Re: [Qemu-devel] Question on aio_poll

[Alex Bligh <alex@alex.org.uk>]

Stefan,

--On 24 July 2013 09:54:39 +0200 Stefan Hajnoczi <stefanha@gmail.com> wrote:

>> I left this how it was in the end (I think), and got round it by
>> creating a bogus pipe for the test to listen to.
>
> Doing that requires the changes in my patch series, otherwise you break
> aio_poll() loops that are waiting for pending I/O requests.  They don't
> want to wait for timers.

Sorry I meant leaving the main code as is, and creating the bogus
pipe solely in tests/test-aio.c in my new test that tests a timer
attached to AioContext - no changes elsewhere.

> I hope we can eventually unify event loops and then the select function
> should behave as you described.  For now though, we need to keep the
> current behavior until my .io_flush() removal series or something
> equivalent is merged, at least.

OK. That's pretty much the way I went with the PATCHv2 series.

>> I note you now have:
>>     if (ctx->pollfds->len == 1) {
>>         return progress;
>>     }
>>
>> Is the '1' there the event notifier? How do we know there is only
>> one of them?
>
> There many be many EventNotifier instances.  That's not what matters.
>
> Rather, it's about the aio_notify() EventNotifier.  Each AioContext has
> its own EventNotifier which can be signalled with aio_notify().  The
> purpose of this function is to kick an event loop that is blocking in
> select()/poll().  This is necessary when another thread modifies
> something that the AioContext needs to act upon, such as adding/removing
> an fd.

Thanks

-- 
Alex Bligh