Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Mon, 18 Jun 2018 18:36:06 -0700
User-Agent: K-9 Mail for Android
In-Reply-To: <20180619010803.GC3593@linux.vnet.ibm.com>
References: <20180610234544.150933-2-joel@joelfernandes.org> <20180613065711.GZ16472@yexl-desktop> <20180613072948.GA142000@joelaf.mtv.corp.google.com> <20180618165646.GA3593@linux.vnet.ibm.com> <20180618222647.GA156172@joelaf.mtv.corp.google.com> <20180619010803.GC3593@linux.vnet.ibm.com>
MIME-Version: 1.0
Content-Type: text/plain;
 charset=utf-8
Content-Transfer-Encoding: quoted-printable
Subject: Re: [lkp-robot] [rcutorture]  46e26223e3: WARNING:at_kernel/rcu/rcutorture.c:#rcu_torture_stats_print
To:     paulmck@linux.vnet.ibm.com,
        "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>,
        Joel Fernandes <joel@joelfernandes.org>
CC:     kernel test robot <xiaolong.ye@intel.com>,
        linux-kernel@vger.kernel.org, kernel-team@android.com,
        Josh Triplett <josh@joshtriplett.org>,
        Lai Jiangshan <jiangshanlai@gmail.com>,
        Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
        mingo@redhat.com, Steven Rostedt <rostedt@goodmis.org>,
        tglx@linutronix.de, lkp@01.org
From:   Joel Fernandes <joelaf@google.com>
Message-ID: <2CDE3A5B-1AD3-4185-9E24-24D4075A9EA6@google.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


On June 18, 2018 6:08:03 PM PDT, "Paul E=2E McKenney" <paulmck@linux=2Evne=
t=2Eibm=2Ecom> wrote:
>On Mon, Jun 18, 2018 at 03:26:47PM -0700, Joel Fernandes wrote:
>> On Mon, Jun 18, 2018 at 09:56:46AM -0700, Paul E=2E McKenney wrote:
>> > > The reason for the rcutorture test failure could be that the
>default
>> > > kthread_prio for the system's RCU threads is set to 1 (unless
>overridden by
>> > > rcutree=2Ekthread_prio) which is also equal to the priority of the
>rcutorture's
>> > > boost threads=2E Due to this the rcutorture test could starve the
>RCU threads
>> > > as well and defeat the boosting mechanism=2E I was able to solve a
>similar
>> > > issue by just passing rcutree=2Ekthread_prio of 50 on the kernel
>command line=2E
>> > >=20
>> > > Paul, would it be ok if we changed the default kthread_prio to
>something > 1
>> > > so that rcutorture can test properly without needing to pass any
>extra
>> > > rcutree=2E* parameters?
>> > >=20
>> > > so something like this in kernel/rcu/tree=2Ec ?
>> > >=20
>> > > static int kthread_prio =3D IS_ENABLED(CONFIG_RCU_BOOST) ? 2 : 0;
>> >=20
>> > Would it be possible to also condition this on rcutorture being
>built
>> > in?  Or are they doing modprobes for rcutorture?
>>=20
>> They seem to be doing built-in rcutorture tests=2E But I believe the
>same
>> problem would occur even if you used modules? I believe the fact that
>> rcutorture is a module or built-in wouldn't matter to the underlying
>issue
>> which is the RCU subsystems's threads are at too low of a priority
>> (rcutree=2Ekthread_prio =3D 1)=2E
>
>Understood=2E=2E=2E
>
>> If you agree with changing the default priority, I have included a
>patch
>> below for rcu/dev=2E
>
>The problem is that without rcutorture, rcutree=2Ekthread_prio=3D1 is a
>legitimate choice, and changing the default globally could be breaking
>someone=2E  So it would be far better to up the priority only during
>known
>rcutorture testing=2E

Oh I see what you're saying=2E I'll work on a patch along these lines then=
=2E Thanks!

- Joel


>
>							Thanx, Paul
>
>> thanks,
>>=20
>>  - Joel
>>=20
>> ---8<-----------------------
>>=20
>> >From b0f4111ef1abd1c481c269fadb3535c83ab43c93 Mon Sep 17 00:00:00
>2001
>> From: "Joel Fernandes (Google)" <joel@joelfernandes=2Eorg>
>> Date: Mon, 18 Jun 2018 15:13:10 -0700
>> Subject: [PATCH] rcu: Change default RCU kthread priority to 2
>>=20
>> The current RT priority of 1 for RCU kthreads makes rcutorture's
>boost test
>> fail on systems where rcutree=2Ekthread_prio isn't passed=2E
>>=20
>> The rcutorture boost kthreads have the same priority as well (RT
>priority of
>> 1)=2E Due to this, the rcutorture kthreads starve the RCU subsystem's
>kthreads
>> and causes rcutorture failures=2E This patch changes the priority of
>the RCU
>> subsystem's threads to a default RT priority of 2 so that
>rcutorture's
>> threads get preempted by them=2E Verified that the boost tests will
>pass with
>> this change=2E
>>=20
>> Reported-by: Xiaolong Ye (via lkp-robot) <xiaolong=2Eye@intel=2Ecom>
>> Signed-off-by: Joel Fernandes (Google) <joel@joelfernandes=2Eorg>
>> ---
>>  kernel/rcu/tree=2Ec | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>=20
>> diff --git a/kernel/rcu/tree=2Ec b/kernel/rcu/tree=2Ec
>> index deb2508be923=2E=2E920c39e3f871 100644
>> --- a/kernel/rcu/tree=2Ec
>> +++ b/kernel/rcu/tree=2Ec
>> @@ -171,7 +171,7 @@ static void rcu_report_exp_rdp(struct rcu_state
>*rsp,
>>  static void sync_sched_exp_online_cleanup(int cpu);
>>=20
>>  /* rcuc/rcub kthread realtime priority */
>> -static int kthread_prio =3D IS_ENABLED(CONFIG_RCU_BOOST) ? 1 : 0;
>> +static int kthread_prio =3D IS_ENABLED(CONFIG_RCU_BOOST) ? 2 : 0;
>>  module_param(kthread_prio, int, 0644);
>>=20
>>  /* Delay in jiffies for grace-period initialization delays, debug
>only=2E */
>> --=20
>> 2=2E18=2E0=2Erc1=2E244=2Egcf134e6275-goog
>>=20

--=20
Sent from my Android device with K-9 Mail=2E Please excuse my brevity=2E