Am 13.03.2013 um 15:03 hat Anthony Liguori geschrieben:
> You are correct, it should look something like:
>
> if (addr & PCI_BASE_ADDRESS_SPACE_IO) {
> size = (1ULL << ctzl(addr & PCI_BASE_ADDRESS_IO_MASK));
> } else {
> size = (1ULL << ctzl(addr & PCI_BASE_ADDRESS_MEM_MASK));
> }
>
> This doesn't deal with 64-bit bars though.
>
> I'll update in the next round. I think this has worked for me because I
> have only done single device testing. I did switch from ffz when I
> rebased but I think ffz would still have this problem.
For when are you planning the next round? Will we get this in time to
make use of it in 1.5?
Kevin
