c/mirror_zfs
1
0
Fork 0
mirror of https://git.proxmox.com/git/mirror_zfs.git synced 2024-11-17 18:11:00 +03:00
Code Issues Packages Projects Releases Wiki Activity
562e1c0327
mirror_zfs/module/zfs/dmu_tx.c

1405 lines
40 KiB
C
Raw Normal View History

Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
/*
* 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
*/
/*
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
* Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
Illumos #1475: zfs spill block hold can access invalid spill blkptr Reviewed by: Dan McDonald <danmcd@nexenta.com> Reviewed by: Gordon Ross <gwr@nexenta.com> Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed by: George Wilson <gwilson@zfsmail.com> Approved by: Garrett D'Amore <garrett@nexenta.com> References to Illumos issue: https://www.illumos.org/issues/1475 Ported-by: Richard Yao <ryao@cs.stonybrook.edu> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #648
2012-04-08 21:10:49 +04:00
* Copyright 2011 Nexenta Systems, Inc. All rights reserved.
OpenZFS 8063 - verify that we do not attempt to access inactive txg Authored by: Matthew Ahrens <mahrens@delphix.com> Reviewed by: Serapheim Dimitropoulos <serapheim@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Approved by: Robert Mustacchi <rm@joyent.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Ported-by: George Melikov <mail@gmelikov.ru> A standard practice in ZFS is to keep track of "per-txg" state. Any of the 3 active TXG's (open, quiescing, syncing) can have different values for this state. We should assert that we do not attempt to modify other (inactive) TXG's. Porting Notes: - ASSERTV added to txg_sync_waiting() for unused variable. OpenZFS-issue: https://www.illumos.org/issues/8063 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/01acb46 Closes #6109
2017-04-24 19:34:36 +03:00
* Copyright (c) 2012, 2017 by Delphix. All rights reserved.
Illumos #1475: zfs spill block hold can access invalid spill blkptr Reviewed by: Dan McDonald <danmcd@nexenta.com> Reviewed by: Gordon Ross <gwr@nexenta.com> Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed by: George Wilson <gwilson@zfsmail.com> Approved by: Garrett D'Amore <garrett@nexenta.com> References to Illumos issue: https://www.illumos.org/issues/1475 Ported-by: Richard Yao <ryao@cs.stonybrook.edu> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #648
2012-04-08 21:10:49 +04:00
*/
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
#include <sys/dmu.h>
#include <sys/dmu_impl.h>
#include <sys/dbuf.h>
#include <sys/dmu_tx.h>
#include <sys/dmu_objset.h>
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
#include <sys/dsl_dataset.h>
#include <sys/dsl_dir.h>
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
#include <sys/dsl_pool.h>
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
#include <sys/zap_impl.h>
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
#include <sys/spa.h>
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
#include <sys/sa.h>
#include <sys/sa_impl.h>
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
#include <sys/zfs_context.h>
Remove duplicate typedefs from trace.h Older versions of GCC (e.g. GCC 4.4.7 on RHEL6) do not allow duplicate typedef declarations with the same type. The trace.h header contains some typedefs to avoid 'unknown type' errors for C files that haven't declared the type in question. But this causes build failures for C files that have already declared the type. Newer versions of GCC (e.g. v4.6) allow duplicate typedefs with the same type unless pedantic error checking is in force. To support the older versions we need to remove the duplicate typedefs. Removal of the typedefs means we can't built tracepoints code using those types unless the required headers have been included. To facilitate this, all tracepoint event declarations have been moved out of trace.h into separate headers. Each new header is explicitly included from the C file that uses the events defined therein. The trace.h header is still indirectly included form zfs_context.h and provides the implementation of the dprintf(), dbgmsg(), and SET_ERROR() interfaces. This makes those interfaces readily available throughout the code base. The macros that redefine DTRACE_PROBE* to use Linux tracepoints are also still provided by trace.h, so it is a prerequisite for the other trace_*.h headers. These new Linux implementation-specific headers do introduce a small divergence from upstream ZFS in several core C files, but this should not present a significant maintenance burden. Signed-off-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #2953
2014-12-13 05:07:39 +03:00
#include <sys/trace_dmu.h>
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
typedef void (*dmu_tx_hold_func_t)(dmu_tx_t *tx, struct dnode *dn,
uint64_t arg1, uint64_t arg2);
Add 'dmu_tx' kstats entry Keep counters for the various reasons that a thread may end up in txg_wait_open() waiting on a new txg. This can be useful when attempting to determine why a particular workload is under performing. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2012-01-20 22:58:57 +04:00
dmu_tx_stats_t dmu_tx_stats = {
{ "dmu_tx_assigned", KSTAT_DATA_UINT64 },
{ "dmu_tx_delay", KSTAT_DATA_UINT64 },
{ "dmu_tx_error", KSTAT_DATA_UINT64 },
{ "dmu_tx_suspended", KSTAT_DATA_UINT64 },
{ "dmu_tx_group", KSTAT_DATA_UINT64 },
{ "dmu_tx_memory_reserve", KSTAT_DATA_UINT64 },
{ "dmu_tx_memory_reclaim", KSTAT_DATA_UINT64 },
{ "dmu_tx_dirty_throttle", KSTAT_DATA_UINT64 },
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@69962b5647e4a8b9b14998733b765925381b727e Ported-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1913
2013-08-29 07:01:20 +04:00
{ "dmu_tx_dirty_delay", KSTAT_DATA_UINT64 },
{ "dmu_tx_dirty_over_max", KSTAT_DATA_UINT64 },
Introduce kstat dmu_tx_dirty_frees_delay It is helpful to tune zfs_per_txg_dirty_frees_percent for commit 539d33c7(OpenZFS 6569 - large file delete can starve out write ops). Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed by: Richard Elling <Richard.Elling@RichardElling.com> Signed-off-by: Feng Sun <loyou85@gmail.com> Closes #7718
2018-07-25 19:52:27 +03:00
{ "dmu_tx_dirty_frees_delay", KSTAT_DATA_UINT64 },
Add 'dmu_tx' kstats entry Keep counters for the various reasons that a thread may end up in txg_wait_open() waiting on a new txg. This can be useful when attempting to determine why a particular workload is under performing. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2012-01-20 22:58:57 +04:00
{ "dmu_tx_quota", KSTAT_DATA_UINT64 },
};
static kstat_t *dmu_tx_ksp;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
dmu_tx_t *
dmu_tx_create_dd(dsl_dir_t *dd)
{
Change KM_PUSHPAGE -> KM_SLEEP By marking DMU transaction processing contexts with PF_FSTRANS we can revert the KM_PUSHPAGE -> KM_SLEEP changes. This brings us back in line with upstream. In some cases this means simply swapping the flags back. For others fnvlist_alloc() was replaced by nvlist_alloc(..., KM_PUSHPAGE) and must be reverted back to fnvlist_alloc() which assumes KM_SLEEP. The one place KM_PUSHPAGE is kept is when allocating ARC buffers which allows us to dip in to reserved memory. This is again the same as upstream. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-11-21 03:09:39 +03:00
dmu_tx_t *tx = kmem_zalloc(sizeof (dmu_tx_t), KM_SLEEP);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
tx->tx_dir = dd;
Illumos #2882, #2883, #2900 2882 implement libzfs_core 2883 changing "canmount" property to "on" should not always remount dataset 2900 "zfs snapshot" should be able to create multiple, arbitrary snapshots at once Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Chris Siden <christopher.siden@delphix.com> Reviewed by: Garrett D'Amore <garrett@damore.org> Reviewed by: Bill Pijewski <wdp@joyent.com> Reviewed by: Dan Kruchinin <dan.kruchinin@gmail.com> Approved by: Eric Schrock <Eric.Schrock@delphix.com> References: https://www.illumos.org/issues/2882 https://www.illumos.org/issues/2883 https://www.illumos.org/issues/2900 illumos/illumos-gate@4445fffbbb1ea25fd0e9ea68b9380dd7a6709025 Ported-by: Tim Chase <tim@chase2k.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1293 Porting notes: WARNING: This patch changes the user/kernel ABI. That means that the zfs/zpool utilities built from master are NOT compatible with the 0.6.2 kernel modules. Ensure you load the matching kernel modules from master after updating the utilities. Otherwise the zfs/zpool commands will be unable to interact with your pool and you will see errors similar to the following: $ zpool list failed to read pool configuration: bad address no pools available $ zfs list no datasets available Add zvol minor device creation to the new zfs_snapshot_nvl function. Remove the logging of the "release" operation in dsl_dataset_user_release_sync(). The logging caused a null dereference because ds->ds_dir is zeroed in dsl_dataset_destroy_sync() and the logging functions try to get the ds name via the dsl_dataset_name() function. I've got no idea why this particular code would have worked in Illumos. This code has subsequently been completely reworked in Illumos commit 3b2aab1 (3464 zfs synctask code needs restructuring). Squash some "may be used uninitialized" warning/erorrs. Fix some printf format warnings for %lld and %llu. Apply a few spa_writeable() changes that were made to Illumos in illumos/illumos-gate.git@cd1c8b8 as part of the 3112, 3113, 3114 and 3115 fixes. Add a missing call to fnvlist_free(nvl) in log_internal() that was added in Illumos to fix issue 3085 but couldn't be ported to ZoL at the time (zfsonlinux/zfs@9e11c73) because it depended on future work.
2013-08-28 15:45:09 +04:00
if (dd != NULL)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
tx->tx_pool = dd->dd_pool;
list_create(&tx->tx_holds, sizeof (dmu_tx_hold_t),
offsetof(dmu_tx_hold_t, txh_node));
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
list_create(&tx->tx_callbacks, sizeof (dmu_tx_callback_t),
offsetof(dmu_tx_callback_t, dcb_node));
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@69962b5647e4a8b9b14998733b765925381b727e Ported-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1913
2013-08-29 07:01:20 +04:00
tx->tx_start = gethrtime();
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
return (tx);
}
dmu_tx_t *
dmu_tx_create(objset_t *os)
{
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
dmu_tx_t *tx = dmu_tx_create_dd(os->os_dsl_dataset->ds_dir);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
tx->tx_objset = os;
return (tx);
}
dmu_tx_t *
dmu_tx_create_assigned(struct dsl_pool *dp, uint64_t txg)
{
dmu_tx_t *tx = dmu_tx_create_dd(NULL);
Small rework of txg_list code This patch simply adds some missing locking to the txg_list functions and refactors txg_verify() so that it is only compiled in for debug builds. Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Tom Caputi <tcaputi@datto.com> Closes #7795
2018-08-20 23:41:53 +03:00
TXG_VERIFY(dp->dp_spa, txg);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
tx->tx_pool = dp;
tx->tx_txg = txg;
tx->tx_anyobj = TRUE;
return (tx);
}
int
dmu_tx_is_syncing(dmu_tx_t *tx)
{
return (tx->tx_anyobj);
}
int
dmu_tx_private_ok(dmu_tx_t *tx)
{
return (tx->tx_anyobj);
}
static dmu_tx_hold_t *
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
dmu_tx_hold_dnode_impl(dmu_tx_t *tx, dnode_t *dn, enum dmu_tx_hold_type type,
uint64_t arg1, uint64_t arg2)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
{
dmu_tx_hold_t *txh;
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
if (dn != NULL) {
Linux 4.19-rc3+ compat: Remove refcount_t compat torvalds/linux@59b57717f ("blkcg: delay blkg destruction until after writeback has finished") added a refcount_t to the blkcg structure. Due to the refcount_t compatibility code, zfs_refcount_t was used by mistake. Resolve this by removing the compatibility code and replacing the occurrences of refcount_t with zfs_refcount_t. Reviewed-by: Franz Pletz <fpletz@fnordicwalking.de> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Tim Schumacher <timschumi@gmx.de> Closes #7885 Closes #7932
2018-09-26 20:29:26 +03:00
(void) zfs_refcount_add(&dn->dn_holds, tx);
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
if (tx->tx_txg != 0) {
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
mutex_enter(&dn->dn_mtx);
/*
* dn->dn_assigned_txg == tx->tx_txg doesn't pose a
* problem, but there's no way for it to happen (for
* now, at least).
*/
ASSERT(dn->dn_assigned_txg == 0);
dn->dn_assigned_txg = tx->tx_txg;
Linux 4.19-rc3+ compat: Remove refcount_t compat torvalds/linux@59b57717f ("blkcg: delay blkg destruction until after writeback has finished") added a refcount_t to the blkcg structure. Due to the refcount_t compatibility code, zfs_refcount_t was used by mistake. Resolve this by removing the compatibility code and replacing the occurrences of refcount_t with zfs_refcount_t. Reviewed-by: Franz Pletz <fpletz@fnordicwalking.de> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Tim Schumacher <timschumi@gmx.de> Closes #7885 Closes #7932
2018-09-26 20:29:26 +03:00
(void) zfs_refcount_add(&dn->dn_tx_holds, tx);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
mutex_exit(&dn->dn_mtx);
}
}
Change KM_PUSHPAGE -> KM_SLEEP By marking DMU transaction processing contexts with PF_FSTRANS we can revert the KM_PUSHPAGE -> KM_SLEEP changes. This brings us back in line with upstream. In some cases this means simply swapping the flags back. For others fnvlist_alloc() was replaced by nvlist_alloc(..., KM_PUSHPAGE) and must be reverted back to fnvlist_alloc() which assumes KM_SLEEP. The one place KM_PUSHPAGE is kept is when allocating ARC buffers which allows us to dip in to reserved memory. This is again the same as upstream. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-11-21 03:09:39 +03:00
txh = kmem_zalloc(sizeof (dmu_tx_hold_t), KM_SLEEP);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
txh->txh_tx = tx;
txh->txh_dnode = dn;
Prefix all refcount functions with zfs_ Recent changes in the Linux kernel made it necessary to prefix the refcount_add() function with zfs_ due to a name collision. To bring the other functions in line with that and to avoid future collisions, prefix the other refcount functions as well. Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Tim Schumacher <timschumi@gmx.de> Closes #7963
2018-10-01 20:42:05 +03:00
zfs_refcount_create(&txh->txh_space_towrite);
zfs_refcount_create(&txh->txh_memory_tohold);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
txh->txh_type = type;
txh->txh_arg1 = arg1;
txh->txh_arg2 = arg2;
list_insert_tail(&tx->tx_holds, txh);
return (txh);
}
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
static dmu_tx_hold_t *
dmu_tx_hold_object_impl(dmu_tx_t *tx, objset_t *os, uint64_t object,
enum dmu_tx_hold_type type, uint64_t arg1, uint64_t arg2)
{
dnode_t *dn = NULL;
dmu_tx_hold_t *txh;
int err;
if (object != DMU_NEW_OBJECT) {
err = dnode_hold(os, object, FTAG, &dn);
Clean up by-dnode code in dmu_tx.c https://github.com/zfsonlinux/zfs/commit/0eef1bde31d67091d3deed23fe2394f5a8bf2276 introduced some changes which we slightly improved the style of when porting to illumos. There is also one minor error-handling fix, in zap_add() the "zap" may become NULL in case of an error re-opening the ZAP. Originally suggested at: https://github.com/openzfs/openzfs/pull/276 Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Signed-off-by: Matthew Ahrens <mahrens@delphix.com> Closes #5805
2017-02-25 00:34:26 +03:00
if (err != 0) {
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
tx->tx_err = err;
return (NULL);
}
}
txh = dmu_tx_hold_dnode_impl(tx, dn, type, arg1, arg2);
if (dn != NULL)
dnode_rele(dn, FTAG);
return (txh);
}
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
void
Clean up by-dnode code in dmu_tx.c https://github.com/zfsonlinux/zfs/commit/0eef1bde31d67091d3deed23fe2394f5a8bf2276 introduced some changes which we slightly improved the style of when porting to illumos. There is also one minor error-handling fix, in zap_add() the "zap" may become NULL in case of an error re-opening the ZAP. Originally suggested at: https://github.com/openzfs/openzfs/pull/276 Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Signed-off-by: Matthew Ahrens <mahrens@delphix.com> Closes #5805
2017-02-25 00:34:26 +03:00
dmu_tx_add_new_object(dmu_tx_t *tx, dnode_t *dn)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
{
/*
* If we're syncing, they can manipulate any object anyhow, and
* the hold on the dnode_t can cause problems.
*/
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
if (!dmu_tx_is_syncing(tx))
(void) dmu_tx_hold_dnode_impl(tx, dn, THT_NEWOBJECT, 0, 0);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
/*
* This function reads specified data from disk. The specified data will
* be needed to perform the transaction -- i.e, it will be read after
* we do dmu_tx_assign(). There are two reasons that we read the data now
* (before dmu_tx_assign()):
*
* 1. Reading it now has potentially better performance. The transaction
* has not yet been assigned, so the TXG is not held open, and also the
* caller typically has less locks held when calling dmu_tx_hold_*() than
* after the transaction has been assigned. This reduces the lock (and txg)
* hold times, thus reducing lock contention.
*
* 2. It is easier for callers (primarily the ZPL) to handle i/o errors
* that are detected before they start making changes to the DMU state
* (i.e. now). Once the transaction has been assigned, and some DMU
* state has been changed, it can be difficult to recover from an i/o
* error (e.g. to undo the changes already made in memory at the DMU
* layer). Typically code to do so does not exist in the caller -- it
* assumes that the data has already been cached and thus i/o errors are
* not possible.
*
* It has been observed that the i/o initiated here can be a performance
* problem, and it appears to be optional, because we don't look at the
* data which is read. However, removing this read would only serve to
* move the work elsewhere (after the dmu_tx_assign()), where it may
* have a greater impact on performance (in addition to the impact on
* fault tolerance noted above).
*/
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
static int
dmu_tx_check_ioerr(zio_t *zio, dnode_t *dn, int level, uint64_t blkid)
{
int err;
dmu_buf_impl_t *db;
rw_enter(&dn->dn_struct_rwlock, RW_READER);
db = dbuf_hold_level(dn, level, blkid, FTAG);
rw_exit(&dn->dn_struct_rwlock);
if (db == NULL)
Illumos #3598 3598 want to dtrace when errors are generated in zfs Reviewed by: Dan Kimmel <dan.kimmel@delphix.com> Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: Christopher Siden <christopher.siden@delphix.com> Approved by: Garrett D'Amore <garrett@damore.org> References: https://www.illumos.org/issues/3598 illumos/illumos-gate@be6fd75a69ae679453d9cda5bff3326111e6d1ca Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775 Porting notes: 1. include/sys/zfs_context.h has been modified to render some new macros inert until dtrace is available on Linux. 2. Linux-specific changes have been adapted to use SET_ERROR(). 3. I'm NOT happy about this change. It does nothing but ugly up the code under Linux. Unfortunately we need to take it to avoid more merge conflicts in the future. -Brian
2013-03-08 22:41:28 +04:00
return (SET_ERROR(EIO));
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
err = dbuf_read(db, zio, DB_RF_CANFAIL | DB_RF_NOPREFETCH);
dbuf_rele(db, FTAG);
return (err);
}
/* ARGSUSED */
static void
dmu_tx_count_write(dmu_tx_hold_t *txh, uint64_t off, uint64_t len)
{
dnode_t *dn = txh->txh_dnode;
int err = 0;
if (len == 0)
return;
Prefix all refcount functions with zfs_ Recent changes in the Linux kernel made it necessary to prefix the refcount_add() function with zfs_ due to a name collision. To bring the other functions in line with that and to avoid future collisions, prefix the other refcount functions as well. Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Tim Schumacher <timschumi@gmx.de> Closes #7963
2018-10-01 20:42:05 +03:00
(void) zfs_refcount_add_many(&txh->txh_space_towrite, len, FTAG);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
Prefix all refcount functions with zfs_ Recent changes in the Linux kernel made it necessary to prefix the refcount_add() function with zfs_ due to a name collision. To bring the other functions in line with that and to avoid future collisions, prefix the other refcount functions as well. Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Tim Schumacher <timschumi@gmx.de> Closes #7963
2018-10-01 20:42:05 +03:00
if (zfs_refcount_count(&txh->txh_space_towrite) > 2 * DMU_MAX_ACCESS)
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
err = SET_ERROR(EFBIG);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
if (dn == NULL)
return;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
/*
* For i/o error checking, read the blocks that will be needed
* to perform the write: the first and last level-0 blocks (if
* they are not aligned, i.e. if they are partial-block writes),
* and all the level-1 blocks.
*/
if (dn->dn_maxblkid == 0) {
if (off < dn->dn_datablksz &&
(off > 0 || len < dn->dn_datablksz)) {
err = dmu_tx_check_ioerr(NULL, dn, 0, 0);
if (err != 0) {
txh->txh_tx->tx_err = err;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
Rebase master to b117
2009-07-03 02:44:48 +04:00
}
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
} else {
zio_t *zio = zio_root(dn->dn_objset->os_spa,
NULL, NULL, ZIO_FLAG_CANFAIL);
Rebase master to b117
2009-07-03 02:44:48 +04:00
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
/* first level-0 block */
uint64_t start = off >> dn->dn_datablkshift;
if (P2PHASE(off, dn->dn_datablksz) || len < dn->dn_datablksz) {
err = dmu_tx_check_ioerr(zio, dn, 0, start);
if (err != 0) {
txh->txh_tx->tx_err = err;
}
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
}
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
/* last level-0 block */
uint64_t end = (off + len - 1) >> dn->dn_datablkshift;
if (end != start && end <= dn->dn_maxblkid &&
P2PHASE(off + len, dn->dn_datablksz)) {
err = dmu_tx_check_ioerr(zio, dn, 0, end);
if (err != 0) {
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
txh->txh_tx->tx_err = err;
Rebase master to b117
2009-07-03 02:44:48 +04:00
}
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
}
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
/* level-1 blocks */
if (dn->dn_nlevels > 1) {
int shft = dn->dn_indblkshift - SPA_BLKPTRSHIFT;
for (uint64_t i = (start >> shft) + 1;
i < end >> shft; i++) {
err = dmu_tx_check_ioerr(zio, dn, 1, i);
if (err != 0) {
txh->txh_tx->tx_err = err;
}
Rebase master to b117
2009-07-03 02:44:48 +04:00
}
}
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
err = zio_wait(zio);
if (err != 0) {
txh->txh_tx->tx_err = err;
Rebase master to b117
2009-07-03 02:44:48 +04:00
}
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
}
static void
dmu_tx_count_dnode(dmu_tx_hold_t *txh)
{
Prefix all refcount functions with zfs_ Recent changes in the Linux kernel made it necessary to prefix the refcount_add() function with zfs_ due to a name collision. To bring the other functions in line with that and to avoid future collisions, prefix the other refcount functions as well. Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Tim Schumacher <timschumi@gmx.de> Closes #7963
2018-10-01 20:42:05 +03:00
(void) zfs_refcount_add_many(&txh->txh_space_towrite,
DNODE_MIN_SIZE, FTAG);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
void
dmu_tx_hold_write(dmu_tx_t *tx, uint64_t object, uint64_t off, int len)
{
dmu_tx_hold_t *txh;
Clean up by-dnode code in dmu_tx.c https://github.com/zfsonlinux/zfs/commit/0eef1bde31d67091d3deed23fe2394f5a8bf2276 introduced some changes which we slightly improved the style of when porting to illumos. There is also one minor error-handling fix, in zap_add() the "zap" may become NULL in case of an error re-opening the ZAP. Originally suggested at: https://github.com/openzfs/openzfs/pull/276 Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Signed-off-by: Matthew Ahrens <mahrens@delphix.com> Closes #5805
2017-02-25 00:34:26 +03:00
ASSERT0(tx->tx_txg);
ASSERT3U(len, <=, DMU_MAX_ACCESS);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
ASSERT(len == 0 || UINT64_MAX - off >= len - 1);
txh = dmu_tx_hold_object_impl(tx, tx->tx_objset,
object, THT_WRITE, off, len);
Clean up by-dnode code in dmu_tx.c https://github.com/zfsonlinux/zfs/commit/0eef1bde31d67091d3deed23fe2394f5a8bf2276 introduced some changes which we slightly improved the style of when porting to illumos. There is also one minor error-handling fix, in zap_add() the "zap" may become NULL in case of an error re-opening the ZAP. Originally suggested at: https://github.com/openzfs/openzfs/pull/276 Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Signed-off-by: Matthew Ahrens <mahrens@delphix.com> Closes #5805
2017-02-25 00:34:26 +03:00
if (txh != NULL) {
dmu_tx_count_write(txh, off, len);
dmu_tx_count_dnode(txh);
}
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
void
dmu_tx_hold_write_by_dnode(dmu_tx_t *tx, dnode_t *dn, uint64_t off, int len)
{
dmu_tx_hold_t *txh;
Clean up by-dnode code in dmu_tx.c https://github.com/zfsonlinux/zfs/commit/0eef1bde31d67091d3deed23fe2394f5a8bf2276 introduced some changes which we slightly improved the style of when porting to illumos. There is also one minor error-handling fix, in zap_add() the "zap" may become NULL in case of an error re-opening the ZAP. Originally suggested at: https://github.com/openzfs/openzfs/pull/276 Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Signed-off-by: Matthew Ahrens <mahrens@delphix.com> Closes #5805
2017-02-25 00:34:26 +03:00
ASSERT0(tx->tx_txg);
ASSERT3U(len, <=, DMU_MAX_ACCESS);
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
ASSERT(len == 0 || UINT64_MAX - off >= len - 1);
txh = dmu_tx_hold_dnode_impl(tx, dn, THT_WRITE, off, len);
Clean up by-dnode code in dmu_tx.c https://github.com/zfsonlinux/zfs/commit/0eef1bde31d67091d3deed23fe2394f5a8bf2276 introduced some changes which we slightly improved the style of when porting to illumos. There is also one minor error-handling fix, in zap_add() the "zap" may become NULL in case of an error re-opening the ZAP. Originally suggested at: https://github.com/openzfs/openzfs/pull/276 Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Signed-off-by: Matthew Ahrens <mahrens@delphix.com> Closes #5805
2017-02-25 00:34:26 +03:00
if (txh != NULL) {
dmu_tx_count_write(txh, off, len);
dmu_tx_count_dnode(txh);
}
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
}
Illumos 4950 - files sometimes can't be removed from a full filesystem 4950 files sometimes can't be removed from a full filesystem Reviewed by: Adam Leventhal <adam.leventhal@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Sebastien Roy <sebastien.roy@delphix.com> Reviewed by: Boris Protopopov <bprotopopov@hotmail.com> Approved by: Dan McDonald <danmcd@omniti.com> References: https://www.illumos.org/issues/4950 https://github.com/illumos/illumos-gate/commit/4bb7380 Porting notes: - ZoL currently does not log discards to zvols, so the portion of this patch that modifies the discard logging to mark it as freeing space has been discarded. 2. may_delete_now had been removed from zfs_remove() in ZoL. It has been reintroduced. 3. We do not try to emulate vnodes, so the following lines are not valid on Linux: mutex_enter(&vp->v_lock); may_delete_now = vp->v_count == 1 && !vn_has_cached_data(vp); mutex_exit(&vp->v_lock); This has been replaced with: mutex_enter(&zp->z_lock); may_delete_now = atomic_read(&ip->i_count) == 1 && !(zp->z_is_mapped); mutex_exit(&zp->z_lock); Ported-by: Richard Yao <richard.yao@clusterhq.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-07-07 23:49:36 +04:00
/*
* This function marks the transaction as being a "net free". The end
* result is that refquotas will be disabled for this transaction, and
* this transaction will be able to use half of the pool space overhead
* (see dsl_pool_adjustedsize()). Therefore this function should only
* be called for transactions that we expect will not cause a net increase
* in the amount of space used (but it's OK if that is occasionally not true).
*/
void
dmu_tx_mark_netfree(dmu_tx_t *tx)
{
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
tx->tx_netfree = B_TRUE;
Illumos 4950 - files sometimes can't be removed from a full filesystem 4950 files sometimes can't be removed from a full filesystem Reviewed by: Adam Leventhal <adam.leventhal@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Sebastien Roy <sebastien.roy@delphix.com> Reviewed by: Boris Protopopov <bprotopopov@hotmail.com> Approved by: Dan McDonald <danmcd@omniti.com> References: https://www.illumos.org/issues/4950 https://github.com/illumos/illumos-gate/commit/4bb7380 Porting notes: - ZoL currently does not log discards to zvols, so the portion of this patch that modifies the discard logging to mark it as freeing space has been discarded. 2. may_delete_now had been removed from zfs_remove() in ZoL. It has been reintroduced. 3. We do not try to emulate vnodes, so the following lines are not valid on Linux: mutex_enter(&vp->v_lock); may_delete_now = vp->v_count == 1 && !vn_has_cached_data(vp); mutex_exit(&vp->v_lock); This has been replaced with: mutex_enter(&zp->z_lock); may_delete_now = atomic_read(&ip->i_count) == 1 && !(zp->z_is_mapped); mutex_exit(&zp->z_lock); Ported-by: Richard Yao <richard.yao@clusterhq.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-07-07 23:49:36 +04:00
}
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
static void
dmu_tx_hold_free_impl(dmu_tx_hold_t *txh, uint64_t off, uint64_t len)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
{
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
dmu_tx_t *tx = txh->txh_tx;
dnode_t *dn = txh->txh_dnode;
Illumos #3834 3834 incremental replication of 'holey' file systems is slow Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Approved by: Richard Lowe <richlowe@richlowe.net> References: https://www.illumos.org/issues/3834 illumos/illumos-gate@ca48f36f20f6098ceb19d5b084b6b3d4b8eca9fa Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775
2013-07-29 22:58:53 +04:00
int err;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
ASSERT(tx->tx_txg == 0);
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@69962b5647e4a8b9b14998733b765925381b727e Ported-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1913
2013-08-29 07:01:20 +04:00
dmu_tx_count_dnode(txh);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
if (off >= (dn->dn_maxblkid + 1) * dn->dn_datablksz)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
return;
if (len == DMU_OBJECT_END)
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
len = (dn->dn_maxblkid + 1) * dn->dn_datablksz - off;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
Illumos #3834 3834 incremental replication of 'holey' file systems is slow Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Approved by: Richard Lowe <richlowe@richlowe.net> References: https://www.illumos.org/issues/3834 illumos/illumos-gate@ca48f36f20f6098ceb19d5b084b6b3d4b8eca9fa Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775
2013-07-29 22:58:53 +04:00
dmu_tx_count_dnode(txh);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
/*
Illumos #3834 3834 incremental replication of 'holey' file systems is slow Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Approved by: Richard Lowe <richlowe@richlowe.net> References: https://www.illumos.org/issues/3834 illumos/illumos-gate@ca48f36f20f6098ceb19d5b084b6b3d4b8eca9fa Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775
2013-07-29 22:58:53 +04:00
* For i/o error checking, we read the first and last level-0
* blocks if they are not aligned, and all the level-1 blocks.
*
* Note: dbuf_free_range() assumes that we have not instantiated
* any level-0 dbufs that will be completely freed. Therefore we must
* exercise care to not read or count the first and last blocks
* if they are blocksize-aligned.
*/
if (dn->dn_datablkshift == 0) {
Illumos #4047 4047 panic from dbuf_free_range() from dmu_free_object() while doing zfs receive Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Approved by: Dan McDonald <danmcd@nexenta.com> References: https://www.illumos.org/issues/4047 illumos/illumos-gate@713d6c208802cfbb806329ec0d154b641b80c355 Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775 Porting notes: 1. The exported symbol dmu_free_object() was renamed to dmu_free_long_object() in Illumos.
2013-08-21 08:11:52 +04:00
if (off != 0 || len < dn->dn_datablksz)
Illumos #4082 4082 zfs receive gets EFBIG from dmu_tx_hold_free() Reviewed by: Eric Schrock <eric.schrock@delphix.com> Reviewed by: Christopher Siden <christopher.siden@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Approved by: Richard Lowe <richlowe@richlowe.net> References: https://www.illumos.org/issues/4082 illumos/illumos-gate@5253393b09789ec67bec153b866d7285a1cf1645 Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775
2013-08-30 13:19:35 +04:00
dmu_tx_count_write(txh, 0, dn->dn_datablksz);
Illumos #3834 3834 incremental replication of 'holey' file systems is slow Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Approved by: Richard Lowe <richlowe@richlowe.net> References: https://www.illumos.org/issues/3834 illumos/illumos-gate@ca48f36f20f6098ceb19d5b084b6b3d4b8eca9fa Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775
2013-07-29 22:58:53 +04:00
} else {
/* first block will be modified if it is not aligned */
if (!IS_P2ALIGNED(off, 1 << dn->dn_datablkshift))
dmu_tx_count_write(txh, off, 1);
/* last block will be modified if it is not aligned */
if (!IS_P2ALIGNED(off + len, 1 << dn->dn_datablkshift))
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
dmu_tx_count_write(txh, off + len, 1);
Illumos #3834 3834 incremental replication of 'holey' file systems is slow Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Approved by: Richard Lowe <richlowe@richlowe.net> References: https://www.illumos.org/issues/3834 illumos/illumos-gate@ca48f36f20f6098ceb19d5b084b6b3d4b8eca9fa Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775
2013-07-29 22:58:53 +04:00
}
/*
* Check level-1 blocks.
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
*/
if (dn->dn_nlevels > 1) {
Illumos #3834 3834 incremental replication of 'holey' file systems is slow Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Approved by: Richard Lowe <richlowe@richlowe.net> References: https://www.illumos.org/issues/3834 illumos/illumos-gate@ca48f36f20f6098ceb19d5b084b6b3d4b8eca9fa Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775
2013-07-29 22:58:53 +04:00
int shift = dn->dn_datablkshift + dn->dn_indblkshift -
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
SPA_BLKPTRSHIFT;
Illumos #3834 3834 incremental replication of 'holey' file systems is slow Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Approved by: Richard Lowe <richlowe@richlowe.net> References: https://www.illumos.org/issues/3834 illumos/illumos-gate@ca48f36f20f6098ceb19d5b084b6b3d4b8eca9fa Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775
2013-07-29 22:58:53 +04:00
uint64_t start = off >> shift;
uint64_t end = (off + len) >> shift;
ASSERT(dn->dn_indblkshift != 0);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
4188 assertion failed in dmu_tx_hold_free(): dn_datablkshift != 0 Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Christopher Siden <christopher.siden@delphix.com> Approved by: Garrett D'Amore <garrett@damore.org> Refences: https://www.illumos.org/issues/4188 illumos/illumos-gate@bb411a08b05466bfe0c7095b6373bbc1587e259a Ported-by: Chris Dunlop <chris@onthe.net.au> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #2091
2014-01-31 19:52:11 +04:00
/*
* dnode_reallocate() can result in an object with indirect
* blocks having an odd data block size. In this case,
* just check the single block.
*/
if (dn->dn_datablkshift == 0)
start = end = 0;
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
zio_t *zio = zio_root(tx->tx_pool->dp_spa,
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
NULL, NULL, ZIO_FLAG_CANFAIL);
Undo c89 workarounds to match with upstream With PR 5756 the zfs module now supports c99 and the remaining past c89 workarounds can be undone. Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed-by: George Melikov <mail@gmelikov.ru> Signed-off-by: Don Brady <don.brady@delphix.com> Closes #6816
2017-11-04 23:25:13 +03:00
for (uint64_t i = start; i <= end; i++) {
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
uint64_t ibyte = i << shift;
Rebase to OpenSolaris b103, in the process we are removing any code which did not originate from the OpenSolaris source. These changes will be reintroduced in topic branches for easier tracking
2008-12-03 23:09:06 +03:00
err = dnode_next_offset(dn, 0, &ibyte, 2, 1, 0);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
i = ibyte >> shift;
Illumos 5911 - ZFS "hangs" while deleting file 5911 ZFS "hangs" while deleting file Reviewed by: Bayard Bell <buffer.g.overflow@gmail.com> Reviewed by: Alek Pinchuk <alek@nexenta.com> Reviewed by: Simon Klinkert <simon.klinkert@gmail.com> Reviewed by: Dan McDonald <danmcd@omniti.com> Approved by: Richard Lowe <richlowe@richlowe.net> References: https://www.illumos.org/issues/5911 https://github.com/illumos/illumos-gate/commit/46e1baa Porting notes: Resolved ISO C90 forbids mixed declarations and code wanting in the dnode_free_range() function. Ported-by: kernelOfTruth kerneloftruth@gmail.com Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #3554
2015-07-02 19:23:20 +03:00
if (err == ESRCH || i > end)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
break;
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
if (err != 0) {
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
tx->tx_err = err;
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
(void) zio_wait(zio);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
return;
}
Prefix all refcount functions with zfs_ Recent changes in the Linux kernel made it necessary to prefix the refcount_add() function with zfs_ due to a name collision. To bring the other functions in line with that and to avoid future collisions, prefix the other refcount functions as well. Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Tim Schumacher <timschumi@gmx.de> Closes #7963
2018-10-01 20:42:05 +03:00
(void) zfs_refcount_add_many(&txh->txh_memory_tohold,
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
1 << dn->dn_indblkshift, FTAG);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
err = dmu_tx_check_ioerr(zio, dn, 1, i);
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
if (err != 0) {
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
tx->tx_err = err;
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
(void) zio_wait(zio);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
return;
}
}
err = zio_wait(zio);
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
if (err != 0) {
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
tx->tx_err = err;
return;
}
}
}
void
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
dmu_tx_hold_free(dmu_tx_t *tx, uint64_t object, uint64_t off, uint64_t len)
{
dmu_tx_hold_t *txh;
txh = dmu_tx_hold_object_impl(tx, tx->tx_objset,
object, THT_FREE, off, len);
Clean up by-dnode code in dmu_tx.c https://github.com/zfsonlinux/zfs/commit/0eef1bde31d67091d3deed23fe2394f5a8bf2276 introduced some changes which we slightly improved the style of when porting to illumos. There is also one minor error-handling fix, in zap_add() the "zap" may become NULL in case of an error re-opening the ZAP. Originally suggested at: https://github.com/openzfs/openzfs/pull/276 Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Signed-off-by: Matthew Ahrens <mahrens@delphix.com> Closes #5805
2017-02-25 00:34:26 +03:00
if (txh != NULL)
(void) dmu_tx_hold_free_impl(txh, off, len);
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
}
void
dmu_tx_hold_free_by_dnode(dmu_tx_t *tx, dnode_t *dn, uint64_t off, uint64_t len)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
{
dmu_tx_hold_t *txh;
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
txh = dmu_tx_hold_dnode_impl(tx, dn, THT_FREE, off, len);
Clean up by-dnode code in dmu_tx.c https://github.com/zfsonlinux/zfs/commit/0eef1bde31d67091d3deed23fe2394f5a8bf2276 introduced some changes which we slightly improved the style of when porting to illumos. There is also one minor error-handling fix, in zap_add() the "zap" may become NULL in case of an error re-opening the ZAP. Originally suggested at: https://github.com/openzfs/openzfs/pull/276 Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Signed-off-by: Matthew Ahrens <mahrens@delphix.com> Closes #5805
2017-02-25 00:34:26 +03:00
if (txh != NULL)
(void) dmu_tx_hold_free_impl(txh, off, len);
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
}
static void
OpenZFS 7801 - add more by-dnode routines (lint) Authored by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: George Melikov <mail@gmelikov.ru> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Ported-by: Giuseppe Di Natale <dinatale2@llnl.gov> OpenZFS-issue: https://www.illumos.org/issues/7801 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/f25efb3 Closes #5894
2017-03-15 15:49:59 +03:00
dmu_tx_hold_zap_impl(dmu_tx_hold_t *txh, const char *name)
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
{
dmu_tx_t *tx = txh->txh_tx;
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
dnode_t *dn = txh->txh_dnode;
OpenZFS 7054 - dmu_tx_hold_t should use refcount_t to track space Authored by: Igor Kozhukhov ikozhukhov@gmail.com Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Paul Dagnelie <pcd@delphix.com> Reviewed by: Igor Kozhukhov <ikozhukhov@gmail.com> Approved by: Dan McDonald <danmcd@omniti.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Ported-by: George Melikov mail@gmelikov.ru OpenZFS-issue: https://www.illumos.org/issues/7054 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/0c779ad Closes #5600
2017-01-23 20:36:24 +03:00
int err;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
ASSERT(tx->tx_txg == 0);
dmu_tx_count_dnode(txh);
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
/*
* Modifying a almost-full microzap is around the worst case (128KB)
*
* If it is a fat zap, the worst case would be 7*16KB=112KB:
* - 3 blocks overwritten: target leaf, ptrtbl block, header block
* - 4 new blocks written if adding:
* - 2 blocks for possibly split leaves,
* - 2 grown ptrtbl blocks
*/
Prefix all refcount functions with zfs_ Recent changes in the Linux kernel made it necessary to prefix the refcount_add() function with zfs_ due to a name collision. To bring the other functions in line with that and to avoid future collisions, prefix the other refcount functions as well. Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Tim Schumacher <timschumi@gmx.de> Closes #7963
2018-10-01 20:42:05 +03:00
(void) zfs_refcount_add_many(&txh->txh_space_towrite,
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
MZAP_MAX_BLKSZ, FTAG);
if (dn == NULL)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
return;
Illumos #2619 and #2747 2619 asynchronous destruction of ZFS file systems 2747 SPA versioning with zfs feature flags Reviewed by: Matt Ahrens <mahrens@delphix.com> Reviewed by: George Wilson <gwilson@delphix.com> Reviewed by: Richard Lowe <richlowe@richlowe.net> Reviewed by: Dan Kruchinin <dan.kruchinin@gmail.com> Approved by: Eric Schrock <Eric.Schrock@delphix.com> References: illumos/illumos-gate@53089ab7c84db6fb76c16ca50076c147cda11757 illumos/illumos-gate@ad135b5d644628e791c3188a6ecbd9c257961ef8 illumos changeset: 13700:2889e2596bd6 https://www.illumos.org/issues/2619 https://www.illumos.org/issues/2747 NOTE: The grub specific changes were not ported. This change must be made to the Linux grub packages. Ported-by: Brian Behlendorf <behlendorf1@llnl.gov>
2012-12-14 03:24:15 +04:00
ASSERT3U(DMU_OT_BYTESWAP(dn->dn_type), ==, DMU_BSWAP_ZAP);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
if (dn->dn_maxblkid == 0 || name == NULL) {
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
/*
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
* This is a microzap (only one block), or we don't know
* the name. Check the first block for i/o errors.
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
*/
err = dmu_tx_check_ioerr(NULL, dn, 0, 0);
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
if (err != 0) {
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
tx->tx_err = err;
OpenZFS 7054 - dmu_tx_hold_t should use refcount_t to track space Authored by: Igor Kozhukhov ikozhukhov@gmail.com Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Paul Dagnelie <pcd@delphix.com> Reviewed by: Igor Kozhukhov <ikozhukhov@gmail.com> Approved by: Dan McDonald <danmcd@omniti.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Ported-by: George Melikov mail@gmelikov.ru OpenZFS-issue: https://www.illumos.org/issues/7054 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/0c779ad Closes #5600
2017-01-23 20:36:24 +03:00
}
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
} else {
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
/*
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
* Access the name so that we'll check for i/o errors to
* the leaf blocks, etc. We ignore ENOENT, as this name
* may not yet exist.
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
*/
OpenZFS 7004 - dmu_tx_hold_zap() does dnode_hold() 7x on same object Using a benchmark which has 32 threads creating 2 million files in the same directory, on a machine with 16 CPU cores, I observed poor performance. I noticed that dmu_tx_hold_zap() was using about 30% of all CPU, and doing dnode_hold() 7 times on the same object (the ZAP object that is being held). dmu_tx_hold_zap() keeps a hold on the dnode_t the entire time it is running, in dmu_tx_hold_t:txh_dnode, so it would be nice to use the dnode_t that we already have in hand, rather than repeatedly calling dnode_hold(). To do this, we need to pass the dnode_t down through all the intermediate calls that dmu_tx_hold_zap() makes, making these routines take the dnode_t* rather than an objset_t* and a uint64_t object number. In particular, the following routines will need to have analogous *_by_dnode() variants created: dmu_buf_hold_noread() dmu_buf_hold() zap_lookup() zap_lookup_norm() zap_count_write() zap_lockdir() zap_count_write() This can improve performance on the benchmark described above by 100%, from 30,000 file creations per second to 60,000. (This improvement is on top of that provided by working around the object allocation issue. Peak performance of ~90,000 creations per second was observed with 8 CPUs; adding CPUs past that decreased performance due to lock contention.) The CPU used by dmu_tx_hold_zap() was reduced by 88%, from 340 CPU-seconds to 40 CPU-seconds. Sponsored by: Intel Corp. Signed-off-by: Matthew Ahrens <mahrens@delphix.com> Signed-off-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> OpenZFS-issue: https://www.illumos.org/issues/7004 OpenZFS-commit: https://github.com/openzfs/openzfs/pull/109 Closes #4641 Closes #4972
2016-07-21 01:42:13 +03:00
err = zap_lookup_by_dnode(dn, name, 8, 0, NULL);
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
if (err == EIO || err == ECKSUM || err == ENXIO) {
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
tx->tx_err = err;
OpenZFS 7054 - dmu_tx_hold_t should use refcount_t to track space Authored by: Igor Kozhukhov ikozhukhov@gmail.com Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Paul Dagnelie <pcd@delphix.com> Reviewed by: Igor Kozhukhov <ikozhukhov@gmail.com> Approved by: Dan McDonald <danmcd@omniti.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Ported-by: George Melikov mail@gmelikov.ru OpenZFS-issue: https://www.illumos.org/issues/7054 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/0c779ad Closes #5600
2017-01-23 20:36:24 +03:00
}
}
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
void
dmu_tx_hold_zap(dmu_tx_t *tx, uint64_t object, int add, const char *name)
{
dmu_tx_hold_t *txh;
Clean up by-dnode code in dmu_tx.c https://github.com/zfsonlinux/zfs/commit/0eef1bde31d67091d3deed23fe2394f5a8bf2276 introduced some changes which we slightly improved the style of when porting to illumos. There is also one minor error-handling fix, in zap_add() the "zap" may become NULL in case of an error re-opening the ZAP. Originally suggested at: https://github.com/openzfs/openzfs/pull/276 Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Signed-off-by: Matthew Ahrens <mahrens@delphix.com> Closes #5805
2017-02-25 00:34:26 +03:00
ASSERT0(tx->tx_txg);
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
txh = dmu_tx_hold_object_impl(tx, tx->tx_objset,
object, THT_ZAP, add, (uintptr_t)name);
Clean up by-dnode code in dmu_tx.c https://github.com/zfsonlinux/zfs/commit/0eef1bde31d67091d3deed23fe2394f5a8bf2276 introduced some changes which we slightly improved the style of when porting to illumos. There is also one minor error-handling fix, in zap_add() the "zap" may become NULL in case of an error re-opening the ZAP. Originally suggested at: https://github.com/openzfs/openzfs/pull/276 Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Signed-off-by: Matthew Ahrens <mahrens@delphix.com> Closes #5805
2017-02-25 00:34:26 +03:00
if (txh != NULL)
OpenZFS 7801 - add more by-dnode routines (lint) Authored by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: George Melikov <mail@gmelikov.ru> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Ported-by: Giuseppe Di Natale <dinatale2@llnl.gov> OpenZFS-issue: https://www.illumos.org/issues/7801 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/f25efb3 Closes #5894
2017-03-15 15:49:59 +03:00
dmu_tx_hold_zap_impl(txh, name);
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
}
void
dmu_tx_hold_zap_by_dnode(dmu_tx_t *tx, dnode_t *dn, int add, const char *name)
{
dmu_tx_hold_t *txh;
Clean up by-dnode code in dmu_tx.c https://github.com/zfsonlinux/zfs/commit/0eef1bde31d67091d3deed23fe2394f5a8bf2276 introduced some changes which we slightly improved the style of when porting to illumos. There is also one minor error-handling fix, in zap_add() the "zap" may become NULL in case of an error re-opening the ZAP. Originally suggested at: https://github.com/openzfs/openzfs/pull/276 Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Signed-off-by: Matthew Ahrens <mahrens@delphix.com> Closes #5805
2017-02-25 00:34:26 +03:00
ASSERT0(tx->tx_txg);
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
ASSERT(dn != NULL);
txh = dmu_tx_hold_dnode_impl(tx, dn, THT_ZAP, add, (uintptr_t)name);
Clean up by-dnode code in dmu_tx.c https://github.com/zfsonlinux/zfs/commit/0eef1bde31d67091d3deed23fe2394f5a8bf2276 introduced some changes which we slightly improved the style of when porting to illumos. There is also one minor error-handling fix, in zap_add() the "zap" may become NULL in case of an error re-opening the ZAP. Originally suggested at: https://github.com/openzfs/openzfs/pull/276 Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Signed-off-by: Matthew Ahrens <mahrens@delphix.com> Closes #5805
2017-02-25 00:34:26 +03:00
if (txh != NULL)
OpenZFS 7801 - add more by-dnode routines (lint) Authored by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: George Melikov <mail@gmelikov.ru> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Ported-by: Giuseppe Di Natale <dinatale2@llnl.gov> OpenZFS-issue: https://www.illumos.org/issues/7801 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/f25efb3 Closes #5894
2017-03-15 15:49:59 +03:00
dmu_tx_hold_zap_impl(txh, name);
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
}
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
void
dmu_tx_hold_bonus(dmu_tx_t *tx, uint64_t object)
{
dmu_tx_hold_t *txh;
ASSERT(tx->tx_txg == 0);
txh = dmu_tx_hold_object_impl(tx, tx->tx_objset,
object, THT_BONUS, 0, 0);
if (txh)
dmu_tx_count_dnode(txh);
}
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
void
dmu_tx_hold_bonus_by_dnode(dmu_tx_t *tx, dnode_t *dn)
{
dmu_tx_hold_t *txh;
Clean up by-dnode code in dmu_tx.c https://github.com/zfsonlinux/zfs/commit/0eef1bde31d67091d3deed23fe2394f5a8bf2276 introduced some changes which we slightly improved the style of when porting to illumos. There is also one minor error-handling fix, in zap_add() the "zap" may become NULL in case of an error re-opening the ZAP. Originally suggested at: https://github.com/openzfs/openzfs/pull/276 Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Signed-off-by: Matthew Ahrens <mahrens@delphix.com> Closes #5805
2017-02-25 00:34:26 +03:00
ASSERT0(tx->tx_txg);
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
txh = dmu_tx_hold_dnode_impl(tx, dn, THT_BONUS, 0, 0);
if (txh)
dmu_tx_count_dnode(txh);
}
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
void
dmu_tx_hold_space(dmu_tx_t *tx, uint64_t space)
{
dmu_tx_hold_t *txh;
dmu_tx: Fix possible NULL pointer dereference dmu_tx_hold_object_impl can return NULL on error. Check for this condition prior to dereferencing pointer. This can only occur if the passed object was invalid or unallocated. Signed-off-by: Nathaniel Clark <Nathaniel.Clark@misrule.us> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1610
2013-07-23 21:32:57 +04:00
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
ASSERT(tx->tx_txg == 0);
txh = dmu_tx_hold_object_impl(tx, tx->tx_objset,
DMU_NEW_OBJECT, THT_SPACE, space, 0);
Prefix all refcount functions with zfs_ Recent changes in the Linux kernel made it necessary to prefix the refcount_add() function with zfs_ due to a name collision. To bring the other functions in line with that and to avoid future collisions, prefix the other refcount functions as well. Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Tim Schumacher <timschumi@gmx.de> Closes #7963
2018-10-01 20:42:05 +03:00
if (txh) {
(void) zfs_refcount_add_many(
&txh->txh_space_towrite, space, FTAG);
}
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
#ifdef ZFS_DEBUG
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
void
dmu_tx_dirty_buf(dmu_tx_t *tx, dmu_buf_impl_t *db)
{
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
boolean_t match_object = B_FALSE;
boolean_t match_offset = B_FALSE;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
DB_DNODE_ENTER(db);
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
dnode_t *dn = DB_DNODE(db);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
ASSERT(tx->tx_txg != 0);
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
ASSERT(tx->tx_objset == NULL || dn->dn_objset == tx->tx_objset);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
ASSERT3U(dn->dn_object, ==, db->db.db_object);
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
if (tx->tx_anyobj) {
DB_DNODE_EXIT(db);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
return;
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
}
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
/* XXX No checking on the meta dnode for now */
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
if (db->db.db_object == DMU_META_DNODE_OBJECT) {
DB_DNODE_EXIT(db);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
return;
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
}
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
for (dmu_tx_hold_t *txh = list_head(&tx->tx_holds); txh != NULL;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
txh = list_next(&tx->tx_holds, txh)) {
Enhance a dmu_tx_dirty_buf() assertion The following assertion is good to validate the correctness of new DMU consumers, but it doesn't quite provide enough information. Slightly rework the assertion so that when it is hit the actual offending values will be included in the output. SPLError: 4787:0:(dmu_tx.c:828:dmu_tx_dirty_buf()) ASSERTION(dn == NULL || dn->dn_assigned_txg == tx->tx_txg) failed Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2012-03-12 23:38:00 +04:00
ASSERT3U(dn->dn_assigned_txg, ==, tx->tx_txg);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
if (txh->txh_dnode == dn && txh->txh_type != THT_NEWOBJECT)
match_object = TRUE;
if (txh->txh_dnode == NULL || txh->txh_dnode == dn) {
int datablkshift = dn->dn_datablkshift ?
dn->dn_datablkshift : SPA_MAXBLOCKSHIFT;
int epbs = dn->dn_indblkshift - SPA_BLKPTRSHIFT;
int shift = datablkshift + epbs * db->db_level;
uint64_t beginblk = shift >= 64 ? 0 :
(txh->txh_arg1 >> shift);
uint64_t endblk = shift >= 64 ? 0 :
((txh->txh_arg1 + txh->txh_arg2 - 1) >> shift);
uint64_t blkid = db->db_blkid;
/* XXX txh_arg2 better not be zero... */
dprintf("found txh type %x beginblk=%llx endblk=%llx\n",
txh->txh_type, beginblk, endblk);
switch (txh->txh_type) {
case THT_WRITE:
if (blkid >= beginblk && blkid <= endblk)
match_offset = TRUE;
/*
* We will let this hold work for the bonus
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
* or spill buffer so that we don't need to
* hold it when creating a new object.
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
*/
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
if (blkid == DMU_BONUS_BLKID ||
blkid == DMU_SPILL_BLKID)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
match_offset = TRUE;
/*
* They might have to increase nlevels,
* thus dirtying the new TLIBs. Or the
* might have to change the block size,
* thus dirying the new lvl=0 blk=0.
*/
if (blkid == 0)
match_offset = TRUE;
break;
case THT_FREE:
Rebase to OpenSolaris b103, in the process we are removing any code which did not originate from the OpenSolaris source. These changes will be reintroduced in topic branches for easier tracking
2008-12-03 23:09:06 +03:00
/*
* We will dirty all the level 1 blocks in
* the free range and perhaps the first and
* last level 0 block.
*/
if (blkid >= beginblk && (blkid <= endblk ||
txh->txh_arg2 == DMU_OBJECT_END))
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
match_offset = TRUE;
break;
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
case THT_SPILL:
if (blkid == DMU_SPILL_BLKID)
match_offset = TRUE;
break;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
case THT_BONUS:
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
if (blkid == DMU_BONUS_BLKID)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
match_offset = TRUE;
break;
case THT_ZAP:
match_offset = TRUE;
break;
case THT_NEWOBJECT:
match_object = TRUE;
break;
default:
Change ASSERT(!"...") to cmn_err(CE_PANIC, ...) There are a handful of ASSERT(!"...")'s throughout the code base for cases which should be impossible. This patch converts them to use cmn_err(CE_PANIC, ...) to ensure they are always enabled and so that additional debugging is logged if they were to occur. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1445
2015-02-27 23:53:35 +03:00
cmn_err(CE_PANIC, "bad txh_type %d",
txh->txh_type);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
}
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
if (match_object && match_offset) {
DB_DNODE_EXIT(db);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
return;
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
}
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
DB_DNODE_EXIT(db);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
panic("dirtying dbuf obj=%llx lvl=%u blkid=%llx but not tx_held\n",
(u_longlong_t)db->db.db_object, db->db_level,
(u_longlong_t)db->db_blkid);
}
#endif
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@69962b5647e4a8b9b14998733b765925381b727e Ported-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1913
2013-08-29 07:01:20 +04:00
/*
* If we can't do 10 iops, something is wrong. Let us go ahead
* and hit zfs_dirty_data_max.
*/
hrtime_t zfs_delay_max_ns = 100 * MICROSEC; /* 100 milliseconds */
int zfs_delay_resolution_ns = 100 * 1000; /* 100 microseconds */
/*
* We delay transactions when we've determined that the backend storage
* isn't able to accommodate the rate of incoming writes.
*
* If there is already a transaction waiting, we delay relative to when
* that transaction finishes waiting. This way the calculated min_time
* is independent of the number of threads concurrently executing
* transactions.
*
* If we are the only waiter, wait relative to when the transaction
* started, rather than the current time. This credits the transaction for
* "time already served", e.g. reading indirect blocks.
*
* The minimum time for a transaction to take is calculated as:
* min_time = scale * (dirty - min) / (max - dirty)
* min_time is then capped at zfs_delay_max_ns.
*
* The delay has two degrees of freedom that can be adjusted via tunables.
* The percentage of dirty data at which we start to delay is defined by
* zfs_delay_min_dirty_percent. This should typically be at or above
* zfs_vdev_async_write_active_max_dirty_percent so that we only start to
* delay after writing at full speed has failed to keep up with the incoming
* write rate. The scale of the curve is defined by zfs_delay_scale. Roughly
* speaking, this variable determines the amount of delay at the midpoint of
* the curve.
*
* delay
* 10ms +-------------------------------------------------------------*+
* | *|
* 9ms + *+
* | *|
* 8ms + *+
* | * |
* 7ms + * +
* | * |
* 6ms + * +
* | * |
* 5ms + * +
* | * |
* 4ms + * +
* | * |
* 3ms + * +
* | * |
* 2ms + (midpoint) * +
* | | ** |
* 1ms + v *** +
* | zfs_delay_scale ----------> ******** |
* 0 +-------------------------------------*********----------------+
* 0% <- zfs_dirty_data_max -> 100%
*
* Note that since the delay is added to the outstanding time remaining on the
* most recent transaction, the delay is effectively the inverse of IOPS.
* Here the midpoint of 500us translates to 2000 IOPS. The shape of the curve
* was chosen such that small changes in the amount of accumulated dirty data
* in the first 3/4 of the curve yield relatively small differences in the
* amount of delay.
*
* The effects can be easier to understand when the amount of delay is
* represented on a log scale:
*
* delay
* 100ms +-------------------------------------------------------------++
* + +
* | |
* + *+
* 10ms + *+
* + ** +
* | (midpoint) ** |
* + | ** +
* 1ms + v **** +
* + zfs_delay_scale ----------> ***** +
* | **** |
* + **** +
* 100us + ** +
* + * +
* | * |
* + * +
* 10us + * +
* + +
* | |
* + +
* +--------------------------------------------------------------+
* 0% <- zfs_dirty_data_max -> 100%
*
* Note here that only as the amount of dirty data approaches its limit does
* the delay start to increase rapidly. The goal of a properly tuned system
* should be to keep the amount of dirty data out of that range by first
* ensuring that the appropriate limits are set for the I/O scheduler to reach
* optimal throughput on the backend storage, and then by changing the value
* of zfs_delay_scale to increase the steepness of the curve.
*/
static void
dmu_tx_delay(dmu_tx_t *tx, uint64_t dirty)
{
dsl_pool_t *dp = tx->tx_pool;
uint64_t delay_min_bytes =
zfs_dirty_data_max * zfs_delay_min_dirty_percent / 100;
hrtime_t wakeup, min_tx_time, now;
if (dirty <= delay_min_bytes)
return;
/*
* The caller has already waited until we are under the max.
* We make them pass us the amount of dirty data so we don't
* have to handle the case of it being >= the max, which could
* cause a divide-by-zero if it's == the max.
*/
ASSERT3U(dirty, <, zfs_dirty_data_max);
now = gethrtime();
min_tx_time = zfs_delay_scale *
(dirty - delay_min_bytes) / (zfs_dirty_data_max - dirty);
min_tx_time = MIN(min_tx_time, zfs_delay_max_ns);
if (now > tx->tx_start + min_tx_time)
return;
DTRACE_PROBE3(delay__mintime, dmu_tx_t *, tx, uint64_t, dirty,
uint64_t, min_tx_time);
mutex_enter(&dp->dp_lock);
wakeup = MAX(tx->tx_start + min_tx_time,
dp->dp_last_wakeup + min_tx_time);
dp->dp_last_wakeup = wakeup;
mutex_exit(&dp->dp_lock);
zfs_sleep_until(wakeup);
}
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
/*
* This routine attempts to assign the transaction to a transaction group.
* To do so, we must determine if there is sufficient free space on disk.
*
* If this is a "netfree" transaction (i.e. we called dmu_tx_mark_netfree()
* on it), then it is assumed that there is sufficient free space,
* unless there's insufficient slop space in the pool (see the comment
* above spa_slop_shift in spa_misc.c).
*
* If it is not a "netfree" transaction, then if the data already on disk
* is over the allowed usage (e.g. quota), this will fail with EDQUOT or
* ENOSPC. Otherwise, if the current rough estimate of pending changes,
* plus the rough estimate of this transaction's changes, may exceed the
* allowed usage, then this will fail with ERESTART, which will cause the
* caller to wait for the pending changes to be written to disk (by waiting
* for the next TXG to open), and then check the space usage again.
*
* The rough estimate of pending changes is comprised of the sum of:
*
* - this transaction's holds' txh_space_towrite
*
* - dd_tempreserved[], which is the sum of in-flight transactions'
* holds' txh_space_towrite (i.e. those transactions that have called
* dmu_tx_assign() but not yet called dmu_tx_commit()).
*
* - dd_space_towrite[], which is the amount of dirtied dbufs.
*
* Note that all of these values are inflated by spa_get_worst_case_asize(),
* which means that we may get ERESTART well before we are actually in danger
* of running out of space, but this also mitigates any small inaccuracies
* in the rough estimate (e.g. txh_space_towrite doesn't take into account
* indirect blocks, and dd_space_towrite[] doesn't take into account changes
* to the MOS).
*
* Note that due to this algorithm, it is possible to exceed the allowed
* usage by one transaction. Also, as we approach the allowed usage,
* we will allow a very limited amount of changes into each TXG, thus
* decreasing performance.
*/
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
static int
OpenZFS 8997 - ztest assertion failure in zil_lwb_write_issue PROBLEM ======= When `dmu_tx_assign` is called from `zil_lwb_write_issue`, it's possible for either `ERESTART` or `EIO` to be returned. If `ERESTART` is returned, this will cause an assertion to fail directly in `zil_lwb_write_issue`, where the code assumes the return value is `EIO` if `dmu_tx_assign` returns a non-zero value. This can occur if the SPA is suspended when `dmu_tx_assign` is called, and most often occurs when running `zloop`. If `EIO` is returned, this can cause assertions to fail elsewhere in the ZIL code. For example, `zil_commit_waiter_timeout` contains the following logic: lwb_t *nlwb = zil_lwb_write_issue(zilog, lwb); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_OPENED); In this case, if `dmu_tx_assign` returned `EIO` from within `zil_lwb_write_issue`, the `lwb` variable passed in will not be issued to disk. Thus, it's `lwb_state` field will remain `LWB_STATE_OPENED` and this assertion will fail. `zil_commit_waiter_timeout` assumes that after it calls `zil_lwb_write_issue`, the `lwb` will be issued to disk, and doesn't handle the case where this is not true; i.e. it doesn't handle the case where `dmu_tx_assign` returns `EIO`. SOLUTION ======== This change modifies the `dmu_tx_assign` function such that `txg_how` is a bitmask, rather than of the `txg_how_t` enum type. Now, the previous `TXG_WAITED` semantics can be used via `TXG_NOTHROTTLE`, along with specifying either `TXG_NOWAIT` or `TXG_WAIT` semantics. Previously, when `TXG_WAITED` was specified, `TXG_NOWAIT` semantics was automatically invoked. This was not ideal when using `TXG_WAITED` within `zil_lwb_write_issued`, leading the problem described above. Rather, we want to achieve the semantics of `TXG_WAIT`, while also preventing the `tx` from being penalized via the dirty delay throttling. With this change, `zil_lwb_write_issued` can acheive the semtantics that it requires by passing in the value `TXG_WAIT | TXG_NOTHROTTLE` to `dmu_tx_assign`. Further, consumers of `dmu_tx_assign` wishing to achieve the old `TXG_WAITED` semantics can pass in the value `TXG_NOWAIT | TXG_NOTHROTTLE`. Authored by: Prakash Surya <prakash.surya@delphix.com> Approved by: Robert Mustacchi <rm@joyent.com> Reviewed by: Matt Ahrens <mahrens@delphix.com> Reviewed by: Andriy Gapon <avg@FreeBSD.org> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Porting Notes: - Additionally updated `zfs_tmpfile` to use `TXG_NOTHROTTLE` OpenZFS-issue: https://www.illumos.org/issues/8997 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/19ea6cb0f9 Closes #7084
2018-01-09 00:45:53 +03:00
dmu_tx_try_assign(dmu_tx_t *tx, uint64_t txg_how)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
{
spa_t *spa = tx->tx_pool->dp_spa;
Illumos #3006 3006 VERIFY[S,U,P] and ASSERT[S,U,P] frequently check if first argument is zero Reviewed by Matt Ahrens <matthew.ahrens@delphix.com> Reviewed by George Wilson <george.wilson@delphix.com> Approved by Eric Schrock <eric.schrock@delphix.com> References: illumos/illumos-gate@fb09f5aad449c97fe309678f3f604982b563a96f https://illumos.org/issues/3006 Requires: zfsonlinux/spl@1c6d149feb4033e4a56fb987004edc5d45288bcb Ported-by: Tim Chase <tim@chase2k.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1509
2013-05-11 01:17:03 +04:00
ASSERT0(tx->tx_txg);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
Add 'dmu_tx' kstats entry Keep counters for the various reasons that a thread may end up in txg_wait_open() waiting on a new txg. This can be useful when attempting to determine why a particular workload is under performing. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2012-01-20 22:58:57 +04:00
if (tx->tx_err) {
DMU_TX_STAT_BUMP(dmu_tx_error);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
return (tx->tx_err);
Add 'dmu_tx' kstats entry Keep counters for the various reasons that a thread may end up in txg_wait_open() waiting on a new txg. This can be useful when attempting to determine why a particular workload is under performing. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2012-01-20 22:58:57 +04:00
}
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
Rebase to OpenSolaris b103, in the process we are removing any code which did not originate from the OpenSolaris source. These changes will be reintroduced in topic branches for easier tracking
2008-12-03 23:09:06 +03:00
if (spa_suspended(spa)) {
Add 'dmu_tx' kstats entry Keep counters for the various reasons that a thread may end up in txg_wait_open() waiting on a new txg. This can be useful when attempting to determine why a particular workload is under performing. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2012-01-20 22:58:57 +04:00
DMU_TX_STAT_BUMP(dmu_tx_suspended);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
/*
* If the user has indicated a blocking failure mode
* then return ERESTART which will block in dmu_tx_wait().
* Otherwise, return EIO so that an error can get
* propagated back to the VOP calls.
*
* Note that we always honor the txg_how flag regardless
* of the failuremode setting.
*/
if (spa_get_failmode(spa) == ZIO_FAILURE_MODE_CONTINUE &&
OpenZFS 8997 - ztest assertion failure in zil_lwb_write_issue PROBLEM ======= When `dmu_tx_assign` is called from `zil_lwb_write_issue`, it's possible for either `ERESTART` or `EIO` to be returned. If `ERESTART` is returned, this will cause an assertion to fail directly in `zil_lwb_write_issue`, where the code assumes the return value is `EIO` if `dmu_tx_assign` returns a non-zero value. This can occur if the SPA is suspended when `dmu_tx_assign` is called, and most often occurs when running `zloop`. If `EIO` is returned, this can cause assertions to fail elsewhere in the ZIL code. For example, `zil_commit_waiter_timeout` contains the following logic: lwb_t *nlwb = zil_lwb_write_issue(zilog, lwb); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_OPENED); In this case, if `dmu_tx_assign` returned `EIO` from within `zil_lwb_write_issue`, the `lwb` variable passed in will not be issued to disk. Thus, it's `lwb_state` field will remain `LWB_STATE_OPENED` and this assertion will fail. `zil_commit_waiter_timeout` assumes that after it calls `zil_lwb_write_issue`, the `lwb` will be issued to disk, and doesn't handle the case where this is not true; i.e. it doesn't handle the case where `dmu_tx_assign` returns `EIO`. SOLUTION ======== This change modifies the `dmu_tx_assign` function such that `txg_how` is a bitmask, rather than of the `txg_how_t` enum type. Now, the previous `TXG_WAITED` semantics can be used via `TXG_NOTHROTTLE`, along with specifying either `TXG_NOWAIT` or `TXG_WAIT` semantics. Previously, when `TXG_WAITED` was specified, `TXG_NOWAIT` semantics was automatically invoked. This was not ideal when using `TXG_WAITED` within `zil_lwb_write_issued`, leading the problem described above. Rather, we want to achieve the semantics of `TXG_WAIT`, while also preventing the `tx` from being penalized via the dirty delay throttling. With this change, `zil_lwb_write_issued` can acheive the semtantics that it requires by passing in the value `TXG_WAIT | TXG_NOTHROTTLE` to `dmu_tx_assign`. Further, consumers of `dmu_tx_assign` wishing to achieve the old `TXG_WAITED` semantics can pass in the value `TXG_NOWAIT | TXG_NOTHROTTLE`. Authored by: Prakash Surya <prakash.surya@delphix.com> Approved by: Robert Mustacchi <rm@joyent.com> Reviewed by: Matt Ahrens <mahrens@delphix.com> Reviewed by: Andriy Gapon <avg@FreeBSD.org> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Porting Notes: - Additionally updated `zfs_tmpfile` to use `TXG_NOTHROTTLE` OpenZFS-issue: https://www.illumos.org/issues/8997 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/19ea6cb0f9 Closes #7084
2018-01-09 00:45:53 +03:00
!(txg_how & TXG_WAIT))
Illumos #3598 3598 want to dtrace when errors are generated in zfs Reviewed by: Dan Kimmel <dan.kimmel@delphix.com> Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: Christopher Siden <christopher.siden@delphix.com> Approved by: Garrett D'Amore <garrett@damore.org> References: https://www.illumos.org/issues/3598 illumos/illumos-gate@be6fd75a69ae679453d9cda5bff3326111e6d1ca Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775 Porting notes: 1. include/sys/zfs_context.h has been modified to render some new macros inert until dtrace is available on Linux. 2. Linux-specific changes have been adapted to use SET_ERROR(). 3. I'm NOT happy about this change. It does nothing but ugly up the code under Linux. Unfortunately we need to take it to avoid more merge conflicts in the future. -Brian
2013-03-08 22:41:28 +04:00
return (SET_ERROR(EIO));
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
Illumos #3598 3598 want to dtrace when errors are generated in zfs Reviewed by: Dan Kimmel <dan.kimmel@delphix.com> Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: Christopher Siden <christopher.siden@delphix.com> Approved by: Garrett D'Amore <garrett@damore.org> References: https://www.illumos.org/issues/3598 illumos/illumos-gate@be6fd75a69ae679453d9cda5bff3326111e6d1ca Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775 Porting notes: 1. include/sys/zfs_context.h has been modified to render some new macros inert until dtrace is available on Linux. 2. Linux-specific changes have been adapted to use SET_ERROR(). 3. I'm NOT happy about this change. It does nothing but ugly up the code under Linux. Unfortunately we need to take it to avoid more merge conflicts in the future. -Brian
2013-03-08 22:41:28 +04:00
return (SET_ERROR(ERESTART));
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
OpenZFS 8997 - ztest assertion failure in zil_lwb_write_issue PROBLEM ======= When `dmu_tx_assign` is called from `zil_lwb_write_issue`, it's possible for either `ERESTART` or `EIO` to be returned. If `ERESTART` is returned, this will cause an assertion to fail directly in `zil_lwb_write_issue`, where the code assumes the return value is `EIO` if `dmu_tx_assign` returns a non-zero value. This can occur if the SPA is suspended when `dmu_tx_assign` is called, and most often occurs when running `zloop`. If `EIO` is returned, this can cause assertions to fail elsewhere in the ZIL code. For example, `zil_commit_waiter_timeout` contains the following logic: lwb_t *nlwb = zil_lwb_write_issue(zilog, lwb); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_OPENED); In this case, if `dmu_tx_assign` returned `EIO` from within `zil_lwb_write_issue`, the `lwb` variable passed in will not be issued to disk. Thus, it's `lwb_state` field will remain `LWB_STATE_OPENED` and this assertion will fail. `zil_commit_waiter_timeout` assumes that after it calls `zil_lwb_write_issue`, the `lwb` will be issued to disk, and doesn't handle the case where this is not true; i.e. it doesn't handle the case where `dmu_tx_assign` returns `EIO`. SOLUTION ======== This change modifies the `dmu_tx_assign` function such that `txg_how` is a bitmask, rather than of the `txg_how_t` enum type. Now, the previous `TXG_WAITED` semantics can be used via `TXG_NOTHROTTLE`, along with specifying either `TXG_NOWAIT` or `TXG_WAIT` semantics. Previously, when `TXG_WAITED` was specified, `TXG_NOWAIT` semantics was automatically invoked. This was not ideal when using `TXG_WAITED` within `zil_lwb_write_issued`, leading the problem described above. Rather, we want to achieve the semantics of `TXG_WAIT`, while also preventing the `tx` from being penalized via the dirty delay throttling. With this change, `zil_lwb_write_issued` can acheive the semtantics that it requires by passing in the value `TXG_WAIT | TXG_NOTHROTTLE` to `dmu_tx_assign`. Further, consumers of `dmu_tx_assign` wishing to achieve the old `TXG_WAITED` semantics can pass in the value `TXG_NOWAIT | TXG_NOTHROTTLE`. Authored by: Prakash Surya <prakash.surya@delphix.com> Approved by: Robert Mustacchi <rm@joyent.com> Reviewed by: Matt Ahrens <mahrens@delphix.com> Reviewed by: Andriy Gapon <avg@FreeBSD.org> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Porting Notes: - Additionally updated `zfs_tmpfile` to use `TXG_NOTHROTTLE` OpenZFS-issue: https://www.illumos.org/issues/8997 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/19ea6cb0f9 Closes #7084
2018-01-09 00:45:53 +03:00
if (!tx->tx_dirty_delayed &&
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@69962b5647e4a8b9b14998733b765925381b727e Ported-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1913
2013-08-29 07:01:20 +04:00
dsl_pool_need_dirty_delay(tx->tx_pool)) {
tx->tx_wait_dirty = B_TRUE;
DMU_TX_STAT_BUMP(dmu_tx_dirty_delay);
Use SET_ERROR for constant non-zero return codes Update many return and assignment statements to follow the convention of using the SET_ERROR macro when returning a hard-coded non-zero value from a function. This aids debugging by recording the error codes in the debug log. Reviewed-by: George Melikov <mail@gmelikov.ru> Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed by: Matthew Ahrens <mahrens@delphix.com> Signed-off-by: Ned Bass <bass6@llnl.gov> Closes #6441
2017-08-03 07:16:12 +03:00
return (SET_ERROR(ERESTART));
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@69962b5647e4a8b9b14998733b765925381b727e Ported-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1913
2013-08-29 07:01:20 +04:00
}
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
tx->tx_txg = txg_hold_open(tx->tx_pool, &tx->tx_txgh);
tx->tx_needassign_txh = NULL;
/*
* NB: No error returns are allowed after txg_hold_open, but
* before processing the dnode holds, due to the
* dmu_tx_unassign() logic.
*/
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
uint64_t towrite = 0;
uint64_t tohold = 0;
for (dmu_tx_hold_t *txh = list_head(&tx->tx_holds); txh != NULL;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
txh = list_next(&tx->tx_holds, txh)) {
dnode_t *dn = txh->txh_dnode;
if (dn != NULL) {
dn_struct_rwlock can not be held in dmu_tx_try_assign() The thread calling dmu_tx_try_assign() can't hold the dn_struct_rwlock while assigning the tx, because this can lead to deadlock. Specifically, if this dnode is already assigned to an earlier txg, this thread may need to wait for that txg to sync (the ERESTART case below). The other thread that has assigned this dnode to an earlier txg prevents this txg from syncing until its tx can complete (calling dmu_tx_commit()), but it may need to acquire the dn_struct_rwlock to do so (e.g. via dmu_buf_hold*()). This commit adds an assertion to dmu_tx_try_assign() to ensure that this deadlock is not inadvertently introduced. Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Matthew Ahrens <mahrens@delphix.com> Closes #8929
2019-06-23 02:48:54 +03:00
/*
* This thread can't hold the dn_struct_rwlock
* while assigning the tx, because this can lead to
* deadlock. Specifically, if this dnode is already
* assigned to an earlier txg, this thread may need
* to wait for that txg to sync (the ERESTART case
* below). The other thread that has assigned this
* dnode to an earlier txg prevents this txg from
* syncing until its tx can complete (calling
* dmu_tx_commit()), but it may need to acquire the
* dn_struct_rwlock to do so (e.g. via
* dmu_buf_hold*()).
*
* Note that this thread can't hold the lock for
* read either, but the rwlock doesn't record
* enough information to make that assertion.
*/
ASSERT(!RW_WRITE_HELD(&dn->dn_struct_rwlock));
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
mutex_enter(&dn->dn_mtx);
if (dn->dn_assigned_txg == tx->tx_txg - 1) {
mutex_exit(&dn->dn_mtx);
tx->tx_needassign_txh = txh;
Add 'dmu_tx' kstats entry Keep counters for the various reasons that a thread may end up in txg_wait_open() waiting on a new txg. This can be useful when attempting to determine why a particular workload is under performing. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2012-01-20 22:58:57 +04:00
DMU_TX_STAT_BUMP(dmu_tx_group);
Illumos #3598 3598 want to dtrace when errors are generated in zfs Reviewed by: Dan Kimmel <dan.kimmel@delphix.com> Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: Christopher Siden <christopher.siden@delphix.com> Approved by: Garrett D'Amore <garrett@damore.org> References: https://www.illumos.org/issues/3598 illumos/illumos-gate@be6fd75a69ae679453d9cda5bff3326111e6d1ca Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775 Porting notes: 1. include/sys/zfs_context.h has been modified to render some new macros inert until dtrace is available on Linux. 2. Linux-specific changes have been adapted to use SET_ERROR(). 3. I'm NOT happy about this change. It does nothing but ugly up the code under Linux. Unfortunately we need to take it to avoid more merge conflicts in the future. -Brian
2013-03-08 22:41:28 +04:00
return (SET_ERROR(ERESTART));
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
if (dn->dn_assigned_txg == 0)
dn->dn_assigned_txg = tx->tx_txg;
ASSERT3U(dn->dn_assigned_txg, ==, tx->tx_txg);
Linux 4.19-rc3+ compat: Remove refcount_t compat torvalds/linux@59b57717f ("blkcg: delay blkg destruction until after writeback has finished") added a refcount_t to the blkcg structure. Due to the refcount_t compatibility code, zfs_refcount_t was used by mistake. Resolve this by removing the compatibility code and replacing the occurrences of refcount_t with zfs_refcount_t. Reviewed-by: Franz Pletz <fpletz@fnordicwalking.de> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Tim Schumacher <timschumi@gmx.de> Closes #7885 Closes #7932
2018-09-26 20:29:26 +03:00
(void) zfs_refcount_add(&dn->dn_tx_holds, tx);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
mutex_exit(&dn->dn_mtx);
}
Prefix all refcount functions with zfs_ Recent changes in the Linux kernel made it necessary to prefix the refcount_add() function with zfs_ due to a name collision. To bring the other functions in line with that and to avoid future collisions, prefix the other refcount functions as well. Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Tim Schumacher <timschumi@gmx.de> Closes #7963
2018-10-01 20:42:05 +03:00
towrite += zfs_refcount_count(&txh->txh_space_towrite);
tohold += zfs_refcount_count(&txh->txh_memory_tohold);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
Rebase to OpenSolaris b103, in the process we are removing any code which did not originate from the OpenSolaris source. These changes will be reintroduced in topic branches for easier tracking
2008-12-03 23:09:06 +03:00
/* needed allocation: worst-case estimate of write space */
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
uint64_t asize = spa_get_worst_case_asize(tx->tx_pool->dp_spa, towrite);
Rebase to OpenSolaris b103, in the process we are removing any code which did not originate from the OpenSolaris source. These changes will be reintroduced in topic branches for easier tracking
2008-12-03 23:09:06 +03:00
/* calculate memory footprint estimate */
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
uint64_t memory = towrite + tohold;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
if (tx->tx_dir != NULL && asize != 0) {
Rebase to OpenSolaris b103, in the process we are removing any code which did not originate from the OpenSolaris source. These changes will be reintroduced in topic branches for easier tracking
2008-12-03 23:09:06 +03:00
int err = dsl_dir_tempreserve_space(tx->tx_dir, memory,
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
asize, tx->tx_netfree, &tx->tx_tempreserve_cookie, tx);
if (err != 0)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
return (err);
}
Add 'dmu_tx' kstats entry Keep counters for the various reasons that a thread may end up in txg_wait_open() waiting on a new txg. This can be useful when attempting to determine why a particular workload is under performing. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2012-01-20 22:58:57 +04:00
DMU_TX_STAT_BUMP(dmu_tx_assigned);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
return (0);
}
static void
dmu_tx_unassign(dmu_tx_t *tx)
{
if (tx->tx_txg == 0)
return;
txg_rele_to_quiesce(&tx->tx_txgh);
Illumos #3741 3741 zfs needs better comments Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed by: Eric Schrock <eric.schrock@delphix.com> Approved by: Christopher Siden <christopher.siden@delphix.com> References: https://www.illumos.org/issues/3741 illumos/illumos-gate@3e30c24aeefdee1631958ecf17f18da671781956 Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775
2013-06-11 21:12:34 +04:00
/*
* Walk the transaction's hold list, removing the hold on the
* associated dnode, and notifying waiters if the refcount drops to 0.
*/
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
for (dmu_tx_hold_t *txh = list_head(&tx->tx_holds);
Fix coverity defects: CID 147553 CID 147553: Type:Dereference null return value Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: cao.xuewen <cao.xuewen@zte.com.cn> Closes #5305
2016-11-01 20:20:24 +03:00
txh && txh != tx->tx_needassign_txh;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
txh = list_next(&tx->tx_holds, txh)) {
dnode_t *dn = txh->txh_dnode;
if (dn == NULL)
continue;
mutex_enter(&dn->dn_mtx);
ASSERT3U(dn->dn_assigned_txg, ==, tx->tx_txg);
Prefix all refcount functions with zfs_ Recent changes in the Linux kernel made it necessary to prefix the refcount_add() function with zfs_ due to a name collision. To bring the other functions in line with that and to avoid future collisions, prefix the other refcount functions as well. Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Tim Schumacher <timschumi@gmx.de> Closes #7963
2018-10-01 20:42:05 +03:00
if (zfs_refcount_remove(&dn->dn_tx_holds, tx) == 0) {
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
dn->dn_assigned_txg = 0;
cv_broadcast(&dn->dn_notxholds);
}
mutex_exit(&dn->dn_mtx);
}
txg_rele_to_sync(&tx->tx_txgh);
tx->tx_lasttried_txg = tx->tx_txg;
tx->tx_txg = 0;
}
/*
OpenZFS 8997 - ztest assertion failure in zil_lwb_write_issue PROBLEM ======= When `dmu_tx_assign` is called from `zil_lwb_write_issue`, it's possible for either `ERESTART` or `EIO` to be returned. If `ERESTART` is returned, this will cause an assertion to fail directly in `zil_lwb_write_issue`, where the code assumes the return value is `EIO` if `dmu_tx_assign` returns a non-zero value. This can occur if the SPA is suspended when `dmu_tx_assign` is called, and most often occurs when running `zloop`. If `EIO` is returned, this can cause assertions to fail elsewhere in the ZIL code. For example, `zil_commit_waiter_timeout` contains the following logic: lwb_t *nlwb = zil_lwb_write_issue(zilog, lwb); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_OPENED); In this case, if `dmu_tx_assign` returned `EIO` from within `zil_lwb_write_issue`, the `lwb` variable passed in will not be issued to disk. Thus, it's `lwb_state` field will remain `LWB_STATE_OPENED` and this assertion will fail. `zil_commit_waiter_timeout` assumes that after it calls `zil_lwb_write_issue`, the `lwb` will be issued to disk, and doesn't handle the case where this is not true; i.e. it doesn't handle the case where `dmu_tx_assign` returns `EIO`. SOLUTION ======== This change modifies the `dmu_tx_assign` function such that `txg_how` is a bitmask, rather than of the `txg_how_t` enum type. Now, the previous `TXG_WAITED` semantics can be used via `TXG_NOTHROTTLE`, along with specifying either `TXG_NOWAIT` or `TXG_WAIT` semantics. Previously, when `TXG_WAITED` was specified, `TXG_NOWAIT` semantics was automatically invoked. This was not ideal when using `TXG_WAITED` within `zil_lwb_write_issued`, leading the problem described above. Rather, we want to achieve the semantics of `TXG_WAIT`, while also preventing the `tx` from being penalized via the dirty delay throttling. With this change, `zil_lwb_write_issued` can acheive the semtantics that it requires by passing in the value `TXG_WAIT | TXG_NOTHROTTLE` to `dmu_tx_assign`. Further, consumers of `dmu_tx_assign` wishing to achieve the old `TXG_WAITED` semantics can pass in the value `TXG_NOWAIT | TXG_NOTHROTTLE`. Authored by: Prakash Surya <prakash.surya@delphix.com> Approved by: Robert Mustacchi <rm@joyent.com> Reviewed by: Matt Ahrens <mahrens@delphix.com> Reviewed by: Andriy Gapon <avg@FreeBSD.org> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Porting Notes: - Additionally updated `zfs_tmpfile` to use `TXG_NOTHROTTLE` OpenZFS-issue: https://www.illumos.org/issues/8997 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/19ea6cb0f9 Closes #7084
2018-01-09 00:45:53 +03:00
* Assign tx to a transaction group; txg_how is a bitmask:
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
*
OpenZFS 8997 - ztest assertion failure in zil_lwb_write_issue PROBLEM ======= When `dmu_tx_assign` is called from `zil_lwb_write_issue`, it's possible for either `ERESTART` or `EIO` to be returned. If `ERESTART` is returned, this will cause an assertion to fail directly in `zil_lwb_write_issue`, where the code assumes the return value is `EIO` if `dmu_tx_assign` returns a non-zero value. This can occur if the SPA is suspended when `dmu_tx_assign` is called, and most often occurs when running `zloop`. If `EIO` is returned, this can cause assertions to fail elsewhere in the ZIL code. For example, `zil_commit_waiter_timeout` contains the following logic: lwb_t *nlwb = zil_lwb_write_issue(zilog, lwb); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_OPENED); In this case, if `dmu_tx_assign` returned `EIO` from within `zil_lwb_write_issue`, the `lwb` variable passed in will not be issued to disk. Thus, it's `lwb_state` field will remain `LWB_STATE_OPENED` and this assertion will fail. `zil_commit_waiter_timeout` assumes that after it calls `zil_lwb_write_issue`, the `lwb` will be issued to disk, and doesn't handle the case where this is not true; i.e. it doesn't handle the case where `dmu_tx_assign` returns `EIO`. SOLUTION ======== This change modifies the `dmu_tx_assign` function such that `txg_how` is a bitmask, rather than of the `txg_how_t` enum type. Now, the previous `TXG_WAITED` semantics can be used via `TXG_NOTHROTTLE`, along with specifying either `TXG_NOWAIT` or `TXG_WAIT` semantics. Previously, when `TXG_WAITED` was specified, `TXG_NOWAIT` semantics was automatically invoked. This was not ideal when using `TXG_WAITED` within `zil_lwb_write_issued`, leading the problem described above. Rather, we want to achieve the semantics of `TXG_WAIT`, while also preventing the `tx` from being penalized via the dirty delay throttling. With this change, `zil_lwb_write_issued` can acheive the semtantics that it requires by passing in the value `TXG_WAIT | TXG_NOTHROTTLE` to `dmu_tx_assign`. Further, consumers of `dmu_tx_assign` wishing to achieve the old `TXG_WAITED` semantics can pass in the value `TXG_NOWAIT | TXG_NOTHROTTLE`. Authored by: Prakash Surya <prakash.surya@delphix.com> Approved by: Robert Mustacchi <rm@joyent.com> Reviewed by: Matt Ahrens <mahrens@delphix.com> Reviewed by: Andriy Gapon <avg@FreeBSD.org> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Porting Notes: - Additionally updated `zfs_tmpfile` to use `TXG_NOTHROTTLE` OpenZFS-issue: https://www.illumos.org/issues/8997 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/19ea6cb0f9 Closes #7084
2018-01-09 00:45:53 +03:00
* If TXG_WAIT is set and the currently open txg is full, this function
* will wait until there's a new txg. This should be used when no locks
* are being held. With this bit set, this function will only fail if
* we're truly out of space (or over quota).
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
*
OpenZFS 8997 - ztest assertion failure in zil_lwb_write_issue PROBLEM ======= When `dmu_tx_assign` is called from `zil_lwb_write_issue`, it's possible for either `ERESTART` or `EIO` to be returned. If `ERESTART` is returned, this will cause an assertion to fail directly in `zil_lwb_write_issue`, where the code assumes the return value is `EIO` if `dmu_tx_assign` returns a non-zero value. This can occur if the SPA is suspended when `dmu_tx_assign` is called, and most often occurs when running `zloop`. If `EIO` is returned, this can cause assertions to fail elsewhere in the ZIL code. For example, `zil_commit_waiter_timeout` contains the following logic: lwb_t *nlwb = zil_lwb_write_issue(zilog, lwb); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_OPENED); In this case, if `dmu_tx_assign` returned `EIO` from within `zil_lwb_write_issue`, the `lwb` variable passed in will not be issued to disk. Thus, it's `lwb_state` field will remain `LWB_STATE_OPENED` and this assertion will fail. `zil_commit_waiter_timeout` assumes that after it calls `zil_lwb_write_issue`, the `lwb` will be issued to disk, and doesn't handle the case where this is not true; i.e. it doesn't handle the case where `dmu_tx_assign` returns `EIO`. SOLUTION ======== This change modifies the `dmu_tx_assign` function such that `txg_how` is a bitmask, rather than of the `txg_how_t` enum type. Now, the previous `TXG_WAITED` semantics can be used via `TXG_NOTHROTTLE`, along with specifying either `TXG_NOWAIT` or `TXG_WAIT` semantics. Previously, when `TXG_WAITED` was specified, `TXG_NOWAIT` semantics was automatically invoked. This was not ideal when using `TXG_WAITED` within `zil_lwb_write_issued`, leading the problem described above. Rather, we want to achieve the semantics of `TXG_WAIT`, while also preventing the `tx` from being penalized via the dirty delay throttling. With this change, `zil_lwb_write_issued` can acheive the semtantics that it requires by passing in the value `TXG_WAIT | TXG_NOTHROTTLE` to `dmu_tx_assign`. Further, consumers of `dmu_tx_assign` wishing to achieve the old `TXG_WAITED` semantics can pass in the value `TXG_NOWAIT | TXG_NOTHROTTLE`. Authored by: Prakash Surya <prakash.surya@delphix.com> Approved by: Robert Mustacchi <rm@joyent.com> Reviewed by: Matt Ahrens <mahrens@delphix.com> Reviewed by: Andriy Gapon <avg@FreeBSD.org> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Porting Notes: - Additionally updated `zfs_tmpfile` to use `TXG_NOTHROTTLE` OpenZFS-issue: https://www.illumos.org/issues/8997 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/19ea6cb0f9 Closes #7084
2018-01-09 00:45:53 +03:00
* If TXG_WAIT is *not* set and we can't assign into the currently open
* txg without blocking, this function will return immediately with
* ERESTART. This should be used whenever locks are being held. On an
* ERESTART error, the caller should drop all locks, call dmu_tx_wait(),
* and try again.
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@69962b5647e4a8b9b14998733b765925381b727e Ported-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1913
2013-08-29 07:01:20 +04:00
*
OpenZFS 8997 - ztest assertion failure in zil_lwb_write_issue PROBLEM ======= When `dmu_tx_assign` is called from `zil_lwb_write_issue`, it's possible for either `ERESTART` or `EIO` to be returned. If `ERESTART` is returned, this will cause an assertion to fail directly in `zil_lwb_write_issue`, where the code assumes the return value is `EIO` if `dmu_tx_assign` returns a non-zero value. This can occur if the SPA is suspended when `dmu_tx_assign` is called, and most often occurs when running `zloop`. If `EIO` is returned, this can cause assertions to fail elsewhere in the ZIL code. For example, `zil_commit_waiter_timeout` contains the following logic: lwb_t *nlwb = zil_lwb_write_issue(zilog, lwb); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_OPENED); In this case, if `dmu_tx_assign` returned `EIO` from within `zil_lwb_write_issue`, the `lwb` variable passed in will not be issued to disk. Thus, it's `lwb_state` field will remain `LWB_STATE_OPENED` and this assertion will fail. `zil_commit_waiter_timeout` assumes that after it calls `zil_lwb_write_issue`, the `lwb` will be issued to disk, and doesn't handle the case where this is not true; i.e. it doesn't handle the case where `dmu_tx_assign` returns `EIO`. SOLUTION ======== This change modifies the `dmu_tx_assign` function such that `txg_how` is a bitmask, rather than of the `txg_how_t` enum type. Now, the previous `TXG_WAITED` semantics can be used via `TXG_NOTHROTTLE`, along with specifying either `TXG_NOWAIT` or `TXG_WAIT` semantics. Previously, when `TXG_WAITED` was specified, `TXG_NOWAIT` semantics was automatically invoked. This was not ideal when using `TXG_WAITED` within `zil_lwb_write_issued`, leading the problem described above. Rather, we want to achieve the semantics of `TXG_WAIT`, while also preventing the `tx` from being penalized via the dirty delay throttling. With this change, `zil_lwb_write_issued` can acheive the semtantics that it requires by passing in the value `TXG_WAIT | TXG_NOTHROTTLE` to `dmu_tx_assign`. Further, consumers of `dmu_tx_assign` wishing to achieve the old `TXG_WAITED` semantics can pass in the value `TXG_NOWAIT | TXG_NOTHROTTLE`. Authored by: Prakash Surya <prakash.surya@delphix.com> Approved by: Robert Mustacchi <rm@joyent.com> Reviewed by: Matt Ahrens <mahrens@delphix.com> Reviewed by: Andriy Gapon <avg@FreeBSD.org> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Porting Notes: - Additionally updated `zfs_tmpfile` to use `TXG_NOTHROTTLE` OpenZFS-issue: https://www.illumos.org/issues/8997 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/19ea6cb0f9 Closes #7084
2018-01-09 00:45:53 +03:00
* If TXG_NOTHROTTLE is set, this indicates that this tx should not be
* delayed due on the ZFS Write Throttle (see comments in dsl_pool.c for
* details on the throttle). This is used by the VFS operations, after
* they have already called dmu_tx_wait() (though most likely on a
* different tx).
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
*/
int
OpenZFS 8997 - ztest assertion failure in zil_lwb_write_issue PROBLEM ======= When `dmu_tx_assign` is called from `zil_lwb_write_issue`, it's possible for either `ERESTART` or `EIO` to be returned. If `ERESTART` is returned, this will cause an assertion to fail directly in `zil_lwb_write_issue`, where the code assumes the return value is `EIO` if `dmu_tx_assign` returns a non-zero value. This can occur if the SPA is suspended when `dmu_tx_assign` is called, and most often occurs when running `zloop`. If `EIO` is returned, this can cause assertions to fail elsewhere in the ZIL code. For example, `zil_commit_waiter_timeout` contains the following logic: lwb_t *nlwb = zil_lwb_write_issue(zilog, lwb); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_OPENED); In this case, if `dmu_tx_assign` returned `EIO` from within `zil_lwb_write_issue`, the `lwb` variable passed in will not be issued to disk. Thus, it's `lwb_state` field will remain `LWB_STATE_OPENED` and this assertion will fail. `zil_commit_waiter_timeout` assumes that after it calls `zil_lwb_write_issue`, the `lwb` will be issued to disk, and doesn't handle the case where this is not true; i.e. it doesn't handle the case where `dmu_tx_assign` returns `EIO`. SOLUTION ======== This change modifies the `dmu_tx_assign` function such that `txg_how` is a bitmask, rather than of the `txg_how_t` enum type. Now, the previous `TXG_WAITED` semantics can be used via `TXG_NOTHROTTLE`, along with specifying either `TXG_NOWAIT` or `TXG_WAIT` semantics. Previously, when `TXG_WAITED` was specified, `TXG_NOWAIT` semantics was automatically invoked. This was not ideal when using `TXG_WAITED` within `zil_lwb_write_issued`, leading the problem described above. Rather, we want to achieve the semantics of `TXG_WAIT`, while also preventing the `tx` from being penalized via the dirty delay throttling. With this change, `zil_lwb_write_issued` can acheive the semtantics that it requires by passing in the value `TXG_WAIT | TXG_NOTHROTTLE` to `dmu_tx_assign`. Further, consumers of `dmu_tx_assign` wishing to achieve the old `TXG_WAITED` semantics can pass in the value `TXG_NOWAIT | TXG_NOTHROTTLE`. Authored by: Prakash Surya <prakash.surya@delphix.com> Approved by: Robert Mustacchi <rm@joyent.com> Reviewed by: Matt Ahrens <mahrens@delphix.com> Reviewed by: Andriy Gapon <avg@FreeBSD.org> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Porting Notes: - Additionally updated `zfs_tmpfile` to use `TXG_NOTHROTTLE` OpenZFS-issue: https://www.illumos.org/issues/8997 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/19ea6cb0f9 Closes #7084
2018-01-09 00:45:53 +03:00
dmu_tx_assign(dmu_tx_t *tx, uint64_t txg_how)
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
{
int err;
ASSERT(tx->tx_txg == 0);
OpenZFS 8997 - ztest assertion failure in zil_lwb_write_issue PROBLEM ======= When `dmu_tx_assign` is called from `zil_lwb_write_issue`, it's possible for either `ERESTART` or `EIO` to be returned. If `ERESTART` is returned, this will cause an assertion to fail directly in `zil_lwb_write_issue`, where the code assumes the return value is `EIO` if `dmu_tx_assign` returns a non-zero value. This can occur if the SPA is suspended when `dmu_tx_assign` is called, and most often occurs when running `zloop`. If `EIO` is returned, this can cause assertions to fail elsewhere in the ZIL code. For example, `zil_commit_waiter_timeout` contains the following logic: lwb_t *nlwb = zil_lwb_write_issue(zilog, lwb); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_OPENED); In this case, if `dmu_tx_assign` returned `EIO` from within `zil_lwb_write_issue`, the `lwb` variable passed in will not be issued to disk. Thus, it's `lwb_state` field will remain `LWB_STATE_OPENED` and this assertion will fail. `zil_commit_waiter_timeout` assumes that after it calls `zil_lwb_write_issue`, the `lwb` will be issued to disk, and doesn't handle the case where this is not true; i.e. it doesn't handle the case where `dmu_tx_assign` returns `EIO`. SOLUTION ======== This change modifies the `dmu_tx_assign` function such that `txg_how` is a bitmask, rather than of the `txg_how_t` enum type. Now, the previous `TXG_WAITED` semantics can be used via `TXG_NOTHROTTLE`, along with specifying either `TXG_NOWAIT` or `TXG_WAIT` semantics. Previously, when `TXG_WAITED` was specified, `TXG_NOWAIT` semantics was automatically invoked. This was not ideal when using `TXG_WAITED` within `zil_lwb_write_issued`, leading the problem described above. Rather, we want to achieve the semantics of `TXG_WAIT`, while also preventing the `tx` from being penalized via the dirty delay throttling. With this change, `zil_lwb_write_issued` can acheive the semtantics that it requires by passing in the value `TXG_WAIT | TXG_NOTHROTTLE` to `dmu_tx_assign`. Further, consumers of `dmu_tx_assign` wishing to achieve the old `TXG_WAITED` semantics can pass in the value `TXG_NOWAIT | TXG_NOTHROTTLE`. Authored by: Prakash Surya <prakash.surya@delphix.com> Approved by: Robert Mustacchi <rm@joyent.com> Reviewed by: Matt Ahrens <mahrens@delphix.com> Reviewed by: Andriy Gapon <avg@FreeBSD.org> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Porting Notes: - Additionally updated `zfs_tmpfile` to use `TXG_NOTHROTTLE` OpenZFS-issue: https://www.illumos.org/issues/8997 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/19ea6cb0f9 Closes #7084
2018-01-09 00:45:53 +03:00
ASSERT0(txg_how & ~(TXG_WAIT | TXG_NOTHROTTLE));
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
ASSERT(!dsl_pool_sync_context(tx->tx_pool));
Illumos #3464 3464 zfs synctask code needs restructuring Reviewed by: Dan Kimmel <dan.kimmel@delphix.com> Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Christopher Siden <christopher.siden@delphix.com> Approved by: Garrett D'Amore <garrett@damore.org> References: https://www.illumos.org/issues/3464 illumos/illumos-gate@3b2aab18808792cbd248a12f1edf139b89833c13 Ported-by: Tim Chase <tim@chase2k.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1495
2013-09-04 16:00:57 +04:00
/* If we might wait, we must not hold the config lock. */
OpenZFS 8997 - ztest assertion failure in zil_lwb_write_issue PROBLEM ======= When `dmu_tx_assign` is called from `zil_lwb_write_issue`, it's possible for either `ERESTART` or `EIO` to be returned. If `ERESTART` is returned, this will cause an assertion to fail directly in `zil_lwb_write_issue`, where the code assumes the return value is `EIO` if `dmu_tx_assign` returns a non-zero value. This can occur if the SPA is suspended when `dmu_tx_assign` is called, and most often occurs when running `zloop`. If `EIO` is returned, this can cause assertions to fail elsewhere in the ZIL code. For example, `zil_commit_waiter_timeout` contains the following logic: lwb_t *nlwb = zil_lwb_write_issue(zilog, lwb); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_OPENED); In this case, if `dmu_tx_assign` returned `EIO` from within `zil_lwb_write_issue`, the `lwb` variable passed in will not be issued to disk. Thus, it's `lwb_state` field will remain `LWB_STATE_OPENED` and this assertion will fail. `zil_commit_waiter_timeout` assumes that after it calls `zil_lwb_write_issue`, the `lwb` will be issued to disk, and doesn't handle the case where this is not true; i.e. it doesn't handle the case where `dmu_tx_assign` returns `EIO`. SOLUTION ======== This change modifies the `dmu_tx_assign` function such that `txg_how` is a bitmask, rather than of the `txg_how_t` enum type. Now, the previous `TXG_WAITED` semantics can be used via `TXG_NOTHROTTLE`, along with specifying either `TXG_NOWAIT` or `TXG_WAIT` semantics. Previously, when `TXG_WAITED` was specified, `TXG_NOWAIT` semantics was automatically invoked. This was not ideal when using `TXG_WAITED` within `zil_lwb_write_issued`, leading the problem described above. Rather, we want to achieve the semantics of `TXG_WAIT`, while also preventing the `tx` from being penalized via the dirty delay throttling. With this change, `zil_lwb_write_issued` can acheive the semtantics that it requires by passing in the value `TXG_WAIT | TXG_NOTHROTTLE` to `dmu_tx_assign`. Further, consumers of `dmu_tx_assign` wishing to achieve the old `TXG_WAITED` semantics can pass in the value `TXG_NOWAIT | TXG_NOTHROTTLE`. Authored by: Prakash Surya <prakash.surya@delphix.com> Approved by: Robert Mustacchi <rm@joyent.com> Reviewed by: Matt Ahrens <mahrens@delphix.com> Reviewed by: Andriy Gapon <avg@FreeBSD.org> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Porting Notes: - Additionally updated `zfs_tmpfile` to use `TXG_NOTHROTTLE` OpenZFS-issue: https://www.illumos.org/issues/8997 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/19ea6cb0f9 Closes #7084
2018-01-09 00:45:53 +03:00
IMPLY((txg_how & TXG_WAIT), !dsl_pool_config_held(tx->tx_pool));
if ((txg_how & TXG_NOTHROTTLE))
tx->tx_dirty_delayed = B_TRUE;
Illumos #3464 3464 zfs synctask code needs restructuring Reviewed by: Dan Kimmel <dan.kimmel@delphix.com> Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Christopher Siden <christopher.siden@delphix.com> Approved by: Garrett D'Amore <garrett@damore.org> References: https://www.illumos.org/issues/3464 illumos/illumos-gate@3b2aab18808792cbd248a12f1edf139b89833c13 Ported-by: Tim Chase <tim@chase2k.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1495
2013-09-04 16:00:57 +04:00
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
while ((err = dmu_tx_try_assign(tx, txg_how)) != 0) {
dmu_tx_unassign(tx);
OpenZFS 8997 - ztest assertion failure in zil_lwb_write_issue PROBLEM ======= When `dmu_tx_assign` is called from `zil_lwb_write_issue`, it's possible for either `ERESTART` or `EIO` to be returned. If `ERESTART` is returned, this will cause an assertion to fail directly in `zil_lwb_write_issue`, where the code assumes the return value is `EIO` if `dmu_tx_assign` returns a non-zero value. This can occur if the SPA is suspended when `dmu_tx_assign` is called, and most often occurs when running `zloop`. If `EIO` is returned, this can cause assertions to fail elsewhere in the ZIL code. For example, `zil_commit_waiter_timeout` contains the following logic: lwb_t *nlwb = zil_lwb_write_issue(zilog, lwb); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_OPENED); In this case, if `dmu_tx_assign` returned `EIO` from within `zil_lwb_write_issue`, the `lwb` variable passed in will not be issued to disk. Thus, it's `lwb_state` field will remain `LWB_STATE_OPENED` and this assertion will fail. `zil_commit_waiter_timeout` assumes that after it calls `zil_lwb_write_issue`, the `lwb` will be issued to disk, and doesn't handle the case where this is not true; i.e. it doesn't handle the case where `dmu_tx_assign` returns `EIO`. SOLUTION ======== This change modifies the `dmu_tx_assign` function such that `txg_how` is a bitmask, rather than of the `txg_how_t` enum type. Now, the previous `TXG_WAITED` semantics can be used via `TXG_NOTHROTTLE`, along with specifying either `TXG_NOWAIT` or `TXG_WAIT` semantics. Previously, when `TXG_WAITED` was specified, `TXG_NOWAIT` semantics was automatically invoked. This was not ideal when using `TXG_WAITED` within `zil_lwb_write_issued`, leading the problem described above. Rather, we want to achieve the semantics of `TXG_WAIT`, while also preventing the `tx` from being penalized via the dirty delay throttling. With this change, `zil_lwb_write_issued` can acheive the semtantics that it requires by passing in the value `TXG_WAIT | TXG_NOTHROTTLE` to `dmu_tx_assign`. Further, consumers of `dmu_tx_assign` wishing to achieve the old `TXG_WAITED` semantics can pass in the value `TXG_NOWAIT | TXG_NOTHROTTLE`. Authored by: Prakash Surya <prakash.surya@delphix.com> Approved by: Robert Mustacchi <rm@joyent.com> Reviewed by: Matt Ahrens <mahrens@delphix.com> Reviewed by: Andriy Gapon <avg@FreeBSD.org> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Porting Notes: - Additionally updated `zfs_tmpfile` to use `TXG_NOTHROTTLE` OpenZFS-issue: https://www.illumos.org/issues/8997 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/19ea6cb0f9 Closes #7084
2018-01-09 00:45:53 +03:00
if (err != ERESTART || !(txg_how & TXG_WAIT))
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
return (err);
dmu_tx_wait(tx);
}
txg_rele_to_quiesce(&tx->tx_txgh);
return (0);
}
void
dmu_tx_wait(dmu_tx_t *tx)
{
spa_t *spa = tx->tx_pool->dp_spa;
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@69962b5647e4a8b9b14998733b765925381b727e Ported-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1913
2013-08-29 07:01:20 +04:00
dsl_pool_t *dp = tx->tx_pool;
Improve reporting of tx assignment wait times Some callers of dmu_tx_assign() use the TXG_NOWAIT flag and call dmu_tx_wait() themselves before retrying if the assignment fails. The wait times for such callers are not accounted for in the dmu_tx_assign kstat histogram, because the histogram only records time spent in dmu_tx_assign(). This change moves the histogram update to dmu_tx_wait() to properly account for all time spent there. One downside of this approach is that it is possible to call dmu_tx_wait() multiple times before successfully assigning a transaction, in which case the cumulative wait time would not be recorded. However, this case should not often arise in practice, because most callers currently use one of these forms: dmu_tx_assign(tx, TXG_WAIT); dmu_tx_assign(tx, waited ? TXG_WAITED : TXG_NOWAIT); The first form should make just one call to dmu_tx_delay() inside of dmu_tx_assign(). The second form retries with TXG_WAITED if the first assignment fails and incurs a delay, in which case no further waiting is performed. Therefore transaction delays normally occur in one call to dmu_tx_wait() so the histogram should be fairly accurate. Another possible downside of this approach is that the histogram will no longer record overhead outside of dmu_tx_wait() such as in dmu_tx_try_assign(). While I'm not aware of any reason for concern on this point, it is conceivable that lock contention, long list traversal, etc. could cause assignment delays that would not be reflected in the histogram. Therefore the histogram should strictly be used for visibility in to the normal delay mechanisms and not as a profiling tool for code performance. Signed-off-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1915
2014-03-01 03:07:00 +04:00
hrtime_t before;
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
ASSERT(tx->tx_txg == 0);
Illumos #3464 3464 zfs synctask code needs restructuring Reviewed by: Dan Kimmel <dan.kimmel@delphix.com> Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Christopher Siden <christopher.siden@delphix.com> Approved by: Garrett D'Amore <garrett@damore.org> References: https://www.illumos.org/issues/3464 illumos/illumos-gate@3b2aab18808792cbd248a12f1edf139b89833c13 Ported-by: Tim Chase <tim@chase2k.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1495
2013-09-04 16:00:57 +04:00
ASSERT(!dsl_pool_config_held(tx->tx_pool));
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
Improve reporting of tx assignment wait times Some callers of dmu_tx_assign() use the TXG_NOWAIT flag and call dmu_tx_wait() themselves before retrying if the assignment fails. The wait times for such callers are not accounted for in the dmu_tx_assign kstat histogram, because the histogram only records time spent in dmu_tx_assign(). This change moves the histogram update to dmu_tx_wait() to properly account for all time spent there. One downside of this approach is that it is possible to call dmu_tx_wait() multiple times before successfully assigning a transaction, in which case the cumulative wait time would not be recorded. However, this case should not often arise in practice, because most callers currently use one of these forms: dmu_tx_assign(tx, TXG_WAIT); dmu_tx_assign(tx, waited ? TXG_WAITED : TXG_NOWAIT); The first form should make just one call to dmu_tx_delay() inside of dmu_tx_assign(). The second form retries with TXG_WAITED if the first assignment fails and incurs a delay, in which case no further waiting is performed. Therefore transaction delays normally occur in one call to dmu_tx_wait() so the histogram should be fairly accurate. Another possible downside of this approach is that the histogram will no longer record overhead outside of dmu_tx_wait() such as in dmu_tx_try_assign(). While I'm not aware of any reason for concern on this point, it is conceivable that lock contention, long list traversal, etc. could cause assignment delays that would not be reflected in the histogram. Therefore the histogram should strictly be used for visibility in to the normal delay mechanisms and not as a profiling tool for code performance. Signed-off-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1915
2014-03-01 03:07:00 +04:00
before = gethrtime();
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@69962b5647e4a8b9b14998733b765925381b727e Ported-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1913
2013-08-29 07:01:20 +04:00
if (tx->tx_wait_dirty) {
uint64_t dirty;
/*
* dmu_tx_try_assign() has determined that we need to wait
* because we've consumed much or all of the dirty buffer
* space.
*/
mutex_enter(&dp->dp_lock);
if (dp->dp_dirty_total >= zfs_dirty_data_max)
DMU_TX_STAT_BUMP(dmu_tx_dirty_over_max);
while (dp->dp_dirty_total >= zfs_dirty_data_max)
cv_wait(&dp->dp_spaceavail_cv, &dp->dp_lock);
dirty = dp->dp_dirty_total;
mutex_exit(&dp->dp_lock);
dmu_tx_delay(tx, dirty);
tx->tx_wait_dirty = B_FALSE;
/*
OpenZFS 8997 - ztest assertion failure in zil_lwb_write_issue PROBLEM ======= When `dmu_tx_assign` is called from `zil_lwb_write_issue`, it's possible for either `ERESTART` or `EIO` to be returned. If `ERESTART` is returned, this will cause an assertion to fail directly in `zil_lwb_write_issue`, where the code assumes the return value is `EIO` if `dmu_tx_assign` returns a non-zero value. This can occur if the SPA is suspended when `dmu_tx_assign` is called, and most often occurs when running `zloop`. If `EIO` is returned, this can cause assertions to fail elsewhere in the ZIL code. For example, `zil_commit_waiter_timeout` contains the following logic: lwb_t *nlwb = zil_lwb_write_issue(zilog, lwb); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_OPENED); In this case, if `dmu_tx_assign` returned `EIO` from within `zil_lwb_write_issue`, the `lwb` variable passed in will not be issued to disk. Thus, it's `lwb_state` field will remain `LWB_STATE_OPENED` and this assertion will fail. `zil_commit_waiter_timeout` assumes that after it calls `zil_lwb_write_issue`, the `lwb` will be issued to disk, and doesn't handle the case where this is not true; i.e. it doesn't handle the case where `dmu_tx_assign` returns `EIO`. SOLUTION ======== This change modifies the `dmu_tx_assign` function such that `txg_how` is a bitmask, rather than of the `txg_how_t` enum type. Now, the previous `TXG_WAITED` semantics can be used via `TXG_NOTHROTTLE`, along with specifying either `TXG_NOWAIT` or `TXG_WAIT` semantics. Previously, when `TXG_WAITED` was specified, `TXG_NOWAIT` semantics was automatically invoked. This was not ideal when using `TXG_WAITED` within `zil_lwb_write_issued`, leading the problem described above. Rather, we want to achieve the semantics of `TXG_WAIT`, while also preventing the `tx` from being penalized via the dirty delay throttling. With this change, `zil_lwb_write_issued` can acheive the semtantics that it requires by passing in the value `TXG_WAIT | TXG_NOTHROTTLE` to `dmu_tx_assign`. Further, consumers of `dmu_tx_assign` wishing to achieve the old `TXG_WAITED` semantics can pass in the value `TXG_NOWAIT | TXG_NOTHROTTLE`. Authored by: Prakash Surya <prakash.surya@delphix.com> Approved by: Robert Mustacchi <rm@joyent.com> Reviewed by: Matt Ahrens <mahrens@delphix.com> Reviewed by: Andriy Gapon <avg@FreeBSD.org> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Porting Notes: - Additionally updated `zfs_tmpfile` to use `TXG_NOTHROTTLE` OpenZFS-issue: https://www.illumos.org/issues/8997 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/19ea6cb0f9 Closes #7084
2018-01-09 00:45:53 +03:00
* Note: setting tx_dirty_delayed only has effect if the
* caller used TX_WAIT. Otherwise they are going to
* destroy this tx and try again. The common case,
* zfs_write(), uses TX_WAIT.
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@69962b5647e4a8b9b14998733b765925381b727e Ported-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1913
2013-08-29 07:01:20 +04:00
*/
OpenZFS 8997 - ztest assertion failure in zil_lwb_write_issue PROBLEM ======= When `dmu_tx_assign` is called from `zil_lwb_write_issue`, it's possible for either `ERESTART` or `EIO` to be returned. If `ERESTART` is returned, this will cause an assertion to fail directly in `zil_lwb_write_issue`, where the code assumes the return value is `EIO` if `dmu_tx_assign` returns a non-zero value. This can occur if the SPA is suspended when `dmu_tx_assign` is called, and most often occurs when running `zloop`. If `EIO` is returned, this can cause assertions to fail elsewhere in the ZIL code. For example, `zil_commit_waiter_timeout` contains the following logic: lwb_t *nlwb = zil_lwb_write_issue(zilog, lwb); ASSERT3S(lwb->lwb_state, !=, LWB_STATE_OPENED); In this case, if `dmu_tx_assign` returned `EIO` from within `zil_lwb_write_issue`, the `lwb` variable passed in will not be issued to disk. Thus, it's `lwb_state` field will remain `LWB_STATE_OPENED` and this assertion will fail. `zil_commit_waiter_timeout` assumes that after it calls `zil_lwb_write_issue`, the `lwb` will be issued to disk, and doesn't handle the case where this is not true; i.e. it doesn't handle the case where `dmu_tx_assign` returns `EIO`. SOLUTION ======== This change modifies the `dmu_tx_assign` function such that `txg_how` is a bitmask, rather than of the `txg_how_t` enum type. Now, the previous `TXG_WAITED` semantics can be used via `TXG_NOTHROTTLE`, along with specifying either `TXG_NOWAIT` or `TXG_WAIT` semantics. Previously, when `TXG_WAITED` was specified, `TXG_NOWAIT` semantics was automatically invoked. This was not ideal when using `TXG_WAITED` within `zil_lwb_write_issued`, leading the problem described above. Rather, we want to achieve the semantics of `TXG_WAIT`, while also preventing the `tx` from being penalized via the dirty delay throttling. With this change, `zil_lwb_write_issued` can acheive the semtantics that it requires by passing in the value `TXG_WAIT | TXG_NOTHROTTLE` to `dmu_tx_assign`. Further, consumers of `dmu_tx_assign` wishing to achieve the old `TXG_WAITED` semantics can pass in the value `TXG_NOWAIT | TXG_NOTHROTTLE`. Authored by: Prakash Surya <prakash.surya@delphix.com> Approved by: Robert Mustacchi <rm@joyent.com> Reviewed by: Matt Ahrens <mahrens@delphix.com> Reviewed by: Andriy Gapon <avg@FreeBSD.org> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Porting Notes: - Additionally updated `zfs_tmpfile` to use `TXG_NOTHROTTLE` OpenZFS-issue: https://www.illumos.org/issues/8997 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/19ea6cb0f9 Closes #7084
2018-01-09 00:45:53 +03:00
tx->tx_dirty_delayed = B_TRUE;
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@69962b5647e4a8b9b14998733b765925381b727e Ported-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1913
2013-08-29 07:01:20 +04:00
} else if (spa_suspended(spa) || tx->tx_lasttried_txg == 0) {
/*
* If the pool is suspended we need to wait until it
* is resumed. Note that it's possible that the pool
* has become active after this thread has tried to
* obtain a tx. If that's the case then tx_lasttried_txg
* would not have been set.
*/
txg_wait_synced(dp, spa_last_synced_txg(spa) + 1);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
} else if (tx->tx_needassign_txh) {
dnode_t *dn = tx->tx_needassign_txh->txh_dnode;
mutex_enter(&dn->dn_mtx);
while (dn->dn_assigned_txg == tx->tx_lasttried_txg - 1)
cv_wait(&dn->dn_notxholds, &dn->dn_mtx);
mutex_exit(&dn->dn_mtx);
tx->tx_needassign_txh = NULL;
} else {
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@69962b5647e4a8b9b14998733b765925381b727e Ported-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1913
2013-08-29 07:01:20 +04:00
/*
OpenZFS 9464 - txg_kick() fails to see that we are quiescing txg_kick() fails to see that we are quiescing, forcing transactions to their next stages without leaving them accumulate changes Creating a fragmented pool in a DCenter VM and continuously writing to it with multiple instances of randwritecomp, we get the following output from txg.d: 0ms 311MB in 4114ms (95% p1) 75MB/s 544MB (76%) 336us 153ms 0ms 0ms 8MB in 51ms ( 0% p1) 163MB/s 474MB (66%) 129us 34ms 0ms 0ms 366MB in 4454ms (93% p1) 82MB/s 572MB (79%) 498us 20ms 0ms 0ms 406MB in 5212ms (95% p1) 77MB/s 591MB (82%) 661us 37ms 0ms 0ms 340MB in 5110ms (94% p1) 66MB/s 622MB (86%) 1048us 41ms 1ms 0ms 3MB in 61ms ( 0% p1) 51MB/s 419MB (58%) 33us 0ms 0ms 0ms 361MB in 3555ms (88% p1) 101MB/s 542MB (75%) 335us 40ms 0ms 0ms 356MB in 4592ms (92% p1) 77MB/s 561MB (78%) 430us 89ms 1ms 0ms 11MB in 129ms (13% p1) 90MB/s 507MB (70%) 222us 15ms 0ms 0ms 281MB in 2520ms (89% p1) 111MB/s 542MB (75%) 334us 42ms 0ms 0ms 383MB in 3666ms (91% p1) 104MB/s 557MB (77%) 411us 133ms 0ms 0ms 404MB in 5757ms (94% p1) 70MB/s 635MB (88%) 1274us 123ms 2ms 4ms 367MB in 4172ms (89% p1) 88MB/s 556MB (77%) 401us 51ms 0ms 0ms 42MB in 470ms (44% p1) 90MB/s 557MB (77%) 412us 43ms 0ms 0ms 261MB in 2273ms (88% p1) 114MB/s 556MB (77%) 407us 27ms 0ms 0ms 394MB in 3646ms (85% p1) 108MB/s 552MB (77%) 393us 304ms 0ms 0ms 275MB in 2416ms (89% p1) 113MB/s 510MB (71%) 200us 53ms 0ms 0ms 9MB in 53ms ( 0% p1) 169MB/s 483MB (67%) 140us 100ms 1ms The TXGs that are getting synced and don't have lots of changes are pushed by txg_kick() which basically forces the current open txg to get to the quiesced state: if (tx->tx_syncing_txg == 0 && tx->tx_quiesce_txg_waiting <= tx->tx_open_txg && tx->tx_sync_txg_waiting <= tx->tx_synced_txg && tx->tx_quiesced_txg <= tx->tx_synced_txg) { tx->tx_quiesce_txg_waiting = tx->tx_open_txg + 1; cv_broadcast(&tx->tx_quiesce_more_cv); } The problem is that the above code doesn't check if we are currently quiescing anything (only if a quiesce or a sync has been requested, ..etc) so the following scenario can happen: 1] We have an open txg A that had enough dirty data (more than zfs_dirty_data_sync) and it was pushed to the quiesced state, and opened a new txg B. No txg is currently being synced. 2] Immediately after the opening of B, txg_kick() was run by some other write (and because of A's dirty data) and saw that we are not currently syncing any txg and no one has requested quiescing so it requests one by bumping tx_quiesce_txg_waiting and broadcasts the quiesce thread. 3] The quiesce thread just passed txg A to be synced and sees that a quiescing request has been sent to it so it immediately grabs B without letting it gather enough data, putting it in a quiesced state and opening a new txg C. In this scenario txg B, is an example of how the entries of interest show up in the txg.d output. Ideally we would like txg_kick() to get triggered only when we are sure that we are not syncing AND not quiescing any txg. This way we can kick an open TXG to the quiescing state when we are sure that there is nothing going on and we would benefit from the different states running concurrently. Authored by: Serapheim Dimitropoulos <serapheim@delphix.com> Reviewed by: Matt Ahrens <matt@delphix.com> Reviewed by: Brad Lewis <brad.lewis@delphix.com> Reviewed by: Andriy Gapon <avg@FreeBSD.org> Approved by: Dan McDonald <danmcd@joyent.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> OpenZFS-issue: https://illumos.org/issues/9464 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/1cd7635b Closes #7587
2017-12-05 20:45:46 +03:00
* If we have a lot of dirty data just wait until we sync
* out a TXG at which point we'll hopefully have synced
* a portion of the changes.
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@69962b5647e4a8b9b14998733b765925381b727e Ported-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1913
2013-08-29 07:01:20 +04:00
*/
OpenZFS 9464 - txg_kick() fails to see that we are quiescing txg_kick() fails to see that we are quiescing, forcing transactions to their next stages without leaving them accumulate changes Creating a fragmented pool in a DCenter VM and continuously writing to it with multiple instances of randwritecomp, we get the following output from txg.d: 0ms 311MB in 4114ms (95% p1) 75MB/s 544MB (76%) 336us 153ms 0ms 0ms 8MB in 51ms ( 0% p1) 163MB/s 474MB (66%) 129us 34ms 0ms 0ms 366MB in 4454ms (93% p1) 82MB/s 572MB (79%) 498us 20ms 0ms 0ms 406MB in 5212ms (95% p1) 77MB/s 591MB (82%) 661us 37ms 0ms 0ms 340MB in 5110ms (94% p1) 66MB/s 622MB (86%) 1048us 41ms 1ms 0ms 3MB in 61ms ( 0% p1) 51MB/s 419MB (58%) 33us 0ms 0ms 0ms 361MB in 3555ms (88% p1) 101MB/s 542MB (75%) 335us 40ms 0ms 0ms 356MB in 4592ms (92% p1) 77MB/s 561MB (78%) 430us 89ms 1ms 0ms 11MB in 129ms (13% p1) 90MB/s 507MB (70%) 222us 15ms 0ms 0ms 281MB in 2520ms (89% p1) 111MB/s 542MB (75%) 334us 42ms 0ms 0ms 383MB in 3666ms (91% p1) 104MB/s 557MB (77%) 411us 133ms 0ms 0ms 404MB in 5757ms (94% p1) 70MB/s 635MB (88%) 1274us 123ms 2ms 4ms 367MB in 4172ms (89% p1) 88MB/s 556MB (77%) 401us 51ms 0ms 0ms 42MB in 470ms (44% p1) 90MB/s 557MB (77%) 412us 43ms 0ms 0ms 261MB in 2273ms (88% p1) 114MB/s 556MB (77%) 407us 27ms 0ms 0ms 394MB in 3646ms (85% p1) 108MB/s 552MB (77%) 393us 304ms 0ms 0ms 275MB in 2416ms (89% p1) 113MB/s 510MB (71%) 200us 53ms 0ms 0ms 9MB in 53ms ( 0% p1) 169MB/s 483MB (67%) 140us 100ms 1ms The TXGs that are getting synced and don't have lots of changes are pushed by txg_kick() which basically forces the current open txg to get to the quiesced state: if (tx->tx_syncing_txg == 0 && tx->tx_quiesce_txg_waiting <= tx->tx_open_txg && tx->tx_sync_txg_waiting <= tx->tx_synced_txg && tx->tx_quiesced_txg <= tx->tx_synced_txg) { tx->tx_quiesce_txg_waiting = tx->tx_open_txg + 1; cv_broadcast(&tx->tx_quiesce_more_cv); } The problem is that the above code doesn't check if we are currently quiescing anything (only if a quiesce or a sync has been requested, ..etc) so the following scenario can happen: 1] We have an open txg A that had enough dirty data (more than zfs_dirty_data_sync) and it was pushed to the quiesced state, and opened a new txg B. No txg is currently being synced. 2] Immediately after the opening of B, txg_kick() was run by some other write (and because of A's dirty data) and saw that we are not currently syncing any txg and no one has requested quiescing so it requests one by bumping tx_quiesce_txg_waiting and broadcasts the quiesce thread. 3] The quiesce thread just passed txg A to be synced and sees that a quiescing request has been sent to it so it immediately grabs B without letting it gather enough data, putting it in a quiesced state and opening a new txg C. In this scenario txg B, is an example of how the entries of interest show up in the txg.d output. Ideally we would like txg_kick() to get triggered only when we are sure that we are not syncing AND not quiescing any txg. This way we can kick an open TXG to the quiescing state when we are sure that there is nothing going on and we would benefit from the different states running concurrently. Authored by: Serapheim Dimitropoulos <serapheim@delphix.com> Reviewed by: Matt Ahrens <matt@delphix.com> Reviewed by: Brad Lewis <brad.lewis@delphix.com> Reviewed by: Andriy Gapon <avg@FreeBSD.org> Approved by: Dan McDonald <danmcd@joyent.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> OpenZFS-issue: https://illumos.org/issues/9464 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/1cd7635b Closes #7587
2017-12-05 20:45:46 +03:00
txg_wait_synced(dp, spa_last_synced_txg(spa) + 1);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
Improve reporting of tx assignment wait times Some callers of dmu_tx_assign() use the TXG_NOWAIT flag and call dmu_tx_wait() themselves before retrying if the assignment fails. The wait times for such callers are not accounted for in the dmu_tx_assign kstat histogram, because the histogram only records time spent in dmu_tx_assign(). This change moves the histogram update to dmu_tx_wait() to properly account for all time spent there. One downside of this approach is that it is possible to call dmu_tx_wait() multiple times before successfully assigning a transaction, in which case the cumulative wait time would not be recorded. However, this case should not often arise in practice, because most callers currently use one of these forms: dmu_tx_assign(tx, TXG_WAIT); dmu_tx_assign(tx, waited ? TXG_WAITED : TXG_NOWAIT); The first form should make just one call to dmu_tx_delay() inside of dmu_tx_assign(). The second form retries with TXG_WAITED if the first assignment fails and incurs a delay, in which case no further waiting is performed. Therefore transaction delays normally occur in one call to dmu_tx_wait() so the histogram should be fairly accurate. Another possible downside of this approach is that the histogram will no longer record overhead outside of dmu_tx_wait() such as in dmu_tx_try_assign(). While I'm not aware of any reason for concern on this point, it is conceivable that lock contention, long list traversal, etc. could cause assignment delays that would not be reflected in the histogram. Therefore the histogram should strictly be used for visibility in to the normal delay mechanisms and not as a profiling tool for code performance. Signed-off-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1915
2014-03-01 03:07:00 +04:00
spa_tx_assign_add_nsecs(spa, gethrtime() - before);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
OpenZFS 7054 - dmu_tx_hold_t should use refcount_t to track space Authored by: Igor Kozhukhov ikozhukhov@gmail.com Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Paul Dagnelie <pcd@delphix.com> Reviewed by: Igor Kozhukhov <ikozhukhov@gmail.com> Approved by: Dan McDonald <danmcd@omniti.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Ported-by: George Melikov mail@gmelikov.ru OpenZFS-issue: https://www.illumos.org/issues/7054 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/0c779ad Closes #5600
2017-01-23 20:36:24 +03:00
static void
dmu_tx_destroy(dmu_tx_t *tx)
{
dmu_tx_hold_t *txh;
while ((txh = list_head(&tx->tx_holds)) != NULL) {
dnode_t *dn = txh->txh_dnode;
list_remove(&tx->tx_holds, txh);
Prefix all refcount functions with zfs_ Recent changes in the Linux kernel made it necessary to prefix the refcount_add() function with zfs_ due to a name collision. To bring the other functions in line with that and to avoid future collisions, prefix the other refcount functions as well. Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Tim Schumacher <timschumi@gmx.de> Closes #7963
2018-10-01 20:42:05 +03:00
zfs_refcount_destroy_many(&txh->txh_space_towrite,
zfs_refcount_count(&txh->txh_space_towrite));
zfs_refcount_destroy_many(&txh->txh_memory_tohold,
zfs_refcount_count(&txh->txh_memory_tohold));
OpenZFS 7054 - dmu_tx_hold_t should use refcount_t to track space Authored by: Igor Kozhukhov ikozhukhov@gmail.com Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Paul Dagnelie <pcd@delphix.com> Reviewed by: Igor Kozhukhov <ikozhukhov@gmail.com> Approved by: Dan McDonald <danmcd@omniti.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Ported-by: George Melikov mail@gmelikov.ru OpenZFS-issue: https://www.illumos.org/issues/7054 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/0c779ad Closes #5600
2017-01-23 20:36:24 +03:00
kmem_free(txh, sizeof (dmu_tx_hold_t));
if (dn != NULL)
dnode_rele(dn, tx);
}
list_destroy(&tx->tx_callbacks);
list_destroy(&tx->tx_holds);
kmem_free(tx, sizeof (dmu_tx_t));
}
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
void
dmu_tx_commit(dmu_tx_t *tx)
{
ASSERT(tx->tx_txg != 0);
Illumos #3741 3741 zfs needs better comments Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed by: Eric Schrock <eric.schrock@delphix.com> Approved by: Christopher Siden <christopher.siden@delphix.com> References: https://www.illumos.org/issues/3741 illumos/illumos-gate@3e30c24aeefdee1631958ecf17f18da671781956 Ported-by: Richard Yao <ryao@gentoo.org> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Issue #1775
2013-06-11 21:12:34 +04:00
/*
* Go through the transaction's hold list and remove holds on
* associated dnodes, notifying waiters if no holds remain.
*/
Undo c89 workarounds to match with upstream With PR 5756 the zfs module now supports c99 and the remaining past c89 workarounds can be undone. Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed-by: George Melikov <mail@gmelikov.ru> Signed-off-by: Don Brady <don.brady@delphix.com> Closes #6816
2017-11-04 23:25:13 +03:00
for (dmu_tx_hold_t *txh = list_head(&tx->tx_holds); txh != NULL;
OpenZFS 7054 - dmu_tx_hold_t should use refcount_t to track space Authored by: Igor Kozhukhov ikozhukhov@gmail.com Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Paul Dagnelie <pcd@delphix.com> Reviewed by: Igor Kozhukhov <ikozhukhov@gmail.com> Approved by: Dan McDonald <danmcd@omniti.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Ported-by: George Melikov mail@gmelikov.ru OpenZFS-issue: https://www.illumos.org/issues/7054 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/0c779ad Closes #5600
2017-01-23 20:36:24 +03:00
txh = list_next(&tx->tx_holds, txh)) {
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
dnode_t *dn = txh->txh_dnode;
if (dn == NULL)
continue;
OpenZFS 7054 - dmu_tx_hold_t should use refcount_t to track space Authored by: Igor Kozhukhov ikozhukhov@gmail.com Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Paul Dagnelie <pcd@delphix.com> Reviewed by: Igor Kozhukhov <ikozhukhov@gmail.com> Approved by: Dan McDonald <danmcd@omniti.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Ported-by: George Melikov mail@gmelikov.ru OpenZFS-issue: https://www.illumos.org/issues/7054 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/0c779ad Closes #5600
2017-01-23 20:36:24 +03:00
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
mutex_enter(&dn->dn_mtx);
ASSERT3U(dn->dn_assigned_txg, ==, tx->tx_txg);
Prefix all refcount functions with zfs_ Recent changes in the Linux kernel made it necessary to prefix the refcount_add() function with zfs_ due to a name collision. To bring the other functions in line with that and to avoid future collisions, prefix the other refcount functions as well. Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Tim Schumacher <timschumi@gmx.de> Closes #7963
2018-10-01 20:42:05 +03:00
if (zfs_refcount_remove(&dn->dn_tx_holds, tx) == 0) {
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
dn->dn_assigned_txg = 0;
cv_broadcast(&dn->dn_notxholds);
}
mutex_exit(&dn->dn_mtx);
}
if (tx->tx_tempreserve_cookie)
dsl_dir_tempreserve_clear(tx->tx_tempreserve_cookie, tx);
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
if (!list_is_empty(&tx->tx_callbacks))
txg_register_callbacks(&tx->tx_txgh, &tx->tx_callbacks);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
if (tx->tx_anyobj == FALSE)
txg_rele_to_sync(&tx->tx_txgh);
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
OpenZFS 7054 - dmu_tx_hold_t should use refcount_t to track space Authored by: Igor Kozhukhov ikozhukhov@gmail.com Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Paul Dagnelie <pcd@delphix.com> Reviewed by: Igor Kozhukhov <ikozhukhov@gmail.com> Approved by: Dan McDonald <danmcd@omniti.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Ported-by: George Melikov mail@gmelikov.ru OpenZFS-issue: https://www.illumos.org/issues/7054 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/0c779ad Closes #5600
2017-01-23 20:36:24 +03:00
dmu_tx_destroy(tx);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
void
dmu_tx_abort(dmu_tx_t *tx)
{
ASSERT(tx->tx_txg == 0);
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
/*
* Call any registered callbacks with an error code.
*/
if (!list_is_empty(&tx->tx_callbacks))
dmu_tx_do_callbacks(&tx->tx_callbacks, ECANCELED);
OpenZFS 7054 - dmu_tx_hold_t should use refcount_t to track space Authored by: Igor Kozhukhov ikozhukhov@gmail.com Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Paul Dagnelie <pcd@delphix.com> Reviewed by: Igor Kozhukhov <ikozhukhov@gmail.com> Approved by: Dan McDonald <danmcd@omniti.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Ported-by: George Melikov mail@gmelikov.ru OpenZFS-issue: https://www.illumos.org/issues/7054 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/0c779ad Closes #5600
2017-01-23 20:36:24 +03:00
dmu_tx_destroy(tx);
Initial Linux ZFS GIT Repo
2008-11-20 23:01:55 +03:00
}
uint64_t
dmu_tx_get_txg(dmu_tx_t *tx)
{
ASSERT(tx->tx_txg != 0);
return (tx->tx_txg);
}
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
Illumos #3464 3464 zfs synctask code needs restructuring Reviewed by: Dan Kimmel <dan.kimmel@delphix.com> Reviewed by: Adam Leventhal <ahl@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Christopher Siden <christopher.siden@delphix.com> Approved by: Garrett D'Amore <garrett@damore.org> References: https://www.illumos.org/issues/3464 illumos/illumos-gate@3b2aab18808792cbd248a12f1edf139b89833c13 Ported-by: Tim Chase <tim@chase2k.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #1495
2013-09-04 16:00:57 +04:00
dsl_pool_t *
dmu_tx_pool(dmu_tx_t *tx)
{
ASSERT(tx->tx_pool != NULL);
return (tx->tx_pool);
}
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
void
dmu_tx_callback_register(dmu_tx_t *tx, dmu_tx_callback_func_t *func, void *data)
{
dmu_tx_callback_t *dcb;
Change KM_PUSHPAGE -> KM_SLEEP By marking DMU transaction processing contexts with PF_FSTRANS we can revert the KM_PUSHPAGE -> KM_SLEEP changes. This brings us back in line with upstream. In some cases this means simply swapping the flags back. For others fnvlist_alloc() was replaced by nvlist_alloc(..., KM_PUSHPAGE) and must be reverted back to fnvlist_alloc() which assumes KM_SLEEP. The one place KM_PUSHPAGE is kept is when allocating ARC buffers which allows us to dip in to reserved memory. This is again the same as upstream. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2014-11-21 03:09:39 +03:00
dcb = kmem_alloc(sizeof (dmu_tx_callback_t), KM_SLEEP);
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
dcb->dcb_func = func;
dcb->dcb_data = data;
list_insert_tail(&tx->tx_callbacks, dcb);
}
/*
* Call all the commit callbacks on a list, with a given error code.
*/
void
dmu_tx_do_callbacks(list_t *cb_list, int error)
{
dmu_tx_callback_t *dcb;
Call commit callbacks from the tail of the list Our zfs backed Lustre MDT had soft lockups while under heavy metadata workloads while handling transaction callbacks from osd_zfs. The problem is zfs is not taking advantage of the fast path in Lustre's trans callback handling, where Lustre will skip the calls to ptlrpc_commit_replies() when it already saw a higher transaction number. This patch corrects this, it also has a positive impact on metadata performance on Lustre with osd_zfs, plus some cleanup in the headers. A similar issue for ext4/ldiskfs is described on: https://jira.hpdd.intel.com/browse/LU-6527 Reviewed-by: Olaf Faaland <faaland1@llnl.gov> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Li Dongyang <dongyang.li@anu.edu.au> Closes #6986
2017-12-22 21:19:51 +03:00
while ((dcb = list_tail(cb_list)) != NULL) {
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
list_remove(cb_list, dcb);
dcb->dcb_func(dcb->dcb_data, error);
kmem_free(dcb, sizeof (dmu_tx_callback_t));
}
}
/*
* Interface to hold a bunch of attributes.
* used for creating new files.
* attrsize is the total size of all attributes
* to be added during object creation
*
* For updating/adding a single attribute dmu_tx_hold_sa() should be used.
*/
/*
* hold necessary attribute name for attribute registration.
* should be a very rare case where this is needed. If it does
* happen it would only happen on the first write to the file system.
*/
static void
dmu_tx_sa_registration_hold(sa_os_t *sa, dmu_tx_t *tx)
{
if (!sa->sa_need_attr_registration)
return;
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
for (int i = 0; i != sa->sa_num_attrs; i++) {
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
if (!sa->sa_attr_table[i].sa_registered) {
if (sa->sa_reg_attr_obj)
dmu_tx_hold_zap(tx, sa->sa_reg_attr_obj,
B_TRUE, sa->sa_attr_table[i].sa_name);
else
dmu_tx_hold_zap(tx, DMU_NEW_OBJECT,
B_TRUE, sa->sa_attr_table[i].sa_name);
}
}
}
void
dmu_tx_hold_spill(dmu_tx_t *tx, uint64_t object)
{
Fix dnode allocation race When performing concurrent object allocations using the new multi-threaded allocator and large dnodes it's possible to allocate overlapping large dnodes. This case should have been handled by detecting an error returned by dnode_hold_impl(). But that logic only checked the returned dnp was not-NULL, and the dnp variable was not reset to NULL when retrying. Resolve this issue by properly checking the return value of dnode_hold_impl(). Additionally, it was possible that dnode_hold_impl() would misreport a dnode as free when it was in fact in use. This could occurs for two reasons: * The per-slot zrl_lock must be held over the entire critical section which includes the alloc/free until the new dnode is assigned to children_dnodes. Additionally, all of the zrl_lock's in the range must be held to protect moving dnodes. * The dn->dn_ot_type cannot be solely relied upon to check the type. When allocating a new dnode its type will be DMU_OT_NONE after dnode_create(). Only latter when dnode_allocate() is called will it transition to the new type. This means there's a window when allocating where it can mistaken for a free dnode. Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov> Reviewed-by: Ned Bass <bass6@llnl.gov> Reviewed-by: Tony Hutter <hutter2@llnl.gov> Reviewed-by: Olaf Faaland <faaland1@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #6414 Closes #6439
2017-08-08 18:38:53 +03:00
dmu_tx_hold_t *txh;
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
Fix dnode allocation race When performing concurrent object allocations using the new multi-threaded allocator and large dnodes it's possible to allocate overlapping large dnodes. This case should have been handled by detecting an error returned by dnode_hold_impl(). But that logic only checked the returned dnp was not-NULL, and the dnp variable was not reset to NULL when retrying. Resolve this issue by properly checking the return value of dnode_hold_impl(). Additionally, it was possible that dnode_hold_impl() would misreport a dnode as free when it was in fact in use. This could occurs for two reasons: * The per-slot zrl_lock must be held over the entire critical section which includes the alloc/free until the new dnode is assigned to children_dnodes. Additionally, all of the zrl_lock's in the range must be held to protect moving dnodes. * The dn->dn_ot_type cannot be solely relied upon to check the type. When allocating a new dnode its type will be DMU_OT_NONE after dnode_create(). Only latter when dnode_allocate() is called will it transition to the new type. This means there's a window when allocating where it can mistaken for a free dnode. Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov> Reviewed-by: Ned Bass <bass6@llnl.gov> Reviewed-by: Tony Hutter <hutter2@llnl.gov> Reviewed-by: Olaf Faaland <faaland1@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #6414 Closes #6439
2017-08-08 18:38:53 +03:00
txh = dmu_tx_hold_object_impl(tx, tx->tx_objset, object,
THT_SPILL, 0, 0);
if (txh != NULL)
Prefix all refcount functions with zfs_ Recent changes in the Linux kernel made it necessary to prefix the refcount_add() function with zfs_ due to a name collision. To bring the other functions in line with that and to avoid future collisions, prefix the other refcount functions as well. Reviewed by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Tim Schumacher <timschumi@gmx.de> Closes #7963
2018-10-01 20:42:05 +03:00
(void) zfs_refcount_add_many(&txh->txh_space_towrite,
Fix dnode allocation race When performing concurrent object allocations using the new multi-threaded allocator and large dnodes it's possible to allocate overlapping large dnodes. This case should have been handled by detecting an error returned by dnode_hold_impl(). But that logic only checked the returned dnp was not-NULL, and the dnp variable was not reset to NULL when retrying. Resolve this issue by properly checking the return value of dnode_hold_impl(). Additionally, it was possible that dnode_hold_impl() would misreport a dnode as free when it was in fact in use. This could occurs for two reasons: * The per-slot zrl_lock must be held over the entire critical section which includes the alloc/free until the new dnode is assigned to children_dnodes. Additionally, all of the zrl_lock's in the range must be held to protect moving dnodes. * The dn->dn_ot_type cannot be solely relied upon to check the type. When allocating a new dnode its type will be DMU_OT_NONE after dnode_create(). Only latter when dnode_allocate() is called will it transition to the new type. This means there's a window when allocating where it can mistaken for a free dnode. Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov> Reviewed-by: Ned Bass <bass6@llnl.gov> Reviewed-by: Tony Hutter <hutter2@llnl.gov> Reviewed-by: Olaf Faaland <faaland1@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #6414 Closes #6439
2017-08-08 18:38:53 +03:00
SPA_OLD_MAXBLOCKSIZE, FTAG);
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
}
void
dmu_tx_hold_sa_create(dmu_tx_t *tx, int attrsize)
{
sa_os_t *sa = tx->tx_objset->os_sa;
dmu_tx_hold_bonus(tx, DMU_NEW_OBJECT);
if (tx->tx_objset->os_sa->sa_master_obj == 0)
return;
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
if (tx->tx_objset->os_sa->sa_layout_attr_obj) {
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
dmu_tx_hold_zap(tx, sa->sa_layout_attr_obj, B_TRUE, NULL);
OpenZFS 7793 - ztest fails assertion in dmu_tx_willuse_space Reviewed by: Steve Gonczi <steve.gonczi@delphix.com> Reviewed by: George Wilson <george.wilson@delphix.com> Reviewed by: Pavel Zakharov <pavel.zakharov@delphix.com> Ported-by: Brian Behlendorf <behlendorf1@llnl.gov> Background information: This assertion about tx_space_* verifies that we are not dirtying more stuff than we thought we would. We “need” to know how much we will dirty so that we can check if we should fail this transaction with ENOSPC/EDQUOT, in dmu_tx_assign(). While the transaction is open (i.e. between dmu_tx_assign() and dmu_tx_commit() — typically less than a millisecond), we call dbuf_dirty() on the exact blocks that will be modified. Once this happens, the temporary accounting in tx_space_* is unnecessary, because we know exactly what blocks are newly dirtied; we call dnode_willuse_space() to track this more exact accounting. The fundamental problem causing this bug is that dmu_tx_hold_*() relies on the current state in the DMU (e.g. dn_nlevels) to predict how much will be dirtied by this transaction, but this state can change before we actually perform the transaction (i.e. call dbuf_dirty()). This bug will be fixed by removing the assertion that the tx_space_* accounting is perfectly accurate (i.e. we never dirty more than was predicted by dmu_tx_hold_*()). By removing the requirement that this accounting be perfectly accurate, we can also vastly simplify it, e.g. removing most of the logic in dmu_tx_count_*(). The new tx space accounting will be very approximate, and may be more or less than what is actually dirtied. It will still be used to determine if this transaction will put us over quota. Transactions that are marked by dmu_tx_mark_netfree() will be excepted from this check. We won’t make an attempt to determine how much space will be freed by the transaction — this was rarely accurate enough to determine if a transaction should be permitted when we are over quota, which is why dmu_tx_mark_netfree() was introduced in 2014. We also won’t attempt to give “credit” when overwriting existing blocks, if those blocks may be freed. This allows us to remove the do_free_accounting logic in dbuf_dirty(), and associated routines. This logic attempted to predict what will be on disk when this txg syncs, to know if the overwritten block will be freed (i.e. exists, and has no snapshots). OpenZFS-issue: https://www.illumos.org/issues/7793 OpenZFS-commit: https://github.com/openzfs/openzfs/commit/3704e0a Upstream bugs: DLPX-32883a Closes #5804 Porting notes: - DNODE_SIZE replaced with DNODE_MIN_SIZE in dmu_tx_count_dnode(), Using the default dnode size would be slightly better. - DEBUG_DMU_TX wrappers and configure option removed. - Resolved _by_dnode() conflicts these changes have not yet been applied to OpenZFS.
2017-03-07 20:51:59 +03:00
} else {
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
dmu_tx_hold_zap(tx, sa->sa_master_obj, B_TRUE, SA_LAYOUTS);
dmu_tx_hold_zap(tx, sa->sa_master_obj, B_TRUE, SA_REGISTRY);
dmu_tx_hold_zap(tx, DMU_NEW_OBJECT, B_TRUE, NULL);
dmu_tx_hold_zap(tx, DMU_NEW_OBJECT, B_TRUE, NULL);
}
dmu_tx_sa_registration_hold(sa, tx);
Implement large_dnode pool feature Justification ------------- This feature adds support for variable length dnodes. Our motivation is to eliminate the overhead associated with using spill blocks. Spill blocks are used to store system attribute data (i.e. file metadata) that does not fit in the dnode's bonus buffer. By allowing a larger bonus buffer area the use of a spill block can be avoided. Spill blocks potentially incur an additional read I/O for every dnode in a dnode block. As a worst case example, reading 32 dnodes from a 16k dnode block and all of the spill blocks could issue 33 separate reads. Now suppose those dnodes have size 1024 and therefore don't need spill blocks. Then the worst case number of blocks read is reduced to from 33 to two--one per dnode block. In practice spill blocks may tend to be co-located on disk with the dnode blocks so the reduction in I/O would not be this drastic. In a badly fragmented pool, however, the improvement could be significant. ZFS-on-Linux systems that make heavy use of extended attributes would benefit from this feature. In particular, ZFS-on-Linux supports the xattr=sa dataset property which allows file extended attribute data to be stored in the dnode bonus buffer as an alternative to the traditional directory-based format. Workloads such as SELinux and the Lustre distributed filesystem often store enough xattr data to force spill bocks when xattr=sa is in effect. Large dnodes may therefore provide a performance benefit to such systems. Other use cases that may benefit from this feature include files with large ACLs and symbolic links with long target names. Furthermore, this feature may be desirable on other platforms in case future applications or features are developed that could make use of a larger bonus buffer area. Implementation -------------- The size of a dnode may be a multiple of 512 bytes up to the size of a dnode block (currently 16384 bytes). A dn_extra_slots field was added to the current on-disk dnode_phys_t structure to describe the size of the physical dnode on disk. The 8 bits for this field were taken from the zero filled dn_pad2 field. The field represents how many "extra" dnode_phys_t slots a dnode consumes in its dnode block. This convention results in a value of 0 for 512 byte dnodes which preserves on-disk format compatibility with older software. Similarly, the in-memory dnode_t structure has a new dn_num_slots field to represent the total number of dnode_phys_t slots consumed on disk. Thus dn->dn_num_slots is 1 greater than the corresponding dnp->dn_extra_slots. This difference in convention was adopted because, unlike on-disk structures, backward compatibility is not a concern for in-memory objects, so we used a more natural way to represent size for a dnode_t. The default size for newly created dnodes is determined by the value of a new "dnodesize" dataset property. By default the property is set to "legacy" which is compatible with older software. Setting the property to "auto" will allow the filesystem to choose the most suitable dnode size. Currently this just sets the default dnode size to 1k, but future code improvements could dynamically choose a size based on observed workload patterns. Dnodes of varying sizes can coexist within the same dataset and even within the same dnode block. For example, to enable automatically-sized dnodes, run # zfs set dnodesize=auto tank/fish The user can also specify literal values for the dnodesize property. These are currently limited to powers of two from 1k to 16k. The power-of-2 limitation is only for simplicity of the user interface. Internally the implementation can handle any multiple of 512 up to 16k, and consumers of the DMU API can specify any legal dnode value. The size of a new dnode is determined at object allocation time and stored as a new field in the znode in-memory structure. New DMU interfaces are added to allow the consumer to specify the dnode size that a newly allocated object should use. Existing interfaces are unchanged to avoid having to update every call site and to preserve compatibility with external consumers such as Lustre. The new interfaces names are given below. The versions of these functions that don't take a dnodesize parameter now just call the _dnsize() versions with a dnodesize of 0, which means use the legacy dnode size. New DMU interfaces: dmu_object_alloc_dnsize() dmu_object_claim_dnsize() dmu_object_reclaim_dnsize() New ZAP interfaces: zap_create_dnsize() zap_create_norm_dnsize() zap_create_flags_dnsize() zap_create_claim_norm_dnsize() zap_create_link_dnsize() The constant DN_MAX_BONUSLEN is renamed to DN_OLD_MAX_BONUSLEN. The spa_maxdnodesize() function should be used to determine the maximum bonus length for a pool. These are a few noteworthy changes to key functions: * The prototype for dnode_hold_impl() now takes a "slots" parameter. When the DNODE_MUST_BE_FREE flag is set, this parameter is used to ensure the hole at the specified object offset is large enough to hold the dnode being created. The slots parameter is also used to ensure a dnode does not span multiple dnode blocks. In both of these cases, if a failure occurs, ENOSPC is returned. Keep in mind, these failure cases are only possible when using DNODE_MUST_BE_FREE. If the DNODE_MUST_BE_ALLOCATED flag is set, "slots" must be 0. dnode_hold_impl() will check if the requested dnode is already consumed as an extra dnode slot by an large dnode, in which case it returns ENOENT. * The function dmu_object_alloc() advances to the next dnode block if dnode_hold_impl() returns an error for a requested object. This is because the beginning of the next dnode block is the only location it can safely assume to either be a hole or a valid starting point for a dnode. * dnode_next_offset_level() and other functions that iterate through dnode blocks may no longer use a simple array indexing scheme. These now use the current dnode's dn_num_slots field to advance to the next dnode in the block. This is to ensure we properly skip the current dnode's bonus area and don't interpret it as a valid dnode. zdb --- The zdb command was updated to display a dnode's size under the "dnsize" column when the object is dumped. For ZIL create log records, zdb will now display the slot count for the object. ztest ----- Ztest chooses a random dnodesize for every newly created object. The random distribution is more heavily weighted toward small dnodes to better simulate real-world datasets. Unused bonus buffer space is filled with non-zero values computed from the object number, dataset id, offset, and generation number. This helps ensure that the dnode traversal code properly skips the interior regions of large dnodes, and that these interior regions are not overwritten by data belonging to other dnodes. A new test visits each object in a dataset. It verifies that the actual dnode size matches what was stored in the ztest block tag when it was created. It also verifies that the unused bonus buffer space is filled with the expected data patterns. ZFS Test Suite -------------- Added six new large dnode-specific tests, and integrated the dnodesize property into existing tests for zfs allow and send/recv. Send/Receive ------------ ZFS send streams for datasets containing large dnodes cannot be received on pools that don't support the large_dnode feature. A send stream with large dnodes sets a DMU_BACKUP_FEATURE_LARGE_DNODE flag which will be unrecognized by an incompatible receiving pool so that the zfs receive will fail gracefully. While not implemented here, it may be possible to generate a backward-compatible send stream from a dataset containing large dnodes. The implementation may be tricky, however, because the send object record for a large dnode would need to be resized to a 512 byte dnode, possibly kicking in a spill block in the process. This means we would need to construct a new SA layout and possibly register it in the SA layout object. The SA layout is normally just sent as an ordinary object record. But if we are constructing new layouts while generating the send stream we'd have to build the SA layout object dynamically and send it at the end of the stream. For sending and receiving between pools that do support large dnodes, the drr_object send record type is extended with a new field to store the dnode slot count. This field was repurposed from unused padding in the structure. ZIL Replay ---------- The dnode slot count is stored in the uppermost 8 bits of the lr_foid field. The bits were unused as the object id is currently capped at 48 bits. Resizing Dnodes --------------- It should be possible to resize a dnode when it is dirtied if the current dnodesize dataset property differs from the dnode's size, but this functionality is not currently implemented. Clearly a dnode can only grow if there are sufficient contiguous unused slots in the dnode block, but it should always be possible to shrink a dnode. Growing dnodes may be useful to reduce fragmentation in a pool with many spill blocks in use. Shrinking dnodes may be useful to allow sending a dataset to a pool that doesn't support the large_dnode feature. Feature Reference Counting -------------------------- The reference count for the large_dnode pool feature tracks the number of datasets that have ever contained a dnode of size larger than 512 bytes. The first time a large dnode is created in a dataset the dataset is converted to an extensible dataset. This is a one-way operation and the only way to decrement the feature count is to destroy the dataset, even if the dataset no longer contains any large dnodes. The complexity of reference counting on a per-dnode basis was too high, so we chose to track it on a per-dataset basis similarly to the large_block feature. Signed-off-by: Ned Bass <bass6@llnl.gov> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #3542
2016-03-17 04:25:34 +03:00
if (attrsize <= DN_OLD_MAX_BONUSLEN && !sa->sa_force_spill)
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
return;
(void) dmu_tx_hold_object_impl(tx, tx->tx_objset, DMU_NEW_OBJECT,
THT_SPILL, 0, 0);
}
/*
* Hold SA attribute
*
* dmu_tx_hold_sa(dmu_tx_t *tx, sa_handle_t *, attribute, add, size)
*
* variable_size is the total size of all variable sized attributes
* passed to this function. It is not the total size of all
* variable size attributes that *may* exist on this object.
*/
void
dmu_tx_hold_sa(dmu_tx_t *tx, sa_handle_t *hdl, boolean_t may_grow)
{
uint64_t object;
sa_os_t *sa = tx->tx_objset->os_sa;
ASSERT(hdl != NULL);
object = sa_handle_object(hdl);
Improve performance by using dmu_tx_hold_*_by_dnode() In zfs_write() and dmu_tx_hold_sa(), we can use dmu_tx_hold_*_by_dnode() instead of dmu_tx_hold_*(), since we already have a dbuf from the target dnode in hand. This eliminates some calls to dnode_hold(), which can be expensive. This is especially impactful if several threads are accessing objects that are in the same block of dnodes, because they will contend for that dbuf's lock. We are seeing 10-20% performance wins for the sequential_writes tests in the performance test suite, when doing >=128K writes to files with recordsize=8K. This also removes some unnecessary casts that are in the area. Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed-by: Tony Nguyen <tony.nguyen@delphix.com> Signed-off-by: Matthew Ahrens <mahrens@delphix.com> Closes #9081
2019-07-30 19:18:30 +03:00
dmu_buf_impl_t *db = (dmu_buf_impl_t *)hdl->sa_bonus;
DB_DNODE_ENTER(db);
dmu_tx_hold_bonus_by_dnode(tx, DB_DNODE(db));
DB_DNODE_EXIT(db);
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
if (tx->tx_objset->os_sa->sa_master_obj == 0)
return;
if (tx->tx_objset->os_sa->sa_reg_attr_obj == 0 ||
tx->tx_objset->os_sa->sa_layout_attr_obj == 0) {
dmu_tx_hold_zap(tx, sa->sa_master_obj, B_TRUE, SA_LAYOUTS);
dmu_tx_hold_zap(tx, sa->sa_master_obj, B_TRUE, SA_REGISTRY);
dmu_tx_hold_zap(tx, DMU_NEW_OBJECT, B_TRUE, NULL);
dmu_tx_hold_zap(tx, DMU_NEW_OBJECT, B_TRUE, NULL);
}
dmu_tx_sa_registration_hold(sa, tx);
if (may_grow && tx->tx_objset->os_sa->sa_layout_attr_obj)
dmu_tx_hold_zap(tx, sa->sa_layout_attr_obj, B_TRUE, NULL);
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
if (sa->sa_force_spill || may_grow || hdl->sa_spill) {
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
ASSERT(tx->tx_txg == 0);
dmu_tx_hold_spill(tx, object);
Update to onnv_147 This is the last official OpenSolaris tag before the public development tree was closed.
2010-08-27 01:24:34 +04:00
} else {
dnode_t *dn;
DB_DNODE_ENTER(db);
dn = DB_DNODE(db);
if (dn->dn_have_spill) {
ASSERT(tx->tx_txg == 0);
dmu_tx_hold_spill(tx, object);
}
DB_DNODE_EXIT(db);
Update core ZFS code from build 121 to build 141.
2010-05-29 00:45:14 +04:00
}
}
Add linux kernel module support Setup linux kernel module support, this includes: - zfs context for kernel/user - kernel module build system integration - kernel module macros - kernel module symbol export - kernel module options Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2010-08-26 22:49:16 +04:00
Add 'dmu_tx' kstats entry Keep counters for the various reasons that a thread may end up in txg_wait_open() waiting on a new txg. This can be useful when attempting to determine why a particular workload is under performing. Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2012-01-20 22:58:57 +04:00
void
dmu_tx_init(void)
{
dmu_tx_ksp = kstat_create("zfs", 0, "dmu_tx", "misc",
KSTAT_TYPE_NAMED, sizeof (dmu_tx_stats) / sizeof (kstat_named_t),
KSTAT_FLAG_VIRTUAL);
if (dmu_tx_ksp != NULL) {
dmu_tx_ksp->ks_data = &dmu_tx_stats;
kstat_install(dmu_tx_ksp);
}
}
void
dmu_tx_fini(void)
{
if (dmu_tx_ksp != NULL) {
kstat_delete(dmu_tx_ksp);
dmu_tx_ksp = NULL;
}
}
Update build system and packaging Minimal changes required to integrate the SPL sources in to the ZFS repository build infrastructure and packaging. Build system and packaging: * Renamed SPL_* autoconf m4 macros to ZFS_*. * Removed redundant SPL_* autoconf m4 macros. * Updated the RPM spec files to remove SPL package dependency. * The zfs package obsoletes the spl package, and the zfs-kmod package obsoletes the spl-kmod package. * The zfs-kmod-devel* packages were updated to add compatibility symlinks under /usr/src/spl-x.y.z until all dependent packages can be updated. They will be removed in a future release. * Updated copy-builtin script for in-kernel builds. * Updated DKMS package to include the spl.ko. * Updated stale AUTHORS file to include all contributors. * Updated stale COPYRIGHT and included the SPL as an exception. * Renamed README.markdown to README.md * Renamed OPENSOLARIS.LICENSE to LICENSE. * Renamed DISCLAIMER to NOTICE. Required code changes: * Removed redundant HAVE_SPL macro. * Removed _BOOT from nvpairs since it doesn't apply for Linux. * Initial header cleanup (removal of empty headers, refactoring). * Remove SPL repository clone/build from zimport.sh. * Use of DEFINE_RATELIMIT_STATE and DEFINE_SPINLOCK removed due to build issues when forcing C99 compilation. * Replaced legacy ACCESS_ONCE with READ_ONCE. * Include needed headers for `current` and `EXPORT_SYMBOL`. Reviewed-by: Tony Hutter <hutter2@llnl.gov> Reviewed-by: Olaf Faaland <faaland1@llnl.gov> Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Pavel Zakharov <pavel.zakharov@delphix.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> TEST_ZIMPORT_SKIP="yes" Closes #7556
2018-02-16 04:53:18 +03:00
#if defined(_KERNEL)
Add linux kernel module support Setup linux kernel module support, this includes: - zfs context for kernel/user - kernel module build system integration - kernel module macros - kernel module symbol export - kernel module options Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2010-08-26 22:49:16 +04:00
EXPORT_SYMBOL(dmu_tx_create);
EXPORT_SYMBOL(dmu_tx_hold_write);
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
EXPORT_SYMBOL(dmu_tx_hold_write_by_dnode);
Add linux kernel module support Setup linux kernel module support, this includes: - zfs context for kernel/user - kernel module build system integration - kernel module macros - kernel module symbol export - kernel module options Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2010-08-26 22:49:16 +04:00
EXPORT_SYMBOL(dmu_tx_hold_free);
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
EXPORT_SYMBOL(dmu_tx_hold_free_by_dnode);
Add linux kernel module support Setup linux kernel module support, this includes: - zfs context for kernel/user - kernel module build system integration - kernel module macros - kernel module symbol export - kernel module options Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2010-08-26 22:49:16 +04:00
EXPORT_SYMBOL(dmu_tx_hold_zap);
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
EXPORT_SYMBOL(dmu_tx_hold_zap_by_dnode);
Add linux kernel module support Setup linux kernel module support, this includes: - zfs context for kernel/user - kernel module build system integration - kernel module macros - kernel module symbol export - kernel module options Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2010-08-26 22:49:16 +04:00
EXPORT_SYMBOL(dmu_tx_hold_bonus);
Add *_by-dnode routines Add *_by_dnode() routines for accessing objects given their dnode_t *, this is more efficient than accessing the object by (objset_t *, uint64_t object). This change converts some but not all of the existing consumers. As performance-sensitive code paths are discovered they should be converted to use these routines. Reviewed-by: Matthew Ahrens <mahrens@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Alex Zhuravlev <bzzz@whamcloud.com> Closes #5534 Issue #4802
2017-01-14 01:58:41 +03:00
EXPORT_SYMBOL(dmu_tx_hold_bonus_by_dnode);
Add linux kernel module support Setup linux kernel module support, this includes: - zfs context for kernel/user - kernel module build system integration - kernel module macros - kernel module symbol export - kernel module options Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2010-08-26 22:49:16 +04:00
EXPORT_SYMBOL(dmu_tx_abort);
EXPORT_SYMBOL(dmu_tx_assign);
EXPORT_SYMBOL(dmu_tx_wait);
EXPORT_SYMBOL(dmu_tx_commit);
Export symbol dmu_tx_mark_netfree() This symbol is needed by Lustre for the same reason it was needed by the ZPL. It should have been exported when the original patch was merged. Reviewed-by: George Melikov <mail@gmelikov.ru> Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov> Reviewed-by: Tony Hutter <hutter2@llnl.gov> Reviewed-by: Alex Zhuravlev <bzzz@whamcloud.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Closes #6660
2017-09-20 19:30:24 +03:00
EXPORT_SYMBOL(dmu_tx_mark_netfree);
Add linux kernel module support Setup linux kernel module support, this includes: - zfs context for kernel/user - kernel module build system integration - kernel module macros - kernel module symbol export - kernel module options Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
2010-08-26 22:49:16 +04:00
EXPORT_SYMBOL(dmu_tx_get_txg);
EXPORT_SYMBOL(dmu_tx_callback_register);
EXPORT_SYMBOL(dmu_tx_do_callbacks);
EXPORT_SYMBOL(dmu_tx_hold_spill);
EXPORT_SYMBOL(dmu_tx_hold_sa_create);
EXPORT_SYMBOL(dmu_tx_hold_sa);
#endif
Reference in New Issue Copy Permalink