On Thu, Apr 21, 2022 at 03:02:54PM +0200, Jerome Brunet wrote:
On Thu 21 Apr 2022 at 13:38, Mark Brown broonie@kernel.org wrote:
I'm also seeing other failures due to these controls having invalid values, eg:
# # AIU ACODEC SRC.0 value 3 more than item count 3
but without documentation I'm not sure how to interpret/fix these - either the value should be fixed up on startup or there should be an extra value there (disconnected possibly?).
Value 3 is an I2S input from a block we don't support. If we did support it, it would be an hostless DPCM BE to BE link I'm not sure how I would represent this in ASoC TBH :/
At the time, I thought it would be easier (for the users) to leave this value out and not give the false impression that the path was somehow supported.
I did not realize it was the reset value nor that it would be a problem.
I can add a element to the enum if you think it is better have it regardless of the actual path support. What do you think ?
I don't have strong feelings TBH - your argument for not supporting it for now makes sense to me and we generally try to avoid changing the hardware defaults. Probably I'd go with changing the default but since it's a valid value in the hardware it's most likely not too bad to leave things as they are for now, though there might be some applications that get confused and explode.