tmlind | no idea on the voice call mic issues.. but i'm slowly rebasing patches for v5.18, there are bunch of n_gsm fixes that just got merged, also fixed a few module reload issues for the d4 serdev drivers but have not yet pushed out anything, maybe over the weekend | 06:20 |
---|---|---|
tmlind | unlikely any of these affect m-l, but being able to reload the modules makes development easier, that's been buggy for a while | 06:21 |
Wizzup | agreed, great | 06:55 |
Wizzup | tmlind: but you have voice calls working on non-speakerphone for you still? | 07:23 |
tmlind | Wizzup: i only have been using speakerphone and some hackish scripts, no ofono so far.. | 07:24 |
Wizzup | ok, how is noice cancelling for you? | 07:24 |
Wizzup | yeah ofono doesn't manage our audio routes either | 07:25 |
Wizzup | sorry, echo cancelling* | 07:25 |
Wizzup | when I tried speakerphone it was not usable due to echo | 07:25 |
Wizzup | I can load the pa echo cancelling module and try that of course | 07:28 |
Wizzup | I had the switch on in alsamixer I think | 07:28 |
tmlind | Wizzup: mdm6600 has built in noise cancellation, i keep that enabled in the alsamixer all the time | 07:42 |
Wizzup | yeah I had that on too | 08:03 |
Wizzup | maybe it needs to be on at the start of the call and I didn't have it then or something | 08:27 |
Wizzup | not sure what else it could be | 08:27 |
uvos__ | Wizzup: we cant add pa noise cancelation or any filter whatsoever since the audio never passes throug the cpu | 09:35 |
uvos__ | that said the built in echo cancelation worked fine when i used it | 09:35 |
uvos__ | s/noise/echo | 09:36 |
Wizzup | uvos: I think we can, since the audio leaves the speakers? | 12:01 |
uvos__ | Wizzup: no the speakers are directly connected to the modem in this mode | 12:03 |
Wizzup | aha | 12:03 |
uvos__ | nothing you do on the cpu can change the audio ouptut | 12:03 |
Wizzup | ok, well, I am pretty sure it was on in alsamixer but it was clearly not working | 12:03 |
uvos__ | its all in hw the omap4 can be off in a call for all the modem cares | 12:03 |
Wizzup | right | 12:03 |
uvos__ | there is a way you could do it tho | 12:03 |
Wizzup | I don't think we need to investigate that way atm, I'd rather figure out why the echo cancel doesn't work even though it's on | 12:04 |
uvos__ | theres a modem<->cpu dai that android uses for call recording | 12:04 |
Wizzup | aha | 12:04 |
uvos__ | you could use that path all the time and disconnect the speakers from the mode | 12:04 |
uvos__ | but this means looksing the awsome talk time the d4 has | 12:04 |
Wizzup | right | 12:05 |
Wizzup | btw, I tried to use ofono + data again on the d4 today while travelling and about every time I do it, it works for a bit, and then bugs in our ofono driver bite me hard | 12:06 |
Wizzup | I'll try to do a few more calls when I get home in an hour or two, see if I can get echo cancellation to work | 12:06 |
uvos__ | iirc cpcap provides eco cancellation not the mdm | 12:07 |
uvos__ | so maybe check the datasheet for the bit | 12:07 |
Wizzup | if you know how I can figure out the required registers for earpiece/hp... that'd be neat | 12:07 |
uvos__ | but im pretty sure it was working | 12:07 |
Wizzup | maybe writing to the same register somehow causes issues | 12:07 |
Wizzup | I will try again | 12:08 |
Wizzup | I guess for input mostly the handset matters, not hp so much | 12:16 |
uvos__ | Wizzup: https://github.com/maemo-leste/sphone/blob/1eaa6f40a86748c327420321c54416e30659a20d/src/modules/route-mapphone-hack.c#L123 | 14:46 |
uvos__ | your freeing an uninitalized pointer | 14:46 |
uvos__ | (you fail to set data in init) | 14:46 |
Wizzup | uvos__: let me fix that | 14:55 |
Wizzup | the code wasn't quite ready for normal usage yet :) | 14:55 |
tmlind | Wizzup, uvos__: pushed out an updated v5.18-rc6 based topic branch for d4 modem serdev stuff: https://github.com/tmlind/linux/commits/serdev-gsm-pending-v5.18 | 15:21 |
tmlind | as always, had to first fix yet another kernel regression for the phy, one for n_gsm debug.. rebased on the recern n_gsm changes, fixed up issues on reloading the d4 serdev modules | 15:22 |
tmlind | probably no need to backport anything to m-l | 15:22 |
tmlind | will do a v5.18-rc based topic branch for the pending pvr patches, i think i have one fix i should improve a bit and then send out, then backport freemangordon's patches in linux next | 15:23 |
tmlind | but that will likely be few days | 15:24 |
tmlind | hmm i also have a musb pm fix for dumb chargers that i need to test | 15:24 |
tmlind | looks like with some dumb chargers musb never idles currently.. | 15:25 |
Wizzup | I can try to rebase on m-l tree on this, I suppose we don't have the pvr stuff on top of it yet? | 15:25 |
tmlind | Wizzup: best to wait few days so we get a much cleaner pvr topic branch | 15:27 |
Wizzup | ok, I can wait | 15:27 |
tmlind | i think we pretty much only have one pvr kernel patch left now :) | 15:27 |
Wizzup | I plan to put me teeth in ofono on the d4 and not let go once it works in a stable manner - it's been the source of much frustration in my day to day usage :P | 15:27 |
Wizzup | right, sweet | 15:27 |
tmlind | well one patch left for omapdrm for mainline i mean, the whole pvr kernel driver is a big mess | 15:27 |
Wizzup | right | 15:28 |
tmlind | Wizzup: yeah ok, i suggest you use some temporary shell script to change the cpcap voice output if you need that for own usage | 15:28 |
tmlind | well could be clock changes are also needed.. uvos knows | 15:29 |
tmlind | hmm so the serial console issue again.. so that happens on shutdown or reboot? | 15:30 |
tmlind | and serial console needs to be detached for it to happen? | 15:31 |
tmlind | i could not reproduce the last time i tried with just a kernel and a minimal filesystem so some more info would be nice | 15:32 |
tmlind | too much computing for me already today, will be mostly offline for the rest of the day though :) | 15:33 |
Wizzup | tmlind: with voice output, do you mean for speakerphone/headphone/earpiece, or you mean input? | 15:33 |
Wizzup | tmlind: if for output from the call, I'm working on a temporary sphone module for that | 15:33 |
Wizzup | if for input from the call, I don't know how to get it to work with anything but speakerphone yet | 15:33 |
Wizzup | input for the call* | 15:33 |
Wizzup | I think we should re-test the serial problems on 5.18 | 15:34 |
Wizzup | no need to try to reproduce it on older | 15:34 |
tmlind | Wizzup: ok yeah let's retest on 5.18 | 16:04 |
tmlind | Wizzup: sounds like you're all set with a temporary sphone module then, that's better than a shell script :) | 16:05 |
Wizzup | tmlind: right, apart from the fact that switching to earpiece or headphone is not useful as the input doesn't work then, so the other side cannot ehar me | 16:06 |
Wizzup | hear me* | 16:06 |
Wizzup | so really only speakerphone 'works' at this point (and I still need to re-test the echo cancellation alsamixer stuff again) | 16:06 |
uvos__ | ucm should set the echo cancellation correctly | 16:08 |
uvos__ | if it dosent its a bug | 16:08 |
Wizzup | it was set but didn't work, but I will re-test to rule out pebcak | 16:09 |
uvos__ | ok | 16:09 |
buZz | ait, borrowing a pinephone soon | 17:56 |
buZz | oh nice, voice routing stuff coming soon? :) my mom will be happy | 17:57 |
buZz | i was discussing yesterday with dreamer a bit, and we both thought it would be cool if we could run asterisk and get the voice routing into that | 17:58 |
dreamer | hehe | 18:00 |
dreamer | something we've been saying for years already | 18:00 |
dreamer | no clue how doable though | 18:00 |
Wizzup | buZz: well I'd be happy when it works with headset, not just speakerphone | 18:03 |
Wizzup | but we'll get there I think | 18:03 |
Wizzup | hmmm dreamer, don't you run android on those things? | 18:03 |
sixwheeledbeast | Does it even need asterisk? | 18:03 |
dreamer | Wizzup: I have a dual boot and a leste | 18:04 |
dreamer | but personal dev focus lately has been nearly entirely on music :# | 18:05 |
buZz | sixwheeledbeast: the idea was doing locally run voicemail mostly | 18:11 |
buZz | and call screening, possibly some IVR to catch spammers etc | 18:11 |
sixwheeledbeast | oh i see, I am not much of a voicemail user. I assume telepathy-gabble will be able to do calls via XMPP like N900? | 18:13 |
sixwheeledbeast | well fremantle | 18:14 |
Wizzup | sixwheeledbeast: yes, planned | 18:21 |
Wizzup | dreamer: ok, maybe you can help get some debug regs later | 18:21 |
dreamer | Wizzup: regs? | 18:26 |
Wizzup | sry mtg | 18:27 |
dreamer | one of my installs is the "dumb user doing dist-upgrade once every 6-12 months" which imo is a good indication for stability ;) | 18:27 |
dreamer | mtg? | 18:27 |
Wizzup | meeting | 18:27 |
dreamer | ah :) | 18:29 |
buZz | dreamer: there's some registers on the modem/baseband or something that determine how/where audio gets routed | 18:49 |
buZz | the android might help to show which settings are used for what :) | 18:50 |
Wizzup | right and I think uvos__ had done that before | 18:51 |
dreamer | buZz: aha | 18:51 |
lel | MerlijnWajer opened an issue: https://github.com/maemo-leste/bugtracker/issues/617 (pinetab image loads pinephone braveheart dts) | 19:47 |
Wizzup | uvos__: thx for that find btw @ not setting data | 20:49 |
Wizzup | uvos__: can you ping me when you're home? | 20:52 |
uvos | Wizzup: im here | 22:04 |
Wizzup | uvos: unrelated but I just made a call, and the n900 started ringing, but ofono didn't let sphone know until I hung up | 22:18 |
Wizzup | sphone: comm-ofono: Error: Timeout was reached | 22:18 |
Wizzup | sphone: comm-ofono: call_added_cb: /motmdm_0/voicecall01 | 22:18 |
Wizzup | sphone: comm-ofono: call: /motmdm_0/voicecall01, state: Active, line identification , emergency: 0 | 22:18 |
Wizzup | sphone: comm-ofono: ofono_voice_call_properties_add_handler: /motmdm_0/voicecall01 | 22:18 |
Wizzup | sphone: manager: check_needed_state: call state Active | 22:18 |
Wizzup | sphone: manager: check_needed_state: incall true, incall_no_route false | 22:18 |
Wizzup | sphone: sphone-mce: call_mode_trigger: active | 22:18 |
Wizzup | one more for the bucket of ofono bugs :) | 22:18 |
Wizzup | also if there's an active call in this state (well, according to ofono or sphone), you can't "hang up" | 22:19 |
uvos | you sure that was ofonos fault? | 22:21 |
uvos | (the n900 thing) | 22:22 |
Wizzup | yes. 100% | 22:22 |
Wizzup | this is entire consistent with it not getting other notifications like data connection ready, sms received, etc | 22:22 |
Wizzup | it's random, of course | 22:22 |
uvos | ok | 22:22 |
uvos | Wizzup: ill do the reg dumping on android later | 22:23 |
Wizzup | ok :) | 22:23 |
uvos | like in 1h | 22:23 |
Wizzup | cool | 22:23 |
Wizzup | ok, this time when I toggled it (off, on) noise cancelling worked | 23:05 |
Wizzup | (had to get another person involved and get them to walk away) | 23:05 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!