[alsa-devel] ALSA calling pcm_pointer excessively?

[Jassi Brar]

On Thu, May 10, 2012 at 10:34 PM, Russell King - ARM Linux
<linux at arm.linux.org.uk> wrote:
> On Thu, May 10, 2012 at 10:30:44PM +0530, Jassi Brar wrote:
>> On Thu, May 10, 2012 at 9:36 PM, Russell King - ARM Linux
>> <linux at arm.linux.org.uk> wrote:
>> >
>> > The question is - what would cause ALSA to hammer away at the buffer
>> > like that - checking it every 16 or 8 DMA bytes transferred?  I'm also
>> > seeing around 75% system CPU time, which really isn't good.
>> >
>> > The above was captured while aplay was running, with no apparant audio
>> > corruption.
>> >
>> > Any ideas?
>> >
>> I believe it's because the pcm's write tries to keep the data in ring
>> buffer filled
>> upto brim and keeps polling the dma position till it keeps reporting
>> some progress.
>>
>> I think if you could avoid that behavior if your driver carefully
>> keeps negligible
>> progresses to itself and report only considerable ones. Basically, core would
>> stop polling and wait on queue when two pointer calls return same value i.e no
>> progress.
>
> You _could_ but why should drivers work around silly behaviour in the
> ALSA core?
>
One reason I could think of is to save us from under-runs for as long
as possible.