On 1/15/21 11:31 AM, Marcelo Diop-Gonzalez wrote:
> On Fri, Jan 15, 2021 at 10:02:12AM -0700, Jens Axboe wrote:
>> On 1/15/21 9:54 AM, Marcelo Diop-Gonzalez wrote:
>>> This patch tries to fix a problem with IORING_OP_TIMEOUT events
>>> not being flushed if they should already have expired. The test below
>>> hangs before this change (unless you run with $ ./a.out ~/somefile 1):
>>>
>>> #include <fcntl.h>
>>> #include <stdio.h>
>>> #include <stdlib.h>
>>> #include <string.h>
>>> #include <unistd.h>
>>>
>>> #include <liburing.h>
>>>
>>> int main(int argc, char **argv) {
>>> if (argc < 2)
>>> return 1;
>>>
>>> int fd = open(argv[1], O_RDONLY);
>>> if (fd < 0) {
>>> perror("open");
>>> return 1;
>>> }
>>>
>>> struct io_uring ring;
>>> io_uring_queue_init(4, &ring, 0);
>>>
>>> struct io_uring_sqe *sqe = io_uring_get_sqe(&ring);
>>>
>>> struct __kernel_timespec ts = { .tv_sec = 9999999 };
>>> io_uring_prep_timeout(sqe, &ts, 1, 0);
>>> sqe->user_data = 123;
>>> int ret = io_uring_submit(&ring);
>>> if (ret < 0) {
>>> fprintf(stderr, "submit(timeout_sqe): %d\n", ret);
>>> return 1;
>>> }
>>>
>>> int n = 2;
>>> if (argc > 2)
>>> n = atoi(argv[2]);
>>>
>>> char buf;
>>> for (int i = 0; i < n; i++) {
>>> sqe = io_uring_get_sqe(&ring);
>>> if (!sqe) {
>>> fprintf(stderr, "too many\n");
>>> exit(1);
>>> }
>>> io_uring_prep_read(sqe, fd, &buf, 1, 0);
>>> }
>>> ret = io_uring_submit(&ring);
>>> if (ret < 0) {
>>> fprintf(stderr, "submit(read_sqe): %d\n", ret);
>>> exit(1);
>>> }
>>>
>>> struct io_uring_cqe *cqe;
>>> for (int i = 0; i < n+1; i++) {
>>> struct io_uring_cqe *cqe;
>>> int ret = io_uring_wait_cqe(&ring, &cqe);
>>> if (ret < 0) {
>>> fprintf(stderr, "wait_cqe(): %d\n", ret);
>>> return 1;
>>> }
>>> if (cqe->user_data == 123)
>>> printf("timeout found\n");
>>> io_uring_cqe_seen(&ring, cqe);
>>> }
>>> }
>>
>> Can you turn this into a test case for liburing? I'll apply the
>> associated patch, thanks (and to Pavel for review as well).
>
> Yup, can do. I'll try to clean it up some first (especially so it
> doesn't just hang when it fails :/)
That'd of course be nice, but not a hard requirement. A lot of the
regressions tests will crash a broken kernel, so...
--
Jens Axboe
