Bug 31169

Created attachment 39815 [details]
screen image

Created attachment 39816 [details]
Xorg.0.log

Please post full dmesg output while corrupted screen.

You should be able to catch it via "dmesg >dmesg.log" before rebooting.

Created attachment 39821 [details]
dmesg

Here it is. If you need additional info, please let me know.

This reporter also filed a bug downstream with Ubuntu about this same issue, at https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-nouveau/+bug/564572

He attached a mmiotrace to that bug, which unfortunately is too large to attach to bugzilla, but here's a link to it in launchpad:

https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-nouveau/+bug/564572/+attachment/1976762/+files/mmiotrace.log.zip

Does booting with "nouveau.uscript_lvds=0x0200" or "nouveau.uscript_lvds=0x0300" help the issue at all?

Err, scratch that: "nouveau.uscript_lvds=0x0100" or "nouveau.uscript_lvds=0x0300" I meant!

Yes! They both work: "nouveau.uscript_lvds=0x0100" and "nouveau.uscript_lvds=0x0300"! Thanks! I haven't seen any difference between the two of them, so I can't say which one is better.

Ok, cool, thanks :)  I'll look at implementing support for panels such as yours when I get to the office on Monday morning.

Ok, nouveau git has a couple of patches now that should make using these options unnecessary.

Are you able to test them please!

Thanks. I'll download the files and try it this evening. I have no Internet connection at home atm, but I should be able to give you a feedback tomorrow or the next day.

Ben, I've just tried to compile it, but I've got some errors during the compilation process (in nouveau_backlight.c). Unfortunately, I'm very busy this period and I don't have much time to experiment. I'll wait for the deb packages on xorg-edgers and let you know.

Hi,

dont know if this is related to this bug, but i dont have output in xorg too.

- linux-2.6.39-rc4 with all changes from http://cgit.freedesktop.org/nouveau/linux-2.6 up to 4a76dc872b9ba0a29399c7149e59190f21797e8e (drm/nouveau/pm: translate ramcfg strap through ram restrict table)
- xf86-video-nouveau git 8378443bd3b26b57ef2ae424a700e01ead813d33 (nv50-nvc0/exa: fix bug causing surface state to not be reemitted after flush)
- libdrm-2.4.25
- Mesa master git 6067a2a67f9a7aab2aee051469bea8af03747a95 (r600g: don't flush the dest caches on every draw)

Xorg.log: http://pastebin.com/5GHkDzp3
dmesg: http://pastebin.com/4f8kRhWf

it seems xorg is running and i cant see anything wrong in xorg.log and dmesg.

Stephan

Created attachment 47853 [details]
Xorg.0.log (git20110411)

Ben, I've just did a test using the latest ubuntu daily build (i386) with xserver-xorg-video-nouveau (1:0.0.16+git20110411+8378443-1); "NOUVEAU driver Date: Thu Mar 24 02:13:12 2011 +1000"

It still doesn't work, I wonder if the changes you were talking about were before or after this git snapshot. "nouveau.uscript_lvds=0x0300" (or 0x0100) worked just fine.

Hi Marius,

The changes mentioned are in the kernel module.

Make sure that your kernel includes the following commits
 * 8c3f6bb970413c6a537736f409a1bc9d8abd671c - drm/nouveau: recognise DCB connector type 0x41 as LVDS
 * b23b9e7109e74a2cb10705396148624016ad8f8f - drm/nv50: respect LVDS link count from EDID on SPWG panels

If you still have issues please provide dmesg and Xorg.log

Thanks

Thanks Emil!

It worked with the latest kernel - 3.0! Sorry, it took so long to confirm. Ben, thanks for the patches!

I'll change the status to fixed, then.