| donihalim | Wondering what else should/can i do on the n900... | 06:39 |
|---|---|---|
| sicelo | freemangordon: Wizzup: upower issue created, https://github.com/maemo-leste/bugtracker/issues/766 | 12:01 |
| arno11 | donihalim: depends of what you mean: testing/optimizing new functionalities or fixing existing issues :P | 12:14 |
| arno11 | (or porting new stuff ?) | 12:14 |
| arno11 | i mean, apart modern web, almost everything else already works but needs most of the time optimization | 12:20 |
| arno11 | (ah i forgot camera and bluetooth, which are the last not working things) | 12:21 |
| sicelo | with the omap3isp patch, camera at least probes correctly now, and shows the required v4l2 endpoints. it should be possible to get some images out of it | 12:26 |
| arno11 | really cool | 12:26 |
| sicelo | with the bluetooth, i would like to investigate how to provide a bluetooth address to the driver. from discussion with sre, there's a timing issue somewhere (firmware perhaps?) which will be very hard to troubleshoot | 12:36 |
| sicelo | anyway you never know ... maybe if we provide the address early enough, it might just make all the difference in the world | 12:36 |
| sicelo | with BT working (even if not perfectly), we would then be able to bring up the FM radio receiver :-) | 12:44 |
| arno11 | :) | 12:46 |
| inky | > with the omap3isp patch, camera at least probes correctly now, and shows the required v4l2 endpoints. it should be possible to get some images out of it | 12:54 |
| inky | omg | 12:54 |
| Wizzup | my droid strangely doesn't want to connect to the network some of the time | 12:55 |
| arno11 | i also got some troubles with chimaera: disconnections, wlan1 issue. not anymore with fresh daedalus | 12:57 |
| Wizzup | sorry, I meant the mobile network | 12:58 |
| arno11 | ah ok | 12:58 |
| arno11 | is it with ofono1 or 2 ? | 12:58 |
| sicelo | arno11: regarding wlan1 ... someday i'll submit patch to make that problem history. we would no longer require the interface to be named wlan0 | 12:59 |
| Wizzup | ofono 2 | 13:00 |
| arno11 | nice | 13:00 |
| Wizzup | sicelo: yeah that does need fixing | 13:00 |
| sicelo | i've looked at it a bit, and can see how we could implement it. -ETIME is the only major problem atm :-) | 13:01 |
| sicelo | on the wlan topic, anyone has an opinion on iwd? | 13:05 |
| Wizzup | as opposed to wpa_supplicant? | 13:14 |
| sicelo | yes | 13:14 |
| Wizzup | I haven't tried it, but I'd be worried about compat with older devices and how we'd do host mode | 13:15 |
| Wizzup | and of course having to rewrite a bunch of C code :P | 13:16 |
| sicelo | it supports hotspot. and works fine on N900, which i guess is oldest device we will ever support | 13:16 |
| Wizzup | what makes it more suited than wpa_supplicant? | 13:17 |
| sicelo | I don't really know 😆 | 13:18 |
| Wizzup | imo, if we wanted to swap out some things I'd want to replace PA with pipewire at some point | 13:19 |
| sicelo | it is said to be mobile optimized (comes from ofono people) | 13:19 |
| Wizzup | if it can do host mode in the same daemon that could be somewhat useful | 13:19 |
| sicelo | yes, hotspot is builtin | 13:21 |
| Wizzup | imo I'd prefer to defer just because it's more work | 13:27 |
| sicelo | yeah, completely agree. it's a project for some future day :-) | 13:28 |
| sicelo | never ending list | 13:28 |
| Wizzup | I hope that at some point it won't be never ending :) | 13:29 |
| Wizzup | but for now, YES :D | 13:29 |
| Wizzup | if we were to set up gitea on say git.maemo.org or src.maemo.org (preferences?), should we allow openid/oauth logins from github and the like? | 13:29 |
| sicelo | sauce.maemo.org :-) | 13:35 |
| sicelo | jokes aside, I'd probably go for git.* | 13:36 |
| arno11 | hmm got again that random issue when wifi burns +30mA...disappearing if i disconnect/reconnect | 13:50 |
| sicelo | how do you correlate that to wifi? | 13:51 |
| arno11 | well, idle is around 45-47mA idle with modem on (3.5g). with wifi connected: 52-58mA | 13:53 |
| arno11 | but sometimes 85mA with wifi | 13:53 |
| arno11 | until i reconnect it | 13:54 |
| arno11 | (or until i disconnect it and idle is back to normal 45-47mA) | 13:55 |
| sicelo | ok. maybe wpa_supplicant? we're still using same kernel from chimaera, right? | 13:55 |
| arno11 | yes same kernel | 13:55 |
| sicelo | maybe htop or dmesg could give you clues? | 13:55 |
| arno11 | 6.6.53 | 13:56 |
| donihalim | arno11: how you measure the power consumtion? | 13:56 |
| arno11 | sicelo: no htop or dmesg unusual stuff | 13:57 |
| arno11 | donihalim: have a look in /sys/class/power_supply/bq27200-0/ | 13:57 |
| arno11 | very useful | 13:57 |
| arno11 | i.e to get quite accurate idle current draw: sleep 30 && cat /sys/class/power_supply/bq27200-0/current_now | 13:58 |
| arno11 | you can even check battery health and temperature | 14:01 |
| donihalim | Cool, thanks for the tips | 14:07 |
| arno11 | np | 14:07 |
| mkf | hello | 14:57 |
| mkf | can the latest images be flashed on n900 or they are borken? | 14:58 |
| arno11 | mkf: hi | 14:58 |
| Wizzup | mkf: hi | 14:59 |
| mkf | mkf: hi | 14:59 |
| arno11 | latest daedalus works fine but you need to edit fstab to switch to sdswap | 14:59 |
| mkf | it uses mmc? | 15:00 |
| Wizzup | arno11: let me see if I can figure out the rootfs-overlay | 15:01 |
| Wizzup | I think we can modify daedalus.blend, where it rsyncs the files, and see if we can have it be more clever | 15:01 |
| arno11 | mkf: yes but just because of building issue | 15:01 |
| arno11 | Wizzup: ok | 15:01 |
| mkf | ok, thanks. | 15:02 |
| mkf | https://maedevu.maemo.org/images/n900/20250316-chimaera/ <- is this what i should download? | 15:04 |
| arno11 | mkf: quite the same issue with last chimaera img: if you modify few files before first boot, it should be ok | 15:04 |
| arno11 | but not OOTB | 15:04 |
| dsc_ | mkf: hi | 15:05 |
| mkf | hi: dsc_ | 15:08 |
| mkf | arno11: any noticable improvment on chimaera vs daedalus? | 15:08 |
| mkf | i.e: better battery or better performance in n900 | 15:08 |
| arno11 | things are quite similar | 15:10 |
| mkf | okay, how should i get daedalus image? | 15:10 |
| donihalim | copying text from surf2 browser now works using https://github.com/cdown/clipmenu | 15:12 |
| Wizzup | mkf: let me see if I can kick off a new build | 15:12 |
| arno11 | mkf: https://maedevu.maemo.org/images/n900/20250209-daedalus | 15:12 |
| arno11 | but better to wait for new build indeed | 15:13 |
| arno11 | donihalim: cool | 15:13 |
| Wizzup | https://phoenix.maemo.org/view/Images/job/leste-image-n900-daedalus/19/console looks like the latest build failed in removing the swap file from the build host lol | 15:14 |
| Wizzup | I'll try to fix today & fix fstab | 15:14 |
| arno11 | Wizzup: ok :) | 15:15 |
| kiva | how newcomers can know what imege should dl with any device? How many had dl some image that does not work and gave up and newer check leste again? | 15:15 |
| donihalim | another n900 user, cool :) | 15:16 |
| kiva | should separate stable and dev images? | 15:17 |
| Wizzup | I think the problem is doing swapon in the postinstall | 15:17 |
| Wizzup | maybe we should not do that | 15:17 |
| Wizzup | arno11 ^ | 15:17 |
| Wizzup | in leste config postinst | 15:17 |
| Wizzup | I think mkswap and fstab writing makes sense, but swapon does it | 15:17 |
| Wizzup | does not* | 15:17 |
| Wizzup | agree? | 15:17 |
| sicelo | yes, i agree | 15:17 |
| Wizzup | will fix | 15:18 |
| arno11 | Wizzup: let me few min to check something | 15:19 |
| arno11 | yes you're right | 15:20 |
| arno11 | that's probably the problem | 15:20 |
| arno11 | 'sw' option in fstab makes the swapon iirc | 15:21 |
| Wizzup | no, we called swapon explicitly | 15:21 |
| arno11 | yes i know | 15:21 |
| sicelo | what is that option? i can't find it in `man 5 fstab` | 15:22 |
| Wizzup | https://unix.stackexchange.com/questions/365953/what-sw-means-in-the-fstab-swap-entry-for-mount-options-column | 15:22 |
| mkf | openbsd's fstab man page does mention sw | 15:22 |
| arno11 | well, we don't need 'swapon /swap' but we need sw option | 15:23 |
| mkf | (alongside of rw, rq, ro and xx options) | 15:23 |
| sicelo | on my N900 i have `/swap none swap defaults 0 0` | 15:23 |
| arno11 | otherwise, there is no swap on boot at all iirc | 15:23 |
| Wizzup | I think we have it don't we? | 15:23 |
| arno11 | yes that's what i said | 15:23 |
| arno11 | sicelo: weird, iirc it didn't work before (without sw) | 15:24 |
| sicelo | devuan fstab also doesn't show that 'sw' :-) | 15:25 |
| arno11 | on my ubuntu fstab, it is | 15:25 |
| arno11 | anyway, the problem seems swapon /swap | 15:26 |
| Wizzup | back in a bit | 15:26 |
| Wizzup | yeah, the image-builder still had it enabled as swap :D | 15:26 |
| Wizzup | that's why the rm failed | 15:26 |
| arno11 | sicelo: oh now i remember how it works with sw | 15:30 |
| arno11 | if you create the swapfile yourself, you don't need the sw option. at the opposit, if the swap file is created during the postinstall, you need it | 15:32 |
| arno11 | otherwise, no swap at all | 15:32 |
| arno11 | *opposite | 15:32 |
| donihalim | /swap is created when dist-upgrade to devel | 15:34 |
| arno11 | Wizzup: ^^^ we had this issue when we switched to emmc swap by default (see irc logs) | 15:34 |
| arno11 | donihalim: yes with chimaera, that's normal, emmc swap is still the default one without devel | 15:35 |
| arno11 | donihalim: on daedalus, default is sdswap (like chimaera -devel) because emmc is buggy with kernel 6.6 | 15:39 |
| Wizzup | Yes, the swap file is created by postinst | 15:47 |
| Wizzup | that is normal | 15:47 |
| arno11 | should work fine now (hopefully) | 16:14 |
| Wizzup | I still need to fix the fstab, but the postinst should be ok | 16:14 |
| arno11 | ah yes @fstab | 16:15 |
| Wizzup | the problem is that our daedalus.blend just overrides the fstab regardless | 16:18 |
| Wizzup | Maybe we have it append instead? | 16:18 |
| Wizzup | hm... actually... there's *two* places where we override fstab, lol | 16:19 |
| arno11 | lol | 16:19 |
| Wizzup | arm-sdk writes it, extra/n900/etc/fstab | 16:19 |
| Wizzup | and daedalus.blend writes it | 16:19 |
| Wizzup | cat <<EOF | sudo tee "$strapdir/etc/fstab" >/dev/null | 16:20 |
| Wizzup | # fstab for arm system | 16:20 |
| Wizzup | none /sys/kernel/debug debugfs defaults 0 0 | 16:20 |
| Wizzup | /dev/mmcblk0p1 /boot ${bootfs} defaults 0 1 | 16:20 |
| Wizzup | EOF | 16:20 |
| arno11 | what a mess | 16:20 |
| Wizzup | yup | 16:20 |
| Wizzup | so I can change arm-sdk not to write the emmc entry, which I think is what is still present, right? | 16:21 |
| arno11 | yes agree | 16:22 |
| sicelo | the workings of the builder seem to be a mystery to me :( | 16:23 |
| sicelo | and i need to understand it in order to add the L5 .. | 16:23 |
| Wizzup | it's actually not that hard, it is just a bit tricky because arm-sdk and vm-sdk aim to work to build images for devuan, regardless of maemo | 16:26 |
| Wizzup | and then 'maemo' is added as a 'blend' | 16:26 |
| Wizzup | so devuan has its own idea of what etc/fstab should look like | 16:26 |
| Wizzup | and then so does maemo with its rootfs overlay, and then our .blend file (mistakenly, imo) also overrides that | 16:26 |
| Wizzup | probably not more than 10-15 minutes of dedicated concentation to fix | 16:27 |
| sicelo | i'll give it a closer look again | 16:43 |
| arno11 | for an hobbyist like me, yeah indeed builder is not so complicated: it is pleasant to troubleshoot compared to deal with valgrind and memory leaks lol | 17:02 |
| inky | my bionic (which is stable) says it got an update for calendar and for some reason from ovi store. Is it expected? | 17:33 |
| inky | (: | 17:33 |
| inky | but the description is 'an open source clone of the default maemo calendar application' | 17:34 |
| dsc_ | yes calendar was updated recently | 17:35 |
| inky | good, thank you | 17:35 |
| dsc_ | they added new days and months | 17:36 |
| inky | the 'ovi store' made me wonder where those words come from. | 17:36 |
| dsc_ | I'm not sure about 'ovi store' | 17:36 |
| dsc_ | where do you read it? | 17:36 |
| dsc_ | what is it? | 17:36 |
| inky | alas i didnt screenshot | 17:36 |
| dsc_ | ah nokia | 17:37 |
| inky | when i tapped on the update indicator on the status bar it opened ham, i think, which mentioned at the top 'ovi store' and below 'calendar'. | 17:37 |
| dsc_ | but yes it was updated | 17:38 |
| inky | good to know, thank you (: | 17:40 |
| dsc_ | yw | 17:41 |
| Wizzup | inky: 'ovi store' is just some translation string I think | 18:16 |
| Wizzup | but yeah it's funny, I think it's due to the domain changes by freemangordon | 18:16 |
| mkf | hm, latest image doesn't work | 19:50 |
| mkf | https://maedevu.maemo.org/images/n900/20250209-daedalus/ | 19:50 |
| mkf | tries to run x11, dies somewhere before i can see a thing | 19:50 |
| mkf | i can only see tty. | 19:50 |
| sicelo | tty allows you to type? | 19:54 |
| mkf | good question. | 19:56 |
| mkf | it allows me to type | 19:58 |
| mkf | backend 'seatd' failde to open seat, skipping | 19:59 |
| mkf | ... <snip> ... | 20:00 |
| mkf | (EE) .......logind.c:137 Could not take device: no such device | 20:00 |
| sicelo | charge the device first :-) | 20:01 |
| mkf | um. how? | 20:02 |
| mkf | (maybe i should buy a seat before using modern linux, they all want seats :D) | 20:02 |
| sicelo | i mean, charge the battery :-) | 20:13 |
| mkf | oh, i read it 'change' | 20:14 |
| mkf | thanks. | 20:14 |
| sicelo | 5-10 minutes should be enough to get it to boot all the way | 20:15 |
| Wizzup | 19:50 < mkf> https://maedevu.maemo.org/images/n900/20250209-daedalus/ | 20:21 |
| Wizzup | 19:50 < mkf> tries to run x11, dies somewhere before i can see a thing | 20:21 |
| Wizzup | 19:50 < mkf> i can only see tty. | 20:21 |
| Wizzup | I think this is upower | 20:21 |
| sicelo | yes, he needs to charge, then upower won't trouble him | 20:22 |
| mkf | hm, i dont have a charger, i can charge batteries | 20:30 |
| mkf | (using a charge module) | 20:31 |
| mkf | i wonder if that casues the phone to not calibrate and think my phone is ran out of battery | 20:31 |
| mkf | i've tried two batteries right now. | 20:31 |
| mkf | i recall when i used leste before it thought my battery had 0 out of 2560(?) mW, which wasn't correct (and i had a regular BL-5J), so i wonder if that's because my battery isn't calibarated in leste (i wonder how can i do so without a charger port) | 20:34 |
| sicelo | mkf: that won't help you (charging externally) ... the problem here is that the fuel gauge, which is inside the N900, is currently reporting 0% *even if* your battery has charge | 20:38 |
| sicelo | pre-daedalus, we had patched upower so it would ignore that 0% and instead make some estimates. we want to drop that, https://github.com/maemo-leste/bugtracker/issues/766 | 20:40 |
| mkf | given there are so many n900s with broken charger ports, is there a way for me to disable that check? | 20:43 |
| mkf | i have two n900s and both of which have broken charger ports. :) | 20:43 |
| sicelo | mkf: that issue details a plan we are working on to disable that check, yes | 20:47 |
| sicelo | in the meantime, you can chmod -x the upower binary, e.g. by mounting the card on a PC | 20:47 |
| mkf | very well. thank you. | 20:48 |
| arno11 | mkf: if you mount the card on pc, you should modify fstab to switch to sdcard swap | 21:00 |
| arno11 | (or you can do the changes from fremantle if it boots btw) | 21:01 |
| mkf | i dont see any swap in fstab | 21:09 |
| mkf | only /boot/efi (?) and / | 21:10 |
| sicelo | just add a new line with `/swap none swap defaults 0 0` | 21:13 |
| mkf | okay. | 21:15 |
| mkf | thank you. | 21:15 |
| arno11 | mkf: /boot/efi ? wtf | 21:16 |
| arno11 | from n900 daedalus img ? | 21:17 |
| mkf | yes. | 21:18 |
| mkf | okay, booting with upower without execution permission still tells me no seat. | 21:18 |
| arno11 | something is wrong you should see at least /boot and emmc swap (/dev/mmcblk1p3) | 21:18 |
| mkf | i dd'd this https://maedevu.maemo.org/images/n900/20250209-daedalus/maemo-leste-5.0-armhf-n900-20250209.img.xz | 21:19 |
| arno11 | it is the correct one | 21:19 |
| arno11 | but /boot/efi is your pc's fstab, no ? | 21:20 |
| mkf | my pc uses bios tho. | 21:20 |
| mkf | might be thumbstick, let me recheck. | 21:20 |
| arno11 | ok | 21:20 |
| arno11 | you should see /dev/mmcblk0p1 /boot ext2 | 21:21 |
| mkf | ah. i see them. | 21:21 |
| mkf | my bad, likely some thumb drive | 21:21 |
| arno11 | ok | 21:22 |
| mkf | enabled swap | 21:23 |
| mkf | i wonder why it still refuses to run x11 | 21:23 |
| sicelo | still fails the same way? | 21:24 |
| mkf | yes. | 21:24 |
| sicelo | sorry i wasn't too clear | 21:25 |
| sicelo | chmod -x the upowerd binary, not upower :-) | 21:25 |
| mkf | ah. no worries, thanks. | 21:25 |
| mkf | x11 is up, thank you arno11, wizzup and sicelo | 21:30 |
| mkf | leste works reasonablely fast | 21:46 |
| mkf | cool. | 21:47 |
| arno11 | mkf: you can even improve things with min freq @500MHz and sudo renice -n -9 $(pgrep Xorg) | 21:48 |
| arno11 | same for systemui, hildon-desktop and mce | 21:48 |
| arno11 | on each boot | 21:48 |
| arno11 | and boost mode ofc | 21:49 |
| arno11 | and low radius in transitions.ini | 21:50 |
| arno11 | (see wiki) | 21:50 |
| arno11 | the only 'major' issue we have with daedalus on n900 is slow Qt5 launch. there is a workaround but if you don't use Qt5 too much, no need | 21:54 |
| dsc_ | i found about this upower issue today with my d4 :)) | 22:13 |
| sicelo | yeah, d4 is affected too | 22:27 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!