aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Add explicit timeout to test stepRich Ercolani2022-02-152-0/+2
| | | | | | | | | | | | | | | | | | If we die from timeout of the whole GH action run, we don't run the collect step afterward, which can make it hard to investigate the timeout. If we timeout first in the test action, though, it qualifies as failure, and collects appropriately. (330 minutes seems like an acceptable tradeoff between the 6h timeout by default on the action and the 4h and change "functional" usually takes.) Reviewed-by: George Melikov <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Rich Ercolani <[email protected]> Closes #12999
* Workaround Debian's fake System.map behaviorRich Ercolani2022-02-101-0/+8
| | | | | | | | | | | | | Debian ships fake System.map files by default, leading to the invocation of depmod with them to flood you with errors about missing symbols. Let's notice and not do that. Reviewed-by: Ahelenia Ziemiańska <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Rich Ercolani <[email protected]> Closes #12862
* Proper support for DESTDIR and INSTALL_MOD_PATHJosé Luis Salvador Rufo2022-02-101-4/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | The environment variables DESTDIR and INSTALL_MOD_PATH must be mutually exclusive. https://www.gnu.org/prep/standards/html_node/DESTDIR.html https://www.kernel.org/doc/Documentation/kbuild/modules.txt This issue was discussed in this Buildroot thread: https://lists.buildroot.org/pipermail/buildroot/2021-August/621350.html I saw this behavior in other different projects, as: - Yocto Project: https://www.yoctoproject.org/pipermail/meta-freescale/2013-August/004307.html - Google IA Coral: https://coral.googlesource.com/linux-imx-debian/+/refs/heads/master/debian/rules For the above reasons, INSTALL_MOD_PATH will be set as DESTDIR by default. Reviewed-by: Tony Hutter <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: José Luis Salvador Rufo <[email protected]> Signed-off-by: Romain Naour <[email protected]> Closes #12577
* ZTS: alloc_class.ksh must wait for the process to exitBrian Behlendorf2022-02-102-5/+1
| | | | | | | | | | | | | | | The alloc_class_* tests may fail on Linux with an EBUSY error if `zfs destroy` is run before the `dd` process has had a chance to terminate. Wait on the pid after the `kill -9` to make sure. When testing I didn't observe any failures for the alloc_class tests. Remove them from the exceptions list, the CI was used to verify the tests pass on all platforms. Reviewed-by: John Kennedy <[email protected]> Reviewed-by: Rich Ercolani <[email protected]> Signed-off-by: Brian Behlendorf <[email protected]> Closes #12873
* ZTS: Avoid piping send directly to /dev/nullRich Ercolani2022-02-101-1/+3
| | | | | | | | | | | | | Unfortunately, #11445 means while we fail gracefully now, we still fail, unless people want to implement a complex workaround just to support /dev/null. So let's just use the cheap workaround in a test for now. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: John Kennedy <[email protected]> Signed-off-by: Rich Ercolani <[email protected]> Closes #12872
* ZTS: Fix zpool_reopen_[1-5] on Fedora 35Tony Hutter2022-02-101-2/+2
| | | | | | | | | | | | | | The zpool_reopen_[1-5] tests are failing Fedora 35 with: zpool_reopen_001_pos.ksh[64]: log_must[67]: log_pos[270]: wait_for_resilver_end[98]: wait_for_action: line 71: func: is read only Renaming 'func' -> 'funct' fixes the issue. Reviewed-by: George Melikov <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Tony Hutter <[email protected]> Closes #12871
* systemd: add weekly and monthly scrub timersGeorgy Yakovlev2022-02-106-1/+64
| | | | | | | | | | | | | | | | | Timers can be enabled as follows: systemctl enable [email protected] --now systemctl enable [email protected] --now Each timer will pull in zfs-scrub@${poolname}.service, which is not schedule-specific. Added PERIODIC SCRUB section to zpool-scrub.8. Reviewed-by: Richard Laager <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Georgy Yakovlev <[email protected]> Closes #12193
* Add init script to load keysogelpre2022-02-107-9/+155
| | | | | | | | | | | Add new init scripts which allow automatic loading of keys if keylocation property is set to a URI. Reviewed-by: Ahelenia Ziemiańska <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Benedikt Neuffer <[email protected]> Closes #11659 Closes #11662
* Notify on UNAVAIL statechangeFrancesco Mazzoli2022-02-101-2/+3
| | | | | | | | | | | | `UNAVAIL` is maybe not quite as concerning as `DEGRADED`, but still an event of notice, in my opinion. For example it is triggered when a drive goes missing. Reviewed-by: Don Brady <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Francesco Mazzoli <[email protected]> Closes #12629 Closes #12630
* Upstream: Add snapshot and zvol eventsJorgen Lundman2022-02-108-1/+70
| | | | | | | | | | | | | | | For kernel to send snapshot mount/unmount events to zed. For kernel to send symlink creates/removes on zvol plumbing. (/dev/run/dsk/zvol/$pool/$zvol -> /dev/diskX) If zed misses the ENODEV, all errors after are EINVAL. Treat any error as kernel module failure. Reviewed-by: Tony Hutter <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Jorgen Lundman <[email protected]> Closes #12416
* zed: Add Pushover notifierScott Colby2022-02-102-0/+99
| | | | | | | | | | Add zed_notify_pushover to zed-functions.sh, along with the necessary configuration variables in zed.rc. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Ahelenia Ziemiańska <[email protected]> Signed-off-by: Scott Colby <[email protected]> Closes #12012
* zed-functions.sh: escape newline to produce valid jsonPhil Kauffman2022-02-071-1/+1
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This was discovered when using Discords Slack compatible webhook. Slack webhooks works without the escape, however Discord rightly refuses the POST as it contains invalid JSON. https://discord.com/developers/docs/resources/webhook#execute-slackcompatible-webhook Valid (while escaping the newline: ``` + msg_json='{"text": "*ZFS scrub_finish error for test on quartz*\nZFS has detected a data error:\n\n eid: 124\n class: scrub_finish\n host: quartz\n time: \n error: \n objid: :\n pool: test\n"}' ``` Invalid (no escape): ``` + msg_json='{"text": "*ZFS scrub_finish error for test on quartz* ZFS has detected a data error:\n\n eid: 124\n class: scrub_finish\n host: quartz\n time: \n error: \n objid: :\n pool: test\n"}' ``` The new line gets rendered and not sent inside the JSON as intended. ``` ++ curl -X POST https://discord.com/api/webhooks/{webhook.id}/{webhook.token}/slack --header 'Content-Type: application/json' --data-binary '{"text": "*ZFS scrub_finish error for test on quartz* ZFS has detected a data error:\n\n eid: 124\n class: scrub_finish\n host: quartz\n time: \n error: \n objid: :\n pool: test\n"}' + msg_out='{"message": "Cannot send an empty message", "code": 50006}' ``` Test method: `root@quartz:/etc/zfs/zed.d# export ZED_ZEDLET_DIR=/etc/zfs/zed.d; export ZEVENT_EID=124; export ZEVENT_SUBCLASS=scrub_finish; export ZEVENT_POOL=test; export ZED_NOTIFY_DATA=1; bash -x ./data-notify.sh` Reviewed-by: Damian Szuberski <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Philip Kauffman <[email protected]> Closes #13049
* zed: send notification email by defaultshodanshok2022-02-071-2/+2
| | | | | | | Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: George Melikov <[email protected]> Signed-off-by: Gionatan Danti <[email protected]> Closes #12806
* Introduce a flag to skip comparing the local mac when raw sendingGeorge Amanakis2022-02-047-14/+163
| | | | | | | | | | | | | | | | | | | | | | | | | Raw receiving a snapshot back to the originating dataset is currently impossible because of user accounting being present in the originating dataset. One solution would be resetting user accounting when raw receiving on the receiving dataset. However, to recalculate it we would have to dirty all dnodes, which may not be preferable on big datasets. Instead, we rely on the os_phys flag OBJSET_FLAG_USERACCOUNTING_COMPLETE to indicate that user accounting is incomplete when raw receiving. Thus, on the next mount of the receiving dataset the local mac protecting user accounting is zeroed out. The flag is then cleared when user accounting of the raw received snapshot is calculated. Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: George Amanakis <[email protected]> Closes #12981 Closes #10523 Closes #11221 Closes #11294 Closes #12594 Issue #11300
* Linux <4.8 compat: submit_bio() rw argFinix19792022-02-041-1/+1
| | | | | | | | | | When using the two argument version of submit_bio() in kernel's prior to 4.8 the first argument should be specified. It's used by block dump to report the bio direction. Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Finix Yan <[email protected]> Closes #13006
* Linux 5.17 compat: PDE_DATA() renamed to pde_data()наб2022-02-043-9/+11
| | | | | | | | | | | | | Upstream commit 359745d78351c6f5442435f81549f0207ece28aa ("proc: remove PDE_DATA() completely") Link: https://lore.kernel.org/all/[email protected]/T/#u Reviewed-by: Tony Hutter <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ahelenia Ziemiańska <[email protected]> Closes #13004 Closes #12989
* Linux 5.17 compat: dequeue_signal() takes a 4th argumentнаб2022-02-042-2/+48
| | | | | | | | | | | | | Linux 5.17's dequeue_signal() takes an additional enum pid_type * output argument Upstream commit 5768d8906bc23d512b1a736c1e198aa833a6daa4 ("signal: Requeue signals in the appropriate queue") Reviewed-by: Tony Hutter <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ahelenia Ziemiańska <[email protected]> Closes #12989
* Linux 5.17 compat: detect complete_and_exit() renameнаб2022-02-043-1/+30
| | | | | | | | | | | | | Linux 5.17 sees a rename from complete_and_exit() to kthread complete_and_exit() Upstream commit cead18552660702a4a46f58e65188fe5f36e9dfe ("exit: Rename complete_and_exit to kthread_complete_and_exit") Reviewed-by: Tony Hutter <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ahelenia Ziemiańska <[email protected]> Closes #12989
* Add support for FALLOC_FL_ZERO_RANGERich Ercolani2022-02-042-2/+24
| | | | | | | | | | For us, I think it's always just FALLOC_FL_PUNCH_HOLE with a fake mustache on. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Coleman Kane <[email protected]> Signed-off-by: Rich Ercolani <[email protected]> Closes #12975
* Linux 5.16 compat: Added mapping for iov_iter_fault_in_readableRich Ercolani2022-02-043-3/+24
| | | | | | | | | | Linux decided to rename this for some reason. At some point, we should probably invert this mapping, but for now... Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Coleman Kane <[email protected]> Signed-off-by: Rich Ercolani <[email protected]> Closes #12975
* Linux 5.16 compat: Added add_disk check for returnRich Ercolani2022-02-043-0/+33
| | | | | | | | | add_disk went from void to must-check int return. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Coleman Kane <[email protected]> Signed-off-by: Rich Ercolani <[email protected]> Closes #12975
* Linux 5.16 compat: Check slab.h for kvmallocRich Ercolani2022-02-041-0/+1
| | | | | | | | | As it says on the tin - the folio work moved a bunch out of mm.h. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Coleman Kane <[email protected]> Signed-off-by: Rich Ercolani <[email protected]> Closes #12975
* Fix handling of errors from dmu_write_uio_dbuf() on FreeBSDMark Johnston2022-02-031-4/+11
| | | | | | | | | | | | | | | | | | | | | FreeBSD's implementation of zfs_uio_fault_move() returns EFAULT when a page fault occurs while copying data in or out of user buffers. The VFS treats such errors specially and will retry the I/O operation (which may have made some partial progress). When the FreeBSD and Linux implementations of zfs_write() were merged, the handling of errors from dmu_write_uio_dbuf() changed such that EFAULT is not handled as a partial write. For example, when appending to a file, the z_size field of the znode is not updated after a partial write resulting in EFAULT. Restore the old handling of errors from dmu_write_uio_dbuf() to fix this. This should have no impact on Linux, which has special handling for EFAULT already. Reviewed-by: Andriy Gapon <[email protected]> Reviewed-by: Ryan Moeller <[email protected]> Signed-off-by: Mark Johnston <[email protected]> Closes #12964
* Avoid memory allocations in the ARC eviction threadMark Johnston2022-02-033-53/+92
| | | | | | | | | | | | | | | | | | | | | | | | | | | When the eviction thread goes to shrink an ARC state, it allocates a set of marker buffers used to hold its place in the state's sublists. This can be problematic in low memory conditions, since 1) the allocation can be substantial, as we allocate NCPU markers; 2) on at least FreeBSD, page reclamation can block in arc_wait_for_eviction() In particular, in stress tests it's possible to hit a deadlock on FreeBSD when the number of free pages is very low, wherein the system is waiting for the page daemon to reclaim memory, the page daemon is waiting for the ARC eviction thread to finish, and the ARC eviction thread is blocked waiting for more memory. Try to reduce the likelihood of such deadlocks by pre-allocating markers for the eviction thread at ARC initialization time. When evicting buffers from an ARC state, check to see if the current thread is the ARC eviction thread, and use the pre-allocated markers for that purpose rather than dynamically allocating them. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Alexander Motin <[email protected]> Reviewed-by: George Amanakis <[email protected]> Signed-off-by: Mark Johnston <[email protected]> Closes #12985
* libzfs_sendrecv: Fix leaked holds nvlistRyan Moeller2022-02-031-4/+3
| | | | | | | | | There is no need to allocate a holds nvlist. lzc_get_holds does that for us. Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #12967
* libzfs_sendrecv: Avoid extra avl_findRyan Moeller2022-02-031-2/+3
| | | | | | | | | avl_add does avl_find internally, then avl_insert. We're already doing the avl_find, so using avl_insert directly avoids repeating the search. Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #12967
* FreeBSD: Fix zvol_cdev_open lockingRyan Moeller2022-02-031-2/+2
| | | | | | | | | | | | First open locking changes were correctly applied to zvol_geom_open but incorrectly applied to zvol_cdev_open, causing spa_namespace_lock to be held indefinitely. Make the first open locking in zvol_cdev_open match zvol_geom_open. Reviewed-by: Alexander Motin <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #13016
* FreeBSD: Fix zvol_*_open() lockingRyan Moeller2022-02-031-31/+59
| | | | | | | | | | | | | | | | These are the changes for FreeBSD corresponding to the changes made for Linux in #12863, see that PR for details. Changes from #12863 are applied for zvol_geom_open and zvol_cdev_open on FreeBSD. This also adds a check for the zvol dying which we had in zvol_geom_open but was missing in zvol_cdev_open. The check causes the open to fail early with ENXIO when we are in the middle of changing volmode. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Alexander Motin <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #12934
* FreeBSD: Fix leaked strings in libspl mnttabRyan Moeller2022-02-031-9/+16
| | | | | | | | | | | | | | | | | | | | The FreeBSD implementations of various libspl functions for getting mounted device information were found to leak several strings which were being allocated in statfs2mnttab but never freed. The Solaris getmntany(3C) and related interfaces are expected to return strings residing in static buffers that need to be copied rather than freed by the caller. Use static thread-local storage to stash the mnttab structure strings from FreeBSD's statfs info rather than strings allocated on the heap by strdup(3). While here, remove some stray commented out lines. Reviewed-by: Alexander Motin <[email protected]> Reviewed-by: Rich Ercolani <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #12961
* linux: libzfs: mount: fix uninitialised flagsнаб2022-02-031-1/+1
| | | | | | | | | | They're later |=d with constants, but never reset Caught by valgrind while investigating https://github.com/openzfs/zfs/pull/12928#issuecomment-1007496550 Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ahelenia Ziemiańska <[email protected]> Closes #12954
* FreeBSD: vfsops: use setgen for error caseнаб2022-02-031-1/+7
| | | | | | | | | Fix from https://github.com/openzfs/zfs/pull/12844#discussion_r774179413 Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Alexander Motin <[email protected]> Reviewed-by: Ryan Moeller <[email protected]> Signed-off-by: Ahelenia Ziemiańska <[email protected]> Closes #12905
* zfs_prune: reset sc.nr_to_scanchrisrd2022-02-031-0/+5
| | | | | | | | | | | | | | | | sc.nr_to_scan is an input to super_cache_clean (via shrinker->scan_objects), used to set the number of objects to scan in the various caches. However super_cache_scan also modifies sc.nr_to_scan, so when used in a loop we need to reset sc.nr_to_scan back to our desired nr_to_scan for the next iteration. Issue discovered and solution suggested by Tenzin Lhakhang @tlhakhan. Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Chris Dunlop <[email protected]> Issue #12433 Closes #12908
* Verify dRAID empty sectorsBrian Behlendorf2022-02-035-10/+67
| | | | | | | | | | | | | | | | | | | | | | | | | | | Verify that all empty sectors are zero filled before using them to calculate parity. Failure to do so can result in incorrect parity columns being generated and written to disk if the contents of an empty sector are non-zero. This was possible because the checksum only protects the data portions of the buffer, not the empty sector padding. This issue has been addressed by updating raidz_parity_verify() to check that all dRAID empty sectors are zero filled. Any sectors which are non-zero will be fixed, repair IO issued, and a checksum error logged. They can then be safely used to verify the parity. This specific type of damage is unlikely to occur since it requires a disk to have silently returned bad data, for an empty sector, while performing a scrub. However, if a pool were to have been damaged in this way, scrubbing the pool with this change applied will repair both the empty sector and parity columns as long as the data checksum is valid. Checksum errors will be reported in the `zpool status` output for any repairs which are made. Reviewed-by: Tony Hutter <[email protected]> Reviewed-by: Mark Maybee <[email protected]> Reviewed-by: Brian Atkinson <[email protected]> Signed-off-by: Brian Behlendorf <[email protected]> Closes #12857
* FreeBSD: fix unpropagated errorнаб2022-02-031-0/+1
| | | | | | | | | | | When performing I/O on FreeBSD using a file based vdev ensure all errors encountered when reading/writing are propagated through the zio pipeline. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Alexander Motin <[email protected]> Reviewed-by: Ryan Moeller <[email protected]> Signed-off-by: Ahelenia Ziemiańska <[email protected]> Closes #12904
* FreeBSD: fix world build after 143476ce8Martin Matuška2022-02-031-1/+1
| | | | | | | | Do not redefine the fallthrough macro when building with libcpp. Reviewed-by: Ryan Moeller <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Martin Matuska <[email protected]> Closes #12880
* Fix error propagation from lzc_send_redactedPhilipp Riederer2022-02-031-1/+1
| | | | | | | | | | Any error from lzc_send_redacted is overwritten by the error of send_conclusion_record; skip writing the conclusion record if there was an earlier error. Reviewed-by: Paul Dagnelie <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Philipp Riederer <[email protected]> Closes #12766
* freebsd/libshare: nfs: don't send SIGHUP to all processesнаб2022-02-031-1/+7
| | | | | | | | | | | | | | pidfile_open() sets *pidptr to -1 if the process currently holding the lock is between pidfile_open() and pidfile_write(), the subsequent kill(mountdpid) would potentially SIGHUP all non-system processes except init: just sleep for half a millisecond and try again in that case Reviewed-by: Don Brady <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: John Kennedy <[email protected]> Signed-off-by: Ahelenia Ziemiańska <[email protected]> Closes #12067
* Fix zvol_open() lock inversionBrian Behlendorf2022-02-031-63/+58
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When restructuring the zvol_open() logic for the Linux 5.13 kernel a lock inversion was accidentally introduced. In the updated code the spa_namespace_lock is now taken before the zv_suspend_lock allowing the following scenario to occur: down_read <=== waiting for zv_suspend_lock zvol_open <=== holds spa_namespace_lock __blkdev_get blkdev_get_by_dev blkdev_open ... mutex_lock <== waiting for spa_namespace_lock spa_open_common spa_open dsl_pool_hold dmu_objset_hold_flags dmu_objset_hold dsl_prop_get dsl_prop_get_integer zvol_create_minor dmu_recv_end zfs_ioc_recv_impl <=== holds zv_suspend_lock via zvol_suspend() zfs_ioc_recv ... This commit resolves the issue by moving the acquisition of the spa_namespace_lock back to after the zv_suspend_lock which restores the original ordering. Additionally, as part of this change the error exit paths were simplified where possible. Reviewed-by: Tony Hutter <[email protected]> Reviewed-by: Rich Ercolani <[email protected]> Signed-off-by: Brian Behlendorf <[email protected]> Closes #12863
* FreeBSD: Update argument types for VOP_READDIRAlan Somers2022-02-031-4/+10
| | | | | | | | | | | | | A recent commit to FreeBSD changed the type of vop_readdir_args.a_cookies to a uint64_t**. There is no functional impact to ZFS because ZFS only uses 32-bit cookies, which will be zero-extended to 64-bits by the existing code. https://github.com/freebsd/freebsd-src/commit/b214fcceacad6b842545150664bd2695c1c2b34f Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Alexander Motin <[email protected]> Signed-off-by: Alan Somers <[email protected]> Closes #12874
* Reduce number of arc_prune threadsAlexander Motin2022-02-032-3/+17
| | | | | | | | | | | | | | | | On FreeBSD vnode reclamation is single-threaded, protected by single global lock. Linux seems to be able to use a thread per mount point, but at this time it creates more harm than good. Reduce number of threads to 1, adding tunable in case somebody wants to try more. Reviewed-by: Ryan Moeller <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Chris Dunlop <[email protected]> Reviewed-by: Ahelenia Ziemiańska <[email protected]> Signed-off-by: Alexander Motin <[email protected]> Closes #12896 Issue #9966
* FreeBSD: Provide correct file generation numberRyan Moeller2022-02-033-3/+2
| | | | | | | | | | | | va_seq was actually a thin veil over va_gen, so z_gen is a more appropriate value than z_seq to populate the field with. Drop the unnecessary compat obfuscation and provide the correct file generation number. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Alexander Motin <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #12851
* Tag zfs-2.1.2zfs-2.1.2Tony Hutter2021-12-131-1/+1
| | | | | | META file and changelog updated. Signed-off-by: Tony Hutter <[email protected]>
* zfs-dkms rpm: Fix scriptlets dependenciesTill Maas2021-12-131-0/+4
| | | | | | | | | | | To ensure that the necessary packages are available during the %post and %preun scriptlets, require them properly. Reviewed-by: Tony Hutter <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Till Maas <[email protected]> Closes #12822 Closes #12832
* FreeBSD: Add vop_standard_writecount_nomsyncRyan Moeller2021-12-132-0/+18
| | | | | | | | | | https://cgit.freebsd.org/src/commit?id=3ffcfa599e29686cf2b3c1a6087408c37acaed78 Reviewed-by: Tony Hutter <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Alexander Motin <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #12828
* FreeBSD: Catch up with more VFS changesRyan Moeller2021-12-131-0/+18
| | | | | | | | | | | | Unused thread argument was removed from NDINIT* https://cgit.freebsd.org/src/commit?id=7e1d3eefd410ca0fbae5a217422821244c3eeee4 Reviewed-by: Tony Hutter <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Alexander Motin <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #12828
* Fix several bugs in the FreeBSD rename VOP implementationMark Johnston2021-12-131-136/+134
| | | | | | | | | | | | | | | | | | | - To avoid a use-after-free, zfsvfs->z_log needs to be loaded after the teardown lock is acquired with ZFS_ENTER(). - Avoid leaking vnode locks in zfs_rename_relock() and zfs_rename_() when the ZFS_ENTER() macros forces an early return. Refactor the rename implementation so that ZFS_ENTER() can be used safely. As a bonus, this lets us use the ZFS_VERIFY_ZP() macro instead of open-coding its implementation. Reported-by: Peter Holm <[email protected]> Tested-by: Peter Holm <[email protected]> Reviewed-by: Ryan Moeller <[email protected]> Reviewed-by: Tony Nguyen <[email protected]> Signed-off-by: Mark Johnston <[email protected]> Sponsored-by: The FreeBSD Foundation Closes #12717
* Remove (now unused) td argument from zfs_lookup()Pawel Jakub Dawidek2021-12-131-11/+10
| | | | | | | Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Ryan Moeller <[email protected]> Signed-off-by: Pawel Jakub Dawidek <[email protected]> Closes #12748
* Exit the teardown section later in rename on FreeBSDMark Johnston2021-12-131-3/+4
| | | | | | | | | | | | We have to hold the teardown lock while dereferencing zfsvfs->z_os and, I believe, when committing to the ZIL. Note that jumping to the "out" label, "error" is always non-zero. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Ryan Moeller <[email protected]> Signed-off-by: Mark Johnston <[email protected]> Closes #12704
* Fix potential use-after-frees in FreeBSD getpages and setattr VOPsMark Johnston2021-12-131-4/+4
| | | | | | | | | | | The objset object is reallocated during certain dataset operations, such as rollbacks, so the objset pointer must be loaded after acquiring the teardown lock. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Ryan Moeller <[email protected]> Signed-off-by: Mark Johnston <[email protected]> Closes #12704
* ZTS: import_rewind_device_replaced reliably failsBrian Behlendorf2021-12-081-2/+2
| | | | | | | | | | | | | | | The import_rewind_device_replaced.ksh test was never entirely reliable because it depends on MOS data not being overwritten. The MOS data is not protected by the snapshot so occasional failures were always expected. However, this test is now failing reliably on all platforms indicating something has changed in the code since the test was marked "maybe". Convert the test to a "known" failure until the root cause is identified and resolved. Reviewed-by: John Kennedy <[email protected]> Reviewed-by: George Melikov <[email protected]> Signed-off-by: Brian Behlendorf <[email protected]> Closes #12821
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-29 23:47:16 +0000