MIME-Version: 1.0
References: <20211120045046.3940942-1-seanjc@google.com> <20211120045046.3940942-10-seanjc@google.com>
In-Reply-To: <20211120045046.3940942-10-seanjc@google.com>
From:   Ben Gardon <bgardon@google.com>
Date:   Mon, 22 Nov 2021 12:10:03 -0800
Message-ID: <CANgfPd9=ce+JT3xEJy=p5MEfvkMGovEaBEu8KmxiZAJ1AA958g@mail.gmail.com>
Subject: Re: [PATCH 09/28] KVM: x86/mmu: Require mmu_lock be held for write in
 unyielding root iter
To:     Sean Christopherson <seanjc@google.com>
Cc:     Paolo Bonzini <pbonzini@redhat.com>,
        Vitaly Kuznetsov <vkuznets@redhat.com>,
        Wanpeng Li <wanpengli@tencent.com>,
        Jim Mattson <jmattson@google.com>,
        Joerg Roedel <joro@8bytes.org>, kvm@vger.kernel.org,
        linux-kernel@vger.kernel.org,
        Hou Wenlong <houwenlong93@linux.alibaba.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Fri, Nov 19, 2021 at 8:51 PM Sean Christopherson <seanjc@google.com> wrote:
>
> Assert that mmu_lock is held for write by users of the yield-unfriendly
> TDP iterator.  The nature of a shared walk means that the caller needs to
> play nice with other tasks modifying the page tables, which is more or
> less the same thing as playing nice with yielding.  Theoretically, KVM
> could gain a flow where it could legitimately take mmu_lock for read in
> a non-preemptible context, but that's highly unlikely and any such case
> should be viewed with a fair amount of scrutiny.
>
> Signed-off-by: Sean Christopherson <seanjc@google.com>
> ---
>  arch/x86/kvm/mmu/tdp_mmu.c | 16 ++++++++++++----
>  1 file changed, 12 insertions(+), 4 deletions(-)
>
> diff --git a/arch/x86/kvm/mmu/tdp_mmu.c b/arch/x86/kvm/mmu/tdp_mmu.c
> index 12a28afce73f..3086c6dc74fb 100644
> --- a/arch/x86/kvm/mmu/tdp_mmu.c
> +++ b/arch/x86/kvm/mmu/tdp_mmu.c
> @@ -159,11 +159,17 @@ static struct kvm_mmu_page *tdp_mmu_next_root(struct kvm *kvm,
>                 if (kvm_mmu_page_as_id(_root) != _as_id) {              \
>                 } else
>
> -#define for_each_tdp_mmu_root(_kvm, _root, _as_id)                             \
> -       list_for_each_entry_rcu(_root, &_kvm->arch.tdp_mmu_roots, link,         \
> -                               lockdep_is_held_type(&kvm->mmu_lock, 0) ||      \
> -                               lockdep_is_held(&kvm->arch.tdp_mmu_pages_lock)) \
> +/*
> + * Iterate over all valid TDP MMU roots.  Requires that mmu_lock be held for
> + * write, the implication being that any flow that holds mmu_lock for read is
> + * inherently yield-friendly and should use the yielf-safe variant above.

Nit: *yield-safe

> + * Holding mmu_lock for write obviates the need for RCU protection as the list
> + * is guaranteed to be stable.
> + */
> +#define for_each_tdp_mmu_root(_kvm, _root, _as_id)                     \
> +       list_for_each_entry(_root, &_kvm->arch.tdp_mmu_roots, link)     \
>                 if (kvm_mmu_page_as_id(_root) != _as_id) {              \
> +                       lockdep_assert_held_write(&(_kvm)->mmu_lock);   \

Did you mean for this lockdep to only be hit in this uncommon
non-matching ASID case?

>                 } else
>
>  static union kvm_mmu_page_role page_role_for_level(struct kvm_vcpu *vcpu,
> @@ -1063,6 +1069,8 @@ static __always_inline bool kvm_tdp_mmu_handle_gfn(struct kvm *kvm,
>         struct tdp_iter iter;
>         bool ret = false;
>
> +       lockdep_assert_held_write(&kvm->mmu_lock);
> +
>         rcu_read_lock();
>
>         /*
> --
> 2.34.0.rc2.393.gf8c9666880-goog
>