diff options
| author | pjd <pjd@FreeBSD.org> | 2008-11-17 20:49:29 +0000 |
|---|---|---|
| committer | pjd <pjd@FreeBSD.org> | 2008-11-17 20:49:29 +0000 |
| commit | bbe899b96e388a8b82439f81ed3707e0d9c6070d (patch) | |
| tree | 81b89fa4ac6467771d5aa291a97f4665981a6108 /cddl | |
| parent | d2f579595c362ce27b4d87e2c40e1c4e09b929e3 (diff) | |
| download | FreeBSD-src-bbe899b96e388a8b82439f81ed3707e0d9c6070d.zip FreeBSD-src-bbe899b96e388a8b82439f81ed3707e0d9c6070d.tar.gz | |
Update ZFS from version 6 to 13 and bring some FreeBSD-specific changes.
This bring huge amount of changes, I'll enumerate only user-visible changes:
- Delegated Administration
Allows regular users to perform ZFS operations, like file system
creation, snapshot creation, etc.
- L2ARC
Level 2 cache for ZFS - allows to use additional disks for cache.
Huge performance improvements mostly for random read of mostly
static content.
- slog
Allow to use additional disks for ZFS Intent Log to speed up
operations like fsync(2).
- vfs.zfs.super_owner
Allows regular users to perform privileged operations on files stored
on ZFS file systems owned by him. Very careful with this one.
- chflags(2)
Not all the flags are supported. This still needs work.
- ZFSBoot
Support to boot off of ZFS pool. Not finished, AFAIK.
Submitted by: dfr
- Snapshot properties
- New failure modes
Before if write requested failed, system paniced. Now one
can select from one of three failure modes:
- panic - panic on write error
- wait - wait for disk to reappear
- continue - serve read requests if possible, block write requests
- Refquota, refreservation properties
Just quota and reservation properties, but don't count space consumed
by children file systems, clones and snapshots.
- Sparse volumes
ZVOLs that don't reserve space in the pool.
- External attributes
Compatible with extattr(2).
- NFSv4-ACLs
Not sure about the status, might not be complete yet.
Submitted by: trasz
- Creation-time properties
- Regression tests for zpool(8) command.
Obtained from: OpenSolaris
Diffstat (limited to 'cddl')
56 files changed, 15477 insertions, 4787 deletions
diff --git a/cddl/compat/opensolaris/include/libshare.h b/cddl/compat/opensolaris/include/libshare.h new file mode 100644 index 0000000..82aacad --- /dev/null +++ b/cddl/compat/opensolaris/include/libshare.h @@ -0,0 +1,36 @@ +/*- + * Copyright (c) 2007 Pawel Jakub Dawidek <pjd@FreeBSD.org> + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + * notice, this list of conditions and the following disclaimer in the + * documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHORS AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE LIABLE + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + * + * $FreeBSD$ + */ + +#ifndef _OPENSOLARIS_LIBSHARE_H_ +#define _OPENSOLARIS_LIBSHARE_H_ + +#define SA_OK 0 + +#define SA_INIT_CONTROL_API 0 + +#endif /* !_OPENSOLARIS_LIBSHARE_H_ */ diff --git a/cddl/compat/opensolaris/include/mnttab.h b/cddl/compat/opensolaris/include/mnttab.h index 9d0bfc7..abd2f9d 100644 --- a/cddl/compat/opensolaris/include/mnttab.h +++ b/cddl/compat/opensolaris/include/mnttab.h @@ -9,6 +9,8 @@ #define MNTTAB _PATH_DEVNULL #define MNT_LINE_MAX 1024 +#define umount2(p, f) unmount(p, f) + struct mnttab { char *mnt_special; char *mnt_mountp; diff --git a/cddl/compat/opensolaris/misc/mnttab.c b/cddl/compat/opensolaris/misc/mnttab.c index c1971ca..8c1c2d6 100644 --- a/cddl/compat/opensolaris/misc/mnttab.c +++ b/cddl/compat/opensolaris/misc/mnttab.c @@ -37,6 +37,8 @@ __FBSDID("$FreeBSD$"); #include <sys/mntent.h> #include <sys/mnttab.h> #include <stdio.h> +#include <stdlib.h> +#include <string.h> static char * mntopt(char **p) diff --git a/cddl/compat/opensolaris/misc/zmount.c b/cddl/compat/opensolaris/misc/zmount.c index d80a3b3..493a4fc 100644 --- a/cddl/compat/opensolaris/misc/zmount.c +++ b/cddl/compat/opensolaris/misc/zmount.c @@ -35,9 +35,10 @@ __FBSDID("$FreeBSD$"); #include <sys/mount.h> #include <sys/uio.h> #include <sys/mntent.h> +#include <assert.h> #include <stdio.h> #include <stdlib.h> -#include <assert.h> +#include <string.h> static void build_iovec(struct iovec **iov, int *iovlen, const char *name, void *val, @@ -86,7 +87,7 @@ zmount(const char *spec, const char *dir, int mflag, char *fstype, assert(optlen > 0); optstr = strdup(optptr); - assert(optptr != NULL); + assert(optstr != NULL); iov = NULL; iovlen = 0; diff --git a/cddl/compat/opensolaris/misc/zone.c b/cddl/compat/opensolaris/misc/zone.c index 1ce77cf..9d48e6d 100644 --- a/cddl/compat/opensolaris/misc/zone.c +++ b/cddl/compat/opensolaris/misc/zone.c @@ -32,7 +32,7 @@ #include <sys/sysctl.h> #include <sys/zone.h> -int +zoneid_t getzoneid(void) { size_t size; @@ -42,5 +42,5 @@ getzoneid(void) size = sizeof(jailid); if (sysctlbyname("security.jail.jailed", &jailid, &size, NULL, 0) == -1) assert(!"No security.jail.jailed sysctl!"); - return (jailid); + return ((zoneid_t)jailid); } diff --git a/cddl/contrib/opensolaris/cmd/zdb/zdb.8 b/cddl/contrib/opensolaris/cmd/zdb/zdb.8 index 87913f6..c9d5aed 100644 --- a/cddl/contrib/opensolaris/cmd/zdb/zdb.8 +++ b/cddl/contrib/opensolaris/cmd/zdb/zdb.8 @@ -28,13 +28,17 @@ zdb \- ZFS debugger .fi .SH DESCRIPTION +.sp .LP The \fBzdb\fR command is used by support engineers to diagnose failures and gather statistics. Since the \fBZFS\fR file system is always consistent on disk and is self-repairing, \fBzdb\fR should only be run under the direction by a support engineer. +.sp .LP If no arguments are specified, \fBzdb\fR, performs basic consistency checks on the pool and associated datasets, and report any problems detected. +.sp .LP Any options supported by this command are internal to Sun and subject to change at any time. .SH EXIT STATUS +.sp .LP The following exit values are returned: .sp @@ -71,6 +75,7 @@ Invalid command line options were specified. .RE .SH ATTRIBUTES +.sp .LP See \fBattributes\fR(5) for descriptions of the following attributes: .sp @@ -89,5 +94,6 @@ Interface StabilityUnstable .TE .SH SEE ALSO +.sp .LP \fBzfs\fR(1M), \fBzpool\fR(1M), \fBattributes\fR(5) diff --git a/cddl/contrib/opensolaris/cmd/zdb/zdb.c b/cddl/contrib/opensolaris/cmd/zdb/zdb.c index 2dc459d..16b2787 100644 --- a/cddl/contrib/opensolaris/cmd/zdb/zdb.c +++ b/cddl/contrib/opensolaris/cmd/zdb/zdb.c @@ -19,12 +19,10 @@ * CDDL HEADER END */ /* - * Copyright 2007 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ -#pragma ident "%Z%%M% %I% %E% SMI" - #include <stdio.h> #include <stdio_ext.h> #include <stdlib.h> @@ -51,6 +49,10 @@ #include <sys/dmu_traverse.h> #include <sys/zio_checksum.h> #include <sys/zio_compress.h> +#include <sys/zfs_fuid.h> +#undef ZFS_MAXNAMELEN +#undef verify +#include <libzfs.h> const char cmdname[] = "zdb"; uint8_t dump_opt[256]; @@ -62,6 +64,9 @@ uint64_t *zopt_object = NULL; int zopt_objects = 0; int zdb_advance = ADVANCE_PRE; zbookmark_t zdb_noread = { 0, 0, ZB_NO_LEVEL, 0 }; +libzfs_handle_t *g_zfs; +boolean_t zdb_sig_user_data = B_TRUE; +int zdb_sig_cksumalg = ZIO_CHECKSUM_SHA256; /* * These libumem hooks provide a reasonable set of defaults for the allocator's @@ -83,12 +88,15 @@ static void usage(void) { (void) fprintf(stderr, - "Usage: %s [-udibcsvLU] [-O order] [-B os:obj:level:blkid] " + "Usage: %s [-udibcsvL] [-U cachefile_path] [-O order] " + "[-B os:obj:level:blkid] [-S user:cksumalg] " "dataset [object...]\n" " %s -C [pool]\n" " %s -l dev\n" - " %s -R vdev:offset:size:flags\n", - cmdname, cmdname, cmdname, cmdname); + " %s -R pool:vdev:offset:size:flags\n" + " %s [-p path_to_vdev_dir]\n" + " %s -e pool | GUID | devid ...\n", + cmdname, cmdname, cmdname, cmdname, cmdname, cmdname); (void) fprintf(stderr, " -u uberblock\n"); (void) fprintf(stderr, " -d datasets\n"); @@ -97,16 +105,22 @@ usage(void) (void) fprintf(stderr, " -b block statistics\n"); (void) fprintf(stderr, " -c checksum all data blocks\n"); (void) fprintf(stderr, " -s report stats on zdb's I/O\n"); + (void) fprintf(stderr, " -S <user|all>:<cksum_alg|all> -- " + "dump blkptr signatures\n"); (void) fprintf(stderr, " -v verbose (applies to all others)\n"); (void) fprintf(stderr, " -l dump label contents\n"); (void) fprintf(stderr, " -L live pool (allows some errors)\n"); (void) fprintf(stderr, " -O [!]<pre|post|prune|data|holes> " "visitation order\n"); - (void) fprintf(stderr, " -U use zpool.cache in /tmp\n"); + (void) fprintf(stderr, " -U cachefile_path -- use alternate " + "cachefile\n"); (void) fprintf(stderr, " -B objset:object:level:blkid -- " "simulate bad block\n"); - (void) fprintf(stderr, " -R read and display block from a" + (void) fprintf(stderr, " -R read and display block from a " "device\n"); + (void) fprintf(stderr, " -e Pool is exported/destroyed/" + "has altroot\n"); + (void) fprintf(stderr, " -p <Path to vdev dir> (use with -e)\n"); (void) fprintf(stderr, "Specify an option more than once (e.g. -bb) " "to make only that option verbose\n"); (void) fprintf(stderr, "Default is to dump everything non-verbosely\n"); @@ -367,6 +381,44 @@ dump_zap(objset_t *os, uint64_t object, void *data, size_t size) zap_cursor_fini(&zc); } +/*ARGSUSED*/ +static void +dump_zpldir(objset_t *os, uint64_t object, void *data, size_t size) +{ + zap_cursor_t zc; + zap_attribute_t attr; + const char *typenames[] = { + /* 0 */ "not specified", + /* 1 */ "FIFO", + /* 2 */ "Character Device", + /* 3 */ "3 (invalid)", + /* 4 */ "Directory", + /* 5 */ "5 (invalid)", + /* 6 */ "Block Device", + /* 7 */ "7 (invalid)", + /* 8 */ "Regular File", + /* 9 */ "9 (invalid)", + /* 10 */ "Symbolic Link", + /* 11 */ "11 (invalid)", + /* 12 */ "Socket", + /* 13 */ "Door", + /* 14 */ "Event Port", + /* 15 */ "15 (invalid)", + }; + + dump_zap_stats(os, object); + (void) printf("\n"); + + for (zap_cursor_init(&zc, os, object); + zap_cursor_retrieve(&zc, &attr) == 0; + zap_cursor_advance(&zc)) { + (void) printf("\t\t%s = %lld (type: %s)\n", + attr.za_name, ZFS_DIRENT_OBJ(attr.za_first_integer), + typenames[ZFS_DIRENT_TYPE(attr.za_first_integer)]); + } + zap_cursor_fini(&zc); +} + static void dump_spacemap(objset_t *os, space_map_obj_t *smo, space_map_t *sm) { @@ -456,10 +508,7 @@ dump_metaslabs(spa_t *spa) for (c = 0; c < rvd->vdev_children; c++) { vd = rvd->vdev_child[c]; - spa_config_enter(spa, RW_READER, FTAG); - (void) printf("\n vdev %llu = %s\n\n", - (u_longlong_t)vd->vdev_id, vdev_description(vd)); - spa_config_exit(spa, FTAG); + (void) printf("\n vdev %llu\n\n", (u_longlong_t)vd->vdev_id); if (dump_opt['d'] <= 5) { (void) printf("\t%10s %10s %5s\n", @@ -477,7 +526,6 @@ static void dump_dtl(vdev_t *vd, int indent) { avl_tree_t *t = &vd->vdev_dtl_map.sm_root; - spa_t *spa = vd->vdev_spa; space_seg_t *ss; vdev_t *pvd; int c; @@ -485,9 +533,10 @@ dump_dtl(vdev_t *vd, int indent) if (indent == 0) (void) printf("\nDirty time logs:\n\n"); - spa_config_enter(spa, RW_READER, FTAG); - (void) printf("\t%*s%s\n", indent, "", vdev_description(vd)); - spa_config_exit(spa, FTAG); + (void) printf("\t%*s%s\n", indent, "", + vd->vdev_path ? vd->vdev_path : + vd->vdev_parent ? vd->vdev_ops->vdev_op_type : + spa_name(vd->vdev_spa)); for (ss = avl_first(t); ss; ss = AVL_NEXT(t, ss)) { /* @@ -670,36 +719,49 @@ dump_dsl_dir(objset_t *os, uint64_t object, void *data, size_t size) { dsl_dir_phys_t *dd = data; time_t crtime; - char used[6], compressed[6], uncompressed[6], quota[6], resv[6]; + char nice[6]; if (dd == NULL) return; - ASSERT(size == sizeof (*dd)); + ASSERT3U(size, >=, sizeof (dsl_dir_phys_t)); crtime = dd->dd_creation_time; - nicenum(dd->dd_used_bytes, used); - nicenum(dd->dd_compressed_bytes, compressed); - nicenum(dd->dd_uncompressed_bytes, uncompressed); - nicenum(dd->dd_quota, quota); - nicenum(dd->dd_reserved, resv); - (void) printf("\t\tcreation_time = %s", ctime(&crtime)); (void) printf("\t\thead_dataset_obj = %llu\n", (u_longlong_t)dd->dd_head_dataset_obj); (void) printf("\t\tparent_dir_obj = %llu\n", (u_longlong_t)dd->dd_parent_obj); - (void) printf("\t\tclone_parent_obj = %llu\n", - (u_longlong_t)dd->dd_clone_parent_obj); + (void) printf("\t\torigin_obj = %llu\n", + (u_longlong_t)dd->dd_origin_obj); (void) printf("\t\tchild_dir_zapobj = %llu\n", (u_longlong_t)dd->dd_child_dir_zapobj); - (void) printf("\t\tused_bytes = %s\n", used); - (void) printf("\t\tcompressed_bytes = %s\n", compressed); - (void) printf("\t\tuncompressed_bytes = %s\n", uncompressed); - (void) printf("\t\tquota = %s\n", quota); - (void) printf("\t\treserved = %s\n", resv); + nicenum(dd->dd_used_bytes, nice); + (void) printf("\t\tused_bytes = %s\n", nice); + nicenum(dd->dd_compressed_bytes, nice); + (void) printf("\t\tcompressed_bytes = %s\n", nice); + nicenum(dd->dd_uncompressed_bytes, nice); + (void) printf("\t\tuncompressed_bytes = %s\n", nice); + nicenum(dd->dd_quota, nice); + (void) printf("\t\tquota = %s\n", nice); + nicenum(dd->dd_reserved, nice); + (void) printf("\t\treserved = %s\n", nice); (void) printf("\t\tprops_zapobj = %llu\n", (u_longlong_t)dd->dd_props_zapobj); + (void) printf("\t\tdeleg_zapobj = %llu\n", + (u_longlong_t)dd->dd_deleg_zapobj); + (void) printf("\t\tflags = %llx\n", + (u_longlong_t)dd->dd_flags); + +#define DO(which) \ + nicenum(dd->dd_used_breakdown[DD_USED_ ## which], nice); \ + (void) printf("\t\tused_breakdown[" #which "] = %s\n", nice) + DO(HEAD); + DO(SNAP); + DO(CHILD); + DO(CHILD_RSRV); + DO(REFRSRV); +#undef DO } /*ARGSUSED*/ @@ -722,7 +784,7 @@ dump_dsl_dataset(objset_t *os, uint64_t object, void *data, size_t size) nicenum(ds->ds_unique_bytes, unique); sprintf_blkptr(blkbuf, BP_SPRINTF_LEN, &ds->ds_bp); - (void) printf("\t\tdataset_obj = %llu\n", + (void) printf("\t\tdir_obj = %llu\n", (u_longlong_t)ds->ds_dir_obj); (void) printf("\t\tprev_snap_obj = %llu\n", (u_longlong_t)ds->ds_prev_snap_obj); @@ -749,6 +811,10 @@ dump_dsl_dataset(objset_t *os, uint64_t object, void *data, size_t size) (u_longlong_t)ds->ds_guid); (void) printf("\t\tflags = %llx\n", (u_longlong_t)ds->ds_flags); + (void) printf("\t\tnext_clones_obj = %llu\n", + (u_longlong_t)ds->ds_next_clones_obj); + (void) printf("\t\tprops_obj = %llu\n", + (u_longlong_t)ds->ds_props_obj); (void) printf("\t\tbp = %s\n", blkbuf); } @@ -765,9 +831,11 @@ dump_bplist(objset_t *mos, uint64_t object, char *name) if (dump_opt['d'] < 3) return; + mutex_init(&bpl.bpl_lock, NULL, MUTEX_DEFAULT, NULL); VERIFY(0 == bplist_open(&bpl, mos, object)); if (bplist_empty(&bpl)) { bplist_close(&bpl); + mutex_destroy(&bpl.bpl_lock); return; } @@ -785,6 +853,7 @@ dump_bplist(objset_t *mos, uint64_t object, char *name) if (dump_opt['d'] < 5) { bplist_close(&bpl); + mutex_destroy(&bpl.bpl_lock); return; } @@ -800,6 +869,65 @@ dump_bplist(objset_t *mos, uint64_t object, char *name) } bplist_close(&bpl); + mutex_destroy(&bpl.bpl_lock); +} + +static avl_tree_t idx_tree; +static avl_tree_t domain_tree; +static boolean_t fuid_table_loaded; + +static void +fuid_table_destroy() +{ + if (fuid_table_loaded) { + zfs_fuid_table_destroy(&idx_tree, &domain_tree); + fuid_table_loaded = B_FALSE; + } +} + +/* + * print uid or gid information. + * For normal POSIX id just the id is printed in decimal format. + * For CIFS files with FUID the fuid is printed in hex followed by + * the doman-rid string. + */ +static void +print_idstr(uint64_t id, const char *id_type) +{ + if (FUID_INDEX(id)) { + char *domain; + + domain = zfs_fuid_idx_domain(&idx_tree, FUID_INDEX(id)); + (void) printf("\t%s %llx [%s-%d]\n", id_type, + (u_longlong_t)id, domain, (int)FUID_RID(id)); + } else { + (void) printf("\t%s %llu\n", id_type, (u_longlong_t)id); + } + +} + +static void +dump_uidgid(objset_t *os, znode_phys_t *zp) +{ + uint32_t uid_idx, gid_idx; + + uid_idx = FUID_INDEX(zp->zp_uid); + gid_idx = FUID_INDEX(zp->zp_gid); + + /* Load domain table, if not already loaded */ + if (!fuid_table_loaded && (uid_idx || gid_idx)) { + uint64_t fuid_obj; + + /* first find the fuid object. It lives in the master node */ + VERIFY(zap_lookup(os, MASTER_NODE_OBJ, ZFS_FUID_TABLES, + 8, 1, &fuid_obj) == 0); + (void) zfs_fuid_table_load(os, fuid_obj, + &idx_tree, &domain_tree); + fuid_table_loaded = B_TRUE; + } + + print_idstr(zp->zp_uid, "uid"); + print_idstr(zp->zp_gid, "gid"); } /*ARGSUSED*/ @@ -830,6 +958,7 @@ dump_znode(objset_t *os, uint64_t object, void *data, size_t size) z_ctime = (time_t)zp->zp_ctime[0]; (void) printf("\tpath %s\n", path); + dump_uidgid(os, zp); (void) printf("\tatime %s", ctime(&z_atime)); (void) printf("\tmtime %s", ctime(&z_mtime)); (void) printf("\tctime %s", ctime(&z_ctime)); @@ -874,9 +1003,9 @@ static object_viewer_t *object_viewer[DMU_OT_NUMTYPES] = { dump_zap, /* DSL props */ dump_dsl_dataset, /* DSL dataset */ dump_znode, /* ZFS znode */ - dump_acl, /* ZFS ACL */ + dump_acl, /* ZFS V0 ACL */ dump_uint8, /* ZFS plain file */ - dump_zap, /* ZFS directory */ + dump_zpldir, /* ZFS directory */ dump_zap, /* ZFS master node */ dump_zap, /* ZFS delete queue */ dump_uint8, /* zvol object */ @@ -888,6 +1017,13 @@ static object_viewer_t *object_viewer[DMU_OT_NUMTYPES] = { dump_uint8, /* SPA history */ dump_uint64, /* SPA history offsets */ dump_zap, /* Pool properties */ + dump_zap, /* DSL permissions */ + dump_acl, /* ZFS ACL */ + dump_uint8, /* ZFS SYSACL */ + dump_none, /* FUID nvlist */ + dump_packed_nvlist, /* FUID nvlist size */ + dump_zap, /* DSL dataset next clones */ + dump_zap, /* DSL scrub queue */ }; static void @@ -930,13 +1066,15 @@ dump_object(objset_t *os, uint64_t object, int verbosity, int *print_header) aux[0] = '\0'; - if (doi.doi_checksum != ZIO_CHECKSUM_INHERIT || verbosity >= 6) + if (doi.doi_checksum != ZIO_CHECKSUM_INHERIT || verbosity >= 6) { (void) snprintf(aux + strlen(aux), sizeof (aux), " (K=%s)", - zio_checksum_table[doi.doi_checksum].ci_name); + zio_checksum_table[doi.doi_checksum].ci_name); + } - if (doi.doi_compress != ZIO_COMPRESS_INHERIT || verbosity >= 6) + if (doi.doi_compress != ZIO_COMPRESS_INHERIT || verbosity >= 6) { (void) snprintf(aux + strlen(aux), sizeof (aux), " (Z=%s)", - zio_compress_table[doi.doi_compress].ci_name); + zio_compress_table[doi.doi_compress].ci_name); + } (void) printf("%10lld %3u %5s %5s %5s %5s %s%s\n", (u_longlong_t)object, doi.doi_indirection, iblk, dblk, lsize, @@ -972,13 +1110,13 @@ dump_object(objset_t *os, uint64_t object, int verbosity, int *print_header) } for (;;) { - error = dnode_next_offset(dn, B_FALSE, &start, minlvl, - blkfill, 0); + error = dnode_next_offset(dn, + 0, &start, minlvl, blkfill, 0); if (error) break; end = start; - error = dnode_next_offset(dn, B_TRUE, &end, minlvl, - blkfill, 0); + error = dnode_next_offset(dn, + DNODE_FIND_HOLE, &end, minlvl, blkfill, 0); nicenum(end - start, segsize); (void) printf("\t\tsegment [%016llx, %016llx)" " size %5s\n", (u_longlong_t)start, @@ -996,7 +1134,6 @@ dump_object(objset_t *os, uint64_t object, int verbosity, int *print_header) static char *objset_types[DMU_OST_NUMTYPES] = { "NONE", "META", "ZPL", "ZVOL", "OTHER", "ANY" }; -/*ARGSUSED*/ static void dump_dir(objset_t *os) { @@ -1019,8 +1156,8 @@ dump_dir(objset_t *os) if (dds.dds_type == DMU_OST_META) { dds.dds_creation_txg = TXG_INITIAL; usedobjs = os->os->os_rootbp->blk_fill; - refdbytes = - os->os->os_spa->spa_dsl_pool->dp_mos_dir->dd_used_bytes; + refdbytes = os->os->os_spa->spa_dsl_pool-> + dp_mos_dir->dd_phys->dd_used_bytes; } else { dmu_objset_space(os, &refdbytes, &scratch, &usedobjs, &scratch); } @@ -1054,6 +1191,9 @@ dump_dir(objset_t *os) if (verbosity < 2) return; + if (os->os->os_rootbp->blk_birth == 0) + return; + if (zopt_objects != 0) { for (i = 0; i < zopt_objects; i++) dump_object(os, zopt_object[i], verbosity, @@ -1115,6 +1255,52 @@ dump_config(const char *pool) } static void +dump_cachefile(const char *cachefile) +{ + int fd; + struct stat64 statbuf; + char *buf; + nvlist_t *config; + + if ((fd = open64(cachefile, O_RDONLY)) < 0) { + (void) printf("cannot open '%s': %s\n", cachefile, + strerror(errno)); + exit(1); + } + + if (fstat64(fd, &statbuf) != 0) { + (void) printf("failed to stat '%s': %s\n", cachefile, + strerror(errno)); + exit(1); + } + + if ((buf = malloc(statbuf.st_size)) == NULL) { + (void) fprintf(stderr, "failed to allocate %llu bytes\n", + (u_longlong_t)statbuf.st_size); + exit(1); + } + + if (read(fd, buf, statbuf.st_size) != statbuf.st_size) { + (void) fprintf(stderr, "failed to read %llu bytes\n", + (u_longlong_t)statbuf.st_size); + exit(1); + } + + (void) close(fd); + + if (nvlist_unpack(buf, statbuf.st_size, &config, 0) != 0) { + (void) fprintf(stderr, "failed to unpack nvlist\n"); + exit(1); + } + + free(buf); + + dump_nvlist(config, 0); + + nvlist_free(config); +} + +static void dump_label(const char *dev) { int fd; @@ -1136,15 +1322,7 @@ dump_label(const char *dev) exit(1); } - if (S_ISCHR(statbuf.st_mode)) { - if (ioctl(fd, DIOCGMEDIASIZE, &psize) != 0) { - (void) printf("failed to get size '%s': %s\n", dev, - strerror(errno)); - exit(1); - } - } else - psize = statbuf.st_size; - + psize = statbuf.st_size; psize = P2ALIGN(psize, (uint64_t)sizeof (vdev_label_t)); for (l = 0; l < VDEV_LABELS; l++) { @@ -1178,170 +1356,84 @@ dump_one_dir(char *dsname, void *arg) objset_t *os; error = dmu_objset_open(dsname, DMU_OST_ANY, - DS_MODE_STANDARD | DS_MODE_READONLY, &os); + DS_MODE_USER | DS_MODE_READONLY, &os); if (error) { (void) printf("Could not open %s\n", dsname); return (0); } dump_dir(os); dmu_objset_close(os); + fuid_table_destroy(); return (0); } static void -zdb_space_map_load(spa_t *spa) +zdb_leak(space_map_t *sm, uint64_t start, uint64_t size) { - vdev_t *rvd = spa->spa_root_vdev; - vdev_t *vd; - int c, m, error; + vdev_t *vd = sm->sm_ppd; - for (c = 0; c < rvd->vdev_children; c++) { - vd = rvd->vdev_child[c]; - for (m = 0; m < vd->vdev_ms_count; m++) { - metaslab_t *msp = vd->vdev_ms[m]; - mutex_enter(&msp->ms_lock); - error = space_map_load(&msp->ms_allocmap[0], NULL, - SM_ALLOC, &msp->ms_smo, spa->spa_meta_objset); - mutex_exit(&msp->ms_lock); - if (error) - fatal("%s bad space map #%d, error %d", - spa->spa_name, c, error); - } - } + (void) printf("leaked space: vdev %llu, offset 0x%llx, size %llu\n", + (u_longlong_t)vd->vdev_id, (u_longlong_t)start, (u_longlong_t)size); } -static int -zdb_space_map_claim(spa_t *spa, blkptr_t *bp, zbookmark_t *zb) +/* ARGSUSED */ +static void +zdb_space_map_load(space_map_t *sm) { - dva_t *dva = bp->blk_dva; - vdev_t *vd; - metaslab_t *msp; - space_map_t *allocmap, *freemap; - int error; - int d; - blkptr_t blk = *bp; - - for (d = 0; d < BP_GET_NDVAS(bp); d++) { - uint64_t vdev = DVA_GET_VDEV(&dva[d]); - uint64_t offset = DVA_GET_OFFSET(&dva[d]); - uint64_t size = DVA_GET_ASIZE(&dva[d]); - - if ((vd = vdev_lookup_top(spa, vdev)) == NULL) - return (ENXIO); - - if ((offset >> vd->vdev_ms_shift) >= vd->vdev_ms_count) - return (ENXIO); - - msp = vd->vdev_ms[offset >> vd->vdev_ms_shift]; - allocmap = &msp->ms_allocmap[0]; - freemap = &msp->ms_freemap[0]; - - /* Prepare our copy of the bp in case we need to read GBHs */ - if (DVA_GET_GANG(&dva[d])) { - size = vdev_psize_to_asize(vd, SPA_GANGBLOCKSIZE); - DVA_SET_ASIZE(&blk.blk_dva[d], size); - DVA_SET_GANG(&blk.blk_dva[d], 0); - } - - mutex_enter(&msp->ms_lock); - if (space_map_contains(freemap, offset, size)) { - mutex_exit(&msp->ms_lock); - return (EAGAIN); /* allocated more than once */ - } - - if (!space_map_contains(allocmap, offset, size)) { - mutex_exit(&msp->ms_lock); - return (ESTALE); /* not allocated at all */ - } - - space_map_remove(allocmap, offset, size); - space_map_add(freemap, offset, size); - - mutex_exit(&msp->ms_lock); - } - - if (BP_IS_GANG(bp)) { - zio_gbh_phys_t gbh; - int g; - - /* LINTED - compile time assert */ - ASSERT(sizeof (zio_gbh_phys_t) == SPA_GANGBLOCKSIZE); - - BP_SET_CHECKSUM(&blk, ZIO_CHECKSUM_GANG_HEADER); - BP_SET_PSIZE(&blk, SPA_GANGBLOCKSIZE); - BP_SET_LSIZE(&blk, SPA_GANGBLOCKSIZE); - BP_SET_COMPRESS(&blk, ZIO_COMPRESS_OFF); - error = zio_wait(zio_read(NULL, spa, &blk, &gbh, - SPA_GANGBLOCKSIZE, NULL, NULL, ZIO_PRIORITY_SYNC_READ, - ZIO_FLAG_CANFAIL | ZIO_FLAG_CONFIG_HELD, zb)); - if (error) - return (error); - if (BP_SHOULD_BYTESWAP(&blk)) - byteswap_uint64_array(&gbh, SPA_GANGBLOCKSIZE); - for (g = 0; g < SPA_GBH_NBLKPTRS; g++) { - if (BP_IS_HOLE(&gbh.zg_blkptr[g])) - break; - error = zdb_space_map_claim(spa, &gbh.zg_blkptr[g], zb); - if (error) - return (error); - } - } - - return (0); } static void -zdb_leak(space_map_t *sm, uint64_t start, uint64_t size) +zdb_space_map_unload(space_map_t *sm) { - metaslab_t *msp; - - /* LINTED */ - msp = (metaslab_t *)((char *)sm - offsetof(metaslab_t, ms_allocmap[0])); + space_map_vacate(sm, zdb_leak, sm); +} - (void) printf("leaked space: vdev %llu, offset 0x%llx, size %llu\n", - (u_longlong_t)msp->ms_group->mg_vd->vdev_id, - (u_longlong_t)start, - (u_longlong_t)size); +/* ARGSUSED */ +static void +zdb_space_map_claim(space_map_t *sm, uint64_t start, uint64_t size) +{ } +static space_map_ops_t zdb_space_map_ops = { + zdb_space_map_load, + zdb_space_map_unload, + NULL, /* alloc */ + zdb_space_map_claim, + NULL /* free */ +}; + static void -zdb_space_map_unload(spa_t *spa) +zdb_leak_init(spa_t *spa) { vdev_t *rvd = spa->spa_root_vdev; - vdev_t *vd; - int c, m; - for (c = 0; c < rvd->vdev_children; c++) { - vd = rvd->vdev_child[c]; - for (m = 0; m < vd->vdev_ms_count; m++) { + for (int c = 0; c < rvd->vdev_children; c++) { + vdev_t *vd = rvd->vdev_child[c]; + for (int m = 0; m < vd->vdev_ms_count; m++) { metaslab_t *msp = vd->vdev_ms[m]; mutex_enter(&msp->ms_lock); - space_map_vacate(&msp->ms_allocmap[0], zdb_leak, - &msp->ms_allocmap[0]); - space_map_unload(&msp->ms_allocmap[0]); - space_map_vacate(&msp->ms_freemap[0], NULL, NULL); + VERIFY(space_map_load(&msp->ms_map, &zdb_space_map_ops, + SM_ALLOC, &msp->ms_smo, spa->spa_meta_objset) == 0); + msp->ms_map.sm_ppd = vd; mutex_exit(&msp->ms_lock); } } } static void -zdb_refresh_ubsync(spa_t *spa) +zdb_leak_fini(spa_t *spa) { - uberblock_t ub = { 0 }; vdev_t *rvd = spa->spa_root_vdev; - zio_t *zio; - /* - * Reload the uberblock. - */ - zio = zio_root(spa, NULL, NULL, - ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE); - vdev_uberblock_load(zio, rvd, &ub); - (void) zio_wait(zio); - - if (ub.ub_txg != 0) - spa->spa_ubsync = ub; + for (int c = 0; c < rvd->vdev_children; c++) { + vdev_t *vd = rvd->vdev_child[c]; + for (int m = 0; m < vd->vdev_ms_count; m++) { + metaslab_t *msp = vd->vdev_ms[m]; + mutex_enter(&msp->ms_lock); + space_map_unload(&msp->ms_map); + mutex_exit(&msp->ms_lock); + } + } } /* @@ -1371,9 +1463,7 @@ typedef struct zdb_cb { static void zdb_count_block(spa_t *spa, zdb_cb_t *zcb, blkptr_t *bp, int type) { - int i, error; - - for (i = 0; i < 4; i++) { + for (int i = 0; i < 4; i++) { int l = (i < 2) ? BP_GET_LEVEL(bp) : ZB_TOTAL; int t = (i & 1) ? type : DMU_OT_TOTAL; zdb_blkstats_t *zb = &zcb->zcb_type[l][t]; @@ -1384,21 +1474,34 @@ zdb_count_block(spa_t *spa, zdb_cb_t *zcb, blkptr_t *bp, int type) zb->zb_count++; } - if (dump_opt['L']) - return; - - error = zdb_space_map_claim(spa, bp, &zcb->zcb_cache->bc_bookmark); - - if (error == 0) - return; - - if (error == EAGAIN) - (void) fatal("double-allocation, bp=%p", bp); - - if (error == ESTALE) - (void) fatal("reference to freed block, bp=%p", bp); + if (dump_opt['S']) { + boolean_t print_sig; + + print_sig = !zdb_sig_user_data || (BP_GET_LEVEL(bp) == 0 && + BP_GET_TYPE(bp) == DMU_OT_PLAIN_FILE_CONTENTS); + + if (BP_GET_CHECKSUM(bp) < zdb_sig_cksumalg) + print_sig = B_FALSE; + + if (print_sig) { + (void) printf("%llu\t%lld\t%lld\t%s\t%s\t%s\t" + "%llx:%llx:%llx:%llx\n", + (u_longlong_t)BP_GET_LEVEL(bp), + (longlong_t)BP_GET_PSIZE(bp), + (longlong_t)BP_GET_NDVAS(bp), + dmu_ot[BP_GET_TYPE(bp)].ot_name, + zio_checksum_table[BP_GET_CHECKSUM(bp)].ci_name, + zio_compress_table[BP_GET_COMPRESS(bp)].ci_name, + (u_longlong_t)bp->blk_cksum.zc_word[0], + (u_longlong_t)bp->blk_cksum.zc_word[1], + (u_longlong_t)bp->blk_cksum.zc_word[2], + (u_longlong_t)bp->blk_cksum.zc_word[3]); + } + } - (void) fatal("fatal error %d in bp %p", error, bp); + if (!dump_opt['L']) + VERIFY(zio_wait(zio_claim(NULL, spa, spa_first_txg(spa), bp, + NULL, NULL, ZIO_FLAG_MUSTSUCCEED)) == 0); } static int @@ -1411,9 +1514,16 @@ zdb_blkptr_cb(traverse_blk_cache_t *bc, spa_t *spa, void *arg) char blkbuf[BP_SPRINTF_LEN]; int error = 0; + ASSERT(!BP_IS_HOLE(bp)); + + zdb_count_block(spa, zcb, bp, type); + if (bc->bc_errno) { if (zcb->zcb_readfails++ < 10 && dump_opt['L']) { - zdb_refresh_ubsync(spa); + uberblock_t ub; + vdev_uberblock_load(NULL, spa->spa_root_vdev, &ub); + if (ub.ub_txg != 0) + spa->spa_ubsync = ub; error = EAGAIN; } else { zcb->zcb_haderrors = 1; @@ -1426,35 +1536,32 @@ zdb_blkptr_cb(traverse_blk_cache_t *bc, spa_t *spa, void *arg) else blkbuf[0] = '\0'; - (void) printf("zdb_blkptr_cb: Got error %d reading " - "<%llu, %llu, %lld, %llx> %s -- %s\n", - bc->bc_errno, - (u_longlong_t)zb->zb_objset, - (u_longlong_t)zb->zb_object, - (u_longlong_t)zb->zb_level, - (u_longlong_t)zb->zb_blkid, - blkbuf, - error == EAGAIN ? "retrying" : "skipping"); + if (!dump_opt['S']) { + (void) printf("zdb_blkptr_cb: Got error %d reading " + "<%llu, %llu, %lld, %llx> %s -- %s\n", + bc->bc_errno, + (u_longlong_t)zb->zb_objset, + (u_longlong_t)zb->zb_object, + (u_longlong_t)zb->zb_level, + (u_longlong_t)zb->zb_blkid, + blkbuf, + error == EAGAIN ? "retrying" : "skipping"); + } return (error); } zcb->zcb_readfails = 0; - ASSERT(!BP_IS_HOLE(bp)); - if (dump_opt['b'] >= 4) { sprintf_blkptr(blkbuf, BP_SPRINTF_LEN, bp); (void) printf("objset %llu object %llu offset 0x%llx %s\n", (u_longlong_t)zb->zb_objset, (u_longlong_t)zb->zb_object, (u_longlong_t)blkid2offset(bc->bc_dnode, - zb->zb_level, zb->zb_blkid), - blkbuf); + zb->zb_level, zb->zb_blkid), blkbuf); } - zdb_count_block(spa, zcb, bp, type); - return (0); } @@ -1465,32 +1572,35 @@ dump_block_stats(spa_t *spa) zdb_cb_t zcb = { 0 }; traverse_blk_cache_t dummy_cache = { 0 }; zdb_blkstats_t *zb, *tzb; - uint64_t alloc, space; + uint64_t alloc, space, logalloc; + vdev_t *rvd = spa->spa_root_vdev; int leaks = 0; int advance = zdb_advance; - int flags; - int e; + int c, e, flags; zcb.zcb_cache = &dummy_cache; - if (dump_opt['c']) + if (dump_opt['c'] || dump_opt['S']) advance |= ADVANCE_DATA; advance |= ADVANCE_PRUNE | ADVANCE_ZIL; - (void) printf("\nTraversing all blocks to %sverify" - " nothing leaked ...\n", - dump_opt['c'] ? "verify checksums and " : ""); + if (!dump_opt['S']) { + (void) printf("\nTraversing all blocks to %sverify" + " nothing leaked ...\n", + dump_opt['c'] ? "verify checksums and " : ""); + } /* - * Load all space maps. As we traverse the pool, if we find a block - * that's not in its space map, that indicates a double-allocation, - * reference to a freed block, or an unclaimed block. Otherwise we - * remove the block from the space map. If the space maps are not - * empty when we're done, that indicates leaked blocks. + * Load all space maps as SM_ALLOC maps, then traverse the pool + * claiming each block we discover. If the pool is perfectly + * consistent, the space maps will be empty when we're done. + * Anything left over is a leak; any block we can't claim (because + * it's not part of any space map) is a double allocation, + * reference to a freed block, or an unclaimed log block. */ if (!dump_opt['L']) - zdb_space_map_load(spa); + zdb_leak_init(spa); /* * If there's a deferred-free bplist, process that first. @@ -1533,7 +1643,7 @@ dump_block_stats(spa_t *spa) traverse_fini(th); - if (zcb.zcb_haderrors) { + if (zcb.zcb_haderrors && !dump_opt['S']) { (void) printf("\nError counts:\n\n"); (void) printf("\t%5s %s\n", "errno", "count"); for (e = 0; e < 256; e++) { @@ -1548,7 +1658,15 @@ dump_block_stats(spa_t *spa) * Report any leaked segments. */ if (!dump_opt['L']) - zdb_space_map_unload(spa); + zdb_leak_fini(spa); + + /* + * If we're interested in printing out the blkptr signatures, + * return now as we don't print out anything else (including + * errors and leaks). + */ + if (dump_opt['S']) + return (zcb.zcb_haderrors ? 3 : 0); if (dump_opt['L']) (void) printf("\n\n *** Live pool traversal; " @@ -1557,17 +1675,27 @@ dump_block_stats(spa_t *spa) alloc = spa_get_alloc(spa); space = spa_get_space(spa); + /* + * Log blocks allocated from a separate log device don't count + * as part of the normal pool space; factor them in here. + */ + logalloc = 0; + + for (c = 0; c < rvd->vdev_children; c++) + if (rvd->vdev_child[c]->vdev_islog) + logalloc += rvd->vdev_child[c]->vdev_stat.vs_alloc; + tzb = &zcb.zcb_type[ZB_TOTAL][DMU_OT_TOTAL]; - if (tzb->zb_asize == alloc) { + if (tzb->zb_asize == alloc + logalloc) { (void) printf("\n\tNo leaks (block sum matches space" " maps exactly)\n"); } else { (void) printf("block traversal size %llu != alloc %llu " "(leaked %lld)\n", (u_longlong_t)tzb->zb_asize, - (u_longlong_t)alloc, - (u_longlong_t)(alloc - tzb->zb_asize)); + (u_longlong_t)alloc + logalloc, + (u_longlong_t)(alloc + logalloc - tzb->zb_asize)); leaks = 1; } @@ -1682,11 +1810,11 @@ dump_zpool(spa_t *spa) dump_dtl(spa->spa_root_vdev, 0); dump_metaslabs(spa); } - (void) dmu_objset_find(spa->spa_name, dump_one_dir, NULL, + (void) dmu_objset_find(spa_name(spa), dump_one_dir, NULL, DS_FIND_SNAPSHOTS | DS_FIND_CHILDREN); } - if (dump_opt['b'] || dump_opt['c']) + if (dump_opt['b'] || dump_opt['c'] || dump_opt['S']) rc = dump_block_stats(spa); if (dump_opt['s']) @@ -1898,12 +2026,12 @@ zdb_read_block(char *thing, spa_t **spap) zio_t *zio; vdev_t *vd; void *buf; - char *s, *p, *dup, *spa_name, *vdev, *flagstr; + char *s, *p, *dup, *pool, *vdev, *flagstr; int i, error, zio_flags; dup = strdup(thing); s = strtok(dup, ":"); - spa_name = s ? s : ""; + pool = s ? s : ""; s = strtok(NULL, ":"); vdev = s ? s : ""; s = strtok(NULL, ":"); @@ -1953,14 +2081,13 @@ zdb_read_block(char *thing, spa_t **spap) } } - if (spa == NULL || spa->spa_name == NULL || - strcmp(spa->spa_name, spa_name)) { - if (spa && spa->spa_name) + if (spa == NULL || strcmp(spa_name(spa), pool) != 0) { + if (spa) spa_close(spa, (void *)zdb_read_block); - error = spa_open(spa_name, spap, (void *)zdb_read_block); + error = spa_open(pool, spap, (void *)zdb_read_block); if (error) fatal("Failed to open pool '%s': %s", - spa_name, strerror(error)); + pool, strerror(error)); spa = *spap; } @@ -1980,16 +2107,15 @@ zdb_read_block(char *thing, spa_t **spap) buf = umem_alloc(size, UMEM_NOFAIL); zio_flags = ZIO_FLAG_DONT_CACHE | ZIO_FLAG_DONT_QUEUE | - ZIO_FLAG_DONT_PROPAGATE | ZIO_FLAG_DONT_RETRY | ZIO_FLAG_NOBOOKMARK; - - if (flags & ZDB_FLAG_PHYS) - zio_flags |= ZIO_FLAG_PHYSICAL; + ZIO_FLAG_DONT_PROPAGATE | ZIO_FLAG_DONT_RETRY; + spa_config_enter(spa, SCL_STATE, FTAG, RW_READER); zio = zio_root(spa, NULL, NULL, 0); /* XXX todo - cons up a BP so RAID-Z will be happy */ zio_nowait(zio_vdev_child_io(zio, NULL, vd, offset, buf, size, ZIO_TYPE_READ, ZIO_PRIORITY_SYNC_READ, zio_flags, NULL, NULL)); error = zio_wait(zio); + spa_config_exit(spa, SCL_STATE, FTAG); if (error) { (void) printf("Read of %s failed, error: %d\n", thing, error); @@ -2014,6 +2140,132 @@ out: free(dup); } +static boolean_t +nvlist_string_match(nvlist_t *config, char *name, char *tgt) +{ + char *s; + + if (nvlist_lookup_string(config, name, &s) != 0) + return (B_FALSE); + + return (strcmp(s, tgt) == 0); +} + +static boolean_t +nvlist_uint64_match(nvlist_t *config, char *name, uint64_t tgt) +{ + uint64_t val; + + if (nvlist_lookup_uint64(config, name, &val) != 0) + return (B_FALSE); + + return (val == tgt); +} + +static boolean_t +vdev_child_guid_match(nvlist_t *vdev, uint64_t guid) +{ + nvlist_t **child; + uint_t c, children; + + verify(nvlist_lookup_nvlist_array(vdev, ZPOOL_CONFIG_CHILDREN, + &child, &children) == 0); + for (c = 0; c < children; ++c) + if (nvlist_uint64_match(child[c], ZPOOL_CONFIG_GUID, guid)) + return (B_TRUE); + return (B_FALSE); +} + +static boolean_t +vdev_child_string_match(nvlist_t *vdev, char *tgt) +{ + nvlist_t **child; + uint_t c, children; + + verify(nvlist_lookup_nvlist_array(vdev, ZPOOL_CONFIG_CHILDREN, + &child, &children) == 0); + for (c = 0; c < children; ++c) { + if (nvlist_string_match(child[c], ZPOOL_CONFIG_PATH, tgt) || + nvlist_string_match(child[c], ZPOOL_CONFIG_DEVID, tgt)) + return (B_TRUE); + } + return (B_FALSE); +} + +static boolean_t +vdev_guid_match(nvlist_t *config, uint64_t guid) +{ + nvlist_t *nvroot; + + verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, + &nvroot) == 0); + + return (nvlist_uint64_match(nvroot, ZPOOL_CONFIG_GUID, guid) || + vdev_child_guid_match(nvroot, guid)); +} + +static boolean_t +vdev_string_match(nvlist_t *config, char *tgt) +{ + nvlist_t *nvroot; + + verify(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, + &nvroot) == 0); + + return (vdev_child_string_match(nvroot, tgt)); +} + +static boolean_t +pool_match(nvlist_t *config, char *tgt) +{ + uint64_t guid = strtoull(tgt, NULL, 0); + + if (guid != 0) { + return ( + nvlist_uint64_match(config, ZPOOL_CONFIG_POOL_GUID, guid) || + vdev_guid_match(config, guid)); + } else { + return ( + nvlist_string_match(config, ZPOOL_CONFIG_POOL_NAME, tgt) || + vdev_string_match(config, tgt)); + } +} + +static int +find_exported_zpool(char *pool_id, nvlist_t **configp, char *vdev_dir) +{ + nvlist_t *pools; + int error = ENOENT; + nvlist_t *match = NULL; + + if (vdev_dir != NULL) + pools = zpool_find_import_activeok(g_zfs, 1, &vdev_dir); + else + pools = zpool_find_import_activeok(g_zfs, 0, NULL); + + if (pools != NULL) { + nvpair_t *elem = NULL; + + while ((elem = nvlist_next_nvpair(pools, elem)) != NULL) { + verify(nvpair_value_nvlist(elem, configp) == 0); + if (pool_match(*configp, pool_id)) { + if (match != NULL) { + (void) fatal( + "More than one matching pool - " + "specify guid/devid/device path."); + } else { + match = *configp; + error = 0; + } + } + } + } + + *configp = error ? NULL : match; + + return (error); +} + int main(int argc, char **argv) { @@ -2026,13 +2278,15 @@ main(int argc, char **argv) int verbose = 0; int error; int flag, set; + int exported = 0; + char *vdev_dir = NULL; (void) setrlimit(RLIMIT_NOFILE, &rl); (void) enable_extended_FILE_stdio(-1, -1); dprintf_setup(&argc, argv); - while ((c = getopt(argc, argv, "udibcsvCLO:B:UlR")) != -1) { + while ((c = getopt(argc, argv, "udibcsvCLO:B:S:U:lRep:")) != -1) { switch (c) { case 'u': case 'd': @@ -2093,7 +2347,31 @@ main(int argc, char **argv) verbose++; break; case 'U': - spa_config_dir = "/tmp"; + spa_config_path = optarg; + break; + case 'e': + exported = 1; + break; + case 'p': + vdev_dir = optarg; + break; + case 'S': + dump_opt[c]++; + dump_all = 0; + zdb_sig_user_data = (strncmp(optarg, "user:", 5) == 0); + if (!zdb_sig_user_data && strncmp(optarg, "all:", 4)) + usage(); + endstr = strchr(optarg, ':') + 1; + if (strcmp(endstr, "fletcher2") == 0) + zdb_sig_cksumalg = ZIO_CHECKSUM_FLETCHER_2; + else if (strcmp(endstr, "fletcher4") == 0) + zdb_sig_cksumalg = ZIO_CHECKSUM_FLETCHER_4; + else if (strcmp(endstr, "sha256") == 0) + zdb_sig_cksumalg = ZIO_CHECKSUM_SHA256; + else if (strcmp(endstr, "all") == 0) + zdb_sig_cksumalg = ZIO_CHECKSUM_FLETCHER_2; + else + usage(); break; default: usage(); @@ -2101,7 +2379,12 @@ main(int argc, char **argv) } } + if (vdev_dir != NULL && exported == 0) + (void) fatal("-p option requires use of -e\n"); + kernel_init(FREAD); + g_zfs = libzfs_init(); + ASSERT(g_zfs != NULL); /* * Disable vdev caching. If we don't do this, live pool traversal @@ -2121,7 +2404,7 @@ main(int argc, char **argv) if (argc < 1) { if (dump_opt['C']) { - dump_config(NULL); + dump_cachefile(spa_config_path); return (0); } usage(); @@ -2156,11 +2439,48 @@ main(int argc, char **argv) if (dump_opt['C']) dump_config(argv[0]); - if (strchr(argv[0], '/') != NULL) { - error = dmu_objset_open(argv[0], DMU_OST_ANY, - DS_MODE_STANDARD | DS_MODE_READONLY, &os); - } else { - error = spa_open(argv[0], &spa, FTAG); + error = 0; + if (exported) { + /* + * Check to see if the name refers to an exported zpool + */ + char *slash; + nvlist_t *exported_conf = NULL; + + if ((slash = strchr(argv[0], '/')) != NULL) + *slash = '\0'; + + error = find_exported_zpool(argv[0], &exported_conf, vdev_dir); + if (error == 0) { + nvlist_t *nvl = NULL; + + if (vdev_dir != NULL) { + if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0) + error = ENOMEM; + else if (nvlist_add_string(nvl, + zpool_prop_to_name(ZPOOL_PROP_ALTROOT), + vdev_dir) != 0) + error = ENOMEM; + } + + if (error == 0) + error = spa_import_faulted(argv[0], + exported_conf, nvl); + + nvlist_free(nvl); + } + + if (slash != NULL) + *slash = '/'; + } + + if (error == 0) { + if (strchr(argv[0], '/') != NULL) { + error = dmu_objset_open(argv[0], DMU_OST_ANY, + DS_MODE_USER | DS_MODE_READONLY, &os); + } else { + error = spa_open(argv[0], &spa, FTAG); + } } if (error) @@ -2187,6 +2507,9 @@ main(int argc, char **argv) spa_close(spa, FTAG); } + fuid_table_destroy(); + + libzfs_fini(g_zfs); kernel_fini(); return (0); diff --git a/cddl/contrib/opensolaris/cmd/zdb/zdb_il.c b/cddl/contrib/opensolaris/cmd/zdb/zdb_il.c index 10dfe20..02d35a0 100644 --- a/cddl/contrib/opensolaris/cmd/zdb/zdb_il.c +++ b/cddl/contrib/opensolaris/cmd/zdb/zdb_il.c @@ -19,7 +19,7 @@ * CDDL HEADER END */ /* - * Copyright 2006 Sun Microsystems, Inc. All rights reserved. + * Copyright 2007 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ @@ -233,19 +233,26 @@ typedef struct zil_rec_info { } zil_rec_info_t; static zil_rec_info_t zil_rec_info[TX_MAX_TYPE] = { - { NULL, "Total " }, - { zil_prt_rec_create, "TX_CREATE " }, - { zil_prt_rec_create, "TX_MKDIR " }, - { zil_prt_rec_create, "TX_MKXATTR " }, - { zil_prt_rec_create, "TX_SYMLINK " }, - { zil_prt_rec_remove, "TX_REMOVE " }, - { zil_prt_rec_remove, "TX_RMDIR " }, - { zil_prt_rec_link, "TX_LINK " }, - { zil_prt_rec_rename, "TX_RENAME " }, - { zil_prt_rec_write, "TX_WRITE " }, - { zil_prt_rec_truncate, "TX_TRUNCATE" }, - { zil_prt_rec_setattr, "TX_SETATTR " }, - { zil_prt_rec_acl, "TX_ACL " }, + { NULL, "Total " }, + { zil_prt_rec_create, "TX_CREATE " }, + { zil_prt_rec_create, "TX_MKDIR " }, + { zil_prt_rec_create, "TX_MKXATTR " }, + { zil_prt_rec_create, "TX_SYMLINK " }, + { zil_prt_rec_remove, "TX_REMOVE " }, + { zil_prt_rec_remove, "TX_RMDIR " }, + { zil_prt_rec_link, "TX_LINK " }, + { zil_prt_rec_rename, "TX_RENAME " }, + { zil_prt_rec_write, "TX_WRITE " }, + { zil_prt_rec_truncate, "TX_TRUNCATE " }, + { zil_prt_rec_setattr, "TX_SETATTR " }, + { zil_prt_rec_acl, "TX_ACL_V0 " }, + { zil_prt_rec_acl, "TX_ACL_ACL " }, + { zil_prt_rec_create, "TX_CREATE_ACL " }, + { zil_prt_rec_create, "TX_CREATE_ATTR " }, + { zil_prt_rec_create, "TX_CREATE_ACL_ATTR " }, + { zil_prt_rec_create, "TX_MKDIR_ACL " }, + { zil_prt_rec_create, "TX_MKDIR_ATTR " }, + { zil_prt_rec_create, "TX_MKDIR_ACL_ATTR " }, }; /* ARGSUSED */ @@ -255,12 +262,14 @@ print_log_record(zilog_t *zilog, lr_t *lr, void *arg, uint64_t claim_txg) int txtype; int verbose = MAX(dump_opt['d'], dump_opt['i']); + /* reduce size of txtype to strip off TX_CI bit */ txtype = lr->lrc_txtype; ASSERT(txtype != 0 && (uint_t)txtype < TX_MAX_TYPE); ASSERT(lr->lrc_txg); - (void) printf("\t\t%s len %6llu, txg %llu, seq %llu\n", + (void) printf("\t\t%s%s len %6llu, txg %llu, seq %llu\n", + (lr->lrc_txtype & TX_CI) ? "CI-" : "", zil_rec_info[txtype].zri_name, (u_longlong_t)lr->lrc_reclen, (u_longlong_t)lr->lrc_txg, diff --git a/cddl/contrib/opensolaris/cmd/zfs/zfs.8 b/cddl/contrib/opensolaris/cmd/zfs/zfs.8 index d49cb87..7217fe8 100644 --- a/cddl/contrib/opensolaris/cmd/zfs/zfs.8 +++ b/cddl/contrib/opensolaris/cmd/zfs/zfs.8 @@ -18,7 +18,7 @@ .\" .\" CDDL HEADER END .\" Copyright (c) 2007 Sun Microsystems, Inc. All Rights Reserved. -.TH zfs 1M "16 Mar 2007" "SunOS 5.11" "System Administration Commands" +.TH zfs 1M "8 Apr 2008" "SunOS 5.11" "System Administration Commands" .SH NAME zfs \- configures ZFS file systems .SH SYNOPSIS @@ -29,12 +29,12 @@ zfs \- configures ZFS file systems .LP .nf -\fBzfs\fR \fBcreate\fR [[\fB-o\fR property=\fIvalue\fR]]... \fIfilesystem\fR +\fBzfs\fR \fBcreate\fR [\fB-p\fR] [\fB-o\fR \fIproperty\fR=\fIvalue\fR] ... \fIfilesystem\fR .fi .LP .nf -\fBzfs\fR \fBcreate\fR [\fB-s\fR] [\fB-b\fR \fIblocksize\fR] [[\fB-o\fR property=\fIvalue\fR]]... \fB-V\fR \fIsize\fR \fIvolume\fR +\fBzfs\fR \fBcreate\fR [\fB-ps\fR] [\fB-b\fR \fIblocksize\fR] [\fB-o\fR \fIproperty\fR=\fIvalue\fR] ... \fB-V\fR \fIsize\fR \fIvolume\fR .fi .LP @@ -44,35 +44,44 @@ zfs \- configures ZFS file systems .LP .nf -\fBzfs\fR \fBclone\fR \fIsnapshot\fR \fIfilesystem\fR|\fIvolume\fR +\fBzfs\fR \fBsnapshot\fR [\fB-r\fR] \fIfilesystem@snapname\fR|\fIvolume@snapname\fR .fi .LP .nf -\fBzfs\fR \fBpromote\fR \fIfilesystem\fR +\fBzfs\fR \fBrollback\fR [\fB-rRf\fR] \fIsnapshot\fR +.fi + +.LP +.nf +\fBzfs\fR \fBclone\fR [\fB-p\fR] \fIsnapshot\fR \fIfilesystem\fR|\fIvolume\fR +.fi + +.LP +.nf +\fBzfs\fR \fBpromote\fR \fIclone-filesystem\fR .fi .LP .nf \fBzfs\fR \fBrename\fR \fIfilesystem\fR|\fIvolume\fR|\fIsnapshot\fR - [\fIfilesystem\fR|\fIvolume\fR|\fIsnapshot\fR] + \fIfilesystem\fR|\fIvolume\fR|\fIsnapshot\fR .fi .LP .nf -\fBzfs\fR \fBsnapshot\fR [\fB-r\fR] \fIfilesystem@name\fR|\fIvolume@name\fR +\fBzfs\fR \fBrename\fR [\fB-p\fR] \fIfilesystem\fR|\fIvolume\fR \fIfilesystem\fR|\fIvolume\fR .fi .LP .nf -\fBzfs\fR \fBrollback\fR [\fB-rRf\fR] \fIsnapshot\fR +\fBzfs\fR \fBrename\fR \fB-r\fR \fIsnapshot\fR \fIsnapshot\fR .fi .LP .nf -\fBzfs\fR \fBlist\fR [\fB-rH\fR] [\fB-o\fR \fIprop\fR[,\fIprop\fR] ]... [ \fB-t\fR \fItype\fR[,\fItype\fR]...] - [ \fB-s\fR \fIprop\fR [\fB-s\fR \fIprop\fR]... [ \fB-S\fR \fIprop\fR [\fB-S\fR \fIprop\fR]... - [\fIfilesystem\fR|\fIvolume\fR|\fIsnapshot\fR|\fI/pathname\fR|.\fI/pathname\fR ... +\fBzfs\fR \fBlist\fR [\fB-rH\fR] [\fB-o\fR \fIproperty\fR[,...]] [\fB-t\fR \fItype\fR[,...]] + [\fB-s\fR \fIproperty\fR] ... [\fB-S\fR \fIproperty\fR ... [\fIfilesystem\fR|\fIvolume\fR|\fIsnapshot\fR] ... .fi .LP @@ -82,14 +91,23 @@ zfs \- configures ZFS file systems .LP .nf -\fBzfs\fR \fBget\fR [\fB-rHp\fR] [\fB-o\fR \fIfield\fR[,\fIfield\fR]...] - [\fB-s\fR \fIsource\fR[,\fIsource\fR]...] \fIall\fR | \fIproperty\fR[,\fIproperty\fR]... +\fBzfs\fR \fBget\fR [\fB-rHp\fR] [\fB-o\fR \fIfield\fR[,...]] [\fB-s\fR \fIsource\fR[,...]] "\fIall\fR" | \fIproperty\fR[,...] \fIfilesystem\fR|\fIvolume\fR|\fIsnapshot\fR ... .fi .LP .nf -\fBzfs\fR \fBinherit\fR [\fB-r\fR] \fIproperty\fR \fIfilesystem\fR|\fIvolume\fR... ... +\fBzfs\fR \fBinherit\fR [\fB-r\fR] \fIproperty\fR \fIfilesystem\fR|\fIvolume\fR ... +.fi + +.LP +.nf +\fBzfs\fR \fBupgrade\fR [\fB-v\fR] +.fi + +.LP +.nf +\fBzfs\fR \fBupgrade\fR [\fB-r\fR] [\fB-V\fR \fIversion\fR] \fB-a\fR | \fIfilesystem\fR .fi .LP @@ -99,58 +117,81 @@ zfs \- configures ZFS file systems .LP .nf -\fBzfs\fR \fBmount\fR [\fB-o \fIoptions\fR\fR] [\fB-O\fR] \fB-a\fR +\fBzfs\fR \fBmount\fR [\fB-vO\fR] [\fB-o \fIoptions\fR\fR] \fB-a\fR | \fIfilesystem\fR .fi .LP .nf -\fBzfs\fR \fBmount\fR [\fB-o \fIoptions\fR\fR] [\fB-O\fR] \fIfilesystem\fR +\fBzfs\fR \fBunmount\fR [\fB-f\fR] \fB-a\fR | \fIfilesystem\fR|\fImountpoint\fR .fi .LP .nf -\fBzfs\fR \fBunmount\fR [\fB-f\fR] \fB-a\fR +\fBzfs\fR \fBshare\fR \fB-a\fR | \fIfilesystem\fR .fi .LP .nf -\fBzfs\fR \fBunmount\fR [\fB-f\fR] \fB\fIfilesystem\fR|\fImountpoint\fR\fR +\fBzfs\fR \fBunshare\fR \fB-a\fR \fIfilesystem\fR|\fImountpoint\fR .fi .LP .nf -\fBzfs\fR \fBshare\fR \fB-a\fR +\fBzfs\fR \fBsend\fR [\fB-vR\fR] [\fB-\fR[\fB-iI\fR] \fIsnapshot\fR] \fIsnapshot\fR .fi .LP .nf -\fBzfs\fR \fBshare\fR \fIfilesystem\fR +\fBzfs\fR \fBreceive\fR [\fB-vnF\fR] \fIfilesystem\fR|\fIvolume\fR|\fIsnapshot\fR .fi .LP .nf -\fBzfs\fR \fBunshare\fR [\fB-f\fR] \fB-a\fR +\fBzfs\fR \fBreceive\fR [\fB-vnF\fR] \fB-d\fR \fIfilesystem\fR .fi .LP .nf -\fBzfs\fR \fBunshare\fR [\fB-f\fR] \fB\fIfilesystem\fR|\fImountpoint\fR\fR +\fBzfs\fR \fBallow\fR [\fB-ldug\fR] "\fIeveryone\fR"|\fIuser\fR|\fIgroup\fR[,...] \fIperm\fR|\fI@setname\fR[,...] + \fIfilesystem\fR|\fIvolume\fR .fi .LP .nf -\fBzfs\fR \fBsend\fR [\fB-i\fR \fIsnapshot1\fR] \fB\fIsnapshot2\fR\fR +\fBzfs\fR \fBallow\fR [\fB-ld\fR] \fB-e\fR \fIperm\fR|@\fIsetname\fR[,...] \fIfilesystem\fR|\fIvolume\fR .fi .LP .nf -\fBzfs\fR \fBreceive\fR [\fB-vnF\fR ] \fIfilesystem\fR|\fIvolume\fR|\fIsnapshot\fR +\fBzfs\fR \fBallow\fR \fB-c\fR \fIperm\fR|@\fIsetname\fR[,...] \fIfilesystem\fR|\fIvolume\fR .fi .LP .nf -\fBzfs\fR \fBreceive\fR [\fB-vnF\fR ] \fB-d\fR \fB\fIfilesystem\fR\fR +\fBzfs\fR \fBallow\fR \fB-s\fR @setname \fIperm\fR|@\fIsetname\fR[,...] \fIfilesystem\fR|\fIvolume\fR .fi + +.LP +.nf +\fBzfs\fR \fBunallow\fR [\fB-rldug\fR] "\fIeveryone\fR"|\fIuser\fR|\fIgroup\fR[,...] [\fIperm\fR|@\fIsetname\fR[,... ]] + \fIfilesystem\fR|\fIvolume\fR +.fi + +.LP +.nf +\fBzfs\fR \fBunallow\fR [\fB-rld\fR] \fB-e\fR [\fIperm\fR|@\fIsetname\fR[,... ]] \fIfilesystem\fR|\fIvolume\fR +.fi + +.LP +.nf +\fBzfs\fR \fBunallow\fR [\fB-r\fR] \fB-c\fR [\fIperm\fR|@\fIsetname\fR[ ... ]] \fIfilesystem\fR|\fIvolume\fR +.fi + +.LP +.nf +\fBzfs\fR \fBunallow\fR [\fB-r\fR] \fB-s\fR @setname [\fIperm\fR|@\fIsetname\fR[,... ]] \fIfilesystem\fR|\fIvolume\fR +.fi + .LP .nf \fBzfs\fR \fBjail\fR \fBjailid\fR \fB\fIfilesystem\fR\fR @@ -161,6 +202,7 @@ zfs \- configures ZFS file systems .fi .SH DESCRIPTION +.sp .LP The \fBzfs\fR command configures \fBZFS\fR datasets within a \fBZFS\fR storage pool, as described in \fBzpool\fR(1M). A dataset is identified by a unique path within the \fBZFS\fR namespace. For example: @@ -172,8 +214,10 @@ pool/{filesystem,volume,snapshot} .in -2 .sp +.sp .LP where the maximum length of a dataset name is \fBMAXNAMELEN\fR (256 bytes). +.sp .LP A dataset can be one of the following: .sp @@ -210,65 +254,88 @@ A read-only version of a file system or volume at a given point in time. It is s .RE .SS "ZFS File System Hierarchy" +.sp .LP A \fBZFS\fR storage pool is a logical collection of devices that provide space for datasets. A storage pool is also the root of the \fBZFS\fR file system hierarchy. +.sp .LP The root of the pool can be accessed as a file system, such as mounting and unmounting, taking snapshots, and setting properties. The physical storage characteristics, however, are managed by the \fBzpool\fR(1M) command. +.sp .LP See \fBzpool\fR(1M) for more information on creating and administering pools. .SS "Snapshots" +.sp .LP A snapshot is a read-only copy of a file system or volume. Snapshots can be created extremely quickly, and initially consume no additional space within the pool. As data within the active dataset changes, the snapshot consumes more data than would otherwise be shared with the active dataset. +.sp .LP Snapshots can have arbitrary names. Snapshots of volumes can be cloned or rolled back, but cannot be accessed independently. +.sp .LP File system snapshots can be accessed under the ".zfs/snapshot" directory in the root of the file system. Snapshots are automatically mounted on demand and may be unmounted at regular intervals. The visibility of the ".zfs" directory can be controlled by the "snapdir" property. .SS "Clones" +.sp .LP A clone is a writable volume or file system whose initial contents are the same as another dataset. As with snapshots, creating a clone is nearly instantaneous, and initially consumes no additional space. +.sp .LP Clones can only be created from a snapshot. When a snapshot is cloned, it creates an implicit dependency between the parent and child. Even though the clone is created somewhere else in the dataset hierarchy, the original snapshot cannot be destroyed as long as a clone exists. The "origin" property exposes this dependency, and the \fBdestroy\fR command lists any such dependencies, if they exist. +.sp .LP The clone parent-child dependency relationship can be reversed by using the "\fBpromote\fR" subcommand. This causes the "origin" file system to become a clone of the specified file system, which makes it possible to destroy the file system that the clone was created from. .SS "Mount Points" +.sp .LP -Creating a \fBZFS\fR file system is a simple operation, so the number of file systems per system will likely be numerous. To cope with this, \fBZFS\fR automatically manages mounting and unmounting file systems without the need to edit the \fB/etc/vfstab\fR file. -All automatically managed file systems are mounted by \fBZFS\fR at boot time. +Creating a \fBZFS\fR file system is a simple operation, so the number of file systems per system is likely to be numerous. To cope with this, \fBZFS\fR automatically manages mounting and unmounting file systems without the need to edit the \fB/etc/vfstab\fR file. All automatically managed file systems are mounted by \fBZFS\fR at boot time. +.sp .LP By default, file systems are mounted under \fB/\fIpath\fR\fR, where \fIpath\fR is the name of the file system in the \fBZFS\fR namespace. Directories are created and destroyed as needed. +.sp .LP A file system can also have a mount point set in the "mountpoint" property. This directory is created as needed, and \fBZFS\fR automatically mounts the file system when the "\fBzfs mount -a\fR" command is invoked (without editing \fB/etc/vfstab\fR). The mountpoint property can be inherited, so if \fBpool/home\fR has a mount point of \fB/export/stuff\fR, then \fBpool/home/user\fR automatically inherits a mount point of \fB/export/stuff/user\fR. +.sp .LP A file system mountpoint property of "none" prevents the file system from being mounted. +.sp .LP If needed, \fBZFS\fR file systems can also be managed with traditional tools (\fBmount\fR, \fBumount\fR, \fB/etc/vfstab\fR). If a file system's mount point is set to "legacy", \fBZFS\fR makes no attempt to manage the file system, and the administrator is responsible for mounting and unmounting the file system. .SS "Zones" +.sp .LP A \fBZFS\fR file system can be added to a non-global zone by using zonecfg's "\fBadd fs\fR" subcommand. A \fBZFS\fR file system that is added to a non-global zone must have its mountpoint property set to legacy. +.sp .LP The physical properties of an added file system are controlled by the global administrator. However, the zone administrator can create, modify, or destroy files within the added file system, depending on how the file system is mounted. +.sp .LP A dataset can also be delegated to a non-global zone by using zonecfg's "\fBadd dataset\fR" subcommand. You cannot delegate a dataset to one zone and the children of the same dataset to another zone. The zone administrator can change properties of the dataset or any of its children. However, the "quota" property is controlled by the global administrator. +.sp .LP A \fBZFS\fR volume can be added as a device to a non-global zone by using zonecfg's "\fBadd device\fR" subcommand. However, its physical properties can only be modified by the global administrator. +.sp .LP For more information about \fBzonecfg\fR syntax, see \fBzonecfg\fR(1M). +.sp .LP After a dataset is delegated to a non-global zone, the "zoned" property is automatically set. A zoned file system cannot be mounted in the global zone, since the zone administrator might have to set the mount point to an unacceptable value. +.sp .LP The global administrator can forcibly clear the "zoned" property, though this should be done with extreme care. The global administrator should verify that all the mount points are acceptable before clearing the property. .SS "Native Properties" +.sp .LP Properties are divided into two types, native properties and user defined properties. Native properties either export internal statistics or control \fBZFS\fR behavior. In addition, native properties are either editable or read-only. User properties have no effect on \fBZFS\fR behavior, but you can use them to annotate datasets in a way that is meaningful in your environment. For more information about user properties, see the "User Properties" section. +.sp .LP Every dataset has a set of properties that export statistics about the dataset as well as control various behavior. Properties are inherited from the parent unless overridden by the child. Snapshot properties can not be edited; they always inherit their inheritable properties. Properties that are not applicable to snapshots are not displayed. +.sp .LP The values of numeric properties can be specified using the following human-readable suffixes (for example, "k", "KB", "M", "Gb", etc, up to Z for zettabyte). The following are all valid (and equal) specifications: .sp @@ -279,19 +346,35 @@ The values of numeric properties can be specified using the following human-read .in -2 .sp +.sp .LP -The values of non-numeric properties are case sensitive and must be lowercase, except for "mountpoint" and "sharenfs". +The values of non-numeric properties are case sensitive and must be lowercase, except for "mountpoint", "sharenfs" and "sharesmb". +.sp .LP -The first set of properties consist of read-only statistics about the dataset. These properties cannot be set, nor are they inherited. Native properties apply to all dataset types unless otherwise noted. +The following native properties consist of read-only statistics about the dataset. These properties cannot be set, nor are they inherited. Native properties apply to all dataset types unless otherwise noted. .sp .ne 2 .mk .na -\fBtype\fR +\fBavailable\fR .ad -.RS 17n -.rt -The type of dataset: "filesystem", "volume", "snapshot", or "clone". +.sp .6 +.RS 4n +The amount of space available to the dataset and all its children, assuming that there is no other activity in the pool. Because space is shared within a pool, availability can be limited by any number of factors, including physical pool size, quotas, reservations, or other datasets +within the pool. +.sp +This property can also be referred to by its shortened column name, "avail". +.RE + +.sp +.ne 2 +.mk +.na +\fBcompressratio\fR +.ad +.sp .6 +.RS 4n +The compression ratio achieved for this dataset, expressed as a multiplier. Compression can be turned on by running "zfs set compression=on \fIdataset\fR". The default value is "off". .RE .sp @@ -300,8 +383,8 @@ The type of dataset: "filesystem", "volume", "snapshot", or "clone". .na \fBcreation\fR .ad -.RS 17n -.rt +.sp .6 +.RS 4n The time this dataset was created. .RE @@ -309,31 +392,22 @@ The time this dataset was created. .ne 2 .mk .na -\fBused\fR +\fBmounted\fR .ad -.RS 17n -.rt -The amount of space consumed by this dataset and all its descendants. This is the value that is checked against this dataset's quota and reservation. The space used does not include this dataset's reservation, but does take into account the reservations of any descendant datasets. -The amount of space that a dataset consumes from its parent, as well as the amount of space that will be freed if this dataset is recursively destroyed, is the greater of its space used and its reservation. -.sp -When snapshots (see the "Snapshots" section) are created, their space is initially shared between the snapshot and the file system, and possibly with previous snapshots. As the file system changes, space that was previously shared becomes unique to the snapshot, and counted in -the snapshot's space used. Additionally, deleting snapshots can increase the amount of space unique to (and used by) other snapshots. -.sp -The amount of space used, available, or referenced does not take into account pending changes. Pending changes are generally accounted for within a few seconds. Committing a change to a disk using \fBfsync\fR(3c) or \fBO_SYNC\fR does not necessarily guarantee that the space usage information is updated immediately. +.sp .6 +.RS 4n +For file systems, indicates whether the file system is currently mounted. This property can be either "yes" or "no". .RE .sp .ne 2 .mk .na -\fBavailable\fR +\fBorigin\fR .ad -.RS 17n -.rt -The amount of space available to the dataset and all its children, assuming that there is no other activity in the pool. Because space is shared within a pool, availability can be limited by any number of factors, including physical pool size, quotas, reservations, or other datasets -within the pool. -.sp -This property can also be referred to by its shortened column name, "avail". +.sp .6 +.RS 4n +For cloned file systems or volumes, the snapshot from which the clone was created. The origin cannot be destroyed (even with the \fB-r\fR or \fB-f\fR options) so long as a clone exists. .RE .sp @@ -342,8 +416,8 @@ This property can also be referred to by its shortened column name, "avail". .na \fBreferenced\fR .ad -.RS 17n -.rt +.sp .6 +.RS 4n The amount of data that is accessible by this dataset, which may or may not be shared with other datasets in the pool. When a snapshot or clone is created, it initially references the same amount of space as the file system or snapshot it was created from, since its contents are identical. .sp @@ -354,116 +428,166 @@ This property can also be referred to by its shortened column name, "refer". .ne 2 .mk .na -\fBcompressratio\fR +\fBtype\fR .ad -.RS 17n -.rt -The compression ratio achieved for this dataset, expressed as a multiplier. Compression can be turned on by running "zfs set compression=on \fIdataset\fR". The default value is "off". +.sp .6 +.RS 4n +The type of dataset: "filesystem", "volume", "snapshot", or "clone". .RE .sp .ne 2 .mk .na -\fBmounted\fR +\fBused\fR .ad -.RS 17n -.rt -For file systems, indicates whether the file system is currently mounted. This property can be either "yes" or "no". +.sp .6 +.RS 4n +The amount of space consumed by this dataset and all its descendents. This is the value that is checked against this dataset's quota and reservation. The space used does not include this dataset's reservation, but does take into account the reservations of any descendent datasets. +The amount of space that a dataset consumes from its parent, as well as the amount of space that are freed if this dataset is recursively destroyed, is the greater of its space used and its reservation. +.sp +When snapshots (see the "Snapshots" section) are created, their space is initially shared between the snapshot and the file system, and possibly with previous snapshots. As the file system changes, space that was previously shared becomes unique to the snapshot, and counted in +the snapshot's space used. Additionally, deleting snapshots can increase the amount of space unique to (and used by) other snapshots. +.sp +The amount of space used, available, or referenced does not take into account pending changes. Pending changes are generally accounted for within a few seconds. Committing a change to a disk using \fBfsync\fR(3c) or \fBO_SYNC\fR does not necessarily guarantee that the space usage information is updated immediately. .RE .sp .ne 2 .mk .na -\fBorigin\fR +\fBvolblocksize=\fIblocksize\fR\fR .ad -.RS 17n -.rt -For cloned file systems or volumes, the snapshot from which the clone was created. The origin cannot be destroyed (even with the \fB-r\fR or \fB-f\fR options) so long as a clone exists. +.sp .6 +.RS 4n +For volumes, specifies the block size of the volume. The \fBblocksize\fR cannot be changed once the volume has been written, so it should be set at volume creation time. The default \fBblocksize\fR for volumes is 8 Kbytes. Any power of 2 from 512 bytes +to 128 Kbytes is valid. +.sp +This property can also be referred to by its shortened column name, "volblock". .RE +.sp .LP -The following two properties can be set to control the way space is allocated between datasets. These properties are not inherited, but do affect their descendants. +The following native properties can be used to change the behavior of a \fBZFS\fR dataset. .sp .ne 2 .mk .na -\fBquota=\fIsize\fR | \fInone\fR\fR +\fBaclinherit=\fBdiscard\fR | \fBnoallow\fR | \fBrestricted\fR | \fBpassthrough\fR\fR .ad .sp .6 .RS 4n -Limits the amount of space a dataset and its descendants can consume. This property enforces a hard limit on the amount of space used. This includes all space consumed by descendants, including file systems and snapshots. Setting a quota on a descendant of a dataset that already -has a quota does not override the ancestor's quota, but rather imposes an additional limit. +Controls how \fBACL\fR entries are inherited when files and directories are created. A file system with an "aclinherit" property of "\fBdiscard\fR" does not inherit any \fBACL\fR entries. A file system with an "aclinherit" +property value of "\fBnoallow\fR" only inherits inheritable \fBACL\fR entries that specify "deny" permissions. The property value "\fBrestricted\fR" (the default) removes the "\fBwrite_acl\fR" and "\fBwrite_owner\fR" permissions when the \fBACL\fR entry is inherited. A file system with an "aclinherit" property value of "\fBpassthrough\fR" inherits all inheritable \fBACL\fR entries without any modifications made to the \fBACL\fR entries when they are inherited. .sp -Quotas cannot be set on volumes, as the "volsize" property acts as an implicit quota. +When the property value is set to "\fBpassthrough\fR," files are created with a mode determined by the inheritable \fBACE\fRs. If no inheritable \fBACE\fRs exist that affect the mode, then the mode is set in accordance to the requested mode +from the application. .RE .sp .ne 2 .mk .na -\fBreservation=\fIsize\fR | \fInone\fR\fR +\fBaclmode=\fBdiscard\fR | \fBgroupmask\fR | \fBpassthrough\fR\fR .ad .sp .6 .RS 4n -The minimum amount of space guaranteed to a dataset and its descendants. When the amount of space used is below this value, the dataset is treated as if it were taking up the amount of space specified by its reservation. Reservations are accounted for in the parent datasets' space -used, and count against the parent datasets' quotas and reservations. +Controls how an \fBACL\fR is modified during \fBchmod\fR(2). A file system with an "aclmode" property of "\fBdiscard\fR" +deletes all \fBACL\fR entries that do not represent the mode of the file. An "aclmode" property of "\fBgroupmask\fR" (the default) reduces user or group permissions. The permissions are reduced, such that they are no greater than the group permission +bits, unless it is a user entry that has the same \fBUID\fR as the owner of the file or directory. In this case, the \fBACL\fR permissions are reduced so that they are no greater than owner permission bits. A file system with an "aclmode" property of "\fBpassthrough\fR" indicates that no changes are made to the \fBACL\fR other than generating the necessary \fBACL\fR entries to represent the new mode of the file or directory. +.RE + .sp -This property can also be referred to by its shortened column name, "reserv". +.ne 2 +.mk +.na +\fBatime=\fIon\fR | \fIoff\fR\fR +.ad +.sp .6 +.RS 4n +Controls whether the access time for files is updated when they are read. Turning this property off avoids producing write traffic when reading files and can result in significant performance gains, though it might confuse mailers and other similar utilities. The default value +is "on". .RE .sp .ne 2 .mk .na -\fBvolsize=\fIsize\fR\fR +\fBcanmount=\fBon\fR | \fBoff\fR | \fBnoauto\fR\fR .ad .sp .6 .RS 4n -For volumes, specifies the logical size of the volume. By default, creating a volume establishes a reservation of equal size. Any changes to \fBvolsize\fR are reflected in an equivalent change to the reservation. The \fBvolsize\fR can only be set to a -multiple of \fBvolblocksize\fR, and cannot be zero. +If this property is set to "\fBoff\fR", the file system cannot be mounted, and is ignored by "\fBzfs mount -a\fR". Setting this property to "\fBoff\fR" is similar to setting the "mountpoint" +property to "\fBnone\fR", except that the dataset still has a normal "mountpoint" property, which can be inherited. Setting this property to "\fBoff\fR" allows datasets to be used solely as a mechanism to inherit properties. One example +of setting canmount=\fBoff\fR is to have two datasets with the same mountpoint, so that the children of both datasets appear in the same directory, but might have different inherited characteristics. .sp -The reservation is kept equal to the volume's logical size to prevent unexpected behavior for consumers. Without the reservation, the volume could run out of space, resulting in undefined behavior or data corruption, depending on how the volume is used. These effects can also occur when -the volume size is changed while it is in use (particularly when shrinking the size). Extreme care should be used when adjusting the volume size. +When the "\fBnoauto\fR" option is set, a dataset can only be mounted and unmounted explicitly. The dataset is not mounted automatically when the dataset is created or imported, nor is it mounted by the "\fBzfs mount -a\fR" command or unmounted +by the "\fBzfs unmount -a\fR" command. .sp -Though not recommended, a "sparse volume" (also known as "thin provisioning") can be created by specifying the \fB-s\fR option to the "\fBzfs create -V\fR" command, or by changing the reservation after the volume has been created. -A "sparse volume" is a volume where the reservation is less then the volume size. Consequently, writes to a sparse volume can fail with \fBENOSPC\fR when the pool is low on space. For a sparse volume, changes to \fBvolsize\fR are not reflected in the reservation. +This property is not inherited. .RE .sp .ne 2 .mk .na -\fBvolblocksize=\fIblocksize\fR\fR +\fBchecksum=\fIon\fR | \fIoff\fR | \fIfletcher2\fR, | \fIfletcher4\fR | \fIsha256\fR\fR .ad .sp .6 .RS 4n -For volumes, specifies the block size of the volume. The \fBblocksize\fR cannot be changed once the volume has been written, so it should be set at volume creation time. The default \fBblocksize\fR for volumes is 8 Kbytes. Any power of 2 from 512 bytes -to 128 Kbytes is valid. -.sp -This property can also be referred to by its shortened column name, "volblock". +Controls the checksum used to verify data integrity. The default value is "on", which automatically selects an appropriate algorithm (currently, \fIfletcher2\fR, but this may change in future releases). The value "off" disables integrity +checking on user data. Disabling checksums is NOT a recommended practice. .RE .sp .ne 2 .mk .na -\fBrecordsize=\fIsize\fR\fR +\fBcompression=\fIon\fR | \fIoff\fR | \fIlzjb\fR | \fIgzip\fR | \fIgzip-N\fR\fR .ad .sp .6 .RS 4n -Specifies a suggested block size for files in the file system. This property is designed solely for use with database workloads that access files in fixed-size records. \fBZFS\fR automatically tunes block sizes according to internal algorithms optimized for typical -access patterns. +Controls the compression algorithm used for this dataset. The "lzjb" compression algorithm is optimized for performance while providing decent data compression. Setting compression to "on" uses the "lzjb" compression algorithm. The "gzip" +compression algorithm uses the same compression as the \fBgzip\fR(1) command. You can specify the "gzip" level by using the value "gzip-\fIN\fR" where \fIN\fR is +an integer from 1 (fastest) to 9 (best compression ratio). Currently, "gzip" is equivalent to "gzip-6" (which is also the default for \fBgzip\fR(1)). .sp -For databases that create very large files but access them in small random chunks, these algorithms may be suboptimal. Specifying a "recordsize" greater than or equal to the record size of the database can result in significant performance gains. Use of this property for general -purpose file systems is strongly discouraged, and may adversely affect performance. +This property can also be referred to by its shortened column name "compress". +.RE + .sp -The size specified must be a power of two greater than or equal to 512 and less than or equal to 128 Kbytes. +.ne 2 +.mk +.na +\fBcopies=\fB1\fR | \fB2\fR | \fB3\fR\fR +.ad +.sp .6 +.RS 4n +Controls the number of copies of data stored for this dataset. These copies are in addition to any redundancy provided by the pool, for example, mirroring or raid-z. The copies are stored on different disks, if possible. The space used by multiple copies is charged to the associated +file and dataset, changing the "used" property and counting against quotas and reservations. .sp -Changing the file system's \fBrecordsize\fR only affects files created afterward; existing files are unaffected. +Changing this property only affects newly-written data. Therefore, set this property at file system creation time by using the "\fB-o\fR copies=" option. +.RE + .sp -This property can also be referred to by its shortened column name, "recsize". +.ne 2 +.mk +.na +\fBdevices=\fIon\fR | \fIoff\fR\fR +.ad +.sp .6 +.RS 4n +Controls whether device nodes can be opened on this file system. The default value is "on". +.RE + +.sp +.ne 2 +.mk +.na +\fBexec=\fIon\fR | \fIoff\fR\fR +.ad +.sp .6 +.RS 4n +Controls whether processes can be executed from within this file system. The default value is "on". .RE .sp @@ -484,89 +608,100 @@ previously "legacy" or "none", or if they were mounted before the property was c .ne 2 .mk .na -\fBsharenfs=\fIon\fR | \fIoff\fR | \fIopts\fR\fR +\fBnbmand=\fIon\fR | \fIoff\fR\fR .ad .sp .6 .RS 4n -Controls whether the file system is shared via \fBNFS\fR, and what options are used. A file system with a sharenfs property of "off" is managed through traditional tools such as \fBshare\fR(1M), \fBunshare\fR(1M), and \fBdfstab\fR(4). Otherwise, the file system is automatically shared and unshared with the "\fBzfs share\fR" and "\fBzfs unshare\fR" commands. If the property is set to "on", the \fBshare\fR(1M) command is invoked with no options. Otherwise, the \fBshare\fR(1M) command is invoked with options equivalent to the contents of this property. -.sp -When the "sharenfs" property is changed for a dataset, the dataset and any children inheriting the property are re-shared with the new options, only if the property was previously "off", or if they were shared before the property was changed. If the new property is "off", -the file systems are unshared. +Controls whether the file system should be mounted with "\fBnbmand\fR" (Non Blocking mandatory locks). This is used for \fBCIFS\fR clients. Changes to this property only take effect when the file system is umounted and remounted. See \fBmount\fR(1M) for more information on "\fBnbmand\fR" mounts. .RE .sp .ne 2 .mk .na -\fBshareiscsi=\fIon\fR | \fIoff\fR\fR +\fBquota=\fIsize\fR | \fInone\fR\fR .ad .sp .6 .RS 4n -Like the "sharenfs" property, "shareiscsi" indicates whether a \fBZFS\fR volume is exported as an \fBiSCSI\fR target. The acceptable values for this property are "on", "off", and "type=disk". -The default value is "off". In the future, other target types might be supported. For example, "tape". +Limits the amount of space a dataset and its descendents can consume. This property enforces a hard limit on the amount of space used. This includes all space consumed by descendents, including file systems and snapshots. Setting a quota on a descendent of a dataset that already +has a quota does not override the ancestor's quota, but rather imposes an additional limit. .sp -You might want to set "shareiscsi=on" for a file system so that all \fBZFS\fR volumes within the file system are shared by default. Setting this property on a file system has no direct effect, however. +Quotas cannot be set on volumes, as the "volsize" property acts as an implicit quota. .RE .sp .ne 2 .mk .na -\fBchecksum=\fIon\fR | \fIoff\fR | \fIfletcher2\fR, | \fIfletcher4\fR | \fIsha256\fR\fR +\fBreadonly=\fIon\fR | \fIoff\fR\fR .ad .sp .6 .RS 4n -Controls the checksum used to verify data integrity. The default value is "on", which automatically selects an appropriate algorithm (currently, \fIfletcher2\fR, but this may change in future releases). The value "off" disables integrity -checking on user data. Disabling checksums is NOT a recommended practice. +Controls whether this dataset can be modified. The default value is "off". +.sp +This property can also be referred to by its shortened column name, "rdonly". .RE .sp .ne 2 .mk .na -\fBcompression=\fIon\fR | \fIoff\fR | \fIlzjb\fR | \fIgzip\fR | \fIgzip-N\fR\fR +\fBrecordsize=\fIsize\fR\fR .ad .sp .6 .RS 4n -Controls the compression algorithm used for this dataset. The "lzjb" compression algorithm is optimized for performance while providing decent data compression. Setting compression to "on" uses the "lzjb" compression algorithm. The "gzip" -compression algorithm uses the same compression as the \fBgzip\fR(1) command. You can specify the "gzip" level by using the value "gzip-\fIN\fR", -where \fIN\fR is an integer from 1 (fastest) to 9 (best compression ratio). Currently, "gzip" is equivalent to "gzip-6" (which is also the default for \fBgzip\fR(1)). +Specifies a suggested block size for files in the file system. This property is designed solely for use with database workloads that access files in fixed-size records. \fBZFS\fR automatically tunes block sizes according to internal algorithms optimized for typical +access patterns. .sp -This property can also be referred to by its shortened column name "compress". +For databases that create very large files but access them in small random chunks, these algorithms may be suboptimal. Specifying a "recordsize" greater than or equal to the record size of the database can result in significant performance gains. Use of this property for general +purpose file systems is strongly discouraged, and may adversely affect performance. +.sp +The size specified must be a power of two greater than or equal to 512 and less than or equal to 128 Kbytes. +.sp +Changing the file system's \fBrecordsize\fR only affects files created afterward; existing files are unaffected. +.sp +This property can also be referred to by its shortened column name, "recsize". .RE .sp .ne 2 .mk .na -\fBatime=\fIon\fR | \fIoff\fR\fR +\fBrefquota=\fIsize\fR | \fInone\fR\fR .ad .sp .6 .RS 4n -Controls whether the access time for files is updated when they are read. Turning this property off avoids producing write traffic when reading files and can result in significant performance gains, though it might confuse mailers and other similar utilities. The default value -is "on". +Limits the amount of space a dataset can consume. This property enforces a hard limit on the amount of space used. This hard limit does not include space used by descendents, including file systems and snapshots. .RE .sp .ne 2 .mk .na -\fBdevices=\fIon\fR | \fIoff\fR\fR +\fBrefreservation=\fIsize\fR | \fInone\fR\fR .ad .sp .6 .RS 4n -Controls whether device nodes can be opened on this file system. The default value is "on". +The minimum amount of space guaranteed to a dataset, not including its descendents. When the amount of space used is below this value, the dataset is treated as if it were taking up the amount of space specified by \fBrefreservation\fR. The \fBrefreservation\fR reservation +is accounted for in the parent datasets' space used, and counts against the parent datasets' quotas and reservations. +.sp +If \fBrefreservation\fR is set, a snapshot is only allowed if there is enough free pool space outside of this reservation to accommodate the current number of "referenced" bytes in the dataset. +.sp +This property can also be referred to by its shortened column name, "refreserv". .RE .sp .ne 2 .mk .na -\fBexec=\fIon\fR | \fIoff\fR\fR +\fBreservation=\fIsize\fR | \fInone\fR\fR .ad .sp .6 .RS 4n -Controls whether processes can be executed from within this file system. The default value is "on". +The minimum amount of space guaranteed to a dataset and its descendents. When the amount of space used is below this value, the dataset is treated as if it were taking up the amount of space specified by its reservation. Reservations are accounted for in the parent datasets' space +used, and count against the parent datasets' quotas and reservations. +.sp +This property can also be referred to by its shortened column name, "reserv". .RE .sp @@ -584,24 +719,48 @@ Controls whether the set-\fBUID\fR bit is respected for the file system. The def .ne 2 .mk .na -\fBreadonly=\fIon\fR | \fIoff\fR\fR +\fBshareiscsi=\fIon\fR | \fIoff\fR\fR .ad .sp .6 .RS 4n -Controls whether this dataset can be modified. The default value is "off". +Like the "sharenfs" property, "shareiscsi" indicates whether a \fBZFS\fR volume is exported as an \fBiSCSI\fR target. The acceptable values for this property are "on", "off", and "type=disk". +The default value is "off". In the future, other target types might be supported. For example, "tape". .sp -This property can also be referred to by its shortened column name, "rdonly". +You might want to set "shareiscsi=on" for a file system so that all \fBZFS\fR volumes within the file system are shared by default. Setting this property on a file system has no direct effect, however. .RE .sp .ne 2 .mk .na -\fBzoned=\fIon\fR | \fIoff\fR\fR +\fBsharesmb=\fIon\fR | \fIoff\fR | \fIopts\fR\fR .ad .sp .6 .RS 4n -Controls whether the dataset is managed from a non-global zone. See the "Zones" section for more information. The default value is "off". +Controls whether the file system is shared by using the Solaris \fBCIFS\fR service, and what options are to be used. A file system with the "\fBsharesmb\fR" property set to "off" is managed through traditional tools such as \fBsharemgr\fR(1M). Otherwise, the file system is automatically shared and unshared with the "zfs share" and "zfs unshare" commands. If the property is set to "on", +the \fBsharemgr\fR(1M) command is invoked with no options. Otherwise, the \fBsharemgr\fR(1M) command is invoked with options equivalent to the contents of this property. +.sp +Because \fBSMB\fR shares requires a resource name, a unique resource name is constructed from the dataset name. The constructed name is a copy of the dataset name except that the characters in the dataset name, which would be illegal in the resource name, are replaced with underscore +(_) characters. A pseudo property "name" is also supported that allows you to replace the data set name with a specified name. The specified name is then used to replace the prefix dataset in the case of inheritance. For example, if the dataset "\fBdata/home/john\fR" +is set to "name=john", then "\fBdata/home/john\fR" has a resource name of "john". If a child dataset of "\fBdata/home/john/backups\fR", it has a resource name of "john_backups". +.sp +When the "sharesmb" property is changed for a dataset, the dataset and any children inheriting the property are re-shared with the new options, only if the property was previously set to "off", or if they were shared before the property was changed. If the new property +is set to "off", the file systems are unshared. +.RE + +.sp +.ne 2 +.mk +.na +\fBsharenfs=\fIon\fR | \fIoff\fR | \fIopts\fR\fR +.ad +.sp .6 +.RS 4n +Controls whether the file system is shared via \fBNFS\fR, and what options are used. A file system with a"\fBsharenfs\fR" property of "off" is managed through traditional tools such as \fBshare\fR(1M), \fBunshare\fR(1M), and \fBdfstab\fR(4). Otherwise, the file system is automatically shared and unshared with the "\fBzfs share\fR" and "\fBzfs unshare\fR" commands. If the property is set to "on", +the \fBshare\fR(1M) command is invoked with no options. Otherwise, the \fBshare\fR(1M) command is invoked with options equivalent to the contents of this property. +.sp +When the "sharenfs" property is changed for a dataset, the dataset and any children inheriting the property are re-shared with the new options, only if the property was previously "off", or if they were shared before the property was changed. If the new property is "off", +the file systems are unshared. .RE .sp @@ -619,40 +778,40 @@ Controls whether the ".zfs" directory is hidden or visible in the root of the fi .ne 2 .mk .na -\fBaclmode=\fBdiscard\fR | \fBgroupmask\fR | \fBpassthrough\fR\fR +\fBversion=\fB1\fR|\fB2\fR|\fBcurrent\fR\fR .ad .sp .6 .RS 4n -Controls how an \fBACL\fR is modified during \fBchmod\fR(2). A file system with an "aclmode" property of "\fBdiscard\fR" -deletes all \fBACL\fR entries that do not represent the mode of the file. An "aclmode" property of "\fBgroupmask\fR" (the default) reduces user or group permissions. The permissions are reduced, such that they are no greater than the group permission -bits, unless it is a user entry that has the same \fBUID\fR as the owner of the file or directory. In this case, the \fBACL\fR permissions are reduced so that they are no greater than owner permission bits. A file system with an "aclmode" property of "\fBpassthrough\fR" indicates that no changes will be made to the \fBACL\fR other than generating the necessary \fBACL\fR entries to represent the new mode of the file or directory. +The on-disk version of this file system, which is independent of the pool version. This property can only be set to later supported versions. See "\fBzfs upgrade\fR". .RE .sp .ne 2 .mk .na -\fBaclinherit=\fBdiscard\fR | \fBnoallow\fR | \fBsecure\fR | \fBpassthrough\fR\fR +\fBvolsize=\fIsize\fR\fR .ad .sp .6 .RS 4n -Controls how \fBACL\fR entries are inherited when files and directories are created. A file system with an "aclinherit" property of "\fBdiscard\fR" does not inherit any \fBACL\fR entries. A file system with an "aclinherit" -property value of "\fBnoallow\fR" only inherits inheritable \fBACL\fR entries that specify "deny" permissions. The property value "\fBsecure\fR" (the default) removes the "\fBwrite_acl\fR" and "\fBwrite_owner\fR" permissions when the \fBACL\fR entry is inherited. A file system with an "aclinherit" property value of "\fBpassthrough\fR" inherits all inheritable \fBACL\fR entries without any modifications made to the \fBACL\fR entries when they are inherited. +For volumes, specifies the logical size of the volume. By default, creating a volume establishes a reservation of equal size. For storage pools with a version number of 9 or higher, a \fBrefreservation\fR is set instead. Any changes to \fBvolsize\fR are +reflected in an equivalent change to the reservation (or \fBrefreservation\fR). The \fBvolsize\fR can only be set to a multiple of \fBvolblocksize\fR, and cannot be zero. +.sp +The reservation is kept equal to the volume's logical size to prevent unexpected behavior for consumers. Without the reservation, the volume could run out of space, resulting in undefined behavior or data corruption, depending on how the volume is used. These effects can also occur when +the volume size is changed while it is in use (particularly when shrinking the size). Extreme care should be used when adjusting the volume size. +.sp +Though not recommended, a "sparse volume" (also known as "thin provisioning") can be created by specifying the \fB-s\fR option to the "\fBzfs create -V\fR" command, or by changing the reservation after the volume has been created. +A "sparse volume" is a volume where the reservation is less then the volume size. Consequently, writes to a sparse volume can fail with \fBENOSPC\fR when the pool is low on space. For a sparse volume, changes to \fBvolsize\fR are not reflected in the reservation. .RE .sp .ne 2 .mk .na -\fBcanmount=\fBon\fR | \fBoff\fR\fR +\fBvscan=\fBon\fR|\fBoff\fR\fR .ad .sp .6 .RS 4n -If this property is set to "\fBoff\fR", the file system cannot be mounted, and is ignored by "\fBzfs mount -a\fR". This is similar to setting the "mountpoint" property to "\fBnone\fR", except -that the dataset still has a normal "mountpoint" property which can be inherited. This allows datasets to be used solely as a mechanism to inherit properties. One use case is to have two logically separate datasets have the same mountpoint, so that the children of both datasets appear -in the same directory, but may have different inherited characteristics. The default value is "\fBon\fR". -.sp -This property is not inherited. +Controls whether regular files should be scanned for viruses when a file is opened and closed. In addition to enabling this property, the virus scan service must also be enabled for virus scanning to occur. The default value is "off". .RE .sp @@ -670,31 +829,74 @@ Controls whether extended attributes are enabled for this file system. The defau .ne 2 .mk .na -\fBcopies=\fB1\fR | \fB2\fR | \fB3\fR\fR +\fBzoned=\fIon\fR | \fIoff\fR\fR .ad .sp .6 .RS 4n -Controls the number of copies of data stored for this dataset. These copies are in addition to any redundancy provided by the pool, for example, mirroring or raid-z. The copies are stored on different disks, if possible. The space used by multiple copies is charged to the associated -file and dataset, changing the "used" property and counting against quotas and reservations. +Controls whether the dataset is managed from a non-global zone. See the "Zones" section for more information. The default value is "off". +.RE + .sp -Changing this property only affects newly-written data. Therefore, set this property at file system creation time by using the "\fB-o\fR copies=" option. +.LP +The following three properties cannot be changed after the file system is created, and therefore, should be set when the file system is created. If the properties are not set with the "\fBzfs create\fR" command, these properties are inherited from the parent dataset. +If the parent dataset lacks these properties due to having been created prior to these features being supported, the new file system will have the default values for these properties. +.sp +.ne 2 +.mk +.na +\fBcasesensitivity = \fBsensitive\fR | \fBinsensitive\fR | \fBmixed\fR\fR +.ad +.sp .6 +.RS 4n +Indicates whether the file name matching algorithm used by the file system should be case-sensitive, case-insensitive, or allow a combination of both styles of matching. The default value for the "\fBcasesensitivity\fR" property is "\fBsensitive\fR." +Traditionally, UNIX and POSIX file systems have case-sensitive file names. +.sp +The "\fBmixed\fR" value for the "\fBcasesensitivity\fR" property indicates that the file system can support requests for both case-sensitive and case-insensitive matching behavior. Currently, case-insensitive matching behavior on a file system +that supports mixed behavior is limited to the Solaris CIFS server product. For more information about the "mixed" value behavior, see the \fIZFS Administration Guide\fR. +.RE + +.sp +.ne 2 +.mk +.na +\fBnormalization =\fBnone\fR | \fBformD\fR | \fBformKCf\fR\fR +.ad +.sp .6 +.RS 4n +Indicates whether the file system should perform a \fBunicode\fR normalization of file names whenever two file names are compared, and which normalization algorithm should be used. File names are always stored unmodified, names are normalized as part of any comparison +process. If this property is set to a legal value other than "\fBnone\fR," and the "\fButf8only\fR" property was left unspecified, the "\fButf8only\fR" property is automatically set to "\fBon\fR." +The default value of the "\fBnormalization\fR" property is "\fBnone\fR." This property cannot be changed after the file system is created. .RE .sp .ne 2 .mk .na -\fBjailed=\fIon\fR | \fIoff\fR\fR +\fBjailed =\fIon\fR | \fIoff\fR\fR .ad .sp .6 .RS 4n Controls whether the dataset is managed from within a jail. The default value is "off". .RE -.SS "iscsioptions" +.sp +.ne 2 +.mk +.na +\fButf8only =\fBon\fR | \fBoff\fR\fR +.ad +.sp .6 +.RS 4n +Indicates whether the file system should reject file names that include characters that are not present in the \fBUTF-8\fR character code set. If this property is explicitly set to "\fBoff\fR," the normalization property must either not be +explicitly set or be set to "\fBnone\fR." The default value for the "\fButf8only\fR" property is "off." This property cannot be changed after the file system is created. +.RE + +.sp .LP -This read-only property, which is hidden, is used by the \fBiSCSI\fR target daemon to store persistent information, such as the \fBIQN\fR. It cannot be viewed or modified using the \fBzfs\fR command. The contents are not intended for external consumers. +The "\fBcasesensitivity\fR," "\fBnormalization\fR," and "\fButf8only\fR" properties are also new permissions that can be assigned to non-privileged users by using the \fBZFS\fR delegated administration +feature. .SS "Temporary Mount Point Properties" +.sp .LP When a file system is mounted, either through \fBmount\fR(1M) for legacy mounts or the "\fBzfs mount\fR" command for normal file systems, its mount options are set according to its properties. The correlation between properties and mount options is as follows: @@ -711,30 +913,39 @@ its mount options are set according to its properties. The correlation between p .in -2 .sp +.sp .LP In addition, these options can be set on a per-mount basis using the \fB-o\fR option, without affecting the property that is stored on disk. The values specified on the command line override the values stored in the dataset. The \fB-nosuid\fR option is an alias for "nodevices,nosetuid". These properties are reported as "temporary" by the "\fBzfs get\fR" command. If the properties are changed while the dataset is mounted, the new setting overrides any temporary settings. .SS "User Properties" +.sp .LP In addition to the standard native properties, \fBZFS\fR supports arbitrary user properties. User properties have no effect on \fBZFS\fR behavior, but applications or administrators can use them to annotate datasets. +.sp .LP User property names must contain a colon (":") character, to distinguish them from native properties. They might contain lowercase letters, numbers, and the following punctuation characters: colon (":"), dash ("-"), period ("."), and underscore ("_"). The expected convention is that the property name is divided into two portions such as "\fImodule\fR:\fIproperty\fR", but this namespace is not enforced by \fBZFS\fR. User property names can be at most 256 characters, and cannot begin with a dash ("-"). +.sp .LP When making programmatic use of user properties, it is strongly suggested to use a reversed \fBDNS\fR domain name for the \fImodule\fR component of property names to reduce the chance that two independently-developed packages use the same property name for different purposes. Property names beginning with "com.sun." are reserved for use by Sun Microsystems. +.sp .LP The values of user properties are arbitrary strings, are always inherited, and are never validated. All of the commands that operate on properties ("zfs list", "zfs get", "zfs set", etc.) can be used to manipulate both native properties and user properties. Use the "\fBzfs inherit\fR" command to clear a user property . If the property is not defined in any parent dataset, it is removed entirely. Property values are limited to 1024 characters. .SS "Volumes as Swap or Dump Devices" +.sp .LP To set up a swap area, create a \fBZFS\fR volume of a specific size and then enable swap on that device. For more information, see the EXAMPLES section. +.sp .LP Do not swap to a file on a \fBZFS\fR file system. A \fBZFS\fR swap file configuration is not supported. +.sp .LP Using a \fBZFS\fR volume as a dump device is not supported. .SH SUBCOMMANDS +.sp .LP All subcommands that modify state are logged persistently to the pool in their original form. .sp @@ -752,7 +963,7 @@ Displays a help message. .ne 2 .mk .na -\fB\fBzfs create\fR [[\fB-o\fR property=value]...] \fIfilesystem\fR\fR +\fB\fBzfs create\fR [\fB-p\fR] [\fB-o\fR \fIproperty\fR=\fIvalue\fR] ... \fIfilesystem\fR\fR .ad .sp .6 .RS 4n @@ -761,7 +972,19 @@ Creates a new \fBZFS\fR file system. The file system is automatically mounted ac .ne 2 .mk .na -\fB\fB-o\fR property=value\fR +\fB\fB-p\fR\fR +.ad +.RS 21n +.rt +Creates all the non-existing parent datasets. Datasets created in this manner are automatically mounted according to the "mountpoint" property inherited from their parent. Any property specified on the command line using the \fB-o\fR option is ignored. If +the target filesystem already exists, the operation completes successfully. +.RE + +.sp +.ne 2 +.mk +.na +\fB\fB-o\fR \fIproperty\fR=\fIvalue\fR\fR .ad .RS 21n .rt @@ -775,7 +998,7 @@ error results if the same property is specified in multiple \fB-o\fR options. .ne 2 .mk .na -\fB\fBzfs create\fR [\fB-s\fR] [\fB-b\fR \fIblocksize\fR] [[\fB-o\fR property=value]...] \fB-V\fR \fIsize\fR \fIvolume\fR\fR +\fB\fBzfs create\fR [\fB-ps\fR] [\fB-b\fR \fIblocksize\fR] [\fB-o\fR \fIproperty\fR=\fIvalue\fR] ... \fB-V\fR \fIsize\fR \fIvolume\fR\fR .ad .sp .6 .RS 4n @@ -787,6 +1010,18 @@ the logical size as exported by the device. By default, a reservation of equal s .ne 2 .mk .na +\fB\fB-p\fR\fR +.ad +.RS 21n +.rt +Creates all the non-existing parent datasets. Datasets created in this manner are automatically mounted according to the "mountpoint" property inherited from their parent. Any property specified on the command line using the \fB-o\fR option is ignored. If +the target filesystem already exists, the operation completes successfully. +.RE + +.sp +.ne 2 +.mk +.na \fB\fB-s\fR\fR .ad .RS 21n @@ -798,7 +1033,7 @@ Creates a sparse volume with no reservation. See "volsize" in the Native Propert .ne 2 .mk .na -\fB\fB-o\fR property=value\fR +\fB\fB-o\fR \fIproperty\fR=\fIvalue\fR\fR .ad .RS 21n .rt @@ -837,7 +1072,7 @@ Destroys the given dataset. By default, the command unshares any file systems th .ad .RS 6n .rt -Recursively destroy all children. If a snapshot is specified, destroy all snapshots with this name in descendant file systems. +Recursively destroy all children. If a snapshot is specified, destroy all snapshots with this name in descendent file systems. .RE .sp @@ -848,7 +1083,7 @@ Recursively destroy all children. If a snapshot is specified, destroy all snapsh .ad .RS 6n .rt -Recursively destroy all dependents, including cloned file systems outside the target hierarchy. If a snapshot is specified, destroy all snapshots with this name in descendant file systems. +Recursively destroy all dependents, including cloned file systems outside the target hierarchy. If a snapshot is specified, destroy all snapshots with this name in descendent file systems. .RE .sp @@ -869,58 +1104,65 @@ Extreme care should be taken when applying either the \fB-r\fR or the \fB-f\fR o .ne 2 .mk .na -\fB\fBzfs clone\fR \fIsnapshot\fR \fIfilesystem\fR|\fIvolume\fR\fR +\fB\fBzfs snapshot\fR [\fB-r\fR] \fIfilesystem@snapname\fR|\fIvolume@snapname\fR\fR .ad .sp .6 .RS 4n -Creates a clone of the given snapshot. See the "Clones" section for details. The target dataset can be located anywhere in the \fBZFS\fR hierarchy, and is created as the same type as the original. -.RE - +Creates a snapshot with the given name. See the "Snapshots" section for details. .sp .ne 2 .mk .na -\fB\fBzfs promote\fR \fIfilesystem\fR\fR +\fB\fB-r\fR\fR .ad -.sp .6 -.RS 4n -Promotes a clone file system to no longer be dependent on its "origin" snapshot. This makes it possible to destroy the file system that the clone was created from. The clone parent-child dependency relationship is reversed, so that the "origin" file system -becomes a clone of the specified file system. -.sp -The snaphot that was cloned, and any snapshots previous to this snapshot, are now owned by the promoted clone. The space they use moves from the "origin" file system to the promoted clone, so enough space must be available to accommodate these snapshots. No new space is consumed -by this operation, but the space accounting is adjusted. The promoted clone must not have any conflicting snapshot names of its own. The "\fBrename\fR" subcommand can be used to rename any conflicting snapshots. +.RS 6n +.rt +Recursively create snapshots of all descendent datasets. Snapshots are taken atomically, so that all recursive snapshots correspond to the same moment in time. +.RE + .RE .sp .ne 2 .mk .na -\fB\fBzfs rename\fR \fIfilesystem\fR|\fIvolume\fR|\fIsnapshot\fR \fIfilesystem\fR|\fIvolume\fR|\fIsnapshot\fR\fR +\fB\fBzfs rollback\fR [\fB-rRf\fR] \fIsnapshot\fR\fR .ad .sp .6 .RS 4n -Renames the given dataset. The new target can be located anywhere in the \fBZFS\fR hierarchy, with the exception of snapshots. Snapshots can only be renamed within the parent file system or volume. When renaming a snapshot, the parent file system of the snapshot does -not need to be specified as part of the second argument. Renamed file systems can inherit new mount points, in which case they are unmounted and remounted at the new mount point. +Roll back the given dataset to a previous snapshot. When a dataset is rolled back, all data that has changed since the snapshot is discarded, and the dataset reverts to the state at the time of the snapshot. By default, the command refuses to roll back to a snapshot other than +the most recent one. In order to do so, all intermediate snapshots must be destroyed by specifying the \fB-r\fR option. +.sp +.ne 2 +.mk +.na +\fB\fB-r\fR\fR +.ad +.RS 6n +.rt +Recursively destroy any snapshots more recent than the one specified. .RE .sp .ne 2 .mk .na -\fB\fBzfs snapshot\fR [\fB-r\fR] \fIfilesystem@name\fR|\fIvolume@name\fR\fR +\fB\fB-R\fR\fR .ad -.sp .6 -.RS 4n -Creates a snapshot with the given name. See the "Snapshots" section for details. +.RS 6n +.rt +Recursively destroy any more recent snapshots, as well as any clones of those snapshots. +.RE + .sp .ne 2 .mk .na -\fB\fB-r\fR\fR +\fB\fB-f\fR\fR .ad .RS 6n .rt -Recursively create snapshots of all descendant datasets. Snapshots are taken atomically, so that all recursive snapshots correspond to the same moment in time. +Used with the \fB-R\fR option to force an unmount of any clone file systems that are to be destroyed. .RE .RE @@ -929,43 +1171,67 @@ Recursively create snapshots of all descendant datasets. Snapshots are taken ato .ne 2 .mk .na -\fB\fBzfs rollback\fR [\fB-rRf\fR] \fIsnapshot\fR\fR +\fB\fBzfs clone\fR [\fB-p\fR] \fIsnapshot\fR \fIfilesystem\fR|\fIvolume\fR\fR .ad .sp .6 .RS 4n -Roll back the given dataset to a previous snapshot. When a dataset is rolled back, all data that has changed since the snapshot is discarded, and the dataset reverts to the state at the time of the snapshot. By default, the command refuses to roll back to a snapshot other than -the most recent one. In order to do so, all intermediate snapshots must be destroyed by specifying the \fB-r\fR option. The file system is unmounted and remounted, if necessary. +Creates a clone of the given snapshot. See the "Clones" section for details. The target dataset can be located anywhere in the \fBZFS\fR hierarchy, and is created as the same type as the original. .sp .ne 2 .mk .na -\fB\fB-r\fR\fR +\fB\fB-p\fR\fR .ad .RS 6n .rt -Recursively destroy any snapshots more recent than the one specified. +Creates all the non-existing parent datasets. Datasets created in this manner are automatically mounted according to the "mountpoint" property inherited from their parent. If the target filesystem or volume already exists, the operation completes successfully. +.RE + .RE .sp .ne 2 .mk .na -\fB\fB-R\fR\fR +\fB\fBzfs promote\fR \fIclone-filesystem\fR\fR .ad -.RS 6n -.rt -Recursively destroy any more recent snapshots, as well as any clones of those snapshots. +.sp .6 +.RS 4n +Promotes a clone file system to no longer be dependent on its "origin" snapshot. This makes it possible to destroy the file system that the clone was created from. The clone parent-child dependency relationship is reversed, so that the "origin" file system +becomes a clone of the specified file system. +.sp +The snapshot that was cloned, and any snapshots previous to this snapshot, are now owned by the promoted clone. The space they use moves from the "origin" file system to the promoted clone, so enough space must be available to accommodate these snapshots. No new space is consumed +by this operation, but the space accounting is adjusted. The promoted clone must not have any conflicting snapshot names of its own. The "\fBrename\fR" subcommand can be used to rename any conflicting snapshots. .RE .sp .ne 2 .mk .na -\fB\fB-f\fR\fR +\fB\fBzfs rename\fR \fIfilesystem\fR|\fIvolume\fR|\fIsnapshot\fR\fR +.ad +.br +.na +\fB\fIfilesystem\fR|\fIvolume\fR|\fIsnapshot\fR\fR +.ad +.br +.na +\fB\fBzfs +rename\fR [\fB-p\fR] \fIfilesystem\fR|\fIvolume\fR \fIfilesystem\fR|\fIvolume\fR\fR +.ad +.sp .6 +.RS 4n +Renames the given dataset. The new target can be located anywhere in the \fBZFS\fR hierarchy, with the exception of snapshots. Snapshots can only be renamed within the parent file system or volume. When renaming a snapshot, the parent file system of the snapshot does +not need to be specified as part of the second argument. Renamed file systems can inherit new mount points, in which case they are unmounted and remounted at the new mount point. +.sp +.ne 2 +.mk +.na +\fB\fB-p\fR\fR .ad .RS 6n .rt -Force an unmount of any file systems using the "\fBunmount -f\fR" command. +Creates all the non-existing parent datasets. Datasets created in this manner are automatically mounted according to the "mountpoint" property inherited from their parent. .RE .RE @@ -974,7 +1240,22 @@ Force an unmount of any file systems using the "\fBunmount -f\fR" command. .ne 2 .mk .na -\fB\fBzfs\fR \fBlist\fR [\fB-rH\fR] [\fB-o\fR \fIprop\fR[,\fIprop\fR] ]... [ \fB-t\fR \fItype\fR[,\fItype\fR]...] [ \fB-s\fR \fIprop\fR [\fB-s\fR \fIprop\fR]... [ \fB-S\fR \fIprop\fR [\fB-S\fR \fIprop\fR]... [\fIfilesystem\fR|\fIvolume\fR|\fIsnapshot\fR|\fI/pathname\fR|.\fI/pathname\fR ...\fR +\fB\fBzfs rename\fR \fB-r\fR \fIsnapshot\fR \fIsnapshot\fR\fR +.ad +.sp .6 +.RS 4n +Recursively rename the snapshots of all descendent datasets. Snapshots are the only dataset that can be renamed recursively. +.RE + +.sp +.ne 2 +.mk +.na +\fB\fBzfs\fR \fBlist\fR [\fB-rH\fR] [\fB-o\fR \fIproperty\fR[,\fI\&...\fR]] [ \fB-t\fR \fItype\fR[,\fI\&...\fR]]\fR +.ad +.br +.na +\fB[ \fB-s\fR \fIproperty\fR ] ... [ \fB-S\fR \fIproperty\fR ] ... [\fIfilesystem\fR|\fIvolume\fR|\fIsnapshot\fR] ...\fR .ad .sp .6 .RS 4n @@ -993,7 +1274,7 @@ name,used,available,referenced,mountpoint .na \fB\fB-H\fR\fR .ad -.RS 11n +.RS 15n .rt Used for scripting mode. Do not print headers and separate fields by a single tab instead of arbitrary whitespace. .RE @@ -1004,7 +1285,7 @@ Used for scripting mode. Do not print headers and separate fields by a single ta .na \fB\fB-r\fR\fR .ad -.RS 11n +.RS 15n .rt Recursively display any children of the dataset on the command line. .RE @@ -1013,9 +1294,9 @@ Recursively display any children of the dataset on the command line. .ne 2 .mk .na -\fB\fB-o\fR \fIprop\fR\fR +\fB\fB-o\fR \fIproperty\fR\fR .ad -.RS 11n +.RS 15n .rt A comma-separated list of properties to display. The property must be one of the properties described in the "Native Properties" section, or the special value "name" to display the dataset name. .RE @@ -1024,12 +1305,12 @@ A comma-separated list of properties to display. The property must be one of the .ne 2 .mk .na -\fB\fB-s\fR \fIprop\fR\fR +\fB\fB-s\fR \fIproperty\fR\fR .ad -.RS 11n +.RS 15n .rt A property to use for sorting the output by column in ascending order based on the value of the property. The property must be one of the properties described in the "Properties" section, or the special value "name" to sort by the dataset name. Multiple -properties can be specified at one time using multiple \fB-s\fR property options. Multiple \fB-s\fR options are evaluated from left to right in decreasing order of importance. +properties can be specified at one time using multiple \fB-s\fR property options. Multiple \fB-s\fR options are evaluated from left to right in decreasing order of importance. .sp The following is a list of sorting criteria: .RS +4 @@ -1062,9 +1343,9 @@ If no sorting options are specified the existing behavior of "\fBzfs list\fR" is .ne 2 .mk .na -\fB\fB-S\fR \fIprop\fR\fR +\fB\fB-S\fR \fIproperty\fR\fR .ad -.RS 11n +.RS 15n .rt Same as the \fB-s\fR option, but sorts by property in descending order. .RE @@ -1075,7 +1356,7 @@ Same as the \fB-s\fR option, but sorts by property in descending order. .na \fB\fB-t\fR \fItype\fR\fR .ad -.RS 11n +.RS 15n .rt A comma-separated list of types to display, where "type" is one of "filesystem", "snapshot" or "volume". For example, specifying "\fB-t snapshot\fR" displays only snapshots. .RE @@ -1098,7 +1379,7 @@ form with a suffix of "B", "K", "M", "G", "T", "P", "E", "Z" (for bytes, Kbytes, .ne 2 .mk .na -\fB\fBzfs get\fR [\fB-rHp\fR] [\fB-o\fR \fIfield\fR[,\fIfield\fR]...] [\fB-s\fR \fIsource\fR[,\fIsource\fR]...] \fIall\fR | \fIproperty\fR[,\fIproperty\fR]... \fIfilesystem\fR|\fIvolume\fR|\fIsnapshot\fR ...\fR +\fB\fBzfs get\fR [\fB-rHp\fR] [\fB-o\fR \fIfield\fR[,...] [\fB-s\fR \fIsource\fR[,...] "\fIall\fR" | \fIproperty\fR[,...] \fIfilesystem\fR|\fIvolume\fR|\fIsnapshot\fR ...\fR .ad .sp .6 .RS 4n @@ -1201,42 +1482,70 @@ Recursively inherit the given property for all children. .ne 2 .mk .na -\fB\fBzfs mount\fR\fR +\fB\fBzfs upgrade\fR [\fB-v\fR]\fR .ad .sp .6 .RS 4n -Displays all \fBZFS\fR file systems currently mounted. +Displays a list of file systems that are not the most recent version. .RE .sp .ne 2 .mk .na -\fB\fBzfs mount\fR[\fB-o\fR \fIopts\fR] [\fB-O\fR] \fB-a\fR\fR +\fB\fBzfs upgrade\fR [\fB-r\fR] [\fB-V\fR \fIversion\fR] [\fB-a\fR | \fIfilesystem\fR]\fR .ad .sp .6 .RS 4n -Mounts all available \fBZFS\fR file systems. Invoked automatically as part of the boot process. +Upgrades file systems to a new on-disk version. Once this is done, the file systems will no longer be accessible on systems running older versions of the software. "\fBzfs send\fR" streams generated from new snapshots of these file systems can not be accessed +on systems running older versions of the software. +.sp +The file system version is independent of the pool version (see \fBzpool\fR(1M) for information on the "\fBzpool upgrade\fR" command). +.sp +The file system version does not have to be upgraded when the pool version is upgraded, and vice versa. .sp .ne 2 .mk .na -\fB\fB-o\fR \fIopts\fR\fR +\fB\fB-a\fR\fR .ad -.RS 11n +.RS 14n .rt -An optional comma-separated list of mount options to use temporarily for the duration of the mount. See the "Temporary Mount Point Properties" section for details. +Upgrade all file systems on all imported pools. .RE .sp .ne 2 .mk .na -\fB\fB-O\fR\fR +\fB\fIfilesystem\fR\fR .ad -.RS 11n +.RS 14n .rt -Perform an overlay mount. See \fBmount\fR(1M) for more information. +Upgrade the specified file system. +.RE + +.sp +.ne 2 +.mk +.na +\fB\fB-r\fR\fR +.ad +.RS 14n +.rt +Upgrade the specified file system and all descendent file systems +.RE + +.sp +.ne 2 +.mk +.na +\fB\fB-V\fR \fIversion\fR\fR +.ad +.RS 14n +.rt +Upgrade to the specified \fIversion\fR. If the \fB-V\fR flag is not specified, this command upgrades to the most recent version. This option can only be used to increase the version number, and only up to the most recent version supported by this +software. .RE .RE @@ -1245,18 +1554,29 @@ Perform an overlay mount. See \fBmount\fR(1M) for more information. .ne 2 .mk .na -\fB\fBzfs mount\fR [\fB-o\fR \fIopts\fR] [\fB-O\fR] \fIfilesystem\fR\fR +\fB\fBzfs mount\fR\fR .ad .sp .6 .RS 4n -Mounts a specific \fBZFS\fR file system. This is typically not necessary, as file systems are automatically mounted when they are created or the mountpoint property has changed. See the "Mount Points" section for details. +Displays all \fBZFS\fR file systems currently mounted. +.RE + .sp .ne 2 .mk .na -\fB\fB-o\fR \fIopts\fR\fR +\fB\fBzfs mount\fR [\fB-vO\fR] [\fB-o\fR \fIoptions\fR] \fB-a\fR | \fIfilesystem\fR\fR .ad -.RS 11n +.sp .6 +.RS 4n +Mounts \fBZFS\fR file systems. Invoked automatically as part of the boot process. +.sp +.ne 2 +.mk +.na +\fB\fB-o\fR \fIoptions\fR\fR +.ad +.RS 14n .rt An optional comma-separated list of mount options to use temporarily for the duration of the mount. See the "Temporary Mount Point Properties" section for details. .RE @@ -1267,97 +1587,152 @@ An optional comma-separated list of mount options to use temporarily for the dur .na \fB\fB-O\fR\fR .ad -.RS 11n +.RS 14n .rt Perform an overlay mount. See \fBmount\fR(1M) for more information. .RE +.sp +.ne 2 +.mk +.na +\fB\fB-v\fR\fR +.ad +.RS 14n +.rt +Report mount progress. .RE .sp .ne 2 .mk .na -\fB\fBzfs unmount\fR \fB-a\fR\fR +\fB\fB-a\fR\fR .ad -.sp .6 -.RS 4n -Unmounts all currently mounted \fBZFS\fR file systems. Invoked automatically as part of the shutdown process. +.RS 14n +.rt +Mount all available \fBZFS\fR file systems. Invoked automatically as part of the boot process. .RE .sp .ne 2 .mk .na -\fB\fBzfs unmount\fR [\fB-f\fR] \fIfilesystem\fR|\fImountpoint\fR\fR +\fB\fIfilesystem\fR\fR +.ad +.RS 14n +.rt +Mount the specified filesystem. +.RE + +.RE + +.sp +.ne 2 +.mk +.na +\fB\fBzfs unmount\fR [\fB-f\fR] \fB-a\fR | \fIfilesystem\fR|\fImountpoint\fR\fR .ad .sp .6 .RS 4n -Unmounts the given file system. The command can also be given a path to a \fBZFS\fR file system mount point on the system. +Unmounts currently mounted \fBZFS\fR file systems. Invoked automatically as part of the shutdown process. .sp .ne 2 .mk .na \fB\fB-f\fR\fR .ad -.RS 6n +.RS 25n .rt Forcefully unmount the file system, even if it is currently in use. .RE +.sp +.ne 2 +.mk +.na +\fB\fB-a\fR\fR +.ad +.RS 25n +.rt +Unmount all available \fBZFS\fR file systems. Invoked automatically as part of the boot process. .RE .sp .ne 2 .mk .na -\fB\fBzfs share\fR \fB-a\fR\fR +\fB\fIfilesystem\fR|\fImountpoint\fR\fR .ad -.sp .6 -.RS 4n -Shares all available \fBZFS\fR file systems. This is invoked automatically as part of the boot process. +.RS 25n +.rt +Unmount the specified filesystem. The command can also be given a path to a \fBZFS\fR file system mount point on the system. +.RE + .RE .sp .ne 2 .mk .na -\fB\fBzfs share\fR \fIfilesystem\fR\fR +\fB\fBzfs share\fR \fB-a\fR | \fIfilesystem\fR\fR .ad .sp .6 .RS 4n -Shares a specific \fBZFS\fR file system according to the "sharenfs" property. File systems are shared when the "sharenfs" property is set. +Shares available \fBZFS\fR file systems. +.sp +.ne 2 +.mk +.na +\fB\fB-a\fR\fR +.ad +.RS 14n +.rt +Share all available \fBZFS\fR file systems. Invoked automatically as part of the boot process. .RE .sp .ne 2 .mk .na -\fB\fBzfs unshare\fR \fB-a\fR\fR +\fB\fIfilesystem\fR\fR .ad -.sp .6 -.RS 4n -Unshares all currently shared \fBZFS\fR file systems. This is invoked automatically as part of the shutdown process. +.RS 14n +.rt +Share the specified filesystem according to the "sharenfs" and "sharesmb" properties. File systems are shared when the "sharenfs" or "sharesmb" property is set. +.RE + .RE .sp .ne 2 .mk .na -\fB\fBzfs unshare\fR [\fB-F\fR] \fIfilesystem\fR|\fImountpoint\fR\fR +\fB\fBzfs unshare\fR \fB-a\fR | \fIfilesystem\fR|\fImountpoint\fR\fR .ad .sp .6 .RS 4n -Unshares the given file system. The command can also be given a path to a \fBZFS\fR file system shared on the system. +Unshares currently shared \fBZFS\fR file systems. This is invoked automatically as part of the shutdown process. .sp .ne 2 .mk .na -\fB\fB-F\fR\fR +\fB\fB-a\fR\fR .ad -.RS 6n +.RS 25n .rt -Forcefully unshare the file system, even if it is currently in use. +Unshare all available \fBZFS\fR file systems. Invoked automatically as part of the boot process. +.RE + +.sp +.ne 2 +.mk +.na +\fB\fIfilesystem\fR|\fImountpoint\fR\fR +.ad +.RS 25n +.rt +Unshare the specified filesystem. The command can also be given a path to a \fBZFS\fR file system shared on the system. .RE .RE @@ -1366,27 +1741,64 @@ Forcefully unshare the file system, even if it is currently in use. .ne 2 .mk .na -\fB\fBzfs send\fR [\fB-i\fR \fIsnapshot1\fR] \fIsnapshot2\fR\fR +\fB\fBzfs send\fR [\fB-vR\fR] [\fB-\fR[\fB-iI\fR] \fIsnapshot\fR] \fIsnapshot\fR\fR .ad .sp .6 .RS 4n -Creates a stream representation of snapshot2, which is written to standard output. The output can be redirected to a file or to a different system (for example, using \fBssh\fR(1). By default, a full stream is generated. +Creates a stream representation of the second \fIsnapshot\fR, which is written to standard output. The output can be redirected to a file or to a different system (for example, using \fBssh\fR(1). By default, a full stream is generated. .sp .ne 2 .mk .na -\fB\fB-i\fR \fIsnapshot1\fR\fR +\fB\fB-i\fR \fIsnapshot\fR\fR .ad -.RS 16n +.RS 15n .rt -Generate an incremental stream from \fIsnapshot1\fR to \fIsnapshot2\fR. The incremental source \fIsnapshot1\fR can be specified as the last component of the snapshot name (for example, the part after the "@"), -and it is assumed to be from the same file system as \fIsnapshot2\fR. +Generate an incremental stream from the first \fIsnapshot\fR to the second \fIsnapshot\fR. The incremental source (the first \fIsnapshot\fR) can be specified as the last component of the snapshot name (for example, +the part after the "@"), and it is assumed to be from the same file system as the second \fIsnapshot\fR. +.sp +If the destination is a clone, the source may be the origin snapshot, which must be fully specified (for example, "pool/fs@origin", not just "@origin"). .RE +.sp +.ne 2 +.mk +.na +\fB\fB-I\fR \fIsnapshot\fR\fR +.ad +.RS 15n +.rt +Generate a stream package that sends all intermediary snapshots from the first snapshot to the second snapshot. For example, "\fB-I @a fs@d\fR" is similar to "\fB-i @a fs@b; -i @b fs@c; -i @c fs@d\fR". The incremental source snapshot +may be specified as with the \fB-i\fR option. +.RE + +.sp +.ne 2 +.mk +.na +\fB\fB-R\fR\fR +.ad +.RS 15n +.rt +Generate a replication stream package, which will replicate the specified filesystem, and all descendant file systems, up to the named snapshot. When received, all properties, snapshots, descendent file systems, and clones are preserved. +.sp +If the \fB-i\fR or \fB-I\fR flags are used in conjunction with the \fB-R\fR flag, an incremental replication stream is generated. The current values of properties, and current snapshot and file system names are set when the stream is received. If the \fB-F\fR flag is specified when this stream is recieved, snapshots and file systems that do not exist on the sending side are destroyed. +.RE + +.sp +.ne 2 +.mk +.na +\fB\fB-v\fR\fR +.ad +.RS 15n +.rt +Print verbose information about the stream package generated. .RE -.LP The format of the stream is evolving. No backwards compatibility is guaranteed. You may not be able to receive your streams on future versions of \fBZFS\fR. +.RE + .sp .ne 2 .mk @@ -1402,7 +1814,8 @@ The format of the stream is evolving. No backwards compatibility is guaranteed. Creates a snapshot whose contents are as specified in the stream provided on standard input. If a full stream is received, then a new file system is created as well. Streams are created using the "\fBzfs send\fR" subcommand, which by default creates a full stream. "\fBzfs recv\fR" can be used as an alias for "\fBzfs receive\fR". .sp -If an incremental stream is received, then the destination file system must already exist, and its most recent snapshot must match the incremental stream's source. The destination file system and all of its child file systems are unmounted and cannot be accessed during the receive operation. +If an incremental stream is received, then the destination file system must already exist, and its most recent snapshot must match the incremental stream's source. For \fBzvols\fR, the destination device link is destroyed and re-created, which means the \fBzvol\fR cannot +be accessed during the \fBreceive\fR operation. .sp The name of the snapshot (and file system, if a full stream is received) that this subcommand creates depends on the argument type and the \fB-d\fR option. .sp @@ -1438,7 +1851,7 @@ Print verbose information about the stream and the time required to perform the .ad .RS 6n .rt -Do not actually receive the stream. This can be useful in conjunction with the \fB-v\fR option to determine what name the receive operation would use. +Do not actually receive the stream. This can be useful in conjunction with the \fB-v\fR option to verify the name the receive operation would use. .RE .sp @@ -1449,7 +1862,8 @@ Do not actually receive the stream. This can be useful in conjunction with the \ .ad .RS 6n .rt -Force a rollback of the \fIfilesystem\fR to the most recent snapshot before performing the receive operation. +Force a rollback of the file system to the most recent snapshot before performing the receive operation. If receiving an incremental replication stream (for example, one generated by "z\fBfs send -R -[iI]\fR"), destroy snapshots and file systems that do +not exist on the sending side. .RE .RE @@ -1458,6 +1872,186 @@ Force a rollback of the \fIfilesystem\fR to the most recent snapshot before perf .ne 2 .mk .na +\fB\fBzfs allow\fR [\fB-ldug\fR] "\fIeveryone\fR"|\fIuser\fR|\fIgroup\fR[,...] \fIperm\fR|@\fIsetname\fR[,...] \fIfilesystem\fR| \fIvolume\fR\fR +.ad +.br +.na +\fB\fBzfs allow\fR [\fB-ld\fR] \fB-e\fR \fIperm\fR|@\fIsetname\fR[,...] \fIfilesystem\fR|\fIvolume\fR\fR +.ad +.sp .6 +.RS 4n +Delegates \fBZFS\fR administration permission for the file systems to non-privileged users. +.sp +.ne 2 +.mk +.na +\fB[\fB-ug\fR] "\fIeveryone\fR"|\fIuser\fR|\fIgroup\fR[,...]\fR +.ad +.sp .6 +.RS 4n +Specifies to whom the permissions are delegated. Multiple entities can be specified as a comma-separated list. If neither of the \fB-ug\fR options are specified, then the argument is interpreted preferentially as the keyword "everyone", then as a user name, +and lastly as a group name. To specify a user or group named "everyone", use the \fB-u\fR or \fB-g\fR options. To specify a group with the same name as a user, use the \fB-g\fR options. +.RE + +.sp +.ne 2 +.mk +.na +\fB[\fB-e\fR] \fIperm\fR|@\fIsetname\fR[,...]\fR +.ad +.sp .6 +.RS 4n +Specifies that the permissions be delegated to "everyone." Multiple permissions may be specified as a comma-separated list. Permission names are the same as \fBZFS\fR subcommand and property names. See the property list below. Property set names, which +begin with an "at sign" ("@") , may be specified. See the \fB-s\fR form below for details. +.RE + +.sp +.ne 2 +.mk +.na +\fB[\fB-ld\fR] \fIfilesystem\fR|\fIvolume\fR\fR +.ad +.sp .6 +.RS 4n +Specifies where the permissions are delegated. If neither of the \fB-ld\fR options are specified, or both are, then the permissions are allowed for the file system or volume, and all of its descendents. If only the \fB-l\fR option is used, then is allowed "locally" +only for the specified file system. If only the \fB-d\fR option is used, then is allowed only for the descendent file systems. +.RE + +.RE + +.sp +.LP +Permissions are generally the ability to use a \fBZFS\fR subcommand or change a \fBZFS\fR property. The following permissions are available: +.sp +.in +2 +.nf +NAME TYPE NOTES +allow subcommand Must also have the permission + that is being allowed. +clone subcommand Must also have the 'create' ability + and the 'mount' ability in the origin + file system. +create subcommand Must also have the 'mount' ability. +destroy subcommand Must also have the 'mount' ability. +mount subcommand Allows mount, unmount, and + create/remove zvol device links. +promote subcommand Must also have the 'mount' ability and + 'promote' ability in the origin file system. +receive subcommand Must also have the 'mount' ability and + the 'create' ability. +rename subcommand Must also have the 'mount' ability and + the 'create' ability in the new parent. +rollback subcommand Must also have the 'mount' ability. +snapshot subcommand Must also have the 'mount' ability. +share subcommand Allows share and unshare. +send subcommand + + +aclinherit property +aclmode property +atime property +canmount property +checksum property +compression property +copies property +devices property +exec property +mountpoint property +quota property +readonly property +recordsize property +reservation property +setuid property +shareiscsi property +sharenfs property +snapdir property +version property +volsize property +xattr property +zoned property +userprop other Allows changing any user property. +.fi +.in -2 +.sp + +.sp +.ne 2 +.mk +.na +\fB\fBzfs allow\fR \fB-c\fR \fIperm\fR|@\fIsetname\fR[,...] \fIfilesystem\fR|\fIvolume\fR\fR +.ad +.sp .6 +.RS 4n +Sets "create time" permissions. These permissions are granted (locally) to the creator of any newly-created descendent file system. +.RE + +.sp +.ne 2 +.mk +.na +\fB\fBzfs allow\fR \fB-s\fR @setname \fIperm\fR|@\fIsetname\fR[,...] \fIfilesystem\fR|\fIvolume\fR\fR +.ad +.sp .6 +.RS 4n +Defines or adds permissions to a permission set. The set can be used by other \fBzfs allow\fR commands for the specified file system and its descendents. Sets are evaluated dynamically, so changes to a set are immediately reflected. Permission sets follow the same +naming restrictions as ZFS file systems, but the name must begin with an "at sign" ("@"), and can be no more than 64 characters long. +.RE + +.sp +.ne 2 +.mk +.na +\fB\fBzfs unallow\fR [\fB-rldug\fR] "\fIeveryone\fR"|\fIuser\fR|\fIgroup\fR[,...] [\fIperm\fR|@\fIsetname\fR[, ...]] \fIfilesystem\fR|\fIvolume\fR\fR +.ad +.br +.na +\fB\fBzfs unallow\fR [\fB-rld\fR] \fB-e\fR [\fIperm\fR|@\fIsetname\fR [,...]] \fIfilesystem\fR|\fIvolume\fR\fR +.ad +.br +.na +\fB\fBzfs unallow\fR [\fB-r\fR] \fB-c\fR [\fIperm\fR|@\fIsetname\fR[,...]]\fR +.ad +.br +.na +\fB\fIfilesystem\fR|\fIvolume\fR\fR +.ad +.sp .6 +.RS 4n +Removes permissions that were granted with the "\fBzfs allow\fR" command. No permissions are explicitly denied, so other permissions granted are still in effect. For example, if the permission is granted by an ancestor. If no permissions are specified, +then all permissions for the specified \fIuser\fR, \fIgroup\fR, or \fIeveryone\fR are removed. Specifying "everyone" (or using the \fB-e\fR option) only removes the permissions that were granted to "everyone", +not all permissions for every user and group. See the "\fBzfs allow\fR" command for a description of the \fB-ldugec\fR options. +.sp +.ne 2 +.mk +.na +\fB\fB-r\fR\fR +.ad +.RS 6n +.rt +Recursively remove the permissions from this file system and all descendents. +.RE + +.RE + +.sp +.ne 2 +.mk +.na +\fB\fBzfs unallow\fR [\fB-r\fR] \fB-s\fR @setname [\fIperm\fR|@\fIsetname\fR[,...]]\fR +.ad +.br +.na +\fB\fIfilesystem\fR|\fIvolume\fR\fR +.ad +.sp .6 +.RS 4n +Removes permissions from a permission set. If no permissions are specified, then all permissions are removed, thus removing the set entirely. +.RE + +.sp +.ne 2 +.mk +.na \fB\fBzfs jail\fR \fIjailid\fR \fIfilesystem\fR\fR .ad .sp .6 @@ -1480,6 +2074,7 @@ Detaches the given file system from the given jail. .SH EXAMPLES .LP \fBExample 1 \fRCreating a ZFS File System Hierarchy +.sp .LP The following commands create a file system named "\fBpool/home\fR" and a file system named "\fBpool/home/bob\fR". The mount point "\fB/export/home\fR" is set for the parent file system, and automatically inherited by the child file system. @@ -1496,6 +2091,7 @@ by the child file system. .LP \fBExample 2 \fRCreating a ZFS Snapshot +.sp .LP The following command creates a snapshot named "yesterday". This snapshot is mounted on demand in the ".zfs/snapshot" directory at the root of the "\fBpool/home/bob\fR" file system. @@ -1509,8 +2105,9 @@ The following command creates a snapshot named "yesterday". This snapshot is mou .LP \fBExample 3 \fRTaking and destroying multiple snapshots +.sp .LP -The following command creates snapshots named "\fByesterday\fR" of "\fBpool/home\fR" and all of its descendant file systems. Each snapshot is mounted on demand in the ".zfs/snapshot" directory at the root of its file system. The +The following command creates snapshots named "\fByesterday\fR" of "\fBpool/home\fR" and all of its descendent file systems. Each snapshot is mounted on demand in the ".zfs/snapshot" directory at the root of its file system. The second command destroys the newly created snapshots. .sp @@ -1524,6 +2121,7 @@ second command destroys the newly created snapshots. .LP \fBExample 4 \fRTurning Off Compression +.sp .LP The following commands turn compression off for all file systems under "\fBpool/home\fR", but explicitly turns it on for "\fBpool/home/anne\fR". @@ -1538,6 +2136,7 @@ The following commands turn compression off for all file systems under "\fBpool/ .LP \fBExample 5 \fRListing ZFS Datasets +.sp .LP The following command lists all active file systems and volumes in the system. @@ -1548,17 +2147,18 @@ The following command lists all active file systems and volumes in the system. NAME USED AVAIL REFER MOUNTPOINT - pool 100G 60G - /pool - pool/home 100G 60G - /export/home - pool/home/bob 40G 60G 40G /export/home/bob - pool/home/bob@yesterday 3M - 40G - - pool/home/anne 60G 60G 40G /export/home/anne + pool 450K 457G 18K /pool + pool/home 315K 457G 21K /export/home + pool/home/anne 18K 457G 18K /export/home/anne + pool/home/bob 276K 457G 276K /export/home/bob + pool/home/bob@yesterday 0 - 276K - .fi .in -2 .sp .LP \fBExample 6 \fRSetting a Quota on a ZFS File System +.sp .LP The following command sets a quota of 50 gbytes for "\fBpool/home/bob\fR". @@ -1572,6 +2172,7 @@ The following command sets a quota of 50 gbytes for "\fBpool/home/bob\fR". .LP \fBExample 7 \fRListing ZFS Properties +.sp .LP The following command lists all properties for "\fBpool/home/bob\fR". @@ -1583,18 +2184,17 @@ The following command lists all properties for "\fBpool/home/bob\fR". NAME PROPERTY VALUE SOURCE pool/home/bob type filesystem - - pool/home/bob creation Fri Feb 23 14:20 2007 - - pool/home/bob used 24.5K - + pool/home/bob creation Thu Jul 12 14:44 2007 - + pool/home/bob used 276K - pool/home/bob available 50.0G - - pool/home/bob referenced 24.5K - + pool/home/bob referenced 276K - pool/home/bob compressratio 1.00x - pool/home/bob mounted yes - pool/home/bob quota 50G local pool/home/bob reservation none default pool/home/bob recordsize 128K default - pool/home/bob mountpoint /pool/home/bob default - pool/home/bob sharenfs off default - pool/home/bob shareiscsi off default + pool/home/bob mountpoint /export/home/bob inherited from + pool/home pool/home/bob checksum on default pool/home/bob compression off default pool/home/bob atime on default @@ -1605,15 +2205,24 @@ The following command lists all properties for "\fBpool/home/bob\fR". pool/home/bob zoned off default pool/home/bob snapdir hidden default pool/home/bob aclmode groupmask default - pool/home/bob aclinherit secure default + pool/home/bob aclinherit restricted default pool/home/bob canmount on default + pool/home/bob nbmand off default + pool/home/bob shareiscsi off default + pool/home/bob sharesmb off default + pool/home/bob sharenfs off default pool/home/bob xattr on default + pool/home/bob refquota 10M local + pool/home/bob refreservation none default + pool/home/bob copies 1 default + pool/home/bob version 1 - .fi .in -2 .sp +.sp .LP The following command gets a single property value. @@ -1626,6 +2235,7 @@ on .in -2 .sp +.sp .LP The following command lists all properties with local settings for "\fBpool/home/bob\fR". @@ -1643,6 +2253,7 @@ The following command lists all properties with local settings for "\fBpool/home .LP \fBExample 8 \fRRolling Back a ZFS File System +.sp .LP The following command reverts the contents of "\fBpool/home/anne\fR" to the snapshot named "\fByesterday\fR", deleting all intermediate snapshots. @@ -1656,6 +2267,7 @@ The following command reverts the contents of "\fBpool/home/anne\fR" to the snap .LP \fBExample 9 \fRCreating a ZFS Clone +.sp .LP The following command creates a writable file system whose initial contents are the same as "\fBpool/home/bob@yesterday\fR". @@ -1669,6 +2281,7 @@ The following command creates a writable file system whose initial contents are .LP \fBExample 10 \fRPromoting a ZFS Clone +.sp .LP The following commands illustrate how to test out changes to a file system, and then replace the original file system with the changed one, using clones, clone promotion, and renaming: @@ -1692,6 +2305,7 @@ The following commands illustrate how to test out changes to a file system, and .LP \fBExample 11 \fRInheriting ZFS Properties +.sp .LP The following command causes "\fBpool/home/bob\fR" and "\fBpool/home/anne\fR" to inherit the "checksum" property from their parent. @@ -1705,6 +2319,7 @@ The following command causes "\fBpool/home/bob\fR" and "\fBpool/home/anne\fR" to .LP \fBExample 12 \fRRemotely Replicating ZFS Data +.sp .LP The following commands send a full stream and then an incremental stream to a remote machine, restoring them into "\fBpoolB/received/fs\fR@a" and "\fBpoolB/received/fs@b\fR", respectively. "\fBpoolB\fR" must contain the file system "\fBpoolB/received\fR", and must not initially contain "\fBpoolB/received/fs\fR". @@ -1721,30 +2336,31 @@ the file system "\fBpoolB/received\fR", and must not initially contain "\fBpoolB .sp .LP -\fBExample 13 \fRUsing the zfs receive -d Option +\fBExample 13 \fRUsing the zfs receive -d Option +.sp .LP The following command sends a full stream of "\fBpoolA/fsA/fsB@snap\fR" to a remote machine, receiving it into "\fBpoolB/received/fsA/fsB@snap\fR". The "\fBfsA/fsB@snap\fR" portion of the received snapshot's name -is determined from the name of the sent snapshot. "\fBpoolB\fR" must contain the file system "\fBpoolB/received\fR". If "\fBpoolB/received/fsA\fR" does not exist, it will be created as an empty file system. +is determined from the name of the sent snapshot. "\fBpoolB\fR" must contain the file system "\fBpoolB/received\fR". If "\fBpoolB/received/fsA\fR" does not exist, it is be created as an empty file system. .sp .in +2 .nf \fB# zfs send poolA/fsA/fsB@snap | \e - ssh host zfs receive -d poolB/received - \fR + ssh host zfs receive -d poolB/received\fR .fi .in -2 .sp .LP \fBExample 14 \fRCreating a ZFS volume as a Swap Device +.sp .LP The following example shows how to create a 5-Gbyte ZFS volume and then add the volume as a swap device. .sp .in +2 .nf -\fB# zfs create -V 5gb tank/vol +\fB# zfs create -V 5gb tank/vol # swap -a /dev/zvol/dsk/tank/vol\fR .fi .in -2 @@ -1752,6 +2368,7 @@ The following example shows how to create a 5-Gbyte ZFS volume and then add the .LP \fBExample 15 \fRSetting User Properties +.sp .LP The following example sets the user defined "com.example:department" property for a dataset. @@ -1765,6 +2382,7 @@ The following example sets the user defined "com.example:department" property fo .LP \fBExample 16 \fRCreating a ZFS Volume as a iSCSI Target Device +.sp .LP The following example shows how to create a \fBZFS\fR volume as an \fBiSCSI\fR target. @@ -1782,9 +2400,173 @@ Connections: 0 .in -2 .sp +.sp .LP After the \fBiSCSI\fR target is created, set up the \fBiSCSI\fR initiator. For more information about the Solaris \fBiSCSI\fR initiator, see the Solaris Administration Guide: Devices and File Systems. +.LP +\fBExample 17 \fRPerforming a Rolling Snapshot +.sp +.LP +The following example shows how to maintain a history of snapshots with a consistent naming scheme. To keep a week's worth of snapshots, the user destroys the oldest snapshot, renames the remaining snapshots, and then creates a new snapshot, as follows: + +.sp +.in +2 +.nf +\fB# zfs destroy -r pool/users@7daysago +# zfs rename -r pool/users@6daysago @7daysago +# zfs rename -r pool/users@5daysago @6daysago +\&... +# zfs rename -r pool/users@yesterday @2daysago +# zfs rename -r pool/users@today @yesterday +# zfs snapshot -r pool/users@today\fR +.fi +.in -2 +.sp + +.LP +\fBExample 18 \fRSetting sharenfs Property Options on a ZFS File System +.sp +.LP +The following commands show how to set "sharenfs" property options to enable \fBrw\fR access for a set of \fBIP\fR addresses and to enable root access for system \fBneo\fR on the \fBtank/home\fR file system. + +.sp +.in +2 +.nf +\fB# zfs set sharenfs='rw=@123.123.0.0/16,root=neo' tank/home\fR + +.fi +.in -2 +.sp + +.sp +.LP +If you are using \fBDNS\fR for host name resolution, specify the fully qualified hostname. + +.LP +\fBExample 19 \fRDelegating ZFS Administration Permissions on a ZFS Dataset +.sp +.LP +The following example shows how to set permissions so that user "\fBcindys\fR" can create, destroy, mount and take snapshots on \fBtank/cindys\fR. The permissions on \fBtank/cindys\fR are also displayed. + +.sp +.in +2 +.nf +\fB# zfs allow cindys create,destroy,mount,snapshot tank/cindys +# zfs allow tank/cindys\fR +------------------------------------------------------------- +Local+Descendent permissions on (tank/cindys) + user cindys create,destroy,mount,snapshot +------------------------------------------------------------- + +.fi +.in -2 +.sp + +.sp +.LP +Because the \fBtank/cindys\fR mount point permission is set to 755 by default, user \fBcindys\fR will be unable to mount file systems under \fBtank/cindys\fR. Set an \fBACL\fR similar to the following syntax to provide mount point access: +.sp +.in +2 +.nf +# chmod A+user:cindys:add_subdirectory:allow /tank/cindys +.fi +.in -2 +.sp + +.LP +\fBExample 20 \fRDelegating Create Time Permissions on a ZFS Dataset +.sp +.LP +The following example shows how to grant anyone in the group \fBstaff\fR to create file systems in \fBtank/users\fR. This syntax also allows staff members to destroy their own file systems, but not destroy anyone else's file system. The permissions on \fBtank/users\fR are also displayed. + +.sp +.in +2 +.nf +\fB# zfs allow staff create,mount tank/users +# zfs allow -c destroy tank/users +# zfs allow tank/users\fR +------------------------------------------------------------- +Create time permissions on (tank/users) + create,destroy +Local+Descendent permissions on (tank/users) + group staff create,mount +------------------------------------------------------------- +.fi +.in -2 +.sp + +.LP +\fBExample 21 \fRDefining and Granting a Permission Set on a ZFS Dataset +.sp +.LP +The following example shows how to define and grant a permission set on the \fBtank/users\fR file system. The permissions on \fBtank/users\fR are also displayed. + +.sp +.in +2 +.nf +\fB# zfs allow -s @pset create,destroy,snapshot,mount tank/users +# zfs allow staff @pset tank/users +# zfs allow tank/users +------------------------------------------------------------- +Permission sets on (tank/users) + @pset create,destroy,mount,snapshot +Create time permissions on (tank/users) + create,destroy +Local+Descendent permissions on (tank/users) + group staff @pset,create,mount +-------------------------------------------------------------\fR +.fi +.in -2 +.sp + +.LP +\fBExample 22 \fRDelegating Property Permissions on a ZFS Dataset +.sp +.LP +The following example shows to grant the ability to set quotas and reservations on the \fBusers/home\fR file system. The permissions on \fBusers/home\fR are also displayed. + +.sp +.in +2 +.nf +\fB# zfs allow cindys quota,reservation users/home +# zfs allow users/home\fR +------------------------------------------------------------- +Local+Descendent permissions on (users/home) + user cindys quota,reservation +------------------------------------------------------------- +cindys% zfs set quota=10G users/home/marks +cindys% zfs get quota users/home/marks +NAME PROPERTY VALUE SOURCE +users/home/marks quota 10G local +.fi +.in -2 +.sp + +.LP +\fBExample 23 \fRRemoving ZFS Delegated Permissions on a ZFS Dataset +.sp +.LP +The following example shows how to remove the snapshot permission from the \fBstaff\fR group on the \fBtank/users\fR file system. The permissions on \fBtank/users\fR are also displayed. + +.sp +.in +2 +.nf +\fB# zfs unallow staff snapshot tank/users +# zfs allow tank/users\fR +------------------------------------------------------------- +Permission sets on (tank/users) + @pset create,destroy,mount,snapshot +Create time permissions on (tank/users) + create,destroy +Local+Descendent permissions on (tank/users) + group staff @pset,create,mount +------------------------------------------------------------- +.fi +.in -2 +.sp + .SH EXIT STATUS +.sp .LP The following exit values are returned: .sp @@ -1821,6 +2603,7 @@ Invalid command line options were specified. .RE .SH ATTRIBUTES +.sp .LP See \fBattributes\fR(5) for descriptions of the following attributes: .sp @@ -1835,9 +2618,13 @@ ATTRIBUTE TYPEATTRIBUTE VALUE _ AvailabilitySUNWzfsu _ -Interface StabilityEvolving +Interface StabilityCommitted .TE .SH SEE ALSO +.sp +.LP +\fBgzip\fR(1), \fBssh\fR(1), \fBmount\fR(1M), \fBshare\fR(1M), \fBsharemgr\fR(1M), \fBunshare\fR(1M), \fBzonecfg\fR(1M), \fBzpool\fR(1M), \fBchmod\fR(2), \fBstat\fR(2), \fBfsync\fR(3c), \fBdfstab\fR(4), \fBattributes\fR(5) +.sp .LP -\fBgzip\fR(1), \fBssh\fR(1), \fBmount\fR(1M), \fBshare\fR(1M), \fBunshare\fR(1M), \fBzonecfg\fR(1M), \fBzpool\fR(1M), \fBchmod\fR(2), \fBstat\fR(2), \fBfsync\fR(3c), \fBdfstab\fR(4), \fBattributes\fR(5) +For information about using the \fBZFS\fR web-based management tool and other \fBZFS\fR features, see the \fIZFS Administration Guide\fR. diff --git a/cddl/contrib/opensolaris/cmd/zfs/zfs_iter.c b/cddl/contrib/opensolaris/cmd/zfs/zfs_iter.c index eb6b8b1..a22370a 100644 --- a/cddl/contrib/opensolaris/cmd/zfs/zfs_iter.c +++ b/cddl/contrib/opensolaris/cmd/zfs/zfs_iter.c @@ -19,12 +19,10 @@ * CDDL HEADER END */ /* - * Copyright 2007 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ -#pragma ident "%Z%%M% %I% %E% SMI" - #include <libintl.h> #include <libuutil.h> #include <stddef.h> @@ -56,28 +54,43 @@ typedef struct zfs_node { typedef struct callback_data { uu_avl_t *cb_avl; - int cb_recurse; + int cb_flags; zfs_type_t cb_types; zfs_sort_column_t *cb_sortcol; - zfs_proplist_t **cb_proplist; + zprop_list_t **cb_proplist; } callback_data_t; uu_avl_pool_t *avl_pool; /* - * Called for each dataset. If the object the object is of an appropriate type, + * Include snaps if they were requested or if this a zfs list where types + * were not specified and the "listsnapshots" property is set on this pool. + */ +static int +zfs_include_snapshots(zfs_handle_t *zhp, callback_data_t *cb) +{ + zpool_handle_t *zph; + + if ((cb->cb_flags & ZFS_ITER_PROP_LISTSNAPS) == 0) + return (cb->cb_types & ZFS_TYPE_SNAPSHOT); + + zph = zfs_get_pool_handle(zhp); + return (zpool_get_prop_int(zph, ZPOOL_PROP_LISTSNAPS, NULL)); +} + +/* + * Called for each dataset. If the object is of an appropriate type, * add it to the avl tree and recurse over any children as necessary. */ -int +static int zfs_callback(zfs_handle_t *zhp, void *data) { callback_data_t *cb = data; int dontclose = 0; + int include_snaps = zfs_include_snapshots(zhp, cb); - /* - * If this object is of the appropriate type, add it to the AVL tree. - */ - if (zfs_get_type(zhp) & cb->cb_types) { + if ((zfs_get_type(zhp) & cb->cb_types) || + ((zfs_get_type(zhp) == ZFS_TYPE_SNAPSHOT) && include_snaps)) { uu_avl_index_t idx; zfs_node_t *node = safe_malloc(sizeof (zfs_node_t)); @@ -100,10 +113,12 @@ zfs_callback(zfs_handle_t *zhp, void *data) /* * Recurse if necessary. */ - if (cb->cb_recurse && (zfs_get_type(zhp) == ZFS_TYPE_FILESYSTEM || - (zfs_get_type(zhp) == ZFS_TYPE_VOLUME && (cb->cb_types & - ZFS_TYPE_SNAPSHOT)))) - (void) zfs_iter_children(zhp, zfs_callback, data); + if (cb->cb_flags & ZFS_ITER_RECURSE) { + if (zfs_get_type(zhp) == ZFS_TYPE_FILESYSTEM) + (void) zfs_iter_filesystems(zhp, zfs_callback, data); + if ((zfs_get_type(zhp) != ZFS_TYPE_SNAPSHOT) && include_snaps) + (void) zfs_iter_snapshots(zhp, zfs_callback, data); + } if (!dontclose) zfs_close(zhp); @@ -118,7 +133,7 @@ zfs_add_sort_column(zfs_sort_column_t **sc, const char *name, zfs_sort_column_t *col; zfs_prop_t prop; - if ((prop = zfs_name_to_prop(name)) == ZFS_PROP_INVAL && + if ((prop = zfs_name_to_prop(name)) == ZPROP_INVAL && !zfs_prop_user(name)) return (-1); @@ -126,7 +141,7 @@ zfs_add_sort_column(zfs_sort_column_t **sc, const char *name, col->sc_prop = prop; col->sc_reverse = reverse; - if (prop == ZFS_PROP_INVAL) { + if (prop == ZPROP_INVAL) { col->sc_user_prop = safe_malloc(strlen(name) + 1); (void) strcpy(col->sc_user_prop, name); } @@ -243,7 +258,7 @@ zfs_sort(const void *larg, const void *rarg, void *data) * Otherwise, we compare 'lnum' and 'rnum'. */ lstr = rstr = NULL; - if (psc->sc_prop == ZFS_PROP_INVAL) { + if (psc->sc_prop == ZPROP_INVAL) { nvlist_t *luser, *ruser; nvlist_t *lval, *rval; @@ -257,10 +272,10 @@ zfs_sort(const void *larg, const void *rarg, void *data) if (lvalid) verify(nvlist_lookup_string(lval, - ZFS_PROP_VALUE, &lstr) == 0); + ZPROP_VALUE, &lstr) == 0); if (rvalid) verify(nvlist_lookup_string(rval, - ZFS_PROP_VALUE, &rstr) == 0); + ZPROP_VALUE, &rstr) == 0); } else if (zfs_prop_is_string(psc->sc_prop)) { lvalid = (zfs_prop_get(l, psc->sc_prop, lbuf, @@ -293,7 +308,7 @@ zfs_sort(const void *larg, const void *rarg, void *data) if (lstr) ret = strcmp(lstr, rstr); - if (lnum < rnum) + else if (lnum < rnum) ret = -1; else if (lnum > rnum) ret = 1; @@ -309,9 +324,9 @@ zfs_sort(const void *larg, const void *rarg, void *data) } int -zfs_for_each(int argc, char **argv, boolean_t recurse, zfs_type_t types, - zfs_sort_column_t *sortcol, zfs_proplist_t **proplist, zfs_iter_f callback, - void *data, boolean_t args_can_be_paths) +zfs_for_each(int argc, char **argv, int flags, zfs_type_t types, + zfs_sort_column_t *sortcol, zprop_list_t **proplist, + zfs_iter_f callback, void *data) { callback_data_t cb; int ret = 0; @@ -328,7 +343,7 @@ zfs_for_each(int argc, char **argv, boolean_t recurse, zfs_type_t types, } cb.cb_sortcol = sortcol; - cb.cb_recurse = recurse; + cb.cb_flags = flags; cb.cb_proplist = proplist; cb.cb_types = types; if ((cb.cb_avl = uu_avl_create(avl_pool, NULL, UU_DEFAULT)) == NULL) { @@ -341,7 +356,7 @@ zfs_for_each(int argc, char **argv, boolean_t recurse, zfs_type_t types, /* * If given no arguments, iterate over all datasets. */ - cb.cb_recurse = 1; + cb.cb_flags |= ZFS_ITER_RECURSE; ret = zfs_iter_root(g_zfs, zfs_callback, &cb); } else { int i; @@ -354,14 +369,14 @@ zfs_for_each(int argc, char **argv, boolean_t recurse, zfs_type_t types, * can take volumes as well. */ argtype = types; - if (recurse) { + if (flags & ZFS_ITER_RECURSE) { argtype |= ZFS_TYPE_FILESYSTEM; if (types & ZFS_TYPE_SNAPSHOT) argtype |= ZFS_TYPE_VOLUME; } for (i = 0; i < argc; i++) { - if (args_can_be_paths) { + if (flags & ZFS_ITER_ARGS_CAN_BE_PATHS) { zhp = zfs_path_to_zhandle(g_zfs, argv[i], argtype); } else { diff --git a/cddl/contrib/opensolaris/cmd/zfs/zfs_iter.h b/cddl/contrib/opensolaris/cmd/zfs/zfs_iter.h index 1f0ce8e..76a1108 100644 --- a/cddl/contrib/opensolaris/cmd/zfs/zfs_iter.h +++ b/cddl/contrib/opensolaris/cmd/zfs/zfs_iter.h @@ -19,15 +19,13 @@ * CDDL HEADER END */ /* - * Copyright 2007 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ #ifndef ZFS_ITER_H #define ZFS_ITER_H -#pragma ident "%Z%%M% %I% %E% SMI" - #ifdef __cplusplus extern "C" { #endif @@ -40,8 +38,12 @@ typedef struct zfs_sort_column { boolean_t sc_reverse; } zfs_sort_column_t; -int zfs_for_each(int, char **, boolean_t, zfs_type_t, zfs_sort_column_t *, - zfs_proplist_t **, zfs_iter_f, void *, boolean_t); +#define ZFS_ITER_RECURSE (1 << 0) +#define ZFS_ITER_ARGS_CAN_BE_PATHS (1 << 1) +#define ZFS_ITER_PROP_LISTSNAPS (1 << 2) + +int zfs_for_each(int, char **, int options, zfs_type_t, + zfs_sort_column_t *, zprop_list_t **, zfs_iter_f, void *); int zfs_add_sort_column(zfs_sort_column_t **, const char *, boolean_t); void zfs_free_sort_columns(zfs_sort_column_t *); diff --git a/cddl/contrib/opensolaris/cmd/zfs/zfs_main.c b/cddl/contrib/opensolaris/cmd/zfs/zfs_main.c index de15b00..6f5f92e 100644 --- a/cddl/contrib/opensolaris/cmd/zfs/zfs_main.c +++ b/cddl/contrib/opensolaris/cmd/zfs/zfs_main.c @@ -20,18 +20,17 @@ */ /* - * Copyright 2007 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ -#pragma ident "%Z%%M% %I% %E% SMI" - #include <assert.h> #include <ctype.h> #include <errno.h> #include <libgen.h> #include <libintl.h> #include <libuutil.h> +#include <libnvpair.h> #include <locale.h> #include <stddef.h> #include <stdio.h> @@ -44,8 +43,10 @@ #include <sys/mnttab.h> #include <sys/mount.h> #include <sys/stat.h> +#include <sys/avl.h> #include <libzfs.h> +#include <libuutil.h> #include "zfs_iter.h" #include "zfs_util.h" @@ -53,6 +54,7 @@ libzfs_handle_t *g_zfs; static FILE *mnttab_file; +static char history_str[HIS_MAX_RECORD_LEN]; static int zfs_do_clone(int argc, char **argv); static int zfs_do_create(int argc, char **argv); @@ -64,6 +66,7 @@ static int zfs_do_mount(int argc, char **argv); static int zfs_do_rename(int argc, char **argv); static int zfs_do_rollback(int argc, char **argv); static int zfs_do_set(int argc, char **argv); +static int zfs_do_upgrade(int argc, char **argv); static int zfs_do_snapshot(int argc, char **argv); static int zfs_do_unmount(int argc, char **argv); static int zfs_do_share(int argc, char **argv); @@ -71,13 +74,16 @@ static int zfs_do_unshare(int argc, char **argv); static int zfs_do_send(int argc, char **argv); static int zfs_do_receive(int argc, char **argv); static int zfs_do_promote(int argc, char **argv); +static int zfs_do_allow(int argc, char **argv); +static int zfs_do_unallow(int argc, char **argv); static int zfs_do_jail(int argc, char **argv); static int zfs_do_unjail(int argc, char **argv); /* - * These libumem hooks provide a reasonable set of defaults for the allocator's - * debugging facilities. + * Enable a reasonable set of defaults for libumem debugging on DEBUG builds. */ + +#ifdef DEBUG const char * _umem_debug_init(void) { @@ -89,6 +95,7 @@ _umem_logging_init(void) { return ("fail,contents"); /* $UMEM_LOGGING setting */ } +#endif typedef enum { HELP_CLONE, @@ -96,6 +103,7 @@ typedef enum { HELP_DESTROY, HELP_GET, HELP_INHERIT, + HELP_UPGRADE, HELP_JAIL, HELP_UNJAIL, HELP_LIST, @@ -109,7 +117,9 @@ typedef enum { HELP_SHARE, HELP_SNAPSHOT, HELP_UNMOUNT, - HELP_UNSHARE + HELP_UNSHARE, + HELP_ALLOW, + HELP_UNALLOW } zfs_help_t; typedef struct zfs_command { @@ -142,18 +152,20 @@ static zfs_command_t command_table[] = { { "set", zfs_do_set, HELP_SET }, { "get", zfs_do_get, HELP_GET }, { "inherit", zfs_do_inherit, HELP_INHERIT }, + { "upgrade", zfs_do_upgrade, HELP_UPGRADE }, { NULL }, { "mount", zfs_do_mount, HELP_MOUNT }, - { NULL }, { "unmount", zfs_do_unmount, HELP_UNMOUNT }, - { NULL }, { "share", zfs_do_share, HELP_SHARE }, - { NULL }, { "unshare", zfs_do_unshare, HELP_UNSHARE }, { NULL }, { "send", zfs_do_send, HELP_SEND }, { "receive", zfs_do_receive, HELP_RECEIVE }, { NULL }, + { "allow", zfs_do_allow, HELP_ALLOW }, + { NULL }, + { "unallow", zfs_do_unallow, HELP_UNALLOW }, + { NULL }, { "jail", zfs_do_jail, HELP_JAIL }, { "unjail", zfs_do_unjail, HELP_UNJAIL }, }; @@ -167,39 +179,41 @@ get_usage(zfs_help_t idx) { switch (idx) { case HELP_CLONE: - return (gettext("\tclone <snapshot> <filesystem|volume>\n")); + return (gettext("\tclone [-p] [-o property=value] ... " + "<snapshot> <filesystem|volume>\n")); case HELP_CREATE: - return (gettext("\tcreate [[-o property=value] ... ] " + return (gettext("\tcreate [-p] [-o property=value] ... " "<filesystem>\n" - "\tcreate [-s] [-b blocksize] [[-o property=value] ...]\n" - "\t -V <size> <volume>\n")); + "\tcreate [-ps] [-b blocksize] [-o property=value] ... " + "-V <size> <volume>\n")); case HELP_DESTROY: return (gettext("\tdestroy [-rRf] " "<filesystem|volume|snapshot>\n")); case HELP_GET: - return (gettext("\tget [-rHp] [-o field[,field]...] " - "[-s source[,source]...]\n" - "\t <all | property[,property]...> " + return (gettext("\tget [-rHp] [-o field[,...]] " + "[-s source[,...]]\n" + "\t <\"all\" | property[,...]> " "[filesystem|volume|snapshot] ...\n")); case HELP_INHERIT: return (gettext("\tinherit [-r] <property> " - "<filesystem|volume> ...\n")); + "<filesystem|volume|snapshot> ...\n")); + case HELP_UPGRADE: + return (gettext("\tupgrade [-v]\n" + "\tupgrade [-r] [-V version] <-a | filesystem ...>\n")); case HELP_JAIL: return (gettext("\tjail <jailid> <filesystem>\n")); case HELP_UNJAIL: return (gettext("\tunjail <jailid> <filesystem>\n")); case HELP_LIST: - return (gettext("\tlist [-rH] [-o property[,property]...] " - "[-t type[,type]...]\n" - "\t [-s property [-s property]...]" - " [-S property [-S property]...]\n" - "\t [filesystem|volume|snapshot] ...\n")); + return (gettext("\tlist [-rH] [-o property[,...]] " + "[-t type[,...]] [-s property] ...\n" + "\t [-S property] ... " + "[filesystem|volume|snapshot] ...\n")); case HELP_MOUNT: return (gettext("\tmount\n" - "\tmount [-o opts] [-O] -a\n" - "\tmount [-o opts] [-O] <filesystem>\n")); + "\tmount [-vO] [-o opts] <-a | filesystem>\n")); case HELP_PROMOTE: - return (gettext("\tpromote <clone filesystem>\n")); + return (gettext("\tpromote <clone-filesystem>\n")); case HELP_RECEIVE: return (gettext("\treceive [-vnF] <filesystem|volume|" "snapshot>\n" @@ -207,26 +221,45 @@ get_usage(zfs_help_t idx) case HELP_RENAME: return (gettext("\trename <filesystem|volume|snapshot> " "<filesystem|volume|snapshot>\n" + "\trename -p <filesystem|volume> <filesystem|volume>\n" "\trename -r <snapshot> <snapshot>")); case HELP_ROLLBACK: return (gettext("\trollback [-rRf] <snapshot>\n")); case HELP_SEND: - return (gettext("\tsend [-i <snapshot>] <snapshot>\n")); + return (gettext("\tsend [-R] [-[iI] snapshot] <snapshot>\n")); case HELP_SET: return (gettext("\tset <property=value> " - "<filesystem|volume> ...\n")); + "<filesystem|volume|snapshot> ...\n")); case HELP_SHARE: - return (gettext("\tshare -a\n" - "\tshare <filesystem>\n")); + return (gettext("\tshare <-a | filesystem>\n")); case HELP_SNAPSHOT: - return (gettext("\tsnapshot [-r] " - "<filesystem@name|volume@name>\n")); + return (gettext("\tsnapshot [-r] [-o property=value] ... " + "<filesystem@snapname|volume@snapname>\n")); case HELP_UNMOUNT: - return (gettext("\tunmount [-f] -a\n" - "\tunmount [-f] <filesystem|mountpoint>\n")); + return (gettext("\tunmount [-f] " + "<-a | filesystem|mountpoint>\n")); case HELP_UNSHARE: - return (gettext("\tunshare [-f] -a\n" - "\tunshare [-f] <filesystem|mountpoint>\n")); + return (gettext("\tunshare [-f] " + "<-a | filesystem|mountpoint>\n")); + case HELP_ALLOW: + return (gettext("\tallow [-ldug] " + "<\"everyone\"|user|group>[,...] <perm|@setname>[,...]\n" + "\t <filesystem|volume>\n" + "\tallow [-ld] -e <perm|@setname>[,...] " + "<filesystem|volume>\n" + "\tallow -c <perm|@setname>[,...] <filesystem|volume>\n" + "\tallow -s @setname <perm|@setname>[,...] " + "<filesystem|volume>\n")); + case HELP_UNALLOW: + return (gettext("\tunallow [-rldug] " + "<\"everyone\"|user|group>[,...]\n" + "\t [<perm|@setname>[,...]] <filesystem|volume>\n" + "\tunallow [-rld] -e [<perm|@setname>[,...]] " + "<filesystem|volume>\n" + "\tunallow [-r] -c [<perm|@setname>[,...]] " + "<filesystem|volume>\n" + "\tunallow [-r] -s @setname [<perm|@setname>[,...]] " + "<filesystem|volume>\n")); } abort(); @@ -250,20 +283,20 @@ safe_malloc(size_t size) } /* - * Callback routinue that will print out information for each of the + * Callback routine that will print out information for each of * the properties. */ -static zfs_prop_t -usage_prop_cb(zfs_prop_t prop, void *cb) +static int +usage_prop_cb(int prop, void *cb) { FILE *fp = cb; - (void) fprintf(fp, "\t%-13s ", zfs_prop_to_name(prop)); + (void) fprintf(fp, "\t%-15s ", zfs_prop_to_name(prop)); if (zfs_prop_readonly(prop)) - (void) fprintf(fp, " NO "); + (void) fprintf(fp, " NO "); else - (void) fprintf(fp, " YES "); + (void) fprintf(fp, "YES "); if (zfs_prop_inheritable(prop)) (void) fprintf(fp, " YES "); @@ -275,7 +308,7 @@ usage_prop_cb(zfs_prop_t prop, void *cb) else (void) fprintf(fp, "%s\n", zfs_prop_values(prop)); - return (ZFS_PROP_CONT); + return (ZPROP_CONT); } /* @@ -288,6 +321,7 @@ usage(boolean_t requested) { int i; boolean_t show_properties = B_FALSE; + boolean_t show_permissions = B_FALSE; FILE *fp = requested ? stdout : stderr; if (current_command == NULL) { @@ -318,29 +352,46 @@ usage(boolean_t requested) strcmp(current_command->name, "list") == 0)) show_properties = B_TRUE; + if (current_command != NULL && + (strcmp(current_command->name, "allow") == 0 || + strcmp(current_command->name, "unallow") == 0)) + show_permissions = B_TRUE; + if (show_properties) { (void) fprintf(fp, gettext("\nThe following properties are supported:\n")); - (void) fprintf(fp, "\n\t%-13s %s %s %s\n\n", + (void) fprintf(fp, "\n\t%-14s %s %s %s\n\n", "PROPERTY", "EDIT", "INHERIT", "VALUES"); /* Iterate over all properties */ - (void) zfs_prop_iter(usage_prop_cb, fp, B_FALSE); + (void) zprop_iter(usage_prop_cb, fp, B_FALSE, B_TRUE, + ZFS_TYPE_DATASET); (void) fprintf(fp, gettext("\nSizes are specified in bytes " "with standard units such as K, M, G, etc.\n")); - (void) fprintf(fp, gettext("\n\nUser-defined properties can " + (void) fprintf(fp, gettext("\nUser-defined properties can " "be specified by using a name containing a colon (:).\n")); + + } else if (show_permissions) { + (void) fprintf(fp, + gettext("\nThe following permissions are supported:\n")); + + zfs_deleg_permissions(); } else { /* * TRANSLATION NOTE: * "zfs set|get" must not be localised this is the * command name and arguments. */ + (void) fprintf(fp, gettext("\nFor the property list, run: zfs set|get\n")); + + (void) fprintf(fp, + gettext("\nFor the delegated permission list, run:" + " zfs allow|unallow\n")); } /* @@ -354,69 +405,143 @@ usage(boolean_t requested) exit(requested ? 0 : 2); } +static int +parseprop(nvlist_t *props) +{ + char *propname = optarg; + char *propval, *strval; + + if ((propval = strchr(propname, '=')) == NULL) { + (void) fprintf(stderr, gettext("missing " + "'=' for -o option\n")); + return (-1); + } + *propval = '\0'; + propval++; + if (nvlist_lookup_string(props, propname, &strval) == 0) { + (void) fprintf(stderr, gettext("property '%s' " + "specified multiple times\n"), propname); + return (-1); + } + if (nvlist_add_string(props, propname, propval) != 0) { + (void) fprintf(stderr, gettext("internal " + "error: out of memory\n")); + return (-1); + } + return (0); + +} + /* - * zfs clone <fs, snap, vol> fs + * zfs clone [-p] [-o prop=value] ... <snap> <fs | vol> * * Given an existing dataset, create a writable copy whose initial contents * are the same as the source. The newly created dataset maintains a * dependency on the original; the original cannot be destroyed so long as * the clone exists. + * + * The '-p' flag creates all the non-existing ancestors of the target first. */ static int zfs_do_clone(int argc, char **argv) { - zfs_handle_t *zhp; + zfs_handle_t *zhp = NULL; + boolean_t parents = B_FALSE; + nvlist_t *props; int ret; + int c; + + if (nvlist_alloc(&props, NV_UNIQUE_NAME, 0) != 0) { + (void) fprintf(stderr, gettext("internal error: " + "out of memory\n")); + return (1); + } /* check options */ - if (argc > 1 && argv[1][0] == '-') { - (void) fprintf(stderr, gettext("invalid option '%c'\n"), - argv[1][1]); - usage(B_FALSE); + while ((c = getopt(argc, argv, "o:p")) != -1) { + switch (c) { + case 'o': + if (parseprop(props)) + return (1); + break; + case 'p': + parents = B_TRUE; + break; + case '?': + (void) fprintf(stderr, gettext("invalid option '%c'\n"), + optopt); + goto usage; + } } + argc -= optind; + argv += optind; + /* check number of arguments */ - if (argc < 2) { + if (argc < 1) { (void) fprintf(stderr, gettext("missing source dataset " "argument\n")); - usage(B_FALSE); + goto usage; } - if (argc < 3) { + if (argc < 2) { (void) fprintf(stderr, gettext("missing target dataset " "argument\n")); - usage(B_FALSE); + goto usage; } - if (argc > 3) { + if (argc > 2) { (void) fprintf(stderr, gettext("too many arguments\n")); - usage(B_FALSE); + goto usage; } /* open the source dataset */ - if ((zhp = zfs_open(g_zfs, argv[1], ZFS_TYPE_SNAPSHOT)) == NULL) + if ((zhp = zfs_open(g_zfs, argv[0], ZFS_TYPE_SNAPSHOT)) == NULL) return (1); + if (parents && zfs_name_valid(argv[1], ZFS_TYPE_FILESYSTEM | + ZFS_TYPE_VOLUME)) { + /* + * Now create the ancestors of the target dataset. If the + * target already exists and '-p' option was used we should not + * complain. + */ + if (zfs_dataset_exists(g_zfs, argv[1], ZFS_TYPE_FILESYSTEM | + ZFS_TYPE_VOLUME)) + return (0); + if (zfs_create_ancestors(g_zfs, argv[1]) != 0) + return (1); + } + /* pass to libzfs */ - ret = zfs_clone(zhp, argv[2], NULL); + ret = zfs_clone(zhp, argv[1], props); /* create the mountpoint if necessary */ if (ret == 0) { - zfs_handle_t *clone = zfs_open(g_zfs, argv[2], ZFS_TYPE_ANY); + zfs_handle_t *clone; + + clone = zfs_open(g_zfs, argv[1], ZFS_TYPE_DATASET); if (clone != NULL) { if ((ret = zfs_mount(clone, NULL, 0)) == 0) ret = zfs_share(clone); zfs_close(clone); } - zpool_log_history(g_zfs, argc, argv, argv[2], B_FALSE, B_FALSE); } zfs_close(zhp); + nvlist_free(props); - return (ret == 0 ? 0 : 1); + return (!!ret); + +usage: + if (zhp) + zfs_close(zhp); + nvlist_free(props); + usage(B_FALSE); + return (-1); } /* - * zfs create [-o prop=value] ... fs - * zfs create [-s] [-b blocksize] [-o prop=value] ... -V vol size + * zfs create [-p] [-o prop=value] ... fs + * zfs create [-ps] [-b blocksize] [-o prop=value] ... -V vol size * * Create a new dataset. This command can be used to create filesystems * and volumes. Snapshot creation is handled by 'zfs snapshot'. @@ -424,7 +549,10 @@ zfs_do_clone(int argc, char **argv) * * The '-s' flag applies only to volumes, and indicates that we should not try * to set the reservation for this volume. By default we set a reservation - * equal to the size for any volume. + * equal to the size for any volume. For pools with SPA_VERSION >= + * SPA_VERSION_REFRESERVATION, we set a refreservation instead. + * + * The '-p' flag creates all the non-existing ancestors of the target first. */ static int zfs_do_create(int argc, char **argv) @@ -434,12 +562,12 @@ zfs_do_create(int argc, char **argv) uint64_t volsize; int c; boolean_t noreserve = B_FALSE; + boolean_t bflag = B_FALSE; + boolean_t parents = B_FALSE; int ret = 1; - nvlist_t *props = NULL; + nvlist_t *props; uint64_t intval; - char *propname; - char *propval = NULL; - char *strval; + int canmount; if (nvlist_alloc(&props, NV_UNIQUE_NAME, 0) != 0) { (void) fprintf(stderr, gettext("internal error: " @@ -448,7 +576,7 @@ zfs_do_create(int argc, char **argv) } /* check options */ - while ((c = getopt(argc, argv, ":V:b:so:")) != -1) { + while ((c = getopt(argc, argv, ":V:b:so:p")) != -1) { switch (c) { case 'V': type = ZFS_TYPE_VOLUME; @@ -468,7 +596,11 @@ zfs_do_create(int argc, char **argv) } volsize = intval; break; + case 'p': + parents = B_TRUE; + break; case 'b': + bflag = B_TRUE; if (zfs_nicestrtonum(g_zfs, optarg, &intval) != 0) { (void) fprintf(stderr, gettext("bad volume " "block size '%s': %s\n"), optarg, @@ -485,25 +617,8 @@ zfs_do_create(int argc, char **argv) } break; case 'o': - propname = optarg; - if ((propval = strchr(propname, '=')) == NULL) { - (void) fprintf(stderr, gettext("missing " - "'=' for -o option\n")); - goto error; - } - *propval = '\0'; - propval++; - if (nvlist_lookup_string(props, propname, - &strval) == 0) { - (void) fprintf(stderr, gettext("property '%s' " - "specified multiple times\n"), propname); - goto error; - } - if (nvlist_add_string(props, propname, propval) != 0) { - (void) fprintf(stderr, gettext("internal " - "error: out of memory\n")); + if (parseprop(props)) goto error; - } break; case 's': noreserve = B_TRUE; @@ -520,9 +635,9 @@ zfs_do_create(int argc, char **argv) } } - if (noreserve && type != ZFS_TYPE_VOLUME) { - (void) fprintf(stderr, gettext("'-s' can only be used when " - "creating a volume\n")); + if ((bflag || noreserve) && type != ZFS_TYPE_VOLUME) { + (void) fprintf(stderr, gettext("'-s' and '-b' can only be " + "used when creating a volume\n")); goto badusage; } @@ -540,46 +655,83 @@ zfs_do_create(int argc, char **argv) goto badusage; } - if (type == ZFS_TYPE_VOLUME && !noreserve && - nvlist_lookup_string(props, zfs_prop_to_name(ZFS_PROP_RESERVATION), - &strval) != 0) { - if (nvlist_add_uint64(props, - zfs_prop_to_name(ZFS_PROP_RESERVATION), - volsize) != 0) { - (void) fprintf(stderr, gettext("internal " - "error: out of memory\n")); - nvlist_free(props); - return (1); + if (type == ZFS_TYPE_VOLUME && !noreserve) { + zpool_handle_t *zpool_handle; + uint64_t spa_version; + char *p; + zfs_prop_t resv_prop; + char *strval; + + if (p = strchr(argv[0], '/')) + *p = '\0'; + zpool_handle = zpool_open(g_zfs, argv[0]); + if (p != NULL) + *p = '/'; + if (zpool_handle == NULL) + goto error; + spa_version = zpool_get_prop_int(zpool_handle, + ZPOOL_PROP_VERSION, NULL); + zpool_close(zpool_handle); + if (spa_version >= SPA_VERSION_REFRESERVATION) + resv_prop = ZFS_PROP_REFRESERVATION; + else + resv_prop = ZFS_PROP_RESERVATION; + + if (nvlist_lookup_string(props, zfs_prop_to_name(resv_prop), + &strval) != 0) { + if (nvlist_add_uint64(props, + zfs_prop_to_name(resv_prop), volsize) != 0) { + (void) fprintf(stderr, gettext("internal " + "error: out of memory\n")); + nvlist_free(props); + return (1); + } } } + if (parents && zfs_name_valid(argv[0], type)) { + /* + * Now create the ancestors of target dataset. If the target + * already exists and '-p' option was used we should not + * complain. + */ + if (zfs_dataset_exists(g_zfs, argv[0], type)) { + ret = 0; + goto error; + } + if (zfs_create_ancestors(g_zfs, argv[0]) != 0) + goto error; + } + /* pass to libzfs */ if (zfs_create(g_zfs, argv[0], type, props) != 0) goto error; - if (propval != NULL) - *(propval - 1) = '='; - zpool_log_history(g_zfs, argc + optind, argv - optind, argv[0], - B_FALSE, B_FALSE); - - if ((zhp = zfs_open(g_zfs, argv[0], ZFS_TYPE_ANY)) == NULL) + if ((zhp = zfs_open(g_zfs, argv[0], ZFS_TYPE_DATASET)) == NULL) goto error; + /* + * if the user doesn't want the dataset automatically mounted, + * then skip the mount/share step + */ + + canmount = zfs_prop_get_int(zhp, ZFS_PROP_CANMOUNT); /* * Mount and/or share the new filesystem as appropriate. We provide a * verbose error message to let the user know that their filesystem was * in fact created, even if we failed to mount or share it. */ - if (zfs_mount(zhp, NULL, 0) != 0) { - (void) fprintf(stderr, gettext("filesystem successfully " - "created, but not mounted\n")); - ret = 1; - } else if (zfs_share(zhp) != 0) { - (void) fprintf(stderr, gettext("filesystem successfully " - "created, but not shared\n")); - ret = 1; - } else { - ret = 0; + ret = 0; + if (canmount == ZFS_CANMOUNT_ON) { + if (zfs_mount(zhp, NULL, 0) != 0) { + (void) fprintf(stderr, gettext("filesystem " + "successfully created, but not mounted\n")); + ret = 1; + } else if (zfs_share(zhp) != 0) { + (void) fprintf(stderr, gettext("filesystem " + "successfully created, but not shared\n")); + ret = 1; + } } error: @@ -789,7 +941,7 @@ zfs_do_destroy(int argc, char **argv) int ret; *cp = '\0'; - if ((zhp = zfs_open(g_zfs, argv[0], ZFS_TYPE_ANY)) == NULL) + if ((zhp = zfs_open(g_zfs, argv[0], ZFS_TYPE_DATASET)) == NULL) return (1); *cp = '@'; cp++; @@ -807,16 +959,13 @@ zfs_do_destroy(int argc, char **argv) if (ret) { (void) fprintf(stderr, gettext("no snapshots destroyed\n")); - } else { - zpool_log_history(g_zfs, argc + optind, argv - optind, - argv[0], B_FALSE, B_FALSE); } return (ret != 0); } /* Open the given dataset */ - if ((zhp = zfs_open(g_zfs, argv[0], ZFS_TYPE_ANY)) == NULL) + if ((zhp = zfs_open(g_zfs, argv[0], ZFS_TYPE_DATASET)) == NULL) return (1); cb.cb_target = zhp; @@ -849,7 +998,6 @@ zfs_do_destroy(int argc, char **argv) return (1); } - if (cb.cb_error || zfs_iter_dependents(zhp, B_FALSE, destroy_callback, &cb) != 0) { zfs_close(zhp); @@ -860,11 +1008,10 @@ zfs_do_destroy(int argc, char **argv) * Do the real thing. The callback will close the handle regardless of * whether it succeeds or not. */ + if (destroy_callback(zhp, &cb) != 0) return (1); - zpool_log_history(g_zfs, argc + optind, argv - optind, argv[0], - B_FALSE, B_FALSE); return (0); } @@ -894,11 +1041,11 @@ static int get_callback(zfs_handle_t *zhp, void *data) { char buf[ZFS_MAXPROPLEN]; - zfs_source_t sourcetype; + zprop_source_t sourcetype; char source[ZFS_MAXNAMELEN]; - libzfs_get_cbdata_t *cbp = data; + zprop_get_cbdata_t *cbp = data; nvlist_t *userprop = zfs_get_user_props(zhp); - zfs_proplist_t *pl = cbp->cb_proplist; + zprop_list_t *pl = cbp->cb_proplist; nvlist_t *propval; char *strval; char *sourceval; @@ -912,7 +1059,7 @@ get_callback(zfs_handle_t *zhp, void *data) pl == cbp->cb_proplist) continue; - if (pl->pl_prop != ZFS_PROP_INVAL) { + if (pl->pl_prop != ZPROP_INVAL) { if (zfs_prop_get(zhp, pl->pl_prop, buf, sizeof (buf), &sourcetype, source, sizeof (source), @@ -920,17 +1067,17 @@ get_callback(zfs_handle_t *zhp, void *data) if (pl->pl_all) continue; if (!zfs_prop_valid_for_type(pl->pl_prop, - ZFS_TYPE_ANY)) { + ZFS_TYPE_DATASET)) { (void) fprintf(stderr, gettext("No such property '%s'\n"), zfs_prop_to_name(pl->pl_prop)); continue; } - sourcetype = ZFS_SRC_NONE; + sourcetype = ZPROP_SRC_NONE; (void) strlcpy(buf, "-", sizeof (buf)); } - libzfs_print_one_property(zfs_get_name(zhp), cbp, + zprop_print_one_property(zfs_get_name(zhp), cbp, zfs_prop_to_name(pl->pl_prop), buf, sourcetype, source); } else { @@ -938,25 +1085,25 @@ get_callback(zfs_handle_t *zhp, void *data) pl->pl_user_prop, &propval) != 0) { if (pl->pl_all) continue; - sourcetype = ZFS_SRC_NONE; + sourcetype = ZPROP_SRC_NONE; strval = "-"; } else { verify(nvlist_lookup_string(propval, - ZFS_PROP_VALUE, &strval) == 0); + ZPROP_VALUE, &strval) == 0); verify(nvlist_lookup_string(propval, - ZFS_PROP_SOURCE, &sourceval) == 0); + ZPROP_SOURCE, &sourceval) == 0); if (strcmp(sourceval, zfs_get_name(zhp)) == 0) { - sourcetype = ZFS_SRC_LOCAL; + sourcetype = ZPROP_SRC_LOCAL; } else { - sourcetype = ZFS_SRC_INHERITED; + sourcetype = ZPROP_SRC_INHERITED; (void) strlcpy(source, sourceval, sizeof (source)); } } - libzfs_print_one_property(zfs_get_name(zhp), cbp, + zprop_print_one_property(zfs_get_name(zhp), cbp, pl->pl_user_prop, strval, sourcetype, source); } @@ -968,21 +1115,21 @@ get_callback(zfs_handle_t *zhp, void *data) static int zfs_do_get(int argc, char **argv) { - libzfs_get_cbdata_t cb = { 0 }; - boolean_t recurse = B_FALSE; - int i, c; + zprop_get_cbdata_t cb = { 0 }; + int i, c, flags = 0; char *value, *fields; int ret; - zfs_proplist_t fake_name = { 0 }; + zprop_list_t fake_name = { 0 }; /* * Set up default columns and sources. */ - cb.cb_sources = ZFS_SRC_ALL; + cb.cb_sources = ZPROP_SRC_ALL; cb.cb_columns[0] = GET_COL_NAME; cb.cb_columns[1] = GET_COL_PROPERTY; cb.cb_columns[2] = GET_COL_VALUE; cb.cb_columns[3] = GET_COL_SOURCE; + cb.cb_type = ZFS_TYPE_DATASET; /* check options */ while ((c = getopt(argc, argv, ":o:s:rHp")) != -1) { @@ -991,7 +1138,7 @@ zfs_do_get(int argc, char **argv) cb.cb_literal = B_TRUE; break; case 'r': - recurse = B_TRUE; + flags |= ZFS_ITER_RECURSE; break; case 'H': cb.cb_scripted = B_TRUE; @@ -1053,19 +1200,19 @@ zfs_do_get(int argc, char **argv) switch (getsubopt(&optarg, source_subopts, &value)) { case 0: - cb.cb_sources |= ZFS_SRC_LOCAL; + cb.cb_sources |= ZPROP_SRC_LOCAL; break; case 1: - cb.cb_sources |= ZFS_SRC_DEFAULT; + cb.cb_sources |= ZPROP_SRC_DEFAULT; break; case 2: - cb.cb_sources |= ZFS_SRC_INHERITED; + cb.cb_sources |= ZPROP_SRC_INHERITED; break; case 3: - cb.cb_sources |= ZFS_SRC_TEMPORARY; + cb.cb_sources |= ZPROP_SRC_TEMPORARY; break; case 4: - cb.cb_sources |= ZFS_SRC_NONE; + cb.cb_sources |= ZPROP_SRC_NONE; break; default: (void) fprintf(stderr, @@ -1094,7 +1241,8 @@ zfs_do_get(int argc, char **argv) fields = argv[0]; - if (zfs_get_proplist(g_zfs, fields, &cb.cb_proplist) != 0) + if (zprop_get_list(g_zfs, fields, &cb.cb_proplist, ZFS_TYPE_DATASET) + != 0) usage(B_FALSE); argc--; @@ -1118,13 +1266,13 @@ zfs_do_get(int argc, char **argv) cb.cb_first = B_TRUE; /* run for each object */ - ret = zfs_for_each(argc, argv, recurse, ZFS_TYPE_ANY, NULL, - &cb.cb_proplist, get_callback, &cb, B_FALSE); + ret = zfs_for_each(argc, argv, flags, ZFS_TYPE_DATASET, NULL, + &cb.cb_proplist, get_callback, &cb); if (cb.cb_proplist == &fake_name) - zfs_free_proplist(fake_name.pl_next); + zprop_free_list(fake_name.pl_next); else - zfs_free_proplist(cb.cb_proplist); + zprop_free_list(cb.cb_proplist); return (ret); } @@ -1140,37 +1288,46 @@ zfs_do_get(int argc, char **argv) * useful for setting a property on a hierarchy-wide basis, regardless of any * local modifications for each dataset. */ -typedef struct inherit_cbdata { - char *cb_propname; - boolean_t cb_any_successful; -} inherit_cbdata_t; static int -inherit_callback(zfs_handle_t *zhp, void *data) +inherit_recurse_cb(zfs_handle_t *zhp, void *data) { - inherit_cbdata_t *cbp = data; - int ret; + char *propname = data; + zfs_prop_t prop = zfs_name_to_prop(propname); - ret = zfs_prop_inherit(zhp, cbp->cb_propname); - if (ret == 0) - cbp->cb_any_successful = B_TRUE; - return (ret != 0); + /* + * If we're doing it recursively, then ignore properties that + * are not valid for this type of dataset. + */ + if (prop != ZPROP_INVAL && + !zfs_prop_valid_for_type(prop, zfs_get_type(zhp))) + return (0); + + return (zfs_prop_inherit(zhp, propname) != 0); +} + +static int +inherit_cb(zfs_handle_t *zhp, void *data) +{ + char *propname = data; + + return (zfs_prop_inherit(zhp, propname) != 0); } static int zfs_do_inherit(int argc, char **argv) { - boolean_t recurse = B_FALSE; int c; zfs_prop_t prop; - inherit_cbdata_t cb; + char *propname; int ret; + int flags = 0; /* check options */ while ((c = getopt(argc, argv, "r")) != -1) { switch (c) { case 'r': - recurse = B_TRUE; + flags |= ZFS_ITER_RECURSE; break; case '?': default: @@ -1193,42 +1350,265 @@ zfs_do_inherit(int argc, char **argv) usage(B_FALSE); } - cb.cb_propname = argv[0]; + propname = argv[0]; argc--; argv++; - if ((prop = zfs_name_to_prop(cb.cb_propname)) != ZFS_PROP_INVAL) { + if ((prop = zfs_name_to_prop(propname)) != ZPROP_INVAL) { if (zfs_prop_readonly(prop)) { (void) fprintf(stderr, gettext( "%s property is read-only\n"), - cb.cb_propname); + propname); return (1); } if (!zfs_prop_inheritable(prop)) { (void) fprintf(stderr, gettext("'%s' property cannot " - "be inherited\n"), cb.cb_propname); + "be inherited\n"), propname); if (prop == ZFS_PROP_QUOTA || - prop == ZFS_PROP_RESERVATION) + prop == ZFS_PROP_RESERVATION || + prop == ZFS_PROP_REFQUOTA || + prop == ZFS_PROP_REFRESERVATION) (void) fprintf(stderr, gettext("use 'zfs set " - "%s=none' to clear\n"), cb.cb_propname); + "%s=none' to clear\n"), propname); return (1); } - } else if (!zfs_prop_user(cb.cb_propname)) { - (void) fprintf(stderr, gettext( - "invalid property '%s'\n"), - cb.cb_propname); + } else if (!zfs_prop_user(propname)) { + (void) fprintf(stderr, gettext("invalid property '%s'\n"), + propname); usage(B_FALSE); } - cb.cb_any_successful = B_FALSE; + if (flags & ZFS_ITER_RECURSE) { + ret = zfs_for_each(argc, argv, flags, ZFS_TYPE_DATASET, + NULL, NULL, inherit_recurse_cb, propname); + } else { + ret = zfs_for_each(argc, argv, flags, ZFS_TYPE_DATASET, + NULL, NULL, inherit_cb, propname); + } + + return (ret); +} + +typedef struct upgrade_cbdata { + uint64_t cb_numupgraded; + uint64_t cb_numsamegraded; + uint64_t cb_numfailed; + uint64_t cb_version; + boolean_t cb_newer; + boolean_t cb_foundone; + char cb_lastfs[ZFS_MAXNAMELEN]; +} upgrade_cbdata_t; + +static int +same_pool(zfs_handle_t *zhp, const char *name) +{ + int len1 = strcspn(name, "/@"); + const char *zhname = zfs_get_name(zhp); + int len2 = strcspn(zhname, "/@"); + + if (len1 != len2) + return (B_FALSE); + return (strncmp(name, zhname, len1) == 0); +} + +static int +upgrade_list_callback(zfs_handle_t *zhp, void *data) +{ + upgrade_cbdata_t *cb = data; + int version = zfs_prop_get_int(zhp, ZFS_PROP_VERSION); + + /* list if it's old/new */ + if ((!cb->cb_newer && version < ZPL_VERSION) || + (cb->cb_newer && version > ZPL_VERSION)) { + char *str; + if (cb->cb_newer) { + str = gettext("The following filesystems are " + "formatted using a newer software version and\n" + "cannot be accessed on the current system.\n\n"); + } else { + str = gettext("The following filesystems are " + "out of date, and can be upgraded. After being\n" + "upgraded, these filesystems (and any 'zfs send' " + "streams generated from\n" + "subsequent snapshots) will no longer be " + "accessible by older software versions.\n\n"); + } + + if (!cb->cb_foundone) { + (void) puts(str); + (void) printf(gettext("VER FILESYSTEM\n")); + (void) printf(gettext("--- ------------\n")); + cb->cb_foundone = B_TRUE; + } + + (void) printf("%2u %s\n", version, zfs_get_name(zhp)); + } + + return (0); +} + +static int +upgrade_set_callback(zfs_handle_t *zhp, void *data) +{ + upgrade_cbdata_t *cb = data; + int version = zfs_prop_get_int(zhp, ZFS_PROP_VERSION); + + if (cb->cb_version >= ZPL_VERSION_FUID) { + int spa_version; + + if (zfs_spa_version(zhp, &spa_version) < 0) + return (-1); + + if (spa_version < SPA_VERSION_FUID) { + /* can't upgrade */ + (void) printf(gettext("%s: can not be upgraded; " + "the pool version needs to first be upgraded\nto " + "version %d\n\n"), + zfs_get_name(zhp), SPA_VERSION_FUID); + cb->cb_numfailed++; + return (0); + } + } + + /* upgrade */ + if (version < cb->cb_version) { + char verstr[16]; + (void) snprintf(verstr, sizeof (verstr), + "%llu", cb->cb_version); + if (cb->cb_lastfs[0] && !same_pool(zhp, cb->cb_lastfs)) { + /* + * If they did "zfs upgrade -a", then we could + * be doing ioctls to different pools. We need + * to log this history once to each pool. + */ + verify(zpool_stage_history(g_zfs, history_str) == 0); + } + if (zfs_prop_set(zhp, "version", verstr) == 0) + cb->cb_numupgraded++; + else + cb->cb_numfailed++; + (void) strcpy(cb->cb_lastfs, zfs_get_name(zhp)); + } else if (version > cb->cb_version) { + /* can't downgrade */ + (void) printf(gettext("%s: can not be downgraded; " + "it is already at version %u\n"), + zfs_get_name(zhp), version); + cb->cb_numfailed++; + } else { + cb->cb_numsamegraded++; + } + return (0); +} - ret = zfs_for_each(argc, argv, recurse, - ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME, NULL, NULL, - inherit_callback, &cb, B_FALSE); +/* + * zfs upgrade + * zfs upgrade -v + * zfs upgrade [-r] [-V <version>] <-a | filesystem> + */ +static int +zfs_do_upgrade(int argc, char **argv) +{ + boolean_t all = B_FALSE; + boolean_t showversions = B_FALSE; + int ret; + upgrade_cbdata_t cb = { 0 }; + char c; + int flags = ZFS_ITER_ARGS_CAN_BE_PATHS; - if (cb.cb_any_successful) { - zpool_log_history(g_zfs, argc + optind + 1, argv - optind - 1, - argv[0], B_FALSE, B_FALSE); + /* check options */ + while ((c = getopt(argc, argv, "rvV:a")) != -1) { + switch (c) { + case 'r': + flags |= ZFS_ITER_RECURSE; + break; + case 'v': + showversions = B_TRUE; + break; + case 'V': + if (zfs_prop_string_to_index(ZFS_PROP_VERSION, + optarg, &cb.cb_version) != 0) { + (void) fprintf(stderr, + gettext("invalid version %s\n"), optarg); + usage(B_FALSE); + } + break; + case 'a': + all = B_TRUE; + break; + case '?': + default: + (void) fprintf(stderr, gettext("invalid option '%c'\n"), + optopt); + usage(B_FALSE); + } + } + + argc -= optind; + argv += optind; + + if ((!all && !argc) && ((flags & ZFS_ITER_RECURSE) | cb.cb_version)) + usage(B_FALSE); + if (showversions && (flags & ZFS_ITER_RECURSE || all || + cb.cb_version || argc)) + usage(B_FALSE); + if ((all || argc) && (showversions)) + usage(B_FALSE); + if (all && argc) + usage(B_FALSE); + + if (showversions) { + /* Show info on available versions. */ + (void) printf(gettext("The following filesystem versions are " + "supported:\n\n")); + (void) printf(gettext("VER DESCRIPTION\n")); + (void) printf("--- -----------------------------------------" + "---------------\n"); + (void) printf(gettext(" 1 Initial ZFS filesystem version\n")); + (void) printf(gettext(" 2 Enhanced directory entries\n")); + (void) printf(gettext(" 3 Case insensitive and File system " + "unique identifer (FUID)\n")); + (void) printf(gettext("\nFor more information on a particular " + "version, including supported releases, see:\n\n")); + (void) printf("http://www.opensolaris.org/os/community/zfs/" + "version/zpl/N\n\n"); + (void) printf(gettext("Where 'N' is the version number.\n")); + ret = 0; + } else if (argc || all) { + /* Upgrade filesystems */ + if (cb.cb_version == 0) + cb.cb_version = ZPL_VERSION; + ret = zfs_for_each(argc, argv, flags, ZFS_TYPE_FILESYSTEM, + NULL, NULL, upgrade_set_callback, &cb); + (void) printf(gettext("%llu filesystems upgraded\n"), + cb.cb_numupgraded); + if (cb.cb_numsamegraded) { + (void) printf(gettext("%llu filesystems already at " + "this version\n"), + cb.cb_numsamegraded); + } + if (cb.cb_numfailed != 0) + ret = 1; + } else { + /* List old-version filesytems */ + boolean_t found; + (void) printf(gettext("This system is currently running " + "ZFS filesystem version %llu.\n\n"), ZPL_VERSION); + + flags |= ZFS_ITER_RECURSE; + ret = zfs_for_each(0, NULL, flags, ZFS_TYPE_FILESYSTEM, + NULL, NULL, upgrade_list_callback, &cb); + + found = cb.cb_foundone; + cb.cb_foundone = B_FALSE; + cb.cb_newer = B_TRUE; + + ret = zfs_for_each(0, NULL, flags, ZFS_TYPE_FILESYSTEM, + NULL, NULL, upgrade_list_callback, &cb); + + if (!cb.cb_foundone && !found) { + (void) printf(gettext("All filesystems are " + "formatted with the current version.\n")); + } } return (ret); @@ -1240,7 +1620,7 @@ zfs_do_inherit(int argc, char **argv) * <dataset> ... * * -r Recurse over all children - * -H Scripted mode; elide headers and separate colums by tabs + * -H Scripted mode; elide headers and separate columns by tabs * -o Control which fields to display. * -t Control which object types to display. * -s Specify sort columns, descending order. @@ -1253,14 +1633,14 @@ zfs_do_inherit(int argc, char **argv) typedef struct list_cbdata { boolean_t cb_first; boolean_t cb_scripted; - zfs_proplist_t *cb_proplist; + zprop_list_t *cb_proplist; } list_cbdata_t; /* * Given a list of columns to display, output appropriate headers for each one. */ static void -print_header(zfs_proplist_t *pl) +print_header(zprop_list_t *pl) { char headerbuf[ZFS_MAXPROPLEN]; const char *header; @@ -1276,7 +1656,7 @@ print_header(zfs_proplist_t *pl) } right_justify = B_FALSE; - if (pl->pl_prop != ZFS_PROP_INVAL) { + if (pl->pl_prop != ZPROP_INVAL) { header = zfs_prop_column_name(pl->pl_prop); right_justify = zfs_prop_align_right(pl->pl_prop); } else { @@ -1302,7 +1682,7 @@ print_header(zfs_proplist_t *pl) * to the described layout. */ static void -print_dataset(zfs_handle_t *zhp, zfs_proplist_t *pl, int scripted) +print_dataset(zfs_handle_t *zhp, zprop_list_t *pl, boolean_t scripted) { boolean_t first = B_TRUE; char property[ZFS_MAXPROPLEN]; @@ -1323,7 +1703,7 @@ print_dataset(zfs_handle_t *zhp, zfs_proplist_t *pl, int scripted) } right_justify = B_FALSE; - if (pl->pl_prop != ZFS_PROP_INVAL) { + if (pl->pl_prop != ZPROP_INVAL) { if (zfs_prop_get(zhp, pl->pl_prop, property, sizeof (property), NULL, NULL, 0, B_FALSE) != 0) propstr = "-"; @@ -1337,7 +1717,7 @@ print_dataset(zfs_handle_t *zhp, zfs_proplist_t *pl, int scripted) propstr = "-"; else verify(nvlist_lookup_string(propval, - ZFS_PROP_VALUE, &propstr) == 0); + ZPROP_VALUE, &propstr) == 0); } width = pl->pl_width; @@ -1381,18 +1761,17 @@ static int zfs_do_list(int argc, char **argv) { int c; - boolean_t recurse = B_FALSE; boolean_t scripted = B_FALSE; static char default_fields[] = "name,used,available,referenced,mountpoint"; - int types = ZFS_TYPE_ANY; + int types = ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME; + boolean_t types_specified = B_FALSE; char *fields = NULL; - char *basic_fields = default_fields; list_cbdata_t cb = { 0 }; char *value; int ret; - char *type_subopts[] = { "filesystem", "volume", "snapshot", NULL }; zfs_sort_column_t *sortcol = NULL; + int flags = ZFS_ITER_PROP_LISTSNAPS | ZFS_ITER_ARGS_CAN_BE_PATHS; /* check options */ while ((c = getopt(argc, argv, ":o:rt:Hs:S:")) != -1) { @@ -1401,7 +1780,7 @@ zfs_do_list(int argc, char **argv) fields = optarg; break; case 'r': - recurse = B_TRUE; + flags |= ZFS_ITER_RECURSE; break; case 'H': scripted = B_TRUE; @@ -1424,7 +1803,12 @@ zfs_do_list(int argc, char **argv) break; case 't': types = 0; + types_specified = B_TRUE; + flags &= ~ZFS_ITER_PROP_LISTSNAPS; while (*optarg != '\0') { + static char *type_subopts[] = { "filesystem", + "volume", "snapshot", "all", NULL }; + switch (getsubopt(&optarg, type_subopts, &value)) { case 0: @@ -1436,6 +1820,10 @@ zfs_do_list(int argc, char **argv) case 2: types |= ZFS_TYPE_SNAPSHOT; break; + case 3: + types = ZFS_TYPE_DATASET; + break; + default: (void) fprintf(stderr, gettext("invalid type '%s'\n"), @@ -1460,35 +1848,46 @@ zfs_do_list(int argc, char **argv) argv += optind; if (fields == NULL) - fields = basic_fields; + fields = default_fields; /* - * If the user specifies '-o all', the zfs_get_proplist() doesn't + * If "-o space" and no types were specified, don't display snapshots. + */ + if (strcmp(fields, "space") == 0 && types_specified == B_FALSE) + types &= ~ZFS_TYPE_SNAPSHOT; + + /* + * If the user specifies '-o all', the zprop_get_list() doesn't * normally include the name of the dataset. For 'zfs list', we always * want this property to be first. */ - if (zfs_get_proplist(g_zfs, fields, &cb.cb_proplist) != 0) + if (zprop_get_list(g_zfs, fields, &cb.cb_proplist, ZFS_TYPE_DATASET) + != 0) usage(B_FALSE); cb.cb_scripted = scripted; cb.cb_first = B_TRUE; - ret = zfs_for_each(argc, argv, recurse, types, sortcol, &cb.cb_proplist, - list_callback, &cb, B_TRUE); + ret = zfs_for_each(argc, argv, flags, types, sortcol, &cb.cb_proplist, + list_callback, &cb); - zfs_free_proplist(cb.cb_proplist); + zprop_free_list(cb.cb_proplist); zfs_free_sort_columns(sortcol); - if (ret == 0 && cb.cb_first) + if (ret == 0 && cb.cb_first && !cb.cb_scripted) (void) printf(gettext("no datasets available\n")); return (ret); } /* - * zfs rename [-r] <fs | snap | vol> <fs | snap | vol> + * zfs rename <fs | snap | vol> <fs | snap | vol> + * zfs rename -p <fs | vol> <fs | vol> + * zfs rename -r <snap> <snap> * * Renames the given dataset to another of the same type. + * + * The '-p' flag creates all the non-existing ancestors of the target first. */ /* ARGSUSED */ static int @@ -1497,13 +1896,17 @@ zfs_do_rename(int argc, char **argv) zfs_handle_t *zhp; int c; int ret; - int recurse = 0; + boolean_t recurse = B_FALSE; + boolean_t parents = B_FALSE; /* check options */ - while ((c = getopt(argc, argv, "r")) != -1) { + while ((c = getopt(argc, argv, "pr")) != -1) { switch (c) { + case 'p': + parents = B_TRUE; + break; case 'r': - recurse = 1; + recurse = B_TRUE; break; case '?': default: @@ -1532,20 +1935,30 @@ zfs_do_rename(int argc, char **argv) usage(B_FALSE); } + if (recurse && parents) { + (void) fprintf(stderr, gettext("-p and -r options are mutually " + "exclusive\n")); + usage(B_FALSE); + } + if (recurse && strchr(argv[0], '@') == 0) { (void) fprintf(stderr, gettext("source dataset for recursive " "rename must be a snapshot\n")); usage(B_FALSE); } - if ((zhp = zfs_open(g_zfs, argv[0], ZFS_TYPE_ANY)) == NULL) + if ((zhp = zfs_open(g_zfs, argv[0], parents ? ZFS_TYPE_FILESYSTEM | + ZFS_TYPE_VOLUME : ZFS_TYPE_DATASET)) == NULL) return (1); - ret = (zfs_rename(zhp, argv[1], recurse) != 0); + /* If we were asked and the name looks good, try to create ancestors. */ + if (parents && zfs_name_valid(argv[1], zfs_get_type(zhp)) && + zfs_create_ancestors(g_zfs, argv[1]) != 0) { + zfs_close(zhp); + return (1); + } - if (!ret) - zpool_log_history(g_zfs, argc + optind, argv - optind, argv[1], - B_FALSE, B_FALSE); + ret = (zfs_rename(zhp, argv[1], recurse) != 0); zfs_close(zhp); return (ret); @@ -1587,19 +2000,17 @@ zfs_do_promote(int argc, char **argv) ret = (zfs_promote(zhp) != 0); - if (!ret) - zpool_log_history(g_zfs, argc, argv, argv[1], B_FALSE, B_FALSE); zfs_close(zhp); return (ret); } /* - * zfs rollback [-rfR] <snapshot> + * zfs rollback [-rRf] <snapshot> * * -r Delete any intervening snapshots before doing rollback * -R Delete any snapshots and their clones - * -f Force unmount filesystems, even if they are in use. + * -f ignored for backwards compatability * * Given a filesystem, rollback to a specific snapshot, discarding any changes * since then and making it the active dataset. If more recent snapshots exist, @@ -1686,18 +2097,15 @@ zfs_do_rollback(int argc, char **argv) { int ret; int c; + boolean_t force = B_FALSE; rollback_cbdata_t cb = { 0 }; zfs_handle_t *zhp, *snap; char parentname[ZFS_MAXNAMELEN]; char *delim; - int force = 0; /* check options */ - while ((c = getopt(argc, argv, "rfR")) != -1) { + while ((c = getopt(argc, argv, "rRf")) != -1) { switch (c) { - case 'f': - force = 1; - break; case 'r': cb.cb_recurse = 1; break; @@ -1705,6 +2113,9 @@ zfs_do_rollback(int argc, char **argv) cb.cb_recurse = 1; cb.cb_doclones = 1; break; + case 'f': + force = B_TRUE; + break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); @@ -1733,7 +2144,7 @@ zfs_do_rollback(int argc, char **argv) (void) strlcpy(parentname, argv[0], sizeof (parentname)); verify((delim = strrchr(parentname, '@')) != NULL); *delim = '\0'; - if ((zhp = zfs_open(g_zfs, parentname, ZFS_TYPE_ANY)) == NULL) { + if ((zhp = zfs_open(g_zfs, parentname, ZFS_TYPE_DATASET)) == NULL) { zfs_close(snap); return (1); } @@ -1757,11 +2168,6 @@ zfs_do_rollback(int argc, char **argv) */ ret = zfs_rollback(zhp, snap, force); - if (!ret) { - zpool_log_history(g_zfs, argc + optind, argv - optind, argv[0], - B_FALSE, B_FALSE); - } - out: zfs_close(snap); zfs_close(zhp); @@ -1780,7 +2186,6 @@ out: typedef struct set_cbdata { char *cb_propname; char *cb_value; - boolean_t cb_any_successful; } set_cbdata_t; static int @@ -1801,7 +2206,6 @@ set_callback(zfs_handle_t *zhp, void *data) } return (1); } - cbp->cb_any_successful = B_TRUE; return (0); } @@ -1831,7 +2235,8 @@ zfs_do_set(int argc, char **argv) /* validate property=value argument */ cb.cb_propname = argv[1]; - if ((cb.cb_value = strchr(cb.cb_propname, '=')) == NULL) { + if (((cb.cb_value = strchr(cb.cb_propname, '=')) == NULL) || + (cb.cb_value[1] == '\0')) { (void) fprintf(stderr, gettext("missing value in " "property=value argument\n")); usage(B_FALSE); @@ -1839,7 +2244,6 @@ zfs_do_set(int argc, char **argv) *cb.cb_value = '\0'; cb.cb_value++; - cb.cb_any_successful = B_FALSE; if (*cb.cb_propname == '\0') { (void) fprintf(stderr, @@ -1847,40 +2251,46 @@ zfs_do_set(int argc, char **argv) usage(B_FALSE); } - ret = zfs_for_each(argc - 2, argv + 2, B_FALSE, - ZFS_TYPE_ANY, NULL, NULL, set_callback, &cb, B_FALSE); - - if (cb.cb_any_successful) { - *(cb.cb_value - 1) = '='; - zpool_log_history(g_zfs, argc, argv, argv[2], B_FALSE, B_FALSE); - } + ret = zfs_for_each(argc - 2, argv + 2, NULL, + ZFS_TYPE_DATASET, NULL, NULL, set_callback, &cb); return (ret); } /* - * zfs snapshot [-r] <fs@snap> + * zfs snapshot [-r] [-o prop=value] ... <fs@snap> * * Creates a snapshot with the given name. While functionally equivalent to - * 'zfs create', it is a separate command to diffferentiate intent. + * 'zfs create', it is a separate command to differentiate intent. */ static int zfs_do_snapshot(int argc, char **argv) { - int recursive = B_FALSE; + boolean_t recursive = B_FALSE; int ret; char c; + nvlist_t *props; + + if (nvlist_alloc(&props, NV_UNIQUE_NAME, 0) != 0) { + (void) fprintf(stderr, gettext("internal error: " + "out of memory\n")); + return (1); + } /* check options */ - while ((c = getopt(argc, argv, ":r")) != -1) { + while ((c = getopt(argc, argv, "ro:")) != -1) { switch (c) { + case 'o': + if (parseprop(props)) + return (1); + break; case 'r': recursive = B_TRUE; break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); - usage(B_FALSE); + goto usage; } } @@ -1890,25 +2300,28 @@ zfs_do_snapshot(int argc, char **argv) /* check number of arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing snapshot argument\n")); - usage(B_FALSE); + goto usage; } if (argc > 1) { (void) fprintf(stderr, gettext("too many arguments\n")); - usage(B_FALSE); + goto usage; } - ret = zfs_snapshot(g_zfs, argv[0], recursive); + ret = zfs_snapshot(g_zfs, argv[0], recursive, props); + nvlist_free(props); if (ret && recursive) (void) fprintf(stderr, gettext("no snapshots were created\n")); - if (!ret) { - zpool_log_history(g_zfs, argc + optind, argv - optind, argv[0], - B_FALSE, B_FALSE); - } return (ret != 0); + +usage: + nvlist_free(props); + usage(B_FALSE); + return (-1); } /* - * zfs send [-i <@snap>] <fs@snap> + * zfs send [-v] -R [-i|-I <@snap>] <fs@snap> + * zfs send [-v] [-i|-I <@snap>] <fs@snap> * * Send a backup stream to stdout. */ @@ -1916,18 +2329,35 @@ static int zfs_do_send(int argc, char **argv) { char *fromname = NULL; + char *toname = NULL; char *cp; zfs_handle_t *zhp; + boolean_t doall = B_FALSE; + boolean_t replicate = B_FALSE; + boolean_t fromorigin = B_FALSE; + boolean_t verbose = B_FALSE; int c, err; /* check options */ - while ((c = getopt(argc, argv, ":i:")) != -1) { + while ((c = getopt(argc, argv, ":i:I:Rv")) != -1) { switch (c) { case 'i': if (fromname) usage(B_FALSE); fromname = optarg; break; + case 'I': + if (fromname) + usage(B_FALSE); + fromname = optarg; + doall = B_TRUE; + break; + case 'R': + replicate = B_TRUE; + break; + case 'v': + verbose = B_TRUE; + break; case ':': (void) fprintf(stderr, gettext("missing argument for " "'%c' option\n"), optopt); @@ -1960,37 +2390,62 @@ zfs_do_send(int argc, char **argv) return (1); } - if ((zhp = zfs_open(g_zfs, argv[0], ZFS_TYPE_SNAPSHOT)) == NULL) + cp = strchr(argv[0], '@'); + if (cp == NULL) { + (void) fprintf(stderr, + gettext("argument must be a snapshot\n")); + usage(B_FALSE); + } + *cp = '\0'; + toname = cp + 1; + zhp = zfs_open(g_zfs, argv[0], ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME); + if (zhp == NULL) return (1); /* * If they specified the full path to the snapshot, chop off - * everything except the short name of the snapshot. + * everything except the short name of the snapshot, but special + * case if they specify the origin. */ if (fromname && (cp = strchr(fromname, '@')) != NULL) { - if (cp != fromname && - strncmp(argv[0], fromname, cp - fromname + 1)) { - (void) fprintf(stderr, - gettext("incremental source must be " - "in same filesystem\n")); - usage(B_FALSE); - } - fromname = cp + 1; - if (strchr(fromname, '@') || strchr(fromname, '/')) { - (void) fprintf(stderr, - gettext("invalid incremental source\n")); - usage(B_FALSE); + char origin[ZFS_MAXNAMELEN]; + zprop_source_t src; + + (void) zfs_prop_get(zhp, ZFS_PROP_ORIGIN, + origin, sizeof (origin), &src, NULL, 0, B_FALSE); + + if (strcmp(origin, fromname) == 0) { + fromname = NULL; + fromorigin = B_TRUE; + } else { + *cp = '\0'; + if (cp != fromname && strcmp(argv[0], fromname)) { + (void) fprintf(stderr, + gettext("incremental source must be " + "in same filesystem\n")); + usage(B_FALSE); + } + fromname = cp + 1; + if (strchr(fromname, '@') || strchr(fromname, '/')) { + (void) fprintf(stderr, + gettext("invalid incremental source\n")); + usage(B_FALSE); + } } } - err = zfs_send(zhp, fromname, STDOUT_FILENO); + if (replicate && fromname == NULL) + doall = B_TRUE; + + err = zfs_send(zhp, fromname, toname, replicate, doall, fromorigin, + verbose, STDOUT_FILENO); zfs_close(zhp); return (err != 0); } /* - * zfs receive <fs@snap> + * zfs receive [-dnvF] <fs@snap> * * Restore a backup stream from stdin. */ @@ -1998,25 +2453,23 @@ static int zfs_do_receive(int argc, char **argv) { int c, err; - boolean_t isprefix = B_FALSE; - boolean_t dryrun = B_FALSE; - boolean_t verbose = B_FALSE; - boolean_t force = B_FALSE; + recvflags_t flags; + bzero(&flags, sizeof (recvflags_t)); /* check options */ while ((c = getopt(argc, argv, ":dnvF")) != -1) { switch (c) { case 'd': - isprefix = B_TRUE; + flags.isprefix = B_TRUE; break; case 'n': - dryrun = B_TRUE; + flags.dryrun = B_TRUE; break; case 'v': - verbose = B_TRUE; + flags.verbose = B_TRUE; break; case 'F': - force = B_TRUE; + flags.force = B_TRUE; break; case ':': (void) fprintf(stderr, gettext("missing argument for " @@ -2051,15 +2504,391 @@ zfs_do_receive(int argc, char **argv) return (1); } - err = zfs_receive(g_zfs, argv[0], isprefix, verbose, dryrun, force, - STDIN_FILENO); + err = zfs_receive(g_zfs, argv[0], flags, STDIN_FILENO, NULL); + + return (err != 0); +} + +typedef struct allow_cb { + int a_permcnt; + size_t a_treeoffset; +} allow_cb_t; - if (!err) { - zpool_log_history(g_zfs, argc + optind, argv - optind, argv[0], - B_FALSE, B_FALSE); +static void +zfs_print_perms(avl_tree_t *tree) +{ + zfs_perm_node_t *permnode; + + permnode = avl_first(tree); + while (permnode != NULL) { + (void) printf("%s", permnode->z_pname); + permnode = AVL_NEXT(tree, permnode); + if (permnode) + (void) printf(","); + else + (void) printf("\n"); } +} - return (err != 0); +/* + * Iterate over user/groups/everyone/... and the call perm_iter + * function to print actual permission when tree has >0 nodes. + */ +static void +zfs_iter_perms(avl_tree_t *tree, const char *banner, allow_cb_t *cb) +{ + zfs_allow_node_t *item; + avl_tree_t *ptree; + + item = avl_first(tree); + while (item) { + ptree = (void *)((char *)item + cb->a_treeoffset); + if (avl_numnodes(ptree)) { + if (cb->a_permcnt++ == 0) + (void) printf("%s\n", banner); + (void) printf("\t%s", item->z_key); + /* + * Avoid an extra space being printed + * for "everyone" which is keyed with a null + * string + */ + if (item->z_key[0] != '\0') + (void) printf(" "); + zfs_print_perms(ptree); + } + item = AVL_NEXT(tree, item); + } +} + +#define LINES "-------------------------------------------------------------\n" +static int +zfs_print_allows(char *ds) +{ + zfs_allow_t *curperms, *perms; + zfs_handle_t *zhp; + allow_cb_t allowcb = { 0 }; + char banner[MAXPATHLEN]; + + if (ds[0] == '-') + usage(B_FALSE); + + if (strrchr(ds, '@')) { + (void) fprintf(stderr, gettext("Snapshots don't have 'allow'" + " permissions\n")); + return (1); + } + if ((zhp = zfs_open(g_zfs, ds, ZFS_TYPE_DATASET)) == NULL) + return (1); + + if (zfs_perm_get(zhp, &perms)) { + (void) fprintf(stderr, + gettext("Failed to retrieve 'allows' on %s\n"), ds); + zfs_close(zhp); + return (1); + } + + zfs_close(zhp); + + if (perms != NULL) + (void) printf("%s", LINES); + for (curperms = perms; curperms; curperms = curperms->z_next) { + + (void) snprintf(banner, sizeof (banner), + "Permission sets on (%s)", curperms->z_setpoint); + allowcb.a_treeoffset = + offsetof(zfs_allow_node_t, z_localdescend); + allowcb.a_permcnt = 0; + zfs_iter_perms(&curperms->z_sets, banner, &allowcb); + + (void) snprintf(banner, sizeof (banner), + "Create time permissions on (%s)", curperms->z_setpoint); + allowcb.a_treeoffset = + offsetof(zfs_allow_node_t, z_localdescend); + allowcb.a_permcnt = 0; + zfs_iter_perms(&curperms->z_crperms, banner, &allowcb); + + + (void) snprintf(banner, sizeof (banner), + "Local permissions on (%s)", curperms->z_setpoint); + allowcb.a_treeoffset = offsetof(zfs_allow_node_t, z_local); + allowcb.a_permcnt = 0; + zfs_iter_perms(&curperms->z_user, banner, &allowcb); + zfs_iter_perms(&curperms->z_group, banner, &allowcb); + zfs_iter_perms(&curperms->z_everyone, banner, &allowcb); + + (void) snprintf(banner, sizeof (banner), + "Descendent permissions on (%s)", curperms->z_setpoint); + allowcb.a_treeoffset = offsetof(zfs_allow_node_t, z_descend); + allowcb.a_permcnt = 0; + zfs_iter_perms(&curperms->z_user, banner, &allowcb); + zfs_iter_perms(&curperms->z_group, banner, &allowcb); + zfs_iter_perms(&curperms->z_everyone, banner, &allowcb); + + (void) snprintf(banner, sizeof (banner), + "Local+Descendent permissions on (%s)", + curperms->z_setpoint); + allowcb.a_treeoffset = + offsetof(zfs_allow_node_t, z_localdescend); + allowcb.a_permcnt = 0; + zfs_iter_perms(&curperms->z_user, banner, &allowcb); + zfs_iter_perms(&curperms->z_group, banner, &allowcb); + zfs_iter_perms(&curperms->z_everyone, banner, &allowcb); + + (void) printf("%s", LINES); + } + zfs_free_allows(perms); + return (0); +} + +#define ALLOWOPTIONS "ldcsu:g:e" +#define UNALLOWOPTIONS "ldcsu:g:er" + +/* + * Validate options, and build necessary datastructure to display/remove/add + * permissions. + * Returns 0 - If permissions should be added/removed + * Returns 1 - If permissions should be displayed. + * Returns -1 - on failure + */ +int +parse_allow_args(int *argc, char **argv[], boolean_t unallow, + char **ds, int *recurse, nvlist_t **zperms) +{ + int c; + char *options = unallow ? UNALLOWOPTIONS : ALLOWOPTIONS; + zfs_deleg_inherit_t deleg_type = ZFS_DELEG_NONE; + zfs_deleg_who_type_t who_type = ZFS_DELEG_WHO_UNKNOWN; + char *who = NULL; + char *perms = NULL; + zfs_handle_t *zhp; + + while ((c = getopt(*argc, *argv, options)) != -1) { + switch (c) { + case 'l': + if (who_type == ZFS_DELEG_CREATE || + who_type == ZFS_DELEG_NAMED_SET) + usage(B_FALSE); + + deleg_type |= ZFS_DELEG_PERM_LOCAL; + break; + case 'd': + if (who_type == ZFS_DELEG_CREATE || + who_type == ZFS_DELEG_NAMED_SET) + usage(B_FALSE); + + deleg_type |= ZFS_DELEG_PERM_DESCENDENT; + break; + case 'r': + *recurse = B_TRUE; + break; + case 'c': + if (who_type != ZFS_DELEG_WHO_UNKNOWN) + usage(B_FALSE); + if (deleg_type) + usage(B_FALSE); + who_type = ZFS_DELEG_CREATE; + break; + case 's': + if (who_type != ZFS_DELEG_WHO_UNKNOWN) + usage(B_FALSE); + if (deleg_type) + usage(B_FALSE); + who_type = ZFS_DELEG_NAMED_SET; + break; + case 'u': + if (who_type != ZFS_DELEG_WHO_UNKNOWN) + usage(B_FALSE); + who_type = ZFS_DELEG_USER; + who = optarg; + break; + case 'g': + if (who_type != ZFS_DELEG_WHO_UNKNOWN) + usage(B_FALSE); + who_type = ZFS_DELEG_GROUP; + who = optarg; + break; + case 'e': + if (who_type != ZFS_DELEG_WHO_UNKNOWN) + usage(B_FALSE); + who_type = ZFS_DELEG_EVERYONE; + break; + default: + usage(B_FALSE); + break; + } + } + + if (deleg_type == 0) + deleg_type = ZFS_DELEG_PERM_LOCALDESCENDENT; + + *argc -= optind; + *argv += optind; + + if (unallow == B_FALSE && *argc == 1) { + /* + * Only print permissions if no options were processed + */ + if (optind == 1) + return (1); + else + usage(B_FALSE); + } + + /* + * initialize variables for zfs_build_perms based on number + * of arguments. + * 3 arguments ==> zfs [un]allow joe perm,perm,perm <dataset> or + * zfs [un]allow -s @set1 perm,perm <dataset> + * 2 arguments ==> zfs [un]allow -c perm,perm <dataset> or + * zfs [un]allow -u|-g <name> perm <dataset> or + * zfs [un]allow -e perm,perm <dataset> + * zfs unallow joe <dataset> + * zfs unallow -s @set1 <dataset> + * 1 argument ==> zfs [un]allow -e <dataset> or + * zfs [un]allow -c <dataset> + */ + + switch (*argc) { + case 3: + perms = (*argv)[1]; + who = (*argv)[0]; + *ds = (*argv)[2]; + + /* + * advance argc/argv for do_allow cases. + * for do_allow case make sure who have a know who type + * and its not a permission set. + */ + if (unallow == B_TRUE) { + *argc -= 2; + *argv += 2; + } else if (who_type != ZFS_DELEG_WHO_UNKNOWN && + who_type != ZFS_DELEG_NAMED_SET) + usage(B_FALSE); + break; + + case 2: + if (unallow == B_TRUE && (who_type == ZFS_DELEG_EVERYONE || + who_type == ZFS_DELEG_CREATE || who != NULL)) { + perms = (*argv)[0]; + *ds = (*argv)[1]; + } else { + if (unallow == B_FALSE && + (who_type == ZFS_DELEG_WHO_UNKNOWN || + who_type == ZFS_DELEG_NAMED_SET)) + usage(B_FALSE); + else if (who_type == ZFS_DELEG_WHO_UNKNOWN || + who_type == ZFS_DELEG_NAMED_SET) + who = (*argv)[0]; + else if (who_type != ZFS_DELEG_NAMED_SET) + perms = (*argv)[0]; + *ds = (*argv)[1]; + } + if (unallow == B_TRUE) { + (*argc)--; + (*argv)++; + } + break; + + case 1: + if (unallow == B_FALSE) + usage(B_FALSE); + if (who == NULL && who_type != ZFS_DELEG_CREATE && + who_type != ZFS_DELEG_EVERYONE) + usage(B_FALSE); + *ds = (*argv)[0]; + break; + + default: + usage(B_FALSE); + } + + if (strrchr(*ds, '@')) { + (void) fprintf(stderr, + gettext("Can't set or remove 'allow' permissions " + "on snapshots.\n")); + return (-1); + } + + if ((zhp = zfs_open(g_zfs, *ds, ZFS_TYPE_DATASET)) == NULL) + return (-1); + + if ((zfs_build_perms(zhp, who, perms, + who_type, deleg_type, zperms)) != 0) { + zfs_close(zhp); + return (-1); + } + zfs_close(zhp); + return (0); +} + +static int +zfs_do_allow(int argc, char **argv) +{ + char *ds; + nvlist_t *zperms = NULL; + zfs_handle_t *zhp; + int unused; + int ret; + + if ((ret = parse_allow_args(&argc, &argv, B_FALSE, &ds, + &unused, &zperms)) == -1) + return (1); + + if (ret == 1) + return (zfs_print_allows(argv[0])); + + if ((zhp = zfs_open(g_zfs, ds, ZFS_TYPE_DATASET)) == NULL) + return (1); + + if (zfs_perm_set(zhp, zperms)) { + zfs_close(zhp); + nvlist_free(zperms); + return (1); + } + nvlist_free(zperms); + zfs_close(zhp); + + return (0); +} + +static int +unallow_callback(zfs_handle_t *zhp, void *data) +{ + nvlist_t *nvp = (nvlist_t *)data; + int error; + + error = zfs_perm_remove(zhp, nvp); + if (error) { + (void) fprintf(stderr, gettext("Failed to remove permissions " + "on %s\n"), zfs_get_name(zhp)); + } + return (error); +} + +static int +zfs_do_unallow(int argc, char **argv) +{ + int recurse = B_FALSE; + char *ds; + int error; + nvlist_t *zperms = NULL; + int flags = 0; + + if (parse_allow_args(&argc, &argv, B_TRUE, + &ds, &recurse, &zperms) == -1) + return (1); + + if (recurse) + flags |= ZFS_ITER_RECURSE; + error = zfs_for_each(argc, argv, flags, + ZFS_TYPE_FILESYSTEM|ZFS_TYPE_VOLUME, NULL, + NULL, unallow_callback, (void *)zperms); + + if (zperms) + nvlist_free(zperms); + + return (error); } typedef struct get_all_cbdata { @@ -2067,14 +2896,35 @@ typedef struct get_all_cbdata { size_t cb_alloc; size_t cb_used; uint_t cb_types; + boolean_t cb_verbose; } get_all_cbdata_t; +#define CHECK_SPINNER 30 +#define SPINNER_TIME 3 /* seconds */ +#define MOUNT_TIME 5 /* seconds */ + static int get_one_dataset(zfs_handle_t *zhp, void *data) { + static char spin[] = { '-', '\\', '|', '/' }; + static int spinval = 0; + static int spincheck = 0; + static time_t last_spin_time = (time_t)0; get_all_cbdata_t *cbp = data; zfs_type_t type = zfs_get_type(zhp); + if (cbp->cb_verbose) { + if (--spincheck < 0) { + time_t now = time(NULL); + if (last_spin_time + SPINNER_TIME < now) { + (void) printf("\b%c", spin[spinval++ % 4]); + (void) fflush(stdout); + last_spin_time = now; + } + spincheck = CHECK_SPINNER; + } + } + /* * Interate over any nested datasets. */ @@ -2117,15 +2967,26 @@ get_one_dataset(zfs_handle_t *zhp, void *data) } static void -get_all_datasets(uint_t types, zfs_handle_t ***dslist, size_t *count) +get_all_datasets(uint_t types, zfs_handle_t ***dslist, size_t *count, + boolean_t verbose) { get_all_cbdata_t cb = { 0 }; cb.cb_types = types; + cb.cb_verbose = verbose; + + if (verbose) { + (void) printf("%s: *", gettext("Reading ZFS config")); + (void) fflush(stdout); + } (void) zfs_iter_root(g_zfs, get_one_dataset, &cb); *dslist = cb.cb_handles; *count = cb.cb_used; + + if (verbose) { + (void) printf("\b%s\n", gettext("done.")); + } } static int @@ -2167,15 +3028,17 @@ dataset_cmp(const void *a, const void *b) * Share or mount a dataset. */ static int -share_mount_one(zfs_handle_t *zhp, int op, int flags, boolean_t explicit, - const char *options) +share_mount_one(zfs_handle_t *zhp, int op, int flags, char *protocol, + boolean_t explicit, const char *options) { char mountpoint[ZFS_MAXPROPLEN]; char shareopts[ZFS_MAXPROPLEN]; + char smbshareopts[ZFS_MAXPROPLEN]; const char *cmdname = op == OP_SHARE ? "share" : "mount"; struct mnttab mnt; uint64_t zoned, canmount; zfs_type_t type = zfs_get_type(zhp); + boolean_t shared_nfs, shared_smb; assert(type & (ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME)); @@ -2216,9 +3079,12 @@ share_mount_one(zfs_handle_t *zhp, int op, int flags, boolean_t explicit, sizeof (mountpoint), NULL, NULL, 0, B_FALSE) == 0); verify(zfs_prop_get(zhp, ZFS_PROP_SHARENFS, shareopts, sizeof (shareopts), NULL, NULL, 0, B_FALSE) == 0); + verify(zfs_prop_get(zhp, ZFS_PROP_SHARESMB, smbshareopts, + sizeof (smbshareopts), NULL, NULL, 0, B_FALSE) == 0); canmount = zfs_prop_get_int(zhp, ZFS_PROP_CANMOUNT); - if (op == OP_SHARE && strcmp(shareopts, "off") == 0) { + if (op == OP_SHARE && strcmp(shareopts, "off") == 0 && + strcmp(smbshareopts, "off") == 0) { if (!explicit) return (0); @@ -2240,9 +3106,8 @@ share_mount_one(zfs_handle_t *zhp, int op, int flags, boolean_t explicit, (void) fprintf(stderr, gettext("cannot %s '%s': " "legacy mountpoint\n"), cmdname, zfs_get_name(zhp)); - (void) fprintf(stderr, gettext("use %s to " - "%s this filesystem\n"), op == OP_SHARE ? - "share(1M)" : "mount(1M)", cmdname); + (void) fprintf(stderr, gettext("use %s(1M) to " + "%s this filesystem\n"), cmdname, cmdname); return (1); } @@ -2255,7 +3120,16 @@ share_mount_one(zfs_handle_t *zhp, int op, int flags, boolean_t explicit, return (1); } - if (!canmount) { + /* + * canmount explicit outcome + * on no pass through + * on yes pass through + * off no return 0 + * off yes display error, return 1 + * noauto no return 0 + * noauto yes pass through + */ + if (canmount == ZFS_CANMOUNT_OFF) { if (!explicit) return (0); @@ -2263,6 +3137,8 @@ share_mount_one(zfs_handle_t *zhp, int op, int flags, boolean_t explicit, "'canmount' property is set to 'off'\n"), cmdname, zfs_get_name(zhp)); return (1); + } else if (canmount == ZFS_CANMOUNT_NOAUTO && !explicit) { + return (0); } /* @@ -2274,7 +3150,15 @@ share_mount_one(zfs_handle_t *zhp, int op, int flags, boolean_t explicit, */ switch (op) { case OP_SHARE: - if (zfs_is_shared_nfs(zhp, NULL)) { + + shared_nfs = zfs_is_shared_nfs(zhp, NULL); + shared_smb = zfs_is_shared_smb(zhp, NULL); + + if (shared_nfs && shared_smb || + (shared_nfs && strcmp(shareopts, "on") == 0 && + strcmp(smbshareopts, "off") == 0) || + (shared_smb && strcmp(smbshareopts, "on") == 0 && + strcmp(shareopts, "off") == 0)) { if (!explicit) return (0); @@ -2288,8 +3172,23 @@ share_mount_one(zfs_handle_t *zhp, int op, int flags, boolean_t explicit, zfs_mount(zhp, NULL, 0) != 0) return (1); - if (zfs_share_nfs(zhp) != 0) + if (protocol == NULL) { + if (zfs_shareall(zhp) != 0) + return (1); + } else if (strcmp(protocol, "nfs") == 0) { + if (zfs_share_nfs(zhp)) + return (1); + } else if (strcmp(protocol, "smb") == 0) { + if (zfs_share_smb(zhp)) + return (1); + } else { + (void) fprintf(stderr, gettext("cannot share " + "'%s': invalid share type '%s' " + "specified\n"), + zfs_get_name(zhp), protocol); return (1); + } + break; case OP_MOUNT: @@ -2352,24 +3251,93 @@ share_mount_one(zfs_handle_t *zhp, int op, int flags, boolean_t explicit, return (0); } +/* + * Reports progress in the form "(current/total)". Not thread-safe. + */ +static void +report_mount_progress(int current, int total) +{ + static int len; + static char *reverse = "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b" + "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b"; + static time_t last_progress_time; + time_t now = time(NULL); + + /* report 1..n instead of 0..n-1 */ + ++current; + + /* display header if we're here for the first time */ + if (current == 1) { + (void) printf(gettext("Mounting ZFS filesystems: ")); + len = 0; + } else if (current != total && last_progress_time + MOUNT_TIME >= now) { + /* too soon to report again */ + return; + } + + last_progress_time = now; + + /* back up to prepare for overwriting */ + if (len) + (void) printf("%*.*s", len, len, reverse); + + /* We put a newline at the end if this is the last one. */ + len = printf("(%d/%d)%s", current, total, current == total ? "\n" : ""); + (void) fflush(stdout); +} + +static void +append_options(char *mntopts, char *newopts) +{ + int len = strlen(mntopts); + + /* original length plus new string to append plus 1 for the comma */ + if (len + 1 + strlen(newopts) >= MNT_LINE_MAX) { + (void) fprintf(stderr, gettext("the opts argument for " + "'%c' option is too long (more than %d chars)\n"), + "-o", MNT_LINE_MAX); + usage(B_FALSE); + } + + if (*mntopts) + mntopts[len++] = ','; + + (void) strcpy(&mntopts[len], newopts); +} + static int share_mount(int op, int argc, char **argv) { int do_all = 0; + boolean_t verbose = B_FALSE; int c, ret = 0; - const char *options = NULL; + char *options = NULL; int types, flags = 0; /* check options */ - while ((c = getopt(argc, argv, op == OP_MOUNT ? ":ao:O" : "a")) + while ((c = getopt(argc, argv, op == OP_MOUNT ? ":avo:O" : "a")) != -1) { switch (c) { case 'a': do_all = 1; break; + case 'v': + verbose = B_TRUE; + break; case 'o': - options = optarg; + if (*optarg == '\0') { + (void) fprintf(stderr, gettext("empty mount " + "options (-o) specified\n")); + usage(B_FALSE); + } + + if (options == NULL) + options = safe_malloc(MNT_LINE_MAX + 1); + + /* option validation is done later */ + append_options(options, optarg); break; + case 'O': warnx("no overlay mounts support on FreeBSD, ignoring"); break; @@ -2392,20 +3360,22 @@ share_mount(int op, int argc, char **argv) if (do_all) { zfs_handle_t **dslist = NULL; size_t i, count = 0; + char *protocol = NULL; if (op == OP_MOUNT) { types = ZFS_TYPE_FILESYSTEM; } else if (argc > 0) { - if (strcmp(argv[0], "nfs") == 0) { + if (strcmp(argv[0], "nfs") == 0 || + strcmp(argv[0], "smb") == 0) { types = ZFS_TYPE_FILESYSTEM; } else if (strcmp(argv[0], "iscsi") == 0) { types = ZFS_TYPE_VOLUME; } else { (void) fprintf(stderr, gettext("share type " - "must be 'nfs' or 'iscsi'\n")); + "must be 'nfs', 'smb' or 'iscsi'\n")); usage(B_FALSE); } - + protocol = argv[0]; argc--; argv++; } else { @@ -2417,7 +3387,7 @@ share_mount(int op, int argc, char **argv) usage(B_FALSE); } - get_all_datasets(types, &dslist, &count); + get_all_datasets(types, &dslist, &count, verbose); if (count == 0) return (0); @@ -2425,8 +3395,11 @@ share_mount(int op, int argc, char **argv) qsort(dslist, count, sizeof (void *), dataset_cmp); for (i = 0; i < count; i++) { - if (share_mount_one(dslist[i], op, flags, B_FALSE, - options) != 0) + if (verbose) + report_mount_progress(i, count); + + if (share_mount_one(dslist[i], op, flags, protocol, + B_FALSE, options) != 0) ret = 1; zfs_close(dslist[i]); } @@ -2436,9 +3409,9 @@ share_mount(int op, int argc, char **argv) struct statfs *sfs; int i, n; - if (op == OP_SHARE) { + if ((op == OP_SHARE) || (options != NULL)) { (void) fprintf(stderr, gettext("missing filesystem " - "argument\n")); + "argument (specify -a for all)\n")); usage(B_FALSE); } @@ -2476,7 +3449,7 @@ share_mount(int op, int argc, char **argv) if ((zhp = zfs_open(g_zfs, argv[0], types)) == NULL) { ret = 1; } else { - ret = share_mount_one(zhp, op, flags, B_TRUE, + ret = share_mount_one(zhp, op, flags, NULL, B_TRUE, options); zfs_close(zhp); } @@ -2498,7 +3471,7 @@ zfs_do_mount(int argc, char **argv) } /* - * zfs share -a [nfs | iscsi] + * zfs share -a [nfs | iscsi | smb] * zfs share filesystem * * Share all filesystems, or share the given filesystem. @@ -2535,9 +3508,23 @@ unshare_unmount_path(int op, char *path, int flags, boolean_t is_manual) { zfs_handle_t *zhp; int ret; + struct stat64 statbuf; struct mnttab search = { 0 }, entry; const char *cmdname = (op == OP_SHARE) ? "unshare" : "unmount"; - char property[ZFS_MAXPROPLEN]; + ino_t path_inode; + + /* + * Search for the path in /etc/mnttab. Rather than looking for the + * specific path, which can be fooled by non-standard paths (i.e. ".." + * or "//"), we stat() the path and search for the corresponding + * (major,minor) device pair. + */ + if (stat64(path, &statbuf) != 0) { + (void) fprintf(stderr, gettext("cannot %s '%s': %s\n"), + cmdname, path, strerror(errno)); + return (1); + } + path_inode = statbuf.st_ino; /* * Search for the given (major,minor) pair in the mount table. @@ -2545,9 +3532,17 @@ unshare_unmount_path(int op, char *path, int flags, boolean_t is_manual) search.mnt_mountp = path; rewind(mnttab_file); if (getmntany(mnttab_file, &entry, &search) != 0) { - (void) fprintf(stderr, gettext("cannot %s '%s': not " - "currently mounted\n"), cmdname, path); - return (1); + if (op == OP_SHARE) { + (void) fprintf(stderr, gettext("cannot %s '%s': not " + "currently mounted\n"), cmdname, path); + return (1); + } + (void) fprintf(stderr, gettext("warning: %s not in mnttab\n"), + path); + if ((ret = umount2(path, flags)) != 0) + (void) fprintf(stderr, gettext("%s: %s\n"), path, + strerror(errno)); + return (ret != 0); } if (strcmp(entry.mnt_fstype, MNTTYPE_ZFS) != 0) { @@ -2560,39 +3555,58 @@ unshare_unmount_path(int op, char *path, int flags, boolean_t is_manual) ZFS_TYPE_FILESYSTEM)) == NULL) return (1); - verify(zfs_prop_get(zhp, op == OP_SHARE ? - ZFS_PROP_SHARENFS : ZFS_PROP_MOUNTPOINT, property, - sizeof (property), NULL, NULL, 0, B_FALSE) == 0); + ret = 1; + if (stat64(entry.mnt_mountp, &statbuf) != 0) { + (void) fprintf(stderr, gettext("cannot %s '%s': %s\n"), + cmdname, path, strerror(errno)); + goto out; + } else if (statbuf.st_ino != path_inode) { + (void) fprintf(stderr, gettext("cannot " + "%s '%s': not a mountpoint\n"), cmdname, path); + goto out; + } if (op == OP_SHARE) { - if (strcmp(property, "off") == 0) { + char nfs_mnt_prop[ZFS_MAXPROPLEN]; + char smbshare_prop[ZFS_MAXPROPLEN]; + + verify(zfs_prop_get(zhp, ZFS_PROP_SHARENFS, nfs_mnt_prop, + sizeof (nfs_mnt_prop), NULL, NULL, 0, B_FALSE) == 0); + verify(zfs_prop_get(zhp, ZFS_PROP_SHARESMB, smbshare_prop, + sizeof (smbshare_prop), NULL, NULL, 0, B_FALSE) == 0); + + if (strcmp(nfs_mnt_prop, "off") == 0 && + strcmp(smbshare_prop, "off") == 0) { (void) fprintf(stderr, gettext("cannot unshare " "'%s': legacy share\n"), path); (void) fprintf(stderr, gettext("use " "unshare(1M) to unshare this filesystem\n")); - ret = 1; - } else if (!zfs_is_shared_nfs(zhp, NULL)) { + } else if (!zfs_is_shared(zhp)) { (void) fprintf(stderr, gettext("cannot unshare '%s': " "not currently shared\n"), path); - ret = 1; } else { - ret = zfs_unshareall_nfs(zhp); + ret = zfs_unshareall_bypath(zhp, path); } } else { + char mtpt_prop[ZFS_MAXPROPLEN]; + + verify(zfs_prop_get(zhp, ZFS_PROP_MOUNTPOINT, mtpt_prop, + sizeof (mtpt_prop), NULL, NULL, 0, B_FALSE) == 0); + if (is_manual) { ret = zfs_unmount(zhp, NULL, flags); - } else if (strcmp(property, "legacy") == 0) { + } else if (strcmp(mtpt_prop, "legacy") == 0) { (void) fprintf(stderr, gettext("cannot unmount " "'%s': legacy mountpoint\n"), zfs_get_name(zhp)); (void) fprintf(stderr, gettext("use umount(1M) " "to unmount this filesystem\n")); - ret = 1; } else { ret = zfs_unmountall(zhp, flags); } } +out: zfs_close(zhp); return (ret != 0); @@ -2609,7 +3623,8 @@ unshare_unmount(int op, int argc, char **argv) int ret = 0; int types, c; zfs_handle_t *zhp; - char property[ZFS_MAXPROPLEN]; + char nfsiscsi_mnt_prop[ZFS_MAXPROPLEN]; + char sharesmb[ZFS_MAXPROPLEN]; /* check options */ while ((c = getopt(argc, argv, op == OP_SHARE ? "a" : "af")) != -1) { @@ -2695,18 +3710,35 @@ unshare_unmount(int op, int argc, char **argv) continue; } - verify(zfs_prop_get(zhp, op == OP_SHARE ? - ZFS_PROP_SHARENFS : ZFS_PROP_MOUNTPOINT, - property, sizeof (property), NULL, NULL, - 0, B_FALSE) == 0); - - /* Ignore legacy mounts and shares */ - if ((op == OP_SHARE && - strcmp(property, "off") == 0) || - (op == OP_MOUNT && - strcmp(property, "legacy") == 0)) { - zfs_close(zhp); - continue; + switch (op) { + case OP_SHARE: + verify(zfs_prop_get(zhp, ZFS_PROP_SHARENFS, + nfsiscsi_mnt_prop, + sizeof (nfsiscsi_mnt_prop), + NULL, NULL, 0, B_FALSE) == 0); + if (strcmp(nfsiscsi_mnt_prop, "off") != 0) + break; + verify(zfs_prop_get(zhp, ZFS_PROP_SHARESMB, + nfsiscsi_mnt_prop, + sizeof (nfsiscsi_mnt_prop), + NULL, NULL, 0, B_FALSE) == 0); + if (strcmp(nfsiscsi_mnt_prop, "off") == 0) + continue; + break; + case OP_MOUNT: + /* Ignore legacy mounts */ + verify(zfs_prop_get(zhp, ZFS_PROP_MOUNTPOINT, + nfsiscsi_mnt_prop, + sizeof (nfsiscsi_mnt_prop), + NULL, NULL, 0, B_FALSE) == 0); + if (strcmp(nfsiscsi_mnt_prop, "legacy") == 0) + continue; + /* Ignore canmount=noauto mounts */ + if (zfs_prop_get_int(zhp, ZFS_PROP_CANMOUNT) == + ZFS_CANMOUNT_NOAUTO) + continue; + default: + break; } node = safe_malloc(sizeof (unshare_unmount_node_t)); @@ -2746,7 +3778,7 @@ unshare_unmount(int op, int argc, char **argv) switch (op) { case OP_SHARE: - if (zfs_unshare_nfs(node->un_zhp, + if (zfs_unshareall_bypath(node->un_zhp, node->un_mountp) != 0) ret = 1; break; @@ -2774,7 +3806,8 @@ unshare_unmount(int op, int argc, char **argv) zfs_handle_t **dslist = NULL; size_t i, count = 0; - get_all_datasets(ZFS_TYPE_VOLUME, &dslist, &count); + get_all_datasets(ZFS_TYPE_VOLUME, &dslist, &count, + B_FALSE); if (count != 0) { qsort(dslist, count, sizeof (void *), @@ -2819,12 +3852,22 @@ unshare_unmount(int op, int argc, char **argv) if (zfs_get_type(zhp) == ZFS_TYPE_FILESYSTEM) { verify(zfs_prop_get(zhp, op == OP_SHARE ? - ZFS_PROP_SHARENFS : ZFS_PROP_MOUNTPOINT, property, - sizeof (property), NULL, NULL, 0, B_FALSE) == 0); + ZFS_PROP_SHARENFS : ZFS_PROP_MOUNTPOINT, + nfsiscsi_mnt_prop, sizeof (nfsiscsi_mnt_prop), NULL, + NULL, 0, B_FALSE) == 0); switch (op) { case OP_SHARE: - if (strcmp(property, "off") == 0) { + verify(zfs_prop_get(zhp, ZFS_PROP_SHARENFS, + nfsiscsi_mnt_prop, + sizeof (nfsiscsi_mnt_prop), + NULL, NULL, 0, B_FALSE) == 0); + verify(zfs_prop_get(zhp, ZFS_PROP_SHARESMB, + sharesmb, sizeof (sharesmb), NULL, NULL, + 0, B_FALSE) == 0); + + if (strcmp(nfsiscsi_mnt_prop, "off") == 0 && + strcmp(sharesmb, "off") == 0) { (void) fprintf(stderr, gettext("cannot " "unshare '%s': legacy share\n"), zfs_get_name(zhp)); @@ -2832,18 +3875,18 @@ unshare_unmount(int op, int argc, char **argv) "unshare(1M) to unshare this " "filesystem\n")); ret = 1; - } else if (!zfs_is_shared_nfs(zhp, NULL)) { + } else if (!zfs_is_shared(zhp)) { (void) fprintf(stderr, gettext("cannot " "unshare '%s': not currently " "shared\n"), zfs_get_name(zhp)); ret = 1; - } else if (zfs_unshareall_nfs(zhp) != 0) { + } else if (zfs_unshareall(zhp) != 0) { ret = 1; } break; case OP_MOUNT: - if (strcmp(property, "legacy") == 0) { + if (strcmp(nfsiscsi_mnt_prop, "legacy") == 0) { (void) fprintf(stderr, gettext("cannot " "unmount '%s': legacy " "mountpoint\n"), zfs_get_name(zhp)); @@ -2865,10 +3908,11 @@ unshare_unmount(int op, int argc, char **argv) } else { assert(op == OP_SHARE); - verify(zfs_prop_get(zhp, ZFS_PROP_SHAREISCSI, property, - sizeof (property), NULL, NULL, 0, B_FALSE) == 0); + verify(zfs_prop_get(zhp, ZFS_PROP_SHAREISCSI, + nfsiscsi_mnt_prop, sizeof (nfsiscsi_mnt_prop), + NULL, NULL, 0, B_FALSE) == 0); - if (strcmp(property, "off") == 0) { + if (strcmp(nfsiscsi_mnt_prop, "off") == 0) { (void) fprintf(stderr, gettext("cannot unshare " "'%s': 'shareiscsi' property not set\n"), zfs_get_name(zhp)); @@ -2948,9 +3992,6 @@ do_jail(int argc, char **argv, int attach) ret = (zfs_jail(zhp, jailid, attach) != 0); - if (!ret) - zpool_log_history(g_zfs, argc, argv, argv[2], B_FALSE, B_FALSE); - zfs_close(zhp); return (ret); } @@ -3141,6 +4182,23 @@ do_volcheck(boolean_t isinit) return (zpool_iter(g_zfs, volcheck, &isinit) ? 1 : 0); } +static int +find_command_idx(char *command, int *idx) +{ + int i; + + for (i = 0; i < NCOMMAND; i++) { + if (command_table[i].name == NULL) + continue; + + if (strcmp(command, command_table[i].name) == 0) { + *idx = i; + return (0); + } + } + return (1); +} + int main(int argc, char **argv) { @@ -3160,6 +4218,9 @@ main(int argc, char **argv) return (1); } + zpool_set_history_str("zfs", argc, argv, history_str); + verify(zpool_stage_history(g_zfs, history_str) == 0); + libzfs_print_on_error(g_zfs, B_TRUE); if ((mnttab_file = fopen(MNTTAB, "r")) == NULL) { @@ -3218,18 +4279,14 @@ main(int argc, char **argv) /* * Run the appropriate command. */ - for (i = 0; i < NCOMMAND; i++) { - if (command_table[i].name == NULL) - continue; - - if (strcmp(cmdname, command_table[i].name) == 0) { - current_command = &command_table[i]; - ret = command_table[i].func(argc - 1, argv + 1); - break; - } - } - - if (i == NCOMMAND) { + if (find_command_idx(cmdname, &i) == 0) { + current_command = &command_table[i]; + ret = command_table[i].func(argc - 1, argv + 1); + } else if (strchr(cmdname, '=') != NULL) { + verify(find_command_idx("set", &i) == 0); + current_command = &command_table[i]; + ret = command_table[i].func(argc, argv); + } else { (void) fprintf(stderr, gettext("unrecognized " "command '%s'\n"), cmdname); usage(B_FALSE); diff --git a/cddl/contrib/opensolaris/cmd/zinject/translate.c b/cddl/contrib/opensolaris/cmd/zinject/translate.c new file mode 100644 index 0000000..da26cd6 --- /dev/null +++ b/cddl/contrib/opensolaris/cmd/zinject/translate.c @@ -0,0 +1,460 @@ +/* + * CDDL HEADER START + * + * The contents of this file are subject to the terms of the + * Common Development and Distribution License (the "License"). + * You may not use this file except in compliance with the License. + * + * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE + * or http://www.opensolaris.org/os/licensing. + * See the License for the specific language governing permissions + * and limitations under the License. + * + * When distributing Covered Code, include this CDDL HEADER in each + * file and include the License file at usr/src/OPENSOLARIS.LICENSE. + * If applicable, add the following below this CDDL HEADER, with the + * fields enclosed by brackets "[]" replaced with your own identifying + * information: Portions Copyright [yyyy] [name of copyright owner] + * + * CDDL HEADER END + */ +/* + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. + * Use is subject to license terms. + */ + +#include <libzfs.h> + +#undef verify /* both libzfs.h and zfs_context.h want to define this */ + +#include <sys/zfs_context.h> + +#include <errno.h> +#include <fcntl.h> +#include <stdarg.h> +#include <stddef.h> +#include <stdio.h> +#include <stdlib.h> +#include <strings.h> +#include <sys/file.h> +#include <sys/mntent.h> +#include <sys/mnttab.h> +#include <sys/param.h> +#include <sys/stat.h> + +#include <sys/dmu.h> +#include <sys/dmu_objset.h> +#include <sys/dnode.h> +#include <sys/vdev_impl.h> + +#include "zinject.h" + +#include <assert.h> +#define verify assert + +extern void kernel_init(int); +extern void kernel_fini(void); + +static int debug; + +static void +ziprintf(const char *fmt, ...) +{ + va_list ap; + + if (!debug) + return; + + va_start(ap, fmt); + (void) vprintf(fmt, ap); + va_end(ap); +} + +/* + * Given a full path to a file, translate into a dataset name and a relative + * path within the dataset. 'dataset' must be at least MAXNAMELEN characters, + * and 'relpath' must be at least MAXPATHLEN characters. We also pass a stat64 + * buffer, which we need later to get the object ID. + */ +static int +parse_pathname(const char *fullpath, char *dataset, char *relpath, + struct stat64 *statbuf) +{ + struct statfs sfs; + const char *rel; + + if (fullpath[0] != '/') { + (void) fprintf(stderr, "invalid object '%s': must be full " + "path\n", fullpath); + usage(); + return (-1); + } + + if (strlen(fullpath) >= MAXPATHLEN) { + (void) fprintf(stderr, "invalid object; pathname too long\n"); + return (-1); + } + + if (stat64(fullpath, statbuf) != 0) { + (void) fprintf(stderr, "cannot open '%s': %s\n", + fullpath, strerror(errno)); + return (-1); + } + + if (statfs(fullpath, &sfs) == -1) { + (void) fprintf(stderr, "cannot find mountpoint for '%s': %s\n", + fullpath, strerror(errno)); + return (-1); + } + + if (strcmp(sfs.f_fstypename, MNTTYPE_ZFS) != 0) { + (void) fprintf(stderr, "invalid path '%s': not a ZFS " + "filesystem\n", fullpath); + return (-1); + } + + if (strncmp(fullpath, sfs.f_mntonname, strlen(sfs.f_mntonname)) != 0) { + (void) fprintf(stderr, "invalid path '%s': mountpoint " + "doesn't match path\n", fullpath); + return (-1); + } + + (void) strcpy(dataset, sfs.f_mntfromname); + + rel = fullpath + strlen(sfs.f_mntonname); + if (rel[0] == '/') + rel++; + (void) strcpy(relpath, rel); + + return (0); +} + +/* + * Convert from a (dataset, path) pair into a (objset, object) pair. Note that + * we grab the object number from the inode number, since looking this up via + * libzpool is a real pain. + */ +/* ARGSUSED */ +static int +object_from_path(const char *dataset, const char *path, struct stat64 *statbuf, + zinject_record_t *record) +{ + objset_t *os; + int err; + + /* + * Before doing any libzpool operations, call sync() to ensure that the + * on-disk state is consistent with the in-core state. + */ + sync(); + + if ((err = dmu_objset_open(dataset, DMU_OST_ZFS, + DS_MODE_USER | DS_MODE_READONLY, &os)) != 0) { + (void) fprintf(stderr, "cannot open dataset '%s': %s\n", + dataset, strerror(err)); + return (-1); + } + + record->zi_objset = dmu_objset_id(os); + record->zi_object = statbuf->st_ino; + + dmu_objset_close(os); + + return (0); +} + +/* + * Calculate the real range based on the type, level, and range given. + */ +static int +calculate_range(const char *dataset, err_type_t type, int level, char *range, + zinject_record_t *record) +{ + objset_t *os = NULL; + dnode_t *dn = NULL; + int err; + int ret = -1; + + /* + * Determine the numeric range from the string. + */ + if (range == NULL) { + /* + * If range is unspecified, set the range to [0,-1], which + * indicates that the whole object should be treated as an + * error. + */ + record->zi_start = 0; + record->zi_end = -1ULL; + } else { + char *end; + + /* XXX add support for suffixes */ + record->zi_start = strtoull(range, &end, 10); + + + if (*end == '\0') + record->zi_end = record->zi_start + 1; + else if (*end == ',') + record->zi_end = strtoull(end + 1, &end, 10); + + if (*end != '\0') { + (void) fprintf(stderr, "invalid range '%s': must be " + "a numeric range of the form 'start[,end]'\n", + range); + goto out; + } + } + + switch (type) { + case TYPE_DATA: + break; + + case TYPE_DNODE: + /* + * If this is a request to inject faults into the dnode, then we + * must translate the current (objset,object) pair into an + * offset within the metadnode for the objset. Specifying any + * kind of range with type 'dnode' is illegal. + */ + if (range != NULL) { + (void) fprintf(stderr, "range cannot be specified when " + "type is 'dnode'\n"); + goto out; + } + + record->zi_start = record->zi_object * sizeof (dnode_phys_t); + record->zi_end = record->zi_start + sizeof (dnode_phys_t); + record->zi_object = 0; + break; + } + + /* + * Get the dnode associated with object, so we can calculate the block + * size. + */ + if ((err = dmu_objset_open(dataset, DMU_OST_ANY, + DS_MODE_USER | DS_MODE_READONLY, &os)) != 0) { + (void) fprintf(stderr, "cannot open dataset '%s': %s\n", + dataset, strerror(err)); + goto out; + } + + if (record->zi_object == 0) { + dn = os->os->os_meta_dnode; + } else { + err = dnode_hold(os->os, record->zi_object, FTAG, &dn); + if (err != 0) { + (void) fprintf(stderr, "failed to hold dnode " + "for object %llu\n", + (u_longlong_t)record->zi_object); + goto out; + } + } + + + ziprintf("data shift: %d\n", (int)dn->dn_datablkshift); + ziprintf(" ind shift: %d\n", (int)dn->dn_indblkshift); + + /* + * Translate range into block IDs. + */ + if (record->zi_start != 0 || record->zi_end != -1ULL) { + record->zi_start >>= dn->dn_datablkshift; + record->zi_end >>= dn->dn_datablkshift; + } + + /* + * Check level, and then translate level 0 blkids into ranges + * appropriate for level of indirection. + */ + record->zi_level = level; + if (level > 0) { + ziprintf("level 0 blkid range: [%llu, %llu]\n", + record->zi_start, record->zi_end); + + if (level >= dn->dn_nlevels) { + (void) fprintf(stderr, "level %d exceeds max level " + "of object (%d)\n", level, dn->dn_nlevels - 1); + goto out; + } + + if (record->zi_start != 0 || record->zi_end != 0) { + int shift = dn->dn_indblkshift - SPA_BLKPTRSHIFT; + + for (; level > 0; level--) { + record->zi_start >>= shift; + record->zi_end >>= shift; + } + } + } + + ret = 0; +out: + if (dn) { + if (dn != os->os->os_meta_dnode) + dnode_rele(dn, FTAG); + } + if (os) + dmu_objset_close(os); + + return (ret); +} + +int +translate_record(err_type_t type, const char *object, const char *range, + int level, zinject_record_t *record, char *poolname, char *dataset) +{ + char path[MAXPATHLEN]; + char *slash; + struct stat64 statbuf; + int ret = -1; + + kernel_init(FREAD); + + debug = (getenv("ZINJECT_DEBUG") != NULL); + + ziprintf("translating: %s\n", object); + + if (MOS_TYPE(type)) { + /* + * MOS objects are treated specially. + */ + switch (type) { + case TYPE_MOS: + record->zi_type = 0; + break; + case TYPE_MOSDIR: + record->zi_type = DMU_OT_OBJECT_DIRECTORY; + break; + case TYPE_METASLAB: + record->zi_type = DMU_OT_OBJECT_ARRAY; + break; + case TYPE_CONFIG: + record->zi_type = DMU_OT_PACKED_NVLIST; + break; + case TYPE_BPLIST: + record->zi_type = DMU_OT_BPLIST; + break; + case TYPE_SPACEMAP: + record->zi_type = DMU_OT_SPACE_MAP; + break; + case TYPE_ERRLOG: + record->zi_type = DMU_OT_ERROR_LOG; + break; + } + + dataset[0] = '\0'; + (void) strcpy(poolname, object); + return (0); + } + + /* + * Convert a full path into a (dataset, file) pair. + */ + if (parse_pathname(object, dataset, path, &statbuf) != 0) + goto err; + + ziprintf(" dataset: %s\n", dataset); + ziprintf(" path: %s\n", path); + + /* + * Convert (dataset, file) into (objset, object) + */ + if (object_from_path(dataset, path, &statbuf, record) != 0) + goto err; + + ziprintf("raw objset: %llu\n", record->zi_objset); + ziprintf("raw object: %llu\n", record->zi_object); + + /* + * For the given object, calculate the real (type, level, range) + */ + if (calculate_range(dataset, type, level, (char *)range, record) != 0) + goto err; + + ziprintf(" objset: %llu\n", record->zi_objset); + ziprintf(" object: %llu\n", record->zi_object); + if (record->zi_start == 0 && + record->zi_end == -1ULL) + ziprintf(" range: all\n"); + else + ziprintf(" range: [%llu, %llu]\n", record->zi_start, + record->zi_end); + + /* + * Copy the pool name + */ + (void) strcpy(poolname, dataset); + if ((slash = strchr(poolname, '/')) != NULL) + *slash = '\0'; + + ret = 0; + +err: + kernel_fini(); + return (ret); +} + +int +translate_raw(const char *str, zinject_record_t *record) +{ + /* + * A raw bookmark of the form objset:object:level:blkid, where each + * number is a hexidecimal value. + */ + if (sscanf(str, "%llx:%llx:%x:%llx", (u_longlong_t *)&record->zi_objset, + (u_longlong_t *)&record->zi_object, &record->zi_level, + (u_longlong_t *)&record->zi_start) != 4) { + (void) fprintf(stderr, "bad raw spec '%s': must be of the form " + "'objset:object:level:blkid'\n", str); + return (-1); + } + + record->zi_end = record->zi_start; + + return (0); +} + +int +translate_device(const char *pool, const char *device, err_type_t label_type, + zinject_record_t *record) +{ + char *end; + zpool_handle_t *zhp; + nvlist_t *tgt; + boolean_t isspare, iscache; + + /* + * Given a device name or GUID, create an appropriate injection record + * with zi_guid set. + */ + if ((zhp = zpool_open(g_zfs, pool)) == NULL) + return (-1); + + record->zi_guid = strtoull(device, &end, 16); + if (record->zi_guid == 0 || *end != '\0') { + tgt = zpool_find_vdev(zhp, device, &isspare, &iscache, NULL); + + if (tgt == NULL) { + (void) fprintf(stderr, "cannot find device '%s' in " + "pool '%s'\n", device, pool); + return (-1); + } + + verify(nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID, + &record->zi_guid) == 0); + } + + switch (label_type) { + case TYPE_LABEL_UBERBLOCK: + record->zi_start = offsetof(vdev_label_t, vl_uberblock[0]); + record->zi_end = record->zi_start + VDEV_UBERBLOCK_RING - 1; + break; + case TYPE_LABEL_NVLIST: + record->zi_start = offsetof(vdev_label_t, vl_vdev_phys); + record->zi_end = record->zi_start + VDEV_PHYS_SIZE - 1; + break; + } + return (0); +} diff --git a/cddl/contrib/opensolaris/cmd/zinject/zinject.c b/cddl/contrib/opensolaris/cmd/zinject/zinject.c new file mode 100644 index 0000000..2302dc4 --- /dev/null +++ b/cddl/contrib/opensolaris/cmd/zinject/zinject.c @@ -0,0 +1,770 @@ +/* + * CDDL HEADER START + * + * The contents of this file are subject to the terms of the + * Common Development and Distribution License (the "License"). + * You may not use this file except in compliance with the License. + * + * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE + * or http://www.opensolaris.org/os/licensing. + * See the License for the specific language governing permissions + * and limitations under the License. + * + * When distributing Covered Code, include this CDDL HEADER in each + * file and include the License file at usr/src/OPENSOLARIS.LICENSE. + * If applicable, add the following below this CDDL HEADER, with the + * fields enclosed by brackets "[]" replaced with your own identifying + * information: Portions Copyright [yyyy] [name of copyright owner] + * + * CDDL HEADER END + */ +/* + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. + * Use is subject to license terms. + */ + +#pragma ident "%Z%%M% %I% %E% SMI" + +/* + * ZFS Fault Injector + * + * This userland component takes a set of options and uses libzpool to translate + * from a user-visible object type and name to an internal representation. + * There are two basic types of faults: device faults and data faults. + * + * + * DEVICE FAULTS + * + * Errors can be injected into a particular vdev using the '-d' option. This + * option takes a path or vdev GUID to uniquely identify the device within a + * pool. There are two types of errors that can be injected, EIO and ENXIO, + * that can be controlled through the '-e' option. The default is ENXIO. For + * EIO failures, any attempt to read data from the device will return EIO, but + * subsequent attempt to reopen the device will succeed. For ENXIO failures, + * any attempt to read from the device will return EIO, but any attempt to + * reopen the device will also return ENXIO. + * For label faults, the -L option must be specified. This allows faults + * to be injected into either the nvlist or uberblock region of all the labels + * for the specified device. + * + * This form of the command looks like: + * + * zinject -d device [-e errno] [-L <uber | nvlist>] pool + * + * + * DATA FAULTS + * + * We begin with a tuple of the form: + * + * <type,level,range,object> + * + * type A string describing the type of data to target. Each type + * implicitly describes how to interpret 'object'. Currently, + * the following values are supported: + * + * data User data for a file + * dnode Dnode for a file or directory + * + * The following MOS objects are special. Instead of injecting + * errors on a particular object or blkid, we inject errors across + * all objects of the given type. + * + * mos Any data in the MOS + * mosdir object directory + * config pool configuration + * bplist blkptr list + * spacemap spacemap + * metaslab metaslab + * errlog persistent error log + * + * level Object level. Defaults to '0', not applicable to all types. If + * a range is given, this corresponds to the indirect block + * corresponding to the specific range. + * + * range A numerical range [start,end) within the object. Defaults to + * the full size of the file. + * + * object A string describing the logical location of the object. For + * files and directories (currently the only supported types), + * this is the path of the object on disk. + * + * This is translated, via libzpool, into the following internal representation: + * + * <type,objset,object,level,range> + * + * These types should be self-explanatory. This tuple is then passed to the + * kernel via a special ioctl() to initiate fault injection for the given + * object. Note that 'type' is not strictly necessary for fault injection, but + * is used when translating existing faults into a human-readable string. + * + * + * The command itself takes one of the forms: + * + * zinject + * zinject <-a | -u pool> + * zinject -c <id|all> + * zinject [-q] <-t type> [-f freq] [-u] [-a] [-m] [-e errno] [-l level] + * [-r range] <object> + * zinject [-f freq] [-a] [-m] [-u] -b objset:object:level:start:end pool + * + * With no arguments, the command prints all currently registered injection + * handlers, with their numeric identifiers. + * + * The '-c' option will clear the given handler, or all handlers if 'all' is + * specified. + * + * The '-e' option takes a string describing the errno to simulate. This must + * be either 'io' or 'checksum'. In most cases this will result in the same + * behavior, but RAID-Z will produce a different set of ereports for this + * situation. + * + * The '-a', '-u', and '-m' flags toggle internal flush behavior. If '-a' is + * specified, then the ARC cache is flushed appropriately. If '-u' is + * specified, then the underlying SPA is unloaded. Either of these flags can be + * specified independently of any other handlers. The '-m' flag automatically + * does an unmount and remount of the underlying dataset to aid in flushing the + * cache. + * + * The '-f' flag controls the frequency of errors injected, expressed as a + * integer percentage between 1 and 100. The default is 100. + * + * The this form is responsible for actually injecting the handler into the + * framework. It takes the arguments described above, translates them to the + * internal tuple using libzpool, and then issues an ioctl() to register the + * handler. + * + * The final form can target a specific bookmark, regardless of whether a + * human-readable interface has been designed. It allows developers to specify + * a particular block by number. + */ + +#include <errno.h> +#include <fcntl.h> +#include <stdio.h> +#include <stdlib.h> +#include <strings.h> +#include <unistd.h> + +#include <sys/fs/zfs.h> +#include <sys/mount.h> + +#include <libzfs.h> + +#undef verify /* both libzfs.h and zfs_context.h want to define this */ + +#include "zinject.h" + +libzfs_handle_t *g_zfs; +int zfs_fd; + +#ifndef ECKSUM +#define ECKSUM EBADE +#endif + +static const char *errtable[TYPE_INVAL] = { + "data", + "dnode", + "mos", + "mosdir", + "metaslab", + "config", + "bplist", + "spacemap", + "errlog", + "uber", + "nvlist" +}; + +static err_type_t +name_to_type(const char *arg) +{ + int i; + for (i = 0; i < TYPE_INVAL; i++) + if (strcmp(errtable[i], arg) == 0) + return (i); + + return (TYPE_INVAL); +} + +static const char * +type_to_name(uint64_t type) +{ + switch (type) { + case DMU_OT_OBJECT_DIRECTORY: + return ("mosdir"); + case DMU_OT_OBJECT_ARRAY: + return ("metaslab"); + case DMU_OT_PACKED_NVLIST: + return ("config"); + case DMU_OT_BPLIST: + return ("bplist"); + case DMU_OT_SPACE_MAP: + return ("spacemap"); + case DMU_OT_ERROR_LOG: + return ("errlog"); + default: + return ("-"); + } +} + + +/* + * Print usage message. + */ +void +usage(void) +{ + (void) printf( + "usage:\n" + "\n" + "\tzinject\n" + "\n" + "\t\tList all active injection records.\n" + "\n" + "\tzinject -c <id|all>\n" + "\n" + "\t\tClear the particular record (if given a numeric ID), or\n" + "\t\tall records if 'all' is specificed.\n" + "\n" + "\tzinject -d device [-e errno] [-L <nvlist|uber>] pool\n" + "\t\tInject a fault into a particular device or the device's\n" + "\t\tlabel. Label injection can either be 'nvlist' or 'uber'.\n" + "\t\t'errno' can either be 'nxio' (the default) or 'io'.\n" + "\n" + "\tzinject -b objset:object:level:blkid pool\n" + "\n" + "\t\tInject an error into pool 'pool' with the numeric bookmark\n" + "\t\tspecified by the remaining tuple. Each number is in\n" + "\t\thexidecimal, and only one block can be specified.\n" + "\n" + "\tzinject [-q] <-t type> [-e errno] [-l level] [-r range]\n" + "\t [-a] [-m] [-u] [-f freq] <object>\n" + "\n" + "\t\tInject an error into the object specified by the '-t' option\n" + "\t\tand the object descriptor. The 'object' parameter is\n" + "\t\tinterperted depending on the '-t' option.\n" + "\n" + "\t\t-q\tQuiet mode. Only print out the handler number added.\n" + "\t\t-e\tInject a specific error. Must be either 'io' or\n" + "\t\t\t'checksum'. Default is 'io'.\n" + "\t\t-l\tInject error at a particular block level. Default is " + "0.\n" + "\t\t-m\tAutomatically remount underlying filesystem.\n" + "\t\t-r\tInject error over a particular logical range of an\n" + "\t\t\tobject. Will be translated to the appropriate blkid\n" + "\t\t\trange according to the object's properties.\n" + "\t\t-a\tFlush the ARC cache. Can be specified without any\n" + "\t\t\tassociated object.\n" + "\t\t-u\tUnload the associated pool. Can be specified with only\n" + "\t\t\ta pool object.\n" + "\t\t-f\tOnly inject errors a fraction of the time. Expressed as\n" + "\t\t\ta percentage between 1 and 100.\n" + "\n" + "\t-t data\t\tInject an error into the plain file contents of a\n" + "\t\t\tfile. The object must be specified as a complete path\n" + "\t\t\tto a file on a ZFS filesystem.\n" + "\n" + "\t-t dnode\tInject an error into the metadnode in the block\n" + "\t\t\tcorresponding to the dnode for a file or directory. The\n" + "\t\t\t'-r' option is incompatible with this mode. The object\n" + "\t\t\tis specified as a complete path to a file or directory\n" + "\t\t\ton a ZFS filesystem.\n" + "\n" + "\t-t <mos>\tInject errors into the MOS for objects of the given\n" + "\t\t\ttype. Valid types are: mos, mosdir, config, bplist,\n" + "\t\t\tspacemap, metaslab, errlog. The only valid <object> is\n" + "\t\t\tthe poolname.\n"); +} + +static int +iter_handlers(int (*func)(int, const char *, zinject_record_t *, void *), + void *data) +{ + zfs_cmd_t zc; + int ret; + + zc.zc_guid = 0; + + while (ioctl(zfs_fd, ZFS_IOC_INJECT_LIST_NEXT, &zc) == 0) + if ((ret = func((int)zc.zc_guid, zc.zc_name, + &zc.zc_inject_record, data)) != 0) + return (ret); + + return (0); +} + +static int +print_data_handler(int id, const char *pool, zinject_record_t *record, + void *data) +{ + int *count = data; + + if (record->zi_guid != 0) + return (0); + + if (*count == 0) { + (void) printf("%3s %-15s %-6s %-6s %-8s %3s %-15s\n", + "ID", "POOL", "OBJSET", "OBJECT", "TYPE", "LVL", "RANGE"); + (void) printf("--- --------------- ------ " + "------ -------- --- ---------------\n"); + } + + *count += 1; + + (void) printf("%3d %-15s %-6llu %-6llu %-8s %3d ", id, pool, + (u_longlong_t)record->zi_objset, (u_longlong_t)record->zi_object, + type_to_name(record->zi_type), record->zi_level); + + if (record->zi_start == 0 && + record->zi_end == -1ULL) + (void) printf("all\n"); + else + (void) printf("[%llu, %llu]\n", (u_longlong_t)record->zi_start, + (u_longlong_t)record->zi_end); + + return (0); +} + +static int +print_device_handler(int id, const char *pool, zinject_record_t *record, + void *data) +{ + int *count = data; + + if (record->zi_guid == 0) + return (0); + + if (*count == 0) { + (void) printf("%3s %-15s %s\n", "ID", "POOL", "GUID"); + (void) printf("--- --------------- ----------------\n"); + } + + *count += 1; + + (void) printf("%3d %-15s %llx\n", id, pool, + (u_longlong_t)record->zi_guid); + + return (0); +} + +/* + * Print all registered error handlers. Returns the number of handlers + * registered. + */ +static int +print_all_handlers(void) +{ + int count = 0; + + (void) iter_handlers(print_device_handler, &count); + (void) printf("\n"); + count = 0; + (void) iter_handlers(print_data_handler, &count); + + return (count); +} + +/* ARGSUSED */ +static int +cancel_one_handler(int id, const char *pool, zinject_record_t *record, + void *data) +{ + zfs_cmd_t zc; + + zc.zc_guid = (uint64_t)id; + + if (ioctl(zfs_fd, ZFS_IOC_CLEAR_FAULT, &zc) != 0) { + (void) fprintf(stderr, "failed to remove handler %d: %s\n", + id, strerror(errno)); + return (1); + } + + return (0); +} + +/* + * Remove all fault injection handlers. + */ +static int +cancel_all_handlers(void) +{ + int ret = iter_handlers(cancel_one_handler, NULL); + + (void) printf("removed all registered handlers\n"); + + return (ret); +} + +/* + * Remove a specific fault injection handler. + */ +static int +cancel_handler(int id) +{ + zfs_cmd_t zc; + + zc.zc_guid = (uint64_t)id; + + if (ioctl(zfs_fd, ZFS_IOC_CLEAR_FAULT, &zc) != 0) { + (void) fprintf(stderr, "failed to remove handler %d: %s\n", + id, strerror(errno)); + return (1); + } + + (void) printf("removed handler %d\n", id); + + return (0); +} + +/* + * Register a new fault injection handler. + */ +static int +register_handler(const char *pool, int flags, zinject_record_t *record, + int quiet) +{ + zfs_cmd_t zc; + + (void) strcpy(zc.zc_name, pool); + zc.zc_inject_record = *record; + zc.zc_guid = flags; + + if (ioctl(zfs_fd, ZFS_IOC_INJECT_FAULT, &zc) != 0) { + (void) fprintf(stderr, "failed to add handler: %s\n", + strerror(errno)); + return (1); + } + + if (flags & ZINJECT_NULL) + return (0); + + if (quiet) { + (void) printf("%llu\n", (u_longlong_t)zc.zc_guid); + } else { + (void) printf("Added handler %llu with the following " + "properties:\n", (u_longlong_t)zc.zc_guid); + (void) printf(" pool: %s\n", pool); + if (record->zi_guid) { + (void) printf(" vdev: %llx\n", + (u_longlong_t)record->zi_guid); + } else { + (void) printf("objset: %llu\n", + (u_longlong_t)record->zi_objset); + (void) printf("object: %llu\n", + (u_longlong_t)record->zi_object); + (void) printf(" type: %llu\n", + (u_longlong_t)record->zi_type); + (void) printf(" level: %d\n", record->zi_level); + if (record->zi_start == 0 && + record->zi_end == -1ULL) + (void) printf(" range: all\n"); + else + (void) printf(" range: [%llu, %llu)\n", + (u_longlong_t)record->zi_start, + (u_longlong_t)record->zi_end); + } + } + + return (0); +} + +int +main(int argc, char **argv) +{ + int c; + char *range = NULL; + char *cancel = NULL; + char *end; + char *raw = NULL; + char *device = NULL; + int level = 0; + int quiet = 0; + int error = 0; + int domount = 0; + err_type_t type = TYPE_INVAL; + err_type_t label = TYPE_INVAL; + zinject_record_t record = { 0 }; + char pool[MAXNAMELEN]; + char dataset[MAXNAMELEN]; + zfs_handle_t *zhp; + int ret; + int flags = 0; + + if ((g_zfs = libzfs_init()) == NULL) { + (void) fprintf(stderr, "internal error: failed to " + "initialize ZFS library\n"); + return (1); + } + + libzfs_print_on_error(g_zfs, B_TRUE); + + if ((zfs_fd = open(ZFS_DEV, O_RDWR)) < 0) { + (void) fprintf(stderr, "failed to open ZFS device\n"); + return (1); + } + + if (argc == 1) { + /* + * No arguments. Print the available handlers. If there are no + * available handlers, direct the user to '-h' for help + * information. + */ + if (print_all_handlers() == 0) { + (void) printf("No handlers registered.\n"); + (void) printf("Run 'zinject -h' for usage " + "information.\n"); + } + + return (0); + } + + while ((c = getopt(argc, argv, ":ab:d:f:qhc:t:l:mr:e:uL:")) != -1) { + switch (c) { + case 'a': + flags |= ZINJECT_FLUSH_ARC; + break; + case 'b': + raw = optarg; + break; + case 'c': + cancel = optarg; + break; + case 'd': + device = optarg; + break; + case 'e': + if (strcasecmp(optarg, "io") == 0) { + error = EIO; + } else if (strcasecmp(optarg, "checksum") == 0) { + error = ECKSUM; + } else if (strcasecmp(optarg, "nxio") == 0) { + error = ENXIO; + } else { + (void) fprintf(stderr, "invalid error type " + "'%s': must be 'io', 'checksum' or " + "'nxio'\n", optarg); + usage(); + return (1); + } + break; + case 'f': + record.zi_freq = atoi(optarg); + if (record.zi_freq < 1 || record.zi_freq > 100) { + (void) fprintf(stderr, "frequency range must " + "be in the range (0, 100]\n"); + return (1); + } + break; + case 'h': + usage(); + return (0); + case 'l': + level = (int)strtol(optarg, &end, 10); + if (*end != '\0') { + (void) fprintf(stderr, "invalid level '%s': " + "must be an integer\n", optarg); + usage(); + return (1); + } + break; + case 'm': + domount = 1; + break; + case 'q': + quiet = 1; + break; + case 'r': + range = optarg; + break; + case 't': + if ((type = name_to_type(optarg)) == TYPE_INVAL && + !MOS_TYPE(type)) { + (void) fprintf(stderr, "invalid type '%s'\n", + optarg); + usage(); + return (1); + } + break; + case 'u': + flags |= ZINJECT_UNLOAD_SPA; + break; + case 'L': + if ((label = name_to_type(optarg)) == TYPE_INVAL && + !LABEL_TYPE(type)) { + (void) fprintf(stderr, "invalid label type " + "'%s'\n", optarg); + usage(); + return (1); + } + break; + case ':': + (void) fprintf(stderr, "option -%c requires an " + "operand\n", optopt); + usage(); + return (1); + case '?': + (void) fprintf(stderr, "invalid option '%c'\n", + optopt); + usage(); + return (2); + } + } + + argc -= optind; + argv += optind; + + if (cancel != NULL) { + /* + * '-c' is invalid with any other options. + */ + if (raw != NULL || range != NULL || type != TYPE_INVAL || + level != 0) { + (void) fprintf(stderr, "cancel (-c) incompatible with " + "any other options\n"); + usage(); + return (2); + } + if (argc != 0) { + (void) fprintf(stderr, "extraneous argument to '-c'\n"); + usage(); + return (2); + } + + if (strcmp(cancel, "all") == 0) { + return (cancel_all_handlers()); + } else { + int id = (int)strtol(cancel, &end, 10); + if (*end != '\0') { + (void) fprintf(stderr, "invalid handle id '%s':" + " must be an integer or 'all'\n", cancel); + usage(); + return (1); + } + return (cancel_handler(id)); + } + } + + if (device != NULL) { + /* + * Device (-d) injection uses a completely different mechanism + * for doing injection, so handle it separately here. + */ + if (raw != NULL || range != NULL || type != TYPE_INVAL || + level != 0) { + (void) fprintf(stderr, "device (-d) incompatible with " + "data error injection\n"); + usage(); + return (2); + } + + if (argc != 1) { + (void) fprintf(stderr, "device (-d) injection requires " + "a single pool name\n"); + usage(); + return (2); + } + + (void) strcpy(pool, argv[0]); + dataset[0] = '\0'; + + if (error == ECKSUM) { + (void) fprintf(stderr, "device error type must be " + "'io' or 'nxio'\n"); + return (1); + } + + if (translate_device(pool, device, label, &record) != 0) + return (1); + if (!error) + error = ENXIO; + } else if (raw != NULL) { + if (range != NULL || type != TYPE_INVAL || level != 0) { + (void) fprintf(stderr, "raw (-b) format with " + "any other options\n"); + usage(); + return (2); + } + + if (argc != 1) { + (void) fprintf(stderr, "raw (-b) format expects a " + "single pool name\n"); + usage(); + return (2); + } + + (void) strcpy(pool, argv[0]); + dataset[0] = '\0'; + + if (error == ENXIO) { + (void) fprintf(stderr, "data error type must be " + "'checksum' or 'io'\n"); + return (1); + } + + if (translate_raw(raw, &record) != 0) + return (1); + if (!error) + error = EIO; + } else if (type == TYPE_INVAL) { + if (flags == 0) { + (void) fprintf(stderr, "at least one of '-b', '-d', " + "'-t', '-a', or '-u' must be specified\n"); + usage(); + return (2); + } + + if (argc == 1 && (flags & ZINJECT_UNLOAD_SPA)) { + (void) strcpy(pool, argv[0]); + dataset[0] = '\0'; + } else if (argc != 0) { + (void) fprintf(stderr, "extraneous argument for " + "'-f'\n"); + usage(); + return (2); + } + + flags |= ZINJECT_NULL; + } else { + if (argc != 1) { + (void) fprintf(stderr, "missing object\n"); + usage(); + return (2); + } + + if (error == ENXIO) { + (void) fprintf(stderr, "data error type must be " + "'checksum' or 'io'\n"); + return (1); + } + + if (translate_record(type, argv[0], range, level, &record, pool, + dataset) != 0) + return (1); + if (!error) + error = EIO; + } + + /* + * If this is pool-wide metadata, unmount everything. The ioctl() will + * unload the pool, so that we trigger spa-wide reopen of metadata next + * time we access the pool. + */ + if (dataset[0] != '\0' && domount) { + if ((zhp = zfs_open(g_zfs, dataset, ZFS_TYPE_DATASET)) == NULL) + return (1); + + if (zfs_unmount(zhp, NULL, 0) != 0) + return (1); + } + + record.zi_error = error; + + ret = register_handler(pool, flags, &record, quiet); + + if (dataset[0] != '\0' && domount) + ret = (zfs_mount(zhp, NULL, 0) != 0); + + libzfs_fini(g_zfs); + + return (ret); +} diff --git a/cddl/contrib/opensolaris/cmd/zinject/zinject.h b/cddl/contrib/opensolaris/cmd/zinject/zinject.h new file mode 100644 index 0000000..adc3efe --- /dev/null +++ b/cddl/contrib/opensolaris/cmd/zinject/zinject.h @@ -0,0 +1,71 @@ +/* + * CDDL HEADER START + * + * The contents of this file are subject to the terms of the + * Common Development and Distribution License (the "License"). + * You may not use this file except in compliance with the License. + * + * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE + * or http://www.opensolaris.org/os/licensing. + * See the License for the specific language governing permissions + * and limitations under the License. + * + * When distributing Covered Code, include this CDDL HEADER in each + * file and include the License file at usr/src/OPENSOLARIS.LICENSE. + * If applicable, add the following below this CDDL HEADER, with the + * fields enclosed by brackets "[]" replaced with your own identifying + * information: Portions Copyright [yyyy] [name of copyright owner] + * + * CDDL HEADER END + */ +/* + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. + * Use is subject to license terms. + */ + +#ifndef _ZINJECT_H +#define _ZINJECT_H + +#pragma ident "%Z%%M% %I% %E% SMI" + +#include <sys/zfs_ioctl.h> + +#ifdef __cplusplus +extern "C" { +#endif + +typedef enum { + TYPE_DATA, /* plain file contents */ + TYPE_DNODE, /* metadnode contents */ + TYPE_MOS, /* all MOS data */ + TYPE_MOSDIR, /* MOS object directory */ + TYPE_METASLAB, /* metaslab objects */ + TYPE_CONFIG, /* MOS config */ + TYPE_BPLIST, /* block pointer list */ + TYPE_SPACEMAP, /* space map objects */ + TYPE_ERRLOG, /* persistent error log */ + TYPE_LABEL_UBERBLOCK, /* label specific uberblock */ + TYPE_LABEL_NVLIST, /* label specific nvlist */ + TYPE_INVAL +} err_type_t; + +#define MOS_TYPE(t) \ + ((t) >= TYPE_MOS && (t) < TYPE_LABEL_UBERBLOCK) + +#define LABEL_TYPE(t) \ + ((t) >= TYPE_LABEL_UBERBLOCK && (t) < TYPE_INVAL) + +int translate_record(err_type_t type, const char *object, const char *range, + int level, zinject_record_t *record, char *poolname, char *dataset); +int translate_raw(const char *raw, zinject_record_t *record); +int translate_device(const char *pool, const char *device, + err_type_t label_type, zinject_record_t *record); +void usage(void); + +extern libzfs_handle_t *g_zfs; + +#ifdef __cplusplus +} +#endif + +#endif /* _ZINJECT_H */ diff --git a/cddl/contrib/opensolaris/cmd/zpool/zpool.8 b/cddl/contrib/opensolaris/cmd/zpool/zpool.8 index 95ae008..a7967d7 100644 --- a/cddl/contrib/opensolaris/cmd/zpool/zpool.8 +++ b/cddl/contrib/opensolaris/cmd/zpool/zpool.8 @@ -17,8 +17,8 @@ .\" information: Portions Copyright [yyyy] [name of copyright owner] .\" .\" CDDL HEADER END -.\" Copyright (c) 2006, Sun Microsystems, Inc. All Rights Reserved. -.TH zpool 1M "14 Nov 2006" "SunOS 5.11" "System Administration Commands" +.\" Copyright (c) 2007, Sun Microsystems, Inc. All Rights Reserved. +.TH zpool 1M "13 Nov 2007" "SunOS 5.11" "System Administration Commands" .SH NAME zpool \- configures ZFS storage pools .SH SYNOPSIS @@ -29,7 +29,8 @@ zpool \- configures ZFS storage pools .LP .nf -\fBzpool create\fR [\fB-fn\fR] [\fB-R\fR \fIroot\fR] [\fB-m\fR \fImountpoint\fR] \fIpool\fR \fIvdev ...\fR +\fBzpool create\fR [\fB-fn\fR] [\fB-o\fR \fIproperty=value\fR] ... [\fB-m\fR \fImountpoint\fR] [\fB-R\fR \fIroot\fR] + \fIpool\fR \fIvdev\fR ... .fi .LP @@ -39,22 +40,22 @@ zpool \- configures ZFS storage pools .LP .nf -\fBzpool add\fR [\fB-fn\fR] \fIpool\fR \fIvdev\fR +\fBzpool add\fR [\fB-fn\fR] \fIpool\fR \fIvdev\fR ... .fi .LP .nf -\fBzpool remove\fR \fIpool\fR \fIvdev\fR +\fBzpool remove\fR \fIpool\fR \fIdevice\fR ... .fi .LP .nf -\fBzpool \fR \fBlist\fR [\fB-H\fR] [\fB-o\fR \fIfield\fR[,\fIfield\fR]*] [\fIpool\fR] ... +\fBzpool list\fR [\fB-H\fR] [\fB-o\fR \fIproperty\fR[,...]] [\fIpool\fR] ... .fi .LP .nf -\fBzpool iostat\fR [\fB-v\fR] [\fIpool\fR] ... [\fIinterval\fR [\fIcount\fR]] +\fBzpool iostat\fR [\fB-v\fR] [\fIpool\fR] ... [\fIinterval\fR[\fIcount\fR]] .fi .LP @@ -64,17 +65,17 @@ zpool \- configures ZFS storage pools .LP .nf -\fBzpool offline\fR [\fB-t\fR] \fIpool\fR \fIdevice\fR ... +\fBzpool online\fR \fIpool\fR \fIdevice\fR ... .fi .LP .nf -\fBzpool online\fR \fIpool\fR \fIdevice\fR ... +\fBzpool offline\fR [\fB-t\fR] \fIpool\fR \fIdevice\fR ... .fi .LP .nf -\fBzpool clear\fR \fIpool\fR [\fIdevice\fR] ... +\fBzpool clear\fR \fIpool\fR [\fIdevice\fR] .fi .LP @@ -99,23 +100,24 @@ zpool \- configures ZFS storage pools .LP .nf -\fBzpool export\fR [\fB-f\fR] \fIpool\fR +\fBzpool import\fR [\fB-d\fR \fIdir\fR] [\fB-D\fR] .fi .LP .nf -\fBzpool import\fR [\fB-d\fR \fIdir\fR] [\fB-D\fR] +\fBzpool import\fR [\fB-o \fImntopts\fR\fR] [\fB-p\fR \fIproperty=value\fR] ... [\fB-d\fR \fIdir\fR | \fB-c\fR \fIcachefile\fR] + [\fB-D\fR] [\fB-f\fR] [\fB-R\fR \fIroot\fR] \fB-a\fR .fi .LP .nf -\fBzpool import\fR [\fB-d\fR \fIdir\fR] [\fB-D\fR] [\fB-f\fR] [\fB-o \fIopts\fR\fR] [\fB-R \fR\fIroot\fR] \fIpool\fR | \fIid\fR - [\fInewpool\fR] +\fBzpool import\fR [\fB-o \fImntopts\fR\fR] [\fB-o\fR \fIproperty=value\fR] ... [\fB-d\fR \fIdir\fR | \fB-c\fR \fIcachefile\fR] + [\fB-D\fR] [\fB-f\fR] [\fB-R\fR \fIroot\fR] \fIpool\fR |\fIid\fR [\fInewpool\fR] .fi .LP .nf -\fBzpool import\fR [\fB-d\fR \fIdir\fR] [\fB-D\fR] [\fB-f\fR] [\fB-a\fR] +\fBzpool export\fR [\fB-f\fR] \fIpool\fR ... .fi .LP @@ -130,20 +132,33 @@ zpool \- configures ZFS storage pools .LP .nf -\fBzpool upgrade\fR [\fB-a\fR | \fIpool\fR] +\fBzpool upgrade\fR [\fB-V\fR \fIversion\fR] \fB-a\fR | \fIpool\fR ... +.fi + +.LP +.nf +\fBzpool history\fR [\fB-il\fR] [\fIpool\fR] ... +.fi + +.LP +.nf +\fBzpool get\fR "\fIall\fR" | \fIproperty\fR[,...] \fIpool\fR ... .fi .LP .nf -\fBzpool history\fR [\fIpool\fR] ... +\fBzpool set\fR \fIproperty\fR=\fIvalue\fR \fIpool\fR .fi .SH DESCRIPTION +.sp .LP The \fBzpool\fR command configures \fBZFS\fR storage pools. A storage pool is a collection of devices that provides physical storage and data replication for \fBZFS\fR datasets. +.sp .LP All datasets within a storage pool share the same space. See \fBzfs\fR(1M) for information on managing datasets. .SS "Virtual Devices (vdevs)" +.sp .LP A "virtual device" describes a single device or a collection of devices organized according to certain performance and fault characteristics. The following virtual devices are supported: .sp @@ -202,8 +217,8 @@ A variation on \fBRAID-5\fR that allows for better distribution of parity and el A \fBraidz\fR group can have either single- or double-parity, meaning that the \fBraidz\fR group can sustain one or two failures respectively without losing any data. The \fBraidz1\fR \fBvdev\fR type specifies a single-parity \fBraidz\fR group and the \fBraidz2\fR \fBvdev\fR type specifies a double-parity \fBraidz\fR group. The \fBraidz\fR \fBvdev\fR type is an alias for \fBraidz1\fR. .sp -A \fBraidz\fR group with \fIN\fR disks of size \fIX\fR with \fIP\fR parity disks can hold approximately (\fIN-P\fR)*\fIX\fR bytes and can withstand one device failing before -data integrity is compromised. The minimum number of devices in a \fBraidz\fR group is one more than the number of parity disks. The recommended number is between 3 and 9. +A \fBraidz\fR group with \fIN\fR disks of size \fIX\fR with \fIP\fR parity disks can hold approximately (\fIN-P\fR)*\fIX\fR bytes and can withstand \fIP\fR device(s) +failing before data integrity is compromised. The minimum number of devices in a \fBraidz\fR group is one more than the number of parity disks. The recommended number is between 3 and 9 to help increase performance. .RE .sp @@ -217,11 +232,37 @@ data integrity is compromised. The minimum number of devices in a \fBraidz\fR gr A special pseudo-\fBvdev\fR which keeps track of available hot spares for a pool. For more information, see the "Hot Spares" section. .RE +.sp +.ne 2 +.mk +.na +\fBlog\fR +.ad +.RS 10n +.rt +A separate intent log device. If more than one log device is specified, then writes are load-balanced between devices. Log devices can be mirrored. However, \fBraidz\fR and \fBraidz2\fR are not supported for the intent log. For more information, see the "Intent +Log" section. +.RE + +.sp +.ne 2 +.mk +.na +\fBcache\fR +.ad +.RS 10n +.rt +A device used to cache storage pool data. A cache device cannot be mirrored or part of a \fBraidz\fR or \fBraidz2\fR configuration. For more information, see the "Cache Devices" section. +.RE + +.sp .LP Virtual devices cannot be nested, so a mirror or \fBraidz\fR virtual device can only contain files or disks. Mirrors of mirrors (or other combinations) are not allowed. +.sp .LP A pool can have any number of virtual devices at the top of the configuration (known as "root vdevs"). Data is dynamically distributed across all top-level devices to balance data among devices. As new virtual devices are added, \fBZFS\fR automatically places data on the newly available devices. +.sp .LP Virtual devices are specified one at a time on the command line, separated by whitespace. The keywords "mirror" and "raidz" are used to distinguish where a group ends and another begins. For example, the following creates two root vdevs, each a mirror of two disks: .sp @@ -233,15 +274,119 @@ Virtual devices are specified one at a time on the command line, separated by wh .sp .SS "Device Failure and Recovery" +.sp .LP \fBZFS\fR supports a rich set of mechanisms for handling device failure and data corruption. All metadata and data is checksummed, and \fBZFS\fR automatically repairs bad data from a good copy when corruption is detected. +.sp .LP In order to take advantage of these features, a pool must make use of some form of redundancy, using either mirrored or \fBraidz\fR groups. While \fBZFS\fR supports running in a non-redundant configuration, where each root vdev is simply a disk or file, this is strongly discouraged. A single case of bit corruption can render some or all of your data unavailable. +.sp .LP A pool's health status is described by one of three states: online, degraded, or faulted. An online pool has all devices operating normally. A degraded pool is one in which one or more devices have failed, but the data is still available due to a redundant configuration. A faulted pool has -one or more failed devices, and there is insufficient redundancy to replicate the missing data. +corrupted metadata, or one or more faulted devices, and insufficient replicas to continue functioning. +.sp +.LP +The health of the top-level vdev, such as mirror or \fBraidz\fR device, is potentially impacted by the state of its associated vdevs, or component devices. A top-level vdev or component device is in one of the following states: +.sp +.ne 2 +.mk +.na +\fB\fBDEGRADED\fR\fR +.ad +.RS 12n +.rt +One or more top-level vdevs is in the degraded state because one or more component devices are offline. Sufficient replicas exist to continue functioning. +.sp +One or more component devices is in the degraded or faulted state, but sufficient replicas exist to continue functioning. The underlying conditions are as follows: +.RS +4 +.TP +.ie t \(bu +.el o +The number of checksum errors exceeds acceptable levels and the device is degraded as an indication that something may be wrong. \fBZFS\fR continues to use the device as necessary. +.RE +.RS +4 +.TP +.ie t \(bu +.el o +The number of I/O errors exceeds acceptable levels. The device could not be marked as faulted because there are insufficient replicas to continue functioning. +.RE +.RE + +.sp +.ne 2 +.mk +.na +\fB\fBFAULTED\fR\fR +.ad +.RS 12n +.rt +One or more top-level vdevs is in the faulted state because one or more component devices are offline. Insufficient replicas exist to continue functioning. +.sp +One or more component devices is in the faulted state, and insufficient replicas exist to continue functioning. The underlying conditions are as follows: +.RS +4 +.TP +.ie t \(bu +.el o +The device could be opened, but the contents did not match expected values. +.RE +.RS +4 +.TP +.ie t \(bu +.el o +The number of I/O errors exceeds acceptable levels and the device is faulted to prevent further use of the device. +.RE +.RE + +.sp +.ne 2 +.mk +.na +\fB\fBOFFLINE\fR\fR +.ad +.RS 12n +.rt +The device was explicitly taken offline by the "\fBzpool offline\fR" command. +.RE + +.sp +.ne 2 +.mk +.na +\fB\fBONLINE\fR\fR +.ad +.RS 12n +.rt +The device is online and functioning. +.RE + +.sp +.ne 2 +.mk +.na +\fB\fBREMOVED\fR\fR +.ad +.RS 12n +.rt +The device was physically removed while the system was running. Device removal detection is hardware-dependent and may not be supported on all platforms. +.RE + +.sp +.ne 2 +.mk +.na +\fB\fBUNAVAIL\fR\fR +.ad +.RS 12n +.rt +The device could not be opened. If a pool is imported when a device was unavailable, then the device will be identified by a unique identifier instead of its path since the path was never correct in the first place. +.RE + +.sp +.LP +If a device is removed and later re-attached to the system, \fBZFS\fR attempts to put the device online automatically. Device attach detection is hardware-dependent and might not be supported on all platforms. .SS "Hot Spares" +.sp .LP \fBZFS\fR allows devices to be associated with pools as "hot spares". These devices are not actively used in the pool, but when an active device fails, it is automatically replaced by a hot spare. To create a pool with hot spares, specify a "spare" \fBvdev\fR with any number of devices. For example, .sp @@ -252,22 +397,261 @@ one or more failed devices, and there is insufficient redundancy to replicate th .in -2 .sp +.sp .LP -Spares can be shared across multiple pools, and can be added with the "zpool add" command and removed with the "zpool remove" command. Once a spare replacement is initiated, a new "spare" \fBvdev\fR is created within the configuration that -will remain there until the original device is replaced. At this point, the hot spare becomes available again if another device fails. +Spares can be shared across multiple pools, and can be added with the "\fBzpool add\fR" command and removed with the "\fBzpool remove\fR" command. Once a spare replacement is initiated, a new "spare" \fBvdev\fR is +created within the configuration that will remain there until the original device is replaced. At this point, the hot spare becomes available again if another device fails. +.sp .LP An in-progress spare replacement can be cancelled by detaching the hot spare. If the original faulted device is detached, then the hot spare assumes its place in the configuration, and is removed from the spare list of all active pools. -.SS "Alternate Root Pools" +.sp +.LP +Spares cannot replace log devices. +.SS "Intent Log" +.sp +.LP +The \fBZFS\fR Intent Log (\fBZIL\fR) satisfies \fBPOSIX\fR requirements for synchronous transactions. For instance, databases often require their transactions to be on stable storage devices when returning from a system call. \fBNFS\fR and +other applications can also use \fBfsync\fR() to ensure data stability. By default, the intent log is allocated from blocks within the main pool. However, it might be possible to get better performance using separate intent log devices such as \fBNVRAM\fR or a dedicated +disk. For example: +.sp +.in +2 +.nf +\fB# zpool create pool c0d0 c1d0 log c2d0\fR +.fi +.in -2 +.sp + +.sp +.LP +Multiple log devices can also be specified, and they can be mirrored. See the EXAMPLES section for an example of mirroring multiple log devices. +.sp +.LP +Log devices can be added, replaced, attached, detached, and imported and exported as part of the larger pool. +.SS "Cache Devices" +.sp +.LP +Devices can be added to a storage pool as "cache devices." These devices provide an additional layer of caching between main memory and disk. For read-heavy workloads, where the working set size is much larger than what can be cached in main memory, using cache devices allow +much more of this working set to be served from low latency media. Using cache devices provides the greatest performance improvement for random read-workloads of mostly static content. +.sp +.LP +To create a pool with cache devices, specify a "cache" \fBvdev\fR with any number of devices. For example: +.sp +.in +2 +.nf +\fB# zpool create pool c0d0 c1d0 cache c2d0 c3d0\fR +.fi +.in -2 +.sp + +.sp +.LP +Cache devices cannot be mirrored or part of a \fBraidz\fR configuration. If a read error is encountered on a cache device, that read \fBI/O\fR is reissued to the original storage pool device, which might be part of a mirrored or \fBraidz\fR configuration. +.sp +.LP +The content of the cache devices is considered volatile, as is the case with other system caches. +.SS "Properties" +.sp +.LP +Each pool has several properties associated with it. Some properties are read-only statistics while others are configurable and change the behavior of the pool. The following are read-only properties: +.sp +.ne 2 +.mk +.na +\fBavailable\fR +.ad +.RS 20n +.rt +Amount of storage available within the pool. This property can also be referred to by its shortened column name, "avail". +.RE + +.sp +.ne 2 +.mk +.na +\fBcapacity\fR +.ad +.RS 20n +.rt +Percentage of pool space used. This property can also be referred to by its shortened column name, "cap". +.RE + +.sp +.ne 2 +.mk +.na +\fBhealth\fR +.ad +.RS 20n +.rt +The current health of the pool. Health can be "\fBONLINE\fR", "\fBDEGRADED\fR", "\fBFAULTED\fR", " \fBOFFLINE\fR", "\fBREMOVED\fR", or "\fBUNAVAIL\fR". +.RE + +.sp +.ne 2 +.mk +.na +\fBguid\fR +.ad +.RS 20n +.rt +A unique identifier for the pool. +.RE + +.sp +.ne 2 +.mk +.na +\fBsize\fR +.ad +.RS 20n +.rt +Total size of the storage pool. +.RE + +.sp +.ne 2 +.mk +.na +\fBused\fR +.ad +.RS 20n +.rt +Amount of storage space used within the pool. +.RE + +.sp .LP -The "zpool create -R" and "zpool import -R" commands allow users to create and import a pool with a different root path. By default, whenever a pool is created or imported on a system, it is permanently added so that it is available whenever the system boots. For -removable media, or when in recovery situations, this may not always be desirable. An alternate root pool does not persist on the system. Instead, it exists only until exported or the system is rebooted, at which point it will have to be imported again. +These space usage properties report actual physical space available to the storage pool. The physical space can be different from the total amount of space that any contained datasets can actually use. The amount of space used in a \fBraidz\fR configuration depends on the characteristics +of the data being written. In addition, \fBZFS\fR reserves some space for internal accounting that the \fBzfs\fR(1M) command takes into account, but the \fBzpool\fR command does not. For non-full pools of a reasonable size, these effects should be invisible. For small pools, or pools that are close to being completely full, these discrepancies may become more noticeable. +.sp .LP -In addition, all mount points in the pool are prefixed with the given root, so a pool can be constrained to a particular area of the file system. This is most useful when importing unknown pools from removable media, as the mount points of any file systems cannot be trusted. +The following property can be set at creation time and import time: +.sp +.ne 2 +.mk +.na +\fB\fBaltroot\fR\fR +.ad +.sp .6 +.RS 4n +Alternate root directory. If set, this directory is prepended to any mount points within the pool. This can be used when examining an unknown pool where the mount points cannot be trusted, or in an alternate boot environment, where the typical paths are not valid. \fBaltroot\fR is +not a persistent property. It is valid only while the system is up. Setting \fBaltroot\fR defaults to using \fBcachefile\fR=none, though this may be overridden using an explicit setting. +.RE + +.sp .LP -When creating an alternate root pool, the default mount point is "/", rather than the normal default "/\fIpool\fR". +The following properties can be set at creation time and import time, and later changed with the "\fBzpool set\fR" command: +.sp +.ne 2 +.mk +.na +\fB\fBautoreplace\fR=on | off\fR +.ad +.sp .6 +.RS 4n +Controls automatic device replacement. If set to "\fBoff\fR", device replacement must be initiated by the administrator by using the "\fBzpool replace\fR" command. If set to "\fBon\fR", any new device, found +in the same physical location as a device that previously belonged to the pool, is automatically formatted and replaced. The default behavior is "\fBoff\fR". This property can also be referred to by its shortened column name, "replace". +.RE + +.sp +.ne 2 +.mk +.na +\fB\fBbootfs\fR=\fIpool\fR/\fIdataset\fR\fR +.ad +.sp .6 +.RS 4n +Identifies the default bootable dataset for the root pool. This property is expected to be set mainly by the installation and upgrade programs. +.RE + +.sp +.ne 2 +.mk +.na +\fB\fBcachefile\fR=\fIpath\fR | "none"\fR +.ad +.sp .6 +.RS 4n +Controls the location of where the pool configuration is cached. Discovering all pools on system startup requires a cached copy of the configuration data that is stored on the root file system. All pools in this cache are automatically imported when the system boots. Some environments, +such as install and clustering, need to cache this information in a different location so that pools are not automatically imported. Setting this property caches the pool configuration in a different location that can later be imported with "\fBzpool import -c\fR". Setting +it to the special value "\fBnone\fR" creates a temporary pool that is never cached, and the special value \fB\&''\fR (empty string) uses the default location. +.sp +Multiple pools can share the same cache file. Because the kernel destroys and recreates this file when pools are added and removed, care should be taken when attempting to access this file. When the last pool using a \fBcachefile\fR is exported or destroyed, the file is removed. +.RE + +.sp +.ne 2 +.mk +.na +\fB\fBdelegation\fR=\fBon\fR | \fBoff\fR\fR +.ad +.sp .6 +.RS 4n +Controls whether a non-privileged user is granted access based on the dataset permissions defined on the dataset. See \fBzfs\fR(1M) for more information +on \fBZFS\fR delegated administration. +.RE + +.sp +.ne 2 +.mk +.na +\fB\fBfailmode\fR=\fBwait\fR | \fBcontinue\fR | \fBpanic\fR\fR +.ad +.sp .6 +.RS 4n +Controls the system behavior in the event of catastrophic pool failure. This condition is typically a result of a loss of connectivity to the underlying storage device(s) or a failure of all devices within the pool. The behavior of such an event is determined as follows: +.sp +.ne 2 +.mk +.na +\fBwait\fR +.ad +.RS 12n +.rt +Blocks all \fBI/O\fR access until the device connectivity is recovered and the errors are cleared. This is the default behavior. +.RE + +.sp +.ne 2 +.mk +.na +\fBcontinue\fR +.ad +.RS 12n +.rt +Returns \fBEIO\fR to any new write \fBI/O\fR requests but allows reads to any of the remaining healthy devices. Any write requests that have yet to be committed to disk would be blocked. +.RE + +.sp +.ne 2 +.mk +.na +\fBpanic\fR +.ad +.RS 12n +.rt +Prints out a message to the console and generates a system crash dump. +.RE + +.RE + +.sp +.ne 2 +.mk +.na +\fB\fBversion\fR=\fIversion\fR\fR +.ad +.sp .6 +.RS 4n +The current on-disk version of the pool. This can be increased, but never decreased. The preferred method of updating pools is with the "\fBzpool upgrade\fR" command, though this property can be used when a specific version is needed for backwards compatibility. +This property can be any number between 1 and the current version reported by "\fBzpool upgrade -v\fR". The special value "\fBcurrent\fR" is an alias for the latest supported version. +.RE + .SS "Subcommands" +.sp .LP All subcommands that modify state are logged persistently to the pool in their original form. +.sp .LP The \fBzpool\fR command provides subcommands to create and destroy storage pools, add capacity to storage pools, and provide information about the storage pools. The following subcommands are supported: .sp @@ -285,12 +669,12 @@ Displays a help message. .ne 2 .mk .na -\fB\fBzpool create\fR [\fB-fn\fR] [\fB-R\fR \fIroot\fR] [\fB-m\fR \fImountpoint\fR] \fIpool\fR \fIvdev ...\fR\fR +\fB\fBzpool create\fR [\fB-fn\fR] [\fB-o\fR \fIproperty=value\fR] ... [\fB-m\fR \fImountpoint\fR] [\fB-R\fR \fIroot\fR] \fIpool\fR \fIvdev\fR ...\fR .ad .sp .6 .RS 4n Creates a new storage pool containing the virtual devices specified on the command line. The pool name must begin with a letter, and can only contain alphanumeric characters as well as underscore ("_"), dash ("-"), and period ("."). The pool -names "mirror", "raidz", and "spare" are reserved, as are names beginning with the pattern "c[0-9]". The \fBvdev\fR specification is described in the "Virtual Devices" section. +names "mirror", "raidz", "spare" and "log" are reserved, as are names beginning with the pattern "c[0-9]". The \fBvdev\fR specification is described in the "Virtual Devices" section. .sp The command verifies that each device specified is accessible and not currently in use by another subsystem. There are some uses, such as being currently mounted, or specified as the dedicated dump device, that prevents a device from ever being used by \fBZFS\fR. Other uses, such as having a preexisting \fBUFS\fR file system, can be overridden with the \fB-f\fR option. @@ -305,8 +689,8 @@ Unless the \fB-R\fR option is specified, the default mount point is "/\fIpool\fR .na \fB\fB-f\fR\fR .ad -.RS 17n -.rt +.sp .6 +.RS 4n Forces use of \fBvdev\fRs, even if they appear in use or specify a conflicting replication level. Not all devices can be overridden in this manner. .RE @@ -316,8 +700,8 @@ Forces use of \fBvdev\fRs, even if they appear in use or specify a conflicting r .na \fB\fB-n\fR\fR .ad -.RS 17n -.rt +.sp .6 +.RS 4n Displays the configuration that would be used without actually creating the pool. The actual pool creation can still fail due to insufficient privileges or device sharing. .RE @@ -325,11 +709,22 @@ Displays the configuration that would be used without actually creating the pool .ne 2 .mk .na +\fB\fB-o\fR \fIproperty=value\fR [\fB-o\fR \fIproperty=value\fR] ...\fR +.ad +.sp .6 +.RS 4n +Sets the given pool properties. See the "Properties" section for a list of valid properties that can be set. +.RE + +.sp +.ne 2 +.mk +.na \fB\fB-R\fR \fIroot\fR\fR .ad -.RS 17n -.rt -Creates the pool with an alternate \fIroot\fR. See the "Alternate Root Pools" section. The root dataset has its mount point set to "/" as part of this operation. +.sp .6 +.RS 4n +Equivalent to "-o cachefile=none,altroot=\fIroot\fR" .RE .sp @@ -338,10 +733,9 @@ Creates the pool with an alternate \fIroot\fR. See the "Alternate Root Pools" se .na \fB\fB-m\fR \fImountpoint\fR\fR .ad -.RS 17n -.rt -Sets the mount point for the root dataset. The default mount point is "/\fIpool\fR". The mount point must be an absolute path, "\fBlegacy\fR", or "\fBnone\fR". For more information on dataset mount -points, see \fBzfs\fR(1M). +.sp .6 +.RS 4n +Sets the mount point for the root dataset. The default mount point is "/\fIpool\fR" or "\fBaltroot\fR/\fIpool\fR" if \fBaltroot\fR is specified. The mount point must be an absolute path, "\fBlegacy\fR", or "\fBnone\fR". For more information on dataset mount points, see \fBzfs\fR(1M). .RE .RE @@ -372,7 +766,7 @@ Forces any active datasets contained within the pool to be unmounted. .ne 2 .mk .na -\fB\fBzpool add\fR [\fB-fn\fR] \fIpool\fR \fIvdev ...\fR\fR +\fB\fBzpool add\fR [\fB-fn\fR] \fIpool\fR \fIvdev\fR ...\fR .ad .sp .6 .RS 4n @@ -400,26 +794,26 @@ Forces use of \fBvdev\fRs, even if they appear in use or specify a conflicting r Displays the configuration that would be used without actually adding the \fBvdev\fRs. The actual pool creation can still fail due to insufficient privileges or device sharing. .RE -Do not add a disk that is currently configured as a quorum device to a zpool. Once a disk is in a zpool, that disk can then be configured as a quorum device. +Do not add a disk that is currently configured as a quorum device to a zpool. After a disk is in the pool, that disk can then be configured as a quorum device. .RE .sp .ne 2 .mk .na -\fB\fBzpool remove\fR \fIpool\fR \fIvdev\fR\fR +\fB\fBzpool remove\fR \fIpool\fR \fIdevice\fR ...\fR .ad .sp .6 .RS 4n -Removes the given \fBvdev\fR from the pool. This command currently only supports removing hot spares. Devices which are part of a mirror can be removed using the "zpool detach" command. \fBRaidz\fR and top-level \fBvdevs\fR cannot -be removed from a pool. +Removes the specified device from the pool. This command currently only supports removing hot spares and cache devices. Devices that are part of a mirrored configuration can be removed using the "\fBzpool detach\fR" command. Non-redundant and \fBraidz\fR devices +cannot be removed from a pool. .RE .sp .ne 2 .mk .na -\fB\fBzpool list\fR [\fB-H\fR] [\fB-o\fR \fIfield\fR[,\fIfield*\fR]] [\fIpool\fR] ...\fR +\fB\fBzpool list\fR [\fB-H\fR] [\fB-o\fR \fIprops\fR[,...]] [\fIpool\fR] ...\fR .ad .sp .6 .RS 4n @@ -439,36 +833,20 @@ Scripted mode. Do not display headers, and separate fields by a single tab inste .ne 2 .mk .na -\fB\fB-o\fR \fIfield\fR\fR +\fB\fB-o\fR \fIprops\fR\fR .ad .RS 12n .rt -Comma-separated list of fields to display. Each field must be one of: -.sp -.in +2 -.nf -name Pool name -size Total size -used Amount of space used -available Amount of space available -capacity Percentage of pool space used -health Health status -.fi -.in -2 -.sp - -The default is all fields. +Comma-separated list of properties to display. See the "Properties" section for a list of valid properties. The default list is "name, size, used, available, capacity, health, altroot" .RE -This command reports actual physical space available to the storage pool. The physical space can be different from the total amount of space that any contained datasets can actually use. The amount of space used in a \fBraidz\fR configuration depends on the characteristics of -the data being written. In addition, \fBZFS\fR reserves some space for internal accounting that the \fBzfs\fR(1M) command takes into account, but the \fBzpool\fR command does not. For non-full pools of a reasonable size, these effects should be invisible. For small pools, or pools that are close to being completely full, these discrepancies may become more noticeable. .RE .sp .ne 2 .mk .na -\fB\fBzpool iostat\fR [\fB-v\fR] [\fIpool\fR] ... [\fIinterval\fR [\fIcount\fR]]\fR +\fB\fBzpool iostat\fR [\fB-v\fR] [\fIpool\fR] ... [\fIinterval\fR[\fIcount\fR]]\fR .ad .sp .6 .RS 4n @@ -495,7 +873,7 @@ Verbose statistics. Reports usage statistics for individual \fIvdevs\fR within t .ad .sp .6 .RS 4n -Displays the detailed health status for the given pools. If no \fIpool\fR is specified, then the status of each pool in the system is displayed. +Displays the detailed health status for the given pools. If no \fIpool\fR is specified, then the status of each pool in the system is displayed. For more information on pool and device health, see the "Device Failure and Recovery" section. .sp If a scrub or resilver is in progress, this command reports the percentage done and the estimated time to completion. Both of these are only approximate, because the amount of data in the pool and the other workloads on the system can change. .sp @@ -526,13 +904,26 @@ Displays verbose data error information, printing out a complete list of all dat .ne 2 .mk .na +\fB\fBzpool online\fR \fIpool\fR \fIdevice\fR ...\fR +.ad +.sp .6 +.RS 4n +Brings the specified physical device online. +.sp +This command is not applicable to spares or cache devices. +.RE + +.sp +.ne 2 +.mk +.na \fB\fBzpool offline\fR [\fB-t\fR] \fIpool\fR \fIdevice\fR ...\fR .ad .sp .6 .RS 4n Takes the specified physical device offline. While the \fIdevice\fR is offline, no attempt is made to read or write to the device. .sp -This command is not applicable to spares. +This command is not applicable to spares or cache devices. .sp .ne 2 .mk @@ -550,19 +941,6 @@ Temporary. Upon reboot, the specified physical device reverts to its previous st .ne 2 .mk .na -\fB\fBzpool online\fR \fIpool\fR \fIdevice\fR ...\fR -.ad -.sp .6 -.RS 4n -Brings the specified physical device online. -.sp -This command is not applicable to spares. -.RE - -.sp -.ne 2 -.mk -.na \fB\fBzpool clear\fR \fIpool\fR [\fIdevice\fR] ...\fR .ad .sp .6 @@ -616,8 +994,8 @@ Replaces \fIold_device\fR with \fInew_device\fR. This is equivalent to attaching .sp The size of \fInew_device\fR must be greater than or equal to the minimum size of all the devices in a mirror or \fBraidz\fR configuration. .sp -If \fInew_device\fR is not specified, it defaults to \fIold_device\fR. This form of replacement is useful after an existing disk has failed and has been physically replaced. In this case, the new disk may have the same \fB/dev/dsk\fR path -as the old device, even though it is actually a different disk. \fBZFS\fR recognizes this. +\fInew_device\fR is required if the pool is not redundant. If \fInew_device\fR is not specified, it defaults to \fIold_device\fR. This form of replacement is useful after an existing disk has failed and has been physically replaced. +In this case, the new disk may have the same \fB/dev/dsk\fR path as the old device, even though it is actually a different disk. \fBZFS\fR recognizes this. .sp .ne 2 .mk @@ -663,24 +1041,45 @@ Stop scrubbing. .ne 2 .mk .na -\fB\fBzpool export\fR [\fB-f\fR] \fIpool\fR ...\fR +\fB\fBzpool import\fR [\fB-d\fR \fIdir\fR | \fB-c\fR \fIcachefile\fR] [\fB-D\fR]\fR .ad .sp .6 .RS 4n -Exports the given pools from the system. All devices are marked as exported, but are still considered in use by other subsystems. The devices can be moved between systems (even those of different endianness) and imported as long as a sufficient number of devices are present. +Lists pools available to import. If the \fB-d\fR option is not specified, this command searches for devices in "/dev/dsk". The \fB-d\fR option can be specified multiple times, and all directories are searched. If the device appears to be part of +an exported pool, this command displays a summary of the pool with the name of the pool, a numeric identifier, as well as the \fIvdev\fR layout and current health of the device for each device or file. Destroyed pools, pools that were previously destroyed with the "\fBzpool destroy\fR" command, are not listed unless the \fB-D\fR option is specified. .sp -Before exporting the pool, all datasets within the pool are unmounted. +The numeric identifier is unique, and can be used instead of the pool name when multiple exported pools of the same name are available. .sp -For pools to be portable, you must give the \fBzpool\fR command whole disks, not just slices, so that \fBZFS\fR can label the disks with portable \fBEFI\fR labels. Otherwise, disk drivers on platforms of different endianness will not recognize the disks. +.ne 2 +.mk +.na +\fB\fB-c\fR \fIcachefile\fR\fR +.ad +.RS 16n +.rt +Reads configuration from the given \fBcachefile\fR that was created with the "\fBcachefile\fR" pool property. This \fBcachefile\fR is used instead of searching for devices. +.RE + .sp .ne 2 .mk .na -\fB\fB-f\fR\fR +\fB\fB-d\fR \fIdir\fR\fR .ad -.RS 6n +.RS 16n .rt -Forcefully unmount all datasets, using the "\fBunmount -f\fR" command. +Searches for devices or files in \fIdir\fR. The \fB-d\fR option can be specified multiple times. +.RE + +.sp +.ne 2 +.mk +.na +\fB\fB-D\fR\fR +.ad +.RS 16n +.rt +Lists destroyed pools only. .RE .RE @@ -689,59 +1088,56 @@ Forcefully unmount all datasets, using the "\fBunmount -f\fR" command. .ne 2 .mk .na -\fB\fBzpool import\fR [\fB-d\fR \fIdir\fR] [\fB-D\fR]\fR +\fB\fBzpool import\fR [\fB-o\fR \fImntopts\fR] [ \fB-o\fR \fIproperty\fR=\fIvalue\fR] ... [\fB-d\fR \fIdir\fR | \fB-c\fR \fIcachefile\fR] +[\fB-D\fR] [\fB-f\fR] [\fB-R\fR \fIroot\fR] \fB-a\fR\fR .ad .sp .6 .RS 4n -Lists pools available to import. If the \fB-d\fR option is not specified, this command searches for devices in "/dev/dsk". The \fB-d\fR option can be specified multiple times, and all directories are searched. If the device appears to be part of -an exported pool, this command displays a summary of the pool with the name of the pool, a numeric identifier, as well as the \fIvdev\fR layout and current health of the device for each device or file. Destroyed pools, pools that were previously destroyed with the "\fB-zpool destroy\fR" command, are not listed unless the \fB-D\fR option is specified. -.sp -The numeric identifier is unique, and can be used instead of the pool name when multiple exported pools of the same name are available. +Imports all pools found in the search directories. Identical to the previous command, except that all pools with a sufficient number of devices available are imported. Destroyed pools, pools that were previously destroyed with the "\fBzpool destroy\fR" +command, will not be imported unless the \fB-D\fR option is specified. .sp .ne 2 .mk .na -\fB\fB-d\fR \fIdir\fR\fR +\fB\fB-o\fR \fImntopts\fR\fR .ad -.RS 10n +.RS 21n .rt -Searches for devices or files in \fIdir\fR. The \fB-d\fR option can be specified multiple times. +Comma-separated list of mount options to use when mounting datasets within the pool. See \fBzfs\fR(1M) for a description of dataset properties and mount +options. .RE .sp .ne 2 .mk .na -\fB\fB-D\fR\fR +\fB\fB-o\fR \fIproperty=value\fR\fR .ad -.RS 10n +.RS 21n .rt -Lists destroyed pools only. -.RE - +Sets the specified property on the imported pool. See the "Properties" section for more information on the available pool properties. .RE .sp .ne 2 .mk .na -\fB\fBzpool import\fR [\fB-d\fR \fIdir\fR] [\fB-D\fR] [\fB-f\fR] [\fB-o\fR \fIopts\fR] [\fB-R\fR \fIroot\fR] \fIpool\fR | \fIid\fR [\fInewpool\fR]\fR +\fB\fB-c\fR \fIcachefile\fR\fR .ad -.sp .6 -.RS 4n -Imports a specific pool. A pool can be identified by its name or the numeric identifier. If \fInewpool\fR is specified, the pool is imported using the name \fInewpool\fR. Otherwise, it is imported with the same name as its exported name. -.sp -If a device is removed from a system without running "\fBzpool export\fR" first, the device appears as potentially active. It cannot be determined if this was a failed export, or whether the device is really in use from another host. To import a pool in this state, -the \fB-f\fR option is required. +.RS 21n +.rt +Reads configuration from the given \fBcachefile\fR that was created with the "\fBcachefile\fR" pool property. This \fBcachefile\fR is used instead of searching for devices. +.RE + .sp .ne 2 .mk .na \fB\fB-d\fR \fIdir\fR\fR .ad -.RS 11n +.RS 21n .rt -Searches for devices or files in \fIdir\fR. The \fB-d\fR option can be specified multiple times. +Searches for devices or files in \fIdir\fR. The \fB-d\fR option can be specified multiple times. This option is incompatible with the \fB-c\fR option. .RE .sp @@ -750,9 +1146,9 @@ Searches for devices or files in \fIdir\fR. The \fB-d\fR option can be specified .na \fB\fB-D\fR\fR .ad -.RS 11n +.RS 21n .rt -Imports destroyed pool. The \fB-f\fR option is also required. +Imports destroyed pools only. The \fB-f\fR option is also required. .RE .sp @@ -761,7 +1157,7 @@ Imports destroyed pool. The \fB-f\fR option is also required. .na \fB\fB-f\fR\fR .ad -.RS 11n +.RS 21n .rt Forces import, even if the pool appears to be potentially active. .RE @@ -770,12 +1166,11 @@ Forces import, even if the pool appears to be potentially active. .ne 2 .mk .na -\fB\fB-o\fR \fIopts\fR\fR +\fB\fB-a\fR\fR .ad -.RS 11n +.RS 21n .rt -Comma-separated list of mount options to use when mounting datasets within the pool. See \fBzfs\fR(1M) for a description of dataset properties and mount -options. +Searches for and imports all pools found. .RE .sp @@ -784,9 +1179,9 @@ options. .na \fB\fB-R\fR \fIroot\fR\fR .ad -.RS 11n +.RS 21n .rt -Imports pool(s) with an alternate \fIroot\fR. See the "Alternate Root Pools" section. +Sets the "\fBcachefile\fR" property to "\fBnone\fR" and the "\fIaltroot\fR" property to "\fIroot\fR". .RE .RE @@ -795,21 +1190,58 @@ Imports pool(s) with an alternate \fIroot\fR. See the "Alternate Root Pools" sec .ne 2 .mk .na -\fB\fBzpool import\fR [\fB-d\fR \fIdir\fR] [\fB-D\fR] [\fB-f\fR] [\fB-a\fR]\fR +\fB\fBzpool import\fR [\fB-o\fR \fImntopts\fR] [ \fB-o\fR \fIproperty\fR=\fIvalue\fR] ... [\fB-d\fR \fIdir\fR | \fB-c\fR \fIcachefile\fR] +[\fB-D\fR] [\fB-f\fR] [\fB-R\fR \fIroot\fR] \fIpool\fR | \fIid\fR [\fInewpool\fR]\fR .ad .sp .6 .RS 4n -Imports all pools found in the search directories. Identical to the previous command, except that all pools with a sufficient number of devices available are imported. Destroyed pools, pools that were previously destroyed with the "\fB-zpool destroy\fR" command, -will not be imported unless the \fB-D\fR option is specified. +Imports a specific pool. A pool can be identified by its name or the numeric identifier. If \fInewpool\fR is specified, the pool is imported using the name \fInewpool\fR. Otherwise, it is imported with the same name as its exported name. +.sp +If a device is removed from a system without running "\fBzpool export\fR" first, the device appears as potentially active. It cannot be determined if this was a failed export, or whether the device is really in use from another host. To import a pool in this state, +the \fB-f\fR option is required. +.sp +.ne 2 +.mk +.na +\fB\fB-o\fR \fImntopts\fR\fR +.ad +.sp .6 +.RS 4n +Comma-separated list of mount options to use when mounting datasets within the pool. See \fBzfs\fR(1M) for a description of dataset properties and mount +options. +.RE + +.sp +.ne 2 +.mk +.na +\fB\fB-o\fR \fIproperty=value\fR\fR +.ad +.sp .6 +.RS 4n +Sets the specified property on the imported pool. See the "Properties" section for more information on the available pool properties. +.RE + +.sp +.ne 2 +.mk +.na +\fB\fB-c\fR \fIcachefile\fR\fR +.ad +.sp .6 +.RS 4n +Reads configuration from the given \fBcachefile\fR that was created with the "\fBcachefile\fR" pool property. This \fBcachefile\fR is used instead of searching for devices. +.RE + .sp .ne 2 .mk .na \fB\fB-d\fR \fIdir\fR\fR .ad -.RS 10n -.rt -Searches for devices or files in \fIdir\fR. The \fB-d\fR option can be specified multiple times. +.sp .6 +.RS 4n +Searches for devices or files in \fIdir\fR. The \fB-d\fR option can be specified multiple times. This option is incompatible with the \fB-c\fR option. .RE .sp @@ -818,9 +1250,9 @@ Searches for devices or files in \fIdir\fR. The \fB-d\fR option can be specified .na \fB\fB-D\fR\fR .ad -.RS 10n -.rt -Imports destroyed pools only. The \fB-f\fR option is also required. +.sp .6 +.RS 4n +Imports destroyed pool. The \fB-f\fR option is also required. .RE .sp @@ -829,11 +1261,48 @@ Imports destroyed pools only. The \fB-f\fR option is also required. .na \fB\fB-f\fR\fR .ad -.RS 10n -.rt +.sp .6 +.RS 4n Forces import, even if the pool appears to be potentially active. .RE +.sp +.ne 2 +.mk +.na +\fB\fB-R\fR \fIroot\fR\fR +.ad +.sp .6 +.RS 4n +Sets the "\fBcachefile\fR" property to "\fBnone\fR" and the "\fIaltroot\fR" property to "\fIroot\fR". +.RE + +.RE + +.sp +.ne 2 +.mk +.na +\fB\fBzpool export\fR [\fB-f\fR] \fIpool\fR ...\fR +.ad +.sp .6 +.RS 4n +Exports the given pools from the system. All devices are marked as exported, but are still considered in use by other subsystems. The devices can be moved between systems (even those of different endianness) and imported as long as a sufficient number of devices are present. +.sp +Before exporting the pool, all datasets within the pool are unmounted. +.sp +For pools to be portable, you must give the \fBzpool\fR command whole disks, not just slices, so that \fBZFS\fR can label the disks with portable \fBEFI\fR labels. Otherwise, disk drivers on platforms of different endianness will not recognize the disks. +.sp +.ne 2 +.mk +.na +\fB\fB-f\fR\fR +.ad +.RS 6n +.rt +Forcefully unmount all datasets, using the "\fBunmount -f\fR" command. +.RE + .RE .sp @@ -856,14 +1325,14 @@ a more recent version are also displayed, although these pools will be inaccessi .ad .sp .6 .RS 4n -Displays \fBZFS\fR versions supported by the current software. The current \fBZFS\fR versions and all previous supportedversions are displayed, along with an explanation of the features provided with each version. +Displays \fBZFS\fR versions supported by the current software. The current \fBZFS\fR versions and all previous supported versions are displayed, along with an explanation of the features provided with each version. .RE .sp .ne 2 .mk .na -\fB\fBzpool upgrade\fR [\fB-a\fR | \fIpool\fR]\fR +\fB\fBzpool upgrade\fR [\fB-V\fR \fIversion\fR] \fB-a\fR | \fIpool\fR ...\fR .ad .sp .6 .RS 4n @@ -874,27 +1343,95 @@ Upgrades the given pool to the latest on-disk version. Once this is done, the po .na \fB\fB-a\fR\fR .ad -.RS 6n +.RS 14n .rt Upgrades all pools. .RE +.sp +.ne 2 +.mk +.na +\fB\fB-V\fR \fIversion\fR\fR +.ad +.RS 14n +.rt +Upgrade to the specified version. If the \fB-V\fR flag is not specified, the pool is upgraded to the most recent version. This option can only be used to increase the version number, and only up to the most recent version supported by this software. +.RE + +.RE + +.sp +.ne 2 +.mk +.na +\fB\fBzpool history\fR [\fB-il\fR] [\fIpool\fR] ...\fR +.ad +.sp .6 +.RS 4n +Displays the command history of the specified pools or all pools if no pool is specified. +.sp +.ne 2 +.mk +.na +\fB\fB-i\fR\fR +.ad +.RS 6n +.rt +Displays internally logged \fBZFS\fR events in addition to user initiated events. +.RE + +.sp +.ne 2 +.mk +.na +\fB\fB-l\fR\fR +.ad +.RS 6n +.rt +Displays log records in long format, which in addition to standard format includes, the user name, the hostname, and the zone in which the operation was performed. +.RE + +.RE + +.sp +.ne 2 +.mk +.na +\fB\fBzpool get\fR "\fIall\fR" | \fIproperty\fR[,...] \fIpool\fR ...\fR +.ad +.sp .6 +.RS 4n +Retrieves the given list of properties (or all properties if "\fBall\fR" is used) for the specified storage pool(s). These properties are displayed with the following fields: +.sp +.in +2 +.nf + name Name of storage pool + property Property name + value Property value + source Property source, either 'default' or 'local'. +.fi +.in -2 +.sp + +See the "Properties" section for more information on the available pool properties. .RE .sp .ne 2 .mk .na -\fB\fBzpool history\fR [\fIpool\fR] ...\fR +\fB\fBzpool set\fR \fIproperty\fR=\fIvalue\fR \fIpool\fR \fR .ad .sp .6 .RS 4n -Displays the command history of the specified pools (or all pools if no pool is specified). +Sets the given property on the specified pool. See the "Properties" section for more information on what properties can be set and acceptable values. .RE .SH EXAMPLES .LP \fBExample 1 \fRCreating a RAID-Z Storage Pool +.sp .LP The following command creates a pool with a single \fBraidz\fR root \fIvdev\fR that consists of six disks. @@ -908,6 +1445,7 @@ The following command creates a pool with a single \fBraidz\fR root \fIvdev\fR t .LP \fBExample 2 \fRCreating a Mirrored Storage Pool +.sp .LP The following command creates a pool with two mirrors, where each mirror contains two disks. @@ -921,6 +1459,7 @@ The following command creates a pool with two mirrors, where each mirror contain .LP \fBExample 3 \fRCreating a ZFS Storage Pool by Using Slices +.sp .LP The following command creates an unmirrored pool using two disk slices. @@ -934,6 +1473,7 @@ The following command creates an unmirrored pool using two disk slices. .LP \fBExample 4 \fRCreating a ZFS Storage Pool by Using Files +.sp .LP The following command creates an unmirrored pool using files. While not recommended, a pool based on files can be useful for experimental purposes. @@ -947,6 +1487,7 @@ The following command creates an unmirrored pool using files. While not recommen .LP \fBExample 5 \fRAdding a Mirror to a ZFS Storage Pool +.sp .LP The following command adds two mirrored disks to the pool "\fItank\fR", assuming the pool is already made up of two-way mirrors. The additional space is immediately available to any datasets within the pool. @@ -960,9 +1501,11 @@ The following command adds two mirrored disks to the pool "\fItank\fR", assuming .LP \fBExample 6 \fRListing Available ZFS Storage Pools +.sp .LP The following command lists all available pools on the system. In this case, the pool \fIzion\fR is faulted due to a missing device. +.sp .LP The results from this command are similar to the following: @@ -980,6 +1523,7 @@ The results from this command are similar to the following: .LP \fBExample 7 \fRDestroying a ZFS Storage Pool +.sp .LP The following command destroys the pool "\fItank\fR" and any datasets contained within. @@ -993,6 +1537,7 @@ The following command destroys the pool "\fItank\fR" and any datasets contained .LP \fBExample 8 \fRExporting a ZFS Storage Pool +.sp .LP The following command exports the devices in pool \fItank\fR so that they can be relocated or later imported. @@ -1006,9 +1551,11 @@ The following command exports the devices in pool \fItank\fR so that they can be .LP \fBExample 9 \fRImporting a ZFS Storage Pool +.sp .LP The following command displays available pools, and then imports the pool "tank" for use on the system. +.sp .LP The results from this command are similar to the following: @@ -1034,6 +1581,7 @@ config: .LP \fBExample 10 \fRUpgrading All ZFS Storage Pools to the Current Version +.sp .LP The following command upgrades all ZFS Storage pools to the current version of the software. @@ -1048,6 +1596,7 @@ This system is currently running ZFS version 2. .LP \fBExample 11 \fRManaging Hot Spares +.sp .LP The following command creates a new pool with an available hot spare: @@ -1059,6 +1608,7 @@ The following command creates a new pool with an available hot spare: .in -2 .sp +.sp .LP If one of the disks were to fail, the pool would be reduced to the degraded state. The failed device can be replaced using the following command: @@ -1070,8 +1620,9 @@ If one of the disks were to fail, the pool would be reduced to the degraded stat .in -2 .sp +.sp .LP -Once the data has been resilvered, the spare is automatically removed and is made available should another device fails. The hot spare can be permanently removed from the pool using the following command: +Once the data has been resilvered, the spare is automatically removed and is made available should another device fails. The hot spare can be permanently removed from the pool using the following command: .sp .in +2 @@ -1081,7 +1632,48 @@ Once the data has been resilvered, the spare is automatically removed and is mad .in -2 .sp +.LP +\fBExample 12 \fRCreating a ZFS Pool with Mirrored Separate Intent Logs +.sp +.LP +The following command creates a ZFS storage pool consisting of two, two-way mirrors and mirrored log devices: + +.sp +.in +2 +.nf +\fB# zpool create pool mirror c0d0 c1d0 mirror c2d0 c3d0 log mirror \e + c4d0 c5d0\fR +.fi +.in -2 +.sp + +.LP +\fBExample 13 \fRAdding Cache Devices to a ZFS Pool +.sp +.LP +The following command adds two disks for use as cache devices to a ZFS storage pool: + +.sp +.in +2 +.nf +\fB# zpool add pool cache c2d0 c3d0\fR +.fi +.in -2 +.sp + +.sp +.LP +Once added, the cache devices gradually fill with content from main memory. Depending on the size of your cache devices, it could take over an hour for them to fill. Capacity and reads can be monitored using the \fBiostat\fR option as follows: +.sp +.in +2 +.nf +\fB# zpool iostat -v pool 5\fR +.fi +.in -2 +.sp + .SH EXIT STATUS +.sp .LP The following exit values are returned: .sp @@ -1118,6 +1710,7 @@ Invalid command line options were specified. .RE .SH ATTRIBUTES +.sp .LP See \fBattributes\fR(5) for descriptions of the following attributes: .sp @@ -1136,5 +1729,6 @@ Interface StabilityEvolving .TE .SH SEE ALSO +.sp .LP \fBzfs\fR(1M), \fBattributes\fR(5) diff --git a/cddl/contrib/opensolaris/cmd/zpool/zpool_iter.c b/cddl/contrib/opensolaris/cmd/zpool/zpool_iter.c index f724179..6ba91b1 100644 --- a/cddl/contrib/opensolaris/cmd/zpool/zpool_iter.c +++ b/cddl/contrib/opensolaris/cmd/zpool/zpool_iter.c @@ -53,6 +53,7 @@ struct zpool_list { boolean_t zl_findall; uu_avl_t *zl_avl; uu_avl_pool_t *zl_pool; + zprop_list_t **zl_proplist; }; /* ARGSUSED */ @@ -81,6 +82,12 @@ add_pool(zpool_handle_t *zhp, void *data) node->zn_handle = zhp; uu_avl_node_init(node, &node->zn_avlnode, zlp->zl_pool); if (uu_avl_find(zlp->zl_avl, node, NULL, &idx) == NULL) { + if (zlp->zl_proplist && + zpool_expand_proplist(zhp, zlp->zl_proplist) != 0) { + zpool_close(zhp); + free(node); + return (-1); + } uu_avl_insert(zlp->zl_avl, node, idx); } else { zpool_close(zhp); @@ -98,7 +105,7 @@ add_pool(zpool_handle_t *zhp, void *data) * line. */ zpool_list_t * -pool_list_get(int argc, char **argv, zpool_proplist_t **proplist, int *err) +pool_list_get(int argc, char **argv, zprop_list_t **proplist, int *err) { zpool_list_t *zlp; @@ -114,6 +121,8 @@ pool_list_get(int argc, char **argv, zpool_proplist_t **proplist, int *err) UU_DEFAULT)) == NULL) zpool_no_memory(); + zlp->zl_proplist = proplist; + if (argc == 0) { (void) zpool_iter(g_zfs, add_pool, zlp); zlp->zl_findall = B_TRUE; @@ -123,13 +132,12 @@ pool_list_get(int argc, char **argv, zpool_proplist_t **proplist, int *err) for (i = 0; i < argc; i++) { zpool_handle_t *zhp; - if ((zhp = zpool_open_canfail(g_zfs, - argv[i])) != NULL && add_pool(zhp, zlp) == 0) { - if (proplist && - zpool_expand_proplist(zhp, proplist) != 0) + if (zhp = zpool_open_canfail(g_zfs, argv[i])) { + if (add_pool(zhp, zlp) != 0) *err = B_TRUE; - } else + } else { *err = B_TRUE; + } } } @@ -228,7 +236,7 @@ pool_list_count(zpool_list_t *zlp) */ int for_each_pool(int argc, char **argv, boolean_t unavail, - zpool_proplist_t **proplist, zpool_iter_f func, void *data) + zprop_list_t **proplist, zpool_iter_f func, void *data) { zpool_list_t *list; int ret = 0; diff --git a/cddl/contrib/opensolaris/cmd/zpool/zpool_main.c b/cddl/contrib/opensolaris/cmd/zpool/zpool_main.c index 5b1d856..2388df9 100644 --- a/cddl/contrib/opensolaris/cmd/zpool/zpool_main.c +++ b/cddl/contrib/opensolaris/cmd/zpool/zpool_main.c @@ -20,12 +20,10 @@ */ /* - * Copyright 2007 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ -#pragma ident "%Z%%M% %I% %E% SMI" - #include <solaris.h> #include <assert.h> #include <ctype.h> @@ -42,6 +40,8 @@ #include <strings.h> #include <unistd.h> #include <priv.h> +#include <pwd.h> +#include <zone.h> #include <sys/time.h> #include <sys/fs/zfs.h> @@ -50,6 +50,7 @@ #include <libzfs.h> #include "zpool_util.h" +#include "zfs_comutil.h" static int zpool_do_create(int, char **); static int zpool_do_destroy(int, char **); @@ -85,6 +86,8 @@ static int zpool_do_set(int, char **); * These libumem hooks provide a reasonable set of defaults for the allocator's * debugging facilities. */ + +#ifdef DEBUG const char * _umem_debug_init(void) { @@ -96,6 +99,7 @@ _umem_logging_init(void) { return ("fail,contents"); /* $UMEM_LOGGING setting */ } +#endif typedef enum { HELP_ADD, @@ -169,6 +173,7 @@ static zpool_command_t command_table[] = { #define NCOMMAND (sizeof (command_table) / sizeof (command_table[0])) zpool_command_t *current_command; +static char history_str[HIS_MAX_RECORD_LEN]; static const char * get_usage(zpool_help_t idx) { @@ -177,12 +182,13 @@ get_usage(zpool_help_t idx) { return (gettext("\tadd [-fn] <pool> <vdev> ...\n")); case HELP_ATTACH: return (gettext("\tattach [-f] <pool> <device> " - "<new_device>\n")); + "<new-device>\n")); case HELP_CLEAR: return (gettext("\tclear <pool> [device]\n")); case HELP_CREATE: - return (gettext("\tcreate [-fn] [-R root] [-m mountpoint] " - "<pool> <vdev> ...\n")); + return (gettext("\tcreate [-fn] [-o property=value] ... \n" + "\t [-O file-system-property=value] ... \n" + "\t [-m mountpoint] [-R root] <pool> <vdev> ...\n")); case HELP_DESTROY: return (gettext("\tdestroy [-f] <pool>\n")); case HELP_DETACH: @@ -190,17 +196,19 @@ get_usage(zpool_help_t idx) { case HELP_EXPORT: return (gettext("\texport [-f] <pool> ...\n")); case HELP_HISTORY: - return (gettext("\thistory [<pool>]\n")); + return (gettext("\thistory [-il] [<pool>] ...\n")); case HELP_IMPORT: return (gettext("\timport [-d dir] [-D]\n" - "\timport [-d dir] [-D] [-f] [-o opts] [-R root] -a\n" - "\timport [-d dir] [-D] [-f] [-o opts] [-R root ]" - " <pool | id> [newpool]\n")); + "\timport [-o mntopts] [-o property=value] ... \n" + "\t [-d dir | -c cachefile] [-D] [-f] [-R root] -a\n" + "\timport [-o mntopts] [-o property=value] ... \n" + "\t [-d dir | -c cachefile] [-D] [-f] [-R root] " + "<pool | id> [newpool]\n")); case HELP_IOSTAT: return (gettext("\tiostat [-v] [pool] ... [interval " "[count]]\n")); case HELP_LIST: - return (gettext("\tlist [-H] [-o field[,field]*] " + return (gettext("\tlist [-H] [-o property[,...]] " "[pool] ...\n")); case HELP_OFFLINE: return (gettext("\toffline [-t] <pool> <device> ...\n")); @@ -208,9 +216,9 @@ get_usage(zpool_help_t idx) { return (gettext("\tonline <pool> <device> ...\n")); case HELP_REPLACE: return (gettext("\treplace [-f] <pool> <device> " - "[new_device]\n")); + "[new-device]\n")); case HELP_REMOVE: - return (gettext("\tremove <pool> <device>\n")); + return (gettext("\tremove <pool> <device> ...\n")); case HELP_SCRUB: return (gettext("\tscrub [-s] <pool> ...\n")); case HELP_STATUS: @@ -218,9 +226,9 @@ get_usage(zpool_help_t idx) { case HELP_UPGRADE: return (gettext("\tupgrade\n" "\tupgrade -v\n" - "\tupgrade <-a | pool>\n")); + "\tupgrade [-V version] <-a | pool ...>\n")); case HELP_GET: - return (gettext("\tget <all | property[,property]...> " + return (gettext("\tget <\"all\" | property[,...]> " "<pool> ...\n")); case HELP_SET: return (gettext("\tset <property=value> <pool> \n")); @@ -230,67 +238,28 @@ get_usage(zpool_help_t idx) { /* NOTREACHED */ } -/* - * Fields available for 'zpool list'. - */ -typedef enum { - ZPOOL_FIELD_NAME, - ZPOOL_FIELD_SIZE, - ZPOOL_FIELD_USED, - ZPOOL_FIELD_AVAILABLE, - ZPOOL_FIELD_CAPACITY, - ZPOOL_FIELD_HEALTH, - ZPOOL_FIELD_ROOT -} zpool_field_t; - -#define MAX_FIELDS 10 - -typedef struct column_def { - const char *cd_title; - size_t cd_width; - enum { - left_justify, - right_justify - } cd_justify; -} column_def_t; - -static column_def_t column_table[] = { - { "NAME", 20, left_justify }, - { "SIZE", 6, right_justify }, - { "USED", 6, right_justify }, - { "AVAIL", 6, right_justify }, - { "CAP", 5, right_justify }, - { "HEALTH", 9, left_justify }, - { "ALTROOT", 15, left_justify } -}; - -static char *column_subopts[] = { - "name", - "size", - "used", - "available", - "capacity", - "health", - "root", - NULL -}; /* * Callback routine that will print out a pool property value. */ -static zpool_prop_t -print_prop_cb(zpool_prop_t prop, void *cb) +static int +print_prop_cb(int prop, void *cb) { FILE *fp = cb; (void) fprintf(fp, "\t%-13s ", zpool_prop_to_name(prop)); + if (zpool_prop_readonly(prop)) + (void) fprintf(fp, " NO "); + else + (void) fprintf(fp, " YES "); + if (zpool_prop_values(prop) == NULL) (void) fprintf(fp, "-\n"); else (void) fprintf(fp, "%s\n", zpool_prop_values(prop)); - return (ZFS_PROP_CONT); + return (ZPROP_CONT); } /* @@ -301,7 +270,6 @@ print_prop_cb(zpool_prop_t prop, void *cb) void usage(boolean_t requested) { - int i; FILE *fp = requested ? stdout : stderr; if (current_command == NULL) { @@ -321,28 +289,22 @@ usage(boolean_t requested) } else { (void) fprintf(fp, gettext("usage:\n")); (void) fprintf(fp, "%s", get_usage(current_command->usage)); - - if (strcmp(current_command->name, "list") == 0) { - (void) fprintf(fp, gettext("\nwhere 'field' is one " - "of the following:\n\n")); - - for (i = 0; column_subopts[i] != NULL; i++) - (void) fprintf(fp, "\t%s\n", column_subopts[i]); - } } if (current_command != NULL && ((strcmp(current_command->name, "set") == 0) || - (strcmp(current_command->name, "get") == 0))) { + (strcmp(current_command->name, "get") == 0) || + (strcmp(current_command->name, "list") == 0))) { (void) fprintf(fp, gettext("\nthe following properties are supported:\n")); - (void) fprintf(fp, "\n\t%-13s %s\n\n", - "PROPERTY", "VALUES"); + (void) fprintf(fp, "\n\t%-13s %s %s\n\n", + "PROPERTY", "EDIT", "VALUES"); /* Iterate over all properties */ - (void) zpool_prop_iter(print_prop_cb, fp, B_FALSE); + (void) zprop_iter(print_prop_cb, fp, B_FALSE, B_TRUE, + ZFS_TYPE_POOL); } /* @@ -356,46 +318,9 @@ usage(boolean_t requested) exit(requested ? 0 : 2); } -const char * -state_to_health(int vs_state) -{ - switch (vs_state) { - case VDEV_STATE_CLOSED: - case VDEV_STATE_CANT_OPEN: - case VDEV_STATE_OFFLINE: - return (dgettext(TEXT_DOMAIN, "FAULTED")); - case VDEV_STATE_DEGRADED: - return (dgettext(TEXT_DOMAIN, "DEGRADED")); - case VDEV_STATE_HEALTHY: - return (dgettext(TEXT_DOMAIN, "ONLINE")); - } - - return (dgettext(TEXT_DOMAIN, "UNKNOWN")); -} - -const char * -state_to_name(vdev_stat_t *vs) -{ - switch (vs->vs_state) { - case VDEV_STATE_CLOSED: - case VDEV_STATE_CANT_OPEN: - if (vs->vs_aux == VDEV_AUX_CORRUPT_DATA) - return (gettext("FAULTED")); - else - return (gettext("UNAVAIL")); - case VDEV_STATE_OFFLINE: - return (gettext("OFFLINE")); - case VDEV_STATE_DEGRADED: - return (gettext("DEGRADED")); - case VDEV_STATE_HEALTHY: - return (gettext("ONLINE")); - } - - return (gettext("UNKNOWN")); -} - void -print_vdev_tree(zpool_handle_t *zhp, const char *name, nvlist_t *nv, int indent) +print_vdev_tree(zpool_handle_t *zhp, const char *name, nvlist_t *nv, int indent, + boolean_t print_logs) { nvlist_t **child; uint_t c, children; @@ -409,13 +334,75 @@ print_vdev_tree(zpool_handle_t *zhp, const char *name, nvlist_t *nv, int indent) return; for (c = 0; c < children; c++) { + uint64_t is_log = B_FALSE; + + (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, + &is_log); + if ((is_log && !print_logs) || (!is_log && print_logs)) + continue; + vname = zpool_vdev_name(g_zfs, zhp, child[c]); - print_vdev_tree(zhp, vname, child[c], indent + 2); + print_vdev_tree(zhp, vname, child[c], indent + 2, + B_FALSE); free(vname); } } /* + * Add a property pair (name, string-value) into a property nvlist. + */ +static int +add_prop_list(const char *propname, char *propval, nvlist_t **props, + boolean_t poolprop) +{ + zpool_prop_t prop = ZPROP_INVAL; + zfs_prop_t fprop; + nvlist_t *proplist; + const char *normnm; + char *strval; + + if (*props == NULL && + nvlist_alloc(props, NV_UNIQUE_NAME, 0) != 0) { + (void) fprintf(stderr, + gettext("internal error: out of memory\n")); + return (1); + } + + proplist = *props; + + if (poolprop) { + if ((prop = zpool_name_to_prop(propname)) == ZPROP_INVAL) { + (void) fprintf(stderr, gettext("property '%s' is " + "not a valid pool property\n"), propname); + return (2); + } + normnm = zpool_prop_to_name(prop); + } else { + if ((fprop = zfs_name_to_prop(propname)) == ZPROP_INVAL) { + (void) fprintf(stderr, gettext("property '%s' is " + "not a valid file system property\n"), propname); + return (2); + } + normnm = zfs_prop_to_name(fprop); + } + + if (nvlist_lookup_string(proplist, normnm, &strval) == 0 && + prop != ZPOOL_PROP_CACHEFILE) { + (void) fprintf(stderr, gettext("property '%s' " + "specified multiple times\n"), propname); + return (2); + } + + if (nvlist_add_string(proplist, normnm, propval) != 0) { + (void) fprintf(stderr, gettext("internal " + "error: out of memory\n")); + return (1); + } + + return (0); +} + +/* * zpool add [-fn] <pool> <vdev> ... * * -f Force addition of devices, even if they appear in use @@ -483,7 +470,8 @@ zpool_do_add(int argc, char **argv) } /* pass off to get_vdev_spec for processing */ - nvroot = make_root_vdev(config, force, !force, B_FALSE, argc, argv); + nvroot = make_root_vdev(zhp, force, !force, B_FALSE, dryrun, + argc, argv); if (nvroot == NULL) { zpool_close(zhp); return (1); @@ -498,16 +486,21 @@ zpool_do_add(int argc, char **argv) (void) printf(gettext("would update '%s' to the following " "configuration:\n"), zpool_get_name(zhp)); - print_vdev_tree(zhp, poolname, poolnvroot, 0); - print_vdev_tree(zhp, NULL, nvroot, 0); + /* print original main pool and new tree */ + print_vdev_tree(zhp, poolname, poolnvroot, 0, B_FALSE); + print_vdev_tree(zhp, NULL, nvroot, 0, B_FALSE); + + /* Do the same for the logs */ + if (num_logs(poolnvroot) > 0) { + print_vdev_tree(zhp, "logs", poolnvroot, 0, B_TRUE); + print_vdev_tree(zhp, NULL, nvroot, 0, B_TRUE); + } else if (num_logs(nvroot) > 0) { + print_vdev_tree(zhp, "logs", nvroot, 0, B_TRUE); + } ret = 0; } else { ret = (zpool_add(zhp, nvroot) != 0); - if (!ret) { - zpool_log_history(g_zfs, argc + 1 + optind, - argv - 1 - optind, poolname, B_TRUE, B_FALSE); - } } nvlist_free(nvroot); @@ -517,17 +510,17 @@ zpool_do_add(int argc, char **argv) } /* - * zpool remove <pool> <vdev> + * zpool remove <pool> <vdev> ... * * Removes the given vdev from the pool. Currently, this only supports removing - * spares from the pool. Eventually, we'll want to support removing leaf vdevs - * (as an alias for 'detach') as well as toplevel vdevs. + * spares and cache devices from the pool. Eventually, we'll want to support + * removing leaf vdevs (as an alias for 'detach') as well as toplevel vdevs. */ int zpool_do_remove(int argc, char **argv) { char *poolname; - int ret; + int i, ret = 0; zpool_handle_t *zhp; argc--; @@ -548,17 +541,18 @@ zpool_do_remove(int argc, char **argv) if ((zhp = zpool_open(g_zfs, poolname)) == NULL) return (1); - ret = (zpool_vdev_remove(zhp, argv[1]) != 0); - if (!ret) { - zpool_log_history(g_zfs, ++argc, --argv, poolname, B_TRUE, - B_FALSE); + for (i = 1; i < argc; i++) { + if (zpool_vdev_remove(zhp, argv[i]) != 0) + ret = 1; } return (ret); } /* - * zpool create [-fn] [-R root] [-m mountpoint] <pool> <dev> ... + * zpool create [-fn] [-o property=value] ... + * [-O file-system-property=value] ... + * [-R root] [-m mountpoint] <pool> <dev> ... * * -f Force creation, even if devices appear in use * -n Do not create the pool, but display the resulting layout if it @@ -566,6 +560,8 @@ zpool_do_remove(int argc, char **argv) * -R Create a pool under an alternate root * -m Set default mountpoint for the root dataset. By default it's * '/<pool>' + * -o Set property=value. + * -O Set fsproperty=value in the pool's root file system * * Creates the named pool according to the given vdev specification. The * bulk of the vdev processing is done in get_vdev_spec() in zpool_vdev.c. Once @@ -578,16 +574,17 @@ zpool_do_create(int argc, char **argv) boolean_t force = B_FALSE; boolean_t dryrun = B_FALSE; int c; - nvlist_t *nvroot; + nvlist_t *nvroot = NULL; char *poolname; - int ret; + int ret = 1; char *altroot = NULL; char *mountpoint = NULL; - nvlist_t **child; - uint_t children; + nvlist_t *fsprops = NULL; + nvlist_t *props = NULL; + char *propval; /* check options */ - while ((c = getopt(argc, argv, ":fnR:m:")) != -1) { + while ((c = getopt(argc, argv, ":fnR:m:o:O:")) != -1) { switch (c) { case 'f': force = B_TRUE; @@ -597,19 +594,52 @@ zpool_do_create(int argc, char **argv) break; case 'R': altroot = optarg; + if (add_prop_list(zpool_prop_to_name( + ZPOOL_PROP_ALTROOT), optarg, &props, B_TRUE)) + goto errout; + if (nvlist_lookup_string(props, + zpool_prop_to_name(ZPOOL_PROP_CACHEFILE), + &propval) == 0) + break; + if (add_prop_list(zpool_prop_to_name( + ZPOOL_PROP_CACHEFILE), "none", &props, B_TRUE)) + goto errout; break; case 'm': mountpoint = optarg; break; + case 'o': + if ((propval = strchr(optarg, '=')) == NULL) { + (void) fprintf(stderr, gettext("missing " + "'=' for -o option\n")); + goto errout; + } + *propval = '\0'; + propval++; + + if (add_prop_list(optarg, propval, &props, B_TRUE)) + goto errout; + break; + case 'O': + if ((propval = strchr(optarg, '=')) == NULL) { + (void) fprintf(stderr, gettext("missing " + "'=' for -O option\n")); + goto errout; + } + *propval = '\0'; + propval++; + + if (add_prop_list(optarg, propval, &fsprops, B_FALSE)) + goto errout; + break; case ':': (void) fprintf(stderr, gettext("missing argument for " "'%c' option\n"), optopt); - usage(B_FALSE); - break; + goto badusage; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); - usage(B_FALSE); + goto badusage; } } @@ -619,11 +649,11 @@ zpool_do_create(int argc, char **argv) /* get pool name and check number of arguments */ if (argc < 1) { (void) fprintf(stderr, gettext("missing pool name argument\n")); - usage(B_FALSE); + goto badusage; } if (argc < 2) { (void) fprintf(stderr, gettext("missing vdev specification\n")); - usage(B_FALSE); + goto badusage; } poolname = argv[0]; @@ -637,31 +667,28 @@ zpool_do_create(int argc, char **argv) "character '/' in pool name\n"), poolname); (void) fprintf(stderr, gettext("use 'zfs create' to " "create a dataset\n")); - return (1); + goto errout; } /* pass off to get_vdev_spec for bulk processing */ - nvroot = make_root_vdev(NULL, force, !force, B_FALSE, argc - 1, - argv + 1); + nvroot = make_root_vdev(NULL, force, !force, B_FALSE, dryrun, + argc - 1, argv + 1); if (nvroot == NULL) - return (1); + goto errout; /* make_root_vdev() allows 0 toplevel children if there are spares */ - verify(nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN, - &child, &children) == 0); - if (children == 0) { + if (!zfs_allocatable_devs(nvroot)) { (void) fprintf(stderr, gettext("invalid vdev " "specification: at least one toplevel vdev must be " "specified\n")); - return (1); + goto errout; } if (altroot != NULL && altroot[0] != '/') { (void) fprintf(stderr, gettext("invalid alternate root '%s': " "must be an absolute path\n"), altroot); - nvlist_free(nvroot); - return (1); + goto errout; } /* @@ -672,14 +699,13 @@ zpool_do_create(int argc, char **argv) (strcmp(mountpoint, ZFS_MOUNTPOINT_LEGACY) != 0 && strcmp(mountpoint, ZFS_MOUNTPOINT_NONE) != 0)) { char buf[MAXPATHLEN]; - struct stat64 statbuf; + DIR *dirp; if (mountpoint && mountpoint[0] != '/') { (void) fprintf(stderr, gettext("invalid mountpoint " "'%s': must be an absolute path, 'legacy', or " "'none'\n"), mountpoint); - nvlist_free(nvroot); - return (1); + goto errout; } if (mountpoint == NULL) { @@ -698,23 +724,30 @@ zpool_do_create(int argc, char **argv) mountpoint); } - if (stat64(buf, &statbuf) == 0 && - statbuf.st_nlink != 2) { - if (mountpoint == NULL) - (void) fprintf(stderr, gettext("default " - "mountpoint '%s' exists and is not " - "empty\n"), buf); - else - (void) fprintf(stderr, gettext("mountpoint " - "'%s' exists and is not empty\n"), buf); + if ((dirp = opendir(buf)) == NULL && errno != ENOENT) { + (void) fprintf(stderr, gettext("mountpoint '%s' : " + "%s\n"), buf, strerror(errno)); (void) fprintf(stderr, gettext("use '-m' " "option to provide a different default\n")); - nvlist_free(nvroot); - return (1); + goto errout; + } else if (dirp) { + int count = 0; + + while (count < 3 && readdir(dirp) != NULL) + count++; + (void) closedir(dirp); + + if (count > 2) { + (void) fprintf(stderr, gettext("mountpoint " + "'%s' exists and is not empty\n"), buf); + (void) fprintf(stderr, gettext("use '-m' " + "option to provide a " + "different default\n")); + goto errout; + } } } - if (dryrun) { /* * For a dry run invocation, print out a basic message and run @@ -724,15 +757,17 @@ zpool_do_create(int argc, char **argv) (void) printf(gettext("would create '%s' with the " "following layout:\n\n"), poolname); - print_vdev_tree(NULL, poolname, nvroot, 0); + print_vdev_tree(NULL, poolname, nvroot, 0, B_FALSE); + if (num_logs(nvroot) > 0) + print_vdev_tree(NULL, "logs", nvroot, 0, B_TRUE); ret = 0; } else { - ret = 1; /* * Hand off to libzfs. */ - if (zpool_create(g_zfs, poolname, nvroot, altroot) == 0) { + if (zpool_create(g_zfs, poolname, + nvroot, props, fsprops) == 0) { zfs_handle_t *pool = zfs_open(g_zfs, poolname, ZFS_TYPE_FILESYSTEM); if (pool != NULL) { @@ -742,20 +777,25 @@ zpool_do_create(int argc, char **argv) ZFS_PROP_MOUNTPOINT), mountpoint) == 0); if (zfs_mount(pool, NULL, 0) == 0) - ret = zfs_share_nfs(pool); + ret = zfs_shareall(pool); zfs_close(pool); } - zpool_log_history(g_zfs, argc + optind, argv - optind, - poolname, B_TRUE, B_TRUE); } else if (libzfs_errno(g_zfs) == EZFS_INVALIDNAME) { (void) fprintf(stderr, gettext("pool name may have " "been omitted\n")); } } +errout: nvlist_free(nvroot); - + nvlist_free(fsprops); + nvlist_free(props); return (ret); +badusage: + nvlist_free(fsprops); + nvlist_free(props); + usage(B_FALSE); + return (2); } /* @@ -819,9 +859,6 @@ zpool_do_destroy(int argc, char **argv) return (1); } - zpool_log_history(g_zfs, argc + optind, argv - optind, pool, B_TRUE, - B_FALSE); - ret = (zpool_destroy(zhp) != 0); zpool_close(zhp); @@ -882,10 +919,7 @@ zpool_do_export(int argc, char **argv) continue; } - zpool_log_history(g_zfs, argc + optind, argv - optind, argv[i], - B_TRUE, B_FALSE); - - if (zpool_export(zhp) != 0) + if (zpool_export(zhp, force) != 0) ret = 1; zpool_close(zhp); @@ -919,6 +953,14 @@ max_width(zpool_handle_t *zhp, nvlist_t *nv, int depth, int max) max = ret; } + if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE, + &child, &children) == 0) { + for (c = 0; c < children; c++) + if ((ret = max_width(zhp, child[c], depth + 2, + max)) > max) + max = ret; + } + if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) == 0) { for (c = 0; c < children; c++) @@ -937,7 +979,8 @@ max_width(zpool_handle_t *zhp, nvlist_t *nv, int depth, int max) * pool, printing out the name and status for each one. */ void -print_import_config(const char *name, nvlist_t *nv, int namewidth, int depth) +print_import_config(const char *name, nvlist_t *nv, int namewidth, int depth, + boolean_t print_logs) { nvlist_t **child; uint_t c, children; @@ -952,9 +995,10 @@ print_import_config(const char *name, nvlist_t *nv, int namewidth, int depth) (uint64_t **)&vs, &c) == 0); (void) printf("\t%*s%-*s", depth, "", namewidth - depth, name); + (void) printf(" %s", zpool_state_to_name(vs->vs_state, vs->vs_aux)); if (vs->vs_aux != 0) { - (void) printf(" %-8s ", state_to_name(vs)); + (void) printf(" "); switch (vs->vs_aux) { case VDEV_AUX_OPEN_FAILED: @@ -973,12 +1017,14 @@ print_import_config(const char *name, nvlist_t *nv, int namewidth, int depth) (void) printf(gettext("newer version")); break; + case VDEV_AUX_ERR_EXCEEDED: + (void) printf(gettext("too many errors")); + break; + default: (void) printf(gettext("corrupted data")); break; } - } else { - (void) printf(" %s", state_to_name(vs)); } (void) printf("\n"); @@ -987,21 +1033,37 @@ print_import_config(const char *name, nvlist_t *nv, int namewidth, int depth) return; for (c = 0; c < children; c++) { + uint64_t is_log = B_FALSE; + + (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, + &is_log); + if ((is_log && !print_logs) || (!is_log && print_logs)) + continue; + vname = zpool_vdev_name(g_zfs, NULL, child[c]); print_import_config(vname, child[c], - namewidth, depth + 2); + namewidth, depth + 2, B_FALSE); free(vname); } - if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES, - &child, &children) != 0) - return; + if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE, + &child, &children) == 0) { + (void) printf(gettext("\tcache\n")); + for (c = 0; c < children; c++) { + vname = zpool_vdev_name(g_zfs, NULL, child[c]); + (void) printf("\t %s\n", vname); + free(vname); + } + } - (void) printf(gettext("\tspares\n")); - for (c = 0; c < children; c++) { - vname = zpool_vdev_name(g_zfs, NULL, child[c]); - (void) printf("\t %s\n", vname); - free(vname); + if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES, + &child, &children) == 0) { + (void) printf(gettext("\tspares\n")); + for (c = 0; c < children; c++) { + vname = zpool_vdev_name(g_zfs, NULL, child[c]); + (void) printf("\t %s\n", vname); + free(vname); + } } } @@ -1033,7 +1095,7 @@ show_import(nvlist_t *config) verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_STATS, (uint64_t **)&vs, &vsc) == 0); - health = state_to_health(vs->vs_state); + health = zpool_state_to_name(vs->vs_state, vs->vs_aux); reason = zpool_import_status(config, &msgid); @@ -1086,6 +1148,18 @@ show_import(nvlist_t *config) (void) printf(gettext("status: The pool was last accessed by " "another system.\n")); break; + + case ZPOOL_STATUS_FAULTED_DEV_R: + case ZPOOL_STATUS_FAULTED_DEV_NR: + (void) printf(gettext("status: One or more devices are " + "faulted.\n")); + break; + + case ZPOOL_STATUS_BAD_LOG: + (void) printf(gettext("status: An intent log record cannot be " + "read.\n")); + break; + default: /* * No other status can be seen when importing pools. @@ -1147,7 +1221,7 @@ show_import(nvlist_t *config) "but can be imported using the '-Df' flags.\n")); else if (pool_state != POOL_STATE_EXPORTED) (void) printf(gettext("\tThe pool may be active on " - "on another system, but can be imported using\n\t" + "another system, but can be imported using\n\t" "the '-f' flag.\n")); } @@ -1160,7 +1234,12 @@ show_import(nvlist_t *config) namewidth = max_width(NULL, nvroot, 0, 0); if (namewidth < 10) namewidth = 10; - print_import_config(name, nvroot, namewidth, 0); + + print_import_config(name, nvroot, namewidth, 0, B_FALSE); + if (num_logs(nvroot) > 0) { + (void) printf(gettext("\tlogs\n")); + print_import_config(name, nvroot, namewidth, 0, B_TRUE); + } if (reason == ZPOOL_STATUS_BAD_GUID_SUM) { (void) printf(gettext("\n\tAdditional devices are known to " @@ -1171,17 +1250,18 @@ show_import(nvlist_t *config) /* * Perform the import for the given configuration. This passes the heavy - * lifting off to zpool_import(), and then mounts the datasets contained within - * the pool. + * lifting off to zpool_import_props(), and then mounts the datasets contained + * within the pool. */ static int do_import(nvlist_t *config, const char *newname, const char *mntopts, - const char *altroot, int force, int argc, char **argv) + int force, nvlist_t *props, boolean_t allowfaulted) { zpool_handle_t *zhp; char *name; uint64_t state; uint64_t version; + int error = 0; verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME, &name) == 0); @@ -1190,7 +1270,7 @@ do_import(nvlist_t *config, const char *newname, const char *mntopts, ZPOOL_CONFIG_POOL_STATE, &state) == 0); verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION, &version) == 0); - if (version > ZFS_VERSION) { + if (version > SPA_VERSION) { (void) fprintf(stderr, gettext("cannot import '%s': pool " "is formatted using a newer ZFS version\n"), name); return (1); @@ -1228,15 +1308,14 @@ do_import(nvlist_t *config, const char *newname, const char *mntopts, } } - if (zpool_import(g_zfs, config, newname, altroot) != 0) + if (zpool_import_props(g_zfs, config, newname, props, + allowfaulted) != 0) return (1); if (newname != NULL) name = (char *)newname; - zpool_log_history(g_zfs, argc, argv, name, B_TRUE, B_FALSE); - - verify((zhp = zpool_open(g_zfs, name)) != NULL); + verify((zhp = zpool_open_canfail(g_zfs, name)) != NULL); if (zpool_enable_datasets(zhp, mntopts, 0) != 0) { zpool_close(zhp); @@ -1244,13 +1323,18 @@ do_import(nvlist_t *config, const char *newname, const char *mntopts, } zpool_close(zhp); - return (0); + return (error); } /* * zpool import [-d dir] [-D] - * import [-R root] [-D] [-d dir] [-f] -a - * import [-R root] [-D] [-d dir] [-f] <pool | id> [newpool] + * import [-o mntopts] [-o prop=value] ... [-R root] [-D] + * [-d dir | -c cachefile] [-f] -a + * import [-o mntopts] [-o prop=value] ... [-R root] [-D] + * [-d dir | -c cachefile] [-f] <pool | id> [newpool] + * + * -c Read pool information from a cachefile instead of searching + * devices. * * -d Scan in a specific directory, other than /dev/dsk. More than * one directory can be specified using multiple '-d' options. @@ -1264,8 +1348,15 @@ do_import(nvlist_t *config, const char *newname, const char *mntopts, * * -f Force import, even if it appears that the pool is active. * + * -F Import even in the presence of faulted vdevs. This is an + * intentionally undocumented option for testing purposes, and + * treats the pool configuration as complete, leaving any bad + * vdevs in the FAULTED state. + * * -a Import all pools found. * + * -o Set property=value and/or temporary mount options (without '='). + * * The import command scans for pools to import, and import pools based on pool * name and GUID. The pool can also be renamed as part of the import process. */ @@ -1276,26 +1367,32 @@ zpool_do_import(int argc, char **argv) int nsearch = 0; int c; int err; - nvlist_t *pools; + nvlist_t *pools = NULL; boolean_t do_all = B_FALSE; boolean_t do_destroyed = B_FALSE; - char *altroot = NULL; char *mntopts = NULL; boolean_t do_force = B_FALSE; nvpair_t *elem; nvlist_t *config; - uint64_t searchguid; - char *searchname; + uint64_t searchguid = 0; + char *searchname = NULL; + char *propval; nvlist_t *found_config; + nvlist_t *props = NULL; boolean_t first; + boolean_t allow_faulted = B_FALSE; uint64_t pool_state; + char *cachefile = NULL; /* check options */ - while ((c = getopt(argc, argv, ":Dfd:R:ao:")) != -1) { + while ((c = getopt(argc, argv, ":ac:d:DfFo:p:R:")) != -1) { switch (c) { case 'a': do_all = B_TRUE; break; + case 'c': + cachefile = optarg; + break; case 'd': if (searchdirs == NULL) { searchdirs = safe_malloc(sizeof (char *)); @@ -1315,11 +1412,31 @@ zpool_do_import(int argc, char **argv) case 'f': do_force = B_TRUE; break; + case 'F': + allow_faulted = B_TRUE; + break; case 'o': - mntopts = optarg; + if ((propval = strchr(optarg, '=')) != NULL) { + *propval = '\0'; + propval++; + if (add_prop_list(optarg, propval, + &props, B_TRUE)) + goto error; + } else { + mntopts = optarg; + } break; case 'R': - altroot = optarg; + if (add_prop_list(zpool_prop_to_name( + ZPOOL_PROP_ALTROOT), optarg, &props, B_TRUE)) + goto error; + if (nvlist_lookup_string(props, + zpool_prop_to_name(ZPOOL_PROP_CACHEFILE), + &propval) == 0) + break; + if (add_prop_list(zpool_prop_to_name( + ZPOOL_PROP_CACHEFILE), "none", &props, B_TRUE)) + goto error; break; case ':': (void) fprintf(stderr, gettext("missing argument for " @@ -1336,9 +1453,14 @@ zpool_do_import(int argc, char **argv) argc -= optind; argv += optind; + if (cachefile && nsearch != 0) { + (void) fprintf(stderr, gettext("-c is incompatible with -d\n")); + usage(B_FALSE); + } + if (searchdirs == NULL) { searchdirs = safe_malloc(sizeof (char *)); - searchdirs[0] = "/dev"; + searchdirs[0] = "/dev/dsk"; nsearch = 1; } @@ -1367,13 +1489,7 @@ zpool_do_import(int argc, char **argv) } } - if ((pools = zpool_find_import(g_zfs, nsearch, searchdirs)) == NULL) { - free(searchdirs); - return (1); - } - /* - * We now have a list of all available pools in the given directories. * Depending on the arguments given, we do one of the following: * * <none> Iterate through all pools and display information about @@ -1393,11 +1509,38 @@ zpool_do_import(int argc, char **argv) searchguid = strtoull(argv[0], &endptr, 10); if (errno != 0 || *endptr != '\0') searchname = argv[0]; - else - searchname = NULL; found_config = NULL; } + if (cachefile) { + pools = zpool_find_import_cached(g_zfs, cachefile, searchname, + searchguid); + } else if (searchname != NULL) { + pools = zpool_find_import_byname(g_zfs, nsearch, searchdirs, + searchname); + } else { + /* + * It's OK to search by guid even if searchguid is 0. + */ + pools = zpool_find_import_byguid(g_zfs, nsearch, searchdirs, + searchguid); + } + + if (pools == NULL) { + if (argc != 0) { + (void) fprintf(stderr, gettext("cannot import '%s': " + "no such pool available\n"), argv[0]); + } + free(searchdirs); + return (1); + } + + /* + * At this point we have a list of import candidate configs. Even if + * we were searching by pool name or guid, we still need to + * post-process the list to deal with pool state and possible + * duplicate names. + */ err = 0; elem = NULL; first = B_TRUE; @@ -1420,8 +1563,7 @@ zpool_do_import(int argc, char **argv) if (do_all) err |= do_import(config, NULL, mntopts, - altroot, do_force, argc + optind, - argv - optind); + do_force, props, allow_faulted); else show_import(config); } else if (searchname != NULL) { @@ -1469,8 +1611,7 @@ zpool_do_import(int argc, char **argv) err = B_TRUE; } else { err |= do_import(found_config, argc == 1 ? NULL : - argv[1], mntopts, altroot, do_force, argc + optind, - argv - optind); + argv[1], mntopts, do_force, props, allow_faulted); } } @@ -1482,6 +1623,8 @@ zpool_do_import(int argc, char **argv) (void) fprintf(stderr, gettext("no pools available to import\n")); +error: + nvlist_free(props); nvlist_free(pools); free(searchdirs); @@ -1518,7 +1661,7 @@ print_iostat_header(iostat_cbdata_t *cb) /* * Display a single statistic. */ -void +static void print_one_stat(uint64_t value) { char buf[64]; @@ -1606,6 +1749,28 @@ print_vdev_stats(zpool_handle_t *zhp, const char *name, nvlist_t *oldnv, newchild[c], cb, depth + 2); free(vname); } + + /* + * Include level 2 ARC devices in iostat output + */ + if (nvlist_lookup_nvlist_array(newnv, ZPOOL_CONFIG_L2CACHE, + &newchild, &children) != 0) + return; + + if (oldnv && nvlist_lookup_nvlist_array(oldnv, ZPOOL_CONFIG_L2CACHE, + &oldchild, &c) != 0) + return; + + if (children > 0) { + (void) printf("%-*s - - - - - " + "-\n", cb->cb_namewidth, "cache"); + for (c = 0; c < children; c++) { + vname = zpool_vdev_name(g_zfs, zhp, newchild[c]); + print_vdev_stats(zhp, vname, oldnv ? oldchild[c] : NULL, + newchild[c], cb, depth + 2); + free(vname); + } + } } static int @@ -1870,160 +2035,122 @@ zpool_do_iostat(int argc, char **argv) typedef struct list_cbdata { boolean_t cb_scripted; boolean_t cb_first; - int cb_fields[MAX_FIELDS]; - int cb_fieldcount; + zprop_list_t *cb_proplist; } list_cbdata_t; /* * Given a list of columns to display, output appropriate headers for each one. */ -void -print_header(int *fields, size_t count) +static void +print_header(zprop_list_t *pl) { - int i; - column_def_t *col; - const char *fmt; + const char *header; + boolean_t first = B_TRUE; + boolean_t right_justify; - for (i = 0; i < count; i++) { - col = &column_table[fields[i]]; - if (i != 0) + for (; pl != NULL; pl = pl->pl_next) { + if (pl->pl_prop == ZPROP_INVAL) + continue; + + if (!first) (void) printf(" "); - if (col->cd_justify == left_justify) - fmt = "%-*s"; else - fmt = "%*s"; + first = B_FALSE; + + header = zpool_prop_column_name(pl->pl_prop); + right_justify = zpool_prop_align_right(pl->pl_prop); - (void) printf(fmt, i == count - 1 ? strlen(col->cd_title) : - col->cd_width, col->cd_title); + if (pl->pl_next == NULL && !right_justify) + (void) printf("%s", header); + else if (right_justify) + (void) printf("%*s", pl->pl_width, header); + else + (void) printf("%-*s", pl->pl_width, header); } (void) printf("\n"); } -int -list_callback(zpool_handle_t *zhp, void *data) +/* + * Given a pool and a list of properties, print out all the properties according + * to the described layout. + */ +static void +print_pool(zpool_handle_t *zhp, zprop_list_t *pl, int scripted) { - list_cbdata_t *cbp = data; - nvlist_t *config; - int i; - char buf[ZPOOL_MAXNAMELEN]; - uint64_t total; - uint64_t used; - const char *fmt; - column_def_t *col; - - if (cbp->cb_first) { - if (!cbp->cb_scripted) - print_header(cbp->cb_fields, cbp->cb_fieldcount); - cbp->cb_first = B_FALSE; - } - - if (zpool_get_state(zhp) == POOL_STATE_UNAVAIL) { - config = NULL; - } else { - config = zpool_get_config(zhp, NULL); - total = zpool_get_space_total(zhp); - used = zpool_get_space_used(zhp); - } - - for (i = 0; i < cbp->cb_fieldcount; i++) { - if (i != 0) { - if (cbp->cb_scripted) + boolean_t first = B_TRUE; + char property[ZPOOL_MAXPROPLEN]; + char *propstr; + boolean_t right_justify; + int width; + + for (; pl != NULL; pl = pl->pl_next) { + if (!first) { + if (scripted) (void) printf("\t"); else (void) printf(" "); + } else { + first = B_FALSE; } - col = &column_table[cbp->cb_fields[i]]; - - switch (cbp->cb_fields[i]) { - case ZPOOL_FIELD_NAME: - (void) strlcpy(buf, zpool_get_name(zhp), sizeof (buf)); - break; - - case ZPOOL_FIELD_SIZE: - if (config == NULL) - (void) strlcpy(buf, "-", sizeof (buf)); - else - zfs_nicenum(total, buf, sizeof (buf)); - break; - - case ZPOOL_FIELD_USED: - if (config == NULL) - (void) strlcpy(buf, "-", sizeof (buf)); + right_justify = B_FALSE; + if (pl->pl_prop != ZPROP_INVAL) { + if (zpool_get_prop(zhp, pl->pl_prop, property, + sizeof (property), NULL) != 0) + propstr = "-"; else - zfs_nicenum(used, buf, sizeof (buf)); - break; + propstr = property; - case ZPOOL_FIELD_AVAILABLE: - if (config == NULL) - (void) strlcpy(buf, "-", sizeof (buf)); - else - zfs_nicenum(total - used, buf, sizeof (buf)); - break; + right_justify = zpool_prop_align_right(pl->pl_prop); + } else { + propstr = "-"; + } - case ZPOOL_FIELD_CAPACITY: - if (config == NULL) { - (void) strlcpy(buf, "-", sizeof (buf)); - } else { - uint64_t capacity = (total == 0 ? 0 : - (used * 100 / total)); - (void) snprintf(buf, sizeof (buf), "%llu%%", - (u_longlong_t)capacity); - } - break; + width = pl->pl_width; - case ZPOOL_FIELD_HEALTH: - if (config == NULL) { - (void) strlcpy(buf, "FAULTED", sizeof (buf)); - } else { - nvlist_t *nvroot; - vdev_stat_t *vs; - uint_t vsc; - - verify(nvlist_lookup_nvlist(config, - ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); - verify(nvlist_lookup_uint64_array(nvroot, - ZPOOL_CONFIG_STATS, (uint64_t **)&vs, - &vsc) == 0); - (void) strlcpy(buf, state_to_name(vs), - sizeof (buf)); - } - break; + /* + * If this is being called in scripted mode, or if this is the + * last column and it is left-justified, don't include a width + * format specifier. + */ + if (scripted || (pl->pl_next == NULL && !right_justify)) + (void) printf("%s", propstr); + else if (right_justify) + (void) printf("%*s", width, propstr); + else + (void) printf("%-*s", width, propstr); + } - case ZPOOL_FIELD_ROOT: - if (config == NULL) - (void) strlcpy(buf, "-", sizeof (buf)); - else if (zpool_get_root(zhp, buf, sizeof (buf)) != 0) - (void) strlcpy(buf, "-", sizeof (buf)); - break; - } + (void) printf("\n"); +} - if (cbp->cb_scripted) - (void) printf("%s", buf); - else { - if (col->cd_justify == left_justify) - fmt = "%-*s"; - else - fmt = "%*s"; +/* + * Generic callback function to list a pool. + */ +int +list_callback(zpool_handle_t *zhp, void *data) +{ + list_cbdata_t *cbp = data; - (void) printf(fmt, i == cbp->cb_fieldcount - 1 ? - strlen(buf) : col->cd_width, buf); - } + if (cbp->cb_first) { + if (!cbp->cb_scripted) + print_header(cbp->cb_proplist); + cbp->cb_first = B_FALSE; } - (void) printf("\n"); + print_pool(zhp, cbp->cb_proplist, cbp->cb_scripted); return (0); } /* - * zpool list [-H] [-o field[,field]*] [pool] ... + * zpool list [-H] [-o prop[,prop]*] [pool] ... * - * -H Scripted mode. Don't display headers, and separate fields by - * a single tab. - * -o List of fields to display. Defaults to all fields, or - * "name,size,used,available,capacity,health,root" + * -H Scripted mode. Don't display headers, and separate properties + * by a single tab. + * -o List of properties to display. Defaults to + * "name,size,used,available,capacity,health,altroot" * * List all pools in the system, whether or not they're healthy. Output space * statistics for each one, as well as health status summary. @@ -2034,10 +2161,9 @@ zpool_do_list(int argc, char **argv) int c; int ret; list_cbdata_t cb = { 0 }; - static char default_fields[] = - "name,size,used,available,capacity,health,root"; - char *fields = default_fields; - char *value; + static char default_props[] = + "name,size,used,available,capacity,health,altroot"; + char *props = default_props; /* check options */ while ((c = getopt(argc, argv, ":Ho:")) != -1) { @@ -2046,7 +2172,7 @@ zpool_do_list(int argc, char **argv) cb.cb_scripted = B_TRUE; break; case 'o': - fields = optarg; + props = optarg; break; case ':': (void) fprintf(stderr, gettext("missing argument for " @@ -2063,29 +2189,17 @@ zpool_do_list(int argc, char **argv) argc -= optind; argv += optind; - while (*fields != '\0') { - if (cb.cb_fieldcount == MAX_FIELDS) { - (void) fprintf(stderr, gettext("too many " - "properties given to -o option\n")); - usage(B_FALSE); - } - - if ((cb.cb_fields[cb.cb_fieldcount] = getsubopt(&fields, - column_subopts, &value)) == -1) { - (void) fprintf(stderr, gettext("invalid property " - "'%s'\n"), value); - usage(B_FALSE); - } - - cb.cb_fieldcount++; - } - + if (zprop_get_list(g_zfs, props, &cb.cb_proplist, ZFS_TYPE_POOL) != 0) + usage(B_FALSE); cb.cb_first = B_TRUE; - ret = for_each_pool(argc, argv, B_TRUE, NULL, list_callback, &cb); + ret = for_each_pool(argc, argv, B_TRUE, &cb.cb_proplist, + list_callback, &cb); - if (argc == 0 && cb.cb_first) { + zprop_free_list(cb.cb_proplist); + + if (argc == 0 && cb.cb_first && !cb.cb_scripted) { (void) printf(gettext("no pools available\n")); return (0); } @@ -2128,10 +2242,7 @@ zpool_do_attach_or_replace(int argc, char **argv, int replacing) nvlist_t *nvroot; char *poolname, *old_disk, *new_disk; zpool_handle_t *zhp; - nvlist_t *config; int ret; - int log_argc; - char **log_argv; /* check options */ while ((c = getopt(argc, argv, "f")) != -1) { @@ -2146,8 +2257,6 @@ zpool_do_attach_or_replace(int argc, char **argv, int replacing) } } - log_argc = argc; - log_argv = argv; argc -= optind; argv += optind; @@ -2190,14 +2299,15 @@ zpool_do_attach_or_replace(int argc, char **argv, int replacing) if ((zhp = zpool_open(g_zfs, poolname)) == NULL) return (1); - if ((config = zpool_get_config(zhp, NULL)) == NULL) { + if (zpool_get_config(zhp, NULL) == NULL) { (void) fprintf(stderr, gettext("pool '%s' is unavailable\n"), poolname); zpool_close(zhp); return (1); } - nvroot = make_root_vdev(config, force, B_FALSE, replacing, argc, argv); + nvroot = make_root_vdev(zhp, force, B_FALSE, replacing, B_FALSE, + argc, argv); if (nvroot == NULL) { zpool_close(zhp); return (1); @@ -2205,11 +2315,6 @@ zpool_do_attach_or_replace(int argc, char **argv, int replacing) ret = zpool_vdev_attach(zhp, old_disk, new_disk, nvroot, replacing); - if (!ret) { - zpool_log_history(g_zfs, log_argc, log_argv, poolname, B_TRUE, - B_FALSE); - } - nvlist_free(nvroot); zpool_close(zhp); @@ -2299,10 +2404,6 @@ zpool_do_detach(int argc, char **argv) ret = zpool_vdev_detach(zhp, path); - if (!ret) { - zpool_log_history(g_zfs, argc + optind, argv - optind, poolname, - B_TRUE, B_FALSE); - } zpool_close(zhp); return (ret); @@ -2311,7 +2412,6 @@ zpool_do_detach(int argc, char **argv) /* * zpool online <pool> <device> ... */ -/* ARGSUSED */ int zpool_do_online(int argc, char **argv) { @@ -2319,6 +2419,7 @@ zpool_do_online(int argc, char **argv) char *poolname; zpool_handle_t *zhp; int ret = 0; + vdev_state_t newstate; /* check options */ while ((c = getopt(argc, argv, "t")) != -1) { @@ -2349,17 +2450,26 @@ zpool_do_online(int argc, char **argv) if ((zhp = zpool_open(g_zfs, poolname)) == NULL) return (1); - for (i = 1; i < argc; i++) - if (zpool_vdev_online(zhp, argv[i]) == 0) - (void) printf(gettext("Bringing device %s online\n"), - argv[i]); - else + for (i = 1; i < argc; i++) { + if (zpool_vdev_online(zhp, argv[i], 0, &newstate) == 0) { + if (newstate != VDEV_STATE_HEALTHY) { + (void) printf(gettext("warning: device '%s' " + "onlined, but remains in faulted state\n"), + argv[i]); + if (newstate == VDEV_STATE_FAULTED) + (void) printf(gettext("use 'zpool " + "clear' to restore a faulted " + "device\n")); + else + (void) printf(gettext("use 'zpool " + "replace' to replace devices " + "that are no longer present\n")); + } + } else { ret = 1; - - if (!ret) { - zpool_log_history(g_zfs, argc + optind, argv - optind, poolname, - B_TRUE, B_FALSE); + } } + zpool_close(zhp); return (ret); @@ -2417,17 +2527,11 @@ zpool_do_offline(int argc, char **argv) if ((zhp = zpool_open(g_zfs, poolname)) == NULL) return (1); - for (i = 1; i < argc; i++) - if (zpool_vdev_offline(zhp, argv[i], istmp) == 0) - (void) printf(gettext("Bringing device %s offline\n"), - argv[i]); - else + for (i = 1; i < argc; i++) { + if (zpool_vdev_offline(zhp, argv[i], istmp) != 0) ret = 1; - - if (!ret) { - zpool_log_history(g_zfs, argc + optind, argv - optind, poolname, - B_TRUE, B_FALSE); } + zpool_close(zhp); return (ret); @@ -2458,14 +2562,12 @@ zpool_do_clear(int argc, char **argv) pool = argv[1]; device = argc == 3 ? argv[2] : NULL; - if ((zhp = zpool_open(g_zfs, pool)) == NULL) + if ((zhp = zpool_open_canfail(g_zfs, pool)) == NULL) return (1); if (zpool_clear(zhp, device) != 0) ret = 1; - if (!ret) - zpool_log_history(g_zfs, argc, argv, pool, B_TRUE, B_FALSE); zpool_close(zhp); return (ret); @@ -2494,11 +2596,6 @@ scrub_callback(zpool_handle_t *zhp, void *data) err = zpool_scrub(zhp, cb->cb_type); - if (!err) { - zpool_log_history(g_zfs, cb->cb_argc, cb->cb_argv, - zpool_get_name(zhp), B_TRUE, B_FALSE); - } - return (err != 0); } @@ -2559,7 +2656,7 @@ print_scrub_status(nvlist_t *nvroot) uint_t vsc; time_t start, end, now; double fraction_done; - uint64_t examined, total, minutes_left; + uint64_t examined, total, minutes_left, minutes_taken; char *scrub_type; verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_STATS, @@ -2583,8 +2680,13 @@ print_scrub_status(nvlist_t *nvroot) total = vs->vs_alloc; if (end != 0) { - (void) printf(gettext("%s %s with %llu errors on %s"), + minutes_taken = (uint64_t)((end - start) / 60); + + (void) printf(gettext("%s %s after %lluh%um with %llu errors " + "on %s"), scrub_type, vs->vs_scrub_complete ? "completed" : "stopped", + (u_longlong_t)(minutes_taken / 60), + (uint_t)(minutes_taken % 60), (u_longlong_t)vs->vs_scrub_errors, ctime(&end)); return; } @@ -2597,9 +2699,12 @@ print_scrub_status(nvlist_t *nvroot) fraction_done = (double)examined / total; minutes_left = (uint64_t)((now - start) * (1 - fraction_done) / fraction_done / 60); + minutes_taken = (uint64_t)((now - start) / 60); - (void) printf(gettext("%s in progress, %.2f%% done, %lluh%um to go\n"), - scrub_type, 100 * fraction_done, + (void) printf(gettext("%s in progress for %lluh%um, %.2f%% done, " + "%lluh%um to go\n"), + scrub_type, (u_longlong_t)(minutes_taken / 60), + (uint_t)(minutes_taken % 60), 100 * fraction_done, (u_longlong_t)(minutes_left / 60), (uint_t)(minutes_left % 60)); } @@ -2653,7 +2758,7 @@ find_spare(zpool_handle_t *zhp, void *data) */ void print_status_config(zpool_handle_t *zhp, const char *name, nvlist_t *nv, - int namewidth, int depth, boolean_t isspare) + int namewidth, int depth, boolean_t isspare, boolean_t print_logs) { nvlist_t **child; uint_t c, children; @@ -2662,7 +2767,7 @@ print_status_config(zpool_handle_t *zhp, const char *name, nvlist_t *nv, char *vname; uint64_t notpresent; spare_cbdata_t cb; - const char *state; + char *state; verify(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_STATS, (uint64_t **)&vs, &c) == 0); @@ -2671,7 +2776,7 @@ print_status_config(zpool_handle_t *zhp, const char *name, nvlist_t *nv, &child, &children) != 0) children = 0; - state = state_to_name(vs); + state = zpool_state_to_name(vs->vs_state, vs->vs_aux); if (isspare) { /* * For hot spares, we use the terms 'INUSE' and 'AVAILABLE' for @@ -2736,6 +2841,18 @@ print_status_config(zpool_handle_t *zhp, const char *name, nvlist_t *nv, } break; + case VDEV_AUX_ERR_EXCEEDED: + (void) printf(gettext("too many errors")); + break; + + case VDEV_AUX_IO_FAILURE: + (void) printf(gettext("experienced I/O failures")); + break; + + case VDEV_AUX_BAD_LOG: + (void) printf(gettext("bad intent log")); + break; + default: (void) printf(gettext("corrupted data")); break; @@ -2753,9 +2870,15 @@ print_status_config(zpool_handle_t *zhp, const char *name, nvlist_t *nv, (void) printf("\n"); for (c = 0; c < children; c++) { + uint64_t is_log = B_FALSE; + + (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, + &is_log); + if ((is_log && !print_logs) || (!is_log && print_logs)) + continue; vname = zpool_vdev_name(g_zfs, zhp, child[c]); print_status_config(zhp, vname, child[c], - namewidth, depth + 2, isspare); + namewidth, depth + 2, isspare, B_FALSE); free(vname); } } @@ -2763,7 +2886,7 @@ print_status_config(zpool_handle_t *zhp, const char *name, nvlist_t *nv, static void print_error_log(zpool_handle_t *zhp) { - nvlist_t *nverrlist; + nvlist_t *nverrlist = NULL; nvpair_t *elem; char *pathname; size_t len = MAXPATHLEN * 2; @@ -2810,7 +2933,27 @@ print_spares(zpool_handle_t *zhp, nvlist_t **spares, uint_t nspares, for (i = 0; i < nspares; i++) { name = zpool_vdev_name(g_zfs, zhp, spares[i]); print_status_config(zhp, name, spares[i], - namewidth, 2, B_TRUE); + namewidth, 2, B_TRUE, B_FALSE); + free(name); + } +} + +static void +print_l2cache(zpool_handle_t *zhp, nvlist_t **l2cache, uint_t nl2cache, + int namewidth) +{ + uint_t i; + char *name; + + if (nl2cache == 0) + return; + + (void) printf(gettext("\tcache\n")); + + for (i = 0; i < nl2cache; i++) { + name = zpool_vdev_name(g_zfs, zhp, l2cache[i]); + print_status_config(zhp, name, l2cache[i], + namewidth, 2, B_FALSE, B_FALSE); free(name); } } @@ -2869,7 +3012,7 @@ status_callback(zpool_handle_t *zhp, void *data) &nvroot) == 0); verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_STATS, (uint64_t **)&vs, &c) == 0); - health = state_to_name(vs); + health = zpool_state_to_name(vs->vs_state, vs->vs_aux); (void) printf(gettext(" pool: %s\n"), zpool_get_name(zhp)); (void) printf(gettext(" state: %s\n"), health); @@ -2972,6 +3115,45 @@ status_callback(zpool_handle_t *zhp, void *data) "backup.\n")); break; + case ZPOOL_STATUS_FAULTED_DEV_R: + (void) printf(gettext("status: One or more devices are " + "faulted in response to persistent errors.\n\tSufficient " + "replicas exist for the pool to continue functioning " + "in a\n\tdegraded state.\n")); + (void) printf(gettext("action: Replace the faulted device, " + "or use 'zpool clear' to mark the device\n\trepaired.\n")); + break; + + case ZPOOL_STATUS_FAULTED_DEV_NR: + (void) printf(gettext("status: One or more devices are " + "faulted in response to persistent errors. There are " + "insufficient replicas for the pool to\n\tcontinue " + "functioning.\n")); + (void) printf(gettext("action: Destroy and re-create the pool " + "from a backup source. Manually marking the device\n" + "\trepaired using 'zpool clear' may allow some data " + "to be recovered.\n")); + break; + + case ZPOOL_STATUS_IO_FAILURE_WAIT: + case ZPOOL_STATUS_IO_FAILURE_CONTINUE: + (void) printf(gettext("status: One or more devices are " + "faulted in response to IO failures.\n")); + (void) printf(gettext("action: Make sure the affected devices " + "are connected, then run 'zpool clear'.\n")); + break; + + case ZPOOL_STATUS_BAD_LOG: + (void) printf(gettext("status: An intent log record " + "could not be read.\n" + "\tWaiting for adminstrator intervention to fix the " + "faulted pool.\n")); + (void) printf(gettext("action: Either restore the affected " + "device(s) and run 'zpool online',\n" + "\tor ignore the intent log records by running " + "'zpool clear'.\n")); + break; + default: /* * The remaining errors can't actually be generated, yet. @@ -2986,8 +3168,8 @@ status_callback(zpool_handle_t *zhp, void *data) if (config != NULL) { int namewidth; uint64_t nerr; - nvlist_t **spares; - uint_t nspares; + nvlist_t **spares, **l2cache; + uint_t nspares, nl2cache; (void) printf(gettext(" scrub: ")); @@ -3001,7 +3183,14 @@ status_callback(zpool_handle_t *zhp, void *data) (void) printf(gettext("\t%-*s %-8s %5s %5s %5s\n"), namewidth, "NAME", "STATE", "READ", "WRITE", "CKSUM"); print_status_config(zhp, zpool_get_name(zhp), nvroot, - namewidth, 0, B_FALSE); + namewidth, 0, B_FALSE, B_FALSE); + if (num_logs(nvroot) > 0) + print_status_config(zhp, "logs", nvroot, namewidth, 0, + B_FALSE, B_TRUE); + + if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE, + &l2cache, &nl2cache) == 0) + print_l2cache(zhp, l2cache, nl2cache, namewidth); if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, &spares, &nspares) == 0) @@ -3016,7 +3205,7 @@ status_callback(zpool_handle_t *zhp, void *data) * precise count by fetching the entire log and * uniquifying the results. */ - if (nerr < 100 && !cbp->cb_verbose && + if (nerr > 0 && nerr < 100 && !cbp->cb_verbose && zpool_get_errlog(zhp, &nverrlist) == 0) { nvpair_t *elem; @@ -3103,6 +3292,7 @@ typedef struct upgrade_cbdata { int cb_first; int cb_newer; int cb_argc; + uint64_t cb_version; char **cb_argv; } upgrade_cbdata_t; @@ -3118,7 +3308,7 @@ upgrade_cb(zpool_handle_t *zhp, void *arg) verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION, &version) == 0); - if (!cbp->cb_newer && version < ZFS_VERSION) { + if (!cbp->cb_newer && version < SPA_VERSION) { if (!cbp->cb_all) { if (cbp->cb_first) { (void) printf(gettext("The following pools are " @@ -3135,16 +3325,13 @@ upgrade_cb(zpool_handle_t *zhp, void *arg) zpool_get_name(zhp)); } else { cbp->cb_first = B_FALSE; - ret = zpool_upgrade(zhp); + ret = zpool_upgrade(zhp, cbp->cb_version); if (!ret) { - zpool_log_history(g_zfs, cbp->cb_argc, - cbp->cb_argv, zpool_get_name(zhp), B_TRUE, - B_FALSE); (void) printf(gettext("Successfully upgraded " - "'%s'\n"), zpool_get_name(zhp)); + "'%s'\n\n"), zpool_get_name(zhp)); } } - } else if (cbp->cb_newer && version > ZFS_VERSION) { + } else if (cbp->cb_newer && version > SPA_VERSION) { assert(!cbp->cb_all); if (cbp->cb_first) { @@ -3168,29 +3355,37 @@ upgrade_cb(zpool_handle_t *zhp, void *arg) static int upgrade_one(zpool_handle_t *zhp, void *data) { - nvlist_t *config; - uint64_t version; - int ret; upgrade_cbdata_t *cbp = data; + uint64_t cur_version; + int ret; - config = zpool_get_config(zhp, NULL); - verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION, - &version) == 0); + if (strcmp("log", zpool_get_name(zhp)) == 0) { + (void) printf(gettext("'log' is now a reserved word\n" + "Pool 'log' must be renamed using export and import" + " to upgrade.\n")); + return (1); + } - if (version == ZFS_VERSION) { + cur_version = zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL); + if (cur_version > cbp->cb_version) { + (void) printf(gettext("Pool '%s' is already formatted " + "using more current version '%llu'.\n"), + zpool_get_name(zhp), cur_version); + return (0); + } + if (cur_version == cbp->cb_version) { (void) printf(gettext("Pool '%s' is already formatted " "using the current version.\n"), zpool_get_name(zhp)); return (0); } - ret = zpool_upgrade(zhp); + ret = zpool_upgrade(zhp, cbp->cb_version); if (!ret) { - zpool_log_history(g_zfs, cbp->cb_argc, cbp->cb_argv, - zpool_get_name(zhp), B_TRUE, B_FALSE); (void) printf(gettext("Successfully upgraded '%s' " - "from version %llu to version %llu\n"), zpool_get_name(zhp), - (u_longlong_t)version, (u_longlong_t)ZFS_VERSION); + "from version %llu to version %llu\n\n"), + zpool_get_name(zhp), (u_longlong_t)cur_version, + (u_longlong_t)cbp->cb_version); } return (ret != 0); @@ -3199,7 +3394,7 @@ upgrade_one(zpool_handle_t *zhp, void *data) /* * zpool upgrade * zpool upgrade -v - * zpool upgrade <-a | pool> + * zpool upgrade [-V version] <-a | pool ...> * * With no arguments, display downrev'd ZFS pool available for upgrade. * Individual pools can be upgraded by specifying the pool, and '-a' will @@ -3212,9 +3407,11 @@ zpool_do_upgrade(int argc, char **argv) upgrade_cbdata_t cb = { 0 }; int ret = 0; boolean_t showversions = B_FALSE; + char *end; + /* check options */ - while ((c = getopt(argc, argv, "av")) != -1) { + while ((c = getopt(argc, argv, "avV:")) != -1) { switch (c) { case 'a': cb.cb_all = B_TRUE; @@ -3222,6 +3419,15 @@ zpool_do_upgrade(int argc, char **argv) case 'v': showversions = B_TRUE; break; + case 'V': + cb.cb_version = strtoll(optarg, &end, 10); + if (*end != '\0' || cb.cb_version > SPA_VERSION || + cb.cb_version < SPA_VERSION_1) { + (void) fprintf(stderr, + gettext("invalid version '%s'\n"), optarg); + usage(B_FALSE); + } + break; case '?': (void) fprintf(stderr, gettext("invalid option '%c'\n"), optopt); @@ -3234,6 +3440,14 @@ zpool_do_upgrade(int argc, char **argv) argc -= optind; argv += optind; + if (cb.cb_version == 0) { + cb.cb_version = SPA_VERSION; + } else if (!cb.cb_all && argc == 0) { + (void) fprintf(stderr, gettext("-V option is " + "incompatible with other arguments\n")); + usage(B_FALSE); + } + if (showversions) { if (cb.cb_all || argc != 0) { (void) fprintf(stderr, gettext("-v option is " @@ -3242,14 +3456,14 @@ zpool_do_upgrade(int argc, char **argv) } } else if (cb.cb_all) { if (argc != 0) { - (void) fprintf(stderr, gettext("-a option is " - "incompatible with other arguments\n")); + (void) fprintf(stderr, gettext("-a option should not " + "be used along with a pool name\n")); usage(B_FALSE); } } - (void) printf(gettext("This system is currently running ZFS version " - "%llu.\n\n"), ZFS_VERSION); + (void) printf(gettext("This system is currently running " + "ZFS pool version %llu.\n\n"), SPA_VERSION); cb.cb_first = B_TRUE; if (showversions) { (void) printf(gettext("The following versions are " @@ -3265,8 +3479,16 @@ zpool_do_upgrade(int argc, char **argv) (void) printf(gettext(" 4 zpool history\n")); (void) printf(gettext(" 5 Compression using the gzip " "algorithm\n")); - (void) printf(gettext(" 6 bootfs pool property ")); - (void) printf(gettext("\nFor more information on a particular " + (void) printf(gettext(" 6 bootfs pool property\n")); + (void) printf(gettext(" 7 Separate intent log devices\n")); + (void) printf(gettext(" 8 Delegated administration\n")); + (void) printf(gettext(" 9 refquota and refreservation " + "properties\n")); + (void) printf(gettext(" 10 Cache devices\n")); + (void) printf(gettext(" 11 Improved scrub performance\n")); + (void) printf(gettext(" 12 Snapshot properties\n")); + (void) printf(gettext(" 13 snapused property\n")); + (void) printf(gettext("For more information on a particular " "version, including supported releases, see:\n\n")); (void) printf("http://www.opensolaris.org/os/community/zfs/" "version/N\n\n"); @@ -3306,6 +3528,53 @@ zpool_do_upgrade(int argc, char **argv) return (ret); } +typedef struct hist_cbdata { + boolean_t first; + int longfmt; + int internal; +} hist_cbdata_t; + +char *hist_event_table[LOG_END] = { + "invalid event", + "pool create", + "vdev add", + "pool remove", + "pool destroy", + "pool export", + "pool import", + "vdev attach", + "vdev replace", + "vdev detach", + "vdev online", + "vdev offline", + "vdev upgrade", + "pool clear", + "pool scrub", + "pool property set", + "create", + "clone", + "destroy", + "destroy_begin_sync", + "inherit", + "property set", + "quota set", + "permission update", + "permission remove", + "permission who remove", + "promote", + "receive", + "rename", + "reservation set", + "replay_inc_sync", + "replay_full_sync", + "rollback", + "snapshot", + "filesystem version upgrade", + "refquota set", + "refreservation set", + "pool scrub done", +}; + /* * Print out the command history for a specific pool. */ @@ -3316,13 +3585,22 @@ get_history_one(zpool_handle_t *zhp, void *data) nvlist_t **records; uint_t numrecords; char *cmdstr; + char *pathstr; uint64_t dst_time; time_t tsec; struct tm t; char tbuf[30]; int ret, i; + uint64_t who; + struct passwd *pwd; + char *hostname; + char *zonename; + char internalstr[MAXPATHLEN]; + hist_cbdata_t *cb = (hist_cbdata_t *)data; + uint64_t txg; + uint64_t ievent; - *(boolean_t *)data = B_FALSE; + cb->first = B_FALSE; (void) printf(gettext("History for '%s':\n"), zpool_get_name(zhp)); @@ -3333,14 +3611,65 @@ get_history_one(zpool_handle_t *zhp, void *data) &records, &numrecords) == 0); for (i = 0; i < numrecords; i++) { if (nvlist_lookup_uint64(records[i], ZPOOL_HIST_TIME, - &dst_time) == 0) { - verify(nvlist_lookup_string(records[i], ZPOOL_HIST_CMD, - &cmdstr) == 0); - tsec = dst_time; - (void) localtime_r(&tsec, &t); - (void) strftime(tbuf, sizeof (tbuf), "%F.%T", &t); - (void) printf("%s %s\n", tbuf, cmdstr); + &dst_time) != 0) + continue; + + /* is it an internal event or a standard event? */ + if (nvlist_lookup_string(records[i], ZPOOL_HIST_CMD, + &cmdstr) != 0) { + if (cb->internal == 0) + continue; + + if (nvlist_lookup_uint64(records[i], + ZPOOL_HIST_INT_EVENT, &ievent) != 0) + continue; + verify(nvlist_lookup_uint64(records[i], + ZPOOL_HIST_TXG, &txg) == 0); + verify(nvlist_lookup_string(records[i], + ZPOOL_HIST_INT_STR, &pathstr) == 0); + if (ievent >= LOG_END) + continue; + (void) snprintf(internalstr, + sizeof (internalstr), + "[internal %s txg:%lld] %s", + hist_event_table[ievent], txg, + pathstr); + cmdstr = internalstr; + } + tsec = dst_time; + (void) localtime_r(&tsec, &t); + (void) strftime(tbuf, sizeof (tbuf), "%F.%T", &t); + (void) printf("%s %s", tbuf, cmdstr); + + if (!cb->longfmt) { + (void) printf("\n"); + continue; } + (void) printf(" ["); + if (nvlist_lookup_uint64(records[i], + ZPOOL_HIST_WHO, &who) == 0) { + pwd = getpwuid((uid_t)who); + if (pwd) + (void) printf("user %s on", + pwd->pw_name); + else + (void) printf("user %d on", + (int)who); + } else { + (void) printf(gettext("no info]\n")); + continue; + } + if (nvlist_lookup_string(records[i], + ZPOOL_HIST_HOST, &hostname) == 0) { + (void) printf(" %s", hostname); + } + if (nvlist_lookup_string(records[i], + ZPOOL_HIST_ZONE, &zonename) == 0) { + (void) printf(":%s", zonename); + } + + (void) printf("]"); + (void) printf("\n"); } (void) printf("\n"); nvlist_free(nvhis); @@ -3353,19 +3682,38 @@ get_history_one(zpool_handle_t *zhp, void *data) * * Displays the history of commands that modified pools. */ + + int zpool_do_history(int argc, char **argv) { - boolean_t first = B_TRUE; + hist_cbdata_t cbdata = { 0 }; int ret; + int c; + cbdata.first = B_TRUE; + /* check options */ + while ((c = getopt(argc, argv, "li")) != -1) { + switch (c) { + case 'l': + cbdata.longfmt = 1; + break; + case 'i': + cbdata.internal = 1; + break; + case '?': + (void) fprintf(stderr, gettext("invalid option '%c'\n"), + optopt); + usage(B_FALSE); + } + } argc -= optind; argv += optind; ret = for_each_pool(argc, argv, B_FALSE, NULL, get_history_one, - &first); + &cbdata); - if (argc == 0 && first == B_TRUE) { + if (argc == 0 && cbdata.first == B_TRUE) { (void) printf(gettext("no pools available\n")); return (0); } @@ -3376,17 +3724,18 @@ zpool_do_history(int argc, char **argv) static int get_callback(zpool_handle_t *zhp, void *data) { - libzfs_get_cbdata_t *cbp = (libzfs_get_cbdata_t *)data; + zprop_get_cbdata_t *cbp = (zprop_get_cbdata_t *)data; char value[MAXNAMELEN]; - zfs_source_t srctype; - zpool_proplist_t *pl; + zprop_source_t srctype; + zprop_list_t *pl; for (pl = cbp->cb_proplist; pl != NULL; pl = pl->pl_next) { /* - * Skip the special fake placeholder. + * Skip the special fake placeholder. This will also skip + * over the name property when 'all' is specified. */ - if (pl->pl_prop == ZFS_PROP_NAME && + if (pl->pl_prop == ZPOOL_PROP_NAME && pl == cbp->cb_proplist) continue; @@ -3394,7 +3743,7 @@ get_callback(zpool_handle_t *zhp, void *data) value, sizeof (value), &srctype) != 0) continue; - libzfs_print_one_property(zpool_get_name(zhp), cbp, + zprop_print_one_property(zpool_get_name(zhp), cbp, zpool_prop_to_name(pl->pl_prop), value, srctype, NULL); } return (0); @@ -3403,25 +3752,27 @@ get_callback(zpool_handle_t *zhp, void *data) int zpool_do_get(int argc, char **argv) { - libzfs_get_cbdata_t cb = { 0 }; - zpool_proplist_t fake_name = { 0 }; + zprop_get_cbdata_t cb = { 0 }; + zprop_list_t fake_name = { 0 }; int ret; if (argc < 3) usage(B_FALSE); cb.cb_first = B_TRUE; - cb.cb_sources = ZFS_SRC_ALL; + cb.cb_sources = ZPROP_SRC_ALL; cb.cb_columns[0] = GET_COL_NAME; cb.cb_columns[1] = GET_COL_PROPERTY; cb.cb_columns[2] = GET_COL_VALUE; cb.cb_columns[3] = GET_COL_SOURCE; + cb.cb_type = ZFS_TYPE_POOL; - if (zpool_get_proplist(g_zfs, argv[1], &cb.cb_proplist) != 0) + if (zprop_get_list(g_zfs, argv[1], &cb.cb_proplist, + ZFS_TYPE_POOL) != 0) usage(B_FALSE); if (cb.cb_proplist != NULL) { - fake_name.pl_prop = ZFS_PROP_NAME; + fake_name.pl_prop = ZPOOL_PROP_NAME; fake_name.pl_width = strlen(gettext("NAME")); fake_name.pl_next = cb.cb_proplist; cb.cb_proplist = &fake_name; @@ -3431,9 +3782,9 @@ zpool_do_get(int argc, char **argv) get_callback, &cb); if (cb.cb_proplist == &fake_name) - zfs_free_proplist(fake_name.pl_next); + zprop_free_list(fake_name.pl_next); else - zfs_free_proplist(cb.cb_proplist); + zprop_free_list(cb.cb_proplist); return (ret); } @@ -3500,11 +3851,6 @@ zpool_do_set(int argc, char **argv) error = for_each_pool(argc - 2, argv + 2, B_TRUE, NULL, set_callback, &cb); - if (cb.cb_any_successful) { - *(cb.cb_value - 1) = '='; - zpool_log_history(g_zfs, argc, argv, argv[2], B_FALSE, B_FALSE); - } - return (error); } @@ -3531,7 +3877,6 @@ main(int argc, char **argv) int ret; int i; char *cmdname; - int found = 0; (void) setlocale(LC_ALL, ""); (void) textdomain(TEXT_DOMAIN); @@ -3562,26 +3907,29 @@ main(int argc, char **argv) if (strcmp(cmdname, "-?") == 0) usage(B_TRUE); + zpool_set_history_str("zpool", argc, argv, history_str); + verify(zpool_stage_history(g_zfs, history_str) == 0); + /* * Run the appropriate command. */ if (find_command_idx(cmdname, &i) == 0) { current_command = &command_table[i]; ret = command_table[i].func(argc - 1, argv + 1); - found++; - } - - /* - * 'freeze' is a vile debugging abomination, so we treat it as such. - */ - if (strcmp(cmdname, "freeze") == 0 && argc == 3) { + } else if (strchr(cmdname, '=')) { + verify(find_command_idx("set", &i) == 0); + current_command = &command_table[i]; + ret = command_table[i].func(argc, argv); + } else if (strcmp(cmdname, "freeze") == 0 && argc == 3) { + /* + * 'freeze' is a vile debugging abomination, so we treat + * it as such. + */ char buf[16384]; int fd = open(ZFS_DEV, O_RDWR); (void) strcpy((void *)buf, argv[2]); return (!!ioctl(fd, ZFS_IOC_POOL_FREEZE, buf)); - } - - if (!found) { + } else { (void) fprintf(stderr, gettext("unrecognized " "command '%s'\n"), cmdname); usage(B_FALSE); diff --git a/cddl/contrib/opensolaris/cmd/zpool/zpool_util.c b/cddl/contrib/opensolaris/cmd/zpool/zpool_util.c index 8eb9c81..f44da4f 100644 --- a/cddl/contrib/opensolaris/cmd/zpool/zpool_util.c +++ b/cddl/contrib/opensolaris/cmd/zpool/zpool_util.c @@ -19,7 +19,7 @@ * CDDL HEADER END */ /* - * Copyright 2006 Sun Microsystems, Inc. All rights reserved. + * Copyright 2007 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ @@ -77,3 +77,28 @@ zpool_no_memory(void) gettext("internal error: out of memory\n")); exit(1); } + +/* + * Return the number of logs in supplied nvlist + */ +uint_t +num_logs(nvlist_t *nv) +{ + uint_t nlogs = 0; + uint_t c, children; + nvlist_t **child; + + if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, + &child, &children) != 0) + return (0); + + for (c = 0; c < children; c++) { + uint64_t is_log = B_FALSE; + + (void) nvlist_lookup_uint64(child[c], ZPOOL_CONFIG_IS_LOG, + &is_log); + if (is_log) + nlogs++; + } + return (nlogs); +} diff --git a/cddl/contrib/opensolaris/cmd/zpool/zpool_util.h b/cddl/contrib/opensolaris/cmd/zpool/zpool_util.h index cb05bda..e82f320 100644 --- a/cddl/contrib/opensolaris/cmd/zpool/zpool_util.h +++ b/cddl/contrib/opensolaris/cmd/zpool/zpool_util.h @@ -19,15 +19,13 @@ * CDDL HEADER END */ /* - * Copyright 2007 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ #ifndef ZPOOL_UTIL_H #define ZPOOL_UTIL_H -#pragma ident "%Z%%M% %I% %E% SMI" - #include <libnvpair.h> #include <libzfs.h> @@ -41,22 +39,24 @@ extern "C" { void *safe_malloc(size_t); char *safe_strdup(const char *); void zpool_no_memory(void); +uint_t num_logs(nvlist_t *nv); /* * Virtual device functions */ -nvlist_t *make_root_vdev(nvlist_t *poolconfig, int force, int check_rep, - boolean_t isreplace, int argc, char **argv); + +nvlist_t *make_root_vdev(zpool_handle_t *zhp, int force, int check_rep, + boolean_t isreplace, boolean_t dryrun, int argc, char **argv); /* * Pool list functions */ -int for_each_pool(int, char **, boolean_t unavail, zpool_proplist_t **, +int for_each_pool(int, char **, boolean_t unavail, zprop_list_t **, zpool_iter_f, void *); typedef struct zpool_list zpool_list_t; -zpool_list_t *pool_list_get(int, char **, zpool_proplist_t **, int *); +zpool_list_t *pool_list_get(int, char **, zprop_list_t **, int *); void pool_list_update(zpool_list_t *); int pool_list_iter(zpool_list_t *, int unavail, zpool_iter_f, void *); void pool_list_free(zpool_list_t *); diff --git a/cddl/contrib/opensolaris/cmd/zpool/zpool_vdev.c b/cddl/contrib/opensolaris/cmd/zpool/zpool_vdev.c index cfed1f0..35a636c 100644 --- a/cddl/contrib/opensolaris/cmd/zpool/zpool_vdev.c +++ b/cddl/contrib/opensolaris/cmd/zpool/zpool_vdev.c @@ -20,12 +20,10 @@ */ /* - * Copyright 2007 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ -#pragma ident "%Z%%M% %I% %E% SMI" - /* * Functions to convert between a list of vdevs and an nvlist representing the * configuration. Each entry in the list can be one of: @@ -48,8 +46,8 @@ * Hot spares are a special case, and passed down as an array of disk vdevs, at * the same level as the root of the vdev tree. * - * The only function exported by this file is 'get_vdev_spec'. The function - * performs several passes: + * The only function exported by this file is 'make_root_vdev'. The + * function performs several passes: * * 1. Construct the vdev specification. Performs syntax validation and * makes sure each device is valid. @@ -59,6 +57,7 @@ * 3. Check for replication errors if the 'force' flag is not specified. * validates that the replication level is consistent across the * entire pool. + * 4. Call libzfs to label any whole disks with an EFI label. */ #include <assert.h> @@ -76,8 +75,6 @@ #include <sys/mntent.h> #include <libgeom.h> -#include <libzfs.h> - #include "zpool_util.h" /* @@ -111,53 +108,105 @@ vdev_error(const char *fmt, ...) } /* - * Validate a GEOM provider. + * Check that a file is valid. All we can do in this case is check that it's + * not in use by another pool, and not in use by swap. */ static int -check_provider(const char *name, boolean_t force, boolean_t isspare) +check_file(const char *file, boolean_t force, boolean_t isspare) { - struct gmesh mesh; - struct gclass *mp; - struct ggeom *gp; - struct gprovider *pp; - int rv; - - /* XXX: What to do with isspare? */ - - if (strncmp(name, _PATH_DEV, sizeof(_PATH_DEV) - 1) == 0) - name += sizeof(_PATH_DEV) - 1; - - rv = geom_gettree(&mesh); - assert(rv == 0); - - pp = NULL; - LIST_FOREACH(mp, &mesh.lg_class, lg_class) { - LIST_FOREACH(gp, &mp->lg_geom, lg_geom) { - LIST_FOREACH(pp, &gp->lg_provider, lg_provider) { - if (strcmp(pp->lg_name, name) == 0) - goto out; + char *name; + int fd; + int ret = 0; + int err; + pool_state_t state; + boolean_t inuse; + +#if 0 + if (dm_inuse_swap(file, &err)) { + if (err) + libdiskmgt_error(err); + else + vdev_error(gettext("%s is currently used by swap. " + "Please see swap(1M).\n"), file); + return (-1); + } +#endif + + if ((fd = open(file, O_RDONLY)) < 0) + return (0); + + if (zpool_in_use(g_zfs, fd, &state, &name, &inuse) == 0 && inuse) { + const char *desc; + + switch (state) { + case POOL_STATE_ACTIVE: + desc = gettext("active"); + break; + + case POOL_STATE_EXPORTED: + desc = gettext("exported"); + break; + + case POOL_STATE_POTENTIALLY_ACTIVE: + desc = gettext("potentially active"); + break; + + default: + desc = gettext("unknown"); + break; + } + + /* + * Allow hot spares to be shared between pools. + */ + if (state == POOL_STATE_SPARE && isspare) + return (0); + + if (state == POOL_STATE_ACTIVE || + state == POOL_STATE_SPARE || !force) { + switch (state) { + case POOL_STATE_SPARE: + vdev_error(gettext("%s is reserved as a hot " + "spare for pool %s\n"), file, name); + break; + default: + vdev_error(gettext("%s is part of %s pool " + "'%s'\n"), file, desc, name); + break; } + ret = -1; } + + free(name); } -out: - rv = -1; - if (pp == NULL) - vdev_error("no such provider %s\n", name); - else { - int acr, acw, ace; - - VERIFY(sscanf(pp->lg_mode, "r%dw%de%d", &acr, &acw, &ace) == 3); - if (acw == 0 && ace == 0) - rv = 0; - else - vdev_error("%s is in use (%s)\n", name, pp->lg_mode); - } - geom_deletetree(&mesh); - return (rv); + + (void) close(fd); + return (ret); +} + +static int +check_provider(const char *name, boolean_t force, boolean_t isspare) +{ + char path[MAXPATHLEN]; + + if (strncmp(name, _PATH_DEV, sizeof(_PATH_DEV) - 1) != 0) + snprintf(path, sizeof(path), "%s%s", _PATH_DEV, name); + else + strlcpy(path, name, sizeof(path)); + + return (check_file(path, force, isspare)); } +/* + * By "whole disk" we mean an entire physical disk (something we can + * label, toggle the write cache on, etc.) as opposed to the full + * capacity of a pseudo-device such as lofi or did. We act as if we + * are labeling the disk, which should be a pretty good test of whether + * it's a viable device or not. Returns B_TRUE if it is and B_FALSE if + * it isn't. + */ static boolean_t -is_provider(const char *name) +is_whole_disk(const char *name) { int fd; @@ -167,8 +216,8 @@ is_provider(const char *name) return (B_TRUE); } return (B_FALSE); - } + /* * Create a leaf vdev. Determine if this is a GEOM provider. * Valid forms for a leaf vdev are: @@ -176,25 +225,81 @@ is_provider(const char *name) * /dev/xxx Complete path to a GEOM provider * xxx Shorthand for /dev/xxx */ -nvlist_t * -make_leaf_vdev(const char *arg) +static nvlist_t * +make_leaf_vdev(const char *arg, uint64_t is_log) { - char ident[DISK_IDENT_SIZE], path[MAXPATHLEN]; + char path[MAXPATHLEN]; struct stat64 statbuf; nvlist_t *vdev = NULL; char *type = NULL; boolean_t wholedisk = B_FALSE; - if (strncmp(arg, _PATH_DEV, sizeof(_PATH_DEV) - 1) == 0) - strlcpy(path, arg, sizeof (path)); - else - snprintf(path, sizeof (path), "%s%s", _PATH_DEV, arg); + /* + * Determine what type of vdev this is, and put the full path into + * 'path'. We detect whether this is a device of file afterwards by + * checking the st_mode of the file. + */ + if (arg[0] == '/') { + /* + * Complete device or file path. Exact type is determined by + * examining the file descriptor afterwards. + */ + wholedisk = is_whole_disk(arg); + if (!wholedisk && (stat64(arg, &statbuf) != 0)) { + (void) fprintf(stderr, + gettext("cannot open '%s': %s\n"), + arg, strerror(errno)); + return (NULL); + } + + (void) strlcpy(path, arg, sizeof (path)); + } else { + /* + * This may be a short path for a device, or it could be total + * gibberish. Check to see if it's a known device in + * /dev/dsk/. As part of this check, see if we've been given a + * an entire disk (minus the slice number). + */ + if (strncmp(arg, _PATH_DEV, sizeof(_PATH_DEV) - 1) == 0) + strlcpy(path, arg, sizeof (path)); + else + snprintf(path, sizeof (path), "%s%s", _PATH_DEV, arg); + wholedisk = is_whole_disk(path); + if (!wholedisk && (stat64(path, &statbuf) != 0)) { + /* + * If we got ENOENT, then the user gave us + * gibberish, so try to direct them with a + * reasonable error message. Otherwise, + * regurgitate strerror() since it's the best we + * can do. + */ + if (errno == ENOENT) { + (void) fprintf(stderr, + gettext("cannot open '%s': no such " + "GEOM provider\n"), arg); + (void) fprintf(stderr, + gettext("must be a full path or " + "shorthand device name\n")); + return (NULL); + } else { + (void) fprintf(stderr, + gettext("cannot open '%s': %s\n"), + path, strerror(errno)); + return (NULL); + } + } + } - if (is_provider(path)) + /* + * Determine whether this is a device or a file. + */ + if (wholedisk) { type = VDEV_TYPE_DISK; - else { + } else if (S_ISREG(statbuf.st_mode)) { + type = VDEV_TYPE_FILE; + } else { (void) fprintf(stderr, gettext("cannot use '%s': must be a " - "GEOM provider\n"), path); + "GEOM provider or regular file\n"), path); return (NULL); } @@ -206,6 +311,7 @@ make_leaf_vdev(const char *arg) verify(nvlist_alloc(&vdev, NV_UNIQUE_NAME, 0) == 0); verify(nvlist_add_string(vdev, ZPOOL_CONFIG_PATH, path) == 0); verify(nvlist_add_string(vdev, ZPOOL_CONFIG_TYPE, type) == 0); + verify(nvlist_add_uint64(vdev, ZPOOL_CONFIG_IS_LOG, is_log) == 0); if (strcmp(type, VDEV_TYPE_DISK) == 0) verify(nvlist_add_uint64(vdev, ZPOOL_CONFIG_WHOLE_DISK, (uint64_t)B_FALSE) == 0); @@ -267,12 +373,14 @@ typedef struct replication_level { uint64_t zprl_parity; } replication_level_t; +#define ZPOOL_FUZZ (16 * 1024 * 1024) + /* * Given a list of toplevel vdevs, return the current replication level. If * the config is inconsistent, then NULL is returned. If 'fatal' is set, then * an error message will be displayed for each self-inconsistent vdev. */ -replication_level_t * +static replication_level_t * get_replication(nvlist_t *nvroot, boolean_t fatal) { nvlist_t **top; @@ -291,10 +399,20 @@ get_replication(nvlist_t *nvroot, boolean_t fatal) lastrep.zprl_type = NULL; for (t = 0; t < toplevels; t++) { + uint64_t is_log = B_FALSE; + nv = top[t]; - verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, &type) == 0); + /* + * For separate logs we ignore the top level vdev replication + * constraints. + */ + (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_IS_LOG, &is_log); + if (is_log) + continue; + verify(nvlist_lookup_string(nv, ZPOOL_CONFIG_TYPE, + &type) == 0); if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) { /* @@ -328,7 +446,7 @@ get_replication(nvlist_t *nvroot, boolean_t fatal) } /* - * The 'dontreport' variable indicatest that we've + * The 'dontreport' variable indicates that we've * already reported an error for this spec, so don't * bother doing it again. */ @@ -349,7 +467,7 @@ get_replication(nvlist_t *nvroot, boolean_t fatal) ZPOOL_CONFIG_TYPE, &childtype) == 0); /* - * If this is a a replacing or spare vdev, then + * If this is a replacing or spare vdev, then * get the real first child of the vdev. */ if (strcmp(childtype, @@ -409,22 +527,30 @@ get_replication(nvlist_t *nvroot, boolean_t fatal) */ if ((fd = open(path, O_RDONLY)) >= 0) { err = fstat64(fd, &statbuf); + if (err == 0 && + S_ISCHR(statbuf.st_mode)) { + err = ioctl(fd, DIOCGMEDIASIZE, + &statbuf.st_size); + } (void) close(fd); } else { err = stat64(path, &statbuf); } - if (err != 0 || statbuf.st_size == 0) continue; size = statbuf.st_size; /* - * Also check the size of each device. If they - * differ, then report an error. + * Also make sure that devices and + * slices have a consistent size. If + * they differ by a significant amount + * (~16MB) then report an error. */ - if (!dontreport && vdev_size != -1ULL && - size != vdev_size) { + if (!dontreport && + (vdev_size != -1ULL && + (labs(size - vdev_size) > + ZPOOL_FUZZ))) { if (ret != NULL) free(ret); ret = NULL; @@ -506,9 +632,11 @@ get_replication(nvlist_t *nvroot, boolean_t fatal) * has a consistent replication level, then we ignore any errors. Otherwise, * report any difference between the two. */ -int +static int check_replication(nvlist_t *config, nvlist_t *newroot) { + nvlist_t **child; + uint_t children; replication_level_t *current = NULL, *new; int ret; @@ -524,6 +652,23 @@ check_replication(nvlist_t *config, nvlist_t *newroot) if ((current = get_replication(nvroot, B_FALSE)) == NULL) return (0); } + /* + * for spares there may be no children, and therefore no + * replication level to check + */ + if ((nvlist_lookup_nvlist_array(newroot, ZPOOL_CONFIG_CHILDREN, + &child, &children) != 0) || (children == 0)) { + free(current); + return (0); + } + + /* + * If all we have is logs then there's no replication level to check. + */ + if (num_logs(newroot) == children) { + free(current); + return (0); + } /* * Get the replication level of the new vdev spec, reporting any @@ -621,7 +766,7 @@ is_spare(nvlist_t *config, const char *path) * Go through and find any devices that are in use. We rely on libdiskmgt for * the majority of this task. */ -int +static int check_in_use(nvlist_t *config, nvlist_t *nv, int force, int isreplacing, int isspare) { @@ -653,6 +798,9 @@ check_in_use(nvlist_t *config, nvlist_t *nv, int force, int isreplacing, if (strcmp(type, VDEV_TYPE_DISK) == 0) ret = check_provider(path, force, isspare); + if (strcmp(type, VDEV_TYPE_FILE) == 0) + ret = check_file(path, force, isspare); + return (ret); } @@ -668,10 +816,17 @@ check_in_use(nvlist_t *config, nvlist_t *nv, int force, int isreplacing, isreplacing, B_TRUE)) != 0) return (ret); + if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE, + &child, &children) == 0) + for (c = 0; c < children; c++) + if ((ret = check_in_use(config, child[c], force, + isreplacing, B_FALSE)) != 0) + return (ret); + return (0); } -const char * +static const char * is_grouping(const char *type, int *mindev) { if (strcmp(type, "raidz") == 0 || strcmp(type, "raidz1") == 0) { @@ -698,6 +853,18 @@ is_grouping(const char *type, int *mindev) return (VDEV_TYPE_SPARE); } + if (strcmp(type, "log") == 0) { + if (mindev != NULL) + *mindev = 1; + return (VDEV_TYPE_LOG); + } + + if (strcmp(type, "cache") == 0) { + if (mindev != NULL) + *mindev = 1; + return (VDEV_TYPE_L2CACHE); + } + return (NULL); } @@ -710,14 +877,21 @@ is_grouping(const char *type, int *mindev) nvlist_t * construct_spec(int argc, char **argv) { - nvlist_t *nvroot, *nv, **top, **spares; - int t, toplevels, mindev, nspares; + nvlist_t *nvroot, *nv, **top, **spares, **l2cache; + int t, toplevels, mindev, nspares, nlogs, nl2cache; const char *type; + uint64_t is_log; + boolean_t seen_logs; top = NULL; toplevels = 0; spares = NULL; + l2cache = NULL; nspares = 0; + nlogs = 0; + nl2cache = 0; + is_log = B_FALSE; + seen_logs = B_FALSE; while (argc > 0) { nv = NULL; @@ -730,12 +904,56 @@ construct_spec(int argc, char **argv) nvlist_t **child = NULL; int c, children = 0; - if (strcmp(type, VDEV_TYPE_SPARE) == 0 && - spares != NULL) { - (void) fprintf(stderr, gettext("invalid vdev " - "specification: 'spare' can be " - "specified only once\n")); - return (NULL); + if (strcmp(type, VDEV_TYPE_SPARE) == 0) { + if (spares != NULL) { + (void) fprintf(stderr, + gettext("invalid vdev " + "specification: 'spare' can be " + "specified only once\n")); + return (NULL); + } + is_log = B_FALSE; + } + + if (strcmp(type, VDEV_TYPE_LOG) == 0) { + if (seen_logs) { + (void) fprintf(stderr, + gettext("invalid vdev " + "specification: 'log' can be " + "specified only once\n")); + return (NULL); + } + seen_logs = B_TRUE; + is_log = B_TRUE; + argc--; + argv++; + /* + * A log is not a real grouping device. + * We just set is_log and continue. + */ + continue; + } + + if (strcmp(type, VDEV_TYPE_L2CACHE) == 0) { + if (l2cache != NULL) { + (void) fprintf(stderr, + gettext("invalid vdev " + "specification: 'cache' can be " + "specified only once\n")); + return (NULL); + } + is_log = B_FALSE; + } + + if (is_log) { + if (strcmp(type, VDEV_TYPE_MIRROR) != 0) { + (void) fprintf(stderr, + gettext("invalid vdev " + "specification: unsupported 'log' " + "device: %s\n"), type); + return (NULL); + } + nlogs++; } for (c = 1; c < argc; c++) { @@ -746,7 +964,8 @@ construct_spec(int argc, char **argv) children * sizeof (nvlist_t *)); if (child == NULL) zpool_no_memory(); - if ((nv = make_leaf_vdev(argv[c])) == NULL) + if ((nv = make_leaf_vdev(argv[c], B_FALSE)) + == NULL) return (NULL); child[children - 1] = nv; } @@ -765,11 +984,17 @@ construct_spec(int argc, char **argv) spares = child; nspares = children; continue; + } else if (strcmp(type, VDEV_TYPE_L2CACHE) == 0) { + l2cache = child; + nl2cache = children; + continue; } else { verify(nvlist_alloc(&nv, NV_UNIQUE_NAME, 0) == 0); verify(nvlist_add_string(nv, ZPOOL_CONFIG_TYPE, type) == 0); + verify(nvlist_add_uint64(nv, + ZPOOL_CONFIG_IS_LOG, is_log) == 0); if (strcmp(type, VDEV_TYPE_RAIDZ) == 0) { verify(nvlist_add_uint64(nv, ZPOOL_CONFIG_NPARITY, @@ -788,8 +1013,10 @@ construct_spec(int argc, char **argv) * We have a device. Pass off to make_leaf_vdev() to * construct the appropriate nvlist describing the vdev. */ - if ((nv = make_leaf_vdev(argv[0])) == NULL) + if ((nv = make_leaf_vdev(argv[0], is_log)) == NULL) return (NULL); + if (is_log) + nlogs++; argc--; argv++; } @@ -801,13 +1028,19 @@ construct_spec(int argc, char **argv) top[toplevels - 1] = nv; } - if (toplevels == 0 && nspares == 0) { + if (toplevels == 0 && nspares == 0 && nl2cache == 0) { (void) fprintf(stderr, gettext("invalid vdev " "specification: at least one toplevel vdev must be " "specified\n")); return (NULL); } + if (seen_logs && nlogs == 0) { + (void) fprintf(stderr, gettext("invalid vdev specification: " + "log requires at least 1 device\n")); + return (NULL); + } + /* * Finally, create nvroot and add all top-level vdevs to it. */ @@ -819,18 +1052,26 @@ construct_spec(int argc, char **argv) if (nspares != 0) verify(nvlist_add_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, spares, nspares) == 0); + if (nl2cache != 0) + verify(nvlist_add_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE, + l2cache, nl2cache) == 0); for (t = 0; t < toplevels; t++) nvlist_free(top[t]); for (t = 0; t < nspares; t++) nvlist_free(spares[t]); + for (t = 0; t < nl2cache; t++) + nvlist_free(l2cache[t]); if (spares) free(spares); + if (l2cache) + free(l2cache); free(top); return (nvroot); } + /* * Get and validate the contents of the given vdev specification. This ensures * that the nvlist returned is well-formed, that all the devices exist, and that @@ -842,11 +1083,11 @@ construct_spec(int argc, char **argv) * added, even if they appear in use. */ nvlist_t * -make_root_vdev(nvlist_t *poolconfig, int force, int check_rep, - boolean_t isreplacing, int argc, char **argv) +make_root_vdev(zpool_handle_t *zhp, int force, int check_rep, + boolean_t isreplacing, boolean_t dryrun, int argc, char **argv) { nvlist_t *newroot; - + nvlist_t *poolconfig = NULL; is_force = force; /* @@ -857,6 +1098,9 @@ make_root_vdev(nvlist_t *poolconfig, int force, int check_rep, if ((newroot = construct_spec(argc, argv)) == NULL) return (NULL); + if (zhp && ((poolconfig = zpool_get_config(zhp, NULL)) == NULL)) + return (NULL); + /* * Validate each device to make sure that its not shared with another * subsystem. We do this even if 'force' is set, because there are some diff --git a/cddl/contrib/opensolaris/cmd/ztest/ztest.c b/cddl/contrib/opensolaris/cmd/ztest/ztest.c index 5d9f028..b7ca302 100644 --- a/cddl/contrib/opensolaris/cmd/ztest/ztest.c +++ b/cddl/contrib/opensolaris/cmd/ztest/ztest.c @@ -19,12 +19,10 @@ * CDDL HEADER END */ /* - * Copyright 2007 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ -#pragma ident "%Z%%M% %I% %E% SMI" - /* * The objective of this program is to provide a DMU/ZAP/SPA stress test * that runs entirely in userland, is easy to use, and easy to extend. @@ -92,6 +90,7 @@ #include <sys/zio_compress.h> #include <sys/zil.h> #include <sys/vdev_impl.h> +#include <sys/vdev_file.h> #include <sys/spa_impl.h> #include <sys/dsl_prop.h> #include <sys/refcount.h> @@ -128,8 +127,18 @@ static char *zopt_dir = "/tmp"; static uint64_t zopt_time = 300; /* 5 minutes */ static int zopt_maxfaults; +typedef struct ztest_block_tag { + uint64_t bt_objset; + uint64_t bt_object; + uint64_t bt_offset; + uint64_t bt_txg; + uint64_t bt_thread; + uint64_t bt_seq; +} ztest_block_tag_t; + typedef struct ztest_args { - char *za_pool; + char za_pool[MAXNAMELEN]; + spa_t *za_spa; objset_t *za_os; zilog_t *za_zilog; thread_t za_thread; @@ -142,6 +151,13 @@ typedef struct ztest_args { hrtime_t za_stop; hrtime_t za_kill; traverse_handle_t *za_th; + /* + * Thread-local variables can go here to aid debugging. + */ + ztest_block_tag_t za_rbt; + ztest_block_tag_t za_wbt; + dmu_object_info_t za_doi; + dmu_buf_t *za_dbuf; } ztest_args_t; typedef void ztest_func_t(ztest_args_t *); @@ -160,14 +176,16 @@ ztest_func_t ztest_dmu_objset_create_destroy; ztest_func_t ztest_dmu_snapshot_create_destroy; ztest_func_t ztest_spa_create_destroy; ztest_func_t ztest_fault_inject; +ztest_func_t ztest_spa_rename; ztest_func_t ztest_vdev_attach_detach; ztest_func_t ztest_vdev_LUN_growth; ztest_func_t ztest_vdev_add_remove; +ztest_func_t ztest_vdev_aux_add_remove; ztest_func_t ztest_scrub; -ztest_func_t ztest_spa_rename; typedef struct ztest_info { ztest_func_t *zi_func; /* test function */ + uint64_t zi_iters; /* iterations per execution */ uint64_t *zi_interval; /* execute every <interval> seconds */ uint64_t zi_calls; /* per-pass count */ uint64_t zi_call_time; /* per-pass time */ @@ -181,22 +199,23 @@ uint64_t zopt_sometimes = 10; /* every 10 seconds */ uint64_t zopt_rarely = 60; /* every 60 seconds */ ztest_info_t ztest_info[] = { - { ztest_dmu_read_write, &zopt_always }, - { ztest_dmu_write_parallel, &zopt_always }, - { ztest_dmu_object_alloc_free, &zopt_always }, - { ztest_zap, &zopt_always }, - { ztest_zap_parallel, &zopt_always }, - { ztest_traverse, &zopt_often }, - { ztest_dsl_prop_get_set, &zopt_sometimes }, - { ztest_dmu_objset_create_destroy, &zopt_sometimes }, - { ztest_dmu_snapshot_create_destroy, &zopt_rarely }, - { ztest_spa_create_destroy, &zopt_sometimes }, - { ztest_fault_inject, &zopt_sometimes }, - { ztest_spa_rename, &zopt_rarely }, - { ztest_vdev_attach_detach, &zopt_rarely }, - { ztest_vdev_LUN_growth, &zopt_rarely }, - { ztest_vdev_add_remove, &zopt_vdevtime }, - { ztest_scrub, &zopt_vdevtime }, + { ztest_dmu_read_write, 1, &zopt_always }, + { ztest_dmu_write_parallel, 30, &zopt_always }, + { ztest_dmu_object_alloc_free, 1, &zopt_always }, + { ztest_zap, 30, &zopt_always }, + { ztest_zap_parallel, 100, &zopt_always }, + { ztest_traverse, 1, &zopt_often }, + { ztest_dsl_prop_get_set, 1, &zopt_sometimes }, + { ztest_dmu_objset_create_destroy, 1, &zopt_sometimes }, + { ztest_dmu_snapshot_create_destroy, 1, &zopt_sometimes }, + { ztest_spa_create_destroy, 1, &zopt_sometimes }, + { ztest_fault_inject, 1, &zopt_sometimes }, + { ztest_spa_rename, 1, &zopt_rarely }, + { ztest_vdev_attach_detach, 1, &zopt_rarely }, + { ztest_vdev_LUN_growth, 1, &zopt_rarely }, + { ztest_vdev_add_remove, 1, &zopt_vdevtime }, + { ztest_vdev_aux_add_remove, 1, &zopt_vdevtime }, + { ztest_scrub, 1, &zopt_vdevtime }, }; #define ZTEST_FUNCS (sizeof (ztest_info) / sizeof (ztest_info_t)) @@ -210,34 +229,27 @@ typedef struct ztest_shared { mutex_t zs_vdev_lock; rwlock_t zs_name_lock; uint64_t zs_vdev_primaries; + uint64_t zs_vdev_aux; uint64_t zs_enospc_count; hrtime_t zs_start_time; hrtime_t zs_stop_time; uint64_t zs_alloc; uint64_t zs_space; - uint64_t zs_txg; ztest_info_t zs_info[ZTEST_FUNCS]; mutex_t zs_sync_lock[ZTEST_SYNC_LOCKS]; uint64_t zs_seq[ZTEST_SYNC_LOCKS]; } ztest_shared_t; -typedef struct ztest_block_tag { - uint64_t bt_objset; - uint64_t bt_object; - uint64_t bt_offset; - uint64_t bt_txg; - uint64_t bt_thread; - uint64_t bt_seq; -} ztest_block_tag_t; - static char ztest_dev_template[] = "%s/%s.%llua"; +static char ztest_aux_template[] = "%s/%s.%s.%llu"; static ztest_shared_t *ztest_shared; static int ztest_random_fd; static int ztest_dump_core = 1; -extern uint64_t zio_gang_bang; -extern uint16_t zio_zil_fail_shift; +static boolean_t ztest_exiting; + +extern uint64_t metaslab_gang_bang; #define ZTEST_DIROBJ 1 #define ZTEST_MICROZAP_OBJ 2 @@ -357,7 +369,7 @@ usage(boolean_t requested) FILE *fp = requested ? stdout : stderr; nicenum(zopt_vdev_size, nice_vdev_size); - nicenum(zio_gang_bang, nice_gang_bang); + nicenum(metaslab_gang_bang, nice_gang_bang); (void) fprintf(fp, "Usage: %s\n" "\t[-v vdevs (default: %llu)]\n" @@ -377,26 +389,24 @@ usage(boolean_t requested) "\t[-E(xisting)] (use existing pool instead of creating new one)\n" "\t[-T time] total run time (default: %llu sec)\n" "\t[-P passtime] time per pass (default: %llu sec)\n" - "\t[-z zil failure rate (default: fail every 2^%llu allocs)]\n" "\t[-h] (print help)\n" "", cmdname, - (u_longlong_t)zopt_vdevs, /* -v */ - nice_vdev_size, /* -s */ - zopt_ashift, /* -a */ - zopt_mirrors, /* -m */ - zopt_raidz, /* -r */ - zopt_raidz_parity, /* -R */ - zopt_datasets, /* -d */ - zopt_threads, /* -t */ - nice_gang_bang, /* -g */ - zopt_init, /* -i */ - (u_longlong_t)zopt_killrate, /* -k */ - zopt_pool, /* -p */ - zopt_dir, /* -f */ - (u_longlong_t)zopt_time, /* -T */ - (u_longlong_t)zopt_passtime, /* -P */ - (u_longlong_t)zio_zil_fail_shift); /* -z */ + (u_longlong_t)zopt_vdevs, /* -v */ + nice_vdev_size, /* -s */ + zopt_ashift, /* -a */ + zopt_mirrors, /* -m */ + zopt_raidz, /* -r */ + zopt_raidz_parity, /* -R */ + zopt_datasets, /* -d */ + zopt_threads, /* -t */ + nice_gang_bang, /* -g */ + zopt_init, /* -i */ + (u_longlong_t)zopt_killrate, /* -k */ + zopt_pool, /* -p */ + zopt_dir, /* -f */ + (u_longlong_t)zopt_time, /* -T */ + (u_longlong_t)zopt_passtime); /* -P */ exit(requested ? 0 : 1); } @@ -431,91 +441,84 @@ process_options(int argc, char **argv) progname = argv[0]; /* By default, test gang blocks for blocks 32K and greater */ - zio_gang_bang = 32 << 10; - - /* Default value, fail every 32nd allocation */ - zio_zil_fail_shift = 5; + metaslab_gang_bang = 32 << 10; while ((opt = getopt(argc, argv, - "v:s:a:m:r:R:d:t:g:i:k:p:f:VET:P:z:h")) != EOF) { + "v:s:a:m:r:R:d:t:g:i:k:p:f:VET:P:h")) != EOF) { value = 0; switch (opt) { - case 'v': - case 's': - case 'a': - case 'm': - case 'r': - case 'R': - case 'd': - case 't': - case 'g': - case 'i': - case 'k': - case 'T': - case 'P': - case 'z': + case 'v': + case 's': + case 'a': + case 'm': + case 'r': + case 'R': + case 'd': + case 't': + case 'g': + case 'i': + case 'k': + case 'T': + case 'P': value = nicenumtoull(optarg); } switch (opt) { - case 'v': + case 'v': zopt_vdevs = value; break; - case 's': + case 's': zopt_vdev_size = MAX(SPA_MINDEVSIZE, value); break; - case 'a': + case 'a': zopt_ashift = value; break; - case 'm': + case 'm': zopt_mirrors = value; break; - case 'r': + case 'r': zopt_raidz = MAX(1, value); break; - case 'R': + case 'R': zopt_raidz_parity = MIN(MAX(value, 1), 2); break; - case 'd': + case 'd': zopt_datasets = MAX(1, value); break; - case 't': + case 't': zopt_threads = MAX(1, value); break; - case 'g': - zio_gang_bang = MAX(SPA_MINBLOCKSIZE << 1, value); + case 'g': + metaslab_gang_bang = MAX(SPA_MINBLOCKSIZE << 1, value); break; - case 'i': + case 'i': zopt_init = value; break; - case 'k': + case 'k': zopt_killrate = value; break; - case 'p': + case 'p': zopt_pool = strdup(optarg); break; - case 'f': + case 'f': zopt_dir = strdup(optarg); break; - case 'V': + case 'V': zopt_verbose++; break; - case 'E': + case 'E': zopt_init = 0; break; - case 'T': + case 'T': zopt_time = value; break; - case 'P': + case 'P': zopt_passtime = MAX(1, value); break; - case 'z': - zio_zil_fail_shift = MIN(value, 16); - break; - case 'h': + case 'h': usage(B_TRUE); break; - case '?': - default: + case '?': + default: usage(B_FALSE); break; } @@ -536,51 +539,58 @@ ztest_get_ashift(void) } static nvlist_t * -make_vdev_file(size_t size) +make_vdev_file(char *path, char *aux, size_t size, uint64_t ashift) { - char dev_name[MAXPATHLEN]; + char pathbuf[MAXPATHLEN]; uint64_t vdev; - uint64_t ashift = ztest_get_ashift(); - int fd; nvlist_t *file; - if (size == 0) { - (void) snprintf(dev_name, sizeof (dev_name), "%s", - "/dev/bogus"); - } else { - vdev = ztest_shared->zs_vdev_primaries++; - (void) sprintf(dev_name, ztest_dev_template, - zopt_dir, zopt_pool, vdev); + if (ashift == 0) + ashift = ztest_get_ashift(); + + if (path == NULL) { + path = pathbuf; + + if (aux != NULL) { + vdev = ztest_shared->zs_vdev_aux; + (void) sprintf(path, ztest_aux_template, + zopt_dir, zopt_pool, aux, vdev); + } else { + vdev = ztest_shared->zs_vdev_primaries++; + (void) sprintf(path, ztest_dev_template, + zopt_dir, zopt_pool, vdev); + } + } - fd = open(dev_name, O_RDWR | O_CREAT | O_TRUNC, 0666); + if (size != 0) { + int fd = open(path, O_RDWR | O_CREAT | O_TRUNC, 0666); if (fd == -1) - fatal(1, "can't open %s", dev_name); + fatal(1, "can't open %s", path); if (ftruncate(fd, size) != 0) - fatal(1, "can't ftruncate %s", dev_name); + fatal(1, "can't ftruncate %s", path); (void) close(fd); } VERIFY(nvlist_alloc(&file, NV_UNIQUE_NAME, 0) == 0); VERIFY(nvlist_add_string(file, ZPOOL_CONFIG_TYPE, VDEV_TYPE_FILE) == 0); - VERIFY(nvlist_add_string(file, ZPOOL_CONFIG_PATH, dev_name) == 0); + VERIFY(nvlist_add_string(file, ZPOOL_CONFIG_PATH, path) == 0); VERIFY(nvlist_add_uint64(file, ZPOOL_CONFIG_ASHIFT, ashift) == 0); return (file); } static nvlist_t * -make_vdev_raidz(size_t size, int r) +make_vdev_raidz(char *path, char *aux, size_t size, uint64_t ashift, int r) { nvlist_t *raidz, **child; int c; if (r < 2) - return (make_vdev_file(size)); - + return (make_vdev_file(path, aux, size, ashift)); child = umem_alloc(r * sizeof (nvlist_t *), UMEM_NOFAIL); for (c = 0; c < r; c++) - child[c] = make_vdev_file(size); + child[c] = make_vdev_file(path, aux, size, ashift); VERIFY(nvlist_alloc(&raidz, NV_UNIQUE_NAME, 0) == 0); VERIFY(nvlist_add_string(raidz, ZPOOL_CONFIG_TYPE, @@ -599,18 +609,19 @@ make_vdev_raidz(size_t size, int r) } static nvlist_t * -make_vdev_mirror(size_t size, int r, int m) +make_vdev_mirror(char *path, char *aux, size_t size, uint64_t ashift, + int r, int m) { nvlist_t *mirror, **child; int c; if (m < 1) - return (make_vdev_raidz(size, r)); + return (make_vdev_raidz(path, aux, size, ashift, r)); child = umem_alloc(m * sizeof (nvlist_t *), UMEM_NOFAIL); for (c = 0; c < m; c++) - child[c] = make_vdev_raidz(size, r); + child[c] = make_vdev_raidz(path, aux, size, ashift, r); VERIFY(nvlist_alloc(&mirror, NV_UNIQUE_NAME, 0) == 0); VERIFY(nvlist_add_string(mirror, ZPOOL_CONFIG_TYPE, @@ -627,7 +638,8 @@ make_vdev_mirror(size_t size, int r, int m) } static nvlist_t * -make_vdev_root(size_t size, int r, int m, int t) +make_vdev_root(char *path, char *aux, size_t size, uint64_t ashift, + int log, int r, int m, int t) { nvlist_t *root, **child; int c; @@ -636,12 +648,15 @@ make_vdev_root(size_t size, int r, int m, int t) child = umem_alloc(t * sizeof (nvlist_t *), UMEM_NOFAIL); - for (c = 0; c < t; c++) - child[c] = make_vdev_mirror(size, r, m); + for (c = 0; c < t; c++) { + child[c] = make_vdev_mirror(path, aux, size, ashift, r, m); + VERIFY(nvlist_add_uint64(child[c], ZPOOL_CONFIG_IS_LOG, + log) == 0); + } VERIFY(nvlist_alloc(&root, NV_UNIQUE_NAME, 0) == 0); VERIFY(nvlist_add_string(root, ZPOOL_CONFIG_TYPE, VDEV_TYPE_ROOT) == 0); - VERIFY(nvlist_add_nvlist_array(root, ZPOOL_CONFIG_CHILDREN, + VERIFY(nvlist_add_nvlist_array(root, aux ? aux : ZPOOL_CONFIG_CHILDREN, child, t) == 0); for (c = 0; c < t; c++) @@ -785,8 +800,8 @@ ztest_spa_create_destroy(ztest_args_t *za) /* * Attempt to create using a bad file. */ - nvroot = make_vdev_root(0, 0, 0, 1); - error = spa_create("ztest_bad_file", nvroot, NULL); + nvroot = make_vdev_root("/dev/bogus", NULL, 0, 0, 0, 0, 0, 1); + error = spa_create("ztest_bad_file", nvroot, NULL, NULL, NULL); nvlist_free(nvroot); if (error != ENOENT) fatal(0, "spa_create(bad_file) = %d", error); @@ -794,8 +809,8 @@ ztest_spa_create_destroy(ztest_args_t *za) /* * Attempt to create using a bad mirror. */ - nvroot = make_vdev_root(0, 0, 2, 1); - error = spa_create("ztest_bad_mirror", nvroot, NULL); + nvroot = make_vdev_root("/dev/bogus", NULL, 0, 0, 0, 0, 2, 1); + error = spa_create("ztest_bad_mirror", nvroot, NULL, NULL, NULL); nvlist_free(nvroot); if (error != ENOENT) fatal(0, "spa_create(bad_mirror) = %d", error); @@ -805,8 +820,8 @@ ztest_spa_create_destroy(ztest_args_t *za) * what's in the nvroot; we should fail with EEXIST. */ (void) rw_rdlock(&ztest_shared->zs_name_lock); - nvroot = make_vdev_root(0, 0, 0, 1); - error = spa_create(za->za_pool, nvroot, NULL); + nvroot = make_vdev_root("/dev/bogus", NULL, 0, 0, 0, 0, 0, 1); + error = spa_create(za->za_pool, nvroot, NULL, NULL, NULL); nvlist_free(nvroot); if (error != EEXIST) fatal(0, "spa_create(whatever) = %d", error); @@ -823,30 +838,48 @@ ztest_spa_create_destroy(ztest_args_t *za) (void) rw_unlock(&ztest_shared->zs_name_lock); } +static vdev_t * +vdev_lookup_by_path(vdev_t *vd, const char *path) +{ + vdev_t *mvd; + + if (vd->vdev_path != NULL && strcmp(path, vd->vdev_path) == 0) + return (vd); + + for (int c = 0; c < vd->vdev_children; c++) + if ((mvd = vdev_lookup_by_path(vd->vdev_child[c], path)) != + NULL) + return (mvd); + + return (NULL); +} + /* * Verify that vdev_add() works as expected. */ void ztest_vdev_add_remove(ztest_args_t *za) { - spa_t *spa = dmu_objset_spa(za->za_os); + spa_t *spa = za->za_spa; uint64_t leaves = MAX(zopt_mirrors, 1) * zopt_raidz; nvlist_t *nvroot; int error; - if (zopt_verbose >= 6) - (void) printf("adding vdev\n"); - (void) mutex_lock(&ztest_shared->zs_vdev_lock); - spa_config_enter(spa, RW_READER, FTAG); + spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER); ztest_shared->zs_vdev_primaries = spa->spa_root_vdev->vdev_children * leaves; - spa_config_exit(spa, FTAG); + spa_config_exit(spa, SCL_VDEV, FTAG); + + /* + * Make 1/4 of the devices be log devices. + */ + nvroot = make_vdev_root(NULL, NULL, zopt_vdev_size, 0, + ztest_random(4) == 0, zopt_raidz, zopt_mirrors, 1); - nvroot = make_vdev_root(zopt_vdev_size, zopt_raidz, zopt_mirrors, 1); error = spa_vdev_add(spa, nvroot); nvlist_free(nvroot); @@ -856,37 +889,86 @@ ztest_vdev_add_remove(ztest_args_t *za) ztest_record_enospc("spa_vdev_add"); else if (error != 0) fatal(0, "spa_vdev_add() = %d", error); - - if (zopt_verbose >= 6) - (void) printf("spa_vdev_add = %d, as expected\n", error); } -static vdev_t * -vdev_lookup_by_path(vdev_t *vd, const char *path) +/* + * Verify that adding/removing aux devices (l2arc, hot spare) works as expected. + */ +void +ztest_vdev_aux_add_remove(ztest_args_t *za) { - int c; - vdev_t *mvd; + spa_t *spa = za->za_spa; + vdev_t *rvd = spa->spa_root_vdev; + spa_aux_vdev_t *sav; + char *aux; + uint64_t guid = 0; + int error; - if (vd->vdev_path != NULL) { - if (vd->vdev_wholedisk == 1) { - /* - * For whole disks, the internal path has 's0', but the - * path passed in by the user doesn't. - */ - if (strlen(path) == strlen(vd->vdev_path) - 2 && - strncmp(path, vd->vdev_path, strlen(path)) == 0) - return (vd); - } else if (strcmp(path, vd->vdev_path) == 0) { - return (vd); + if (ztest_random(2) == 0) { + sav = &spa->spa_spares; + aux = ZPOOL_CONFIG_SPARES; + } else { + sav = &spa->spa_l2cache; + aux = ZPOOL_CONFIG_L2CACHE; + } + + (void) mutex_lock(&ztest_shared->zs_vdev_lock); + + spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER); + + if (sav->sav_count != 0 && ztest_random(4) == 0) { + /* + * Pick a random device to remove. + */ + guid = sav->sav_vdevs[ztest_random(sav->sav_count)]->vdev_guid; + } else { + /* + * Find an unused device we can add. + */ + ztest_shared->zs_vdev_aux = 0; + for (;;) { + char path[MAXPATHLEN]; + int c; + (void) sprintf(path, ztest_aux_template, zopt_dir, + zopt_pool, aux, ztest_shared->zs_vdev_aux); + for (c = 0; c < sav->sav_count; c++) + if (strcmp(sav->sav_vdevs[c]->vdev_path, + path) == 0) + break; + if (c == sav->sav_count && + vdev_lookup_by_path(rvd, path) == NULL) + break; + ztest_shared->zs_vdev_aux++; } } - for (c = 0; c < vd->vdev_children; c++) - if ((mvd = vdev_lookup_by_path(vd->vdev_child[c], path)) != - NULL) - return (mvd); + spa_config_exit(spa, SCL_VDEV, FTAG); - return (NULL); + if (guid == 0) { + /* + * Add a new device. + */ + nvlist_t *nvroot = make_vdev_root(NULL, aux, + (zopt_vdev_size * 5) / 4, 0, 0, 0, 0, 1); + error = spa_vdev_add(spa, nvroot); + if (error != 0) + fatal(0, "spa_vdev_add(%p) = %d", nvroot, error); + nvlist_free(nvroot); + } else { + /* + * Remove an existing device. Sometimes, dirty its + * vdev state first to make sure we handle removal + * of devices that have pending state changes. + */ + if (ztest_random(2) == 0) + (void) vdev_online(spa, guid, B_FALSE, NULL); + + error = spa_vdev_remove(spa, guid, B_FALSE); + if (error != 0 && error != EBUSY) + fatal(0, "spa_vdev_remove(%llu) = %d", guid, error); + } + + (void) mutex_unlock(&ztest_shared->zs_vdev_lock); } /* @@ -895,22 +977,26 @@ vdev_lookup_by_path(vdev_t *vd, const char *path) void ztest_vdev_attach_detach(ztest_args_t *za) { - spa_t *spa = dmu_objset_spa(za->za_os); + spa_t *spa = za->za_spa; + spa_aux_vdev_t *sav = &spa->spa_spares; vdev_t *rvd = spa->spa_root_vdev; vdev_t *oldvd, *newvd, *pvd; - nvlist_t *root, *file; + nvlist_t *root; uint64_t leaves = MAX(zopt_mirrors, 1) * zopt_raidz; uint64_t leaf, top; uint64_t ashift = ztest_get_ashift(); + uint64_t oldguid; size_t oldsize, newsize; char oldpath[MAXPATHLEN], newpath[MAXPATHLEN]; int replacing; + int oldvd_has_siblings = B_FALSE; + int newvd_is_spare = B_FALSE; + int oldvd_is_log; int error, expected_error; - int fd; (void) mutex_lock(&ztest_shared->zs_vdev_lock); - spa_config_enter(spa, RW_READER, FTAG); + spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER); /* * Decide whether to do an attach or a replace. @@ -928,39 +1014,70 @@ ztest_vdev_attach_detach(ztest_args_t *za) leaf = ztest_random(leaves); /* - * Generate the path to this leaf. The filename will end with 'a'. - * We'll alternate replacements with a filename that ends with 'b'. + * Locate this vdev. */ - (void) snprintf(oldpath, sizeof (oldpath), - ztest_dev_template, zopt_dir, zopt_pool, top * leaves + leaf); - - bcopy(oldpath, newpath, MAXPATHLEN); + oldvd = rvd->vdev_child[top]; + if (zopt_mirrors >= 1) + oldvd = oldvd->vdev_child[leaf / zopt_raidz]; + if (zopt_raidz > 1) + oldvd = oldvd->vdev_child[leaf % zopt_raidz]; /* - * If the 'a' file isn't part of the pool, the 'b' file must be. + * If we're already doing an attach or replace, oldvd may be a + * mirror vdev -- in which case, pick a random child. */ - if (vdev_lookup_by_path(rvd, oldpath) == NULL) - oldpath[strlen(oldpath) - 1] = 'b'; - else - newpath[strlen(newpath) - 1] = 'b'; + while (oldvd->vdev_children != 0) { + oldvd_has_siblings = B_TRUE; + ASSERT(oldvd->vdev_children == 2); + oldvd = oldvd->vdev_child[ztest_random(2)]; + } + + oldguid = oldvd->vdev_guid; + oldsize = vdev_get_rsize(oldvd); + oldvd_is_log = oldvd->vdev_top->vdev_islog; + (void) strcpy(oldpath, oldvd->vdev_path); + pvd = oldvd->vdev_parent; /* - * Now oldpath represents something that's already in the pool, - * and newpath is the thing we'll try to attach. + * If oldvd has siblings, then half of the time, detach it. */ - oldvd = vdev_lookup_by_path(rvd, oldpath); - newvd = vdev_lookup_by_path(rvd, newpath); - ASSERT(oldvd != NULL); - pvd = oldvd->vdev_parent; + if (oldvd_has_siblings && ztest_random(2) == 0) { + spa_config_exit(spa, SCL_VDEV, FTAG); + error = spa_vdev_detach(spa, oldguid, B_FALSE); + if (error != 0 && error != ENODEV && error != EBUSY) + fatal(0, "detach (%s) returned %d", + oldpath, error); + (void) mutex_unlock(&ztest_shared->zs_vdev_lock); + return; + } /* - * Make newsize a little bigger or smaller than oldsize. - * If it's smaller, the attach should fail. - * If it's larger, and we're doing a replace, - * we should get dynamic LUN growth when we're done. + * For the new vdev, choose with equal probability between the two + * standard paths (ending in either 'a' or 'b') or a random hot spare. */ - oldsize = vdev_get_rsize(oldvd); - newsize = 10 * oldsize / (9 + ztest_random(3)); + if (sav->sav_count != 0 && ztest_random(3) == 0) { + newvd = sav->sav_vdevs[ztest_random(sav->sav_count)]; + newvd_is_spare = B_TRUE; + (void) strcpy(newpath, newvd->vdev_path); + } else { + (void) snprintf(newpath, sizeof (newpath), ztest_dev_template, + zopt_dir, zopt_pool, top * leaves + leaf); + if (ztest_random(2) == 0) + newpath[strlen(newpath) - 1] = 'b'; + newvd = vdev_lookup_by_path(rvd, newpath); + } + + if (newvd) { + newsize = vdev_get_rsize(newvd); + } else { + /* + * Make newsize a little bigger or smaller than oldsize. + * If it's smaller, the attach should fail. + * If it's larger, and we're doing a replace, + * we should get dynamic LUN growth when we're done. + */ + newsize = 10 * oldsize / (9 + ztest_random(3)); + } /* * If pvd is not a mirror or root, the attach should fail with ENOTSUP, @@ -970,12 +1087,17 @@ ztest_vdev_attach_detach(ztest_args_t *za) * * If newvd is too small, it should fail with EOVERFLOW. */ - if (newvd != NULL) - expected_error = EBUSY; - else if (pvd->vdev_ops != &vdev_mirror_ops && - pvd->vdev_ops != &vdev_root_ops && - (!replacing || pvd->vdev_ops == &vdev_replacing_ops)) + if (pvd->vdev_ops != &vdev_mirror_ops && + pvd->vdev_ops != &vdev_root_ops && (!replacing || + pvd->vdev_ops == &vdev_replacing_ops || + pvd->vdev_ops == &vdev_spare_ops)) + expected_error = ENOTSUP; + else if (newvd_is_spare && (!replacing || oldvd_is_log)) expected_error = ENOTSUP; + else if (newvd == oldvd) + expected_error = replacing ? 0 : EBUSY; + else if (vdev_lookup_by_path(rvd, newpath) != NULL) + expected_error = EBUSY; else if (newsize < oldsize) expected_error = EOVERFLOW; else if (ashift > oldvd->vdev_top->vdev_ashift) @@ -983,36 +1105,16 @@ ztest_vdev_attach_detach(ztest_args_t *za) else expected_error = 0; - /* - * If newvd isn't already part of the pool, create it. - */ - if (newvd == NULL) { - fd = open(newpath, O_RDWR | O_CREAT | O_TRUNC, 0666); - if (fd == -1) - fatal(1, "can't open %s", newpath); - if (ftruncate(fd, newsize) != 0) - fatal(1, "can't ftruncate %s", newpath); - (void) close(fd); - } - - spa_config_exit(spa, FTAG); + spa_config_exit(spa, SCL_VDEV, FTAG); /* * Build the nvlist describing newpath. */ - VERIFY(nvlist_alloc(&file, NV_UNIQUE_NAME, 0) == 0); - VERIFY(nvlist_add_string(file, ZPOOL_CONFIG_TYPE, VDEV_TYPE_FILE) == 0); - VERIFY(nvlist_add_string(file, ZPOOL_CONFIG_PATH, newpath) == 0); - VERIFY(nvlist_add_uint64(file, ZPOOL_CONFIG_ASHIFT, ashift) == 0); + root = make_vdev_root(newpath, NULL, newvd == NULL ? newsize : 0, + ashift, 0, 0, 0, 1); - VERIFY(nvlist_alloc(&root, NV_UNIQUE_NAME, 0) == 0); - VERIFY(nvlist_add_string(root, ZPOOL_CONFIG_TYPE, VDEV_TYPE_ROOT) == 0); - VERIFY(nvlist_add_nvlist_array(root, ZPOOL_CONFIG_CHILDREN, - &file, 1) == 0); - - error = spa_vdev_attach(spa, oldvd->vdev_guid, root, replacing); + error = spa_vdev_attach(spa, oldguid, root, replacing); - nvlist_free(file); nvlist_free(root); /* @@ -1027,12 +1129,15 @@ ztest_vdev_attach_detach(ztest_args_t *za) /* * If someone grew the LUN, the replacement may be too small. */ - if (error == EOVERFLOW) + if (error == EOVERFLOW || error == EBUSY) expected_error = error; - if (error != expected_error) { - fatal(0, "attach (%s, %s, %d) returned %d, expected %d", - oldpath, newpath, replacing, error, expected_error); + /* XXX workaround 6690467 */ + if (error != expected_error && expected_error != EBUSY) { + fatal(0, "attach (%s %llu, %s %llu, %d) " + "returned %d, expected %d", + oldpath, (longlong_t)oldsize, newpath, + (longlong_t)newsize, replacing, error, expected_error); } (void) mutex_unlock(&ztest_shared->zs_vdev_lock); @@ -1045,7 +1150,7 @@ ztest_vdev_attach_detach(ztest_args_t *za) void ztest_vdev_LUN_growth(ztest_args_t *za) { - spa_t *spa = dmu_objset_spa(za->za_os); + spa_t *spa = za->za_spa; char dev_name[MAXPATHLEN]; uint64_t leaves = MAX(zopt_mirrors, 1) * zopt_raidz; uint64_t vdev; @@ -1057,9 +1162,9 @@ ztest_vdev_LUN_growth(ztest_args_t *za) /* * Pick a random leaf vdev. */ - spa_config_enter(spa, RW_READER, FTAG); + spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER); vdev = ztest_random(spa->spa_root_vdev->vdev_children * leaves); - spa_config_exit(spa, FTAG); + spa_config_exit(spa, SCL_VDEV, FTAG); (void) sprintf(dev_name, ztest_dev_template, zopt_dir, zopt_pool, vdev); @@ -1088,14 +1193,14 @@ ztest_vdev_LUN_growth(ztest_args_t *za) /* ARGSUSED */ static void -ztest_create_cb(objset_t *os, void *arg, dmu_tx_t *tx) +ztest_create_cb(objset_t *os, void *arg, cred_t *cr, dmu_tx_t *tx) { /* * Create the directory object. */ VERIFY(dmu_object_claim(os, ZTEST_DIROBJ, DMU_OT_UINT64_OTHER, ZTEST_DIROBJ_BLOCKSIZE, - DMU_OT_UINT64_OTHER, sizeof (ztest_block_tag_t), tx) == 0); + DMU_OT_UINT64_OTHER, 5 * sizeof (ztest_block_tag_t), tx) == 0); VERIFY(zap_create_claim(os, ZTEST_MICROZAP_OBJ, DMU_OT_ZAP_OTHER, DMU_OT_NONE, 0, tx) == 0); @@ -1104,26 +1209,26 @@ ztest_create_cb(objset_t *os, void *arg, dmu_tx_t *tx) DMU_OT_ZAP_OTHER, DMU_OT_NONE, 0, tx) == 0); } -/* ARGSUSED */ static int ztest_destroy_cb(char *name, void *arg) { + ztest_args_t *za = arg; objset_t *os; - dmu_object_info_t doi; + dmu_object_info_t *doi = &za->za_doi; int error; /* * Verify that the dataset contains a directory object. */ error = dmu_objset_open(name, DMU_OST_OTHER, - DS_MODE_STANDARD | DS_MODE_READONLY, &os); + DS_MODE_USER | DS_MODE_READONLY, &os); ASSERT3U(error, ==, 0); - error = dmu_object_info(os, ZTEST_DIROBJ, &doi); + error = dmu_object_info(os, ZTEST_DIROBJ, doi); if (error != ENOENT) { /* We could have crashed in the middle of destroying it */ ASSERT3U(error, ==, 0); - ASSERT3U(doi.doi_type, ==, DMU_OT_UINT64_OTHER); - ASSERT3S(doi.doi_physical_blks, >=, 0); + ASSERT3U(doi->doi_type, ==, DMU_OT_UINT64_OTHER); + ASSERT3S(doi->doi_physical_blks, >=, 0); } dmu_objset_close(os); @@ -1131,7 +1236,11 @@ ztest_destroy_cb(char *name, void *arg) * Destroy the dataset. */ error = dmu_objset_destroy(name); - ASSERT3U(error, ==, 0); + if (error) { + (void) dmu_objset_open(name, DMU_OST_OTHER, + DS_MODE_USER | DS_MODE_READONLY, &os); + fatal(0, "dmu_objset_destroy(os=%p) = %d\n", &os, error); + } return (0); } @@ -1171,9 +1280,9 @@ void ztest_dmu_objset_create_destroy(ztest_args_t *za) { int error; - objset_t *os; + objset_t *os, *os2; char name[100]; - int mode, basemode, expected_error; + int basemode, expected_error; zilog_t *zilog; uint64_t seq; uint64_t objects; @@ -1183,9 +1292,9 @@ ztest_dmu_objset_create_destroy(ztest_args_t *za) (void) snprintf(name, 100, "%s/%s_temp_%llu", za->za_pool, za->za_pool, (u_longlong_t)za->za_instance); - basemode = DS_MODE_LEVEL(za->za_instance); - if (basemode == DS_MODE_NONE) - basemode++; + basemode = DS_MODE_TYPE(za->za_instance); + if (basemode != DS_MODE_USER && basemode != DS_MODE_OWNER) + basemode = DS_MODE_USER; /* * If this dataset exists from a previous run, process its replay log @@ -1193,9 +1302,9 @@ ztest_dmu_objset_create_destroy(ztest_args_t *za) * (invoked from ztest_destroy_cb() below) should just throw it away. */ if (ztest_random(2) == 0 && - dmu_objset_open(name, DMU_OST_OTHER, DS_MODE_PRIMARY, &os) == 0) { + dmu_objset_open(name, DMU_OST_OTHER, DS_MODE_OWNER, &os) == 0) { zr.zr_os = os; - zil_replay(os, &zr, &zr.zr_assign, ztest_replay_vector); + zil_replay(os, &zr, &zr.zr_assign, ztest_replay_vector, NULL); dmu_objset_close(os); } @@ -1204,7 +1313,7 @@ ztest_dmu_objset_create_destroy(ztest_args_t *za) * create lying around from a previous run. If so, destroy it * and all of its snapshots. */ - (void) dmu_objset_find(name, ztest_destroy_cb, NULL, + (void) dmu_objset_find(name, ztest_destroy_cb, za, DS_FIND_CHILDREN | DS_FIND_SNAPSHOTS); /* @@ -1218,8 +1327,8 @@ ztest_dmu_objset_create_destroy(ztest_args_t *za) /* * Verify that we can create a new dataset. */ - error = dmu_objset_create(name, DMU_OST_OTHER, NULL, ztest_create_cb, - NULL); + error = dmu_objset_create(name, DMU_OST_OTHER, NULL, 0, + ztest_create_cb, NULL); if (error) { if (error == ENOSPC) { ztest_record_enospc("dmu_objset_create"); @@ -1274,26 +1383,29 @@ ztest_dmu_objset_create_destroy(ztest_args_t *za) /* * Verify that we cannot create an existing dataset. */ - error = dmu_objset_create(name, DMU_OST_OTHER, NULL, NULL, NULL); + error = dmu_objset_create(name, DMU_OST_OTHER, NULL, 0, NULL, NULL); if (error != EEXIST) fatal(0, "created existing dataset, error = %d", error); /* - * Verify that multiple dataset opens are allowed, but only when + * Verify that multiple dataset holds are allowed, but only when * the new access mode is compatible with the base mode. - * We use a mixture of typed and typeless opens, and when the - * open succeeds, verify that the discovered type is correct. - */ - for (mode = DS_MODE_STANDARD; mode < DS_MODE_LEVELS; mode++) { - objset_t *os2; - error = dmu_objset_open(name, DMU_OST_OTHER, mode, &os2); - expected_error = (basemode + mode < DS_MODE_LEVELS) ? 0 : EBUSY; - if (error != expected_error) - fatal(0, "dmu_objset_open('%s') = %d, expected %d", - name, error, expected_error); - if (error == 0) + */ + if (basemode == DS_MODE_OWNER) { + error = dmu_objset_open(name, DMU_OST_OTHER, DS_MODE_USER, + &os2); + if (error) + fatal(0, "dmu_objset_open('%s') = %d", name, error); + else dmu_objset_close(os2); } + error = dmu_objset_open(name, DMU_OST_OTHER, DS_MODE_OWNER, &os2); + expected_error = (basemode == DS_MODE_OWNER) ? EBUSY : 0; + if (error != expected_error) + fatal(0, "dmu_objset_open('%s') = %d, expected %d", + name, error, expected_error); + if (error == 0) + dmu_objset_close(os2); zil_close(zilog); dmu_objset_close(os); @@ -1417,7 +1529,7 @@ ztest_blk_cb(traverse_blk_cache_t *bc, spa_t *spa, void *arg) void ztest_traverse(ztest_args_t *za) { - spa_t *spa = dmu_objset_spa(za->za_os); + spa_t *spa = za->za_spa; traverse_handle_t *th = za->za_th; int rc, advance; uint64_t cbstart, cblimit; @@ -1489,7 +1601,7 @@ ztest_dmu_object_alloc_free(ztest_args_t *za) dmu_tx_t *tx; uint64_t batchobj, object, batchsize, endoff, temp; int b, c, error, bonuslen; - dmu_object_info_t doi; + dmu_object_info_t *doi = &za->za_doi; char osname[MAXNAMELEN]; dmu_objset_name(os, osname); @@ -1500,7 +1612,7 @@ ztest_dmu_object_alloc_free(ztest_args_t *za) /* * Create a batch object if necessary, and record it in the directory. */ - VERIFY(0 == dmu_read(os, ZTEST_DIROBJ, za->za_diroff, + VERIFY3U(0, ==, dmu_read(os, ZTEST_DIROBJ, za->za_diroff, sizeof (uint64_t), &batchobj)); if (batchobj == 0) { tx = dmu_tx_create(os); @@ -1525,7 +1637,7 @@ ztest_dmu_object_alloc_free(ztest_args_t *za) * Destroy the previous batch of objects. */ for (b = 0; b < batchsize; b++) { - VERIFY(0 == dmu_read(os, batchobj, b * sizeof (uint64_t), + VERIFY3U(0, ==, dmu_read(os, batchobj, b * sizeof (uint64_t), sizeof (uint64_t), &object)); if (object == 0) continue; @@ -1534,13 +1646,14 @@ ztest_dmu_object_alloc_free(ztest_args_t *za) * We expect the nth byte of the bonus buffer to be n. */ VERIFY(0 == dmu_bonus_hold(os, object, FTAG, &db)); + za->za_dbuf = db; - dmu_object_info_from_db(db, &doi); - ASSERT(doi.doi_type == DMU_OT_UINT64_OTHER); - ASSERT(doi.doi_bonus_type == DMU_OT_PLAIN_OTHER); - ASSERT3S(doi.doi_physical_blks, >=, 0); + dmu_object_info_from_db(db, doi); + ASSERT(doi->doi_type == DMU_OT_UINT64_OTHER); + ASSERT(doi->doi_bonus_type == DMU_OT_PLAIN_OTHER); + ASSERT3S(doi->doi_physical_blks, >=, 0); - bonuslen = db->db_size; + bonuslen = doi->doi_bonus_size; for (c = 0; c < bonuslen; c++) { if (((uint8_t *)db->db_data)[c] != @@ -1554,6 +1667,7 @@ ztest_dmu_object_alloc_free(ztest_args_t *za) } dmu_buf_rele(db, FTAG); + za->za_dbuf = NULL; /* * We expect the word at endoff to be our object number. @@ -1658,8 +1772,9 @@ ztest_dmu_object_alloc_free(ztest_args_t *za) /* * Write to both the bonus buffer and the regular data. */ - VERIFY(0 == dmu_bonus_hold(os, object, FTAG, &db)); - ASSERT3U(bonuslen, ==, db->db_size); + VERIFY(dmu_bonus_hold(os, object, FTAG, &db) == 0); + za->za_dbuf = db; + ASSERT3U(bonuslen, <=, db->db_size); dmu_object_size_from_db(db, &va_blksize, &va_nblocks); ASSERT3S(va_nblocks, >=, 0); @@ -1670,10 +1785,11 @@ ztest_dmu_object_alloc_free(ztest_args_t *za) * See comments above regarding the contents of * the bonus buffer and the word at endoff. */ - for (c = 0; c < db->db_size; c++) + for (c = 0; c < bonuslen; c++) ((uint8_t *)db->db_data)[c] = (uint8_t)(c + bonuslen); dmu_buf_rele(db, FTAG); + za->za_dbuf = NULL; /* * Write to a large offset to increase indirection. @@ -1928,226 +2044,240 @@ ztest_dmu_read_write(ztest_args_t *za) } void -ztest_dmu_check_future_leak(objset_t *os, uint64_t txg) +ztest_dmu_check_future_leak(ztest_args_t *za) { + objset_t *os = za->za_os; dmu_buf_t *db; - ztest_block_tag_t rbt; - - if (zopt_verbose >= 3) { - char osname[MAXNAMELEN]; - dmu_objset_name(os, osname); - (void) printf("checking %s for future leaks in txg %lld...\n", - osname, (u_longlong_t)txg); - } + ztest_block_tag_t *bt; + dmu_object_info_t *doi = &za->za_doi; /* * Make sure that, if there is a write record in the bonus buffer * of the ZTEST_DIROBJ, that the txg for this record is <= the * last synced txg of the pool. */ - - VERIFY(0 == dmu_bonus_hold(os, ZTEST_DIROBJ, FTAG, &db)); - ASSERT3U(db->db_size, ==, sizeof (rbt)); - bcopy(db->db_data, &rbt, db->db_size); - if (rbt.bt_objset != 0) { - ASSERT3U(rbt.bt_objset, ==, dmu_objset_id(os)); - ASSERT3U(rbt.bt_object, ==, ZTEST_DIROBJ); - ASSERT3U(rbt.bt_offset, ==, -1ULL); - if (rbt.bt_txg > txg) { - fatal(0, - "future leak: got %llx, last synced txg is %llx", - rbt.bt_txg, txg); - } + VERIFY(dmu_bonus_hold(os, ZTEST_DIROBJ, FTAG, &db) == 0); + za->za_dbuf = db; + VERIFY(dmu_object_info(os, ZTEST_DIROBJ, doi) == 0); + ASSERT3U(doi->doi_bonus_size, >=, sizeof (*bt)); + ASSERT3U(doi->doi_bonus_size, <=, db->db_size); + ASSERT3U(doi->doi_bonus_size % sizeof (*bt), ==, 0); + bt = (void *)((char *)db->db_data + doi->doi_bonus_size - sizeof (*bt)); + if (bt->bt_objset != 0) { + ASSERT3U(bt->bt_objset, ==, dmu_objset_id(os)); + ASSERT3U(bt->bt_object, ==, ZTEST_DIROBJ); + ASSERT3U(bt->bt_offset, ==, -1ULL); + ASSERT3U(bt->bt_txg, <, spa_first_txg(za->za_spa)); } dmu_buf_rele(db, FTAG); + za->za_dbuf = NULL; } void ztest_dmu_write_parallel(ztest_args_t *za) { objset_t *os = za->za_os; - dmu_tx_t *tx; + ztest_block_tag_t *rbt = &za->za_rbt; + ztest_block_tag_t *wbt = &za->za_wbt; + const size_t btsize = sizeof (ztest_block_tag_t); dmu_buf_t *db; - int i, b, error, do_free, bs; - uint64_t off, txg_how, txg; + int b, error; + int bs = ZTEST_DIROBJ_BLOCKSIZE; + int do_free = 0; + uint64_t off, txg, txg_how; mutex_t *lp; char osname[MAXNAMELEN]; char iobuf[SPA_MAXBLOCKSIZE]; - ztest_block_tag_t rbt, wbt; + blkptr_t blk = { 0 }; + uint64_t blkoff; + zbookmark_t zb; + dmu_tx_t *tx = dmu_tx_create(os); dmu_objset_name(os, osname); - bs = ZTEST_DIROBJ_BLOCKSIZE; /* * Have multiple threads write to large offsets in ZTEST_DIROBJ * to verify that having multiple threads writing to the same object * in parallel doesn't cause any trouble. - * Also do parallel writes to the bonus buffer on occasion. */ - for (i = 0; i < 50; i++) { + if (ztest_random(4) == 0) { + /* + * Do the bonus buffer instead of a regular block. + * We need a lock to serialize resize vs. others, + * so we hash on the objset ID. + */ + b = dmu_objset_id(os) % ZTEST_SYNC_LOCKS; + off = -1ULL; + dmu_tx_hold_bonus(tx, ZTEST_DIROBJ); + } else { b = ztest_random(ZTEST_SYNC_LOCKS); - lp = &ztest_shared->zs_sync_lock[b]; - - do_free = (ztest_random(4) == 0); - - off = za->za_diroff_shared + ((uint64_t)b << SPA_MAXBLOCKSHIFT); - + off = za->za_diroff_shared + (b << SPA_MAXBLOCKSHIFT); if (ztest_random(4) == 0) { - /* - * Do the bonus buffer instead of a regular block. - */ - do_free = 0; - off = -1ULL; - } - - tx = dmu_tx_create(os); - - if (off == -1ULL) - dmu_tx_hold_bonus(tx, ZTEST_DIROBJ); - else if (do_free) + do_free = 1; dmu_tx_hold_free(tx, ZTEST_DIROBJ, off, bs); - else + } else { dmu_tx_hold_write(tx, ZTEST_DIROBJ, off, bs); + } + } - txg_how = ztest_random(2) == 0 ? TXG_WAIT : TXG_NOWAIT; - error = dmu_tx_assign(tx, txg_how); - if (error) { - if (error == ERESTART) { - ASSERT(txg_how == TXG_NOWAIT); - dmu_tx_wait(tx); - dmu_tx_abort(tx); - continue; - } - dmu_tx_abort(tx); + txg_how = ztest_random(2) == 0 ? TXG_WAIT : TXG_NOWAIT; + error = dmu_tx_assign(tx, txg_how); + if (error) { + if (error == ERESTART) { + ASSERT(txg_how == TXG_NOWAIT); + dmu_tx_wait(tx); + } else { ztest_record_enospc("dmu write parallel"); - return; } - txg = dmu_tx_get_txg(tx); + dmu_tx_abort(tx); + return; + } + txg = dmu_tx_get_txg(tx); - if (do_free) { - (void) mutex_lock(lp); - VERIFY(0 == dmu_free_range(os, ZTEST_DIROBJ, off, - bs, tx)); - (void) mutex_unlock(lp); - dmu_tx_commit(tx); - continue; + lp = &ztest_shared->zs_sync_lock[b]; + (void) mutex_lock(lp); + + wbt->bt_objset = dmu_objset_id(os); + wbt->bt_object = ZTEST_DIROBJ; + wbt->bt_offset = off; + wbt->bt_txg = txg; + wbt->bt_thread = za->za_instance; + wbt->bt_seq = ztest_shared->zs_seq[b]++; /* protected by lp */ + + /* + * Occasionally, write an all-zero block to test the behavior + * of blocks that compress into holes. + */ + if (off != -1ULL && ztest_random(8) == 0) + bzero(wbt, btsize); + + if (off == -1ULL) { + dmu_object_info_t *doi = &za->za_doi; + char *dboff; + + VERIFY(dmu_bonus_hold(os, ZTEST_DIROBJ, FTAG, &db) == 0); + za->za_dbuf = db; + dmu_object_info_from_db(db, doi); + ASSERT3U(doi->doi_bonus_size, <=, db->db_size); + ASSERT3U(doi->doi_bonus_size, >=, btsize); + ASSERT3U(doi->doi_bonus_size % btsize, ==, 0); + dboff = (char *)db->db_data + doi->doi_bonus_size - btsize; + bcopy(dboff, rbt, btsize); + if (rbt->bt_objset != 0) { + ASSERT3U(rbt->bt_objset, ==, wbt->bt_objset); + ASSERT3U(rbt->bt_object, ==, wbt->bt_object); + ASSERT3U(rbt->bt_offset, ==, wbt->bt_offset); + ASSERT3U(rbt->bt_txg, <=, wbt->bt_txg); } - - wbt.bt_objset = dmu_objset_id(os); - wbt.bt_object = ZTEST_DIROBJ; - wbt.bt_offset = off; - wbt.bt_txg = txg; - wbt.bt_thread = za->za_instance; - - if (off == -1ULL) { - wbt.bt_seq = 0; - VERIFY(0 == dmu_bonus_hold(os, ZTEST_DIROBJ, - FTAG, &db)); - ASSERT3U(db->db_size, ==, sizeof (wbt)); - bcopy(db->db_data, &rbt, db->db_size); - if (rbt.bt_objset != 0) { - ASSERT3U(rbt.bt_objset, ==, wbt.bt_objset); - ASSERT3U(rbt.bt_object, ==, wbt.bt_object); - ASSERT3U(rbt.bt_offset, ==, wbt.bt_offset); - ASSERT3U(rbt.bt_txg, <=, wbt.bt_txg); - } - dmu_buf_will_dirty(db, tx); - bcopy(&wbt, db->db_data, db->db_size); - dmu_buf_rele(db, FTAG); - dmu_tx_commit(tx); - continue; + if (ztest_random(10) == 0) { + int newsize = (ztest_random(db->db_size / + btsize) + 1) * btsize; + + ASSERT3U(newsize, >=, btsize); + ASSERT3U(newsize, <=, db->db_size); + VERIFY3U(dmu_set_bonus(db, newsize, tx), ==, 0); + dboff = (char *)db->db_data + newsize - btsize; } + dmu_buf_will_dirty(db, tx); + bcopy(wbt, dboff, btsize); + dmu_buf_rele(db, FTAG); + za->za_dbuf = NULL; + } else if (do_free) { + VERIFY(dmu_free_range(os, ZTEST_DIROBJ, off, bs, tx) == 0); + } else { + dmu_write(os, ZTEST_DIROBJ, off, btsize, wbt, tx); + } - (void) mutex_lock(lp); + (void) mutex_unlock(lp); - wbt.bt_seq = ztest_shared->zs_seq[b]++; + if (ztest_random(1000) == 0) + (void) poll(NULL, 0, 1); /* open dn_notxholds window */ - dmu_write(os, ZTEST_DIROBJ, off, sizeof (wbt), &wbt, tx); + dmu_tx_commit(tx); + + if (ztest_random(10000) == 0) + txg_wait_synced(dmu_objset_pool(os), txg); + + if (off == -1ULL || do_free) + return; + if (ztest_random(2) != 0) + return; + + /* + * dmu_sync() the block we just wrote. + */ + (void) mutex_lock(lp); + + blkoff = P2ALIGN_TYPED(off, bs, uint64_t); + error = dmu_buf_hold(os, ZTEST_DIROBJ, blkoff, FTAG, &db); + za->za_dbuf = db; + if (error) { + dprintf("dmu_buf_hold(%s, %d, %llx) = %d\n", + osname, ZTEST_DIROBJ, blkoff, error); (void) mutex_unlock(lp); + return; + } + blkoff = off - blkoff; + error = dmu_sync(NULL, db, &blk, txg, NULL, NULL); + dmu_buf_rele(db, FTAG); + za->za_dbuf = NULL; - if (ztest_random(100) == 0) - (void) poll(NULL, 0, 1); /* open dn_notxholds window */ + (void) mutex_unlock(lp); - dmu_tx_commit(tx); + if (error) { + dprintf("dmu_sync(%s, %d, %llx) = %d\n", + osname, ZTEST_DIROBJ, off, error); + return; + } - if (ztest_random(1000) == 0) - txg_wait_synced(dmu_objset_pool(os), txg); - - if (ztest_random(2) == 0) { - blkptr_t blk = { 0 }; - uint64_t blkoff; - zbookmark_t zb; - - (void) mutex_lock(lp); - blkoff = P2ALIGN_TYPED(off, bs, uint64_t); - error = dmu_buf_hold(os, - ZTEST_DIROBJ, blkoff, FTAG, &db); - if (error) { - dprintf("dmu_buf_hold(%s, %d, %llx) = %d\n", - osname, ZTEST_DIROBJ, blkoff, error); - (void) mutex_unlock(lp); - continue; - } - blkoff = off - blkoff; - error = dmu_sync(NULL, db, &blk, txg, NULL, NULL); - dmu_buf_rele(db, FTAG); - (void) mutex_unlock(lp); - if (error) { - dprintf("dmu_sync(%s, %d, %llx) = %d\n", - osname, ZTEST_DIROBJ, off, error); - continue; - } + if (blk.blk_birth == 0) /* concurrent free */ + return; - if (blk.blk_birth == 0) { /* concurrent free */ - continue; - } - txg_suspend(dmu_objset_pool(os)); + txg_suspend(dmu_objset_pool(os)); - ASSERT(blk.blk_fill == 1); - ASSERT3U(BP_GET_TYPE(&blk), ==, DMU_OT_UINT64_OTHER); - ASSERT3U(BP_GET_LEVEL(&blk), ==, 0); - ASSERT3U(BP_GET_LSIZE(&blk), ==, bs); + ASSERT(blk.blk_fill == 1); + ASSERT3U(BP_GET_TYPE(&blk), ==, DMU_OT_UINT64_OTHER); + ASSERT3U(BP_GET_LEVEL(&blk), ==, 0); + ASSERT3U(BP_GET_LSIZE(&blk), ==, bs); - /* - * Read the block that dmu_sync() returned to - * make sure its contents match what we wrote. - * We do this while still txg_suspend()ed to ensure - * that the block can't be reused before we read it. - */ - zb.zb_objset = dmu_objset_id(os); - zb.zb_object = ZTEST_DIROBJ; - zb.zb_level = 0; - zb.zb_blkid = off / bs; - error = zio_wait(zio_read(NULL, dmu_objset_spa(os), - &blk, iobuf, bs, NULL, NULL, - ZIO_PRIORITY_SYNC_READ, ZIO_FLAG_MUSTSUCCEED, &zb)); - ASSERT(error == 0); + /* + * Read the block that dmu_sync() returned to make sure its contents + * match what we wrote. We do this while still txg_suspend()ed + * to ensure that the block can't be reused before we read it. + */ + zb.zb_objset = dmu_objset_id(os); + zb.zb_object = ZTEST_DIROBJ; + zb.zb_level = 0; + zb.zb_blkid = off / bs; + error = zio_wait(zio_read(NULL, za->za_spa, &blk, iobuf, bs, + NULL, NULL, ZIO_PRIORITY_SYNC_READ, ZIO_FLAG_MUSTSUCCEED, &zb)); + ASSERT3U(error, ==, 0); - txg_resume(dmu_objset_pool(os)); + txg_resume(dmu_objset_pool(os)); - bcopy(&iobuf[blkoff], &rbt, sizeof (rbt)); + bcopy(&iobuf[blkoff], rbt, btsize); - if (rbt.bt_objset == 0) /* concurrent free */ - continue; + if (rbt->bt_objset == 0) /* concurrent free */ + return; - ASSERT3U(rbt.bt_objset, ==, wbt.bt_objset); - ASSERT3U(rbt.bt_object, ==, wbt.bt_object); - ASSERT3U(rbt.bt_offset, ==, wbt.bt_offset); + if (wbt->bt_objset == 0) /* all-zero overwrite */ + return; - /* - * The semantic of dmu_sync() is that we always - * push the most recent version of the data, - * so in the face of concurrent updates we may - * see a newer version of the block. That's OK. - */ - ASSERT3U(rbt.bt_txg, >=, wbt.bt_txg); - if (rbt.bt_thread == wbt.bt_thread) - ASSERT3U(rbt.bt_seq, ==, wbt.bt_seq); - else - ASSERT3U(rbt.bt_seq, >, wbt.bt_seq); - } - } + ASSERT3U(rbt->bt_objset, ==, wbt->bt_objset); + ASSERT3U(rbt->bt_object, ==, wbt->bt_object); + ASSERT3U(rbt->bt_offset, ==, wbt->bt_offset); + + /* + * The semantic of dmu_sync() is that we always push the most recent + * version of the data, so in the face of concurrent updates we may + * see a newer version of the block. That's OK. + */ + ASSERT3U(rbt->bt_txg, >=, wbt->bt_txg); + if (rbt->bt_thread == wbt->bt_thread) + ASSERT3U(rbt->bt_seq, ==, wbt->bt_seq); + else + ASSERT3U(rbt->bt_seq, >, wbt->bt_seq); } /* @@ -2166,7 +2296,6 @@ ztest_zap(ztest_args_t *za) uint64_t value[ZTEST_ZAP_MAX_INTS]; uint64_t zl_ints, zl_intsize, prop; int i, ints; - int iters = 100; dmu_tx_t *tx; char propname[100], txgname[100]; int error; @@ -2230,122 +2359,113 @@ ztest_zap(ztest_args_t *za) ints = MAX(ZTEST_ZAP_MIN_INTS, object % ZTEST_ZAP_MAX_INTS); - while (--iters >= 0) { - prop = ztest_random(ZTEST_ZAP_MAX_PROPS); - (void) sprintf(propname, "prop_%llu", (u_longlong_t)prop); - (void) sprintf(txgname, "txg_%llu", (u_longlong_t)prop); - bzero(value, sizeof (value)); - last_txg = 0; - - /* - * If these zap entries already exist, validate their contents. - */ - error = zap_length(os, object, txgname, &zl_intsize, &zl_ints); - if (error == 0) { - ASSERT3U(zl_intsize, ==, sizeof (uint64_t)); - ASSERT3U(zl_ints, ==, 1); - - error = zap_lookup(os, object, txgname, zl_intsize, - zl_ints, &last_txg); + prop = ztest_random(ZTEST_ZAP_MAX_PROPS); + (void) sprintf(propname, "prop_%llu", (u_longlong_t)prop); + (void) sprintf(txgname, "txg_%llu", (u_longlong_t)prop); + bzero(value, sizeof (value)); + last_txg = 0; - ASSERT3U(error, ==, 0); + /* + * If these zap entries already exist, validate their contents. + */ + error = zap_length(os, object, txgname, &zl_intsize, &zl_ints); + if (error == 0) { + ASSERT3U(zl_intsize, ==, sizeof (uint64_t)); + ASSERT3U(zl_ints, ==, 1); - error = zap_length(os, object, propname, &zl_intsize, - &zl_ints); + VERIFY(zap_lookup(os, object, txgname, zl_intsize, + zl_ints, &last_txg) == 0); - ASSERT3U(error, ==, 0); - ASSERT3U(zl_intsize, ==, sizeof (uint64_t)); - ASSERT3U(zl_ints, ==, ints); + VERIFY(zap_length(os, object, propname, &zl_intsize, + &zl_ints) == 0); - error = zap_lookup(os, object, propname, zl_intsize, - zl_ints, value); + ASSERT3U(zl_intsize, ==, sizeof (uint64_t)); + ASSERT3U(zl_ints, ==, ints); - ASSERT3U(error, ==, 0); + VERIFY(zap_lookup(os, object, propname, zl_intsize, + zl_ints, value) == 0); - for (i = 0; i < ints; i++) { - ASSERT3U(value[i], ==, last_txg + object + i); - } - } else { - ASSERT3U(error, ==, ENOENT); + for (i = 0; i < ints; i++) { + ASSERT3U(value[i], ==, last_txg + object + i); } + } else { + ASSERT3U(error, ==, ENOENT); + } - /* - * Atomically update two entries in our zap object. - * The first is named txg_%llu, and contains the txg - * in which the property was last updated. The second - * is named prop_%llu, and the nth element of its value - * should be txg + object + n. - */ - tx = dmu_tx_create(os); - dmu_tx_hold_zap(tx, object, TRUE, NULL); - error = dmu_tx_assign(tx, TXG_WAIT); - if (error) { - ztest_record_enospc("create zap entry"); - dmu_tx_abort(tx); - return; - } - txg = dmu_tx_get_txg(tx); + /* + * Atomically update two entries in our zap object. + * The first is named txg_%llu, and contains the txg + * in which the property was last updated. The second + * is named prop_%llu, and the nth element of its value + * should be txg + object + n. + */ + tx = dmu_tx_create(os); + dmu_tx_hold_zap(tx, object, TRUE, NULL); + error = dmu_tx_assign(tx, TXG_WAIT); + if (error) { + ztest_record_enospc("create zap entry"); + dmu_tx_abort(tx); + return; + } + txg = dmu_tx_get_txg(tx); - if (last_txg > txg) - fatal(0, "zap future leak: old %llu new %llu", - last_txg, txg); + if (last_txg > txg) + fatal(0, "zap future leak: old %llu new %llu", last_txg, txg); - for (i = 0; i < ints; i++) - value[i] = txg + object + i; + for (i = 0; i < ints; i++) + value[i] = txg + object + i; - error = zap_update(os, object, txgname, sizeof (uint64_t), - 1, &txg, tx); - if (error) - fatal(0, "zap_update('%s', %llu, '%s') = %d", - osname, object, txgname, error); + error = zap_update(os, object, txgname, sizeof (uint64_t), 1, &txg, tx); + if (error) + fatal(0, "zap_update('%s', %llu, '%s') = %d", + osname, object, txgname, error); - error = zap_update(os, object, propname, sizeof (uint64_t), - ints, value, tx); - if (error) - fatal(0, "zap_update('%s', %llu, '%s') = %d", - osname, object, propname, error); + error = zap_update(os, object, propname, sizeof (uint64_t), + ints, value, tx); + if (error) + fatal(0, "zap_update('%s', %llu, '%s') = %d", + osname, object, propname, error); - dmu_tx_commit(tx); + dmu_tx_commit(tx); - /* - * Remove a random pair of entries. - */ - prop = ztest_random(ZTEST_ZAP_MAX_PROPS); - (void) sprintf(propname, "prop_%llu", (u_longlong_t)prop); - (void) sprintf(txgname, "txg_%llu", (u_longlong_t)prop); + /* + * Remove a random pair of entries. + */ + prop = ztest_random(ZTEST_ZAP_MAX_PROPS); + (void) sprintf(propname, "prop_%llu", (u_longlong_t)prop); + (void) sprintf(txgname, "txg_%llu", (u_longlong_t)prop); - error = zap_length(os, object, txgname, &zl_intsize, &zl_ints); + error = zap_length(os, object, txgname, &zl_intsize, &zl_ints); - if (error == ENOENT) - continue; + if (error == ENOENT) + return; - ASSERT3U(error, ==, 0); + ASSERT3U(error, ==, 0); - tx = dmu_tx_create(os); - dmu_tx_hold_zap(tx, object, TRUE, NULL); - error = dmu_tx_assign(tx, TXG_WAIT); - if (error) { - ztest_record_enospc("remove zap entry"); - dmu_tx_abort(tx); - return; - } - error = zap_remove(os, object, txgname, tx); - if (error) - fatal(0, "zap_remove('%s', %llu, '%s') = %d", - osname, object, txgname, error); + tx = dmu_tx_create(os); + dmu_tx_hold_zap(tx, object, TRUE, NULL); + error = dmu_tx_assign(tx, TXG_WAIT); + if (error) { + ztest_record_enospc("remove zap entry"); + dmu_tx_abort(tx); + return; + } + error = zap_remove(os, object, txgname, tx); + if (error) + fatal(0, "zap_remove('%s', %llu, '%s') = %d", + osname, object, txgname, error); - error = zap_remove(os, object, propname, tx); - if (error) - fatal(0, "zap_remove('%s', %llu, '%s') = %d", - osname, object, propname, error); + error = zap_remove(os, object, propname, tx); + if (error) + fatal(0, "zap_remove('%s', %llu, '%s') = %d", + osname, object, propname, error); - dmu_tx_commit(tx); - } + dmu_tx_commit(tx); /* * Once in a while, destroy the object. */ - if (ztest_random(100) != 0) + if (ztest_random(1000) != 0) return; tx = dmu_tx_create(os); @@ -2372,111 +2492,107 @@ ztest_zap_parallel(ztest_args_t *za) { objset_t *os = za->za_os; uint64_t txg, object, count, wsize, wc, zl_wsize, zl_wc; - int iters = 100; dmu_tx_t *tx; int i, namelen, error; char name[20], string_value[20]; void *data; - while (--iters >= 0) { - /* - * Generate a random name of the form 'xxx.....' where each - * x is a random printable character and the dots are dots. - * There are 94 such characters, and the name length goes from - * 6 to 20, so there are 94^3 * 15 = 12,458,760 possible names. - */ - namelen = ztest_random(sizeof (name) - 5) + 5 + 1; + /* + * Generate a random name of the form 'xxx.....' where each + * x is a random printable character and the dots are dots. + * There are 94 such characters, and the name length goes from + * 6 to 20, so there are 94^3 * 15 = 12,458,760 possible names. + */ + namelen = ztest_random(sizeof (name) - 5) + 5 + 1; - for (i = 0; i < 3; i++) - name[i] = '!' + ztest_random('~' - '!' + 1); - for (; i < namelen - 1; i++) - name[i] = '.'; - name[i] = '\0'; + for (i = 0; i < 3; i++) + name[i] = '!' + ztest_random('~' - '!' + 1); + for (; i < namelen - 1; i++) + name[i] = '.'; + name[i] = '\0'; - if (ztest_random(2) == 0) - object = ZTEST_MICROZAP_OBJ; - else - object = ZTEST_FATZAP_OBJ; + if (ztest_random(2) == 0) + object = ZTEST_MICROZAP_OBJ; + else + object = ZTEST_FATZAP_OBJ; - if ((namelen & 1) || object == ZTEST_MICROZAP_OBJ) { - wsize = sizeof (txg); - wc = 1; - data = &txg; - } else { - wsize = 1; - wc = namelen; - data = string_value; - } + if ((namelen & 1) || object == ZTEST_MICROZAP_OBJ) { + wsize = sizeof (txg); + wc = 1; + data = &txg; + } else { + wsize = 1; + wc = namelen; + data = string_value; + } - count = -1ULL; - VERIFY(zap_count(os, object, &count) == 0); - ASSERT(count != -1ULL); + count = -1ULL; + VERIFY(zap_count(os, object, &count) == 0); + ASSERT(count != -1ULL); - /* - * Select an operation: length, lookup, add, update, remove. - */ - i = ztest_random(5); - - if (i >= 2) { - tx = dmu_tx_create(os); - dmu_tx_hold_zap(tx, object, TRUE, NULL); - error = dmu_tx_assign(tx, TXG_WAIT); - if (error) { - ztest_record_enospc("zap parallel"); - dmu_tx_abort(tx); - return; - } - txg = dmu_tx_get_txg(tx); - bcopy(name, string_value, namelen); - } else { - tx = NULL; - txg = 0; - bzero(string_value, namelen); - } + /* + * Select an operation: length, lookup, add, update, remove. + */ + i = ztest_random(5); - switch (i) { + if (i >= 2) { + tx = dmu_tx_create(os); + dmu_tx_hold_zap(tx, object, TRUE, NULL); + error = dmu_tx_assign(tx, TXG_WAIT); + if (error) { + ztest_record_enospc("zap parallel"); + dmu_tx_abort(tx); + return; + } + txg = dmu_tx_get_txg(tx); + bcopy(name, string_value, namelen); + } else { + tx = NULL; + txg = 0; + bzero(string_value, namelen); + } - case 0: - error = zap_length(os, object, name, &zl_wsize, &zl_wc); - if (error == 0) { - ASSERT3U(wsize, ==, zl_wsize); - ASSERT3U(wc, ==, zl_wc); - } else { - ASSERT3U(error, ==, ENOENT); - } - break; + switch (i) { - case 1: - error = zap_lookup(os, object, name, wsize, wc, data); - if (error == 0) { - if (data == string_value && - bcmp(name, data, namelen) != 0) - fatal(0, "name '%s' != val '%s' len %d", - name, data, namelen); - } else { - ASSERT3U(error, ==, ENOENT); - } - break; + case 0: + error = zap_length(os, object, name, &zl_wsize, &zl_wc); + if (error == 0) { + ASSERT3U(wsize, ==, zl_wsize); + ASSERT3U(wc, ==, zl_wc); + } else { + ASSERT3U(error, ==, ENOENT); + } + break; - case 2: - error = zap_add(os, object, name, wsize, wc, data, tx); - ASSERT(error == 0 || error == EEXIST); - break; + case 1: + error = zap_lookup(os, object, name, wsize, wc, data); + if (error == 0) { + if (data == string_value && + bcmp(name, data, namelen) != 0) + fatal(0, "name '%s' != val '%s' len %d", + name, data, namelen); + } else { + ASSERT3U(error, ==, ENOENT); + } + break; - case 3: - VERIFY(zap_update(os, object, name, wsize, wc, - data, tx) == 0); - break; + case 2: + error = zap_add(os, object, name, wsize, wc, data, tx); + ASSERT(error == 0 || error == EEXIST); + break; - case 4: - error = zap_remove(os, object, name, tx); - ASSERT(error == 0 || error == ENOENT); - break; - } + case 3: + VERIFY(zap_update(os, object, name, wsize, wc, data, tx) == 0); + break; - if (tx != NULL) - dmu_tx_commit(tx); + case 4: + error = zap_remove(os, object, name, tx); + ASSERT(error == 0 || error == ENOENT); + break; } + + if (tx != NULL) + dmu_tx_commit(tx); } void @@ -2532,21 +2648,6 @@ ztest_dsl_prop_get_set(ztest_args_t *za) (void) rw_unlock(&ztest_shared->zs_name_lock); } -static void -ztest_error_setup(vdev_t *vd, int mode, int mask, uint64_t arg) -{ - int c; - - for (c = 0; c < vd->vdev_children; c++) - ztest_error_setup(vd->vdev_child[c], mode, mask, arg); - - if (vd->vdev_path != NULL) { - vd->vdev_fault_mode = mode; - vd->vdev_fault_mask = mask; - vd->vdev_fault_arg = arg; - } -} - /* * Inject random faults into the on-disk data. */ @@ -2561,67 +2662,97 @@ ztest_fault_inject(ztest_args_t *za) char path0[MAXPATHLEN]; char pathrand[MAXPATHLEN]; size_t fsize; - spa_t *spa = dmu_objset_spa(za->za_os); + spa_t *spa = za->za_spa; int bshift = SPA_MAXBLOCKSHIFT + 2; /* don't scrog all labels */ int iters = 1000; - vdev_t *vd0; + int maxfaults = zopt_maxfaults; + vdev_t *vd0 = NULL; uint64_t guid0 = 0; - /* - * We can't inject faults when we have no fault tolerance. - */ - if (zopt_maxfaults == 0) - return; - - ASSERT(leaves >= 2); + ASSERT(leaves >= 1); /* - * Pick a random top-level vdev. + * We need SCL_STATE here because we're going to look at vd0->vdev_tsd. */ - spa_config_enter(spa, RW_READER, FTAG); - top = ztest_random(spa->spa_root_vdev->vdev_children); - spa_config_exit(spa, FTAG); + spa_config_enter(spa, SCL_STATE, FTAG, RW_READER); - /* - * Pick a random leaf. - */ - leaf = ztest_random(leaves); + if (ztest_random(2) == 0) { + /* + * Inject errors on a normal data device. + */ + top = ztest_random(spa->spa_root_vdev->vdev_children); + leaf = ztest_random(leaves); - /* - * Generate paths to the first two leaves in this top-level vdev, - * and to the random leaf we selected. We'll induce transient - * I/O errors and random online/offline activity on leaf 0, - * and we'll write random garbage to the randomly chosen leaf. - */ - (void) snprintf(path0, sizeof (path0), - ztest_dev_template, zopt_dir, zopt_pool, top * leaves + 0); - (void) snprintf(pathrand, sizeof (pathrand), - ztest_dev_template, zopt_dir, zopt_pool, top * leaves + leaf); + /* + * Generate paths to the first leaf in this top-level vdev, + * and to the random leaf we selected. We'll induce transient + * write failures and random online/offline activity on leaf 0, + * and we'll write random garbage to the randomly chosen leaf. + */ + (void) snprintf(path0, sizeof (path0), ztest_dev_template, + zopt_dir, zopt_pool, top * leaves + 0); + (void) snprintf(pathrand, sizeof (pathrand), ztest_dev_template, + zopt_dir, zopt_pool, top * leaves + leaf); - dprintf("damaging %s and %s\n", path0, pathrand); + vd0 = vdev_lookup_by_path(spa->spa_root_vdev, path0); + if (vd0 != NULL && maxfaults != 1) { + /* + * Make vd0 explicitly claim to be unreadable, + * or unwriteable, or reach behind its back + * and close the underlying fd. We can do this if + * maxfaults == 0 because we'll fail and reexecute, + * and we can do it if maxfaults >= 2 because we'll + * have enough redundancy. If maxfaults == 1, the + * combination of this with injection of random data + * corruption below exceeds the pool's fault tolerance. + */ + vdev_file_t *vf = vd0->vdev_tsd; - spa_config_enter(spa, RW_READER, FTAG); + if (vf != NULL && ztest_random(3) == 0) { + (void) close(vf->vf_vnode->v_fd); + vf->vf_vnode->v_fd = -1; + } else if (ztest_random(2) == 0) { + vd0->vdev_cant_read = B_TRUE; + } else { + vd0->vdev_cant_write = B_TRUE; + } + guid0 = vd0->vdev_guid; + } + } else { + /* + * Inject errors on an l2cache device. + */ + spa_aux_vdev_t *sav = &spa->spa_l2cache; - /* - * If we can tolerate two or more faults, make vd0 fail randomly. - */ - vd0 = vdev_lookup_by_path(spa->spa_root_vdev, path0); - if (vd0 != NULL && zopt_maxfaults >= 2) { + if (sav->sav_count == 0) { + spa_config_exit(spa, SCL_STATE, FTAG); + return; + } + vd0 = sav->sav_vdevs[ztest_random(sav->sav_count)]; guid0 = vd0->vdev_guid; - ztest_error_setup(vd0, VDEV_FAULT_COUNT, - (1U << ZIO_TYPE_READ) | (1U << ZIO_TYPE_WRITE), 100); + (void) strcpy(path0, vd0->vdev_path); + (void) strcpy(pathrand, vd0->vdev_path); + + leaf = 0; + leaves = 1; + maxfaults = INT_MAX; /* no limit on cache devices */ } - spa_config_exit(spa, FTAG); + dprintf("damaging %s and %s\n", path0, pathrand); + + spa_config_exit(spa, SCL_STATE, FTAG); + + if (maxfaults == 0) + return; /* * If we can tolerate two or more faults, randomly online/offline vd0. */ - if (zopt_maxfaults >= 2 && guid0 != 0) { + if (maxfaults >= 2 && guid0 != 0) { if (ztest_random(10) < 6) (void) vdev_offline(spa, guid0, B_TRUE); else - (void) vdev_online(spa, guid0); + (void) vdev_online(spa, guid0, B_FALSE, NULL); } /* @@ -2660,11 +2791,11 @@ ztest_fault_inject(ztest_args_t *za) void ztest_scrub(ztest_args_t *za) { - spa_t *spa = dmu_objset_spa(za->za_os); + spa_t *spa = za->za_spa; - (void) spa_scrub(spa, POOL_SCRUB_EVERYTHING, B_FALSE); + (void) spa_scrub(spa, POOL_SCRUB_EVERYTHING); (void) poll(NULL, 0, 1000); /* wait a second, then force a restart */ - (void) spa_scrub(spa, POOL_SCRUB_EVERYTHING, B_FALSE); + (void) spa_scrub(spa, POOL_SCRUB_EVERYTHING); } /* @@ -2706,7 +2837,7 @@ ztest_spa_rename(ztest_args_t *za) if (error != 0) fatal(0, "spa_open('%s') = %d", newname, error); - ASSERT(spa == dmu_objset_spa(za->za_os)); + ASSERT(spa == za->za_spa); spa_close(spa, FTAG); /* @@ -2724,7 +2855,7 @@ ztest_spa_rename(ztest_args_t *za) if (error != 0) fatal(0, "spa_open('%s') = %d", oldname, error); - ASSERT(spa == dmu_objset_spa(za->za_os)); + ASSERT(spa == za->za_spa); spa_close(spa, FTAG); umem_free(newname, strlen(newname) + 1); @@ -2778,10 +2909,9 @@ static void ztest_replace_one_disk(spa_t *spa, uint64_t vdev) { char dev_name[MAXPATHLEN]; - nvlist_t *file, *root; + nvlist_t *root; int error; uint64_t guid; - uint64_t ashift = ztest_get_ashift(); vdev_t *vd; (void) sprintf(dev_name, ztest_dev_template, zopt_dir, zopt_pool, vdev); @@ -2789,22 +2919,14 @@ ztest_replace_one_disk(spa_t *spa, uint64_t vdev) /* * Build the nvlist describing dev_name. */ - VERIFY(nvlist_alloc(&file, NV_UNIQUE_NAME, 0) == 0); - VERIFY(nvlist_add_string(file, ZPOOL_CONFIG_TYPE, VDEV_TYPE_FILE) == 0); - VERIFY(nvlist_add_string(file, ZPOOL_CONFIG_PATH, dev_name) == 0); - VERIFY(nvlist_add_uint64(file, ZPOOL_CONFIG_ASHIFT, ashift) == 0); - - VERIFY(nvlist_alloc(&root, NV_UNIQUE_NAME, 0) == 0); - VERIFY(nvlist_add_string(root, ZPOOL_CONFIG_TYPE, VDEV_TYPE_ROOT) == 0); - VERIFY(nvlist_add_nvlist_array(root, ZPOOL_CONFIG_CHILDREN, - &file, 1) == 0); + root = make_vdev_root(dev_name, NULL, 0, 0, 0, 0, 0, 1); - spa_config_enter(spa, RW_READER, FTAG); + spa_config_enter(spa, SCL_VDEV, FTAG, RW_READER); if ((vd = vdev_lookup_by_path(spa->spa_root_vdev, dev_name)) == NULL) guid = 0; else guid = vd->vdev_guid; - spa_config_exit(spa, FTAG); + spa_config_exit(spa, SCL_VDEV, FTAG); error = spa_vdev_attach(spa, guid, root, B_TRUE); if (error != 0 && error != EBUSY && @@ -2813,7 +2935,6 @@ ztest_replace_one_disk(spa_t *spa, uint64_t vdev) error != EDOM) fatal(0, "spa_vdev_attach(in-place) = %d", error); - nvlist_free(file); nvlist_free(root); } @@ -2824,6 +2945,9 @@ ztest_verify_blocks(char *pool) char zdb[MAXPATHLEN + MAXNAMELEN + 20]; char zbuf[1024]; char *bin; + char *ztest; + char *isa; + int isalen; FILE *fp; if (realpath(progname, zdb) == NULL) @@ -2831,13 +2955,19 @@ ztest_verify_blocks(char *pool) /* zdb lives in /usr/sbin, while ztest lives in /usr/bin */ bin = strstr(zdb, "/usr/bin/"); - if (bin == NULL) - bin = zdb; + ztest = strstr(bin, "/ztest"); + isa = bin + 8; + isalen = ztest - isa; + isa = strdup(isa); /* LINTED */ - (void) sprintf(bin, "/usr/sbin/zdb -bc%s%s -U -O %s %s", + (void) sprintf(bin, + "/usr/sbin%.*s/zdb -bc%s%s -U /tmp/zpool.cache -O %s %s", + isalen, + isa, zopt_verbose >= 3 ? "s" : "", zopt_verbose >= 4 ? "v" : "", ztest_random(2) == 0 ? "pre" : "post", pool); + free(isa); if (zopt_verbose >= 5) (void) printf("Executing %s\n", strstr(zdb, "zdb ")); @@ -2909,7 +3039,7 @@ ztest_spa_import_export(char *oldname, char *newname) /* * Export it. */ - error = spa_export(oldname, &config); + error = spa_export(oldname, &config, B_FALSE); if (error) fatal(0, "spa_export('%s') = %d", oldname, error); @@ -2958,35 +3088,41 @@ ztest_spa_import_export(char *oldname, char *newname) } static void * +ztest_resume(void *arg) +{ + spa_t *spa = arg; + + while (!ztest_exiting) { + (void) poll(NULL, 0, 1000); + + if (!spa_suspended(spa)) + continue; + + spa_vdev_state_enter(spa); + vdev_clear(spa, NULL); + (void) spa_vdev_state_exit(spa, NULL, 0); + + zio_resume(spa); + } + return (NULL); +} + +static void * ztest_thread(void *arg) { ztest_args_t *za = arg; ztest_shared_t *zs = ztest_shared; hrtime_t now, functime; ztest_info_t *zi; - int f; + int f, i; while ((now = gethrtime()) < za->za_stop) { /* * See if it's time to force a crash. */ if (now > za->za_kill) { - dmu_tx_t *tx; - uint64_t txg; - - mutex_enter(&spa_namespace_lock); - tx = dmu_tx_create(za->za_os); - VERIFY(0 == dmu_tx_assign(tx, TXG_NOWAIT)); - txg = dmu_tx_get_txg(tx); - dmu_tx_commit(tx); - zs->zs_txg = txg; - if (zopt_verbose >= 3) - (void) printf( - "killing process after txg %lld\n", - (u_longlong_t)txg); - txg_wait_synced(dmu_objset_pool(za->za_os), txg); - zs->zs_alloc = spa_get_alloc(dmu_objset_spa(za->za_os)); - zs->zs_space = spa_get_space(dmu_objset_spa(za->za_os)); + zs->zs_alloc = spa_get_alloc(za->za_spa); + zs->zs_space = spa_get_space(za->za_spa); (void) kill(getpid(), SIGKILL); } @@ -3011,9 +3147,8 @@ ztest_thread(void *arg) ZTEST_DIRSIZE; za->za_diroff_shared = (1ULL << 63); - ztest_dmu_write_parallel(za); - - zi->zi_func(za); + for (i = 0; i < zi->zi_iters; i++) + zi->zi_func(za); functime = gethrtime() - now; @@ -3047,6 +3182,9 @@ ztest_run(char *pool) ztest_args_t *za; spa_t *spa; char name[100]; + thread_t resume_tid; + + ztest_exiting = B_FALSE; (void) _mutex_init(&zs->zs_vdev_lock, USYNC_THREAD, NULL); (void) rwlock_init(&zs->zs_name_lock, USYNC_THREAD, NULL); @@ -3071,9 +3209,7 @@ ztest_run(char *pool) * Kick off a replacement of the disk we just obliterated. */ kernel_init(FREAD | FWRITE); - error = spa_open(pool, &spa, FTAG); - if (error) - fatal(0, "spa_open(%s) = %d", pool, error); + VERIFY(spa_open(pool, &spa, FTAG) == 0); ztest_replace_one_disk(spa, 0); if (zopt_verbose >= 5) show_pool_stats(spa); @@ -3106,9 +3242,13 @@ ztest_run(char *pool) /* * Open our pool. */ - error = spa_open(pool, &spa, FTAG); - if (error) - fatal(0, "spa_open() = %d", error); + VERIFY(spa_open(pool, &spa, FTAG) == 0); + + /* + * Create a thread to periodically resume suspended I/O. + */ + VERIFY(thr_create(0, 0, ztest_resume, spa, THR_BOUND, + &resume_tid) == 0); /* * Verify that we can safely inquire about about any object, @@ -3144,71 +3284,62 @@ ztest_run(char *pool) for (t = 0; t < zopt_threads; t++) { d = t % zopt_datasets; + + (void) strcpy(za[t].za_pool, pool); + za[t].za_os = za[d].za_os; + za[t].za_spa = spa; + za[t].za_zilog = za[d].za_zilog; + za[t].za_instance = t; + za[t].za_random = ztest_random(-1ULL); + za[t].za_start = za[0].za_start; + za[t].za_stop = za[0].za_stop; + za[t].za_kill = za[0].za_kill; + if (t < zopt_datasets) { ztest_replay_t zr; int test_future = FALSE; (void) rw_rdlock(&ztest_shared->zs_name_lock); (void) snprintf(name, 100, "%s/%s_%d", pool, pool, d); - error = dmu_objset_create(name, DMU_OST_OTHER, NULL, + error = dmu_objset_create(name, DMU_OST_OTHER, NULL, 0, ztest_create_cb, NULL); if (error == EEXIST) { test_future = TRUE; + } else if (error == ENOSPC) { + zs->zs_enospc_count++; + (void) rw_unlock(&ztest_shared->zs_name_lock); + break; } else if (error != 0) { - if (error == ENOSPC) { - zs->zs_enospc_count++; - (void) rw_unlock( - &ztest_shared->zs_name_lock); - break; - } fatal(0, "dmu_objset_create(%s) = %d", name, error); } error = dmu_objset_open(name, DMU_OST_OTHER, - DS_MODE_STANDARD, &za[d].za_os); + DS_MODE_USER, &za[d].za_os); if (error) fatal(0, "dmu_objset_open('%s') = %d", name, error); (void) rw_unlock(&ztest_shared->zs_name_lock); - if (test_future && ztest_shared->zs_txg > 0) - ztest_dmu_check_future_leak(za[d].za_os, - ztest_shared->zs_txg); + if (test_future) + ztest_dmu_check_future_leak(&za[t]); zr.zr_os = za[d].za_os; zil_replay(zr.zr_os, &zr, &zr.zr_assign, - ztest_replay_vector); + ztest_replay_vector, NULL); za[d].za_zilog = zil_open(za[d].za_os, NULL); } - za[t].za_pool = spa_strdup(pool); - za[t].za_os = za[d].za_os; - za[t].za_zilog = za[d].za_zilog; - za[t].za_instance = t; - za[t].za_random = ztest_random(-1ULL); - za[t].za_start = za[0].za_start; - za[t].za_stop = za[0].za_stop; - za[t].za_kill = za[0].za_kill; - error = thr_create(0, 0, ztest_thread, &za[t], THR_BOUND, - &za[t].za_thread); - if (error) - fatal(0, "can't create thread %d: error %d", - t, error); + VERIFY(thr_create(0, 0, ztest_thread, &za[t], THR_BOUND, + &za[t].za_thread) == 0); } - ztest_shared->zs_txg = 0; while (--t >= 0) { - error = thr_join(za[t].za_thread, NULL, NULL); - if (error) - fatal(0, "thr_join(%d) = %d", t, error); + VERIFY(thr_join(za[t].za_thread, NULL, NULL) == 0); if (za[t].za_th) traverse_fini(za[t].za_th); if (t < zopt_datasets) { zil_close(za[t].za_zilog); dmu_objset_close(za[t].za_os); } - spa_strfree(za[t].za_pool); } - umem_free(za, zopt_threads * sizeof (ztest_args_t)); - if (zopt_verbose >= 3) show_pool_stats(spa); @@ -3218,21 +3349,27 @@ ztest_run(char *pool) zs->zs_space = spa_get_space(spa); /* - * Did we have out-of-space errors? If so, destroy a random objset. + * If we had out-of-space errors, destroy a random objset. */ if (zs->zs_enospc_count != 0) { (void) rw_rdlock(&ztest_shared->zs_name_lock); - (void) snprintf(name, 100, "%s/%s_%d", pool, pool, - (int)ztest_random(zopt_datasets)); + d = (int)ztest_random(zopt_datasets); + (void) snprintf(name, 100, "%s/%s_%d", pool, pool, d); if (zopt_verbose >= 3) (void) printf("Destroying %s to free up space\n", name); - (void) dmu_objset_find(name, ztest_destroy_cb, NULL, + (void) dmu_objset_find(name, ztest_destroy_cb, &za[d], DS_FIND_SNAPSHOTS | DS_FIND_CHILDREN); (void) rw_unlock(&ztest_shared->zs_name_lock); } txg_wait_synced(spa_get_dsl(spa), 0); + umem_free(za, zopt_threads * sizeof (ztest_args_t)); + + /* Kill the resume thread */ + ztest_exiting = B_TRUE; + VERIFY(thr_join(resume_tid, NULL, NULL) == 0); + /* * Right before closing the pool, kick off a bunch of async I/O; * spa_close() should wait for it to complete. @@ -3288,8 +3425,9 @@ ztest_init(char *pool) */ (void) spa_destroy(pool); ztest_shared->zs_vdev_primaries = 0; - nvroot = make_vdev_root(zopt_vdev_size, zopt_raidz, zopt_mirrors, 1); - error = spa_create(pool, nvroot, NULL); + nvroot = make_vdev_root(NULL, NULL, zopt_vdev_size, 0, + 0, zopt_raidz, zopt_mirrors, 1); + error = spa_create(pool, nvroot, NULL, NULL, NULL); nvlist_free(nvroot); if (error) @@ -3320,7 +3458,7 @@ main(int argc, char **argv) (void) setvbuf(stdout, NULL, _IOLBF, 0); /* Override location of zpool.cache */ - spa_config_dir = "/tmp"; + spa_config_path = "/tmp/zpool.cache"; ztest_random_fd = open("/dev/urandom", O_RDONLY); diff --git a/cddl/contrib/opensolaris/head/assert.h b/cddl/contrib/opensolaris/head/assert.h index b088033..394820a 100644 --- a/cddl/contrib/opensolaris/head/assert.h +++ b/cddl/contrib/opensolaris/head/assert.h @@ -39,7 +39,7 @@ extern "C" { #if defined(__STDC__) #if __STDC_VERSION__ - 0 >= 199901L -extern void __assert_c99(const char *, const char *, int, const char *); +extern void __assert(const char *, const char *, int); #else extern void __assert(const char *, const char *, int); #endif /* __STDC_VERSION__ - 0 >= 199901L */ @@ -70,8 +70,7 @@ extern void _assert(); #if defined(__STDC__) #if __STDC_VERSION__ - 0 >= 199901L -#define assert(EX) (void)((EX) || \ - (__assert_c99(#EX, __FILE__, __LINE__, __func__), 0)) +#define assert(EX) (void)((EX) || (__assert(#EX, __FILE__, __LINE__), 0)) #else #define assert(EX) (void)((EX) || (__assert(#EX, __FILE__, __LINE__), 0)) #endif /* __STDC_VERSION__ - 0 >= 199901L */ diff --git a/cddl/contrib/opensolaris/head/libintl.h b/cddl/contrib/opensolaris/head/libintl.h index 94b4d03..e649668 100644 --- a/cddl/contrib/opensolaris/head/libintl.h +++ b/cddl/contrib/opensolaris/head/libintl.h @@ -2,9 +2,8 @@ * CDDL HEADER START * * The contents of this file are subject to the terms of the - * Common Development and Distribution License, Version 1.0 only - * (the "License"). You may not use this file except in compliance - * with the License. + * Common Development and Distribution License (the "License"). + * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. @@ -20,11 +19,10 @@ * CDDL HEADER END */ /* - * Copyright 2005 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ -/* Libintl is a library of advanced internationalization functions. */ #ifndef _LIBINTL_H #define _LIBINTL_H @@ -63,6 +61,9 @@ typedef long wchar_t; #define TEXTDOMAINMAX 256 +#define __GNU_GETTEXT_SUPPORTED_REVISION(m) \ + ((((m) == 0) || ((m) == 1)) ? 1 : -1) + #ifdef __STDC__ extern char *dcgettext(const char *, const char *, const int); extern char *dgettext(const char *, const char *); diff --git a/cddl/contrib/opensolaris/head/synch.h b/cddl/contrib/opensolaris/head/synch.h index 8d825d5..eab9de8 100644 --- a/cddl/contrib/opensolaris/head/synch.h +++ b/cddl/contrib/opensolaris/head/synch.h @@ -2,9 +2,8 @@ * CDDL HEADER START * * The contents of this file are subject to the terms of the - * Common Development and Distribution License, Version 1.0 only - * (the "License"). You may not use this file except in compliance - * with the License. + * Common Development and Distribution License (the "License"). + * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. @@ -19,8 +18,9 @@ * * CDDL HEADER END */ + /* - * Copyright 2003 Sun Microsystems, Inc. All rights reserved. + * Copyright 2007 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ @@ -81,12 +81,12 @@ typedef lwp_cond_t cond_t; * Because we have to deal with C++, we cannot redefine this one as that one. */ typedef struct _rwlock { - int32_t readers; /* -1 == writer else # of readers */ + int32_t readers; /* rwstate word */ uint16_t type; uint16_t magic; - mutex_t mutex; /* used to indicate ownership */ - cond_t readercv; /* unused */ - cond_t writercv; /* unused */ + mutex_t mutex; /* used with process-shared rwlocks */ + cond_t readercv; /* used only to indicate ownership */ + cond_t writercv; /* used only to indicate ownership */ } rwlock_t; #ifdef __STDC__ @@ -111,6 +111,7 @@ int cond_signal(cond_t *); int cond_broadcast(cond_t *); int mutex_init(mutex_t *, int, void *); int mutex_destroy(mutex_t *); +int mutex_consistent(mutex_t *); int mutex_lock(mutex_t *); int mutex_trylock(mutex_t *); int mutex_unlock(mutex_t *); @@ -152,6 +153,7 @@ int cond_signal(); int cond_broadcast(); int mutex_init(); int mutex_destroy(); +int mutex_consistent(); int mutex_lock(); int mutex_trylock(); int mutex_unlock(); diff --git a/cddl/contrib/opensolaris/head/thread.h b/cddl/contrib/opensolaris/head/thread.h index b9ed952..b374b8a 100644 --- a/cddl/contrib/opensolaris/head/thread.h +++ b/cddl/contrib/opensolaris/head/thread.h @@ -30,6 +30,7 @@ #pragma ident "%Z%%M% %I% %E% SMI" #include <pthread.h> +#include <pthread_np.h> #include <assert.h> /* @@ -52,6 +53,7 @@ typedef pthread_rwlock_t rwlock_t; #define mutex_lock(l) pthread_mutex_lock(l) #define mutex_trylock(l) pthread_mutex_trylock(l) #define mutex_unlock(l) pthread_mutex_unlock(l) +#define mutex_owned(l) pthread_mutex_isowned_np(l) #define rwlock_init(l,f,a) pthread_rwlock_init(l,NULL) #define rwlock_destroy(l) pthread_rwlock_destroy(l) #define rw_rdlock(l) pthread_rwlock_rdlock(l) diff --git a/cddl/contrib/opensolaris/lib/libnvpair/libnvpair.c b/cddl/contrib/opensolaris/lib/libnvpair/libnvpair.c index 58a47f1..89e01dd 100644 --- a/cddl/contrib/opensolaris/lib/libnvpair/libnvpair.c +++ b/cddl/contrib/opensolaris/lib/libnvpair/libnvpair.c @@ -2,9 +2,8 @@ * CDDL HEADER START * * The contents of this file are subject to the terms of the - * Common Development and Distribution License, Version 1.0 only - * (the "License"). You may not use this file except in compliance - * with the License. + * Common Development and Distribution License (the "License"). + * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. @@ -20,12 +19,13 @@ * CDDL HEADER END */ /* - * Copyright 2004 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ #pragma ident "%Z%%M% %I% %E% SMI" +#include <inttypes.h> #include <unistd.h> #include <strings.h> #include "libnvpair.h" @@ -137,6 +137,12 @@ nvlist_print_with_indent(FILE *fp, nvlist_t *nvl, int depth) (void) fprintf(fp, " 0x%llx", (u_longlong_t)val); break; } + case DATA_TYPE_DOUBLE: { + double val; + (void) nvpair_value_double(nvp, &val); + (void) fprintf(fp, " 0x%llf", val); + break; + } case DATA_TYPE_STRING: { char *val; (void) nvpair_value_string(nvp, &val); @@ -264,3 +270,348 @@ nvlist_print(FILE *fp, nvlist_t *nvl) { nvlist_print_with_indent(fp, nvl, 0); } + +/* + * Determine if string 'value' matches 'nvp' value. The 'value' string is + * converted, depending on the type of 'nvp', prior to match. For numeric + * types, a radix independent sscanf conversion of 'value' is used. If 'nvp' + * is an array type, 'ai' is the index into the array against which we are + * checking for match. If nvp is of DATA_TYPE_STRING*, the caller can pass + * in a regex_t compilation of value in 'value_regex' to trigger regular + * expression string match instead of simple strcmp(). + * + * Return 1 on match, 0 on no-match, and -1 on error. If the error is + * related to value syntax error and 'ep' is non-NULL, *ep will point into + * the 'value' string at the location where the error exists. + * + * NOTE: It may be possible to move the non-regex_t version of this into + * common code used by library/kernel/boot. + */ +int +nvpair_value_match_regex(nvpair_t *nvp, int ai, + char *value, regex_t *value_regex, char **ep) +{ + char *evalue; + uint_t a_len; + int sr; + + if (ep) + *ep = NULL; + + if ((nvp == NULL) || (value == NULL)) + return (-1); /* error fail match - invalid args */ + + /* make sure array and index combination make sense */ + if ((nvpair_type_is_array(nvp) && (ai < 0)) || + (!nvpair_type_is_array(nvp) && (ai >= 0))) + return (-1); /* error fail match - bad index */ + + /* non-string values should be single 'chunk' */ + if ((nvpair_type(nvp) != DATA_TYPE_STRING) && + (nvpair_type(nvp) != DATA_TYPE_STRING_ARRAY)) { + value += strspn(value, " \t"); + evalue = value + strcspn(value, " \t"); + if (*evalue) { + if (ep) + *ep = evalue; + return (-1); /* error fail match - syntax */ + } + } + + sr = EOF; + switch (nvpair_type(nvp)) { + case DATA_TYPE_STRING: { + char *val; + + /* check string value for match */ + if (nvpair_value_string(nvp, &val) == 0) { + if (value_regex) { + if (regexec(value_regex, val, + (size_t)0, NULL, 0) == 0) + return (1); /* match */ + } else { + if (strcmp(value, val) == 0) + return (1); /* match */ + } + } + break; + } + case DATA_TYPE_STRING_ARRAY: { + char **val_array; + + /* check indexed string value of array for match */ + if ((nvpair_value_string_array(nvp, &val_array, &a_len) == 0) && + (ai < a_len)) { + if (value_regex) { + if (regexec(value_regex, val_array[ai], + (size_t)0, NULL, 0) == 0) + return (1); + } else { + if (strcmp(value, val_array[ai]) == 0) + return (1); + } + } + break; + } + case DATA_TYPE_BYTE: { + uchar_t val, val_arg; + + /* scanf uchar_t from value and check for match */ + sr = sscanf(value, "%c", &val_arg); + if ((sr == 1) && (nvpair_value_byte(nvp, &val) == 0) && + (val == val_arg)) + return (1); + break; + } + case DATA_TYPE_BYTE_ARRAY: { + uchar_t *val_array, val_arg; + + + /* check indexed value of array for match */ + sr = sscanf(value, "%c", &val_arg); + if ((sr == 1) && + (nvpair_value_byte_array(nvp, &val_array, &a_len) == 0) && + (ai < a_len) && + (val_array[ai] == val_arg)) + return (1); + break; + } + case DATA_TYPE_INT8: { + int8_t val, val_arg; + + /* scanf int8_t from value and check for match */ + sr = sscanf(value, "%"SCNi8, &val_arg); + if ((sr == 1) && + (nvpair_value_int8(nvp, &val) == 0) && + (val == val_arg)) + return (1); + break; + } + case DATA_TYPE_INT8_ARRAY: { + int8_t *val_array, val_arg; + + /* check indexed value of array for match */ + sr = sscanf(value, "%"SCNi8, &val_arg); + if ((sr == 1) && + (nvpair_value_int8_array(nvp, &val_array, &a_len) == 0) && + (ai < a_len) && + (val_array[ai] == val_arg)) + return (1); + break; + } + case DATA_TYPE_UINT8: { + uint8_t val, val_arg; + + /* scanf uint8_t from value and check for match */ + sr = sscanf(value, "%"SCNi8, (int8_t *)&val_arg); + if ((sr == 1) && + (nvpair_value_uint8(nvp, &val) == 0) && + (val == val_arg)) + return (1); + break; + } + case DATA_TYPE_UINT8_ARRAY: { + uint8_t *val_array, val_arg; + + /* check indexed value of array for match */ + sr = sscanf(value, "%"SCNi8, (int8_t *)&val_arg); + if ((sr == 1) && + (nvpair_value_uint8_array(nvp, &val_array, &a_len) == 0) && + (ai < a_len) && + (val_array[ai] == val_arg)) + return (1); + break; + } + case DATA_TYPE_INT16: { + int16_t val, val_arg; + + /* scanf int16_t from value and check for match */ + sr = sscanf(value, "%"SCNi16, &val_arg); + if ((sr == 1) && + (nvpair_value_int16(nvp, &val) == 0) && + (val == val_arg)) + return (1); + break; + } + case DATA_TYPE_INT16_ARRAY: { + int16_t *val_array, val_arg; + + /* check indexed value of array for match */ + sr = sscanf(value, "%"SCNi16, &val_arg); + if ((sr == 1) && + (nvpair_value_int16_array(nvp, &val_array, &a_len) == 0) && + (ai < a_len) && + (val_array[ai] == val_arg)) + return (1); + break; + } + case DATA_TYPE_UINT16: { + uint16_t val, val_arg; + + /* scanf uint16_t from value and check for match */ + sr = sscanf(value, "%"SCNi16, (int16_t *)&val_arg); + if ((sr == 1) && + (nvpair_value_uint16(nvp, &val) == 0) && + (val == val_arg)) + return (1); + break; + } + case DATA_TYPE_UINT16_ARRAY: { + uint16_t *val_array, val_arg; + + /* check indexed value of array for match */ + sr = sscanf(value, "%"SCNi16, (int16_t *)&val_arg); + if ((sr == 1) && + (nvpair_value_uint16_array(nvp, &val_array, &a_len) == 0) && + (ai < a_len) && + (val_array[ai] == val_arg)) + return (1); + break; + } + case DATA_TYPE_INT32: { + int32_t val, val_arg; + + /* scanf int32_t from value and check for match */ + sr = sscanf(value, "%"SCNi32, &val_arg); + if ((sr == 1) && + (nvpair_value_int32(nvp, &val) == 0) && + (val == val_arg)) + return (1); + break; + } + case DATA_TYPE_INT32_ARRAY: { + int32_t *val_array, val_arg; + + /* check indexed value of array for match */ + sr = sscanf(value, "%"SCNi32, &val_arg); + if ((sr == 1) && + (nvpair_value_int32_array(nvp, &val_array, &a_len) == 0) && + (ai < a_len) && + (val_array[ai] == val_arg)) + return (1); + break; + } + case DATA_TYPE_UINT32: { + uint32_t val, val_arg; + + /* scanf uint32_t from value and check for match */ + sr = sscanf(value, "%"SCNi32, (int32_t *)&val_arg); + if ((sr == 1) && + (nvpair_value_uint32(nvp, &val) == 0) && + (val == val_arg)) + return (1); + break; + } + case DATA_TYPE_UINT32_ARRAY: { + uint32_t *val_array, val_arg; + + /* check indexed value of array for match */ + sr = sscanf(value, "%"SCNi32, (int32_t *)&val_arg); + if ((sr == 1) && + (nvpair_value_uint32_array(nvp, &val_array, &a_len) == 0) && + (ai < a_len) && + (val_array[ai] == val_arg)) + return (1); + break; + } + case DATA_TYPE_INT64: { + int64_t val, val_arg; + + /* scanf int64_t from value and check for match */ + sr = sscanf(value, "%"SCNi64, &val_arg); + if ((sr == 1) && + (nvpair_value_int64(nvp, &val) == 0) && + (val == val_arg)) + return (1); + break; + } + case DATA_TYPE_INT64_ARRAY: { + int64_t *val_array, val_arg; + + /* check indexed value of array for match */ + sr = sscanf(value, "%"SCNi64, &val_arg); + if ((sr == 1) && + (nvpair_value_int64_array(nvp, &val_array, &a_len) == 0) && + (ai < a_len) && + (val_array[ai] == val_arg)) + return (1); + break; + } + case DATA_TYPE_UINT64: { + uint64_t val_arg, val; + + /* scanf uint64_t from value and check for match */ + sr = sscanf(value, "%"SCNi64, (int64_t *)&val_arg); + if ((sr == 1) && + (nvpair_value_uint64(nvp, &val) == 0) && + (val == val_arg)) + return (1); + break; + } + case DATA_TYPE_UINT64_ARRAY: { + uint64_t *val_array, val_arg; + + /* check indexed value of array for match */ + sr = sscanf(value, "%"SCNi64, (int64_t *)&val_arg); + if ((sr == 1) && + (nvpair_value_uint64_array(nvp, &val_array, &a_len) == 0) && + (ai < a_len) && + (val_array[ai] == val_arg)) + return (1); + break; + } + case DATA_TYPE_BOOLEAN_VALUE: { + boolean_t val, val_arg; + + /* scanf boolean_t from value and check for match */ + sr = sscanf(value, "%"SCNi32, &val_arg); + if ((sr == 1) && + (nvpair_value_boolean_value(nvp, &val) == 0) && + (val == val_arg)) + return (1); + break; + } + case DATA_TYPE_BOOLEAN_ARRAY: { + boolean_t *val_array, val_arg; + + /* check indexed value of array for match */ + sr = sscanf(value, "%"SCNi32, &val_arg); + if ((sr == 1) && + (nvpair_value_boolean_array(nvp, + &val_array, &a_len) == 0) && + (ai < a_len) && + (val_array[ai] == val_arg)) + return (1); + break; + } + case DATA_TYPE_HRTIME: + case DATA_TYPE_NVLIST: + case DATA_TYPE_NVLIST_ARRAY: + case DATA_TYPE_BOOLEAN: + case DATA_TYPE_DOUBLE: + case DATA_TYPE_UNKNOWN: + default: + /* + * unknown/unsupported data type + */ + return (-1); /* error fail match */ + } + + /* + * check to see if sscanf failed conversion, return approximate + * pointer to problem + */ + if (sr != 1) { + if (ep) + *ep = value; + return (-1); /* error fail match - syntax */ + } + + return (0); /* fail match */ +} + +int +nvpair_value_match(nvpair_t *nvp, int ai, char *value, char **ep) +{ + return (nvpair_value_match_regex(nvp, ai, value, NULL, ep)); +} diff --git a/cddl/contrib/opensolaris/lib/libnvpair/libnvpair.h b/cddl/contrib/opensolaris/lib/libnvpair/libnvpair.h index d1d25ea..e655e0d 100644 --- a/cddl/contrib/opensolaris/lib/libnvpair/libnvpair.h +++ b/cddl/contrib/opensolaris/lib/libnvpair/libnvpair.h @@ -2,9 +2,8 @@ * CDDL HEADER START * * The contents of this file are subject to the terms of the - * Common Development and Distribution License, Version 1.0 only - * (the "License"). You may not use this file except in compliance - * with the License. + * Common Development and Distribution License (the "License"). + * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. @@ -20,7 +19,7 @@ * CDDL HEADER END */ /* - * Copyright 2005 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ @@ -32,12 +31,15 @@ #include <sys/nvpair.h> #include <stdlib.h> #include <stdio.h> +#include <regex.h> #ifdef __cplusplus extern "C" { #endif void nvlist_print(FILE *, nvlist_t *); +int nvpair_value_match(nvpair_t *, int, char *, char **); +int nvpair_value_match_regex(nvpair_t *, int, char *, regex_t *, char **); #ifdef __cplusplus } diff --git a/cddl/contrib/opensolaris/lib/libuutil/common/libuutil.h b/cddl/contrib/opensolaris/lib/libuutil/common/libuutil.h index fcd5e79..269687e 100644 --- a/cddl/contrib/opensolaris/lib/libuutil/common/libuutil.h +++ b/cddl/contrib/opensolaris/lib/libuutil/common/libuutil.h @@ -2,9 +2,8 @@ * CDDL HEADER START * * The contents of this file are subject to the terms of the - * Common Development and Distribution License, Version 1.0 only - * (the "License"). You may not use this file except in compliance - * with the License. + * Common Development and Distribution License (the "License"). + * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. @@ -20,15 +19,13 @@ * CDDL HEADER END */ /* - * Copyright 2005 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ #ifndef _LIBUUTIL_H #define _LIBUUTIL_H -#pragma ident "%Z%%M% %I% %E% SMI" - #include <solaris.h> #include <sys/types.h> #include <stdarg.h> @@ -149,6 +146,7 @@ extern int uu_open_tmp(const char *dir, uint_t uflags); /*PRINTFLIKE1*/ extern char *uu_msprintf(const char *format, ...); extern void *uu_zalloc(size_t); +extern char *uu_strdup(const char *); extern void uu_free(void *); /* diff --git a/cddl/contrib/opensolaris/lib/libuutil/common/libuutil_common.h b/cddl/contrib/opensolaris/lib/libuutil/common/libuutil_common.h index fc9cc7c..9ebaaed 100644 --- a/cddl/contrib/opensolaris/lib/libuutil/common/libuutil_common.h +++ b/cddl/contrib/opensolaris/lib/libuutil/common/libuutil_common.h @@ -2,9 +2,8 @@ * CDDL HEADER START * * The contents of this file are subject to the terms of the - * Common Development and Distribution License, Version 1.0 only - * (the "License"). You may not use this file except in compliance - * with the License. + * Common Development and Distribution License (the "License"). + * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. @@ -21,7 +20,7 @@ */ /* - * Copyright 2006 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ @@ -30,16 +29,6 @@ #pragma ident "%Z%%M% %I% %E% SMI" -#include <solaris.h> - -/* - * We don't bind to the internal libc interfaces if this is a - * native build. - */ -#ifndef NATIVE_BUILD -#include "c_synonyms.h" -#endif - #include <libuutil.h> #include <libuutil_impl.h> diff --git a/cddl/contrib/opensolaris/lib/libuutil/common/uu_alloc.c b/cddl/contrib/opensolaris/lib/libuutil/common/uu_alloc.c index 7cdbf01..05d8622 100644 --- a/cddl/contrib/opensolaris/lib/libuutil/common/uu_alloc.c +++ b/cddl/contrib/opensolaris/lib/libuutil/common/uu_alloc.c @@ -2,9 +2,8 @@ * CDDL HEADER START * * The contents of this file are subject to the terms of the - * Common Development and Distribution License, Version 1.0 only - * (the "License"). You may not use this file except in compliance - * with the License. + * Common Development and Distribution License (the "License"). + * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. @@ -20,12 +19,10 @@ * CDDL HEADER END */ /* - * Copyright 2004 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ -#pragma ident "%Z%%M% %I% %E% SMI" - #include "libuutil_common.h" #include <stdarg.h> @@ -55,6 +52,22 @@ uu_free(void *p) } char * +uu_strdup(const char *str) +{ + char *buf = NULL; + + if (str != NULL) { + size_t sz; + + sz = strlen(str) + 1; + buf = uu_zalloc(sz); + if (buf != NULL) + (void) memcpy(buf, str, sz); + } + return (buf); +} + +char * uu_msprintf(const char *format, ...) { va_list args; diff --git a/cddl/contrib/opensolaris/lib/libuutil/common/uu_avl.c b/cddl/contrib/opensolaris/lib/libuutil/common/uu_avl.c index 93feea9..308e920 100644 --- a/cddl/contrib/opensolaris/lib/libuutil/common/uu_avl.c +++ b/cddl/contrib/opensolaris/lib/libuutil/common/uu_avl.c @@ -19,7 +19,7 @@ * CDDL HEADER END */ /* - * Copyright 2006 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ @@ -120,7 +120,8 @@ uu_avl_pool_destroy(uu_avl_pool_t *pp) UU_PTR_ENCODE(&pp->uap_null_avl)) { uu_panic("uu_avl_pool_destroy: Pool \"%.*s\" (%p) has " "outstanding avls, or is corrupt.\n", - sizeof (pp->uap_name), pp->uap_name, pp); + (int)sizeof (pp->uap_name), pp->uap_name, + (void *)pp); } } (void) pthread_mutex_lock(&uu_apool_list_lock); @@ -142,14 +143,14 @@ uu_avl_node_init(void *base, uu_avl_node_t *np, uu_avl_pool_t *pp) if (offset + sizeof (*np) > pp->uap_objsize) { uu_panic("uu_avl_node_init(%p, %p, %p (\"%s\")): " "offset %ld doesn't fit in object (size %ld)\n", - base, np, pp, pp->uap_name, offset, - pp->uap_objsize); + base, (void *)np, (void *)pp, pp->uap_name, + (long)offset, (long)pp->uap_objsize); } if (offset != pp->uap_nodeoffset) { uu_panic("uu_avl_node_init(%p, %p, %p (\"%s\")): " "offset %ld doesn't match pool's offset (%ld)\n", - base, np, pp, pp->uap_name, offset, - pp->uap_objsize); + base, (void *)np, (void *)pp, pp->uap_name, + (long)offset, (long)pp->uap_objsize); } } @@ -166,12 +167,12 @@ uu_avl_node_fini(void *base, uu_avl_node_t *np, uu_avl_pool_t *pp) if (na[0] == DEAD_MARKER && na[1] == DEAD_MARKER) { uu_panic("uu_avl_node_fini(%p, %p, %p (\"%s\")): " "node already finied\n", - base, np, pp, pp->uap_name); + base, (void *)np, (void *)pp, pp->uap_name); } if (na[0] != POOL_TO_MARKER(pp) || na[1] != 0) { uu_panic("uu_avl_node_fini(%p, %p, %p (\"%s\")): " "node corrupt, in tree, or in different pool\n", - base, np, pp, pp->uap_name); + base, (void *)np, (void *)pp, pp->uap_name); } } @@ -251,12 +252,13 @@ uu_avl_destroy(uu_avl_t *ap) if (ap->ua_debug) { if (avl_numnodes(&ap->ua_tree) != 0) { - uu_panic("uu_avl_destroy(%p): tree not empty\n", ap); + uu_panic("uu_avl_destroy(%p): tree not empty\n", + (void *)ap); } if (ap->ua_null_walk.uaw_next != &ap->ua_null_walk || ap->ua_null_walk.uaw_prev != &ap->ua_null_walk) { uu_panic("uu_avl_destroy(%p): outstanding walkers\n", - ap); + (void *)ap); } } (void) pthread_mutex_lock(&pp->uap_lock); @@ -441,7 +443,7 @@ uu_avl_remove(uu_avl_t *ap, void *elem) (void) _avl_walk_advance(wp, ap); } else if (wp->uaw_next_result != NULL) { uu_panic("uu_avl_remove(%p, %p): active non-robust " - "walker\n", ap, elem); + "walker\n", (void *)ap, elem); } } @@ -497,19 +499,19 @@ uu_avl_insert(uu_avl_t *ap, void *elem, uu_avl_index_t idx) if (na[1] != 0) uu_panic("uu_avl_insert(%p, %p, %p): node already " "in tree, or corrupt\n", - ap, elem, idx); + (void *)ap, elem, (void *)idx); if (na[0] == 0) uu_panic("uu_avl_insert(%p, %p, %p): node not " "initialized\n", - ap, elem, idx); + (void *)ap, elem, (void *)idx); if (na[0] != POOL_TO_MARKER(pp)) uu_panic("uu_avl_insert(%p, %p, %p): node from " "other pool, or corrupt\n", - ap, elem, idx); + (void *)ap, elem, (void *)idx); if (!INDEX_VALID(ap, idx)) uu_panic("uu_avl_insert(%p, %p, %p): %s\n", - ap, elem, idx, + (void *)ap, elem, (void *)idx, INDEX_CHECK(idx)? "outdated index" : "invalid index"); @@ -526,8 +528,8 @@ uu_avl_nearest_next(uu_avl_t *ap, uu_avl_index_t idx) { if (ap->ua_debug && !INDEX_VALID(ap, idx)) uu_panic("uu_avl_nearest_next(%p, %p): %s\n", - ap, idx, INDEX_CHECK(idx)? "outdated index" : - "invalid index"); + (void *)ap, (void *)idx, INDEX_CHECK(idx)? + "outdated index" : "invalid index"); return (avl_nearest(&ap->ua_tree, INDEX_DECODE(idx), AVL_AFTER)); } @@ -536,8 +538,8 @@ uu_avl_nearest_prev(uu_avl_t *ap, uu_avl_index_t idx) { if (ap->ua_debug && !INDEX_VALID(ap, idx)) uu_panic("uu_avl_nearest_prev(%p, %p): %s\n", - ap, idx, INDEX_CHECK(idx)? "outdated index" : - "invalid index"); + (void *)ap, (void *)idx, INDEX_CHECK(idx)? + "outdated index" : "invalid index"); return (avl_nearest(&ap->ua_tree, INDEX_DECODE(idx), AVL_BEFORE)); } diff --git a/cddl/contrib/opensolaris/lib/libuutil/common/uu_dprintf.c b/cddl/contrib/opensolaris/lib/libuutil/common/uu_dprintf.c index 5b990a5..528c3e7 100644 --- a/cddl/contrib/opensolaris/lib/libuutil/common/uu_dprintf.c +++ b/cddl/contrib/opensolaris/lib/libuutil/common/uu_dprintf.c @@ -33,7 +33,7 @@ #include <stdarg.h> #include <stdio.h> #include <stdlib.h> -#include <strings.h> +#include <string.h> #define FACILITY_FMT "%s (%s): " diff --git a/cddl/contrib/opensolaris/lib/libuutil/common/uu_list.c b/cddl/contrib/opensolaris/lib/libuutil/common/uu_list.c index d9dc86f..35c7ba8 100644 --- a/cddl/contrib/opensolaris/lib/libuutil/common/uu_list.c +++ b/cddl/contrib/opensolaris/lib/libuutil/common/uu_list.c @@ -2,9 +2,8 @@ * CDDL HEADER START * * The contents of this file are subject to the terms of the - * Common Development and Distribution License, Version 1.0 only - * (the "License"). You may not use this file except in compliance - * with the License. + * Common Development and Distribution License (the "License"). + * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. @@ -20,7 +19,7 @@ * CDDL HEADER END */ /* - * Copyright 2005 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ @@ -117,7 +116,8 @@ uu_list_pool_destroy(uu_list_pool_t *pp) UU_PTR_ENCODE(&pp->ulp_null_list)) { uu_panic("uu_list_pool_destroy: Pool \"%.*s\" (%p) has " "outstanding lists, or is corrupt.\n", - sizeof (pp->ulp_name), pp->ulp_name, pp); + (int)sizeof (pp->ulp_name), pp->ulp_name, + (void *)pp); } } (void) pthread_mutex_lock(&uu_lpool_list_lock); @@ -139,14 +139,14 @@ uu_list_node_init(void *base, uu_list_node_t *np_arg, uu_list_pool_t *pp) if (offset + sizeof (*np) > pp->ulp_objsize) { uu_panic("uu_list_node_init(%p, %p, %p (\"%s\")): " "offset %ld doesn't fit in object (size %ld)\n", - base, np, pp, pp->ulp_name, offset, - pp->ulp_objsize); + base, (void *)np, (void *)pp, pp->ulp_name, + (long)offset, (long)pp->ulp_objsize); } if (offset != pp->ulp_nodeoffset) { uu_panic("uu_list_node_init(%p, %p, %p (\"%s\")): " "offset %ld doesn't match pool's offset (%ld)\n", - base, np, pp, pp->ulp_name, offset, - pp->ulp_objsize); + base, (void *)np, (void *)pp, pp->ulp_name, + (long)offset, (long)pp->ulp_objsize); } } np->uln_next = POOL_TO_MARKER(pp); @@ -163,13 +163,13 @@ uu_list_node_fini(void *base, uu_list_node_t *np_arg, uu_list_pool_t *pp) np->uln_prev == NULL) { uu_panic("uu_list_node_fini(%p, %p, %p (\"%s\")): " "node already finied\n", - base, np_arg, pp, pp->ulp_name); + base, (void *)np_arg, (void *)pp, pp->ulp_name); } if (np->uln_next != POOL_TO_MARKER(pp) || np->uln_prev != NULL) { uu_panic("uu_list_node_fini(%p, %p, %p (\"%s\")): " "node corrupt or on list\n", - base, np_arg, pp, pp->ulp_name); + base, (void *)np_arg, (void *)pp, pp->ulp_name); } } np->uln_next = NULL; @@ -190,7 +190,7 @@ uu_list_create(uu_list_pool_t *pp, void *parent, uint32_t flags) if (pp->ulp_debug) uu_panic("uu_list_create(%p, ...): requested " "UU_LIST_SORTED, but pool has no comparison func\n", - pp); + (void *)pp); uu_set_error(UU_ERROR_NOT_SUPPORTED); return (NULL); } @@ -236,16 +236,16 @@ uu_list_destroy(uu_list_t *lp) if (lp->ul_null_node.uln_next != &lp->ul_null_node || lp->ul_null_node.uln_prev != &lp->ul_null_node) { uu_panic("uu_list_destroy(%p): list not empty\n", - lp); + (void *)lp); } if (lp->ul_numnodes != 0) { uu_panic("uu_list_destroy(%p): numnodes is nonzero, " - "but list is empty\n", lp); + "but list is empty\n", (void *)lp); } if (lp->ul_null_walk.ulw_next != &lp->ul_null_walk || lp->ul_null_walk.ulw_prev != &lp->ul_null_walk) { uu_panic("uu_list_destroy(%p): outstanding walkers\n", - lp); + (void *)lp); } } @@ -266,13 +266,14 @@ list_insert(uu_list_t *lp, uu_list_node_impl_t *np, uu_list_node_impl_t *prev, if (lp->ul_debug) { if (next->uln_prev != prev || prev->uln_next != next) uu_panic("insert(%p): internal error: %p and %p not " - "neighbors\n", lp, next, prev); + "neighbors\n", (void *)lp, (void *)next, + (void *)prev); if (np->uln_next != POOL_TO_MARKER(lp->ul_pool) || np->uln_prev != NULL) { uu_panic("insert(%p): elem %p node %p corrupt, " "not initialized, or already in a list.\n", - lp, NODE_TO_ELEM(lp, np), np); + (void *)lp, NODE_TO_ELEM(lp, np), (void *)np); } /* * invalidate outstanding uu_list_index_ts. @@ -299,12 +300,12 @@ uu_list_insert(uu_list_t *lp, void *elem, uu_list_index_t idx) if (lp->ul_debug) { if (!INDEX_VALID(lp, idx)) uu_panic("uu_list_insert(%p, %p, %p): %s\n", - lp, elem, idx, + (void *)lp, elem, (void *)idx, INDEX_CHECK(idx)? "outdated index" : "invalid index"); if (np->uln_prev == NULL) uu_panic("uu_list_insert(%p, %p, %p): out-of-date " - "index\n", lp, elem, idx); + "index\n", (void *)lp, elem, (void *)idx); } list_insert(lp, ELEM_TO_NODE(lp, elem), np->uln_prev, np); @@ -354,11 +355,12 @@ uu_list_nearest_next(uu_list_t *lp, uu_list_index_t idx) if (lp->ul_debug) { if (!INDEX_VALID(lp, idx)) uu_panic("uu_list_nearest_next(%p, %p): %s\n", - lp, idx, INDEX_CHECK(idx)? "outdated index" : + (void *)lp, (void *)idx, + INDEX_CHECK(idx)? "outdated index" : "invalid index"); if (np->uln_prev == NULL) uu_panic("uu_list_nearest_next(%p, %p): out-of-date " - "index\n", lp, idx); + "index\n", (void *)lp, (void *)idx); } if (np == &lp->ul_null_node) @@ -378,11 +380,11 @@ uu_list_nearest_prev(uu_list_t *lp, uu_list_index_t idx) if (lp->ul_debug) { if (!INDEX_VALID(lp, idx)) uu_panic("uu_list_nearest_prev(%p, %p): %s\n", - lp, idx, INDEX_CHECK(idx)? "outdated index" : - "invalid index"); + (void *)lp, (void *)idx, INDEX_CHECK(idx)? + "outdated index" : "invalid index"); if (np->uln_prev == NULL) uu_panic("uu_list_nearest_prev(%p, %p): out-of-date " - "index\n", lp, idx); + "index\n", (void *)lp, (void *)idx); } if ((np = np->uln_prev) == &lp->ul_null_node) @@ -409,6 +411,11 @@ list_walk_init(uu_list_walk_t *wp, uu_list_t *lp, uint32_t flags) wp->ulw_next_result = lp->ul_null_node.uln_prev; if (lp->ul_debug || robust) { + /* + * Add this walker to the list's list of walkers so + * uu_list_remove() can advance us if somebody tries to + * remove ulw_next_result. + */ wp->ulw_next = next = &lp->ul_null_walk; wp->ulw_prev = prev = next->ulw_prev; next->ulw_prev = wp; @@ -538,7 +545,7 @@ uu_list_remove(uu_list_t *lp, void *elem) if (lp->ul_debug) { if (np->uln_prev == NULL) uu_panic("uu_list_remove(%p, %p): elem not on list\n", - lp, elem); + (void *)lp, elem); /* * invalidate outstanding uu_list_index_ts. */ @@ -556,7 +563,7 @@ uu_list_remove(uu_list_t *lp, void *elem) (void) list_walk_advance(wp, lp); } else if (wp->ulw_next_result != NULL) { uu_panic("uu_list_remove(%p, %p): active non-robust " - "walker\n", lp, elem); + "walker\n", (void *)lp, elem); } } @@ -578,8 +585,8 @@ uu_list_teardown(uu_list_t *lp, void **cookie) * XXX: disable list modification until list is empty */ if (lp->ul_debug && *cookie != NULL) - uu_panic("uu_list_teardown(%p, %p): unexpected cookie\n", lp, - cookie); + uu_panic("uu_list_teardown(%p, %p): unexpected cookie\n", + (void *)lp, (void *)cookie); ep = uu_list_first(lp); if (ep) @@ -599,12 +606,12 @@ uu_list_insert_before(uu_list_t *lp, void *target, void *elem) if (np->uln_prev == NULL) uu_panic("uu_list_insert_before(%p, %p, %p): %p is " "not currently on a list\n", - lp, target, elem, target); + (void *)lp, target, elem, target); } if (lp->ul_sorted) { if (lp->ul_debug) uu_panic("uu_list_insert_before(%p, ...): list is " - "UU_LIST_SORTED\n", lp); + "UU_LIST_SORTED\n", (void *)lp); uu_set_error(UU_ERROR_NOT_SUPPORTED); return (-1); } @@ -625,12 +632,12 @@ uu_list_insert_after(uu_list_t *lp, void *target, void *elem) if (np->uln_prev == NULL) uu_panic("uu_list_insert_after(%p, %p, %p): %p is " "not currently on a list\n", - lp, target, elem, target); + (void *)lp, target, elem, target); } if (lp->ul_sorted) { if (lp->ul_debug) uu_panic("uu_list_insert_after(%p, ...): list is " - "UU_LIST_SORTED\n", lp); + "UU_LIST_SORTED\n", (void *)lp); uu_set_error(UU_ERROR_NOT_SUPPORTED); return (-1); } diff --git a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs.h b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs.h index 232324e..71aab1c 100644 --- a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs.h +++ b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs.h @@ -20,21 +20,20 @@ */ /* - * Copyright 2007 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ #ifndef _LIBZFS_H #define _LIBZFS_H -#pragma ident "%Z%%M% %I% %E% SMI" - #include <assert.h> #include <libnvpair.h> #include <sys/param.h> #include <sys/types.h> #include <sys/varargs.h> #include <sys/fs/zfs.h> +#include <sys/avl.h> #include <sys/zfs_ioctl.h> #ifdef __cplusplus @@ -47,6 +46,7 @@ extern "C" { #define ZFS_MAXNAMELEN MAXNAMELEN #define ZPOOL_MAXNAMELEN MAXNAMELEN #define ZFS_MAXPROPLEN MAXPATHLEN +#define ZPOOL_MAXPROPLEN MAXPATHLEN /* * libzfs errors @@ -99,10 +99,62 @@ enum { EZFS_POOL_NOTSUP, /* ops not supported for this type of pool */ EZFS_POOL_INVALARG, /* invalid argument for this pool operation */ EZFS_NAMETOOLONG, /* dataset name is too long */ + EZFS_OPENFAILED, /* open of device failed */ + EZFS_NOCAP, /* couldn't get capacity */ + EZFS_LABELFAILED, /* write of label failed */ + EZFS_ISCSISVCUNAVAIL, /* iscsi service unavailable */ + EZFS_BADWHO, /* invalid permission who */ + EZFS_BADPERM, /* invalid permission */ + EZFS_BADPERMSET, /* invalid permission set name */ + EZFS_NODELEGATION, /* delegated administration is disabled */ + EZFS_PERMRDONLY, /* pemissions are readonly */ + EZFS_UNSHARESMBFAILED, /* failed to unshare over smb */ + EZFS_SHARESMBFAILED, /* failed to share over smb */ + EZFS_BADCACHE, /* bad cache file */ + EZFS_ISL2CACHE, /* device is for the level 2 ARC */ + EZFS_VDEVNOTSUP, /* unsupported vdev type */ + EZFS_NOTSUP, /* ops not supported on this dataset */ + EZFS_ACTIVE_SPARE, /* pool has active shared spare devices */ EZFS_UNKNOWN }; /* + * The following data structures are all part + * of the zfs_allow_t data structure which is + * used for printing 'allow' permissions. + * It is a linked list of zfs_allow_t's which + * then contain avl tree's for user/group/sets/... + * and each one of the entries in those trees have + * avl tree's for the permissions they belong to and + * whether they are local,descendent or local+descendent + * permissions. The AVL trees are used primarily for + * sorting purposes, but also so that we can quickly find + * a given user and or permission. + */ +typedef struct zfs_perm_node { + avl_node_t z_node; + char z_pname[MAXPATHLEN]; +} zfs_perm_node_t; + +typedef struct zfs_allow_node { + avl_node_t z_node; + char z_key[MAXPATHLEN]; /* name, such as joe */ + avl_tree_t z_localdescend; /* local+descendent perms */ + avl_tree_t z_local; /* local permissions */ + avl_tree_t z_descend; /* descendent permissions */ +} zfs_allow_node_t; + +typedef struct zfs_allow { + struct zfs_allow *z_next; + char z_setpoint[MAXPATHLEN]; + avl_tree_t z_sets; + avl_tree_t z_crperms; + avl_tree_t z_user; + avl_tree_t z_group; + avl_tree_t z_everyone; +} zfs_allow_t; + +/* * Basic handle types */ typedef struct zfs_handle zfs_handle_t; @@ -131,12 +183,9 @@ extern zpool_handle_t *zpool_open(libzfs_handle_t *, const char *); extern zpool_handle_t *zpool_open_canfail(libzfs_handle_t *, const char *); extern void zpool_close(zpool_handle_t *); extern const char *zpool_get_name(zpool_handle_t *); -extern uint64_t zpool_get_guid(zpool_handle_t *); -extern uint64_t zpool_get_space_used(zpool_handle_t *); -extern uint64_t zpool_get_space_total(zpool_handle_t *); -extern int zpool_get_root(zpool_handle_t *, char *, size_t); extern int zpool_get_state(zpool_handle_t *); -extern uint64_t zpool_get_version(zpool_handle_t *); +extern char *zpool_state_to_name(vdev_state_t, vdev_aux_t); +extern void zpool_free_handles(libzfs_handle_t *); /* * Iterate over all active pools in the system. @@ -148,7 +197,7 @@ extern int zpool_iter(libzfs_handle_t *, zpool_iter_f, void *); * Functions to create and destroy pools */ extern int zpool_create(libzfs_handle_t *, const char *, nvlist_t *, - const char *); + nvlist_t *, nvlist_t *); extern int zpool_destroy(zpool_handle_t *); extern int zpool_add(zpool_handle_t *, nvlist_t *); @@ -156,22 +205,33 @@ extern int zpool_add(zpool_handle_t *, nvlist_t *); * Functions to manipulate pool and vdev state */ extern int zpool_scrub(zpool_handle_t *, pool_scrub_type_t); +extern int zpool_clear(zpool_handle_t *, const char *); -extern int zpool_vdev_online(zpool_handle_t *, const char *); -extern int zpool_vdev_offline(zpool_handle_t *, const char *, int); -extern int zpool_vdev_attach(zpool_handle_t *, const char *, const char *, - nvlist_t *, int); +extern int zpool_vdev_online(zpool_handle_t *, const char *, int, + vdev_state_t *); +extern int zpool_vdev_offline(zpool_handle_t *, const char *, boolean_t); +extern int zpool_vdev_attach(zpool_handle_t *, const char *, + const char *, nvlist_t *, int); extern int zpool_vdev_detach(zpool_handle_t *, const char *); extern int zpool_vdev_remove(zpool_handle_t *, const char *); -extern int zpool_clear(zpool_handle_t *, const char *); -extern nvlist_t *zpool_find_vdev(zpool_handle_t *, const char *, boolean_t *); + +extern int zpool_vdev_fault(zpool_handle_t *, uint64_t); +extern int zpool_vdev_degrade(zpool_handle_t *, uint64_t); +extern int zpool_vdev_clear(zpool_handle_t *, uint64_t); + +extern nvlist_t *zpool_find_vdev(zpool_handle_t *, const char *, boolean_t *, + boolean_t *, boolean_t *); +extern int zpool_label_disk(libzfs_handle_t *, zpool_handle_t *, char *); /* * Functions to manage pool properties */ extern int zpool_set_prop(zpool_handle_t *, const char *, const char *); -extern int zpool_get_prop(zpool_handle_t *, zfs_prop_t, char *, - size_t proplen, zfs_source_t *); +extern int zpool_get_prop(zpool_handle_t *, zpool_prop_t, char *, + size_t proplen, zprop_source_t *); +extern uint64_t zpool_get_prop_int(zpool_handle_t *, zpool_prop_t, + zprop_source_t *); + extern const char *zpool_prop_to_name(zpool_prop_t); extern const char *zpool_prop_values(zpool_prop_t); @@ -194,6 +254,11 @@ typedef enum { ZPOOL_STATUS_FAILING_DEV, /* device experiencing errors */ ZPOOL_STATUS_VERSION_NEWER, /* newer on-disk version */ ZPOOL_STATUS_HOSTID_MISMATCH, /* last accessed by another system */ + ZPOOL_STATUS_IO_FAILURE_WAIT, /* failed I/O, failmode 'wait' */ + ZPOOL_STATUS_IO_FAILURE_CONTINUE, /* failed I/O, failmode 'continue' */ + ZPOOL_STATUS_FAULTED_DEV_R, /* faulted device with replicas */ + ZPOOL_STATUS_FAULTED_DEV_NR, /* faulted device with no replicas */ + ZPOOL_STATUS_BAD_LOG, /* cannot read log chain(s) */ /* * The following are not faults per se, but still an error possibly @@ -223,26 +288,39 @@ extern int zpool_get_errlog(zpool_handle_t *, nvlist_t **); /* * Import and export functions */ -extern int zpool_export(zpool_handle_t *); +extern int zpool_export(zpool_handle_t *, boolean_t); extern int zpool_import(libzfs_handle_t *, nvlist_t *, const char *, - const char *); + char *altroot); +extern int zpool_import_props(libzfs_handle_t *, nvlist_t *, const char *, + nvlist_t *, boolean_t); /* * Search for pools to import */ extern nvlist_t *zpool_find_import(libzfs_handle_t *, int, char **); +extern nvlist_t *zpool_find_import_cached(libzfs_handle_t *, const char *, + char *, uint64_t); +extern nvlist_t *zpool_find_import_byname(libzfs_handle_t *, int, char **, + char *); +extern nvlist_t *zpool_find_import_byguid(libzfs_handle_t *, int, char **, + uint64_t); +extern nvlist_t *zpool_find_import_activeok(libzfs_handle_t *, int, char **); /* * Miscellaneous pool functions */ +struct zfs_cmd; + extern char *zpool_vdev_name(libzfs_handle_t *, zpool_handle_t *, nvlist_t *); -extern int zpool_upgrade(zpool_handle_t *); +extern int zpool_upgrade(zpool_handle_t *, uint64_t); extern int zpool_get_history(zpool_handle_t *, nvlist_t **); -extern void zpool_log_history(libzfs_handle_t *, int, char **, const char *, - boolean_t, boolean_t); +extern void zpool_set_history_str(const char *subcommand, int argc, + char **argv, char *history_str); +extern int zpool_stage_history(libzfs_handle_t *, const char *); extern void zpool_obj_to_path(zpool_handle_t *, uint64_t, uint64_t, char *, size_t len); - +extern int zfs_ioctl(libzfs_handle_t *, int, struct zfs_cmd *); +extern int zpool_get_physpath(zpool_handle_t *, char *); /* * Basic handle manipulations. These functions do not create or destroy the * underlying datasets, only the references to them. @@ -251,65 +329,84 @@ extern zfs_handle_t *zfs_open(libzfs_handle_t *, const char *, int); extern void zfs_close(zfs_handle_t *); extern zfs_type_t zfs_get_type(const zfs_handle_t *); extern const char *zfs_get_name(const zfs_handle_t *); +extern zpool_handle_t *zfs_get_pool_handle(const zfs_handle_t *); /* * Property management functions. Some functions are shared with the kernel, * and are found in sys/fs/zfs.h. */ + +/* + * zfs dataset property management + */ +extern const char *zfs_prop_default_string(zfs_prop_t); +extern uint64_t zfs_prop_default_numeric(zfs_prop_t); +extern const char *zfs_prop_column_name(zfs_prop_t); +extern boolean_t zfs_prop_align_right(zfs_prop_t); + +extern nvlist_t *zfs_valid_proplist(libzfs_handle_t *, zfs_type_t, + nvlist_t *, uint64_t, zfs_handle_t *, const char *); + extern const char *zfs_prop_to_name(zfs_prop_t); extern int zfs_prop_set(zfs_handle_t *, const char *, const char *); extern int zfs_prop_get(zfs_handle_t *, zfs_prop_t, char *, size_t, - zfs_source_t *, char *, size_t, boolean_t); + zprop_source_t *, char *, size_t, boolean_t); extern int zfs_prop_get_numeric(zfs_handle_t *, zfs_prop_t, uint64_t *, - zfs_source_t *, char *, size_t); + zprop_source_t *, char *, size_t); extern uint64_t zfs_prop_get_int(zfs_handle_t *, zfs_prop_t); -extern const char *zfs_prop_get_string(zfs_handle_t *, zfs_prop_t); extern int zfs_prop_inherit(zfs_handle_t *, const char *); extern const char *zfs_prop_values(zfs_prop_t); -extern int zfs_prop_valid_for_type(zfs_prop_t, int); -extern const char *zfs_prop_default_string(zfs_prop_t prop); -extern uint64_t zfs_prop_default_numeric(zfs_prop_t); extern int zfs_prop_is_string(zfs_prop_t prop); -extern const char *zfs_prop_column_name(zfs_prop_t); -extern boolean_t zfs_prop_align_right(zfs_prop_t); -extern void nicebool(int value, char *buf, size_t buflen); +extern nvlist_t *zfs_get_user_props(zfs_handle_t *); -typedef struct zfs_proplist { - zfs_prop_t pl_prop; +typedef struct zprop_list { + int pl_prop; char *pl_user_prop; - struct zfs_proplist *pl_next; + struct zprop_list *pl_next; boolean_t pl_all; size_t pl_width; boolean_t pl_fixed; -} zfs_proplist_t; +} zprop_list_t; -typedef zfs_proplist_t zpool_proplist_t; - -extern int zfs_get_proplist(libzfs_handle_t *, char *, zfs_proplist_t **); -extern int zpool_get_proplist(libzfs_handle_t *, char *, zpool_proplist_t **); -extern int zfs_expand_proplist(zfs_handle_t *, zfs_proplist_t **); -extern int zpool_expand_proplist(zpool_handle_t *, zpool_proplist_t **); -extern void zfs_free_proplist(zfs_proplist_t *); -extern nvlist_t *zfs_get_user_props(zfs_handle_t *); +extern int zfs_expand_proplist(zfs_handle_t *, zprop_list_t **); #define ZFS_MOUNTPOINT_NONE "none" #define ZFS_MOUNTPOINT_LEGACY "legacy" /* - * Functions for printing properties from zfs/zpool + * zpool property management + */ +extern int zpool_expand_proplist(zpool_handle_t *, zprop_list_t **); +extern const char *zpool_prop_default_string(zpool_prop_t); +extern uint64_t zpool_prop_default_numeric(zpool_prop_t); +extern const char *zpool_prop_column_name(zpool_prop_t); +extern boolean_t zpool_prop_align_right(zpool_prop_t); + +/* + * Functions shared by zfs and zpool property management. */ -typedef struct libzfs_get_cbdata { +extern int zprop_iter(zprop_func func, void *cb, boolean_t show_all, + boolean_t ordered, zfs_type_t type); +extern int zprop_get_list(libzfs_handle_t *, char *, zprop_list_t **, + zfs_type_t); +extern void zprop_free_list(zprop_list_t *); + +/* + * Functions for printing zfs or zpool properties + */ +typedef struct zprop_get_cbdata { int cb_sources; int cb_columns[4]; int cb_colwidths[5]; boolean_t cb_scripted; boolean_t cb_literal; boolean_t cb_first; - zfs_proplist_t *cb_proplist; -} libzfs_get_cbdata_t; + zprop_list_t *cb_proplist; + zfs_type_t cb_type; +} zprop_get_cbdata_t; -void libzfs_print_one_property(const char *, libzfs_get_cbdata_t *, - const char *, const char *, zfs_source_t, const char *); +void zprop_print_one_property(const char *, zprop_get_cbdata_t *, + const char *, const char *, zprop_source_t, const char *); #define GET_COL_NAME 1 #define GET_COL_PROPERTY 2 @@ -331,26 +428,61 @@ extern int zfs_iter_snapshots(zfs_handle_t *, zfs_iter_f, void *); */ extern int zfs_create(libzfs_handle_t *, const char *, zfs_type_t, nvlist_t *); +extern int zfs_create_ancestors(libzfs_handle_t *, const char *); extern int zfs_destroy(zfs_handle_t *); extern int zfs_destroy_snaps(zfs_handle_t *, char *); extern int zfs_clone(zfs_handle_t *, const char *, nvlist_t *); -extern int zfs_snapshot(libzfs_handle_t *, const char *, boolean_t); -extern int zfs_rollback(zfs_handle_t *, zfs_handle_t *, int); -extern int zfs_rename(zfs_handle_t *, const char *, int); -extern int zfs_send(zfs_handle_t *, const char *, int); -extern int zfs_receive(libzfs_handle_t *, const char *, int, int, int, - boolean_t, int); +extern int zfs_snapshot(libzfs_handle_t *, const char *, boolean_t, nvlist_t *); +extern int zfs_rollback(zfs_handle_t *, zfs_handle_t *, boolean_t); +extern int zfs_rename(zfs_handle_t *, const char *, boolean_t); +extern int zfs_send(zfs_handle_t *, const char *, const char *, + boolean_t, boolean_t, boolean_t, boolean_t, int); extern int zfs_promote(zfs_handle_t *); +typedef struct recvflags { + /* print informational messages (ie, -v was specified) */ + int verbose : 1; + + /* the destination is a prefix, not the exact fs (ie, -d) */ + int isprefix : 1; + + /* do not actually do the recv, just check if it would work (ie, -n) */ + int dryrun : 1; + + /* rollback/destroy filesystems as necessary (eg, -F) */ + int force : 1; + + /* set "canmount=off" on all modified filesystems */ + int canmountoff : 1; + + /* byteswap flag is used internally; callers need not specify */ + int byteswap : 1; +} recvflags_t; + +extern int zfs_receive(libzfs_handle_t *, const char *, recvflags_t, + int, avl_tree_t *); + /* * Miscellaneous functions. */ extern const char *zfs_type_to_name(zfs_type_t); extern void zfs_refresh_properties(zfs_handle_t *); extern int zfs_name_valid(const char *, zfs_type_t); -extern int zfs_disable(zfs_handle_t *); -extern int zfs_enable(zfs_handle_t *); extern zfs_handle_t *zfs_path_to_zhandle(libzfs_handle_t *, char *, zfs_type_t); +extern boolean_t zfs_dataset_exists(libzfs_handle_t *, const char *, + zfs_type_t); +extern int zfs_spa_version(zfs_handle_t *, int *); + +/* + * dataset permission functions. + */ +extern int zfs_perm_set(zfs_handle_t *, nvlist_t *); +extern int zfs_perm_remove(zfs_handle_t *, nvlist_t *); +extern int zfs_build_perms(zfs_handle_t *, char *, char *, + zfs_deleg_who_type_t, zfs_deleg_inherit_t, nvlist_t **nvlist_t); +extern int zfs_perm_get(zfs_handle_t *, zfs_allow_t **); +extern void zfs_free_allows(zfs_allow_t *); +extern void zfs_deleg_permissions(void); /* * Mount support functions. @@ -369,15 +501,27 @@ extern int zfs_share(zfs_handle_t *); extern int zfs_unshare(zfs_handle_t *); /* - * Protocol-specifc share support functions. + * Protocol-specific share support functions. */ extern boolean_t zfs_is_shared_nfs(zfs_handle_t *, char **); +extern boolean_t zfs_is_shared_smb(zfs_handle_t *, char **); extern int zfs_share_nfs(zfs_handle_t *); +extern int zfs_share_smb(zfs_handle_t *); +extern int zfs_shareall(zfs_handle_t *); extern int zfs_unshare_nfs(zfs_handle_t *, const char *); +extern int zfs_unshare_smb(zfs_handle_t *, const char *); extern int zfs_unshareall_nfs(zfs_handle_t *); +extern int zfs_unshareall_smb(zfs_handle_t *); +extern int zfs_unshareall_bypath(zfs_handle_t *, const char *); +extern int zfs_unshareall(zfs_handle_t *); extern boolean_t zfs_is_shared_iscsi(zfs_handle_t *); extern int zfs_share_iscsi(zfs_handle_t *); extern int zfs_unshare_iscsi(zfs_handle_t *); +#ifdef TODO +extern int zfs_iscsi_perm_check(libzfs_handle_t *, char *, ucred_t *); +#endif +extern int zfs_deleg_share_nfs(libzfs_handle_t *, char *, char *, + void *, void *, int, zfs_share_op_t); /* * FreeBSD-specific jail support function. @@ -402,12 +546,6 @@ extern void zfs_nicenum(uint64_t, char *, size_t); extern int zfs_nicestrtonum(libzfs_handle_t *, const char *, uint64_t *); /* - * Pool destroy special. Remove the device information without destroying - * the underlying dataset. - */ -extern int zfs_remove_link(zfs_handle_t *); - -/* * Given a device or file, determine if it is part of a pool. */ extern int zpool_in_use(libzfs_handle_t *, int, pool_state_t *, char **, @@ -424,6 +562,9 @@ extern int zpool_read_label(int, nvlist_t **); extern int zpool_create_zvol_links(zpool_handle_t *); extern int zpool_remove_zvol_links(zpool_handle_t *); +/* is this zvol valid for use as a dump device? */ +extern int zvol_check_dump_config(char *); + /* * Enable and disable datasets within a pool by mounting/unmounting and * sharing/unsharing them. diff --git a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_changelist.c b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_changelist.c index 5e6de6d..b905bc6 100644 --- a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_changelist.c +++ b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_changelist.c @@ -20,12 +20,12 @@ */ /* - * Copyright 2007 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. + * + * Portions Copyright 2007 Ramprakash Jelari */ -#pragma ident "%Z%%M% %I% %E% SMI" - #include <libintl.h> #include <libuutil.h> #include <stddef.h> @@ -65,18 +65,21 @@ typedef struct prop_changenode { int cn_shared; int cn_mounted; int cn_zoned; + boolean_t cn_needpost; /* is postfix() needed? */ uu_list_node_t cn_listnode; } prop_changenode_t; struct prop_changelist { zfs_prop_t cl_prop; zfs_prop_t cl_realprop; + zfs_prop_t cl_shareprop; /* used with sharenfs/sharesmb */ uu_list_pool_t *cl_pool; uu_list_t *cl_list; boolean_t cl_waslegacy; boolean_t cl_allchildren; boolean_t cl_alldependents; - int cl_flags; + int cl_mflags; /* Mount flags */ + int cl_gflags; /* Gather request flags */ boolean_t cl_haszonedchild; boolean_t cl_sorted; }; @@ -84,7 +87,8 @@ struct prop_changelist { /* * If the property is 'mountpoint', go through and unmount filesystems as * necessary. We don't do the same for 'sharenfs', because we can just re-share - * with different options without interrupting service. + * with different options without interrupting service. We do handle 'sharesmb' + * since there may be old resource names that need to be removed. */ int changelist_prefix(prop_changelist_t *clp) @@ -92,11 +96,19 @@ changelist_prefix(prop_changelist_t *clp) prop_changenode_t *cn; int ret = 0; - if (clp->cl_prop != ZFS_PROP_MOUNTPOINT) + if (clp->cl_prop != ZFS_PROP_MOUNTPOINT && + clp->cl_prop != ZFS_PROP_SHARESMB) return (0); for (cn = uu_list_first(clp->cl_list); cn != NULL; cn = uu_list_next(clp->cl_list, cn)) { + + /* if a previous loop failed, set the remaining to false */ + if (ret == -1) { + cn->cn_needpost = B_FALSE; + continue; + } + /* * If we are in the global zone, but this dataset is exported * to a local zone, do nothing. @@ -114,8 +126,11 @@ changelist_prefix(prop_changelist_t *clp) (void) zfs_unshare_iscsi(cn->cn_handle); if (zvol_remove_link(cn->cn_handle->zfs_hdl, - cn->cn_handle->zfs_name) != 0) + cn->cn_handle->zfs_name) != 0) { ret = -1; + cn->cn_needpost = B_FALSE; + (void) zfs_share_iscsi(cn->cn_handle); + } break; case ZFS_PROP_VOLSIZE: @@ -126,10 +141,28 @@ changelist_prefix(prop_changelist_t *clp) (void) zfs_unshare_iscsi(cn->cn_handle); break; } - } else if (zfs_unmount(cn->cn_handle, NULL, clp->cl_flags) != 0) - ret = -1; + } else { + /* + * Do the property specific processing. + */ + switch (clp->cl_prop) { + case ZFS_PROP_MOUNTPOINT: + if (zfs_unmount(cn->cn_handle, NULL, + clp->cl_mflags) != 0) { + ret = -1; + cn->cn_needpost = B_FALSE; + } + break; + case ZFS_PROP_SHARESMB: + (void) zfs_unshare_smb(cn->cn_handle, NULL); + break; + } + } } + if (ret == -1) + (void) changelist_postfix(clp); + return (ret); } @@ -147,7 +180,8 @@ changelist_postfix(prop_changelist_t *clp) { prop_changenode_t *cn; char shareopts[ZFS_MAXPROPLEN]; - int ret = 0; + int errors = 0; + libzfs_handle_t *hdl; /* * If we're changing the mountpoint, attempt to destroy the underlying @@ -163,11 +197,28 @@ changelist_postfix(prop_changelist_t *clp) remove_mountpoint(cn->cn_handle); /* + * It is possible that the changelist_prefix() used libshare + * to unshare some entries. Since libshare caches data, an + * attempt to reshare during postfix can fail unless libshare + * is uninitialized here so that it will reinitialize later. + */ + if (cn->cn_handle != NULL) { + hdl = cn->cn_handle->zfs_hdl; + assert(hdl != NULL); + zfs_uninit_libshare(hdl); + } + + /* * We walk the datasets in reverse, because we want to mount any parent - * datasets before mounting the children. + * datasets before mounting the children. We walk all datasets even if + * there are errors. */ for (cn = uu_list_last(clp->cl_list); cn != NULL; cn = uu_list_prev(clp->cl_list, cn)) { + + boolean_t sharenfs; + boolean_t sharesmb; + /* * If we are in the global zone, but this dataset is exported * to a local zone, do nothing. @@ -175,6 +226,11 @@ changelist_postfix(prop_changelist_t *clp) if (getzoneid() == GLOBAL_ZONEID && cn->cn_zoned) continue; + /* Only do post-processing if it's required */ + if (!cn->cn_needpost) + continue; + cn->cn_needpost = B_FALSE; + zfs_refresh_properties(cn->cn_handle); if (ZFS_IS_VOLUME(cn->cn_handle)) { @@ -185,7 +241,7 @@ changelist_postfix(prop_changelist_t *clp) if (clp->cl_realprop == ZFS_PROP_NAME && zvol_create_link(cn->cn_handle->zfs_hdl, cn->cn_handle->zfs_name) != 0) { - ret = -1; + errors++; } else if (cn->cn_shared || clp->cl_prop == ZFS_PROP_SHAREISCSI) { if (zfs_prop_get(cn->cn_handle, @@ -193,43 +249,55 @@ changelist_postfix(prop_changelist_t *clp) sizeof (shareopts), NULL, NULL, 0, B_FALSE) == 0 && strcmp(shareopts, "off") == 0) { - ret = zfs_unshare_iscsi(cn->cn_handle); + errors += + zfs_unshare_iscsi(cn->cn_handle); } else { - ret = zfs_share_iscsi(cn->cn_handle); + errors += + zfs_share_iscsi(cn->cn_handle); } } continue; } - if ((clp->cl_waslegacy || cn->cn_mounted) && - !zfs_is_mounted(cn->cn_handle, NULL) && + /* + * Remount if previously mounted or mountpoint was legacy, + * or sharenfs or sharesmb property is set. + */ + sharenfs = ((zfs_prop_get(cn->cn_handle, ZFS_PROP_SHARENFS, + shareopts, sizeof (shareopts), NULL, NULL, 0, + B_FALSE) == 0) && (strcmp(shareopts, "off") != 0)); + + sharesmb = ((zfs_prop_get(cn->cn_handle, ZFS_PROP_SHARESMB, + shareopts, sizeof (shareopts), NULL, NULL, 0, + B_FALSE) == 0) && (strcmp(shareopts, "off") != 0)); + + if ((cn->cn_mounted || clp->cl_waslegacy || sharenfs || + sharesmb) && !zfs_is_mounted(cn->cn_handle, NULL) && zfs_mount(cn->cn_handle, NULL, 0) != 0) - ret = -1; + errors++; /* * We always re-share even if the filesystem is currently * shared, so that we can adopt any new options. */ - if (cn->cn_shared || - (clp->cl_prop == ZFS_PROP_SHARENFS && clp->cl_waslegacy)) { - if (zfs_prop_get(cn->cn_handle, ZFS_PROP_SHARENFS, - shareopts, sizeof (shareopts), NULL, NULL, 0, - B_FALSE) == 0 && strcmp(shareopts, "off") == 0) { - ret = zfs_unshare_nfs(cn->cn_handle, NULL); - } else { - ret = zfs_share_nfs(cn->cn_handle); - } - } + if (sharenfs) + errors += zfs_share_nfs(cn->cn_handle); + else if (cn->cn_shared || clp->cl_waslegacy) + errors += zfs_unshare_nfs(cn->cn_handle, NULL); + if (sharesmb) + errors += zfs_share_smb(cn->cn_handle); + else if (cn->cn_shared || clp->cl_waslegacy) + errors += zfs_unshare_smb(cn->cn_handle, NULL); } - return (ret); + return (errors ? -1 : 0); } /* * Is this "dataset" a child of "parent"? */ -static boolean_t +boolean_t isa_child_of(const char *dataset, const char *parent) { int len; @@ -280,21 +348,22 @@ changelist_rename(prop_changelist_t *clp, const char *src, const char *dst) } /* - * Given a gathered changelist for the 'sharenfs' property, unshare all the - * datasets in the list. + * Given a gathered changelist for the 'sharenfs' or 'sharesmb' property, + * unshare all the datasets in the list. */ int -changelist_unshare(prop_changelist_t *clp) +changelist_unshare(prop_changelist_t *clp, zfs_share_proto_t *proto) { prop_changenode_t *cn; int ret = 0; - if (clp->cl_prop != ZFS_PROP_SHARENFS) + if (clp->cl_prop != ZFS_PROP_SHARENFS && + clp->cl_prop != ZFS_PROP_SHARESMB) return (0); for (cn = uu_list_first(clp->cl_list); cn != NULL; cn = uu_list_next(clp->cl_list, cn)) { - if (zfs_unshare_nfs(cn->cn_handle, NULL) != 0) + if (zfs_unshare_proto(cn->cn_handle, NULL, proto) != 0) ret = -1; } @@ -316,14 +385,14 @@ changelist_haszonedchild(prop_changelist_t *clp) * Remove a node from a gathered list. */ void -changelist_remove(zfs_handle_t *zhp, prop_changelist_t *clp) +changelist_remove(prop_changelist_t *clp, const char *name) { prop_changenode_t *cn; for (cn = uu_list_first(clp->cl_list); cn != NULL; cn = uu_list_next(clp->cl_list, cn)) { - if (strcmp(cn->cn_handle->zfs_name, zhp->zfs_name) == 0) { + if (strcmp(cn->cn_handle->zfs_name, name) == 0) { uu_list_remove(clp->cl_list, cn); zfs_close(cn->cn_handle); free(cn); @@ -363,7 +432,8 @@ change_one(zfs_handle_t *zhp, void *data) char property[ZFS_MAXPROPLEN]; char where[64]; prop_changenode_t *cn; - zfs_source_t sourcetype; + zprop_source_t sourcetype; + zprop_source_t share_sourcetype; /* * We only want to unmount/unshare those filesystems that may inherit @@ -383,8 +453,25 @@ change_one(zfs_handle_t *zhp, void *data) return (0); } + /* + * If we are "watching" sharenfs or sharesmb + * then check out the companion property which is tracked + * in cl_shareprop + */ + if (clp->cl_shareprop != ZPROP_INVAL && + zfs_prop_get(zhp, clp->cl_shareprop, property, + sizeof (property), &share_sourcetype, where, sizeof (where), + B_FALSE) != 0) { + zfs_close(zhp); + return (0); + } + if (clp->cl_alldependents || clp->cl_allchildren || - sourcetype == ZFS_SRC_DEFAULT || sourcetype == ZFS_SRC_INHERITED) { + sourcetype == ZPROP_SRC_DEFAULT || + sourcetype == ZPROP_SRC_INHERITED || + (clp->cl_shareprop != ZPROP_INVAL && + (share_sourcetype == ZPROP_SRC_DEFAULT || + share_sourcetype == ZPROP_SRC_INHERITED))) { if ((cn = zfs_alloc(zfs_get_handle(zhp), sizeof (prop_changenode_t))) == NULL) { zfs_close(zhp); @@ -392,9 +479,11 @@ change_one(zfs_handle_t *zhp, void *data) } cn->cn_handle = zhp; - cn->cn_mounted = zfs_is_mounted(zhp, NULL); + cn->cn_mounted = (clp->cl_gflags & CL_GATHER_MOUNT_ALWAYS) || + zfs_is_mounted(zhp, NULL); cn->cn_shared = zfs_is_shared(zhp); cn->cn_zoned = zfs_prop_get_int(zhp, ZFS_PROP_ZONED); + cn->cn_needpost = B_TRUE; /* Indicate if any child is exported to a local zone. */ if (getzoneid() == GLOBAL_ZONEID && cn->cn_zoned) @@ -467,7 +556,8 @@ compare_mountpoints(const void *a, const void *b, void *unused) * mark whether it was shared beforehand. */ prop_changelist_t * -changelist_gather(zfs_handle_t *zhp, zfs_prop_t prop, int flags) +changelist_gather(zfs_handle_t *zhp, zfs_prop_t prop, int gather_flags, + int mnt_flags) { prop_changelist_t *clp; prop_changenode_t *cn; @@ -484,7 +574,8 @@ changelist_gather(zfs_handle_t *zhp, zfs_prop_t prop, int flags) * order, regardless of their position in the hierarchy. */ if (prop == ZFS_PROP_NAME || prop == ZFS_PROP_ZONED || - prop == ZFS_PROP_MOUNTPOINT || prop == ZFS_PROP_SHARENFS) { + prop == ZFS_PROP_MOUNTPOINT || prop == ZFS_PROP_SHARENFS || + prop == ZFS_PROP_SHARESMB) { compare = compare_mountpoints; clp->cl_sorted = B_TRUE; } @@ -502,7 +593,8 @@ changelist_gather(zfs_handle_t *zhp, zfs_prop_t prop, int flags) clp->cl_list = uu_list_create(clp->cl_pool, NULL, clp->cl_sorted ? UU_LIST_SORTED : 0); - clp->cl_flags = flags; + clp->cl_gflags = gather_flags; + clp->cl_mflags = mnt_flags; if (clp->cl_list == NULL) { assert(uu_error() == UU_ERROR_NO_MEMORY); @@ -529,6 +621,8 @@ changelist_gather(zfs_handle_t *zhp, zfs_prop_t prop, int flags) clp->cl_prop = ZFS_PROP_MOUNTPOINT; } else if (prop == ZFS_PROP_VOLSIZE) { clp->cl_prop = ZFS_PROP_MOUNTPOINT; + } else if (prop == ZFS_PROP_VERSION) { + clp->cl_prop = ZFS_PROP_MOUNTPOINT; } else { clp->cl_prop = prop; } @@ -536,9 +630,19 @@ changelist_gather(zfs_handle_t *zhp, zfs_prop_t prop, int flags) if (clp->cl_prop != ZFS_PROP_MOUNTPOINT && clp->cl_prop != ZFS_PROP_SHARENFS && + clp->cl_prop != ZFS_PROP_SHARESMB && clp->cl_prop != ZFS_PROP_SHAREISCSI) return (clp); + /* + * If watching SHARENFS or SHARESMB then + * also watch its companion property. + */ + if (clp->cl_prop == ZFS_PROP_SHARENFS) + clp->cl_shareprop = ZFS_PROP_SHARESMB; + else if (clp->cl_prop == ZFS_PROP_SHARESMB) + clp->cl_shareprop = ZFS_PROP_SHARENFS; + if (clp->cl_alldependents) { if (zfs_iter_dependents(zhp, B_TRUE, change_one, clp) != 0) { changelist_free(clp); @@ -554,7 +658,7 @@ changelist_gather(zfs_handle_t *zhp, zfs_prop_t prop, int flags) * and can't tell the difference. */ if ((temp = zfs_open(zhp->zfs_hdl, zfs_get_name(zhp), - ZFS_TYPE_ANY)) == NULL) { + ZFS_TYPE_DATASET)) == NULL) { changelist_free(clp); return (NULL); } @@ -571,9 +675,11 @@ changelist_gather(zfs_handle_t *zhp, zfs_prop_t prop, int flags) } cn->cn_handle = temp; - cn->cn_mounted = zfs_is_mounted(temp, NULL); + cn->cn_mounted = (clp->cl_gflags & CL_GATHER_MOUNT_ALWAYS) || + zfs_is_mounted(temp, NULL); cn->cn_shared = zfs_is_shared(temp); cn->cn_zoned = zfs_prop_get_int(zhp, ZFS_PROP_ZONED); + cn->cn_needpost = B_TRUE; uu_list_node_init(cn, &cn->cn_listnode, clp->cl_pool); if (clp->cl_sorted) { @@ -586,14 +692,22 @@ changelist_gather(zfs_handle_t *zhp, zfs_prop_t prop, int flags) } /* - * If the property was previously 'legacy' or 'none', record this fact, - * as the behavior of changelist_postfix() will be different. + * If the mountpoint property was previously 'legacy', or 'none', + * record it as the behavior of changelist_postfix() will be different. */ - if (zfs_prop_get(zhp, prop, property, sizeof (property), + if ((clp->cl_prop == ZFS_PROP_MOUNTPOINT) && + (zfs_prop_get(zhp, prop, property, sizeof (property), NULL, NULL, 0, B_FALSE) == 0 && - (strcmp(property, "legacy") == 0 || strcmp(property, "none") == 0 || - strcmp(property, "off") == 0)) - clp->cl_waslegacy = B_TRUE; + (strcmp(property, "legacy") == 0 || + strcmp(property, "none") == 0))) { + /* + * do not automatically mount ex-legacy datasets if + * we specifically set canmount to noauto + */ + if (zfs_prop_get_int(zhp, ZFS_PROP_CANMOUNT) != + ZFS_CANMOUNT_NOAUTO) + clp->cl_waslegacy = B_TRUE; + } return (clp); } diff --git a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_dataset.c b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_dataset.c index 4fc441a..58ce6c8 100644 --- a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_dataset.c +++ b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_dataset.c @@ -20,12 +20,10 @@ */ /* - * Copyright 2007 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ -#pragma ident "%Z%%M% %I% %E% SMI" - #include <assert.h> #include <ctype.h> #include <errno.h> @@ -35,20 +33,26 @@ #include <stdlib.h> #include <strings.h> #include <unistd.h> +#include <stddef.h> #include <zone.h> #include <fcntl.h> #include <sys/mntent.h> #include <sys/mnttab.h> #include <sys/mount.h> +#include <sys/avl.h> +#include <priv.h> +#include <pwd.h> +#include <grp.h> +#include <stddef.h> #include <sys/spa.h> -#include <sys/zio.h> #include <sys/zap.h> #include <libzfs.h> #include "zfs_namecheck.h" #include "zfs_prop.h" #include "libzfs_impl.h" +#include "zfs_deleg.h" static int zvol_create_link_common(libzfs_handle_t *, const char *, int); @@ -121,7 +125,8 @@ path_to_str(const char *path, int types) * 'buf' detailing exactly why the name was not valid. */ static int -zfs_validate_name(libzfs_handle_t *hdl, const char *path, int type) +zfs_validate_name(libzfs_handle_t *hdl, const char *path, int type, + boolean_t modifying) { namecheck_err_t why; char what; @@ -194,43 +199,118 @@ zfs_validate_name(libzfs_handle_t *hdl, const char *path, int type) return (0); } + if (modifying && strchr(path, '%') != NULL) { + if (hdl != NULL) + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "invalid character %c in name"), '%'); + return (0); + } + return (-1); } int zfs_name_valid(const char *name, zfs_type_t type) { - return (zfs_validate_name(NULL, name, type)); + if (type == ZFS_TYPE_POOL) + return (zpool_name_valid(NULL, B_FALSE, name)); + return (zfs_validate_name(NULL, name, type, B_FALSE)); } /* * This function takes the raw DSL properties, and filters out the user-defined * properties into a separate nvlist. */ -static int -process_user_props(zfs_handle_t *zhp) +static nvlist_t * +process_user_props(zfs_handle_t *zhp, nvlist_t *props) { libzfs_handle_t *hdl = zhp->zfs_hdl; nvpair_t *elem; nvlist_t *propval; + nvlist_t *nvl; - nvlist_free(zhp->zfs_user_props); - - if (nvlist_alloc(&zhp->zfs_user_props, NV_UNIQUE_NAME, 0) != 0) - return (no_memory(hdl)); + if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0) { + (void) no_memory(hdl); + return (NULL); + } elem = NULL; - while ((elem = nvlist_next_nvpair(zhp->zfs_props, elem)) != NULL) { + while ((elem = nvlist_next_nvpair(props, elem)) != NULL) { if (!zfs_prop_user(nvpair_name(elem))) continue; verify(nvpair_value_nvlist(elem, &propval) == 0); - if (nvlist_add_nvlist(zhp->zfs_user_props, - nvpair_name(elem), propval) != 0) - return (no_memory(hdl)); + if (nvlist_add_nvlist(nvl, nvpair_name(elem), propval) != 0) { + nvlist_free(nvl); + (void) no_memory(hdl); + return (NULL); + } } - return (0); + return (nvl); +} + +static zpool_handle_t * +zpool_add_handle(zfs_handle_t *zhp, const char *pool_name) +{ + libzfs_handle_t *hdl = zhp->zfs_hdl; + zpool_handle_t *zph; + + if ((zph = zpool_open_canfail(hdl, pool_name)) != NULL) { + if (hdl->libzfs_pool_handles != NULL) + zph->zpool_next = hdl->libzfs_pool_handles; + hdl->libzfs_pool_handles = zph; + } + return (zph); +} + +static zpool_handle_t * +zpool_find_handle(zfs_handle_t *zhp, const char *pool_name, int len) +{ + libzfs_handle_t *hdl = zhp->zfs_hdl; + zpool_handle_t *zph = hdl->libzfs_pool_handles; + + while ((zph != NULL) && + (strncmp(pool_name, zpool_get_name(zph), len) != 0)) + zph = zph->zpool_next; + return (zph); +} + +/* + * Returns a handle to the pool that contains the provided dataset. + * If a handle to that pool already exists then that handle is returned. + * Otherwise, a new handle is created and added to the list of handles. + */ +static zpool_handle_t * +zpool_handle(zfs_handle_t *zhp) +{ + char *pool_name; + int len; + zpool_handle_t *zph; + + len = strcspn(zhp->zfs_name, "/@") + 1; + pool_name = zfs_alloc(zhp->zfs_hdl, len); + (void) strlcpy(pool_name, zhp->zfs_name, len); + + zph = zpool_find_handle(zhp, pool_name, len); + if (zph == NULL) + zph = zpool_add_handle(zhp, pool_name); + + free(pool_name); + return (zph); +} + +void +zpool_free_handles(libzfs_handle_t *hdl) +{ + zpool_handle_t *next, *zph = hdl->libzfs_pool_handles; + + while (zph != NULL) { + next = zph->zpool_next; + zpool_close(zph); + zph = next; + } + hdl->libzfs_pool_handles = NULL; } /* @@ -241,6 +321,7 @@ get_stats(zfs_handle_t *zhp) { zfs_cmd_t zc = { 0 }; libzfs_handle_t *hdl = zhp->zfs_hdl; + nvlist_t *allprops, *userprops; (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); @@ -261,22 +342,23 @@ get_stats(zfs_handle_t *zhp) zhp->zfs_dmustats = zc.zc_objset_stats; /* structure assignment */ - (void) strlcpy(zhp->zfs_root, zc.zc_value, sizeof (zhp->zfs_root)); - - if (zhp->zfs_props) { - nvlist_free(zhp->zfs_props); - zhp->zfs_props = NULL; - } - - if (zcmd_read_dst_nvlist(hdl, &zc, &zhp->zfs_props) != 0) { + if (zcmd_read_dst_nvlist(hdl, &zc, &allprops) != 0) { zcmd_free_nvlists(&zc); return (-1); } zcmd_free_nvlists(&zc); - if (process_user_props(zhp) != 0) + if ((userprops = process_user_props(zhp, allprops)) == NULL) { + nvlist_free(allprops); return (-1); + } + + nvlist_free(zhp->zfs_props); + nvlist_free(zhp->zfs_user_props); + + zhp->zfs_props = allprops; + zhp->zfs_user_props = userprops; return (0); } @@ -298,16 +380,25 @@ zfs_handle_t * make_dataset_handle(libzfs_handle_t *hdl, const char *path) { zfs_handle_t *zhp = calloc(sizeof (zfs_handle_t), 1); + char *logstr; if (zhp == NULL) return (NULL); zhp->zfs_hdl = hdl; + /* + * Preserve history log string. + * any changes performed here will be + * logged as an internal event. + */ + logstr = zhp->zfs_hdl->libzfs_log_str; + zhp->zfs_hdl->libzfs_log_str = NULL; top: (void) strlcpy(zhp->zfs_name, path, sizeof (zhp->zfs_name)); if (get_stats(zhp) != 0) { + zhp->zfs_hdl->libzfs_log_str = logstr; free(zhp); return (NULL); } @@ -339,18 +430,19 @@ top: zc.zc_objset_type = DMU_OST_ZFS; } - /* If we can successfully roll it back, reget the stats */ - if (ioctl(hdl->libzfs_fd, ZFS_IOC_ROLLBACK, &zc) == 0) - goto top; /* - * If we can sucessfully destroy it, pretend that it + * If we can successfully destroy it, pretend that it * never existed. */ if (ioctl(hdl->libzfs_fd, ZFS_IOC_DESTROY, &zc) == 0) { + zhp->zfs_hdl->libzfs_log_str = logstr; free(zhp); errno = ENOENT; return (NULL); } + /* If we can successfully roll it back, reget the stats */ + if (ioctl(hdl->libzfs_fd, ZFS_IOC_ROLLBACK, &zc) == 0) + goto top; } /* @@ -373,6 +465,8 @@ top: else abort(); /* we should never see any other types */ + zhp->zfs_hdl->libzfs_log_str = logstr; + zhp->zpool_hdl = zpool_handle(zhp); return (zhp); } @@ -393,7 +487,7 @@ zfs_open(libzfs_handle_t *hdl, const char *path, int types) /* * Validate the name before we even try to open it. */ - if (!zfs_validate_name(hdl, path, ZFS_TYPE_ANY)) { + if (!zfs_validate_name(hdl, path, ZFS_TYPE_DATASET, B_FALSE)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid dataset name")); (void) zfs_error(hdl, EZFS_INVALIDNAME, errbuf); @@ -431,350 +525,92 @@ zfs_close(zfs_handle_t *zhp) free(zhp); } -/* - * Given a numeric suffix, convert the value into a number of bits that the - * resulting value must be shifted. - */ -static int -str2shift(libzfs_handle_t *hdl, const char *buf) -{ - const char *ends = "BKMGTPEZ"; - int i; - - if (buf[0] == '\0') - return (0); - for (i = 0; i < strlen(ends); i++) { - if (toupper(buf[0]) == ends[i]) - break; - } - if (i == strlen(ends)) { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "invalid numeric suffix '%s'"), buf); - return (-1); - } - - /* - * We want to allow trailing 'b' characters for 'GB' or 'Mb'. But don't - * allow 'BB' - that's just weird. - */ - if (buf[1] == '\0' || (toupper(buf[1]) == 'B' && buf[2] == '\0' && - toupper(buf[0]) != 'B')) - return (10*i); - - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "invalid numeric suffix '%s'"), buf); - return (-1); -} - -/* - * Convert a string of the form '100G' into a real number. Used when setting - * properties or creating a volume. 'buf' is used to place an extended error - * message for the caller to use. - */ -static int -nicestrtonum(libzfs_handle_t *hdl, const char *value, uint64_t *num) +int +zfs_spa_version(zfs_handle_t *zhp, int *spa_version) { - char *end; - int shift; - - *num = 0; - - /* Check to see if this looks like a number. */ - if ((value[0] < '0' || value[0] > '9') && value[0] != '.') { - if (hdl) - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "bad numeric value '%s'"), value); - return (-1); - } - - /* Rely on stroll() to process the numeric portion. */ - errno = 0; - *num = strtoll(value, &end, 10); + zpool_handle_t *zpool_handle = zhp->zpool_hdl; - /* - * Check for ERANGE, which indicates that the value is too large to fit - * in a 64-bit value. - */ - if (errno == ERANGE) { - if (hdl) - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "numeric value is too large")); + if (zpool_handle == NULL) return (-1); - } - - /* - * If we have a decimal value, then do the computation with floating - * point arithmetic. Otherwise, use standard arithmetic. - */ - if (*end == '.') { - double fval = strtod(value, &end); - - if ((shift = str2shift(hdl, end)) == -1) - return (-1); - - fval *= pow(2, shift); - - if (fval > UINT64_MAX) { - if (hdl) - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "numeric value is too large")); - return (-1); - } - - *num = (uint64_t)fval; - } else { - if ((shift = str2shift(hdl, end)) == -1) - return (-1); - - /* Check for overflow */ - if (shift >= 64 || (*num << shift) >> shift != *num) { - if (hdl) - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "numeric value is too large")); - return (-1); - } - - *num <<= shift; - } + *spa_version = zpool_get_prop_int(zpool_handle, + ZPOOL_PROP_VERSION, NULL); return (0); } -int -zfs_nicestrtonum(libzfs_handle_t *hdl, const char *str, uint64_t *val) -{ - return (nicestrtonum(hdl, str, val)); -} - /* - * The prop_parse_*() functions are designed to allow flexibility in callers - * when setting properties. At the DSL layer, all properties are either 64-bit - * numbers or strings. We want the user to be able to ignore this fact and - * specify properties as native values (boolean, for example) or as strings (to - * simplify command line utilities). This also handles converting index types - * (compression, checksum, etc) from strings to their on-disk index. + * The choice of reservation property depends on the SPA version. */ - static int -prop_parse_boolean(libzfs_handle_t *hdl, nvpair_t *elem, uint64_t *val) +zfs_which_resv_prop(zfs_handle_t *zhp, zfs_prop_t *resv_prop) { - uint64_t ret; + int spa_version; - switch (nvpair_type(elem)) { - case DATA_TYPE_STRING: - { - char *value; - verify(nvpair_value_string(elem, &value) == 0); - - if (strcmp(value, "on") == 0) { - ret = 1; - } else if (strcmp(value, "off") == 0) { - ret = 0; - } else { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "property '%s' must be 'on' or 'off'"), - nvpair_name(elem)); - return (-1); - } - break; - } - - case DATA_TYPE_UINT64: - { - verify(nvpair_value_uint64(elem, &ret) == 0); - if (ret > 1) { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "'%s' must be a boolean value"), - nvpair_name(elem)); - return (-1); - } - break; - } - - case DATA_TYPE_BOOLEAN_VALUE: - { - boolean_t value; - verify(nvpair_value_boolean_value(elem, &value) == 0); - ret = value; - break; - } - - default: - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "'%s' must be a boolean value"), - nvpair_name(elem)); + if (zfs_spa_version(zhp, &spa_version) < 0) return (-1); - } - - *val = ret; - return (0); -} - -static int -prop_parse_number(libzfs_handle_t *hdl, nvpair_t *elem, zfs_prop_t prop, - uint64_t *val) -{ - uint64_t ret; - boolean_t isnone = B_FALSE; - switch (nvpair_type(elem)) { - case DATA_TYPE_STRING: - { - char *value; - (void) nvpair_value_string(elem, &value); - if (strcmp(value, "none") == 0) { - isnone = B_TRUE; - ret = 0; - } else if (nicestrtonum(hdl, value, &ret) != 0) { - return (-1); - } - break; - } - - case DATA_TYPE_UINT64: - (void) nvpair_value_uint64(elem, &ret); - break; - - default: - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "'%s' must be a number"), - nvpair_name(elem)); - return (-1); - } - - /* - * Quota special: force 'none' and don't allow 0. - */ - if (ret == 0 && !isnone && prop == ZFS_PROP_QUOTA) { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "use 'none' to disable quota")); - return (-1); - } - - *val = ret; - return (0); -} - -static int -prop_parse_index(libzfs_handle_t *hdl, nvpair_t *elem, zfs_prop_t prop, - uint64_t *val) -{ - char *propname = nvpair_name(elem); - char *value; - - if (nvpair_type(elem) != DATA_TYPE_STRING) { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "'%s' must be a string"), propname); - return (-1); - } - - (void) nvpair_value_string(elem, &value); - - if (zfs_prop_string_to_index(prop, value, val) != 0) { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "'%s' must be one of '%s'"), propname, - zfs_prop_values(prop)); - return (-1); - } + if (spa_version >= SPA_VERSION_REFRESERVATION) + *resv_prop = ZFS_PROP_REFRESERVATION; + else + *resv_prop = ZFS_PROP_RESERVATION; return (0); } /* - * Check if the bootfs name has the same pool name as it is set to. - * Assuming bootfs is a valid dataset name. - */ -static boolean_t -bootfs_poolname_valid(char *pool, char *bootfs) -{ - char ch, *pname; - - /* get the pool name from the bootfs name */ - pname = bootfs; - while (*bootfs && !isspace(*bootfs) && *bootfs != '/') - bootfs++; - - ch = *bootfs; - *bootfs = 0; - - if (strcmp(pool, pname) == 0) { - *bootfs = ch; - return (B_TRUE); - } - - *bootfs = ch; - return (B_FALSE); -} - -/* * Given an nvlist of properties to set, validates that they are correct, and * parses any numeric properties (index, boolean, etc) if they are specified as * strings. */ nvlist_t * -zfs_validate_properties(libzfs_handle_t *hdl, zfs_type_t type, char *pool_name, - nvlist_t *nvl, uint64_t zoned, zfs_handle_t *zhp, const char *errbuf) +zfs_valid_proplist(libzfs_handle_t *hdl, zfs_type_t type, nvlist_t *nvl, + uint64_t zoned, zfs_handle_t *zhp, const char *errbuf) { nvpair_t *elem; - const char *propname; - zfs_prop_t prop; uint64_t intval; char *strval; + zfs_prop_t prop; nvlist_t *ret; - int isuser; + int chosen_normal = -1; + int chosen_utf = -1; if (nvlist_alloc(&ret, NV_UNIQUE_NAME, 0) != 0) { (void) no_memory(hdl); return (NULL); } - if (type == ZFS_TYPE_SNAPSHOT) { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "snapshot properties cannot be modified")); - (void) zfs_error(hdl, EZFS_PROPTYPE, errbuf); - goto error; - } - elem = NULL; while ((elem = nvlist_next_nvpair(nvl, elem)) != NULL) { - propname = nvpair_name(elem); + const char *propname = nvpair_name(elem); /* * Make sure this property is valid and applies to this type. */ - if ((prop = zfs_name_to_prop_common(propname, type)) - == ZFS_PROP_INVAL) { - isuser = zfs_prop_user(propname); - if (!isuser || (isuser && (type & ZFS_TYPE_POOL))) { + if ((prop = zfs_name_to_prop(propname)) == ZPROP_INVAL) { + if (!zfs_prop_user(propname)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "invalid property '%s'"), - propname); + "invalid property '%s'"), propname); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; - } else { - /* - * If this is a user property, make sure it's a - * string, and that it's less than - * ZAP_MAXNAMELEN. - */ - if (nvpair_type(elem) != DATA_TYPE_STRING) { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "'%s' must be a string"), - propname); - (void) zfs_error(hdl, EZFS_BADPROP, - errbuf); - goto error; - } + } - if (strlen(nvpair_name(elem)) >= - ZAP_MAXNAMELEN) { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "property name '%s' is too long"), - propname); - (void) zfs_error(hdl, EZFS_BADPROP, - errbuf); - goto error; - } + /* + * If this is a user property, make sure it's a + * string, and that it's less than ZAP_MAXNAMELEN. + */ + if (nvpair_type(elem) != DATA_TYPE_STRING) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "'%s' must be a string"), propname); + (void) zfs_error(hdl, EZFS_BADPROP, errbuf); + goto error; + } + + if (strlen(nvpair_name(elem)) >= ZAP_MAXNAMELEN) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "property name '%s' is too long"), + propname); + (void) zfs_error(hdl, EZFS_BADPROP, errbuf); + goto error; } (void) nvpair_value_string(elem, &strval); @@ -785,10 +621,12 @@ zfs_validate_properties(libzfs_handle_t *hdl, zfs_type_t type, char *pool_name, continue; } - /* - * Normalize the name, to get rid of shorthand abbrevations. - */ - propname = zfs_prop_to_name(prop); + if (type == ZFS_TYPE_SNAPSHOT) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "this property can not be modified for snapshots")); + (void) zfs_error(hdl, EZFS_PROPTYPE, errbuf); + goto error; + } if (!zfs_prop_valid_for_type(prop, type)) { zfs_error_aux(hdl, @@ -799,7 +637,7 @@ zfs_validate_properties(libzfs_handle_t *hdl, zfs_type_t type, char *pool_name, } if (zfs_prop_readonly(prop) && - (prop != ZFS_PROP_VOLBLOCKSIZE || zhp != NULL)) { + (!zfs_prop_setonce(prop) || zhp != NULL)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' is readonly"), propname); @@ -807,70 +645,31 @@ zfs_validate_properties(libzfs_handle_t *hdl, zfs_type_t type, char *pool_name, goto error; } + if (zprop_parse_value(hdl, elem, prop, type, ret, + &strval, &intval, errbuf) != 0) + goto error; + /* - * Convert any properties to the internal DSL value types. + * Perform some additional checks for specific properties. */ - strval = NULL; - switch (zfs_prop_get_type(prop)) { - case prop_type_boolean: - if (prop_parse_boolean(hdl, elem, &intval) != 0) { - (void) zfs_error(hdl, EZFS_BADPROP, errbuf); - goto error; - } - break; + switch (prop) { + case ZFS_PROP_VERSION: + { + int version; - case prop_type_string: - if (nvpair_type(elem) != DATA_TYPE_STRING) { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "'%s' must be a string"), - propname); - (void) zfs_error(hdl, EZFS_BADPROP, errbuf); - goto error; - } - (void) nvpair_value_string(elem, &strval); - if (strlen(strval) >= ZFS_MAXPROPLEN) { + if (zhp == NULL) + break; + version = zfs_prop_get_int(zhp, ZFS_PROP_VERSION); + if (intval < version) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "'%s' is too long"), propname); + "Can not downgrade; already at version %u"), + version); (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } break; - - case prop_type_number: - if (prop_parse_number(hdl, elem, prop, &intval) != 0) { - (void) zfs_error(hdl, EZFS_BADPROP, errbuf); - goto error; - } - break; - - case prop_type_index: - if (prop_parse_index(hdl, elem, prop, &intval) != 0) { - (void) zfs_error(hdl, EZFS_BADPROP, errbuf); - goto error; - } - break; - - default: - abort(); } - /* - * Add the result to our return set of properties. - */ - if (strval) { - if (nvlist_add_string(ret, propname, strval) != 0) { - (void) no_memory(hdl); - goto error; - } - } else if (nvlist_add_uint64(ret, propname, intval) != 0) { - (void) no_memory(hdl); - goto error; - } - - /* - * Perform some additional checks for specific properties. - */ - switch (prop) { case ZFS_PROP_RECORDSIZE: case ZFS_PROP_VOLBLOCKSIZE: /* must be power of two within SPA_{MIN,MAX}BLOCKSIZE */ @@ -900,32 +699,52 @@ zfs_validate_properties(libzfs_handle_t *hdl, zfs_type_t type, char *pool_name, break; case ZFS_PROP_MOUNTPOINT: + { + namecheck_err_t why; + if (strcmp(strval, ZFS_MOUNTPOINT_NONE) == 0 || strcmp(strval, ZFS_MOUNTPOINT_LEGACY) == 0) break; - if (strval[0] != '/') { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "'%s' must be an absolute path, " - "'none', or 'legacy'"), propname); + if (mountpoint_namecheck(strval, &why)) { + switch (why) { + case NAME_ERR_LEADING_SLASH: + zfs_error_aux(hdl, + dgettext(TEXT_DOMAIN, + "'%s' must be an absolute path, " + "'none', or 'legacy'"), propname); + break; + case NAME_ERR_TOOLONG: + zfs_error_aux(hdl, + dgettext(TEXT_DOMAIN, + "component of '%s' is too long"), + propname); + break; + } (void) zfs_error(hdl, EZFS_BADPROP, errbuf); goto error; } + } + /*FALLTHRU*/ + case ZFS_PROP_SHARESMB: case ZFS_PROP_SHARENFS: /* - * For the mountpoint and sharenfs properties, check if - * it can be set in a global/non-global zone based on + * For the mountpoint and sharenfs or sharesmb + * properties, check if it can be set in a + * global/non-global zone based on * the zoned property value: * * global zone non-global zone * -------------------------------------------------- * zoned=on mountpoint (no) mountpoint (yes) * sharenfs (no) sharenfs (no) + * sharesmb (no) sharesmb (no) * * zoned=off mountpoint (yes) N/A * sharenfs (yes) + * sharesmb (yes) */ if (zoned) { if (getzoneid() == GLOBAL_ZONEID) { @@ -936,7 +755,8 @@ zfs_validate_properties(libzfs_handle_t *hdl, zfs_type_t type, char *pool_name, (void) zfs_error(hdl, EZFS_ZONED, errbuf); goto error; - } else if (prop == ZFS_PROP_SHARENFS) { + } else if (prop == ZFS_PROP_SHARENFS || + prop == ZFS_PROP_SHARESMB) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' cannot be set in " "a non-global zone"), propname); @@ -956,22 +776,73 @@ zfs_validate_properties(libzfs_handle_t *hdl, zfs_type_t type, char *pool_name, goto error; } - break; - - case ZFS_PROP_BOOTFS: /* - * bootfs property value has to be a dataset name and - * the dataset has to be in the same pool as it sets to. + * At this point, it is legitimate to set the + * property. Now we want to make sure that the + * property value is valid if it is sharenfs. */ - if (strval[0] != '\0' && (!zfs_name_valid(strval, - ZFS_TYPE_FILESYSTEM) || !bootfs_poolname_valid( - pool_name, strval))) { + if ((prop == ZFS_PROP_SHARENFS || + prop == ZFS_PROP_SHARESMB) && + strcmp(strval, "on") != 0 && + strcmp(strval, "off") != 0) { + zfs_share_proto_t proto; - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' " - "is an invalid name"), strval); - (void) zfs_error(hdl, EZFS_INVALIDNAME, errbuf); - goto error; + if (prop == ZFS_PROP_SHARESMB) + proto = PROTO_SMB; + else + proto = PROTO_NFS; + + /* + * Must be an valid sharing protocol + * option string so init the libshare + * in order to enable the parser and + * then parse the options. We use the + * control API since we don't care about + * the current configuration and don't + * want the overhead of loading it + * until we actually do something. + */ + + if (zfs_init_libshare(hdl, + SA_INIT_CONTROL_API) != SA_OK) { + /* + * An error occurred so we can't do + * anything + */ + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "'%s' cannot be set: problem " + "in share initialization"), + propname); + (void) zfs_error(hdl, EZFS_BADPROP, + errbuf); + goto error; + } + + if (zfs_parse_options(strval, proto) != SA_OK) { + /* + * There was an error in parsing so + * deal with it by issuing an error + * message and leaving after + * uninitializing the the libshare + * interface. + */ + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "'%s' cannot be set to invalid " + "options"), propname); + (void) zfs_error(hdl, EZFS_BADPROP, + errbuf); + zfs_uninit_libshare(hdl); + goto error; + } + zfs_uninit_libshare(hdl); } + + break; + case ZFS_PROP_UTF8ONLY: + chosen_utf = (int)intval; + break; + case ZFS_PROP_NORMALIZE: + chosen_normal = (int)intval; break; } @@ -988,6 +859,7 @@ zfs_validate_properties(libzfs_handle_t *hdl, zfs_type_t type, char *pool_name, switch (prop) { case ZFS_PROP_RESERVATION: + case ZFS_PROP_REFRESERVATION: if (intval > volsize) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' is greater than current " @@ -1025,6 +897,27 @@ zfs_validate_properties(libzfs_handle_t *hdl, zfs_type_t type, char *pool_name, } /* + * If normalization was chosen, but no UTF8 choice was made, + * enforce rejection of non-UTF8 names. + * + * If normalization was chosen, but rejecting non-UTF8 names + * was explicitly not chosen, it is an error. + */ + if (chosen_normal > 0 && chosen_utf < 0) { + if (nvlist_add_uint64(ret, + zfs_prop_to_name(ZFS_PROP_UTF8ONLY), 1) != 0) { + (void) no_memory(hdl); + goto error; + } + } else if (chosen_normal > 0 && chosen_utf == 0) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "'%s' must be set 'on' if normalization chosen"), + zfs_prop_to_name(ZFS_PROP_UTF8ONLY)); + (void) zfs_error(hdl, EZFS_BADPROP, errbuf); + goto error; + } + + /* * If this is an existing volume, and someone is setting the volsize, * make sure that it matches the reservation, or add it if necessary. */ @@ -1033,23 +926,24 @@ zfs_validate_properties(libzfs_handle_t *hdl, zfs_type_t type, char *pool_name, &intval) == 0) { uint64_t old_volsize = zfs_prop_get_int(zhp, ZFS_PROP_VOLSIZE); - uint64_t old_reservation = zfs_prop_get_int(zhp, - ZFS_PROP_RESERVATION); + uint64_t old_reservation; uint64_t new_reservation; + zfs_prop_t resv_prop; + + if (zfs_which_resv_prop(zhp, &resv_prop) < 0) + goto error; + old_reservation = zfs_prop_get_int(zhp, resv_prop); if (old_volsize == old_reservation && - nvlist_lookup_uint64(ret, - zfs_prop_to_name(ZFS_PROP_RESERVATION), + nvlist_lookup_uint64(ret, zfs_prop_to_name(resv_prop), &new_reservation) != 0) { if (nvlist_add_uint64(ret, - zfs_prop_to_name(ZFS_PROP_RESERVATION), - intval) != 0) { + zfs_prop_to_name(resv_prop), intval) != 0) { (void) no_memory(hdl); goto error; } } } - return (ret); error: @@ -1057,6 +951,808 @@ error: return (NULL); } +static int +zfs_get_perm_who(const char *who, zfs_deleg_who_type_t *who_type, + uint64_t *ret_who) +{ + struct passwd *pwd; + struct group *grp; + uid_t id; + + if (*who_type == ZFS_DELEG_EVERYONE || *who_type == ZFS_DELEG_CREATE || + *who_type == ZFS_DELEG_NAMED_SET) { + *ret_who = -1; + return (0); + } + if (who == NULL && !(*who_type == ZFS_DELEG_EVERYONE)) + return (EZFS_BADWHO); + + if (*who_type == ZFS_DELEG_WHO_UNKNOWN && + strcmp(who, "everyone") == 0) { + *ret_who = -1; + *who_type = ZFS_DELEG_EVERYONE; + return (0); + } + + pwd = getpwnam(who); + grp = getgrnam(who); + + if ((*who_type == ZFS_DELEG_USER) && pwd) { + *ret_who = pwd->pw_uid; + } else if ((*who_type == ZFS_DELEG_GROUP) && grp) { + *ret_who = grp->gr_gid; + } else if (pwd) { + *ret_who = pwd->pw_uid; + *who_type = ZFS_DELEG_USER; + } else if (grp) { + *ret_who = grp->gr_gid; + *who_type = ZFS_DELEG_GROUP; + } else { + char *end; + + id = strtol(who, &end, 10); + if (errno != 0 || *end != '\0') { + return (EZFS_BADWHO); + } else { + *ret_who = id; + if (*who_type == ZFS_DELEG_WHO_UNKNOWN) + *who_type = ZFS_DELEG_USER; + } + } + + return (0); +} + +static void +zfs_perms_add_to_nvlist(nvlist_t *who_nvp, char *name, nvlist_t *perms_nvp) +{ + if (perms_nvp != NULL) { + verify(nvlist_add_nvlist(who_nvp, + name, perms_nvp) == 0); + } else { + verify(nvlist_add_boolean(who_nvp, name) == 0); + } +} + +static void +helper(zfs_deleg_who_type_t who_type, uint64_t whoid, char *whostr, + zfs_deleg_inherit_t inherit, nvlist_t *who_nvp, nvlist_t *perms_nvp, + nvlist_t *sets_nvp) +{ + boolean_t do_perms, do_sets; + char name[ZFS_MAX_DELEG_NAME]; + + do_perms = (nvlist_next_nvpair(perms_nvp, NULL) != NULL); + do_sets = (nvlist_next_nvpair(sets_nvp, NULL) != NULL); + + if (!do_perms && !do_sets) + do_perms = do_sets = B_TRUE; + + if (do_perms) { + zfs_deleg_whokey(name, who_type, inherit, + (who_type == ZFS_DELEG_NAMED_SET) ? + whostr : (void *)&whoid); + zfs_perms_add_to_nvlist(who_nvp, name, perms_nvp); + } + if (do_sets) { + zfs_deleg_whokey(name, toupper(who_type), inherit, + (who_type == ZFS_DELEG_NAMED_SET) ? + whostr : (void *)&whoid); + zfs_perms_add_to_nvlist(who_nvp, name, sets_nvp); + } +} + +static void +zfs_perms_add_who_nvlist(nvlist_t *who_nvp, uint64_t whoid, void *whostr, + nvlist_t *perms_nvp, nvlist_t *sets_nvp, + zfs_deleg_who_type_t who_type, zfs_deleg_inherit_t inherit) +{ + if (who_type == ZFS_DELEG_NAMED_SET || who_type == ZFS_DELEG_CREATE) { + helper(who_type, whoid, whostr, 0, + who_nvp, perms_nvp, sets_nvp); + } else { + if (inherit & ZFS_DELEG_PERM_LOCAL) { + helper(who_type, whoid, whostr, ZFS_DELEG_LOCAL, + who_nvp, perms_nvp, sets_nvp); + } + if (inherit & ZFS_DELEG_PERM_DESCENDENT) { + helper(who_type, whoid, whostr, ZFS_DELEG_DESCENDENT, + who_nvp, perms_nvp, sets_nvp); + } + } +} + +/* + * Construct nvlist to pass down to kernel for setting/removing permissions. + * + * The nvlist is constructed as a series of nvpairs with an optional embedded + * nvlist of permissions to remove or set. The topmost nvpairs are the actual + * base attribute named stored in the dsl. + * Arguments: + * + * whostr: is a comma separated list of users, groups, or a single set name. + * whostr may be null for everyone or create perms. + * who_type: is the type of entry in whostr. Typically this will be + * ZFS_DELEG_WHO_UNKNOWN. + * perms: common separated list of permissions. May be null if user + * is requested to remove permissions by who. + * inherit: Specifies the inheritance of the permissions. Will be either + * ZFS_DELEG_PERM_LOCAL and/or ZFS_DELEG_PERM_DESCENDENT. + * nvp The constructed nvlist to pass to zfs_perm_set(). + * The output nvp will look something like this. + * ul$1234 -> {create ; destroy } + * Ul$1234 -> { @myset } + * s-$@myset - { snapshot; checksum; compression } + */ +int +zfs_build_perms(zfs_handle_t *zhp, char *whostr, char *perms, + zfs_deleg_who_type_t who_type, zfs_deleg_inherit_t inherit, nvlist_t **nvp) +{ + nvlist_t *who_nvp; + nvlist_t *perms_nvp = NULL; + nvlist_t *sets_nvp = NULL; + char errbuf[1024]; + char *who_tok, *perm; + int error; + + *nvp = NULL; + + if (perms) { + if ((error = nvlist_alloc(&perms_nvp, + NV_UNIQUE_NAME, 0)) != 0) { + return (1); + } + if ((error = nvlist_alloc(&sets_nvp, + NV_UNIQUE_NAME, 0)) != 0) { + nvlist_free(perms_nvp); + return (1); + } + } + + if ((error = nvlist_alloc(&who_nvp, NV_UNIQUE_NAME, 0)) != 0) { + if (perms_nvp) + nvlist_free(perms_nvp); + if (sets_nvp) + nvlist_free(sets_nvp); + return (1); + } + + if (who_type == ZFS_DELEG_NAMED_SET) { + namecheck_err_t why; + char what; + + if ((error = permset_namecheck(whostr, &why, &what)) != 0) { + nvlist_free(who_nvp); + if (perms_nvp) + nvlist_free(perms_nvp); + if (sets_nvp) + nvlist_free(sets_nvp); + + switch (why) { + case NAME_ERR_NO_AT: + zfs_error_aux(zhp->zfs_hdl, + dgettext(TEXT_DOMAIN, + "set definition must begin with an '@' " + "character")); + } + return (zfs_error(zhp->zfs_hdl, + EZFS_BADPERMSET, whostr)); + } + } + + /* + * Build up nvlist(s) of permissions. Two nvlists are maintained. + * The first nvlist perms_nvp will have normal permissions and the + * other sets_nvp will have only permssion set names in it. + */ + for (perm = strtok(perms, ","); perm; perm = strtok(NULL, ",")) { + const char *perm_canonical = zfs_deleg_canonicalize_perm(perm); + + if (perm_canonical) { + verify(nvlist_add_boolean(perms_nvp, + perm_canonical) == 0); + } else if (perm[0] == '@') { + verify(nvlist_add_boolean(sets_nvp, perm) == 0); + } else { + nvlist_free(who_nvp); + nvlist_free(perms_nvp); + nvlist_free(sets_nvp); + return (zfs_error(zhp->zfs_hdl, EZFS_BADPERM, perm)); + } + } + + if (whostr && who_type != ZFS_DELEG_CREATE) { + who_tok = strtok(whostr, ","); + if (who_tok == NULL) { + nvlist_free(who_nvp); + if (perms_nvp) + nvlist_free(perms_nvp); + if (sets_nvp) + nvlist_free(sets_nvp); + (void) snprintf(errbuf, sizeof (errbuf), + dgettext(TEXT_DOMAIN, "Who string is NULL"), + whostr); + return (zfs_error(zhp->zfs_hdl, EZFS_BADWHO, errbuf)); + } + } + + /* + * Now create the nvlist(s) + */ + do { + uint64_t who_id; + + error = zfs_get_perm_who(who_tok, &who_type, + &who_id); + if (error) { + nvlist_free(who_nvp); + if (perms_nvp) + nvlist_free(perms_nvp); + if (sets_nvp) + nvlist_free(sets_nvp); + (void) snprintf(errbuf, sizeof (errbuf), + dgettext(TEXT_DOMAIN, + "Unable to determine uid/gid for " + "%s "), who_tok); + return (zfs_error(zhp->zfs_hdl, EZFS_BADWHO, errbuf)); + } + + /* + * add entries for both local and descendent when required + */ + zfs_perms_add_who_nvlist(who_nvp, who_id, who_tok, + perms_nvp, sets_nvp, who_type, inherit); + + } while (who_tok = strtok(NULL, ",")); + *nvp = who_nvp; + return (0); +} + +static int +zfs_perm_set_common(zfs_handle_t *zhp, nvlist_t *nvp, boolean_t unset) +{ + zfs_cmd_t zc = { 0 }; + int error; + char errbuf[1024]; + + (void) snprintf(errbuf, sizeof (errbuf), + dgettext(TEXT_DOMAIN, "Cannot update 'allows' for '%s'"), + zhp->zfs_name); + + if (zcmd_write_src_nvlist(zhp->zfs_hdl, &zc, nvp)) + return (-1); + + (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); + zc.zc_perm_action = unset; + + error = zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_SET_FSACL, &zc); + if (error && errno == ENOTSUP) { + (void) snprintf(errbuf, sizeof (errbuf), + gettext("Pool must be upgraded to use 'allow/unallow'")); + zcmd_free_nvlists(&zc); + return (zfs_error(zhp->zfs_hdl, EZFS_BADVERSION, errbuf)); + } else if (error) { + return (zfs_standard_error(zhp->zfs_hdl, errno, errbuf)); + } + zcmd_free_nvlists(&zc); + + return (error); +} + +int +zfs_perm_set(zfs_handle_t *zhp, nvlist_t *nvp) +{ + return (zfs_perm_set_common(zhp, nvp, B_FALSE)); +} + +int +zfs_perm_remove(zfs_handle_t *zhp, nvlist_t *perms) +{ + return (zfs_perm_set_common(zhp, perms, B_TRUE)); +} + +static int +perm_compare(const void *arg1, const void *arg2) +{ + const zfs_perm_node_t *node1 = arg1; + const zfs_perm_node_t *node2 = arg2; + int ret; + + ret = strcmp(node1->z_pname, node2->z_pname); + + if (ret > 0) + return (1); + if (ret < 0) + return (-1); + else + return (0); +} + +static void +zfs_destroy_perm_tree(avl_tree_t *tree) +{ + zfs_perm_node_t *permnode; + void *cookie = NULL; + + while ((permnode = avl_destroy_nodes(tree, &cookie)) != NULL) + free(permnode); + avl_destroy(tree); +} + +static void +zfs_destroy_tree(avl_tree_t *tree) +{ + zfs_allow_node_t *allownode; + void *cookie = NULL; + + while ((allownode = avl_destroy_nodes(tree, &cookie)) != NULL) { + zfs_destroy_perm_tree(&allownode->z_localdescend); + zfs_destroy_perm_tree(&allownode->z_local); + zfs_destroy_perm_tree(&allownode->z_descend); + free(allownode); + } + avl_destroy(tree); +} + +void +zfs_free_allows(zfs_allow_t *allow) +{ + zfs_allow_t *allownext; + zfs_allow_t *freeallow; + + allownext = allow; + while (allownext) { + zfs_destroy_tree(&allownext->z_sets); + zfs_destroy_tree(&allownext->z_crperms); + zfs_destroy_tree(&allownext->z_user); + zfs_destroy_tree(&allownext->z_group); + zfs_destroy_tree(&allownext->z_everyone); + freeallow = allownext; + allownext = allownext->z_next; + free(freeallow); + } +} + +static zfs_allow_t * +zfs_alloc_perm_tree(zfs_handle_t *zhp, zfs_allow_t *prev, char *setpoint) +{ + zfs_allow_t *ptree; + + if ((ptree = zfs_alloc(zhp->zfs_hdl, + sizeof (zfs_allow_t))) == NULL) { + return (NULL); + } + + (void) strlcpy(ptree->z_setpoint, setpoint, sizeof (ptree->z_setpoint)); + avl_create(&ptree->z_sets, + perm_compare, sizeof (zfs_allow_node_t), + offsetof(zfs_allow_node_t, z_node)); + avl_create(&ptree->z_crperms, + perm_compare, sizeof (zfs_allow_node_t), + offsetof(zfs_allow_node_t, z_node)); + avl_create(&ptree->z_user, + perm_compare, sizeof (zfs_allow_node_t), + offsetof(zfs_allow_node_t, z_node)); + avl_create(&ptree->z_group, + perm_compare, sizeof (zfs_allow_node_t), + offsetof(zfs_allow_node_t, z_node)); + avl_create(&ptree->z_everyone, + perm_compare, sizeof (zfs_allow_node_t), + offsetof(zfs_allow_node_t, z_node)); + + if (prev) + prev->z_next = ptree; + ptree->z_next = NULL; + return (ptree); +} + +/* + * Add permissions to the appropriate AVL permission tree. + * The appropriate tree may not be the requested tree. + * For example if ld indicates a local permission, but + * same permission also exists as a descendent permission + * then the permission will be removed from the descendent + * tree and add the the local+descendent tree. + */ +static int +zfs_coalesce_perm(zfs_handle_t *zhp, zfs_allow_node_t *allownode, + char *perm, char ld) +{ + zfs_perm_node_t pnode, *permnode, *permnode2; + zfs_perm_node_t *newnode; + avl_index_t where, where2; + avl_tree_t *tree, *altree; + + (void) strlcpy(pnode.z_pname, perm, sizeof (pnode.z_pname)); + + if (ld == ZFS_DELEG_NA) { + tree = &allownode->z_localdescend; + altree = &allownode->z_descend; + } else if (ld == ZFS_DELEG_LOCAL) { + tree = &allownode->z_local; + altree = &allownode->z_descend; + } else { + tree = &allownode->z_descend; + altree = &allownode->z_local; + } + permnode = avl_find(tree, &pnode, &where); + permnode2 = avl_find(altree, &pnode, &where2); + + if (permnode2) { + avl_remove(altree, permnode2); + free(permnode2); + if (permnode == NULL) { + tree = &allownode->z_localdescend; + } + } + + /* + * Now insert new permission in either requested location + * local/descendent or into ld when perm will exist in both. + */ + if (permnode == NULL) { + if ((newnode = zfs_alloc(zhp->zfs_hdl, + sizeof (zfs_perm_node_t))) == NULL) { + return (-1); + } + *newnode = pnode; + avl_add(tree, newnode); + } + return (0); +} + +/* + * Uggh, this is going to be a bit complicated. + * we have an nvlist coming out of the kernel that + * will indicate where the permission is set and then + * it will contain allow of the various "who's", and what + * their permissions are. To further complicate this + * we will then have to coalesce the local,descendent + * and local+descendent permissions where appropriate. + * The kernel only knows about a permission as being local + * or descendent, but not both. + * + * In order to make this easier for zfs_main to deal with + * a series of AVL trees will be used to maintain + * all of this, primarily for sorting purposes as well + * as the ability to quickly locate a specific entry. + * + * What we end up with are tree's for sets, create perms, + * user, groups and everyone. With each of those trees + * we have subtrees for local, descendent and local+descendent + * permissions. + */ +int +zfs_perm_get(zfs_handle_t *zhp, zfs_allow_t **zfs_perms) +{ + zfs_cmd_t zc = { 0 }; + int error; + nvlist_t *nvlist; + nvlist_t *permnv, *sourcenv; + nvpair_t *who_pair, *source_pair; + nvpair_t *perm_pair; + char errbuf[1024]; + zfs_allow_t *zallowp, *newallowp; + char ld; + char *nvpname; + uid_t uid; + gid_t gid; + avl_tree_t *tree; + avl_index_t where; + + (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); + + if (zcmd_alloc_dst_nvlist(zhp->zfs_hdl, &zc, 0) != 0) + return (-1); + + while (ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_GET_FSACL, &zc) != 0) { + if (errno == ENOMEM) { + if (zcmd_expand_dst_nvlist(zhp->zfs_hdl, &zc) != 0) { + zcmd_free_nvlists(&zc); + return (-1); + } + } else if (errno == ENOTSUP) { + zcmd_free_nvlists(&zc); + (void) snprintf(errbuf, sizeof (errbuf), + gettext("Pool must be upgraded to use 'allow'")); + return (zfs_error(zhp->zfs_hdl, + EZFS_BADVERSION, errbuf)); + } else { + zcmd_free_nvlists(&zc); + return (-1); + } + } + + if (zcmd_read_dst_nvlist(zhp->zfs_hdl, &zc, &nvlist) != 0) { + zcmd_free_nvlists(&zc); + return (-1); + } + + zcmd_free_nvlists(&zc); + + source_pair = nvlist_next_nvpair(nvlist, NULL); + + if (source_pair == NULL) { + *zfs_perms = NULL; + return (0); + } + + *zfs_perms = zfs_alloc_perm_tree(zhp, NULL, nvpair_name(source_pair)); + if (*zfs_perms == NULL) { + return (0); + } + + zallowp = *zfs_perms; + + for (;;) { + struct passwd *pwd; + struct group *grp; + zfs_allow_node_t *allownode; + zfs_allow_node_t findallownode; + zfs_allow_node_t *newallownode; + + (void) strlcpy(zallowp->z_setpoint, + nvpair_name(source_pair), + sizeof (zallowp->z_setpoint)); + + if ((error = nvpair_value_nvlist(source_pair, &sourcenv)) != 0) + goto abort; + + /* + * Make sure nvlist is composed correctly + */ + if (zfs_deleg_verify_nvlist(sourcenv)) { + goto abort; + } + + who_pair = nvlist_next_nvpair(sourcenv, NULL); + if (who_pair == NULL) { + goto abort; + } + + do { + error = nvpair_value_nvlist(who_pair, &permnv); + if (error) { + goto abort; + } + + /* + * First build up the key to use + * for looking up in the various + * who trees. + */ + ld = nvpair_name(who_pair)[1]; + nvpname = nvpair_name(who_pair); + switch (nvpair_name(who_pair)[0]) { + case ZFS_DELEG_USER: + case ZFS_DELEG_USER_SETS: + tree = &zallowp->z_user; + uid = atol(&nvpname[3]); + pwd = getpwuid(uid); + (void) snprintf(findallownode.z_key, + sizeof (findallownode.z_key), "user %s", + (pwd) ? pwd->pw_name : + &nvpair_name(who_pair)[3]); + break; + case ZFS_DELEG_GROUP: + case ZFS_DELEG_GROUP_SETS: + tree = &zallowp->z_group; + gid = atol(&nvpname[3]); + grp = getgrgid(gid); + (void) snprintf(findallownode.z_key, + sizeof (findallownode.z_key), "group %s", + (grp) ? grp->gr_name : + &nvpair_name(who_pair)[3]); + break; + case ZFS_DELEG_CREATE: + case ZFS_DELEG_CREATE_SETS: + tree = &zallowp->z_crperms; + (void) strlcpy(findallownode.z_key, "", + sizeof (findallownode.z_key)); + break; + case ZFS_DELEG_EVERYONE: + case ZFS_DELEG_EVERYONE_SETS: + (void) snprintf(findallownode.z_key, + sizeof (findallownode.z_key), "everyone"); + tree = &zallowp->z_everyone; + break; + case ZFS_DELEG_NAMED_SET: + case ZFS_DELEG_NAMED_SET_SETS: + (void) snprintf(findallownode.z_key, + sizeof (findallownode.z_key), "%s", + &nvpair_name(who_pair)[3]); + tree = &zallowp->z_sets; + break; + } + + /* + * Place who in tree + */ + allownode = avl_find(tree, &findallownode, &where); + if (allownode == NULL) { + if ((newallownode = zfs_alloc(zhp->zfs_hdl, + sizeof (zfs_allow_node_t))) == NULL) { + goto abort; + } + avl_create(&newallownode->z_localdescend, + perm_compare, + sizeof (zfs_perm_node_t), + offsetof(zfs_perm_node_t, z_node)); + avl_create(&newallownode->z_local, + perm_compare, + sizeof (zfs_perm_node_t), + offsetof(zfs_perm_node_t, z_node)); + avl_create(&newallownode->z_descend, + perm_compare, + sizeof (zfs_perm_node_t), + offsetof(zfs_perm_node_t, z_node)); + (void) strlcpy(newallownode->z_key, + findallownode.z_key, + sizeof (findallownode.z_key)); + avl_insert(tree, newallownode, where); + allownode = newallownode; + } + + /* + * Now iterate over the permissions and + * place them in the appropriate local, + * descendent or local+descendent tree. + * + * The permissions are added to the tree + * via zfs_coalesce_perm(). + */ + perm_pair = nvlist_next_nvpair(permnv, NULL); + if (perm_pair == NULL) + goto abort; + do { + if (zfs_coalesce_perm(zhp, allownode, + nvpair_name(perm_pair), ld) != 0) + goto abort; + } while (perm_pair = nvlist_next_nvpair(permnv, + perm_pair)); + } while (who_pair = nvlist_next_nvpair(sourcenv, who_pair)); + + source_pair = nvlist_next_nvpair(nvlist, source_pair); + if (source_pair == NULL) + break; + + /* + * allocate another node from the link list of + * zfs_allow_t structures + */ + newallowp = zfs_alloc_perm_tree(zhp, zallowp, + nvpair_name(source_pair)); + if (newallowp == NULL) { + goto abort; + } + zallowp = newallowp; + } + nvlist_free(nvlist); + return (0); +abort: + zfs_free_allows(*zfs_perms); + nvlist_free(nvlist); + return (-1); +} + +static char * +zfs_deleg_perm_note(zfs_deleg_note_t note) +{ + /* + * Don't put newlines on end of lines + */ + switch (note) { + case ZFS_DELEG_NOTE_CREATE: + return (dgettext(TEXT_DOMAIN, + "Must also have the 'mount' ability")); + case ZFS_DELEG_NOTE_DESTROY: + return (dgettext(TEXT_DOMAIN, + "Must also have the 'mount' ability")); + case ZFS_DELEG_NOTE_SNAPSHOT: + return (dgettext(TEXT_DOMAIN, + "Must also have the 'mount' ability")); + case ZFS_DELEG_NOTE_ROLLBACK: + return (dgettext(TEXT_DOMAIN, + "Must also have the 'mount' ability")); + case ZFS_DELEG_NOTE_CLONE: + return (dgettext(TEXT_DOMAIN, "Must also have the 'create' " + "ability and 'mount'\n" + "\t\t\t\tability in the origin file system")); + case ZFS_DELEG_NOTE_PROMOTE: + return (dgettext(TEXT_DOMAIN, "Must also have the 'mount'\n" + "\t\t\t\tand 'promote' ability in the origin file system")); + case ZFS_DELEG_NOTE_RENAME: + return (dgettext(TEXT_DOMAIN, "Must also have the 'mount' " + "and 'create' \n\t\t\t\tability in the new parent")); + case ZFS_DELEG_NOTE_RECEIVE: + return (dgettext(TEXT_DOMAIN, "Must also have the 'mount'" + " and 'create' ability")); + case ZFS_DELEG_NOTE_USERPROP: + return (dgettext(TEXT_DOMAIN, + "Allows changing any user property")); + case ZFS_DELEG_NOTE_ALLOW: + return (dgettext(TEXT_DOMAIN, + "Must also have the permission that is being\n" + "\t\t\t\tallowed")); + case ZFS_DELEG_NOTE_MOUNT: + return (dgettext(TEXT_DOMAIN, + "Allows mount/umount of ZFS datasets")); + case ZFS_DELEG_NOTE_SHARE: + return (dgettext(TEXT_DOMAIN, + "Allows sharing file systems over NFS or SMB\n" + "\t\t\t\tprotocols")); + case ZFS_DELEG_NOTE_NONE: + default: + return (dgettext(TEXT_DOMAIN, "")); + } +} + +typedef enum { + ZFS_DELEG_SUBCOMMAND, + ZFS_DELEG_PROP, + ZFS_DELEG_OTHER +} zfs_deleg_perm_type_t; + +/* + * is the permission a subcommand or other? + */ +zfs_deleg_perm_type_t +zfs_deleg_perm_type(const char *perm) +{ + if (strcmp(perm, "userprop") == 0) + return (ZFS_DELEG_OTHER); + else + return (ZFS_DELEG_SUBCOMMAND); +} + +static char * +zfs_deleg_perm_type_str(zfs_deleg_perm_type_t type) +{ + switch (type) { + case ZFS_DELEG_SUBCOMMAND: + return (dgettext(TEXT_DOMAIN, "subcommand")); + case ZFS_DELEG_PROP: + return (dgettext(TEXT_DOMAIN, "property")); + case ZFS_DELEG_OTHER: + return (dgettext(TEXT_DOMAIN, "other")); + } + return (""); +} + +/*ARGSUSED*/ +static int +zfs_deleg_prop_cb(int prop, void *cb) +{ + if (zfs_prop_delegatable(prop)) + (void) fprintf(stderr, "%-15s %-15s\n", zfs_prop_to_name(prop), + zfs_deleg_perm_type_str(ZFS_DELEG_PROP)); + + return (ZPROP_CONT); +} + +void +zfs_deleg_permissions(void) +{ + int i; + + (void) fprintf(stderr, "\n%-15s %-15s\t%s\n\n", "NAME", + "TYPE", "NOTES"); + + /* + * First print out the subcommands + */ + for (i = 0; zfs_deleg_perm_tab[i].z_perm != NULL; i++) { + (void) fprintf(stderr, "%-15s %-15s\t%s\n", + zfs_deleg_perm_tab[i].z_perm, + zfs_deleg_perm_type_str( + zfs_deleg_perm_type(zfs_deleg_perm_tab[i].z_perm)), + zfs_deleg_perm_note(zfs_deleg_perm_tab[i].z_note)); + } + + (void) zprop_iter(zfs_deleg_prop_cb, NULL, B_FALSE, B_TRUE, + ZFS_TYPE_DATASET|ZFS_TYPE_VOLUME); +} + /* * Given a property name and value, set the property for the given dataset. */ @@ -1070,6 +1766,8 @@ zfs_prop_set(zfs_handle_t *zhp, const char *propname, const char *propval) libzfs_handle_t *hdl = zhp->zfs_hdl; nvlist_t *nvl = NULL, *realprops; zfs_prop_t prop; + boolean_t do_prefix; + uint64_t idx; (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot set property for '%s'"), @@ -1081,9 +1779,10 @@ zfs_prop_set(zfs_handle_t *zhp, const char *propname, const char *propval) goto error; } - if ((realprops = zfs_validate_properties(hdl, zhp->zfs_type, NULL, nvl, + if ((realprops = zfs_valid_proplist(hdl, zhp->zfs_type, nvl, zfs_prop_get_int(zhp, ZFS_PROP_ZONED), zhp, errbuf)) == NULL) goto error; + nvlist_free(nvl); nvl = realprops; @@ -1102,7 +1801,7 @@ zfs_prop_set(zfs_handle_t *zhp, const char *propname, const char *propval) goto error; } - if ((cl = changelist_gather(zhp, prop, 0)) == NULL) + if ((cl = changelist_gather(zhp, prop, 0, 0)) == NULL) goto error; if (prop == ZFS_PROP_MOUNTPOINT && changelist_haszonedchild(cl)) { @@ -1113,7 +1812,15 @@ zfs_prop_set(zfs_handle_t *zhp, const char *propname, const char *propval) goto error; } - if ((ret = changelist_prefix(cl)) != 0) + /* + * If the dataset's canmount property is being set to noauto, + * then we want to prevent unmounting & remounting it. + */ + do_prefix = !((prop == ZFS_PROP_CANMOUNT) && + (zprop_string_to_index(prop, propval, &idx, + ZFS_TYPE_DATASET) == 0) && (idx == ZFS_CANMOUNT_NOAUTO)); + + if (do_prefix && (ret = changelist_prefix(cl)) != 0) goto error; /* @@ -1121,11 +1828,10 @@ zfs_prop_set(zfs_handle_t *zhp, const char *propname, const char *propval) */ (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); - if (zcmd_write_src_nvlist(hdl, &zc, nvl, NULL) != 0) + if (zcmd_write_src_nvlist(hdl, &zc, nvl) != 0) goto error; - ret = ioctl(hdl->libzfs_fd, ZFS_IOC_SET_PROP, &zc); - + ret = zfs_ioctl(hdl, ZFS_IOC_SET_PROP, &zc); if (ret != 0) { switch (errno) { @@ -1137,6 +1843,7 @@ zfs_prop_set(zfs_handle_t *zhp, const char *propname, const char *propval) */ switch (prop) { case ZFS_PROP_QUOTA: + case ZFS_PROP_REFQUOTA: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "size is less than current used or " "reserved space")); @@ -1144,6 +1851,7 @@ zfs_prop_set(zfs_handle_t *zhp, const char *propname, const char *propval) break; case ZFS_PROP_RESERVATION: + case ZFS_PROP_REFRESERVATION: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "size is greater than available space")); (void) zfs_error(hdl, EZFS_PROPSPACE, errbuf); @@ -1168,10 +1876,22 @@ zfs_prop_set(zfs_handle_t *zhp, const char *propname, const char *propval) case ENOTSUP: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "pool must be upgraded to allow gzip compression")); + "pool and or dataset must be upgraded to set this " + "property or value")); (void) zfs_error(hdl, EZFS_BADVERSION, errbuf); break; + case ERANGE: + if (prop == ZFS_PROP_COMPRESSION) { + (void) zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "property setting is not allowed on " + "bootable datasets")); + (void) zfs_error(hdl, EZFS_NOTSUP, errbuf); + } else { + (void) zfs_standard_error(hdl, errno, errbuf); + } + break; + case EOVERFLOW: /* * This platform can't address a volume this big. @@ -1187,11 +1907,14 @@ zfs_prop_set(zfs_handle_t *zhp, const char *propname, const char *propval) (void) zfs_standard_error(hdl, errno, errbuf); } } else { + if (do_prefix) + ret = changelist_postfix(cl); + /* * Refresh the statistics so the new property value * is reflected. */ - if ((ret = changelist_postfix(cl)) == 0) + if (ret == 0) (void) get_stats(zhp); } @@ -1219,7 +1942,7 @@ zfs_prop_inherit(zfs_handle_t *zhp, const char *propname) (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, "cannot inherit %s for '%s'"), propname, zhp->zfs_name); - if ((prop = zfs_name_to_prop(propname)) == ZFS_PROP_INVAL) { + if ((prop = zfs_name_to_prop(propname)) == ZPROP_INVAL) { /* * For user properties, the amount of work we have to do is very * small, so just do it here. @@ -1233,8 +1956,7 @@ zfs_prop_inherit(zfs_handle_t *zhp, const char *propname) (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); (void) strlcpy(zc.zc_value, propname, sizeof (zc.zc_value)); - if (ioctl(zhp->zfs_hdl->libzfs_fd, - ZFS_IOC_SET_PROP, &zc) != 0) + if (zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_INHERIT_PROP, &zc) != 0) return (zfs_standard_error(hdl, errno, errbuf)); return (0); @@ -1272,7 +1994,7 @@ zfs_prop_inherit(zfs_handle_t *zhp, const char *propname) /* * Determine datasets which will be affected by this change, if any. */ - if ((cl = changelist_gather(zhp, prop, 0)) == NULL) + if ((cl = changelist_gather(zhp, prop, 0, 0)) == NULL) return (-1); if (prop == ZFS_PROP_MOUNTPOINT && changelist_haszonedchild(cl)) { @@ -1286,8 +2008,7 @@ zfs_prop_inherit(zfs_handle_t *zhp, const char *propname) if ((ret = changelist_prefix(cl)) != 0) goto error; - if ((ret = ioctl(zhp->zfs_hdl->libzfs_fd, - ZFS_IOC_SET_PROP, &zc)) != 0) { + if ((ret = zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_INHERIT_PROP, &zc)) != 0) { return (zfs_standard_error(hdl, errno, errbuf)); } else { @@ -1305,15 +2026,6 @@ error: return (ret); } -void -nicebool(int value, char *buf, size_t buflen) -{ - if (value) - (void) strlcpy(buf, "on", buflen); - else - (void) strlcpy(buf, "off", buflen); -} - /* * True DSL properties are stored in an nvlist. The following two functions * extract them appropriately. @@ -1327,8 +2039,8 @@ getprop_uint64(zfs_handle_t *zhp, zfs_prop_t prop, char **source) *source = NULL; if (nvlist_lookup_nvlist(zhp->zfs_props, zfs_prop_to_name(prop), &nv) == 0) { - verify(nvlist_lookup_uint64(nv, ZFS_PROP_VALUE, &value) == 0); - (void) nvlist_lookup_string(nv, ZFS_PROP_SOURCE, source); + verify(nvlist_lookup_uint64(nv, ZPROP_VALUE, &value) == 0); + (void) nvlist_lookup_string(nv, ZPROP_SOURCE, source); } else { value = zfs_prop_default_numeric(prop); *source = ""; @@ -1346,8 +2058,8 @@ getprop_string(zfs_handle_t *zhp, zfs_prop_t prop, char **source) *source = NULL; if (nvlist_lookup_nvlist(zhp->zfs_props, zfs_prop_to_name(prop), &nv) == 0) { - verify(nvlist_lookup_string(nv, ZFS_PROP_VALUE, &value) == 0); - (void) nvlist_lookup_string(nv, ZFS_PROP_SOURCE, source); + verify(nvlist_lookup_string(nv, ZPROP_VALUE, &value) == 0); + (void) nvlist_lookup_string(nv, ZPROP_SOURCE, source); } else { if ((value = (char *)zfs_prop_default_string(prop)) == NULL) value = ""; @@ -1367,9 +2079,11 @@ getprop_string(zfs_handle_t *zhp, zfs_prop_t prop, char **source) * the source "temporary". */ static int -get_numeric_property(zfs_handle_t *zhp, zfs_prop_t prop, zfs_source_t *src, +get_numeric_property(zfs_handle_t *zhp, zfs_prop_t prop, zprop_source_t *src, char **source, uint64_t *val) { + zfs_cmd_t zc = { 0 }; + nvlist_t *zplprops = NULL; struct mnttab mnt; char *mntopt_on = NULL; char *mntopt_off = NULL; @@ -1406,6 +2120,11 @@ get_numeric_property(zfs_handle_t *zhp, zfs_prop_t prop, zfs_source_t *src, mntopt_on = MNTOPT_XATTR; mntopt_off = MNTOPT_NOXATTR; break; + + case ZFS_PROP_NBMAND: + mntopt_on = MNTOPT_NBMAND; + mntopt_off = MNTOPT_NONBMAND; + break; } /* @@ -1444,43 +2163,32 @@ get_numeric_property(zfs_handle_t *zhp, zfs_prop_t prop, zfs_source_t *src, case ZFS_PROP_READONLY: case ZFS_PROP_SETUID: case ZFS_PROP_XATTR: + case ZFS_PROP_NBMAND: *val = getprop_uint64(zhp, prop, source); if (hasmntopt(&mnt, mntopt_on) && !*val) { *val = B_TRUE; if (src) - *src = ZFS_SRC_TEMPORARY; + *src = ZPROP_SRC_TEMPORARY; } else if (hasmntopt(&mnt, mntopt_off) && *val) { *val = B_FALSE; if (src) - *src = ZFS_SRC_TEMPORARY; + *src = ZPROP_SRC_TEMPORARY; } break; - case ZFS_PROP_RECORDSIZE: - case ZFS_PROP_COMPRESSION: - case ZFS_PROP_ZONED: - case ZFS_PROP_CREATION: - case ZFS_PROP_COMPRESSRATIO: - case ZFS_PROP_REFERENCED: - case ZFS_PROP_USED: - case ZFS_PROP_CREATETXG: - case ZFS_PROP_AVAILABLE: - case ZFS_PROP_VOLSIZE: - case ZFS_PROP_VOLBLOCKSIZE: - *val = getprop_uint64(zhp, prop, source); - break; - case ZFS_PROP_CANMOUNT: *val = getprop_uint64(zhp, prop, source); - if (*val == 0) + if (*val != ZFS_CANMOUNT_ON) *source = zhp->zfs_name; else *source = ""; /* default */ break; case ZFS_PROP_QUOTA: + case ZFS_PROP_REFQUOTA: case ZFS_PROP_RESERVATION: + case ZFS_PROP_REFRESERVATION: *val = getprop_uint64(zhp, prop, source); if (*val == 0) *source = ""; /* default */ @@ -1496,11 +2204,60 @@ get_numeric_property(zfs_handle_t *zhp, zfs_prop_t prop, zfs_source_t *src, *val = zhp->zfs_dmustats.dds_num_clones; break; + case ZFS_PROP_VERSION: + case ZFS_PROP_NORMALIZE: + case ZFS_PROP_UTF8ONLY: + case ZFS_PROP_CASE: + if (!zfs_prop_valid_for_type(prop, zhp->zfs_head_type) || + zcmd_alloc_dst_nvlist(zhp->zfs_hdl, &zc, 0) != 0) + return (-1); + (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); + if (zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_OBJSET_ZPLPROPS, &zc)) { + zcmd_free_nvlists(&zc); + zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN, + "unable to get %s property"), + zfs_prop_to_name(prop)); + return (zfs_error(zhp->zfs_hdl, EZFS_BADVERSION, + dgettext(TEXT_DOMAIN, "internal error"))); + } + if (zcmd_read_dst_nvlist(zhp->zfs_hdl, &zc, &zplprops) != 0 || + nvlist_lookup_uint64(zplprops, zfs_prop_to_name(prop), + val) != 0) { + zcmd_free_nvlists(&zc); + zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN, + "unable to get %s property"), + zfs_prop_to_name(prop)); + return (zfs_error(zhp->zfs_hdl, EZFS_NOMEM, + dgettext(TEXT_DOMAIN, "internal error"))); + } + if (zplprops) + nvlist_free(zplprops); + zcmd_free_nvlists(&zc); + break; + default: - zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN, - "cannot get non-numeric property")); - return (zfs_error(zhp->zfs_hdl, EZFS_BADPROP, - dgettext(TEXT_DOMAIN, "internal error"))); + switch (zfs_prop_get_type(prop)) { + case PROP_TYPE_NUMBER: + case PROP_TYPE_INDEX: + *val = getprop_uint64(zhp, prop, source); + /* + * If we tried to use a defalut value for a + * readonly property, it means that it was not + * present; return an error. + */ + if (zfs_prop_readonly(prop) && + *source && (*source)[0] == '\0') { + return (-1); + } + break; + + case PROP_TYPE_STRING: + default: + zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN, + "cannot get non-numeric property")); + return (zfs_error(zhp->zfs_hdl, EZFS_BADPROP, + dgettext(TEXT_DOMAIN, "internal error"))); + } } return (0); @@ -1510,22 +2267,22 @@ get_numeric_property(zfs_handle_t *zhp, zfs_prop_t prop, zfs_source_t *src, * Calculate the source type, given the raw source string. */ static void -get_source(zfs_handle_t *zhp, zfs_source_t *srctype, char *source, +get_source(zfs_handle_t *zhp, zprop_source_t *srctype, char *source, char *statbuf, size_t statlen) { - if (statbuf == NULL || *srctype == ZFS_SRC_TEMPORARY) + if (statbuf == NULL || *srctype == ZPROP_SRC_TEMPORARY) return; if (source == NULL) { - *srctype = ZFS_SRC_NONE; + *srctype = ZPROP_SRC_NONE; } else if (source[0] == '\0') { - *srctype = ZFS_SRC_DEFAULT; + *srctype = ZPROP_SRC_DEFAULT; } else { if (strcmp(source, zhp->zfs_name) == 0) { - *srctype = ZFS_SRC_LOCAL; + *srctype = ZPROP_SRC_LOCAL; } else { (void) strlcpy(statbuf, source, statlen); - *srctype = ZFS_SRC_INHERITED; + *srctype = ZPROP_SRC_INHERITED; } } @@ -1540,12 +2297,11 @@ get_source(zfs_handle_t *zhp, zfs_source_t *srctype, char *source, */ int zfs_prop_get(zfs_handle_t *zhp, zfs_prop_t prop, char *propbuf, size_t proplen, - zfs_source_t *src, char *statbuf, size_t statlen, boolean_t literal) + zprop_source_t *src, char *statbuf, size_t statlen, boolean_t literal) { char *source = NULL; uint64_t val; char *str; - const char *root; const char *strval; /* @@ -1555,66 +2311,9 @@ zfs_prop_get(zfs_handle_t *zhp, zfs_prop_t prop, char *propbuf, size_t proplen, return (-1); if (src) - *src = ZFS_SRC_NONE; + *src = ZPROP_SRC_NONE; switch (prop) { - case ZFS_PROP_ATIME: - case ZFS_PROP_READONLY: - case ZFS_PROP_SETUID: - case ZFS_PROP_ZONED: - case ZFS_PROP_DEVICES: - case ZFS_PROP_EXEC: - case ZFS_PROP_CANMOUNT: - case ZFS_PROP_XATTR: - /* - * Basic boolean values are built on top of - * get_numeric_property(). - */ - if (get_numeric_property(zhp, prop, src, &source, &val) != 0) - return (-1); - nicebool(val, propbuf, proplen); - - break; - - case ZFS_PROP_AVAILABLE: - case ZFS_PROP_RECORDSIZE: - case ZFS_PROP_CREATETXG: - case ZFS_PROP_REFERENCED: - case ZFS_PROP_USED: - case ZFS_PROP_VOLSIZE: - case ZFS_PROP_VOLBLOCKSIZE: - case ZFS_PROP_NUMCLONES: - /* - * Basic numeric values are built on top of - * get_numeric_property(). - */ - if (get_numeric_property(zhp, prop, src, &source, &val) != 0) - return (-1); - if (literal) - (void) snprintf(propbuf, proplen, "%llu", - (u_longlong_t)val); - else - zfs_nicenum(val, propbuf, proplen); - break; - - case ZFS_PROP_COMPRESSION: - case ZFS_PROP_CHECKSUM: - case ZFS_PROP_SNAPDIR: -#ifdef ZFS_NO_ACL - case ZFS_PROP_ACLMODE: - case ZFS_PROP_ACLINHERIT: - case ZFS_PROP_COPIES: - val = getprop_uint64(zhp, prop, &source); - verify(zfs_prop_index_to_string(prop, val, &strval) == 0); - (void) strlcpy(propbuf, strval, proplen); - break; -#else /* ZFS_NO_ACL */ - case ZFS_PROP_ACLMODE: - case ZFS_PROP_ACLINHERIT: - (void) strlcpy(propbuf, "<unsupported>", proplen); - break; -#endif /* ZFS_NO_ACL */ - case ZFS_PROP_CREATION: /* * 'creation' is a time_t stored in the statistics. We convert @@ -1638,25 +2337,42 @@ zfs_prop_get(zfs_handle_t *zhp, zfs_prop_t prop, char *propbuf, size_t proplen, * Getting the precise mountpoint can be tricky. * * - for 'none' or 'legacy', return those values. - * - for default mountpoints, construct it as /zfs/<dataset> * - for inherited mountpoints, we want to take everything * after our ancestor and append it to the inherited value. * * If the pool has an alternate root, we want to prepend that * root to any values we return. */ - root = zhp->zfs_root; + str = getprop_string(zhp, prop, &source); - if (str[0] == '\0') { - (void) snprintf(propbuf, proplen, "%s/zfs/%s", - root, zhp->zfs_name); - } else if (str[0] == '/') { + if (str[0] == '/') { + char buf[MAXPATHLEN]; + char *root = buf; const char *relpath = zhp->zfs_name + strlen(source); if (relpath[0] == '/') relpath++; - if (str[1] == '\0') + + if ((zpool_get_prop(zhp->zpool_hdl, + ZPOOL_PROP_ALTROOT, buf, MAXPATHLEN, NULL)) || + (strcmp(root, "-") == 0)) + root[0] = '\0'; + /* + * Special case an alternate root of '/'. This will + * avoid having multiple leading slashes in the + * mountpoint path. + */ + if (strcmp(root, "/") == 0) + root++; + + /* + * If the mountpoint is '/' then skip over this + * if we are obtaining either an alternate root or + * an inherited mountpoint. + */ + if (str[1] == '\0' && (root[0] != '\0' || + relpath[0] != '\0')) str++; if (relpath[0] == '\0') @@ -1673,13 +2389,6 @@ zfs_prop_get(zfs_handle_t *zhp, zfs_prop_t prop, char *propbuf, size_t proplen, break; - case ZFS_PROP_SHARENFS: - case ZFS_PROP_SHAREISCSI: - case ZFS_PROP_ISCSIOPTIONS: - (void) strlcpy(propbuf, getprop_string(zhp, prop, &source), - proplen); - break; - case ZFS_PROP_ORIGIN: (void) strlcpy(propbuf, getprop_string(zhp, prop, &source), proplen); @@ -1692,7 +2401,10 @@ zfs_prop_get(zfs_handle_t *zhp, zfs_prop_t prop, char *propbuf, size_t proplen, break; case ZFS_PROP_QUOTA: + case ZFS_PROP_REFQUOTA: case ZFS_PROP_RESERVATION: + case ZFS_PROP_REFRESERVATION: + if (get_numeric_property(zhp, prop, src, &source, &val) != 0) return (-1); @@ -1766,7 +2478,35 @@ zfs_prop_get(zfs_handle_t *zhp, zfs_prop_t prop, char *propbuf, size_t proplen, break; default: - abort(); + switch (zfs_prop_get_type(prop)) { + case PROP_TYPE_NUMBER: + if (get_numeric_property(zhp, prop, src, + &source, &val) != 0) + return (-1); + if (literal) + (void) snprintf(propbuf, proplen, "%llu", + (u_longlong_t)val); + else + zfs_nicenum(val, propbuf, proplen); + break; + + case PROP_TYPE_STRING: + (void) strlcpy(propbuf, + getprop_string(zhp, prop, &source), proplen); + break; + + case PROP_TYPE_INDEX: + if (get_numeric_property(zhp, prop, src, + &source, &val) != 0) + return (-1); + if (zfs_prop_index_to_string(prop, val, &strval) != 0) + return (-1); + (void) strlcpy(propbuf, strval, proplen); + break; + + default: + abort(); + } } get_source(zhp, src, source, statbuf, statlen); @@ -1783,33 +2523,42 @@ uint64_t zfs_prop_get_int(zfs_handle_t *zhp, zfs_prop_t prop) { char *source; - zfs_source_t sourcetype = ZFS_SRC_NONE; uint64_t val; - (void) get_numeric_property(zhp, prop, &sourcetype, &source, &val); + (void) get_numeric_property(zhp, prop, NULL, &source, &val); return (val); } +int +zfs_prop_set_int(zfs_handle_t *zhp, zfs_prop_t prop, uint64_t val) +{ + char buf[64]; + + zfs_nicenum(val, buf, sizeof (buf)); + return (zfs_prop_set(zhp, zfs_prop_to_name(prop), buf)); +} + /* * Similar to zfs_prop_get(), but returns the value as an integer. */ int zfs_prop_get_numeric(zfs_handle_t *zhp, zfs_prop_t prop, uint64_t *value, - zfs_source_t *src, char *statbuf, size_t statlen) + zprop_source_t *src, char *statbuf, size_t statlen) { char *source; /* * Check to see if this property applies to our object */ - if (!zfs_prop_valid_for_type(prop, zhp->zfs_type)) + if (!zfs_prop_valid_for_type(prop, zhp->zfs_type)) { return (zfs_error_fmt(zhp->zfs_hdl, EZFS_PROPTYPE, dgettext(TEXT_DOMAIN, "cannot get property '%s'"), zfs_prop_to_name(prop))); + } if (src) - *src = ZFS_SRC_NONE; + *src = ZPROP_SRC_NONE; if (get_numeric_property(zhp, prop, src, &source, value) != 0) return (-1); @@ -1847,6 +2596,9 @@ zfs_iter_filesystems(zfs_handle_t *zhp, zfs_iter_f func, void *data) zfs_handle_t *nzhp; int ret; + if (zhp->zfs_type != ZFS_TYPE_FILESYSTEM) + return (0); + for ((void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_DATASET_LIST_NEXT, &zc) == 0; (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name))) { @@ -1890,6 +2642,9 @@ zfs_iter_snapshots(zfs_handle_t *zhp, zfs_iter_f func, void *data) zfs_handle_t *nzhp; int ret; + if (zhp->zfs_type == ZFS_TYPE_SNAPSHOT) + return (0); + for ((void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_SNAPSHOT_LIST_NEXT, &zc) == 0; @@ -1948,12 +2703,16 @@ parent_name(const char *path, char *buf, size_t buflen) } /* - * Checks to make sure that the given path has a parent, and that it exists. We - * also fetch the 'zoned' property, which is used to validate property settings - * when creating new datasets. + * If accept_ancestor is false, then check to make sure that the given path has + * a parent, and that it exists. If accept_ancestor is true, then find the + * closest existing ancestor for the given path. In prefixlen return the + * length of already existing prefix of the given path. We also fetch the + * 'zoned' property, which is used to validate property settings when creating + * new datasets. */ static int -check_parents(libzfs_handle_t *hdl, const char *path, uint64_t *zoned) +check_parents(libzfs_handle_t *hdl, const char *path, uint64_t *zoned, + boolean_t accept_ancestor, int *prefixlen) { zfs_cmd_t zc = { 0 }; char parent[ZFS_MAXNAMELEN]; @@ -1984,16 +2743,22 @@ check_parents(libzfs_handle_t *hdl, const char *path, uint64_t *zoned) } /* check to see if the parent dataset exists */ - if ((zhp = make_dataset_handle(hdl, parent)) == NULL) { - switch (errno) { - case ENOENT: + while ((zhp = make_dataset_handle(hdl, parent)) == NULL) { + if (errno == ENOENT && accept_ancestor) { + /* + * Go deeper to find an ancestor, give up on top level. + */ + if (parent_name(parent, parent, sizeof (parent)) != 0) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "no such pool '%s'"), zc.zc_name); + return (zfs_error(hdl, EZFS_NOENT, errbuf)); + } + } else if (errno == ENOENT) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "parent does not exist")); return (zfs_error(hdl, EZFS_NOENT, errbuf)); - - default: + } else return (zfs_standard_error(hdl, errno, errbuf)); - } } *zoned = zfs_prop_get_int(zhp, ZFS_PROP_ZONED); @@ -2014,6 +2779,136 @@ check_parents(libzfs_handle_t *hdl, const char *path, uint64_t *zoned) } zfs_close(zhp); + if (prefixlen != NULL) + *prefixlen = strlen(parent); + return (0); +} + +/* + * Finds whether the dataset of the given type(s) exists. + */ +boolean_t +zfs_dataset_exists(libzfs_handle_t *hdl, const char *path, zfs_type_t types) +{ + zfs_handle_t *zhp; + + if (!zfs_validate_name(hdl, path, types, B_FALSE)) + return (B_FALSE); + + /* + * Try to get stats for the dataset, which will tell us if it exists. + */ + if ((zhp = make_dataset_handle(hdl, path)) != NULL) { + int ds_type = zhp->zfs_type; + + zfs_close(zhp); + if (types & ds_type) + return (B_TRUE); + } + return (B_FALSE); +} + +/* + * Given a path to 'target', create all the ancestors between + * the prefixlen portion of the path, and the target itself. + * Fail if the initial prefixlen-ancestor does not already exist. + */ +int +create_parents(libzfs_handle_t *hdl, char *target, int prefixlen) +{ + zfs_handle_t *h; + char *cp; + const char *opname; + + /* make sure prefix exists */ + cp = target + prefixlen; + if (*cp != '/') { + assert(strchr(cp, '/') == NULL); + h = zfs_open(hdl, target, ZFS_TYPE_FILESYSTEM); + } else { + *cp = '\0'; + h = zfs_open(hdl, target, ZFS_TYPE_FILESYSTEM); + *cp = '/'; + } + if (h == NULL) + return (-1); + zfs_close(h); + + /* + * Attempt to create, mount, and share any ancestor filesystems, + * up to the prefixlen-long one. + */ + for (cp = target + prefixlen + 1; + cp = strchr(cp, '/'); *cp = '/', cp++) { + char *logstr; + + *cp = '\0'; + + h = make_dataset_handle(hdl, target); + if (h) { + /* it already exists, nothing to do here */ + zfs_close(h); + continue; + } + + logstr = hdl->libzfs_log_str; + hdl->libzfs_log_str = NULL; + if (zfs_create(hdl, target, ZFS_TYPE_FILESYSTEM, + NULL) != 0) { + hdl->libzfs_log_str = logstr; + opname = dgettext(TEXT_DOMAIN, "create"); + goto ancestorerr; + } + + hdl->libzfs_log_str = logstr; + h = zfs_open(hdl, target, ZFS_TYPE_FILESYSTEM); + if (h == NULL) { + opname = dgettext(TEXT_DOMAIN, "open"); + goto ancestorerr; + } + + if (zfs_mount(h, NULL, 0) != 0) { + opname = dgettext(TEXT_DOMAIN, "mount"); + goto ancestorerr; + } + + if (zfs_share(h) != 0) { + opname = dgettext(TEXT_DOMAIN, "share"); + goto ancestorerr; + } + + zfs_close(h); + } + + return (0); + +ancestorerr: + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "failed to %s ancestor '%s'"), opname, target); + return (-1); +} + +/* + * Creates non-existing ancestors of the given path. + */ +int +zfs_create_ancestors(libzfs_handle_t *hdl, const char *path) +{ + int prefix; + uint64_t zoned; + char *path_copy; + int rc; + + if (check_parents(hdl, path, &zoned, B_TRUE, &prefix) != 0) + return (-1); + + if ((path_copy = strdup(path)) != NULL) { + rc = create_parents(hdl, path_copy, prefix); + free(path_copy); + } + if (path_copy == NULL || rc != 0) + return (-1); + return (0); } @@ -2035,11 +2930,11 @@ zfs_create(libzfs_handle_t *hdl, const char *path, zfs_type_t type, "cannot create '%s'"), path); /* validate the path, taking care to note the extended error message */ - if (!zfs_validate_name(hdl, path, type)) + if (!zfs_validate_name(hdl, path, type, B_TRUE)) return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); /* validate parents exist */ - if (check_parents(hdl, path, &zoned) != 0) + if (check_parents(hdl, path, &zoned, B_FALSE, NULL) != 0) return (-1); /* @@ -2050,7 +2945,7 @@ zfs_create(libzfs_handle_t *hdl, const char *path, zfs_type_t type, * first try to see if the dataset exists. */ (void) strlcpy(zc.zc_name, path, sizeof (zc.zc_name)); - if (ioctl(hdl->libzfs_fd, ZFS_IOC_OBJSET_STATS, &zc) == 0) { + if (zfs_dataset_exists(hdl, zc.zc_name, ZFS_TYPE_DATASET)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "dataset already exists")); return (zfs_error(hdl, EZFS_EXISTS, errbuf)); @@ -2061,7 +2956,7 @@ zfs_create(libzfs_handle_t *hdl, const char *path, zfs_type_t type, else zc.zc_objset_type = DMU_OST_ZFS; - if (props && (props = zfs_validate_properties(hdl, type, NULL, props, + if (props && (props = zfs_valid_proplist(hdl, type, props, zoned, NULL, errbuf)) == 0) return (-1); @@ -2111,13 +3006,12 @@ zfs_create(libzfs_handle_t *hdl, const char *path, zfs_type_t type, } } - if (props && - zcmd_write_src_nvlist(hdl, &zc, props, NULL) != 0) + if (props && zcmd_write_src_nvlist(hdl, &zc, props) != 0) return (-1); nvlist_free(props); /* create the dataset */ - ret = ioctl(hdl->libzfs_fd, ZFS_IOC_CREATE, &zc); + ret = zfs_ioctl(hdl, ZFS_IOC_CREATE, &zc); if (ret == 0 && type == ZFS_TYPE_VOLUME) { ret = zvol_create_link(hdl, path); @@ -2158,6 +3052,11 @@ zfs_create(libzfs_handle_t *hdl, const char *path, zfs_type_t type, return (zfs_error(hdl, EZFS_BADPROP, errbuf)); + case ENOTSUP: + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "pool must be upgraded to set this " + "property or value")); + return (zfs_error(hdl, EZFS_BADVERSION, errbuf)); #ifdef _ILP32 case EOVERFLOW: /* @@ -2189,10 +3088,13 @@ zfs_destroy(zfs_handle_t *zhp) if (ZFS_IS_VOLUME(zhp)) { /* - * Unconditionally unshare this zvol ignoring failure as it - * indicates only that the volume wasn't shared initially. + * If user doesn't have permissions to unshare volume, then + * abort the request. This would only happen for a + * non-privileged user. */ - (void) zfs_unshare_iscsi(zhp); + if (zfs_unshare_iscsi(zhp) != 0) { + return (-1); + } if (zvol_remove_link(zhp->zfs_hdl, zhp->zfs_name) != 0) return (-1); @@ -2202,7 +3104,7 @@ zfs_destroy(zfs_handle_t *zhp) zc.zc_objset_type = DMU_OST_ZFS; } - if (ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_DESTROY, &zc) != 0) { + if (zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_DESTROY, &zc) != 0) { return (zfs_standard_error_fmt(zhp->zfs_hdl, errno, dgettext(TEXT_DOMAIN, "cannot destroy '%s'"), zhp->zfs_name)); @@ -2276,7 +3178,7 @@ zfs_destroy_snaps(zfs_handle_t *zhp, char *snapname) (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); (void) strlcpy(zc.zc_value, snapname, sizeof (zc.zc_value)); - ret = ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_DESTROY_SNAPS, &zc); + ret = zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_DESTROY_SNAPS, &zc); if (ret != 0) { char errbuf[1024]; @@ -2318,11 +3220,11 @@ zfs_clone(zfs_handle_t *zhp, const char *target, nvlist_t *props) "cannot create '%s'"), target); /* validate the target name */ - if (!zfs_validate_name(hdl, target, ZFS_TYPE_FILESYSTEM)) + if (!zfs_validate_name(hdl, target, ZFS_TYPE_FILESYSTEM, B_TRUE)) return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); /* validate parents exist */ - if (check_parents(hdl, target, &zoned) != 0) + if (check_parents(hdl, target, &zoned, B_FALSE, NULL) != 0) return (-1); (void) parent_name(target, parent, sizeof (parent)); @@ -2337,11 +3239,11 @@ zfs_clone(zfs_handle_t *zhp, const char *target, nvlist_t *props) } if (props) { - if ((props = zfs_validate_properties(hdl, type, NULL, props, - zoned, zhp, errbuf)) == NULL) + if ((props = zfs_valid_proplist(hdl, type, props, zoned, + zhp, errbuf)) == NULL) return (-1); - if (zcmd_write_src_nvlist(hdl, &zc, props, NULL) != 0) { + if (zcmd_write_src_nvlist(hdl, &zc, props) != 0) { nvlist_free(props); return (-1); } @@ -2351,7 +3253,7 @@ zfs_clone(zfs_handle_t *zhp, const char *target, nvlist_t *props) (void) strlcpy(zc.zc_name, target, sizeof (zc.zc_name)); (void) strlcpy(zc.zc_value, zhp->zfs_name, sizeof (zc.zc_value)); - ret = ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_CREATE, &zc); + ret = zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_CREATE, &zc); zcmd_free_nvlists(&zc); @@ -2470,7 +3372,7 @@ zfs_promote(zfs_handle_t *zhp) return (zfs_error(hdl, EZFS_BADTYPE, errbuf)); } - (void) strlcpy(parent, zhp->zfs_dmustats.dds_clone_of, sizeof (parent)); + (void) strlcpy(parent, zhp->zfs_dmustats.dds_origin, sizeof (parent)); if (parent[0] == '\0') { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "not a cloned filesystem")); @@ -2480,14 +3382,14 @@ zfs_promote(zfs_handle_t *zhp) *cp = '\0'; /* Walk the snapshots we will be moving */ - pzhp = zfs_open(hdl, zhp->zfs_dmustats.dds_clone_of, ZFS_TYPE_SNAPSHOT); + pzhp = zfs_open(hdl, zhp->zfs_dmustats.dds_origin, ZFS_TYPE_SNAPSHOT); if (pzhp == NULL) return (-1); pd.cb_pivot_txg = zfs_prop_get_int(pzhp, ZFS_PROP_CREATETXG); zfs_close(pzhp); pd.cb_target = zhp->zfs_name; pd.cb_errbuf = errbuf; - pzhp = zfs_open(hdl, parent, ZFS_TYPE_ANY); + pzhp = zfs_open(hdl, parent, ZFS_TYPE_DATASET); if (pzhp == NULL) return (-1); (void) zfs_prop_get(pzhp, ZFS_PROP_MOUNTPOINT, pd.cb_mountpoint, @@ -2499,10 +3401,10 @@ zfs_promote(zfs_handle_t *zhp) } /* issue the ioctl */ - (void) strlcpy(zc.zc_value, zhp->zfs_dmustats.dds_clone_of, + (void) strlcpy(zc.zc_value, zhp->zfs_dmustats.dds_origin, sizeof (zc.zc_value)); (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); - ret = ioctl(hdl->libzfs_fd, ZFS_IOC_PROMOTE, &zc); + ret = zfs_ioctl(hdl, ZFS_IOC_PROMOTE, &zc); if (ret != 0) { int save_errno = errno; @@ -2570,10 +3472,11 @@ zfs_create_link_cb(zfs_handle_t *zhp, void *arg) * Takes a snapshot of the given dataset. */ int -zfs_snapshot(libzfs_handle_t *hdl, const char *path, boolean_t recursive) +zfs_snapshot(libzfs_handle_t *hdl, const char *path, boolean_t recursive, + nvlist_t *props) { const char *delim; - char *parent; + char parent[ZFS_MAXNAMELEN]; zfs_handle_t *zhp; zfs_cmd_t zc = { 0 }; int ret; @@ -2583,33 +3486,52 @@ zfs_snapshot(libzfs_handle_t *hdl, const char *path, boolean_t recursive) "cannot snapshot '%s'"), path); /* validate the target name */ - if (!zfs_validate_name(hdl, path, ZFS_TYPE_SNAPSHOT)) + if (!zfs_validate_name(hdl, path, ZFS_TYPE_SNAPSHOT, B_TRUE)) return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); + if (props) { + if ((props = zfs_valid_proplist(hdl, ZFS_TYPE_SNAPSHOT, + props, B_FALSE, NULL, errbuf)) == NULL) + return (-1); + + if (zcmd_write_src_nvlist(hdl, &zc, props) != 0) { + nvlist_free(props); + return (-1); + } + + nvlist_free(props); + } + /* make sure the parent exists and is of the appropriate type */ delim = strchr(path, '@'); - if ((parent = zfs_alloc(hdl, delim - path + 1)) == NULL) - return (-1); (void) strncpy(parent, path, delim - path); parent[delim - path] = '\0'; if ((zhp = zfs_open(hdl, parent, ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME)) == NULL) { - free(parent); + zcmd_free_nvlists(&zc); return (-1); } (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); (void) strlcpy(zc.zc_value, delim+1, sizeof (zc.zc_value)); + if (ZFS_IS_VOLUME(zhp)) + zc.zc_objset_type = DMU_OST_ZVOL; + else + zc.zc_objset_type = DMU_OST_ZFS; zc.zc_cookie = recursive; - ret = ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_SNAPSHOT, &zc); + ret = zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_SNAPSHOT, &zc); + + zcmd_free_nvlists(&zc); /* * if it was recursive, the one that actually failed will be in * zc.zc_name. */ - (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, - "cannot create snapshot '%s@%s'"), zc.zc_name, zc.zc_value); + if (ret != 0) + (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, + "cannot create snapshot '%s@%s'"), zc.zc_name, zc.zc_value); + if (ret == 0 && recursive) { struct createdata cd; @@ -2620,408 +3542,24 @@ zfs_snapshot(libzfs_handle_t *hdl, const char *path, boolean_t recursive) if (ret == 0 && zhp->zfs_type == ZFS_TYPE_VOLUME) { ret = zvol_create_link(zhp->zfs_hdl, path); if (ret != 0) { - (void) ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_DESTROY, - &zc); + (void) zfs_standard_error(hdl, errno, + dgettext(TEXT_DOMAIN, + "Volume successfully snapshotted, but device links " + "were not created")); + zfs_close(zhp); + return (-1); } } if (ret != 0) (void) zfs_standard_error(hdl, errno, errbuf); - free(parent); zfs_close(zhp); return (ret); } /* - * Dumps a backup of the given snapshot (incremental from fromsnap if it's not - * NULL) to the file descriptor specified by outfd. - */ -int -zfs_send(zfs_handle_t *zhp, const char *fromsnap, int outfd) -{ - zfs_cmd_t zc = { 0 }; - char errbuf[1024]; - libzfs_handle_t *hdl = zhp->zfs_hdl; - - assert(zhp->zfs_type == ZFS_TYPE_SNAPSHOT); - - (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); - if (fromsnap) - (void) strlcpy(zc.zc_value, fromsnap, sizeof (zc.zc_name)); - zc.zc_cookie = outfd; - - if (ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_SENDBACKUP, &zc) != 0) { - (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, - "cannot send '%s'"), zhp->zfs_name); - - switch (errno) { - - case EXDEV: - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "not an earlier snapshot from the same fs")); - return (zfs_error(hdl, EZFS_CROSSTARGET, errbuf)); - - case EDQUOT: - case EFBIG: - case EIO: - case ENOLINK: - case ENOSPC: - case ENXIO: - case EPIPE: - case ERANGE: - case EFAULT: - case EROFS: - zfs_error_aux(hdl, strerror(errno)); - return (zfs_error(hdl, EZFS_BADBACKUP, errbuf)); - - default: - return (zfs_standard_error(hdl, errno, errbuf)); - } - } - - return (0); -} - -/* - * Create ancestors of 'target', but not target itself, and not - * ancestors whose names are shorter than prefixlen. Die if - * prefixlen-ancestor does not exist. - */ -static int -create_parents(libzfs_handle_t *hdl, char *target, int prefixlen) -{ - zfs_handle_t *h; - char *cp; - - /* make sure prefix exists */ - cp = strchr(target + prefixlen, '/'); - *cp = '\0'; - h = zfs_open(hdl, target, ZFS_TYPE_FILESYSTEM); - *cp = '/'; - if (h == NULL) - return (-1); - zfs_close(h); - - /* - * Attempt to create, mount, and share any ancestor filesystems, - * up to the prefixlen-long one. - */ - for (cp = target + prefixlen + 1; - cp = strchr(cp, '/'); *cp = '/', cp++) { - const char *opname; - - *cp = '\0'; - - h = make_dataset_handle(hdl, target); - if (h) { - /* it already exists, nothing to do here */ - zfs_close(h); - continue; - } - - opname = dgettext(TEXT_DOMAIN, "create"); - if (zfs_create(hdl, target, ZFS_TYPE_FILESYSTEM, - NULL) != 0) - goto ancestorerr; - - opname = dgettext(TEXT_DOMAIN, "open"); - h = zfs_open(hdl, target, ZFS_TYPE_FILESYSTEM); - if (h == NULL) - goto ancestorerr; - - opname = dgettext(TEXT_DOMAIN, "mount"); - if (zfs_mount(h, NULL, 0) != 0) - goto ancestorerr; - - opname = dgettext(TEXT_DOMAIN, "share"); - if (zfs_share(h) != 0) - goto ancestorerr; - - zfs_close(h); - - continue; -ancestorerr: - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "failed to %s ancestor '%s'"), opname, target); - return (-1); - } - - return (0); -} - -/* - * Restores a backup of tosnap from the file descriptor specified by infd. - */ -int -zfs_receive(libzfs_handle_t *hdl, const char *tosnap, int isprefix, - int verbose, int dryrun, boolean_t force, int infd) -{ - zfs_cmd_t zc = { 0 }; - time_t begin_time; - int ioctl_err, err, bytes, size, choplen; - char *cp; - dmu_replay_record_t drr; - struct drr_begin *drrb = &zc.zc_begin_record; - char errbuf[1024]; - prop_changelist_t *clp; - char chopprefix[ZFS_MAXNAMELEN]; - - begin_time = time(NULL); - - (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, - "cannot receive")); - - /* read in the BEGIN record */ - cp = (char *)&drr; - bytes = 0; - do { - size = read(infd, cp, sizeof (drr) - bytes); - cp += size; - bytes += size; - } while (size > 0); - - if (size < 0 || bytes != sizeof (drr)) { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid " - "stream (failed to read first record)")); - return (zfs_error(hdl, EZFS_BADSTREAM, errbuf)); - } - - zc.zc_begin_record = drr.drr_u.drr_begin; - - if (drrb->drr_magic != DMU_BACKUP_MAGIC && - drrb->drr_magic != BSWAP_64(DMU_BACKUP_MAGIC)) { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid " - "stream (bad magic number)")); - return (zfs_error(hdl, EZFS_BADSTREAM, errbuf)); - } - - if (drrb->drr_version != DMU_BACKUP_VERSION && - drrb->drr_version != BSWAP_64(DMU_BACKUP_VERSION)) { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "only version " - "0x%llx is supported (stream is version 0x%llx)"), - DMU_BACKUP_VERSION, drrb->drr_version); - return (zfs_error(hdl, EZFS_BADSTREAM, errbuf)); - } - - if (strchr(drr.drr_u.drr_begin.drr_toname, '@') == NULL) { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid " - "stream (bad snapshot name)")); - return (zfs_error(hdl, EZFS_BADSTREAM, errbuf)); - } - /* - * Determine how much of the snapshot name stored in the stream - * we are going to tack on to the name they specified on the - * command line, and how much we are going to chop off. - * - * If they specified a snapshot, chop the entire name stored in - * the stream. - */ - (void) strcpy(chopprefix, drr.drr_u.drr_begin.drr_toname); - if (isprefix) { - /* - * They specified a fs with -d, we want to tack on - * everything but the pool name stored in the stream - */ - if (strchr(tosnap, '@')) { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid " - "argument - snapshot not allowed with -d")); - return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); - } - cp = strchr(chopprefix, '/'); - if (cp == NULL) - cp = strchr(chopprefix, '@'); - *cp = '\0'; - } else if (strchr(tosnap, '@') == NULL) { - /* - * If they specified a filesystem without -d, we want to - * tack on everything after the fs specified in the - * first name from the stream. - */ - cp = strchr(chopprefix, '@'); - *cp = '\0'; - } - choplen = strlen(chopprefix); - - /* - * Determine name of destination snapshot, store in zc_value. - */ - (void) strcpy(zc.zc_value, tosnap); - (void) strncat(zc.zc_value, drr.drr_u.drr_begin.drr_toname+choplen, - sizeof (zc.zc_value)); - if (!zfs_validate_name(hdl, zc.zc_value, ZFS_TYPE_SNAPSHOT)) - return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); - - (void) strcpy(zc.zc_name, zc.zc_value); - if (drrb->drr_fromguid) { - /* incremental backup stream */ - zfs_handle_t *h; - - /* do the recvbackup ioctl to the containing fs */ - *strchr(zc.zc_name, '@') = '\0'; - - /* make sure destination fs exists */ - h = zfs_open(hdl, zc.zc_name, - ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME); - if (h == NULL) - return (-1); - if (!dryrun) { - /* - * We need to unmount all the dependents of the dataset - * and the dataset itself. If it's a volume - * then remove device link. - */ - if (h->zfs_type == ZFS_TYPE_FILESYSTEM) { - clp = changelist_gather(h, ZFS_PROP_NAME, 0); - if (clp == NULL) - return (-1); - if (changelist_prefix(clp) != 0) { - changelist_free(clp); - return (-1); - } - } else { - (void) zvol_remove_link(hdl, h->zfs_name); - } - } - zfs_close(h); - } else { - /* full backup stream */ - - /* Make sure destination fs does not exist */ - *strchr(zc.zc_name, '@') = '\0'; - if (ioctl(hdl->libzfs_fd, ZFS_IOC_OBJSET_STATS, &zc) == 0) { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "destination '%s' exists"), zc.zc_name); - return (zfs_error(hdl, EZFS_EXISTS, errbuf)); - } - - if (strchr(zc.zc_name, '/') == NULL) { - /* - * they're trying to do a recv into a - * nonexistant topmost filesystem. - */ - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "destination does not exist"), zc.zc_name); - return (zfs_error(hdl, EZFS_EXISTS, errbuf)); - } - - /* Do the recvbackup ioctl to the fs's parent. */ - *strrchr(zc.zc_name, '/') = '\0'; - - if (isprefix && (err = create_parents(hdl, - zc.zc_value, strlen(tosnap))) != 0) { - return (zfs_error(hdl, EZFS_BADRESTORE, errbuf)); - } - - } - - zc.zc_cookie = infd; - zc.zc_guid = force; - if (verbose) { - (void) printf("%s %s stream of %s into %s\n", - dryrun ? "would receive" : "receiving", - drrb->drr_fromguid ? "incremental" : "full", - drr.drr_u.drr_begin.drr_toname, - zc.zc_value); - (void) fflush(stdout); - } - if (dryrun) - return (0); - err = ioctl_err = ioctl(hdl->libzfs_fd, ZFS_IOC_RECVBACKUP, &zc); - if (ioctl_err != 0) { - switch (errno) { - case ENODEV: - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "most recent snapshot does not match incremental " - "source")); - (void) zfs_error(hdl, EZFS_BADRESTORE, errbuf); - break; - case ETXTBSY: - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "destination has been modified since most recent " - "snapshot")); - (void) zfs_error(hdl, EZFS_BADRESTORE, errbuf); - break; - case EEXIST: - if (drrb->drr_fromguid == 0) { - /* it's the containing fs that exists */ - cp = strchr(zc.zc_value, '@'); - *cp = '\0'; - } - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "destination already exists")); - (void) zfs_error_fmt(hdl, EZFS_EXISTS, - dgettext(TEXT_DOMAIN, "cannot restore to %s"), - zc.zc_value); - break; - case EINVAL: - (void) zfs_error(hdl, EZFS_BADSTREAM, errbuf); - break; - case ECKSUM: - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "invalid stream (checksum mismatch)")); - (void) zfs_error(hdl, EZFS_BADSTREAM, errbuf); - break; - default: - (void) zfs_standard_error(hdl, errno, errbuf); - } - } - - /* - * Mount or recreate the /dev links for the target filesystem - * (if created, or if we tore them down to do an incremental - * restore), and the /dev links for the new snapshot (if - * created). Also mount any children of the target filesystem - * if we did an incremental receive. - */ - cp = strchr(zc.zc_value, '@'); - if (cp && (ioctl_err == 0 || drrb->drr_fromguid)) { - zfs_handle_t *h; - - *cp = '\0'; - h = zfs_open(hdl, zc.zc_value, - ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME); - *cp = '@'; - if (h) { - if (h->zfs_type == ZFS_TYPE_VOLUME) { - err = zvol_create_link(hdl, h->zfs_name); - if (err == 0 && ioctl_err == 0) - err = zvol_create_link(hdl, - zc.zc_value); - } else { - if (drrb->drr_fromguid) { - err = changelist_postfix(clp); - changelist_free(clp); - } else { - err = zfs_mount(h, NULL, 0); - } - } - zfs_close(h); - } - } - - if (err || ioctl_err) - return (-1); - - if (verbose) { - char buf1[64]; - char buf2[64]; - uint64_t bytes = zc.zc_cookie; - time_t delta = time(NULL) - begin_time; - if (delta == 0) - delta = 1; - zfs_nicenum(bytes, buf1, sizeof (buf1)); - zfs_nicenum(bytes/delta, buf2, sizeof (buf1)); - - (void) printf("received %sb stream in %lu seconds (%sb/sec)\n", - buf1, delta, buf2); - } - - return (0); -} - -/* * Destroy any more recent snapshots. We invoke this callback on any dependents * of the snapshot first. If the 'cb_dependent' member is non-zero, then this * is a dependent and we should just destroy it without checking the transaction @@ -3030,9 +3568,9 @@ zfs_receive(libzfs_handle_t *hdl, const char *tosnap, int isprefix, typedef struct rollback_data { const char *cb_target; /* the snapshot */ uint64_t cb_create; /* creation time reference */ - prop_changelist_t *cb_clp; /* changelist pointer */ - int cb_error; + boolean_t cb_error; boolean_t cb_dependent; + boolean_t cb_force; } rollback_data_t; static int @@ -3045,23 +3583,35 @@ rollback_destroy(zfs_handle_t *zhp, void *data) zfs_get_type(zhp) == ZFS_TYPE_SNAPSHOT && zfs_prop_get_int(zhp, ZFS_PROP_CREATETXG) > cbp->cb_create) { + char *logstr; cbp->cb_dependent = B_TRUE; - if (zfs_iter_dependents(zhp, B_FALSE, rollback_destroy, - cbp) != 0) - cbp->cb_error = 1; + cbp->cb_error |= zfs_iter_dependents(zhp, B_FALSE, + rollback_destroy, cbp); cbp->cb_dependent = B_FALSE; - if (zfs_destroy(zhp) != 0) - cbp->cb_error = 1; - else - changelist_remove(zhp, cbp->cb_clp); + logstr = zhp->zfs_hdl->libzfs_log_str; + zhp->zfs_hdl->libzfs_log_str = NULL; + cbp->cb_error |= zfs_destroy(zhp); + zhp->zfs_hdl->libzfs_log_str = logstr; } } else { + /* We must destroy this clone; first unmount it */ + prop_changelist_t *clp; + + clp = changelist_gather(zhp, ZFS_PROP_NAME, 0, + cbp->cb_force ? MS_FORCE: 0); + if (clp == NULL || changelist_prefix(clp) != 0) { + cbp->cb_error = B_TRUE; + zfs_close(zhp); + return (0); + } if (zfs_destroy(zhp) != 0) - cbp->cb_error = 1; + cbp->cb_error = B_TRUE; else - changelist_remove(zhp, cbp->cb_clp); + changelist_remove(clp, zhp->zfs_name); + (void) changelist_postfix(clp); + changelist_free(clp); } zfs_close(zhp); @@ -3069,48 +3619,6 @@ rollback_destroy(zfs_handle_t *zhp, void *data) } /* - * Rollback the dataset to its latest snapshot. - */ -static int -do_rollback(zfs_handle_t *zhp) -{ - int ret; - zfs_cmd_t zc = { 0 }; - - assert(zhp->zfs_type == ZFS_TYPE_FILESYSTEM || - zhp->zfs_type == ZFS_TYPE_VOLUME); - - if (zhp->zfs_type == ZFS_TYPE_VOLUME && - zvol_remove_link(zhp->zfs_hdl, zhp->zfs_name) != 0) - return (-1); - - (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); - - if (ZFS_IS_VOLUME(zhp)) - zc.zc_objset_type = DMU_OST_ZVOL; - else - zc.zc_objset_type = DMU_OST_ZFS; - - /* - * We rely on the consumer to verify that there are no newer snapshots - * for the given dataset. Given these constraints, we can simply pass - * the name on to the ioctl() call. There is still an unlikely race - * condition where the user has taken a snapshot since we verified that - * this was the most recent. - */ - if ((ret = ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_ROLLBACK, - &zc)) != 0) { - (void) zfs_standard_error_fmt(zhp->zfs_hdl, errno, - dgettext(TEXT_DOMAIN, "cannot rollback '%s'"), - zhp->zfs_name); - } else if (zhp->zfs_type == ZFS_TYPE_VOLUME) { - ret = zvol_create_link(zhp->zfs_hdl, zhp->zfs_name); - } - - return (ret); -} - -/* * Given a dataset, rollback to a specific snapshot, discarding any * data changes since then and making it the active dataset. * @@ -3118,56 +3626,87 @@ do_rollback(zfs_handle_t *zhp) * their dependents. */ int -zfs_rollback(zfs_handle_t *zhp, zfs_handle_t *snap, int flag) +zfs_rollback(zfs_handle_t *zhp, zfs_handle_t *snap, boolean_t force) { - int ret; rollback_data_t cb = { 0 }; - prop_changelist_t *clp; - - /* - * Unmount all dependendents of the dataset and the dataset itself. - * The list we need to gather is the same as for doing rename - */ - clp = changelist_gather(zhp, ZFS_PROP_NAME, flag ? MS_FORCE: 0); - if (clp == NULL) - return (-1); + int err; + zfs_cmd_t zc = { 0 }; + boolean_t restore_resv = 0; + uint64_t old_volsize, new_volsize; + zfs_prop_t resv_prop; - if ((ret = changelist_prefix(clp)) != 0) - goto out; + assert(zhp->zfs_type == ZFS_TYPE_FILESYSTEM || + zhp->zfs_type == ZFS_TYPE_VOLUME); /* * Destroy all recent snapshots and its dependends. */ + cb.cb_force = force; cb.cb_target = snap->zfs_name; cb.cb_create = zfs_prop_get_int(snap, ZFS_PROP_CREATETXG); - cb.cb_clp = clp; (void) zfs_iter_children(zhp, rollback_destroy, &cb); - if ((ret = cb.cb_error) != 0) { - (void) changelist_postfix(clp); - goto out; - } + if (cb.cb_error) + return (-1); /* * Now that we have verified that the snapshot is the latest, * rollback to the given snapshot. */ - ret = do_rollback(zhp); - if (ret != 0) { - (void) changelist_postfix(clp); - goto out; + if (zhp->zfs_type == ZFS_TYPE_VOLUME) { + if (zvol_remove_link(zhp->zfs_hdl, zhp->zfs_name) != 0) + return (-1); + if (zfs_which_resv_prop(zhp, &resv_prop) < 0) + return (-1); + old_volsize = zfs_prop_get_int(zhp, ZFS_PROP_VOLSIZE); + restore_resv = + (old_volsize == zfs_prop_get_int(zhp, resv_prop)); } + (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); + + if (ZFS_IS_VOLUME(zhp)) + zc.zc_objset_type = DMU_OST_ZVOL; + else + zc.zc_objset_type = DMU_OST_ZFS; + /* - * We only want to re-mount the filesystem if it was mounted in the - * first place. + * We rely on zfs_iter_children() to verify that there are no + * newer snapshots for the given dataset. Therefore, we can + * simply pass the name on to the ioctl() call. There is still + * an unlikely race condition where the user has taken a + * snapshot since we verified that this was the most recent. + * */ - ret = changelist_postfix(clp); + if ((err = zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_ROLLBACK, &zc)) != 0) { + (void) zfs_standard_error_fmt(zhp->zfs_hdl, errno, + dgettext(TEXT_DOMAIN, "cannot rollback '%s'"), + zhp->zfs_name); + return (err); + } -out: - changelist_free(clp); - return (ret); + /* + * For volumes, if the pre-rollback volsize matched the pre- + * rollback reservation and the volsize has changed then set + * the reservation property to the post-rollback volsize. + * Make a new handle since the rollback closed the dataset. + */ + if ((zhp->zfs_type == ZFS_TYPE_VOLUME) && + (zhp = make_dataset_handle(zhp->zfs_hdl, zhp->zfs_name))) { + if (err = zvol_create_link(zhp->zfs_hdl, zhp->zfs_name)) { + zfs_close(zhp); + return (err); + } + if (restore_resv) { + new_volsize = zfs_prop_get_int(zhp, ZFS_PROP_VOLSIZE); + if (old_volsize != new_volsize) + err = zfs_prop_set_int(zhp, resv_prop, + new_volsize); + } + zfs_close(zhp); + } + return (err); } /* @@ -3210,7 +3749,7 @@ zfs_iter_dependents(zfs_handle_t *zhp, boolean_t allowrecursion, * Renames the given dataset. */ int -zfs_rename(zfs_handle_t *zhp, const char *target, int recursive) +zfs_rename(zfs_handle_t *zhp, const char *target, boolean_t recursive) { int ret; zfs_cmd_t zc = { 0 }; @@ -3262,7 +3801,7 @@ zfs_rename(zfs_handle_t *zhp, const char *target, int recursive) errbuf)); } } - if (!zfs_validate_name(hdl, target, zhp->zfs_type)) + if (!zfs_validate_name(hdl, target, zhp->zfs_type, B_TRUE)) return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); } else { if (recursive) { @@ -3271,12 +3810,12 @@ zfs_rename(zfs_handle_t *zhp, const char *target, int recursive) return (zfs_error(hdl, EZFS_BADTYPE, errbuf)); } - if (!zfs_validate_name(hdl, target, zhp->zfs_type)) + if (!zfs_validate_name(hdl, target, zhp->zfs_type, B_TRUE)) return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); uint64_t unused; /* validate parents */ - if (check_parents(hdl, target, &unused) != 0) + if (check_parents(hdl, target, &unused, B_FALSE, NULL) != 0) return (-1); (void) parent_name(target, parent, sizeof (parent)); @@ -3313,17 +3852,22 @@ zfs_rename(zfs_handle_t *zhp, const char *target, int recursive) if (recursive) { struct destroydata dd; - parentname = strdup(zhp->zfs_name); + parentname = zfs_strdup(zhp->zfs_hdl, zhp->zfs_name); + if (parentname == NULL) { + ret = -1; + goto error; + } delim = strchr(parentname, '@'); *delim = '\0'; - zhrp = zfs_open(zhp->zfs_hdl, parentname, ZFS_TYPE_ANY); + zhrp = zfs_open(zhp->zfs_hdl, parentname, ZFS_TYPE_DATASET); if (zhrp == NULL) { - return (-1); + ret = -1; + goto error; } dd.snapname = delim + 1; dd.gotone = B_FALSE; - dd.closezhp = B_FALSE; + dd.closezhp = B_TRUE; /* We remove any zvol links prior to renaming them */ ret = zfs_iter_filesystems(zhrp, zfs_remove_link_cb, &dd); @@ -3331,7 +3875,7 @@ zfs_rename(zfs_handle_t *zhp, const char *target, int recursive) goto error; } } else { - if ((cl = changelist_gather(zhp, ZFS_PROP_NAME, 0)) == NULL) + if ((cl = changelist_gather(zhp, ZFS_PROP_NAME, 0, 0)) == NULL) return (-1); if (changelist_haszonedchild(cl)) { @@ -3356,19 +3900,19 @@ zfs_rename(zfs_handle_t *zhp, const char *target, int recursive) zc.zc_cookie = recursive; - if ((ret = ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_RENAME, &zc)) != 0) { + if ((ret = zfs_ioctl(zhp->zfs_hdl, ZFS_IOC_RENAME, &zc)) != 0) { /* * if it was recursive, the one that actually failed will * be in zc.zc_name */ (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, - "cannot rename to '%s'"), zc.zc_name); + "cannot rename '%s'"), zc.zc_name); if (recursive && errno == EEXIST) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "a child dataset already has a snapshot " "with the new name")); - (void) zfs_error(hdl, EZFS_CROSSTARGET, errbuf); + (void) zfs_error(hdl, EZFS_EXISTS, errbuf); } else { (void) zfs_standard_error(zhp->zfs_hdl, errno, errbuf); } @@ -3432,6 +3976,8 @@ zvol_create_link_common(libzfs_handle_t *hdl, const char *dataset, int ifexists) zfs_cmd_t zc = { 0 }; #if 0 di_devlink_handle_t dhdl; + priv_set_t *priv_effective; + int privileged; #endif (void) strlcpy(zc.zc_name, dataset, sizeof (zc.zc_name)); @@ -3470,17 +4016,51 @@ zvol_create_link_common(libzfs_handle_t *hdl, const char *dataset, int ifexists) #if 0 /* - * Call devfsadm and wait for the links to magically appear. + * If privileged call devfsadm and wait for the links to + * magically appear. + * Otherwise, print out an informational message. */ - if ((dhdl = di_devlink_init(ZFS_DRIVER, DI_MAKE_LINK)) == NULL) { - zfs_error_aux(hdl, strerror(errno)); - (void) zfs_error_fmt(hdl, EZFS_DEVLINKS, - dgettext(TEXT_DOMAIN, "cannot create device links " - "for '%s'"), dataset); - (void) ioctl(hdl->libzfs_fd, ZFS_IOC_REMOVE_MINOR, &zc); - return (-1); + + priv_effective = priv_allocset(); + (void) getppriv(PRIV_EFFECTIVE, priv_effective); + privileged = (priv_isfullset(priv_effective) == B_TRUE); + priv_freeset(priv_effective); + + if (privileged) { + if ((dhdl = di_devlink_init(ZFS_DRIVER, + DI_MAKE_LINK)) == NULL) { + zfs_error_aux(hdl, strerror(errno)); + (void) zfs_error_fmt(hdl, errno, + dgettext(TEXT_DOMAIN, "cannot create device links " + "for '%s'"), dataset); + (void) ioctl(hdl->libzfs_fd, ZFS_IOC_REMOVE_MINOR, &zc); + return (-1); + } else { + (void) di_devlink_fini(&dhdl); + } } else { - (void) di_devlink_fini(&dhdl); + char pathname[MAXPATHLEN]; + struct stat64 statbuf; + int i; + +#define MAX_WAIT 10 + + /* + * This is the poor mans way of waiting for the link + * to show up. If after 10 seconds we still don't + * have it, then print out a message. + */ + (void) snprintf(pathname, sizeof (pathname), "/dev/zvol/dsk/%s", + dataset); + + for (i = 0; i != MAX_WAIT; i++) { + if (stat64(pathname, &statbuf) == 0) + break; + (void) sleep(1); + } + if (i == MAX_WAIT) + (void) printf(gettext("%s may not be immediately " + "available\n"), pathname); } #endif @@ -3524,200 +4104,6 @@ zfs_get_user_props(zfs_handle_t *zhp) } /* - * Given a comma-separated list of properties, contruct a property list - * containing both user-defined and native properties. This function will - * return a NULL list if 'all' is specified, which can later be expanded on a - * per-dataset basis by zfs_expand_proplist(). - */ -int -zfs_get_proplist_common(libzfs_handle_t *hdl, char *fields, - zfs_proplist_t **listp, zfs_type_t type) -{ - size_t len; - char *s, *p; - char c; - zfs_prop_t prop; - zfs_proplist_t *entry; - zfs_proplist_t **last; - - *listp = NULL; - last = listp; - - /* - * If 'all' is specified, return a NULL list. - */ - if (strcmp(fields, "all") == 0) - return (0); - - /* - * If no fields were specified, return an error. - */ - if (fields[0] == '\0') { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "no properties specified")); - return (zfs_error(hdl, EZFS_BADPROP, dgettext(TEXT_DOMAIN, - "bad property list"))); - } - - /* - * It would be nice to use getsubopt() here, but the inclusion of column - * aliases makes this more effort than it's worth. - */ - s = fields; - while (*s != '\0') { - if ((p = strchr(s, ',')) == NULL) { - len = strlen(s); - p = s + len; - } else { - len = p - s; - } - - /* - * Check for empty options. - */ - if (len == 0) { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "empty property name")); - return (zfs_error(hdl, EZFS_BADPROP, - dgettext(TEXT_DOMAIN, "bad property list"))); - } - - /* - * Check all regular property names. - */ - c = s[len]; - s[len] = '\0'; - prop = zfs_name_to_prop_common(s, type); - - if (prop != ZFS_PROP_INVAL && - !zfs_prop_valid_for_type(prop, type)) - prop = ZFS_PROP_INVAL; - - /* - * When no property table entry can be found, return failure if - * this is a pool property or if this isn't a user-defined - * dataset property, - */ - if (prop == ZFS_PROP_INVAL && - (type & ZFS_TYPE_POOL || !zfs_prop_user(s))) { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "invalid property '%s'"), s); - return (zfs_error(hdl, EZFS_BADPROP, - dgettext(TEXT_DOMAIN, "bad property list"))); - } - - if ((entry = zfs_alloc(hdl, sizeof (zfs_proplist_t))) == NULL) - return (-1); - - entry->pl_prop = prop; - if (prop == ZFS_PROP_INVAL) { - if ((entry->pl_user_prop = - zfs_strdup(hdl, s)) == NULL) { - free(entry); - return (-1); - } - entry->pl_width = strlen(s); - } else { - entry->pl_width = zfs_prop_width(prop, - &entry->pl_fixed); - } - - *last = entry; - last = &entry->pl_next; - - s = p; - if (c == ',') - s++; - } - - return (0); -} - -int -zfs_get_proplist(libzfs_handle_t *hdl, char *fields, zfs_proplist_t **listp) -{ - return (zfs_get_proplist_common(hdl, fields, listp, ZFS_TYPE_ANY)); -} - -void -zfs_free_proplist(zfs_proplist_t *pl) -{ - zfs_proplist_t *next; - - while (pl != NULL) { - next = pl->pl_next; - free(pl->pl_user_prop); - free(pl); - pl = next; - } -} - -typedef struct expand_data { - zfs_proplist_t **last; - libzfs_handle_t *hdl; -} expand_data_t; - -static zfs_prop_t -zfs_expand_proplist_cb(zfs_prop_t prop, void *cb) -{ - zfs_proplist_t *entry; - expand_data_t *edp = cb; - - if ((entry = zfs_alloc(edp->hdl, sizeof (zfs_proplist_t))) == NULL) - return (ZFS_PROP_INVAL); - - entry->pl_prop = prop; - entry->pl_width = zfs_prop_width(prop, &entry->pl_fixed); - entry->pl_all = B_TRUE; - - *(edp->last) = entry; - edp->last = &entry->pl_next; - - return (ZFS_PROP_CONT); -} - -int -zfs_expand_proplist_common(libzfs_handle_t *hdl, zfs_proplist_t **plp, - zfs_type_t type) -{ - zfs_proplist_t *entry; - zfs_proplist_t **last; - expand_data_t exp; - - if (*plp == NULL) { - /* - * If this is the very first time we've been called for an 'all' - * specification, expand the list to include all native - * properties. - */ - last = plp; - - exp.last = last; - exp.hdl = hdl; - - if (zfs_prop_iter_common(zfs_expand_proplist_cb, &exp, type, - B_FALSE) == ZFS_PROP_INVAL) - return (-1); - - /* - * Add 'name' to the beginning of the list, which is handled - * specially. - */ - if ((entry = zfs_alloc(hdl, - sizeof (zfs_proplist_t))) == NULL) - return (-1); - - entry->pl_prop = ZFS_PROP_NAME; - entry->pl_width = zfs_prop_width(ZFS_PROP_NAME, - &entry->pl_fixed); - entry->pl_all = B_TRUE; - entry->pl_next = *plp; - *plp = entry; - } - return (0); -} - -/* * This function is used by 'zfs list' to determine the exact set of columns to * display, and their maximum widths. This does two main things: * @@ -3729,17 +4115,17 @@ zfs_expand_proplist_common(libzfs_handle_t *hdl, zfs_proplist_t **plp, * so that we can size the column appropriately. */ int -zfs_expand_proplist(zfs_handle_t *zhp, zfs_proplist_t **plp) +zfs_expand_proplist(zfs_handle_t *zhp, zprop_list_t **plp) { libzfs_handle_t *hdl = zhp->zfs_hdl; - zfs_proplist_t *entry; - zfs_proplist_t **last, **start; + zprop_list_t *entry; + zprop_list_t **last, **start; nvlist_t *userprops, *propval; nvpair_t *elem; char *strval; char buf[ZFS_MAXPROPLEN]; - if (zfs_expand_proplist_common(hdl, plp, ZFS_TYPE_ANY) != 0) + if (zprop_expand_list(hdl, plp, ZFS_TYPE_DATASET) != 0) return (-1); userprops = zfs_get_user_props(zhp); @@ -3753,7 +4139,7 @@ zfs_expand_proplist(zfs_handle_t *zhp, zfs_proplist_t **plp) */ start = plp; while (*start != NULL) { - if ((*start)->pl_prop == ZFS_PROP_INVAL) + if ((*start)->pl_prop == ZPROP_INVAL) break; start = &(*start)->pl_next; } @@ -3772,14 +4158,14 @@ zfs_expand_proplist(zfs_handle_t *zhp, zfs_proplist_t **plp) if (*last == NULL) { if ((entry = zfs_alloc(hdl, - sizeof (zfs_proplist_t))) == NULL || + sizeof (zprop_list_t))) == NULL || ((entry->pl_user_prop = zfs_strdup(hdl, nvpair_name(elem)))) == NULL) { free(entry); return (-1); } - entry->pl_prop = ZFS_PROP_INVAL; + entry->pl_prop = ZPROP_INVAL; entry->pl_width = strlen(nvpair_name(elem)); entry->pl_all = B_TRUE; *last = entry; @@ -3794,7 +4180,7 @@ zfs_expand_proplist(zfs_handle_t *zhp, zfs_proplist_t **plp) if (entry->pl_fixed) continue; - if (entry->pl_prop != ZFS_PROP_INVAL) { + if (entry->pl_prop != ZPROP_INVAL) { if (zfs_prop_get(zhp, entry->pl_prop, buf, sizeof (buf), NULL, NULL, 0, B_FALSE) == 0) { if (strlen(buf) > entry->pl_width) @@ -3803,7 +4189,7 @@ zfs_expand_proplist(zfs_handle_t *zhp, zfs_proplist_t **plp) } else if (nvlist_lookup_nvlist(userprops, entry->pl_user_prop, &propval) == 0) { verify(nvlist_lookup_string(propval, - ZFS_PROP_VALUE, &strval) == 0); + ZPROP_VALUE, &strval) == 0); if (strlen(strval) > entry->pl_width) entry->pl_width = strlen(strval); } @@ -3812,6 +4198,72 @@ zfs_expand_proplist(zfs_handle_t *zhp, zfs_proplist_t **plp) return (0); } +#ifdef TODO +int +zfs_iscsi_perm_check(libzfs_handle_t *hdl, char *dataset, ucred_t *cred) +{ + zfs_cmd_t zc = { 0 }; + nvlist_t *nvp; + gid_t gid; + uid_t uid; + const gid_t *groups; + int group_cnt; + int error; + + if (nvlist_alloc(&nvp, NV_UNIQUE_NAME, 0) != 0) + return (no_memory(hdl)); + + uid = ucred_geteuid(cred); + gid = ucred_getegid(cred); + group_cnt = ucred_getgroups(cred, &groups); + + if (uid == (uid_t)-1 || gid == (uid_t)-1 || group_cnt == (uid_t)-1) + return (1); + + if (nvlist_add_uint32(nvp, ZFS_DELEG_PERM_UID, uid) != 0) { + nvlist_free(nvp); + return (1); + } + + if (nvlist_add_uint32(nvp, ZFS_DELEG_PERM_GID, gid) != 0) { + nvlist_free(nvp); + return (1); + } + + if (nvlist_add_uint32_array(nvp, + ZFS_DELEG_PERM_GROUPS, (uint32_t *)groups, group_cnt) != 0) { + nvlist_free(nvp); + return (1); + } + (void) strlcpy(zc.zc_name, dataset, sizeof (zc.zc_name)); + + if (zcmd_write_src_nvlist(hdl, &zc, nvp)) + return (-1); + + error = ioctl(hdl->libzfs_fd, ZFS_IOC_ISCSI_PERM_CHECK, &zc); + nvlist_free(nvp); + return (error); +} +#endif + +int +zfs_deleg_share_nfs(libzfs_handle_t *hdl, char *dataset, char *path, + void *export, void *sharetab, int sharemax, zfs_share_op_t operation) +{ + zfs_cmd_t zc = { 0 }; + int error; + + (void) strlcpy(zc.zc_name, dataset, sizeof (zc.zc_name)); + (void) strlcpy(zc.zc_value, path, sizeof (zc.zc_value)); + zc.zc_share.z_sharedata = (uint64_t)(uintptr_t)sharetab; + zc.zc_share.z_exportdata = (uint64_t)(uintptr_t)export; + zc.zc_share.z_sharetype = operation; + zc.zc_share.z_sharemax = sharemax; + + error = ioctl(hdl->libzfs_fd, ZFS_IOC_SHARE, &zc); + return (error); +} + /* * Attach/detach the given filesystem to/from the given jail. */ diff --git a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_graph.c b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_graph.c index c283016..e7cbf23 100644 --- a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_graph.c +++ b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_graph.c @@ -19,7 +19,7 @@ * CDDL HEADER END */ /* - * Copyright 2006 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ @@ -126,6 +126,8 @@ typedef struct zfs_graph { zfs_vertex_t **zg_hash; size_t zg_size; size_t zg_nvertex; + const char *zg_root; + int zg_clone_count; } zfs_graph_t; /* @@ -255,7 +257,7 @@ zfs_vertex_sort_edges(zfs_vertex_t *zvp) * datasets in the pool. */ static zfs_graph_t * -zfs_graph_create(libzfs_handle_t *hdl, size_t size) +zfs_graph_create(libzfs_handle_t *hdl, const char *dataset, size_t size) { zfs_graph_t *zgp = zfs_alloc(hdl, sizeof (zfs_graph_t)); @@ -269,6 +271,9 @@ zfs_graph_create(libzfs_handle_t *hdl, size_t size) return (NULL); } + zgp->zg_root = dataset; + zgp->zg_clone_count = 0; + return (zgp); } @@ -367,17 +372,16 @@ zfs_graph_add(libzfs_handle_t *hdl, zfs_graph_t *zgp, const char *source, } /* - * Iterate over all children of the given dataset, adding any vertices as - * necessary. Returns 0 if no cloned snapshots were seen, -1 if there was an - * error, or 1 otherwise. This is a simple recursive algorithm - the ZFS - * namespace typically is very flat. We manually invoke the necessary ioctl() - * calls to avoid the overhead and additional semantics of zfs_open(). + * Iterate over all children of the given dataset, adding any vertices + * as necessary. Returns -1 if there was an error, or 0 otherwise. + * This is a simple recursive algorithm - the ZFS namespace typically + * is very flat. We manually invoke the necessary ioctl() calls to + * avoid the overhead and additional semantics of zfs_open(). */ static int iterate_children(libzfs_handle_t *hdl, zfs_graph_t *zgp, const char *dataset) { zfs_cmd_t zc = { 0 }; - int ret = 0, err; zfs_vertex_t *zvp; /* @@ -390,18 +394,8 @@ iterate_children(libzfs_handle_t *hdl, zfs_graph_t *zgp, const char *dataset) return (0); /* - * We check the clone parent here instead of within the loop, so that if - * the root dataset has been promoted from a clone, we find its parent - * appropriately. + * Iterate over all children */ - (void) strlcpy(zc.zc_name, dataset, sizeof (zc.zc_name)); - if (ioctl(hdl->libzfs_fd, ZFS_IOC_OBJSET_STATS, &zc) == 0 && - zc.zc_objset_stats.dds_clone_of[0] != '\0') { - if (zfs_graph_add(hdl, zgp, zc.zc_objset_stats.dds_clone_of, - zc.zc_name, zc.zc_objset_stats.dds_creation_txg) != 0) - return (-1); - } - for ((void) strlcpy(zc.zc_name, dataset, sizeof (zc.zc_name)); ioctl(hdl->libzfs_fd, ZFS_IOC_DATASET_LIST_NEXT, &zc) == 0; (void) strlcpy(zc.zc_name, dataset, sizeof (zc.zc_name))) { @@ -417,9 +411,23 @@ iterate_children(libzfs_handle_t *hdl, zfs_graph_t *zgp, const char *dataset) * dataset and clone statistics. If this fails, the dataset has * since been removed, and we're pretty much screwed anyway. */ + zc.zc_objset_stats.dds_origin[0] = '\0'; if (ioctl(hdl->libzfs_fd, ZFS_IOC_OBJSET_STATS, &zc) != 0) continue; + if (zc.zc_objset_stats.dds_origin[0] != '\0') { + if (zfs_graph_add(hdl, zgp, + zc.zc_objset_stats.dds_origin, zc.zc_name, + zc.zc_objset_stats.dds_creation_txg) != 0) + return (-1); + /* + * Count origins only if they are contained in the graph + */ + if (isa_child_of(zc.zc_objset_stats.dds_origin, + zgp->zg_root)) + zgp->zg_clone_count--; + } + /* * Add an edge between the parent and the child. */ @@ -428,19 +436,10 @@ iterate_children(libzfs_handle_t *hdl, zfs_graph_t *zgp, const char *dataset) return (-1); /* - * Iterate over all children + * Recursively visit child */ - err = iterate_children(hdl, zgp, zc.zc_name); - if (err == -1) + if (iterate_children(hdl, zgp, zc.zc_name)) return (-1); - else if (err == 1) - ret = 1; - - /* - * Indicate if we found a dataset with a non-zero clone count. - */ - if (zc.zc_objset_stats.dds_num_clones != 0) - ret = 1; } /* @@ -467,67 +466,84 @@ iterate_children(libzfs_handle_t *hdl, zfs_graph_t *zgp, const char *dataset) zc.zc_objset_stats.dds_creation_txg) != 0) return (-1); - /* - * Indicate if we found a dataset with a non-zero clone count. - */ - if (zc.zc_objset_stats.dds_num_clones != 0) - ret = 1; + zgp->zg_clone_count += zc.zc_objset_stats.dds_num_clones; } zvp->zv_visited = VISIT_SEEN; - return (ret); + return (0); } /* - * Construct a complete graph of all necessary vertices. First, we iterate over - * only our object's children. If we don't find any cloned snapshots, then we - * simple return that. Otherwise, we have to start at the pool root and iterate - * over all datasets. + * Returns false if there are no snapshots with dependent clones in this + * subtree or if all of those clones are also in this subtree. Returns + * true if there is an error or there are external dependents. + */ +static boolean_t +external_dependents(libzfs_handle_t *hdl, zfs_graph_t *zgp, const char *dataset) +{ + zfs_cmd_t zc = { 0 }; + + /* + * Check whether this dataset is a clone or has clones since + * iterate_children() only checks the children. + */ + (void) strlcpy(zc.zc_name, dataset, sizeof (zc.zc_name)); + if (ioctl(hdl->libzfs_fd, ZFS_IOC_OBJSET_STATS, &zc) != 0) + return (B_TRUE); + + if (zc.zc_objset_stats.dds_origin[0] != '\0') { + if (zfs_graph_add(hdl, zgp, + zc.zc_objset_stats.dds_origin, zc.zc_name, + zc.zc_objset_stats.dds_creation_txg) != 0) + return (B_TRUE); + if (isa_child_of(zc.zc_objset_stats.dds_origin, dataset)) + zgp->zg_clone_count--; + } + + if ((zc.zc_objset_stats.dds_num_clones) || + iterate_children(hdl, zgp, dataset)) + return (B_TRUE); + + return (zgp->zg_clone_count != 0); +} + +/* + * Construct a complete graph of all necessary vertices. First, iterate over + * only our object's children. If no cloned snapshots are found, or all of + * the cloned snapshots are in this subtree then return a graph of the subtree. + * Otherwise, start at the root of the pool and iterate over all datasets. */ static zfs_graph_t * construct_graph(libzfs_handle_t *hdl, const char *dataset) { - zfs_graph_t *zgp = zfs_graph_create(hdl, ZFS_GRAPH_SIZE); - zfs_cmd_t zc = { 0 }; + zfs_graph_t *zgp = zfs_graph_create(hdl, dataset, ZFS_GRAPH_SIZE); int ret = 0; if (zgp == NULL) return (zgp); - /* - * We need to explicitly check whether this dataset has clones or not, - * since iterate_children() only checks the children. - */ - (void) strlcpy(zc.zc_name, dataset, sizeof (zc.zc_name)); - (void) ioctl(hdl->libzfs_fd, ZFS_IOC_OBJSET_STATS, &zc); - - if (zc.zc_objset_stats.dds_num_clones != 0 || - (ret = iterate_children(hdl, zgp, dataset)) != 0) { + if ((strchr(dataset, '/') == NULL) || + (external_dependents(hdl, zgp, dataset))) { /* * Determine pool name and try again. */ - char *pool, *slash; - - if ((slash = strchr(dataset, '/')) != NULL || - (slash = strchr(dataset, '@')) != NULL) { - pool = zfs_alloc(hdl, slash - dataset + 1); - if (pool == NULL) { - zfs_graph_destroy(zgp); - return (NULL); - } - (void) strncpy(pool, dataset, slash - dataset); - pool[slash - dataset] = '\0'; - - if (iterate_children(hdl, zgp, pool) == -1 || - zfs_graph_add(hdl, zgp, pool, NULL, 0) != 0) { - free(pool); - zfs_graph_destroy(zgp); - return (NULL); - } + int len = strcspn(dataset, "/@") + 1; + char *pool = zfs_alloc(hdl, len); + + if (pool == NULL) { + zfs_graph_destroy(zgp); + return (NULL); + } + (void) strlcpy(pool, dataset, len); + if (iterate_children(hdl, zgp, pool) == -1 || + zfs_graph_add(hdl, zgp, pool, NULL, 0) != 0) { free(pool); + zfs_graph_destroy(zgp); + return (NULL); } + free(pool); } if (ret == -1 || zfs_graph_add(hdl, zgp, dataset, NULL, 0) != 0) { diff --git a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_impl.h b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_impl.h index 9581331..afe71f3 100644 --- a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_impl.h +++ b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_impl.h @@ -1,5 +1,5 @@ /* - * CDDL HEADER START + * CDDL HEADER SART * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). @@ -20,21 +20,21 @@ */ /* - * Copyright 2007 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ #ifndef _LIBFS_IMPL_H #define _LIBFS_IMPL_H -#pragma ident "%Z%%M% %I% %E% SMI" - #include <sys/dmu.h> #include <sys/fs/zfs.h> #include <sys/zfs_ioctl.h> #include <sys/zfs_acl.h> +#include <sys/spa.h> #include <sys/nvpair.h> +#include <libshare.h> #include <libuutil.h> #include <libzfs.h> @@ -42,22 +42,33 @@ extern "C" { #endif +#ifdef VERIFY +#undef VERIFY +#endif +#define VERIFY verify + struct libzfs_handle { int libzfs_error; int libzfs_fd; FILE *libzfs_mnttab; FILE *libzfs_sharetab; + zpool_handle_t *libzfs_pool_handles; uu_avl_pool_t *libzfs_ns_avlpool; uu_avl_t *libzfs_ns_avl; uint64_t libzfs_ns_gen; int libzfs_desc_active; char libzfs_action[1024]; char libzfs_desc[1024]; + char *libzfs_log_str; int libzfs_printerr; + void *libzfs_sharehdl; /* libshare handle */ + uint_t libzfs_shareflags; }; +#define ZFSSHARE_MISS 0x01 /* Didn't find entry in cache */ struct zfs_handle { libzfs_handle_t *zfs_hdl; + zpool_handle_t *zpool_hdl; char zfs_name[ZFS_MAXNAMELEN]; zfs_type_t zfs_type; /* type including snapshot */ zfs_type_t zfs_head_type; /* type excluding snapshot */ @@ -66,7 +77,6 @@ struct zfs_handle { nvlist_t *zfs_user_props; boolean_t zfs_mntcheck; char *zfs_mntopts; - char zfs_root[MAXPATHLEN]; }; /* @@ -77,14 +87,33 @@ struct zfs_handle { struct zpool_handle { libzfs_handle_t *zpool_hdl; + zpool_handle_t *zpool_next; char zpool_name[ZPOOL_MAXNAMELEN]; int zpool_state; size_t zpool_config_size; nvlist_t *zpool_config; nvlist_t *zpool_old_config; nvlist_t *zpool_props; + diskaddr_t zpool_start_block; }; +typedef enum { + PROTO_NFS = 0, + PROTO_SMB = 1, + PROTO_END = 2 +} zfs_share_proto_t; + +/* + * The following can be used as a bitmask and any new values + * added must preserve that capability. + */ +typedef enum { + SHARED_NOT_SHARED = 0x0, + SHARED_ISCSI = 0x1, + SHARED_NFS = 0x2, + SHARED_SMB = 0x4 +} zfs_share_type_t; + int zfs_error(libzfs_handle_t *, int, const char *); int zfs_error_fmt(libzfs_handle_t *, int, const char *, ...); void zfs_error_aux(libzfs_handle_t *, const char *, ...); @@ -101,20 +130,24 @@ int zpool_standard_error_fmt(libzfs_handle_t *, int, const char *, ...); int get_dependents(libzfs_handle_t *, boolean_t, const char *, char ***, size_t *); -int zfs_expand_proplist_common(libzfs_handle_t *, zfs_proplist_t **, - zfs_type_t); -int zfs_get_proplist_common(libzfs_handle_t *, char *, zfs_proplist_t **, - zfs_type_t); -zfs_prop_t zfs_prop_iter_common(zfs_prop_f, void *, zfs_type_t, boolean_t); -zfs_prop_t zfs_name_to_prop_common(const char *, zfs_type_t); -nvlist_t *zfs_validate_properties(libzfs_handle_t *, zfs_type_t, char *, - nvlist_t *, uint64_t, zfs_handle_t *zhp, const char *errbuf); +int zprop_parse_value(libzfs_handle_t *, nvpair_t *, int, zfs_type_t, + nvlist_t *, char **, uint64_t *, const char *); +int zprop_expand_list(libzfs_handle_t *hdl, zprop_list_t **plp, + zfs_type_t type); + +/* + * Use this changelist_gather() flag to force attempting mounts + * on each change node regardless of whether or not it is currently + * mounted. + */ +#define CL_GATHER_MOUNT_ALWAYS 1 typedef struct prop_changelist prop_changelist_t; int zcmd_alloc_dst_nvlist(libzfs_handle_t *, zfs_cmd_t *, size_t); -int zcmd_write_src_nvlist(libzfs_handle_t *, zfs_cmd_t *, nvlist_t *, size_t *); +int zcmd_write_src_nvlist(libzfs_handle_t *, zfs_cmd_t *, nvlist_t *); +int zcmd_write_conf_nvlist(libzfs_handle_t *, zfs_cmd_t *, nvlist_t *); int zcmd_expand_dst_nvlist(libzfs_handle_t *, zfs_cmd_t *); int zcmd_read_dst_nvlist(libzfs_handle_t *, zfs_cmd_t *, nvlist_t **); void zcmd_free_nvlists(zfs_cmd_t *); @@ -122,13 +155,15 @@ void zcmd_free_nvlists(zfs_cmd_t *); int changelist_prefix(prop_changelist_t *); int changelist_postfix(prop_changelist_t *); void changelist_rename(prop_changelist_t *, const char *, const char *); -void changelist_remove(zfs_handle_t *, prop_changelist_t *); +void changelist_remove(prop_changelist_t *, const char *); void changelist_free(prop_changelist_t *); -prop_changelist_t *changelist_gather(zfs_handle_t *, zfs_prop_t, int); -int changelist_unshare(prop_changelist_t *); +prop_changelist_t *changelist_gather(zfs_handle_t *, zfs_prop_t, int, int); +int changelist_unshare(prop_changelist_t *, zfs_share_proto_t *); int changelist_haszonedchild(prop_changelist_t *); void remove_mountpoint(zfs_handle_t *); +int create_parents(libzfs_handle_t *, char *, int); +boolean_t isa_child_of(const char *dataset, const char *parent); zfs_handle_t *make_dataset_handle(libzfs_handle_t *, const char *); @@ -137,10 +172,23 @@ int zpool_open_silent(libzfs_handle_t *, const char *, zpool_handle_t **); int zvol_create_link(libzfs_handle_t *, const char *); int zvol_remove_link(libzfs_handle_t *, const char *); int zpool_iter_zvol(zpool_handle_t *, int (*)(const char *, void *), void *); +boolean_t zpool_name_valid(libzfs_handle_t *, boolean_t, const char *); void namespace_clear(libzfs_handle_t *); +/* + * libshare (sharemgr) interfaces used internally. + */ + +extern int zfs_init_libshare(libzfs_handle_t *, int); +extern void zfs_uninit_libshare(libzfs_handle_t *); +extern int zfs_parse_options(char *, zfs_share_proto_t); + +extern int zfs_unshare_proto(zfs_handle_t *zhp, + const char *, zfs_share_proto_t *); + #ifdef __FreeBSD__ + /* * This is FreeBSD version of ioctl, because Solaris' ioctl() updates * zc_nvlist_dst_size even if an error is returned, on FreeBSD if an diff --git a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_import.c b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_import.c index 1c77045..166c831 100644 --- a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_import.c +++ b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_import.c @@ -19,7 +19,7 @@ * CDDL HEADER END */ /* - * Copyright 2007 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ @@ -213,11 +213,13 @@ add_config(libzfs_handle_t *hdl, pool_list_t *pl, const char *path, name_entry_t *ne; /* - * If this is a hot spare not currently in use, add it to the list of - * names to translate, but don't do anything else. + * If this is a hot spare not currently in use or level 2 cache + * device, add it to the list of names to translate, but don't do + * anything else. */ if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_STATE, - &state) == 0 && state == POOL_STATE_SPARE && + &state) == 0 && + (state == POOL_STATE_SPARE || state == POOL_STATE_L2CACHE) && nvlist_lookup_uint64(config, ZPOOL_CONFIG_GUID, &vdev_guid) == 0) { if ((ne = zfs_alloc(hdl, sizeof (name_entry_t))) == NULL) return (-1); @@ -361,6 +363,46 @@ pool_active(libzfs_handle_t *hdl, const char *name, uint64_t guid, return (0); } +static nvlist_t * +refresh_config(libzfs_handle_t *hdl, nvlist_t *config) +{ + nvlist_t *nvl; + zfs_cmd_t zc = { 0 }; + int err; + + if (zcmd_write_conf_nvlist(hdl, &zc, config) != 0) + return (NULL); + + if (zcmd_alloc_dst_nvlist(hdl, &zc, + zc.zc_nvlist_conf_size * 2) != 0) { + zcmd_free_nvlists(&zc); + return (NULL); + } + + while ((err = ioctl(hdl->libzfs_fd, ZFS_IOC_POOL_TRYIMPORT, + &zc)) != 0 && errno == ENOMEM) { + if (zcmd_expand_dst_nvlist(hdl, &zc) != 0) { + zcmd_free_nvlists(&zc); + return (NULL); + } + } + + if (err) { + (void) zpool_standard_error(hdl, errno, + dgettext(TEXT_DOMAIN, "cannot discover pools")); + zcmd_free_nvlists(&zc); + return (NULL); + } + + if (zcmd_read_dst_nvlist(hdl, &zc, &nvl) != 0) { + zcmd_free_nvlists(&zc); + return (NULL); + } + + zcmd_free_nvlists(&zc); + return (nvl); +} + /* * Convert our list of pools into the definitive set of configurations. We * start by picking the best config for each toplevel vdev. Once that's done, @@ -369,26 +411,25 @@ pool_active(libzfs_handle_t *hdl, const char *name, uint64_t guid, * return to the user. */ static nvlist_t * -get_configs(libzfs_handle_t *hdl, pool_list_t *pl) +get_configs(libzfs_handle_t *hdl, pool_list_t *pl, boolean_t active_ok) { pool_entry_t *pe; vdev_entry_t *ve; config_entry_t *ce; nvlist_t *ret = NULL, *config = NULL, *tmp, *nvtop, *nvroot; - nvlist_t **spares; - uint_t i, nspares; + nvlist_t **spares, **l2cache; + uint_t i, nspares, nl2cache; boolean_t config_seen; uint64_t best_txg; char *name, *hostname; - zfs_cmd_t zc = { 0 }; uint64_t version, guid; - size_t len; - int err; uint_t children = 0; nvlist_t **child = NULL; uint_t c; boolean_t isactive; uint64_t hostid; + nvlist_t *nvl; + boolean_t found_one = B_FALSE; if (nvlist_alloc(&ret, 0, 0) != 0) goto nomem; @@ -571,6 +612,13 @@ get_configs(libzfs_handle_t *hdl, pool_list_t *pl) nvlist_free(nvroot); /* + * zdb uses this path to report on active pools that were + * imported or created using -R. + */ + if (active_ok) + goto add_pool; + + /* * Determine if this pool is currently active, in which case we * can't actually import it. */ @@ -588,41 +636,11 @@ get_configs(libzfs_handle_t *hdl, pool_list_t *pl) continue; } - /* - * Try to do the import in order to get vdev state. - */ - if (zcmd_write_src_nvlist(hdl, &zc, config, &len) != 0) + if ((nvl = refresh_config(hdl, config)) == NULL) goto error; nvlist_free(config); - config = NULL; - - if (zcmd_alloc_dst_nvlist(hdl, &zc, len * 2) != 0) { - zcmd_free_nvlists(&zc); - goto error; - } - - while ((err = ioctl(hdl->libzfs_fd, ZFS_IOC_POOL_TRYIMPORT, - &zc)) != 0 && errno == ENOMEM) { - if (zcmd_expand_dst_nvlist(hdl, &zc) != 0) { - zcmd_free_nvlists(&zc); - goto error; - } - } - - if (err) { - (void) zpool_standard_error(hdl, errno, - dgettext(TEXT_DOMAIN, "cannot discover pools")); - zcmd_free_nvlists(&zc); - goto error; - } - - if (zcmd_read_dst_nvlist(hdl, &zc, &config) != 0) { - zcmd_free_nvlists(&zc); - goto error; - } - - zcmd_free_nvlists(&zc); + config = nvl; /* * Go through and update the paths for spares, now that we have @@ -639,6 +657,17 @@ get_configs(libzfs_handle_t *hdl, pool_list_t *pl) } /* + * Update the paths for l2cache devices. + */ + if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE, + &l2cache, &nl2cache) == 0) { + for (i = 0; i < nl2cache; i++) { + if (fix_paths(l2cache[i], pl->names) != 0) + goto nomem; + } + } + + /* * Restore the original information read from the actual label. */ (void) nvlist_remove(config, ZPOOL_CONFIG_HOSTID, @@ -652,6 +681,7 @@ get_configs(libzfs_handle_t *hdl, pool_list_t *pl) hostname) == 0); } +add_pool: /* * Add this pool to the list of configs. */ @@ -660,10 +690,16 @@ get_configs(libzfs_handle_t *hdl, pool_list_t *pl) if (nvlist_add_nvlist(ret, name, config) != 0) goto nomem; + found_one = B_TRUE; nvlist_free(config); config = NULL; } + if (!found_one) { + nvlist_free(ret); + ret = NULL; + } + return (ret); nomem: @@ -682,8 +718,9 @@ error: * Return the offset of the given label. */ static uint64_t -label_offset(size_t size, int l) +label_offset(uint64_t size, int l) { + ASSERT(P2PHASE_TYPED(size, sizeof (vdev_label_t), uint64_t) == 0); return (l * sizeof (vdev_label_t) + (l < VDEV_LABELS / 2 ? 0 : size - VDEV_LABELS * sizeof (vdev_label_t))); } @@ -698,19 +735,20 @@ zpool_read_label(int fd, nvlist_t **config) struct stat64 statbuf; int l; vdev_label_t *label; - uint64_t state, txg; + uint64_t state, txg, size; *config = NULL; if (fstat64(fd, &statbuf) == -1) return (0); + size = P2ALIGN_TYPED(statbuf.st_size, sizeof (vdev_label_t), uint64_t); if ((label = malloc(sizeof (vdev_label_t))) == NULL) return (-1); for (l = 0; l < VDEV_LABELS; l++) { - if (pread(fd, label, sizeof (vdev_label_t), - label_offset(statbuf.st_size, l)) != sizeof (vdev_label_t)) + if (pread64(fd, label, sizeof (vdev_label_t), + label_offset(size, l)) != sizeof (vdev_label_t)) continue; if (nvlist_unpack(label->vl_vdev_phys.vp_nvlist, @@ -718,12 +756,12 @@ zpool_read_label(int fd, nvlist_t **config) continue; if (nvlist_lookup_uint64(*config, ZPOOL_CONFIG_POOL_STATE, - &state) != 0 || state > POOL_STATE_SPARE) { + &state) != 0 || state > POOL_STATE_L2CACHE) { nvlist_free(*config); continue; } - if (state != POOL_STATE_SPARE && + if (state != POOL_STATE_SPARE && state != POOL_STATE_L2CACHE && (nvlist_lookup_uint64(*config, ZPOOL_CONFIG_POOL_TXG, &txg) != 0 || txg == 0)) { nvlist_free(*config); @@ -739,31 +777,20 @@ zpool_read_label(int fd, nvlist_t **config) return (0); } -/* - * Given a list of directories to search, find all pools stored on disk. This - * includes partial pools which are not available to import. If no args are - * given (argc is 0), then the default directory (/dev) is searched. - */ -nvlist_t * -zpool_find_import(libzfs_handle_t *hdl, int argc, char **argv) +static int +geom_find_import(libzfs_handle_t *hdl, pool_list_t *pools) { - int i; char path[MAXPATHLEN]; - nvlist_t *ret = NULL, *config; - int fd; - pool_list_t pools = { 0 }; - pool_entry_t *pe, *penext; - vdev_entry_t *ve, *venext; - config_entry_t *ce, *cenext; - name_entry_t *ne, *nenext; struct gmesh mesh; struct gclass *mp; struct ggeom *gp; struct gprovider *pp; + nvlist_t *config; + int fd, ret = 0; /* * Go through and read the label configuration information from every - * possible device, organizing the information according to pool GUID + * GEOM provider, organizing the information according to pool GUID * and toplevel GUID. */ @@ -773,32 +800,183 @@ zpool_find_import(libzfs_handle_t *hdl, int argc, char **argv) LIST_FOREACH(mp, &mesh.lg_class, lg_class) { LIST_FOREACH(gp, &mp->lg_geom, lg_geom) { LIST_FOREACH(pp, &gp->lg_provider, lg_provider) { + if ((fd = g_open(pp->lg_name, 0)) < 0) + continue; (void) snprintf(path, sizeof (path), "%s%s", _PATH_DEV, pp->lg_name); - if ((fd = open64(path, O_RDONLY)) < 0) - continue; - if ((zpool_read_label(fd, &config)) != 0) { + (void) g_close(fd); (void) no_memory(hdl); goto error; } - (void) close(fd); + (void) g_close(fd); if (config == NULL) continue; - if (add_config(hdl, &pools, path, config) != 0) + if (add_config(hdl, pools, path, config) != 0) { + ret = -1; goto error; + } } } } - +error: geom_deletetree(&mesh); + return (ret); +} + +/* + * Given a list of directories to search, find all pools stored on disk. This + * includes partial pools which are not available to import. If no args are + * given (argc is 0), then the default directory (/dev/dsk) is searched. + * poolname or guid (but not both) are provided by the caller when trying + * to import a specific pool. + */ +static nvlist_t * +zpool_find_import_impl(libzfs_handle_t *hdl, int argc, char **argv, + boolean_t active_ok, char *poolname, uint64_t guid) +{ + int i; + DIR *dirp = NULL; + struct dirent64 *dp; + char path[MAXPATHLEN]; + char *end; + size_t pathleft; + struct stat64 statbuf; + nvlist_t *ret = NULL, *config; + static char *default_dir = "/dev/dsk"; + int fd; + pool_list_t pools = { 0 }; + pool_entry_t *pe, *penext; + vdev_entry_t *ve, *venext; + config_entry_t *ce, *cenext; + name_entry_t *ne, *nenext; + + verify(poolname == NULL || guid == 0); + + if (argc == 0) { + argc = 1; + argv = &default_dir; + } + + /* + * Go through and read the label configuration information from every + * possible device, organizing the information according to pool GUID + * and toplevel GUID. + */ + for (i = 0; i < argc; i++) { + char *rdsk; + int dfd; + + /* use realpath to normalize the path */ + if (realpath(argv[i], path) == 0) { + (void) zfs_error_fmt(hdl, EZFS_BADPATH, + dgettext(TEXT_DOMAIN, "cannot open '%s'"), + argv[i]); + goto error; + } + end = &path[strlen(path)]; + *end++ = '/'; + *end = 0; + pathleft = &path[sizeof (path)] - end; - ret = get_configs(hdl, &pools); + if (strcmp(argv[i], default_dir) == 0) { + geom_find_import(hdl, &pools); + continue; + } + + /* + * Using raw devices instead of block devices when we're + * reading the labels skips a bunch of slow operations during + * close(2) processing, so we replace /dev/dsk with /dev/rdsk. + */ + if (strcmp(path, "/dev/dsk/") == 0) + rdsk = "/dev/rdsk/"; + else + rdsk = path; + + if ((dirp = opendir(rdsk)) == NULL) { + zfs_error_aux(hdl, strerror(errno)); + (void) zfs_error_fmt(hdl, EZFS_BADPATH, + dgettext(TEXT_DOMAIN, "cannot open '%s'"), + rdsk); + goto error; + } + + /* + * This is not MT-safe, but we have no MT consumers of libzfs + */ + while ((dp = readdir64(dirp)) != NULL) { + const char *name = dp->d_name; + if (name[0] == '.' && + (name[1] == 0 || (name[1] == '.' && name[2] == 0))) + continue; + + (void) snprintf(path, sizeof (path), "%s/%s", rdsk, + dp->d_name); + + if ((fd = open64(path, O_RDONLY)) < 0) + continue; + + /* + * Ignore failed stats. We only want regular + * files, character devs and block devs. + */ + if (fstat64(fd, &statbuf) != 0 || + (!S_ISREG(statbuf.st_mode) && + !S_ISCHR(statbuf.st_mode) && + !S_ISBLK(statbuf.st_mode))) { + (void) close(fd); + continue; + } + + if ((zpool_read_label(fd, &config)) != 0) { + (void) close(fd); + (void) no_memory(hdl); + goto error; + } + + (void) close(fd); + + if (config != NULL) { + boolean_t matched = B_TRUE; + + if (poolname != NULL) { + char *pname; + + matched = nvlist_lookup_string(config, + ZPOOL_CONFIG_POOL_NAME, + &pname) == 0 && + strcmp(poolname, pname) == 0; + } else if (guid != 0) { + uint64_t this_guid; + + matched = nvlist_lookup_uint64(config, + ZPOOL_CONFIG_POOL_GUID, + &this_guid) == 0 && + guid == this_guid; + } + if (!matched) { + nvlist_free(config); + config = NULL; + continue; + } + /* use the non-raw path for the config */ + (void) strlcpy(end, name, pathleft); + if (add_config(hdl, &pools, path, config) != 0) + goto error; + } + } + + (void) closedir(dirp); + dirp = NULL; + } + + ret = get_configs(hdl, &pools, active_ok); error: for (pe = pools.pools; pe != NULL; pe = penext) { @@ -823,9 +1001,158 @@ error: free(ne); } + if (dirp) + (void) closedir(dirp); + return (ret); } +nvlist_t * +zpool_find_import(libzfs_handle_t *hdl, int argc, char **argv) +{ + return (zpool_find_import_impl(hdl, argc, argv, B_FALSE, NULL, 0)); +} + +nvlist_t * +zpool_find_import_byname(libzfs_handle_t *hdl, int argc, char **argv, + char *pool) +{ + return (zpool_find_import_impl(hdl, argc, argv, B_FALSE, pool, 0)); +} + +nvlist_t * +zpool_find_import_byguid(libzfs_handle_t *hdl, int argc, char **argv, + uint64_t guid) +{ + return (zpool_find_import_impl(hdl, argc, argv, B_FALSE, NULL, guid)); +} + +nvlist_t * +zpool_find_import_activeok(libzfs_handle_t *hdl, int argc, char **argv) +{ + return (zpool_find_import_impl(hdl, argc, argv, B_TRUE, NULL, 0)); +} + +/* + * Given a cache file, return the contents as a list of importable pools. + * poolname or guid (but not both) are provided by the caller when trying + * to import a specific pool. + */ +nvlist_t * +zpool_find_import_cached(libzfs_handle_t *hdl, const char *cachefile, + char *poolname, uint64_t guid) +{ + char *buf; + int fd; + struct stat64 statbuf; + nvlist_t *raw, *src, *dst; + nvlist_t *pools; + nvpair_t *elem; + char *name; + uint64_t this_guid; + boolean_t active; + + verify(poolname == NULL || guid == 0); + + if ((fd = open(cachefile, O_RDONLY)) < 0) { + zfs_error_aux(hdl, "%s", strerror(errno)); + (void) zfs_error(hdl, EZFS_BADCACHE, + dgettext(TEXT_DOMAIN, "failed to open cache file")); + return (NULL); + } + + if (fstat64(fd, &statbuf) != 0) { + zfs_error_aux(hdl, "%s", strerror(errno)); + (void) close(fd); + (void) zfs_error(hdl, EZFS_BADCACHE, + dgettext(TEXT_DOMAIN, "failed to get size of cache file")); + return (NULL); + } + + if ((buf = zfs_alloc(hdl, statbuf.st_size)) == NULL) { + (void) close(fd); + return (NULL); + } + + if (read(fd, buf, statbuf.st_size) != statbuf.st_size) { + (void) close(fd); + free(buf); + (void) zfs_error(hdl, EZFS_BADCACHE, + dgettext(TEXT_DOMAIN, + "failed to read cache file contents")); + return (NULL); + } + + (void) close(fd); + + if (nvlist_unpack(buf, statbuf.st_size, &raw, 0) != 0) { + free(buf); + (void) zfs_error(hdl, EZFS_BADCACHE, + dgettext(TEXT_DOMAIN, + "invalid or corrupt cache file contents")); + return (NULL); + } + + free(buf); + + /* + * Go through and get the current state of the pools and refresh their + * state. + */ + if (nvlist_alloc(&pools, 0, 0) != 0) { + (void) no_memory(hdl); + nvlist_free(raw); + return (NULL); + } + + elem = NULL; + while ((elem = nvlist_next_nvpair(raw, elem)) != NULL) { + verify(nvpair_value_nvlist(elem, &src) == 0); + + verify(nvlist_lookup_string(src, ZPOOL_CONFIG_POOL_NAME, + &name) == 0); + if (poolname != NULL && strcmp(poolname, name) != 0) + continue; + + verify(nvlist_lookup_uint64(src, ZPOOL_CONFIG_POOL_GUID, + &this_guid) == 0); + if (guid != 0) { + verify(nvlist_lookup_uint64(src, ZPOOL_CONFIG_POOL_GUID, + &this_guid) == 0); + if (guid != this_guid) + continue; + } + + if (pool_active(hdl, name, this_guid, &active) != 0) { + nvlist_free(raw); + nvlist_free(pools); + return (NULL); + } + + if (active) + continue; + + if ((dst = refresh_config(hdl, src)) == NULL) { + nvlist_free(raw); + nvlist_free(pools); + return (NULL); + } + + if (nvlist_add_nvlist(pools, nvpair_name(elem), dst) != 0) { + (void) no_memory(hdl); + nvlist_free(dst); + nvlist_free(raw); + nvlist_free(pools); + return (NULL); + } + nvlist_free(dst); + } + + nvlist_free(raw); + return (pools); +} + + boolean_t find_guid(nvlist_t *nv, uint64_t guid) { @@ -847,27 +1174,28 @@ find_guid(nvlist_t *nv, uint64_t guid) return (B_FALSE); } -typedef struct spare_cbdata { +typedef struct aux_cbdata { + const char *cb_type; uint64_t cb_guid; zpool_handle_t *cb_zhp; -} spare_cbdata_t; +} aux_cbdata_t; static int -find_spare(zpool_handle_t *zhp, void *data) +find_aux(zpool_handle_t *zhp, void *data) { - spare_cbdata_t *cbp = data; - nvlist_t **spares; - uint_t i, nspares; + aux_cbdata_t *cbp = data; + nvlist_t **list; + uint_t i, count; uint64_t guid; nvlist_t *nvroot; verify(nvlist_lookup_nvlist(zhp->zpool_config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); - if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, - &spares, &nspares) == 0) { - for (i = 0; i < nspares; i++) { - verify(nvlist_lookup_uint64(spares[i], + if (nvlist_lookup_nvlist_array(nvroot, cbp->cb_type, + &list, &count) == 0) { + for (i = 0; i < count; i++) { + verify(nvlist_lookup_uint64(list[i], ZPOOL_CONFIG_GUID, &guid) == 0); if (guid == cbp->cb_guid) { cbp->cb_zhp = zhp; @@ -896,7 +1224,7 @@ zpool_in_use(libzfs_handle_t *hdl, int fd, pool_state_t *state, char **namestr, zpool_handle_t *zhp; nvlist_t *pool_config; uint64_t stateval, isspare; - spare_cbdata_t cb = { 0 }; + aux_cbdata_t cb = { 0 }; boolean_t isactive; *inuse = B_FALSE; @@ -914,7 +1242,7 @@ zpool_in_use(libzfs_handle_t *hdl, int fd, pool_state_t *state, char **namestr, verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_GUID, &vdev_guid) == 0); - if (stateval != POOL_STATE_SPARE) { + if (stateval != POOL_STATE_SPARE && stateval != POOL_STATE_L2CACHE) { verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME, &name) == 0); verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID, @@ -993,7 +1321,24 @@ zpool_in_use(libzfs_handle_t *hdl, int fd, pool_state_t *state, char **namestr, */ cb.cb_zhp = NULL; cb.cb_guid = vdev_guid; - if (zpool_iter(hdl, find_spare, &cb) == 1) { + cb.cb_type = ZPOOL_CONFIG_SPARES; + if (zpool_iter(hdl, find_aux, &cb) == 1) { + name = (char *)zpool_get_name(cb.cb_zhp); + ret = TRUE; + } else { + ret = FALSE; + } + break; + + case POOL_STATE_L2CACHE: + + /* + * Check if any pool is currently using this l2cache device. + */ + cb.cb_zhp = NULL; + cb.cb_guid = vdev_guid; + cb.cb_type = ZPOOL_CONFIG_L2CACHE; + if (zpool_iter(hdl, find_aux, &cb) == 1) { name = (char *)zpool_get_name(cb.cb_zhp); ret = TRUE; } else { @@ -1008,6 +1353,8 @@ zpool_in_use(libzfs_handle_t *hdl, int fd, pool_state_t *state, char **namestr, if (ret) { if ((*namestr = zfs_strdup(hdl, name)) == NULL) { + if (cb.cb_zhp) + zpool_close(cb.cb_zhp); nvlist_free(config); return (-1); } diff --git a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_mount.c b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_mount.c index b4bc945..84a8a57 100644 --- a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_mount.c +++ b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_mount.c @@ -20,12 +20,10 @@ */ /* - * Copyright 2007 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ -#pragma ident "%Z%%M% %I% %E% SMI" - /* * Routines to manage ZFS mounts. We separate all the nasty routines that have * to deal with the OS. The following functions are the main entry points -- @@ -45,11 +43,17 @@ * zfs_unshare() * * zfs_is_shared_nfs() - * zfs_share_nfs() - * zfs_unshare_nfs() - * zfs_unshareall_nfs() + * zfs_is_shared_smb() * zfs_is_shared_iscsi() + * zfs_share_proto() + * zfs_shareall(); * zfs_share_iscsi() + * zfs_unshare_nfs() + * zfs_unshare_smb() + * zfs_unshareall_nfs() + * zfs_unshareall_smb() + * zfs_unshareall() + * zfs_unshareall_bypath() * zfs_unshare_iscsi() * * The following functions are available for pool consumers, and will @@ -78,9 +82,49 @@ #include "libzfs_impl.h" +#include <libshare.h> + +#define MAXISALEN 257 /* based on sysinfo(2) man page */ + +static int zfs_share_proto(zfs_handle_t *, zfs_share_proto_t *); +zfs_share_type_t zfs_is_shared_proto(zfs_handle_t *, char **, + zfs_share_proto_t); + static int (*iscsitgt_zfs_share)(const char *); static int (*iscsitgt_zfs_unshare)(const char *); static int (*iscsitgt_zfs_is_shared)(const char *); +static int (*iscsitgt_svc_online)(); + +/* + * The share protocols table must be in the same order as the zfs_share_prot_t + * enum in libzfs_impl.h + */ +typedef struct { + zfs_prop_t p_prop; + char *p_name; + int p_share_err; + int p_unshare_err; +} proto_table_t; + +proto_table_t proto_table[PROTO_END] = { + {ZFS_PROP_SHARENFS, "nfs", EZFS_SHARENFSFAILED, EZFS_UNSHARENFSFAILED}, + {ZFS_PROP_SHARESMB, "smb", EZFS_SHARESMBFAILED, EZFS_UNSHARESMBFAILED}, +}; + +zfs_share_proto_t nfs_only[] = { + PROTO_NFS, + PROTO_END +}; + +zfs_share_proto_t smb_only[] = { + PROTO_SMB, + PROTO_END +}; +zfs_share_proto_t share_all_proto[] = { + PROTO_NFS, + PROTO_SMB, + PROTO_END +}; #pragma init(zfs_iscsi_init) static void @@ -95,37 +139,65 @@ zfs_iscsi_init(void) (iscsitgt_zfs_unshare = (int (*)(const char *))dlsym(libiscsitgt, "iscsitgt_zfs_unshare")) == NULL || (iscsitgt_zfs_is_shared = (int (*)(const char *))dlsym(libiscsitgt, - "iscsitgt_zfs_is_shared")) == NULL) { + "iscsitgt_zfs_is_shared")) == NULL || + (iscsitgt_svc_online = (int (*)(const char *))dlsym(libiscsitgt, + "iscsitgt_svc_online")) == NULL) { iscsitgt_zfs_share = NULL; iscsitgt_zfs_unshare = NULL; iscsitgt_zfs_is_shared = NULL; + iscsitgt_svc_online = NULL; } } /* - * Search the sharetab for the given mountpoint, returning true if it is found. + * Search the sharetab for the given mountpoint and protocol, returning + * a zfs_share_type_t value. */ -static boolean_t -is_shared(libzfs_handle_t *hdl, const char *mountpoint) +static zfs_share_type_t +is_shared(libzfs_handle_t *hdl, const char *mountpoint, zfs_share_proto_t proto) { char buf[MAXPATHLEN], *tab; + char *ptr; if (hdl->libzfs_sharetab == NULL) - return (0); + return (SHARED_NOT_SHARED); (void) fseek(hdl->libzfs_sharetab, 0, SEEK_SET); while (fgets(buf, sizeof (buf), hdl->libzfs_sharetab) != NULL) { /* the mountpoint is the first entry on each line */ - if ((tab = strchr(buf, '\t')) != NULL) { + if ((tab = strchr(buf, '\t')) == NULL) + continue; + + *tab = '\0'; + if (strcmp(buf, mountpoint) == 0) { + /* + * the protocol field is the third field + * skip over second field + */ + ptr = ++tab; + if ((tab = strchr(ptr, '\t')) == NULL) + continue; + ptr = ++tab; + if ((tab = strchr(ptr, '\t')) == NULL) + continue; *tab = '\0'; - if (strcmp(buf, mountpoint) == 0) - return (B_TRUE); + if (strcmp(ptr, + proto_table[proto].p_name) == 0) { + switch (proto) { + case PROTO_NFS: + return (SHARED_NFS); + case PROTO_SMB: + return (SHARED_SMB); + default: + return (0); + } + } } } - return (B_FALSE); + return (SHARED_NOT_SHARED); } #if 0 @@ -198,10 +270,10 @@ zfs_is_mounted(zfs_handle_t *zhp, char **where) */ static boolean_t zfs_is_mountable(zfs_handle_t *zhp, char *buf, size_t buflen, - zfs_source_t *source) + zprop_source_t *source) { char sourceloc[ZFS_MAXNAMELEN]; - zfs_source_t sourcetype; + zprop_source_t sourcetype; if (!zfs_prop_valid_for_type(ZFS_PROP_MOUNTPOINT, zhp->zfs_type)) return (B_FALSE); @@ -213,7 +285,7 @@ zfs_is_mountable(zfs_handle_t *zhp, char *buf, size_t buflen, strcmp(buf, ZFS_MOUNTPOINT_LEGACY) == 0) return (B_FALSE); - if (!zfs_prop_get_int(zhp, ZFS_PROP_CANMOUNT)) + if (zfs_prop_get_int(zhp, ZFS_PROP_CANMOUNT) == ZFS_CANMOUNT_OFF) return (B_FALSE); if (zfs_prop_get_int(zhp, ZFS_PROP_ZONED) && @@ -281,11 +353,15 @@ zfs_mount(zfs_handle_t *zhp, const char *options, int flags) * from mount(), and they're well-understood. We pick a few * common ones to improve upon. */ - if (errno == EBUSY) + if (errno == EBUSY) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "mountpoint or dataset is busy")); - else + } else if (errno == EPERM) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "Insufficient privileges")); + } else { zfs_error_aux(hdl, strerror(errno)); + } return (zfs_error_fmt(hdl, EZFS_MOUNTFAILED, dgettext(TEXT_DOMAIN, "cannot mount '%s'"), @@ -318,6 +394,7 @@ int zfs_unmount(zfs_handle_t *zhp, const char *mountpoint, int flags) { struct mnttab search = { 0 }, entry; + char *mntpt = NULL; /* check to see if need to unmount the filesystem */ search.mnt_special = zhp->zfs_name; @@ -326,15 +403,29 @@ zfs_unmount(zfs_handle_t *zhp, const char *mountpoint, int flags) if (mountpoint != NULL || ((zfs_get_type(zhp) == ZFS_TYPE_FILESYSTEM) && getmntany(zhp->zfs_hdl->libzfs_mnttab, &entry, &search) == 0)) { + /* + * mountpoint may have come from a call to + * getmnt/getmntany if it isn't NULL. If it is NULL, + * we know it comes from getmntany which can then get + * overwritten later. We strdup it to play it safe. + */ if (mountpoint == NULL) - mountpoint = entry.mnt_mountp; + mntpt = zfs_strdup(zhp->zfs_hdl, entry.mnt_mountp); + else + mntpt = zfs_strdup(zhp->zfs_hdl, mountpoint); /* * Unshare and unmount the filesystem */ - if (zfs_unshare_nfs(zhp, mountpoint) != 0 || - unmount_one(zhp->zfs_hdl, mountpoint, flags) != 0) + if (zfs_unshare_proto(zhp, mntpt, share_all_proto) != 0) + return (-1); + + if (unmount_one(zhp->zfs_hdl, mntpt, flags) != 0) { + free(mntpt); + (void) zfs_shareall(zhp); return (-1); + } + free(mntpt); } return (0); @@ -351,7 +442,7 @@ zfs_unmountall(zfs_handle_t *zhp, int flags) prop_changelist_t *clp; int ret; - clp = changelist_gather(zhp, ZFS_PROP_MOUNTPOINT, flags); + clp = changelist_gather(zhp, ZFS_PROP_MOUNTPOINT, 0, flags); if (clp == NULL) return (-1); @@ -364,10 +455,17 @@ zfs_unmountall(zfs_handle_t *zhp, int flags) boolean_t zfs_is_shared(zfs_handle_t *zhp) { + zfs_share_type_t rc = 0; + zfs_share_proto_t *curr_proto; + if (ZFS_IS_VOLUME(zhp)) return (zfs_is_shared_iscsi(zhp)); - return (zfs_is_shared_nfs(zhp, NULL)); + for (curr_proto = share_all_proto; *curr_proto != PROTO_END; + curr_proto++) + rc |= zfs_is_shared_proto(zhp, NULL, *curr_proto); + + return (rc ? B_TRUE : B_FALSE); } int @@ -376,7 +474,7 @@ zfs_share(zfs_handle_t *zhp) if (ZFS_IS_VOLUME(zhp)) return (zfs_share_iscsi(zhp)); - return (zfs_share_nfs(zhp)); + return (zfs_share_proto(zhp, share_all_proto)); } int @@ -385,139 +483,399 @@ zfs_unshare(zfs_handle_t *zhp) if (ZFS_IS_VOLUME(zhp)) return (zfs_unshare_iscsi(zhp)); - return (zfs_unshare_nfs(zhp, NULL)); + return (zfs_unshareall(zhp)); } /* * Check to see if the filesystem is currently shared. */ -boolean_t -zfs_is_shared_nfs(zfs_handle_t *zhp, char **where) +zfs_share_type_t +zfs_is_shared_proto(zfs_handle_t *zhp, char **where, zfs_share_proto_t proto) { char *mountpoint; + zfs_share_type_t rc; if (!zfs_is_mounted(zhp, &mountpoint)) - return (B_FALSE); + return (SHARED_NOT_SHARED); - if (is_shared(zhp->zfs_hdl, mountpoint)) { + if (rc = is_shared(zhp->zfs_hdl, mountpoint, proto)) { if (where != NULL) *where = mountpoint; else free(mountpoint); - return (B_TRUE); + return (rc); } else { free(mountpoint); - return (B_FALSE); + return (SHARED_NOT_SHARED); + } +} + +boolean_t +zfs_is_shared_nfs(zfs_handle_t *zhp, char **where) +{ + return (zfs_is_shared_proto(zhp, where, + PROTO_NFS) != SHARED_NOT_SHARED); +} + +boolean_t +zfs_is_shared_smb(zfs_handle_t *zhp, char **where) +{ + return (zfs_is_shared_proto(zhp, where, + PROTO_SMB) != SHARED_NOT_SHARED); +} + +/* + * Make sure things will work if libshare isn't installed by using + * wrapper functions that check to see that the pointers to functions + * initialized in _zfs_init_libshare() are actually present. + */ + +#if 0 +static sa_handle_t (*_sa_init)(int); +static void (*_sa_fini)(sa_handle_t); +static sa_share_t (*_sa_find_share)(sa_handle_t, char *); +static int (*_sa_enable_share)(sa_share_t, char *); +static int (*_sa_disable_share)(sa_share_t, char *); +static char *(*_sa_errorstr)(int); +static int (*_sa_parse_legacy_options)(sa_group_t, char *, char *); +static boolean_t (*_sa_needs_refresh)(sa_handle_t *); +static libzfs_handle_t *(*_sa_get_zfs_handle)(sa_handle_t); +static int (*_sa_zfs_process_share)(sa_handle_t, sa_group_t, sa_share_t, + char *, char *, zprop_source_t, char *, char *, char *); +static void (*_sa_update_sharetab_ts)(sa_handle_t); +#endif + +/* + * _zfs_init_libshare() + * + * Find the libshare.so.1 entry points that we use here and save the + * values to be used later. This is triggered by the runtime loader. + * Make sure the correct ISA version is loaded. + */ + +#pragma init(_zfs_init_libshare) +static void +_zfs_init_libshare(void) +{ +#if 0 + void *libshare; + char path[MAXPATHLEN]; + char isa[MAXISALEN]; + +#if defined(_LP64) + if (sysinfo(SI_ARCHITECTURE_64, isa, MAXISALEN) == -1) + isa[0] = '\0'; +#else + isa[0] = '\0'; +#endif + (void) snprintf(path, MAXPATHLEN, + "/usr/lib/%s/libshare.so.1", isa); + + if ((libshare = dlopen(path, RTLD_LAZY | RTLD_GLOBAL)) != NULL) { + _sa_init = (sa_handle_t (*)(int))dlsym(libshare, "sa_init"); + _sa_fini = (void (*)(sa_handle_t))dlsym(libshare, "sa_fini"); + _sa_find_share = (sa_share_t (*)(sa_handle_t, char *)) + dlsym(libshare, "sa_find_share"); + _sa_enable_share = (int (*)(sa_share_t, char *))dlsym(libshare, + "sa_enable_share"); + _sa_disable_share = (int (*)(sa_share_t, char *))dlsym(libshare, + "sa_disable_share"); + _sa_errorstr = (char *(*)(int))dlsym(libshare, "sa_errorstr"); + _sa_parse_legacy_options = (int (*)(sa_group_t, char *, char *)) + dlsym(libshare, "sa_parse_legacy_options"); + _sa_needs_refresh = (boolean_t (*)(sa_handle_t *)) + dlsym(libshare, "sa_needs_refresh"); + _sa_get_zfs_handle = (libzfs_handle_t *(*)(sa_handle_t)) + dlsym(libshare, "sa_get_zfs_handle"); + _sa_zfs_process_share = (int (*)(sa_handle_t, sa_group_t, + sa_share_t, char *, char *, zprop_source_t, char *, + char *, char *))dlsym(libshare, "sa_zfs_process_share"); + _sa_update_sharetab_ts = (void (*)(sa_handle_t)) + dlsym(libshare, "sa_update_sharetab_ts"); + if (_sa_init == NULL || _sa_fini == NULL || + _sa_find_share == NULL || _sa_enable_share == NULL || + _sa_disable_share == NULL || _sa_errorstr == NULL || + _sa_parse_legacy_options == NULL || + _sa_needs_refresh == NULL || _sa_get_zfs_handle == NULL || + _sa_zfs_process_share == NULL || + _sa_update_sharetab_ts == NULL) { + _sa_init = NULL; + _sa_fini = NULL; + _sa_disable_share = NULL; + _sa_enable_share = NULL; + _sa_errorstr = NULL; + _sa_parse_legacy_options = NULL; + (void) dlclose(libshare); + _sa_needs_refresh = NULL; + _sa_get_zfs_handle = NULL; + _sa_zfs_process_share = NULL; + _sa_update_sharetab_ts = NULL; + } } +#endif } /* - * Share the given filesystem according to the options in 'sharenfs'. We rely - * on share(1M) to the dirty work for us. + * zfs_init_libshare(zhandle, service) + * + * Initialize the libshare API if it hasn't already been initialized. + * In all cases it returns 0 if it succeeded and an error if not. The + * service value is which part(s) of the API to initialize and is a + * direct map to the libshare sa_init(service) interface. */ int -zfs_share_nfs(zfs_handle_t *zhp) +zfs_init_libshare(libzfs_handle_t *zhandle, int service) { - char mountpoint[ZFS_MAXPROPLEN]; - char shareopts[ZFS_MAXPROPLEN]; - char buf[MAXPATHLEN]; - FILE *fp; - libzfs_handle_t *hdl = zhp->zfs_hdl; + int ret = SA_OK; - if (!zfs_is_mountable(zhp, mountpoint, sizeof (mountpoint), NULL)) - return (0); +#if 0 + if (_sa_init == NULL) + ret = SA_CONFIG_ERR; - /* - * Return success if there are no share options. - */ - if (zfs_prop_get(zhp, ZFS_PROP_SHARENFS, shareopts, sizeof (shareopts), - NULL, NULL, 0, B_FALSE) != 0 || - strcmp(shareopts, "off") == 0) - return (0); + if (ret == SA_OK && zhandle->libzfs_shareflags & ZFSSHARE_MISS) { + /* + * We had a cache miss. Most likely it is a new ZFS + * dataset that was just created. We want to make sure + * so check timestamps to see if a different process + * has updated any of the configuration. If there was + * some non-ZFS change, we need to re-initialize the + * internal cache. + */ + zhandle->libzfs_shareflags &= ~ZFSSHARE_MISS; + if (_sa_needs_refresh != NULL && + _sa_needs_refresh(zhandle->libzfs_sharehdl)) { + zfs_uninit_libshare(zhandle); + zhandle->libzfs_sharehdl = _sa_init(service); + } + } - /* - * If the 'zoned' property is set, then zfs_is_mountable() will have - * already bailed out if we are in the global zone. But local - * zones cannot be NFS servers, so we ignore it for local zones as well. - */ - if (zfs_prop_get_int(zhp, ZFS_PROP_ZONED)) - return (0); + if (ret == SA_OK && zhandle && zhandle->libzfs_sharehdl == NULL) + zhandle->libzfs_sharehdl = _sa_init(service); -#ifdef __FreeBSD__ - { - int error; + if (ret == SA_OK && zhandle->libzfs_sharehdl == NULL) + ret = SA_NO_MEMORY; +#endif - if (strcmp(shareopts, "on") == 0) - error = fsshare(ZFS_EXPORTS_PATH, mountpoint, ""); - else - error = fsshare(ZFS_EXPORTS_PATH, mountpoint, shareopts); - if (error != 0) { - zfs_error_aux(hdl, "%s", strerror(error)); - (void) zfs_error_fmt(hdl, EZFS_SHARENFSFAILED, - dgettext(TEXT_DOMAIN, "cannot share '%s'"), - zfs_get_name(zhp)); - return (-1); + return (ret); +} + +/* + * zfs_uninit_libshare(zhandle) + * + * Uninitialize the libshare API if it hasn't already been + * uninitialized. It is OK to call multiple times. + */ +void +zfs_uninit_libshare(libzfs_handle_t *zhandle) +{ + if (zhandle != NULL && zhandle->libzfs_sharehdl != NULL) { +#if 0 + if (_sa_fini != NULL) + _sa_fini(zhandle->libzfs_sharehdl); +#endif + zhandle->libzfs_sharehdl = NULL; } +} + +/* + * zfs_parse_options(options, proto) + * + * Call the legacy parse interface to get the protocol specific + * options using the NULL arg to indicate that this is a "parse" only. + */ +int +zfs_parse_options(char *options, zfs_share_proto_t proto) +{ +#if 0 + if (_sa_parse_legacy_options != NULL) { + return (_sa_parse_legacy_options(NULL, options, + proto_table[proto].p_name)); } + return (SA_CONFIG_ERR); #else - /* - * Invoke the share(1M) command. We always do this, even if it's - * currently shared, as the options may have changed. - */ - if (strcmp(shareopts, "on") == 0) - (void) snprintf(buf, sizeof (buf), "/usr/sbin/share " - "-F nfs \"%s\" 2>&1", mountpoint); - else - (void) snprintf(buf, sizeof (buf), "/usr/sbin/share " - "-F nfs -o \"%s\" \"%s\" 2>&1", shareopts, - mountpoint); + return (SA_OK); +#endif +} - if ((fp = popen(buf, "r")) == NULL) - return (zfs_error_fmt(hdl, EZFS_SHARENFSFAILED, - dgettext(TEXT_DOMAIN, "cannot share '%s'"), - zfs_get_name(zhp))); +#if 0 +/* + * zfs_sa_find_share(handle, path) + * + * wrapper around sa_find_share to find a share path in the + * configuration. + */ +static sa_share_t +zfs_sa_find_share(sa_handle_t handle, char *path) +{ + if (_sa_find_share != NULL) + return (_sa_find_share(handle, path)); + return (NULL); +} - /* - * share(1M) should only produce output if there is some kind - * of error. All output begins with "share_nfs: ", so we trim - * this off to get to the real error. - */ - if (fgets(buf, sizeof (buf), fp) != NULL) { - char *colon = strchr(buf, ':'); +/* + * zfs_sa_enable_share(share, proto) + * + * Wrapper for sa_enable_share which enables a share for a specified + * protocol. + */ +static int +zfs_sa_enable_share(sa_share_t share, char *proto) +{ + if (_sa_enable_share != NULL) + return (_sa_enable_share(share, proto)); + return (SA_CONFIG_ERR); +} - while (buf[strlen(buf) - 1] == '\n') - buf[strlen(buf) - 1] = '\0'; +/* + * zfs_sa_disable_share(share, proto) + * + * Wrapper for sa_enable_share which disables a share for a specified + * protocol. + */ +static int +zfs_sa_disable_share(sa_share_t share, char *proto) +{ + if (_sa_disable_share != NULL) + return (_sa_disable_share(share, proto)); + return (SA_CONFIG_ERR); +} +#endif - if (colon != NULL) - zfs_error_aux(hdl, colon + 2); +/* + * Share the given filesystem according to the options in the specified + * protocol specific properties (sharenfs, sharesmb). We rely + * on "libshare" to the dirty work for us. + */ +static int +zfs_share_proto(zfs_handle_t *zhp, zfs_share_proto_t *proto) +{ + char mountpoint[ZFS_MAXPROPLEN]; + char shareopts[ZFS_MAXPROPLEN]; + char sourcestr[ZFS_MAXPROPLEN]; + libzfs_handle_t *hdl = zhp->zfs_hdl; + zfs_share_proto_t *curr_proto; + zprop_source_t sourcetype; + int error, ret; - (void) zfs_error_fmt(hdl, EZFS_SHARENFSFAILED, - dgettext(TEXT_DOMAIN, "cannot share '%s'"), - zfs_get_name(zhp)); + if (!zfs_is_mountable(zhp, mountpoint, sizeof (mountpoint), NULL)) + return (0); - verify(pclose(fp) != 0); - return (-1); - } + for (curr_proto = proto; *curr_proto != PROTO_END; curr_proto++) { + /* + * Return success if there are no share options. + */ + if (zfs_prop_get(zhp, proto_table[*curr_proto].p_prop, + shareopts, sizeof (shareopts), &sourcetype, sourcestr, + ZFS_MAXPROPLEN, B_FALSE) != 0 || + strcmp(shareopts, "off") == 0) + continue; - verify(pclose(fp) == 0); -#endif + /* + * If the 'zoned' property is set, then zfs_is_mountable() + * will have already bailed out if we are in the global zone. + * But local zones cannot be NFS servers, so we ignore it for + * local zones as well. + */ + if (zfs_prop_get_int(zhp, ZFS_PROP_ZONED)) + continue; + if (*curr_proto != PROTO_NFS) { + fprintf(stderr, "Unsupported share protocol: %d.\n", + *curr_proto); + continue; + } + +#if 0 + share = zfs_sa_find_share(hdl->libzfs_sharehdl, mountpoint); + if (share == NULL) { + /* + * This may be a new file system that was just + * created so isn't in the internal cache + * (second time through). Rather than + * reloading the entire configuration, we can + * assume ZFS has done the checking and it is + * safe to add this to the internal + * configuration. + */ + if (_sa_zfs_process_share(hdl->libzfs_sharehdl, + NULL, NULL, mountpoint, + proto_table[*curr_proto].p_name, sourcetype, + shareopts, sourcestr, zhp->zfs_name) != SA_OK) { + (void) zfs_error_fmt(hdl, + proto_table[*curr_proto].p_share_err, + dgettext(TEXT_DOMAIN, "cannot share '%s'"), + zfs_get_name(zhp)); + return (-1); + } + hdl->libzfs_shareflags |= ZFSSHARE_MISS; + share = zfs_sa_find_share(hdl->libzfs_sharehdl, + mountpoint); + } + if (share != NULL) { + int err; + err = zfs_sa_enable_share(share, + proto_table[*curr_proto].p_name); + if (err != SA_OK) { + (void) zfs_error_fmt(hdl, + proto_table[*curr_proto].p_share_err, + dgettext(TEXT_DOMAIN, "cannot share '%s'"), + zfs_get_name(zhp)); + return (-1); + } + } else +#else + if (strcmp(shareopts, "on") == 0) + error = fsshare(ZFS_EXPORTS_PATH, mountpoint, ""); + else + error = fsshare(ZFS_EXPORTS_PATH, mountpoint, shareopts); + if (error != 0) +#endif + { + (void) zfs_error_fmt(hdl, + proto_table[*curr_proto].p_share_err, + dgettext(TEXT_DOMAIN, "cannot share '%s'"), + zfs_get_name(zhp)); + return (-1); + } + } return (0); } + +int +zfs_share_nfs(zfs_handle_t *zhp) +{ + return (zfs_share_proto(zhp, nfs_only)); +} + +int +zfs_share_smb(zfs_handle_t *zhp) +{ + return (zfs_share_proto(zhp, smb_only)); +} + +int +zfs_shareall(zfs_handle_t *zhp) +{ + return (zfs_share_proto(zhp, share_all_proto)); +} + /* * Unshare a filesystem by mountpoint. */ static int -unshare_one(libzfs_handle_t *hdl, const char *name, const char *mountpoint) +unshare_one(libzfs_handle_t *hdl, const char *name, const char *mountpoint, + zfs_share_proto_t proto) { char buf[MAXPATHLEN]; FILE *fp; - -#ifdef __FreeBSD__ - { int error; + if (proto != PROTO_NFS) { + fprintf(stderr, "No SMB support in FreeBSD yet.\n"); + return (EOPNOTSUPP); + } + error = fsunshare(ZFS_EXPORTS_PATH, mountpoint); if (error != 0) { zfs_error_aux(hdl, "%s", strerror(error)); @@ -525,40 +883,6 @@ unshare_one(libzfs_handle_t *hdl, const char *name, const char *mountpoint) dgettext(TEXT_DOMAIN, "cannot unshare '%s'"), name)); } - } -#else - (void) snprintf(buf, sizeof (buf), - "/usr/sbin/unshare \"%s\" 2>&1", - mountpoint); - - if ((fp = popen(buf, "r")) == NULL) - return (zfs_error_fmt(hdl, EZFS_UNSHARENFSFAILED, - dgettext(TEXT_DOMAIN, - "cannot unshare '%s'"), name)); - - /* - * unshare(1M) should only produce output if there is - * some kind of error. All output begins with "unshare - * nfs: ", so we trim this off to get to the real error. - */ - if (fgets(buf, sizeof (buf), fp) != NULL) { - char *colon = strchr(buf, ':'); - - while (buf[strlen(buf) - 1] == '\n') - buf[strlen(buf) - 1] = '\0'; - - if (colon != NULL) - zfs_error_aux(hdl, colon + 2); - - verify(pclose(fp) != 0); - - return (zfs_error_fmt(hdl, EZFS_UNSHARENFSFAILED, - dgettext(TEXT_DOMAIN, - "cannot unshare '%s'"), name)); - } - - verify(pclose(fp) == 0); -#endif return (0); } @@ -567,47 +891,99 @@ unshare_one(libzfs_handle_t *hdl, const char *name, const char *mountpoint) * Unshare the given filesystem. */ int -zfs_unshare_nfs(zfs_handle_t *zhp, const char *mountpoint) +zfs_unshare_proto(zfs_handle_t *zhp, const char *mountpoint, + zfs_share_proto_t *proto) { struct mnttab search = { 0 }, entry; + char *mntpt = NULL; /* check to see if need to unmount the filesystem */ search.mnt_special = (char *)zfs_get_name(zhp); search.mnt_fstype = MNTTYPE_ZFS; rewind(zhp->zfs_hdl->libzfs_mnttab); + if (mountpoint != NULL) + mntpt = zfs_strdup(zhp->zfs_hdl, mountpoint); + if (mountpoint != NULL || ((zfs_get_type(zhp) == ZFS_TYPE_FILESYSTEM) && getmntany(zhp->zfs_hdl->libzfs_mnttab, &entry, &search) == 0)) { + zfs_share_proto_t *curr_proto; if (mountpoint == NULL) - mountpoint = entry.mnt_mountp; + mntpt = zfs_strdup(zhp->zfs_hdl, entry.mnt_mountp); - if (is_shared(zhp->zfs_hdl, mountpoint) && - unshare_one(zhp->zfs_hdl, zhp->zfs_name, mountpoint) != 0) - return (-1); + for (curr_proto = proto; *curr_proto != PROTO_END; + curr_proto++) { + + if (is_shared(zhp->zfs_hdl, mntpt, *curr_proto) && + unshare_one(zhp->zfs_hdl, zhp->zfs_name, + mntpt, *curr_proto) != 0) { + if (mntpt != NULL) + free(mntpt); + return (-1); + } + } } + if (mntpt != NULL) + free(mntpt); return (0); } +int +zfs_unshare_nfs(zfs_handle_t *zhp, const char *mountpoint) +{ + return (zfs_unshare_proto(zhp, mountpoint, nfs_only)); +} + +int +zfs_unshare_smb(zfs_handle_t *zhp, const char *mountpoint) +{ + return (zfs_unshare_proto(zhp, mountpoint, smb_only)); +} + /* - * Same as zfs_unmountall(), but for NFS unshares. + * Same as zfs_unmountall(), but for NFS and SMB unshares. */ int -zfs_unshareall_nfs(zfs_handle_t *zhp) +zfs_unshareall_proto(zfs_handle_t *zhp, zfs_share_proto_t *proto) { prop_changelist_t *clp; int ret; - clp = changelist_gather(zhp, ZFS_PROP_SHARENFS, 0); + clp = changelist_gather(zhp, ZFS_PROP_SHARENFS, 0, 0); if (clp == NULL) return (-1); - ret = changelist_unshare(clp); + ret = changelist_unshare(clp, proto); changelist_free(clp); return (ret); } +int +zfs_unshareall_nfs(zfs_handle_t *zhp) +{ + return (zfs_unshareall_proto(zhp, nfs_only)); +} + +int +zfs_unshareall_smb(zfs_handle_t *zhp) +{ + return (zfs_unshareall_proto(zhp, smb_only)); +} + +int +zfs_unshareall(zfs_handle_t *zhp) +{ + return (zfs_unshareall_proto(zhp, share_all_proto)); +} + +int +zfs_unshareall_bypath(zfs_handle_t *zhp, const char *mountpoint) +{ + return (zfs_unshare_proto(zhp, mountpoint, share_all_proto)); +} + /* * Remove the mountpoint associated with the current dataset, if necessary. * We only remove the underlying directory if: @@ -623,14 +999,14 @@ void remove_mountpoint(zfs_handle_t *zhp) { char mountpoint[ZFS_MAXPROPLEN]; - zfs_source_t source; + zprop_source_t source; if (!zfs_is_mountable(zhp, mountpoint, sizeof (mountpoint), &source)) return; - if (source == ZFS_SRC_DEFAULT || - source == ZFS_SRC_INHERITED) { + if (source == ZPROP_SRC_DEFAULT || + source == ZPROP_SRC_INHERITED) { /* * Try to remove the directory, silently ignoring any errors. * The filesystem may have since been removed or moved around, @@ -644,8 +1020,15 @@ remove_mountpoint(zfs_handle_t *zhp) boolean_t zfs_is_shared_iscsi(zfs_handle_t *zhp) { - return (iscsitgt_zfs_is_shared != NULL && - iscsitgt_zfs_is_shared(zhp->zfs_name) != 0); + + /* + * If iscsi deamon isn't running then we aren't shared + */ + if (iscsitgt_svc_online && iscsitgt_svc_online() == 1) + return (B_FALSE); + else + return (iscsitgt_zfs_is_shared != NULL && + iscsitgt_zfs_is_shared(zhp->zfs_name) != 0); } int @@ -665,9 +1048,20 @@ zfs_share_iscsi(zfs_handle_t *zhp) /* We don't support iSCSI on FreeBSD yet. */ #ifdef TODO - if (iscsitgt_zfs_share == NULL || iscsitgt_zfs_share(dataset) != 0) - return (zfs_error_fmt(hdl, EZFS_SHAREISCSIFAILED, + if (iscsitgt_zfs_share == NULL || iscsitgt_zfs_share(dataset) != 0) { + int error = EZFS_SHAREISCSIFAILED; + + /* + * If service isn't availabele and EPERM was + * returned then use special error. + */ + if (iscsitgt_svc_online && errno == EPERM && + (iscsitgt_svc_online() != 0)) + error = EZFS_ISCSISVCUNAVAIL; + + return (zfs_error_fmt(hdl, error, dgettext(TEXT_DOMAIN, "cannot share '%s'"), dataset)); + } #endif return (0); @@ -684,7 +1078,7 @@ zfs_unshare_iscsi(zfs_handle_t *zhp) /* * Return if the volume is not shared */ - if (!zfs_is_shared_iscsi(zhp)) + if (zfs_is_shared_iscsi(zhp) != SHARED_ISCSI) return (0); /* @@ -692,9 +1086,13 @@ zfs_unshare_iscsi(zfs_handle_t *zhp) * we should return success in that case. */ if (iscsitgt_zfs_unshare == NULL || - (iscsitgt_zfs_unshare(dataset) != 0 && errno != ENODEV)) + (iscsitgt_zfs_unshare(dataset) != 0 && errno != ENODEV)) { + if (errno == EPERM) + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "Insufficient privileges to unshare iscsi")); return (zfs_error_fmt(hdl, EZFS_UNSHAREISCSIFAILED, dgettext(TEXT_DOMAIN, "cannot unshare '%s'"), dataset)); + } #endif return (0); @@ -716,6 +1114,11 @@ mount_cb(zfs_handle_t *zhp, void *data) return (0); } + if (zfs_prop_get_int(zhp, ZFS_PROP_CANMOUNT) == ZFS_CANMOUNT_NOAUTO) { + zfs_close(zhp); + return (0); + } + if (cbp->cb_alloc == cbp->cb_used) { void *ptr; @@ -730,7 +1133,7 @@ mount_cb(zfs_handle_t *zhp, void *data) cbp->cb_datasets[cbp->cb_used++] = zhp; - return (zfs_iter_children(zhp, mount_cb, cbp)); + return (zfs_iter_filesystems(zhp, mount_cb, cbp)); } static int @@ -776,21 +1179,22 @@ zpool_enable_datasets(zpool_handle_t *zhp, const char *mntopts, int flags) libzfs_handle_t *hdl = zhp->zpool_hdl; zfs_handle_t *zfsp; int i, ret = -1; + int *good; /* - * Gather all datasets within the pool. + * Gather all non-snap datasets within the pool. */ if ((cb.cb_datasets = zfs_alloc(hdl, 4 * sizeof (void *))) == NULL) return (-1); cb.cb_alloc = 4; - if ((zfsp = zfs_open(hdl, zhp->zpool_name, ZFS_TYPE_ANY)) == NULL) + if ((zfsp = zfs_open(hdl, zhp->zpool_name, ZFS_TYPE_DATASET)) == NULL) goto out; cb.cb_datasets[0] = zfsp; cb.cb_used = 1; - if (zfs_iter_children(zfsp, mount_cb, &cb) != 0) + if (zfs_iter_filesystems(zfsp, mount_cb, &cb) != 0) goto out; /* @@ -799,15 +1203,32 @@ zpool_enable_datasets(zpool_handle_t *zhp, const char *mntopts, int flags) qsort(cb.cb_datasets, cb.cb_used, sizeof (void *), dataset_cmp); /* - * And mount all the datasets. + * And mount all the datasets, keeping track of which ones + * succeeded or failed. By using zfs_alloc(), the good pointer + * will always be non-NULL. */ + good = zfs_alloc(zhp->zpool_hdl, cb.cb_used * sizeof (int)); ret = 0; for (i = 0; i < cb.cb_used; i++) { - if (zfs_mount(cb.cb_datasets[i], mntopts, flags) != 0 || - zfs_share(cb.cb_datasets[i]) != 0) + if (zfs_mount(cb.cb_datasets[i], mntopts, flags) != 0) + ret = -1; + else + good[i] = 1; + } + + /* + * Then share all the ones that need to be shared. This needs + * to be a separate pass in order to avoid excessive reloading + * of the configuration. Good should never be NULL since + * zfs_alloc is supposed to exit if memory isn't available. + */ + for (i = 0; i < cb.cb_used; i++) { + if (good[i] && zfs_share(cb.cb_datasets[i]) != 0) ret = -1; } + free(good); + out: for (i = 0; i < cb.cb_used; i++) zfs_close(cb.cb_datasets[i]); @@ -830,7 +1251,8 @@ zvol_cb(const char *dataset, void *data) (zhp = zfs_open(hdl, dataset, ZFS_TYPE_VOLUME)) == NULL) return (0); - (void) zfs_unshare_iscsi(zhp); + if (zfs_unshare_iscsi(zhp) != 0) + return (-1); zfs_close(zhp); @@ -953,9 +1375,14 @@ zpool_disable_datasets(zpool_handle_t *zhp, boolean_t force) * Walk through and first unshare everything. */ for (i = 0; i < used; i++) { - if (is_shared(hdl, mountpoints[i]) && - unshare_one(hdl, mountpoints[i], mountpoints[i]) != 0) - goto out; + zfs_share_proto_t *curr_proto; + for (curr_proto = share_all_proto; *curr_proto != PROTO_END; + curr_proto++) { + if (is_shared(hdl, mountpoints[i], *curr_proto) && + unshare_one(hdl, mountpoints[i], + mountpoints[i], *curr_proto) != 0) + goto out; + } } /* diff --git a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_pool.c b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_pool.c index 8580837..c3cb547 100644 --- a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_pool.c +++ b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_pool.c @@ -20,12 +20,10 @@ */ /* - * Copyright 2007 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ -#pragma ident "%Z%%M% %I% %E% SMI" - #include <sys/types.h> #include <sys/stat.h> #include <assert.h> @@ -39,6 +37,7 @@ #include <stdlib.h> #include <strings.h> #include <unistd.h> +#include <zone.h> #include <sys/zfs_ioctl.h> #include <sys/zio.h> #include <strings.h> @@ -48,11 +47,554 @@ #include "zfs_prop.h" #include "libzfs_impl.h" +static int read_efi_label(nvlist_t *config, diskaddr_t *sb); + +/* + * ==================================================================== + * zpool property functions + * ==================================================================== + */ + +static int +zpool_get_all_props(zpool_handle_t *zhp) +{ + zfs_cmd_t zc = { 0 }; + libzfs_handle_t *hdl = zhp->zpool_hdl; + + (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); + + if (zcmd_alloc_dst_nvlist(hdl, &zc, 0) != 0) + return (-1); + + while (ioctl(hdl->libzfs_fd, ZFS_IOC_POOL_GET_PROPS, &zc) != 0) { + if (errno == ENOMEM) { + if (zcmd_expand_dst_nvlist(hdl, &zc) != 0) { + zcmd_free_nvlists(&zc); + return (-1); + } + } else { + zcmd_free_nvlists(&zc); + return (-1); + } + } + + if (zcmd_read_dst_nvlist(hdl, &zc, &zhp->zpool_props) != 0) { + zcmd_free_nvlists(&zc); + return (-1); + } + + zcmd_free_nvlists(&zc); + + return (0); +} + +static int +zpool_props_refresh(zpool_handle_t *zhp) +{ + nvlist_t *old_props; + + old_props = zhp->zpool_props; + + if (zpool_get_all_props(zhp) != 0) + return (-1); + + nvlist_free(old_props); + return (0); +} + +static char * +zpool_get_prop_string(zpool_handle_t *zhp, zpool_prop_t prop, + zprop_source_t *src) +{ + nvlist_t *nv, *nvl; + uint64_t ival; + char *value; + zprop_source_t source; + + nvl = zhp->zpool_props; + if (nvlist_lookup_nvlist(nvl, zpool_prop_to_name(prop), &nv) == 0) { + verify(nvlist_lookup_uint64(nv, ZPROP_SOURCE, &ival) == 0); + source = ival; + verify(nvlist_lookup_string(nv, ZPROP_VALUE, &value) == 0); + } else { + source = ZPROP_SRC_DEFAULT; + if ((value = (char *)zpool_prop_default_string(prop)) == NULL) + value = "-"; + } + + if (src) + *src = source; + + return (value); +} + +uint64_t +zpool_get_prop_int(zpool_handle_t *zhp, zpool_prop_t prop, zprop_source_t *src) +{ + nvlist_t *nv, *nvl; + uint64_t value; + zprop_source_t source; + + if (zhp->zpool_props == NULL && zpool_get_all_props(zhp)) { + /* + * zpool_get_all_props() has most likely failed because + * the pool is faulted, but if all we need is the top level + * vdev's guid then get it from the zhp config nvlist. + */ + if ((prop == ZPOOL_PROP_GUID) && + (nvlist_lookup_nvlist(zhp->zpool_config, + ZPOOL_CONFIG_VDEV_TREE, &nv) == 0) && + (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &value) + == 0)) { + return (value); + } + return (zpool_prop_default_numeric(prop)); + } + + nvl = zhp->zpool_props; + if (nvlist_lookup_nvlist(nvl, zpool_prop_to_name(prop), &nv) == 0) { + verify(nvlist_lookup_uint64(nv, ZPROP_SOURCE, &value) == 0); + source = value; + verify(nvlist_lookup_uint64(nv, ZPROP_VALUE, &value) == 0); + } else { + source = ZPROP_SRC_DEFAULT; + value = zpool_prop_default_numeric(prop); + } + + if (src) + *src = source; + + return (value); +} + +/* + * Map VDEV STATE to printed strings. + */ +char * +zpool_state_to_name(vdev_state_t state, vdev_aux_t aux) +{ + switch (state) { + case VDEV_STATE_CLOSED: + case VDEV_STATE_OFFLINE: + return (gettext("OFFLINE")); + case VDEV_STATE_REMOVED: + return (gettext("REMOVED")); + case VDEV_STATE_CANT_OPEN: + if (aux == VDEV_AUX_CORRUPT_DATA || aux == VDEV_AUX_BAD_LOG) + return (gettext("FAULTED")); + else + return (gettext("UNAVAIL")); + case VDEV_STATE_FAULTED: + return (gettext("FAULTED")); + case VDEV_STATE_DEGRADED: + return (gettext("DEGRADED")); + case VDEV_STATE_HEALTHY: + return (gettext("ONLINE")); + } + + return (gettext("UNKNOWN")); +} + +/* + * Get a zpool property value for 'prop' and return the value in + * a pre-allocated buffer. + */ +int +zpool_get_prop(zpool_handle_t *zhp, zpool_prop_t prop, char *buf, size_t len, + zprop_source_t *srctype) +{ + uint64_t intval; + const char *strval; + zprop_source_t src = ZPROP_SRC_NONE; + nvlist_t *nvroot; + vdev_stat_t *vs; + uint_t vsc; + + if (zpool_get_state(zhp) == POOL_STATE_UNAVAIL) { + if (prop == ZPOOL_PROP_NAME) + (void) strlcpy(buf, zpool_get_name(zhp), len); + else if (prop == ZPOOL_PROP_HEALTH) + (void) strlcpy(buf, "FAULTED", len); + else + (void) strlcpy(buf, "-", len); + return (0); + } + + if (zhp->zpool_props == NULL && zpool_get_all_props(zhp) && + prop != ZPOOL_PROP_NAME) + return (-1); + + switch (zpool_prop_get_type(prop)) { + case PROP_TYPE_STRING: + (void) strlcpy(buf, zpool_get_prop_string(zhp, prop, &src), + len); + break; + + case PROP_TYPE_NUMBER: + intval = zpool_get_prop_int(zhp, prop, &src); + + switch (prop) { + case ZPOOL_PROP_SIZE: + case ZPOOL_PROP_USED: + case ZPOOL_PROP_AVAILABLE: + (void) zfs_nicenum(intval, buf, len); + break; + + case ZPOOL_PROP_CAPACITY: + (void) snprintf(buf, len, "%llu%%", + (u_longlong_t)intval); + break; + + case ZPOOL_PROP_HEALTH: + verify(nvlist_lookup_nvlist(zpool_get_config(zhp, NULL), + ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); + verify(nvlist_lookup_uint64_array(nvroot, + ZPOOL_CONFIG_STATS, (uint64_t **)&vs, &vsc) == 0); + + (void) strlcpy(buf, zpool_state_to_name(intval, + vs->vs_aux), len); + break; + default: + (void) snprintf(buf, len, "%llu", intval); + } + break; + + case PROP_TYPE_INDEX: + intval = zpool_get_prop_int(zhp, prop, &src); + if (zpool_prop_index_to_string(prop, intval, &strval) + != 0) + return (-1); + (void) strlcpy(buf, strval, len); + break; + + default: + abort(); + } + + if (srctype) + *srctype = src; + + return (0); +} + +/* + * Check if the bootfs name has the same pool name as it is set to. + * Assuming bootfs is a valid dataset name. + */ +static boolean_t +bootfs_name_valid(const char *pool, char *bootfs) +{ + int len = strlen(pool); + + if (!zfs_name_valid(bootfs, ZFS_TYPE_FILESYSTEM|ZFS_TYPE_SNAPSHOT)) + return (B_FALSE); + + if (strncmp(pool, bootfs, len) == 0 && + (bootfs[len] == '/' || bootfs[len] == '\0')) + return (B_TRUE); + + return (B_FALSE); +} + +#if defined(sun) +/* + * Inspect the configuration to determine if any of the devices contain + * an EFI label. + */ +static boolean_t +pool_uses_efi(nvlist_t *config) +{ + nvlist_t **child; + uint_t c, children; + + if (nvlist_lookup_nvlist_array(config, ZPOOL_CONFIG_CHILDREN, + &child, &children) != 0) + return (read_efi_label(config, NULL) >= 0); + + for (c = 0; c < children; c++) { + if (pool_uses_efi(child[c])) + return (B_TRUE); + } + return (B_FALSE); +} +#endif + +/* + * Given an nvlist of zpool properties to be set, validate that they are + * correct, and parse any numeric properties (index, boolean, etc) if they are + * specified as strings. + */ +static nvlist_t * +zpool_valid_proplist(libzfs_handle_t *hdl, const char *poolname, + nvlist_t *props, uint64_t version, boolean_t create_or_import, char *errbuf) +{ + nvpair_t *elem; + nvlist_t *retprops; + zpool_prop_t prop; + char *strval; + uint64_t intval; + char *slash; + struct stat64 statbuf; + zpool_handle_t *zhp; + nvlist_t *nvroot; + + if (nvlist_alloc(&retprops, NV_UNIQUE_NAME, 0) != 0) { + (void) no_memory(hdl); + return (NULL); + } + + elem = NULL; + while ((elem = nvlist_next_nvpair(props, elem)) != NULL) { + const char *propname = nvpair_name(elem); + + /* + * Make sure this property is valid and applies to this type. + */ + if ((prop = zpool_name_to_prop(propname)) == ZPROP_INVAL) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "invalid property '%s'"), propname); + (void) zfs_error(hdl, EZFS_BADPROP, errbuf); + goto error; + } + + if (zpool_prop_readonly(prop)) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' " + "is readonly"), propname); + (void) zfs_error(hdl, EZFS_PROPREADONLY, errbuf); + goto error; + } + + if (zprop_parse_value(hdl, elem, prop, ZFS_TYPE_POOL, retprops, + &strval, &intval, errbuf) != 0) + goto error; + + /* + * Perform additional checking for specific properties. + */ + switch (prop) { + case ZPOOL_PROP_VERSION: + if (intval < version || intval > SPA_VERSION) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "property '%s' number %d is invalid."), + propname, intval); + (void) zfs_error(hdl, EZFS_BADVERSION, errbuf); + goto error; + } + break; + + case ZPOOL_PROP_BOOTFS: + if (create_or_import) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "property '%s' cannot be set at creation " + "or import time"), propname); + (void) zfs_error(hdl, EZFS_BADPROP, errbuf); + goto error; + } + + if (version < SPA_VERSION_BOOTFS) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "pool must be upgraded to support " + "'%s' property"), propname); + (void) zfs_error(hdl, EZFS_BADVERSION, errbuf); + goto error; + } + + /* + * bootfs property value has to be a dataset name and + * the dataset has to be in the same pool as it sets to. + */ + if (strval[0] != '\0' && !bootfs_name_valid(poolname, + strval)) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "'%s' " + "is an invalid name"), strval); + (void) zfs_error(hdl, EZFS_INVALIDNAME, errbuf); + goto error; + } + + if ((zhp = zpool_open_canfail(hdl, poolname)) == NULL) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "could not open pool '%s'"), poolname); + (void) zfs_error(hdl, EZFS_OPENFAILED, errbuf); + goto error; + } + verify(nvlist_lookup_nvlist(zpool_get_config(zhp, NULL), + ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); + +#if defined(sun) + /* + * bootfs property cannot be set on a disk which has + * been EFI labeled. + */ + if (pool_uses_efi(nvroot)) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "property '%s' not supported on " + "EFI labeled devices"), propname); + (void) zfs_error(hdl, EZFS_POOL_NOTSUP, errbuf); + zpool_close(zhp); + goto error; + } +#endif + zpool_close(zhp); + break; + + case ZPOOL_PROP_ALTROOT: + if (!create_or_import) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "property '%s' can only be set during pool " + "creation or import"), propname); + (void) zfs_error(hdl, EZFS_BADPROP, errbuf); + goto error; + } + + if (strval[0] != '/') { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "bad alternate root '%s'"), strval); + (void) zfs_error(hdl, EZFS_BADPATH, errbuf); + goto error; + } + break; + + case ZPOOL_PROP_CACHEFILE: + if (strval[0] == '\0') + break; + + if (strcmp(strval, "none") == 0) + break; + + if (strval[0] != '/') { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "property '%s' must be empty, an " + "absolute path, or 'none'"), propname); + (void) zfs_error(hdl, EZFS_BADPATH, errbuf); + goto error; + } + + slash = strrchr(strval, '/'); + + if (slash[1] == '\0' || strcmp(slash, "/.") == 0 || + strcmp(slash, "/..") == 0) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "'%s' is not a valid file"), strval); + (void) zfs_error(hdl, EZFS_BADPATH, errbuf); + goto error; + } + + *slash = '\0'; + + if (strval[0] != '\0' && + (stat64(strval, &statbuf) != 0 || + !S_ISDIR(statbuf.st_mode))) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "'%s' is not a valid directory"), + strval); + (void) zfs_error(hdl, EZFS_BADPATH, errbuf); + goto error; + } + + *slash = '/'; + break; + } + } + + return (retprops); +error: + nvlist_free(retprops); + return (NULL); +} + +/* + * Set zpool property : propname=propval. + */ +int +zpool_set_prop(zpool_handle_t *zhp, const char *propname, const char *propval) +{ + zfs_cmd_t zc = { 0 }; + int ret = -1; + char errbuf[1024]; + nvlist_t *nvl = NULL; + nvlist_t *realprops; + uint64_t version; + + (void) snprintf(errbuf, sizeof (errbuf), + dgettext(TEXT_DOMAIN, "cannot set property for '%s'"), + zhp->zpool_name); + + if (zhp->zpool_props == NULL && zpool_get_all_props(zhp)) + return (zfs_error(zhp->zpool_hdl, EZFS_POOLPROPS, errbuf)); + + if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0) + return (no_memory(zhp->zpool_hdl)); + + if (nvlist_add_string(nvl, propname, propval) != 0) { + nvlist_free(nvl); + return (no_memory(zhp->zpool_hdl)); + } + + version = zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL); + if ((realprops = zpool_valid_proplist(zhp->zpool_hdl, + zhp->zpool_name, nvl, version, B_FALSE, errbuf)) == NULL) { + nvlist_free(nvl); + return (-1); + } + + nvlist_free(nvl); + nvl = realprops; + + /* + * Execute the corresponding ioctl() to set this property. + */ + (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); + + if (zcmd_write_src_nvlist(zhp->zpool_hdl, &zc, nvl) != 0) { + nvlist_free(nvl); + return (-1); + } + + ret = zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_POOL_SET_PROPS, &zc); + + zcmd_free_nvlists(&zc); + nvlist_free(nvl); + + if (ret) + (void) zpool_standard_error(zhp->zpool_hdl, errno, errbuf); + else + (void) zpool_props_refresh(zhp); + + return (ret); +} + +int +zpool_expand_proplist(zpool_handle_t *zhp, zprop_list_t **plp) +{ + libzfs_handle_t *hdl = zhp->zpool_hdl; + zprop_list_t *entry; + char buf[ZFS_MAXPROPLEN]; + + if (zprop_expand_list(hdl, plp, ZFS_TYPE_POOL) != 0) + return (-1); + + for (entry = *plp; entry != NULL; entry = entry->pl_next) { + + if (entry->pl_fixed) + continue; + + if (entry->pl_prop != ZPROP_INVAL && + zpool_get_prop(zhp, entry->pl_prop, buf, sizeof (buf), + NULL) == 0) { + if (strlen(buf) > entry->pl_width) + entry->pl_width = strlen(buf); + } + } + + return (0); +} + + /* * Validate the given pool name, optionally putting an extended error message in * 'buf'. */ -static boolean_t +boolean_t zpool_name_valid(libzfs_handle_t *hdl, boolean_t isopen, const char *pool) { namecheck_err_t why; @@ -70,9 +612,11 @@ zpool_name_valid(libzfs_handle_t *hdl, boolean_t isopen, const char *pool) if (ret == 0 && !isopen && (strncmp(pool, "mirror", 6) == 0 || strncmp(pool, "raidz", 5) == 0 || - strncmp(pool, "spare", 5) == 0)) { - zfs_error_aux(hdl, - dgettext(TEXT_DOMAIN, "name is reserved")); + strncmp(pool, "spare", 5) == 0 || + strcmp(pool, "log") == 0)) { + if (hdl != NULL) + zfs_error_aux(hdl, + dgettext(TEXT_DOMAIN, "name is reserved")); return (B_FALSE); } @@ -134,39 +678,6 @@ zpool_name_valid(libzfs_handle_t *hdl, boolean_t isopen, const char *pool) return (B_TRUE); } -static int -zpool_get_all_props(zpool_handle_t *zhp) -{ - zfs_cmd_t zc = { 0 }; - libzfs_handle_t *hdl = zhp->zpool_hdl; - - (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); - - if (zcmd_alloc_dst_nvlist(hdl, &zc, 0) != 0) - return (-1); - - while (ioctl(hdl->libzfs_fd, ZFS_IOC_POOL_GET_PROPS, &zc) != 0) { - if (errno == ENOMEM) { - if (zcmd_expand_dst_nvlist(hdl, &zc) != 0) { - zcmd_free_nvlists(&zc); - return (-1); - } - } else { - zcmd_free_nvlists(&zc); - return (-1); - } - } - - if (zcmd_read_dst_nvlist(hdl, &zc, &zhp->zpool_props) != 0) { - zcmd_free_nvlists(&zc); - return (-1); - } - - zcmd_free_nvlists(&zc); - - return (0); -} - /* * Open a handle to the given pool, even if the pool is currently in the FAULTED * state. @@ -199,11 +710,9 @@ zpool_open_canfail(libzfs_handle_t *hdl, const char *pool) } if (missing) { - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "no such pool")); + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "no such pool")); (void) zfs_error_fmt(hdl, EZFS_NOENT, - dgettext(TEXT_DOMAIN, "cannot open '%s'"), - pool); + dgettext(TEXT_DOMAIN, "cannot open '%s'"), pool); zpool_close(zhp); return (NULL); } @@ -288,86 +797,6 @@ zpool_get_name(zpool_handle_t *zhp) return (zhp->zpool_name); } -/* - * Return the GUID of the pool. - */ -uint64_t -zpool_get_guid(zpool_handle_t *zhp) -{ - uint64_t guid; - - verify(nvlist_lookup_uint64(zhp->zpool_config, ZPOOL_CONFIG_POOL_GUID, - &guid) == 0); - return (guid); -} - -/* - * Return the version of the pool. - */ -uint64_t -zpool_get_version(zpool_handle_t *zhp) -{ - uint64_t version; - - verify(nvlist_lookup_uint64(zhp->zpool_config, ZPOOL_CONFIG_VERSION, - &version) == 0); - - return (version); -} - -/* - * Return the amount of space currently consumed by the pool. - */ -uint64_t -zpool_get_space_used(zpool_handle_t *zhp) -{ - nvlist_t *nvroot; - vdev_stat_t *vs; - uint_t vsc; - - verify(nvlist_lookup_nvlist(zhp->zpool_config, ZPOOL_CONFIG_VDEV_TREE, - &nvroot) == 0); - verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_STATS, - (uint64_t **)&vs, &vsc) == 0); - - return (vs->vs_alloc); -} - -/* - * Return the total space in the pool. - */ -uint64_t -zpool_get_space_total(zpool_handle_t *zhp) -{ - nvlist_t *nvroot; - vdev_stat_t *vs; - uint_t vsc; - - verify(nvlist_lookup_nvlist(zhp->zpool_config, ZPOOL_CONFIG_VDEV_TREE, - &nvroot) == 0); - verify(nvlist_lookup_uint64_array(nvroot, ZPOOL_CONFIG_STATS, - (uint64_t **)&vs, &vsc) == 0); - - return (vs->vs_space); -} - -/* - * Return the alternate root for this pool, if any. - */ -int -zpool_get_root(zpool_handle_t *zhp, char *buf, size_t buflen) -{ - zfs_cmd_t zc = { 0 }; - - (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); - if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_OBJSET_STATS, &zc) != 0 || - zc.zc_value[0] == '\0') - return (-1); - - (void) strlcpy(buf, zc.zc_value, buflen); - - return (0); -} /* * Return the state of the pool (ACTIVE or UNAVAILABLE) @@ -385,10 +814,14 @@ zpool_get_state(zpool_handle_t *zhp) */ int zpool_create(libzfs_handle_t *hdl, const char *pool, nvlist_t *nvroot, - const char *altroot) + nvlist_t *props, nvlist_t *fsprops) { zfs_cmd_t zc = { 0 }; + nvlist_t *zc_fsprops = NULL; + nvlist_t *zc_props = NULL; char msg[1024]; + char *altroot; + int ret = -1; (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot create '%s'"), pool); @@ -396,20 +829,48 @@ zpool_create(libzfs_handle_t *hdl, const char *pool, nvlist_t *nvroot, if (!zpool_name_valid(hdl, B_FALSE, pool)) return (zfs_error(hdl, EZFS_INVALIDNAME, msg)); - if (altroot != NULL && altroot[0] != '/') - return (zfs_error_fmt(hdl, EZFS_BADPATH, - dgettext(TEXT_DOMAIN, "bad alternate root '%s'"), altroot)); - - if (zcmd_write_src_nvlist(hdl, &zc, nvroot, NULL) != 0) + if (zcmd_write_conf_nvlist(hdl, &zc, nvroot) != 0) return (-1); + if (props) { + if ((zc_props = zpool_valid_proplist(hdl, pool, props, + SPA_VERSION_1, B_TRUE, msg)) == NULL) { + goto create_failed; + } + } + + if (fsprops) { + uint64_t zoned; + char *zonestr; + + zoned = ((nvlist_lookup_string(fsprops, + zfs_prop_to_name(ZFS_PROP_ZONED), &zonestr) == 0) && + strcmp(zonestr, "on") == 0); + + if ((zc_fsprops = zfs_valid_proplist(hdl, + ZFS_TYPE_FILESYSTEM, fsprops, zoned, NULL, msg)) == NULL) { + goto create_failed; + } + if (!zc_props && + (nvlist_alloc(&zc_props, NV_UNIQUE_NAME, 0) != 0)) { + goto create_failed; + } + if (nvlist_add_nvlist(zc_props, + ZPOOL_ROOTFS_PROPS, zc_fsprops) != 0) { + goto create_failed; + } + } + + if (zc_props && zcmd_write_src_nvlist(hdl, &zc, zc_props) != 0) + goto create_failed; + (void) strlcpy(zc.zc_name, pool, sizeof (zc.zc_name)); - if (altroot != NULL) - (void) strlcpy(zc.zc_value, altroot, sizeof (zc.zc_value)); + if ((ret = zfs_ioctl(hdl, ZFS_IOC_POOL_CREATE, &zc)) != 0) { - if (ioctl(hdl->libzfs_fd, ZFS_IOC_POOL_CREATE, &zc) != 0) { zcmd_free_nvlists(&zc); + nvlist_free(zc_props); + nvlist_free(zc_fsprops); switch (errno) { case EBUSY: @@ -446,28 +907,36 @@ zpool_create(libzfs_handle_t *hdl, const char *pool, nvlist_t *nvroot, "one or more devices is out of space")); return (zfs_error(hdl, EZFS_BADDEV, msg)); + case ENOTBLK: + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "cache device must be a disk or disk slice")); + return (zfs_error(hdl, EZFS_BADDEV, msg)); + default: return (zpool_standard_error(hdl, errno, msg)); } } - zcmd_free_nvlists(&zc); - /* * If this is an alternate root pool, then we automatically set the * mountpoint of the root dataset to be '/'. */ - if (altroot != NULL) { + if (nvlist_lookup_string(props, zpool_prop_to_name(ZPOOL_PROP_ALTROOT), + &altroot) == 0) { zfs_handle_t *zhp; - verify((zhp = zfs_open(hdl, pool, ZFS_TYPE_ANY)) != NULL); + verify((zhp = zfs_open(hdl, pool, ZFS_TYPE_DATASET)) != NULL); verify(zfs_prop_set(zhp, zfs_prop_to_name(ZFS_PROP_MOUNTPOINT), "/") == 0); zfs_close(zhp); } - return (0); +create_failed: + zcmd_free_nvlists(&zc); + nvlist_free(zc_props); + nvlist_free(zc_fsprops); + return (ret); } /* @@ -492,7 +961,7 @@ zpool_destroy(zpool_handle_t *zhp) (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); - if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_POOL_DESTROY, &zc) != 0) { + if (zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_POOL_DESTROY, &zc) != 0) { (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot destroy '%s'"), zhp->zpool_name); @@ -528,13 +997,14 @@ zpool_add(zpool_handle_t *zhp, nvlist_t *nvroot) int ret; libzfs_handle_t *hdl = zhp->zpool_hdl; char msg[1024]; - nvlist_t **spares; - uint_t nspares; + nvlist_t **spares, **l2cache; + uint_t nspares, nl2cache; (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot add to '%s'"), zhp->zpool_name); - if (zpool_get_version(zhp) < ZFS_VERSION_SPARES && + if (zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL) < + SPA_VERSION_SPARES && nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, &spares, &nspares) == 0) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool must be " @@ -542,11 +1012,20 @@ zpool_add(zpool_handle_t *zhp, nvlist_t *nvroot) return (zfs_error(hdl, EZFS_BADVERSION, msg)); } - if (zcmd_write_src_nvlist(hdl, &zc, nvroot, NULL) != 0) + if (zpool_get_prop_int(zhp, ZPOOL_PROP_VERSION, NULL) < + SPA_VERSION_L2CACHE && + nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_L2CACHE, + &l2cache, &nl2cache) == 0) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool must be " + "upgraded to add cache devices")); + return (zfs_error(hdl, EZFS_BADVERSION, msg)); + } + + if (zcmd_write_conf_nvlist(hdl, &zc, nvroot) != 0) return (-1); (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); - if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_VDEV_ADD, &zc) != 0) { + if (zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_VDEV_ADD, &zc) != 0) { switch (errno) { case EBUSY: /* @@ -581,16 +1060,23 @@ zpool_add(zpool_handle_t *zhp, nvlist_t *nvroot) case ENOTSUP: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "pool must be upgraded to add raidz2 vdevs")); + "pool must be upgraded to add these vdevs")); (void) zfs_error(hdl, EZFS_BADVERSION, msg); break; case EDOM: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "root pool can not have concatenated devices")); + "root pool can not have multiple vdevs" + " or separate logs")); (void) zfs_error(hdl, EZFS_POOL_NOTSUP, msg); break; + case ENOTBLK: + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "cache device must be a disk or disk slice")); + (void) zfs_error(hdl, EZFS_BADDEV, msg); + break; + default: (void) zpool_standard_error(hdl, errno, msg); } @@ -610,40 +1096,97 @@ zpool_add(zpool_handle_t *zhp, nvlist_t *nvroot) * mounted datasets in the pool. */ int -zpool_export(zpool_handle_t *zhp) +zpool_export(zpool_handle_t *zhp, boolean_t force) { zfs_cmd_t zc = { 0 }; + char msg[1024]; if (zpool_remove_zvol_links(zhp) != 0) return (-1); + (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, + "cannot export '%s'"), zhp->zpool_name); + (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); + zc.zc_cookie = force; + + if (zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_POOL_EXPORT, &zc) != 0) { + switch (errno) { + case EXDEV: + zfs_error_aux(zhp->zpool_hdl, dgettext(TEXT_DOMAIN, + "use '-f' to override the following errors:\n" + "'%s' has an active shared spare which could be" + " used by other pools once '%s' is exported."), + zhp->zpool_name, zhp->zpool_name); + return (zfs_error(zhp->zpool_hdl, EZFS_ACTIVE_SPARE, + msg)); + default: + return (zpool_standard_error_fmt(zhp->zpool_hdl, errno, + msg)); + } + } - if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_POOL_EXPORT, &zc) != 0) - return (zpool_standard_error_fmt(zhp->zpool_hdl, errno, - dgettext(TEXT_DOMAIN, "cannot export '%s'"), - zhp->zpool_name)); return (0); } /* - * Import the given pool using the known configuration. The configuration - * should have come from zpool_find_import(). The 'newname' and 'altroot' - * parameters control whether the pool is imported with a different name or with - * an alternate root, respectively. + * zpool_import() is a contracted interface. Should be kept the same + * if possible. + * + * Applications should use zpool_import_props() to import a pool with + * new properties value to be set. */ int zpool_import(libzfs_handle_t *hdl, nvlist_t *config, const char *newname, - const char *altroot) + char *altroot) +{ + nvlist_t *props = NULL; + int ret; + + if (altroot != NULL) { + if (nvlist_alloc(&props, NV_UNIQUE_NAME, 0) != 0) { + return (zfs_error_fmt(hdl, EZFS_NOMEM, + dgettext(TEXT_DOMAIN, "cannot import '%s'"), + newname)); + } + + if (nvlist_add_string(props, + zpool_prop_to_name(ZPOOL_PROP_ALTROOT), altroot) != 0) { + nvlist_free(props); + return (zfs_error_fmt(hdl, EZFS_NOMEM, + dgettext(TEXT_DOMAIN, "cannot import '%s'"), + newname)); + } + } + + ret = zpool_import_props(hdl, config, newname, props, B_FALSE); + if (props) + nvlist_free(props); + return (ret); +} + +/* + * Import the given pool using the known configuration and a list of + * properties to be set. The configuration should have come from + * zpool_find_import(). The 'newname' parameters control whether the pool + * is imported with a different name. + */ +int +zpool_import_props(libzfs_handle_t *hdl, nvlist_t *config, const char *newname, + nvlist_t *props, boolean_t importfaulted) { zfs_cmd_t zc = { 0 }; char *thename; char *origname; int ret; + char errbuf[1024]; verify(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME, &origname) == 0); + (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, + "cannot import pool '%s'"), origname); + if (newname != NULL) { if (!zpool_name_valid(hdl, B_FALSE, newname)) return (zfs_error_fmt(hdl, EZFS_INVALIDNAME, @@ -654,26 +1197,34 @@ zpool_import(libzfs_handle_t *hdl, nvlist_t *config, const char *newname, thename = origname; } - if (altroot != NULL && altroot[0] != '/') - return (zfs_error_fmt(hdl, EZFS_BADPATH, - dgettext(TEXT_DOMAIN, "bad alternate root '%s'"), - altroot)); + if (props) { + uint64_t version; - (void) strlcpy(zc.zc_name, thename, sizeof (zc.zc_name)); + verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION, + &version) == 0); - if (altroot != NULL) - (void) strlcpy(zc.zc_value, altroot, sizeof (zc.zc_value)); - else - zc.zc_value[0] = '\0'; + if ((props = zpool_valid_proplist(hdl, origname, + props, version, B_TRUE, errbuf)) == NULL) { + return (-1); + } else if (zcmd_write_src_nvlist(hdl, &zc, props) != 0) { + nvlist_free(props); + return (-1); + } + } + + (void) strlcpy(zc.zc_name, thename, sizeof (zc.zc_name)); verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_POOL_GUID, &zc.zc_guid) == 0); - if (zcmd_write_src_nvlist(hdl, &zc, config, NULL) != 0) + if (zcmd_write_conf_nvlist(hdl, &zc, config) != 0) { + nvlist_free(props); return (-1); + } + zc.zc_cookie = (uint64_t)importfaulted; ret = 0; - if (ioctl(hdl->libzfs_fd, ZFS_IOC_POOL_IMPORT, &zc) != 0) { + if (zfs_ioctl(hdl, ZFS_IOC_POOL_IMPORT, &zc) != 0) { char desc[1024]; if (newname == NULL) (void) snprintf(desc, sizeof (desc), @@ -703,6 +1254,7 @@ zpool_import(libzfs_handle_t *hdl, nvlist_t *config, const char *newname, ret = -1; } else { zpool_handle_t *zhp; + /* * This should never fail, but play it safe anyway. */ @@ -712,9 +1264,12 @@ zpool_import(libzfs_handle_t *hdl, nvlist_t *config, const char *newname, ret = zpool_create_zvol_links(zhp); zpool_close(zhp); } + } zcmd_free_nvlists(&zc); + nvlist_free(props); + return (ret); } @@ -731,7 +1286,7 @@ zpool_scrub(zpool_handle_t *zhp, pool_scrub_type_t type) (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); zc.zc_cookie = type; - if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_POOL_SCRUB, &zc) == 0) + if (zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_POOL_SCRUB, &zc) == 0) return (0); (void) snprintf(msg, sizeof (msg), @@ -749,7 +1304,7 @@ zpool_scrub(zpool_handle_t *zhp, pool_scrub_type_t type) */ static nvlist_t * vdev_to_nvlist_iter(nvlist_t *nv, const char *search, uint64_t guid, - boolean_t *avail_spare) + boolean_t *avail_spare, boolean_t *l2cache, boolean_t *log) { uint_t c, children; nvlist_t **child; @@ -757,6 +1312,7 @@ vdev_to_nvlist_iter(nvlist_t *nv, const char *search, uint64_t guid, char *path; uint64_t wholedisk = 0; nvlist_t *ret; + uint64_t is_log; verify(nvlist_lookup_uint64(nv, ZPOOL_CONFIG_GUID, &theguid) == 0); @@ -789,27 +1345,53 @@ vdev_to_nvlist_iter(nvlist_t *nv, const char *search, uint64_t guid, &child, &children) != 0) return (NULL); - for (c = 0; c < children; c++) + for (c = 0; c < children; c++) { if ((ret = vdev_to_nvlist_iter(child[c], search, guid, - avail_spare)) != NULL) + avail_spare, l2cache, NULL)) != NULL) { + /* + * The 'is_log' value is only set for the toplevel + * vdev, not the leaf vdevs. So we always lookup the + * log device from the root of the vdev tree (where + * 'log' is non-NULL). + */ + if (log != NULL && + nvlist_lookup_uint64(child[c], + ZPOOL_CONFIG_IS_LOG, &is_log) == 0 && + is_log) { + *log = B_TRUE; + } return (ret); + } + } if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_SPARES, &child, &children) == 0) { for (c = 0; c < children; c++) { if ((ret = vdev_to_nvlist_iter(child[c], search, guid, - avail_spare)) != NULL) { + avail_spare, l2cache, NULL)) != NULL) { *avail_spare = B_TRUE; return (ret); } } } + if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_L2CACHE, + &child, &children) == 0) { + for (c = 0; c < children; c++) { + if ((ret = vdev_to_nvlist_iter(child[c], search, guid, + avail_spare, l2cache, NULL)) != NULL) { + *l2cache = B_TRUE; + return (ret); + } + } + } + return (NULL); } nvlist_t * -zpool_find_vdev(zpool_handle_t *zhp, const char *path, boolean_t *avail_spare) +zpool_find_vdev(zpool_handle_t *zhp, const char *path, boolean_t *avail_spare, + boolean_t *l2cache, boolean_t *log) { char buf[MAXPATHLEN]; const char *search; @@ -831,29 +1413,124 @@ zpool_find_vdev(zpool_handle_t *zhp, const char *path, boolean_t *avail_spare) &nvroot) == 0); *avail_spare = B_FALSE; - return (vdev_to_nvlist_iter(nvroot, search, guid, avail_spare)); + *l2cache = B_FALSE; + if (log != NULL) + *log = B_FALSE; + return (vdev_to_nvlist_iter(nvroot, search, guid, avail_spare, + l2cache, log)); +} + +static int +vdev_online(nvlist_t *nv) +{ + uint64_t ival; + + if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_OFFLINE, &ival) == 0 || + nvlist_lookup_uint64(nv, ZPOOL_CONFIG_FAULTED, &ival) == 0 || + nvlist_lookup_uint64(nv, ZPOOL_CONFIG_REMOVED, &ival) == 0) + return (0); + + return (1); +} + +/* + * Get phys_path for a root pool + * Return 0 on success; non-zeron on failure. + */ +int +zpool_get_physpath(zpool_handle_t *zhp, char *physpath) +{ + char bootfs[ZPOOL_MAXNAMELEN]; + nvlist_t *vdev_root; + nvlist_t **child; + uint_t count; + int i; + + /* + * Make sure this is a root pool, as phys_path doesn't mean + * anything to a non-root pool. + */ + if (zpool_get_prop(zhp, ZPOOL_PROP_BOOTFS, bootfs, + sizeof (bootfs), NULL) != 0) + return (-1); + + verify(nvlist_lookup_nvlist(zhp->zpool_config, + ZPOOL_CONFIG_VDEV_TREE, &vdev_root) == 0); + + if (nvlist_lookup_nvlist_array(vdev_root, ZPOOL_CONFIG_CHILDREN, + &child, &count) != 0) + return (-2); + + for (i = 0; i < count; i++) { + nvlist_t **child2; + uint_t count2; + char *type; + char *tmppath; + int j; + + if (nvlist_lookup_string(child[i], ZPOOL_CONFIG_TYPE, &type) + != 0) + return (-3); + + if (strcmp(type, VDEV_TYPE_DISK) == 0) { + if (!vdev_online(child[i])) + return (-8); + verify(nvlist_lookup_string(child[i], + ZPOOL_CONFIG_PHYS_PATH, &tmppath) == 0); + (void) strncpy(physpath, tmppath, strlen(tmppath)); + } else if (strcmp(type, VDEV_TYPE_MIRROR) == 0) { + if (nvlist_lookup_nvlist_array(child[i], + ZPOOL_CONFIG_CHILDREN, &child2, &count2) != 0) + return (-4); + + for (j = 0; j < count2; j++) { + if (!vdev_online(child2[j])) + return (-8); + if (nvlist_lookup_string(child2[j], + ZPOOL_CONFIG_PHYS_PATH, &tmppath) != 0) + return (-5); + + if ((strlen(physpath) + strlen(tmppath)) > + MAXNAMELEN) + return (-6); + + if (strlen(physpath) == 0) { + (void) strncpy(physpath, tmppath, + strlen(tmppath)); + } else { + (void) strcat(physpath, " "); + (void) strcat(physpath, tmppath); + } + } + } else { + return (-7); + } + } + + return (0); } /* - * Returns TRUE if the given guid corresponds to a spare (INUSE or not). + * Returns TRUE if the given guid corresponds to the given type. + * This is used to check for hot spares (INUSE or not), and level 2 cache + * devices. */ static boolean_t -is_spare(zpool_handle_t *zhp, uint64_t guid) +is_guid_type(zpool_handle_t *zhp, uint64_t guid, const char *type) { - uint64_t spare_guid; + uint64_t target_guid; nvlist_t *nvroot; - nvlist_t **spares; - uint_t nspares; + nvlist_t **list; + uint_t count; int i; verify(nvlist_lookup_nvlist(zhp->zpool_config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); - if (nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_SPARES, - &spares, &nspares) == 0) { - for (i = 0; i < nspares; i++) { - verify(nvlist_lookup_uint64(spares[i], - ZPOOL_CONFIG_GUID, &spare_guid) == 0); - if (guid == spare_guid) + if (nvlist_lookup_nvlist_array(nvroot, type, &list, &count) == 0) { + for (i = 0; i < count; i++) { + verify(nvlist_lookup_uint64(list[i], ZPOOL_CONFIG_GUID, + &target_guid) == 0); + if (guid == target_guid) return (B_TRUE); } } @@ -862,62 +1539,106 @@ is_spare(zpool_handle_t *zhp, uint64_t guid) } /* - * Bring the specified vdev online + * Bring the specified vdev online. The 'flags' parameter is a set of the + * ZFS_ONLINE_* flags. */ int -zpool_vdev_online(zpool_handle_t *zhp, const char *path) +zpool_vdev_online(zpool_handle_t *zhp, const char *path, int flags, + vdev_state_t *newstate) { zfs_cmd_t zc = { 0 }; char msg[1024]; nvlist_t *tgt; - boolean_t avail_spare; + boolean_t avail_spare, l2cache; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot online %s"), path); (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); - if ((tgt = zpool_find_vdev(zhp, path, &avail_spare)) == NULL) + if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache, + NULL)) == NULL) return (zfs_error(hdl, EZFS_NODEVICE, msg)); verify(nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID, &zc.zc_guid) == 0); - if (avail_spare || is_spare(zhp, zc.zc_guid) == B_TRUE) + if (avail_spare || + is_guid_type(zhp, zc.zc_guid, ZPOOL_CONFIG_SPARES) == B_TRUE) return (zfs_error(hdl, EZFS_ISSPARE, msg)); - if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_VDEV_ONLINE, &zc) == 0) - return (0); + zc.zc_cookie = VDEV_STATE_ONLINE; + zc.zc_obj = flags; - return (zpool_standard_error(hdl, errno, msg)); + if (zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_VDEV_SET_STATE, &zc) != 0) + return (zpool_standard_error(hdl, errno, msg)); + + *newstate = zc.zc_cookie; + return (0); } /* * Take the specified vdev offline */ int -zpool_vdev_offline(zpool_handle_t *zhp, const char *path, int istmp) +zpool_vdev_offline(zpool_handle_t *zhp, const char *path, boolean_t istmp) { zfs_cmd_t zc = { 0 }; char msg[1024]; nvlist_t *tgt; - boolean_t avail_spare; + boolean_t avail_spare, l2cache; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot offline %s"), path); (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); - if ((tgt = zpool_find_vdev(zhp, path, &avail_spare)) == NULL) + if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache, + NULL)) == NULL) return (zfs_error(hdl, EZFS_NODEVICE, msg)); verify(nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID, &zc.zc_guid) == 0); - if (avail_spare || is_spare(zhp, zc.zc_guid) == B_TRUE) + if (avail_spare || + is_guid_type(zhp, zc.zc_guid, ZPOOL_CONFIG_SPARES) == B_TRUE) return (zfs_error(hdl, EZFS_ISSPARE, msg)); - zc.zc_cookie = istmp; + zc.zc_cookie = VDEV_STATE_OFFLINE; + zc.zc_obj = istmp ? ZFS_OFFLINE_TEMPORARY : 0; + + if (zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_VDEV_SET_STATE, &zc) == 0) + return (0); + + switch (errno) { + case EBUSY: + + /* + * There are no other replicas of this device. + */ + return (zfs_error(hdl, EZFS_NOREPLICAS, msg)); + + default: + return (zpool_standard_error(hdl, errno, msg)); + } +} + +/* + * Mark the given vdev faulted. + */ +int +zpool_vdev_fault(zpool_handle_t *zhp, uint64_t guid) +{ + zfs_cmd_t zc = { 0 }; + char msg[1024]; + libzfs_handle_t *hdl = zhp->zpool_hdl; + + (void) snprintf(msg, sizeof (msg), + dgettext(TEXT_DOMAIN, "cannot fault %llu"), guid); + + (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); + zc.zc_guid = guid; + zc.zc_cookie = VDEV_STATE_FAULTED; - if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_VDEV_OFFLINE, &zc) == 0) + if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_VDEV_SET_STATE, &zc) == 0) return (0); switch (errno) { @@ -931,6 +1652,30 @@ zpool_vdev_offline(zpool_handle_t *zhp, const char *path, int istmp) default: return (zpool_standard_error(hdl, errno, msg)); } + +} + +/* + * Mark the given vdev degraded. + */ +int +zpool_vdev_degrade(zpool_handle_t *zhp, uint64_t guid) +{ + zfs_cmd_t zc = { 0 }; + char msg[1024]; + libzfs_handle_t *hdl = zhp->zpool_hdl; + + (void) snprintf(msg, sizeof (msg), + dgettext(TEXT_DOMAIN, "cannot degrade %llu"), guid); + + (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); + zc.zc_guid = guid; + zc.zc_cookie = VDEV_STATE_DEGRADED; + + if (ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_VDEV_SET_STATE, &zc) == 0) + return (0); + + return (zpool_standard_error(hdl, errno, msg)); } /* @@ -963,7 +1708,7 @@ is_replacing_spare(nvlist_t *search, nvlist_t *tgt, int which) /* * Attach new_disk (fully described by nvroot) to old_disk. - * If 'replacing' is specified, tne new disk will replace the old one. + * If 'replacing' is specified, the new disk will replace the old one. */ int zpool_vdev_attach(zpool_handle_t *zhp, @@ -973,9 +1718,9 @@ zpool_vdev_attach(zpool_handle_t *zhp, char msg[1024]; int ret; nvlist_t *tgt; - boolean_t avail_spare; + boolean_t avail_spare, l2cache, islog; uint64_t val; - char *path; + char *path, *newname; nvlist_t **child; uint_t children; nvlist_t *config_root; @@ -989,12 +1734,16 @@ zpool_vdev_attach(zpool_handle_t *zhp, "cannot attach %s to %s"), new_disk, old_disk); (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); - if ((tgt = zpool_find_vdev(zhp, old_disk, &avail_spare)) == 0) + if ((tgt = zpool_find_vdev(zhp, old_disk, &avail_spare, &l2cache, + &islog)) == 0) return (zfs_error(hdl, EZFS_NODEVICE, msg)); if (avail_spare) return (zfs_error(hdl, EZFS_ISSPARE, msg)); + if (l2cache) + return (zfs_error(hdl, EZFS_ISL2CACHE, msg)); + verify(nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID, &zc.zc_guid) == 0); zc.zc_cookie = replacing; @@ -1008,17 +1757,21 @@ zpool_vdev_attach(zpool_handle_t *zhp, verify(nvlist_lookup_nvlist(zpool_get_config(zhp, NULL), ZPOOL_CONFIG_VDEV_TREE, &config_root) == 0); + if ((newname = zpool_vdev_name(NULL, NULL, child[0])) == NULL) + return (-1); + /* * If the target is a hot spare that has been swapped in, we can only * replace it with another hot spare. */ if (replacing && nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_IS_SPARE, &val) == 0 && - nvlist_lookup_string(child[0], ZPOOL_CONFIG_PATH, &path) == 0 && - (zpool_find_vdev(zhp, path, &avail_spare) == NULL || - !avail_spare) && is_replacing_spare(config_root, tgt, 1)) { + (zpool_find_vdev(zhp, newname, &avail_spare, &l2cache, + NULL) == NULL || !avail_spare) && + is_replacing_spare(config_root, tgt, 1)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "can only be replaced by another hot spare")); + free(newname); return (zfs_error(hdl, EZFS_BADTARGET, msg)); } @@ -1028,17 +1781,21 @@ zpool_vdev_attach(zpool_handle_t *zhp, */ if (replacing && nvlist_lookup_string(child[0], ZPOOL_CONFIG_PATH, &path) == 0 && - zpool_find_vdev(zhp, path, &avail_spare) != NULL && avail_spare && + zpool_find_vdev(zhp, newname, &avail_spare, + &l2cache, NULL) != NULL && avail_spare && is_replacing_spare(config_root, tgt, 0)) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "device has already been replaced with a spare")); + free(newname); return (zfs_error(hdl, EZFS_BADTARGET, msg)); } - if (zcmd_write_src_nvlist(hdl, &zc, nvroot, NULL) != 0) + free(newname); + + if (zcmd_write_conf_nvlist(hdl, &zc, nvroot) != 0) return (-1); - ret = ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_VDEV_ATTACH, &zc); + ret = zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_VDEV_ATTACH, &zc); zcmd_free_nvlists(&zc); @@ -1050,13 +1807,18 @@ zpool_vdev_attach(zpool_handle_t *zhp, /* * Can't attach to or replace this type of vdev. */ - if (replacing) - zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "cannot replace a replacing device")); - else + if (replacing) { + if (islog) + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "cannot replace a log with a spare")); + else + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "cannot replace a replacing device")); + } else { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "can only attach to mirrors and top-level " "disks")); + } (void) zfs_error(hdl, EZFS_BADTARGET, msg); break; @@ -1116,22 +1878,26 @@ zpool_vdev_detach(zpool_handle_t *zhp, const char *path) zfs_cmd_t zc = { 0 }; char msg[1024]; nvlist_t *tgt; - boolean_t avail_spare; + boolean_t avail_spare, l2cache; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot detach %s"), path); (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); - if ((tgt = zpool_find_vdev(zhp, path, &avail_spare)) == 0) + if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache, + NULL)) == 0) return (zfs_error(hdl, EZFS_NODEVICE, msg)); if (avail_spare) return (zfs_error(hdl, EZFS_ISSPARE, msg)); + if (l2cache) + return (zfs_error(hdl, EZFS_ISL2CACHE, msg)); + verify(nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID, &zc.zc_guid) == 0); - if (ioctl(hdl->libzfs_fd, ZFS_IOC_VDEV_DETACH, &zc) == 0) + if (zfs_ioctl(hdl, ZFS_IOC_VDEV_DETACH, &zc) == 0) return (0); switch (errno) { @@ -1160,7 +1926,8 @@ zpool_vdev_detach(zpool_handle_t *zhp, const char *path) } /* - * Remove the given device. Currently, this is supported only for hot spares. + * Remove the given device. Currently, this is supported only for hot spares + * and level 2 cache devices. */ int zpool_vdev_remove(zpool_handle_t *zhp, const char *path) @@ -1168,25 +1935,27 @@ zpool_vdev_remove(zpool_handle_t *zhp, const char *path) zfs_cmd_t zc = { 0 }; char msg[1024]; nvlist_t *tgt; - boolean_t avail_spare; + boolean_t avail_spare, l2cache; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, "cannot remove %s"), path); (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); - if ((tgt = zpool_find_vdev(zhp, path, &avail_spare)) == 0) + if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, &l2cache, + NULL)) == 0) return (zfs_error(hdl, EZFS_NODEVICE, msg)); - if (!avail_spare) { + if (!avail_spare && !l2cache) { zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, - "only inactive hot spares can be removed")); + "only inactive hot spares or cache devices " + "can be removed")); return (zfs_error(hdl, EZFS_NODEVICE, msg)); } verify(nvlist_lookup_uint64(tgt, ZPOOL_CONFIG_GUID, &zc.zc_guid) == 0); - if (ioctl(hdl->libzfs_fd, ZFS_IOC_VDEV_REMOVE, &zc) == 0) + if (zfs_ioctl(hdl, ZFS_IOC_VDEV_REMOVE, &zc) == 0) return (0); return (zpool_standard_error(hdl, errno, msg)); @@ -1201,7 +1970,7 @@ zpool_clear(zpool_handle_t *zhp, const char *path) zfs_cmd_t zc = { 0 }; char msg[1024]; nvlist_t *tgt; - boolean_t avail_spare; + boolean_t avail_spare, l2cache; libzfs_handle_t *hdl = zhp->zpool_hdl; if (path) @@ -1215,9 +1984,14 @@ zpool_clear(zpool_handle_t *zhp, const char *path) (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); if (path) { - if ((tgt = zpool_find_vdev(zhp, path, &avail_spare)) == 0) + if ((tgt = zpool_find_vdev(zhp, path, &avail_spare, + &l2cache, NULL)) == 0) return (zfs_error(hdl, EZFS_NODEVICE, msg)); + /* + * Don't allow error clearing for hot spares. Do allow + * error clearing for l2cache devices. + */ if (avail_spare) return (zfs_error(hdl, EZFS_ISSPARE, msg)); @@ -1225,6 +1999,29 @@ zpool_clear(zpool_handle_t *zhp, const char *path) &zc.zc_guid) == 0); } + if (zfs_ioctl(hdl, ZFS_IOC_CLEAR, &zc) == 0) + return (0); + + return (zpool_standard_error(hdl, errno, msg)); +} + +/* + * Similar to zpool_clear(), but takes a GUID (used by fmd). + */ +int +zpool_vdev_clear(zpool_handle_t *zhp, uint64_t guid) +{ + zfs_cmd_t zc = { 0 }; + char msg[1024]; + libzfs_handle_t *hdl = zhp->zpool_hdl; + + (void) snprintf(msg, sizeof (msg), + dgettext(TEXT_DOMAIN, "cannot clear errors for %llx"), + guid); + + (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); + zc.zc_guid = guid; + if (ioctl(hdl->libzfs_fd, ZFS_IOC_CLEAR, &zc) == 0) return (0); @@ -1342,12 +2139,15 @@ typedef struct zvol_cb { static int do_zvol_create(zfs_handle_t *zhp, void *data) { - int ret; + int ret = 0; - if (ZFS_IS_VOLUME(zhp)) + if (ZFS_IS_VOLUME(zhp)) { (void) zvol_create_link(zhp->zfs_hdl, zhp->zfs_name); + ret = zfs_iter_snapshots(zhp, do_zvol_create, NULL); + } - ret = zfs_iter_children(zhp, do_zvol_create, NULL); + if (ret == 0) + ret = zfs_iter_filesystems(zhp, do_zvol_create, NULL); zfs_close(zhp); @@ -1370,7 +2170,7 @@ zpool_create_zvol_links(zpool_handle_t *zhp) zhp->zpool_name)) == NULL) return (0); - ret = zfs_iter_children(zfp, do_zvol_create, NULL); + ret = zfs_iter_filesystems(zfp, do_zvol_create, NULL); zfs_close(zfp); return (ret); @@ -1492,6 +2292,8 @@ zpool_vdev_name(libzfs_handle_t *hdl, zpool_handle_t *zhp, nvlist_t *nv) char *path, *devid; uint64_t value; char buf[64]; + vdev_stat_t *vs; + uint_t vsc; if (nvlist_lookup_uint64(nv, ZPOOL_CONFIG_NOT_PRESENT, &value) == 0) { @@ -1502,7 +2304,16 @@ zpool_vdev_name(libzfs_handle_t *hdl, zpool_handle_t *zhp, nvlist_t *nv) path = buf; } else if (nvlist_lookup_string(nv, ZPOOL_CONFIG_PATH, &path) == 0) { - if (zhp != NULL && + /* + * If the device is dead (faulted, offline, etc) then don't + * bother opening it. Otherwise we may be forcing the user to + * open a misbehaving device, which can have undesirable + * effects. + */ + if ((nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_STATS, + (uint64_t **)&vs, &vsc) != 0 || + vs->vs_state >= VDEV_STATE_DEGRADED) && + zhp != NULL && nvlist_lookup_string(nv, ZPOOL_CONFIG_DEVID, &devid) == 0) { /* * Determine if the current path is correct. @@ -1585,6 +2396,8 @@ zpool_get_errlog(zpool_handle_t *zhp, nvlist_t **nverrlistp) */ verify(nvlist_lookup_uint64(zhp->zpool_config, ZPOOL_CONFIG_ERRCOUNT, &count) == 0); + if (count == 0) + return (0); if ((zc.zc_nvlist_dst = (uintptr_t)zfs_alloc(zhp->zpool_hdl, count * sizeof (zbookmark_t))) == (uintptr_t)NULL) return (-1); @@ -1665,62 +2478,56 @@ nomem: * Upgrade a ZFS pool to the latest on-disk version. */ int -zpool_upgrade(zpool_handle_t *zhp) +zpool_upgrade(zpool_handle_t *zhp, uint64_t new_version) { zfs_cmd_t zc = { 0 }; libzfs_handle_t *hdl = zhp->zpool_hdl; (void) strcpy(zc.zc_name, zhp->zpool_name); - if (ioctl(hdl->libzfs_fd, ZFS_IOC_POOL_UPGRADE, &zc) != 0) + zc.zc_cookie = new_version; + + if (zfs_ioctl(hdl, ZFS_IOC_POOL_UPGRADE, &zc) != 0) return (zpool_standard_error_fmt(hdl, errno, dgettext(TEXT_DOMAIN, "cannot upgrade '%s'"), zhp->zpool_name)); - return (0); } -/* - * Log command history. - * - * 'pool' is B_TRUE if we are logging a command for 'zpool'; B_FALSE - * otherwise ('zfs'). 'pool_create' is B_TRUE if we are logging the creation - * of the pool; B_FALSE otherwise. 'path' is the pathanme containing the - * poolname. 'argc' and 'argv' are used to construct the command string. - */ void -zpool_log_history(libzfs_handle_t *hdl, int argc, char **argv, const char *path, - boolean_t pool, boolean_t pool_create) +zpool_set_history_str(const char *subcommand, int argc, char **argv, + char *history_str) { - char cmd_buf[HIS_MAX_RECORD_LEN]; - char *dspath; - zfs_cmd_t zc = { 0 }; int i; - /* construct the command string */ - (void) strcpy(cmd_buf, pool ? "zpool" : "zfs"); - for (i = 0; i < argc; i++) { - if (strlen(cmd_buf) + 1 + strlen(argv[i]) > HIS_MAX_RECORD_LEN) + (void) strlcpy(history_str, subcommand, HIS_MAX_RECORD_LEN); + for (i = 1; i < argc; i++) { + if (strlen(history_str) + 1 + strlen(argv[i]) > + HIS_MAX_RECORD_LEN) break; - (void) strcat(cmd_buf, " "); - (void) strcat(cmd_buf, argv[i]); + (void) strlcat(history_str, " ", HIS_MAX_RECORD_LEN); + (void) strlcat(history_str, argv[i], HIS_MAX_RECORD_LEN); } +} - /* figure out the poolname */ - dspath = strpbrk(path, "/@"); - if (dspath == NULL) { - (void) strcpy(zc.zc_name, path); - } else { - (void) strncpy(zc.zc_name, path, dspath - path); - zc.zc_name[dspath-path] = '\0'; - } +/* + * Stage command history for logging. + */ +int +zpool_stage_history(libzfs_handle_t *hdl, const char *history_str) +{ + if (history_str == NULL) + return (EINVAL); + + if (strlen(history_str) > HIS_MAX_RECORD_LEN) + return (EINVAL); - zc.zc_history = (uint64_t)(uintptr_t)cmd_buf; - zc.zc_history_len = strlen(cmd_buf); + if (hdl->libzfs_log_str != NULL) + free(hdl->libzfs_log_str); - /* overloading zc_history_offset */ - zc.zc_history_offset = pool_create; + if ((hdl->libzfs_log_str = strdup(history_str)) == NULL) + return (no_memory(hdl)); - (void) ioctl(hdl->libzfs_fd, ZFS_IOC_POOL_LOG_HISTORY, &zc); + return (0); } /* @@ -1906,150 +2713,296 @@ zpool_obj_to_path(zpool_handle_t *zhp, uint64_t dsobj, uint64_t obj, free(mntpnt); } -int -zpool_set_prop(zpool_handle_t *zhp, const char *propname, const char *propval) +#define RDISK_ROOT "/dev/rdsk" +#define BACKUP_SLICE "s2" +/* + * Don't start the slice at the default block of 34; many storage + * devices will use a stripe width of 128k, so start there instead. + */ +#define NEW_START_BLOCK 256 + +#if defined(sun) +/* + * Read the EFI label from the config, if a label does not exist then + * pass back the error to the caller. If the caller has passed a non-NULL + * diskaddr argument then we set it to the starting address of the EFI + * partition. + */ +static int +read_efi_label(nvlist_t *config, diskaddr_t *sb) { - zfs_cmd_t zc = { 0 }; - int ret = -1; - char errbuf[1024]; - nvlist_t *nvl = NULL; - nvlist_t *realprops; + char *path; + int fd; + char diskname[MAXPATHLEN]; + int err = -1; - (void) snprintf(errbuf, sizeof (errbuf), - dgettext(TEXT_DOMAIN, "cannot set property for '%s'"), - zhp->zpool_name); + if (nvlist_lookup_string(config, ZPOOL_CONFIG_PATH, &path) != 0) + return (err); - if (zpool_get_version(zhp) < ZFS_VERSION_BOOTFS) { - zfs_error_aux(zhp->zpool_hdl, - dgettext(TEXT_DOMAIN, "pool must be " - "upgraded to support pool properties")); - return (zfs_error(zhp->zpool_hdl, EZFS_BADVERSION, errbuf)); - } + (void) snprintf(diskname, sizeof (diskname), "%s%s", RDISK_ROOT, + strrchr(path, '/')); + if ((fd = open(diskname, O_RDONLY|O_NDELAY)) >= 0) { + struct dk_gpt *vtoc; - if (zhp->zpool_props == NULL && zpool_get_all_props(zhp)) - return (zfs_error(zhp->zpool_hdl, EZFS_POOLPROPS, errbuf)); + if ((err = efi_alloc_and_read(fd, &vtoc)) >= 0) { + if (sb != NULL) + *sb = vtoc->efi_parts[0].p_start; + efi_free(vtoc); + } + (void) close(fd); + } + return (err); +} - if (nvlist_alloc(&nvl, NV_UNIQUE_NAME, 0) != 0 || - nvlist_add_string(nvl, propname, propval) != 0) { - return (no_memory(zhp->zpool_hdl)); +/* + * determine where a partition starts on a disk in the current + * configuration + */ +static diskaddr_t +find_start_block(nvlist_t *config) +{ + nvlist_t **child; + uint_t c, children; + diskaddr_t sb = MAXOFFSET_T; + uint64_t wholedisk; + + if (nvlist_lookup_nvlist_array(config, + ZPOOL_CONFIG_CHILDREN, &child, &children) != 0) { + if (nvlist_lookup_uint64(config, + ZPOOL_CONFIG_WHOLE_DISK, + &wholedisk) != 0 || !wholedisk) { + return (MAXOFFSET_T); + } + if (read_efi_label(config, &sb) < 0) + sb = MAXOFFSET_T; + return (sb); } - if ((realprops = zfs_validate_properties(zhp->zpool_hdl, ZFS_TYPE_POOL, - zhp->zpool_name, nvl, 0, NULL, errbuf)) == NULL) { - nvlist_free(nvl); - return (-1); + for (c = 0; c < children; c++) { + sb = find_start_block(child[c]); + if (sb != MAXOFFSET_T) { + return (sb); + } } + return (MAXOFFSET_T); +} +#endif /* sun */ - nvlist_free(nvl); - nvl = realprops; +/* + * Label an individual disk. The name provided is the short name, + * stripped of any leading /dev path. + */ +int +zpool_label_disk(libzfs_handle_t *hdl, zpool_handle_t *zhp, char *name) +{ +#if defined(sun) + char path[MAXPATHLEN]; + struct dk_gpt *vtoc; + int fd; + size_t resv = EFI_MIN_RESV_SIZE; + uint64_t slice_size; + diskaddr_t start_block; + char errbuf[1024]; - /* - * Execute the corresponding ioctl() to set this property. - */ - (void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name)); + /* prepare an error message just in case */ + (void) snprintf(errbuf, sizeof (errbuf), + dgettext(TEXT_DOMAIN, "cannot label '%s'"), name); - if (zcmd_write_src_nvlist(zhp->zpool_hdl, &zc, nvl, NULL) != 0) - return (-1); + if (zhp) { + nvlist_t *nvroot; - ret = ioctl(zhp->zpool_hdl->libzfs_fd, ZFS_IOC_POOL_SET_PROPS, &zc); - zcmd_free_nvlists(&zc); + verify(nvlist_lookup_nvlist(zhp->zpool_config, + ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); - if (ret) - (void) zpool_standard_error(zhp->zpool_hdl, errno, errbuf); + if (zhp->zpool_start_block == 0) + start_block = find_start_block(nvroot); + else + start_block = zhp->zpool_start_block; + zhp->zpool_start_block = start_block; + } else { + /* new pool */ + start_block = NEW_START_BLOCK; + } - return (ret); -} + (void) snprintf(path, sizeof (path), "%s/%s%s", RDISK_ROOT, name, + BACKUP_SLICE); -int -zpool_get_prop(zpool_handle_t *zhp, zpool_prop_t prop, char *propbuf, - size_t proplen, zfs_source_t *srctype) -{ - uint64_t value; - char msg[1024], *strvalue; - nvlist_t *nvp; - zfs_source_t src = ZFS_SRC_NONE; + if ((fd = open(path, O_RDWR | O_NDELAY)) < 0) { + /* + * This shouldn't happen. We've long since verified that this + * is a valid device. + */ + zfs_error_aux(hdl, + dgettext(TEXT_DOMAIN, "unable to open device")); + return (zfs_error(hdl, EZFS_OPENFAILED, errbuf)); + } - (void) snprintf(msg, sizeof (msg), dgettext(TEXT_DOMAIN, - "cannot get property '%s'"), zpool_prop_to_name(prop)); + if (efi_alloc_and_init(fd, EFI_NUMPAR, &vtoc) != 0) { + /* + * The only way this can fail is if we run out of memory, or we + * were unable to read the disk's capacity + */ + if (errno == ENOMEM) + (void) no_memory(hdl); + + (void) close(fd); + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "unable to read disk capacity"), name); - if (zpool_get_version(zhp) < ZFS_VERSION_BOOTFS) { - zfs_error_aux(zhp->zpool_hdl, - dgettext(TEXT_DOMAIN, "pool must be " - "upgraded to support pool properties")); - return (zfs_error(zhp->zpool_hdl, EZFS_BADVERSION, msg)); + return (zfs_error(hdl, EZFS_NOCAP, errbuf)); } - if (zhp->zpool_props == NULL && zpool_get_all_props(zhp)) - return (zfs_error(zhp->zpool_hdl, EZFS_POOLPROPS, msg)); + slice_size = vtoc->efi_last_u_lba + 1; + slice_size -= EFI_MIN_RESV_SIZE; + if (start_block == MAXOFFSET_T) + start_block = NEW_START_BLOCK; + slice_size -= start_block; + + vtoc->efi_parts[0].p_start = start_block; + vtoc->efi_parts[0].p_size = slice_size; /* - * the "name" property is special cased + * Why we use V_USR: V_BACKUP confuses users, and is considered + * disposable by some EFI utilities (since EFI doesn't have a backup + * slice). V_UNASSIGNED is supposed to be used only for zero size + * partitions, and efi_write() will fail if we use it. V_ROOT, V_BOOT, + * etc. were all pretty specific. V_USR is as close to reality as we + * can get, in the absence of V_OTHER. */ - if (!zfs_prop_valid_for_type(prop, ZFS_TYPE_POOL) && - prop != ZFS_PROP_NAME) - return (-1); + vtoc->efi_parts[0].p_tag = V_USR; + (void) strcpy(vtoc->efi_parts[0].p_name, "zfs"); - switch (prop) { - case ZFS_PROP_NAME: - (void) strlcpy(propbuf, zhp->zpool_name, proplen); - break; + vtoc->efi_parts[8].p_start = slice_size + start_block; + vtoc->efi_parts[8].p_size = resv; + vtoc->efi_parts[8].p_tag = V_RESERVED; - case ZFS_PROP_BOOTFS: - if (nvlist_lookup_nvlist(zhp->zpool_props, - zpool_prop_to_name(prop), &nvp) != 0) { - strvalue = (char *)zfs_prop_default_string(prop); - if (strvalue == NULL) - strvalue = "-"; - src = ZFS_SRC_DEFAULT; - } else { - VERIFY(nvlist_lookup_uint64(nvp, - ZFS_PROP_SOURCE, &value) == 0); - src = value; - VERIFY(nvlist_lookup_string(nvp, ZFS_PROP_VALUE, - &strvalue) == 0); - if (strlen(strvalue) >= proplen) - return (-1); - } - (void) strcpy(propbuf, strvalue); - break; + if (efi_write(fd, vtoc) != 0) { + /* + * Some block drivers (like pcata) may not support EFI + * GPT labels. Print out a helpful error message dir- + * ecting the user to manually label the disk and give + * a specific slice. + */ + (void) close(fd); + efi_free(vtoc); - default: - return (-1); + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "try using fdisk(1M) and then provide a specific slice")); + return (zfs_error(hdl, EZFS_LABELFAILED, errbuf)); } - if (srctype) - *srctype = src; + + (void) close(fd); + efi_free(vtoc); +#endif /* sun */ return (0); } -int -zpool_get_proplist(libzfs_handle_t *hdl, char *fields, zpool_proplist_t **listp) +static boolean_t +supported_dump_vdev_type(libzfs_handle_t *hdl, nvlist_t *config, char *errbuf) { - return (zfs_get_proplist_common(hdl, fields, listp, ZFS_TYPE_POOL)); -} + char *type; + nvlist_t **child; + uint_t children, c; + verify(nvlist_lookup_string(config, ZPOOL_CONFIG_TYPE, &type) == 0); + if (strcmp(type, VDEV_TYPE_RAIDZ) == 0 || + strcmp(type, VDEV_TYPE_FILE) == 0 || + strcmp(type, VDEV_TYPE_LOG) == 0 || + strcmp(type, VDEV_TYPE_MISSING) == 0) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "vdev type '%s' is not supported"), type); + (void) zfs_error(hdl, EZFS_VDEVNOTSUP, errbuf); + return (B_FALSE); + } + if (nvlist_lookup_nvlist_array(config, ZPOOL_CONFIG_CHILDREN, + &child, &children) == 0) { + for (c = 0; c < children; c++) { + if (!supported_dump_vdev_type(hdl, child[c], errbuf)) + return (B_FALSE); + } + } + return (B_TRUE); +} +/* + * check if this zvol is allowable for use as a dump device; zero if + * it is, > 0 if it isn't, < 0 if it isn't a zvol + */ int -zpool_expand_proplist(zpool_handle_t *zhp, zpool_proplist_t **plp) +zvol_check_dump_config(char *arg) { - libzfs_handle_t *hdl = zhp->zpool_hdl; - zpool_proplist_t *entry; - char buf[ZFS_MAXPROPLEN]; + zpool_handle_t *zhp = NULL; + nvlist_t *config, *nvroot; + char *p, *volname; + nvlist_t **top; + uint_t toplevels; + libzfs_handle_t *hdl; + char errbuf[1024]; + char poolname[ZPOOL_MAXNAMELEN]; + int pathlen = strlen(ZVOL_FULL_DEV_DIR); + int ret = 1; - if (zfs_expand_proplist_common(hdl, plp, ZFS_TYPE_POOL) != 0) + if (strncmp(arg, ZVOL_FULL_DEV_DIR, pathlen)) { return (-1); + } - for (entry = *plp; entry != NULL; entry = entry->pl_next) { + (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, + "dump is not supported on device '%s'"), arg); - if (entry->pl_fixed) - continue; + if ((hdl = libzfs_init()) == NULL) + return (1); + libzfs_print_on_error(hdl, B_TRUE); - if (entry->pl_prop != ZFS_PROP_INVAL && - zpool_get_prop(zhp, entry->pl_prop, buf, sizeof (buf), - NULL) == 0) { - if (strlen(buf) > entry->pl_width) - entry->pl_width = strlen(buf); - } + volname = arg + pathlen; + + /* check the configuration of the pool */ + if ((p = strchr(volname, '/')) == NULL) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "malformed dataset name")); + (void) zfs_error(hdl, EZFS_INVALIDNAME, errbuf); + return (1); + } else if (p - volname >= ZFS_MAXNAMELEN) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "dataset name is too long")); + (void) zfs_error(hdl, EZFS_NAMETOOLONG, errbuf); + return (1); + } else { + (void) strncpy(poolname, volname, p - volname); + poolname[p - volname] = '\0'; } - return (0); + if ((zhp = zpool_open(hdl, poolname)) == NULL) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "could not open pool '%s'"), poolname); + (void) zfs_error(hdl, EZFS_OPENFAILED, errbuf); + goto out; + } + config = zpool_get_config(zhp, NULL); + if (nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, + &nvroot) != 0) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "could not obtain vdev configuration for '%s'"), poolname); + (void) zfs_error(hdl, EZFS_INVALCONFIG, errbuf); + goto out; + } + + verify(nvlist_lookup_nvlist_array(nvroot, ZPOOL_CONFIG_CHILDREN, + &top, &toplevels) == 0); + if (toplevels != 1) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "'%s' has multiple top level vdevs"), poolname); + (void) zfs_error(hdl, EZFS_DEVOVERFLOW, errbuf); + goto out; + } + + if (!supported_dump_vdev_type(hdl, top[0], errbuf)) { + goto out; + } + ret = 0; + +out: + if (zhp) + zpool_close(zhp); + libzfs_fini(hdl); + return (ret); } diff --git a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_sendrecv.c b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_sendrecv.c new file mode 100644 index 0000000..d1163ce --- /dev/null +++ b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_sendrecv.c @@ -0,0 +1,2103 @@ +/* + * CDDL HEADER START + * + * The contents of this file are subject to the terms of the + * Common Development and Distribution License (the "License"). + * You may not use this file except in compliance with the License. + * + * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE + * or http://www.opensolaris.org/os/licensing. + * See the License for the specific language governing permissions + * and limitations under the License. + * + * When distributing Covered Code, include this CDDL HEADER in each + * file and include the License file at usr/src/OPENSOLARIS.LICENSE. + * If applicable, add the following below this CDDL HEADER, with the + * fields enclosed by brackets "[]" replaced with your own identifying + * information: Portions Copyright [yyyy] [name of copyright owner] + * + * CDDL HEADER END + */ + +/* + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. + * Use is subject to license terms. + */ + +#include <assert.h> +#include <ctype.h> +#include <errno.h> +#include <libintl.h> +#include <stdio.h> +#include <stdlib.h> +#include <strings.h> +#include <unistd.h> +#include <stddef.h> +#include <fcntl.h> +#include <sys/mount.h> +#include <sys/mntent.h> +#include <sys/mnttab.h> +#include <sys/avl.h> +#include <stddef.h> + +#include <libzfs.h> + +#include "zfs_namecheck.h" +#include "zfs_prop.h" +#include "libzfs_impl.h" + +#include <fletcher.c> /* XXX */ + +/* We need to use something for ENODATA. */ +#define ENODATA EIDRM + +static int zfs_receive_impl(libzfs_handle_t *, const char *, recvflags_t, + int, avl_tree_t *, char **); + +/* + * Routines for dealing with the AVL tree of fs-nvlists + */ +typedef struct fsavl_node { + avl_node_t fn_node; + nvlist_t *fn_nvfs; + char *fn_snapname; + uint64_t fn_guid; +} fsavl_node_t; + +static int +fsavl_compare(const void *arg1, const void *arg2) +{ + const fsavl_node_t *fn1 = arg1; + const fsavl_node_t *fn2 = arg2; + + if (fn1->fn_guid > fn2->fn_guid) + return (+1); + else if (fn1->fn_guid < fn2->fn_guid) + return (-1); + else + return (0); +} + +/* + * Given the GUID of a snapshot, find its containing filesystem and + * (optionally) name. + */ +static nvlist_t * +fsavl_find(avl_tree_t *avl, uint64_t snapguid, char **snapname) +{ + fsavl_node_t fn_find; + fsavl_node_t *fn; + + fn_find.fn_guid = snapguid; + + fn = avl_find(avl, &fn_find, NULL); + if (fn) { + if (snapname) + *snapname = fn->fn_snapname; + return (fn->fn_nvfs); + } + return (NULL); +} + +static void +fsavl_destroy(avl_tree_t *avl) +{ + fsavl_node_t *fn; + void *cookie; + + if (avl == NULL) + return; + + cookie = NULL; + while ((fn = avl_destroy_nodes(avl, &cookie)) != NULL) + free(fn); + avl_destroy(avl); + free(avl); +} + +static avl_tree_t * +fsavl_create(nvlist_t *fss) +{ + avl_tree_t *fsavl; + nvpair_t *fselem = NULL; + + if ((fsavl = malloc(sizeof (avl_tree_t))) == NULL) + return (NULL); + + avl_create(fsavl, fsavl_compare, sizeof (fsavl_node_t), + offsetof(fsavl_node_t, fn_node)); + + while ((fselem = nvlist_next_nvpair(fss, fselem)) != NULL) { + nvlist_t *nvfs, *snaps; + nvpair_t *snapelem = NULL; + + VERIFY(0 == nvpair_value_nvlist(fselem, &nvfs)); + VERIFY(0 == nvlist_lookup_nvlist(nvfs, "snaps", &snaps)); + + while ((snapelem = + nvlist_next_nvpair(snaps, snapelem)) != NULL) { + fsavl_node_t *fn; + uint64_t guid; + + VERIFY(0 == nvpair_value_uint64(snapelem, &guid)); + if ((fn = malloc(sizeof (fsavl_node_t))) == NULL) { + fsavl_destroy(fsavl); + return (NULL); + } + fn->fn_nvfs = nvfs; + fn->fn_snapname = nvpair_name(snapelem); + fn->fn_guid = guid; + + /* + * Note: if there are multiple snaps with the + * same GUID, we ignore all but one. + */ + if (avl_find(fsavl, fn, NULL) == NULL) + avl_add(fsavl, fn); + else + free(fn); + } + } + + return (fsavl); +} + +/* + * Routines for dealing with the giant nvlist of fs-nvlists, etc. + */ +typedef struct send_data { + uint64_t parent_fromsnap_guid; + nvlist_t *parent_snaps; + nvlist_t *fss; + nvlist_t *snapprops; + const char *fromsnap; + const char *tosnap; + + /* + * The header nvlist is of the following format: + * { + * "tosnap" -> string + * "fromsnap" -> string (if incremental) + * "fss" -> { + * id -> { + * + * "name" -> string (full name; for debugging) + * "parentfromsnap" -> number (guid of fromsnap in parent) + * + * "props" -> { name -> value (only if set here) } + * "snaps" -> { name (lastname) -> number (guid) } + * "snapprops" -> { name (lastname) -> { name -> value } } + * + * "origin" -> number (guid) (if clone) + * "sent" -> boolean (not on-disk) + * } + * } + * } + * + */ +} send_data_t; + +static void send_iterate_prop(zfs_handle_t *zhp, nvlist_t *nv); + +static int +send_iterate_snap(zfs_handle_t *zhp, void *arg) +{ + send_data_t *sd = arg; + uint64_t guid = zhp->zfs_dmustats.dds_guid; + char *snapname; + nvlist_t *nv; + + snapname = strrchr(zhp->zfs_name, '@')+1; + + VERIFY(0 == nvlist_add_uint64(sd->parent_snaps, snapname, guid)); + /* + * NB: if there is no fromsnap here (it's a newly created fs in + * an incremental replication), we will substitute the tosnap. + */ + if ((sd->fromsnap && strcmp(snapname, sd->fromsnap) == 0) || + (sd->parent_fromsnap_guid == 0 && sd->tosnap && + strcmp(snapname, sd->tosnap) == 0)) { + sd->parent_fromsnap_guid = guid; + } + + VERIFY(0 == nvlist_alloc(&nv, NV_UNIQUE_NAME, 0)); + send_iterate_prop(zhp, nv); + VERIFY(0 == nvlist_add_nvlist(sd->snapprops, snapname, nv)); + nvlist_free(nv); + + zfs_close(zhp); + return (0); +} + +static void +send_iterate_prop(zfs_handle_t *zhp, nvlist_t *nv) +{ + nvpair_t *elem = NULL; + + while ((elem = nvlist_next_nvpair(zhp->zfs_props, elem)) != NULL) { + char *propname = nvpair_name(elem); + zfs_prop_t prop = zfs_name_to_prop(propname); + nvlist_t *propnv; + + if (!zfs_prop_user(propname) && zfs_prop_readonly(prop)) + continue; + + verify(nvpair_value_nvlist(elem, &propnv) == 0); + if (prop == ZFS_PROP_QUOTA || prop == ZFS_PROP_RESERVATION) { + /* these guys are modifyable, but have no source */ + uint64_t value; + verify(nvlist_lookup_uint64(propnv, + ZPROP_VALUE, &value) == 0); + if (zhp->zfs_type == ZFS_TYPE_SNAPSHOT) + continue; + } else { + char *source; + if (nvlist_lookup_string(propnv, + ZPROP_SOURCE, &source) != 0) + continue; + if (strcmp(source, zhp->zfs_name) != 0) + continue; + } + + if (zfs_prop_user(propname) || + zfs_prop_get_type(prop) == PROP_TYPE_STRING) { + char *value; + verify(nvlist_lookup_string(propnv, + ZPROP_VALUE, &value) == 0); + VERIFY(0 == nvlist_add_string(nv, propname, value)); + } else { + uint64_t value; + verify(nvlist_lookup_uint64(propnv, + ZPROP_VALUE, &value) == 0); + VERIFY(0 == nvlist_add_uint64(nv, propname, value)); + } + } +} + +static int +send_iterate_fs(zfs_handle_t *zhp, void *arg) +{ + send_data_t *sd = arg; + nvlist_t *nvfs, *nv; + int rv; + uint64_t parent_fromsnap_guid_save = sd->parent_fromsnap_guid; + uint64_t guid = zhp->zfs_dmustats.dds_guid; + char guidstring[64]; + + VERIFY(0 == nvlist_alloc(&nvfs, NV_UNIQUE_NAME, 0)); + VERIFY(0 == nvlist_add_string(nvfs, "name", zhp->zfs_name)); + VERIFY(0 == nvlist_add_uint64(nvfs, "parentfromsnap", + sd->parent_fromsnap_guid)); + + if (zhp->zfs_dmustats.dds_origin[0]) { + zfs_handle_t *origin = zfs_open(zhp->zfs_hdl, + zhp->zfs_dmustats.dds_origin, ZFS_TYPE_SNAPSHOT); + if (origin == NULL) + return (-1); + VERIFY(0 == nvlist_add_uint64(nvfs, "origin", + origin->zfs_dmustats.dds_guid)); + } + + /* iterate over props */ + VERIFY(0 == nvlist_alloc(&nv, NV_UNIQUE_NAME, 0)); + send_iterate_prop(zhp, nv); + VERIFY(0 == nvlist_add_nvlist(nvfs, "props", nv)); + nvlist_free(nv); + + /* iterate over snaps, and set sd->parent_fromsnap_guid */ + sd->parent_fromsnap_guid = 0; + VERIFY(0 == nvlist_alloc(&sd->parent_snaps, NV_UNIQUE_NAME, 0)); + VERIFY(0 == nvlist_alloc(&sd->snapprops, NV_UNIQUE_NAME, 0)); + (void) zfs_iter_snapshots(zhp, send_iterate_snap, sd); + VERIFY(0 == nvlist_add_nvlist(nvfs, "snaps", sd->parent_snaps)); + VERIFY(0 == nvlist_add_nvlist(nvfs, "snapprops", sd->snapprops)); + nvlist_free(sd->parent_snaps); + nvlist_free(sd->snapprops); + + /* add this fs to nvlist */ + (void) snprintf(guidstring, sizeof (guidstring), + "0x%llx", (longlong_t)guid); + VERIFY(0 == nvlist_add_nvlist(sd->fss, guidstring, nvfs)); + nvlist_free(nvfs); + + /* iterate over children */ + rv = zfs_iter_filesystems(zhp, send_iterate_fs, sd); + + sd->parent_fromsnap_guid = parent_fromsnap_guid_save; + + zfs_close(zhp); + return (rv); +} + +static int +gather_nvlist(libzfs_handle_t *hdl, const char *fsname, const char *fromsnap, + const char *tosnap, nvlist_t **nvlp, avl_tree_t **avlp) +{ + zfs_handle_t *zhp; + send_data_t sd = { 0 }; + int error; + + zhp = zfs_open(hdl, fsname, ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME); + if (zhp == NULL) + return (EZFS_BADTYPE); + + VERIFY(0 == nvlist_alloc(&sd.fss, NV_UNIQUE_NAME, 0)); + sd.fromsnap = fromsnap; + sd.tosnap = tosnap; + + if ((error = send_iterate_fs(zhp, &sd)) != 0) { + nvlist_free(sd.fss); + if (avlp != NULL) + *avlp = NULL; + *nvlp = NULL; + return (error); + } + + if (avlp != NULL && (*avlp = fsavl_create(sd.fss)) == NULL) { + nvlist_free(sd.fss); + *nvlp = NULL; + return (EZFS_NOMEM); + } + + *nvlp = sd.fss; + return (0); +} + +/* + * Routines for dealing with the sorted snapshot functionality + */ +typedef struct zfs_node { + zfs_handle_t *zn_handle; + avl_node_t zn_avlnode; +} zfs_node_t; + +static int +zfs_sort_snaps(zfs_handle_t *zhp, void *data) +{ + avl_tree_t *avl = data; + zfs_node_t *node = zfs_alloc(zhp->zfs_hdl, sizeof (zfs_node_t)); + + node->zn_handle = zhp; + avl_add(avl, node); + return (0); +} + +/* ARGSUSED */ +static int +zfs_snapshot_compare(const void *larg, const void *rarg) +{ + zfs_handle_t *l = ((zfs_node_t *)larg)->zn_handle; + zfs_handle_t *r = ((zfs_node_t *)rarg)->zn_handle; + uint64_t lcreate, rcreate; + + /* + * Sort them according to creation time. We use the hidden + * CREATETXG property to get an absolute ordering of snapshots. + */ + lcreate = zfs_prop_get_int(l, ZFS_PROP_CREATETXG); + rcreate = zfs_prop_get_int(r, ZFS_PROP_CREATETXG); + + if (lcreate < rcreate) + return (-1); + else if (lcreate > rcreate) + return (+1); + else + return (0); +} + +static int +zfs_iter_snapshots_sorted(zfs_handle_t *zhp, zfs_iter_f callback, void *data) +{ + int ret = 0; + zfs_node_t *node; + avl_tree_t avl; + void *cookie = NULL; + + avl_create(&avl, zfs_snapshot_compare, + sizeof (zfs_node_t), offsetof(zfs_node_t, zn_avlnode)); + + ret = zfs_iter_snapshots(zhp, zfs_sort_snaps, &avl); + + for (node = avl_first(&avl); node != NULL; node = AVL_NEXT(&avl, node)) + ret |= callback(node->zn_handle, data); + + while ((node = avl_destroy_nodes(&avl, &cookie)) != NULL) + free(node); + + avl_destroy(&avl); + + return (ret); +} + +/* + * Routines specific to "zfs send" + */ +typedef struct send_dump_data { + /* these are all just the short snapname (the part after the @) */ + const char *fromsnap; + const char *tosnap; + char lastsnap[ZFS_MAXNAMELEN]; + boolean_t seenfrom, seento, replicate, doall, fromorigin; + boolean_t verbose; + int outfd; + boolean_t err; + nvlist_t *fss; + avl_tree_t *fsavl; +} send_dump_data_t; + +/* + * Dumps a backup of the given snapshot (incremental from fromsnap if it's not + * NULL) to the file descriptor specified by outfd. + */ +static int +dump_ioctl(zfs_handle_t *zhp, const char *fromsnap, boolean_t fromorigin, + int outfd) +{ + zfs_cmd_t zc = { 0 }; + libzfs_handle_t *hdl = zhp->zfs_hdl; + + assert(zhp->zfs_type == ZFS_TYPE_SNAPSHOT); + assert(fromsnap == NULL || fromsnap[0] == '\0' || !fromorigin); + + (void) strlcpy(zc.zc_name, zhp->zfs_name, sizeof (zc.zc_name)); + if (fromsnap) + (void) strlcpy(zc.zc_value, fromsnap, sizeof (zc.zc_value)); + zc.zc_cookie = outfd; + zc.zc_obj = fromorigin; + + if (ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_SEND, &zc) != 0) { + char errbuf[1024]; + (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, + "warning: cannot send '%s'"), zhp->zfs_name); + + switch (errno) { + + case EXDEV: + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "not an earlier snapshot from the same fs")); + return (zfs_error(hdl, EZFS_CROSSTARGET, errbuf)); + + case ENOENT: + if (zfs_dataset_exists(hdl, zc.zc_name, + ZFS_TYPE_SNAPSHOT)) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "incremental source (@%s) does not exist"), + zc.zc_value); + } + return (zfs_error(hdl, EZFS_NOENT, errbuf)); + + case EDQUOT: + case EFBIG: + case EIO: + case ENOLINK: + case ENOSPC: + case ENXIO: + case EPIPE: + case ERANGE: + case EFAULT: + case EROFS: + zfs_error_aux(hdl, strerror(errno)); + return (zfs_error(hdl, EZFS_BADBACKUP, errbuf)); + + default: + return (zfs_standard_error(hdl, errno, errbuf)); + } + } + + return (0); +} + +static int +dump_snapshot(zfs_handle_t *zhp, void *arg) +{ + send_dump_data_t *sdd = arg; + const char *thissnap; + int err; + + thissnap = strchr(zhp->zfs_name, '@') + 1; + + if (sdd->fromsnap && !sdd->seenfrom && + strcmp(sdd->fromsnap, thissnap) == 0) { + sdd->seenfrom = B_TRUE; + (void) strcpy(sdd->lastsnap, thissnap); + zfs_close(zhp); + return (0); + } + + if (sdd->seento || !sdd->seenfrom) { + zfs_close(zhp); + return (0); + } + + /* send it */ + if (sdd->verbose) { + (void) fprintf(stderr, "sending from @%s to %s\n", + sdd->lastsnap, zhp->zfs_name); + } + + err = dump_ioctl(zhp, sdd->lastsnap, + sdd->lastsnap[0] == '\0' && (sdd->fromorigin || sdd->replicate), + sdd->outfd); + + if (!sdd->seento && strcmp(sdd->tosnap, thissnap) == 0) + sdd->seento = B_TRUE; + + (void) strcpy(sdd->lastsnap, thissnap); + zfs_close(zhp); + return (err); +} + +static int +dump_filesystem(zfs_handle_t *zhp, void *arg) +{ + int rv = 0; + send_dump_data_t *sdd = arg; + boolean_t missingfrom = B_FALSE; + zfs_cmd_t zc = { 0 }; + + (void) snprintf(zc.zc_name, sizeof (zc.zc_name), "%s@%s", + zhp->zfs_name, sdd->tosnap); + if (ioctl(zhp->zfs_hdl->libzfs_fd, ZFS_IOC_OBJSET_STATS, &zc) != 0) { + (void) fprintf(stderr, "WARNING: " + "could not send %s@%s: does not exist\n", + zhp->zfs_name, sdd->tosnap); + sdd->err = B_TRUE; + return (0); + } + + if (sdd->replicate && sdd->fromsnap) { + /* + * If this fs does not have fromsnap, and we're doing + * recursive, we need to send a full stream from the + * beginning (or an incremental from the origin if this + * is a clone). If we're doing non-recursive, then let + * them get the error. + */ + (void) snprintf(zc.zc_name, sizeof (zc.zc_name), "%s@%s", + zhp->zfs_name, sdd->fromsnap); + if (ioctl(zhp->zfs_hdl->libzfs_fd, + ZFS_IOC_OBJSET_STATS, &zc) != 0) { + missingfrom = B_TRUE; + } + } + + if (sdd->doall) { + sdd->seenfrom = sdd->seento = sdd->lastsnap[0] = 0; + if (sdd->fromsnap == NULL || missingfrom) + sdd->seenfrom = B_TRUE; + + rv = zfs_iter_snapshots_sorted(zhp, dump_snapshot, arg); + if (!sdd->seenfrom) { + (void) fprintf(stderr, + "WARNING: could not send %s@%s:\n" + "incremental source (%s@%s) does not exist\n", + zhp->zfs_name, sdd->tosnap, + zhp->zfs_name, sdd->fromsnap); + sdd->err = B_TRUE; + } else if (!sdd->seento) { + (void) fprintf(stderr, + "WARNING: could not send %s@%s:\n" + "incremental source (%s@%s) " + "is not earlier than it\n", + zhp->zfs_name, sdd->tosnap, + zhp->zfs_name, sdd->fromsnap); + sdd->err = B_TRUE; + } + } else { + zfs_handle_t *snapzhp; + char snapname[ZFS_MAXNAMELEN]; + + (void) snprintf(snapname, sizeof (snapname), "%s@%s", + zfs_get_name(zhp), sdd->tosnap); + snapzhp = zfs_open(zhp->zfs_hdl, snapname, ZFS_TYPE_SNAPSHOT); + if (snapzhp == NULL) { + rv = -1; + } else { + rv = dump_ioctl(snapzhp, + missingfrom ? NULL : sdd->fromsnap, + sdd->fromorigin || missingfrom, + sdd->outfd); + sdd->seento = B_TRUE; + zfs_close(snapzhp); + } + } + + return (rv); +} + +static int +dump_filesystems(zfs_handle_t *rzhp, void *arg) +{ + send_dump_data_t *sdd = arg; + nvpair_t *fspair; + boolean_t needagain, progress; + + if (!sdd->replicate) + return (dump_filesystem(rzhp, sdd)); + +again: + needagain = progress = B_FALSE; + for (fspair = nvlist_next_nvpair(sdd->fss, NULL); fspair; + fspair = nvlist_next_nvpair(sdd->fss, fspair)) { + nvlist_t *fslist; + char *fsname; + zfs_handle_t *zhp; + int err; + uint64_t origin_guid = 0; + nvlist_t *origin_nv; + + VERIFY(nvpair_value_nvlist(fspair, &fslist) == 0); + if (nvlist_lookup_boolean(fslist, "sent") == 0) + continue; + + VERIFY(nvlist_lookup_string(fslist, "name", &fsname) == 0); + (void) nvlist_lookup_uint64(fslist, "origin", &origin_guid); + + origin_nv = fsavl_find(sdd->fsavl, origin_guid, NULL); + if (origin_nv && + nvlist_lookup_boolean(origin_nv, "sent") == ENOENT) { + /* + * origin has not been sent yet; + * skip this clone. + */ + needagain = B_TRUE; + continue; + } + + zhp = zfs_open(rzhp->zfs_hdl, fsname, ZFS_TYPE_DATASET); + if (zhp == NULL) + return (-1); + err = dump_filesystem(zhp, sdd); + VERIFY(nvlist_add_boolean(fslist, "sent") == 0); + progress = B_TRUE; + zfs_close(zhp); + if (err) + return (err); + } + if (needagain) { + assert(progress); + goto again; + } + return (0); +} + +/* + * Dumps a backup of tosnap, incremental from fromsnap if it isn't NULL. + * If 'doall', dump all intermediate snaps. + * If 'replicate', dump special header and do recursively. + */ +int +zfs_send(zfs_handle_t *zhp, const char *fromsnap, const char *tosnap, + boolean_t replicate, boolean_t doall, boolean_t fromorigin, + boolean_t verbose, int outfd) +{ + char errbuf[1024]; + send_dump_data_t sdd = { 0 }; + int err; + nvlist_t *fss = NULL; + avl_tree_t *fsavl = NULL; + + (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, + "cannot send '%s'"), zhp->zfs_name); + + if (fromsnap && fromsnap[0] == '\0') { + zfs_error_aux(zhp->zfs_hdl, dgettext(TEXT_DOMAIN, + "zero-length incremental source")); + return (zfs_error(zhp->zfs_hdl, EZFS_NOENT, errbuf)); + } + + if (replicate || doall) { + dmu_replay_record_t drr = { 0 }; + char *packbuf = NULL; + size_t buflen = 0; + zio_cksum_t zc = { 0 }; + + assert(fromsnap || doall); + + if (replicate) { + nvlist_t *hdrnv; + + VERIFY(0 == nvlist_alloc(&hdrnv, NV_UNIQUE_NAME, 0)); + if (fromsnap) { + VERIFY(0 == nvlist_add_string(hdrnv, + "fromsnap", fromsnap)); + } + VERIFY(0 == nvlist_add_string(hdrnv, "tosnap", tosnap)); + + err = gather_nvlist(zhp->zfs_hdl, zhp->zfs_name, + fromsnap, tosnap, &fss, &fsavl); + if (err) + return (err); + VERIFY(0 == nvlist_add_nvlist(hdrnv, "fss", fss)); + err = nvlist_pack(hdrnv, &packbuf, &buflen, + NV_ENCODE_XDR, 0); + nvlist_free(hdrnv); + if (err) { + fsavl_destroy(fsavl); + nvlist_free(fss); + return (zfs_standard_error(zhp->zfs_hdl, + err, errbuf)); + } + } + + /* write first begin record */ + drr.drr_type = DRR_BEGIN; + drr.drr_u.drr_begin.drr_magic = DMU_BACKUP_MAGIC; + drr.drr_u.drr_begin.drr_version = DMU_BACKUP_HEADER_VERSION; + (void) snprintf(drr.drr_u.drr_begin.drr_toname, + sizeof (drr.drr_u.drr_begin.drr_toname), + "%s@%s", zhp->zfs_name, tosnap); + drr.drr_payloadlen = buflen; + fletcher_4_incremental_native(&drr, sizeof (drr), &zc); + err = write(outfd, &drr, sizeof (drr)); + + /* write header nvlist */ + if (err != -1) { + fletcher_4_incremental_native(packbuf, buflen, &zc); + err = write(outfd, packbuf, buflen); + } + free(packbuf); + if (err == -1) { + fsavl_destroy(fsavl); + nvlist_free(fss); + return (zfs_standard_error(zhp->zfs_hdl, + errno, errbuf)); + } + + /* write end record */ + if (err != -1) { + bzero(&drr, sizeof (drr)); + drr.drr_type = DRR_END; + drr.drr_u.drr_end.drr_checksum = zc; + err = write(outfd, &drr, sizeof (drr)); + if (err == -1) { + fsavl_destroy(fsavl); + nvlist_free(fss); + return (zfs_standard_error(zhp->zfs_hdl, + errno, errbuf)); + } + } + } + + /* dump each stream */ + sdd.fromsnap = fromsnap; + sdd.tosnap = tosnap; + sdd.outfd = outfd; + sdd.replicate = replicate; + sdd.doall = doall; + sdd.fromorigin = fromorigin; + sdd.fss = fss; + sdd.fsavl = fsavl; + sdd.verbose = verbose; + err = dump_filesystems(zhp, &sdd); + fsavl_destroy(fsavl); + nvlist_free(fss); + + if (replicate || doall) { + /* + * write final end record. NB: want to do this even if + * there was some error, because it might not be totally + * failed. + */ + dmu_replay_record_t drr = { 0 }; + drr.drr_type = DRR_END; + if (write(outfd, &drr, sizeof (drr)) == -1) { + return (zfs_standard_error(zhp->zfs_hdl, + errno, errbuf)); + } + } + + return (err || sdd.err); +} + +/* + * Routines specific to "zfs recv" + */ + +static int +recv_read(libzfs_handle_t *hdl, int fd, void *buf, int ilen, + boolean_t byteswap, zio_cksum_t *zc) +{ + char *cp = buf; + int rv; + int len = ilen; + + do { + rv = read(fd, cp, len); + cp += rv; + len -= rv; + } while (rv > 0); + + if (rv < 0 || len != 0) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "failed to read from stream")); + return (zfs_error(hdl, EZFS_BADSTREAM, dgettext(TEXT_DOMAIN, + "cannot receive"))); + } + + if (zc) { + if (byteswap) + fletcher_4_incremental_byteswap(buf, ilen, zc); + else + fletcher_4_incremental_native(buf, ilen, zc); + } + return (0); +} + +static int +recv_read_nvlist(libzfs_handle_t *hdl, int fd, int len, nvlist_t **nvp, + boolean_t byteswap, zio_cksum_t *zc) +{ + char *buf; + int err; + + buf = zfs_alloc(hdl, len); + if (buf == NULL) + return (ENOMEM); + + err = recv_read(hdl, fd, buf, len, byteswap, zc); + if (err != 0) { + free(buf); + return (err); + } + + err = nvlist_unpack(buf, len, nvp, 0); + free(buf); + if (err != 0) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid " + "stream (malformed nvlist)")); + return (EINVAL); + } + return (0); +} + +static int +recv_rename(libzfs_handle_t *hdl, const char *name, const char *tryname, + int baselen, char *newname, recvflags_t flags) +{ + static int seq; + zfs_cmd_t zc = { 0 }; + int err; + prop_changelist_t *clp; + zfs_handle_t *zhp; + + zhp = zfs_open(hdl, name, ZFS_TYPE_DATASET); + if (zhp == NULL) + return (-1); + clp = changelist_gather(zhp, ZFS_PROP_NAME, 0, + flags.force ? MS_FORCE : 0); + zfs_close(zhp); + if (clp == NULL) + return (-1); + err = changelist_prefix(clp); + if (err) + return (err); + + if (tryname) { + (void) strcpy(newname, tryname); + + zc.zc_objset_type = DMU_OST_ZFS; + (void) strlcpy(zc.zc_name, name, sizeof (zc.zc_name)); + (void) strlcpy(zc.zc_value, tryname, sizeof (zc.zc_value)); + + if (flags.verbose) { + (void) printf("attempting rename %s to %s\n", + zc.zc_name, zc.zc_value); + } + err = ioctl(hdl->libzfs_fd, ZFS_IOC_RENAME, &zc); + if (err == 0) + changelist_rename(clp, name, tryname); + } else { + err = ENOENT; + } + + if (err != 0 && strncmp(name+baselen, "recv-", 5) != 0) { + seq++; + + (void) strncpy(newname, name, baselen); + (void) snprintf(newname+baselen, ZFS_MAXNAMELEN-baselen, + "recv-%u-%u", getpid(), seq); + (void) strlcpy(zc.zc_value, newname, sizeof (zc.zc_value)); + + if (flags.verbose) { + (void) printf("failed - trying rename %s to %s\n", + zc.zc_name, zc.zc_value); + } + err = ioctl(hdl->libzfs_fd, ZFS_IOC_RENAME, &zc); + if (err == 0) + changelist_rename(clp, name, newname); + if (err && flags.verbose) { + (void) printf("failed (%u) - " + "will try again on next pass\n", errno); + } + err = EAGAIN; + } else if (flags.verbose) { + if (err == 0) + (void) printf("success\n"); + else + (void) printf("failed (%u)\n", errno); + } + + (void) changelist_postfix(clp); + changelist_free(clp); + + return (err); +} + +static int +recv_destroy(libzfs_handle_t *hdl, const char *name, int baselen, + char *newname, recvflags_t flags) +{ + zfs_cmd_t zc = { 0 }; + int err = 0; + prop_changelist_t *clp; + zfs_handle_t *zhp; + + zhp = zfs_open(hdl, name, ZFS_TYPE_DATASET); + if (zhp == NULL) + return (-1); + clp = changelist_gather(zhp, ZFS_PROP_NAME, 0, + flags.force ? MS_FORCE : 0); + zfs_close(zhp); + if (clp == NULL) + return (-1); + err = changelist_prefix(clp); + if (err) + return (err); + + zc.zc_objset_type = DMU_OST_ZFS; + (void) strlcpy(zc.zc_name, name, sizeof (zc.zc_name)); + + if (flags.verbose) + (void) printf("attempting destroy %s\n", zc.zc_name); + err = ioctl(hdl->libzfs_fd, ZFS_IOC_DESTROY, &zc); + + if (err == 0) { + if (flags.verbose) + (void) printf("success\n"); + changelist_remove(clp, zc.zc_name); + } + + (void) changelist_postfix(clp); + changelist_free(clp); + + if (err != 0) + err = recv_rename(hdl, name, NULL, baselen, newname, flags); + + return (err); +} + +typedef struct guid_to_name_data { + uint64_t guid; + char *name; +} guid_to_name_data_t; + +static int +guid_to_name_cb(zfs_handle_t *zhp, void *arg) +{ + guid_to_name_data_t *gtnd = arg; + int err; + + if (zhp->zfs_dmustats.dds_guid == gtnd->guid) { + (void) strcpy(gtnd->name, zhp->zfs_name); + return (EEXIST); + } + err = zfs_iter_children(zhp, guid_to_name_cb, gtnd); + zfs_close(zhp); + return (err); +} + +static int +guid_to_name(libzfs_handle_t *hdl, const char *parent, uint64_t guid, + char *name) +{ + /* exhaustive search all local snapshots */ + guid_to_name_data_t gtnd; + int err = 0; + zfs_handle_t *zhp; + char *cp; + + gtnd.guid = guid; + gtnd.name = name; + + if (strchr(parent, '@') == NULL) { + zhp = make_dataset_handle(hdl, parent); + if (zhp != NULL) { + err = zfs_iter_children(zhp, guid_to_name_cb, >nd); + zfs_close(zhp); + if (err == EEXIST) + return (0); + } + } + + cp = strchr(parent, '/'); + if (cp) + *cp = '\0'; + zhp = make_dataset_handle(hdl, parent); + if (cp) + *cp = '/'; + + if (zhp) { + err = zfs_iter_children(zhp, guid_to_name_cb, >nd); + zfs_close(zhp); + } + + return (err == EEXIST ? 0 : ENOENT); + +} + +/* + * Return true if dataset guid1 is created before guid2. + */ +static int +created_before(libzfs_handle_t *hdl, avl_tree_t *avl, + uint64_t guid1, uint64_t guid2) +{ + nvlist_t *nvfs; + char *fsname, *snapname; + char buf[ZFS_MAXNAMELEN]; + int rv; + zfs_node_t zn1, zn2; + + if (guid2 == 0) + return (0); + if (guid1 == 0) + return (1); + + nvfs = fsavl_find(avl, guid1, &snapname); + VERIFY(0 == nvlist_lookup_string(nvfs, "name", &fsname)); + (void) snprintf(buf, sizeof (buf), "%s@%s", fsname, snapname); + zn1.zn_handle = zfs_open(hdl, buf, ZFS_TYPE_SNAPSHOT); + if (zn1.zn_handle == NULL) + return (-1); + + nvfs = fsavl_find(avl, guid2, &snapname); + VERIFY(0 == nvlist_lookup_string(nvfs, "name", &fsname)); + (void) snprintf(buf, sizeof (buf), "%s@%s", fsname, snapname); + zn2.zn_handle = zfs_open(hdl, buf, ZFS_TYPE_SNAPSHOT); + if (zn2.zn_handle == NULL) { + zfs_close(zn2.zn_handle); + return (-1); + } + + rv = (zfs_snapshot_compare(&zn1, &zn2) == -1); + + zfs_close(zn1.zn_handle); + zfs_close(zn2.zn_handle); + + return (rv); +} + +static int +recv_incremental_replication(libzfs_handle_t *hdl, const char *tofs, + recvflags_t flags, nvlist_t *stream_nv, avl_tree_t *stream_avl) +{ + nvlist_t *local_nv; + avl_tree_t *local_avl; + nvpair_t *fselem, *nextfselem; + char *tosnap, *fromsnap; + char newname[ZFS_MAXNAMELEN]; + int error; + boolean_t needagain, progress; + + VERIFY(0 == nvlist_lookup_string(stream_nv, "fromsnap", &fromsnap)); + VERIFY(0 == nvlist_lookup_string(stream_nv, "tosnap", &tosnap)); + + if (flags.dryrun) + return (0); + +again: + needagain = progress = B_FALSE; + + if ((error = gather_nvlist(hdl, tofs, fromsnap, NULL, + &local_nv, &local_avl)) != 0) + return (error); + + /* + * Process deletes and renames + */ + for (fselem = nvlist_next_nvpair(local_nv, NULL); + fselem; fselem = nextfselem) { + nvlist_t *nvfs, *snaps; + nvlist_t *stream_nvfs = NULL; + nvpair_t *snapelem, *nextsnapelem; + uint64_t fromguid = 0; + uint64_t originguid = 0; + uint64_t stream_originguid = 0; + uint64_t parent_fromsnap_guid, stream_parent_fromsnap_guid; + char *fsname, *stream_fsname; + + nextfselem = nvlist_next_nvpair(local_nv, fselem); + + VERIFY(0 == nvpair_value_nvlist(fselem, &nvfs)); + VERIFY(0 == nvlist_lookup_nvlist(nvfs, "snaps", &snaps)); + VERIFY(0 == nvlist_lookup_string(nvfs, "name", &fsname)); + VERIFY(0 == nvlist_lookup_uint64(nvfs, "parentfromsnap", + &parent_fromsnap_guid)); + (void) nvlist_lookup_uint64(nvfs, "origin", &originguid); + + /* + * First find the stream's fs, so we can check for + * a different origin (due to "zfs promote") + */ + for (snapelem = nvlist_next_nvpair(snaps, NULL); + snapelem; snapelem = nvlist_next_nvpair(snaps, snapelem)) { + uint64_t thisguid; + + VERIFY(0 == nvpair_value_uint64(snapelem, &thisguid)); + stream_nvfs = fsavl_find(stream_avl, thisguid, NULL); + + if (stream_nvfs != NULL) + break; + } + + /* check for promote */ + (void) nvlist_lookup_uint64(stream_nvfs, "origin", + &stream_originguid); + if (stream_nvfs && originguid != stream_originguid) { + switch (created_before(hdl, local_avl, + stream_originguid, originguid)) { + case 1: { + /* promote it! */ + zfs_cmd_t zc = { 0 }; + nvlist_t *origin_nvfs; + char *origin_fsname; + + if (flags.verbose) + (void) printf("promoting %s\n", fsname); + + origin_nvfs = fsavl_find(local_avl, originguid, + NULL); + VERIFY(0 == nvlist_lookup_string(origin_nvfs, + "name", &origin_fsname)); + (void) strlcpy(zc.zc_value, origin_fsname, + sizeof (zc.zc_value)); + (void) strlcpy(zc.zc_name, fsname, + sizeof (zc.zc_name)); + error = zfs_ioctl(hdl, ZFS_IOC_PROMOTE, &zc); + if (error == 0) + progress = B_TRUE; + break; + } + default: + break; + case -1: + fsavl_destroy(local_avl); + nvlist_free(local_nv); + return (-1); + } + /* + * We had/have the wrong origin, therefore our + * list of snapshots is wrong. Need to handle + * them on the next pass. + */ + needagain = B_TRUE; + continue; + } + + for (snapelem = nvlist_next_nvpair(snaps, NULL); + snapelem; snapelem = nextsnapelem) { + uint64_t thisguid; + char *stream_snapname; + nvlist_t *found, *props; + + nextsnapelem = nvlist_next_nvpair(snaps, snapelem); + + VERIFY(0 == nvpair_value_uint64(snapelem, &thisguid)); + found = fsavl_find(stream_avl, thisguid, + &stream_snapname); + + /* check for delete */ + if (found == NULL) { + char name[ZFS_MAXNAMELEN]; + + if (!flags.force) + continue; + + (void) snprintf(name, sizeof (name), "%s@%s", + fsname, nvpair_name(snapelem)); + + error = recv_destroy(hdl, name, + strlen(fsname)+1, newname, flags); + if (error) + needagain = B_TRUE; + else + progress = B_TRUE; + continue; + } + + stream_nvfs = found; + + if (0 == nvlist_lookup_nvlist(stream_nvfs, "snapprops", + &props) && 0 == nvlist_lookup_nvlist(props, + stream_snapname, &props)) { + zfs_cmd_t zc = { 0 }; + + zc.zc_cookie = B_TRUE; /* clear current props */ + (void) snprintf(zc.zc_name, sizeof (zc.zc_name), + "%s@%s", fsname, nvpair_name(snapelem)); + if (zcmd_write_src_nvlist(hdl, &zc, + props) == 0) { + (void) zfs_ioctl(hdl, + ZFS_IOC_SET_PROP, &zc); + zcmd_free_nvlists(&zc); + } + } + + /* check for different snapname */ + if (strcmp(nvpair_name(snapelem), + stream_snapname) != 0) { + char name[ZFS_MAXNAMELEN]; + char tryname[ZFS_MAXNAMELEN]; + + (void) snprintf(name, sizeof (name), "%s@%s", + fsname, nvpair_name(snapelem)); + (void) snprintf(tryname, sizeof (name), "%s@%s", + fsname, stream_snapname); + + error = recv_rename(hdl, name, tryname, + strlen(fsname)+1, newname, flags); + if (error) + needagain = B_TRUE; + else + progress = B_TRUE; + } + + if (strcmp(stream_snapname, fromsnap) == 0) + fromguid = thisguid; + } + + /* check for delete */ + if (stream_nvfs == NULL) { + if (!flags.force) + continue; + + error = recv_destroy(hdl, fsname, strlen(tofs)+1, + newname, flags); + if (error) + needagain = B_TRUE; + else + progress = B_TRUE; + continue; + } + + if (fromguid == 0 && flags.verbose) { + (void) printf("local fs %s does not have fromsnap " + "(%s in stream); must have been deleted locally; " + "ignoring\n", fsname, fromsnap); + continue; + } + + VERIFY(0 == nvlist_lookup_string(stream_nvfs, + "name", &stream_fsname)); + VERIFY(0 == nvlist_lookup_uint64(stream_nvfs, + "parentfromsnap", &stream_parent_fromsnap_guid)); + + /* check for rename */ + if ((stream_parent_fromsnap_guid != 0 && + stream_parent_fromsnap_guid != parent_fromsnap_guid) || + strcmp(strrchr(fsname, '/'), + strrchr(stream_fsname, '/')) != 0) { + nvlist_t *parent; + char tryname[ZFS_MAXNAMELEN]; + + parent = fsavl_find(local_avl, + stream_parent_fromsnap_guid, NULL); + /* + * NB: parent might not be found if we used the + * tosnap for stream_parent_fromsnap_guid, + * because the parent is a newly-created fs; + * we'll be able to rename it after we recv the + * new fs. + */ + if (parent != NULL) { + char *pname; + + VERIFY(0 == nvlist_lookup_string(parent, "name", + &pname)); + (void) snprintf(tryname, sizeof (tryname), + "%s%s", pname, strrchr(stream_fsname, '/')); + } else { + tryname[0] = '\0'; + if (flags.verbose) { + (void) printf("local fs %s new parent " + "not found\n", fsname); + } + } + + error = recv_rename(hdl, fsname, tryname, + strlen(tofs)+1, newname, flags); + if (error) + needagain = B_TRUE; + else + progress = B_TRUE; + } + } + + fsavl_destroy(local_avl); + nvlist_free(local_nv); + + if (needagain && progress) { + /* do another pass to fix up temporary names */ + if (flags.verbose) + (void) printf("another pass:\n"); + goto again; + } + + return (needagain); +} + +static int +zfs_receive_package(libzfs_handle_t *hdl, int fd, const char *destname, + recvflags_t flags, dmu_replay_record_t *drr, zio_cksum_t *zc, + char **top_zfs) +{ + nvlist_t *stream_nv = NULL; + avl_tree_t *stream_avl = NULL; + char *fromsnap = NULL; + char tofs[ZFS_MAXNAMELEN]; + char errbuf[1024]; + dmu_replay_record_t drre; + int error; + boolean_t anyerr = B_FALSE; + boolean_t softerr = B_FALSE; + + (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, + "cannot receive")); + + if (strchr(destname, '@')) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "can not specify snapshot name for multi-snapshot stream")); + return (zfs_error(hdl, EZFS_BADSTREAM, errbuf)); + } + + assert(drr->drr_type == DRR_BEGIN); + assert(drr->drr_u.drr_begin.drr_magic == DMU_BACKUP_MAGIC); + assert(drr->drr_u.drr_begin.drr_version == DMU_BACKUP_HEADER_VERSION); + + /* + * Read in the nvlist from the stream. + */ + if (drr->drr_payloadlen != 0) { + if (!flags.isprefix) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "must use -d to receive replication " + "(send -R) stream")); + return (zfs_error(hdl, EZFS_BADSTREAM, errbuf)); + } + + error = recv_read_nvlist(hdl, fd, drr->drr_payloadlen, + &stream_nv, flags.byteswap, zc); + if (error) { + error = zfs_error(hdl, EZFS_BADSTREAM, errbuf); + goto out; + } + } + + /* + * Read in the end record and verify checksum. + */ + if (0 != (error = recv_read(hdl, fd, &drre, sizeof (drre), + flags.byteswap, NULL))) + goto out; + if (flags.byteswap) { + drre.drr_type = BSWAP_32(drre.drr_type); + drre.drr_u.drr_end.drr_checksum.zc_word[0] = + BSWAP_64(drre.drr_u.drr_end.drr_checksum.zc_word[0]); + drre.drr_u.drr_end.drr_checksum.zc_word[1] = + BSWAP_64(drre.drr_u.drr_end.drr_checksum.zc_word[1]); + drre.drr_u.drr_end.drr_checksum.zc_word[2] = + BSWAP_64(drre.drr_u.drr_end.drr_checksum.zc_word[2]); + drre.drr_u.drr_end.drr_checksum.zc_word[3] = + BSWAP_64(drre.drr_u.drr_end.drr_checksum.zc_word[3]); + } + if (drre.drr_type != DRR_END) { + error = zfs_error(hdl, EZFS_BADSTREAM, errbuf); + goto out; + } + if (!ZIO_CHECKSUM_EQUAL(drre.drr_u.drr_end.drr_checksum, *zc)) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "incorrect header checksum")); + error = zfs_error(hdl, EZFS_BADSTREAM, errbuf); + goto out; + } + + (void) nvlist_lookup_string(stream_nv, "fromsnap", &fromsnap); + + if (drr->drr_payloadlen != 0) { + nvlist_t *stream_fss; + + VERIFY(0 == nvlist_lookup_nvlist(stream_nv, "fss", + &stream_fss)); + if ((stream_avl = fsavl_create(stream_fss)) == NULL) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "couldn't allocate avl tree")); + error = zfs_error(hdl, EZFS_NOMEM, errbuf); + goto out; + } + + if (fromsnap != NULL) { + (void) strlcpy(tofs, destname, ZFS_MAXNAMELEN); + if (flags.isprefix) { + int i = strcspn(drr->drr_u.drr_begin.drr_toname, + "/@"); + /* zfs_receive_one() will create_parents() */ + (void) strlcat(tofs, + &drr->drr_u.drr_begin.drr_toname[i], + ZFS_MAXNAMELEN); + *strchr(tofs, '@') = '\0'; + } + softerr = recv_incremental_replication(hdl, tofs, + flags, stream_nv, stream_avl); + } + } + + + /* Finally, receive each contained stream */ + do { + /* + * we should figure out if it has a recoverable + * error, in which case do a recv_skip() and drive on. + * Note, if we fail due to already having this guid, + * zfs_receive_one() will take care of it (ie, + * recv_skip() and return 0). + */ + error = zfs_receive_impl(hdl, destname, flags, fd, + stream_avl, top_zfs); + if (error == ENODATA) { + error = 0; + break; + } + anyerr |= error; + } while (error == 0); + + if (drr->drr_payloadlen != 0 && fromsnap != NULL) { + /* + * Now that we have the fs's they sent us, try the + * renames again. + */ + softerr = recv_incremental_replication(hdl, tofs, flags, + stream_nv, stream_avl); + } + +out: + fsavl_destroy(stream_avl); + if (stream_nv) + nvlist_free(stream_nv); + if (softerr) + error = -2; + if (anyerr) + error = -1; + return (error); +} + +static int +recv_skip(libzfs_handle_t *hdl, int fd, boolean_t byteswap) +{ + dmu_replay_record_t *drr; + void *buf = malloc(1<<20); + + /* XXX would be great to use lseek if possible... */ + drr = buf; + + while (recv_read(hdl, fd, drr, sizeof (dmu_replay_record_t), + byteswap, NULL) == 0) { + if (byteswap) + drr->drr_type = BSWAP_32(drr->drr_type); + + switch (drr->drr_type) { + case DRR_BEGIN: + /* NB: not to be used on v2 stream packages */ + assert(drr->drr_payloadlen == 0); + break; + + case DRR_END: + free(buf); + return (0); + + case DRR_OBJECT: + if (byteswap) { + drr->drr_u.drr_object.drr_bonuslen = + BSWAP_32(drr->drr_u.drr_object. + drr_bonuslen); + } + (void) recv_read(hdl, fd, buf, + P2ROUNDUP(drr->drr_u.drr_object.drr_bonuslen, 8), + B_FALSE, NULL); + break; + + case DRR_WRITE: + if (byteswap) { + drr->drr_u.drr_write.drr_length = + BSWAP_64(drr->drr_u.drr_write.drr_length); + } + (void) recv_read(hdl, fd, buf, + drr->drr_u.drr_write.drr_length, B_FALSE, NULL); + break; + + case DRR_FREEOBJECTS: + case DRR_FREE: + break; + + default: + assert(!"invalid record type"); + } + } + + free(buf); + return (-1); +} + +/* + * Restores a backup of tosnap from the file descriptor specified by infd. + */ +static int +zfs_receive_one(libzfs_handle_t *hdl, int infd, const char *tosnap, + recvflags_t flags, dmu_replay_record_t *drr, + dmu_replay_record_t *drr_noswap, avl_tree_t *stream_avl, + char **top_zfs) +{ + zfs_cmd_t zc = { 0 }; + time_t begin_time; + int ioctl_err, ioctl_errno, err, choplen; + char *cp; + struct drr_begin *drrb = &drr->drr_u.drr_begin; + char errbuf[1024]; + char chopprefix[ZFS_MAXNAMELEN]; + boolean_t newfs = B_FALSE; + boolean_t stream_wantsnewfs; + uint64_t parent_snapguid = 0; + prop_changelist_t *clp = NULL; + nvlist_t *snapprops_nvlist = NULL; + + begin_time = time(NULL); + + (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, + "cannot receive")); + + if (stream_avl != NULL) { + char *snapname; + nvlist_t *fs = fsavl_find(stream_avl, drrb->drr_toguid, + &snapname); + nvlist_t *props; + int ret; + + (void) nvlist_lookup_uint64(fs, "parentfromsnap", + &parent_snapguid); + err = nvlist_lookup_nvlist(fs, "props", &props); + if (err) + VERIFY(0 == nvlist_alloc(&props, NV_UNIQUE_NAME, 0)); + + if (flags.canmountoff) { + VERIFY(0 == nvlist_add_uint64(props, + zfs_prop_to_name(ZFS_PROP_CANMOUNT), 0)); + } + ret = zcmd_write_src_nvlist(hdl, &zc, props); + if (err) + nvlist_free(props); + + if (0 == nvlist_lookup_nvlist(fs, "snapprops", &props)) { + VERIFY(0 == nvlist_lookup_nvlist(props, + snapname, &snapprops_nvlist)); + } + + if (ret != 0) + return (-1); + } + + /* + * Determine how much of the snapshot name stored in the stream + * we are going to tack on to the name they specified on the + * command line, and how much we are going to chop off. + * + * If they specified a snapshot, chop the entire name stored in + * the stream. + */ + (void) strcpy(chopprefix, drrb->drr_toname); + if (flags.isprefix) { + /* + * They specified a fs with -d, we want to tack on + * everything but the pool name stored in the stream + */ + if (strchr(tosnap, '@')) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid " + "argument - snapshot not allowed with -d")); + return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); + } + cp = strchr(chopprefix, '/'); + if (cp == NULL) + cp = strchr(chopprefix, '@'); + *cp = '\0'; + } else if (strchr(tosnap, '@') == NULL) { + /* + * If they specified a filesystem without -d, we want to + * tack on everything after the fs specified in the + * first name from the stream. + */ + cp = strchr(chopprefix, '@'); + *cp = '\0'; + } + choplen = strlen(chopprefix); + + /* + * Determine name of destination snapshot, store in zc_value. + */ + (void) strcpy(zc.zc_value, tosnap); + (void) strncat(zc.zc_value, drrb->drr_toname+choplen, + sizeof (zc.zc_value)); + if (!zfs_name_valid(zc.zc_value, ZFS_TYPE_SNAPSHOT)) { + zcmd_free_nvlists(&zc); + return (zfs_error(hdl, EZFS_INVALIDNAME, errbuf)); + } + + /* + * Determine the name of the origin snapshot, store in zc_string. + */ + if (drrb->drr_flags & DRR_FLAG_CLONE) { + if (guid_to_name(hdl, tosnap, + drrb->drr_fromguid, zc.zc_string) != 0) { + zcmd_free_nvlists(&zc); + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "local origin for clone %s does not exist"), + zc.zc_value); + return (zfs_error(hdl, EZFS_NOENT, errbuf)); + } + if (flags.verbose) + (void) printf("found clone origin %s\n", zc.zc_string); + } + + stream_wantsnewfs = (drrb->drr_fromguid == 0 || + (drrb->drr_flags & DRR_FLAG_CLONE)); + + if (stream_wantsnewfs) { + /* + * if the parent fs does not exist, look for it based on + * the parent snap GUID + */ + (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, + "cannot receive new filesystem stream")); + + (void) strcpy(zc.zc_name, zc.zc_value); + cp = strrchr(zc.zc_name, '/'); + if (cp) + *cp = '\0'; + if (cp && + !zfs_dataset_exists(hdl, zc.zc_name, ZFS_TYPE_DATASET)) { + char suffix[ZFS_MAXNAMELEN]; + (void) strcpy(suffix, strrchr(zc.zc_value, '/')); + if (guid_to_name(hdl, tosnap, parent_snapguid, + zc.zc_value) == 0) { + *strchr(zc.zc_value, '@') = '\0'; + (void) strcat(zc.zc_value, suffix); + } + } + } else { + /* + * if the fs does not exist, look for it based on the + * fromsnap GUID + */ + (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, + "cannot receive incremental stream")); + + (void) strcpy(zc.zc_name, zc.zc_value); + *strchr(zc.zc_name, '@') = '\0'; + + if (!zfs_dataset_exists(hdl, zc.zc_name, ZFS_TYPE_DATASET)) { + char snap[ZFS_MAXNAMELEN]; + (void) strcpy(snap, strchr(zc.zc_value, '@')); + if (guid_to_name(hdl, tosnap, drrb->drr_fromguid, + zc.zc_value) == 0) { + *strchr(zc.zc_value, '@') = '\0'; + (void) strcat(zc.zc_value, snap); + } + } + } + + (void) strcpy(zc.zc_name, zc.zc_value); + *strchr(zc.zc_name, '@') = '\0'; + + if (zfs_dataset_exists(hdl, zc.zc_name, ZFS_TYPE_DATASET)) { + zfs_handle_t *zhp; + /* + * Destination fs exists. Therefore this should either + * be an incremental, or the stream specifies a new fs + * (full stream or clone) and they want us to blow it + * away (and have therefore specified -F and removed any + * snapshots). + */ + + if (stream_wantsnewfs) { + if (!flags.force) { + zcmd_free_nvlists(&zc); + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "destination '%s' exists\n" + "must specify -F to overwrite it"), + zc.zc_name); + return (zfs_error(hdl, EZFS_EXISTS, errbuf)); + } + if (ioctl(hdl->libzfs_fd, ZFS_IOC_SNAPSHOT_LIST_NEXT, + &zc) == 0) { + zcmd_free_nvlists(&zc); + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "destination has snapshots (eg. %s)\n" + "must destroy them to overwrite it"), + zc.zc_name); + return (zfs_error(hdl, EZFS_EXISTS, errbuf)); + } + } + + if ((zhp = zfs_open(hdl, zc.zc_name, + ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME)) == NULL) { + zcmd_free_nvlists(&zc); + return (-1); + } + + if (stream_wantsnewfs && + zhp->zfs_dmustats.dds_origin[0]) { + zcmd_free_nvlists(&zc); + zfs_close(zhp); + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "destination '%s' is a clone\n" + "must destroy it to overwrite it"), + zc.zc_name); + return (zfs_error(hdl, EZFS_EXISTS, errbuf)); + } + + if (!flags.dryrun && zhp->zfs_type == ZFS_TYPE_FILESYSTEM && + stream_wantsnewfs) { + /* We can't do online recv in this case */ + clp = changelist_gather(zhp, ZFS_PROP_NAME, 0, 0); + if (clp == NULL) { + zcmd_free_nvlists(&zc); + return (-1); + } + if (changelist_prefix(clp) != 0) { + changelist_free(clp); + zcmd_free_nvlists(&zc); + return (-1); + } + } + if (!flags.dryrun && zhp->zfs_type == ZFS_TYPE_VOLUME && + zvol_remove_link(hdl, zhp->zfs_name) != 0) { + zfs_close(zhp); + zcmd_free_nvlists(&zc); + return (-1); + } + zfs_close(zhp); + } else { + /* + * Destination filesystem does not exist. Therefore we better + * be creating a new filesystem (either from a full backup, or + * a clone). It would therefore be invalid if the user + * specified only the pool name (i.e. if the destination name + * contained no slash character). + */ + if (!stream_wantsnewfs || + (cp = strrchr(zc.zc_name, '/')) == NULL) { + zcmd_free_nvlists(&zc); + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "destination '%s' does not exist"), zc.zc_name); + return (zfs_error(hdl, EZFS_NOENT, errbuf)); + } + + /* + * Trim off the final dataset component so we perform the + * recvbackup ioctl to the filesystems's parent. + */ + *cp = '\0'; + + if (flags.isprefix && !flags.dryrun && + create_parents(hdl, zc.zc_value, strlen(tosnap)) != 0) { + zcmd_free_nvlists(&zc); + return (zfs_error(hdl, EZFS_BADRESTORE, errbuf)); + } + + newfs = B_TRUE; + } + + zc.zc_begin_record = drr_noswap->drr_u.drr_begin; + zc.zc_cookie = infd; + zc.zc_guid = flags.force; + if (flags.verbose) { + (void) printf("%s %s stream of %s into %s\n", + flags.dryrun ? "would receive" : "receiving", + drrb->drr_fromguid ? "incremental" : "full", + drrb->drr_toname, zc.zc_value); + (void) fflush(stdout); + } + + if (flags.dryrun) { + zcmd_free_nvlists(&zc); + return (recv_skip(hdl, infd, flags.byteswap)); + } + + err = ioctl_err = zfs_ioctl(hdl, ZFS_IOC_RECV, &zc); + ioctl_errno = errno; + zcmd_free_nvlists(&zc); + + if (err == 0 && snapprops_nvlist) { + zfs_cmd_t zc2 = { 0 }; + + (void) strcpy(zc2.zc_name, zc.zc_value); + if (zcmd_write_src_nvlist(hdl, &zc2, snapprops_nvlist) == 0) { + (void) zfs_ioctl(hdl, ZFS_IOC_SET_PROP, &zc2); + zcmd_free_nvlists(&zc2); + } + } + + if (err && (ioctl_errno == ENOENT || ioctl_errno == ENODEV)) { + /* + * It may be that this snapshot already exists, + * in which case we want to consume & ignore it + * rather than failing. + */ + avl_tree_t *local_avl; + nvlist_t *local_nv, *fs; + char *cp = strchr(zc.zc_value, '@'); + + /* + * XXX Do this faster by just iterating over snaps in + * this fs. Also if zc_value does not exist, we will + * get a strange "does not exist" error message. + */ + *cp = '\0'; + if (gather_nvlist(hdl, zc.zc_value, NULL, NULL, + &local_nv, &local_avl) == 0) { + *cp = '@'; + fs = fsavl_find(local_avl, drrb->drr_toguid, NULL); + fsavl_destroy(local_avl); + nvlist_free(local_nv); + + if (fs != NULL) { + if (flags.verbose) { + (void) printf("snap %s already exists; " + "ignoring\n", zc.zc_value); + } + ioctl_err = recv_skip(hdl, infd, + flags.byteswap); + } + } + *cp = '@'; + } + + + if (ioctl_err != 0) { + switch (ioctl_errno) { + case ENODEV: + cp = strchr(zc.zc_value, '@'); + *cp = '\0'; + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "most recent snapshot of %s does not\n" + "match incremental source"), zc.zc_value); + (void) zfs_error(hdl, EZFS_BADRESTORE, errbuf); + *cp = '@'; + break; + case ETXTBSY: + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "destination %s has been modified\n" + "since most recent snapshot"), zc.zc_name); + (void) zfs_error(hdl, EZFS_BADRESTORE, errbuf); + break; + case EEXIST: + cp = strchr(zc.zc_value, '@'); + if (newfs) { + /* it's the containing fs that exists */ + *cp = '\0'; + } + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "destination already exists")); + (void) zfs_error_fmt(hdl, EZFS_EXISTS, + dgettext(TEXT_DOMAIN, "cannot restore to %s"), + zc.zc_value); + *cp = '@'; + break; + case EINVAL: + (void) zfs_error(hdl, EZFS_BADSTREAM, errbuf); + break; + case ECKSUM: + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "invalid stream (checksum mismatch)")); + (void) zfs_error(hdl, EZFS_BADSTREAM, errbuf); + break; + default: + (void) zfs_standard_error(hdl, ioctl_errno, errbuf); + } + } + + /* + * Mount or recreate the /dev links for the target filesystem + * (if created, or if we tore them down to do an incremental + * restore), and the /dev links for the new snapshot (if + * created). Also mount any children of the target filesystem + * if we did an incremental receive. + */ + cp = strchr(zc.zc_value, '@'); + if (cp && (ioctl_err == 0 || !newfs)) { + zfs_handle_t *h; + + *cp = '\0'; + h = zfs_open(hdl, zc.zc_value, + ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME); + if (h != NULL) { + if (h->zfs_type == ZFS_TYPE_VOLUME) { + *cp = '@'; + err = zvol_create_link(hdl, h->zfs_name); + if (err == 0 && ioctl_err == 0) + err = zvol_create_link(hdl, + zc.zc_value); + } else if (newfs) { + /* + * Track the first/top of hierarchy fs, + * for mounting and sharing later. + */ + if (top_zfs && *top_zfs == NULL) + *top_zfs = zfs_strdup(hdl, zc.zc_value); + } + zfs_close(h); + } + *cp = '@'; + } + + if (clp) { + err |= changelist_postfix(clp); + changelist_free(clp); + } + + if (err || ioctl_err) + return (-1); + + if (flags.verbose) { + char buf1[64]; + char buf2[64]; + uint64_t bytes = zc.zc_cookie; + time_t delta = time(NULL) - begin_time; + if (delta == 0) + delta = 1; + zfs_nicenum(bytes, buf1, sizeof (buf1)); + zfs_nicenum(bytes/delta, buf2, sizeof (buf1)); + + (void) printf("received %sB stream in %lu seconds (%sB/sec)\n", + buf1, delta, buf2); + } + + return (0); +} + +static int +zfs_receive_impl(libzfs_handle_t *hdl, const char *tosnap, recvflags_t flags, + int infd, avl_tree_t *stream_avl, char **top_zfs) +{ + int err; + dmu_replay_record_t drr, drr_noswap; + struct drr_begin *drrb = &drr.drr_u.drr_begin; + char errbuf[1024]; + zio_cksum_t zcksum = { 0 }; + + (void) snprintf(errbuf, sizeof (errbuf), dgettext(TEXT_DOMAIN, + "cannot receive")); + + if (flags.isprefix && + !zfs_dataset_exists(hdl, tosnap, ZFS_TYPE_DATASET)) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "specified fs " + "(%s) does not exist"), tosnap); + return (zfs_error(hdl, EZFS_NOENT, errbuf)); + } + + /* read in the BEGIN record */ + if (0 != (err = recv_read(hdl, infd, &drr, sizeof (drr), B_FALSE, + &zcksum))) + return (err); + + if (drr.drr_type == DRR_END || drr.drr_type == BSWAP_32(DRR_END)) { + /* It's the double end record at the end of a package */ + return (ENODATA); + } + + /* the kernel needs the non-byteswapped begin record */ + drr_noswap = drr; + + flags.byteswap = B_FALSE; + if (drrb->drr_magic == BSWAP_64(DMU_BACKUP_MAGIC)) { + /* + * We computed the checksum in the wrong byteorder in + * recv_read() above; do it again correctly. + */ + bzero(&zcksum, sizeof (zio_cksum_t)); + fletcher_4_incremental_byteswap(&drr, sizeof (drr), &zcksum); + flags.byteswap = B_TRUE; + + drr.drr_type = BSWAP_32(drr.drr_type); + drr.drr_payloadlen = BSWAP_32(drr.drr_payloadlen); + drrb->drr_magic = BSWAP_64(drrb->drr_magic); + drrb->drr_version = BSWAP_64(drrb->drr_version); + drrb->drr_creation_time = BSWAP_64(drrb->drr_creation_time); + drrb->drr_type = BSWAP_32(drrb->drr_type); + drrb->drr_flags = BSWAP_32(drrb->drr_flags); + drrb->drr_toguid = BSWAP_64(drrb->drr_toguid); + drrb->drr_fromguid = BSWAP_64(drrb->drr_fromguid); + } + + if (drrb->drr_magic != DMU_BACKUP_MAGIC || drr.drr_type != DRR_BEGIN) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid " + "stream (bad magic number)")); + return (zfs_error(hdl, EZFS_BADSTREAM, errbuf)); + } + + if (strchr(drrb->drr_toname, '@') == NULL) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "invalid " + "stream (bad snapshot name)")); + return (zfs_error(hdl, EZFS_BADSTREAM, errbuf)); + } + + if (drrb->drr_version == DMU_BACKUP_STREAM_VERSION) { + return (zfs_receive_one(hdl, infd, tosnap, flags, + &drr, &drr_noswap, stream_avl, top_zfs)); + } else if (drrb->drr_version == DMU_BACKUP_HEADER_VERSION) { + return (zfs_receive_package(hdl, infd, tosnap, flags, + &drr, &zcksum, top_zfs)); + } else { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "stream is unsupported version %llu"), + drrb->drr_version); + return (zfs_error(hdl, EZFS_BADSTREAM, errbuf)); + } +} + +/* + * Restores a backup of tosnap from the file descriptor specified by infd. + * Return 0 on total success, -2 if some things couldn't be + * destroyed/renamed/promoted, -1 if some things couldn't be received. + * (-1 will override -2). + */ +int +zfs_receive(libzfs_handle_t *hdl, const char *tosnap, recvflags_t flags, + int infd, avl_tree_t *stream_avl) +{ + char *top_zfs = NULL; + int err; + + err = zfs_receive_impl(hdl, tosnap, flags, infd, stream_avl, &top_zfs); + + if (err == 0 && top_zfs) { + zfs_handle_t *zhp; + prop_changelist_t *clp; + + zhp = zfs_open(hdl, top_zfs, ZFS_TYPE_FILESYSTEM); + if (zhp != NULL) { + clp = changelist_gather(zhp, ZFS_PROP_MOUNTPOINT, + CL_GATHER_MOUNT_ALWAYS, 0); + zfs_close(zhp); + if (clp != NULL) { + /* mount and share received datasets */ + err = changelist_postfix(clp); + changelist_free(clp); + } + } + if (zhp == NULL || clp == NULL || err) + err = -1; + } + if (top_zfs) + free(top_zfs); + + return (err); +} diff --git a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_status.c b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_status.c index 3eba97a..c7eb04e 100644 --- a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_status.c +++ b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_status.c @@ -19,18 +19,16 @@ * CDDL HEADER END */ /* - * Copyright 2007 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ -#pragma ident "%Z%%M% %I% %E% SMI" - /* * This file contains the functions which analyze the status of a pool. This * include both the status of an active pool, as well as the status exported * pools. Returns one of the ZPOOL_STATUS_* defines describing the status of * the pool. This status is independent (to a certain degree) from the state of - * the pool. A pool's state descsribes only whether or not it is capable of + * the pool. A pool's state describes only whether or not it is capable of * providing the necessary fault tolerance for data. The status describes the * overall status of devices. A pool that is online can still have a device * that is experiencing errors. @@ -47,7 +45,7 @@ #include "libzfs_impl.h" /* - * Message ID table. This must be kep in sync with the ZPOOL_STATUS_* defines + * Message ID table. This must be kept in sync with the ZPOOL_STATUS_* defines * in libzfs.h. Note that there are some status results which go past the end * of this table, and hence have no associated message ID. */ @@ -62,26 +60,10 @@ static char *zfs_msgid_table[] = { "ZFS-8000-8A", "ZFS-8000-9P", "ZFS-8000-A5", - "ZFS-8000-EY" -}; - -/* - * If the pool is active, a certain class of static errors is overridden by the - * faults as analayzed by FMA. These faults have separate knowledge articles, - * and the article referred to by 'zpool status' must match that indicated by - * the syslog error message. We override missing data as well as corrupt pool. - */ -static char *zfs_msgid_table_active[] = { - "ZFS-8000-14", - "ZFS-8000-D3", /* overridden */ - "ZFS-8000-D3", /* overridden */ - "ZFS-8000-4J", - "ZFS-8000-5E", - "ZFS-8000-6X", - "ZFS-8000-CS", /* overridden */ - "ZFS-8000-8A", - "ZFS-8000-9P", - "ZFS-8000-CS", /* overridden */ + "ZFS-8000-EY", + "ZFS-8000-HC", + "ZFS-8000-JQ", + "ZFS-8000-K4", }; #define NMSGID (sizeof (zfs_msgid_table) / sizeof (zfs_msgid_table[0])) @@ -96,9 +78,16 @@ vdev_missing(uint64_t state, uint64_t aux, uint64_t errs) /* ARGSUSED */ static int +vdev_faulted(uint64_t state, uint64_t aux, uint64_t errs) +{ + return (state == VDEV_STATE_FAULTED); +} + +/* ARGSUSED */ +static int vdev_errors(uint64_t state, uint64_t aux, uint64_t errs) { - return (errs != 0); + return (state == VDEV_STATE_DEGRADED || errs != 0); } /* ARGSUSED */ @@ -163,9 +152,9 @@ find_vdev_problem(nvlist_t *vdev, int (*func)(uint64_t, uint64_t, uint64_t)) * following: * * - Check for a complete and valid configuration - * - Look for any missing devices in a non-replicated config + * - Look for any faulted or missing devices in a non-replicated config * - Check for any data errors - * - Check for any missing devices in a replicated config + * - Check for any faulted or missing devices in a replicated config * - Look for any devices showing errors * - Check for any resilvering devices * @@ -181,6 +170,7 @@ check_status(nvlist_t *config, boolean_t isimport) uint64_t nerr; uint64_t version; uint64_t stateval; + uint64_t suspended; uint64_t hostid = 0; verify(nvlist_lookup_uint64(config, ZPOOL_CONFIG_VERSION, @@ -215,9 +205,31 @@ check_status(nvlist_t *config, boolean_t isimport) return (ZPOOL_STATUS_BAD_GUID_SUM); /* - * Missing devices in non-replicated config. + * Check whether the pool has suspended due to failed I/O. + */ + if (nvlist_lookup_uint64(config, ZPOOL_CONFIG_SUSPENDED, + &suspended) == 0) { + if (suspended == ZIO_FAILURE_MODE_CONTINUE) + return (ZPOOL_STATUS_IO_FAILURE_CONTINUE); + return (ZPOOL_STATUS_IO_FAILURE_WAIT); + } + + /* + * Could not read a log. + */ + if (vs->vs_state == VDEV_STATE_CANT_OPEN && + vs->vs_aux == VDEV_AUX_BAD_LOG) { + return (ZPOOL_STATUS_BAD_LOG); + } + + /* + * Bad devices in non-replicated config. */ if (vs->vs_state == VDEV_STATE_CANT_OPEN && + find_vdev_problem(nvroot, vdev_faulted)) + return (ZPOOL_STATUS_FAULTED_DEV_NR); + + if (vs->vs_state == VDEV_STATE_CANT_OPEN && find_vdev_problem(nvroot, vdev_missing)) return (ZPOOL_STATUS_MISSING_DEV_NR); @@ -244,6 +256,8 @@ check_status(nvlist_t *config, boolean_t isimport) /* * Missing devices in a replicated config. */ + if (find_vdev_problem(nvroot, vdev_faulted)) + return (ZPOOL_STATUS_FAULTED_DEV_R); if (find_vdev_problem(nvroot, vdev_missing)) return (ZPOOL_STATUS_MISSING_DEV_R); if (find_vdev_problem(nvroot, vdev_broken)) @@ -270,7 +284,7 @@ check_status(nvlist_t *config, boolean_t isimport) /* * Outdated, but usable, version */ - if (version < ZFS_VERSION) + if (version < SPA_VERSION) return (ZPOOL_STATUS_VERSION_OLDER); return (ZPOOL_STATUS_OK); @@ -284,7 +298,7 @@ zpool_get_status(zpool_handle_t *zhp, char **msgid) if (ret >= NMSGID) *msgid = NULL; else - *msgid = zfs_msgid_table_active[ret]; + *msgid = zfs_msgid_table[ret]; return (ret); } diff --git a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_util.c b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_util.c index c706126..9d60d60 100644 --- a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_util.c +++ b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_util.c @@ -19,12 +19,10 @@ * CDDL HEADER END */ /* - * Copyright 2007 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ -#pragma ident "%Z%%M% %I% %E% SMI" - /* * Internal utility routines for the ZFS library. */ @@ -37,6 +35,8 @@ #include <stdlib.h> #include <strings.h> #include <unistd.h> +#include <ctype.h> +#include <math.h> #include <sys/mnttab.h> #include <sys/mntent.h> #include <sys/types.h> @@ -44,6 +44,7 @@ #include <libzfs.h> #include "libzfs_impl.h" +#include "zfs_prop.h" int libzfs_errno(libzfs_handle_t *hdl) @@ -133,6 +134,14 @@ libzfs_error_description(libzfs_handle_t *hdl) return (dgettext(TEXT_DOMAIN, "unshare(1M) failed")); case EZFS_SHARENFSFAILED: return (dgettext(TEXT_DOMAIN, "share(1M) failed")); + case EZFS_UNSHARESMBFAILED: + return (dgettext(TEXT_DOMAIN, "smb remove share failed")); + case EZFS_SHARESMBFAILED: + return (dgettext(TEXT_DOMAIN, "smb add share failed")); + case EZFS_ISCSISVCUNAVAIL: + return (dgettext(TEXT_DOMAIN, + "iscsitgt service need to be enabled by " + "a privileged user")); case EZFS_DEVLINKS: return (dgettext(TEXT_DOMAIN, "failed to create /dev links")); case EZFS_PERM: @@ -169,6 +178,38 @@ libzfs_error_description(libzfs_handle_t *hdl) "this pool operation")); case EZFS_NAMETOOLONG: return (dgettext(TEXT_DOMAIN, "dataset name is too long")); + case EZFS_OPENFAILED: + return (dgettext(TEXT_DOMAIN, "open failed")); + case EZFS_NOCAP: + return (dgettext(TEXT_DOMAIN, + "disk capacity information could not be retrieved")); + case EZFS_LABELFAILED: + return (dgettext(TEXT_DOMAIN, "write of label failed")); + case EZFS_BADWHO: + return (dgettext(TEXT_DOMAIN, "invalid user/group")); + case EZFS_BADPERM: + return (dgettext(TEXT_DOMAIN, "invalid permission")); + case EZFS_BADPERMSET: + return (dgettext(TEXT_DOMAIN, "invalid permission set name")); + case EZFS_NODELEGATION: + return (dgettext(TEXT_DOMAIN, "delegated administration is " + "disabled on pool")); + case EZFS_PERMRDONLY: + return (dgettext(TEXT_DOMAIN, "snapshot permissions cannot be" + " modified")); + case EZFS_BADCACHE: + return (dgettext(TEXT_DOMAIN, "invalid or missing cache file")); + case EZFS_ISL2CACHE: + return (dgettext(TEXT_DOMAIN, "device is in use as a cache")); + case EZFS_VDEVNOTSUP: + return (dgettext(TEXT_DOMAIN, "vdev specification is not " + "supported")); + case EZFS_NOTSUP: + return (dgettext(TEXT_DOMAIN, "operation not supported " + "on this dataset")); + case EZFS_ACTIVE_SPARE: + return (dgettext(TEXT_DOMAIN, "pool has active shared spare " + "device")); case EZFS_UNKNOWN: return (dgettext(TEXT_DOMAIN, "unknown error")); default: @@ -249,6 +290,10 @@ zfs_common_error(libzfs_handle_t *hdl, int error, const char *fmt, zfs_verror(hdl, EZFS_PERM, fmt, ap); return (-1); + case ECANCELED: + zfs_verror(hdl, EZFS_NODELEGATION, fmt, ap); + return (-1); + case EIO: zfs_verror(hdl, EZFS_IO, fmt, ap); return (-1); @@ -280,9 +325,9 @@ zfs_standard_error_fmt(libzfs_handle_t *hdl, int error, const char *fmt, ...) return (-1); } - switch (error) { case ENXIO: + case ENODEV: zfs_verror(hdl, EZFS_IO, fmt, ap); break; @@ -308,11 +353,17 @@ zfs_standard_error_fmt(libzfs_handle_t *hdl, int error, const char *fmt, ...) "dataset is busy")); zfs_verror(hdl, EZFS_BUSY, fmt, ap); break; - + case EROFS: + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "snapshot permissions cannot be modified")); + zfs_verror(hdl, EZFS_PERMRDONLY, fmt, ap); + break; case ENAMETOOLONG: zfs_verror(hdl, EZFS_NAMETOOLONG, fmt, ap); break; - + case ENOTSUP: + zfs_verror(hdl, EZFS_BADVERSION, fmt, ap); + break; default: zfs_error_aux(hdl, strerror(errno)); zfs_verror(hdl, EZFS_UNKNOWN, fmt, ap); @@ -361,7 +412,7 @@ zpool_standard_error_fmt(libzfs_handle_t *hdl, int error, const char *fmt, ...) case EBUSY: zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, "pool is busy")); - zfs_verror(hdl, EZFS_EXISTS, fmt, ap); + zfs_verror(hdl, EZFS_BUSY, fmt, ap); break; case ENXIO: @@ -382,6 +433,11 @@ zpool_standard_error_fmt(libzfs_handle_t *hdl, int error, const char *fmt, ...) zfs_verror(hdl, EZFS_POOL_INVALARG, fmt, ap); break; + case ENOSPC: + case EDQUOT: + zfs_verror(hdl, EZFS_NOSPC, fmt, ap); + return (-1); + default: zfs_error_aux(hdl, strerror(error)); zfs_verror(hdl, EZFS_UNKNOWN, fmt, ap); @@ -483,9 +539,8 @@ zfs_nicenum(uint64_t num, char *buf, size_t buflen) */ int i; for (i = 2; i >= 0; i--) { - (void) snprintf(buf, buflen, "%.*f%c", i, - (double)num / (1ULL << 10 * index), u); - if (strlen(buf) <= 5) + if (snprintf(buf, buflen, "%.*f%c", i, + (double)num / (1ULL << 10 * index), u) <= 5) break; } } @@ -538,6 +593,9 @@ libzfs_init(void) hdl->libzfs_sharetab = fopen(ZFS_EXPORTS_PATH, "r"); + zfs_prop_init(); + zpool_prop_init(); + return (hdl); } @@ -549,6 +607,10 @@ libzfs_fini(libzfs_handle_t *hdl) (void) fclose(hdl->libzfs_mnttab); if (hdl->libzfs_sharetab) (void) fclose(hdl->libzfs_sharetab); + zfs_uninit_libshare(hdl); + if (hdl->libzfs_log_str) + (void) free(hdl->libzfs_log_str); + zpool_free_handles(hdl); namespace_clear(hdl); free(hdl); } @@ -565,6 +627,12 @@ zfs_get_handle(zfs_handle_t *zhp) return (zhp->zfs_hdl); } +zpool_handle_t * +zfs_get_pool_handle(const zfs_handle_t *zhp) +{ + return (zhp->zpool_hdl); +} + /* * Given a name, determine whether or not it's a valid path * (starts with '/' or "./"). If so, walk the mnttab trying @@ -637,13 +705,14 @@ zcmd_expand_dst_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc) void zcmd_free_nvlists(zfs_cmd_t *zc) { + free((void *)(uintptr_t)zc->zc_nvlist_conf); free((void *)(uintptr_t)zc->zc_nvlist_src); free((void *)(uintptr_t)zc->zc_nvlist_dst); } -int -zcmd_write_src_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, nvlist_t *nvl, - size_t *size) +static int +zcmd_write_nvlist_com(libzfs_handle_t *hdl, uint64_t *outnv, uint64_t *outlen, + nvlist_t *nvl) { char *packed; size_t len; @@ -655,14 +724,26 @@ zcmd_write_src_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, nvlist_t *nvl, verify(nvlist_pack(nvl, &packed, &len, NV_ENCODE_NATIVE, 0) == 0); - zc->zc_nvlist_src = (uint64_t)(uintptr_t)packed; - zc->zc_nvlist_src_size = len; + *outnv = (uint64_t)(uintptr_t)packed; + *outlen = len; - if (size) - *size = len; return (0); } +int +zcmd_write_conf_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, nvlist_t *nvl) +{ + return (zcmd_write_nvlist_com(hdl, &zc->zc_nvlist_conf, + &zc->zc_nvlist_conf_size, nvl)); +} + +int +zcmd_write_src_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, nvlist_t *nvl) +{ + return (zcmd_write_nvlist_com(hdl, &zc->zc_nvlist_src, + &zc->zc_nvlist_src_size, nvl)); +} + /* * Unpacks an nvlist from the ZFS ioctl command structure. */ @@ -676,10 +757,32 @@ zcmd_read_dst_nvlist(libzfs_handle_t *hdl, zfs_cmd_t *zc, nvlist_t **nvlp) return (0); } +int +zfs_ioctl(libzfs_handle_t *hdl, int request, zfs_cmd_t *zc) +{ + int error; + + zc->zc_history = (uint64_t)(uintptr_t)hdl->libzfs_log_str; + error = ioctl(hdl->libzfs_fd, request, zc); + if (hdl->libzfs_log_str) { + free(hdl->libzfs_log_str); + hdl->libzfs_log_str = NULL; + } + zc->zc_history = 0; + + return (error); +} + +/* + * ================================================================ + * API shared by zfs and zpool property management + * ================================================================ + */ + static void -zfs_print_prop_headers(libzfs_get_cbdata_t *cbp) +zprop_print_headers(zprop_get_cbdata_t *cbp, zfs_type_t type) { - zfs_proplist_t *pl = cbp->cb_proplist; + zprop_list_t *pl = cbp->cb_proplist; int i; char *title; size_t len; @@ -711,8 +814,12 @@ zfs_print_prop_headers(libzfs_get_cbdata_t *cbp) /* * 'PROPERTY' column */ - if (pl->pl_prop != ZFS_PROP_INVAL) { - len = strlen(zfs_prop_to_name(pl->pl_prop)); + if (pl->pl_prop != ZPROP_INVAL) { + const char *propname = (type == ZFS_TYPE_POOL) ? + zpool_prop_to_name(pl->pl_prop) : + zfs_prop_to_name(pl->pl_prop); + + len = strlen(propname); if (len > cbp->cb_colwidths[GET_COL_PROPERTY]) cbp->cb_colwidths[GET_COL_PROPERTY] = len; } else { @@ -731,7 +838,8 @@ zfs_print_prop_headers(libzfs_get_cbdata_t *cbp) /* * 'NAME' and 'SOURCE' columns */ - if (pl->pl_prop == ZFS_PROP_NAME && + if (pl->pl_prop == (type == ZFS_TYPE_POOL ? ZPOOL_PROP_NAME : + ZFS_PROP_NAME) && pl->pl_width > cbp->cb_colwidths[GET_COL_NAME]) { cbp->cb_colwidths[GET_COL_NAME] = pl->pl_width; cbp->cb_colwidths[GET_COL_SOURCE] = pl->pl_width + @@ -777,8 +885,8 @@ zfs_print_prop_headers(libzfs_get_cbdata_t *cbp) * structure. */ void -libzfs_print_one_property(const char *name, libzfs_get_cbdata_t *cbp, - const char *propname, const char *value, zfs_source_t sourcetype, +zprop_print_one_property(const char *name, zprop_get_cbdata_t *cbp, + const char *propname, const char *value, zprop_source_t sourcetype, const char *source) { int i; @@ -792,7 +900,7 @@ libzfs_print_one_property(const char *name, libzfs_get_cbdata_t *cbp, return; if (cbp->cb_first) - zfs_print_prop_headers(cbp); + zprop_print_headers(cbp, cbp->cb_type); for (i = 0; i < 4; i++) { switch (cbp->cb_columns[i]) { @@ -810,23 +918,23 @@ libzfs_print_one_property(const char *name, libzfs_get_cbdata_t *cbp, case GET_COL_SOURCE: switch (sourcetype) { - case ZFS_SRC_NONE: + case ZPROP_SRC_NONE: str = "-"; break; - case ZFS_SRC_DEFAULT: + case ZPROP_SRC_DEFAULT: str = "default"; break; - case ZFS_SRC_LOCAL: + case ZPROP_SRC_LOCAL: str = "local"; break; - case ZFS_SRC_TEMPORARY: + case ZPROP_SRC_TEMPORARY: str = "temporary"; break; - case ZFS_SRC_INHERITED: + case ZPROP_SRC_INHERITED: (void) snprintf(buf, sizeof (buf), "inherited from %s", source); str = buf; @@ -851,3 +959,451 @@ libzfs_print_one_property(const char *name, libzfs_get_cbdata_t *cbp, (void) printf("\n"); } + +/* + * Given a numeric suffix, convert the value into a number of bits that the + * resulting value must be shifted. + */ +static int +str2shift(libzfs_handle_t *hdl, const char *buf) +{ + const char *ends = "BKMGTPEZ"; + int i; + + if (buf[0] == '\0') + return (0); + for (i = 0; i < strlen(ends); i++) { + if (toupper(buf[0]) == ends[i]) + break; + } + if (i == strlen(ends)) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "invalid numeric suffix '%s'"), buf); + return (-1); + } + + /* + * We want to allow trailing 'b' characters for 'GB' or 'Mb'. But don't + * allow 'BB' - that's just weird. + */ + if (buf[1] == '\0' || (toupper(buf[1]) == 'B' && buf[2] == '\0' && + toupper(buf[0]) != 'B')) + return (10*i); + + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "invalid numeric suffix '%s'"), buf); + return (-1); +} + +/* + * Convert a string of the form '100G' into a real number. Used when setting + * properties or creating a volume. 'buf' is used to place an extended error + * message for the caller to use. + */ +int +zfs_nicestrtonum(libzfs_handle_t *hdl, const char *value, uint64_t *num) +{ + char *end; + int shift; + + *num = 0; + + /* Check to see if this looks like a number. */ + if ((value[0] < '0' || value[0] > '9') && value[0] != '.') { + if (hdl) + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "bad numeric value '%s'"), value); + return (-1); + } + + /* Rely on stroll() to process the numeric portion. */ + errno = 0; + *num = strtoll(value, &end, 10); + + /* + * Check for ERANGE, which indicates that the value is too large to fit + * in a 64-bit value. + */ + if (errno == ERANGE) { + if (hdl) + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "numeric value is too large")); + return (-1); + } + + /* + * If we have a decimal value, then do the computation with floating + * point arithmetic. Otherwise, use standard arithmetic. + */ + if (*end == '.') { + double fval = strtod(value, &end); + + if ((shift = str2shift(hdl, end)) == -1) + return (-1); + + fval *= pow(2, shift); + + if (fval > UINT64_MAX) { + if (hdl) + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "numeric value is too large")); + return (-1); + } + + *num = (uint64_t)fval; + } else { + if ((shift = str2shift(hdl, end)) == -1) + return (-1); + + /* Check for overflow */ + if (shift >= 64 || (*num << shift) >> shift != *num) { + if (hdl) + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "numeric value is too large")); + return (-1); + } + + *num <<= shift; + } + + return (0); +} + +/* + * Given a propname=value nvpair to set, parse any numeric properties + * (index, boolean, etc) if they are specified as strings and add the + * resulting nvpair to the returned nvlist. + * + * At the DSL layer, all properties are either 64-bit numbers or strings. + * We want the user to be able to ignore this fact and specify properties + * as native values (numbers, for example) or as strings (to simplify + * command line utilities). This also handles converting index types + * (compression, checksum, etc) from strings to their on-disk index. + */ +int +zprop_parse_value(libzfs_handle_t *hdl, nvpair_t *elem, int prop, + zfs_type_t type, nvlist_t *ret, char **svalp, uint64_t *ivalp, + const char *errbuf) +{ + data_type_t datatype = nvpair_type(elem); + zprop_type_t proptype; + const char *propname; + char *value; + boolean_t isnone = B_FALSE; + + if (type == ZFS_TYPE_POOL) { + proptype = zpool_prop_get_type(prop); + propname = zpool_prop_to_name(prop); + } else { + proptype = zfs_prop_get_type(prop); + propname = zfs_prop_to_name(prop); + } + + /* + * Convert any properties to the internal DSL value types. + */ + *svalp = NULL; + *ivalp = 0; + + switch (proptype) { + case PROP_TYPE_STRING: + if (datatype != DATA_TYPE_STRING) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "'%s' must be a string"), nvpair_name(elem)); + goto error; + } + (void) nvpair_value_string(elem, svalp); + if (strlen(*svalp) >= ZFS_MAXPROPLEN) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "'%s' is too long"), nvpair_name(elem)); + goto error; + } + break; + + case PROP_TYPE_NUMBER: + if (datatype == DATA_TYPE_STRING) { + (void) nvpair_value_string(elem, &value); + if (strcmp(value, "none") == 0) { + isnone = B_TRUE; + } else if (zfs_nicestrtonum(hdl, value, ivalp) + != 0) { + goto error; + } + } else if (datatype == DATA_TYPE_UINT64) { + (void) nvpair_value_uint64(elem, ivalp); + } else { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "'%s' must be a number"), nvpair_name(elem)); + goto error; + } + + /* + * Quota special: force 'none' and don't allow 0. + */ + if ((type & ZFS_TYPE_DATASET) && *ivalp == 0 && !isnone && + (prop == ZFS_PROP_QUOTA || prop == ZFS_PROP_REFQUOTA)) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "use 'none' to disable quota/refquota")); + goto error; + } + break; + + case PROP_TYPE_INDEX: + if (datatype != DATA_TYPE_STRING) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "'%s' must be a string"), nvpair_name(elem)); + goto error; + } + + (void) nvpair_value_string(elem, &value); + + if (zprop_string_to_index(prop, value, ivalp, type) != 0) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "'%s' must be one of '%s'"), propname, + zprop_values(prop, type)); + goto error; + } + break; + + default: + abort(); + } + + /* + * Add the result to our return set of properties. + */ + if (*svalp != NULL) { + if (nvlist_add_string(ret, propname, *svalp) != 0) { + (void) no_memory(hdl); + return (-1); + } + } else { + if (nvlist_add_uint64(ret, propname, *ivalp) != 0) { + (void) no_memory(hdl); + return (-1); + } + } + + return (0); +error: + (void) zfs_error(hdl, EZFS_BADPROP, errbuf); + return (-1); +} + +static int +addlist(libzfs_handle_t *hdl, char *propname, zprop_list_t **listp, + zfs_type_t type) +{ + int prop; + zprop_list_t *entry; + + prop = zprop_name_to_prop(propname, type); + + if (prop != ZPROP_INVAL && !zprop_valid_for_type(prop, type)) + prop = ZPROP_INVAL; + + /* + * When no property table entry can be found, return failure if + * this is a pool property or if this isn't a user-defined + * dataset property, + */ + if (prop == ZPROP_INVAL && (type == ZFS_TYPE_POOL || + !zfs_prop_user(propname))) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "invalid property '%s'"), propname); + return (zfs_error(hdl, EZFS_BADPROP, + dgettext(TEXT_DOMAIN, "bad property list"))); + } + + if ((entry = zfs_alloc(hdl, sizeof (zprop_list_t))) == NULL) + return (-1); + + entry->pl_prop = prop; + if (prop == ZPROP_INVAL) { + if ((entry->pl_user_prop = zfs_strdup(hdl, propname)) == NULL) { + free(entry); + return (-1); + } + entry->pl_width = strlen(propname); + } else { + entry->pl_width = zprop_width(prop, &entry->pl_fixed, + type); + } + + *listp = entry; + + return (0); +} + +/* + * Given a comma-separated list of properties, construct a property list + * containing both user-defined and native properties. This function will + * return a NULL list if 'all' is specified, which can later be expanded + * by zprop_expand_list(). + */ +int +zprop_get_list(libzfs_handle_t *hdl, char *props, zprop_list_t **listp, + zfs_type_t type) +{ + *listp = NULL; + + /* + * If 'all' is specified, return a NULL list. + */ + if (strcmp(props, "all") == 0) + return (0); + + /* + * If no props were specified, return an error. + */ + if (props[0] == '\0') { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "no properties specified")); + return (zfs_error(hdl, EZFS_BADPROP, dgettext(TEXT_DOMAIN, + "bad property list"))); + } + + /* + * It would be nice to use getsubopt() here, but the inclusion of column + * aliases makes this more effort than it's worth. + */ + while (*props != '\0') { + size_t len; + char *p; + char c; + + if ((p = strchr(props, ',')) == NULL) { + len = strlen(props); + p = props + len; + } else { + len = p - props; + } + + /* + * Check for empty options. + */ + if (len == 0) { + zfs_error_aux(hdl, dgettext(TEXT_DOMAIN, + "empty property name")); + return (zfs_error(hdl, EZFS_BADPROP, + dgettext(TEXT_DOMAIN, "bad property list"))); + } + + /* + * Check all regular property names. + */ + c = props[len]; + props[len] = '\0'; + + if (strcmp(props, "space") == 0) { + static char *spaceprops[] = { + "name", "avail", "used", "usedbysnapshots", + "usedbydataset", "usedbyrefreservation", + "usedbychildren", NULL + }; + int i; + + for (i = 0; spaceprops[i]; i++) { + if (addlist(hdl, spaceprops[i], listp, type)) + return (-1); + listp = &(*listp)->pl_next; + } + } else { + if (addlist(hdl, props, listp, type)) + return (-1); + listp = &(*listp)->pl_next; + } + + props = p; + if (c == ',') + props++; + } + + return (0); +} + +void +zprop_free_list(zprop_list_t *pl) +{ + zprop_list_t *next; + + while (pl != NULL) { + next = pl->pl_next; + free(pl->pl_user_prop); + free(pl); + pl = next; + } +} + +typedef struct expand_data { + zprop_list_t **last; + libzfs_handle_t *hdl; + zfs_type_t type; +} expand_data_t; + +int +zprop_expand_list_cb(int prop, void *cb) +{ + zprop_list_t *entry; + expand_data_t *edp = cb; + + if ((entry = zfs_alloc(edp->hdl, sizeof (zprop_list_t))) == NULL) + return (ZPROP_INVAL); + + entry->pl_prop = prop; + entry->pl_width = zprop_width(prop, &entry->pl_fixed, edp->type); + entry->pl_all = B_TRUE; + + *(edp->last) = entry; + edp->last = &entry->pl_next; + + return (ZPROP_CONT); +} + +int +zprop_expand_list(libzfs_handle_t *hdl, zprop_list_t **plp, zfs_type_t type) +{ + zprop_list_t *entry; + zprop_list_t **last; + expand_data_t exp; + + if (*plp == NULL) { + /* + * If this is the very first time we've been called for an 'all' + * specification, expand the list to include all native + * properties. + */ + last = plp; + + exp.last = last; + exp.hdl = hdl; + exp.type = type; + + if (zprop_iter_common(zprop_expand_list_cb, &exp, B_FALSE, + B_FALSE, type) == ZPROP_INVAL) + return (-1); + + /* + * Add 'name' to the beginning of the list, which is handled + * specially. + */ + if ((entry = zfs_alloc(hdl, sizeof (zprop_list_t))) == NULL) + return (-1); + + entry->pl_prop = (type == ZFS_TYPE_POOL) ? ZPOOL_PROP_NAME : + ZFS_PROP_NAME; + entry->pl_width = zprop_width(entry->pl_prop, + &entry->pl_fixed, type); + entry->pl_all = B_TRUE; + entry->pl_next = *plp; + *plp = entry; + } + return (0); +} + +int +zprop_iter(zprop_func func, void *cb, boolean_t show_all, boolean_t ordered, + zfs_type_t type) +{ + return (zprop_iter_common(func, cb, show_all, ordered, type)); +} diff --git a/cddl/contrib/opensolaris/lib/libzpool/common/kernel.c b/cddl/contrib/opensolaris/lib/libzpool/common/kernel.c index 30c5a0c..467cf5c 100644 --- a/cddl/contrib/opensolaris/lib/libzpool/common/kernel.c +++ b/cddl/contrib/opensolaris/lib/libzpool/common/kernel.c @@ -19,7 +19,7 @@ * CDDL HEADER END */ /* - * Copyright 2007 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ @@ -101,20 +101,24 @@ void zmutex_init(kmutex_t *mp) { mp->m_owner = NULL; + mp->initialized = B_TRUE; (void) _mutex_init(&mp->m_lock, USYNC_THREAD, NULL); } void zmutex_destroy(kmutex_t *mp) { + ASSERT(mp->initialized == B_TRUE); ASSERT(mp->m_owner == NULL); (void) _mutex_destroy(&(mp)->m_lock); mp->m_owner = (void *)-1UL; + mp->initialized = B_FALSE; } void mutex_enter(kmutex_t *mp) { + ASSERT(mp->initialized == B_TRUE); ASSERT(mp->m_owner != (void *)-1UL); ASSERT(mp->m_owner != curthread); VERIFY(mutex_lock(&mp->m_lock) == 0); @@ -125,6 +129,7 @@ mutex_enter(kmutex_t *mp) int mutex_tryenter(kmutex_t *mp) { + ASSERT(mp->initialized == B_TRUE); ASSERT(mp->m_owner != (void *)-1UL); if (mutex_trylock(&mp->m_lock) == 0) { ASSERT(mp->m_owner == NULL); @@ -138,6 +143,7 @@ mutex_tryenter(kmutex_t *mp) void mutex_exit(kmutex_t *mp) { + ASSERT(mp->initialized == B_TRUE); ASSERT(mp->m_owner == curthread); mp->m_owner = NULL; VERIFY(mutex_unlock(&mp->m_lock) == 0); @@ -146,6 +152,7 @@ mutex_exit(kmutex_t *mp) void * mutex_owner(kmutex_t *mp) { + ASSERT(mp->initialized == B_TRUE); return (mp->m_owner); } @@ -160,7 +167,7 @@ rw_init(krwlock_t *rwlp, char *name, int type, void *arg) { rwlock_init(&rwlp->rw_lock, USYNC_THREAD, NULL); rwlp->rw_owner = NULL; - rwlp->rw_count = 0; + rwlp->initialized = B_TRUE; } void @@ -168,22 +175,23 @@ rw_destroy(krwlock_t *rwlp) { rwlock_destroy(&rwlp->rw_lock); rwlp->rw_owner = (void *)-1UL; - rwlp->rw_count = -2; + rwlp->initialized = B_FALSE; } void rw_enter(krwlock_t *rwlp, krw_t rw) { //ASSERT(!RW_LOCK_HELD(rwlp)); + ASSERT(rwlp->initialized == B_TRUE); ASSERT(rwlp->rw_owner != (void *)-1UL); ASSERT(rwlp->rw_owner != curthread); if (rw == RW_READER) { - (void) rw_rdlock(&rwlp->rw_lock); + VERIFY(rw_rdlock(&rwlp->rw_lock) == 0); ASSERT(rwlp->rw_count >= 0); atomic_add_int(&rwlp->rw_count, 1); } else { - (void) rw_wrlock(&rwlp->rw_lock); + VERIFY(rw_wrlock(&rwlp->rw_lock) == 0); ASSERT(rwlp->rw_count == 0); rwlp->rw_count = -1; rwlp->rw_owner = curthread; @@ -193,6 +201,7 @@ rw_enter(krwlock_t *rwlp, krw_t rw) void rw_exit(krwlock_t *rwlp) { + ASSERT(rwlp->initialized == B_TRUE); ASSERT(rwlp->rw_owner != (void *)-1UL); if (rwlp->rw_owner == curthread) { @@ -205,7 +214,7 @@ rw_exit(krwlock_t *rwlp) ASSERT(rwlp->rw_count > 0); atomic_add_int(&rwlp->rw_count, -1); } - (void) rw_unlock(&rwlp->rw_lock); + VERIFY(rw_unlock(&rwlp->rw_lock) == 0); } int @@ -213,6 +222,7 @@ rw_tryenter(krwlock_t *rwlp, krw_t rw) { int rv; + ASSERT(rwlp->initialized == B_TRUE); ASSERT(rwlp->rw_owner != (void *)-1UL); ASSERT(rwlp->rw_owner != curthread); @@ -241,6 +251,7 @@ rw_tryenter(krwlock_t *rwlp, krw_t rw) int rw_tryupgrade(krwlock_t *rwlp) { + ASSERT(rwlp->initialized == B_TRUE); ASSERT(rwlp->rw_owner != (void *)-1UL); return (0); @@ -422,9 +433,10 @@ vn_open(char *path, int x1, int flags, int mode, vnode_t **vpp, int x2, int x3) return (0); } +/*ARGSUSED*/ int vn_openat(char *path, int x1, int flags, int mode, vnode_t **vpp, int x2, - int x3, vnode_t *startvp) + int x3, vnode_t *startvp, int fd) { char *realpath = umem_alloc(strlen(path) + 2, UMEM_NOFAIL); int ret; @@ -432,6 +444,7 @@ vn_openat(char *path, int x1, int flags, int mode, vnode_t **vpp, int x2, ASSERT(startvp == rootdir); (void) sprintf(realpath, "/%s", path); + /* fd ignored for now, need if want to simulate nbmand support */ ret = vn_open(realpath, x1, flags, mode, vpp, x2, x3); umem_free(realpath, strlen(path) + 2); @@ -469,7 +482,7 @@ vn_rdwr(int uio, vnode_t *vp, void *addr, ssize_t len, offset_t offset, } void -vn_close(vnode_t *vp) +vn_close(vnode_t *vp, int openflag, cred_t *cr, kthread_t *td) { close(vp->v_fd); spa_strfree(vp->v_path); @@ -657,7 +670,8 @@ kobj_open_file(char *name) vnode_t *vp; /* set vp as the _fd field of the file */ - if (vn_openat(name, UIO_SYSSPACE, FREAD, 0, &vp, 0, 0, rootdir) != 0) + if (vn_openat(name, UIO_SYSSPACE, FREAD, 0, &vp, 0, 0, rootdir, + -1) != 0) return ((void *)-1UL); file = umem_zalloc(sizeof (struct _buf), UMEM_NOFAIL); @@ -679,7 +693,7 @@ kobj_read_file(struct _buf *file, char *buf, unsigned size, unsigned off) void kobj_close_file(struct _buf *file) { - vn_close((vnode_t *)file->_fd); + vn_close((vnode_t *)file->_fd, 0, NULL, NULL); umem_free(file, sizeof (struct _buf)); } @@ -690,7 +704,7 @@ kobj_get_filesize(struct _buf *file, uint64_t *size) vnode_t *vp = (vnode_t *)file->_fd; if (fstat64(vp->v_fd, &st) == -1) { - vn_close(vp); + vn_close(vp, 0, NULL, NULL); return (errno); } *size = st.st_size; @@ -746,10 +760,11 @@ highbit(ulong_t i) } #endif +static int random_fd = -1, urandom_fd = -1; + static int -random_get_bytes_common(uint8_t *ptr, size_t len, char *devname) +random_get_bytes_common(uint8_t *ptr, size_t len, int fd) { - int fd = open(devname, O_RDONLY); size_t resid = len; ssize_t bytes; @@ -757,26 +772,24 @@ random_get_bytes_common(uint8_t *ptr, size_t len, char *devname) while (resid != 0) { bytes = read(fd, ptr, resid); - ASSERT(bytes >= 0); + ASSERT3S(bytes, >=, 0); ptr += bytes; resid -= bytes; } - close(fd); - return (0); } int random_get_bytes(uint8_t *ptr, size_t len) { - return (random_get_bytes_common(ptr, len, "/dev/random")); + return (random_get_bytes_common(ptr, len, random_fd)); } int random_get_pseudo_bytes(uint8_t *ptr, size_t len) { - return (random_get_bytes_common(ptr, len, "/dev/urandom")); + return (random_get_bytes_common(ptr, len, urandom_fd)); } int @@ -815,7 +828,11 @@ kernel_init(int mode) dprintf("physmem = %llu pages (%.2f GB)\n", physmem, (double)physmem * sysconf(_SC_PAGE_SIZE) / (1ULL << 30)); - snprintf(hw_serial, sizeof (hw_serial), "%ld", gethostid()); + snprintf(hw_serial, sizeof (hw_serial), "%lu", + (unsigned long)gethostid()); + + VERIFY((random_fd = open("/dev/random", O_RDONLY)) != -1); + VERIFY((urandom_fd = open("/dev/urandom", O_RDONLY)) != -1); spa_init(mode); } @@ -824,6 +841,12 @@ void kernel_fini(void) { spa_fini(); + + close(random_fd); + close(urandom_fd); + + random_fd = -1; + urandom_fd = -1; } int @@ -850,3 +873,62 @@ z_compress_level(void *dst, size_t *dstlen, const void *src, size_t srclen, return (ret); } + +uid_t +crgetuid(cred_t *cr) +{ + return (0); +} + +gid_t +crgetgid(cred_t *cr) +{ + return (0); +} + +int +crgetngroups(cred_t *cr) +{ + return (0); +} + +gid_t * +crgetgroups(cred_t *cr) +{ + return (NULL); +} + +int +zfs_secpolicy_snapshot_perms(const char *name, cred_t *cr) +{ + return (0); +} + +int +zfs_secpolicy_rename_perms(const char *from, const char *to, cred_t *cr) +{ + return (0); +} + +int +zfs_secpolicy_destroy_perms(const char *name, cred_t *cr) +{ + return (0); +} + +ksiddomain_t * +ksid_lookupdomain(const char *dom) +{ + ksiddomain_t *kd; + + kd = umem_zalloc(sizeof (ksiddomain_t), UMEM_NOFAIL); + kd->kd_name = spa_strdup(dom); + return (kd); +} + +void +ksiddomain_rele(ksiddomain_t *ksid) +{ + spa_strfree(ksid->kd_name); + umem_free(ksid, sizeof (ksiddomain_t)); +} diff --git a/cddl/contrib/opensolaris/lib/libzpool/common/sys/zfs_context.h b/cddl/contrib/opensolaris/lib/libzpool/common/sys/zfs_context.h index 24572c9..971584d 100644 --- a/cddl/contrib/opensolaris/lib/libzpool/common/sys/zfs_context.h +++ b/cddl/contrib/opensolaris/lib/libzpool/common/sys/zfs_context.h @@ -19,15 +19,13 @@ * CDDL HEADER END */ /* - * Copyright 2007 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ #ifndef _SYS_ZFS_CONTEXT_H #define _SYS_ZFS_CONTEXT_H -#pragma ident "%Z%%M% %I% %E% SMI" - #ifdef __cplusplus extern "C" { #endif @@ -64,6 +62,7 @@ extern "C" { #include <fsshare.h> #include <sys/note.h> #include <sys/types.h> +#include <sys/cred.h> #include <sys/atomic.h> #include <sys/sysmacros.h> #include <sys/bitmap.h> @@ -78,8 +77,10 @@ extern "C" { #include <sys/debug.h> #include <sys/sdt.h> #include <sys/kstat.h> +#include <sys/u8_textprep.h> #include <sys/kernel.h> #include <sys/disk.h> +#include <sys/sysevent/eventdefs.h> #include <machine/atomic.h> #define ZFS_EXPORTS_PATH "/etc/zfs/exports" @@ -116,11 +117,12 @@ extern void vcmn_err(int, const char *, __va_list); extern void panic(const char *, ...); extern void vpanic(const char *, __va_list); +#define fm_panic panic + /* This definition is copied from assert.h. */ #if defined(__STDC__) #if __STDC_VERSION__ - 0 >= 199901L -#define verify(EX) (void)((EX) || \ - (__assert_c99(#EX, __FILE__, __LINE__, __func__), 0)) +#define verify(EX) (void)((EX) || (__assert(#EX, __FILE__, __LINE__), 0)) #else #define verify(EX) (void)((EX) || (__assert(#EX, __FILE__, __LINE__), 0)) #endif /* __STDC_VERSION__ - 0 >= 199901L */ @@ -167,11 +169,16 @@ _NOTE(CONSTCOND) } while (0) #endif /* - * Dtrace SDT probes have different signatures in userland than they do in + * DTrace SDT probes have different signatures in userland than they do in * kernel. If they're being used in kernel code, re-define them out of * existence for their counterparts in libzpool. */ +#ifdef DTRACE_PROBE +#undef DTRACE_PROBE +#define DTRACE_PROBE(a) ((void)0) +#endif /* DTRACE_PROBE */ + #ifdef DTRACE_PROBE1 #undef DTRACE_PROBE1 #define DTRACE_PROBE1(a, b, c) ((void)0) @@ -212,8 +219,9 @@ extern kthread_t *zk_thread_create(void (*func)(), void *arg); * Mutexes */ typedef struct kmutex { - void *m_owner; - mutex_t m_lock; + void *m_owner; + boolean_t initialized; + mutex_t m_lock; } kmutex_t; #define MUTEX_DEFAULT USYNC_THREAD @@ -243,6 +251,7 @@ extern void *mutex_owner(kmutex_t *mp); typedef struct krwlock { int rw_count; void *rw_owner; + boolean_t initialized; rwlock_t rw_lock; } krwlock_t; @@ -253,6 +262,7 @@ typedef int krw_t; #define RW_DEFAULT USYNC_THREAD #undef RW_READ_HELD +#define RW_READ_HELD(x) ((x)->rw_owner == NULL && (x)->rw_count > 0) #undef RW_WRITE_HELD #define RW_WRITE_HELD(x) ((x)->rw_owner == curthread) @@ -267,6 +277,11 @@ extern void rw_exit(krwlock_t *rwlp); extern int rw_lock_held(krwlock_t *rwlp); #define rw_downgrade(rwlp) do { } while (0) +extern uid_t crgetuid(cred_t *cr); +extern gid_t crgetgid(cred_t *cr); +extern int crgetngroups(cred_t *cr); +extern gid_t *crgetgroups(cred_t *cr); + /* * Condition variables */ @@ -285,6 +300,7 @@ extern void cv_broadcast(kcondvar_t *cv); * Kernel memory */ #define KM_SLEEP UMEM_NOFAIL +#define KM_PUSHPAGE KM_SLEEP #define KM_NOSLEEP UMEM_DEFAULT #define KMC_NODEBUG UMC_NODEBUG #define kmem_alloc(_s, _f) umem_alloc(_s, _f) @@ -322,6 +338,9 @@ extern void taskq_destroy(taskq_t *); extern void taskq_wait(taskq_t *); extern int taskq_member(taskq_t *, void *); +#define XVA_MAPSIZE 3 +#define XVA_MAGIC 0x78766174 + /* * vnodes */ @@ -331,44 +350,93 @@ typedef struct vnode { char *v_path; } vnode_t; + +typedef struct xoptattr { + timestruc_t xoa_createtime; /* Create time of file */ + uint8_t xoa_archive; + uint8_t xoa_system; + uint8_t xoa_readonly; + uint8_t xoa_hidden; + uint8_t xoa_nounlink; + uint8_t xoa_immutable; + uint8_t xoa_appendonly; + uint8_t xoa_nodump; + uint8_t xoa_settable; + uint8_t xoa_opaque; + uint8_t xoa_av_quarantined; + uint8_t xoa_av_modified; +} xoptattr_t; + typedef struct vattr { uint_t va_mask; /* bit-mask of attributes */ u_offset_t va_size; /* file size in bytes */ } vattr_t; -#define AT_TYPE 0x0001 -#define AT_MODE 0x0002 -#define AT_UID 0x0004 -#define AT_GID 0x0008 -#define AT_FSID 0x0010 -#define AT_NODEID 0x0020 -#define AT_NLINK 0x0040 -#define AT_SIZE 0x0080 -#define AT_ATIME 0x0100 -#define AT_MTIME 0x0200 -#define AT_CTIME 0x0400 -#define AT_RDEV 0x0800 -#define AT_BLKSIZE 0x1000 -#define AT_NBLOCKS 0x2000 -#define AT_SEQ 0x8000 + +typedef struct xvattr { + vattr_t xva_vattr; /* Embedded vattr structure */ + uint32_t xva_magic; /* Magic Number */ + uint32_t xva_mapsize; /* Size of attr bitmap (32-bit words) */ + uint32_t *xva_rtnattrmapp; /* Ptr to xva_rtnattrmap[] */ + uint32_t xva_reqattrmap[XVA_MAPSIZE]; /* Requested attrs */ + uint32_t xva_rtnattrmap[XVA_MAPSIZE]; /* Returned attrs */ + xoptattr_t xva_xoptattrs; /* Optional attributes */ +} xvattr_t; + +typedef struct vsecattr { + uint_t vsa_mask; /* See below */ + int vsa_aclcnt; /* ACL entry count */ + void *vsa_aclentp; /* pointer to ACL entries */ + int vsa_dfaclcnt; /* default ACL entry count */ + void *vsa_dfaclentp; /* pointer to default ACL entries */ + size_t vsa_aclentsz; /* ACE size in bytes of vsa_aclentp */ +} vsecattr_t; + +#define AT_TYPE 0x00001 +#define AT_MODE 0x00002 +#define AT_UID 0x00004 +#define AT_GID 0x00008 +#define AT_FSID 0x00010 +#define AT_NODEID 0x00020 +#define AT_NLINK 0x00040 +#define AT_SIZE 0x00080 +#define AT_ATIME 0x00100 +#define AT_MTIME 0x00200 +#define AT_CTIME 0x00400 +#define AT_RDEV 0x00800 +#define AT_BLKSIZE 0x01000 +#define AT_NBLOCKS 0x02000 +#define AT_SEQ 0x08000 +#define AT_XVATTR 0x10000 #define CRCREAT 0 -#define VOP_CLOSE(vp, f, c, o, cr) 0 -#define VOP_PUTPAGE(vp, of, sz, fl, cr) 0 -#define VOP_GETATTR(vp, vap, fl) ((vap)->va_size = (vp)->v_size, 0) +#define VOP_CLOSE(vp, f, c, o, cr, ct) 0 +#define VOP_PUTPAGE(vp, of, sz, fl, cr, ct) 0 +#define VOP_GETATTR(vp, vap, cr) ((vap)->va_size = (vp)->v_size, 0) -#define VOP_FSYNC(vp, f, cr) fsync((vp)->v_fd) +#define VOP_FSYNC(vp, f, cr, ct) fsync((vp)->v_fd) -#define VN_RELE(vp) vn_close(vp) +#define VN_RELE(vp) vn_close(vp, 0, NULL, NULL) + +#define vn_lock(vp, type) +#define VOP_UNLOCK(vp, type) +#ifdef VFS_LOCK_GIANT +#undef VFS_LOCK_GIANT +#endif +#define VFS_LOCK_GIANT(mp) 0 +#ifdef VFS_UNLOCK_GIANT +#undef VFS_UNLOCK_GIANT +#endif +#define VFS_UNLOCK_GIANT(vfslocked) extern int vn_open(char *path, int x1, int oflags, int mode, vnode_t **vpp, int x2, int x3); extern int vn_openat(char *path, int x1, int oflags, int mode, vnode_t **vpp, - int x2, int x3, vnode_t *vp); + int x2, int x3, vnode_t *vp, int fd); extern int vn_rdwr(int uio, vnode_t *vp, void *addr, ssize_t len, offset_t offset, int x1, int x2, rlim64_t x3, void *x4, ssize_t *residp); -extern void vn_close(vnode_t *vp); +extern void vn_close(vnode_t *vp, int openflag, cred_t *cr, kthread_t *td); #define vn_remove(path, x1, x2) remove(path) #define vn_rename(from, to, seg) rename((from), (to)) @@ -397,8 +465,9 @@ extern void delay(clock_t ticks); #define CPU_SEQID (thr_self() & (max_ncpus - 1)) -#define kcred NULL -#define CRED() NULL +#ifndef ptob +#define ptob(x) ((x) * PAGESIZE) +#endif extern uint64_t physmem; @@ -455,11 +524,31 @@ struct bootstat { uint64_t st_size; }; +typedef struct ace_object { + uid_t a_who; + uint32_t a_access_mask; + uint16_t a_flags; + uint16_t a_type; + uint8_t a_obj_type[16]; + uint8_t a_inherit_obj_type[16]; +} ace_object_t; + + +#define ACE_ACCESS_ALLOWED_OBJECT_ACE_TYPE 0x05 +#define ACE_ACCESS_DENIED_OBJECT_ACE_TYPE 0x06 +#define ACE_SYSTEM_AUDIT_OBJECT_ACE_TYPE 0x07 +#define ACE_SYSTEM_ALARM_OBJECT_ACE_TYPE 0x08 + extern struct _buf *kobj_open_file(char *name); extern int kobj_read_file(struct _buf *file, char *buf, unsigned size, unsigned off); extern void kobj_close_file(struct _buf *file); extern int kobj_get_filesize(struct _buf *file, uint64_t *size); +extern int zfs_secpolicy_snapshot_perms(const char *name, cred_t *cr); +extern int zfs_secpolicy_rename_perms(const char *from, const char *to, + cred_t *cr); +extern int zfs_secpolicy_destroy_perms(const char *name, cred_t *cr); +extern zoneid_t getzoneid(void); /* Random compatibility stuff. */ #define lbolt (gethrtime() >> 23) #define lbolt64 (gethrtime() >> 23) @@ -482,18 +571,32 @@ struct file { #define FCREAT O_CREAT #define FOFFMAX 0x0 +/* SID stuff */ +typedef struct ksiddomain { + uint_t kd_ref; + uint_t kd_len; + char *kd_name; +} ksiddomain_t; + +ksiddomain_t *ksid_lookupdomain(const char *); +void ksiddomain_rele(ksiddomain_t *); + #define SX_SYSINIT(name, lock, desc) #define SYSCTL_DECL(...) #define SYSCTL_NODE(...) #define SYSCTL_INT(...) +#define SYSCTL_UINT(...) #define SYSCTL_ULONG(...) +#define SYSCTL_QUAD(...) #ifdef TUNABLE_INT #undef TUNABLE_INT #undef TUNABLE_ULONG +#undef TUNABLE_QUAD #endif #define TUNABLE_INT(...) #define TUNABLE_ULONG(...) +#define TUNABLE_QUAD(...) /* Errors */ diff --git a/cddl/contrib/opensolaris/lib/libzpool/common/taskq.c b/cddl/contrib/opensolaris/lib/libzpool/common/taskq.c index f7b6571..ccf5b4d 100644 --- a/cddl/contrib/opensolaris/lib/libzpool/common/taskq.c +++ b/cddl/contrib/opensolaris/lib/libzpool/common/taskq.c @@ -2,9 +2,8 @@ * CDDL HEADER START * * The contents of this file are subject to the terms of the - * Common Development and Distribution License, Version 1.0 only - * (the "License"). You may not use this file except in compliance - * with the License. + * Common Development and Distribution License (the "License"). + * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. @@ -20,7 +19,7 @@ * CDDL HEADER END */ /* - * Copyright 2005 Sun Microsystems, Inc. All rights reserved. + * Copyright 2007 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ @@ -177,6 +176,9 @@ taskq_create(const char *name, int nthreads, pri_t pri, int t; rw_init(&tq->tq_threadlock, NULL, RW_DEFAULT, NULL); + mutex_init(&tq->tq_lock, NULL, MUTEX_DEFAULT, NULL); + cv_init(&tq->tq_dispatch_cv, NULL, CV_DEFAULT, NULL); + cv_init(&tq->tq_wait_cv, NULL, CV_DEFAULT, NULL); tq->tq_flags = flags | TASKQ_ACTIVE; tq->tq_active = nthreads; tq->tq_nthreads = nthreads; @@ -230,6 +232,9 @@ taskq_destroy(taskq_t *tq) kmem_free(tq->tq_threadlist, nthreads * sizeof (thread_t)); rw_destroy(&tq->tq_threadlock); + mutex_destroy(&tq->tq_lock); + cv_destroy(&tq->tq_dispatch_cv); + cv_destroy(&tq->tq_wait_cv); kmem_free(tq, sizeof (taskq_t)); } diff --git a/cddl/contrib/opensolaris/lib/libzpool/common/util.c b/cddl/contrib/opensolaris/lib/libzpool/common/util.c index df49adb..781edb6 100644 --- a/cddl/contrib/opensolaris/lib/libzpool/common/util.c +++ b/cddl/contrib/opensolaris/lib/libzpool/common/util.c @@ -19,12 +19,10 @@ * CDDL HEADER END */ /* - * Copyright 2006 Sun Microsystems, Inc. All rights reserved. + * Copyright 2008 Sun Microsystems, Inc. All rights reserved. * Use is subject to license terms. */ -#pragma ident "%Z%%M% %I% %E% SMI" - #include <assert.h> #include <sys/zfs_context.h> #include <sys/avl.h> @@ -67,46 +65,58 @@ nicenum(uint64_t num, char *buf) } static void -show_vdev_stats(const char *desc, nvlist_t *nv, int indent) +show_vdev_stats(const char *desc, const char *ctype, nvlist_t *nv, int indent) { - nvlist_t **child; - uint_t c, children; vdev_stat_t *vs; + vdev_stat_t v0 = { 0 }; uint64_t sec; + uint64_t is_log = 0; + nvlist_t **child; + uint_t c, children; char used[6], avail[6]; char rops[6], wops[6], rbytes[6], wbytes[6], rerr[6], werr[6], cerr[6]; + char *prefix = ""; - if (indent == 0) { - (void) printf(" " + if (indent == 0 && desc != NULL) { + (void) printf(" " " capacity operations bandwidth ---- errors ----\n"); - (void) printf("description " + (void) printf("description " "used avail read write read write read write cksum\n"); } - VERIFY(nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_STATS, - (uint64_t **)&vs, &c) == 0); - - sec = MAX(1, vs->vs_timestamp / NANOSEC); - - nicenum(vs->vs_alloc, used); - nicenum(vs->vs_space - vs->vs_alloc, avail); - nicenum(vs->vs_ops[ZIO_TYPE_READ] / sec, rops); - nicenum(vs->vs_ops[ZIO_TYPE_WRITE] / sec, wops); - nicenum(vs->vs_bytes[ZIO_TYPE_READ] / sec, rbytes); - nicenum(vs->vs_bytes[ZIO_TYPE_WRITE] / sec, wbytes); - nicenum(vs->vs_read_errors, rerr); - nicenum(vs->vs_write_errors, werr); - nicenum(vs->vs_checksum_errors, cerr); - - (void) printf("%*s%*s%*s%*s %5s %5s %5s %5s %5s %5s %5s\n", - indent, "", - indent - 19 - (vs->vs_space ? 0 : 12), desc, - vs->vs_space ? 6 : 0, vs->vs_space ? used : "", - vs->vs_space ? 6 : 0, vs->vs_space ? avail : "", - rops, wops, rbytes, wbytes, rerr, werr, cerr); - - if (nvlist_lookup_nvlist_array(nv, ZPOOL_CONFIG_CHILDREN, - &child, &children) != 0) + if (desc != NULL) { + (void) nvlist_lookup_uint64(nv, ZPOOL_CONFIG_IS_LOG, &is_log); + + if (is_log) + prefix = "log "; + + if (nvlist_lookup_uint64_array(nv, ZPOOL_CONFIG_STATS, + (uint64_t **)&vs, &c) != 0) + vs = &v0; + + sec = MAX(1, vs->vs_timestamp / NANOSEC); + + nicenum(vs->vs_alloc, used); + nicenum(vs->vs_space - vs->vs_alloc, avail); + nicenum(vs->vs_ops[ZIO_TYPE_READ] / sec, rops); + nicenum(vs->vs_ops[ZIO_TYPE_WRITE] / sec, wops); + nicenum(vs->vs_bytes[ZIO_TYPE_READ] / sec, rbytes); + nicenum(vs->vs_bytes[ZIO_TYPE_WRITE] / sec, wbytes); + nicenum(vs->vs_read_errors, rerr); + nicenum(vs->vs_write_errors, werr); + nicenum(vs->vs_checksum_errors, cerr); + + (void) printf("%*s%s%*s%*s%*s %5s %5s %5s %5s %5s %5s %5s\n", + indent, "", + prefix, + indent + strlen(prefix) - 25 - (vs->vs_space ? 0 : 12), + desc, + vs->vs_space ? 6 : 0, vs->vs_space ? used : "", + vs->vs_space ? 6 : 0, vs->vs_space ? avail : "", + rops, wops, rbytes, wbytes, rerr, werr, cerr); + } + + if (nvlist_lookup_nvlist_array(nv, ctype, &child, &children) != 0) return; for (c = 0; c < children; c++) { @@ -120,7 +130,7 @@ show_vdev_stats(const char *desc, nvlist_t *nv, int indent) (void) strcpy(tname, cname); if (nvlist_lookup_uint64(cnv, ZPOOL_CONFIG_NPARITY, &np) == 0) tname[strlen(tname)] = '0' + np; - show_vdev_stats(tname, cnv, indent + 2); + show_vdev_stats(tname, ctype, cnv, indent + 2); free(tname); } } @@ -131,14 +141,16 @@ show_pool_stats(spa_t *spa) nvlist_t *config, *nvroot; char *name; - spa_config_enter(spa, RW_READER, FTAG); - config = spa_config_generate(spa, NULL, -1ULL, B_TRUE); - spa_config_exit(spa, FTAG); + VERIFY(spa_get_stats(spa_name(spa), &config, NULL, 0) == 0); VERIFY(nvlist_lookup_nvlist(config, ZPOOL_CONFIG_VDEV_TREE, &nvroot) == 0); VERIFY(nvlist_lookup_string(config, ZPOOL_CONFIG_POOL_NAME, &name) == 0); - show_vdev_stats(name, nvroot, 0); + show_vdev_stats(name, ZPOOL_CONFIG_CHILDREN, nvroot, 0); + show_vdev_stats(NULL, ZPOOL_CONFIG_L2CACHE, nvroot, 0); + show_vdev_stats(NULL, ZPOOL_CONFIG_SPARES, nvroot, 0); + + nvlist_free(config); } diff --git a/cddl/lib/libzfs/Makefile b/cddl/lib/libzfs/Makefile index a896b8f..30c516b 100644 --- a/cddl/lib/libzfs/Makefile +++ b/cddl/lib/libzfs/Makefile @@ -10,14 +10,17 @@ DPADD= ${LIBUTIL} LDADD= -lutil SRCS= deviceid.c \ - mnttab.c \ + fsshare.c \ mkdirp.c \ + mnttab.c \ zmount.c \ - fsshare.c \ zone.c -SRCS+= zfs_namecheck.c \ +SRCS+= zfs_deleg.c \ + zfs_namecheck.c \ zfs_prop.c \ + zpool_prop.c \ + zprop_common.c \ libzfs_dataset.c \ libzfs_util.c \ libzfs_graph.c \ @@ -26,6 +29,7 @@ SRCS+= zfs_namecheck.c \ libzfs_changelist.c \ libzfs_config.c \ libzfs_import.c \ + libzfs_sendrecv.c \ libzfs_status.c CFLAGS+= -DZFS_NO_ACL diff --git a/cddl/lib/libzpool/Makefile b/cddl/lib/libzpool/Makefile index 0890073..26d5499 100644 --- a/cddl/lib/libzpool/Makefile +++ b/cddl/lib/libzpool/Makefile @@ -18,6 +18,8 @@ ATOMIC_SRCS= atomic.S .PATH: ${.CURDIR}/../../../sys/cddl/compat/opensolaris/kern ATOMIC_SRCS= opensolaris_atomic.c .endif +# UNICODE_SRCS +.PATH: ${.CURDIR}/../../../sys/cddl/contrib/opensolaris/common/unicode LIB= zpool @@ -25,9 +27,13 @@ ZFS_COMMON_SRCS= ${ZFS_COMMON_OBJS:C/.o$/.c/} vdev_file.c ZFS_SHARED_SRCS= ${ZFS_SHARED_OBJS:C/.o$/.c/} KERNEL_SRCS= kernel.c taskq.c util.c LIST_SRCS= list.c +UNICODE_SRCS= u8_textprep.c SRCS= ${ZFS_COMMON_SRCS} ${ZFS_SHARED_SRCS} \ - ${KERNEL_SRCS} ${LIST_SRCS} ${ATOMIC_SRCS} + ${KERNEL_SRCS} ${LIST_SRCS} ${ATOMIC_SRCS} \ + ${UNICODE_SRCS} + +CFLAGS+= -std=c99 CFLAGS+= -I${.CURDIR}/../../../sys/cddl/compat/opensolaris CFLAGS+= -I${.CURDIR}/../../../cddl/compat/opensolaris/include @@ -40,6 +46,13 @@ CFLAGS+= -I${.CURDIR}/../../../sys/cddl/contrib/opensolaris/uts/common CFLAGS+= -I${.CURDIR}/../../../cddl/contrib/opensolaris/head CFLAGS+= -I${.CURDIR}/../../../cddl/lib/libumem CFLAGS+= -I${.CURDIR}/../../../cddl/contrib/opensolaris/lib/libnvpair +# XXX: pthread doesn't have mutex_owned() equivalent, so we need to look +# into libthr private structures. That's sooo evil, but it's only for +# ZFS debugging tools needs. +CFLAGS+= -DWANTS_MUTEX_OWNED +CFLAGS+= -I${.CURDIR}/../../../lib/libpthread/thread +CFLAGS+= -I${.CURDIR}/../../../lib/libpthread/sys +CFLAGS+= -I${.CURDIR}/../../../lib/libthr/arch/${MACHINE_ARCH}/include DPADD= ${LIBPTHREAD} ${LIBZ} LDADD= -lpthread -lz diff --git a/cddl/sbin/zpool/Makefile b/cddl/sbin/zpool/Makefile index 04d6c13..4b24dfb 100644 --- a/cddl/sbin/zpool/Makefile +++ b/cddl/sbin/zpool/Makefile @@ -1,10 +1,11 @@ # $FreeBSD$ -.PATH: ${.CURDIR}/../../../cddl/contrib/opensolaris/cmd/zpool +.PATH: ${.CURDIR}/../../../cddl/contrib/opensolaris/cmd/zpool \ + ${.CURDIR}/../../../sys/cddl/contrib/opensolaris/common/zfs PROG= zpool MAN= zpool.8 -SRCS= zpool_main.c zpool_vdev.c zpool_iter.c zpool_util.c +SRCS= zpool_main.c zpool_vdev.c zpool_iter.c zpool_util.c zfs_comutil.c CFLAGS+= -I${.CURDIR}/../../../cddl/contrib/opensolaris/lib/libzpool/common CFLAGS+= -I${.CURDIR}/../../../cddl/compat/opensolaris/include @@ -15,6 +16,7 @@ CFLAGS+= -I${.CURDIR}/../../../cddl/contrib/opensolaris/lib/libuutil/common CFLAGS+= -I${.CURDIR}/../../../cddl/contrib/opensolaris/lib/libumem/common CFLAGS+= -I${.CURDIR}/../../../cddl/contrib/opensolaris/lib/libzfs/common CFLAGS+= -I${.CURDIR}/../../../cddl/contrib/opensolaris/lib/libnvpair +CFLAGS+= -I${.CURDIR}/../../../sys/cddl/contrib/opensolaris/common/zfs CFLAGS+= -I${.CURDIR}/../../../sys/cddl/contrib/opensolaris/uts/common CFLAGS+= -I${.CURDIR}/../../../sys/cddl/contrib/opensolaris/uts/common/fs/zfs CFLAGS+= -I${.CURDIR}/../../../sys/cddl/contrib/opensolaris/uts/common/sys diff --git a/cddl/usr.bin/Makefile b/cddl/usr.bin/Makefile index 4112e8a..c6b1341 100644 --- a/cddl/usr.bin/Makefile +++ b/cddl/usr.bin/Makefile @@ -7,9 +7,11 @@ SUBDIR= \ ctfdump \ ctfmerge \ sgsmsg \ + ${_zinject} \ ${_ztest} .if ${MK_ZFS} != "no" +_zinject= zinject .if ${MK_LIBTHR} != "no" _ztest= ztest .endif diff --git a/cddl/usr.bin/zinject/Makefile b/cddl/usr.bin/zinject/Makefile new file mode 100644 index 0000000..c8693fe --- /dev/null +++ b/cddl/usr.bin/zinject/Makefile @@ -0,0 +1,25 @@ +# $FreeBSD$ + +.PATH: ${.CURDIR}/../../contrib/opensolaris/cmd/zinject + +PROG= zinject +SRCS= zinject.c translate.c +NO_MAN= + +CFLAGS+= -I${.CURDIR}/../../../sys/cddl/compat/opensolaris +CFLAGS+= -I${.CURDIR}/../../compat/opensolaris/include +CFLAGS+= -I${.CURDIR}/../../compat/opensolaris/lib/libumem +CFLAGS+= -I${.CURDIR}/../../contrib/opensolaris/lib/libzfs/common +CFLAGS+= -I${.CURDIR}/../../contrib/opensolaris/lib/libzpool/common +CFLAGS+= -I${.CURDIR}/../../contrib/opensolaris/lib/libnvpair +CFLAGS+= -I${.CURDIR}/../../../sys/cddl/contrib/opensolaris/uts/common/fs/zfs +CFLAGS+= -I${.CURDIR}/../../../sys/cddl/contrib/opensolaris/uts/common/sys +CFLAGS+= -I${.CURDIR}/../../../sys/cddl/contrib/opensolaris/uts/common +CFLAGS+= -I${.CURDIR}/../../contrib/opensolaris/head +CFLAGS+= -I${.CURDIR}/../../lib/libumem + +DPADD= ${LIBAVL} ${LIBGEOM} ${LIBM} ${LIBNVPAIR} ${LIBUMEM} ${LIBUUTIL} \ + ${LIBZFS} ${LIBZPOOL} ${LIBUUTIL} +LDADD= -lavl -lgeom -lm -lnvpair -lumem -luutil -lzfs -lzpool + +.include <bsd.prog.mk> diff --git a/cddl/usr.bin/ztest/Makefile b/cddl/usr.bin/ztest/Makefile index 0bdded6..cc0bca4 100644 --- a/cddl/usr.bin/ztest/Makefile +++ b/cddl/usr.bin/ztest/Makefile @@ -1,18 +1,20 @@ # $FreeBSD$ -.PATH: ${.CURDIR}/../../../cddl/contrib/opensolaris/cmd/ztest +.PATH: ${.CURDIR}/../..//contrib/opensolaris/cmd/ztest PROG= ztest NO_MAN= +CFLAGS+= -std=c99 + CFLAGS+= -I${.CURDIR}/../../../sys/cddl/compat/opensolaris -CFLAGS+= -I${.CURDIR}/../../../cddl/compat/opensolaris/include -CFLAGS+= -I${.CURDIR}/../../../cddl/compat/opensolaris/lib/libumem -CFLAGS+= -I${.CURDIR}/../../../cddl/contrib/opensolaris/lib/libzpool/common +CFLAGS+= -I${.CURDIR}/../../compat/opensolaris/include +CFLAGS+= -I${.CURDIR}/../../compat/opensolaris/lib/libumem +CFLAGS+= -I${.CURDIR}/../../contrib/opensolaris/lib/libzpool/common CFLAGS+= -I${.CURDIR}/../../../sys/cddl/contrib/opensolaris/uts/common/fs/zfs CFLAGS+= -I${.CURDIR}/../../../sys/cddl/contrib/opensolaris/uts/common/sys CFLAGS+= -I${.CURDIR}/../../../sys/cddl/contrib/opensolaris/uts/common -CFLAGS+= -I${.CURDIR}/../../../cddl/contrib/opensolaris/head +CFLAGS+= -I${.CURDIR}/../../contrib/opensolaris/head CFLAGS+= -I${.CURDIR}/../../lib/libumem DPADD= ${LIBM} ${LIBNVPAIR} ${LIBUMEM} ${LIBZPOOL} \ diff --git a/cddl/usr.sbin/zdb/Makefile b/cddl/usr.sbin/zdb/Makefile index 92a7352..f0c1b22 100644 --- a/cddl/usr.sbin/zdb/Makefile +++ b/cddl/usr.sbin/zdb/Makefile @@ -6,9 +6,14 @@ PROG= zdb MAN= zdb.8 SRCS= zdb.c zdb_il.c +CFLAGS+= -std=c99 + CFLAGS+= -I${.CURDIR}/../../../sys/cddl/compat/opensolaris CFLAGS+= -I${.CURDIR}/../../../cddl/compat/opensolaris/include CFLAGS+= -I${.CURDIR}/../../../cddl/compat/opensolaris/lib/libumem +CFLAGS+= -I${.CURDIR}/../../../cddl/contrib/opensolaris/lib/libnvpair +CFLAGS+= -I${.CURDIR}/../../../cddl/contrib/opensolaris/lib/libuutil/common +CFLAGS+= -I${.CURDIR}/../../../cddl/contrib/opensolaris/lib/libzfs/common CFLAGS+= -I${.CURDIR}/../../../cddl/contrib/opensolaris/lib/libzpool/common CFLAGS+= -I${.CURDIR}/../../../sys/cddl/contrib/opensolaris/uts/common/fs/zfs CFLAGS+= -I${.CURDIR}/../../../sys/cddl/contrib/opensolaris/uts/common @@ -16,8 +21,8 @@ CFLAGS+= -I${.CURDIR}/../../../sys/cddl/contrib/opensolaris/uts/common/sys CFLAGS+= -I${.CURDIR}/../../../cddl/contrib/opensolaris/head CFLAGS+= -I${.CURDIR}/../../lib/libumem -DPADD= ${LIBM} ${LIBNVPAIR} ${LIBUMEM} ${LIBZPOOL} \ - ${LIBPTHREAD} ${LIBZ} ${LIBAVL} -LDADD= -lm -lnvpair -lumem -lzpool -lpthread -lz -lavl +DPADD= ${LIBAVL} ${LIBGEOM} ${LIBM} ${LIBNVPAIR} ${LIBPTHREAD} ${LIBUMEM} \ + ${LIBUUTIL} ${LIBZ} ${LIBZFS} ${LIBZPOOL} +LDADD= -lavl -lgeom -lm -lnvpair -lpthread -lumem -luutil -lz -lzfs -lzpool .include <bsd.prog.mk> |
