Bug 199367 - Uninitialized function pointer (lookup) triggered when mounting and operating a crafted xfs image
Summary: Uninitialized function pointer (lookup) triggered when mounting and operating...
Status: RESOLVED CODE_FIX
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: 2018-04-12 02:33 UTC by Wen Xu
Modified: 2018-06-02 17:11 UTC (History)
4 users (show)

See Also:
Kernel Version: 4.16.x, 4.15.x
Tree: Mainline
Regression: No


Attachments
The (compressed) crafted image which causes crash (24.47 KB, application/zip)
2018-04-12 02:33 UTC, Wen Xu
Details
poc.c (3.18 KB, text/plain)
2018-04-12 02:34 UTC, Wen Xu
Details
config-4.16.0-rc1+ (130.18 KB, text/x-mpsub)
2018-04-12 04:06 UTC, Wen Xu
Details
The (compressed) crafted image which causes crash (22.17 KB, application/zip)
2018-04-13 04:46 UTC, Wen Xu
Details
Simplified poc for 53.img to trigger the bug (895 bytes, text/plain)
2018-04-15 05:57 UTC, Wen Xu
Details

Description Wen Xu 2018-04-12 02:33:39 UTC
Created attachment 275303 [details]
The (compressed) crafted image which causes crash

- Overview
Uninitialized function pointer (lookup) triggered when mounting and operating a crafted xfs image

- Reproduce
# mkdir mnt
# mount -t xfs 5.img mnt
# gcc -o poc poc.c
# ./poc ./mnt

- Reason
https://elixir.bootlin.com/linux/v4.15/source/fs/namei.c#L1625
XFS does not properly initialize an image's lookup() operator.

- Crash dump
[   62.197531] XFS (loop0): Mounting V4 Filesystem
[   62.204072] XFS (loop0): Ending clean mount
[   66.091635] XFS (loop0): Metadata corruption detected at xfs_attr_shortform_verify+0xb5/0xf0, inode 0x6c2 attr fork
[   66.093297] XFS (loop0): Unmount and run xfs_repair
[   66.094125] XFS (loop0): Metadata corruption detected at xfs_attr_shortform_verify+0xb5/0xf0, inode 0x6c2 attr fork
[   66.095742] XFS (loop0): Unmount and run xfs_repair
[   66.096742] XFS (loop0): Metadata corruption detected at xfs_attr_shortform_verify+0xb5/0xf0, inode 0x6c2 attr fork
[   66.102112] XFS (loop0): Unmount and run xfs_repair
[   66.102928] XFS (loop0): Metadata corruption detected at xfs_attr_shortform_verify+0xb5/0xf0, inode 0x6c2 attr fork
[   66.104585] XFS (loop0): Unmount and run xfs_repair
[   66.105399] XFS (loop0): Metadata corruption detected at xfs_attr_shortform_verify+0xb5/0xf0, inode 0x6c2 attr fork
[   66.107035] XFS (loop0): Unmount and run xfs_repair
[   66.107893] BUG: unable to handle kernel NULL pointer dereference at           (null)
[   66.109121] IP:           (null)
[   66.109654] PGD 800000013938f067 P4D 800000013938f067 PUD 133bb7067 PMD 0
[   66.110727] Oops: 0010 [#1] SMP PTI
[   66.111289] Modules linked in: snd_hda_codec_generic snd_hda_intel snd_hda_codec snd_hwdep snd_hda_core snd_pcm snd_timer snd soundcore i2c_piix4 mac_hid ib_iser rdma_cm iw_cm ib_cm ib_core iscsi_tcp libiscsi_tcp libiscsi scsi_transport_iscsi autofs4 raid10 raid456 async_raid6_recov async_memcpy async_pq async_xor async_tx raid1 raid0 multipath linear crct10dif_pclmul crc32_pclmul 8139too qxl aesni_intel drm_kms_helper aes_x86_64 crypto_simd syscopyarea cryptd sysfillrect sysimgblt fb_sys_fops glue_helper ttm drm 8139cp mii floppy pata_acpi
[   66.118963] CPU: 1 PID: 1362 Comm: poc Not tainted 4.16.0-rc1+ #2
[   66.119910] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Ubuntu-1.8.2-1ubuntu1 04/01/2014
[   66.121366] RIP: 0010:          (null)
[   66.121963] RSP: 0018:ffffba76c0aabca0 EFLAGS: 00010206
[   66.122794] RAX: 0000000000000000 RBX: ffff9c473943c180 RCX: 0000000000000000
[   66.123899] RDX: 0000000000000001 RSI: ffff9c473943c180 RDI: ffff9c4738dc44f8
[   66.124999] RBP: ffff9c4738dc45a0 R08: 0000000000000002 R09: 0000000000000005
[   66.126124] R10: 000000004cbf3f44 R11: 000000004cbf3f44 R12: ffff9c473943ca80
[   66.127234] R13: 0000000000000001 R14: ffffba76c0aabde0 R15: ffff9c4738dc44f8
[   66.128339] FS:  00007ff5c2e51700(0000) GS:ffff9c473fd00000(0000) knlGS:0000000000000000
[   66.129601] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   66.130496] CR2: 0000000000000000 CR3: 0000000138730000 CR4: 00000000000006e0
[   66.131601] Call Trace:
[   66.132024]  ? lookup_slow+0xa9/0x170
[   66.132618]  ? walk_component+0x1c6/0x350
[   66.133258]  ? path_lookupat+0x76/0x230
[   66.134002]  ? filename_lookup+0xb8/0x1a0
[   66.134662]  ? _cond_resched+0x16/0x40
[   66.135264]  ? kmem_cache_alloc+0xd9/0x1b0
[   66.135929]  ? path_listxattr+0x41/0xa0
[   66.136545]  ? path_listxattr+0x41/0xa0
[   66.137170]  ? do_syscall_64+0x71/0x120
[   66.137795]  ? entry_SYSCALL_64_after_hwframe+0x21/0x86
[   66.138615] Code:  Bad RIP value.
[   66.139154] RIP:           (null) RSP: ffffba76c0aabca0
[   66.139975] CR2: 0000000000000000
[   66.140538] ---[ end trace 8491db2208736519 ]---

- Credit
Reported by Wen Xu at SSLab, Gatech
Comment 1 Wen Xu 2018-04-12 02:34:38 UTC
Created attachment 275305 [details]
poc.c
Comment 2 Eric Sandeen 2018-04-12 03:21:28 UTC
If you're going to be running the fuzzer script for a while, it would be super to compile your kernel w/ frame pointers so that the backtraces you provide are more useful:

CONFIG_FRAME_POINTER=y

For that matter, please include your entire kernel config, because this specific bug does not reproduce on my 4.16 kernel.

Thanks,
-Eric
Comment 3 Wen Xu 2018-04-12 03:30:31 UTC
Oh, really...md5sum(5.img) = f839e50539770fa2faab324fa5d34564

Let me first paste the message of latest Ubuntu 18.04 I also test on which uses 4.15.0-12-generic;

[ 1081.146990] XFS (loop0): Mounting V4 Filesystem
[ 1081.149756] XFS (loop0): Ending clean mount
[ 1087.546113] XFS (loop0): corrupt inode 1730 (bad attr fork size 0).
[ 1087.546150] 00000000d07ef4d1: 49 4e 81 a4 02 02 00 00 00 00 00 00 00 00 00 00  IN..............
[ 1087.546187] 00000000c038e632: 00 00 00 00 ff ff dd 00 00 00 00 00 00 00 00 01  ................
[ 1087.546213] 00000000600bcf94: 5a b2 9b 71 07 03 42 1b 5a b2 9b 71 07 03 42 1b  Z..q..B.Z..q..B.
[ 1087.546239] 000000006ff0297d: 5a b2 9b 71 07 03 42 1b 00 00 00 00 00 00 00 00  Z..q..B.........
[ 1087.549984] XFS (loop0): Internal error xfs_iformat(8) at line 218 of file /build/linux-LL9XfP/linux-4.15.0/fs/xfs/libxfs/xfs_inode_fork.c.  Caller xfs_iread+0x1fd/0x220 [xfs]
[ 1087.550161] CPU: 0 PID: 52410 Comm: poc Not tainted 4.15.0-12-generic #13-Ubuntu
[ 1087.550163] Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 07/02/2015
[ 1087.550164] Call Trace:
[ 1087.551789]  dump_stack+0x63/0x8b
[ 1087.551844]  xfs_corruption_error+0x87/0x90 [xfs]
[ 1087.551866]  ? xfs_iread+0x1fd/0x220 [xfs]
[ 1087.551886]  xfs_iformat_fork+0x61c/0x630 [xfs]
[ 1087.551904]  ? xfs_iread+0x1fd/0x220 [xfs]
[ 1087.551922]  ? xfs_inode_from_disk+0x35/0x230 [xfs]
[ 1087.551940]  xfs_iread+0x1fd/0x220 [xfs]
[ 1087.551963]  xfs_iget+0x2b1/0x930 [xfs]
[ 1087.551981]  ? xfs_da_compname+0x1f/0x30 [xfs]
[ 1087.552005]  xfs_lookup+0xe1/0x120 [xfs]
[ 1087.552026]  xfs_vn_lookup+0x70/0xb0 [xfs]
[ 1087.552678]  path_openat+0xcab/0x16c0
[ 1087.552729]  ? xfs_trans_free_items+0x87/0xb0 [xfs]
[ 1087.552734]  do_filp_open+0x9b/0x110
[ 1087.552910]  ? __check_object_size+0xaf/0x1b0
[ 1087.552948]  ? __alloc_fd+0x46/0x170
[ 1087.552985]  do_sys_open+0x1bb/0x2c0
[ 1087.552988]  ? do_sys_open+0x1bb/0x2c0
[ 1087.553025]  ? _cond_resched+0x19/0x40
[ 1087.553028]  SyS_openat+0x14/0x20
[ 1087.553153]  do_syscall_64+0x73/0x130
[ 1087.553157]  entry_SYSCALL_64_after_hwframe+0x3d/0xa2
[ 1087.553222] RIP: 0033:0x7fa7faecfc8e
[ 1087.553226] RSP: 002b:00007ffcb19b4bf0 EFLAGS: 00000246 ORIG_RAX: 0000000000000101
[ 1087.553228] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007fa7faecfc8e
[ 1087.553229] RDX: 0000000000000000 RSI: 0000561f966992d0 RDI: 00000000ffffff9c
[ 1087.553230] RBP: 00007ffcb19b4dc0 R08: 0000000000000078 R09: 0000000000000000
[ 1087.553230] R10: 0000000000000000 R11: 0000000000000246 R12: 0000561f9544ed30
[ 1087.553231] R13: 00007ffcb19b4ec0 R14: 0000000000000000 R15: 0000000000000000
[ 1087.553269] XFS (loop0): Corruption detected. Unmount and run xfs_repair
[ 1087.554820] XFS (loop0): corrupt inode 1730 (bad attr fork size 0).
[ 1087.554823] 00000000d07ef4d1: 49 4e 81 a4 02 02 00 00 00 00 00 00 00 00 00 00  IN..............
[ 1087.554854] 00000000c038e632: 00 00 00 00 ff ff dd 00 00 00 00 00 00 00 00 01  ................
[ 1087.554880] 00000000600bcf94: 5a b2 9b 71 07 03 42 1b 5a b2 9b 71 07 03 42 1b  Z..q..B.Z..q..B.
[ 1087.554907] 000000006ff0297d: 5a b2 9b 71 07 03 42 1b 00 00 00 00 00 00 00 00  Z..q..B.........
[ 1087.554990] XFS (loop0): Internal error xfs_iformat(8) at line 218 of file /build/linux-LL9XfP/linux-4.15.0/fs/xfs/libxfs/xfs_inode_fork.c.  Caller xfs_iread+0x1fd/0x220 [xfs]
[ 1087.555038] CPU: 0 PID: 52410 Comm: poc Not tainted 4.15.0-12-generic #13-Ubuntu
[ 1087.555103] Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 07/02/2015
[ 1087.555105] Call Trace:
[ 1087.555110]  dump_stack+0x63/0x8b
[ 1087.555136]  xfs_corruption_error+0x87/0x90 [xfs]
[ 1087.555156]  ? xfs_iread+0x1fd/0x220 [xfs]
[ 1087.555176]  xfs_iformat_fork+0x61c/0x630 [xfs]
[ 1087.555194]  ? xfs_iread+0x1fd/0x220 [xfs]
[ 1087.555212]  ? xfs_inode_from_disk+0x35/0x230 [xfs]
[ 1087.555230]  xfs_iread+0x1fd/0x220 [xfs]
[ 1087.555253]  xfs_iget+0x2b1/0x930 [xfs]
[ 1087.555271]  ? xfs_da_compname+0x1f/0x30 [xfs]
[ 1087.555295]  xfs_lookup+0xe1/0x120 [xfs]
[ 1087.555316]  xfs_vn_lookup+0x70/0xb0 [xfs]
[ 1087.555318]  path_openat+0xcab/0x16c0
[ 1087.555320]  do_filp_open+0x9b/0x110
[ 1087.555322]  ? __check_object_size+0xaf/0x1b0
[ 1087.555324]  ? __alloc_fd+0x46/0x170
[ 1087.555326]  do_sys_open+0x1bb/0x2c0
[ 1087.555328]  ? do_sys_open+0x1bb/0x2c0
[ 1087.555330]  ? _cond_resched+0x19/0x40
[ 1087.555332]  SyS_openat+0x14/0x20
[ 1087.555334]  do_syscall_64+0x73/0x130
[ 1087.555336]  entry_SYSCALL_64_after_hwframe+0x3d/0xa2
[ 1087.555338] RIP: 0033:0x7fa7faecfc8e
[ 1087.555339] RSP: 002b:00007ffcb19b4bf0 EFLAGS: 00000246 ORIG_RAX: 0000000000000101
[ 1087.555341] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007fa7faecfc8e
[ 1087.555341] RDX: 0000000000000202 RSI: 0000561f966992d0 RDI: 00000000ffffff9c
[ 1087.555342] RBP: 00007ffcb19b4dc0 R08: 0000000000000078 R09: 0000000000000000
[ 1087.555343] R10: 0000000000000000 R11: 0000000000000246 R12: 0000561f9544ed30
[ 1087.555344] R13: 00007ffcb19b4ec0 R14: 0000000000000000 R15: 0000000000000000
[ 1087.555346] XFS (loop0): Corruption detected. Unmount and run xfs_repair
[ 1087.557979] XFS (loop0): corrupt inode 1730 (bad attr fork size 0).
[ 1087.557981] 00000000d07ef4d1: 49 4e 81 a4 02 02 00 00 00 00 00 00 00 00 00 00  IN..............
[ 1087.558008] 00000000c038e632: 00 00 00 00 ff ff dd 00 00 00 00 00 00 00 00 01  ................
[ 1087.558031] 00000000600bcf94: 5a b2 9b 71 07 03 42 1b 5a b2 9b 71 07 03 42 1b  Z..q..B.Z..q..B.
[ 1087.558054] 000000006ff0297d: 5a b2 9b 71 07 03 42 1b 00 00 00 00 00 00 00 00  Z..q..B.........
[ 1087.558113] XFS (loop0): Internal error xfs_iformat(8) at line 218 of file /build/linux-LL9XfP/linux-4.15.0/fs/xfs/libxfs/xfs_inode_fork.c.  Caller xfs_iread+0x1fd/0x220 [xfs]
[ 1087.558190] CPU: 0 PID: 52410 Comm: poc Not tainted 4.15.0-12-generic #13-Ubuntu
[ 1087.558191] Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 07/02/2015
[ 1087.558191] Call Trace:
[ 1087.558195]  dump_stack+0x63/0x8b
[ 1087.558229]  xfs_corruption_error+0x87/0x90 [xfs]
[ 1087.558247]  ? xfs_iread+0x1fd/0x220 [xfs]
[ 1087.558263]  xfs_iformat_fork+0x61c/0x630 [xfs]
[ 1087.558277]  ? xfs_iread+0x1fd/0x220 [xfs]
[ 1087.558291]  ? xfs_inode_from_disk+0x35/0x230 [xfs]
[ 1087.558304]  xfs_iread+0x1fd/0x220 [xfs]
[ 1087.558323]  xfs_iget+0x2b1/0x930 [xfs]
[ 1087.558338]  ? xfs_da_compname+0x1f/0x30 [xfs]
[ 1087.558357]  xfs_lookup+0xe1/0x120 [xfs]
[ 1087.558375]  xfs_vn_lookup+0x70/0xb0 [xfs]
[ 1087.558377]  path_openat+0xcab/0x16c0
[ 1087.558379]  do_filp_open+0x9b/0x110
[ 1087.558381]  ? __check_object_size+0xaf/0x1b0
[ 1087.558383]  ? __alloc_fd+0x46/0x170
[ 1087.558385]  do_sys_open+0x1bb/0x2c0
[ 1087.558386]  ? do_sys_open+0x1bb/0x2c0
[ 1087.558388]  ? _cond_resched+0x19/0x40
[ 1087.558389]  SyS_openat+0x14/0x20
[ 1087.558391]  do_syscall_64+0x73/0x130
[ 1087.558393]  entry_SYSCALL_64_after_hwframe+0x3d/0xa2
[ 1087.558394] RIP: 0033:0x7fa7faecfc8e
[ 1087.558395] RSP: 002b:00007ffcb19b4bf0 EFLAGS: 00000246 ORIG_RAX: 0000000000000101
[ 1087.558397] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007fa7faecfc8e
[ 1087.558398] RDX: 0000000000000202 RSI: 0000561f966992d0 RDI: 00000000ffffff9c
[ 1087.558398] RBP: 00007ffcb19b4dc0 R08: 0000000000000078 R09: 0000000000000000
[ 1087.558399] R10: 0000000000000000 R11: 0000000000000246 R12: 0000561f9544ed30
[ 1087.558452] R13: 00007ffcb19b4ec0 R14: 0000000000000000 R15: 0000000000000000
[ 1087.558476] XFS (loop0): Corruption detected. Unmount and run xfs_repair
[ 1087.558851] XFS (loop0): corrupt inode 1730 (bad attr fork size 0).
[ 1087.558853] 00000000d07ef4d1: 49 4e 81 a4 02 02 00 00 00 00 00 00 00 00 00 00  IN..............
[ 1087.558876] 00000000c038e632: 00 00 00 00 ff ff dd 00 00 00 00 00 00 00 00 01  ................
[ 1087.558897] 00000000600bcf94: 5a b2 9b 71 07 03 42 1b 5a b2 9b 71 07 03 42 1b  Z..q..B.Z..q..B.
[ 1087.558918] 000000006ff0297d: 5a b2 9b 71 07 03 42 1b 00 00 00 00 00 00 00 00  Z..q..B.........
[ 1087.558956] XFS (loop0): Internal error xfs_iformat(8) at line 218 of file /build/linux-LL9XfP/linux-4.15.0/fs/xfs/libxfs/xfs_inode_fork.c.  Caller xfs_iread+0x1fd/0x220 [xfs]
[ 1087.558994] CPU: 0 PID: 52410 Comm: poc Not tainted 4.15.0-12-generic #13-Ubuntu
[ 1087.558994] Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 07/02/2015
[ 1087.558995] Call Trace:
[ 1087.559008]  dump_stack+0x63/0x8b
[ 1087.559026]  xfs_corruption_error+0x87/0x90 [xfs]
[ 1087.559040]  ? xfs_iread+0x1fd/0x220 [xfs]
[ 1087.559054]  xfs_iformat_fork+0x61c/0x630 [xfs]
[ 1087.559067]  ? xfs_iread+0x1fd/0x220 [xfs]
[ 1087.559090]  ? xfs_inode_from_disk+0x35/0x230 [xfs]
[ 1087.559105]  xfs_iread+0x1fd/0x220 [xfs]
[ 1087.559122]  xfs_iget+0x2b1/0x930 [xfs]
[ 1087.559136]  ? xfs_da_compname+0x1f/0x30 [xfs]
[ 1087.559153]  xfs_lookup+0xe1/0x120 [xfs]
[ 1087.559169]  xfs_vn_lookup+0x70/0xb0 [xfs]
[ 1087.559171]  __lookup_hash+0x6a/0xa0
[ 1087.559173]  SyS_rename+0x1fb/0x3c0
[ 1087.559175]  do_syscall_64+0x73/0x130
[ 1087.559177]  entry_SYSCALL_64_after_hwframe+0x3d/0xa2
[ 1087.559178] RIP: 0033:0x7fa7fae3bd37
[ 1087.559178] RSP: 002b:00007ffcb19b4c58 EFLAGS: 00000286 ORIG_RAX: 0000000000000052
[ 1087.559180] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007fa7fae3bd37
[ 1087.559180] RDX: 0000561f966992f0 RSI: 0000561f966992f0 RDI: 0000561f966992d0
[ 1087.559181] RBP: 00007ffcb19b4dc0 R08: 0000000000000078 R09: 0000000000000000
[ 1087.559181] R10: 0000000000000000 R11: 0000000000000286 R12: 0000561f9544ed30
[ 1087.559182] R13: 00007ffcb19b4ec0 R14: 0000000000000000 R15: 0000000000000000
[ 1087.559205] XFS (loop0): Corruption detected. Unmount and run xfs_repair
[ 1087.561587] XFS (loop0): corrupt inode 1730 (bad attr fork size 0).
[ 1087.561591] 00000000d07ef4d1: 49 4e 81 a4 02 02 00 00 00 00 00 00 00 00 00 00  IN..............
[ 1087.562100] 00000000c038e632: 00 00 00 00 ff ff dd 00 00 00 00 00 00 00 00 01  ................
[ 1087.562547] 00000000600bcf94: 5a b2 9b 71 07 03 42 1b 5a b2 9b 71 07 03 42 1b  Z..q..B.Z..q..B.
[ 1087.563000] 000000006ff0297d: 5a b2 9b 71 07 03 42 1b 00 00 00 00 00 00 00 00  Z..q..B.........
[ 1087.563506] XFS (loop0): Internal error xfs_iformat(8) at line 218 of file /build/linux-LL9XfP/linux-4.15.0/fs/xfs/libxfs/xfs_inode_fork.c.  Caller xfs_iread+0x1fd/0x220 [xfs]
[ 1087.564471] CPU: 0 PID: 52410 Comm: poc Not tainted 4.15.0-12-generic #13-Ubuntu
[ 1087.564471] Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 07/02/2015
[ 1087.564472] Call Trace:
[ 1087.564478]  dump_stack+0x63/0x8b
[ 1087.564504]  xfs_corruption_error+0x87/0x90 [xfs]
[ 1087.564517]  ? xfs_iread+0x1fd/0x220 [xfs]
[ 1087.564529]  xfs_iformat_fork+0x61c/0x630 [xfs]
[ 1087.564541]  ? xfs_iread+0x1fd/0x220 [xfs]
[ 1087.564552]  ? xfs_inode_from_disk+0x35/0x230 [xfs]
[ 1087.564567]  xfs_iread+0x1fd/0x220 [xfs]
[ 1087.564581]  xfs_iget+0x2b1/0x930 [xfs]
[ 1087.564593]  ? xfs_da_compname+0x1f/0x30 [xfs]
[ 1087.564608]  xfs_lookup+0xe1/0x120 [xfs]
[ 1087.564622]  xfs_vn_lookup+0x70/0xb0 [xfs]
[ 1087.564625]  __lookup_hash+0x6a/0xa0
[ 1087.564626]  do_unlinkat+0x137/0x320
[ 1087.564627]  SyS_unlink+0x1f/0x30
[ 1087.564630]  do_syscall_64+0x73/0x130
[ 1087.564632]  entry_SYSCALL_64_after_hwframe+0x3d/0xa2
[ 1087.564634] RIP: 0033:0x7fa7faed1d47
[ 1087.564634] RSP: 002b:00007ffcb19b4c58 EFLAGS: 00000286 ORIG_RAX: 0000000000000057
[ 1087.564698] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007fa7faed1d47
[ 1087.564698] RDX: 0000000000000001 RSI: 0000000000000001 RDI: 0000561f966992d0
[ 1087.564699] RBP: 00007ffcb19b4dc0 R08: 0000000000000078 R09: 0000000000000000
[ 1087.564699] R10: 0000000000000000 R11: 0000000000000286 R12: 0000561f9544ed30
[ 1087.564700] R13: 00007ffcb19b4ec0 R14: 0000000000000000 R15: 0000000000000000
[ 1087.564737] XFS (loop0): Corruption detected. Unmount and run xfs_repair
[ 1087.568680] BUG: unable to handle kernel NULL pointer dereference at           (null)
[ 1087.569170] IP:           (null)
[ 1087.569639] PGD 0 P4D 0
[ 1087.570095] Oops: 0010 [#1] SMP PTI
[ 1087.570624] Modules linked in: ufs qnx4 hfsplus hfs minix ntfs msdos jfs xfs vmw_balloon coretemp intel_rapl_perf input_leds joydev serio_raw snd_ens1371 btusb snd_ac97_codec uvcvideo btrtl videobuf2_vmalloc btbcm btintel gameport snd_rawmidi videobuf2_memops bluetooth videobuf2_v4l2 videobuf2_core snd_seq_device ac97_bus snd_pcm videodev media ecdh_generic snd_timer snd soundcore shpchp mac_hid vmw_vsock_vmci_transport vsock vmw_vmci sch_fq_codel ib_iser rdma_cm iw_cm ib_cm ib_core iscsi_tcp libiscsi_tcp libiscsi scsi_transport_iscsi ip_tables x_tables autofs4 btrfs zstd_compress raid10 raid456 async_raid6_recov async_memcpy async_pq async_xor async_tx xor raid6_pq libcrc32c raid1 raid0 multipath linear hid_generic usbhid hid crct10dif_pclmul crc32_pclmul ghash_clmulni_intel pcbc aesni_intel aes_x86_64
[ 1087.577770]  crypto_simd glue_helper cryptd vmwgfx psmouse ttm drm_kms_helper mptspi mptscsih ahci libahci e1000 mptbase scsi_transport_spi syscopyarea sysfillrect sysimgblt fb_sys_fops drm i2c_piix4 pata_acpi
[ 1087.579878] CPU: 0 PID: 52410 Comm: poc Not tainted 4.15.0-12-generic #13-Ubuntu
[ 1087.580479] Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 07/02/2015
[ 1087.581694] RIP: 0010:          (null)
[ 1087.582306] RSP: 0018:ffffb9074151fc40 EFLAGS: 00010206
[ 1087.582898] RAX: 0000000000000000 RBX: ffff8b4b8e9001e0 RCX: 0000000000000000
[ 1087.583500] RDX: 0000000000000001 RSI: ffff8b4bb8d3dec0 RDI: ffff8b4b8e900138
[ 1087.584059] RBP: ffffb9074151fca0 R08: 0000000000000000 R09: ffff8b4bb8d3dec0
[ 1087.584511] R10: ffff8b4bb5ac702a R11: 14077336e65e3568 R12: ffff8b4bb8d3d440
[ 1087.584961] R13: 0000000000000001 R14: ffff8b4b8e900138 R15: ffff8b4bb8d3dec0
[ 1087.585414] FS:  00007fa7fb3ce540(0000) GS:ffff8b4bbc600000(0000) knlGS:0000000000000000
[ 1087.585862] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1087.586301] CR2: 0000000000000000 CR3: 000000003a242006 CR4: 00000000001606f0
[ 1087.586774] Call Trace:
[ 1087.587211]  ? lookup_slow+0xab/0x170
[ 1087.587650]  walk_component+0x1c3/0x350
[ 1087.588184]  ? __xfs_trans_commit+0xa0/0x260 [xfs]
[ 1087.588676]  path_lookupat+0x84/0x1f0
[ 1087.589731]  ? __cgroup_account_cputime+0x28/0x30
[ 1087.590341]  filename_lookup+0xb6/0x190
[ 1087.590995]  ? __check_object_size+0xaf/0x1b0
[ 1087.591921]  ? strncpy_from_user+0x4d/0x170
[ 1087.592413]  user_path_at_empty+0x36/0x40
[ 1087.592897]  ? user_path_at_empty+0x36/0x40
[ 1087.593384]  path_listxattr+0x4b/0xb0
[ 1087.593807]  SyS_listxattr+0x13/0x20
[ 1087.594246]  do_syscall_64+0x73/0x130
[ 1087.594916]  entry_SYSCALL_64_after_hwframe+0x3d/0xa2
[ 1087.595314] RIP: 0033:0x7fa7faedf907
[ 1087.595675] RSP: 002b:00007ffcb19b4c58 EFLAGS: 00000283 ORIG_RAX: 00000000000000c2
[ 1087.596038] RAX: ffffffffffffffda RBX: 0000000000000000 RCX: 00007fa7faedf907
[ 1087.596394] RDX: 0000000000000071 RSI: 00007ffcb19b4d40 RDI: 0000561f96699310
[ 1087.596740] RBP: 00007ffcb19b4dc0 R08: 0000561f966992f0 R09: 0000000000000000
[ 1087.597152] R10: 0000000000000000 R11: 0000000000000283 R12: 0000561f9544ed30
[ 1087.597483] R13: 00007ffcb19b4ec0 R14: 0000000000000000 R15: 0000000000000000
[ 1087.597804] Code:  Bad RIP value.
[ 1087.598119] RIP:           (null) RSP: ffffb9074151fc40
[ 1087.598477] CR2: 0000000000000000
[ 1087.598813] ---[ end trace 1ecf08f3cdf242f0 ]---
Comment 4 Wen Xu 2018-04-12 03:32:55 UTC
For 4.16, I uses 
commit 7928b2cbe55b2a410a0f5c1f154610059c57b1b2
Author: Linus Torvalds <torvalds@linux-foundation.org>
Date:   Sun Feb 11 15:04:29 2018 -0800

    Linux 4.16-rc1

Here is the config:
test@ubuntu:~/ext4$ cat /boot/config-4.16.0-rc1+ | grep XFS
CONFIG_XFS_FS=y
CONFIG_XFS_QUOTA=y
CONFIG_XFS_POSIX_ACL=y
CONFIG_XFS_RT=y
CONFIG_XFS_ONLINE_SCRUB=y
# CONFIG_XFS_WARN is not set
# CONFIG_XFS_DEBUG is not set
# CONFIG_VXFS_FS is not set

If you want to check others, I will dump for you!
Comment 5 Wen Xu 2018-04-12 03:34:13 UTC
Sorry! The md5sum of the image is c467a649a12c5ea5d5e93c487a71131e.
Comment 6 Eric Sandeen 2018-04-12 04:02:31 UTC
That's not your kernel config; that is only a tiny portion of your config.  Attaching your actual, full config to these bugs may be helpful, so that the problems you encounter can be reliably reproduced.  Based on what I see, it looks like your test kernel may not have list debugging enabled.  Please attach your full kernel config for reference.
Comment 7 Wen Xu 2018-04-12 04:06:49 UTC
Created attachment 275331 [details]
config-4.16.0-rc1+

I see, here it is.
Comment 8 Dave Chinner 2018-04-13 04:11:18 UTC
This doesn't reproduce on a TOT kernel because it detects other corruptions in the inode buffer and so prevents teh code from exercising whatever is being tripped over here. Can you please reproduce this on a TOT kernel and attach new images and POC code.

The corruption verifiers and hardening code is a quickly moving target in XFS at the moment, so you really need to be fuzzing the bleeding edge dev kernels to catch stuff before other XFS developers do....

-Dave.
Comment 9 Wen Xu 2018-04-13 04:46:49 UTC
Created attachment 275345 [details]
The (compressed) crafted image which causes crash

Hi Dave, could you please try this image? 

# mkdir mnt
# mount -t xfs 53.img mnt
# gcc -o poc poc.c
# ./poc ./mnt

I uses this version to reproduce:

commit 16e205cf42da1f497b10a4a24f563e6c0d574eec
Merge: affb028 a10beab
Author: Linus Torvalds <torvalds@linux-foundation.org>
Date:   Thu Apr 12 20:56:10 2018 -0700

    Merge tag 'drm-fixes-for-v4.17-rc1' of git://people.freedesktop.org/~airlied/linux

Kernel message:
[  154.995063] XFS (loop0): Mounting V4 Filesystem
[  154.997133] XFS (loop0): Ending clean mount
[  160.384068] XFS (loop0): Metadata corruption detected at xfs_buf_ioend+0x56/0x1d0, xfs_inode block 0x1b00 xfs_inode_buf_verify
[  160.385875] XFS (loop0): Unmount and run xfs_repair
[  160.386641] XFS (loop0): First 128 bytes of corrupted metadata buffer:
[  160.387687] 000000006691f35a: 49 4e 41 ed 02 01 00 00 00 00 00 00 00 00 00 00  INA.............
[  160.389072] 000000006bd52143: 15 00 00 02 00 00 00 00 00 00 00 00 00 00 00 01  ................
[  160.391452] 000000000a915b74: 5a b2 9b 71 07 03 42 1b 5a b2 9b 71 07 7d 54 6d  Z..q..B.Z..q.}Tm
[  160.392850] 00000000ca4be198: 5a b2 9b 71 07 7d 54 6d 00 00 00 00 00 00 00 59  Z..q.}Tm.......Y
[  160.394216] 000000009412a18a: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
[  160.395587] 00000000fcedee95: 00 00 00 02 00 00 00 00 00 00 00 00 00 00 00 00  ................
[  160.396982] 0000000006278573: ff 00 ff ff 07 00 00 00 06 bf 03 00 30 62 61 7a  ............0baz
[  160.398345] 000000003e07be63: 01 00 00 06 c1 05 00 40 78 61 74 74 72 01 00 00  .......@xattr...
[  160.399727] XFS (loop0): metadata I/O error in "xfs_trans_read_buf_map" at daddr 0x1b00 len 16 error 117
[  160.401958] XFS (loop0): xfs_imap_to_bp: xfs_trans_read_buf() returned error -117.
[  160.402491] XFS (loop0): Metadata corruption detected at xfs_buf_ioend+0x56/0x1d0, xfs_inode block 0x1b00 xfs_inode_buf_verify
[  160.404367] XFS (loop0): Unmount and run xfs_repair
[  160.405797] XFS (loop0): First 128 bytes of corrupted metadata buffer:
[  160.407188] 00000000da1e6dc6: 49 4e 41 ed 02 01 00 00 00 00 00 00 00 00 00 00  INA.............
[  160.408601] 000000008e853d8c: 15 00 00 02 00 00 00 00 00 00 00 00 00 00 00 01  ................
[  160.409969] 00000000ae81435b: 5a b2 9b 71 07 03 42 1b 5a b2 9b 71 07 7d 54 6d  Z..q..B.Z..q.}Tm
[  160.412162] 000000001b28b3ab: 5a b2 9b 71 07 7d 54 6d 00 00 00 00 00 00 00 59  Z..q.}Tm.......Y
[  160.413641] 000000007f4a8866: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
[  160.415060] 00000000dc7dadb8: 00 00 00 02 00 00 00 00 00 00 00 00 00 00 00 00  ................
[  160.416895] 0000000020e200d7: ff 00 ff ff 07 00 00 00 06 bf 03 00 30 62 61 7a  ............0baz
[  160.418310] 000000009fa38e58: 01 00 00 06 c1 05 00 40 78 61 74 74 72 01 00 00  .......@xattr...
[  160.419916] XFS (loop0): metadata I/O error in "xfs_trans_read_buf_map" at daddr 0x1b00 len 16 error 117
[  160.421579] XFS (loop0): xfs_imap_to_bp: xfs_trans_read_buf() returned error -117.
[  160.423367] XFS (loop0): Metadata corruption detected at xfs_buf_ioend+0x56/0x1d0, xfs_inode block 0x1b00 xfs_inode_buf_verify
[  160.425311] XFS (loop0): Unmount and run xfs_repair
[  160.426075] XFS (loop0): First 128 bytes of corrupted metadata buffer:
[  160.427056] 000000006691f35a: 49 4e 41 ed 02 01 00 00 00 00 00 00 00 00 00 00  INA.............
[  160.428573] 000000006bd52143: 15 00 00 02 00 00 00 00 00 00 00 00 00 00 00 01  ................
[  160.430032] 000000000a915b74: 5a b2 9b 71 07 03 42 1b 5a b2 9b 71 07 7d 54 6d  Z..q..B.Z..q.}Tm
[  160.440711] 00000000ca4be198: 5a b2 9b 71 07 7d 54 6d 00 00 00 00 00 00 00 59  Z..q.}Tm.......Y
[  160.442076] 000000009412a18a: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
[  160.443457] 00000000fcedee95: 00 00 00 02 00 00 00 00 00 00 00 00 00 00 00 00  ................
[  160.444829] 0000000006278573: ff 00 ff ff 07 00 00 00 06 bf 03 00 30 62 61 7a  ............0baz
[  160.446188] 000000003e07be63: 01 00 00 06 c1 05 00 40 78 61 74 74 72 01 00 00  .......@xattr...
[  160.447609] XFS (loop0): metadata I/O error in "xfs_trans_read_buf_map" at daddr 0x1b00 len 16 error 117
[  160.449167] XFS (loop0): xfs_imap_to_bp: xfs_trans_read_buf() returned error -117.
[  160.450058] XFS (loop0): Metadata corruption detected at xfs_buf_ioend+0x56/0x1d0, xfs_inode block 0x1b00 xfs_inode_buf_verify
[  160.451907] XFS (loop0): Unmount and run xfs_repair
[  160.452717] XFS (loop0): First 128 bytes of corrupted metadata buffer:
[  160.453739] 00000000da1e6dc6: 49 4e 41 ed 02 01 00 00 00 00 00 00 00 00 00 00  INA.............
[  160.455078] 000000008e853d8c: 15 00 00 02 00 00 00 00 00 00 00 00 00 00 00 01  ................
[  160.456465] 00000000ae81435b: 5a b2 9b 71 07 03 42 1b 5a b2 9b 71 07 7d 54 6d  Z..q..B.Z..q.}Tm
[  160.457822] 000000001b28b3ab: 5a b2 9b 71 07 7d 54 6d 00 00 00 00 00 00 00 59  Z..q.}Tm.......Y
[  160.459182] 000000007f4a8866: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
[  160.460572] 00000000dc7dadb8: 00 00 00 02 00 00 00 00 00 00 00 00 00 00 00 00  ................
[  160.461938] 0000000020e200d7: ff 00 ff ff 07 00 00 00 06 bf 03 00 30 62 61 7a  ............0baz
[  160.463299] 000000009fa38e58: 01 00 00 06 c1 05 00 40 78 61 74 74 72 01 00 00  .......@xattr...
[  160.464698] XFS (loop0): metadata I/O error in "xfs_trans_read_buf_map" at daddr 0x1b00 len 16 error 117
[  160.466180] XFS (loop0): xfs_imap_to_bp: xfs_trans_read_buf() returned error -117.
[  160.466478] XFS (loop0): Metadata corruption detected at xfs_buf_ioend+0x56/0x1d0, xfs_inode block 0x1b00 xfs_inode_buf_verify
[  160.468310] XFS (loop0): Unmount and run xfs_repair
[  160.469091] XFS (loop0): First 128 bytes of corrupted metadata buffer:
[  160.470126] 000000006691f35a: 49 4e 41 ed 02 01 00 00 00 00 00 00 00 00 00 00  INA.............
[  160.471500] 000000006bd52143: 15 00 00 02 00 00 00 00 00 00 00 00 00 00 00 01  ................
[  160.472891] 000000000a915b74: 5a b2 9b 71 07 03 42 1b 5a b2 9b 71 07 7d 54 6d  Z..q..B.Z..q.}Tm
[  160.474255] 00000000ca4be198: 5a b2 9b 71 07 7d 54 6d 00 00 00 00 00 00 00 59  Z..q.}Tm.......Y
[  160.475623] 000000009412a18a: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
[  160.477010] 00000000fcedee95: 00 00 00 02 00 00 00 00 00 00 00 00 00 00 00 00  ................
[  160.478366] 0000000006278573: ff 00 ff ff 07 00 00 00 06 bf 03 00 30 62 61 7a  ............0baz
[  160.479728] 000000003e07be63: 01 00 00 06 c1 05 00 40 78 61 74 74 72 01 00 00  .......@xattr...
[  160.481123] XFS (loop0): metadata I/O error in "xfs_trans_read_buf_map" at daddr 0x1b00 len 16 error 117
[  160.482628] XFS (loop0): xfs_imap_to_bp: xfs_trans_read_buf() returned error -117.
[  160.482809] BUG: unable to handle kernel NULL pointer dereference at 0000000000000000
[  160.484073] PGD 80000002330a2067 P4D 80000002330a2067 PUD 233853067 PMD 0
[  160.485148] Oops: 0010 [#1] SMP PTI
[  160.485717] Modules linked in: snd_hda_codec_generic snd_hda_intel snd_hda_codec snd_hwdep snd_hda_core snd_pcm snd_timer snd soundcore i2c_piix4 mac_hid ib_iser rdma_cm iw_cm ib_cm ib_core iscsi_tcp libiscsi_tcp libiscsi scsi_transport_iscsi autofs4 raid10 raid456 async_raid6_recov async_memcpy async_pq async_xor async_tx raid1 raid0 multipath linear qxl drm_kms_helper syscopyarea sysfillrect sysimgblt fb_sys_fops ttm drm crct10dif_pclmul crc32_pclmul 8139too aesni_intel aes_x86_64 crypto_simd cryptd glue_helper floppy 8139cp mii pata_acpi
[  160.493367] CPU: 1 PID: 1482 Comm: poc Not tainted 4.16.0+ #1
[  160.494269] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Ubuntu-1.8.2-1ubuntu1 04/01/2014
[  160.495757] RIP: 0010:          (null)
[  160.496352] RSP: 0018:ffffa0f901663c10 EFLAGS: 00010206
[  160.497173] RAX: 0000000000000000 RBX: ffff8bd2b36ff540 RCX: 0000000000000000
[  160.498285] RDX: 0000000000000011 RSI: ffff8bd2b36ff540 RDI: ffff8bd2b36684f8
[  160.499414] RBP: ffff8bd2b36684f8 R08: ffffc0f8ffc69f80 R09: 0000000000000000
[  160.500528] R10: 61c8864680b583eb R11: 00000000e34ee8a9 R12: ffff8bd2b3fefd80
[  160.501644] R13: ffffa0f901663de0 R14: 0000000000000011 R15: 0000000000000000
[  160.502756] FS:  00007f486d655700(0000) GS:ffff8bd2bfc40000(0000) knlGS:0000000000000000
[  160.504029] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[  160.504951] CR2: 0000000000000000 CR3: 0000000233a22000 CR4: 00000000000006e0
[  160.506101] Call Trace:
[  160.506542]  ? __lookup_slow+0x97/0x150
[  160.507177]  ? lookup_slow+0x35/0x50
[  160.507750]  ? walk_component+0x1c6/0x350
[  160.508389]  ? link_path_walk+0x184/0x510
[  160.509038]  ? __xfs_trans_commit+0x9a/0x250
[  160.509718]  ? path_lookupat+0x9c/0x230
[  160.510331]  ? filename_lookup+0xb8/0x1a0
[  160.510971]  ? iput+0x6f/0x1f0
[  160.511497]  ? _cond_resched+0x16/0x40
[  160.512111]  ? kmem_cache_alloc+0x160/0x1d0
[  160.512791]  ? path_listxattr+0x41/0xa0
[  160.513404]  ? path_listxattr+0x41/0xa0
[  160.514035]  ? do_syscall_64+0x6e/0x120
[  160.514659]  ? entry_SYSCALL_64_after_hwframe+0x3d/0xa2
[  160.515493] Code:  Bad RIP value.
[  160.516041] RIP:           (null) RSP: ffffa0f901663c10
[  160.516856] CR2: 0000000000000000
[  160.517422] ---[ end trace 990c8b847d277377 ]---
Comment 10 Wen Xu 2018-04-15 05:57:29 UTC
Created attachment 275379 [details]
Simplified poc for 53.img to trigger the bug

Here is a simplified poc to reproduce the bug with mounted 53.img.
Comment 11 Darrick J. Wong 2018-06-02 16:33:04 UTC
Does this still reproduce in for-next?  I think Dave's series to inspect newly allocated inode records for signs that the inode isn't really free fixes this.

https://git.kernel.org/pub/scm/fs/xfs/xfs-linux.git/commit/?h=for-next&id=afca6c5b2595fc44383919fba740c194b0b76aff
Comment 12 Wen Xu 2018-06-02 17:11:29 UTC
(In reply to Darrick J. Wong from comment #11)
> Does this still reproduce in for-next?  I think Dave's series to inspect
> newly allocated inode records for signs that the inode isn't really free
> fixes this.
> 
> https://git.kernel.org/pub/scm/fs/xfs/xfs-linux.git/commit/?h=for-
> next&id=afca6c5b2595fc44383919fba740c194b0b76aff

I see, sorry for that I did not follow the recent commit.

Note You need to log in before you can comment on or make changes to this bug.