linux-kernel - Re: [PATCH 10/23] KVM: MMU: split cpu_role from 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:   Mon, 7 Feb 2022 13:38:48 -0800
From:   David Matlack <dmatlack@...gle.com>
To:     Paolo Bonzini <pbonzini@...hat.com>
Cc:     LKML <linux-kernel@...r.kernel.org>,
        kvm list <kvm@...r.kernel.org>,
        Sean Christopherson <seanjc@...gle.com>,
        Vitaly Kuznetsov <vkuznets@...hat.com>
Subject: Re: [PATCH 10/23] KVM: MMU: split cpu_role from mmu_role

On Sat, Feb 5, 2022 at 6:49 AM Paolo Bonzini <pbonzini@...hat.com> wrote:
>
> On 2/4/22 22:57, David Matlack wrote:
> >> +    vcpu->arch.root_mmu.cpu_role.base.level = 0;
> >> +    vcpu->arch.guest_mmu.cpu_role.base.level = 0;
> >> +    vcpu->arch.nested_mmu.cpu_role.base.level = 0;
> > Will cpu_role.base.level already be 0 if CR0.PG=0 && !tdp_enabled? i.e.
> > setting cpu_role.base.level to 0 might not have the desired effect.
> >
> > It might not matter in practice since the shadow_mmu_init_context() and
> > kvm_calc_mmu_role_common() check both the mmu_role and cpu_role, but does
> > make this reset code confusing.
> >
>
> Good point.  The (still unrealized) purpose of this series is to be able
> to check mmu_role only, so for now I'll just keep the valid bit in the
> ext part of the cpu_role.  The mmu_role's level however is never zero,
> so I can already use the level when I remove the ext part from the mmu_role.

Agreed.

>
> I'll remove the valid bit of the ext part only after the cpu_role check
> is removed, because then it can trivially go.

Ok sounds good.

>
> Paolo
>