| + erlehmann_ (~erlehmann@x4db71e06.dyn.telefonica.de) | 02:13 | |
| - erlehmann (QUIT: Ping timeout: 250 seconds) (~erlehmann@x590c2982.dyn.telefonica.de) | 02:17 | |
| - erlehmann_ (QUIT: Quit: Leaving) (~erlehmann@x4db71e06.dyn.telefonica.de) | 02:34 | |
| + erlehmann (~erlehmann@x4db71e06.dyn.telefonica.de) | 02:34 | |
| - erlehmann (QUIT: Quit: Leaving) (~erlehmann@x4db71e06.dyn.telefonica.de) | 02:50 | |
| + erlehmann (~erlehmann@x4db71e06.dyn.telefonica.de) | 03:34 | |
| + pretec_ (~pretec@mue-88-130-49-112.dsl.tropolys.de) | 04:24 | |
| - pretec (QUIT: Ping timeout: 250 seconds) (~pretec@mue-88-130-49-173.dsl.tropolys.de) | 04:26 | |
| + dest__ (~pretec@mue-88-130-49-089.dsl.tropolys.de) | 05:07 | |
| - pretec_ (QUIT: Ping timeout: 240 seconds) (~pretec@mue-88-130-49-112.dsl.tropolys.de) | 05:09 | |
| - erlehmann (QUIT: Quit: Leaving) (~erlehmann@x4db71e06.dyn.telefonica.de) | 08:17 | |
| - dest__ (QUIT: Quit: Leaving) (~pretec@mue-88-130-49-089.dsl.tropolys.de) | 10:09 | |
| + pretec (~pretec@mue-88-130-49-089.dsl.tropolys.de) | 10:09 | |
| mntmn | funny article: http://www.naughtycomputer.uk/do_i_really_need_to_get_out_the_soldering_iron_again.html i still need a good speaker amp circuit for the lineout header inside of reform btw | 14:10 |
|---|---|---|
| pretec | haha, really good :) | 20:10 |
| vifino | mntmn: the kmscube works if you're in the video group, once you add tty xorg tries to start up as the user, fails because of tty7 | 20:18 |
| vifino | setting the group read permission bit makes it attempt a little more | 20:18 |
| vifino | but it fails with drmSetMaster after switching buffers (?) | 20:19 |
| mntmn | ah | 20:19 |
| mntmn | vifino, did you try starting X with a vt1 argument or similar? | 20:19 |
| mntmn | like, setting it to your logged in vt maybe | 20:19 |
| vifino | i did not, can do | 20:19 |
| mntmn | also, try: mv /dev/dri/card0 /tmp | 20:20 |
| vifino | ah, you think it's using the wrong card? | 20:20 |
| mntmn | normally the settings in xorg.conf should take care of that | 20:21 |
| mntmn | but maybe X still struggles with something card0 (imx-drm) related when not root | 20:22 |
| mntmn | vifino, for reference, here are the required xorg.conf settings https://mntmn.com/media/tech_notes_md/2018-19-11-using-modesetting-with-etnaviv.html | 20:22 |
| vifino | yeah, the mv didn't do anything, the xorg settings make it use the right device | 20:23 |
| vifino | at this point i'd like to note that xinit fails with tty7 argument, but startx fails with... something? | 20:24 |
| mntmn | yeah, startx cannot work | 20:24 |
| mntmn | try Xorg directly | 20:24 |
| vifino | `Xorg vt1` fails similarly to what it failed like after settting tty7 to have g+r | 20:26 |
| vifino | (EE) modeset(0): drmSetMaster failed: Permission denied | 20:26 |
| mntmn | ah ok | 20:26 |
| mntmn | sounds like it needs to access some other file | 20:27 |
| mntmn | seems to be a common problem | 20:27 |
| vifino | unsure *what* it fails to access, given that dri and such have the correct bits | 20:27 |
| vifino | let me strace it | 20:27 |
| mntmn | yeah strace is a good idea | 20:28 |
| mntmn | maybe systemd/pam is blocking something? | 20:28 |
| vifino | could be, because it fails to call drmSetMaster on /dev/dri/card1 | 20:32 |
| vifino | or rather fd 10, which is /dev/dri/card1 | 20:32 |
| mntmn | vifino, https://stackoverflow.com/a/29711077 | 20:33 |
| vifino | err, ioctl(10, DRM_IOCTL_SET_MASTER, 0) | 20:33 |
| mntmn | looks like that _needs_ root permissions | 20:33 |
| vifino | ugh. | 20:35 |
| mntmn | so you need suid bit on Xorg or systemd-logind | 20:35 |
| mntmn | sad. | 20:35 |
| vifino | that it is. | 20:35 |
| mntmn | pick your poison :D | 20:35 |
| vifino | being one of the people who exploited that xorg font path thing.. haha, just kidding. it's better than systemd. | 20:36 |
| vifino | still disappointing. | 20:37 |
| mntmn | interesting, gentoo has a page about this https://wiki.gentoo.org/wiki/Non_root_Xorg | 20:39 |
| vifino | unfortunately that won't help with drmSetMaster, it seems. | 20:41 |
| mntmn | hmm i have a almost-completed reform here, i'll quickly give it a try | 20:44 |
| mntmn | so i did: chmod 4755 /usr/local/bin/Xorg | 20:47 |
| mntmn | then ran Xorg vt1 as a user | 20:48 |
| mntmn | and i get a black screen | 20:48 |
| mntmn | xinit works | 20:49 |
| mntmn | running openbox as a regular user | 20:49 |
| mntmn | vifino, did you try chmod 4755 /usr/local/bin/Xorg | 20:49 |
| mntmn | ? | 20:49 |
| vifino | yeah, it works as root. | 20:50 |
| vifino | with root, startx works too. | 20:50 |
| vifino | still sad. | 20:50 |
| mntmn | yeah, but that's just the design. we could submit a pull request to change that :D | 20:51 |
| mntmn | i mean it's pretty nonsensical that other access rights is mediated through file permissions but not this | 20:51 |
| vifino | yeah, lets have a trip to the LKMS and change the DRM_IOCTL_SET_MASTER permission structure :P | 20:52 |
| mntmn | :D | 20:53 |
| vifino | not sure if we'd get roasted for that, though. | 20:53 |
| vifino | kinda seems it'd make sense to have that change. | 20:54 |
| vifino | but.. there's probably a reason why it wasn't changed before. | 20:54 |
| mntmn | maybe i'll address it in #xorg-devel at some point | 20:54 |
| vifino | yeah, might be a good idea. | 20:55 |
| vifino | by the way, does the reform kernel have much patches in it or is it just very recent mainline? | 20:57 |
| mntmn | vifino: currently only 1 patch https://source.mntmn.com/MNT/reform-system-image/src/branch/master/mkkernel.sh | 20:59 |
| mntmn | i couldn't find out so far why this isn't upstreamed (pcie suspend patch) | 21:00 |
| vifino | lots of things seem to get lost somewhere.. | 21:00 |
| mntmn | this is the patch https://github.com/mntmn/reform-linux/blob/master/0017-pci-fix-suspend-on-i.MX6.patch | 21:02 |
| mntmn | it's needed when you have a mini pcie card inserted and use suspend/wake. it won't wake up without it | 21:04 |
| vifino | yeah, that sounds lame. but good to know that mainline should boot just fine. | 21:04 |
| mntmn | yep | 21:05 |
| vifino | i guess an archlinux arm port would be relatively easy. just need to build a patched uboot, different config and such. | 21:05 |
| mntmn | (if it has imx6qp etc activated) | 21:05 |
| vifino | alarm does. | 21:06 |
| mntmn | yeah, you can just use my uboot i guess? | 21:06 |
| mntmn | + dts/dtb | 21:06 |
| mntmn | you can just overwrite the debian userland on the sd card with the arch default userland for alarm maybe | 21:07 |
| vifino | nope, the environment needs some changes i think | 21:07 |
| mntmn | ah | 21:07 |
| vifino | different kernel paths and such. | 21:07 |
| mntmn | ah but that's trivial to do | 21:07 |
| vifino | might set up a boot partition, that could be neat. | 21:07 |
| vifino | oh, yeah, definitly. | 21:07 |
| vifino | i'll probably do it today. | 21:07 |
| vifino | just need to find more sdcards. | 21:08 |
| mntmn | you just need to edit https://github.com/mntmn/u-boot/blob/mntreform/reformenv.txt probably | 21:08 |
| vifino | yeah, saw that. | 21:08 |
| mntmn | i recommend to get an msata ssd, then you can use the sd as your rescue system | 21:09 |
| vifino | what we could do is unify that, so that the uboot image opens, say, the first partition with ext4, loads the small mnt image and then looks for a boot.scr in either / or /boot, for example | 21:10 |
| vifino | that way you can change the kernel parameters without rebuilding uboot | 21:10 |
| mntmn | sounds good | 21:10 |
| mntmn | is it possible to have them in plaintext or it always needs to be binary stuff? | 21:11 |
| vifino | good question, i am not sure | 21:11 |
| vifino | i've only seen the .scr's used | 21:11 |
| mntmn | just remember vaguely that on the raspi you have a config txt file | 21:11 |
| vifino | yeah, but that's not uboot :P | 21:11 |
| mntmn | ah sure :D | 21:11 |
| vifino | that's the vc4 doing.. whatever it is it does. | 21:12 |
| vifino | the cool thing about the looking-for-boot.scr is that installing, say, alarm, is something very trivial | 21:12 |
| vifino | make partition, install rootfs, drop in boot.scr, dd uboot over, done | 21:13 |
| mntmn | ok... but the reform dtb would need to be upstreamed first i guess? or can you drop in a custom one? | 21:19 |
| vifino | well, could just be an archive that needs to be extracted on the boot partition, but eventually, that could be put in an alarm repo to keep things up to date | 21:20 |
| vifino | having a seperate boot partition is also handy. could boot non-linuxy stuff! | 21:21 |
| mntmn | true true | 21:22 |
| mntmn | i had this kind of setup before, not sure why i changed it tbh... i guess i wanted the imaging process to be easier | 21:22 |
| mntmn | if you figure something out, i'll accept PRs ;) | 21:23 |
| vifino | :P | 21:24 |
| vifino | will gladly do so. | 21:25 |
| vifino | if i get too annoyed with drmSetMaster/suid, expect a patch flying your way as well ^^ | 21:25 |
| mntmn | haha | 21:26 |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!