On Thursday 14 January 2010 12:16:32 am Yinghai Lu wrote:
> Index: linux-2.6/arch/x86/pci/intel_bus.c
> ===================================================================
> --- linux-2.6.orig/arch/x86/pci/intel_bus.c
> +++ linux-2.6/arch/x86/pci/intel_bus.c
> @@ -41,6 +41,9 @@ static inline void print_ioh_resources(s
> #define IOH_LMMIOH_LIMITU 0x118
> #define IOH_LCFGBUS 0x11c
>
> +#define IOH_VTBAR 0x180
> +#define IOH_VTSIZE 0x2000 /* Fixed HW size (not programmable) */
> +
> static void __devinit pci_root_bus_res(struct pci_dev *dev)
> {
> u16 word;
> @@ -52,7 +55,7 @@ static void __devinit pci_root_bus_res(s
> int bus_base, bus_end;
>
> /* some sys doesn't get mmconf enabled */
> - if (dev->cfg_size < 0x120)
> + if (dev->cfg_size < 0x200)
> return;
I think the idea is that the config registers we're about to read:
IOH_LCFGBUS at 0x11c
IOH_LIO at 0x108
IOH_LMMIOL at 0x10c
and now
IOH_VTBAR at 0x180
are in extended config space, so the legacy PCI config accessors
might not be able to read them.
But I think this is an ugly way to check. Can't we just check whether
pci_read_config_dword() returns failure?
This actually raises an interesting point: if we ever reassign any PCI
resources, it is essential that we have correct information about these
host bridge windows. If we have to bail out early because we can't
access extended config space, e.g., maybe MMCONFIG is disabled for some
reason, our idea of the windows will be wrong.
This is a case where we'd be better off paying attention to the _CRS
information from ACPI.
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
