MIME-Version: 1.0
References: <20230203192822.106773-1-vipinsh@google.com> <20230203192822.106773-4-vipinsh@google.com>
 <Y+GTyVHvpskr8YxD@google.com>
In-Reply-To: <Y+GTyVHvpskr8YxD@google.com>
From:   Vipin Sharma <vipinsh@google.com>
Date:   Tue, 7 Feb 2023 09:50:15 -0800
Message-ID: <CAHVum0dxAnm7SP9iCBk2eMeTNr5u_H=TTgZZJSJpx0Fxw5P-Ag@mail.gmail.com>
Subject: Re: [Patch v2 3/5] KVM: x86/mmu: Optimize SPTE change for aging gfn range
To:     David Matlack <dmatlack@google.com>
Cc:     seanjc@google.com, pbonzini@redhat.com, bgardon@google.com,
        kvm@vger.kernel.org, linux-kernel@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Mon, Feb 6, 2023 at 3:57 PM David Matlack <dmatlack@google.com> wrote:
>
> On Fri, Feb 03, 2023 at 11:28:20AM -0800, Vipin Sharma wrote:
> > No need to check all of the conditions in __handle_changed_spte(). Aging
> > a gfn range implies resetting access bit or marking spte for access
> > tracking.
> >
> > Use atomic operation to only reset those bits. This avoids checking many
> > conditions in __handle_changed_spte() API. Also, clean up code by
> > removing dead code and API parameters.
>
> Suggest splitting out the dead code cleanup to make it easier to review.
>

Sounds good.

> >
> > Signed-off-by: Vipin Sharma <vipinsh@google.com>
> > ---
> >  arch/x86/kvm/mmu/tdp_mmu.c | 68 ++++++++++++++------------------------
> >  1 file changed, 25 insertions(+), 43 deletions(-)
> >
> > diff --git a/arch/x86/kvm/mmu/tdp_mmu.c b/arch/x86/kvm/mmu/tdp_mmu.c
> > index 83f15052aa6c..18630a06fa1f 100644
> > --- a/arch/x86/kvm/mmu/tdp_mmu.c
> > +++ b/arch/x86/kvm/mmu/tdp_mmu.c
> > @@ -1251,32 +1228,37 @@ static __always_inline bool kvm_tdp_mmu_handle_gfn(struct kvm *kvm,
> >  /*
> >   * Mark the SPTEs range of GFNs [start, end) unaccessed and return non-zero
> >   * if any of the GFNs in the range have been accessed.
> > + *
> > + * No need to mark corresponding PFN as accessed as this call is coming from
> > + * MMU notifier for that page via HVA.
>
> Thanks for adding this comment.
>
> Can you just extend it to mention that the information is passed via the
> return value? e.g.
>
>  * No need to mark corresponding PFN as accessed as this call is coming
>  * from the clear_young() or clear_flush_young() notifier, which uses
>  * the return value to determine if the page has been accessed.
>

Sure.