FS#11854 - New Intel driver does not work
Attached to Project:
Arch Linux
Opened by Simone (simo91) - Wednesday, 22 October 2008, 19:04 GMT
Last edited by Jan de Groot (JGC) - Thursday, 06 November 2008, 22:00 GMT
Opened by Simone (simo91) - Wednesday, 22 October 2008, 19:04 GMT
Last edited by Jan de Groot (JGC) - Thursday, 06 November 2008, 22:00 GMT
|
Details
Description:
After upgrading xf86-video-intel by version 2.4.2-1 to 2.5.0-1 my DE (Gnome) won't start. I downgrade that package and now everything works. Additional info: * package version(s) * config and/or log files etc. When I start X Server with "startx" it prints an error repeated for many lines that I can't count. Steps to reproduce: Upgrade the package and after logging in GDM you will see a blank screen. |
This task depends upon
Closed by Jan de Groot (JGC)
Thursday, 06 November 2008, 22:00 GMT
Reason for closing: Fixed
Additional comments about closing: Problem was caused by broken DRI. This is fixed in 2.4.2-2.
Thursday, 06 November 2008, 22:00 GMT
Reason for closing: Fixed
Additional comments about closing: Problem was caused by broken DRI. This is fixed in 2.4.2-2.
INTEL 0
INTEL 0
INTEL 0
...
Last working packages are xorg-server-1.5.1-1 and xf86-video-intel-2.4.2-1.
After 20-30 seconds i can read this:
(EE) intel (0) underrun on pipe B!
(EE) intel (0) underrun on pipe B!
(EE) intel (0) underrun on pipe B!
... ecc
And later a single line with:
Error: Malformed line (no filename) - Ignoring entry.
Then I read again (EE) intel ... for (too) many lines..
Attached is an Xorg log with tiling disabled.
lscpi output and xorg log are attached.
In the meantime, downgrading to xf86-video-intel 2.4.2-1 solves the problem.
lspci (2 KiB)
The real problem is Compiz!! Perhaps we have to update it..
Try starting X keeping compiz disabled!
00:02.0 VGA compatible controller: Intel Corporation Mobile 945GME Express Integrated Graphics Controller (rev 03)
00:02.1 Display controller: Intel Corporation Mobile 945GM/GMS/GME, 943/940GML Express Integrated Graphics Controller (rev 03)
What could i do? Wait for support?
Some interesting bits from the Xorg.0.log:
(EE) intel(0): I830 Dma Initialization Failed
(II) AIGLX: Screen 0 is not DRI capable
(II) AIGLX: Loaded and initialized /usr/lib/xorg/modules/dri/swrast_dri.so
(II) GLX: Initialized DRISWRAST GL provider for screen 0
There's a kernel patch that goes hand in hand with 2.5.0 that affects the G4x series. Maybe that's what's needed to fix the G4x problems? http://lists.freedesktop.org/archives/xorg/2008-October/039599.html I checked my kernel git repo, this patch went into Linus' tree Oct 14. This patch hasn't made it into the 2.6.27.y maintenance tree (at least not as of v2.6.27.4) so I don't think it's in the kernel arch is using.
Should I open a new bug, or keep my issue here in the "intel 2.5.0 is made of fail" bug?
There's a kernel patch that goes hand in hand with 2.5.0 that affects the G4x series. Maybe that's what's needed to fix the G4x problems? http://lists.freedesktop.org/archives/xorg/2008-October/039599.html I checked my kernel git repo, this patch went into Linus' tree Oct 14. This patch hasn't made it into the 2.6.27.y maintenance tree (at least not as of v2.6.27.4) so I don't think it's in the kernel arch is using.
Should I open a new bug, or keep my issue here in the "intel 2.5.0 is made of fail" bug?
Either way, I think your bug is different but I'm a noob at this. However, I've learnt it's better to file a new bug if you are unsure. They are easy to close if it turns out to be a duplicate, and it generally helps keeping things organized. The other scenario, posting different bugs under same ticket is more of a hassle to work with for everybody.
However, this patch only solves the crash on G4x but not the other issues like broken DRI or this one here...
Joel: As I understand it, this version of the intel driver requires a corresponding patch to the kernel that fixes G4x memory allocation problems. That is, this fix requires complementary fixes in both the kernel (intel-agp) and the driver, and it seems that in shipping this version of the driver without a patched kernel, the G4x series has been disabled.
From the 2.4.98 changelog:
Eric Anholt (12):
*snip*
Fix broken stolen memory counting on G4X.
*snip*
This fix requires this patch: http://git.kernel.org/?p=linux/kernel/git/anholt/drm-intel.git;a=commit;h=2052746fc8397130c120f0194a89938b0b62b6cb The patch is from Anholt's tree, and he states on the commit log that this patch corresponds to the patch introduced into 2.4.98.
I think I'll go open a new bug for the 845 problem so that this bug can be closed if/when the G4x problems are resolved.