[alsa-devel] More snd_pcm_ioplug_avail_update() questions

[Rob Duncan]

At 00:04 on Thu, Jul 26 2018, Takashi wrote:
>> Is it required that all available data has been committed between calls
>> to snd_pcm_ioplug_avail_update() on an IO plugin capture stream?
>
> Yes, the capture data is committed.
> The snd_pcm_ioplug_hw_ptr_update() should have reported the amount of
> data the slave PCM can actually transfer.  Hence the transfer call
> thereafter must fulfill the whole requested data.

But what I see is that snd_pcm_rate_avail_update() commits data in the
slave PCM in units of slave->period_size (via
snd_pcm_rate_grab_next_period()), which means that if there is a partial
period in the slave PCM mmap it will not be committed.

In other words, not all the available data will be commmitted before the
next call to snd_pcm_ioplug_avail_update().  This means that the data
will be discarded the next time that snd_pcm_ioplug_avail_update() is
called.

Rob.