linux-kernel - Re: [PATCH] KVM: x86/mmu: Do not create SPTEs for GFNs that exceed host.MAXPHYADDR

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: <YmwL87h6klEC4UKV@google.com>
Date:   Fri, 29 Apr 2022 16:01:55 +0000
From:   Sean Christopherson <seanjc@...gle.com>
To:     Paolo Bonzini <pbonzini@...hat.com>
Cc:     Vitaly Kuznetsov <vkuznets@...hat.com>,
        Wanpeng Li <wanpengli@...cent.com>,
        Jim Mattson <jmattson@...gle.com>,
        Joerg Roedel <joro@...tes.org>, kvm@...r.kernel.org,
        linux-kernel@...r.kernel.org, Maxim Levitsky <mlevitsk@...hat.com>,
        Ben Gardon <bgardon@...gle.com>,
        David Matlack <dmatlack@...gle.com>
Subject: Re: [PATCH] KVM: x86/mmu: Do not create SPTEs for GFNs that exceed
 host.MAXPHYADDR

On Fri, Apr 29, 2022, Paolo Bonzini wrote:
> On 4/29/22 16:42, Sean Christopherson wrote:
> > On Fri, Apr 29, 2022, Paolo Bonzini wrote:
> > > On 4/29/22 16:24, Sean Christopherson wrote:
> > > > I don't love the divergent memslot behavior, but it's technically correct, so I
> > > > can't really argue.  Do we want to "officially" document the memslot behavior?
> > > > 
> > > 
> > > I don't know what you mean by officially document,
> > 
> > Something in kvm/api.rst under KVM_SET_USER_MEMORY_REGION.
> 
> Not sure if the API documentation is the best place because userspace does
> not know whether shadow paging is on (except indirectly through other
> capabilities, perhaps)?

Hrm, true, it's not like the userspace VMM can rewrite itself at runtime.

> It could even be programmatic, such as returning 52 for CPUID[0x80000008].
> A nested KVM on L1 would not be able to use the #PF(RSVD) trick to detect
> MMIO faults.  That's not a big price to pay, however I'm not sure it's a
> good idea in general...

Agreed, messing with CPUID is likely to end in tears.