My default output sound device is HDMI from NVIDIA GTX-1060. After suspend to RAM output device changes from HDMI to S/PDIF. Reverting to HDMI breaks right channel, no longer can hear sound from the right speaker. Details on https://bugs.launchpad.net/ubuntu/+source/alsa-driver/+bug/1877154 I was asked to report this upstream so I'm doing it. My ALSA information is located at http://alsa-project.org/db/?f=25acdb7332de1b80a0cc4cb0abfeff2d21cccd4a
After the first wake from suspend you can find in messages from dmesg: [ 68.362807] pcieport 0000:00:01.3: PME: Spurious native interrupt! [ 68.527845] nvme nvme0: 15/0/0 default/read/poll queues [ 68.533836] snd_hda_intel 0000:0a:00.1: spurious response 0x80000000:0x0, last cmd=0x6f2f03
There was a known regression in Nvidia HDMI, and it must have been fixed in the latest 5.7-rc. It might be that stable trees are still behind, so please check with the latest 5.7-rc.
It will take some time until I'm able to test the latest kernel with zfs support. My current file system is zfs.
Created attachment 289155 [details] Dmesg output for no sound in 5.7.0 kernel
I've tested with 5.7.0-rc5+ and now output device stays after suspend to ram on HDMI. It doesn't change to S/PDIF so indeed it was fixed. However there is no sound at all coming from the speakers after power on. You can see in PAVU control that there is signal on the device but can't hear the the sound.It only starts to work after waking from suspend to ram, then you can hear the sound.
@Takashi Iwai do you need anything from me to do? I did what was requested from me. Should we carry on or close this bug?
It sounds like a different bug, so better to close this one, yes. But beware that such a bug is often caused by the graphics driver side and it's a Nvidia-binary, and if that's the case, it's not debuggable.