09:25 <leio> jannau: the screen now blanks occasionally when touching sound settings stuff in gnome (e.g. now when I expanded the device selection to switch from hdmi to headset)

09:26 <leio> looks like renegotiating stuff, etc; https://bpa.st/ACIA

09:27 <leio> around timestamp 47820 or so

09:49 <chadmed> i can go try write a ucm profile for it

09:54 <jannau> we might want to change things in the driver first

09:54 <leio> now it just did it without me doing anything, was idle

09:55 <leio> http://dpaste.com/BV7XMUPNN

09:55 <leio> (some notification sounds came in, but it was on headset)

09:57 <jannau> I'm also not sure how we want to handle it in general. I guess we would want a sound card which is always present and jacks for each possible hdmi/dp audio out

09:59 <jannau> in both cases 5 second before: "RTKit: syslog message: DCPAVService.cpp:954: [AFK]Restaring interfaces with reason - Capabilities changed.."

10:01 <chadmed> how does e.g. i915 do it

10:02 <chadmed> i had a look last week when we discussed this but didnt dig too deep as i had stuff to do

10:10 <jannau> mostly in snd_hda_intel and the drm driver provides just utility functions as component. I haven't looked into snd_hda_intel

10:20 <chadmed> seems as though amd does the same with just hardware init done in the actual gpu driver

10:21 <chadmed> each crtc gets its own pcm it seems too

14:59 <leio> and now lost display permanently with some page flip failed fun after accidentally having turned off the monitor for a bit

15:00 <leio> gnome-shell[1353]: Page flip failed: drmModeAtomicCommit: Invalid argument

15:00 <leio> gnome-shell[1353]: Failed to post KMS update: drmModeAtomicCommit: Invalid argument