| sicelo | yay! progress ... now x server is crashing. but at least that means we got somewhere | 00:37 |
|---|---|---|
| Wizzup | cool :) | 00:40 |
| sicelo | https://paste.debian.net/1311092/ | 00:44 |
| sicelo | i think i must sleep now and see this next time :-) | 00:44 |
| Wizzup | seems like you might need some xorg conf | 00:47 |
| Wizzup | either that or our mesa lacks the driver | 00:47 |
| Wizzup | but I don't think it does | 00:47 |
| sicelo | we might be missing it. i was just doing a find, and this is what i can find: | 00:54 |
| sicelo | sudo find . -name "*etnaviv*" | 00:54 |
| sicelo | [sudo] password for mobian: | 00:54 |
| sicelo | ./usr/lib/aarch64-linux-gnu/dri/etnaviv_dri.so | 00:54 |
| sicelo | ./usr/src/linux-headers-5.10.0-28-common/include/uapi/drm/etnaviv_drm.h | 00:54 |
| sicelo | ./usr/src/linux-headers-6.1.15/include/uapi/drm/etnaviv_drm.h | 00:54 |
| sicelo | altough i think i saw it in the sources. maybe we don't have config option to actually build it | 00:54 |
| sicelo | i.e. mesa | 00:54 |
| sicelo | ah, the first one is from mesa, so we have it | 00:58 |
| sicelo | added a xorg.conf from postmarketOS. result is https://paste.debian.net/1311099/ | 01:09 |
| sicelo | here's the conf https://gitlab.com/postmarketOS/pmaports/-/blob/master/device/main/device-purism-librem5/xorg.conf?ref_type=heads | 01:09 |
| Wizzup | sicelo: is this X being started by maemo init script? | 03:20 |
| Wizzup | sicelo: it looks like it is being started as user, right? | 03:21 |
| Wizzup | (which might be fine, but the permission denied is a bit suspicious) | 03:21 |
| Wizzup | what are the privs on /dev/dri/card* | 03:21 |
| freemangordon | Wizzup: sicelo: what about the idea to do 'color coding' to match signal strength vs operator name in status bar, when there is more than one modem? | 08:05 |
| freemangordon | like, bars/name of modem 1 are green, bars/name of modem 2 are blue (ignore the particular colors)? | 08:06 |
| freemangordon | what colors to use will be the next question if you like the idea | 08:07 |
| gnarface | is that a problem you run into a lot? phones with two modems? | 08:07 |
| sicelo | we might end up with a rainbow. at first glance, doesn't seem to be good option | 08:07 |
| freemangordon | gnarface: no | 08:07 |
| freemangordon | sicelo: I plan support for 2 modems only ni status menu | 08:07 |
| freemangordon | that would not make a rainbow :) | 08:07 |
| sicelo | also, what if the person has a theme that changes colors? | 08:08 |
| freemangordon | again, ignore the particular colors | 08:08 |
| sicelo | ah, *maybe* if we're talking about only two then | 08:08 |
| freemangordon | we already agreed there will be no more than 2 oiperator names in status bar | 08:08 |
| sicelo | but with two, why not just use position ... left icon is related to left name, etc. | 08:08 |
| freemangordon | because I plan to use 2 rows | 08:08 |
| freemangordon | not append oerator names | 08:09 |
| freemangordon | hmm... | 08:09 |
| sicelo | freemangordon: any idea/suggestion about https://paste.debian.net/1311099/ ? | 08:09 |
| freemangordon | user is not mamber of video group? | 08:09 |
| freemangordon | *member | 08:09 |
| sicelo | i think the same could work ... top name is for leftmost signal bars | 08:09 |
| freemangordon | sicelo: but how do you know that by the look of it? | 08:10 |
| freemangordon | I can use primary and secondary text colors for 1st and 2nd modem | 08:10 |
| freemangordon | but ok | 08:11 |
| sicelo | i actually don't have an issue for the colors ... but we need to be sure it doesn't break due to themes (i don't use them, so i'm not even sure what they do) | 08:14 |
| inky | on jolla btrfs had huge problems. people say those are solved now and don't happen with modern kernels. but on jolla btrfs didn't report that the space is exhausted, but it was, and nothing was working, and the solution was to balance the trees. | 12:00 |
| inky | i am afraid of btrfs. maybe i just don't know how to use it. but i don't know how to use xfs, and xfs never betrayed me or put in bad situation. | 12:00 |
| inky | i am also afraid btrfs will be slow. i am ignorant in this though. | 12:01 |
| inky | so mz617 can be recommended? it won't be very unstable? | 12:03 |
| uvos__ | mz617 isent unstable, its not in anyway really ready yet either | 12:15 |
| inky | okay! | 12:16 |
| uvos__ | btfs is mostly fine, some edgecases aisde | 12:16 |
| uvos__ | we need it on mz617 to span volumes, not really mutch of a choice there since we cant change the partition table | 12:17 |
| uvos__ | i gues lvm could be used instead of btrfs | 12:18 |
| uvos__ | but im not sure about the pros and cons of that | 12:18 |
| inky | yes lvm is much much more usable, i love it. | 12:18 |
| inky | also i think designwise lvm is a right tool in unix way. | 12:18 |
| inky | it does what it does and all the file systems can use it. while btrfs/zfs can do same things but only with their volumes. | 12:18 |
| buZz | lvm2 i hope ;) | 12:30 |
| sicelo | 09:09 < freemangordon> user is not mamber of video group? ... | 13:42 |
| sicelo | i've chrooted into the sdcard from mobian: | 13:43 |
| sicelo | root@mobian:/# id user | 13:43 |
| sicelo | uid=1000(user) gid=1000(user) groups=1000(user),27(sudo),29(audio),44(video),102(netdev),111(pulse-access) | 13:43 |
| sicelo | i *think* this is the true state of the SD card installation, because user 'user' does not exist in mobian | 13:43 |
| uvos__ | sicelo: how did you mount /dev /sys /proc inside the chroot? | 13:47 |
| uvos__ | sicelo: host udev propubly sets different permissions on those than the chroot expects | 13:47 |
| sicelo | sudo mount -t proc /proc proc/ ... same for /sys. and finally sudo mount --bind /dev dev | 13:51 |
| sicelo | anyway | 13:52 |
| sicelo | i chrooted so i could check if `user` is in the correct group. this is a user who doesn't exist in mobian ... so i think it's correct to conclude that the result is from Leste | 13:52 |
| sicelo | for context, i'm chrooting into a Leste SD card, from mobian | 13:53 |
| uvos__ | the problem here is that it may be in the right grou | 14:00 |
| uvos__ | p | 14:00 |
| uvos__ | but the right group only grants permissions if the /sys /proc or /dev files relateing to the feature have the right permissions set | 14:01 |
| uvos__ | which is done by udev | 14:01 |
| uvos__ | which in turn is a place where host stuff leaks into the chroot | 14:01 |
| uvos__ | since host udev is involved here | 14:01 |
| uvos__ | and if its configure differently it wont set the permissions the way the chroot expects | 14:01 |
| sicelo | i think i should find a way to make USB network start much earlier. will appreciate suggestions, but can look at it later in the day | 14:13 |
| uvos__ | what do you need early usb network for? the kernel can write the console to a udp endpoint | 14:14 |
| uvos__ | this is the erliest you can get any output if you cant use serial | 14:15 |
| uvos__ | i think you can make this work via wifi | 14:15 |
| uvos__ | another fun idea for a device without any accessable serial output would be to use a usb otg -> serial or usb otg -> ethernet adapter | 14:16 |
| uvos__ | and use that | 14:16 |
| sicelo | wifi is harder than usbnet, isn't it ... | 14:17 |
| uvos__ | sure im just saying this can be done before userspace even loads | 14:19 |
| uvos__ | maybe not via wifi im not sure about that | 14:19 |
| uvos__ | but in generall | 14:19 |
| sicelo | userspace solution is fine. the problem here is X11 failing to start, and i don't have a way to interact with the device afterwards | 14:28 |
| uvos__ | on d4/xt875 i have just pluged in a usb keyboard and switched to a different vt via ctrl-alt-fX to look around in this case | 14:34 |
| uvos__ | that works pretty well | 14:34 |
| uvos__ | dose the device have working otg/ | 14:34 |
| uvos__ | ? | 14:34 |
| sicelo | i don't know how it works ... anyway, i don't have usb-c cable either. | 14:43 |
| Wizzup | freemangordon: fine by me @ colours | 16:00 |
| Wizzup | inky: btrfs space issues are long solved for the most parts, like years I think | 16:00 |
| Wizzup | but yeah it's not a requiremen for mz617 and the images I am building won't use it | 16:00 |
| Wizzup | sicelo: check /dev/dri privs | 16:01 |
| Wizzup | maybe the entry isn't root:video but something else | 16:01 |
| Wizzup | serial would be helpful indeed | 16:02 |
| sicelo | it appears to be root:video ... but of course i can't see it from the card when OS is not running :-) | 16:11 |
| Wizzup | yeah sounds like it's time to set up some remote access | 16:16 |
| Wizzup | you could try to set up a network for icd2 wiht gconf, assuming the interface is called wlan0 (there's a kernel command line param for this but I think we set up udev correctly so it does not rename) | 16:17 |
| Wizzup | sicelo: seems like you should get a usb c cable for otg as well | 16:17 |
| sicelo | i'll try to get the otg cable/adapter soon | 22:36 |
| sicelo | mmm, maybe it's because pinephone image doesn't have libdrm-etnaviv1 :-p | 22:55 |
| sicelo | let me chroot, install, and see | 22:56 |
| Wizzup | sicelo: yes that could be problem | 23:00 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!