| freemangordon | Wizzup: got it working without hacks, it even works properly when call is received while OMP plays music | 14:29 |
|---|---|---|
| freemangordon | but, can we jump to newer than 6.6? | 14:29 |
| freemangordon | as otherwise I will not be able to send a patch upstream, as audio-card-graph2 has changes after 6.6 will will not allow the patch for 6.6 to be applied | 14:30 |
| freemangordon | s/will will/which will/ | 14:31 |
| freemangordon | BTW, I am going to push the chages, if you don't mind | 14:31 |
| freemangordon | uvos: ^^^ ? | 14:32 |
| Wizzup | please do push, we *can* go for newer than 6.6, but I would like to aim for 6.6 firs | 14:32 |
| freemangordon | ok, will do | 14:32 |
| freemangordon | and will think of upstreaming afterwards | 14:32 |
| Wizzup | so we can have a 6.6 branch for us and then work on a 6.newer branch? | 14:32 |
| freemangordon | yes | 14:32 |
| Wizzup | freemangordon: do you test with sphone's ofono backend or the voicecallmanager one? | 14:32 |
| freemangordon | whatever comes as default | 14:32 |
| Wizzup | also you cannot comprehend how happy I am with this news, truly | 14:32 |
| Wizzup | in devel that is voicecallmanager | 14:33 |
| freemangordon | mhm | 14:33 |
| freemangordon | BTW, sphone does not survive PA restart | 14:33 |
| freemangordon | ok, lemme clean-up the patches and push | 14:33 |
| Wizzup | \o/ \o/ | 14:35 |
| freemangordon | Wizzup: https://github.com/maemo-leste/droid4-linux/commits/maemo-6.6.y/ | 15:48 |
| freemangordon | tmlind: please have a look if you have some spare time | 15:49 |
| freemangordon | I have no idea if that's the proper way to do it :) | 15:49 |
| freemangordon | also, I think set_tdm shall be removed too | 15:49 |
| freemangordon | from cpcap that is | 15:49 |
| freemangordon | though that might be needed for BT etc | 15:50 |
| freemangordon | Wizzup: please, make new kernel release for experimental, so the others to be able to test | 15:52 |
| freemangordon | we need uvos to merge messing patches though | 15:53 |
| freemangordon | *missing | 15:53 |
| freemangordon | tmlind: I think https://github.com/maemo-leste/droid4-linux/commit/3e9ab86f6b6c90495a12e3bc92487b211bded86b shall be dropped, however, I was not able to find anything about https://github.com/maemo-leste/droid4-linux/commit/3e9ab86f6b6c90495a12e3bc92487b211bded86b#diff-3943bf3a33be67e51248be2256236bc49d57d158b6963870c0410b4853303cf7R1396 | 15:58 |
| freemangordon | do we have a knob for that? or, shall I create one? | 15:58 |
| Wizzup | freemangordon: the charging patches? | 16:11 |
| freemangordon | Wizzup: yes, see yesterday's backscroll | 16:13 |
| freemangordon | "(17,29,37) uvos: yeah i missed merging in the whole cpcap-charger feature branch" | 16:13 |
| Wizzup | yeah | 17:14 |
| Wizzup | I don't know where this branch | 17:21 |
| freemangordon | lets wait for uvos maybe | 17:25 |
| freemangordon | he said he'll merge them | 17:25 |
| Wizzup | imo we can just build one now and do it again later | 17:28 |
| Wizzup | I'd like to toy with it | 17:28 |
| Wizzup | what branch did you push to? | 17:28 |
| arno11 | Wizzup: btw for the n900 dts mess, i know what to do to avoid that, so if uvos have no time to check, i can make a PR if you want (and remove the 850 freq) | 17:29 |
| freemangordon | Wizzup: 6.6.y | 17:32 |
| Wizzup | do we need modified UCMs? | 17:32 |
| freemangordon | no | 17:32 |
| Wizzup | sweet | 17:32 |
| freemangordon | however, I am not really sure about this https://github.com/maemo-leste/droid4-linux/commit/d935fcb8b3fcb87947317073c88a7aa7816518c8#diff-ebae04ce1c30a63fd60be92eb12f924f07a3e95128805c3fdc9b2e94a0bc04f0R107 | 17:33 |
| freemangordon | like, it works, but I have NFC why :) | 17:33 |
| freemangordon | IOW - I don't know if we need ("Voice Call Capture", "Internal Mic") c2c route | 17:34 |
| Wizzup | I'd have to read what the routing part in the soundcard does | 17:34 |
| freemangordon | please do, I am having hard times understanding that | 17:35 |
| freemangordon | I am almost sure this route should not be there, however, all my tests were performed with that present | 17:39 |
| freemangordon | Wizzup: another issue with sphone, besides crash on PA restarted - there is no ringing tone if OMP plays music when call comes. OMP properly pauses the music though | 17:41 |
| freemangordon | bbl | 17:42 |
| uvos | i dont think this is a issue with sphone | 18:18 |
| uvos | rather mafw dose something wierd or pa breaks. for the ringtone sphone dosent do anything with pa at all no switching of any kind no restoreing of volumes or anything like that | 18:19 |
| uvos | it simply delegates gst to play a file - thats it | 18:19 |
| uvos | i see no reason why omp would pause besides pa breaking and omp pausing because pa no longer accepts samples on its soccet | 18:20 |
| uvos | (due to it being in a bad state or it crashing) | 18:21 |
| uvos | or its the kernels fault, with the kernel changeing something due to the modem comeing in voicecall state early maybe | 18:23 |
| uvos | but again here sphone dosent post a single call to pa's interfaces untill the call enters the active state | 18:23 |
| uvos | you can double check this by not loading sphones pa module at all - the ringtone is played by its gst module which only communicates with gst so that will still work | 18:26 |
| uvos | you will note that the behavior is exactly the same | 18:26 |
| Wizzup | uvos: can you link to the cpcap charger branch so I/you can merge it in? | 21:05 |
| Wizzup | arno11: then we can also get the n900 fix in | 21:05 |
| arno11 | yep | 21:14 |
| arno11 | Wizzup: so do you want a PR ? | 21:15 |
| Wizzup | arno11: that works for me | 21:18 |
| arno11 | ok | 21:18 |
| uvos | dont bother | 21:40 |
| uvos | i have done this allready | 21:40 |
| arno11 | argh i just opened a PR lol | 21:40 |
| uvos | well i said id do it :P | 21:41 |
| arno11 | ok thx :P | 21:41 |
| uvos | freemangordon: uh can you not do this sort of thing https://github.com/maemo-leste/droid4-linux/commit/351358e32794e60ee39e895a01b1153c73e7492d | 21:42 |
| uvos | regarding moveing beyond 6.6 | 21:43 |
| uvos | i am against it | 21:43 |
| uvos | since we do tend to get stuck on kernel versions for this reason or that reason | 21:44 |
| uvos | and i really really want to be stuck on a lts | 21:44 |
| uvos | since that gives us plenty of time to figure out whats wrong with the new kernel | 21:44 |
| uvos | we can upgrade again in nov/december when this years lts gets released | 21:45 |
| uvos | Wizzup: freemangordon: is there something else missing here or shal i tag a release? | 21:45 |
| Wizzup | uvos: ok, where is the stuff pushed? | 21:50 |
| Wizzup | there's n900 dts and charger stuff for cpcap | 21:50 |
| uvos | 6.6.yh | 21:50 |
| uvos | btw vcm has some code that seams to deal with audio routing | 21:53 |
| uvos | possibly that its breaking things by doing things we dont want | 21:53 |
| arno11 | Wizzup: btw on n900 there is still the 6.6 headers error with iphb-dkms | 21:56 |
| arno11 | but according to irc logs you already fixed a similar issue in march (with arm64) | 21:57 |
| Wizzup | uvos: thanks, I think we should do an experimental build | 22:21 |
| Wizzup | arno11: ok, thanks for doing the digging, I'll promise to take a look tomorrow (got up at 4am so I'm kind of wrecked now :)) | 22:21 |
| Wizzup | I wonder if dpkg -L shows anything obviously wrong between the header packages for 6.1.x an 6.6x | 22:22 |
| arno11 | ok no probs man ;) | 22:26 |
| freemangordon | uvos: you are against what? the patch you linked basically implements the specs | 22:40 |
| freemangordon | and I will send similar patch for upstreaming | 22:40 |
| uvos | freemangordon: i dislike the fact that its called just xxx, when i inevitably have to rebase this stuff for the next kernel i will be looking at this patch trying to figure out if we still need it | 22:42 |
| uvos | it being called just xxx wont help | 22:42 |
| freemangordon | oh, did I really do that? | 22:42 |
| freemangordon | sorry | 22:42 |
| uvos | np | 22:42 |
| freemangordon | I just named it loke that to merge it to upstream tree and forgot to rename it afterwords | 22:43 |
| uvos | ok, i see | 22:43 |
| freemangordon | maybe rename it and rebase | 22:43 |
| freemangordon | or I can do that | 22:43 |
| uvos | i dont want to force push this atm | 22:43 |
| uvos | as Wizzup and arno are working on things i think | 22:43 |
| freemangordon | again, sorry, that was just an overlook | 22:44 |
| uvos | no worrys | 22:45 |
| uvos | btw i found the bug that causes sphone to fail on pa restart | 22:45 |
| uvos | will fix | 22:45 |
| freemangordon | I was so happy that everything works properly that forgot to rename the patch | 22:45 |
| freemangordon | ok, thanks | 22:45 |
| freemangordon | uvos: mafw/omp behave properly: OMP *pauses* playback on incoming call and *resumes* on hangup | 22:46 |
| freemangordon | there is no bug there | 22:46 |
| uvos | it should pause when pa disconnects the stream | 22:47 |
| freemangordon | you can see the playback/pause buttons change | 22:47 |
| uvos | but this should only happen when sphone switches ucm profile | 22:47 |
| freemangordon | no, it pauses before that ;) | 22:47 |
| freemangordon | which is the correct behaviour | 22:47 |
| uvos | no one is telling it to do this | 22:47 |
| uvos | afaik | 22:47 |
| uvos | i would not know who | 22:47 |
| uvos | sphone dosent do anything before the call is accepted | 22:48 |
| freemangordon | I have to check how it know, but, music should stop playing on call, no? | 22:48 |
| uvos | sure but its not implemented | 22:48 |
| freemangordon | it is | 22:48 |
| freemangordon | in OMP at least | 22:48 |
| freemangordon | you can check it by yourself | 22:48 |
| uvos | sure | 22:48 |
| uvos | but the other side is not implemented | 22:48 |
| uvos | or at least nothing in the stack i know of dose this | 22:49 |
| uvos | for sure not sphone | 22:49 |
| freemangordon | yeah | 22:49 |
| freemangordon | maybe if it use notifications (as you proposed back then) it will work properly | 22:49 |
| uvos | well notifications dont do ringger | 22:49 |
| uvos | that was just for sphone-sms which just i now use anyhow | 22:50 |
| uvos | regardless the ringer not working is very puzzeling as sphone only calls gstreamer | 22:50 |
| freemangordon | actually they do | 22:50 |
| uvos | as a totaly normal application and dose not do any profile switching until the call is accepted | 22:51 |
| uvos | besides i can repo the problem with sphones pa module not even loaded | 22:51 |
| freemangordon | yes, I understand | 22:51 |
| uvos | so something is up somewhere else, but i dont know what | 22:51 |
| freemangordon | mhm | 22:51 |
| sicelo | i assume OMP gets info via mce? | 22:51 |
| freemangordon | maybe libplayback or dunno | 22:51 |
| freemangordon | yes, most probably | 22:52 |
| sicelo | s/info/call info/ | 22:52 |
| uvos | iirc mce recives call state | 22:52 |
| uvos | but no interface for anyone to recive the call state from mce | 22:52 |
| uvos | nvm it dose | 22:53 |
| freemangordon | sig_call_state_ind | 22:53 |
| uvos | ok maybe omp direcly asks mce for a the call state | 22:54 |
| freemangordon | com.nokia.mce.signal | 22:54 |
| freemangordon | no, mce sends signal | 22:54 |
| uvos | sure | 22:54 |
| freemangordon | https://github.com/maemo-leste-extras/openmediaplayer/blob/master/player/missioncontrol.cpp#L62 | 22:54 |
| uvos | so yes ok maybe omp recives the state from mce | 22:54 |
| uvos | sphone dose tell mce about the ringging state | 22:55 |
| uvos | but that dosent explain why gst fails to play sphones audio, but only in the ringging state (and call ofc but thats expected) | 22:55 |
| uvos | it works fine if you trigger sphones ringger by hand | 22:55 |
| uvos | the code in sphones gst plugin goes down the happy path with no errors too when this happens | 22:56 |
| uvos | so yeah not sure whats going on | 22:57 |
| freemangordon | well, that's not a big issue | 22:58 |
| freemangordon | but what I think is a biggher issue is the power usage | 22:59 |
| freemangordon | my battery went fat for 3 hours or so | 22:59 |
| freemangordon | maybe because of the missing patches, dunno | 22:59 |
| freemangordon | but Wizzup said something about cpufreq | 22:59 |
| freemangordon | hmm, IIRC on 6.1 scaling governor was menu | 23:14 |
| freemangordon | now it is ondemand | 23:14 |
| Wizzup | I think we can make this a config change easily | 23:14 |
| uvos | this makes essentally no difference for pm | 23:15 |
| arno11 | it was already ondemand in 6.1 btw | 23:15 |
| uvos | if its dieing in 3h there is some other problem | 23:15 |
| uvos | but the scaleing frequencies are almost irrelivant | 23:15 |
| uvos | because the cpu should spend all its time in ret anyhow | 23:15 |
| freemangordon | maybe missing iphb is the reason | 23:25 |
| freemangordon | will check again with a properly built kernel | 23:26 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!