Re: [PATCH 2/2] io_uring: fix cancellation taking mutex while TASK_UNINTERRUPTIBLE

[Jens Axboe <axboe@kernel.dk>]

On 1/26/21 8:56 AM, Pavel Begunkov wrote:
> On 26/01/2021 15:28, Pavel Begunkov wrote:
>> do not call blocking ops when !TASK_RUNNING; state=2 set at
>> 	[<00000000ced9dbfc>] prepare_to_wait+0x1f4/0x3b0
>> 	kernel/sched/wait.c:262
>> WARNING: CPU: 1 PID: 19888 at kernel/sched/core.c:7853
>> 	__might_sleep+0xed/0x100 kernel/sched/core.c:7848
>> RIP: 0010:__might_sleep+0xed/0x100 kernel/sched/core.c:7848
>> Call Trace:
>>  __mutex_lock_common+0xc4/0x2ef0 kernel/locking/mutex.c:935
>>  __mutex_lock kernel/locking/mutex.c:1103 [inline]
>>  mutex_lock_nested+0x1a/0x20 kernel/locking/mutex.c:1118
>>  io_wq_submit_work+0x39a/0x720 fs/io_uring.c:6411
>>  io_run_cancel fs/io-wq.c:856 [inline]
>>  io_wqe_cancel_pending_work fs/io-wq.c:990 [inline]
>>  io_wq_cancel_cb+0x614/0xcb0 fs/io-wq.c:1027
>>  io_uring_cancel_files fs/io_uring.c:8874 [inline]
>>  io_uring_cancel_task_requests fs/io_uring.c:8952 [inline]
>>  __io_uring_files_cancel+0x115d/0x19e0 fs/io_uring.c:9038
>>  io_uring_files_cancel include/linux/io_uring.h:51 [inline]
>>  do_exit+0x2e6/0x2490 kernel/exit.c:780
>>  do_group_exit+0x168/0x2d0 kernel/exit.c:922
>>  get_signal+0x16b5/0x2030 kernel/signal.c:2770
>>  arch_do_signal_or_restart+0x8e/0x6a0 arch/x86/kernel/signal.c:811
>>  handle_signal_work kernel/entry/common.c:147 [inline]
>>  exit_to_user_mode_loop kernel/entry/common.c:171 [inline]
>>  exit_to_user_mode_prepare+0xac/0x1e0 kernel/entry/common.c:201
>>  __syscall_exit_to_user_mode_work kernel/entry/common.c:291 [inline]
>>  syscall_exit_to_user_mode+0x48/0x190 kernel/entry/common.c:302
>>  entry_SYSCALL_64_after_hwframe+0x44/0xa9
>>
>> Rewrite io_uring_cancel_files() to mimic __io_uring_task_cancel()'s
>> counting scheme, so it does all the heavy work before setting
>> TASK_UNINTERRUPTIBLE.
>>
>> Cc: stable@vger.kernel.org # 5.9+
>> Reported-by: syzbot+f655445043a26a7cfab8@syzkaller.appspotmail.com
>> Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
>> ---
>>  fs/io_uring.c | 41 ++++++++++++++++++++++++-----------------
>>  1 file changed, 24 insertions(+), 17 deletions(-)
>>
>> diff --git a/fs/io_uring.c b/fs/io_uring.c
>> index 09aada153a71..f3f2b37e7021 100644
>> --- a/fs/io_uring.c
>> +++ b/fs/io_uring.c
>> @@ -8873,30 +8873,33 @@ static void io_cancel_defer_files(struct io_ring_ctx *ctx,
>>  	}
>>  }
>>  
>> +static int io_uring_count_inflight(struct io_ring_ctx *ctx,
>> +				   struct task_struct *task,
>> +				   struct files_struct *files)
>> +{
>> +	struct io_kiocb *req;
>> +	int cnt = 0;
>> +
>> +	spin_lock_irq(&ctx->inflight_lock);
>> +	list_for_each_entry(req, &ctx->inflight_list, inflight_entry) {
>> +		if (!io_match_task(req, task, files))
> 
> This condition should be inversed. Jens, please drop it
> 
> p.s. I wonder how tests didn't catch that

Probably just make it cnt += io_match_task(req, task_files) to simplify
it. But yes, it looks reversed.

-- 
Jens Axboe