Re: [PATCH v2 8/9] KVM: X86: Optimize pte_list_desc with per-array counter

From: Peter Xu
Date: Wed Jul 28 2021 - 17:51:26 EST


On Wed, Jul 28, 2021 at 09:04:30PM +0000, Sean Christopherson wrote:
> > struct pte_list_desc {
> > u64 *sptes[PTE_LIST_EXT];
> > + /*
> > + * Stores number of entries stored in the pte_list_desc. No need to be
> > + * u64 but just for easier alignment. When PTE_LIST_EXT, means full.
> > + */
> > + u64 spte_count;
>
> Per my feedback to the previous patch, this should be above sptes[] so that rmaps
> with <8 SPTEs only touch one cache line. No idea if it actually matters in
> practice, but I can't see how it would harm anything.

Reasonable. Not sure whether this would change the numbers a bit in the commit
message; it can be slightly better but also possible to be non-observable.
Paolo, let me know if you want me to repost/retest with the change (along with
keeping the comment in the other patch).

Thanks for looking!

--
Peter Xu