David Laight <david@l8s.co.uk> writes:

> On Thu, Nov 08, 2007 at 02:56:35PM -0500, Stephen Degler wrote:
>> >
>> >It may have more address space, but there is still a physical address
>> >boundary at 4GB.
>> 
>> David, can you provide some insight into that?
>
> The one I was thinking of is that there are 32bit PCI bus devices...

I am not entirely clear on how tihs works, but I thought that the
cardbus/rbus controlled virtual addresses to which areas of pci space
were mapped.  So given a 32-bit cardbus device, I don't see any reason
why that couldn't be mapped at a >32-bit virtual address.

But, I would not be surprised to find trouble from doing this due to
things I don't understand.