summaryrefslogtreecommitdiffstats
path: root/module/os
Commit message (Collapse)AuthorAgeFilesLines
* dmu_objset_from_ds must be called with dp_config_rwlock heldMatthew Ahrens2020-03-121-0/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The normal lock order is that the dp_config_rwlock must be held before the ds_opening_lock. For example, dmu_objset_hold() does this. However, dmu_objset_open_impl() is called with the ds_opening_lock held, and if the dp_config_rwlock is not already held, it will attempt to acquire it. This may lead to deadlock, since the lock order is reversed. Looking at all the callers of dmu_objset_open_impl() (which is principally the callers of dmu_objset_from_ds()), almost all callers already have the dp_config_rwlock. However, there are a few places in the send and receive code paths that do not. For example: dsl_crypto_populate_key_nvlist, send_cb, dmu_recv_stream, receive_write_byref, redact_traverse_thread. This commit resolves the problem by requiring all callers ot dmu_objset_from_ds() to hold the dp_config_rwlock. In most cases, the code has been restructured such that we call dmu_objset_from_ds() earlier on in the send and receive processes, when we already have the dp_config_rwlock, and save the objset_t until we need it in the middle of the send or receive (similar to what we already do with the dsl_dataset_t). Thus we do not need to acquire the dp_config_rwlock in many new places. I also cleaned up code in dmu_redact_snap() and send_traverse_thread(). Reviewed-by: Paul Dagnelie <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Paul Zuchowski <[email protected]> Signed-off-by: Matthew Ahrens <[email protected]> Closes #9662 Closes #10115
* Improve performance of zio_taskq_memberMatthew Ahrens2020-03-031-0/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | __zio_execute() calls zio_taskq_member() to determine if we are running in a zio interrupt taskq, in which case we may need to switch to processing this zio in a zio issue taskq. The call to zio_taskq_member() can become a performance bottleneck when we are processing a high rate of zio's. zio_taskq_member() calls taskq_member() on each of the zio interrupt taskqs, of which there are 21. This is slow because each call to taskq_member() does tsd_get(taskq_tsd), which on Linux is relatively slow. This commit improves the performance of zio_taskq_member() by having it cache the value of tsd_get(taskq_tsd), reducing the number of those calls to 1/21th of the current behavior. In a test case running `zfs send -c >/dev/null` of a filesystem with small blocks (average 2.5KB/block), zio_taskq_member() was using 6.7% of one CPU, and with this change it is reduced to 1.3%. Overall time to perform the `zfs send` reduced by 10% (~150,000 block/sec to ~165,000 blocks/sec). Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Serapheim Dimitropoulos <[email protected]> Reviewed-by: Ryan Moeller <[email protected]> Reviewed-by: Tony Nguyen <[email protected]> Signed-off-by: Matthew Ahrens <[email protected]> Closes #10070
* Make spa_history_zone platform-dependent in kernelRyan Moeller2020-03-021-0/+6
| | | | | | | | | | | This function should only return "linux" on Linux. Move the kernel part of the function out of common code. Fix the tests for FreeBSD. Reviewed-by: George Melikov <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #10079
* Re-share zfsdev_getminor and zfs_onexit_fd_holdMatthew Macy2020-02-284-100/+13
| | | | | | | | | | By adding a zfs_file_private accessor to the common interfaces and some extensions to FreeBSD platform code it is now possible to share the implementations for the aforementioned functions. Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Matt Macy <[email protected]> Closes #10073
* Linux 5.5 compat: blkg_tryget()Brian Behlendorf2020-02-281-1/+31
| | | | | | | | | | | | | Commit https://github.com/torvalds/linux/commit/9e8d42a0f accidentally converted the static inline function blkg_tryget() to GPL-only for kernels built with CONFIG_PREEMPT_RCU=y and CONFIG_BLK_CGROUP=y. Resolve the build issue by providing our own equivalent functionality when needed which uses rcu_read_lock_sched() internally as before. Reviewed-by: Tony Hutter <[email protected]> Signed-off-by: Brian Behlendorf <[email protected]> Closes #9745 Closes #10072
* Linux 5.6 compat: time_tBrian Behlendorf2020-02-271-1/+1
| | | | | | | | | | | | | | | | | | | As part of the Linux kernel's y2038 changes the time_t type has been fully retired. Callers are now required to use the time64_t type. Rather than move to the new type, I've removed the few remaining places where a time_t is used in the kernel code. They've been replaced with a uint64_t which is already how ZFS internally handled these values. Going forward we should work towards updating the remaining user space time_t consumers to the 64-bit interfaces. Reviewed-by: Matthew Macy <[email protected]> Reviewed-by: Tony Hutter <[email protected]> Signed-off-by: Brian Behlendorf <[email protected]> Closes #10052 Closes #10064
* Refactor dnode dirty context from dbuf_dirtyMatthew Macy2020-02-261-1/+1
| | | | | | | | | | | * Add dedicated donde_set_dirtyctx routine. * Add empty dirty record on destroy assertion. * Make much more extensive use of the SET_ERROR macro. Reviewed-by: Will Andrews <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Matthew Ahrens <[email protected]> Signed-off-by: Matt Macy <[email protected]> Closes #9924
* Remove zfs_getattr and convoff dead codeDirkjan Bussink2020-02-241-266/+0
| | | | | | | | | | | | The `convoff` function is called only in one code path in `zfs_space`. Each caller of `zfs_space` is called with a `flock64_t` that has `l_whence` set to `SEEK_SET`. This means that `convoff` always results in a no-op as the `bfp` parameter has `l_whence` set to `SEEK_SET` and `int whence` is `SEEK_SET` as well. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Ryan Moeller <[email protected]> Signed-off-by: Dirkjan Bussink <[email protected]> Closes #10006
* Honour sync=disabled when relinking tpmfilesDeHackEd2020-02-161-1/+1
| | | | | | | | | | | Unlinked files don't respect synchronous flush commands, but when they get relinked their state is unknown. Previously we force flushed all such files even when sync=disabled. Correct this case. Reviewed-by: Chunwei Chen <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: George Melikov <[email protected]> Signed-off-by: DHE <[email protected]> Closes #10005
* Missed wakeup when growing kmem cacheMatthew Ahrens2020-02-131-3/+6
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When growing the size of a (VMEM or KVMEM) kmem cache, spl_cache_grow() always does taskq_dispatch(spl_cache_grow_work), and then waits for the KMC_BIT_GROWING to be cleared by the taskq thread. The taskq thread (spl_cache_grow_work()) does: 1. allocate new slab and add to list 2. wake_up_all(skc_waitq) 3. clear_bit(KMC_BIT_GROWING) Therefore, the waiting thread can wake up before GROWING has been cleared. It will see that the growing has not yet completed, and go back to sleep until it hits the 100ms timeout. This can have an extreme performance impact on workloads that alloc/free more than fits in the (statically-sized) magazines. These workloads allocate and free slabs with high frequency. The problem can be observed with `funclatency spl_cache_grow`, which on some workloads shows that 99.5% of the time it takes <64us to allocate slabs, but we spend ~70% of our time in outliers, waiting for the 100ms timeout. The fix is to do `clear_bit(KMC_BIT_GROWING)` before `wake_up_all(skc_waitq)`. A future investigation should evaluate if we still actually need to taskq_dispatch() at all, and if so on which kernel versions. Reviewed-by: Paul Dagnelie <[email protected]> Reviewed-by: Pavel Zakharov <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: George Wilson <[email protected]> Signed-off-by: Matthew Ahrens <[email protected]> Closes #9989
* Share some code for spa deadman tunablesRyan Moeller2020-02-101-16/+2
| | | | | | | | | | | | | We need to do the same thing to update all spas on any OS for these tunables, so let's share the code. While here let's match the types of the literals initializing the variables with the type of the variable. Reviewed-by: Allan Jude <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Olaf Faaland <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #9964
* Linux 5.6 compat: timestamp_truncate()Brian Behlendorf2020-02-072-8/+6
| | | | | | | | | | | | The timestamp_truncate() function was added, it replaces the existing timespec64_trunc() function. This change renames our wrapper function to be consistent with the upstream name and updates the compatibility code for older kernels accordingly. Reviewed-by: Tony Hutter <[email protected]> Signed-off-by: Brian Behlendorf <[email protected]> Closes #9956 Closes #9961
* Linux 5.6 compat: struct proc_opsBrian Behlendorf2020-02-073-11/+47
| | | | | | | | | | | | | | The proc_ops structure was introduced to replace the use of of the file_operations structure when registering proc handlers. This change creates a new kstat_proc_op_t typedef for compatibility which can be used to pass around the correct structure. This change additionally adds the 'const' keyword to all of the existing proc operations structures. Reviewed-by: Tony Hutter <[email protected]> Signed-off-by: Brian Behlendorf <[email protected]> Closes #9961
* Replace static per-cpu with dynamic per-cpu dataRomain Dolbeau2020-02-061-4/+14
| | | | | | | This solves the issue of loading the spl module on RISC-V. Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Romain Dolbeau <[email protected]> Closes #9942
* Prepare ks_data before calling kstat_install()Alexander Motin2020-02-041-3/+2
| | | | | | | | | | | | It violated sequence described in kstat.h, and at least on FreeBSD kstat_install() uses provided names to create the sysctls. If the names are not available at the time, it ends up bad. Reviewed-by: Igor Kozhukhov <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Ryan Moeller <[email protected]> Signed-off-by: Alexander Motin <[email protected]> Sponsored-By: iXsystems, Inc. Closes #9933
* async zvol minor node creation interferes with receiveMatthew Ahrens2020-02-031-3/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | When we finish a zfs receive, dmu_recv_end_sync() calls zvol_create_minors(async=TRUE). This kicks off some other threads that create the minor device nodes (in /dev/zvol/poolname/...). These async threads call zvol_prefetch_minors_impl() and zvol_create_minor(), which both call dmu_objset_own(), which puts a "long hold" on the dataset. Since the zvol minor node creation is asynchronous, this can happen after the `ZFS_IOC_RECV[_NEW]` ioctl and `zfs receive` process have completed. After the first receive ioctl has completed, userland may attempt to do another receive into the same dataset (e.g. the next incremental stream). This second receive and the asynchronous minor node creation can interfere with one another in several different ways, because they both require exclusive access to the dataset: 1. When the second receive is finishing up, dmu_recv_end_check() does dsl_dataset_handoff_check(), which can fail with EBUSY if the async minor node creation already has a "long hold" on this dataset. This causes the 2nd receive to fail. 2. The async udev rule can fail if zvol_id and/or systemd-udevd try to open the device while the the second receive's async attempt at minor node creation owns the dataset (via zvol_prefetch_minors_impl). This causes the minor node (/dev/zd*) to exist, but the udev-generated /dev/zvol/... to not exist. 3. The async minor node creation can silently fail with EBUSY if the first receive's zvol_create_minor() trys to own the dataset while the second receive's zvol_prefetch_minors_impl already owns the dataset. To address these problems, this change synchronously creates the minor node. To avoid the lock ordering problems that the asynchrony was introduced to fix (see #3681), we create the minor nodes from open context, with no locks held, rather than from syncing contex as was originally done. Implementation notes: We generally do not need to traverse children or prefetch anything (e.g. when running the recv, snapshot, create, or clone subcommands of zfs). We only need recursion when importing/opening a pool and when loading encryption keys. The existing recursive, asynchronous, prefetching code is preserved for use in these cases. Channel programs may need to create zvol minor nodes, when creating a snapshot of a zvol with the snapdev property set. We figure out what snapshots are created when running the LUA program in syncing context. In this case we need to remember what snapshots were created, and then try to create their minor nodes from open context, after the LUA code has completed. There are additional zvol use cases that asynchronously own the dataset, which can cause similar problems. E.g. changing the volmode or snapdev properties. These are less problematic because they are not recursive and don't touch datasets that are not involved in the operation, there is still potential for interference with subsequent operations. In the future, these cases should be similarly converted to create the zvol minor node synchronously from open context. The async tasks of removing and renaming minors do not own the objset, so they do not have this problem. However, it may make sense to also convert these operations to happen synchronously from open context, in the future. Reviewed-by: Paul Dagnelie <[email protected]> Reviewed-by: Prakash Surya <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Matthew Ahrens <[email protected]> External-issue: DLPX-65948 Closes #7863 Closes #9885
* Re-consolidate zio_delay_interruptMatthew Macy2020-01-212-104/+0
| | | | | | | | With recent SPL changes there is no longer any need for a per platform version. Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Matt Macy <[email protected]> Closes #9860
* Unify target_cpu handlingBrian Behlendorf2020-01-171-3/+1
| | | | | | | | | | | | | Over the years several slightly different approaches were used in the Makefiles to determine the target architecture. This change updates both the build system and Makefile to handle this in a consistent fashion. TARGET_CPU is set to i386, x86_64, powerpc, aarch6 or sparc64 and made available in the Makefiles to be used as appropriate. Reviewed-by: Ryan Moeller <[email protected]> Signed-off-by: Brian Behlendorf <[email protected]> Closes #9848
* KMC_KVMEM disrupts kv_alloc() memory alignment expectationsloli10K2020-01-141-20/+2
| | | | | | | | | | | | | | | | | | | | | On kernels with KASAN enabled the following failure can be observed as soon as the zfs module is loaded: VERIFY(IS_P2ALIGNED(ptr, PAGE_SIZE)) failed PANIC at spl-kmem-cache.c:228:kv_alloc() The problem is kmalloc() has never guaranteed aligned allocations; this requirement resulted in zfsonlinux/spl@8b45dda which removed all kmalloc() usage in kv_alloc(). Until a GFP_ALIGNED flag (or equivalent functionality) is provided by the kernel this commit partially reverts 66955885 and 6d948c35 to prevent k(v)malloc() allocations in kv_alloc(). Reviewed-by: Kjeld Schouten <[email protected]> Reviewed-by: Michael Niewöhner <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: loli10K <[email protected]> Closes #9813
* Change http://zfsonlinux.org links to https://zfsonlinux.orgBrian Behlendorf2020-01-131-1/+1
| | | | | | | | | | | Update the project website links contained in to repository to reference the secure https://zfsonlinux.org address. Reviewed-By: Richard Laager <[email protected]> Reviewed-by: George Melikov <[email protected]> Reviewed-by: Garrett Fields <[email protected]> Reviewed-by: Kjeld Schouten <[email protected]> Signed-off-by: Brian Behlendorf <[email protected]> Closes #9837
* Fix QAT allocation failure return valueBrian Behlendorf2020-01-061-15/+18
| | | | | | | | | | | | | | | When qat_compress() fails to allocate the required contiguous memory it mistakenly returns success. This prevents the fallback software compression from taking over and (un)compressing the block. Resolve the issue by correctly setting the local 'status' variable on all exit paths. Furthermore, initialize it to CPA_STATUS_FAIL to ensure qat_compress() always fails safe to guard against any similar bugs in the future. Reviewed-by: Tony Hutter <[email protected]> Signed-off-by: Brian Behlendorf <[email protected]> Closes #9784 Closes #9788
* cppcheck: (error) Shifting signed 64-bit value by 63 bitsUbuntu2019-12-181-0/+2
| | | | | | | | | As of cppcheck 1.82 surpress the warning regarding shifting too many bits for __divdi3() implemention. The algorithm used here is correct. Reviewed-by: Tony Hutter <[email protected]> Signed-off-by: Brian Behlendorf <[email protected]> Closes #9732
* cppcheck: (error) Uninitialized variableUbuntu2019-12-188-22/+23
| | | | | | | | | | As of cppcheck 1.82 warnings are issued when using the list_for_each_* functions with an uninitialized variable. Functionally, this is fine but to resolve the warning initialize these variables. Reviewed-by: Tony Hutter <[email protected]> Signed-off-by: Brian Behlendorf <[email protected]> Closes #9732
* Don't fail to apply umask for O_TMPFILE filesTomohiro Kusumi2019-12-131-0/+6
| | | | | | | | | | | | | Apply umask to `mode` which will eventually be applied to inode. This is needed since VFS doesn't apply umask for O_TMPFILE files. (Note that zpl_init_acl() applies `ip->i_mode &= ~current_umask();` only when POSIX ACL is used.) Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Tony Hutter <[email protected]> Signed-off-by: Tomohiro Kusumi <[email protected]> Closes #8997 Closes #8998
* Make zfs_replay.c work on FreeBSDMatthew Macy2019-12-132-0/+42
| | | | | | | | | | | | | FreeBSD's vfs currently doesn't permit file systems to do their own locking. To avoid having to have duplicate zfs functions with and without locking add locking here. With luck these changes can be removed in the future. Reviewed-by: Sean Eric Fagan <[email protected]> Reviewed-by: Jorgen Lundman <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Matt Macy <[email protected]> Closes #9715
* Relocate common quota functions to shared codeRyan Moeller2019-12-113-455/+5
| | | | | | | | | | | The quota functions are common to all implementations and can be moved to common code. As a simplification they were moved to the Linux platform code in the initial refactoring. Reviewed-by: Jorgen Lundman <[email protected]> Reviewed-by: Igor Kozhukhov <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Ryan Moeller <[email protected]> Closes #9710
* Eliminate Linux specific inode usage from common code Matthew Macy2019-12-1111-263/+272
| | | | | | | | | | Change many of the znops routines to take a znode rather than an inode so that zfs_replay code can be largely shared and in the future the much of the znops code may be shared. Reviewed-by: Jorgen Lundman <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Matt Macy <[email protected]> Closes #9708
* Abstract away platform specific superblock referencesMatthew Macy2019-12-101-0/+12
| | | | | | | | The zfsvfs->z_sb field is Linux specified and should be abstracted. Reviewed-by: Richard Laager <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Matt Macy <[email protected]> Closes #9697
* ZTS: Fix zpool_reopen_001_posBrian Behlendorf2019-12-091-10/+25
| | | | | | | | | | | | Update the vdev_disk_open() retry logic to use a specified number of milliseconds to be more robust. Additionally, on failure log both the time waited and requested timeout to the internal log. The default maximum allowed open retry time has been increased from 500ms to 1000ms. Reviewed-by: Kjeld Schouten <[email protected]> Signed-off-by: Brian Behlendorf <[email protected]> Closes #9680
* Refactor deadman set failmode to be cross platformMatthew Macy2019-12-051-0/+13
| | | | | | | | | Update zfs_deadman_failmode to use the ZFS_MODULE_PARAM_CALL wrapper, and split the common and platform specific portions. Reviewed-by: Jorgen Lundman <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Matt Macy <[email protected]> Closes #9670
* Replace ASSERTV macro with compiler annotationMatthew Macy2019-12-051-3/+3
| | | | | | | | | | | Remove the ASSERTV macro and handle suppressing unused compiler warnings for variables only in ASSERTs using the __attribute__((unused)) compiler annotation. The annotation is understood by both gcc and clang. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Jorgen Lundman <[email protected]> Signed-off-by: Matt Macy <[email protected]> Closes #9671
* Move zfs_cmd_t copyin/copyout to platform codeMatthew Macy2019-12-021-1/+17
| | | | | | | | | | | FreeBSD needs to cope with multiple version of the zfs_cmd_t structure. Allowing the platform code to pre and post process the cmd structure makes it possible to work with legacy tooling. Reviewed-by: Jorgen Lundman <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Matt Macy <[email protected]> Closes #9624
* Mark Linux fallocate extensions as specific to LinuxMatthew Macy2019-11-301-3/+5
| | | | | | | | | fallocate(2) is a Linux-specific system call which in unavailable on other platforms. Reviewed-by: Jorgen Lundman <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Matt Macy <[email protected]> Closes #9633
* Resolve ZoF differences in zfs_ioctl.hMatthew Macy2019-11-301-1/+0
| | | | | | | | | | FreeBSD needs to be able to pass the jail id to the jail/unjail ioctls and the struct file in the device structure is unused. Reviewed-by: Kjeld Schouten <[email protected]> Reviewed-by: Jorgen Lundman <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Matt Macy <[email protected]> Closes #9625
* Remove zfs_vdev_elevator module optionBrian Behlendorf2019-11-271-111/+19
| | | | | | | | | | | | | As described in commit f81d5ef6 the zfs_vdev_elevator module option is being removed. Users who require this functionality should update their systems to set the disk scheduler using a udev rule. Reviewed-by: Richard Laager <[email protected]> Reviewed-by: loli10K <[email protected]> Signed-off-by: Brian Behlendorf <[email protected]> Issue #8664 Closes #9417 Closes #9609
* Check for unlinked znodes after igrab()Mauricio Faria de Oliveira2019-11-211-5/+7
| | | | | | | | | | | | | | | | | | The changes in commit 41e1aa2a / PR #9583 introduced a regression on tmpfile_001_pos: fsetxattr() on a O_TMPFILE file descriptor started to fail with errno ENODATA: openat(AT_FDCWD, "/test", O_RDWR|O_TMPFILE, 0666) = 3 <...> fsetxattr(3, "user.test", <...>, 64, 0) = -1 ENODATA The originally proposed change on PR #9583 is not susceptible to it, so just move the code/if-checks around back in that way, to fix it. Reviewed-by: Pavel Snajdr <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Original-patch-by: Heitor Alves de Siqueira <[email protected]> Signed-off-by: Mauricio Faria de Oliveira <[email protected]> Closes #9602
* Add zfs_file_* interface, remove vnodesMatthew Macy2019-11-2118-876/+575
| | | | | | | | | | | | | | | | | | Provide a common zfs_file_* interface which can be implemented on all platforms to perform normal file access from either the kernel module or the libzpool library. This allows all non-portable vnode_t usage in the common code to be replaced by the new portable zfs_file_t. The associated vnode and kobj compatibility functions, types, and macros have been removed from the SPL. Moving forward, vnodes should only be used in platform specific code when provided by the native operating system. Reviewed-by: Sean Eric Fagan <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Igor Kozhukhov <[email protected]> Reviewed-by: Jorgen Lundman <[email protected]> Signed-off-by: Matt Macy <[email protected]> Closes #9556
* Partially revert 5a6ac4cBrian Behlendorf2019-11-182-0/+45
| | | | | | | | | | | | | | | | | | | | | | | | Reinstate the zpl_revalidate() functionality to resolve a regression where dentries for open files during a rollback are not invalidated. The unrelated functionality for automatically unmounting .zfs/snapshots was not reverted. Nor was the addition of shrink_dcache_sb() to the zfs_resume_fs() function. This issue was not immediately caught by the CI because the test case intended to catch it was included in the list of ZTS tests which may occasionally fail for unrelated reasons. Remove all of the rollback tests from this list to help identify the frequency of any spurious failures. The rollback_003_pos.ksh test case exposes a real issue with the long standing code which needs to be investigated. Regardless, it has been enable with a small workaround in the test case itself. Reviewed-by: Matt Ahrens <[email protected]> Reviewed-by: Pavel Snajdr <[email protected]> Signed-off-by: Brian Behlendorf <[email protected]> Closes #9587 Closes #9592
* Break out of zfs_zget early if unlinked znodeHeitor Alves de Siqueira2019-11-151-9/+17
| | | | | | | | | | | If zp->z_unlinked is set, we're working with a znode that has been marked for deletion. If that's the case, we can skip the "goto again" loop and return ENOENT, as the znode should not be discovered. Reviewed-by: Richard Yao <[email protected]> Reviewed-by: Matt Ahrens <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Heitor Alves de Siqueira <[email protected]> Closes #9583
* Prevent NULL pointer dereference in blkg_tryget() on EL8 kernelsloli10K2019-11-131-1/+1
| | | | | | | | | | | | | | blkg_tryget() as shipped in EL8 kernels does not seem to handle NULL @blkg as input; this is different from its mainline counterpart where NULL is accepted. To prevent dereferencing a NULL pointer when dealing with block devices which do not set a root_blkg on the request queue perform the NULL check in vdev_bio_associate_blkg(). Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Kjeld Schouten <[email protected]> Reviewed-by: Tony Hutter <[email protected]> Signed-off-by: loli10K <[email protected]> Closes #9546 Closes #9577
* Check for __GFP_RECLAIM instead of GFP_KERNELMichael Niewöhner2019-11-131-3/+8
| | | | | | | | | | | Check for __GFP_RECLAIM instead of GFP_KERNEL because zfs modifies IO and FS flags which breaks the check for GFP_KERNEL. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Matt Ahrens <[email protected]> Reviewed-by: Sebastian Gottschall <[email protected]> Signed-off-by: Michael Niewöhner <[email protected]> Closes #9034
* Add kmem_cache flag for forcing kvmallocMichael Niewöhner2019-11-133-9/+38
| | | | | | | | | | | | | | This adds a new KMC_KVMEM flag was added to enforce use of the kvmalloc allocator in kmem_cache_create even for large blocks, which may also increase performance in some specific cases (e.g. zstd), too. Default to KVMEM instead of VMEM in spl_kmem_cache_create. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Matt Ahrens <[email protected]> Signed-off-by: Sebastian Gottschall <[email protected]> Signed-off-by: Michael Niewöhner <[email protected]> Closes #9034
* Make use of kvmalloc if available and fix vmem_alloc implementationMichael Niewöhner2019-11-132-11/+95
| | | | | | | | | | | | | | | | | This patch implements use of kvmalloc for GFP_KERNEL allocations, which may increase performance if the allocator is able to allocate physical memory, if kvmalloc is available as a public kernel interface (since v4.12). Otherwise it will simply fall back to virtual memory (vmalloc). Also fix vmem_alloc implementation which can lead to slow allocations since the first attempt with kmalloc does not make use of the noretry flag but tells the linux kernel to retry several times before it fails. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Matt Ahrens <[email protected]> Signed-off-by: Sebastian Gottschall <[email protected]> Signed-off-by: Michael Niewöhner <[email protected]> Closes #9034
* Add missing documentation for some KMC flagsMichael Niewöhner2019-11-131-5/+5
| | | | | | | | Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Matt Ahrens <[email protected]> Signed-off-by: Sebastian Gottschall <[email protected]> Signed-off-by: Michael Niewöhner <[email protected]> Closes #9034
* Linux compat: Minimum kernel version 3.10Brian Behlendorf2019-11-1217-541/+47
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Increase the minimum supported kernel version from 2.6.32 to 3.10. This removes support for the following Linux enterprise distributions. Distribution | Kernel | End of Life ---------------- | ------ | ------------- Ubuntu 12.04 LTS | 3.2 | Apr 28, 2017 SLES 11 | 3.0 | Mar 32, 2019 RHEL / CentOS 6 | 2.6.32 | Nov 30, 2020 The following changes were made as part of removing support. * Updated `configure` to enforce a minimum kernel version as specified in the META file (Linux-Minimum: 3.10). configure: error: *** Cannot build against kernel version 2.6.32. *** The minimum supported kernel version is 3.10. * Removed all `configure` kABI checks and matching C code for interfaces which solely predate the Linux 3.10 kernel. * Updated all `configure` kABI checks to fail when an interface is missing which was in the 3.10 kernel up to the latest 5.1 kernel. Removed the HAVE_* preprocessor defines for these checks and updated the code to unconditionally use the verified interface. * Inverted the detection logic in several kABI checks to match the new interface as it appears in 3.10 and newer and not the legacy interface. * Consolidated the following checks in to individual files. Due the large number of changes in the checks it made sense to handle this now. It would be desirable to group other related checks in the same fashion, but this as left as future work. - config/kernel-blkdev.m4 - Block device kABI checks - config/kernel-blk-queue.m4 - Block queue kABI checks - config/kernel-bio.m4 - Bio interface kABI checks * Removed the kABI checks for sops->nr_cached_objects() and sops->free_cached_objects(). These interfaces are currently unused. Signed-off-by: Brian Behlendorf <[email protected]> Closes #9566
* Remove zpl_revalidatePavel Snajdr2019-11-112-63/+30
| | | | | | | | | | | | | | | | | | | | | | | This patch removes the need for zpl_revalidate altogether. There were 3 main reasons why we used d_revalidate: 1. periodic automounted snapshots umount deferral 2. negative dentries created before snapshot rollback 3. stale inodes referenced by dentry cache after snapshot rollback Periodic snapshots deferral solution introduces zfs_exit_fs function, which is called as a part of ZFS_EXIT(zfsvfs_t) macro. Negative dentries and stale inodes are solved by flushing the dcache for the particular dataset on zfs_resume_fs call. This patch also removes now unused HAVE_S_D_OP configure test. Reviewed-by: Aleksa Sarai <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Pavel Snajdr <[email protected]> Closes #8774 Closes #9549
* Move platform specific parts of zfs_znode.h to platform codeMatthew Macy2019-11-061-2/+2
| | | | | | | | | Some of the znode fields are different and functions consuming an inode don't exist on FreeBSD. Reviewed-by: Brian Behlendorf <[email protected]> Reviewed-by: Jorgen Lundman <[email protected]> Signed-off-by: Matt Macy <[email protected]> Closes #9536
* Add tracepoints for taskq entry lifetime eventsPrakash Surya2019-11-012-4/+18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | This adds some new DTRACE_PROBE* endpoints so that we can observe taskq latencies on a system. Additionally, a new "taskqlatency.bt" script is added to do this observation via "bpftrace". Lastly, a "zfs-trace.sh" script is added to wrap "bpftrace" with the proper options required to run and use "taskqlatency.bt". For example, with these changes in place, a user can run the following: $ cd ./contrib/bpftrace $ sudo ./zfs-trace.sh taskqlatency.bt Attaching 6 probes... ^C Here's some example output, showing latency information for time spent executing the taskq entry's function: @exec_lat_us[dp_sync_taskq, userquota_updates_task]: [2, 4) 5 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@| [4, 8) 0 | | [8, 16) 1 |@@@@@@@@@@ | [16, 32) 2 |@@@@@@@@@@@@@@@@@@@@ | @exec_lat_us[z_wr_int_h, zio_execute]: [8, 16) 16 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@| [16, 32) 2 |@@@@@@ | @exec_lat_us[z_wr_iss_h, zio_execute]: [16, 32) 4 |@@@@@@@@@@@@@@@@ | [32, 64) 13 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@| [64, 128) 1 |@@@@ | @exec_lat_us[z_ioctl_int, zio_execute]: [2, 4) 1 |@@@@ | [4, 8) 11 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@| [8, 16) 8 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ | @exec_lat_us[dp_sync_taskq, sync_dnodes_task]: [2, 4) 1 |@@@@@@ | [4, 8) 7 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ | [8, 16) 8 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@| [16, 32) 2 |@@@@@@@@@@@@@ | [32, 64) 4 |@@@@@@@@@@@@@@@@@@@@@@@@@@ | [64, 128) 1 |@@@@@@ | [128, 256) 0 | | [256, 512) 1 |@@@@@@ Here's some example output, showing latency information for time spent waiting on the taskq, prior to starting execution of entry's function: @queue_lat_us[dp_sync_taskq]: [2, 4) 1 |@@@@ | [4, 8) 7 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ | [8, 16) 2 |@@@@@@@@ | [16, 32) 3 |@@@@@@@@@@@@@ | [32, 64) 12 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@| [64, 128) 6 |@@@@@@@@@@@@@@@@@@@@@@@@@@ | [128, 256) 0 | | [256, 512) 1 |@@@@ | @queue_lat_us[z_wr_iss]: [4, 8) 4 |@@@@ | [8, 16) 13 |@@@@@@@@@@@@@@@ | [16, 32) 6 |@@@@@@@ | [32, 64) 2 |@@ | [64, 128) 12 |@@@@@@@@@@@@@@ | [128, 256) 15 |@@@@@@@@@@@@@@@@@@ | [256, 512) 33 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ | [512, 1K) 27 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ | [1K, 2K) 7 |@@@@@@@@ | [2K, 4K) 14 |@@@@@@@@@@@@@@@@ | [4K, 8K) 14 |@@@@@@@@@@@@@@@@ | [8K, 16K) 23 |@@@@@@@@@@@@@@@@@@@@@@@@@@@ | [16K, 32K) 43 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@| @queue_lat_us[z_wr_int]: [2, 4) 10 |@@@@@ | [4, 8) 71 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ | [8, 16) 88 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@| [16, 32) 50 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ | [32, 64) 65 |@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@ | [64, 128) 43 |@@@@@@@@@@@@@@@@@@@@@@@@@ | [128, 256) 19 |@@@@@@@@@@@ | [256, 512) 3 |@ | [512, 1K) 1 | | Reviewed by: Brad Lewis <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Prakash Surya <[email protected]> Closes #9525
* Enable use of DTRACE_PROBE* macros in "spl" modulePrakash Surya2019-11-016-2/+35
| | | | | | | | | | | | | | | | | | | | | | | | | | | This change modifies some of the infrastructure for enabling the use of the DTRACE_PROBE* macros, such that we can use tehm in the "spl" module. Currently, when the DTRACE_PROBE* macros are used, they get expanded to create new functions, and these dynamically generated functions become part of the "zfs" module. Since the "spl" module does not depend on the "zfs" module, the use of DTRACE_PROBE* in the "spl" module would result in undefined symbols being used in the "spl" module. Specifically, DTRACE_PROBE* would turn into a function call, and the function being called would be a symbol only contained in the "zfs" module; which results in a linker and/or runtime error. Thus, this change adds the necessary logic to the "spl" module, to mirror the tracing functionality available to the "zfs" module. After this change, we'll have a "trace_zfs.h" header file which defines the probes available only to the "zfs" module, and a "trace_spl.h" header file which defines the probes available only to the "spl" module. Reviewed by: Brad Lewis <[email protected]> Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Prakash Surya <[email protected]> Closes #9525
* Wrap Linux module macrosMatthew Macy2019-11-012-9/+10
| | | | | | | | | | MODULE_VERSION is already defined on FreeBSD. Wrap all of the used MODULE_* macros for the sake of consistency and portability. Add a user space noop version to reduce the need for _KERNEL ifdefs. Reviewed-by: Brian Behlendorf <[email protected]> Signed-off-by: Matt Macy <[email protected]> Closes #9542
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-21 09:49:15 +0000