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: <CAABZP2xJRGhPmfB-PrfesQKzP7fsuZsj+3TewAiLLW8u=YK4dg@mail.gmail.com>
 <CAEXW_YSSGYgqTpxqbYikCFS9t=2f+L-0phbU+gAAngB5z-FbyA@mail.gmail.com>
 <ZEXOMC2casTlobE1@boqun-archlinux> <87fs8pzalj.fsf@mail.concordia>
 <20230424151351.GP19790@gate.crashing.org> <ZEagN1jJwg+rUzX4@boqun-archlinux>
 <CAEXW_YRfetnhgCw5OgnwhgZF_U+UkHN=uy=L8ovGLqn1UCtfTg@mail.gmail.com>
 <20230425101324.GD1331236@hirez.programming.kicks-ass.net>
 <CAABZP2ypJ98T3XAqPnLrxxzrYckSQ6sn3woEmpigQ+cRRaw=Zw@mail.gmail.com>
 <CAEXW_YQEarLt7YGQZdwmcSyZcGRCGKf89ovxjQdXBO-TgXAk-w@mail.gmail.com> <877ctyzv08.fsf@mail.concordia>
In-Reply-To: <877ctyzv08.fsf@mail.concordia>
From:   Joel Fernandes <joel@joelfernandes.org>
Date:   Wed, 26 Apr 2023 09:44:22 -0400
Message-ID: <CAEXW_YR4rmNY5JF-077EUpriF=PP9uML+_WMDSVi-L3G6_FOdw@mail.gmail.com>
Subject: Re: BUG : PowerPC RCU: torture test failed with __stack_chk_fail
To:     Michael Ellerman <mpe@ellerman.id.au>
Cc:     Zhouyi Zhou <zhouzhouyi@gmail.com>,
        Christophe Leroy <christophe.leroy@c-s.fr>,
        Peter Zijlstra <peterz@infradead.org>,
        Boqun Feng <boqun.feng@gmail.com>,
        Segher Boessenkool <segher@kernel.crashing.org>,
        linuxppc-dev <linuxppc-dev@lists.ozlabs.org>,
        rcu <rcu@vger.kernel.org>,
        linux-kernel <linux-kernel@vger.kernel.org>, lance@osuosl.org,
        "Paul E. McKenney" <paulmck@kernel.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Wed, Apr 26, 2023 at 8:30=E2=80=AFAM Michael Ellerman <mpe@ellerman.id.a=
u> wrote:
>
> Joel Fernandes <joel@joelfernandes.org> writes:
> > On Tue, Apr 25, 2023 at 6:58=E2=80=AFAM Zhouyi Zhou <zhouzhouyi@gmail.c=
om> wrote:
> ...
> >
> > Out of curiosity for PPC folks, why cannot 64-bit PPC use per-task
> > canary? Michael, is this an optimization? Adding Christophe as well
> > since it came in a few years ago via the following commit:
>
> I think Christophe also answered these in his reply.
>
> We do use a per-task canary, but because we don't have "current" in a
> register, we can't use the value in current for GCC.
>
> In one of my replies I said a possible solution would be to keep current
> in a register on 64-bit, but we'd need to do that in addition to the
> paca, so that would consume another GPR which we'd need to think hard
> about.

Makes sense. I'd think it is not worth allocating a separate GPR just
for this, and may cause similar register optimization issues as well.

> There's another reason to have it in the paca, which is that the paca is
> always accessible, even when the MMU is off, whereas current isn't (in
> some situations).
>
> In general we don't want to use stack protector in code that runs with
> the MMU off, but if the canary wasn't in the paca then we'd have a hard
> requirement to not use stack protector in that code.

How could you control which code paths don't have the stack protector?
Just curious.

thanks,

 - Joel