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;
MIME-Version: 1.0
References: <20230515130553.2311248-1-jeffxu@chromium.org> <20230515130553.2311248-2-jeffxu@chromium.org>
 <6cb7df56-0479-30be-5389-b4b819572deb@intel.com>
In-Reply-To: <6cb7df56-0479-30be-5389-b4b819572deb@intel.com>
From:   Jeff Xu <jeffxu@google.com>
Date:   Tue, 16 May 2023 16:55:28 -0700
Message-ID: <CALmYWFszHvHa2_DqMPZcWkVqhAtKxrbpJ4Mzg+VWLnPHQD_fsw@mail.gmail.com>
Subject: Re: [PATCH 1/6] PKEY: Introduce PKEY_ENFORCE_API flag
To:     Dave Hansen <dave.hansen@intel.com>
Cc:     jeffxu@chromium.org, luto@kernel.org, jorgelo@chromium.org,
        keescook@chromium.org, groeck@chromium.org, jannh@google.com,
        sroettger@google.com, akpm@linux-foundation.org,
        linux-kernel@vger.kernel.org, linux-kselftest@vger.kernel.org,
        linux-mm@kvack.org, linux-hardening@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Tue, May 16, 2023 at 4:14=E2=80=AFPM Dave Hansen <dave.hansen@intel.com>=
 wrote:
>
> On 5/15/23 06:05, jeffxu@chromium.org wrote:
> > --- a/arch/x86/mm/pkeys.c
> > +++ b/arch/x86/mm/pkeys.c
> > @@ -20,7 +20,7 @@ int __execute_only_pkey(struct mm_struct *mm)
> >       /* Do we need to assign a pkey for mm's execute-only maps? */
> >       if (execute_only_pkey =3D=3D -1) {
> >               /* Go allocate one to use, which might fail */
> > -             execute_only_pkey =3D mm_pkey_alloc(mm);
> > +             execute_only_pkey =3D mm_pkey_alloc(mm, 0);
> >               if (execute_only_pkey < 0)
> >                       return -1;
> >               need_to_set_mm_pkey =3D true;
>
> In your threat model, what mechanism prevents the attacker from
> modifying executable mappings?
>
> I was trying to figure out if the implicit execute-only pkey should have
> the PKEY_ENFORCE_API bit set.  I think that in particular would probably
> cause some kind of ABI breakage, but it still reminded me that I have an
> incomplete picture of the threat model.
Yes. The main reason for not adding it now is the ABI breakage.
As a next step,  we could potentially develop mseal(), which fits more
to the code segment.
The PKEY_ENFORCE_API allows munmap(), so the user case is slightly differen=
t.

I will leave the threat model / V8 specific question to Stephan.