Description of problem: Invoking Googleearth, I got a error log in dmesg as belows. "[drm:i915_vblank_swap] *ERROR* Invalid pipe 0" And driconf gives a message "Could not detect any configurable direct-rendering capable devices. DRIconf will be started in expert mode." But there exists /dev/dri/card0. Chipset : intel 82865G Version-Release number of selected component (if applicable): linux-2.6.24-0.41.rc3.git1.fc9 libdrm-2.4.0-0.1.fc9 driconf-0.9.1-6.fc9 googleearth-4.2.0198.2451 (beta) How reproducible: always Steps to Reproduce: 1. To invoke googleearth or driconf 2. 3. Actual results: Expected results: Additional info:
Created attachment 267941 [details] cat /proc/dri/0/*
Created attachment 269511 [details] Xorg.0.log
I have tested on linux-2.4.26-0.45.rc3.git1.fc9. But, X server cannot recognize dri. Xorg.0.log is attached
Could you please attach a copy of your "dmesg" and "lsmod" so we can determine if the correct modules are loaded?
I generate kernel with config options such as "CONFIG_DRM=y" and "CONFIG_DRM_i915=y". I tried to generate kernel with "CONFIG_DRM=m" and "CONFIG_DRM_I915=m". But, I get a same result. I attach dmesg and ".config" and a lsmod log
Created attachment 270581 [details] configuration file configuration file
Created attachment 270591 [details] dmesg log
Created attachment 270601 [details] lsmod log
Created attachment 270641 [details] dmesg with verbose drm debug messages on linux-2.6.24-0.45.rc3.git1.fc9
(In reply to comment #4) > Could you please attach a copy of your "dmesg" and "lsmod" so we can determine > if the correct modules are loaded? I attached already. And Now, I tested linux-2.6.24-0.69.rc3.git7.fc9. The "drm:i915_vblank_swap *ERROR Invalid pipe 0" error is still alive when invoking googleearth. My hardware is as follows pentium 4 2.6GHz 512MB memory 8MB AGP memory 82865G intel chip
Created attachment 276591 [details] more verbose drm debug messages
Created attachment 276601 [details] more verbose X.org.0.log
Can you retry with the latest kernel in rawhide? cheers, Kyle
I retry with linux-2.6.24-0.138.rc7.fc9. The error message does not appered. But, "glxgear" is failed displaying following message. "glxgears: intel_tris.c:114: intelStartInlinePrimitive: Assertion `intel->batch->id == intel->last_state_batch_id' failed. Aborted" Perhaps, mesa library has problem. I attach X.org.0.log .
Created attachment 291107 [details] X.org.0.log with linux-2.6.24-0.138.rc7.fc9 xorg-x11-server-1.4.99.1-0.12.fc9 xorg-x11-drv-i810-2.2.0-2.fc9
Please try the rpms at http://koji.fedoraproject.org/koji/taskinfo?taskID=336149 they incorporate a possible fix. cheers, Kyle
Dear Mr. Kyle. I tried with the package that you recommended above. 2.6.24-0.138.rc7.bz397631.fc9 But, glxgears is failed with same error message. And there are no warning messages in X.org.0.log such as (WW) intel(0): PRB0_CTL (0x0000f001) indicates ring buffer enabled (WW) intel(0): Existing errors found in hardware state. Please see an attached log.
Created attachment 291217 [details] X.org.0.log with 2.6.24-0.138.rc7.bz397631.fc9
I have tested again. !. environment linux-2.6.25-0.73.rc3.git9.fc9 mesa-7.1-0.17.fc9 libdrm-2.4.0-0.4.fc9 2. results Google worked well. The error message was gone.
I correct a mesa revision. mesa-7.1-0.17.fc9 has a problem which can not enable dri. mesa-7.1-0.16.fc9 is good.
It seems that the problem is gone away. I close this report. The another problem concerned with mesa package. The gcc option "-fvisibility=hidden" is restored again in mesa-7.1-0.17. This option causes an undefined symbol error while loading "i915_dri.so". undefined symbol are "_glthread_xxx". After generating mesa without "-fvisibility=hidden" options, the above error is not displayed and "dri" is enabled.