Hey,
On Mon, 17 Feb 2020, Takashi Iwai wrote:
On Mon, 17 Feb 2020 10:11:59 +0100, Kai Vehmanen wrote:
With my patchset, if i915 is not present or the init fails, only the idisp codec is dropped and the probe continues for other codecs.
Your patch will partly conflict with mine as I rely on i915_init to report failure in this case. Ok if we wait until my patch gets in?
Sure, mine is just a quick fix. If there is already a work on it, let's take it instead.
ok, I'll try to fix the remaining issue with module reload today.
But, this HP machine still seems broken even after ignoring i915, so we'd need more work in anyway?
I needed to do some changes in the machine driver as well, but it should work on devices like this HP now. SOF will create PCM nodes for HDMI/DP, but they will never be connected -- these ports will be handled by the DGPU driver.
The one thing I haven't checked is how UCM behaves when the "Jack" nodes for HDMI/DP are not created. That may require some further changes to UCM (fortunately with UCMv2 we can now additionl conditional statements to cover this case as well). Another option is to create dummy mixer nodes, but that's not yet in my patch series.
Br, Kai