Created attachment 32042 [details] kms_rv790_blurry_startup Occasionally (so far twice in the past couple of months), after a reboot, kms fails to properly setup the display (see attached camera photo) and blurry output is seen during system init and remains when launching X and for the entire session until the next reboot. Seems to be the output only as taking a screen snapshot inside kde when the problem occurs does not capture the issue. attaching dmesg from a blurry boot however I see nothing of interest when doing a diff of a good vs bad dmesg. 2.6.32 stock kernel mesa/ddx/drm git master rv790 card 5 metre dual link DVI cable NEC LCD3090WQXi @ 2560x1600
Created attachment 32043 [details] dmesg_from_blurry_boot
Does it get back to normal (and occasionally screw up again) if you turn the display off and on again like this: xrandr --output DVI-I --off; xrandr --output DVI-I --auto (not sure if DVI-I is the right identifier, xrandr will tell). I have similar issues with rv410 (bug 22755) and this is at least a better workaround than rebooting.
(In reply to comment #2) I just tried to trigger this again by continual rebooting. It happened again on the 6th reboot so I guess it isn't *that* occasional. Yes, xrandr --output DVI-0 --off;xrandr --output DVI-0 --auto does resurrect the display and rid the blurriness, though there are side effects - plasma-desktop crashes and needs manual restart, windows get resized to very small and desktop effects are disabled.
adjusting summary, not actually occasional at all, rather is happening every few reboots. I just wasn't rebooting frequently enough to notice.
Created attachment 32138 [details] kms_rv790_blurry_on_restore_from_dpms corruption on restore from dpms sleep. seen on 2.6.32 + drm-linus patch (http://files.iniza.org/drm-linus/vanilla-2.6.32/) + V6 pm patch. nothing in dmesg.
I've noticed on dpms restore that when the corruption occurs, my monitor OSD shows 1280x1600, not 2560x1600. this seems different to the boot blur, where the osd reports 2560x1600. The dpms restore issue is happening very regularly.
this still occurs under 2.6.33, but the monitor now consistently reports 2560x1600. I have also seen this under Windows 7 with catalyst 10.2 on the same machine. is this really a software issue?
unreproducible on current git d-r-t.
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.