On Fri, 14 Jan 2022 19:37:20 +0100, Alexander Sergeyev wrote: > > On Fri, Jan 14, 2022 at 05:37:42PM +0100, Takashi Iwai wrote: > > > I assume that the referred repository is the one at [1]. I've tried > > > 081c73701ef0 "ALSA: hda: intel-dsp-config: reorder the config > > > table". It crashed with nearly identical logs. > > > > OK, then it's still something to do with the led cdev > > unregisteration. > > > > Could you try the patch below? > > This patch solved the BUG problem. But after unbind/bind micmute LED stopped > working. Speakers and mute LED are fine though. Does the corresponding sysfs entry exist in /sys/class/leds/*? And can you control LED over there? > Dmesg: > snd_hda_intel 0000:05:00.6: AMD-Vi: Event logged [IO_PAGE_FAULT domain=0x0015 address=0x1fffff800 flags=0x0020] Hmm, that looks bad. Something must be accessing out of bound. > snd_hda_codec_realtek hdaudioC1D0: out of range cmd 0:20:400:90170118 > snd_hda_codec_realtek hdaudioC1D0: out of range cmd 0:20:400:411111f0 > snd_hda_codec_realtek hdaudioC1D0: out of range cmd 0:20:400:270300 This seems to be a bogus COEF. But I have no idea from where this comes. The values look completely wrong. I guess you'd need to put some debug prints to trace down how those are triggered... Takashi
