| freemangordon | arno11: is it possible to set stream reate from ucm? | 11:52 |
|---|---|---|
| freemangordon | *rate | 11:52 |
| arno11 | freemangordon: not possible imo | 13:32 |
| freemangordon | arno11: ani idea why I cannot select device 1 from UCM? | 13:36 |
| freemangordon | *any | 13:36 |
| freemangordon | see https://pastebin.com/cbSu2Vcx | 13:36 |
| arno11 | ok let me check | 13:36 |
| freemangordon | but when I set PlaybackPCM "hw:${CardId},1" in UCM, I get "[pulseaudio] (alsa-lib)conf.c: Unknown parameter 1" | 13:37 |
| freemangordon | maybe I don;t understand what it is about | 13:37 |
| freemangordon | when it is 0, it sets on cpcap-hifi-0 | 13:38 |
| freemangordon | but I want to set on cpcap-voice-1 | 13:38 |
| arno11 | ok, need few min | 13:40 |
| freemangordon | ok | 13:40 |
| freemangordon | LMK if you need more info | 13:41 |
| arno11 | (it is so different from n900, i need to check mapphone ucm) | 13:41 |
| arno11 | ok | 13:41 |
| freemangordon | ok | 13:42 |
| freemangordon | I want VoiceCall.conf to refer to device 1, not device 0 | 13:42 |
| freemangordon | maybe not possible, as it sets controls, dunno | 13:42 |
| freemangordon | but, something sets rate to 8000 during call | 13:43 |
| freemangordon | I am not able to identify who | 13:43 |
| arno11 | so for voicecall.conf, indeed it uses card 0. must be PlaybackPCM "hw:${CardId},1" in each SectionDevice | 13:47 |
| arno11 | to use card1 | 13:47 |
| freemangordon | no, it uses card 1 | 13:47 |
| freemangordon | ,0 is device | 13:47 |
| freemangordon | so, it uses card 1,0 | 13:48 |
| freemangordon | which is hifi | 13:48 |
| freemangordon | but I want to use 1,1 | 13:48 |
| freemangordon | which is voice | 13:48 |
| freemangordon | card 0 is HDMI | 13:48 |
| arno11 | ah yes i forgot that (Wizzup already told me...) | 13:49 |
| arno11 | so really don't know atm | 13:50 |
| freemangordon | ok | 13:50 |
| arno11 | let me few min again | 13:50 |
| arno11 | looking @mappphone ucm again, you should try hw:${CardId},1 imo | 13:56 |
| freemangordon | I did | 13:56 |
| freemangordon | "but when I set PlaybackPCM "hw:${CardId},1" in UCM, I get "[pulseaudio] (alsa-lib)conf.c: Unknown parameter 1"" | 13:57 |
| arno11 | ah yes sorry, ok | 13:57 |
| arno11 | (just need time to understand mapphone stuff :P) | 13:57 |
| arno11 | btw the noise in earpeace is a static noise or ? | 13:59 |
| freemangordon | no | 13:59 |
| freemangordon | it changes when the other side speaks | 13:59 |
| freemangordon | deffinitely a wrong freq | 13:59 |
| arno11 | ah ok | 13:59 |
| arno11 | yep | 13:59 |
| freemangordon | but I can't get kernel to set it, as I can't get alsa/pa/ucm/whatever to do anything about t | 14:00 |
| freemangordon | well... | 14:00 |
| arno11 | indeed | 14:00 |
| freemangordon | cpcap_voice_hw_params() (in the diver is called) | 14:00 |
| freemangordon | but I don;t want to hack code there, but want everything to come through DT | 14:01 |
| freemangordon | not sure it is possible though | 14:01 |
| freemangordon | tmlind: ping | 14:02 |
| tmlind | freemangordon: pong | 14:03 |
| freemangordon | tmlind: I was able to get calls working on d4 without motmdm hack (the one that sets dai tdm etc) | 14:06 |
| freemangordon | by using audio-graph-card2 and codec2codec link | 14:06 |
| freemangordon | so, voice dai gets properly powered up etc on call | 14:06 |
| freemangordon | however, I am not able to make kernel set proper tdm/clocl/format | 14:07 |
| freemangordon | so, shall I hack cpcap_voice_hw_params() to do that, by either calling asoc_simple_hw_params() or directly? | 14:08 |
| freemangordon | or this is not a hack, but the proper way? | 14:08 |
| freemangordon | tmlind: ^^^ | 14:09 |
| arno11 | freemangordon: codec2codec uses 48000 rate ? | 14:14 |
| tmlind | freemangordon: cool that's great news | 14:15 |
| freemangordon | arno11: no, but I have no idea what rate it uses | 14:16 |
| freemangordon | arno11: but that's irrelevant anyways | 14:16 |
| freemangordon | tmlind: with kernel 6.6 that is, 6.1 does not support c2c links | 14:16 |
| tmlind | freemangordon: so with the set_tdm whatever calls you can configure the mcbsp time divsion multiplexing, not sure what clock needs to be set to | 14:17 |
| freemangordon | yes, I know | 14:17 |
| tmlind | i recall voice calls need to change the clock to a different value from hifi and voice codecs? | 14:18 |
| freemangordon | yes, correct | 14:18 |
| freemangordon | but the point is - shall I do that from cpcap_voice_hw_params() ? | 14:18 |
| freemangordon | somehow set_tdm gets called for hifi | 14:19 |
| tmlind | hmm not sure how that should be handled properly.. so if a voice call needs to change the mcbsp clock, what are hifi and voice going to do? | 14:19 |
| freemangordon | but not for voice | 14:19 |
| tmlind | has voice registered the set_tdm callback? | 14:20 |
| freemangordon | yes | 14:20 |
| freemangordon | but it is not called | 14:20 |
| tmlind | weird, maybe there's some bug that it only gets called to the first or last registered codec and not all of them? | 14:20 |
| freemangordon | could be | 14:20 |
| tmlind | so it might be able to write the voice call codec to just do qmi call with no serial n_gsm stuff.. that is if the settings are just modem nvram value changes | 14:21 |
| tmlind | might be worth diffing the nvram before and after voice call volume change for example | 14:22 |
| freemangordon | yeah, but I want it first working as it is, witjout the hack | 14:22 |
| tmlind | rigth | 14:22 |
| tmlind | it should not matter what the control bus is :) | 14:22 |
| tmlind | no idea about the codec2codec link | 14:23 |
| freemangordon | see v | 14:24 |
| freemangordon | https://pastebin.com/JGEKkyrj | 14:24 |
| freemangordon | yeah c2c is something new | 14:24 |
| freemangordon | https://kernel.org/doc/html/latest/sound/soc/codec-to-codec.html | 14:24 |
| freemangordon | more or less our use case - audio playback without CPU involved | 14:25 |
| freemangordon | so I made it work in terms of DAPM | 14:25 |
| tmlind | ok cool | 14:25 |
| freemangordon | I can hear noise and the other side hears me more or less ok | 14:25 |
| freemangordon | noise changes when the other side speaks | 14:26 |
| tmlind | so the old hack you're trying to fix is the voice call codec just rudely reconfiguring registers? | 14:26 |
| freemangordon | the hack I am trying to fix is motmdm calling set_tdm, set_ckl etc of cpcap | 14:26 |
| freemangordon | when call is registered | 14:27 |
| tmlind | ok | 14:27 |
| freemangordon | this https://github.com/maemo-leste/droid4-linux/blob/maemo-6.6.y/sound/soc/codecs/motmdm.c#L548 | 14:28 |
| freemangordon | so, without dt changes, nothing happens on call | 14:28 |
| freemangordon | with then cpcap_voice_hw_params() gets called | 14:29 |
| freemangordon | *with them | 14:29 |
| tmlind | oh ok | 14:29 |
| freemangordon | so, basically, I want to know if I shall do like this https://github.com/maemo-leste/droid4-linux/blob/maemo-6.6.y/sound/soc/tegra/tegra_audio_graph_card.c#L165 | 14:30 |
| tmlind | are you seeing similar voice call modem status messages also over the ttyUSB4 or ttyUSB5 whatever the at port is? | 14:31 |
| freemangordon | no idea | 14:31 |
| freemangordon | I didn't check as I am focused on making it work as is (without the hack ofc) | 14:31 |
| tmlind | ok | 14:32 |
| freemangordon | I guess we'll be able to get rid of dlci alltogether | 14:32 |
| freemangordon | but not now :_) | 14:32 |
| freemangordon | :) | 14:32 |
| freemangordon | so, what do you think about https://github.com/maemo-leste/droid4-linux/blob/maemo-6.6.y/sound/soc/tegra/tegra_audio_graph_card.c#L165 ? | 14:32 |
| freemangordon | shall we do the same? | 14:32 |
| freemangordon | if asoc_simple_hw_params() is called, that would set TDM/clocks, whatever from DT | 14:33 |
| freemangordon | but I don't know if this is the proper way | 14:33 |
| freemangordon | thus my question :) | 14:33 |
| tmlind | sure the more generic the easier to maintain, i just don't know enought about also to comment really | 14:33 |
| tmlind | maybe there's some tdm configuration that would work for everything.. | 14:34 |
| freemangordon | ugh, asoc_simple_hw_params() was dropped in 6.7 | 14:35 |
| tmlind | heh ok | 14:35 |
| freemangordon | well, it was replaced with simple_util_hw_params | 14:35 |
| freemangordon | which I guess does the same | 14:35 |
| freemangordon | yeah, they just renamed it | 14:36 |
| freemangordon | ok, I'll do that | 14:37 |
| freemangordon | will ping you when I'm ready to see if you like it :) | 14:37 |
| tmlind | ok yeah i'll check the irc every few days at least | 14:38 |
| freemangordon | thanks! | 14:38 |
| tmlind | thank you | 14:40 |
| arno11 | Wizzup: i built drnoksnes a bit differently to avoid the buggy additional deps. it works fine from HAM :) (only for armhf and no gui atm) | 15:20 |
| arno11 | works ootb using '/usr/bin/drnoksnes -c /usr/games/drnoksnes.conf /game_path' | 15:23 |
| arno11 | fullscreen works (4:3) | 15:24 |
| arno11 | Ctl+Shift+N for compo OFF and 'f' for fullscreen | 15:29 |
| Wizzup | great, back in a few ours | 15:35 |
| Wizzup | phoenix is back, leste is back, will fix build vm today | 19:41 |
| freemangordon | Wizzup: not that great, as I still can't make it work properly, but yeah, some progress | 22:43 |
| freemangordon | please, when you have some spare time, try to understand https://elixir.bootlin.com/linux/v6.10-rc6/source/sound/soc/generic/audio-graph-card2-custom-sample.dtsi#L113 | 22:44 |
| Wizzup | ok, probably another day or two out :) | 22:52 |
| Wizzup | at leas tmost infra is back | 22:52 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!