Jochen Voss wrote:
Maybe you should use binary search and apply it directly to the git tree or something, then you have a direct correspondence to one small patch as opposed to the large rc patches.My only idea is to take the 2.6.12-rc kernels one by one, to find out in which one it breaks and then to look at the corresponding patch. Maybe something suspicious can be spotted there. I guess this whole procedure would take me about one day, so I did not yet find time to do this.
johannes