Re: [PATCH -mm] 1/2: MMCONFIG: validate against ACPI motherboardresources

[Robert Hancock]

Olivier Galibert wrote:
> On Tue, May 29, 2007 at 10:03:32PM -0600, Robert Hancock wrote:
> > -Validate that the area is reserved even if we read it from the
> > chipset directly and not from the MCFG table. This catches the case
> > where the BIOS didn't set the location properly in the chipset and
> > has mapped it over other things it shouldn't have.
>
> Just for the record, I still fundamentally disagree with that part.
> You're not catching what you think you're catching, since the chipset
> tells you what it is going to decode as mmconfig, no matter what is
> connected to it.

But we don't - not in the case where it's overlapped some other area 
with the MMCONFIG region. We saw a case of this on some of the Intel 
boards, where the MMCONFIG is a 128MB area at 0xf0000000, and when 
sizing the BARs on a PCI Express video device with a 256MB region, it 
ended up being located momentarily at f0000000-ffffffff, which 
overlapped the MMCONFIG area. That caused MMCONFIG to stop working, so 
apparently on that chipset, PCI Express devices connected to the 
northbridge have a higher decode priority than the MMCONFIG area.

If the BIOS is so screwed up that this becomes an issue, I don't think 
we can sanely try to use the table, since we can't anticipate all the 
potential problems that might result.

--
Robert Hancock      Saskatoon, SK, Canada
To email, remove "nospam" from hancockr@xxxxxxxxxxxxx
Home Page: http://www.roberthancock.com/

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/