Re: [PATCH v3 1/3] x86/hyperv: Use slow_virt_to_phys() in page transition hypervisor callback

["Edgecombe, Rick P" <rick.p.edgecombe@intel.com>]

On Fri, 2024-01-05 at 10:30 -0800, mhkelley58@gmail.com wrote:
> + * It is also used in callbacks for CoCo VM page transitions between
> private
> + * and shared because it works when the PRESENT bit is not set in
> the leaf
> + * PTE. In such cases, the state of the PTEs, including the PFN, is
> otherwise
> + * known to be valid, so the returned physical address is correct.
> The similar
> + * function vmalloc_to_pfn() can't be used because it requires the
> PRESENT bit.

I'm not sure about this comment. It is mostly about callers far away
and other functions in vmalloc. Probably a decent chance to get stale.
It also kind of begs the question of why vmalloc_to_pfn() requires the
present bit in the leaf.

It seems the first part of the comment is about why this is needed when
__pa() exists. One reason given is that __pa() doesn't work with
vmalloc memory. Then the next bit talks about another similar function
that works with vmalloc memory.

So the comment is a risk to get stale, and leaves me a little confused
why this function exists.

I think the reason is because vmalloc_to_pfn() *only* works with
vmalloc memory and this is needed to work on other alias mappings.