Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   ebiederm@xmission.com (Eric W. Biederman)
To:     Kees Cook <keescook@chromium.org>
Cc:     Andy Lutomirski <luto@amacapital.net>,
        Will Drewry <wad@chromium.org>, linux-kernel@vger.kernel.org,
        linux-kselftest@vger.kernel.org, linux-hardening@vger.kernel.org
References: <20211103163039.2104830-1-keescook@chromium.org>
        <87lf253x1c.fsf@disp2133> <202111031139.80CE97C532@keescook>
Date:   Wed, 03 Nov 2021 14:17:56 -0500
In-Reply-To: <202111031139.80CE97C532@keescook> (Kees Cook's message of "Wed,
        3 Nov 2021 11:40:23 -0700")
Message-ID: <871r3xyrob.fsf@disp2133>
User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/26.1 (gnu/linux)
MIME-Version: 1.0
Content-Type: text/plain
Subject: Re: [PATCH 0/2] selftests/seccomp: Report event mismatches more clearly
Precedence: bulk

Kees Cook <keescook@chromium.org> writes:

> On Wed, Nov 03, 2021 at 01:37:51PM -0500, Eric W. Biederman wrote:
>> Kees Cook <keescook@chromium.org> writes:
>> 
>> > Hi,
>> >
>> > This expands the seccomp selftests slightly to add additional debug
>> > reporting detail and a new "immediate fatal SIGSYS under tracing" test.
>> > I expect to be taking these via my seccomp tree.
>> 
>> Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>
>> 
>> I am a little fuzzy on the details but I understand what and why
>> you are testing (I broken it).  So this is my 10,000 foot ack.
>
> Thanks! Yeah, and the other tests did catch it, but it was kind of a
> "side effect", so I added the specific "direct" case where it can be
> seen more clearly.

Hey.  Did you happen to understand the bit about racing with sigaction?

As much as I care about not braking ptrace.  What really decided me was
the on SA_IMMUTABLE was closing the race with sigaction changing the
signal handler.  Especially for something like seccomp.

It is a race so probably very fickle to write a test for, but if we can
figure out how to write a reliable test I expect it will be a good idea.
Do you have any ideas?

I am concerned there is some threaded program somewhere using seccomp
that is allowed to call sigaction, can use sigaction to keep from
being killed (before I send the fix to Linus). 

Eric