A lot has changed in X lately... GEM, DRI2, without even talking about kms... So many things has changed that its hard to diagnose where the problem might be. It could even be many problems... :S 2009/4/14 Jonathan Brown

what if the problem may not be the intel or nvidia drivers, but the way all of those drivers might be interplaying with the latest new xorg?

Because I am seeing significant slowdowns in my X environment overall with the new xorg (I use xfce), and my use the nvidia driver.

----- Original Message ---- From: János Illés To: General Discusson about Arch Linux Sent: Tuesday, April 14, 2009 5:13:46 PM Subject: Re: [arch-general] Corruption with new intel driver

On Tue, Apr 14, 2009 at 22:51, Nicolas Bigaouette wrote:

...

Hi all, I just fully upgraded on two of my machines. These packages related to xorg were updated:

Unfortunately, I am facing some screen corruptions, as seen on this screenshot:

http://nbigaouette.inrs-emt.homelinux.net/linux/arch/xf86_video_intel_corrup...

this happens for me too with intel-legacy driver and QT4 apps.

-- IJ

The new X server will automatically load dri2 if you don't explicitly define dri in load module section, and dri2 needs uxa. So if you try to use dri+uxa or dri2+exa, X server would hang up just like you described. I met the same problem last night, and this is my conclusion from Xorg.log.0. On Thu, Apr 16, 2009 at 7:29 PM, Rafa Griman wrote:

Hi :)

On Wednesday 15 April 2009 00:56:27 Simon Braunstein wrote:

...

Hello,

...

Unfortunately, I am facing some screen corruptions, as seen on this screenshot: http://nbigaouette.inrs-emt.homelinux.net/linux/arch/xf86_video_intel_cor ru ption.png

Same problem here. It seems using Option "AccelMethod" "UXA" in the device section of xorg.conf solves the problem for me.

I'm also having issues. Tried using UXA and everything works fine until I log off: screen goes black and no response via keyboard (no ctrl+alt+Fn nor ctrl+alt+backspace).

Hardware is a Dell Latitude D610:

       00:02.1 Display controller: Intel Corporation Mobile 915GM/GMS/910GML                Express Graphics Controller (rev 03)

Haven't tried to ssh from another laptop.

I've removed the UXA line and added this to the xorg.conf:

       Option      "AccelMethod"  "exa"        Option      "MigrationHeuristic"  "greedy"

Now it seems to be working fine.

  Rafa

-- "We cannot treat computers as Humans. Computers need love."

rgriman@skype.com rgriman@jabberes.org

-- Best regards, Li Ye

I reported my problem upstream[1] and someone suggested adding: Option "ExaOptimizeMigration" "off" Maybe it can help for you... [1] http://bugs.freedesktop.org/show_bug.cgi?id=21191 2009/4/17 Sergi Pons Freixes

El Thursday 16 April 2009 13:43:58 李业 va escriure:

...

The new X server will automatically load dri2 if you don't explicitly define dri in load module section, and dri2 needs uxa. So if you try to use dri+uxa or dri2+exa, X server would hang up just like you described. I met the same problem last night, and this is my conclusion from Xorg.log.0.

I am working with no x.org file, so the X server should automagically choose the rigth combination (dri+exa, or dri2+uxa). But I am still having these "screen corruptions" in Qt4 apps. The last line of Xorg.0.log states: exaCopyDirty: Pending damage region empty!

Any clue?

I switched to no xorg.conf file too not long ago. But needing to change one value (here, the ExaOptimizeMigration option) will mean you will have to create a "full" minimalist xorg.conf. Here's am example: Section "ServerLayout" Identifier "Layout0" Screen 0 "Screen0" 0 0 EndSection Section "Device" Identifier "Device0" Driver "intel" Option "ExaOptimizeMigration" "off" EndSection Section "Screen" Identifier "Screen0" Device "Device0" EndSection (adapted from http://wiki.archlinux.org/index.php/Xorg#Without_xorg.conf ) 2009/4/18 Sergi Pons Freixes

El Friday 17 April 2009 16:16:19 Nicolas Bigaouette va escriure:

...

I reported my problem upstream[1] and someone suggested adding:

Option "ExaOptimizeMigration" "off"

Maybe it can help for you...

[1] http://bugs.freedesktop.org/show_bug.cgi?id=21191

Thank you for the suggestion. Now I have a doubt: as I am not using an xorg.conf file, should I create one with only:

Section "ServerFlags" Option "ExaOptimizeMigration" "off" EndSection

as contents? Or should I recreate the device section for the intel card (or worst of all, _all_ the sections) ? I was used to the old approach of hand- making the whole xorg.conf file, and this new "it just works" approach is confusing me a bit :).

Regards, Sergi

Hi :) On Thursday 16 April 2009 13:43:58 李业 wrote:

The new X server will automatically load dri2 if you don't explicitly define dri in load module section, and dri2 needs uxa. So if you try to use dri+uxa or dri2+exa, X server would hang up just like you described. I met the same problem last night, and this is my conclusion from Xorg.log.0.

Haven't been able to try this before but I've tried it now and still get the same issue (yup, latest patches installed). I've tried: - NOT using an xorg.conf file - using a minimal xorg.conf file WITHOUT the dri option - using a minimal xorg.conf file defining dri2 - any other combination my brain could work out But it still hangs. No matter what I do, I see the following in the Xorg.log (_without_ the Load "dri" option, minimal xorg.conf with AccelMethod UXA only): ----- *** ----- *** ----- *** ----- *** ----- *** ----- # grep -i dri /var/log/Xorg.0.log (II) "dri" will be loaded by default. (II) "dri2" will be loaded. This was enabled by default and also specified in the config file. (II) LoadModule: "dri2" (II) Loading /usr/lib/xorg/modules/extensions//libdri2.so (II) Module dri2: vendor="X.Org Foundation" (II) Loading extension DRI2 (II) LoadModule: "dri" (II) Loading /usr/lib/xorg/modules/extensions//libdri.so (II) Module dri: vendor="X.Org Foundation" (II) Loading extension XFree86-DRI (II) Loading /usr/lib/xorg/modules/drivers//intel_drv.so (II) intel(0): [DRI2] Setup complete (II) UXA(0): Driver registered support for the following operations: (II) intel(0): 0x007bf000-0x0fff4fff: DRI memory manager (254168 kB) (II) intel(0): direct rendering: DRI2 Enabled (II) AIGLX: Loaded and initialized /usr/lib/xorg/modules/dri/i915_dri.so (II) GLX: Initialized DRI2 GL provider for screen 0 drmOpenDevice: node name is /dev/dri/card0 drmOpenDevice: node name is /dev/dri/card0 (II) intel(0): [DRI2] Setup complete ----- *** ----- *** ----- *** ----- *** ----- *** ----- As you can see, dri gets loaded as well. I'm not sure if that's OK. And I am using UXA, but it still hangs up. My UXA testing was done using the following options (tried them all at once, one by one, skipping one or the other, ... any combination you can imagine): ----- *** ----- *** ----- *** ----- *** ----- *** ----- #Option "AccelMethod" "UXA" #Option "Framebuffer" "false" #Option "Tiling" "false" #Option "EnablePageFlip" "true" ----- *** ----- *** ----- *** ----- *** ----- *** ----- I don't use my laptop for gaming or 3D or scientific viz so I'm fine using the EXA acceleration. I'm just giving some feedback in case the maintainer needs/wants it. Rafa -- "We cannot treat computers as Humans. Computers need love." rgriman@skype.com rgriman@jabberes.org