Re: [RFC Part2 PATCH 02/30] x86/sev-snp: add RMP entry lookup helpers

From: Borislav Petkov
Date: Thu Apr 15 2021 - 15:50:33 EST


On Thu, Apr 15, 2021 at 01:08:09PM -0500, Brijesh Singh wrote:
> This is from Family 19h Model 01h Rev B01. The processor which
> introduces the SNP feature. Yes, I have already upload the PPR on the BZ.
>
> The PPR is also available at AMD: https://www.amd.com/en/support/tech-docs

Please add the link in the bugzilla to the comments here - this is the
reason why stuff is being uploaded in the first place, because those
vendor sites tend to change and those links become stale with time.

> I guess I was trying to shorten the name. I am good with struct rmpentry;

Yes please - typedefs are used only in very specific cases.

> All those magic numbers are documented in the PPR.

We use defines - not magic numbers. For example

#define RMPTABLE_ENTRIES_OFFSET 0x4000

The 8 is probably

PAGE_SHIFT - RMPENTRY_SHIFT

because you have GPA bits [50:12] and an RMP entry is 16 bytes, i.e., 1 << 4.

With defines it is actually clear what the computation is doing - with
naked numbers not really.

> APM does not provide the offset of the entry inside the RMP table.

It does, kinda, but in the pseudocode of those new insns in APM v3. From
PVALIDATE pseudo:

RMP_ENTRY_PA = RMP_BASE + 0x4000 + (SYSTEM_PA / 0x1000) * 16

and that last

/ 0x1000 * 16

is actually

>> 12 - 4

i.e., the >> 8 shift.

Thx.

--
Regards/Gruss,
Boris.

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