Re: [PATCH 0/4] ASoC: Add support for GPIOs driven amplifiers
From: Christophe Leroy (CS GROUP)
Date: Sun Apr 05 2026 - 13:00:34 EST
Hi Mark,
Le 30/03/2026 à 18:41, Herve Codina a écrit :
On Mon, 30 Mar 2026 16:48:54 +0100
Mark Brown <broonie@xxxxxxxxxx> wrote:
On Mon, Mar 30, 2026 at 05:39:44PM +0200, Herve Codina wrote:
Mark Brown <broonie@xxxxxxxxxx> wrote:
This sounds a lot like simple-amplifier.c?
The gpio driven amplifier proposed is more generic and can handle
more complex design. I.e. op-amp + resistor and/or line (mute,
bypass) switching. Hardwares handled by this driver are a superset
of just dio2125 and so simple-amplifier.c.
IMHO, it makes sense to have a specific driver for those kind
of hardware design.
Right, and if it's a superset it feels like it should all be one driver
rather than two separate ones.
Also, it is worth noting that simple-amplifier.c considered a stereo
amplifier (left + right).
Considering the two op-amp available in dio2125 as just two op-amp used
in two separated mono channel with additional component to switch related
to resistors (independant switching for each channel) means that almost
everything proposed in audio-gpio-amp have to be duplicated (gain, mute,
bypass per channel) instead of just instantiate two audio-gpio-amps.
On the other hand, there is no reason to handle a stereo component in
audio-gpio-amp. Further more with the hardware I have handling a stereo
component doesn't make sense. Indeed, I only have independent mono lines
with their own resistor switched amplification circuitry.
I could merge everything in one .c file but only a few part of source code
will be common to simple-amplifier and audio-gpio-amp. IMHO the resulting
merged code will look like two different drivers merged in one .c file.
Following explanation from Herve I have the feeling that combining the two drivers into a single one will bring more complexity for little benefit.
Do you still think it is worth having a combined driver allthough they address quite different setups ?
Thanks
Christophe