On Wed, Mar 12, 2025 at 12:54:51AM +0200, Cristian Ciocaltea wrote: > On 3/11/25 6:17 PM, Maxime Ripard wrote: > > On Tue, Mar 11, 2025 at 12:57:38PM +0200, Cristian Ciocaltea wrote: > >> Provide tests to verify that drm_atomic_helper_connector_hdmi_check() > >> helper behaviour when using YUV420 output format is to always set the > >> limited RGB quantization range to 'limited', no matter what the value of > >> Broadcast RGB property is. > > [...] > > > We need more tests than that to test the various combinations, whether > > the fallback to YUV420 should work or not depending on the EDID, the > > driver capabilities, YUV420-only vs YUV420-also, etc. > > Some fallback tests were provided in the next patch, including checks like: > > KUNIT_ASSERT_TRUE(test, drm_mode_is_420_only(info, yuv420_only_mode)); > KUNIT_ASSERT_TRUE(test, drm_mode_is_420_also(info, preferred)); > > I'll try to further extend the test coverage. Yeah, sorry, I saw them after reviewing this patch. Still, I think we'll need more, especially to deal with the fallback cases. IIRC, you were testing the cases where you're forced to take YUV420 because of the monitors (or drivers) limit, but we should also try the tests where you're forced to take YUV420 but the driver implements HDMI 1.4 only for example, and thus we can't. Maxime
Attachment:
signature.asc
Description: PGP signature
