On 23/02/2024 16:47, Mark Brown wrote:
On Fri, Feb 23, 2024 at 03:39:10PM +0000, Richard Fitzgerald wrote:
+config SND_SOC_CS_AMP_LIB_TEST
- tristate "KUnit test for Cirrus Logic cs-amp-lib"
- depends on ACPI || COMPILE_TEST || KUNIT
Should this not depend unconditionally on KUNIT rather than KUNIT or some other stuff? ie
depends on ACPI || COMPILE_TEST depends on KUNIT
or equivalent.
Now I look at it again, it's not correct. It's the ACPI || COMPILE_TEST that is bogus.
+#define TYPESAFE_ACTIVATE_STATIC_STUB_PTR(test, fn_ptr, replacement_fn) \
- do { \
typecheck_fn(typeof(fn_ptr), replacement_fn); \__kunit_activate_static_stub(test, fn_ptr, replacement_fn); \- } while (0)
Should this be somewhere more generic in the kunit headers?
Damn, I meant to change this. I did a patch a while ago to fix kunit_activate_static_stub() so that it worked but it takes a really long time for kunit patches to end up in the mainline kernel. Hence this was a temporary workaround. I noticed my fix has gone in at last but I forgot to remove this workaround.
Can you skip this patch and take the others (assuming you are happy with them) and I'll fixup and resubmit this test later.