davinci-pcm.c : snd_pcm_hardware.periods_min value

[Christophe Aeschlimann <c.aeschlimann-5/WUhD8XThOVQ4qyx6sgCw@public.gmane.org>]

Hi,

We are working with an OMAP-L138 attached to a TI stereo codec
(TLV320AIC3107) using McASP and we are trying to reduce the latency on
the playback path.

We are using the real-time patches on a 3.0.y longterm stable kernel.

In /sound/soc/davinci/davinci-pcm.c I can see the following structure:

> static struct snd_pcm_hardware pcm_hardware_playback = {
> 	.info = (SNDRV_PCM_INFO_INTERLEAVED | SNDRV_PCM_INFO_BLOCK_TRANSFER |
> 		 SNDRV_PCM_INFO_MMAP | SNDRV_PCM_INFO_MMAP_VALID |
> 		 SNDRV_PCM_INFO_PAUSE | SNDRV_PCM_INFO_RESUME),
> 	.formats = (SNDRV_PCM_FMTBIT_S16_LE),
> 	.rates = (SNDRV_PCM_RATE_8000 | SNDRV_PCM_RATE_16000 |
> 		  SNDRV_PCM_RATE_22050 | SNDRV_PCM_RATE_32000 |
> 		  SNDRV_PCM_RATE_44100 | SNDRV_PCM_RATE_48000 |
> 		  SNDRV_PCM_RATE_88200 | SNDRV_PCM_RATE_96000 |
> 		  SNDRV_PCM_RATE_KNOT),
> 	.rate_min = 8000,
> 	.rate_max = 96000,
> 	.channels_min = 2,
> 	.channels_max = 2,
> 	.buffer_bytes_max = 128 * 1024,
> 	.period_bytes_min = 32,
> 	.period_bytes_max = 8 * 1024,
> 	.periods_min = 16,
> 	.periods_max = 255,
> 	.fifo_size = 0,
> };

Which I think affects the minimum latency achievable.

I would like to know if there is a reason why the periods_min is set to
such a high value (16) ? (comparatively to most drivers).

I'd welcome any information :)

Best regards,
-- 
Christophe Aeschlimann

Embedded Software Engineer & IT Manager
ACN Advanced Communications Networks S.A.
2000 - Neuchâtel, Switzerland
Tel. +41 32 724 74 31