popcornmix

kms support is in upstream kernel for pi0-3. Pi4 support has been submitted but not fully merged yet. https://lore.kernel.org/dri-devel/cover.6c896ace9a5a7840e9cec008b553cbb004ca1f91.1582533919.git-series.maxime@cerno.tech/T/#m0275b55e1518be1fb2154d3c95e13c1b7de1f347

Does launching `kodi --pulse` help?

> BTW, Why is that happens, Is there any config or settings I can do? pulse prevents advanced alsa features like passthough audio which can be important for kodi users,...

It's more likely to be a kms vs fkms issue than 32-bit vs 64-bit. Just to confirm, @broken613, do you have `dtoverlay=vc4-kms-v3d` in config.txt? Does changing it to `dtoverlay=vc4-fkms-v3d` help?

`GET_CLOCK_RATE` has always been problematic, as the firmware can change the clocks for a variety of reasons. e.g. core clock changes after `initial_turbo` has expired, during hdmi mode sets, when...

I'd expect setting clock and retrieving it to give the value you set. Let me now if you can confirm that.

MAX clock may be safer if you don't want a divided down value to exceed a specced value.

> I would also like clarification that SET_CLOCK_RATE will act as a clamp and the firmware won't exceed the requested value, as well as it will try and honor that...

> @popcornmix what is the expected behavior of GET_CLOCK_RATE at this point? It returns the value requested by the arm. > I've just retested with commit f9bc224 and for CLOCK_ID_EMMC...

> let me check what that is doing. I can reproduce. For those clocks it's not behaving as designed. I'll have a think about best way to fix it.