Re: [PATCH v4 06/15] x86/sev: Perform PVALIDATE using the SVSM when not at VMPL0

[Borislav Petkov]

On Wed, May 22, 2024 at 04:14:54PM -0500, Tom Lendacky wrote:
> Right, changing that from previous comment. But are you also asking that the
> if / else style be used?

Yeah, please. It is trivial and thus more readable this way.

> Sure, I'll add more comments or expand the comment above.

Yes, and pls split it into helpers. Those bunch of nested loops are
kinda begging to be separate function helpers.

> That's because you can only pass a certain number of entries to the SVSM to
> handle at one time. If the kernel is in the process of validating, say,
> 1,000 entries, but the CA can only hold 511 at a time, then after it reaches
> the 511th entry, the SVSM must be called. Upon return, the kernel resets the
> CA area and builds up the entries in there again, calling the SVSM again
> when the area is again full or we reach the last entry to be validated.
> 
> I'll add more detail in the comments.

Yeah, that "portioning" must be explained there.

Thx.

-- 
Regards/Gruss,
    Boris.

https://people.kernel.org/tglx/notes-about-netiquette