OpenZFS decided that ignore_hole_birth was too imprecise and
incorrect a name (and went with send_holes_without_birth_time).
Rename it in ZoL too, while keeping the name "ignore_hole_birth"
pointing to the same variable for existing consumers.
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Rich Ercolani <rincebrain@gmail.com>
Closes #5239
Håkan Johansson [Fri, 7 Oct 2016 20:25:35 +0000 (22:25 +0200)]
Fix vdev_open_child() race on updating vdev_parent->vdev_nonrot
Updating vd->vdev_parent->vdev_nonrot in vdev_open_child()
is a race when vdev_open_child is called for many children
from a task queue.
vdev_open_child() is only called by vdev_open_children(), let
the latter update the parent vdev_nonrot member. The update
was already there, so done twice previously. Thus using the
same logic at the end in vdev_open_children() to update
vdev_nonrot, either we are vdev_uses_zvols() or not.
Reviewed-by: Richard Elling <Richard.Elling@RichardElling.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Haakan T Johansson <f96hajo@chalmers.se>
Closes #5162
Introduce a make recipe for flake8 to enable python
style checking. Ensure all python scripts pass flake8.
Return an error code of 0 for arcstat.py -v and
dbufstat.py -v. Add test cases for python scripts.
Reviewed by: Richard Laager <rlaager@wiktel.com> Reviewed-by: Richard Elling <Richard.Elling@RichardElling.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed-by: Ian Lee <IanLee1521@gmail.com>
Closes #5230
OpenZFS 6988 spa_sync() spends half its time in dmu_objset_do_userquota_updates
Using a benchmark which creates 2 million files in one TXG, I observe
that the thread running spa_sync() is on CPU almost the entire time we
are syncing, and therefore can be a performance bottleneck. About 50% of
the time in spa_sync() is in dmu_objset_do_userquota_updates().
The problem is that dmu_objset_do_userquota_updates() calls
zap_increment_int(DMU_USERUSED_OBJECT) once for every file that was
modified (or created). In this benchmark, all the files are owned by the
same user/group, so all 2 million calls to zap_increment_int() are
modifying the same entry in the zap. The same issue exists for the
DMU_GROUPUSED_OBJECT.
We should keep an in-memory map from user to space delta while we are
syncing, and when we finish, iterate over the in-memory map and modify
the ZAP once per entry. This reduces the number of calls to
zap_increment_int() from "number of objects modified" to "number of
owners/groups of modified files".
This reduced the time spent in spa_sync() in the file create benchmark
by ~33%, from 11 seconds to 7 seconds.
Upstream bugs: DLPX-44799
Ported by: Ned Bass <bass6@llnl.gov>
Porting notes:
- Added curly braces around declaration of userquota_cache_t cache to
quiet compiler warning;
- Handled the userobj accounting the same way it proposed in this path.
Jinshan Xiong [Tue, 4 Oct 2016 18:46:10 +0000 (11:46 -0700)]
Add support for user/group dnode accounting & quota
This patch tracks dnode usage for each user/group in the
DMU_USER/GROUPUSED_OBJECT ZAPs. ZAP entries dedicated to dnode
accounting have the key prefixed with "obj-" followed by the UID/GID
in string format (as done for the block accounting).
A new SPA feature has been added for dnode accounting as well as
a new ZPL version. The SPA feature must be enabled in the pool
before upgrading the zfs filesystem. During the zfs version upgrade,
a "quotacheck" will be executed by marking all dnode as dirty.
Implement tests to ensure that python scripts
that are distributed with ZFS continue to at
minimum run without errors. This will help prevent
accidental breaking of these scripts.
Signed-off-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Rudd-O [Thu, 6 Oct 2016 17:26:47 +0000 (17:26 +0000)]
Use a different technique to detect whether to mount-zfs
The behavior of the Dracut module was very wrong before.
The correct behavior: initramfs should not run `zfs-mount` to completion
if the two generator files exist. If, however, one of them is missing,
it indicates one of three cases:
* The kernel command line did not specify a root ZFS file system, and
another Dracut module is already handling root mount (via systemd).
`mount-zfs` can run, but it will do nothing.
* There is no systemd to run `sysroot.mount` to begin with.
`mount-zfs` must run.
* The root parameter is zfs:AUTO, which cannot be run in sysroot.mount.
`mount-zfs` must run.
In any of these three cases, it is safe to run `zfs-mount` to completion.
`zfs-mount` must also delete itself if it determines it should not run,
or else Dracut will do the insane thing of running it over and over again.
Literally, the definition of insanity, doing the same thing that did not
work before, expecting different results. Doing that may have had a great
result before, when we had a race between devices appearing and pools
being mounted, and `mount-zfs` was tasked with the full responsibility
of importing the needed pool, but nowadays it is wrong behavior and
should be suppressed.
I deduced that self-deletion was the correct thing to do by looking at
other Dracut code, because (as we all are very fully aware of) Dracut
is entirely, ahem, "implementation-defined".
Tested-by: @wphilips Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Manuel Amador (Rudd-O) <rudd-o@rudd-o.com>
Closes #5157
Closes #5204
lorddoskias [Wed, 5 Oct 2016 21:47:29 +0000 (00:47 +0300)]
Refactor updating of immutable/appendonly flags
Move the synchronization of inode/znode i_flgas/pflags into
the respective internal zfs function. This is mostly
mechanical work and shouldn't introduce any functional
changes.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Nikolay Borisov <n.borisov.lkml@gmail.com>
Issue #227
Closes #5223
Init, compute, and fini methods are changed to work on internal context object.
This is necessary because ABI does not guarantee that SIMD registers will be preserved
on function calls. This is technically the case in Linux kernel in between
`kfpu_begin()/kfpu_end()`, but it breaks user-space tests and some kernels that
don't require disabling preemption for using SIMD (osx).
Use scalar compute methods in-place for small buffers, and when the buffer size
does not meet SIMD size alignment.
Combine incrementally computed fletcher4 checksums. Checksums are combined
a posteriori, allowing for parallel computation on chunks to be implemented if
required. The algorithm is general, and does not add changes in each SIMD
implementation.
New test in ztest verifies incremental fletcher computations.
Checksum combining matrix for two buffers `a` and `b`, where `Ca` and `Cb` are
respective fletcher4 checksums, `Cab` is combined checksum, `s` is size of buffer
`b` (divided by sizeof(uint32_t)) is:
Tony Hutter [Mon, 15 Aug 2016 23:34:02 +0000 (19:34 -0400)]
Use 100MB pool for filetest_001_pos.ksh checksum test
As part of its tests, filetest_001_pos.ksh wipes the entire vdev to
create checksum errors. This patch uses the setup/cleanup scripts from
the scrub_mirror test to create a custom 100MB pool, rather than
using the entire device size that is passed into zfs-tests.sh
(which defaults to 2GB). This speeds up the buildbot tests, and also
makes it possible for someone to use real disks (say, 1TB) without the
test taking an insanely long amount of time.
ilovezfs [Thu, 28 Jan 2016 12:51:19 +0000 (04:51 -0800)]
OpenZFS 6585 - sha512, skein, and edonr have an unenforced dependency on extensible dataset
Authored by: ilovezfs <ilovezfs@icloud.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Richard Laager <rlaager@wiktel.com>
Approved by: Robert Mustacchi <rm@joyent.com>
Ported by: Tony Hutter <hutter2@llnl.gov>
In any pool without the extensible dataset feature flag already enabled,
creating a dataset with dedup set to use one of the new checksums would
result in the following panic as soon as any data was added:
Inpsection showed that feature->fi_feature was 7, which is the value of
SPA_FEATURE_EXTENSIBLE_DATASET in the spa_feature enum. This commit
adds extensible dataset as a dependency for the sha512, edonr, and skein
feature flags, which prevents the panic.
OpenZFS-issue: https://www.illumos.org/issues/6585
OpenZFS-commit: https://github.com/illumos/illumos-gate/commit/892586e8a147c02d7f4053cc405229a13e796928
Porting Notes:
This code was originally from Illumos, but I actually ported it from:
openzfsonosx/zfs@b62a652
ilovezfs [Tue, 26 Jan 2016 07:41:11 +0000 (23:41 -0800)]
OpenZFS 6541 - Pool feature-flag check defeated if "verify" is included in the dedup property value
Authored by: ilovezfs <ilovezfs@icloud.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Richard Laager <rlaager@wiktel.com>
Approved by: Robert Mustacchi <rm@joyent.com> Ported-by: Tony Hutter <hutter2@llnl.gov>
zio_checksum_to_feature() expects a zio_checksum enum not a raw property
intval, so the new checksums weren't being detected when the
ZIO_CHECKSUM_VERIFY flag got in the way.
would incorrectly succeed because ZIO_CHECKSUM_VERIFY would be in the
way, the raw intval would not be a member of the enum, and
zio_checksum_to_feature() would return SPA_FEATURE_NONE, with the result
that spa_feature_is_enabled() would never be called.
This was first detected with edonr, since in that case verify is
required.
This commit clears the ZIO_CHECKSUM_VERIFY flag before calling
zio_checksum_to_feature() using the ZIO_CHECKSUM_MASK and verifies in
zio_checksum_to_feature() that ZIO_CHECKSUM_MASK has been applied by the
caller to attempt to prevent the same bug from occurring again in the
future.
- Copied module/icp/include/sha2/sha2.h directly from illumos
- Removed from module/icp/algs/sha2/sha2.c:
#pragma inline(SHA256Init, SHA384Init, SHA512Init)
- Added 'ctx' to lib/libzfs/libzfs_sendrecv.c:zio_checksum_SHA256() since
it now takes in an extra parameter.
- Added CTASSERT() to assert.h from for module/zfs/edonr_zfs.c
- Added skein & edonr to libicp/Makefile.am
- Added sha512.S. It was generated from sha512-x86_64.pl in Illumos.
- Updated ztest.c with new fletcher_4_*() args; used NULL for new CTX argument.
- In icp/algs/edonr/edonr_byteorder.h, Removed the #if defined(__linux) section
to not #include the non-existant endian.h.
- In skein_test.c, renane NULL to 0 in "no test vector" array entries to get
around a compiler warning.
- Fixup test files:
- Rename <sys/varargs.h> -> <varargs.h>, <strings.h> -> <string.h>,
- Remove <note.h> and define NOTE() as NOP.
- Define u_longlong_t
- Rename "#!/usr/bin/ksh" -> "#!/bin/ksh -p"
- Rename NULL to 0 in "no test vector" array entries to get around a
compiler warning.
- Remove "for isa in $($ISAINFO); do" stuff
- Add/update Makefiles
- Add some userspace headers like stdio.h/stdlib.h in places of
sys/types.h.
- EXPORT_SYMBOL *_Init/*_Update/*_Final... routines in ICP modules.
- Update scripts/zfs2zol-patch.sed
- include <sys/sha2.h> in sha2_impl.h
- Add sha2.h to include/sys/Makefile.am
- Add skein and edonr dirs to icp Makefile
- Add new checksums to zpool_get.cfg
- Move checksum switch block from zfs_secpolicy_setprop() to
zfs_check_settable()
- Fix -Wuninitialized error in edonr_byteorder.h on PPC
- Fix stack frame size errors on ARM32
- Don't unroll loops in Skein on 32-bit to save stack space
- Add memory barriers in sha2.c on 32-bit to save stack space
- Add filetest_001_pos.ksh checksum sanity test
- Add option to write psudorandom data in file_write utility
Romain Dolbeau [Mon, 3 Oct 2016 16:44:00 +0000 (18:44 +0200)]
Add parity generation/rebuild using 128-bits NEON for Aarch64
This re-use the framework established for SSE2, SSSE3 and
AVX2. However, GCC is using FP registers on Aarch64, so
unlike SSE/AVX2 we can't rely on the registers being left alone
between ASM statements. So instead, the NEON code uses
C variables and GCC extended ASM syntax. Note that since
the kernel explicitly disable vector registers, they
have to be locally re-enabled explicitly.
As we use the variable's number to define the symbolic
name, and GCC won't allow duplicate symbolic names,
numbers have to be unique. Even when the code is not
going to be used (e.g. the case for 4 registers when
using the macro with only 2). Only the actually used
variables should be declared, otherwise the build
will fails in debug mode.
This requires the replacement of the XOR(X,X) syntax
by a new ZERO(X) macro, which does the same thing but
without repeating the argument. And perhaps someday
there will be a machine where there is a more efficient
way to zero a register than XOR with itself. This affects
scalar, SSE2, SSSE3 and AVX2 as they need the new macro.
It's possible to write faster implementations (different
scheduling, different unrolling, interleaving NEON and
scalar, ...) for various cores, but this one has the
advantage of fitting in the current state of the code,
and thus is likely easier to review/check/merge.
The only difference between aarch64-neon and aarch64-neonx2
is that aarch64-neonx2 unroll some functions some more.
Richard Laager [Sun, 2 Oct 2016 18:34:17 +0000 (13:34 -0500)]
Correct zpool_vdev_remove() error message
The error message in zpool_vdev_remove() said top-level devices
could be removed, but that has never been true.
Reported-by: Colin Ian King <colin.king@canonical.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Richard Laager <rlaager@wiktel.com>
Closes #4506
Closes #5213
coverity scan CID:147531,type: Argument cannot be negative
- may copy data with negative size
coverity scan CID:147532,type: resource leaks
- may close a fd which is negative
coverity scan CID:147533,type: resource leaks
- may call pwrite64 with a negative size
coverity scan CID:147535,type: resource leaks
- may call fdopen with a negative fd
Reviewed-by: Richard Laager <rlaager@wiktel.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: GeLiXin <ge.lixin@zte.com.cn>
Closes #5176
coverity scan CID:147536, type: Argument cannot be negative
- may write or close fd which is negative
coverity scan CID:147537, type: Argument cannot be negative
- may call dup2 with a negative fd
coverity scan CID:147538, type: Argument cannot be negative
- may read or fchown with a negative fd
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: GeLiXin <ge.lixin@zte.com.cn>
Closes #5185
Brian Behlendorf [Fri, 30 Sep 2016 22:04:21 +0000 (15:04 -0700)]
Fix cppcheck warning in buf_init()
Cppcheck 1.63 erroneously complains about an uninitialized value
in buf_init(). Newer versions of cppcheck (1.72) handle this
correctly but we'll initialize the value anyway to silence the
warning.
Reviewed-by: Richard Elling <Richard.Elling@RichardElling.com> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes #5203
Brian Behlendorf [Fri, 30 Sep 2016 19:12:53 +0000 (12:12 -0700)]
Disable zpool_import_002_pos and ro_props_001_pos
These test cases fail some percentage of the time resulting
in automated testing failures. Disable the offending tests
until they can be made reliable.
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Issue #5201
Issue #5202
Closes #5194
Gvozden Neskovic [Wed, 31 Aug 2016 08:12:08 +0000 (10:12 +0200)]
fix: Shift exponent too large
Undefined operation is reported by running ztest (or zloop) compiled with GCC
UndefinedBehaviorSanitizer. Error only happens on top level of dnode indirection
with large enough offset values. Logically, left shift operation would work,
but bit shift semantics in C, and limitation of uint64_t, do not produce desired
result.
Isaac Huang [Thu, 29 Sep 2016 20:13:31 +0000 (14:13 -0600)]
Explicit block device plugging when submitting multiple BIOs
Without plugging, the default 'noop' scheduler will not merge
the BIOs which are part of a large ZIO.
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Isaac Huang <he.huang@intel.com>
Closes #5181
Fix zfs_clone_010_pos.ksh to verify zfs clones property displays right
Because the macro ZFS_MAXPROPLEN used in function print_dataset
differs between platforms set it appropriately and calculate the expected
number of passes.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed-by: Richard Laager <rlaager@wiktel.com> Signed-off-by: yuxiang <guo.yong33@zte.com.cn>
Closes #5154
Fix zfs_clone_010_pos.ksh to verify the space used by multiple copies
The default blocksize in Linux is 1024 due to a GNU-ism. Setting the
expected blocksize resolves the issue. As mentioned in the PR an
alternate solution would be to set POSIXLY_CORRECT=1.
Reviewed-by: Richard Laager <rlaager@wiktel.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: yuxiang <guo.yong33@zte.com.cn>
Closes #5167
Refactor the code in such a way so that inode->i_mode is being set
at the same time zp->z_mode is being changed. This has the effect of
keeping both in sync without relying on zfs_inode_update.
Reviewed-by: Richard Laager <rlaager@wiktel.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Nikolay Borisov <n.borisov.lkml@gmail.com>
Closes #5158
Linux 4.7 compat: Fix deadlock during lookup on case-insensitive
We must not use d_add_ci if the dentry already has the real name. Otherwise,
d_add_ci()->d_alloc_parallel() will find itself on the lookup hash and wait
on itself causing deadlock.
OpenZFS 6111 - zfs send should ignore datasets created after the ending snapshot
Authored by: Alex Deiter <alex.deiter@gmail.com>
Reviewed by: Alex Aizman alex.aizman@nexenta.com
Reviewed by: Alek Pinchuk alek.pinchuk@nexenta.com
Reviewed by: Roman Strashkin roman.strashkin@nexenta.com
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: Paul Dagnelie <pcd@delphix.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Approved by: Garrett D'Amore <garrett@damore.org> Ported-by: kernelOfTruth <kerneloftruth@gmail.com>
OpenZFS-issue: https://www.illumos.org/issues/6111
OpenZFS-commit: https://github.com/illumos/illumos-gate/commit/4a20c933
Closes #5110
Porting notes:
There were changes from upstream due to the following commits:
- zfs send -p send properties only for snapshots that are actually sent
https://github.com/zfsonlinux/zfs/commit/057485504e3a4502c265813ab58e9ec8ffc2a3be
- Produce a full snapshot list for zfs send -p
https://github.com/zfsonlinux/zfs/commit/e890dd85a7522730ad46daf68150aafd3952d0c1
- Implement zfs_ioc_recv_new() for OpenZFS 2605
https://github.com/zfsonlinux/zfs/commit/43e52eddb13d8accbd052fac9a242ce979531aa4
- OpenZFS 6314 - buffer overflow in dsl_dataset_name
ZFS_MAXNAMELEN was changed to the now used ZFS_MAX_DATASET_NAME_LEN since
https://github.com/zfsonlinux/zfs/commit/eca7b76001a7d33f78bd98884aef8325bdbf98e7
Due to how the Linux VFS was designed busy mount points
cannot be destroyed even when given the force option. Update
the zfs_destroy_001_pos test case to expect this behavior when
running under Linux.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: candychencan <chen.can2@zte.com.cn>
Closes #5132
Brian Behlendorf [Wed, 21 Sep 2016 20:45:21 +0000 (13:45 -0700)]
Fix automatically generated release number
When building from the head of a branch a release number is
automatically generated with `git describe` using the last tag
on that branch as the base. For this to work the last tag on the
branch needs to be predictable given the current META file.
This logic was accidentally broken when an -rcX tag was added to
the branch. Update it to search for a VERSION or VERSION-RELEASE
tag.
Reviewed-by: Chris Siebenmann <cks.git01@cs.toronto.edu> Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes #5105
Closes #5140
Moritz Maxeiner [Wed, 21 Sep 2016 20:35:16 +0000 (22:35 +0200)]
Fix regression that broke dracut initramfs generation
Based upon @ryao's initial fix for 1c73494394fc9de9283b3fd4f00bcdf4bd300a7
( 5e9843405f63fdabe76e87b92b81a127d488abc7 ) this one also uses
`command -v` instead of `type`, but additionally only applies the
fix to close zfsonlinux/zfs#4749 when `libgcc_s.so.1` has not been included
by dracut automatically (verified by whether `zpool` links directly to
`libgcc_s.so`), as well as change the fallback option to match `libgcc_s.so*`.
Tested-by: Ben Jencks <ben@bjencks.net> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Moritz Maxeiner <moritz@ucworks.org>
Closes #5089
Closed #5138
zfs_commands.cfg have printed "No such file or directory", When executing
script/zfs-test.sh. The script is a symlink to ../../../zfs-script-config.sh
So delete the symlink, and directly source $SRCDIR/zfs-script-config.sh
when it exists from default.cfg.in
Reviewed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: legend-hua <liu.hua130@zte.com.cn>
Closes #5133
The FALLOC_FL_PUNCH_HOLE flag was introduced in the 2.6.38
kernel. To prevent breaking the build on older systems wrap its use
in a conditional. When FALLOC_FL_PUNCH_HOLE isn't available
return a non-zero status and error message.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: legend-hua <liu.hua130@zte.com.cn>
Closes #5101
Nikolay Borisov [Mon, 12 Sep 2016 19:35:56 +0000 (22:35 +0300)]
Simplify time handling logic in zfs_settattr
Simplify time handling in zfs_setattr by mimicking the logic in
setattr_copy from the linux kernel. In order to achieve this
in the case when ZFS' log is being replayed it is necessary
to unconditionally set the ctime in zfs_replay_setattr.
Also use the timespec_trunc function when assigning values to the
generic inode struct. This is currently a noop since zfs sets
s_time_gran to 1, however in the future rules about precision might
change.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed-by: Chunwei Chen <david.chen@osnexus.com> Signed-off-by: Nikolay Borisov <n.borisov.lkml@gmail.com>
Closes #4916
Nikolay Borisov [Mon, 1 Aug 2016 20:02:25 +0000 (23:02 +0300)]
Refactor generic inode time updating
ZFS doesn't provide a custom update_time method meaning it delegates
this job to the generic VFS layer. The only time when it needs to
set the various *time values is when the inode is being marshalled
to/from the disk. Do this by moving the relevant code from
zfs_inode_update_impl to zfs_node_alloc and zfs_rezget. As a result
from this change it is no longer necessary to have multiple versions
of the zfs_inode_update function - so just nuke them and leave only
one.
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed-by: Chunwei Chen <david.chen@osnexus.com> Signed-off-by: Nikolay Borisov <n.borisov.lkml@gmail.com>
Issue #227
Closes #4916
George Wilson [Thu, 2 Jun 2016 04:04:53 +0000 (00:04 -0400)]
OpenZFS 6950 - ARC should cache compressed data
Authored by: George Wilson <george.wilson@delphix.com>
Reviewed by: Prakash Surya <prakash.surya@delphix.com>
Reviewed by: Dan Kimmel <dan.kimmel@delphix.com>
Reviewed by: Matt Ahrens <mahrens@delphix.com>
Reviewed by: Paul Dagnelie <pcd@delphix.com>
Reviewed by: Tom Caputi <tcaputi@datto.com>
Reviewed by: Brian Behlendorf <behlendorf1@llnl.gov>
Ported by: David Quigley <david.quigley@intel.com>
This review covers the reading and writing of compressed arc headers, sharing
data between the arc_hdr_t and the arc_buf_t, and the implementation of a new
dbuf cache to keep frequently access data uncompressed.
I've added a new member to l1 arc hdr called b_pdata. The b_pdata always hangs
off the arc_buf_hdr_t (if an L1 hdr is in use) and points to the physical block
for that DVA. The physical block may or may not be compressed. If compressed
arc is enabled and the block on-disk is compressed, then the b_pdata will match
the block on-disk and remain compressed in memory. If the block on disk is not
compressed, then neither will the b_pdata. Lastly, if compressed arc is
disabled, then b_pdata will always be an uncompressed version of the on-disk
block.
Typically the arc will cache only the arc_buf_hdr_t and will aggressively evict
any arc_buf_t's that are no longer referenced. This means that the arc will
primarily have compressed blocks as the arc_buf_t's are considered overhead and
are always uncompressed. When a consumer reads a block we first look to see if
the arc_buf_hdr_t is cached. If the hdr is cached then we allocate a new
arc_buf_t and decompress the b_pdata contents into the arc_buf_t's b_data. If
the hdr already has a arc_buf_t, then we will allocate an additional arc_buf_t
and bcopy the uncompressed contents from the first arc_buf_t to the new one.
Writing to the compressed arc requires that we first discard the b_pdata since
the physical block is about to be rewritten. The new data contents will be
passed in via an arc_buf_t (uncompressed) and during the I/O pipeline stages we
will copy the physical block contents to a newly allocated b_pdata.
When an l2arc is inuse it will also take advantage of the b_pdata. Now the
l2arc will always write the contents of b_pdata to the l2arc. This means that
when compressed arc is enabled that the l2arc blocks are identical to those
stored in the main data pool. This provides a significant advantage since we
can leverage the bp's checksum when reading from the l2arc to determine if the
contents are valid. If the compressed arc is disabled, then we must first
transform the read block to look like the physical block in the main data pool
before comparing the checksum and determining it's valid.
Nikolay Borisov [Sat, 10 Sep 2016 20:06:17 +0000 (23:06 +0300)]
Refactor spa_load_l2cache to make build happy
In case sav->sav_config was NULL the body of the function
would skip the iteration of the l2 cache devices and will
just cleanup the old devices. However, this wasn't very obvious
since the null check was performed after the loop body and after
the old devices were cleaned. Refactor the code so that it's now
obvious when the iteration of the l2cache devices is skipped.
This fixes the following cppcheck warning:
[module/zfs/spa.c:1552]: (error) Possible null pointer dereference: newvdevs
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Nikolay Borisov <n.borisov.lkml@gmail.com>
Closes #5087
Tim Chase [Sat, 10 Sep 2016 15:16:13 +0000 (10:16 -0500)]
Free property names with spa_strfree() rather than strfree()
Since they're allocated with spa_strdup(), they should be freed with
spa_strfree() so the proper length buffer is freed.
Reviewed-by: Richard Yao <ryao@gentoo.org> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Tim Chase <tim@chase2k.com>
Closes #5082
Closes #5086
OpenZFS - Performance regression suite for zfstest
Author: John Wren Kennedy <john.kennedy@delphix.com>
Reviewed by: Prakash Surya <prakash.surya@delphix.com>
Reviewed by: Dan Kimmel <dan.kimmel@delphix.com>
Reviewed by: Matt Ahrens <mahrens@delphix.com>
Reviewed by: Paul Dagnelie <pcd@delphix.com>
Reviewed by: Don Brady <don.brady@intel.com>
Reviewed by: Richard Elling <Richard.Elling@RichardElling.com>
Reviewed by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed-by: David Quigley <david.quigley@intel.com>
Approved by: Richard Lowe <richlowe@richlowe.net> Ported-by: Don Brady <don.brady@intel.com>
OpenZFS-issue: https://www.illumos.org/issues/6950
OpenZFS-commit: https://github.com/openzfs/openzfs/commit/dcbf3bd6
Delphix-commit: https://github.com/delphix/delphix-os/commit/978ed49
Closes #4929
ZFS Test Suite Performance Regression Tests
This was pulled into OpenZFS via the compressed arc featureand was
separated out in zfsonlinux as a separate pull request from PR-4768.
It originally came in as QA-4903 in Delphix-OS from John Kennedy.
Porting Notes:
1. Added assertions in the setup script to make sure required tools
(fio, mpstat, ...) are present.
2. For the config.json generation in perf.shlib used arcstats and
other binaries instead of dtrace to query the values.
3. For the perf data collection:
- use "zpool iostat -lpvyL" instead of the io.d dtrace script
(currently not collecting zfs_read/write latency stats)
- mpstat and iostat take different arguments
- prefetch_io.sh is a placeholder that uses arcstats instead of
dtrace
4. Build machines require fio, mdadm and sysstat pakage (YMMV).
Future Work:
- Need a way to measure zfs_read and zfs_write latencies per pool.
- Need tools to takes two sets of output and display/graph the
differences
- Bring over additional regression tests from Delphix
Sydney Vanda [Fri, 22 Jul 2016 15:07:04 +0000 (15:07 +0000)]
Real disk partitioning now enabled in test suite for Linux
When using real devices, specify DISKS="sdb sdc sdd" opposed to
/dev/sdb in zfs-tests.sh - otherwise errors with directory names and
disk names registering as "/dev//dev/sdb" for some tests. The same
goes for mpath: DISK="mpatha mpathad mpathb"
Expected Usage:
$ DISKS="sdb sdc sdd" zfs-tests.sh
SLICE_PREFIX is now set as "p" for a loop device (ie loop0p2) or
"" for a real device (ie sdb2), or either for multipath devices
(ie mpatha1 or mpath1p1) instead of only "p" by default. Note that
kpartx partitioning is not currently supported in this patch
(ie "partx") and may need to be disabled on Debian distributions.
Functions added for determining test directory (/dev or /dev/mapper)
as well as slice prefix are determined and exported mostly in the cfg
file of each test group directory.
Currently zpools cannot be created on whole mpath devices that have
been partitioned. In order to fix this tests have either been revised
to use a partition instead, or if there is a size constraint and the
pool needs to be created on the whole disk, partitions are then deleted
if the device is a multipath device. This functionality is added to
default_cleanup() or to individual cleanup scripts if a non-default
cleanup method is used.
The max partitions is currently set at 8 to account for all of the
tests thus far.
Patch changes are generally encompassed in "if is_linux" construct.
Signed-off-by: Sydney Vanda <sydney.m.vanda@intel.com> Reviewed-by: John Salinas <John.Salinas@intel.com> Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov> Reviewed-by: David Quigley <david.quigley@intel.com>
Closes #4447
Closes #4964
Closes #5074
Don Brady [Wed, 31 Aug 2016 21:46:58 +0000 (15:46 -0600)]
Bring over illumos ZFS FMA logic -- phase 1
This first phase brings over the ZFS SLM module, zfs_mod.c, to handle
auto operations in response to disk events. Disk event monitoring is
provided from libudev and generates the expected payload schema for
zfs_mod. This work leverages the recently added devid and phys_path
strings in the vdev label.
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov> Signed-off-by: Don Brady <don.brady@intel.com> Signed-off-by: Tony Hutter <hutter2@llnl.gov>
Closes #4673