| Wizzup | sicelo: I would look towards debian | 00:24 |
|---|---|---|
| Wizzup | sicelo: I don't think devuan forks upower | 00:24 |
| sicelo | mmm, what about the build-depend against systemd-dev? but sure, i'll test properly in the morning hopefully. now must sleep | 00:29 |
| sicelo | freemangordon: Wizzup: sneak preview of what i've been working on for our upower predicament - depends on https://gitlab.freedesktop.org/upower/upower/-/merge_requests/259 | 00:29 |
| sicelo | status-area-battery-applet: https://github.com/sicelo/status-area-applet-battery/commit/c96ae44f9d311aefc089b0b52d03231332700d7b | 00:30 |
| sicelo | mce: https://github.com/sicelo/mce/commit/cc8b4b279498bd0ec622bb73bf1b2946a8f9bab5 | 00:30 |
| sicelo | for now I haven't actually built and tested both of these, hence not submitting the PR just yet | 00:31 |
| Wizzup | sicelo: I think that build depend is ok | 00:31 |
| Wizzup | sicelo: just look at what we have currently, I think we have a -dev pkg for it | 00:31 |
| Wizzup | sicelo: cool @ chagnes :) | 00:32 |
| sicelo | oh cool then. will give it a go | 00:32 |
| Wizzup | I'm no battery expert but fallbacks are great | 00:32 |
| mkf | okay now this thing is stable, should i upgrade to daedalus? | 00:34 |
| sicelo | you really seem to know your way around this, so yes, do upgrade :-) | 00:35 |
| Wizzup | maybe make a backup of the sd card | 00:37 |
| freemangordon | sicelo: do you expect a review? | 06:28 |
| freemangordon | Wizzup: I'll give ananicy guys chance to reply till weekend | 06:31 |
| freemangordon | if they don't, I'll write our own daemon. or will add functionality as part of dsme | 06:33 |
| freemangordon | it already supports niceness from the command line | 06:33 |
| freemangordon | for cgroup we will simply use cgrulesengd | 06:38 |
| sicelo | freemangordon: maybe let me build all three, give it a quick test, then submit the MR for full review | 08:36 |
| freemangordon | I meant code review, as I saw some nits already :) | 09:06 |
| sicelo | sure, please share them | 09:32 |
| sicelo | i.e. yes, it is open to review :-) | 09:32 |
| Wizzup | sicelo: https://git.maemo.org/explore/code?q=exit&fuzzy=false | 12:10 |
| Wizzup | might get a bit painful on our few linux forks, but yeah :D | 12:21 |
| mkf | good evening. | 12:48 |
| mkf | internal error means a seg fault? | 14:09 |
| mkf | where should i look for core dump? | 14:09 |
| sicelo | Wizzup: cool. Will have a look at that. Otherwise, search per repository is still fine overall. | 14:45 |
| sicelo | back to my inhibit monologue 🤠... what happens if, while doing system updates via HAM, the user initiates shutdown? | 14:47 |
| sicelo | Everything is a nail when all you have is a hammer (and my hammer is inhibit locks) ... this is one area where an inhibit lock would fit well. when HAM UI is open, it would take the lock until it is done | 14:49 |
| sicelo | anyhow, long term project, but maybe I should open an issue so it doesn't get forgotten someday | 14:50 |
| arno11 | mkf: core dumps appear in /home/user | 15:05 |
| arno11 | Wizzup: @calls issue: seems to be a pulseaudio crash when it stops ringing (according to sphone logs). if no ringtone (silent mode), no crash | 15:08 |
| arno11 | i looked @PA changes between PA 14 and 16, the only diff is module-gconf (missing in PA 16), all other config files seem similar | 15:10 |
| arno11 | i tried to increase shm a bit more (and a lot more), same issue | 15:11 |
| arno11 | if i play a music file and call the n900 (silent mode), no issue and remixing between call audio and music works as usual | 15:15 |
| arno11 | so the issue seems related to ringtone but i really don't know how to monitor what's going on | 15:16 |
| sicelo | maybe stop PA and run it manually in debug mode, then hopefully it'll tell you what causes the crash? | 15:17 |
| arno11 | i see it in sphone logs: it crash @/src/pulse/thread-mainloop.c | 15:19 |
| arno11 | function pa_threaded_mainloop_stop() | 15:20 |
| arno11 | line 166 | 15:20 |
| arno11 | will try later with PA debug mode | 15:21 |
| arno11 | hmm in fact PA is aborting and sphone crash and restart with no issue | 15:25 |
| arno11 | any ringtone issue on daedalus d4 btw ? | 15:30 |
| freemangordon | sicelo: made one comment on mce patch, not sure how to link to it | 17:36 |
| sicelo | thanks! saw the email. will implement | 17:37 |
| dsc_ | freemangordon: ehm.. I have a AMD 3600x laying around if you want it for free | 18:00 |
| dsc_ | it will be faster than this 2011 cpu :x | 18:01 |
| Wizzup | switched this to forgejo: https://git.maemo.org/ | 18:58 |
| Wizzup | will add some repos and oauth2 a bit later | 18:58 |
| dsc_ | nice | 19:00 |
| arno11 | cool | 19:07 |
| mkftab | hello from converstions | 19:32 |
| mkfx | Hello mkf | 19:32 |
| mkftab | Wonderful, this thing works much better than chimaera | 19:32 |
| mkf | do we have debug symbols in repositories? | 20:53 |
| mkf | for maemo packages | 20:53 |
| arno11 | usually yes, pkg_name-dbgsym | 21:02 |
| mkf | wonderful. | 21:10 |
| mkf | going from beowulf to chimaera to daedalus have been quite a big bump on reousrces used in this poor tablet. hm. | 21:14 |
| mkf | i wonder if i can somehow conserve ram and cpu usage. | 21:15 |
| Wizzup | mkf: there should not have been an increase | 21:18 |
| arno11 | indeed, and you can even expect better perfs (a bit) with daedalus | 21:20 |
| mkf | i think it's not in leste side, but general gnu/linux stuff getting fatter over the releases | 21:22 |
| arno11 | hmm on n900, ram and cpu usage is quite low on daedalus, i mean not higher than chimaera | 21:27 |
| arno11 | btw, how things work with qt5 on your tablet ? | 21:27 |
| mkf | i'm not sure. | 21:27 |
| mkf | telegram wrecks havoc, converstions is peaceful | 21:28 |
| mkf | telegram used to be better in chimaera | 21:28 |
| arno11 | ah | 21:28 |
| arno11 | maybe you have troubles with Qt5 | 21:28 |
| mkf | could be. how can i check? | 21:28 |
| arno11 | i.e are qt apps very slow to launch ? | 21:29 |
| arno11 | like calculator or calendar ? | 21:29 |
| mkf | i'll test. | 21:31 |
| mkf | should i test right after boot or i can wait a while until things are fully loaded | 21:31 |
| mkf | (in other words, do maemo-launcher loads qt5 stuff?) | 21:32 |
| arno11 | no maemo-launcher for qt unfortunately | 21:32 |
| mkf | alright. | 21:33 |
| arno11 | yes you can wait to be sure all things are fully loaded | 21:33 |
| sicelo | i wonder if there's some way to have 'signal' that things have settled :-) | 21:34 |
| arno11 | good question | 21:35 |
| sicelo | then we could simply display a banner on screen ("Booting. Please wait ..."), and when everything settles, we can auto-close this banner. It's ugly, but might help us not end up with N900s thrown against walls | 21:36 |
| arno11 | yes definitely | 21:37 |
| sicelo | the answer probably depends on knowing what is causing that initial 'hang' | 21:38 |
| arno11 | mhm | 21:42 |
| sicelo | Wizzup: any idea how to 'restart' the status applets without reboot? | 21:43 |
| arno11 | you mean hildon-status-menu ? | 21:44 |
| sicelo | yes | 21:45 |
| arno11 | just killall it, it restarts immediately | 21:45 |
| sicelo | cool. i remember now, from when you had some other problem ... was it the IM account Online indication? | 21:46 |
| arno11 | yes :) | 21:46 |
| mkf | btw sicelo, do you have problems with telegram on leste? | 21:46 |
| sicelo | i haven't run telegram in ages. arno11 does though .. | 21:46 |
| mkf | ah | 21:46 |
| mkf | mine barely works. going to settings panel gives a segfault, which i couldn't debug since debian lacks a lot of debugsyms | 21:47 |
| arno11 | telegram is very slow to start | 21:47 |
| arno11 | around 40-60 sec on n900 and hang for 2 min on daedalus and then works fine | 21:47 |
| arno11 | no hang on chimaera but it was an older tg version and no qt5 troubles | 21:48 |
| arno11 | mkf: sorry, didn t see your segfault msg | 21:49 |
| arno11 | yes the segfault is common issue | 21:50 |
| arno11 | you have to do a long press on buttons to avoid segfault | 21:50 |
| mkf | odd. | 21:51 |
| arno11 | yep | 21:51 |
| arno11 | and you need a proper alsoft.conf file in /.config to avoid audio issues | 21:53 |
| mkf | there is also a error of no seat in that tablet during the boot in daedalus | 21:54 |
| arno11 | ah | 21:54 |
| arno11 | battery related | 21:54 |
| sicelo | mkf: consistent? or intermittent? (no seat during boot) | 21:55 |
| mkf | qalendar loads in like 5s | 21:56 |
| mkf | sicelo: consistent | 21:57 |
| mkf | also no matter how many times i charge it, it always needs calibration | 21:57 |
| mkf | it got really better recently, but i think that might be a bug somewhere. (hi james) | 21:58 |
| sicelo | what do you have in /sys/class/power_supply? | 21:58 |
| mkf | axp20x-battery/ axp20x-usb/ axp22x-ac/ | 21:58 |
| mkf | blindly enabled them all. | 21:58 |
| sicelo | oh, i think PP has that one too (or at least same family) | 21:59 |
| sicelo | now, `ls -l /sys/class/power_supply/axp20x-battery/` | 21:59 |
| mkf | https://cloud9.org/paste/fd72 | 22:01 |
| mkf | https://cloud9p.org/paste/fd72 | 22:01 |
| sicelo | you're out of luck :-) | 22:02 |
| mkf | yay :D | 22:03 |
| mkf | what should i be looking for? | 22:03 |
| sicelo | i was hoping `capacity` would be writeable ... your device has same problem as Droid 4, where on boot, calibration is lost. at least on D4 you are able to write the capacity, so there's a service which stores the last calibration when you poweroff, and restores it at boot, before upower starts | 22:03 |
| sicelo | what does `health` contain? | 22:04 |
| mkf | Good | 22:04 |
| mkf | fwiw, this device has very generic batteries | 22:05 |
| sicelo | in fact, show what's in `uevent` | 22:05 |
| mkf | idk if they are smart or dumb | 22:05 |
| mkf | also capacity tells `100` | 22:05 |
| Wizzup | mkf: maybe the dist upgraded added some weight | 22:05 |
| mkf | https://cloud9p.org/paste/148e | 22:06 |
| mkf | wizzup: could be. | 22:06 |
| mkf | wizzup: could you add a repo to git? | 22:08 |
| mkf | for browser tests :) | 22:08 |
| Wizzup | https://git.maemo.org/merlijn/mce | 22:10 |
| Wizzup | mkf: ^ | 22:10 |
| sicelo | mkf: at least it looks like the planned changes to upower, mce, and status-area-applet-battery will help your device too ;-) | 22:12 |
| mkf | \o/ | 22:12 |
| mkf | you guys are wonderful, thanks! | 22:13 |
| sicelo | i'm currently building the last one of them, then will submit final MR | 22:14 |
| mkf | i have so many questions :D | 22:21 |
| mkf | sorry if it's a bit overwhelming. | 22:21 |
| mkf | how can i reduce the dpi? | 22:22 |
| mkf | forgejo looks usable in microb too. cool. | 22:25 |
| sicelo | microb? | 22:25 |
| mkf | from framentle. i still use it in my n900 because notalgia and i don't have a usb port :) | 22:26 |
| sicelo | oh nice :-) | 22:28 |
| sicelo | arno11: Wizzup: oddly, the module blacklist for stuff like twl4030_usb, bq27000 battery, etc. isn't working | 22:31 |
| sicelo | in daedalus, that is | 22:32 |
| sicelo | iirc it worked in chimaera | 22:32 |
| arno11 | ah | 22:36 |
| sicelo | ok i see. i botched it. it never worked :-D | 22:37 |
| sicelo | will submit fix soon | 22:37 |
| arno11 | ok cool | 22:37 |
| arno11 | Wizzup: forgejo seems also usable with links2 gui (no js) | 22:39 |
| sicelo | i suppose dillo too? | 22:40 |
| arno11 | idk, let's try | 22:40 |
| mkf | Wizzup: while you are at it, could you either move blog to maemo.org or at least disable strict https? | 22:41 |
| arno11 | sicelo: yeah, seems to work fine with new dillo | 22:44 |
| mkf | ok this sips ~1200mW roughly | 22:52 |
| mkf | with screen off | 22:52 |
| mkf | quite a lot. | 22:52 |
| arno11 | wow | 22:53 |
| mkf | i'm multiplying voltage reported by battery to current reported by battery, is that correct? | 22:53 |
| arno11 | probably, yeah | 22:54 |
| arno11 | so you see around 300mA idle, right ? | 22:55 |
| arno11 | it is uA in battery stuff | 22:55 |
| mkf | in0: 3.95 V (min = +3.30 V, max = +4.20 V) | 22:56 |
| mkf | curr1: -410.00 mA | 22:56 |
| mkf | from sensors command output | 22:56 |
| arno11 | 400mA is a lot | 22:58 |
| arno11 | how is cpu @idle ? | 22:58 |
| mkf | indeed. | 22:58 |
| mkf | 99.9 | 22:58 |
| arno11 | ah so no surprise | 22:58 |
| mkf | yes surprise? | 22:59 |
| arno11 | i mean, cpu @99.9 and 400mA is normal | 22:59 |
| mkf | 99.9 idle | 23:00 |
| arno11 | but the question is what is overloading your cpu ? | 23:00 |
| arno11 | maybe tracker ? | 23:00 |
| mkf | cpu is @ 99.9% idle. | 23:00 |
| mkf | not doing anything at all :) | 23:01 |
| arno11 | something is wrong | 23:01 |
| arno11 | you should check htop or something to see what processes are involved | 23:02 |
| sicelo | i think the SoC just doesn't do good (or any?) power management | 23:02 |
| arno11 | maybe | 23:03 |
| mkf | possible | 23:04 |
| arno11 | but first you should check if some processes use cpu or not | 23:04 |
| arno11 | i.e telegram after a crash | 23:05 |
| arno11 | it happened to me several times :P | 23:05 |
| mkf | nothing intersting. | 23:06 |
| mkf | it might be a kernel issue. | 23:06 |
| Wizzup | mkf: blog being our github pages? | 23:06 |
| mkf | yes | 23:06 |
| Wizzup | we'll have to figure out what to do with that in time | 23:06 |
| arno11 | mkf: @kernel, yep like cpufreq missing or something ? | 23:08 |
| mkf | well cpufreq dir is empty at /sys if thats what you mean | 23:18 |
| mkf | looked around kernel options, nothing to find. | 23:18 |
| arno11 | no cpufreq-dt ? | 23:24 |
| mkf | is that a module or a /sys file? | 23:26 |
| mkf | ah yeah added those as modules. | 23:27 |
| mkf | interestingly freemangordon have added cpu frequencies in their dts, i wonder why | 23:29 |
| mkf | might be related to this issue | 23:29 |
| mkf | freemangordon: care to repost those patches again? they expired. :( | 23:34 |
| arno11 | mkf: btw cpufreq missing in /sys is an issue we had last year, i wonder if it is similar on your tablet | 23:37 |
| arno11 | what is your kernel, 6.6 or ? | 23:38 |
| mkf | 6.12 | 23:39 |
| mkf | root@sunxi:/sys# ls ./devices/system/cpu/cpufreq | 23:40 |
| mkf | root@sunxi:/sys# | 23:40 |
| Wizzup | does the dt have frews? | 23:42 |
| Wizzup | freqs | 23:42 |
| mkf | my device tree? no. | 23:44 |
| Wizzup | I don't know if this is necessary fwiw | 23:51 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!