Hey all, I'm getting two blank screens (but still powered up, so I guess they're getting some signal) on Xorg startup, and if I SSH in and kill the X server, the workstation reboots. Debian Sid, xserver-xorg-core 1.6.5, xserver-xorg-video-radeonhd 1.3.0 I've attached what I hope are the relevant logs, I've tried testing with RandR both enabled and disabled (and this should be reflected in the filenames). Thanks very much and please let me know if there's anything I can do to help, David Harris
Created attachment 31957 [details] Output of 'lspci -vvv'
Created attachment 31958 [details] Log from 'Xorg -logverbose 7 :0' before server kill/shutdown, with RandR enabled
Created attachment 31959 [details] Copy of xorg.conf with RandR enabled
Created attachment 31960 [details] dmesg log after Xorg start, before server kill/shutdown, with RandR enabled
Created attachment 31961 [details] Log from 'Xorg -logverbose 7 :0' before server kill shutdown, with RandR disabled
Created attachment 31962 [details] dmesg log after Xorg start, before server kill/shutdown, with RandR disabled
Created attachment 31963 [details] Copy of xorg.conf with RandR disabled
And I suck for not mentioning that the card is an MSI R4850, model R4850-2D512-OC. Thanks again, Dave
Lastly (I hope :), I should mention that the card has two DVI-I outputs. The top one is connected to an LCD via a DVI cable, the bottom one is connected to an LCD via a DVI-I->15-pin DSUB adapter. Thanks, Dave
You seem to be running with KMS, which radeonhd does not support.
Okay, I'll try re-running with KMS explicitly disabled (I presume 'modprobe radeon modeset=0' will work, please let me know if there's a better way - the radeonhd manpage didn't seem to provide an option). Thanks, Dave
Hey, yep, loading the module manually with 'modeset=0' seemed to fix the issues completely. Thanks a bunch! :) If it's possible for the Xorg driver to decide whether to use KMS or not, that'd be great too. Thanks again, Dave
You cannot select to not use KMS in the driver - if it's active, it's active. Unfortunately. It would be much better if radeonhd would support KMS as well.
(In reply to comment #13) > You cannot select to not use KMS in the driver - if it's active, it's active. > Unfortunately. > > It would be much better if radeonhd would support KMS as well. Or at least if radeonhd would cleanly abort initialization when KMS is detected. Egbert, do you need some help with that KMS detection patch of mine?
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.