Received-SPF: pass (google.com: domain of linux-kernel+bounces-188272-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Precedence: bulk
MIME-Version: 1.0
References: <20240524033933.135049-1-jeffxu@google.com> <20240524033933.135049-3-jeffxu@google.com>
In-Reply-To: <20240524033933.135049-3-jeffxu@google.com>
From: Jeff Xu <jeffxu@google.com>
Date: Thu, 23 May 2024 20:42:27 -0700
Message-ID: <CALmYWFs8e4r390v5pepxZ4H6E=x_21jG1vjVYici5pgt5rTGxg@mail.gmail.com>
Subject: Re: [PATCH v2 2/2] memfd:add MEMFD_NOEXEC_SEAL documentation
To: jeffxu@chromium.org
Cc: akpm@linux-foundation.org, cyphar@cyphar.com, dmitry.torokhov@gmail.com, 
	dverkamp@chromium.org, hughd@google.com, jorgelo@chromium.org, 
	keescook@chromium.org, linux-kernel@vger.kernel.org, 
	linux-kselftest@vger.kernel.org, linux-mm@kvack.org, pobrn@protonmail.com, 
	skhan@linuxfoundation.org, stable@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Hi Aleksa

On Thu, May 23, 2024 at 8:39=E2=80=AFPM <jeffxu@chromium.org> wrote:
>
> From: Jeff Xu <jeffxu@google.com>
>
> Add documentation for MFD_NOEXEC_SEAL and MFD_EXEC
>
> Cc: stable@vger.kernel.org
> Signed-off-by: Jeff Xu <jeffxu@google.com>
> ---
>  Documentation/userspace-api/index.rst      |  1 +
>  Documentation/userspace-api/mfd_noexec.rst | 90 ++++++++++++++++++++++
>  2 files changed, 91 insertions(+)
>  create mode 100644 Documentation/userspace-api/mfd_noexec.rst
>
> diff --git a/Documentation/userspace-api/index.rst b/Documentation/usersp=
ace-api/index.rst
> index 5926115ec0ed..8a251d71fa6e 100644
> --- a/Documentation/userspace-api/index.rst
> +++ b/Documentation/userspace-api/index.rst
> @@ -32,6 +32,7 @@ Security-related interfaces
>     seccomp_filter
>     landlock
>     lsm
> +   mfd_noexec
>     spec_ctrl
>     tee
>
> diff --git a/Documentation/userspace-api/mfd_noexec.rst b/Documentation/u=
serspace-api/mfd_noexec.rst
> new file mode 100644
> index 000000000000..6f11ad86b076
> --- /dev/null
> +++ b/Documentation/userspace-api/mfd_noexec.rst
> @@ -0,0 +1,90 @@
> +.. SPDX-License-Identifier: GPL-2.0
> +
> +=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> +Introduction of non executable mfd
> +=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> +:Author:
> +    Daniel Verkamp <dverkamp@chromium.org>
> +    Jeff Xu <jeffxu@google.com>
> +
> +:Contributor:
> +       Aleksa Sarai <cyphar@cyphar.com>
> +       Barnab=C3=A1s P=C5=91cze <pobrn@protonmail.com>
> +       David Rheinsberg <david@readahead.eu>
> +
> +Since Linux introduced the memfd feature, memfd have always had their
> +execute bit set, and the memfd_create() syscall doesn't allow setting
> +it differently.
> +
> +However, in a secure by default system, such as ChromeOS, (where all
> +executables should come from the rootfs, which is protected by Verified
> +boot), this executable nature of memfd opens a door for NoExec bypass
> +and enables =E2=80=9Cconfused deputy attack=E2=80=9D.  E.g, in VRP bug [=
1]: cros_vm
> +process created a memfd to share the content with an external process,
> +however the memfd is overwritten and used for executing arbitrary code
> +and root escalation. [2] lists more VRP in this kind.
> +
> +On the other hand, executable memfd has its legit use, runc uses memfd=
=E2=80=99s
> +seal and executable feature to copy the contents of the binary then
> +execute them, for such system, we need a solution to differentiate runc'=
s
> +use of  executable memfds and an attacker's [3].
> +
> +To address those above.
> + - Let memfd_create() set X bit at creation time.
> + - Let memfd be sealed for modifying X bit when NX is set.
> + - A new pid namespace sysctl: vm.memfd_noexec to help applications to
> +   migrating and enforcing non-executable MFD.
> +
> +User API
> +=3D=3D=3D=3D=3D=3D=3D=3D
> +``int memfd_create(const char *name, unsigned int flags)``
> +
> +``MFD_NOEXEC_SEAL``
> +       When MFD_NOEXEC_SEAL bit is set in the ``flags``, memfd is create=
d
> +       with NX. F_SEAL_EXEC is set and the memfd can't be modified to
> +       add X later.
> +       This is the most common case for the application to use memfd.
> +
> +``MFD_EXEC``
> +       When MFD_EXEC bit is set in the ``flags``, memfd is created with =
X.
> +
> +Note:
> +       ``MFD_NOEXEC_SEAL`` and ``MFD_EXEC`` doesn't change the sealable
> +       characteristic of memfd, which is controlled by ``MFD_ALLOW_SEALI=
NG``.
> +
> +
> +Sysctl:
> +=3D=3D=3D=3D=3D=3D=3D=3D
> +``pid namespaced sysctl vm.memfd_noexec``
> +
> +The new pid namespaced sysctl vm.memfd_noexec has 3 values:
> +
> + - 0: MEMFD_NOEXEC_SCOPE_EXEC
> +       memfd_create() without MFD_EXEC nor MFD_NOEXEC_SEAL acts like
> +       MFD_EXEC was set.
> +
> + - 1: MEMFD_NOEXEC_SCOPE_NOEXEC_SEAL
> +       memfd_create() without MFD_EXEC nor MFD_NOEXEC_SEAL acts like
> +       MFD_NOEXEC_SEAL was set.
> +
> + - 2: MEMFD_NOEXEC_SCOPE_NOEXEC_ENFORCED
> +       memfd_create() without MFD_NOEXEC_SEAL will be rejected.
> +
> +The sysctl allows finer control of memfd_create for old-software that
> +doesn't set the executable bit, for example, a container with
> +vm.memfd_noexec=3D1 means the old-software will create non-executable me=
mfd
> +by default while new-software can create executable memfd by setting
> +MFD_EXEC.
> +
> +The value of memfd_noexec is passed to child namespace at creation time,
> +in addition, the setting is hierarchical, i.e. during memfd_create,
> +we will search from current ns to root ns and use the most restrictive
> +setting.
> +

Can you please help to review the sysctl part to check if  I captured
your change correctly ?

Thanks
-Jeff


> +Reference:
> +=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> +[1] https://crbug.com/1305267
> +
> +[2] https://bugs.chromium.org/p/chromium/issues/list?q=3Dtype%3Dbug-secu=
rity%20memfd%20escalation&can=3D1
> +
> +[3] https://lwn.net/Articles/781013/
> --
> 2.45.1.288.g0e0cd299f1-goog
>