[PATCH] acpi: mcfg: Set MCFGMMIONotReserved error to low

Colin Ian King colin.king at canonical.com
Thu Oct 25 08:48:33 UTC 2012


On 25/10/12 06:42, Alex Hung wrote:
> In PCI Firmware Specification, it claims "If the operating system does not natively
> comprehend reserving the MMCFG region, the MMCFG region must be reserved by firmware."
> However, Linux is an operating system that comprehend MMCFG, and therefore it is not
> a problem for not declaring it.
>
> Signed-off-by: Alex Hung <alex.hung at canonical.com>
> ---
>   src/acpi/mcfg/mcfg.c |    2 +-
>   1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/src/acpi/mcfg/mcfg.c b/src/acpi/mcfg/mcfg.c
> index 9dfd9fd..878c0f1 100644
> --- a/src/acpi/mcfg/mcfg.c
> +++ b/src/acpi/mcfg/mcfg.c
> @@ -200,7 +200,7 @@ static int mcfg_test1(fwts_framework *fw)
>   		if ((memory_map_list != NULL) &&
>   		    (!fwts_memory_map_is_reserved(memory_map_list, config->base_address))) {
>
> -			fwts_failed(fw, LOG_LEVEL_MEDIUM, "MCFGMMIONotReserved",
> +			fwts_failed(fw, LOG_LEVEL_LOW, "MCFGMMIONotReserved",
>   				"MCFG mmio config space at 0x%" PRIx64
>   				" is not reserved in the memory map table",
>   				config->base_address);
>

I'm not 100% sure how significant this is.  Can you take a look at what 
the kernel does if it isn't reserved by firmware?  I think the relevant 
code is in arch/x86/pci/*

Personally I don't mind dropping this to LOW, but I'd like to know what 
the kernel does nowadays for this issue.

Colin




More information about the fwts-devel mailing list