linux-kernel - Re: [RFC PATCH 1/5] KVM: Split tdp_mmu

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Tue, 18 Apr 2023 10:14:22 -0700
From:   Sagi Shahar <sagis@...gle.com>
To:     Zhi Wang <zhi.wang.linux@...il.com>
Cc:     kvm@...r.kernel.org, linux-kernel@...r.kernel.org, x86@...nel.org,
        Sean Christopherson <seanjc@...gle.com>,
        Paolo Bonzini <pbonzini@...hat.com>,
        Isaku Yamahata <isaku.yamahata@...el.com>,
        Erdem Aktas <erdemaktas@...gle.com>,
        David Matlack <dmatlack@...gle.com>,
        Kai Huang <kai.huang@...el.com>,
        Chao Peng <chao.p.peng@...ux.intel.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        Dave Hansen <dave.hansen@...ux.intel.com>
Subject: Re: [RFC PATCH 1/5] KVM: Split tdp_mmu_pages to private and shared lists

On Mon, Apr 17, 2023 at 12:37 PM Zhi Wang <zhi.wang.linux@...il.com> wrote:
>
> On Fri,  7 Apr 2023 20:19:17 +0000
> Sagi Shahar <sagis@...gle.com> wrote:
>
> This patch is actually adding a separate counter for accounting private
> tdp mmu page not really introducing a new tdp_mmu_pages list for private
> pages. I guess better refine the tittle to reflect what this patch
> is doing.

Thanks for catching this. tdp_mmu_pages actually used to be a list of
pages when I first developed this code but was replaced with a counter
in https://lore.kernel.org/all/20221019165618.927057-6-seanjc@google.com/

>
> > tdp_mmu_pages holds all the active pages used by the mmu. When we
> > transfer the state during intra-host migration we need to transfer the
> > private pages but not the shared ones.
> >
> Maybe explain a little bit about how the shared one is processed. Guess
> one sentence is enough.

How about:
tdp_mmu_pages holds all the active pages used by the mmu. When we
transfer the state during intra-host migration we need to transfer the
private pages but not the shared ones. The shared pages are going to
be re-faulted as needed on the destination, but that approach doesn't
work for private pages which stores information in the secure EPT.

> > Keeping them in separate counters makes this transfer more efficient.
> >
> > Signed-off-by: Sagi Shahar <sagis@...gle.com>
> > ---
> >  arch/x86/include/asm/kvm_host.h |  5 ++++-
> >  arch/x86/kvm/mmu/tdp_mmu.c      | 11 +++++++++--
> >  2 files changed, 13 insertions(+), 3 deletions(-)
> >
> > diff --git a/arch/x86/include/asm/kvm_host.h b/arch/x86/include/asm/kvm_host.h
> > index ae377eec81987..5ed70cd9d74bf 100644
> > --- a/arch/x86/include/asm/kvm_host.h
> > +++ b/arch/x86/include/asm/kvm_host.h
> > @@ -1426,9 +1426,12 @@ struct kvm_arch {
> >       struct task_struct *nx_huge_page_recovery_thread;
> >
> >  #ifdef CONFIG_X86_64
> > -     /* The number of TDP MMU pages across all roots. */
> > +     /* The number of non-private TDP MMU pages across all roots. */
> >       atomic64_t tdp_mmu_pages;
> >
> > +     /* Same as tdp_mmu_pages but only for private pages. */
> > +     atomic64_t tdp_private_mmu_pages;
> > +
> >       /*
> >        * List of struct kvm_mmu_pages being used as roots.
> >        * All struct kvm_mmu_pages in the list should have
> > diff --git a/arch/x86/kvm/mmu/tdp_mmu.c b/arch/x86/kvm/mmu/tdp_mmu.c
> > index 58a236a69ec72..327dee4f6170e 100644
> > --- a/arch/x86/kvm/mmu/tdp_mmu.c
> > +++ b/arch/x86/kvm/mmu/tdp_mmu.c
> > @@ -44,6 +44,7 @@ void kvm_mmu_uninit_tdp_mmu(struct kvm *kvm)
> >       destroy_workqueue(kvm->arch.tdp_mmu_zap_wq);
> >
> >       WARN_ON(atomic64_read(&kvm->arch.tdp_mmu_pages));
> > +     WARN_ON(atomic64_read(&kvm->arch.tdp_private_mmu_pages));
> >       WARN_ON(!list_empty(&kvm->arch.tdp_mmu_roots));
> >
> >       /*
> > @@ -373,13 +374,19 @@ static void handle_changed_spte_dirty_log(struct kvm *kvm, int as_id, gfn_t gfn,
> >  static void tdp_account_mmu_page(struct kvm *kvm, struct kvm_mmu_page *sp)
> >  {
> >       kvm_account_pgtable_pages((void *)sp->spt, +1);
> > -     atomic64_inc(&kvm->arch.tdp_mmu_pages);
> > +     if (is_private_sp(sp))
> > +             atomic64_inc(&kvm->arch.tdp_private_mmu_pages);
> > +     else
> > +             atomic64_inc(&kvm->arch.tdp_mmu_pages);
> >  }
> >
> >  static void tdp_unaccount_mmu_page(struct kvm *kvm, struct kvm_mmu_page *sp)
> >  {
> >       kvm_account_pgtable_pages((void *)sp->spt, -1);
> > -     atomic64_dec(&kvm->arch.tdp_mmu_pages);
> > +     if (is_private_sp(sp))
> > +             atomic64_dec(&kvm->arch.tdp_private_mmu_pages);
> > +     else
> > +             atomic64_dec(&kvm->arch.tdp_mmu_pages);
> >  }
> >
> >  /**
>