On Sun, Oct 09, 2022 at 05:58:33PM +0000, bugzilla-daemon@xxxxxxxxxx wrote: > https://bugzilla.kernel.org/show_bug.cgi?id=216567 > > --- Comment #1 from Zorro Lang (zlang@xxxxxxxxxx) --- > Hmm... besides this panic, g/451 just hit another panic when I tried to > reproduce this bug: > > [ 1084.111233] run fstests generic/451 at 2022-10-09 11:12:39 > [ 1099.015616] restraintd[2581]: *** Current Time: Sun Oct 09 11:12:56 2022 > Localwatchdog at: Tue Oct 11 10:57:56 2022 > [ 1101.932132] ------------[ cut here ]------------ > [ 1101.932220] ------------[ cut here ]------------ > [ 1101.936972] kernel BUG at include/linux/pagemap.h:1247! > [ 1101.936985] invalid opcode: 0000 [#1] PREEMPT SMP KASAN NOPTI > [ 1101.941681] kernel BUG at include/linux/pagemap.h:1247! > [ 1101.946825] CPU: 19 PID: 557513 Comm: xfs_io Kdump: loaded Not tainted > 6.0.0+ #1 > [ 1101.946831] Hardware name: Dell Inc. PowerEdge R750/0PJ80M, BIOS 1.5.4 > 12/17/2021 > [ 1101.946833] RIP: 0010:read_pages+0xa29/0xda0 > [ 1101.976950] Code: ff ff be 01 00 00 00 e9 87 fe ff ff 0f b6 d0 be ff ff ff > ff 4c 89 ff 88 44 24 18 e8 11 74 25 00 0f b6 44 24 18 e9 f1 fe ff ff <0f> 0b 4c > 89 ff e8 1d 86 00 00 e9 ea fe ff ff 48 c7 c6 c0 85 55 99 > [ 1101.995693] RSP: 0018:ffa00000396ef7f0 EFLAGS: 00010202 > [ 1102.000921] RAX: 0000000000000002 RBX: dffffc0000000000 RCX: > 0000000000000001 > [ 1102.008054] RDX: 1fe220003427d324 RSI: 0000000000000004 RDI: > ffd40000095e8500 > [ 1102.015186] RBP: ffffffffc13f66c0 R08: 0000000000000000 R09: > ffffffff9aa44067 > [ 1102.022321] R10: fffffbfff354880c R11: 0000000000000001 R12: > fff3fc00072ddf4a > [ 1102.029451] R13: ffa00000396efa54 R14: ffa00000396efa30 R15: > 0000000000000003 > [ 1102.036584] FS: 00007f1de484b740(0000) GS:ff11002033400000(0000) > knlGS:0000000000000000 > [ 1102.044671] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 > [ 1102.050418] CR2: 0000000001c81ff8 CR3: 000000016171e004 CR4: > 0000000000771ee0 > [ 1102.057549] DR0: 0000000000000000 DR1: 0000000000000000 DR2: > 0000000000000000 > [ 1102.064681] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: > 0000000000000400 > [ 1102.071815] PKRU: 55555554 > [ 1102.074527] Call Trace: > [ 1102.076982] <TASK> > [ 1102.079092] ? file_ra_state_init+0xe0/0xe0 > [ 1102.083283] ? __xa_clear_mark+0x100/0x100 > [ 1102.087385] page_cache_ra_unbounded+0x269/0x510 > [ 1102.092013] filemap_get_pages+0x26d/0x980 > [ 1102.096121] ? filemap_add_folio+0x150/0x150 > [ 1102.100403] filemap_read+0x2a9/0xae0 > [ 1102.104074] ? lock_acquire+0x1d8/0x620 > [ 1102.107921] ? find_held_lock+0x33/0x120 > [ 1102.111850] ? filemap_get_pages+0x980/0x980 > [ 1102.116121] ? validate_chain+0x154/0xdf0 > [ 1102.120133] ? __lock_contended+0x980/0x980 > [ 1102.124320] ? xfs_ilock+0x1d0/0x4d0 [xfs] > [ 1102.128582] ? xfs_ilock+0x1d0/0x4d0 [xfs] > [ 1102.132816] xfs_file_buffered_read+0x16f/0x390 [xfs] > [ 1102.137995] xfs_file_read_iter+0x274/0x560 [xfs] > [ 1102.142831] vfs_read+0x585/0x810 This is also a problem with the page cache, and doesn't seem related to XFS or directory block size configuration: BUG_ON(ractl->_batch_count > ractl->_nr_pages); Also, there haven't been any changes to XFS code so far in 6.1-rc0, so this isn't a recent XFS regression, either. Perhaps a bisect would be in order? -Dave. -- Dave Chinner david@xxxxxxxxxxxxx
