]> granicus.if.org Git - zfs/log
zfs
6 years agoOpenZFS 9286 - want refreservation=auto
Mike Gerdts [Wed, 11 Apr 2018 16:14:45 +0000 (10:14 -0600)]
OpenZFS 9286 - want refreservation=auto

Authored by: Mike Gerdts <mike.gerdts@joyent.com>
Reviewed by: Allan Jude <allanjude@freebsd.org>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Reviewed by: John Kennedy <john.kennedy@delphix.com>
Reviewed by: Andy Stormont <astormont@racktopsystems.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Approved by: Richard Lowe <richlowe@richlowe.net>
Ported-by: Don Brady <don.brady@delphix.com>
Porting Notes:
* Adopted destroy_dataset in ZTS test cleanup
* Use ksh shebang instead of bash for new tests

OpenZFS-issue: https://www.illumos.org/issues/9286
OpenZFS-commit: https://github.com/openzfs/openzfs/commit/723d0c85
Closes #7387

6 years agoFix zpool set feature@<feature>=disabled
LOLi [Wed, 11 Apr 2018 21:45:58 +0000 (23:45 +0200)]
Fix zpool set feature@<feature>=disabled

Commit e4010f2 accidentally allows zpool to set pool features to
"disabled"; this should only be allowed at pool creation. This commit
adds additional checks and test coverage to 'zpool set'.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: loli10K <ezomori.nozomu@gmail.com>
Closes #7402

6 years agozfs(8): fix dedup omission during mdoc overhaul
DeHackEd [Wed, 11 Apr 2018 00:19:01 +0000 (20:19 -0400)]
zfs(8): fix dedup omission during mdoc overhaul

The property description has been updated with new algorithms as well.

Reviewed-by: Matt Ahrens <matt@delphix.com>
Reviewed-by: Richard Elling <Richard.Elling@RichardElling.com>
Reviewed-by: loli10K <ezomori.nozomu@gmail.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: George Melikov <mail@gmelikov.ru>
Signed-off-by: DHE <git@dehacked.net>
Closes #7377

6 years agoFix race in dnode_check_slots_free()
Tom Caputi [Tue, 10 Apr 2018 18:15:05 +0000 (14:15 -0400)]
Fix race in dnode_check_slots_free()

Currently, dnode_check_slots_free() works by checking dn->dn_type
in the dnode to determine if the dnode is reclaimable. However,
there is a small window of time between dnode_free_sync() in the
first call to dsl_dataset_sync() and when the useraccounting code
is run when the type is set DMU_OT_NONE, but the dnode is not yet
evictable, leading to crashes. This patch adds the ability for
dnodes to track which txg they were last dirtied in and adds a
check for this before performing the reclaim.

This patch also corrects several instances when dn_dirty_link was
treated as a list_node_t when it is technically a multilist_node_t.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes #7147
Closes #7388

6 years agoLinux compat 4.16: blk_queue_flag_{set,clear}
Giuseppe Di Natale [Tue, 10 Apr 2018 17:32:14 +0000 (10:32 -0700)]
Linux compat 4.16: blk_queue_flag_{set,clear}

queue_flag_{set,clear}_unlocked are now private interfaces in
the Linux kernel (https://github.com/torvalds/linux/commit/8a0ac14).
Use blk_queue_flag_{set,clear} interfaces which were introduced as
of https://github.com/torvalds/linux/commit/8814ce8.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Closes #7410

6 years agoCorrect swapped keylocation error messages
Tom Caputi [Tue, 10 Apr 2018 04:11:17 +0000 (00:11 -0400)]
Correct swapped keylocation error messages

This patch corrects a small issue where two error messages
in the code that checks for invalid keylocations were
swapped.

Reviewed by: Matt Ahrens <mahrens@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes #7418

6 years agoRevert "Handle zap_add() failures in mixed ... "
Tony Hutter [Mon, 9 Apr 2018 21:24:46 +0000 (14:24 -0700)]
Revert "Handle zap_add() failures in mixed ... "

This reverts commit cc63068e95ee725cce03b1b7ce50179825a6cda5.

Under certain circumstances this change can result in an ENOSPC
error when adding new files to a directory.  See #7401 for full
details.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tony Hutter <hutter2@llnl.gov>
Issue #7401
Cloes #7416

6 years agoFix 'zfs send/recv' hang with 16M blocks
Brian Behlendorf [Mon, 9 Apr 2018 02:41:15 +0000 (19:41 -0700)]
Fix 'zfs send/recv' hang with 16M blocks

When using 16MB blocks the send/recv queue's aren't quite big
enough.  This change leaves the default 16M queue size which a
good value for most pools.  But it additionally ensures that the
queue sizes are at least twice the allowed zfs_max_recordsize.

Reviewed-by: loli10K <ezomori.nozomu@gmail.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes #7365
Closes #7404

6 years agoClean up (k)shlib and cfg file shebangs
Giuseppe Di Natale [Mon, 9 Apr 2018 02:37:22 +0000 (19:37 -0700)]
Clean up (k)shlib and cfg file shebangs

Most kshlib files are imported by other scripts
and do not have a shebang at the top of their files.
Make all kshlib follow this convention.

Remove shebangs from cfg files as well.

Reviewed-by: loli10K <ezomori.nozomu@gmail.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Close #7406

6 years agoFix "file is executable, but no shebang" warnings
Tony Hutter [Fri, 6 Apr 2018 23:34:21 +0000 (16:34 -0700)]
Fix "file is executable, but no shebang" warnings

Fedora 28's RPM build checks warn when executable files don't have a
shebang line.  These warnings are caused when we (incorrectly)
include data & config files in the_SCRIPTS automake lines. Files in
_SCRIPTS are marked executable by automake. This patch fixes the
issue by including non-executable scripts in a _DATA line instead.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Signed-off-by: Tony Hutter <hutter2@llnl.gov>
Closes #7359
Closes #7395

6 years agoExclude python scripts from RPM shebang check
Tony Hutter [Fri, 6 Apr 2018 23:32:58 +0000 (16:32 -0700)]
Exclude python scripts from RPM shebang check

The newest Fedora packaging rules print warnings for scripts using the
/usr/bin/python shebang:

    *** WARNING: mangling shebang in /usr/bin/arc_summary.py from
    #!/usr/bin/python to #!/usr/bin/python2. This will become an ERROR,
    fix it manually!

Fedora wants all cross compatible scripts to pick python3.  Since we
don't want our users to have to pick a specific version of python, we
exclude our scripts from the RPM build check.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Signed-off-by: Tony Hutter <hutter2@llnl.gov>
Closes #7360
Closes #7399

6 years agosystemd mount generator and tracking ZEDLET
Antonio Russo [Fri, 6 Apr 2018 21:11:09 +0000 (17:11 -0400)]
systemd mount generator and tracking ZEDLET

zfs-mount-generator implements the "systemd generator" protocol,
producing systemd.mount units from the cached outputs of zfs list,
during early boot, integrating with systemd.

Each pool has an indpendent cache of the command

  zfs list -H -oname,mountpoint,canmount -tfilesystem -r $pool

which is kept synchronized by the ZEDLET

  history_event-zfs-list-cacher.sh

Datasets not in the cache will be loaded later in the boot process by
zfs-mount.service, including pools without a cache.

Among other things, this allows for complex mount hierarchies.

Reviewed-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
Reviewed-by: Richard Laager <rlaager@wiktel.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Antonio Russo <antonio.e.russo@gmail.com>
Closes #7329

6 years agoFixes for SNPRINTF_BLKPTR with encrypted BP's
Matthew Ahrens [Fri, 6 Apr 2018 20:30:26 +0000 (13:30 -0700)]
Fixes for SNPRINTF_BLKPTR with encrypted BP's

mdb doesn't have dmu_ot[], so we need a different mechanism for its
SNPRINTF_BLKPTR() to determine if the BP is encrypted vs authenticated.

Additionally, since it already relies on BP_IS_ENCRYPTED (etc),
SNPRINTF_BLKPTR might as well figure out the "crypt_type" on its own,
rather than making the caller do so.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Tom Caputi <tcaputi@datto.com>
Signed-off-by: Matthew Ahrens <mahrens@delphix.com>
Closes #7390

6 years agoFix divide-by-zero in mmp_delay_update()
Olaf Faaland [Fri, 6 Apr 2018 20:29:11 +0000 (13:29 -0700)]
Fix divide-by-zero in mmp_delay_update()

vdev_count_leaves() in the denominator may return 0, caught by Coverity.
Introduced by

533ea04 Update mmp_delay on sync or skipped, failed write

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Reviewed-by: George Melikov <mail@gmelikov.ru>
Signed-off-by: Olaf Faaland <faaland1@llnl.gov>
Closes #7391

6 years agoMake encrypted "zfs mount -a" failures consistent
Tom Caputi [Fri, 6 Apr 2018 20:28:15 +0000 (16:28 -0400)]
Make encrypted "zfs mount -a" failures consistent

Currently, "zfs mount -a" will print a warning and fail to mount
any encrypted datasets that do not have a key loaded. This patch
makes the behavior of this failure consistent with other failure
modes ("zfs mount -a" will silently continue, explict "zfs mount"
will print a message and return an error code.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes #7382

6 years agoUpdate mmp_delay on sync or skipped, failed write
Olaf Faaland [Wed, 4 Apr 2018 23:38:44 +0000 (16:38 -0700)]
Update mmp_delay on sync or skipped, failed write

When an MMP write is skipped, or fails, and time since
mts->mmp_last_write is already greater than mts->mmp_delay, increase
mts->mmp_delay.  The original code only updated mts->mmp_delay when a
write succeeded, but this results in the write(s) after delays and
failed write(s) reporting an ub_mmp_delay which is too low.

Update mmp_last_write and mmp_delay if a txg sync was successful.  At
least one uberblock was written, thus extending the time we can be sure
the pool will not be imported by another host.

Do not allow mmp_delay to go below (MSEC2NSEC(zfs_multihost_interval) /
vdev_count_leaves()) so that a period of frequent successful MMP writes,
e.g. due to frequent txg syncs, does not result in an import activity
check so short it is not reliable based on mmp thread writes alone.

Remove unnecessary local variable, start.  We do not use the start time
of the loop iteration.

Add a debug message in spa_activity_check() to allow verification of the
import_delay value and to prove the activity check occurred.

Alter the tests that import pools and attempt to detect an activity
check.  Calculate the expected duration of spa_activity_check() based on
module parameters at the time the import is performed, rather than a
fixed time set in mmp.cfg.  The fixed time may be wrong.  Also, use the
default zfs_multihost_interval value so the activity check is longer and
easier to recognize.

Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Signed-off-by: Olaf Faaland <faaland1@llnl.gov>
Closes #7330

6 years agoFedora 28: Fix misc bounds check compiler warnings
Tony Hutter [Wed, 4 Apr 2018 17:16:47 +0000 (10:16 -0700)]
Fedora 28: Fix misc bounds check compiler warnings

Fix a bunch of (mostly) sprintf/snprintf truncation compiler
warnings that show up on Fedora 28 (GCC 8.0.1).

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tony Hutter <hutter2@llnl.gov>
Closes #7361
Closes #7368

6 years agoFix spa reference leak in zfs_ioc_pool_scan
LOLi [Wed, 4 Apr 2018 00:31:30 +0000 (02:31 +0200)]
Fix spa reference leak in zfs_ioc_pool_scan

zfs_ioc_pool_scan leaks a spa reference when zc->zc_flags is not a
valid pool_scrub_cmd_t: this could happen if the userland binaries
and ZFS kernel module differ in version and would prevent the pool from
being exported.

Reviewed by: Matt Ahrens <matt@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Signed-off-by: loli10K <ezomori.nozomu@gmail.com>
Closes #7380

6 years agoFix add_nested_replacing_spare test case
LOLi [Wed, 4 Apr 2018 00:30:14 +0000 (02:30 +0200)]
Fix add_nested_replacing_spare test case

Use 'zpool reopen' instead of 'zpool scrub' to kick in the spare device:
this is required to avoid spurious failures caused by a race condition
in events processing by the ZFS Event Daemon:

P1 (zpool scrub)                            P2 (zed)
---
zfs_ioc_pool_scan()
 -> dsl_scan()
  -> vdev_reopen()
   -> vdev_set_state(VDEV_STATE_CANT_OPEN)
                                            zfs_ioc_vdev_attach()
                                             -> spa_vdev_attach()
                                              -> dsl_resilver_restart()
  -> dsl_sync_task()
   -> dsl_scan_setup_check()
   <- dsl_scan_setup_check(): EBUSY

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: loli10K <ezomori.nozomu@gmail.com>
Closes #7247
Closes #7342

6 years agoRemove ASSERT() in l2arc_apply_transforms()
Tim Chase [Sat, 31 Mar 2018 22:14:21 +0000 (17:14 -0500)]
Remove ASSERT() in l2arc_apply_transforms()

The ASSERT was erroneously copied from the next section of code.
The buffer's size should be expanded from "psize" to "asize"
if necessary.

Reviewed-by: Tom Caputi <tcaputi@datto.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tim Chase <tim@chase2k.com>
Closes #7375

6 years agoDecryption error handling improvements
Tom Caputi [Sat, 31 Mar 2018 18:12:51 +0000 (14:12 -0400)]
Decryption error handling improvements

Currently, the decryption and block authentication code in
the ZIO / ARC layers is a bit inconsistent with regards to
the ereports that are produces and the error codes that are
passed to calling functions. This patch ensures that all of
these errors (which begin as ECKSUM) are converted to EIO
before they leave the ZIO or ARC layer and that ereports
are correctly generated on each decryption / authentication
failure.

In addition, this patch fixes a bug in zio_decrypt() where
ECKSUM never gets written to zio->io_error.

Reviewed by: Matt Ahrens <matt@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes #7372

6 years agoEncrypted dnode blocks should be prefetched raw
Tom Caputi [Sat, 31 Mar 2018 18:11:48 +0000 (14:11 -0400)]
Encrypted dnode blocks should be prefetched raw

Encrypted dnode blocks are always initially read as raw data and
converted to decrypted data when an encrypted bonus buffer is
needed. This allows the DMU to be used for things like fetching
the DMU master node without requiring keys to be loaded. However,
dbuf_issue_final_prefetch() does not currently read the data as
raw. The end result of this is that prefetched dnode blocks are
read twice from disk: once decrypted and then again as raw data.
This patch corrects the issue by adding the flag when appropriate.

Reviewed by: Matt Ahrens <matt@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes #7362

6 years agoFix hung z_zvol tasks during 'zfs receive'
LOLi [Fri, 30 Mar 2018 19:10:01 +0000 (21:10 +0200)]
Fix hung z_zvol tasks during 'zfs receive'

During a receive operation zvol_create_minors_impl() can wait
needlessly for the prefetch thread because both share the same tasks
queue.  This results in hung tasks:

<3>INFO: task z_zvol:5541 blocked for more than 120 seconds.
<3>      Tainted: P           O  3.16.0-4-amd64
<3>"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.

The first z_zvol:5541 (zvol_task_cb) is waiting for the long running
traverse_prefetch_thread:260

root@linux:~# cat /proc/spl/taskq
taskq                       act  nthr  spwn  maxt   pri  mina
spl_system_taskq/0            1     2     0    64   100     1
active: [260]traverse_prefetch_thread [zfs](0xffff88003347ae40)
wait: 5541
spl_delay_taskq/0             0     1     0     4   100     1
delay: spa_deadman [zfs](0xffff880039924000)
z_zvol/1                      1     1     0     1   120     1
active: [5541]zvol_task_cb [zfs](0xffff88001fde6400)
pend: zvol_task_cb [zfs](0xffff88001fde6800)

This change adds a dedicated, per-pool, prefetch taskq to prevent the
traverse code from monopolizing the global (and limited) system_taskq by
inappropriately scheduling long running tasks on it.

Reviewed-by: Albert Lee <trisk@forkgnu.org>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: loli10K <ezomori.nozomu@gmail.com>
Closes #6330
Closes #6890
Closes #7343

6 years agozfs-functions: skip lines where mntpnt is 'none'
Georgy Yakovlev [Fri, 30 Mar 2018 19:05:24 +0000 (11:05 -0800)]
zfs-functions: skip lines where mntpnt is 'none'

This fixes zfs-mount initscript trying to mount swap volumes
as filesystems or anything that has 'none' as a mountpoint
in /etc/fstab.  Additionally, fixes trying to mount swap volumes
as a filesystem on RHEL.  RHEL defines mountpoint for swap
as `swap`.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Reviewed-by: George Melikov <mail@gmelikov.ru>
Signed-off-by: Georgy Yakovlev <ya@sysdump.net>
Closes #7346
Closes #7347

6 years agoOpenZFS 9164 - assert: newds == os->os_dsl_dataset
Andriy Gapon [Wed, 21 Feb 2018 12:55:55 +0000 (14:55 +0200)]
OpenZFS 9164 - assert: newds == os->os_dsl_dataset

Authored by: Andriy Gapon <avg@FreeBSD.org>
Reviewed by: Matt Ahrens <mahrens@delphix.com>
Reviewed by: Don Brady <don.brady@delphix.com>
Reviewed-by: loli10K <ezomori.nozomu@gmail.com>
Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Approved by: Richard Lowe <richlowe@richlowe.net>
Ported-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Porting Notes:
* Re-enabled and tweaked the zpool_upgrade_007_pos test case
  to successfully run in under 5 minutes.

OpenZFS-issue: https://www.illumos.org/issues/9164
OpenZFS-commit: https://github.com/openzfs/openzfs/commit/0e776dc06a
Closes #6112
Closes #7336

6 years agoAdd support for nvme based devids
Don Brady [Fri, 30 Mar 2018 00:43:25 +0000 (18:43 -0600)]
Add support for nvme based devids

Adds a devid for nvme devices. This is very similar to how the
other 'bus' (scsi|sata|usb) devids are generated. The devid
resides in a name/value pair in the leaf vdevs in a zpool config.

Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Richard Elling <Richard.Elling@RichardElling.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Don Brady <don.brady@delphix.com>
Closes #7356

6 years agoResolve QAT issues with incompressible data
Tom Caputi [Fri, 30 Mar 2018 00:40:34 +0000 (20:40 -0400)]
Resolve QAT issues with incompressible data

Currently, when ZFS wants to accelerate compression with QAT, it
passes a destination buffer of the same size as the source buffer.
Unfortunately, if the data is incompressible, QAT can actually
"compress" the data to be larger than the source buffer. When this
happens, the QAT driver will return a FAILED error code and print
warnings to dmesg. This patch fixes these issues by providing the
QAT driver with an additional buffer to work with so that even
completely incompressible source data will not cause an overflow.

This patch also resolves an error handling issue where
incompressible data attempts compression twice: once by QAT and
once in software. To fix this issue, a new (and fake) error code
CPA_STATUS_INOMPRESSIBLE has been added so that the calling code
can correctly account for the difference between a hardware
failure and data that simply cannot be compressed.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Weigang Li <weigang.li@intel.com>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes #7338

6 years agoFix ASSERT in dsl_scan_fini() and cleanup comments
Tom Caputi [Thu, 29 Mar 2018 01:30:44 +0000 (21:30 -0400)]
Fix ASSERT in dsl_scan_fini() and cleanup comments

This patch fixes an issue where dsl_scan_prefetch_cb() might
add more prefetch I/Os to the prefetch queue after prefetching
has been completed. This was happening because that code was
checking scn->scn_suspending instead of scn->scn_prefetch_stop.
This occasionally triggered an ASSERT during ztest runs in
dsl_scan_fini() when the code attempted to destroy an AVL tree
that still had entires in it. This patch also includes a number
of spelling corrections and comment cleanups throughout
dsl_scan.c

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes #7353

6 years agochmod -x on etc/init.d/zfs-*.in automake files
Tony Hutter [Wed, 28 Mar 2018 19:08:21 +0000 (12:08 -0700)]
chmod -x on etc/init.d/zfs-*.in automake files

Clear executable bit on zfs-import.in, zfs-mount.in,
zfs-share.in, and zfs-zed.in.  These are automake files and
should not be marked executable.  This fixes a RPM build error
on Fedora 28.

Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tony Hutter <hutter2@llnl.gov>
Closes #7355
Closes #7327

6 years agoFix mmap / libaio deadlock
Brian Behlendorf [Wed, 28 Mar 2018 17:19:22 +0000 (10:19 -0700)]
Fix mmap / libaio deadlock

Calling uiomove() in mappedread() under the page lock can result
in a deadlock if the user space page needs to be faulted in.

Resolve the issue by dropping the page lock before the uiomove().
The inode range lock protects against concurrent updates via
zfs_read() and zfs_write().

Reviewed-by: Albert Lee <trisk@forkgnu.org>
Reviewed-by: Chunwei Chen <david.chen@nutanix.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes #7335
Closes #7339

6 years agoRemove libattr requirement
DeHackEd [Tue, 27 Mar 2018 23:51:33 +0000 (19:51 -0400)]
Remove libattr requirement

RHEL/CentOS 6 supports sys/xattr.h eliminating the need for
libattr-devel as a dependency.

Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: DHE <git@dehacked.net>
Closes #7344
Closes #7351

6 years agoOpenZFS 9321 - arc_loan_compressed_buf() can increment arc_loaned_bytes by the wrong...
Allan Jude [Tue, 20 Feb 2018 05:31:14 +0000 (00:31 -0500)]
OpenZFS 9321 - arc_loan_compressed_buf() can increment arc_loaned_bytes by the wrong value

Authored by: Allan Jude <allanjude@freebsd.org>
Reviewed by: Matt Ahrens <matt@delphix.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Approved by: Garrett D'Amore <garrett@damore.org>
Ported-by: Giuseppe Di Natale <dinatale2@llnl.gov>
OpenZFS-issue: https://www.illumos.org/issues/9321
OpenZFS-commit: https://github.com/openzfs/openzfs/commit/92b05f3a18
Closes #7333

6 years agoFedora 28: Fix "Macro %_dracutdir has empty body"
Tony Hutter [Sun, 25 Mar 2018 22:00:47 +0000 (15:00 -0700)]
Fedora 28: Fix "Macro %_dracutdir has empty body"

If you run ./configure --with-config=srpm, it will not trigger
the user m4 scripts to populate the dracut and udev directories.
This causes a build error on Fedora 28.  Make the dracut and
udev lines conditional to get around this.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tony Hutter <hutter2@llnl.gov>
Closes #7326
Closes #7328

6 years agoDon't count embedded bps in read stats
Tom Caputi [Sat, 24 Mar 2018 04:35:19 +0000 (00:35 -0400)]
Don't count embedded bps in read stats

Currently, ZFS tracks statistics about calls to arc_read()
via the /proc/spl/kstat/zfs/<pool>/reads file for debugging.
Unfortunately, this file currently counts embedded bps as
disk reads since they are technically processed by the ZIO
layer. This pollutes the log since the ARC will never cache
embedded bps. This patch  corrects this issue by preventing
the logging of embedded bp reads.

Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes #7334

6 years agoOpenZFS 9193 - bootcfg -C doesn't work
Paul Dagnelie [Thu, 23 Mar 2017 22:28:22 +0000 (15:28 -0700)]
OpenZFS 9193 - bootcfg -C doesn't work

When given an empty string as a rootds value, bootcfg -C fails with
the error message 'could not set nextboot: '' is an invalid name'.
This should be allowed because it represents clearing the nextboot
configuration.

Authored by: Paul Dagnelie <pcd@delphix.com>
Reviewed by: Chris Williamson <chris.williamson@delphix.com>
Reviewed by: Sebastien Roy <sebastien.roy@delphix.com>
Reviewed by: Igor Kozhukhov <igor@dilos.org>
Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Approved by: Robert Mustacchi <rm@joyent.com>
Ported-by: Giuseppe Di Natale <dinatale2@llnl.gov>
OpenZFS-issue: https://www.illumos.org/issues/9193
OpenZFS-commit: https://github.com/openzfs/openzfs/commit/504645d227
Closes #7230

6 years agoClarify zpool actions for an intent log device
Peter Ashford [Thu, 22 Mar 2018 22:12:08 +0000 (15:12 -0700)]
Clarify zpool actions for an intent log device

Updated the "Intent Log" section of the "zpool" manual page to
properly reflect the actions that may be performed.

Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Peter Ashford <ashford@accs.com>
Closes #6938
Closes #7318

6 years agomodprobe zfs during dracut mount
kpande [Thu, 22 Mar 2018 17:14:29 +0000 (13:14 -0400)]
modprobe zfs during dracut mount

Resolves importing root pool during boot in dracut.  This case was
inadvertently broken with the module autoloading change in #7287.

Reviewed-by: Matthew Thode <prometheanfire@gentoo.org>
Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Signed-off-by: Kash Pande <kash@tripleback.net>
Closes #7322

6 years agoenable zfs_dbgmsg() by default, without dprintf()
Matthew Ahrens [Wed, 21 Mar 2018 22:37:32 +0000 (15:37 -0700)]
enable zfs_dbgmsg() by default, without dprintf()

zfs_dbgmsg() should record a message by default.  As a general
principal, these messages shouldn't be too verbose.  Furthermore, the
amount of memory used is limited to 4MB (by default).

dprintf() should only record a message if this is a debug build, and
ZFS_DEBUG_DPRINTF is set in zfs_flags.  This flag is not set by default
(even on debug builds).  These messages are extremely verbose, and
sometimes nontrivial to compute.

SET_ERROR() should only record a message if ZFS_DEBUG_SET_ERROR is set
in zfs_flags.  This flag is not set by default (even on debug builds).

This brings our behavior in line with illumos.  Note that the message
format is unchanged (including file, line, and function, even though
these are not recorded on illumos).

Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Prakash Surya <prakash.surya@delphix.com>
Signed-off-by: Matthew Ahrens <mahrens@delphix.com>
Closes #7314

6 years agoFix spelling errors in comments
Tom Caputi [Wed, 21 Mar 2018 15:42:13 +0000 (11:42 -0400)]
Fix spelling errors in comments

This patch simply corrects some spelling / grammar errors in
the QAT and encryption code comments. No functional changes

Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes #7319

6 years agoAdd support for nvme disk detection
timor [Wed, 21 Mar 2018 15:35:20 +0000 (16:35 +0100)]
Add support for nvme disk detection

This treats /dev/nvme.. devices the same way as /dev/sd... devices.  The
motivation behind this is that whole disk detection did not work on nvme
SSDs without that, because it DKC_UNKNOWN was returned for such devices.

Perhaps there should be a separate DKC_ type for this, but I don't know
enough about the code to know the implications of that.

Reviewed-by: Don Brady <don.brady@delphix.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: timor <timor.dd@googlemail.com>
Closes #7304

6 years agoAdd comments for portable dnode / objset flags
Tom Caputi [Tue, 20 Mar 2018 18:55:21 +0000 (14:55 -0400)]
Add comments for portable dnode / objset flags

This patch adds some comments describing the purpose of "portable"
dnode and objset flags so that it is clear when new flags should
be added to the repective flag masks. This patch includes no
functional changes.

Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes #7313

6 years agoAdd JSON output support to channel programs
Alek P [Mon, 19 Mar 2018 19:40:58 +0000 (15:40 -0400)]
Add JSON output support to channel programs

The changes piggyback JSON output support on top of channel programs
(#6558).  This way the JSON output support is targeted to scripting
use cases and is easily maintainable since it really only touches
one function (zfs_do_channel_program()).

This patch ports Joyent's JSON nvlist library from illumos to enable
easy JSON printing of channel program output nvlist.  To keep the
delta small I also took advantage of the fact that printing in
zfs_do_channel_program() was almost always done before exiting
the program.

Reviewed by: Matt Ahrens <mahrens@delphix.com>
Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Richard Elling <Richard.Elling@RichardElling.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Alek Pinchuk <apinchuk@datto.com>
Closes #7281

6 years agoFix deadlock in IO pipeline
Brian Behlendorf [Fri, 16 Mar 2018 23:46:06 +0000 (16:46 -0700)]
Fix deadlock in IO pipeline

In vdev_queue_aggregate() the zio_execute() bypass should not be
called under the vdev queue lock.  This can result in a deadlock
as shown in the stack traces below.

Drop the vdev queue lock then walk the parents of the aggregate IO
to determine the list of component IOs to be bypassed.  This can
be done safely without holding the io_lock since the new aggregate
IO has not yet been returned and its parents cannot change.

---  THREAD 1 ---
arc_read()
  zio_nowait()
    zio_vdev_io_start()
      vdev_queue_io() <--- mutex_enter(vq->vq_lock)
        vdev_queue_io_to_issue()
          vdev_queue_aggregate()
            zio_execute()
              zio_vdev_io_assess()
                zio_wait_for_children() <- mutex_enter(zio->io_lock)

--- THREAD 2 --- (inverse order)
arc_read()
  zio_change_priority() <- mutex_enter(zio->zio_lock)
    vdev_queue_change_io_priority() <- mutex_enter(vq->vq_lock)

Reviewed-by: Tom Caputi <tcaputi@datto.com>
Reviewed-by: Don Brady <don.brady@delphix.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes #7307

6 years agoAllow QAT to handle non page-aligned buffers
Tom Caputi [Fri, 16 Mar 2018 23:02:49 +0000 (19:02 -0400)]
Allow QAT to handle non page-aligned buffers

This patch adds some handling to the QAT acceleration functions
that allows them to handle buffers that are not aligned with the
page cache. At the moment this never happens since callers only
happen to work with page-aligned buffers, but this code should
prevent headaches if this isn't always true in the future. This
patch also adds some cleanups to align the QAT compression code
with the encryption and checksumming code.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Weigang Li <weigang.li@intel.com>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes #7305

6 years agoReport pool suspended due to MMP
Olaf Faaland [Thu, 15 Mar 2018 17:56:55 +0000 (10:56 -0700)]
Report pool suspended due to MMP

When the pool is suspended, record whether it was due to an I/O error or
due to MMP writes failing to succeed within the required time.

Change spa_suspended from uint8_t to zio_suspend_reason_t to store the
reason.

When userspace queries pool status via spa_tryimport(), report the
reason the pool was suspended in a new key,
ZPOOL_CONFIG_SUSPENDED_REASON.

In libzfs, when interpreting the returned config nvlist, report
suspension due to MMP with a new pool status enum value,
ZPOOL_STATUS_IO_FAILURE_MMP.

In status_callback(), which generates and emits the message when 'zpool
status' is executed, add a case to print an appropriate message for the
new pool status enum value.

Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Signed-off-by: Olaf Faaland <faaland1@llnl.gov>
Closes #7296

6 years agoSHA256 QAT acceleration
Tom Caputi [Thu, 15 Mar 2018 17:53:58 +0000 (13:53 -0400)]
SHA256 QAT acceleration

This patch enables acceleration of SHA256 checksums using Intel
Quick Assist Technology. This patch also fixes up and refactors
some of the code from QAT encryption to make the behavior
consistent.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Chengfeix Zhu <chengfeix.zhu@intel.com>
Signed-off-by: Weigang Li <weigang.li@intel.com>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes #7295

6 years agoOpenZFS 9076 - Adjust perf test concurrency settings
Stephen Blinick [Sun, 11 Feb 2018 23:11:59 +0000 (16:11 -0700)]
OpenZFS 9076 - Adjust perf test concurrency settings

ZFS Performance test concurrency should be lowered for better latency

Work by Stephen Blinick.

Nightly performance runs typically consist of two levels of concurrency;
and both are fairly high.

Since the IO runs are to a ZFS filesystem, within a zpool, which is
based on some variable number of vdev's, the amount of IO driven to each
device is variable. Additionally, different device types (HDD vs SSD,
etc) can generally handle a different amount of concurrent IO before
saturating.

Nevertheless, in practice, it appears that most tests are well past the
concurrency saturation point and therefore both perform with the same
throughput, the maximum of the device. Because the queuedepth to the
device(s) is so high however, the latency is much higher than the best
possible at that throughput, and increases linearly with the increase in
concurrency.

This means that changes in code that impact latency during normal
operation (before saturation) may not be apparent when a large component
of the measured latency is from the IO sitting in a queue to be
serviced. Therefore, changing the concurrency settings is recommended

Authored by: Stephen Blinick <stephen.blinick@delphix.com>
Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed by: Dan Kimmel <dan.kimmel@delphix.com>
Reviewed by: John Wren Kennedy <jwk404@gmail.com>
Ported-by: John Wren Kennedy <jwk404@gmail.com>
OpenZFS-issue: https://www.illumos.org/issues/9076
OpenZFS-commit: https://github.com/openzfs/openzfs/pull/562
Upstream bug: DLPX-45477
Closes #7302

6 years agoreceive_spill does not byte swap spill contents
Paul Zuchowski [Thu, 15 Mar 2018 17:29:51 +0000 (13:29 -0400)]
receive_spill does not byte swap spill contents

In zfs receive, the function receive_spill should account
for spill block endian conversion as a defensive measure.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Tom Caputi <tcaputi@datto.com>
Signed-off-by: Paul Zuchowski <pzuchowski@datto.com>
Closes #7300

6 years agoOpenZFS 9188 - increase size of dbuf cache to reduce indirect block decompression
Brian Behlendorf [Tue, 13 Mar 2018 17:52:48 +0000 (10:52 -0700)]
OpenZFS 9188 - increase size of dbuf cache to reduce indirect block decompression

With compressed ARC (bug 6950) we use up to 25% of our CPU to decompress
indirect blocks, under a workload of random cached reads. To reduce this
decompression cost, we would like to increase the size of the dbuf cache so
that more indirect blocks can be stored uncompressed.

If we are caching entire large files of recordsize=8K, the indirect blocks
use 1/64th as much memory as the data blocks (assuming they have the same
compression ratio). We suggest making the dbuf cache be 1/32nd of all memory,
so that in this scenario we should be able to keep all the indirect blocks
decompressed in the dbuf cache. (We want it to be more than the 1/64th that
the indirect blocks would use because we need to cache other stuff in the dbuf
cache as well.)

In real world workloads, this won't help as dramatically as the example above,
but we think it's still worth it because the risk of decreasing performance is
low. The potential negative performance impact is that we will be slightly
reducing the size of the ARC (by ~3%).

Porting Notes:
* Added modules options to zfs-module-parameters.5 man page.
* Preserved scaling based on target ARC size rather than max ARC size.

Authored by: George Wilson <george.wilson@delphix.com>
Reviewed by: Dan Kimmel <dan.kimmel@delphix.com>
Reviewed by: Prashanth Sreenivasa <pks@delphix.com>
Reviewed by: Paul Dagnelie <pcd@delphix.com>
Reviewed-by: Richard Elling <Richard.Elling@RichardElling.com>
Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Ported-by: Brian Behlendorf <behlendorf1@llnl.gov>
OpenZFS-issue: https://www.illumos.org/issues/9188
OpenZFS-commit: https://github.com/openzfs/openzfs/pull/564
Upstream bug: DLPX-46942
Closes #7273

6 years agoAdd kernel module auto-loading
Brian Behlendorf [Tue, 13 Mar 2018 17:45:55 +0000 (10:45 -0700)]
Add kernel module auto-loading

Historically a dynamic misc minor number was registered for the
/dev/zfs device in order to prevent minor number collisions.  This
was fine but it prevented us from being able to use the kernel
module auto-loaded which requires a known reserved value.

Resolve this issue by adding a configure test to find an available
misc minor number which can then be used in MODULE_ALIAS_MISCDEV at
build time.  By adding this alias the zfs kmod is added to the list
of known static-nodes and the systemd-tmpfiles-setup-dev service
will create a /dev/zfs character device at boot time.

This in turn allows us to update the 90-zfs.rules file to make it
aware this is a static node.  The upshot of this is that whenever
a process (zpool, zfs, zed) opens the /dev/zfs the kmods will be
automatic loaded.  This even works for unprivileged users so there
is no longer a need to manually load the modules at boot time.

As an additional bonus the zed now no longer needs to start after
the zfs-import.service since it will trigger the module load.

In the unlikely event the minor number we selected conflicts with
another out of tree unregistered minor number the code falls back
to dynamically allocating it.  In this case the modules again
must be manually loaded.

Note that due to the change in the method of registering the minor
number the zimport.sh test case may incorrectly fail when the
static node for the installed packages is created instead of the
dynamic one.  This issue will only transiently impact zimport.sh
for this single commit when we transition and are mixing and
matching methods.

Reviewed-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
TEST_ZIMPORT_SKIP="yes"
Closes #7287

6 years agoAdd zfs_scan_ignore_errors tunable
Tim Chase [Tue, 13 Mar 2018 17:43:14 +0000 (12:43 -0500)]
Add zfs_scan_ignore_errors tunable

When it's set, a DTL range will be cleared even if its scan/scrub had
errors.  This allows to work around resilver/scrub upon import when the
pool has errors.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tim Chase <tim@chase2k.com>
Closes #7293

6 years agoDestroy makes full snap list before destroying
Paul Zuchowski [Mon, 12 Mar 2018 22:24:08 +0000 (18:24 -0400)]
Destroy makes full snap list before destroying

Change zfs destroy logic so destroying begins before
the entire list of snapshots is built.

Reviewed-by: loli10K <ezomori.nozomu@gmail.com>
Reviewed-by: Kash Pande <kash@tripleback.net>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Paul Zuchowski <pzuchowski@datto.com>
Closes #7271

6 years agoFix race in trace point in zrl_add_impl
Chunwei Chen [Mon, 12 Mar 2018 18:27:02 +0000 (11:27 -0700)]
Fix race in trace point in zrl_add_impl

We hit an illegal memory access in the zrlock trace point. The problem
is that zrl->zr_owner and zrl->zr_caller are assigned locklessly. And if
zrl->zr_owner got assigned a longer string between when __string()
calculate the strlen, and when __assign_str() does strcpy. The copy will
overflow the buffer.

==
For example:

Initial condition:
zrl->zr_owner = A
zrl->zr_caller = "abc"

Thread A                                 Thread B
-------------------------------------------------
if (zrl->zr_owner == A) {
  DTRACE_PROBE2() {
    __string() {
      strlen(zrl->zr_caller) -> 3
      allocate buf[4]
    }

                                        zrl->zr_owner = B
        zrl->zr_caller = "abcd"

    __assign_str() {
      strcpy(buf, zrl->zr_caller) <- buffer overflow
==

Dereferencing zrl->zr_owner->pid may also be problematic, in that the
zrl->zr_owner got changed to other task, and that task exits, freeing
the task_struct. This should be very unlikely, as the other task need to
zrl_remove and exit between the dereferencing zr->zr_owner and
zr->zr_owner->pid. Nevertheless, we'll deal with it as well.

To fix the zrl->zr_caller issue, instead of copy the string content, we
just copy the pointer, this is safe because it always points to
__func__, which is static. As for the zrl->zr_owner issue, we pass in
curthread instead of using zrl->zr_owner.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Chunwei Chen <david.chen@nutanix.com>
Closes #7291

6 years agoFix MMP write frequency for large pools
Brian Behlendorf [Mon, 12 Mar 2018 18:26:05 +0000 (11:26 -0700)]
Fix MMP write frequency for large pools

When a single pool contains more vdevs than the CONFIG_HZ for
for the kernel the mmp thread will not delay properly.  Switch
to using cv_timedwait_sig_hires() to handle higher resolution
delays.

This issue was reported on Arch Linux where HZ defaults to only
100 and this could be fairly easily reproduced with a reasonably
large pool.  Most distribution kernels set CONFIG_HZ=250 or
CONFIG_HZ=1000 and thus are unlikely to be impacted.

Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Reviewed-by: Olaf Faaland <faaland1@llnl.gov>
Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes #7205
Closes #7289

6 years agoHold SCL_VDEV when counting leaves
Olaf Faaland [Thu, 8 Mar 2018 23:39:07 +0000 (15:39 -0800)]
Hold SCL_VDEV when counting leaves

A config lock should be held while vdev_count_leaves() walks the tree,
otherwise the pointers reference may become invalid during the walk.

SCL_VDEV is a minimal lock provided for such uses cases.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Signed-off-by: Olaf Faaland <faaland1@llnl.gov>
Closes #7286

6 years agoHandle zio_resume and mmp => off
Olaf Faaland [Thu, 8 Mar 2018 23:21:54 +0000 (15:21 -0800)]
Handle zio_resume and mmp => off

When multihost is disabled on a pool, and the pool is resumed via zpool
clear, within a single cycle of the mmp thread's loop (e.g.  while it's
in the cv_timedwait call), both mmp_last_write and mmp_delay should be
updated.

The original code mistakenly treated the two cases as if they could not
occur at the same time.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Signed-off-by: Olaf Faaland <faaland1@llnl.gov>
Closes #7286

6 years agoDocument allowed pool names
Tomohiro Kusumi [Thu, 22 Feb 2018 17:31:34 +0000 (02:31 +0900)]
Document allowed pool names

PR #7208 was a patch to allow non-reserved pool names which begin with
mirror, raidz, spare (but do not equal), however we'd rather document
it in the man page for compatibility with other OpenZFS implementations,
to avoid pool names that may not work on non-Linux platforms.

Reviewed-by: Richard Laager <rlaager@wiktel.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tomohiro Kusumi <kusumi.tomohiro@osnexus.com>
Closes #7216

6 years agoFix zfs-kmod builds when using rpm >= 4.14
LOLi [Fri, 9 Mar 2018 21:52:37 +0000 (22:52 +0100)]
Fix zfs-kmod builds when using rpm >= 4.14

With rpm-software-management/rpm@5e94633 a package version containing
invalid characters (most commonly a double '-') causes the kmod package
generation to terminate with an error.  This change takes advantage of
the newly introduced rpm macro "_wrong_version_format_terminate_build"
to allow kmod packages to be built.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: loli10K <ezomori.nozomu@gmail.com>
Closes #7284

6 years agoChange functions which return literals to return `const char*`
Tomohiro Kusumi [Fri, 9 Mar 2018 21:47:32 +0000 (06:47 +0900)]
Change functions which return literals to return `const char*`

get_format_prompt_string() and zpool_state_to_name() return
a string literal which is read-only, thus they should return
`const char*`.

zpool_get_prop_string() returns a non-const string after
successful nv-lookup, and returns a string literal otherwise.
Since this function is designed to be used for read-only purpose,
the return type should also be `const char*`.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tomohiro Kusumi <kusumi.tomohiro@osnexus.com>
Closes #7285

6 years agoQAT support for AES-GCM
Tom Caputi [Fri, 9 Mar 2018 21:37:15 +0000 (16:37 -0500)]
QAT support for AES-GCM

This patch adds support for acceleration of AES-GCM encryption
with Intel Quick Assist Technology.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Chengfeix Zhu <chengfeix.zhu@intel.com>
Signed-off-by: Weigang Li <weigang.li@intel.com>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes #7282

6 years agozdb and inuse tests don't pass with real disks
Paul Zuchowski [Thu, 8 Mar 2018 01:03:33 +0000 (20:03 -0500)]
zdb and inuse tests don't pass with real disks

Due to zpool create auto-partioning in Linux (i.e. sdb1),
certain utilities need to use the parition (sdb1) while
others use the whole disk name (sdb).

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Paul Zuchowski <pzuchowski@datto.com>
Closes #6939
Closes #7261

6 years agoTake user namespaces into account in policy checks
Wolfgang Bumiller [Wed, 7 Mar 2018 23:40:42 +0000 (00:40 +0100)]
Take user namespaces into account in policy checks

Change file related checks to use user namespaces and make
sure involved uids/gids are mappable in the current
namespace.

Note that checks without file ownership information will
still not take user namespaces into account, as some of
these should be handled via 'zfs allow' (otherwise root in a
user namespace could issue commands such as `zpool export`).

This also adds an initial user namespace regression test
for the setgid bit loss, with a user_ns_exec helper usable
in further tests.

Additionally, configure checks for the required user
namespace related features are added for:
  * ns_capable
  * kuid/kgid_has_mapping()
  * user_ns in cred_t

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Wolfgang Bumiller <w.bumiller@proxmox.com>
Closes #6800
Closes #7270

6 years agoZTS: fix send-c_stream_size_estimate
Brian Behlendorf [Wed, 7 Mar 2018 17:55:54 +0000 (09:55 -0800)]
ZTS: fix send-c_stream_size_estimate

The test could fail when attempting to write to a newly created
volume which was missing its device node.  Resolve the issue by
calling block_device_wait() which blocks until udev creates the
needed  entry.

Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes #7276
Closes #7277

6 years agoFix dbufstats_001_pos
Giuseppe Di Natale [Wed, 7 Mar 2018 17:53:04 +0000 (09:53 -0800)]
Fix dbufstats_001_pos

Implement a new helper within_tolerance to test if a value
is within range of a target.

Because the dbufstats and dbufs kstat file are being read
at slightly different times, it is possible for stats to be
slightly off. Use within_tolerance to determine if the value
is "close enough" to the target.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Closes #7239
Closes #7266

6 years agoAllow to limit zed's syslog chattiness
Tony Hutter [Tue, 6 Mar 2018 23:41:52 +0000 (15:41 -0800)]
Allow to limit zed's syslog chattiness

Some usage patterns like send/recv of replication streams can
produce a large number of events. In such a case, the current
all-syslog.sh zedlet will hold up to its name, and flood the
logs with mostly redundant information. Two mitigate this
situation, this changeset introduces to new variables
ZED_SYSLOG_SUBCLASS_INCLUDE and ZED_SYSLOG_SUBCLASS_EXCLUDE
to zed.rc that give more control over which event classes end
up in the syslog.

Reviewed-by: loli10K <ezomori.nozomu@gmail.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Signed-off-by: Tony Hutter <hutter2@llnl.gov>
Signed-off-by: Daniel Kobras <d.kobras@science-computing.de>
Closes #6886
Closes #7260

6 years agoRecord skipped MMP writes in multihost_history
Olaf Faaland [Tue, 27 Feb 2018 01:32:49 +0000 (20:32 -0500)]
Record skipped MMP writes in multihost_history

Once per pass through the MMP thread's loop, the vdev tree is walked to
find a suitable leaf to write the next MMP block to.  If no such leaf is
found, the thread sleeps for a while and resumes at the top of the loop.

Add an entry to multihost_history when no leaf can be found, and record
the reason in the error column.  The error code for such entries is a
bitfield, displayed in hex:

0x1  At least one vdev (interior or leaf) was not writeable.
0x2  At least one writeable leaf vdev was found, but it had a pending
MMP write.

timestamp = the time in seconds since the epoch when no leaf could be
found originally.

duration = the time (in ns) during which no MMP block was written for
this reason.  This does not include the preceeding inter-write period
nor the following inter-write period.

vdev_guid = the number of sequential cycles of the MMP thread looop when
this occurred.

Sample output, truncated to fit:

For records of skipped MMP writes the right-most column, vdev_path, is
reported as "-".

id   txg  timestamp   error  duration    mmp_delay  vdev_guid     ...
936  11   1520036441  0      146264      891422313  1740883117838 ...
937  11   1520036441  0      163956      888356657  7320395061548 ...
938  11   1520036442  0      130690      885314969  7320395061548 ...
939  11   1520036442  0      2001068577  882296582  1740883117838 ...
940  11   1520036443  0      161806      882296582  7320395061548 ...
941  11   1520036443  0x2    0           998020546  1             ...
942  11   1520036444  0      136585      998020546  7320395061548 ...
943  11   1520036444  0x2    0           998020257  1             ...
944  11   1520036445  5      2002662964  994160219  1740883117838 ...
945  11   1520036445  0x2    998073118   994160219  3             ...
946  11   1520036447  0      247136      994160219  7320395061548 ...

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Olaf Faaland <faaland1@llnl.gov>
Closes #7212

6 years agoDetect long config lock acquisition in mmp
Olaf Faaland [Wed, 21 Feb 2018 01:33:51 +0000 (17:33 -0800)]
Detect long config lock acquisition in mmp

If something holds the config lock as a writer for too long, MMP will
fail to issue MMP writes in a timely manner.  This will result either in
the pool being suspended, or in an extreme case, in the pool not being
protected.

If the time to acquire the config lock exceeds 1/10 of the minimum
zfs_multihost_interval, report it in the zfs debug log.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Olaf Faaland <faaland1@llnl.gov>
Closes #7212

6 years agoIntroduce a destroy_dataset helper
Giuseppe Di Natale [Tue, 6 Mar 2018 22:54:57 +0000 (14:54 -0800)]
Introduce a destroy_dataset helper

Datasets can be busy when calling zfs destroy. Introduce
a helper function to destroy datasets and use it to destroy
datasets in zfs_allow_004_pos, zfs_promote_008_pos, and
zfs_destroy_002_pos.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Closes #7224
Closes #7246
Closes #7249
Closes #7267

6 years agoMisc fixes and cleanup for project quota
Nasf-Fan [Mon, 5 Mar 2018 20:56:27 +0000 (04:56 +0800)]
Misc fixes and cleanup for project quota

1) The Coverity Scan reports some issues for the project
   quota patch, including:

1.1) zfs_prop_get_userquota() directly uses the const quota
   type value as the condition check by wrong.

1.2) dmu_objset_userquota_get_ids() may cause dnode::dn_newgid
   to be overwritten by dnode::dn->dn_oldprojid.

2) This patch fixes related issues. It also enhances the logic
   for zfs_project_item_alloc() to avoid buffer overflow.

3) Skip project quota ability check if does not change project
   quota related things (id or flag). Otherwise, it will cause
   chattr (for other non project quota flags) operation failed
   if project quota disabled.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Fan Yong <fan.yong@intel.com>
Closes #7251
Closes #7265

6 years agoLinux 4.16 compat: get_disk_and_module()
Giuseppe Di Natale [Mon, 5 Mar 2018 20:44:35 +0000 (12:44 -0800)]
Linux 4.16 compat: get_disk_and_module()

As of https://github.com/torvalds/linux/commit/fb6d47a, get_disk()
is now get_disk_and_module(). Add a configure check to determine
if we need to use get_disk_and_module().

Reviewed-by: loli10K <ezomori.nozomu@gmail.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Closes #7264

6 years agoChange checksum & IO delay ratelimit values
Tony Hutter [Mon, 5 Mar 2018 01:34:51 +0000 (17:34 -0800)]
Change checksum & IO delay ratelimit values

Change checksum & IO delay ratelimit thresholds from 5/sec to 20/sec.
This allows zed to actually trigger if a bunch of these events arrive in
a short period of time (zed has a threshold of 10 events in 10 sec).
Previously, if you had, say, 100 checksum errors in 1 sec, it would get
ratelimited to 5/sec which wouldn't trigger zed to fault the drive.

Also, convert the checksum and IO delay thresholds to module params for
easy testing.

Reviewed-by: loli10K <ezomori.nozomu@gmail.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Signed-off-by: Tony Hutter <hutter2@llnl.gov>
Closes #7252

6 years agoIncrement zil_itx_needcopy_bytes properly
chrisrd [Fri, 2 Mar 2018 18:01:53 +0000 (05:01 +1100)]
Increment zil_itx_needcopy_bytes properly

In zil_lwb_commit() with TX_WRITE, we copy the log write record (lrw)
into the log write block (lwb) and send it off using zil_lwb_add_txg().
If we also have WR_NEED_COPY, we additionally copy the lwr's data into
the lwb to be sent off.  If the lwr + data doesn't fit into the lwb, we
send the lrw and as much data as will fit (dnow bytes), then go back
and do the same with the remaining data.

Each time through this loop we're sending dnow data bytes. I.e.
zil_itx_needcopy_bytes should be incremented by dnow.

Reviewed-by: Richard Elling <Richard.Elling@RichardElling.com>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Chris Dunlop <chris@onthe.net.au>
Closes #6988
Closes #7176

6 years agoZTS: fix spurious failures in mv_files
chrisrd [Fri, 2 Mar 2018 17:57:29 +0000 (04:57 +1100)]
ZTS: fix spurious failures in mv_files

The test could fail because of a race condition between the files being
generated in the background and attempting to move the files. Wait for
all file generation to complete before trying to move the files around.

Also, clean up the waiting: the 'wait' command without arguments waits
for all child pids.

Reviewed-by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Chris Dunlop <chris@onthe.net.au>
Closes #7220
Closes #7242
Closes #7258

6 years agoAdd ZFS perf test for dbuf cache
John Wren Kennedy [Wed, 28 Feb 2018 18:38:37 +0000 (10:38 -0800)]
Add ZFS perf test for dbuf cache

This change adds a test for sequential reads out of the dbuf cache.
It's essentially a copy of sequential_reads_cached, using a smaller
data set. The sequential read tests are renamed to differentiate them.

Authored by: Dan Kimmel <dan.kimmel@delphix.com>
Reviewed by: Paul Dagnelie <pcd@delphix.com>
Reviewed by: Matt Ahrens <mahrens@delphix.com>
Reviewed by: George Wilson <george.wilson@delphix.com>
Reviewed by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: John Wren Kennedy <john.kennedy@delphix.com>
Closes #7225

6 years agoFix some typos
John Eismeier [Wed, 28 Feb 2018 16:57:10 +0000 (11:57 -0500)]
Fix some typos

Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed by: George Melikov <mail@gmelikov.ru>
Signed-off-by: John Eismeier <john.eismeier@gmail.com>
Closes #7237

6 years agoFix zpool(8) list example to match actual format
Tomohiro Kusumi [Wed, 28 Feb 2018 16:54:53 +0000 (01:54 +0900)]
Fix zpool(8) list example to match actual format

a05dfd00 (Illumos 5147) has swapped FRAG and EXPANDSZ,
so it's natural to modify these examples.

 # zpool list | head -1
 NAME     SIZE  ALLOC   FREE  EXPANDSZ   FRAG    CAP  DEDUP  HEALTH  ALTROOT
                              ^^^^^^^^^^^^^^^

Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tomohiro Kusumi <kusumi.tomohiro@osnexus.com>
Closes #7244

6 years agoAdd Python 3 rewrite of arc_summary.py
Scot W. Stevenson [Wed, 28 Feb 2018 16:52:34 +0000 (17:52 +0100)]
Add Python 3 rewrite of arc_summary.py

Add new script arc_summary3.py as a complete rewrite of the
arc_summary.py tool (see issue #6873)

Add new options:

        -g/--graph    - Display crude graphic representation
                        of ARC status and quit
        -r/--raw      - Print all available information as
                        minimally formatted list (for grep)
        -s/--section  - Print a single section. This
                        replaces -p/--page, which is kept for
                        backwards use but marked as
                        depreciated

Add new sections with information on ZIL and SPL. Notify user
if sections L2ARC and VDEV are skipped instead of failing
silently. Add warning that -p/--page option is depreciated.

Developed for Python 3.5.

Reviewed-by: Richard Laager <rlaager@wiktel.com>
Reviewed-by: Richard Elling <Richard.Elling@RichardElling.com>
Reviewed by: George Melikov <mail@gmelikov.ru>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Scot W. Stevenson <scot.stevenson@gmail.com>
Closes #6873
Closes #6892

6 years agoAdd SMART self-test results to zpool status -c
Tony Hutter [Tue, 27 Feb 2018 17:31:27 +0000 (09:31 -0800)]
Add SMART self-test results to zpool status -c

Add in SMART self-test results to zpool status|iostat -c.  This
works for both SAS and SATA drives.

Also, add plumbing to allow the 'smart' script to take smartctl
output from a directory of output text files instead of running
it against the vdevs.

Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tony Hutter <hutter2@llnl.gov>
Closes #7178

6 years agoRaw DRR_OBJECT records must write raw data
Tom Caputi [Tue, 27 Feb 2018 17:04:05 +0000 (12:04 -0500)]
Raw DRR_OBJECT records must write raw data

b1d21733 made it possible for empty metadnode blocks to be
compressed to a hole, fixing a bug that would cause invalid
metadnode MACs when a send stream attempted to free objects
and allowing the blocks to be reclaimed when they were no
longer needed. However, this patch also introduced a race
condition; if a txg sync occurred after a DRR_OBJECT_RANGE
record was received but before any objects were added, the
metadnode block would be compressed to a hole and lose all
of its encryption parameters. This would cause subsequent
DRR_OBJECT records to fail when they attempted to write
their data into an unencrypted block. This patch defers the
DRR_OBJECT_RANGE handling to receive_object() so that the
encryption parameters are set with each object that is
written into that block.

Reviewed-by: Kash Pande <kash@tripleback.net>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes #7215
Closes #7236

6 years agoIncorrect maximum DVA value in DDE_GET_NDVAS()
Tim Chase [Mon, 26 Feb 2018 22:20:12 +0000 (16:20 -0600)]
Incorrect maximum DVA value in DDE_GET_NDVAS()

The conditional was reversed which caused garbage values to be used when
calculating dds_ref_dsize.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Tom Caputi <tcaputi@datto.com>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Signed-off-by: Tim Chase <tim@chase2k.com>
Closes #7234

6 years agoFix segfault in zfs_do_bookmark()
LOLi [Mon, 26 Feb 2018 17:55:18 +0000 (18:55 +0100)]
Fix segfault in zfs_do_bookmark()

When invoked with wrong parameters 'zfs bookmark' fails to gracefully
validate user input and crashes.

This is a regression accidentally introduced in 587e228; this commit
adds additional tests to the ZFS Test Suite to exercise this codepath.

Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: KireinaHoro <i@jsteward.moe>
Signed-off-by: loli10K <ezomori.nozomu@gmail.com>
Closes #7228
Closes #7229

6 years agoZTS: Fix zfs_share_* test case failures
Brian Behlendorf [Sat, 24 Feb 2018 18:07:12 +0000 (10:07 -0800)]
ZTS: Fix zfs_share_* test case failures

Prevent false positives when running the zfs_share_* test
cases due to leftover stale /var/lib/nfs/etab entries.  When
starting the test group re-synchronize the /var/lib/nfs/etab
file with /etc/exports.  At this point in the testing there
will be no additional `zfs share` entries to add.

Reviewed by: George Melikov <mail@gmelikov.ru>
Reviewed-by: loli10K <ezomori.nozomu@gmail.com>
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
Closes #7226

6 years agoShellcheck cleanup for initrd scripts
Kash Pande [Thu, 22 Feb 2018 01:54:54 +0000 (20:54 -0500)]
Shellcheck cleanup for initrd scripts

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Co-authored-by: Kash Pande <kash@tripleback.net>
Co-authored-by: Matthew Thode <mthode@mthode.org>
Signed-off-by: Kash Pande <kash@tripleback.net>
Signed-off-by: Matthew Thode <mthode@mthode.org>
Closes #7214

6 years agoEnable booting from nested encrypted datasets
Kash Pande [Tue, 20 Feb 2018 18:13:20 +0000 (13:13 -0500)]
Enable booting from nested encrypted datasets

- enable booting from nested encrypted datasets
- fix plymouth boot splash passphrase entry
- optimize unlock process

Co-authored-by: Kash Pande <kash@tripleback.net>
Co-authored-by: Matthew Thode <mthode@mthode.org>
Signed-off-by: Kash Pande <kash@tripleback.net>
Signed-off-by: Matthew Thode <mthode@mthode.org>
Closes #7214

6 years agoAdd scrub after resilver zed script
Tony Hutter [Fri, 23 Feb 2018 19:38:05 +0000 (11:38 -0800)]
Add scrub after resilver zed script

* Add a zed script to kick off a scrub after a resilver.  The script is
disabled by default.

* Add a optional $PATH (-P) option to zed to allow it to use a custom
$PATH for its zedlets.  This is needed when you're running zed under
the ZTS in a local workspace.

* Update test scripts to not copy in all-debug.sh and all-syslog.sh by
default.  They can be optionally copied in as part of zed_setup().
These scripts slow down zed considerably under heavy events loads and
can cause events to be dropped or their delivery delayed. This was
causing some sporadic failures in the 'fault' tests.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: Richard Laager <rlaager@wiktel.com>
Signed-off-by: Tony Hutter <hutter2@llnl.gov>
Closes #4662
Closes #7086

6 years agoFix free memory calculation on v3.14+
chrisrd [Fri, 23 Feb 2018 16:50:06 +0000 (03:50 +1100)]
Fix free memory calculation on v3.14+

Provide infrastructure to auto-configure to enum and API changes in the
global page stats used for our free memory calculations.

arc_free_memory has been broken since an API change in Linux v3.14:

2016-07-28 v4.8 599d0c95 mm, vmscan: move LRU lists to node
2016-07-28 v4.8 75ef7184 mm, vmstat: add infrastructure for per-node
  vmstats

These commits moved some of global_page_state() into
global_node_page_state(). The API change was particularly egregious as,
instead of breaking the old code, it silently did the wrong thing and we
continued using global_page_state() where we should have been using
global_node_page_state(), thus indexing into the wrong array via
NR_SLAB_RECLAIMABLE et al.

There have been further API changes along the way:

2017-07-06 v4.13 385386cf mm: vmstat: move slab statistics from zone to
  node counters
2017-09-06 v4.14 c41f012a mm: rename global_page_state to
  global_zone_page_state

...and various (incomplete, as it turns out) attempts to accomodate
these changes in ZoL:

2017-08-24 2209e409 Linux 4.8+ compatibility fix for vm stats
2017-09-16 787acae0 Linux 3.14 compat: IO acct, global_page_state, etc
2017-09-19 661907e6 Linux 4.14 compat: IO acct, global_page_state, etc

The config infrastructure provided here resolves these issues going back
to the original API change in v3.14 and is robust against further Linux
changes in this area.

Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: George Melikov <mail@gmelikov.ru>
Signed-off-by: Chris Dunlop <chris@onthe.net.au>
Closes #7170

6 years agoReport duration and error in mmp_history entries
Olaf Faaland [Thu, 22 Feb 2018 23:34:34 +0000 (15:34 -0800)]
Report duration and error in mmp_history entries

After an MMP write completes, update the relevant mmp_history entry
with the time between submission and completion, and the error
status of the write.

[faaland1@toss3a zfs]$ cat /proc/spl/kstat/zfs/pool/multihost
39 0 0x01 100 8800 69147946270893 72723903122926
id       txg     timestamp  error  duration   mmp_delay    vdev_guid
10607    1166    1518985089 0      138301     637785455    4882...
10608    1166    1518985089 0      136154     635407747    1151...
10609    1166    1518985089 0      803618560  633048078    9740...
10610    1166    1518985090 0      144826     633048078    4882...
10611    1166    1518985090 0      164527     666187671    1151...

Where duration = gethrtime_in_done_fn - gethrtime_at_submission, and
error = zio->io_error.

Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Olaf Faaland <faaland1@llnl.gov>
Closes #7190

6 years agoDo not initiate MMP writes while pool is suspended
Olaf Faaland [Thu, 22 Feb 2018 17:14:46 +0000 (09:14 -0800)]
Do not initiate MMP writes while pool is suspended

While the pool is suspended on host A, it may be imported on host B.
If host A continued to write MMP blocks, it would be blindly
overwriting MMP blocks written by host B, and the blocks written by
host A would have outdated txg information.

Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Olaf Faaland <faaland1@llnl.gov>
Closes #7182

6 years agoLinux 4.16 compat: use correct *_dec_and_test()
Tony Hutter [Thu, 22 Feb 2018 17:02:06 +0000 (09:02 -0800)]
Linux 4.16 compat: use correct *_dec_and_test()

Use refcount_dec_and_test() on 4.16+ kernels, atomic_dec_and_test()
on older kernels.  https://lwn.net/Articles/714974/

Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tony Hutter <hutter2@llnl.gov>
Closes: #7179
Closes: #7211
6 years agoFix bounds check in zio_crypt_do_objset_hmacs
Tom Caputi [Thu, 22 Feb 2018 16:50:14 +0000 (11:50 -0500)]
Fix bounds check in zio_crypt_do_objset_hmacs

The current bounds check in zio_crypt_do_objset_hmacs() does not
properly handle the possible sizes of the objset_phys_t and
can therefore read outside the buffer's memory. If that memory
happened to match what the check was actually looking for, the
objset would fail to be owned, complaining that the MAC was
invalid.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes #7210

6 years agoOpenZFS 9035 - zfs: this statement may fall through
Toomas Soome [Sat, 1 Jul 2017 07:13:08 +0000 (10:13 +0300)]
OpenZFS 9035 - zfs: this statement may fall through

Authored by: Toomas Soome <tsoome@me.com>
Reviewed by: Yuri Pankov <yuripv@yuripv.net>
Reviewed by: Andy Fiddaman <omnios@citrus-it.co.uk>
Reviewed by: Matthew Ahrens <mahrens@delphix.com>
Approved by: Dan McDonald <danmcd@joyent.com>
Ported-by: Giuseppe Di Natale <dinatale2@llnl.gov>
OpenZFS-issue: https://www.illumos.org/issues/9035
OpenZFS-commit: https://github.com/openzfs/openzfs/commit/46ac8fdfc5
Closes #7206

6 years agoAllow modprobe to fail when called within systemd
Matthew Thode [Wed, 21 Feb 2018 22:45:35 +0000 (22:45 +0000)]
Allow modprobe to fail when called within systemd

This allows for systems with zfs built into the kernel manually to run
these services.  Otherwise the service will fail to start.

Reviewed-by: loli10K <ezomori.nozomu@gmail.com>
Reviewed-by: Kash Pande <kash@tripleback.net>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Matthew Thode <mthode@mthode.org>
Closes #7174

6 years agoAdd SMART attributes for SSD and NVMe
bunder2015 [Wed, 21 Feb 2018 21:52:47 +0000 (16:52 -0500)]
Add SMART attributes for SSD and NVMe

This adds the SMART attributes required to probe Samsung SSD and NVMe
(and possibly others) disks when using the "zpool status -c" command.

Reviewed-by: loli10K <ezomori.nozomu@gmail.com>
Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Reviewed-by: Tony Hutter <hutter2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: bunder2015 <omfgbunder@gmail.com>
Closes #7183
Closes #7193

6 years agoAllow make checkstyle and paxscript in build dir
chrisrd [Wed, 21 Feb 2018 20:35:59 +0000 (07:35 +1100)]
Allow make checkstyle and paxscript in build dir

Reviewed-by: Giuseppe Di Natale <dinatale2@llnl.gov>
Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Reviewed-by: George Melikov <mail@gmelikov.ru>
Signed-off-by: Chris Dunlop <chris@onthe.net.au>
Closes #7202

6 years agoWant 'zfs send -b'
LOLi [Wed, 21 Feb 2018 20:32:06 +0000 (21:32 +0100)]
Want 'zfs send -b'

This change implements 'zfs send -b' which can be used to send only
received property values whether or not they are overridden by local
settings.

This can be very useful during "restore" operations from a backup pool
because it allows to send only the property values originally sent
from the backup source, even though they were later modified on the
destination either by a 'zfs set' operation, explicit 'zfs inherit' or
overridden during the receive process via 'zfs receive -o|-x'.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: loli10K <ezomori.nozomu@gmail.com>
Closes #7156

6 years agoRaw receive should change key atomically
Tom Caputi [Wed, 21 Feb 2018 20:31:03 +0000 (15:31 -0500)]
Raw receive should change key atomically

Currently, raw zfs sends transfer the encrypted master keys and
objset_phys_t encryption parameters in the DRR_BEGIN payload of
each send file. Both of these are processed as soon as they are
read in dmu_recv_stream(), meaning that the new keys are set
before the new snapshot is received. In addition to the fact that
this changes the user's keys for the dataset earlier than they
might expect, the keys were never reset to what they originally
were in the event that the receive failed. This patch splits the
processing into objset handling and key handling, the later of
which is moved to dmu_recv_end() so that they key change can be
done atomically.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes #7200

6 years agoPrevent raw zfs recv -F if dataset is unencrypted
Tom Caputi [Wed, 21 Feb 2018 20:30:11 +0000 (15:30 -0500)]
Prevent raw zfs recv -F if dataset is unencrypted

The current design of ZFS encryption only allows a dataset to
have one DSL Crypto Key at a time. As a result, it is important
that the zfs receive code ensures that only one key can be in use
at a time for a given DSL Directory. zfs receive -F complicates
this, since the new dataset is received as a clone of the existing
one so that an atomic switch can be done at the end. To prevent
confusion about which dataset is actually encrypted a check was
added to ensure that encrypted datasets cannot use zfs recv -F to
completely replace existing datasets. Unfortunately, the check did
not take into account unencrypted datasets being overriden by
encrypted ones as a case.

Along the same lines, the code also failed to ensure that raw
recieves could not be done on top of existing unencrypted
datasets, which causes amny problems since the new stream cannot
be decrypted.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes #7199

6 years agoRaw receives must compress metadnode blocks
Tom Caputi [Wed, 21 Feb 2018 20:28:52 +0000 (15:28 -0500)]
Raw receives must compress metadnode blocks

Currently, the DMU relies on ZIO layer compression to free LO
dnode blocks that no longer have objects in them. However,
raw receives disable all compression, meaning that these blocks
can never be freed. In addition to the obvious space concerns,
this could also cause incremental raw receives to fail to mount
since the MAC of a hole is different from that of a completely
zeroed block.

This patch corrects this issue by adding a special case in
zio_write_compress() which will attempt to compress these blocks
to a hole even if ZIO_FLAG_RAW_ENCRYPT is set. This patch also
removes the zfs_mdcomp_disable tunable, since tuning it could
cause these same issues.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes #7198

6 years agoRemove unnecessary txg syncs from receive_object()
Tom Caputi [Wed, 21 Feb 2018 20:26:51 +0000 (15:26 -0500)]
Remove unnecessary txg syncs from receive_object()

1b66810b introduced serveral changes which improved the reliability
of zfs sends when large dnodes were involved. However, these fixes
required adding a few calls to txg_wait_synced() in the DRR_OBJECT
handling code. Although most of them are currently necessary, this
patch allows the code to continue without waiting in some cases
where it doesn't have to.

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes #7197

6 years agoAdd omitted set for os->os_next_write_raw
Tom Caputi [Wed, 21 Feb 2018 20:24:37 +0000 (15:24 -0500)]
Add omitted set for os->os_next_write_raw

This one line patch adds adds a set to os->os_next_write_raw
that was omitted when the code was updated in 1b66810. Without
it, the code (in some instances) could attempt to write raw
encrypted data as regular unencrypted data without the keys
being loaded, triggering an ASSERT in zio_encrypt().

Reviewed-by: Brian Behlendorf <behlendorf1@llnl.gov>
Signed-off-by: Tom Caputi <tcaputi@datto.com>
Closes #7196