| * demoveu -> debacle | 00:02 | |
| + freakazoid0223 (~matt@pool-108-2-68-211.phlapa.fios.verizon.net) | 00:12 | |
| devyl | mntmn doesn't run xorg on the reform? (i saw your twitter comment about xfce) | 00:13 |
|---|---|---|
| * debacle -> demoveu | 00:51 | |
| - erlehmann (QUIT: Ping timeout: 240 seconds) (~erlehmann@dynamic-046-114-036-085.46.114.pool.telefonica.de) | 01:31 | |
| + erlehmann (~erlehmann@dynamic-046-114-033-249.46.114.pool.telefonica.de) | 01:44 | |
| + B[] (~Thunderbi@219-89-228-78-adsl.sparkbb.co.nz) | 01:56 | |
| - Guest37872 (QUIT: Quit: Reconnecting) (~blast007@bzexcess.com) | 02:36 | |
| + blast007 (~blast007@bzflag/developer/Blast) | 02:37 | |
| - oomono (QUIT: Quit: Connection closed for inactivity) (uid328183@gateway/web/irccloud.com/x-zaoqaoengdljwgtj) | 03:04 | |
| - freakazoid0223 (QUIT: Ping timeout: 246 seconds) (~matt@pool-108-2-68-211.phlapa.fios.verizon.net) | 06:41 | |
| + freakazoid0223 (~matt@pool-108-2-68-211.phlapa.fios.verizon.net) | 06:42 | |
| * demoveu -> debacle | 07:52 | |
| * debacle -> demoveu | 07:52 | |
| * demoveu -> debacle | 07:52 | |
| * debacle -> demoveu | 07:52 | |
| + futarisIRCcloud (uid222239@gateway/web/irccloud.com/x-kzrcosrqntwaexum) | 08:53 | |
| + oomono (uid328183@gateway/web/irccloud.com/x-zkriaajwwtbobusp) | 10:53 | |
| * demoveu -> debacle | 11:34 | |
| * debacle -> demoveu | 11:34 | |
| * demoveu -> debacle | 11:34 | |
| - futarisIRCcloud (QUIT: Quit: Connection closed for inactivity) (uid222239@gateway/web/irccloud.com/x-kzrcosrqntwaexum) | 13:02 | |
| + rasmus-mk (~rasmus@c83-253-223-217.bredband.comhem.se) | 13:50 | |
| - freakazoid0223 (QUIT: Ping timeout: 240 seconds) (~matt@pool-108-2-68-211.phlapa.fios.verizon.net) | 14:26 | |
| - oomono (QUIT: Quit: Connection closed for inactivity) (uid328183@gateway/web/irccloud.com/x-zkriaajwwtbobusp) | 14:33 | |
| - B[] (QUIT: Ping timeout: 265 seconds) (~Thunderbi@219-89-228-78-adsl.sparkbb.co.nz) | 15:11 | |
| + freakazoid0223 (~matt@pool-108-2-68-211.phlapa.fios.verizon.net) | 17:08 | |
| - freakazoid0223 (QUIT: Ping timeout: 265 seconds) (~matt@pool-108-2-68-211.phlapa.fios.verizon.net) | 17:50 | |
| - terpri (QUIT: Remote host closed the connection) (~terpri@2603-6081-2241-0752-96b8-6dff-feb3-87d3.res6.spectrum.com) | 18:16 | |
| + terpri (~terpri@2603-6081-2241-0752-96b8-6dff-feb3-87d3.res6.spectrum.com) | 18:16 | |
| * debacle -> demoveu | 18:31 | |
| * demoveu -> debacle | 18:31 | |
| * debacle -> demoveu | 18:31 | |
| mntmn | devyl: no, except if you run it on the framebuffer without acceleration i think | 18:50 |
| mntmn | devyl: last time i tried (long time ago) running Xorg directly on modesetting did not work | 18:51 |
| mntmn | i also don't really understand why one would prefer to run Xorg | 18:51 |
| + oomono (uid328183@gateway/web/irccloud.com/x-qyvydxildvfgyqxb) | 18:54 | |
| swivel | for me, answering the question of "does xorg w/modesetting work?" is the easiest and most likely known form of answer to the real question of "does the mainline driver in use support modesetting?" which applies to far more than just xorg... plenty low-level graphics tools support xorg-less operation w/libdrm and expect modesetting to function | 19:11 |
| mntmn | modesetting works | 19:17 |
| mntmn | for example, kmscube works | 19:17 |
| mntmn | but Xorg just shows a black screen. no errors | 19:18 |
| mntmn | (just tried it) | 19:18 |
| mntmn | i just googled it and found an old email thread by myself https://xorg-devel.x.narkive.com/DoRvbgMq/all-black-x-windows-with-etnaviv-xwayland-glamor | 19:20 |
| mntmn | apparently several years ago i already spent a lot of time trying to debug X issues | 19:20 |
| mntmn | on the other hand, wayland compositors just work, and Xwayland on top. so you can still use any X window manager you want | 19:21 |
| swivel | hrm, i wonder why drmModeGetResources() fails w/ENOTSUPP on my buddies librem5 then | 19:22 |
| devyl | I would like to use e24 and xfce :( so i would need xorg | 19:22 |
| swivel | when i investigated the etnaviv kernel driver to check, it didn't set the DRIVER_MODESET bit: | 19:23 |
| swivel | etnaviv_drv.c: .driver_features = DRIVER_GEM | DRIVER_RENDER, | 19:23 |
| mntmn | devyl: afaik enlightenment is a wayland compositor too | 19:23 |
| mntmn | swivel: you're looking in the wrong place | 19:24 |
| mntmn | swivel: etnaviv is not a display device | 19:24 |
| mntmn | swivel: etnaviv is just a renderer | 19:24 |
| swivel | mntmn: what's the right place for imx8m? | 19:24 |
| mntmn | swivel: either mxsfb or dcss | 19:24 |
| mntmn | swivel: both are kms devices | 19:24 |
| mntmn | swivel: it depends on how the librem is structured, i think they use lcdif for the internal display, hence mxsfb | 19:24 |
| swivel | mntmn: hrm, ok, thanks for the pointers | 19:25 |
| mntmn | swivel: try kmscube -D /dev/dri/card1 for example | 19:25 |
| mntmn | (assuming that exists) | 19:26 |
| mntmn | swivel: the display device creates the framebuffer. etnaviv instructs the gpu to render to it. it can also render headless / to some other memory | 19:27 |
| mntmn | so, to configure your display/mode, you don't talk to etnaviv in the first place, but to one of the two display engines. | 19:27 |
| devyl | If i apt e24 on debiab testing it still installs xorg. Maybe i need to compile it myself on reform once i get it ;) | 19:27 |
| swivel | ah so there may be multiple nodes in /dev/dri/ the experiment probably used card0 | 19:27 |
| mntmn | swivel: yep | 19:28 |
| swivel | good to know | 19:28 |
| mntmn | devyl: anyway, you can use for example sway or cage as your xorg substitute | 19:29 |
| swivel | thanks again for explaining that | 19:29 |
| mntmn | devyl: https://github.com/Hjdskes/cage | 19:29 |
| devyl | Thx | 19:29 |
| mntmn | devyl: this is a compositor to run just one app fullscreen. that can be Xwayland (non-rootless) + xfce4 for example | 19:30 |
| mntmn | bonus: no tearing | 19:30 |
| devyl | Ah ok... Maybe i could just use a compositor instead of the xfce wm to run xfce session and apps (...)!? | 19:32 |
| mntmn | devyl: well, you still want the xfce wm i guess | 19:33 |
| mntmn | i'll put together a quick recipe | 19:33 |
| devyl | 👍 | 19:33 |
| swivel | mntmn: do you happen to know what happens on the imx8 when things get too hot for the SoC? does linux put anything in the journal about it and shut it down via userland awareness, or does it just power off to protect itself independent of the OS? | 19:52 |
| mntmn | devyl: http://dump.mntmn.com/reform-xfce.png | 19:53 |
| mntmn | swivel: it powers itself off | 19:53 |
| mntmn | swivel: but before that, it can also signal the kernel to do that | 19:53 |
| devyl | Looks good :) | 19:54 |
| * demoveu -> debacle | 19:54 | |
| mntmn | devyl: here is how i start it http://dump.mntmn.com/xfce4.sh.txt | 19:55 |
| mntmn | devyl: these exports are only needed because i have wayland specific stuff in my bashrc | 19:55 |
| devyl | Thx! | 19:56 |
| mntmn | works absolutely fine | 19:56 |
| mntmn | will try enlightenment next... | 20:04 |
| swivel | mntmn: hmmm, ok, my buddy tested a drm raytracer on his librem5 to verify /dev/dri/card1 supports modesetting (it does!) and not long after it shutdown, but the journal doesn't mention thermals as a cause, but it also appears to have shutdown the wireless hardware independently so it looks like it overheated. the ray tracer is threaded | 20:05 |
| swivel | mntmn: im curious if the reform would tolerate the sustained cpu burn better, probably | 20:06 |
| mntmn | swivel: how long until it shut down? | 20:16 |
| mntmn | swivel: i don't have any thermal problems so far... but our heatsink is pretty massive | 20:16 |
| mntmn | devyl: enlightenment works fine too with the same method https://files.mastodon.social/media_attachments/files/105/765/177/975/659/441/original/ff2c15aef312b6f4.jpeg | 20:17 |
| swivel | mntmn: something like a couple mins max | 20:19 |
| mntmn | swivel: oof | 20:22 |
| mntmn | swivel: naja looking at the librem5 not sure where all the heat is going | 20:22 |
| swivel | to be fair, it's not exactly unusual for phones to overheat when you saturate all the cores | 20:23 |
| swivel | but it'd be nice for it to be a graceful shutdown with stuff in the journal at least | 20:24 |
| mntmn | swivel: maybe you need to install some daemon for that? i mean, it is easy to watch the temperature from userspace | 20:36 |
| eery | hmm, I guess it'd be possible to run sway with no/minimal info bar, and run Xwayland + WM along with native wayland clients | 20:45 |
| eery | never thought about it tbh | 20:45 |
| + rvense (~rvense@212.97.250.176) | 20:46 | |
| swivel | mntmn: apparently it didn't shut itself down after all, there looks to be a bug with journalctl relative boot ids, a retest on the librem has it running over 10mins without rebooting, false alarm | 20:54 |
| swivel | though it's still unclear why the wireless stuff shutdown and wouldn't come back up without pulling the battery | 20:54 |
| mntmn | i don't have a librem so i don't know how it works in detail | 20:55 |
| mntmn | eery: yes. i have a script that runs me a fullscreen Xwayland+blackbox session in a sway workspace when i need it. | 20:56 |
| mntmn | eery: to run some odd apps that don't work well directly through rootless Xwayland | 20:56 |
| mntmn | for example visicut (a java application for running our lasercutter) | 20:56 |
| mntmn | another cool thing is that sway can do output scaling for you (hardware accelerated) | 20:57 |
| - _bleb (QUIT: Ping timeout: 240 seconds) (cm@faeroes.sdf.org) | 21:01 | |
| - oomono (QUIT: Quit: Connection closed for inactivity) (uid328183@gateway/web/irccloud.com/x-qyvydxildvfgyqxb) | 21:43 | |
| * debacle -> demoveu | 21:53 | |
| * demoveu -> debacle | 21:53 | |
| * debacle -> demoveu | 21:53 | |
| + freakazoid0223 (~matt@pool-108-2-68-211.phlapa.fios.verizon.net) | 21:56 | |
| * demoveu -> debacle | 22:59 | |
| - rasmus-mk (QUIT: Quit: Konversation terminated!) (~rasmus@c83-253-223-217.bredband.comhem.se) | 23:24 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!
