[alsa-devel] [PATCH] ASoC: rt5677: handle reset_pin logical state correctly

[Anatol Pomozov]

Hi

Returning back to this discussion that slipped away from my mailbox.

On Fri, Aug 14, 2015 at 12:01 PM, Mark Brown <broonie at kernel.org> wrote:
> On Fri, Jul 17, 2015 at 10:38:56AM -0700, Anatol Pomozov wrote:
>> According to the datasheet RESET is active low pin, i.e. system goes to reset
>> state when pin is low.
>>
>> Handle logic state correctly - set reset_pin to logical 0 at boot time, and
>> set it to logical 1 when we need to reset the chip.
>
> I'm a bit confused here - how was the original code (written by you it
> seems...) tested?  This looks like it's fixing the device being held in
> reset when it should be operation which seems like something that'd get
> noticed.  Are there existing systems that the current code works for
> which we need to handle here?

While working with the original patch you mentioned I missed the
active low description at RESET pin. I configured gpio in the dts as
ACTIVE_HIGH and set the logic state to 0 at suspend mode. This puts
the chip into reset state correctly.

Later I discovered correct polarity and given change modifies gpio
properties accordingly. I also changed our dts gpio to ACTIVE_LOW.
Chip reset state still works as expected. The change does not modify
anything at the signalling level, it just adds double negative to the
code. Compare:

pin configured as ACTIVE_HIGH - to move codec out of reset we need to
set gpio logic level to 1. Physical signal is 1.
pin configured as ACTIVE_LOW - to move codec out of reset we need to
set gpio logic level to 0. Gpio becomes non-active that is represented
by physical signal level 1.