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;
MIME-Version: 1.0
References: <20210701083842.580466-1-elver@google.com> <87h7hdn24k.fsf@disp2133>
In-Reply-To: <87h7hdn24k.fsf@disp2133>
From:   Marco Elver <elver@google.com>
Date:   Fri, 2 Jul 2021 09:20:56 +0200
Message-ID: <CANpmjNMtK53SiZwm0N9VuwGJthY0unZ_1_mZ=gXdMH0_LAFr5A@mail.gmail.com>
Subject: Re: [PATCH v2] perf: Require CAP_KILL if sigtrap is requested
To:     "Eric W. Biederman" <ebiederm@xmission.com>
Cc:     peterz@infradead.org, tglx@linutronix.de, mingo@kernel.org,
        kasan-dev@googlegroups.com, linux-kernel@vger.kernel.org,
        mingo@redhat.com, acme@kernel.org, mark.rutland@arm.com,
        alexander.shishkin@linux.intel.com, jolsa@redhat.com,
        namhyung@kernel.org, linux-perf-users@vger.kernel.org,
        omosnace@redhat.com, serge@hallyn.com,
        linux-security-module@vger.kernel.org, stable@vger.kernel.org,
        Dmitry Vyukov <dvyukov@google.com>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Thu, 1 Jul 2021 at 23:41, Eric W. Biederman <ebiederm@xmission.com> wrote:
>
> Marco Elver <elver@google.com> writes:
>
> > If perf_event_open() is called with another task as target and
> > perf_event_attr::sigtrap is set, and the target task's user does not
> > match the calling user, also require the CAP_KILL capability.
> >
> > Otherwise, with the CAP_PERFMON capability alone it would be possible
> > for a user to send SIGTRAP signals via perf events to another user's
> > tasks. This could potentially result in those tasks being terminated if
> > they cannot handle SIGTRAP signals.
> >
> > Note: The check complements the existing capability check, but is not
> > supposed to supersede the ptrace_may_access() check. At a high level we
> > now have:
> >
> >       capable of CAP_PERFMON and (CAP_KILL if sigtrap)
> >               OR
> >       ptrace_may_access() // also checks for same thread-group and uid
>
> Is there anyway we could have a comment that makes the required
> capability checks clear?
>
> Basically I see an inlined version of kill_ok_by_cred being implemented
> without the comments on why the various pieces make sense.

I'll add more comments. It probably also makes sense to factor the
code here into its own helper.

> Certainly ptrace_may_access(task, PTRACE_MODE_READ_REALCREDS) should not
> be a check to allow writing/changing a task.  It needs to be
> PTRACE_MODE_ATTACH_REALCREDS, like /proc/self/mem uses.

So if attr.sigtrap the checked ptrace mode needs to switch to
PTRACE_MODE_ATTACH_REALCREDS. Otherwise, it is possible to send a
signal if only read-ptrace permissions are granted.

Is my assumption here correct?

> Now in practice I think your patch probably has the proper checks in
> place for sending a signal but it is far from clear.

Thanks,
-- Marco