On 09/05/16 09:56, Ard Biesheuvel wrote:
> GCC 4.8 in RELEASE mode complains about GetPciIoTranslation () potentially
> not assigning IoTranslation, but does not notice that it returns failure in
> this case, which means IoTranslation is never referenced *unless* it has
> been assigned. So simply set IoTranslation to zero to help the compiler.
>
> Contributed-under: TianoCore Contribution Agreement 1.0
> Signed-off-by: Ard Biesheuvel <[email protected]>
> ---
> ArmVirtPkg/Library/FdtPciPcdProducerLib/FdtPciPcdProducerLib.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/ArmVirtPkg/Library/FdtPciPcdProducerLib/FdtPciPcdProducerLib.c
> b/ArmVirtPkg/Library/FdtPciPcdProducerLib/FdtPciPcdProducerLib.c
> index 10b47560cb9c..ea27cda7b77c 100644
> --- a/ArmVirtPkg/Library/FdtPciPcdProducerLib/FdtPciPcdProducerLib.c
> +++ b/ArmVirtPkg/Library/FdtPciPcdProducerLib/FdtPciPcdProducerLib.c
> @@ -128,6 +128,7 @@ FdtPciPcdProducerLibConstructor (
>
> PcdSetBool (PcdPciDisableBusEnumeration, FALSE);
>
> + IoTranslation = 0;
> RetStatus = GetPciIoTranslation (FdtClient, Node, &IoTranslation);
> if (!RETURN_ERROR (RetStatus)) {
> PcdSet64 (PcdPciIoTranslation, IoTranslation);
>
Ouch. This is exactly why "IoTranslation = 0" appeared at the beginning,
in commit 65bb13b0fd7f ("ArmVirtualizationPkg/VirtFdtDxe: parse
"pci-host-ecam-generic" properties"), and why we've been carrying it
around, most recently in commit d4cb9a30494d ("ArmVirtPkg: implement
FdtPciHostBridgeLib"). We forgot about it in commit c8f1a75aa997
("ArmVirtPkg/FdtPciPcdProducerLib: add handling of PcdPciIoTranslation").
Reviewed-by: Laszlo Ersek <[email protected]>
Thanks
Laszlo
_______________________________________________
edk2-devel mailing list
[email protected]
https://lists.01.org/mailman/listinfo/edk2-devel
