Re: [PATCH 1/1] x86: Exclude E820_RESERVED regions and memory holes above 4 GB from direct mapping.

["H. Peter Anvin" <hpa@zytor.com>]

On 12/16/2011 09:42 AM, Yinghai Lu wrote:
>
> no, you change the meaning max_low_pfn_mapped and max_pfn_mapped for x86_64 at least.
>
> before your patch:
> max_low_pfn_mapped is the mapped pfn beblow 4g.
> max_pfn_mapped: is mapped pfn.
>
> after your patch, those two variables does not mean the memory [0, max_low_pfn_mapped) and [4g<<12, max_pfn_mapped)
> are really mapped.
>

And that's exactly the problem.  It is BROKEN -- as in fundamentally 
dangerous -- for these mappings to exist.  It is because the model is 
too inflexible.

> so in arch/x86/platform/efi/efi.c
>
>                  if (end_pfn<= max_low_pfn_mapped
>                      || (end_pfn>  (1UL<<  (32 - PAGE_SHIFT))
>                          &&  end_pfn<= max_pfn_mapped))
>                          va = __va(md->phys_addr);
>                  else
>                          va = efi_ioremap(md->phys_addr, size, md->type);
>
>
> and others will have problem.
>
> to solve your problem:
> 1. unmap the HT range ?
> 2. or introduce mapped_pfn_range array?

1 is fundamentally a braindead hack that solves one case without solving 
the overall problem.

For 2 - why can't we simply make the invariant that E820_RAM is mapped 
and nothing else, with the sole exceptions being the 1 MiB (fixed MTRR)?

For things like efi.c we should make sure to have interfaces instead of 
open-code this kind of stuff.

	-hpa