Re: [PATCH] ArmPkg ArmVirtPkg: fix the GIC base address variables as 64-bit

[Leif Lindholm <leif.lindholm@linaro.org>]

On Mon, Oct 17, 2016 at 10:20:30AM +0100, Ard Biesheuvel wrote:
> On 17 October 2016 at 09:33, Leif Lindholm <leif.lindholm@linaro.org> wrote:
> > On Mon, Oct 17, 2016 at 08:28:50AM +0100, Ard Biesheuvel wrote:
> >> > diff --git a/ArmVirtPkg/Library/ArmVirtGicArchLib/ArmVirtGicArchLib.c b/ArmVirtPkg/Library/ArmVirtGicArchLib/ArmVirtGicArchLib.c
> >> > index 64afc4d..16683ef 100644
> >> > --- a/ArmVirtPkg/Library/ArmVirtGicArchLib/ArmVirtGicArchLib.c
> >> > +++ b/ArmVirtPkg/Library/ArmVirtGicArchLib/ArmVirtGicArchLib.c
> >> > @@ -79,11 +79,11 @@ ArmVirtGicArchLibConstructor (
> >> >
> >> >      // RegProp[0..1] == { GICD base, GICD size }
> >> >      DistBase = SwapBytes64 (Reg[0]);
> >> > -    ASSERT (DistBase < MAX_UINT32);
> >> > +    ASSERT (DistBase < MAX_UINT64);
> >> >
> >>
> >> This becomes equivalent to 'DistBase != MAX_UINT64' given that a
> >> UINT64 cannot exceed MAX_UINT64. That is a nonsensical thing to
> >> assert, so it is better to simply drop it
> >
> > Random thought:
> > Could we keep the assert(s) and change the test to MAX_UINTN, to have
> > a sanity test over whether a 32-bit plaform ends up with a duff
> > address?
> 
> That seems like a useful thing in general, but given that we don't do
> that anywhere else, I'd rather we just remove them.

I won't argue with that.

/
    Leif