On Sun, May 14, 2023 at 1:41 AM Bagas Sanjaya wrote:
Can you attach dmesg messages so that we can know what exactly the error is?
Attached are both warm.txt (where it's broken) and cold.txt (where it's working fine, for comparison). These messages (only in warm.txt) stand out to me immediately:
[ 5.482071] snd_hda_intel 0000:03:00.6: azx_get_response timeout, switching to polling mode: last cmd=0x800f0005 [ 6.482536] snd_hda_intel 0000:03:00.6: No response from codec, disabling MSI: last cmd=0x800f0005 [ 7.485529] snd_hda_intel 0000:03:00.6: azx_get_response timeout, switching to single_cmd mode: last cmd=0x800f0005 [ 13.550516] snd_hda_intel 0000:03:00.6: no codecs initialized [ 13.550500] hdaudio hdaudioC1D0: no AFG or MFG node found
I bisected this to commit 5aec989, which first appeared in 5.15-rc6. In addition, I confirmed by doing a custom build of 6.3.2 with reverts of both f30741c (just to avoid a conflict) and 5aec989, which works correctly for me.
Note: when you refer commits, write at least 12 first characters of the commit's hash. You may want to set `core.abbrev=12` in your `~/.gitconfig`. Also the canonical commit reference is `--pretty=format:%h ("%s")`.
Got it. To avoid any ambiguity, here's the two commits I mentioned that way: 5aec98913095 ("ALSA: hda/realtek - ALC236 headset MIC recording issue") f30741cded62 ("ALSA: hda/realtek: Fix audio regression on Mi Notebook Pro 2020")
Regards,
Joseph C. Sible