"Nathan J. Williams" wrote:

> A round of debugging on the uhci machine check (with key assists from
> Jason) led to finding and fixing a flaw in the usb system's handling
> of DMA addresses. This problem would affect any system where PCI DMA
> addresses are different from memory physical addresses (everything
> besides i386?).

Great!  I hope we'll those fixes committed soon. :)


--

        -- Lennart