On 12/17/2009 07:52 PM, crdaley@gmail.com wrote:

Are you sure the problem lies within the driver itself? The fact that you can't locate a regression with git bisect leads me to believe the issue would be with some other software, and with so much graphics work being done in so many different area areas it would very well be possible. Off the top of my head it could also be in mesa, the kernel, or X.Org.

Yes, I'm sure it is the driver. I haven't completed the bisecting despite having spent hours on it. With the driver there is no way to automate the process other than building, installing, restarting X and checking and then marking good/bad. To my shagrin, there were only 90+ commits within the good/bad window I have. -- David C. Rankin, J.D.,P.E. Rankin Law Firm, PLLC 510 Ochiltree Street Nacogdoches, Texas 75961 Telephone: (936) 715-9333 Facsimile: (936) 715-9339 www.rankinlawfirm.com