linux/fs/xfs
Omar Sandoval f63a5b3769 xfs: fix internal error from AGFL exhaustion
We've been seeing XFS errors like the following:

XFS: Internal error i != 1 at line 3526 of file fs/xfs/libxfs/xfs_btree.c.  Caller xfs_btree_insert+0x1ec/0x280
...
Call Trace:
 xfs_corruption_error+0x94/0xa0
 xfs_btree_insert+0x221/0x280
 xfs_alloc_fixup_trees+0x104/0x3e0
 xfs_alloc_ag_vextent_size+0x667/0x820
 xfs_alloc_fix_freelist+0x5d9/0x750
 xfs_free_extent_fix_freelist+0x65/0xa0
 __xfs_free_extent+0x57/0x180
...

This is the XFS_IS_CORRUPT() check in xfs_btree_insert() when
xfs_btree_insrec() fails.

After converting this into a panic and dissecting the core dump, I found
that xfs_btree_insrec() is failing because it's trying to split a leaf
node in the cntbt when the AG free list is empty. In particular, it's
failing to get a block from the AGFL _while trying to refill the AGFL_.

If a single operation splits every level of the bnobt and the cntbt (and
the rmapbt if it is enabled) at once, the free list will be empty. Then,
when the next operation tries to refill the free list, it allocates
space. If the allocation does not use a full extent, it will need to
insert records for the remaining space in the bnobt and cntbt. And if
those new records go in full leaves, the leaves (and potentially more
nodes up to the old root) need to be split.

Fix it by accounting for the additional splits that may be required to
refill the free list in the calculation for the minimum free list size.

P.S. As far as I can tell, this bug has existed for a long time -- maybe
back to xfs-history commit afdf80ae7405 ("Add XFS_AG_MAXLEVELS macros
...") in April 1994! It requires a very unlucky sequence of events, and
in fact we didn't hit it until a particular sparse mmap workload updated
from 5.12 to 5.19. But this bug existed in 5.12, so it must've been
exposed by some other change in allocation or writeback patterns. It's
also much less likely to be hit with the rmapbt enabled, since that
increases the minimum free list size and is unlikely to split at the
same time as the bnobt and cntbt.

Reviewed-by: "Darrick J. Wong" <djwong@kernel.org>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Chandan Babu R <chandanbabu@kernel.org>
2023-11-13 09:11:40 +05:30
..
libxfs xfs: fix internal error from AGFL exhaustion 2023-11-13 09:11:40 +05:30
scrub xfs: simplify rt bitmap/summary block accessor functions 2023-10-19 08:33:42 -07:00
Kconfig xfs: fix select in config XFS_ONLINE_SCRUB_STATS 2023-09-11 14:52:23 +05:30
kmem.c mm: introduce memalloc_retry_wait() 2022-01-15 16:30:29 +02:00
kmem.h
Makefile xfs: move the realtime summary file scrubber to a separate source file 2023-08-10 07:48:09 -07:00
mrlock.h
xfs_acl.c xfs: convert to ctime accessor functions 2023-07-24 10:30:06 +02:00
xfs_acl.h fs: port ->set_acl() to pass mnt_idmap 2023-01-19 09:24:27 +01:00
xfs_aops.c New code for 6.6: 2023-08-30 12:34:12 -07:00
xfs_aops.h
xfs_attr_inactive.c xfs: make inode unlinked bucket recovery work with quotacheck 2023-09-12 10:31:07 -07:00
xfs_attr_item.c xfs: reserve less log space when recovering log intent items 2023-09-12 10:31:07 -07:00
xfs_attr_item.h xfs: share xattr name and value buffers when logging xattr updates 2022-05-23 08:43:46 +10:00
xfs_attr_list.c xfs: use XFS_IFORK_Q to determine the presence of an xattr fork 2022-07-09 15:17:21 -07:00
xfs_bio_io.c fs/xfs: Use the enum req_op and blk_opf_t types 2022-07-14 12:14:33 -06:00
xfs_bmap_item.c xfs: reserve less log space when recovering log intent items 2023-09-12 10:31:07 -07:00
xfs_bmap_item.h
xfs_bmap_util.c New code for 6.7: 2023-11-08 13:22:16 -08:00
xfs_bmap_util.h xfs: xfs_bmap_punch_delalloc_range() should take a byte range 2022-11-29 09:09:17 +11:00
xfs_buf_item_recover.c xfs: verify buffer contents when we skip log replay 2023-04-12 15:49:23 +10:00
xfs_buf_item.c xfs: buffer pins need to hold a buffer reference 2023-06-05 04:05:27 +10:00
xfs_buf_item.h xfs: convert buffer log item flags to unsigned. 2022-04-21 10:46:40 +10:00
xfs_buf.c Many singleton patches against the MM code. The patch series which are 2023-11-02 19:38:47 -10:00
xfs_buf.h Many singleton patches against the MM code. The patch series which are 2023-11-02 19:38:47 -10:00
xfs_dahash_test.c xfs: test the ascii case-insensitive hash 2023-04-11 19:05:05 -07:00
xfs_dahash_test.h xfs: test dir/attr hash when loading module 2023-03-19 09:55:49 -07:00
xfs_dir2_readdir.c xfs: rearrange the logic and remove the broken comment for xfs_dir2_isxx 2022-10-04 16:39:58 +11:00
xfs_discard.c xfs: abort fstrim if kernel is suspending 2023-10-04 09:25:04 +11:00
xfs_discard.h xfs: move log discard work to xfs_discard.c 2023-10-04 09:24:02 +11:00
xfs_dquot_item_recover.c
xfs_dquot_item.c
xfs_dquot_item.h
xfs_dquot.c xfs: fix dqiterate thinko 2023-08-18 13:42:36 +05:30
xfs_dquot.h xfs: remove warning counters from struct xfs_dquot_res 2022-05-11 17:12:09 +10:00
xfs_drain.c xfs: minimize overhead of drain wakeups by using jump labels 2023-04-11 18:59:59 -07:00
xfs_drain.h xfs: minimize overhead of drain wakeups by using jump labels 2023-04-11 18:59:59 -07:00
xfs_error.c xfs: make kobj_type structures constant 2023-02-10 08:59:48 -08:00
xfs_error.h xfs: allow setting full range of panic tags 2023-02-09 18:36:17 -08:00
xfs_export.c xfs: fix reloading entire unlinked bucket lists 2023-09-24 18:12:13 -07:00
xfs_export.h
xfs_extent_busy.c xfs: process free extents to busy list in FIFO order 2023-10-11 12:35:21 -07:00
xfs_extent_busy.h xfs: reduce AGF hold times during fstrim operations 2023-10-04 09:24:52 +11:00
xfs_extfree_item.c xfs: reserve less log space when recovering log intent items 2023-09-12 10:31:07 -07:00
xfs_extfree_item.h xfs: refactor all the EFI/EFD log item sizeof logic 2022-10-31 08:58:20 -07:00
xfs_file.c xfs: allow read IO and FICLONE to run concurrently 2023-10-23 12:02:26 +05:30
xfs_filestream.c xfs: fix double xfs_perag_rele() in xfs_filestream_pick_ag() 2023-06-05 14:48:15 +10:00
xfs_filestream.h xfs: pass perag to filestreams tracing 2023-02-13 09:14:56 +11:00
xfs_fsmap.c xfs: convert do_div calls to xfs_rtb_to_rtx helper calls 2023-10-17 16:25:55 -07:00
xfs_fsmap.h
xfs_fsops.c Minor cleanups for 6.5: 2023-07-09 09:50:42 -07:00
xfs_fsops.h
xfs_globals.c xfs: allow setting full range of panic tags 2023-02-09 18:36:17 -08:00
xfs_health.c
xfs_icache.c xfs: dynamically allocate the xfs-inodegc shrinker 2023-10-04 10:32:26 -07:00
xfs_icache.h xfs: use per-mount cpumask to track nonempty percpu inodegc lists 2023-09-11 08:39:03 -07:00
xfs_icreate_item.c xfs: fix potential log item leak 2022-05-04 11:45:11 +10:00
xfs_icreate_item.h
xfs_inode_item_recover.c xfs: clean up "%Ld/%Lu" which doesn't meet C standard 2022-09-19 06:47:14 +10:00
xfs_inode_item.c New code for 6.7: 2023-11-08 13:22:16 -08:00
xfs_inode_item.h xfs: fix AGF vs inode cluster buffer deadlock 2023-06-05 04:08:27 +10:00
xfs_inode.c New code for 6.7: 2023-11-08 13:22:16 -08:00
xfs_inode.h xfs: allow read IO and FICLONE to run concurrently 2023-10-23 12:02:26 +05:30
xfs_ioctl32.c fs: port i_{g,u}id_into_vfs{g,u}id() to mnt_idmap 2023-01-19 09:24:29 +01:00
xfs_ioctl32.h arch: Remove Itanium (IA-64) architecture 2023-09-11 08:13:17 +00:00
xfs_ioctl.c xfs: create a helper to compute leftovers of realtime extents 2023-10-17 16:24:22 -07:00
xfs_ioctl.h fs: port ->fileattr_set() to pass mnt_idmap 2023-01-19 09:24:27 +01:00
xfs_iomap.c xfs: don't allocate into the data fork for an unshare request 2023-05-02 09:14:51 +10:00
xfs_iomap.h xfs: use iomap_valid method to detect stale cached iomaps 2022-11-29 09:09:17 +11:00
xfs_iops.c vfs-6.7.ctime 2023-10-30 09:47:13 -10:00
xfs_iops.h fs: port ->setattr() to pass mnt_idmap 2023-01-19 09:24:02 +01:00
xfs_itable.c xfs: convert to new timestamp accessors 2023-10-18 14:08:29 +02:00
xfs_itable.h fs: port i_{g,u}id_into_vfs{g,u}id() to mnt_idmap 2023-01-19 09:24:29 +01:00
xfs_iunlink_item.c xfs: create traced helper to get extra perag references 2023-04-11 18:59:55 -07:00
xfs_iunlink_item.h xfs: add in-memory iunlink log item 2022-07-14 11:47:42 +10:00
xfs_iwalk.c xfs: create traced helper to get extra perag references 2023-04-11 18:59:55 -07:00
xfs_iwalk.h xfs: Decouple XFS_IBULK flags from XFS_IWALK flags 2022-04-13 07:02:44 +00:00
xfs_linux.h xfs: use shifting and masking when converting rt extents, if possible 2023-10-17 16:26:25 -07:00
xfs_log_cil.c xfs: move log discard work to xfs_discard.c 2023-10-04 09:24:02 +11:00
xfs_log_priv.h xfs: move log discard work to xfs_discard.c 2023-10-04 09:24:02 +11:00
xfs_log_recover.c xfs: abort intent items when recovery intents fail 2023-11-13 09:08:34 +05:30
xfs_log.c xfs: up(ic_sema) if flushing data device fails 2023-11-13 09:11:40 +05:30
xfs_log.h xfs: move CIL ordering to the logvec chain 2022-07-07 18:56:08 +10:00
xfs_message.c Merge branch 'guilt/xfs-unsigned-flags-5.18' into xfs-5.19-for-next 2022-04-21 16:45:03 +10:00
xfs_message.h xfs: implement per-mount warnings for scrub and shrink usage 2022-05-27 10:31:34 +10:00
xfs_mount.c xfs: dynamically allocate the xfs-inodegc shrinker 2023-10-04 10:32:26 -07:00
xfs_mount.h New code for 6.7: 2023-11-08 13:22:16 -08:00
xfs_mru_cache.c
xfs_mru_cache.h
xfs_notify_failure.c xfs: correct calculation for agend and blockcount 2023-10-12 10:11:56 +05:30
xfs_ondisk.h xfs: use accessor functions for summary info words 2023-10-18 16:53:00 -07:00
xfs_pnfs.c fs: port ->setattr() to pass mnt_idmap 2023-01-19 09:24:02 +01:00
xfs_pnfs.h
xfs_pwork.c
xfs_pwork.h
xfs_qm_bhv.c
xfs_qm_syscalls.c xfs: introduce xfs_inodegc_push() 2022-06-23 13:34:38 -07:00
xfs_qm.c xfs: dynamically allocate the xfs-qm shrinker 2023-10-04 10:32:26 -07:00
xfs_qm.h xfs: dynamically allocate the xfs-qm shrinker 2023-10-04 10:32:26 -07:00
xfs_quota.h
xfs_quotaops.c xfs: don't set quota warning values 2022-05-11 17:12:09 +10:00
xfs_refcount_item.c xfs: reserve less log space when recovering log intent items 2023-09-12 10:31:07 -07:00
xfs_refcount_item.h
xfs_reflink.c xfs: only remap the written blocks in xfs_reflink_end_cow_extent 2023-11-13 09:09:19 +05:30
xfs_reflink.h xfs: pass perag to xfs_alloc_read_agf() 2022-07-07 19:07:40 +10:00
xfs_rmap_item.c xfs: reserve less log space when recovering log intent items 2023-09-12 10:31:07 -07:00
xfs_rmap_item.h
xfs_rtalloc.c New code for 6.7: 2023-11-08 13:22:16 -08:00
xfs_rtalloc.h xfs: convert rt extent numbers to xfs_rtxnum_t 2023-10-17 16:24:22 -07:00
xfs_stats.c xfs: replace unnecessary seq_printf with seq_puts 2022-09-19 06:48:14 +10:00
xfs_stats.h
xfs_super.c New code for 6.7: 2023-11-08 13:22:16 -08:00
xfs_super.h xfs: create scaffolding for creating debugfs entries 2023-08-10 07:48:07 -07:00
xfs_symlink.c fs: port fs{g,u}id helpers to mnt_idmap 2023-01-19 09:24:30 +01:00
xfs_symlink.h fs: port inode_init_owner() to mnt_idmap 2023-01-19 09:24:28 +01:00
xfs_sysctl.c xfs: simplify two-level sysctl registration for xfs_table 2023-04-13 11:49:35 -07:00
xfs_sysctl.h xfs: Add larp debug option 2022-05-11 17:01:22 +10:00
xfs_sysfs.c xfs: make kobj_type structures constant 2023-02-10 08:59:48 -08:00
xfs_sysfs.h xfs: make kobj_type structures constant 2023-02-10 08:59:48 -08:00
xfs_trace.c xfs: add debug knob to slow down writeback for fun 2022-11-28 17:24:35 -08:00
xfs_trace.h xfs: load uncached unlinked inodes into memory on demand 2023-09-12 10:31:07 -07:00
xfs_trans_ail.c xfs: don't reverse order of items in bulk AIL insertion 2023-06-29 09:28:23 -07:00
xfs_trans_buf.c
xfs_trans_dquot.c xfs: remove quota warning limit from struct xfs_quota_limits 2022-05-11 17:12:09 +10:00
xfs_trans_priv.h xfs: convert log vector chain to use list heads 2022-07-07 18:55:59 +10:00
xfs_trans.c xfs: use shifting and masking when converting rt extents, if possible 2023-10-17 16:26:25 -07:00
xfs_trans.h xfs: t_firstblock is tracking AGs not blocks 2023-02-11 04:11:06 +11:00
xfs_xattr.c vfs-6.7.xattr 2023-10-30 09:29:44 -10:00
xfs_xattr.h xfs: move xfs_xattr_handlers to .rodata 2023-10-10 13:49:20 +02:00
xfs.h