WARNING in blk_mq_init_sched

[syzbot]

Hello,

syzbot found the following crash on:

HEAD commit: f7c3bf8f Merge tag 'gfs2-for-5.4' of git://git.kernel.org/..
git tree: upstream
console output: https://syzkaller.appspot.com/x/log.txt?x=15f5baf9600000
kernel config: https://syzkaller.appspot.com/x/.config?x=50d4af03d68a470c
dashboard link: https://syzkaller.appspot.com/bug?extid=b2c197f98f86543b69c8
compiler: clang version 9.0.0 (/home/glider/llvm/clang
80fee25776c2fb61e74c1ecb1a523375c2500b69)

Unfortunately, I don't have any reproducer for this crash yet.

IMPORTANT: if you fix the bug, please add the following tag to the commit:
Reported-by: syzbot+b2c197...@syzkaller.appspotmail.com

------------[ cut here ]------------
WARNING: CPU: 1 PID: 25817 at block/blk-mq-sched.c:558
blk_mq_sched_free_requests block/blk-mq-sched.c:558 [inline]
WARNING: CPU: 1 PID: 25817 at block/blk-mq-sched.c:558
blk_mq_init_sched+0xad6/0xc00 block/blk-mq-sched.c:543
Kernel panic - not syncing: panic_on_warn set ...
CPU: 1 PID: 25817 Comm: syz-executor.4 Not tainted 5.3.0+ #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS
Google 01/01/2011
Call Trace:
__dump_stack lib/dump_stack.c:77 [inline]
dump_stack+0x1d8/0x2f8 lib/dump_stack.c:113
panic+0x25c/0x799 kernel/panic.c:219
__warn+0x22f/0x230 kernel/panic.c:576
report_bug+0x190/0x290 lib/bug.c:186
fixup_bug arch/x86/kernel/traps.c:179 [inline]
do_error_trap+0xd7/0x440 arch/x86/kernel/traps.c:272
do_invalid_op+0x36/0x40 arch/x86/kernel/traps.c:291
invalid_op+0x23/0x30 arch/x86/entry/entry_64.S:1028
RIP: 0010:blk_mq_sched_free_requests block/blk-mq-sched.c:558 [inline]
RIP: 0010:blk_mq_init_sched+0xad6/0xc00 block/blk-mq-sched.c:543
Code: f6 e8 9e 03 00 00 49 83 c6 10 4c 89 f7 e8 82 08 37 04 e9 ce fd ff ff
e8 c8 81 3f fe 48 c7 c7 72 5c 35 88 31 c0 e8 1d ae 28 fe <0f> 0b e9 ce f9
ff ff e8 ae 81 3f fe 48 c7 c7 72 5c 35 88 31 c0 e8
RSP: 0018:ffff88802225fbb8 EFLAGS: 00010246
RAX: 0000000000000024 RBX: 0000000000000000 RCX: 489d2508ed9c7100
RDX: ffffc9000e9a6000 RSI: 0000000000009af7 RDI: 0000000000009af8
RBP: ffff88802225fc50 R08: ffffffff815c9744 R09: ffffed1015d66090
R10: ffffed1015d66090 R11: 0000000000000000 R12: dffffc0000000000
R13: ffff888026958990 R14: ffff888026958080 R15: ffff8880269580d0
elevator_init_mq+0x317/0x450 block/elevator.c:719
__device_add_disk+0x6d/0x1140 block/genhd.c:705
device_add_disk+0x2a/0x40 block/genhd.c:763
add_disk include/linux/genhd.h:429 [inline]
loop_add+0x5d1/0x780 drivers/block/loop.c:2051
loop_control_ioctl+0x422/0x640 drivers/block/loop.c:2174
do_vfs_ioctl+0x744/0x1730 fs/ioctl.c:46
ksys_ioctl fs/ioctl.c:713 [inline]
__do_sys_ioctl fs/ioctl.c:720 [inline]
__se_sys_ioctl fs/ioctl.c:718 [inline]
__x64_sys_ioctl+0xe3/0x120 fs/ioctl.c:718
do_syscall_64+0xf7/0x1c0 arch/x86/entry/common.c:290
entry_SYSCALL_64_after_hwframe+0x49/0xbe
RIP: 0033:0x459a09
Code: fd b7 fb ff c3 66 2e 0f 1f 84 00 00 00 00 00 66 90 48 89 f8 48 89 f7
48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff
ff 0f 83 cb b7 fb ff c3 66 2e 0f 1f 84 00 00 00 00
RSP: 002b:00007fce60497c78 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
RAX: ffffffffffffffda RBX: 0000000000000003 RCX: 0000000000459a09
RDX: 0000000000000000 RSI: 0000000000004c80 RDI: 0000000000000006
RBP: 000000000075bfc8 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 00007fce604986d4
R13: 00000000004c3118 R14: 00000000004d69f8 R15: 00000000ffffffff
Kernel Offset: disabled
Rebooting in 86400 seconds..


---
This bug 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 bug report. See:
https://goo.gl/tpsmEJ#status for how to communicate with syzbot.

[Damien Le Moal]

On 2019/09/25 9:56, syzbot wrote:
> Hello,
>
> syzbot found the following crash on:
>
> HEAD commit: f7c3bf8f Merge tag 'gfs2-for-5.4' of git://git.kernel.org/..
> git tree: upstream
> console output: https://syzkaller.appspot.com/x/log.txt?x=15f5baf9600000
> kernel config: https://syzkaller.appspot.com/x/.config?x=50d4af03d68a470c
> dashboard link: https://syzkaller.appspot.com/bug?extid=b2c197f98f86543b69c8
> compiler: clang version 9.0.0 (/home/glider/llvm/clang
> 80fee25776c2fb61e74c1ecb1a523375c2500b69)
>
> Unfortunately, I don't have any reproducer for this crash yet.
>
> IMPORTANT: if you fix the bug, please add the following tag to the commit:
> Reported-by: syzbot+b2c197...@syzkaller.appspotmail.com

Oh... When the queue is initialized and the elevator initialization done by
elevator_init_mq() is executed without the queue sysfs lock held. In that step,
if the elevator initialization fails, blk_mq_sched_free_requests() is called and
will trip on the lockdep_assert_held(&q->sysfs_lock) check on entry. I guess
that is what is causing the crash ? But I thought lockdep_assert_held() only
spits out warnings...

Ming,

Your patch c48dac137a62 ("block: don't hold q->sysfs_lock in elevator_init_mq")
removed the sysfs_lock use in elevator_init_mq(). With that, should we move the
lockdep_assert_held(&q->sysfs_lock) call out of blk_mq_sched_free_requests() and
directly call it lockdep before calling that function (that's ugly) or do you
see a nice trick for handling the special case that is the first initialization ?

Cheers.

--
Damien Le Moal
Western Digital Research

[Damien Le Moal]

Please ignore. It looks like the gfs2 tree tested does not have commit
954b4a5ce4a8 ("block: Change elevator_init_mq() to always succeed") which
removes the possibility of having blk_mq_sched_free_requests() being called
during the first elevator initialization without the sysfs lock being held.

So if the crash is indeed triggered by the lockdep_assert_held() call, then this
problem will be fixed after a rebase on 5.4-rc1.

[Eric Biggers]

No, as the report says, this occurred on commit f7c3bf8f. Commit 954b4a5ce4a8
("block: Change elevator_init_mq() to always succeed") was already merged then.

- Eric

[syzbot]

Auto-closing this bug as obsolete.
Crashes did not happen for a while, no reproducer and no activity.