Created attachment 31654 [details] dmesg | tail -n 150 I'm experiencing X crashes, graphical corruption and sudden deactivations of a VGA monitor using nouveau and dual head with my NV43. I'm seeing the following from dmesg each time: [drm:edid_is_valid] *ERROR* EDID checksum is invalid, remainder is N [drm:edid_is_valid] *ERROR* Raw EDID: for varying N. There can be one or several consecutive occurrences of EDID errors for each problem. The remainder displayed (N) is not constant in the latter case. Every printed raw EDID appears correct however. Though commonly occurring when starting an X session or resuming from suspend, this also happens at seemingly random intervals with no apparent correlation with my uptime or how long my X server's been running. Graphical corruption is only visible after s2ram. N.B. Using current Fedora 12 packages, no modifications. Problems did not occur with Fedora 11 snapshots.
Created attachment 31655 [details] Xorg log
Created attachment 31656 [details] xorg.conf
X crashes and monitor deactivations no longer occur with latest upstream code. EDID errors still reported upon boot, and when resuming from suspend. Non-static graphical corruption still visible after resume.
I have worked around the problem as follows. Background: The nouveau driver interferes with the probing of nvidia's driver. Going into init 1 or init 3 from init 5 will result in nouveau still being loaded in the kernel. One must boot into init 1 or init 3 before nouveau has a chance to load for the NVIDIA installer to work correctly. Solution: Restart and boot into single mode by appending the capital letter S to grub's kernel line. run the installer as before, ignoring the warning about potential problems with init 1. Verified as a solution in UBUNTU 10.04 LTS and NVIDIA x86-195-32-24. Also, these newer (post 190) drivers have a ignore-EDID option; you can Google the readme of the driver if need be, that should solve these problems. --Tom G.
(In reply to comment #4) > I have worked around the problem as follows. > > Background: The nouveau driver interferes with the probing of nvidia's driver. > Going into init 1 or init 3 from init 5 will result in nouveau still being > loaded in the kernel. One must boot into init 1 or init 3 before nouveau has a > chance to load for the NVIDIA installer to work correctly. Thanks Tom; I'll keep that in mind if I ever want to use the proprietary driver with bad EDID devices.
Created attachment 39099 [details] dmesg output after (suspend to RAM)→(resume)
With the latest DRM and DDX, I only see these errors when suspending to RAM. The error only shows up in my logs a couple of times (still with different remainders), suspend/sleep is working great (no graphical corruption/flickering), and I'm no longer noticing any 'symptoms' other than kernel message buffer usage. I have attached dmesg output from a suspend→resume cycle in case anyone wants a look, but I've reniced the bug as it no longer bothers me.
It appears that this bug report has laid dormant for quite a while. Sorry we haven't gotten to it. Since we fix bugs all the time, chances are pretty good that your issue has been fixed with the latest software. Please give it a shot. (Linux kernel 3.10.7, xf86-video-nouveau 1.0.9, mesa 9.1.6, or their git versions.) If upgrading to the latest isn't an option for you, your distro's bugzilla is probably the right destination for your bug report. In an effort to clean up our bug list, we're pre-emptively closing all bugs that haven't seen updates since 2011. If the original issue remains, please make sure to provide fresh info, see http://nouveau.freedesktop.org/wiki/Bugs/ for what we need to see, and re-open this one. Thanks, The Nouveau Team
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.