On 2019/3/29 下午2:20, Nikolay Borisov wrote:
>
>
> On 29.03.19 г. 1:18 ч., David Sterba wrote:
>> On Wed, Mar 27, 2019 at 02:24:03PM +0200, Nikolay Borisov wrote:
>>> Here is the (hopefully final) v4 of the fitrim patches. Main changes since
>>> v3:
>>
>> Nope, tests don't like it. It's v4 applied from mails on top of misc-next.
>>
>> generic/475 [20:36:10][11571.724402] run fstests generic/475 at
>> 2019-03-28 20:36:10
> <snip>
>
>> [11576.260681] BTRFS error (device dm-0): error reading free space cache
>> [11576.260832] BUG: unable to handle kernel paging request at
>> fffffffffffffffb
>> [11576.262301] BTRFS warning (device dm-0): failed to load free space cache
>> for block group 1104150528, rebuilding it now
>> [11576.265687] #PF error: [normal kernel read fault]
>> [11576.265690] PGD 2b016067 P4D 2b016067 PUD 2b018067 PMD 0
>> [11576.265696] Oops: 0000 [#1] PREEMPT SMP
>> [11576.265700] CPU: 0 PID: 25594 Comm: fsstress Not tainted
>> 5.1.0-rc2-default+ #531
>> [11576.265702] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS
>> rel-1.12.0-0-ga698c89-prebuilt.qemu.org 04/01/2014
>> [11576.265753] RIP: 0010:btrfs_lookup_dentry+0x4ce/0x5d0 [btrfs]
>> [11576.265756] Code: 4c 63 f8 85 c0 0f 88 90 fb ff ff 80 7d ae 01 0f 85 8c
>> fc ff ff 48 8b 7b 28 45 31 c0 31 c9 4c 89 f2 48 8d 75 a6 e8 e2 f0 ff ff <0f>
>> b7 38 49 89 c7 e8 c7 01 e9 d2 3a 45 98 0f 84 5d fb ff ff 41
>> [11576.271036] BTRFS: error (device dm-0) in __btrfs_free_extent:7103:
>> errno=-5 IO failure
>> [11576.271144] RSP: 0018:ffffa82cc3fe7db8 EFLAGS: 00010292
>> [11576.271149] RAX: fffffffffffffffb RBX: ffff94f41ea66218 RCX:
>> 0000001603c56600
>> [11576.273851] BTRFS info (device dm-0): forced readonly
>> [11576.274694] RDX: 0000000000000011 RSI: ffff94f41ea63ed8 RDI:
>> ffffffff930dacb4
>> [11576.274696] RBP: ffffa82cc3fe7e40 R08: 0000000000000001 R09:
>> fffffffffffffffb
>> [11576.274698] R10: 0000000000000000 R11: ffff94f46f36a910 R12:
>> ffff94f414e94500
>> [11576.275912] BTRFS: error (device dm-0) in btrfs_run_delayed_refs:3001:
>> errno=-5 IO failure
>> [11576.278115] R13: ffff94f474880000 R14: ffff94f403833000 R15:
>> ffff94f4563eaa10
>> [11576.278117] FS: 00007ff34f4f4b80(0000) GS:ffff94f47d400000(0000)
>> knlGS:0000000000000000
>> [11576.278121] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> [11576.278123] CR2: fffffffffffffffb CR3: 000000001f60c000 CR4:
>> 00000000000006f0
>> [11576.299608] Call Trace:
>> [11576.300463] btrfs_lookup+0xe/0x30 [btrfs]
>> [11576.301666] __lookup_hash+0x6b/0x90
>> [11576.302748] filename_create+0xa9/0x180
>> [11576.303865] do_mkdirat+0x53/0xf0
>> [11576.304844] do_syscall_64+0x54/0x180
>> [11576.305924] entry_SYSCALL_64_after_hwframe+0x49/0xbe
>> [11576.307131] RIP: 0033:0x7ff34f5e5307
>> [11576.307952] Code: 1f 40 00 48 8b 05 91 eb 0c 00 64 c7 00 5f 00 00 00 b8
>> ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 66 90 b8 53 00 00 00 0f 05 <48>
>> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 61 eb 0c 00 f7 d8 64 89 01
>
> My Code doesn't touch the lookup path, furthermore it doesn't touch the
> extent freeing logic, could this be something unrelated? How reliable it
> is to repro with/without this patchset?
OK, it's my inode mode checker patchset triggering all these.
It's btrfs_inode_type() call causing all the problems.
I'll investigate how this happens.
Thanks,
Qu
>
>> [11576.311502] RSP: 002b:00007fff14fd0a98 EFLAGS: 00000206 ORIG_RAX:
>> 0000000000000053
>> [11576.313415] RAX: ffffffffffffffda RBX: 0000000000000014 RCX:
>> 00007ff34f5e5307
>> [11576.314809] RDX: 0000000000000000 RSI: 00000000000001ff RDI:
>> 00000000004153d0
>> [11576.316224] RBP: 00007fff14fd0bf0 R08: fefefefefefefeff R09:
>> fefefefefeff3263
>> [11576.318087] R10: 0000000000000000 R11: 0000000000000206 R12:
>> 00000000000001ff
>> [11576.319955] R13: 0000000000405390 R14: 0000000000000000 R15:
>> 0000000000000000
>> [11576.321746] Modules linked in: dm_thin_pool dm_persistent_data
>> dm_bio_prison dm_snapshot dm_bufio btrfs libcrc32c xor zstd_decompress
>> zstd_compress xxhash raid6_pq dm_flakey dm_mod loop [last unloaded: scsi_d
>> [11576.325773] CR2: fffffffffffffffb
>> [11576.326498] ---[ end trace fe71251c3619f317 ]---
>> [11576.327514] RIP: 0010:btrfs_lookup_dentry+0x4ce/0x5d0 [btrfs]
>> [11576.328841] Code: 4c 63 f8 85 c0 0f 88 90 fb ff ff 80 7d ae 01 0f 85 8c
>> fc ff ff 48 8b 7b 28 45 31 c0 31 c9 4c 89 f2 48 8d 75 a6 e8 e2 f0 ff ff <0f>
>> b7 38 49 89 c7 e8 c7 01 e9 d2 3a 45 98 0f 84 5d fb ff ff 41
>> [11576.332993] RSP: 0018:ffffa82cc3fe7db8 EFLAGS: 00010292
>> [11576.334430] RAX: fffffffffffffffb RBX: ffff94f41ea66218 RCX:
>> 0000001603c56600
>> [11576.336263] RDX: 0000000000000011 RSI: ffff94f41ea63ed8 RDI:
>> ffffffff930dacb4
>> [11576.338080] RBP: ffffa82cc3fe7e40 R08: 0000000000000001 R09:
>> fffffffffffffffb
>> [11576.339907] R10: 0000000000000000 R11: ffff94f46f36a910 R12:
>> ffff94f414e94500
>> [11576.341535] R13: ffff94f474880000 R14: ffff94f403833000 R15:
>> ffff94f4563eaa10
>> [11576.342561] FS: 00007ff34f4f4b80(0000) GS:ffff94f47d400000(0000)
>> knlGS:0000000000000000
>> [11576.343689] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> [11576.345041] CR2: fffffffffffffffb CR3: 000000001f60c000 CR4:
>> 00000000000006f0
>> [11576.346593] BUG: sleeping function called from invalid context at
>> include/linux/cgroup-defs.h:721
>> [11576.348556] in_atomic(): 0, irqs_disabled(): 1, pid: 25594, name: fsstress
>> [11576.350006] INFO: lockdep is turned off.
>> [11576.350894] irq event stamp: 0
>> [11576.351608] hardirqs last enabled at (0): [<0000000000000000>]
>> (null)
>> [11576.353254] hardirqs last disabled at (0): [<ffffffff9305cc77>]
>> copy_process.part.72+0x847/0x1df0
>> [11576.355056] softirqs last enabled at (0): [<ffffffff9305cc77>]
>> copy_process.part.72+0x847/0x1df0
>> [11576.356705] softirqs last disabled at (0): [<0000000000000000>]
>> (null)
>> [11576.357848] CPU: 0 PID: 25594 Comm: fsstress Tainted: G D
>> 5.1.0-rc2-default+ #531
>> [11576.359000] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS
>> rel-1.12.0-0-ga698c89-prebuilt.qemu.org 04/01/2014
>> [11576.360913] Call Trace:
>> [11576.361480] dump_stack+0x67/0x90
>> [11576.362231] ___might_sleep.cold.85+0x9f/0xb1
>> [11576.363300] exit_signals+0x30/0x130
>> [11576.364220] do_exit+0xac/0xc10
>> [11576.365042] ? do_mkdirat+0x53/0xf0
>> [11576.365937] rewind_stack_do_exit+0x17/0x20
>> [failed, exit status 1] [20:36:15]- output mismatch (see
>> /tmp/fstests/results//generic/475.out.bad)
>>
