Apply similar options to BLAKE3 as it is done for zfs_fletcher_4_impl.
The zfs module parameter on Linux changes from icp_blake3_impl to
zfs_blake3_impl.
You can check and set it on Linux via sysfs like this:
```
[bash]# cat /sys/module/zfs/parameters/zfs_blake3_impl
cycle [fastest] generic sse2 sse41 avx2
[bash]# echo sse2 > /sys/module/zfs/parameters/zfs_blake3_impl
[bash]# cat /sys/module/zfs/parameters/zfs_blake3_impl
cycle fastest generic [sse2] sse41 avx2
```
The modprobe module parameters may also be used now:
```
[bash]# modprobe zfs zfs_blake3_impl=sse41
[bash]# cat /sys/module/zfs/parameters/zfs_blake3_impl
cycle fastest generic sse2 [sse41] avx2
```
On FreeBSD the BLAKE3 implementation can be set via sysctl like this:
```
[bsd]# sysctl vfs.zfs.blake3_impl
vfs.zfs.blake3_impl: cycle [fastest] generic sse2 sse41 avx2
[bsd]# sysctl vfs.zfs.blake3_impl=sse2
vfs.zfs.blake3_impl: cycle [fastest] generic sse2 sse41 avx2 \
-> cycle fastest generic [sse2] sse41 avx2
```
This commit changes also some Blake3 internals like these:
- blake3_impl_ops_t was renamed to blake3_ops_t
- all functions are named blake3_impl_NAME() now
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Co-authored-by: Ryan Moeller <ryan@iXsystems.com>
Signed-off-by: Tino Reichardt <milky-zfs@mcmilk.de>
Closes#13725
- Some optimizations for bqueue enqueue/dequeue.
- Added a fix to prevent deadlock when both bqueue_enqueue_impl()
and bqueue_dequeue() waits for signal to be triggered.
Reviewed-by: Alexander Motin <mav@FreeBSD.org>
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Signed-off-by: Ameer Hamza <ahamza@ixsystems.com>
Closes#13855
Replace ZFS_ENTER and ZFS_VERIFY_ZP, which have hidden returns, with
functions that return error code. The reason we want to do this is
because hidden returns are not obvious and had caused some missing fail
path unwinding.
This patch changes the common, linux, and freebsd parts. Also fixes
fail path unwinding in zfs_fsync, zpl_fsync, zpl_xattr_{list,get,set}, and
zfs_lookup().
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Signed-off-by: Chunwei Chen <david.chen@nutanix.com>
Closes#13831
Unused code detected by coverity.
Reviewed-by: Allan Jude <allan@klarasystems.com>
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Neal Gompa <ngompa@datto.com>
Signed-off-by: Richard Yao <richard.yao@alumni.stonybrook.edu>
Closes#13868
In our codebase, `cond_resched() and `schedule()` are Linux kernel
functions that have replaced the OpenSolaris `kpreempt()` functions in
the codebase to such an extent that `kpreempt()` in zfs_context.h was
broken. Nobody noticed because we did not actually use it. The header
had defined `kpreempt()` as `yield()`, which works on OpenSolaris and
Illumos where `sched_yield()` is a wrapper for `yield()`, but that does
not work on any other platform.
The FreeBSD platform specific code implemented shims for these, but the
shim for `schedule()` forced us to wait, which is different than merely
rescheduling to another thread as the original Linux code does, while
the shim for `cond_resched()` had the same definition as its kernel
kpreempt() shim.
After studying this, I have concluded that we should reintroduce the
kpreempt() function in platform independent code with the following
definitions:
- In the Linux kernel:
kpreempt(unused) -> cond_resched()
- In the FreeBSD kernel:
kpreempt(unused) -> kern_yield(PRI_USER)
- In userspace:
kpreempt(unused) -> sched_yield()
In userspace, nothing changes from this cleanup. In the kernels, the
function `fm_fini()` will now call `kern_yield(PRI_USER)` on FreeBSD and
`cond_resched()` on Linux. This is instead of `pause("schedule", 1)` on
FreeBSD and `schedule()` on Linux. This makes our behavior consistent
across platforms.
Note that Linux's SPL continues to use `cond_resched()` and
`schedule()`. However, those functions have been removed from both the
FreeBSD code and userspace code.
This should have the benefit of making it slightly easier to port the
code to new platforms by making how things should be mapped less
confusing.
Reviewed-by: Alexander Motin <mav@FreeBSD.org>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Neal Gompa <ngompa@datto.com>
Signed-off-by: Richard Yao <richard.yao@alumni.stonybrook.edu>
Closes#13845
Users were seeing floods of `config_sync` events when autoexpand was
enabled. This happened because all "disk status change" udev events
invoke the autoexpand codepath, which calls zpool_relabel_disk(),
which in turn cause another "disk status change" event to happen,
in a feedback loop. Note that "disk status change" happens every time
a user calls close() on a block device.
This commit breaks the feedback loop by only allowing an autoexpand
to happen if the disk actually changed size.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tony Hutter <hutter2@llnl.gov>
Closes: #7132Closes: #7366Closes#13729
When iterating through children physical ashifts for vdev, prefer
ones above the maximum logical ashift, that we can actually use,
but within the administrator defined maximum.
When selecting top-level vdev ashift, do not set it to the defined
maximum in case physical ashift is even higher, but just ignore one.
Using the maximum does not prevent misaligned writes, but reduces
space efficiency. Since ZFS tries to write data sequentially and
aggregates the writes, in many cases large misanigned writes may be
not as bad as the space penalty otherwise.
Allow internal physical ashifts for vdevs higher than SHIFT_MAX.
May be one day allocator or aggregation could benefit from that.
Reduce zfs_vdev_max_auto_ashift default from 16 (64KB) to 14 (16KB),
so that ZFS may still use bigger ashifts up to SHIFT_MAX (64KB),
but only if it really has to or explicitly told to, but not as an
"optimization".
There are some read-intensive NVMe SSDs that report Preferred Write
Alignment of 64KB, and attempt to build RAIDZ2 of those leads to a
space inefficiency that can't be justified. Instead these changes
make ZFS fall back to logical ashift of 12 (4KB) by default and
only warn user that it may be suboptimal for performance.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Signed-off-by: Alexander Motin <mav@FreeBSD.org>
Sponsored by: iXsystems, Inc.
Closes#13798
Without this patch, the
ASSERT3U(dbuf_is_metadata(db), ==, arc_is_metadata(buf));
at the beginning of dbuf_assign_arcbuf can panic
if the object type is a DMU_OT_NEWTYPE that has
DMU_OT_METADATA set.
While we're at it, fix DMU_OT_IS_ENCRYPTED as well.
Reviewed-by: Richard Yao <richard.yao@alumni.stonybrook.edu>
Reviewed-by: Alexander Motin <mav@FreeBSD.org>
Signed-off-by: Christian Schwarz <christian.schwarz@nutanix.com>
Closes#13842
This commit adds DD_FIELD string used in extensified dsl_dir zap object
for snapshots_changed property.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Signed-off-by: Umer Saleem <usaleem@ixsystems.com>
Closes#13819
Only the single snapshot rename is provided.
The recursive or more complex rename can be scripted.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: George Melikov <mail@gmelikov.ru>
Signed-off-by: Andriy Gapon <avg@FreeBSD.org>
Closes#13802
This reverts commit 80a650b7bb. This change
inadvertently introduced a regression in ztest where one of the new ASSERTs
is triggered in dsl_scan_visitbp().
Reviewed-by: George Amanakis <gamanakis@gmail.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Issue #12275Closes#13799
When scrubbing an encrypted filesystem with unloaded key still report an
error in zpool status.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Alek Pinchuk <apinchuk@axcient.com>
Signed-off-by: George Amanakis <gamanakis@gmail.com>
Closes#13675Closes#13717
Make dd_snap_cmtime property persistent across mount and unmount
operations by storing in ZAP and restore the value from ZAP on hold
into dd_snap_cmtime instead of updating it.
Expose dd_snap_cmtime as 'snapshots_changed' property that provides a
mechanism to quickly determine whether snapshot list for dataset has
changed without having to mount a dataset or iterate the snapshot list.
It specifies the time at which a snapshot for a dataset was last
created or deleted. This allows us to be more efficient how often we
query snapshots.
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Reviewed-by: Alexander Motin <mav@FreeBSD.org>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Umer Saleem <usaleem@ixsystems.com>
Closes#13635
This type of recv is used to heal corrupted data when a replica
of the data already exists (in the form of a send file for example).
With the provided send stream, corrective receive will read from
disk blocks described by the WRITE records. When any of the reads
come back with ECKSUM we use the data from the corresponding WRITE
record to rewrite the corrupted block.
Reviewed-by: Paul Dagnelie <pcd@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Paul Zuchowski <pzuchowski@datto.com>
Signed-off-by: Alek Pinchuk <apinchuk@axcient.com>
Closes#9372
ZIL kstats are reported in an inclusive way, i.e., same counters are
shared to capture all the activities happening in zil. Added support
to report zil stats for every datset individually by combining them
with already exposed dataset kstats.
Wmsum uses per cpu counters and provide less overhead as compared
to atomic operations. Updated zil kstats to replace wmsum counters
to avoid atomic operations.
Reviewed-by: Christian Schwarz <christian.schwarz@nutanix.com>
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Reviewed-by: Alexander Motin <mav@FreeBSD.org>
Signed-off-by: Ameer Hamza <ahamza@ixsystems.com>
Closes#13636
It may happen that scan bookmark points to a block that was turned
into a part of a big hole. In such case dsl_scan_visitbp() may skip
it and dsl_scan_check_resume() will not be called for it. As result
new scan suspend won't be possible until the end of the object, that
may take hours if the object is a multi-terabyte ZVOL on a slow HDD
pool, stretching TXG to all that time, creating all sorts of problems.
This patch changes the resume condition to any greater or equal block,
so even if we miss the bookmarked block, the next one we find will
delete the bookmark, allowing new suspend.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Signed-off-by: Alexander Motin <mav@FreeBSD.org>
Sponsored-By: iXsystems, Inc.
Closes#13643
Makes the case sensitivity setting visible on Linux in /proc/mounts.
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Reviewed-by: Alexander Motin <mav@FreeBSD.org>
Signed-off-by: Ameer Hamza <ahamza@ixsystems.com>
Closes#13607
Block statistics calculation during scrub I/O issue in case of sorted
scrub accounted ditto blocks several times. Embedded blocks on other
side were not accounted at all. This change moves the accounting from
issue to scan stage, that fixes both problems and also allows to avoid
pool-wide locking and the lock contention it created.
Since this statistics is quite specific and is not even exposed now
anywhere, disable its calculation by default to not waste CPU time.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Alexander Motin <mav@FreeBSD.org>
Sponsored-By: iXsystems, Inc.
Closes#13579
Restructure the code in zfs_log_xvattr() to use a lr_attr_end
structure when accessing lr_attr_t elements located after the
variable sized array. This makes the code more understandable
and resolves the accessing beyond the end of the field warnings.
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Reviewed-by: Alexander Motin <mav@FreeBSD.org>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#13528Closes#13575
The current codebase does not support raw sending buffers with block
size > 128kB when large_blocks is not active. This can happen in the
codepath dsl_dataset_sync()->dmu_objset_sync()->zio_nowait() which
calls back dmu_objset_write_done()->dsl_dataset_block_born(). If
dsl_dataset_sync() completes its run before dsl_dataset_block_born() is
called, we will end up not activating some of the necessary flags, while
having blocks based on those flags written in the filesystem. A
subsequent send will then panic.
Fix this by directly deciding in dmu_objset_sync() whether these flags
need to be activated later by dsl_dataset_sync(). Instead of panicking
due to a NULL pointer dereference in dmu_dump_write() in case of a send,
print out an error message. Also during scrub verify there are no
contradicting filesystem flags.
Reviewed-by: Paul Dagnelie <pcd@delphix.com>
Signed-off-by: George Amanakis <gamanakis@gmail.com>
Closes#12275Closes#12438
- Introduce first element offset within a leaf. It allows to reduce
by ~50% average memmove() size when adding/removing elements. If the
added/removed element is in the first half of the leaf, we may shift
elements before it and adjust the bth_first instead of moving more
elements after it.
- Use memcpy() instead of memmove() when we know there is no overlap.
- Switch from uint64_t to uint32_t. It does not limit anything,
but 32-bit arches should appreciate it greatly in hot paths.
- Store leaf capacity in struct btree to avoid 64-bit divisions.
- Adjust zfs_btree_insert_into_leaf() to always result in balanced
leaves after splitting, no matter where the new element was inserted.
Not that we care about it much, but it should also allow B-trees with
as little as two elements per leaf instead of 4 previously.
When scrubbing pool of 12 SSDs, storing 1.5TB of 4KB zvol blocks this
reduces amount of time spent in memmove() inside the scan thread from
13.7% to 5.7% and total scrub time by ~15 seconds out of 9 minutes.
It should also reduce spacemaps load time, but I haven't measured it.
Reviewed-by: Paul Dagnelie <pcd@delphix.com>
Signed-off-by: Alexander Motin <mav@FreeBSD.org>
Sponsored-By: iXsystems, Inc.
Closes#13582
- Reduce size and comparison complexity of q_exts_by_size B-tree.
Previous code used two 64-bit divisions and many other operations to
compare two B-tree elements. It created enormous overhead. This
implementation moves the math to the upper level and stores the score
in the B-tree elements themselves. Since all that we need to store in
that B-tree is the extent score and offset, those can fit into single
8 byte value instead of 24 bytes of q_exts_by_addr element and can be
compared with single operation.
- Better decouple secondary tree logic from main range_tree by moving
rt_btree_ops and related functions into dsl_scan.c as ext_size_ops.
Those functions are very small to worry about the code duplication and
range_tree does not need to know details such as rt_btree_compare.
- Instead of accounting number of pending bytes per pool, that needs
atomic on global variable per block, account the number of non-empty
per-vdev queues, that change much more rarely.
- When extent scan is interrupted by TXG end, continue it in the next
TXG instead of selecting next best extent. It allows to avoid leaving
one truncated (and so likely not the best any more) extent each TXG.
On top of some other optimizations this saves about 1.5 minutes out of
10 to scrub pool of 12 SSDs, storing 1.5TB of 4KB zvol blocks.
Reviewed-by: Paul Dagnelie <pcd@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Tom Caputi <caputit1@tcnj.edu>
Signed-off-by: Alexander Motin <mav@FreeBSD.org>
Sponsored-By: iXsystems, Inc.
Closes#13576
The kmem_alloc(sizeof (*ctx), KM_NOSLEEP) call on FreeBSD can't be
used in this code segment. Work around this by pre-allocating a percpu
context array for later use.
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tino Reichardt <milky-zfs@mcmilk.de>
Closes#13568
This allows ZFS datasets to be delegated to a user/mount namespace
Within that namespace, only the delegated datasets are visible
Works very similarly to Zones/Jailes on other ZFS OSes
As a user:
```
$ unshare -Um
$ zfs list
no datasets available
$ echo $$
1234
```
As root:
```
# zfs list
NAME ZONED MOUNTPOINT
containers off /containers
containers/host off /containers/host
containers/host/child off /containers/host/child
containers/host/child/gchild off /containers/host/child/gchild
containers/unpriv on /unpriv
containers/unpriv/child on /unpriv/child
containers/unpriv/child/gchild on /unpriv/child/gchild
# zfs zone /proc/1234/ns/user containers/unpriv
```
Back to the user namespace:
```
$ zfs list
NAME USED AVAIL REFER MOUNTPOINT
containers 129M 47.8G 24K /containers
containers/unpriv 128M 47.8G 24K /unpriv
containers/unpriv/child 128M 47.8G 128M /unpriv/child
```
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Will Andrews <will.andrews@klarasystems.com>
Signed-off-by: Allan Jude <allan@klarasystems.com>
Signed-off-by: Mateusz Piotrowski <mateusz.piotrowski@klarasystems.com>
Co-authored-by: Allan Jude <allan@klarasystems.com>
Co-authored-by: Mateusz Piotrowski <mateusz.piotrowski@klarasystems.com>
Sponsored-by: Buddy <https://buddy.works>
Closes#12263
This commit adds BLAKE3 checksums to OpenZFS, it has similar
performance to Edon-R, but without the caveats around the latter.
Homepage of BLAKE3: https://github.com/BLAKE3-team/BLAKE3
Wikipedia: https://en.wikipedia.org/wiki/BLAKE_(hash_function)#BLAKE3
Short description of Wikipedia:
BLAKE3 is a cryptographic hash function based on Bao and BLAKE2,
created by Jack O'Connor, Jean-Philippe Aumasson, Samuel Neves, and
Zooko Wilcox-O'Hearn. It was announced on January 9, 2020, at Real
World Crypto. BLAKE3 is a single algorithm with many desirable
features (parallelism, XOF, KDF, PRF and MAC), in contrast to BLAKE
and BLAKE2, which are algorithm families with multiple variants.
BLAKE3 has a binary tree structure, so it supports a practically
unlimited degree of parallelism (both SIMD and multithreading) given
enough input. The official Rust and C implementations are
dual-licensed as public domain (CC0) and the Apache License.
Along with adding the BLAKE3 hash into the OpenZFS infrastructure a
new benchmarking file called chksum_bench was introduced. When read
it reports the speed of the available checksum functions.
On Linux: cat /proc/spl/kstat/zfs/chksum_bench
On FreeBSD: sysctl kstat.zfs.misc.chksum_bench
This is an example output of an i3-1005G1 test system with Debian 11:
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1196 1602 1761 1749 1762 1759 1751
skein-generic 546 591 608 615 619 612 616
sha256-generic 240 300 316 314 304 285 276
sha512-generic 353 441 467 476 472 467 426
blake3-generic 308 313 313 313 312 313 312
blake3-sse2 402 1289 1423 1446 1432 1458 1413
blake3-sse41 427 1470 1625 1704 1679 1607 1629
blake3-avx2 428 1920 3095 3343 3356 3318 3204
blake3-avx512 473 2687 4905 5836 5844 5643 5374
Output on Debian 5.10.0-10-amd64 system: (Ryzen 7 5800X)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1840 2458 2665 2719 2711 2723 2693
skein-generic 870 966 996 992 1003 1005 1009
sha256-generic 415 442 453 455 457 457 457
sha512-generic 608 690 711 718 719 720 721
blake3-generic 301 313 311 309 309 310 310
blake3-sse2 343 1865 2124 2188 2180 2181 2186
blake3-sse41 364 2091 2396 2509 2463 2482 2488
blake3-avx2 365 2590 4399 4971 4915 4802 4764
Output on Debian 5.10.0-9-powerpc64le system: (POWER 9)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 1213 1703 1889 1918 1957 1902 1907
skein-generic 434 492 520 522 511 525 525
sha256-generic 167 183 187 188 188 187 188
sha512-generic 186 216 222 221 225 224 224
blake3-generic 153 152 154 153 151 153 153
blake3-sse2 391 1170 1366 1406 1428 1426 1414
blake3-sse41 352 1049 1212 1174 1262 1258 1259
Output on Debian 5.10.0-11-arm64 system: (Pi400)
implementation 1k 4k 16k 64k 256k 1m 4m
edonr-generic 487 603 629 639 643 641 641
skein-generic 271 299 303 308 309 309 307
sha256-generic 117 127 128 130 130 129 130
sha512-generic 145 165 170 172 173 174 175
blake3-generic 81 29 71 89 89 89 89
blake3-sse2 112 323 368 379 380 371 374
blake3-sse41 101 315 357 368 369 364 360
Structurally, the new code is mainly split into these parts:
- 1x cross platform generic c variant: blake3_generic.c
- 4x assembly for X86-64 (SSE2, SSE4.1, AVX2, AVX512)
- 2x assembly for ARMv8 (NEON converted from SSE2)
- 2x assembly for PPC64-LE (POWER8 converted from SSE2)
- one file for switching between the implementations
Note the PPC64 assembly requires the VSX instruction set and the
kfpu_begin() / kfpu_end() calls on PowerPC were updated accordingly.
Reviewed-by: Felix Dörre <felix@dogcraft.de>
Reviewed-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tino Reichardt <milky-zfs@mcmilk.de>
Co-authored-by: Rich Ercolani <rincebrain@gmail.com>
Closes#10058Closes#12918
As of the Linux 5.19 kernel an identically named zap_flags_t typedef
is declared in the include/linux/mm_types.h linux header. Sadly,
the inclusion of this header cannot be easily avoided. To resolve
the conflict a #define is used to remap the name in the OpenZFS
sources when building against the Linux kernel.
Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#13515
In current zil_commit() process, transaction lwb_tx is assigned in
zil_lwb_write_issue(), and is committed in zil_lwb_flush_vdevs_done().
Thus, during lwb write out process, the txg is held in open or quiesing
state, until zil_lwb_flush_vdevs_done() is called. If the zil's zio
latency is high, it will cause txg_sync_thread() to starve.
The goal here is to defer waiting for zil_lwb_flush_vdevs_done to the
'syncing' txg state. That is, in zil_sync().
In this patch, it achieves the goal without holding transaction.
A new function zil_lwb_flush_wait_all() is introduced. It waits for
the completion of all the zil_lwb_flush_vdevs_done() by given txg.
Reviewed-by: Alexander Motin <mav@FreeBSD.org>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Prakash Surya <prakash.surya@delphix.com>
Signed-off-by: jxdking <lostking2008@hotmail.com>
Closes#12321
This was breaking the kmod port build on FreeBSD with Clang 13.
Use the same trick as we do for ASSERT() to make DNODE_VERIFY() use
its parameter at compile time without actually using it at run time
in non-debug builds.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ryan Moeller <ryan@iXsystems.com>
Closes#13507
- Make prefetch distance adaptive: up to 4MB prefetch doubles for
every, hit same as before, but after that it grows by 1/8 every time
the prefetch read does not complete in time to satisfy the demand.
My tests show that 4MB is sufficient for wide NVMe pool to saturate
single reader thread at 2.5GB/s, while new 64MB maximum allows the
same thread to reach 1.5GB/s on wide HDD pool. Further distance
increase may increase speed even more, but less dramatic and with
higher latency.
- Allow early reuse of inactive prefetch streams: streams that never
saw hits can be reused immediately if there is a demand, while others
can be reused after 1s of inactivity, starting with the oldest. After
2s of inactivity streams are deleted to free resources same as before.
This allows by several times increase strided read performance on HDD
pool in presence of simultaneous random reads, previously filling the
zfetch_max_streams limit for seconds and so blocking most of prefetch.
- Always issue intermediate indirect block reads with SYNC priority.
Each of those reads if delayed for longer may delay up to 1024 other
block prefetches, that may be not good for wide pools.
Reviewed-by: Allan Jude <allan@klarasystems.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Alexander Motin <mav@FreeBSD.org>
Sponsored-By: iXsystems, Inc.
Closes#13452
It turns out that wrapping the atomic macro in () breaks build
on Linux/SPARC64. Oops.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Rich Ercolani <rincebrain@gmail.com>
Closes#13506
Original Log Size Limit implementation blocked all writes in case of
limit reached until the TXG is committed and the log is freed. It
caused huge delays and following speed spikes in application writes.
This implementation instead smoothly throttles writes, using exactly
the same mechanism as used for dirty data.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: jxdking <lostking2008@hotmail.com>
Signed-off-by: Alexander Motin <mav@FreeBSD.org>
Sponsored-By: iXsystems, Inc.
Issue #12284Closes#13476
It turns out that "do LZ4 and zstd-1 both fail" is a great heuristic
for "don't even bother trying higher zstd tiers".
By way of illustration:
$ cat /incompress | mbuffer | zfs recv -o compression=zstd-12 evenfaster/lowcomp_1M_zstd12_normal
summary: 39.8 GiByte in 3min 40.2sec - average of 185 MiB/s
$ echo 3 | sudo tee /sys/module/zzstd/parameters/zstd_lz4_pass
3
$ cat /incompress | mbuffer -m 4G | zfs recv -o compression=zstd-12 evenfaster/lowcomp_1M_zstd12_patched
summary: 39.8 GiByte in 48.6sec - average of 839 MiB/s
$ sudo zfs list -p -o name,used,lused,ratio evenfaster/lowcomp_1M_zstd12_normal evenfaster/lowcomp_1M_zstd12_patched
NAME USED LUSED RATIO
evenfaster/lowcomp_1M_zstd12_normal 39549931520 42721221632 1.08
evenfaster/lowcomp_1M_zstd12_patched 39626399744 42721217536 1.07
$ python3 -c "print(39626399744 - 39549931520)"
76468224
$
I'll take 76 MB out of 42 GB for > 4x speedup.
Reviewed-by: Allan Jude <allan@klarasystems.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Kjeld Schouten <kjeld@schouten-lebbing.nl>
Reviewed-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Signed-off-by: Rich Ercolani <rincebrain@gmail.com>
Closes#13244
The short-path is now one access() call,
we always modprobe zfs (ZFS_MODULE_LOADING which doesn't use the libzfs
boolean parsing is gone),
and we use a simple inotify IN_CREATE loop with a timerfd timeout
rather than 10ms kernel-style polling
There's one substantial difference: ZFS_MODULE_TIMEOUT=-1
now means "never give up", rather than "wait 10 minutes"
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Closes#13330
There are times when end-users may wish to have
a fast and convenient method to get zpool guid
without having to use libzfs. This commit
exposes the zpool guid via kstats in similar
manner to the zpool state.
Reviewed-by: Alexander Motin <mav@FreeBSD.org>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Andrew Walker <awalker@ixsystems.com>
Closes#13466
Still descend, but only once: we get a lot of mileage out of nodist_
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Closes#13316
Holding a dbuf is a common operation which can become highly contended
in dbuf_find() when acquiring the dbuf hash mutex. This is particularly
true on Linux when reading/writing volumes since by default up to 32
threads from the zvol_taskq may be taking a hold of the same dbuf.
This should also be observable on FreeBSD as long as there are enough
processes accessing the volume concurrently.
This is further aggregrated by the fact that only the block id will
be unique when calculating the dbuf hash for a single volume. The
objset id, object id, and level will be the same for data blocks.
This has been observed to result in a somehwat less than uniform hash
distribution and a longer than expected max hash chain depth (~20)
on a large memory system (256 GB) using volumes.
This commit improves the siutation by switching the hash mutex to
an rwlock to allow concurrent lookups, and increasing DBUF_RWLOCKS
from 2048 to 8192 to further reduce the odds of a hash collision.
Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Alexander Motin <mav@FreeBSD.org>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#13405
Page writebacks with WB_SYNC_NONE can take several seconds to complete
since they wait for the transaction group to close before being
committed. This is usually not a problem since the caller does not
need to wait. However, if we're simultaneously doing a writeback
with WB_SYNC_ALL (e.g via msync), the latter can block for several
seconds (up to zfs_txg_timeout) due to the active WB_SYNC_NONE
writeback since it needs to wait for the transaction to complete
and the PG_writeback bit to be cleared.
This commit deals with 2 cases:
- No page writeback is active. A WB_SYNC_ALL page writeback starts
and even completes. But when it's about to check if the PG_writeback
bit has been cleared, another writeback with WB_SYNC_NONE starts.
The sync page writeback ends up waiting for the non-sync page
writeback to complete.
- A page writeback with WB_SYNC_NONE is already active when a
WB_SYNC_ALL writeback starts. The WB_SYNC_ALL writeback ends up
waiting for the WB_SYNC_NONE writeback.
The fix works by carefully keeping track of active sync/non-sync
writebacks and committing when beneficial.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Shaan Nobee <sniper111@gmail.com>
Closes#12662Closes#12790
Previous flushing algorithm limited only total number of log blocks to
the minimum of 256K and 4x number of metaslabs in the pool. As result,
system with 1500 disks with 1000 metaslabs each, touching several new
metaslabs each TXG could grow spacemap log to huge size without much
benefits. We've observed one of such systems importing pool for about
45 minutes.
This patch improves the situation from five sides:
- By limiting maximum period for each metaslab to be flushed to 1000
TXGs, that effectively limits maximum number of per-TXG spacemap logs
to load to the same number.
- By making flushing more smooth via accounting number of metaslabs
that were touched after the last flush and actually need another flush,
not just ms_unflushed_txg bump.
- By applying zfs_unflushed_log_block_pct to the number of metaslabs
that were touched after the last flush, not all metaslabs in the pool.
- By aggressively prefetching per-TXG spacemap logs up to 16 TXGs in
advance, making log spacemap load process for wide HDD pool CPU-bound,
accelerating it by many times.
- By reducing zfs_unflushed_log_block_max from 256K to 128K, reducing
single-threaded by nature log processing time from ~10 to ~5 minutes.
As further optimization we could skip bumping ms_unflushed_txg for
metaslabs not touched since the last flush, but that would be an
incompatible change, requiring new pool feature.
Reviewed-by: Matthew Ahrens <mahrens@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Alexander Motin <mav@FreeBSD.org>
Sponsored-By: iXsystems, Inc.
Closes#12789
Currently, determining which datasets are affected by corruption is
a manual process.
The primary difficulty in reporting the list of affected snapshots is
that since the error was initially found, the snapshot where the error
originally occurred in, may have been deleted. To solve this issue, we
add the ID of the head dataset of the original snapshot which the error
was detected in, to the stored error report. Then any time a filesystem
is deleted, the errors associated with it are deleted as well. Any time
a clone promote occurs, we modify reports associated with the original
head to refer to the new head. The stored error reports are identified
by this head ID, the birth time of the block which the error occurred
in, as well as some information about the error itself are also stored.
Once this information is stored, we can find the set of datasets
affected by an error by walking back the list of snapshots in the given
head until we find one with the appropriate birth txg, and then traverse
through the snapshots of the clone family, terminating a branch if the
block was replaced in a given snapshot. Then we report this information
back to libzfs, and to the zpool status command, where it is displayed
as follows:
pool: test
state: ONLINE
status: One or more devices has experienced an error resulting in data
corruption. Applications may be affected.
action: Restore the file in question if possible. Otherwise restore the
entire pool from backup.
see: https://openzfs.github.io/openzfs-docs/msg/ZFS-8000-8A
scan: scrub repaired 0B in 00:00:00 with 800 errors on Fri Dec 3
08:27:57 2021
config:
NAME STATE READ WRITE CKSUM
test ONLINE 0 0 0
sdb ONLINE 0 0 1.58K
errors: Permanent errors have been detected in the following files:
test@1:/test.0.0
/test/test.0.0
/test/1clone/test.0.0
A new feature flag is introduced to mark the presence of this change, as
well as promotion and backwards compatibility logic. This is an updated
version of #9175. Rebase required fixing the tests, updating the ABI of
libzfs, updating the man pages, fixing bugs, fixing the error returns,
and updating the old on-disk error logs to the new format when
activating the feature.
Reviewed-by: Matthew Ahrens <mahrens@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Mark Maybee <mark.maybee@delphix.com>
Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Co-authored-by: TulsiJain <tulsi.jain@delphix.com>
Signed-off-by: George Amanakis <gamanakis@gmail.com>
Closes#9175Closes#12812
Originally it was thought it would be useful to split up the kmods
by functionality. This would allow external consumers to only load
what was needed. However, in practice we've never had a case where
this functionality would be needed, and conversely managing multiple
kmods can be awkward. Therefore, this change merges all but the
spl.ko kmod in to a single zfs.ko kmod.
Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Closes#13274
Parts of the Linux kernel build system struggle with _Noreturn. This
results in the following warnings when building on RHEL 8.5, and likely
other environments. Switch to using the __attribute__((noreturn)).
warning: objtool: dbuf_free_range()+0x2b8:
return with modified stack frame
warning: objtool: dbuf_free_range()+0x0:
stack state mismatch: cfa1=7+40 cfa2=7+8
...
WARNING: EXPORT symbol "arc_buf_size" [zfs.ko] version generation
failed, symbol will not be versioned.
WARNING: EXPORT symbol "spa_open" [zfs.ko] version generation
failed, symbol will not be versioned.
...
Additionally, __thread_exit() has been renamed spl_thread_exit() and
made a static inline function. This was needed because the kernel
will generate a warning for symbols which are __attribute__((noreturn))
and then exported with EXPORT_SYMBOL.
While we could continue to use _Noreturn in user space I've also
switched it to __attribute__((noreturn)) purely for consistency
throughout the code base.
Reviewed-by: Ryan Moeller <freqlabs@FreeBSD.org>
Reviewed-by: Brian Atkinson <batkinson@lanl.gov>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes#13238
bcopy() has a confusing argument order and is actually a move, not a
copy; they're all deprecated since POSIX.1-2001 and removed in -2008,
and we shim them out to mem*() on Linux anyway
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Closes#12996
Tino Reichardt
Ameer Hamza
Chunwei Chen
Richard Yao
Tony Hutter
Alexander Motin
Christian Schwarz
Umer Saleem
Andriy Gapon
Brian Behlendorf
George Amanakis
Alek P
наб
Allan Jude
Will Andrews
Kevin Jin
Ryan Moeller
Rich Ercolani
Andrew
Shaan Nobee