Comment 46 for bug 195051

I commented on the upstream bug, and will do so here. I showed this bug to Alex Deucher; he hadn't seen this particular bug but has an idea what the underlying cause is. When -ati was developed they did not have access to the Radeon register documentation so made a lot of guesses, that have been proven wrong since then as the Radeon documentation has become available. As a result, he believes a lot of registers have been incorrectly mapped in -ati, causing random failures such as have been described. He plans to get these mappings corrected within the next 3-6 months. So there's no chance this will be fixed for Hardy, but hopefully will be in place for Intrepid.

I don't know how hard the register mapping work is, but suspect it's pretty straightforward grunt coding work. If anyone is anxious to see these issues fixed and interested in putting some time into it, please contact me or Alex (aka agd5f on #xorg-devel/Freenode) for details.