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;
MIME-Version: 1.0
In-Reply-To: <9f2e445b-15b0-d1fa-832c-f801efc34d03@lge.com>
References: <1526027434-21237-1-git-send-email-byungchul.park@lge.com>
 <3af4cec0-4019-e3ac-77f9-8631252fb6da@lge.com> <20180511161746.GX26088@linux.vnet.ibm.com>
 <20180511224138.GA89902@joelaf.mtv.corp.google.com> <9f2e445b-15b0-d1fa-832c-f801efc34d03@lge.com>
From:   Byungchul Park <max.byungchul.park@gmail.com>
Date:   Mon, 14 May 2018 23:25:10 +0900
Message-ID: <CANrsvRNsKHdVBXMZraNMTzuVoU443uU1F34iVHvxYUKw73esRg@mail.gmail.com>
Subject: Re: [PATCH] rcu: Report a quiescent state when it's exactly in the state
To:     Byungchul Park <byungchul.park@lge.com>
Cc:     Joel Fernandes <joel@joelfernandes.org>,
        "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>,
        jiangshanlai@gmail.com, josh@joshtriplett.org, rostedt@goodmis.org,
        Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
        linux-kernel@vger.kernel.org, kernel-team@lge.com,
        Peter Zijlstra <peterz@infradead.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Mon, May 14, 2018 at 11:59 AM, Byungchul Park <byungchul.park@lge.com> w=
rote:
>
>
> On 2018-05-12 =EC=98=A4=EC=A0=84 7:41, Joel Fernandes wrote:
>>
>> On Fri, May 11, 2018 at 09:17:46AM -0700, Paul E. McKenney wrote:
>>>
>>> On Fri, May 11, 2018 at 09:57:54PM +0900, Byungchul Park wrote:
>>>>
>>>> Hello folks,
>>>>
>>>> I think I wrote the title in a misleading way.
>>>>
>>>> Please change the title to something else such as,
>>>> "rcu: Report a quiescent state when it's in the state" or,
>>>> "rcu: Add points reporting quiescent states where proper" or so on.
>>>>
>>>> On 2018-05-11 =EC=98=A4=ED=9B=84 5:30, Byungchul Park wrote:
>>>>>
>>>>> We expect a quiescent state of TASKS_RCU when
>>>>> cond_resched_tasks_rcu_qs()
>>>>> is called, no matter whether it actually be scheduled or not. However=
,
>>>>> it currently doesn't report the quiescent state when the task enters
>>>>> into __schedule() as it's called with preempt =3D true. So make it re=
port
>>>>> the quiescent state unconditionally when cond_resched_tasks_rcu_qs() =
is
>>>>> called.
>>>>>
>>>>> And in TINY_RCU, even though the quiescent state of rcu_bh also shoul=
d
>>>>> be reported when the tick interrupt comes from user, it doesn't. So
>>>>> make
>>>>> it reported.
>>>>>
>>>>> Lastly in TREE_RCU, rcu_note_voluntary_context_switch() should be
>>>>> reported when the tick interrupt comes from not only user but also
>>>>> idle,
>>>>> as an extended quiescent state.
>>>>>
>>>>> Signed-off-by: Byungchul Park <byungchul.park@lge.com>
>>>>> ---
>>>>>   include/linux/rcupdate.h | 4 ++--
>>>>>   kernel/rcu/tiny.c        | 6 +++---
>>>>>   kernel/rcu/tree.c        | 4 ++--
>>>>>   3 files changed, 7 insertions(+), 7 deletions(-)
>>>>>
>>>>> diff --git a/include/linux/rcupdate.h b/include/linux/rcupdate.h
>>>>> index ee8cf5fc..7432261 100644
>>>>> --- a/include/linux/rcupdate.h
>>>>> +++ b/include/linux/rcupdate.h
>>>>> @@ -195,8 +195,8 @@ static inline void exit_tasks_rcu_finish(void) { =
}
>>>>>    */
>>>>>   #define cond_resched_tasks_rcu_qs() \
>>>>>   do { \
>>>>> -       if (!cond_resched()) \
>>>>> -               rcu_note_voluntary_context_switch_lite(current); \
>>>>> +       rcu_note_voluntary_context_switch_lite(current); \
>>>>> +       cond_resched(); \
>>>
>>>
>>> Ah, good point.
>>>
>>> Peter, I have to ask...  Why is "cond_resched()" considered a preemptio=
n
>>> while "schedule()" is not?
>>
>>
>> Infact something interesting I inferred from the __schedule loop related
>> to
>> your question:
>>
>> switch_count can either be set to prev->invcsw or prev->nvcsw. If we can
>> assume that switch_count reflects whether the context switch is
>> involuntary
>> or voluntary,
>>                        task-running-state       preempt
>> switch_count
>> 0 (running)             1               involuntary
>> 0                       0               involuntary
>> 1                       0               voluntary
>> 1                       1               involuntary
>>
>> According to the above table, both the task's running state and the
>> preempt
>> parameter to __schedule should be used together to determine if the swit=
ch
>> is
>> a voluntary one or not.
>>
>> So this code in rcu_note_context_switch should really be:
>> if (!preempt && !(current->state & TASK_RUNNING))
>>         rcu_note_voluntary_context_switch_lite(current);
>>
>> According to the above table, cond_resched always classifies as an
>> involuntary switch which makes sense to me. Even though cond_resched is
>
>
> Hello guys,
>
> The classification for nivcsw/nvcsw used in scheduler core, Joel, you
> showed us is different from that used in when we distinguish between
> non preemption/voluntary preemption/preemption/full and so on, even
> they use the same word, "voluntary" though.
>
> The name, rcu_note_voluntary_context_switch_lite() used in RCU has
> a lot to do with the latter, the term of preemption. Furthermore, I
> think the function should be called even when calling schedule() for
> sleep as well. I think it would be better to change the function
> name to something else to prevent confusing, it's up to Paul tho. :)

Let me explain more what I did earlier.

In the scheduler core when classifying nivcsw/nvcsw, they classify the
tries as "voluntary", which go to the inactivate state i.e. sleep through
a normal path w/o any disturbed e.g. by interrupt preemption.

However, in RCU, it's for indicating the places trying to explicitly call
scheduler which are quiescent states anyway for TASKS_RCU. Any
explicit tries including voluntary preemption points are the cases.

That 's why I said they have different meaning from each other. But
anyway I also think it would be much better if we can make them
consistent by renaming or modifying both scheduler and rcu code.

>> explicitly called, its still sort of involuntary in the sense its not
>> called
>> into the scheduler for sleeping, but rather for seeing if something else
>> can
>> run instead (a preemption point). Infact none of the task deactivation i=
n
>> the
>> __schedule loop will run if cond_resched is used.
>>
>> I agree that if schedule was called directly but with TASK_RUNNING=3D1, =
then
>> that could probably be classified an involuntary switch too...
>>
>> Also since we're deciding to call rcu_note_voluntary_context_switch_lite
>> unconditionally, then IMO this comment on that macro:
>>
>> /*
>>   * Note a voluntary context switch for RCU-tasks benefit.  This is a
>>   * macro rather than an inline function to avoid #include hell.
>>   */
>>   #ifdef CONFIG_TASKS_RCU
>>   #define rcu_note_voluntary_context_switch_lite(t)
>>
>> Should be changed to:
>>
>> /*
>>   * Note a attempt to perform a voluntary context switch for RCU-tasks
>>   * benefit.  This is called even in situations where a context switch
>>   * didn't really happen even though it was requested. This is a
>>   * macro rather than an inline function to avoid #include hell.
>>   */
>>   #ifdef CONFIG_TASKS_RCU
>>   #define rcu_note_voluntary_context_switch_lite(t)
>>
>> Right?
>>
>> Correct me if I'm wrong about anything, thanks,
>>
>> - Joel
>>
>>
>
> --
> Thanks,
> Byungchul


--=20
Thanks,
Byungchul