/* 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();
err = dmu_objset_own(dataset, DMU_OST_ZFS, B_TRUE, FTAG, &os);
if (err != 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_disown(os, FTAG);
return (0);
}
int
__dbuf_stats_hash_table_data(char *buf, size_t size, dmu_buf_impl_t *db)
{
arc_buf_info_t abi = { 0 };
dmu_object_info_t doi = { 0 };
dnode_t *dn = DB_DNODE(db);
if (db->db_buf)
arc_buf_info(db->db_buf, &abi, zfs_dbuf_state_index);
if (dn)
__dmu_object_info_from_dnode(dn, &doi);
size = snprintf(buf, size - 1,
"%-16s %-8llu %-8lld %-8lld %-8lld %-8llu %-8llu %-5d %-5d %-5lu | "
"%-5d %-5d %-6lld 0x%-6x %-6lu %-8llu %-12llu "
"%-6lu %-6lu %-6lu %-6lu %-6lu %-8llu %-8llu %-8d %-5lu | "
"%-6d %-6d %-8lu %-8lu %-6llu %-6lu %-5lu %-8llu %-8llu\n",
/* dmu_buf_impl_t */
spa_name(dn->dn_objset->os_spa),
(u_longlong_t)dmu_objset_id(db->db_objset),
(longlong_t)db->db.db_object,
(longlong_t)db->db_level,
(longlong_t)db->db_blkid,
(u_longlong_t)db->db.db_offset,
(u_longlong_t)db->db.db_size,
!!dbuf_is_metadata(db),
db->db_state,
(ulong_t)refcount_count(&db->db_holds),
/* arc_buf_info_t */
abi.abi_state_type,
abi.abi_state_contents,
(longlong_t)abi.abi_state_index,
abi.abi_flags,
(ulong_t)abi.abi_datacnt,
(u_longlong_t)abi.abi_size,
(u_longlong_t)abi.abi_access,
(ulong_t)abi.abi_mru_hits,
(ulong_t)abi.abi_mru_ghost_hits,
(ulong_t)abi.abi_mfu_hits,
(ulong_t)abi.abi_mfu_ghost_hits,
(ulong_t)abi.abi_l2arc_hits,
(u_longlong_t)abi.abi_l2arc_dattr,
(u_longlong_t)abi.abi_l2arc_asize,
abi.abi_l2arc_compress,
(ulong_t)abi.abi_holds,
/* dmu_object_info_t */
doi.doi_type,
doi.doi_bonus_type,
(ulong_t)doi.doi_data_block_size,
(ulong_t)doi.doi_metadata_block_size,
(u_longlong_t)doi.doi_bonus_size,
(ulong_t)doi.doi_indirection,
(ulong_t)refcount_count(&dn->dn_holds),
(u_longlong_t)doi.doi_fill_count,
(u_longlong_t)doi.doi_max_offset);
buf[size] = '\0';
return (size);
}
/*
* Check if this super block has a matching objset id.
*/
static int
zfsctl_test_super(struct super_block *sb, void *objsetidp)
{
zfs_sb_t *zsb = sb->s_fs_info;
uint64_t objsetid = *(uint64_t *)objsetidp;
return (dmu_objset_id(zsb->z_os) == objsetid);
}
static void
zil_init_log_chain(zilog_t *zilog, blkptr_t *bp)
{
zio_cksum_t *zc = &bp->blk_cksum;
zc->zc_word[ZIL_ZC_GUID_0] = spa_get_random(-1ULL);
zc->zc_word[ZIL_ZC_GUID_1] = spa_get_random(-1ULL);
zc->zc_word[ZIL_ZC_OBJSET] = dmu_objset_id(zilog->zl_os);
zc->zc_word[ZIL_ZC_SEQ] = 1ULL;
}
static int
#ifdef HAVE_D_REVALIDATE_NAMEIDATA
zpl_revalidate(struct dentry *dentry, struct nameidata *nd)
{
unsigned int flags = (nd ? nd->flags : 0);
#else
zpl_revalidate(struct dentry *dentry, unsigned int flags)
{
#endif /* HAVE_D_REVALIDATE_NAMEIDATA */
zfs_sb_t *zsb = dentry->d_sb->s_fs_info;
int error;
if (flags & LOOKUP_RCU)
return (-ECHILD);
/*
* Automounted snapshots rely on periodic dentry revalidation
* to defer snapshots from being automatically unmounted.
*/
if (zsb->z_issnap) {
if (time_after(jiffies, zsb->z_snap_defer_time +
MAX(zfs_expire_snapshot * HZ / 2, HZ))) {
zsb->z_snap_defer_time = jiffies;
zfsctl_snapshot_unmount_delay(zsb->z_os->os_spa,
dmu_objset_id(zsb->z_os), zfs_expire_snapshot);
}
}
/*
* After a rollback negative dentries created before the rollback
* time must be invalidated. Otherwise they can obscure files which
* are only present in the rolled back dataset.
*/
if (dentry->d_inode == NULL) {
spin_lock(&dentry->d_lock);
error = time_before(dentry->d_time, zsb->z_rollback_time);
spin_unlock(&dentry->d_lock);
if (error)
return (0);
}
/*
* The dentry may reference a stale inode if a mounted file system
* was rolled back to a point in time where the object didn't exist.
*/
if (dentry->d_inode && ITOZ(dentry->d_inode)->z_is_stale)
return (0);
return (1);
}
int
zfs_set_version(const char *name, uint64_t newvers)
{
int error;
objset_t *os;
dmu_tx_t *tx;
uint64_t curvers;
/*
* XXX for now, require that the filesystem be unmounted. Would
* be nice to find the zfsvfs_t and just update that if
* possible.
*/
if (newvers < ZPL_VERSION_INITIAL || newvers > ZPL_VERSION)
return (EINVAL);
error = dmu_objset_open(name, DMU_OST_ZFS, DS_MODE_PRIMARY, &os);
if (error)
return (error);
error = zap_lookup(os, MASTER_NODE_OBJ, ZPL_VERSION_STR,
8, 1, &curvers);
if (error)
goto out;
if (newvers < curvers) {
error = EINVAL;
goto out;
}
tx = dmu_tx_create(os);
dmu_tx_hold_zap(tx, MASTER_NODE_OBJ, 0, ZPL_VERSION_STR);
error = dmu_tx_assign(tx, TXG_WAIT);
if (error) {
dmu_tx_abort(tx);
goto out;
}
error = zap_update(os, MASTER_NODE_OBJ, ZPL_VERSION_STR, 8, 1,
&newvers, tx);
spa_history_internal_log(LOG_DS_UPGRADE,
dmu_objset_spa(os), tx, CRED(),
"oldver=%llu newver=%llu dataset = %llu", curvers, newvers,
dmu_objset_id(os));
dmu_tx_commit(tx);
out:
dmu_objset_close(os);
return (error);
}
/*
* Read in the data for the dmu_sync()ed block, and change the log
* record to write this whole block.
*/
void
zil_get_replay_data(zilog_t *zilog, lr_write_t *lr)
{
blkptr_t *wbp = &lr->lr_blkptr;
char *wbuf = (char *)(lr + 1); /* data follows lr_write_t */
uint64_t blksz;
if (BP_IS_HOLE(wbp)) { /* compressed to a hole */
blksz = BP_GET_LSIZE(&lr->lr_blkptr);
/*
* If the blksz is zero then we must be replaying a log
* from an version prior to setting the blksize of null blocks.
* So we just zero the actual write size reqeusted.
*/
if (blksz == 0) {
bzero(wbuf, lr->lr_length);
return;
}
bzero(wbuf, blksz);
} else {
/*
* A subsequent write may have overwritten this block, in which
* case wbp may have been been freed and reallocated, and our
* read of wbp may fail with a checksum error. We can safely
* ignore this because the later write will provide the
* correct data.
*/
zbookmark_t zb;
zb.zb_objset = dmu_objset_id(zilog->zl_os);
zb.zb_object = lr->lr_foid;
zb.zb_level = 0;
zb.zb_blkid = -1; /* unknown */
blksz = BP_GET_LSIZE(&lr->lr_blkptr);
(void) zio_wait(zio_read(NULL, zilog->zl_spa, wbp, wbuf, blksz,
NULL, NULL, ZIO_PRIORITY_SYNC_READ,
ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE, &zb));
}
lr->lr_offset -= lr->lr_offset % blksz;
lr->lr_length = blksz;
}
/* ARGSUSED */
static void
zil_prt_rec_write(zilog_t *zilog, int txtype, lr_write_t *lr)
{
char *data, *dlimit;
blkptr_t *bp = &lr->lr_blkptr;
zbookmark_phys_t zb;
char buf[SPA_MAXBLOCKSIZE];
int verbose = MAX(dump_opt['d'], dump_opt['i']);
int error;
(void) printf("%sfoid %llu, offset %llx, length %llx\n", prefix,
(u_longlong_t)lr->lr_foid, (u_longlong_t)lr->lr_offset,
(u_longlong_t)lr->lr_length);
if (txtype == TX_WRITE2 || verbose < 5)
return;
if (lr->lr_common.lrc_reclen == sizeof (lr_write_t)) {
(void) printf("%shas blkptr, %s\n", prefix,
!BP_IS_HOLE(bp) &&
bp->blk_birth >= spa_first_txg(zilog->zl_spa) ?
"will claim" : "won't claim");
print_log_bp(bp, prefix);
if (BP_IS_HOLE(bp)) {
(void) printf("\t\t\tLSIZE 0x%llx\n",
(u_longlong_t)BP_GET_LSIZE(bp));
bzero(buf, sizeof (buf));
(void) printf("%s<hole>\n", prefix);
return;
}
if (bp->blk_birth < zilog->zl_header->zh_claim_txg) {
(void) printf("%s<block already committed>\n", prefix);
return;
}
SET_BOOKMARK(&zb, dmu_objset_id(zilog->zl_os),
lr->lr_foid, ZB_ZIL_LEVEL,
lr->lr_offset / BP_GET_LSIZE(bp));
error = zio_wait(zio_read(NULL, zilog->zl_spa,
bp, buf, BP_GET_LSIZE(bp), NULL, NULL,
ZIO_PRIORITY_SYNC_READ, ZIO_FLAG_CANFAIL, &zb));
if (error)
return;
data = buf;
} else {
data = (char *)(lr + 1);
}
dlimit = data + MIN(lr->lr_length,
(verbose < 6 ? 20 : SPA_MAXBLOCKSIZE));
(void) printf("%s", prefix);
while (data < dlimit) {
if (isprint(*data))
(void) printf("%c ", *data);
else
(void) printf("%2hhX", *data);
data++;
}
(void) printf("\n");
}
/* ARGSUSED */
static int
zfsctl_snapdir_lookup(vnode_t *dvp, char *nm, vnode_t **vpp, pathname_t *pnp,
int flags, vnode_t *rdir, cred_t *cr, caller_context_t *ct,
int *direntflags, pathname_t *realpnp)
{
zfsctl_snapdir_t *sdp = dvp->v_data;
objset_t *snap;
char snapname[MAXNAMELEN];
char real[MAXNAMELEN];
char *mountpoint;
zfs_snapentry_t *sep, search;
struct mounta margs;
vfs_t *vfsp;
size_t mountpoint_len;
avl_index_t where;
zfsvfs_t *zfsvfs = dvp->v_vfsp->vfs_data;
int err;
/*
* No extended attributes allowed under .zfs
*/
if (flags & LOOKUP_XATTR)
return (EINVAL);
ASSERT(dvp->v_type == VDIR);
/*
* If we get a recursive call, that means we got called
* from the domount() code while it was trying to look up the
* spec (which looks like a local path for zfs). We need to
* add some flag to domount() to tell it not to do this lookup.
*/
if (MUTEX_HELD(&sdp->sd_lock))
return (ENOENT);
ZFS_ENTER(zfsvfs);
if (gfs_lookup_dot(vpp, dvp, zfsvfs->z_ctldir, nm) == 0) {
ZFS_EXIT(zfsvfs);
return (0);
}
if (flags & FIGNORECASE) {
boolean_t conflict = B_FALSE;
err = dmu_snapshot_realname(zfsvfs->z_os, nm, real,
MAXNAMELEN, &conflict);
if (err == 0) {
nm = real;
} else if (err != ENOTSUP) {
ZFS_EXIT(zfsvfs);
return (err);
}
if (realpnp)
(void) strlcpy(realpnp->pn_buf, nm,
realpnp->pn_bufsize);
if (conflict && direntflags)
*direntflags = ED_CASE_CONFLICT;
}
mutex_enter(&sdp->sd_lock);
search.se_name = (char *)nm;
if ((sep = avl_find(&sdp->sd_snaps, &search, &where)) != NULL) {
*vpp = sep->se_root;
VN_HOLD(*vpp);
err = traverse(vpp);
if (err) {
VN_RELE(*vpp);
*vpp = NULL;
} else if (*vpp == sep->se_root) {
/*
* The snapshot was unmounted behind our backs,
* try to remount it.
*/
goto domount;
} else {
/*
* VROOT was set during the traverse call. We need
* to clear it since we're pretending to be part
* of our parent's vfs.
*/
(*vpp)->v_flag &= ~VROOT;
}
mutex_exit(&sdp->sd_lock);
ZFS_EXIT(zfsvfs);
return (err);
}
/*
* The requested snapshot is not currently mounted, look it up.
*/
err = zfsctl_snapshot_zname(dvp, nm, MAXNAMELEN, snapname);
if (err) {
mutex_exit(&sdp->sd_lock);
ZFS_EXIT(zfsvfs);
/*
* handle "ls *" or "?" in a graceful manner,
* forcing EILSEQ to ENOENT.
* Since shell ultimately passes "*" or "?" as name to lookup
*/
return (err == EILSEQ ? ENOENT : err);
}
if (dmu_objset_hold(snapname, FTAG, &snap) != 0) {
mutex_exit(&sdp->sd_lock);
ZFS_EXIT(zfsvfs);
return (ENOENT);
}
sep = kmem_alloc(sizeof (zfs_snapentry_t), KM_SLEEP);
sep->se_name = kmem_alloc(strlen(nm) + 1, KM_SLEEP);
(void) strcpy(sep->se_name, nm);
*vpp = sep->se_root = zfsctl_snapshot_mknode(dvp, dmu_objset_id(snap));
avl_insert(&sdp->sd_snaps, sep, where);
dmu_objset_rele(snap, FTAG);
domount:
mountpoint_len = strlen(refstr_value(dvp->v_vfsp->vfs_mntpt)) +
strlen("/.zfs/snapshot/") + strlen(nm) + 1;
mountpoint = kmem_alloc(mountpoint_len, KM_SLEEP);
(void) snprintf(mountpoint, mountpoint_len, "%s/.zfs/snapshot/%s",
refstr_value(dvp->v_vfsp->vfs_mntpt), nm);
margs.spec = snapname;
margs.dir = mountpoint;
margs.flags = MS_SYSSPACE | MS_NOMNTTAB;
margs.fstype = "zfs";
margs.dataptr = NULL;
margs.datalen = 0;
margs.optptr = NULL;
margs.optlen = 0;
err = domount("zfs", &margs, *vpp, kcred, &vfsp);
kmem_free(mountpoint, mountpoint_len);
if (err == 0) {
/*
* Return the mounted root rather than the covered mount point.
* Takes the GFS vnode at .zfs/snapshot/<snapname> and returns
* the ZFS vnode mounted on top of the GFS node. This ZFS
* vnode is the root of the newly created vfsp.
*/
VFS_RELE(vfsp);
err = traverse(vpp);
}
if (err == 0) {
/*
* Fix up the root vnode mounted on .zfs/snapshot/<snapname>.
*
* This is where we lie about our v_vfsp in order to
* make .zfs/snapshot/<snapname> accessible over NFS
* without requiring manual mounts of <snapname>.
*/
ASSERT(VTOZ(*vpp)->z_zfsvfs != zfsvfs);
VTOZ(*vpp)->z_zfsvfs->z_parent = zfsvfs;
(*vpp)->v_vfsp = zfsvfs->z_vfs;
(*vpp)->v_flag &= ~VROOT;
}
mutex_exit(&sdp->sd_lock);
ZFS_EXIT(zfsvfs);
/*
* If we had an error, drop our hold on the vnode and
* zfsctl_snapshot_inactive() will clean up.
*/
if (err) {
VN_RELE(*vpp);
*vpp = NULL;
}
return (err);
}
void
dataset_kstats_create(dataset_kstats_t *dk, objset_t *objset)
{
/*
* There should not be anything wrong with having kstats for
* snapshots. Since we are not sure how useful they would be
* though nor how much their memory overhead would matter in
* a filesystem with many snapshots, we skip them for now.
*/
if (dmu_objset_is_snapshot(objset))
return;
/*
* At the time of this writing, KSTAT_STRLEN is 255 in Linux,
* and the spa_name can theoretically be up to 256 characters.
* In reality though the spa_name can be 240 characters max
* [see origin directory name check in pool_namecheck()]. Thus,
* the naming scheme for the module name below should not cause
* any truncations. In the event that a truncation does happen
* though, due to some future change, we silently skip creating
* the kstat and log the event.
*/
char kstat_module_name[KSTAT_STRLEN];
int n = snprintf(kstat_module_name, sizeof (kstat_module_name),
"zfs/%s", spa_name(dmu_objset_spa(objset)));
if (n < 0) {
zfs_dbgmsg("failed to create dataset kstat for objset %lld: "
" snprintf() for kstat module name returned %d",
(unsigned long long)dmu_objset_id(objset), n);
return;
} else if (n >= KSTAT_STRLEN) {
zfs_dbgmsg("failed to create dataset kstat for objset %lld: "
"kstat module name length (%d) exceeds limit (%d)",
(unsigned long long)dmu_objset_id(objset),
n, KSTAT_STRLEN);
return;
}
char kstat_name[KSTAT_STRLEN];
n = snprintf(kstat_name, sizeof (kstat_name), "objset-0x%llx",
(unsigned long long)dmu_objset_id(objset));
if (n < 0) {
zfs_dbgmsg("failed to create dataset kstat for objset %lld: "
" snprintf() for kstat name returned %d",
(unsigned long long)dmu_objset_id(objset), n);
return;
}
ASSERT3U(n, <, KSTAT_STRLEN);
kstat_t *kstat = kstat_create(kstat_module_name, 0, kstat_name,
"dataset", KSTAT_TYPE_NAMED,
sizeof (empty_dataset_kstats) / sizeof (kstat_named_t),
KSTAT_FLAG_VIRTUAL);
if (kstat == NULL)
return;
dataset_kstat_values_t *dk_kstats =
kmem_alloc(sizeof (empty_dataset_kstats), KM_SLEEP);
bcopy(&empty_dataset_kstats, dk_kstats,
sizeof (empty_dataset_kstats));
char *ds_name = kmem_zalloc(ZFS_MAX_DATASET_NAME_LEN, KM_SLEEP);
dsl_dataset_name(objset->os_dsl_dataset, ds_name);
KSTAT_NAMED_STR_PTR(&dk_kstats->dkv_ds_name) = ds_name;
KSTAT_NAMED_STR_BUFLEN(&dk_kstats->dkv_ds_name) =
ZFS_MAX_DATASET_NAME_LEN;
kstat->ks_data = dk_kstats;
kstat->ks_update = dataset_kstats_update;
kstat->ks_private = dk;
kstat_install(kstat);
dk->dk_kstats = kstat;
aggsum_init(&dk->dk_aggsums.das_writes, 0);
aggsum_init(&dk->dk_aggsums.das_nwritten, 0);
aggsum_init(&dk->dk_aggsums.das_reads, 0);
aggsum_init(&dk->dk_aggsums.das_nread, 0);
aggsum_init(&dk->dk_aggsums.das_nunlinks, 0);
aggsum_init(&dk->dk_aggsums.das_nunlinked, 0);
}
/*
* 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 stat
* buffer, which we need later to get the object ID.
*/
static int
parse_pathname(const char *inpath, char *dataset, char *relpath,
struct stat *statbuf)
{
struct extmnttab mp;
FILE *fp;
int match;
const char *rel;
char fullpath[MAXPATHLEN];
compress_slashes(inpath, fullpath);
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 (stat(fullpath, statbuf) != 0) {
(void) fprintf(stderr, "cannot open '%s': %s\n",
fullpath, strerror(errno));
return (-1);
}
#ifdef HAVE_SETMNTENT
if ((fp = setmntent(MNTTAB, "r")) == NULL) {
#else
if ((fp = fopen(MNTTAB, "r")) == NULL) {
#endif
(void) fprintf(stderr, "cannot open /etc/mtab\n");
return (-1);
}
match = 0;
while (getextmntent(fp, &mp, sizeof (mp)) == 0) {
if (makedev(mp.mnt_major, mp.mnt_minor) == statbuf->st_dev) {
match = 1;
break;
}
}
if (!match) {
(void) fprintf(stderr, "cannot find mountpoint for '%s'\n",
fullpath);
return (-1);
}
if (strcmp(mp.mnt_fstype, MNTTYPE_ZFS) != 0) {
(void) fprintf(stderr, "invalid path '%s': not a ZFS "
"filesystem\n", fullpath);
return (-1);
}
if (strncmp(fullpath, mp.mnt_mountp, strlen(mp.mnt_mountp)) != 0) {
(void) fprintf(stderr, "invalid path '%s': mountpoint "
"doesn't match path\n", fullpath);
return (-1);
}
(void) strcpy(dataset, mp.mnt_special);
rel = fullpath + strlen(mp.mnt_mountp);
if (rel[0] == '/')
rel++;
(void) strcpy(relpath, rel);
return (0);
}
#endif
//From FreeBSD
static int
parse_pathname(const char *inpath, char *dataset, char *relpath,
struct stat *statbuf)
{
struct statfs sfs;
const char *rel;
char fullpath[MAXPATHLEN];
compress_slashes(inpath, fullpath);
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 (stat(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 stat *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();
err = dmu_objset_own(dataset, DMU_OST_ZFS, B_TRUE, B_FALSE, FTAG, &os);
if (err != 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_disown(os, B_FALSE, FTAG);
return (0);
}
static void
zil_replay_log_record(zilog_t *zilog, lr_t *lr, void *zra, uint64_t claim_txg)
{
zil_replay_arg_t *zr = zra;
const zil_header_t *zh = zilog->zl_header;
uint64_t reclen = lr->lrc_reclen;
uint64_t txtype = lr->lrc_txtype;
char *name;
int pass, error;
if (!zilog->zl_replay) /* giving up */
return;
if (lr->lrc_txg < claim_txg) /* already committed */
return;
if (lr->lrc_seq <= zh->zh_replay_seq) /* already replayed */
return;
/* Strip case-insensitive bit, still present in log record */
txtype &= ~TX_CI;
if (txtype == 0 || txtype >= TX_MAX_TYPE) {
error = EINVAL;
goto bad;
}
/*
* Make a copy of the data so we can revise and extend it.
*/
bcopy(lr, zr->zr_lrbuf, reclen);
/*
* The log block containing this lr may have been byteswapped
* so that we can easily examine common fields like lrc_txtype.
* However, the log is a mix of different data types, and only the
* replay vectors know how to byteswap their records. Therefore, if
* the lr was byteswapped, undo it before invoking the replay vector.
*/
if (zr->zr_byteswap)
byteswap_uint64_array(zr->zr_lrbuf, reclen);
/*
* If this is a TX_WRITE with a blkptr, suck in the data.
*/
if (txtype == TX_WRITE && reclen == sizeof (lr_write_t)) {
lr_write_t *lrw = (lr_write_t *)lr;
blkptr_t *wbp = &lrw->lr_blkptr;
uint64_t wlen = lrw->lr_length;
char *wbuf = zr->zr_lrbuf + reclen;
if (BP_IS_HOLE(wbp)) { /* compressed to a hole */
bzero(wbuf, wlen);
} else {
/*
* A subsequent write may have overwritten this block,
* in which case wbp may have been been freed and
* reallocated, and our read of wbp may fail with a
* checksum error. We can safely ignore this because
* the later write will provide the correct data.
*/
zbookmark_t zb;
zb.zb_objset = dmu_objset_id(zilog->zl_os);
zb.zb_object = lrw->lr_foid;
zb.zb_level = -1;
zb.zb_blkid = lrw->lr_offset / BP_GET_LSIZE(wbp);
(void) zio_wait(zio_read(NULL, zilog->zl_spa,
wbp, wbuf, BP_GET_LSIZE(wbp), NULL, NULL,
ZIO_PRIORITY_SYNC_READ,
ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE, &zb));
(void) memmove(wbuf, wbuf + lrw->lr_blkoff, wlen);
}
}
/*
* We must now do two things atomically: replay this log record,
* and update the log header sequence number to reflect the fact that
* we did so. At the end of each replay function the sequence number
* is updated if we are in replay mode.
*/
for (pass = 1; pass <= 2; pass++) {
zilog->zl_replaying_seq = lr->lrc_seq;
/* Only byteswap (if needed) on the 1st pass. */
error = zr->zr_replay[txtype](zr->zr_arg, zr->zr_lrbuf,
zr->zr_byteswap && pass == 1);
if (!error)
return;
/*
* The DMU's dnode layer doesn't see removes until the txg
* commits, so a subsequent claim can spuriously fail with
* EEXIST. So if we receive any error we try syncing out
* any removes then retry the transaction.
*/
if (pass == 1)
txg_wait_synced(spa_get_dsl(zilog->zl_spa), 0);
}
bad:
ASSERT(error);
name = kmem_alloc(MAXNAMELEN, KM_SLEEP);
dmu_objset_name(zr->zr_os, name);
cmn_err(CE_WARN, "ZFS replay transaction error %d, "
"dataset %s, seq 0x%llx, txtype %llu %s\n",
error, name, (u_longlong_t)lr->lrc_seq, (u_longlong_t)txtype,
(lr->lrc_txtype & TX_CI) ? "CI" : "");
zilog->zl_replay = B_FALSE;
kmem_free(name, MAXNAMELEN);
}
ASSERT3U(spa_version(dmu_objset_spa(zsb->z_os)), >=, SPA_VERSION_SA); sa_obj = zap_create(os, DMU_OT_SA_MASTER_NODE, DMU_OT_NONE, 0, tx); error = zap_add(os, MASTER_NODE_OBJ, ZFS_SA_ATTRS, 8, 1, &sa_obj, tx); ASSERT0(error); VERIFY(0 == sa_set_sa_object(os, sa_obj)); sa_register_update_callback(os, zfs_sa_upgrade); } spa_history_log_internal(LOG_DS_UPGRADE, dmu_objset_spa(os), tx, "oldver=%llu newver=%llu dataset = %llu", zsb->z_version, newvers, dmu_objset_id(os)); dmu_tx_commit(tx); zsb->z_version = newvers; if (zsb->z_version >= ZPL_VERSION_FUID) zfs_set_fuid_feature(zsb); return (0); } EXPORT_SYMBOL(zfs_set_version); /* * Read a property stored within the master node. */
