d/freebsd-nq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
ce37ebd2eb
freebsd-nq/module/zfs/dsl_pool.c
Matthew Ahrens e8b96c6007 Illumos #4045 write throttle & i/o scheduler performance work 
4045 zfs write throttle & i/o scheduler performance work

1. The ZFS i/o scheduler (vdev_queue.c) now divides i/os into 5 classes: sync
read, sync write, async read, async write, and scrub/resilver.  The scheduler
issues a number of concurrent i/os from each class to the device.  Once a class
has been selected, an i/o is selected from this class using either an elevator
algorithem (async, scrub classes) or FIFO (sync classes).  The number of
concurrent async write i/os is tuned dynamically based on i/o load, to achieve
good sync i/o latency when there is not a high load of writes, and good write
throughput when there is.  See the block comment in vdev_queue.c (reproduced
below) for more details.

2. The write throttle (dsl_pool_tempreserve_space() and
txg_constrain_throughput()) is rewritten to produce much more consistent delays
when under constant load.  The new write throttle is based on the amount of
dirty data, rather than guesses about future performance of the system.  When
there is a lot of dirty data, each transaction (e.g. write() syscall) will be
delayed by the same small amount.  This eliminates the "brick wall of wait"
that the old write throttle could hit, causing all transactions to wait several
seconds until the next txg opens.  One of the keys to the new write throttle is
decrementing the amount of dirty data as i/o completes, rather than at the end
of spa_sync().  Note that the write throttle is only applied once the i/o
scheduler is issuing the maximum number of outstanding async writes.  See the
block comments in dsl_pool.c and above dmu_tx_delay() (reproduced below) for
more details.

This diff has several other effects, including:

 * the commonly-tuned global variable zfs_vdev_max_pending has been removed;
use per-class zfs_vdev_*_max_active values or zfs_vdev_max_active instead.

 * the size of each txg (meaning the amount of dirty data written, and thus the
time it takes to write out) is now controlled differently.  There is no longer
an explicit time goal; the primary determinant is amount of dirty data.
Systems that are under light or medium load will now often see that a txg is
always syncing, but the impact to performance (e.g. read latency) is minimal.
Tune zfs_dirty_data_max and zfs_dirty_data_sync to control this.

 * zio_taskq_batch_pct = 75 -- Only use 75% of all CPUs for compression,
checksum, etc.  This improves latency by not allowing these CPU-intensive tasks
to consume all CPU (on machines with at least 4 CPU's; the percentage is
rounded up).

--matt

APPENDIX: problems with the current i/o scheduler

The current ZFS i/o scheduler (vdev_queue.c) is deadline based.  The problem
with this is that if there are always i/os pending, then certain classes of
i/os can see very long delays.

For example, if there are always synchronous reads outstanding, then no async
writes will be serviced until they become "past due".  One symptom of this
situation is that each pass of the txg sync takes at least several seconds
(typically 3 seconds).

If many i/os become "past due" (their deadline is in the past), then we must
service all of these overdue i/os before any new i/os.  This happens when we
enqueue a batch of async writes for the txg sync, with deadlines 2.5 seconds in
the future.  If we can't complete all the i/os in 2.5 seconds (e.g. because
there were always reads pending), then these i/os will become past due.  Now we
must service all the "async" writes (which could be hundreds of megabytes)
before we service any reads, introducing considerable latency to synchronous
i/os (reads or ZIL writes).

Notes on porting to ZFS on Linux:

- zio_t gained new members io_physdone and io_phys_children.  Because
  object caches in the Linux port call the constructor only once at
  allocation time, objects may contain residual data when retrieved
  from the cache. Therefore zio_create() was updated to zero out the two
  new fields.

- vdev_mirror_pending() relied on the depth of the per-vdev pending queue
  (vq->vq_pending_tree) to select the least-busy leaf vdev to read from.
  This tree has been replaced by vq->vq_active_tree which is now used
  for the same purpose.

- vdev_queue_init() used the value of zfs_vdev_max_pending to determine
  the number of vdev I/O buffers to pre-allocate.  That global no longer
  exists, so we instead use the sum of the *_max_active values for each of
  the five I/O classes described above.

- The Illumos implementation of dmu_tx_delay() delays a transaction by
  sleeping in condition variable embedded in the thread
  (curthread->t_delay_cv).  We do not have an equivalent CV to use in
  Linux, so this change replaced the delay logic with a wrapper called
  zfs_sleep_until(). This wrapper could be adopted upstream and in other
  downstream ports to abstract away operating system-specific delay logic.

- These tunables are added as module parameters, and descriptions added
  to the zfs-module-parameters.5 man page.

  spa_asize_inflation
  zfs_deadman_synctime_ms
  zfs_vdev_max_active
  zfs_vdev_async_write_active_min_dirty_percent
  zfs_vdev_async_write_active_max_dirty_percent
  zfs_vdev_async_read_max_active
  zfs_vdev_async_read_min_active
  zfs_vdev_async_write_max_active
  zfs_vdev_async_write_min_active
  zfs_vdev_scrub_max_active
  zfs_vdev_scrub_min_active
  zfs_vdev_sync_read_max_active
  zfs_vdev_sync_read_min_active
  zfs_vdev_sync_write_max_active
  zfs_vdev_sync_write_min_active
  zfs_dirty_data_max_percent
  zfs_delay_min_dirty_percent
  zfs_dirty_data_max_max_percent
  zfs_dirty_data_max
  zfs_dirty_data_max_max
  zfs_dirty_data_sync
  zfs_delay_scale

  The latter four have type unsigned long, whereas they are uint64_t in
  Illumos.  This accommodates Linux's module_param() supported types, but
  means they may overflow on 32-bit architectures.

  The values zfs_dirty_data_max and zfs_dirty_data_max_max are the most
  likely to overflow on 32-bit systems, since they express physical RAM
  sizes in bytes.  In fact, Illumos initializes zfs_dirty_data_max_max to
  2^32 which does overflow. To resolve that, this port instead initializes
  it in arc_init() to 25% of physical RAM, and adds the tunable
  zfs_dirty_data_max_max_percent to override that percentage.  While this
  solution doesn't completely avoid the overflow issue, it should be a
  reasonable default for most systems, and the minority of affected
  systems can work around the issue by overriding the defaults.

- Fixed reversed logic in comment above zfs_delay_scale declaration.

- Clarified comments in vdev_queue.c regarding when per-queue minimums take
  effect.

- Replaced dmu_tx_write_limit in the dmu_tx kstat file
  with dmu_tx_dirty_delay and dmu_tx_dirty_over_max.  The first counts
  how many times a transaction has been delayed because the pool dirty
  data has exceeded zfs_delay_min_dirty_percent.  The latter counts how
  many times the pool dirty data has exceeded zfs_dirty_data_max (which
  we expect to never happen).

- The original patch would have regressed the bug fixed in
  zfsonlinux/zfs@c418410, which prevented users from setting the
  zfs_vdev_aggregation_limit tuning larger than SPA_MAXBLOCKSIZE.
  A similar fix is added to vdev_queue_aggregate().

- In vdev_queue_io_to_issue(), dynamically allocate 'zio_t search' on the
  heap instead of the stack.  In Linux we can't afford such large
  structures on the stack.

Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Adam Leventhal <ahl@delphix.com>
Reviewed by: Christopher Siden <christopher.siden@delphix.com>
Reviewed by: Ned Bass <bass6@llnl.gov>
Reviewed by: Brendan Gregg <brendan.gregg@joyent.com>
Approved by: Robert Mustacchi <rm@joyent.com>

References:
  http://www.illumos.org/issues/4045
  illumos/illumos-gate@69962b5647

Ported-by: Ned Bass <bass6@llnl.gov>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes #1913
2013-12-06 09:32:43 -08:00

1079 lines
31 KiB
C
Raw Blame History

/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License (the "License").
* You may not use this file except in compliance with the License.
*
* You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
* or http://www.opensolaris.org/os/licensing.
* See the License for the specific language governing permissions
* and limitations under the License.
*
* When distributing Covered Code, include this CDDL HEADER in each
* file and include the License file at usr/src/OPENSOLARIS.LICENSE.
* If applicable, add the following below this CDDL HEADER, with the
* fields enclosed by brackets "[]" replaced with your own identifying
* information: Portions Copyright [yyyy] [name of copyright owner]
*
* CDDL HEADER END
*/
/*
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
* Copyright (c) 2013 by Delphix. All rights reserved.
* Copyright (c) 2013 Steven Hartland. All rights reserved.
*/
#include <sys/dsl_pool.h>
#include <sys/dsl_dataset.h>
#include <sys/dsl_prop.h>
#include <sys/dsl_dir.h>
#include <sys/dsl_synctask.h>
#include <sys/dsl_scan.h>
#include <sys/dnode.h>
#include <sys/dmu_tx.h>
#include <sys/dmu_objset.h>
#include <sys/arc.h>
#include <sys/zap.h>
#include <sys/zio.h>
#include <sys/zfs_context.h>
#include <sys/fs/zfs.h>
#include <sys/zfs_znode.h>
#include <sys/spa_impl.h>
#include <sys/dsl_deadlist.h>
#include <sys/bptree.h>
#include <sys/zfeature.h>
#include <sys/zil_impl.h>
#include <sys/dsl_userhold.h>
/*
* ZFS Write Throttle
* ------------------
*
* ZFS must limit the rate of incoming writes to the rate at which it is able
* to sync data modifications to the backend storage. Throttling by too much
* creates an artificial limit; throttling by too little can only be sustained
* for short periods and would lead to highly lumpy performance. On a per-pool
* basis, ZFS tracks the amount of modified (dirty) data. As operations change
* data, the amount of dirty data increases; as ZFS syncs out data, the amount
* of dirty data decreases. When the amount of dirty data exceeds a
* predetermined threshold further modifications are blocked until the amount
* of dirty data decreases (as data is synced out).
*
* The limit on dirty data is tunable, and should be adjusted according to
* both the IO capacity and available memory of the system. The larger the
* window, the more ZFS is able to aggregate and amortize metadata (and data)
* changes. However, memory is a limited resource, and allowing for more dirty
* data comes at the cost of keeping other useful data in memory (for example
* ZFS data cached by the ARC).
*
* Implementation
*
* As buffers are modified dsl_pool_willuse_space() increments both the per-
* txg (dp_dirty_pertxg[]) and poolwide (dp_dirty_total) accounting of
* dirty space used; dsl_pool_dirty_space() decrements those values as data
* is synced out from dsl_pool_sync(). While only the poolwide value is
* relevant, the per-txg value is useful for debugging. The tunable
* zfs_dirty_data_max determines the dirty space limit. Once that value is
* exceeded, new writes are halted until space frees up.
*
* The zfs_dirty_data_sync tunable dictates the threshold at which we
* ensure that there is a txg syncing (see the comment in txg.c for a full
* description of transaction group stages).
*
* The IO scheduler uses both the dirty space limit and current amount of
* dirty data as inputs. Those values affect the number of concurrent IOs ZFS
* issues. See the comment in vdev_queue.c for details of the IO scheduler.
*
* The delay is also calculated based on the amount of dirty data. See the
* comment above dmu_tx_delay() for details.
*/
/*
* zfs_dirty_data_max will be set to zfs_dirty_data_max_percent% of all memory,
* capped at zfs_dirty_data_max_max. It can also be overridden with a module
* parameter.
*/
unsigned long zfs_dirty_data_max = 0;
unsigned long zfs_dirty_data_max_max = 0;
int zfs_dirty_data_max_percent = 10;
int zfs_dirty_data_max_max_percent = 25;
/*
* If there is at least this much dirty data, push out a txg.
*/
unsigned long zfs_dirty_data_sync = 64 * 1024 * 1024;
/*
* Once there is this amount of dirty data, the dmu_tx_delay() will kick in
* and delay each transaction.
* This value should be >= zfs_vdev_async_write_active_max_dirty_percent.
*/
int zfs_delay_min_dirty_percent = 60;
/*
* This controls how quickly the delay approaches infinity.
* Larger values cause it to delay more for a given amount of dirty data.
* Therefore larger values will cause there to be less dirty data for a
* given throughput.
*
* For the smoothest delay, this value should be about 1 billion divided
* by the maximum number of operations per second. This will smoothly
* handle between 10x and 1/10th this number.
*
* Note: zfs_delay_scale * zfs_dirty_data_max must be < 2^64, due to the
* multiply in dmu_tx_delay().
*/
unsigned long zfs_delay_scale = 1000 * 1000 * 1000 / 2000;
hrtime_t zfs_throttle_delay = MSEC2NSEC(10);
hrtime_t zfs_throttle_resolution = MSEC2NSEC(10);
int
dsl_pool_open_special_dir(dsl_pool_t *dp, const char *name, dsl_dir_t **ddp)
{
uint64_t obj;
int err;
err = zap_lookup(dp->dp_meta_objset,
dp->dp_root_dir->dd_phys->dd_child_dir_zapobj,
name, sizeof (obj), 1, &obj);
if (err)
return (err);
return (dsl_dir_hold_obj(dp, obj, name, dp, ddp));
}
static dsl_pool_t *
dsl_pool_open_impl(spa_t *spa, uint64_t txg)
{
dsl_pool_t *dp;
blkptr_t *bp = spa_get_rootblkptr(spa);
dp = kmem_zalloc(sizeof (dsl_pool_t), KM_SLEEP);
dp->dp_spa = spa;
dp->dp_meta_rootbp = *bp;
rrw_init(&dp->dp_config_rwlock, B_TRUE);
txg_init(dp, txg);
txg_list_create(&dp->dp_dirty_datasets,
offsetof(dsl_dataset_t, ds_dirty_link));
txg_list_create(&dp->dp_dirty_zilogs,
offsetof(zilog_t, zl_dirty_link));
txg_list_create(&dp->dp_dirty_dirs,
offsetof(dsl_dir_t, dd_dirty_link));
txg_list_create(&dp->dp_sync_tasks,
offsetof(dsl_sync_task_t, dst_node));
mutex_init(&dp->dp_lock, NULL, MUTEX_DEFAULT, NULL);
cv_init(&dp->dp_spaceavail_cv, NULL, CV_DEFAULT, NULL);
dp->dp_iput_taskq = taskq_create("zfs_iput_taskq", 1, minclsyspri,
1, 4, 0);
return (dp);
}
int
dsl_pool_init(spa_t *spa, uint64_t txg, dsl_pool_t **dpp)
{
int err;
dsl_pool_t *dp = dsl_pool_open_impl(spa, txg);
err = dmu_objset_open_impl(spa, NULL, &dp->dp_meta_rootbp,
&dp->dp_meta_objset);
if (err != 0)
dsl_pool_close(dp);
else
*dpp = dp;
return (err);
}
int
dsl_pool_open(dsl_pool_t *dp)
{
int err;
dsl_dir_t *dd;
dsl_dataset_t *ds;
uint64_t obj;
rrw_enter(&dp->dp_config_rwlock, RW_WRITER, FTAG);
err = zap_lookup(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_ROOT_DATASET, sizeof (uint64_t), 1,
&dp->dp_root_dir_obj);
if (err)
goto out;
err = dsl_dir_hold_obj(dp, dp->dp_root_dir_obj,
NULL, dp, &dp->dp_root_dir);
if (err)
goto out;
err = dsl_pool_open_special_dir(dp, MOS_DIR_NAME, &dp->dp_mos_dir);
if (err)
goto out;
if (spa_version(dp->dp_spa) >= SPA_VERSION_ORIGIN) {
err = dsl_pool_open_special_dir(dp, ORIGIN_DIR_NAME, &dd);
if (err)
goto out;
err = dsl_dataset_hold_obj(dp, dd->dd_phys->dd_head_dataset_obj,
FTAG, &ds);
if (err == 0) {
err = dsl_dataset_hold_obj(dp,
ds->ds_phys->ds_prev_snap_obj, dp,
&dp->dp_origin_snap);
dsl_dataset_rele(ds, FTAG);
}
dsl_dir_rele(dd, dp);
if (err)
goto out;
}
if (spa_version(dp->dp_spa) >= SPA_VERSION_DEADLISTS) {
err = dsl_pool_open_special_dir(dp, FREE_DIR_NAME,
&dp->dp_free_dir);
if (err)
goto out;
err = zap_lookup(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_FREE_BPOBJ, sizeof (uint64_t), 1, &obj);
if (err)
goto out;
VERIFY0(bpobj_open(&dp->dp_free_bpobj,
dp->dp_meta_objset, obj));
}
if (spa_feature_is_active(dp->dp_spa,
&spa_feature_table[SPA_FEATURE_ASYNC_DESTROY])) {
err = zap_lookup(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_BPTREE_OBJ, sizeof (uint64_t), 1,
&dp->dp_bptree_obj);
if (err != 0)
goto out;
}
if (spa_feature_is_active(dp->dp_spa,
&spa_feature_table[SPA_FEATURE_EMPTY_BPOBJ])) {
err = zap_lookup(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_EMPTY_BPOBJ, sizeof (uint64_t), 1,
&dp->dp_empty_bpobj);
if (err != 0)
goto out;
}
err = zap_lookup(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_TMP_USERREFS, sizeof (uint64_t), 1,
&dp->dp_tmp_userrefs_obj);
if (err == ENOENT)
err = 0;
if (err)
goto out;
err = dsl_scan_init(dp, dp->dp_tx.tx_open_txg);
out:
rrw_exit(&dp->dp_config_rwlock, FTAG);
return (err);
}
void
dsl_pool_close(dsl_pool_t *dp)
{
/*
* Drop our references from dsl_pool_open().
*
* Since we held the origin_snap from "syncing" context (which
* includes pool-opening context), it actually only got a "ref"
* and not a hold, so just drop that here.
*/
if (dp->dp_origin_snap)
dsl_dataset_rele(dp->dp_origin_snap, dp);
if (dp->dp_mos_dir)
dsl_dir_rele(dp->dp_mos_dir, dp);
if (dp->dp_free_dir)
dsl_dir_rele(dp->dp_free_dir, dp);
if (dp->dp_root_dir)
dsl_dir_rele(dp->dp_root_dir, dp);
bpobj_close(&dp->dp_free_bpobj);
/* undo the dmu_objset_open_impl(mos) from dsl_pool_open() */
if (dp->dp_meta_objset)
dmu_objset_evict(dp->dp_meta_objset);
txg_list_destroy(&dp->dp_dirty_datasets);
txg_list_destroy(&dp->dp_dirty_zilogs);
txg_list_destroy(&dp->dp_sync_tasks);
txg_list_destroy(&dp->dp_dirty_dirs);
arc_flush(dp->dp_spa);
txg_fini(dp);
dsl_scan_fini(dp);
rrw_destroy(&dp->dp_config_rwlock);
mutex_destroy(&dp->dp_lock);
taskq_destroy(dp->dp_iput_taskq);
if (dp->dp_blkstats)
kmem_free(dp->dp_blkstats, sizeof (zfs_all_blkstats_t));
kmem_free(dp, sizeof (dsl_pool_t));
}
dsl_pool_t *
dsl_pool_create(spa_t *spa, nvlist_t *zplprops, uint64_t txg)
{
int err;
dsl_pool_t *dp = dsl_pool_open_impl(spa, txg);
dmu_tx_t *tx = dmu_tx_create_assigned(dp, txg);
objset_t *os;
dsl_dataset_t *ds;
uint64_t obj;
rrw_enter(&dp->dp_config_rwlock, RW_WRITER, FTAG);
/* create and open the MOS (meta-objset) */
dp->dp_meta_objset = dmu_objset_create_impl(spa,
NULL, &dp->dp_meta_rootbp, DMU_OST_META, tx);
/* create the pool directory */
err = zap_create_claim(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_OT_OBJECT_DIRECTORY, DMU_OT_NONE, 0, tx);
ASSERT0(err);
/* Initialize scan structures */
VERIFY0(dsl_scan_init(dp, txg));
/* create and open the root dir */
dp->dp_root_dir_obj = dsl_dir_create_sync(dp, NULL, NULL, tx);
VERIFY0(dsl_dir_hold_obj(dp, dp->dp_root_dir_obj,
NULL, dp, &dp->dp_root_dir));
/* create and open the meta-objset dir */
(void) dsl_dir_create_sync(dp, dp->dp_root_dir, MOS_DIR_NAME, tx);
VERIFY0(dsl_pool_open_special_dir(dp,
MOS_DIR_NAME, &dp->dp_mos_dir));
if (spa_version(spa) >= SPA_VERSION_DEADLISTS) {
/* create and open the free dir */
(void) dsl_dir_create_sync(dp, dp->dp_root_dir,
FREE_DIR_NAME, tx);
VERIFY0(dsl_pool_open_special_dir(dp,
FREE_DIR_NAME, &dp->dp_free_dir));
/* create and open the free_bplist */
obj = bpobj_alloc(dp->dp_meta_objset, SPA_MAXBLOCKSIZE, tx);
VERIFY(zap_add(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_FREE_BPOBJ, sizeof (uint64_t), 1, &obj, tx) == 0);
VERIFY0(bpobj_open(&dp->dp_free_bpobj,
dp->dp_meta_objset, obj));
}
if (spa_version(spa) >= SPA_VERSION_DSL_SCRUB)
dsl_pool_create_origin(dp, tx);
/* create the root dataset */
obj = dsl_dataset_create_sync_dd(dp->dp_root_dir, NULL, 0, tx);
/* create the root objset */
VERIFY0(dsl_dataset_hold_obj(dp, obj, FTAG, &ds));
VERIFY(NULL != (os = dmu_objset_create_impl(dp->dp_spa, ds,
dsl_dataset_get_blkptr(ds), DMU_OST_ZFS, tx)));
#ifdef _KERNEL
zfs_create_fs(os, kcred, zplprops, tx);
#endif
dsl_dataset_rele(ds, FTAG);
dmu_tx_commit(tx);
rrw_exit(&dp->dp_config_rwlock, FTAG);
return (dp);
}
/*
* Account for the meta-objset space in its placeholder dsl_dir.
*/
void
dsl_pool_mos_diduse_space(dsl_pool_t *dp,
int64_t used, int64_t comp, int64_t uncomp)
{
ASSERT3U(comp, ==, uncomp); /* it's all metadata */
mutex_enter(&dp->dp_lock);
dp->dp_mos_used_delta += used;
dp->dp_mos_compressed_delta += comp;
dp->dp_mos_uncompressed_delta += uncomp;
mutex_exit(&dp->dp_lock);
}
static int
deadlist_enqueue_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx)
{
dsl_deadlist_t *dl = arg;
dsl_deadlist_insert(dl, bp, tx);
return (0);
}
static void
dsl_pool_sync_mos(dsl_pool_t *dp, dmu_tx_t *tx)
{
zio_t *zio = zio_root(dp->dp_spa, NULL, NULL, ZIO_FLAG_MUSTSUCCEED);
dmu_objset_sync(dp->dp_meta_objset, zio, tx);
VERIFY0(zio_wait(zio));
dprintf_bp(&dp->dp_meta_rootbp, "meta objset rootbp is %s", "");
spa_set_rootblkptr(dp->dp_spa, &dp->dp_meta_rootbp);
}
static void
dsl_pool_dirty_delta(dsl_pool_t *dp, int64_t delta)
{
ASSERT(MUTEX_HELD(&dp->dp_lock));
if (delta < 0)
ASSERT3U(-delta, <=, dp->dp_dirty_total);
dp->dp_dirty_total += delta;
/*
* Note: we signal even when increasing dp_dirty_total.
* This ensures forward progress -- each thread wakes the next waiter.
*/
if (dp->dp_dirty_total <= zfs_dirty_data_max)
cv_signal(&dp->dp_spaceavail_cv);
}
void
dsl_pool_sync(dsl_pool_t *dp, uint64_t txg)
{
zio_t *zio;
dmu_tx_t *tx;
dsl_dir_t *dd;
dsl_dataset_t *ds;
objset_t *mos = dp->dp_meta_objset;
list_t synced_datasets;
list_create(&synced_datasets, sizeof (dsl_dataset_t),
offsetof(dsl_dataset_t, ds_synced_link));
tx = dmu_tx_create_assigned(dp, txg);
/*
* Write out all dirty blocks of dirty datasets.
*/
zio = zio_root(dp->dp_spa, NULL, NULL, ZIO_FLAG_MUSTSUCCEED);
while ((ds = txg_list_remove(&dp->dp_dirty_datasets, txg)) != NULL) {
/*
* We must not sync any non-MOS datasets twice, because
* we may have taken a snapshot of them. However, we
* may sync newly-created datasets on pass 2.
*/
ASSERT(!list_link_active(&ds->ds_synced_link));
list_insert_tail(&synced_datasets, ds);
dsl_dataset_sync(ds, zio, tx);
}
VERIFY0(zio_wait(zio));
/*
* We have written all of the accounted dirty data, so our
* dp_space_towrite should now be zero. However, some seldom-used
* code paths do not adhere to this (e.g. dbuf_undirty(), also
* rounding error in dbuf_write_physdone).
* Shore up the accounting of any dirtied space now.
*/
dsl_pool_undirty_space(dp, dp->dp_dirty_pertxg[txg & TXG_MASK], txg);
/*
* After the data blocks have been written (ensured by the zio_wait()
* above), update the user/group space accounting.
*/
for (ds = list_head(&synced_datasets); ds != NULL;
ds = list_next(&synced_datasets, ds)) {
dmu_objset_do_userquota_updates(ds->ds_objset, tx);
}
/*
* Sync the datasets again to push out the changes due to
* userspace updates. This must be done before we process the
* sync tasks, so that any snapshots will have the correct
* user accounting information (and we won't get confused
* about which blocks are part of the snapshot).
*/
zio = zio_root(dp->dp_spa, NULL, NULL, ZIO_FLAG_MUSTSUCCEED);
while ((ds = txg_list_remove(&dp->dp_dirty_datasets, txg)) != NULL) {
ASSERT(list_link_active(&ds->ds_synced_link));
dmu_buf_rele(ds->ds_dbuf, ds);
dsl_dataset_sync(ds, zio, tx);
}
VERIFY0(zio_wait(zio));
/*
* Now that the datasets have been completely synced, we can
* clean up our in-memory structures accumulated while syncing:
*
* - move dead blocks from the pending deadlist to the on-disk deadlist
* - release hold from dsl_dataset_dirty()
*/
while ((ds = list_remove_head(&synced_datasets)) != NULL) {
ASSERTV(objset_t *os = ds->ds_objset);
bplist_iterate(&ds->ds_pending_deadlist,
deadlist_enqueue_cb, &ds->ds_deadlist, tx);
ASSERT(!dmu_objset_is_dirty(os, txg));
dmu_buf_rele(ds->ds_dbuf, ds);
}
while ((dd = txg_list_remove(&dp->dp_dirty_dirs, txg)) != NULL) {
dsl_dir_sync(dd, tx);
}
/*
* The MOS's space is accounted for in the pool/$MOS
* (dp_mos_dir). We can't modify the mos while we're syncing
* it, so we remember the deltas and apply them here.
*/
if (dp->dp_mos_used_delta != 0 || dp->dp_mos_compressed_delta != 0 ||
dp->dp_mos_uncompressed_delta != 0) {
dsl_dir_diduse_space(dp->dp_mos_dir, DD_USED_HEAD,
dp->dp_mos_used_delta,
dp->dp_mos_compressed_delta,
dp->dp_mos_uncompressed_delta, tx);
dp->dp_mos_used_delta = 0;
dp->dp_mos_compressed_delta = 0;
dp->dp_mos_uncompressed_delta = 0;
}
if (list_head(&mos->os_dirty_dnodes[txg & TXG_MASK]) != NULL ||
list_head(&mos->os_free_dnodes[txg & TXG_MASK]) != NULL) {
dsl_pool_sync_mos(dp, tx);
}
/*
* If we modify a dataset in the same txg that we want to destroy it,
* its dsl_dir's dd_dbuf will be dirty, and thus have a hold on it.
* dsl_dir_destroy_check() will fail if there are unexpected holds.
* Therefore, we want to sync the MOS (thus syncing the dd_dbuf
* and clearing the hold on it) before we process the sync_tasks.
* The MOS data dirtied by the sync_tasks will be synced on the next
* pass.
*/
if (!txg_list_empty(&dp->dp_sync_tasks, txg)) {
dsl_sync_task_t *dst;
/*
* No more sync tasks should have been added while we
* were syncing.
*/
ASSERT3U(spa_sync_pass(dp->dp_spa), ==, 1);
while ((dst = txg_list_remove(&dp->dp_sync_tasks, txg)) != NULL)
dsl_sync_task_sync(dst, tx);
}
dmu_tx_commit(tx);
DTRACE_PROBE2(dsl_pool_sync__done, dsl_pool_t *dp, dp, uint64_t, txg);
}
void
dsl_pool_sync_done(dsl_pool_t *dp, uint64_t txg)
{
zilog_t *zilog;
while ((zilog = txg_list_remove(&dp->dp_dirty_zilogs, txg))) {
dsl_dataset_t *ds = dmu_objset_ds(zilog->zl_os);
zil_clean(zilog, txg);
ASSERT(!dmu_objset_is_dirty(zilog->zl_os, txg));
dmu_buf_rele(ds->ds_dbuf, zilog);
}
ASSERT(!dmu_objset_is_dirty(dp->dp_meta_objset, txg));
}
/*
* TRUE if the current thread is the tx_sync_thread or if we
* are being called from SPA context during pool initialization.
*/
int
dsl_pool_sync_context(dsl_pool_t *dp)
{
return (curthread == dp->dp_tx.tx_sync_thread ||
spa_is_initializing(dp->dp_spa));
}
uint64_t
dsl_pool_adjustedsize(dsl_pool_t *dp, boolean_t netfree)
{
uint64_t space, resv;
/*
* Reserve about 1.6% (1/64), or at least 32MB, for allocation
* efficiency.
* XXX The intent log is not accounted for, so it must fit
* within this slop.
*
* If we're trying to assess whether it's OK to do a free,
* cut the reservation in half to allow forward progress
* (e.g. make it possible to rm(1) files from a full pool).
*/
space = spa_get_dspace(dp->dp_spa);
resv = MAX(space >> 6, SPA_MINDEVSIZE >> 1);
if (netfree)
resv >>= 1;
return (space - resv);
}
boolean_t
dsl_pool_need_dirty_delay(dsl_pool_t *dp)
{
uint64_t delay_min_bytes =
zfs_dirty_data_max * zfs_delay_min_dirty_percent / 100;
boolean_t rv;
mutex_enter(&dp->dp_lock);
if (dp->dp_dirty_total > zfs_dirty_data_sync)
txg_kick(dp);
rv = (dp->dp_dirty_total > delay_min_bytes);
mutex_exit(&dp->dp_lock);
return (rv);
}
void
dsl_pool_dirty_space(dsl_pool_t *dp, int64_t space, dmu_tx_t *tx)
{
if (space > 0) {
mutex_enter(&dp->dp_lock);
dp->dp_dirty_pertxg[tx->tx_txg & TXG_MASK] += space;
dsl_pool_dirty_delta(dp, space);
mutex_exit(&dp->dp_lock);
}
}
void
dsl_pool_undirty_space(dsl_pool_t *dp, int64_t space, uint64_t txg)
{
ASSERT3S(space, >=, 0);
if (space == 0)
return;
mutex_enter(&dp->dp_lock);
if (dp->dp_dirty_pertxg[txg & TXG_MASK] < space) {
/* XXX writing something we didn't dirty? */
space = dp->dp_dirty_pertxg[txg & TXG_MASK];
}
ASSERT3U(dp->dp_dirty_pertxg[txg & TXG_MASK], >=, space);
dp->dp_dirty_pertxg[txg & TXG_MASK] -= space;
ASSERT3U(dp->dp_dirty_total, >=, space);
dsl_pool_dirty_delta(dp, -space);
mutex_exit(&dp->dp_lock);
}
/* ARGSUSED */
static int
upgrade_clones_cb(dsl_pool_t *dp, dsl_dataset_t *hds, void *arg)
{
dmu_tx_t *tx = arg;
dsl_dataset_t *ds, *prev = NULL;
int err;
err = dsl_dataset_hold_obj(dp, hds->ds_object, FTAG, &ds);
if (err)
return (err);
while (ds->ds_phys->ds_prev_snap_obj != 0) {
err = dsl_dataset_hold_obj(dp, ds->ds_phys->ds_prev_snap_obj,
FTAG, &prev);
if (err) {
dsl_dataset_rele(ds, FTAG);
return (err);
}
if (prev->ds_phys->ds_next_snap_obj != ds->ds_object)
break;
dsl_dataset_rele(ds, FTAG);
ds = prev;
prev = NULL;
}
if (prev == NULL) {
prev = dp->dp_origin_snap;
/*
* The $ORIGIN can't have any data, or the accounting
* will be wrong.
*/
ASSERT0(prev->ds_phys->ds_bp.blk_birth);
/* The origin doesn't get attached to itself */
if (ds->ds_object == prev->ds_object) {
dsl_dataset_rele(ds, FTAG);
return (0);
}
dmu_buf_will_dirty(ds->ds_dbuf, tx);
ds->ds_phys->ds_prev_snap_obj = prev->ds_object;
ds->ds_phys->ds_prev_snap_txg = prev->ds_phys->ds_creation_txg;
dmu_buf_will_dirty(ds->ds_dir->dd_dbuf, tx);
ds->ds_dir->dd_phys->dd_origin_obj = prev->ds_object;
dmu_buf_will_dirty(prev->ds_dbuf, tx);
prev->ds_phys->ds_num_children++;
if (ds->ds_phys->ds_next_snap_obj == 0) {
ASSERT(ds->ds_prev == NULL);
VERIFY0(dsl_dataset_hold_obj(dp,
ds->ds_phys->ds_prev_snap_obj, ds, &ds->ds_prev));
}
}
ASSERT3U(ds->ds_dir->dd_phys->dd_origin_obj, ==, prev->ds_object);
ASSERT3U(ds->ds_phys->ds_prev_snap_obj, ==, prev->ds_object);
if (prev->ds_phys->ds_next_clones_obj == 0) {
dmu_buf_will_dirty(prev->ds_dbuf, tx);
prev->ds_phys->ds_next_clones_obj =
zap_create(dp->dp_meta_objset,
DMU_OT_NEXT_CLONES, DMU_OT_NONE, 0, tx);
}
VERIFY0(zap_add_int(dp->dp_meta_objset,
prev->ds_phys->ds_next_clones_obj, ds->ds_object, tx));
dsl_dataset_rele(ds, FTAG);
if (prev != dp->dp_origin_snap)
dsl_dataset_rele(prev, FTAG);
return (0);
}
void
dsl_pool_upgrade_clones(dsl_pool_t *dp, dmu_tx_t *tx)
{
ASSERT(dmu_tx_is_syncing(tx));
ASSERT(dp->dp_origin_snap != NULL);
VERIFY0(dmu_objset_find_dp(dp, dp->dp_root_dir_obj, upgrade_clones_cb,
tx, DS_FIND_CHILDREN));
}
/* ARGSUSED */
static int
upgrade_dir_clones_cb(dsl_pool_t *dp, dsl_dataset_t *ds, void *arg)
{
dmu_tx_t *tx = arg;
objset_t *mos = dp->dp_meta_objset;
if (ds->ds_dir->dd_phys->dd_origin_obj != 0) {
dsl_dataset_t *origin;
VERIFY0(dsl_dataset_hold_obj(dp,
ds->ds_dir->dd_phys->dd_origin_obj, FTAG, &origin));
if (origin->ds_dir->dd_phys->dd_clones == 0) {
dmu_buf_will_dirty(origin->ds_dir->dd_dbuf, tx);
origin->ds_dir->dd_phys->dd_clones = zap_create(mos,
DMU_OT_DSL_CLONES, DMU_OT_NONE, 0, tx);
}
VERIFY0(zap_add_int(dp->dp_meta_objset,
origin->ds_dir->dd_phys->dd_clones, ds->ds_object, tx));
dsl_dataset_rele(origin, FTAG);
}
return (0);
}
void
dsl_pool_upgrade_dir_clones(dsl_pool_t *dp, dmu_tx_t *tx)
{
uint64_t obj;
ASSERT(dmu_tx_is_syncing(tx));
(void) dsl_dir_create_sync(dp, dp->dp_root_dir, FREE_DIR_NAME, tx);
VERIFY0(dsl_pool_open_special_dir(dp,
FREE_DIR_NAME, &dp->dp_free_dir));
/*
* We can't use bpobj_alloc(), because spa_version() still
* returns the old version, and we need a new-version bpobj with
* subobj support. So call dmu_object_alloc() directly.
*/
obj = dmu_object_alloc(dp->dp_meta_objset, DMU_OT_BPOBJ,
SPA_MAXBLOCKSIZE, DMU_OT_BPOBJ_HDR, sizeof (bpobj_phys_t), tx);
VERIFY0(zap_add(dp->dp_meta_objset, DMU_POOL_DIRECTORY_OBJECT,
DMU_POOL_FREE_BPOBJ, sizeof (uint64_t), 1, &obj, tx));
VERIFY0(bpobj_open(&dp->dp_free_bpobj, dp->dp_meta_objset, obj));
VERIFY0(dmu_objset_find_dp(dp, dp->dp_root_dir_obj,
upgrade_dir_clones_cb, tx, DS_FIND_CHILDREN));
}
void
dsl_pool_create_origin(dsl_pool_t *dp, dmu_tx_t *tx)
{
uint64_t dsobj;
dsl_dataset_t *ds;
ASSERT(dmu_tx_is_syncing(tx));
ASSERT(dp->dp_origin_snap == NULL);
ASSERT(rrw_held(&dp->dp_config_rwlock, RW_WRITER));
/* create the origin dir, ds, & snap-ds */
dsobj = dsl_dataset_create_sync(dp->dp_root_dir, ORIGIN_DIR_NAME,
NULL, 0, kcred, tx);
VERIFY0(dsl_dataset_hold_obj(dp, dsobj, FTAG, &ds));
dsl_dataset_snapshot_sync_impl(ds, ORIGIN_DIR_NAME, tx);
VERIFY0(dsl_dataset_hold_obj(dp, ds->ds_phys->ds_prev_snap_obj,
dp, &dp->dp_origin_snap));
dsl_dataset_rele(ds, FTAG);
}
taskq_t *
dsl_pool_iput_taskq(dsl_pool_t *dp)
{
return (dp->dp_iput_taskq);
}
/*
* Walk through the pool-wide zap object of temporary snapshot user holds
* and release them.
*/
void
dsl_pool_clean_tmp_userrefs(dsl_pool_t *dp)
{
zap_attribute_t za;
zap_cursor_t zc;
objset_t *mos = dp->dp_meta_objset;
uint64_t zapobj = dp->dp_tmp_userrefs_obj;
nvlist_t *holds;
if (zapobj == 0)
return;
ASSERT(spa_version(dp->dp_spa) >= SPA_VERSION_USERREFS);
holds = fnvlist_alloc();
for (zap_cursor_init(&zc, mos, zapobj);
zap_cursor_retrieve(&zc, &za) == 0;
zap_cursor_advance(&zc)) {
char *htag;
nvlist_t *tags;
htag = strchr(za.za_name, '-');
*htag = '\0';
++htag;
if (nvlist_lookup_nvlist(holds, za.za_name, &tags) != 0) {
tags = fnvlist_alloc();
fnvlist_add_boolean(tags, htag);
fnvlist_add_nvlist(holds, za.za_name, tags);
fnvlist_free(tags);
} else {
fnvlist_add_boolean(tags, htag);
}
}
dsl_dataset_user_release_tmp(dp, holds);
fnvlist_free(holds);
zap_cursor_fini(&zc);
}
/*
* Create the pool-wide zap object for storing temporary snapshot holds.
*/
void
dsl_pool_user_hold_create_obj(dsl_pool_t *dp, dmu_tx_t *tx)
{
objset_t *mos = dp->dp_meta_objset;
ASSERT(dp->dp_tmp_userrefs_obj == 0);
ASSERT(dmu_tx_is_syncing(tx));
dp->dp_tmp_userrefs_obj = zap_create_link(mos, DMU_OT_USERREFS,
DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_TMP_USERREFS, tx);
}
static int
dsl_pool_user_hold_rele_impl(dsl_pool_t *dp, uint64_t dsobj,
const char *tag, uint64_t now, dmu_tx_t *tx, boolean_t holding)
{
objset_t *mos = dp->dp_meta_objset;
uint64_t zapobj = dp->dp_tmp_userrefs_obj;
char *name;
int error;
ASSERT(spa_version(dp->dp_spa) >= SPA_VERSION_USERREFS);
ASSERT(dmu_tx_is_syncing(tx));
/*
* If the pool was created prior to SPA_VERSION_USERREFS, the
* zap object for temporary holds might not exist yet.
*/
if (zapobj == 0) {
if (holding) {
dsl_pool_user_hold_create_obj(dp, tx);
zapobj = dp->dp_tmp_userrefs_obj;
} else {
return (SET_ERROR(ENOENT));
}
}
name = kmem_asprintf("%llx-%s", (u_longlong_t)dsobj, tag);
if (holding)
error = zap_add(mos, zapobj, name, 8, 1, &now, tx);
else
error = zap_remove(mos, zapobj, name, tx);
strfree(name);
return (error);
}
/*
* Add a temporary hold for the given dataset object and tag.
*/
int
dsl_pool_user_hold(dsl_pool_t *dp, uint64_t dsobj, const char *tag,
uint64_t now, dmu_tx_t *tx)
{
return (dsl_pool_user_hold_rele_impl(dp, dsobj, tag, now, tx, B_TRUE));
}
/*
* Release a temporary hold for the given dataset object and tag.
*/
int
dsl_pool_user_release(dsl_pool_t *dp, uint64_t dsobj, const char *tag,
dmu_tx_t *tx)
{
return (dsl_pool_user_hold_rele_impl(dp, dsobj, tag, 0,
tx, B_FALSE));
}
/*
* DSL Pool Configuration Lock
*
* The dp_config_rwlock protects against changes to DSL state (e.g. dataset
* creation / destruction / rename / property setting). It must be held for
* read to hold a dataset or dsl_dir. I.e. you must call
* dsl_pool_config_enter() or dsl_pool_hold() before calling
* dsl_{dataset,dir}_hold{_obj}. In most circumstances, the dp_config_rwlock
* must be held continuously until all datasets and dsl_dirs are released.
*
* The only exception to this rule is that if a "long hold" is placed on
* a dataset, then the dp_config_rwlock may be dropped while the dataset
* is still held. The long hold will prevent the dataset from being
* destroyed -- the destroy will fail with EBUSY. A long hold can be
* obtained by calling dsl_dataset_long_hold(), or by "owning" a dataset
* (by calling dsl_{dataset,objset}_{try}own{_obj}).
*
* Legitimate long-holders (including owners) should be long-running, cancelable
* tasks that should cause "zfs destroy" to fail. This includes DMU
* consumers (i.e. a ZPL filesystem being mounted or ZVOL being open),
* "zfs send", and "zfs diff". There are several other long-holders whose
* uses are suboptimal (e.g. "zfs promote", and zil_suspend()).
*
* The usual formula for long-holding would be:
* dsl_pool_hold()
* dsl_dataset_hold()
* ... perform checks ...
* dsl_dataset_long_hold()
* dsl_pool_rele()
* ... perform long-running task ...
* dsl_dataset_long_rele()
* dsl_dataset_rele()
*
* Note that when the long hold is released, the dataset is still held but
* the pool is not held. The dataset may change arbitrarily during this time
* (e.g. it could be destroyed). Therefore you shouldn't do anything to the
* dataset except release it.
*
* User-initiated operations (e.g. ioctls, zfs_ioc_*()) are either read-only
* or modifying operations.
*
* Modifying operations should generally use dsl_sync_task(). The synctask
* infrastructure enforces proper locking strategy with respect to the
* dp_config_rwlock. See the comment above dsl_sync_task() for details.
*
* Read-only operations will manually hold the pool, then the dataset, obtain
* information from the dataset, then release the pool and dataset.
* dmu_objset_{hold,rele}() are convenience routines that also do the pool
* hold/rele.
*/
int
dsl_pool_hold(const char *name, void *tag, dsl_pool_t **dp)
{
spa_t *spa;
int error;
error = spa_open(name, &spa, tag);
if (error == 0) {
*dp = spa_get_dsl(spa);
dsl_pool_config_enter(*dp, tag);
}
return (error);
}
void
dsl_pool_rele(dsl_pool_t *dp, void *tag)
{
dsl_pool_config_exit(dp, tag);
spa_close(dp->dp_spa, tag);
}
void
dsl_pool_config_enter(dsl_pool_t *dp, void *tag)
{
/*
* We use a "reentrant" reader-writer lock, but not reentrantly.
*
* The rrwlock can (with the track_all flag) track all reading threads,
* which is very useful for debugging which code path failed to release
* the lock, and for verifying that the *current* thread does hold
* the lock.
*
* (Unlike a rwlock, which knows that N threads hold it for
* read, but not *which* threads, so rw_held(RW_READER) returns TRUE
* if any thread holds it for read, even if this thread doesn't).
*/
ASSERT(!rrw_held(&dp->dp_config_rwlock, RW_READER));
rrw_enter(&dp->dp_config_rwlock, RW_READER, tag);
}
void
dsl_pool_config_exit(dsl_pool_t *dp, void *tag)
{
rrw_exit(&dp->dp_config_rwlock, tag);
}
boolean_t
dsl_pool_config_held(dsl_pool_t *dp)
{
return (RRW_LOCK_HELD(&dp->dp_config_rwlock));
}
#if defined(_KERNEL) && defined(HAVE_SPL)
EXPORT_SYMBOL(dsl_pool_config_enter);
EXPORT_SYMBOL(dsl_pool_config_exit);
/* zfs_dirty_data_max_percent only applied at module load time in arc_init(). */
module_param(zfs_dirty_data_max_percent, int, 0444);
MODULE_PARM_DESC(zfs_dirty_data_max_percent, "percent of ram can be dirty");
/* zfs_dirty_data_max_max_percent only applied at module load time in
* arc_init(). */
module_param(zfs_dirty_data_max_max_percent, int, 0444);
MODULE_PARM_DESC(zfs_dirty_data_max_max_percent,
"zfs_dirty_data_max upper bound as % of RAM");
module_param(zfs_delay_min_dirty_percent, int, 0644);
MODULE_PARM_DESC(zfs_delay_min_dirty_percent, "transaction delay threshold");
module_param(zfs_dirty_data_max, ulong, 0644);
MODULE_PARM_DESC(zfs_dirty_data_max, "determines the dirty space limit");
/* zfs_dirty_data_max_max only applied at module load time in arc_init(). */
module_param(zfs_dirty_data_max_max, ulong, 0444);
MODULE_PARM_DESC(zfs_dirty_data_max_max,
"zfs_dirty_data_max upper bound in bytes");
module_param(zfs_dirty_data_sync, ulong, 0644);
MODULE_PARM_DESC(zfs_dirty_data_sync, "sync txg when this much dirty data");
module_param(zfs_delay_scale, ulong, 0644);
MODULE_PARM_DESC(zfs_delay_scale, "how quickly delay approaches infinity");
#endif
Reference in New Issue View Git Blame Copy Permalink