linux-kernel - Re: [PATCH 1/7] kvm: x86/mmu: Use KVM_MMU_ROOT_XXX for kvm_mmu_invalidate

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]

Message-ID: <CAJhGHyCgSzH5x9=OutgOcxCEg784woz9VouAwJ=K0xqXTF9Avw@mail.gmail.com>
Date:   Fri, 3 Feb 2023 22:51:15 +0800
From:   Lai Jiangshan <jiangshanlai@...il.com>
To:     Sean Christopherson <seanjc@...gle.com>
Cc:     linux-kernel@...r.kernel.org, Paolo Bonzini <pbonzini@...hat.com>,
        Lai Jiangshan <jiangshan.ljs@...group.com>,
        Thomas Gleixner <tglx@...utronix.de>,
        Ingo Molnar <mingo@...hat.com>, Borislav Petkov <bp@...en8.de>,
        Dave Hansen <dave.hansen@...ux.intel.com>, x86@...nel.org,
        "H. Peter Anvin" <hpa@...or.com>, kvm@...r.kernel.org
Subject: Re: [PATCH 1/7] kvm: x86/mmu: Use KVM_MMU_ROOT_XXX for kvm_mmu_invalidate_gva()

On Thu, Feb 2, 2023 at 9:21 AM Sean Christopherson <seanjc@...gle.com> wrote:

>
> This is logically correct, but there's potential (weird) functional change here.
> If this is called with an invalid root, then KVM will invalidate the GVA in all
> roots prior to this patch, but in no roots after this patch.
>
> I _think_ it should be impossible get here with an invalid root.  Can you try
> adding a prep patch to assert that the root is valid so that this patch can
> reasonably assert that there's no functional change?
>
>
> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
> index 508074e47bc0..fffd9b610196 100644
> --- a/arch/x86/kvm/x86.c
> +++ b/arch/x86/kvm/x86.c
> @@ -792,6 +792,8 @@ void kvm_inject_emulated_page_fault(struct kvm_vcpu *vcpu,
>         fault_mmu = fault->nested_page_fault ? vcpu->arch.mmu :
>                                                vcpu->arch.walk_mmu;
>
> +       WARN_ON_ONCE(!VALID_PAGE(fault_mmu->root.hpa));
> +

I've been updating the patches as per your suggestions.

And I suddenly realized that when fault->nested_page_fault=false
with nested EPT, fault_mmu->root.hpa is always unset.

fault_mmu->root.hpa is just meaningless when fault_mmu is not
vcpu->arch.mmu.

I will add it as one of the reasons for replacing the argument
with KVM_MMU_ROOT_XXX.

>         /*
>          * Invalidate the TLB entry for the faulting address, if it exists,
>          * else the access will fault indefinitely (and to emulate hardware).
>