Re: [PATCH V2] efi_high_alloc: use EFI_ALLOCATE_MAX_ADDRESS

[Matt Fleming <matt@console-pimps.org>]

On Mon, 25 Aug, at 04:33:54PM, Harald Hoyer wrote:
> 
> So is that ok, if other callers of efi_high_alloc() get an address >
> 4GB?

Yeah, the issue at hand is that some x86 EFI firmware cannot perform a
file read into a buffer above 4G. The actual allocation works fine.

If other callers want a high address buffer, they're not going to hit
that issue. We've no reason to believe that this bug affects arm64 at
all, for instance.

Working around this problem in efi_high_alloc() gives the impression
that there's something wrong with AllocatePages(), which isn't the case.

> Will the buggy EFI implementation work with the usage of the other
> caller's allocated memory? Or will they run into the same issues as
> the initramfs loader?

Based on the information we've gathered so far, there's no reason to
believe other callers will hit the same issue as the file reading code.

-- 
Matt Fleming, Intel Open Source Technology Center