| Wizzup | arno11: yea it's supposed to run in system mode | 01:16 |
|---|---|---|
| Wizzup | CynicalDebian: great | 01:57 |
| Wizzup | uvos: over the next few days I would like to try to make some progress on additional devices - droid3, razr, etc | 16:16 |
| Wizzup | if you're around I'd appreciate some help/thoughts etc | 16:16 |
| arno11 | Wizzup: sicelo: finally i fixed 2g calls issue. Sound is clear and latency is...well...very low...arround 5ms ! the problem is PA cpu usage: it causes audio sync troubles if other apps are running. 3g works with the same config with 300ms latency as usual. | 17:37 |
| arno11 | the 15 sec weird bug in incoming calls is still there | 17:38 |
| arno11 | (my 10 sec sleep trick doesn't work anymore) | 17:38 |
| arno11 | so i think i need help to solve this weird bug and need to talk about why PA is running in system mode (too much cpu usage for n900) | 17:41 |
| Wizzup | arno11: 5ms latency on 2g seem impossible? | 17:47 |
| Wizzup | do you mean PA latency | 17:47 |
| arno11 | yes pa latency | 17:48 |
| Wizzup | pa in system mode shouldn't use more cpu fwiw | 17:48 |
| arno11 | are you sure ? | 17:49 |
| Wizzup | why would it? | 17:53 |
| arno11 | i just remember reading stuff about it and the high cpu usage but maybe i'm wrong | 17:56 |
| arno11 | anyway the cpu usage is really to high | 17:56 |
| bencoh | well, I often get high cpu usage and crackling audio when using twinkle (sip client) on droid4/leste | 17:56 |
| bencoh | not sure if it is related | 17:56 |
| Wizzup | I don't see why system pa would take more cpu tbh | 17:57 |
| bencoh | (pulseaudio* cpu usage) | 17:57 |
| arno11 | interesting | 17:57 |
| bencoh | no idea about system vs user mode | 17:57 |
| arno11 | bencoh: did you noticed high xorg cpu usage using twinkle too ? | 17:58 |
| bencoh | not that I remember | 17:58 |
| arno11 | ok | 17:58 |
| bencoh | (I use it mostly in commandline mode anyway) | 17:59 |
| arno11 | yes but for example on n900 during calls, xorg burns 40% cpu in command line | 18:00 |
| arno11 | so weird | 18:00 |
| Wizzup | xorg burning cpu is weird yes | 18:00 |
| Wizzup | that we want to figure out for sure | 18:00 |
| arno11 | yep | 18:00 |
| arno11 | i'll try to monitor what's going on with xorg | 18:03 |
| arno11 | during calls | 18:03 |
| Wizzup | freemangordon might have some idueas | 18:05 |
| Wizzup | arno11: does it also happen if you don't use sphone btw? | 18:05 |
| Wizzup | or if the screen is off? | 18:05 |
| arno11 | oh good question | 18:05 |
| arno11 | pretty sure it happens with screen off but not sure for sphone | 18:06 |
| arno11 | let me check | 18:06 |
| arno11 | so weird: 2 pulse processes, 20 and 40%. xorg 25 | 18:14 |
| bencoh | uh | 18:14 |
| arno11 | even without sphone | 18:15 |
| arno11 | using ofono scripts | 18:15 |
| arno11 | Wizzup: i already asked freemangordon about xorg few weeks ago | 18:19 |
| arno11 | something is really wrong: creating a loopback using arecord/aplay, PA burns only 10% and xorg 10 | 18:24 |
| arno11 | Wizzup: the main question is why things are working differently with calls since the last kernel update ? or maybe PA reacts differently when we use UCM files ? | 18:36 |
| Wizzup | arno11: I don't think pulseaudio and x really interact | 18:48 |
| Wizzup | interesting though @ 25% | 18:48 |
| arno11 | i think so | 18:49 |
| arno11 | anyway we have few remaining troubles but 2g calls quality is really good now | 18:53 |
| Wizzup | great | 18:54 |
| Wizzup | what did you change? | 18:54 |
| Wizzup | when you say sync troubles with other apps, do you mean when other applications are playing sound? | 18:54 |
| arno11 | just if apps burn cpu | 18:55 |
| arno11 | not necessary playing sound | 18:55 |
| arno11 | oh didn't see your question about changes | 18:58 |
| arno11 | few lines in PA daemon.conf again, one sleep in cmtspeech_ofono_test and fragsize/tlength set to 256 ! | 19:00 |
| arno11 | this way 2g calls work well and no inpact on 3g calls | 19:01 |
| arno11 | *fragsize/minreq/tlength | 19:03 |
| arno11 | in pulse.c | 19:03 |
| Wizzup | hm, ok | 19:11 |
| Wizzup | can you share a diff or git patch against something? | 19:11 |
| arno11 | sure | 19:13 |
| freemangordon | sorry guys, I am very short on spare time recently :( | 19:21 |
| Wizzup | uvos: yo yo | 22:24 |
| Wizzup | uvos: so I hope to make some time for the d3 and razr soon | 22:34 |
| Wizzup | I'll be in the US for three weeks where I can't really use the d4 | 22:34 |
| uvos | Wizzup: okay | 22:38 |
| uvos | great | 22:38 |
| uvos | us razr wont work either | 22:40 |
| uvos | eu razr will | 22:40 |
| Wizzup | because of the spi modem? | 22:41 |
| Wizzup | oh | 22:41 |
| Wizzup | right | 22:41 |
| Wizzup | uvos: yeah, but I will probably need some guidance from you :D | 22:41 |
| Wizzup | I might have this one sent to the office https://www.ebay.com/itm/364123083812 because I think I mostly have xt912's even though they were called xt910's | 22:43 |
| uvos | okay | 22:45 |
| uvos | well for d3 dumping emif regs would be usfull | 22:45 |
| uvos | the same is usefull for the crtc controller on xt91x | 22:46 |
| Wizzup | okay, so this requires flashing a certain android version, rooting it, and then running a certain tool? | 22:46 |
| Wizzup | right, the razr had some display update issues right? | 22:46 |
| Wizzup | (I'm asking now so then I'll update the github issues) | 22:51 |
| uvos | display roles around | 22:51 |
| uvos | like a tv out of hsync | 22:51 |
| Wizzup | I see | 22:52 |
| Wizzup | I don't remember, did I send a droid 3 to both you and tmlind ? | 22:52 |
| uvos | i have one yes | 22:53 |
| Wizzup | There is a bullet point for me to do that here https://github.com/maemo-leste/bugtracker/issues/647 | 22:53 |
| uvos | no idea if you sent one to tmlind | 22:53 |
| uvos | ofc | 22:53 |
| Wizzup | https://github.com/maemo-leste/bugtracker/issues/600 | 22:53 |
| Wizzup | I guess this is what you meant? | 22:53 |
| uvos | ye | 22:53 |
| Wizzup | (with emif?) | 22:53 |
| uvos | s | 22:53 |
| uvos | yes | 22:53 |
| uvos | no | 22:53 |
| uvos | not sure what that was about at this point | 22:53 |
| uvos | but its about dumping cpcap not emif | 22:54 |
| Wizzup | and I guess also backlight control? | 22:54 |
| Wizzup | sorry, that is razr | 22:54 |
| Wizzup | (was looking at https://github.com/maemo-leste/bugtracker/issues/648 too) | 22:54 |
| uvos | i mean razr has no backlight | 22:57 |
| Wizzup | do you mean backlight doesn't work currently? | 22:58 |
| uvos | no it dosent have one | 22:58 |
| uvos | its oled | 22:58 |
| uvos | you have to speak to the display controller to set bightness | 22:58 |
| uvos | no idea how this work exactly | 22:58 |
| Wizzup | through xrandr? | 23:01 |
| Wizzup | xrandr --output foo --brightness 0.5 | 23:01 |
| uvos | no | 23:01 |
| uvos | you can tell the display to change brighness, this has mutch more dynamic range than the 255 usual values | 23:02 |
| Wizzup | ah, so it is through /sys | 23:02 |
| uvos | well no - what i mean is i dont know how to tell the pannel to do this on razr | 23:02 |
| Wizzup | ok | 23:03 |
| uvos | like in the electrical interface and kernel sense | 23:03 |
| Wizzup | ok | 23:03 |
| uvos | maybe it goes via dsi commands, maybe there is some i2c side channel or so | 23:04 |
| uvos | really no idea | 23:04 |
| uvos | but looking at the pannel code in android should do it | 23:04 |
| Wizzup | probably in the android kernel somewhere | 23:04 |
| Wizzup | yeah | 23:04 |
| * sicelo got an N9 the past week | 23:05 | |
| Wizzup | :) | 23:06 |
| sicelo | will try mainline once i have some time to spare | 23:07 |
| Wizzup | cool | 23:10 |
| Wizzup | we have this issue for them https://github.com/maemo-leste/bugtracker/issues/535 | 23:10 |
| Wizzup | I'm trying to make screenshots of maemo weather for the news post but the font has all these ants | 23:17 |
| Wizzup | :D | 23:17 |
| Wizzup | anki works ok with default theme and 2.0 scale | 23:28 |
| Wizzup | maybe 1.5 is better | 23:28 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!