#asahi-dev · 2022-10-22 09:43
chadmednot sure how we would go about implementing that, i even tried constraining pipewire's maximum channel count to 6 for any given device at compile time, and it still opened all 8/32... #asahi-dev · 2022-10-22 09:35
chadmedthats unfortunate#asahi-dev · 2022-10-22 09:35
chadmedright#asahi-dev · 2022-10-22 08:47
chadmedcpufreq has been working fine here for the past few days#asahi-dev · 2022-10-22 06:53
chadmedprobably still sounds bad on j316 since im pretty certain theyre completely different speakers, but a test on one of those wouldnt go astray either :) #asahi-dev · 2022-10-22 06:52
chadmedclarity is greatly improved and ive managed to squish most of the audible boxiness present in the last version #asahi-dev · 2022-10-22 06:51
chadmeddoes someone with j314 speakers enabled mind testing the latest version of asahi-audio? ive completely redone the IRs and they sound much better now#asahi-dev · 2022-10-22 03:36
chadmedsounds like i might need to redo the firs at some point though, the safety stuff has changed the sound profile a little bit#asahi-dev · 2022-10-22 03:35
chadmedthen we get a frame size of 32 for each speaker channel which matches the jack #asahi-dev · 2022-10-22 03:34
chadmedsetting the bclk ratio to 192 fixes everything#asahi-dev · 2022-10-22 03:33
chadmedpovik: 256 % 6 != 0 #asahi-dev · 2022-10-18 07:33
chadmedpipewire can do enough to route the convolvers to their correct channels, but those two extra phantom ones still seem to be blowing something up#asahi-dev · 2022-10-18 07:32
chadmedso same failure mode, just less severe #asahi-dev · 2022-10-18 07:32
chadmedrunning samples through at ~half speed and cutting out every couple of seconds#asahi-dev · 2022-10-18 07:32
chadmedno it sounded like a vapourwave album (which like i said is progress over before)#asahi-dev · 2022-10-18 07:31
chadmedyeah intelligible music did come out#asahi-dev · 2022-10-18 07:17
chadmedif not then nslots should be fine... right? #asahi-dev · 2022-10-18 07:16
chadmeddoes num_codecs include the jack channels?#asahi-dev · 2022-10-18 07:14
chadmedno idea about the extra channels though, thats a relatively new thing#asahi-dev · 2022-10-18 07:14
chadmedthe failure mode re samples being messed up does resemble that. i did some testing with various mclk values when they were still in the DT and the same thing happened back then#asahi-dev · 2022-10-18 06:52
chadmedwell _less_ extra channels is still progress#asahi-dev · 2022-10-18 06:51
chadmedplayback_AUX6, playback_AUX7 :P #asahi-dev · 2022-10-18 06:49
chadmedgit rev-list v6.1-rc1..asahi-6.1-rc1-2 | wc -l gave 198#asahi-dev · 2022-10-18 06:47
chadmedyeah i noticed that, samples are still going through at ~half the speed they should even when everythings clamped to 96kHz #asahi-dev · 2022-10-18 06:46
chadmedoh ok it doesnt fix it fully but the samples buffer faster than they used to#asahi-dev · 2022-10-18 06:43
chadmeducm stuff still borked with pipewire but i can live without it for now until we work out whats going on there#asahi-dev · 2022-10-18 06:43
chadmedcould it actually have been i2c being too slow without interrupts?#asahi-dev · 2022-10-18 06:43
chadmedoh hey new kernel fixes speakers on j314 too! #asahi-dev · 2022-10-18 06:24
chadmedhey hey we're down to 198 patches on top of 6.1-rc1 vs 440 on top of 6.0-rc6#asahi-dev · 2022-10-17 06:51
chadmedjudging by dmesg output, that seems to be what is happening#asahi-dev · 2022-10-17 06:45
chadmedpovik: could the dapm null pointer deref be caused by the driver catching the jack unplug irq before cs42l84_component_probe() has run? #asahi-dev · 2022-10-16 12:54
chadmedhow fragile is asoc...#asahi-dev · 2022-10-16 12:53
chadmedok no nvm kernel null pointer if speakers are enabled???#asahi-dev · 2022-10-16 12:47
chadmednvm that was a dtb issue, the jack works with ucm and pipewire with the branch you linked povik #asahi-dev · 2022-10-16 12:35
chadmedinteresting, that HSBIAS pin disable results in a null pointer deref on j314#asahi-dev · 2022-10-16 11:29
chadmedhttps://gist.github.com/chadmed/2559fc4818a6a022209671a1da354d99 here you go #asahi-dev · 2022-10-16 11:26
chadmedyeah im trying, its a bit slow to paste :P #asahi-dev · 2022-10-16 11:25
chadmeddebugging pw is an absolute nightmare lmao#asahi-dev · 2022-10-16 11:24
chadmedthe next loglevel up makes over 100k lines..#asahi-dev · 2022-10-16 11:21
chadmedyeah the pw log doesnt really show much, it opens the device and then just closes it...#asahi-dev · 2022-10-16 11:15
chadmedthe second FE still suffers from the 32 channel thing#asahi-dev · 2022-10-16 11:14
chadmedJACK works with the ucm stuff in place, i can send vlc's JACK output module to the sink and it comes out of my headphones#asahi-dev · 2022-10-16 11:11
chadmedi dont recall ever having pulse on the asahi desktop image but maybe i just got rid of it really early on? #asahi-dev · 2022-10-16 11:10
chadmediirc#asahi-dev · 2022-10-16 11:10
chadmedasahi ships pipewire#asahi-dev · 2022-10-16 11:08
chadmedi dont get any sinks, just the mic#asahi-dev · 2022-10-16 11:07
chadmedhttps://tpaste.us/m69j #asahi-dev · 2022-10-16 11:06
chadmedhold on ill snip a pipewire log #asahi-dev · 2022-10-16 11:06
chadmeda profile called n/a appears which lets me use my headset mic lmao #asahi-dev · 2022-10-16 11:02
chadmeddevice is forced into "Off" and i cant select any other profile