Changes in Kconfig

b261d222 - lib/crc: remove CONFIG_LIBCRC32C

Tue, 01 Apr 2025 22:16:00 +0000

lib/crc: remove CONFIG_LIBCRC32CNow that LIBCRC32C does nothing besides select CRC32, make every optionthat selects LIBCRC32C instead select CRC32 directly. Then removeLIBCRC32C.Reviewed-by: Christoph Hellwig Reviewed-by: "Martin K. Petersen" Acked-by: Ard Biesheuvel Link: https://lore.kernel.org/r/20250401221600.24878-8-ebiggers@kernel.orgSigned-off-by: Eric Biggers List of files: /linux-6.15/fs/xfs/Kconfig

150bb10a - xfs: verify buffer, inode, and dquot items every tx commit

Tue, 02 Jul 2024 18:23:23 +0000

xfs: verify buffer, inode, and dquot items every tx commitgeneric/388 has an annoying tendency to fail like this during logrecovery:XFS (sda4): Unmounting Filesystem 435fe39b-82b6-46ef-be56-819499585130XFS (sda4): Mounting V5 Filesystem 435fe39b-82b6-46ef-be56-819499585130XFS (sda4): Starting recovery (logdev: internal)00000000: 49 4e 81 b6 03 02 00 00 00 00 00 07 00 00 00 07 IN..............00000010: 00 00 00 01 00 00 00 00 00 00 00 00 00 00 00 10 ................00000020: 35 9a 8b c1 3e 6e 81 00 35 9a 8b c1 3f dc b7 00 5...>n..5...?...00000030: 35 9a 8b c1 3f dc b7 00 00 00 00 00 00 3c 86 4f 5...?........<.O00000040: 00 00 00 00 00 00 02 f3 00 00 00 00 00 00 00 00 ................00000050: 00 00 1f 01 00 00 00 00 00 00 00 02 b2 74 c9 0b .............t..00000060: ff ff ff ff d7 45 73 10 00 00 00 00 00 00 00 2d .....Es........-00000070: 00 00 07 92 00 01 fe 30 00 00 00 00 00 00 00 1a .......0........00000080: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................00000090: 35 9a 8b c1 3b 55 0c 00 00 00 00 00 04 27 b2 d1 5...;U.......'..000000a0: 43 5f e3 9b 82 b6 46 ef be 56 81 94 99 58 51 30 C_....F..V...XQ0XFS (sda4): Internal error Bad dinode after recovery at line 539 of file fs/xfs/xfs_inode_item_recover.c. Caller xlog_recover_items_pass2+0x4e/0xc0 [xfs]CPU: 0 PID: 2189311 Comm: mount Not tainted 6.9.0-rc4-djwx #rc4Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS ?-20171121_152543-x86-ol7-builder-01.us.oracle.com-4.el7.1 04/01/2014Call Trace: dump_stack_lvl+0x4f/0x60 xfs_corruption_error+0x90/0xa0 xlog_recover_inode_commit_pass2+0x5f1/0xb00 xlog_recover_items_pass2+0x4e/0xc0 xlog_recover_commit_trans+0x2db/0x350 xlog_recovery_process_trans+0xab/0xe0 xlog_recover_process_data+0xa7/0x130 xlog_do_recovery_pass+0x398/0x840 xlog_do_log_recovery+0x62/0xc0 xlog_do_recover+0x34/0x1d0 xlog_recover+0xe9/0x1a0 xfs_log_mount+0xff/0x260 xfs_mountfs+0x5d9/0xb60 xfs_fs_fill_super+0x76b/0xa30 get_tree_bdev+0x124/0x1d0 vfs_get_tree+0x17/0xa0 path_mount+0x72b/0xa90 __x64_sys_mount+0x112/0x150 do_syscall_64+0x49/0x100 entry_SYSCALL_64_after_hwframe+0x4b/0x53 XFS (sda4): Corruption detected. Unmount and run xfs_repairXFS (sda4): Metadata corruption detected at xfs_dinode_verify.part.0+0x739/0x920 [xfs], inode 0x427b2d1XFS (sda4): Filesystem has been shut down due to log error (0x2).XFS (sda4): Please unmount the filesystem and rectify the problem(s).XFS (sda4): log mount/recovery failed: error -117XFS (sda4): log mount failedThis inode log item recovery failing the dinode verifier afterreplaying the contents of the inode log item into the ondisk inode.Looking back into what the kernel was doing at the time of the fsshutdown, a thread was in the middle of running a series oftransactions, each of which committed changes to the inode.At some point in the middle of that chain, an invalid (at leastaccording to the verifier) change was committed. Had the filesystem notshut down in the middle of the chain, a subsequent transaction wouldhave corrected the invalid state and nobody would have noticed. Butthat's not what happened here. Instead, the invalid inode state wascommitted to the ondisk log, so log recovery tripped over it.The actual defect here was an overzealous inode verifier, which wasfixed in a separate patch. This patch adds some transaction precommitfunctions for CONFIG_XFS_DEBUG=y mode so that we can detect these kindsof transient errors at transaction commit time, where it's much easierto find the root cause.Signed-off-by: Darrick J. Wong Reviewed-by: Christoph Hellwig List of files: /linux-6.15/fs/xfs/Kconfig

a095686a - xfs: support in-memory btrees

Thu, 22 Feb 2024 20:43:35 +0000

xfs: support in-memory btreesAdapt the generic btree cursor code to be able to create a btree whosebuffers come from a (presumably in-memory) buftarg with a header blockthat's specific to in-memory btrees. We'll connect this to other partsof online scrub in the next patches.Note that in-memory btrees always have a block size matching the systemmemory page size for efficiency reasons. There are also a few things weneed to do to finalize a btree update; that's covered in the next patch.Signed-off-by: Darrick J. Wong Reviewed-by: Christoph Hellwig List of files: /linux-6.15/fs/xfs/Kconfig

5076a604 - xfs: support in-memory buffer cache targets

Thu, 22 Feb 2024 20:43:21 +0000

xfs: support in-memory buffer cache targetsAllow the buffer cache to target in-memory files by making it possibleto have a buftarg that maps pages from private shmem files. As theprevous patch alludes, the in-memory buftarg contains its own cache,points to a shmem file, and does not point to a block_device.The next few patches will make it possible to construct an xfs_btree inpageable memory by using this buftarg.Signed-off-by: Darrick J. Wong Reviewed-by: Christoph Hellwig List of files: /linux-6.15/fs/xfs/Kconfig

4e98cc90 - xfs: allow scrub to hook metadata updates in other writers

Thu, 22 Feb 2024 20:30:45 +0000

xfs: allow scrub to hook metadata updates in other writersCertain types of filesystem metadata can only be checked by scanningevery file in the entire filesystem. Specific examples of this includequota counts, file link counts, and reverse mappings of file extents.Directory and parent pointer reconstruction may also fall into thiscategory. File scanning is much trickier than scanning AG metadatabecause we have to take inode locks in the same order as the rest of[VX]FS, we can't be holding buffer locks when we do that, and scanningthe whole filesystem takes time.Earlier versions of the online repair patchset relied heavily onfsfreeze as a means to quiesce the filesystem so that we could takelocks in the proper order without worrying about concurrent updates fromother writers. Reviewers of those patches opined that freezing theentire fs to check and repair something was not sufficiently better thanunmounting to run fsck offline. I don't agree with that 100%, but themessage was clear: find a way to repair things that minimizes thequiet period where nobody can write to the filesystem.Generally, building btree indexes online can be split into two phases: acollection phase where we compute the records that will be put into thenew btree; and a construction phase, where we construct the physicalbtree blocks and persist them. While it's simple to hold resource locksfor the entirety of the two phases to ensure that the new index isconsistent with the rest of the system, we don't need to hold resourcelocks during the collection phase if we have a means to receive liveupdates of other work going on elsewhere in the system.The goal of this patch, then, is to enable online fsck to learn aboutmetadata updates going on in other threads while it constructs a shadowcopy of the metadata records to verify or correct the real metadata. Tominimize the overhead when online fsck isn't running, we use srcunotifiers because they prioritize fast access to the notifier call chain(particularly when the chain is empty) at a cost to configuringnotifiers. Online fsck should be relatively infrequent, so this isacceptable.The intended usage model is fairly simple. Code that modifies ametadata structure of interest should declare a xfs_hook_chain structurein some well defined place, and call xfs_hook_call whenever an updatehappens. Online fsck code should define a struct notifier_block and usexfs_hook_add to attach the block to the chain, along with a function tobe called. This function should synchronize with the fsck scanner toupdate whatever in-memory data the scanner is collecting. Whenfinished, xfs_hook_del removes the notifier from the list and waits forthem all to complete.Originally, I selected srcu notifiers over blocking notifiers toimplement live hooks because they seemed to have fewer impacts toscalability. The per-call cost of srcu_notifier_call_chain is higher(19ns) than blocking_notifier_ (4ns) in the single threaded case, butblocking notifiers use an rwsem to stabilize the list. Cachelinebouncing for that rwsem is costly to runtime code when there are a lotof CPUs running regular filesystem operations. If there are no hooksinstalled, this is a total waste of CPU time.Therefore, I stuck with srcu notifiers, despite trading off singlethreaded performance for multithreaded performance. I also wasn'tthrilled with the very high teardown time for srcu notifiers, since thecaller has to wait for the next rcu grace period. This can take a longtime if there are a lot of CPUs.Then I discovered the jump label implementation of static keys.Jump labels use kernel code patching to replace a branch with a nop sledwhen the key is disabled. IOWs, they can eliminate the overhead of_call_chain when there are no hooks enabled. This makes blockingnotifiers competitive again -- scrub runs faster because teardown of thechain is a lot cheaper, and runtime code only pays the rwsem lockingoverhead when scrub is actually running.With jump labels enabled, calls to empty notifier chains are elided fromthe call sites when there are no hooks registered, which means that theoverhead is 0.36ns when fsck is not running. This is perfect for mostof the architectures that XFS is expected to run on (e.g. x86, powerpc,arm64, s390x, riscv).For architectures that don't support jump labels (e.g. m68k) the runtimeoverhead of checking the static key is an atomic counter read. Thisisn't great, but it's still cheaper than taking a shared rwsem.Signed-off-by: Darrick J. Wong Reviewed-by: Christoph Hellwig List of files: /linux-6.15/fs/xfs/Kconfig

a2e4388a - xfs: fix again select in kconfig XFS_ONLINE_SCRUB_STATS

Sun, 05 Nov 2023 19:23:18 +0000

xfs: fix again select in kconfig XFS_ONLINE_SCRUB_STATSCommit 57c0f4a8ea3a attempted to fix the select in the kconfig entryXFS_ONLINE_SCRUB_STATS by selecting XFS_DEBUG, but the originalintention was to select DEBUG_FS, since the feature relies on debugfs toexport the related scrub statistics.Fixes: 57c0f4a8ea3a ("xfs: fix select in config XFS_ONLINE_SCRUB_STATS")Reported-by: Holger Hoffstätte Signed-off-by: Anthony Iliopoulos Reviewed-by: Dave Chinner Reviewed-by: "Darrick J. Wong" Signed-off-by: Chandan Babu R List of files: /linux-6.15/fs/xfs/Kconfig

57c0f4a8 - xfs: fix select in config XFS_ONLINE_SCRUB_STATS

Fri, 25 Aug 2023 12:05:13 +0000

xfs: fix select in config XFS_ONLINE_SCRUB_STATSCommit d7a74cad8f45 ("xfs: track usage statistics of online fsck")introduces config XFS_ONLINE_SCRUB_STATS, which selects the non-existingconfig FS_DEBUG. It is probably intended to select the existing configXFS_DEBUG.Fix the select in config XFS_ONLINE_SCRUB_STATS.Fixes: d7a74cad8f45 ("xfs: track usage statistics of online fsck")Signed-off-by: Lukas Bulwahn Reviewed-by: "Darrick J. Wong" Signed-off-by: Chandan Babu R List of files: /linux-6.15/fs/xfs/Kconfig

d7a74cad - xfs: track usage statistics of online fsck

Thu, 10 Aug 2023 14:48:07 +0000

xfs: track usage statistics of online fsckTrack the usage, outcomes, and run times of the online fsck code, andreport these values via debugfs. The columns in the file are: * scrubber name * number of scrub invocations * clean objects found * corruptions found * optimizations found * cross referencing failures * inconsistencies found during cross referencing * incomplete scrubs * warnings * number of time scrub had to retry * cumulative amount of time spent scrubbing (microseconds) * number of repair inovcations * successfully repaired objects * cumuluative amount of time spent repairing (microseconds)Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner List of files: /linux-6.15/fs/xfs/Kconfig

3934e8eb - xfs: create a big array data structure

Thu, 10 Aug 2023 14:48:04 +0000

xfs: create a big array data structureCreate a simple 'big array' data structure for storage of fixed-sizemetadata records that will be used to reconstruct a btree index. Forrepair operations, the most important operations are append, iterate,and sort.Earlier implementations of the big array used linked lists and sufferedfrom severe problems -- pinning all records in kernel memory was not agood idea and frequently lead to OOM situations; random access was veryinefficient; and record overhead for the lists was unacceptably high at40-60%.Therefore, the big memory array relies on the 'xfile' abstraction, whichcreates a memfd file and stores the records in page cache pages. Sincethe memfd is created in tmpfs, the memory pages can be pushed out todisk if necessary and we have a built-in usage limit of 50% of physicalmemory.Signed-off-by: Darrick J. Wong Reviewed-by: Kent Overstreet Reviewed-by: Dave Chinner List of files: /linux-6.15/fs/xfs/Kconfig

7ba83850 - xfs: deprecate the ascii-ci feature

Wed, 12 Apr 2023 02:05:19 +0000

xfs: deprecate the ascii-ci featureThis feature is a mess -- the hash function has been broken for theentire 15 years of its existence if you create names with extended asciibytes; metadump name obfuscation has silently failed for just as long;and the feature clashes horribly with the UTF8 encodings that mostsystems use today. There is exactly one fstest for this feature.In other words, this feature is crap. Let's deprecate it now so we canremove it from the codebase in 2030.Signed-off-by: Darrick J. Wong Reviewed-by: Christoph Hellwig List of files: /linux-6.15/fs/xfs/Kconfig

466c525d - xfs: minimize overhead of drain wakeups by using jump labels

Wed, 12 Apr 2023 01:59:59 +0000

xfs: minimize overhead of drain wakeups by using jump labelsTo reduce the runtime overhead even further when online fsck isn'trunning, use a static branch key to decide if we call wake_up on thedrain. For compilers that support jump labels, the call to wake_up isreplaced by a nop sled when nobody is waiting for intents to drain.From my initial microbenchmarking, every transition of the static keybetween the on and off states takes about 22000ns to complete; this ispaid entirely by the xfs_scrub process. When the static key is off(which it should be when fsck isn't running), the nop sled adds anoverhead of approximately 0.36ns to runtime code. The post-atomiclockless waiter check adds about 0.03ns, which is basically free.For the few compilers that don't support jump labels, runtime code paysthe cost of calling wake_up on an empty waitqueue, which was observed tobe about 30ns. However, most architectures that have sufficient memoryand CPU capacity to run XFS also support jump labels, so this is notmuch of a worry.Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner List of files: /linux-6.15/fs/xfs/Kconfig

d5c88131 - xfs: allow queued AG intents to drain before scrubbing

Wed, 12 Apr 2023 01:59:58 +0000

xfs: allow queued AG intents to drain before scrubbingWhen a writer thread executes a chain of log intent items, the AG headerbuffer locks will cycle during a transaction roll to get from one intentitem to the next in a chain. Although scrub takes all AG header bufferlocks, this isn't sufficient to guard against scrub checking an AG whilethat writer thread is in the middle of finishing a chain because there'sno higher level locking primitive guarding allocation groups.When there's a collision, cross-referencing between data structures(e.g. rmapbt and refcountbt) yields false corruption events; if repairis running, this results in incorrect repairs, which is catastrophic.Fix this by adding to the perag structure the count of active intentsand make scrub wait until it has both AG header buffer locks and theintent counter reaches zero.One quirk of the drain code is that deferred bmap updates also bump anddrop the intent counter. A fundamental decision made during the designphase of the reverse mapping feature is that updates to the rmapbtrecords are always made by the same code that updates the primarymetadata. In other words, callers of bmapi functions expect that thebmapi functions will queue deferred rmap updates.Some parts of the reflink code queue deferred refcount (CUI) and bmap(BUI) updates in the same head transaction, but the deferred workmanager completely finishes the CUI before the BUI work is started. Asa result, the CUI drops the intent count long before the deferred rmap(RUI) update even has a chance to bump the intent count. The only wayto keep the intent count elevated between the CUI and RUI is for the BUIto bump the counter until the RUI has been created.A second quirk of the intent drain code is that deferred work items mustincrement the intent counter as soon as the work item is added to thetransaction. When a BUI completes and queues an RUI, the RUI mustincrement the counter before the BUI decrements it. The only way toaccomplish this is to require that the counter be bumped as soon as thedeferred work item is created in memory.In the next patches we'll improve on this facility, but this patchprovides the basic functionality.Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner List of files: /linux-6.15/fs/xfs/Kconfig

89464554 - xfs: fix Kconfig asking about XFS_SUPPORT_V4 when XFS_FS=n

Mon, 12 Oct 2020 21:10:03 +0000

xfs: fix Kconfig asking about XFS_SUPPORT_V4 when XFS_FS=nPavel Machek complained that the question about supporting deprecatedXFS v4 comes up even when XFS is disabled. This clearly makes no sense,so fix Kconfig.Reported-by: Pavel Machek Signed-off-by: Darrick J. Wong Reviewed-by: Eric Sandeen List of files: /linux-6.15/fs/xfs/Kconfig

b96cb835 - xfs: deprecate the V4 format

Thu, 10 Sep 2020 17:57:17 +0000

xfs: deprecate the V4 formatThe V4 filesystem format contains known weaknesses in the on-disk formatthat make metadata verification diffiult. In addition, the format doesnot support dates past 2038 and will not be upgraded to do so. Weshould start the process of retiring the old format to close off attacksurfaces and to encourage users to migrate onto V5.Therefore, make XFS V4 support a configurable option. For the firstperiod it will be default Y in case some distributors want to withdrawsupport early; for the second period it will be default N so that anyonewho wishes to continue support can do so; and after that, support willbe removed from the kernel. Dates for these events have been added tothe upstream kernel.Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner Reviewed-by: Eric Sandeen List of files: /linux-6.15/fs/xfs/Kconfig

ec8f24b7 - treewide: Add SPDX license identifier - Makefile/Kconfig

Sun, 19 May 2019 12:07:45 +0000

treewide: Add SPDX license identifier - Makefile/KconfigAdd SPDX license identifiers to all Make/Kconfig files which: - Have no license information of any formThese files fall under the project license, GPL v2 only. The resulting SPDXlicense identifier is: GPL-2.0-onlySigned-off-by: Thomas Gleixner Signed-off-by: Greg Kroah-Hartman List of files: /linux-6.15/fs/xfs/Kconfig

72deb455 - block: remove CONFIG_LBDAF

Fri, 05 Apr 2019 16:08:59 +0000

block: remove CONFIG_LBDAFCurrently support for 64-bit sector_t and blkcnt_t is optional on 32-bitarchitectures. These types are required to support block device and/orfile sizes larger than 2 TiB, and have generally defaulted to on fora long time. Enabling the option only increases the i386 tinyconfigsize by 145 bytes, and many data structures already always use64-bit values for their in-core and on-disk data structures anyway,so there should not be a large change in dynamic memory usage either.Dropping this option removes a somewhat weird non-default config thathas cause various bugs or compiler warnings when actually used.Signed-off-by: Christoph Hellwig Signed-off-by: Jens Axboe List of files: /linux-6.15/fs/xfs/Kconfig

84d42ea6 - xfs: implement the metadata repair ioctl flag

Mon, 14 May 2018 13:34:36 +0000

xfs: implement the metadata repair ioctl flagPlumb in the pieces necessary to make the "scrub" subfunction ofthe scrub ioctl actually work. This means that we make the IFLAG_REPAIRflag to the scrub ioctl actually do something, and we add an errortagknob so that xfstests can force the kernel to rebuild a metadatastructure even if there's nothing wrong with it.Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner List of files: /linux-6.15/fs/xfs/Kconfig

91581e4c - fs/*/Kconfig: drop links to 404-compliant http://acl.bestbits.at

Wed, 20 Dec 2017 13:58:52 +0000

fs/*/Kconfig: drop links to 404-compliant http://acl.bestbits.atThis link is replicated in most filesystems' config stanzas. Referringto an archived version of that site is pointless as it mostly deals withpatches; user documentation is available elsewhere.Signed-off-by: Adam Borowski CC: Alexander Viro Reviewed-by: Darrick J. Wong Acked-by: Jan Kara Acked-by: Dave Kleikamp Acked-by: David Sterba Acked-by: "Yan, Zheng" Acked-by: Chao Yu Acked-by: Jaegeuk Kim Acked-by: Steve French Signed-off-by: Jonathan Corbet List of files: /linux-6.15/fs/xfs/Kconfig

36fd6e86 - xfs: create an ioctl to scrub AG metadata

Wed, 18 Oct 2017 04:37:34 +0000

xfs: create an ioctl to scrub AG metadataCreate an ioctl that can be used to scrub internal filesystem metadata.The new ioctl takes the metadata type, an (optional) AG number, an(optional) inode number and generation, and a flags argument. This willbe used by the upcoming XFS online scrub tool.Signed-off-by: Darrick J. Wong Reviewed-by: Dave Chinner List of files: /linux-6.15/fs/xfs/Kconfig

1040960e - xfs: define fatal assert build time tunable

Thu, 15 Jun 2017 04:29:13 +0000

xfs: define fatal assert build time tunableWhile configurable at runtime, the DEBUG mode assert failurebehavior is usually either desired or not for a particularsituation. For example, developers using kernel modules may preferfor fatal asserts to remain disabled across module reloads while QEengineers doing broad regression testing may prefer to have fatalasserts enabled on boot to facilitate data collection for bugreports.To provide a compromise/convenience for developers, create a Kconfigoption that sets the default value of the DEBUG mode 'bug_on_assert'sysfs tunable. The default behavior remains to trigger kernel BUGson assert failures to preserve existing behavior across kernelconfiguration updates with DEBUG mode enabled.Signed-off-by: Brian Foster Reviewed-by: Darrick J. Wong Signed-off-by: Darrick J. Wong List of files: /linux-6.15/fs/xfs/Kconfig