This sort of sounds similar to the 965 issue we solved by switching to EXA+greedy migration heuristic. Can you test this by launching X in hardy configured to use EXA and the "greedy" migration heuristic? For example:
Section "Device" .... Driver "intel" Option "AccelMethod" "EXA" Option "MigrationHeuristic" "greedy" .... EndSection
If this solves the issue, then we should extend the fix for bug 177492 to cover this chipset as well.
This sort of sounds similar to the 965 issue we solved by switching to EXA+greedy migration heuristic. Can you test this by launching X in hardy configured to use EXA and the "greedy" migration heuristic? For example:
Section "Device" stic" "greedy"
....
Driver "intel"
Option "AccelMethod" "EXA"
Option "MigrationHeuri
....
EndSection
If this solves the issue, then we should extend the fix for bug 177492 to cover this chipset as well.