[syzbot] [btrfs?] WARNING in __btrfs_free_extent (2)

[syzbot]

Hello,

syzbot found the following issue on:

HEAD commit: 2ccbdf43d5e7 Merge tag 'for-linus' of git://git.kernel.org..
git tree: upstream
console+strace: https://syzkaller.appspot.com/x/log.txt?x=10aed446980000
kernel config: https://syzkaller.appspot.com/x/.config?x=c79815c08cc14227
dashboard link: https://syzkaller.appspot.com/bug?extid=fe3566bcb509ae7764ef
compiler: Debian clang version 15.0.6, GNU ld (GNU Binutils for Debian) 2.40
syz repro: https://syzkaller.appspot.com/x/repro.syz?x=10ebe941980000
C reproducer: https://syzkaller.appspot.com/x/repro.c?x=1374b3de980000

Downloadable assets:
disk image: https://storage.googleapis.com/syzbot-assets/35e32e9073a7/disk-2ccbdf43.raw.xz
vmlinux: https://storage.googleapis.com/syzbot-assets/6c6e34658d16/vmlinux-2ccbdf43.xz
kernel image: https://storage.googleapis.com/syzbot-assets/4417e7ef76ed/bzImage-2ccbdf43.xz
mounted in repro: https://storage.googleapis.com/syzbot-assets/b1cd2f1f0c7e/mount_0.gz

The issue was bisected to:

commit cecbb533b5fcec4ff77e786b7f94457f6cacd9e7
Author: Boris Burkov <bo...@bur.io>
Date: Wed Jun 28 18:00:15 2023 +0000

btrfs: record simple quota deltas in delayed refs

bisection log: https://syzkaller.appspot.com/x/bisect.txt?x=15a47be2980000
final oops: https://syzkaller.appspot.com/x/report.txt?x=17a47be2980000
console output: https://syzkaller.appspot.com/x/log.txt?x=13a47be2980000

IMPORTANT: if you fix the issue, please add the following tag to the commit:
Reported-by: syzbot+fe3566...@syzkaller.appspotmail.com
Fixes: cecbb533b5fc ("btrfs: record simple quota deltas in delayed refs")

BTRFS: Transaction aborted (error -2)
WARNING: CPU: 0 PID: 5085 at fs/btrfs/extent-tree.c:2984 do_free_extent_accounting fs/btrfs/extent-tree.c:2984 [inline]
WARNING: CPU: 0 PID: 5085 at fs/btrfs/extent-tree.c:2984 __btrfs_free_extent+0x32d1/0x3a10 fs/btrfs/extent-tree.c:3358
Modules linked in:
CPU: 0 PID: 5085 Comm: syz-executor845 Not tainted 6.10.0-rc3-syzkaller-00044-g2ccbdf43d5e7 #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 06/07/2024
RIP: 0010:do_free_extent_accounting fs/btrfs/extent-tree.c:2984 [inline]
RIP: 0010:__btrfs_free_extent+0x32d1/0x3a10 fs/btrfs/extent-tree.c:3358
Code: e8 64 b0 b0 fd 90 0f 0b 90 90 e9 3c f3 ff ff e8 b5 81 ee fd 90 48 c7 c7 00 2e 0b 8c 44 8b 6c 24 18 44 89 ee e8 40 b0 b0 fd 90 <0f> 0b 90 90 4c 8b 24 24 e9 4f f3 ff ff e8 8d 81 ee fd 90 48 c7 c7
RSP: 0018:ffffc9000352f220 EFLAGS: 00010246
RAX: 7e1377ca92db5900 RBX: ffff888024e3c001 RCX: ffff88807c1f0000
RDX: 0000000000000000 RSI: 0000000000000001 RDI: 0000000000000000
RBP: ffffc9000352f3f0 R08: ffffffff81585742 R09: fffffbfff1c39994
R10: dffffc0000000000 R11: fffffbfff1c39994 R12: dffffc0000000000
R13: 00000000fffffffe R14: 0000000000000000 R15: ffff888063a8b5c8
FS: 000055556fa3c3c0(0000) GS:ffff8880b9400000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00005648b40a4798 CR3: 0000000023a7e000 CR4: 00000000003506f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
<TASK>
run_delayed_tree_ref fs/btrfs/extent-tree.c:1736 [inline]
run_one_delayed_ref fs/btrfs/extent-tree.c:1762 [inline]
btrfs_run_delayed_refs_for_head fs/btrfs/extent-tree.c:2027 [inline]
__btrfs_run_delayed_refs+0x117c/0x4670 fs/btrfs/extent-tree.c:2097
btrfs_run_delayed_refs+0xe3/0x2c0 fs/btrfs/extent-tree.c:2209
btrfs_commit_transaction+0xf5d/0x3740 fs/btrfs/transaction.c:2400
sync_filesystem+0x1c8/0x230 fs/sync.c:66
generic_shutdown_super+0x72/0x2d0 fs/super.c:621
kill_anon_super+0x3b/0x70 fs/super.c:1226
btrfs_kill_super+0x41/0x50 fs/btrfs/super.c:2096
deactivate_locked_super+0xc4/0x130 fs/super.c:473
cleanup_mnt+0x41f/0x4b0 fs/namespace.c:1267
task_work_run+0x24f/0x310 kernel/task_work.c:180
ptrace_notify+0x2d2/0x380 kernel/signal.c:2402
ptrace_report_syscall include/linux/ptrace.h:415 [inline]
ptrace_report_syscall_exit include/linux/ptrace.h:477 [inline]
syscall_exit_work+0xc6/0x190 kernel/entry/common.c:173
syscall_exit_to_user_mode_prepare kernel/entry/common.c:200 [inline]
__syscall_exit_to_user_mode_work kernel/entry/common.c:205 [inline]
syscall_exit_to_user_mode+0x273/0x370 kernel/entry/common.c:218
do_syscall_64+0x100/0x230 arch/x86/entry/common.c:89
entry_SYSCALL_64_after_hwframe+0x77/0x7f
RIP: 0033:0x7fda22d8de67
Code: 08 00 48 83 c4 08 5b 5d c3 66 2e 0f 1f 84 00 00 00 00 00 c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 b8 a6 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 01 c3 48 c7 c2 b0 ff ff ff f7 d8 64 89 02 b8
RSP: 002b:00007fffaba15888 EFLAGS: 00000206 ORIG_RAX: 00000000000000a6
RAX: 0000000000000000 RBX: 0000000000000000 RCX: 00007fda22d8de67
RDX: 0000000000000000 RSI: 0000000000000009 RDI: 00007fffaba15940
RBP: 00007fffaba15940 R08: 0000000000000000 R09: 0000000000000000
R10: 00000000ffffffff R11: 0000000000000206 R12: 00007fffaba16a00
R13: 000055556fa3d700 R14: 431bde82d7b634db R15: 00007fffaba169a4
</TASK>


---
This report is generated by a bot. It may contain errors.
See https://goo.gl/tpsmEJ for more information about syzbot.
syzbot engineers can be reached at syzk...@googlegroups.com.

syzbot will keep track of this issue. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.
For information about bisection process see: https://goo.gl/tpsmEJ#bisection

If the report is already addressed, let syzbot know by replying with:
#syz fix: exact-commit-title

If you want syzbot to run the reproducer, reply with:
#syz test: git://repo/address.git branch-or-commit-hash
If you attach or paste a git patch, syzbot will apply it before testing.

If you want to overwrite report's subsystems, reply with:
#syz set subsystems: new-subsystem
(See the list of subsystem names on the web dashboard)

If the report is a duplicate of another one, reply with:
#syz dup: exact-subject-of-another-report

If you want to undo deduplication, reply with:
#syz undup

[Lizhi Xu]

#syz test: upstream 2ccbdf43d5e7

diff --git a/fs/btrfs/extent-tree.c b/fs/btrfs/extent-tree.c
index 3774c191e36d..8cab95333f35 100644
--- a/fs/btrfs/extent-tree.c
+++ b/fs/btrfs/extent-tree.c
@@ -2979,10 +2979,12 @@ static int do_free_extent_accounting(struct btrfs_trans_handle *trans,
}
}

- ret = btrfs_record_squota_delta(trans->fs_info, delta);
- if (ret) {
- btrfs_abort_transaction(trans, ret);
- return ret;
+ if (btrfs_qgroup_enabled(trans->fs_info)) {
+ ret = btrfs_record_squota_delta(trans->fs_info, delta);
+ if (ret) {
+ btrfs_abort_transaction(trans, ret);
+ return ret;
+ }
}

ret = add_to_free_space_tree(trans, bytenr, num_bytes);

[syzbot]

Hello,

syzbot has tested the proposed patch but the reproducer is still triggering an issue:
WARNING in __btrfs_free_extent

------------[ cut here ]------------

BTRFS: Transaction aborted (error -2)

WARNING: CPU: 1 PID: 5913 at fs/btrfs/extent-tree.c:2985 do_free_extent_accounting fs/btrfs/extent-tree.c:2985 [inline]
WARNING: CPU: 1 PID: 5913 at fs/btrfs/extent-tree.c:2985 __btrfs_free_extent+0x3564/0x3a40 fs/btrfs/extent-tree.c:3360
Modules linked in:
CPU: 1 PID: 5913 Comm: syz-executor.2 Not tainted 6.10.0-rc3-syzkaller-00044-g2ccbdf43d5e7-dirty #0

Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 06/07/2024

RIP: 0010:do_free_extent_accounting fs/btrfs/extent-tree.c:2985 [inline]
RIP: 0010:__btrfs_free_extent+0x3564/0x3a40 fs/btrfs/extent-tree.c:3360
Code: ff ff 48 89 df e8 9c f9 50 fe e9 9d e0 ff ff e8 22 7f ee fd 90 48 c7 c7 00 2e 0b 8c 44 8b 6c 24 18 44 89 ee e8 ad ad b0 fd 90 <0f> 0b 90 90 4c 8b 24 24 e9 13 f5 ff ff 48 8b 4c 24 20 80 e1 07 80
RSP: 0018:ffffc900042af340 EFLAGS: 00010246

RAX: 6c4a53fef53c7400 RBX: ffff88807135c001 RCX: ffff88802ae55a00

RDX: 0000000000000000 RSI: 0000000000000001 RDI: 0000000000000000

RBP: ffffc900042af510 R08: ffffffff81585742 R09: 1ffff92000855e08
R10: dffffc0000000000 R11: fffff52000855e09 R12: dffffc0000000000
R13: 00000000fffffffe R14: 0000000000000000 R15: ffff88806f7cf818
FS: 0000555588813480(0000) GS:ffff8880b9500000(0000) knlGS:0000000000000000

CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033

CR2: 000055558881c938 CR3: 000000007124e000 CR4: 00000000003506f0

DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
<TASK>
run_delayed_tree_ref fs/btrfs/extent-tree.c:1736 [inline]
run_one_delayed_ref fs/btrfs/extent-tree.c:1762 [inline]
btrfs_run_delayed_refs_for_head fs/btrfs/extent-tree.c:2027 [inline]
__btrfs_run_delayed_refs+0x117c/0x4670 fs/btrfs/extent-tree.c:2097
btrfs_run_delayed_refs+0xe3/0x2c0 fs/btrfs/extent-tree.c:2209
btrfs_commit_transaction+0xf5d/0x3740 fs/btrfs/transaction.c:2400
sync_filesystem+0x1c8/0x230 fs/sync.c:66
generic_shutdown_super+0x72/0x2d0 fs/super.c:621
kill_anon_super+0x3b/0x70 fs/super.c:1226
btrfs_kill_super+0x41/0x50 fs/btrfs/super.c:2096
deactivate_locked_super+0xc4/0x130 fs/super.c:473
cleanup_mnt+0x41f/0x4b0 fs/namespace.c:1267
task_work_run+0x24f/0x310 kernel/task_work.c:180

resume_user_mode_work include/linux/resume_user_mode.h:50 [inline]
exit_to_user_mode_loop kernel/entry/common.c:114 [inline]
exit_to_user_mode_prepare include/linux/entry-common.h:328 [inline]
__syscall_exit_to_user_mode_work kernel/entry/common.c:207 [inline]
syscall_exit_to_user_mode+0x168/0x370 kernel/entry/common.c:218
do_syscall_64+0x100/0x230 arch/x86/entry/common.c:89
entry_SYSCALL_64_after_hwframe+0x77/0x7f
RIP: 0033:0x7f9668a7e257
Code: b0 ff ff ff f7 d8 64 89 01 48 83 c8 ff c3 0f 1f 44 00 00 31 f6 e9 09 00 00 00 66 0f 1f 84 00 00 00 00 00 b8 a6 00 00 00 0f 05 <48> 3d 00 f0 ff ff 77 01 c3 48 c7 c2 b0 ff ff ff f7 d8 64 89 02 b8
RSP: 002b:00007ffdf71b8e28 EFLAGS: 00000246 ORIG_RAX: 00000000000000a6
RAX: 0000000000000000 RBX: 0000000000000000 RCX: 00007f9668a7e257
RDX: 0000000000000000 RSI: 0000000000000009 RDI: 00007ffdf71b8ee0
RBP: 00007ffdf71b8ee0 R08: 0000000000000000 R09: 0000000000000000
R10: 00000000ffffffff R11: 0000000000000246 R12: 00007ffdf71b9f90
R13: 00007f9668ad96c6 R14: 0000000000019832 R15: 0000000000000002
</TASK>


Tested on:

commit: 2ccbdf43 Merge tag 'for-linus' of git://git.kernel.org..
git tree: git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
console output: https://syzkaller.appspot.com/x/log.txt?x=16868cd1980000

kernel config: https://syzkaller.appspot.com/x/.config?x=c79815c08cc14227
dashboard link: https://syzkaller.appspot.com/bug?extid=fe3566bcb509ae7764ef
compiler: Debian clang version 15.0.6, GNU ld (GNU Binutils for Debian) 2.40

patch: https://syzkaller.appspot.com/x/patch.diff?x=1666bda9980000

[Boris Burkov]

This is a legitimate issue. The basic idea is that when you enable
simple quotas on an existing fs, the existing subvols start at 0 usage.
Subsequently, we can add and remove extents and occasionally hit 0
again.

If 0 usage is the only saving grace for rejecting qgroup deletion, we
can hit a race where we pass the 0 check in the deletion code, then
record a positive squota delta, then actually delete the qgroup, then
delete and try to record the delete for that extent by which point the
qgroup is gone and we blow up on ENOENT.

The possible fixes I can think of are:
- lock something during qgroup deletion that record_squota_delta also
takes
- mark the qgroup ro after the point of no return on deleting it
- disallow deleting qgroups whose subvol still exists

The third option has already been implemented and is why this doesn't
reproduce on the tip of btrfs/for-next. The first option could be an
interesting option to be extra safe, even with the third in place.

Therefore,
#syz fix: btrfs: slightly loosen the requirement for qgroup removal