-cached state (i.e. not l2c_only or * anonymous), we realloc the header to add an L1hdr * beforehand. */ ASSERT(HDR_HAS_L1HDR(hdr)); multilist_insert(&new_state->arcs_list[buftype], hdr); /* ghost elements have a ghost size */ if (GHOST_STATE(new_state)) { ASSERT0(datacnt); ASSERT(hdr->b_l1hdr.b_buf == NULL); to_delta = hdr->b_size; } atomic_add_64(size, to_delta); } } ASSERT(!BUF_EMPTY(hdr)); if (new_state == arc_anon && HDR_IN_HASH_TABLE(hdr)) buf_hash_remove(hdr); /* adjust state sizes (ignore arc_l2c_only) */ if (to_delta && new_state != arc_l2c_only) { ASSERT(HDR_HAS_L1HDR(hdr)); if (GHOST_STATE(new_state)) { ASSERT0(datacnt); /* * We moving a header to a ghost state, we first * remove all arc buffers. Thus, we'll have a * datacnt of zero, and no arc buffer to use for * the reference. As a result, we use the arc * header pointer for the reference. */ (void) refcount_add_many(&new_state->arcs_size, hdr->b_size, hdr); } else { arc_buf_t *buf; ASSERT3U(datacnt, !=, 0); /* * Each individual buffer holds a unique reference, * thus we must remove each of these references one * at a time. */ for (buf = hdr->b_l1hdr.b_buf; buf != NULL; buf = buf->b_next) { (void) refcount_add_many(&new_state->arcs_size, hdr->b_size, buf); } } } if (from_delta && old_state != arc_l2c_only) { ASSERT(HDR_HAS_L1HDR(hdr)); if (GHOST_STATE(old_state)) { /* * When moving a header off of a ghost state, * there's the possibility for datacnt to be * non-zero. This is because we first add the * arc buffer to the header prior to changing * the header's state. Since we used the header * for the reference when putting the header on * the ghost state, we must balance that and use * the header when removing off the ghost state * (even though datacnt is non zero). */ IMPLY(datacnt == 0, new_state == arc_anon || new_state == arc_l2c_only); (void) refcount_remove_many(&old_state->arcs_size, hdr->b_size, hdr); } else { arc_buf_t *buf; ASSERT3U(datacnt, !=, 0); /* * Each individual buffer holds a unique reference, * thus we must remove each of these references one * at a time. */ for (buf = hdr->b_l1hdr.b_buf; buf != NULL; buf = buf->b_next) { (void) refcount_remove_many( &old_state->arcs_size, hdr->b_size, buf); } } } if (HDR_HAS_L1HDR(hdr)) hdr->b_l1hdr.b_state = new_state; /* * L2 headers should never be on the L2 state list since they don't * have L1 headers allocated. */ ASSERT(multilist_is_empty(&arc_l2c_only->arcs_list[ARC_BUFC_DATA]) && multilist_is_empty(&arc_l2c_only->arcs_list[ARC_BUFC_METADATA])); } void arc_space_consume(uint64_t space, arc_space_type_t type) { ASSERT(type >= 0 && type < ARC_SPACE_NUMTYPES); switch (type) { default: break; case ARC_SPACE_DATA: ARCSTAT_INCR(arcstat_data_size, space); break; case ARC_SPACE_META: ARCSTAT_INCR(arcstat_metadata_size, space); break; case ARC_SPACE_OTHER: ARCSTAT_INCR(arcstat_other_size, space); break; case ARC_SPACE_HDRS: ARCSTAT_INCR(arcstat_hdr_size, space); break; case ARC_SPACE_L2HDRS: ARCSTAT_INCR(arcstat_l2_hdr_size, space); break; } if (type != ARC_SPACE_DATA) ARCSTAT_INCR(arcstat_meta_used, space); atomic_add_64(&arc_size, space); } void arc_space_return(uint64_t space, arc_space_type_t type) { ASSERT(type >= 0 && type < ARC_SPACE_NUMTYPES); switch (type) { default: break; case ARC_SPACE_DATA: ARCSTAT_INCR(arcstat_data_size, -space); break; case ARC_SPACE_META: ARCSTAT_INCR(arcstat_metadata_size, -space); break; case ARC_SPACE_OTHER: ARCSTAT_INCR(arcstat_other_size, -space); break; case ARC_SPACE_HDRS: ARCSTAT_INCR(arcstat_hdr_size, -space); break; case ARC_SPACE_L2HDRS: ARCSTAT_INCR(arcstat_l2_hdr_size, -space); break; } if (type != ARC_SPACE_DATA) { ASSERT(arc_meta_used >= space); if (arc_meta_max < arc_meta_used) arc_meta_max = arc_meta_used; ARCSTAT_INCR(arcstat_meta_used, -space); } ASSERT(arc_size >= space); atomic_add_64(&arc_size, -space); } arc_buf_t * arc_buf_alloc(spa_t *spa, uint64_t size, void *tag, arc_buf_contents_t type) { arc_buf_hdr_t *hdr; arc_buf_t *buf; VERIFY3U(size, <=, spa_maxblocksize(spa)); hdr = kmem_cache_alloc(hdr_full_cache, KM_PUSHPAGE); ASSERT(BUF_EMPTY(hdr)); ASSERT3P(hdr->b_freeze_cksum, ==, NULL); hdr->b_size = size; hdr->b_spa = spa_load_guid(spa); hdr->b_l1hdr.b_mru_hits = 0; hdr->b_l1hdr.b_mru_ghost_hits = 0; hdr->b_l1hdr.b_mfu_hits = 0; hdr->b_l1hdr.b_mfu_ghost_hits = 0; hdr->b_l1hdr.b_l2_hits = 0; buf = kmem_cache_alloc(buf_cache, KM_PUSHPAGE); buf->b_hdr = hdr; buf->b_data = NULL; buf->b_efunc = NULL; buf->b_private = NULL; buf->b_next = NULL; hdr->b_flags = arc_bufc_to_flags(type); hdr->b_flags |= ARC_FLAG_HAS_L1HDR; hdr->b_l1hdr.b_buf = buf; hdr->b_l1hdr.b_state = arc_anon; hdr->b_l1hdr.b_arc_access = 0; hdr->b_l1hdr.b_datacnt = 1; hdr->b_l1hdr.b_tmp_cdata = NULL; arc_get_data_buf(buf); ASSERT(refcount_is_zero(&hdr->b_l1hdr.b_refcnt)); (void) refcount_add(&hdr->b_l1hdr.b_refcnt, tag); return (buf); } static char *arc_onloan_tag = "onloan"; /* * Loan out an anonymous arc buffer. Loaned buffers are not counted as in * flight data by arc_tempreserve_space() until they are "returned". Loaned * buffers must be returned to the arc before they can be used by the DMU or * freed. */ arc_buf_t * arc_loan_buf(spa_t *spa, uint64_t size) { arc_buf_t *buf; buf = arc_buf_alloc(spa, size, arc_onloan_tag, ARC_BUFC_DATA); atomic_add_64(&arc_loaned_bytes, size); return (buf); } /* * Return a loaned arc buffer to the arc. */ void arc_return_buf(arc_buf_t *buf, void *tag) { arc_buf_hdr_t *hdr = buf->b_hdr; ASSERT(buf->b_data != NULL); ASSERT(HDR_HAS_L1HDR(hdr)); (void) refcount_add(&hdr->b_l1hdr.b_refcnt, tag); (void) refcount_remove(&hdr->b_l1hdr.b_refcnt, arc_onloan_tag); atomic_add_64(&arc_loaned_bytes, -hdr->b_size); } /* Detach an arc_buf from a dbuf (tag) */ void arc_loan_inuse_buf(arc_buf_t *buf, void *tag) { arc_buf_hdr_t *hdr = buf->b_hdr; ASSERT(buf->b_data != NULL); ASSERT(HDR_HAS_L1HDR(hdr)); (void) refcount_add(&hdr->b_l1hdr.b_refcnt, arc_onloan_tag); (void) refcount_remove(&hdr->b_l1hdr.b_refcnt, tag); buf->b_efunc = NULL; buf->b_private = NULL; atomic_add_64(&arc_loaned_bytes, hdr->b_size); } static arc_buf_t * arc_buf_clone(arc_buf_t *from) { arc_buf_t *buf; arc_buf_hdr_t *hdr = from->b_hdr; uint64_t size = hdr->b_size; ASSERT(HDR_HAS_L1HDR(hdr)); ASSERT(hdr->b_l1hdr.b_state != arc_anon); buf = kmem_cache_alloc(buf_cache, KM_PUSHPAGE); buf->b_hdr = hdr; buf->b_data = NULL; buf->b_efunc = NULL; buf->b_private = NULL; buf->b_next = hdr->b_l1hdr.b_buf; hdr->b_l1hdr.b_buf = buf; arc_get_data_buf(buf); bcopy(from->b_data, buf->b_data, size); /* * This buffer already exists in the arc so create a duplicate * copy for the caller. If the buffer is associated with user data * then track the size and number of duplicates. These stats will be * updated as duplicate buffers are created and destroyed. */ if (HDR_ISTYPE_DATA(hdr)) { ARCSTAT_BUMP(arcstat_duplicate_buffers); ARCSTAT_INCR(arcstat_duplicate_buffers_size, size); } hdr->b_l1hdr.b_datacnt += 1; return (buf); } void arc_buf_add_ref(arc_buf_t *buf, void* tag) { arc_buf_hdr_t *hdr; kmutex_t *hash_lock; /* * Check to see if this buffer is evicted. Callers * must verify b_data != NULL to know if the add_ref * was successful. */ mutex_enter(&buf->b_evict_lock); if (buf->b_data == NULL) { mutex_exit(&buf->b_evict_lock); return; } hash_lock = HDR_LOCK(buf->b_hdr); mutex_enter(hash_lock); hdr = buf->b_hdr; ASSERT(HDR_HAS_L1HDR(hdr)); ASSERT3P(hash_lock, ==, HDR_LOCK(hdr)); mutex_exit(&buf->b_evict_lock); ASSERT(hdr->b_l1hdr.b_state == arc_mru || hdr->b_l1hdr.b_state == arc_mfu); add_reference(hdr, hash_lock, tag); DTRACE_PROBE1(arc__hit, arc_buf_hdr_t *, hdr); arc_access(hdr, hash_lock); mutex_exit(hash_lock); ARCSTAT_BUMP(arcstat_hits); ARCSTAT_CONDSTAT(!HDR_PREFETCH(hdr), demand, prefetch, !HDR_ISTYPE_METADATA(hdr), data, metadata, hits); } static void arc_buf_free_on_write(void *data, size_t size, void (*free_func)(void *, size_t)) { l2arc_data_free_t *df; df = kmem_alloc(sizeof (*df), KM_SLEEP); df->l2df_data = data; df->l2df_size = size; df->l2df_func = free_func; mutex_enter(&l2arc_free_on_write_mtx); list_insert_head(l2arc_free_on_write, df); mutex_exit(&l2arc_free_on_write_mtx); } /* * Free the arc data buffer. If it is an l2arc write in progress, * the buffer is placed on l2arc_free_on_write to be freed later. */ static void arc_buf_data_free(arc_buf_t *buf, void (*free_func)(void *, size_t)) { arc_buf_hdr_t *hdr = buf->b_hdr; if (HDR_L2_WRITING(hdr)) { arc_buf_free_on_write(buf->b_data, hdr->b_size, free_func); ARCSTAT_BUMP(arcstat_l2_free_on_write); } else { free_func(buf->b_data, hdr->b_size); } } static void arc_buf_l2_cdata_free(arc_buf_hdr_t *hdr) { ASSERT(HDR_HAS_L2HDR(hdr)); ASSERT(MUTEX_HELD(&hdr->b_l2hdr.b_dev->l2ad_mtx)); /* * The b_tmp_cdata field is linked off of the b_l1hdr, so if * that doesn't exist, the header is in the arc_l2c_only state, * and there isn't anything to free (it's already been freed). */ if (!HDR_HAS_L1HDR(hdr)) return; /* * The header isn't being written to the l2arc device, thus it * shouldn't have a b_tmp_cdata to free. */ if (!HDR_L2_WRITING(hdr)) { ASSERT3P(hdr->b_l1hdr.b_tmp_cdata, ==, NULL); return; } /* * The header does not have compression enabled. This can be due * to the buffer not being compressible, or because we're * freeing the buffer before the second phase of * l2arc_write_buffer() has started (which does the compression * step). In either case, b_tmp_cdata does not point to a * separately compressed buffer, so there's nothing to free (it * points to the same buffer as the arc_buf_t's b_data field). */ if (hdr->b_l2hdr.b_compress == ZIO_COMPRESS_OFF) { hdr->b_l1hdr.b_tmp_cdata = NULL; return; } /* * There's nothing to free since the buffer was all zero's and * compressed to a zero length buffer. */ if (hdr->b_l2hdr.b_compress == ZIO_COMPRESS_EMPTY) { ASSERT3P(hdr->b_l1hdr.b_tmp_cdata, ==, NULL); return; } ASSERT(L2ARC_IS_VALID_COMPRESS(hdr->b_l2hdr.b_compress)); arc_buf_free_on_write(hdr->b_l1hdr.b_tmp_cdata, hdr->b_size, zio_data_buf_free); ARCSTAT_BUMP(arcstat_l2_cdata_free_on_write); hdr->b_l1hdr.b_tmp_cdata = NULL; } /* * Free up buf->b_data and if 'remove' is set, then pull the * arc_buf_t off of the the arc_buf_hdr_t's list and free it. */ static void arc_buf_destroy(arc_buf_t *buf, boolean_t remove) { arc_buf_t **bufp; /* free up data associated with the buf */ if (buf->b_data != NULL) { arc_state_t *state = buf->b_hdr->b_l1hdr.b_state; uint64_t size = buf->b_hdr->b_size; arc_buf_contents_t type = arc_buf_type(buf->b_hdr); arc_cksum_verify(buf); arc_buf_unwatch(buf); if (type == ARC_BUFC_METADATA) { arc_buf_data_free(buf, zio_buf_free); arc_space_return(size, ARC_SPACE_META); } else { ASSERT(type == ARC_BUFC_DATA); arc_buf_data_free(buf, zio_data_buf_free); arc_space_return(size, ARC_SPACE_DATA); } /* protected by hash lock, if in the hash table */ if (multilist_link_active(&buf->b_hdr->b_l1hdr.b_arc_node)) { uint64_t *cnt = &state->arcs_lsize[type]; ASSERT(refcount_is_zero( &buf->b_hdr->b_l1hdr.b_refcnt)); ASSERT(state != arc_anon && state != arc_l2c_only); ASSERT3U(*cnt, >=, size); atomic_add_64(cnt, -size); } (void) refcount_remove_many(&state->arcs_size, size, buf); buf->b_data = NULL; /* * If we're destroying a duplicate buffer make sure * that the appropriate statistics are updated. */ if (buf->b_hdr->b_l1hdr.b_datacnt > 1 && HDR_ISTYPE_DATA(buf->b_hdr)) { ARCSTAT_BUMPDOWN(arcstat_duplicate_buffers); ARCSTAT_INCR(arcstat_duplicate_buffers_size, -size); } ASSERT(buf->b_hdr->b_l1hdr.b_datacnt > 0); buf->b_hdr->b_l1hdr.b_datacnt -= 1; } /* only remove the buf if requested */ if (!remove) return; /* remove the buf from the hdr list */ for (bufp = &buf->b_hdr->b_l1hdr.b_buf; *bufp != buf; bufp = &(*bufp)->b_next) continue; *bufp = buf->b_next; buf->b_next = NULL; ASSERT(buf->b_efunc == NULL); /* clean up the buf */ buf->b_hdr = NULL; kmem_cache_free(buf_cache, buf); } static void arc_hdr_l2hdr_destroy(arc_buf_hdr_t *hdr) { l2arc_buf_hdr_t *l2hdr = &hdr->b_l2hdr; l2arc_dev_t *dev = l2hdr->b_dev; ASSERT(MUTEX_HELD(&dev->l2ad_mtx)); ASSERT(HDR_HAS_L2HDR(hdr)); list_remove(&dev->l2ad_buflist, hdr); /* * We don't want to leak the b_tmp_cdata buffer that was * allocated in l2arc_write_buffers() */ arc_buf_l2_cdata_free(hdr); /* * If the l2hdr's b_daddr is equal to L2ARC_ADDR_UNSET, then * this header is being processed by l2arc_write_buffers() (i.e. * it's in the first stage of l2arc_write_buffers()). * Re-affirming that truth here, just to serve as a reminder. If * b_daddr does not equal L2ARC_ADDR_UNSET, then the header may or * may not have its HDR_L2_WRITING flag set. (the write may have * completed, in which case HDR_L2_WRITING will be false and the * b_daddr field will point to the address of the buffer on disk). */ IMPLY(l2hdr->b_daddr == L2ARC_ADDR_UNSET, HDR_L2_WRITING(hdr)); /* * If b_daddr is equal to L2ARC_ADDR_UNSET, we're racing with * l2arc_write_buffers(). Since we've just removed this header * from the l2arc buffer list, this header will never reach the * second stage of l2arc_write_buffers(), which increments the * accounting stats for this header. Thus, we must be careful * not to decrement them for this header either. */ if (l2hdr->b_daddr != L2ARC_ADDR_UNSET) { ARCSTAT_INCR(arcstat_l2_asize, -l2hdr->b_asize); ARCSTAT_INCR(arcstat_l2_size, -hdr->b_size); vdev_space_update(dev->l2ad_vdev, -l2hdr->b_asize, 0, 0); (void) refcount_remove_many(&dev->l2ad_alloc, l2hdr->b_asize, hdr); } hdr->b_flags &= ~ARC_FLAG_HAS_L2HDR; } static void arc_hdr_destroy(arc_buf_hdr_t *hdr) { if (HDR_HAS_L1HDR(hdr)) { ASSERT(hdr->b_l1hdr.b_buf == NULL || hdr->b_l1hdr.b_datacnt > 0); ASSERT(refcount_is_zero(&hdr->b_l1hdr.b_refcnt)); ASSERT3P(hdr->b_l1hdr.b_state, ==, arc_anon); } ASSERT(!HDR_IO_IN_PROGRESS(hdr)); ASSERT(!HDR_IN_HASH_TABLE(hdr)); if (HDR_HAS_L2HDR(hdr)) { l2arc_dev_t *dev = hdr->b_l2hdr.b_dev; boolean_t buflist_held = MUTEX_HELD(&dev->l2ad_mtx); if (!buflist_held) mutex_enter(&dev->l2ad_mtx); /* * Even though we checked this conditional above, we * need to check this again now that we have the * l2ad_mtx. This is because we could be racing with * another thread calling l2arc_evict() which might have * destroyed this header's L2 portion as we were waiting * to acquire the l2ad_mtx. If that happens, we don't * want to re-destroy the header's L2 portion. */ if (HDR_HAS_L2HDR(hdr)) arc_hdr_l2hdr_destroy(hdr); if (!buflist_held) mutex_exit(&dev->l2ad_mtx); } if (!BUF_EMPTY(hdr)) buf_discard_identity(hdr); if (hdr->b_freeze_cksum != NULL) { kmem_free(hdr->b_freeze_cksum, sizeof (zio_cksum_t)); hdr->b_freeze_cksum = NULL; } if (HDR_HAS_L1HDR(hdr)) { while (hdr->b_l1hdr.b_buf) { arc_buf_t *buf = hdr->b_l1hdr.b_buf; if (buf->b_efunc != NULL) { mutex_enter(&arc_user_evicts_lock); mutex_enter(&buf->b_evict_lock); ASSERT(buf->b_hdr != NULL); arc_buf_destroy(hdr->b_l1hdr.b_buf, FALSE); hdr->b_l1hdr.b_buf = buf->b_next; buf->b_hdr = &arc_eviction_hdr; buf->b_next = arc_eviction_list; arc_eviction_list = buf; mutex_exit(&buf->b_evict_lock); cv_signal(&arc_user_evicts_cv); mutex_exit(&arc_user_evicts_lock); } else { arc_buf_destroy(hdr->b_l1hdr.b_buf, TRUE); } } } ASSERT3P(hdr->b_hash_next, ==, NULL); if (HDR_HAS_L1HDR(hdr)) { ASSERT(!multilist_link_active(&hdr->b_l1hdr.b_arc_node)); ASSERT3P(hdr->b_l1hdr.b_acb, ==, NULL); kmem_cache_free(hdr_full_cache, hdr); } else { kmem_cache_free(hdr_l2only_cache, hdr); } } void arc_buf_free(arc_buf_t *buf, void *tag) { arc_buf_hdr_t *hdr = buf->b_hdr; int hashed = hdr->b_l1hdr.b_state != arc_anon; ASSERT(buf->b_efunc == NULL); ASSERT(buf->b_data != NULL); if (hashed) { kmutex_t *hash_lock = HDR_LOCK(hdr); mutex_enter(hash_lock); hdr = buf->b_hdr; ASSERT3P(hash_lock, ==, HDR_LOCK(hdr)); (void) remove_reference(hdr, hash_lock, tag); if (hdr->b_l1hdr.b_datacnt > 1) { arc_buf_destroy(buf, TRUE); } else { ASSERT(buf == hdr->b_l1hdr.b_buf); ASSERT(buf->b_efunc == NULL); hdr->b_flags |= ARC_FLAG_BUF_AVAILABLE; } mutex_exit(hash_lock); } else if (HDR_IO_IN_PROGRESS(hdr)) { int destroy_hdr; /* * We are in the middle of an async write. Don't destroy * this buffer unless the write completes before we finish * decrementing the reference count. */ mutex_enter(&arc_user_evicts_lock); (void) remove_reference(hdr, NULL, tag); ASSERT(refcount_is_zero(&hdr->b_l1hdr.b_refcnt)); destroy_hdr = !HDR_IO_IN_PROGRESS(hdr); mutex_exit(&arc_user_evicts_lock); if (destroy_hdr) arc_hdr_destroy(hdr); } else { if (remove_reference(hdr, NULL, tag) > 0) arc_buf_destroy(buf, TRUE); else arc_hdr_destroy(hdr); } } boolean_t arc_buf_remove_ref(arc_buf_t *buf, void* tag) { arc_buf_hdr_t *hdr = buf->b_hdr; kmutex_t *hash_lock = HDR_LOCK(hdr); boolean_t no_callback = (buf->b_efunc == NULL); if (hdr->b_l1hdr.b_state == arc_anon) { ASSERT(hdr->b_l1hdr.b_datacnt == 1); arc_buf_free(buf, tag); return (no_callback); } mutex_enter(hash_lock); hdr = buf->b_hdr; ASSERT(hdr->b_l1hdr.b_datacnt > 0); ASSERT3P(hash_lock, ==, HDR_LOCK(hdr)); ASSERT(hdr->b_l1hdr.b_state != arc_anon); ASSERT(buf->b_data != NULL); (void) remove_reference(hdr, hash_lock, tag); if (hdr->b_l1hdr.b_datacnt > 1) { if (no_callback) arc_buf_destroy(buf, TRUE); } else if (no_callback) { ASSERT(hdr->b_l1hdr.b_buf == buf && buf->b_next == NULL); ASSERT(buf->b_efunc == NULL); hdr->b_flags |= ARC_FLAG_BUF_AVAILABLE; } ASSERT(no_callback || hdr->b_l1hdr.b_datacnt > 1 || refcount_is_zero(&hdr->b_l1hdr.b_refcnt)); mutex_exit(hash_lock); return (no_callback); } uint64_t arc_buf_size(arc_buf_t *buf) { return (buf->b_hdr->b_size); } /* * Called from the DMU to determine if the current buffer should be * evicted. In order to ensure proper locking, the eviction must be initiated * from the DMU. Return true if the buffer is associated with user data and * duplicate buffers still exist. */ boolean_t arc_buf_eviction_needed(arc_buf_t *buf) { arc_buf_hdr_t *hdr; boolean_t evict_needed = B_FALSE; if (zfs_disable_dup_eviction) return (B_FALSE); mutex_enter(&buf->b_evict_lock); hdr = buf->b_hdr; if (hdr == NULL) { /* * We are in arc_do_user_evicts(); let that function * perform the eviction. */ ASSERT(buf->b_data == NULL); mutex_exit(&buf->b_evict_lock); return (B_FALSE); } else if (buf->b_data == NULL) { /* * We have already been added to the arc eviction list; * recommend eviction. */ ASSERT3P(hdr, ==, &arc_eviction_hdr); mutex_exit(&buf->b_evict_lock); return (B_TRUE); } if (hdr->b_l1hdr.b_datacnt > 1 && HDR_ISTYPE_DATA(hdr)) evict_needed = B_TRUE; mutex_exit(&buf->b_evict_lock); return (evict_needed); } /* * Evict the arc_buf_hdr that is provided as a parameter. The resultant * state of the header is dependent on its state prior to entering this * function. The following transitions are possible: * * - arc_mru -> arc_mru_ghost * - arc_mfu -> arc_mfu_ghost * - arc_mru_ghost -> arc_l2c_only * - arc_mru_ghost -> deleted * - arc_mfu_ghost -> arc_l2c_only * - arc_mfu_ghost -> deleted */ static int64_t arc_evict_hdr(arc_buf_hdr_t *hdr, kmutex_t *hash_lock) { arc_state_t *evicted_state, *state; int64_t bytes_evicted = 0; ASSERT(MUTEX_HELD(hash_lock)); ASSERT(HDR_HAS_L1HDR(hdr)); state = hdr->b_l1hdr.b_state; if (GHOST_STATE(state)) { ASSERT(!HDR_IO_IN_PROGRESS(hdr)); ASSERT(hdr->b_l1hdr.b_buf == NULL); /* * l2arc_write_buffers() relies on a header's L1 portion * (i.e. its b_tmp_cdata field) during its write phase. * Thus, we cannot push a header onto the arc_l2c_only * state (removing its L1 piece) until the header is * done being written to the l2arc. */ if (HDR_HAS_L2HDR(hdr) && HDR_L2_WRITING(hdr)) { ARCSTAT_BUMP(arcstat_evict_l2_skip); return (bytes_evicted); } ARCSTAT_BUMP(arcstat_deleted); bytes_evicted += hdr->b_size; DTRACE_PROBE1(arc__delete, arc_buf_hdr_t *, hdr); if (HDR_HAS_L2HDR(hdr)) { /* * This buffer is cached on the 2nd Level ARC; * don't destroy the header. */ arc_change_state(arc_l2c_only, hdr, hash_lock); /* * dropping from L1+L2 cached to L2-only, * realloc to remove the L1 header. */ hdr = arc_hdr_realloc(hdr, hdr_full_cache, hdr_l2only_cache); } else { arc_change_state(arc_anon, hdr, hash_lock); arc_hdr_destroy(hdr); } return (bytes_evicted); } ASSERT(state == arc_mru || state == arc_mfu); evicted_state = (state == arc_mru) ? arc_mru_ghost : arc_mfu_ghost; /* prefetch buffers have a minimum lifespan */ if (HDR_IO_IN_PROGRESS(hdr) || ((hdr->b_flags & (ARC_FLAG_PREFETCH | ARC_FLAG_INDIRECT)) && ddi_get_lbolt() - hdr->b_l1hdr.b_arc_access < arc_min_prefetch_lifespan)) { ARCSTAT_BUMP(arcstat_evict_skip); return (bytes_evicted); } ASSERT0(refcount_count(&hdr->b_l1hdr.b_refcnt)); ASSERT3U(hdr->b_l1hdr.b_datacnt, >, 0); while (hdr->b_l1hdr.b_buf) { arc_buf_t *buf = hdr->b_l1hdr.b_buf; if (!mutex_tryenter(&buf->b_evict_lock)) { ARCSTAT_BUMP(arcstat_mutex_miss); break; } if (buf->b_data != NULL) bytes_evicted += hdr->b_size; if (buf->b_efunc != NULL) { mutex_enter(&arc_user_evicts_lock); arc_buf_destroy(buf, FALSE); hdr->b_l1hdr.b_buf = buf->b_next; buf->b_hdr = &arc_eviction_hdr; buf->b_next = arc_eviction_list; arc_eviction_list = buf; cv_signal(&arc_user_evicts_cv); mutex_exit(&arc_user_evicts_lock); mutex_exit(&buf->b_evict_lock); } else { mutex_exit(&buf->b_evict_lock); arc_buf_destroy(buf, TRUE); } } if (HDR_HAS_L2HDR(hdr)) { ARCSTAT_INCR(arcstat_evict_l2_cached, hdr->b_size); } else { if (l2arc_write_eligible(hdr->b_spa, hdr)) ARCSTAT_INCR(arcstat_evict_l2_eligible, hdr->b_size); else ARCSTAT_INCR(arcstat_evict_l2_ineligible, hdr->b_size); } if (hdr->b_l1hdr.b_datacnt == 0) { arc_change_state(evicted_state, hdr, hash_lock); ASSERT(HDR_IN_HASH_TABLE(hdr)); hdr->b_flags |= ARC_FLAG_IN_HASH_TABLE; hdr->b_flags &= ~ARC_FLAG_BUF_AVAILABLE; DTRACE_PROBE1(arc__evict, arc_buf_hdr_t *, hdr); } return (bytes_evicted); } static uint64_t arc_evict_state_impl(multilist_t *ml, int idx, arc_buf_hdr_t *marker, uint64_t spa, int64_t bytes) { multilist_sublist_t *mls; uint64_t bytes_evicted = 0; arc_buf_hdr_t *hdr; kmutex_t *hash_lock; int evict_count = 0; ASSERT3P(marker, !=, NULL); IMPLY(bytes < 0, bytes == ARC_EVICT_ALL); mls = multilist_sublist_lock(ml, idx); for (hdr = multilist_sublist_prev(mls, marker); hdr != NULL; hdr = multilist_sublist_prev(mls, marker)) { if ((bytes != ARC_EVICT_ALL && bytes_evicted >= bytes) || (evict_count >= zfs_arc_evict_batch_limit)) break; /* * To keep our iteration location, move the marker * forward. Since we're not holding hdr's hash lock, we * must be very careful and not remove 'hdr' from the * sublist. Otherwise, other consumers might mistake the * 'hdr' as not being on a sublist when they call the * multilist_link_active() function (they all rely on * the hash lock protecting concurrent insertions and * removals). multilist_sublist_move_forward() was * specifically implemented to ensure this is the case * (only 'marker' will be removed and re-inserted). */ multilist_sublist_move_forward(mls, marker); /* * The only case where the b_spa field should ever be * zero, is the marker headers inserted by * arc_evict_state(). It's possible for multiple threads * to be calling arc_evict_state() concurrently (e.g. * dsl_pool_close() and zio_inject_fault()), so we must * skip any markers we see from these other threads. */ if (hdr->b_spa == 0) continue; /* we're only interested in evicting buffers of a certain spa */ if (spa != 0 && hdr->b_spa != spa) { ARCSTAT_BUMP(arcstat_evict_skip); continue; } hash_lock = HDR_LOCK(hdr); /* * We aren't calling this function from any code path * that would already be holding a hash lock, so we're * asserting on this assumption to be defensive in case * this ever changes. Without this check, it would be * possible to incorrectly increment arcstat_mutex_miss * below (e.g. if the code changed such that we called * this function with a hash lock held). */ ASSERT(!MUTEX_HELD(hash_lock)); if (mutex_tryenter(hash_lock)) { uint64_t evicted = arc_evict_hdr(hdr, hash_lock); mutex_exit(hash_lock); bytes_evicted += evicted; /* * If evicted is zero, arc_evict_hdr() must have * decided to skip this header, don't increment * evict_count in this case. */ if (evicted != 0) evict_count++; /* * If arc_size isn't overflowing, signal any * threads that might happen to be waiting. * * For each header evicted, we wake up a single * thread. If we used cv_broadcast, we could * wake up "too many" threads causing arc_size * to significantly overflow arc_c; since * arc_get_data_buf() doesn't check for overflow * when it's woken up (it doesn't because it's * possible for the ARC to be overflowing while * full of un-evictable buffers, and the * function should proceed in this case). * * If threads are left sleeping, due to not * using cv_broadcast, they will be woken up * just before arc_reclaim_thread() sleeps. */ mutex_enter(&arc_reclaim_lock); if (!arc_is_overflowing()) cv_signal(&arc_reclaim_waiters_cv); mutex_exit(&arc_reclaim_lock); } else { ARCSTAT_BUMP(arcstat_mutex_miss); } } multilist_sublist_unlock(mls); return (bytes_evicted); } /* * Evict buffers from the given arc state, until we've removed the * specified number of bytes. Move the removed buffers to the * appropriate evict state. * * This function makes a "best effort". It skips over any buffers * it can't get a hash_lock on, and so, may not catch all candidates. * It may also return without evicting as much space as requested. * * If bytes is specified using the special value ARC_EVICT_ALL, this * will evict all available (i.e. unlocked and evictable) buffers from * the given arc state; which is used by arc_flush(). */ static uint64_t arc_evict_state(arc_state_t *state, uint64_t spa, int64_t bytes, arc_buf_contents_t type) { uint64_t total_evicted = 0; multilist_t *ml = &state->arcs_list[type]; int num_sublists; arc_buf_hdr_t **markers; int i; IMPLY(bytes < 0, bytes == ARC_EVICT_ALL); num_sublists = multilist_get_num_sublists(ml); /* * If we've tried to evict from each sublist, made some * progress, but still have not hit the target number of bytes * to evict, we want to keep trying. The markers allow us to * pick up where we left off for each individual sublist, rather * than starting from the tail each time. */ markers = kmem_zalloc(sizeof (*markers) * num_sublists, KM_SLEEP); for (i = 0; i < num_sublists; i++) { multilist_sublist_t *mls; markers[i] = kmem_cache_alloc(hdr_full_cache, KM_SLEEP); /* * A b_spa of 0 is used to indicate that this header is * a marker. This fact is used in arc_adjust_type() and * arc_evict_state_impl(). */ markers[i]->b_spa = 0; mls = multilist_sublist_lock(ml, i); multilist_sublist_insert_tail(mls, markers[i]); multilist_sublist_unlock(mls); } /* * While we haven't hit our target number of bytes to evict, or * we're evicting all available buffers. */ while (total_evicted < bytes || bytes == ARC_EVICT_ALL) { /* * Start eviction using a randomly selected sublist, * this is to try and evenly balance eviction across all * sublists. Always starting at the same sublist * (e.g. index 0) would cause evictions to favor certain * sublists over others. */ int sublist_idx = multilist_get_random_index(ml); uint64_t scan_evicted = 0; for (i = 0; i < num_sublists; i++) { uint64_t bytes_remaining; uint64_t bytes_evicted; if (bytes == ARC_EVICT_ALL) bytes_remaining = ARC_EVICT_ALL; else if (total_evicted < bytes) bytes_remaining = bytes - total_evicted; else break; bytes_evicted = arc_evict_state_impl(ml, sublist_idx, markers[sublist_idx], spa, bytes_remaining); scan_evicted += bytes_evicted; total_evicted += bytes_evicted; /* we've reached the end, wrap to the beginning */ if (++sublist_idx >= num_sublists) sublist_idx = 0; } /* * If we didn't evict anything during this scan, we have * no reason to believe we'll evict more during another * scan, so break the loop. */ if (scan_evicted == 0) { /* This isn't possible, let's make that obvious */ ASSERT3S(bytes, !=, 0); /* * When bytes is ARC_EVICT_ALL, the only way to * break the loop is when scan_evicted is zero. * In that case, we actually have evicted enough, * so we don't want to increment the kstat. */ if (bytes != ARC_EVICT_ALL) { ASSERT3S(total_evicted, <, bytes); ARCSTAT_BUMP(arcstat_evict_not_enough); } break; } } for (i = 0; i < num_sublists; i++) { multilist_sublist_t *mls = multilist_sublist_lock(ml, i); multilist_sublist_remove(mls, markers[i]); multilist_sublist_unlock(mls); kmem_cache_free(hdr_full_cache, markers[i]); } kmem_free(markers, sizeof (*markers) * num_sublists); return (total_evicted); } /* * Flush all "evictable" data of the given type from the arc state * specified. This will not evict any "active" buffers (i.e. referenced). * * When 'retry' is set to FALSE, the function will make a single pass * over the state and evict any buffers that it can. Since it doesn't * continually retry the eviction, it might end up leaving some buffers * in the ARC due to lock misses. * * When 'retry' is set to TRUE, the function will continually retry the * eviction until *all* evictable buffers have been removed from the * state. As a result, if concurrent insertions into the state are * allowed (e.g. if the ARC isn't shutting down), this function might * wind up in an infinite loop, continually trying to evict buffers. */ static uint64_t arc_flush_state(arc_state_t *state, uint64_t spa, arc_buf_contents_t type, boolean_t retry) { uint64_t evicted = 0; while (state->arcs_lsize[type] != 0) { evicted += arc_evict_state(state, spa, ARC_EVICT_ALL, type); if (!retry) break; } return (evicted); } /* * Helper function for arc_prune_async() it is responsible for safely * handling the execution of a registered arc_prune_func_t. */ static void arc_prune_task(void *ptr) { arc_prune_t *ap = (arc_prune_t *)ptr; arc_prune_func_t *func = ap->p_pfunc; if (func != NULL) func(ap->p_adjust, ap->p_private); /* Callback unregistered concurrently with execution */ if (refcount_remove(&ap->p_refcnt, func) == 0) { ASSERT(!list_link_active(&ap->p_node)); refcount_destroy(&ap->p_refcnt); kmem_free(ap, sizeof (*ap)); } } /* * Notify registered consumers they must drop holds on a portion of the ARC * buffered they reference. This provides a mechanism to ensure the ARC can * honor the arc_meta_limit and reclaim otherwise pinned ARC buffers. This * is analogous to dnlc_reduce_cache() but more generic. * * This operation is performed asynchronously so it may be safely called * in the context of the arc_reclaim_thread(). A reference is taken here * for each registered arc_prune_t and the arc_prune_task() is responsible * for releasing it once the registered arc_prune_func_t has completed. */ static void arc_prune_async(int64_t adjust) { arc_prune_t *ap; mutex_enter(&arc_prune_mtx); for (ap = list_head(&arc_prune_list); ap != NULL; ap = list_next(&arc_prune_list, ap)) { if (refcount_count(&ap->p_refcnt) >= 2) continue; refcount_add(&ap->p_refcnt, ap->p_pfunc); ap->p_adjust = adjust; taskq_dispatch(arc_prune_taskq, arc_prune_task, ap, TQ_SLEEP); ARCSTAT_BUMP(arcstat_prune); } mutex_exit(&arc_prune_mtx); } /* * Evict the specified number of bytes from the state specified, * restricting eviction to the spa and type given. This function * prevents us from trying to evict more from a state's list than * is "evictable", and to skip evicting altogether when passed a * negative value for "bytes". In contrast, arc_evict_state() will * evict everything it can, when passed a negative value for "bytes". */ static uint64_t arc_adjust_impl(arc_state_t *state, uint64_t spa, int64_t bytes, arc_buf_contents_t type) { int64_t delta; if (bytes > 0 && state->arcs_lsize[type] > 0) { delta = MIN(state->arcs_lsize[type], bytes); return (arc_evict_state(state, spa, delta, type)); } return (0); } /* * The goal of this function is to evict enough meta data buffers from the * ARC in order to enforce the arc_meta_limit. Achieving this is slightly * more complicated than it appears because it is common for data buffers * to have holds on meta data buffers. In addition, dnode meta data buffers * will be held by the dnodes in the block preventing them from being freed. * This means we can't simply traverse the ARC and expect to always find * enough unheld meta data buffer to release. * * Therefore, this function has been updated to make alternating passes * over the ARC releasing data buffers and then newly unheld meta data * buffers. This ensures forward progress is maintained and arc_meta_used * will decrease. Normally this is sufficient, but if required the ARC * will call the registered prune callbacks causing dentry and inodes to * be dropped from the VFS cache. This will make dnode meta data buffers * available for reclaim. */ static uint64_t arc_adjust_meta_balanced(void) { int64_t adjustmnt, delta, prune = 0; uint64_t total_evicted = 0; arc_buf_contents_t type = ARC_BUFC_DATA; int restarts = MAX(zfs_arc_meta_adjust_restarts, 0); restart: /* * This slightly differs than the way we evict from the mru in * arc_adjust because we don't have a "target" value (i.e. no * "meta" arc_p). As a result, I think we can completely * cannibalize the metadata in the MRU before we evict the * metadata from the MFU. I think we probably need to implement a * "metadata arc_p" value to do this properly. */ adjustmnt = arc_meta_used - arc_meta_limit; if (adjustmnt > 0 && arc_mru->arcs_lsize[type] > 0) { delta = MIN(arc_mru->arcs_lsize[type], adjustmnt); total_evicted += arc_adjust_impl(arc_mru, 0, delta, type); adjustmnt -= delta; } /* * We can't afford to recalculate adjustmnt here. If we do, * new metadata buffers can sneak into the MRU or ANON lists, * thus penalize the MFU metadata. Although the fudge factor is * small, it has been empirically shown to be significant for * certain workloads (e.g. creating many empty directories). As * such, we use the original calculation for adjustmnt, and * simply decrement the amount of data evicted from the MRU. */ if (adjustmnt > 0 && arc_mfu->arcs_lsize[type] > 0) { delta = MIN(arc_mfu->arcs_lsize[type], adjustmnt); total_evicted += arc_adjust_impl(arc_mfu, 0, delta, type); } adjustmnt = arc_meta_used - arc_meta_limit; if (adjustmnt > 0 && arc_mru_ghost->arcs_lsize[type] > 0) { delta = MIN(adjustmnt, arc_mru_ghost->arcs_lsize[type]); total_evicted += arc_adjust_impl(arc_mru_ghost, 0, delta, type); adjustmnt -= delta; } if (adjustmnt > 0 && arc_mfu_ghost->arcs_lsize[type] > 0) { delta = MIN(adjustmnt, arc_mfu_ghost->arcs_lsize[type]); total_evicted += arc_adjust_impl(arc_mfu_ghost, 0, delta, type); } /* * If after attempting to make the requested adjustment to the ARC * the meta limit is still being exceeded then request that the * higher layers drop some cached objects which have holds on ARC * meta buffers. Requests to the upper layers will be made with * increasingly large scan sizes until the ARC is below the limit. */ if (arc_meta_used > arc_meta_limit) { if (type == ARC_BUFC_DATA) { type = ARC_BUFC_METADATA; } else { type = ARC_BUFC_DATA; if (zfs_arc_meta_prune) { prune += zfs_arc_meta_prune; arc_prune_async(prune); } } if (restarts > 0) { restarts--; goto restart; } } return (total_evicted); } /* * Evict metadata buffers from the cache, such that arc_meta_used is * capped by the arc_meta_limit tunable. */ static uint64_t arc_adjust_meta_only(void) { uint64_t total_evicted = 0; int64_t target; /* * If we're over the meta limit, we want to evict enough * metadata to get back under the meta limit. We don't want to * evict so much that we drop the MRU below arc_p, though. If * we're over the meta limit more than we're over arc_p, we * evict some from the MRU here, and some from the MFU below. */ target = MIN((int64_t)(arc_meta_used - arc_meta_limit), (int64_t)(refcount_count(&arc_anon->arcs_size) + refcount_count(&arc_mru->arcs_size) - arc_p)); total_evicted += arc_adjust_impl(arc_mru, 0, target, ARC_BUFC_METADATA); /* * Similar to the above, we want to evict enough bytes to get us * below the meta limit, but not so much as to drop us below the * space alloted to the MFU (which is defined as arc_c - arc_p). */ target = MIN((int64_t)(arc_meta_used - arc_meta_limit), (int64_t)(refcount_count(&arc_mfu->arcs_size) - (arc_c - arc_p))); total_evicted += arc_adjust_impl(arc_mfu, 0, target, ARC_BUFC_METADATA); return (total_evicted); } static uint64_t arc_adjust_meta(void) { if (zfs_arc_meta_strategy == ARC_STRATEGY_META_ONLY) return (arc_adjust_meta_only()); else return (arc_adjust_meta_balanced()); } /* * Return the type of the oldest buffer in the given arc state * * This function will select a random sublist of type ARC_BUFC_DATA and * a random sublist of type ARC_BUFC_METADATA. The tail of each sublist * is compared, and the type which contains the "older" buffer will be * returned. */ static arc_buf_contents_t arc_adjust_type(arc_state_t *state) { multilist_t *data_ml = &state->arcs_list[ARC_BUFC_DATA]; multilist_t *meta_ml = &state->arcs_list[ARC_BUFC_METADATA]; int data_idx = multilist_get_random_index(data_ml); int meta_idx = multilist_get_random_index(meta_ml); multilist_sublist_t *data_mls; multilist_sublist_t *meta_mls; arc_buf_contents_t type; arc_buf_hdr_t *data_hdr; arc_buf_hdr_t *meta_hdr; /* * We keep the sublist lock until we're finished, to prevent * the headers from being destroyed via arc_evict_state(). */ data_mls = multilist_sublist_lock(data_ml, data_idx); meta_mls = multilist_sublist_lock(meta_ml, meta_idx); /* * These two loops are to ensure we skip any markers that * might be at the tail of the lists due to arc_evict_state(). */ for (data_hdr = multilist_sublist_tail(data_mls); data_hdr != NULL; data_hdr = multilist_sublist_prev(data_mls, data_hdr)) { if (data_hdr->b_spa != 0) break; } for (meta_hdr = multilist_sublist_tail(meta_mls); meta_hdr != NULL; meta_hdr = multilist_sublist_prev(meta_mls, meta_hdr)) { if (meta_hdr->b_spa != 0) break; } if (data_hdr == NULL && meta_hdr == NULL) { type = ARC_BUFC_DATA; } else if (data_hdr == NULL) { ASSERT3P(meta_hdr, !=, NULL); type = ARC_BUFC_METADATA; } else if (meta_hdr == NULL) { ASSERT3P(data_hdr, !=, NULL); type = ARC_BUFC_DATA; } else { ASSERT3P(data_hdr, !=, NULL); ASSERT3P(meta_hdr, !=, NULL); /* The headers can't be on the sublist without an L1 header */ ASSERT(HDR_HAS_L1HDR(data_hdr)); ASSERT(HDR_HAS_L1HDR(meta_hdr)); if (data_hdr->b_l1hdr.b_arc_access < meta_hdr->b_l1hdr.b_arc_access) { type = ARC_BUFC_DATA; } else { type = ARC_BUFC_METADATA; } } multilist_sublist_unlock(meta_mls); multilist_sublist_unlock(data_mls); return (type); } /* * Evict buffers from the cache, such that arc_size is capped by arc_c. */ static uint64_t arc_adjust(void) { uint64_t total_evicted = 0; uint64_t bytes; int64_t target; /* * If we're over arc_meta_limit, we want to correct that before * potentially evicting data buffers below. */ total_evicted += arc_adjust_meta(); /* * Adjust MRU size * * If we're over the target cache size, we want to evict enough * from the list to get back to our target size. We don't want * to evict too much from the MRU, such that it drops below * arc_p. So, if we're over our target cache size more than * the MRU is over arc_p, we'll evict enough to get back to * arc_p here, and then evict more from the MFU below. */ target = MIN((int64_t)(arc_size - arc_c), (int64_t)(refcount_count(&arc_anon->arcs_size) + refcount_count(&arc_mru->arcs_size) + arc_meta_used - arc_p)); /* * If we're below arc_meta_min, always prefer to evict data. * Otherwise, try to satisfy the requested number of bytes to * evict from the type which contains older buffers; in an * effort to keep newer buffers in the cache regardless of their * type. If we cannot satisfy the number of bytes from this * type, spill over into the next type. */ if (arc_adjust_type(arc_mru) == ARC_BUFC_METADATA && arc_meta_used > arc_meta_min) { bytes = arc_adjust_impl(arc_mru, 0, target, ARC_BUFC_METADATA); total_evicted += bytes; /* * If we couldn't evict our target number of bytes from * metadata, we try to get the rest from data. */ target -= bytes; total_evicted += arc_adjust_impl(arc_mru, 0, target, ARC_BUFC_DATA); } else { bytes = arc_adjust_impl(arc_mru, 0, target, ARC_BUFC_DATA); total_evicted += bytes; /* * If we couldn't evict our target number of bytes from * data, we try to get the rest from metadata. */ target -= bytes; total_evicted += arc_adjust_impl(arc_mru, 0, target, ARC_BUFC_METADATA); } /* * Adjust MFU size * * Now that we've tried to evict enough from the MRU to get its * size back to arc_p, if we're still above the target cache * size, we evict the rest from the MFU. */ target = arc_size - arc_c; if (arc_adjust_type(arc_mfu) == ARC_BUFC_METADATA && arc_meta_used > arc_meta_min) { bytes = arc_adjust_impl(arc_mfu, 0, target, ARC_BUFC_METADATA); total_evicted += bytes; /* * If we couldn't evict our target number of bytes from * metadata, we try to get the rest from data. */ target -= bytes; total_evicted += arc_adjust_impl(arc_mfu, 0, target, ARC_BUFC_DATA); } else { bytes = arc_adjust_impl(arc_mfu, 0, target, ARC_BUFC_DATA); total_evicted += bytes; /* * If we couldn't evict our target number of bytes from * data, we try to get the rest from data. */ target -= bytes; total_evicted += arc_adjust_impl(arc_mfu, 0, target, ARC_BUFC_METADATA); } /* * Adjust ghost lists * * In addition to the above, the ARC also defines target values * for the ghost lists. The sum of the mru list and mru ghost * list should never exceed the target size of the cache, and * the sum of the mru list, mfu list, mru ghost list, and mfu * ghost list should never exceed twice the target size of the * cache. The following logic enforces these limits on the ghost * caches, and evicts from them as needed. */ target = refcount_count(&arc_mru->arcs_size) + refcount_count(&arc_mru_ghost->arcs_size) - arc_c; bytes = arc_adjust_impl(arc_mru_ghost, 0, target, ARC_BUFC_DATA); total_evicted += bytes; target -= bytes; total_evicted += arc_adjust_impl(arc_mru_ghost, 0, target, ARC_BUFC_METADATA); /* * We assume the sum of the mru list and mfu list is less than * or equal to arc_c (we enforced this above), which means we * can use the simpler of the two equations below: * * mru + mfu + mru ghost + mfu ghost <= 2 * arc_c * mru ghost + mfu ghost <= arc_c */ target = refcount_count(&arc_mru_ghost->arcs_size) + refcount_count(&arc_mfu_ghost->arcs_size) - arc_c; bytes = arc_adjust_impl(arc_mfu_ghost, 0, target, ARC_BUFC_DATA); total_evicted += bytes; target -= bytes; total_evicted += arc_adjust_impl(arc_mfu_ghost, 0, target, ARC_BUFC_METADATA); return (total_evicted); } static void arc_do_user_evicts(void) { mutex_enter(&arc_user_evicts_lock); while (arc_eviction_list != NULL) { arc_buf_t *buf = arc_eviction_list; arc_eviction_list = buf->b_next; mutex_enter(&buf->b_evict_lock); buf->b_hdr = NULL; mutex_exit(&buf->b_evict_lock); mutex_exit(&arc_user_evicts_lock); if (buf->b_efunc != NULL) VERIFY0(buf->b_efunc(buf->b_private)); buf->b_efunc = NULL; buf->b_private = NULL; kmem_cache_free(buf_cache, buf); mutex_enter(&arc_user_evicts_lock); } mutex_exit(&arc_user_evicts_lock); } void arc_flush(spa_t *spa, boolean_t retry) { uint64_t guid = 0; /* * If retry is TRUE, a spa must not be specified since we have * no good way to determine if all of a spa's buffers have been * evicted from an arc state. */ ASSERT(!retry || spa == 0); if (spa != NULL) guid = spa_load_guid(spa); (void) arc_flush_state(arc_mru, guid, ARC_BUFC_DATA, retry); (void) arc_flush_state(arc_mru, guid, ARC_BUFC_METADATA, retry); (void) arc_flush_state(arc_mfu, guid, ARC_BUFC_DATA, retry); (void) arc_flush_state(arc_mfu, guid, ARC_BUFC_METADATA, retry); (void) arc_flush_state(arc_mru_ghost, guid, ARC_BUFC_DATA, retry); (void) arc_flush_state(arc_mru_ghost, guid, ARC_BUFC_METADATA, retry); (void) arc_flush_state(arc_mfu_ghost, guid, ARC_BUFC_DATA, retry); (void) arc_flush_state(arc_mfu_ghost, guid, ARC_BUFC_METADATA, retry); arc_do_user_evicts(); ASSERT(spa || arc_eviction_list == NULL); } void arc_shrink(int64_t to_free) { if (arc_c > arc_c_min) { if (arc_c > arc_c_min + to_free) atomic_add_64(&arc_c, -to_free); else arc_c = arc_c_min; atomic_add_64(&arc_p, -(arc_p >> arc_shrink_shift)); if (arc_c > arc_size) arc_c = MAX(arc_size, arc_c_min); if (arc_p > arc_c) arc_p = (arc_c >> 1); ASSERT(arc_c >= arc_c_min); ASSERT((int64_t)arc_p >= 0); } if (arc_size > arc_c) (void) arc_adjust(); } typedef enum free_memory_reason_t { FMR_UNKNOWN, FMR_NEEDFREE, FMR_LOTSFREE, FMR_SWAPFS_MINFREE, FMR_PAGES_PP_MAXIMUM, FMR_HEAP_ARENA, FMR_ZIO_ARENA, } free_memory_reason_t; int64_t last_free_memory; free_memory_reason_t last_free_reason; #ifdef _KERNEL /* * Additional reserve of pages for pp_reserve. */ int64_t arc_pages_pp_reserve = 64; /* * Additional reserve of pages for swapfs. */ int64_t arc_swapfs_reserve = 64; #endif /* _KERNEL */ /* * Return the amount of memory that can be consumed before reclaim will be * needed. Positive if there is sufficient free memory, negative indicates * the amount of memory that needs to be freed up. */ static int64_t arc_available_memory(void) { int64_t lowest = INT64_MAX; free_memory_reason_t r = FMR_UNKNOWN; #ifdef _KERNEL int64_t n; #ifdef __linux__ pgcnt_t needfree = btop(arc_need_free); pgcnt_t lotsfree = btop(arc_sys_free); pgcnt_t desfree = 0; #endif if (needfree > 0) { n = PAGESIZE * (-needfree); if (n < lowest) { lowest = n; r = FMR_NEEDFREE; } } /* * check that we're out of range of the pageout scanner. It starts to * schedule paging if freemem is less than lotsfree and needfree. * lotsfree is the high-water mark for pageout, and needfree is the * number of needed free pages. We add extra pages here to make sure * the scanner doesn't start up while we're freeing memory. */ n = PAGESIZE * (freemem - lotsfree - needfree - desfree); if (n < lowest) { lowest = n; r = FMR_LOTSFREE; } #ifndef __linux__ /* * check to make sure that swapfs has enough space so that anon * reservations can still succeed. anon_resvmem() checks that the * availrmem is greater than swapfs_minfree, and the number of reserved * swap pages. We also add a bit of extra here just to prevent * circumstances from getting really dire. */ n = PAGESIZE * (availrmem - swapfs_minfree - swapfs_reserve - desfree - arc_swapfs_reserve); if (n < lowest) { lowest = n; r = FMR_SWAPFS_MINFREE; } /* * Check that we have enough availrmem that memory locking (e.g., via * mlock(3C) or memcntl(2)) can still succeed. (pages_pp_maximum * stores the number of pages that cannot be locked; when availrmem * drops below pages_pp_maximum, page locking mechanisms such as * page_pp_lock() will fail.) */ n = PAGESIZE * (availrmem - pages_pp_maximum - arc_pages_pp_reserve); if (n < lowest) { lowest = n; r = FMR_PAGES_PP_MAXIMUM; } #endif #if defined(__i386) /* * If we're on an i386 platform, it's possible that we'll exhaust the * kernel heap space before we ever run out of available physical * memory. Most checks of the size of the heap_area compare against * tune.t_minarmem, which is the minimum available real memory that we * can have in the system. However, this is generally fixed at 25 pages * which is so low that it's useless. In this comparison, we seek to * calculate the total heap-size, and reclaim if more than 3/4ths of the * heap is allocated. (Or, in the calculation, if less than 1/4th is * free) */ n = vmem_size(heap_arena, VMEM_FREE) - (vmem_size(heap_arena, VMEM_FREE | VMEM_ALLOC) >> 2); if (n < lowest) { lowest = n; r = FMR_HEAP_ARENA; } #endif /* * If zio data pages are being allocated out of a separate heap segment, * then enforce that the size of available vmem for this arena remains * above about 1/16th free. * * Note: The 1/16th arena free requirement was put in place * to aggressively evict memory from the arc in order to avoid * memory fragmentation issues. */ if (zio_arena != NULL) { n = vmem_size(zio_arena, VMEM_FREE) - (vmem_size(zio_arena, VMEM_ALLOC) >> 4); if (n < lowest) { lowest = n; r = FMR_ZIO_ARENA; } } #else /* _KERNEL */ /* Every 100 calls, free a small amount */ if (spa_get_random(100) == 0) lowest = -1024; #endif /* _KERNEL */ last_free_memory = lowest; last_free_reason = r; return (lowest); } /* * Determine if the system is under memory pressure and is asking * to reclaim memory. A return value of TRUE indicates that the system * is under memory pressure and that the arc should adjust accordingly. */ static boolean_t arc_reclaim_needed(void) { return (arc_available_memory() < 0); } static void arc_kmem_reap_now(void) { size_t i; kmem_cache_t *prev_cache = NULL; kmem_cache_t *prev_data_cache = NULL; extern kmem_cache_t *zio_buf_cache[]; extern kmem_cache_t *zio_data_buf_cache[]; extern kmem_cache_t *range_seg_cache; if ((arc_meta_used >= arc_meta_limit) && zfs_arc_meta_prune) { /* * We are exceeding our meta-data cache limit. * Prune some entries to release holds on meta-data. */ arc_prune_async(zfs_arc_meta_prune); } for (i = 0; i < SPA_MAXBLOCKSIZE >> SPA_MINBLOCKSHIFT; i++) { #ifdef _ILP32 /* reach upper limit of cache size on 32-bit */ if (zio_buf_cache[i] == NULL) break; #endif if (zio_buf_cache[i] != prev_cache) { prev_cache = zio_buf_cache[i]; kmem_cache_reap_now(zio_buf_cache[i]); } if (zio_data_buf_cache[i] != prev_data_cache) { prev_data_cache = zio_data_buf_cache[i]; kmem_cache_reap_now(zio_data_buf_cache[i]); } } kmem_cache_reap_now(buf_cache); kmem_cache_reap_now(hdr_full_cache); kmem_cache_reap_now(hdr_l2only_cache); kmem_cache_reap_now(range_seg_cache); if (zio_arena != NULL) { /* * Ask the vmem arena to reclaim unused memory from its * quantum caches. */ vmem_qcache_reap(zio_arena); } } /* * Threads can block in arc_get_data_buf() waiting for this thread to evict * enough data and signal them to proceed. When this happens, the threads in * arc_get_data_buf() are sleeping while holding the hash lock for their * particular arc header. Thus, we must be careful to never sleep on a * hash lock in this thread. This is to prevent the following deadlock: * * - Thread A sleeps on CV in arc_get_data_buf() holding hash lock "L", * waiting for the reclaim thread to signal it. * * - arc_reclaim_thread() tries to acquire hash lock "L" using mutex_enter, * fails, and goes to sleep forever. * * This possible deadlock is avoided by always acquiring a hash lock * using mutex_tryenter() from arc_reclaim_thread(). */ static void arc_reclaim_thread(void) { fstrans_cookie_t cookie = spl_fstrans_mark(); clock_t growtime = 0; callb_cpr_t cpr; CALLB_CPR_INIT(&cpr, &arc_reclaim_lock, callb_generic_cpr, FTAG); mutex_enter(&arc_reclaim_lock); while (!arc_reclaim_thread_exit) { int64_t to_free; int64_t free_memory = arc_available_memory(); uint64_t evicted = 0; arc_tuning_update(); mutex_exit(&arc_reclaim_lock); if (free_memory < 0) { arc_no_grow = B_TRUE; arc_warm = B_TRUE; /* * Wait at least zfs_grow_retry (default 5) seconds * before considering growing. */ growtime = ddi_get_lbolt() + (arc_grow_retry * hz); arc_kmem_reap_now(); /* * If we are still low on memory, shrink the ARC * so that we have arc_shrink_min free space. */ free_memory = arc_available_memory(); to_free = (arc_c >> arc_shrink_shift) - free_memory; if (to_free > 0) { #ifdef _KERNEL to_free = MAX(to_free, arc_need_free); #endif arc_shrink(to_free); } } else if (free_memory < arc_c >> arc_no_grow_shift) { arc_no_grow = B_TRUE; } else if (ddi_get_lbolt() >= growtime) { arc_no_grow = B_FALSE; } evicted = arc_adjust(); mutex_enter(&arc_reclaim_lock); /* * If evicted is zero, we couldn't evict anything via * arc_adjust(). This could be due to hash lock * collisions, but more likely due to the majority of * arc buffers being unevictable. Therefore, even if * arc_size is above arc_c, another pass is unlikely to * be helpful and could potentially cause us to enter an * infinite loop. */ if (arc_size <= arc_c || evicted == 0) { /* * We're either no longer overflowing, or we * can't evict anything more, so we should wake * up any threads before we go to sleep and clear * arc_need_free since nothing more can be done. */ cv_broadcast(&arc_reclaim_waiters_cv); arc_need_free = 0; /* * Block until signaled, or after one second (we * might need to perform arc_kmem_reap_now() * even if we aren't being signalled) */ CALLB_CPR_SAFE_BEGIN(&cpr); (void) cv_timedwait_sig(&arc_reclaim_thread_cv, &arc_reclaim_lock, ddi_get_lbolt() + hz); CALLB_CPR_SAFE_END(&cpr, &arc_reclaim_lock); } } arc_reclaim_thread_exit = FALSE; cv_broadcast(&arc_reclaim_thread_cv); CALLB_CPR_EXIT(&cpr); /* drops arc_reclaim_lock */ spl_fstrans_unmark(cookie); thread_exit(); } static void arc_user_evicts_thread(void) { fstrans_cookie_t cookie = spl_fstrans_mark(); callb_cpr_t cpr; CALLB_CPR_INIT(&cpr, &arc_user_evicts_lock, callb_generic_cpr, FTAG); mutex_enter(&arc_user_evicts_lock); while (!arc_user_evicts_thread_exit) { mutex_exit(&arc_user_evicts_lock); arc_do_user_evicts(); /* * This is necessary in order for the mdb ::arc dcmd to * show up to date information. Since the ::arc command * does not call the kstat's update function, without * this call, the command may show stale stats for the * anon, mru, mru_ghost, mfu, and mfu_ghost lists. Even * with this change, the data might be up to 1 second * out of date; but that should suffice. The arc_state_t * structures can be queried directly if more accurate * information is needed. */ if (arc_ksp != NULL) arc_ksp->ks_update(arc_ksp, KSTAT_READ); mutex_enter(&arc_user_evicts_lock); /* * Block until signaled, or after one second (we need to * call the arc's kstat update function regularly). */ CALLB_CPR_SAFE_BEGIN(&cpr); (void) cv_timedwait_sig(&arc_user_evicts_cv, &arc_user_evicts_lock, ddi_get_lbolt() + hz); CALLB_CPR_SAFE_END(&cpr, &arc_user_evicts_lock); } arc_user_evicts_thread_exit = FALSE; cv_broadcast(&arc_user_evicts_cv); CALLB_CPR_EXIT(&cpr); /* drops arc_user_evicts_lock */ spl_fstrans_unmark(cookie); thread_exit(); } #ifdef _KERNEL /* * Determine the amount of memory eligible for eviction contained in the * ARC. All clean data reported by the ghost lists can always be safely * evicted. Due to arc_c_min, the same does not hold for all clean data * contained by the regular mru and mfu lists. * * In the case of the regular mru and mfu lists, we need to report as * much clean data as possible, such that evicting that same reported * data will not bring arc_size below arc_c_min. Thus, in certain * circumstances, the total amount of clean data in the mru and mfu * lists might not actually be evictable. * * The following two distinct cases are accounted for: * * 1. The sum of the amount of dirty data contained by both the mru and * mfu lists, plus the ARC's other accounting (e.g. the anon list), * is greater than or equal to arc_c_min. * (i.e. amount of dirty data >= arc_c_min) * * This is the easy case; all clean data contained by the mru and mfu * lists is evictable. Evicting all clean data can only drop arc_size * to the amount of dirty data, which is greater than arc_c_min. * * 2. The sum of the amount of dirty data contained by both the mru and * mfu lists, plus the ARC's other accounting (e.g. the anon list), * is less than arc_c_min. * (i.e. arc_c_min > amount of dirty data) * * 2.1. arc_size is greater than or equal arc_c_min. * (i.e. arc_size >= arc_c_min > amount of dirty data) * * In this case, not all clean data from the regular mru and mfu * lists is actually evictable; we must leave enough clean data * to keep arc_size above arc_c_min. Thus, the maximum amount of * evictable data from the two lists combined, is exactly the * difference between arc_size and arc_c_min. * * 2.2. arc_size is less than arc_c_min * (i.e. arc_c_min > arc_size > amount of dirty data) * * In this case, none of the data contained in the mru and mfu * lists is evictable, even if it's clean. Since arc_size is * already below arc_c_min, evicting any more would only * increase this negative difference. */ static uint64_t arc_evictable_memory(void) { uint64_t arc_clean = arc_mru->arcs_lsize[ARC_BUFC_DATA] + arc_mru->arcs_lsize[ARC_BUFC_METADATA] + arc_mfu->arcs_lsize[ARC_BUFC_DATA] + arc_mfu->arcs_lsize[ARC_BUFC_METADATA]; uint64_t ghost_clean = arc_mru_ghost->arcs_lsize[ARC_BUFC_DATA] + arc_mru_ghost->arcs_lsize[ARC_BUFC_METADATA] + arc_mfu_ghost->arcs_lsize[ARC_BUFC_DATA] + arc_mfu_ghost->arcs_lsize[ARC_BUFC_METADATA]; uint64_t arc_dirty = MAX((int64_t)arc_size - (int64_t)arc_clean, 0); if (arc_dirty >= arc_c_min) return (ghost_clean + arc_clean); return (ghost_clean + MAX((int64_t)arc_size - (int64_t)arc_c_min, 0)); } /* * If sc->nr_to_scan is zero, the caller is requesting a query of the * number of objects which can potentially be freed. If it is nonzero, * the request is to free that many objects. * * Linux kernels >= 3.12 have the count_objects and scan_objects callbacks * in struct shrinker and also require the shrinker to return the number * of objects freed. * * Older kernels require the shrinker to return the number of freeable * objects following the freeing of nr_to_free. */ static spl_shrinker_t __arc_shrinker_func(struct shrinker *shrink, struct shrink_control *sc) { int64_t pages; /* The arc is considered warm once reclaim has occurred */ if (unlikely(arc_warm == B_FALSE)) arc_warm = B_TRUE; /* Return the potential number of reclaimable pages */ pages = btop((int64_t)arc_evictable_memory()); if (sc->nr_to_scan == 0) return (pages); /* Not allowed to perform filesystem reclaim */ if (!(sc->gfp_mask & __GFP_FS)) return (SHRINK_STOP); /* Reclaim in progress */ if (mutex_tryenter(&arc_reclaim_lock) == 0) return (SHRINK_STOP); mutex_exit(&arc_reclaim_lock); /* * Evict the requested number of pages by shrinking arc_c the * requested amount. If there is nothing left to evict just * reap whatever we can from the various arc slabs. */ if (pages > 0) { arc_shrink(ptob(sc->nr_to_scan)); arc_kmem_reap_now(); #ifdef HAVE_SPLIT_SHRINKER_CALLBACK pages = MAX(pages - btop(arc_evictable_memory()), 0); #else pages = btop(arc_evictable_memory()); #endif } else { arc_kmem_reap_now(); pages = SHRINK_STOP; } /* * We've reaped what we can, wake up threads. */ cv_broadcast(&arc_reclaim_waiters_cv); /* * When direct reclaim is observed it usually indicates a rapid * increase in memory pressure. This occurs because the kswapd * threads were unable to asynchronously keep enough free memory * available. In this case set arc_no_grow to briefly pause arc * growth to avoid compounding the memory pressure. */ if (current_is_kswapd()) { ARCSTAT_BUMP(arcstat_memory_indirect_count); } else { arc_no_grow = B_TRUE; arc_need_free = ptob(sc->nr_to_scan); ARCSTAT_BUMP(arcstat_memory_direct_count); } return (pages); } SPL_SHRINKER_CALLBACK_WRAPPER(arc_shrinker_func); SPL_SHRINKER_DECLARE(arc_shrinker, arc_shrinker_func, DEFAULT_SEEKS); #endif /* _KERNEL */ /* * Adapt arc info given the number of bytes we are trying to add and * the state that we are comming from. This function is only called * when we are adding new content to the cache. */ static void arc_adapt(int bytes, arc_state_t *state) { int mult; uint64_t arc_p_min = (arc_c >> arc_p_min_shift); int64_t mrug_size = refcount_count(&arc_mru_ghost->arcs_size); int64_t mfug_size = refcount_count(&arc_mfu_ghost->arcs_size); if (state == arc_l2c_only) return; ASSERT(bytes > 0); /* * Adapt the target size of the MRU list: * - if we just hit in the MRU ghost list, then increase * the target size of the MRU list. * - if we just hit in the MFU ghost list, then increase * the target size of the MFU list by decreasing the * target size of the MRU list. */ if (state == arc_mru_ghost) { mult = (mrug_size >= mfug_size) ? 1 : (mfug_size / mrug_size); if (!zfs_arc_p_dampener_disable) mult = MIN(mult, 10); /* avoid wild arc_p adjustment */ arc_p = MIN(arc_c - arc_p_min, arc_p + bytes * mult); } else if (state == arc_mfu_ghost) { uint64_t delta; mult = (mfug_size >= mrug_size) ? 1 : (mrug_size / mfug_size); if (!zfs_arc_p_dampener_disable) mult = MIN(mult, 10); delta = MIN(bytes * mult, arc_p); arc_p = MAX(arc_p_min, arc_p - delta); } ASSERT((int64_t)arc_p >= 0); if (arc_reclaim_needed()) { cv_signal(&arc_reclaim_thread_cv); return; } if (arc_no_grow) return; if (arc_c >= arc_c_max) return; /* * If we're within (2 * maxblocksize) bytes of the target * cache size, increment the target cache size */ ASSERT3U(arc_c, >=, 2ULL << SPA_MAXBLOCKSHIFT); arc_c = MAX(arc_c, 2ULL << SPA_MAXBLOCKSHIFT); if (arc_size >= arc_c - (2ULL << SPA_MAXBLOCKSHIFT)) { atomic_add_64(&arc_c, (int64_t)bytes); if (arc_c > arc_c_max) arc_c = arc_c_max; else if (state == arc_anon) atomic_add_64(&arc_p, (int64_t)bytes); if (arc_p > arc_c) arc_p = arc_c; } ASSERT((int64_t)arc_p >= 0); } /* * Check if arc_size has grown past our upper threshold, determined by * zfs_arc_overflow_shift. */ static boolean_t arc_is_overflowing(void) { /* Always allow at least one block of overflow */ uint64_t overflow = MAX(SPA_MAXBLOCKSIZE, arc_c >> zfs_arc_overflow_shift); return (arc_size >= arc_c + overflow); } /* * The buffer, supplied as the first argument, needs a data block. If we * are hitting the hard limit for the cache size, we must sleep, waiting * for the eviction thread to catch up. If we're past the target size * but below the hard limit, we'll only signal the reclaim thread and * continue on. */ static void arc_get_data_buf(arc_buf_t *buf) { arc_state_t *state = buf->b_hdr->b_l1hdr.b_state; uint64_t size = buf->b_hdr->b_size; arc_buf_contents_t type = arc_buf_type(buf->b_hdr); arc_adapt(size, state); /* * If arc_size is currently overflowing, and has grown past our * upper limit, we must be adding data faster than the evict * thread can evict. Thus, to ensure we don't compound the * problem by adding more data and forcing arc_size to grow even * further past it's target size, we halt and wait for the * eviction thread to catch up. * * It's also possible that the reclaim thread is unable to evict * enough buffers to get arc_size below the overflow limit (e.g. * due to buffers being un-evictable, or hash lock collisions). * In this case, we want to proceed regardless if we're * overflowing; thus we don't use a while loop here. */ if (arc_is_overflowing()) { mutex_enter(&arc_reclaim_lock); /* * Now that we've acquired the lock, we may no longer be * over the overflow limit, lets check. * * We're ignoring the case of spurious wake ups. If that * were to happen, it'd let this thread consume an ARC * buffer before it should have (i.e. before we're under * the overflow limit and were signalled by the reclaim * thread). As long as that is a rare occurrence, it * shouldn't cause any harm. */ if (arc_is_overflowing()) { cv_signal(&arc_reclaim_thread_cv); cv_wait(&arc_reclaim_waiters_cv, &arc_reclaim_lock); } mutex_exit(&arc_reclaim_lock); } if (type == ARC_BUFC_METADATA) { buf->b_data = zio_buf_alloc(size); arc_space_consume(size, ARC_SPACE_META); } else { ASSERT(type == ARC_BUFC_DATA); buf->b_data = zio_data_buf_alloc(size); arc_space_consume(size, ARC_SPACE_DATA); } /* * Update the state size. Note that ghost states have a * "ghost size" and so don't need to be updated. */ if (!GHOST_STATE(buf->b_hdr->b_l1hdr.b_state)) { arc_buf_hdr_t *hdr = buf->b_hdr; arc_state_t *state = hdr->b_l1hdr.b_state; (void) refcount_add_many(&state->arcs_size, size, buf); /* * If this is reached via arc_read, the link is * protected by the hash lock. If reached via * arc_buf_alloc, the header should not be accessed by * any other thread. And, if reached via arc_read_done, * the hash lock will protect it if it's found in the * hash table; otherwise no other thread should be * trying to [add|remove]_reference it. */ if (multilist_link_active(&hdr->b_l1hdr.b_arc_node)) { ASSERT(refcount_is_zero(&hdr->b_l1hdr.b_refcnt)); atomic_add_64(&hdr->b_l1hdr.b_state->arcs_lsize[type], size); } /* * If we are growing the cache, and we are adding anonymous * data, and we have outgrown arc_p, update arc_p */ if (arc_size < arc_c && hdr->b_l1hdr.b_state == arc_anon && (refcount_count(&arc_anon->arcs_size) + refcount_count(&arc_mru->arcs_size) > arc_p)) arc_p = MIN(arc_c, arc_p + size); } } /* * This routine is called whenever a buffer is accessed. * NOTE: the hash lock is dropped in this function. */ static void arc_access(arc_buf_hdr_t *hdr, kmutex_t *hash_lock) { clock_t now; ASSERT(MUTEX_HELD(hash_lock)); ASSERT(HDR_HAS_L1HDR(hdr)); if (hdr->b_l1hdr.b_state == arc_anon) { /* * This buffer is not in the cache, and does not * appear in our "ghost" list. Add the new buffer * to the MRU state. */ ASSERT0(hdr->b_l1hdr.b_arc_access); hdr->b_l1hdr.b_arc_access = ddi_get_lbolt(); DTRACE_PROBE1(new_state__mru, arc_buf_hdr_t *, hdr); arc_change_state(arc_mru, hdr, hash_lock); } else if (hdr->b_l1hdr.b_state == arc_mru) { now = ddi_get_lbolt(); /* * If this buffer is here because of a prefetch, then either: * - clear the flag if this is a "referencing" read * (any subsequent access will bump this into the MFU state). * or * - move the buffer to the head of the list if this is * another prefetch (to make it less likely to be evicted). */ if (HDR_PREFETCH(hdr)) { if (refcount_count(&hdr->b_l1hdr.b_refcnt) == 0) { /* link protected by hash lock */ ASSERT(multilist_link_active( &hdr->b_l1hdr.b_arc_node)); } else { hdr->b_flags &= ~ARC_FLAG_PREFETCH; atomic_inc_32(&hdr->b_l1hdr.b_mru_hits); ARCSTAT_BUMP(arcstat_mru_hits); } hdr->b_l1hdr.b_arc_access = now; return; } /* * This buffer has been "accessed" only once so far, * but it is still in the cache. Move it to the MFU * state. */ if (ddi_time_after(now, hdr->b_l1hdr.b_arc_access + ARC_MINTIME)) { /* * More than 125ms have passed since we * instantiated this buffer. Move it to the * most frequently used state. */ hdr->b_l1hdr.b_arc_access = now; DTRACE_PROBE1(new_state__mfu, arc_buf_hdr_t *, hdr); arc_change_state(arc_mfu, hdr, hash_lock); } atomic_inc_32(&hdr->b_l1hdr.b_mru_hits); ARCSTAT_BUMP(arcstat_mru_hits); } else if (hdr->b_l1hdr.b_state == arc_mru_ghost) { arc_state_t *new_state; /* * This buffer has been "accessed" recently, but * was evicted from the cache. Move it to the * MFU state. */ if (HDR_PREFETCH(hdr)) { new_state = arc_mru; if (refcount_count(&hdr->b_l1hdr.b_refcnt) > 0) hdr->b_flags &= ~ARC_FLAG_PREFETCH; DTRACE_PROBE1(new_state__mru, arc_buf_hdr_t *, hdr); } else { new_state = arc_mfu; DTRACE_PROBE1(new_state__mfu, arc_buf_hdr_t *, hdr); } hdr->b_l1hdr.b_arc_access = ddi_get_lbolt(); arc_change_state(new_state, hdr, hash_lock); atomic_inc_32(&hdr->b_l1hdr.b_mru_ghost_hits); ARCSTAT_BUMP(arcstat_mru_ghost_hits); } else if (hdr->b_l1hdr.b_state == arc_mfu) { /* * This buffer has been accessed more than once and is * still in the cache. Keep it in the MFU state. * * NOTE: an add_reference() that occurred when we did * the arc_read() will have kicked this off the list. * If it was a prefetch, we will explicitly move it to * the head of the list now. */ if ((HDR_PREFETCH(hdr)) != 0) { ASSERT(refcount_is_zero(&hdr->b_l1hdr.b_refcnt)); /* link protected by hash_lock */ ASSERT(multilist_link_active(&hdr->b_l1hdr.b_arc_node)); } atomic_inc_32(&hdr->b_l1hdr.b_mfu_hits); ARCSTAT_BUMP(arcstat_mfu_hits); hdr->b_l1hdr.b_arc_access = ddi_get_lbolt(); } else if (hdr->b_l1hdr.b_state == arc_mfu_ghost) { arc_state_t *new_state = arc_mfu; /* * This buffer has been accessed more than once but has * been evicted from the cache. Move it back to the * MFU state. */ if (HDR_PREFETCH(hdr)) { /* * This is a prefetch access... * move this block back to the MRU state. */ ASSERT0(refcount_count(&hdr->b_l1hdr.b_refcnt)); new_state = arc_mru; } hdr->b_l1hdr.b_arc_access = ddi_get_lbolt(); DTRACE_PROBE1(new_state__mfu, arc_buf_hdr_t *, hdr); arc_change_state(new_state, hdr, hash_lock); atomic_inc_32(&hdr->b_l1hdr.b_mfu_ghost_hits); ARCSTAT_BUMP(arcstat_mfu_ghost_hits); } else if (hdr->b_l1hdr.b_state == arc_l2c_only) { /* * This buffer is on the 2nd Level ARC. */ hdr->b_l1hdr.b_arc_access = ddi_get_lbolt(); DTRACE_PROBE1(new_state__mfu, arc_buf_hdr_t *, hdr); arc_change_state(arc_mfu, hdr, hash_lock); } else { cmn_err(CE_PANIC, "invalid arc state 0x%p", hdr->b_l1hdr.b_state); } } /* a generic arc_done_func_t which you can use */ /* ARGSUSED */ void arc_bcopy_func(zio_t *zio, arc_buf_t *buf, void *arg) { if (zio == NULL || zio->io_error == 0) bcopy(buf->b_data, arg, buf->b_hdr->b_size); VERIFY(arc_buf_remove_ref(buf, arg)); } /* a generic arc_done_func_t */ void arc_getbuf_func(zio_t *zio, arc_buf_t *buf, void *arg) { arc_buf_t **bufp = arg; if (zio && zio->io_error) { VERIFY(arc_buf_remove_ref(buf, arg)); *bufp = NULL; } else { *bufp = buf; ASSERT(buf->b_data); } } static void arc_read_done(zio_t *zio) { arc_buf_hdr_t *hdr; arc_buf_t *buf; arc_buf_t *abuf; /* buffer we're assigning to callback */ kmutex_t *hash_lock = NULL; arc_callback_t *callback_list, *acb; int freeable = FALSE; buf = zio->io_private; hdr = buf->b_hdr; /* * The hdr was inserted into hash-table and removed from lists * prior to starting I/O. We should find this header, since * it's in the hash table, and it should be legit since it's * not possible to evict it during the I/O. The only possible * reason for it not to be found is if we were freed during the * read. */ if (HDR_IN_HASH_TABLE(hdr)) { arc_buf_hdr_t *found; ASSERT3U(hdr->b_birth, ==, BP_PHYSICAL_BIRTH(zio->io_bp)); ASSERT3U(hdr->b_dva.dva_word[0], ==, BP_IDENTITY(zio->io_bp)->dva_word[0]); ASSERT3U(hdr->b_dva.dva_word[1], ==, BP_IDENTITY(zio->io_bp)->dva_word[1]); found = buf_hash_find(hdr->b_spa, zio->io_bp, &hash_lock); ASSERT((found == NULL && HDR_FREED_IN_READ(hdr) && hash_lock == NULL) || (found == hdr && DVA_EQUAL(&hdr->b_dva, BP_IDENTITY(zio->io_bp))) || (found == hdr && HDR_L2_READING(hdr))); } hdr->b_flags &= ~ARC_FLAG_L2_EVICTED; if (l2arc_noprefetch && HDR_PREFETCH(hdr)) hdr->b_flags &= ~ARC_FLAG_L2CACHE; /* byteswap if necessary */ callback_list = hdr->b_l1hdr.b_acb; ASSERT(callback_list != NULL); if (BP_SHOULD_BYTESWAP(zio->io_bp) && zio->io_error == 0) { dmu_object_byteswap_t bswap = DMU_OT_BYTESWAP(BP_GET_TYPE(zio->io_bp)); if (BP_GET_LEVEL(zio->io_bp) > 0) byteswap_uint64_array(buf->b_data, hdr->b_size); else dmu_ot_byteswap[bswap].ob_func(buf->b_data, hdr->b_size); } arc_cksum_compute(buf, B_FALSE); arc_buf_watch(buf); if (hash_lock && zio->io_error == 0 && hdr->b_l1hdr.b_state == arc_anon) { /* * Only call arc_access on anonymous buffers. This is because * if we've issued an I/O for an evicted buffer, we've already * called arc_access (to prevent any simultaneous readers from * getting confused). */ arc_access(hdr, hash_lock); } /* create copies of the data buffer for the callers */ abuf = buf; for (acb = callback_list; acb; acb = acb->acb_next) { if (acb->acb_done) { if (abuf == NULL) { ARCSTAT_BUMP(arcstat_duplicate_reads); abuf = arc_buf_clone(buf); } acb->acb_buf = abuf; abuf = NULL; } } hdr->b_l1hdr.b_acb = NULL; hdr->b_flags &= ~ARC_FLAG_IO_IN_PROGRESS; ASSERT(!HDR_BUF_AVAILABLE(hdr)); if (abuf == buf) { ASSERT(buf->b_efunc == NULL); ASSERT(hdr->b_l1hdr.b_datacnt == 1); hdr->b_flags |= ARC_FLAG_BUF_AVAILABLE; } ASSERT(refcount_is_zero(&hdr->b_l1hdr.b_refcnt) || callback_list != NULL); if (zio->io_error != 0) { hdr->b_flags |= ARC_FLAG_IO_ERROR; if (hdr->b_l1hdr.b_state != arc_anon) arc_change_state(arc_anon, hdr, hash_lock); if (HDR_IN_HASH_TABLE(hdr)) buf_hash_remove(hdr); freeable = refcount_is_zero(&hdr->b_l1hdr.b_refcnt); } /* * Broadcast before we drop the hash_lock to avoid the possibility * that the hdr (and hence the cv) might be freed before we get to * the cv_broadcast(). */ cv_broadcast(&hdr->b_l1hdr.b_cv); if (hash_lock != NULL) { mutex_exit(hash_lock); } else { /* * This block was freed while we waited for the read to * complete. It has been removed from the hash table and * moved to the anonymous state (so that it won't show up * in the cache). */ ASSERT3P(hdr->b_l1hdr.b_state, ==, arc_anon); freeable = refcount_is_zero(&hdr->b_l1hdr.b_refcnt); } /* execute each callback and free its structure */ while ((acb = callback_list) != NULL) { if (acb->acb_done) acb->acb_done(zio, acb->acb_buf, acb->acb_private); if (acb->acb_zio_dummy != NULL) { acb->acb_zio_dummy->io_error = zio->io_error; zio_nowait(acb->acb_zio_dummy); } callback_list = acb->acb_next; kmem_free(acb, sizeof (arc_callback_t)); } if (freeable) arc_hdr_destroy(hdr); } /* * "Read" the block at the specified DVA (in bp) via the * cache. If the block is found in the cache, invoke the provided * callback immediately and return. Note that the `zio' parameter * in the callback will be NULL in this case, since no IO was * required. If the block is not in the cache pass the read request * on to the spa with a substitute callback function, so that the * requested block will be added to the cache. * * If a read request arrives for a block that has a read in-progress, * either wait for the in-progress read to complete (and return the * results); or, if this is a read with a "done" func, add a record * to the read to invoke the "done" func when the read completes, * and return; or just return. * * arc_read_done() will invoke all the requested "done" functions * for readers of this block. */ int arc_read(zio_t *pio, spa_t *spa, const blkptr_t *bp, arc_done_func_t *done, void *private, zio_priority_t priority, int zio_flags, arc_flags_t *arc_flags, const zbookmark_phys_t *zb) { arc_buf_hdr_t *hdr = NULL; arc_buf_t *buf = NULL; kmutex_t *hash_lock = NULL; zio_t *rzio; uint64_t guid = spa_load_guid(spa); int rc = 0; ASSERT(!BP_IS_EMBEDDED(bp) || BPE_GET_ETYPE(bp) == BP_EMBEDDED_TYPE_DATA); top: if (!BP_IS_EMBEDDED(bp)) { /* * Embedded BP's have no DVA and require no I/O to "read". * Create an anonymous arc buf to back it. */ hdr = buf_hash_find(guid, bp, &hash_lock); } if (hdr != NULL && HDR_HAS_L1HDR(hdr) && hdr->b_l1hdr.b_datacnt > 0) { *arc_flags |= ARC_FLAG_CACHED; if (HDR_IO_IN_PROGRESS(hdr)) { + if ((hdr->b_flags & ARC_FLAG_PRIO_ASYNC_READ) && + priority == ZIO_PRIORITY_SYNC_READ) { + /* + * This sync read must wait for an + * in-progress async read (e.g. a predictive + * prefetch). Async reads are queued + * separately at the vdev_queue layer, so + * this is a form of priority inversion. + * Ideally, we would "inherit" the demand + * i/o's priority by moving the i/o from + * the async queue to the synchronous queue, + * but there is currently no mechanism to do + * so. Track this so that we can evaluate + * the magnitude of this potential performance + * problem. + * + * Note that if the prefetch i/o is already + * active (has been issued to the device), + * the prefetch improved performance, because + * we issued it sooner than we would have + * without the prefetch. + */ + DTRACE_PROBE1(arc__sync__wait__for__async, + arc_buf_hdr_t *, hdr); + ARCSTAT_BUMP(arcstat_sync_wait_for_async); + } + if (hdr->b_flags & ARC_FLAG_PREDICTIVE_PREFETCH) { + hdr->b_flags &= ~ARC_FLAG_PREDICTIVE_PREFETCH; + } + if (*arc_flags & ARC_FLAG_WAIT) { cv_wait(&hdr->b_l1hdr.b_cv, hash_lock); mutex_exit(hash_lock); goto top; } ASSERT(*arc_flags & ARC_FLAG_NOWAIT); if (done) { - arc_callback_t *acb = NULL; + arc_callback_t *acb = NULL; acb = kmem_zalloc(sizeof (arc_callback_t), KM_SLEEP); acb->acb_done = done; acb->acb_private = private; if (pio != NULL) acb->acb_zio_dummy = zio_null(pio, spa, NULL, NULL, NULL, zio_flags); ASSERT(acb->acb_done != NULL); acb->acb_next = hdr->b_l1hdr.b_acb; hdr->b_l1hdr.b_acb = acb; add_reference(hdr, hash_lock, private); mutex_exit(hash_lock); goto out; } mutex_exit(hash_lock); goto out; } ASSERT(hdr->b_l1hdr.b_state == arc_mru || hdr->b_l1hdr.b_state == arc_mfu); if (done) { + if (hdr->b_flags & ARC_FLAG_PREDICTIVE_PREFETCH) { + /* + * This is a demand read which does not have to + * wait for i/o because we did a predictive + * prefetch i/o for it, which has completed. + */ + DTRACE_PROBE1( + arc__demand__hit__predictive__prefetch, + arc_buf_hdr_t *, hdr); + ARCSTAT_BUMP( + arcstat_demand_hit_predictive_prefetch); + hdr->b_flags &= ~ARC_FLAG_PREDICTIVE_PREFETCH; + } add_reference(hdr, hash_lock, private); /* * If this block is already in use, create a new * copy of the data so that we will be guaranteed * that arc_release() will always succeed. */ buf = hdr->b_l1hdr.b_buf; ASSERT(buf); ASSERT(buf->b_data); if (HDR_BUF_AVAILABLE(hdr)) { ASSERT(buf->b_efunc == NULL); hdr->b_flags &= ~ARC_FLAG_BUF_AVAILABLE; } else { buf = arc_buf_clone(buf); } } else if (*arc_flags & ARC_FLAG_PREFETCH && refcount_count(&hdr->b_l1hdr.b_refcnt) == 0) { hdr->b_flags |= ARC_FLAG_PREFETCH; } DTRACE_PROBE1(arc__hit, arc_buf_hdr_t *, hdr); arc_access(hdr, hash_lock); if (*arc_flags & ARC_FLAG_L2CACHE) hdr->b_flags |= ARC_FLAG_L2CACHE; if (*arc_flags & ARC_FLAG_L2COMPRESS) hdr->b_flags |= ARC_FLAG_L2COMPRESS; mutex_exit(hash_lock); ARCSTAT_BUMP(arcstat_hits); ARCSTAT_CONDSTAT(!HDR_PREFETCH(hdr), demand, prefetch, !HDR_ISTYPE_METADATA(hdr), data, metadata, hits); if (done) done(NULL, buf, private); } else { uint64_t size = BP_GET_LSIZE(bp); arc_callback_t *acb; vdev_t *vd = NULL; uint64_t addr = 0; boolean_t devw = B_FALSE; enum zio_compress b_compress = ZIO_COMPRESS_OFF; int32_t b_asize = 0; /* * Gracefully handle a damaged logical block size as a * checksum error. */ if (size > spa_maxblocksize(spa)) { ASSERT3P(buf, ==, NULL); rc = SET_ERROR(ECKSUM); goto out; } if (hdr == NULL) { /* this block is not in the cache */ arc_buf_hdr_t *exists = NULL; arc_buf_contents_t type = BP_GET_BUFC_TYPE(bp); buf = arc_buf_alloc(spa, size, private, type); hdr = buf->b_hdr; if (!BP_IS_EMBEDDED(bp)) { hdr->b_dva = *BP_IDENTITY(bp); hdr->b_birth = BP_PHYSICAL_BIRTH(bp); exists = buf_hash_insert(hdr, &hash_lock); } if (exists != NULL) { /* somebody beat us to the hash insert */ mutex_exit(hash_lock); buf_discard_identity(hdr); (void) arc_buf_remove_ref(buf, private); goto top; /* restart the IO request */ } - /* if this is a prefetch, we don't have a reference */ - if (*arc_flags & ARC_FLAG_PREFETCH) { + /* + * If there is a callback, we pass our reference to + * it; otherwise we remove our reference. + */ + if (done == NULL) { (void) remove_reference(hdr, hash_lock, private); - hdr->b_flags |= ARC_FLAG_PREFETCH; } + if (*arc_flags & ARC_FLAG_PREFETCH) + hdr->b_flags |= ARC_FLAG_PREFETCH; if (*arc_flags & ARC_FLAG_L2CACHE) hdr->b_flags |= ARC_FLAG_L2CACHE; if (*arc_flags & ARC_FLAG_L2COMPRESS) hdr->b_flags |= ARC_FLAG_L2COMPRESS; if (BP_GET_LEVEL(bp) > 0) hdr->b_flags |= ARC_FLAG_INDIRECT; } else { /* * This block is in the ghost cache. If it was L2-only * (and thus didn't have an L1 hdr), we realloc the * header to add an L1 hdr. */ if (!HDR_HAS_L1HDR(hdr)) { hdr = arc_hdr_realloc(hdr, hdr_l2only_cache, hdr_full_cache); } ASSERT(GHOST_STATE(hdr->b_l1hdr.b_state)); ASSERT(!HDR_IO_IN_PROGRESS(hdr)); ASSERT(refcount_is_zero(&hdr->b_l1hdr.b_refcnt)); ASSERT3P(hdr->b_l1hdr.b_buf, ==, NULL); - /* if this is a prefetch, we don't have a reference */ + /* + * If there is a callback, we pass a reference to it. + */ + if (done != NULL) + add_reference(hdr, hash_lock, private); if (*arc_flags & ARC_FLAG_PREFETCH) hdr->b_flags |= ARC_FLAG_PREFETCH; - else - add_reference(hdr, hash_lock, private); if (*arc_flags & ARC_FLAG_L2CACHE) hdr->b_flags |= ARC_FLAG_L2CACHE; if (*arc_flags & ARC_FLAG_L2COMPRESS) hdr->b_flags |= ARC_FLAG_L2COMPRESS; buf = kmem_cache_alloc(buf_cache, KM_PUSHPAGE); buf->b_hdr = hdr; buf->b_data = NULL; buf->b_efunc = NULL; buf->b_private = NULL; buf->b_next = NULL; hdr->b_l1hdr.b_buf = buf; ASSERT0(hdr->b_l1hdr.b_datacnt); hdr->b_l1hdr.b_datacnt = 1; arc_get_data_buf(buf); arc_access(hdr, hash_lock); } + if (*arc_flags & ARC_FLAG_PREDICTIVE_PREFETCH) + hdr->b_flags |= ARC_FLAG_PREDICTIVE_PREFETCH; ASSERT(!GHOST_STATE(hdr->b_l1hdr.b_state)); acb = kmem_zalloc(sizeof (arc_callback_t), KM_SLEEP); acb->acb_done = done; acb->acb_private = private; ASSERT(hdr->b_l1hdr.b_acb == NULL); hdr->b_l1hdr.b_acb = acb; hdr->b_flags |= ARC_FLAG_IO_IN_PROGRESS; if (HDR_HAS_L2HDR(hdr) && (vd = hdr->b_l2hdr.b_dev->l2ad_vdev) != NULL) { devw = hdr->b_l2hdr.b_dev->l2ad_writing; addr = hdr->b_l2hdr.b_daddr; b_compress = hdr->b_l2hdr.b_compress; b_asize = hdr->b_l2hdr.b_asize; /* * Lock out device removal. */ if (vdev_is_dead(vd) || !spa_config_tryenter(spa, SCL_L2ARC, vd, RW_READER)) vd = NULL; } if (hash_lock != NULL) mutex_exit(hash_lock); /* * At this point, we have a level 1 cache miss. Try again in * L2ARC if possible. */ ASSERT3U(hdr->b_size, ==, size); DTRACE_PROBE4(arc__miss, arc_buf_hdr_t *, hdr, blkptr_t *, bp, uint64_t, size, zbookmark_phys_t *, zb); ARCSTAT_BUMP(arcstat_misses); ARCSTAT_CONDSTAT(!HDR_PREFETCH(hdr), demand, prefetch, !HDR_ISTYPE_METADATA(hdr), data, metadata, misses); + if (priority == ZIO_PRIORITY_ASYNC_READ) + hdr->b_flags |= ARC_FLAG_PRIO_ASYNC_READ; + else + hdr->b_flags &= ~ARC_FLAG_PRIO_ASYNC_READ; + if (vd != NULL && l2arc_ndev != 0 && !(l2arc_norw && devw)) { /* * Read from the L2ARC if the following are true: * 1. The L2ARC vdev was previously cached. * 2. This buffer still has L2ARC metadata. * 3. This buffer isn't currently writing to the L2ARC. * 4. The L2ARC entry wasn't evicted, which may * also have invalidated the vdev. * 5. This isn't prefetch and l2arc_noprefetch is set. */ if (HDR_HAS_L2HDR(hdr) && !HDR_L2_WRITING(hdr) && !HDR_L2_EVICTED(hdr) && !(l2arc_noprefetch && HDR_PREFETCH(hdr))) { l2arc_read_callback_t *cb; DTRACE_PROBE1(l2arc__hit, arc_buf_hdr_t *, hdr); ARCSTAT_BUMP(arcstat_l2_hits); atomic_inc_32(&hdr->b_l2hdr.b_hits); cb = kmem_zalloc(sizeof (l2arc_read_callback_t), KM_SLEEP); cb->l2rcb_buf = buf; cb->l2rcb_spa = spa; cb->l2rcb_bp = *bp; cb->l2rcb_zb = *zb; cb->l2rcb_flags = zio_flags; cb->l2rcb_compress = b_compress; ASSERT(addr >= VDEV_LABEL_START_SIZE && addr + size < vd->vdev_psize - VDEV_LABEL_END_SIZE); /* * l2arc read. The SCL_L2ARC lock will be * released by l2arc_read_done(). * Issue a null zio if the underlying buffer * was squashed to zero size by compression. */ if (b_compress == ZIO_COMPRESS_EMPTY) { rzio = zio_null(pio, spa, vd, l2arc_read_done, cb, zio_flags | ZIO_FLAG_DONT_CACHE | ZIO_FLAG_CANFAIL | ZIO_FLAG_DONT_PROPAGATE | ZIO_FLAG_DONT_RETRY); } else { rzio = zio_read_phys(pio, vd, addr, b_asize, buf->b_data, ZIO_CHECKSUM_OFF, l2arc_read_done, cb, priority, zio_flags | ZIO_FLAG_DONT_CACHE | ZIO_FLAG_CANFAIL | ZIO_FLAG_DONT_PROPAGATE | ZIO_FLAG_DONT_RETRY, B_FALSE); } DTRACE_PROBE2(l2arc__read, vdev_t *, vd, zio_t *, rzio); ARCSTAT_INCR(arcstat_l2_read_bytes, b_asize); if (*arc_flags & ARC_FLAG_NOWAIT) { zio_nowait(rzio); goto out; } ASSERT(*arc_flags & ARC_FLAG_WAIT); if (zio_wait(rzio) == 0) goto out; /* l2arc read error; goto zio_read() */ } else { DTRACE_PROBE1(l2arc__miss, arc_buf_hdr_t *, hdr); ARCSTAT_BUMP(arcstat_l2_misses); if (HDR_L2_WRITING(hdr)) ARCSTAT_BUMP(arcstat_l2_rw_clash); spa_config_exit(spa, SCL_L2ARC, vd); } } else { if (vd != NULL) spa_config_exit(spa, SCL_L2ARC, vd); if (l2arc_ndev != 0) { DTRACE_PROBE1(l2arc__miss, arc_buf_hdr_t *, hdr); ARCSTAT_BUMP(arcstat_l2_misses); } } rzio = zio_read(pio, spa, bp, buf->b_data, size, arc_read_done, buf, priority, zio_flags, zb); if (*arc_flags & ARC_FLAG_WAIT) { rc = zio_wait(rzio); goto out; } ASSERT(*arc_flags & ARC_FLAG_NOWAIT); zio_nowait(rzio); } out: spa_read_history_add(spa, zb, *arc_flags); return (rc); } arc_prune_t * arc_add_prune_callback(arc_prune_func_t *func, void *private) { arc_prune_t *p; p = kmem_alloc(sizeof (*p), KM_SLEEP); p->p_pfunc = func; p->p_private = private; list_link_init(&p->p_node); refcount_create(&p->p_refcnt); mutex_enter(&arc_prune_mtx); refcount_add(&p->p_refcnt, &arc_prune_list); list_insert_head(&arc_prune_list, p); mutex_exit(&arc_prune_mtx); return (p); } void arc_remove_prune_callback(arc_prune_t *p) { mutex_enter(&arc_prune_mtx); list_remove(&arc_prune_list, p); if (refcount_remove(&p->p_refcnt, &arc_prune_list) == 0) { refcount_destroy(&p->p_refcnt); kmem_free(p, sizeof (*p)); } mutex_exit(&arc_prune_mtx); } void arc_set_callback(arc_buf_t *buf, arc_evict_func_t *func, void *private) { ASSERT(buf->b_hdr != NULL); ASSERT(buf->b_hdr->b_l1hdr.b_state != arc_anon); ASSERT(!refcount_is_zero(&buf->b_hdr->b_l1hdr.b_refcnt) || func == NULL); ASSERT(buf->b_efunc == NULL); ASSERT(!HDR_BUF_AVAILABLE(buf->b_hdr)); buf->b_efunc = func; buf->b_private = private; } /* * Notify the arc that a block was freed, and thus will never be used again. */ void arc_freed(spa_t *spa, const blkptr_t *bp) { arc_buf_hdr_t *hdr; kmutex_t *hash_lock; uint64_t guid = spa_load_guid(spa); ASSERT(!BP_IS_EMBEDDED(bp)); hdr = buf_hash_find(guid, bp, &hash_lock); if (hdr == NULL) return; if (HDR_BUF_AVAILABLE(hdr)) { arc_buf_t *buf = hdr->b_l1hdr.b_buf; add_reference(hdr, hash_lock, FTAG); hdr->b_flags &= ~ARC_FLAG_BUF_AVAILABLE; mutex_exit(hash_lock); arc_release(buf, FTAG); (void) arc_buf_remove_ref(buf, FTAG); } else { mutex_exit(hash_lock); } } /* * Clear the user eviction callback set by arc_set_callback(), first calling * it if it exists. Because the presence of a callback keeps an arc_buf cached * clearing the callback may result in the arc_buf being destroyed. However, * it will not result in the *last* arc_buf being destroyed, hence the data * will remain cached in the ARC. We make a copy of the arc buffer here so * that we can process the callback without holding any locks. * * It's possible that the callback is already in the process of being cleared * by another thread. In this case we can not clear the callback. * * Returns B_TRUE if the callback was successfully called and cleared. */ boolean_t arc_clear_callback(arc_buf_t *buf) { arc_buf_hdr_t *hdr; kmutex_t *hash_lock; arc_evict_func_t *efunc = buf->b_efunc; void *private = buf->b_private; mutex_enter(&buf->b_evict_lock); hdr = buf->b_hdr; if (hdr == NULL) { /* * We are in arc_do_user_evicts(). */ ASSERT(buf->b_data == NULL); mutex_exit(&buf->b_evict_lock); return (B_FALSE); } else if (buf->b_data == NULL) { /* * We are on the eviction list; process this buffer now * but let arc_do_user_evicts() do the reaping. */ buf->b_efunc = NULL; mutex_exit(&buf->b_evict_lock); VERIFY0(efunc(private)); return (B_TRUE); } hash_lock = HDR_LOCK(hdr); mutex_enter(hash_lock); hdr = buf->b_hdr; ASSERT3P(hash_lock, ==, HDR_LOCK(hdr)); ASSERT3U(refcount_count(&hdr->b_l1hdr.b_refcnt), <, hdr->b_l1hdr.b_datacnt); ASSERT(hdr->b_l1hdr.b_state == arc_mru || hdr->b_l1hdr.b_state == arc_mfu); buf->b_efunc = NULL; buf->b_private = NULL; if (hdr->b_l1hdr.b_datacnt > 1) { mutex_exit(&buf->b_evict_lock); arc_buf_destroy(buf, TRUE); } else { ASSERT(buf == hdr->b_l1hdr.b_buf); hdr->b_flags |= ARC_FLAG_BUF_AVAILABLE; mutex_exit(&buf->b_evict_lock); } mutex_exit(hash_lock); VERIFY0(efunc(private)); return (B_TRUE); } /* * Release this buffer from the cache, making it an anonymous buffer. This * must be done after a read and prior to modifying the buffer contents. * If the buffer has more than one reference, we must make * a new hdr for the buffer. */ void arc_release(arc_buf_t *buf, void *tag) { kmutex_t *hash_lock; arc_state_t *state; arc_buf_hdr_t *hdr = buf->b_hdr; /* * It would be nice to assert that if its DMU metadata (level > * 0 || it's the dnode file), then it must be syncing context. * But we don't know that information at this level. */ mutex_enter(&buf->b_evict_lock); ASSERT(HDR_HAS_L1HDR(hdr)); /* * We don't grab the hash lock prior to this check, because if * the buffer's header is in the arc_anon state, it won't be * linked into the hash table. */ if (hdr->b_l1hdr.b_state == arc_anon) { mutex_exit(&buf->b_evict_lock); ASSERT(!HDR_IO_IN_PROGRESS(hdr)); ASSERT(!HDR_IN_HASH_TABLE(hdr)); ASSERT(!HDR_HAS_L2HDR(hdr)); ASSERT(BUF_EMPTY(hdr)); ASSERT3U(hdr->b_l1hdr.b_datacnt, ==, 1); ASSERT3S(refcount_count(&hdr->b_l1hdr.b_refcnt), ==, 1); ASSERT(!list_link_active(&hdr->b_l1hdr.b_arc_node)); ASSERT3P(buf->b_efunc, ==, NULL); ASSERT3P(buf->b_private, ==, NULL); hdr->b_l1hdr.b_arc_access = 0; arc_buf_thaw(buf); return; } hash_lock = HDR_LOCK(hdr); mutex_enter(hash_lock); /* * This assignment is only valid as long as the hash_lock is * held, we must be careful not to reference state or the * b_state field after dropping the lock. */ state = hdr->b_l1hdr.b_state; ASSERT3P(hash_lock, ==, HDR_LOCK(hdr)); ASSERT3P(state, !=, arc_anon); /* this buffer is not on any list */ ASSERT(refcount_count(&hdr->b_l1hdr.b_refcnt) > 0); if (HDR_HAS_L2HDR(hdr)) { mutex_enter(&hdr->b_l2hdr.b_dev->l2ad_mtx); /* * We have to recheck this conditional again now that * we're holding the l2ad_mtx to prevent a race with * another thread which might be concurrently calling * l2arc_evict(). In that case, l2arc_evict() might have * destroyed the header's L2 portion as we were waiting * to acquire the l2ad_mtx. */ if (HDR_HAS_L2HDR(hdr)) arc_hdr_l2hdr_destroy(hdr); mutex_exit(&hdr->b_l2hdr.b_dev->l2ad_mtx); } /* * Do we have more than one buf? */ if (hdr->b_l1hdr.b_datacnt > 1) { arc_buf_hdr_t *nhdr; arc_buf_t **bufp; uint64_t blksz = hdr->b_size; uint64_t spa = hdr->b_spa; arc_buf_contents_t type = arc_buf_type(hdr); uint32_t flags = hdr->b_flags; ASSERT(hdr->b_l1hdr.b_buf != buf || buf->b_next != NULL); /* * Pull the data off of this hdr and attach it to * a new anonymous hdr. */ (void) remove_reference(hdr, hash_lock, tag); bufp = &hdr->b_l1hdr.b_buf; while (*bufp != buf) bufp = &(*bufp)->b_next; *bufp = buf->b_next; buf->b_next = NULL; ASSERT3P(state, !=, arc_l2c_only); (void) refcount_remove_many( &state->arcs_size, hdr->b_size, buf); if (refcount_is_zero(&hdr->b_l1hdr.b_refcnt)) { uint64_t *size; ASSERT3P(state, !=, arc_l2c_only); size = &state->arcs_lsize[type]; ASSERT3U(*size, >=, hdr->b_size); atomic_add_64(size, -hdr->b_size); } /* * We're releasing a duplicate user data buffer, update * our statistics accordingly. */ if (HDR_ISTYPE_DATA(hdr)) { ARCSTAT_BUMPDOWN(arcstat_duplicate_buffers); ARCSTAT_INCR(arcstat_duplicate_buffers_size, -hdr->b_size); } hdr->b_l1hdr.b_datacnt -= 1; arc_cksum_verify(buf); arc_buf_unwatch(buf); mutex_exit(hash_lock); nhdr = kmem_cache_alloc(hdr_full_cache, KM_PUSHPAGE); nhdr->b_size = blksz; nhdr->b_spa = spa; nhdr->b_l1hdr.b_mru_hits = 0; nhdr->b_l1hdr.b_mru_ghost_hits = 0; nhdr->b_l1hdr.b_mfu_hits = 0; nhdr->b_l1hdr.b_mfu_ghost_hits = 0; nhdr->b_l1hdr.b_l2_hits = 0; nhdr->b_flags = flags & ARC_FLAG_L2_WRITING; nhdr->b_flags |= arc_bufc_to_flags(type); nhdr->b_flags |= ARC_FLAG_HAS_L1HDR; nhdr->b_l1hdr.b_buf = buf; nhdr->b_l1hdr.b_datacnt = 1; nhdr->b_l1hdr.b_state = arc_anon; nhdr->b_l1hdr.b_arc_access = 0; nhdr->b_l1hdr.b_tmp_cdata = NULL; nhdr->b_freeze_cksum = NULL; (void) refcount_add(&nhdr->b_l1hdr.b_refcnt, tag); buf->b_hdr = nhdr; mutex_exit(&buf->b_evict_lock); (void) refcount_add_many(&arc_anon->arcs_size, blksz, buf); } else { mutex_exit(&buf->b_evict_lock); ASSERT(refcount_count(&hdr->b_l1hdr.b_refcnt) == 1); /* protected by hash lock, or hdr is on arc_anon */ ASSERT(!multilist_link_active(&hdr->b_l1hdr.b_arc_node)); ASSERT(!HDR_IO_IN_PROGRESS(hdr)); hdr->b_l1hdr.b_mru_hits = 0; hdr->b_l1hdr.b_mru_ghost_hits = 0; hdr->b_l1hdr.b_mfu_hits = 0; hdr->b_l1hdr.b_mfu_ghost_hits = 0; hdr->b_l1hdr.b_l2_hits = 0; arc_change_state(arc_anon, hdr, hash_lock); hdr->b_l1hdr.b_arc_access = 0; mutex_exit(hash_lock); buf_discard_identity(hdr); arc_buf_thaw(buf); } buf->b_efunc = NULL; buf->b_private = NULL; } int arc_released(arc_buf_t *buf) { int released; mutex_enter(&buf->b_evict_lock); released = (buf->b_data != NULL && buf->b_hdr->b_l1hdr.b_state == arc_anon); mutex_exit(&buf->b_evict_lock); return (released); } #ifdef ZFS_DEBUG int arc_referenced(arc_buf_t *buf) { int referenced; mutex_enter(&buf->b_evict_lock); referenced = (refcount_count(&buf->b_hdr->b_l1hdr.b_refcnt)); mutex_exit(&buf->b_evict_lock); return (referenced); } #endif static void arc_write_ready(zio_t *zio) { arc_write_callback_t *callback = zio->io_private; arc_buf_t *buf = callback->awcb_buf; arc_buf_hdr_t *hdr = buf->b_hdr; ASSERT(HDR_HAS_L1HDR(hdr)); ASSERT(!refcount_is_zero(&buf->b_hdr->b_l1hdr.b_refcnt)); ASSERT(hdr->b_l1hdr.b_datacnt > 0); callback->awcb_ready(zio, buf, callback->awcb_private); /* * If the IO is already in progress, then this is a re-write * attempt, so we need to thaw and re-compute the cksum. * It is the responsibility of the callback to handle the * accounting for any re-write attempt. */ if (HDR_IO_IN_PROGRESS(hdr)) { mutex_enter(&hdr->b_l1hdr.b_freeze_lock); if (hdr->b_freeze_cksum != NULL) { kmem_free(hdr->b_freeze_cksum, sizeof (zio_cksum_t)); hdr->b_freeze_cksum = NULL; } mutex_exit(&hdr->b_l1hdr.b_freeze_lock); } arc_cksum_compute(buf, B_FALSE); hdr->b_flags |= ARC_FLAG_IO_IN_PROGRESS; } /* * The SPA calls this callback for each physical write that happens on behalf * of a logical write. See the comment in dbuf_write_physdone() for details. */ static void arc_write_physdone(zio_t *zio) { arc_write_callback_t *cb = zio->io_private; if (cb->awcb_physdone != NULL) cb->awcb_physdone(zio, cb->awcb_buf, cb->awcb_private); } static void arc_write_done(zio_t *zio) { arc_write_callback_t *callback = zio->io_private; arc_buf_t *buf = callback->awcb_buf; arc_buf_hdr_t *hdr = buf->b_hdr; ASSERT(hdr->b_l1hdr.b_acb == NULL); if (zio->io_error == 0) { if (BP_IS_HOLE(zio->io_bp) || BP_IS_EMBEDDED(zio->io_bp)) { buf_discard_identity(hdr); } else { hdr->b_dva = *BP_IDENTITY(zio->io_bp); hdr->b_birth = BP_PHYSICAL_BIRTH(zio->io_bp); } } else { ASSERT(BUF_EMPTY(hdr)); } /* * If the block to be written was all-zero or compressed enough to be * embedded in the BP, no write was performed so there will be no * dva/birth/checksum. The buffer must therefore remain anonymous * (and uncached). */ if (!BUF_EMPTY(hdr)) { arc_buf_hdr_t *exists; kmutex_t *hash_lock; ASSERT(zio->io_error == 0); arc_cksum_verify(buf); exists = buf_hash_insert(hdr, &hash_lock); if (exists != NULL) { /* * This can only happen if we overwrite for * sync-to-convergence, because we remove * buffers from the hash table when we arc_free(). */ if (zio->io_flags & ZIO_FLAG_IO_REWRITE) { if (!BP_EQUAL(&zio->io_bp_orig, zio->io_bp)) panic("bad overwrite, hdr=%p exists=%p", (void *)hdr, (void *)exists); ASSERT(refcount_is_zero( &exists->b_l1hdr.b_refcnt)); arc_change_state(arc_anon, exists, hash_lock); mutex_exit(hash_lock); arc_hdr_destroy(exists); exists = buf_hash_insert(hdr, &hash_lock); ASSERT3P(exists, ==, NULL); } else if (zio->io_flags & ZIO_FLAG_NOPWRITE) { /* nopwrite */ ASSERT(zio->io_prop.zp_nopwrite); if (!BP_EQUAL(&zio->io_bp_orig, zio->io_bp)) panic("bad nopwrite, hdr=%p exists=%p", (void *)hdr, (void *)exists); } else { /* Dedup */ ASSERT(hdr->b_l1hdr.b_datacnt == 1); ASSERT(hdr->b_l1hdr.b_state == arc_anon); ASSERT(BP_GET_DEDUP(zio->io_bp)); ASSERT(BP_GET_LEVEL(zio->io_bp) == 0); } } hdr->b_flags &= ~ARC_FLAG_IO_IN_PROGRESS; /* if it's not anon, we are doing a scrub */ if (exists == NULL && hdr->b_l1hdr.b_state == arc_anon) arc_access(hdr, hash_lock); mutex_exit(hash_lock); } else { hdr->b_flags &= ~ARC_FLAG_IO_IN_PROGRESS; } ASSERT(!refcount_is_zero(&hdr->b_l1hdr.b_refcnt)); callback->awcb_done(zio, buf, callback->awcb_private); kmem_free(callback, sizeof (arc_write_callback_t)); } zio_t * arc_write(zio_t *pio, spa_t *spa, uint64_t txg, blkptr_t *bp, arc_buf_t *buf, boolean_t l2arc, boolean_t l2arc_compress, const zio_prop_t *zp, arc_done_func_t *ready, arc_done_func_t *physdone, arc_done_func_t *done, void *private, zio_priority_t priority, int zio_flags, const zbookmark_phys_t *zb) { arc_buf_hdr_t *hdr = buf->b_hdr; arc_write_callback_t *callback; zio_t *zio; ASSERT(ready != NULL); ASSERT(done != NULL); ASSERT(!HDR_IO_ERROR(hdr)); ASSERT(!HDR_IO_IN_PROGRESS(hdr)); ASSERT(hdr->b_l1hdr.b_acb == NULL); ASSERT(hdr->b_l1hdr.b_datacnt > 0); if (l2arc) hdr->b_flags |= ARC_FLAG_L2CACHE; if (l2arc_compress) hdr->b_flags |= ARC_FLAG_L2COMPRESS; callback = kmem_zalloc(sizeof (arc_write_callback_t), KM_SLEEP); callback->awcb_ready = ready; callback->awcb_physdone = physdone; callback->awcb_done = done; callback->awcb_private = private; callback->awcb_buf = buf; zio = zio_write(pio, spa, txg, bp, buf->b_data, hdr->b_size, zp, arc_write_ready, arc_write_physdone, arc_write_done, callback, priority, zio_flags, zb); return (zio); } static int arc_memory_throttle(uint64_t reserve, uint64_t txg) { #ifdef _KERNEL uint64_t available_memory = ptob(freemem); static uint64_t page_load = 0; static uint64_t last_txg = 0; #ifdef __linux__ pgcnt_t minfree = btop(arc_sys_free / 4); #endif if (freemem > physmem * arc_lotsfree_percent / 100) return (0); if (txg > last_txg) { last_txg = txg; page_load = 0; } /* * If we are in pageout, we know that memory is already tight, * the arc is already going to be evicting, so we just want to * continue to let page writes occur as quickly as possible. */ if (current_is_kswapd()) { if (page_load > MAX(ptob(minfree), available_memory) / 4) { DMU_TX_STAT_BUMP(dmu_tx_memory_reclaim); return (SET_ERROR(ERESTART)); } /* Note: reserve is inflated, so we deflate */ page_load += reserve / 8; return (0); } else if (page_load > 0 && arc_reclaim_needed()) { /* memory is low, delay before restarting */ ARCSTAT_INCR(arcstat_memory_throttle_count, 1); DMU_TX_STAT_BUMP(dmu_tx_memory_reclaim); return (SET_ERROR(EAGAIN)); } page_load = 0; #endif return (0); } void arc_tempreserve_clear(uint64_t reserve) { atomic_add_64(&arc_tempreserve, -reserve); ASSERT((int64_t)arc_tempreserve >= 0); } int arc_tempreserve_space(uint64_t reserve, uint64_t txg) { int error; uint64_t anon_size; if (reserve > arc_c/4 && !arc_no_grow) arc_c = MIN(arc_c_max, reserve * 4); /* * Throttle when the calculated memory footprint for the TXG * exceeds the target ARC size. */ if (reserve > arc_c) { DMU_TX_STAT_BUMP(dmu_tx_memory_reserve); return (SET_ERROR(ERESTART)); } /* * Don't count loaned bufs as in flight dirty data to prevent long * network delays from blocking transactions that are ready to be * assigned to a txg. */ anon_size = MAX((int64_t)(refcount_count(&arc_anon->arcs_size) - arc_loaned_bytes), 0); /* * Writes will, almost always, require additional memory allocations * in order to compress/encrypt/etc the data. We therefore need to * make sure that there is sufficient available memory for this. */ error = arc_memory_throttle(reserve, txg); if (error != 0) return (error); /* * Throttle writes when the amount of dirty data in the cache * gets too large. We try to keep the cache less than half full * of dirty blocks so that our sync times don't grow too large. * Note: if two requests come in concurrently, we might let them * both succeed, when one of them should fail. Not a huge deal. */ if (reserve + arc_tempreserve + anon_size > arc_c / 2 && anon_size > arc_c / 4) { dprintf("failing, arc_tempreserve=%lluK anon_meta=%lluK " "anon_data=%lluK tempreserve=%lluK arc_c=%lluK\n", arc_tempreserve>>10, arc_anon->arcs_lsize[ARC_BUFC_METADATA]>>10, arc_anon->arcs_lsize[ARC_BUFC_DATA]>>10, reserve>>10, arc_c>>10); DMU_TX_STAT_BUMP(dmu_tx_dirty_throttle); return (SET_ERROR(ERESTART)); } atomic_add_64(&arc_tempreserve, reserve); return (0); } static void arc_kstat_update_state(arc_state_t *state, kstat_named_t *size, kstat_named_t *evict_data, kstat_named_t *evict_metadata) { size->value.ui64 = refcount_count(&state->arcs_size); evict_data->value.ui64 = state->arcs_lsize[ARC_BUFC_DATA]; evict_metadata->value.ui64 = state->arcs_lsize[ARC_BUFC_METADATA]; } static int arc_kstat_update(kstat_t *ksp, int rw) { arc_stats_t *as = ksp->ks_data; if (rw == KSTAT_WRITE) { return (EACCES); } else { arc_kstat_update_state(arc_anon, &as->arcstat_anon_size, &as->arcstat_anon_evictable_data, &as->arcstat_anon_evictable_metadata); arc_kstat_update_state(arc_mru, &as->arcstat_mru_size, &as->arcstat_mru_evictable_data, &as->arcstat_mru_evictable_metadata); arc_kstat_update_state(arc_mru_ghost, &as->arcstat_mru_ghost_size, &as->arcstat_mru_ghost_evictable_data, &as->arcstat_mru_ghost_evictable_metadata); arc_kstat_update_state(arc_mfu, &as->arcstat_mfu_size, &as->arcstat_mfu_evictable_data, &as->arcstat_mfu_evictable_metadata); arc_kstat_update_state(arc_mfu_ghost, &as->arcstat_mfu_ghost_size, &as->arcstat_mfu_ghost_evictable_data, &as->arcstat_mfu_ghost_evictable_metadata); } return (0); } /* * This function *must* return indices evenly distributed between all * sublists of the multilist. This is needed due to how the ARC eviction * code is laid out; arc_evict_state() assumes ARC buffers are evenly * distributed between all sublists and uses this assumption when * deciding which sublist to evict from and how much to evict from it. */ unsigned int arc_state_multilist_index_func(multilist_t *ml, void *obj) { arc_buf_hdr_t *hdr = obj; /* * We rely on b_dva to generate evenly distributed index * numbers using buf_hash below. So, as an added precaution, * let's make sure we never add empty buffers to the arc lists. */ ASSERT(!BUF_EMPTY(hdr)); /* * The assumption here, is the hash value for a given * arc_buf_hdr_t will remain constant throughout its lifetime * (i.e. its b_spa, b_dva, and b_birth fields don't change). * Thus, we don't need to store the header'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. */ return (buf_hash(hdr->b_spa, &hdr->b_dva, hdr->b_birth) % multilist_get_num_sublists(ml)); } /* * Called during module initialization and periodically thereafter to * apply reasonable changes to the exposed performance tunings. Non-zero * zfs_* values which differ from the currently set values will be applied. */ static void arc_tuning_update(void) { /* Valid range: 64M - */ if ((zfs_arc_max) && (zfs_arc_max != arc_c_max) && (zfs_arc_max > 64 << 20) && (zfs_arc_max < ptob(physmem)) && (zfs_arc_max > arc_c_min)) { arc_c_max = zfs_arc_max; arc_c = arc_c_max; arc_p = (arc_c >> 1); arc_meta_limit = MIN(arc_meta_limit, (3 * arc_c_max) / 4); } /* Valid range: 32M - */ if ((zfs_arc_min) && (zfs_arc_min != arc_c_min) && (zfs_arc_min >= 2ULL << SPA_MAXBLOCKSHIFT) && (zfs_arc_min <= arc_c_max)) { arc_c_min = zfs_arc_min; arc_c = MAX(arc_c, arc_c_min); } /* Valid range: 16M - */ if ((zfs_arc_meta_min) && (zfs_arc_meta_min != arc_meta_min) && (zfs_arc_meta_min >= 1ULL << SPA_MAXBLOCKSHIFT) && (zfs_arc_meta_min <= arc_c_max)) { arc_meta_min = zfs_arc_meta_min; arc_meta_limit = MAX(arc_meta_limit, arc_meta_min); } /* Valid range: - */ if ((zfs_arc_meta_limit) && (zfs_arc_meta_limit != arc_meta_limit) && (zfs_arc_meta_limit >= zfs_arc_meta_min) && (zfs_arc_meta_limit <= arc_c_max)) arc_meta_limit = zfs_arc_meta_limit; /* Valid range: 1 - N */ if (zfs_arc_grow_retry) arc_grow_retry = zfs_arc_grow_retry; /* Valid range: 1 - N */ if (zfs_arc_shrink_shift) { arc_shrink_shift = zfs_arc_shrink_shift; arc_no_grow_shift = MIN(arc_no_grow_shift, arc_shrink_shift -1); } /* Valid range: 1 - N */ if (zfs_arc_p_min_shift) arc_p_min_shift = zfs_arc_p_min_shift; /* Valid range: 1 - N ticks */ if (zfs_arc_min_prefetch_lifespan) arc_min_prefetch_lifespan = zfs_arc_min_prefetch_lifespan; /* Valid range: 0 - 100 */ if ((zfs_arc_lotsfree_percent >= 0) && (zfs_arc_lotsfree_percent <= 100)) arc_lotsfree_percent = zfs_arc_lotsfree_percent; /* Valid range: 0 - */ if ((zfs_arc_sys_free) && (zfs_arc_sys_free != arc_sys_free)) arc_sys_free = MIN(MAX(zfs_arc_sys_free, 0), ptob(physmem)); } void arc_init(void) { /* * allmem is "all memory that we could possibly use". */ #ifdef _KERNEL uint64_t allmem = ptob(physmem); #else uint64_t allmem = (physmem * PAGESIZE) / 2; #endif mutex_init(&arc_reclaim_lock, NULL, MUTEX_DEFAULT, NULL); cv_init(&arc_reclaim_thread_cv, NULL, CV_DEFAULT, NULL); cv_init(&arc_reclaim_waiters_cv, NULL, CV_DEFAULT, NULL); mutex_init(&arc_user_evicts_lock, NULL, MUTEX_DEFAULT, NULL); cv_init(&arc_user_evicts_cv, NULL, CV_DEFAULT, NULL); /* Convert seconds to clock ticks */ arc_min_prefetch_lifespan = 1 * hz; /* Start out with 1/8 of all memory */ arc_c = allmem / 8; #ifdef _KERNEL /* * On architectures where the physical memory can be larger * than the addressable space (intel in 32-bit mode), we may * need to limit the cache to 1/8 of VM size. */ arc_c = MIN(arc_c, vmem_size(heap_arena, VMEM_ALLOC | VMEM_FREE) / 8); /* * Register a shrinker to support synchronous (direct) memory * reclaim from the arc. This is done to prevent kswapd from * swapping out pages when it is preferable to shrink the arc. */ spl_register_shrinker(&arc_shrinker); /* Set to 1/64 of all memory or a minimum of 512K */ arc_sys_free = MAX(ptob(physmem / 64), (512 * 1024)); arc_need_free = 0; #endif /* * In userland, there's only the memory pressure that we artificially * create (see arc_available_memory()). Don't let arc_c get too * small, because it can cause transactions to be larger than * arc_c, causing arc_tempreserve_space() to fail. */ #ifndef _KERNEL arc_c_min = arc_c_max / 2; #else arc_c_min = 2ULL << SPA_MAXBLOCKSHIFT; #endif /* Set max to 1/2 of all memory */ arc_c_max = allmem / 2; arc_c = arc_c_max; arc_p = (arc_c >> 1); /* Set min to 1/2 of arc_c_min */ arc_meta_min = 1ULL << SPA_MAXBLOCKSHIFT; /* Initialize maximum observed usage to zero */ arc_meta_max = 0; /* Set limit to 3/4 of arc_c_max with a floor of arc_meta_min */ arc_meta_limit = MAX((3 * arc_c_max) / 4, arc_meta_min); /* Apply user specified tunings */ arc_tuning_update(); if (zfs_arc_num_sublists_per_state < 1) zfs_arc_num_sublists_per_state = MAX(boot_ncpus, 1); /* if kmem_flags are set, lets try to use less memory */ if (kmem_debugging()) arc_c = arc_c / 2; if (arc_c < arc_c_min) arc_c = arc_c_min; arc_anon = &ARC_anon; arc_mru = &ARC_mru; arc_mru_ghost = &ARC_mru_ghost; arc_mfu = &ARC_mfu; arc_mfu_ghost = &ARC_mfu_ghost; arc_l2c_only = &ARC_l2c_only; arc_size = 0; multilist_create(&arc_mru->arcs_list[ARC_BUFC_METADATA], sizeof (arc_buf_hdr_t), offsetof(arc_buf_hdr_t, b_l1hdr.b_arc_node), zfs_arc_num_sublists_per_state, arc_state_multilist_index_func); multilist_create(&arc_mru->arcs_list[ARC_BUFC_DATA], sizeof (arc_buf_hdr_t), offsetof(arc_buf_hdr_t, b_l1hdr.b_arc_node), zfs_arc_num_sublists_per_state, arc_state_multilist_index_func); multilist_create(&arc_mru_ghost->arcs_list[ARC_BUFC_METADATA], sizeof (arc_buf_hdr_t), offsetof(arc_buf_hdr_t, b_l1hdr.b_arc_node), zfs_arc_num_sublists_per_state, arc_state_multilist_index_func); multilist_create(&arc_mru_ghost->arcs_list[ARC_BUFC_DATA], sizeof (arc_buf_hdr_t), offsetof(arc_buf_hdr_t, b_l1hdr.b_arc_node), zfs_arc_num_sublists_per_state, arc_state_multilist_index_func); multilist_create(&arc_mfu->arcs_list[ARC_BUFC_METADATA], sizeof (arc_buf_hdr_t), offsetof(arc_buf_hdr_t, b_l1hdr.b_arc_node), zfs_arc_num_sublists_per_state, arc_state_multilist_index_func); multilist_create(&arc_mfu->arcs_list[ARC_BUFC_DATA], sizeof (arc_buf_hdr_t), offsetof(arc_buf_hdr_t, b_l1hdr.b_arc_node), zfs_arc_num_sublists_per_state, arc_state_multilist_index_func); multilist_create(&arc_mfu_ghost->arcs_list[ARC_BUFC_METADATA], sizeof (arc_buf_hdr_t), offsetof(arc_buf_hdr_t, b_l1hdr.b_arc_node), zfs_arc_num_sublists_per_state, arc_state_multilist_index_func); multilist_create(&arc_mfu_ghost->arcs_list[ARC_BUFC_DATA], sizeof (arc_buf_hdr_t), offsetof(arc_buf_hdr_t, b_l1hdr.b_arc_node), zfs_arc_num_sublists_per_state, arc_state_multilist_index_func); multilist_create(&arc_l2c_only->arcs_list[ARC_BUFC_METADATA], sizeof (arc_buf_hdr_t), offsetof(arc_buf_hdr_t, b_l1hdr.b_arc_node), zfs_arc_num_sublists_per_state, arc_state_multilist_index_func); multilist_create(&arc_l2c_only->arcs_list[ARC_BUFC_DATA], sizeof (arc_buf_hdr_t), offsetof(arc_buf_hdr_t, b_l1hdr.b_arc_node), zfs_arc_num_sublists_per_state, arc_state_multilist_index_func); arc_anon->arcs_state = ARC_STATE_ANON; arc_mru->arcs_state = ARC_STATE_MRU; arc_mru_ghost->arcs_state = ARC_STATE_MRU_GHOST; arc_mfu->arcs_state = ARC_STATE_MFU; arc_mfu_ghost->arcs_state = ARC_STATE_MFU_GHOST; arc_l2c_only->arcs_state = ARC_STATE_L2C_ONLY; refcount_create(&arc_anon->arcs_size); refcount_create(&arc_mru->arcs_size); refcount_create(&arc_mru_ghost->arcs_size); refcount_create(&arc_mfu->arcs_size); refcount_create(&arc_mfu_ghost->arcs_size); refcount_create(&arc_l2c_only->arcs_size); buf_init(); arc_reclaim_thread_exit = FALSE; arc_user_evicts_thread_exit = FALSE; list_create(&arc_prune_list, sizeof (arc_prune_t), offsetof(arc_prune_t, p_node)); arc_eviction_list = NULL; mutex_init(&arc_prune_mtx, NULL, MUTEX_DEFAULT, NULL); bzero(&arc_eviction_hdr, sizeof (arc_buf_hdr_t)); arc_prune_taskq = taskq_create("arc_prune", max_ncpus, defclsyspri, max_ncpus, INT_MAX, TASKQ_PREPOPULATE | TASKQ_DYNAMIC); arc_ksp = kstat_create("zfs", 0, "arcstats", "misc", KSTAT_TYPE_NAMED, sizeof (arc_stats) / sizeof (kstat_named_t), KSTAT_FLAG_VIRTUAL); if (arc_ksp != NULL) { arc_ksp->ks_data = &arc_stats; arc_ksp->ks_update = arc_kstat_update; kstat_install(arc_ksp); } (void) thread_create(NULL, 0, arc_reclaim_thread, NULL, 0, &p0, TS_RUN, defclsyspri); (void) thread_create(NULL, 0, arc_user_evicts_thread, NULL, 0, &p0, TS_RUN, defclsyspri); arc_dead = FALSE; arc_warm = B_FALSE; /* * Calculate maximum amount of dirty data per pool. * * If it has been set by a module parameter, take that. * Otherwise, use a percentage of physical memory defined by * zfs_dirty_data_max_percent (default 10%) with a cap at * zfs_dirty_data_max_max (default 25% of physical memory). */ if (zfs_dirty_data_max_max == 0) zfs_dirty_data_max_max = (uint64_t)physmem * PAGESIZE * zfs_dirty_data_max_max_percent / 100; if (zfs_dirty_data_max == 0) { zfs_dirty_data_max = (uint64_t)physmem * PAGESIZE * zfs_dirty_data_max_percent / 100; zfs_dirty_data_max = MIN(zfs_dirty_data_max, zfs_dirty_data_max_max); } } void arc_fini(void) { arc_prune_t *p; #ifdef _KERNEL spl_unregister_shrinker(&arc_shrinker); #endif /* _KERNEL */ mutex_enter(&arc_reclaim_lock); arc_reclaim_thread_exit = TRUE; /* * The reclaim thread will set arc_reclaim_thread_exit back to * FALSE when it is finished exiting; we're waiting for that. */ while (arc_reclaim_thread_exit) { cv_signal(&arc_reclaim_thread_cv); cv_wait(&arc_reclaim_thread_cv, &arc_reclaim_lock); } mutex_exit(&arc_reclaim_lock); mutex_enter(&arc_user_evicts_lock); arc_user_evicts_thread_exit = TRUE; /* * The user evicts thread will set arc_user_evicts_thread_exit * to FALSE when it is finished exiting; we're waiting for that. */ while (arc_user_evicts_thread_exit) { cv_signal(&arc_user_evicts_cv); cv_wait(&arc_user_evicts_cv, &arc_user_evicts_lock); } mutex_exit(&arc_user_evicts_lock); /* Use TRUE to ensure *all* buffers are evicted */ arc_flush(NULL, TRUE); arc_dead = TRUE; if (arc_ksp != NULL) { kstat_delete(arc_ksp); arc_ksp = NULL; } taskq_wait(arc_prune_taskq); taskq_destroy(arc_prune_taskq); mutex_enter(&arc_prune_mtx); while ((p = list_head(&arc_prune_list)) != NULL) { list_remove(&arc_prune_list, p); refcount_remove(&p->p_refcnt, &arc_prune_list); refcount_destroy(&p->p_refcnt); kmem_free(p, sizeof (*p)); } mutex_exit(&arc_prune_mtx); list_destroy(&arc_prune_list); mutex_destroy(&arc_prune_mtx); mutex_destroy(&arc_reclaim_lock); cv_destroy(&arc_reclaim_thread_cv); cv_destroy(&arc_reclaim_waiters_cv); mutex_destroy(&arc_user_evicts_lock); cv_destroy(&arc_user_evicts_cv); refcount_destroy(&arc_anon->arcs_size); refcount_destroy(&arc_mru->arcs_size); refcount_destroy(&arc_mru_ghost->arcs_size); refcount_destroy(&arc_mfu->arcs_size); refcount_destroy(&arc_mfu_ghost->arcs_size); refcount_destroy(&arc_l2c_only->arcs_size); multilist_destroy(&arc_mru->arcs_list[ARC_BUFC_METADATA]); multilist_destroy(&arc_mru_ghost->arcs_list[ARC_BUFC_METADATA]); multilist_destroy(&arc_mfu->arcs_list[ARC_BUFC_METADATA]); multilist_destroy(&arc_mfu_ghost->arcs_list[ARC_BUFC_METADATA]); multilist_destroy(&arc_mru->arcs_list[ARC_BUFC_DATA]); multilist_destroy(&arc_mru_ghost->arcs_list[ARC_BUFC_DATA]); multilist_destroy(&arc_mfu->arcs_list[ARC_BUFC_DATA]); multilist_destroy(&arc_mfu_ghost->arcs_list[ARC_BUFC_DATA]); multilist_destroy(&arc_l2c_only->arcs_list[ARC_BUFC_METADATA]); multilist_destroy(&arc_l2c_only->arcs_list[ARC_BUFC_DATA]); buf_fini(); ASSERT0(arc_loaned_bytes); } /* * Level 2 ARC * * The level 2 ARC (L2ARC) is a cache layer in-between main memory and disk. * It uses dedicated storage devices to hold cached data, which are populated * using large infrequent writes. The main role of this cache is to boost * the performance of random read workloads. The intended L2ARC devices * include short-stroked disks, solid state disks, and other media with * substantially faster read latency than disk. * * +-----------------------+ * | ARC | * +-----------------------+ * | ^ ^ * | | | * l2arc_feed_thread() arc_read() * | | | * | l2arc read | * V | | * +---------------+ | * | L2ARC | | * +---------------+ | * | ^ | * l2arc_write() | | * | | | * V | | * +-------+ +-------+ * | vdev | | vdev | * | cache | | cache | * +-------+ +-------+ * +=========+ .-----. * : L2ARC : |-_____-| * : devices : | Disks | * +=========+ `-_____-' * * Read requests are satisfied from the following sources, in order: * * 1) ARC * 2) vdev cache of L2ARC devices * 3) L2ARC devices * 4) vdev cache of disks * 5) disks * * Some L2ARC device types exhibit extremely slow write performance. * To accommodate for this there are some significant differences between * the L2ARC and traditional cache design: * * 1. There is no eviction path from the ARC to the L2ARC. Evictions from * the ARC behave as usual, freeing buffers and placing headers on ghost * lists. The ARC does not send buffers to the L2ARC during eviction as * this would add inflated write latencies for all ARC memory pressure. * * 2. The L2ARC attempts to cache data from the ARC before it is evicted. * It does this by periodically scanning buffers from the eviction-end of * the MFU and MRU ARC lists, copying them to the L2ARC devices if they are * not already there. It scans until a headroom of buffers is satisfied, * which itself is a buffer for ARC eviction. If a compressible buffer is * found during scanning and selected for writing to an L2ARC device, we * temporarily boost scanning headroom during the next scan cycle to make * sure we adapt to compression effects (which might significantly reduce * the data volume we write to L2ARC). The thread that does this is * l2arc_feed_thread(), illustrated below; example sizes are included to * provide a better sense of ratio than this diagram: * * head --> tail * +---------------------+----------+ * ARC_mfu |:::::#:::::::::::::::|o#o###o###|-->. # already on L2ARC * +---------------------+----------+ | o L2ARC eligible * ARC_mru |:#:::::::::::::::::::|#o#ooo####|-->| : ARC buffer * +---------------------+----------+ | * 15.9 Gbytes ^ 32 Mbytes | * headroom | * l2arc_feed_thread() * | * l2arc write hand <--[oooo]--' * | 8 Mbyte * | write max * V * +==============================+ * L2ARC dev |####|#|###|###| |####| ... | * +==============================+ * 32 Gbytes * * 3. If an ARC buffer is copied to the L2ARC but then hit instead of * evicted, then the L2ARC has cached a buffer much sooner than it probably * needed to, potentially wasting L2ARC device bandwidth and storage. It is * safe to say that this is an uncommon case, since buffers at the end of * the ARC lists have moved there due to inactivity. * * 4. If the ARC evicts faster than the L2ARC can maintain a headroom, * then the L2ARC simply misses copying some buffers. This serves as a * pressure valve to prevent heavy read workloads from both stalling the ARC * with waits and clogging the L2ARC with writes. This also helps prevent * the potential for the L2ARC to churn if it attempts to cache content too * quickly, such as during backups of the entire pool. * * 5. After system boot and before the ARC has filled main memory, there are * no evictions from the ARC and so the tails of the ARC_mfu and ARC_mru * lists can remain mostly static. Instead of searching from tail of these * lists as pictured, the l2arc_feed_thread() will search from the list heads * for eligible buffers, greatly increasing its chance of finding them. * * The L2ARC device write speed is also boosted during this time so that * the L2ARC warms up faster. Since there have been no ARC evictions yet, * there are no L2ARC reads, and no fear of degrading read performance * through increased writes. * * 6. Writes to the L2ARC devices are grouped and sent in-sequence, so that * the vdev queue can aggregate them into larger and fewer writes. Each * device is written to in a rotor fashion, sweeping writes through * available space then repeating. * * 7. The L2ARC does not store dirty content. It never needs to flush * write buffers back to disk based storage. * * 8. If an ARC buffer is written (and dirtied) which also exists in the * L2ARC, the now stale L2ARC buffer is immediately dropped. * * The performance of the L2ARC can be tweaked by a number of tunables, which * may be necessary for different workloads: * * l2arc_write_max max write bytes per interval * l2arc_write_boost extra write bytes during device warmup * l2arc_noprefetch skip caching prefetched buffers * l2arc_nocompress skip compressing buffers * l2arc_headroom number of max device writes to precache * l2arc_headroom_boost when we find compressed buffers during ARC * scanning, we multiply headroom by this * percentage factor for the next scan cycle, * since more compressed buffers are likely to * be present * l2arc_feed_secs seconds between L2ARC writing * * Tunables may be removed or added as future performance improvements are * integrated, and also may become zpool properties. * * There are three key functions that control how the L2ARC warms up: * * l2arc_write_eligible() check if a buffer is eligible to cache * l2arc_write_size() calculate how much to write * l2arc_write_interval() calculate sleep delay between writes * * These three functions determine what to write, how much, and how quickly * to send writes. */ static boolean_t l2arc_write_eligible(uint64_t spa_guid, arc_buf_hdr_t *hdr) { /* * A buffer is *not* eligible for the L2ARC if it: * 1. belongs to a different spa. * 2. is already cached on the L2ARC. * 3. has an I/O in progress (it may be an incomplete read). * 4. is flagged not eligible (zfs property). */ if (hdr->b_spa != spa_guid || HDR_HAS_L2HDR(hdr) || HDR_IO_IN_PROGRESS(hdr) || !HDR_L2CACHE(hdr)) return (B_FALSE); return (B_TRUE); } static uint64_t l2arc_write_size(void) { uint64_t size; /* * Make sure our globals have meaningful values in case the user * altered them. */ size = l2arc_write_max; if (size == 0) { cmn_err(CE_NOTE, "Bad value for l2arc_write_max, value must " "be greater than zero, resetting it to the default (%d)", L2ARC_WRITE_SIZE); size = l2arc_write_max = L2ARC_WRITE_SIZE; } if (arc_warm == B_FALSE) size += l2arc_write_boost; return (size); } static clock_t l2arc_write_interval(clock_t began, uint64_t wanted, uint64_t wrote) { clock_t interval, next, now; /* * If the ARC lists are busy, increase our write rate; if the * lists are stale, idle back. This is achieved by checking * how much we previously wrote - if it was more than half of * what we wanted, schedule the next write much sooner. */ if (l2arc_feed_again && wrote > (wanted / 2)) interval = (hz * l2arc_feed_min_ms) / 1000; else interval = hz * l2arc_feed_secs; now = ddi_get_lbolt(); next = MAX(now, MIN(now + interval, began + interval)); return (next); } /* * Cycle through L2ARC devices. This is how L2ARC load balances. * If a device is returned, this also returns holding the spa config lock. */ static l2arc_dev_t * l2arc_dev_get_next(void) { l2arc_dev_t *first, *next = NULL; /* * Lock out the removal of spas (spa_namespace_lock), then removal * of cache devices (l2arc_dev_mtx). Once a device has been selected, * both locks will be dropped and a spa config lock held instead. */ mutex_enter(&spa_namespace_lock); mutex_enter(&l2arc_dev_mtx); /* if there are no vdevs, there is nothing to do */ if (l2arc_ndev == 0) goto out; first = NULL; next = l2arc_dev_last; do { /* loop around the list looking for a non-faulted vdev */ if (next == NULL) { next = list_head(l2arc_dev_list); } else { next = list_next(l2arc_dev_list, next); if (next == NULL) next = list_head(l2arc_dev_list); } /* if we have come back to the start, bail out */ if (first == NULL) first = next; else if (next == first) break; } while (vdev_is_dead(next->l2ad_vdev)); /* if we were unable to find any usable vdevs, return NULL */ if (vdev_is_dead(next->l2ad_vdev)) next = NULL; l2arc_dev_last = next; out: mutex_exit(&l2arc_dev_mtx); /* * Grab the config lock to prevent the 'next' device from being * removed while we are writing to it. */ if (next != NULL) spa_config_enter(next->l2ad_spa, SCL_L2ARC, next, RW_READER); mutex_exit(&spa_namespace_lock); return (next); } /* * Free buffers that were tagged for destruction. */ static void l2arc_do_free_on_write(void) { list_t *buflist; l2arc_data_free_t *df, *df_prev; mutex_enter(&l2arc_free_on_write_mtx); buflist = l2arc_free_on_write; for (df = list_tail(buflist); df; df = df_prev) { df_prev = list_prev(buflist, df); ASSERT(df->l2df_data != NULL); ASSERT(df->l2df_func != NULL); df->l2df_func(df->l2df_data, df->l2df_size); list_remove(buflist, df); kmem_free(df, sizeof (l2arc_data_free_t)); } mutex_exit(&l2arc_free_on_write_mtx); } /* * A write to a cache device has completed. Update all headers to allow * reads from these buffers to begin. */ static void l2arc_write_done(zio_t *zio) { l2arc_write_callback_t *cb; l2arc_dev_t *dev; list_t *buflist; arc_buf_hdr_t *head, *hdr, *hdr_prev; kmutex_t *hash_lock; int64_t bytes_dropped = 0; cb = zio->io_private; ASSERT(cb != NULL); dev = cb->l2wcb_dev; ASSERT(dev != NULL); head = cb->l2wcb_head; ASSERT(head != NULL); buflist = &dev->l2ad_buflist; ASSERT(buflist != NULL); DTRACE_PROBE2(l2arc__iodone, zio_t *, zio, l2arc_write_callback_t *, cb); if (zio->io_error != 0) ARCSTAT_BUMP(arcstat_l2_writes_error); /* * All writes completed, or an error was hit. */ top: mutex_enter(&dev->l2ad_mtx); for (hdr = list_prev(buflist, head); hdr; hdr = hdr_prev) { hdr_prev = list_prev(buflist, hdr); hash_lock = HDR_LOCK(hdr); /* * We cannot use mutex_enter or else we can deadlock * with l2arc_write_buffers (due to swapping the order * the hash lock and l2ad_mtx are taken). */ if (!mutex_tryenter(hash_lock)) { /* * Missed the hash lock. We must retry so we * don't leave the ARC_FLAG_L2_WRITING bit set. */ ARCSTAT_BUMP(arcstat_l2_writes_lock_retry); /* * We don't want to rescan the headers we've * already marked as having been written out, so * we reinsert the head node so we can pick up * where we left off. */ list_remove(buflist, head); list_insert_after(buflist, hdr, head); mutex_exit(&dev->l2ad_mtx); /* * We wait for the hash lock to become available * to try and prevent busy waiting, and increase * the chance we'll be able to acquire the lock * the next time around. */ mutex_enter(hash_lock); mutex_exit(hash_lock); goto top; } /* * We could not have been moved into the arc_l2c_only * state while in-flight due to our ARC_FLAG_L2_WRITING * bit being set. Let's just ensure that's being enforced. */ ASSERT(HDR_HAS_L1HDR(hdr)); /* * We may have allocated a buffer for L2ARC compression, * we must release it to avoid leaking this data. */ l2arc_release_cdata_buf(hdr); if (zio->io_error != 0) { /* * Error - drop L2ARC entry. */ list_remove(buflist, hdr); hdr->b_flags &= ~ARC_FLAG_HAS_L2HDR; ARCSTAT_INCR(arcstat_l2_asize, -hdr->b_l2hdr.b_asize); ARCSTAT_INCR(arcstat_l2_size, -hdr->b_size); bytes_dropped += hdr->b_l2hdr.b_asize; (void) refcount_remove_many(&dev->l2ad_alloc, hdr->b_l2hdr.b_asize, hdr); } /* * Allow ARC to begin reads and ghost list evictions to * this L2ARC entry. */ hdr->b_flags &= ~ARC_FLAG_L2_WRITING; mutex_exit(hash_lock); } atomic_inc_64(&l2arc_writes_done); list_remove(buflist, head); ASSERT(!HDR_HAS_L1HDR(head)); kmem_cache_free(hdr_l2only_cache, head); mutex_exit(&dev->l2ad_mtx); vdev_space_update(dev->l2ad_vdev, -bytes_dropped, 0, 0); l2arc_do_free_on_write(); kmem_free(cb, sizeof (l2arc_write_callback_t)); } /* * A read to a cache device completed. Validate buffer contents before * handing over to the regular ARC routines. */ static void l2arc_read_done(zio_t *zio) { l2arc_read_callback_t *cb; arc_buf_hdr_t *hdr; arc_buf_t *buf; kmutex_t *hash_lock; int equal; ASSERT(zio->io_vd != NULL); ASSERT(zio->io_flags & ZIO_FLAG_DONT_PROPAGATE); spa_config_exit(zio->io_spa, SCL_L2ARC, zio->io_vd); cb = zio->io_private; ASSERT(cb != NULL); buf = cb->l2rcb_buf; ASSERT(buf != NULL); hash_lock = HDR_LOCK(buf->b_hdr); mutex_enter(hash_lock); hdr = buf->b_hdr; ASSERT3P(hash_lock, ==, HDR_LOCK(hdr)); /* * If the buffer was compressed, decompress it first. */ if (cb->l2rcb_compress != ZIO_COMPRESS_OFF) l2arc_decompress_zio(zio, hdr, cb->l2rcb_compress); ASSERT(zio->io_data != NULL); ASSERT3U(zio->io_size, ==, hdr->b_size); ASSERT3U(BP_GET_LSIZE(&cb->l2rcb_bp), ==, hdr->b_size); /* * Check this survived the L2ARC journey. */ equal = arc_cksum_equal(buf); if (equal && zio->io_error == 0 && !HDR_L2_EVICTED(hdr)) { mutex_exit(hash_lock); zio->io_private = buf; zio->io_bp_copy = cb->l2rcb_bp; /* XXX fix in L2ARC 2.0 */ zio->io_bp = &zio->io_bp_copy; /* XXX fix in L2ARC 2.0 */ arc_read_done(zio); } else { mutex_exit(hash_lock); /* * Buffer didn't survive caching. Increment stats and * reissue to the original storage device. */ if (zio->io_error != 0) { ARCSTAT_BUMP(arcstat_l2_io_error); } else { zio->io_error = SET_ERROR(EIO); } if (!equal) ARCSTAT_BUMP(arcstat_l2_cksum_bad); /* * If there's no waiter, issue an async i/o to the primary * storage now. If there *is* a waiter, the caller must * issue the i/o in a context where it's OK to block. */ if (zio->io_waiter == NULL) { zio_t *pio = zio_unique_parent(zio); ASSERT(!pio || pio->io_child_type == ZIO_CHILD_LOGICAL); zio_nowait(zio_read(pio, cb->l2rcb_spa, &cb->l2rcb_bp, buf->b_data, hdr->b_size, arc_read_done, buf, zio->io_priority, cb->l2rcb_flags, &cb->l2rcb_zb)); } } kmem_free(cb, sizeof (l2arc_read_callback_t)); } /* * This is the list priority from which the L2ARC will search for pages to * cache. This is used within loops (0..3) to cycle through lists in the * desired order. This order can have a significant effect on cache * performance. * * Currently the metadata lists are hit first, MFU then MRU, followed by * the data lists. This function returns a locked list, and also returns * the lock pointer. */ static multilist_sublist_t * l2arc_sublist_lock(int list_num) { multilist_t *ml = NULL; unsigned int idx; ASSERT(list_num >= 0 && list_num <= 3); switch (list_num) { case 0: ml = &arc_mfu->arcs_list[ARC_BUFC_METADATA]; break; case 1: ml = &arc_mru->arcs_list[ARC_BUFC_METADATA]; break; case 2: ml = &arc_mfu->arcs_list[ARC_BUFC_DATA]; break; case 3: ml = &arc_mru->arcs_list[ARC_BUFC_DATA]; break; } /* * Return a randomly-selected sublist. This is acceptable * because the caller feeds only a little bit of data for each * call (8MB). Subsequent calls will result in different * sublists being selected. */ idx = multilist_get_random_index(ml); return (multilist_sublist_lock(ml, idx)); } /* * Evict buffers from the device write hand to the distance specified in * bytes. This distance may span populated buffers, it may span nothing. * This is clearing a region on the L2ARC device ready for writing. * If the 'all' boolean is set, every buffer is evicted. */ static void l2arc_evict(l2arc_dev_t *dev, uint64_t distance, boolean_t all) { list_t *buflist; arc_buf_hdr_t *hdr, *hdr_prev; kmutex_t *hash_lock; uint64_t taddr; buflist = &dev->l2ad_buflist; if (!all && dev->l2ad_first) { /* * This is the first sweep through the device. There is * nothing to evict. */ return; } if (dev->l2ad_hand >= (dev->l2ad_end - (2 * distance))) { /* * When nearing the end of the device, evict to the end * before the device write hand jumps to the start. */ taddr = dev->l2ad_end; } else { taddr = dev->l2ad_hand + distance; } DTRACE_PROBE4(l2arc__evict, l2arc_dev_t *, dev, list_t *, buflist, uint64_t, taddr, boolean_t, all); top: mutex_enter(&dev->l2ad_mtx); for (hdr = list_tail(buflist); hdr; hdr = hdr_prev) { hdr_prev = list_prev(buflist, hdr); hash_lock = HDR_LOCK(hdr); /* * We cannot use mutex_enter or else we can deadlock * with l2arc_write_buffers (due to swapping the order * the hash lock and l2ad_mtx are taken). */ if (!mutex_tryenter(hash_lock)) { /* * Missed the hash lock. Retry. */ ARCSTAT_BUMP(arcstat_l2_evict_lock_retry); mutex_exit(&dev->l2ad_mtx); mutex_enter(hash_lock); mutex_exit(hash_lock); goto top; } if (HDR_L2_WRITE_HEAD(hdr)) { /* * We hit a write head node. Leave it for * l2arc_write_done(). */ list_remove(buflist, hdr); mutex_exit(hash_lock); continue; } if (!all && HDR_HAS_L2HDR(hdr) && (hdr->b_l2hdr.b_daddr > taddr || hdr->b_l2hdr.b_daddr < dev->l2ad_hand)) { /* * We've evicted to the target address, * or the end of the device. */ mutex_exit(hash_lock); break; } ASSERT(HDR_HAS_L2HDR(hdr)); if (!HDR_HAS_L1HDR(hdr)) { ASSERT(!HDR_L2_READING(hdr)); /* * This doesn't exist in the ARC. Destroy. * arc_hdr_destroy() will call list_remove() * and decrement arcstat_l2_size. */ arc_change_state(arc_anon, hdr, hash_lock); arc_hdr_destroy(hdr); } else { ASSERT(hdr->b_l1hdr.b_state != arc_l2c_only); ARCSTAT_BUMP(arcstat_l2_evict_l1cached); /* * Invalidate issued or about to be issued * reads, since we may be about to write * over this location. */ if (HDR_L2_READING(hdr)) { ARCSTAT_BUMP(arcstat_l2_evict_reading); hdr->b_flags |= ARC_FLAG_L2_EVICTED; } /* Ensure this header has finished being written */ ASSERT(!HDR_L2_WRITING(hdr)); ASSERT3P(hdr->b_l1hdr.b_tmp_cdata, ==, NULL); arc_hdr_l2hdr_destroy(hdr); } mutex_exit(hash_lock); } mutex_exit(&dev->l2ad_mtx); } /* * Find and write ARC buffers to the L2ARC device. * * An ARC_FLAG_L2_WRITING flag is set so that the L2ARC buffers are not valid * for reading until they have completed writing. * The headroom_boost is an in-out parameter used to maintain headroom boost * state between calls to this function. * * Returns the number of bytes actually written (which may be smaller than * the delta by which the device hand has changed due to alignment). */ static uint64_t l2arc_write_buffers(spa_t *spa, l2arc_dev_t *dev, uint64_t target_sz, boolean_t *headroom_boost) { arc_buf_hdr_t *hdr, *hdr_prev, *head; uint64_t write_asize, write_sz, headroom, buf_compress_minsz, stats_size; void *buf_data; boolean_t full; l2arc_write_callback_t *cb; zio_t *pio, *wzio; uint64_t guid = spa_load_guid(spa); int try; const boolean_t do_headroom_boost = *headroom_boost; ASSERT(dev->l2ad_vdev != NULL); /* Lower the flag now, we might want to raise it again later. */ *headroom_boost = B_FALSE; pio = NULL; write_sz = write_asize = 0; full = B_FALSE; head = kmem_cache_alloc(hdr_l2only_cache, KM_PUSHPAGE); head->b_flags |= ARC_FLAG_L2_WRITE_HEAD; head->b_flags |= ARC_FLAG_HAS_L2HDR; /* * We will want to try to compress buffers that are at least 2x the * device sector size. */ buf_compress_minsz = 2 << dev->l2ad_vdev->vdev_ashift; /* * Copy buffers for L2ARC writing. */ for (try = 0; try <= 3; try++) { multilist_sublist_t *mls = l2arc_sublist_lock(try); uint64_t passed_sz = 0; /* * L2ARC fast warmup. * * Until the ARC is warm and starts to evict, read from the * head of the ARC lists rather than the tail. */ if (arc_warm == B_FALSE) hdr = multilist_sublist_head(mls); else hdr = multilist_sublist_tail(mls); headroom = target_sz * l2arc_headroom; if (do_headroom_boost) headroom = (headroom * l2arc_headroom_boost) / 100; for (; hdr; hdr = hdr_prev) { kmutex_t *hash_lock; uint64_t buf_sz; uint64_t buf_a_sz; if (arc_warm == B_FALSE) hdr_prev = multilist_sublist_next(mls, hdr); else hdr_prev = multilist_sublist_prev(mls, hdr); hash_lock = HDR_LOCK(hdr); if (!mutex_tryenter(hash_lock)) { /* * Skip this buffer rather than waiting. */ continue; } passed_sz += hdr->b_size; if (passed_sz > headroom) { /* * Searched too far. */ mutex_exit(hash_lock); break; } if (!l2arc_write_eligible(guid, hdr)) { mutex_exit(hash_lock); continue; } /* * Assume that the buffer is not going to be compressed * and could take more space on disk because of a larger * disk block size. */ buf_sz = hdr->b_size; buf_a_sz = vdev_psize_to_asize(dev->l2ad_vdev, buf_sz); if ((write_asize + buf_a_sz) > target_sz) { full = B_TRUE; mutex_exit(hash_lock); break; } if (pio == NULL) { /* * Insert a dummy header on the buflist so * l2arc_write_done() can find where the * write buffers begin without searching. */ mutex_enter(&dev->l2ad_mtx); list_insert_head(&dev->l2ad_buflist, head); mutex_exit(&dev->l2ad_mtx); cb = kmem_alloc( sizeof (l2arc_write_callback_t), KM_SLEEP); cb->l2wcb_dev = dev; cb->l2wcb_head = head; pio = zio_root(spa, l2arc_write_done, cb, ZIO_FLAG_CANFAIL); } /* * Create and add a new L2ARC header. */ hdr->b_l2hdr.b_dev = dev; hdr->b_flags |= ARC_FLAG_L2_WRITING; /* * Temporarily stash the data buffer in b_tmp_cdata. * The subsequent write step will pick it up from * there. This is because can't access b_l1hdr.b_buf * without holding the hash_lock, which we in turn * can't access without holding the ARC list locks * (which we want to avoid during compression/writing) */ hdr->b_l2hdr.b_compress = ZIO_COMPRESS_OFF; hdr->b_l2hdr.b_asize = hdr->b_size; hdr->b_l2hdr.b_hits = 0; hdr->b_l1hdr.b_tmp_cdata = hdr->b_l1hdr.b_buf->b_data; /* * Explicitly set the b_daddr field to a known * value which means "invalid address". This * enables us to differentiate which stage of * l2arc_write_buffers() the particular header * is in (e.g. this loop, or the one below). * ARC_FLAG_L2_WRITING is not enough to make * this distinction, and we need to know in * order to do proper l2arc vdev accounting in * arc_release() and arc_hdr_destroy(). * * Note, we can't use a new flag to distinguish * the two stages because we don't hold the * header's hash_lock below, in the second stage * of this function. Thus, we can't simply * change the b_flags field to denote that the * IO has been sent. We can change the b_daddr * field of the L2 portion, though, since we'll * be holding the l2ad_mtx; which is why we're * using it to denote the header's state change. */ hdr->b_l2hdr.b_daddr = L2ARC_ADDR_UNSET; hdr->b_flags |= ARC_FLAG_HAS_L2HDR; mutex_enter(&dev->l2ad_mtx); list_insert_head(&dev->l2ad_buflist, hdr); mutex_exit(&dev->l2ad_mtx); /* * Compute and store the buffer cksum before * writing. On debug the cksum is verified first. */ arc_cksum_verify(hdr->b_l1hdr.b_buf); arc_cksum_compute(hdr->b_l1hdr.b_buf, B_TRUE); mutex_exit(hash_lock); write_sz += buf_sz; write_asize += buf_a_sz; } multilist_sublist_unlock(mls); if (full == B_TRUE) break; } /* No buffers selected for writing? */ if (pio == NULL) { ASSERT0(write_sz); ASSERT(!HDR_HAS_L1HDR(head)); kmem_cache_free(hdr_l2only_cache, head); return (0); } mutex_enter(&dev->l2ad_mtx); /* * Note that elsewhere in this file arcstat_l2_asize * and the used space on l2ad_vdev are updated using b_asize, * which is not necessarily rounded up to the device block size. * Too keep accounting consistent we do the same here as well: * stats_size accumulates the sum of b_asize of the written buffers, * while write_asize accumulates the sum of b_asize rounded up * to the device block size. * The latter sum is used only to validate the corectness of the code. */ stats_size = 0; write_asize = 0; /* * Now start writing the buffers. We're starting at the write head * and work backwards, retracing the course of the buffer selector * loop above. */ for (hdr = list_prev(&dev->l2ad_buflist, head); hdr; hdr = list_prev(&dev->l2ad_buflist, hdr)) { uint64_t buf_sz; /* * We rely on the L1 portion of the header below, so * it's invalid for this header to have been evicted out * of the ghost cache, prior to being written out. The * ARC_FLAG_L2_WRITING bit ensures this won't happen. */ ASSERT(HDR_HAS_L1HDR(hdr)); /* * We shouldn't need to lock the buffer here, since we flagged * it as ARC_FLAG_L2_WRITING in the previous step, but we must * take care to only access its L2 cache parameters. In * particular, hdr->l1hdr.b_buf may be invalid by now due to * ARC eviction. */ hdr->b_l2hdr.b_daddr = dev->l2ad_hand; if ((!l2arc_nocompress && HDR_L2COMPRESS(hdr)) && hdr->b_l2hdr.b_asize >= buf_compress_minsz) { if (l2arc_compress_buf(hdr)) { /* * If compression succeeded, enable headroom * boost on the next scan cycle. */ *headroom_boost = B_TRUE; } } /* * Pick up the buffer data we had previously stashed away * (and now potentially also compressed). */ buf_data = hdr->b_l1hdr.b_tmp_cdata; buf_sz = hdr->b_l2hdr.b_asize; /* * We need to do this regardless if buf_sz is zero or * not, otherwise, when this l2hdr is evicted we'll * remove a reference that was never added. */ (void) refcount_add_many(&dev->l2ad_alloc, buf_sz, hdr); /* Compression may have squashed the buffer to zero length. */ if (buf_sz != 0) { uint64_t buf_a_sz; wzio = zio_write_phys(pio, dev->l2ad_vdev, dev->l2ad_hand, buf_sz, buf_data, ZIO_CHECKSUM_OFF, NULL, NULL, ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_CANFAIL, B_FALSE); DTRACE_PROBE2(l2arc__write, vdev_t *, dev->l2ad_vdev, zio_t *, wzio); (void) zio_nowait(wzio); stats_size += buf_sz; /* * Keep the clock hand suitably device-aligned. */ buf_a_sz = vdev_psize_to_asize(dev->l2ad_vdev, buf_sz); write_asize += buf_a_sz; dev->l2ad_hand += buf_a_sz; } } mutex_exit(&dev->l2ad_mtx); ASSERT3U(write_asize, <=, target_sz); ARCSTAT_BUMP(arcstat_l2_writes_sent); ARCSTAT_INCR(arcstat_l2_write_bytes, write_asize); ARCSTAT_INCR(arcstat_l2_size, write_sz); ARCSTAT_INCR(arcstat_l2_asize, stats_size); vdev_space_update(dev->l2ad_vdev, stats_size, 0, 0); /* * Bump device hand to the device start if it is approaching the end. * l2arc_evict() will already have evicted ahead for this case. */ if (dev->l2ad_hand >= (dev->l2ad_end - target_sz)) { dev->l2ad_hand = dev->l2ad_start; dev->l2ad_first = B_FALSE; } dev->l2ad_writing = B_TRUE; (void) zio_wait(pio); dev->l2ad_writing = B_FALSE; return (write_asize); } /* * Compresses an L2ARC buffer. * The data to be compressed must be prefilled in l1hdr.b_tmp_cdata and its * size in l2hdr->b_asize. This routine tries to compress the data and * depending on the compression result there are three possible outcomes: * *) The buffer was incompressible. The original l2hdr contents were left * untouched and are ready for writing to an L2 device. * *) The buffer was all-zeros, so there is no need to write it to an L2 * device. To indicate this situation b_tmp_cdata is NULL'ed, b_asize is * set to zero and b_compress is set to ZIO_COMPRESS_EMPTY. * *) Compression succeeded and b_tmp_cdata was replaced with a temporary * data buffer which holds the compressed data to be written, and b_asize * tells us how much data there is. b_compress is set to the appropriate * compression algorithm. Once writing is done, invoke * l2arc_release_cdata_buf on this l2hdr to free this temporary buffer. * * Returns B_TRUE if compression succeeded, or B_FALSE if it didn't (the * buffer was incompressible). */ static boolean_t l2arc_compress_buf(arc_buf_hdr_t *hdr) { void *cdata; size_t csize, len, rounded; l2arc_buf_hdr_t *l2hdr; ASSERT(HDR_HAS_L2HDR(hdr)); l2hdr = &hdr->b_l2hdr; ASSERT(HDR_HAS_L1HDR(hdr)); ASSERT3U(l2hdr->b_compress, ==, ZIO_COMPRESS_OFF); ASSERT(hdr->b_l1hdr.b_tmp_cdata != NULL); len = l2hdr->b_asize; cdata = zio_data_buf_alloc(len); ASSERT3P(cdata, !=, NULL); csize = zio_compress_data(ZIO_COMPRESS_LZ4, hdr->b_l1hdr.b_tmp_cdata, cdata, l2hdr->b_asize); rounded = P2ROUNDUP(csize, (size_t)SPA_MINBLOCKSIZE); if (rounded > csize) { bzero((char *)cdata + csize, rounded - csize); csize = rounded; } if (csize == 0) { /* zero block, indicate that there's nothing to write */ zio_data_buf_free(cdata, len); l2hdr->b_compress = ZIO_COMPRESS_EMPTY; l2hdr->b_asize = 0; hdr->b_l1hdr.b_tmp_cdata = NULL; ARCSTAT_BUMP(arcstat_l2_compress_zeros); return (B_TRUE); } else if (csize > 0 && csize < len) { /* * Compression succeeded, we'll keep the cdata around for * writing and release it afterwards. */ l2hdr->b_compress = ZIO_COMPRESS_LZ4; l2hdr->b_asize = csize; hdr->b_l1hdr.b_tmp_cdata = cdata; ARCSTAT_BUMP(arcstat_l2_compress_successes); return (B_TRUE); } else { /* * Compression failed, release the compressed buffer. * l2hdr will be left unmodified. */ zio_data_buf_free(cdata, len); ARCSTAT_BUMP(arcstat_l2_compress_failures); return (B_FALSE); } } /* * Decompresses a zio read back from an l2arc device. On success, the * underlying zio's io_data buffer is overwritten by the uncompressed * version. On decompression error (corrupt compressed stream), the * zio->io_error value is set to signal an I/O error. * * Please note that the compressed data stream is not checksummed, so * if the underlying device is experiencing data corruption, we may feed * corrupt data to the decompressor, so the decompressor needs to be * able to handle this situation (LZ4 does). */ static void l2arc_decompress_zio(zio_t *zio, arc_buf_hdr_t *hdr, enum zio_compress c) { uint64_t csize; void *cdata; ASSERT(L2ARC_IS_VALID_COMPRESS(c)); if (zio->io_error != 0) { /* * An io error has occured, just restore the original io * size in preparation for a main pool read. */ zio->io_orig_size = zio->io_size = hdr->b_size; return; } if (c == ZIO_COMPRESS_EMPTY) { /* * An empty buffer results in a null zio, which means we * need to fill its io_data after we're done restoring the * buffer's contents. */ ASSERT(hdr->b_l1hdr.b_buf != NULL); bzero(hdr->b_l1hdr.b_buf->b_data, hdr->b_size); zio->io_data = zio->io_orig_data = hdr->b_l1hdr.b_buf->b_data; } else { ASSERT(zio->io_data != NULL); /* * We copy the compressed data from the start of the arc buffer * (the zio_read will have pulled in only what we need, the * rest is garbage which we will overwrite at decompression) * and then decompress back to the ARC data buffer. This way we * can minimize copying by simply decompressing back over the * original compressed data (rather than decompressing to an * aux buffer and then copying back the uncompressed buffer, * which is likely to be much larger). */ csize = zio->io_size; cdata = zio_data_buf_alloc(csize); bcopy(zio->io_data, cdata, csize); if (zio_decompress_data(c, cdata, zio->io_data, csize, hdr->b_size) != 0) zio->io_error = EIO; zio_data_buf_free(cdata, csize); } /* Restore the expected uncompressed IO size. */ zio->io_orig_size = zio->io_size = hdr->b_size; } /* * Releases the temporary b_tmp_cdata buffer in an l2arc header structure. * This buffer serves as a temporary holder of compressed data while * the buffer entry is being written to an l2arc device. Once that is * done, we can dispose of it. */ static void l2arc_release_cdata_buf(arc_buf_hdr_t *hdr) { enum zio_compress comp; ASSERT(HDR_HAS_L1HDR(hdr)); ASSERT(HDR_HAS_L2HDR(hdr)); comp = hdr->b_l2hdr.b_compress; ASSERT(comp == ZIO_COMPRESS_OFF || L2ARC_IS_VALID_COMPRESS(comp)); if (comp == ZIO_COMPRESS_OFF) { /* * In this case, b_tmp_cdata points to the same buffer * as the arc_buf_t's b_data field. We don't want to * free it, since the arc_buf_t will handle that. */ hdr->b_l1hdr.b_tmp_cdata = NULL; } else if (comp == ZIO_COMPRESS_EMPTY) { /* * In this case, b_tmp_cdata was compressed to an empty * buffer, thus there's nothing to free and b_tmp_cdata * should have been set to NULL in l2arc_write_buffers(). */ ASSERT3P(hdr->b_l1hdr.b_tmp_cdata, ==, NULL); } else { /* * If the data was compressed, then we've allocated a * temporary buffer for it, so now we need to release it. */ ASSERT(hdr->b_l1hdr.b_tmp_cdata != NULL); zio_data_buf_free(hdr->b_l1hdr.b_tmp_cdata, hdr->b_size); hdr->b_l1hdr.b_tmp_cdata = NULL; } } /* * This thread feeds the L2ARC at regular intervals. This is the beating * heart of the L2ARC. */ static void l2arc_feed_thread(void) { callb_cpr_t cpr; l2arc_dev_t *dev; spa_t *spa; uint64_t size, wrote; clock_t begin, next = ddi_get_lbolt(); boolean_t headroom_boost = B_FALSE; fstrans_cookie_t cookie; CALLB_CPR_INIT(&cpr, &l2arc_feed_thr_lock, callb_generic_cpr, FTAG); mutex_enter(&l2arc_feed_thr_lock); cookie = spl_fstrans_mark(); while (l2arc_thread_exit == 0) { CALLB_CPR_SAFE_BEGIN(&cpr); (void) cv_timedwait_sig(&l2arc_feed_thr_cv, &l2arc_feed_thr_lock, next); CALLB_CPR_SAFE_END(&cpr, &l2arc_feed_thr_lock); next = ddi_get_lbolt() + hz; /* * Quick check for L2ARC devices. */ mutex_enter(&l2arc_dev_mtx); if (l2arc_ndev == 0) { mutex_exit(&l2arc_dev_mtx); continue; } mutex_exit(&l2arc_dev_mtx); begin = ddi_get_lbolt(); /* * This selects the next l2arc device to write to, and in * doing so the next spa to feed from: dev->l2ad_spa. This * will return NULL if there are now no l2arc devices or if * they are all faulted. * * If a device is returned, its spa's config lock is also * held to prevent device removal. l2arc_dev_get_next() * will grab and release l2arc_dev_mtx. */ if ((dev = l2arc_dev_get_next()) == NULL) continue; spa = dev->l2ad_spa; ASSERT(spa != NULL); /* * If the pool is read-only then force the feed thread to * sleep a little longer. */ if (!spa_writeable(spa)) { next = ddi_get_lbolt() + 5 * l2arc_feed_secs * hz; spa_config_exit(spa, SCL_L2ARC, dev); continue; } /* * Avoid contributing to memory pressure. */ if (arc_reclaim_needed()) { ARCSTAT_BUMP(arcstat_l2_abort_lowmem); spa_config_exit(spa, SCL_L2ARC, dev); continue; } ARCSTAT_BUMP(arcstat_l2_feeds); size = l2arc_write_size(); /* * Evict L2ARC buffers that will be overwritten. */ l2arc_evict(dev, size, B_FALSE); /* * Write ARC buffers. */ wrote = l2arc_write_buffers(spa, dev, size, &headroom_boost); /* * Calculate interval between writes. */ next = l2arc_write_interval(begin, size, wrote); spa_config_exit(spa, SCL_L2ARC, dev); } spl_fstrans_unmark(cookie); l2arc_thread_exit = 0; cv_broadcast(&l2arc_feed_thr_cv); CALLB_CPR_EXIT(&cpr); /* drops l2arc_feed_thr_lock */ thread_exit(); } boolean_t l2arc_vdev_present(vdev_t *vd) { l2arc_dev_t *dev; mutex_enter(&l2arc_dev_mtx); for (dev = list_head(l2arc_dev_list); dev != NULL; dev = list_next(l2arc_dev_list, dev)) { if (dev->l2ad_vdev == vd) break; } mutex_exit(&l2arc_dev_mtx); return (dev != NULL); } /* * Add a vdev for use by the L2ARC. By this point the spa has already * validated the vdev and opened it. */ void l2arc_add_vdev(spa_t *spa, vdev_t *vd) { l2arc_dev_t *adddev; ASSERT(!l2arc_vdev_present(vd)); /* * Create a new l2arc device entry. */ adddev = kmem_zalloc(sizeof (l2arc_dev_t), KM_SLEEP); adddev->l2ad_spa = spa; adddev->l2ad_vdev = vd; adddev->l2ad_start = VDEV_LABEL_START_SIZE; adddev->l2ad_end = VDEV_LABEL_START_SIZE + vdev_get_min_asize(vd); adddev->l2ad_hand = adddev->l2ad_start; adddev->l2ad_first = B_TRUE; adddev->l2ad_writing = B_FALSE; list_link_init(&adddev->l2ad_node); mutex_init(&adddev->l2ad_mtx, NULL, MUTEX_DEFAULT, NULL); /* * This is a list of all ARC buffers that are still valid on the * device. */ list_create(&adddev->l2ad_buflist, sizeof (arc_buf_hdr_t), offsetof(arc_buf_hdr_t, b_l2hdr.b_l2node)); vdev_space_update(vd, 0, 0, adddev->l2ad_end - adddev->l2ad_hand); refcount_create(&adddev->l2ad_alloc); /* * Add device to global list */ mutex_enter(&l2arc_dev_mtx); list_insert_head(l2arc_dev_list, adddev); atomic_inc_64(&l2arc_ndev); mutex_exit(&l2arc_dev_mtx); } /* * Remove a vdev from the L2ARC. */ void l2arc_remove_vdev(vdev_t *vd) { l2arc_dev_t *dev, *nextdev, *remdev = NULL; /* * Find the device by vdev */ mutex_enter(&l2arc_dev_mtx); for (dev = list_head(l2arc_dev_list); dev; dev = nextdev) { nextdev = list_next(l2arc_dev_list, dev); if (vd == dev->l2ad_vdev) { remdev = dev; break; } } ASSERT(remdev != NULL); /* * Remove device from global list */ list_remove(l2arc_dev_list, remdev); l2arc_dev_last = NULL; /* may have been invalidated */ atomic_dec_64(&l2arc_ndev); mutex_exit(&l2arc_dev_mtx); /* * Clear all buflists and ARC references. L2ARC device flush. */ l2arc_evict(remdev, 0, B_TRUE); list_destroy(&remdev->l2ad_buflist); mutex_destroy(&remdev->l2ad_mtx); refcount_destroy(&remdev->l2ad_alloc); kmem_free(remdev, sizeof (l2arc_dev_t)); } void l2arc_init(void) { l2arc_thread_exit = 0; l2arc_ndev = 0; l2arc_writes_sent = 0; l2arc_writes_done = 0; mutex_init(&l2arc_feed_thr_lock, NULL, MUTEX_DEFAULT, NULL); cv_init(&l2arc_feed_thr_cv, NULL, CV_DEFAULT, NULL); mutex_init(&l2arc_dev_mtx, NULL, MUTEX_DEFAULT, NULL); mutex_init(&l2arc_free_on_write_mtx, NULL, MUTEX_DEFAULT, NULL); l2arc_dev_list = &L2ARC_dev_list; l2arc_free_on_write = &L2ARC_free_on_write; list_create(l2arc_dev_list, sizeof (l2arc_dev_t), offsetof(l2arc_dev_t, l2ad_node)); list_create(l2arc_free_on_write, sizeof (l2arc_data_free_t), offsetof(l2arc_data_free_t, l2df_list_node)); } void l2arc_fini(void) { /* * This is called from dmu_fini(), which is called from spa_fini(); * Because of this, we can assume that all l2arc devices have * already been removed when the pools themselves were removed. */ l2arc_do_free_on_write(); mutex_destroy(&l2arc_feed_thr_lock); cv_destroy(&l2arc_feed_thr_cv); mutex_destroy(&l2arc_dev_mtx); mutex_destroy(&l2arc_free_on_write_mtx); list_destroy(l2arc_dev_list); list_destroy(l2arc_free_on_write); } void l2arc_start(void) { if (!(spa_mode_global & FWRITE)) return; (void) thread_create(NULL, 0, l2arc_feed_thread, NULL, 0, &p0, TS_RUN, defclsyspri); } void l2arc_stop(void) { if (!(spa_mode_global & FWRITE)) return; mutex_enter(&l2arc_feed_thr_lock); cv_signal(&l2arc_feed_thr_cv); /* kick thread out of startup */ l2arc_thread_exit = 1; while (l2arc_thread_exit != 0) cv_wait(&l2arc_feed_thr_cv, &l2arc_feed_thr_lock); mutex_exit(&l2arc_feed_thr_lock); } #if defined(_KERNEL) && defined(HAVE_SPL) EXPORT_SYMBOL(arc_buf_size); EXPORT_SYMBOL(arc_write); EXPORT_SYMBOL(arc_read); EXPORT_SYMBOL(arc_buf_remove_ref); EXPORT_SYMBOL(arc_buf_info); EXPORT_SYMBOL(arc_getbuf_func); EXPORT_SYMBOL(arc_add_prune_callback); EXPORT_SYMBOL(arc_remove_prune_callback); module_param(zfs_arc_min, ulong, 0644); MODULE_PARM_DESC(zfs_arc_min, "Min arc size"); module_param(zfs_arc_max, ulong, 0644); MODULE_PARM_DESC(zfs_arc_max, "Max arc size"); module_param(zfs_arc_meta_limit, ulong, 0644); MODULE_PARM_DESC(zfs_arc_meta_limit, "Meta limit for arc size"); module_param(zfs_arc_meta_min, ulong, 0644); MODULE_PARM_DESC(zfs_arc_meta_min, "Min arc metadata"); module_param(zfs_arc_meta_prune, int, 0644); MODULE_PARM_DESC(zfs_arc_meta_prune, "Meta objects to scan for prune"); module_param(zfs_arc_meta_adjust_restarts, int, 0644); MODULE_PARM_DESC(zfs_arc_meta_adjust_restarts, "Limit number of restarts in arc_adjust_meta"); module_param(zfs_arc_meta_strategy, int, 0644); MODULE_PARM_DESC(zfs_arc_meta_strategy, "Meta reclaim strategy"); module_param(zfs_arc_grow_retry, int, 0644); MODULE_PARM_DESC(zfs_arc_grow_retry, "Seconds before growing arc size"); module_param(zfs_arc_p_aggressive_disable, int, 0644); MODULE_PARM_DESC(zfs_arc_p_aggressive_disable, "disable aggressive arc_p grow"); module_param(zfs_arc_p_dampener_disable, int, 0644); MODULE_PARM_DESC(zfs_arc_p_dampener_disable, "disable arc_p adapt dampener"); module_param(zfs_arc_shrink_shift, int, 0644); MODULE_PARM_DESC(zfs_arc_shrink_shift, "log2(fraction of arc to reclaim)"); module_param(zfs_arc_p_min_shift, int, 0644); MODULE_PARM_DESC(zfs_arc_p_min_shift, "arc_c shift to calc min/max arc_p"); module_param(zfs_disable_dup_eviction, int, 0644); MODULE_PARM_DESC(zfs_disable_dup_eviction, "disable duplicate buffer eviction"); module_param(zfs_arc_average_blocksize, int, 0444); MODULE_PARM_DESC(zfs_arc_average_blocksize, "Target average block size"); module_param(zfs_arc_min_prefetch_lifespan, int, 0644); MODULE_PARM_DESC(zfs_arc_min_prefetch_lifespan, "Min life of prefetch block"); module_param(zfs_arc_num_sublists_per_state, int, 0644); MODULE_PARM_DESC(zfs_arc_num_sublists_per_state, "Number of sublists used in each of the ARC state lists"); module_param(l2arc_write_max, ulong, 0644); MODULE_PARM_DESC(l2arc_write_max, "Max write bytes per interval"); module_param(l2arc_write_boost, ulong, 0644); MODULE_PARM_DESC(l2arc_write_boost, "Extra write bytes during device warmup"); module_param(l2arc_headroom, ulong, 0644); MODULE_PARM_DESC(l2arc_headroom, "Number of max device writes to precache"); module_param(l2arc_headroom_boost, ulong, 0644); MODULE_PARM_DESC(l2arc_headroom_boost, "Compressed l2arc_headroom multiplier"); module_param(l2arc_feed_secs, ulong, 0644); MODULE_PARM_DESC(l2arc_feed_secs, "Seconds between L2ARC writing"); module_param(l2arc_feed_min_ms, ulong, 0644); MODULE_PARM_DESC(l2arc_feed_min_ms, "Min feed interval in milliseconds"); module_param(l2arc_noprefetch, int, 0644); MODULE_PARM_DESC(l2arc_noprefetch, "Skip caching prefetched buffers"); module_param(l2arc_nocompress, int, 0644); MODULE_PARM_DESC(l2arc_nocompress, "Skip compressing L2ARC buffers"); module_param(l2arc_feed_again, int, 0644); MODULE_PARM_DESC(l2arc_feed_again, "Turbo L2ARC warmup"); module_param(l2arc_norw, int, 0644); MODULE_PARM_DESC(l2arc_norw, "No reads during writes"); module_param(zfs_arc_lotsfree_percent, int, 0644); MODULE_PARM_DESC(zfs_arc_lotsfree_percent, "System free memory I/O throttle in bytes"); module_param(zfs_arc_sys_free, ulong, 0644); MODULE_PARM_DESC(zfs_arc_sys_free, "System free memory target size in bytes"); #endif diff --git a/module/zfs/dbuf.c b/module/zfs/dbuf.c index e08dcc4a3d27..d53d8607aba9 100644 --- a/module/zfs/dbuf.c +++ b/module/zfs/dbuf.c @@ -1,3357 +1,3354 @@ /* * 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) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright 2011 Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2012, 2015 by Delphix. All rights reserved. * Copyright (c) 2013 by Saso Kiselkov. All rights reserved. * Copyright (c) 2014 Spectra Logic Corporation, All rights reserved. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include struct dbuf_hold_impl_data { /* Function arguments */ dnode_t *dh_dn; uint8_t dh_level; uint64_t dh_blkid; boolean_t dh_fail_sparse; boolean_t dh_fail_uncached; void *dh_tag; dmu_buf_impl_t **dh_dbp; /* Local variables */ dmu_buf_impl_t *dh_db; dmu_buf_impl_t *dh_parent; blkptr_t *dh_bp; int dh_err; dbuf_dirty_record_t *dh_dr; arc_buf_contents_t dh_type; int dh_depth; }; static void __dbuf_hold_impl_init(struct dbuf_hold_impl_data *dh, dnode_t *dn, uint8_t level, uint64_t blkid, boolean_t fail_sparse, boolean_t fail_uncached, void *tag, dmu_buf_impl_t **dbp, int depth); static int __dbuf_hold_impl(struct dbuf_hold_impl_data *dh); /* * Number of times that zfs_free_range() took the slow path while doing * a zfs receive. A nonzero value indicates a potential performance problem. */ uint64_t zfs_free_range_recv_miss; static void dbuf_destroy(dmu_buf_impl_t *db); static boolean_t dbuf_undirty(dmu_buf_impl_t *db, dmu_tx_t *tx); static void dbuf_write(dbuf_dirty_record_t *dr, arc_buf_t *data, dmu_tx_t *tx); #ifndef __lint extern inline void dmu_buf_init_user(dmu_buf_user_t *dbu, dmu_buf_evict_func_t *evict_func, dmu_buf_t **clear_on_evict_dbufp); #endif /* ! __lint */ /* * Global data structures and functions for the dbuf cache. */ static kmem_cache_t *dbuf_cache; static taskq_t *dbu_evict_taskq; /* ARGSUSED */ static int dbuf_cons(void *vdb, void *unused, int kmflag) { dmu_buf_impl_t *db = vdb; bzero(db, sizeof (dmu_buf_impl_t)); mutex_init(&db->db_mtx, NULL, MUTEX_DEFAULT, NULL); cv_init(&db->db_changed, NULL, CV_DEFAULT, NULL); refcount_create(&db->db_holds); return (0); } /* ARGSUSED */ static void dbuf_dest(void *vdb, void *unused) { dmu_buf_impl_t *db = vdb; mutex_destroy(&db->db_mtx); cv_destroy(&db->db_changed); refcount_destroy(&db->db_holds); } /* * dbuf hash table routines */ static dbuf_hash_table_t dbuf_hash_table; static uint64_t dbuf_hash_count; static uint64_t dbuf_hash(void *os, uint64_t obj, uint8_t lvl, uint64_t blkid) { uintptr_t osv = (uintptr_t)os; uint64_t crc = -1ULL; ASSERT(zfs_crc64_table[128] == ZFS_CRC64_POLY); crc = (crc >> 8) ^ zfs_crc64_table[(crc ^ (lvl)) & 0xFF]; crc = (crc >> 8) ^ zfs_crc64_table[(crc ^ (osv >> 6)) & 0xFF]; crc = (crc >> 8) ^ zfs_crc64_table[(crc ^ (obj >> 0)) & 0xFF]; crc = (crc >> 8) ^ zfs_crc64_table[(crc ^ (obj >> 8)) & 0xFF]; crc = (crc >> 8) ^ zfs_crc64_table[(crc ^ (blkid >> 0)) & 0xFF]; crc = (crc >> 8) ^ zfs_crc64_table[(crc ^ (blkid >> 8)) & 0xFF]; crc ^= (osv>>14) ^ (obj>>16) ^ (blkid>>16); return (crc); } #define DBUF_HASH(os, obj, level, blkid) dbuf_hash(os, obj, level, blkid); #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) { 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)); 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, hv, idx; dmu_buf_impl_t *dbf; blkid = db->db_blkid; hv = DBUF_HASH(os, obj, level, blkid); idx = hv & h->hash_table_mask; mutex_enter(DBUF_HASH_MUTEX(h, idx)); for (dbf = h->hash_table[idx]; dbf != NULL; dbf = dbf->db_hash_next) { 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); } } 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)); atomic_add_64(&dbuf_hash_count, 1); return (NULL); } /* * 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 hv, idx; dmu_buf_impl_t *dbf, **dbp; hv = DBUF_HASH(db->db_objset, db->db.db_object, db->db_level, db->db_blkid); idx = hv & h->hash_table_mask; /* * We musn't hold db_mtx to maintain lock ordering: * DBUF_HASH_MUTEX > db_mtx. */ ASSERT(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; mutex_exit(DBUF_HASH_MUTEX(h, idx)); atomic_add_64(&dbuf_hash_count, -1); } static arc_evict_func_t dbuf_do_evict; 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 = 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 /* * Invoke the callback from a taskq to avoid lock order reversals * and limit stack depth. */ taskq_dispatch_ent(dbu_evict_taskq, dbu->dbu_evict_func, 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); } } void dbuf_evict(dmu_buf_impl_t *db) { ASSERT(MUTEX_HELD(&db->db_mtx)); ASSERT(db->db_buf == NULL); ASSERT(db->db_data_pending == NULL); dbuf_clear(db); dbuf_destroy(db); } void dbuf_init(void) { uint64_t hsize = 1ULL << 16; dbuf_hash_table_t *h = &dbuf_hash_table; int i; /* * The hash table is big enough to fill all 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 < physmem * PAGESIZE) hsize <<= 1; retry: h->hash_table_mask = hsize - 1; #if defined(_KERNEL) && defined(HAVE_SPL) /* * Large allocations which do not require contiguous pages * should be using vmem_alloc() in the linux kernel */ h->hash_table = vmem_zalloc(hsize * sizeof (void *), KM_SLEEP); #else h->hash_table = kmem_zalloc(hsize * sizeof (void *), KM_NOSLEEP); #endif if (h->hash_table == NULL) { /* XXX - we should really return an error instead of assert */ ASSERT(hsize > (1ULL << 10)); hsize >>= 1; goto retry; } dbuf_cache = kmem_cache_create("dmu_buf_impl_t", sizeof (dmu_buf_impl_t), 0, dbuf_cons, dbuf_dest, NULL, NULL, NULL, 0); for (i = 0; i < DBUF_MUTEXES; 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); } void dbuf_fini(void) { dbuf_hash_table_t *h = &dbuf_hash_table; int i; dbuf_stats_destroy(); for (i = 0; i < DBUF_MUTEXES; i++) mutex_destroy(&h->hash_mutexes[i]); #if defined(_KERNEL) && defined(HAVE_SPL) /* * Large allocations which do not require contiguous pages * should be using vmem_free() in the linux kernel */ vmem_free(h->hash_table, (h->hash_table_mask + 1) * sizeof (void *)); #else kmem_free(h->hash_table, (h->hash_table_mask + 1) * sizeof (void *)); #endif kmem_cache_destroy(dbuf_cache); taskq_destroy(dbu_evict_taskq); } /* * Other stuff. */ #ifdef ZFS_DEBUG static void dbuf_verify(dmu_buf_impl_t *db) { dnode_t *dn; dbuf_dirty_record_t *dr; 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); ASSERT3U(db->db.db_size, >=, dn->dn_bonuslen); ASSERT0(db->db.db_offset); } else { ASSERT3U(db->db.db_offset, ==, db->db_blkid * db->db.db_size); } for (dr = db->db_data_pending; dr != NULL; dr = dr->dr_next) ASSERT(dr->dr_dbuf == db); for (dr = db->db_last_dirty; dr != NULL; dr = dr->dr_next) ASSERT(dr->dr_dbuf == db); /* * 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 */ ASSERTV(int epb = 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 struct_rwlock. XXX indblksz no longer * grows. safe to do this now? */ if (RW_WRITE_HELD(&dn->dn_struct_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->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. */ if (db->db_dirtycnt == 0) { ASSERTV(uint64_t *buf = db->db.db_data); int i; for (i = 0; i < db->db.db_size >> 3; i++) { ASSERT(buf[i] == 0); } } } 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); db->db_buf = NULL; db->db.db_data = NULL; if (db->db_state != DB_NOFILL) db->db_state = DB_UNCACHED; } 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; if (!arc_released(buf)) arc_set_callback(buf, dbuf_do_evict, db); } /* * 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; mutex_enter(&db->db_mtx); if (arc_released(db->db_buf) || 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, blksz); bcopy(db->db.db_data, abuf->b_data, blksz); } else { abuf = db->db_buf; arc_loan_inuse_buf(abuf, db); 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(dnode_t *dn, int64_t level, 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)) */ return (offset >> (dn->dn_datablkshift + level * (dn->dn_indblkshift - SPA_BLKPTRSHIFT))); } else { ASSERT3U(offset, <, dn->dn_datablksz); return (0); } } static void dbuf_read_done(zio_t *zio, arc_buf_t *buf, void *vdb) { 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(refcount_count(&db->db_holds) > 0); ASSERT(db->db_buf == NULL); ASSERT(db->db.db_data == NULL); if (db->db_level == 0 && db->db_freed_in_flight) { /* we were freed in flight; disregard any error */ arc_release(buf, db); bzero(buf->b_data, db->db.db_size); arc_buf_freeze(buf); db->db_freed_in_flight = FALSE; dbuf_set_data(db, buf); db->db_state = DB_CACHED; } else if (zio == NULL || zio->io_error == 0) { dbuf_set_data(db, buf); db->db_state = DB_CACHED; } else { ASSERT(db->db_blkid != DMU_BONUS_BLKID); ASSERT3P(db->db_buf, ==, NULL); VERIFY(arc_buf_remove_ref(buf, db)); db->db_state = DB_UNCACHED; } cv_broadcast(&db->db_changed); dbuf_rele_and_unlock(db, NULL); } static int -dbuf_read_impl(dmu_buf_impl_t *db, zio_t *zio, uint32_t *flags) +dbuf_read_impl(dmu_buf_impl_t *db, zio_t *zio, uint32_t flags) { dnode_t *dn; zbookmark_phys_t zb; uint32_t aflags = ARC_FLAG_NOWAIT; int err; DB_DNODE_ENTER(db); dn = DB_DNODE(db); ASSERT(!refcount_is_zero(&db->db_holds)); /* We need the struct_rwlock to prevent db_blkptr from changing. */ ASSERT(RW_LOCK_HELD(&dn->dn_struct_rwlock)); ASSERT(MUTEX_HELD(&db->db_mtx)); ASSERT(db->db_state == DB_UNCACHED); ASSERT(db->db_buf == NULL); if (db->db_blkid == DMU_BONUS_BLKID) { int bonuslen = MIN(dn->dn_bonuslen, dn->dn_phys->dn_bonuslen); ASSERT3U(bonuslen, <=, db->db.db_size); db->db.db_data = zio_buf_alloc(DN_MAX_BONUSLEN); arc_space_consume(DN_MAX_BONUSLEN, ARC_SPACE_OTHER); if (bonuslen < DN_MAX_BONUSLEN) bzero(db->db.db_data, DN_MAX_BONUSLEN); if (bonuslen) bcopy(DN_BONUS(dn->dn_phys), db->db.db_data, bonuslen); DB_DNODE_EXIT(db); db->db_state = DB_CACHED; mutex_exit(&db->db_mtx); return (0); } /* * 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 (db->db_blkptr == NULL || BP_IS_HOLE(db->db_blkptr) || (db->db_level == 0 && (dnode_block_freed(dn, db->db_blkid) || BP_IS_HOLE(db->db_blkptr)))) { arc_buf_contents_t type = DBUF_GET_BUFC_TYPE(db); DB_DNODE_EXIT(db); dbuf_set_data(db, arc_buf_alloc(db->db_objset->os_spa, db->db.db_size, db, type)); bzero(db->db.db_data, db->db.db_size); db->db_state = DB_CACHED; - *flags |= DB_RF_CACHED; mutex_exit(&db->db_mtx); return (0); } DB_DNODE_EXIT(db); db->db_state = DB_READ; mutex_exit(&db->db_mtx); if (DBUF_IS_L2CACHEABLE(db)) aflags |= ARC_FLAG_L2CACHE; if (DBUF_IS_L2COMPRESSIBLE(db)) aflags |= ARC_FLAG_L2COMPRESS; SET_BOOKMARK(&zb, db->db_objset->os_dsl_dataset ? db->db_objset->os_dsl_dataset->ds_object : DMU_META_OBJSET, db->db.db_object, db->db_level, db->db_blkid); dbuf_add_ref(db, NULL); err = arc_read(zio, db->db_objset->os_spa, db->db_blkptr, dbuf_read_done, db, ZIO_PRIORITY_SYNC_READ, - (*flags & DB_RF_CANFAIL) ? ZIO_FLAG_CANFAIL : ZIO_FLAG_MUSTSUCCEED, + (flags & DB_RF_CANFAIL) ? ZIO_FLAG_CANFAIL : ZIO_FLAG_MUSTSUCCEED, &aflags, &zb); - if (aflags & ARC_FLAG_CACHED) - *flags |= DB_RF_CACHED; return (SET_ERROR(err)); } int dbuf_read(dmu_buf_impl_t *db, zio_t *zio, uint32_t flags) { int err = 0; boolean_t havepzio = (zio != NULL); 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(!refcount_is_zero(&db->db_holds)); if (db->db_state == DB_NOFILL) return (SET_ERROR(EIO)); DB_DNODE_ENTER(db); dn = DB_DNODE(db); if ((flags & DB_RF_HAVESTRUCT) == 0) rw_enter(&dn->dn_struct_rwlock, RW_READER); prefetch = db->db_level == 0 && db->db_blkid != DMU_BONUS_BLKID && (flags & DB_RF_NOPREFETCH) == 0 && dn != NULL && DBUF_IS_CACHEABLE(db); mutex_enter(&db->db_mtx); if (db->db_state == DB_CACHED) { mutex_exit(&db->db_mtx); if (prefetch) - dmu_zfetch(&dn->dn_zfetch, db->db.db_offset, - db->db.db_size, TRUE); + dmu_zfetch(&dn->dn_zfetch, db->db_blkid, 1); if ((flags & DB_RF_HAVESTRUCT) == 0) rw_exit(&dn->dn_struct_rwlock); DB_DNODE_EXIT(db); } else if (db->db_state == DB_UNCACHED) { spa_t *spa = dn->dn_objset->os_spa; if (zio == NULL) zio = zio_root(spa, NULL, NULL, ZIO_FLAG_CANFAIL); - err = dbuf_read_impl(db, zio, &flags); + err = dbuf_read_impl(db, zio, flags); /* dbuf_read_impl has dropped db_mtx for us */ if (!err && prefetch) - dmu_zfetch(&dn->dn_zfetch, db->db.db_offset, - db->db.db_size, flags & DB_RF_CACHED); + dmu_zfetch(&dn->dn_zfetch, db->db_blkid, 1); if ((flags & DB_RF_HAVESTRUCT) == 0) rw_exit(&dn->dn_struct_rwlock); DB_DNODE_EXIT(db); if (!err && !havepzio) err = 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.db_offset, - db->db.db_size, TRUE); + dmu_zfetch(&dn->dn_zfetch, db->db_blkid, 1); if ((flags & DB_RF_HAVESTRUCT) == 0) rw_exit(&dn->dn_struct_rwlock); DB_DNODE_EXIT(db); /* Skip the wait per the caller's request. */ mutex_enter(&db->db_mtx); if ((flags & DB_RF_NEVERWAIT) == 0) { 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); } ASSERT(err || havepzio || db->db_state == DB_CACHED); return (err); } static void dbuf_noread(dmu_buf_impl_t *db) { ASSERT(!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) { arc_buf_contents_t type = DBUF_GET_BUFC_TYPE(db); spa_t *spa = db->db_objset->os_spa; ASSERT(db->db_buf == NULL); ASSERT(db->db.db_data == NULL); dbuf_set_data(db, arc_buf_alloc(spa, db->db.db_size, db, type)); db->db_state = DB_FILL; } else if (db->db_state == DB_NOFILL) { dbuf_clear_data(db); } else { ASSERT3U(db->db_state, ==, DB_CACHED); } mutex_exit(&db->db_mtx); } /* * 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 modify them in the current active group. * * This function is used in two places: when we are dirtying a * buffer for the first time in a txg, and when we are freeing * a range in a dnode that includes this buffer. * * 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 = db->db_last_dirty; 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. */ ASSERT(dr->dr_txg >= txg - 2); if (db->db_blkid == DMU_BONUS_BLKID) { /* Note that the data bufs here are zio_bufs */ dr->dt.dl.dr_data = zio_buf_alloc(DN_MAX_BONUSLEN); arc_space_consume(DN_MAX_BONUSLEN, ARC_SPACE_OTHER); bcopy(db->db.db_data, dr->dt.dl.dr_data, DN_MAX_BONUSLEN); } else if (refcount_count(&db->db_holds) > db->db_dirtycnt) { int size = db->db.db_size; arc_buf_contents_t type = DBUF_GET_BUFC_TYPE(db); spa_t *spa = db->db_objset->os_spa; dr->dt.dl.dr_data = arc_buf_alloc(spa, size, db, type); bcopy(db->db.db_data, dr->dt.dl.dr_data->b_data, size); } else { dbuf_clear_data(db); } } 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; ASSERT(MUTEX_HELD(&db->db_mtx)); 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); dr->dt.dl.dr_override_state = DR_NOT_OVERRIDDEN; dr->dt.dl.dr_nopwrite = 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. */ 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. * * This is a no-op if the dataset is in the middle of an incremental * receive; see comment below for details. */ 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; boolean_t freespill = (start_blkid == DMU_SPILL_BLKID || end_blkid == DMU_SPILL_BLKID); if (end_blkid > dn->dn_maxblkid && !freespill) end_blkid = dn->dn_maxblkid; dprintf_dnode(dn, "start=%llu end=%llu\n", start_blkid, 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); if (start_blkid >= dn->dn_unlisted_l0_blkid && !freespill) { /* There can't be any dbufs in this range; no need to search. */ #ifdef DEBUG db = avl_find(&dn->dn_dbufs, db_search, &where); ASSERT3P(db, ==, NULL); db = avl_nearest(&dn->dn_dbufs, where, AVL_AFTER); ASSERT(db == NULL || db->db_level > 0); #endif goto out; } else if (dmu_objset_is_receiving(dn->dn_objset)) { /* * If we are receiving, we expect there to be no dbufs in * the range to be freed, because receive modifies each * block at most once, and in offset order. If this is * not the case, it can lead to performance problems, * so note that we unexpectedly took the slow path. */ atomic_inc_64(&zfs_free_range_recv_miss); } 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 (refcount_count(&db->db_holds) == 0) { ASSERT(db->db_buf); dbuf_clear(db); continue; } /* The dbuf is referenced */ if (db->db_last_dirty != NULL) { dbuf_dirty_record_t *dr = db->db_last_dirty; 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); bzero(db->db.db_data, db->db.db_size); arc_buf_freeze(db->db_buf); } mutex_exit(&db->db_mtx); } out: kmem_free(db_search, sizeof (dmu_buf_impl_t)); mutex_exit(&dn->dn_dbufs_mtx); } static int dbuf_block_freeable(dmu_buf_impl_t *db) { dsl_dataset_t *ds = db->db_objset->os_dsl_dataset; uint64_t birth_txg = 0; /* * We don't need any locking to protect db_blkptr: * If it's syncing, then db_last_dirty will be set * so we'll ignore db_blkptr. * * This logic ensures that only block births for * filled blocks are considered. */ ASSERT(MUTEX_HELD(&db->db_mtx)); if (db->db_last_dirty && (db->db_blkptr == NULL || !BP_IS_HOLE(db->db_blkptr))) { birth_txg = db->db_last_dirty->dr_txg; } else if (db->db_blkptr != NULL && !BP_IS_HOLE(db->db_blkptr)) { birth_txg = db->db_blkptr->blk_birth; } /* * If this block don't exist or is in a snapshot, it can't be freed. * Don't pass the bp to dsl_dataset_block_freeable() since we * are holding the db_mtx lock and might deadlock if we are * prefetching a dedup-ed block. */ if (birth_txg != 0) return (ds == NULL || dsl_dataset_block_freeable(ds, NULL, birth_txg)); else return (B_FALSE); } void dbuf_new_size(dmu_buf_impl_t *db, int size, dmu_tx_t *tx) { arc_buf_t *buf, *obuf; 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 does *this* func really need the lock? */ ASSERT(RW_WRITE_HELD(&dn->dn_struct_rwlock)); /* * This call to dmu_buf_will_dirty() with the dn_struct_rwlock held * is OK, because there can be no other references to the db * when we are changing its size, so no concurrent DB_FILL can * be happening. */ /* * 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_buf_alloc(dn->dn_objset->os_spa, size, db, type); /* copy old block data to the new block */ obuf = db->db_buf; bcopy(obuf->b_data, buf->b_data, MIN(osize, size)); /* zero the remainder */ if (size > osize) bzero((uint8_t *)buf->b_data + osize, size - osize); mutex_enter(&db->db_mtx); dbuf_set_data(db, buf); VERIFY(arc_buf_remove_ref(obuf, db)); db->db.db_size = size; if (db->db_level == 0) { ASSERT3U(db->db_last_dirty->dr_txg, ==, tx->tx_txg); db->db_last_dirty->dt.dl.dr_data = buf; } mutex_exit(&db->db_mtx); dnode_willuse_space(dn, size-osize, tx); DB_DNODE_EXIT(db); } void dbuf_release_bp(dmu_buf_impl_t *db) { ASSERTV(objset_t *os = 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); } 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 **drp, *dr; int drop_struct_lock = FALSE; boolean_t do_free_accounting = B_FALSE; int txgoff = tx->tx_txg & TXG_MASK; ASSERT(tx->tx_txg != 0); ASSERT(!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. */ 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); /* * 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); /* * Don't set dirtyctx to SYNC if we're just modifying this as we * initialize the objset. */ if (dn->dn_dirtyctx == DN_UNDIRTIED && !BP_IS_HOLE(dn->dn_objset->os_rootbp)) { dn->dn_dirtyctx = (dmu_tx_is_syncing(tx) ? DN_DIRTY_SYNC : DN_DIRTY_OPEN); ASSERT(dn->dn_dirtyctx_firstset == NULL); dn->dn_dirtyctx_firstset = kmem_alloc(1, KM_SLEEP); } 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. */ drp = &db->db_last_dirty; ASSERT(*drp == NULL || (*drp)->dr_txg <= tx->tx_txg || db->db.db_object == DMU_META_DNODE_OBJECT); while ((dr = *drp) != NULL && dr->dr_txg > tx->tx_txg) drp = &dr->dr_next; if (dr && dr->dr_txg == tx->tx_txg) { DB_DNODE_EXIT(db); 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) arc_buf_thaw(db->db_buf); } mutex_exit(&db->db_mtx); return (dr); } /* * 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); 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); /* * 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; ASSERT(!dmu_tx_is_syncing(tx) || DMU_OBJECT_IS_SPECIAL(dn->dn_object) || os->os_dsl_dataset == NULL || BP_IS_HOLE(os->os_rootbp)); 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) { /* * Update the accounting. * Note: we delay "free accounting" until after we drop * the db_mtx. This keeps us from grabbing other locks * (and possibly deadlocking) in bp_get_dsize() while * also holding the db_mtx. */ dnode_willuse_space(dn, db->db.db_size, tx); do_free_accounting = dbuf_block_freeable(db); } /* * 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); 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 && os->os_dsl_dataset != NULL) dr->dr_accounted = db->db.db_size; dr->dr_dbuf = db; dr->dr_txg = tx->tx_txg; dr->dr_next = *drp; *drp = 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); } else if (do_free_accounting) { blkptr_t *bp = db->db_blkptr; int64_t willfree = (bp && !BP_IS_HOLE(bp)) ? bp_get_dsize(os->os_spa, bp) : db->db.db_size; /* * This is only a guess -- if the dbuf is dirty * in a previous txg, we don't know how much * space it will use on disk yet. We should * really have the struct_rwlock to access * db_blkptr, but since this is just a guess, * it's OK if we get an odd answer. */ ddt_prefetch(os->os_spa, bp); dnode_willuse_space(dn, -willfree, tx); } if (!RW_WRITE_HELD(&dn->dn_struct_rwlock)) { rw_enter(&dn->dn_struct_rwlock, RW_READER); drop_struct_lock = TRUE; } if (db->db_level == 0) { dnode_new_blkid(dn, db->db_blkid, tx, drop_struct_lock); 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_lock) 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 (db->db_last_dirty == 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_lock) rw_exit(&dn->dn_struct_rwlock); } dnode_setdirty(dn, tx); DB_DNODE_EXIT(db); return (dr); } /* * Undirty a buffer in the transaction group referenced by the given * transaction. Return whether this evicted the dbuf. */ static boolean_t dbuf_undirty(dmu_buf_impl_t *db, dmu_tx_t *tx) { dnode_t *dn; uint64_t txg = tx->tx_txg; dbuf_dirty_record_t *dr, **drp; 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. */ for (drp = &db->db_last_dirty; (dr = *drp) != NULL; drp = &dr->dr_next) if (dr->dr_txg <= txg) break; if (dr == NULL || dr->dr_txg < txg) return (B_FALSE); ASSERT(dr->dr_txg == txg); ASSERT(dr->dr_dbuf == db); DB_DNODE_ENTER(db); dn = DB_DNODE(db); 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); *drp = dr->dr_next; /* * 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); } DB_DNODE_EXIT(db); if (db->db_state != DB_NOFILL) { dbuf_unoverride(dr); ASSERT(db->db_buf != NULL); ASSERT(dr->dt.dl.dr_data != NULL); if (dr->dt.dl.dr_data != db->db_buf) VERIFY(arc_buf_remove_ref(dr->dt.dl.dr_data, db)); } kmem_free(dr, sizeof (dbuf_dirty_record_t)); ASSERT(db->db_dirtycnt > 0); db->db_dirtycnt -= 1; if (refcount_remove(&db->db_holds, (void *)(uintptr_t)txg) == 0) { arc_buf_t *buf = db->db_buf; ASSERT(db->db_state == DB_NOFILL || arc_released(buf)); dbuf_clear_data(db); VERIFY(arc_buf_remove_ref(buf, db)); dbuf_evict(db); return (B_TRUE); } return (B_FALSE); } void dmu_buf_will_dirty(dmu_buf_t *db_fake, dmu_tx_t *tx) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake; int rf = DB_RF_MUST_SUCCEED | DB_RF_NOPREFETCH; ASSERT(tx->tx_txg != 0); ASSERT(!refcount_is_zero(&db->db_holds)); DB_DNODE_ENTER(db); if (RW_WRITE_HELD(&DB_DNODE(db)->dn_struct_rwlock)) rf |= DB_RF_HAVESTRUCT; DB_DNODE_EXIT(db); (void) dbuf_read(db, NULL, rf); (void) dbuf_dirty(db, 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; db->db_state = DB_NOFILL; dmu_buf_will_fill(db_fake, 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(!refcount_is_zero(&db->db_holds)); ASSERT(db->db.db_object != DMU_META_DNODE_OBJECT || dmu_tx_private_ok(tx)); dbuf_noread(db); (void) dbuf_dirty(db, tx); } #pragma weak dmu_buf_fill_done = dbuf_fill_done /* ARGSUSED */ void dbuf_fill_done(dmu_buf_impl_t *db, dmu_tx_t *tx) { mutex_enter(&db->db_mtx); DBUF_VERIFY(db); if (db->db_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? */ bzero(db->db.db_data, db->db.db_size); db->db_freed_in_flight = FALSE; } db->db_state = DB_CACHED; 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; 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); ASSERT3U(db->db_last_dirty->dr_txg, ==, tx->tx_txg); dl = &db->db_last_dirty->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 = db->db_last_dirty->dr_txg; } /* * 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(!refcount_is_zero(&db->db_holds)); ASSERT(db->db_blkid != DMU_BONUS_BLKID); ASSERT(db->db_level == 0); ASSERT(DBUF_GET_BUFC_TYPE(db) == ARC_BUFC_DATA); ASSERT(buf != NULL); ASSERT(arc_buf_size(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 && refcount_count(&db->db_holds) - 1 > db->db_dirtycnt) { mutex_exit(&db->db_mtx); (void) dbuf_dirty(db, tx); bcopy(buf->b_data, db->db.db_data, db->db.db_size); VERIFY(arc_buf_remove_ref(buf, db)); xuio_stat_wbuf_copied(); return; } xuio_stat_wbuf_nocopy(); if (db->db_state == DB_CACHED) { dbuf_dirty_record_t *dr = db->db_last_dirty; 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; VERIFY(arc_buf_remove_ref(db->db_buf, db)); } else if (dr == NULL || dr->dt.dl.dr_data != db->db_buf) { arc_release(db->db_buf, db); VERIFY(arc_buf_remove_ref(db->db_buf, db)); } db->db_buf = NULL; } ASSERT(db->db_buf == NULL); dbuf_set_data(db, buf); db->db_state = DB_FILL; mutex_exit(&db->db_mtx); (void) dbuf_dirty(db, tx); dmu_buf_fill_done(&db->db, tx); } /* * "Clear" the contents of this dbuf. This will mark the dbuf * EVICTING and clear *most* of its references. Unfortunately, * when we are not holding the dn_dbufs_mtx, we can't clear the * entry in the dn_dbufs list. We have to wait until dbuf_destroy() * in this case. For callers from the DMU we will usually see: * dbuf_clear()->arc_clear_callback()->dbuf_do_evict()->dbuf_destroy() * For the arc callback, we will usually see: * dbuf_do_evict()->dbuf_clear();dbuf_destroy() * Sometimes, though, we will get a mix of these two: * DMU: dbuf_clear()->arc_clear_callback() * ARC: dbuf_do_evict()->dbuf_destroy() * * This routine will dissociate the dbuf from the arc, by calling * arc_clear_callback(), but will not evict the data from the ARC. */ void dbuf_clear(dmu_buf_impl_t *db) { dnode_t *dn; dmu_buf_impl_t *parent = db->db_parent; dmu_buf_impl_t *dndb; boolean_t dbuf_gone = B_FALSE; ASSERT(MUTEX_HELD(&db->db_mtx)); ASSERT(refcount_is_zero(&db->db_holds)); dbuf_evict_user(db); if (db->db_state == DB_CACHED) { ASSERT(db->db.db_data != NULL); if (db->db_blkid == DMU_BONUS_BLKID) { zio_buf_free(db->db.db_data, DN_MAX_BONUSLEN); arc_space_return(DN_MAX_BONUSLEN, ARC_SPACE_OTHER); } db->db.db_data = NULL; db->db_state = DB_UNCACHED; } ASSERT(db->db_state == DB_UNCACHED || db->db_state == DB_NOFILL); ASSERT(db->db_data_pending == NULL); db->db_state = DB_EVICTING; db->db_blkptr = NULL; DB_DNODE_ENTER(db); dn = DB_DNODE(db); dndb = dn->dn_dbuf; if (db->db_blkid != DMU_BONUS_BLKID && MUTEX_HELD(&dn->dn_dbufs_mtx)) { avl_remove(&dn->dn_dbufs, db); atomic_dec_32(&dn->dn_dbufs_count); membar_producer(); DB_DNODE_EXIT(db); /* * 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. */ dnode_rele(dn, db); db->db_dnode_handle = NULL; } else { DB_DNODE_EXIT(db); } if (db->db_buf) dbuf_gone = arc_clear_callback(db->db_buf); if (!dbuf_gone) mutex_exit(&db->db_mtx); /* * If this dbuf is referenced from an indirect dbuf, * decrement the ref count on the indirect dbuf. */ if (parent && parent != dndb) dbuf_rele(parent, db); } /* * 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 a userused or groupused * 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, struct dbuf_hold_impl_data *dh) { int nlevels, epbs; *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->dn_phys->dn_spill; else *bpp = NULL; dbuf_add_ref(dn->dn_dbuf, NULL); *parentp = dn->dn_dbuf; mutex_exit(&dn->dn_mtx); return (0); } if (dn->dn_phys->dn_nlevels == 0) nlevels = 1; else nlevels = dn->dn_phys->dn_nlevels; epbs = dn->dn_indblkshift - SPA_BLKPTRSHIFT; ASSERT3U(level * epbs, <, 64); ASSERT(RW_LOCK_HELD(&dn->dn_struct_rwlock)); if (level >= nlevels || (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; if (dh == NULL) { err = dbuf_hold_impl(dn, level+1, blkid >> epbs, fail_sparse, FALSE, NULL, parentp); } else { __dbuf_hold_impl_init(dh + 1, dn, dh->dh_level + 1, blkid >> epbs, fail_sparse, FALSE, NULL, parentp, dh->dh_depth + 1); err = __dbuf_hold_impl(dh + 1); } 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); } *bpp = ((blkptr_t *)(*parentp)->db.db_data) + (blkid & ((1ULL << epbs) - 1)); 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) { 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_cache, KM_SLEEP); db->db_objset = os; db->db.db_object = dn->dn_object; db->db_level = level; db->db_blkid = blkid; db->db_last_dirty = NULL; db->db_dirtycnt = 0; db->db_dnode_handle = dn->dn_handle; db->db_parent = parent; db->db_blkptr = blkptr; 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_MAX_BONUSLEN - (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; /* the bonus dbuf is not placed in the hash table */ arc_space_consume(sizeof (dmu_buf_impl_t), ARC_SPACE_OTHER); 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 its added to the * dn_dbufs list. */ mutex_enter(&dn->dn_dbufs_mtx); db->db_state = DB_EVICTING; if ((odb = dbuf_hash_insert(db)) != NULL) { /* someone else inserted it first */ kmem_cache_free(dbuf_cache, db); mutex_exit(&dn->dn_dbufs_mtx); return (odb); } avl_add(&dn->dn_dbufs, db); if (db->db_level == 0 && db->db_blkid >= dn->dn_unlisted_l0_blkid) dn->dn_unlisted_l0_blkid = db->db_blkid + 1; db->db_state = DB_UNCACHED; mutex_exit(&dn->dn_dbufs_mtx); arc_space_consume(sizeof (dmu_buf_impl_t), ARC_SPACE_OTHER); if (parent && parent != dn->dn_dbuf) dbuf_add_ref(parent, db); ASSERT(dn->dn_object == DMU_META_DNODE_OBJECT || refcount_count(&dn->dn_holds) > 0); (void) refcount_add(&dn->dn_holds, db); atomic_inc_32(&dn->dn_dbufs_count); dprintf_dbuf(db, "db=%p\n", db); return (db); } static int dbuf_do_evict(void *private) { dmu_buf_impl_t *db = private; if (!MUTEX_HELD(&db->db_mtx)) mutex_enter(&db->db_mtx); ASSERT(refcount_is_zero(&db->db_holds)); if (db->db_state != DB_EVICTING) { ASSERT(db->db_state == DB_CACHED); DBUF_VERIFY(db); db->db_buf = NULL; dbuf_evict(db); } else { mutex_exit(&db->db_mtx); dbuf_destroy(db); } return (0); } static void dbuf_destroy(dmu_buf_impl_t *db) { ASSERT(refcount_is_zero(&db->db_holds)); if (db->db_blkid != DMU_BONUS_BLKID) { /* * If this dbuf is still on the dn_dbufs list, * remove it from that list. */ if (db->db_dnode_handle != NULL) { dnode_t *dn; DB_DNODE_ENTER(db); dn = DB_DNODE(db); mutex_enter(&dn->dn_dbufs_mtx); avl_remove(&dn->dn_dbufs, db); atomic_dec_32(&dn->dn_dbufs_count); mutex_exit(&dn->dn_dbufs_mtx); DB_DNODE_EXIT(db); /* * 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. */ dnode_rele(dn, db); db->db_dnode_handle = NULL; } dbuf_hash_remove(db); } db->db_parent = NULL; 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); kmem_cache_free(dbuf_cache, db); arc_space_return(sizeof (dmu_buf_impl_t), ARC_SPACE_OTHER); } 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 */ 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_arg_t; /* * Actually issue the prefetch read for the block given. */ static void dbuf_issue_final_prefetch(dbuf_prefetch_arg_t *dpa, blkptr_t *bp) { arc_flags_t aflags; if (BP_IS_HOLE(bp) || BP_IS_EMBEDDED(bp)) return; aflags = dpa->dpa_aflags | ARC_FLAG_NOWAIT | ARC_FLAG_PREFETCH; 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, NULL, NULL, dpa->dpa_prio, ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE, &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, arc_buf_t *abuf, void *private) { dbuf_prefetch_arg_t *dpa = private; uint64_t nextblkid; blkptr_t *bp; ASSERT3S(dpa->dpa_zb.zb_level, <, dpa->dpa_curlevel); ASSERT3S(dpa->dpa_curlevel, >, 0); if (zio != NULL) { ASSERT3S(BP_GET_LEVEL(zio->io_bp), ==, dpa->dpa_curlevel); ASSERT3U(BP_GET_LSIZE(zio->io_bp), ==, zio->io_size); ASSERT3P(zio->io_spa, ==, dpa->dpa_spa); } dpa->dpa_curlevel--; nextblkid = dpa->dpa_zb.zb_blkid >> (dpa->dpa_epbs * (dpa->dpa_curlevel - dpa->dpa_zb.zb_level)); bp = ((blkptr_t *)abuf->b_data) + P2PHASE(nextblkid, 1ULL << dpa->dpa_epbs); if (BP_IS_HOLE(bp) || (zio != NULL && zio->io_error != 0)) { kmem_free(dpa, sizeof (*dpa)); } else if (dpa->dpa_curlevel == dpa->dpa_zb.zb_level) { ASSERT3U(nextblkid, ==, dpa->dpa_zb.zb_blkid); dbuf_issue_final_prefetch(dpa, bp); kmem_free(dpa, sizeof (*dpa)); } else { arc_flags_t iter_aflags = ARC_FLAG_NOWAIT; zbookmark_phys_t zb; 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, dpa->dpa_prio, ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE, &iter_aflags, &zb); } (void) arc_buf_remove_ref(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. */ void dbuf_prefetch(dnode_t *dn, int64_t level, uint64_t blkid, zio_priority_t prio, arc_flags_t aflags) { blkptr_t bp; int epbs, nlevels, curlevel; uint64_t curblkid; dmu_buf_impl_t *db; zio_t *pio; dbuf_prefetch_arg_t *dpa; dsl_dataset_t *ds; ASSERT(blkid != DMU_BONUS_BLKID); ASSERT(RW_LOCK_HELD(&dn->dn_struct_rwlock)); + if (blkid > dn->dn_maxblkid) + return; + if (dnode_block_freed(dn, blkid)) return; /* * 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) return; epbs = dn->dn_phys->dn_indblkshift - SPA_BLKPTRSHIFT; if (dn->dn_phys->dn_maxblkid < blkid << (epbs * level)) return; db = dbuf_find(dn->dn_objset, dn->dn_object, level, blkid); 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. */ return; } /* * 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]; } if (BP_IS_HOLE(&bp)) return; ASSERT3U(curlevel, ==, BP_GET_LEVEL(&bp)); pio = zio_root(dmu_objset_spa(dn->dn_objset), NULL, NULL, ZIO_FLAG_CANFAIL); dpa = kmem_zalloc(sizeof (*dpa), KM_SLEEP); 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_epbs = epbs; dpa->dpa_zio = pio; /* * 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); kmem_free(dpa, sizeof (*dpa)); } else { arc_flags_t iter_aflags = ARC_FLAG_NOWAIT; zbookmark_phys_t zb; 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, prio, 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); } #define DBUF_HOLD_IMPL_MAX_DEPTH 20 /* * Returns with db_holds incremented, and db_mtx not held. * Note: dn_struct_rwlock must be held. */ static int __dbuf_hold_impl(struct dbuf_hold_impl_data *dh) { ASSERT3S(dh->dh_depth, <, DBUF_HOLD_IMPL_MAX_DEPTH); dh->dh_parent = NULL; ASSERT(dh->dh_blkid != DMU_BONUS_BLKID); ASSERT(RW_LOCK_HELD(&dh->dh_dn->dn_struct_rwlock)); ASSERT3U(dh->dh_dn->dn_nlevels, >, dh->dh_level); *(dh->dh_dbp) = NULL; top: /* dbuf_find() returns with db_mtx held */ dh->dh_db = dbuf_find(dh->dh_dn->dn_objset, dh->dh_dn->dn_object, dh->dh_level, dh->dh_blkid); if (dh->dh_db == NULL) { dh->dh_bp = NULL; if (dh->dh_fail_uncached) return (SET_ERROR(ENOENT)); ASSERT3P(dh->dh_parent, ==, NULL); dh->dh_err = dbuf_findbp(dh->dh_dn, dh->dh_level, dh->dh_blkid, dh->dh_fail_sparse, &dh->dh_parent, &dh->dh_bp, dh); if (dh->dh_fail_sparse) { if (dh->dh_err == 0 && dh->dh_bp && BP_IS_HOLE(dh->dh_bp)) dh->dh_err = SET_ERROR(ENOENT); if (dh->dh_err) { if (dh->dh_parent) dbuf_rele(dh->dh_parent, NULL); return (dh->dh_err); } } if (dh->dh_err && dh->dh_err != ENOENT) return (dh->dh_err); dh->dh_db = dbuf_create(dh->dh_dn, dh->dh_level, dh->dh_blkid, dh->dh_parent, dh->dh_bp); } if (dh->dh_fail_uncached && dh->dh_db->db_state != DB_CACHED) { mutex_exit(&dh->dh_db->db_mtx); return (SET_ERROR(ENOENT)); } if (dh->dh_db->db_buf && refcount_is_zero(&dh->dh_db->db_holds)) { arc_buf_add_ref(dh->dh_db->db_buf, dh->dh_db); if (dh->dh_db->db_buf->b_data == NULL) { dbuf_clear(dh->dh_db); if (dh->dh_parent) { dbuf_rele(dh->dh_parent, NULL); dh->dh_parent = NULL; } goto top; } ASSERT3P(dh->dh_db->db.db_data, ==, dh->dh_db->db_buf->b_data); } ASSERT(dh->dh_db->db_buf == NULL || arc_referenced(dh->dh_db->db_buf)); /* * If this buffer is currently syncing out, and we are 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 (dh->dh_db->db_level == 0 && dh->dh_db->db_blkid != DMU_BONUS_BLKID && dh->dh_dn->dn_object != DMU_META_DNODE_OBJECT && dh->dh_db->db_state == DB_CACHED && dh->dh_db->db_data_pending) { dh->dh_dr = dh->dh_db->db_data_pending; if (dh->dh_dr->dt.dl.dr_data == dh->dh_db->db_buf) { dh->dh_type = DBUF_GET_BUFC_TYPE(dh->dh_db); dbuf_set_data(dh->dh_db, arc_buf_alloc(dh->dh_dn->dn_objset->os_spa, dh->dh_db->db.db_size, dh->dh_db, dh->dh_type)); bcopy(dh->dh_dr->dt.dl.dr_data->b_data, dh->dh_db->db.db_data, dh->dh_db->db.db_size); } } (void) refcount_add(&dh->dh_db->db_holds, dh->dh_tag); DBUF_VERIFY(dh->dh_db); mutex_exit(&dh->dh_db->db_mtx); /* NOTE: we can't rele the parent until after we drop the db_mtx */ if (dh->dh_parent) dbuf_rele(dh->dh_parent, NULL); ASSERT3P(DB_DNODE(dh->dh_db), ==, dh->dh_dn); ASSERT3U(dh->dh_db->db_blkid, ==, dh->dh_blkid); ASSERT3U(dh->dh_db->db_level, ==, dh->dh_level); *(dh->dh_dbp) = dh->dh_db; return (0); } /* * The following code preserves the recursive function dbuf_hold_impl() * but moves the local variables AND function arguments to the heap to * minimize the stack frame size. Enough space is initially allocated * on the stack for 20 levels of recursion. */ int dbuf_hold_impl(dnode_t *dn, uint8_t level, uint64_t blkid, boolean_t fail_sparse, boolean_t fail_uncached, void *tag, dmu_buf_impl_t **dbp) { struct dbuf_hold_impl_data *dh; int error; dh = kmem_zalloc(sizeof (struct dbuf_hold_impl_data) * DBUF_HOLD_IMPL_MAX_DEPTH, KM_SLEEP); __dbuf_hold_impl_init(dh, dn, level, blkid, fail_sparse, fail_uncached, tag, dbp, 0); error = __dbuf_hold_impl(dh); kmem_free(dh, sizeof (struct dbuf_hold_impl_data) * DBUF_HOLD_IMPL_MAX_DEPTH); return (error); } static void __dbuf_hold_impl_init(struct dbuf_hold_impl_data *dh, dnode_t *dn, uint8_t level, uint64_t blkid, boolean_t fail_sparse, boolean_t fail_uncached, void *tag, dmu_buf_impl_t **dbp, int depth) { dh->dh_dn = dn; dh->dh_level = level; dh->dh_blkid = blkid; dh->dh_fail_sparse = fail_sparse; dh->dh_fail_uncached = fail_uncached; dh->dh_tag = tag; dh->dh_dbp = dbp; dh->dh_depth = depth; } dmu_buf_impl_t * dbuf_hold(dnode_t *dn, uint64_t blkid, 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, 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); } 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; dnode_t *dn; 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); DB_DNODE_ENTER(db); dn = DB_DNODE(db); rw_enter(&dn->dn_struct_rwlock, RW_WRITER); dbuf_new_size(db, blksz, tx); rw_exit(&dn->dn_struct_rwlock); DB_DNODE_EXIT(db); 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, void *tag) { VERIFY(refcount_add(&db->db_holds, tag) > 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, 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); if (found_db != NULL) { if (db == found_db && dbuf_refcount(db) > db->db_dirtycnt) { (void) 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, void *tag) { mutex_enter(&db->db_mtx); dbuf_rele_and_unlock(db, tag); } void dmu_buf_rele(dmu_buf_t *db, 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. */ void dbuf_rele_and_unlock(dmu_buf_impl_t *db, void *tag) { int64_t holds; 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 = 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 && 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_evict(db); } else if (arc_released(db->db_buf)) { arc_buf_t *buf = db->db_buf; /* * This dbuf has anonymous data associated with it. */ dbuf_clear_data(db); VERIFY(arc_buf_remove_ref(buf, db)); dbuf_evict(db); } else { VERIFY(!arc_buf_remove_ref(db->db_buf, db)); /* * A dbuf will be eligible for eviction if either the * 'primarycache' property is set or a duplicate * copy of this buffer is already cached in the arc. * * In the case of the 'primarycache' a buffer * is considered for eviction if it matches the * criteria set in the property. * * To decide if our buffer is considered a * duplicate, we must call into the arc to determine * if multiple buffers are referencing the same * block on-disk. If so, then we simply evict * ourselves. */ if (!DBUF_IS_CACHEABLE(db)) { if (db->db_blkptr != NULL && !BP_IS_HOLE(db->db_blkptr) && !BP_IS_EMBEDDED(db->db_blkptr)) { spa_t *spa = dmu_objset_spa(db->db_objset); blkptr_t bp = *db->db_blkptr; dbuf_clear(db); arc_freed(spa, &bp); } else { dbuf_clear(db); } } else if (db->db_pending_evict || arc_buf_eviction_needed(db->db_buf)) { dbuf_clear(db); } else { mutex_exit(&db->db_mtx); } } } else { mutex_exit(&db->db_mtx); } } #pragma weak dmu_buf_refcount = dbuf_refcount uint64_t dbuf_refcount(dmu_buf_impl_t *db) { return (refcount_count(&db->db_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() { taskq_wait(dbu_evict_taskq); } boolean_t dmu_buf_freeable(dmu_buf_t *dbuf) { boolean_t res = B_FALSE; dmu_buf_impl_t *db = (dmu_buf_impl_t *)dbuf; if (db->db_blkptr) res = dsl_dataset_block_freeable(db->db_objset->os_dsl_dataset, db->db_blkptr, db->db_blkptr->blk_birth); return (res); } blkptr_t * dmu_buf_get_blkptr(dmu_buf_t *db) { dmu_buf_impl_t *dbi = (dmu_buf_impl_t *)db; return (dbi->db_blkptr); } 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->dn_phys->dn_spill; 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 mis-match). */ 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); } } /* * 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; zio_t *zio; 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); DB_DNODE_ENTER(db); dn = DB_DNODE(db); /* 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); DB_DNODE_EXIT(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 = 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); } /* * 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; 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); DB_DNODE_ENTER(db); dn = DB_DNODE(db); if (db->db_blkid == DMU_SPILL_BLKID) { mutex_enter(&dn->dn_mtx); 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) { dbuf_dirty_record_t **drp; ASSERT(*datap != NULL); ASSERT0(db->db_level); ASSERT3U(dn->dn_phys->dn_bonuslen, <=, DN_MAX_BONUSLEN); bcopy(*datap, DN_BONUS(dn->dn_phys), dn->dn_phys->dn_bonuslen); DB_DNODE_EXIT(db); if (*datap != db->db.db_data) { zio_buf_free(*datap, DN_MAX_BONUSLEN); arc_space_return(DN_MAX_BONUSLEN, ARC_SPACE_OTHER); } db->db_data_pending = NULL; drp = &db->db_last_dirty; while (*drp != dr) drp = &(*drp)->dr_next; ASSERT(dr->dr_next == NULL); ASSERT(dr->dr_dbuf == db); *drp = dr->dr_next; 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)); ASSERT(db->db_dirtycnt > 0); db->db_dirtycnt -= 1; dbuf_rele_and_unlock(db, (void *)(uintptr_t)txg); 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); ASSERT(dr->dt.dl.dr_override_state != DR_NOT_OVERRIDDEN); } if (db->db_state != DB_NOFILL && dn->dn_object != DMU_META_DNODE_OBJECT && 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 blksz = arc_buf_size(*datap); arc_buf_contents_t type = DBUF_GET_BUFC_TYPE(db); *datap = arc_buf_alloc(os->os_spa, blksz, db, type); bcopy(db->db.db_data, (*datap)->b_data, blksz); } 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); DB_DNODE_EXIT(db); } else { /* * Although zio_nowait() does not "wait for an IO", it does * initiate the IO. If this is an empty write it seems plausible * that the IO could actually be completed before the nowait * returns. We need to DB_DNODE_EXIT() first in case * zio_nowait() invalidates the dbuf. */ DB_DNODE_EXIT(db); 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; } if (dr->dr_dbuf->db_blkid != DMU_BONUS_BLKID && dr->dr_dbuf->db_blkid != DMU_SPILL_BLKID) { VERIFY3U(dr->dr_dbuf->db_level, ==, level); } list_remove(list, dr); if (dr->dr_dbuf->db_level > 0) dbuf_sync_indirect(dr, tx); else dbuf_sync_leaf(dr, tx); } } /* ARGSUSED */ static void dbuf_write_ready(zio_t *zio, arc_buf_t *buf, void *vdb) { 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, ==, 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(db->db_blkptr)) && db->db_blkptr == &dn->dn_phys->dn_spill); } #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) dn->dn_phys->dn_maxblkid = db->db_blkid; mutex_exit(&dn->dn_mtx); if (dn->dn_type == DMU_OT_DNODE) { dnode_phys_t *dnp = db->db.db_data; for (i = db->db.db_size >> DNODE_SHIFT; i > 0; i--, dnp++) { if (dnp->dn_type != DMU_OT_NONE) fill++; } } 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; fill += BP_GET_FILL(ibp); } } DB_DNODE_EXIT(db); if (!BP_IS_EMBEDDED(bp)) bp->blk_fill = fill; mutex_exit(&db->db_mtx); } /* * 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. */ /* ARGSUSED */ static void dbuf_write_physdone(zio_t *zio, arc_buf_t *buf, void *arg) { 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 dsl_pool_sync()'s call to * dsl_pool_undirty_space(). */ delta = dr->dr_accounted / zio->io_phys_children; dsl_pool_undirty_space(dp, delta, zio->io_txg); } /* ARGSUSED */ static void dbuf_write_done(zio_t *zio, arc_buf_t *buf, void *vdb) { 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; dbuf_dirty_record_t **drp, *dr; 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); drp = &db->db_last_dirty; while ((dr = *drp) != db->db_data_pending) drp = &dr->dr_next; ASSERT(!list_link_active(&dr->dr_dirty_node)); ASSERT(dr->dr_dbuf == db); ASSERT(dr->dr_next == NULL); *drp = dr->dr_next; #ifdef ZFS_DEBUG if (db->db_blkid == DMU_SPILL_BLKID) { dnode_t *dn; DB_DNODE_ENTER(db); dn = DB_DNODE(db); ASSERT(dn->dn_phys->dn_flags & DNODE_FLAG_SPILL_BLKPTR); ASSERT(!(BP_IS_HOLE(db->db_blkptr)) && db->db_blkptr == &dn->dn_phys->dn_spill); DB_DNODE_EXIT(db); } #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 != db->db_buf) VERIFY(arc_buf_remove_ref(dr->dt.dl.dr_data, db)); else if (!arc_released(db->db_buf)) arc_set_callback(db->db_buf, dbuf_do_evict, db); } } else { dnode_t *dn; DB_DNODE_ENTER(db); dn = DB_DNODE(db); 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)) { ASSERTV(int epbs = 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); if (!arc_released(db->db_buf)) arc_set_callback(db->db_buf, dbuf_do_evict, db); } DB_DNODE_EXIT(db); mutex_destroy(&dr->dt.di.dr_mtx); list_destroy(&dr->dt.di.dr_children); } kmem_free(dr, sizeof (dbuf_dirty_record_t)); 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); } 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); } /* 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; 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 *zio; int wp_flag = 0; DB_DNODE_ENTER(db); dn = DB_DNODE(db); 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); } } } 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)); zio = 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]); zio = dn->dn_zio; } ASSERT(db->db_level == 0 || data == db->db_buf); ASSERT3U(db->db_blkptr->blk_birth, <=, txg); ASSERT(zio); 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); DB_DNODE_EXIT(db); 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()). */ void *contents = (data != NULL) ? data->b_data : NULL; dr->dr_zio = zio_write(zio, os->os_spa, txg, db->db_blkptr, contents, db->db.db_size, &zp, dbuf_write_override_ready, 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); mutex_exit(&db->db_mtx); } else if (db->db_state == DB_NOFILL) { ASSERT(zp.zp_checksum == ZIO_CHECKSUM_OFF); dr->dr_zio = zio_write(zio, os->os_spa, txg, db->db_blkptr, NULL, db->db.db_size, &zp, dbuf_write_nofill_ready, NULL, dbuf_write_nofill_done, db, ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_MUSTSUCCEED | ZIO_FLAG_NODATA, &zb); } else { ASSERT(arc_released(data)); dr->dr_zio = arc_write(zio, os->os_spa, txg, db->db_blkptr, data, DBUF_IS_L2CACHEABLE(db), DBUF_IS_L2COMPRESSIBLE(db), &zp, dbuf_write_ready, dbuf_write_physdone, dbuf_write_done, db, ZIO_PRIORITY_ASYNC_WRITE, ZIO_FLAG_MUSTSUCCEED, &zb); } } #if defined(_KERNEL) && defined(HAVE_SPL) EXPORT_SYMBOL(dbuf_find); EXPORT_SYMBOL(dbuf_is_metadata); EXPORT_SYMBOL(dbuf_evict); 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_will_dirty); 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_clear); 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_freeable); EXPORT_SYMBOL(dmu_buf_get_blkptr); #endif diff --git a/module/zfs/dmu.c b/module/zfs/dmu.c index b4133f0e4205..1bf108bfeabb 100644 --- a/module/zfs/dmu.c +++ b/module/zfs/dmu.c @@ -1,2024 +1,2031 @@ /* * 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) 2005, 2010, Oracle and/or its affiliates. All rights reserved. - * Copyright (c) 2011, 2014 by Delphix. All rights reserved. + * Copyright (c) 2011, 2015 by Delphix. All rights reserved. * Copyright (c) 2013 by Saso Kiselkov. All rights reserved. * Copyright (c) 2014, Nexenta Systems, Inc. All rights reserved. * Copyright (c) 2015 by Chunwei Chen. All rights reserved. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef _KERNEL #include #include #endif /* * Enable/disable nopwrite feature. */ int zfs_nopwrite_enabled = 1; const dmu_object_type_info_t dmu_ot[DMU_OT_NUMTYPES] = { { DMU_BSWAP_UINT8, TRUE, "unallocated" }, { DMU_BSWAP_ZAP, TRUE, "object directory" }, { DMU_BSWAP_UINT64, TRUE, "object array" }, { DMU_BSWAP_UINT8, TRUE, "packed nvlist" }, { DMU_BSWAP_UINT64, TRUE, "packed nvlist size" }, { DMU_BSWAP_UINT64, TRUE, "bpobj" }, { DMU_BSWAP_UINT64, TRUE, "bpobj header" }, { DMU_BSWAP_UINT64, TRUE, "SPA space map header" }, { DMU_BSWAP_UINT64, TRUE, "SPA space map" }, { DMU_BSWAP_UINT64, TRUE, "ZIL intent log" }, { DMU_BSWAP_DNODE, TRUE, "DMU dnode" }, { DMU_BSWAP_OBJSET, TRUE, "DMU objset" }, { DMU_BSWAP_UINT64, TRUE, "DSL directory" }, { DMU_BSWAP_ZAP, TRUE, "DSL directory child map"}, { DMU_BSWAP_ZAP, TRUE, "DSL dataset snap map" }, { DMU_BSWAP_ZAP, TRUE, "DSL props" }, { DMU_BSWAP_UINT64, TRUE, "DSL dataset" }, { DMU_BSWAP_ZNODE, TRUE, "ZFS znode" }, { DMU_BSWAP_OLDACL, TRUE, "ZFS V0 ACL" }, { DMU_BSWAP_UINT8, FALSE, "ZFS plain file" }, { DMU_BSWAP_ZAP, TRUE, "ZFS directory" }, { DMU_BSWAP_ZAP, TRUE, "ZFS master node" }, { DMU_BSWAP_ZAP, TRUE, "ZFS delete queue" }, { DMU_BSWAP_UINT8, FALSE, "zvol object" }, { DMU_BSWAP_ZAP, TRUE, "zvol prop" }, { DMU_BSWAP_UINT8, FALSE, "other uint8[]" }, { DMU_BSWAP_UINT64, FALSE, "other uint64[]" }, { DMU_BSWAP_ZAP, TRUE, "other ZAP" }, { DMU_BSWAP_ZAP, TRUE, "persistent error log" }, { DMU_BSWAP_UINT8, TRUE, "SPA history" }, { DMU_BSWAP_UINT64, TRUE, "SPA history offsets" }, { DMU_BSWAP_ZAP, TRUE, "Pool properties" }, { DMU_BSWAP_ZAP, TRUE, "DSL permissions" }, { DMU_BSWAP_ACL, TRUE, "ZFS ACL" }, { DMU_BSWAP_UINT8, TRUE, "ZFS SYSACL" }, { DMU_BSWAP_UINT8, TRUE, "FUID table" }, { DMU_BSWAP_UINT64, TRUE, "FUID table size" }, { DMU_BSWAP_ZAP, TRUE, "DSL dataset next clones"}, { DMU_BSWAP_ZAP, TRUE, "scan work queue" }, { DMU_BSWAP_ZAP, TRUE, "ZFS user/group used" }, { DMU_BSWAP_ZAP, TRUE, "ZFS user/group quota" }, { DMU_BSWAP_ZAP, TRUE, "snapshot refcount tags"}, { DMU_BSWAP_ZAP, TRUE, "DDT ZAP algorithm" }, { DMU_BSWAP_ZAP, TRUE, "DDT statistics" }, { DMU_BSWAP_UINT8, TRUE, "System attributes" }, { DMU_BSWAP_ZAP, TRUE, "SA master node" }, { DMU_BSWAP_ZAP, TRUE, "SA attr registration" }, { DMU_BSWAP_ZAP, TRUE, "SA attr layouts" }, { DMU_BSWAP_ZAP, TRUE, "scan translations" }, { DMU_BSWAP_UINT8, FALSE, "deduplicated block" }, { DMU_BSWAP_ZAP, TRUE, "DSL deadlist map" }, { DMU_BSWAP_UINT64, TRUE, "DSL deadlist map hdr" }, { DMU_BSWAP_ZAP, TRUE, "DSL dir clones" }, { DMU_BSWAP_UINT64, TRUE, "bpobj subobj" } }; const dmu_object_byteswap_info_t dmu_ot_byteswap[DMU_BSWAP_NUMFUNCS] = { { byteswap_uint8_array, "uint8" }, { byteswap_uint16_array, "uint16" }, { byteswap_uint32_array, "uint32" }, { byteswap_uint64_array, "uint64" }, { zap_byteswap, "zap" }, { dnode_buf_byteswap, "dnode" }, { dmu_objset_byteswap, "objset" }, { zfs_znode_byteswap, "znode" }, { zfs_oldacl_byteswap, "oldacl" }, { zfs_acl_byteswap, "acl" } }; int dmu_buf_hold_noread(objset_t *os, uint64_t object, uint64_t offset, void *tag, dmu_buf_t **dbp) { dnode_t *dn; uint64_t blkid; dmu_buf_impl_t *db; int err; err = dnode_hold(os, object, FTAG, &dn); if (err) return (err); blkid = dbuf_whichblock(dn, 0, offset); rw_enter(&dn->dn_struct_rwlock, RW_READER); db = dbuf_hold(dn, blkid, tag); rw_exit(&dn->dn_struct_rwlock); dnode_rele(dn, FTAG); if (db == NULL) { *dbp = NULL; return (SET_ERROR(EIO)); } *dbp = &db->db; return (err); } int dmu_buf_hold(objset_t *os, uint64_t object, uint64_t offset, void *tag, dmu_buf_t **dbp, int flags) { int err; int db_flags = DB_RF_CANFAIL; if (flags & DMU_READ_NO_PREFETCH) db_flags |= DB_RF_NOPREFETCH; err = dmu_buf_hold_noread(os, object, offset, tag, dbp); if (err == 0) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)(*dbp); err = dbuf_read(db, NULL, db_flags); if (err != 0) { dbuf_rele(db, tag); *dbp = NULL; } } return (err); } int dmu_bonus_max(void) { return (DN_MAX_BONUSLEN); } int dmu_set_bonus(dmu_buf_t *db_fake, int newsize, dmu_tx_t *tx) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake; dnode_t *dn; int error; DB_DNODE_ENTER(db); dn = DB_DNODE(db); if (dn->dn_bonus != db) { error = SET_ERROR(EINVAL); } else if (newsize < 0 || newsize > db_fake->db_size) { error = SET_ERROR(EINVAL); } else { dnode_setbonuslen(dn, newsize, tx); error = 0; } DB_DNODE_EXIT(db); return (error); } int dmu_set_bonustype(dmu_buf_t *db_fake, dmu_object_type_t type, dmu_tx_t *tx) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake; dnode_t *dn; int error; DB_DNODE_ENTER(db); dn = DB_DNODE(db); if (!DMU_OT_IS_VALID(type)) { error = SET_ERROR(EINVAL); } else if (dn->dn_bonus != db) { error = SET_ERROR(EINVAL); } else { dnode_setbonus_type(dn, type, tx); error = 0; } DB_DNODE_EXIT(db); return (error); } dmu_object_type_t dmu_get_bonustype(dmu_buf_t *db_fake) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake; dnode_t *dn; dmu_object_type_t type; DB_DNODE_ENTER(db); dn = DB_DNODE(db); type = dn->dn_bonustype; DB_DNODE_EXIT(db); return (type); } int dmu_rm_spill(objset_t *os, uint64_t object, dmu_tx_t *tx) { dnode_t *dn; int error; error = dnode_hold(os, object, FTAG, &dn); dbuf_rm_spill(dn, tx); rw_enter(&dn->dn_struct_rwlock, RW_WRITER); dnode_rm_spill(dn, tx); rw_exit(&dn->dn_struct_rwlock); dnode_rele(dn, FTAG); return (error); } /* * returns ENOENT, EIO, or 0. */ int dmu_bonus_hold(objset_t *os, uint64_t object, void *tag, dmu_buf_t **dbp) { dnode_t *dn; dmu_buf_impl_t *db; int error; error = dnode_hold(os, object, FTAG, &dn); if (error) return (error); rw_enter(&dn->dn_struct_rwlock, RW_READER); if (dn->dn_bonus == NULL) { rw_exit(&dn->dn_struct_rwlock); rw_enter(&dn->dn_struct_rwlock, RW_WRITER); if (dn->dn_bonus == NULL) dbuf_create_bonus(dn); } db = dn->dn_bonus; /* as long as the bonus buf is held, the dnode will be held */ if (refcount_add(&db->db_holds, tag) == 1) { VERIFY(dnode_add_ref(dn, db)); atomic_inc_32(&dn->dn_dbufs_count); } /* * Wait to drop dn_struct_rwlock until after adding the bonus dbuf's * hold and incrementing the dbuf count to ensure that dnode_move() sees * a dnode hold for every dbuf. */ rw_exit(&dn->dn_struct_rwlock); dnode_rele(dn, FTAG); VERIFY(0 == dbuf_read(db, NULL, DB_RF_MUST_SUCCEED | DB_RF_NOPREFETCH)); *dbp = &db->db; return (0); } /* * returns ENOENT, EIO, or 0. * * This interface will allocate a blank spill dbuf when a spill blk * doesn't already exist on the dnode. * * if you only want to find an already existing spill db, then * dmu_spill_hold_existing() should be used. */ int dmu_spill_hold_by_dnode(dnode_t *dn, uint32_t flags, void *tag, dmu_buf_t **dbp) { dmu_buf_impl_t *db = NULL; int err; if ((flags & DB_RF_HAVESTRUCT) == 0) rw_enter(&dn->dn_struct_rwlock, RW_READER); db = dbuf_hold(dn, DMU_SPILL_BLKID, tag); if ((flags & DB_RF_HAVESTRUCT) == 0) rw_exit(&dn->dn_struct_rwlock); ASSERT(db != NULL); err = dbuf_read(db, NULL, flags); if (err == 0) *dbp = &db->db; else dbuf_rele(db, tag); return (err); } int dmu_spill_hold_existing(dmu_buf_t *bonus, void *tag, dmu_buf_t **dbp) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)bonus; dnode_t *dn; int err; DB_DNODE_ENTER(db); dn = DB_DNODE(db); if (spa_version(dn->dn_objset->os_spa) < SPA_VERSION_SA) { err = SET_ERROR(EINVAL); } else { rw_enter(&dn->dn_struct_rwlock, RW_READER); if (!dn->dn_have_spill) { err = SET_ERROR(ENOENT); } else { err = dmu_spill_hold_by_dnode(dn, DB_RF_HAVESTRUCT | DB_RF_CANFAIL, tag, dbp); } rw_exit(&dn->dn_struct_rwlock); } DB_DNODE_EXIT(db); return (err); } int dmu_spill_hold_by_bonus(dmu_buf_t *bonus, void *tag, dmu_buf_t **dbp) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)bonus; dnode_t *dn; int err; DB_DNODE_ENTER(db); dn = DB_DNODE(db); err = dmu_spill_hold_by_dnode(dn, DB_RF_CANFAIL, tag, dbp); DB_DNODE_EXIT(db); return (err); } /* * Note: longer-term, we should modify all of the dmu_buf_*() interfaces * to take a held dnode rather than -- the lookup is wasteful, * and can induce severe lock contention when writing to several files * whose dnodes are in the same block. */ static int dmu_buf_hold_array_by_dnode(dnode_t *dn, uint64_t offset, uint64_t length, - int read, void *tag, int *numbufsp, dmu_buf_t ***dbpp, uint32_t flags) + boolean_t read, void *tag, int *numbufsp, dmu_buf_t ***dbpp, uint32_t flags) { dmu_buf_t **dbp; uint64_t blkid, nblks, i; uint32_t dbuf_flags; int err; zio_t *zio; ASSERT(length <= DMU_MAX_ACCESS); - dbuf_flags = DB_RF_CANFAIL | DB_RF_NEVERWAIT | DB_RF_HAVESTRUCT; - if (flags & DMU_READ_NO_PREFETCH || length > zfetch_array_rd_sz) - dbuf_flags |= DB_RF_NOPREFETCH; + /* + * Note: We directly notify the prefetch code of this read, so that + * we can tell it about the multi-block read. dbuf_read() only knows + * about the one block it is accessing. + */ + dbuf_flags = DB_RF_CANFAIL | DB_RF_NEVERWAIT | DB_RF_HAVESTRUCT | + DB_RF_NOPREFETCH; rw_enter(&dn->dn_struct_rwlock, RW_READER); if (dn->dn_datablkshift) { int blkshift = dn->dn_datablkshift; - nblks = (P2ROUNDUP(offset+length, 1ULL<> blkshift; + nblks = (P2ROUNDUP(offset + length, 1ULL << blkshift) - + P2ALIGN(offset, 1ULL << blkshift)) >> blkshift; } else { if (offset + length > dn->dn_datablksz) { zfs_panic_recover("zfs: accessing past end of object " "%llx/%llx (size=%u access=%llu+%llu)", (longlong_t)dn->dn_objset-> os_dsl_dataset->ds_object, (longlong_t)dn->dn_object, dn->dn_datablksz, (longlong_t)offset, (longlong_t)length); rw_exit(&dn->dn_struct_rwlock); return (SET_ERROR(EIO)); } nblks = 1; } dbp = kmem_zalloc(sizeof (dmu_buf_t *) * nblks, KM_SLEEP); zio = zio_root(dn->dn_objset->os_spa, NULL, NULL, ZIO_FLAG_CANFAIL); blkid = dbuf_whichblock(dn, 0, offset); for (i = 0; i < nblks; i++) { - dmu_buf_impl_t *db = dbuf_hold(dn, blkid+i, tag); + dmu_buf_impl_t *db = dbuf_hold(dn, blkid + i, tag); if (db == NULL) { rw_exit(&dn->dn_struct_rwlock); dmu_buf_rele_array(dbp, nblks, tag); zio_nowait(zio); return (SET_ERROR(EIO)); } + /* initiate async i/o */ - if (read) { + if (read) (void) dbuf_read(db, zio, dbuf_flags); - } dbp[i] = &db->db; } + + if ((flags & DMU_READ_NO_PREFETCH) == 0 && read && + length < zfetch_array_rd_sz) { + dmu_zfetch(&dn->dn_zfetch, blkid, nblks); + } rw_exit(&dn->dn_struct_rwlock); /* wait for async i/o */ err = zio_wait(zio); if (err) { dmu_buf_rele_array(dbp, nblks, tag); return (err); } /* wait for other io to complete */ if (read) { for (i = 0; i < nblks; i++) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)dbp[i]; 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) err = SET_ERROR(EIO); mutex_exit(&db->db_mtx); if (err) { dmu_buf_rele_array(dbp, nblks, tag); return (err); } } } *numbufsp = nblks; *dbpp = dbp; return (0); } static int dmu_buf_hold_array(objset_t *os, uint64_t object, uint64_t offset, uint64_t length, int read, void *tag, int *numbufsp, dmu_buf_t ***dbpp) { dnode_t *dn; int err; err = dnode_hold(os, object, FTAG, &dn); if (err) return (err); err = dmu_buf_hold_array_by_dnode(dn, offset, length, read, tag, numbufsp, dbpp, DMU_READ_PREFETCH); dnode_rele(dn, FTAG); return (err); } int dmu_buf_hold_array_by_bonus(dmu_buf_t *db_fake, uint64_t offset, - uint64_t length, int read, void *tag, int *numbufsp, dmu_buf_t ***dbpp) + uint64_t length, boolean_t read, void *tag, int *numbufsp, + dmu_buf_t ***dbpp) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake; dnode_t *dn; int err; DB_DNODE_ENTER(db); dn = DB_DNODE(db); err = dmu_buf_hold_array_by_dnode(dn, offset, length, read, tag, numbufsp, dbpp, DMU_READ_PREFETCH); DB_DNODE_EXIT(db); return (err); } void dmu_buf_rele_array(dmu_buf_t **dbp_fake, int numbufs, void *tag) { int i; dmu_buf_impl_t **dbp = (dmu_buf_impl_t **)dbp_fake; if (numbufs == 0) return; for (i = 0; i < numbufs; i++) { if (dbp[i]) dbuf_rele(dbp[i], tag); } kmem_free(dbp, sizeof (dmu_buf_t *) * numbufs); } /* * Issue prefetch i/os for the given blocks. If level is greater than 0, the * indirect blocks prefeteched will be those that point to the blocks containing * the data starting at offset, and continuing to offset + len. * * Note that if the indirect blocks above the blocks being prefetched are not in * cache, they will be asychronously read in. */ void dmu_prefetch(objset_t *os, uint64_t object, int64_t level, uint64_t offset, uint64_t len, zio_priority_t pri) { dnode_t *dn; uint64_t blkid; int nblks, err; - if (zfs_prefetch_disable) - return; - if (len == 0) { /* they're interested in the bonus buffer */ dn = DMU_META_DNODE(os); if (object == 0 || object >= DN_MAX_OBJECT) return; rw_enter(&dn->dn_struct_rwlock, RW_READER); blkid = dbuf_whichblock(dn, level, object * sizeof (dnode_phys_t)); dbuf_prefetch(dn, level, blkid, pri, 0); rw_exit(&dn->dn_struct_rwlock); return; } /* * XXX - Note, if the dnode for the requested object is not * already cached, we will do a *synchronous* read in the * dnode_hold() call. The same is true for any indirects. */ err = dnode_hold(os, object, FTAG, &dn); if (err != 0) return; rw_enter(&dn->dn_struct_rwlock, RW_READER); /* * offset + len - 1 is the last byte we want to prefetch for, and offset * is the first. Then dbuf_whichblk(dn, level, off + len - 1) is the * last block we want to prefetch, and dbuf_whichblock(dn, level, * offset) is the first. Then the number we need to prefetch is the * last - first + 1. */ if (level > 0 || dn->dn_datablkshift != 0) { nblks = dbuf_whichblock(dn, level, offset + len - 1) - dbuf_whichblock(dn, level, offset) + 1; } else { nblks = (offset < dn->dn_datablksz); } if (nblks != 0) { int i; blkid = dbuf_whichblock(dn, level, offset); for (i = 0; i < nblks; i++) dbuf_prefetch(dn, level, blkid + i, pri, 0); } rw_exit(&dn->dn_struct_rwlock); dnode_rele(dn, FTAG); } /* * Get the next "chunk" of file data to free. We traverse the file from * the end so that the file gets shorter over time (if we crashes in the * middle, this will leave us in a better state). We find allocated file * data by simply searching the allocated level 1 indirects. * * On input, *start should be the first offset that does not need to be * freed (e.g. "offset + length"). On return, *start will be the first * offset that should be freed. */ static int get_next_chunk(dnode_t *dn, uint64_t *start, uint64_t minimum) { uint64_t maxblks = DMU_MAX_ACCESS >> (dn->dn_indblkshift + 1); /* bytes of data covered by a level-1 indirect block */ uint64_t iblkrange = dn->dn_datablksz * EPB(dn->dn_indblkshift, SPA_BLKPTRSHIFT); uint64_t blks; ASSERT3U(minimum, <=, *start); if (*start - minimum <= iblkrange * maxblks) { *start = minimum; return (0); } ASSERT(ISP2(iblkrange)); for (blks = 0; *start > minimum && blks < maxblks; blks++) { int err; /* * dnode_next_offset(BACKWARDS) will find an allocated L1 * indirect block at or before the input offset. We must * decrement *start so that it is at the end of the region * to search. */ (*start)--; err = dnode_next_offset(dn, DNODE_FIND_BACKWARDS, start, 2, 1, 0); /* if there are no indirect blocks before start, we are done */ if (err == ESRCH) { *start = minimum; break; } else if (err != 0) { return (err); } /* set start to the beginning of this L1 indirect */ *start = P2ALIGN(*start, iblkrange); } if (*start < minimum) *start = minimum; return (0); } static int dmu_free_long_range_impl(objset_t *os, dnode_t *dn, uint64_t offset, uint64_t length) { uint64_t object_size; int err; if (dn == NULL) return (SET_ERROR(EINVAL)); object_size = (dn->dn_maxblkid + 1) * dn->dn_datablksz; if (offset >= object_size) return (0); if (length == DMU_OBJECT_END || offset + length > object_size) length = object_size - offset; while (length != 0) { uint64_t chunk_end, chunk_begin; dmu_tx_t *tx; chunk_end = chunk_begin = offset + length; /* move chunk_begin backwards to the beginning of this chunk */ err = get_next_chunk(dn, &chunk_begin, offset); if (err) return (err); ASSERT3U(chunk_begin, >=, offset); ASSERT3U(chunk_begin, <=, chunk_end); tx = dmu_tx_create(os); dmu_tx_hold_free(tx, dn->dn_object, chunk_begin, chunk_end - chunk_begin); err = dmu_tx_assign(tx, TXG_WAIT); if (err) { dmu_tx_abort(tx); return (err); } dnode_free_range(dn, chunk_begin, chunk_end - chunk_begin, tx); dmu_tx_commit(tx); length -= chunk_end - chunk_begin; } return (0); } int dmu_free_long_range(objset_t *os, uint64_t object, uint64_t offset, uint64_t length) { dnode_t *dn; int err; err = dnode_hold(os, object, FTAG, &dn); if (err != 0) return (err); err = dmu_free_long_range_impl(os, dn, offset, length); /* * It is important to zero out the maxblkid when freeing the entire * file, so that (a) subsequent calls to dmu_free_long_range_impl() * will take the fast path, and (b) dnode_reallocate() can verify * that the entire file has been freed. */ if (err == 0 && offset == 0 && length == DMU_OBJECT_END) dn->dn_maxblkid = 0; dnode_rele(dn, FTAG); return (err); } int dmu_free_long_object(objset_t *os, uint64_t object) { dmu_tx_t *tx; int err; err = dmu_free_long_range(os, object, 0, DMU_OBJECT_END); if (err != 0) return (err); tx = dmu_tx_create(os); dmu_tx_hold_bonus(tx, object); dmu_tx_hold_free(tx, object, 0, DMU_OBJECT_END); err = dmu_tx_assign(tx, TXG_WAIT); if (err == 0) { err = dmu_object_free(os, object, tx); dmu_tx_commit(tx); } else { dmu_tx_abort(tx); } return (err); } int dmu_free_range(objset_t *os, uint64_t object, uint64_t offset, uint64_t size, dmu_tx_t *tx) { dnode_t *dn; int err = dnode_hold(os, object, FTAG, &dn); if (err) return (err); ASSERT(offset < UINT64_MAX); ASSERT(size == -1ULL || size <= UINT64_MAX - offset); dnode_free_range(dn, offset, size, tx); dnode_rele(dn, FTAG); return (0); } int dmu_read(objset_t *os, uint64_t object, uint64_t offset, uint64_t size, void *buf, uint32_t flags) { dnode_t *dn; dmu_buf_t **dbp; int numbufs, err; err = dnode_hold(os, object, FTAG, &dn); if (err) return (err); /* * Deal with odd block sizes, where there can't be data past the first * block. If we ever do the tail block optimization, we will need to * handle that here as well. */ if (dn->dn_maxblkid == 0) { uint64_t newsz = offset > dn->dn_datablksz ? 0 : MIN(size, dn->dn_datablksz - offset); bzero((char *)buf + newsz, size - newsz); size = newsz; } while (size > 0) { uint64_t mylen = MIN(size, DMU_MAX_ACCESS / 2); int i; /* * NB: we could do this block-at-a-time, but it's nice * to be reading in parallel. */ err = dmu_buf_hold_array_by_dnode(dn, offset, mylen, TRUE, FTAG, &numbufs, &dbp, flags); if (err) break; for (i = 0; i < numbufs; i++) { uint64_t tocpy; int64_t bufoff; dmu_buf_t *db = dbp[i]; ASSERT(size > 0); bufoff = offset - db->db_offset; tocpy = MIN(db->db_size - bufoff, size); (void) memcpy(buf, (char *)db->db_data + bufoff, tocpy); offset += tocpy; size -= tocpy; buf = (char *)buf + tocpy; } dmu_buf_rele_array(dbp, numbufs, FTAG); } dnode_rele(dn, FTAG); return (err); } void dmu_write(objset_t *os, uint64_t object, uint64_t offset, uint64_t size, const void *buf, dmu_tx_t *tx) { dmu_buf_t **dbp; int numbufs, i; if (size == 0) return; VERIFY0(dmu_buf_hold_array(os, object, offset, size, FALSE, FTAG, &numbufs, &dbp)); for (i = 0; i < numbufs; i++) { uint64_t tocpy; int64_t bufoff; dmu_buf_t *db = dbp[i]; ASSERT(size > 0); bufoff = offset - db->db_offset; tocpy = MIN(db->db_size - bufoff, size); ASSERT(i == 0 || i == numbufs-1 || tocpy == db->db_size); if (tocpy == db->db_size) dmu_buf_will_fill(db, tx); else dmu_buf_will_dirty(db, tx); (void) memcpy((char *)db->db_data + bufoff, buf, tocpy); if (tocpy == db->db_size) dmu_buf_fill_done(db, tx); offset += tocpy; size -= tocpy; buf = (char *)buf + tocpy; } dmu_buf_rele_array(dbp, numbufs, FTAG); } void dmu_prealloc(objset_t *os, uint64_t object, uint64_t offset, uint64_t size, dmu_tx_t *tx) { dmu_buf_t **dbp; int numbufs, i; if (size == 0) return; VERIFY(0 == dmu_buf_hold_array(os, object, offset, size, FALSE, FTAG, &numbufs, &dbp)); for (i = 0; i < numbufs; i++) { dmu_buf_t *db = dbp[i]; dmu_buf_will_not_fill(db, tx); } dmu_buf_rele_array(dbp, numbufs, FTAG); } 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) { dmu_buf_t *db; ASSERT3U(etype, <, NUM_BP_EMBEDDED_TYPES); ASSERT3U(comp, <, ZIO_COMPRESS_FUNCTIONS); VERIFY0(dmu_buf_hold_noread(os, object, offset, FTAG, &db)); dmu_buf_write_embedded(db, data, (bp_embedded_type_t)etype, (enum zio_compress)comp, uncompressed_size, compressed_size, byteorder, tx); dmu_buf_rele(db, FTAG); } /* * DMU support for xuio */ kstat_t *xuio_ksp = NULL; typedef struct xuio_stats { /* loaned yet not returned arc_buf */ kstat_named_t xuiostat_onloan_rbuf; kstat_named_t xuiostat_onloan_wbuf; /* whether a copy is made when loaning out a read buffer */ kstat_named_t xuiostat_rbuf_copied; kstat_named_t xuiostat_rbuf_nocopy; /* whether a copy is made when assigning a write buffer */ kstat_named_t xuiostat_wbuf_copied; kstat_named_t xuiostat_wbuf_nocopy; } xuio_stats_t; static xuio_stats_t xuio_stats = { { "onloan_read_buf", KSTAT_DATA_UINT64 }, { "onloan_write_buf", KSTAT_DATA_UINT64 }, { "read_buf_copied", KSTAT_DATA_UINT64 }, { "read_buf_nocopy", KSTAT_DATA_UINT64 }, { "write_buf_copied", KSTAT_DATA_UINT64 }, { "write_buf_nocopy", KSTAT_DATA_UINT64 } }; #define XUIOSTAT_INCR(stat, val) \ atomic_add_64(&xuio_stats.stat.value.ui64, (val)) #define XUIOSTAT_BUMP(stat) XUIOSTAT_INCR(stat, 1) int dmu_xuio_init(xuio_t *xuio, int nblk) { dmu_xuio_t *priv; uio_t *uio = &xuio->xu_uio; uio->uio_iovcnt = nblk; uio->uio_iov = kmem_zalloc(nblk * sizeof (iovec_t), KM_SLEEP); priv = kmem_zalloc(sizeof (dmu_xuio_t), KM_SLEEP); priv->cnt = nblk; priv->bufs = kmem_zalloc(nblk * sizeof (arc_buf_t *), KM_SLEEP); priv->iovp = (iovec_t *)uio->uio_iov; XUIO_XUZC_PRIV(xuio) = priv; if (XUIO_XUZC_RW(xuio) == UIO_READ) XUIOSTAT_INCR(xuiostat_onloan_rbuf, nblk); else XUIOSTAT_INCR(xuiostat_onloan_wbuf, nblk); return (0); } void dmu_xuio_fini(xuio_t *xuio) { dmu_xuio_t *priv = XUIO_XUZC_PRIV(xuio); int nblk = priv->cnt; kmem_free(priv->iovp, nblk * sizeof (iovec_t)); kmem_free(priv->bufs, nblk * sizeof (arc_buf_t *)); kmem_free(priv, sizeof (dmu_xuio_t)); if (XUIO_XUZC_RW(xuio) == UIO_READ) XUIOSTAT_INCR(xuiostat_onloan_rbuf, -nblk); else XUIOSTAT_INCR(xuiostat_onloan_wbuf, -nblk); } /* * Initialize iov[priv->next] and priv->bufs[priv->next] with { off, n, abuf } * and increase priv->next by 1. */ int dmu_xuio_add(xuio_t *xuio, arc_buf_t *abuf, offset_t off, size_t n) { struct iovec *iov; uio_t *uio = &xuio->xu_uio; dmu_xuio_t *priv = XUIO_XUZC_PRIV(xuio); int i = priv->next++; ASSERT(i < priv->cnt); ASSERT(off + n <= arc_buf_size(abuf)); iov = (iovec_t *)uio->uio_iov + i; iov->iov_base = (char *)abuf->b_data + off; iov->iov_len = n; priv->bufs[i] = abuf; return (0); } int dmu_xuio_cnt(xuio_t *xuio) { dmu_xuio_t *priv = XUIO_XUZC_PRIV(xuio); return (priv->cnt); } arc_buf_t * dmu_xuio_arcbuf(xuio_t *xuio, int i) { dmu_xuio_t *priv = XUIO_XUZC_PRIV(xuio); ASSERT(i < priv->cnt); return (priv->bufs[i]); } void dmu_xuio_clear(xuio_t *xuio, int i) { dmu_xuio_t *priv = XUIO_XUZC_PRIV(xuio); ASSERT(i < priv->cnt); priv->bufs[i] = NULL; } static void xuio_stat_init(void) { xuio_ksp = kstat_create("zfs", 0, "xuio_stats", "misc", KSTAT_TYPE_NAMED, sizeof (xuio_stats) / sizeof (kstat_named_t), KSTAT_FLAG_VIRTUAL); if (xuio_ksp != NULL) { xuio_ksp->ks_data = &xuio_stats; kstat_install(xuio_ksp); } } static void xuio_stat_fini(void) { if (xuio_ksp != NULL) { kstat_delete(xuio_ksp); xuio_ksp = NULL; } } void xuio_stat_wbuf_copied() { XUIOSTAT_BUMP(xuiostat_wbuf_copied); } void xuio_stat_wbuf_nocopy() { XUIOSTAT_BUMP(xuiostat_wbuf_nocopy); } #ifdef _KERNEL static int dmu_read_uio_dnode(dnode_t *dn, uio_t *uio, uint64_t size) { dmu_buf_t **dbp; int numbufs, i, err; xuio_t *xuio = NULL; /* * NB: we could do this block-at-a-time, but it's nice * to be reading in parallel. */ err = dmu_buf_hold_array_by_dnode(dn, uio->uio_loffset, size, TRUE, FTAG, &numbufs, &dbp, 0); if (err) return (err); for (i = 0; i < numbufs; i++) { uint64_t tocpy; int64_t bufoff; dmu_buf_t *db = dbp[i]; ASSERT(size > 0); bufoff = uio->uio_loffset - db->db_offset; tocpy = MIN(db->db_size - bufoff, size); if (xuio) { dmu_buf_impl_t *dbi = (dmu_buf_impl_t *)db; arc_buf_t *dbuf_abuf = dbi->db_buf; arc_buf_t *abuf = dbuf_loan_arcbuf(dbi); err = dmu_xuio_add(xuio, abuf, bufoff, tocpy); if (!err) { uio->uio_resid -= tocpy; uio->uio_loffset += tocpy; } if (abuf == dbuf_abuf) XUIOSTAT_BUMP(xuiostat_rbuf_nocopy); else XUIOSTAT_BUMP(xuiostat_rbuf_copied); } else { err = uiomove((char *)db->db_data + bufoff, tocpy, UIO_READ, uio); } if (err) break; size -= tocpy; } dmu_buf_rele_array(dbp, numbufs, FTAG); return (err); } /* * Read 'size' bytes into the uio buffer. * From object zdb->db_object. * Starting at offset uio->uio_loffset. * * If the caller already has a dbuf in the target object * (e.g. its bonus buffer), this routine is faster than dmu_read_uio(), * because we don't have to find the dnode_t for the object. */ int dmu_read_uio_dbuf(dmu_buf_t *zdb, uio_t *uio, uint64_t size) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)zdb; dnode_t *dn; int err; if (size == 0) return (0); DB_DNODE_ENTER(db); dn = DB_DNODE(db); err = dmu_read_uio_dnode(dn, uio, size); DB_DNODE_EXIT(db); return (err); } /* * Read 'size' bytes into the uio buffer. * From the specified object * Starting at offset uio->uio_loffset. */ int dmu_read_uio(objset_t *os, uint64_t object, uio_t *uio, uint64_t size) { dnode_t *dn; int err; if (size == 0) return (0); err = dnode_hold(os, object, FTAG, &dn); if (err) return (err); err = dmu_read_uio_dnode(dn, uio, size); dnode_rele(dn, FTAG); return (err); } static int dmu_write_uio_dnode(dnode_t *dn, uio_t *uio, uint64_t size, dmu_tx_t *tx) { dmu_buf_t **dbp; int numbufs; int err = 0; int i; err = dmu_buf_hold_array_by_dnode(dn, uio->uio_loffset, size, FALSE, FTAG, &numbufs, &dbp, DMU_READ_PREFETCH); if (err) return (err); for (i = 0; i < numbufs; i++) { uint64_t tocpy; int64_t bufoff; dmu_buf_t *db = dbp[i]; ASSERT(size > 0); bufoff = uio->uio_loffset - db->db_offset; tocpy = MIN(db->db_size - bufoff, size); ASSERT(i == 0 || i == numbufs-1 || tocpy == db->db_size); if (tocpy == db->db_size) dmu_buf_will_fill(db, tx); else dmu_buf_will_dirty(db, tx); /* * XXX uiomove could block forever (eg.nfs-backed * pages). There needs to be a uiolockdown() function * to lock the pages in memory, so that uiomove won't * block. */ err = uiomove((char *)db->db_data + bufoff, tocpy, UIO_WRITE, uio); if (tocpy == db->db_size) dmu_buf_fill_done(db, tx); if (err) break; size -= tocpy; } dmu_buf_rele_array(dbp, numbufs, FTAG); return (err); } /* * Write 'size' bytes from the uio buffer. * To object zdb->db_object. * Starting at offset uio->uio_loffset. * * If the caller already has a dbuf in the target object * (e.g. its bonus buffer), this routine is faster than dmu_write_uio(), * because we don't have to find the dnode_t for the object. */ int dmu_write_uio_dbuf(dmu_buf_t *zdb, uio_t *uio, uint64_t size, dmu_tx_t *tx) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)zdb; dnode_t *dn; int err; if (size == 0) return (0); DB_DNODE_ENTER(db); dn = DB_DNODE(db); err = dmu_write_uio_dnode(dn, uio, size, tx); DB_DNODE_EXIT(db); return (err); } /* * Write 'size' bytes from the uio buffer. * To the specified object. * Starting at offset uio->uio_loffset. */ int dmu_write_uio(objset_t *os, uint64_t object, uio_t *uio, uint64_t size, dmu_tx_t *tx) { dnode_t *dn; int err; if (size == 0) return (0); err = dnode_hold(os, object, FTAG, &dn); if (err) return (err); err = dmu_write_uio_dnode(dn, uio, size, tx); dnode_rele(dn, FTAG); return (err); } #endif /* _KERNEL */ /* * Allocate a loaned anonymous arc buffer. */ arc_buf_t * dmu_request_arcbuf(dmu_buf_t *handle, int size) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)handle; return (arc_loan_buf(db->db_objset->os_spa, size)); } /* * Free a loaned arc buffer. */ void dmu_return_arcbuf(arc_buf_t *buf) { arc_return_buf(buf, FTAG); VERIFY(arc_buf_remove_ref(buf, FTAG)); } /* * When possible directly assign passed loaned arc buffer to a dbuf. * If this is not possible copy the contents of passed arc buf via * dmu_write(). */ void dmu_assign_arcbuf(dmu_buf_t *handle, uint64_t offset, arc_buf_t *buf, dmu_tx_t *tx) { dmu_buf_impl_t *dbuf = (dmu_buf_impl_t *)handle; dnode_t *dn; dmu_buf_impl_t *db; uint32_t blksz = (uint32_t)arc_buf_size(buf); uint64_t blkid; DB_DNODE_ENTER(dbuf); dn = DB_DNODE(dbuf); rw_enter(&dn->dn_struct_rwlock, RW_READER); blkid = dbuf_whichblock(dn, 0, offset); VERIFY((db = dbuf_hold(dn, blkid, FTAG)) != NULL); rw_exit(&dn->dn_struct_rwlock); DB_DNODE_EXIT(dbuf); /* * We can only assign if the offset is aligned, the arc buf is the * same size as the dbuf, and the dbuf is not metadata. It * can't be metadata because the loaned arc buf comes from the * user-data kmem area. */ if (offset == db->db.db_offset && blksz == db->db.db_size && DBUF_GET_BUFC_TYPE(db) == ARC_BUFC_DATA) { dbuf_assign_arcbuf(db, buf, tx); dbuf_rele(db, FTAG); } else { objset_t *os; uint64_t object; DB_DNODE_ENTER(dbuf); dn = DB_DNODE(dbuf); os = dn->dn_objset; object = dn->dn_object; DB_DNODE_EXIT(dbuf); dbuf_rele(db, FTAG); dmu_write(os, object, offset, blksz, buf->b_data, tx); dmu_return_arcbuf(buf); XUIOSTAT_BUMP(xuiostat_wbuf_copied); } } typedef struct { dbuf_dirty_record_t *dsa_dr; dmu_sync_cb_t *dsa_done; zgd_t *dsa_zgd; dmu_tx_t *dsa_tx; } dmu_sync_arg_t; /* ARGSUSED */ static void dmu_sync_ready(zio_t *zio, arc_buf_t *buf, void *varg) { dmu_sync_arg_t *dsa = varg; dmu_buf_t *db = dsa->dsa_zgd->zgd_db; blkptr_t *bp = zio->io_bp; if (zio->io_error == 0) { if (BP_IS_HOLE(bp)) { /* * A block of zeros may compress to a hole, but the * block size still needs to be known for replay. */ BP_SET_LSIZE(bp, db->db_size); } else if (!BP_IS_EMBEDDED(bp)) { ASSERT(BP_GET_LEVEL(bp) == 0); bp->blk_fill = 1; } } } static void dmu_sync_late_arrival_ready(zio_t *zio) { dmu_sync_ready(zio, NULL, zio->io_private); } /* ARGSUSED */ static void dmu_sync_done(zio_t *zio, arc_buf_t *buf, void *varg) { dmu_sync_arg_t *dsa = varg; dbuf_dirty_record_t *dr = dsa->dsa_dr; dmu_buf_impl_t *db = dr->dr_dbuf; mutex_enter(&db->db_mtx); ASSERT(dr->dt.dl.dr_override_state == DR_IN_DMU_SYNC); if (zio->io_error == 0) { dr->dt.dl.dr_nopwrite = !!(zio->io_flags & ZIO_FLAG_NOPWRITE); if (dr->dt.dl.dr_nopwrite) { ASSERTV(blkptr_t *bp = zio->io_bp); ASSERTV(blkptr_t *bp_orig = &zio->io_bp_orig); ASSERTV(uint8_t chksum = BP_GET_CHECKSUM(bp_orig)); ASSERT(BP_EQUAL(bp, bp_orig)); ASSERT(zio->io_prop.zp_compress != ZIO_COMPRESS_OFF); ASSERT(zio_checksum_table[chksum].ci_dedup); } dr->dt.dl.dr_overridden_by = *zio->io_bp; dr->dt.dl.dr_override_state = DR_OVERRIDDEN; dr->dt.dl.dr_copies = zio->io_prop.zp_copies; /* * Old style holes are filled with all zeros, whereas * new-style holes maintain their lsize, type, level, * and birth time (see zio_write_compress). While we * need to reset the BP_SET_LSIZE() call that happened * in dmu_sync_ready for old style holes, we do *not* * want to wipe out the information contained in new * style holes. Thus, only zero out the block pointer if * it's an old style hole. */ if (BP_IS_HOLE(&dr->dt.dl.dr_overridden_by) && dr->dt.dl.dr_overridden_by.blk_birth == 0) BP_ZERO(&dr->dt.dl.dr_overridden_by); } else { dr->dt.dl.dr_override_state = DR_NOT_OVERRIDDEN; } cv_broadcast(&db->db_changed); mutex_exit(&db->db_mtx); dsa->dsa_done(dsa->dsa_zgd, zio->io_error); kmem_free(dsa, sizeof (*dsa)); } static void dmu_sync_late_arrival_done(zio_t *zio) { blkptr_t *bp = zio->io_bp; dmu_sync_arg_t *dsa = zio->io_private; ASSERTV(blkptr_t *bp_orig = &zio->io_bp_orig); if (zio->io_error == 0 && !BP_IS_HOLE(bp)) { /* * If we didn't allocate a new block (i.e. ZIO_FLAG_NOPWRITE) * then there is nothing to do here. Otherwise, free the * newly allocated block in this txg. */ if (zio->io_flags & ZIO_FLAG_NOPWRITE) { ASSERT(BP_EQUAL(bp, bp_orig)); } else { ASSERT(BP_IS_HOLE(bp_orig) || !BP_EQUAL(bp, bp_orig)); ASSERT(zio->io_bp->blk_birth == zio->io_txg); ASSERT(zio->io_txg > spa_syncing_txg(zio->io_spa)); zio_free(zio->io_spa, zio->io_txg, zio->io_bp); } } dmu_tx_commit(dsa->dsa_tx); dsa->dsa_done(dsa->dsa_zgd, zio->io_error); kmem_free(dsa, sizeof (*dsa)); } static int dmu_sync_late_arrival(zio_t *pio, objset_t *os, dmu_sync_cb_t *done, zgd_t *zgd, zio_prop_t *zp, zbookmark_phys_t *zb) { dmu_sync_arg_t *dsa; dmu_tx_t *tx; tx = dmu_tx_create(os); dmu_tx_hold_space(tx, zgd->zgd_db->db_size); if (dmu_tx_assign(tx, TXG_WAIT) != 0) { dmu_tx_abort(tx); /* Make zl_get_data do txg_waited_synced() */ return (SET_ERROR(EIO)); } dsa = kmem_alloc(sizeof (dmu_sync_arg_t), KM_SLEEP); dsa->dsa_dr = NULL; dsa->dsa_done = done; dsa->dsa_zgd = zgd; dsa->dsa_tx = tx; zio_nowait(zio_write(pio, os->os_spa, dmu_tx_get_txg(tx), zgd->zgd_bp, zgd->zgd_db->db_data, zgd->zgd_db->db_size, zp, dmu_sync_late_arrival_ready, NULL, dmu_sync_late_arrival_done, dsa, ZIO_PRIORITY_SYNC_WRITE, ZIO_FLAG_CANFAIL|ZIO_FLAG_FASTWRITE, zb)); return (0); } /* * Intent log support: sync the block associated with db to disk. * N.B. and XXX: the caller is responsible for making sure that the * data isn't changing while dmu_sync() is writing it. * * Return values: * * EEXIST: this txg has already been synced, so there's nothing to do. * The caller should not log the write. * * ENOENT: the block was dbuf_free_range()'d, so there's nothing to do. * The caller should not log the write. * * EALREADY: this block is already in the process of being synced. * The caller should track its progress (somehow). * * EIO: could not do the I/O. * The caller should do a txg_wait_synced(). * * 0: the I/O has been initiated. * The caller should log this blkptr in the done callback. * It is possible that the I/O will fail, in which case * the error will be reported to the done callback and * propagated to pio from zio_done(). */ int dmu_sync(zio_t *pio, uint64_t txg, dmu_sync_cb_t *done, zgd_t *zgd) { blkptr_t *bp = zgd->zgd_bp; dmu_buf_impl_t *db = (dmu_buf_impl_t *)zgd->zgd_db; objset_t *os = db->db_objset; dsl_dataset_t *ds = os->os_dsl_dataset; dbuf_dirty_record_t *dr; dmu_sync_arg_t *dsa; zbookmark_phys_t zb; zio_prop_t zp; dnode_t *dn; ASSERT(pio != NULL); ASSERT(txg != 0); SET_BOOKMARK(&zb, ds->ds_object, db->db.db_object, db->db_level, db->db_blkid); DB_DNODE_ENTER(db); dn = DB_DNODE(db); dmu_write_policy(os, dn, db->db_level, WP_DMU_SYNC, &zp); DB_DNODE_EXIT(db); /* * If we're frozen (running ziltest), we always need to generate a bp. */ if (txg > spa_freeze_txg(os->os_spa)) return (dmu_sync_late_arrival(pio, os, done, zgd, &zp, &zb)); /* * Grabbing db_mtx now provides a barrier between dbuf_sync_leaf() * and us. If we determine that this txg is not yet syncing, * but it begins to sync a moment later, that's OK because the * sync thread will block in dbuf_sync_leaf() until we drop db_mtx. */ mutex_enter(&db->db_mtx); if (txg <= spa_last_synced_txg(os->os_spa)) { /* * This txg has already synced. There's nothing to do. */ mutex_exit(&db->db_mtx); return (SET_ERROR(EEXIST)); } if (txg <= spa_syncing_txg(os->os_spa)) { /* * This txg is currently syncing, so we can't mess with * the dirty record anymore; just write a new log block. */ mutex_exit(&db->db_mtx); return (dmu_sync_late_arrival(pio, os, done, zgd, &zp, &zb)); } dr = db->db_last_dirty; while (dr && dr->dr_txg != txg) dr = dr->dr_next; if (dr == NULL) { /* * There's no dr for this dbuf, so it must have been freed. * There's no need to log writes to freed blocks, so we're done. */ mutex_exit(&db->db_mtx); return (SET_ERROR(ENOENT)); } ASSERT(dr->dr_next == NULL || dr->dr_next->dr_txg < txg); /* * Assume the on-disk data is X, the current syncing data (in * txg - 1) is Y, and the current in-memory data is Z (currently * in dmu_sync). * * We usually want to perform a nopwrite if X and Z are the * same. However, if Y is different (i.e. the BP is going to * change before this write takes effect), then a nopwrite will * be incorrect - we would override with X, which could have * been freed when Y was written. * * (Note that this is not a concern when we are nop-writing from * syncing context, because X and Y must be identical, because * all previous txgs have been synced.) * * Therefore, we disable nopwrite if the current BP could change * before this TXG. There are two ways it could change: by * being dirty (dr_next is non-NULL), or by being freed * (dnode_block_freed()). This behavior is verified by * zio_done(), which VERIFYs that the override BP is identical * to the on-disk BP. */ DB_DNODE_ENTER(db); dn = DB_DNODE(db); if (dr->dr_next != NULL || dnode_block_freed(dn, db->db_blkid)) zp.zp_nopwrite = B_FALSE; DB_DNODE_EXIT(db); ASSERT(dr->dr_txg == txg); if (dr->dt.dl.dr_override_state == DR_IN_DMU_SYNC || dr->dt.dl.dr_override_state == DR_OVERRIDDEN) { /* * We have already issued a sync write for this buffer, * or this buffer has already been synced. It could not * have been dirtied since, or we would have cleared the state. */ mutex_exit(&db->db_mtx); return (SET_ERROR(EALREADY)); } ASSERT(dr->dt.dl.dr_override_state == DR_NOT_OVERRIDDEN); dr->dt.dl.dr_override_state = DR_IN_DMU_SYNC; mutex_exit(&db->db_mtx); dsa = kmem_alloc(sizeof (dmu_sync_arg_t), KM_SLEEP); dsa->dsa_dr = dr; dsa->dsa_done = done; dsa->dsa_zgd = zgd; dsa->dsa_tx = NULL; zio_nowait(arc_write(pio, os->os_spa, txg, bp, dr->dt.dl.dr_data, DBUF_IS_L2CACHEABLE(db), DBUF_IS_L2COMPRESSIBLE(db), &zp, dmu_sync_ready, NULL, dmu_sync_done, dsa, ZIO_PRIORITY_SYNC_WRITE, ZIO_FLAG_CANFAIL, &zb)); return (0); } int dmu_object_set_blocksize(objset_t *os, uint64_t object, uint64_t size, int ibs, dmu_tx_t *tx) { dnode_t *dn; int err; err = dnode_hold(os, object, FTAG, &dn); if (err) return (err); err = dnode_set_blksz(dn, size, ibs, tx); dnode_rele(dn, FTAG); return (err); } void dmu_object_set_checksum(objset_t *os, uint64_t object, uint8_t checksum, dmu_tx_t *tx) { dnode_t *dn; /* * Send streams include each object's checksum function. This * check ensures that the receiving system can understand the * checksum function transmitted. */ ASSERT3U(checksum, <, ZIO_CHECKSUM_LEGACY_FUNCTIONS); VERIFY0(dnode_hold(os, object, FTAG, &dn)); ASSERT3U(checksum, <, ZIO_CHECKSUM_FUNCTIONS); dn->dn_checksum = checksum; dnode_setdirty(dn, tx); dnode_rele(dn, FTAG); } void dmu_object_set_compress(objset_t *os, uint64_t object, uint8_t compress, dmu_tx_t *tx) { dnode_t *dn; /* * Send streams include each object's compression function. This * check ensures that the receiving system can understand the * compression function transmitted. */ ASSERT3U(compress, <, ZIO_COMPRESS_LEGACY_FUNCTIONS); VERIFY0(dnode_hold(os, object, FTAG, &dn)); dn->dn_compress = compress; dnode_setdirty(dn, tx); dnode_rele(dn, FTAG); } int zfs_mdcomp_disable = 0; /* * When the "redundant_metadata" property is set to "most", only indirect * blocks of this level and higher will have an additional ditto block. */ int zfs_redundant_metadata_most_ditto_level = 2; void dmu_write_policy(objset_t *os, dnode_t *dn, int level, int wp, zio_prop_t *zp) { dmu_object_type_t type = dn ? dn->dn_type : DMU_OT_OBJSET; boolean_t ismd = (level > 0 || DMU_OT_IS_METADATA(type) || (wp & WP_SPILL)); enum zio_checksum checksum = os->os_checksum; enum zio_compress compress = os->os_compress; enum zio_checksum dedup_checksum = os->os_dedup_checksum; boolean_t dedup = B_FALSE; boolean_t nopwrite = B_FALSE; boolean_t dedup_verify = os->os_dedup_verify; int copies = os->os_copies; /* * We maintain different write policies for each of the following * types of data: * 1. metadata * 2. preallocated blocks (i.e. level-0 blocks of a dump device) * 3. all other level 0 blocks */ if (ismd) { if (zfs_mdcomp_disable) { compress = ZIO_COMPRESS_EMPTY; } else { /* * XXX -- we should design a compression algorithm * that specializes in arrays of bps. */ compress = zio_compress_select(os->os_spa, ZIO_COMPRESS_ON, ZIO_COMPRESS_ON); } /* * Metadata always gets checksummed. If the data * checksum is multi-bit correctable, and it's not a * ZBT-style checksum, then it's suitable for metadata * as well. Otherwise, the metadata checksum defaults * to fletcher4. */ if (zio_checksum_table[checksum].ci_correctable < 1 || zio_checksum_table[checksum].ci_eck) checksum = ZIO_CHECKSUM_FLETCHER_4; if (os->os_redundant_metadata == ZFS_REDUNDANT_METADATA_ALL || (os->os_redundant_metadata == ZFS_REDUNDANT_METADATA_MOST && (level >= zfs_redundant_metadata_most_ditto_level || DMU_OT_IS_METADATA(type) || (wp & WP_SPILL)))) copies++; } else if (wp & WP_NOFILL) { ASSERT(level == 0); /* * If we're writing preallocated blocks, we aren't actually * writing them so don't set any policy properties. These * blocks are currently only used by an external subsystem * outside of zfs (i.e. dump) and not written by the zio * pipeline. */ compress = ZIO_COMPRESS_OFF; checksum = ZIO_CHECKSUM_OFF; } else { compress = zio_compress_select(os->os_spa, dn->dn_compress, compress); checksum = (dedup_checksum == ZIO_CHECKSUM_OFF) ? zio_checksum_select(dn->dn_checksum, checksum) : dedup_checksum; /* * Determine dedup setting. If we are in dmu_sync(), * we won't actually dedup now because that's all * done in syncing context; but we do want to use the * dedup checkum. If the checksum is not strong * enough to ensure unique signatures, force * dedup_verify. */ if (dedup_checksum != ZIO_CHECKSUM_OFF) { dedup = (wp & WP_DMU_SYNC) ? B_FALSE : B_TRUE; if (!zio_checksum_table[checksum].ci_dedup) dedup_verify = B_TRUE; } /* * Enable nopwrite if we have a cryptographically secure * checksum that has no known collisions (i.e. SHA-256) * and compression is enabled. We don't enable nopwrite if * dedup is enabled as the two features are mutually exclusive. */ nopwrite = (!dedup && zio_checksum_table[checksum].ci_dedup && compress != ZIO_COMPRESS_OFF && zfs_nopwrite_enabled); } zp->zp_checksum = checksum; zp->zp_compress = compress; zp->zp_type = (wp & WP_SPILL) ? dn->dn_bonustype : type; zp->zp_level = level; zp->zp_copies = MIN(copies, spa_max_replication(os->os_spa)); zp->zp_dedup = dedup; zp->zp_dedup_verify = dedup && dedup_verify; zp->zp_nopwrite = nopwrite; } int dmu_offset_next(objset_t *os, uint64_t object, boolean_t hole, uint64_t *off) { dnode_t *dn; int i, err; err = dnode_hold(os, object, FTAG, &dn); if (err) return (err); /* * Sync any current changes before * we go trundling through the block pointers. */ for (i = 0; i < TXG_SIZE; i++) { if (list_link_active(&dn->dn_dirty_link[i])) break; } if (i != TXG_SIZE) { dnode_rele(dn, FTAG); txg_wait_synced(dmu_objset_pool(os), 0); err = dnode_hold(os, object, FTAG, &dn); if (err) return (err); } err = dnode_next_offset(dn, (hole ? DNODE_FIND_HOLE : 0), off, 1, 1, 0); dnode_rele(dn, FTAG); return (err); } void __dmu_object_info_from_dnode(dnode_t *dn, dmu_object_info_t *doi) { dnode_phys_t *dnp = dn->dn_phys; int i; doi->doi_data_block_size = dn->dn_datablksz; doi->doi_metadata_block_size = dn->dn_indblkshift ? 1ULL << dn->dn_indblkshift : 0; doi->doi_type = dn->dn_type; doi->doi_bonus_type = dn->dn_bonustype; doi->doi_bonus_size = dn->dn_bonuslen; doi->doi_indirection = dn->dn_nlevels; doi->doi_checksum = dn->dn_checksum; doi->doi_compress = dn->dn_compress; doi->doi_nblkptr = dn->dn_nblkptr; doi->doi_physical_blocks_512 = (DN_USED_BYTES(dnp) + 256) >> 9; doi->doi_max_offset = (dn->dn_maxblkid + 1) * dn->dn_datablksz; doi->doi_fill_count = 0; for (i = 0; i < dnp->dn_nblkptr; i++) doi->doi_fill_count += BP_GET_FILL(&dnp->dn_blkptr[i]); } void dmu_object_info_from_dnode(dnode_t *dn, dmu_object_info_t *doi) { rw_enter(&dn->dn_struct_rwlock, RW_READER); mutex_enter(&dn->dn_mtx); __dmu_object_info_from_dnode(dn, doi); mutex_exit(&dn->dn_mtx); rw_exit(&dn->dn_struct_rwlock); } /* * Get information on a DMU object. * 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) { dnode_t *dn; int err = dnode_hold(os, object, FTAG, &dn); if (err) return (err); if (doi != NULL) dmu_object_info_from_dnode(dn, doi); dnode_rele(dn, FTAG); return (0); } /* * As above, but faster; can be used when you have a held dbuf in hand. */ void dmu_object_info_from_db(dmu_buf_t *db_fake, dmu_object_info_t *doi) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake; DB_DNODE_ENTER(db); dmu_object_info_from_dnode(DB_DNODE(db), doi); DB_DNODE_EXIT(db); } /* * Faster still when you only care about the size. * This is specifically optimized for zfs_getattr(). */ void dmu_object_size_from_db(dmu_buf_t *db_fake, uint32_t *blksize, u_longlong_t *nblk512) { dmu_buf_impl_t *db = (dmu_buf_impl_t *)db_fake; dnode_t *dn; DB_DNODE_ENTER(db); dn = DB_DNODE(db); *blksize = dn->dn_datablksz; /* add 1 for dnode space */ *nblk512 = ((DN_USED_BYTES(dn->dn_phys) + SPA_MINBLOCKSIZE/2) >> SPA_MINBLOCKSHIFT) + 1; DB_DNODE_EXIT(db); } void byteswap_uint64_array(void *vbuf, size_t size) { uint64_t *buf = vbuf; size_t count = size >> 3; int i; ASSERT((size & 7) == 0); for (i = 0; i < count; i++) buf[i] = BSWAP_64(buf[i]); } void byteswap_uint32_array(void *vbuf, size_t size) { uint32_t *buf = vbuf; size_t count = size >> 2; int i; ASSERT((size & 3) == 0); for (i = 0; i < count; i++) buf[i] = BSWAP_32(buf[i]); } void byteswap_uint16_array(void *vbuf, size_t size) { uint16_t *buf = vbuf; size_t count = size >> 1; int i; ASSERT((size & 1) == 0); for (i = 0; i < count; i++) buf[i] = BSWAP_16(buf[i]); } /* ARGSUSED */ void byteswap_uint8_array(void *vbuf, size_t size) { } void dmu_init(void) { zfs_dbgmsg_init(); sa_cache_init(); xuio_stat_init(); dmu_objset_init(); dnode_init(); dbuf_init(); zfetch_init(); dmu_tx_init(); l2arc_init(); arc_init(); } void dmu_fini(void) { arc_fini(); /* arc depends on l2arc, so arc must go first */ l2arc_fini(); dmu_tx_fini(); zfetch_fini(); dbuf_fini(); dnode_fini(); dmu_objset_fini(); xuio_stat_fini(); sa_cache_fini(); zfs_dbgmsg_fini(); } #if defined(_KERNEL) && defined(HAVE_SPL) EXPORT_SYMBOL(dmu_bonus_hold); EXPORT_SYMBOL(dmu_buf_hold_array_by_bonus); EXPORT_SYMBOL(dmu_buf_rele_array); EXPORT_SYMBOL(dmu_prefetch); EXPORT_SYMBOL(dmu_free_range); EXPORT_SYMBOL(dmu_free_long_range); EXPORT_SYMBOL(dmu_free_long_object); EXPORT_SYMBOL(dmu_read); EXPORT_SYMBOL(dmu_write); EXPORT_SYMBOL(dmu_prealloc); EXPORT_SYMBOL(dmu_object_info); EXPORT_SYMBOL(dmu_object_info_from_dnode); EXPORT_SYMBOL(dmu_object_info_from_db); EXPORT_SYMBOL(dmu_object_size_from_db); EXPORT_SYMBOL(dmu_object_set_blocksize); EXPORT_SYMBOL(dmu_object_set_checksum); EXPORT_SYMBOL(dmu_object_set_compress); EXPORT_SYMBOL(dmu_write_policy); EXPORT_SYMBOL(dmu_sync); EXPORT_SYMBOL(dmu_request_arcbuf); EXPORT_SYMBOL(dmu_return_arcbuf); EXPORT_SYMBOL(dmu_assign_arcbuf); EXPORT_SYMBOL(dmu_buf_hold); EXPORT_SYMBOL(dmu_ot); module_param(zfs_mdcomp_disable, int, 0644); MODULE_PARM_DESC(zfs_mdcomp_disable, "Disable meta data compression"); module_param(zfs_nopwrite_enabled, int, 0644); MODULE_PARM_DESC(zfs_nopwrite_enabled, "Enable NOP writes"); #endif diff --git a/module/zfs/dmu_zfetch.c b/module/zfs/dmu_zfetch.c index 9d41bdc0f132..b25b0e00772e 100644 --- a/module/zfs/dmu_zfetch.c +++ b/module/zfs/dmu_zfetch.c @@ -1,749 +1,306 @@ /* * 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 2009 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ /* - * Copyright (c) 2013 by Delphix. All rights reserved. + * Copyright (c) 2013, 2014 by Delphix. All rights reserved. */ #include #include #include #include #include #include #include /* - * I'm against tune-ables, but these should probably exist as tweakable globals - * until we can get this working the way we want it to. + * This tunable disables predictive prefetch. Note that it leaves "prescient" + * prefetch (e.g. prefetch for zfs send) intact. Unlike predictive prefetch, + * prescient prefetch never issues i/os that end up not being needed, + * so it can't hurt performance. */ -int zfs_prefetch_disable = 0; +int zfs_prefetch_disable = B_FALSE; /* max # of streams per zfetch */ unsigned int zfetch_max_streams = 8; /* min time before stream reclaim */ unsigned int zfetch_min_sec_reap = 2; -/* max number of blocks to fetch at a time */ -unsigned int zfetch_block_cap = 256; -/* number of bytes in a array_read at which we stop prefetching (1Mb) */ +/* max bytes to prefetch per stream (default 8MB) */ +unsigned int zfetch_max_distance = 8 * 1024 * 1024; +/* number of bytes in a array_read at which we stop prefetching (1MB) */ unsigned long zfetch_array_rd_sz = 1024 * 1024; -/* forward decls for static routines */ -static boolean_t dmu_zfetch_colinear(zfetch_t *, zstream_t *); -static void dmu_zfetch_dofetch(zfetch_t *, zstream_t *); -static uint64_t dmu_zfetch_fetch(dnode_t *, uint64_t, uint64_t); -static uint64_t dmu_zfetch_fetchsz(dnode_t *, uint64_t, uint64_t); -static boolean_t dmu_zfetch_find(zfetch_t *, zstream_t *, int); -static int dmu_zfetch_stream_insert(zfetch_t *, zstream_t *); -static zstream_t *dmu_zfetch_stream_reclaim(zfetch_t *); -static void dmu_zfetch_stream_remove(zfetch_t *, zstream_t *); -static int dmu_zfetch_streams_equal(zstream_t *, zstream_t *); - typedef struct zfetch_stats { kstat_named_t zfetchstat_hits; kstat_named_t zfetchstat_misses; - kstat_named_t zfetchstat_colinear_hits; - kstat_named_t zfetchstat_colinear_misses; - kstat_named_t zfetchstat_stride_hits; - kstat_named_t zfetchstat_stride_misses; - kstat_named_t zfetchstat_reclaim_successes; - kstat_named_t zfetchstat_reclaim_failures; - kstat_named_t zfetchstat_stream_resets; - kstat_named_t zfetchstat_stream_noresets; - kstat_named_t zfetchstat_bogus_streams; + kstat_named_t zfetchstat_max_streams; } zfetch_stats_t; static zfetch_stats_t zfetch_stats = { { "hits", KSTAT_DATA_UINT64 }, { "misses", KSTAT_DATA_UINT64 }, - { "colinear_hits", KSTAT_DATA_UINT64 }, - { "colinear_misses", KSTAT_DATA_UINT64 }, - { "stride_hits", KSTAT_DATA_UINT64 }, - { "stride_misses", KSTAT_DATA_UINT64 }, - { "reclaim_successes", KSTAT_DATA_UINT64 }, - { "reclaim_failures", KSTAT_DATA_UINT64 }, - { "streams_resets", KSTAT_DATA_UINT64 }, - { "streams_noresets", KSTAT_DATA_UINT64 }, - { "bogus_streams", KSTAT_DATA_UINT64 }, + { "max_streams", KSTAT_DATA_UINT64 }, }; -#define ZFETCHSTAT_INCR(stat, val) \ - atomic_add_64(&zfetch_stats.stat.value.ui64, (val)); - -#define ZFETCHSTAT_BUMP(stat) ZFETCHSTAT_INCR(stat, 1); +#define ZFETCHSTAT_BUMP(stat) \ + atomic_inc_64(&zfetch_stats.stat.value.ui64); kstat_t *zfetch_ksp; -/* - * Given a zfetch structure and a zstream structure, determine whether the - * blocks to be read are part of a co-linear pair of existing prefetch - * streams. If a set is found, coalesce the streams, removing one, and - * configure the prefetch so it looks for a strided access pattern. - * - * In other words: if we find two sequential access streams that are - * the same length and distance N appart, and this read is N from the - * last stream, then we are probably in a strided access pattern. So - * combine the two sequential streams into a single strided stream. - * - * Returns whether co-linear streams were found. - */ -static boolean_t -dmu_zfetch_colinear(zfetch_t *zf, zstream_t *zh) -{ - zstream_t *z_walk; - zstream_t *z_comp; - - if (! rw_tryenter(&zf->zf_rwlock, RW_WRITER)) - return (0); - - if (zh == NULL) { - rw_exit(&zf->zf_rwlock); - return (0); - } - - for (z_walk = list_head(&zf->zf_stream); z_walk; - z_walk = list_next(&zf->zf_stream, z_walk)) { - for (z_comp = list_next(&zf->zf_stream, z_walk); z_comp; - z_comp = list_next(&zf->zf_stream, z_comp)) { - int64_t diff; - - if (z_walk->zst_len != z_walk->zst_stride || - z_comp->zst_len != z_comp->zst_stride) { - continue; - } - - diff = z_comp->zst_offset - z_walk->zst_offset; - if (z_comp->zst_offset + diff == zh->zst_offset) { - z_walk->zst_offset = zh->zst_offset; - z_walk->zst_direction = diff < 0 ? - ZFETCH_BACKWARD : ZFETCH_FORWARD; - z_walk->zst_stride = - diff * z_walk->zst_direction; - z_walk->zst_ph_offset = - zh->zst_offset + z_walk->zst_stride; - dmu_zfetch_stream_remove(zf, z_comp); - mutex_destroy(&z_comp->zst_lock); - kmem_free(z_comp, sizeof (zstream_t)); - - dmu_zfetch_dofetch(zf, z_walk); - - rw_exit(&zf->zf_rwlock); - return (1); - } - - diff = z_walk->zst_offset - z_comp->zst_offset; - if (z_walk->zst_offset + diff == zh->zst_offset) { - z_walk->zst_offset = zh->zst_offset; - z_walk->zst_direction = diff < 0 ? - ZFETCH_BACKWARD : ZFETCH_FORWARD; - z_walk->zst_stride = - diff * z_walk->zst_direction; - z_walk->zst_ph_offset = - zh->zst_offset + z_walk->zst_stride; - dmu_zfetch_stream_remove(zf, z_comp); - mutex_destroy(&z_comp->zst_lock); - kmem_free(z_comp, sizeof (zstream_t)); - - dmu_zfetch_dofetch(zf, z_walk); - - rw_exit(&zf->zf_rwlock); - return (1); - } - } - } - - rw_exit(&zf->zf_rwlock); - return (0); -} - -/* - * Given a zstream_t, determine the bounds of the prefetch. Then call the - * routine that actually prefetches the individual blocks. - */ -static void -dmu_zfetch_dofetch(zfetch_t *zf, zstream_t *zs) -{ - uint64_t prefetch_tail; - uint64_t prefetch_limit; - uint64_t prefetch_ofst; - uint64_t prefetch_len; - uint64_t blocks_fetched; - - zs->zst_stride = MAX((int64_t)zs->zst_stride, zs->zst_len); - zs->zst_cap = MIN(zfetch_block_cap, 2 * zs->zst_cap); - - prefetch_tail = MAX((int64_t)zs->zst_ph_offset, - (int64_t)(zs->zst_offset + zs->zst_stride)); - /* - * XXX: use a faster division method? - */ - prefetch_limit = zs->zst_offset + zs->zst_len + - (zs->zst_cap * zs->zst_stride) / zs->zst_len; - - while (prefetch_tail < prefetch_limit) { - prefetch_ofst = zs->zst_offset + zs->zst_direction * - (prefetch_tail - zs->zst_offset); - - prefetch_len = zs->zst_len; - - /* - * Don't prefetch beyond the end of the file, if working - * backwards. - */ - if ((zs->zst_direction == ZFETCH_BACKWARD) && - (prefetch_ofst > prefetch_tail)) { - prefetch_len += prefetch_ofst; - prefetch_ofst = 0; - } - - /* don't prefetch more than we're supposed to */ - if (prefetch_len > zs->zst_len) - break; - - blocks_fetched = dmu_zfetch_fetch(zf->zf_dnode, - prefetch_ofst, zs->zst_len); - - prefetch_tail += zs->zst_stride; - /* stop if we've run out of stuff to prefetch */ - if (blocks_fetched < zs->zst_len) - break; - } - zs->zst_ph_offset = prefetch_tail; - zs->zst_last = ddi_get_lbolt(); -} - void zfetch_init(void) { - zfetch_ksp = kstat_create("zfs", 0, "zfetchstats", "misc", KSTAT_TYPE_NAMED, sizeof (zfetch_stats) / sizeof (kstat_named_t), KSTAT_FLAG_VIRTUAL); if (zfetch_ksp != NULL) { zfetch_ksp->ks_data = &zfetch_stats; kstat_install(zfetch_ksp); } } void zfetch_fini(void) { if (zfetch_ksp != NULL) { kstat_delete(zfetch_ksp); zfetch_ksp = NULL; } } /* * This takes a pointer to a zfetch structure and a dnode. It performs the * necessary setup for the zfetch structure, grokking data from the * associated dnode. */ void dmu_zfetch_init(zfetch_t *zf, dnode_t *dno) { - if (zf == NULL) { + if (zf == NULL) return; - } zf->zf_dnode = dno; - zf->zf_stream_cnt = 0; - zf->zf_alloc_fail = 0; list_create(&zf->zf_stream, sizeof (zstream_t), - offsetof(zstream_t, zst_node)); + offsetof(zstream_t, zs_node)); rw_init(&zf->zf_rwlock, NULL, RW_DEFAULT, NULL); } -/* - * This function computes the actual size, in blocks, that can be prefetched, - * and fetches it. - */ -static uint64_t -dmu_zfetch_fetch(dnode_t *dn, uint64_t blkid, uint64_t nblks) -{ - uint64_t fetchsz; - uint64_t i; - - fetchsz = dmu_zfetch_fetchsz(dn, blkid, nblks); - - for (i = 0; i < fetchsz; i++) { - dbuf_prefetch(dn, 0, blkid + i, ZIO_PRIORITY_ASYNC_READ, - ARC_FLAG_PREFETCH); - } - - return (fetchsz); -} - -/* - * this function returns the number of blocks that would be prefetched, based - * upon the supplied dnode, blockid, and nblks. This is used so that we can - * update streams in place, and then prefetch with their old value after the - * fact. This way, we can delay the prefetch, but subsequent accesses to the - * stream won't result in the same data being prefetched multiple times. - */ -static uint64_t -dmu_zfetch_fetchsz(dnode_t *dn, uint64_t blkid, uint64_t nblks) -{ - uint64_t fetchsz; - - if (blkid > dn->dn_maxblkid) { - return (0); - } - - /* compute fetch size */ - if (blkid + nblks + 1 > dn->dn_maxblkid) { - fetchsz = (dn->dn_maxblkid - blkid) + 1; - ASSERT(blkid + fetchsz - 1 <= dn->dn_maxblkid); - } else { - fetchsz = nblks; - } - - - return (fetchsz); -} - -/* - * given a zfetch and a zstream structure, see if there is an associated zstream - * for this block read. If so, it starts a prefetch for the stream it - * located and returns true, otherwise it returns false - */ -static boolean_t -dmu_zfetch_find(zfetch_t *zf, zstream_t *zh, int prefetched) +static void +dmu_zfetch_stream_remove(zfetch_t *zf, zstream_t *zs) { - zstream_t *zs; - int64_t diff; - int reset = !prefetched; - int rc = 0; - - if (zh == NULL) - return (0); - - /* - * XXX: This locking strategy is a bit coarse; however, it's impact has - * yet to be tested. If this turns out to be an issue, it can be - * modified in a number of different ways. - */ - - rw_enter(&zf->zf_rwlock, RW_READER); -top: - - for (zs = list_head(&zf->zf_stream); zs; - zs = list_next(&zf->zf_stream, zs)) { - - /* - * XXX - should this be an assert? - */ - if (zs->zst_len == 0) { - /* bogus stream */ - ZFETCHSTAT_BUMP(zfetchstat_bogus_streams); - continue; - } - - /* - * We hit this case when we are in a strided prefetch stream: - * we will read "len" blocks before "striding". - */ - if (zh->zst_offset >= zs->zst_offset && - zh->zst_offset < zs->zst_offset + zs->zst_len) { - if (prefetched) { - /* already fetched */ - ZFETCHSTAT_BUMP(zfetchstat_stride_hits); - rc = 1; - goto out; - } else { - ZFETCHSTAT_BUMP(zfetchstat_stride_misses); - } - } - - /* - * This is the forward sequential read case: we increment - * len by one each time we hit here, so we will enter this - * case on every read. - */ - if (zh->zst_offset == zs->zst_offset + zs->zst_len) { - - reset = !prefetched && zs->zst_len > 1; - - mutex_enter(&zs->zst_lock); - - if (zh->zst_offset != zs->zst_offset + zs->zst_len) { - mutex_exit(&zs->zst_lock); - goto top; - } - zs->zst_len += zh->zst_len; - diff = zs->zst_len - zfetch_block_cap; - if (diff > 0) { - zs->zst_offset += diff; - zs->zst_len = zs->zst_len > diff ? - zs->zst_len - diff : 0; - } - zs->zst_direction = ZFETCH_FORWARD; - - break; - - /* - * Same as above, but reading backwards through the file. - */ - } else if (zh->zst_offset == zs->zst_offset - zh->zst_len) { - /* backwards sequential access */ - - reset = !prefetched && zs->zst_len > 1; - - mutex_enter(&zs->zst_lock); - - if (zh->zst_offset != zs->zst_offset - zh->zst_len) { - mutex_exit(&zs->zst_lock); - goto top; - } - - zs->zst_offset = zs->zst_offset > zh->zst_len ? - zs->zst_offset - zh->zst_len : 0; - zs->zst_ph_offset = zs->zst_ph_offset > zh->zst_len ? - zs->zst_ph_offset - zh->zst_len : 0; - zs->zst_len += zh->zst_len; - - diff = zs->zst_len - zfetch_block_cap; - if (diff > 0) { - zs->zst_ph_offset = zs->zst_ph_offset > diff ? - zs->zst_ph_offset - diff : 0; - zs->zst_len = zs->zst_len > diff ? - zs->zst_len - diff : zs->zst_len; - } - zs->zst_direction = ZFETCH_BACKWARD; - - break; - - } else if ((zh->zst_offset - zs->zst_offset - zs->zst_stride < - zs->zst_len) && (zs->zst_len != zs->zst_stride)) { - /* strided forward access */ - - mutex_enter(&zs->zst_lock); - - if ((zh->zst_offset - zs->zst_offset - zs->zst_stride >= - zs->zst_len) || (zs->zst_len == zs->zst_stride)) { - mutex_exit(&zs->zst_lock); - goto top; - } - - zs->zst_offset += zs->zst_stride; - zs->zst_direction = ZFETCH_FORWARD; - - break; - - } else if ((zh->zst_offset - zs->zst_offset + zs->zst_stride < - zs->zst_len) && (zs->zst_len != zs->zst_stride)) { - /* strided reverse access */ - - mutex_enter(&zs->zst_lock); - - if ((zh->zst_offset - zs->zst_offset + zs->zst_stride >= - zs->zst_len) || (zs->zst_len == zs->zst_stride)) { - mutex_exit(&zs->zst_lock); - goto top; - } - - zs->zst_offset = zs->zst_offset > zs->zst_stride ? - zs->zst_offset - zs->zst_stride : 0; - zs->zst_ph_offset = (zs->zst_ph_offset > - (2 * zs->zst_stride)) ? - (zs->zst_ph_offset - (2 * zs->zst_stride)) : 0; - zs->zst_direction = ZFETCH_BACKWARD; - - break; - } - } - - if (zs) { - if (reset) { - zstream_t *remove = zs; - - ZFETCHSTAT_BUMP(zfetchstat_stream_resets); - rc = 0; - mutex_exit(&zs->zst_lock); - rw_exit(&zf->zf_rwlock); - rw_enter(&zf->zf_rwlock, RW_WRITER); - /* - * Relocate the stream, in case someone removes - * it while we were acquiring the WRITER lock. - */ - for (zs = list_head(&zf->zf_stream); zs; - zs = list_next(&zf->zf_stream, zs)) { - if (zs == remove) { - dmu_zfetch_stream_remove(zf, zs); - mutex_destroy(&zs->zst_lock); - kmem_free(zs, sizeof (zstream_t)); - break; - } - } - } else { - ZFETCHSTAT_BUMP(zfetchstat_stream_noresets); - rc = 1; - dmu_zfetch_dofetch(zf, zs); - mutex_exit(&zs->zst_lock); - } - } -out: - rw_exit(&zf->zf_rwlock); - return (rc); + ASSERT(RW_WRITE_HELD(&zf->zf_rwlock)); + list_remove(&zf->zf_stream, zs); + mutex_destroy(&zs->zs_lock); + kmem_free(zs, sizeof (*zs)); } /* - * Clean-up state associated with a zfetch structure. This frees allocated - * structure members, empties the zf_stream tree, and generally makes things - * nice. This doesn't free the zfetch_t itself, that's left to the caller. + * Clean-up state associated with a zfetch structure (e.g. destroy the + * streams). This doesn't free the zfetch_t itself, that's left to the caller. */ void -dmu_zfetch_rele(zfetch_t *zf) +dmu_zfetch_fini(zfetch_t *zf) { - zstream_t *zs; - zstream_t *zs_next; + zstream_t *zs; ASSERT(!RW_LOCK_HELD(&zf->zf_rwlock)); - for (zs = list_head(&zf->zf_stream); zs; zs = zs_next) { - zs_next = list_next(&zf->zf_stream, zs); - - list_remove(&zf->zf_stream, zs); - mutex_destroy(&zs->zst_lock); - kmem_free(zs, sizeof (zstream_t)); - } + rw_enter(&zf->zf_rwlock, RW_WRITER); + while ((zs = list_head(&zf->zf_stream)) != NULL) + dmu_zfetch_stream_remove(zf, zs); + rw_exit(&zf->zf_rwlock); list_destroy(&zf->zf_stream); rw_destroy(&zf->zf_rwlock); zf->zf_dnode = NULL; } /* - * Given a zfetch and zstream structure, insert the zstream structure into the - * AVL tree contained within the zfetch structure. Peform the appropriate - * book-keeping. It is possible that another thread has inserted a stream which - * matches one that we are about to insert, so we must be sure to check for this - * case. If one is found, return failure, and let the caller cleanup the - * duplicates. + * If there aren't too many streams already, create a new stream. + * The "blkid" argument is the next block that we expect this stream to access. + * While we're here, clean up old streams (which haven't been + * accessed for at least zfetch_min_sec_reap seconds). */ -static int -dmu_zfetch_stream_insert(zfetch_t *zf, zstream_t *zs) +static void +dmu_zfetch_stream_create(zfetch_t *zf, uint64_t blkid) { - zstream_t *zs_walk; - zstream_t *zs_next; + zstream_t *zs; + zstream_t *zs_next; + int numstreams = 0; + uint32_t max_streams; ASSERT(RW_WRITE_HELD(&zf->zf_rwlock)); - for (zs_walk = list_head(&zf->zf_stream); zs_walk; zs_walk = zs_next) { - zs_next = list_next(&zf->zf_stream, zs_walk); - - if (dmu_zfetch_streams_equal(zs_walk, zs)) { - return (0); - } - } - - list_insert_head(&zf->zf_stream, zs); - zf->zf_stream_cnt++; - return (1); -} - - -/* - * Walk the list of zstreams in the given zfetch, find an old one (by time), and - * reclaim it for use by the caller. - */ -static zstream_t * -dmu_zfetch_stream_reclaim(zfetch_t *zf) -{ - zstream_t *zs; - - if (! rw_tryenter(&zf->zf_rwlock, RW_WRITER)) - return (0); - - for (zs = list_head(&zf->zf_stream); zs; - zs = list_next(&zf->zf_stream, zs)) { - - if (((ddi_get_lbolt() - zs->zst_last)/hz) > zfetch_min_sec_reap) - break; + /* + * Clean up old streams. + */ + for (zs = list_head(&zf->zf_stream); + zs != NULL; zs = zs_next) { + zs_next = list_next(&zf->zf_stream, zs); + if (((gethrtime() - zs->zs_atime) / NANOSEC) > + zfetch_min_sec_reap) + dmu_zfetch_stream_remove(zf, zs); + else + numstreams++; } - if (zs) { - dmu_zfetch_stream_remove(zf, zs); - mutex_destroy(&zs->zst_lock); - bzero(zs, sizeof (zstream_t)); - } else { - zf->zf_alloc_fail++; + /* + * The maximum number of streams is normally zfetch_max_streams, + * but for small files we lower it such that it's at least possible + * for all the streams to be non-overlapping. + * + * If we are already at the maximum number of streams for this file, + * even after removing old streams, then don't create this stream. + */ + max_streams = MAX(1, MIN(zfetch_max_streams, + zf->zf_dnode->dn_maxblkid * zf->zf_dnode->dn_datablksz / + zfetch_max_distance)); + if (numstreams >= max_streams) { + ZFETCHSTAT_BUMP(zfetchstat_max_streams); + return; } - rw_exit(&zf->zf_rwlock); - - return (zs); -} -/* - * Given a zfetch and zstream structure, remove the zstream structure from its - * container in the zfetch structure. Perform the appropriate book-keeping. - */ -static void -dmu_zfetch_stream_remove(zfetch_t *zf, zstream_t *zs) -{ - ASSERT(RW_WRITE_HELD(&zf->zf_rwlock)); + zs = kmem_zalloc(sizeof (*zs), KM_SLEEP); + zs->zs_blkid = blkid; + zs->zs_pf_blkid = blkid; + zs->zs_atime = gethrtime(); + mutex_init(&zs->zs_lock, NULL, MUTEX_DEFAULT, NULL); - list_remove(&zf->zf_stream, zs); - zf->zf_stream_cnt--; -} - -static int -dmu_zfetch_streams_equal(zstream_t *zs1, zstream_t *zs2) -{ - if (zs1->zst_offset != zs2->zst_offset) - return (0); - - if (zs1->zst_len != zs2->zst_len) - return (0); - - if (zs1->zst_stride != zs2->zst_stride) - return (0); - - if (zs1->zst_ph_offset != zs2->zst_ph_offset) - return (0); - - if (zs1->zst_cap != zs2->zst_cap) - return (0); - - if (zs1->zst_direction != zs2->zst_direction) - return (0); - - return (1); + list_insert_head(&zf->zf_stream, zs); } /* * This is the prefetch entry point. It calls all of the other dmu_zfetch * routines to create, delete, find, or operate upon prefetch streams. */ void -dmu_zfetch(zfetch_t *zf, uint64_t offset, uint64_t size, int prefetched) +dmu_zfetch(zfetch_t *zf, uint64_t blkid, uint64_t nblks) { - zstream_t zst; - zstream_t *newstream; - boolean_t fetched; - int inserted; - unsigned int blkshft; - uint64_t blksz; + zstream_t *zs; + int64_t pf_start; + int pf_nblks; + int i; if (zfs_prefetch_disable) return; - /* files that aren't ln2 blocksz are only one block -- nothing to do */ - if (!zf->zf_dnode->dn_datablkshift) + /* + * As a fast path for small (single-block) files, ignore access + * to the first block. + */ + if (blkid == 0) return; - /* convert offset and size, into blockid and nblocks */ - blkshft = zf->zf_dnode->dn_datablkshift; - blksz = (1 << blkshft); - - bzero(&zst, sizeof (zstream_t)); - zst.zst_offset = offset >> blkshft; - zst.zst_len = (P2ROUNDUP(offset + size, blksz) - - P2ALIGN(offset, blksz)) >> blkshft; + rw_enter(&zf->zf_rwlock, RW_READER); - fetched = dmu_zfetch_find(zf, &zst, prefetched); - if (fetched) { - ZFETCHSTAT_BUMP(zfetchstat_hits); - } else { - ZFETCHSTAT_BUMP(zfetchstat_misses); - if ((fetched = dmu_zfetch_colinear(zf, &zst))) { - ZFETCHSTAT_BUMP(zfetchstat_colinear_hits); - } else { - ZFETCHSTAT_BUMP(zfetchstat_colinear_misses); + for (zs = list_head(&zf->zf_stream); zs != NULL; + zs = list_next(&zf->zf_stream, zs)) { + if (blkid == zs->zs_blkid) { + mutex_enter(&zs->zs_lock); + /* + * zs_blkid could have changed before we + * acquired zs_lock; re-check them here. + */ + if (blkid != zs->zs_blkid) { + mutex_exit(&zs->zs_lock); + continue; + } + break; } } - if (!fetched) { - newstream = dmu_zfetch_stream_reclaim(zf); - + if (zs == NULL) { /* - * we still couldn't find a stream, drop the lock, and allocate - * one if possible. Otherwise, give up and go home. + * This access is not part of any existing stream. Create + * a new stream for it. */ - if (newstream) { - ZFETCHSTAT_BUMP(zfetchstat_reclaim_successes); - } else { - uint64_t maxblocks; - uint32_t max_streams; - uint32_t cur_streams; - - ZFETCHSTAT_BUMP(zfetchstat_reclaim_failures); - cur_streams = zf->zf_stream_cnt; - maxblocks = zf->zf_dnode->dn_maxblkid; - - max_streams = MIN(zfetch_max_streams, - (maxblocks / zfetch_block_cap)); - if (max_streams == 0) { - max_streams++; - } - - if (cur_streams >= max_streams) { - return; - } - newstream = - kmem_zalloc(sizeof (zstream_t), KM_SLEEP); - } + ZFETCHSTAT_BUMP(zfetchstat_misses); + if (rw_tryupgrade(&zf->zf_rwlock)) + dmu_zfetch_stream_create(zf, blkid + nblks); + rw_exit(&zf->zf_rwlock); + return; + } - newstream->zst_offset = zst.zst_offset; - newstream->zst_len = zst.zst_len; - newstream->zst_stride = zst.zst_len; - newstream->zst_ph_offset = zst.zst_len + zst.zst_offset; - newstream->zst_cap = zst.zst_len; - newstream->zst_direction = ZFETCH_FORWARD; - newstream->zst_last = ddi_get_lbolt(); + /* + * This access was to a block that we issued a prefetch for on + * behalf of this stream. Issue further prefetches for this stream. + * + * Normally, we start prefetching where we stopped + * prefetching last (zs_pf_blkid). But when we get our first + * hit on this stream, zs_pf_blkid == zs_blkid, we don't + * want to prefetch to block we just accessed. In this case, + * start just after the block we just accessed. + */ + pf_start = MAX(zs->zs_pf_blkid, blkid + nblks); - mutex_init(&newstream->zst_lock, NULL, MUTEX_DEFAULT, NULL); + /* + * Double our amount of prefetched data, but don't let the + * prefetch get further ahead than zfetch_max_distance. + */ + pf_nblks = + MIN((int64_t)zs->zs_pf_blkid - zs->zs_blkid + nblks, + zs->zs_blkid + nblks + + (zfetch_max_distance >> zf->zf_dnode->dn_datablkshift) - pf_start); - rw_enter(&zf->zf_rwlock, RW_WRITER); - inserted = dmu_zfetch_stream_insert(zf, newstream); - rw_exit(&zf->zf_rwlock); + zs->zs_pf_blkid = pf_start + pf_nblks; + zs->zs_atime = gethrtime(); + zs->zs_blkid = blkid + nblks; - if (!inserted) { - mutex_destroy(&newstream->zst_lock); - kmem_free(newstream, sizeof (zstream_t)); - } + /* + * dbuf_prefetch() issues the prefetch i/o + * asynchronously, but it may need to wait for an + * indirect block to be read from disk. Therefore + * we do not want to hold any locks while we call it. + */ + mutex_exit(&zs->zs_lock); + rw_exit(&zf->zf_rwlock); + for (i = 0; i < pf_nblks; i++) { + dbuf_prefetch(zf->zf_dnode, 0, pf_start + i, + ZIO_PRIORITY_ASYNC_READ, ARC_FLAG_PREDICTIVE_PREFETCH); } + ZFETCHSTAT_BUMP(zfetchstat_hits); } #if defined(_KERNEL) && defined(HAVE_SPL) module_param(zfs_prefetch_disable, int, 0644); MODULE_PARM_DESC(zfs_prefetch_disable, "Disable all ZFS prefetching"); module_param(zfetch_max_streams, uint, 0644); MODULE_PARM_DESC(zfetch_max_streams, "Max number of streams per zfetch"); module_param(zfetch_min_sec_reap, uint, 0644); MODULE_PARM_DESC(zfetch_min_sec_reap, "Min time before stream reclaim"); -module_param(zfetch_block_cap, uint, 0644); -MODULE_PARM_DESC(zfetch_block_cap, "Max number of blocks to fetch at a time"); +module_param(zfetch_max_distance, uint, 0644); +MODULE_PARM_DESC(zfetch_max_distance, + "Max bytes to prefetch per stream (default 8MB)"); module_param(zfetch_array_rd_sz, ulong, 0644); MODULE_PARM_DESC(zfetch_array_rd_sz, "Number of bytes in a array_read"); #endif diff --git a/module/zfs/dnode.c b/module/zfs/dnode.c index 190e66b744c0..38bcecd467e3 100644 --- a/module/zfs/dnode.c +++ b/module/zfs/dnode.c @@ -1,2044 +1,2042 @@ /* * 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) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2012, 2015 by Delphix. All rights reserved. * Copyright (c) 2014 Spectra Logic Corporation, All rights reserved. */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include static kmem_cache_t *dnode_cache; /* * Define DNODE_STATS to turn on statistic gathering. By default, it is only * turned on when DEBUG is also defined. */ #ifdef DEBUG #define DNODE_STATS #endif /* DEBUG */ #ifdef DNODE_STATS #define DNODE_STAT_ADD(stat) ((stat)++) #else #define DNODE_STAT_ADD(stat) /* nothing */ #endif /* DNODE_STATS */ ASSERTV(static dnode_phys_t dnode_phys_zero); int zfs_default_bs = SPA_MINBLOCKSHIFT; int zfs_default_ibs = DN_MAX_INDBLKSHIFT; #ifdef _KERNEL static kmem_cbrc_t dnode_move(void *, void *, size_t, void *); #endif /* _KERNEL */ static int dbuf_compare(const void *x1, const void *x2) { const dmu_buf_impl_t *d1 = x1; const dmu_buf_impl_t *d2 = x2; if (d1->db_level < d2->db_level) { return (-1); } if (d1->db_level > d2->db_level) { return (1); } if (d1->db_blkid < d2->db_blkid) { return (-1); } if (d1->db_blkid > d2->db_blkid) { return (1); } if (d1->db_state == DB_SEARCH) { ASSERT3S(d2->db_state, !=, DB_SEARCH); return (-1); } else if (d2->db_state == DB_SEARCH) { ASSERT3S(d1->db_state, !=, DB_SEARCH); return (1); } if ((uintptr_t)d1 < (uintptr_t)d2) { return (-1); } if ((uintptr_t)d1 > (uintptr_t)d2) { return (1); } return (0); } /* ARGSUSED */ static int dnode_cons(void *arg, void *unused, int kmflag) { dnode_t *dn = arg; int i; rw_init(&dn->dn_struct_rwlock, NULL, RW_NOLOCKDEP, NULL); mutex_init(&dn->dn_mtx, NULL, MUTEX_DEFAULT, NULL); mutex_init(&dn->dn_dbufs_mtx, NULL, MUTEX_DEFAULT, NULL); cv_init(&dn->dn_notxholds, NULL, CV_DEFAULT, NULL); /* * Every dbuf has a reference, and dropping a tracked reference is * O(number of references), so don't track dn_holds. */ refcount_create_untracked(&dn->dn_holds); refcount_create(&dn->dn_tx_holds); list_link_init(&dn->dn_link); bzero(&dn->dn_next_nblkptr[0], sizeof (dn->dn_next_nblkptr)); bzero(&dn->dn_next_nlevels[0], sizeof (dn->dn_next_nlevels)); bzero(&dn->dn_next_indblkshift[0], sizeof (dn->dn_next_indblkshift)); bzero(&dn->dn_next_bonustype[0], sizeof (dn->dn_next_bonustype)); bzero(&dn->dn_rm_spillblk[0], sizeof (dn->dn_rm_spillblk)); bzero(&dn->dn_next_bonuslen[0], sizeof (dn->dn_next_bonuslen)); bzero(&dn->dn_next_blksz[0], sizeof (dn->dn_next_blksz)); for (i = 0; i < TXG_SIZE; i++) { list_link_init(&dn->dn_dirty_link[i]); dn->dn_free_ranges[i] = NULL; list_create(&dn->dn_dirty_records[i], sizeof (dbuf_dirty_record_t), offsetof(dbuf_dirty_record_t, dr_dirty_node)); } dn->dn_allocated_txg = 0; dn->dn_free_txg = 0; dn->dn_assigned_txg = 0; dn->dn_dirtyctx = 0; dn->dn_dirtyctx_firstset = NULL; dn->dn_bonus = NULL; dn->dn_have_spill = B_FALSE; dn->dn_zio = NULL; dn->dn_oldused = 0; dn->dn_oldflags = 0; dn->dn_olduid = 0; dn->dn_oldgid = 0; dn->dn_newuid = 0; dn->dn_newgid = 0; dn->dn_id_flags = 0; dn->dn_dbufs_count = 0; dn->dn_unlisted_l0_blkid = 0; avl_create(&dn->dn_dbufs, dbuf_compare, sizeof (dmu_buf_impl_t), offsetof(dmu_buf_impl_t, db_link)); dn->dn_moved = 0; return (0); } /* ARGSUSED */ static void dnode_dest(void *arg, void *unused) { int i; dnode_t *dn = arg; rw_destroy(&dn->dn_struct_rwlock); mutex_destroy(&dn->dn_mtx); mutex_destroy(&dn->dn_dbufs_mtx); cv_destroy(&dn->dn_notxholds); refcount_destroy(&dn->dn_holds); refcount_destroy(&dn->dn_tx_holds); ASSERT(!list_link_active(&dn->dn_link)); for (i = 0; i < TXG_SIZE; i++) { ASSERT(!list_link_active(&dn->dn_dirty_link[i])); ASSERT3P(dn->dn_free_ranges[i], ==, NULL); list_destroy(&dn->dn_dirty_records[i]); ASSERT0(dn->dn_next_nblkptr[i]); ASSERT0(dn->dn_next_nlevels[i]); ASSERT0(dn->dn_next_indblkshift[i]); ASSERT0(dn->dn_next_bonustype[i]); ASSERT0(dn->dn_rm_spillblk[i]); ASSERT0(dn->dn_next_bonuslen[i]); ASSERT0(dn->dn_next_blksz[i]); } ASSERT0(dn->dn_allocated_txg); ASSERT0(dn->dn_free_txg); ASSERT0(dn->dn_assigned_txg); ASSERT0(dn->dn_dirtyctx); ASSERT3P(dn->dn_dirtyctx_firstset, ==, NULL); ASSERT3P(dn->dn_bonus, ==, NULL); ASSERT(!dn->dn_have_spill); ASSERT3P(dn->dn_zio, ==, NULL); ASSERT0(dn->dn_oldused); ASSERT0(dn->dn_oldflags); ASSERT0(dn->dn_olduid); ASSERT0(dn->dn_oldgid); ASSERT0(dn->dn_newuid); ASSERT0(dn->dn_newgid); ASSERT0(dn->dn_id_flags); ASSERT0(dn->dn_dbufs_count); ASSERT0(dn->dn_unlisted_l0_blkid); avl_destroy(&dn->dn_dbufs); } void dnode_init(void) { ASSERT(dnode_cache == NULL); dnode_cache = kmem_cache_create("dnode_t", sizeof (dnode_t), 0, dnode_cons, dnode_dest, NULL, NULL, NULL, 0); kmem_cache_set_move(dnode_cache, dnode_move); } void dnode_fini(void) { kmem_cache_destroy(dnode_cache); dnode_cache = NULL; } #ifdef ZFS_DEBUG void dnode_verify(dnode_t *dn) { int drop_struct_lock = FALSE; ASSERT(dn->dn_phys); ASSERT(dn->dn_objset); ASSERT(dn->dn_handle->dnh_dnode == dn); ASSERT(DMU_OT_IS_VALID(dn->dn_phys->dn_type)); if (!(zfs_flags & ZFS_DEBUG_DNODE_VERIFY)) return; if (!RW_WRITE_HELD(&dn->dn_struct_rwlock)) { rw_enter(&dn->dn_struct_rwlock, RW_READER); drop_struct_lock = TRUE; } if (dn->dn_phys->dn_type != DMU_OT_NONE || dn->dn_allocated_txg != 0) { int i; ASSERT3U(dn->dn_indblkshift, <=, SPA_MAXBLOCKSHIFT); if (dn->dn_datablkshift) { ASSERT3U(dn->dn_datablkshift, >=, SPA_MINBLOCKSHIFT); ASSERT3U(dn->dn_datablkshift, <=, SPA_MAXBLOCKSHIFT); ASSERT3U(1<dn_datablkshift, ==, dn->dn_datablksz); } ASSERT3U(dn->dn_nlevels, <=, 30); ASSERT(DMU_OT_IS_VALID(dn->dn_type)); ASSERT3U(dn->dn_nblkptr, >=, 1); ASSERT3U(dn->dn_nblkptr, <=, DN_MAX_NBLKPTR); ASSERT3U(dn->dn_bonuslen, <=, DN_MAX_BONUSLEN); ASSERT3U(dn->dn_datablksz, ==, dn->dn_datablkszsec << SPA_MINBLOCKSHIFT); ASSERT3U(ISP2(dn->dn_datablksz), ==, dn->dn_datablkshift != 0); ASSERT3U((dn->dn_nblkptr - 1) * sizeof (blkptr_t) + dn->dn_bonuslen, <=, DN_MAX_BONUSLEN); for (i = 0; i < TXG_SIZE; i++) { ASSERT3U(dn->dn_next_nlevels[i], <=, dn->dn_nlevels); } } if (dn->dn_phys->dn_type != DMU_OT_NONE) ASSERT3U(dn->dn_phys->dn_nlevels, <=, dn->dn_nlevels); ASSERT(DMU_OBJECT_IS_SPECIAL(dn->dn_object) || dn->dn_dbuf != NULL); if (dn->dn_dbuf != NULL) { ASSERT3P(dn->dn_phys, ==, (dnode_phys_t *)dn->dn_dbuf->db.db_data + (dn->dn_object % (dn->dn_dbuf->db.db_size >> DNODE_SHIFT))); } if (drop_struct_lock) rw_exit(&dn->dn_struct_rwlock); } #endif void dnode_byteswap(dnode_phys_t *dnp) { uint64_t *buf64 = (void*)&dnp->dn_blkptr; int i; if (dnp->dn_type == DMU_OT_NONE) { bzero(dnp, sizeof (dnode_phys_t)); return; } dnp->dn_datablkszsec = BSWAP_16(dnp->dn_datablkszsec); dnp->dn_bonuslen = BSWAP_16(dnp->dn_bonuslen); dnp->dn_maxblkid = BSWAP_64(dnp->dn_maxblkid); dnp->dn_used = BSWAP_64(dnp->dn_used); /* * dn_nblkptr is only one byte, so it's OK to read it in either * byte order. We can't read dn_bouslen. */ ASSERT(dnp->dn_indblkshift <= SPA_MAXBLOCKSHIFT); ASSERT(dnp->dn_nblkptr <= DN_MAX_NBLKPTR); for (i = 0; i < dnp->dn_nblkptr * sizeof (blkptr_t)/8; i++) buf64[i] = BSWAP_64(buf64[i]); /* * OK to check dn_bonuslen for zero, because it won't matter if * we have the wrong byte order. This is necessary because the * dnode dnode is smaller than a regular dnode. */ if (dnp->dn_bonuslen != 0) { /* * Note that the bonus length calculated here may be * longer than the actual bonus buffer. This is because * we always put the bonus buffer after the last block * pointer (instead of packing it against the end of the * dnode buffer). */ int off = (dnp->dn_nblkptr-1) * sizeof (blkptr_t); size_t len = DN_MAX_BONUSLEN - off; dmu_object_byteswap_t byteswap; ASSERT(DMU_OT_IS_VALID(dnp->dn_bonustype)); byteswap = DMU_OT_BYTESWAP(dnp->dn_bonustype); dmu_ot_byteswap[byteswap].ob_func(dnp->dn_bonus + off, len); } /* Swap SPILL block if we have one */ if (dnp->dn_flags & DNODE_FLAG_SPILL_BLKPTR) byteswap_uint64_array(&dnp->dn_spill, sizeof (blkptr_t)); } void dnode_buf_byteswap(void *vbuf, size_t size) { dnode_phys_t *buf = vbuf; int i; ASSERT3U(sizeof (dnode_phys_t), ==, (1<>= DNODE_SHIFT; for (i = 0; i < size; i++) { dnode_byteswap(buf); buf++; } } void dnode_setbonuslen(dnode_t *dn, int newsize, dmu_tx_t *tx) { ASSERT3U(refcount_count(&dn->dn_holds), >=, 1); dnode_setdirty(dn, tx); rw_enter(&dn->dn_struct_rwlock, RW_WRITER); ASSERT3U(newsize, <=, DN_MAX_BONUSLEN - (dn->dn_nblkptr-1) * sizeof (blkptr_t)); dn->dn_bonuslen = newsize; if (newsize == 0) dn->dn_next_bonuslen[tx->tx_txg & TXG_MASK] = DN_ZERO_BONUSLEN; else dn->dn_next_bonuslen[tx->tx_txg & TXG_MASK] = dn->dn_bonuslen; rw_exit(&dn->dn_struct_rwlock); } void dnode_setbonus_type(dnode_t *dn, dmu_object_type_t newtype, dmu_tx_t *tx) { ASSERT3U(refcount_count(&dn->dn_holds), >=, 1); dnode_setdirty(dn, tx); rw_enter(&dn->dn_struct_rwlock, RW_WRITER); dn->dn_bonustype = newtype; dn->dn_next_bonustype[tx->tx_txg & TXG_MASK] = dn->dn_bonustype; rw_exit(&dn->dn_struct_rwlock); } void dnode_rm_spill(dnode_t *dn, dmu_tx_t *tx) { ASSERT3U(refcount_count(&dn->dn_holds), >=, 1); ASSERT(RW_WRITE_HELD(&dn->dn_struct_rwlock)); dnode_setdirty(dn, tx); dn->dn_rm_spillblk[tx->tx_txg&TXG_MASK] = DN_KILL_SPILLBLK; dn->dn_have_spill = B_FALSE; } static void dnode_setdblksz(dnode_t *dn, int size) { ASSERT0(P2PHASE(size, SPA_MINBLOCKSIZE)); ASSERT3U(size, <=, SPA_MAXBLOCKSIZE); ASSERT3U(size, >=, SPA_MINBLOCKSIZE); ASSERT3U(size >> SPA_MINBLOCKSHIFT, <, 1<<(sizeof (dn->dn_phys->dn_datablkszsec) * 8)); dn->dn_datablksz = size; dn->dn_datablkszsec = size >> SPA_MINBLOCKSHIFT; dn->dn_datablkshift = ISP2(size) ? highbit64(size - 1) : 0; } static dnode_t * dnode_create(objset_t *os, dnode_phys_t *dnp, dmu_buf_impl_t *db, uint64_t object, dnode_handle_t *dnh) { dnode_t *dn; dn = kmem_cache_alloc(dnode_cache, KM_SLEEP); ASSERT(!POINTER_IS_VALID(dn->dn_objset)); dn->dn_moved = 0; /* * Defer setting dn_objset until the dnode is ready to be a candidate * for the dnode_move() callback. */ dn->dn_object = object; dn->dn_dbuf = db; dn->dn_handle = dnh; dn->dn_phys = dnp; if (dnp->dn_datablkszsec) { dnode_setdblksz(dn, dnp->dn_datablkszsec << SPA_MINBLOCKSHIFT); } else { dn->dn_datablksz = 0; dn->dn_datablkszsec = 0; dn->dn_datablkshift = 0; } dn->dn_indblkshift = dnp->dn_indblkshift; dn->dn_nlevels = dnp->dn_nlevels; dn->dn_type = dnp->dn_type; dn->dn_nblkptr = dnp->dn_nblkptr; dn->dn_checksum = dnp->dn_checksum; dn->dn_compress = dnp->dn_compress; dn->dn_bonustype = dnp->dn_bonustype; dn->dn_bonuslen = dnp->dn_bonuslen; dn->dn_maxblkid = dnp->dn_maxblkid; dn->dn_have_spill = ((dnp->dn_flags & DNODE_FLAG_SPILL_BLKPTR) != 0); dn->dn_id_flags = 0; dmu_zfetch_init(&dn->dn_zfetch, dn); ASSERT(DMU_OT_IS_VALID(dn->dn_phys->dn_type)); mutex_enter(&os->os_lock); if (dnh->dnh_dnode != NULL) { /* Lost the allocation race. */ mutex_exit(&os->os_lock); kmem_cache_free(dnode_cache, dn); return (dnh->dnh_dnode); } /* * Exclude special dnodes from os_dnodes so an empty os_dnodes * signifies that the special dnodes have no references from * their children (the entries in os_dnodes). This allows * dnode_destroy() to easily determine if the last child has * been removed and then complete eviction of the objset. */ if (!DMU_OBJECT_IS_SPECIAL(object)) list_insert_head(&os->os_dnodes, dn); membar_producer(); /* * Everything else must be valid before assigning dn_objset * makes the dnode eligible for dnode_move(). */ dn->dn_objset = os; dnh->dnh_dnode = dn; mutex_exit(&os->os_lock); arc_space_consume(sizeof (dnode_t), ARC_SPACE_OTHER); return (dn); } /* * Caller must be holding the dnode handle, which is released upon return. */ static void dnode_destroy(dnode_t *dn) { objset_t *os = dn->dn_objset; boolean_t complete_os_eviction = B_FALSE; ASSERT((dn->dn_id_flags & DN_ID_NEW_EXIST) == 0); mutex_enter(&os->os_lock); POINTER_INVALIDATE(&dn->dn_objset); if (!DMU_OBJECT_IS_SPECIAL(dn->dn_object)) { list_remove(&os->os_dnodes, dn); complete_os_eviction = list_is_empty(&os->os_dnodes) && list_link_active(&os->os_evicting_node); } mutex_exit(&os->os_lock); /* the dnode can no longer move, so we can release the handle */ zrl_remove(&dn->dn_handle->dnh_zrlock); dn->dn_allocated_txg = 0; dn->dn_free_txg = 0; dn->dn_assigned_txg = 0; dn->dn_dirtyctx = 0; if (dn->dn_dirtyctx_firstset != NULL) { kmem_free(dn->dn_dirtyctx_firstset, 1); dn->dn_dirtyctx_firstset = NULL; } if (dn->dn_bonus != NULL) { mutex_enter(&dn->dn_bonus->db_mtx); dbuf_evict(dn->dn_bonus); dn->dn_bonus = NULL; } dn->dn_zio = NULL; dn->dn_have_spill = B_FALSE; dn->dn_oldused = 0; dn->dn_oldflags = 0; dn->dn_olduid = 0; dn->dn_oldgid = 0; dn->dn_newuid = 0; dn->dn_newgid = 0; dn->dn_id_flags = 0; dn->dn_unlisted_l0_blkid = 0; - dmu_zfetch_rele(&dn->dn_zfetch); + dmu_zfetch_fini(&dn->dn_zfetch); kmem_cache_free(dnode_cache, dn); arc_space_return(sizeof (dnode_t), ARC_SPACE_OTHER); if (complete_os_eviction) dmu_objset_evict_done(os); } void dnode_allocate(dnode_t *dn, dmu_object_type_t ot, int blocksize, int ibs, dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx) { int i; ASSERT3U(blocksize, <=, spa_maxblocksize(dmu_objset_spa(dn->dn_objset))); if (blocksize == 0) blocksize = 1 << zfs_default_bs; else blocksize = P2ROUNDUP(blocksize, SPA_MINBLOCKSIZE); if (ibs == 0) ibs = zfs_default_ibs; ibs = MIN(MAX(ibs, DN_MIN_INDBLKSHIFT), DN_MAX_INDBLKSHIFT); dprintf("os=%p obj=%llu txg=%llu blocksize=%d ibs=%d\n", dn->dn_objset, dn->dn_object, tx->tx_txg, blocksize, ibs); ASSERT(dn->dn_type == DMU_OT_NONE); ASSERT(bcmp(dn->dn_phys, &dnode_phys_zero, sizeof (dnode_phys_t)) == 0); ASSERT(dn->dn_phys->dn_type == DMU_OT_NONE); ASSERT(ot != DMU_OT_NONE); ASSERT(DMU_OT_IS_VALID(ot)); ASSERT((bonustype == DMU_OT_NONE && bonuslen == 0) || (bonustype == DMU_OT_SA && bonuslen == 0) || (bonustype != DMU_OT_NONE && bonuslen != 0)); ASSERT(DMU_OT_IS_VALID(bonustype)); ASSERT3U(bonuslen, <=, DN_MAX_BONUSLEN); ASSERT(dn->dn_type == DMU_OT_NONE); ASSERT0(dn->dn_maxblkid); ASSERT0(dn->dn_allocated_txg); ASSERT0(dn->dn_assigned_txg); ASSERT(refcount_is_zero(&dn->dn_tx_holds)); ASSERT3U(refcount_count(&dn->dn_holds), <=, 1); ASSERT(avl_is_empty(&dn->dn_dbufs)); for (i = 0; i < TXG_SIZE; i++) { ASSERT0(dn->dn_next_nblkptr[i]); ASSERT0(dn->dn_next_nlevels[i]); ASSERT0(dn->dn_next_indblkshift[i]); ASSERT0(dn->dn_next_bonuslen[i]); ASSERT0(dn->dn_next_bonustype[i]); ASSERT0(dn->dn_rm_spillblk[i]); ASSERT0(dn->dn_next_blksz[i]); ASSERT(!list_link_active(&dn->dn_dirty_link[i])); ASSERT3P(list_head(&dn->dn_dirty_records[i]), ==, NULL); ASSERT3P(dn->dn_free_ranges[i], ==, NULL); } dn->dn_type = ot; dnode_setdblksz(dn, blocksize); dn->dn_indblkshift = ibs; dn->dn_nlevels = 1; if (bonustype == DMU_OT_SA) /* Maximize bonus space for SA */ dn->dn_nblkptr = 1; else dn->dn_nblkptr = 1 + ((DN_MAX_BONUSLEN - bonuslen) >> SPA_BLKPTRSHIFT); dn->dn_bonustype = bonustype; dn->dn_bonuslen = bonuslen; dn->dn_checksum = ZIO_CHECKSUM_INHERIT; dn->dn_compress = ZIO_COMPRESS_INHERIT; dn->dn_dirtyctx = 0; dn->dn_free_txg = 0; if (dn->dn_dirtyctx_firstset) { kmem_free(dn->dn_dirtyctx_firstset, 1); dn->dn_dirtyctx_firstset = NULL; } dn->dn_allocated_txg = tx->tx_txg; dn->dn_id_flags = 0; dnode_setdirty(dn, tx); dn->dn_next_indblkshift[tx->tx_txg & TXG_MASK] = ibs; dn->dn_next_bonuslen[tx->tx_txg & TXG_MASK] = dn->dn_bonuslen; dn->dn_next_bonustype[tx->tx_txg & TXG_MASK] = dn->dn_bonustype; dn->dn_next_blksz[tx->tx_txg & TXG_MASK] = dn->dn_datablksz; } void dnode_reallocate(dnode_t *dn, dmu_object_type_t ot, int blocksize, dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx) { int nblkptr; ASSERT3U(blocksize, >=, SPA_MINBLOCKSIZE); ASSERT3U(blocksize, <=, spa_maxblocksize(dmu_objset_spa(dn->dn_objset))); ASSERT0(blocksize % SPA_MINBLOCKSIZE); ASSERT(dn->dn_object != DMU_META_DNODE_OBJECT || dmu_tx_private_ok(tx)); ASSERT(tx->tx_txg != 0); ASSERT((bonustype == DMU_OT_NONE && bonuslen == 0) || (bonustype != DMU_OT_NONE && bonuslen != 0) || (bonustype == DMU_OT_SA && bonuslen == 0)); ASSERT(DMU_OT_IS_VALID(bonustype)); ASSERT3U(bonuslen, <=, DN_MAX_BONUSLEN); /* clean up any unreferenced dbufs */ dnode_evict_dbufs(dn); dn->dn_id_flags = 0; rw_enter(&dn->dn_struct_rwlock, RW_WRITER); dnode_setdirty(dn, tx); if (dn->dn_datablksz != blocksize) { /* change blocksize */ ASSERT(dn->dn_maxblkid == 0 && (BP_IS_HOLE(&dn->dn_phys->dn_blkptr[0]) || dnode_block_freed(dn, 0))); dnode_setdblksz(dn, blocksize); dn->dn_next_blksz[tx->tx_txg&TXG_MASK] = blocksize; } if (dn->dn_bonuslen != bonuslen) dn->dn_next_bonuslen[tx->tx_txg&TXG_MASK] = bonuslen; if (bonustype == DMU_OT_SA) /* Maximize bonus space for SA */ nblkptr = 1; else nblkptr = 1 + ((DN_MAX_BONUSLEN - bonuslen) >> SPA_BLKPTRSHIFT); if (dn->dn_bonustype != bonustype) dn->dn_next_bonustype[tx->tx_txg&TXG_MASK] = bonustype; if (dn->dn_nblkptr != nblkptr) dn->dn_next_nblkptr[tx->tx_txg&TXG_MASK] = nblkptr; if (dn->dn_phys->dn_flags & DNODE_FLAG_SPILL_BLKPTR) { dbuf_rm_spill(dn, tx); dnode_rm_spill(dn, tx); } rw_exit(&dn->dn_struct_rwlock); /* change type */ dn->dn_type = ot; /* change bonus size and type */ mutex_enter(&dn->dn_mtx); dn->dn_bonustype = bonustype; dn->dn_bonuslen = bonuslen; dn->dn_nblkptr = nblkptr; dn->dn_checksum = ZIO_CHECKSUM_INHERIT; dn->dn_compress = ZIO_COMPRESS_INHERIT; ASSERT3U(dn->dn_nblkptr, <=, DN_MAX_NBLKPTR); /* fix up the bonus db_size */ if (dn->dn_bonus) { dn->dn_bonus->db.db_size = DN_MAX_BONUSLEN - (dn->dn_nblkptr-1) * sizeof (blkptr_t); ASSERT(dn->dn_bonuslen <= dn->dn_bonus->db.db_size); } dn->dn_allocated_txg = tx->tx_txg; mutex_exit(&dn->dn_mtx); } #ifdef _KERNEL #ifdef DNODE_STATS static struct { uint64_t dms_dnode_invalid; uint64_t dms_dnode_recheck1; uint64_t dms_dnode_recheck2; uint64_t dms_dnode_special; uint64_t dms_dnode_handle; uint64_t dms_dnode_rwlock; uint64_t dms_dnode_active; } dnode_move_stats; #endif /* DNODE_STATS */ static void dnode_move_impl(dnode_t *odn, dnode_t *ndn) { int i; ASSERT(!RW_LOCK_HELD(&odn->dn_struct_rwlock)); ASSERT(MUTEX_NOT_HELD(&odn->dn_mtx)); ASSERT(MUTEX_NOT_HELD(&odn->dn_dbufs_mtx)); ASSERT(!RW_LOCK_HELD(&odn->dn_zfetch.zf_rwlock)); /* Copy fields. */ ndn->dn_objset = odn->dn_objset; ndn->dn_object = odn->dn_object; ndn->dn_dbuf = odn->dn_dbuf; ndn->dn_handle = odn->dn_handle; ndn->dn_phys = odn->dn_phys; ndn->dn_type = odn->dn_type; ndn->dn_bonuslen = odn->dn_bonuslen; ndn->dn_bonustype = odn->dn_bonustype; ndn->dn_nblkptr = odn->dn_nblkptr; ndn->dn_checksum = odn->dn_checksum; ndn->dn_compress = odn->dn_compress; ndn->dn_nlevels = odn->dn_nlevels; ndn->dn_indblkshift = odn->dn_indblkshift; ndn->dn_datablkshift = odn->dn_datablkshift; ndn->dn_datablkszsec = odn->dn_datablkszsec; ndn->dn_datablksz = odn->dn_datablksz; ndn->dn_maxblkid = odn->dn_maxblkid; bcopy(&odn->dn_next_nblkptr[0], &ndn->dn_next_nblkptr[0], sizeof (odn->dn_next_nblkptr)); bcopy(&odn->dn_next_nlevels[0], &ndn->dn_next_nlevels[0], sizeof (odn->dn_next_nlevels)); bcopy(&odn->dn_next_indblkshift[0], &ndn->dn_next_indblkshift[0], sizeof (odn->dn_next_indblkshift)); bcopy(&odn->dn_next_bonustype[0], &ndn->dn_next_bonustype[0], sizeof (odn->dn_next_bonustype)); bcopy(&odn->dn_rm_spillblk[0], &ndn->dn_rm_spillblk[0], sizeof (odn->dn_rm_spillblk)); bcopy(&odn->dn_next_bonuslen[0], &ndn->dn_next_bonuslen[0], sizeof (odn->dn_next_bonuslen)); bcopy(&odn->dn_next_blksz[0], &ndn->dn_next_blksz[0], sizeof (odn->dn_next_blksz)); for (i = 0; i < TXG_SIZE; i++) { list_move_tail(&ndn->dn_dirty_records[i], &odn->dn_dirty_records[i]); } bcopy(&odn->dn_free_ranges[0], &ndn->dn_free_ranges[0], sizeof (odn->dn_free_ranges)); ndn->dn_allocated_txg = odn->dn_allocated_txg; ndn->dn_free_txg = odn->dn_free_txg; ndn->dn_assigned_txg = odn->dn_assigned_txg; ndn->dn_dirtyctx = odn->dn_dirtyctx; ndn->dn_dirtyctx_firstset = odn->dn_dirtyctx_firstset; ASSERT(refcount_count(&odn->dn_tx_holds) == 0); refcount_transfer(&ndn->dn_holds, &odn->dn_holds); ASSERT(avl_is_empty(&ndn->dn_dbufs)); avl_swap(&ndn->dn_dbufs, &odn->dn_dbufs); ndn->dn_dbufs_count = odn->dn_dbufs_count; ndn->dn_unlisted_l0_blkid = odn->dn_unlisted_l0_blkid; ndn->dn_bonus = odn->dn_bonus; ndn->dn_have_spill = odn->dn_have_spill; ndn->dn_zio = odn->dn_zio; ndn->dn_oldused = odn->dn_oldused; ndn->dn_oldflags = odn->dn_oldflags; ndn->dn_olduid = odn->dn_olduid; ndn->dn_oldgid = odn->dn_oldgid; ndn->dn_newuid = odn->dn_newuid; ndn->dn_newgid = odn->dn_newgid; ndn->dn_id_flags = odn->dn_id_flags; dmu_zfetch_init(&ndn->dn_zfetch, NULL); list_move_tail(&ndn->dn_zfetch.zf_stream, &odn->dn_zfetch.zf_stream); ndn->dn_zfetch.zf_dnode = odn->dn_zfetch.zf_dnode; - ndn->dn_zfetch.zf_stream_cnt = odn->dn_zfetch.zf_stream_cnt; - ndn->dn_zfetch.zf_alloc_fail = odn->dn_zfetch.zf_alloc_fail; /* * Update back pointers. Updating the handle fixes the back pointer of * every descendant dbuf as well as the bonus dbuf. */ ASSERT(ndn->dn_handle->dnh_dnode == odn); ndn->dn_handle->dnh_dnode = ndn; if (ndn->dn_zfetch.zf_dnode == odn) { ndn->dn_zfetch.zf_dnode = ndn; } /* * Invalidate the original dnode by clearing all of its back pointers. */ odn->dn_dbuf = NULL; odn->dn_handle = NULL; avl_create(&odn->dn_dbufs, dbuf_compare, sizeof (dmu_buf_impl_t), offsetof(dmu_buf_impl_t, db_link)); odn->dn_dbufs_count = 0; odn->dn_unlisted_l0_blkid = 0; odn->dn_bonus = NULL; odn->dn_zfetch.zf_dnode = NULL; /* * Set the low bit of the objset pointer to ensure that dnode_move() * recognizes the dnode as invalid in any subsequent callback. */ POINTER_INVALIDATE(&odn->dn_objset); /* * Satisfy the destructor. */ for (i = 0; i < TXG_SIZE; i++) { list_create(&odn->dn_dirty_records[i], sizeof (dbuf_dirty_record_t), offsetof(dbuf_dirty_record_t, dr_dirty_node)); odn->dn_free_ranges[i] = NULL; odn->dn_next_nlevels[i] = 0; odn->dn_next_indblkshift[i] = 0; odn->dn_next_bonustype[i] = 0; odn->dn_rm_spillblk[i] = 0; odn->dn_next_bonuslen[i] = 0; odn->dn_next_blksz[i] = 0; } odn->dn_allocated_txg = 0; odn->dn_free_txg = 0; odn->dn_assigned_txg = 0; odn->dn_dirtyctx = 0; odn->dn_dirtyctx_firstset = NULL; odn->dn_have_spill = B_FALSE; odn->dn_zio = NULL; odn->dn_oldused = 0; odn->dn_oldflags = 0; odn->dn_olduid = 0; odn->dn_oldgid = 0; odn->dn_newuid = 0; odn->dn_newgid = 0; odn->dn_id_flags = 0; /* * Mark the dnode. */ ndn->dn_moved = 1; odn->dn_moved = (uint8_t)-1; } /*ARGSUSED*/ static kmem_cbrc_t dnode_move(void *buf, void *newbuf, size_t size, void *arg) { dnode_t *odn = buf, *ndn = newbuf; objset_t *os; int64_t refcount; uint32_t dbufs; /* * The dnode is on the objset's list of known dnodes if the objset * pointer is valid. We set the low bit of the objset pointer when * freeing the dnode to invalidate it, and the memory patterns written * by kmem (baddcafe and deadbeef) set at least one of the two low bits. * A newly created dnode sets the objset pointer last of all to indicate * that the dnode is known and in a valid state to be moved by this * function. */ os = odn->dn_objset; if (!POINTER_IS_VALID(os)) { DNODE_STAT_ADD(dnode_move_stats.dms_dnode_invalid); return (KMEM_CBRC_DONT_KNOW); } /* * Ensure that the objset does not go away during the move. */ rw_enter(&os_lock, RW_WRITER); if (os != odn->dn_objset) { rw_exit(&os_lock); DNODE_STAT_ADD(dnode_move_stats.dms_dnode_recheck1); return (KMEM_CBRC_DONT_KNOW); } /* * If the dnode is still valid, then so is the objset. We know that no * valid objset can be freed while we hold os_lock, so we can safely * ensure that the objset remains in use. */ mutex_enter(&os->os_lock); /* * Recheck the objset pointer in case the dnode was removed just before * acquiring the lock. */ if (os != odn->dn_objset) { mutex_exit(&os->os_lock); rw_exit(&os_lock); DNODE_STAT_ADD(dnode_move_stats.dms_dnode_recheck2); return (KMEM_CBRC_DONT_KNOW); } /* * At this point we know that as long as we hold os->os_lock, the dnode * cannot be freed and fields within the dnode can be safely accessed. * The objset listing this dnode cannot go away as long as this dnode is * on its list. */ rw_exit(&os_lock); if (DMU_OBJECT_IS_SPECIAL(odn->dn_object)) { mutex_exit(&os->os_lock); DNODE_STAT_ADD(dnode_move_stats.dms_dnode_special); return (KMEM_CBRC_NO); } ASSERT(odn->dn_dbuf != NULL); /* only "special" dnodes have no parent */ /* * Lock the dnode handle to prevent the dnode from obtaining any new * holds. This also prevents the descendant dbufs and the bonus dbuf * from accessing the dnode, so that we can discount their holds. The * handle is safe to access because we know that while the dnode cannot * go away, neither can its handle. Once we hold dnh_zrlock, we can * safely move any dnode referenced only by dbufs. */ if (!zrl_tryenter(&odn->dn_handle->dnh_zrlock)) { mutex_exit(&os->os_lock); DNODE_STAT_ADD(dnode_move_stats.dms_dnode_handle); return (KMEM_CBRC_LATER); } /* * Ensure a consistent view of the dnode's holds and the dnode's dbufs. * We need to guarantee that there is a hold for every dbuf in order to * determine whether the dnode is actively referenced. Falsely matching * a dbuf to an active hold would lead to an unsafe move. It's possible * that a thread already having an active dnode hold is about to add a * dbuf, and we can't compare hold and dbuf counts while the add is in * progress. */ if (!rw_tryenter(&odn->dn_struct_rwlock, RW_WRITER)) { zrl_exit(&odn->dn_handle->dnh_zrlock); mutex_exit(&os->os_lock); DNODE_STAT_ADD(dnode_move_stats.dms_dnode_rwlock); return (KMEM_CBRC_LATER); } /* * A dbuf may be removed (evicted) without an active dnode hold. In that * case, the dbuf count is decremented under the handle lock before the * dbuf's hold is released. This order ensures that if we count the hold * after the dbuf is removed but before its hold is released, we will * treat the unmatched hold as active and exit safely. If we count the * hold before the dbuf is removed, the hold is discounted, and the * removal is blocked until the move completes. */ refcount = refcount_count(&odn->dn_holds); ASSERT(refcount >= 0); dbufs = odn->dn_dbufs_count; /* We can't have more dbufs than dnode holds. */ ASSERT3U(dbufs, <=, refcount); DTRACE_PROBE3(dnode__move, dnode_t *, odn, int64_t, refcount, uint32_t, dbufs); if (refcount > dbufs) { rw_exit(&odn->dn_struct_rwlock); zrl_exit(&odn->dn_handle->dnh_zrlock); mutex_exit(&os->os_lock); DNODE_STAT_ADD(dnode_move_stats.dms_dnode_active); return (KMEM_CBRC_LATER); } rw_exit(&odn->dn_struct_rwlock); /* * At this point we know that anyone with a hold on the dnode is not * actively referencing it. The dnode is known and in a valid state to * move. We're holding the locks needed to execute the critical section. */ dnode_move_impl(odn, ndn); list_link_replace(&odn->dn_link, &ndn->dn_link); /* If the dnode was safe to move, the refcount cannot have changed. */ ASSERT(refcount == refcount_count(&ndn->dn_holds)); ASSERT(dbufs == ndn->dn_dbufs_count); zrl_exit(&ndn->dn_handle->dnh_zrlock); /* handle has moved */ mutex_exit(&os->os_lock); return (KMEM_CBRC_YES); } #endif /* _KERNEL */ void dnode_special_close(dnode_handle_t *dnh) { dnode_t *dn = dnh->dnh_dnode; /* * Wait for final references to the dnode to clear. This can * only happen if the arc is asyncronously evicting state that * has a hold on this dnode while we are trying to evict this * dnode. */ while (refcount_count(&dn->dn_holds) > 0) delay(1); ASSERT(dn->dn_dbuf == NULL || dmu_buf_get_user(&dn->dn_dbuf->db) == NULL); zrl_add(&dnh->dnh_zrlock); dnode_destroy(dn); /* implicit zrl_remove() */ zrl_destroy(&dnh->dnh_zrlock); dnh->dnh_dnode = NULL; } void dnode_special_open(objset_t *os, dnode_phys_t *dnp, uint64_t object, dnode_handle_t *dnh) { dnode_t *dn; dn = dnode_create(os, dnp, NULL, object, dnh); zrl_init(&dnh->dnh_zrlock); DNODE_VERIFY(dn); } static void dnode_buf_pageout(void *dbu) { dnode_children_t *children_dnodes = dbu; int i; for (i = 0; i < children_dnodes->dnc_count; i++) { dnode_handle_t *dnh = &children_dnodes->dnc_children[i]; dnode_t *dn; /* * The dnode handle lock guards against the dnode moving to * another valid address, so there is no need here to guard * against changes to or from NULL. */ if (dnh->dnh_dnode == NULL) { zrl_destroy(&dnh->dnh_zrlock); continue; } zrl_add(&dnh->dnh_zrlock); dn = dnh->dnh_dnode; /* * If there are holds on this dnode, then there should * be holds on the dnode's containing dbuf as well; thus * it wouldn't be eligible for eviction and this function * would not have been called. */ ASSERT(refcount_is_zero(&dn->dn_holds)); ASSERT(refcount_is_zero(&dn->dn_tx_holds)); dnode_destroy(dn); /* implicit zrl_remove() */ zrl_destroy(&dnh->dnh_zrlock); dnh->dnh_dnode = NULL; } kmem_free(children_dnodes, sizeof (dnode_children_t) + children_dnodes->dnc_count * sizeof (dnode_handle_t)); } /* * errors: * EINVAL - invalid object number. * EIO - i/o error. * succeeds even for free dnodes. */ int dnode_hold_impl(objset_t *os, uint64_t object, int flag, void *tag, dnode_t **dnp) { int epb, idx, err; int drop_struct_lock = FALSE; int type; uint64_t blk; dnode_t *mdn, *dn; dmu_buf_impl_t *db; dnode_children_t *children_dnodes; dnode_handle_t *dnh; /* * If you are holding the spa config lock as writer, you shouldn't * be asking the DMU to do *anything* unless it's the root pool * which may require us to read from the root filesystem while * holding some (not all) of the locks as writer. */ ASSERT(spa_config_held(os->os_spa, SCL_ALL, RW_WRITER) == 0 || (spa_is_root(os->os_spa) && spa_config_held(os->os_spa, SCL_STATE, RW_WRITER))); if (object == DMU_USERUSED_OBJECT || object == DMU_GROUPUSED_OBJECT) { dn = (object == DMU_USERUSED_OBJECT) ? DMU_USERUSED_DNODE(os) : DMU_GROUPUSED_DNODE(os); if (dn == NULL) return (SET_ERROR(ENOENT)); type = dn->dn_type; if ((flag & DNODE_MUST_BE_ALLOCATED) && type == DMU_OT_NONE) return (SET_ERROR(ENOENT)); if ((flag & DNODE_MUST_BE_FREE) && type != DMU_OT_NONE) return (SET_ERROR(EEXIST)); DNODE_VERIFY(dn); (void) refcount_add(&dn->dn_holds, tag); *dnp = dn; return (0); } if (object == 0 || object >= DN_MAX_OBJECT) return (SET_ERROR(EINVAL)); mdn = DMU_META_DNODE(os); ASSERT(mdn->dn_object == DMU_META_DNODE_OBJECT); DNODE_VERIFY(mdn); if (!RW_WRITE_HELD(&mdn->dn_struct_rwlock)) { rw_enter(&mdn->dn_struct_rwlock, RW_READER); drop_struct_lock = TRUE; } blk = dbuf_whichblock(mdn, 0, object * sizeof (dnode_phys_t)); db = dbuf_hold(mdn, blk, FTAG); if (drop_struct_lock) rw_exit(&mdn->dn_struct_rwlock); if (db == NULL) return (SET_ERROR(EIO)); err = dbuf_read(db, NULL, DB_RF_CANFAIL); if (err) { dbuf_rele(db, FTAG); return (err); } ASSERT3U(db->db.db_size, >=, 1<db.db_size >> DNODE_SHIFT; idx = object & (epb-1); ASSERT(DB_DNODE(db)->dn_type == DMU_OT_DNODE); children_dnodes = dmu_buf_get_user(&db->db); if (children_dnodes == NULL) { int i; dnode_children_t *winner; children_dnodes = kmem_zalloc(sizeof (dnode_children_t) + epb * sizeof (dnode_handle_t), KM_SLEEP); children_dnodes->dnc_count = epb; dnh = &children_dnodes->dnc_children[0]; for (i = 0; i < epb; i++) { zrl_init(&dnh[i].dnh_zrlock); } dmu_buf_init_user(&children_dnodes->dnc_dbu, dnode_buf_pageout, NULL); winner = dmu_buf_set_user(&db->db, &children_dnodes->dnc_dbu); if (winner != NULL) { for (i = 0; i < epb; i++) { zrl_destroy(&dnh[i].dnh_zrlock); } kmem_free(children_dnodes, sizeof (dnode_children_t) + epb * sizeof (dnode_handle_t)); children_dnodes = winner; } } ASSERT(children_dnodes->dnc_count == epb); dnh = &children_dnodes->dnc_children[idx]; zrl_add(&dnh->dnh_zrlock); dn = dnh->dnh_dnode; if (dn == NULL) { dnode_phys_t *phys = (dnode_phys_t *)db->db.db_data+idx; dn = dnode_create(os, phys, db, object, dnh); } mutex_enter(&dn->dn_mtx); type = dn->dn_type; if (dn->dn_free_txg || ((flag & DNODE_MUST_BE_ALLOCATED) && type == DMU_OT_NONE) || ((flag & DNODE_MUST_BE_FREE) && (type != DMU_OT_NONE || !refcount_is_zero(&dn->dn_holds)))) { mutex_exit(&dn->dn_mtx); zrl_remove(&dnh->dnh_zrlock); dbuf_rele(db, FTAG); return (type == DMU_OT_NONE ? ENOENT : EEXIST); } if (refcount_add(&dn->dn_holds, tag) == 1) dbuf_add_ref(db, dnh); mutex_exit(&dn->dn_mtx); /* Now we can rely on the hold to prevent the dnode from moving. */ zrl_remove(&dnh->dnh_zrlock); DNODE_VERIFY(dn); ASSERT3P(dn->dn_dbuf, ==, db); ASSERT3U(dn->dn_object, ==, object); dbuf_rele(db, FTAG); *dnp = dn; return (0); } /* * Return held dnode if the object is allocated, NULL if not. */ int dnode_hold(objset_t *os, uint64_t object, void *tag, dnode_t **dnp) { return (dnode_hold_impl(os, object, DNODE_MUST_BE_ALLOCATED, tag, dnp)); } /* * Can only add a reference if there is already at least one * reference on the dnode. Returns FALSE if unable to add a * new reference. */ boolean_t dnode_add_ref(dnode_t *dn, void *tag) { mutex_enter(&dn->dn_mtx); if (refcount_is_zero(&dn->dn_holds)) { mutex_exit(&dn->dn_mtx); return (FALSE); } VERIFY(1 < refcount_add(&dn->dn_holds, tag)); mutex_exit(&dn->dn_mtx); return (TRUE); } void dnode_rele(dnode_t *dn, void *tag) { mutex_enter(&dn->dn_mtx); dnode_rele_and_unlock(dn, tag); } void dnode_rele_and_unlock(dnode_t *dn, void *tag) { uint64_t refs; /* Get while the hold prevents the dnode from moving. */ dmu_buf_impl_t *db = dn->dn_dbuf; dnode_handle_t *dnh = dn->dn_handle; refs = refcount_remove(&dn->dn_holds, tag); mutex_exit(&dn->dn_mtx); /* * It's unsafe to release the last hold on a dnode by dnode_rele() or * indirectly by dbuf_rele() while relying on the dnode handle to * prevent the dnode from moving, since releasing the last hold could * result in the dnode's parent dbuf evicting its dnode handles. For * that reason anyone calling dnode_rele() or dbuf_rele() without some * other direct or indirect hold on the dnode must first drop the dnode * handle. */ ASSERT(refs > 0 || dnh->dnh_zrlock.zr_owner != curthread); /* NOTE: the DNODE_DNODE does not have a dn_dbuf */ if (refs == 0 && db != NULL) { /* * Another thread could add a hold to the dnode handle in * dnode_hold_impl() while holding the parent dbuf. Since the * hold on the parent dbuf prevents the handle from being * destroyed, the hold on the handle is OK. We can't yet assert * that the handle has zero references, but that will be * asserted anyway when the handle gets destroyed. */ dbuf_rele(db, dnh); } } void dnode_setdirty(dnode_t *dn, dmu_tx_t *tx) { objset_t *os = dn->dn_objset; uint64_t txg = tx->tx_txg; if (DMU_OBJECT_IS_SPECIAL(dn->dn_object)) { dsl_dataset_dirty(os->os_dsl_dataset, tx); return; } DNODE_VERIFY(dn); #ifdef ZFS_DEBUG mutex_enter(&dn->dn_mtx); ASSERT(dn->dn_phys->dn_type || dn->dn_allocated_txg); ASSERT(dn->dn_free_txg == 0 || dn->dn_free_txg >= txg); mutex_exit(&dn->dn_mtx); #endif /* * Determine old uid/gid when necessary */ dmu_objset_userquota_get_ids(dn, B_TRUE, tx); mutex_enter(&os->os_lock); /* * If we are already marked dirty, we're done. */ if (list_link_active(&dn->dn_dirty_link[txg & TXG_MASK])) { mutex_exit(&os->os_lock); return; } ASSERT(!refcount_is_zero(&dn->dn_holds) || !avl_is_empty(&dn->dn_dbufs)); ASSERT(dn->dn_datablksz != 0); ASSERT0(dn->dn_next_bonuslen[txg&TXG_MASK]); ASSERT0(dn->dn_next_blksz[txg&TXG_MASK]); ASSERT0(dn->dn_next_bonustype[txg&TXG_MASK]); dprintf_ds(os->os_dsl_dataset, "obj=%llu txg=%llu\n", dn->dn_object, txg); if (dn->dn_free_txg > 0 && dn->dn_free_txg <= txg) { list_insert_tail(&os->os_free_dnodes[txg&TXG_MASK], dn); } else { list_insert_tail(&os->os_dirty_dnodes[txg&TXG_MASK], dn); } mutex_exit(&os->os_lock); /* * The dnode maintains a hold on its containing dbuf as * long as there are holds on it. Each instantiated child * dbuf maintains a hold on the dnode. When the last child * drops its hold, the dnode will drop its hold on the * containing dbuf. We add a "dirty hold" here so that the * dnode will hang around after we finish processing its * children. */ VERIFY(dnode_add_ref(dn, (void *)(uintptr_t)tx->tx_txg)); (void) dbuf_dirty(dn->dn_dbuf, tx); dsl_dataset_dirty(os->os_dsl_dataset, tx); } void dnode_free(dnode_t *dn, dmu_tx_t *tx) { int txgoff = tx->tx_txg & TXG_MASK; dprintf("dn=%p txg=%llu\n", dn, tx->tx_txg); /* we should be the only holder... hopefully */ /* ASSERT3U(refcount_count(&dn->dn_holds), ==, 1); */ mutex_enter(&dn->dn_mtx); if (dn->dn_type == DMU_OT_NONE || dn->dn_free_txg) { mutex_exit(&dn->dn_mtx); return; } dn->dn_free_txg = tx->tx_txg; mutex_exit(&dn->dn_mtx); /* * If the dnode is already dirty, it needs to be moved from * the dirty list to the free list. */ mutex_enter(&dn->dn_objset->os_lock); if (list_link_active(&dn->dn_dirty_link[txgoff])) { list_remove(&dn->dn_objset->os_dirty_dnodes[txgoff], dn); list_insert_tail(&dn->dn_objset->os_free_dnodes[txgoff], dn); mutex_exit(&dn->dn_objset->os_lock); } else { mutex_exit(&dn->dn_objset->os_lock); dnode_setdirty(dn, tx); } } /* * Try to change the block size for the indicated dnode. This can only * succeed if there are no blocks allocated or dirty beyond first block */ int dnode_set_blksz(dnode_t *dn, uint64_t size, int ibs, dmu_tx_t *tx) { dmu_buf_impl_t *db; int err; ASSERT3U(size, <=, spa_maxblocksize(dmu_objset_spa(dn->dn_objset))); if (size == 0) size = SPA_MINBLOCKSIZE; else size = P2ROUNDUP(size, SPA_MINBLOCKSIZE); if (ibs == dn->dn_indblkshift) ibs = 0; if (size >> SPA_MINBLOCKSHIFT == dn->dn_datablkszsec && ibs == 0) return (0); rw_enter(&dn->dn_struct_rwlock, RW_WRITER); /* Check for any allocated blocks beyond the first */ if (dn->dn_maxblkid != 0) goto fail; mutex_enter(&dn->dn_dbufs_mtx); for (db = avl_first(&dn->dn_dbufs); db != NULL; db = AVL_NEXT(&dn->dn_dbufs, db)) { if (db->db_blkid != 0 && db->db_blkid != DMU_BONUS_BLKID && db->db_blkid != DMU_SPILL_BLKID) { mutex_exit(&dn->dn_dbufs_mtx); goto fail; } } mutex_exit(&dn->dn_dbufs_mtx); if (ibs && dn->dn_nlevels != 1) goto fail; /* resize the old block */ err = dbuf_hold_impl(dn, 0, 0, TRUE, FALSE, FTAG, &db); if (err == 0) dbuf_new_size(db, size, tx); else if (err != ENOENT) goto fail; dnode_setdblksz(dn, size); dnode_setdirty(dn, tx); dn->dn_next_blksz[tx->tx_txg&TXG_MASK] = size; if (ibs) { dn->dn_indblkshift = ibs; dn->dn_next_indblkshift[tx->tx_txg&TXG_MASK] = ibs; } /* rele after we have fixed the blocksize in the dnode */ if (db) dbuf_rele(db, FTAG); rw_exit(&dn->dn_struct_rwlock); return (0); fail: rw_exit(&dn->dn_struct_rwlock); return (SET_ERROR(ENOTSUP)); } /* read-holding callers must not rely on the lock being continuously held */ void dnode_new_blkid(dnode_t *dn, uint64_t blkid, dmu_tx_t *tx, boolean_t have_read) { uint64_t txgoff = tx->tx_txg & TXG_MASK; int epbs, new_nlevels; uint64_t sz; ASSERT(blkid != DMU_BONUS_BLKID); ASSERT(have_read ? RW_READ_HELD(&dn->dn_struct_rwlock) : RW_WRITE_HELD(&dn->dn_struct_rwlock)); /* * if we have a read-lock, check to see if we need to do any work * before upgrading to a write-lock. */ if (have_read) { if (blkid <= dn->dn_maxblkid) return; if (!rw_tryupgrade(&dn->dn_struct_rwlock)) { rw_exit(&dn->dn_struct_rwlock); rw_enter(&dn->dn_struct_rwlock, RW_WRITER); } } if (blkid <= dn->dn_maxblkid) goto out; dn->dn_maxblkid = blkid; /* * Compute the number of levels necessary to support the new maxblkid. */ new_nlevels = 1; epbs = dn->dn_indblkshift - SPA_BLKPTRSHIFT; for (sz = dn->dn_nblkptr; sz <= blkid && sz >= dn->dn_nblkptr; sz <<= epbs) new_nlevels++; if (new_nlevels > dn->dn_nlevels) { int old_nlevels = dn->dn_nlevels; dmu_buf_impl_t *db; list_t *list; dbuf_dirty_record_t *new, *dr, *dr_next; dn->dn_nlevels = new_nlevels; ASSERT3U(new_nlevels, >, dn->dn_next_nlevels[txgoff]); dn->dn_next_nlevels[txgoff] = new_nlevels; /* dirty the left indirects */ db = dbuf_hold_level(dn, old_nlevels, 0, FTAG); ASSERT(db != NULL); new = dbuf_dirty(db, tx); dbuf_rele(db, FTAG); /* transfer the dirty records to the new indirect */ mutex_enter(&dn->dn_mtx); mutex_enter(&new->dt.di.dr_mtx); list = &dn->dn_dirty_records[txgoff]; for (dr = list_head(list); dr; dr = dr_next) { dr_next = list_next(&dn->dn_dirty_records[txgoff], dr); if (dr->dr_dbuf->db_level != new_nlevels-1 && dr->dr_dbuf->db_blkid != DMU_BONUS_BLKID && dr->dr_dbuf->db_blkid != DMU_SPILL_BLKID) { ASSERT(dr->dr_dbuf->db_level == old_nlevels-1); list_remove(&dn->dn_dirty_records[txgoff], dr); list_insert_tail(&new->dt.di.dr_children, dr); dr->dr_parent = new; } } mutex_exit(&new->dt.di.dr_mtx); mutex_exit(&dn->dn_mtx); } out: if (have_read) rw_downgrade(&dn->dn_struct_rwlock); } static void dnode_dirty_l1(dnode_t *dn, uint64_t l1blkid, dmu_tx_t *tx) { dmu_buf_impl_t *db = dbuf_hold_level(dn, 1, l1blkid, FTAG); if (db != NULL) { dmu_buf_will_dirty(&db->db, tx); dbuf_rele(db, FTAG); } } void dnode_free_range(dnode_t *dn, uint64_t off, uint64_t len, dmu_tx_t *tx) { dmu_buf_impl_t *db; uint64_t blkoff, blkid, nblks; int blksz, blkshift, head, tail; int trunc = FALSE; int epbs; rw_enter(&dn->dn_struct_rwlock, RW_WRITER); blksz = dn->dn_datablksz; blkshift = dn->dn_datablkshift; epbs = dn->dn_indblkshift - SPA_BLKPTRSHIFT; if (len == DMU_OBJECT_END) { len = UINT64_MAX - off; trunc = TRUE; } /* * First, block align the region to free: */ if (ISP2(blksz)) { head = P2NPHASE(off, blksz); blkoff = P2PHASE(off, blksz); if ((off >> blkshift) > dn->dn_maxblkid) goto out; } else { ASSERT(dn->dn_maxblkid == 0); if (off == 0 && len >= blksz) { /* * Freeing the whole block; fast-track this request. * Note that we won't dirty any indirect blocks, * which is fine because we will be freeing the entire * file and thus all indirect blocks will be freed * by free_children(). */ blkid = 0; nblks = 1; goto done; } else if (off >= blksz) { /* Freeing past end-of-data */ goto out; } else { /* Freeing part of the block. */ head = blksz - off; ASSERT3U(head, >, 0); } blkoff = off; } /* zero out any partial block data at the start of the range */ if (head) { ASSERT3U(blkoff + head, ==, blksz); if (len < head) head = len; if (dbuf_hold_impl(dn, 0, dbuf_whichblock(dn, 0, off), TRUE, FALSE, FTAG, &db) == 0) { caddr_t data; /* don't dirty if it isn't on disk and isn't dirty */ if (db->db_last_dirty || (db->db_blkptr && !BP_IS_HOLE(db->db_blkptr))) { rw_exit(&dn->dn_struct_rwlock); dmu_buf_will_dirty(&db->db, tx); rw_enter(&dn->dn_struct_rwlock, RW_WRITER); data = db->db.db_data; bzero(data + blkoff, head); } dbuf_rele(db, FTAG); } off += head; len -= head; } /* If the range was less than one block, we're done */ if (len == 0) goto out; /* If the remaining range is past end of file, we're done */ if ((off >> blkshift) > dn->dn_maxblkid) goto out; ASSERT(ISP2(blksz)); if (trunc) tail = 0; else tail = P2PHASE(len, blksz); ASSERT0(P2PHASE(off, blksz)); /* zero out any partial block data at the end of the range */ if (tail) { if (len < tail) tail = len; if (dbuf_hold_impl(dn, 0, dbuf_whichblock(dn, 0, off+len), TRUE, FALSE, FTAG, &db) == 0) { /* don't dirty if not on disk and not dirty */ if (db->db_last_dirty || (db->db_blkptr && !BP_IS_HOLE(db->db_blkptr))) { rw_exit(&dn->dn_struct_rwlock); dmu_buf_will_dirty(&db->db, tx); rw_enter(&dn->dn_struct_rwlock, RW_WRITER); bzero(db->db.db_data, tail); } dbuf_rele(db, FTAG); } len -= tail; } /* If the range did not include a full block, we are done */ if (len == 0) goto out; ASSERT(IS_P2ALIGNED(off, blksz)); ASSERT(trunc || IS_P2ALIGNED(len, blksz)); blkid = off >> blkshift; nblks = len >> blkshift; if (trunc) nblks += 1; /* * Dirty all the indirect blocks in this range. Note that only * the first and last indirect blocks can actually be written * (if they were partially freed) -- they must be dirtied, even if * they do not exist on disk yet. The interior blocks will * be freed by free_children(), so they will not actually be written. * Even though these interior blocks will not be written, we * dirty them for two reasons: * * - It ensures that the indirect blocks remain in memory until * syncing context. (They have already been prefetched by * dmu_tx_hold_free(), so we don't have to worry about reading * them serially here.) * * - The dirty space accounting will put pressure on the txg sync * mechanism to begin syncing, and to delay transactions if there * is a large amount of freeing. Even though these indirect * blocks will not be written, we could need to write the same * amount of space if we copy the freed BPs into deadlists. */ if (dn->dn_nlevels > 1) { uint64_t first, last, i, ibyte; int shift, err; first = blkid >> epbs; dnode_dirty_l1(dn, first, tx); if (trunc) last = dn->dn_maxblkid >> epbs; else last = (blkid + nblks - 1) >> epbs; if (last != first) dnode_dirty_l1(dn, last, tx); shift = dn->dn_datablkshift + dn->dn_indblkshift - SPA_BLKPTRSHIFT; for (i = first + 1; i < last; i++) { /* * Set i to the blockid of the next non-hole * level-1 indirect block at or after i. Note * that dnode_next_offset() operates in terms of * level-0-equivalent bytes. */ ibyte = i << shift; err = dnode_next_offset(dn, DNODE_FIND_HAVELOCK, &ibyte, 2, 1, 0); i = ibyte >> shift; if (i >= last) break; /* * Normally we should not see an error, either * from dnode_next_offset() or dbuf_hold_level() * (except for ESRCH from dnode_next_offset). * If there is an i/o error, then when we read * this block in syncing context, it will use * ZIO_FLAG_MUSTSUCCEED, and thus hang/panic according * to the "failmode" property. dnode_next_offset() * doesn't have a flag to indicate MUSTSUCCEED. */ if (err != 0) break; dnode_dirty_l1(dn, i, tx); } } done: /* * Add this range to the dnode range list. * We will finish up this free operation in the syncing phase. */ mutex_enter(&dn->dn_mtx); { int txgoff = tx->tx_txg & TXG_MASK; if (dn->dn_free_ranges[txgoff] == NULL) { dn->dn_free_ranges[txgoff] = range_tree_create(NULL, NULL, &dn->dn_mtx); } range_tree_clear(dn->dn_free_ranges[txgoff], blkid, nblks); range_tree_add(dn->dn_free_ranges[txgoff], blkid, nblks); } dprintf_dnode(dn, "blkid=%llu nblks=%llu txg=%llu\n", blkid, nblks, tx->tx_txg); mutex_exit(&dn->dn_mtx); dbuf_free_range(dn, blkid, blkid + nblks - 1, tx); dnode_setdirty(dn, tx); out: rw_exit(&dn->dn_struct_rwlock); } static boolean_t dnode_spill_freed(dnode_t *dn) { int i; mutex_enter(&dn->dn_mtx); for (i = 0; i < TXG_SIZE; i++) { if (dn->dn_rm_spillblk[i] == DN_KILL_SPILLBLK) break; } mutex_exit(&dn->dn_mtx); return (i < TXG_SIZE); } /* return TRUE if this blkid was freed in a recent txg, or FALSE if it wasn't */ uint64_t dnode_block_freed(dnode_t *dn, uint64_t blkid) { void *dp = spa_get_dsl(dn->dn_objset->os_spa); int i; if (blkid == DMU_BONUS_BLKID) return (FALSE); /* * If we're in the process of opening the pool, dp will not be * set yet, but there shouldn't be anything dirty. */ if (dp == NULL) return (FALSE); if (dn->dn_free_txg) return (TRUE); if (blkid == DMU_SPILL_BLKID) return (dnode_spill_freed(dn)); mutex_enter(&dn->dn_mtx); for (i = 0; i < TXG_SIZE; i++) { if (dn->dn_free_ranges[i] != NULL && range_tree_contains(dn->dn_free_ranges[i], blkid, 1)) break; } mutex_exit(&dn->dn_mtx); return (i < TXG_SIZE); } /* call from syncing context when we actually write/free space for this dnode */ void dnode_diduse_space(dnode_t *dn, int64_t delta) { uint64_t space; dprintf_dnode(dn, "dn=%p dnp=%p used=%llu delta=%lld\n", dn, dn->dn_phys, (u_longlong_t)dn->dn_phys->dn_used, (longlong_t)delta); mutex_enter(&dn->dn_mtx); space = DN_USED_BYTES(dn->dn_phys); if (delta > 0) { ASSERT3U(space + delta, >=, space); /* no overflow */ } else { ASSERT3U(space, >=, -delta); /* no underflow */ } space += delta; if (spa_version(dn->dn_objset->os_spa) < SPA_VERSION_DNODE_BYTES) { ASSERT((dn->dn_phys->dn_flags & DNODE_FLAG_USED_BYTES) == 0); ASSERT0(P2PHASE(space, 1<dn_phys->dn_used = space >> DEV_BSHIFT; } else { dn->dn_phys->dn_used = space; dn->dn_phys->dn_flags |= DNODE_FLAG_USED_BYTES; } mutex_exit(&dn->dn_mtx); } /* * Call when we think we're going to write/free space in open context to track * the amount of memory in use by the currently open txg. */ void dnode_willuse_space(dnode_t *dn, int64_t space, dmu_tx_t *tx) { objset_t *os = dn->dn_objset; dsl_dataset_t *ds = os->os_dsl_dataset; int64_t aspace = spa_get_asize(os->os_spa, space); if (ds != NULL) { dsl_dir_willuse_space(ds->ds_dir, aspace, tx); dsl_pool_dirty_space(dmu_tx_pool(tx), space, tx); } dmu_tx_willuse_space(tx, aspace); } /* * Scans a block at the indicated "level" looking for a hole or data, * depending on 'flags'. * * If level > 0, then we are scanning an indirect block looking at its * pointers. If level == 0, then we are looking at a block of dnodes. * * If we don't find what we are looking for in the block, we return ESRCH. * Otherwise, return with *offset pointing to the beginning (if searching * forwards) or end (if searching backwards) of the range covered by the * block pointer we matched on (or dnode). * * The basic search algorithm used below by dnode_next_offset() is to * use this function to search up the block tree (widen the search) until * we find something (i.e., we don't return ESRCH) and then search back * down the tree (narrow the search) until we reach our original search * level. */ static int dnode_next_offset_level(dnode_t *dn, int flags, uint64_t *offset, int lvl, uint64_t blkfill, uint64_t txg) { dmu_buf_impl_t *db = NULL; void *data = NULL; uint64_t epbs = dn->dn_phys->dn_indblkshift - SPA_BLKPTRSHIFT; uint64_t epb = 1ULL << epbs; uint64_t minfill, maxfill; boolean_t hole; int i, inc, error, span; dprintf("probing object %llu offset %llx level %d of %u\n", dn->dn_object, *offset, lvl, dn->dn_phys->dn_nlevels); hole = ((flags & DNODE_FIND_HOLE) != 0); inc = (flags & DNODE_FIND_BACKWARDS) ? -1 : 1; ASSERT(txg == 0 || !hole); if (lvl == dn->dn_phys->dn_nlevels) { error = 0; epb = dn->dn_phys->dn_nblkptr; data = dn->dn_phys->dn_blkptr; } else { uint64_t blkid = dbuf_whichblock(dn, lvl, *offset); error = dbuf_hold_impl(dn, lvl, blkid, TRUE, FALSE, FTAG, &db); if (error) { if (error != ENOENT) return (error); if (hole) return (0); /* * This can only happen when we are searching up * the block tree for data. We don't really need to * adjust the offset, as we will just end up looking * at the pointer to this block in its parent, and its * going to be unallocated, so we will skip over it. */ return (SET_ERROR(ESRCH)); } error = dbuf_read(db, NULL, DB_RF_CANFAIL | DB_RF_HAVESTRUCT); if (error) { dbuf_rele(db, FTAG); return (error); } data = db->db.db_data; } if (db != NULL && txg != 0 && (db->db_blkptr == NULL || db->db_blkptr->blk_birth <= txg || BP_IS_HOLE(db->db_blkptr))) { /* * This can only happen when we are searching up the tree * and these conditions mean that we need to keep climbing. */ error = SET_ERROR(ESRCH); } else if (lvl == 0) { dnode_phys_t *dnp = data; span = DNODE_SHIFT; ASSERT(dn->dn_type == DMU_OT_DNODE); for (i = (*offset >> span) & (blkfill - 1); i >= 0 && i < blkfill; i += inc) { if ((dnp[i].dn_type == DMU_OT_NONE) == hole) break; *offset += (1ULL << span) * inc; } if (i < 0 || i == blkfill) error = SET_ERROR(ESRCH); } else { blkptr_t *bp = data; uint64_t start = *offset; span = (lvl - 1) * epbs + dn->dn_datablkshift; minfill = 0; maxfill = blkfill << ((lvl - 1) * epbs); if (hole) maxfill--; else minfill++; *offset = *offset >> span; for (i = BF64_GET(*offset, 0, epbs); i >= 0 && i < epb; i += inc) { if (BP_GET_FILL(&bp[i]) >= minfill && BP_GET_FILL(&bp[i]) <= maxfill && (hole || bp[i].blk_birth > txg)) break; if (inc > 0 || *offset > 0) *offset += inc; } *offset = *offset << span; if (inc < 0) { /* traversing backwards; position offset at the end */ ASSERT3U(*offset, <=, start); *offset = MIN(*offset + (1ULL << span) - 1, start); } else if (*offset < start) { *offset = start; } if (i < 0 || i >= epb) error = SET_ERROR(ESRCH); } if (db) dbuf_rele(db, FTAG); return (error); } /* * Find the next hole, data, or sparse region at or after *offset. * The value 'blkfill' tells us how many items we expect to find * in an L0 data block; this value is 1 for normal objects, * DNODES_PER_BLOCK for the meta dnode, and some fraction of * DNODES_PER_BLOCK when searching for sparse regions thereof. * * Examples: * * dnode_next_offset(dn, flags, offset, 1, 1, 0); * Finds the next/previous hole/data in a file. * Used in dmu_offset_next(). * * dnode_next_offset(mdn, flags, offset, 0, DNODES_PER_BLOCK, txg); * Finds the next free/allocated dnode an objset's meta-dnode. * Only finds objects that have new contents since txg (ie. * bonus buffer changes and content removal are ignored). * Used in dmu_object_next(). * * dnode_next_offset(mdn, DNODE_FIND_HOLE, offset, 2, DNODES_PER_BLOCK >> 2, 0); * Finds the next L2 meta-dnode bp that's at most 1/4 full. * Used in dmu_object_alloc(). */ int dnode_next_offset(dnode_t *dn, int flags, uint64_t *offset, int minlvl, uint64_t blkfill, uint64_t txg) { uint64_t initial_offset = *offset; int lvl, maxlvl; int error = 0; if (!(flags & DNODE_FIND_HAVELOCK)) rw_enter(&dn->dn_struct_rwlock, RW_READER); if (dn->dn_phys->dn_nlevels == 0) { error = SET_ERROR(ESRCH); goto out; } if (dn->dn_datablkshift == 0) { if (*offset < dn->dn_datablksz) { if (flags & DNODE_FIND_HOLE) *offset = dn->dn_datablksz; } else { error = SET_ERROR(ESRCH); } goto out; } maxlvl = dn->dn_phys->dn_nlevels; for (lvl = minlvl; lvl <= maxlvl; lvl++) { error = dnode_next_offset_level(dn, flags, offset, lvl, blkfill, txg); if (error != ESRCH) break; } while (error == 0 && --lvl >= minlvl) { error = dnode_next_offset_level(dn, flags, offset, lvl, blkfill, txg); } /* * There's always a "virtual hole" at the end of the object, even * if all BP's which physically exist are non-holes. */ if ((flags & DNODE_FIND_HOLE) && error == ESRCH && txg == 0 && minlvl == 1 && blkfill == 1 && !(flags & DNODE_FIND_BACKWARDS)) { error = 0; } if (error == 0 && (flags & DNODE_FIND_BACKWARDS ? initial_offset < *offset : initial_offset > *offset)) error = SET_ERROR(ESRCH); out: if (!(flags & DNODE_FIND_HAVELOCK)) rw_exit(&dn->dn_struct_rwlock); return (error); }