int
dmu_objset_destroy(const char *name)
{
objset_t *os;
int error;
/*
* If it looks like we'll be able to destroy it, and there's
* an unplayed replay log sitting around, destroy the log.
* It would be nicer to do this in dsl_dataset_destroy_sync(),
* but the replay log objset is modified in open context.
*/
error = dmu_objset_open(name, DMU_OST_ANY,
DS_MODE_OWNER|DS_MODE_READONLY|DS_MODE_INCONSISTENT, &os);
if (error == 0) {
dsl_dataset_t *ds = os->os->os_dsl_dataset;
zil_destroy(dmu_objset_zil(os), B_FALSE);
error = dsl_dataset_destroy(ds, os);
/*
* dsl_dataset_destroy() closes the ds.
*/
kmem_free(os, sizeof (objset_t));
}
return (error);
}
/*
* If this dataset has a non-empty intent log, replay it and destroy it.
*/
void
zil_replay(objset_t *os, void *arg, zil_replay_func_t *replay_func[TX_MAX_TYPE])
{
zilog_t *zilog = dmu_objset_zil(os);
const zil_header_t *zh = zilog->zl_header;
zil_replay_arg_t zr;
if ((zh->zh_flags & ZIL_REPLAY_NEEDED) == 0) {
zil_destroy(zilog, B_TRUE);
return;
}
zr.zr_os = os;
zr.zr_replay = replay_func;
zr.zr_arg = arg;
zr.zr_byteswap = BP_SHOULD_BYTESWAP(&zh->zh_log);
/* XXX: Changed to use vmem_alloc instead of kmem_alloc for
* large allocation size (I think this is safe here).
*/
zr.zr_lrbuf = vmem_alloc(2 * SPA_MAXBLOCKSIZE, KM_SLEEP);
/*
* Wait for in-progress removes to sync before starting replay.
*/
txg_wait_synced(zilog->zl_dmu_pool, 0);
zilog->zl_replay = B_TRUE;
zilog->zl_replay_time = lbolt;
ASSERT(zilog->zl_replay_blks == 0);
(void) zil_parse(zilog, zil_incr_blks, zil_replay_log_record, &zr,
zh->zh_claim_txg);
vmem_free(zr.zr_lrbuf, 2 * SPA_MAXBLOCKSIZE);
zil_destroy(zilog, B_FALSE);
txg_wait_synced(zilog->zl_dmu_pool, zilog->zl_destroy_txg);
zilog->zl_replay = B_FALSE;
}
/*
* If this dataset has a non-empty intent log, replay it and destroy it.
*/
void
zil_replay(objset_t *os, void *arg, zil_replay_func_t *replay_func[TX_MAX_TYPE])
{
zilog_t *zilog = dmu_objset_zil(os);
const zil_header_t *zh = zilog->zl_header;
zil_replay_arg_t zr;
if ((zh->zh_flags & ZIL_REPLAY_NEEDED) == 0) {
zil_destroy(zilog, B_TRUE);
return;
}
//printf("ZFS: Replaying ZIL on %s...\n", os->os->os_spa->spa_name);
zr.zr_os = os;
zr.zr_replay = replay_func;
zr.zr_arg = arg;
zr.zr_byteswap = BP_SHOULD_BYTESWAP(&zh->zh_log);
zr.zr_lrbuf = kmem_alloc(2 * SPA_MAXBLOCKSIZE, KM_SLEEP);
/*
* Wait for in-progress removes to sync before starting replay.
*/
txg_wait_synced(zilog->zl_dmu_pool, 0);
zilog->zl_replay = B_TRUE;
zilog->zl_replay_time = LBOLT;
ASSERT(zilog->zl_replay_blks == 0);
(void) zil_parse(zilog, zil_incr_blks, zil_replay_log_record, &zr,
zh->zh_claim_txg);
kmem_free(zr.zr_lrbuf, 2 * SPA_MAXBLOCKSIZE);
zil_destroy(zilog, B_FALSE);
txg_wait_synced(zilog->zl_dmu_pool, zilog->zl_destroy_txg);
zilog->zl_replay = B_FALSE;
//printf("ZFS: Replay of ZIL on %s finished.\n", os->os->os_spa->spa_name);
}
/*
* Suspend an intent log. While in suspended mode, we still honor
* synchronous semantics, but we rely on txg_wait_synced() to do it.
* We suspend the log briefly when taking a snapshot so that the snapshot
* contains all the data it's supposed to, and has an empty intent log.
*/
int
zil_suspend(zilog_t *zilog)
{
const zil_header_t *zh = zilog->zl_header;
mutex_enter(&zilog->zl_lock);
if (zh->zh_flags & ZIL_REPLAY_NEEDED) { /* unplayed log */
mutex_exit(&zilog->zl_lock);
return (EBUSY);
}
if (zilog->zl_suspend++ != 0) {
/*
* Someone else already began a suspend.
* Just wait for them to finish.
*/
while (zilog->zl_suspending)
cv_wait(&zilog->zl_cv_suspend, &zilog->zl_lock);
mutex_exit(&zilog->zl_lock);
return (0);
}
zilog->zl_suspending = B_TRUE;
mutex_exit(&zilog->zl_lock);
zil_commit(zilog, UINT64_MAX, 0);
/*
* Wait for any in-flight log writes to complete.
*/
mutex_enter(&zilog->zl_lock);
while (zilog->zl_writer)
cv_wait(&zilog->zl_cv_writer, &zilog->zl_lock);
mutex_exit(&zilog->zl_lock);
zil_destroy(zilog, B_FALSE);
mutex_enter(&zilog->zl_lock);
zilog->zl_suspending = B_FALSE;
cv_broadcast(&zilog->zl_cv_suspend);
mutex_exit(&zilog->zl_lock);
return (0);
}
/*
* Create a block device minor node and setup the linkage between it
* and the specified volume. Once this function returns the block
* device is live and ready for use.
*/
static int
zvol_create_minor_impl(const char *name)
{
zvol_state_t *zv;
objset_t *os;
dmu_object_info_t *doi;
uint64_t volsize;
uint64_t len;
unsigned minor = 0;
int error = 0;
mutex_enter(&zvol_state_lock);
zv = zvol_find_by_name(name);
if (zv) {
error = SET_ERROR(EEXIST);
goto out;
}
doi = kmem_alloc(sizeof (dmu_object_info_t), KM_SLEEP);
error = dmu_objset_own(name, DMU_OST_ZVOL, B_TRUE, zvol_tag, &os);
if (error)
goto out_doi;
error = dmu_object_info(os, ZVOL_OBJ, doi);
if (error)
goto out_dmu_objset_disown;
error = zap_lookup(os, ZVOL_ZAP_OBJ, "size", 8, 1, &volsize);
if (error)
goto out_dmu_objset_disown;
error = zvol_find_minor(&minor);
if (error)
goto out_dmu_objset_disown;
zv = zvol_alloc(MKDEV(zvol_major, minor), name);
if (zv == NULL) {
error = SET_ERROR(EAGAIN);
goto out_dmu_objset_disown;
}
if (dmu_objset_is_snapshot(os))
zv->zv_flags |= ZVOL_RDONLY;
zv->zv_volblocksize = doi->doi_data_block_size;
zv->zv_volsize = volsize;
zv->zv_objset = os;
set_capacity(zv->zv_disk, zv->zv_volsize >> 9);
blk_queue_max_hw_sectors(zv->zv_queue, (DMU_MAX_ACCESS / 4) >> 9);
blk_queue_max_segments(zv->zv_queue, UINT16_MAX);
blk_queue_max_segment_size(zv->zv_queue, UINT_MAX);
blk_queue_physical_block_size(zv->zv_queue, zv->zv_volblocksize);
blk_queue_io_opt(zv->zv_queue, zv->zv_volblocksize);
blk_queue_max_discard_sectors(zv->zv_queue,
(zvol_max_discard_blocks * zv->zv_volblocksize) >> 9);
blk_queue_discard_granularity(zv->zv_queue, zv->zv_volblocksize);
queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, zv->zv_queue);
#ifdef QUEUE_FLAG_NONROT
queue_flag_set_unlocked(QUEUE_FLAG_NONROT, zv->zv_queue);
#endif
#ifdef QUEUE_FLAG_ADD_RANDOM
queue_flag_clear_unlocked(QUEUE_FLAG_ADD_RANDOM, zv->zv_queue);
#endif
if (spa_writeable(dmu_objset_spa(os))) {
if (zil_replay_disable)
zil_destroy(dmu_objset_zil(os), B_FALSE);
else
zil_replay(os, zv, zvol_replay_vector);
}
/*
* When udev detects the addition of the device it will immediately
* invoke blkid(8) to determine the type of content on the device.
* Prefetching the blocks commonly scanned by blkid(8) will speed
* up this process.
*/
len = MIN(MAX(zvol_prefetch_bytes, 0), SPA_MAXBLOCKSIZE);
if (len > 0) {
dmu_prefetch(os, ZVOL_OBJ, 0, 0, len, ZIO_PRIORITY_SYNC_READ);
dmu_prefetch(os, ZVOL_OBJ, 0, volsize - len, len,
ZIO_PRIORITY_SYNC_READ);
}
zv->zv_objset = NULL;
out_dmu_objset_disown:
dmu_objset_disown(os, zvol_tag);
out_doi:
kmem_free(doi, sizeof (dmu_object_info_t));
out:
if (error == 0) {
zvol_insert(zv);
/*
* Drop the lock to prevent deadlock with sys_open() ->
* zvol_open(), which first takes bd_disk->bd_mutex and then
* takes zvol_state_lock, whereas this code path first takes
* zvol_state_lock, and then takes bd_disk->bd_mutex.
*/
mutex_exit(&zvol_state_lock);
add_disk(zv->zv_disk);
} else {
mutex_exit(&zvol_state_lock);
}
return (SET_ERROR(error));
}
static int
__zvol_create_minor(const char *name)
{
zvol_state_t *zv;
objset_t *os;
dmu_object_info_t *doi;
uint64_t volsize;
unsigned minor = 0;
int error = 0;
ASSERT(MUTEX_HELD(&zvol_state_lock));
zv = zvol_find_by_name(name);
if (zv) {
error = EEXIST;
goto out;
}
doi = kmem_alloc(sizeof(dmu_object_info_t), KM_SLEEP);
error = dmu_objset_own(name, DMU_OST_ZVOL, B_TRUE, zvol_tag, &os);
if (error)
goto out_doi;
/* Make sure we have the key loaded if we need one. */
error = dsl_crypto_key_inherit(name);
if (error != 0 && error != EEXIST)
goto out_dmu_objset_disown;
error = dmu_object_info(os, ZVOL_OBJ, doi);
if (error)
goto out_dmu_objset_disown;
error = zap_lookup(os, ZVOL_ZAP_OBJ, "size", 8, 1, &volsize);
if (error)
goto out_dmu_objset_disown;
error = zvol_find_minor(&minor);
if (error)
goto out_dmu_objset_disown;
zv = zvol_alloc(MKDEV(zvol_major, minor), name);
if (zv == NULL) {
error = EAGAIN;
goto out_dmu_objset_disown;
}
if (dmu_objset_is_snapshot(os))
zv->zv_flags |= ZVOL_RDONLY;
zv->zv_volblocksize = doi->doi_data_block_size;
zv->zv_volsize = volsize;
zv->zv_objset = os;
set_capacity(zv->zv_disk, zv->zv_volsize >> 9);
blk_queue_max_hw_sectors(zv->zv_queue, UINT_MAX);
blk_queue_max_segments(zv->zv_queue, UINT16_MAX);
blk_queue_max_segment_size(zv->zv_queue, UINT_MAX);
blk_queue_physical_block_size(zv->zv_queue, zv->zv_volblocksize);
blk_queue_io_opt(zv->zv_queue, zv->zv_volblocksize);
#ifdef HAVE_BLK_QUEUE_DISCARD
blk_queue_max_discard_sectors(zv->zv_queue,
(zvol_max_discard_blocks * zv->zv_volblocksize) >> 9);
blk_queue_discard_granularity(zv->zv_queue, zv->zv_volblocksize);
queue_flag_set_unlocked(QUEUE_FLAG_DISCARD, zv->zv_queue);
#endif
#ifdef HAVE_BLK_QUEUE_NONROT
queue_flag_set_unlocked(QUEUE_FLAG_NONROT, zv->zv_queue);
#endif
if (zil_replay_disable)
zil_destroy(dmu_objset_zil(os), B_FALSE);
else
zil_replay(os, zv, zvol_replay_vector);
out_dmu_objset_disown:
dmu_objset_disown(os, zvol_tag);
zv->zv_objset = NULL;
out_doi:
kmem_free(doi, sizeof(dmu_object_info_t));
out:
if (error == 0) {
zvol_insert(zv);
add_disk(zv->zv_disk);
}
return (error);
}
int
zfs_sb_setup(zfs_sb_t *zsb, boolean_t mounting)
{
int error;
error = zfs_register_callbacks(zsb);
if (error)
return (error);
/*
* Set the objset user_ptr to track its zsb.
*/
mutex_enter(&zsb->z_os->os_user_ptr_lock);
dmu_objset_set_user(zsb->z_os, zsb);
mutex_exit(&zsb->z_os->os_user_ptr_lock);
zsb->z_log = zil_open(zsb->z_os, zfs_get_data);
/*
* If we are not mounting (ie: online recv), then we don't
* have to worry about replaying the log as we blocked all
* operations out since we closed the ZIL.
*/
if (mounting) {
boolean_t readonly;
/*
* During replay we remove the read only flag to
* allow replays to succeed.
*/
readonly = zfs_is_readonly(zsb);
if (readonly != 0)
readonly_changed_cb(zsb, B_FALSE);
else
zfs_unlinked_drain(zsb);
/*
* Parse and replay the intent log.
*
* Because of ziltest, this must be done after
* zfs_unlinked_drain(). (Further note: ziltest
* doesn't use readonly mounts, where
* zfs_unlinked_drain() isn't called.) This is because
* ziltest causes spa_sync() to think it's committed,
* but actually it is not, so the intent log contains
* many txg's worth of changes.
*
* In particular, if object N is in the unlinked set in
* the last txg to actually sync, then it could be
* actually freed in a later txg and then reallocated
* in a yet later txg. This would write a "create
* object N" record to the intent log. Normally, this
* would be fine because the spa_sync() would have
* written out the fact that object N is free, before
* we could write the "create object N" intent log
* record.
*
* But when we are in ziltest mode, we advance the "open
* txg" without actually spa_sync()-ing the changes to
* disk. So we would see that object N is still
* allocated and in the unlinked set, and there is an
* intent log record saying to allocate it.
*/
if (spa_writeable(dmu_objset_spa(zsb->z_os))) {
if (zil_replay_disable) {
zil_destroy(zsb->z_log, B_FALSE);
} else {
zsb->z_replay = B_TRUE;
zil_replay(zsb->z_os, zsb,
zfs_replay_vector);
zsb->z_replay = B_FALSE;
}
}
/* restore readonly bit */
if (readonly != 0)
readonly_changed_cb(zsb, B_TRUE);
}
return (0);
}
static int
__zvol_create_minor(const char *name)
{
zvol_state_t *zv;
objset_t *os;
dmu_object_info_t *doi;
uint64_t volsize;
unsigned minor = 0;
int error = 0;
ASSERT(MUTEX_HELD(&zvol_state_lock));
zv = zvol_find_by_name(name);
if (zv) {
error = EEXIST;
goto out;
}
doi = kmem_alloc(sizeof(dmu_object_info_t), KM_SLEEP);
error = dmu_objset_own(name, DMU_OST_ZVOL, B_TRUE, zvol_tag, &os);
if (error)
goto out_doi;
error = dmu_object_info(os, ZVOL_OBJ, doi);
if (error)
goto out_dmu_objset_disown;
error = zap_lookup(os, ZVOL_ZAP_OBJ, "size", 8, 1, &volsize);
if (error)
goto out_dmu_objset_disown;
error = zvol_find_minor(&minor);
if (error)
goto out_dmu_objset_disown;
zv = zvol_alloc(MKDEV(zvol_major, minor), name);
if (zv == NULL) {
error = EAGAIN;
goto out_dmu_objset_disown;
}
if (dmu_objset_is_snapshot(os))
zv->zv_flags |= ZVOL_RDONLY;
zv->zv_volblocksize = doi->doi_data_block_size;
zv->zv_volsize = volsize;
zv->zv_objset = os;
set_capacity(zv->zv_disk, zv->zv_volsize >> 9);
if (zil_replay_disable)
zil_destroy(dmu_objset_zil(os), B_FALSE);
else
zil_replay(os, zv, zvol_replay_vector);
out_dmu_objset_disown:
dmu_objset_disown(os, zvol_tag);
zv->zv_objset = NULL;
out_doi:
kmem_free(doi, sizeof(dmu_object_info_t));
out:
if (error == 0) {
zvol_insert(zv);
add_disk(zv->zv_disk);
}
return (error);
}
