diff --git a/include/os/freebsd/zfs/sys/zfs_vfsops_os.h b/include/os/freebsd/zfs/sys/zfs_vfsops_os.h index 7f0f24325d59..e7ebcccbe0ce 100644 --- a/include/os/freebsd/zfs/sys/zfs_vfsops_os.h +++ b/include/os/freebsd/zfs/sys/zfs_vfsops_os.h @@ -1,309 +1,310 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2011 Pawel Jakub Dawidek . * All rights reserved. */ #ifndef _SYS_FS_ZFS_VFSOPS_H #define _SYS_FS_ZFS_VFSOPS_H #if __FreeBSD_version >= 1300125 #define TEARDOWN_RMS #endif #if __FreeBSD_version >= 1300109 #define TEARDOWN_INACTIVE_RMS #endif #include #include #include #include #include #include #ifdef TEARDOWN_INACTIVE_RMS #include #endif #include #ifdef __cplusplus extern "C" { #endif #ifdef TEARDOWN_RMS typedef struct rmslock zfs_teardown_lock_t; #else #define zfs_teardown_lock_t rrmlock_t #endif #ifdef TEARDOWN_INACTIVE_RMS typedef struct rmslock zfs_teardown_inactive_lock_t; #else #define zfs_teardown_inactive_lock_t krwlock_t #endif typedef struct zfsvfs zfsvfs_t; struct znode; struct zfsvfs { vfs_t *z_vfs; /* generic fs struct */ zfsvfs_t *z_parent; /* parent fs */ objset_t *z_os; /* objset reference */ uint64_t z_flags; /* super_block flags */ uint64_t z_root; /* id of root znode */ uint64_t z_unlinkedobj; /* id of unlinked zapobj */ uint64_t z_max_blksz; /* maximum block size for files */ uint64_t z_fuid_obj; /* fuid table object number */ uint64_t z_fuid_size; /* fuid table size */ avl_tree_t z_fuid_idx; /* fuid tree keyed by index */ avl_tree_t z_fuid_domain; /* fuid tree keyed by domain */ krwlock_t z_fuid_lock; /* fuid lock */ boolean_t z_fuid_loaded; /* fuid tables are loaded */ boolean_t z_fuid_dirty; /* need to sync fuid table ? */ struct zfs_fuid_info *z_fuid_replay; /* fuid info for replay */ zilog_t *z_log; /* intent log pointer */ uint_t z_acl_type; /* type of acl usable on this fs */ uint_t z_acl_mode; /* acl chmod/mode behavior */ uint_t z_acl_inherit; /* acl inheritance behavior */ zfs_case_t z_case; /* case-sense */ boolean_t z_utf8; /* utf8-only */ int z_norm; /* normalization flags */ boolean_t z_atime; /* enable atimes mount option */ boolean_t z_unmounted; /* unmounted */ zfs_teardown_lock_t z_teardown_lock; zfs_teardown_inactive_lock_t z_teardown_inactive_lock; list_t z_all_znodes; /* all vnodes in the fs */ kmutex_t z_znodes_lock; /* lock for z_all_znodes */ struct zfsctl_root *z_ctldir; /* .zfs directory pointer */ boolean_t z_show_ctldir; /* expose .zfs in the root dir */ boolean_t z_issnap; /* true if this is a snapshot */ boolean_t z_use_fuids; /* version allows fuids */ boolean_t z_replay; /* set during ZIL replay */ boolean_t z_use_sa; /* version allow system attributes */ boolean_t z_xattr_sa; /* allow xattrs to be stores as SA */ boolean_t z_use_namecache; /* make use of FreeBSD name cache */ uint8_t z_xattr; /* xattr type in use */ uint64_t z_version; /* ZPL version */ uint64_t z_shares_dir; /* hidden shares dir */ dataset_kstats_t z_kstat; /* fs kstats */ kmutex_t z_lock; uint64_t z_userquota_obj; uint64_t z_groupquota_obj; uint64_t z_userobjquota_obj; uint64_t z_groupobjquota_obj; uint64_t z_projectquota_obj; uint64_t z_projectobjquota_obj; uint64_t z_replay_eof; /* New end of file - replay only */ sa_attr_type_t *z_attr_table; /* SA attr mapping->id */ #define ZFS_OBJ_MTX_SZ 64 kmutex_t z_hold_mtx[ZFS_OBJ_MTX_SZ]; /* znode hold locks */ struct task z_unlinked_drain_task; }; #ifdef TEARDOWN_RMS #define ZFS_TEARDOWN_INIT(zfsvfs) \ rms_init(&(zfsvfs)->z_teardown_lock, "zfs teardown") #define ZFS_TEARDOWN_DESTROY(zfsvfs) \ rms_destroy(&(zfsvfs)->z_teardown_lock) #define ZFS_TEARDOWN_ENTER_READ(zfsvfs, tag) \ rms_rlock(&(zfsvfs)->z_teardown_lock); #define ZFS_TEARDOWN_EXIT_READ(zfsvfs, tag) \ rms_runlock(&(zfsvfs)->z_teardown_lock) #define ZFS_TEARDOWN_ENTER_WRITE(zfsvfs, tag) \ rms_wlock(&(zfsvfs)->z_teardown_lock) #define ZFS_TEARDOWN_EXIT_WRITE(zfsvfs) \ rms_wunlock(&(zfsvfs)->z_teardown_lock) #define ZFS_TEARDOWN_EXIT(zfsvfs, tag) \ rms_unlock(&(zfsvfs)->z_teardown_lock) #define ZFS_TEARDOWN_READ_HELD(zfsvfs) \ rms_rowned(&(zfsvfs)->z_teardown_lock) #define ZFS_TEARDOWN_WRITE_HELD(zfsvfs) \ rms_wowned(&(zfsvfs)->z_teardown_lock) #define ZFS_TEARDOWN_HELD(zfsvfs) \ rms_owned_any(&(zfsvfs)->z_teardown_lock) #else #define ZFS_TEARDOWN_INIT(zfsvfs) \ rrm_init(&(zfsvfs)->z_teardown_lock, B_FALSE) #define ZFS_TEARDOWN_DESTROY(zfsvfs) \ rrm_destroy(&(zfsvfs)->z_teardown_lock) #define ZFS_TEARDOWN_ENTER_READ(zfsvfs, tag) \ rrm_enter_read(&(zfsvfs)->z_teardown_lock, tag); #define ZFS_TEARDOWN_EXIT_READ(zfsvfs, tag) \ rrm_exit(&(zfsvfs)->z_teardown_lock, tag) #define ZFS_TEARDOWN_ENTER_WRITE(zfsvfs, tag) \ rrm_enter(&(zfsvfs)->z_teardown_lock, RW_WRITER, tag) #define ZFS_TEARDOWN_EXIT_WRITE(zfsvfs) \ rrm_exit(&(zfsvfs)->z_teardown_lock, tag) #define ZFS_TEARDOWN_EXIT(zfsvfs, tag) \ rrm_exit(&(zfsvfs)->z_teardown_lock, tag) #define ZFS_TEARDOWN_READ_HELD(zfsvfs) \ RRM_READ_HELD(&(zfsvfs)->z_teardown_lock) #define ZFS_TEARDOWN_WRITE_HELD(zfsvfs) \ RRM_WRITE_HELD(&(zfsvfs)->z_teardown_lock) #define ZFS_TEARDOWN_HELD(zfsvfs) \ RRM_LOCK_HELD(&(zfsvfs)->z_teardown_lock) #endif #ifdef TEARDOWN_INACTIVE_RMS #define ZFS_TEARDOWN_INACTIVE_INIT(zfsvfs) \ rms_init(&(zfsvfs)->z_teardown_inactive_lock, "zfs teardown inactive") #define ZFS_TEARDOWN_INACTIVE_DESTROY(zfsvfs) \ rms_destroy(&(zfsvfs)->z_teardown_inactive_lock) #define ZFS_TEARDOWN_INACTIVE_TRY_ENTER_READ(zfsvfs) \ rms_try_rlock(&(zfsvfs)->z_teardown_inactive_lock) #define ZFS_TEARDOWN_INACTIVE_ENTER_READ(zfsvfs) \ rms_rlock(&(zfsvfs)->z_teardown_inactive_lock) #define ZFS_TEARDOWN_INACTIVE_EXIT_READ(zfsvfs) \ rms_runlock(&(zfsvfs)->z_teardown_inactive_lock) #define ZFS_TEARDOWN_INACTIVE_ENTER_WRITE(zfsvfs) \ rms_wlock(&(zfsvfs)->z_teardown_inactive_lock) #define ZFS_TEARDOWN_INACTIVE_EXIT_WRITE(zfsvfs) \ rms_wunlock(&(zfsvfs)->z_teardown_inactive_lock) #define ZFS_TEARDOWN_INACTIVE_WRITE_HELD(zfsvfs) \ rms_wowned(&(zfsvfs)->z_teardown_inactive_lock) #else #define ZFS_TEARDOWN_INACTIVE_INIT(zfsvfs) \ rw_init(&(zfsvfs)->z_teardown_inactive_lock, NULL, RW_DEFAULT, NULL) #define ZFS_TEARDOWN_INACTIVE_DESTROY(zfsvfs) \ rw_destroy(&(zfsvfs)->z_teardown_inactive_lock) #define ZFS_TEARDOWN_INACTIVE_TRY_ENTER_READ(zfsvfs) \ rw_tryenter(&(zfsvfs)->z_teardown_inactive_lock, RW_READER) #define ZFS_TEARDOWN_INACTIVE_ENTER_READ(zfsvfs) \ rw_enter(&(zfsvfs)->z_teardown_inactive_lock, RW_READER) #define ZFS_TEARDOWN_INACTIVE_EXIT_READ(zfsvfs) \ rw_exit(&(zfsvfs)->z_teardown_inactive_lock) #define ZFS_TEARDOWN_INACTIVE_ENTER_WRITE(zfsvfs) \ rw_enter(&(zfsvfs)->z_teardown_inactive_lock, RW_WRITER) #define ZFS_TEARDOWN_INACTIVE_EXIT_WRITE(zfsvfs) \ rw_exit(&(zfsvfs)->z_teardown_inactive_lock) #define ZFS_TEARDOWN_INACTIVE_WRITE_HELD(zfsvfs) \ RW_WRITE_HELD(&(zfsvfs)->z_teardown_inactive_lock) #endif #define ZSB_XATTR 0x0001 /* Enable user xattrs */ /* * Normal filesystems (those not under .zfs/snapshot) have a total * file ID size limited to 12 bytes (including the length field) due to * NFSv2 protocol's limitation of 32 bytes for a filehandle. For historical * reasons, this same limit is being imposed by the Solaris NFSv3 implementation * (although the NFSv3 protocol actually permits a maximum of 64 bytes). It * is not possible to expand beyond 12 bytes without abandoning support * of NFSv2. * * For normal filesystems, we partition up the available space as follows: * 2 bytes fid length (required) * 6 bytes object number (48 bits) * 4 bytes generation number (32 bits) * * We reserve only 48 bits for the object number, as this is the limit * currently defined and imposed by the DMU. */ typedef struct zfid_short { uint16_t zf_len; uint8_t zf_object[6]; /* obj[i] = obj >> (8 * i) */ uint8_t zf_gen[4]; /* gen[i] = gen >> (8 * i) */ } zfid_short_t; /* * Filesystems under .zfs/snapshot have a total file ID size of 22[*] bytes * (including the length field). This makes files under .zfs/snapshot * accessible by NFSv3 and NFSv4, but not NFSv2. * * For files under .zfs/snapshot, we partition up the available space * as follows: * 2 bytes fid length (required) * 6 bytes object number (48 bits) * 4 bytes generation number (32 bits) * 6 bytes objset id (48 bits) * 4 bytes[**] currently just zero (32 bits) * * We reserve only 48 bits for the object number and objset id, as these are * the limits currently defined and imposed by the DMU. * * [*] 20 bytes on FreeBSD to fit into the size of struct fid. * [**] 2 bytes on FreeBSD for the above reason. */ typedef struct zfid_long { zfid_short_t z_fid; uint8_t zf_setid[6]; /* obj[i] = obj >> (8 * i) */ uint8_t zf_setgen[2]; /* gen[i] = gen >> (8 * i) */ } zfid_long_t; #define SHORT_FID_LEN (sizeof (zfid_short_t) - sizeof (uint16_t)) #define LONG_FID_LEN (sizeof (zfid_long_t) - sizeof (uint16_t)) extern int zfs_super_owner; +extern int zfs_bclone_enabled; extern void zfs_init(void); extern void zfs_fini(void); extern int zfs_suspend_fs(zfsvfs_t *zfsvfs); extern int zfs_resume_fs(zfsvfs_t *zfsvfs, struct dsl_dataset *ds); extern int zfs_end_fs(zfsvfs_t *zfsvfs, struct dsl_dataset *ds); extern int zfs_set_version(zfsvfs_t *zfsvfs, uint64_t newvers); extern int zfsvfs_create(const char *name, boolean_t readonly, zfsvfs_t **zfvp); extern int zfsvfs_create_impl(zfsvfs_t **zfvp, zfsvfs_t *zfsvfs, objset_t *os); extern void zfsvfs_free(zfsvfs_t *zfsvfs); extern int zfs_check_global_label(const char *dsname, const char *hexsl); extern boolean_t zfs_is_readonly(zfsvfs_t *zfsvfs); extern int zfs_get_temporary_prop(struct dsl_dataset *ds, zfs_prop_t zfs_prop, uint64_t *val, char *setpoint); extern int zfs_busy(void); #ifdef __cplusplus } #endif #endif /* _SYS_FS_ZFS_VFSOPS_H */ diff --git a/include/os/linux/zfs/sys/zfs_vfsops_os.h b/include/os/linux/zfs/sys/zfs_vfsops_os.h index b4d5db21f5e5..220466550258 100644 --- a/include/os/linux/zfs/sys/zfs_vfsops_os.h +++ b/include/os/linux/zfs/sys/zfs_vfsops_os.h @@ -1,255 +1,257 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2013, 2018 by Delphix. All rights reserved. */ #ifndef _SYS_FS_ZFS_VFSOPS_H #define _SYS_FS_ZFS_VFSOPS_H #include #include #include #include #include #include #include #include #include #include #include #ifdef __cplusplus extern "C" { #endif typedef struct zfsvfs zfsvfs_t; struct znode; +extern int zfs_bclone_enabled; + /* * This structure emulates the vfs_t from other platforms. It's purpose * is to facilitate the handling of mount options and minimize structural * differences between the platforms. */ typedef struct vfs { struct zfsvfs *vfs_data; char *vfs_mntpoint; /* Primary mount point */ uint64_t vfs_xattr; boolean_t vfs_readonly; boolean_t vfs_do_readonly; boolean_t vfs_setuid; boolean_t vfs_do_setuid; boolean_t vfs_exec; boolean_t vfs_do_exec; boolean_t vfs_devices; boolean_t vfs_do_devices; boolean_t vfs_do_xattr; boolean_t vfs_atime; boolean_t vfs_do_atime; boolean_t vfs_relatime; boolean_t vfs_do_relatime; boolean_t vfs_nbmand; boolean_t vfs_do_nbmand; } vfs_t; typedef struct zfs_mnt { const char *mnt_osname; /* Objset name */ char *mnt_data; /* Raw mount options */ } zfs_mnt_t; struct zfsvfs { vfs_t *z_vfs; /* generic fs struct */ struct super_block *z_sb; /* generic super_block */ struct zfsvfs *z_parent; /* parent fs */ objset_t *z_os; /* objset reference */ uint64_t z_flags; /* super_block flags */ uint64_t z_root; /* id of root znode */ uint64_t z_unlinkedobj; /* id of unlinked zapobj */ uint64_t z_max_blksz; /* maximum block size for files */ uint64_t z_fuid_obj; /* fuid table object number */ uint64_t z_fuid_size; /* fuid table size */ avl_tree_t z_fuid_idx; /* fuid tree keyed by index */ avl_tree_t z_fuid_domain; /* fuid tree keyed by domain */ krwlock_t z_fuid_lock; /* fuid lock */ boolean_t z_fuid_loaded; /* fuid tables are loaded */ boolean_t z_fuid_dirty; /* need to sync fuid table ? */ struct zfs_fuid_info *z_fuid_replay; /* fuid info for replay */ zilog_t *z_log; /* intent log pointer */ uint_t z_acl_mode; /* acl chmod/mode behavior */ uint_t z_acl_inherit; /* acl inheritance behavior */ uint_t z_acl_type; /* type of ACL usable on this FS */ zfs_case_t z_case; /* case-sense */ boolean_t z_utf8; /* utf8-only */ int z_norm; /* normalization flags */ boolean_t z_relatime; /* enable relatime mount option */ boolean_t z_unmounted; /* unmounted */ rrmlock_t z_teardown_lock; krwlock_t z_teardown_inactive_lock; list_t z_all_znodes; /* all znodes in the fs */ unsigned long z_rollback_time; /* last online rollback time */ unsigned long z_snap_defer_time; /* last snapshot unmount deferral */ kmutex_t z_znodes_lock; /* lock for z_all_znodes */ arc_prune_t *z_arc_prune; /* called by ARC to prune caches */ struct inode *z_ctldir; /* .zfs directory inode */ boolean_t z_show_ctldir; /* expose .zfs in the root dir */ boolean_t z_issnap; /* true if this is a snapshot */ boolean_t z_use_fuids; /* version allows fuids */ boolean_t z_replay; /* set during ZIL replay */ boolean_t z_use_sa; /* version allow system attributes */ boolean_t z_xattr_sa; /* allow xattrs to be stores as SA */ boolean_t z_draining; /* is true when drain is active */ boolean_t z_drain_cancel; /* signal the unlinked drain to stop */ uint64_t z_version; /* ZPL version */ uint64_t z_shares_dir; /* hidden shares dir */ dataset_kstats_t z_kstat; /* fs kstats */ kmutex_t z_lock; uint64_t z_userquota_obj; uint64_t z_groupquota_obj; uint64_t z_userobjquota_obj; uint64_t z_groupobjquota_obj; uint64_t z_projectquota_obj; uint64_t z_projectobjquota_obj; uint64_t z_replay_eof; /* New end of file - replay only */ sa_attr_type_t *z_attr_table; /* SA attr mapping->id */ uint64_t z_hold_size; /* znode hold array size */ avl_tree_t *z_hold_trees; /* znode hold trees */ kmutex_t *z_hold_locks; /* znode hold locks */ taskqid_t z_drain_task; /* task id for the unlink drain task */ }; #define ZFS_TEARDOWN_INIT(zfsvfs) \ rrm_init(&(zfsvfs)->z_teardown_lock, B_FALSE) #define ZFS_TEARDOWN_DESTROY(zfsvfs) \ rrm_destroy(&(zfsvfs)->z_teardown_lock) #define ZFS_TEARDOWN_ENTER_READ(zfsvfs, tag) \ rrm_enter_read(&(zfsvfs)->z_teardown_lock, tag); #define ZFS_TEARDOWN_EXIT_READ(zfsvfs, tag) \ rrm_exit(&(zfsvfs)->z_teardown_lock, tag) #define ZFS_TEARDOWN_ENTER_WRITE(zfsvfs, tag) \ rrm_enter(&(zfsvfs)->z_teardown_lock, RW_WRITER, tag) #define ZFS_TEARDOWN_EXIT_WRITE(zfsvfs) \ rrm_exit(&(zfsvfs)->z_teardown_lock, tag) #define ZFS_TEARDOWN_EXIT(zfsvfs, tag) \ rrm_exit(&(zfsvfs)->z_teardown_lock, tag) #define ZFS_TEARDOWN_READ_HELD(zfsvfs) \ RRM_READ_HELD(&(zfsvfs)->z_teardown_lock) #define ZFS_TEARDOWN_WRITE_HELD(zfsvfs) \ RRM_WRITE_HELD(&(zfsvfs)->z_teardown_lock) #define ZFS_TEARDOWN_HELD(zfsvfs) \ RRM_LOCK_HELD(&(zfsvfs)->z_teardown_lock) #define ZSB_XATTR 0x0001 /* Enable user xattrs */ /* * Allow a maximum number of links. While ZFS does not internally limit * this the inode->i_nlink member is defined as an unsigned int. To be * safe we use 2^31-1 as the limit. */ #define ZFS_LINK_MAX ((1U << 31) - 1U) /* * Normal filesystems (those not under .zfs/snapshot) have a total * file ID size limited to 12 bytes (including the length field) due to * NFSv2 protocol's limitation of 32 bytes for a filehandle. For historical * reasons, this same limit is being imposed by the Solaris NFSv3 implementation * (although the NFSv3 protocol actually permits a maximum of 64 bytes). It * is not possible to expand beyond 12 bytes without abandoning support * of NFSv2. * * For normal filesystems, we partition up the available space as follows: * 2 bytes fid length (required) * 6 bytes object number (48 bits) * 4 bytes generation number (32 bits) * * We reserve only 48 bits for the object number, as this is the limit * currently defined and imposed by the DMU. */ typedef struct zfid_short { uint16_t zf_len; uint8_t zf_object[6]; /* obj[i] = obj >> (8 * i) */ uint8_t zf_gen[4]; /* gen[i] = gen >> (8 * i) */ } zfid_short_t; /* * Filesystems under .zfs/snapshot have a total file ID size of 22 bytes * (including the length field). This makes files under .zfs/snapshot * accessible by NFSv3 and NFSv4, but not NFSv2. * * For files under .zfs/snapshot, we partition up the available space * as follows: * 2 bytes fid length (required) * 6 bytes object number (48 bits) * 4 bytes generation number (32 bits) * 6 bytes objset id (48 bits) * 4 bytes currently just zero (32 bits) * * We reserve only 48 bits for the object number and objset id, as these are * the limits currently defined and imposed by the DMU. */ typedef struct zfid_long { zfid_short_t z_fid; uint8_t zf_setid[6]; /* obj[i] = obj >> (8 * i) */ uint8_t zf_setgen[4]; /* gen[i] = gen >> (8 * i) */ } zfid_long_t; #define SHORT_FID_LEN (sizeof (zfid_short_t) - sizeof (uint16_t)) #define LONG_FID_LEN (sizeof (zfid_long_t) - sizeof (uint16_t)) extern void zfs_init(void); extern void zfs_fini(void); extern int zfs_suspend_fs(zfsvfs_t *zfsvfs); extern int zfs_resume_fs(zfsvfs_t *zfsvfs, struct dsl_dataset *ds); extern int zfs_end_fs(zfsvfs_t *zfsvfs, struct dsl_dataset *ds); extern void zfs_exit_fs(zfsvfs_t *zfsvfs); extern int zfs_set_version(zfsvfs_t *zfsvfs, uint64_t newvers); extern int zfsvfs_create(const char *name, boolean_t readony, zfsvfs_t **zfvp); extern int zfsvfs_create_impl(zfsvfs_t **zfvp, zfsvfs_t *zfsvfs, objset_t *os); extern void zfsvfs_free(zfsvfs_t *zfsvfs); extern int zfs_check_global_label(const char *dsname, const char *hexsl); extern boolean_t zfs_is_readonly(zfsvfs_t *zfsvfs); extern int zfs_domount(struct super_block *sb, zfs_mnt_t *zm, int silent); extern void zfs_preumount(struct super_block *sb); extern int zfs_umount(struct super_block *sb); extern int zfs_remount(struct super_block *sb, int *flags, zfs_mnt_t *zm); extern int zfs_statvfs(struct inode *ip, struct kstatfs *statp); extern int zfs_vget(struct super_block *sb, struct inode **ipp, fid_t *fidp); extern int zfs_prune(struct super_block *sb, unsigned long nr_to_scan, int *objects); extern int zfs_get_temporary_prop(dsl_dataset_t *ds, zfs_prop_t zfs_prop, uint64_t *val, char *setpoint); #ifdef __cplusplus } #endif #endif /* _SYS_FS_ZFS_VFSOPS_H */ diff --git a/man/man4/zfs.4 b/man/man4/zfs.4 index 4a0441ed9469..5daf27e9d536 100644 --- a/man/man4/zfs.4 +++ b/man/man4/zfs.4 @@ -1,2615 +1,2620 @@ .\" .\" Copyright (c) 2013 by Turbo Fredriksson . All rights reserved. .\" Copyright (c) 2019, 2021 by Delphix. All rights reserved. .\" Copyright (c) 2019 Datto Inc. .\" The contents of this file are subject to the terms of the Common Development .\" and Distribution License (the "License"). You may not use this file except .\" in compliance with the License. You can obtain a copy of the license at .\" usr/src/OPENSOLARIS.LICENSE or https://opensource.org/licenses/CDDL-1.0. .\" .\" See the License for the specific language governing permissions and .\" limitations under the License. When distributing Covered Code, include this .\" CDDL HEADER in each file and include the License file at .\" usr/src/OPENSOLARIS.LICENSE. If applicable, add the following below this .\" CDDL HEADER, with the fields enclosed by brackets "[]" replaced with your .\" own identifying information: .\" Portions Copyright [yyyy] [name of copyright owner] .\" .Dd July 21, 2023 .Dt ZFS 4 .Os . .Sh NAME .Nm zfs .Nd tuning of the ZFS kernel module . .Sh DESCRIPTION The ZFS module supports these parameters: .Bl -tag -width Ds .It Sy dbuf_cache_max_bytes Ns = Ns Sy UINT64_MAX Ns B Pq u64 Maximum size in bytes of the dbuf cache. The target size is determined by the MIN versus .No 1/2^ Ns Sy dbuf_cache_shift Pq 1/32nd of the target ARC size. The behavior of the dbuf cache and its associated settings can be observed via the .Pa /proc/spl/kstat/zfs/dbufstats kstat. . .It Sy dbuf_metadata_cache_max_bytes Ns = Ns Sy UINT64_MAX Ns B Pq u64 Maximum size in bytes of the metadata dbuf cache. The target size is determined by the MIN versus .No 1/2^ Ns Sy dbuf_metadata_cache_shift Pq 1/64th of the target ARC size. The behavior of the metadata dbuf cache and its associated settings can be observed via the .Pa /proc/spl/kstat/zfs/dbufstats kstat. . .It Sy dbuf_cache_hiwater_pct Ns = Ns Sy 10 Ns % Pq uint The percentage over .Sy dbuf_cache_max_bytes when dbufs must be evicted directly. . .It Sy dbuf_cache_lowater_pct Ns = Ns Sy 10 Ns % Pq uint The percentage below .Sy dbuf_cache_max_bytes when the evict thread stops evicting dbufs. . .It Sy dbuf_cache_shift Ns = Ns Sy 5 Pq uint Set the size of the dbuf cache .Pq Sy dbuf_cache_max_bytes to a log2 fraction of the target ARC size. . .It Sy dbuf_metadata_cache_shift Ns = Ns Sy 6 Pq uint Set the size of the dbuf metadata cache .Pq Sy dbuf_metadata_cache_max_bytes to a log2 fraction of the target ARC size. . .It Sy dbuf_mutex_cache_shift Ns = Ns Sy 0 Pq uint Set the size of the mutex array for the dbuf cache. When set to .Sy 0 the array is dynamically sized based on total system memory. . .It Sy dmu_object_alloc_chunk_shift Ns = Ns Sy 7 Po 128 Pc Pq uint dnode slots allocated in a single operation as a power of 2. The default value minimizes lock contention for the bulk operation performed. . .It Sy dmu_prefetch_max Ns = Ns Sy 134217728 Ns B Po 128 MiB Pc Pq uint Limit the amount we can prefetch with one call to this amount in bytes. This helps to limit the amount of memory that can be used by prefetching. . .It Sy ignore_hole_birth Pq int Alias for .Sy send_holes_without_birth_time . . .It Sy l2arc_feed_again Ns = Ns Sy 1 Ns | Ns 0 Pq int Turbo L2ARC warm-up. When the L2ARC is cold the fill interval will be set as fast as possible. . .It Sy l2arc_feed_min_ms Ns = Ns Sy 200 Pq u64 Min feed interval in milliseconds. Requires .Sy l2arc_feed_again Ns = Ns Ar 1 and only applicable in related situations. . .It Sy l2arc_feed_secs Ns = Ns Sy 1 Pq u64 Seconds between L2ARC writing. . .It Sy l2arc_headroom Ns = Ns Sy 8 Pq u64 How far through the ARC lists to search for L2ARC cacheable content, expressed as a multiplier of .Sy l2arc_write_max . ARC persistence across reboots can be achieved with persistent L2ARC by setting this parameter to .Sy 0 , allowing the full length of ARC lists to be searched for cacheable content. . .It Sy l2arc_headroom_boost Ns = Ns Sy 200 Ns % Pq u64 Scales .Sy l2arc_headroom by this percentage when L2ARC contents are being successfully compressed before writing. A value of .Sy 100 disables this feature. . .It Sy l2arc_exclude_special Ns = Ns Sy 0 Ns | Ns 1 Pq int Controls whether buffers present on special vdevs are eligible for caching into L2ARC. If set to 1, exclude dbufs on special vdevs from being cached to L2ARC. . .It Sy l2arc_mfuonly Ns = Ns Sy 0 Ns | Ns 1 Pq int Controls whether only MFU metadata and data are cached from ARC into L2ARC. This may be desired to avoid wasting space on L2ARC when reading/writing large amounts of data that are not expected to be accessed more than once. .Pp The default is off, meaning both MRU and MFU data and metadata are cached. When turning off this feature, some MRU buffers will still be present in ARC and eventually cached on L2ARC. .No If Sy l2arc_noprefetch Ns = Ns Sy 0 , some prefetched buffers will be cached to L2ARC, and those might later transition to MRU, in which case the .Sy l2arc_mru_asize No arcstat will not be Sy 0 . .Pp Regardless of .Sy l2arc_noprefetch , some MFU buffers might be evicted from ARC, accessed later on as prefetches and transition to MRU as prefetches. If accessed again they are counted as MRU and the .Sy l2arc_mru_asize No arcstat will not be Sy 0 . .Pp The ARC status of L2ARC buffers when they were first cached in L2ARC can be seen in the .Sy l2arc_mru_asize , Sy l2arc_mfu_asize , No and Sy l2arc_prefetch_asize arcstats when importing the pool or onlining a cache device if persistent L2ARC is enabled. .Pp The .Sy evict_l2_eligible_mru arcstat does not take into account if this option is enabled as the information provided by the .Sy evict_l2_eligible_m[rf]u arcstats can be used to decide if toggling this option is appropriate for the current workload. . .It Sy l2arc_meta_percent Ns = Ns Sy 33 Ns % Pq uint Percent of ARC size allowed for L2ARC-only headers. Since L2ARC buffers are not evicted on memory pressure, too many headers on a system with an irrationally large L2ARC can render it slow or unusable. This parameter limits L2ARC writes and rebuilds to achieve the target. . .It Sy l2arc_trim_ahead Ns = Ns Sy 0 Ns % Pq u64 Trims ahead of the current write size .Pq Sy l2arc_write_max on L2ARC devices by this percentage of write size if we have filled the device. If set to .Sy 100 we TRIM twice the space required to accommodate upcoming writes. A minimum of .Sy 64 MiB will be trimmed. It also enables TRIM of the whole L2ARC device upon creation or addition to an existing pool or if the header of the device is invalid upon importing a pool or onlining a cache device. A value of .Sy 0 disables TRIM on L2ARC altogether and is the default as it can put significant stress on the underlying storage devices. This will vary depending of how well the specific device handles these commands. . .It Sy l2arc_noprefetch Ns = Ns Sy 1 Ns | Ns 0 Pq int Do not write buffers to L2ARC if they were prefetched but not used by applications. In case there are prefetched buffers in L2ARC and this option is later set, we do not read the prefetched buffers from L2ARC. Unsetting this option is useful for caching sequential reads from the disks to L2ARC and serve those reads from L2ARC later on. This may be beneficial in case the L2ARC device is significantly faster in sequential reads than the disks of the pool. .Pp Use .Sy 1 to disable and .Sy 0 to enable caching/reading prefetches to/from L2ARC. . .It Sy l2arc_norw Ns = Ns Sy 0 Ns | Ns 1 Pq int No reads during writes. . .It Sy l2arc_write_boost Ns = Ns Sy 33554432 Ns B Po 32 MiB Pc Pq u64 Cold L2ARC devices will have .Sy l2arc_write_max increased by this amount while they remain cold. . .It Sy l2arc_write_max Ns = Ns Sy 33554432 Ns B Po 32 MiB Pc Pq u64 Max write bytes per interval. . .It Sy l2arc_rebuild_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Rebuild the L2ARC when importing a pool (persistent L2ARC). This can be disabled if there are problems importing a pool or attaching an L2ARC device (e.g. the L2ARC device is slow in reading stored log metadata, or the metadata has become somehow fragmented/unusable). . .It Sy l2arc_rebuild_blocks_min_l2size Ns = Ns Sy 1073741824 Ns B Po 1 GiB Pc Pq u64 Mininum size of an L2ARC device required in order to write log blocks in it. The log blocks are used upon importing the pool to rebuild the persistent L2ARC. .Pp For L2ARC devices less than 1 GiB, the amount of data .Fn l2arc_evict evicts is significant compared to the amount of restored L2ARC data. In this case, do not write log blocks in L2ARC in order not to waste space. . .It Sy metaslab_aliquot Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64 Metaslab granularity, in bytes. This is roughly similar to what would be referred to as the "stripe size" in traditional RAID arrays. In normal operation, ZFS will try to write this amount of data to each disk before moving on to the next top-level vdev. . .It Sy metaslab_bias_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable metaslab group biasing based on their vdevs' over- or under-utilization relative to the pool. . .It Sy metaslab_force_ganging Ns = Ns Sy 16777217 Ns B Po 16 MiB + 1 B Pc Pq u64 Make some blocks above a certain size be gang blocks. This option is used by the test suite to facilitate testing. . .It Sy metaslab_force_ganging_pct Ns = Ns Sy 3 Ns % Pq uint For blocks that could be forced to be a gang block (due to .Sy metaslab_force_ganging ) , force this many of them to be gang blocks. . .It Sy zfs_ddt_zap_default_bs Ns = Ns Sy 15 Po 32 KiB Pc Pq int Default DDT ZAP data block size as a power of 2. Note that changing this after creating a DDT on the pool will not affect existing DDTs, only newly created ones. . .It Sy zfs_ddt_zap_default_ibs Ns = Ns Sy 15 Po 32 KiB Pc Pq int Default DDT ZAP indirect block size as a power of 2. Note that changing this after creating a DDT on the pool will not affect existing DDTs, only newly created ones. . .It Sy zfs_default_bs Ns = Ns Sy 9 Po 512 B Pc Pq int Default dnode block size as a power of 2. . .It Sy zfs_default_ibs Ns = Ns Sy 17 Po 128 KiB Pc Pq int Default dnode indirect block size as a power of 2. . .It Sy zfs_history_output_max Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64 When attempting to log an output nvlist of an ioctl in the on-disk history, the output will not be stored if it is larger than this size (in bytes). This must be less than .Sy DMU_MAX_ACCESS Pq 64 MiB . This applies primarily to .Fn zfs_ioc_channel_program Pq cf. Xr zfs-program 8 . . .It Sy zfs_keep_log_spacemaps_at_export Ns = Ns Sy 0 Ns | Ns 1 Pq int Prevent log spacemaps from being destroyed during pool exports and destroys. . .It Sy zfs_metaslab_segment_weight_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable/disable segment-based metaslab selection. . .It Sy zfs_metaslab_switch_threshold Ns = Ns Sy 2 Pq int When using segment-based metaslab selection, continue allocating from the active metaslab until this option's worth of buckets have been exhausted. . .It Sy metaslab_debug_load Ns = Ns Sy 0 Ns | Ns 1 Pq int Load all metaslabs during pool import. . .It Sy metaslab_debug_unload Ns = Ns Sy 0 Ns | Ns 1 Pq int Prevent metaslabs from being unloaded. . .It Sy metaslab_fragmentation_factor_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable use of the fragmentation metric in computing metaslab weights. . .It Sy metaslab_df_max_search Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq uint Maximum distance to search forward from the last offset. Without this limit, fragmented pools can see .Em >100`000 iterations and .Fn metaslab_block_picker becomes the performance limiting factor on high-performance storage. .Pp With the default setting of .Sy 16 MiB , we typically see less than .Em 500 iterations, even with very fragmented .Sy ashift Ns = Ns Sy 9 pools. The maximum number of iterations possible is .Sy metaslab_df_max_search / 2^(ashift+1) . With the default setting of .Sy 16 MiB this is .Em 16*1024 Pq with Sy ashift Ns = Ns Sy 9 or .Em 2*1024 Pq with Sy ashift Ns = Ns Sy 12 . . .It Sy metaslab_df_use_largest_segment Ns = Ns Sy 0 Ns | Ns 1 Pq int If not searching forward (due to .Sy metaslab_df_max_search , metaslab_df_free_pct , .No or Sy metaslab_df_alloc_threshold ) , this tunable controls which segment is used. If set, we will use the largest free segment. If unset, we will use a segment of at least the requested size. . .It Sy zfs_metaslab_max_size_cache_sec Ns = Ns Sy 3600 Ns s Po 1 hour Pc Pq u64 When we unload a metaslab, we cache the size of the largest free chunk. We use that cached size to determine whether or not to load a metaslab for a given allocation. As more frees accumulate in that metaslab while it's unloaded, the cached max size becomes less and less accurate. After a number of seconds controlled by this tunable, we stop considering the cached max size and start considering only the histogram instead. . .It Sy zfs_metaslab_mem_limit Ns = Ns Sy 25 Ns % Pq uint When we are loading a new metaslab, we check the amount of memory being used to store metaslab range trees. If it is over a threshold, we attempt to unload the least recently used metaslab to prevent the system from clogging all of its memory with range trees. This tunable sets the percentage of total system memory that is the threshold. . .It Sy zfs_metaslab_try_hard_before_gang Ns = Ns Sy 0 Ns | Ns 1 Pq int .Bl -item -compact .It If unset, we will first try normal allocation. .It If that fails then we will do a gang allocation. .It If that fails then we will do a "try hard" gang allocation. .It If that fails then we will have a multi-layer gang block. .El .Pp .Bl -item -compact .It If set, we will first try normal allocation. .It If that fails then we will do a "try hard" allocation. .It If that fails we will do a gang allocation. .It If that fails we will do a "try hard" gang allocation. .It If that fails then we will have a multi-layer gang block. .El . .It Sy zfs_metaslab_find_max_tries Ns = Ns Sy 100 Pq uint When not trying hard, we only consider this number of the best metaslabs. This improves performance, especially when there are many metaslabs per vdev and the allocation can't actually be satisfied (so we would otherwise iterate all metaslabs). . .It Sy zfs_vdev_default_ms_count Ns = Ns Sy 200 Pq uint When a vdev is added, target this number of metaslabs per top-level vdev. . .It Sy zfs_vdev_default_ms_shift Ns = Ns Sy 29 Po 512 MiB Pc Pq uint Default lower limit for metaslab size. . .It Sy zfs_vdev_max_ms_shift Ns = Ns Sy 34 Po 16 GiB Pc Pq uint Default upper limit for metaslab size. . .It Sy zfs_vdev_max_auto_ashift Ns = Ns Sy 14 Pq uint Maximum ashift used when optimizing for logical \[->] physical sector size on new top-level vdevs. May be increased up to .Sy ASHIFT_MAX Po 16 Pc , but this may negatively impact pool space efficiency. . .It Sy zfs_vdev_min_auto_ashift Ns = Ns Sy ASHIFT_MIN Po 9 Pc Pq uint Minimum ashift used when creating new top-level vdevs. . .It Sy zfs_vdev_min_ms_count Ns = Ns Sy 16 Pq uint Minimum number of metaslabs to create in a top-level vdev. . .It Sy vdev_validate_skip Ns = Ns Sy 0 Ns | Ns 1 Pq int Skip label validation steps during pool import. Changing is not recommended unless you know what you're doing and are recovering a damaged label. . .It Sy zfs_vdev_ms_count_limit Ns = Ns Sy 131072 Po 128k Pc Pq uint Practical upper limit of total metaslabs per top-level vdev. . .It Sy metaslab_preload_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable metaslab group preloading. . .It Sy metaslab_preload_limit Ns = Ns Sy 10 Pq uint Maximum number of metaslabs per group to preload . .It Sy metaslab_preload_pct Ns = Ns Sy 50 Pq uint Percentage of CPUs to run a metaslab preload taskq . .It Sy metaslab_lba_weighting_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Give more weight to metaslabs with lower LBAs, assuming they have greater bandwidth, as is typically the case on a modern constant angular velocity disk drive. . .It Sy metaslab_unload_delay Ns = Ns Sy 32 Pq uint After a metaslab is used, we keep it loaded for this many TXGs, to attempt to reduce unnecessary reloading. Note that both this many TXGs and .Sy metaslab_unload_delay_ms milliseconds must pass before unloading will occur. . .It Sy metaslab_unload_delay_ms Ns = Ns Sy 600000 Ns ms Po 10 min Pc Pq uint After a metaslab is used, we keep it loaded for this many milliseconds, to attempt to reduce unnecessary reloading. Note, that both this many milliseconds and .Sy metaslab_unload_delay TXGs must pass before unloading will occur. . .It Sy reference_history Ns = Ns Sy 3 Pq uint Maximum reference holders being tracked when reference_tracking_enable is active. .It Sy raidz_expand_max_copy_bytes Ns = Ns Sy 160MB Pq ulong Max amount of memory to use for RAID-Z expansion I/O. This limits how much I/O can be outstanding at once. . .It Sy raidz_expand_max_reflow_bytes Ns = Ns Sy 0 Pq ulong For testing, pause RAID-Z expansion when reflow amount reaches this value. . .It Sy raidz_io_aggregate_rows Ns = Ns Sy 4 Pq ulong For expanded RAID-Z, aggregate reads that have more rows than this. . .It Sy reference_history Ns = Ns Sy 3 Pq int Maximum reference holders being tracked when reference_tracking_enable is active. . .It Sy reference_tracking_enable Ns = Ns Sy 0 Ns | Ns 1 Pq int Track reference holders to .Sy refcount_t objects (debug builds only). . .It Sy send_holes_without_birth_time Ns = Ns Sy 1 Ns | Ns 0 Pq int When set, the .Sy hole_birth optimization will not be used, and all holes will always be sent during a .Nm zfs Cm send . This is useful if you suspect your datasets are affected by a bug in .Sy hole_birth . . .It Sy spa_config_path Ns = Ns Pa /etc/zfs/zpool.cache Pq charp SPA config file. . .It Sy spa_asize_inflation Ns = Ns Sy 24 Pq uint Multiplication factor used to estimate actual disk consumption from the size of data being written. The default value is a worst case estimate, but lower values may be valid for a given pool depending on its configuration. Pool administrators who understand the factors involved may wish to specify a more realistic inflation factor, particularly if they operate close to quota or capacity limits. . .It Sy spa_load_print_vdev_tree Ns = Ns Sy 0 Ns | Ns 1 Pq int Whether to print the vdev tree in the debugging message buffer during pool import. . .It Sy spa_load_verify_data Ns = Ns Sy 1 Ns | Ns 0 Pq int Whether to traverse data blocks during an "extreme rewind" .Pq Fl X import. .Pp An extreme rewind import normally performs a full traversal of all blocks in the pool for verification. If this parameter is unset, the traversal skips non-metadata blocks. It can be toggled once the import has started to stop or start the traversal of non-metadata blocks. . .It Sy spa_load_verify_metadata Ns = Ns Sy 1 Ns | Ns 0 Pq int Whether to traverse blocks during an "extreme rewind" .Pq Fl X pool import. .Pp An extreme rewind import normally performs a full traversal of all blocks in the pool for verification. If this parameter is unset, the traversal is not performed. It can be toggled once the import has started to stop or start the traversal. . .It Sy spa_load_verify_shift Ns = Ns Sy 4 Po 1/16th Pc Pq uint Sets the maximum number of bytes to consume during pool import to the log2 fraction of the target ARC size. . .It Sy spa_slop_shift Ns = Ns Sy 5 Po 1/32nd Pc Pq int Normally, we don't allow the last .Sy 3.2% Pq Sy 1/2^spa_slop_shift of space in the pool to be consumed. This ensures that we don't run the pool completely out of space, due to unaccounted changes (e.g. to the MOS). It also limits the worst-case time to allocate space. If we have less than this amount of free space, most ZPL operations (e.g. write, create) will return .Sy ENOSPC . . .It Sy spa_num_allocators Ns = Ns Sy 4 Pq int Determines the number of block alloctators to use per spa instance. Capped by the number of actual CPUs in the system. .Pp Note that setting this value too high could result in performance degredation and/or excess fragmentation. . .It Sy spa_upgrade_errlog_limit Ns = Ns Sy 0 Pq uint Limits the number of on-disk error log entries that will be converted to the new format when enabling the .Sy head_errlog feature. The default is to convert all log entries. . .It Sy vdev_removal_max_span Ns = Ns Sy 32768 Ns B Po 32 KiB Pc Pq uint During top-level vdev removal, chunks of data are copied from the vdev which may include free space in order to trade bandwidth for IOPS. This parameter determines the maximum span of free space, in bytes, which will be included as "unnecessary" data in a chunk of copied data. .Pp The default value here was chosen to align with .Sy zfs_vdev_read_gap_limit , which is a similar concept when doing regular reads (but there's no reason it has to be the same). . .It Sy vdev_file_logical_ashift Ns = Ns Sy 9 Po 512 B Pc Pq u64 Logical ashift for file-based devices. . .It Sy vdev_file_physical_ashift Ns = Ns Sy 9 Po 512 B Pc Pq u64 Physical ashift for file-based devices. . .It Sy zap_iterate_prefetch Ns = Ns Sy 1 Ns | Ns 0 Pq int If set, when we start iterating over a ZAP object, prefetch the entire object (all leaf blocks). However, this is limited by .Sy dmu_prefetch_max . . .It Sy zap_micro_max_size Ns = Ns Sy 131072 Ns B Po 128 KiB Pc Pq int Maximum micro ZAP size. A micro ZAP is upgraded to a fat ZAP, once it grows beyond the specified size. . .It Sy zfetch_min_distance Ns = Ns Sy 4194304 Ns B Po 4 MiB Pc Pq uint Min bytes to prefetch per stream. Prefetch distance starts from the demand access size and quickly grows to this value, doubling on each hit. After that it may grow further by 1/8 per hit, but only if some prefetch since last time haven't completed in time to satisfy demand request, i.e. prefetch depth didn't cover the read latency or the pool got saturated. . .It Sy zfetch_max_distance Ns = Ns Sy 67108864 Ns B Po 64 MiB Pc Pq uint Max bytes to prefetch per stream. . .It Sy zfetch_max_idistance Ns = Ns Sy 67108864 Ns B Po 64 MiB Pc Pq uint Max bytes to prefetch indirects for per stream. . .It Sy zfetch_max_streams Ns = Ns Sy 8 Pq uint Max number of streams per zfetch (prefetch streams per file). . .It Sy zfetch_min_sec_reap Ns = Ns Sy 1 Pq uint Min time before inactive prefetch stream can be reclaimed . .It Sy zfetch_max_sec_reap Ns = Ns Sy 2 Pq uint Max time before inactive prefetch stream can be deleted . .It Sy zfs_abd_scatter_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Enables ARC from using scatter/gather lists and forces all allocations to be linear in kernel memory. Disabling can improve performance in some code paths at the expense of fragmented kernel memory. . .It Sy zfs_abd_scatter_max_order Ns = Ns Sy MAX_ORDER\-1 Pq uint Maximum number of consecutive memory pages allocated in a single block for scatter/gather lists. .Pp The value of .Sy MAX_ORDER depends on kernel configuration. . .It Sy zfs_abd_scatter_min_size Ns = Ns Sy 1536 Ns B Po 1.5 KiB Pc Pq uint This is the minimum allocation size that will use scatter (page-based) ABDs. Smaller allocations will use linear ABDs. . .It Sy zfs_arc_dnode_limit Ns = Ns Sy 0 Ns B Pq u64 When the number of bytes consumed by dnodes in the ARC exceeds this number of bytes, try to unpin some of it in response to demand for non-metadata. This value acts as a ceiling to the amount of dnode metadata, and defaults to .Sy 0 , which indicates that a percent which is based on .Sy zfs_arc_dnode_limit_percent of the ARC meta buffers that may be used for dnodes. .It Sy zfs_arc_dnode_limit_percent Ns = Ns Sy 10 Ns % Pq u64 Percentage that can be consumed by dnodes of ARC meta buffers. .Pp See also .Sy zfs_arc_dnode_limit , which serves a similar purpose but has a higher priority if nonzero. . .It Sy zfs_arc_dnode_reduce_percent Ns = Ns Sy 10 Ns % Pq u64 Percentage of ARC dnodes to try to scan in response to demand for non-metadata when the number of bytes consumed by dnodes exceeds .Sy zfs_arc_dnode_limit . . .It Sy zfs_arc_average_blocksize Ns = Ns Sy 8192 Ns B Po 8 KiB Pc Pq uint The ARC's buffer hash table is sized based on the assumption of an average block size of this value. This works out to roughly 1 MiB of hash table per 1 GiB of physical memory with 8-byte pointers. For configurations with a known larger average block size, this value can be increased to reduce the memory footprint. . .It Sy zfs_arc_eviction_pct Ns = Ns Sy 200 Ns % Pq uint When .Fn arc_is_overflowing , .Fn arc_get_data_impl waits for this percent of the requested amount of data to be evicted. For example, by default, for every .Em 2 KiB that's evicted, .Em 1 KiB of it may be "reused" by a new allocation. Since this is above .Sy 100 Ns % , it ensures that progress is made towards getting .Sy arc_size No under Sy arc_c . Since this is finite, it ensures that allocations can still happen, even during the potentially long time that .Sy arc_size No is more than Sy arc_c . . .It Sy zfs_arc_evict_batch_limit Ns = Ns Sy 10 Pq uint Number ARC headers to evict per sub-list before proceeding to another sub-list. This batch-style operation prevents entire sub-lists from being evicted at once but comes at a cost of additional unlocking and locking. . .It Sy zfs_arc_grow_retry Ns = Ns Sy 0 Ns s Pq uint If set to a non zero value, it will replace the .Sy arc_grow_retry value with this value. The .Sy arc_grow_retry .No value Pq default Sy 5 Ns s is the number of seconds the ARC will wait before trying to resume growth after a memory pressure event. . .It Sy zfs_arc_lotsfree_percent Ns = Ns Sy 10 Ns % Pq int Throttle I/O when free system memory drops below this percentage of total system memory. Setting this value to .Sy 0 will disable the throttle. . .It Sy zfs_arc_max Ns = Ns Sy 0 Ns B Pq u64 Max size of ARC in bytes. If .Sy 0 , then the max size of ARC is determined by the amount of system memory installed. The larger of .Sy all_system_memory No \- Sy 1 GiB and .Sy 5/8 No \(mu Sy all_system_memory will be used as the limit. This value must be at least .Sy 67108864 Ns B Pq 64 MiB . .Pp This value can be changed dynamically, with some caveats. It cannot be set back to .Sy 0 while running, and reducing it below the current ARC size will not cause the ARC to shrink without memory pressure to induce shrinking. . .It Sy zfs_arc_meta_balance Ns = Ns Sy 500 Pq uint Balance between metadata and data on ghost hits. Values above 100 increase metadata caching by proportionally reducing effect of ghost data hits on target data/metadata rate. . .It Sy zfs_arc_min Ns = Ns Sy 0 Ns B Pq u64 Min size of ARC in bytes. .No If set to Sy 0 , arc_c_min will default to consuming the larger of .Sy 32 MiB and .Sy all_system_memory No / Sy 32 . . .It Sy zfs_arc_min_prefetch_ms Ns = Ns Sy 0 Ns ms Ns Po Ns ≡ Ns 1s Pc Pq uint Minimum time prefetched blocks are locked in the ARC. . .It Sy zfs_arc_min_prescient_prefetch_ms Ns = Ns Sy 0 Ns ms Ns Po Ns ≡ Ns 6s Pc Pq uint Minimum time "prescient prefetched" blocks are locked in the ARC. These blocks are meant to be prefetched fairly aggressively ahead of the code that may use them. . .It Sy zfs_arc_prune_task_threads Ns = Ns Sy 1 Pq int Number of arc_prune threads. .Fx does not need more than one. Linux may theoretically use one per mount point up to number of CPUs, but that was not proven to be useful. . .It Sy zfs_max_missing_tvds Ns = Ns Sy 0 Pq int Number of missing top-level vdevs which will be allowed during pool import (only in read-only mode). . .It Sy zfs_max_nvlist_src_size Ns = Sy 0 Pq u64 Maximum size in bytes allowed to be passed as .Sy zc_nvlist_src_size for ioctls on .Pa /dev/zfs . This prevents a user from causing the kernel to allocate an excessive amount of memory. When the limit is exceeded, the ioctl fails with .Sy EINVAL and a description of the error is sent to the .Pa zfs-dbgmsg log. This parameter should not need to be touched under normal circumstances. If .Sy 0 , equivalent to a quarter of the user-wired memory limit under .Fx and to .Sy 134217728 Ns B Pq 128 MiB under Linux. . .It Sy zfs_multilist_num_sublists Ns = Ns Sy 0 Pq uint To allow more fine-grained locking, each ARC state contains a series of lists for both data and metadata objects. Locking is performed at the level of these "sub-lists". This parameters controls the number of sub-lists per ARC state, and also applies to other uses of the multilist data structure. .Pp If .Sy 0 , equivalent to the greater of the number of online CPUs and .Sy 4 . . .It Sy zfs_arc_overflow_shift Ns = Ns Sy 8 Pq int The ARC size is considered to be overflowing if it exceeds the current ARC target size .Pq Sy arc_c by thresholds determined by this parameter. Exceeding by .Sy ( arc_c No >> Sy zfs_arc_overflow_shift ) No / Sy 2 starts ARC reclamation process. If that appears insufficient, exceeding by .Sy ( arc_c No >> Sy zfs_arc_overflow_shift ) No \(mu Sy 1.5 blocks new buffer allocation until the reclaim thread catches up. Started reclamation process continues till ARC size returns below the target size. .Pp The default value of .Sy 8 causes the ARC to start reclamation if it exceeds the target size by .Em 0.2% of the target size, and block allocations by .Em 0.6% . . .It Sy zfs_arc_shrink_shift Ns = Ns Sy 0 Pq uint If nonzero, this will update .Sy arc_shrink_shift Pq default Sy 7 with the new value. . .It Sy zfs_arc_pc_percent Ns = Ns Sy 0 Ns % Po off Pc Pq uint Percent of pagecache to reclaim ARC to. .Pp This tunable allows the ZFS ARC to play more nicely with the kernel's LRU pagecache. It can guarantee that the ARC size won't collapse under scanning pressure on the pagecache, yet still allows the ARC to be reclaimed down to .Sy zfs_arc_min if necessary. This value is specified as percent of pagecache size (as measured by .Sy NR_FILE_PAGES ) , where that percent may exceed .Sy 100 . This only operates during memory pressure/reclaim. . .It Sy zfs_arc_shrinker_limit Ns = Ns Sy 10000 Pq int This is a limit on how many pages the ARC shrinker makes available for eviction in response to one page allocation attempt. Note that in practice, the kernel's shrinker can ask us to evict up to about four times this for one allocation attempt. .Pp The default limit of .Sy 10000 Pq in practice, Em 160 MiB No per allocation attempt with 4 KiB pages limits the amount of time spent attempting to reclaim ARC memory to less than 100 ms per allocation attempt, even with a small average compressed block size of ~8 KiB. .Pp The parameter can be set to 0 (zero) to disable the limit, and only applies on Linux. . .It Sy zfs_arc_sys_free Ns = Ns Sy 0 Ns B Pq u64 The target number of bytes the ARC should leave as free memory on the system. If zero, equivalent to the bigger of .Sy 512 KiB No and Sy all_system_memory/64 . . .It Sy zfs_autoimport_disable Ns = Ns Sy 1 Ns | Ns 0 Pq int Disable pool import at module load by ignoring the cache file .Pq Sy spa_config_path . . .It Sy zfs_checksum_events_per_second Ns = Ns Sy 20 Ns /s Pq uint Rate limit checksum events to this many per second. Note that this should not be set below the ZED thresholds (currently 10 checksums over 10 seconds) or else the daemon may not trigger any action. . .It Sy zfs_commit_timeout_pct Ns = Ns Sy 10 Ns % Pq uint This controls the amount of time that a ZIL block (lwb) will remain "open" when it isn't "full", and it has a thread waiting for it to be committed to stable storage. The timeout is scaled based on a percentage of the last lwb latency to avoid significantly impacting the latency of each individual transaction record (itx). . .It Sy zfs_condense_indirect_commit_entry_delay_ms Ns = Ns Sy 0 Ns ms Pq int Vdev indirection layer (used for device removal) sleeps for this many milliseconds during mapping generation. Intended for use with the test suite to throttle vdev removal speed. . .It Sy zfs_condense_indirect_obsolete_pct Ns = Ns Sy 25 Ns % Pq uint Minimum percent of obsolete bytes in vdev mapping required to attempt to condense .Pq see Sy zfs_condense_indirect_vdevs_enable . Intended for use with the test suite to facilitate triggering condensing as needed. . .It Sy zfs_condense_indirect_vdevs_enable Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable condensing indirect vdev mappings. When set, attempt to condense indirect vdev mappings if the mapping uses more than .Sy zfs_condense_min_mapping_bytes bytes of memory and if the obsolete space map object uses more than .Sy zfs_condense_max_obsolete_bytes bytes on-disk. The condensing process is an attempt to save memory by removing obsolete mappings. . .It Sy zfs_condense_max_obsolete_bytes Ns = Ns Sy 1073741824 Ns B Po 1 GiB Pc Pq u64 Only attempt to condense indirect vdev mappings if the on-disk size of the obsolete space map object is greater than this number of bytes .Pq see Sy zfs_condense_indirect_vdevs_enable . . .It Sy zfs_condense_min_mapping_bytes Ns = Ns Sy 131072 Ns B Po 128 KiB Pc Pq u64 Minimum size vdev mapping to attempt to condense .Pq see Sy zfs_condense_indirect_vdevs_enable . . .It Sy zfs_dbgmsg_enable Ns = Ns Sy 1 Ns | Ns 0 Pq int Internally ZFS keeps a small log to facilitate debugging. The log is enabled by default, and can be disabled by unsetting this option. The contents of the log can be accessed by reading .Pa /proc/spl/kstat/zfs/dbgmsg . Writing .Sy 0 to the file clears the log. .Pp This setting does not influence debug prints due to .Sy zfs_flags . . .It Sy zfs_dbgmsg_maxsize Ns = Ns Sy 4194304 Ns B Po 4 MiB Pc Pq uint Maximum size of the internal ZFS debug log. . .It Sy zfs_dbuf_state_index Ns = Ns Sy 0 Pq int Historically used for controlling what reporting was available under .Pa /proc/spl/kstat/zfs . No effect. . .It Sy zfs_deadman_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int When a pool sync operation takes longer than .Sy zfs_deadman_synctime_ms , or when an individual I/O operation takes longer than .Sy zfs_deadman_ziotime_ms , then the operation is considered to be "hung". If .Sy zfs_deadman_enabled is set, then the deadman behavior is invoked as described by .Sy zfs_deadman_failmode . By default, the deadman is enabled and set to .Sy wait which results in "hung" I/O operations only being logged. The deadman is automatically disabled when a pool gets suspended. . .It Sy zfs_deadman_failmode Ns = Ns Sy wait Pq charp Controls the failure behavior when the deadman detects a "hung" I/O operation. Valid values are: .Bl -tag -compact -offset 4n -width "continue" .It Sy wait Wait for a "hung" operation to complete. For each "hung" operation a "deadman" event will be posted describing that operation. .It Sy continue Attempt to recover from a "hung" operation by re-dispatching it to the I/O pipeline if possible. .It Sy panic Panic the system. This can be used to facilitate automatic fail-over to a properly configured fail-over partner. .El . .It Sy zfs_deadman_checktime_ms Ns = Ns Sy 60000 Ns ms Po 1 min Pc Pq u64 Check time in milliseconds. This defines the frequency at which we check for hung I/O requests and potentially invoke the .Sy zfs_deadman_failmode behavior. . .It Sy zfs_deadman_synctime_ms Ns = Ns Sy 600000 Ns ms Po 10 min Pc Pq u64 Interval in milliseconds after which the deadman is triggered and also the interval after which a pool sync operation is considered to be "hung". Once this limit is exceeded the deadman will be invoked every .Sy zfs_deadman_checktime_ms milliseconds until the pool sync completes. . .It Sy zfs_deadman_ziotime_ms Ns = Ns Sy 300000 Ns ms Po 5 min Pc Pq u64 Interval in milliseconds after which the deadman is triggered and an individual I/O operation is considered to be "hung". As long as the operation remains "hung", the deadman will be invoked every .Sy zfs_deadman_checktime_ms milliseconds until the operation completes. . .It Sy zfs_dedup_prefetch Ns = Ns Sy 0 Ns | Ns 1 Pq int Enable prefetching dedup-ed blocks which are going to be freed. . .It Sy zfs_delay_min_dirty_percent Ns = Ns Sy 60 Ns % Pq uint Start to delay each transaction once there is this amount of dirty data, expressed as a percentage of .Sy zfs_dirty_data_max . This value should be at least .Sy zfs_vdev_async_write_active_max_dirty_percent . .No See Sx ZFS TRANSACTION DELAY . . .It Sy zfs_delay_scale Ns = Ns Sy 500000 Pq int This controls how quickly the transaction delay approaches infinity. Larger values cause longer delays for a given amount of dirty data. .Pp For the smoothest delay, this value should be about 1 billion divided by the maximum number of operations per second. This will smoothly handle between ten times and a tenth of this number. .No See Sx ZFS TRANSACTION DELAY . .Pp .Sy zfs_delay_scale No \(mu Sy zfs_dirty_data_max Em must No be smaller than Sy 2^64 . . .It Sy zfs_disable_ivset_guid_check Ns = Ns Sy 0 Ns | Ns 1 Pq int Disables requirement for IVset GUIDs to be present and match when doing a raw receive of encrypted datasets. Intended for users whose pools were created with OpenZFS pre-release versions and now have compatibility issues. . .It Sy zfs_key_max_salt_uses Ns = Ns Sy 400000000 Po 4*10^8 Pc Pq ulong Maximum number of uses of a single salt value before generating a new one for encrypted datasets. The default value is also the maximum. . .It Sy zfs_object_mutex_size Ns = Ns Sy 64 Pq uint Size of the znode hashtable used for holds. .Pp Due to the need to hold locks on objects that may not exist yet, kernel mutexes are not created per-object and instead a hashtable is used where collisions will result in objects waiting when there is not actually contention on the same object. . .It Sy zfs_slow_io_events_per_second Ns = Ns Sy 20 Ns /s Pq int Rate limit delay and deadman zevents (which report slow I/O operations) to this many per second. . .It Sy zfs_unflushed_max_mem_amt Ns = Ns Sy 1073741824 Ns B Po 1 GiB Pc Pq u64 Upper-bound limit for unflushed metadata changes to be held by the log spacemap in memory, in bytes. . .It Sy zfs_unflushed_max_mem_ppm Ns = Ns Sy 1000 Ns ppm Po 0.1% Pc Pq u64 Part of overall system memory that ZFS allows to be used for unflushed metadata changes by the log spacemap, in millionths. . .It Sy zfs_unflushed_log_block_max Ns = Ns Sy 131072 Po 128k Pc Pq u64 Describes the maximum number of log spacemap blocks allowed for each pool. The default value means that the space in all the log spacemaps can add up to no more than .Sy 131072 blocks (which means .Em 16 GiB of logical space before compression and ditto blocks, assuming that blocksize is .Em 128 KiB ) . .Pp This tunable is important because it involves a trade-off between import time after an unclean export and the frequency of flushing metaslabs. The higher this number is, the more log blocks we allow when the pool is active which means that we flush metaslabs less often and thus decrease the number of I/O operations for spacemap updates per TXG. At the same time though, that means that in the event of an unclean export, there will be more log spacemap blocks for us to read, inducing overhead in the import time of the pool. The lower the number, the amount of flushing increases, destroying log blocks quicker as they become obsolete faster, which leaves less blocks to be read during import time after a crash. .Pp Each log spacemap block existing during pool import leads to approximately one extra logical I/O issued. This is the reason why this tunable is exposed in terms of blocks rather than space used. . .It Sy zfs_unflushed_log_block_min Ns = Ns Sy 1000 Pq u64 If the number of metaslabs is small and our incoming rate is high, we could get into a situation that we are flushing all our metaslabs every TXG. Thus we always allow at least this many log blocks. . .It Sy zfs_unflushed_log_block_pct Ns = Ns Sy 400 Ns % Pq u64 Tunable used to determine the number of blocks that can be used for the spacemap log, expressed as a percentage of the total number of unflushed metaslabs in the pool. . .It Sy zfs_unflushed_log_txg_max Ns = Ns Sy 1000 Pq u64 Tunable limiting maximum time in TXGs any metaslab may remain unflushed. It effectively limits maximum number of unflushed per-TXG spacemap logs that need to be read after unclean pool export. . .It Sy zfs_unlink_suspend_progress Ns = Ns Sy 0 Ns | Ns 1 Pq uint When enabled, files will not be asynchronously removed from the list of pending unlinks and the space they consume will be leaked. Once this option has been disabled and the dataset is remounted, the pending unlinks will be processed and the freed space returned to the pool. This option is used by the test suite. . .It Sy zfs_delete_blocks Ns = Ns Sy 20480 Pq ulong This is the used to define a large file for the purposes of deletion. Files containing more than .Sy zfs_delete_blocks will be deleted asynchronously, while smaller files are deleted synchronously. Decreasing this value will reduce the time spent in an .Xr unlink 2 system call, at the expense of a longer delay before the freed space is available. This only applies on Linux. . .It Sy zfs_dirty_data_max Ns = Pq int Determines the dirty space limit in bytes. Once this limit is exceeded, new writes are halted until space frees up. This parameter takes precedence over .Sy zfs_dirty_data_max_percent . .No See Sx ZFS TRANSACTION DELAY . .Pp Defaults to .Sy physical_ram/10 , capped at .Sy zfs_dirty_data_max_max . . .It Sy zfs_dirty_data_max_max Ns = Pq int Maximum allowable value of .Sy zfs_dirty_data_max , expressed in bytes. This limit is only enforced at module load time, and will be ignored if .Sy zfs_dirty_data_max is later changed. This parameter takes precedence over .Sy zfs_dirty_data_max_max_percent . .No See Sx ZFS TRANSACTION DELAY . .Pp Defaults to .Sy min(physical_ram/4, 4GiB) , or .Sy min(physical_ram/4, 1GiB) for 32-bit systems. . .It Sy zfs_dirty_data_max_max_percent Ns = Ns Sy 25 Ns % Pq uint Maximum allowable value of .Sy zfs_dirty_data_max , expressed as a percentage of physical RAM. This limit is only enforced at module load time, and will be ignored if .Sy zfs_dirty_data_max is later changed. The parameter .Sy zfs_dirty_data_max_max takes precedence over this one. .No See Sx ZFS TRANSACTION DELAY . . .It Sy zfs_dirty_data_max_percent Ns = Ns Sy 10 Ns % Pq uint Determines the dirty space limit, expressed as a percentage of all memory. Once this limit is exceeded, new writes are halted until space frees up. The parameter .Sy zfs_dirty_data_max takes precedence over this one. .No See Sx ZFS TRANSACTION DELAY . .Pp Subject to .Sy zfs_dirty_data_max_max . . .It Sy zfs_dirty_data_sync_percent Ns = Ns Sy 20 Ns % Pq uint Start syncing out a transaction group if there's at least this much dirty data .Pq as a percentage of Sy zfs_dirty_data_max . This should be less than .Sy zfs_vdev_async_write_active_min_dirty_percent . . .It Sy zfs_wrlog_data_max Ns = Pq int The upper limit of write-transaction zil log data size in bytes. Write operations are throttled when approaching the limit until log data is cleared out after transaction group sync. Because of some overhead, it should be set at least 2 times the size of .Sy zfs_dirty_data_max .No to prevent harming normal write throughput . It also should be smaller than the size of the slog device if slog is present. .Pp Defaults to .Sy zfs_dirty_data_max*2 . .It Sy zfs_fallocate_reserve_percent Ns = Ns Sy 110 Ns % Pq uint Since ZFS is a copy-on-write filesystem with snapshots, blocks cannot be preallocated for a file in order to guarantee that later writes will not run out of space. Instead, .Xr fallocate 2 space preallocation only checks that sufficient space is currently available in the pool or the user's project quota allocation, and then creates a sparse file of the requested size. The requested space is multiplied by .Sy zfs_fallocate_reserve_percent to allow additional space for indirect blocks and other internal metadata. Setting this to .Sy 0 disables support for .Xr fallocate 2 and causes it to return .Sy EOPNOTSUPP . . .It Sy zfs_fletcher_4_impl Ns = Ns Sy fastest Pq string Select a fletcher 4 implementation. .Pp Supported selectors are: .Sy fastest , scalar , sse2 , ssse3 , avx2 , avx512f , avx512bw , .No and Sy aarch64_neon . All except .Sy fastest No and Sy scalar require instruction set extensions to be available, and will only appear if ZFS detects that they are present at runtime. If multiple implementations of fletcher 4 are available, the .Sy fastest will be chosen using a micro benchmark. Selecting .Sy scalar results in the original CPU-based calculation being used. Selecting any option other than .Sy fastest No or Sy scalar results in vector instructions from the respective CPU instruction set being used. . +.It Sy zfs_bclone_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int +Enable the experimental block cloning feature. +If this setting is 0, then even if feature@block_cloning is enabled, +attempts to clone blocks will act as though the feature is disabled. +. .It Sy zfs_blake3_impl Ns = Ns Sy fastest Pq string Select a BLAKE3 implementation. .Pp Supported selectors are: .Sy cycle , fastest , generic , sse2 , sse41 , avx2 , avx512 . All except .Sy cycle , fastest No and Sy generic require instruction set extensions to be available, and will only appear if ZFS detects that they are present at runtime. If multiple implementations of BLAKE3 are available, the .Sy fastest will be chosen using a micro benchmark. You can see the benchmark results by reading this kstat file: .Pa /proc/spl/kstat/zfs/chksum_bench . . .It Sy zfs_free_bpobj_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable/disable the processing of the free_bpobj object. . .It Sy zfs_async_block_max_blocks Ns = Ns Sy UINT64_MAX Po unlimited Pc Pq u64 Maximum number of blocks freed in a single TXG. . .It Sy zfs_max_async_dedup_frees Ns = Ns Sy 100000 Po 10^5 Pc Pq u64 Maximum number of dedup blocks freed in a single TXG. . .It Sy zfs_vdev_async_read_max_active Ns = Ns Sy 3 Pq uint Maximum asynchronous read I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_async_read_min_active Ns = Ns Sy 1 Pq uint Minimum asynchronous read I/O operation active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_async_write_active_max_dirty_percent Ns = Ns Sy 60 Ns % Pq uint When the pool has more than this much dirty data, use .Sy zfs_vdev_async_write_max_active to limit active async writes. If the dirty data is between the minimum and maximum, the active I/O limit is linearly interpolated. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_async_write_active_min_dirty_percent Ns = Ns Sy 30 Ns % Pq uint When the pool has less than this much dirty data, use .Sy zfs_vdev_async_write_min_active to limit active async writes. If the dirty data is between the minimum and maximum, the active I/O limit is linearly interpolated. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_async_write_max_active Ns = Ns Sy 10 Pq uint Maximum asynchronous write I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_async_write_min_active Ns = Ns Sy 2 Pq uint Minimum asynchronous write I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . .Pp Lower values are associated with better latency on rotational media but poorer resilver performance. The default value of .Sy 2 was chosen as a compromise. A value of .Sy 3 has been shown to improve resilver performance further at a cost of further increasing latency. . .It Sy zfs_vdev_initializing_max_active Ns = Ns Sy 1 Pq uint Maximum initializing I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_initializing_min_active Ns = Ns Sy 1 Pq uint Minimum initializing I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_max_active Ns = Ns Sy 1000 Pq uint The maximum number of I/O operations active to each device. Ideally, this will be at least the sum of each queue's .Sy max_active . .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_open_timeout_ms Ns = Ns Sy 1000 Pq uint Timeout value to wait before determining a device is missing during import. This is helpful for transient missing paths due to links being briefly removed and recreated in response to udev events. . .It Sy zfs_vdev_rebuild_max_active Ns = Ns Sy 3 Pq uint Maximum sequential resilver I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_rebuild_min_active Ns = Ns Sy 1 Pq uint Minimum sequential resilver I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_removal_max_active Ns = Ns Sy 2 Pq uint Maximum removal I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_removal_min_active Ns = Ns Sy 1 Pq uint Minimum removal I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_scrub_max_active Ns = Ns Sy 2 Pq uint Maximum scrub I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_scrub_min_active Ns = Ns Sy 1 Pq uint Minimum scrub I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_sync_read_max_active Ns = Ns Sy 10 Pq uint Maximum synchronous read I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_sync_read_min_active Ns = Ns Sy 10 Pq uint Minimum synchronous read I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_sync_write_max_active Ns = Ns Sy 10 Pq uint Maximum synchronous write I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_sync_write_min_active Ns = Ns Sy 10 Pq uint Minimum synchronous write I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_trim_max_active Ns = Ns Sy 2 Pq uint Maximum trim/discard I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_trim_min_active Ns = Ns Sy 1 Pq uint Minimum trim/discard I/O operations active to each device. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_nia_delay Ns = Ns Sy 5 Pq uint For non-interactive I/O (scrub, resilver, removal, initialize and rebuild), the number of concurrently-active I/O operations is limited to .Sy zfs_*_min_active , unless the vdev is "idle". When there are no interactive I/O operations active (synchronous or otherwise), and .Sy zfs_vdev_nia_delay operations have completed since the last interactive operation, then the vdev is considered to be "idle", and the number of concurrently-active non-interactive operations is increased to .Sy zfs_*_max_active . .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_nia_credit Ns = Ns Sy 5 Pq uint Some HDDs tend to prioritize sequential I/O so strongly, that concurrent random I/O latency reaches several seconds. On some HDDs this happens even if sequential I/O operations are submitted one at a time, and so setting .Sy zfs_*_max_active Ns = Sy 1 does not help. To prevent non-interactive I/O, like scrub, from monopolizing the device, no more than .Sy zfs_vdev_nia_credit operations can be sent while there are outstanding incomplete interactive operations. This enforced wait ensures the HDD services the interactive I/O within a reasonable amount of time. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_queue_depth_pct Ns = Ns Sy 1000 Ns % Pq uint Maximum number of queued allocations per top-level vdev expressed as a percentage of .Sy zfs_vdev_async_write_max_active , which allows the system to detect devices that are more capable of handling allocations and to allocate more blocks to those devices. This allows for dynamic allocation distribution when devices are imbalanced, as fuller devices will tend to be slower than empty devices. .Pp Also see .Sy zio_dva_throttle_enabled . . .It Sy zfs_vdev_def_queue_depth Ns = Ns Sy 32 Pq uint Default queue depth for each vdev IO allocator. Higher values allow for better coalescing of sequential writes before sending them to the disk, but can increase transaction commit times. . .It Sy zfs_vdev_failfast_mask Ns = Ns Sy 1 Pq uint Defines if the driver should retire on a given error type. The following options may be bitwise-ored together: .TS box; lbz r l l . Value Name Description _ 1 Device No driver retries on device errors 2 Transport No driver retries on transport errors. 4 Driver No driver retries on driver errors. .TE . .It Sy zfs_expire_snapshot Ns = Ns Sy 300 Ns s Pq int Time before expiring .Pa .zfs/snapshot . . .It Sy zfs_admin_snapshot Ns = Ns Sy 0 Ns | Ns 1 Pq int Allow the creation, removal, or renaming of entries in the .Sy .zfs/snapshot directory to cause the creation, destruction, or renaming of snapshots. When enabled, this functionality works both locally and over NFS exports which have the .Em no_root_squash option set. . .It Sy zfs_flags Ns = Ns Sy 0 Pq int Set additional debugging flags. The following flags may be bitwise-ored together: .TS box; lbz r l l . Value Name Description _ 1 ZFS_DEBUG_DPRINTF Enable dprintf entries in the debug log. * 2 ZFS_DEBUG_DBUF_VERIFY Enable extra dbuf verifications. * 4 ZFS_DEBUG_DNODE_VERIFY Enable extra dnode verifications. 8 ZFS_DEBUG_SNAPNAMES Enable snapshot name verification. * 16 ZFS_DEBUG_MODIFY Check for illegally modified ARC buffers. 64 ZFS_DEBUG_ZIO_FREE Enable verification of block frees. 128 ZFS_DEBUG_HISTOGRAM_VERIFY Enable extra spacemap histogram verifications. 256 ZFS_DEBUG_METASLAB_VERIFY Verify space accounting on disk matches in-memory \fBrange_trees\fP. 512 ZFS_DEBUG_SET_ERROR Enable \fBSET_ERROR\fP and dprintf entries in the debug log. 1024 ZFS_DEBUG_INDIRECT_REMAP Verify split blocks created by device removal. 2048 ZFS_DEBUG_TRIM Verify TRIM ranges are always within the allocatable range tree. 4096 ZFS_DEBUG_LOG_SPACEMAP Verify that the log summary is consistent with the spacemap log and enable \fBzfs_dbgmsgs\fP for metaslab loading and flushing. .TE .Sy \& * No Requires debug build . . .It Sy zfs_btree_verify_intensity Ns = Ns Sy 0 Pq uint Enables btree verification. The following settings are culminative: .TS box; lbz r l l . Value Description 1 Verify height. 2 Verify pointers from children to parent. 3 Verify element counts. 4 Verify element order. (expensive) * 5 Verify unused memory is poisoned. (expensive) .TE .Sy \& * No Requires debug build . . .It Sy zfs_free_leak_on_eio Ns = Ns Sy 0 Ns | Ns 1 Pq int If destroy encounters an .Sy EIO while reading metadata (e.g. indirect blocks), space referenced by the missing metadata can not be freed. Normally this causes the background destroy to become "stalled", as it is unable to make forward progress. While in this stalled state, all remaining space to free from the error-encountering filesystem is "temporarily leaked". Set this flag to cause it to ignore the .Sy EIO , permanently leak the space from indirect blocks that can not be read, and continue to free everything else that it can. .Pp The default "stalling" behavior is useful if the storage partially fails (i.e. some but not all I/O operations fail), and then later recovers. In this case, we will be able to continue pool operations while it is partially failed, and when it recovers, we can continue to free the space, with no leaks. Note, however, that this case is actually fairly rare. .Pp Typically pools either .Bl -enum -compact -offset 4n -width "1." .It fail completely (but perhaps temporarily, e.g. due to a top-level vdev going offline), or .It have localized, permanent errors (e.g. disk returns the wrong data due to bit flip or firmware bug). .El In the former case, this setting does not matter because the pool will be suspended and the sync thread will not be able to make forward progress regardless. In the latter, because the error is permanent, the best we can do is leak the minimum amount of space, which is what setting this flag will do. It is therefore reasonable for this flag to normally be set, but we chose the more conservative approach of not setting it, so that there is no possibility of leaking space in the "partial temporary" failure case. . .It Sy zfs_free_min_time_ms Ns = Ns Sy 1000 Ns ms Po 1s Pc Pq uint During a .Nm zfs Cm destroy operation using the .Sy async_destroy feature, a minimum of this much time will be spent working on freeing blocks per TXG. . .It Sy zfs_obsolete_min_time_ms Ns = Ns Sy 500 Ns ms Pq uint Similar to .Sy zfs_free_min_time_ms , but for cleanup of old indirection records for removed vdevs. . .It Sy zfs_immediate_write_sz Ns = Ns Sy 32768 Ns B Po 32 KiB Pc Pq s64 Largest data block to write to the ZIL. Larger blocks will be treated as if the dataset being written to had the .Sy logbias Ns = Ns Sy throughput property set. . .It Sy zfs_initialize_value Ns = Ns Sy 16045690984833335022 Po 0xDEADBEEFDEADBEEE Pc Pq u64 Pattern written to vdev free space by .Xr zpool-initialize 8 . . .It Sy zfs_initialize_chunk_size Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64 Size of writes used by .Xr zpool-initialize 8 . This option is used by the test suite. . .It Sy zfs_livelist_max_entries Ns = Ns Sy 500000 Po 5*10^5 Pc Pq u64 The threshold size (in block pointers) at which we create a new sub-livelist. Larger sublists are more costly from a memory perspective but the fewer sublists there are, the lower the cost of insertion. . .It Sy zfs_livelist_min_percent_shared Ns = Ns Sy 75 Ns % Pq int If the amount of shared space between a snapshot and its clone drops below this threshold, the clone turns off the livelist and reverts to the old deletion method. This is in place because livelists no long give us a benefit once a clone has been overwritten enough. . .It Sy zfs_livelist_condense_new_alloc Ns = Ns Sy 0 Pq int Incremented each time an extra ALLOC blkptr is added to a livelist entry while it is being condensed. This option is used by the test suite to track race conditions. . .It Sy zfs_livelist_condense_sync_cancel Ns = Ns Sy 0 Pq int Incremented each time livelist condensing is canceled while in .Fn spa_livelist_condense_sync . This option is used by the test suite to track race conditions. . .It Sy zfs_livelist_condense_sync_pause Ns = Ns Sy 0 Ns | Ns 1 Pq int When set, the livelist condense process pauses indefinitely before executing the synctask \(em .Fn spa_livelist_condense_sync . This option is used by the test suite to trigger race conditions. . .It Sy zfs_livelist_condense_zthr_cancel Ns = Ns Sy 0 Pq int Incremented each time livelist condensing is canceled while in .Fn spa_livelist_condense_cb . This option is used by the test suite to track race conditions. . .It Sy zfs_livelist_condense_zthr_pause Ns = Ns Sy 0 Ns | Ns 1 Pq int When set, the livelist condense process pauses indefinitely before executing the open context condensing work in .Fn spa_livelist_condense_cb . This option is used by the test suite to trigger race conditions. . .It Sy zfs_lua_max_instrlimit Ns = Ns Sy 100000000 Po 10^8 Pc Pq u64 The maximum execution time limit that can be set for a ZFS channel program, specified as a number of Lua instructions. . .It Sy zfs_lua_max_memlimit Ns = Ns Sy 104857600 Po 100 MiB Pc Pq u64 The maximum memory limit that can be set for a ZFS channel program, specified in bytes. . .It Sy zfs_max_dataset_nesting Ns = Ns Sy 50 Pq int The maximum depth of nested datasets. This value can be tuned temporarily to fix existing datasets that exceed the predefined limit. . .It Sy zfs_max_log_walking Ns = Ns Sy 5 Pq u64 The number of past TXGs that the flushing algorithm of the log spacemap feature uses to estimate incoming log blocks. . .It Sy zfs_max_logsm_summary_length Ns = Ns Sy 10 Pq u64 Maximum number of rows allowed in the summary of the spacemap log. . .It Sy zfs_max_recordsize Ns = Ns Sy 16777216 Po 16 MiB Pc Pq uint We currently support block sizes from .Em 512 Po 512 B Pc No to Em 16777216 Po 16 MiB Pc . The benefits of larger blocks, and thus larger I/O, need to be weighed against the cost of COWing a giant block to modify one byte. Additionally, very large blocks can have an impact on I/O latency, and also potentially on the memory allocator. Therefore, we formerly forbade creating blocks larger than 1M. Larger blocks could be created by changing it, and pools with larger blocks can always be imported and used, regardless of this setting. . .It Sy zfs_allow_redacted_dataset_mount Ns = Ns Sy 0 Ns | Ns 1 Pq int Allow datasets received with redacted send/receive to be mounted. Normally disabled because these datasets may be missing key data. . .It Sy zfs_min_metaslabs_to_flush Ns = Ns Sy 1 Pq u64 Minimum number of metaslabs to flush per dirty TXG. . .It Sy zfs_metaslab_fragmentation_threshold Ns = Ns Sy 70 Ns % Pq uint Allow metaslabs to keep their active state as long as their fragmentation percentage is no more than this value. An active metaslab that exceeds this threshold will no longer keep its active status allowing better metaslabs to be selected. . .It Sy zfs_mg_fragmentation_threshold Ns = Ns Sy 95 Ns % Pq uint Metaslab groups are considered eligible for allocations if their fragmentation metric (measured as a percentage) is less than or equal to this value. If a metaslab group exceeds this threshold then it will be skipped unless all metaslab groups within the metaslab class have also crossed this threshold. . .It Sy zfs_mg_noalloc_threshold Ns = Ns Sy 0 Ns % Pq uint Defines a threshold at which metaslab groups should be eligible for allocations. The value is expressed as a percentage of free space beyond which a metaslab group is always eligible for allocations. If a metaslab group's free space is less than or equal to the threshold, the allocator will avoid allocating to that group unless all groups in the pool have reached the threshold. Once all groups have reached the threshold, all groups are allowed to accept allocations. The default value of .Sy 0 disables the feature and causes all metaslab groups to be eligible for allocations. .Pp This parameter allows one to deal with pools having heavily imbalanced vdevs such as would be the case when a new vdev has been added. Setting the threshold to a non-zero percentage will stop allocations from being made to vdevs that aren't filled to the specified percentage and allow lesser filled vdevs to acquire more allocations than they otherwise would under the old .Sy zfs_mg_alloc_failures facility. . .It Sy zfs_ddt_data_is_special Ns = Ns Sy 1 Ns | Ns 0 Pq int If enabled, ZFS will place DDT data into the special allocation class. . .It Sy zfs_user_indirect_is_special Ns = Ns Sy 1 Ns | Ns 0 Pq int If enabled, ZFS will place user data indirect blocks into the special allocation class. . .It Sy zfs_multihost_history Ns = Ns Sy 0 Pq uint Historical statistics for this many latest multihost updates will be available in .Pa /proc/spl/kstat/zfs/ Ns Ao Ar pool Ac Ns Pa /multihost . . .It Sy zfs_multihost_interval Ns = Ns Sy 1000 Ns ms Po 1 s Pc Pq u64 Used to control the frequency of multihost writes which are performed when the .Sy multihost pool property is on. This is one of the factors used to determine the length of the activity check during import. .Pp The multihost write period is .Sy zfs_multihost_interval No / Sy leaf-vdevs . On average a multihost write will be issued for each leaf vdev every .Sy zfs_multihost_interval milliseconds. In practice, the observed period can vary with the I/O load and this observed value is the delay which is stored in the uberblock. . .It Sy zfs_multihost_import_intervals Ns = Ns Sy 20 Pq uint Used to control the duration of the activity test on import. Smaller values of .Sy zfs_multihost_import_intervals will reduce the import time but increase the risk of failing to detect an active pool. The total activity check time is never allowed to drop below one second. .Pp On import the activity check waits a minimum amount of time determined by .Sy zfs_multihost_interval No \(mu Sy zfs_multihost_import_intervals , or the same product computed on the host which last had the pool imported, whichever is greater. The activity check time may be further extended if the value of MMP delay found in the best uberblock indicates actual multihost updates happened at longer intervals than .Sy zfs_multihost_interval . A minimum of .Em 100 ms is enforced. .Pp .Sy 0 No is equivalent to Sy 1 . . .It Sy zfs_multihost_fail_intervals Ns = Ns Sy 10 Pq uint Controls the behavior of the pool when multihost write failures or delays are detected. .Pp When .Sy 0 , multihost write failures or delays are ignored. The failures will still be reported to the ZED which depending on its configuration may take action such as suspending the pool or offlining a device. .Pp Otherwise, the pool will be suspended if .Sy zfs_multihost_fail_intervals No \(mu Sy zfs_multihost_interval milliseconds pass without a successful MMP write. This guarantees the activity test will see MMP writes if the pool is imported. .Sy 1 No is equivalent to Sy 2 ; this is necessary to prevent the pool from being suspended due to normal, small I/O latency variations. . .It Sy zfs_no_scrub_io Ns = Ns Sy 0 Ns | Ns 1 Pq int Set to disable scrub I/O. This results in scrubs not actually scrubbing data and simply doing a metadata crawl of the pool instead. . .It Sy zfs_no_scrub_prefetch Ns = Ns Sy 0 Ns | Ns 1 Pq int Set to disable block prefetching for scrubs. . .It Sy zfs_nocacheflush Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable cache flush operations on disks when writing. Setting this will cause pool corruption on power loss if a volatile out-of-order write cache is enabled. . .It Sy zfs_nopwrite_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Allow no-operation writes. The occurrence of nopwrites will further depend on other pool properties .Pq i.a. the checksumming and compression algorithms . . .It Sy zfs_dmu_offset_next_sync Ns = Ns Sy 1 Ns | Ns 0 Pq int Enable forcing TXG sync to find holes. When enabled forces ZFS to sync data when .Sy SEEK_HOLE No or Sy SEEK_DATA flags are used allowing holes in a file to be accurately reported. When disabled holes will not be reported in recently dirtied files. . .It Sy zfs_pd_bytes_max Ns = Ns Sy 52428800 Ns B Po 50 MiB Pc Pq int The number of bytes which should be prefetched during a pool traversal, like .Nm zfs Cm send or other data crawling operations. . .It Sy zfs_traverse_indirect_prefetch_limit Ns = Ns Sy 32 Pq uint The number of blocks pointed by indirect (non-L0) block which should be prefetched during a pool traversal, like .Nm zfs Cm send or other data crawling operations. . .It Sy zfs_per_txg_dirty_frees_percent Ns = Ns Sy 30 Ns % Pq u64 Control percentage of dirtied indirect blocks from frees allowed into one TXG. After this threshold is crossed, additional frees will wait until the next TXG. .Sy 0 No disables this throttle . . .It Sy zfs_prefetch_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable predictive prefetch. Note that it leaves "prescient" prefetch .Pq for, e.g., Nm zfs Cm send intact. Unlike predictive prefetch, prescient prefetch never issues I/O that ends up not being needed, so it can't hurt performance. . .It Sy zfs_qat_checksum_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable QAT hardware acceleration for SHA256 checksums. May be unset after the ZFS modules have been loaded to initialize the QAT hardware as long as support is compiled in and the QAT driver is present. . .It Sy zfs_qat_compress_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable QAT hardware acceleration for gzip compression. May be unset after the ZFS modules have been loaded to initialize the QAT hardware as long as support is compiled in and the QAT driver is present. . .It Sy zfs_qat_encrypt_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable QAT hardware acceleration for AES-GCM encryption. May be unset after the ZFS modules have been loaded to initialize the QAT hardware as long as support is compiled in and the QAT driver is present. . .It Sy zfs_vnops_read_chunk_size Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64 Bytes to read per chunk. . .It Sy zfs_read_history Ns = Ns Sy 0 Pq uint Historical statistics for this many latest reads will be available in .Pa /proc/spl/kstat/zfs/ Ns Ao Ar pool Ac Ns Pa /reads . . .It Sy zfs_read_history_hits Ns = Ns Sy 0 Ns | Ns 1 Pq int Include cache hits in read history . .It Sy zfs_rebuild_max_segment Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq u64 Maximum read segment size to issue when sequentially resilvering a top-level vdev. . .It Sy zfs_rebuild_scrub_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Automatically start a pool scrub when the last active sequential resilver completes in order to verify the checksums of all blocks which have been resilvered. This is enabled by default and strongly recommended. . .It Sy zfs_rebuild_vdev_limit Ns = Ns Sy 67108864 Ns B Po 64 MiB Pc Pq u64 Maximum amount of I/O that can be concurrently issued for a sequential resilver per leaf device, given in bytes. . .It Sy zfs_reconstruct_indirect_combinations_max Ns = Ns Sy 4096 Pq int If an indirect split block contains more than this many possible unique combinations when being reconstructed, consider it too computationally expensive to check them all. Instead, try at most this many randomly selected combinations each time the block is accessed. This allows all segment copies to participate fairly in the reconstruction when all combinations cannot be checked and prevents repeated use of one bad copy. . .It Sy zfs_recover Ns = Ns Sy 0 Ns | Ns 1 Pq int Set to attempt to recover from fatal errors. This should only be used as a last resort, as it typically results in leaked space, or worse. . .It Sy zfs_removal_ignore_errors Ns = Ns Sy 0 Ns | Ns 1 Pq int Ignore hard I/O errors during device removal. When set, if a device encounters a hard I/O error during the removal process the removal will not be cancelled. This can result in a normally recoverable block becoming permanently damaged and is hence not recommended. This should only be used as a last resort when the pool cannot be returned to a healthy state prior to removing the device. . .It Sy zfs_removal_suspend_progress Ns = Ns Sy 0 Ns | Ns 1 Pq uint This is used by the test suite so that it can ensure that certain actions happen while in the middle of a removal. . .It Sy zfs_remove_max_segment Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq uint The largest contiguous segment that we will attempt to allocate when removing a device. If there is a performance problem with attempting to allocate large blocks, consider decreasing this. The default value is also the maximum. . .It Sy zfs_resilver_disable_defer Ns = Ns Sy 0 Ns | Ns 1 Pq int Ignore the .Sy resilver_defer feature, causing an operation that would start a resilver to immediately restart the one in progress. . .It Sy zfs_resilver_min_time_ms Ns = Ns Sy 3000 Ns ms Po 3 s Pc Pq uint Resilvers are processed by the sync thread. While resilvering, it will spend at least this much time working on a resilver between TXG flushes. . .It Sy zfs_scan_ignore_errors Ns = Ns Sy 0 Ns | Ns 1 Pq int If set, remove the DTL (dirty time list) upon completion of a pool scan (scrub), even if there were unrepairable errors. Intended to be used during pool repair or recovery to stop resilvering when the pool is next imported. . .It Sy zfs_scrub_after_expand Ns = Ns Sy 1 Ns | Ns 0 Pq int Automatically start a pool scrub after a RAIDZ expansion completes in order to verify the checksums of all blocks which have been copied during the expansion. This is enabled by default and strongly recommended. . .It Sy zfs_scrub_min_time_ms Ns = Ns Sy 1000 Ns ms Po 1 s Pc Pq uint Scrubs are processed by the sync thread. While scrubbing, it will spend at least this much time working on a scrub between TXG flushes. . .It Sy zfs_scrub_error_blocks_per_txg Ns = Ns Sy 4096 Pq uint Error blocks to be scrubbed in one txg. . .It Sy zfs_scan_checkpoint_intval Ns = Ns Sy 7200 Ns s Po 2 hour Pc Pq uint To preserve progress across reboots, the sequential scan algorithm periodically needs to stop metadata scanning and issue all the verification I/O to disk. The frequency of this flushing is determined by this tunable. . .It Sy zfs_scan_fill_weight Ns = Ns Sy 3 Pq uint This tunable affects how scrub and resilver I/O segments are ordered. A higher number indicates that we care more about how filled in a segment is, while a lower number indicates we care more about the size of the extent without considering the gaps within a segment. This value is only tunable upon module insertion. Changing the value afterwards will have no effect on scrub or resilver performance. . .It Sy zfs_scan_issue_strategy Ns = Ns Sy 0 Pq uint Determines the order that data will be verified while scrubbing or resilvering: .Bl -tag -compact -offset 4n -width "a" .It Sy 1 Data will be verified as sequentially as possible, given the amount of memory reserved for scrubbing .Pq see Sy zfs_scan_mem_lim_fact . This may improve scrub performance if the pool's data is very fragmented. .It Sy 2 The largest mostly-contiguous chunk of found data will be verified first. By deferring scrubbing of small segments, we may later find adjacent data to coalesce and increase the segment size. .It Sy 0 .No Use strategy Sy 1 No during normal verification .No and strategy Sy 2 No while taking a checkpoint . .El . .It Sy zfs_scan_legacy Ns = Ns Sy 0 Ns | Ns 1 Pq int If unset, indicates that scrubs and resilvers will gather metadata in memory before issuing sequential I/O. Otherwise indicates that the legacy algorithm will be used, where I/O is initiated as soon as it is discovered. Unsetting will not affect scrubs or resilvers that are already in progress. . .It Sy zfs_scan_max_ext_gap Ns = Ns Sy 2097152 Ns B Po 2 MiB Pc Pq int Sets the largest gap in bytes between scrub/resilver I/O operations that will still be considered sequential for sorting purposes. Changing this value will not affect scrubs or resilvers that are already in progress. . .It Sy zfs_scan_mem_lim_fact Ns = Ns Sy 20 Ns ^-1 Pq uint Maximum fraction of RAM used for I/O sorting by sequential scan algorithm. This tunable determines the hard limit for I/O sorting memory usage. When the hard limit is reached we stop scanning metadata and start issuing data verification I/O. This is done until we get below the soft limit. . .It Sy zfs_scan_mem_lim_soft_fact Ns = Ns Sy 20 Ns ^-1 Pq uint The fraction of the hard limit used to determined the soft limit for I/O sorting by the sequential scan algorithm. When we cross this limit from below no action is taken. When we cross this limit from above it is because we are issuing verification I/O. In this case (unless the metadata scan is done) we stop issuing verification I/O and start scanning metadata again until we get to the hard limit. . .It Sy zfs_scan_report_txgs Ns = Ns Sy 0 Ns | Ns 1 Pq uint When reporting resilver throughput and estimated completion time use the performance observed over roughly the last .Sy zfs_scan_report_txgs TXGs. When set to zero performance is calculated over the time between checkpoints. . .It Sy zfs_scan_strict_mem_lim Ns = Ns Sy 0 Ns | Ns 1 Pq int Enforce tight memory limits on pool scans when a sequential scan is in progress. When disabled, the memory limit may be exceeded by fast disks. . .It Sy zfs_scan_suspend_progress Ns = Ns Sy 0 Ns | Ns 1 Pq int Freezes a scrub/resilver in progress without actually pausing it. Intended for testing/debugging. . .It Sy zfs_scan_vdev_limit Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq int Maximum amount of data that can be concurrently issued at once for scrubs and resilvers per leaf device, given in bytes. . .It Sy zfs_send_corrupt_data Ns = Ns Sy 0 Ns | Ns 1 Pq int Allow sending of corrupt data (ignore read/checksum errors when sending). . .It Sy zfs_send_unmodified_spill_blocks Ns = Ns Sy 1 Ns | Ns 0 Pq int Include unmodified spill blocks in the send stream. Under certain circumstances, previous versions of ZFS could incorrectly remove the spill block from an existing object. Including unmodified copies of the spill blocks creates a backwards-compatible stream which will recreate a spill block if it was incorrectly removed. . .It Sy zfs_send_no_prefetch_queue_ff Ns = Ns Sy 20 Ns ^\-1 Pq uint The fill fraction of the .Nm zfs Cm send internal queues. The fill fraction controls the timing with which internal threads are woken up. . .It Sy zfs_send_no_prefetch_queue_length Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq uint The maximum number of bytes allowed in .Nm zfs Cm send Ns 's internal queues. . .It Sy zfs_send_queue_ff Ns = Ns Sy 20 Ns ^\-1 Pq uint The fill fraction of the .Nm zfs Cm send prefetch queue. The fill fraction controls the timing with which internal threads are woken up. . .It Sy zfs_send_queue_length Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq uint The maximum number of bytes allowed that will be prefetched by .Nm zfs Cm send . This value must be at least twice the maximum block size in use. . .It Sy zfs_recv_queue_ff Ns = Ns Sy 20 Ns ^\-1 Pq uint The fill fraction of the .Nm zfs Cm receive queue. The fill fraction controls the timing with which internal threads are woken up. . .It Sy zfs_recv_queue_length Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq uint The maximum number of bytes allowed in the .Nm zfs Cm receive queue. This value must be at least twice the maximum block size in use. . .It Sy zfs_recv_write_batch_size Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq uint The maximum amount of data, in bytes, that .Nm zfs Cm receive will write in one DMU transaction. This is the uncompressed size, even when receiving a compressed send stream. This setting will not reduce the write size below a single block. Capped at a maximum of .Sy 32 MiB . . .It Sy zfs_recv_best_effort_corrective Ns = Ns Sy 0 Pq int When this variable is set to non-zero a corrective receive: .Bl -enum -compact -offset 4n -width "1." .It Does not enforce the restriction of source & destination snapshot GUIDs matching. .It If there is an error during healing, the healing receive is not terminated instead it moves on to the next record. .El . .It Sy zfs_override_estimate_recordsize Ns = Ns Sy 0 Ns | Ns 1 Pq uint Setting this variable overrides the default logic for estimating block sizes when doing a .Nm zfs Cm send . The default heuristic is that the average block size will be the current recordsize. Override this value if most data in your dataset is not of that size and you require accurate zfs send size estimates. . .It Sy zfs_sync_pass_deferred_free Ns = Ns Sy 2 Pq uint Flushing of data to disk is done in passes. Defer frees starting in this pass. . .It Sy zfs_spa_discard_memory_limit Ns = Ns Sy 16777216 Ns B Po 16 MiB Pc Pq int Maximum memory used for prefetching a checkpoint's space map on each vdev while discarding the checkpoint. . .It Sy zfs_special_class_metadata_reserve_pct Ns = Ns Sy 25 Ns % Pq uint Only allow small data blocks to be allocated on the special and dedup vdev types when the available free space percentage on these vdevs exceeds this value. This ensures reserved space is available for pool metadata as the special vdevs approach capacity. . .It Sy zfs_sync_pass_dont_compress Ns = Ns Sy 8 Pq uint Starting in this sync pass, disable compression (including of metadata). With the default setting, in practice, we don't have this many sync passes, so this has no effect. .Pp The original intent was that disabling compression would help the sync passes to converge. However, in practice, disabling compression increases the average number of sync passes; because when we turn compression off, many blocks' size will change, and thus we have to re-allocate (not overwrite) them. It also increases the number of .Em 128 KiB allocations (e.g. for indirect blocks and spacemaps) because these will not be compressed. The .Em 128 KiB allocations are especially detrimental to performance on highly fragmented systems, which may have very few free segments of this size, and may need to load new metaslabs to satisfy these allocations. . .It Sy zfs_sync_pass_rewrite Ns = Ns Sy 2 Pq uint Rewrite new block pointers starting in this pass. . .It Sy zfs_trim_extent_bytes_max Ns = Ns Sy 134217728 Ns B Po 128 MiB Pc Pq uint Maximum size of TRIM command. Larger ranges will be split into chunks no larger than this value before issuing. . .It Sy zfs_trim_extent_bytes_min Ns = Ns Sy 32768 Ns B Po 32 KiB Pc Pq uint Minimum size of TRIM commands. TRIM ranges smaller than this will be skipped, unless they're part of a larger range which was chunked. This is done because it's common for these small TRIMs to negatively impact overall performance. . .It Sy zfs_trim_metaslab_skip Ns = Ns Sy 0 Ns | Ns 1 Pq uint Skip uninitialized metaslabs during the TRIM process. This option is useful for pools constructed from large thinly-provisioned devices where TRIM operations are slow. As a pool ages, an increasing fraction of the pool's metaslabs will be initialized, progressively degrading the usefulness of this option. This setting is stored when starting a manual TRIM and will persist for the duration of the requested TRIM. . .It Sy zfs_trim_queue_limit Ns = Ns Sy 10 Pq uint Maximum number of queued TRIMs outstanding per leaf vdev. The number of concurrent TRIM commands issued to the device is controlled by .Sy zfs_vdev_trim_min_active No and Sy zfs_vdev_trim_max_active . . .It Sy zfs_trim_txg_batch Ns = Ns Sy 32 Pq uint The number of transaction groups' worth of frees which should be aggregated before TRIM operations are issued to the device. This setting represents a trade-off between issuing larger, more efficient TRIM operations and the delay before the recently trimmed space is available for use by the device. .Pp Increasing this value will allow frees to be aggregated for a longer time. This will result is larger TRIM operations and potentially increased memory usage. Decreasing this value will have the opposite effect. The default of .Sy 32 was determined to be a reasonable compromise. . .It Sy zfs_txg_history Ns = Ns Sy 0 Pq uint Historical statistics for this many latest TXGs will be available in .Pa /proc/spl/kstat/zfs/ Ns Ao Ar pool Ac Ns Pa /TXGs . . .It Sy zfs_txg_timeout Ns = Ns Sy 5 Ns s Pq uint Flush dirty data to disk at least every this many seconds (maximum TXG duration). . .It Sy zfs_vdev_aggregation_limit Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq uint Max vdev I/O aggregation size. . .It Sy zfs_vdev_aggregation_limit_non_rotating Ns = Ns Sy 131072 Ns B Po 128 KiB Pc Pq uint Max vdev I/O aggregation size for non-rotating media. . .It Sy zfs_vdev_mirror_rotating_inc Ns = Ns Sy 0 Pq int A number by which the balancing algorithm increments the load calculation for the purpose of selecting the least busy mirror member when an I/O operation immediately follows its predecessor on rotational vdevs for the purpose of making decisions based on load. . .It Sy zfs_vdev_mirror_rotating_seek_inc Ns = Ns Sy 5 Pq int A number by which the balancing algorithm increments the load calculation for the purpose of selecting the least busy mirror member when an I/O operation lacks locality as defined by .Sy zfs_vdev_mirror_rotating_seek_offset . Operations within this that are not immediately following the previous operation are incremented by half. . .It Sy zfs_vdev_mirror_rotating_seek_offset Ns = Ns Sy 1048576 Ns B Po 1 MiB Pc Pq int The maximum distance for the last queued I/O operation in which the balancing algorithm considers an operation to have locality. .No See Sx ZFS I/O SCHEDULER . . .It Sy zfs_vdev_mirror_non_rotating_inc Ns = Ns Sy 0 Pq int A number by which the balancing algorithm increments the load calculation for the purpose of selecting the least busy mirror member on non-rotational vdevs when I/O operations do not immediately follow one another. . .It Sy zfs_vdev_mirror_non_rotating_seek_inc Ns = Ns Sy 1 Pq int A number by which the balancing algorithm increments the load calculation for the purpose of selecting the least busy mirror member when an I/O operation lacks locality as defined by the .Sy zfs_vdev_mirror_rotating_seek_offset . Operations within this that are not immediately following the previous operation are incremented by half. . .It Sy zfs_vdev_read_gap_limit Ns = Ns Sy 32768 Ns B Po 32 KiB Pc Pq uint Aggregate read I/O operations if the on-disk gap between them is within this threshold. . .It Sy zfs_vdev_write_gap_limit Ns = Ns Sy 4096 Ns B Po 4 KiB Pc Pq uint Aggregate write I/O operations if the on-disk gap between them is within this threshold. . .It Sy zfs_vdev_raidz_impl Ns = Ns Sy fastest Pq string Select the raidz parity implementation to use. .Pp Variants that don't depend on CPU-specific features may be selected on module load, as they are supported on all systems. The remaining options may only be set after the module is loaded, as they are available only if the implementations are compiled in and supported on the running system. .Pp Once the module is loaded, .Pa /sys/module/zfs/parameters/zfs_vdev_raidz_impl will show the available options, with the currently selected one enclosed in square brackets. .Pp .TS lb l l . fastest selected by built-in benchmark original original implementation scalar scalar implementation sse2 SSE2 instruction set 64-bit x86 ssse3 SSSE3 instruction set 64-bit x86 avx2 AVX2 instruction set 64-bit x86 avx512f AVX512F instruction set 64-bit x86 avx512bw AVX512F & AVX512BW instruction sets 64-bit x86 aarch64_neon NEON Aarch64/64-bit ARMv8 aarch64_neonx2 NEON with more unrolling Aarch64/64-bit ARMv8 powerpc_altivec Altivec PowerPC .TE . .It Sy zfs_vdev_scheduler Pq charp .Sy DEPRECATED . Prints warning to kernel log for compatibility. . .It Sy zfs_zevent_len_max Ns = Ns Sy 512 Pq uint Max event queue length. Events in the queue can be viewed with .Xr zpool-events 8 . . .It Sy zfs_zevent_retain_max Ns = Ns Sy 2000 Pq int Maximum recent zevent records to retain for duplicate checking. Setting this to .Sy 0 disables duplicate detection. . .It Sy zfs_zevent_retain_expire_secs Ns = Ns Sy 900 Ns s Po 15 min Pc Pq int Lifespan for a recent ereport that was retained for duplicate checking. . .It Sy zfs_zil_clean_taskq_maxalloc Ns = Ns Sy 1048576 Pq int The maximum number of taskq entries that are allowed to be cached. When this limit is exceeded transaction records (itxs) will be cleaned synchronously. . .It Sy zfs_zil_clean_taskq_minalloc Ns = Ns Sy 1024 Pq int The number of taskq entries that are pre-populated when the taskq is first created and are immediately available for use. . .It Sy zfs_zil_clean_taskq_nthr_pct Ns = Ns Sy 100 Ns % Pq int This controls the number of threads used by .Sy dp_zil_clean_taskq . The default value of .Sy 100% will create a maximum of one thread per cpu. . .It Sy zil_maxblocksize Ns = Ns Sy 131072 Ns B Po 128 KiB Pc Pq uint This sets the maximum block size used by the ZIL. On very fragmented pools, lowering this .Pq typically to Sy 36 KiB can improve performance. . .It Sy zil_maxcopied Ns = Ns Sy 7680 Ns B Po 7.5 KiB Pc Pq uint This sets the maximum number of write bytes logged via WR_COPIED. It tunes a tradeoff between additional memory copy and possibly worse log space efficiency vs additional range lock/unlock. . .It Sy zil_nocacheflush Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable the cache flush commands that are normally sent to disk by the ZIL after an LWB write has completed. Setting this will cause ZIL corruption on power loss if a volatile out-of-order write cache is enabled. . .It Sy zil_replay_disable Ns = Ns Sy 0 Ns | Ns 1 Pq int Disable intent logging replay. Can be disabled for recovery from corrupted ZIL. . .It Sy zil_slog_bulk Ns = Ns Sy 67108864 Ns B Po 64 MiB Pc Pq u64 Limit SLOG write size per commit executed with synchronous priority. Any writes above that will be executed with lower (asynchronous) priority to limit potential SLOG device abuse by single active ZIL writer. . .It Sy zfs_zil_saxattr Ns = Ns Sy 1 Ns | Ns 0 Pq int Setting this tunable to zero disables ZIL logging of new .Sy xattr Ns = Ns Sy sa records if the .Sy org.openzfs:zilsaxattr feature is enabled on the pool. This would only be necessary to work around bugs in the ZIL logging or replay code for this record type. The tunable has no effect if the feature is disabled. . .It Sy zfs_embedded_slog_min_ms Ns = Ns Sy 64 Pq uint Usually, one metaslab from each normal-class vdev is dedicated for use by the ZIL to log synchronous writes. However, if there are fewer than .Sy zfs_embedded_slog_min_ms metaslabs in the vdev, this functionality is disabled. This ensures that we don't set aside an unreasonable amount of space for the ZIL. . .It Sy zstd_earlyabort_pass Ns = Ns Sy 1 Pq uint Whether heuristic for detection of incompressible data with zstd levels >= 3 using LZ4 and zstd-1 passes is enabled. . .It Sy zstd_abort_size Ns = Ns Sy 131072 Pq uint Minimal uncompressed size (inclusive) of a record before the early abort heuristic will be attempted. . .It Sy zio_deadman_log_all Ns = Ns Sy 0 Ns | Ns 1 Pq int If non-zero, the zio deadman will produce debugging messages .Pq see Sy zfs_dbgmsg_enable for all zios, rather than only for leaf zios possessing a vdev. This is meant to be used by developers to gain diagnostic information for hang conditions which don't involve a mutex or other locking primitive: typically conditions in which a thread in the zio pipeline is looping indefinitely. . .It Sy zio_slow_io_ms Ns = Ns Sy 30000 Ns ms Po 30 s Pc Pq int When an I/O operation takes more than this much time to complete, it's marked as slow. Each slow operation causes a delay zevent. Slow I/O counters can be seen with .Nm zpool Cm status Fl s . . .It Sy zio_dva_throttle_enabled Ns = Ns Sy 1 Ns | Ns 0 Pq int Throttle block allocations in the I/O pipeline. This allows for dynamic allocation distribution when devices are imbalanced. When enabled, the maximum number of pending allocations per top-level vdev is limited by .Sy zfs_vdev_queue_depth_pct . . .It Sy zfs_xattr_compat Ns = Ns 0 Ns | Ns 1 Pq int Control the naming scheme used when setting new xattrs in the user namespace. If .Sy 0 .Pq the default on Linux , user namespace xattr names are prefixed with the namespace, to be backwards compatible with previous versions of ZFS on Linux. If .Sy 1 .Pq the default on Fx , user namespace xattr names are not prefixed, to be backwards compatible with previous versions of ZFS on illumos and .Fx . .Pp Either naming scheme can be read on this and future versions of ZFS, regardless of this tunable, but legacy ZFS on illumos or .Fx are unable to read user namespace xattrs written in the Linux format, and legacy versions of ZFS on Linux are unable to read user namespace xattrs written in the legacy ZFS format. .Pp An existing xattr with the alternate naming scheme is removed when overwriting the xattr so as to not accumulate duplicates. . .It Sy zio_requeue_io_start_cut_in_line Ns = Ns Sy 0 Ns | Ns 1 Pq int Prioritize requeued I/O. . .It Sy zio_taskq_batch_pct Ns = Ns Sy 80 Ns % Pq uint Percentage of online CPUs which will run a worker thread for I/O. These workers are responsible for I/O work such as compression and checksum calculations. Fractional number of CPUs will be rounded down. .Pp The default value of .Sy 80% was chosen to avoid using all CPUs which can result in latency issues and inconsistent application performance, especially when slower compression and/or checksumming is enabled. . .It Sy zio_taskq_batch_tpq Ns = Ns Sy 0 Pq uint Number of worker threads per taskq. Lower values improve I/O ordering and CPU utilization, while higher reduces lock contention. .Pp If .Sy 0 , generate a system-dependent value close to 6 threads per taskq. . .It Sy zio_taskq_wr_iss_ncpus Ns = Ns Sy 0 Pq uint Determines the number of CPUs to run write issue taskqs. .Pp When 0 (the default), the value to use is computed internally as the number of actual CPUs in the system divided by the .Sy spa_num_allocators value. . .It Sy zvol_inhibit_dev Ns = Ns Sy 0 Ns | Ns 1 Pq uint Do not create zvol device nodes. This may slightly improve startup time on systems with a very large number of zvols. . .It Sy zvol_major Ns = Ns Sy 230 Pq uint Major number for zvol block devices. . .It Sy zvol_max_discard_blocks Ns = Ns Sy 16384 Pq long Discard (TRIM) operations done on zvols will be done in batches of this many blocks, where block size is determined by the .Sy volblocksize property of a zvol. . .It Sy zvol_prefetch_bytes Ns = Ns Sy 131072 Ns B Po 128 KiB Pc Pq uint When adding a zvol to the system, prefetch this many bytes from the start and end of the volume. Prefetching these regions of the volume is desirable, because they are likely to be accessed immediately by .Xr blkid 8 or the kernel partitioner. . .It Sy zvol_request_sync Ns = Ns Sy 0 Ns | Ns 1 Pq uint When processing I/O requests for a zvol, submit them synchronously. This effectively limits the queue depth to .Em 1 for each I/O submitter. When unset, requests are handled asynchronously by a thread pool. The number of requests which can be handled concurrently is controlled by .Sy zvol_threads . .Sy zvol_request_sync is ignored when running on a kernel that supports block multiqueue .Pq Li blk-mq . . .It Sy zvol_threads Ns = Ns Sy 0 Pq uint The number of system wide threads to use for processing zvol block IOs. If .Sy 0 (the default) then internally set .Sy zvol_threads to the number of CPUs present or 32 (whichever is greater). . .It Sy zvol_blk_mq_threads Ns = Ns Sy 0 Pq uint The number of threads per zvol to use for queuing IO requests. This parameter will only appear if your kernel supports .Li blk-mq and is only read and assigned to a zvol at zvol load time. If .Sy 0 (the default) then internally set .Sy zvol_blk_mq_threads to the number of CPUs present. . .It Sy zvol_use_blk_mq Ns = Ns Sy 0 Ns | Ns 1 Pq uint Set to .Sy 1 to use the .Li blk-mq API for zvols. Set to .Sy 0 (the default) to use the legacy zvol APIs. This setting can give better or worse zvol performance depending on the workload. This parameter will only appear if your kernel supports .Li blk-mq and is only read and assigned to a zvol at zvol load time. . .It Sy zvol_blk_mq_blocks_per_thread Ns = Ns Sy 8 Pq uint If .Sy zvol_use_blk_mq is enabled, then process this number of .Sy volblocksize Ns -sized blocks per zvol thread. This tunable can be use to favor better performance for zvol reads (lower values) or writes (higher values). If set to .Sy 0 , then the zvol layer will process the maximum number of blocks per thread that it can. This parameter will only appear if your kernel supports .Li blk-mq and is only applied at each zvol's load time. . .It Sy zvol_blk_mq_queue_depth Ns = Ns Sy 0 Pq uint The queue_depth value for the zvol .Li blk-mq interface. This parameter will only appear if your kernel supports .Li blk-mq and is only applied at each zvol's load time. If .Sy 0 (the default) then use the kernel's default queue depth. Values are clamped to the kernel's .Dv BLKDEV_MIN_RQ and .Dv BLKDEV_MAX_RQ Ns / Ns Dv BLKDEV_DEFAULT_RQ limits. . .It Sy zvol_volmode Ns = Ns Sy 1 Pq uint Defines zvol block devices behaviour when .Sy volmode Ns = Ns Sy default : .Bl -tag -compact -offset 4n -width "a" .It Sy 1 .No equivalent to Sy full .It Sy 2 .No equivalent to Sy dev .It Sy 3 .No equivalent to Sy none .El . .It Sy zvol_enforce_quotas Ns = Ns Sy 0 Ns | Ns 1 Pq uint Enable strict ZVOL quota enforcement. The strict quota enforcement may have a performance impact. .El . .Sh ZFS I/O SCHEDULER ZFS issues I/O operations to leaf vdevs to satisfy and complete I/O operations. The scheduler determines when and in what order those operations are issued. The scheduler divides operations into five I/O classes, prioritized in the following order: sync read, sync write, async read, async write, and scrub/resilver. Each queue defines the minimum and maximum number of concurrent operations that may be issued to the device. In addition, the device has an aggregate maximum, .Sy zfs_vdev_max_active . Note that the sum of the per-queue minima must not exceed the aggregate maximum. If the sum of the per-queue maxima exceeds the aggregate maximum, then the number of active operations may reach .Sy zfs_vdev_max_active , in which case no further operations will be issued, regardless of whether all per-queue minima have been met. .Pp For many physical devices, throughput increases with the number of concurrent operations, but latency typically suffers. Furthermore, physical devices typically have a limit at which more concurrent operations have no effect on throughput or can actually cause it to decrease. .Pp The scheduler selects the next operation to issue by first looking for an I/O class whose minimum has not been satisfied. Once all are satisfied and the aggregate maximum has not been hit, the scheduler looks for classes whose maximum has not been satisfied. Iteration through the I/O classes is done in the order specified above. No further operations are issued if the aggregate maximum number of concurrent operations has been hit, or if there are no operations queued for an I/O class that has not hit its maximum. Every time an I/O operation is queued or an operation completes, the scheduler looks for new operations to issue. .Pp In general, smaller .Sy max_active Ns s will lead to lower latency of synchronous operations. Larger .Sy max_active Ns s may lead to higher overall throughput, depending on underlying storage. .Pp The ratio of the queues' .Sy max_active Ns s determines the balance of performance between reads, writes, and scrubs. For example, increasing .Sy zfs_vdev_scrub_max_active will cause the scrub or resilver to complete more quickly, but reads and writes to have higher latency and lower throughput. .Pp All I/O classes have a fixed maximum number of outstanding operations, except for the async write class. Asynchronous writes represent the data that is committed to stable storage during the syncing stage for transaction groups. Transaction groups enter the syncing state periodically, so the number of queued async writes will quickly burst up and then bleed down to zero. Rather than servicing them as quickly as possible, the I/O scheduler changes the maximum number of active async write operations according to the amount of dirty data in the pool. Since both throughput and latency typically increase with the number of concurrent operations issued to physical devices, reducing the burstiness in the number of simultaneous operations also stabilizes the response time of operations from other queues, in particular synchronous ones. In broad strokes, the I/O scheduler will issue more concurrent operations from the async write queue as there is more dirty data in the pool. . .Ss Async Writes The number of concurrent operations issued for the async write I/O class follows a piece-wise linear function defined by a few adjustable points: .Bd -literal | o---------| <-- \fBzfs_vdev_async_write_max_active\fP ^ | /^ | | | / | | active | / | | I/O | / | | count | / | | | / | | |-------o | | <-- \fBzfs_vdev_async_write_min_active\fP 0|_______^______|_________| 0% | | 100% of \fBzfs_dirty_data_max\fP | | | `-- \fBzfs_vdev_async_write_active_max_dirty_percent\fP `--------- \fBzfs_vdev_async_write_active_min_dirty_percent\fP .Ed .Pp Until the amount of dirty data exceeds a minimum percentage of the dirty data allowed in the pool, the I/O scheduler will limit the number of concurrent operations to the minimum. As that threshold is crossed, the number of concurrent operations issued increases linearly to the maximum at the specified maximum percentage of the dirty data allowed in the pool. .Pp Ideally, the amount of dirty data on a busy pool will stay in the sloped part of the function between .Sy zfs_vdev_async_write_active_min_dirty_percent and .Sy zfs_vdev_async_write_active_max_dirty_percent . If it exceeds the maximum percentage, this indicates that the rate of incoming data is greater than the rate that the backend storage can handle. In this case, we must further throttle incoming writes, as described in the next section. . .Sh ZFS TRANSACTION DELAY We delay transactions when we've determined that the backend storage isn't able to accommodate the rate of incoming writes. .Pp If there is already a transaction waiting, we delay relative to when that transaction will finish waiting. This way the calculated delay time is independent of the number of threads concurrently executing transactions. .Pp If we are the only waiter, wait relative to when the transaction started, rather than the current time. This credits the transaction for "time already served", e.g. reading indirect blocks. .Pp The minimum time for a transaction to take is calculated as .D1 min_time = min( Ns Sy zfs_delay_scale No \(mu Po Sy dirty No \- Sy min Pc / Po Sy max No \- Sy dirty Pc , 100ms) .Pp The delay has two degrees of freedom that can be adjusted via tunables. The percentage of dirty data at which we start to delay is defined by .Sy zfs_delay_min_dirty_percent . This should typically be at or above .Sy zfs_vdev_async_write_active_max_dirty_percent , so that we only start to delay after writing at full speed has failed to keep up with the incoming write rate. The scale of the curve is defined by .Sy zfs_delay_scale . Roughly speaking, this variable determines the amount of delay at the midpoint of the curve. .Bd -literal delay 10ms +-------------------------------------------------------------*+ | *| 9ms + *+ | *| 8ms + *+ | * | 7ms + * + | * | 6ms + * + | * | 5ms + * + | * | 4ms + * + | * | 3ms + * + | * | 2ms + (midpoint) * + | | ** | 1ms + v *** + | \fBzfs_delay_scale\fP ----------> ******** | 0 +-------------------------------------*********----------------+ 0% <- \fBzfs_dirty_data_max\fP -> 100% .Ed .Pp Note, that since the delay is added to the outstanding time remaining on the most recent transaction it's effectively the inverse of IOPS. Here, the midpoint of .Em 500 us translates to .Em 2000 IOPS . The shape of the curve was chosen such that small changes in the amount of accumulated dirty data in the first three quarters of the curve yield relatively small differences in the amount of delay. .Pp The effects can be easier to understand when the amount of delay is represented on a logarithmic scale: .Bd -literal delay 100ms +-------------------------------------------------------------++ + + | | + *+ 10ms + *+ + ** + | (midpoint) ** | + | ** + 1ms + v **** + + \fBzfs_delay_scale\fP ----------> ***** + | **** | + **** + 100us + ** + + * + | * | + * + 10us + * + + + | | + + +--------------------------------------------------------------+ 0% <- \fBzfs_dirty_data_max\fP -> 100% .Ed .Pp Note here that only as the amount of dirty data approaches its limit does the delay start to increase rapidly. The goal of a properly tuned system should be to keep the amount of dirty data out of that range by first ensuring that the appropriate limits are set for the I/O scheduler to reach optimal throughput on the back-end storage, and then by changing the value of .Sy zfs_delay_scale to increase the steepness of the curve. diff --git a/module/os/freebsd/zfs/zfs_vfsops.c b/module/os/freebsd/zfs/zfs_vfsops.c index a972c720dfdb..f2d5391037c4 100644 --- a/module/os/freebsd/zfs/zfs_vfsops.c +++ b/module/os/freebsd/zfs/zfs_vfsops.c @@ -1,2547 +1,2551 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2011 Pawel Jakub Dawidek . * All rights reserved. * Copyright (c) 2012, 2015 by Delphix. All rights reserved. * Copyright (c) 2014 Integros [integros.com] * Copyright 2016 Nexenta Systems, Inc. All rights reserved. */ /* Portions Copyright 2010 Robert Milkowski */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "zfs_comutil.h" #ifndef MNTK_VMSETSIZE_BUG #define MNTK_VMSETSIZE_BUG 0 #endif #ifndef MNTK_NOMSYNC #define MNTK_NOMSYNC 8 #endif struct mtx zfs_debug_mtx; MTX_SYSINIT(zfs_debug_mtx, &zfs_debug_mtx, "zfs_debug", MTX_DEF); SYSCTL_NODE(_vfs, OID_AUTO, zfs, CTLFLAG_RW, 0, "ZFS file system"); int zfs_super_owner; SYSCTL_INT(_vfs_zfs, OID_AUTO, super_owner, CTLFLAG_RW, &zfs_super_owner, 0, "File system owners can perform privileged operation on file systems"); int zfs_debug_level; SYSCTL_INT(_vfs_zfs, OID_AUTO, debug, CTLFLAG_RWTUN, &zfs_debug_level, 0, "Debug level"); +int zfs_bclone_enabled = 1; +SYSCTL_INT(_vfs_zfs, OID_AUTO, bclone_enabled, CTLFLAG_RWTUN, + &zfs_bclone_enabled, 0, "Enable block cloning"); + struct zfs_jailparam { int mount_snapshot; }; static struct zfs_jailparam zfs_jailparam0 = { .mount_snapshot = 0, }; static int zfs_jailparam_slot; SYSCTL_JAIL_PARAM_SYS_NODE(zfs, CTLFLAG_RW, "Jail ZFS parameters"); SYSCTL_JAIL_PARAM(_zfs, mount_snapshot, CTLTYPE_INT | CTLFLAG_RW, "I", "Allow mounting snapshots in the .zfs directory for unjailed datasets"); SYSCTL_NODE(_vfs_zfs, OID_AUTO, version, CTLFLAG_RD, 0, "ZFS versions"); static int zfs_version_acl = ZFS_ACL_VERSION; SYSCTL_INT(_vfs_zfs_version, OID_AUTO, acl, CTLFLAG_RD, &zfs_version_acl, 0, "ZFS_ACL_VERSION"); static int zfs_version_spa = SPA_VERSION; SYSCTL_INT(_vfs_zfs_version, OID_AUTO, spa, CTLFLAG_RD, &zfs_version_spa, 0, "SPA_VERSION"); static int zfs_version_zpl = ZPL_VERSION; SYSCTL_INT(_vfs_zfs_version, OID_AUTO, zpl, CTLFLAG_RD, &zfs_version_zpl, 0, "ZPL_VERSION"); #if __FreeBSD_version >= 1400018 static int zfs_quotactl(vfs_t *vfsp, int cmds, uid_t id, void *arg, bool *mp_busy); #else static int zfs_quotactl(vfs_t *vfsp, int cmds, uid_t id, void *arg); #endif static int zfs_mount(vfs_t *vfsp); static int zfs_umount(vfs_t *vfsp, int fflag); static int zfs_root(vfs_t *vfsp, int flags, vnode_t **vpp); static int zfs_statfs(vfs_t *vfsp, struct statfs *statp); static int zfs_vget(vfs_t *vfsp, ino_t ino, int flags, vnode_t **vpp); static int zfs_sync(vfs_t *vfsp, int waitfor); #if __FreeBSD_version >= 1300098 static int zfs_checkexp(vfs_t *vfsp, struct sockaddr *nam, uint64_t *extflagsp, struct ucred **credanonp, int *numsecflavors, int *secflavors); #else static int zfs_checkexp(vfs_t *vfsp, struct sockaddr *nam, int *extflagsp, struct ucred **credanonp, int *numsecflavors, int **secflavors); #endif static int zfs_fhtovp(vfs_t *vfsp, fid_t *fidp, int flags, vnode_t **vpp); static void zfs_freevfs(vfs_t *vfsp); struct vfsops zfs_vfsops = { .vfs_mount = zfs_mount, .vfs_unmount = zfs_umount, #if __FreeBSD_version >= 1300049 .vfs_root = vfs_cache_root, .vfs_cachedroot = zfs_root, #else .vfs_root = zfs_root, #endif .vfs_statfs = zfs_statfs, .vfs_vget = zfs_vget, .vfs_sync = zfs_sync, .vfs_checkexp = zfs_checkexp, .vfs_fhtovp = zfs_fhtovp, .vfs_quotactl = zfs_quotactl, }; #ifdef VFCF_CROSS_COPY_FILE_RANGE VFS_SET(zfs_vfsops, zfs, VFCF_DELEGADMIN | VFCF_JAIL | VFCF_CROSS_COPY_FILE_RANGE); #else VFS_SET(zfs_vfsops, zfs, VFCF_DELEGADMIN | VFCF_JAIL); #endif /* * We need to keep a count of active fs's. * This is necessary to prevent our module * from being unloaded after a umount -f */ static uint32_t zfs_active_fs_count = 0; int zfs_get_temporary_prop(dsl_dataset_t *ds, zfs_prop_t zfs_prop, uint64_t *val, char *setpoint) { int error; zfsvfs_t *zfvp; vfs_t *vfsp; objset_t *os; uint64_t tmp = *val; error = dmu_objset_from_ds(ds, &os); if (error != 0) return (error); error = getzfsvfs_impl(os, &zfvp); if (error != 0) return (error); if (zfvp == NULL) return (ENOENT); vfsp = zfvp->z_vfs; switch (zfs_prop) { case ZFS_PROP_ATIME: if (vfs_optionisset(vfsp, MNTOPT_NOATIME, NULL)) tmp = 0; if (vfs_optionisset(vfsp, MNTOPT_ATIME, NULL)) tmp = 1; break; case ZFS_PROP_DEVICES: if (vfs_optionisset(vfsp, MNTOPT_NODEVICES, NULL)) tmp = 0; if (vfs_optionisset(vfsp, MNTOPT_DEVICES, NULL)) tmp = 1; break; case ZFS_PROP_EXEC: if (vfs_optionisset(vfsp, MNTOPT_NOEXEC, NULL)) tmp = 0; if (vfs_optionisset(vfsp, MNTOPT_EXEC, NULL)) tmp = 1; break; case ZFS_PROP_SETUID: if (vfs_optionisset(vfsp, MNTOPT_NOSETUID, NULL)) tmp = 0; if (vfs_optionisset(vfsp, MNTOPT_SETUID, NULL)) tmp = 1; break; case ZFS_PROP_READONLY: if (vfs_optionisset(vfsp, MNTOPT_RW, NULL)) tmp = 0; if (vfs_optionisset(vfsp, MNTOPT_RO, NULL)) tmp = 1; break; case ZFS_PROP_XATTR: if (zfvp->z_flags & ZSB_XATTR) tmp = zfvp->z_xattr; break; case ZFS_PROP_NBMAND: if (vfs_optionisset(vfsp, MNTOPT_NONBMAND, NULL)) tmp = 0; if (vfs_optionisset(vfsp, MNTOPT_NBMAND, NULL)) tmp = 1; break; default: vfs_unbusy(vfsp); return (ENOENT); } vfs_unbusy(vfsp); if (tmp != *val) { if (setpoint) (void) strcpy(setpoint, "temporary"); *val = tmp; } return (0); } static int zfs_getquota(zfsvfs_t *zfsvfs, uid_t id, int isgroup, struct dqblk64 *dqp) { int error = 0; char buf[32]; uint64_t usedobj, quotaobj; uint64_t quota, used = 0; timespec_t now; usedobj = isgroup ? DMU_GROUPUSED_OBJECT : DMU_USERUSED_OBJECT; quotaobj = isgroup ? zfsvfs->z_groupquota_obj : zfsvfs->z_userquota_obj; if (quotaobj == 0 || zfsvfs->z_replay) { error = ENOENT; goto done; } (void) sprintf(buf, "%llx", (longlong_t)id); if ((error = zap_lookup(zfsvfs->z_os, quotaobj, buf, sizeof (quota), 1, "a)) != 0) { dprintf("%s(%d): quotaobj lookup failed\n", __FUNCTION__, __LINE__); goto done; } /* * quota(8) uses bsoftlimit as "quoota", and hardlimit as "limit". * So we set them to be the same. */ dqp->dqb_bsoftlimit = dqp->dqb_bhardlimit = btodb(quota); error = zap_lookup(zfsvfs->z_os, usedobj, buf, sizeof (used), 1, &used); if (error && error != ENOENT) { dprintf("%s(%d): usedobj failed; %d\n", __FUNCTION__, __LINE__, error); goto done; } dqp->dqb_curblocks = btodb(used); dqp->dqb_ihardlimit = dqp->dqb_isoftlimit = 0; vfs_timestamp(&now); /* * Setting this to 0 causes FreeBSD quota(8) to print * the number of days since the epoch, which isn't * particularly useful. */ dqp->dqb_btime = dqp->dqb_itime = now.tv_sec; done: return (error); } static int #if __FreeBSD_version >= 1400018 zfs_quotactl(vfs_t *vfsp, int cmds, uid_t id, void *arg, bool *mp_busy) #else zfs_quotactl(vfs_t *vfsp, int cmds, uid_t id, void *arg) #endif { zfsvfs_t *zfsvfs = vfsp->vfs_data; struct thread *td; int cmd, type, error = 0; int bitsize; zfs_userquota_prop_t quota_type; struct dqblk64 dqblk = { 0 }; td = curthread; cmd = cmds >> SUBCMDSHIFT; type = cmds & SUBCMDMASK; if ((error = zfs_enter(zfsvfs, FTAG)) != 0) return (error); if (id == -1) { switch (type) { case USRQUOTA: id = td->td_ucred->cr_ruid; break; case GRPQUOTA: id = td->td_ucred->cr_rgid; break; default: error = EINVAL; #if __FreeBSD_version < 1400018 if (cmd == Q_QUOTAON || cmd == Q_QUOTAOFF) vfs_unbusy(vfsp); #endif goto done; } } /* * Map BSD type to: * ZFS_PROP_USERUSED, * ZFS_PROP_USERQUOTA, * ZFS_PROP_GROUPUSED, * ZFS_PROP_GROUPQUOTA */ switch (cmd) { case Q_SETQUOTA: case Q_SETQUOTA32: if (type == USRQUOTA) quota_type = ZFS_PROP_USERQUOTA; else if (type == GRPQUOTA) quota_type = ZFS_PROP_GROUPQUOTA; else error = EINVAL; break; case Q_GETQUOTA: case Q_GETQUOTA32: if (type == USRQUOTA) quota_type = ZFS_PROP_USERUSED; else if (type == GRPQUOTA) quota_type = ZFS_PROP_GROUPUSED; else error = EINVAL; break; } /* * Depending on the cmd, we may need to get * the ruid and domain (see fuidstr_to_sid?), * the fuid (how?), or other information. * Create fuid using zfs_fuid_create(zfsvfs, id, * ZFS_OWNER or ZFS_GROUP, cr, &fuidp)? * I think I can use just the id? * * Look at zfs_id_overquota() to look up a quota. * zap_lookup(something, quotaobj, fuidstring, * sizeof (long long), 1, "a) * * See zfs_set_userquota() to set a quota. */ if ((uint32_t)type >= MAXQUOTAS) { error = EINVAL; goto done; } switch (cmd) { case Q_GETQUOTASIZE: bitsize = 64; error = copyout(&bitsize, arg, sizeof (int)); break; case Q_QUOTAON: // As far as I can tell, you can't turn quotas on or off on zfs error = 0; #if __FreeBSD_version < 1400018 vfs_unbusy(vfsp); #endif break; case Q_QUOTAOFF: error = ENOTSUP; #if __FreeBSD_version < 1400018 vfs_unbusy(vfsp); #endif break; case Q_SETQUOTA: error = copyin(arg, &dqblk, sizeof (dqblk)); if (error == 0) error = zfs_set_userquota(zfsvfs, quota_type, "", id, dbtob(dqblk.dqb_bhardlimit)); break; case Q_GETQUOTA: error = zfs_getquota(zfsvfs, id, type == GRPQUOTA, &dqblk); if (error == 0) error = copyout(&dqblk, arg, sizeof (dqblk)); break; default: error = EINVAL; break; } done: zfs_exit(zfsvfs, FTAG); return (error); } boolean_t zfs_is_readonly(zfsvfs_t *zfsvfs) { return (!!(zfsvfs->z_vfs->vfs_flag & VFS_RDONLY)); } static int zfs_sync(vfs_t *vfsp, int waitfor) { /* * Data integrity is job one. We don't want a compromised kernel * writing to the storage pool, so we never sync during panic. */ if (panicstr) return (0); /* * Ignore the system syncher. ZFS already commits async data * at zfs_txg_timeout intervals. */ if (waitfor == MNT_LAZY) return (0); if (vfsp != NULL) { /* * Sync a specific filesystem. */ zfsvfs_t *zfsvfs = vfsp->vfs_data; dsl_pool_t *dp; int error; if ((error = zfs_enter(zfsvfs, FTAG)) != 0) return (error); dp = dmu_objset_pool(zfsvfs->z_os); /* * If the system is shutting down, then skip any * filesystems which may exist on a suspended pool. */ if (rebooting && spa_suspended(dp->dp_spa)) { zfs_exit(zfsvfs, FTAG); return (0); } if (zfsvfs->z_log != NULL) zil_commit(zfsvfs->z_log, 0); zfs_exit(zfsvfs, FTAG); } else { /* * Sync all ZFS filesystems. This is what happens when you * run sync(8). Unlike other filesystems, ZFS honors the * request by waiting for all pools to commit all dirty data. */ spa_sync_allpools(); } return (0); } static void atime_changed_cb(void *arg, uint64_t newval) { zfsvfs_t *zfsvfs = arg; if (newval == TRUE) { zfsvfs->z_atime = TRUE; zfsvfs->z_vfs->vfs_flag &= ~MNT_NOATIME; vfs_clearmntopt(zfsvfs->z_vfs, MNTOPT_NOATIME); vfs_setmntopt(zfsvfs->z_vfs, MNTOPT_ATIME, NULL, 0); } else { zfsvfs->z_atime = FALSE; zfsvfs->z_vfs->vfs_flag |= MNT_NOATIME; vfs_clearmntopt(zfsvfs->z_vfs, MNTOPT_ATIME); vfs_setmntopt(zfsvfs->z_vfs, MNTOPT_NOATIME, NULL, 0); } } static void xattr_changed_cb(void *arg, uint64_t newval) { zfsvfs_t *zfsvfs = arg; if (newval == ZFS_XATTR_OFF) { zfsvfs->z_flags &= ~ZSB_XATTR; } else { zfsvfs->z_flags |= ZSB_XATTR; if (newval == ZFS_XATTR_SA) zfsvfs->z_xattr_sa = B_TRUE; else zfsvfs->z_xattr_sa = B_FALSE; } } static void blksz_changed_cb(void *arg, uint64_t newval) { zfsvfs_t *zfsvfs = arg; ASSERT3U(newval, <=, spa_maxblocksize(dmu_objset_spa(zfsvfs->z_os))); ASSERT3U(newval, >=, SPA_MINBLOCKSIZE); ASSERT(ISP2(newval)); zfsvfs->z_max_blksz = newval; zfsvfs->z_vfs->mnt_stat.f_iosize = newval; } static void readonly_changed_cb(void *arg, uint64_t newval) { zfsvfs_t *zfsvfs = arg; if (newval) { /* XXX locking on vfs_flag? */ zfsvfs->z_vfs->vfs_flag |= VFS_RDONLY; vfs_clearmntopt(zfsvfs->z_vfs, MNTOPT_RW); vfs_setmntopt(zfsvfs->z_vfs, MNTOPT_RO, NULL, 0); } else { /* XXX locking on vfs_flag? */ zfsvfs->z_vfs->vfs_flag &= ~VFS_RDONLY; vfs_clearmntopt(zfsvfs->z_vfs, MNTOPT_RO); vfs_setmntopt(zfsvfs->z_vfs, MNTOPT_RW, NULL, 0); } } static void setuid_changed_cb(void *arg, uint64_t newval) { zfsvfs_t *zfsvfs = arg; if (newval == FALSE) { zfsvfs->z_vfs->vfs_flag |= VFS_NOSETUID; vfs_clearmntopt(zfsvfs->z_vfs, MNTOPT_SETUID); vfs_setmntopt(zfsvfs->z_vfs, MNTOPT_NOSETUID, NULL, 0); } else { zfsvfs->z_vfs->vfs_flag &= ~VFS_NOSETUID; vfs_clearmntopt(zfsvfs->z_vfs, MNTOPT_NOSETUID); vfs_setmntopt(zfsvfs->z_vfs, MNTOPT_SETUID, NULL, 0); } } static void exec_changed_cb(void *arg, uint64_t newval) { zfsvfs_t *zfsvfs = arg; if (newval == FALSE) { zfsvfs->z_vfs->vfs_flag |= VFS_NOEXEC; vfs_clearmntopt(zfsvfs->z_vfs, MNTOPT_EXEC); vfs_setmntopt(zfsvfs->z_vfs, MNTOPT_NOEXEC, NULL, 0); } else { zfsvfs->z_vfs->vfs_flag &= ~VFS_NOEXEC; vfs_clearmntopt(zfsvfs->z_vfs, MNTOPT_NOEXEC); vfs_setmntopt(zfsvfs->z_vfs, MNTOPT_EXEC, NULL, 0); } } /* * The nbmand mount option can be changed at mount time. * We can't allow it to be toggled on live file systems or incorrect * behavior may be seen from cifs clients * * This property isn't registered via dsl_prop_register(), but this callback * will be called when a file system is first mounted */ static void nbmand_changed_cb(void *arg, uint64_t newval) { zfsvfs_t *zfsvfs = arg; if (newval == FALSE) { vfs_clearmntopt(zfsvfs->z_vfs, MNTOPT_NBMAND); vfs_setmntopt(zfsvfs->z_vfs, MNTOPT_NONBMAND, NULL, 0); } else { vfs_clearmntopt(zfsvfs->z_vfs, MNTOPT_NONBMAND); vfs_setmntopt(zfsvfs->z_vfs, MNTOPT_NBMAND, NULL, 0); } } static void snapdir_changed_cb(void *arg, uint64_t newval) { zfsvfs_t *zfsvfs = arg; zfsvfs->z_show_ctldir = newval; } static void acl_mode_changed_cb(void *arg, uint64_t newval) { zfsvfs_t *zfsvfs = arg; zfsvfs->z_acl_mode = newval; } static void acl_inherit_changed_cb(void *arg, uint64_t newval) { zfsvfs_t *zfsvfs = arg; zfsvfs->z_acl_inherit = newval; } static void acl_type_changed_cb(void *arg, uint64_t newval) { zfsvfs_t *zfsvfs = arg; zfsvfs->z_acl_type = newval; } static int zfs_register_callbacks(vfs_t *vfsp) { struct dsl_dataset *ds = NULL; objset_t *os = NULL; zfsvfs_t *zfsvfs = NULL; uint64_t nbmand; boolean_t readonly = B_FALSE; boolean_t do_readonly = B_FALSE; boolean_t setuid = B_FALSE; boolean_t do_setuid = B_FALSE; boolean_t exec = B_FALSE; boolean_t do_exec = B_FALSE; boolean_t xattr = B_FALSE; boolean_t atime = B_FALSE; boolean_t do_atime = B_FALSE; boolean_t do_xattr = B_FALSE; int error = 0; ASSERT3P(vfsp, !=, NULL); zfsvfs = vfsp->vfs_data; ASSERT3P(zfsvfs, !=, NULL); os = zfsvfs->z_os; /* * This function can be called for a snapshot when we update snapshot's * mount point, which isn't really supported. */ if (dmu_objset_is_snapshot(os)) return (EOPNOTSUPP); /* * The act of registering our callbacks will destroy any mount * options we may have. In order to enable temporary overrides * of mount options, we stash away the current values and * restore them after we register the callbacks. */ if (vfs_optionisset(vfsp, MNTOPT_RO, NULL) || !spa_writeable(dmu_objset_spa(os))) { readonly = B_TRUE; do_readonly = B_TRUE; } else if (vfs_optionisset(vfsp, MNTOPT_RW, NULL)) { readonly = B_FALSE; do_readonly = B_TRUE; } if (vfs_optionisset(vfsp, MNTOPT_NOSETUID, NULL)) { setuid = B_FALSE; do_setuid = B_TRUE; } else if (vfs_optionisset(vfsp, MNTOPT_SETUID, NULL)) { setuid = B_TRUE; do_setuid = B_TRUE; } if (vfs_optionisset(vfsp, MNTOPT_NOEXEC, NULL)) { exec = B_FALSE; do_exec = B_TRUE; } else if (vfs_optionisset(vfsp, MNTOPT_EXEC, NULL)) { exec = B_TRUE; do_exec = B_TRUE; } if (vfs_optionisset(vfsp, MNTOPT_NOXATTR, NULL)) { zfsvfs->z_xattr = xattr = ZFS_XATTR_OFF; do_xattr = B_TRUE; } else if (vfs_optionisset(vfsp, MNTOPT_XATTR, NULL)) { zfsvfs->z_xattr = xattr = ZFS_XATTR_DIR; do_xattr = B_TRUE; } else if (vfs_optionisset(vfsp, MNTOPT_DIRXATTR, NULL)) { zfsvfs->z_xattr = xattr = ZFS_XATTR_DIR; do_xattr = B_TRUE; } else if (vfs_optionisset(vfsp, MNTOPT_SAXATTR, NULL)) { zfsvfs->z_xattr = xattr = ZFS_XATTR_SA; do_xattr = B_TRUE; } if (vfs_optionisset(vfsp, MNTOPT_NOATIME, NULL)) { atime = B_FALSE; do_atime = B_TRUE; } else if (vfs_optionisset(vfsp, MNTOPT_ATIME, NULL)) { atime = B_TRUE; do_atime = B_TRUE; } /* * We need to enter pool configuration here, so that we can use * dsl_prop_get_int_ds() to handle the special nbmand property below. * dsl_prop_get_integer() can not be used, because it has to acquire * spa_namespace_lock and we can not do that because we already hold * z_teardown_lock. The problem is that spa_write_cachefile() is called * with spa_namespace_lock held and the function calls ZFS vnode * operations to write the cache file and thus z_teardown_lock is * acquired after spa_namespace_lock. */ ds = dmu_objset_ds(os); dsl_pool_config_enter(dmu_objset_pool(os), FTAG); /* * nbmand is a special property. It can only be changed at * mount time. * * This is weird, but it is documented to only be changeable * at mount time. */ if (vfs_optionisset(vfsp, MNTOPT_NONBMAND, NULL)) { nbmand = B_FALSE; } else if (vfs_optionisset(vfsp, MNTOPT_NBMAND, NULL)) { nbmand = B_TRUE; } else if ((error = dsl_prop_get_int_ds(ds, "nbmand", &nbmand)) != 0) { dsl_pool_config_exit(dmu_objset_pool(os), FTAG); return (error); } /* * Register property callbacks. * * It would probably be fine to just check for i/o error from * the first prop_register(), but I guess I like to go * overboard... */ error = dsl_prop_register(ds, zfs_prop_to_name(ZFS_PROP_ATIME), atime_changed_cb, zfsvfs); error = error ? error : dsl_prop_register(ds, zfs_prop_to_name(ZFS_PROP_XATTR), xattr_changed_cb, zfsvfs); error = error ? error : dsl_prop_register(ds, zfs_prop_to_name(ZFS_PROP_RECORDSIZE), blksz_changed_cb, zfsvfs); error = error ? error : dsl_prop_register(ds, zfs_prop_to_name(ZFS_PROP_READONLY), readonly_changed_cb, zfsvfs); error = error ? error : dsl_prop_register(ds, zfs_prop_to_name(ZFS_PROP_SETUID), setuid_changed_cb, zfsvfs); error = error ? error : dsl_prop_register(ds, zfs_prop_to_name(ZFS_PROP_EXEC), exec_changed_cb, zfsvfs); error = error ? error : dsl_prop_register(ds, zfs_prop_to_name(ZFS_PROP_SNAPDIR), snapdir_changed_cb, zfsvfs); error = error ? error : dsl_prop_register(ds, zfs_prop_to_name(ZFS_PROP_ACLTYPE), acl_type_changed_cb, zfsvfs); error = error ? error : dsl_prop_register(ds, zfs_prop_to_name(ZFS_PROP_ACLMODE), acl_mode_changed_cb, zfsvfs); error = error ? error : dsl_prop_register(ds, zfs_prop_to_name(ZFS_PROP_ACLINHERIT), acl_inherit_changed_cb, zfsvfs); dsl_pool_config_exit(dmu_objset_pool(os), FTAG); if (error) goto unregister; /* * Invoke our callbacks to restore temporary mount options. */ if (do_readonly) readonly_changed_cb(zfsvfs, readonly); if (do_setuid) setuid_changed_cb(zfsvfs, setuid); if (do_exec) exec_changed_cb(zfsvfs, exec); if (do_xattr) xattr_changed_cb(zfsvfs, xattr); if (do_atime) atime_changed_cb(zfsvfs, atime); nbmand_changed_cb(zfsvfs, nbmand); return (0); unregister: dsl_prop_unregister_all(ds, zfsvfs); return (error); } /* * Associate this zfsvfs with the given objset, which must be owned. * This will cache a bunch of on-disk state from the objset in the * zfsvfs. */ static int zfsvfs_init(zfsvfs_t *zfsvfs, objset_t *os) { int error; uint64_t val; zfsvfs->z_max_blksz = SPA_OLD_MAXBLOCKSIZE; zfsvfs->z_show_ctldir = ZFS_SNAPDIR_VISIBLE; zfsvfs->z_os = os; error = zfs_get_zplprop(os, ZFS_PROP_VERSION, &zfsvfs->z_version); if (error != 0) return (error); if (zfsvfs->z_version > zfs_zpl_version_map(spa_version(dmu_objset_spa(os)))) { (void) printf("Can't mount a version %lld file system " "on a version %lld pool\n. Pool must be upgraded to mount " "this file system.", (u_longlong_t)zfsvfs->z_version, (u_longlong_t)spa_version(dmu_objset_spa(os))); return (SET_ERROR(ENOTSUP)); } error = zfs_get_zplprop(os, ZFS_PROP_NORMALIZE, &val); if (error != 0) return (error); zfsvfs->z_norm = (int)val; error = zfs_get_zplprop(os, ZFS_PROP_UTF8ONLY, &val); if (error != 0) return (error); zfsvfs->z_utf8 = (val != 0); error = zfs_get_zplprop(os, ZFS_PROP_CASE, &val); if (error != 0) return (error); zfsvfs->z_case = (uint_t)val; error = zfs_get_zplprop(os, ZFS_PROP_ACLTYPE, &val); if (error != 0) return (error); zfsvfs->z_acl_type = (uint_t)val; /* * Fold case on file systems that are always or sometimes case * insensitive. */ if (zfsvfs->z_case == ZFS_CASE_INSENSITIVE || zfsvfs->z_case == ZFS_CASE_MIXED) zfsvfs->z_norm |= U8_TEXTPREP_TOUPPER; zfsvfs->z_use_fuids = USE_FUIDS(zfsvfs->z_version, zfsvfs->z_os); zfsvfs->z_use_sa = USE_SA(zfsvfs->z_version, zfsvfs->z_os); uint64_t sa_obj = 0; if (zfsvfs->z_use_sa) { /* should either have both of these objects or none */ error = zap_lookup(os, MASTER_NODE_OBJ, ZFS_SA_ATTRS, 8, 1, &sa_obj); if (error != 0) return (error); error = zfs_get_zplprop(os, ZFS_PROP_XATTR, &val); if (error == 0 && val == ZFS_XATTR_SA) zfsvfs->z_xattr_sa = B_TRUE; } error = sa_setup(os, sa_obj, zfs_attr_table, ZPL_END, &zfsvfs->z_attr_table); if (error != 0) return (error); if (zfsvfs->z_version >= ZPL_VERSION_SA) sa_register_update_callback(os, zfs_sa_upgrade); error = zap_lookup(os, MASTER_NODE_OBJ, ZFS_ROOT_OBJ, 8, 1, &zfsvfs->z_root); if (error != 0) return (error); ASSERT3U(zfsvfs->z_root, !=, 0); error = zap_lookup(os, MASTER_NODE_OBJ, ZFS_UNLINKED_SET, 8, 1, &zfsvfs->z_unlinkedobj); if (error != 0) return (error); error = zap_lookup(os, MASTER_NODE_OBJ, zfs_userquota_prop_prefixes[ZFS_PROP_USERQUOTA], 8, 1, &zfsvfs->z_userquota_obj); if (error == ENOENT) zfsvfs->z_userquota_obj = 0; else if (error != 0) return (error); error = zap_lookup(os, MASTER_NODE_OBJ, zfs_userquota_prop_prefixes[ZFS_PROP_GROUPQUOTA], 8, 1, &zfsvfs->z_groupquota_obj); if (error == ENOENT) zfsvfs->z_groupquota_obj = 0; else if (error != 0) return (error); error = zap_lookup(os, MASTER_NODE_OBJ, zfs_userquota_prop_prefixes[ZFS_PROP_PROJECTQUOTA], 8, 1, &zfsvfs->z_projectquota_obj); if (error == ENOENT) zfsvfs->z_projectquota_obj = 0; else if (error != 0) return (error); error = zap_lookup(os, MASTER_NODE_OBJ, zfs_userquota_prop_prefixes[ZFS_PROP_USEROBJQUOTA], 8, 1, &zfsvfs->z_userobjquota_obj); if (error == ENOENT) zfsvfs->z_userobjquota_obj = 0; else if (error != 0) return (error); error = zap_lookup(os, MASTER_NODE_OBJ, zfs_userquota_prop_prefixes[ZFS_PROP_GROUPOBJQUOTA], 8, 1, &zfsvfs->z_groupobjquota_obj); if (error == ENOENT) zfsvfs->z_groupobjquota_obj = 0; else if (error != 0) return (error); error = zap_lookup(os, MASTER_NODE_OBJ, zfs_userquota_prop_prefixes[ZFS_PROP_PROJECTOBJQUOTA], 8, 1, &zfsvfs->z_projectobjquota_obj); if (error == ENOENT) zfsvfs->z_projectobjquota_obj = 0; else if (error != 0) return (error); error = zap_lookup(os, MASTER_NODE_OBJ, ZFS_FUID_TABLES, 8, 1, &zfsvfs->z_fuid_obj); if (error == ENOENT) zfsvfs->z_fuid_obj = 0; else if (error != 0) return (error); error = zap_lookup(os, MASTER_NODE_OBJ, ZFS_SHARES_DIR, 8, 1, &zfsvfs->z_shares_dir); if (error == ENOENT) zfsvfs->z_shares_dir = 0; else if (error != 0) return (error); /* * Only use the name cache if we are looking for a * name on a file system that does not require normalization * or case folding. We can also look there if we happen to be * on a non-normalizing, mixed sensitivity file system IF we * are looking for the exact name (which is always the case on * FreeBSD). */ zfsvfs->z_use_namecache = !zfsvfs->z_norm || ((zfsvfs->z_case == ZFS_CASE_MIXED) && !(zfsvfs->z_norm & ~U8_TEXTPREP_TOUPPER)); return (0); } taskq_t *zfsvfs_taskq; static void zfsvfs_task_unlinked_drain(void *context, int pending __unused) { zfs_unlinked_drain((zfsvfs_t *)context); } int zfsvfs_create(const char *osname, boolean_t readonly, zfsvfs_t **zfvp) { objset_t *os; zfsvfs_t *zfsvfs; int error; boolean_t ro = (readonly || (strchr(osname, '@') != NULL)); /* * XXX: Fix struct statfs so this isn't necessary! * * The 'osname' is used as the filesystem's special node, which means * it must fit in statfs.f_mntfromname, or else it can't be * enumerated, so libzfs_mnttab_find() returns NULL, which causes * 'zfs unmount' to think it's not mounted when it is. */ if (strlen(osname) >= MNAMELEN) return (SET_ERROR(ENAMETOOLONG)); zfsvfs = kmem_zalloc(sizeof (zfsvfs_t), KM_SLEEP); error = dmu_objset_own(osname, DMU_OST_ZFS, ro, B_TRUE, zfsvfs, &os); if (error != 0) { kmem_free(zfsvfs, sizeof (zfsvfs_t)); return (error); } error = zfsvfs_create_impl(zfvp, zfsvfs, os); return (error); } int zfsvfs_create_impl(zfsvfs_t **zfvp, zfsvfs_t *zfsvfs, objset_t *os) { int error; zfsvfs->z_vfs = NULL; zfsvfs->z_parent = zfsvfs; mutex_init(&zfsvfs->z_znodes_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&zfsvfs->z_lock, NULL, MUTEX_DEFAULT, NULL); list_create(&zfsvfs->z_all_znodes, sizeof (znode_t), offsetof(znode_t, z_link_node)); TASK_INIT(&zfsvfs->z_unlinked_drain_task, 0, zfsvfs_task_unlinked_drain, zfsvfs); ZFS_TEARDOWN_INIT(zfsvfs); ZFS_TEARDOWN_INACTIVE_INIT(zfsvfs); rw_init(&zfsvfs->z_fuid_lock, NULL, RW_DEFAULT, NULL); for (int i = 0; i != ZFS_OBJ_MTX_SZ; i++) mutex_init(&zfsvfs->z_hold_mtx[i], NULL, MUTEX_DEFAULT, NULL); error = zfsvfs_init(zfsvfs, os); if (error != 0) { dmu_objset_disown(os, B_TRUE, zfsvfs); *zfvp = NULL; kmem_free(zfsvfs, sizeof (zfsvfs_t)); return (error); } *zfvp = zfsvfs; return (0); } static int zfsvfs_setup(zfsvfs_t *zfsvfs, boolean_t mounting) { int error; /* * Check for a bad on-disk format version now since we * lied about owning the dataset readonly before. */ if (!(zfsvfs->z_vfs->vfs_flag & VFS_RDONLY) && dmu_objset_incompatible_encryption_version(zfsvfs->z_os)) return (SET_ERROR(EROFS)); error = zfs_register_callbacks(zfsvfs->z_vfs); if (error) return (error); /* * If we are not mounting (ie: online recv), then we don't * have to worry about replaying the log as we blocked all * operations out since we closed the ZIL. */ if (mounting) { boolean_t readonly; ASSERT3P(zfsvfs->z_kstat.dk_kstats, ==, NULL); error = dataset_kstats_create(&zfsvfs->z_kstat, zfsvfs->z_os); if (error) return (error); zfsvfs->z_log = zil_open(zfsvfs->z_os, zfs_get_data, &zfsvfs->z_kstat.dk_zil_sums); /* * During replay we remove the read only flag to * allow replays to succeed. */ readonly = zfsvfs->z_vfs->vfs_flag & VFS_RDONLY; if (readonly != 0) { zfsvfs->z_vfs->vfs_flag &= ~VFS_RDONLY; } else { dsl_dir_t *dd; zap_stats_t zs; if (zap_get_stats(zfsvfs->z_os, zfsvfs->z_unlinkedobj, &zs) == 0) { dataset_kstats_update_nunlinks_kstat( &zfsvfs->z_kstat, zs.zs_num_entries); dprintf_ds(zfsvfs->z_os->os_dsl_dataset, "num_entries in unlinked set: %llu", (u_longlong_t)zs.zs_num_entries); } zfs_unlinked_drain(zfsvfs); dd = zfsvfs->z_os->os_dsl_dataset->ds_dir; dd->dd_activity_cancelled = B_FALSE; } /* * Parse and replay the intent log. * * Because of ziltest, this must be done after * zfs_unlinked_drain(). (Further note: ziltest * doesn't use readonly mounts, where * zfs_unlinked_drain() isn't called.) This is because * ziltest causes spa_sync() to think it's committed, * but actually it is not, so the intent log contains * many txg's worth of changes. * * In particular, if object N is in the unlinked set in * the last txg to actually sync, then it could be * actually freed in a later txg and then reallocated * in a yet later txg. This would write a "create * object N" record to the intent log. Normally, this * would be fine because the spa_sync() would have * written out the fact that object N is free, before * we could write the "create object N" intent log * record. * * But when we are in ziltest mode, we advance the "open * txg" without actually spa_sync()-ing the changes to * disk. So we would see that object N is still * allocated and in the unlinked set, and there is an * intent log record saying to allocate it. */ if (spa_writeable(dmu_objset_spa(zfsvfs->z_os))) { if (zil_replay_disable) { zil_destroy(zfsvfs->z_log, B_FALSE); } else { boolean_t use_nc = zfsvfs->z_use_namecache; zfsvfs->z_use_namecache = B_FALSE; zfsvfs->z_replay = B_TRUE; zil_replay(zfsvfs->z_os, zfsvfs, zfs_replay_vector); zfsvfs->z_replay = B_FALSE; zfsvfs->z_use_namecache = use_nc; } } /* restore readonly bit */ if (readonly != 0) zfsvfs->z_vfs->vfs_flag |= VFS_RDONLY; } else { ASSERT3P(zfsvfs->z_kstat.dk_kstats, !=, NULL); zfsvfs->z_log = zil_open(zfsvfs->z_os, zfs_get_data, &zfsvfs->z_kstat.dk_zil_sums); } /* * Set the objset user_ptr to track its zfsvfs. */ mutex_enter(&zfsvfs->z_os->os_user_ptr_lock); dmu_objset_set_user(zfsvfs->z_os, zfsvfs); mutex_exit(&zfsvfs->z_os->os_user_ptr_lock); return (0); } void zfsvfs_free(zfsvfs_t *zfsvfs) { int i; zfs_fuid_destroy(zfsvfs); mutex_destroy(&zfsvfs->z_znodes_lock); mutex_destroy(&zfsvfs->z_lock); list_destroy(&zfsvfs->z_all_znodes); ZFS_TEARDOWN_DESTROY(zfsvfs); ZFS_TEARDOWN_INACTIVE_DESTROY(zfsvfs); rw_destroy(&zfsvfs->z_fuid_lock); for (i = 0; i != ZFS_OBJ_MTX_SZ; i++) mutex_destroy(&zfsvfs->z_hold_mtx[i]); dataset_kstats_destroy(&zfsvfs->z_kstat); kmem_free(zfsvfs, sizeof (zfsvfs_t)); } static void zfs_set_fuid_feature(zfsvfs_t *zfsvfs) { zfsvfs->z_use_fuids = USE_FUIDS(zfsvfs->z_version, zfsvfs->z_os); zfsvfs->z_use_sa = USE_SA(zfsvfs->z_version, zfsvfs->z_os); } static int zfs_domount(vfs_t *vfsp, char *osname) { uint64_t recordsize, fsid_guid; int error = 0; zfsvfs_t *zfsvfs; ASSERT3P(vfsp, !=, NULL); ASSERT3P(osname, !=, NULL); error = zfsvfs_create(osname, vfsp->mnt_flag & MNT_RDONLY, &zfsvfs); if (error) return (error); zfsvfs->z_vfs = vfsp; if ((error = dsl_prop_get_integer(osname, "recordsize", &recordsize, NULL))) goto out; zfsvfs->z_vfs->vfs_bsize = SPA_MINBLOCKSIZE; zfsvfs->z_vfs->mnt_stat.f_iosize = recordsize; vfsp->vfs_data = zfsvfs; vfsp->mnt_flag |= MNT_LOCAL; vfsp->mnt_kern_flag |= MNTK_LOOKUP_SHARED; vfsp->mnt_kern_flag |= MNTK_SHARED_WRITES; vfsp->mnt_kern_flag |= MNTK_EXTENDED_SHARED; /* * This can cause a loss of coherence between ARC and page cache * on ZoF - unclear if the problem is in FreeBSD or ZoF */ vfsp->mnt_kern_flag |= MNTK_NO_IOPF; /* vn_io_fault can be used */ vfsp->mnt_kern_flag |= MNTK_NOMSYNC; vfsp->mnt_kern_flag |= MNTK_VMSETSIZE_BUG; #if defined(_KERNEL) && !defined(KMEM_DEBUG) vfsp->mnt_kern_flag |= MNTK_FPLOOKUP; #endif /* * The fsid is 64 bits, composed of an 8-bit fs type, which * separates our fsid from any other filesystem types, and a * 56-bit objset unique ID. The objset unique ID is unique to * all objsets open on this system, provided by unique_create(). * The 8-bit fs type must be put in the low bits of fsid[1] * because that's where other Solaris filesystems put it. */ fsid_guid = dmu_objset_fsid_guid(zfsvfs->z_os); ASSERT3U((fsid_guid & ~((1ULL << 56) - 1)), ==, 0); vfsp->vfs_fsid.val[0] = fsid_guid; vfsp->vfs_fsid.val[1] = ((fsid_guid >> 32) << 8) | (vfsp->mnt_vfc->vfc_typenum & 0xFF); /* * Set features for file system. */ zfs_set_fuid_feature(zfsvfs); if (dmu_objset_is_snapshot(zfsvfs->z_os)) { uint64_t pval; atime_changed_cb(zfsvfs, B_FALSE); readonly_changed_cb(zfsvfs, B_TRUE); if ((error = dsl_prop_get_integer(osname, "xattr", &pval, NULL))) goto out; xattr_changed_cb(zfsvfs, pval); if ((error = dsl_prop_get_integer(osname, "acltype", &pval, NULL))) goto out; acl_type_changed_cb(zfsvfs, pval); zfsvfs->z_issnap = B_TRUE; zfsvfs->z_os->os_sync = ZFS_SYNC_DISABLED; mutex_enter(&zfsvfs->z_os->os_user_ptr_lock); dmu_objset_set_user(zfsvfs->z_os, zfsvfs); mutex_exit(&zfsvfs->z_os->os_user_ptr_lock); } else { if ((error = zfsvfs_setup(zfsvfs, B_TRUE))) goto out; } vfs_mountedfrom(vfsp, osname); if (!zfsvfs->z_issnap) zfsctl_create(zfsvfs); out: if (error) { dmu_objset_disown(zfsvfs->z_os, B_TRUE, zfsvfs); zfsvfs_free(zfsvfs); } else { atomic_inc_32(&zfs_active_fs_count); } return (error); } static void zfs_unregister_callbacks(zfsvfs_t *zfsvfs) { objset_t *os = zfsvfs->z_os; if (!dmu_objset_is_snapshot(os)) dsl_prop_unregister_all(dmu_objset_ds(os), zfsvfs); } static int getpoolname(const char *osname, char *poolname) { char *p; p = strchr(osname, '/'); if (p == NULL) { if (strlen(osname) >= MAXNAMELEN) return (ENAMETOOLONG); (void) strcpy(poolname, osname); } else { if (p - osname >= MAXNAMELEN) return (ENAMETOOLONG); (void) strlcpy(poolname, osname, p - osname + 1); } return (0); } static void fetch_osname_options(char *name, bool *checkpointrewind) { if (name[0] == '!') { *checkpointrewind = true; memmove(name, name + 1, strlen(name)); } else { *checkpointrewind = false; } } static int zfs_mount(vfs_t *vfsp) { kthread_t *td = curthread; vnode_t *mvp = vfsp->mnt_vnodecovered; cred_t *cr = td->td_ucred; char *osname; int error = 0; int canwrite; bool checkpointrewind, isctlsnap = false; if (vfs_getopt(vfsp->mnt_optnew, "from", (void **)&osname, NULL)) return (SET_ERROR(EINVAL)); /* * If full-owner-access is enabled and delegated administration is * turned on, we must set nosuid. */ if (zfs_super_owner && dsl_deleg_access(osname, ZFS_DELEG_PERM_MOUNT, cr) != ECANCELED) { secpolicy_fs_mount_clearopts(cr, vfsp); } fetch_osname_options(osname, &checkpointrewind); isctlsnap = (mvp != NULL && zfsctl_is_node(mvp) && strchr(osname, '@') != NULL); /* * Check for mount privilege? * * If we don't have privilege then see if * we have local permission to allow it */ error = secpolicy_fs_mount(cr, mvp, vfsp); if (error && isctlsnap) { secpolicy_fs_mount_clearopts(cr, vfsp); } else if (error) { if (dsl_deleg_access(osname, ZFS_DELEG_PERM_MOUNT, cr) != 0) goto out; if (!(vfsp->vfs_flag & MS_REMOUNT)) { vattr_t vattr; /* * Make sure user is the owner of the mount point * or has sufficient privileges. */ vattr.va_mask = AT_UID; vn_lock(mvp, LK_SHARED | LK_RETRY); if (VOP_GETATTR(mvp, &vattr, cr)) { VOP_UNLOCK1(mvp); goto out; } if (secpolicy_vnode_owner(mvp, cr, vattr.va_uid) != 0 && VOP_ACCESS(mvp, VWRITE, cr, td) != 0) { VOP_UNLOCK1(mvp); goto out; } VOP_UNLOCK1(mvp); } secpolicy_fs_mount_clearopts(cr, vfsp); } /* * Refuse to mount a filesystem if we are in a local zone and the * dataset is not visible. */ if (!INGLOBALZONE(curproc) && (!zone_dataset_visible(osname, &canwrite) || !canwrite)) { boolean_t mount_snapshot = B_FALSE; /* * Snapshots may be mounted in .zfs for unjailed datasets * if allowed by the jail param zfs.mount_snapshot. */ if (isctlsnap) { struct prison *pr; struct zfs_jailparam *zjp; pr = curthread->td_ucred->cr_prison; mtx_lock(&pr->pr_mtx); zjp = osd_jail_get(pr, zfs_jailparam_slot); mtx_unlock(&pr->pr_mtx); if (zjp && zjp->mount_snapshot) mount_snapshot = B_TRUE; } if (!mount_snapshot) { error = SET_ERROR(EPERM); goto out; } } vfsp->vfs_flag |= MNT_NFS4ACLS; /* * When doing a remount, we simply refresh our temporary properties * according to those options set in the current VFS options. */ if (vfsp->vfs_flag & MS_REMOUNT) { zfsvfs_t *zfsvfs = vfsp->vfs_data; /* * Refresh mount options with z_teardown_lock blocking I/O while * the filesystem is in an inconsistent state. * The lock also serializes this code with filesystem * manipulations between entry to zfs_suspend_fs() and return * from zfs_resume_fs(). */ ZFS_TEARDOWN_ENTER_WRITE(zfsvfs, FTAG); zfs_unregister_callbacks(zfsvfs); error = zfs_register_callbacks(vfsp); ZFS_TEARDOWN_EXIT(zfsvfs, FTAG); goto out; } /* Initial root mount: try hard to import the requested root pool. */ if ((vfsp->vfs_flag & MNT_ROOTFS) != 0 && (vfsp->vfs_flag & MNT_UPDATE) == 0) { char pname[MAXNAMELEN]; error = getpoolname(osname, pname); if (error == 0) error = spa_import_rootpool(pname, checkpointrewind); if (error) goto out; } DROP_GIANT(); error = zfs_domount(vfsp, osname); PICKUP_GIANT(); out: return (error); } static int zfs_statfs(vfs_t *vfsp, struct statfs *statp) { zfsvfs_t *zfsvfs = vfsp->vfs_data; uint64_t refdbytes, availbytes, usedobjs, availobjs; int error; statp->f_version = STATFS_VERSION; if ((error = zfs_enter(zfsvfs, FTAG)) != 0) return (error); dmu_objset_space(zfsvfs->z_os, &refdbytes, &availbytes, &usedobjs, &availobjs); /* * The underlying storage pool actually uses multiple block sizes. * We report the fragsize as the smallest block size we support, * and we report our blocksize as the filesystem's maximum blocksize. */ statp->f_bsize = SPA_MINBLOCKSIZE; statp->f_iosize = zfsvfs->z_vfs->mnt_stat.f_iosize; /* * The following report "total" blocks of various kinds in the * file system, but reported in terms of f_frsize - the * "fragment" size. */ statp->f_blocks = (refdbytes + availbytes) >> SPA_MINBLOCKSHIFT; statp->f_bfree = availbytes / statp->f_bsize; statp->f_bavail = statp->f_bfree; /* no root reservation */ /* * statvfs() should really be called statufs(), because it assumes * static metadata. ZFS doesn't preallocate files, so the best * we can do is report the max that could possibly fit in f_files, * and that minus the number actually used in f_ffree. * For f_ffree, report the smaller of the number of object available * and the number of blocks (each object will take at least a block). */ statp->f_ffree = MIN(availobjs, statp->f_bfree); statp->f_files = statp->f_ffree + usedobjs; /* * We're a zfs filesystem. */ strlcpy(statp->f_fstypename, "zfs", sizeof (statp->f_fstypename)); strlcpy(statp->f_mntfromname, vfsp->mnt_stat.f_mntfromname, sizeof (statp->f_mntfromname)); strlcpy(statp->f_mntonname, vfsp->mnt_stat.f_mntonname, sizeof (statp->f_mntonname)); statp->f_namemax = MAXNAMELEN - 1; zfs_exit(zfsvfs, FTAG); return (0); } static int zfs_root(vfs_t *vfsp, int flags, vnode_t **vpp) { zfsvfs_t *zfsvfs = vfsp->vfs_data; znode_t *rootzp; int error; if ((error = zfs_enter(zfsvfs, FTAG)) != 0) return (error); error = zfs_zget(zfsvfs, zfsvfs->z_root, &rootzp); if (error == 0) *vpp = ZTOV(rootzp); zfs_exit(zfsvfs, FTAG); if (error == 0) { error = vn_lock(*vpp, flags); if (error != 0) { VN_RELE(*vpp); *vpp = NULL; } } return (error); } /* * Teardown the zfsvfs::z_os. * * Note, if 'unmounting' is FALSE, we return with the 'z_teardown_lock' * and 'z_teardown_inactive_lock' held. */ static int zfsvfs_teardown(zfsvfs_t *zfsvfs, boolean_t unmounting) { znode_t *zp; dsl_dir_t *dd; /* * If someone has not already unmounted this file system, * drain the zrele_taskq to ensure all active references to the * zfsvfs_t have been handled only then can it be safely destroyed. */ if (zfsvfs->z_os) { /* * If we're unmounting we have to wait for the list to * drain completely. * * If we're not unmounting there's no guarantee the list * will drain completely, but zreles run from the taskq * may add the parents of dir-based xattrs to the taskq * so we want to wait for these. * * We can safely check z_all_znodes for being empty because the * VFS has already blocked operations which add to it. */ int round = 0; while (!list_is_empty(&zfsvfs->z_all_znodes)) { taskq_wait_outstanding(dsl_pool_zrele_taskq( dmu_objset_pool(zfsvfs->z_os)), 0); if (++round > 1 && !unmounting) break; } } ZFS_TEARDOWN_ENTER_WRITE(zfsvfs, FTAG); if (!unmounting) { /* * We purge the parent filesystem's vfsp as the parent * filesystem and all of its snapshots have their vnode's * v_vfsp set to the parent's filesystem's vfsp. Note, * 'z_parent' is self referential for non-snapshots. */ #ifdef FREEBSD_NAMECACHE #if __FreeBSD_version >= 1300117 cache_purgevfs(zfsvfs->z_parent->z_vfs); #else cache_purgevfs(zfsvfs->z_parent->z_vfs, true); #endif #endif } /* * Close the zil. NB: Can't close the zil while zfs_inactive * threads are blocked as zil_close can call zfs_inactive. */ if (zfsvfs->z_log) { zil_close(zfsvfs->z_log); zfsvfs->z_log = NULL; } ZFS_TEARDOWN_INACTIVE_ENTER_WRITE(zfsvfs); /* * If we are not unmounting (ie: online recv) and someone already * unmounted this file system while we were doing the switcheroo, * or a reopen of z_os failed then just bail out now. */ if (!unmounting && (zfsvfs->z_unmounted || zfsvfs->z_os == NULL)) { ZFS_TEARDOWN_INACTIVE_EXIT_WRITE(zfsvfs); ZFS_TEARDOWN_EXIT(zfsvfs, FTAG); return (SET_ERROR(EIO)); } /* * At this point there are no vops active, and any new vops will * fail with EIO since we have z_teardown_lock for writer (only * relevant for forced unmount). * * Release all holds on dbufs. */ mutex_enter(&zfsvfs->z_znodes_lock); for (zp = list_head(&zfsvfs->z_all_znodes); zp != NULL; zp = list_next(&zfsvfs->z_all_znodes, zp)) { if (zp->z_sa_hdl != NULL) { zfs_znode_dmu_fini(zp); } } mutex_exit(&zfsvfs->z_znodes_lock); /* * If we are unmounting, set the unmounted flag and let new vops * unblock. zfs_inactive will have the unmounted behavior, and all * other vops will fail with EIO. */ if (unmounting) { zfsvfs->z_unmounted = B_TRUE; ZFS_TEARDOWN_INACTIVE_EXIT_WRITE(zfsvfs); ZFS_TEARDOWN_EXIT(zfsvfs, FTAG); } /* * z_os will be NULL if there was an error in attempting to reopen * zfsvfs, so just return as the properties had already been * unregistered and cached data had been evicted before. */ if (zfsvfs->z_os == NULL) return (0); /* * Unregister properties. */ zfs_unregister_callbacks(zfsvfs); /* * Evict cached data */ if (!zfs_is_readonly(zfsvfs)) txg_wait_synced(dmu_objset_pool(zfsvfs->z_os), 0); dmu_objset_evict_dbufs(zfsvfs->z_os); dd = zfsvfs->z_os->os_dsl_dataset->ds_dir; dsl_dir_cancel_waiters(dd); return (0); } static int zfs_umount(vfs_t *vfsp, int fflag) { kthread_t *td = curthread; zfsvfs_t *zfsvfs = vfsp->vfs_data; objset_t *os; cred_t *cr = td->td_ucred; int ret; ret = secpolicy_fs_unmount(cr, vfsp); if (ret) { if (dsl_deleg_access((char *)vfsp->vfs_resource, ZFS_DELEG_PERM_MOUNT, cr)) return (ret); } /* * Unmount any snapshots mounted under .zfs before unmounting the * dataset itself. */ if (zfsvfs->z_ctldir != NULL) { if ((ret = zfsctl_umount_snapshots(vfsp, fflag, cr)) != 0) return (ret); } if (fflag & MS_FORCE) { /* * Mark file system as unmounted before calling * vflush(FORCECLOSE). This way we ensure no future vnops * will be called and risk operating on DOOMED vnodes. */ ZFS_TEARDOWN_ENTER_WRITE(zfsvfs, FTAG); zfsvfs->z_unmounted = B_TRUE; ZFS_TEARDOWN_EXIT(zfsvfs, FTAG); } /* * Flush all the files. */ ret = vflush(vfsp, 0, (fflag & MS_FORCE) ? FORCECLOSE : 0, td); if (ret != 0) return (ret); while (taskqueue_cancel(zfsvfs_taskq->tq_queue, &zfsvfs->z_unlinked_drain_task, NULL) != 0) taskqueue_drain(zfsvfs_taskq->tq_queue, &zfsvfs->z_unlinked_drain_task); VERIFY0(zfsvfs_teardown(zfsvfs, B_TRUE)); os = zfsvfs->z_os; /* * z_os will be NULL if there was an error in * attempting to reopen zfsvfs. */ if (os != NULL) { /* * Unset the objset user_ptr. */ mutex_enter(&os->os_user_ptr_lock); dmu_objset_set_user(os, NULL); mutex_exit(&os->os_user_ptr_lock); /* * Finally release the objset */ dmu_objset_disown(os, B_TRUE, zfsvfs); } /* * We can now safely destroy the '.zfs' directory node. */ if (zfsvfs->z_ctldir != NULL) zfsctl_destroy(zfsvfs); zfs_freevfs(vfsp); return (0); } static int zfs_vget(vfs_t *vfsp, ino_t ino, int flags, vnode_t **vpp) { zfsvfs_t *zfsvfs = vfsp->vfs_data; znode_t *zp; int err; /* * zfs_zget() can't operate on virtual entries like .zfs/ or * .zfs/snapshot/ directories, that's why we return EOPNOTSUPP. * This will make NFS to switch to LOOKUP instead of using VGET. */ if (ino == ZFSCTL_INO_ROOT || ino == ZFSCTL_INO_SNAPDIR || (zfsvfs->z_shares_dir != 0 && ino == zfsvfs->z_shares_dir)) return (EOPNOTSUPP); if ((err = zfs_enter(zfsvfs, FTAG)) != 0) return (err); err = zfs_zget(zfsvfs, ino, &zp); if (err == 0 && zp->z_unlinked) { vrele(ZTOV(zp)); err = EINVAL; } if (err == 0) *vpp = ZTOV(zp); zfs_exit(zfsvfs, FTAG); if (err == 0) { err = vn_lock(*vpp, flags); if (err != 0) vrele(*vpp); } if (err != 0) *vpp = NULL; return (err); } static int #if __FreeBSD_version >= 1300098 zfs_checkexp(vfs_t *vfsp, struct sockaddr *nam, uint64_t *extflagsp, struct ucred **credanonp, int *numsecflavors, int *secflavors) #else zfs_checkexp(vfs_t *vfsp, struct sockaddr *nam, int *extflagsp, struct ucred **credanonp, int *numsecflavors, int **secflavors) #endif { zfsvfs_t *zfsvfs = vfsp->vfs_data; /* * If this is regular file system vfsp is the same as * zfsvfs->z_parent->z_vfs, but if it is snapshot, * zfsvfs->z_parent->z_vfs represents parent file system * which we have to use here, because only this file system * has mnt_export configured. */ return (vfs_stdcheckexp(zfsvfs->z_parent->z_vfs, nam, extflagsp, credanonp, numsecflavors, secflavors)); } _Static_assert(sizeof (struct fid) >= SHORT_FID_LEN, "struct fid bigger than SHORT_FID_LEN"); _Static_assert(sizeof (struct fid) >= LONG_FID_LEN, "struct fid bigger than LONG_FID_LEN"); static int zfs_fhtovp(vfs_t *vfsp, fid_t *fidp, int flags, vnode_t **vpp) { struct componentname cn; zfsvfs_t *zfsvfs = vfsp->vfs_data; znode_t *zp; vnode_t *dvp; uint64_t object = 0; uint64_t fid_gen = 0; uint64_t setgen = 0; uint64_t gen_mask; uint64_t zp_gen; int i, err; *vpp = NULL; if ((err = zfs_enter(zfsvfs, FTAG)) != 0) return (err); /* * On FreeBSD we can get snapshot's mount point or its parent file * system mount point depending if snapshot is already mounted or not. */ if (zfsvfs->z_parent == zfsvfs && fidp->fid_len == LONG_FID_LEN) { zfid_long_t *zlfid = (zfid_long_t *)fidp; uint64_t objsetid = 0; for (i = 0; i < sizeof (zlfid->zf_setid); i++) objsetid |= ((uint64_t)zlfid->zf_setid[i]) << (8 * i); for (i = 0; i < sizeof (zlfid->zf_setgen); i++) setgen |= ((uint64_t)zlfid->zf_setgen[i]) << (8 * i); zfs_exit(zfsvfs, FTAG); err = zfsctl_lookup_objset(vfsp, objsetid, &zfsvfs); if (err) return (SET_ERROR(EINVAL)); if ((err = zfs_enter(zfsvfs, FTAG)) != 0) return (err); } if (fidp->fid_len == SHORT_FID_LEN || fidp->fid_len == LONG_FID_LEN) { zfid_short_t *zfid = (zfid_short_t *)fidp; for (i = 0; i < sizeof (zfid->zf_object); i++) object |= ((uint64_t)zfid->zf_object[i]) << (8 * i); for (i = 0; i < sizeof (zfid->zf_gen); i++) fid_gen |= ((uint64_t)zfid->zf_gen[i]) << (8 * i); } else { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EINVAL)); } if (fidp->fid_len == LONG_FID_LEN && setgen != 0) { zfs_exit(zfsvfs, FTAG); dprintf("snapdir fid: fid_gen (%llu) and setgen (%llu)\n", (u_longlong_t)fid_gen, (u_longlong_t)setgen); return (SET_ERROR(EINVAL)); } /* * A zero fid_gen means we are in .zfs or the .zfs/snapshot * directory tree. If the object == zfsvfs->z_shares_dir, then * we are in the .zfs/shares directory tree. */ if ((fid_gen == 0 && (object == ZFSCTL_INO_ROOT || object == ZFSCTL_INO_SNAPDIR)) || (zfsvfs->z_shares_dir != 0 && object == zfsvfs->z_shares_dir)) { zfs_exit(zfsvfs, FTAG); VERIFY0(zfsctl_root(zfsvfs, LK_SHARED, &dvp)); if (object == ZFSCTL_INO_SNAPDIR) { cn.cn_nameptr = "snapshot"; cn.cn_namelen = strlen(cn.cn_nameptr); cn.cn_nameiop = LOOKUP; cn.cn_flags = ISLASTCN | LOCKLEAF; cn.cn_lkflags = flags; VERIFY0(VOP_LOOKUP(dvp, vpp, &cn)); vput(dvp); } else if (object == zfsvfs->z_shares_dir) { /* * XXX This branch must not be taken, * if it is, then the lookup below will * explode. */ cn.cn_nameptr = "shares"; cn.cn_namelen = strlen(cn.cn_nameptr); cn.cn_nameiop = LOOKUP; cn.cn_flags = ISLASTCN; cn.cn_lkflags = flags; VERIFY0(VOP_LOOKUP(dvp, vpp, &cn)); vput(dvp); } else { *vpp = dvp; } return (err); } gen_mask = -1ULL >> (64 - 8 * i); dprintf("getting %llu [%llu mask %llx]\n", (u_longlong_t)object, (u_longlong_t)fid_gen, (u_longlong_t)gen_mask); if ((err = zfs_zget(zfsvfs, object, &zp))) { zfs_exit(zfsvfs, FTAG); return (err); } (void) sa_lookup(zp->z_sa_hdl, SA_ZPL_GEN(zfsvfs), &zp_gen, sizeof (uint64_t)); zp_gen = zp_gen & gen_mask; if (zp_gen == 0) zp_gen = 1; if (zp->z_unlinked || zp_gen != fid_gen) { dprintf("znode gen (%llu) != fid gen (%llu)\n", (u_longlong_t)zp_gen, (u_longlong_t)fid_gen); vrele(ZTOV(zp)); zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EINVAL)); } *vpp = ZTOV(zp); zfs_exit(zfsvfs, FTAG); err = vn_lock(*vpp, flags); if (err == 0) vnode_create_vobject(*vpp, zp->z_size, curthread); else *vpp = NULL; return (err); } /* * Block out VOPs and close zfsvfs_t::z_os * * Note, if successful, then we return with the 'z_teardown_lock' and * 'z_teardown_inactive_lock' write held. We leave ownership of the underlying * dataset and objset intact so that they can be atomically handed off during * a subsequent rollback or recv operation and the resume thereafter. */ int zfs_suspend_fs(zfsvfs_t *zfsvfs) { int error; if ((error = zfsvfs_teardown(zfsvfs, B_FALSE)) != 0) return (error); return (0); } /* * Rebuild SA and release VOPs. Note that ownership of the underlying dataset * is an invariant across any of the operations that can be performed while the * filesystem was suspended. Whether it succeeded or failed, the preconditions * are the same: the relevant objset and associated dataset are owned by * zfsvfs, held, and long held on entry. */ int zfs_resume_fs(zfsvfs_t *zfsvfs, dsl_dataset_t *ds) { int err; znode_t *zp; ASSERT(ZFS_TEARDOWN_WRITE_HELD(zfsvfs)); ASSERT(ZFS_TEARDOWN_INACTIVE_WRITE_HELD(zfsvfs)); /* * We already own this, so just update the objset_t, as the one we * had before may have been evicted. */ objset_t *os; VERIFY3P(ds->ds_owner, ==, zfsvfs); VERIFY(dsl_dataset_long_held(ds)); dsl_pool_t *dp = spa_get_dsl(dsl_dataset_get_spa(ds)); dsl_pool_config_enter(dp, FTAG); VERIFY0(dmu_objset_from_ds(ds, &os)); dsl_pool_config_exit(dp, FTAG); err = zfsvfs_init(zfsvfs, os); if (err != 0) goto bail; ds->ds_dir->dd_activity_cancelled = B_FALSE; VERIFY0(zfsvfs_setup(zfsvfs, B_FALSE)); zfs_set_fuid_feature(zfsvfs); /* * Attempt to re-establish all the active znodes with * their dbufs. If a zfs_rezget() fails, then we'll let * any potential callers discover that via zfs_enter_verify_zp * when they try to use their znode. */ mutex_enter(&zfsvfs->z_znodes_lock); for (zp = list_head(&zfsvfs->z_all_znodes); zp; zp = list_next(&zfsvfs->z_all_znodes, zp)) { (void) zfs_rezget(zp); } mutex_exit(&zfsvfs->z_znodes_lock); bail: /* release the VOPs */ ZFS_TEARDOWN_INACTIVE_EXIT_WRITE(zfsvfs); ZFS_TEARDOWN_EXIT(zfsvfs, FTAG); if (err) { /* * Since we couldn't setup the sa framework, try to force * unmount this file system. */ if (vn_vfswlock(zfsvfs->z_vfs->vfs_vnodecovered) == 0) { vfs_ref(zfsvfs->z_vfs); (void) dounmount(zfsvfs->z_vfs, MS_FORCE, curthread); } } return (err); } static void zfs_freevfs(vfs_t *vfsp) { zfsvfs_t *zfsvfs = vfsp->vfs_data; zfsvfs_free(zfsvfs); atomic_dec_32(&zfs_active_fs_count); } #ifdef __i386__ static int desiredvnodes_backup; #include #include #include #include #include #endif static void zfs_vnodes_adjust(void) { #ifdef __i386__ int newdesiredvnodes; desiredvnodes_backup = desiredvnodes; /* * We calculate newdesiredvnodes the same way it is done in * vntblinit(). If it is equal to desiredvnodes, it means that * it wasn't tuned by the administrator and we can tune it down. */ newdesiredvnodes = min(maxproc + vm_cnt.v_page_count / 4, 2 * vm_kmem_size / (5 * (sizeof (struct vm_object) + sizeof (struct vnode)))); if (newdesiredvnodes == desiredvnodes) desiredvnodes = (3 * newdesiredvnodes) / 4; #endif } static void zfs_vnodes_adjust_back(void) { #ifdef __i386__ desiredvnodes = desiredvnodes_backup; #endif } #if __FreeBSD_version >= 1300139 static struct sx zfs_vnlru_lock; static struct vnode *zfs_vnlru_marker; #endif static arc_prune_t *zfs_prune; static void zfs_prune_task(uint64_t nr_to_scan, void *arg __unused) { if (nr_to_scan > INT_MAX) nr_to_scan = INT_MAX; #if __FreeBSD_version >= 1300139 sx_xlock(&zfs_vnlru_lock); vnlru_free_vfsops(nr_to_scan, &zfs_vfsops, zfs_vnlru_marker); sx_xunlock(&zfs_vnlru_lock); #else vnlru_free(nr_to_scan, &zfs_vfsops); #endif } void zfs_init(void) { printf("ZFS filesystem version: " ZPL_VERSION_STRING "\n"); /* * Initialize .zfs directory structures */ zfsctl_init(); /* * Initialize znode cache, vnode ops, etc... */ zfs_znode_init(); /* * Reduce number of vnodes. Originally number of vnodes is calculated * with UFS inode in mind. We reduce it here, because it's too big for * ZFS/i386. */ zfs_vnodes_adjust(); dmu_objset_register_type(DMU_OST_ZFS, zpl_get_file_info); zfsvfs_taskq = taskq_create("zfsvfs", 1, minclsyspri, 0, 0, 0); #if __FreeBSD_version >= 1300139 zfs_vnlru_marker = vnlru_alloc_marker(); sx_init(&zfs_vnlru_lock, "zfs vnlru lock"); #endif zfs_prune = arc_add_prune_callback(zfs_prune_task, NULL); } void zfs_fini(void) { arc_remove_prune_callback(zfs_prune); #if __FreeBSD_version >= 1300139 vnlru_free_marker(zfs_vnlru_marker); sx_destroy(&zfs_vnlru_lock); #endif taskq_destroy(zfsvfs_taskq); zfsctl_fini(); zfs_znode_fini(); zfs_vnodes_adjust_back(); } int zfs_busy(void) { return (zfs_active_fs_count != 0); } /* * Release VOPs and unmount a suspended filesystem. */ int zfs_end_fs(zfsvfs_t *zfsvfs, dsl_dataset_t *ds) { ASSERT(ZFS_TEARDOWN_WRITE_HELD(zfsvfs)); ASSERT(ZFS_TEARDOWN_INACTIVE_WRITE_HELD(zfsvfs)); /* * We already own this, so just hold and rele it to update the * objset_t, as the one we had before may have been evicted. */ objset_t *os; VERIFY3P(ds->ds_owner, ==, zfsvfs); VERIFY(dsl_dataset_long_held(ds)); dsl_pool_t *dp = spa_get_dsl(dsl_dataset_get_spa(ds)); dsl_pool_config_enter(dp, FTAG); VERIFY0(dmu_objset_from_ds(ds, &os)); dsl_pool_config_exit(dp, FTAG); zfsvfs->z_os = os; /* release the VOPs */ ZFS_TEARDOWN_INACTIVE_EXIT_WRITE(zfsvfs); ZFS_TEARDOWN_EXIT(zfsvfs, FTAG); /* * Try to force unmount this file system. */ (void) zfs_umount(zfsvfs->z_vfs, 0); zfsvfs->z_unmounted = B_TRUE; return (0); } int zfs_set_version(zfsvfs_t *zfsvfs, uint64_t newvers) { int error; objset_t *os = zfsvfs->z_os; dmu_tx_t *tx; if (newvers < ZPL_VERSION_INITIAL || newvers > ZPL_VERSION) return (SET_ERROR(EINVAL)); if (newvers < zfsvfs->z_version) return (SET_ERROR(EINVAL)); if (zfs_spa_version_map(newvers) > spa_version(dmu_objset_spa(zfsvfs->z_os))) return (SET_ERROR(ENOTSUP)); tx = dmu_tx_create(os); dmu_tx_hold_zap(tx, MASTER_NODE_OBJ, B_FALSE, ZPL_VERSION_STR); if (newvers >= ZPL_VERSION_SA && !zfsvfs->z_use_sa) { dmu_tx_hold_zap(tx, MASTER_NODE_OBJ, B_TRUE, ZFS_SA_ATTRS); dmu_tx_hold_zap(tx, DMU_NEW_OBJECT, FALSE, NULL); } error = dmu_tx_assign(tx, TXG_WAIT); if (error) { dmu_tx_abort(tx); return (error); } error = zap_update(os, MASTER_NODE_OBJ, ZPL_VERSION_STR, 8, 1, &newvers, tx); if (error) { dmu_tx_commit(tx); return (error); } if (newvers >= ZPL_VERSION_SA && !zfsvfs->z_use_sa) { uint64_t sa_obj; ASSERT3U(spa_version(dmu_objset_spa(zfsvfs->z_os)), >=, SPA_VERSION_SA); sa_obj = zap_create(os, DMU_OT_SA_MASTER_NODE, DMU_OT_NONE, 0, tx); error = zap_add(os, MASTER_NODE_OBJ, ZFS_SA_ATTRS, 8, 1, &sa_obj, tx); ASSERT0(error); VERIFY0(sa_set_sa_object(os, sa_obj)); sa_register_update_callback(os, zfs_sa_upgrade); } spa_history_log_internal_ds(dmu_objset_ds(os), "upgrade", tx, "from %ju to %ju", (uintmax_t)zfsvfs->z_version, (uintmax_t)newvers); dmu_tx_commit(tx); zfsvfs->z_version = newvers; os->os_version = newvers; zfs_set_fuid_feature(zfsvfs); return (0); } /* * Return true if the corresponding vfs's unmounted flag is set. * Otherwise return false. * If this function returns true we know VFS unmount has been initiated. */ boolean_t zfs_get_vfs_flag_unmounted(objset_t *os) { zfsvfs_t *zfvp; boolean_t unmounted = B_FALSE; ASSERT3U(dmu_objset_type(os), ==, DMU_OST_ZFS); mutex_enter(&os->os_user_ptr_lock); zfvp = dmu_objset_get_user(os); if (zfvp != NULL && zfvp->z_vfs != NULL && (zfvp->z_vfs->mnt_kern_flag & MNTK_UNMOUNT)) unmounted = B_TRUE; mutex_exit(&os->os_user_ptr_lock); return (unmounted); } #ifdef _KERNEL void zfsvfs_update_fromname(const char *oldname, const char *newname) { char tmpbuf[MAXPATHLEN]; struct mount *mp; char *fromname; size_t oldlen; oldlen = strlen(oldname); mtx_lock(&mountlist_mtx); TAILQ_FOREACH(mp, &mountlist, mnt_list) { fromname = mp->mnt_stat.f_mntfromname; if (strcmp(fromname, oldname) == 0) { (void) strlcpy(fromname, newname, sizeof (mp->mnt_stat.f_mntfromname)); continue; } if (strncmp(fromname, oldname, oldlen) == 0 && (fromname[oldlen] == '/' || fromname[oldlen] == '@')) { (void) snprintf(tmpbuf, sizeof (tmpbuf), "%s%s", newname, fromname + oldlen); (void) strlcpy(fromname, tmpbuf, sizeof (mp->mnt_stat.f_mntfromname)); continue; } } mtx_unlock(&mountlist_mtx); } #endif /* * Find a prison with ZFS info. * Return the ZFS info and the (locked) prison. */ static struct zfs_jailparam * zfs_jailparam_find(struct prison *spr, struct prison **prp) { struct prison *pr; struct zfs_jailparam *zjp; for (pr = spr; ; pr = pr->pr_parent) { mtx_lock(&pr->pr_mtx); if (pr == &prison0) { zjp = &zfs_jailparam0; break; } zjp = osd_jail_get(pr, zfs_jailparam_slot); if (zjp != NULL) break; mtx_unlock(&pr->pr_mtx); } *prp = pr; return (zjp); } /* * Ensure a prison has its own ZFS info. If zjpp is non-null, point it to the * ZFS info and lock the prison. */ static void zfs_jailparam_alloc(struct prison *pr, struct zfs_jailparam **zjpp) { struct prison *ppr; struct zfs_jailparam *zjp, *nzjp; void **rsv; /* If this prison already has ZFS info, return that. */ zjp = zfs_jailparam_find(pr, &ppr); if (ppr == pr) goto done; /* * Allocate a new info record. Then check again, in case something * changed during the allocation. */ mtx_unlock(&ppr->pr_mtx); nzjp = malloc(sizeof (struct zfs_jailparam), M_PRISON, M_WAITOK); rsv = osd_reserve(zfs_jailparam_slot); zjp = zfs_jailparam_find(pr, &ppr); if (ppr == pr) { free(nzjp, M_PRISON); osd_free_reserved(rsv); goto done; } /* Inherit the initial values from the ancestor. */ mtx_lock(&pr->pr_mtx); (void) osd_jail_set_reserved(pr, zfs_jailparam_slot, rsv, nzjp); (void) memcpy(nzjp, zjp, sizeof (*zjp)); zjp = nzjp; mtx_unlock(&ppr->pr_mtx); done: if (zjpp != NULL) *zjpp = zjp; else mtx_unlock(&pr->pr_mtx); } /* * Jail OSD methods for ZFS VFS info. */ static int zfs_jailparam_create(void *obj, void *data) { struct prison *pr = obj; struct vfsoptlist *opts = data; int jsys; if (vfs_copyopt(opts, "zfs", &jsys, sizeof (jsys)) == 0 && jsys == JAIL_SYS_INHERIT) return (0); /* * Inherit a prison's initial values from its parent * (different from JAIL_SYS_INHERIT which also inherits changes). */ zfs_jailparam_alloc(pr, NULL); return (0); } static int zfs_jailparam_get(void *obj, void *data) { struct prison *ppr, *pr = obj; struct vfsoptlist *opts = data; struct zfs_jailparam *zjp; int jsys, error; zjp = zfs_jailparam_find(pr, &ppr); jsys = (ppr == pr) ? JAIL_SYS_NEW : JAIL_SYS_INHERIT; error = vfs_setopt(opts, "zfs", &jsys, sizeof (jsys)); if (error != 0 && error != ENOENT) goto done; if (jsys == JAIL_SYS_NEW) { error = vfs_setopt(opts, "zfs.mount_snapshot", &zjp->mount_snapshot, sizeof (zjp->mount_snapshot)); if (error != 0 && error != ENOENT) goto done; } else { /* * If this prison is inheriting its ZFS info, report * empty/zero parameters. */ static int mount_snapshot = 0; error = vfs_setopt(opts, "zfs.mount_snapshot", &mount_snapshot, sizeof (mount_snapshot)); if (error != 0 && error != ENOENT) goto done; } error = 0; done: mtx_unlock(&ppr->pr_mtx); return (error); } static int zfs_jailparam_set(void *obj, void *data) { struct prison *pr = obj; struct prison *ppr; struct vfsoptlist *opts = data; int error, jsys, mount_snapshot; /* Set the parameters, which should be correct. */ error = vfs_copyopt(opts, "zfs", &jsys, sizeof (jsys)); if (error == ENOENT) jsys = -1; error = vfs_copyopt(opts, "zfs.mount_snapshot", &mount_snapshot, sizeof (mount_snapshot)); if (error == ENOENT) mount_snapshot = -1; else jsys = JAIL_SYS_NEW; switch (jsys) { case JAIL_SYS_NEW: { /* "zfs=new" or "zfs.*": the prison gets its own ZFS info. */ struct zfs_jailparam *zjp; /* * A child jail cannot have more permissions than its parent */ if (pr->pr_parent != &prison0) { zjp = zfs_jailparam_find(pr->pr_parent, &ppr); mtx_unlock(&ppr->pr_mtx); if (zjp->mount_snapshot < mount_snapshot) { return (EPERM); } } zfs_jailparam_alloc(pr, &zjp); if (mount_snapshot != -1) zjp->mount_snapshot = mount_snapshot; mtx_unlock(&pr->pr_mtx); break; } case JAIL_SYS_INHERIT: /* "zfs=inherit": inherit the parent's ZFS info. */ mtx_lock(&pr->pr_mtx); osd_jail_del(pr, zfs_jailparam_slot); mtx_unlock(&pr->pr_mtx); break; case -1: /* * If the setting being changed is not ZFS related * then do nothing. */ break; } return (0); } static int zfs_jailparam_check(void *obj __unused, void *data) { struct vfsoptlist *opts = data; int error, jsys, mount_snapshot; /* Check that the parameters are correct. */ error = vfs_copyopt(opts, "zfs", &jsys, sizeof (jsys)); if (error != ENOENT) { if (error != 0) return (error); if (jsys != JAIL_SYS_NEW && jsys != JAIL_SYS_INHERIT) return (EINVAL); } error = vfs_copyopt(opts, "zfs.mount_snapshot", &mount_snapshot, sizeof (mount_snapshot)); if (error != ENOENT) { if (error != 0) return (error); if (mount_snapshot != 0 && mount_snapshot != 1) return (EINVAL); } return (0); } static void zfs_jailparam_destroy(void *data) { free(data, M_PRISON); } static void zfs_jailparam_sysinit(void *arg __unused) { struct prison *pr; osd_method_t methods[PR_MAXMETHOD] = { [PR_METHOD_CREATE] = zfs_jailparam_create, [PR_METHOD_GET] = zfs_jailparam_get, [PR_METHOD_SET] = zfs_jailparam_set, [PR_METHOD_CHECK] = zfs_jailparam_check, }; zfs_jailparam_slot = osd_jail_register(zfs_jailparam_destroy, methods); /* Copy the defaults to any existing prisons. */ sx_slock(&allprison_lock); TAILQ_FOREACH(pr, &allprison, pr_list) zfs_jailparam_alloc(pr, NULL); sx_sunlock(&allprison_lock); } static void zfs_jailparam_sysuninit(void *arg __unused) { osd_jail_deregister(zfs_jailparam_slot); } SYSINIT(zfs_jailparam_sysinit, SI_SUB_DRIVERS, SI_ORDER_ANY, zfs_jailparam_sysinit, NULL); SYSUNINIT(zfs_jailparam_sysuninit, SI_SUB_DRIVERS, SI_ORDER_ANY, zfs_jailparam_sysuninit, NULL); diff --git a/module/os/freebsd/zfs/zfs_vnops_os.c b/module/os/freebsd/zfs/zfs_vnops_os.c index c37f543cea02..107cd69c756c 100644 --- a/module/os/freebsd/zfs/zfs_vnops_os.c +++ b/module/os/freebsd/zfs/zfs_vnops_os.c @@ -1,6442 +1,6447 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2012, 2015 by Delphix. All rights reserved. * Copyright (c) 2014 Integros [integros.com] * Copyright 2017 Nexenta Systems, Inc. */ /* Portions Copyright 2007 Jeremy Teo */ /* Portions Copyright 2010 Robert Milkowski */ #include #include #include #include #include #include #include #include #include #include #if __FreeBSD_version >= 1300102 #include #endif #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifndef VN_OPEN_INVFS #define VN_OPEN_INVFS 0x0 #endif VFS_SMR_DECLARE; #if __FreeBSD_version < 1300103 #define NDFREE_PNBUF(ndp) NDFREE((ndp), NDF_ONLY_PNBUF) #endif #if __FreeBSD_version >= 1300047 #define vm_page_wire_lock(pp) #define vm_page_wire_unlock(pp) #else #define vm_page_wire_lock(pp) vm_page_lock(pp) #define vm_page_wire_unlock(pp) vm_page_unlock(pp) #endif #ifdef DEBUG_VFS_LOCKS #define VNCHECKREF(vp) \ VNASSERT((vp)->v_holdcnt > 0 && (vp)->v_usecount > 0, vp, \ ("%s: wrong ref counts", __func__)); #else #define VNCHECKREF(vp) #endif #if __FreeBSD_version >= 1400045 typedef uint64_t cookie_t; #else typedef ulong_t cookie_t; #endif /* * Programming rules. * * Each vnode op performs some logical unit of work. To do this, the ZPL must * properly lock its in-core state, create a DMU transaction, do the work, * record this work in the intent log (ZIL), commit the DMU transaction, * and wait for the intent log to commit if it is a synchronous operation. * Moreover, the vnode ops must work in both normal and log replay context. * The ordering of events is important to avoid deadlocks and references * to freed memory. The example below illustrates the following Big Rules: * * (1) A check must be made in each zfs thread for a mounted file system. * This is done avoiding races using zfs_enter(zfsvfs). * A zfs_exit(zfsvfs) is needed before all returns. Any znodes * must be checked with zfs_verify_zp(zp). Both of these macros * can return EIO from the calling function. * * (2) VN_RELE() should always be the last thing except for zil_commit() * (if necessary) and zfs_exit(). This is for 3 reasons: * First, if it's the last reference, the vnode/znode * can be freed, so the zp may point to freed memory. Second, the last * reference will call zfs_zinactive(), which may induce a lot of work -- * pushing cached pages (which acquires range locks) and syncing out * cached atime changes. Third, zfs_zinactive() may require a new tx, * which could deadlock the system if you were already holding one. * If you must call VN_RELE() within a tx then use VN_RELE_ASYNC(). * * (3) All range locks must be grabbed before calling dmu_tx_assign(), * as they can span dmu_tx_assign() calls. * * (4) If ZPL locks are held, pass TXG_NOWAIT as the second argument to * dmu_tx_assign(). This is critical because we don't want to block * while holding locks. * * If no ZPL locks are held (aside from zfs_enter()), use TXG_WAIT. This * reduces lock contention and CPU usage when we must wait (note that if * throughput is constrained by the storage, nearly every transaction * must wait). * * Note, in particular, that if a lock is sometimes acquired before * the tx assigns, and sometimes after (e.g. z_lock), then failing * to use a non-blocking assign can deadlock the system. The scenario: * * Thread A has grabbed a lock before calling dmu_tx_assign(). * Thread B is in an already-assigned tx, and blocks for this lock. * Thread A calls dmu_tx_assign(TXG_WAIT) and blocks in txg_wait_open() * forever, because the previous txg can't quiesce until B's tx commits. * * If dmu_tx_assign() returns ERESTART and zfsvfs->z_assign is TXG_NOWAIT, * then drop all locks, call dmu_tx_wait(), and try again. On subsequent * calls to dmu_tx_assign(), pass TXG_NOTHROTTLE in addition to TXG_NOWAIT, * to indicate that this operation has already called dmu_tx_wait(). * This will ensure that we don't retry forever, waiting a short bit * each time. * * (5) If the operation succeeded, generate the intent log entry for it * before dropping locks. This ensures that the ordering of events * in the intent log matches the order in which they actually occurred. * During ZIL replay the zfs_log_* functions will update the sequence * number to indicate the zil transaction has replayed. * * (6) At the end of each vnode op, the DMU tx must always commit, * regardless of whether there were any errors. * * (7) After dropping all locks, invoke zil_commit(zilog, foid) * to ensure that synchronous semantics are provided when necessary. * * In general, this is how things should be ordered in each vnode op: * * zfs_enter(zfsvfs); // exit if unmounted * top: * zfs_dirent_lookup(&dl, ...) // lock directory entry (may VN_HOLD()) * rw_enter(...); // grab any other locks you need * tx = dmu_tx_create(...); // get DMU tx * dmu_tx_hold_*(); // hold each object you might modify * error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT); * if (error) { * rw_exit(...); // drop locks * zfs_dirent_unlock(dl); // unlock directory entry * VN_RELE(...); // release held vnodes * if (error == ERESTART) { * waited = B_TRUE; * dmu_tx_wait(tx); * dmu_tx_abort(tx); * goto top; * } * dmu_tx_abort(tx); // abort DMU tx * zfs_exit(zfsvfs); // finished in zfs * return (error); // really out of space * } * error = do_real_work(); // do whatever this VOP does * if (error == 0) * zfs_log_*(...); // on success, make ZIL entry * dmu_tx_commit(tx); // commit DMU tx -- error or not * rw_exit(...); // drop locks * zfs_dirent_unlock(dl); // unlock directory entry * VN_RELE(...); // release held vnodes * zil_commit(zilog, foid); // synchronous when necessary * zfs_exit(zfsvfs); // finished in zfs * return (error); // done, report error */ static int zfs_open(vnode_t **vpp, int flag, cred_t *cr) { (void) cr; znode_t *zp = VTOZ(*vpp); zfsvfs_t *zfsvfs = zp->z_zfsvfs; int error; if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); if ((flag & FWRITE) && (zp->z_pflags & ZFS_APPENDONLY) && ((flag & FAPPEND) == 0)) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EPERM)); } /* * Keep a count of the synchronous opens in the znode. On first * synchronous open we must convert all previous async transactions * into sync to keep correct ordering. */ if (flag & O_SYNC) { if (atomic_inc_32_nv(&zp->z_sync_cnt) == 1) zil_async_to_sync(zfsvfs->z_log, zp->z_id); } zfs_exit(zfsvfs, FTAG); return (0); } static int zfs_close(vnode_t *vp, int flag, int count, offset_t offset, cred_t *cr) { (void) offset, (void) cr; znode_t *zp = VTOZ(vp); zfsvfs_t *zfsvfs = zp->z_zfsvfs; int error; if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); /* Decrement the synchronous opens in the znode */ if ((flag & O_SYNC) && (count == 1)) atomic_dec_32(&zp->z_sync_cnt); zfs_exit(zfsvfs, FTAG); return (0); } static int zfs_ioctl(vnode_t *vp, ulong_t com, intptr_t data, int flag, cred_t *cred, int *rvalp) { (void) flag, (void) cred, (void) rvalp; loff_t off; int error; switch (com) { case _FIOFFS: { return (0); /* * The following two ioctls are used by bfu. Faking out, * necessary to avoid bfu errors. */ } case _FIOGDIO: case _FIOSDIO: { return (0); } case F_SEEK_DATA: case F_SEEK_HOLE: { off = *(offset_t *)data; /* offset parameter is in/out */ error = zfs_holey(VTOZ(vp), com, &off); if (error) return (error); *(offset_t *)data = off; return (0); } } return (SET_ERROR(ENOTTY)); } static vm_page_t page_busy(vnode_t *vp, int64_t start, int64_t off, int64_t nbytes) { vm_object_t obj; vm_page_t pp; int64_t end; /* * At present vm_page_clear_dirty extends the cleared range to DEV_BSIZE * aligned boundaries, if the range is not aligned. As a result a * DEV_BSIZE subrange with partially dirty data may get marked as clean. * It may happen that all DEV_BSIZE subranges are marked clean and thus * the whole page would be considered clean despite have some * dirty data. * For this reason we should shrink the range to DEV_BSIZE aligned * boundaries before calling vm_page_clear_dirty. */ end = rounddown2(off + nbytes, DEV_BSIZE); off = roundup2(off, DEV_BSIZE); nbytes = end - off; obj = vp->v_object; zfs_vmobject_assert_wlocked_12(obj); #if __FreeBSD_version < 1300050 for (;;) { if ((pp = vm_page_lookup(obj, OFF_TO_IDX(start))) != NULL && pp->valid) { if (vm_page_xbusied(pp)) { /* * Reference the page before unlocking and * sleeping so that the page daemon is less * likely to reclaim it. */ vm_page_reference(pp); vm_page_lock(pp); zfs_vmobject_wunlock(obj); vm_page_busy_sleep(pp, "zfsmwb", true); zfs_vmobject_wlock(obj); continue; } vm_page_sbusy(pp); } else if (pp != NULL) { ASSERT(!pp->valid); pp = NULL; } if (pp != NULL) { ASSERT3U(pp->valid, ==, VM_PAGE_BITS_ALL); vm_object_pip_add(obj, 1); pmap_remove_write(pp); if (nbytes != 0) vm_page_clear_dirty(pp, off, nbytes); } break; } #else vm_page_grab_valid_unlocked(&pp, obj, OFF_TO_IDX(start), VM_ALLOC_NOCREAT | VM_ALLOC_SBUSY | VM_ALLOC_NORMAL | VM_ALLOC_IGN_SBUSY); if (pp != NULL) { ASSERT3U(pp->valid, ==, VM_PAGE_BITS_ALL); vm_object_pip_add(obj, 1); pmap_remove_write(pp); if (nbytes != 0) vm_page_clear_dirty(pp, off, nbytes); } #endif return (pp); } static void page_unbusy(vm_page_t pp) { vm_page_sunbusy(pp); #if __FreeBSD_version >= 1300041 vm_object_pip_wakeup(pp->object); #else vm_object_pip_subtract(pp->object, 1); #endif } #if __FreeBSD_version > 1300051 static vm_page_t page_hold(vnode_t *vp, int64_t start) { vm_object_t obj; vm_page_t m; obj = vp->v_object; vm_page_grab_valid_unlocked(&m, obj, OFF_TO_IDX(start), VM_ALLOC_NOCREAT | VM_ALLOC_WIRED | VM_ALLOC_IGN_SBUSY | VM_ALLOC_NOBUSY); return (m); } #else static vm_page_t page_hold(vnode_t *vp, int64_t start) { vm_object_t obj; vm_page_t pp; obj = vp->v_object; zfs_vmobject_assert_wlocked(obj); for (;;) { if ((pp = vm_page_lookup(obj, OFF_TO_IDX(start))) != NULL && pp->valid) { if (vm_page_xbusied(pp)) { /* * Reference the page before unlocking and * sleeping so that the page daemon is less * likely to reclaim it. */ vm_page_reference(pp); vm_page_lock(pp); zfs_vmobject_wunlock(obj); vm_page_busy_sleep(pp, "zfsmwb", true); zfs_vmobject_wlock(obj); continue; } ASSERT3U(pp->valid, ==, VM_PAGE_BITS_ALL); vm_page_wire_lock(pp); vm_page_hold(pp); vm_page_wire_unlock(pp); } else pp = NULL; break; } return (pp); } #endif static void page_unhold(vm_page_t pp) { vm_page_wire_lock(pp); #if __FreeBSD_version >= 1300035 vm_page_unwire(pp, PQ_ACTIVE); #else vm_page_unhold(pp); #endif vm_page_wire_unlock(pp); } /* * When a file is memory mapped, we must keep the IO data synchronized * between the DMU cache and the memory mapped pages. What this means: * * On Write: If we find a memory mapped page, we write to *both* * the page and the dmu buffer. */ void update_pages(znode_t *zp, int64_t start, int len, objset_t *os) { vm_object_t obj; struct sf_buf *sf; vnode_t *vp = ZTOV(zp); caddr_t va; int off; ASSERT3P(vp->v_mount, !=, NULL); obj = vp->v_object; ASSERT3P(obj, !=, NULL); off = start & PAGEOFFSET; zfs_vmobject_wlock_12(obj); #if __FreeBSD_version >= 1300041 vm_object_pip_add(obj, 1); #endif for (start &= PAGEMASK; len > 0; start += PAGESIZE) { vm_page_t pp; int nbytes = imin(PAGESIZE - off, len); if ((pp = page_busy(vp, start, off, nbytes)) != NULL) { zfs_vmobject_wunlock_12(obj); va = zfs_map_page(pp, &sf); (void) dmu_read(os, zp->z_id, start + off, nbytes, va + off, DMU_READ_PREFETCH); zfs_unmap_page(sf); zfs_vmobject_wlock_12(obj); page_unbusy(pp); } len -= nbytes; off = 0; } #if __FreeBSD_version >= 1300041 vm_object_pip_wakeup(obj); #else vm_object_pip_wakeupn(obj, 0); #endif zfs_vmobject_wunlock_12(obj); } /* * Read with UIO_NOCOPY flag means that sendfile(2) requests * ZFS to populate a range of page cache pages with data. * * NOTE: this function could be optimized to pre-allocate * all pages in advance, drain exclusive busy on all of them, * map them into contiguous KVA region and populate them * in one single dmu_read() call. */ int mappedread_sf(znode_t *zp, int nbytes, zfs_uio_t *uio) { vnode_t *vp = ZTOV(zp); objset_t *os = zp->z_zfsvfs->z_os; struct sf_buf *sf; vm_object_t obj; vm_page_t pp; int64_t start; caddr_t va; int len = nbytes; int error = 0; ASSERT3U(zfs_uio_segflg(uio), ==, UIO_NOCOPY); ASSERT3P(vp->v_mount, !=, NULL); obj = vp->v_object; ASSERT3P(obj, !=, NULL); ASSERT0(zfs_uio_offset(uio) & PAGEOFFSET); zfs_vmobject_wlock_12(obj); for (start = zfs_uio_offset(uio); len > 0; start += PAGESIZE) { int bytes = MIN(PAGESIZE, len); pp = vm_page_grab_unlocked(obj, OFF_TO_IDX(start), VM_ALLOC_SBUSY | VM_ALLOC_NORMAL | VM_ALLOC_IGN_SBUSY); if (vm_page_none_valid(pp)) { zfs_vmobject_wunlock_12(obj); va = zfs_map_page(pp, &sf); error = dmu_read(os, zp->z_id, start, bytes, va, DMU_READ_PREFETCH); if (bytes != PAGESIZE && error == 0) memset(va + bytes, 0, PAGESIZE - bytes); zfs_unmap_page(sf); zfs_vmobject_wlock_12(obj); #if __FreeBSD_version >= 1300081 if (error == 0) { vm_page_valid(pp); vm_page_activate(pp); vm_page_do_sunbusy(pp); } else { zfs_vmobject_wlock(obj); if (!vm_page_wired(pp) && pp->valid == 0 && vm_page_busy_tryupgrade(pp)) vm_page_free(pp); else vm_page_sunbusy(pp); zfs_vmobject_wunlock(obj); } #else vm_page_do_sunbusy(pp); vm_page_lock(pp); if (error) { if (pp->wire_count == 0 && pp->valid == 0 && !vm_page_busied(pp)) vm_page_free(pp); } else { pp->valid = VM_PAGE_BITS_ALL; vm_page_activate(pp); } vm_page_unlock(pp); #endif } else { ASSERT3U(pp->valid, ==, VM_PAGE_BITS_ALL); vm_page_do_sunbusy(pp); } if (error) break; zfs_uio_advance(uio, bytes); len -= bytes; } zfs_vmobject_wunlock_12(obj); return (error); } /* * When a file is memory mapped, we must keep the IO data synchronized * between the DMU cache and the memory mapped pages. What this means: * * On Read: We "read" preferentially from memory mapped pages, * else we default from the dmu buffer. * * NOTE: We will always "break up" the IO into PAGESIZE uiomoves when * the file is memory mapped. */ int mappedread(znode_t *zp, int nbytes, zfs_uio_t *uio) { vnode_t *vp = ZTOV(zp); vm_object_t obj; int64_t start; int len = nbytes; int off; int error = 0; ASSERT3P(vp->v_mount, !=, NULL); obj = vp->v_object; ASSERT3P(obj, !=, NULL); start = zfs_uio_offset(uio); off = start & PAGEOFFSET; zfs_vmobject_wlock_12(obj); for (start &= PAGEMASK; len > 0; start += PAGESIZE) { vm_page_t pp; uint64_t bytes = MIN(PAGESIZE - off, len); if ((pp = page_hold(vp, start))) { struct sf_buf *sf; caddr_t va; zfs_vmobject_wunlock_12(obj); va = zfs_map_page(pp, &sf); error = vn_io_fault_uiomove(va + off, bytes, GET_UIO_STRUCT(uio)); zfs_unmap_page(sf); zfs_vmobject_wlock_12(obj); page_unhold(pp); } else { zfs_vmobject_wunlock_12(obj); error = dmu_read_uio_dbuf(sa_get_db(zp->z_sa_hdl), uio, bytes); zfs_vmobject_wlock_12(obj); } len -= bytes; off = 0; if (error) break; } zfs_vmobject_wunlock_12(obj); return (error); } int zfs_write_simple(znode_t *zp, const void *data, size_t len, loff_t pos, size_t *presid) { int error = 0; ssize_t resid; error = vn_rdwr(UIO_WRITE, ZTOV(zp), __DECONST(void *, data), len, pos, UIO_SYSSPACE, IO_SYNC, kcred, NOCRED, &resid, curthread); if (error) { return (SET_ERROR(error)); } else if (presid == NULL) { if (resid != 0) { error = SET_ERROR(EIO); } } else { *presid = resid; } return (error); } void zfs_zrele_async(znode_t *zp) { vnode_t *vp = ZTOV(zp); objset_t *os = ITOZSB(vp)->z_os; VN_RELE_ASYNC(vp, dsl_pool_zrele_taskq(dmu_objset_pool(os))); } static int zfs_dd_callback(struct mount *mp, void *arg, int lkflags, struct vnode **vpp) { int error; *vpp = arg; error = vn_lock(*vpp, lkflags); if (error != 0) vrele(*vpp); return (error); } static int zfs_lookup_lock(vnode_t *dvp, vnode_t *vp, const char *name, int lkflags) { znode_t *zdp = VTOZ(dvp); zfsvfs_t *zfsvfs __unused = zdp->z_zfsvfs; int error; int ltype; if (zfsvfs->z_replay == B_FALSE) ASSERT_VOP_LOCKED(dvp, __func__); if (name[0] == 0 || (name[0] == '.' && name[1] == 0)) { ASSERT3P(dvp, ==, vp); vref(dvp); ltype = lkflags & LK_TYPE_MASK; if (ltype != VOP_ISLOCKED(dvp)) { if (ltype == LK_EXCLUSIVE) vn_lock(dvp, LK_UPGRADE | LK_RETRY); else /* if (ltype == LK_SHARED) */ vn_lock(dvp, LK_DOWNGRADE | LK_RETRY); /* * Relock for the "." case could leave us with * reclaimed vnode. */ if (VN_IS_DOOMED(dvp)) { vrele(dvp); return (SET_ERROR(ENOENT)); } } return (0); } else if (name[0] == '.' && name[1] == '.' && name[2] == 0) { /* * Note that in this case, dvp is the child vnode, and we * are looking up the parent vnode - exactly reverse from * normal operation. Unlocking dvp requires some rather * tricky unlock/relock dance to prevent mp from being freed; * use vn_vget_ino_gen() which takes care of all that. * * XXX Note that there is a time window when both vnodes are * unlocked. It is possible, although highly unlikely, that * during that window the parent-child relationship between * the vnodes may change, for example, get reversed. * In that case we would have a wrong lock order for the vnodes. * All other filesystems seem to ignore this problem, so we * do the same here. * A potential solution could be implemented as follows: * - using LK_NOWAIT when locking the second vnode and retrying * if necessary * - checking that the parent-child relationship still holds * after locking both vnodes and retrying if it doesn't */ error = vn_vget_ino_gen(dvp, zfs_dd_callback, vp, lkflags, &vp); return (error); } else { error = vn_lock(vp, lkflags); if (error != 0) vrele(vp); return (error); } } /* * Lookup an entry in a directory, or an extended attribute directory. * If it exists, return a held vnode reference for it. * * IN: dvp - vnode of directory to search. * nm - name of entry to lookup. * pnp - full pathname to lookup [UNUSED]. * flags - LOOKUP_XATTR set if looking for an attribute. * rdir - root directory vnode [UNUSED]. * cr - credentials of caller. * ct - caller context * * OUT: vpp - vnode of located entry, NULL if not found. * * RETURN: 0 on success, error code on failure. * * Timestamps: * NA */ static int zfs_lookup(vnode_t *dvp, const char *nm, vnode_t **vpp, struct componentname *cnp, int nameiop, cred_t *cr, int flags, boolean_t cached) { znode_t *zdp = VTOZ(dvp); znode_t *zp; zfsvfs_t *zfsvfs = zdp->z_zfsvfs; #if __FreeBSD_version > 1300124 seqc_t dvp_seqc; #endif int error = 0; /* * Fast path lookup, however we must skip DNLC lookup * for case folding or normalizing lookups because the * DNLC code only stores the passed in name. This means * creating 'a' and removing 'A' on a case insensitive * file system would work, but DNLC still thinks 'a' * exists and won't let you create it again on the next * pass through fast path. */ if (!(flags & LOOKUP_XATTR)) { if (dvp->v_type != VDIR) { return (SET_ERROR(ENOTDIR)); } else if (zdp->z_sa_hdl == NULL) { return (SET_ERROR(EIO)); } } DTRACE_PROBE2(zfs__fastpath__lookup__miss, vnode_t *, dvp, const char *, nm); if ((error = zfs_enter_verify_zp(zfsvfs, zdp, FTAG)) != 0) return (error); #if __FreeBSD_version > 1300124 dvp_seqc = vn_seqc_read_notmodify(dvp); #endif *vpp = NULL; if (flags & LOOKUP_XATTR) { /* * If the xattr property is off, refuse the lookup request. */ if (!(zfsvfs->z_flags & ZSB_XATTR)) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EOPNOTSUPP)); } /* * We don't allow recursive attributes.. * Maybe someday we will. */ if (zdp->z_pflags & ZFS_XATTR) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EINVAL)); } if ((error = zfs_get_xattrdir(VTOZ(dvp), &zp, cr, flags))) { zfs_exit(zfsvfs, FTAG); return (error); } *vpp = ZTOV(zp); /* * Do we have permission to get into attribute directory? */ error = zfs_zaccess(zp, ACE_EXECUTE, 0, B_FALSE, cr, NULL); if (error) { vrele(ZTOV(zp)); } zfs_exit(zfsvfs, FTAG); return (error); } /* * Check accessibility of directory if we're not coming in via * VOP_CACHEDLOOKUP. */ if (!cached) { #ifdef NOEXECCHECK if ((cnp->cn_flags & NOEXECCHECK) != 0) { cnp->cn_flags &= ~NOEXECCHECK; } else #endif if ((error = zfs_zaccess(zdp, ACE_EXECUTE, 0, B_FALSE, cr, NULL))) { zfs_exit(zfsvfs, FTAG); return (error); } } if (zfsvfs->z_utf8 && u8_validate(nm, strlen(nm), NULL, U8_VALIDATE_ENTIRE, &error) < 0) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EILSEQ)); } /* * First handle the special cases. */ if ((cnp->cn_flags & ISDOTDOT) != 0) { /* * If we are a snapshot mounted under .zfs, return * the vp for the snapshot directory. */ if (zdp->z_id == zfsvfs->z_root && zfsvfs->z_parent != zfsvfs) { struct componentname cn; vnode_t *zfsctl_vp; int ltype; zfs_exit(zfsvfs, FTAG); ltype = VOP_ISLOCKED(dvp); VOP_UNLOCK1(dvp); error = zfsctl_root(zfsvfs->z_parent, LK_SHARED, &zfsctl_vp); if (error == 0) { cn.cn_nameptr = "snapshot"; cn.cn_namelen = strlen(cn.cn_nameptr); cn.cn_nameiop = cnp->cn_nameiop; cn.cn_flags = cnp->cn_flags & ~ISDOTDOT; cn.cn_lkflags = cnp->cn_lkflags; error = VOP_LOOKUP(zfsctl_vp, vpp, &cn); vput(zfsctl_vp); } vn_lock(dvp, ltype | LK_RETRY); return (error); } } if (zfs_has_ctldir(zdp) && strcmp(nm, ZFS_CTLDIR_NAME) == 0) { zfs_exit(zfsvfs, FTAG); if ((cnp->cn_flags & ISLASTCN) != 0 && nameiop != LOOKUP) return (SET_ERROR(ENOTSUP)); error = zfsctl_root(zfsvfs, cnp->cn_lkflags, vpp); return (error); } /* * The loop is retry the lookup if the parent-child relationship * changes during the dot-dot locking complexities. */ for (;;) { uint64_t parent; error = zfs_dirlook(zdp, nm, &zp); if (error == 0) *vpp = ZTOV(zp); zfs_exit(zfsvfs, FTAG); if (error != 0) break; error = zfs_lookup_lock(dvp, *vpp, nm, cnp->cn_lkflags); if (error != 0) { /* * If we've got a locking error, then the vnode * got reclaimed because of a force unmount. * We never enter doomed vnodes into the name cache. */ *vpp = NULL; return (error); } if ((cnp->cn_flags & ISDOTDOT) == 0) break; if ((error = zfs_enter(zfsvfs, FTAG)) != 0) { vput(ZTOV(zp)); *vpp = NULL; return (error); } if (zdp->z_sa_hdl == NULL) { error = SET_ERROR(EIO); } else { error = sa_lookup(zdp->z_sa_hdl, SA_ZPL_PARENT(zfsvfs), &parent, sizeof (parent)); } if (error != 0) { zfs_exit(zfsvfs, FTAG); vput(ZTOV(zp)); break; } if (zp->z_id == parent) { zfs_exit(zfsvfs, FTAG); break; } vput(ZTOV(zp)); } if (error != 0) *vpp = NULL; /* Translate errors and add SAVENAME when needed. */ if (cnp->cn_flags & ISLASTCN) { switch (nameiop) { case CREATE: case RENAME: if (error == ENOENT) { error = EJUSTRETURN; #if __FreeBSD_version < 1400068 cnp->cn_flags |= SAVENAME; #endif break; } zfs_fallthrough; case DELETE: #if __FreeBSD_version < 1400068 if (error == 0) cnp->cn_flags |= SAVENAME; #endif break; } } #if __FreeBSD_version > 1300124 if ((cnp->cn_flags & ISDOTDOT) != 0) { /* * FIXME: zfs_lookup_lock relocks vnodes and does nothing to * handle races. In particular different callers may end up * with different vnodes and will try to add conflicting * entries to the namecache. * * While finding different result may be acceptable in face * of concurrent modification, adding conflicting entries * trips over an assert in the namecache. * * Ultimately let an entry through once everything settles. */ if (!vn_seqc_consistent(dvp, dvp_seqc)) { cnp->cn_flags &= ~MAKEENTRY; } } #endif /* Insert name into cache (as non-existent) if appropriate. */ if (zfsvfs->z_use_namecache && !zfsvfs->z_replay && error == ENOENT && (cnp->cn_flags & MAKEENTRY) != 0) cache_enter(dvp, NULL, cnp); /* Insert name into cache if appropriate. */ if (zfsvfs->z_use_namecache && !zfsvfs->z_replay && error == 0 && (cnp->cn_flags & MAKEENTRY)) { if (!(cnp->cn_flags & ISLASTCN) || (nameiop != DELETE && nameiop != RENAME)) { cache_enter(dvp, *vpp, cnp); } } return (error); } /* * Attempt to create a new entry in a directory. If the entry * already exists, truncate the file if permissible, else return * an error. Return the vp of the created or trunc'd file. * * IN: dvp - vnode of directory to put new file entry in. * name - name of new file entry. * vap - attributes of new file. * excl - flag indicating exclusive or non-exclusive mode. * mode - mode to open file with. * cr - credentials of caller. * flag - large file flag [UNUSED]. * ct - caller context * vsecp - ACL to be set * mnt_ns - Unused on FreeBSD * * OUT: vpp - vnode of created or trunc'd entry. * * RETURN: 0 on success, error code on failure. * * Timestamps: * dvp - ctime|mtime updated if new entry created * vp - ctime|mtime always, atime if new */ int zfs_create(znode_t *dzp, const char *name, vattr_t *vap, int excl, int mode, znode_t **zpp, cred_t *cr, int flag, vsecattr_t *vsecp, zidmap_t *mnt_ns) { (void) excl, (void) mode, (void) flag; znode_t *zp; zfsvfs_t *zfsvfs = dzp->z_zfsvfs; zilog_t *zilog; objset_t *os; dmu_tx_t *tx; int error; uid_t uid = crgetuid(cr); gid_t gid = crgetgid(cr); uint64_t projid = ZFS_DEFAULT_PROJID; zfs_acl_ids_t acl_ids; boolean_t fuid_dirtied; uint64_t txtype; #ifdef DEBUG_VFS_LOCKS vnode_t *dvp = ZTOV(dzp); #endif /* * If we have an ephemeral id, ACL, or XVATTR then * make sure file system is at proper version */ if (zfsvfs->z_use_fuids == B_FALSE && (vsecp || (vap->va_mask & AT_XVATTR) || IS_EPHEMERAL(uid) || IS_EPHEMERAL(gid))) return (SET_ERROR(EINVAL)); if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0) return (error); os = zfsvfs->z_os; zilog = zfsvfs->z_log; if (zfsvfs->z_utf8 && u8_validate(name, strlen(name), NULL, U8_VALIDATE_ENTIRE, &error) < 0) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EILSEQ)); } if (vap->va_mask & AT_XVATTR) { if ((error = secpolicy_xvattr(ZTOV(dzp), (xvattr_t *)vap, crgetuid(cr), cr, vap->va_type)) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } } *zpp = NULL; if ((vap->va_mode & S_ISVTX) && secpolicy_vnode_stky_modify(cr)) vap->va_mode &= ~S_ISVTX; error = zfs_dirent_lookup(dzp, name, &zp, ZNEW); if (error) { zfs_exit(zfsvfs, FTAG); return (error); } ASSERT3P(zp, ==, NULL); /* * Create a new file object and update the directory * to reference it. */ if ((error = zfs_zaccess(dzp, ACE_ADD_FILE, 0, B_FALSE, cr, mnt_ns))) { goto out; } /* * We only support the creation of regular files in * extended attribute directories. */ if ((dzp->z_pflags & ZFS_XATTR) && (vap->va_type != VREG)) { error = SET_ERROR(EINVAL); goto out; } if ((error = zfs_acl_ids_create(dzp, 0, vap, cr, vsecp, &acl_ids, NULL)) != 0) goto out; if (S_ISREG(vap->va_mode) || S_ISDIR(vap->va_mode)) projid = zfs_inherit_projid(dzp); if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, projid)) { zfs_acl_ids_free(&acl_ids); error = SET_ERROR(EDQUOT); goto out; } getnewvnode_reserve_(); tx = dmu_tx_create(os); dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes + ZFS_SA_BASE_ATTR_SIZE); fuid_dirtied = zfsvfs->z_fuid_dirty; if (fuid_dirtied) zfs_fuid_txhold(zfsvfs, tx); dmu_tx_hold_zap(tx, dzp->z_id, TRUE, name); dmu_tx_hold_sa(tx, dzp->z_sa_hdl, B_FALSE); if (!zfsvfs->z_use_sa && acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) { dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, acl_ids.z_aclp->z_acl_bytes); } error = dmu_tx_assign(tx, TXG_WAIT); if (error) { zfs_acl_ids_free(&acl_ids); dmu_tx_abort(tx); getnewvnode_drop_reserve(); zfs_exit(zfsvfs, FTAG); return (error); } zfs_mknode(dzp, vap, tx, cr, 0, &zp, &acl_ids); if (fuid_dirtied) zfs_fuid_sync(zfsvfs, tx); (void) zfs_link_create(dzp, name, zp, tx, ZNEW); txtype = zfs_log_create_txtype(Z_FILE, vsecp, vap); zfs_log_create(zilog, tx, txtype, dzp, zp, name, vsecp, acl_ids.z_fuidp, vap); zfs_acl_ids_free(&acl_ids); dmu_tx_commit(tx); getnewvnode_drop_reserve(); out: VNCHECKREF(dvp); if (error == 0) { *zpp = zp; } if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS) zil_commit(zilog, 0); zfs_exit(zfsvfs, FTAG); return (error); } /* * Remove an entry from a directory. * * IN: dvp - vnode of directory to remove entry from. * name - name of entry to remove. * cr - credentials of caller. * ct - caller context * flags - case flags * * RETURN: 0 on success, error code on failure. * * Timestamps: * dvp - ctime|mtime * vp - ctime (if nlink > 0) */ static int zfs_remove_(vnode_t *dvp, vnode_t *vp, const char *name, cred_t *cr) { znode_t *dzp = VTOZ(dvp); znode_t *zp; znode_t *xzp; zfsvfs_t *zfsvfs = dzp->z_zfsvfs; zilog_t *zilog; uint64_t xattr_obj; uint64_t obj = 0; dmu_tx_t *tx; boolean_t unlinked; uint64_t txtype; int error; if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0) return (error); zp = VTOZ(vp); if ((error = zfs_verify_zp(zp)) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } zilog = zfsvfs->z_log; xattr_obj = 0; xzp = NULL; if ((error = zfs_zaccess_delete(dzp, zp, cr, NULL))) { goto out; } /* * Need to use rmdir for removing directories. */ if (vp->v_type == VDIR) { error = SET_ERROR(EPERM); goto out; } vnevent_remove(vp, dvp, name, ct); obj = zp->z_id; /* are there any extended attributes? */ error = sa_lookup(zp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs), &xattr_obj, sizeof (xattr_obj)); if (error == 0 && xattr_obj) { error = zfs_zget(zfsvfs, xattr_obj, &xzp); ASSERT0(error); } /* * We may delete the znode now, or we may put it in the unlinked set; * it depends on whether we're the last link, and on whether there are * other holds on the vnode. So we dmu_tx_hold() the right things to * allow for either case. */ tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_zap(tx, dzp->z_id, FALSE, name); dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE); zfs_sa_upgrade_txholds(tx, zp); zfs_sa_upgrade_txholds(tx, dzp); if (xzp) { dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE); dmu_tx_hold_sa(tx, xzp->z_sa_hdl, B_FALSE); } /* charge as an update -- would be nice not to charge at all */ dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL); /* * Mark this transaction as typically resulting in a net free of space */ dmu_tx_mark_netfree(tx); error = dmu_tx_assign(tx, TXG_WAIT); if (error) { dmu_tx_abort(tx); zfs_exit(zfsvfs, FTAG); return (error); } /* * Remove the directory entry. */ error = zfs_link_destroy(dzp, name, zp, tx, ZEXISTS, &unlinked); if (error) { dmu_tx_commit(tx); goto out; } if (unlinked) { zfs_unlinked_add(zp, tx); vp->v_vflag |= VV_NOSYNC; } /* XXX check changes to linux vnops */ txtype = TX_REMOVE; zfs_log_remove(zilog, tx, txtype, dzp, name, obj, unlinked); dmu_tx_commit(tx); out: if (xzp) vrele(ZTOV(xzp)); if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS) zil_commit(zilog, 0); zfs_exit(zfsvfs, FTAG); return (error); } static int zfs_lookup_internal(znode_t *dzp, const char *name, vnode_t **vpp, struct componentname *cnp, int nameiop) { zfsvfs_t *zfsvfs = dzp->z_zfsvfs; int error; cnp->cn_nameptr = __DECONST(char *, name); cnp->cn_namelen = strlen(name); cnp->cn_nameiop = nameiop; cnp->cn_flags = ISLASTCN; #if __FreeBSD_version < 1400068 cnp->cn_flags |= SAVENAME; #endif cnp->cn_lkflags = LK_EXCLUSIVE | LK_RETRY; cnp->cn_cred = kcred; #if __FreeBSD_version < 1400037 cnp->cn_thread = curthread; #endif if (zfsvfs->z_use_namecache && !zfsvfs->z_replay) { struct vop_lookup_args a; a.a_gen.a_desc = &vop_lookup_desc; a.a_dvp = ZTOV(dzp); a.a_vpp = vpp; a.a_cnp = cnp; error = vfs_cache_lookup(&a); } else { error = zfs_lookup(ZTOV(dzp), name, vpp, cnp, nameiop, kcred, 0, B_FALSE); } #ifdef ZFS_DEBUG if (error) { printf("got error %d on name %s on op %d\n", error, name, nameiop); kdb_backtrace(); } #endif return (error); } int zfs_remove(znode_t *dzp, const char *name, cred_t *cr, int flags) { vnode_t *vp; int error; struct componentname cn; if ((error = zfs_lookup_internal(dzp, name, &vp, &cn, DELETE))) return (error); error = zfs_remove_(ZTOV(dzp), vp, name, cr); vput(vp); return (error); } /* * Create a new directory and insert it into dvp using the name * provided. Return a pointer to the inserted directory. * * IN: dvp - vnode of directory to add subdir to. * dirname - name of new directory. * vap - attributes of new directory. * cr - credentials of caller. * ct - caller context * flags - case flags * vsecp - ACL to be set * mnt_ns - Unused on FreeBSD * * OUT: vpp - vnode of created directory. * * RETURN: 0 on success, error code on failure. * * Timestamps: * dvp - ctime|mtime updated * vp - ctime|mtime|atime updated */ int zfs_mkdir(znode_t *dzp, const char *dirname, vattr_t *vap, znode_t **zpp, cred_t *cr, int flags, vsecattr_t *vsecp, zidmap_t *mnt_ns) { (void) flags, (void) vsecp; znode_t *zp; zfsvfs_t *zfsvfs = dzp->z_zfsvfs; zilog_t *zilog; uint64_t txtype; dmu_tx_t *tx; int error; uid_t uid = crgetuid(cr); gid_t gid = crgetgid(cr); zfs_acl_ids_t acl_ids; boolean_t fuid_dirtied; ASSERT3U(vap->va_type, ==, VDIR); /* * If we have an ephemeral id, ACL, or XVATTR then * make sure file system is at proper version */ if (zfsvfs->z_use_fuids == B_FALSE && ((vap->va_mask & AT_XVATTR) || IS_EPHEMERAL(uid) || IS_EPHEMERAL(gid))) return (SET_ERROR(EINVAL)); if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0) return (error); zilog = zfsvfs->z_log; if (dzp->z_pflags & ZFS_XATTR) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EINVAL)); } if (zfsvfs->z_utf8 && u8_validate(dirname, strlen(dirname), NULL, U8_VALIDATE_ENTIRE, &error) < 0) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EILSEQ)); } if (vap->va_mask & AT_XVATTR) { if ((error = secpolicy_xvattr(ZTOV(dzp), (xvattr_t *)vap, crgetuid(cr), cr, vap->va_type)) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } } if ((error = zfs_acl_ids_create(dzp, 0, vap, cr, NULL, &acl_ids, NULL)) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } /* * First make sure the new directory doesn't exist. * * Existence is checked first to make sure we don't return * EACCES instead of EEXIST which can cause some applications * to fail. */ *zpp = NULL; if ((error = zfs_dirent_lookup(dzp, dirname, &zp, ZNEW))) { zfs_acl_ids_free(&acl_ids); zfs_exit(zfsvfs, FTAG); return (error); } ASSERT3P(zp, ==, NULL); if ((error = zfs_zaccess(dzp, ACE_ADD_SUBDIRECTORY, 0, B_FALSE, cr, mnt_ns))) { zfs_acl_ids_free(&acl_ids); zfs_exit(zfsvfs, FTAG); return (error); } if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, zfs_inherit_projid(dzp))) { zfs_acl_ids_free(&acl_ids); zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EDQUOT)); } /* * Add a new entry to the directory. */ getnewvnode_reserve_(); tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_zap(tx, dzp->z_id, TRUE, dirname); dmu_tx_hold_zap(tx, DMU_NEW_OBJECT, FALSE, NULL); fuid_dirtied = zfsvfs->z_fuid_dirty; if (fuid_dirtied) zfs_fuid_txhold(zfsvfs, tx); if (!zfsvfs->z_use_sa && acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) { dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, acl_ids.z_aclp->z_acl_bytes); } dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes + ZFS_SA_BASE_ATTR_SIZE); error = dmu_tx_assign(tx, TXG_WAIT); if (error) { zfs_acl_ids_free(&acl_ids); dmu_tx_abort(tx); getnewvnode_drop_reserve(); zfs_exit(zfsvfs, FTAG); return (error); } /* * Create new node. */ zfs_mknode(dzp, vap, tx, cr, 0, &zp, &acl_ids); if (fuid_dirtied) zfs_fuid_sync(zfsvfs, tx); /* * Now put new name in parent dir. */ (void) zfs_link_create(dzp, dirname, zp, tx, ZNEW); *zpp = zp; txtype = zfs_log_create_txtype(Z_DIR, NULL, vap); zfs_log_create(zilog, tx, txtype, dzp, zp, dirname, NULL, acl_ids.z_fuidp, vap); zfs_acl_ids_free(&acl_ids); dmu_tx_commit(tx); getnewvnode_drop_reserve(); if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS) zil_commit(zilog, 0); zfs_exit(zfsvfs, FTAG); return (0); } #if __FreeBSD_version < 1300124 static void cache_vop_rmdir(struct vnode *dvp, struct vnode *vp) { cache_purge(dvp); cache_purge(vp); } #endif /* * Remove a directory subdir entry. If the current working * directory is the same as the subdir to be removed, the * remove will fail. * * IN: dvp - vnode of directory to remove from. * name - name of directory to be removed. * cwd - vnode of current working directory. * cr - credentials of caller. * ct - caller context * flags - case flags * * RETURN: 0 on success, error code on failure. * * Timestamps: * dvp - ctime|mtime updated */ static int zfs_rmdir_(vnode_t *dvp, vnode_t *vp, const char *name, cred_t *cr) { znode_t *dzp = VTOZ(dvp); znode_t *zp = VTOZ(vp); zfsvfs_t *zfsvfs = dzp->z_zfsvfs; zilog_t *zilog; dmu_tx_t *tx; int error; if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0) return (error); if ((error = zfs_verify_zp(zp)) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } zilog = zfsvfs->z_log; if ((error = zfs_zaccess_delete(dzp, zp, cr, NULL))) { goto out; } if (vp->v_type != VDIR) { error = SET_ERROR(ENOTDIR); goto out; } vnevent_rmdir(vp, dvp, name, ct); tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_zap(tx, dzp->z_id, FALSE, name); dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE); dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL); zfs_sa_upgrade_txholds(tx, zp); zfs_sa_upgrade_txholds(tx, dzp); dmu_tx_mark_netfree(tx); error = dmu_tx_assign(tx, TXG_WAIT); if (error) { dmu_tx_abort(tx); zfs_exit(zfsvfs, FTAG); return (error); } error = zfs_link_destroy(dzp, name, zp, tx, ZEXISTS, NULL); if (error == 0) { uint64_t txtype = TX_RMDIR; zfs_log_remove(zilog, tx, txtype, dzp, name, ZFS_NO_OBJECT, B_FALSE); } dmu_tx_commit(tx); if (zfsvfs->z_use_namecache) cache_vop_rmdir(dvp, vp); out: if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS) zil_commit(zilog, 0); zfs_exit(zfsvfs, FTAG); return (error); } int zfs_rmdir(znode_t *dzp, const char *name, znode_t *cwd, cred_t *cr, int flags) { struct componentname cn; vnode_t *vp; int error; if ((error = zfs_lookup_internal(dzp, name, &vp, &cn, DELETE))) return (error); error = zfs_rmdir_(ZTOV(dzp), vp, name, cr); vput(vp); return (error); } /* * Read as many directory entries as will fit into the provided * buffer from the given directory cursor position (specified in * the uio structure). * * IN: vp - vnode of directory to read. * uio - structure supplying read location, range info, * and return buffer. * cr - credentials of caller. * ct - caller context * * OUT: uio - updated offset and range, buffer filled. * eofp - set to true if end-of-file detected. * ncookies- number of entries in cookies * cookies - offsets to directory entries * * RETURN: 0 on success, error code on failure. * * Timestamps: * vp - atime updated * * Note that the low 4 bits of the cookie returned by zap is always zero. * This allows us to use the low range for "special" directory entries: * We use 0 for '.', and 1 for '..'. If this is the root of the filesystem, * we use the offset 2 for the '.zfs' directory. */ static int zfs_readdir(vnode_t *vp, zfs_uio_t *uio, cred_t *cr, int *eofp, int *ncookies, cookie_t **cookies) { znode_t *zp = VTOZ(vp); iovec_t *iovp; dirent64_t *odp; zfsvfs_t *zfsvfs = zp->z_zfsvfs; objset_t *os; caddr_t outbuf; size_t bufsize; zap_cursor_t zc; zap_attribute_t zap; uint_t bytes_wanted; uint64_t offset; /* must be unsigned; checks for < 1 */ uint64_t parent; int local_eof; int outcount; int error; uint8_t prefetch; uint8_t type; int ncooks; cookie_t *cooks = NULL; if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_PARENT(zfsvfs), &parent, sizeof (parent))) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } /* * If we are not given an eof variable, * use a local one. */ if (eofp == NULL) eofp = &local_eof; /* * Check for valid iov_len. */ if (GET_UIO_STRUCT(uio)->uio_iov->iov_len <= 0) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EINVAL)); } /* * Quit if directory has been removed (posix) */ if ((*eofp = zp->z_unlinked) != 0) { zfs_exit(zfsvfs, FTAG); return (0); } error = 0; os = zfsvfs->z_os; offset = zfs_uio_offset(uio); prefetch = zp->z_zn_prefetch; /* * Initialize the iterator cursor. */ if (offset <= 3) { /* * Start iteration from the beginning of the directory. */ zap_cursor_init(&zc, os, zp->z_id); } else { /* * The offset is a serialized cursor. */ zap_cursor_init_serialized(&zc, os, zp->z_id, offset); } /* * Get space to change directory entries into fs independent format. */ iovp = GET_UIO_STRUCT(uio)->uio_iov; bytes_wanted = iovp->iov_len; if (zfs_uio_segflg(uio) != UIO_SYSSPACE || zfs_uio_iovcnt(uio) != 1) { bufsize = bytes_wanted; outbuf = kmem_alloc(bufsize, KM_SLEEP); odp = (struct dirent64 *)outbuf; } else { bufsize = bytes_wanted; outbuf = NULL; odp = (struct dirent64 *)iovp->iov_base; } if (ncookies != NULL) { /* * Minimum entry size is dirent size and 1 byte for a file name. */ ncooks = zfs_uio_resid(uio) / (sizeof (struct dirent) - sizeof (((struct dirent *)NULL)->d_name) + 1); cooks = malloc(ncooks * sizeof (*cooks), M_TEMP, M_WAITOK); *cookies = cooks; *ncookies = ncooks; } /* * Transform to file-system independent format */ outcount = 0; while (outcount < bytes_wanted) { ino64_t objnum; ushort_t reclen; off64_t *next = NULL; /* * Special case `.', `..', and `.zfs'. */ if (offset == 0) { (void) strcpy(zap.za_name, "."); zap.za_normalization_conflict = 0; objnum = zp->z_id; type = DT_DIR; } else if (offset == 1) { (void) strcpy(zap.za_name, ".."); zap.za_normalization_conflict = 0; objnum = parent; type = DT_DIR; } else if (offset == 2 && zfs_show_ctldir(zp)) { (void) strcpy(zap.za_name, ZFS_CTLDIR_NAME); zap.za_normalization_conflict = 0; objnum = ZFSCTL_INO_ROOT; type = DT_DIR; } else { /* * Grab next entry. */ if ((error = zap_cursor_retrieve(&zc, &zap))) { if ((*eofp = (error == ENOENT)) != 0) break; else goto update; } if (zap.za_integer_length != 8 || zap.za_num_integers != 1) { cmn_err(CE_WARN, "zap_readdir: bad directory " "entry, obj = %lld, offset = %lld\n", (u_longlong_t)zp->z_id, (u_longlong_t)offset); error = SET_ERROR(ENXIO); goto update; } objnum = ZFS_DIRENT_OBJ(zap.za_first_integer); /* * MacOS X can extract the object type here such as: * uint8_t type = ZFS_DIRENT_TYPE(zap.za_first_integer); */ type = ZFS_DIRENT_TYPE(zap.za_first_integer); } reclen = DIRENT64_RECLEN(strlen(zap.za_name)); /* * Will this entry fit in the buffer? */ if (outcount + reclen > bufsize) { /* * Did we manage to fit anything in the buffer? */ if (!outcount) { error = SET_ERROR(EINVAL); goto update; } break; } /* * Add normal entry: */ odp->d_ino = objnum; odp->d_reclen = reclen; odp->d_namlen = strlen(zap.za_name); /* NOTE: d_off is the offset for the *next* entry. */ next = &odp->d_off; strlcpy(odp->d_name, zap.za_name, odp->d_namlen + 1); odp->d_type = type; dirent_terminate(odp); odp = (dirent64_t *)((intptr_t)odp + reclen); outcount += reclen; ASSERT3S(outcount, <=, bufsize); if (prefetch) dmu_prefetch_dnode(os, objnum, ZIO_PRIORITY_SYNC_READ); /* * Move to the next entry, fill in the previous offset. */ if (offset > 2 || (offset == 2 && !zfs_show_ctldir(zp))) { zap_cursor_advance(&zc); offset = zap_cursor_serialize(&zc); } else { offset += 1; } /* Fill the offset right after advancing the cursor. */ if (next != NULL) *next = offset; if (cooks != NULL) { *cooks++ = offset; ncooks--; KASSERT(ncooks >= 0, ("ncookies=%d", ncooks)); } } zp->z_zn_prefetch = B_FALSE; /* a lookup will re-enable pre-fetching */ /* Subtract unused cookies */ if (ncookies != NULL) *ncookies -= ncooks; if (zfs_uio_segflg(uio) == UIO_SYSSPACE && zfs_uio_iovcnt(uio) == 1) { iovp->iov_base += outcount; iovp->iov_len -= outcount; zfs_uio_resid(uio) -= outcount; } else if ((error = zfs_uiomove(outbuf, (long)outcount, UIO_READ, uio))) { /* * Reset the pointer. */ offset = zfs_uio_offset(uio); } update: zap_cursor_fini(&zc); if (zfs_uio_segflg(uio) != UIO_SYSSPACE || zfs_uio_iovcnt(uio) != 1) kmem_free(outbuf, bufsize); if (error == ENOENT) error = 0; ZFS_ACCESSTIME_STAMP(zfsvfs, zp); zfs_uio_setoffset(uio, offset); zfs_exit(zfsvfs, FTAG); if (error != 0 && cookies != NULL) { free(*cookies, M_TEMP); *cookies = NULL; *ncookies = 0; } return (error); } /* * Get the requested file attributes and place them in the provided * vattr structure. * * IN: vp - vnode of file. * vap - va_mask identifies requested attributes. * If AT_XVATTR set, then optional attrs are requested * flags - ATTR_NOACLCHECK (CIFS server context) * cr - credentials of caller. * * OUT: vap - attribute values. * * RETURN: 0 (always succeeds). */ static int zfs_getattr(vnode_t *vp, vattr_t *vap, int flags, cred_t *cr) { znode_t *zp = VTOZ(vp); zfsvfs_t *zfsvfs = zp->z_zfsvfs; int error = 0; uint32_t blksize; u_longlong_t nblocks; uint64_t mtime[2], ctime[2], crtime[2], rdev; xvattr_t *xvap = (xvattr_t *)vap; /* vap may be an xvattr_t * */ xoptattr_t *xoap = NULL; boolean_t skipaclchk = (flags & ATTR_NOACLCHECK) ? B_TRUE : B_FALSE; sa_bulk_attr_t bulk[4]; int count = 0; if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); zfs_fuid_map_ids(zp, cr, &vap->va_uid, &vap->va_gid); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs), NULL, &mtime, 16); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL, &ctime, 16); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CRTIME(zfsvfs), NULL, &crtime, 16); if (vp->v_type == VBLK || vp->v_type == VCHR) SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_RDEV(zfsvfs), NULL, &rdev, 8); if ((error = sa_bulk_lookup(zp->z_sa_hdl, bulk, count)) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } /* * If ACL is trivial don't bother looking for ACE_READ_ATTRIBUTES. * Also, if we are the owner don't bother, since owner should * always be allowed to read basic attributes of file. */ if (!(zp->z_pflags & ZFS_ACL_TRIVIAL) && (vap->va_uid != crgetuid(cr))) { if ((error = zfs_zaccess(zp, ACE_READ_ATTRIBUTES, 0, skipaclchk, cr, NULL))) { zfs_exit(zfsvfs, FTAG); return (error); } } /* * Return all attributes. It's cheaper to provide the answer * than to determine whether we were asked the question. */ vap->va_type = IFTOVT(zp->z_mode); vap->va_mode = zp->z_mode & ~S_IFMT; vn_fsid(vp, vap); vap->va_nodeid = zp->z_id; vap->va_nlink = zp->z_links; if ((vp->v_flag & VROOT) && zfs_show_ctldir(zp) && zp->z_links < ZFS_LINK_MAX) vap->va_nlink++; vap->va_size = zp->z_size; if (vp->v_type == VBLK || vp->v_type == VCHR) vap->va_rdev = zfs_cmpldev(rdev); vap->va_gen = zp->z_gen; vap->va_flags = 0; /* FreeBSD: Reset chflags(2) flags. */ vap->va_filerev = zp->z_seq; /* * Add in any requested optional attributes and the create time. * Also set the corresponding bits in the returned attribute bitmap. */ if ((xoap = xva_getxoptattr(xvap)) != NULL && zfsvfs->z_use_fuids) { if (XVA_ISSET_REQ(xvap, XAT_ARCHIVE)) { xoap->xoa_archive = ((zp->z_pflags & ZFS_ARCHIVE) != 0); XVA_SET_RTN(xvap, XAT_ARCHIVE); } if (XVA_ISSET_REQ(xvap, XAT_READONLY)) { xoap->xoa_readonly = ((zp->z_pflags & ZFS_READONLY) != 0); XVA_SET_RTN(xvap, XAT_READONLY); } if (XVA_ISSET_REQ(xvap, XAT_SYSTEM)) { xoap->xoa_system = ((zp->z_pflags & ZFS_SYSTEM) != 0); XVA_SET_RTN(xvap, XAT_SYSTEM); } if (XVA_ISSET_REQ(xvap, XAT_HIDDEN)) { xoap->xoa_hidden = ((zp->z_pflags & ZFS_HIDDEN) != 0); XVA_SET_RTN(xvap, XAT_HIDDEN); } if (XVA_ISSET_REQ(xvap, XAT_NOUNLINK)) { xoap->xoa_nounlink = ((zp->z_pflags & ZFS_NOUNLINK) != 0); XVA_SET_RTN(xvap, XAT_NOUNLINK); } if (XVA_ISSET_REQ(xvap, XAT_IMMUTABLE)) { xoap->xoa_immutable = ((zp->z_pflags & ZFS_IMMUTABLE) != 0); XVA_SET_RTN(xvap, XAT_IMMUTABLE); } if (XVA_ISSET_REQ(xvap, XAT_APPENDONLY)) { xoap->xoa_appendonly = ((zp->z_pflags & ZFS_APPENDONLY) != 0); XVA_SET_RTN(xvap, XAT_APPENDONLY); } if (XVA_ISSET_REQ(xvap, XAT_NODUMP)) { xoap->xoa_nodump = ((zp->z_pflags & ZFS_NODUMP) != 0); XVA_SET_RTN(xvap, XAT_NODUMP); } if (XVA_ISSET_REQ(xvap, XAT_OPAQUE)) { xoap->xoa_opaque = ((zp->z_pflags & ZFS_OPAQUE) != 0); XVA_SET_RTN(xvap, XAT_OPAQUE); } if (XVA_ISSET_REQ(xvap, XAT_AV_QUARANTINED)) { xoap->xoa_av_quarantined = ((zp->z_pflags & ZFS_AV_QUARANTINED) != 0); XVA_SET_RTN(xvap, XAT_AV_QUARANTINED); } if (XVA_ISSET_REQ(xvap, XAT_AV_MODIFIED)) { xoap->xoa_av_modified = ((zp->z_pflags & ZFS_AV_MODIFIED) != 0); XVA_SET_RTN(xvap, XAT_AV_MODIFIED); } if (XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP) && vp->v_type == VREG) { zfs_sa_get_scanstamp(zp, xvap); } if (XVA_ISSET_REQ(xvap, XAT_REPARSE)) { xoap->xoa_reparse = ((zp->z_pflags & ZFS_REPARSE) != 0); XVA_SET_RTN(xvap, XAT_REPARSE); } if (XVA_ISSET_REQ(xvap, XAT_GEN)) { xoap->xoa_generation = zp->z_gen; XVA_SET_RTN(xvap, XAT_GEN); } if (XVA_ISSET_REQ(xvap, XAT_OFFLINE)) { xoap->xoa_offline = ((zp->z_pflags & ZFS_OFFLINE) != 0); XVA_SET_RTN(xvap, XAT_OFFLINE); } if (XVA_ISSET_REQ(xvap, XAT_SPARSE)) { xoap->xoa_sparse = ((zp->z_pflags & ZFS_SPARSE) != 0); XVA_SET_RTN(xvap, XAT_SPARSE); } if (XVA_ISSET_REQ(xvap, XAT_PROJINHERIT)) { xoap->xoa_projinherit = ((zp->z_pflags & ZFS_PROJINHERIT) != 0); XVA_SET_RTN(xvap, XAT_PROJINHERIT); } if (XVA_ISSET_REQ(xvap, XAT_PROJID)) { xoap->xoa_projid = zp->z_projid; XVA_SET_RTN(xvap, XAT_PROJID); } } ZFS_TIME_DECODE(&vap->va_atime, zp->z_atime); ZFS_TIME_DECODE(&vap->va_mtime, mtime); ZFS_TIME_DECODE(&vap->va_ctime, ctime); ZFS_TIME_DECODE(&vap->va_birthtime, crtime); sa_object_size(zp->z_sa_hdl, &blksize, &nblocks); vap->va_blksize = blksize; vap->va_bytes = nblocks << 9; /* nblocks * 512 */ if (zp->z_blksz == 0) { /* * Block size hasn't been set; suggest maximal I/O transfers. */ vap->va_blksize = zfsvfs->z_max_blksz; } zfs_exit(zfsvfs, FTAG); return (0); } /* * Set the file attributes to the values contained in the * vattr structure. * * IN: zp - znode of file to be modified. * vap - new attribute values. * If AT_XVATTR set, then optional attrs are being set * flags - ATTR_UTIME set if non-default time values provided. * - ATTR_NOACLCHECK (CIFS context only). * cr - credentials of caller. * mnt_ns - Unused on FreeBSD * * RETURN: 0 on success, error code on failure. * * Timestamps: * vp - ctime updated, mtime updated if size changed. */ int zfs_setattr(znode_t *zp, vattr_t *vap, int flags, cred_t *cr, zidmap_t *mnt_ns) { vnode_t *vp = ZTOV(zp); zfsvfs_t *zfsvfs = zp->z_zfsvfs; objset_t *os; zilog_t *zilog; dmu_tx_t *tx; vattr_t oldva; xvattr_t tmpxvattr; uint_t mask = vap->va_mask; uint_t saved_mask = 0; uint64_t saved_mode; int trim_mask = 0; uint64_t new_mode; uint64_t new_uid, new_gid; uint64_t xattr_obj; uint64_t mtime[2], ctime[2]; uint64_t projid = ZFS_INVALID_PROJID; znode_t *attrzp; int need_policy = FALSE; int err, err2; zfs_fuid_info_t *fuidp = NULL; xvattr_t *xvap = (xvattr_t *)vap; /* vap may be an xvattr_t * */ xoptattr_t *xoap; zfs_acl_t *aclp; boolean_t skipaclchk = (flags & ATTR_NOACLCHECK) ? B_TRUE : B_FALSE; boolean_t fuid_dirtied = B_FALSE; sa_bulk_attr_t bulk[7], xattr_bulk[7]; int count = 0, xattr_count = 0; if (mask == 0) return (0); if (mask & AT_NOSET) return (SET_ERROR(EINVAL)); if ((err = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (err); os = zfsvfs->z_os; zilog = zfsvfs->z_log; /* * Make sure that if we have ephemeral uid/gid or xvattr specified * that file system is at proper version level */ if (zfsvfs->z_use_fuids == B_FALSE && (((mask & AT_UID) && IS_EPHEMERAL(vap->va_uid)) || ((mask & AT_GID) && IS_EPHEMERAL(vap->va_gid)) || (mask & AT_XVATTR))) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EINVAL)); } if (mask & AT_SIZE && vp->v_type == VDIR) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EISDIR)); } if (mask & AT_SIZE && vp->v_type != VREG && vp->v_type != VFIFO) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EINVAL)); } /* * If this is an xvattr_t, then get a pointer to the structure of * optional attributes. If this is NULL, then we have a vattr_t. */ xoap = xva_getxoptattr(xvap); xva_init(&tmpxvattr); /* * Immutable files can only alter immutable bit and atime */ if ((zp->z_pflags & ZFS_IMMUTABLE) && ((mask & (AT_SIZE|AT_UID|AT_GID|AT_MTIME|AT_MODE)) || ((mask & AT_XVATTR) && XVA_ISSET_REQ(xvap, XAT_CREATETIME)))) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EPERM)); } /* * Note: ZFS_READONLY is handled in zfs_zaccess_common. */ /* * Verify timestamps doesn't overflow 32 bits. * ZFS can handle large timestamps, but 32bit syscalls can't * handle times greater than 2039. This check should be removed * once large timestamps are fully supported. */ if (mask & (AT_ATIME | AT_MTIME)) { if (((mask & AT_ATIME) && TIMESPEC_OVERFLOW(&vap->va_atime)) || ((mask & AT_MTIME) && TIMESPEC_OVERFLOW(&vap->va_mtime))) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EOVERFLOW)); } } if (xoap != NULL && (mask & AT_XVATTR)) { if (XVA_ISSET_REQ(xvap, XAT_CREATETIME) && TIMESPEC_OVERFLOW(&vap->va_birthtime)) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EOVERFLOW)); } if (XVA_ISSET_REQ(xvap, XAT_PROJID)) { if (!dmu_objset_projectquota_enabled(os) || (!S_ISREG(zp->z_mode) && !S_ISDIR(zp->z_mode))) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EOPNOTSUPP)); } projid = xoap->xoa_projid; if (unlikely(projid == ZFS_INVALID_PROJID)) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EINVAL)); } if (projid == zp->z_projid && zp->z_pflags & ZFS_PROJID) projid = ZFS_INVALID_PROJID; else need_policy = TRUE; } if (XVA_ISSET_REQ(xvap, XAT_PROJINHERIT) && (xoap->xoa_projinherit != ((zp->z_pflags & ZFS_PROJINHERIT) != 0)) && (!dmu_objset_projectquota_enabled(os) || (!S_ISREG(zp->z_mode) && !S_ISDIR(zp->z_mode)))) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EOPNOTSUPP)); } } attrzp = NULL; aclp = NULL; if (zfsvfs->z_vfs->vfs_flag & VFS_RDONLY) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EROFS)); } /* * First validate permissions */ if (mask & AT_SIZE) { /* * XXX - Note, we are not providing any open * mode flags here (like FNDELAY), so we may * block if there are locks present... this * should be addressed in openat(). */ /* XXX - would it be OK to generate a log record here? */ err = zfs_freesp(zp, vap->va_size, 0, 0, FALSE); if (err) { zfs_exit(zfsvfs, FTAG); return (err); } } if (mask & (AT_ATIME|AT_MTIME) || ((mask & AT_XVATTR) && (XVA_ISSET_REQ(xvap, XAT_HIDDEN) || XVA_ISSET_REQ(xvap, XAT_READONLY) || XVA_ISSET_REQ(xvap, XAT_ARCHIVE) || XVA_ISSET_REQ(xvap, XAT_OFFLINE) || XVA_ISSET_REQ(xvap, XAT_SPARSE) || XVA_ISSET_REQ(xvap, XAT_CREATETIME) || XVA_ISSET_REQ(xvap, XAT_SYSTEM)))) { need_policy = zfs_zaccess(zp, ACE_WRITE_ATTRIBUTES, 0, skipaclchk, cr, mnt_ns); } if (mask & (AT_UID|AT_GID)) { int idmask = (mask & (AT_UID|AT_GID)); int take_owner; int take_group; /* * NOTE: even if a new mode is being set, * we may clear S_ISUID/S_ISGID bits. */ if (!(mask & AT_MODE)) vap->va_mode = zp->z_mode; /* * Take ownership or chgrp to group we are a member of */ take_owner = (mask & AT_UID) && (vap->va_uid == crgetuid(cr)); take_group = (mask & AT_GID) && zfs_groupmember(zfsvfs, vap->va_gid, cr); /* * If both AT_UID and AT_GID are set then take_owner and * take_group must both be set in order to allow taking * ownership. * * Otherwise, send the check through secpolicy_vnode_setattr() * */ if (((idmask == (AT_UID|AT_GID)) && take_owner && take_group) || ((idmask == AT_UID) && take_owner) || ((idmask == AT_GID) && take_group)) { if (zfs_zaccess(zp, ACE_WRITE_OWNER, 0, skipaclchk, cr, mnt_ns) == 0) { /* * Remove setuid/setgid for non-privileged users */ secpolicy_setid_clear(vap, vp, cr); trim_mask = (mask & (AT_UID|AT_GID)); } else { need_policy = TRUE; } } else { need_policy = TRUE; } } oldva.va_mode = zp->z_mode; zfs_fuid_map_ids(zp, cr, &oldva.va_uid, &oldva.va_gid); if (mask & AT_XVATTR) { /* * Update xvattr mask to include only those attributes * that are actually changing. * * the bits will be restored prior to actually setting * the attributes so the caller thinks they were set. */ if (XVA_ISSET_REQ(xvap, XAT_APPENDONLY)) { if (xoap->xoa_appendonly != ((zp->z_pflags & ZFS_APPENDONLY) != 0)) { need_policy = TRUE; } else { XVA_CLR_REQ(xvap, XAT_APPENDONLY); XVA_SET_REQ(&tmpxvattr, XAT_APPENDONLY); } } if (XVA_ISSET_REQ(xvap, XAT_PROJINHERIT)) { if (xoap->xoa_projinherit != ((zp->z_pflags & ZFS_PROJINHERIT) != 0)) { need_policy = TRUE; } else { XVA_CLR_REQ(xvap, XAT_PROJINHERIT); XVA_SET_REQ(&tmpxvattr, XAT_PROJINHERIT); } } if (XVA_ISSET_REQ(xvap, XAT_NOUNLINK)) { if (xoap->xoa_nounlink != ((zp->z_pflags & ZFS_NOUNLINK) != 0)) { need_policy = TRUE; } else { XVA_CLR_REQ(xvap, XAT_NOUNLINK); XVA_SET_REQ(&tmpxvattr, XAT_NOUNLINK); } } if (XVA_ISSET_REQ(xvap, XAT_IMMUTABLE)) { if (xoap->xoa_immutable != ((zp->z_pflags & ZFS_IMMUTABLE) != 0)) { need_policy = TRUE; } else { XVA_CLR_REQ(xvap, XAT_IMMUTABLE); XVA_SET_REQ(&tmpxvattr, XAT_IMMUTABLE); } } if (XVA_ISSET_REQ(xvap, XAT_NODUMP)) { if (xoap->xoa_nodump != ((zp->z_pflags & ZFS_NODUMP) != 0)) { need_policy = TRUE; } else { XVA_CLR_REQ(xvap, XAT_NODUMP); XVA_SET_REQ(&tmpxvattr, XAT_NODUMP); } } if (XVA_ISSET_REQ(xvap, XAT_AV_MODIFIED)) { if (xoap->xoa_av_modified != ((zp->z_pflags & ZFS_AV_MODIFIED) != 0)) { need_policy = TRUE; } else { XVA_CLR_REQ(xvap, XAT_AV_MODIFIED); XVA_SET_REQ(&tmpxvattr, XAT_AV_MODIFIED); } } if (XVA_ISSET_REQ(xvap, XAT_AV_QUARANTINED)) { if ((vp->v_type != VREG && xoap->xoa_av_quarantined) || xoap->xoa_av_quarantined != ((zp->z_pflags & ZFS_AV_QUARANTINED) != 0)) { need_policy = TRUE; } else { XVA_CLR_REQ(xvap, XAT_AV_QUARANTINED); XVA_SET_REQ(&tmpxvattr, XAT_AV_QUARANTINED); } } if (XVA_ISSET_REQ(xvap, XAT_REPARSE)) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EPERM)); } if (need_policy == FALSE && (XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP) || XVA_ISSET_REQ(xvap, XAT_OPAQUE))) { need_policy = TRUE; } } if (mask & AT_MODE) { if (zfs_zaccess(zp, ACE_WRITE_ACL, 0, skipaclchk, cr, mnt_ns) == 0) { err = secpolicy_setid_setsticky_clear(vp, vap, &oldva, cr); if (err) { zfs_exit(zfsvfs, FTAG); return (err); } trim_mask |= AT_MODE; } else { need_policy = TRUE; } } if (need_policy) { /* * If trim_mask is set then take ownership * has been granted or write_acl is present and user * has the ability to modify mode. In that case remove * UID|GID and or MODE from mask so that * secpolicy_vnode_setattr() doesn't revoke it. */ if (trim_mask) { saved_mask = vap->va_mask; vap->va_mask &= ~trim_mask; if (trim_mask & AT_MODE) { /* * Save the mode, as secpolicy_vnode_setattr() * will overwrite it with ova.va_mode. */ saved_mode = vap->va_mode; } } err = secpolicy_vnode_setattr(cr, vp, vap, &oldva, flags, (int (*)(void *, int, cred_t *))zfs_zaccess_unix, zp); if (err) { zfs_exit(zfsvfs, FTAG); return (err); } if (trim_mask) { vap->va_mask |= saved_mask; if (trim_mask & AT_MODE) { /* * Recover the mode after * secpolicy_vnode_setattr(). */ vap->va_mode = saved_mode; } } } /* * secpolicy_vnode_setattr, or take ownership may have * changed va_mask */ mask = vap->va_mask; if ((mask & (AT_UID | AT_GID)) || projid != ZFS_INVALID_PROJID) { err = sa_lookup(zp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs), &xattr_obj, sizeof (xattr_obj)); if (err == 0 && xattr_obj) { err = zfs_zget(zp->z_zfsvfs, xattr_obj, &attrzp); if (err == 0) { err = vn_lock(ZTOV(attrzp), LK_EXCLUSIVE); if (err != 0) vrele(ZTOV(attrzp)); } if (err) goto out2; } if (mask & AT_UID) { new_uid = zfs_fuid_create(zfsvfs, (uint64_t)vap->va_uid, cr, ZFS_OWNER, &fuidp); if (new_uid != zp->z_uid && zfs_id_overquota(zfsvfs, DMU_USERUSED_OBJECT, new_uid)) { if (attrzp) vput(ZTOV(attrzp)); err = SET_ERROR(EDQUOT); goto out2; } } if (mask & AT_GID) { new_gid = zfs_fuid_create(zfsvfs, (uint64_t)vap->va_gid, cr, ZFS_GROUP, &fuidp); if (new_gid != zp->z_gid && zfs_id_overquota(zfsvfs, DMU_GROUPUSED_OBJECT, new_gid)) { if (attrzp) vput(ZTOV(attrzp)); err = SET_ERROR(EDQUOT); goto out2; } } if (projid != ZFS_INVALID_PROJID && zfs_id_overquota(zfsvfs, DMU_PROJECTUSED_OBJECT, projid)) { if (attrzp) vput(ZTOV(attrzp)); err = SET_ERROR(EDQUOT); goto out2; } } tx = dmu_tx_create(os); if (mask & AT_MODE) { uint64_t pmode = zp->z_mode; uint64_t acl_obj; new_mode = (pmode & S_IFMT) | (vap->va_mode & ~S_IFMT); if (zp->z_zfsvfs->z_acl_mode == ZFS_ACL_RESTRICTED && !(zp->z_pflags & ZFS_ACL_TRIVIAL)) { err = SET_ERROR(EPERM); goto out; } if ((err = zfs_acl_chmod_setattr(zp, &aclp, new_mode))) goto out; if (!zp->z_is_sa && ((acl_obj = zfs_external_acl(zp)) != 0)) { /* * Are we upgrading ACL from old V0 format * to V1 format? */ if (zfsvfs->z_version >= ZPL_VERSION_FUID && zfs_znode_acl_version(zp) == ZFS_ACL_VERSION_INITIAL) { dmu_tx_hold_free(tx, acl_obj, 0, DMU_OBJECT_END); dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, aclp->z_acl_bytes); } else { dmu_tx_hold_write(tx, acl_obj, 0, aclp->z_acl_bytes); } } else if (!zp->z_is_sa && aclp->z_acl_bytes > ZFS_ACE_SPACE) { dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, aclp->z_acl_bytes); } dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE); } else { if (((mask & AT_XVATTR) && XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP)) || (projid != ZFS_INVALID_PROJID && !(zp->z_pflags & ZFS_PROJID))) dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE); else dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE); } if (attrzp) { dmu_tx_hold_sa(tx, attrzp->z_sa_hdl, B_FALSE); } fuid_dirtied = zfsvfs->z_fuid_dirty; if (fuid_dirtied) zfs_fuid_txhold(zfsvfs, tx); zfs_sa_upgrade_txholds(tx, zp); err = dmu_tx_assign(tx, TXG_WAIT); if (err) goto out; count = 0; /* * Set each attribute requested. * We group settings according to the locks they need to acquire. * * Note: you cannot set ctime directly, although it will be * updated as a side-effect of calling this function. */ if (projid != ZFS_INVALID_PROJID && !(zp->z_pflags & ZFS_PROJID)) { /* * For the existed object that is upgraded from old system, * its on-disk layout has no slot for the project ID attribute. * But quota accounting logic needs to access related slots by * offset directly. So we need to adjust old objects' layout * to make the project ID to some unified and fixed offset. */ if (attrzp) err = sa_add_projid(attrzp->z_sa_hdl, tx, projid); if (err == 0) err = sa_add_projid(zp->z_sa_hdl, tx, projid); if (unlikely(err == EEXIST)) err = 0; else if (err != 0) goto out; else projid = ZFS_INVALID_PROJID; } if (mask & (AT_UID|AT_GID|AT_MODE)) mutex_enter(&zp->z_acl_lock); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL, &zp->z_pflags, sizeof (zp->z_pflags)); if (attrzp) { if (mask & (AT_UID|AT_GID|AT_MODE)) mutex_enter(&attrzp->z_acl_lock); SA_ADD_BULK_ATTR(xattr_bulk, xattr_count, SA_ZPL_FLAGS(zfsvfs), NULL, &attrzp->z_pflags, sizeof (attrzp->z_pflags)); if (projid != ZFS_INVALID_PROJID) { attrzp->z_projid = projid; SA_ADD_BULK_ATTR(xattr_bulk, xattr_count, SA_ZPL_PROJID(zfsvfs), NULL, &attrzp->z_projid, sizeof (attrzp->z_projid)); } } if (mask & (AT_UID|AT_GID)) { if (mask & AT_UID) { SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_UID(zfsvfs), NULL, &new_uid, sizeof (new_uid)); zp->z_uid = new_uid; if (attrzp) { SA_ADD_BULK_ATTR(xattr_bulk, xattr_count, SA_ZPL_UID(zfsvfs), NULL, &new_uid, sizeof (new_uid)); attrzp->z_uid = new_uid; } } if (mask & AT_GID) { SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_GID(zfsvfs), NULL, &new_gid, sizeof (new_gid)); zp->z_gid = new_gid; if (attrzp) { SA_ADD_BULK_ATTR(xattr_bulk, xattr_count, SA_ZPL_GID(zfsvfs), NULL, &new_gid, sizeof (new_gid)); attrzp->z_gid = new_gid; } } if (!(mask & AT_MODE)) { SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MODE(zfsvfs), NULL, &new_mode, sizeof (new_mode)); new_mode = zp->z_mode; } err = zfs_acl_chown_setattr(zp); ASSERT0(err); if (attrzp) { vn_seqc_write_begin(ZTOV(attrzp)); err = zfs_acl_chown_setattr(attrzp); vn_seqc_write_end(ZTOV(attrzp)); ASSERT0(err); } } if (mask & AT_MODE) { SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MODE(zfsvfs), NULL, &new_mode, sizeof (new_mode)); zp->z_mode = new_mode; ASSERT3P(aclp, !=, NULL); err = zfs_aclset_common(zp, aclp, cr, tx); ASSERT0(err); if (zp->z_acl_cached) zfs_acl_free(zp->z_acl_cached); zp->z_acl_cached = aclp; aclp = NULL; } if (mask & AT_ATIME) { ZFS_TIME_ENCODE(&vap->va_atime, zp->z_atime); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_ATIME(zfsvfs), NULL, &zp->z_atime, sizeof (zp->z_atime)); } if (mask & AT_MTIME) { ZFS_TIME_ENCODE(&vap->va_mtime, mtime); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs), NULL, mtime, sizeof (mtime)); } if (projid != ZFS_INVALID_PROJID) { zp->z_projid = projid; SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_PROJID(zfsvfs), NULL, &zp->z_projid, sizeof (zp->z_projid)); } /* XXX - shouldn't this be done *before* the ATIME/MTIME checks? */ if (mask & AT_SIZE && !(mask & AT_MTIME)) { SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs), NULL, mtime, sizeof (mtime)); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL, &ctime, sizeof (ctime)); zfs_tstamp_update_setup(zp, CONTENT_MODIFIED, mtime, ctime); } else if (mask != 0) { SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL, &ctime, sizeof (ctime)); zfs_tstamp_update_setup(zp, STATE_CHANGED, mtime, ctime); if (attrzp) { SA_ADD_BULK_ATTR(xattr_bulk, xattr_count, SA_ZPL_CTIME(zfsvfs), NULL, &ctime, sizeof (ctime)); zfs_tstamp_update_setup(attrzp, STATE_CHANGED, mtime, ctime); } } /* * Do this after setting timestamps to prevent timestamp * update from toggling bit */ if (xoap && (mask & AT_XVATTR)) { if (XVA_ISSET_REQ(xvap, XAT_CREATETIME)) xoap->xoa_createtime = vap->va_birthtime; /* * restore trimmed off masks * so that return masks can be set for caller. */ if (XVA_ISSET_REQ(&tmpxvattr, XAT_APPENDONLY)) { XVA_SET_REQ(xvap, XAT_APPENDONLY); } if (XVA_ISSET_REQ(&tmpxvattr, XAT_NOUNLINK)) { XVA_SET_REQ(xvap, XAT_NOUNLINK); } if (XVA_ISSET_REQ(&tmpxvattr, XAT_IMMUTABLE)) { XVA_SET_REQ(xvap, XAT_IMMUTABLE); } if (XVA_ISSET_REQ(&tmpxvattr, XAT_NODUMP)) { XVA_SET_REQ(xvap, XAT_NODUMP); } if (XVA_ISSET_REQ(&tmpxvattr, XAT_AV_MODIFIED)) { XVA_SET_REQ(xvap, XAT_AV_MODIFIED); } if (XVA_ISSET_REQ(&tmpxvattr, XAT_AV_QUARANTINED)) { XVA_SET_REQ(xvap, XAT_AV_QUARANTINED); } if (XVA_ISSET_REQ(&tmpxvattr, XAT_PROJINHERIT)) { XVA_SET_REQ(xvap, XAT_PROJINHERIT); } if (XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP)) ASSERT3S(vp->v_type, ==, VREG); zfs_xvattr_set(zp, xvap, tx); } if (fuid_dirtied) zfs_fuid_sync(zfsvfs, tx); if (mask != 0) zfs_log_setattr(zilog, tx, TX_SETATTR, zp, vap, mask, fuidp); if (mask & (AT_UID|AT_GID|AT_MODE)) mutex_exit(&zp->z_acl_lock); if (attrzp) { if (mask & (AT_UID|AT_GID|AT_MODE)) mutex_exit(&attrzp->z_acl_lock); } out: if (err == 0 && attrzp) { err2 = sa_bulk_update(attrzp->z_sa_hdl, xattr_bulk, xattr_count, tx); ASSERT0(err2); } if (attrzp) vput(ZTOV(attrzp)); if (aclp) zfs_acl_free(aclp); if (fuidp) { zfs_fuid_info_free(fuidp); fuidp = NULL; } if (err) { dmu_tx_abort(tx); } else { err2 = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx); dmu_tx_commit(tx); } out2: if (os->os_sync == ZFS_SYNC_ALWAYS) zil_commit(zilog, 0); zfs_exit(zfsvfs, FTAG); return (err); } /* * Look up the directory entries corresponding to the source and target * directory/name pairs. */ static int zfs_rename_relock_lookup(znode_t *sdzp, const struct componentname *scnp, znode_t **szpp, znode_t *tdzp, const struct componentname *tcnp, znode_t **tzpp) { zfsvfs_t *zfsvfs; znode_t *szp, *tzp; int error; /* * Before using sdzp and tdzp we must ensure that they are live. * As a porting legacy from illumos we have two things to worry * about. One is typical for FreeBSD and it is that the vnode is * not reclaimed (doomed). The other is that the znode is live. * The current code can invalidate the znode without acquiring the * corresponding vnode lock if the object represented by the znode * and vnode is no longer valid after a rollback or receive operation. * z_teardown_lock hidden behind zfs_enter and zfs_exit is the lock * that protects the znodes from the invalidation. */ zfsvfs = sdzp->z_zfsvfs; ASSERT3P(zfsvfs, ==, tdzp->z_zfsvfs); if ((error = zfs_enter_verify_zp(zfsvfs, sdzp, FTAG)) != 0) return (error); if ((error = zfs_verify_zp(tdzp)) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } /* * Re-resolve svp to be certain it still exists and fetch the * correct vnode. */ error = zfs_dirent_lookup(sdzp, scnp->cn_nameptr, &szp, ZEXISTS); if (error != 0) { /* Source entry invalid or not there. */ if ((scnp->cn_flags & ISDOTDOT) != 0 || (scnp->cn_namelen == 1 && scnp->cn_nameptr[0] == '.')) error = SET_ERROR(EINVAL); goto out; } *szpp = szp; /* * Re-resolve tvp, if it disappeared we just carry on. */ error = zfs_dirent_lookup(tdzp, tcnp->cn_nameptr, &tzp, 0); if (error != 0) { vrele(ZTOV(szp)); if ((tcnp->cn_flags & ISDOTDOT) != 0) error = SET_ERROR(EINVAL); goto out; } *tzpp = tzp; out: zfs_exit(zfsvfs, FTAG); return (error); } /* * We acquire all but fdvp locks using non-blocking acquisitions. If we * fail to acquire any lock in the path we will drop all held locks, * acquire the new lock in a blocking fashion, and then release it and * restart the rename. This acquire/release step ensures that we do not * spin on a lock waiting for release. On error release all vnode locks * and decrement references the way tmpfs_rename() would do. */ static int zfs_rename_relock(struct vnode *sdvp, struct vnode **svpp, struct vnode *tdvp, struct vnode **tvpp, const struct componentname *scnp, const struct componentname *tcnp) { struct vnode *nvp, *svp, *tvp; znode_t *sdzp, *tdzp, *szp, *tzp; int error; VOP_UNLOCK1(tdvp); if (*tvpp != NULL && *tvpp != tdvp) VOP_UNLOCK1(*tvpp); relock: error = vn_lock(sdvp, LK_EXCLUSIVE); if (error) goto out; error = vn_lock(tdvp, LK_EXCLUSIVE | LK_NOWAIT); if (error != 0) { VOP_UNLOCK1(sdvp); if (error != EBUSY) goto out; error = vn_lock(tdvp, LK_EXCLUSIVE); if (error) goto out; VOP_UNLOCK1(tdvp); goto relock; } tdzp = VTOZ(tdvp); sdzp = VTOZ(sdvp); error = zfs_rename_relock_lookup(sdzp, scnp, &szp, tdzp, tcnp, &tzp); if (error != 0) { VOP_UNLOCK1(sdvp); VOP_UNLOCK1(tdvp); goto out; } svp = ZTOV(szp); tvp = tzp != NULL ? ZTOV(tzp) : NULL; /* * Now try acquire locks on svp and tvp. */ nvp = svp; error = vn_lock(nvp, LK_EXCLUSIVE | LK_NOWAIT); if (error != 0) { VOP_UNLOCK1(sdvp); VOP_UNLOCK1(tdvp); if (tvp != NULL) vrele(tvp); if (error != EBUSY) { vrele(nvp); goto out; } error = vn_lock(nvp, LK_EXCLUSIVE); if (error != 0) { vrele(nvp); goto out; } VOP_UNLOCK1(nvp); /* * Concurrent rename race. * XXX ? */ if (nvp == tdvp) { vrele(nvp); error = SET_ERROR(EINVAL); goto out; } vrele(*svpp); *svpp = nvp; goto relock; } vrele(*svpp); *svpp = nvp; if (*tvpp != NULL) vrele(*tvpp); *tvpp = NULL; if (tvp != NULL) { nvp = tvp; error = vn_lock(nvp, LK_EXCLUSIVE | LK_NOWAIT); if (error != 0) { VOP_UNLOCK1(sdvp); VOP_UNLOCK1(tdvp); VOP_UNLOCK1(*svpp); if (error != EBUSY) { vrele(nvp); goto out; } error = vn_lock(nvp, LK_EXCLUSIVE); if (error != 0) { vrele(nvp); goto out; } vput(nvp); goto relock; } *tvpp = nvp; } return (0); out: return (error); } /* * Note that we must use VRELE_ASYNC in this function as it walks * up the directory tree and vrele may need to acquire an exclusive * lock if a last reference to a vnode is dropped. */ static int zfs_rename_check(znode_t *szp, znode_t *sdzp, znode_t *tdzp) { zfsvfs_t *zfsvfs; znode_t *zp, *zp1; uint64_t parent; int error; zfsvfs = tdzp->z_zfsvfs; if (tdzp == szp) return (SET_ERROR(EINVAL)); if (tdzp == sdzp) return (0); if (tdzp->z_id == zfsvfs->z_root) return (0); zp = tdzp; for (;;) { ASSERT(!zp->z_unlinked); if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_PARENT(zfsvfs), &parent, sizeof (parent))) != 0) break; if (parent == szp->z_id) { error = SET_ERROR(EINVAL); break; } if (parent == zfsvfs->z_root) break; if (parent == sdzp->z_id) break; error = zfs_zget(zfsvfs, parent, &zp1); if (error != 0) break; if (zp != tdzp) VN_RELE_ASYNC(ZTOV(zp), dsl_pool_zrele_taskq( dmu_objset_pool(zfsvfs->z_os))); zp = zp1; } if (error == ENOTDIR) panic("checkpath: .. not a directory\n"); if (zp != tdzp) VN_RELE_ASYNC(ZTOV(zp), dsl_pool_zrele_taskq(dmu_objset_pool(zfsvfs->z_os))); return (error); } #if __FreeBSD_version < 1300124 static void cache_vop_rename(struct vnode *fdvp, struct vnode *fvp, struct vnode *tdvp, struct vnode *tvp, struct componentname *fcnp, struct componentname *tcnp) { cache_purge(fvp); if (tvp != NULL) cache_purge(tvp); cache_purge_negative(tdvp); } #endif static int zfs_do_rename_impl(vnode_t *sdvp, vnode_t **svpp, struct componentname *scnp, vnode_t *tdvp, vnode_t **tvpp, struct componentname *tcnp, cred_t *cr); /* * Move an entry from the provided source directory to the target * directory. Change the entry name as indicated. * * IN: sdvp - Source directory containing the "old entry". * scnp - Old entry name. * tdvp - Target directory to contain the "new entry". * tcnp - New entry name. * cr - credentials of caller. * INOUT: svpp - Source file * tvpp - Target file, may point to NULL initially * * RETURN: 0 on success, error code on failure. * * Timestamps: * sdvp,tdvp - ctime|mtime updated */ static int zfs_do_rename(vnode_t *sdvp, vnode_t **svpp, struct componentname *scnp, vnode_t *tdvp, vnode_t **tvpp, struct componentname *tcnp, cred_t *cr) { int error; ASSERT_VOP_ELOCKED(tdvp, __func__); if (*tvpp != NULL) ASSERT_VOP_ELOCKED(*tvpp, __func__); /* Reject renames across filesystems. */ if ((*svpp)->v_mount != tdvp->v_mount || ((*tvpp) != NULL && (*svpp)->v_mount != (*tvpp)->v_mount)) { error = SET_ERROR(EXDEV); goto out; } if (zfsctl_is_node(tdvp)) { error = SET_ERROR(EXDEV); goto out; } /* * Lock all four vnodes to ensure safety and semantics of renaming. */ error = zfs_rename_relock(sdvp, svpp, tdvp, tvpp, scnp, tcnp); if (error != 0) { /* no vnodes are locked in the case of error here */ return (error); } error = zfs_do_rename_impl(sdvp, svpp, scnp, tdvp, tvpp, tcnp, cr); VOP_UNLOCK1(sdvp); VOP_UNLOCK1(*svpp); out: if (*tvpp != NULL) VOP_UNLOCK1(*tvpp); if (tdvp != *tvpp) VOP_UNLOCK1(tdvp); return (error); } static int zfs_do_rename_impl(vnode_t *sdvp, vnode_t **svpp, struct componentname *scnp, vnode_t *tdvp, vnode_t **tvpp, struct componentname *tcnp, cred_t *cr) { dmu_tx_t *tx; zfsvfs_t *zfsvfs; zilog_t *zilog; znode_t *tdzp, *sdzp, *tzp, *szp; const char *snm = scnp->cn_nameptr; const char *tnm = tcnp->cn_nameptr; int error; tdzp = VTOZ(tdvp); sdzp = VTOZ(sdvp); zfsvfs = tdzp->z_zfsvfs; if ((error = zfs_enter_verify_zp(zfsvfs, tdzp, FTAG)) != 0) return (error); if ((error = zfs_verify_zp(sdzp)) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } zilog = zfsvfs->z_log; if (zfsvfs->z_utf8 && u8_validate(tnm, strlen(tnm), NULL, U8_VALIDATE_ENTIRE, &error) < 0) { error = SET_ERROR(EILSEQ); goto out; } /* If source and target are the same file, there is nothing to do. */ if ((*svpp) == (*tvpp)) { error = 0; goto out; } if (((*svpp)->v_type == VDIR && (*svpp)->v_mountedhere != NULL) || ((*tvpp) != NULL && (*tvpp)->v_type == VDIR && (*tvpp)->v_mountedhere != NULL)) { error = SET_ERROR(EXDEV); goto out; } szp = VTOZ(*svpp); if ((error = zfs_verify_zp(szp)) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } tzp = *tvpp == NULL ? NULL : VTOZ(*tvpp); if (tzp != NULL) { if ((error = zfs_verify_zp(tzp)) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } } /* * This is to prevent the creation of links into attribute space * by renaming a linked file into/outof an attribute directory. * See the comment in zfs_link() for why this is considered bad. */ if ((tdzp->z_pflags & ZFS_XATTR) != (sdzp->z_pflags & ZFS_XATTR)) { error = SET_ERROR(EINVAL); goto out; } /* * If we are using project inheritance, means if the directory has * ZFS_PROJINHERIT set, then its descendant directories will inherit * not only the project ID, but also the ZFS_PROJINHERIT flag. Under * such case, we only allow renames into our tree when the project * IDs are the same. */ if (tdzp->z_pflags & ZFS_PROJINHERIT && tdzp->z_projid != szp->z_projid) { error = SET_ERROR(EXDEV); goto out; } /* * Must have write access at the source to remove the old entry * and write access at the target to create the new entry. * Note that if target and source are the same, this can be * done in a single check. */ if ((error = zfs_zaccess_rename(sdzp, szp, tdzp, tzp, cr, NULL))) goto out; if ((*svpp)->v_type == VDIR) { /* * Avoid ".", "..", and aliases of "." for obvious reasons. */ if ((scnp->cn_namelen == 1 && scnp->cn_nameptr[0] == '.') || sdzp == szp || (scnp->cn_flags | tcnp->cn_flags) & ISDOTDOT) { error = EINVAL; goto out; } /* * Check to make sure rename is valid. * Can't do a move like this: /usr/a/b to /usr/a/b/c/d */ if ((error = zfs_rename_check(szp, sdzp, tdzp))) goto out; } /* * Does target exist? */ if (tzp) { /* * Source and target must be the same type. */ if ((*svpp)->v_type == VDIR) { if ((*tvpp)->v_type != VDIR) { error = SET_ERROR(ENOTDIR); goto out; } else { cache_purge(tdvp); if (sdvp != tdvp) cache_purge(sdvp); } } else { if ((*tvpp)->v_type == VDIR) { error = SET_ERROR(EISDIR); goto out; } } } vn_seqc_write_begin(*svpp); vn_seqc_write_begin(sdvp); if (*tvpp != NULL) vn_seqc_write_begin(*tvpp); if (tdvp != *tvpp) vn_seqc_write_begin(tdvp); vnevent_rename_src(*svpp, sdvp, scnp->cn_nameptr, ct); if (tzp) vnevent_rename_dest(*tvpp, tdvp, tnm, ct); /* * notify the target directory if it is not the same * as source directory. */ if (tdvp != sdvp) { vnevent_rename_dest_dir(tdvp, ct); } tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_sa(tx, szp->z_sa_hdl, B_FALSE); dmu_tx_hold_sa(tx, sdzp->z_sa_hdl, B_FALSE); dmu_tx_hold_zap(tx, sdzp->z_id, FALSE, snm); dmu_tx_hold_zap(tx, tdzp->z_id, TRUE, tnm); if (sdzp != tdzp) { dmu_tx_hold_sa(tx, tdzp->z_sa_hdl, B_FALSE); zfs_sa_upgrade_txholds(tx, tdzp); } if (tzp) { dmu_tx_hold_sa(tx, tzp->z_sa_hdl, B_FALSE); zfs_sa_upgrade_txholds(tx, tzp); } zfs_sa_upgrade_txholds(tx, szp); dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL); error = dmu_tx_assign(tx, TXG_WAIT); if (error) { dmu_tx_abort(tx); goto out_seq; } if (tzp) /* Attempt to remove the existing target */ error = zfs_link_destroy(tdzp, tnm, tzp, tx, 0, NULL); if (error == 0) { error = zfs_link_create(tdzp, tnm, szp, tx, ZRENAMING); if (error == 0) { szp->z_pflags |= ZFS_AV_MODIFIED; error = sa_update(szp->z_sa_hdl, SA_ZPL_FLAGS(zfsvfs), (void *)&szp->z_pflags, sizeof (uint64_t), tx); ASSERT0(error); error = zfs_link_destroy(sdzp, snm, szp, tx, ZRENAMING, NULL); if (error == 0) { zfs_log_rename(zilog, tx, TX_RENAME, sdzp, snm, tdzp, tnm, szp); } else { /* * At this point, we have successfully created * the target name, but have failed to remove * the source name. Since the create was done * with the ZRENAMING flag, there are * complications; for one, the link count is * wrong. The easiest way to deal with this * is to remove the newly created target, and * return the original error. This must * succeed; fortunately, it is very unlikely to * fail, since we just created it. */ VERIFY0(zfs_link_destroy(tdzp, tnm, szp, tx, ZRENAMING, NULL)); } } if (error == 0) { cache_vop_rename(sdvp, *svpp, tdvp, *tvpp, scnp, tcnp); } } dmu_tx_commit(tx); out_seq: vn_seqc_write_end(*svpp); vn_seqc_write_end(sdvp); if (*tvpp != NULL) vn_seqc_write_end(*tvpp); if (tdvp != *tvpp) vn_seqc_write_end(tdvp); out: if (error == 0 && zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS) zil_commit(zilog, 0); zfs_exit(zfsvfs, FTAG); return (error); } int zfs_rename(znode_t *sdzp, const char *sname, znode_t *tdzp, const char *tname, cred_t *cr, int flags, uint64_t rflags, vattr_t *wo_vap, zidmap_t *mnt_ns) { struct componentname scn, tcn; vnode_t *sdvp, *tdvp; vnode_t *svp, *tvp; int error; svp = tvp = NULL; if (rflags != 0 || wo_vap != NULL) return (SET_ERROR(EINVAL)); sdvp = ZTOV(sdzp); tdvp = ZTOV(tdzp); error = zfs_lookup_internal(sdzp, sname, &svp, &scn, DELETE); if (sdzp->z_zfsvfs->z_replay == B_FALSE) VOP_UNLOCK1(sdvp); if (error != 0) goto fail; VOP_UNLOCK1(svp); vn_lock(tdvp, LK_EXCLUSIVE | LK_RETRY); error = zfs_lookup_internal(tdzp, tname, &tvp, &tcn, RENAME); if (error == EJUSTRETURN) tvp = NULL; else if (error != 0) { VOP_UNLOCK1(tdvp); goto fail; } error = zfs_do_rename(sdvp, &svp, &scn, tdvp, &tvp, &tcn, cr); fail: if (svp != NULL) vrele(svp); if (tvp != NULL) vrele(tvp); return (error); } /* * Insert the indicated symbolic reference entry into the directory. * * IN: dvp - Directory to contain new symbolic link. * link - Name for new symlink entry. * vap - Attributes of new entry. * cr - credentials of caller. * ct - caller context * flags - case flags * mnt_ns - Unused on FreeBSD * * RETURN: 0 on success, error code on failure. * * Timestamps: * dvp - ctime|mtime updated */ int zfs_symlink(znode_t *dzp, const char *name, vattr_t *vap, const char *link, znode_t **zpp, cred_t *cr, int flags, zidmap_t *mnt_ns) { (void) flags; znode_t *zp; dmu_tx_t *tx; zfsvfs_t *zfsvfs = dzp->z_zfsvfs; zilog_t *zilog; uint64_t len = strlen(link); int error; zfs_acl_ids_t acl_ids; boolean_t fuid_dirtied; uint64_t txtype = TX_SYMLINK; ASSERT3S(vap->va_type, ==, VLNK); if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0) return (error); zilog = zfsvfs->z_log; if (zfsvfs->z_utf8 && u8_validate(name, strlen(name), NULL, U8_VALIDATE_ENTIRE, &error) < 0) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EILSEQ)); } if (len > MAXPATHLEN) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(ENAMETOOLONG)); } if ((error = zfs_acl_ids_create(dzp, 0, vap, cr, NULL, &acl_ids, NULL)) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } /* * Attempt to lock directory; fail if entry already exists. */ error = zfs_dirent_lookup(dzp, name, &zp, ZNEW); if (error) { zfs_acl_ids_free(&acl_ids); zfs_exit(zfsvfs, FTAG); return (error); } if ((error = zfs_zaccess(dzp, ACE_ADD_FILE, 0, B_FALSE, cr, mnt_ns))) { zfs_acl_ids_free(&acl_ids); zfs_exit(zfsvfs, FTAG); return (error); } if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, 0 /* projid */)) { zfs_acl_ids_free(&acl_ids); zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EDQUOT)); } getnewvnode_reserve_(); tx = dmu_tx_create(zfsvfs->z_os); fuid_dirtied = zfsvfs->z_fuid_dirty; dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, MAX(1, len)); dmu_tx_hold_zap(tx, dzp->z_id, TRUE, name); dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes + ZFS_SA_BASE_ATTR_SIZE + len); dmu_tx_hold_sa(tx, dzp->z_sa_hdl, B_FALSE); if (!zfsvfs->z_use_sa && acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) { dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, acl_ids.z_aclp->z_acl_bytes); } if (fuid_dirtied) zfs_fuid_txhold(zfsvfs, tx); error = dmu_tx_assign(tx, TXG_WAIT); if (error) { zfs_acl_ids_free(&acl_ids); dmu_tx_abort(tx); getnewvnode_drop_reserve(); zfs_exit(zfsvfs, FTAG); return (error); } /* * Create a new object for the symlink. * for version 4 ZPL datasets the symlink will be an SA attribute */ zfs_mknode(dzp, vap, tx, cr, 0, &zp, &acl_ids); if (fuid_dirtied) zfs_fuid_sync(zfsvfs, tx); if (zp->z_is_sa) error = sa_update(zp->z_sa_hdl, SA_ZPL_SYMLINK(zfsvfs), __DECONST(void *, link), len, tx); else zfs_sa_symlink(zp, __DECONST(char *, link), len, tx); zp->z_size = len; (void) sa_update(zp->z_sa_hdl, SA_ZPL_SIZE(zfsvfs), &zp->z_size, sizeof (zp->z_size), tx); /* * Insert the new object into the directory. */ (void) zfs_link_create(dzp, name, zp, tx, ZNEW); zfs_log_symlink(zilog, tx, txtype, dzp, zp, name, link); *zpp = zp; zfs_acl_ids_free(&acl_ids); dmu_tx_commit(tx); getnewvnode_drop_reserve(); if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS) zil_commit(zilog, 0); zfs_exit(zfsvfs, FTAG); return (error); } /* * Return, in the buffer contained in the provided uio structure, * the symbolic path referred to by vp. * * IN: vp - vnode of symbolic link. * uio - structure to contain the link path. * cr - credentials of caller. * ct - caller context * * OUT: uio - structure containing the link path. * * RETURN: 0 on success, error code on failure. * * Timestamps: * vp - atime updated */ static int zfs_readlink(vnode_t *vp, zfs_uio_t *uio, cred_t *cr, caller_context_t *ct) { (void) cr, (void) ct; znode_t *zp = VTOZ(vp); zfsvfs_t *zfsvfs = zp->z_zfsvfs; int error; if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); if (zp->z_is_sa) error = sa_lookup_uio(zp->z_sa_hdl, SA_ZPL_SYMLINK(zfsvfs), uio); else error = zfs_sa_readlink(zp, uio); ZFS_ACCESSTIME_STAMP(zfsvfs, zp); zfs_exit(zfsvfs, FTAG); return (error); } /* * Insert a new entry into directory tdvp referencing svp. * * IN: tdvp - Directory to contain new entry. * svp - vnode of new entry. * name - name of new entry. * cr - credentials of caller. * * RETURN: 0 on success, error code on failure. * * Timestamps: * tdvp - ctime|mtime updated * svp - ctime updated */ int zfs_link(znode_t *tdzp, znode_t *szp, const char *name, cred_t *cr, int flags) { (void) flags; znode_t *tzp; zfsvfs_t *zfsvfs = tdzp->z_zfsvfs; zilog_t *zilog; dmu_tx_t *tx; int error; uint64_t parent; uid_t owner; ASSERT3S(ZTOV(tdzp)->v_type, ==, VDIR); if ((error = zfs_enter_verify_zp(zfsvfs, tdzp, FTAG)) != 0) return (error); zilog = zfsvfs->z_log; /* * POSIX dictates that we return EPERM here. * Better choices include ENOTSUP or EISDIR. */ if (ZTOV(szp)->v_type == VDIR) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EPERM)); } if ((error = zfs_verify_zp(szp)) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } /* * If we are using project inheritance, means if the directory has * ZFS_PROJINHERIT set, then its descendant directories will inherit * not only the project ID, but also the ZFS_PROJINHERIT flag. Under * such case, we only allow hard link creation in our tree when the * project IDs are the same. */ if (tdzp->z_pflags & ZFS_PROJINHERIT && tdzp->z_projid != szp->z_projid) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EXDEV)); } if (szp->z_pflags & (ZFS_APPENDONLY | ZFS_IMMUTABLE | ZFS_READONLY)) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EPERM)); } /* Prevent links to .zfs/shares files */ if ((error = sa_lookup(szp->z_sa_hdl, SA_ZPL_PARENT(zfsvfs), &parent, sizeof (uint64_t))) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } if (parent == zfsvfs->z_shares_dir) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EPERM)); } if (zfsvfs->z_utf8 && u8_validate(name, strlen(name), NULL, U8_VALIDATE_ENTIRE, &error) < 0) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EILSEQ)); } /* * We do not support links between attributes and non-attributes * because of the potential security risk of creating links * into "normal" file space in order to circumvent restrictions * imposed in attribute space. */ if ((szp->z_pflags & ZFS_XATTR) != (tdzp->z_pflags & ZFS_XATTR)) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EINVAL)); } owner = zfs_fuid_map_id(zfsvfs, szp->z_uid, cr, ZFS_OWNER); if (owner != crgetuid(cr) && secpolicy_basic_link(ZTOV(szp), cr) != 0) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EPERM)); } if ((error = zfs_zaccess(tdzp, ACE_ADD_FILE, 0, B_FALSE, cr, NULL))) { zfs_exit(zfsvfs, FTAG); return (error); } /* * Attempt to lock directory; fail if entry already exists. */ error = zfs_dirent_lookup(tdzp, name, &tzp, ZNEW); if (error) { zfs_exit(zfsvfs, FTAG); return (error); } tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_sa(tx, szp->z_sa_hdl, B_FALSE); dmu_tx_hold_zap(tx, tdzp->z_id, TRUE, name); zfs_sa_upgrade_txholds(tx, szp); zfs_sa_upgrade_txholds(tx, tdzp); error = dmu_tx_assign(tx, TXG_WAIT); if (error) { dmu_tx_abort(tx); zfs_exit(zfsvfs, FTAG); return (error); } error = zfs_link_create(tdzp, name, szp, tx, 0); if (error == 0) { uint64_t txtype = TX_LINK; zfs_log_link(zilog, tx, txtype, tdzp, szp, name); } dmu_tx_commit(tx); if (error == 0) { vnevent_link(ZTOV(szp), ct); } if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS) zil_commit(zilog, 0); zfs_exit(zfsvfs, FTAG); return (error); } /* * Free or allocate space in a file. Currently, this function only * supports the `F_FREESP' command. However, this command is somewhat * misnamed, as its functionality includes the ability to allocate as * well as free space. * * IN: ip - inode of file to free data in. * cmd - action to take (only F_FREESP supported). * bfp - section of file to free/alloc. * flag - current file open mode flags. * offset - current file offset. * cr - credentials of caller. * * RETURN: 0 on success, error code on failure. * * Timestamps: * ip - ctime|mtime updated */ int zfs_space(znode_t *zp, int cmd, flock64_t *bfp, int flag, offset_t offset, cred_t *cr) { (void) offset; zfsvfs_t *zfsvfs = ZTOZSB(zp); uint64_t off, len; int error; if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); if (cmd != F_FREESP) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EINVAL)); } /* * Callers might not be able to detect properly that we are read-only, * so check it explicitly here. */ if (zfs_is_readonly(zfsvfs)) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EROFS)); } if (bfp->l_len < 0) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EINVAL)); } /* * Permissions aren't checked on Solaris because on this OS * zfs_space() can only be called with an opened file handle. * On Linux we can get here through truncate_range() which * operates directly on inodes, so we need to check access rights. */ if ((error = zfs_zaccess(zp, ACE_WRITE_DATA, 0, B_FALSE, cr, NULL))) { zfs_exit(zfsvfs, FTAG); return (error); } off = bfp->l_start; len = bfp->l_len; /* 0 means from off to end of file */ error = zfs_freesp(zp, off, len, flag, TRUE); zfs_exit(zfsvfs, FTAG); return (error); } static void zfs_inactive(vnode_t *vp, cred_t *cr, caller_context_t *ct) { (void) cr, (void) ct; znode_t *zp = VTOZ(vp); zfsvfs_t *zfsvfs = zp->z_zfsvfs; int error; ZFS_TEARDOWN_INACTIVE_ENTER_READ(zfsvfs); if (zp->z_sa_hdl == NULL) { /* * The fs has been unmounted, or we did a * suspend/resume and this file no longer exists. */ ZFS_TEARDOWN_INACTIVE_EXIT_READ(zfsvfs); vrecycle(vp); return; } if (zp->z_unlinked) { /* * Fast path to recycle a vnode of a removed file. */ ZFS_TEARDOWN_INACTIVE_EXIT_READ(zfsvfs); vrecycle(vp); return; } if (zp->z_atime_dirty && zp->z_unlinked == 0) { dmu_tx_t *tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE); zfs_sa_upgrade_txholds(tx, zp); error = dmu_tx_assign(tx, TXG_WAIT); if (error) { dmu_tx_abort(tx); } else { (void) sa_update(zp->z_sa_hdl, SA_ZPL_ATIME(zfsvfs), (void *)&zp->z_atime, sizeof (zp->z_atime), tx); zp->z_atime_dirty = 0; dmu_tx_commit(tx); } } ZFS_TEARDOWN_INACTIVE_EXIT_READ(zfsvfs); } _Static_assert(sizeof (struct zfid_short) <= sizeof (struct fid), "struct zfid_short bigger than struct fid"); _Static_assert(sizeof (struct zfid_long) <= sizeof (struct fid), "struct zfid_long bigger than struct fid"); static int zfs_fid(vnode_t *vp, fid_t *fidp, caller_context_t *ct) { (void) ct; znode_t *zp = VTOZ(vp); zfsvfs_t *zfsvfs = zp->z_zfsvfs; uint32_t gen; uint64_t gen64; uint64_t object = zp->z_id; zfid_short_t *zfid; int size, i, error; if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_GEN(zfsvfs), &gen64, sizeof (uint64_t))) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } gen = (uint32_t)gen64; size = (zfsvfs->z_parent != zfsvfs) ? LONG_FID_LEN : SHORT_FID_LEN; fidp->fid_len = size; zfid = (zfid_short_t *)fidp; zfid->zf_len = size; for (i = 0; i < sizeof (zfid->zf_object); i++) zfid->zf_object[i] = (uint8_t)(object >> (8 * i)); /* Must have a non-zero generation number to distinguish from .zfs */ if (gen == 0) gen = 1; for (i = 0; i < sizeof (zfid->zf_gen); i++) zfid->zf_gen[i] = (uint8_t)(gen >> (8 * i)); if (size == LONG_FID_LEN) { uint64_t objsetid = dmu_objset_id(zfsvfs->z_os); zfid_long_t *zlfid; zlfid = (zfid_long_t *)fidp; for (i = 0; i < sizeof (zlfid->zf_setid); i++) zlfid->zf_setid[i] = (uint8_t)(objsetid >> (8 * i)); /* XXX - this should be the generation number for the objset */ for (i = 0; i < sizeof (zlfid->zf_setgen); i++) zlfid->zf_setgen[i] = 0; } zfs_exit(zfsvfs, FTAG); return (0); } static int zfs_pathconf(vnode_t *vp, int cmd, ulong_t *valp, cred_t *cr, caller_context_t *ct) { znode_t *zp; zfsvfs_t *zfsvfs; int error; switch (cmd) { case _PC_LINK_MAX: *valp = MIN(LONG_MAX, ZFS_LINK_MAX); return (0); case _PC_FILESIZEBITS: *valp = 64; return (0); case _PC_MIN_HOLE_SIZE: *valp = (int)SPA_MINBLOCKSIZE; return (0); case _PC_ACL_EXTENDED: #if 0 /* POSIX ACLs are not implemented for ZFS on FreeBSD yet. */ zp = VTOZ(vp); zfsvfs = zp->z_zfsvfs; if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); *valp = zfsvfs->z_acl_type == ZFSACLTYPE_POSIX ? 1 : 0; zfs_exit(zfsvfs, FTAG); #else *valp = 0; #endif return (0); case _PC_ACL_NFS4: zp = VTOZ(vp); zfsvfs = zp->z_zfsvfs; if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); *valp = zfsvfs->z_acl_type == ZFS_ACLTYPE_NFSV4 ? 1 : 0; zfs_exit(zfsvfs, FTAG); return (0); case _PC_ACL_PATH_MAX: *valp = ACL_MAX_ENTRIES; return (0); default: return (EOPNOTSUPP); } } static int zfs_getpages(struct vnode *vp, vm_page_t *ma, int count, int *rbehind, int *rahead) { znode_t *zp = VTOZ(vp); zfsvfs_t *zfsvfs = zp->z_zfsvfs; zfs_locked_range_t *lr; vm_object_t object; off_t start, end, obj_size; uint_t blksz; int pgsin_b, pgsin_a; int error; if (zfs_enter_verify_zp(zfsvfs, zp, FTAG) != 0) return (zfs_vm_pagerret_error); start = IDX_TO_OFF(ma[0]->pindex); end = IDX_TO_OFF(ma[count - 1]->pindex + 1); /* * Lock a range covering all required and optional pages. * Note that we need to handle the case of the block size growing. */ for (;;) { blksz = zp->z_blksz; lr = zfs_rangelock_tryenter(&zp->z_rangelock, rounddown(start, blksz), roundup(end, blksz) - rounddown(start, blksz), RL_READER); if (lr == NULL) { if (rahead != NULL) { *rahead = 0; rahead = NULL; } if (rbehind != NULL) { *rbehind = 0; rbehind = NULL; } break; } if (blksz == zp->z_blksz) break; zfs_rangelock_exit(lr); } object = ma[0]->object; zfs_vmobject_wlock(object); obj_size = object->un_pager.vnp.vnp_size; zfs_vmobject_wunlock(object); if (IDX_TO_OFF(ma[count - 1]->pindex) >= obj_size) { if (lr != NULL) zfs_rangelock_exit(lr); zfs_exit(zfsvfs, FTAG); return (zfs_vm_pagerret_bad); } pgsin_b = 0; if (rbehind != NULL) { pgsin_b = OFF_TO_IDX(start - rounddown(start, blksz)); pgsin_b = MIN(*rbehind, pgsin_b); } pgsin_a = 0; if (rahead != NULL) { pgsin_a = OFF_TO_IDX(roundup(end, blksz) - end); if (end + IDX_TO_OFF(pgsin_a) >= obj_size) pgsin_a = OFF_TO_IDX(round_page(obj_size) - end); pgsin_a = MIN(*rahead, pgsin_a); } /* * NB: we need to pass the exact byte size of the data that we expect * to read after accounting for the file size. This is required because * ZFS will panic if we request DMU to read beyond the end of the last * allocated block. */ error = dmu_read_pages(zfsvfs->z_os, zp->z_id, ma, count, &pgsin_b, &pgsin_a, MIN(end, obj_size) - (end - PAGE_SIZE)); if (lr != NULL) zfs_rangelock_exit(lr); ZFS_ACCESSTIME_STAMP(zfsvfs, zp); dataset_kstats_update_read_kstats(&zfsvfs->z_kstat, count*PAGE_SIZE); zfs_exit(zfsvfs, FTAG); if (error != 0) return (zfs_vm_pagerret_error); VM_CNT_INC(v_vnodein); VM_CNT_ADD(v_vnodepgsin, count + pgsin_b + pgsin_a); if (rbehind != NULL) *rbehind = pgsin_b; if (rahead != NULL) *rahead = pgsin_a; return (zfs_vm_pagerret_ok); } #ifndef _SYS_SYSPROTO_H_ struct vop_getpages_args { struct vnode *a_vp; vm_page_t *a_m; int a_count; int *a_rbehind; int *a_rahead; }; #endif static int zfs_freebsd_getpages(struct vop_getpages_args *ap) { return (zfs_getpages(ap->a_vp, ap->a_m, ap->a_count, ap->a_rbehind, ap->a_rahead)); } static int zfs_putpages(struct vnode *vp, vm_page_t *ma, size_t len, int flags, int *rtvals) { znode_t *zp = VTOZ(vp); zfsvfs_t *zfsvfs = zp->z_zfsvfs; zfs_locked_range_t *lr; dmu_tx_t *tx; struct sf_buf *sf; vm_object_t object; vm_page_t m; caddr_t va; size_t tocopy; size_t lo_len; vm_ooffset_t lo_off; vm_ooffset_t off; uint_t blksz; int ncount; int pcount; int err; int i; object = vp->v_object; KASSERT(ma[0]->object == object, ("mismatching object")); KASSERT(len > 0 && (len & PAGE_MASK) == 0, ("unexpected length")); pcount = btoc(len); ncount = pcount; for (i = 0; i < pcount; i++) rtvals[i] = zfs_vm_pagerret_error; if (zfs_enter_verify_zp(zfsvfs, zp, FTAG) != 0) return (zfs_vm_pagerret_error); off = IDX_TO_OFF(ma[0]->pindex); blksz = zp->z_blksz; lo_off = rounddown(off, blksz); lo_len = roundup(len + (off - lo_off), blksz); lr = zfs_rangelock_enter(&zp->z_rangelock, lo_off, lo_len, RL_WRITER); zfs_vmobject_wlock(object); if (len + off > object->un_pager.vnp.vnp_size) { if (object->un_pager.vnp.vnp_size > off) { int pgoff; len = object->un_pager.vnp.vnp_size - off; ncount = btoc(len); if ((pgoff = (int)len & PAGE_MASK) != 0) { /* * If the object is locked and the following * conditions hold, then the page's dirty * field cannot be concurrently changed by a * pmap operation. */ m = ma[ncount - 1]; vm_page_assert_sbusied(m); KASSERT(!pmap_page_is_write_mapped(m), ("zfs_putpages: page %p is not read-only", m)); vm_page_clear_dirty(m, pgoff, PAGE_SIZE - pgoff); } } else { len = 0; ncount = 0; } if (ncount < pcount) { for (i = ncount; i < pcount; i++) { rtvals[i] = zfs_vm_pagerret_bad; } } } zfs_vmobject_wunlock(object); boolean_t commit = (flags & (zfs_vm_pagerput_sync | zfs_vm_pagerput_inval)) != 0 || zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS; if (ncount == 0) goto out; if (zfs_id_overblockquota(zfsvfs, DMU_USERUSED_OBJECT, zp->z_uid) || zfs_id_overblockquota(zfsvfs, DMU_GROUPUSED_OBJECT, zp->z_gid) || (zp->z_projid != ZFS_DEFAULT_PROJID && zfs_id_overblockquota(zfsvfs, DMU_PROJECTUSED_OBJECT, zp->z_projid))) { goto out; } tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_write(tx, zp->z_id, off, len); dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE); zfs_sa_upgrade_txholds(tx, zp); err = dmu_tx_assign(tx, TXG_WAIT); if (err != 0) { dmu_tx_abort(tx); goto out; } if (zp->z_blksz < PAGE_SIZE) { for (i = 0; len > 0; off += tocopy, len -= tocopy, i++) { tocopy = len > PAGE_SIZE ? PAGE_SIZE : len; va = zfs_map_page(ma[i], &sf); dmu_write(zfsvfs->z_os, zp->z_id, off, tocopy, va, tx); zfs_unmap_page(sf); } } else { err = dmu_write_pages(zfsvfs->z_os, zp->z_id, off, len, ma, tx); } if (err == 0) { uint64_t mtime[2], ctime[2]; sa_bulk_attr_t bulk[3]; int count = 0; SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs), NULL, &mtime, 16); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL, &ctime, 16); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL, &zp->z_pflags, 8); zfs_tstamp_update_setup(zp, CONTENT_MODIFIED, mtime, ctime); err = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx); ASSERT0(err); /* * XXX we should be passing a callback to undirty * but that would make the locking messier */ zfs_log_write(zfsvfs->z_log, tx, TX_WRITE, zp, off, len, commit, NULL, NULL); zfs_vmobject_wlock(object); for (i = 0; i < ncount; i++) { rtvals[i] = zfs_vm_pagerret_ok; vm_page_undirty(ma[i]); } zfs_vmobject_wunlock(object); VM_CNT_INC(v_vnodeout); VM_CNT_ADD(v_vnodepgsout, ncount); } dmu_tx_commit(tx); out: zfs_rangelock_exit(lr); if (commit) zil_commit(zfsvfs->z_log, zp->z_id); dataset_kstats_update_write_kstats(&zfsvfs->z_kstat, len); zfs_exit(zfsvfs, FTAG); return (rtvals[0]); } #ifndef _SYS_SYSPROTO_H_ struct vop_putpages_args { struct vnode *a_vp; vm_page_t *a_m; int a_count; int a_sync; int *a_rtvals; }; #endif static int zfs_freebsd_putpages(struct vop_putpages_args *ap) { return (zfs_putpages(ap->a_vp, ap->a_m, ap->a_count, ap->a_sync, ap->a_rtvals)); } #ifndef _SYS_SYSPROTO_H_ struct vop_bmap_args { struct vnode *a_vp; daddr_t a_bn; struct bufobj **a_bop; daddr_t *a_bnp; int *a_runp; int *a_runb; }; #endif static int zfs_freebsd_bmap(struct vop_bmap_args *ap) { if (ap->a_bop != NULL) *ap->a_bop = &ap->a_vp->v_bufobj; if (ap->a_bnp != NULL) *ap->a_bnp = ap->a_bn; if (ap->a_runp != NULL) *ap->a_runp = 0; if (ap->a_runb != NULL) *ap->a_runb = 0; return (0); } #ifndef _SYS_SYSPROTO_H_ struct vop_open_args { struct vnode *a_vp; int a_mode; struct ucred *a_cred; struct thread *a_td; }; #endif static int zfs_freebsd_open(struct vop_open_args *ap) { vnode_t *vp = ap->a_vp; znode_t *zp = VTOZ(vp); int error; error = zfs_open(&vp, ap->a_mode, ap->a_cred); if (error == 0) vnode_create_vobject(vp, zp->z_size, ap->a_td); return (error); } #ifndef _SYS_SYSPROTO_H_ struct vop_close_args { struct vnode *a_vp; int a_fflag; struct ucred *a_cred; struct thread *a_td; }; #endif static int zfs_freebsd_close(struct vop_close_args *ap) { return (zfs_close(ap->a_vp, ap->a_fflag, 1, 0, ap->a_cred)); } #ifndef _SYS_SYSPROTO_H_ struct vop_ioctl_args { struct vnode *a_vp; ulong_t a_command; caddr_t a_data; int a_fflag; struct ucred *cred; struct thread *td; }; #endif static int zfs_freebsd_ioctl(struct vop_ioctl_args *ap) { return (zfs_ioctl(ap->a_vp, ap->a_command, (intptr_t)ap->a_data, ap->a_fflag, ap->a_cred, NULL)); } static int ioflags(int ioflags) { int flags = 0; if (ioflags & IO_APPEND) flags |= O_APPEND; if (ioflags & IO_NDELAY) flags |= O_NONBLOCK; if (ioflags & IO_SYNC) flags |= O_SYNC; return (flags); } #ifndef _SYS_SYSPROTO_H_ struct vop_read_args { struct vnode *a_vp; struct uio *a_uio; int a_ioflag; struct ucred *a_cred; }; #endif static int zfs_freebsd_read(struct vop_read_args *ap) { zfs_uio_t uio; zfs_uio_init(&uio, ap->a_uio); return (zfs_read(VTOZ(ap->a_vp), &uio, ioflags(ap->a_ioflag), ap->a_cred)); } #ifndef _SYS_SYSPROTO_H_ struct vop_write_args { struct vnode *a_vp; struct uio *a_uio; int a_ioflag; struct ucred *a_cred; }; #endif static int zfs_freebsd_write(struct vop_write_args *ap) { zfs_uio_t uio; zfs_uio_init(&uio, ap->a_uio); return (zfs_write(VTOZ(ap->a_vp), &uio, ioflags(ap->a_ioflag), ap->a_cred)); } #if __FreeBSD_version >= 1300102 /* * VOP_FPLOOKUP_VEXEC routines are subject to special circumstances, see * the comment above cache_fplookup for details. */ static int zfs_freebsd_fplookup_vexec(struct vop_fplookup_vexec_args *v) { vnode_t *vp; znode_t *zp; uint64_t pflags; vp = v->a_vp; zp = VTOZ_SMR(vp); if (__predict_false(zp == NULL)) return (EAGAIN); pflags = atomic_load_64(&zp->z_pflags); if (pflags & ZFS_AV_QUARANTINED) return (EAGAIN); if (pflags & ZFS_XATTR) return (EAGAIN); if ((pflags & ZFS_NO_EXECS_DENIED) == 0) return (EAGAIN); return (0); } #endif #if __FreeBSD_version >= 1300139 static int zfs_freebsd_fplookup_symlink(struct vop_fplookup_symlink_args *v) { vnode_t *vp; znode_t *zp; char *target; vp = v->a_vp; zp = VTOZ_SMR(vp); if (__predict_false(zp == NULL)) { return (EAGAIN); } target = atomic_load_consume_ptr(&zp->z_cached_symlink); if (target == NULL) { return (EAGAIN); } return (cache_symlink_resolve(v->a_fpl, target, strlen(target))); } #endif #ifndef _SYS_SYSPROTO_H_ struct vop_access_args { struct vnode *a_vp; accmode_t a_accmode; struct ucred *a_cred; struct thread *a_td; }; #endif static int zfs_freebsd_access(struct vop_access_args *ap) { vnode_t *vp = ap->a_vp; znode_t *zp = VTOZ(vp); accmode_t accmode; int error = 0; if (ap->a_accmode == VEXEC) { if (zfs_fastaccesschk_execute(zp, ap->a_cred) == 0) return (0); } /* * ZFS itself only knowns about VREAD, VWRITE, VEXEC and VAPPEND, */ accmode = ap->a_accmode & (VREAD|VWRITE|VEXEC|VAPPEND); if (accmode != 0) error = zfs_access(zp, accmode, 0, ap->a_cred); /* * VADMIN has to be handled by vaccess(). */ if (error == 0) { accmode = ap->a_accmode & ~(VREAD|VWRITE|VEXEC|VAPPEND); if (accmode != 0) { #if __FreeBSD_version >= 1300105 error = vaccess(vp->v_type, zp->z_mode, zp->z_uid, zp->z_gid, accmode, ap->a_cred); #else error = vaccess(vp->v_type, zp->z_mode, zp->z_uid, zp->z_gid, accmode, ap->a_cred, NULL); #endif } } /* * For VEXEC, ensure that at least one execute bit is set for * non-directories. */ if (error == 0 && (ap->a_accmode & VEXEC) != 0 && vp->v_type != VDIR && (zp->z_mode & (S_IXUSR | S_IXGRP | S_IXOTH)) == 0) { error = EACCES; } return (error); } #ifndef _SYS_SYSPROTO_H_ struct vop_lookup_args { struct vnode *a_dvp; struct vnode **a_vpp; struct componentname *a_cnp; }; #endif static int zfs_freebsd_lookup(struct vop_lookup_args *ap, boolean_t cached) { struct componentname *cnp = ap->a_cnp; char nm[NAME_MAX + 1]; ASSERT3U(cnp->cn_namelen, <, sizeof (nm)); strlcpy(nm, cnp->cn_nameptr, MIN(cnp->cn_namelen + 1, sizeof (nm))); return (zfs_lookup(ap->a_dvp, nm, ap->a_vpp, cnp, cnp->cn_nameiop, cnp->cn_cred, 0, cached)); } static int zfs_freebsd_cachedlookup(struct vop_cachedlookup_args *ap) { return (zfs_freebsd_lookup((struct vop_lookup_args *)ap, B_TRUE)); } #ifndef _SYS_SYSPROTO_H_ struct vop_lookup_args { struct vnode *a_dvp; struct vnode **a_vpp; struct componentname *a_cnp; }; #endif static int zfs_cache_lookup(struct vop_lookup_args *ap) { zfsvfs_t *zfsvfs; zfsvfs = ap->a_dvp->v_mount->mnt_data; if (zfsvfs->z_use_namecache) return (vfs_cache_lookup(ap)); else return (zfs_freebsd_lookup(ap, B_FALSE)); } #ifndef _SYS_SYSPROTO_H_ struct vop_create_args { struct vnode *a_dvp; struct vnode **a_vpp; struct componentname *a_cnp; struct vattr *a_vap; }; #endif static int zfs_freebsd_create(struct vop_create_args *ap) { zfsvfs_t *zfsvfs; struct componentname *cnp = ap->a_cnp; vattr_t *vap = ap->a_vap; znode_t *zp = NULL; int rc, mode; #if __FreeBSD_version < 1400068 ASSERT(cnp->cn_flags & SAVENAME); #endif vattr_init_mask(vap); mode = vap->va_mode & ALLPERMS; zfsvfs = ap->a_dvp->v_mount->mnt_data; *ap->a_vpp = NULL; rc = zfs_create(VTOZ(ap->a_dvp), cnp->cn_nameptr, vap, 0, mode, &zp, cnp->cn_cred, 0 /* flag */, NULL /* vsecattr */, NULL); if (rc == 0) *ap->a_vpp = ZTOV(zp); if (zfsvfs->z_use_namecache && rc == 0 && (cnp->cn_flags & MAKEENTRY) != 0) cache_enter(ap->a_dvp, *ap->a_vpp, cnp); return (rc); } #ifndef _SYS_SYSPROTO_H_ struct vop_remove_args { struct vnode *a_dvp; struct vnode *a_vp; struct componentname *a_cnp; }; #endif static int zfs_freebsd_remove(struct vop_remove_args *ap) { #if __FreeBSD_version < 1400068 ASSERT(ap->a_cnp->cn_flags & SAVENAME); #endif return (zfs_remove_(ap->a_dvp, ap->a_vp, ap->a_cnp->cn_nameptr, ap->a_cnp->cn_cred)); } #ifndef _SYS_SYSPROTO_H_ struct vop_mkdir_args { struct vnode *a_dvp; struct vnode **a_vpp; struct componentname *a_cnp; struct vattr *a_vap; }; #endif static int zfs_freebsd_mkdir(struct vop_mkdir_args *ap) { vattr_t *vap = ap->a_vap; znode_t *zp = NULL; int rc; #if __FreeBSD_version < 1400068 ASSERT(ap->a_cnp->cn_flags & SAVENAME); #endif vattr_init_mask(vap); *ap->a_vpp = NULL; rc = zfs_mkdir(VTOZ(ap->a_dvp), ap->a_cnp->cn_nameptr, vap, &zp, ap->a_cnp->cn_cred, 0, NULL, NULL); if (rc == 0) *ap->a_vpp = ZTOV(zp); return (rc); } #ifndef _SYS_SYSPROTO_H_ struct vop_rmdir_args { struct vnode *a_dvp; struct vnode *a_vp; struct componentname *a_cnp; }; #endif static int zfs_freebsd_rmdir(struct vop_rmdir_args *ap) { struct componentname *cnp = ap->a_cnp; #if __FreeBSD_version < 1400068 ASSERT(cnp->cn_flags & SAVENAME); #endif return (zfs_rmdir_(ap->a_dvp, ap->a_vp, cnp->cn_nameptr, cnp->cn_cred)); } #ifndef _SYS_SYSPROTO_H_ struct vop_readdir_args { struct vnode *a_vp; struct uio *a_uio; struct ucred *a_cred; int *a_eofflag; int *a_ncookies; cookie_t **a_cookies; }; #endif static int zfs_freebsd_readdir(struct vop_readdir_args *ap) { zfs_uio_t uio; zfs_uio_init(&uio, ap->a_uio); return (zfs_readdir(ap->a_vp, &uio, ap->a_cred, ap->a_eofflag, ap->a_ncookies, ap->a_cookies)); } #ifndef _SYS_SYSPROTO_H_ struct vop_fsync_args { struct vnode *a_vp; int a_waitfor; struct thread *a_td; }; #endif static int zfs_freebsd_fsync(struct vop_fsync_args *ap) { return (zfs_fsync(VTOZ(ap->a_vp), 0, ap->a_td->td_ucred)); } #ifndef _SYS_SYSPROTO_H_ struct vop_getattr_args { struct vnode *a_vp; struct vattr *a_vap; struct ucred *a_cred; }; #endif static int zfs_freebsd_getattr(struct vop_getattr_args *ap) { vattr_t *vap = ap->a_vap; xvattr_t xvap; ulong_t fflags = 0; int error; xva_init(&xvap); xvap.xva_vattr = *vap; xvap.xva_vattr.va_mask |= AT_XVATTR; /* Convert chflags into ZFS-type flags. */ /* XXX: what about SF_SETTABLE?. */ XVA_SET_REQ(&xvap, XAT_IMMUTABLE); XVA_SET_REQ(&xvap, XAT_APPENDONLY); XVA_SET_REQ(&xvap, XAT_NOUNLINK); XVA_SET_REQ(&xvap, XAT_NODUMP); XVA_SET_REQ(&xvap, XAT_READONLY); XVA_SET_REQ(&xvap, XAT_ARCHIVE); XVA_SET_REQ(&xvap, XAT_SYSTEM); XVA_SET_REQ(&xvap, XAT_HIDDEN); XVA_SET_REQ(&xvap, XAT_REPARSE); XVA_SET_REQ(&xvap, XAT_OFFLINE); XVA_SET_REQ(&xvap, XAT_SPARSE); error = zfs_getattr(ap->a_vp, (vattr_t *)&xvap, 0, ap->a_cred); if (error != 0) return (error); /* Convert ZFS xattr into chflags. */ #define FLAG_CHECK(fflag, xflag, xfield) do { \ if (XVA_ISSET_RTN(&xvap, (xflag)) && (xfield) != 0) \ fflags |= (fflag); \ } while (0) FLAG_CHECK(SF_IMMUTABLE, XAT_IMMUTABLE, xvap.xva_xoptattrs.xoa_immutable); FLAG_CHECK(SF_APPEND, XAT_APPENDONLY, xvap.xva_xoptattrs.xoa_appendonly); FLAG_CHECK(SF_NOUNLINK, XAT_NOUNLINK, xvap.xva_xoptattrs.xoa_nounlink); FLAG_CHECK(UF_ARCHIVE, XAT_ARCHIVE, xvap.xva_xoptattrs.xoa_archive); FLAG_CHECK(UF_NODUMP, XAT_NODUMP, xvap.xva_xoptattrs.xoa_nodump); FLAG_CHECK(UF_READONLY, XAT_READONLY, xvap.xva_xoptattrs.xoa_readonly); FLAG_CHECK(UF_SYSTEM, XAT_SYSTEM, xvap.xva_xoptattrs.xoa_system); FLAG_CHECK(UF_HIDDEN, XAT_HIDDEN, xvap.xva_xoptattrs.xoa_hidden); FLAG_CHECK(UF_REPARSE, XAT_REPARSE, xvap.xva_xoptattrs.xoa_reparse); FLAG_CHECK(UF_OFFLINE, XAT_OFFLINE, xvap.xva_xoptattrs.xoa_offline); FLAG_CHECK(UF_SPARSE, XAT_SPARSE, xvap.xva_xoptattrs.xoa_sparse); #undef FLAG_CHECK *vap = xvap.xva_vattr; vap->va_flags = fflags; return (0); } #ifndef _SYS_SYSPROTO_H_ struct vop_setattr_args { struct vnode *a_vp; struct vattr *a_vap; struct ucred *a_cred; }; #endif static int zfs_freebsd_setattr(struct vop_setattr_args *ap) { vnode_t *vp = ap->a_vp; vattr_t *vap = ap->a_vap; cred_t *cred = ap->a_cred; xvattr_t xvap; ulong_t fflags; uint64_t zflags; vattr_init_mask(vap); vap->va_mask &= ~AT_NOSET; xva_init(&xvap); xvap.xva_vattr = *vap; zflags = VTOZ(vp)->z_pflags; if (vap->va_flags != VNOVAL) { zfsvfs_t *zfsvfs = VTOZ(vp)->z_zfsvfs; int error; if (zfsvfs->z_use_fuids == B_FALSE) return (EOPNOTSUPP); fflags = vap->va_flags; /* * XXX KDM * We need to figure out whether it makes sense to allow * UF_REPARSE through, since we don't really have other * facilities to handle reparse points and zfs_setattr() * doesn't currently allow setting that attribute anyway. */ if ((fflags & ~(SF_IMMUTABLE|SF_APPEND|SF_NOUNLINK|UF_ARCHIVE| UF_NODUMP|UF_SYSTEM|UF_HIDDEN|UF_READONLY|UF_REPARSE| UF_OFFLINE|UF_SPARSE)) != 0) return (EOPNOTSUPP); /* * Unprivileged processes are not permitted to unset system * flags, or modify flags if any system flags are set. * Privileged non-jail processes may not modify system flags * if securelevel > 0 and any existing system flags are set. * Privileged jail processes behave like privileged non-jail * processes if the PR_ALLOW_CHFLAGS permission bit is set; * otherwise, they behave like unprivileged processes. */ if (secpolicy_fs_owner(vp->v_mount, cred) == 0 || spl_priv_check_cred(cred, PRIV_VFS_SYSFLAGS) == 0) { if (zflags & (ZFS_IMMUTABLE | ZFS_APPENDONLY | ZFS_NOUNLINK)) { error = securelevel_gt(cred, 0); if (error != 0) return (error); } } else { /* * Callers may only modify the file flags on * objects they have VADMIN rights for. */ if ((error = VOP_ACCESS(vp, VADMIN, cred, curthread)) != 0) return (error); if (zflags & (ZFS_IMMUTABLE | ZFS_APPENDONLY | ZFS_NOUNLINK)) { return (EPERM); } if (fflags & (SF_IMMUTABLE | SF_APPEND | SF_NOUNLINK)) { return (EPERM); } } #define FLAG_CHANGE(fflag, zflag, xflag, xfield) do { \ if (((fflags & (fflag)) && !(zflags & (zflag))) || \ ((zflags & (zflag)) && !(fflags & (fflag)))) { \ XVA_SET_REQ(&xvap, (xflag)); \ (xfield) = ((fflags & (fflag)) != 0); \ } \ } while (0) /* Convert chflags into ZFS-type flags. */ /* XXX: what about SF_SETTABLE?. */ FLAG_CHANGE(SF_IMMUTABLE, ZFS_IMMUTABLE, XAT_IMMUTABLE, xvap.xva_xoptattrs.xoa_immutable); FLAG_CHANGE(SF_APPEND, ZFS_APPENDONLY, XAT_APPENDONLY, xvap.xva_xoptattrs.xoa_appendonly); FLAG_CHANGE(SF_NOUNLINK, ZFS_NOUNLINK, XAT_NOUNLINK, xvap.xva_xoptattrs.xoa_nounlink); FLAG_CHANGE(UF_ARCHIVE, ZFS_ARCHIVE, XAT_ARCHIVE, xvap.xva_xoptattrs.xoa_archive); FLAG_CHANGE(UF_NODUMP, ZFS_NODUMP, XAT_NODUMP, xvap.xva_xoptattrs.xoa_nodump); FLAG_CHANGE(UF_READONLY, ZFS_READONLY, XAT_READONLY, xvap.xva_xoptattrs.xoa_readonly); FLAG_CHANGE(UF_SYSTEM, ZFS_SYSTEM, XAT_SYSTEM, xvap.xva_xoptattrs.xoa_system); FLAG_CHANGE(UF_HIDDEN, ZFS_HIDDEN, XAT_HIDDEN, xvap.xva_xoptattrs.xoa_hidden); FLAG_CHANGE(UF_REPARSE, ZFS_REPARSE, XAT_REPARSE, xvap.xva_xoptattrs.xoa_reparse); FLAG_CHANGE(UF_OFFLINE, ZFS_OFFLINE, XAT_OFFLINE, xvap.xva_xoptattrs.xoa_offline); FLAG_CHANGE(UF_SPARSE, ZFS_SPARSE, XAT_SPARSE, xvap.xva_xoptattrs.xoa_sparse); #undef FLAG_CHANGE } if (vap->va_birthtime.tv_sec != VNOVAL) { xvap.xva_vattr.va_mask |= AT_XVATTR; XVA_SET_REQ(&xvap, XAT_CREATETIME); } return (zfs_setattr(VTOZ(vp), (vattr_t *)&xvap, 0, cred, NULL)); } #ifndef _SYS_SYSPROTO_H_ struct vop_rename_args { struct vnode *a_fdvp; struct vnode *a_fvp; struct componentname *a_fcnp; struct vnode *a_tdvp; struct vnode *a_tvp; struct componentname *a_tcnp; }; #endif static int zfs_freebsd_rename(struct vop_rename_args *ap) { vnode_t *fdvp = ap->a_fdvp; vnode_t *fvp = ap->a_fvp; vnode_t *tdvp = ap->a_tdvp; vnode_t *tvp = ap->a_tvp; int error; #if __FreeBSD_version < 1400068 ASSERT(ap->a_fcnp->cn_flags & (SAVENAME|SAVESTART)); ASSERT(ap->a_tcnp->cn_flags & (SAVENAME|SAVESTART)); #endif error = zfs_do_rename(fdvp, &fvp, ap->a_fcnp, tdvp, &tvp, ap->a_tcnp, ap->a_fcnp->cn_cred); vrele(fdvp); vrele(fvp); vrele(tdvp); if (tvp != NULL) vrele(tvp); return (error); } #ifndef _SYS_SYSPROTO_H_ struct vop_symlink_args { struct vnode *a_dvp; struct vnode **a_vpp; struct componentname *a_cnp; struct vattr *a_vap; char *a_target; }; #endif static int zfs_freebsd_symlink(struct vop_symlink_args *ap) { struct componentname *cnp = ap->a_cnp; vattr_t *vap = ap->a_vap; znode_t *zp = NULL; #if __FreeBSD_version >= 1300139 char *symlink; size_t symlink_len; #endif int rc; #if __FreeBSD_version < 1400068 ASSERT(cnp->cn_flags & SAVENAME); #endif vap->va_type = VLNK; /* FreeBSD: Syscall only sets va_mode. */ vattr_init_mask(vap); *ap->a_vpp = NULL; rc = zfs_symlink(VTOZ(ap->a_dvp), cnp->cn_nameptr, vap, ap->a_target, &zp, cnp->cn_cred, 0 /* flags */, NULL); if (rc == 0) { *ap->a_vpp = ZTOV(zp); ASSERT_VOP_ELOCKED(ZTOV(zp), __func__); #if __FreeBSD_version >= 1300139 MPASS(zp->z_cached_symlink == NULL); symlink_len = strlen(ap->a_target); symlink = cache_symlink_alloc(symlink_len + 1, M_WAITOK); if (symlink != NULL) { memcpy(symlink, ap->a_target, symlink_len); symlink[symlink_len] = '\0'; atomic_store_rel_ptr((uintptr_t *)&zp->z_cached_symlink, (uintptr_t)symlink); } #endif } return (rc); } #ifndef _SYS_SYSPROTO_H_ struct vop_readlink_args { struct vnode *a_vp; struct uio *a_uio; struct ucred *a_cred; }; #endif static int zfs_freebsd_readlink(struct vop_readlink_args *ap) { zfs_uio_t uio; int error; #if __FreeBSD_version >= 1300139 znode_t *zp = VTOZ(ap->a_vp); char *symlink, *base; size_t symlink_len; bool trycache; #endif zfs_uio_init(&uio, ap->a_uio); #if __FreeBSD_version >= 1300139 trycache = false; if (zfs_uio_segflg(&uio) == UIO_SYSSPACE && zfs_uio_iovcnt(&uio) == 1) { base = zfs_uio_iovbase(&uio, 0); symlink_len = zfs_uio_iovlen(&uio, 0); trycache = true; } #endif error = zfs_readlink(ap->a_vp, &uio, ap->a_cred, NULL); #if __FreeBSD_version >= 1300139 if (atomic_load_ptr(&zp->z_cached_symlink) != NULL || error != 0 || !trycache) { return (error); } symlink_len -= zfs_uio_resid(&uio); symlink = cache_symlink_alloc(symlink_len + 1, M_WAITOK); if (symlink != NULL) { memcpy(symlink, base, symlink_len); symlink[symlink_len] = '\0'; if (!atomic_cmpset_rel_ptr((uintptr_t *)&zp->z_cached_symlink, (uintptr_t)NULL, (uintptr_t)symlink)) { cache_symlink_free(symlink, symlink_len + 1); } } #endif return (error); } #ifndef _SYS_SYSPROTO_H_ struct vop_link_args { struct vnode *a_tdvp; struct vnode *a_vp; struct componentname *a_cnp; }; #endif static int zfs_freebsd_link(struct vop_link_args *ap) { struct componentname *cnp = ap->a_cnp; vnode_t *vp = ap->a_vp; vnode_t *tdvp = ap->a_tdvp; if (tdvp->v_mount != vp->v_mount) return (EXDEV); #if __FreeBSD_version < 1400068 ASSERT(cnp->cn_flags & SAVENAME); #endif return (zfs_link(VTOZ(tdvp), VTOZ(vp), cnp->cn_nameptr, cnp->cn_cred, 0)); } #ifndef _SYS_SYSPROTO_H_ struct vop_inactive_args { struct vnode *a_vp; struct thread *a_td; }; #endif static int zfs_freebsd_inactive(struct vop_inactive_args *ap) { vnode_t *vp = ap->a_vp; #if __FreeBSD_version >= 1300123 zfs_inactive(vp, curthread->td_ucred, NULL); #else zfs_inactive(vp, ap->a_td->td_ucred, NULL); #endif return (0); } #if __FreeBSD_version >= 1300042 #ifndef _SYS_SYSPROTO_H_ struct vop_need_inactive_args { struct vnode *a_vp; struct thread *a_td; }; #endif static int zfs_freebsd_need_inactive(struct vop_need_inactive_args *ap) { vnode_t *vp = ap->a_vp; znode_t *zp = VTOZ(vp); zfsvfs_t *zfsvfs = zp->z_zfsvfs; int need; if (vn_need_pageq_flush(vp)) return (1); if (!ZFS_TEARDOWN_INACTIVE_TRY_ENTER_READ(zfsvfs)) return (1); need = (zp->z_sa_hdl == NULL || zp->z_unlinked || zp->z_atime_dirty); ZFS_TEARDOWN_INACTIVE_EXIT_READ(zfsvfs); return (need); } #endif #ifndef _SYS_SYSPROTO_H_ struct vop_reclaim_args { struct vnode *a_vp; struct thread *a_td; }; #endif static int zfs_freebsd_reclaim(struct vop_reclaim_args *ap) { vnode_t *vp = ap->a_vp; znode_t *zp = VTOZ(vp); zfsvfs_t *zfsvfs = zp->z_zfsvfs; ASSERT3P(zp, !=, NULL); #if __FreeBSD_version < 1300042 /* Destroy the vm object and flush associated pages. */ vnode_destroy_vobject(vp); #endif /* * z_teardown_inactive_lock protects from a race with * zfs_znode_dmu_fini in zfsvfs_teardown during * force unmount. */ ZFS_TEARDOWN_INACTIVE_ENTER_READ(zfsvfs); if (zp->z_sa_hdl == NULL) zfs_znode_free(zp); else zfs_zinactive(zp); ZFS_TEARDOWN_INACTIVE_EXIT_READ(zfsvfs); vp->v_data = NULL; return (0); } #ifndef _SYS_SYSPROTO_H_ struct vop_fid_args { struct vnode *a_vp; struct fid *a_fid; }; #endif static int zfs_freebsd_fid(struct vop_fid_args *ap) { return (zfs_fid(ap->a_vp, (void *)ap->a_fid, NULL)); } #ifndef _SYS_SYSPROTO_H_ struct vop_pathconf_args { struct vnode *a_vp; int a_name; register_t *a_retval; } *ap; #endif static int zfs_freebsd_pathconf(struct vop_pathconf_args *ap) { ulong_t val; int error; error = zfs_pathconf(ap->a_vp, ap->a_name, &val, curthread->td_ucred, NULL); if (error == 0) { *ap->a_retval = val; return (error); } if (error != EOPNOTSUPP) return (error); switch (ap->a_name) { case _PC_NAME_MAX: *ap->a_retval = NAME_MAX; return (0); #if __FreeBSD_version >= 1400032 case _PC_DEALLOC_PRESENT: *ap->a_retval = 1; return (0); #endif case _PC_PIPE_BUF: if (ap->a_vp->v_type == VDIR || ap->a_vp->v_type == VFIFO) { *ap->a_retval = PIPE_BUF; return (0); } return (EINVAL); default: return (vop_stdpathconf(ap)); } } static int zfs_xattr_compat = 1; static int zfs_check_attrname(const char *name) { /* We don't allow '/' character in attribute name. */ if (strchr(name, '/') != NULL) return (SET_ERROR(EINVAL)); /* We don't allow attribute names that start with a namespace prefix. */ if (ZFS_XA_NS_PREFIX_FORBIDDEN(name)) return (SET_ERROR(EINVAL)); return (0); } /* * FreeBSD's extended attributes namespace defines file name prefix for ZFS' * extended attribute name: * * NAMESPACE XATTR_COMPAT PREFIX * system * freebsd:system: * user 1 (none, can be used to access ZFS * fsattr(5) attributes created on Solaris) * user 0 user. */ static int zfs_create_attrname(int attrnamespace, const char *name, char *attrname, size_t size, boolean_t compat) { const char *namespace, *prefix, *suffix; memset(attrname, 0, size); switch (attrnamespace) { case EXTATTR_NAMESPACE_USER: if (compat) { /* * This is the default namespace by which we can access * all attributes created on Solaris. */ prefix = namespace = suffix = ""; } else { /* * This is compatible with the user namespace encoding * on Linux prior to xattr_compat, but nothing * else. */ prefix = ""; namespace = "user"; suffix = "."; } break; case EXTATTR_NAMESPACE_SYSTEM: prefix = "freebsd:"; namespace = EXTATTR_NAMESPACE_SYSTEM_STRING; suffix = ":"; break; case EXTATTR_NAMESPACE_EMPTY: default: return (SET_ERROR(EINVAL)); } if (snprintf(attrname, size, "%s%s%s%s", prefix, namespace, suffix, name) >= size) { return (SET_ERROR(ENAMETOOLONG)); } return (0); } static int zfs_ensure_xattr_cached(znode_t *zp) { int error = 0; ASSERT(RW_LOCK_HELD(&zp->z_xattr_lock)); if (zp->z_xattr_cached != NULL) return (0); if (rw_write_held(&zp->z_xattr_lock)) return (zfs_sa_get_xattr(zp)); if (!rw_tryupgrade(&zp->z_xattr_lock)) { rw_exit(&zp->z_xattr_lock); rw_enter(&zp->z_xattr_lock, RW_WRITER); } if (zp->z_xattr_cached == NULL) error = zfs_sa_get_xattr(zp); rw_downgrade(&zp->z_xattr_lock); return (error); } #ifndef _SYS_SYSPROTO_H_ struct vop_getextattr { IN struct vnode *a_vp; IN int a_attrnamespace; IN const char *a_name; INOUT struct uio *a_uio; OUT size_t *a_size; IN struct ucred *a_cred; IN struct thread *a_td; }; #endif static int zfs_getextattr_dir(struct vop_getextattr_args *ap, const char *attrname) { struct thread *td = ap->a_td; struct nameidata nd; struct vattr va; vnode_t *xvp = NULL, *vp; int error, flags; error = zfs_lookup(ap->a_vp, NULL, &xvp, NULL, 0, ap->a_cred, LOOKUP_XATTR, B_FALSE); if (error != 0) return (error); flags = FREAD; #if __FreeBSD_version < 1400043 NDINIT_ATVP(&nd, LOOKUP, NOFOLLOW, UIO_SYSSPACE, attrname, xvp, td); #else NDINIT_ATVP(&nd, LOOKUP, NOFOLLOW, UIO_SYSSPACE, attrname, xvp); #endif error = vn_open_cred(&nd, &flags, 0, VN_OPEN_INVFS, ap->a_cred, NULL); if (error != 0) return (SET_ERROR(error)); vp = nd.ni_vp; NDFREE_PNBUF(&nd); if (ap->a_size != NULL) { error = VOP_GETATTR(vp, &va, ap->a_cred); if (error == 0) *ap->a_size = (size_t)va.va_size; } else if (ap->a_uio != NULL) error = VOP_READ(vp, ap->a_uio, IO_UNIT, ap->a_cred); VOP_UNLOCK1(vp); vn_close(vp, flags, ap->a_cred, td); return (error); } static int zfs_getextattr_sa(struct vop_getextattr_args *ap, const char *attrname) { znode_t *zp = VTOZ(ap->a_vp); uchar_t *nv_value; uint_t nv_size; int error; error = zfs_ensure_xattr_cached(zp); if (error != 0) return (error); ASSERT(RW_LOCK_HELD(&zp->z_xattr_lock)); ASSERT3P(zp->z_xattr_cached, !=, NULL); error = nvlist_lookup_byte_array(zp->z_xattr_cached, attrname, &nv_value, &nv_size); if (error != 0) return (SET_ERROR(error)); if (ap->a_size != NULL) *ap->a_size = nv_size; else if (ap->a_uio != NULL) error = uiomove(nv_value, nv_size, ap->a_uio); if (error != 0) return (SET_ERROR(error)); return (0); } static int zfs_getextattr_impl(struct vop_getextattr_args *ap, boolean_t compat) { znode_t *zp = VTOZ(ap->a_vp); zfsvfs_t *zfsvfs = ZTOZSB(zp); char attrname[EXTATTR_MAXNAMELEN+1]; int error; error = zfs_create_attrname(ap->a_attrnamespace, ap->a_name, attrname, sizeof (attrname), compat); if (error != 0) return (error); error = ENOENT; if (zfsvfs->z_use_sa && zp->z_is_sa) error = zfs_getextattr_sa(ap, attrname); if (error == ENOENT) error = zfs_getextattr_dir(ap, attrname); return (error); } /* * Vnode operation to retrieve a named extended attribute. */ static int zfs_getextattr(struct vop_getextattr_args *ap) { znode_t *zp = VTOZ(ap->a_vp); zfsvfs_t *zfsvfs = ZTOZSB(zp); int error; /* * If the xattr property is off, refuse the request. */ if (!(zfsvfs->z_flags & ZSB_XATTR)) return (SET_ERROR(EOPNOTSUPP)); error = extattr_check_cred(ap->a_vp, ap->a_attrnamespace, ap->a_cred, ap->a_td, VREAD); if (error != 0) return (SET_ERROR(error)); error = zfs_check_attrname(ap->a_name); if (error != 0) return (error); if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); error = ENOENT; rw_enter(&zp->z_xattr_lock, RW_READER); error = zfs_getextattr_impl(ap, zfs_xattr_compat); if ((error == ENOENT || error == ENOATTR) && ap->a_attrnamespace == EXTATTR_NAMESPACE_USER) { /* * Fall back to the alternate namespace format if we failed to * find a user xattr. */ error = zfs_getextattr_impl(ap, !zfs_xattr_compat); } rw_exit(&zp->z_xattr_lock); zfs_exit(zfsvfs, FTAG); if (error == ENOENT) error = SET_ERROR(ENOATTR); return (error); } #ifndef _SYS_SYSPROTO_H_ struct vop_deleteextattr { IN struct vnode *a_vp; IN int a_attrnamespace; IN const char *a_name; IN struct ucred *a_cred; IN struct thread *a_td; }; #endif static int zfs_deleteextattr_dir(struct vop_deleteextattr_args *ap, const char *attrname) { struct nameidata nd; vnode_t *xvp = NULL, *vp; int error; error = zfs_lookup(ap->a_vp, NULL, &xvp, NULL, 0, ap->a_cred, LOOKUP_XATTR, B_FALSE); if (error != 0) return (error); #if __FreeBSD_version < 1400043 NDINIT_ATVP(&nd, DELETE, NOFOLLOW | LOCKPARENT | LOCKLEAF, UIO_SYSSPACE, attrname, xvp, ap->a_td); #else NDINIT_ATVP(&nd, DELETE, NOFOLLOW | LOCKPARENT | LOCKLEAF, UIO_SYSSPACE, attrname, xvp); #endif error = namei(&nd); if (error != 0) return (SET_ERROR(error)); vp = nd.ni_vp; error = VOP_REMOVE(nd.ni_dvp, vp, &nd.ni_cnd); NDFREE_PNBUF(&nd); vput(nd.ni_dvp); if (vp == nd.ni_dvp) vrele(vp); else vput(vp); return (error); } static int zfs_deleteextattr_sa(struct vop_deleteextattr_args *ap, const char *attrname) { znode_t *zp = VTOZ(ap->a_vp); nvlist_t *nvl; int error; error = zfs_ensure_xattr_cached(zp); if (error != 0) return (error); ASSERT(RW_WRITE_HELD(&zp->z_xattr_lock)); ASSERT3P(zp->z_xattr_cached, !=, NULL); nvl = zp->z_xattr_cached; error = nvlist_remove(nvl, attrname, DATA_TYPE_BYTE_ARRAY); if (error != 0) error = SET_ERROR(error); else error = zfs_sa_set_xattr(zp, attrname, NULL, 0); if (error != 0) { zp->z_xattr_cached = NULL; nvlist_free(nvl); } return (error); } static int zfs_deleteextattr_impl(struct vop_deleteextattr_args *ap, boolean_t compat) { znode_t *zp = VTOZ(ap->a_vp); zfsvfs_t *zfsvfs = ZTOZSB(zp); char attrname[EXTATTR_MAXNAMELEN+1]; int error; error = zfs_create_attrname(ap->a_attrnamespace, ap->a_name, attrname, sizeof (attrname), compat); if (error != 0) return (error); error = ENOENT; if (zfsvfs->z_use_sa && zp->z_is_sa) error = zfs_deleteextattr_sa(ap, attrname); if (error == ENOENT) error = zfs_deleteextattr_dir(ap, attrname); return (error); } /* * Vnode operation to remove a named attribute. */ static int zfs_deleteextattr(struct vop_deleteextattr_args *ap) { znode_t *zp = VTOZ(ap->a_vp); zfsvfs_t *zfsvfs = ZTOZSB(zp); int error; /* * If the xattr property is off, refuse the request. */ if (!(zfsvfs->z_flags & ZSB_XATTR)) return (SET_ERROR(EOPNOTSUPP)); error = extattr_check_cred(ap->a_vp, ap->a_attrnamespace, ap->a_cred, ap->a_td, VWRITE); if (error != 0) return (SET_ERROR(error)); error = zfs_check_attrname(ap->a_name); if (error != 0) return (error); if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); rw_enter(&zp->z_xattr_lock, RW_WRITER); error = zfs_deleteextattr_impl(ap, zfs_xattr_compat); if ((error == ENOENT || error == ENOATTR) && ap->a_attrnamespace == EXTATTR_NAMESPACE_USER) { /* * Fall back to the alternate namespace format if we failed to * find a user xattr. */ error = zfs_deleteextattr_impl(ap, !zfs_xattr_compat); } rw_exit(&zp->z_xattr_lock); zfs_exit(zfsvfs, FTAG); if (error == ENOENT) error = SET_ERROR(ENOATTR); return (error); } #ifndef _SYS_SYSPROTO_H_ struct vop_setextattr { IN struct vnode *a_vp; IN int a_attrnamespace; IN const char *a_name; INOUT struct uio *a_uio; IN struct ucred *a_cred; IN struct thread *a_td; }; #endif static int zfs_setextattr_dir(struct vop_setextattr_args *ap, const char *attrname) { struct thread *td = ap->a_td; struct nameidata nd; struct vattr va; vnode_t *xvp = NULL, *vp; int error, flags; error = zfs_lookup(ap->a_vp, NULL, &xvp, NULL, 0, ap->a_cred, LOOKUP_XATTR | CREATE_XATTR_DIR, B_FALSE); if (error != 0) return (error); flags = FFLAGS(O_WRONLY | O_CREAT); #if __FreeBSD_version < 1400043 NDINIT_ATVP(&nd, LOOKUP, NOFOLLOW, UIO_SYSSPACE, attrname, xvp, td); #else NDINIT_ATVP(&nd, LOOKUP, NOFOLLOW, UIO_SYSSPACE, attrname, xvp); #endif error = vn_open_cred(&nd, &flags, 0600, VN_OPEN_INVFS, ap->a_cred, NULL); if (error != 0) return (SET_ERROR(error)); vp = nd.ni_vp; NDFREE_PNBUF(&nd); VATTR_NULL(&va); va.va_size = 0; error = VOP_SETATTR(vp, &va, ap->a_cred); if (error == 0) VOP_WRITE(vp, ap->a_uio, IO_UNIT, ap->a_cred); VOP_UNLOCK1(vp); vn_close(vp, flags, ap->a_cred, td); return (error); } static int zfs_setextattr_sa(struct vop_setextattr_args *ap, const char *attrname) { znode_t *zp = VTOZ(ap->a_vp); nvlist_t *nvl; size_t sa_size; int error; error = zfs_ensure_xattr_cached(zp); if (error != 0) return (error); ASSERT(RW_WRITE_HELD(&zp->z_xattr_lock)); ASSERT3P(zp->z_xattr_cached, !=, NULL); nvl = zp->z_xattr_cached; size_t entry_size = ap->a_uio->uio_resid; if (entry_size > DXATTR_MAX_ENTRY_SIZE) return (SET_ERROR(EFBIG)); error = nvlist_size(nvl, &sa_size, NV_ENCODE_XDR); if (error != 0) return (SET_ERROR(error)); if (sa_size > DXATTR_MAX_SA_SIZE) return (SET_ERROR(EFBIG)); uchar_t *buf = kmem_alloc(entry_size, KM_SLEEP); error = uiomove(buf, entry_size, ap->a_uio); if (error != 0) { error = SET_ERROR(error); } else { error = nvlist_add_byte_array(nvl, attrname, buf, entry_size); if (error != 0) error = SET_ERROR(error); } if (error == 0) error = zfs_sa_set_xattr(zp, attrname, buf, entry_size); kmem_free(buf, entry_size); if (error != 0) { zp->z_xattr_cached = NULL; nvlist_free(nvl); } return (error); } static int zfs_setextattr_impl(struct vop_setextattr_args *ap, boolean_t compat) { znode_t *zp = VTOZ(ap->a_vp); zfsvfs_t *zfsvfs = ZTOZSB(zp); char attrname[EXTATTR_MAXNAMELEN+1]; int error; error = zfs_create_attrname(ap->a_attrnamespace, ap->a_name, attrname, sizeof (attrname), compat); if (error != 0) return (error); struct vop_deleteextattr_args vda = { .a_vp = ap->a_vp, .a_attrnamespace = ap->a_attrnamespace, .a_name = ap->a_name, .a_cred = ap->a_cred, .a_td = ap->a_td, }; error = ENOENT; if (zfsvfs->z_use_sa && zp->z_is_sa && zfsvfs->z_xattr_sa) { error = zfs_setextattr_sa(ap, attrname); if (error == 0) { /* * Successfully put into SA, we need to clear the one * in dir if present. */ zfs_deleteextattr_dir(&vda, attrname); } } if (error != 0) { error = zfs_setextattr_dir(ap, attrname); if (error == 0 && zp->z_is_sa) { /* * Successfully put into dir, we need to clear the one * in SA if present. */ zfs_deleteextattr_sa(&vda, attrname); } } if (error == 0 && ap->a_attrnamespace == EXTATTR_NAMESPACE_USER) { /* * Also clear all versions of the alternate compat name. */ zfs_deleteextattr_impl(&vda, !compat); } return (error); } /* * Vnode operation to set a named attribute. */ static int zfs_setextattr(struct vop_setextattr_args *ap) { znode_t *zp = VTOZ(ap->a_vp); zfsvfs_t *zfsvfs = ZTOZSB(zp); int error; /* * If the xattr property is off, refuse the request. */ if (!(zfsvfs->z_flags & ZSB_XATTR)) return (SET_ERROR(EOPNOTSUPP)); error = extattr_check_cred(ap->a_vp, ap->a_attrnamespace, ap->a_cred, ap->a_td, VWRITE); if (error != 0) return (SET_ERROR(error)); error = zfs_check_attrname(ap->a_name); if (error != 0) return (error); if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); rw_enter(&zp->z_xattr_lock, RW_WRITER); error = zfs_setextattr_impl(ap, zfs_xattr_compat); rw_exit(&zp->z_xattr_lock); zfs_exit(zfsvfs, FTAG); return (error); } #ifndef _SYS_SYSPROTO_H_ struct vop_listextattr { IN struct vnode *a_vp; IN int a_attrnamespace; INOUT struct uio *a_uio; OUT size_t *a_size; IN struct ucred *a_cred; IN struct thread *a_td; }; #endif static int zfs_listextattr_dir(struct vop_listextattr_args *ap, const char *attrprefix) { struct thread *td = ap->a_td; struct nameidata nd; uint8_t dirbuf[sizeof (struct dirent)]; struct iovec aiov; struct uio auio; vnode_t *xvp = NULL, *vp; int error, eof; error = zfs_lookup(ap->a_vp, NULL, &xvp, NULL, 0, ap->a_cred, LOOKUP_XATTR, B_FALSE); if (error != 0) { /* * ENOATTR means that the EA directory does not yet exist, * i.e. there are no extended attributes there. */ if (error == ENOATTR) error = 0; return (error); } #if __FreeBSD_version < 1400043 NDINIT_ATVP(&nd, LOOKUP, NOFOLLOW | LOCKLEAF | LOCKSHARED, UIO_SYSSPACE, ".", xvp, td); #else NDINIT_ATVP(&nd, LOOKUP, NOFOLLOW | LOCKLEAF | LOCKSHARED, UIO_SYSSPACE, ".", xvp); #endif error = namei(&nd); if (error != 0) return (SET_ERROR(error)); vp = nd.ni_vp; NDFREE_PNBUF(&nd); auio.uio_iov = &aiov; auio.uio_iovcnt = 1; auio.uio_segflg = UIO_SYSSPACE; auio.uio_td = td; auio.uio_rw = UIO_READ; auio.uio_offset = 0; size_t plen = strlen(attrprefix); do { aiov.iov_base = (void *)dirbuf; aiov.iov_len = sizeof (dirbuf); auio.uio_resid = sizeof (dirbuf); error = VOP_READDIR(vp, &auio, ap->a_cred, &eof, NULL, NULL); if (error != 0) break; int done = sizeof (dirbuf) - auio.uio_resid; for (int pos = 0; pos < done; ) { struct dirent *dp = (struct dirent *)(dirbuf + pos); pos += dp->d_reclen; /* * XXX: Temporarily we also accept DT_UNKNOWN, as this * is what we get when attribute was created on Solaris. */ if (dp->d_type != DT_REG && dp->d_type != DT_UNKNOWN) continue; else if (plen == 0 && ZFS_XA_NS_PREFIX_FORBIDDEN(dp->d_name)) continue; else if (strncmp(dp->d_name, attrprefix, plen) != 0) continue; uint8_t nlen = dp->d_namlen - plen; if (ap->a_size != NULL) { *ap->a_size += 1 + nlen; } else if (ap->a_uio != NULL) { /* * Format of extattr name entry is one byte for * length and the rest for name. */ error = uiomove(&nlen, 1, ap->a_uio); if (error == 0) { char *namep = dp->d_name + plen; error = uiomove(namep, nlen, ap->a_uio); } if (error != 0) { error = SET_ERROR(error); break; } } } } while (!eof && error == 0); vput(vp); return (error); } static int zfs_listextattr_sa(struct vop_listextattr_args *ap, const char *attrprefix) { znode_t *zp = VTOZ(ap->a_vp); int error; error = zfs_ensure_xattr_cached(zp); if (error != 0) return (error); ASSERT(RW_LOCK_HELD(&zp->z_xattr_lock)); ASSERT3P(zp->z_xattr_cached, !=, NULL); size_t plen = strlen(attrprefix); nvpair_t *nvp = NULL; while ((nvp = nvlist_next_nvpair(zp->z_xattr_cached, nvp)) != NULL) { ASSERT3U(nvpair_type(nvp), ==, DATA_TYPE_BYTE_ARRAY); const char *name = nvpair_name(nvp); if (plen == 0 && ZFS_XA_NS_PREFIX_FORBIDDEN(name)) continue; else if (strncmp(name, attrprefix, plen) != 0) continue; uint8_t nlen = strlen(name) - plen; if (ap->a_size != NULL) { *ap->a_size += 1 + nlen; } else if (ap->a_uio != NULL) { /* * Format of extattr name entry is one byte for * length and the rest for name. */ error = uiomove(&nlen, 1, ap->a_uio); if (error == 0) { char *namep = __DECONST(char *, name) + plen; error = uiomove(namep, nlen, ap->a_uio); } if (error != 0) { error = SET_ERROR(error); break; } } } return (error); } static int zfs_listextattr_impl(struct vop_listextattr_args *ap, boolean_t compat) { znode_t *zp = VTOZ(ap->a_vp); zfsvfs_t *zfsvfs = ZTOZSB(zp); char attrprefix[16]; int error; error = zfs_create_attrname(ap->a_attrnamespace, "", attrprefix, sizeof (attrprefix), compat); if (error != 0) return (error); if (zfsvfs->z_use_sa && zp->z_is_sa) error = zfs_listextattr_sa(ap, attrprefix); if (error == 0) error = zfs_listextattr_dir(ap, attrprefix); return (error); } /* * Vnode operation to retrieve extended attributes on a vnode. */ static int zfs_listextattr(struct vop_listextattr_args *ap) { znode_t *zp = VTOZ(ap->a_vp); zfsvfs_t *zfsvfs = ZTOZSB(zp); int error; if (ap->a_size != NULL) *ap->a_size = 0; /* * If the xattr property is off, refuse the request. */ if (!(zfsvfs->z_flags & ZSB_XATTR)) return (SET_ERROR(EOPNOTSUPP)); error = extattr_check_cred(ap->a_vp, ap->a_attrnamespace, ap->a_cred, ap->a_td, VREAD); if (error != 0) return (SET_ERROR(error)); if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); rw_enter(&zp->z_xattr_lock, RW_READER); error = zfs_listextattr_impl(ap, zfs_xattr_compat); if (error == 0 && ap->a_attrnamespace == EXTATTR_NAMESPACE_USER) { /* Also list user xattrs with the alternate format. */ error = zfs_listextattr_impl(ap, !zfs_xattr_compat); } rw_exit(&zp->z_xattr_lock); zfs_exit(zfsvfs, FTAG); return (error); } #ifndef _SYS_SYSPROTO_H_ struct vop_getacl_args { struct vnode *vp; acl_type_t type; struct acl *aclp; struct ucred *cred; struct thread *td; }; #endif static int zfs_freebsd_getacl(struct vop_getacl_args *ap) { int error; vsecattr_t vsecattr; if (ap->a_type != ACL_TYPE_NFS4) return (EINVAL); vsecattr.vsa_mask = VSA_ACE | VSA_ACECNT; if ((error = zfs_getsecattr(VTOZ(ap->a_vp), &vsecattr, 0, ap->a_cred))) return (error); error = acl_from_aces(ap->a_aclp, vsecattr.vsa_aclentp, vsecattr.vsa_aclcnt); if (vsecattr.vsa_aclentp != NULL) kmem_free(vsecattr.vsa_aclentp, vsecattr.vsa_aclentsz); return (error); } #ifndef _SYS_SYSPROTO_H_ struct vop_setacl_args { struct vnode *vp; acl_type_t type; struct acl *aclp; struct ucred *cred; struct thread *td; }; #endif static int zfs_freebsd_setacl(struct vop_setacl_args *ap) { int error; vsecattr_t vsecattr; int aclbsize; /* size of acl list in bytes */ aclent_t *aaclp; if (ap->a_type != ACL_TYPE_NFS4) return (EINVAL); if (ap->a_aclp == NULL) return (EINVAL); if (ap->a_aclp->acl_cnt < 1 || ap->a_aclp->acl_cnt > MAX_ACL_ENTRIES) return (EINVAL); /* * With NFSv4 ACLs, chmod(2) may need to add additional entries, * splitting every entry into two and appending "canonical six" * entries at the end. Don't allow for setting an ACL that would * cause chmod(2) to run out of ACL entries. */ if (ap->a_aclp->acl_cnt * 2 + 6 > ACL_MAX_ENTRIES) return (ENOSPC); error = acl_nfs4_check(ap->a_aclp, ap->a_vp->v_type == VDIR); if (error != 0) return (error); vsecattr.vsa_mask = VSA_ACE; aclbsize = ap->a_aclp->acl_cnt * sizeof (ace_t); vsecattr.vsa_aclentp = kmem_alloc(aclbsize, KM_SLEEP); aaclp = vsecattr.vsa_aclentp; vsecattr.vsa_aclentsz = aclbsize; aces_from_acl(vsecattr.vsa_aclentp, &vsecattr.vsa_aclcnt, ap->a_aclp); error = zfs_setsecattr(VTOZ(ap->a_vp), &vsecattr, 0, ap->a_cred); kmem_free(aaclp, aclbsize); return (error); } #ifndef _SYS_SYSPROTO_H_ struct vop_aclcheck_args { struct vnode *vp; acl_type_t type; struct acl *aclp; struct ucred *cred; struct thread *td; }; #endif static int zfs_freebsd_aclcheck(struct vop_aclcheck_args *ap) { return (EOPNOTSUPP); } static int zfs_vptocnp(struct vop_vptocnp_args *ap) { vnode_t *covered_vp; vnode_t *vp = ap->a_vp; zfsvfs_t *zfsvfs = vp->v_vfsp->vfs_data; znode_t *zp = VTOZ(vp); int ltype; int error; if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); /* * If we are a snapshot mounted under .zfs, run the operation * on the covered vnode. */ if (zp->z_id != zfsvfs->z_root || zfsvfs->z_parent == zfsvfs) { char name[MAXNAMLEN + 1]; znode_t *dzp; size_t len; error = zfs_znode_parent_and_name(zp, &dzp, name); if (error == 0) { len = strlen(name); if (*ap->a_buflen < len) error = SET_ERROR(ENOMEM); } if (error == 0) { *ap->a_buflen -= len; memcpy(ap->a_buf + *ap->a_buflen, name, len); *ap->a_vpp = ZTOV(dzp); } zfs_exit(zfsvfs, FTAG); return (error); } zfs_exit(zfsvfs, FTAG); covered_vp = vp->v_mount->mnt_vnodecovered; #if __FreeBSD_version >= 1300045 enum vgetstate vs = vget_prep(covered_vp); #else vhold(covered_vp); #endif ltype = VOP_ISLOCKED(vp); VOP_UNLOCK1(vp); #if __FreeBSD_version >= 1300045 error = vget_finish(covered_vp, LK_SHARED, vs); #else error = vget(covered_vp, LK_SHARED | LK_VNHELD, curthread); #endif if (error == 0) { #if __FreeBSD_version >= 1300123 error = VOP_VPTOCNP(covered_vp, ap->a_vpp, ap->a_buf, ap->a_buflen); #else error = VOP_VPTOCNP(covered_vp, ap->a_vpp, ap->a_cred, ap->a_buf, ap->a_buflen); #endif vput(covered_vp); } vn_lock(vp, ltype | LK_RETRY); if (VN_IS_DOOMED(vp)) error = SET_ERROR(ENOENT); return (error); } #if __FreeBSD_version >= 1400032 static int zfs_deallocate(struct vop_deallocate_args *ap) { znode_t *zp = VTOZ(ap->a_vp); zfsvfs_t *zfsvfs = zp->z_zfsvfs; zilog_t *zilog; off_t off, len, file_sz; int error; if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); /* * Callers might not be able to detect properly that we are read-only, * so check it explicitly here. */ if (zfs_is_readonly(zfsvfs)) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EROFS)); } zilog = zfsvfs->z_log; off = *ap->a_offset; len = *ap->a_len; file_sz = zp->z_size; if (off + len > file_sz) len = file_sz - off; /* Fast path for out-of-range request. */ if (len <= 0) { *ap->a_len = 0; zfs_exit(zfsvfs, FTAG); return (0); } error = zfs_freesp(zp, off, len, O_RDWR, TRUE); if (error == 0) { if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS || (ap->a_ioflag & IO_SYNC) != 0) zil_commit(zilog, zp->z_id); *ap->a_offset = off + len; *ap->a_len = 0; } zfs_exit(zfsvfs, FTAG); return (error); } #endif #ifndef _SYS_SYSPROTO_H_ struct vop_copy_file_range_args { struct vnode *a_invp; off_t *a_inoffp; struct vnode *a_outvp; off_t *a_outoffp; size_t *a_lenp; unsigned int a_flags; struct ucred *a_incred; struct ucred *a_outcred; struct thread *a_fsizetd; } #endif /* * TODO: FreeBSD will only call file system-specific copy_file_range() if both * files resides under the same mountpoint. In case of ZFS we want to be called * even is files are in different datasets (but on the same pools, but we need * to check that ourselves). */ static int zfs_freebsd_copy_file_range(struct vop_copy_file_range_args *ap) { zfsvfs_t *outzfsvfs; struct vnode *invp = ap->a_invp; struct vnode *outvp = ap->a_outvp; struct mount *mp; struct uio io; int error; uint64_t len = *ap->a_lenp; + if (!zfs_bclone_enabled) { + mp = NULL; + goto bad_write_fallback; + } + /* * TODO: If offset/length is not aligned to recordsize, use * vn_generic_copy_file_range() on this fragment. * It would be better to do this after we lock the vnodes, but then we * need something else than vn_generic_copy_file_range(). */ vn_start_write(outvp, &mp, V_WAIT); if (__predict_true(mp == outvp->v_mount)) { outzfsvfs = (zfsvfs_t *)mp->mnt_data; if (!spa_feature_is_enabled(dmu_objset_spa(outzfsvfs->z_os), SPA_FEATURE_BLOCK_CLONING)) { goto bad_write_fallback; } } if (invp == outvp) { if (vn_lock(outvp, LK_EXCLUSIVE) != 0) { goto bad_write_fallback; } } else { #if (__FreeBSD_version >= 1302506 && __FreeBSD_version < 1400000) || \ __FreeBSD_version >= 1400086 vn_lock_pair(invp, false, LK_EXCLUSIVE, outvp, false, LK_EXCLUSIVE); #else vn_lock_pair(invp, false, outvp, false); #endif if (VN_IS_DOOMED(invp) || VN_IS_DOOMED(outvp)) { goto bad_locked_fallback; } } #ifdef MAC error = mac_vnode_check_write(curthread->td_ucred, ap->a_outcred, outvp); if (error != 0) goto out_locked; #endif io.uio_offset = *ap->a_outoffp; io.uio_resid = *ap->a_lenp; error = vn_rlimit_fsize(outvp, &io, ap->a_fsizetd); if (error != 0) goto out_locked; error = zfs_clone_range(VTOZ(invp), ap->a_inoffp, VTOZ(outvp), ap->a_outoffp, &len, ap->a_outcred); if (error == EXDEV || error == EAGAIN || error == EINVAL || error == EOPNOTSUPP) goto bad_locked_fallback; *ap->a_lenp = (size_t)len; out_locked: if (invp != outvp) VOP_UNLOCK(invp); VOP_UNLOCK(outvp); if (mp != NULL) vn_finished_write(mp); return (error); bad_locked_fallback: if (invp != outvp) VOP_UNLOCK(invp); VOP_UNLOCK(outvp); bad_write_fallback: if (mp != NULL) vn_finished_write(mp); error = vn_generic_copy_file_range(ap->a_invp, ap->a_inoffp, ap->a_outvp, ap->a_outoffp, ap->a_lenp, ap->a_flags, ap->a_incred, ap->a_outcred, ap->a_fsizetd); return (error); } struct vop_vector zfs_vnodeops; struct vop_vector zfs_fifoops; struct vop_vector zfs_shareops; struct vop_vector zfs_vnodeops = { .vop_default = &default_vnodeops, .vop_inactive = zfs_freebsd_inactive, #if __FreeBSD_version >= 1300042 .vop_need_inactive = zfs_freebsd_need_inactive, #endif .vop_reclaim = zfs_freebsd_reclaim, #if __FreeBSD_version >= 1300102 .vop_fplookup_vexec = zfs_freebsd_fplookup_vexec, #endif #if __FreeBSD_version >= 1300139 .vop_fplookup_symlink = zfs_freebsd_fplookup_symlink, #endif .vop_access = zfs_freebsd_access, .vop_allocate = VOP_EINVAL, #if __FreeBSD_version >= 1400032 .vop_deallocate = zfs_deallocate, #endif .vop_lookup = zfs_cache_lookup, .vop_cachedlookup = zfs_freebsd_cachedlookup, .vop_getattr = zfs_freebsd_getattr, .vop_setattr = zfs_freebsd_setattr, .vop_create = zfs_freebsd_create, .vop_mknod = (vop_mknod_t *)zfs_freebsd_create, .vop_mkdir = zfs_freebsd_mkdir, .vop_readdir = zfs_freebsd_readdir, .vop_fsync = zfs_freebsd_fsync, .vop_open = zfs_freebsd_open, .vop_close = zfs_freebsd_close, .vop_rmdir = zfs_freebsd_rmdir, .vop_ioctl = zfs_freebsd_ioctl, .vop_link = zfs_freebsd_link, .vop_symlink = zfs_freebsd_symlink, .vop_readlink = zfs_freebsd_readlink, .vop_read = zfs_freebsd_read, .vop_write = zfs_freebsd_write, .vop_remove = zfs_freebsd_remove, .vop_rename = zfs_freebsd_rename, .vop_pathconf = zfs_freebsd_pathconf, .vop_bmap = zfs_freebsd_bmap, .vop_fid = zfs_freebsd_fid, .vop_getextattr = zfs_getextattr, .vop_deleteextattr = zfs_deleteextattr, .vop_setextattr = zfs_setextattr, .vop_listextattr = zfs_listextattr, .vop_getacl = zfs_freebsd_getacl, .vop_setacl = zfs_freebsd_setacl, .vop_aclcheck = zfs_freebsd_aclcheck, .vop_getpages = zfs_freebsd_getpages, .vop_putpages = zfs_freebsd_putpages, .vop_vptocnp = zfs_vptocnp, #if __FreeBSD_version >= 1300064 .vop_lock1 = vop_lock, .vop_unlock = vop_unlock, .vop_islocked = vop_islocked, #endif #if __FreeBSD_version >= 1400043 .vop_add_writecount = vop_stdadd_writecount_nomsync, #endif .vop_copy_file_range = zfs_freebsd_copy_file_range, }; VFS_VOP_VECTOR_REGISTER(zfs_vnodeops); struct vop_vector zfs_fifoops = { .vop_default = &fifo_specops, .vop_fsync = zfs_freebsd_fsync, #if __FreeBSD_version >= 1300102 .vop_fplookup_vexec = zfs_freebsd_fplookup_vexec, #endif #if __FreeBSD_version >= 1300139 .vop_fplookup_symlink = zfs_freebsd_fplookup_symlink, #endif .vop_access = zfs_freebsd_access, .vop_getattr = zfs_freebsd_getattr, .vop_inactive = zfs_freebsd_inactive, .vop_read = VOP_PANIC, .vop_reclaim = zfs_freebsd_reclaim, .vop_setattr = zfs_freebsd_setattr, .vop_write = VOP_PANIC, .vop_pathconf = zfs_freebsd_pathconf, .vop_fid = zfs_freebsd_fid, .vop_getacl = zfs_freebsd_getacl, .vop_setacl = zfs_freebsd_setacl, .vop_aclcheck = zfs_freebsd_aclcheck, #if __FreeBSD_version >= 1400043 .vop_add_writecount = vop_stdadd_writecount_nomsync, #endif }; VFS_VOP_VECTOR_REGISTER(zfs_fifoops); /* * special share hidden files vnode operations template */ struct vop_vector zfs_shareops = { .vop_default = &default_vnodeops, #if __FreeBSD_version >= 1300121 .vop_fplookup_vexec = VOP_EAGAIN, #endif #if __FreeBSD_version >= 1300139 .vop_fplookup_symlink = VOP_EAGAIN, #endif .vop_access = zfs_freebsd_access, .vop_inactive = zfs_freebsd_inactive, .vop_reclaim = zfs_freebsd_reclaim, .vop_fid = zfs_freebsd_fid, .vop_pathconf = zfs_freebsd_pathconf, #if __FreeBSD_version >= 1400043 .vop_add_writecount = vop_stdadd_writecount_nomsync, #endif }; VFS_VOP_VECTOR_REGISTER(zfs_shareops); ZFS_MODULE_PARAM(zfs, zfs_, xattr_compat, INT, ZMOD_RW, "Use legacy ZFS xattr naming for writing new user namespace xattrs"); diff --git a/module/os/linux/zfs/zfs_vnops_os.c b/module/os/linux/zfs/zfs_vnops_os.c index 1ae8023adc32..e990f7055f8a 100644 --- a/module/os/linux/zfs/zfs_vnops_os.c +++ b/module/os/linux/zfs/zfs_vnops_os.c @@ -1,4252 +1,4256 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2012, 2018 by Delphix. All rights reserved. * Copyright (c) 2015 by Chunwei Chen. All rights reserved. * Copyright 2017 Nexenta Systems, Inc. */ /* Portions Copyright 2007 Jeremy Teo */ /* Portions Copyright 2010 Robert Milkowski */ #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include /* * Programming rules. * * Each vnode op performs some logical unit of work. To do this, the ZPL must * properly lock its in-core state, create a DMU transaction, do the work, * record this work in the intent log (ZIL), commit the DMU transaction, * and wait for the intent log to commit if it is a synchronous operation. * Moreover, the vnode ops must work in both normal and log replay context. * The ordering of events is important to avoid deadlocks and references * to freed memory. The example below illustrates the following Big Rules: * * (1) A check must be made in each zfs thread for a mounted file system. * This is done avoiding races using zfs_enter(zfsvfs). * A zfs_exit(zfsvfs) is needed before all returns. Any znodes * must be checked with zfs_verify_zp(zp). Both of these macros * can return EIO from the calling function. * * (2) zrele() should always be the last thing except for zil_commit() (if * necessary) and zfs_exit(). This is for 3 reasons: First, if it's the * last reference, the vnode/znode can be freed, so the zp may point to * freed memory. Second, the last reference will call zfs_zinactive(), * which may induce a lot of work -- pushing cached pages (which acquires * range locks) and syncing out cached atime changes. Third, * zfs_zinactive() may require a new tx, which could deadlock the system * if you were already holding one. This deadlock occurs because the tx * currently being operated on prevents a txg from syncing, which * prevents the new tx from progressing, resulting in a deadlock. If you * must call zrele() within a tx, use zfs_zrele_async(). Note that iput() * is a synonym for zrele(). * * (3) All range locks must be grabbed before calling dmu_tx_assign(), * as they can span dmu_tx_assign() calls. * * (4) If ZPL locks are held, pass TXG_NOWAIT as the second argument to * dmu_tx_assign(). This is critical because we don't want to block * while holding locks. * * If no ZPL locks are held (aside from zfs_enter()), use TXG_WAIT. This * reduces lock contention and CPU usage when we must wait (note that if * throughput is constrained by the storage, nearly every transaction * must wait). * * Note, in particular, that if a lock is sometimes acquired before * the tx assigns, and sometimes after (e.g. z_lock), then failing * to use a non-blocking assign can deadlock the system. The scenario: * * Thread A has grabbed a lock before calling dmu_tx_assign(). * Thread B is in an already-assigned tx, and blocks for this lock. * Thread A calls dmu_tx_assign(TXG_WAIT) and blocks in txg_wait_open() * forever, because the previous txg can't quiesce until B's tx commits. * * If dmu_tx_assign() returns ERESTART and zfsvfs->z_assign is TXG_NOWAIT, * then drop all locks, call dmu_tx_wait(), and try again. On subsequent * calls to dmu_tx_assign(), pass TXG_NOTHROTTLE in addition to TXG_NOWAIT, * to indicate that this operation has already called dmu_tx_wait(). * This will ensure that we don't retry forever, waiting a short bit * each time. * * (5) If the operation succeeded, generate the intent log entry for it * before dropping locks. This ensures that the ordering of events * in the intent log matches the order in which they actually occurred. * During ZIL replay the zfs_log_* functions will update the sequence * number to indicate the zil transaction has replayed. * * (6) At the end of each vnode op, the DMU tx must always commit, * regardless of whether there were any errors. * * (7) After dropping all locks, invoke zil_commit(zilog, foid) * to ensure that synchronous semantics are provided when necessary. * * In general, this is how things should be ordered in each vnode op: * * zfs_enter(zfsvfs); // exit if unmounted * top: * zfs_dirent_lock(&dl, ...) // lock directory entry (may igrab()) * rw_enter(...); // grab any other locks you need * tx = dmu_tx_create(...); // get DMU tx * dmu_tx_hold_*(); // hold each object you might modify * error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT); * if (error) { * rw_exit(...); // drop locks * zfs_dirent_unlock(dl); // unlock directory entry * zrele(...); // release held znodes * if (error == ERESTART) { * waited = B_TRUE; * dmu_tx_wait(tx); * dmu_tx_abort(tx); * goto top; * } * dmu_tx_abort(tx); // abort DMU tx * zfs_exit(zfsvfs); // finished in zfs * return (error); // really out of space * } * error = do_real_work(); // do whatever this VOP does * if (error == 0) * zfs_log_*(...); // on success, make ZIL entry * dmu_tx_commit(tx); // commit DMU tx -- error or not * rw_exit(...); // drop locks * zfs_dirent_unlock(dl); // unlock directory entry * zrele(...); // release held znodes * zil_commit(zilog, foid); // synchronous when necessary * zfs_exit(zfsvfs); // finished in zfs * return (error); // done, report error */ int zfs_open(struct inode *ip, int mode, int flag, cred_t *cr) { (void) cr; znode_t *zp = ITOZ(ip); zfsvfs_t *zfsvfs = ITOZSB(ip); int error; if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); /* Honor ZFS_APPENDONLY file attribute */ if (blk_mode_is_open_write(mode) && (zp->z_pflags & ZFS_APPENDONLY) && ((flag & O_APPEND) == 0)) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EPERM)); } /* * Keep a count of the synchronous opens in the znode. On first * synchronous open we must convert all previous async transactions * into sync to keep correct ordering. */ if (flag & O_SYNC) { if (atomic_inc_32_nv(&zp->z_sync_cnt) == 1) zil_async_to_sync(zfsvfs->z_log, zp->z_id); } zfs_exit(zfsvfs, FTAG); return (0); } int zfs_close(struct inode *ip, int flag, cred_t *cr) { (void) cr; znode_t *zp = ITOZ(ip); zfsvfs_t *zfsvfs = ITOZSB(ip); int error; if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); /* Decrement the synchronous opens in the znode */ if (flag & O_SYNC) atomic_dec_32(&zp->z_sync_cnt); zfs_exit(zfsvfs, FTAG); return (0); } #if defined(_KERNEL) static int zfs_fillpage(struct inode *ip, struct page *pp); /* * When a file is memory mapped, we must keep the IO data synchronized * between the DMU cache and the memory mapped pages. Update all mapped * pages with the contents of the coresponding dmu buffer. */ void update_pages(znode_t *zp, int64_t start, int len, objset_t *os) { struct address_space *mp = ZTOI(zp)->i_mapping; int64_t off = start & (PAGE_SIZE - 1); for (start &= PAGE_MASK; len > 0; start += PAGE_SIZE) { uint64_t nbytes = MIN(PAGE_SIZE - off, len); struct page *pp = find_lock_page(mp, start >> PAGE_SHIFT); if (pp) { if (mapping_writably_mapped(mp)) flush_dcache_page(pp); void *pb = kmap(pp); int error = dmu_read(os, zp->z_id, start + off, nbytes, pb + off, DMU_READ_PREFETCH); kunmap(pp); if (error) { SetPageError(pp); ClearPageUptodate(pp); } else { ClearPageError(pp); SetPageUptodate(pp); if (mapping_writably_mapped(mp)) flush_dcache_page(pp); mark_page_accessed(pp); } unlock_page(pp); put_page(pp); } len -= nbytes; off = 0; } } /* * When a file is memory mapped, we must keep the I/O data synchronized * between the DMU cache and the memory mapped pages. Preferentially read * from memory mapped pages, otherwise fallback to reading through the dmu. */ int mappedread(znode_t *zp, int nbytes, zfs_uio_t *uio) { struct inode *ip = ZTOI(zp); struct address_space *mp = ip->i_mapping; int64_t start = uio->uio_loffset; int64_t off = start & (PAGE_SIZE - 1); int len = nbytes; int error = 0; for (start &= PAGE_MASK; len > 0; start += PAGE_SIZE) { uint64_t bytes = MIN(PAGE_SIZE - off, len); struct page *pp = find_lock_page(mp, start >> PAGE_SHIFT); if (pp) { /* * If filemap_fault() retries there exists a window * where the page will be unlocked and not up to date. * In this case we must try and fill the page. */ if (unlikely(!PageUptodate(pp))) { error = zfs_fillpage(ip, pp); if (error) { unlock_page(pp); put_page(pp); return (error); } } ASSERT(PageUptodate(pp) || PageDirty(pp)); unlock_page(pp); void *pb = kmap(pp); error = zfs_uiomove(pb + off, bytes, UIO_READ, uio); kunmap(pp); if (mapping_writably_mapped(mp)) flush_dcache_page(pp); mark_page_accessed(pp); put_page(pp); } else { error = dmu_read_uio_dbuf(sa_get_db(zp->z_sa_hdl), uio, bytes); } len -= bytes; off = 0; if (error) break; } return (error); } #endif /* _KERNEL */ static unsigned long zfs_delete_blocks = DMU_MAX_DELETEBLKCNT; /* * Write the bytes to a file. * * IN: zp - znode of file to be written to * data - bytes to write * len - number of bytes to write * pos - offset to start writing at * * OUT: resid - remaining bytes to write * * RETURN: 0 if success * positive error code if failure. EIO is returned * for a short write when residp isn't provided. * * Timestamps: * zp - ctime|mtime updated if byte count > 0 */ int zfs_write_simple(znode_t *zp, const void *data, size_t len, loff_t pos, size_t *residp) { fstrans_cookie_t cookie; int error; struct iovec iov; iov.iov_base = (void *)data; iov.iov_len = len; zfs_uio_t uio; zfs_uio_iovec_init(&uio, &iov, 1, pos, UIO_SYSSPACE, len, 0); cookie = spl_fstrans_mark(); error = zfs_write(zp, &uio, 0, kcred); spl_fstrans_unmark(cookie); if (error == 0) { if (residp != NULL) *residp = zfs_uio_resid(&uio); else if (zfs_uio_resid(&uio) != 0) error = SET_ERROR(EIO); } return (error); } static void zfs_rele_async_task(void *arg) { iput(arg); } void zfs_zrele_async(znode_t *zp) { struct inode *ip = ZTOI(zp); objset_t *os = ITOZSB(ip)->z_os; ASSERT(atomic_read(&ip->i_count) > 0); ASSERT(os != NULL); /* * If decrementing the count would put us at 0, we can't do it inline * here, because that would be synchronous. Instead, dispatch an iput * to run later. * * For more information on the dangers of a synchronous iput, see the * header comment of this file. */ if (!atomic_add_unless(&ip->i_count, -1, 1)) { VERIFY(taskq_dispatch(dsl_pool_zrele_taskq(dmu_objset_pool(os)), zfs_rele_async_task, ip, TQ_SLEEP) != TASKQID_INVALID); } } /* * Lookup an entry in a directory, or an extended attribute directory. * If it exists, return a held inode reference for it. * * IN: zdp - znode of directory to search. * nm - name of entry to lookup. * flags - LOOKUP_XATTR set if looking for an attribute. * cr - credentials of caller. * direntflags - directory lookup flags * realpnp - returned pathname. * * OUT: zpp - znode of located entry, NULL if not found. * * RETURN: 0 on success, error code on failure. * * Timestamps: * NA */ int zfs_lookup(znode_t *zdp, char *nm, znode_t **zpp, int flags, cred_t *cr, int *direntflags, pathname_t *realpnp) { zfsvfs_t *zfsvfs = ZTOZSB(zdp); int error = 0; /* * Fast path lookup, however we must skip DNLC lookup * for case folding or normalizing lookups because the * DNLC code only stores the passed in name. This means * creating 'a' and removing 'A' on a case insensitive * file system would work, but DNLC still thinks 'a' * exists and won't let you create it again on the next * pass through fast path. */ if (!(flags & (LOOKUP_XATTR | FIGNORECASE))) { if (!S_ISDIR(ZTOI(zdp)->i_mode)) { return (SET_ERROR(ENOTDIR)); } else if (zdp->z_sa_hdl == NULL) { return (SET_ERROR(EIO)); } if (nm[0] == 0 || (nm[0] == '.' && nm[1] == '\0')) { error = zfs_fastaccesschk_execute(zdp, cr); if (!error) { *zpp = zdp; zhold(*zpp); return (0); } return (error); } } if ((error = zfs_enter_verify_zp(zfsvfs, zdp, FTAG)) != 0) return (error); *zpp = NULL; if (flags & LOOKUP_XATTR) { /* * We don't allow recursive attributes.. * Maybe someday we will. */ if (zdp->z_pflags & ZFS_XATTR) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EINVAL)); } if ((error = zfs_get_xattrdir(zdp, zpp, cr, flags))) { zfs_exit(zfsvfs, FTAG); return (error); } /* * Do we have permission to get into attribute directory? */ if ((error = zfs_zaccess(*zpp, ACE_EXECUTE, 0, B_TRUE, cr, zfs_init_idmap))) { zrele(*zpp); *zpp = NULL; } zfs_exit(zfsvfs, FTAG); return (error); } if (!S_ISDIR(ZTOI(zdp)->i_mode)) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(ENOTDIR)); } /* * Check accessibility of directory. */ if ((error = zfs_zaccess(zdp, ACE_EXECUTE, 0, B_FALSE, cr, zfs_init_idmap))) { zfs_exit(zfsvfs, FTAG); return (error); } if (zfsvfs->z_utf8 && u8_validate(nm, strlen(nm), NULL, U8_VALIDATE_ENTIRE, &error) < 0) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EILSEQ)); } error = zfs_dirlook(zdp, nm, zpp, flags, direntflags, realpnp); if ((error == 0) && (*zpp)) zfs_znode_update_vfs(*zpp); zfs_exit(zfsvfs, FTAG); return (error); } /* * Attempt to create a new entry in a directory. If the entry * already exists, truncate the file if permissible, else return * an error. Return the ip of the created or trunc'd file. * * IN: dzp - znode of directory to put new file entry in. * name - name of new file entry. * vap - attributes of new file. * excl - flag indicating exclusive or non-exclusive mode. * mode - mode to open file with. * cr - credentials of caller. * flag - file flag. * vsecp - ACL to be set * mnt_ns - user namespace of the mount * * OUT: zpp - znode of created or trunc'd entry. * * RETURN: 0 on success, error code on failure. * * Timestamps: * dzp - ctime|mtime updated if new entry created * zp - ctime|mtime always, atime if new */ int zfs_create(znode_t *dzp, char *name, vattr_t *vap, int excl, int mode, znode_t **zpp, cred_t *cr, int flag, vsecattr_t *vsecp, zidmap_t *mnt_ns) { znode_t *zp; zfsvfs_t *zfsvfs = ZTOZSB(dzp); zilog_t *zilog; objset_t *os; zfs_dirlock_t *dl; dmu_tx_t *tx; int error; uid_t uid; gid_t gid; zfs_acl_ids_t acl_ids; boolean_t fuid_dirtied; boolean_t have_acl = B_FALSE; boolean_t waited = B_FALSE; boolean_t skip_acl = (flag & ATTR_NOACLCHECK) ? B_TRUE : B_FALSE; /* * If we have an ephemeral id, ACL, or XVATTR then * make sure file system is at proper version */ gid = crgetgid(cr); uid = crgetuid(cr); if (zfsvfs->z_use_fuids == B_FALSE && (vsecp || IS_EPHEMERAL(uid) || IS_EPHEMERAL(gid))) return (SET_ERROR(EINVAL)); if (name == NULL) return (SET_ERROR(EINVAL)); if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0) return (error); os = zfsvfs->z_os; zilog = zfsvfs->z_log; if (zfsvfs->z_utf8 && u8_validate(name, strlen(name), NULL, U8_VALIDATE_ENTIRE, &error) < 0) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EILSEQ)); } if (vap->va_mask & ATTR_XVATTR) { if ((error = secpolicy_xvattr((xvattr_t *)vap, crgetuid(cr), cr, vap->va_mode)) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } } top: *zpp = NULL; if (*name == '\0') { /* * Null component name refers to the directory itself. */ zhold(dzp); zp = dzp; dl = NULL; error = 0; } else { /* possible igrab(zp) */ int zflg = 0; if (flag & FIGNORECASE) zflg |= ZCILOOK; error = zfs_dirent_lock(&dl, dzp, name, &zp, zflg, NULL, NULL); if (error) { if (have_acl) zfs_acl_ids_free(&acl_ids); if (strcmp(name, "..") == 0) error = SET_ERROR(EISDIR); zfs_exit(zfsvfs, FTAG); return (error); } } if (zp == NULL) { uint64_t txtype; uint64_t projid = ZFS_DEFAULT_PROJID; /* * Create a new file object and update the directory * to reference it. */ if ((error = zfs_zaccess(dzp, ACE_ADD_FILE, 0, skip_acl, cr, mnt_ns))) { if (have_acl) zfs_acl_ids_free(&acl_ids); goto out; } /* * We only support the creation of regular files in * extended attribute directories. */ if ((dzp->z_pflags & ZFS_XATTR) && !S_ISREG(vap->va_mode)) { if (have_acl) zfs_acl_ids_free(&acl_ids); error = SET_ERROR(EINVAL); goto out; } if (!have_acl && (error = zfs_acl_ids_create(dzp, 0, vap, cr, vsecp, &acl_ids, mnt_ns)) != 0) goto out; have_acl = B_TRUE; if (S_ISREG(vap->va_mode) || S_ISDIR(vap->va_mode)) projid = zfs_inherit_projid(dzp); if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, projid)) { zfs_acl_ids_free(&acl_ids); error = SET_ERROR(EDQUOT); goto out; } tx = dmu_tx_create(os); dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes + ZFS_SA_BASE_ATTR_SIZE); fuid_dirtied = zfsvfs->z_fuid_dirty; if (fuid_dirtied) zfs_fuid_txhold(zfsvfs, tx); dmu_tx_hold_zap(tx, dzp->z_id, TRUE, name); dmu_tx_hold_sa(tx, dzp->z_sa_hdl, B_FALSE); if (!zfsvfs->z_use_sa && acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) { dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, acl_ids.z_aclp->z_acl_bytes); } error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT); if (error) { zfs_dirent_unlock(dl); if (error == ERESTART) { waited = B_TRUE; dmu_tx_wait(tx); dmu_tx_abort(tx); goto top; } zfs_acl_ids_free(&acl_ids); dmu_tx_abort(tx); zfs_exit(zfsvfs, FTAG); return (error); } zfs_mknode(dzp, vap, tx, cr, 0, &zp, &acl_ids); error = zfs_link_create(dl, zp, tx, ZNEW); if (error != 0) { /* * Since, we failed to add the directory entry for it, * delete the newly created dnode. */ zfs_znode_delete(zp, tx); remove_inode_hash(ZTOI(zp)); zfs_acl_ids_free(&acl_ids); dmu_tx_commit(tx); goto out; } if (fuid_dirtied) zfs_fuid_sync(zfsvfs, tx); txtype = zfs_log_create_txtype(Z_FILE, vsecp, vap); if (flag & FIGNORECASE) txtype |= TX_CI; zfs_log_create(zilog, tx, txtype, dzp, zp, name, vsecp, acl_ids.z_fuidp, vap); zfs_acl_ids_free(&acl_ids); dmu_tx_commit(tx); } else { int aflags = (flag & O_APPEND) ? V_APPEND : 0; if (have_acl) zfs_acl_ids_free(&acl_ids); /* * A directory entry already exists for this name. */ /* * Can't truncate an existing file if in exclusive mode. */ if (excl) { error = SET_ERROR(EEXIST); goto out; } /* * Can't open a directory for writing. */ if (S_ISDIR(ZTOI(zp)->i_mode)) { error = SET_ERROR(EISDIR); goto out; } /* * Verify requested access to file. */ if (mode && (error = zfs_zaccess_rwx(zp, mode, aflags, cr, mnt_ns))) { goto out; } mutex_enter(&dzp->z_lock); dzp->z_seq++; mutex_exit(&dzp->z_lock); /* * Truncate regular files if requested. */ if (S_ISREG(ZTOI(zp)->i_mode) && (vap->va_mask & ATTR_SIZE) && (vap->va_size == 0)) { /* we can't hold any locks when calling zfs_freesp() */ if (dl) { zfs_dirent_unlock(dl); dl = NULL; } error = zfs_freesp(zp, 0, 0, mode, TRUE); } } out: if (dl) zfs_dirent_unlock(dl); if (error) { if (zp) zrele(zp); } else { zfs_znode_update_vfs(dzp); zfs_znode_update_vfs(zp); *zpp = zp; } if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS) zil_commit(zilog, 0); zfs_exit(zfsvfs, FTAG); return (error); } int zfs_tmpfile(struct inode *dip, vattr_t *vap, int excl, int mode, struct inode **ipp, cred_t *cr, int flag, vsecattr_t *vsecp, zidmap_t *mnt_ns) { (void) excl, (void) mode, (void) flag; znode_t *zp = NULL, *dzp = ITOZ(dip); zfsvfs_t *zfsvfs = ITOZSB(dip); objset_t *os; dmu_tx_t *tx; int error; uid_t uid; gid_t gid; zfs_acl_ids_t acl_ids; uint64_t projid = ZFS_DEFAULT_PROJID; boolean_t fuid_dirtied; boolean_t have_acl = B_FALSE; boolean_t waited = B_FALSE; /* * If we have an ephemeral id, ACL, or XVATTR then * make sure file system is at proper version */ gid = crgetgid(cr); uid = crgetuid(cr); if (zfsvfs->z_use_fuids == B_FALSE && (vsecp || IS_EPHEMERAL(uid) || IS_EPHEMERAL(gid))) return (SET_ERROR(EINVAL)); if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0) return (error); os = zfsvfs->z_os; if (vap->va_mask & ATTR_XVATTR) { if ((error = secpolicy_xvattr((xvattr_t *)vap, crgetuid(cr), cr, vap->va_mode)) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } } top: *ipp = NULL; /* * Create a new file object and update the directory * to reference it. */ if ((error = zfs_zaccess(dzp, ACE_ADD_FILE, 0, B_FALSE, cr, mnt_ns))) { if (have_acl) zfs_acl_ids_free(&acl_ids); goto out; } if (!have_acl && (error = zfs_acl_ids_create(dzp, 0, vap, cr, vsecp, &acl_ids, mnt_ns)) != 0) goto out; have_acl = B_TRUE; if (S_ISREG(vap->va_mode) || S_ISDIR(vap->va_mode)) projid = zfs_inherit_projid(dzp); if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, projid)) { zfs_acl_ids_free(&acl_ids); error = SET_ERROR(EDQUOT); goto out; } tx = dmu_tx_create(os); dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes + ZFS_SA_BASE_ATTR_SIZE); dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL); fuid_dirtied = zfsvfs->z_fuid_dirty; if (fuid_dirtied) zfs_fuid_txhold(zfsvfs, tx); if (!zfsvfs->z_use_sa && acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) { dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, acl_ids.z_aclp->z_acl_bytes); } error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT); if (error) { if (error == ERESTART) { waited = B_TRUE; dmu_tx_wait(tx); dmu_tx_abort(tx); goto top; } zfs_acl_ids_free(&acl_ids); dmu_tx_abort(tx); zfs_exit(zfsvfs, FTAG); return (error); } zfs_mknode(dzp, vap, tx, cr, IS_TMPFILE, &zp, &acl_ids); if (fuid_dirtied) zfs_fuid_sync(zfsvfs, tx); /* Add to unlinked set */ zp->z_unlinked = B_TRUE; zfs_unlinked_add(zp, tx); zfs_acl_ids_free(&acl_ids); dmu_tx_commit(tx); out: if (error) { if (zp) zrele(zp); } else { zfs_znode_update_vfs(dzp); zfs_znode_update_vfs(zp); *ipp = ZTOI(zp); } zfs_exit(zfsvfs, FTAG); return (error); } /* * Remove an entry from a directory. * * IN: dzp - znode of directory to remove entry from. * name - name of entry to remove. * cr - credentials of caller. * flags - case flags. * * RETURN: 0 if success * error code if failure * * Timestamps: * dzp - ctime|mtime * ip - ctime (if nlink > 0) */ static uint64_t null_xattr = 0; int zfs_remove(znode_t *dzp, char *name, cred_t *cr, int flags) { znode_t *zp; znode_t *xzp; zfsvfs_t *zfsvfs = ZTOZSB(dzp); zilog_t *zilog; uint64_t acl_obj, xattr_obj; uint64_t xattr_obj_unlinked = 0; uint64_t obj = 0; uint64_t links; zfs_dirlock_t *dl; dmu_tx_t *tx; boolean_t may_delete_now, delete_now = FALSE; boolean_t unlinked, toobig = FALSE; uint64_t txtype; pathname_t *realnmp = NULL; pathname_t realnm; int error; int zflg = ZEXISTS; boolean_t waited = B_FALSE; if (name == NULL) return (SET_ERROR(EINVAL)); if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0) return (error); zilog = zfsvfs->z_log; if (flags & FIGNORECASE) { zflg |= ZCILOOK; pn_alloc(&realnm); realnmp = &realnm; } top: xattr_obj = 0; xzp = NULL; /* * Attempt to lock directory; fail if entry doesn't exist. */ if ((error = zfs_dirent_lock(&dl, dzp, name, &zp, zflg, NULL, realnmp))) { if (realnmp) pn_free(realnmp); zfs_exit(zfsvfs, FTAG); return (error); } if ((error = zfs_zaccess_delete(dzp, zp, cr, zfs_init_idmap))) { goto out; } /* * Need to use rmdir for removing directories. */ if (S_ISDIR(ZTOI(zp)->i_mode)) { error = SET_ERROR(EPERM); goto out; } mutex_enter(&zp->z_lock); may_delete_now = atomic_read(&ZTOI(zp)->i_count) == 1 && !zn_has_cached_data(zp, 0, LLONG_MAX); mutex_exit(&zp->z_lock); /* * We may delete the znode now, or we may put it in the unlinked set; * it depends on whether we're the last link, and on whether there are * other holds on the inode. So we dmu_tx_hold() the right things to * allow for either case. */ obj = zp->z_id; tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_zap(tx, dzp->z_id, FALSE, name); dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE); zfs_sa_upgrade_txholds(tx, zp); zfs_sa_upgrade_txholds(tx, dzp); if (may_delete_now) { toobig = zp->z_size > zp->z_blksz * zfs_delete_blocks; /* if the file is too big, only hold_free a token amount */ dmu_tx_hold_free(tx, zp->z_id, 0, (toobig ? DMU_MAX_ACCESS : DMU_OBJECT_END)); } /* are there any extended attributes? */ error = sa_lookup(zp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs), &xattr_obj, sizeof (xattr_obj)); if (error == 0 && xattr_obj) { error = zfs_zget(zfsvfs, xattr_obj, &xzp); ASSERT0(error); dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE); dmu_tx_hold_sa(tx, xzp->z_sa_hdl, B_FALSE); } mutex_enter(&zp->z_lock); if ((acl_obj = zfs_external_acl(zp)) != 0 && may_delete_now) dmu_tx_hold_free(tx, acl_obj, 0, DMU_OBJECT_END); mutex_exit(&zp->z_lock); /* charge as an update -- would be nice not to charge at all */ dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL); /* * Mark this transaction as typically resulting in a net free of space */ dmu_tx_mark_netfree(tx); error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT); if (error) { zfs_dirent_unlock(dl); if (error == ERESTART) { waited = B_TRUE; dmu_tx_wait(tx); dmu_tx_abort(tx); zrele(zp); if (xzp) zrele(xzp); goto top; } if (realnmp) pn_free(realnmp); dmu_tx_abort(tx); zrele(zp); if (xzp) zrele(xzp); zfs_exit(zfsvfs, FTAG); return (error); } /* * Remove the directory entry. */ error = zfs_link_destroy(dl, zp, tx, zflg, &unlinked); if (error) { dmu_tx_commit(tx); goto out; } if (unlinked) { /* * Hold z_lock so that we can make sure that the ACL obj * hasn't changed. Could have been deleted due to * zfs_sa_upgrade(). */ mutex_enter(&zp->z_lock); (void) sa_lookup(zp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs), &xattr_obj_unlinked, sizeof (xattr_obj_unlinked)); delete_now = may_delete_now && !toobig && atomic_read(&ZTOI(zp)->i_count) == 1 && !zn_has_cached_data(zp, 0, LLONG_MAX) && xattr_obj == xattr_obj_unlinked && zfs_external_acl(zp) == acl_obj; VERIFY_IMPLY(xattr_obj_unlinked, xzp); } if (delete_now) { if (xattr_obj_unlinked) { ASSERT3U(ZTOI(xzp)->i_nlink, ==, 2); mutex_enter(&xzp->z_lock); xzp->z_unlinked = B_TRUE; clear_nlink(ZTOI(xzp)); links = 0; error = sa_update(xzp->z_sa_hdl, SA_ZPL_LINKS(zfsvfs), &links, sizeof (links), tx); ASSERT3U(error, ==, 0); mutex_exit(&xzp->z_lock); zfs_unlinked_add(xzp, tx); if (zp->z_is_sa) error = sa_remove(zp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs), tx); else error = sa_update(zp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs), &null_xattr, sizeof (uint64_t), tx); ASSERT0(error); } /* * Add to the unlinked set because a new reference could be * taken concurrently resulting in a deferred destruction. */ zfs_unlinked_add(zp, tx); mutex_exit(&zp->z_lock); } else if (unlinked) { mutex_exit(&zp->z_lock); zfs_unlinked_add(zp, tx); } txtype = TX_REMOVE; if (flags & FIGNORECASE) txtype |= TX_CI; zfs_log_remove(zilog, tx, txtype, dzp, name, obj, unlinked); dmu_tx_commit(tx); out: if (realnmp) pn_free(realnmp); zfs_dirent_unlock(dl); zfs_znode_update_vfs(dzp); zfs_znode_update_vfs(zp); if (delete_now) zrele(zp); else zfs_zrele_async(zp); if (xzp) { zfs_znode_update_vfs(xzp); zfs_zrele_async(xzp); } if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS) zil_commit(zilog, 0); zfs_exit(zfsvfs, FTAG); return (error); } /* * Create a new directory and insert it into dzp using the name * provided. Return a pointer to the inserted directory. * * IN: dzp - znode of directory to add subdir to. * dirname - name of new directory. * vap - attributes of new directory. * cr - credentials of caller. * flags - case flags. * vsecp - ACL to be set * mnt_ns - user namespace of the mount * * OUT: zpp - znode of created directory. * * RETURN: 0 if success * error code if failure * * Timestamps: * dzp - ctime|mtime updated * zpp - ctime|mtime|atime updated */ int zfs_mkdir(znode_t *dzp, char *dirname, vattr_t *vap, znode_t **zpp, cred_t *cr, int flags, vsecattr_t *vsecp, zidmap_t *mnt_ns) { znode_t *zp; zfsvfs_t *zfsvfs = ZTOZSB(dzp); zilog_t *zilog; zfs_dirlock_t *dl; uint64_t txtype; dmu_tx_t *tx; int error; int zf = ZNEW; uid_t uid; gid_t gid = crgetgid(cr); zfs_acl_ids_t acl_ids; boolean_t fuid_dirtied; boolean_t waited = B_FALSE; ASSERT(S_ISDIR(vap->va_mode)); /* * If we have an ephemeral id, ACL, or XVATTR then * make sure file system is at proper version */ uid = crgetuid(cr); if (zfsvfs->z_use_fuids == B_FALSE && (vsecp || IS_EPHEMERAL(uid) || IS_EPHEMERAL(gid))) return (SET_ERROR(EINVAL)); if (dirname == NULL) return (SET_ERROR(EINVAL)); if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0) return (error); zilog = zfsvfs->z_log; if (dzp->z_pflags & ZFS_XATTR) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EINVAL)); } if (zfsvfs->z_utf8 && u8_validate(dirname, strlen(dirname), NULL, U8_VALIDATE_ENTIRE, &error) < 0) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EILSEQ)); } if (flags & FIGNORECASE) zf |= ZCILOOK; if (vap->va_mask & ATTR_XVATTR) { if ((error = secpolicy_xvattr((xvattr_t *)vap, crgetuid(cr), cr, vap->va_mode)) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } } if ((error = zfs_acl_ids_create(dzp, 0, vap, cr, vsecp, &acl_ids, mnt_ns)) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } /* * First make sure the new directory doesn't exist. * * Existence is checked first to make sure we don't return * EACCES instead of EEXIST which can cause some applications * to fail. */ top: *zpp = NULL; if ((error = zfs_dirent_lock(&dl, dzp, dirname, &zp, zf, NULL, NULL))) { zfs_acl_ids_free(&acl_ids); zfs_exit(zfsvfs, FTAG); return (error); } if ((error = zfs_zaccess(dzp, ACE_ADD_SUBDIRECTORY, 0, B_FALSE, cr, mnt_ns))) { zfs_acl_ids_free(&acl_ids); zfs_dirent_unlock(dl); zfs_exit(zfsvfs, FTAG); return (error); } if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, zfs_inherit_projid(dzp))) { zfs_acl_ids_free(&acl_ids); zfs_dirent_unlock(dl); zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EDQUOT)); } /* * Add a new entry to the directory. */ tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_zap(tx, dzp->z_id, TRUE, dirname); dmu_tx_hold_zap(tx, DMU_NEW_OBJECT, FALSE, NULL); fuid_dirtied = zfsvfs->z_fuid_dirty; if (fuid_dirtied) zfs_fuid_txhold(zfsvfs, tx); if (!zfsvfs->z_use_sa && acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) { dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, acl_ids.z_aclp->z_acl_bytes); } dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes + ZFS_SA_BASE_ATTR_SIZE); error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT); if (error) { zfs_dirent_unlock(dl); if (error == ERESTART) { waited = B_TRUE; dmu_tx_wait(tx); dmu_tx_abort(tx); goto top; } zfs_acl_ids_free(&acl_ids); dmu_tx_abort(tx); zfs_exit(zfsvfs, FTAG); return (error); } /* * Create new node. */ zfs_mknode(dzp, vap, tx, cr, 0, &zp, &acl_ids); /* * Now put new name in parent dir. */ error = zfs_link_create(dl, zp, tx, ZNEW); if (error != 0) { zfs_znode_delete(zp, tx); remove_inode_hash(ZTOI(zp)); goto out; } if (fuid_dirtied) zfs_fuid_sync(zfsvfs, tx); *zpp = zp; txtype = zfs_log_create_txtype(Z_DIR, vsecp, vap); if (flags & FIGNORECASE) txtype |= TX_CI; zfs_log_create(zilog, tx, txtype, dzp, zp, dirname, vsecp, acl_ids.z_fuidp, vap); out: zfs_acl_ids_free(&acl_ids); dmu_tx_commit(tx); zfs_dirent_unlock(dl); if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS) zil_commit(zilog, 0); if (error != 0) { zrele(zp); } else { zfs_znode_update_vfs(dzp); zfs_znode_update_vfs(zp); } zfs_exit(zfsvfs, FTAG); return (error); } /* * Remove a directory subdir entry. If the current working * directory is the same as the subdir to be removed, the * remove will fail. * * IN: dzp - znode of directory to remove from. * name - name of directory to be removed. * cwd - inode of current working directory. * cr - credentials of caller. * flags - case flags * * RETURN: 0 on success, error code on failure. * * Timestamps: * dzp - ctime|mtime updated */ int zfs_rmdir(znode_t *dzp, char *name, znode_t *cwd, cred_t *cr, int flags) { znode_t *zp; zfsvfs_t *zfsvfs = ZTOZSB(dzp); zilog_t *zilog; zfs_dirlock_t *dl; dmu_tx_t *tx; int error; int zflg = ZEXISTS; boolean_t waited = B_FALSE; if (name == NULL) return (SET_ERROR(EINVAL)); if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0) return (error); zilog = zfsvfs->z_log; if (flags & FIGNORECASE) zflg |= ZCILOOK; top: zp = NULL; /* * Attempt to lock directory; fail if entry doesn't exist. */ if ((error = zfs_dirent_lock(&dl, dzp, name, &zp, zflg, NULL, NULL))) { zfs_exit(zfsvfs, FTAG); return (error); } if ((error = zfs_zaccess_delete(dzp, zp, cr, zfs_init_idmap))) { goto out; } if (!S_ISDIR(ZTOI(zp)->i_mode)) { error = SET_ERROR(ENOTDIR); goto out; } if (zp == cwd) { error = SET_ERROR(EINVAL); goto out; } /* * Grab a lock on the directory to make sure that no one is * trying to add (or lookup) entries while we are removing it. */ rw_enter(&zp->z_name_lock, RW_WRITER); /* * Grab a lock on the parent pointer to make sure we play well * with the treewalk and directory rename code. */ rw_enter(&zp->z_parent_lock, RW_WRITER); tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_zap(tx, dzp->z_id, FALSE, name); dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE); dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL); zfs_sa_upgrade_txholds(tx, zp); zfs_sa_upgrade_txholds(tx, dzp); dmu_tx_mark_netfree(tx); error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT); if (error) { rw_exit(&zp->z_parent_lock); rw_exit(&zp->z_name_lock); zfs_dirent_unlock(dl); if (error == ERESTART) { waited = B_TRUE; dmu_tx_wait(tx); dmu_tx_abort(tx); zrele(zp); goto top; } dmu_tx_abort(tx); zrele(zp); zfs_exit(zfsvfs, FTAG); return (error); } error = zfs_link_destroy(dl, zp, tx, zflg, NULL); if (error == 0) { uint64_t txtype = TX_RMDIR; if (flags & FIGNORECASE) txtype |= TX_CI; zfs_log_remove(zilog, tx, txtype, dzp, name, ZFS_NO_OBJECT, B_FALSE); } dmu_tx_commit(tx); rw_exit(&zp->z_parent_lock); rw_exit(&zp->z_name_lock); out: zfs_dirent_unlock(dl); zfs_znode_update_vfs(dzp); zfs_znode_update_vfs(zp); zrele(zp); if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS) zil_commit(zilog, 0); zfs_exit(zfsvfs, FTAG); return (error); } /* * Read directory entries from the given directory cursor position and emit * name and position for each entry. * * IN: ip - inode of directory to read. * ctx - directory entry context. * cr - credentials of caller. * * RETURN: 0 if success * error code if failure * * Timestamps: * ip - atime updated * * Note that the low 4 bits of the cookie returned by zap is always zero. * This allows us to use the low range for "special" directory entries: * We use 0 for '.', and 1 for '..'. If this is the root of the filesystem, * we use the offset 2 for the '.zfs' directory. */ int zfs_readdir(struct inode *ip, zpl_dir_context_t *ctx, cred_t *cr) { (void) cr; znode_t *zp = ITOZ(ip); zfsvfs_t *zfsvfs = ITOZSB(ip); objset_t *os; zap_cursor_t zc; zap_attribute_t zap; int error; uint8_t prefetch; uint8_t type; int done = 0; uint64_t parent; uint64_t offset; /* must be unsigned; checks for < 1 */ if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_PARENT(zfsvfs), &parent, sizeof (parent))) != 0) goto out; /* * Quit if directory has been removed (posix) */ if (zp->z_unlinked) goto out; error = 0; os = zfsvfs->z_os; offset = ctx->pos; prefetch = zp->z_zn_prefetch; /* * Initialize the iterator cursor. */ if (offset <= 3) { /* * Start iteration from the beginning of the directory. */ zap_cursor_init(&zc, os, zp->z_id); } else { /* * The offset is a serialized cursor. */ zap_cursor_init_serialized(&zc, os, zp->z_id, offset); } /* * Transform to file-system independent format */ while (!done) { uint64_t objnum; /* * Special case `.', `..', and `.zfs'. */ if (offset == 0) { (void) strcpy(zap.za_name, "."); zap.za_normalization_conflict = 0; objnum = zp->z_id; type = DT_DIR; } else if (offset == 1) { (void) strcpy(zap.za_name, ".."); zap.za_normalization_conflict = 0; objnum = parent; type = DT_DIR; } else if (offset == 2 && zfs_show_ctldir(zp)) { (void) strcpy(zap.za_name, ZFS_CTLDIR_NAME); zap.za_normalization_conflict = 0; objnum = ZFSCTL_INO_ROOT; type = DT_DIR; } else { /* * Grab next entry. */ if ((error = zap_cursor_retrieve(&zc, &zap))) { if (error == ENOENT) break; else goto update; } /* * Allow multiple entries provided the first entry is * the object id. Non-zpl consumers may safely make * use of the additional space. * * XXX: This should be a feature flag for compatibility */ if (zap.za_integer_length != 8 || zap.za_num_integers == 0) { cmn_err(CE_WARN, "zap_readdir: bad directory " "entry, obj = %lld, offset = %lld, " "length = %d, num = %lld\n", (u_longlong_t)zp->z_id, (u_longlong_t)offset, zap.za_integer_length, (u_longlong_t)zap.za_num_integers); error = SET_ERROR(ENXIO); goto update; } objnum = ZFS_DIRENT_OBJ(zap.za_first_integer); type = ZFS_DIRENT_TYPE(zap.za_first_integer); } done = !zpl_dir_emit(ctx, zap.za_name, strlen(zap.za_name), objnum, type); if (done) break; if (prefetch) dmu_prefetch_dnode(os, objnum, ZIO_PRIORITY_SYNC_READ); /* * Move to the next entry, fill in the previous offset. */ if (offset > 2 || (offset == 2 && !zfs_show_ctldir(zp))) { zap_cursor_advance(&zc); offset = zap_cursor_serialize(&zc); } else { offset += 1; } ctx->pos = offset; } zp->z_zn_prefetch = B_FALSE; /* a lookup will re-enable pre-fetching */ update: zap_cursor_fini(&zc); if (error == ENOENT) error = 0; out: zfs_exit(zfsvfs, FTAG); return (error); } /* * Get the basic file attributes and place them in the provided kstat * structure. The inode is assumed to be the authoritative source * for most of the attributes. However, the znode currently has the * authoritative atime, blksize, and block count. * * IN: ip - inode of file. * * OUT: sp - kstat values. * * RETURN: 0 (always succeeds) */ int #ifdef HAVE_GENERIC_FILLATTR_IDMAP_REQMASK zfs_getattr_fast(zidmap_t *user_ns, u32 request_mask, struct inode *ip, struct kstat *sp) #else zfs_getattr_fast(zidmap_t *user_ns, struct inode *ip, struct kstat *sp) #endif { znode_t *zp = ITOZ(ip); zfsvfs_t *zfsvfs = ITOZSB(ip); uint32_t blksize; u_longlong_t nblocks; int error; if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); mutex_enter(&zp->z_lock); #ifdef HAVE_GENERIC_FILLATTR_IDMAP_REQMASK zpl_generic_fillattr(user_ns, request_mask, ip, sp); #else zpl_generic_fillattr(user_ns, ip, sp); #endif /* * +1 link count for root inode with visible '.zfs' directory. */ if ((zp->z_id == zfsvfs->z_root) && zfs_show_ctldir(zp)) if (sp->nlink < ZFS_LINK_MAX) sp->nlink++; sa_object_size(zp->z_sa_hdl, &blksize, &nblocks); sp->blksize = blksize; sp->blocks = nblocks; if (unlikely(zp->z_blksz == 0)) { /* * Block size hasn't been set; suggest maximal I/O transfers. */ sp->blksize = zfsvfs->z_max_blksz; } mutex_exit(&zp->z_lock); /* * Required to prevent NFS client from detecting different inode * numbers of snapshot root dentry before and after snapshot mount. */ if (zfsvfs->z_issnap) { if (ip->i_sb->s_root->d_inode == ip) sp->ino = ZFSCTL_INO_SNAPDIRS - dmu_objset_id(zfsvfs->z_os); } zfs_exit(zfsvfs, FTAG); return (0); } /* * For the operation of changing file's user/group/project, we need to * handle not only the main object that is assigned to the file directly, * but also the ones that are used by the file via hidden xattr directory. * * Because the xattr directory may contains many EA entries, as to it may * be impossible to change all of them via the transaction of changing the * main object's user/group/project attributes. Then we have to change them * via other multiple independent transactions one by one. It may be not good * solution, but we have no better idea yet. */ static int zfs_setattr_dir(znode_t *dzp) { struct inode *dxip = ZTOI(dzp); struct inode *xip = NULL; zfsvfs_t *zfsvfs = ZTOZSB(dzp); objset_t *os = zfsvfs->z_os; zap_cursor_t zc; zap_attribute_t zap; zfs_dirlock_t *dl; znode_t *zp = NULL; dmu_tx_t *tx = NULL; uint64_t uid, gid; sa_bulk_attr_t bulk[4]; int count; int err; zap_cursor_init(&zc, os, dzp->z_id); while ((err = zap_cursor_retrieve(&zc, &zap)) == 0) { count = 0; if (zap.za_integer_length != 8 || zap.za_num_integers != 1) { err = ENXIO; break; } err = zfs_dirent_lock(&dl, dzp, (char *)zap.za_name, &zp, ZEXISTS, NULL, NULL); if (err == ENOENT) goto next; if (err) break; xip = ZTOI(zp); if (KUID_TO_SUID(xip->i_uid) == KUID_TO_SUID(dxip->i_uid) && KGID_TO_SGID(xip->i_gid) == KGID_TO_SGID(dxip->i_gid) && zp->z_projid == dzp->z_projid) goto next; tx = dmu_tx_create(os); if (!(zp->z_pflags & ZFS_PROJID)) dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE); else dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE); err = dmu_tx_assign(tx, TXG_WAIT); if (err) break; mutex_enter(&dzp->z_lock); if (KUID_TO_SUID(xip->i_uid) != KUID_TO_SUID(dxip->i_uid)) { xip->i_uid = dxip->i_uid; uid = zfs_uid_read(dxip); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_UID(zfsvfs), NULL, &uid, sizeof (uid)); } if (KGID_TO_SGID(xip->i_gid) != KGID_TO_SGID(dxip->i_gid)) { xip->i_gid = dxip->i_gid; gid = zfs_gid_read(dxip); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_GID(zfsvfs), NULL, &gid, sizeof (gid)); } if (zp->z_projid != dzp->z_projid) { if (!(zp->z_pflags & ZFS_PROJID)) { zp->z_pflags |= ZFS_PROJID; SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL, &zp->z_pflags, sizeof (zp->z_pflags)); } zp->z_projid = dzp->z_projid; SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_PROJID(zfsvfs), NULL, &zp->z_projid, sizeof (zp->z_projid)); } mutex_exit(&dzp->z_lock); if (likely(count > 0)) { err = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx); dmu_tx_commit(tx); } else { dmu_tx_abort(tx); } tx = NULL; if (err != 0 && err != ENOENT) break; next: if (zp) { zrele(zp); zp = NULL; zfs_dirent_unlock(dl); } zap_cursor_advance(&zc); } if (tx) dmu_tx_abort(tx); if (zp) { zrele(zp); zfs_dirent_unlock(dl); } zap_cursor_fini(&zc); return (err == ENOENT ? 0 : err); } /* * Set the file attributes to the values contained in the * vattr structure. * * IN: zp - znode of file to be modified. * vap - new attribute values. * If ATTR_XVATTR set, then optional attrs are being set * flags - ATTR_UTIME set if non-default time values provided. * - ATTR_NOACLCHECK (CIFS context only). * cr - credentials of caller. * mnt_ns - user namespace of the mount * * RETURN: 0 if success * error code if failure * * Timestamps: * ip - ctime updated, mtime updated if size changed. */ int zfs_setattr(znode_t *zp, vattr_t *vap, int flags, cred_t *cr, zidmap_t *mnt_ns) { struct inode *ip; zfsvfs_t *zfsvfs = ZTOZSB(zp); objset_t *os = zfsvfs->z_os; zilog_t *zilog; dmu_tx_t *tx; vattr_t oldva; xvattr_t *tmpxvattr; uint_t mask = vap->va_mask; uint_t saved_mask = 0; int trim_mask = 0; uint64_t new_mode; uint64_t new_kuid = 0, new_kgid = 0, new_uid, new_gid; uint64_t xattr_obj; uint64_t mtime[2], ctime[2], atime[2]; uint64_t projid = ZFS_INVALID_PROJID; znode_t *attrzp; int need_policy = FALSE; int err, err2 = 0; zfs_fuid_info_t *fuidp = NULL; xvattr_t *xvap = (xvattr_t *)vap; /* vap may be an xvattr_t * */ xoptattr_t *xoap; zfs_acl_t *aclp; boolean_t skipaclchk = (flags & ATTR_NOACLCHECK) ? B_TRUE : B_FALSE; boolean_t fuid_dirtied = B_FALSE; boolean_t handle_eadir = B_FALSE; sa_bulk_attr_t *bulk, *xattr_bulk; int count = 0, xattr_count = 0, bulks = 8; if (mask == 0) return (0); if ((err = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (err); ip = ZTOI(zp); /* * If this is a xvattr_t, then get a pointer to the structure of * optional attributes. If this is NULL, then we have a vattr_t. */ xoap = xva_getxoptattr(xvap); if (xoap != NULL && (mask & ATTR_XVATTR)) { if (XVA_ISSET_REQ(xvap, XAT_PROJID)) { if (!dmu_objset_projectquota_enabled(os) || (!S_ISREG(ip->i_mode) && !S_ISDIR(ip->i_mode))) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(ENOTSUP)); } projid = xoap->xoa_projid; if (unlikely(projid == ZFS_INVALID_PROJID)) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EINVAL)); } if (projid == zp->z_projid && zp->z_pflags & ZFS_PROJID) projid = ZFS_INVALID_PROJID; else need_policy = TRUE; } if (XVA_ISSET_REQ(xvap, XAT_PROJINHERIT) && (xoap->xoa_projinherit != ((zp->z_pflags & ZFS_PROJINHERIT) != 0)) && (!dmu_objset_projectquota_enabled(os) || (!S_ISREG(ip->i_mode) && !S_ISDIR(ip->i_mode)))) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(ENOTSUP)); } } zilog = zfsvfs->z_log; /* * Make sure that if we have ephemeral uid/gid or xvattr specified * that file system is at proper version level */ if (zfsvfs->z_use_fuids == B_FALSE && (((mask & ATTR_UID) && IS_EPHEMERAL(vap->va_uid)) || ((mask & ATTR_GID) && IS_EPHEMERAL(vap->va_gid)) || (mask & ATTR_XVATTR))) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EINVAL)); } if (mask & ATTR_SIZE && S_ISDIR(ip->i_mode)) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EISDIR)); } if (mask & ATTR_SIZE && !S_ISREG(ip->i_mode) && !S_ISFIFO(ip->i_mode)) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EINVAL)); } tmpxvattr = kmem_alloc(sizeof (xvattr_t), KM_SLEEP); xva_init(tmpxvattr); bulk = kmem_alloc(sizeof (sa_bulk_attr_t) * bulks, KM_SLEEP); xattr_bulk = kmem_alloc(sizeof (sa_bulk_attr_t) * bulks, KM_SLEEP); /* * Immutable files can only alter immutable bit and atime */ if ((zp->z_pflags & ZFS_IMMUTABLE) && ((mask & (ATTR_SIZE|ATTR_UID|ATTR_GID|ATTR_MTIME|ATTR_MODE)) || ((mask & ATTR_XVATTR) && XVA_ISSET_REQ(xvap, XAT_CREATETIME)))) { err = SET_ERROR(EPERM); goto out3; } if ((mask & ATTR_SIZE) && (zp->z_pflags & ZFS_READONLY)) { err = SET_ERROR(EPERM); goto out3; } /* * Verify timestamps doesn't overflow 32 bits. * ZFS can handle large timestamps, but 32bit syscalls can't * handle times greater than 2039. This check should be removed * once large timestamps are fully supported. */ if (mask & (ATTR_ATIME | ATTR_MTIME)) { if (((mask & ATTR_ATIME) && TIMESPEC_OVERFLOW(&vap->va_atime)) || ((mask & ATTR_MTIME) && TIMESPEC_OVERFLOW(&vap->va_mtime))) { err = SET_ERROR(EOVERFLOW); goto out3; } } top: attrzp = NULL; aclp = NULL; /* Can this be moved to before the top label? */ if (zfs_is_readonly(zfsvfs)) { err = SET_ERROR(EROFS); goto out3; } /* * First validate permissions */ if (mask & ATTR_SIZE) { err = zfs_zaccess(zp, ACE_WRITE_DATA, 0, skipaclchk, cr, mnt_ns); if (err) goto out3; /* * XXX - Note, we are not providing any open * mode flags here (like FNDELAY), so we may * block if there are locks present... this * should be addressed in openat(). */ /* XXX - would it be OK to generate a log record here? */ err = zfs_freesp(zp, vap->va_size, 0, 0, FALSE); if (err) goto out3; } if (mask & (ATTR_ATIME|ATTR_MTIME) || ((mask & ATTR_XVATTR) && (XVA_ISSET_REQ(xvap, XAT_HIDDEN) || XVA_ISSET_REQ(xvap, XAT_READONLY) || XVA_ISSET_REQ(xvap, XAT_ARCHIVE) || XVA_ISSET_REQ(xvap, XAT_OFFLINE) || XVA_ISSET_REQ(xvap, XAT_SPARSE) || XVA_ISSET_REQ(xvap, XAT_CREATETIME) || XVA_ISSET_REQ(xvap, XAT_SYSTEM)))) { need_policy = zfs_zaccess(zp, ACE_WRITE_ATTRIBUTES, 0, skipaclchk, cr, mnt_ns); } if (mask & (ATTR_UID|ATTR_GID)) { int idmask = (mask & (ATTR_UID|ATTR_GID)); int take_owner; int take_group; uid_t uid; gid_t gid; /* * NOTE: even if a new mode is being set, * we may clear S_ISUID/S_ISGID bits. */ if (!(mask & ATTR_MODE)) vap->va_mode = zp->z_mode; /* * Take ownership or chgrp to group we are a member of */ uid = zfs_uid_to_vfsuid(mnt_ns, zfs_i_user_ns(ip), vap->va_uid); gid = zfs_gid_to_vfsgid(mnt_ns, zfs_i_user_ns(ip), vap->va_gid); take_owner = (mask & ATTR_UID) && (uid == crgetuid(cr)); take_group = (mask & ATTR_GID) && zfs_groupmember(zfsvfs, gid, cr); /* * If both ATTR_UID and ATTR_GID are set then take_owner and * take_group must both be set in order to allow taking * ownership. * * Otherwise, send the check through secpolicy_vnode_setattr() * */ if (((idmask == (ATTR_UID|ATTR_GID)) && take_owner && take_group) || ((idmask == ATTR_UID) && take_owner) || ((idmask == ATTR_GID) && take_group)) { if (zfs_zaccess(zp, ACE_WRITE_OWNER, 0, skipaclchk, cr, mnt_ns) == 0) { /* * Remove setuid/setgid for non-privileged users */ (void) secpolicy_setid_clear(vap, cr); trim_mask = (mask & (ATTR_UID|ATTR_GID)); } else { need_policy = TRUE; } } else { need_policy = TRUE; } } mutex_enter(&zp->z_lock); oldva.va_mode = zp->z_mode; zfs_fuid_map_ids(zp, cr, &oldva.va_uid, &oldva.va_gid); if (mask & ATTR_XVATTR) { /* * Update xvattr mask to include only those attributes * that are actually changing. * * the bits will be restored prior to actually setting * the attributes so the caller thinks they were set. */ if (XVA_ISSET_REQ(xvap, XAT_APPENDONLY)) { if (xoap->xoa_appendonly != ((zp->z_pflags & ZFS_APPENDONLY) != 0)) { need_policy = TRUE; } else { XVA_CLR_REQ(xvap, XAT_APPENDONLY); XVA_SET_REQ(tmpxvattr, XAT_APPENDONLY); } } if (XVA_ISSET_REQ(xvap, XAT_PROJINHERIT)) { if (xoap->xoa_projinherit != ((zp->z_pflags & ZFS_PROJINHERIT) != 0)) { need_policy = TRUE; } else { XVA_CLR_REQ(xvap, XAT_PROJINHERIT); XVA_SET_REQ(tmpxvattr, XAT_PROJINHERIT); } } if (XVA_ISSET_REQ(xvap, XAT_NOUNLINK)) { if (xoap->xoa_nounlink != ((zp->z_pflags & ZFS_NOUNLINK) != 0)) { need_policy = TRUE; } else { XVA_CLR_REQ(xvap, XAT_NOUNLINK); XVA_SET_REQ(tmpxvattr, XAT_NOUNLINK); } } if (XVA_ISSET_REQ(xvap, XAT_IMMUTABLE)) { if (xoap->xoa_immutable != ((zp->z_pflags & ZFS_IMMUTABLE) != 0)) { need_policy = TRUE; } else { XVA_CLR_REQ(xvap, XAT_IMMUTABLE); XVA_SET_REQ(tmpxvattr, XAT_IMMUTABLE); } } if (XVA_ISSET_REQ(xvap, XAT_NODUMP)) { if (xoap->xoa_nodump != ((zp->z_pflags & ZFS_NODUMP) != 0)) { need_policy = TRUE; } else { XVA_CLR_REQ(xvap, XAT_NODUMP); XVA_SET_REQ(tmpxvattr, XAT_NODUMP); } } if (XVA_ISSET_REQ(xvap, XAT_AV_MODIFIED)) { if (xoap->xoa_av_modified != ((zp->z_pflags & ZFS_AV_MODIFIED) != 0)) { need_policy = TRUE; } else { XVA_CLR_REQ(xvap, XAT_AV_MODIFIED); XVA_SET_REQ(tmpxvattr, XAT_AV_MODIFIED); } } if (XVA_ISSET_REQ(xvap, XAT_AV_QUARANTINED)) { if ((!S_ISREG(ip->i_mode) && xoap->xoa_av_quarantined) || xoap->xoa_av_quarantined != ((zp->z_pflags & ZFS_AV_QUARANTINED) != 0)) { need_policy = TRUE; } else { XVA_CLR_REQ(xvap, XAT_AV_QUARANTINED); XVA_SET_REQ(tmpxvattr, XAT_AV_QUARANTINED); } } if (XVA_ISSET_REQ(xvap, XAT_REPARSE)) { mutex_exit(&zp->z_lock); err = SET_ERROR(EPERM); goto out3; } if (need_policy == FALSE && (XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP) || XVA_ISSET_REQ(xvap, XAT_OPAQUE))) { need_policy = TRUE; } } mutex_exit(&zp->z_lock); if (mask & ATTR_MODE) { if (zfs_zaccess(zp, ACE_WRITE_ACL, 0, skipaclchk, cr, mnt_ns) == 0) { err = secpolicy_setid_setsticky_clear(ip, vap, &oldva, cr, mnt_ns, zfs_i_user_ns(ip)); if (err) goto out3; trim_mask |= ATTR_MODE; } else { need_policy = TRUE; } } if (need_policy) { /* * If trim_mask is set then take ownership * has been granted or write_acl is present and user * has the ability to modify mode. In that case remove * UID|GID and or MODE from mask so that * secpolicy_vnode_setattr() doesn't revoke it. */ if (trim_mask) { saved_mask = vap->va_mask; vap->va_mask &= ~trim_mask; } err = secpolicy_vnode_setattr(cr, ip, vap, &oldva, flags, zfs_zaccess_unix, zp); if (err) goto out3; if (trim_mask) vap->va_mask |= saved_mask; } /* * secpolicy_vnode_setattr, or take ownership may have * changed va_mask */ mask = vap->va_mask; if ((mask & (ATTR_UID | ATTR_GID)) || projid != ZFS_INVALID_PROJID) { handle_eadir = B_TRUE; err = sa_lookup(zp->z_sa_hdl, SA_ZPL_XATTR(zfsvfs), &xattr_obj, sizeof (xattr_obj)); if (err == 0 && xattr_obj) { err = zfs_zget(ZTOZSB(zp), xattr_obj, &attrzp); if (err) goto out2; } if (mask & ATTR_UID) { new_kuid = zfs_fuid_create(zfsvfs, (uint64_t)vap->va_uid, cr, ZFS_OWNER, &fuidp); if (new_kuid != KUID_TO_SUID(ZTOI(zp)->i_uid) && zfs_id_overquota(zfsvfs, DMU_USERUSED_OBJECT, new_kuid)) { if (attrzp) zrele(attrzp); err = SET_ERROR(EDQUOT); goto out2; } } if (mask & ATTR_GID) { new_kgid = zfs_fuid_create(zfsvfs, (uint64_t)vap->va_gid, cr, ZFS_GROUP, &fuidp); if (new_kgid != KGID_TO_SGID(ZTOI(zp)->i_gid) && zfs_id_overquota(zfsvfs, DMU_GROUPUSED_OBJECT, new_kgid)) { if (attrzp) zrele(attrzp); err = SET_ERROR(EDQUOT); goto out2; } } if (projid != ZFS_INVALID_PROJID && zfs_id_overquota(zfsvfs, DMU_PROJECTUSED_OBJECT, projid)) { if (attrzp) zrele(attrzp); err = EDQUOT; goto out2; } } tx = dmu_tx_create(os); if (mask & ATTR_MODE) { uint64_t pmode = zp->z_mode; uint64_t acl_obj; new_mode = (pmode & S_IFMT) | (vap->va_mode & ~S_IFMT); if (ZTOZSB(zp)->z_acl_mode == ZFS_ACL_RESTRICTED && !(zp->z_pflags & ZFS_ACL_TRIVIAL)) { err = EPERM; goto out; } if ((err = zfs_acl_chmod_setattr(zp, &aclp, new_mode))) goto out; mutex_enter(&zp->z_lock); if (!zp->z_is_sa && ((acl_obj = zfs_external_acl(zp)) != 0)) { /* * Are we upgrading ACL from old V0 format * to V1 format? */ if (zfsvfs->z_version >= ZPL_VERSION_FUID && zfs_znode_acl_version(zp) == ZFS_ACL_VERSION_INITIAL) { dmu_tx_hold_free(tx, acl_obj, 0, DMU_OBJECT_END); dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, aclp->z_acl_bytes); } else { dmu_tx_hold_write(tx, acl_obj, 0, aclp->z_acl_bytes); } } else if (!zp->z_is_sa && aclp->z_acl_bytes > ZFS_ACE_SPACE) { dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, aclp->z_acl_bytes); } mutex_exit(&zp->z_lock); dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE); } else { if (((mask & ATTR_XVATTR) && XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP)) || (projid != ZFS_INVALID_PROJID && !(zp->z_pflags & ZFS_PROJID))) dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_TRUE); else dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE); } if (attrzp) { dmu_tx_hold_sa(tx, attrzp->z_sa_hdl, B_FALSE); } fuid_dirtied = zfsvfs->z_fuid_dirty; if (fuid_dirtied) zfs_fuid_txhold(zfsvfs, tx); zfs_sa_upgrade_txholds(tx, zp); err = dmu_tx_assign(tx, TXG_WAIT); if (err) goto out; count = 0; /* * Set each attribute requested. * We group settings according to the locks they need to acquire. * * Note: you cannot set ctime directly, although it will be * updated as a side-effect of calling this function. */ if (projid != ZFS_INVALID_PROJID && !(zp->z_pflags & ZFS_PROJID)) { /* * For the existed object that is upgraded from old system, * its on-disk layout has no slot for the project ID attribute. * But quota accounting logic needs to access related slots by * offset directly. So we need to adjust old objects' layout * to make the project ID to some unified and fixed offset. */ if (attrzp) err = sa_add_projid(attrzp->z_sa_hdl, tx, projid); if (err == 0) err = sa_add_projid(zp->z_sa_hdl, tx, projid); if (unlikely(err == EEXIST)) err = 0; else if (err != 0) goto out; else projid = ZFS_INVALID_PROJID; } if (mask & (ATTR_UID|ATTR_GID|ATTR_MODE)) mutex_enter(&zp->z_acl_lock); mutex_enter(&zp->z_lock); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL, &zp->z_pflags, sizeof (zp->z_pflags)); if (attrzp) { if (mask & (ATTR_UID|ATTR_GID|ATTR_MODE)) mutex_enter(&attrzp->z_acl_lock); mutex_enter(&attrzp->z_lock); SA_ADD_BULK_ATTR(xattr_bulk, xattr_count, SA_ZPL_FLAGS(zfsvfs), NULL, &attrzp->z_pflags, sizeof (attrzp->z_pflags)); if (projid != ZFS_INVALID_PROJID) { attrzp->z_projid = projid; SA_ADD_BULK_ATTR(xattr_bulk, xattr_count, SA_ZPL_PROJID(zfsvfs), NULL, &attrzp->z_projid, sizeof (attrzp->z_projid)); } } if (mask & (ATTR_UID|ATTR_GID)) { if (mask & ATTR_UID) { ZTOI(zp)->i_uid = SUID_TO_KUID(new_kuid); new_uid = zfs_uid_read(ZTOI(zp)); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_UID(zfsvfs), NULL, &new_uid, sizeof (new_uid)); if (attrzp) { SA_ADD_BULK_ATTR(xattr_bulk, xattr_count, SA_ZPL_UID(zfsvfs), NULL, &new_uid, sizeof (new_uid)); ZTOI(attrzp)->i_uid = SUID_TO_KUID(new_uid); } } if (mask & ATTR_GID) { ZTOI(zp)->i_gid = SGID_TO_KGID(new_kgid); new_gid = zfs_gid_read(ZTOI(zp)); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_GID(zfsvfs), NULL, &new_gid, sizeof (new_gid)); if (attrzp) { SA_ADD_BULK_ATTR(xattr_bulk, xattr_count, SA_ZPL_GID(zfsvfs), NULL, &new_gid, sizeof (new_gid)); ZTOI(attrzp)->i_gid = SGID_TO_KGID(new_kgid); } } if (!(mask & ATTR_MODE)) { SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MODE(zfsvfs), NULL, &new_mode, sizeof (new_mode)); new_mode = zp->z_mode; } err = zfs_acl_chown_setattr(zp); ASSERT(err == 0); if (attrzp) { err = zfs_acl_chown_setattr(attrzp); ASSERT(err == 0); } } if (mask & ATTR_MODE) { SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MODE(zfsvfs), NULL, &new_mode, sizeof (new_mode)); zp->z_mode = ZTOI(zp)->i_mode = new_mode; ASSERT3P(aclp, !=, NULL); err = zfs_aclset_common(zp, aclp, cr, tx); ASSERT0(err); if (zp->z_acl_cached) zfs_acl_free(zp->z_acl_cached); zp->z_acl_cached = aclp; aclp = NULL; } if ((mask & ATTR_ATIME) || zp->z_atime_dirty) { zp->z_atime_dirty = B_FALSE; ZFS_TIME_ENCODE(&ip->i_atime, atime); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_ATIME(zfsvfs), NULL, &atime, sizeof (atime)); } if (mask & (ATTR_MTIME | ATTR_SIZE)) { ZFS_TIME_ENCODE(&vap->va_mtime, mtime); ZTOI(zp)->i_mtime = zpl_inode_timestamp_truncate( vap->va_mtime, ZTOI(zp)); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs), NULL, mtime, sizeof (mtime)); } if (mask & (ATTR_CTIME | ATTR_SIZE)) { ZFS_TIME_ENCODE(&vap->va_ctime, ctime); zpl_inode_set_ctime_to_ts(ZTOI(zp), zpl_inode_timestamp_truncate(vap->va_ctime, ZTOI(zp))); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL, ctime, sizeof (ctime)); } if (projid != ZFS_INVALID_PROJID) { zp->z_projid = projid; SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_PROJID(zfsvfs), NULL, &zp->z_projid, sizeof (zp->z_projid)); } if (attrzp && mask) { SA_ADD_BULK_ATTR(xattr_bulk, xattr_count, SA_ZPL_CTIME(zfsvfs), NULL, &ctime, sizeof (ctime)); } /* * Do this after setting timestamps to prevent timestamp * update from toggling bit */ if (xoap && (mask & ATTR_XVATTR)) { /* * restore trimmed off masks * so that return masks can be set for caller. */ if (XVA_ISSET_REQ(tmpxvattr, XAT_APPENDONLY)) { XVA_SET_REQ(xvap, XAT_APPENDONLY); } if (XVA_ISSET_REQ(tmpxvattr, XAT_NOUNLINK)) { XVA_SET_REQ(xvap, XAT_NOUNLINK); } if (XVA_ISSET_REQ(tmpxvattr, XAT_IMMUTABLE)) { XVA_SET_REQ(xvap, XAT_IMMUTABLE); } if (XVA_ISSET_REQ(tmpxvattr, XAT_NODUMP)) { XVA_SET_REQ(xvap, XAT_NODUMP); } if (XVA_ISSET_REQ(tmpxvattr, XAT_AV_MODIFIED)) { XVA_SET_REQ(xvap, XAT_AV_MODIFIED); } if (XVA_ISSET_REQ(tmpxvattr, XAT_AV_QUARANTINED)) { XVA_SET_REQ(xvap, XAT_AV_QUARANTINED); } if (XVA_ISSET_REQ(tmpxvattr, XAT_PROJINHERIT)) { XVA_SET_REQ(xvap, XAT_PROJINHERIT); } if (XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP)) ASSERT(S_ISREG(ip->i_mode)); zfs_xvattr_set(zp, xvap, tx); } if (fuid_dirtied) zfs_fuid_sync(zfsvfs, tx); if (mask != 0) zfs_log_setattr(zilog, tx, TX_SETATTR, zp, vap, mask, fuidp); mutex_exit(&zp->z_lock); if (mask & (ATTR_UID|ATTR_GID|ATTR_MODE)) mutex_exit(&zp->z_acl_lock); if (attrzp) { if (mask & (ATTR_UID|ATTR_GID|ATTR_MODE)) mutex_exit(&attrzp->z_acl_lock); mutex_exit(&attrzp->z_lock); } out: if (err == 0 && xattr_count > 0) { err2 = sa_bulk_update(attrzp->z_sa_hdl, xattr_bulk, xattr_count, tx); ASSERT(err2 == 0); } if (aclp) zfs_acl_free(aclp); if (fuidp) { zfs_fuid_info_free(fuidp); fuidp = NULL; } if (err) { dmu_tx_abort(tx); if (attrzp) zrele(attrzp); if (err == ERESTART) goto top; } else { if (count > 0) err2 = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx); dmu_tx_commit(tx); if (attrzp) { if (err2 == 0 && handle_eadir) err = zfs_setattr_dir(attrzp); zrele(attrzp); } zfs_znode_update_vfs(zp); } out2: if (os->os_sync == ZFS_SYNC_ALWAYS) zil_commit(zilog, 0); out3: kmem_free(xattr_bulk, sizeof (sa_bulk_attr_t) * bulks); kmem_free(bulk, sizeof (sa_bulk_attr_t) * bulks); kmem_free(tmpxvattr, sizeof (xvattr_t)); zfs_exit(zfsvfs, FTAG); return (err); } typedef struct zfs_zlock { krwlock_t *zl_rwlock; /* lock we acquired */ znode_t *zl_znode; /* znode we held */ struct zfs_zlock *zl_next; /* next in list */ } zfs_zlock_t; /* * Drop locks and release vnodes that were held by zfs_rename_lock(). */ static void zfs_rename_unlock(zfs_zlock_t **zlpp) { zfs_zlock_t *zl; while ((zl = *zlpp) != NULL) { if (zl->zl_znode != NULL) zfs_zrele_async(zl->zl_znode); rw_exit(zl->zl_rwlock); *zlpp = zl->zl_next; kmem_free(zl, sizeof (*zl)); } } /* * Search back through the directory tree, using the ".." entries. * Lock each directory in the chain to prevent concurrent renames. * Fail any attempt to move a directory into one of its own descendants. * XXX - z_parent_lock can overlap with map or grow locks */ static int zfs_rename_lock(znode_t *szp, znode_t *tdzp, znode_t *sdzp, zfs_zlock_t **zlpp) { zfs_zlock_t *zl; znode_t *zp = tdzp; uint64_t rootid = ZTOZSB(zp)->z_root; uint64_t oidp = zp->z_id; krwlock_t *rwlp = &szp->z_parent_lock; krw_t rw = RW_WRITER; /* * First pass write-locks szp and compares to zp->z_id. * Later passes read-lock zp and compare to zp->z_parent. */ do { if (!rw_tryenter(rwlp, rw)) { /* * Another thread is renaming in this path. * Note that if we are a WRITER, we don't have any * parent_locks held yet. */ if (rw == RW_READER && zp->z_id > szp->z_id) { /* * Drop our locks and restart */ zfs_rename_unlock(&zl); *zlpp = NULL; zp = tdzp; oidp = zp->z_id; rwlp = &szp->z_parent_lock; rw = RW_WRITER; continue; } else { /* * Wait for other thread to drop its locks */ rw_enter(rwlp, rw); } } zl = kmem_alloc(sizeof (*zl), KM_SLEEP); zl->zl_rwlock = rwlp; zl->zl_znode = NULL; zl->zl_next = *zlpp; *zlpp = zl; if (oidp == szp->z_id) /* We're a descendant of szp */ return (SET_ERROR(EINVAL)); if (oidp == rootid) /* We've hit the top */ return (0); if (rw == RW_READER) { /* i.e. not the first pass */ int error = zfs_zget(ZTOZSB(zp), oidp, &zp); if (error) return (error); zl->zl_znode = zp; } (void) sa_lookup(zp->z_sa_hdl, SA_ZPL_PARENT(ZTOZSB(zp)), &oidp, sizeof (oidp)); rwlp = &zp->z_parent_lock; rw = RW_READER; } while (zp->z_id != sdzp->z_id); return (0); } /* * Move an entry from the provided source directory to the target * directory. Change the entry name as indicated. * * IN: sdzp - Source directory containing the "old entry". * snm - Old entry name. * tdzp - Target directory to contain the "new entry". * tnm - New entry name. * cr - credentials of caller. * flags - case flags * rflags - RENAME_* flags * wa_vap - attributes for RENAME_WHITEOUT (must be a char 0:0). * mnt_ns - user namespace of the mount * * RETURN: 0 on success, error code on failure. * * Timestamps: * sdzp,tdzp - ctime|mtime updated */ int zfs_rename(znode_t *sdzp, char *snm, znode_t *tdzp, char *tnm, cred_t *cr, int flags, uint64_t rflags, vattr_t *wo_vap, zidmap_t *mnt_ns) { znode_t *szp, *tzp; zfsvfs_t *zfsvfs = ZTOZSB(sdzp); zilog_t *zilog; zfs_dirlock_t *sdl, *tdl; dmu_tx_t *tx; zfs_zlock_t *zl; int cmp, serr, terr; int error = 0; int zflg = 0; boolean_t waited = B_FALSE; /* Needed for whiteout inode creation. */ boolean_t fuid_dirtied; zfs_acl_ids_t acl_ids; boolean_t have_acl = B_FALSE; znode_t *wzp = NULL; if (snm == NULL || tnm == NULL) return (SET_ERROR(EINVAL)); if (rflags & ~(RENAME_NOREPLACE | RENAME_EXCHANGE | RENAME_WHITEOUT)) return (SET_ERROR(EINVAL)); /* Already checked by Linux VFS, but just to make sure. */ if (rflags & RENAME_EXCHANGE && (rflags & (RENAME_NOREPLACE | RENAME_WHITEOUT))) return (SET_ERROR(EINVAL)); /* * Make sure we only get wo_vap iff. RENAME_WHITEOUT and that it's the * right kind of vattr_t for the whiteout file. These are set * internally by ZFS so should never be incorrect. */ VERIFY_EQUIV(rflags & RENAME_WHITEOUT, wo_vap != NULL); VERIFY_IMPLY(wo_vap, wo_vap->va_mode == S_IFCHR); VERIFY_IMPLY(wo_vap, wo_vap->va_rdev == makedevice(0, 0)); if ((error = zfs_enter_verify_zp(zfsvfs, sdzp, FTAG)) != 0) return (error); zilog = zfsvfs->z_log; if ((error = zfs_verify_zp(tdzp)) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } /* * We check i_sb because snapshots and the ctldir must have different * super blocks. */ if (ZTOI(tdzp)->i_sb != ZTOI(sdzp)->i_sb || zfsctl_is_node(ZTOI(tdzp))) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EXDEV)); } if (zfsvfs->z_utf8 && u8_validate(tnm, strlen(tnm), NULL, U8_VALIDATE_ENTIRE, &error) < 0) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EILSEQ)); } if (flags & FIGNORECASE) zflg |= ZCILOOK; top: szp = NULL; tzp = NULL; zl = NULL; /* * This is to prevent the creation of links into attribute space * by renaming a linked file into/outof an attribute directory. * See the comment in zfs_link() for why this is considered bad. */ if ((tdzp->z_pflags & ZFS_XATTR) != (sdzp->z_pflags & ZFS_XATTR)) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EINVAL)); } /* * Lock source and target directory entries. To prevent deadlock, * a lock ordering must be defined. We lock the directory with * the smallest object id first, or if it's a tie, the one with * the lexically first name. */ if (sdzp->z_id < tdzp->z_id) { cmp = -1; } else if (sdzp->z_id > tdzp->z_id) { cmp = 1; } else { /* * First compare the two name arguments without * considering any case folding. */ int nofold = (zfsvfs->z_norm & ~U8_TEXTPREP_TOUPPER); cmp = u8_strcmp(snm, tnm, 0, nofold, U8_UNICODE_LATEST, &error); ASSERT(error == 0 || !zfsvfs->z_utf8); if (cmp == 0) { /* * POSIX: "If the old argument and the new argument * both refer to links to the same existing file, * the rename() function shall return successfully * and perform no other action." */ zfs_exit(zfsvfs, FTAG); return (0); } /* * If the file system is case-folding, then we may * have some more checking to do. A case-folding file * system is either supporting mixed case sensitivity * access or is completely case-insensitive. Note * that the file system is always case preserving. * * In mixed sensitivity mode case sensitive behavior * is the default. FIGNORECASE must be used to * explicitly request case insensitive behavior. * * If the source and target names provided differ only * by case (e.g., a request to rename 'tim' to 'Tim'), * we will treat this as a special case in the * case-insensitive mode: as long as the source name * is an exact match, we will allow this to proceed as * a name-change request. */ if ((zfsvfs->z_case == ZFS_CASE_INSENSITIVE || (zfsvfs->z_case == ZFS_CASE_MIXED && flags & FIGNORECASE)) && u8_strcmp(snm, tnm, 0, zfsvfs->z_norm, U8_UNICODE_LATEST, &error) == 0) { /* * case preserving rename request, require exact * name matches */ zflg |= ZCIEXACT; zflg &= ~ZCILOOK; } } /* * If the source and destination directories are the same, we should * grab the z_name_lock of that directory only once. */ if (sdzp == tdzp) { zflg |= ZHAVELOCK; rw_enter(&sdzp->z_name_lock, RW_READER); } if (cmp < 0) { serr = zfs_dirent_lock(&sdl, sdzp, snm, &szp, ZEXISTS | zflg, NULL, NULL); terr = zfs_dirent_lock(&tdl, tdzp, tnm, &tzp, ZRENAMING | zflg, NULL, NULL); } else { terr = zfs_dirent_lock(&tdl, tdzp, tnm, &tzp, zflg, NULL, NULL); serr = zfs_dirent_lock(&sdl, sdzp, snm, &szp, ZEXISTS | ZRENAMING | zflg, NULL, NULL); } if (serr) { /* * Source entry invalid or not there. */ if (!terr) { zfs_dirent_unlock(tdl); if (tzp) zrele(tzp); } if (sdzp == tdzp) rw_exit(&sdzp->z_name_lock); if (strcmp(snm, "..") == 0) serr = EINVAL; zfs_exit(zfsvfs, FTAG); return (serr); } if (terr) { zfs_dirent_unlock(sdl); zrele(szp); if (sdzp == tdzp) rw_exit(&sdzp->z_name_lock); if (strcmp(tnm, "..") == 0) terr = EINVAL; zfs_exit(zfsvfs, FTAG); return (terr); } /* * If we are using project inheritance, means if the directory has * ZFS_PROJINHERIT set, then its descendant directories will inherit * not only the project ID, but also the ZFS_PROJINHERIT flag. Under * such case, we only allow renames into our tree when the project * IDs are the same. */ if (tdzp->z_pflags & ZFS_PROJINHERIT && tdzp->z_projid != szp->z_projid) { error = SET_ERROR(EXDEV); goto out; } /* * Must have write access at the source to remove the old entry * and write access at the target to create the new entry. * Note that if target and source are the same, this can be * done in a single check. */ if ((error = zfs_zaccess_rename(sdzp, szp, tdzp, tzp, cr, mnt_ns))) goto out; if (S_ISDIR(ZTOI(szp)->i_mode)) { /* * Check to make sure rename is valid. * Can't do a move like this: /usr/a/b to /usr/a/b/c/d */ if ((error = zfs_rename_lock(szp, tdzp, sdzp, &zl))) goto out; } /* * Does target exist? */ if (tzp) { if (rflags & RENAME_NOREPLACE) { error = SET_ERROR(EEXIST); goto out; } /* * Source and target must be the same type (unless exchanging). */ if (!(rflags & RENAME_EXCHANGE)) { boolean_t s_is_dir = S_ISDIR(ZTOI(szp)->i_mode) != 0; boolean_t t_is_dir = S_ISDIR(ZTOI(tzp)->i_mode) != 0; if (s_is_dir != t_is_dir) { error = SET_ERROR(s_is_dir ? ENOTDIR : EISDIR); goto out; } } /* * POSIX dictates that when the source and target * entries refer to the same file object, rename * must do nothing and exit without error. */ if (szp->z_id == tzp->z_id) { error = 0; goto out; } } else if (rflags & RENAME_EXCHANGE) { /* Target must exist for RENAME_EXCHANGE. */ error = SET_ERROR(ENOENT); goto out; } /* Set up inode creation for RENAME_WHITEOUT. */ if (rflags & RENAME_WHITEOUT) { /* * Whiteout files are not regular files or directories, so to * match zfs_create() we do not inherit the project id. */ uint64_t wo_projid = ZFS_DEFAULT_PROJID; error = zfs_zaccess(sdzp, ACE_ADD_FILE, 0, B_FALSE, cr, mnt_ns); if (error) goto out; if (!have_acl) { error = zfs_acl_ids_create(sdzp, 0, wo_vap, cr, NULL, &acl_ids, mnt_ns); if (error) goto out; have_acl = B_TRUE; } if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, wo_projid)) { error = SET_ERROR(EDQUOT); goto out; } } tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_sa(tx, szp->z_sa_hdl, B_FALSE); dmu_tx_hold_sa(tx, sdzp->z_sa_hdl, B_FALSE); dmu_tx_hold_zap(tx, sdzp->z_id, (rflags & RENAME_EXCHANGE) ? TRUE : FALSE, snm); dmu_tx_hold_zap(tx, tdzp->z_id, TRUE, tnm); if (sdzp != tdzp) { dmu_tx_hold_sa(tx, tdzp->z_sa_hdl, B_FALSE); zfs_sa_upgrade_txholds(tx, tdzp); } if (tzp) { dmu_tx_hold_sa(tx, tzp->z_sa_hdl, B_FALSE); zfs_sa_upgrade_txholds(tx, tzp); } if (rflags & RENAME_WHITEOUT) { dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes + ZFS_SA_BASE_ATTR_SIZE); dmu_tx_hold_zap(tx, sdzp->z_id, TRUE, snm); dmu_tx_hold_sa(tx, sdzp->z_sa_hdl, B_FALSE); if (!zfsvfs->z_use_sa && acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) { dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, acl_ids.z_aclp->z_acl_bytes); } } fuid_dirtied = zfsvfs->z_fuid_dirty; if (fuid_dirtied) zfs_fuid_txhold(zfsvfs, tx); zfs_sa_upgrade_txholds(tx, szp); dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL); error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT); if (error) { if (zl != NULL) zfs_rename_unlock(&zl); zfs_dirent_unlock(sdl); zfs_dirent_unlock(tdl); if (sdzp == tdzp) rw_exit(&sdzp->z_name_lock); if (error == ERESTART) { waited = B_TRUE; dmu_tx_wait(tx); dmu_tx_abort(tx); zrele(szp); if (tzp) zrele(tzp); goto top; } dmu_tx_abort(tx); zrele(szp); if (tzp) zrele(tzp); zfs_exit(zfsvfs, FTAG); return (error); } /* * Unlink the source. */ szp->z_pflags |= ZFS_AV_MODIFIED; if (tdzp->z_pflags & ZFS_PROJINHERIT) szp->z_pflags |= ZFS_PROJINHERIT; error = sa_update(szp->z_sa_hdl, SA_ZPL_FLAGS(zfsvfs), (void *)&szp->z_pflags, sizeof (uint64_t), tx); VERIFY0(error); error = zfs_link_destroy(sdl, szp, tx, ZRENAMING, NULL); if (error) goto commit; /* * Unlink the target. */ if (tzp) { int tzflg = zflg; if (rflags & RENAME_EXCHANGE) { /* This inode will be re-linked soon. */ tzflg |= ZRENAMING; tzp->z_pflags |= ZFS_AV_MODIFIED; if (sdzp->z_pflags & ZFS_PROJINHERIT) tzp->z_pflags |= ZFS_PROJINHERIT; error = sa_update(tzp->z_sa_hdl, SA_ZPL_FLAGS(zfsvfs), (void *)&tzp->z_pflags, sizeof (uint64_t), tx); ASSERT0(error); } error = zfs_link_destroy(tdl, tzp, tx, tzflg, NULL); if (error) goto commit_link_szp; } /* * Create the new target links: * * We always link the target. * * RENAME_EXCHANGE: Link the old target to the source. * * RENAME_WHITEOUT: Create a whiteout inode in-place of the source. */ error = zfs_link_create(tdl, szp, tx, ZRENAMING); if (error) { /* * If we have removed the existing target, a subsequent call to * zfs_link_create() to add back the same entry, but with a new * dnode (szp), should not fail. */ ASSERT3P(tzp, ==, NULL); goto commit_link_tzp; } switch (rflags & (RENAME_EXCHANGE | RENAME_WHITEOUT)) { case RENAME_EXCHANGE: error = zfs_link_create(sdl, tzp, tx, ZRENAMING); /* * The same argument as zfs_link_create() failing for * szp applies here, since the source directory must * have had an entry we are replacing. */ ASSERT0(error); if (error) goto commit_unlink_td_szp; break; case RENAME_WHITEOUT: zfs_mknode(sdzp, wo_vap, tx, cr, 0, &wzp, &acl_ids); error = zfs_link_create(sdl, wzp, tx, ZNEW); if (error) { zfs_znode_delete(wzp, tx); remove_inode_hash(ZTOI(wzp)); goto commit_unlink_td_szp; } break; } if (fuid_dirtied) zfs_fuid_sync(zfsvfs, tx); switch (rflags & (RENAME_EXCHANGE | RENAME_WHITEOUT)) { case RENAME_EXCHANGE: zfs_log_rename_exchange(zilog, tx, (flags & FIGNORECASE ? TX_CI : 0), sdzp, sdl->dl_name, tdzp, tdl->dl_name, szp); break; case RENAME_WHITEOUT: zfs_log_rename_whiteout(zilog, tx, (flags & FIGNORECASE ? TX_CI : 0), sdzp, sdl->dl_name, tdzp, tdl->dl_name, szp, wzp); break; default: ASSERT0(rflags & ~RENAME_NOREPLACE); zfs_log_rename(zilog, tx, (flags & FIGNORECASE ? TX_CI : 0), sdzp, sdl->dl_name, tdzp, tdl->dl_name, szp); break; } commit: dmu_tx_commit(tx); out: if (have_acl) zfs_acl_ids_free(&acl_ids); zfs_znode_update_vfs(sdzp); if (sdzp == tdzp) rw_exit(&sdzp->z_name_lock); if (sdzp != tdzp) zfs_znode_update_vfs(tdzp); zfs_znode_update_vfs(szp); zrele(szp); if (wzp) { zfs_znode_update_vfs(wzp); zrele(wzp); } if (tzp) { zfs_znode_update_vfs(tzp); zrele(tzp); } if (zl != NULL) zfs_rename_unlock(&zl); zfs_dirent_unlock(sdl); zfs_dirent_unlock(tdl); if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS) zil_commit(zilog, 0); zfs_exit(zfsvfs, FTAG); return (error); /* * Clean-up path for broken link state. * * At this point we are in a (very) bad state, so we need to do our * best to correct the state. In particular, all of the nlinks are * wrong because we were destroying and creating links with ZRENAMING. * * In some form, all of these operations have to resolve the state: * * * link_destroy() *must* succeed. Fortunately, this is very likely * since we only just created it. * * * link_create()s are allowed to fail (though they shouldn't because * we only just unlinked them and are putting the entries back * during clean-up). But if they fail, we can just forcefully drop * the nlink value to (at the very least) avoid broken nlink values * -- though in the case of non-empty directories we will have to * panic (otherwise we'd have a leaked directory with a broken ..). */ commit_unlink_td_szp: VERIFY0(zfs_link_destroy(tdl, szp, tx, ZRENAMING, NULL)); commit_link_tzp: if (tzp) { if (zfs_link_create(tdl, tzp, tx, ZRENAMING)) VERIFY0(zfs_drop_nlink(tzp, tx, NULL)); } commit_link_szp: if (zfs_link_create(sdl, szp, tx, ZRENAMING)) VERIFY0(zfs_drop_nlink(szp, tx, NULL)); goto commit; } /* * Insert the indicated symbolic reference entry into the directory. * * IN: dzp - Directory to contain new symbolic link. * name - Name of directory entry in dip. * vap - Attributes of new entry. * link - Name for new symlink entry. * cr - credentials of caller. * flags - case flags * mnt_ns - user namespace of the mount * * OUT: zpp - Znode for new symbolic link. * * RETURN: 0 on success, error code on failure. * * Timestamps: * dip - ctime|mtime updated */ int zfs_symlink(znode_t *dzp, char *name, vattr_t *vap, char *link, znode_t **zpp, cred_t *cr, int flags, zidmap_t *mnt_ns) { znode_t *zp; zfs_dirlock_t *dl; dmu_tx_t *tx; zfsvfs_t *zfsvfs = ZTOZSB(dzp); zilog_t *zilog; uint64_t len = strlen(link); int error; int zflg = ZNEW; zfs_acl_ids_t acl_ids; boolean_t fuid_dirtied; uint64_t txtype = TX_SYMLINK; boolean_t waited = B_FALSE; ASSERT(S_ISLNK(vap->va_mode)); if (name == NULL) return (SET_ERROR(EINVAL)); if ((error = zfs_enter_verify_zp(zfsvfs, dzp, FTAG)) != 0) return (error); zilog = zfsvfs->z_log; if (zfsvfs->z_utf8 && u8_validate(name, strlen(name), NULL, U8_VALIDATE_ENTIRE, &error) < 0) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EILSEQ)); } if (flags & FIGNORECASE) zflg |= ZCILOOK; if (len > MAXPATHLEN) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(ENAMETOOLONG)); } if ((error = zfs_acl_ids_create(dzp, 0, vap, cr, NULL, &acl_ids, mnt_ns)) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } top: *zpp = NULL; /* * Attempt to lock directory; fail if entry already exists. */ error = zfs_dirent_lock(&dl, dzp, name, &zp, zflg, NULL, NULL); if (error) { zfs_acl_ids_free(&acl_ids); zfs_exit(zfsvfs, FTAG); return (error); } if ((error = zfs_zaccess(dzp, ACE_ADD_FILE, 0, B_FALSE, cr, mnt_ns))) { zfs_acl_ids_free(&acl_ids); zfs_dirent_unlock(dl); zfs_exit(zfsvfs, FTAG); return (error); } if (zfs_acl_ids_overquota(zfsvfs, &acl_ids, ZFS_DEFAULT_PROJID)) { zfs_acl_ids_free(&acl_ids); zfs_dirent_unlock(dl); zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EDQUOT)); } tx = dmu_tx_create(zfsvfs->z_os); fuid_dirtied = zfsvfs->z_fuid_dirty; dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, MAX(1, len)); dmu_tx_hold_zap(tx, dzp->z_id, TRUE, name); dmu_tx_hold_sa_create(tx, acl_ids.z_aclp->z_acl_bytes + ZFS_SA_BASE_ATTR_SIZE + len); dmu_tx_hold_sa(tx, dzp->z_sa_hdl, B_FALSE); if (!zfsvfs->z_use_sa && acl_ids.z_aclp->z_acl_bytes > ZFS_ACE_SPACE) { dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, acl_ids.z_aclp->z_acl_bytes); } if (fuid_dirtied) zfs_fuid_txhold(zfsvfs, tx); error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT); if (error) { zfs_dirent_unlock(dl); if (error == ERESTART) { waited = B_TRUE; dmu_tx_wait(tx); dmu_tx_abort(tx); goto top; } zfs_acl_ids_free(&acl_ids); dmu_tx_abort(tx); zfs_exit(zfsvfs, FTAG); return (error); } /* * Create a new object for the symlink. * for version 4 ZPL datasets the symlink will be an SA attribute */ zfs_mknode(dzp, vap, tx, cr, 0, &zp, &acl_ids); if (fuid_dirtied) zfs_fuid_sync(zfsvfs, tx); mutex_enter(&zp->z_lock); if (zp->z_is_sa) error = sa_update(zp->z_sa_hdl, SA_ZPL_SYMLINK(zfsvfs), link, len, tx); else zfs_sa_symlink(zp, link, len, tx); mutex_exit(&zp->z_lock); zp->z_size = len; (void) sa_update(zp->z_sa_hdl, SA_ZPL_SIZE(zfsvfs), &zp->z_size, sizeof (zp->z_size), tx); /* * Insert the new object into the directory. */ error = zfs_link_create(dl, zp, tx, ZNEW); if (error != 0) { zfs_znode_delete(zp, tx); remove_inode_hash(ZTOI(zp)); } else { if (flags & FIGNORECASE) txtype |= TX_CI; zfs_log_symlink(zilog, tx, txtype, dzp, zp, name, link); zfs_znode_update_vfs(dzp); zfs_znode_update_vfs(zp); } zfs_acl_ids_free(&acl_ids); dmu_tx_commit(tx); zfs_dirent_unlock(dl); if (error == 0) { *zpp = zp; if (zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS) zil_commit(zilog, 0); } else { zrele(zp); } zfs_exit(zfsvfs, FTAG); return (error); } /* * Return, in the buffer contained in the provided uio structure, * the symbolic path referred to by ip. * * IN: ip - inode of symbolic link * uio - structure to contain the link path. * cr - credentials of caller. * * RETURN: 0 if success * error code if failure * * Timestamps: * ip - atime updated */ int zfs_readlink(struct inode *ip, zfs_uio_t *uio, cred_t *cr) { (void) cr; znode_t *zp = ITOZ(ip); zfsvfs_t *zfsvfs = ITOZSB(ip); int error; if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); mutex_enter(&zp->z_lock); if (zp->z_is_sa) error = sa_lookup_uio(zp->z_sa_hdl, SA_ZPL_SYMLINK(zfsvfs), uio); else error = zfs_sa_readlink(zp, uio); mutex_exit(&zp->z_lock); zfs_exit(zfsvfs, FTAG); return (error); } /* * Insert a new entry into directory tdzp referencing szp. * * IN: tdzp - Directory to contain new entry. * szp - znode of new entry. * name - name of new entry. * cr - credentials of caller. * flags - case flags. * * RETURN: 0 if success * error code if failure * * Timestamps: * tdzp - ctime|mtime updated * szp - ctime updated */ int zfs_link(znode_t *tdzp, znode_t *szp, char *name, cred_t *cr, int flags) { struct inode *sip = ZTOI(szp); znode_t *tzp; zfsvfs_t *zfsvfs = ZTOZSB(tdzp); zilog_t *zilog; zfs_dirlock_t *dl; dmu_tx_t *tx; int error; int zf = ZNEW; uint64_t parent; uid_t owner; boolean_t waited = B_FALSE; boolean_t is_tmpfile = 0; uint64_t txg; #ifdef HAVE_TMPFILE is_tmpfile = (sip->i_nlink == 0 && (sip->i_state & I_LINKABLE)); #endif ASSERT(S_ISDIR(ZTOI(tdzp)->i_mode)); if (name == NULL) return (SET_ERROR(EINVAL)); if ((error = zfs_enter_verify_zp(zfsvfs, tdzp, FTAG)) != 0) return (error); zilog = zfsvfs->z_log; /* * POSIX dictates that we return EPERM here. * Better choices include ENOTSUP or EISDIR. */ if (S_ISDIR(sip->i_mode)) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EPERM)); } if ((error = zfs_verify_zp(szp)) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } /* * If we are using project inheritance, means if the directory has * ZFS_PROJINHERIT set, then its descendant directories will inherit * not only the project ID, but also the ZFS_PROJINHERIT flag. Under * such case, we only allow hard link creation in our tree when the * project IDs are the same. */ if (tdzp->z_pflags & ZFS_PROJINHERIT && tdzp->z_projid != szp->z_projid) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EXDEV)); } /* * We check i_sb because snapshots and the ctldir must have different * super blocks. */ if (sip->i_sb != ZTOI(tdzp)->i_sb || zfsctl_is_node(sip)) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EXDEV)); } /* Prevent links to .zfs/shares files */ if ((error = sa_lookup(szp->z_sa_hdl, SA_ZPL_PARENT(zfsvfs), &parent, sizeof (uint64_t))) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } if (parent == zfsvfs->z_shares_dir) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EPERM)); } if (zfsvfs->z_utf8 && u8_validate(name, strlen(name), NULL, U8_VALIDATE_ENTIRE, &error) < 0) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EILSEQ)); } if (flags & FIGNORECASE) zf |= ZCILOOK; /* * We do not support links between attributes and non-attributes * because of the potential security risk of creating links * into "normal" file space in order to circumvent restrictions * imposed in attribute space. */ if ((szp->z_pflags & ZFS_XATTR) != (tdzp->z_pflags & ZFS_XATTR)) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EINVAL)); } owner = zfs_fuid_map_id(zfsvfs, KUID_TO_SUID(sip->i_uid), cr, ZFS_OWNER); if (owner != crgetuid(cr) && secpolicy_basic_link(cr) != 0) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EPERM)); } if ((error = zfs_zaccess(tdzp, ACE_ADD_FILE, 0, B_FALSE, cr, zfs_init_idmap))) { zfs_exit(zfsvfs, FTAG); return (error); } top: /* * Attempt to lock directory; fail if entry already exists. */ error = zfs_dirent_lock(&dl, tdzp, name, &tzp, zf, NULL, NULL); if (error) { zfs_exit(zfsvfs, FTAG); return (error); } tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_sa(tx, szp->z_sa_hdl, B_FALSE); dmu_tx_hold_zap(tx, tdzp->z_id, TRUE, name); if (is_tmpfile) dmu_tx_hold_zap(tx, zfsvfs->z_unlinkedobj, FALSE, NULL); zfs_sa_upgrade_txholds(tx, szp); zfs_sa_upgrade_txholds(tx, tdzp); error = dmu_tx_assign(tx, (waited ? TXG_NOTHROTTLE : 0) | TXG_NOWAIT); if (error) { zfs_dirent_unlock(dl); if (error == ERESTART) { waited = B_TRUE; dmu_tx_wait(tx); dmu_tx_abort(tx); goto top; } dmu_tx_abort(tx); zfs_exit(zfsvfs, FTAG); return (error); } /* unmark z_unlinked so zfs_link_create will not reject */ if (is_tmpfile) szp->z_unlinked = B_FALSE; error = zfs_link_create(dl, szp, tx, 0); if (error == 0) { uint64_t txtype = TX_LINK; /* * tmpfile is created to be in z_unlinkedobj, so remove it. * Also, we don't log in ZIL, because all previous file * operation on the tmpfile are ignored by ZIL. Instead we * always wait for txg to sync to make sure all previous * operation are sync safe. */ if (is_tmpfile) { VERIFY(zap_remove_int(zfsvfs->z_os, zfsvfs->z_unlinkedobj, szp->z_id, tx) == 0); } else { if (flags & FIGNORECASE) txtype |= TX_CI; zfs_log_link(zilog, tx, txtype, tdzp, szp, name); } } else if (is_tmpfile) { /* restore z_unlinked since when linking failed */ szp->z_unlinked = B_TRUE; } txg = dmu_tx_get_txg(tx); dmu_tx_commit(tx); zfs_dirent_unlock(dl); if (!is_tmpfile && zfsvfs->z_os->os_sync == ZFS_SYNC_ALWAYS) zil_commit(zilog, 0); if (is_tmpfile && zfsvfs->z_os->os_sync != ZFS_SYNC_DISABLED) txg_wait_synced(dmu_objset_pool(zfsvfs->z_os), txg); zfs_znode_update_vfs(tdzp); zfs_znode_update_vfs(szp); zfs_exit(zfsvfs, FTAG); return (error); } static void zfs_putpage_sync_commit_cb(void *arg) { struct page *pp = arg; ClearPageError(pp); end_page_writeback(pp); } static void zfs_putpage_async_commit_cb(void *arg) { struct page *pp = arg; znode_t *zp = ITOZ(pp->mapping->host); ClearPageError(pp); end_page_writeback(pp); atomic_dec_32(&zp->z_async_writes_cnt); } /* * Push a page out to disk, once the page is on stable storage the * registered commit callback will be run as notification of completion. * * IN: ip - page mapped for inode. * pp - page to push (page is locked) * wbc - writeback control data * for_sync - does the caller intend to wait synchronously for the * page writeback to complete? * * RETURN: 0 if success * error code if failure * * Timestamps: * ip - ctime|mtime updated */ int zfs_putpage(struct inode *ip, struct page *pp, struct writeback_control *wbc, boolean_t for_sync) { znode_t *zp = ITOZ(ip); zfsvfs_t *zfsvfs = ITOZSB(ip); loff_t offset; loff_t pgoff; unsigned int pglen; dmu_tx_t *tx; caddr_t va; int err = 0; uint64_t mtime[2], ctime[2]; inode_timespec_t tmp_ctime; sa_bulk_attr_t bulk[3]; int cnt = 0; struct address_space *mapping; if ((err = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (err); ASSERT(PageLocked(pp)); pgoff = page_offset(pp); /* Page byte-offset in file */ offset = i_size_read(ip); /* File length in bytes */ pglen = MIN(PAGE_SIZE, /* Page length in bytes */ P2ROUNDUP(offset, PAGE_SIZE)-pgoff); /* Page is beyond end of file */ if (pgoff >= offset) { unlock_page(pp); zfs_exit(zfsvfs, FTAG); return (0); } /* Truncate page length to end of file */ if (pgoff + pglen > offset) pglen = offset - pgoff; #if 0 /* * FIXME: Allow mmap writes past its quota. The correct fix * is to register a page_mkwrite() handler to count the page * against its quota when it is about to be dirtied. */ if (zfs_id_overblockquota(zfsvfs, DMU_USERUSED_OBJECT, KUID_TO_SUID(ip->i_uid)) || zfs_id_overblockquota(zfsvfs, DMU_GROUPUSED_OBJECT, KGID_TO_SGID(ip->i_gid)) || (zp->z_projid != ZFS_DEFAULT_PROJID && zfs_id_overblockquota(zfsvfs, DMU_PROJECTUSED_OBJECT, zp->z_projid))) { err = EDQUOT; } #endif /* * The ordering here is critical and must adhere to the following * rules in order to avoid deadlocking in either zfs_read() or * zfs_free_range() due to a lock inversion. * * 1) The page must be unlocked prior to acquiring the range lock. * This is critical because zfs_read() calls find_lock_page() * which may block on the page lock while holding the range lock. * * 2) Before setting or clearing write back on a page the range lock * must be held in order to prevent a lock inversion with the * zfs_free_range() function. * * This presents a problem because upon entering this function the * page lock is already held. To safely acquire the range lock the * page lock must be dropped. This creates a window where another * process could truncate, invalidate, dirty, or write out the page. * * Therefore, after successfully reacquiring the range and page locks * the current page state is checked. In the common case everything * will be as is expected and it can be written out. However, if * the page state has changed it must be handled accordingly. */ mapping = pp->mapping; redirty_page_for_writepage(wbc, pp); unlock_page(pp); zfs_locked_range_t *lr = zfs_rangelock_enter(&zp->z_rangelock, pgoff, pglen, RL_WRITER); lock_page(pp); /* Page mapping changed or it was no longer dirty, we're done */ if (unlikely((mapping != pp->mapping) || !PageDirty(pp))) { unlock_page(pp); zfs_rangelock_exit(lr); zfs_exit(zfsvfs, FTAG); return (0); } /* Another process started write block if required */ if (PageWriteback(pp)) { unlock_page(pp); zfs_rangelock_exit(lr); if (wbc->sync_mode != WB_SYNC_NONE) { /* * Speed up any non-sync page writebacks since * they may take several seconds to complete. * Refer to the comment in zpl_fsync() (when * HAVE_FSYNC_RANGE is defined) for details. */ if (atomic_load_32(&zp->z_async_writes_cnt) > 0) { zil_commit(zfsvfs->z_log, zp->z_id); } if (PageWriteback(pp)) #ifdef HAVE_PAGEMAP_FOLIO_WAIT_BIT folio_wait_bit(page_folio(pp), PG_writeback); #else wait_on_page_bit(pp, PG_writeback); #endif } zfs_exit(zfsvfs, FTAG); return (0); } /* Clear the dirty flag the required locks are held */ if (!clear_page_dirty_for_io(pp)) { unlock_page(pp); zfs_rangelock_exit(lr); zfs_exit(zfsvfs, FTAG); return (0); } /* * Counterpart for redirty_page_for_writepage() above. This page * was in fact not skipped and should not be counted as if it were. */ wbc->pages_skipped--; if (!for_sync) atomic_inc_32(&zp->z_async_writes_cnt); set_page_writeback(pp); unlock_page(pp); tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_write(tx, zp->z_id, pgoff, pglen); dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE); zfs_sa_upgrade_txholds(tx, zp); err = dmu_tx_assign(tx, TXG_NOWAIT); if (err != 0) { if (err == ERESTART) dmu_tx_wait(tx); dmu_tx_abort(tx); #ifdef HAVE_VFS_FILEMAP_DIRTY_FOLIO filemap_dirty_folio(page_mapping(pp), page_folio(pp)); #else __set_page_dirty_nobuffers(pp); #endif ClearPageError(pp); end_page_writeback(pp); if (!for_sync) atomic_dec_32(&zp->z_async_writes_cnt); zfs_rangelock_exit(lr); zfs_exit(zfsvfs, FTAG); return (err); } va = kmap(pp); ASSERT3U(pglen, <=, PAGE_SIZE); dmu_write(zfsvfs->z_os, zp->z_id, pgoff, pglen, va, tx); kunmap(pp); SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_MTIME(zfsvfs), NULL, &mtime, 16); SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_CTIME(zfsvfs), NULL, &ctime, 16); SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_FLAGS(zfsvfs), NULL, &zp->z_pflags, 8); /* Preserve the mtime and ctime provided by the inode */ ZFS_TIME_ENCODE(&ip->i_mtime, mtime); tmp_ctime = zpl_inode_get_ctime(ip); ZFS_TIME_ENCODE(&tmp_ctime, ctime); zp->z_atime_dirty = B_FALSE; zp->z_seq++; err = sa_bulk_update(zp->z_sa_hdl, bulk, cnt, tx); boolean_t commit = B_FALSE; if (wbc->sync_mode != WB_SYNC_NONE) { /* * Note that this is rarely called under writepages(), because * writepages() normally handles the entire commit for * performance reasons. */ commit = B_TRUE; } else if (!for_sync && atomic_load_32(&zp->z_sync_writes_cnt) > 0) { /* * If the caller does not intend to wait synchronously * for this page writeback to complete and there are active * synchronous calls on this file, do a commit so that * the latter don't accidentally end up waiting for * our writeback to complete. Refer to the comment in * zpl_fsync() (when HAVE_FSYNC_RANGE is defined) for details. */ commit = B_TRUE; } zfs_log_write(zfsvfs->z_log, tx, TX_WRITE, zp, pgoff, pglen, commit, for_sync ? zfs_putpage_sync_commit_cb : zfs_putpage_async_commit_cb, pp); dmu_tx_commit(tx); zfs_rangelock_exit(lr); if (commit) zil_commit(zfsvfs->z_log, zp->z_id); dataset_kstats_update_write_kstats(&zfsvfs->z_kstat, pglen); zfs_exit(zfsvfs, FTAG); return (err); } /* * Update the system attributes when the inode has been dirtied. For the * moment we only update the mode, atime, mtime, and ctime. */ int zfs_dirty_inode(struct inode *ip, int flags) { znode_t *zp = ITOZ(ip); zfsvfs_t *zfsvfs = ITOZSB(ip); dmu_tx_t *tx; uint64_t mode, atime[2], mtime[2], ctime[2]; inode_timespec_t tmp_ctime; sa_bulk_attr_t bulk[4]; int error = 0; int cnt = 0; if (zfs_is_readonly(zfsvfs) || dmu_objset_is_snapshot(zfsvfs->z_os)) return (0); if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); #ifdef I_DIRTY_TIME /* * This is the lazytime semantic introduced in Linux 4.0 * This flag will only be called from update_time when lazytime is set. * (Note, I_DIRTY_SYNC will also set if not lazytime) * Fortunately mtime and ctime are managed within ZFS itself, so we * only need to dirty atime. */ if (flags == I_DIRTY_TIME) { zp->z_atime_dirty = B_TRUE; goto out; } #endif tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE); zfs_sa_upgrade_txholds(tx, zp); error = dmu_tx_assign(tx, TXG_WAIT); if (error) { dmu_tx_abort(tx); goto out; } mutex_enter(&zp->z_lock); zp->z_atime_dirty = B_FALSE; SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_MODE(zfsvfs), NULL, &mode, 8); SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_ATIME(zfsvfs), NULL, &atime, 16); SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_MTIME(zfsvfs), NULL, &mtime, 16); SA_ADD_BULK_ATTR(bulk, cnt, SA_ZPL_CTIME(zfsvfs), NULL, &ctime, 16); /* Preserve the mode, mtime and ctime provided by the inode */ ZFS_TIME_ENCODE(&ip->i_atime, atime); ZFS_TIME_ENCODE(&ip->i_mtime, mtime); tmp_ctime = zpl_inode_get_ctime(ip); ZFS_TIME_ENCODE(&tmp_ctime, ctime); mode = ip->i_mode; zp->z_mode = mode; error = sa_bulk_update(zp->z_sa_hdl, bulk, cnt, tx); mutex_exit(&zp->z_lock); dmu_tx_commit(tx); out: zfs_exit(zfsvfs, FTAG); return (error); } void zfs_inactive(struct inode *ip) { znode_t *zp = ITOZ(ip); zfsvfs_t *zfsvfs = ITOZSB(ip); uint64_t atime[2]; int error; int need_unlock = 0; /* Only read lock if we haven't already write locked, e.g. rollback */ if (!RW_WRITE_HELD(&zfsvfs->z_teardown_inactive_lock)) { need_unlock = 1; rw_enter(&zfsvfs->z_teardown_inactive_lock, RW_READER); } if (zp->z_sa_hdl == NULL) { if (need_unlock) rw_exit(&zfsvfs->z_teardown_inactive_lock); return; } if (zp->z_atime_dirty && zp->z_unlinked == B_FALSE) { dmu_tx_t *tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE); zfs_sa_upgrade_txholds(tx, zp); error = dmu_tx_assign(tx, TXG_WAIT); if (error) { dmu_tx_abort(tx); } else { ZFS_TIME_ENCODE(&ip->i_atime, atime); mutex_enter(&zp->z_lock); (void) sa_update(zp->z_sa_hdl, SA_ZPL_ATIME(zfsvfs), (void *)&atime, sizeof (atime), tx); zp->z_atime_dirty = B_FALSE; mutex_exit(&zp->z_lock); dmu_tx_commit(tx); } } zfs_zinactive(zp); if (need_unlock) rw_exit(&zfsvfs->z_teardown_inactive_lock); } /* * Fill pages with data from the disk. */ static int zfs_fillpage(struct inode *ip, struct page *pp) { zfsvfs_t *zfsvfs = ITOZSB(ip); loff_t i_size = i_size_read(ip); u_offset_t io_off = page_offset(pp); size_t io_len = PAGE_SIZE; ASSERT3U(io_off, <, i_size); if (io_off + io_len > i_size) io_len = i_size - io_off; void *va = kmap(pp); int error = dmu_read(zfsvfs->z_os, ITOZ(ip)->z_id, io_off, io_len, va, DMU_READ_PREFETCH); if (io_len != PAGE_SIZE) memset((char *)va + io_len, 0, PAGE_SIZE - io_len); kunmap(pp); if (error) { /* convert checksum errors into IO errors */ if (error == ECKSUM) error = SET_ERROR(EIO); SetPageError(pp); ClearPageUptodate(pp); } else { ClearPageError(pp); SetPageUptodate(pp); } return (error); } /* * Uses zfs_fillpage to read data from the file and fill the page. * * IN: ip - inode of file to get data from. * pp - page to read * * RETURN: 0 on success, error code on failure. * * Timestamps: * vp - atime updated */ int zfs_getpage(struct inode *ip, struct page *pp) { zfsvfs_t *zfsvfs = ITOZSB(ip); znode_t *zp = ITOZ(ip); int error; if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); error = zfs_fillpage(ip, pp); if (error == 0) dataset_kstats_update_read_kstats(&zfsvfs->z_kstat, PAGE_SIZE); zfs_exit(zfsvfs, FTAG); return (error); } /* * Check ZFS specific permissions to memory map a section of a file. * * IN: ip - inode of the file to mmap * off - file offset * addrp - start address in memory region * len - length of memory region * vm_flags- address flags * * RETURN: 0 if success * error code if failure */ int zfs_map(struct inode *ip, offset_t off, caddr_t *addrp, size_t len, unsigned long vm_flags) { (void) addrp; znode_t *zp = ITOZ(ip); zfsvfs_t *zfsvfs = ITOZSB(ip); int error; if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); if ((vm_flags & VM_WRITE) && (vm_flags & VM_SHARED) && (zp->z_pflags & (ZFS_IMMUTABLE | ZFS_READONLY | ZFS_APPENDONLY))) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EPERM)); } if ((vm_flags & (VM_READ | VM_EXEC)) && (zp->z_pflags & ZFS_AV_QUARANTINED)) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EACCES)); } if (off < 0 || len > MAXOFFSET_T - off) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(ENXIO)); } zfs_exit(zfsvfs, FTAG); return (0); } /* * Free or allocate space in a file. Currently, this function only * supports the `F_FREESP' command. However, this command is somewhat * misnamed, as its functionality includes the ability to allocate as * well as free space. * * IN: zp - znode of file to free data in. * cmd - action to take (only F_FREESP supported). * bfp - section of file to free/alloc. * flag - current file open mode flags. * offset - current file offset. * cr - credentials of caller. * * RETURN: 0 on success, error code on failure. * * Timestamps: * zp - ctime|mtime updated */ int zfs_space(znode_t *zp, int cmd, flock64_t *bfp, int flag, offset_t offset, cred_t *cr) { (void) offset; zfsvfs_t *zfsvfs = ZTOZSB(zp); uint64_t off, len; int error; if ((error = zfs_enter_verify_zp(zfsvfs, zp, FTAG)) != 0) return (error); if (cmd != F_FREESP) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EINVAL)); } /* * Callers might not be able to detect properly that we are read-only, * so check it explicitly here. */ if (zfs_is_readonly(zfsvfs)) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EROFS)); } if (bfp->l_len < 0) { zfs_exit(zfsvfs, FTAG); return (SET_ERROR(EINVAL)); } /* * Permissions aren't checked on Solaris because on this OS * zfs_space() can only be called with an opened file handle. * On Linux we can get here through truncate_range() which * operates directly on inodes, so we need to check access rights. */ if ((error = zfs_zaccess(zp, ACE_WRITE_DATA, 0, B_FALSE, cr, zfs_init_idmap))) { zfs_exit(zfsvfs, FTAG); return (error); } off = bfp->l_start; len = bfp->l_len; /* 0 means from off to end of file */ error = zfs_freesp(zp, off, len, flag, TRUE); zfs_exit(zfsvfs, FTAG); return (error); } int zfs_fid(struct inode *ip, fid_t *fidp) { znode_t *zp = ITOZ(ip); zfsvfs_t *zfsvfs = ITOZSB(ip); uint32_t gen; uint64_t gen64; uint64_t object = zp->z_id; zfid_short_t *zfid; int size, i, error; if ((error = zfs_enter(zfsvfs, FTAG)) != 0) return (error); if (fidp->fid_len < SHORT_FID_LEN) { fidp->fid_len = SHORT_FID_LEN; zfs_exit(zfsvfs, FTAG); return (SET_ERROR(ENOSPC)); } if ((error = zfs_verify_zp(zp)) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_GEN(zfsvfs), &gen64, sizeof (uint64_t))) != 0) { zfs_exit(zfsvfs, FTAG); return (error); } gen = (uint32_t)gen64; size = SHORT_FID_LEN; zfid = (zfid_short_t *)fidp; zfid->zf_len = size; for (i = 0; i < sizeof (zfid->zf_object); i++) zfid->zf_object[i] = (uint8_t)(object >> (8 * i)); /* Must have a non-zero generation number to distinguish from .zfs */ if (gen == 0) gen = 1; for (i = 0; i < sizeof (zfid->zf_gen); i++) zfid->zf_gen[i] = (uint8_t)(gen >> (8 * i)); zfs_exit(zfsvfs, FTAG); return (0); } #if defined(_KERNEL) EXPORT_SYMBOL(zfs_open); EXPORT_SYMBOL(zfs_close); EXPORT_SYMBOL(zfs_lookup); EXPORT_SYMBOL(zfs_create); EXPORT_SYMBOL(zfs_tmpfile); EXPORT_SYMBOL(zfs_remove); EXPORT_SYMBOL(zfs_mkdir); EXPORT_SYMBOL(zfs_rmdir); EXPORT_SYMBOL(zfs_readdir); EXPORT_SYMBOL(zfs_getattr_fast); EXPORT_SYMBOL(zfs_setattr); EXPORT_SYMBOL(zfs_rename); EXPORT_SYMBOL(zfs_symlink); EXPORT_SYMBOL(zfs_readlink); EXPORT_SYMBOL(zfs_link); EXPORT_SYMBOL(zfs_inactive); EXPORT_SYMBOL(zfs_space); EXPORT_SYMBOL(zfs_fid); EXPORT_SYMBOL(zfs_getpage); EXPORT_SYMBOL(zfs_putpage); EXPORT_SYMBOL(zfs_dirty_inode); EXPORT_SYMBOL(zfs_map); /* CSTYLED */ module_param(zfs_delete_blocks, ulong, 0644); MODULE_PARM_DESC(zfs_delete_blocks, "Delete files larger than N blocks async"); +/* CSTYLED */ +module_param(zfs_bclone_enabled, uint, 0644); +MODULE_PARM_DESC(zfs_bclone_enabled, "Enable block cloning"); + #endif diff --git a/module/os/linux/zfs/zpl_file_range.c b/module/os/linux/zfs/zpl_file_range.c index c47fe99dacff..73476ff40ebf 100644 --- a/module/os/linux/zfs/zpl_file_range.c +++ b/module/os/linux/zfs/zpl_file_range.c @@ -1,276 +1,281 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or https://opensource.org/licenses/CDDL-1.0. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2023, Klara Inc. */ #ifdef CONFIG_COMPAT #include #endif #include #include #include #include #include +int zfs_bclone_enabled = 1; + /* * Clone part of a file via block cloning. * * Note that we are not required to update file offsets; the kernel will take * care of that depending on how it was called. */ static ssize_t __zpl_clone_file_range(struct file *src_file, loff_t src_off, struct file *dst_file, loff_t dst_off, size_t len) { struct inode *src_i = file_inode(src_file); struct inode *dst_i = file_inode(dst_file); uint64_t src_off_o = (uint64_t)src_off; uint64_t dst_off_o = (uint64_t)dst_off; uint64_t len_o = (uint64_t)len; cred_t *cr = CRED(); fstrans_cookie_t cookie; int err; + if (!zfs_bclone_enabled) + return (-EOPNOTSUPP); + if (!spa_feature_is_enabled( dmu_objset_spa(ITOZSB(dst_i)->z_os), SPA_FEATURE_BLOCK_CLONING)) return (-EOPNOTSUPP); if (src_i != dst_i) spl_inode_lock_shared(src_i); spl_inode_lock(dst_i); crhold(cr); cookie = spl_fstrans_mark(); err = -zfs_clone_range(ITOZ(src_i), &src_off_o, ITOZ(dst_i), &dst_off_o, &len_o, cr); spl_fstrans_unmark(cookie); crfree(cr); spl_inode_unlock(dst_i); if (src_i != dst_i) spl_inode_unlock_shared(src_i); if (err < 0) return (err); return ((ssize_t)len_o); } #if defined(HAVE_VFS_COPY_FILE_RANGE) || \ defined(HAVE_VFS_FILE_OPERATIONS_EXTEND) /* * Entry point for copy_file_range(). Copy len bytes from src_off in src_file * to dst_off in dst_file. We are permitted to do this however we like, so we * try to just clone the blocks, and if we can't support it, fall back to the * kernel's generic byte copy function. */ ssize_t zpl_copy_file_range(struct file *src_file, loff_t src_off, struct file *dst_file, loff_t dst_off, size_t len, unsigned int flags) { ssize_t ret; if (flags != 0) return (-EINVAL); /* Try to do it via zfs_clone_range() */ ret = __zpl_clone_file_range(src_file, src_off, dst_file, dst_off, len); #ifdef HAVE_VFS_GENERIC_COPY_FILE_RANGE /* * Since Linux 5.3 the filesystem driver is responsible for executing * an appropriate fallback, and a generic fallback function is provided. */ if (ret == -EOPNOTSUPP || ret == -EINVAL || ret == -EXDEV || ret == -EAGAIN) ret = generic_copy_file_range(src_file, src_off, dst_file, dst_off, len, flags); #else /* * Before Linux 5.3 the filesystem has to return -EOPNOTSUPP to signal * to the kernel that it should fallback to a content copy. */ if (ret == -EINVAL || ret == -EXDEV || ret == -EAGAIN) ret = -EOPNOTSUPP; #endif /* HAVE_VFS_GENERIC_COPY_FILE_RANGE */ return (ret); } #endif /* HAVE_VFS_COPY_FILE_RANGE || HAVE_VFS_FILE_OPERATIONS_EXTEND */ #ifdef HAVE_VFS_REMAP_FILE_RANGE /* * Entry point for FICLONE/FICLONERANGE/FIDEDUPERANGE. * * FICLONE and FICLONERANGE are basically the same as copy_file_range(), except * that they must clone - they cannot fall back to copying. FICLONE is exactly * FICLONERANGE, for the entire file. We don't need to try to tell them apart; * the kernel will sort that out for us. * * FIDEDUPERANGE is for turning a non-clone into a clone, that is, compare the * range in both files and if they're the same, arrange for them to be backed * by the same storage. */ loff_t zpl_remap_file_range(struct file *src_file, loff_t src_off, struct file *dst_file, loff_t dst_off, loff_t len, unsigned int flags) { if (flags & ~(REMAP_FILE_DEDUP | REMAP_FILE_CAN_SHORTEN)) return (-EINVAL); /* * REMAP_FILE_CAN_SHORTEN lets us know we can clone less than the given * range if we want. Its designed for filesystems that make data past * EOF available, and don't want it to be visible in both files. ZFS * doesn't do that, so we just turn the flag off. */ flags &= ~REMAP_FILE_CAN_SHORTEN; if (flags & REMAP_FILE_DEDUP) /* No support for dedup yet */ return (-EOPNOTSUPP); /* Zero length means to clone everything to the end of the file */ if (len == 0) len = i_size_read(file_inode(src_file)) - src_off; return (__zpl_clone_file_range(src_file, src_off, dst_file, dst_off, len)); } #endif /* HAVE_VFS_REMAP_FILE_RANGE */ #if defined(HAVE_VFS_CLONE_FILE_RANGE) || \ defined(HAVE_VFS_FILE_OPERATIONS_EXTEND) /* * Entry point for FICLONE and FICLONERANGE, before Linux 4.20. */ int zpl_clone_file_range(struct file *src_file, loff_t src_off, struct file *dst_file, loff_t dst_off, uint64_t len) { /* Zero length means to clone everything to the end of the file */ if (len == 0) len = i_size_read(file_inode(src_file)) - src_off; return (__zpl_clone_file_range(src_file, src_off, dst_file, dst_off, len)); } #endif /* HAVE_VFS_CLONE_FILE_RANGE || HAVE_VFS_FILE_OPERATIONS_EXTEND */ #ifdef HAVE_VFS_DEDUPE_FILE_RANGE /* * Entry point for FIDEDUPERANGE, before Linux 4.20. */ int zpl_dedupe_file_range(struct file *src_file, loff_t src_off, struct file *dst_file, loff_t dst_off, uint64_t len) { /* No support for dedup yet */ return (-EOPNOTSUPP); } #endif /* HAVE_VFS_DEDUPE_FILE_RANGE */ /* Entry point for FICLONE, before Linux 4.5. */ long zpl_ioctl_ficlone(struct file *dst_file, void *arg) { unsigned long sfd = (unsigned long)arg; struct file *src_file = fget(sfd); if (src_file == NULL) return (-EBADF); if (dst_file->f_op != src_file->f_op) { fput(src_file); return (-EXDEV); } size_t len = i_size_read(file_inode(src_file)); ssize_t ret = __zpl_clone_file_range(src_file, 0, dst_file, 0, len); fput(src_file); if (ret < 0) { if (ret == -EOPNOTSUPP) return (-ENOTTY); return (ret); } if (ret != len) return (-EINVAL); return (0); } /* Entry point for FICLONERANGE, before Linux 4.5. */ long zpl_ioctl_ficlonerange(struct file *dst_file, void __user *arg) { zfs_ioc_compat_file_clone_range_t fcr; if (copy_from_user(&fcr, arg, sizeof (fcr))) return (-EFAULT); struct file *src_file = fget(fcr.fcr_src_fd); if (src_file == NULL) return (-EBADF); if (dst_file->f_op != src_file->f_op) { fput(src_file); return (-EXDEV); } size_t len = fcr.fcr_src_length; if (len == 0) len = i_size_read(file_inode(src_file)) - fcr.fcr_src_offset; ssize_t ret = __zpl_clone_file_range(src_file, fcr.fcr_src_offset, dst_file, fcr.fcr_dest_offset, len); fput(src_file); if (ret < 0) { if (ret == -EOPNOTSUPP) return (-ENOTTY); return (ret); } if (ret != len) return (-EINVAL); return (0); } /* Entry point for FIDEDUPERANGE, before Linux 4.5. */ long zpl_ioctl_fideduperange(struct file *filp, void *arg) { (void) arg; /* No support for dedup yet */ return (-ENOTTY); } diff --git a/tests/zfs-tests/include/libtest.shlib b/tests/zfs-tests/include/libtest.shlib index 844caa17d8ed..d5d7bb6c8360 100644 --- a/tests/zfs-tests/include/libtest.shlib +++ b/tests/zfs-tests/include/libtest.shlib @@ -1,3875 +1,3890 @@ # # CDDL HEADER START # # The contents of this file are subject to the terms of the # Common Development and Distribution License (the "License"). # You may not use this file except in compliance with the License. # # You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE # or https://opensource.org/licenses/CDDL-1.0. # See the License for the specific language governing permissions # and limitations under the License. # # When distributing Covered Code, include this CDDL HEADER in each # file and include the License file at usr/src/OPENSOLARIS.LICENSE. # If applicable, add the following below this CDDL HEADER, with the # fields enclosed by brackets "[]" replaced with your own identifying # information: Portions Copyright [yyyy] [name of copyright owner] # # CDDL HEADER END # # # Copyright (c) 2009, Sun Microsystems Inc. All rights reserved. # Copyright (c) 2012, 2020, Delphix. All rights reserved. # Copyright (c) 2017, Tim Chase. All rights reserved. # Copyright (c) 2017, Nexenta Systems Inc. All rights reserved. # Copyright (c) 2017, Lawrence Livermore National Security LLC. # Copyright (c) 2017, Datto Inc. All rights reserved. # Copyright (c) 2017, Open-E Inc. All rights reserved. # Copyright (c) 2021, The FreeBSD Foundation. # Use is subject to license terms. # . ${STF_SUITE}/include/tunables.cfg . ${STF_TOOLS}/include/logapi.shlib . ${STF_SUITE}/include/math.shlib . ${STF_SUITE}/include/blkdev.shlib # # Apply constrained path when available. This is required since the # PATH may have been modified by sudo's secure_path behavior. # if [ -n "$STF_PATH" ]; then export PATH="$STF_PATH" fi # # Generic dot version comparison function # # Returns success when version $1 is greater than or equal to $2. # function compare_version_gte { [ "$(printf "$1\n$2" | sort -V | tail -n1)" = "$1" ] } # Linux kernel version comparison function # # $1 Linux version ("4.10", "2.6.32") or blank for installed Linux version # # Used for comparison: if [ $(linux_version) -ge $(linux_version "2.6.32") ] # function linux_version { typeset ver="$1" [ -z "$ver" ] && ver=$(uname -r | grep -Eo "^[0-9]+\.[0-9]+\.[0-9]+") typeset version major minor _ IFS='.' read -r version major minor _ <<<"$ver" [ -z "$version" ] && version=0 [ -z "$major" ] && major=0 [ -z "$minor" ] && minor=0 echo $((version * 100000 + major * 1000 + minor)) } # Determine if this is a Linux test system # # Return 0 if platform Linux, 1 if otherwise function is_linux { [ "$UNAME" = "Linux" ] } # Determine if this is an illumos test system # # Return 0 if platform illumos, 1 if otherwise function is_illumos { [ "$UNAME" = "illumos" ] } # Determine if this is a FreeBSD test system # # Return 0 if platform FreeBSD, 1 if otherwise function is_freebsd { [ "$UNAME" = "FreeBSD" ] } # Determine if this is a 32-bit system # # Return 0 if platform is 32-bit, 1 if otherwise function is_32bit { [ $(getconf LONG_BIT) = "32" ] } # Determine if kmemleak is enabled # # Return 0 if kmemleak is enabled, 1 if otherwise function is_kmemleak { is_linux && [ -e /sys/kernel/debug/kmemleak ] } # Determine whether a dataset is mounted # # $1 dataset name # $2 filesystem type; optional - defaulted to zfs # # Return 0 if dataset is mounted; 1 if unmounted; 2 on error function ismounted { typeset fstype=$2 [[ -z $fstype ]] && fstype=zfs typeset out dir name case $fstype in zfs) if [[ "$1" == "/"* ]] ; then ! zfs mount | awk -v fs="$1" '$2 == fs {exit 1}' else ! zfs mount | awk -v ds="$1" '$1 == ds {exit 1}' fi ;; ufs|nfs) if is_freebsd; then mount -pt $fstype | while read dev dir _t _flags; do [[ "$1" == "$dev" || "$1" == "$dir" ]] && return 0 done else out=$(df -F $fstype $1 2>/dev/null) || return dir=${out%%\(*} dir=${dir%% *} name=${out##*\(} name=${name%%\)*} name=${name%% *} [[ "$1" == "$dir" || "$1" == "$name" ]] && return 0 fi ;; ext*) df -t $fstype $1 > /dev/null 2>&1 ;; zvol) if [[ -L "$ZVOL_DEVDIR/$1" ]]; then link=$(readlink -f $ZVOL_DEVDIR/$1) [[ -n "$link" ]] && \ mount | grep -q "^$link" && \ return 0 fi ;; *) false ;; esac } # Return 0 if a dataset is mounted; 1 otherwise # # $1 dataset name # $2 filesystem type; optional - defaulted to zfs function mounted { ismounted $1 $2 } # Return 0 if a dataset is unmounted; 1 otherwise # # $1 dataset name # $2 filesystem type; optional - defaulted to zfs function unmounted { ! ismounted $1 $2 } function default_setup { default_setup_noexit "$@" log_pass } function default_setup_no_mountpoint { default_setup_noexit "$1" "$2" "$3" "yes" log_pass } # # Given a list of disks, setup storage pools and datasets. # function default_setup_noexit { typeset disklist=$1 typeset container=$2 typeset volume=$3 typeset no_mountpoint=$4 log_note begin default_setup_noexit if is_global_zone; then if poolexists $TESTPOOL ; then destroy_pool $TESTPOOL fi [[ -d /$TESTPOOL ]] && rm -rf /$TESTPOOL log_must zpool create -f $TESTPOOL $disklist else reexport_pool fi rm -rf $TESTDIR || log_unresolved Could not remove $TESTDIR mkdir -p $TESTDIR || log_unresolved Could not create $TESTDIR log_must zfs create $TESTPOOL/$TESTFS if [[ -z $no_mountpoint ]]; then log_must zfs set mountpoint=$TESTDIR $TESTPOOL/$TESTFS fi if [[ -n $container ]]; then rm -rf $TESTDIR1 || \ log_unresolved Could not remove $TESTDIR1 mkdir -p $TESTDIR1 || \ log_unresolved Could not create $TESTDIR1 log_must zfs create $TESTPOOL/$TESTCTR log_must zfs set canmount=off $TESTPOOL/$TESTCTR log_must zfs create $TESTPOOL/$TESTCTR/$TESTFS1 if [[ -z $no_mountpoint ]]; then log_must zfs set mountpoint=$TESTDIR1 \ $TESTPOOL/$TESTCTR/$TESTFS1 fi fi if [[ -n $volume ]]; then if is_global_zone ; then log_must zfs create -V $VOLSIZE $TESTPOOL/$TESTVOL block_device_wait else log_must zfs create $TESTPOOL/$TESTVOL fi fi } # # Given a list of disks, setup a storage pool, file system and # a container. # function default_container_setup { typeset disklist=$1 default_setup "$disklist" "true" } # # Given a list of disks, setup a storage pool,file system # and a volume. # function default_volume_setup { typeset disklist=$1 default_setup "$disklist" "" "true" } # # Given a list of disks, setup a storage pool,file system, # a container and a volume. # function default_container_volume_setup { typeset disklist=$1 default_setup "$disklist" "true" "true" } # # Create a snapshot on a filesystem or volume. Defaultly create a snapshot on # filesystem # # $1 Existing filesystem or volume name. Default, $TESTPOOL/$TESTFS # $2 snapshot name. Default, $TESTSNAP # function create_snapshot { typeset fs_vol=${1:-$TESTPOOL/$TESTFS} typeset snap=${2:-$TESTSNAP} [[ -z $fs_vol ]] && log_fail "Filesystem or volume's name is undefined." [[ -z $snap ]] && log_fail "Snapshot's name is undefined." if snapexists $fs_vol@$snap; then log_fail "$fs_vol@$snap already exists." fi datasetexists $fs_vol || \ log_fail "$fs_vol must exist." log_must zfs snapshot $fs_vol@$snap } # # Create a clone from a snapshot, default clone name is $TESTCLONE. # # $1 Existing snapshot, $TESTPOOL/$TESTFS@$TESTSNAP is default. # $2 Clone name, $TESTPOOL/$TESTCLONE is default. # function create_clone # snapshot clone { typeset snap=${1:-$TESTPOOL/$TESTFS@$TESTSNAP} typeset clone=${2:-$TESTPOOL/$TESTCLONE} [[ -z $snap ]] && \ log_fail "Snapshot name is undefined." [[ -z $clone ]] && \ log_fail "Clone name is undefined." log_must zfs clone $snap $clone } # # Create a bookmark of the given snapshot. Defaultly create a bookmark on # filesystem. # # $1 Existing filesystem or volume name. Default, $TESTFS # $2 Existing snapshot name. Default, $TESTSNAP # $3 bookmark name. Default, $TESTBKMARK # function create_bookmark { typeset fs_vol=${1:-$TESTFS} typeset snap=${2:-$TESTSNAP} typeset bkmark=${3:-$TESTBKMARK} [[ -z $fs_vol ]] && log_fail "Filesystem or volume's name is undefined." [[ -z $snap ]] && log_fail "Snapshot's name is undefined." [[ -z $bkmark ]] && log_fail "Bookmark's name is undefined." if bkmarkexists $fs_vol#$bkmark; then log_fail "$fs_vol#$bkmark already exists." fi datasetexists $fs_vol || \ log_fail "$fs_vol must exist." snapexists $fs_vol@$snap || \ log_fail "$fs_vol@$snap must exist." log_must zfs bookmark $fs_vol@$snap $fs_vol#$bkmark } # # Create a temporary clone result of an interrupted resumable 'zfs receive' # $1 Destination filesystem name. Must not exist, will be created as the result # of this function along with its %recv temporary clone # $2 Source filesystem name. Must not exist, will be created and destroyed # function create_recv_clone { typeset recvfs="$1" typeset sendfs="${2:-$TESTPOOL/create_recv_clone}" typeset snap="$sendfs@snap1" typeset incr="$sendfs@snap2" typeset mountpoint="$TESTDIR/create_recv_clone" typeset sendfile="$TESTDIR/create_recv_clone.zsnap" [[ -z $recvfs ]] && log_fail "Recv filesystem's name is undefined." datasetexists $recvfs && log_fail "Recv filesystem must not exist." datasetexists $sendfs && log_fail "Send filesystem must not exist." log_must zfs create -o compression=off -o mountpoint="$mountpoint" $sendfs log_must zfs snapshot $snap log_must eval "zfs send $snap | zfs recv -u $recvfs" log_must mkfile 1m "$mountpoint/data" log_must zfs snapshot $incr log_must eval "zfs send -i $snap $incr | dd bs=10K count=1 \ iflag=fullblock > $sendfile" log_mustnot eval "zfs recv -su $recvfs < $sendfile" destroy_dataset "$sendfs" "-r" log_must rm -f "$sendfile" if [[ $(get_prop 'inconsistent' "$recvfs/%recv") -ne 1 ]]; then log_fail "Error creating temporary $recvfs/%recv clone" fi } function default_mirror_setup { default_mirror_setup_noexit $1 $2 $3 log_pass } # # Given a pair of disks, set up a storage pool and dataset for the mirror # @parameters: $1 the primary side of the mirror # $2 the secondary side of the mirror # @uses: ZPOOL ZFS TESTPOOL TESTFS function default_mirror_setup_noexit { readonly func="default_mirror_setup_noexit" typeset primary=$1 typeset secondary=$2 [[ -z $primary ]] && \ log_fail "$func: No parameters passed" [[ -z $secondary ]] && \ log_fail "$func: No secondary partition passed" [[ -d /$TESTPOOL ]] && rm -rf /$TESTPOOL log_must zpool create -f $TESTPOOL mirror $@ log_must zfs create $TESTPOOL/$TESTFS log_must zfs set mountpoint=$TESTDIR $TESTPOOL/$TESTFS } # # Destroy the configured testpool mirrors. # the mirrors are of the form ${TESTPOOL}{number} # @uses: ZPOOL ZFS TESTPOOL function destroy_mirrors { default_cleanup_noexit log_pass } function default_raidz_setup { default_raidz_setup_noexit "$*" log_pass } # # Given a minimum of two disks, set up a storage pool and dataset for the raid-z # $1 the list of disks # function default_raidz_setup_noexit { typeset disklist="$*" disks=(${disklist[*]}) if [[ ${#disks[*]} -lt 2 ]]; then log_fail "A raid-z requires a minimum of two disks." fi [[ -d /$TESTPOOL ]] && rm -rf /$TESTPOOL log_must zpool create -f $TESTPOOL raidz $disklist log_must zfs create $TESTPOOL/$TESTFS log_must zfs set mountpoint=$TESTDIR $TESTPOOL/$TESTFS } # # Common function used to cleanup storage pools and datasets. # # Invoked at the start of the test suite to ensure the system # is in a known state, and also at the end of each set of # sub-tests to ensure errors from one set of tests doesn't # impact the execution of the next set. function default_cleanup { default_cleanup_noexit log_pass } # # Utility function used to list all available pool names. # # NOTE: $KEEP is a variable containing pool names, separated by a newline # character, that must be excluded from the returned list. # function get_all_pools { zpool list -H -o name | grep -Fvx "$KEEP" | grep -v "$NO_POOLS" } function default_cleanup_noexit { typeset pool="" # # Destroying the pool will also destroy any # filesystems it contains. # if is_global_zone; then zfs unmount -a > /dev/null 2>&1 ALL_POOLS=$(get_all_pools) # Here, we loop through the pools we're allowed to # destroy, only destroying them if it's safe to do # so. while [ ! -z ${ALL_POOLS} ] do for pool in ${ALL_POOLS} do if safe_to_destroy_pool $pool ; then destroy_pool $pool fi done ALL_POOLS=$(get_all_pools) done zfs mount -a else typeset fs="" for fs in $(zfs list -H -o name \ | grep "^$ZONE_POOL/$ZONE_CTR[01234]/"); do destroy_dataset "$fs" "-Rf" done # Need cleanup here to avoid garbage dir left. for fs in $(zfs list -H -o name); do [[ $fs == /$ZONE_POOL ]] && continue [[ -d $fs ]] && log_must rm -rf $fs/* done # # Reset the $ZONE_POOL/$ZONE_CTR[01234] file systems property to # the default value # for fs in $(zfs list -H -o name); do if [[ $fs == $ZONE_POOL/$ZONE_CTR[01234] ]]; then log_must zfs set reservation=none $fs log_must zfs set recordsize=128K $fs log_must zfs set mountpoint=/$fs $fs typeset enc=$(get_prop encryption $fs) if [ -z "$enc" ] || [ "$enc" = "off" ]; then log_must zfs set checksum=on $fs fi log_must zfs set compression=off $fs log_must zfs set atime=on $fs log_must zfs set devices=off $fs log_must zfs set exec=on $fs log_must zfs set setuid=on $fs log_must zfs set readonly=off $fs log_must zfs set snapdir=hidden $fs log_must zfs set aclmode=groupmask $fs log_must zfs set aclinherit=secure $fs fi done fi [[ -d $TESTDIR ]] && \ log_must rm -rf $TESTDIR disk1=${DISKS%% *} if is_mpath_device $disk1; then delete_partitions fi rm -f $TEST_BASE_DIR/{err,out} } # # Common function used to cleanup storage pools, file systems # and containers. # function default_container_cleanup { if ! is_global_zone; then reexport_pool fi ismounted $TESTPOOL/$TESTCTR/$TESTFS1 && log_must zfs unmount $TESTPOOL/$TESTCTR/$TESTFS1 destroy_dataset "$TESTPOOL/$TESTCTR/$TESTFS1" "-R" destroy_dataset "$TESTPOOL/$TESTCTR" "-Rf" [[ -e $TESTDIR1 ]] && \ log_must rm -rf $TESTDIR1 default_cleanup } # # Common function used to cleanup snapshot of file system or volume. Default to # delete the file system's snapshot # # $1 snapshot name # function destroy_snapshot { typeset snap=${1:-$TESTPOOL/$TESTFS@$TESTSNAP} if ! snapexists $snap; then log_fail "'$snap' does not exist." fi # # For the sake of the value which come from 'get_prop' is not equal # to the really mountpoint when the snapshot is unmounted. So, firstly # check and make sure this snapshot's been mounted in current system. # typeset mtpt="" if ismounted $snap; then mtpt=$(get_prop mountpoint $snap) fi destroy_dataset "$snap" [[ $mtpt != "" && -d $mtpt ]] && \ log_must rm -rf $mtpt } # # Common function used to cleanup clone. # # $1 clone name # function destroy_clone { typeset clone=${1:-$TESTPOOL/$TESTCLONE} if ! datasetexists $clone; then log_fail "'$clone' does not existed." fi # With the same reason in destroy_snapshot typeset mtpt="" if ismounted $clone; then mtpt=$(get_prop mountpoint $clone) fi destroy_dataset "$clone" [[ $mtpt != "" && -d $mtpt ]] && \ log_must rm -rf $mtpt } # # Common function used to cleanup bookmark of file system or volume. Default # to delete the file system's bookmark. # # $1 bookmark name # function destroy_bookmark { typeset bkmark=${1:-$TESTPOOL/$TESTFS#$TESTBKMARK} if ! bkmarkexists $bkmark; then log_fail "'$bkmarkp' does not existed." fi destroy_dataset "$bkmark" } # Return 0 if a snapshot exists; $? otherwise # # $1 - snapshot name function snapexists { zfs list -H -t snapshot "$1" > /dev/null 2>&1 } # # Return 0 if a bookmark exists; $? otherwise # # $1 - bookmark name # function bkmarkexists { zfs list -H -t bookmark "$1" > /dev/null 2>&1 } # # Return 0 if a hold exists; $? otherwise # # $1 - hold tag # $2 - snapshot name # function holdexists { ! zfs holds "$2" | awk -v t="$1" '$2 ~ t { exit 1 }' } # # Set a property to a certain value on a dataset. # Sets a property of the dataset to the value as passed in. # @param: # $1 dataset who's property is being set # $2 property to set # $3 value to set property to # @return: # 0 if the property could be set. # non-zero otherwise. # @use: ZFS # function dataset_setprop { typeset fn=dataset_setprop if (($# < 3)); then log_note "$fn: Insufficient parameters (need 3, had $#)" return 1 fi typeset output= output=$(zfs set $2=$3 $1 2>&1) typeset rv=$? if ((rv != 0)); then log_note "Setting property on $1 failed." log_note "property $2=$3" log_note "Return Code: $rv" log_note "Output: $output" return $rv fi return 0 } # # Check a numeric assertion # @parameter: $@ the assertion to check # @output: big loud notice if assertion failed # @use: log_fail # function assert { (($@)) || log_fail "$@" } # # Function to format partition size of a disk # Given a disk cxtxdx reduces all partitions # to 0 size # function zero_partitions # { typeset diskname=$1 typeset i if is_freebsd; then gpart destroy -F $diskname elif is_linux; then DSK=$DEV_DSKDIR/$diskname DSK=$(echo $DSK | sed -e "s|//|/|g") log_must parted $DSK -s -- mklabel gpt blockdev --rereadpt $DSK 2>/dev/null block_device_wait else for i in 0 1 3 4 5 6 7 do log_must set_partition $i "" 0mb $diskname done fi return 0 } # # Given a slice, size and disk, this function # formats the slice to the specified size. # Size should be specified with units as per # the `format` command requirements eg. 100mb 3gb # # NOTE: This entire interface is problematic for the Linux parted utility # which requires the end of the partition to be specified. It would be # best to retire this interface and replace it with something more flexible. # At the moment a best effort is made. # # arguments: function set_partition { typeset -i slicenum=$1 typeset start=$2 typeset size=$3 typeset disk=${4#$DEV_DSKDIR/} disk=${disk#$DEV_RDSKDIR/} case "$UNAME" in Linux) if [[ -z $size || -z $disk ]]; then log_fail "The size or disk name is unspecified." fi disk=$DEV_DSKDIR/$disk typeset size_mb=${size%%[mMgG]} size_mb=${size_mb%%[mMgG][bB]} if [[ ${size:1:1} == 'g' ]]; then ((size_mb = size_mb * 1024)) fi # Create GPT partition table when setting slice 0 or # when the device doesn't already contain a GPT label. parted $disk -s -- print 1 >/dev/null typeset ret_val=$? if [[ $slicenum -eq 0 || $ret_val -ne 0 ]]; then if ! parted $disk -s -- mklabel gpt; then log_note "Failed to create GPT partition table on $disk" return 1 fi fi # When no start is given align on the first cylinder. if [[ -z "$start" ]]; then start=1 fi # Determine the cylinder size for the device and using # that calculate the end offset in cylinders. typeset -i cly_size_kb=0 cly_size_kb=$(parted -m $disk -s -- unit cyl print | awk -F '[:k.]' 'NR == 3 {print $4}') ((end = (size_mb * 1024 / cly_size_kb) + start)) parted $disk -s -- \ mkpart part$slicenum ${start}cyl ${end}cyl typeset ret_val=$? if [[ $ret_val -ne 0 ]]; then log_note "Failed to create partition $slicenum on $disk" return 1 fi blockdev --rereadpt $disk 2>/dev/null block_device_wait $disk ;; FreeBSD) if [[ -z $size || -z $disk ]]; then log_fail "The size or disk name is unspecified." fi disk=$DEV_DSKDIR/$disk if [[ $slicenum -eq 0 ]] || ! gpart show $disk >/dev/null 2>&1; then gpart destroy -F $disk >/dev/null 2>&1 if ! gpart create -s GPT $disk; then log_note "Failed to create GPT partition table on $disk" return 1 fi fi typeset index=$((slicenum + 1)) if [[ -n $start ]]; then start="-b $start" fi gpart add -t freebsd-zfs $start -s $size -i $index $disk if [[ $ret_val -ne 0 ]]; then log_note "Failed to create partition $slicenum on $disk" return 1 fi block_device_wait $disk ;; *) if [[ -z $slicenum || -z $size || -z $disk ]]; then log_fail "The slice, size or disk name is unspecified." fi typeset format_file=/var/tmp/format_in.$$ echo "partition" >$format_file echo "$slicenum" >> $format_file echo "" >> $format_file echo "" >> $format_file echo "$start" >> $format_file echo "$size" >> $format_file echo "label" >> $format_file echo "" >> $format_file echo "q" >> $format_file echo "q" >> $format_file format -e -s -d $disk -f $format_file typeset ret_val=$? rm -f $format_file ;; esac if [[ $ret_val -ne 0 ]]; then log_note "Unable to format $disk slice $slicenum to $size" return 1 fi return 0 } # # Delete all partitions on all disks - this is specifically for the use of multipath # devices which currently can only be used in the test suite as raw/un-partitioned # devices (ie a zpool cannot be created on a whole mpath device that has partitions) # function delete_partitions { typeset disk if [[ -z $DISKSARRAY ]]; then DISKSARRAY=$DISKS fi if is_linux; then typeset -i part for disk in $DISKSARRAY; do for (( part = 1; part < MAX_PARTITIONS; part++ )); do typeset partition=${disk}${SLICE_PREFIX}${part} parted $DEV_DSKDIR/$disk -s rm $part > /dev/null 2>&1 if lsblk | grep -qF ${partition}; then log_fail "Partition ${partition} not deleted" else log_note "Partition ${partition} deleted" fi done done elif is_freebsd; then for disk in $DISKSARRAY; do if gpart destroy -F $disk; then log_note "Partitions for ${disk} deleted" else log_fail "Partitions for ${disk} not deleted" fi done fi } # # Get the end cyl of the given slice # function get_endslice # { typeset disk=$1 typeset slice=$2 if [[ -z $disk || -z $slice ]] ; then log_fail "The disk name or slice number is unspecified." fi case "$UNAME" in Linux) endcyl=$(parted -s $DEV_DSKDIR/$disk -- unit cyl print | \ awk "/part${slice}/"' {sub(/cyl/, "", $3); print $3}') ((endcyl = (endcyl + 1))) ;; FreeBSD) disk=${disk#/dev/zvol/} disk=${disk%p*} slice=$((slice + 1)) endcyl=$(gpart show $disk | \ awk -v slice=$slice '$3 == slice { print $1 + $2 }') ;; *) disk=${disk#/dev/dsk/} disk=${disk#/dev/rdsk/} disk=${disk%s*} typeset -i ratio=0 ratio=$(prtvtoc /dev/rdsk/${disk}s2 | \ awk '/sectors\/cylinder/ {print $2}') if ((ratio == 0)); then return fi typeset -i endcyl=$(prtvtoc -h /dev/rdsk/${disk}s2 | awk -v token="$slice" '$1 == token {print $6}') ((endcyl = (endcyl + 1) / ratio)) ;; esac echo $endcyl } # # Given a size,disk and total slice number, this function formats the # disk slices from 0 to the total slice number with the same specified # size. # function partition_disk # { typeset -i i=0 typeset slice_size=$1 typeset disk_name=$2 typeset total_slices=$3 typeset cyl zero_partitions $disk_name while ((i < $total_slices)); do if ! is_linux; then if ((i == 2)); then ((i = i + 1)) continue fi fi log_must set_partition $i "$cyl" $slice_size $disk_name cyl=$(get_endslice $disk_name $i) ((i = i+1)) done } # # This function continues to write to a filenum number of files into dirnum # number of directories until either file_write returns an error or the # maximum number of files per directory have been written. # # Usage: # fill_fs [destdir] [dirnum] [filenum] [bytes] [num_writes] [data] # # Return value: 0 on success # non 0 on error # # Where : # destdir: is the directory where everything is to be created under # dirnum: the maximum number of subdirectories to use, -1 no limit # filenum: the maximum number of files per subdirectory # bytes: number of bytes to write # num_writes: number of types to write out bytes # data: the data that will be written # # E.g. # fill_fs /testdir 20 25 1024 256 0 # # Note: bytes * num_writes equals the size of the testfile # function fill_fs # destdir dirnum filenum bytes num_writes data { typeset destdir=${1:-$TESTDIR} typeset -i dirnum=${2:-50} typeset -i filenum=${3:-50} typeset -i bytes=${4:-8192} typeset -i num_writes=${5:-10240} typeset data=${6:-0} mkdir -p $destdir/{1..$dirnum} for f in $destdir/{1..$dirnum}/$TESTFILE{1..$filenum}; do file_write -o create -f $f -b $bytes -c $num_writes -d $data \ || return done } # Get the specified dataset property in parsable format or fail function get_prop # property dataset { typeset prop=$1 typeset dataset=$2 zfs get -Hpo value "$prop" "$dataset" || log_fail "zfs get $prop $dataset" } # Get the specified pool property in parsable format or fail function get_pool_prop # property pool { typeset prop=$1 typeset pool=$2 zpool get -Hpo value "$prop" "$pool" || log_fail "zpool get $prop $pool" } # Return 0 if a pool exists; $? otherwise # # $1 - pool name function poolexists { typeset pool=$1 if [[ -z $pool ]]; then log_note "No pool name given." return 1 fi zpool get name "$pool" > /dev/null 2>&1 } # Return 0 if all the specified datasets exist; $? otherwise # # $1-n dataset name function datasetexists { if (($# == 0)); then log_note "No dataset name given." return 1 fi zfs get name "$@" > /dev/null 2>&1 } # return 0 if none of the specified datasets exists, otherwise return 1. # # $1-n dataset name function datasetnonexists { if (($# == 0)); then log_note "No dataset name given." return 1 fi while (($# > 0)); do zfs list -H -t filesystem,snapshot,volume $1 > /dev/null 2>&1 \ && return 1 shift done return 0 } # FreeBSD breaks exports(5) at whitespace and doesn't process escapes # Solaris just breaks # # cf. https://github.com/openzfs/zfs/pull/13165#issuecomment-1059845807 # # Linux can have spaces (which are \OOO-escaped), # but can't have backslashes because they're parsed recursively function shares_can_have_whitespace { is_linux } function is_shared_freebsd { typeset fs=$1 pgrep -q mountd && showmount -E | grep -qx "$fs" } function is_shared_illumos { typeset fs=$1 typeset mtpt for mtpt in `share | awk '{print $2}'` ; do if [[ $mtpt == $fs ]] ; then return 0 fi done typeset stat=$(svcs -H -o STA nfs/server:default) if [[ $stat != "ON" ]]; then log_note "Current nfs/server status: $stat" fi return 1 } function is_shared_linux { typeset fs=$1 ! exportfs -s | awk -v fs="${fs//\\/\\\\}" '/^\// && $1 == fs {exit 1}' } # # Given a mountpoint, or a dataset name, determine if it is shared via NFS. # # Returns 0 if shared, 1 otherwise. # function is_shared { typeset fs=$1 typeset mtpt if [[ $fs != "/"* ]] ; then if datasetnonexists "$fs" ; then return 1 else mtpt=$(get_prop mountpoint "$fs") case "$mtpt" in none|legacy|-) return 1 ;; *) fs=$mtpt ;; esac fi fi case "$UNAME" in FreeBSD) is_shared_freebsd "$fs" ;; Linux) is_shared_linux "$fs" ;; *) is_shared_illumos "$fs" ;; esac } function is_exported_illumos { typeset fs=$1 typeset mtpt _ while read -r mtpt _; do [ "$mtpt" = "$fs" ] && return done < /etc/dfs/sharetab return 1 } function is_exported_freebsd { typeset fs=$1 typeset mtpt _ while read -r mtpt _; do [ "$mtpt" = "$fs" ] && return done < /etc/zfs/exports return 1 } function is_exported_linux { typeset fs=$1 typeset mtpt _ while read -r mtpt _; do [ "$(printf "$mtpt")" = "$fs" ] && return done < /etc/exports.d/zfs.exports return 1 } # # Given a mountpoint, or a dataset name, determine if it is exported via # the os-specific NFS exports file. # # Returns 0 if exported, 1 otherwise. # function is_exported { typeset fs=$1 typeset mtpt if [[ $fs != "/"* ]] ; then if datasetnonexists "$fs" ; then return 1 else mtpt=$(get_prop mountpoint "$fs") case $mtpt in none|legacy|-) return 1 ;; *) fs=$mtpt ;; esac fi fi case "$UNAME" in FreeBSD) is_exported_freebsd "$fs" ;; Linux) is_exported_linux "$fs" ;; *) is_exported_illumos "$fs" ;; esac } # # Given a dataset name determine if it is shared via SMB. # # Returns 0 if shared, 1 otherwise. # function is_shared_smb { typeset fs=$1 datasetexists "$fs" || return if is_linux; then net usershare list | grep -xFq "${fs//[-\/]/_}" else log_note "SMB on $UNAME currently unsupported by the test framework" return 1 fi } # # Given a mountpoint, determine if it is not shared via NFS. # # Returns 0 if not shared, 1 otherwise. # function not_shared { ! is_shared $1 } # # Given a dataset determine if it is not shared via SMB. # # Returns 0 if not shared, 1 otherwise. # function not_shared_smb { ! is_shared_smb $1 } # # Helper function to unshare a mountpoint. # function unshare_fs #fs { typeset fs=$1 if is_shared $fs || is_shared_smb $fs; then log_must zfs unshare $fs fi } # # Helper function to share a NFS mountpoint. # function share_nfs #fs { typeset fs=$1 is_shared "$fs" && return case "$UNAME" in Linux) log_must exportfs "*:$fs" ;; FreeBSD) typeset mountd read -r mountd < /var/run/mountd.pid log_must eval "printf '%s\t\n' \"$fs\" >> /etc/zfs/exports" log_must kill -s HUP "$mountd" ;; *) log_must share -F nfs "$fs" ;; esac return 0 } # # Helper function to unshare a NFS mountpoint. # function unshare_nfs #fs { typeset fs=$1 ! is_shared "$fs" && return case "$UNAME" in Linux) log_must exportfs -u "*:$fs" ;; FreeBSD) typeset mountd read -r mountd < /var/run/mountd.pid awk -v fs="${fs//\\/\\\\}" '$1 != fs' /etc/zfs/exports > /etc/zfs/exports.$$ log_must mv /etc/zfs/exports.$$ /etc/zfs/exports log_must kill -s HUP "$mountd" ;; *) log_must unshare -F nfs $fs ;; esac return 0 } # # Helper function to show NFS shares. # function showshares_nfs { case "$UNAME" in Linux) exportfs -v ;; FreeBSD) showmount ;; *) share -F nfs ;; esac } function check_nfs { case "$UNAME" in Linux) exportfs -s ;; FreeBSD) showmount -e ;; *) log_unsupported "Unknown platform" ;; esac || log_unsupported "The NFS utilities are not installed" } # # Check NFS server status and trigger it online. # function setup_nfs_server { # Cannot share directory in non-global zone. # if ! is_global_zone; then log_note "Cannot trigger NFS server by sharing in LZ." return fi if is_linux; then # # Re-synchronize /var/lib/nfs/etab with /etc/exports and # /etc/exports.d./* to provide a clean test environment. # log_must exportfs -r log_note "NFS server must be started prior to running ZTS." return elif is_freebsd; then log_must kill -s HUP $(/dev/null) [ $cur_zone = "global" ] fi } # # Verify whether test is permitted to run from # global zone, local zone, or both # # $1 zone limit, could be "global", "local", or "both"(no limit) # # Return 0 if permitted, otherwise exit with log_unsupported # function verify_runnable # zone limit { typeset limit=$1 [[ -z $limit ]] && return 0 if is_global_zone ; then case $limit in global|both) ;; local) log_unsupported "Test is unable to run from "\ "global zone." ;; *) log_note "Warning: unknown limit $limit - " \ "use both." ;; esac else case $limit in local|both) ;; global) log_unsupported "Test is unable to run from "\ "local zone." ;; *) log_note "Warning: unknown limit $limit - " \ "use both." ;; esac reexport_pool fi return 0 } # Return 0 if create successfully or the pool exists; $? otherwise # Note: In local zones, this function should return 0 silently. # # $1 - pool name # $2-n - [keyword] devs_list function create_pool #pool devs_list { typeset pool=${1%%/*} shift if [[ -z $pool ]]; then log_note "Missing pool name." return 1 fi if poolexists $pool ; then destroy_pool $pool fi if is_global_zone ; then [[ -d /$pool ]] && rm -rf /$pool log_must zpool create -f $pool $@ fi return 0 } # Return 0 if destroy successfully or the pool exists; $? otherwise # Note: In local zones, this function should return 0 silently. # # $1 - pool name # Destroy pool with the given parameters. function destroy_pool #pool { typeset pool=${1%%/*} typeset mtpt if [[ -z $pool ]]; then log_note "No pool name given." return 1 fi if is_global_zone ; then if poolexists "$pool" ; then mtpt=$(get_prop mountpoint "$pool") # At times, syseventd/udev activity can cause attempts # to destroy a pool to fail with EBUSY. We retry a few # times allowing failures before requiring the destroy # to succeed. log_must_busy zpool destroy -f $pool [[ -d $mtpt ]] && \ log_must rm -rf $mtpt else log_note "Pool does not exist. ($pool)" return 1 fi fi return 0 } # Return 0 if created successfully; $? otherwise # # $1 - dataset name # $2-n - dataset options function create_dataset #dataset dataset_options { typeset dataset=$1 shift if [[ -z $dataset ]]; then log_note "Missing dataset name." return 1 fi if datasetexists $dataset ; then destroy_dataset $dataset fi log_must zfs create $@ $dataset return 0 } # Return 0 if destroy successfully or the dataset exists; $? otherwise # Note: In local zones, this function should return 0 silently. # # $1 - dataset name # $2 - custom arguments for zfs destroy # Destroy dataset with the given parameters. function destroy_dataset # dataset [args] { typeset dataset=$1 typeset mtpt typeset args=${2:-""} if [[ -z $dataset ]]; then log_note "No dataset name given." return 1 fi if is_global_zone ; then if datasetexists "$dataset" ; then mtpt=$(get_prop mountpoint "$dataset") log_must_busy zfs destroy $args $dataset [ -d $mtpt ] && log_must rm -rf $mtpt else log_note "Dataset does not exist. ($dataset)" return 1 fi fi return 0 } # # Reexport TESTPOOL & TESTPOOL(1-4) # function reexport_pool { typeset -i cntctr=5 typeset -i i=0 while ((i < cntctr)); do if ((i == 0)); then TESTPOOL=$ZONE_POOL/$ZONE_CTR$i if ! ismounted $TESTPOOL; then log_must zfs mount $TESTPOOL fi else eval TESTPOOL$i=$ZONE_POOL/$ZONE_CTR$i if eval ! ismounted \$TESTPOOL$i; then log_must eval zfs mount \$TESTPOOL$i fi fi ((i += 1)) done } # # Verify a given disk or pool state # # Return 0 is pool/disk matches expected state, 1 otherwise # function check_state # pool disk state{online,offline,degraded} { typeset pool=$1 typeset disk=${2#$DEV_DSKDIR/} typeset state=$3 [[ -z $pool ]] || [[ -z $state ]] \ && log_fail "Arguments invalid or missing" if [[ -z $disk ]]; then #check pool state only zpool get -H -o value health $pool | grep -qi "$state" else zpool status -v $pool | grep "$disk" | grep -qi "$state" fi } # # Get the mountpoint of snapshot # For the snapshot use /.zfs/snapshot/ # as its mountpoint # function snapshot_mountpoint { typeset dataset=${1:-$TESTPOOL/$TESTFS@$TESTSNAP} if [[ $dataset != *@* ]]; then log_fail "Error name of snapshot '$dataset'." fi typeset fs=${dataset%@*} typeset snap=${dataset#*@} if [[ -z $fs || -z $snap ]]; then log_fail "Error name of snapshot '$dataset'." fi echo $(get_prop mountpoint $fs)/.zfs/snapshot/$snap } # # Given a device and 'ashift' value verify it's correctly set on every label # function verify_ashift # device ashift { typeset device="$1" typeset ashift="$2" zdb -e -lll $device | awk -v ashift=$ashift ' /ashift: / { if (ashift != $2) exit 1; else count++; } END { exit (count != 4); }' } # # Given a pool and file system, this function will verify the file system # using the zdb internal tool. Note that the pool is exported and imported # to ensure it has consistent state. # function verify_filesys # pool filesystem dir { typeset pool="$1" typeset filesys="$2" typeset zdbout="/tmp/zdbout.$$" shift shift typeset dirs=$@ typeset search_path="" log_note "Calling zdb to verify filesystem '$filesys'" zfs unmount -a > /dev/null 2>&1 log_must zpool export $pool if [[ -n $dirs ]] ; then for dir in $dirs ; do search_path="$search_path -d $dir" done fi log_must zpool import $search_path $pool if ! zdb -cudi $filesys > $zdbout 2>&1; then log_note "Output: zdb -cudi $filesys" cat $zdbout rm -f $zdbout log_fail "zdb detected errors with: '$filesys'" fi log_must zfs mount -a log_must rm -rf $zdbout } # # Given a pool issue a scrub and verify that no checksum errors are reported. # function verify_pool { typeset pool=${1:-$TESTPOOL} log_must zpool scrub $pool log_must wait_scrubbed $pool typeset -i cksum=$(zpool status $pool | awk ' !NF { isvdev = 0 } isvdev { errors += $NF } /CKSUM$/ { isvdev = 1 } END { print errors } ') if [[ $cksum != 0 ]]; then log_must zpool status -v log_fail "Unexpected CKSUM errors found on $pool ($cksum)" fi } # # Given a pool, and this function list all disks in the pool # function get_disklist # pool { echo $(zpool iostat -v $1 | awk '(NR > 4) {print $1}' | \ grep -vEe '^-----' -e "^(mirror|raidz[1-3]|draid[1-3]|spare|log|cache|special|dedup)|\-[0-9]$") } # # Given a pool, and this function list all disks in the pool with their full # path (like "/dev/sda" instead of "sda"). # function get_disklist_fullpath # pool { get_disklist "-P $1" } # /** # This function kills a given list of processes after a time period. We use # this in the stress tests instead of STF_TIMEOUT so that we can have processes # run for a fixed amount of time, yet still pass. Tests that hit STF_TIMEOUT # would be listed as FAIL, which we don't want : we're happy with stress tests # running for a certain amount of time, then finishing. # # @param $1 the time in seconds after which we should terminate these processes # @param $2..$n the processes we wish to terminate. # */ function stress_timeout { typeset -i TIMEOUT=$1 shift typeset cpids="$@" log_note "Waiting for child processes($cpids). " \ "It could last dozens of minutes, please be patient ..." log_must sleep $TIMEOUT log_note "Killing child processes after ${TIMEOUT} stress timeout." typeset pid for pid in $cpids; do ps -p $pid > /dev/null 2>&1 && log_must kill -USR1 $pid done } # # Verify a given hotspare disk is inuse or avail # # Return 0 is pool/disk matches expected state, 1 otherwise # function check_hotspare_state # pool disk state{inuse,avail} { typeset pool=$1 typeset disk=${2#$DEV_DSKDIR/} typeset state=$3 cur_state=$(get_device_state $pool $disk "spares") [ $state = $cur_state ] } # # Wait until a hotspare transitions to a given state or times out. # # Return 0 when pool/disk matches expected state, 1 on timeout. # function wait_hotspare_state # pool disk state timeout { typeset pool=$1 typeset disk=${2#*$DEV_DSKDIR/} typeset state=$3 typeset timeout=${4:-60} typeset -i i=0 while [[ $i -lt $timeout ]]; do if check_hotspare_state $pool $disk $state; then return 0 fi i=$((i+1)) sleep 1 done return 1 } # # Verify a given vdev disk is inuse or avail # # Return 0 is pool/disk matches expected state, 1 otherwise # function check_vdev_state # pool disk state{online,offline,unavail,removed} { typeset pool=$1 typeset disk=${2#*$DEV_DSKDIR/} typeset state=$3 cur_state=$(get_device_state $pool $disk) [ $state = $cur_state ] } # # Wait until a vdev transitions to a given state or times out. # # Return 0 when pool/disk matches expected state, 1 on timeout. # function wait_vdev_state # pool disk state timeout { typeset pool=$1 typeset disk=${2#*$DEV_DSKDIR/} typeset state=$3 typeset timeout=${4:-60} typeset -i i=0 while [[ $i -lt $timeout ]]; do if check_vdev_state $pool $disk $state; then return 0 fi i=$((i+1)) sleep 1 done return 1 } # # Check the output of 'zpool status -v ', # and to see if the content of contain the specified. # # Return 0 is contain, 1 otherwise # function check_pool_status # pool token keyword { typeset pool=$1 typeset token=$2 typeset keyword=$3 typeset verbose=${4:-false} scan=$(zpool status -v "$pool" 2>/dev/null | awk -v token="$token:" '$1==token') if [[ $verbose == true ]]; then log_note $scan fi echo $scan | grep -qi "$keyword" } # # The following functions are instance of check_pool_status() # is_pool_resilvering - to check if the pool resilver is in progress # is_pool_resilvered - to check if the pool resilver is completed # is_pool_scrubbing - to check if the pool scrub is in progress # is_pool_scrubbed - to check if the pool scrub is completed # is_pool_scrub_stopped - to check if the pool scrub is stopped # is_pool_scrub_paused - to check if the pool scrub has paused # is_pool_removing - to check if the pool removing is a vdev # is_pool_removed - to check if the pool remove is completed # is_pool_discarding - to check if the pool checkpoint is being discarded # is_pool_replacing - to check if the pool is performing a replacement # function is_pool_resilvering #pool { check_pool_status "$1" "scan" \ "resilver[ ()0-9A-Za-z:_-]* in progress since" $2 } function is_pool_resilvered #pool { check_pool_status "$1" "scan" "resilvered " $2 } function is_pool_scrubbing #pool { check_pool_status "$1" "scan" "scrub in progress since " $2 } function is_pool_error_scrubbing #pool { check_pool_status "$1" "scrub" "error scrub in progress since " $2 return $? } function is_pool_scrubbed #pool { check_pool_status "$1" "scan" "scrub repaired" $2 } function is_pool_scrub_stopped #pool { check_pool_status "$1" "scan" "scrub canceled" $2 } function is_pool_error_scrub_stopped #pool { check_pool_status "$1" "scrub" "error scrub canceled on " $2 return $? } function is_pool_scrub_paused #pool { check_pool_status "$1" "scan" "scrub paused since " $2 } function is_pool_error_scrub_paused #pool { check_pool_status "$1" "scrub" "error scrub paused since " $2 return $? } function is_pool_removing #pool { check_pool_status "$1" "remove" "in progress since " } function is_pool_removed #pool { check_pool_status "$1" "remove" "completed on" } function is_pool_discarding #pool { check_pool_status "$1" "checkpoint" "discarding" } function is_pool_replacing #pool { zpool status "$1" | grep -qE 'replacing-[0-9]+' } function wait_for_degraded { typeset pool=$1 typeset timeout=${2:-30} typeset t0=$SECONDS while :; do [[ $(get_pool_prop health $pool) == "DEGRADED" ]] && break log_note "$pool is not yet degraded." sleep 1 if ((SECONDS - t0 > $timeout)); then log_note "$pool not degraded after $timeout seconds." return 1 fi done return 0 } # # Use create_pool()/destroy_pool() to clean up the information in # in the given disk to avoid slice overlapping. # function cleanup_devices #vdevs { typeset pool="foopool$$" for vdev in $@; do zero_partitions $vdev done poolexists $pool && destroy_pool $pool create_pool $pool $@ destroy_pool $pool return 0 } #/** # A function to find and locate free disks on a system or from given # disks as the parameter. It works by locating disks that are in use # as swap devices and dump devices, and also disks listed in /etc/vfstab # # $@ given disks to find which are free, default is all disks in # the test system # # @return a string containing the list of available disks #*/ function find_disks { # Trust provided list, no attempt is made to locate unused devices. if is_linux || is_freebsd; then echo "$@" return fi sfi=/tmp/swaplist.$$ dmpi=/tmp/dumpdev.$$ max_finddisksnum=${MAX_FINDDISKSNUM:-6} swap -l > $sfi dumpadm > $dmpi 2>/dev/null disks=${@:-$(echo "" | format -e 2>/dev/null | awk ' BEGIN { FS="."; } /^Specify disk/{ searchdisks=0; } { if (searchdisks && $2 !~ "^$"){ split($2,arr," "); print arr[1]; } } /^AVAILABLE DISK SELECTIONS:/{ searchdisks=1; } ')} unused="" for disk in $disks; do # Check for mounted grep -q "${disk}[sp]" /etc/mnttab && continue # Check for swap grep -q "${disk}[sp]" $sfi && continue # check for dump device grep -q "${disk}[sp]" $dmpi && continue # check to see if this disk hasn't been explicitly excluded # by a user-set environment variable echo "${ZFS_HOST_DEVICES_IGNORE}" | grep -q "${disk}" && continue unused_candidates="$unused_candidates $disk" done rm $sfi $dmpi # now just check to see if those disks do actually exist # by looking for a device pointing to the first slice in # each case. limit the number to max_finddisksnum count=0 for disk in $unused_candidates; do if is_disk_device $DEV_DSKDIR/${disk}s0 && \ [ $count -lt $max_finddisksnum ]; then unused="$unused $disk" # do not impose limit if $@ is provided [[ -z $@ ]] && ((count = count + 1)) fi done # finally, return our disk list echo $unused } function add_user_freebsd # { typeset group=$1 typeset user=$2 typeset basedir=$3 # Check to see if the user exists. if id $user > /dev/null 2>&1; then return 0 fi # Assign 1000 as the base uid typeset -i uid=1000 while true; do pw useradd -u $uid -g $group -d $basedir/$user -m -n $user case $? in 0) break ;; # The uid is not unique 65) ((uid += 1)) ;; *) return 1 ;; esac if [[ $uid == 65000 ]]; then log_fail "No user id available under 65000 for $user" fi done # Silence MOTD touch $basedir/$user/.hushlogin return 0 } # # Delete the specified user. # # $1 login name # function del_user_freebsd # { typeset user=$1 if id $user > /dev/null 2>&1; then log_must pw userdel $user fi return 0 } # # Select valid gid and create specified group. # # $1 group name # function add_group_freebsd # { typeset group=$1 # See if the group already exists. if pw groupshow $group >/dev/null 2>&1; then return 0 fi # Assign 1000 as the base gid typeset -i gid=1000 while true; do pw groupadd -g $gid -n $group > /dev/null 2>&1 case $? in 0) return 0 ;; # The gid is not unique 65) ((gid += 1)) ;; *) return 1 ;; esac if [[ $gid == 65000 ]]; then log_fail "No user id available under 65000 for $group" fi done } # # Delete the specified group. # # $1 group name # function del_group_freebsd # { typeset group=$1 pw groupdel -n $group > /dev/null 2>&1 case $? in # Group does not exist, or was deleted successfully. 0|6|65) return 0 ;; # Name already exists as a group name 9) log_must pw groupdel $group ;; *) return 1 ;; esac return 0 } function add_user_illumos # { typeset group=$1 typeset user=$2 typeset basedir=$3 log_must useradd -g $group -d $basedir/$user -m $user return 0 } function del_user_illumos # { typeset user=$1 if id $user > /dev/null 2>&1; then log_must_retry "currently used" 6 userdel $user fi return 0 } function add_group_illumos # { typeset group=$1 typeset -i gid=100 while true; do groupadd -g $gid $group > /dev/null 2>&1 case $? in 0) return 0 ;; # The gid is not unique 4) ((gid += 1)) ;; *) return 1 ;; esac done } function del_group_illumos # { typeset group=$1 groupmod -n $grp $grp > /dev/null 2>&1 case $? in # Group does not exist. 6) return 0 ;; # Name already exists as a group name 9) log_must groupdel $grp ;; *) return 1 ;; esac } function add_user_linux # { typeset group=$1 typeset user=$2 typeset basedir=$3 log_must useradd -g $group -d $basedir/$user -m $user # Add new users to the same group and the command line utils. # This allows them to be run out of the original users home # directory as long as it permissioned to be group readable. cmd_group=$(stat --format="%G" $(command -v zfs)) log_must usermod -a -G $cmd_group $user return 0 } function del_user_linux # { typeset user=$1 if id $user > /dev/null 2>&1; then log_must_retry "currently used" 6 userdel $user fi } function add_group_linux # { typeset group=$1 # Assign 100 as the base gid, a larger value is selected for # Linux because for many distributions 1000 and under are reserved. while true; do groupadd $group > /dev/null 2>&1 case $? in 0) return 0 ;; *) return 1 ;; esac done } function del_group_linux # { typeset group=$1 getent group $group > /dev/null 2>&1 case $? in # Group does not exist. 2) return 0 ;; # Name already exists as a group name 0) log_must groupdel $group ;; *) return 1 ;; esac return 0 } # # Add specified user to specified group # # $1 group name # $2 user name # $3 base of the homedir (optional) # function add_user # { typeset group=$1 typeset user=$2 typeset basedir=${3:-"/var/tmp"} if ((${#group} == 0 || ${#user} == 0)); then log_fail "group name or user name are not defined." fi case "$UNAME" in FreeBSD) add_user_freebsd "$group" "$user" "$basedir" ;; Linux) add_user_linux "$group" "$user" "$basedir" ;; *) add_user_illumos "$group" "$user" "$basedir" ;; esac return 0 } # # Delete the specified user. # # $1 login name # $2 base of the homedir (optional) # function del_user # { typeset user=$1 typeset basedir=${2:-"/var/tmp"} if ((${#user} == 0)); then log_fail "login name is necessary." fi case "$UNAME" in FreeBSD) del_user_freebsd "$user" ;; Linux) del_user_linux "$user" ;; *) del_user_illumos "$user" ;; esac [[ -d $basedir/$user ]] && rm -fr $basedir/$user return 0 } # # Select valid gid and create specified group. # # $1 group name # function add_group # { typeset group=$1 if ((${#group} == 0)); then log_fail "group name is necessary." fi case "$UNAME" in FreeBSD) add_group_freebsd "$group" ;; Linux) add_group_linux "$group" ;; *) add_group_illumos "$group" ;; esac return 0 } # # Delete the specified group. # # $1 group name # function del_group # { typeset group=$1 if ((${#group} == 0)); then log_fail "group name is necessary." fi case "$UNAME" in FreeBSD) del_group_freebsd "$group" ;; Linux) del_group_linux "$group" ;; *) del_group_illumos "$group" ;; esac return 0 } # # This function will return true if it's safe to destroy the pool passed # as argument 1. It checks for pools based on zvols and files, and also # files contained in a pool that may have a different mountpoint. # function safe_to_destroy_pool { # $1 the pool name typeset pool="" typeset DONT_DESTROY="" # We check that by deleting the $1 pool, we're not # going to pull the rug out from other pools. Do this # by looking at all other pools, ensuring that they # aren't built from files or zvols contained in this pool. for pool in $(zpool list -H -o name) do ALTMOUNTPOOL="" # this is a list of the top-level directories in each of the # files that make up the path to the files the pool is based on FILEPOOL=$(zpool status -v $pool | awk -v pool="/$1/" '$0 ~ pool {print $1}') # this is a list of the zvols that make up the pool ZVOLPOOL=$(zpool status -v $pool | awk -v zvols="$ZVOL_DEVDIR/$1$" '$0 ~ zvols {print $1}') # also want to determine if it's a file-based pool using an # alternate mountpoint... POOL_FILE_DIRS=$(zpool status -v $pool | \ awk '/\// {print $1}' | \ awk -F/ '!/dev/ {print $2}') for pooldir in $POOL_FILE_DIRS do OUTPUT=$(zfs list -H -r -o mountpoint $1 | \ awk -v pd="${pooldir}$" '$0 ~ pd {print $1}') ALTMOUNTPOOL="${ALTMOUNTPOOL}${OUTPUT}" done if [ ! -z "$ZVOLPOOL" ] then DONT_DESTROY="true" log_note "Pool $pool is built from $ZVOLPOOL on $1" fi if [ ! -z "$FILEPOOL" ] then DONT_DESTROY="true" log_note "Pool $pool is built from $FILEPOOL on $1" fi if [ ! -z "$ALTMOUNTPOOL" ] then DONT_DESTROY="true" log_note "Pool $pool is built from $ALTMOUNTPOOL on $1" fi done if [ -z "${DONT_DESTROY}" ] then return 0 else log_note "Warning: it is not safe to destroy $1!" return 1 fi } # # Verify zfs operation with -p option work as expected # $1 operation, value could be create, clone or rename # $2 dataset type, value could be fs or vol # $3 dataset name # $4 new dataset name # function verify_opt_p_ops { typeset ops=$1 typeset datatype=$2 typeset dataset=$3 typeset newdataset=$4 if [[ $datatype != "fs" && $datatype != "vol" ]]; then log_fail "$datatype is not supported." fi # check parameters accordingly case $ops in create) newdataset=$dataset dataset="" if [[ $datatype == "vol" ]]; then ops="create -V $VOLSIZE" fi ;; clone) if [[ -z $newdataset ]]; then log_fail "newdataset should not be empty" \ "when ops is $ops." fi log_must datasetexists $dataset log_must snapexists $dataset ;; rename) if [[ -z $newdataset ]]; then log_fail "newdataset should not be empty" \ "when ops is $ops." fi log_must datasetexists $dataset ;; *) log_fail "$ops is not supported." ;; esac # make sure the upper level filesystem does not exist destroy_dataset "${newdataset%/*}" "-rRf" # without -p option, operation will fail log_mustnot zfs $ops $dataset $newdataset log_mustnot datasetexists $newdataset ${newdataset%/*} # with -p option, operation should succeed log_must zfs $ops -p $dataset $newdataset block_device_wait if ! datasetexists $newdataset ; then log_fail "-p option does not work for $ops" fi # when $ops is create or clone, redo the operation still return zero if [[ $ops != "rename" ]]; then log_must zfs $ops -p $dataset $newdataset fi return 0 } # # Get configuration of pool # $1 pool name # $2 config name # function get_config { typeset pool=$1 typeset config=$2 if ! poolexists "$pool" ; then return 1 fi if [ "$(get_pool_prop cachefile "$pool")" = "none" ]; then zdb -e $pool else zdb -C $pool fi | awk -F: -v cfg="$config:" '$0 ~ cfg {sub(/^'\''/, $2); sub(/'\''$/, $2); print $2}' } # # Privated function. Random select one of items from arguments. # # $1 count # $2-n string # function _random_get { typeset cnt=$1 shift typeset str="$@" typeset -i ind ((ind = RANDOM % cnt + 1)) echo "$str" | cut -f $ind -d ' ' } # # Random select one of item from arguments which include NONE string # function random_get_with_non { typeset -i cnt=$# ((cnt =+ 1)) _random_get "$cnt" "$@" } # # Random select one of item from arguments which doesn't include NONE string # function random_get { _random_get "$#" "$@" } # # The function will generate a dataset name with specific length # $1, the length of the name # $2, the base string to construct the name # function gen_dataset_name { typeset -i len=$1 typeset basestr="$2" typeset -i baselen=${#basestr} typeset -i iter=0 typeset l_name="" if ((len % baselen == 0)); then ((iter = len / baselen)) else ((iter = len / baselen + 1)) fi while ((iter > 0)); do l_name="${l_name}$basestr" ((iter -= 1)) done echo $l_name } # # Get cksum tuple of dataset # $1 dataset name # # sample zdb output: # Dataset data/test [ZPL], ID 355, cr_txg 2413856, 31.0K, 7 objects, rootbp # DVA[0]=<0:803046400:200> DVA[1]=<0:81199000:200> [L0 DMU objset] fletcher4 # lzjb LE contiguous unique double size=800L/200P birth=2413856L/2413856P # fill=7 cksum=11ce125712:643a9c18ee2:125e25238fca0:254a3f74b59744 function datasetcksum { typeset cksum sync sync_all_pools zdb -vvv $1 | awk -F= -v ds="^Dataset $1 "'\\[' '$0 ~ ds && /cksum/ {print $7}' } # # Get the given disk/slice state from the specific field of the pool # function get_device_state #pool disk field("", "spares","logs") { typeset pool=$1 typeset disk=${2#$DEV_DSKDIR/} typeset field=${3:-$pool} zpool status -v "$pool" 2>/dev/null | \ awk -v device=$disk -v pool=$pool -v field=$field \ 'BEGIN {startconfig=0; startfield=0; } /config:/ {startconfig=1} (startconfig==1) && ($1==field) {startfield=1; next;} (startfield==1) && ($1==device) {print $2; exit;} (startfield==1) && ($1==field || $1 ~ "^spares$" || $1 ~ "^logs$") {startfield=0}' } # # get the root filesystem name if it's zfsroot system. # # return: root filesystem name function get_rootfs { typeset rootfs="" if is_freebsd; then rootfs=$(mount -p | awk '$2 == "/" && $3 == "zfs" {print $1}') elif ! is_linux; then rootfs=$(awk '$2 == "/" && $3 == "zfs" {print $1}' \ /etc/mnttab) fi if [[ -z "$rootfs" ]]; then log_fail "Can not get rootfs" fi if datasetexists $rootfs; then echo $rootfs else log_fail "This is not a zfsroot system." fi } # # get the rootfs's pool name # return: # rootpool name # function get_rootpool { typeset rootfs=$(get_rootfs) echo ${rootfs%%/*} } # # To verify if the require numbers of disks is given # function verify_disk_count { typeset -i min=${2:-1} typeset -i count=$(echo "$1" | wc -w) if ((count < min)); then log_untested "A minimum of $min disks is required to run." \ " You specified $count disk(s)" fi } function ds_is_volume { typeset type=$(get_prop type $1) [ $type = "volume" ] } function ds_is_filesystem { typeset type=$(get_prop type $1) [ $type = "filesystem" ] } # # Check if Trusted Extensions are installed and enabled # function is_te_enabled { svcs -H -o state labeld 2>/dev/null | grep -q "enabled" } # Return the number of CPUs (cross-platform) function get_num_cpus { if is_linux ; then grep -c '^processor' /proc/cpuinfo elif is_freebsd; then sysctl -n kern.smp.cpus else psrinfo | wc -l fi } # Utility function to determine if a system has multiple cpus. function is_mp { [[ $(get_num_cpus) -gt 1 ]] } function get_cpu_freq { if is_linux; then lscpu | awk '/CPU MHz/ { print $3 }' elif is_freebsd; then sysctl -n hw.clockrate else psrinfo -v 0 | awk '/processor operates at/ {print $6}' fi } # Run the given command as the user provided. function user_run { typeset user=$1 shift log_note "user: $user" log_note "cmd: $*" typeset out=$TEST_BASE_DIR/out typeset err=$TEST_BASE_DIR/err sudo -Eu $user env PATH="$PATH" ksh <<<"$*" >$out 2>$err typeset res=$? log_note "out: $(<$out)" log_note "err: $(<$err)" return $res } # # Check if the pool contains the specified vdevs # # $1 pool # $2..n ... # # Return 0 if the vdevs are contained in the pool, 1 if any of the specified # vdevs is not in the pool, and 2 if pool name is missing. # function vdevs_in_pool { typeset pool=$1 typeset vdev if [[ -z $pool ]]; then log_note "Missing pool name." return 2 fi shift # We could use 'zpool list' to only get the vdevs of the pool but we # can't reference a mirror/raidz vdev using its ID (i.e mirror-0), # therefore we use the 'zpool status' output. typeset tmpfile=$(mktemp) zpool status -v "$pool" | grep -A 1000 "config:" >$tmpfile for vdev in "$@"; do grep -wq ${vdev##*/} $tmpfile || return 1 done rm -f $tmpfile return 0 } function get_max { typeset -l i max=$1 shift for i in "$@"; do max=$((max > i ? max : i)) done echo $max } # Write data that can be compressed into a directory function write_compressible { typeset dir=$1 typeset megs=$2 typeset nfiles=${3:-1} typeset bs=${4:-1024k} typeset fname=${5:-file} [[ -d $dir ]] || log_fail "No directory: $dir" # Under Linux fio is not currently used since its behavior can # differ significantly across versions. This includes missing # command line options and cases where the --buffer_compress_* # options fail to behave as expected. if is_linux; then typeset file_bytes=$(to_bytes $megs) typeset bs_bytes=4096 typeset blocks=$(($file_bytes / $bs_bytes)) for (( i = 0; i < $nfiles; i++ )); do truncate -s $file_bytes $dir/$fname.$i # Write every third block to get 66% compression. for (( j = 0; j < $blocks; j += 3 )); do dd if=/dev/urandom of=$dir/$fname.$i \ seek=$j bs=$bs_bytes count=1 \ conv=notrunc >/dev/null 2>&1 done done else command -v fio > /dev/null || log_unsupported "fio missing" log_must eval fio \ --name=job \ --fallocate=0 \ --minimal \ --randrepeat=0 \ --buffer_compress_percentage=66 \ --buffer_compress_chunk=4096 \ --directory="$dir" \ --numjobs="$nfiles" \ --nrfiles="$nfiles" \ --rw=write \ --bs="$bs" \ --filesize="$megs" \ "--filename_format='$fname.\$jobnum' >/dev/null" fi } function get_objnum { typeset pathname=$1 typeset objnum [[ -e $pathname ]] || log_fail "No such file or directory: $pathname" if is_freebsd; then objnum=$(stat -f "%i" $pathname) else objnum=$(stat -c %i $pathname) fi echo $objnum } # # Sync data to the pool # # $1 pool name # $2 boolean to force uberblock (and config including zpool cache file) update # function sync_pool #pool { typeset pool=${1:-$TESTPOOL} typeset force=${2:-false} if [[ $force == true ]]; then log_must zpool sync -f $pool else log_must zpool sync $pool fi return 0 } # # Sync all pools # # $1 boolean to force uberblock (and config including zpool cache file) update # function sync_all_pools # { typeset force=${1:-false} if [[ $force == true ]]; then log_must zpool sync -f else log_must zpool sync fi return 0 } # # Wait for zpool 'freeing' property drops to zero. # # $1 pool name # function wait_freeing #pool { typeset pool=${1:-$TESTPOOL} while true; do [[ "0" == "$(zpool list -Ho freeing $pool)" ]] && break log_must sleep 1 done } # # Wait for every device replace operation to complete # # $1 pool name # $2 timeout # function wait_replacing #pool timeout { typeset timeout=${2:-300} typeset pool=${1:-$TESTPOOL} for (( timer = 0; timer < $timeout; timer++ )); do is_pool_replacing $pool || break; sleep 1; done } # Wait for a pool to be scrubbed # # $1 pool name # $2 timeout # function wait_scrubbed #pool timeout { typeset timeout=${2:-300} typeset pool=${1:-$TESTPOOL} for (( timer = 0; timer < $timeout; timer++ )); do is_pool_scrubbed $pool && break; sleep 1; done } # Backup the zed.rc in our test directory so that we can edit it for our test. # # Returns: Backup file name. You will need to pass this to zed_rc_restore(). function zed_rc_backup { zedrc_backup="$(mktemp)" cp $ZEDLET_DIR/zed.rc $zedrc_backup echo $zedrc_backup } function zed_rc_restore { mv $1 $ZEDLET_DIR/zed.rc } # # Setup custom environment for the ZED. # # $@ Optional list of zedlets to run under zed. function zed_setup { if ! is_linux; then log_unsupported "No zed on $UNAME" fi if [[ ! -d $ZEDLET_DIR ]]; then log_must mkdir $ZEDLET_DIR fi if [[ ! -e $VDEVID_CONF ]]; then log_must touch $VDEVID_CONF fi if [[ -e $VDEVID_CONF_ETC ]]; then log_fail "Must not have $VDEVID_CONF_ETC file present on system" fi EXTRA_ZEDLETS=$@ # Create a symlink for /etc/zfs/vdev_id.conf file. log_must ln -s $VDEVID_CONF $VDEVID_CONF_ETC # Setup minimal ZED configuration. Individual test cases should # add additional ZEDLETs as needed for their specific test. log_must cp ${ZEDLET_ETC_DIR}/zed.rc $ZEDLET_DIR log_must cp ${ZEDLET_ETC_DIR}/zed-functions.sh $ZEDLET_DIR # Scripts must only be user writable. if [[ -n "$EXTRA_ZEDLETS" ]] ; then saved_umask=$(umask) log_must umask 0022 for i in $EXTRA_ZEDLETS ; do log_must cp ${ZEDLET_LIBEXEC_DIR}/$i $ZEDLET_DIR done log_must umask $saved_umask fi # Customize the zed.rc file to enable the full debug log. log_must sed -i '/\#ZED_DEBUG_LOG=.*/d' $ZEDLET_DIR/zed.rc echo "ZED_DEBUG_LOG=$ZED_DEBUG_LOG" >>$ZEDLET_DIR/zed.rc } # # Cleanup custom ZED environment. # # $@ Optional list of zedlets to remove from our test zed.d directory. function zed_cleanup { if ! is_linux; then return fi for extra_zedlet; do log_must rm -f ${ZEDLET_DIR}/$extra_zedlet done log_must rm -fd ${ZEDLET_DIR}/zed.rc ${ZEDLET_DIR}/zed-functions.sh ${ZEDLET_DIR}/all-syslog.sh ${ZEDLET_DIR}/all-debug.sh ${ZEDLET_DIR}/state \ $ZED_LOG $ZED_DEBUG_LOG $VDEVID_CONF_ETC $VDEVID_CONF \ $ZEDLET_DIR } # # Check if ZED is currently running; if so, returns PIDs # function zed_check { if ! is_linux; then return fi zedpids="$(pgrep -x zed)" zedpids2="$(pgrep -x lt-zed)" echo ${zedpids} ${zedpids2} } # # Check if ZED is currently running, if not start ZED. # function zed_start { if ! is_linux; then return fi # ZEDLET_DIR=/var/tmp/zed if [[ ! -d $ZEDLET_DIR ]]; then log_must mkdir $ZEDLET_DIR fi # Verify the ZED is not already running. zedpids=$(zed_check) if [ -n "$zedpids" ]; then # We never, ever, really want it to just keep going if zed # is already running - usually this implies our test cases # will break very strangely because whatever we wanted to # configure zed for won't be listening to our changes in the # tmpdir log_fail "ZED already running - ${zedpids}" else log_note "Starting ZED" # run ZED in the background and redirect foreground logging # output to $ZED_LOG. log_must truncate -s 0 $ZED_DEBUG_LOG log_must eval "zed -vF -d $ZEDLET_DIR -P $PATH" \ "-s $ZEDLET_DIR/state -j 1 2>$ZED_LOG &" fi return 0 } # # Kill ZED process # function zed_stop { if ! is_linux; then return "" fi log_note "Stopping ZED" while true; do zedpids=$(zed_check) [ ! -n "$zedpids" ] && break log_must kill $zedpids sleep 1 done return 0 } # # Drain all zevents # function zed_events_drain { while [ $(zpool events -H | wc -l) -ne 0 ]; do sleep 1 zpool events -c >/dev/null done } # Set a variable in zed.rc to something, un-commenting it in the process. # # $1 variable # $2 value function zed_rc_set { var="$1" val="$2" # Remove the line cmd="'/$var/d'" eval sed -i $cmd $ZEDLET_DIR/zed.rc # Add it at the end echo "$var=$val" >> $ZEDLET_DIR/zed.rc } # # Check is provided device is being active used as a swap device. # function is_swap_inuse { typeset device=$1 if [[ -z $device ]] ; then log_note "No device specified." return 1 fi case "$UNAME" in Linux) swapon -s | grep -wq $(readlink -f $device) ;; FreeBSD) swapctl -l | grep -wq $device ;; *) swap -l | grep -wq $device ;; esac } # # Setup a swap device using the provided device. # function swap_setup { typeset swapdev=$1 case "$UNAME" in Linux) log_must eval "mkswap $swapdev > /dev/null 2>&1" log_must swapon $swapdev ;; FreeBSD) log_must swapctl -a $swapdev ;; *) log_must swap -a $swapdev ;; esac return 0 } # # Cleanup a swap device on the provided device. # function swap_cleanup { typeset swapdev=$1 if is_swap_inuse $swapdev; then if is_linux; then log_must swapoff $swapdev elif is_freebsd; then log_must swapoff $swapdev else log_must swap -d $swapdev fi fi return 0 } # # Set a global system tunable (64-bit value) # # $1 tunable name (use a NAME defined in tunables.cfg) # $2 tunable values # function set_tunable64 { set_tunable_impl "$1" "$2" Z } # # Set a global system tunable (32-bit value) # # $1 tunable name (use a NAME defined in tunables.cfg) # $2 tunable values # function set_tunable32 { set_tunable_impl "$1" "$2" W } function set_tunable_impl { typeset name="$1" typeset value="$2" typeset mdb_cmd="$3" eval "typeset tunable=\$$name" case "$tunable" in UNSUPPORTED) log_unsupported "Tunable '$name' is unsupported on $UNAME" ;; "") log_fail "Tunable '$name' must be added to tunables.cfg" ;; *) ;; esac [[ -z "$value" ]] && return 1 [[ -z "$mdb_cmd" ]] && return 1 case "$UNAME" in Linux) typeset zfs_tunables="/sys/module/zfs/parameters" echo "$value" >"$zfs_tunables/$tunable" ;; FreeBSD) sysctl vfs.zfs.$tunable=$value ;; SunOS) echo "${tunable}/${mdb_cmd}0t${value}" | mdb -kw ;; esac } +function save_tunable +{ + [[ ! -d $TEST_BASE_DIR ]] && return 1 + [[ -e $TEST_BASE_DIR/tunable-$1 ]] && return 2 + echo "$(get_tunable """$1""")" > "$TEST_BASE_DIR"/tunable-"$1" +} + +function restore_tunable +{ + [[ ! -e $TEST_BASE_DIR/tunable-$1 ]] && return 1 + val="$(cat $TEST_BASE_DIR/tunable-"""$1""")" + set_tunable64 "$1" "$val" + rm $TEST_BASE_DIR/tunable-$1 +} + # # Get a global system tunable # # $1 tunable name (use a NAME defined in tunables.cfg) # function get_tunable { get_tunable_impl "$1" } function get_tunable_impl { typeset name="$1" typeset module="${2:-zfs}" typeset check_only="$3" eval "typeset tunable=\$$name" case "$tunable" in UNSUPPORTED) if [ -z "$check_only" ] ; then log_unsupported "Tunable '$name' is unsupported on $UNAME" else return 1 fi ;; "") if [ -z "$check_only" ] ; then log_fail "Tunable '$name' must be added to tunables.cfg" else return 1 fi ;; *) ;; esac case "$UNAME" in Linux) typeset zfs_tunables="/sys/module/$module/parameters" cat $zfs_tunables/$tunable ;; FreeBSD) sysctl -n vfs.zfs.$tunable ;; SunOS) [[ "$module" -eq "zfs" ]] || return 1 ;; esac } # Does a tunable exist? # # $1: Tunable name function tunable_exists { get_tunable_impl $1 "zfs" 1 } # # Compute MD5 digest for given file or stdin if no file given. # Note: file path must not contain spaces # function md5digest { typeset file=$1 case "$UNAME" in FreeBSD) md5 -q $file ;; *) typeset sum _ read -r sum _ < <(md5sum -b $file) echo $sum ;; esac } # # Compute SHA256 digest for given file or stdin if no file given. # Note: file path must not contain spaces # function sha256digest { typeset file=$1 case "$UNAME" in FreeBSD) sha256 -q $file ;; *) typeset sum _ read -r sum _ < <(sha256sum -b $file) echo $sum ;; esac } function new_fs # { case "$UNAME" in FreeBSD) newfs "$@" ;; *) echo y | newfs -v "$@" ;; esac } function stat_size # { typeset path=$1 case "$UNAME" in FreeBSD) stat -f %z "$path" ;; *) stat -c %s "$path" ;; esac } function stat_mtime # { typeset path=$1 case "$UNAME" in FreeBSD) stat -f %m "$path" ;; *) stat -c %Y "$path" ;; esac } function stat_ctime # { typeset path=$1 case "$UNAME" in FreeBSD) stat -f %c "$path" ;; *) stat -c %Z "$path" ;; esac } function stat_crtime # { typeset path=$1 case "$UNAME" in FreeBSD) stat -f %B "$path" ;; *) stat -c %W "$path" ;; esac } function stat_generation # { typeset path=$1 case "$UNAME" in Linux) getversion "${path}" ;; *) stat -f %v "${path}" ;; esac } # Run a command as if it was being run in a TTY. # # Usage: # # faketty command # function faketty { if is_freebsd; then script -q /dev/null env "$@" else script --return --quiet -c "$*" /dev/null fi } # # Produce a random permutation of the integers in a given range (inclusive). # function range_shuffle # begin end { typeset -i begin=$1 typeset -i end=$2 seq ${begin} ${end} | sort -R } # # Cross-platform xattr helpers # function get_xattr # name path { typeset name=$1 typeset path=$2 case "$UNAME" in FreeBSD) getextattr -qq user "${name}" "${path}" ;; *) attr -qg "${name}" "${path}" ;; esac } function set_xattr # name value path { typeset name=$1 typeset value=$2 typeset path=$3 case "$UNAME" in FreeBSD) setextattr user "${name}" "${value}" "${path}" ;; *) attr -qs "${name}" -V "${value}" "${path}" ;; esac } function set_xattr_stdin # name value { typeset name=$1 typeset path=$2 case "$UNAME" in FreeBSD) setextattr -i user "${name}" "${path}" ;; *) attr -qs "${name}" "${path}" ;; esac } function rm_xattr # name path { typeset name=$1 typeset path=$2 case "$UNAME" in FreeBSD) rmextattr -q user "${name}" "${path}" ;; *) attr -qr "${name}" "${path}" ;; esac } function ls_xattr # path { typeset path=$1 case "$UNAME" in FreeBSD) lsextattr -qq user "${path}" ;; *) attr -ql "${path}" ;; esac } function kstat # stat flags? { typeset stat=$1 typeset flags=${2-"-n"} case "$UNAME" in FreeBSD) sysctl $flags kstat.zfs.misc.$stat ;; Linux) cat "/proc/spl/kstat/zfs/$stat" 2>/dev/null ;; *) false ;; esac } function get_arcstat # stat { typeset stat=$1 case "$UNAME" in FreeBSD) kstat arcstats.$stat ;; Linux) kstat arcstats | awk "/$stat/"' { print $3 }' ;; *) false ;; esac } function punch_hole # offset length file { typeset offset=$1 typeset length=$2 typeset file=$3 case "$UNAME" in FreeBSD) truncate -d -o $offset -l $length "$file" ;; Linux) fallocate --punch-hole --offset $offset --length $length "$file" ;; *) false ;; esac } function zero_range # offset length file { typeset offset=$1 typeset length=$2 typeset file=$3 case "$UNAME" in Linux) fallocate --zero-range --offset $offset --length $length "$file" ;; *) false ;; esac } # # Wait for the specified arcstat to reach non-zero quiescence. # If echo is 1 echo the value after reaching quiescence, otherwise # if echo is 0 print the arcstat we are waiting on. # function arcstat_quiescence # stat echo { typeset stat=$1 typeset echo=$2 typeset do_once=true if [[ $echo -eq 0 ]]; then echo "Waiting for arcstat $1 quiescence." fi while $do_once || [ $stat1 -ne $stat2 ] || [ $stat2 -eq 0 ]; do typeset stat1=$(get_arcstat $stat) sleep 0.5 typeset stat2=$(get_arcstat $stat) do_once=false done if [[ $echo -eq 1 ]]; then echo $stat2 fi } function arcstat_quiescence_noecho # stat { typeset stat=$1 arcstat_quiescence $stat 0 } function arcstat_quiescence_echo # stat { typeset stat=$1 arcstat_quiescence $stat 1 } # # Given an array of pids, wait until all processes # have completed and check their return status. # function wait_for_children #children { rv=0 children=("$@") for child in "${children[@]}" do child_exit=0 wait ${child} || child_exit=$? if [ $child_exit -ne 0 ]; then echo "child ${child} failed with ${child_exit}" rv=1 fi done return $rv } # # Compare two directory trees recursively in a manner similar to diff(1), but # using rsync. If there are any discrepancies, a summary of the differences are # output and a non-zero error is returned. # # If you're comparing a directory after a ZIL replay, you should set # LIBTEST_DIFF_ZIL_REPLAY=1 or use replay_directory_diff which will cause # directory_diff to ignore mtime changes (the ZIL replay won't fix up mtime # information). # function directory_diff # dir_a dir_b { dir_a="$1" dir_b="$2" zil_replay="${LIBTEST_DIFF_ZIL_REPLAY:-0}" # If one of the directories doesn't exist, return 2. This is to match the # semantics of diff. if ! [ -d "$dir_a" -a -d "$dir_b" ]; then return 2 fi # Run rsync with --dry-run --itemize-changes to get something akin to diff # output, but rsync is far more thorough in detecting differences (diff # doesn't compare file metadata, and cannot handle special files). # # Also make sure to filter out non-user.* xattrs when comparing. On # SELinux-enabled systems the copied tree will probably have different # SELinux labels. args=("-nicaAHX" '--filter=-x! user.*' "--delete") # NOTE: Quite a few rsync builds do not support --crtimes which would be # necessary to verify that creation times are being maintained properly. # Unfortunately because of this we cannot use it unconditionally but we can # check if this rsync build supports it and use it then. This check is # based on the same check in the rsync test suite (testsuite/crtimes.test). # # We check ctimes even with zil_replay=1 because the ZIL does store # creation times and we should make sure they match (if the creation times # do not match there is a "c" entry in one of the columns). if rsync --version | grep -q "[, ] crtimes"; then args+=("--crtimes") else log_note "This rsync package does not support --crtimes (-N)." fi # If we are testing a ZIL replay, we need to ignore timestamp changes. # Unfortunately --no-times doesn't do what we want -- it will still tell # you if the timestamps don't match but rsync will set the timestamps to # the current time (leading to an itemised change entry). It's simpler to # just filter out those lines. if [ "$zil_replay" -eq 0 ]; then filter=("cat") else # Different rsync versions have different numbers of columns. So just # require that aside from the first two, all other columns must be # blank (literal ".") or a timestamp field ("[tT]"). filter=("grep" "-v" '^\..[.Tt]\+ ') fi diff="$(rsync "${args[@]}" "$dir_a/" "$dir_b/" | "${filter[@]}")" rv=0 if [ -n "$diff" ]; then echo "$diff" rv=1 fi return $rv } # # Compare two directory trees recursively, without checking whether the mtimes # match (creation times will be checked if the available rsync binary supports # it). This is necessary for ZIL replay checks (because the ZIL does not # contain mtimes and thus after a ZIL replay, mtimes won't match). # # This is shorthand for LIBTEST_DIFF_ZIL_REPLAY=1 directory_diff <...>. # function replay_directory_diff # dir_a dir_b { LIBTEST_DIFF_ZIL_REPLAY=1 directory_diff "$@" } # # Put coredumps into $1/core.{basename} # # Output must be saved and passed to pop_coredump_pattern on cleanup # function push_coredump_pattern # dir { ulimit -c unlimited case "$UNAME" in Linux) cat /proc/sys/kernel/core_pattern /proc/sys/kernel/core_uses_pid echo "$1/core.%e" >/proc/sys/kernel/core_pattern && echo 0 >/proc/sys/kernel/core_uses_pid ;; FreeBSD) sysctl -n kern.corefile sysctl kern.corefile="$1/core.%N" >/dev/null ;; *) # Nothing to output – set only for this shell coreadm -p "$1/core.%f" ;; esac } # # Put coredumps back into the default location # function pop_coredump_pattern { [ -s "$1" ] || return 0 case "$UNAME" in Linux) typeset pat pid { read -r pat; read -r pid; } < "$1" echo "$pat" >/proc/sys/kernel/core_pattern && echo "$pid" >/proc/sys/kernel/core_uses_pid ;; FreeBSD) sysctl kern.corefile="$(<"$1")" >/dev/null ;; esac } diff --git a/tests/zfs-tests/include/tunables.cfg b/tests/zfs-tests/include/tunables.cfg index fb861f1a285c..e4e380aa7fd5 100644 --- a/tests/zfs-tests/include/tunables.cfg +++ b/tests/zfs-tests/include/tunables.cfg @@ -1,104 +1,105 @@ # This file exports variables for each tunable used in the test suite. # # Different platforms use different names for most tunables. To avoid littering # the tests with conditional logic for deciding how to set each tunable, the # logic is instead consolidated to this one file. # # Any use of tunables in tests must use a name defined here. New entries # should be added to the table as needed. Please keep the table sorted # alphabetically for ease of maintenance. # # Platform-specific tunables should still use a NAME from this table for # consistency. Enter UNSUPPORTED in the column for platforms on which the # tunable is not implemented. UNAME=$(uname) # NAME FreeBSD tunable Linux tunable cat <<%%%% | ADMIN_SNAPSHOT UNSUPPORTED zfs_admin_snapshot ALLOW_REDACTED_DATASET_MOUNT allow_redacted_dataset_mount zfs_allow_redacted_dataset_mount ARC_MAX arc.max zfs_arc_max ARC_MIN arc.min zfs_arc_min ASYNC_BLOCK_MAX_BLOCKS async_block_max_blocks zfs_async_block_max_blocks CHECKSUM_EVENTS_PER_SECOND checksum_events_per_second zfs_checksum_events_per_second COMMIT_TIMEOUT_PCT commit_timeout_pct zfs_commit_timeout_pct COMPRESSED_ARC_ENABLED compressed_arc_enabled zfs_compressed_arc_enabled CONDENSE_INDIRECT_COMMIT_ENTRY_DELAY_MS condense.indirect_commit_entry_delay_ms zfs_condense_indirect_commit_entry_delay_ms CONDENSE_INDIRECT_OBSOLETE_PCT condense.indirect_obsolete_pct zfs_condense_indirect_obsolete_pct CONDENSE_MIN_MAPPING_BYTES condense.min_mapping_bytes zfs_condense_min_mapping_bytes DBUF_CACHE_SHIFT dbuf.cache_shift dbuf_cache_shift DEADMAN_CHECKTIME_MS deadman.checktime_ms zfs_deadman_checktime_ms DEADMAN_FAILMODE deadman.failmode zfs_deadman_failmode DEADMAN_SYNCTIME_MS deadman.synctime_ms zfs_deadman_synctime_ms DEADMAN_ZIOTIME_MS deadman.ziotime_ms zfs_deadman_ziotime_ms DISABLE_IVSET_GUID_CHECK disable_ivset_guid_check zfs_disable_ivset_guid_check DMU_OFFSET_NEXT_SYNC dmu_offset_next_sync zfs_dmu_offset_next_sync EMBEDDED_SLOG_MIN_MS embedded_slog_min_ms zfs_embedded_slog_min_ms INITIALIZE_CHUNK_SIZE initialize_chunk_size zfs_initialize_chunk_size INITIALIZE_VALUE initialize_value zfs_initialize_value KEEP_LOG_SPACEMAPS_AT_EXPORT keep_log_spacemaps_at_export zfs_keep_log_spacemaps_at_export LUA_MAX_MEMLIMIT lua.max_memlimit zfs_lua_max_memlimit L2ARC_MFUONLY l2arc.mfuonly l2arc_mfuonly L2ARC_NOPREFETCH l2arc.noprefetch l2arc_noprefetch L2ARC_REBUILD_BLOCKS_MIN_L2SIZE l2arc.rebuild_blocks_min_l2size l2arc_rebuild_blocks_min_l2size L2ARC_REBUILD_ENABLED l2arc.rebuild_enabled l2arc_rebuild_enabled L2ARC_TRIM_AHEAD l2arc.trim_ahead l2arc_trim_ahead L2ARC_WRITE_BOOST l2arc.write_boost l2arc_write_boost L2ARC_WRITE_MAX l2arc.write_max l2arc_write_max LIVELIST_CONDENSE_NEW_ALLOC livelist.condense.new_alloc zfs_livelist_condense_new_alloc LIVELIST_CONDENSE_SYNC_CANCEL livelist.condense.sync_cancel zfs_livelist_condense_sync_cancel LIVELIST_CONDENSE_SYNC_PAUSE livelist.condense.sync_pause zfs_livelist_condense_sync_pause LIVELIST_CONDENSE_ZTHR_CANCEL livelist.condense.zthr_cancel zfs_livelist_condense_zthr_cancel LIVELIST_CONDENSE_ZTHR_PAUSE livelist.condense.zthr_pause zfs_livelist_condense_zthr_pause LIVELIST_MAX_ENTRIES livelist.max_entries zfs_livelist_max_entries LIVELIST_MIN_PERCENT_SHARED livelist.min_percent_shared zfs_livelist_min_percent_shared MAX_DATASET_NESTING max_dataset_nesting zfs_max_dataset_nesting MAX_MISSING_TVDS max_missing_tvds zfs_max_missing_tvds METASLAB_DEBUG_LOAD metaslab.debug_load metaslab_debug_load METASLAB_FORCE_GANGING metaslab.force_ganging metaslab_force_ganging MULTIHOST_FAIL_INTERVALS multihost.fail_intervals zfs_multihost_fail_intervals MULTIHOST_HISTORY multihost.history zfs_multihost_history MULTIHOST_IMPORT_INTERVALS multihost.import_intervals zfs_multihost_import_intervals MULTIHOST_INTERVAL multihost.interval zfs_multihost_interval OVERRIDE_ESTIMATE_RECORDSIZE send.override_estimate_recordsize zfs_override_estimate_recordsize PREFETCH_DISABLE prefetch.disable zfs_prefetch_disable RAIDZ_EXPAND_MAX_REFLOW_BYTES vdev.expand_max_reflow_bytes raidz_expand_max_reflow_bytes REBUILD_SCRUB_ENABLED rebuild_scrub_enabled zfs_rebuild_scrub_enabled REMOVAL_SUSPEND_PROGRESS removal_suspend_progress zfs_removal_suspend_progress REMOVE_MAX_SEGMENT remove_max_segment zfs_remove_max_segment RESILVER_MIN_TIME_MS resilver_min_time_ms zfs_resilver_min_time_ms SCAN_LEGACY scan_legacy zfs_scan_legacy SCAN_SUSPEND_PROGRESS scan_suspend_progress zfs_scan_suspend_progress SCAN_VDEV_LIMIT scan_vdev_limit zfs_scan_vdev_limit SCRUB_AFTER_EXPAND scrub_after_expand zfs_scrub_after_expand SEND_HOLES_WITHOUT_BIRTH_TIME send_holes_without_birth_time send_holes_without_birth_time SLOW_IO_EVENTS_PER_SECOND slow_io_events_per_second zfs_slow_io_events_per_second SPA_ASIZE_INFLATION spa.asize_inflation spa_asize_inflation SPA_DISCARD_MEMORY_LIMIT spa.discard_memory_limit zfs_spa_discard_memory_limit SPA_LOAD_VERIFY_DATA spa.load_verify_data spa_load_verify_data SPA_LOAD_VERIFY_METADATA spa.load_verify_metadata spa_load_verify_metadata TRIM_EXTENT_BYTES_MIN trim.extent_bytes_min zfs_trim_extent_bytes_min TRIM_METASLAB_SKIP trim.metaslab_skip zfs_trim_metaslab_skip TRIM_TXG_BATCH trim.txg_batch zfs_trim_txg_batch TXG_HISTORY txg.history zfs_txg_history TXG_TIMEOUT txg.timeout zfs_txg_timeout UNLINK_SUSPEND_PROGRESS UNSUPPORTED zfs_unlink_suspend_progress VDEV_FILE_LOGICAL_ASHIFT vdev.file.logical_ashift vdev_file_logical_ashift VDEV_FILE_PHYSICAL_ASHIFT vdev.file.physical_ashift vdev_file_physical_ashift VDEV_MAX_AUTO_ASHIFT vdev.max_auto_ashift zfs_vdev_max_auto_ashift VDEV_MIN_MS_COUNT vdev.min_ms_count zfs_vdev_min_ms_count VDEV_VALIDATE_SKIP vdev.validate_skip vdev_validate_skip VOL_INHIBIT_DEV UNSUPPORTED zvol_inhibit_dev VOL_MODE vol.mode zvol_volmode VOL_RECURSIVE vol.recursive UNSUPPORTED VOL_USE_BLK_MQ UNSUPPORTED zvol_use_blk_mq +BCLONE_ENABLED zfs_bclone_enabled zfs_bclone_enabled XATTR_COMPAT xattr_compat zfs_xattr_compat ZEVENT_LEN_MAX zevent.len_max zfs_zevent_len_max ZEVENT_RETAIN_MAX zevent.retain_max zfs_zevent_retain_max ZIO_SLOW_IO_MS zio.slow_io_ms zio_slow_io_ms ZIL_SAXATTR zil_saxattr zfs_zil_saxattr %%%% while read name FreeBSD Linux; do eval "export ${name}=\$${UNAME}" done diff --git a/tests/zfs-tests/tests/functional/block_cloning/cleanup.ksh b/tests/zfs-tests/tests/functional/block_cloning/cleanup.ksh index 7ac13adb6325..b985445a5d12 100755 --- a/tests/zfs-tests/tests/functional/block_cloning/cleanup.ksh +++ b/tests/zfs-tests/tests/functional/block_cloning/cleanup.ksh @@ -1,34 +1,38 @@ #!/bin/ksh -p # # CDDL HEADER START # # The contents of this file are subject to the terms of the # Common Development and Distribution License (the "License"). # You may not use this file except in compliance with the License. # # You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE # or https://opensource.org/licenses/CDDL-1.0. # See the License for the specific language governing permissions # and limitations under the License. # # When distributing Covered Code, include this CDDL HEADER in each # file and include the License file at usr/src/OPENSOLARIS.LICENSE. # If applicable, add the following below this CDDL HEADER, with the # fields enclosed by brackets "[]" replaced with your own identifying # information: Portions Copyright [yyyy] [name of copyright owner] # # CDDL HEADER END # # # Copyright (c) 2023, Klara Inc. # . $STF_SUITE/include/libtest.shlib . $STF_SUITE/tests/functional/block_cloning/block_cloning.kshlib verify_runnable "global" default_cleanup_noexit +if tunable_exists BCLONE_ENABLED ; then + log_must restore_tunable BCLONE_ENABLED +fi + log_pass diff --git a/tests/zfs-tests/tests/functional/block_cloning/setup.ksh b/tests/zfs-tests/tests/functional/block_cloning/setup.ksh index 512f5a0644df..58441bf8f3ad 100755 --- a/tests/zfs-tests/tests/functional/block_cloning/setup.ksh +++ b/tests/zfs-tests/tests/functional/block_cloning/setup.ksh @@ -1,36 +1,41 @@ #!/bin/ksh -p # # CDDL HEADER START # # The contents of this file are subject to the terms of the # Common Development and Distribution License (the "License"). # You may not use this file except in compliance with the License. # # You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE # or https://opensource.org/licenses/CDDL-1.0. # See the License for the specific language governing permissions # and limitations under the License. # # When distributing Covered Code, include this CDDL HEADER in each # file and include the License file at usr/src/OPENSOLARIS.LICENSE. # If applicable, add the following below this CDDL HEADER, with the # fields enclosed by brackets "[]" replaced with your own identifying # information: Portions Copyright [yyyy] [name of copyright owner] # # CDDL HEADER END # # # Copyright (c) 2023, Klara Inc. # . $STF_SUITE/include/libtest.shlib . $STF_SUITE/tests/functional/block_cloning/block_cloning.kshlib if ! command -v clonefile > /dev/null ; then log_unsupported "clonefile program required to test block cloning" fi verify_runnable "global" +if tunable_exists BCLONE_ENABLED ; then + log_must save_tunable BCLONE_ENABLED + log_must set_tunable32 BCLONE_ENABLED 1 +fi + log_pass