fix: io_uring sqpoll issue_time empty when kernel not yet read sq

[yangjueji]

In io_uring sqpoll mode, when kernel side thread has not yet read the sqring before second fio_ioring_commit() called, the sq_ring.head will remain the same. The second
fio_ioring_commit() will initialize the wrong io_u's issue_time. The old(in head) io_u‘s issue_time will to be initialized twice and new(in tail - 1) io_u's issue_time will not to be initialized. This problem will cause clat is weird, sometimes larger than lat.

Situation:

1. fio_ioring_queue(td, io_u_a) // head -> io_u_a; tail -> io_u_a + 1
2. fio_ioring_commit(td) // initial io_u_a.issue_time
3. fio_ioring_queue(td, io_u_b) // head -> io_u_a; tail -> io_u_b + 1
4. fio_ioring_commit(td) // initial io_u_a.issue_time again
5. kernel thread read sq_ring, // head += 2

io_u_b.issue_time is never initilized.

[ankit-sam]

Hi @yangjueji yeah I also observed the issue you mentioned with higher queue depth, thanks for submitting the change. It works.
c011bf1 partially fixed the sqpoll clat reporting issue (iodepth = 1). it never worked properly for iodepth > 1.