We always call it twice with JUSTLOOKING and then FORREAL.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Pawel Jakub Dawidek <pawel@dawidek.net>
Closes#16225
- Reduce number of allocators on small system down to one per 4
CPU cores, keeping maximum at 4 on 16+ core systems. Small systems
should not have the lock contention multiple allocators supposed
to solve, while having several metaslabs open and modified each
TXG is not free.
- Reduce number of write issue taskqs down to one per 16 CPU
cores and an integer fraction of number of allocators. On mid-
sized systems, where multiple allocators already make sense, too
many write issue taskqs may reduce write speed on single-file
workloads, since single file is handled by only one taskq to
reduce fragmentation. On large systems, that can actually benefit
from many taskq's better IOPS, the bottleneck is less important,
since in worst case there will be at least 16 cores to handle it.
- Distribute dnodes between allocators (and taskqs) in a round-
robin fashion instead of relying on sync taskqs to be balanced.
The last is not guarantied and may depend on scheduling.
- Remove io_wr_iss_tq from struct zio. io_allocator is enough.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Alexander Motin <mav@FreeBSD.org>
Sponsored by: iXsystems, Inc.
Closes#16130
As I understand just for being less predictable dnode hash includes
8 bits of objset pointer, starting at 6. But since objset_t is
more than 1KB in size, its allocations are likely aligned to 2KB,
that means 11 lower bits provide no entropy. Just take the 8 bits
starting from 11.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Alexander Motin <mav@FreeBSD.org>
Sponsored by: iXsystems, Inc.
Closes#16131
Previous code held ARC state sublist lock throughout all L2ARC
write process, which included number of allocations and even ZIO
issues. Being blocked in any of those places the code could also
block ARC eviction, that could cause OOM activation or even dead-
lock if system is low on memory or one is too fragmented.
Fix it by dropping the lock as soon as we see a block eligible
for L2ARC writing and pick it up later using earlier inserted
marker. While there, also reduce scope of hash lock, moving
ZIO allocation and other operations not requiring header access
out of it. All operations requiring header access move under
hash lock, since L2_WRITING flag does not prevent header eviction
only transition to arc_l2c_only state with L1 header.
To be able to manipulate sublist lock and marker as needed add few
more multilist functions and modify one.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Alexander Motin <mav@FreeBSD.org>
Sponsored by: iXsystems, Inc.
Closes#16040
As part of transaction group commit, dsl_pool_sync() sequentially calls
dsl_dataset_sync() for each dirty dataset, which subsequently calls
dmu_objset_sync(). dmu_objset_sync() in turn uses up to 75% of CPU
cores to run sync_dnodes_task() in taskq threads to sync the dirty
dnodes (files).
There are two problems:
1. Each ZVOL in a pool is a separate dataset/objset having a single
dnode. This means the objsets are synchronized serially, which
leads to a bottleneck of ~330K blocks written per second per pool.
2. In the case of multiple dirty dnodes/files on a dataset/objset on a
big system they will be sync'd in parallel taskq threads. However,
it is inefficient to to use 75% of CPU cores of a big system to do
that, because of (a) bottlenecks on a single write issue taskq, and
(b) allocation throttling. In addition, if not for the allocation
throttling sorting write requests by bookmarks (logical address),
writes for different files may reach space allocators interleaved,
leading to unwanted fragmentation.
The solution to both problems is to always sync no more and (if
possible) no fewer dnodes at the same time than there are allocators
the pool.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Alexander Motin <mav@FreeBSD.org>
Signed-off-by: Edmund Nadolski <edmund.nadolski@ixsystems.com>
Closes#15197
ZFS prefetch is currently governed by the zfs_prefetch_disable
tunable. However, this is a module-wide settings - if a specific
dataset benefits from prefetch, while others have issue with it,
an optimal solution does not exists.
This commit introduce the "prefetch" tri-state property, which enable
granular control (at dataset/volume level) for prefetching.
This patch does not remove the zfs_prefetch_disable, which remains
a system-wide switch for enable/disable prefetch. However, to avoid
duplication, it would be preferable to deprecate and then remove
the module tunable.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Alexander Motin <mav@FreeBSD.org>
Reviewed-by: Ameer Hamza <ahamza@ixsystems.com>
Signed-off-by: Gionatan Danti <g.danti@assyoma.it>
Co-authored-by: Gionatan Danti <g.danti@assyoma.it>
Closes#15237Closes#15436
Those callbacks were introduced many years ago as part of a bigger
patch to smoothen the write throttling within a txg. They allow to
account completion of individual physical writes within a logical
one, improving cases when some of physical writes complete much
sooner than others, gradually opening the write throttle.
Few years after that ZFS got allocation throttling, working on a
level of logical writes and limiting number of writes queued to
vdevs at any point, and so limiting latency distribution between
the physical writes and especially writes of multiple copies.
The addition of scheduling deadline I proposed in #14925 should
further reduce the latency distribution. Grown memory sizes over
the past 10 years should also reduce importance of the smoothing.
While the use of physdone callback may still in theory provide
some smoother throttling, there are cases where we simply can not
afford it. Since dirty data accounting is protected by pool-wide
lock, in case of 6-wide RAIDZ, for example, it requires us to take
it 8 times per logical block write, creating huge lock contention.
My tests of this patch show radical reduction of the lock spinning
time on workloads when smaller blocks are written to RAIDZ pools,
when each of the disks receives 8-16KB chunks, but the total rate
reaching 100K+ blocks per second. Same time attempts to measure
any write time fluctuations didn't show anything noticeable.
While there, remove also io_child_count/io_parent_count counters.
They are used only for couple assertions that can be avoided.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Alexander Motin <mav@FreeBSD.org>
Sponsored by: iXsystems, Inc.
Closes#14948
... instead of list_head() + list_remove(). On FreeBSD the list
functions are not inlined, so in addition to more compact code
this also saves another function call.
Reviewed-by: Brian Atkinson <batkinson@lanl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Alexander Motin <mav@FreeBSD.org>
Sponsored by: iXsystems, Inc.
Closes#14955
ae7e700650 added an assertion to suppress
a complaint from Clang's static analyzer. Unfortunately, it missed
another way for Clang to complain about this function. This adds another
assertion to handle that.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Richard Yao <richard.yao@alumni.stonybrook.edu>
Closes#14575
by placing the most common use case (no special vdevs) first and avoid
allocating new variables.
Reviewed-by: Alexander Motin <mav@FreeBSD.org>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: George Amanakis <gamanakis@gmail.com>
Closes#14494Closes#14563
Clang's static analyzer claims that dereferencing ds in
dmu_objset_create_impl_dnstats() could cause a NULL pointer dereference
when a previous NULL check confirms that it is NULL. It is only NULL on
the MOS, for which dmu_objset_userused_enabled(os) should always return
false, so ds will never be dereferenced when it is NULL. We add an
assertion to suppress this warning.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Brian Atkinson <batkinson@lanl.gov>
Signed-off-by: Richard Yao <richard.yao@alumni.stonybrook.edu>
Closes#14470
Otherwise the dataset may be freed after the last dmu_buf_rele() leading
to a panic.
Reviewed-by: Mark Maybee <mark.maybee@delphix.com>
Reviewed-by: Matthew Ahrens <mahrens@delphix.com>
Signed-off-by: George Amanakis <gamanakis@gmail.com>
Closes#14522Closes#14523
With commit 34ce4c42f applied, there is no need for eee9362a7.
Revert that aside from the test. All tests introduced in those commits
pass.
Reviewed-by: Richard Yao <richard.yao@alumni.stonybrook.edu>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: George Amanakis <gamanakis@gmail.com>
Closes#14502
When activating filesystem features after receiving a snapshot, do
so only in syncing context.
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Reviewed-by: Richard Yao <richard.yao@alumni.stonybrook.edu>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: George Amanakis <gamanakis@gmail.com>
Closes#14304Closes#14252
Previously the primarycache property was handled only in the dbuf
layer. Since the speculative prefetcher is implemented in the ARC,
it had to be disabled for uncacheable buffers.
This change gives the ARC knowledge about uncacheable buffers
via arc_read() and arc_write(). So when remove_reference() drops
the last reference on the ARC header, it can either immediately destroy
it, or if it is marked as prefetch, put it into a new arc_uncached state.
That state is scanned every second, evicting stale buffers that were
not demand read.
This change also tracks dbufs that were read from the beginning,
but not to the end. It is assumed that such buffers may receive further
reads, and so they are stored in dbuf cache. If a following
reads reaches the end of the buffer, it is immediately evicted.
Otherwise it will follow regular dbuf cache eviction. Since the dbuf
layer does not know actual file sizes, this logic is not applied to
the final buffer of a dnode.
Since uncacheable buffers should no longer stay in the ARC for long,
this patch also tries to optimize I/O by allocating ARC physical
buffers as linear to allow buffer sharing.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: George Wilson <george.wilson@delphix.com>
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Signed-off-by: Alexander Motin <mav@FreeBSD.org>
Sponsored by: iXsystems, Inc.
Closes#14243
We ran out of space in enum zio_flag for additional flags. Rather than
introduce enum zio_flag2 and then modify a bunch of functions to take a
second flags variable, we expand the type to 64 bits via `typedef
uint64_t zio_flag_t`.
Reviewed-by: Allan Jude <allan@klarasystems.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Richard Yao <richard.yao@klarasystems.com>
Signed-off-by: Allan Jude <allan@klarasystems.com>
Co-authored-by: Richard Yao <richard.yao@klarasystems.com>
Closes#14086
Currently, additional/extra copies are created for metadata in
addition to the redundancy provided by the pool(mirror/raidz/draid),
due to this 2 times more space is utilized per inode and this decreases
the total number of inodes that can be created in the filesystem. By
setting redundant_metadata to none, no additional copies of metadata
are created, hence can reduce the space consumed by the additional
metadata copies and increase the total number of inodes that can be
created in the filesystem. Additionally, this can improve file create
performance due to the reduced amount of metadata which needs
to be written.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Dipak Ghosh <dipak.ghosh@hpe.com>
Signed-off-by: Akash B <akash-b@hpe.com>
Closes#13680
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
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
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
Evaluated every variable that lives in .data (and globals in .rodata)
in the kernel modules, and constified/eliminated/localised them
appropriately. This means that all read-only data is now actually
read-only data, and, if possible, at file scope. A lot of previously-
global-symbols became inlinable (and inlined!) constants. Probably
not in a big Wowee Performance Moment, but hey.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ahelenia Ziemiańska <nabijaczleweli@nabijaczleweli.xyz>
Closes#12899
Special allocation class or dedup vdevs may have roughly the same
performance as L2ARC vdevs. Introduce a new tunable to exclude those
buffers from being cacheable on L2ARC.
Reviewed-by: Don Brady <don.brady@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: George Amanakis <gamanakis@gmail.com>
Closes#11761Closes#12285
The number of sublists in a multilist is relatively small. We dont need
64 bits to calculate an index. 32 bits is sufficient and makes the
code more efficient.
Reviewed-by: Matthew Ahrens <mahrens@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Mark Maybee <mark.maybee@delphix.com>
Signed-off-by: Alexander Motin <mav@FreeBSD.org>
Sponsored-By: iXsystems, Inc.
Closes#12288
ZFS loves using %llu for uint64_t, but that requires a cast to not
be noisy - which is even done in many, though not all, places.
Also a couple places used %u for uint64_t, which were promoted
to %llu.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Rich Ercolani <rincebrain@gmail.com>
Closes#12233
This commit partially reverts changes to multilists in PR 7968
(multi-threaded spa-sync()) and adds some cache line alignments to
separate read-only multilists and heavily modified refcount's to different
cache lines.
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#12158
Follow up for commit 624222a, value asserted <= SPA_OLD_MAXBLOCKSIZE
instead of SPA_MAXBLOCKSIZE as it should be after the previous change.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Alexander Motin <mav@FreeBSD.org>
Closes#11501
Build error on illumos with gcc 10 did reveal:
In function 'dmu_objset_refresh_ownership':
../../common/fs/zfs/dmu_objset.c:857:25: error: implicit conversion
from 'boolean_t' to 'ds_hold_flags_t' {aka 'enum ds_hold_flags'}
[-Werror=enum-conversion]
857 | dsl_dataset_disown(ds, decrypt, tag);
| ^~~~~~~
cc1: all warnings being treated as errors
libzfs_input_check.c: In function 'zfs_ioc_input_tests':
libzfs_input_check.c:754:28: error: implicit conversion from
'enum dmu_objset_type' to 'enum lzc_dataset_type'
[-Werror=enum-conversion]
754 | err = lzc_create(dataset, DMU_OST_ZFS, NULL, NULL, 0);
| ^~~~~~~~~~~
cc1: all warnings being treated as errors
The same issue is present in openzfs, and also the same issue about
ds_hold_flags_t, which currently defines exactly one valid value.
Reviewed-by: Igor Kozhukhov <igor@dilos.org>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Toomas Soome <tsoome@me.com>
Closes#11406
After porting the fix for https://github.com/openzfs/zfs/issues/5295
over to illumos, we started hitting an assertion failure when running
the testsuite:
assertion failed: rc->rc_count == number, file: .../refcount.c
and the unexpected hold has this stack:
dsl_dataset_long_hold+0x59 dmu_objset_upgrade+0x73
dmu_objset_id_quota_upgrade+0x15 dmu_objset_own+0x14f
The simplest reproducer for this in illumos is
zpool create -f -O version=1 testpool c3t0d0; zpool destroy testpool
which is run as part of the zpool_create_tempname test, but I can't get
this to trigger on FreeBSD. This appears to be because of the call to
txg_wait_synced() in dmu_objset_upgrade_stop() (which was missing in
illumos), slows down dmu_objset_disown() enough to avoid the condition.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Andy Fiddaman <andy@omnios.org>
Closes#11368
The performance of `zfs receive` can be bottlenecked on the CPU consumed
by the `receive_writer` thread, especially when receiving streams with
small compressed block sizes. Much of the CPU is spent creating and
destroying dbuf's and arc buf's, one for each `WRITE` record in the send
stream.
This commit introduces the concept of "lightweight writes", which allows
`zfs receive` to write to the DMU by providing an ABD, and instantiating
only a new type of `dbuf_dirty_record_t`. The dbuf and arc buf for this
"dirty leaf block" are not instantiated.
Because there is no dbuf with the dirty data, this mechanism doesn't
support reading from "lightweight-dirty" blocks (they would see the
on-disk state rather than the dirty data). Since the dedup-receive code
has been removed, `zfs receive` is write-only, so this works fine.
Because there are no arc bufs for the received data, the received data
is no longer cached in the ARC.
Testing a receive of a stream with average compressed block size of 4KB,
this commit improves performance by 50%, while also reducing CPU usage
by 50% of a CPU. On a per-block basis, CPU consumed by receive_writer()
and dbuf_evict() is now 1/7th (14%) of what it was.
Baseline: 450MB/s, CPU in receive_writer() 40% + dbuf_evict() 35%
New: 670MB/s, CPU in receive_writer() 17% + dbuf_evict() 0%
The code is also restructured in a few ways:
Added a `dr_dnode` field to the dbuf_dirty_record_t. This simplifies
some existing code that no longer needs `DB_DNODE_ENTER()` and related
routines. The new field is needed by the lightweight-type dirty record.
To ensure that the `dr_dnode` field remains valid until the dirty record
is freed, we have to ensure that the `dnode_move()` doesn't relocate the
dnode_t. To do this we keep a hold on the dnode until it's zio's have
completed. This is already done by the user-accounting code
(`userquota_updates_task()`), this commit extends that so that it always
keeps the dnode hold until zio completion (see `dnode_rele_task()`).
`dn_dirty_txg` was previously zeroed when the dnode was synced. This
was not necessary, since its meaning can be "when was this dnode last
dirtied". This change simplifies the new `dnode_rele_task()` code.
Removed some dead code related to `DRR_WRITE_BYREF` (dedup receive).
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Paul Dagnelie <pcd@delphix.com>
Reviewed-by: George Wilson <gwilson@delphix.com>
Signed-off-by: Matthew Ahrens <mahrens@delphix.com>
Closes#11105
For encrypted receives, where user accounting is initially disabled on
creation, both 'zfs userspace' and 'zfs groupspace' fails with
EOPNOTSUPP: this is because dmu_objset_id_quota_upgrade_cb() forgets to
set OBJSET_FLAG_USERACCOUNTING_COMPLETE on the objset flags after a
successful dmu_objset_space_upgrade().
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Co-authored-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: loli10K <ezomori.nozomu@gmail.com>
Closes#9501Closes#9596
In C, const indicates to the reader that mutation will not occur.
It can also serve as a hint about ownership.
Add const in a few places where it makes sense.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Ryan Moeller <freqlabs@FreeBSD.org>
Closes#10997
This PR adds two new compression types, based on ZStandard:
- zstd: A basic ZStandard compression algorithm Available compression.
Levels for zstd are zstd-1 through zstd-19, where the compression
increases with every level, but speed decreases.
- zstd-fast: A faster version of the ZStandard compression algorithm
zstd-fast is basically a "negative" level of zstd. The compression
decreases with every level, but speed increases.
Available compression levels for zstd-fast:
- zstd-fast-1 through zstd-fast-10
- zstd-fast-20 through zstd-fast-100 (in increments of 10)
- zstd-fast-500 and zstd-fast-1000
For more information check the man page.
Implementation details:
Rather than treat each level of zstd as a different algorithm (as was
done historically with gzip), the block pointer `enum zio_compress`
value is simply zstd for all levels, including zstd-fast, since they all
use the same decompression function.
The compress= property (a 64bit unsigned integer) uses the lower 7 bits
to store the compression algorithm (matching the number of bits used in
a block pointer, as the 8th bit was borrowed for embedded block
pointers). The upper bits are used to store the compression level.
It is necessary to be able to determine what compression level was used
when later reading a block back, so the concept used in LZ4, where the
first 32bits of the on-disk value are the size of the compressed data
(since the allocation is rounded up to the nearest ashift), was
extended, and we store the version of ZSTD and the level as well as the
compressed size. This value is returned when decompressing a block, so
that if the block needs to be recompressed (L2ARC, nop-write, etc), that
the same parameters will be used to result in the matching checksum.
All of the internal ZFS code ( `arc_buf_hdr_t`, `objset_t`,
`zio_prop_t`, etc.) uses the separated _compress and _complevel
variables. Only the properties ZAP contains the combined/bit-shifted
value. The combined value is split when the compression_changed_cb()
callback is called, and sets both objset members (os_compress and
os_complevel).
The userspace tools all use the combined/bit-shifted value.
Additional notes:
zdb can now also decode the ZSTD compression header (flag -Z) and
inspect the size, version and compression level saved in that header.
For each record, if it is ZSTD compressed, the parameters of the decoded
compression header get printed.
ZSTD is included with all current tests and new tests are added
as-needed.
Per-dataset feature flags now get activated when the property is set.
If a compression algorithm requires a feature flag, zfs activates the
feature when the property is set, rather than waiting for the first
block to be born. This is currently only used by zstd but can be
extended as needed.
Portions-Sponsored-By: The FreeBSD Foundation
Co-authored-by: Allan Jude <allanjude@freebsd.org>
Co-authored-by: Brian Behlendorf <behlendorf1@llnl.gov>
Co-authored-by: Sebastian Gottschall <s.gottschall@dd-wrt.com>
Co-authored-by: Kjeld Schouten-Lebbing <kjeld@schouten-lebbing.nl>
Co-authored-by: Michael Niewöhner <foss@mniewoehner.de>
Signed-off-by: Allan Jude <allan@klarasystems.com>
Signed-off-by: Allan Jude <allanjude@freebsd.org>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Sebastian Gottschall <s.gottschall@dd-wrt.com>
Signed-off-by: Kjeld Schouten-Lebbing <kjeld@schouten-lebbing.nl>
Signed-off-by: Michael Niewöhner <foss@mniewoehner.de>
Closes#6247Closes#9024Closes#10277Closes#10278
Reviewed-by: Ryan Moeller <ryan@ixsystems.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Allan Jude <allan@klarasystems.com>
Closes#10636
The filesystem_limit and snapshot_limit properties limit the number of
filesystems or snapshots that can be created below this dataset.
According to the manpage, "The limit is not enforced if the user is
allowed to change the limit." Two types of users are allowed to change
the limit:
1. Those that have been delegated the `filesystem_limit` or
`snapshot_limit` permission, e.g. with
`zfs allow USER filesystem_limit DATASET`. This works properly.
2. A user with elevated system privileges (e.g. root). This does not
work - the root user will incorrectly get an error when trying to create
a snapshot/filesystem, if it exceeds the `_limit` property.
The problem is that `priv_policy_ns()` does not work if the `cred_t` is
not that of the current process. This happens when
`dsl_enforce_ds_ss_limits()` is called in syncing context (as part of a
sync task's check func) to determine the permissions of the
corresponding user process.
This commit fixes the issue by passing the `task_struct` (typedef'ed as
a `proc_t`) to syncing context, and then using `has_capability()` to
determine if that process is privileged. Note that we still need to
pass the `cred_t` to syncing context so that we can check if the user
was delegated this permission with `zfs allow`.
This problem only impacts Linux. Wrappers are added to FreeBSD but it
continues to use `priv_check_cred()`, which works on arbitrary `cred_t`.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Ryan Moeller <ryan@ixsystems.com>
Signed-off-by: Matthew Ahrens <mahrens@delphix.com>
Closes#8226Closes#10545
Mark functions used only in the same translation unit as static. This
only includes functions that do not have a prototype in a header file
either.
Reviewed-by: Ryan Moeller <ryan@iXsystems.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Arvind Sankar <nivedita@alum.mit.edu>
Closes#10470
Background:
By increasing the recordsize property above the default of 128KB, a
filesystem may have "large" blocks. By default, a send stream of such a
filesystem does not contain large WRITE records, instead it decreases
objects' block sizes to 128KB and splits the large blocks into 128KB
blocks, allowing the large-block filesystem to be received by a system
that does not support the `large_blocks` feature. A send stream
generated by `zfs send -L` (or `--large-block`) preserves the large
block size on the receiving system, by using large WRITE records.
When receiving an incremental send stream for a filesystem with large
blocks, if the send stream's -L flag was toggled, a bug is encountered
in which the file's contents are incorrectly zeroed out. The contents
of any blocks that were not modified by this send stream will be lost.
"Toggled" means that the previous send used `-L`, but this incremental
does not use `-L` (-L to no-L); or that the previous send did not use
`-L`, but this incremental does use `-L` (no-L to -L).
Changes:
This commit addresses the problem with several changes to the semantics
of zfs send/receive:
1. "-L to no-L" incrementals are rejected. If the previous send used
`-L`, but this incremental does not use `-L`, the `zfs receive` will
fail with this error message:
incremental send stream requires -L (--large-block), to match
previous receive.
2. "no-L to -L" incrementals are handled correctly, preserving the
smaller (128KB) block size of any already-received files that used large
blocks on the sending system but were split by `zfs send` without the
`-L` flag.
3. A new send stream format flag is added, `SWITCH_TO_LARGE_BLOCKS`.
This feature indicates that we can correctly handle "no-L to -L"
incrementals. This flag is currently not set on any send streams. In
the future, we intend for incremental send streams of snapshots that
have large blocks to use `-L` by default, and these streams will also
have the `SWITCH_TO_LARGE_BLOCKS` feature set. This ensures that streams
from the default use of `zfs send` won't encounter the bug mentioned
above, because they can't be received by software with the bug.
Implementation notes:
To facilitate accessing the ZPL's generation number,
`zfs_space_delta_cb()` has been renamed to `zpl_get_file_info()` and
restructured to fill in a struct with ZPL-specific info including owner
and generation.
In the "no-L to -L" case, if this is a compressed send stream (from
`zfs send -cL`), large WRITE records that are being written to small
(128KB) blocksize files need to be decompressed so that they can be
written split up into multiple blocks. The zio pipeline will recompress
each smaller block individually.
A new test case, `send-L_toggle`, is added, which tests the "no-L to -L"
case and verifies that we get an error for the "-L to no-L" case.
Reviewed-by: Paul Dagnelie <pcd@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Matthew Ahrens <mahrens@delphix.com>
Closes#6224Closes#10383
The strcpy() and sprintf() functions are deprecated on some platforms.
Care is needed to ensure correct size is used. If some platforms
miss snprintf, we can add a #define to sprintf, likewise strlcpy().
The biggest change is adding a size parameter to zfs_id_to_fuidstr().
The various *_impl_get() functions are only used on linux and have
not yet been updated.
Reviewed by: Sean Eric Fagan <sef@ixsystems.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Jorgen Lundman <lundman@lundman.net>
Closes#10400
The normal lock order is that the dp_config_rwlock must be held before
the ds_opening_lock. For example, dmu_objset_hold() does this.
However, dmu_objset_open_impl() is called with the ds_opening_lock held,
and if the dp_config_rwlock is not already held, it will attempt to
acquire it. This may lead to deadlock, since the lock order is
reversed.
Looking at all the callers of dmu_objset_open_impl() (which is
principally the callers of dmu_objset_from_ds()), almost all callers
already have the dp_config_rwlock. However, there are a few places in
the send and receive code paths that do not. For example:
dsl_crypto_populate_key_nvlist, send_cb, dmu_recv_stream,
receive_write_byref, redact_traverse_thread.
This commit resolves the problem by requiring all callers ot
dmu_objset_from_ds() to hold the dp_config_rwlock. In most cases, the
code has been restructured such that we call dmu_objset_from_ds()
earlier on in the send and receive processes, when we already have the
dp_config_rwlock, and save the objset_t until we need it in the middle
of the send or receive (similar to what we already do with the
dsl_dataset_t). Thus we do not need to acquire the dp_config_rwlock in
many new places.
I also cleaned up code in dmu_redact_snap() and send_traverse_thread().
Reviewed-by: Paul Dagnelie <pcd@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Paul Zuchowski <pzuchowski@datto.com>
Signed-off-by: Matthew Ahrens <mahrens@delphix.com>
Closes#9662Closes#10115
When "zfs destroy" is run, it completes quickly, and in the background
we locate the blocks to free and free them. This background activity
can be observed with `zpool get freeing` and `zpool wait -t free ...`.
This background activity is processed by a single thread (the spa_sync
thread) which calls zio_free() on each of the blocks to free. With even
modest storage performance, the CPU consumption of zio_free() can be the
performance bottleneck.
Performance of zio_free() can be improved by not actually creating a
zio_t in the common case (non-dedup, non-gang), instead calling
metaslab_free() directly. This avoids the CPU cost of allocating the
zio_t, and more importantly the cost of adding and later removing this
zio_t from the parent zio's child list.
The result is that performance of background freeing more than doubles,
from 0.6 million blocks per second to 1.3 million blocks per second.
Reviewed-by: Paul Dagnelie <pcd@delphix.com>
Reviewed-by: Serapheim Dimitropoulos <serapheim@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: George Wilson <gwilson@delphix.com>
Signed-off-by: Matthew Ahrens <mahrens@delphix.com>
Closes#10034
Additionally pull in state machine comments about
upcoming async cow work.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Matt Ahrens <matt@delphix.com>
Signed-off-by: Matt Macy <mmacy@FreeBSD.org>
Closes#9902
When we finish a zfs receive, dmu_recv_end_sync() calls
zvol_create_minors(async=TRUE). This kicks off some other threads that
create the minor device nodes (in /dev/zvol/poolname/...). These async
threads call zvol_prefetch_minors_impl() and zvol_create_minor(), which
both call dmu_objset_own(), which puts a "long hold" on the dataset.
Since the zvol minor node creation is asynchronous, this can happen
after the `ZFS_IOC_RECV[_NEW]` ioctl and `zfs receive` process have
completed.
After the first receive ioctl has completed, userland may attempt to do
another receive into the same dataset (e.g. the next incremental
stream). This second receive and the asynchronous minor node creation
can interfere with one another in several different ways, because they
both require exclusive access to the dataset:
1. When the second receive is finishing up, dmu_recv_end_check() does
dsl_dataset_handoff_check(), which can fail with EBUSY if the async
minor node creation already has a "long hold" on this dataset. This
causes the 2nd receive to fail.
2. The async udev rule can fail if zvol_id and/or systemd-udevd try to
open the device while the the second receive's async attempt at minor
node creation owns the dataset (via zvol_prefetch_minors_impl). This
causes the minor node (/dev/zd*) to exist, but the udev-generated
/dev/zvol/... to not exist.
3. The async minor node creation can silently fail with EBUSY if the
first receive's zvol_create_minor() trys to own the dataset while the
second receive's zvol_prefetch_minors_impl already owns the dataset.
To address these problems, this change synchronously creates the minor
node. To avoid the lock ordering problems that the asynchrony was
introduced to fix (see #3681), we create the minor nodes from open
context, with no locks held, rather than from syncing contex as was
originally done.
Implementation notes:
We generally do not need to traverse children or prefetch anything (e.g.
when running the recv, snapshot, create, or clone subcommands of zfs).
We only need recursion when importing/opening a pool and when loading
encryption keys. The existing recursive, asynchronous, prefetching code
is preserved for use in these cases.
Channel programs may need to create zvol minor nodes, when creating a
snapshot of a zvol with the snapdev property set. We figure out what
snapshots are created when running the LUA program in syncing context.
In this case we need to remember what snapshots were created, and then
try to create their minor nodes from open context, after the LUA code
has completed.
There are additional zvol use cases that asynchronously own the dataset,
which can cause similar problems. E.g. changing the volmode or snapdev
properties. These are less problematic because they are not recursive
and don't touch datasets that are not involved in the operation, there
is still potential for interference with subsequent operations. In the
future, these cases should be similarly converted to create the zvol
minor node synchronously from open context.
The async tasks of removing and renaming minors do not own the objset,
so they do not have this problem. However, it may make sense to also
convert these operations to happen synchronously from open context, in
the future.
Reviewed-by: Paul Dagnelie <pcd@delphix.com>
Reviewed-by: Prakash Surya <prakash.surya@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Matthew Ahrens <mahrens@delphix.com>
External-issue: DLPX-65948
Closes#7863Closes#9885
In the FreeBSD kernel the strdup signature is:
```
char *strdup(const char *__restrict, struct malloc_type *);
```
It's unfortunate that the developers have chosen to change
the signature of libc functions - but it's what I have to
deal with.
Reviewed-by: Jorgen Lundman <lundman@lundman.net>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Matt Macy <mmacy@FreeBSD.org>
Closes#9433
This patch implements a new tree structure for ZFS, and uses it to
store range trees more efficiently.
The new structure is approximately a B-tree, though there are some
small differences from the usual characterizations. The tree has core
nodes and leaf nodes; each contain data elements, which the elements
in the core nodes acting as separators between its children. The
difference between core and leaf nodes is that the core nodes have an
array of children, while leaf nodes don't. Every node in the tree may
be only partially full; in most cases, they are all at least 50% full
(in terms of element count) except for the root node, which can be
less full. Underfull nodes will steal from their neighbors or merge to
remain full enough, while overfull nodes will split in two. The data
elements are contained in tree-controlled buffers; they are copied
into these on insertion, and overwritten on deletion. This means that
the elements are not independently allocated, which reduces overhead,
but also means they can't be shared between trees (and also that
pointers to them are only valid until a side-effectful tree operation
occurs). The overhead varies based on how dense the tree is, but is
usually on the order of about 50% of the element size; the per-node
overheads are very small, and so don't make a significant difference.
The trees can accept arbitrary records; they accept a size and a
comparator to allow them to be used for a variety of purposes.
The new trees replace the AVL trees used in the range trees today.
Currently, the range_seg_t structure contains three 8 byte integers
of payload and two 24 byte avl_tree_node_ts to handle its storage in
both an offset-sorted tree and a size-sorted tree (total size: 64
bytes). In the new model, the range seg structures are usually two 4
byte integers, but a separate one needs to exist for the size-sorted
and offset-sorted tree. Between the raw size, the 50% overhead, and
the double storage, the new btrees are expected to use 8*1.5*2 = 24
bytes per record, or 33.3% as much memory as the AVL trees (this is
for the purposes of storing metaslab range trees; for other purposes,
like scrubs, they use ~50% as much memory).
We reduced the size of the payload in the range segments by teaching
range trees about starting offsets and shifts; since metaslabs have a
fixed starting offset, and they all operate in terms of disk sectors,
we can store the ranges using 4-byte integers as long as the size of
the metaslab divided by the sector size is less than 2^32. For 512-byte
sectors, this is a 2^41 (or 2TB) metaslab, which with the default
settings corresponds to a 256PB disk. 4k sector disks can handle
metaslabs up to 2^46 bytes, or 2^63 byte disks. Since we do not
anticipate disks of this size in the near future, there should be
almost no cases where metaslabs need 64-byte integers to store their
ranges. We do still have the capability to store 64-byte integer ranges
to account for cases where we are storing per-vdev (or per-dnode) trees,
which could reasonably go above the limits discussed. We also do not
store fill information in the compact version of the node, since it
is only used for sorted scrub.
We also optimized the metaslab loading process in various other ways
to offset some inefficiencies in the btree model. While individual
operations (find, insert, remove_from) are faster for the btree than
they are for the avl tree, remove usually requires a find operation,
while in the AVL tree model the element itself suffices. Some clever
changes actually caused an overall speedup in metaslab loading; we use
approximately 40% less cpu to load metaslabs in our tests on Illumos.
Another memory and performance optimization was achieved by changing
what is stored in the size-sorted trees. When a disk is heavily
fragmented, the df algorithm used by default in ZFS will almost always
find a number of small regions in its initial cursor-based search; it
will usually only fall back to the size-sorted tree to find larger
regions. If we increase the size of the cursor-based search slightly,
and don't store segments that are smaller than a tunable size floor
in the size-sorted tree, we can further cut memory usage down to
below 20% of what the AVL trees store. This also results in further
reductions in CPU time spent loading metaslabs.
The 16KiB size floor was chosen because it results in substantial memory
usage reduction while not usually resulting in situations where we can't
find an appropriate chunk with the cursor and are forced to use an
oversized chunk from the size-sorted tree. In addition, even if we do
have to use an oversized chunk from the size-sorted tree, the chunk
would be too small to use for ZIL allocations, so it isn't as big of a
loss as it might otherwise be. And often, more small allocations will
follow the initial one, and the cursor search will now find the
remainder of the chunk we didn't use all of and use it for subsequent
allocations. Practical testing has shown little or no change in
fragmentation as a result of this change.
If the size-sorted tree becomes empty while the offset sorted one still
has entries, it will load all the entries from the offset sorted tree
and disregard the size floor until it is unloaded again. This operation
occurs rarely with the default setting, only on incredibly thoroughly
fragmented pools.
There are some other small changes to zdb to teach it to handle btrees,
but nothing major.
Reviewed-by: George Wilson <gwilson@delphix.com>
Reviewed-by: Matt Ahrens <matt@delphix.com>
Reviewed by: Sebastien Roy seb@delphix.com
Reviewed-by: Igor Kozhukhov <igor@dilos.org>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Paul Dagnelie <pcd@delphix.com>
Closes#9181
Refactor the zvol in to platform dependent and independent bits.
Reviewed-by: Allan Jude <allanjude@freebsd.org>
Reviewed-by: Jorgen Lundman <lundman@lundman.net>
Reviewed-by: Igor Kozhukhov <igor@dilos.org>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Matt Macy <mmacy@FreeBSD.org>
Closes#9295
Reviewed-by: Matt Ahrens <matt@delphix.com>
Reviewed-by: Ryan Moeller <ryan@ixsystems.com>
Reviewed-by: Richard Laager <rlaager@wiktel.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Andrea Gelmini <andrea.gelmini@gelma.net>
Closes#9240
Even though the bug's writeup (Github issue #9136) is very detailed,
we still don't know exactly how we got to that state, thus I wasn't
able to reproduce the bug. That said, we can make an educated guess
combining the information on filled issue with the code.
From the fact that `dp_dirty_total` was 0 (which is less than
`zfs_dirty_data_max`) we know that there was one thread that set it to
0 and then signaled one of the waiters of `dp_spaceavail_cv` [see
`dsl_pool_dirty_delta()` which is also the only place that
`dp_dirty_total` is changed]. Thus, the only logical explaination
then for the bug being hit is that the waiter that just got awaken
didn't go through `dsl_pool_dirty_data()`. Given that this function
is only called by `dsl_pool_dirty_space()` or `dsl_pool_undirty_space()`
I can only think of two possible ways of the above scenario happening:
[1] The waiter didn't call into any of the two functions - which I
find highly unlikely (i.e. why wait on `dp_spaceavail_cv` to begin
with?).
[2] The waiter did call in one of the above function but it passed 0 as
the space/delta to be dirtied (or undirtied) and then the callee
returned immediately (e.g both `dsl_pool_dirty_space()` and
`dsl_pool_undirty_space()` return immediately when space is 0).
In any case and no matter how we got there, the easy fix would be to
just broadcast to all waiters whenever `dp_dirty_total` hits 0. That
said and given that we've never hit this before, it would make sense
to think more on why the above situation occured.
Attempting to mimic what Prakash was doing in the issue filed, I
created a dataset with `sync=always` and started doing contiguous
writes in a file within that dataset. I observed with DTrace that even
though we update the pool's dirty data accounting when we would dirty
stuff, the accounting wouldn't be decremented incrementally as we were
done with the ZIOs of those writes (the reason being that
`dbuf_write_physdone()` isn't be called as we go through the override
code paths, and thus `dsl_pool_undirty_space()` is never called). As a
result we'd have to wait until we get to `dsl_pool_sync()` where we
zero out all dirty data accounting for the pool and the current TXG's
metadata.
In addition, as Matt noted and I later verified, the same issue would
arise when using dedup.
In both cases (sync & dedup) we shouldn't have to wait until
`dsl_pool_sync()` zeros out the accounting data. According to the
comment in that part of the code, the reasons why we do the zeroing,
have nothing to do with what we observe:
````
/*
* We have written all of the accounted dirty data, so our
* dp_space_towrite should now be zero. However, some seldom-used
* code paths do not adhere to this (e.g. dbuf_undirty(), also
* rounding error in dbuf_write_physdone).
* Shore up the accounting of any dirtied space now.
*/
dsl_pool_undirty_space(dp, dp->dp_dirty_pertxg[txg & TXG_MASK], txg);
````
Ideally what we want to do is to undirty in the accounting exactly what
we dirty (I use the word ideally as we can still have rounding errors).
This would make the behavior of the system more clear and predictable.
Another interesting issue that I observed with DTrace was that we
wouldn't update any of the pool's dirty data accounting whenever we
would dirty and/or undirty MOS data. In addition, every time we would
change the size of a dbuf through `dbuf_new_size()` we wouldn't update
the accounted space dirtied in the appropriate dirty record, so when
ZIOs are done we would undirty less that we dirtied from the pool's
accounting point of view.
For the first two issues observed (sync & dedup) this patch ensures
that we still update the pool's accounting when we undirty data,
regardless of the write being physical or not.
For changes in the MOS, we first ensure to zero out the pool's dirty
data accounting in `dsl_pool_sync()` after we synced the MOS. Then we
can go ahead and enable the update of the pool's dirty data accounting
wheneve we change MOS data.
Another fix is that we now update the accounting explicitly for
counting errors in `dbuf_write_done()`.
Finally, `dbuf_new_size()` updates the accounted space of the
appropriate dirty record correctly now.
The problem is that we still don't know how the bug came up in the
issue filled. That said the issues fixed seem to be very relevant, so
instead of going with the broadcasting solution right away,
I decided to leave this patch as is.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Prakash Surya <prakash.surya@delphix.com>
Signed-off-by: Serapheim Dimitropoulos <serapheim@delphix.com>
External-issue: DLPX-47285
Closes#9137
