On 11/11/2010 12:50 AM, Jan Beulich wrote:
quoted text
>>>> On 11.11.10 at 01:54, Yinghai Lu <yinghai@kernel.org> wrote:
>> @@ -191,10 +208,12 @@ void __cpuinit fam10h_check_enable_mmcfg
>>  		/* only trust the one handle 256 buses, if acpi=off */
>>  		if (!acpi_pci_disabled || busnbits >= 8) {
>>  			u64 base;
>> -			base = val & (0xffffULL << 32);
>> +			base = val & (FAM10H_MMIO_CONF_BASE_MASK <<
>> +					FAM10H_MMIO_CONF_BASE_SHIFT);
>>  			if (fam10h_pci_mmconf_base_status <= 0) {
>>  				fam10h_pci_mmconf_base = base;
>>  				fam10h_pci_mmconf_base_status = 1;
>> +				e820_add_mmconf_range(busnbits);
>>  				return;
>>  			} else if (fam10h_pci_mmconf_base ==  base)
>>  				return;
> 
> I don't think adding the range in this case is correct: Here, we
> found that the BIOS enabled the feature, and we're doing
> nothing to "correct" it, so we also should expect the range
> to be reserved by the BIOS.

this function is protected by PCI_CHECK_ENABLE_AMD_MMCONF

void __cpuinit fam10h_check_enable_mmcfg(void)
{
        u64 val;
        u32 address;

        if (!(pci_probe & PCI_CHECK_ENABLE_AMD_MMCONF))
                return;

and could have systems that setting in hw config is right, but BIOS doesn't put
them with e820 reserved entries.

Thanks

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