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;
Content-Type: text/plain;
        charset=utf-8
Mime-Version: 1.0 (Mac OS X Mail 13.4 \(3608.80.23.2.2\))
Subject: Re: [PATCH -next v2] locking/osq_lock: annotate a data race in
 osq_lock
From:   Qian Cai <cai@lca.pw>
In-Reply-To: <20200509043308.GL2869@paulmck-ThinkPad-P72>
Date:   Sat, 9 May 2020 09:01:53 -0400
Cc:     Will Deacon <will@kernel.org>, Elver Marco <elver@google.com>,
        LKML <linux-kernel@vger.kernel.org>,
        Ingo Molnar <mingo@redhat.com>,
        "Peter Zijlstra (Intel)" <peterz@infradead.org>
Content-Transfer-Encoding: quoted-printable
Message-Id: <9FBC2A17-2821-49F7-9A0F-5E41116BBC70@lca.pw>
References: <1581429255-12542-1-git-send-email-cai@lca.pw>
 <2C8BF141-5384-475F-B713-6D473557C65C@lca.pw>
 <20200509043308.GL2869@paulmck-ThinkPad-P72>
To:     "Paul E. McKenney" <paulmck@kernel.org>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


> On May 9, 2020, at 12:33 AM, Paul E. McKenney <paulmck@kernel.org> =
wrote:
>=20
> On Fri, May 08, 2020 at 04:59:05PM -0400, Qian Cai wrote:
>>=20
>>=20
>>> On Feb 11, 2020, at 8:54 AM, Qian Cai <cai@lca.pw> wrote:
>>>=20
>>> prev->next could be accessed concurrently as noticed by KCSAN,
>>>=20
>>> write (marked) to 0xffff9d3370dbbe40 of 8 bytes by task 3294 on cpu =
107:
>>> osq_lock+0x25f/0x350
>>> osq_wait_next at kernel/locking/osq_lock.c:79
>>> (inlined by) osq_lock at kernel/locking/osq_lock.c:185
>>> rwsem_optimistic_spin
>>> <snip>
>>>=20
>>> read to 0xffff9d3370dbbe40 of 8 bytes by task 3398 on cpu 100:
>>> osq_lock+0x196/0x350
>>> osq_lock at kernel/locking/osq_lock.c:157
>>> rwsem_optimistic_spin
>>> <snip>
>>>=20
>>> Since the write only stores NULL to prev->next and the read tests if
>>> prev->next equals to this_cpu_ptr(&osq_node). Even if the value is
>>> shattered, the code is still working correctly. Thus, mark it as an
>>> intentional data race using the data_race() macro.
>>>=20
>>> Signed-off-by: Qian Cai <cai@lca.pw>
>>=20
>> Hmm, this patch has been dropped from linux-next from some reasons.
>>=20
>> Paul, can you pick this up along with KCSAN fixes?
>>=20
>> =
https://lore.kernel.org/lkml/1581429255-12542-1-git-send-email-cai@lca.pw/=

>=20
> I have queued it on -rcu, but it is too late for v5.8 via the -rcu =
tree,
> so this is v5.9 at the earliest.  Plus I would need an ack from one of
> the locking folks.

Peter, Will, can you give an ACK? This v2 should incorporate all the =
feedback from Peter,

=
https://lore.kernel.org/lkml/20200211124753.GP14914@hirez.programming.kick=
s-ass.net/

V5.9 is fine. All I care about is it is always in linux-next (so the =
testing bots won=E2=80=99t trigger this over and over again) and to be =
in mainline at some point in the future.

>=20
> 							Thanx, Paul
>=20
>>> ---
>>>=20
>>> v2: insert some code comments.
>>>=20
>>> kernel/locking/osq_lock.c | 6 +++++-
>>> 1 file changed, 5 insertions(+), 1 deletion(-)
>>>=20
>>> diff --git a/kernel/locking/osq_lock.c b/kernel/locking/osq_lock.c
>>> index 1f7734949ac8..f733bcd99e8a 100644
>>> --- a/kernel/locking/osq_lock.c
>>> +++ b/kernel/locking/osq_lock.c
>>> @@ -154,7 +154,11 @@ bool osq_lock(struct optimistic_spin_queue =
*lock)
>>> 	 */
>>>=20
>>> 	for (;;) {
>>> -		if (prev->next =3D=3D node &&
>>> +		/*
>>> +		 * cpu_relax() below implies a compiler barrier which =
would
>>> +		 * prevent this comparison being optimized away.
>>> +		 */
>>> +		if (data_race(prev->next =3D=3D node) &&
>>> 		    cmpxchg(&prev->next, node, NULL) =3D=3D node)
>>> 			break;
>>>=20
>>> --=20
>>> 1.8.3.1