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: <20220210025321.787113-1-keescook@chromium.org>
 <871r0a8u29.fsf@email.froward.int.ebiederm.org> <202202101033.9C04563D9@keescook>
 <87pmnu5z28.fsf@email.froward.int.ebiederm.org> <202202101137.B48D02138@keescook>
 <87k0e249tt.fsf@email.froward.int.ebiederm.org> <202202101710.668EDCDC@keescook>
 <875ypm41kb.fsf@email.froward.int.ebiederm.org> <202202101827.4B16DF54@keescook>
 <CAP145phAg3ZSPJw7x2kKVQe86puy-XyKatVoByVoM27RP4aw_g@mail.gmail.com> <87a6ex1ek2.fsf@email.froward.int.ebiederm.org>
In-Reply-To: <87a6ex1ek2.fsf@email.froward.int.ebiederm.org>
From:   =?UTF-8?B?Um9iZXJ0IMWad2nEmWNraQ==?= <robert@swiecki.net>
Date:   Fri, 11 Feb 2022 19:57:02 +0100
Message-ID: <CAP145pii+zCyoQBw778P-Jw9YSOWhQEgB7OTHUaRwfGy7BYWCA@mail.gmail.com>
Subject: Re: [PATCH 0/3] signal: HANDLER_EXIT should clear SIGNAL_UNKILLABLE
To:     "Eric W. Biederman" <ebiederm@xmission.com>
Cc:     Kees Cook <keescook@chromium.org>,
        Andy Lutomirski <luto@amacapital.net>,
        Will Drewry <wad@chromium.org>, linux-kernel@vger.kernel.org,
        linux-hardening@vger.kernel.org
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

pt., 11 lut 2022 o 18:47 Eric W. Biederman <ebiederm@xmission.com> napisa=
=C5=82(a):

> > I think it'd be good to have some way of doing it from the perspective
> > of a parent process - it'd simplify development of sandboxing managers
> > (eg nsjail, minijail, firejail), and creation of good seccomp
> > policies.
>
> By development do you mean debugging sandbox managers?  Or do you mean
> something that sandbox managers can use on a routine basis?

On a routine basis for 2 purposes

a). development of seccomp policies - the manager (regular parent of
sandboxed process) can tell which syscall (and arguments) failed and
it can be then added to policy (though, 'strace -f -c' might be a
better option here)
b). in case of actual seccomp SIGSYS kill, it could then inform users
about what exactly and where happened (syscall no, cpu arch,
arguments, maybe also eip + stack ptr)

But, I don't want to derail the current bug fixing effort, so I just
wanted to mention all of this quickly, and maybe we can follow up on
this in the future.