Re: [PATCH] kunit: cs_dsp: Depend on FW_CS_DSP rather then enabling it

[Richard Fitzgerald]

On 21/3/25 15:01, Nico Pache wrote:
> I'm a little confused how the FW_CS_DSP config which was added in
> v5.16 is reliant (library code that is only used by KUNIT) on a config
> that was added in v6.14. Presumably the library is not just for the
> KUNIT test. What was the purpose of this config before the
> introduction of the KUNIT test. Im guessing it was not user selectable
> back then too.
All this cs_dsp code was originally in sound/soc/codecs/wm_adsp.c.
The main DSP support was moved out into the generic cs_dsp library
so we can support using the DSP with non-ASoC drivers, and on devices
that are not sound chips. The drivers that use the cs_dsp library
select the Kconfig.
(ISTR that the config option may have been renamed to FW_CS_DSP at some
point, so could have a different name in older kernels.)

> > > It also looks like FW_CS_DSP_KUNIT_TEST_UTILS and FW_CS_DSP_KUNIT_TEST
> > > are redundant.
> >
> > Possibly there's more tests to come that'll use them?
Yes, that is explained in the commit message that added this config
symbol. As this has been split out of wm_adsp, we still need to test
the functionality that was left in wm_adsp. I wrote the test utils
functions as a library so that a wm_adsp kunit test can re-use it.
I am still working on the wm_adsp kunit test.