Tom Horsley wrote: > > Oh, I did all that. The memory has identical timing and > the i386 kernel as well as memtest86 has no problems, it > is only the x86_64 kernel that has USB troubles. I suspect > some sort of IOMMU nonsense, but the simplest fix is to > just leave only 2 gig in the machine in any case. > It is interesting that it works with the i386 but not the X86_64 kernel. I have not worked with X86_64 kernels, so this is a shot in the dark, but could there be a compile time option for the amount of memory it supports like there is for the i386 kernels? I would not expect a problem at the 2G boundary with a 64bit processor. Mikkel -- Do not meddle in the affairs of dragons, for thou art crunchy and taste good with Ketchup!
