mirror_zfs/module/zfs
Paul Dagnelie 350646563f Concurrent small allocation defeats large allocation
With the new parallel allocators scheme, there is a possibility for
a problem where two threads, allocating from the same allocator at
the same time, conflict with each other. There are two primary cases
to worry about. First, another thread working on another allocator
activates the same metaslab that the first thread was trying to
activate. This results in the first thread needing to go back and
reselect a new metaslab, even though it may have waited a long time
for this metaslab to load. Second, another thread working on the same
allocator may have activated a different metaslab while the first
thread was waiting for its metaslab to load. Both of these cases
can cause the first thread to be significantly delayed in issuing
its IOs. The second case can also cause metaslab load/unload churn;
because the metaslab is loaded but not fully activated, we never set
the selected_txg, which results in the metaslab being immediately
unloaded again. This process can repeat many times, wasting disk and
cpu resources. This is more likely to happen when the IO of the first
thread is a larger one (like a ZIL write) and the other thread is
doing a smaller write, because it is more likely to find an
acceptable metaslab quickly.

There are two primary changes. The first is to always proceed with
the allocation when returning from metaslab_activate if we were
preempted in either of the ways described in the previous section.
The second change is to set the selected_txg before we do the call
to activate so that even if the metaslab is not used for an
allocation, we won't immediately attempt to unload it.

Reviewed by: Jerry Jelinek <jerry.jelinek@joyent.com>
Reviewed by: Matt Ahrens <matt@delphix.com>
Reviewed by: Serapheim Dimitropoulos <serapheim.dimitro@delphix.com>
Reviewed by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Paul Dagnelie <pcd@delphix.com>
External-issue: DLPX-61314
Closes #8843
2020-01-22 13:48:56 -08:00
..
abd.c single-chunk scatter ABDs can be treated as linear 2020-01-22 13:48:56 -08:00
aggsum.c OpenZFS 9688 - aggsum_fini leaks memory 2018-10-19 12:08:03 -07:00
arc.c single-chunk scatter ABDs can be treated as linear 2020-01-22 13:48:56 -08:00
blkptr.c Undo c89 workarounds to match with upstream 2017-11-04 13:25:13 -07:00
bplist.c Change KM_PUSHPAGE -> KM_SLEEP 2015-01-16 14:41:26 -08:00
bpobj.c Stack overflow in recursive bpobj_iterate_impl 2019-03-06 09:50:55 -08:00
bptree.c Native Encryption for ZFS on Linux 2017-08-14 10:36:48 -07:00
bqueue.c Wait in 'S' state when send/recv pipe is blocking 2019-06-07 12:45:40 -07:00
cityhash.c OpenZFS 8484 - Implement aggregate sum and use for arc counters 2018-06-06 09:35:59 -07:00
dataset_kstats.c port async unlinked drain from illumos-nexenta 2019-02-12 10:41:15 -08:00
dbuf_stats.c Prefix all refcount functions with zfs_ 2018-10-01 10:42:05 -07:00
dbuf.c Fix lockdep recursive locking false positive in dbuf_destroy 2019-09-25 11:27:49 -07:00
ddt_zap.c fat zap should prefetch when iterating 2019-09-25 11:27:47 -07:00
ddt.c ztest: scrub ddt repair 2019-01-17 15:25:00 -08:00
dmu_diff.c Fix issues found with zfs diff 2018-05-01 11:24:20 -07:00
dmu_object.c Fix send/recv lost spill block 2019-05-07 15:18:44 -07:00
dmu_objset.c Avoid extra taskq_dispatch() calls by DMU 2019-09-25 11:27:48 -07:00
dmu_recv.c Always refuse receving non-resume stream when resume state exists 2019-09-25 11:27:51 -07:00
dmu_send.c Fix send/recv lost spill block 2019-05-07 15:18:44 -07:00
dmu_traverse.c Fix traverse_impl() kmem leak 2018-08-15 09:53:44 -07:00
dmu_tx.c Improve performance by using dmu_tx_hold_*_by_dnode() 2019-09-25 11:27:50 -07:00
dmu_zfetch.c Linux 5.2 compat: rw_tryupgrade() 2019-05-23 13:46:33 -07:00
dmu.c fat zap should prefetch when iterating 2019-09-25 11:27:47 -07:00
dnode_sync.c Reinstate raw receive check when truncating 2019-06-07 12:45:40 -07:00
dnode.c Fix zil replay panic when TX_REMOVE followed by TX_CREATE 2019-09-25 11:27:51 -07:00
dsl_bookmark.c Detect and prevent mixed raw and non-raw sends 2019-03-13 11:00:43 -07:00
dsl_crypt.c Remove VERIFY from dsl_dataset_crypt_stats() 2019-09-25 11:27:49 -07:00
dsl_dataset.c Mutex leak in dsl_dataset_hold_obj() 2019-03-21 10:36:58 -07:00
dsl_deadlist.c OpenZFS 7614, 9064 - zfs device evacuation/removal 2018-04-14 12:16:17 -07:00
dsl_deleg.c Update build system and packaging 2018-05-29 16:00:33 -07:00
dsl_destroy.c Ensure dsl_destroy_head() decrypts objsets 2019-09-25 11:27:49 -07:00
dsl_dir.c Fix TXG_MASK cstyle 2019-04-12 11:30:59 -07:00
dsl_pool.c port async unlinked drain from illumos-nexenta 2019-02-12 10:41:15 -08:00
dsl_prop.c Update build system and packaging 2018-05-29 16:00:33 -07:00
dsl_scan.c Disabled resilver_defer feature leads to looping resilvers 2019-09-25 11:27:51 -07:00
dsl_synctask.c OpenZFS 9425 - channel programs can be interrupted 2020-01-22 13:48:56 -08:00
dsl_userhold.c zfs should optionally send holds 2019-02-15 12:41:38 -08:00
edonr_zfs.c DLPX-44812 integrate EP-220 large memory scalability 2016-11-29 14:34:27 -08:00
fm.c OpenZFS 9580 - Add a hash-table on top of nvlist to speed-up operations 2018-07-30 11:30:03 -07:00
gzip.c Update build system and packaging 2018-05-29 16:00:33 -07:00
hkdf.c Encryption patch follow-up 2017-10-11 16:54:48 -04:00
lz4.c Reword comment in lz4_compress_zfs 2019-05-02 16:46:04 -07:00
lzjb.c Change KM_PUSHPAGE -> KM_SLEEP 2015-01-16 14:41:26 -08:00
Makefile.in Add TRIM support 2019-03-29 09:13:20 -07:00
metaslab.c Concurrent small allocation defeats large allocation 2020-01-22 13:48:56 -08:00
mmp.c MMP interval and fail_intervals in uberblock 2019-03-21 12:47:57 -07:00
multilist.c Avoid extra taskq_dispatch() calls by DMU 2019-09-25 11:27:48 -07:00
pathname.c Disable unused pathname::pn_path* (unneeded in Linux) 2019-09-25 11:27:49 -07:00
policy.c Implement secpolicy_vnode_setid_retain() 2019-09-25 11:27:50 -07:00
qat_compress.c QAT related bug fixes 2019-09-25 11:27:51 -07:00
qat_crypt.c QAT related bug fixes 2019-09-25 11:27:51 -07:00
qat.c QAT related bug fixes 2019-09-25 11:27:51 -07:00
qat.h QAT related bug fixes 2019-09-25 11:27:51 -07:00
range_tree.c Restrict kstats and print real pointers 2019-04-04 18:57:06 -07:00
refcount.c Add zfs_refcount_transfer_ownership_many() 2018-10-09 10:05:48 -07:00
rrwlock.c Prefix all refcount functions with zfs_ 2018-10-01 10:42:05 -07:00
sa.c Improve performance by using dmu_tx_hold_*_by_dnode() 2019-09-25 11:27:50 -07:00
sha256.c SHA256 QAT acceleration 2018-03-15 10:53:58 -07:00
skein_zfs.c DLPX-44812 integrate EP-220 large memory scalability 2016-11-29 14:34:27 -08:00
spa_boot.c Add linux kernel module support 2010-08-31 13:41:58 -07:00
spa_checkpoint.c Get rid of space_map_update() for ms_synced_length 2019-02-12 10:38:11 -08:00
spa_config.c Fix /etc/hostid on root pool deadlock 2019-09-25 11:27:51 -07:00
spa_errlog.c Update build system and packaging 2018-05-29 16:00:33 -07:00
spa_history.c Create /proc/sys/kernel/spl/gitrev with git hash 2018-10-08 21:57:02 -07:00
spa_misc.c Fix /etc/hostid on root pool deadlock 2019-09-25 11:27:51 -07:00
spa_stats.c Restrict kstats and print real pointers 2019-04-04 18:57:06 -07:00
spa.c Fix /etc/hostid on root pool deadlock 2019-09-25 11:27:51 -07:00
space_map.c Restrict kstats and print real pointers 2019-04-04 18:57:06 -07:00
space_reftree.c OpenZFS 7614, 9064 - zfs device evacuation/removal 2018-04-14 12:16:17 -07:00
THIRDPARTYLICENSE.cityhash OpenZFS 8484 - Implement aggregate sum and use for arc counters 2018-06-06 09:35:59 -07:00
THIRDPARTYLICENSE.cityhash.descrip OpenZFS 8484 - Implement aggregate sum and use for arc counters 2018-06-06 09:35:59 -07:00
trace.c OpenZFS 7614, 9064 - zfs device evacuation/removal 2018-04-14 12:16:17 -07:00
txg.c OpenZFS 9425 - channel programs can be interrupted 2020-01-22 13:48:56 -08:00
uberblock.c MMP interval and fail_intervals in uberblock 2019-03-21 12:47:57 -07:00
unique.c Performance optimization of AVL tree comparator functions 2016-08-31 14:35:34 -07:00
vdev_cache.c Update build system and packaging 2018-05-29 16:00:33 -07:00
vdev_disk.c Scrubbing root pools may deadlock on kernels without elevator_change() (#9321) 2019-09-25 11:27:51 -07:00
vdev_file.c Update vdev_ops_t from illumos 2019-09-25 11:27:48 -07:00
vdev_indirect_births.c Fixes: #8934 Large kmem_alloc 2019-09-25 11:27:49 -07:00
vdev_indirect_mapping.c Get rid of space_map_update() for ms_synced_length 2019-02-12 10:38:11 -08:00
vdev_indirect.c Update vdev_ops_t from illumos 2019-09-25 11:27:48 -07:00
vdev_initialize.c Add TRIM support 2019-03-29 09:13:20 -07:00
vdev_label.c panic in removal_remap test on 4K devices 2019-09-25 11:27:47 -07:00
vdev_mirror.c Update vdev_ops_t from illumos 2019-09-25 11:27:48 -07:00
vdev_missing.c Update vdev_ops_t from illumos 2019-09-25 11:27:48 -07:00
vdev_queue.c Move write aggregation memory copy out of vq_lock 2019-09-25 11:27:47 -07:00
vdev_raidz_math_aarch64_neon_common.h Linux 5.0 compat: ASM_BUG macro 2019-05-08 10:18:40 -07:00
vdev_raidz_math_aarch64_neon.c codebase style improvements for OpenZFS 6459 port 2017-01-22 13:25:40 -08:00
vdev_raidz_math_aarch64_neonx2.c ABD raidz NEON support 2016-11-29 14:34:33 -08:00
vdev_raidz_math_avx2.c Linux 5.0 compat: ASM_BUG macro 2019-05-08 10:18:40 -07:00
vdev_raidz_math_avx512bw.c Linux 5.0 compat: ASM_BUG macro 2019-05-08 10:18:40 -07:00
vdev_raidz_math_avx512f.c Use cstyle -cpP in make cstyle check 2016-12-12 10:46:26 -08:00
vdev_raidz_math_impl.h codebase style improvements for OpenZFS 6459 port 2017-01-22 13:25:40 -08:00
vdev_raidz_math_scalar.c Linux 5.3: Fix switch() fall though compiler errors 2019-09-25 11:27:50 -07:00
vdev_raidz_math_sse2.c ABD raidz avx512f support 2016-11-29 14:34:33 -08:00
vdev_raidz_math_ssse3.c Linux 5.0 compat: ASM_BUG macro 2019-05-08 10:18:40 -07:00
vdev_raidz_math.c Fix typo in vdev_raidz_math.c 2019-09-25 11:27:47 -07:00
vdev_raidz.c Update vdev_ops_t from illumos 2019-09-25 11:27:48 -07:00
vdev_removal.c panic in removal_remap test on 4K devices 2019-09-25 11:27:47 -07:00
vdev_root.c Update vdev_ops_t from illumos 2019-09-25 11:27:48 -07:00
vdev_trim.c Add TRIM support 2019-03-29 09:13:20 -07:00
vdev.c Allow metaslab to be unloaded even when not freed from 2019-09-25 11:27:47 -07:00
zap_leaf.c Off-by-one in zap_leaf_array_create() 2019-01-18 09:58:46 -08:00
zap_micro.c fat zap should prefetch when iterating 2019-09-25 11:27:47 -07:00
zap.c fat zap should prefetch when iterating 2019-09-25 11:27:47 -07:00
zcp_get.c Fix get_special_prop() build failure 2019-09-25 11:27:49 -07:00
zcp_global.c OpenZFS 8600 - ZFS channel programs - snapshot 2018-02-08 15:29:24 -08:00
zcp_iter.c OpenZFS 9337 - zfs get all is slow due to uncached metadata 2018-07-12 10:49:27 -07:00
zcp_synctask.c OpenZFS 9166 - zfs storage pool checkpoint 2018-06-26 10:07:42 -07:00
zcp.c OpenZFS 9425 - channel programs can be interrupted 2020-01-22 13:48:56 -08:00
zfeature.c Consistently captialize GUID for features 2019-04-16 10:01:51 -07:00
zfs_acl.c Update build system and packaging 2018-05-29 16:00:33 -07:00
zfs_byteswap.c Update build system and packaging 2018-05-29 16:00:33 -07:00
zfs_ctldir.c Use zfsctl_snapshot_hold() wrapper 2019-09-25 11:27:49 -07:00
zfs_debug.c Restrict kstats and print real pointers 2019-04-04 18:57:06 -07:00
zfs_dir.c port async unlinked drain from illumos-nexenta 2019-02-12 10:41:15 -08:00
zfs_fm.c Add zpool status -s (slow I/Os) and -p (parseable) 2018-11-08 16:47:24 -08:00
zfs_fuid.c Update build system and packaging 2018-05-29 16:00:33 -07:00
zfs_ioctl.c zfs_ioc_snapshot: check user-prop permissions on snapshotted datasets 2019-09-25 11:27:50 -07:00
zfs_log.c make zil max block size tunable 2020-01-22 13:48:56 -08:00
zfs_onexit.c Update build system and packaging 2018-05-29 16:00:33 -07:00
zfs_ratelimit.c Change checksum & IO delay ratelimit values 2018-03-04 17:34:51 -08:00
zfs_replay.c Fix zil replay panic when TX_REMOVE followed by TX_CREATE 2019-09-25 11:27:51 -07:00
zfs_rlock.c OpenZFS 9689 - zfs range lock code should not be zpl-specific 2018-10-11 10:19:33 -07:00
zfs_sa.c Project Quota on ZFS 2018-02-13 14:54:54 -08:00
zfs_sysfs.c Prevent pointer to an out-of-scope local variable 2019-09-25 11:27:48 -07:00
zfs_vfsops.c Fix deadlock in 'zfs rollback' 2019-09-25 11:27:50 -07:00
zfs_vnops.c Fix out-of-order ZIL txtype lost on hardlinked files 2019-09-25 11:27:50 -07:00
zfs_znode.c Fix deadlock in 'zfs rollback' 2019-09-25 11:27:50 -07:00
zil.c make zil max block size tunable 2020-01-22 13:48:56 -08:00
zio_checksum.c Undo c89 workarounds to match with upstream 2017-11-04 13:25:13 -07:00
zio_compress.c OpenZFS 9403 - assertion failed in arc_buf_destroy() 2018-08-29 11:33:33 -07:00
zio_crypt.c Always call rw_init in zio_crypt_key_unwrap 2019-04-10 15:39:40 -07:00
zio_inject.c Multiple DVA Scrubbing Fix 2019-03-15 14:14:31 -07:00
zio.c Fix bp_embedded_type enum definition 2020-01-22 13:48:56 -08:00
zle.c Fix zle_decompress out of bound access 2018-02-09 10:08:05 -08:00
zpl_ctldir.c RHEL 7.5 compat: FMODE_KABI_ITERATE 2018-05-02 15:01:24 -07:00
zpl_export.c Use cstyle -cpP in make cstyle check 2016-12-12 10:46:26 -08:00
zpl_file.c Fix errant EFAULT during writes (#8719) 2019-05-08 10:04:04 -07:00
zpl_inode.c Fix errant EFAULT during writes (#8719) 2019-05-08 10:04:04 -07:00
zpl_super.c Fix statfs(2) for 32-bit user space 2018-09-24 17:11:25 -07:00
zpl_xattr.c Drop redundant POSIX ACL check in zpl_init_acl() 2019-09-25 11:27:49 -07:00
zrlock.c Update build system and packaging 2018-05-29 16:00:33 -07:00
zthr.c Fix txg_wait_open() load average inflation 2019-04-04 09:44:46 -07:00
zvol.c make zil max block size tunable 2020-01-22 13:48:56 -08:00