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> <20230515130553.2311248-6-jeffxu@chromium.org>
 <4f14a645-3569-2e3b-f55c-3b17b567845a@intel.com>
In-Reply-To: <4f14a645-3569-2e3b-f55c-3b17b567845a@intel.com>
From:   Jeff Xu <jeffxu@google.com>
Date:   Tue, 16 May 2023 17:08:42 -0700
Message-ID: <CALmYWFvcZ+fGrCjfdx6DWNbYj0WQCJ-i4HFKHgSCqRU6Q4YbBQ@mail.gmail.com>
Subject: Re: [PATCH 5/6] KEY: Apply PKEY_ENFORCE_API to munmap
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 4:24=E2=80=AFPM Dave Hansen <dave.hansen@intel.com>=
 wrote:
>
> On 5/15/23 06:05, jeffxu@chromium.org wrote:
> > From: Jeff Xu <jeffxu@google.com>
> >
> > This patch enables PKEY_ENFORCE_API for the munmap
> > syscall.
>
> The basic problem here is how we know when the set of syscalls that are
> patched here is good enough and how we catch future functionality that
> might need to be captured as well.
>
> This mechanism really needs to be able to defend against *any* changes
> to the address space.  I assume that folks are using syscall filtering
> to prevent new syscalls from causing havoc, but is there anything that
> can be done for, say, things like madvise()?  I bet it was harmless for
> a long time until MADV_DONTNEED showed up and made it able to
> effectively zero memory.

Not any change, just a limited set of syscall from user space.
I think it is reasonable to hope that any kind of syscall ABI change that
affects VMA will get reviewed thoroughly from now on.

Also, if we continue to add mseal() to the kernel, we will have to pay more
attention to syscalls related to VMA.

Thanks
-Jeff Xu