On Wednesday, February 23, 2011 02:15:44 pm Jesse Barnes wrote:
>
> But maybe using MCFG's bus numbers *is* the canonical way of getting
> the max bus... I don't know what Windows does here, but I can imagine
> that it exposes all these devices to various drivers. Perhaps they're
> hidden so that earlier versions of Windows won't "yellow bang" the
> unknown devices, while newer versions use MCFG.
[I added linux-acpi to the cc: list.]
MCFG is for communicating memory-mapped configuration space addresses.
I don't see anything in the specs about getting bus numbers from it.
I would think _CRS is the canonical way of getting bus ranges because:
- PNP0A03/08 devices have _CRS indicating bus range
[PCI Firmware 3.0, sec 4.3.2.1].
- PNP0A03 devices (PCI or PCI-X host bridges) need not support
the enhanced configuration access mechanism, so wouldn't need
MCFG, but we still need the bus range.
- DIG64 systems may have PNP0A08 devices but no MCFG, but again,
we still need the bus range [PCI Firmware 3.0, sec. 4.1.5].
> Does anyone have an affected machine with Windows 7 or something so we
> can verify?
If we added MCFG support to qemu, it should be pretty easy to see how
Windows handles it, but I'm not in a position to do that right now.
Bjorn
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
