Comment 9 for bug 611951

I have forwarded this bug the intel developers at https://bugs.freedesktop.org/show_bug.cgi?id=29345 . Please register at bugs.freedesktop.org, add yourself to the CC: field of that bug report, and respond to any requests made by the intel developers. If you need help with fulfilling any of these requests, you may ask for help here.

There are a couple of things you may do to enhance the upstream bug report:

* Verify that the problem exists with the very latest and greatest xorg/libdrm etc. You can do this by using the xorg-edgers PPA at https://launchpad.net/~xorg-edgers/+archive/ppa .
* Verify that the problem exists with the very latest and greatest kernel. The current Maverick kernel is pretty up-to-date, but it is comforting for the developers to know that the problem exists in a non-distribution-specific kernel. You can test both the latest drm-intel-next kernel and the latest 2.6.35-rcX (now rc6, and the next is probably the 2.6.35 release itself). Look at the mainline builds page at https://wiki.ubuntu.com/Kernel/MainlineBuilds .
* Use drm.debug=0x02 kernel parameter to get a more verbose dmesg output. It may include more details about what went wrong during suspend. Attach resulting dmesg output before and after suspend.
* Grab the output of `sudo intel_reg_dumper` before suspend and after resume. There may be important differences there (but probably not).

Instead of relying on apport to tell you that there is a GPU error, you can see it in the dmesg output:

[17706.150645] render error detected, EIR: 0x00000010
[17706.150645] page table error
[17706.150645] PGTBL_ER: 0x00000001
[17706.150645] [drm:i915_report_and_clear_eir] *ERROR* EIR stuck: 0x00000010, masking
[17706.150703] render error detected, EIR: 0x00000010
[17706.150703] page table error
[17706.150703] PGTBL_ER: 0x00000001

Actually, it would be nice if you could let the computer be suspended for a few minutes, so that is easier to distinguish what happens on suspend and what happens on the following resume.