Re: [Qemu-devel] [PATCH 3/3] linux-aio: process completions from ioq_submit()

[Paolo Bonzini]

On 19/07/2016 12:25, Roman Pen wrote:
>      if (laiocb->co) {
> -        qemu_coroutine_enter(laiocb->co, NULL);
> +        if (laiocb->co == qemu_coroutine_self()) {
> +            laiocb->self_completed = true;

No need for this new field.  You can just do nothing here and check
laiocb.ret == -EINPROGRESS here in laio_co_submit.

> +        } else {
> +            qemu_coroutine_enter(laiocb->co, NULL);
> +        }
>      } else {
>          laiocb->common.cb(laiocb->common.opaque, ret);
>          qemu_aio_unref(laiocb);
> @@ -312,6 +317,12 @@ static void ioq_submit(LinuxAioState *s)
>          QSIMPLEQ_SPLIT_AFTER(&s->io_q.pending, aiocb, next, &completed);
>      } while (ret == len && !QSIMPLEQ_EMPTY(&s->io_q.pending));
>      s->io_q.blocked = (s->io_q.in_queue > 0);
> +
> +    if (s->io_q.in_flight) {
> +        /* We can try to complete something just right away if there are
> +         * still requests in-flight. */
> +        qemu_laio_process_completions(s);
> +    }

Can this leave I/O stuck if in_queue > 0 && in_flight == 0 after the
return from qemu_laio_process_completions?  I think you need to goto the
beginning of the function to submit more I/O requests in that case.

In fact, perhaps it's useful to always do so if any I/Os were completed.
 Should qemu_laio_process_completions return the number of completed
requests?

Paolo

>  }
>  
>  void laio_io_plug(BlockDriverState *bs, LinuxAioState *s)