Re: [PATCH 5.10 1/5] io_uring: always let io_iopoll_complete() complete polled io.

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

On 09/12/2020 20:17, Pavel Begunkov wrote:
> On 08/12/2020 21:10, Jens Axboe wrote:
>> On 12/8/20 12:24 PM, Pavel Begunkov wrote:
>>> On 08/12/2020 19:17, Jens Axboe wrote:
>>>> On 12/8/20 12:12 PM, Pavel Begunkov wrote:
>>>>> On 07/12/2020 16:28, Jens Axboe wrote:
>>>>>> On Sun, Dec 6, 2020 at 3:26 PM Pavel Begunkov <asml.silence@gmail.com> wrote:
>>>>>>> From: Xiaoguang Wang <xiaoguang.wang@linux.alibaba.com>
>>>>>>>
>>>>>>> The reason is that once we got a non EAGAIN error in io_wq_submit_work(),
>>>>>>> we'll complete req by calling io_req_complete(), which will hold completion_lock
>>>>>>> to call io_commit_cqring(), but for polled io, io_iopoll_complete() won't
>>>>>>> hold completion_lock to call io_commit_cqring(), then there maybe concurrent
>>>>>>> access to ctx->defer_list, double free may happen.
>>>>>>>
>>>>>>> To fix this bug, we always let io_iopoll_complete() complete polled io.
>>>>>>
>>>>>> This patch is causing hangs with iopoll testing, if you end up getting
>>>>>> -EAGAIN on request submission. I've dropped it.
>>>>>
>>>>> I fail to understand without debugging how does it happen, especially since
>>>>> it shouldn't even get out of the while in io_wq_submit_work(). Is that
>>>>> something obvious I've missed?
>>>>
>>>> I didn't have time to look into it, and haven't yet, just reporting thation.
>>>> it very reliably fails (and under what conditions).
>>>
>>> Yeah, I get it, asked just in case.
>>> I'll see what's going on if Xiaoguang wouldn't handle it before.
>>
>> Should be trivial to reproduce on eg nvme by doing:
>>
>> echo mq-deadline > /sys/block/nvme0n1/queue/scheduler
>> echo 2 > /sys/block/nvme0n1/queue/nr_requests
>>
>> and then run test/iopoll on that device. I'll try and take a look
>> tomorrow unless someone beats me to it.
> 
> Tried out with iopoll-enabled null_blk. test/iopoll fails with
> "test_io_uring_submit_enters failed", but if remove iteration limit
> from the test, it completes... eventually.
> 
> premise: io_complete_rw_iopoll() gets -EAGAIN but returns 0 to
> io_wq_submit_work().
> The old version happily completes IO with that 0, but the patch delays
> it to do_iopoll(), which retries and so all that repeats. And, I believe,
> that the behaviour that io_wq_submit_work()'s -EGAIN check was trying to
> achieve...
> 
> The question left is why no one progresses. May even be something in block.
> Need to trace further.

test_io_uring_submit_enters()'s io_uring_submit never goes into the kernel,
IMHO it's saner to not expect to get any CQE, that's also implied in a comment
above the function. I guess before we were getting blk-mq/etc. them back
because of timers in blk-mq/etc.

So I guess it should have been be more like the diff below, which still
doesn't match the comment though.

diff --git a/test/iopoll.c b/test/iopoll.c
index d70ae56..d6f2f3e 100644
--- a/test/iopoll.c
+++ b/test/iopoll.c
@@ -269,13 +269,13 @@ static int test_io_uring_submit_enters(const char *file)
 	/* submit manually to avoid adding IORING_ENTER_GETEVENTS */
 	ret = __sys_io_uring_enter(ring.ring_fd, __io_uring_flush_sq(&ring), 0,
 						0, NULL);
-	if (ret < 0)
+	if (ret != BUFFERS)
 		goto err;
 
 	for (i = 0; i < 500; i++) {
-		ret = io_uring_submit(&ring);
-		if (ret != 0) {
-			fprintf(stderr, "still had %d sqes to submit, this is unexpected", ret);
+		ret = io_uring_wait_cqe(&ring, &cqe);
+		if (ret < 0) {
+			fprintf(stderr, "wait cqe failed %i\n", ret);
 			goto err;
 		}

-- 
Pavel Begunkov