Re: [syzbot] INFO: task hung in io_uring_release (2)

[Pavel Begunkov <asml.silence@gmail.com>]

On 12/27/21 19:12, syzbot wrote:
> Hello,
> 
> syzbot found the following issue on:
> 
> HEAD commit:    bc491fb12513 Merge tag 'fixes-2021-12-22' of git://git.ker..
> git tree:       upstream
> console output: https://syzkaller.appspot.com/x/log.txt?x=165b1a5db00000
> kernel config:  https://syzkaller.appspot.com/x/.config?x=6104739ac5f067ea
> dashboard link: https://syzkaller.appspot.com/bug?extid=f4dee1c474a6e3f68100
> compiler:       Debian clang version 11.0.1-2, GNU ld (GNU Binutils for Debian) 2.35.2
> 
> Unfortunately, I don't have any reproducer for this issue yet.
> 
> IMPORTANT: if you fix the issue, please add the following tag to the commit:
> Reported-by: syzbot+f4dee1c474a6e3f68100@syzkaller.appspotmail.com

Neither io_uring_release() nor io_uring_del_tctx_node() can take
->uring_lock and so hung. Apparently, the lock is held by

1 lock held by syz-executor.2/8302:
   #0: ffff88801de6e0a8 (&ctx->uring_lock){+.+.}-{3:3}, at: tctx_task_work+0x2e5/0x560 fs/io_uring.c:2242


The most likely scenario is that some lower layer doesn't honour
NOWAIT/NOBLOCK flags/etc. E.g. I'd guess it can happen with
io_uring using O_NONBLOCK pipes.

Also possible that tctx_task_work() screwed locking somehow, but
it's 5.16 and related bits shouldn't have been changed for a couple
of releases and so less likely.

Would be lovely to find a repro and/or see what ops were issued/etc.


> INFO: task syz-executor.2:8284 blocked for more than 143 seconds.
>        Not tainted 5.16.0-rc6-syzkaller #0
> "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
> task:syz-executor.2  state:D stack:26704 pid: 8284 ppid:  3729 flags:0x00004004
> Call Trace:
>   <TASK>
>   context_switch kernel/sched/core.c:4972 [inline]
>   __schedule+0xb72/0x1460 kernel/sched/core.c:6253
>   schedule+0x12b/0x1f0 kernel/sched/core.c:6326
>   schedule_preempt_disabled+0xf/0x20 kernel/sched/core.c:6385
>   __mutex_lock_common+0xd1f/0x2590 kernel/locking/mutex.c:680
>   __mutex_lock kernel/locking/mutex.c:740 [inline]
>   mutex_lock_nested+0x1a/0x20 kernel/locking/mutex.c:792
>   io_ring_ctx_wait_and_kill+0xa0/0x36a fs/io_uring.c:9565
>   io_uring_release+0x59/0x63 fs/io_uring.c:9594
>   __fput+0x3fc/0x870 fs/file_table.c:280
>   task_work_run+0x146/0x1c0 kernel/task_work.c:164
>   tracehook_notify_resume include/linux/tracehook.h:189 [inline]
>   exit_to_user_mode_loop kernel/entry/common.c:175 [inline]
>   exit_to_user_mode_prepare+0x209/0x220 kernel/entry/common.c:207
>   __syscall_exit_to_user_mode_work kernel/entry/common.c:289 [inline]
>   syscall_exit_to_user_mode+0x2e/0x70 kernel/entry/common.c:300
>   do_syscall_64+0x53/0xd0 arch/x86/entry/common.c:86
>   entry_SYSCALL_64_after_hwframe+0x44/0xae
> RIP: 0033:0x7f7b18378adb
> RSP: 002b:00007ffe8c741150 EFLAGS: 00000293 ORIG_RAX: 0000000000000003
> RAX: 0000000000000000 RBX: 0000000000000004 RCX: 00007f7b18378adb
> RDX: 0000001b2cc20000 RSI: 0000000000000000 RDI: 0000000000000003
> RBP: 00007f7b184da960 R08: 0000000000000000 R09: 0000000000000010
> R10: 00007ffe8c76b0b8 R11: 0000000000000293 R12: 000000000007cb81
> R13: 00007f7b184dfb50 R14: 00007f7b184d9370 R15: 0000000000000005
>   </TASK>
> INFO: task syz-executor.2:8285 blocked for more than 143 seconds.
>        Not tainted 5.16.0-rc6-syzkaller #0
> "echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
> task:syz-executor.2  state:D stack:23728 pid: 8285 ppid:  3729 flags:0x00004004
> Call Trace:
>   <TASK>
>   context_switch kernel/sched/core.c:4972 [inline]
>   __schedule+0xb72/0x1460 kernel/sched/core.c:6253
>   schedule+0x12b/0x1f0 kernel/sched/core.c:6326
>   schedule_preempt_disabled+0xf/0x20 kernel/sched/core.c:6385
>   __mutex_lock_common+0xd1f/0x2590 kernel/locking/mutex.c:680
>   __mutex_lock kernel/locking/mutex.c:740 [inline]
>   mutex_lock_nested+0x1a/0x20 kernel/locking/mutex.c:792
>   io_uring_del_tctx_node+0xe4/0x2a9 fs/io_uring.c:9777
>   io_uring_clean_tctx+0x192/0x1d5 fs/io_uring.c:9793
>   io_uring_cancel_generic+0x629/0x671 fs/io_uring.c:9884
>   io_uring_files_cancel include/linux/io_uring.h:16 [inline]
>   do_exit+0x281/0x24f0 kernel/exit.c:787
>   do_group_exit+0x168/0x2d0 kernel/exit.c:929
>   get_signal+0x1740/0x2120 kernel/signal.c:2852
>   arch_do_signal_or_restart+0x9c/0x730 arch/x86/kernel/signal.c:868
>   handle_signal_work kernel/entry/common.c:148 [inline]
>   exit_to_user_mode_loop kernel/entry/common.c:172 [inline]
>   exit_to_user_mode_prepare+0x191/0x220 kernel/entry/common.c:207
>   __syscall_exit_to_user_mode_work kernel/entry/common.c:289 [inline]
>   syscall_exit_to_user_mode+0x2e/0x70 kernel/entry/common.c:300
>   do_syscall_64+0x53/0xd0 arch/x86/entry/common.c:86
>   entry_SYSCALL_64_after_hwframe+0x44/0xae
> RIP: 0033:0x7f7b183c5e99
> RSP: 002b:00007f7b16d3b218 EFLAGS: 00000246 ORIG_RAX: 00000000000000ca
> RAX: fffffffffffffe00 RBX: 00007f7b184d8f68 RCX: 00007f7b183c5e99
> RDX: 0000000000000000 RSI: 0000000000000080 RDI: 00007f7b184d8f68
> RBP: 00007f7b184d8f60 R08: 0000000000000000 R09: 0000000000000000
> R10: 0000000000000000 R11: 0000000000000246 R12: 00007f7b184d8f6c
> R13: 00007ffe8c7410ef R14: 00007f7b16d3b300 R15: 0000000000022000
>   </TASK>
> 
> Showing all locks held in the system:
> 1 lock held by khungtaskd/27:
>   #0: ffffffff8cb1de00 (rcu_read_lock){....}-{1:2}, at: rcu_lock_acquire+0x0/0x30
> 2 locks held by getty/3280:
>   #0: ffff88807f219098 (&tty->ldisc_sem){++++}-{0:0}, at: tty_ldisc_ref_wait+0x21/0x70 drivers/tty/tty_ldisc.c:252
>   #1: ffffc90002b962e8 (&ldata->atomic_read_lock){+.+.}-{3:3}, at: n_tty_read+0x6c5/0x1c60 drivers/tty/n_tty.c:2113
> 1 lock held by syz-executor.2/8284:
>   #0: ffff88801de6e0a8 (&ctx->uring_lock){+.+.}-{3:3}, at: io_ring_ctx_wait_and_kill+0xa0/0x36a fs/io_uring.c:9565
> 1 lock held by syz-executor.2/8285:
>   #0: ffff88801de6e0a8 (&ctx->uring_lock){+.+.}-{3:3}, at: io_uring_del_tctx_node+0xe4/0x2a9 fs/io_uring.c:9777
> 1 lock held by syz-executor.2/8302:
>   #0: ffff88801de6e0a8 (&ctx->uring_lock){+.+.}-{3:3}, at: tctx_task_work+0x2e5/0x560 fs/io_uring.c:2242
> 
[...]
-- 
Pavel Begunkov