Bug 199899 - Mounting broken XFS partition causes infinite dmesg flood (system still responds)
Summary: Mounting broken XFS partition causes infinite dmesg flood (system still respo...
Status: NEW
Alias: None
Product: File System
Classification: Unclassified
Component: XFS (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: FileSystem/XFS Default Virtual Assignee
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2018-06-01 11:09 UTC by Anatoly Trosinenko
Modified: 2018-07-15 07:52 UTC (History)
1 user (show)

See Also:
Kernel Version: v4.18-rc4
Subsystem:
Regression: No
Bisected commit-id:

Attachments
XFS image triggering this bug (128 Mb unpacked) (5.15 KB, application/x-bzip)
2018-06-01 11:09 UTC, Anatoly Trosinenko 		Details
Kernel config (111.26 KB, text/x-mpsub)
2018-06-01 11:09 UTC, Anatoly Trosinenko 		Details
Kernel log truncated to about 2000 lines (111.57 KB, text/plain)
2018-06-01 11:10 UTC, Anatoly Trosinenko 		Details
Kernel config (112.67 KB, text/x-mpsub)
2018-07-15 07:36 UTC, Anatoly Trosinenko 		Details
Show Obsolete (1) Add an attachment (proposed patch, testcase, etc.)

Description Anatoly Trosinenko 2018-06-01 11:09:13 UTC 
Created attachment 276293 [details]
XFS image triggering this bug (128 Mb unpacked)

Mounting broken XFS filesystem on recent v4.17-rc7 causes mount to never return and dmesg is flooded. System still responds but kernel CPU usage is very high (when testing, I observed 90%).

How to reproduce:
1. Compile the kernel with the config attached
2. Unpack the attached image (128 Mb) and mount it as XFS

What happens:

Mount command does not return, dmesg is flooded with

[    1.584607] XFS (sda): Internal error xlog_valid_rec_header(2) at line 5280 of file fs/xfs/xfs_log_recover.c.  Caller xlog_do_recovery_pass+0x4da/0x5e0
[    1.584775] CPU: 0 PID: 989 Comm: exe Not tainted 4.17.0-rc7 #2
[    1.584846] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.10.2-1ubuntu1 04/01/2014
[    1.584948] Call Trace:
[    1.584995]  dump_stack+0x46/0x5b
[    1.585058]  xlog_valid_rec_header.isra.21+0xd1/0xe0
[    1.585124]  xlog_do_recovery_pass+0x4da/0x5e0
[    1.585187]  ? xlog_bread_noalign+0x7c/0xf0
[    1.585246]  xlog_verify_tail+0x139/0x1d0
[    1.585305]  xlog_verify_head+0xcf/0x190
[    1.585376]  xlog_find_tail+0x255/0x340
[    1.585432]  xlog_recover+0x26/0x170
[    1.585504]  xfs_log_mount+0x175/0x2a0
[    1.585593]  xfs_mountfs+0x524/0x910
[    1.585680]  xfs_fs_fill_super+0x417/0x580
[    1.585736]  ? xfs_test_remount_options+0x50/0x50
[    1.585799]  mount_bdev+0x16e/0x1a0
[    1.585888]  mount_fs+0x30/0x150
[    1.585936]  vfs_kern_mount.part.28+0x4f/0xf0
[    1.585998]  do_mount+0x5d0/0xc60
[    1.586045]  ? _copy_from_user+0x37/0x60
[    1.586097]  ? memdup_user+0x39/0x60
[    1.586146]  ksys_mount+0x7b/0xd0
[    1.586192]  __x64_sys_mount+0x1c/0x20
[    1.586242]  do_syscall_64+0x43/0xf0
[    1.586295]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
[    1.586360] RIP: 0033:0x48d31a
[    1.586401] RSP: 002b:00007ffeaee5d0d8 EFLAGS: 00000246 ORIG_RAX: 00000000000000a5
[    1.586495] RAX: ffffffffffffffda RBX: 0000000000008000 RCX: 000000000048d31a
[    1.586579] RDX: 00007ffeaee5efa2 RSI: 00007ffeaee5ef9a RDI: 00007ffeaee5ef91
[    1.586663] RBP: 00000000017328a0 R08: 0000000000000000 R09: 0000000000000000
[    1.586746] R10: 0000000000008000 R11: 0000000000000246 R12: 0000000000000000
[    1.586828] R13: 0000000000000000 R14: 00007ffeaee5d348 R15: 0000000000000000
Comment 1 Anatoly Trosinenko 2018-06-01 11:09:49 UTC 
Created attachment 276295 [details]
Kernel config
Comment 2 Anatoly Trosinenko 2018-06-01 11:10:37 UTC 
Created attachment 276297 [details]
Kernel log truncated to about 2000 lines
Comment 3 Anatoly Trosinenko 2018-07-15 07:36:03 UTC 
Created attachment 277357 [details]
Kernel config

Updated config for v4.18-rc4
Comment 4 Anatoly Trosinenko 2018-07-15 07:52:48 UTC 
Reproducible on xfs-linux/for-next (commit efe80327)
Note You need to log in before you can comment on or make changes to this bug.