Re: [PATCH 3/3] winecoreaudio.drv: Use AUHAL API instead of AudioQueue

[Francois Gouget]

On Fri, 27 Jun 2014, Andrew Eikum wrote:
[...]
> +        if(This->wrap_bufsize_frames < *nframes){
> +            free(This->wrap_buffer);
> +            This->wrap_buffer = malloc(data->mBuffers[0].mDataByteSize);
> +            This->wrap_bufsize_frames = *nframes;

Does the AUHAL API require that we use malloc() / free() here (and in 
the other locations dealing with wrap_buffer) or could this be switched 
to HeapAlloc() / HeapFree()?

If we're free to use the latter doing so would be more consistent with 
the rest of Wine.


-- 
Francois Gouget <fgouget(a)free.fr>              http://fgouget.free.fr/
      Sufficently advanced incompetence is indistinguishable from malice.

[PATCH 3/3] winecoreaudio.drv: Use AUHAL API instead of AudioQueue

[Ken Thomases]

On Jul 5, 2014, at 4:13 AM, Francois Gouget wrote:
> On Fri, 27 Jun 2014, Andrew Eikum wrote:
> [...]
> > +        if(This->wrap_bufsize_frames < *nframes){
> > +            free(This->wrap_buffer);
> > +            This->wrap_buffer = malloc(data->mBuffers[0].mDataByteSize);
> > +            This->wrap_bufsize_frames = *nframes;
>
> Does the AUHAL API require that we use malloc() / free() here (and in 
> the other locations dealing with wrap_buffer) or could this be switched 
> to HeapAlloc() / HeapFree()?
>
> If we're free to use the latter doing so would be more consistent with 
> the rest of Wine.

This code is in a callback that Core Audio calls on a thread of its own creation.  That thread is not set up as a Wine thread, so you can't call other Wine code.

My argument was that this allocation can be completely avoided in this context.  It is possible to know in advance the upper limit of the buffer size required and allocate it during setup.  In that case, HeapAlloc() could be used.

-Ken