Session restart when viewing a specific PDF
- Inicie sesión o regístrese para enviar comentarios
This is probably a different facet of these two:
https://trisquel.info/es/forum/x11-close-session-when-opening-file-or-browsing-web
https://trisquel.info/en/forum/frequent-crashes-trisquel-11
I have a particular PDF, which is attached. Opening it (in Atril, as is the default) restarts the session.
I have a gdb debug script, for use on a virtual console, and the output, also attached.
$ uname -a
Linux user-ThinkPad-X200 5.15.0-76-generic #83+11.0trisquel24 SMP Thu Jun 29 05:37:25 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux
Librebooted. Current version of Trisquel, modulo abrowser, which I think that I am not running.
I am open to adding packages or modifying the debug script, to collect more information.
Adjunto | Tamaño |
---|---|
religions-13-00275-replika-as-spiritual-enhancement.pdf | 255.44 KB |
xdbg.sh | 663 bytes |
gdb_log.709.txt | 4.06 KB |
Crediting the core of my debug script:
https://www.x.org/wiki/Development/Documentation/ServerDebugging/#OneMachine
This work-around suppresses my symptoms:
https://trisquel.info/es/forum/x11-close-session-when-opening-file-or-browsing-web#comment-171662
which I am summarising as: create the file between the ruled lines, with the path in the top ruled line:
============== /etc/X11/xorg.conf.d/20-intel.conf =====================
Section "Device
Identifier "Intel Graphics"
Driver "Intel
Option "TearFree" "true
EndSection
======================================
I used a terminal, sudo (because the directory is owned by root and the GUI runs as user) and cat, but anything that creates the same file should have the same effect.
Trying to understand the workaround got me to here:
https://manpages.debian.org/bookworm/xserver-xorg-video-intel/intel.4.en.html
which got me to these bug reporting instructions:
https://www.intel.com/content/www/us/en/docs/graphics-for-linux/developer-reference/1-0/report-bugs.html
but really I'm defeated unless someone wants to do a lot of (probably not very productive) hand-holding.
>The solution is to create a file called 20-intel.conf and upload it to the /etc/X11/xorg.conf.d folder with the following lines:
I need to read the thread a bit better next time X)
Thanks I'll try and I'll update.
It works no. Creating the file and adding the lines solved everything (except for noclip/WebGL 2). I just had to crash/restart xorg once.
Again thank you :)
Coincidently I'm having a similar issue.
I had prepared yesterday text and documents but it seems to me that most won't be necessary.
Not only I'm having Xorg crashes and the session restarting when I open pdfs but also when I open certain .mp4 containers be it with vlc/mpv, but also on specific webpages like on danbooru, example with https://danbooru.donmai.us/posts/1317198 and select download on the bottom left. But also twitter and other heavy JS websites.
The issue is less common the moment I disable the hardware decoding in vlc, but still happens. And also when I deactivate hardware acceleration in FF nightly.
I'll also note that this website https://noclip.website/ cannot load anymore because of no WebGL 2 support which is very strange because it worked in past versions of trisquel...I think, I just migrated on this T510 that I just got back from someone that was in urgent need of a laptop for a year and the one I was using an asus a52J is having tons of i/o issues related to a few chips that needs replacement.
Surprisingly I don't have a intel.conf even tho I'm also using intel graphics.
============== /etc/X11/xorg.conf.d/ =====================
10-radeon.conf
======================================
00:02.0 VGA compatible controller [0300]: Intel Corporation Core Processor Integrated Graphics Controller [8086:0046] (rev 02) (prog-if 00 [VGA controller])
Kernel driver in use: i915
Kernel modules: i915
However I have errors in /var/log/lightdm/x-0.log which I added to this post. No dmesg err but lots of warnings.
Adjunto | Tamaño |
---|---|
x-0.log | 9.21 KB |
lspci_vnnk.txt | 10.09 KB |
dmesg_warn.log | 3.71 KB |
Sorry, the fault has gone underground. I have fairly good evidence in /var/log/X.0.log that I undid the workaround. qpdf produced readable output that was modestly larger than the original. Neither exercises the restart, so I don't know whether linearising would have suppressed it earlier.
- Inicie sesión o regístrese para enviar comentarios