Re: [Qemu-devel] [PATCH 1/2] Use fd signal trick to break us out of select; do not sigwait

[Ian Jackson]

Anthony Liguori writes ("Re: [Qemu-devel] [PATCH 1/2] Use fd signal trick to 
break us out of select; do not sigwait"):
> Somehow, I didn't see this until today--sorry for the delayed response.  
> I personally prefer the patch I posted.

Did you see Jamie Lokier's message about the behaviour of glibc 2.3.1
with respect to signal blocking and the glibc aio helper threads ?
That is, it doesn't always block them.

I think the result is that with your patch you risk having the signal
delivered on the aio helper thread.

>  I've also got another patch on top of it that refactors the aio API
> so that multiple aio implementations can be used.

That doesn't seem unreasonable.  But we have to have an answer which
works around the glibc bug.

>  I notice you leave the qemu_aio_poll() in place in the main loop,
> is there a reason you left it there?

This is essential as it does the actual work for AIO completion.

Ian.