Created attachment 17316 [details] Xorg.log file with the non-working version Hi, In all version after git commit 566ba690de22f60059e9aee1b6b942088b3b6f01 I cannot move my external display above my panel interface. With the working version I just use: xrandr --output VGA_1 --above PANEL --mode 1440x900 with the non-working version the screen flashes and then X crashes with a sig 11.
Created attachment 17317 [details] Xorg.log with the working version
Created attachment 17318 [details] xrandr -q before changing the resolution and moving the external display
Created attachment 17319 [details] xrandr output after the switch
The commit has nothing to do with where the crash (according to the log file) actually happens. The commit actually dealt with PLL assignments while the log claimed that the crash happened somewhere in the fb code (underneath the xaa wrapper). Please try: Option "AccelMethod" "shadowfb" for testing.
Thanks. I tried this and now there is no crash anymore. I see: pevaneyn@sharrow:~ :) $ xrandr -q --verbose > xrandr-q-verbose-before pevaneyn@sharrow:~ :( $ xrandr --output VGA_1 --above PANEL xrandr: Configure crtc 0 failed pevaneyn@sharrow:~ :( $ xrandr --output VGA_1 --mode 1440x900 pevaneyn@sharrow:~ :) $ xrandr -q --verbose > xrandr-q-verbose-after pevaneyn@sharrow:~ :) $ xrandr --version Server reports RandR version 1.2 I'm attaching the two xrandr output files, the xorg.conf file and the Xorg.log file.
Created attachment 17342 [details] xrandr -q --verbose output with git HEAD before uncloning displays
Created attachment 17343 [details] xrandr -q --verbose output with git HEAD after uncloning displays
Created attachment 17344 [details] Xorg.0.log with Option "AccelMethod" "shadowfb"
Created attachment 17345 [details] current xorg.conf with Option "AccelMethod" "shadowfb"
Peter, I cannot reproduce this - it works here without any problems. The result is the expected one and I don't get an error like: > pevaneyn@sharrow:~ :( $ xrandr --output VGA_1 --above PANEL > xrandr: Configure crtc 0 failed ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ I've added some more debugging stuff to the driver. Please refetch the driver, run X in verbose 7 mode and re-execute the command sequence to bring the panel above the VGA_1 display. The failure should show in the log file. Please attach the resulting log. Thanks!
Created attachment 17927 [details] Test with git commit b0f470aa Hello, sorry for the delay: I was on holiday. I've taken the debugs I think, please see the attached (edited) typescript with the commands and the logoutput of the X server. If you need anything more do not hesitate to ask. Best regards, Peter
In the log file I see these lines: (II) RADEONHD(0): FUNCTION: rhdModeValidateCrtc (II) RADEONHD(0): rhdRROutputModeFixup: 1024x768 FAILED: [0xe] no clock available for mode (II) RADEONHD(0): rhdRROutputDpms: Output PANEL : Off I think I know what the problem is here. I will prepare a patch.
I've just pushed a bunch of fixes. Please retest. I might have found the issue that's causing the problem seen in the previous commit.
Created attachment 17988 [details] New test with git commit 996d309f Hi I retried and the problem remains. I've repeated the tests and attached the logs again. It seems the T60 is a difficult customer :-(. Best regards, Peter
OK, I cannot reproduce this here and there seems to be a plethora of Xservers with slightly different behavior of the RandR layer around. The only way I can find out what's going on here is to clutter teh code with a load of debug messages. Maybe I should provide you with a patch directly.
Created attachment 17999 [details] [review] Test code.
Created attachment 18001 [details] typescript with extra debugs Hello again, I ran the tests again with your extra debugs and attached the results. Best regards, Peter.
Created attachment 18006 [details] [review] Possible fix. Peter, could you please try if this fixes your problem? This patch should apply on top of the previous one. Thanks!
Hello Egbert, The patches work. I can now enable my external screen and use the laptop panel at the same time. Thanks! I'll now again track HEAD and report if it should break again. Best regards, Peter
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.