Re: [PATCH 0/2] ASoC: cs35l56/57/62/63: Use the correct SoundWire DP for feedback

[Pierre-Louis Bossart <pierre-louis.bossart@linux.dev>]

On 10/6/26 11:53, Richard Fitzgerald wrote:
> On 05/10/2026 7:33 pm, Pierre-Louis Bossart wrote:
>> On 10/5/26 17:52, Richard Fitzgerald wrote:
>>> The amp feedback path (AEC) was incorrectly using SoundWire DP3.
>>> The firmware outputs SDCA OT25 feedback on DP4. DP3 is reserved for
>>> SDCA companion amp.
>>>
>>> This series adds a DAI for OT25 and switches the sdw machine driver
>>> to use the new DAI for the feedback path.
>>
>> Is there any merit in keeping this DP3 exposed as a capture DAI, if
>> indeed it's intended to be a side connection for a companion chip?
> 
> I kept this patch to a minimum because it is a Fixes: so it must apply
> to older kernels.
> 
> However, you are correct. DP3 isn't used now. Not even on companion amp.
> DP3 is a companion output from the amp (SDCA OT127), but Cirrus
> companion amp goes codec->amp so uses DP1 on the amp.
> 
> When the driver was being written the SDCA spec and firmware behavior
> were not fully defined so pairing of the symmetrical DP1+DP3
> (for play+capture) was carried forward from previous amps. But SDCA
> doesn't provide any sort of UCM capability - widgets on the device have
> defined, fixed functionality that has to be preconfigured by the
> firmware to match what SDCA needs. Ultimately DP4 was more suitable for
> OT25. DP4 also works for loopback testing.
> 
> Removing DP3 would also solve some other problems that the driver and
> firmware are both "owning" the registers. We originally attempted to
> sync this up, then gave up on that complexity and overwrite the register
> settings the firmware made.

thanks for the explanations, I guess we'll see a follow-up patch to
remove DP3 altogether at some point.