[alsa-devel] Semantics of a single period per buffer?

[Dave Dillow]

On Sun, 2007-11-25 at 21:09 +0100, Jaroslav Kysela wrote:
> On Sun, 25 Nov 2007, Dave Dillow wrote:
> 
> > While refining the sis7019 for submission, I found a possible issue with
> > having a single period per buffer. The easy way out is to set the
> > minimum number of periods to 2, but I see other drivers allow a single
> > period. This leads to a question about the semantics of a single period
> > that I've not been able to resolve using Google.
> > 
> > Any one have any thoughts or pointers to a discussion on having a single
> > period per buffer? Should it be a one-shot play, or just let it hit XRUN
> > every time? Other ideas?
> 
> I think, one period buffer can be only useable when XRUN check is disabled 
> (endless ring buffer playing) and application has own timing and polls for 
> the actual position in the ring buffer.

Thanks for the clarification -- I'll code accordingly.