[alsa-devel] [PATCH 2/3] ASoC: tas571x: New driver for TI TAS571x power amplifiers

[Kevin Cernekee]

On Fri, Apr 24, 2015 at 2:28 AM, Mark Brown <broonie at kernel.org> wrote:
> On Thu, Apr 23, 2015 at 05:47:49PM -0700, Kevin Cernekee wrote:
>
>> This is mostly working OK, but regcache_sync() assumes that the
>> hardware registers have been reset back to the default values.  The
>> "pdn" GPIO doesn't actually reset the state of the tas571x; it just
>> makes I2C inaccessible and inhibits audio output.  So if the factory
>> default for mute is 0, corner cases like this fail:
>
> ...
>
>> Aside from unnecessarily pulsing the reset GPIO when transitioning
>> back from SND_SOC_BIAS_OFF or overriding regcache_default_sync(), can
>> you think of a way to work around this?
>
> Do you need to work around it?  If the register map is being perserved
> you don't need to sync so just don't do it - it's just that the normal
> expectation would be that power down would cause the register map to be
> reset.

How do I tell regcache to write out any updates that happened while
the hardware was inaccessible?  I see that regmap->cache_dirty is 1,
but nothing flushes it automatically when exiting cache_only mode.