[alsa-devel] [RFC][PATCH 1/3] ASoC: core: Move the default regmap I/O setting to snd_soc_register_codec()

[Li.Xiubo at freescale.com]

> Subject: Re: [alsa-devel] [RFC][PATCH 1/3] ASoC: core: Move the default regmap
> I/O setting to snd_soc_register_codec()
> 
> On 03/04/2014 09:25 AM, Xiubo Li wrote:
> > Add the default regmap I/O setting to snd_soc_register_codec() while
> > the CODEC is initialising, which will be called by CODEC driver device
> > probe(), and then we can make set_cache_io() go away entirely from each
> > CODEC ASoC probe.
> >
> > Signed-off-by: Xiubo Li <Li.Xiubo at freescale.com>
> > ---
> >   include/sound/soc.h  |  3 +++
> >   sound/soc/soc-core.c | 11 +++++++++++
> >   2 files changed, 14 insertions(+)
> >
> > diff --git a/include/sound/soc.h b/include/sound/soc.h
> > index 4c4d7e1..94bc1c4 100644
> > --- a/include/sound/soc.h
> > +++ b/include/sound/soc.h
> > @@ -749,6 +749,9 @@ struct snd_soc_codec_driver {
> >   	int (*set_pll)(struct snd_soc_codec *codec, int pll_id, int source,
> >   		unsigned int freq_in, unsigned int freq_out);
> >
> > +	/* codec regmap */
> > +	struct regmap *regmap;
> > +
> 
> Nope. The driver struct is globally shared between all device instances, the
> regmap struct is device instance specific. The proper way to solve this is
> to have a function like snd_soc_register_codec_with_io() which takes a
> pointer to the regmap struct.
> 

@Lars
Yes, this maybe a better choice, and I will think of it.

Thanks very much,

--
Best Regards,
Xiubo