Am 07.03.2024 um 18:19 hat Prasad Pandit geschrieben:
> Hi,
>
> On Thu, 7 Mar 2024 at 19:21, Kevin Wolf <kw...@redhat.com> wrote:
> > Kernel support for this is "relatively" new, added in 2018. Should we
> > fall back to the thread pool if we receive -EINVAL?
>
> laio_co_submit
> laio_do_submit
> ioq_submit
> io_submit
>
> * When an aio operation is not supported by the kernel, io_submit()
> call would return '-EINVAL', indicating that the specified (_FDSYNC)
> operation is invalid for the file descriptor. ie. an error (-EINVAL)
> needs to reach the 'laio_co_submit' routine above, to make its caller
> fall back on the thread-pool functionality as default.
Hm... This might make it more challenging because then not only one
specific request fails, but the whole submission batch. Do we know if it
can include unrelated requests?
> * Strangely the 'ioq_submit' routine above ignores the return value
> from io_submit and returns void. I think we need to change that to be
> able to fall back on the thread-pool functionality. I'll try to see if
> such a change works as expected. Please let me know if there's another
> way to fix this.
It passes the return value to the request:
if (ret < 0) {
/* Fail the first request, retry the rest */
aiocb = QSIMPLEQ_FIRST(&s->io_q.pending);
QSIMPLEQ_REMOVE_HEAD(&s->io_q.pending, next);
s->io_q.in_queue--;
aiocb->ret = ret;
qemu_laio_process_completion(aiocb);
continue;
}
laio_co_submit() then fetches it from laiocb.ret and returns it to its
caller (in this case, your new code).
Kevin
