Please read this before reporting a bug:
https://wiki.archlinux.org/title/Bug_reporting_guidelines
Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.
REPEAT: Do NOT report bugs for outdated packages!
https://wiki.archlinux.org/title/Bug_reporting_guidelines
Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.
REPEAT: Do NOT report bugs for outdated packages!
FS#46566 - [xorg-server] Segfault with fullscreen GL or VA-API window on reverse PRIME output
Attached to Project:
Arch Linux
Opened by Matthias Schiffer (NeoRaider) - Monday, 05 October 2015, 09:36 GMT
Last edited by Andreas Radke (AndyRTR) - Monday, 04 January 2016, 18:48 GMT
Opened by Matthias Schiffer (NeoRaider) - Monday, 05 October 2015, 09:36 GMT
Last edited by Andreas Radke (AndyRTR) - Monday, 04 January 2016, 18:48 GMT
|
DetailsDescription:
I'm using reverse prime (xrandr --setprovideroutputsource) to access the outputs of my NVIDIA card (using the nouveau driver), while my primary card uses the intel driver. When I put a window using OpenGL or VA-API onto the reverse PRIME screeen in fullscreen, Xorg segfaults. Additional info: * This bug affects version 1.17.2-4; downgrading to 1.17.2-2 fixes the issue (but there are other ways to crash a 1.17.2-2 Xorg using reverse PRIME... as those aren't caused by Arch-specific patches, I'll report them upstream after further investigation) I can always reproduce this if I try directly after booting my machine; when I try this after having used it for a while, it sometimes doesn't crash. Xorg.log attached. 
Xorg.log.dri2
(82.8 KiB)
|
This task depends upon
Closed by Andreas Radke (AndyRTR)
Monday, 04 January 2016, 18:48 GMT
Reason for closing: Upstream
Additional comments about closing: Our devs are cc'ed to the upstream report. Currently we can't fix this downstream.
Monday, 04 January 2016, 18:48 GMT
Reason for closing: Upstream
Additional comments about closing: Our devs are cc'ed to the upstream report. Currently we can't fix this downstream.
I also got a similar random crash without a fullscreen window involved, backtrace attached.
I've also tried to reproduce this with the current Xorg master, but XineramaQueryScreens seems to be completely broken there (as least with reverse PRIME), so I couldn't even get my window manager (xmonad) to set a window to fullscreen there... (I'll report that upstream as well)