[alsa-devel] [PATCH] ASoC: arizona: Add support for SNDRV_PCM_RATE_KNOT

[Charles Keepax]

On Thu, Feb 04, 2016 at 01:55:58PM +0100, Lars-Peter Clausen wrote:
> On 02/04/2016 01:45 PM, Charles Keepax wrote:
> > On Wed, Feb 03, 2016 at 09:24:25AM +0100, Lars-Peter Clausen wrote:
> >> On 02/02/2016 08:30 PM, Mark Brown wrote:
> >>> On Tue, Feb 02, 2016 at 04:38:04PM +0000, Charles Keepax wrote:
> >>>
> >>>> +#define ARIZONA_48K_RATES 12000, 24000, 48000, 96000, 192000, 4000, 8000, \
> >>>> +			  16000, 32000
> >>>
> >>> This just looks wrong.  It's valid code but with the combination of the
> >>> split line and the list without any sort of parenthesis it sets off
> >>> alarm bells.  Duplicating might not be the worst thing ever...
> >>
> >> For list constraints it is possible to specify a mask for which of the
> >> entries in the rates array should be considered. So you could use the same
> >> rate array, but use different masks. The adau1977 drive for example does this.
> >>
> > 
> > Ok so looking at this a little more I think you can only apply
> > the mask thing to values that are part of the basic defines
> > rather than masking off entries in a constraint list. The
> > adau1977 is using the mask to limit the supported formats which
> > is all done through the defines in that driver.
> 
> The list constraint struct looks like this:
> 
> struct snd_pcm_hw_constraint_list {
>         unsigned int count;
>         const unsigned int *list;
>         unsigned int mask;
> };
> 
> list is your rates array, count is the size of the rates array and mask
> specifies which of the entries in the rates array are valid.
> 
> E.g.
> 
> unsigned int rates[] = {
> 	44100,
> 	48000,
> };
> 
> struct snd_pcm_hw_constraint_list constr_44100 {
> 	.count = ARRAY_SIZE(rates),
> 	.list = rates,
> 	.mask = 0x1,
> };
> 
> struct snd_pcm_hw_constraint_list constr_48000 {
> 	.count = ARRAY_SIZE(rates),
> 	.list = rates,
> 	.mask = 0x2,
> };
> 
> struct snd_pcm_hw_constraint_list constr_all {
> 	.count = ARRAY_SIZE(rates),
> 	.list = rates,
> 	.mask = 0, /* 0 = All */
> };
> 

Ah my bad I was looking at entirely the wrong mask.

Thanks,
Charles