Re: [PATCH, RESEND] x86/pat: Simplifying the PAT programming protocol
From: Kirill A. Shutemov
Date: Tue Feb 13 2024 - 11:57:29 EST
On Tue, Feb 13, 2024 at 05:15:14PM +0100, Borislav Petkov wrote:
> On Thu, Feb 01, 2024 at 12:17:32AM +0200, Kirill A. Shutemov wrote:
> > > So the "relaxation" is the removal of that CR0.CD requirement?
>
> So I'm looking at the SDM, revisions 081US and 082US.
>
> Section
>
> "12.11.8 MTRR Considerations in MP Systems"
The point is that PAT programming doesn't need to follow MTRR
considerations anymore.
Previously "Programming the PAT" section had this:
The operating system is responsible for ensuring that changes to a PAT
entry occur in a manner that maintains the consistency of the processor
caches and translation lookaside buffers (TLB). This is accomplished by
following the procedure as specified in Section 12.11.8, “MTRR
Considerations in MP Systems,” for changing the value of an MTRR in a
multiple processor system. It requires a specific sequence of
operations that includes flushing the processors caches and TLBs.
The new version points to MTTR consideration as one of possible way to
invalidate TLB and caches:
The operating system (OS) is responsible for ensuring that changes to a
PAT entry occur in a manner that maintains the consistency of the
processor caches and translation lookaside buffers (TLB). It requires the
OS to invalidate all affected TLB entries (including global entries) and
all entries in all paging-structure caches. It may also require flushing
of the processor caches in certain situations. This can be accomplished
in various ways, including the sequence below or by following the
procedure specified in Section 12.11.8, “MTRR Considerations in MP
Systems.” (See Section 4.10.4, “Invalidation of TLBs and
Paging-Structure Caches” for additional background information.) Also
note that in a multi-processor environment, it is the software's
responsibility to resolve differences in conflicting memory types across
logical processors that may arise from changes to the PAT (e.g., if two
logical processors map a linear address to the same physical address but
have PATs that specify a different memory type for that physical
address).
The new text follows with example of sequence that flushes TLB and
caches. And it doesn't touch CR0.CD.
--
Kiryl Shutsemau / Kirill A. Shutemov