Re: [Qemu-devel] Questions on audio_atexit(), possibly bugs

[malc]

On Wed, 30 Sep 2009, Markus Armbruster wrote:

> Excuse my ignorance on all things audio, but I stumbled over something
> that could be wrong.
> 
> audio_vm_change_state_handler() stops voices when the VM stops, and
> starts them when it continues.
> 
> audio_atexit() unconditionally stops them.  When a stopped VM exits,
> this stops voices that are already stopped.
> 
> Does the audio driver contract allow stopping a stopped voice?  If yes,
> I figured starting a running voice is fine, too.  If no, we have a bug
> in audio_atexit().

This should answer the question audio_atexit existed long before vm
change state handlers. Those were actually added to stop the host from
looping the same audio fragment over and over again (can/will happen
with DirectSound, mmapped OSS, fmod too if i'm not mistaken).

> 
> Why is audio_atexit() needed at all?  Doesn't the OS clean up just fine
> all by itself?  If we do need manual cleanup, why do we have to stop
> voices before we run fini_out() and fini_in()?

audio_atexit is needed for WAV, the OS will not write the final
information for us.

> 
> Unrelated, but nearby: audio_vm_change_state_handler() calls the
> ctl_out() callback with three arguments:
> 
>         hwo->pcm_ops->ctl_out (hwo, op, conf.try_poll_out);
> 
> (op is either VOICE_ENABLE or VOICE_DISABLE here), while audio_atexit()
> calls it with two:
> 
>         hwo->pcm_ops->ctl_out (hwo, VOICE_DISABLE);
> 
> Same for ctl_in().  Doesn't look kosher.  A quick check of oss_ctl_out()
> and oss_ctl_in() shows use of three parameters.

Yes, not kosher, but harmless, conf.try_poll_out is only applicable to
VOICE_ENABLE and is simply ignored by the handler of VOICE_DISABLE, this
is a vararg function, so it's okay, though i'd probably change this to
avoid further confusion.

-- 
mailto:address@hidden