ASoC for TI OMAP - alsa-devel submission?

ASoC for TI OMAP - alsa-devel submission?

[Jarkko Nikula]

I've been doing generic ASoC drivers for Texas Instruments OMAP
processors and I recently released the first version on linux-omap
mailing list:

http://marc.info/?l=linux-omap&m=120643770810683&w=2
http://marc.info/?l=linux-omap&m=120643766610623&w=2

Currently patch tries to cover OMAPs from 1x10 to 2420 and is tested
on Nokia N810 having OMAP2420 and TLV320AIC33 codec.

I'm going to release it here also but there are few issues I would like
to discuss before it:

- McBSP used in OMAP2430 and OMAP3430 are bit different than other
1510-2420 OMAPs and that support we have currently missing in
linux-omap either. However, changes to my drivers should be quite
minimal in order to support them. E.g. base address definitions and
different DMA data type.
- Solving the channel synchronization problem might require API change
to linux-omap McBSP driver.
- ASoC v2 considerations? Like multiformat DAI.
- Multiformat McBSP DAI link driver (I'll send an another mail on that).

So what I mean that should I post the patch like as it now, rebased to
dev branch of linux-2.6-asoc tree and with some minor changes, or wait a
bit when all linux-omap and ASoC v2 evolve? What you think?

-- 
Jarkko

Re: ASoC for TI OMAP - alsa-devel submission?

[Mark Brown]

On Wed, Apr 02, 2008 at 01:45:30PM +0300, Jarkko Nikula wrote:

> So what I mean that should I post the patch like as it now, rebased to
> dev branch of linux-2.6-asoc tree and with some minor changes, or wait a
> bit when all linux-omap and ASoC v2 evolve? What you think?

Whatever is most convenient for you.  From the point of view of getting
stuff in immediately it's best to base things on ASoC v1 though from an
ongoing development point of view ASoC v2 is a better target if you can
cope with any churn and hopefully we should be starting to push it into
the kernel within the next couple of release cycles so if it's going to
take a while to get the required BSP changes in...

Re: ASoC for TI OMAP - alsa-devel submission?

[Jarkko Nikula]

On Wed, 2 Apr 2008 12:00:54 +0100
"ext Mark Brown" <broonie@opensource.wolfsonmicro.com> wrote:

> On Wed, Apr 02, 2008 at 01:45:30PM +0300, Jarkko Nikula wrote:
> 
> > So what I mean that should I post the patch like as it now, rebased
> > to dev branch of linux-2.6-asoc tree and with some minor changes,
> > or wait a bit when all linux-omap and ASoC v2 evolve? What you
> > think?
> 
> Whatever is most convenient for you.  From the point of view of
> getting stuff in immediately it's best to base things on ASoC v1
> though from an ongoing development point of view ASoC v2 is a better
> target if you can cope with any churn and hopefully we should be
> starting to push it into the kernel within the next couple of release
> cycles so if it's going to take a while to get the required BSP
> changes in...
>
I'll think I have to go for both :-)

The first ASoC v1 based version I would mostly like to see as an
enabler for getting n different ALSA implementation we have in
linux-omap tree to be ported for ASoC. Especially codec drivers. Or to
have an another DAI link driver for EAC (Enhanced Audio Controller)
found in OMAP2420.

The v2 I see also as mandatory path. We already do have boards with
multiple DAI links and codecs: e.g. audio codec + audio link to
bluetooth chip.

I'll simplify a bit my patch, base it to alsa-dev and post it here soon.

-- 
Jarkko