Re: [RFC PATCH 7/7] MdePkg/ProcessorBind AARCH64: limit MAX_ALLOC_ADDRESS to 48 bits

[Laszlo Ersek <lersek@redhat.com>]

On 12/07/18 12:23, Ard Biesheuvel wrote:
> Limit MAX_ALLOC_ADDRESS to 48 bits on AArch64 so that the memory
> handling routines running at boot time take care not to allocate
> memory that the CPU itself cannot access due to the fact that it
> runs with 4 KB pages and thus an address space that is limited to
> 256 TB.
> 
> Contributed-under: TianoCore Contribution Agreement 1.1
> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
> ---
>  MdePkg/Include/AArch64/ProcessorBind.h | 5 +++++
>  1 file changed, 5 insertions(+)
> 
> diff --git a/MdePkg/Include/AArch64/ProcessorBind.h b/MdePkg/Include/AArch64/ProcessorBind.h
> index 968c18f915ae..a8c698484a1d 100644
> --- a/MdePkg/Include/AArch64/ProcessorBind.h
> +++ b/MdePkg/Include/AArch64/ProcessorBind.h
> @@ -142,6 +142,11 @@ typedef INT64   INTN;
>  ///
>  #define MAX_ADDRESS   0xFFFFFFFFFFFFFFFFULL
>  
> +///
> +/// Maximum address usable at boot services time (48 bits for 4 KB pages)
> +///
> +#define MAX_ALLOC_ADDRESS   0xFFFFFFFFFFFFULL
> +
>  ///
>  /// Maximum legal AArch64 INTN and UINTN values.
>  ///
> 

I think this patch is safer than the previous variant. If this series
omits some replacements, from MAX_ADDRESS to MAX_ALLOC_ADDRESS, at
various spots in the edk2 tree, we should be able to fix those up
incrementally. Whereas changing MAX_ADDRESS was a heavy hammer that hit
everything simultaneously.

Reviewed-by: Laszlo Ersek <lersek@redhat.com>

Thanks
Laszlo