#
✕ nick:povik
#asahi · 2022-12-02 10:58
povikthat's what all the fuss in suspend/resume functions is about
#asahi · 2022-12-02 10:57
povikno, you need special sequencing of the pstate restore
#asahi · 2022-12-02 10:54
povikbut there's no special sequencing of a pstate register change
#asahi · 2022-12-02 10:53
povikit's put into shutdown by SDZ (if it has one)
#asahi · 2022-12-02 10:50
poviktas2562 doesn't touch the pstate register on suspend
#asahi · 2022-12-02 10:48
poviklooks like tas2552 has it too
#asahi · 2022-12-02 10:47
povikyeah, hit on tas2780 (this one i don't have :/)
#asahi · 2022-12-02 10:45
povikand i can eventually carry over the fix to others once i hook up the tower of codec evaluation boards TI sent me :p
#asahi · 2022-12-02 10:45
povikprobably best to fix tas2764/2770 now
#asahi · 2022-12-02 10:44
povikfrom experience there will be a bunch of TI codec drivers with this same faulty code
#asahi · 2022-12-02 10:38
povikprobably until *we* revive shared gpio handling
#asahi · 2022-12-02 10:37
poviknah, just give me a suggested-by
#asahi · 2022-12-02 10:36
poviknotwithstanding the other suspend fixes since those probably aren't upstreamable
#asahi · 2022-12-02 10:36
povikchadmed_: you can send this one straight to the list if you feel like it
#asahi · 2022-12-02 10:30
povik:D
#asahi · 2022-12-02 10:30
povikgot an email from github: @chadmed pushed 0 commits.
#asahi · 2022-12-02 10:30
povikah right
#asahi · 2022-12-02 10:29
povikjoske maybe means patches from PR, not the first link
#asahi · 2022-12-02 10:28
povikchadmed_: probably best putting the cache_only(false) above sync, but yeah, like that
#asahi · 2022-12-02 10:22
povikthat could explaine the tas2764/tas2770 difference
#asahi · 2022-12-02 10:22
povikbecause the ordering of registers will be different, so different registers will have been restored by the time sync gets to the power register
#asahi · 2022-12-02 10:21
povikwell, but that could be explained by the bug i suspect
#asahi · 2022-12-02 10:21
povikand the code is indeed the same you say...
#asahi · 2022-12-02 10:20
poviki was thinking of code
#asahi · 2022-12-02 10:20
povikah, as in actual sections in datasheet :)
#asahi · 2022-12-02 10:20
povikdatasheet sections?
#asahi · 2022-12-02 10:19
povikit should do sync, *then* up the power state
#asahi · 2022-12-02 10:19
povikif that's the case, i suspect there's a bug in the original resume function
#asahi · 2022-12-02 10:18
povikchadmed_: do you know if with your patch "ASoC: tas2770: software shutdown before suspend" the power-related code paths between tas2764/tas2770 are essentially the same?
#asahi · 2022-12-02 10:11
povikif you override that during the suspend cycle you lose some info
#asahi · 2022-12-02 10:11
povikthe fields are supposed to reflect the desired power state as far as commanded by the sound card
#asahi · 2022-12-02 10:07
povikyou shouldn't touch dac_powered/unmuted in the resume/suspend functions
#asahi · 2022-12-02 10:06
povikchadmed_: sorry to be a party pooper but that patch is wrong
#asahi · 2022-11-24 16:02
povikmust be some service
#asahi · 2022-11-24 16:02
povikis that the vanilla asahi distro?
#asahi · 2022-11-24 16:01
poviksomething on your system remembers the settings and restores them on restart
#asahi · 2022-11-24 16:01
povikyours should now work without any change to ucm unless you touch the mux again
#asahi · 2022-11-24 15:57
poviklet me do that before i get dinner :)
#asahi · 2022-11-24 15:57
povikyeah, ucm should probably set the headphone mux that's wreaking havoc here
#asahi · 2022-11-24 15:55
povikif not, see if you can reset that saved alsamixer settings. if you find out what service/daemon does the restoring, please let me know
#asahi · 2022-11-24 15:55
povikset it to primary and reset pipewire, see if that helps
#asahi · 2022-11-24 15:55
povikas you probably pieced together too
#asahi · 2022-11-24 15:55
povikcy8aer: ok. i think some userspace service restored the alsamixer settings from when you were trying it earlier. and that's what making pipewire reject the ucm settings, since it can't open hw:0,0 where it expects the headphones
#asahi · 2022-11-24 15:53
povikhuh, i seem to have left it in, it's allowed
#asahi · 2022-11-24 15:51
poviki thought i left headphones unroutable to hw:0,1. let me check the macaudio sources
#asahi · 2022-11-24 15:51
povikif it works with hw:0,1 you must have touched the Primary/Secondary muxes in alsamixer
#asahi · 2022-11-24 15:49
povikstrange. are you passing anything else to speaker-test? is there something noteworthy you changed on your setup?
#asahi · 2022-11-24 15:46
povikyou can do quick speaker-test -D hw:0,0 test to check the headphone is ok. beware it will be loud if you have Jack DAC at 100%!
#asahi · 2022-11-24 15:45
povikthen i think there's some pipewire library you are missing for devices to show up in gnome settings
#asahi · 2022-11-24 15:44
povikhm, and pipewire/pulseaudio?

plus ›