Re: [PATCH 5/5] KVM: x86/mmu: Move kvm_tdp_mmu_map()'s prolog and epilog to its caller

[David Matlack]

On Thu, Dec 29, 2022 at 1:06 PM Paolo Bonzini <pbonzini@xxxxxxxxxx> wrote:
>
> On 12/29/22 20:51, David Matlack wrote:
> > Your proposal (below) to split out the "lower half" of the page fault
> > handling routine works now because that's where all the divergence is.
> > But with the common MMU there's also going to be divergence in the fast
> > page fault handler. So I prefer to just keep the routines separate to
> > avoid thrashing down the road.
>
> Can you put the changes at the beginning of the common MMU series?

Can do. By "the changes" I assume you mean the yet-to-be-written
changes to split out a fast_page_fault() handler for the TDP MMU? Or
do you mean Sean's changes (this series)?

> Large parts of the whole common MMU refactoring can be merged piece by
> piece, so they can be taken as soon as they're ready.

Ack.