[G45,DG45ID] No GEM, DRI, XvMC on PAE x86 server kernel.

Bug #388886 reported by Jonathan Sambrook
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
xserver-xorg-video-intel (Ubuntu)
Won't Fix
Undecided
Unassigned

Bug Description

Binary package hint: xserver-xorg-video-intel

I've been up to bleeding edge Version: 2:2.7.99.901+git20090617.b5cd2130-0ubuntu0sarvatt to try and get some stability, but things are worse than ever.

Compiz is a white screen no go. Dropping down to Metacity gives me a totatlly
busy CPU core with Xorg running ~98% when anything major happens on screen (eg.
loading a page in Firefox). Typing in terminal doesn't seem to sweat X.

Kernel: "2.6.30-9-server #10-Ubuntu SMP Fri Jun 12 14:12:48 UTC 2009 i686 GNU/Linux"

The i915.ko is loaded, but has a use count of zero and I can rmmod it whilst X stays running.

From lsmod with X running (full file is attached):

  i915 174056 0
  drm 158624 1 i915
  intel_agp 26340 1
  video 18848 1 i915

From dmesg (full file is attached):

  [ 15.654947] agpgart-intel 0000:00:00.0: Intel G45/G43 Chipset
  [ 15.655890] agpgart-intel 0000:00:00.0: detected 32764K stolen memory
  [ 15.658289] agpgart-intel 0000:00:00.0: AGP aperture is 256M @ 0xd0000000
  [ 15.847995] [drm] Initialized drm 1.1.0 20060810
  [ 16.114523] pci 0000:00:02.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
  [ 16.114571] pci 0000:00:02.0: setting latency timer to 64
  [ 16.115507] mtrr: type mismatch for d0000000,10000000 old: write-back new: write-combining
  [ 16.115567] [drm] MTRR allocation failed. Graphics performance may suffer.
  [ 16.115622] alloc irq_desc for 26 on cpu 0 node 0
  [ 16.115624] alloc kstat_irqs on cpu 0 node 0
  [ 16.115632] pci 0000:00:02.0: irq 26 for MSI/MSI-X
  [ 16.115674] [drm] Initialized i915 1.6.0 20080730 for 0000:00:02.0 on minor 0

  [ 52.480642] mtrr: type mismatch for d0000000,10000000 old: write-back new: write-combining

Points that jump out at me from Xorg.0.log (full file is attached):

  drmOpenDevice: node name is /dev/dri/card0
  drmOpenDevice: open result is 10, (OK)
  drmOpenByBusid: Searching for BusID pci:0000:00:02.0
  drmOpenDevice: node name is /dev/dri/card0
  drmOpenDevice: open result is 10, (OK)
  drmOpenByBusid: drmOpenMinor returns 10
  drmOpenByBusid: drmGetBusid reports pci:0000:00:02.0
  (EE) intel(0): [drm] Failed to detect GEM. Kernel 2.6.28 required.

  (II) intel(0): Integrated Graphics Chipset: Intel(R) G45/G43
  (--) intel(0): Chipset: "G45/G43"

  (--) intel(0): Using UXA for acceleration

  (II) intel(0): I830CheckAvailableMemory: 4991996 kB available
  (WW) intel(0): DRI2: failed to open drm device

  (==) intel(0): Intel XvMC decoder disabled
  (II) intel(0): Set up textured video
  (II) intel(0): direct rendering: Failed

Revision history for this message
Jonathan Sambrook (jonathan-hmmn) wrote :
Revision history for this message
Jonathan Sambrook (jonathan-hmmn) wrote :
Revision history for this message
Jonathan Sambrook (jonathan-hmmn) wrote :
Revision history for this message
Jonathan Sambrook (jonathan-hmmn) wrote :
Revision history for this message
Jonathan Sambrook (jonathan-hmmn) wrote :
Revision history for this message
Robert Hooker (sarvatt) wrote :

GEM does not work on PAE kernels currently, and the -server kernel has PAE enabled. There is a kernel you can try here with an experimental patch to enable GEM to work with PAE, but keep in mind that switching to x64 is by far a better option for system performance than using i386 with PAE.

https://edge.launchpad.net/~apw/+archive/red

tags: added: xorg-needs-kernel-fix
Robert Hooker (sarvatt)
summary: - [G45,DG45ID] No GEM, DRI, XvMC. Busy CPU.
+ [G45,DG45ID] No GEM, DRI, XvMC on PAE x86 server kernel.
Revision history for this message
Bryce Harrington (bryce) wrote :

The PAE kernel does not support GEM; this isn't an -intel issue, you're welcome to re-file a request for adding it to the PAE kernel

Changed in xserver-xorg-video-intel (Ubuntu):
status: New → Won't Fix
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.