Skip to contentSkip to conversation

occasional failures in generic/455 and generic/457

DISCUSSION Matthew Wilcox 2 participants 2 messages Aug 19, 2020 – Aug 20, 2020

lore
Share & tools
lore
#0 lore raw

This is with 5.8.0-rc1 plus the series of seven patches to modify how pagevec_lookup_entries + find_get_entries() work [1]. I think it's unlikely those patches caused this, but it's sporadic so I can't be sure.

This one is from generic/457 but I've seen something similar from generic/455. I'm using the following:
+ echo 'MKFS_OPTIONS="-m reflink=1,rmapbt=1 -i sparse=1 -b size=1024"' >> /ktest/tests/xfstests/local.config
+ mkfs.$FSTYP -q $TEST_DEV -m reflink=1,rmapbt=1 -i sparse=1 -b size=1024

[1] https://lore.kernel.org/linux-mm/[email protected]/T/

6946 XFS (sdc): Internal error ltrec.rm_startblock > bno || ltrec.rm_startblock + ltrec.rm_blockcount < bno + len at line 575 of file fs/xfs/libxfs/xfs_rmap.c. Caller xfs_rmap_unmap+0x5fc/0x900
6946 CPU: 2 PID: 13591 Comm: mount Kdump: loaded Tainted: G W 5.9.0-rc1-00007-g96650d19d84b #452
6946 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.14.0-1 04/01/2014
6946 Call Trace:

6946  dump_stack+0x5e/0x7a
6946  xfs_corruption_error+0x7c/0x80
6946  ? xfs_rmap_unmap+0x5fc/0x900
6946  xfs_rmap_unmap+0x626/0x900
6946  ? xfs_rmap_unmap+0x5fc/0x900
6946  ? xfs_free_extent_fix_freelist+0x81/0xc0
6946  ? xfs_rmapbt_init_cursor+0x31/0x90
6946  xfs_rmap_finish_one+0x1dd/0x330
6946  xfs_rmap_update_finish_item+0x3f/0x70
6946  xfs_defer_finish_noroll+0x153/0x400
6946  ? kmem_alloc+0x74/0x120
6946  ? xfs_trans_commit+0xb/0x10
6946  __xfs_trans_commit+0x138/0x340
6946  xfs_trans_commit+0xb/0x10
6946  xfs_refcount_recover_cow_leftovers+0x1b9/0x300
6946  xfs_reflink_recover_cow+0x36/0x50
6946  xfs_mountfs+0x5a4/0x910
6946  xfs_fc_fill_super+0x34c/0x560
6946  get_tree_bdev+0x169/0x260
6946  ? xfs_setup_devices+0x80/0x80
6946  xfs_fc_get_tree+0x10/0x20
6946  vfs_get_tree+0x19/0x80
6946  path_mount+0x6ba/0x9f0
6946  __x64_sys_mount+0xe5/0x120
6946  do_syscall_64+0x32/0x50
6946  entry_SYSCALL_64_after_hwframe+0x44/0xa9

6946 RIP: 0033:0x7fa229f96fea
6946 Code: 48 8b 0d a9 0e 0c 00 f7 d8 64 89 01 48 83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 49 89 ca b8 a5 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 76 0e 0c 00 f7 d8 64 89 01 48 6946 RSP: 002b:00007ffcb19cf128 EFLAGS: 00000246 ORIG_RAX: 00000000000000a5
6946 RAX: ffffffffffffffda RBX: 000055eb7a037970 RCX: 00007fa229f96fea
6946 RDX: 000055eb7a037b80 RSI: 000055eb7a037bc0 RDI: 000055eb7a037ba0
6946 RBP: 00007fa22a2e41c4 R08: 0000000000000000 R09: 000055eb7a03a890
6946 R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
6946 R13: 0000000000000000 R14: 000055eb7a037ba0 R15: 000055eb7a037b80
6946 XFS (sdc): Corruption detected. Unmount and run xfs_repair
6946 XFS (sdc): xfs_do_force_shutdown(0x8) called from line 450 of file fs/xfs/libxfs/xfs_defer.c. Return address = ffffffff812ee23b
6946 XFS (sdc): Corruption of in-memory data detected. Shutting down filesystem
6946 XFS (sdc): Please unmount the filesystem and rectify the problem(s)
6946 XFS (sdc): Error -117 recovering leftover CoW allocations.
6946 [failed, exit status 1]XFS (sdb): Unmounting Filesystem

#1 lore raw
Quoted 9 lines · Matthew Wilcox

This is with 5.8.0-rc1 plus the series of seven patches to modify how pagevec_lookup_entries + find_get_entries() work [1]. I think it's unlikely those patches caused this, but it's sporadic so I can't be sure.

This one is from generic/457 but I've seen something similar from generic/455. I'm using the following:

(quoted diff) +2 -0
+    echo 'MKFS_OPTIONS="-m reflink=1,rmapbt=1 -i sparse=1 -b size=1024"' >> /ktest/tests/xfstests/local.config
+    mkfs.$FSTYP -q $TEST_DEV -m reflink=1,rmapbt=1 -i sparse=1 -b size=1024
Quoted 5 lines

[1] https://lore.kernel.org/linux-mm/[email protected]/T/

6946 XFS (sdc): Internal error ltrec.rm_startblock > bno || ltrec.rm_startblock + ltrec.rm_blockcount < bno + len at line 575 of file fs/xfs/libxfs/xfs_rmap.c. Caller xfs_rmap_unmap+0x5fc/0x900
6946 CPU: 2 PID: 13591 Comm: mount Kdump: loaded Tainted: G W 5.9.0-rc1-00007-g96650d19d84b #452
6946 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.14.0-1 04/01/2014

This looks similar to rmapbt related failures I saw with generic/455 and addressed with the following patch:

https://lore.kernel.org/linux-xfs/[email protected]/

Care to give that a whirl?

Brian

Quoted 43 lines

6946 Call Trace:

6946  dump_stack+0x5e/0x7a
6946  xfs_corruption_error+0x7c/0x80
6946  ? xfs_rmap_unmap+0x5fc/0x900
6946  xfs_rmap_unmap+0x626/0x900
6946  ? xfs_rmap_unmap+0x5fc/0x900
6946  ? xfs_free_extent_fix_freelist+0x81/0xc0
6946  ? xfs_rmapbt_init_cursor+0x31/0x90
6946  xfs_rmap_finish_one+0x1dd/0x330
6946  xfs_rmap_update_finish_item+0x3f/0x70
6946  xfs_defer_finish_noroll+0x153/0x400
6946  ? kmem_alloc+0x74/0x120
6946  ? xfs_trans_commit+0xb/0x10
6946  __xfs_trans_commit+0x138/0x340
6946  xfs_trans_commit+0xb/0x10
6946  xfs_refcount_recover_cow_leftovers+0x1b9/0x300
6946  xfs_reflink_recover_cow+0x36/0x50
6946  xfs_mountfs+0x5a4/0x910
6946  xfs_fc_fill_super+0x34c/0x560
6946  get_tree_bdev+0x169/0x260
6946  ? xfs_setup_devices+0x80/0x80
6946  xfs_fc_get_tree+0x10/0x20
6946  vfs_get_tree+0x19/0x80
6946  path_mount+0x6ba/0x9f0
6946  __x64_sys_mount+0xe5/0x120
6946  do_syscall_64+0x32/0x50
6946  entry_SYSCALL_64_after_hwframe+0x44/0xa9

6946 RIP: 0033:0x7fa229f96fea
6946 Code: 48 8b 0d a9 0e 0c 00 f7 d8 64 89 01 48 83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 49 89 ca b8 a5 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 76 0e 0c 00 f7 d8 64 89 01 48 6946 RSP: 002b:00007ffcb19cf128 EFLAGS: 00000246 ORIG_RAX: 00000000000000a5
6946 RAX: ffffffffffffffda RBX: 000055eb7a037970 RCX: 00007fa229f96fea
6946 RDX: 000055eb7a037b80 RSI: 000055eb7a037bc0 RDI: 000055eb7a037ba0
6946 RBP: 00007fa22a2e41c4 R08: 0000000000000000 R09: 000055eb7a03a890
6946 R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000
6946 R13: 0000000000000000 R14: 000055eb7a037ba0 R15: 000055eb7a037b80
6946 XFS (sdc): Corruption detected. Unmount and run xfs_repair
6946 XFS (sdc): xfs_do_force_shutdown(0x8) called from line 450 of file fs/xfs/libxfs/xfs_defer.c. Return address = ffffffff812ee23b
6946 XFS (sdc): Corruption of in-memory data detected. Shutting down filesystem
6946 XFS (sdc): Please unmount the filesystem and rectify the problem(s)
6946 XFS (sdc): Error -117 recovering leftover CoW allocations.
6946 [failed, exit status 1]XFS (sdb): Unmounting Filesystem