Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Fri, 16 Jun 2023 06:56:17 -0700
In-Reply-To: <001aa2ed-2f78-4361-451d-e31a4d4abaa0@semihalf.com>
Mime-Version: 1.0
References: <20230612164727.3935657-1-carlos.bilbao@amd.com>
 <ZIihRqZljMaMRGcK@google.com> <DM8PR11MB5750D16B08B769173DCD740AE75AA@DM8PR11MB5750.namprd11.prod.outlook.com>
 <ZInLlkXsgnej9ZpT@google.com> <001aa2ed-2f78-4361-451d-e31a4d4abaa0@semihalf.com>
Message-ID: <ZIxqAXhy1tCVpzz7@google.com>
Subject: Re: [PATCH v2] docs: security: Confidential computing intro and
 threat model for x86 virtualization
From:   Sean Christopherson <seanjc@google.com>
To:     Dmytro Maluka <dmy@semihalf.com>
Cc:     Elena Reshetova <elena.reshetova@intel.com>,
        Carlos Bilbao <carlos.bilbao@amd.com>,
        Jason CJ Chen <jason.cj.chen@intel.com>,
        "corbet@lwn.net" <corbet@lwn.net>,
        "linux-doc@vger.kernel.org" <linux-doc@vger.kernel.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        "ardb@kernel.org" <ardb@kernel.org>,
        "kraxel@redhat.com" <kraxel@redhat.com>,
        "dovmurik@linux.ibm.com" <dovmurik@linux.ibm.com>,
        "dave.hansen@linux.intel.com" <dave.hansen@linux.intel.com>,
        "Dhaval.Giani@amd.com" <Dhaval.Giani@amd.com>,
        "michael.day@amd.com" <michael.day@amd.com>,
        "pavankumar.paluri@amd.com" <pavankumar.paluri@amd.com>,
        "David.Kaplan@amd.com" <David.Kaplan@amd.com>,
        "Reshma.Lal@amd.com" <Reshma.Lal@amd.com>,
        "Jeremy.Powell@amd.com" <Jeremy.Powell@amd.com>,
        "sathyanarayanan.kuppuswamy@linux.intel.com" 
        <sathyanarayanan.kuppuswamy@linux.intel.com>,
        "alexander.shishkin@linux.intel.com" 
        <alexander.shishkin@linux.intel.com>,
        "thomas.lendacky@amd.com" <thomas.lendacky@amd.com>,
        "tglx@linutronix.de" <tglx@linutronix.de>,
        "dgilbert@redhat.com" <dgilbert@redhat.com>,
        "gregkh@linuxfoundation.org" <gregkh@linuxfoundation.org>,
        "dinechin@redhat.com" <dinechin@redhat.com>,
        "linux-coco@lists.linux.dev" <linux-coco@lists.linux.dev>,
        "berrange@redhat.com" <berrange@redhat.com>,
        "mst@redhat.com" <mst@redhat.com>, "tytso@mit.edu" <tytso@mit.edu>,
        "jikos@kernel.org" <jikos@kernel.org>,
        "joro@8bytes.org" <joro@8bytes.org>,
        "leon@kernel.org" <leon@kernel.org>,
        "richard.weinberger@gmail.com" <richard.weinberger@gmail.com>,
        "lukas@wunner.de" <lukas@wunner.de>,
        "jejb@linux.ibm.com" <jejb@linux.ibm.com>,
        "cdupontd@redhat.com" <cdupontd@redhat.com>,
        "jasowang@redhat.com" <jasowang@redhat.com>,
        "sameo@rivosinc.com" <sameo@rivosinc.com>,
        "bp@alien8.de" <bp@alien8.de>,
        "security@kernel.org" <security@kernel.org>,
        Larry Dewey <larry.dewey@amd.com>, android-kvm@google.com,
        Dmitry Torokhov <dtor@google.com>,
        Allen Webb <allenwebb@google.com>,
        Tomasz Nowicki <tn@semihalf.com>,
        Grzegorz Jaszczyk <jaz@semihalf.com>,
        Patryk Duda <pdk@semihalf.com>
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Fri, Jun 16, 2023, Dmytro Maluka wrote:
> On 6/14/23 16:15, Sean Christopherson wrote:
> > On Wed, Jun 14, 2023, Elena Reshetova wrote:
> >> Not having a network access requirement doesn=E2=80=99t implicitly inv=
alidate the=20
> >> separation guarantees between the host and guest, it just makes it eas=
ier
> >> since you have one interface less between the host and guest.
> >=20
> > My point is that if the protected guest doesn't need any I/O beyond the=
 hardware
> > device that it accesses, then the threat model is different because man=
y of the
> > new/novel attack surfaces that come with the TDX/SNP threat model don't=
 exist.
> > E.g. the hardening that people want to do for VirtIO drivers may not be=
 at all
> > relevant to pKVM.

...

> But I think I get what you mean: there is no data transfer whereby the
> host is not an endpoint but an intermediary between the guest and some
> device. In simple words, things like virtio-net or virtio-blk are out of
> scope. Yes, I think that's correct for pKVM-on-x86 use cases (and I
> suppose it is correct for pKVM-on-ARM use cases as well). I guess it
> means that "guest data attacks" may not be relevant to pKVM, and perhaps
> this makes its threat model substantially different from cloud use
> cases.

Yes. =20

> >>>> +This new type of adversary may be viewed as a more powerful type
> >>>> +of external attacker, as it resides locally on the same physical ma=
chine
> >>>> +-in contrast to a remote network attacker- and has control over the=
 guest
> >>>> +kernel communication with most of the HW::
> >>>
> >>> IIUC, this last statement doesn't hold true for the pKVM on x86 use c=
ase, which
> >>> specifically aims to give a "guest" exclusive access to hardware reso=
urces.
> >>
> >> Does it hold for *all* HW resources? If yes, indeed this would make pK=
VM on
> >> x86 considerably different.
> >=20
> > Heh, the original says "most", so it doesn't have to hold for all hardw=
are resources,
> > just a simple majority.
>=20
> Again, pedantic mode on, I find it difficult to agree with the wording
> that the guest owns "most of" the HW resources it uses. It controls the
> data communication with its hardware device, but other resources (e.g.
> CPU time, interrupts, timers, PCI config space, ACPI) are owned by the
> host and virtualized by it for the guest.

I wasn't saying that the guest owns most resources, I was saying that the *=
untrusted*
host does *not* own most resources that are exposed to the guest.  My under=
standing
is that everything in your list is owned by the trusted hypervisor in the p=
KVM model.

What I was pointing out is related to the above discussion about the guest =
needing
access to hardware that is effectively owned by the untrusted host, e.g. ne=
twork
access.