Re: [PATCH, resend] x86/PCI: don't export a __devinit function

[Jesse Barnes <jbarnes@virtuousgeek.org>]

On Fri, 18 Feb 2011 09:12:11 +0000
"Jan Beulich" <JBeulich@novell.com> wrote:

> Like this:
> 
> --- a/arch/x86/pci/mmconfig-shared.c
> +++ b/arch/x86/pci/mmconfig-shared.c
> @@ -606,6 +606,16 @@ static void __init __pci_mmcfg_init(int 
>  	if (list_empty(&pci_mmcfg_list))
>  		return;
>  
> +	if (pcibios_last_bus < 0) {
> +		const struct pci_mmcfg_region *cfg;
> +
> +		list_for_each_entry(cfg, &pci_mmcfg_list, list) {
> +			if (cfg->segment)
> +				break;
> +			pcibios_last_bus = cfg->end_bus;
> +		}
> +	}
> +
>  	if (pci_mmcfg_arch_init())
>  		pci_probe = (pci_probe & ~PCI_PROBE_MASK) | PCI_PROBE_MMCONF;
>  	else {
> 
> I (unsuccessfully so far) tried to find someone at Intel who could
> confirm that this approach is usable namely on those Xeon 55xx
> systems that the edac driver tries to deal with, which is why I
> didn't submit this so far.

What's the actual user of pcibios_last_bus we're concerned about here?
Presumably pcibios_fixup_peer_bridges(), since that's the only place
that might be affected by a positive value...

So in that sense, the above seems like a good fix; we should initialize
pcibios_last_bus to a reasonable value based on the configuration we
discover in the MCFG table.

-- 
Jesse Barnes, Intel Open Source Technology Center