After the commit: commit 62f44052573b475a7b4c24f8e2da9bd6a8f1bd70 Author: Michel Dänzer <michel.daenzer@amd.com> Date: Tue Jun 21 15:41:45 2016 +0900 xfree86/modes: Move gamma initialization to xf86RandR12Init12 v2 With xinerama enabled, every screen on xinerama uses "different" random color on startup instead of default black color. The effect is that eg one screen is orange, next one is blue etc.
Created attachment 130335 [details] xorg.conf
Created attachment 130336 [details] log
Only for your information when you're at this bug: Reverting the changes from file: xf86Crtc.c are sufficient to make it all work back again - the monitors are black when starting Xorg, no gamma problems on xinerama. Overall - I have to revert the xf86Crtc.c changes and also apply a patch from bug: https://bugs.freedesktop.org/show_activity.cgi?id=100293 to have my Xorg working properly.
I get updates about this report via the xorg-team mailing list. Thanks for the report, I'll look into it when I get a chance.
Created attachment 130656 [details] [review] a reverting patch to have it back working (fwiw)
The issue here appears to be that the colormap isn't applied properly with Xinerama enabled. The code moved by the bisected commit merely papered over this, AFAICT e.g. changing gamma wouldn't have any effect on any screens but the first one before either. FWIW, changing gamma e.g. with xgamma works around the issue on the first screen for now.
Created attachment 130733 [details] First monitor of the xinerama set
Created attachment 130734 [details] Second monitor of the xinerama set
Created attachment 130735 [details] A clean desktop for the comparision
Created attachment 130773 [details] [review] Example fix for the radeon driver Actually, this happens because the driver's drmmode_set_mode_major uses crtc->gamma_red/green/blue directly. Since Xorg 1.19, those do not contain the final gamma values to be loaded to the hardware anymore[0]. This is an example fix for the radeon driver, should be easy to adapt to the nouveau driver. [0] Even with RandR enabled; this can result in other incorrect behaviour, e.g. when switching modes or VTs while something like redshift is setting a non-default gamma.
Never mind, I'll send a different patch or patches to fix this bug and bug 100293.
Comment on attachment 130773 [details] [review] Example fix for the radeon driver Fix posted for review: https://patchwork.freedesktop.org/patch/149863/
The patch is working. Please close this bug after merging. Thank you.
Thanks for the report, fixed in Git master: https://cgit.freedesktop.org/xorg/xserver/commit/?id=41dafcc2a2942fc4c94ce3cbafc4a1b413c460c3
Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.