Re: [RFC PATCH 1/1] rcu: Use rcu_gp_kthread_wake() to wake up kthreads

[Pranith Kumar]

On Fri, Jul 25, 2014 at 10:44 AM, Paul E. McKenney
<paulmck@xxxxxxxxxxxxxxxxxx> wrote:
> On Fri, Jul 25, 2014 at 01:06:58AM -0400, Pranith Kumar wrote:
>> The rcu_gp_kthread_wake() function checks for three conditions before waking up
>> grace period kthreads:
>>
>> *  Is the thread we are trying to wake up the current thread?
>> *  Are the gp_flags zero? (all threads wait on non-zero gp_flags condition)
>> *  Is there no thread created for this flavour, hence nothing to wake up?
>>
>> If any one of these condition is true, we do not call wake_up().
>>
>> In rcu_report_qs_rsp(), I added a pr_info() call testing if any of the above
>> conditions is true, in which case we can avoid calling wake_up(). It turns out
>> that quite a few actually are. Most of the cases where we can avoid is condition 2
>> above and condition 1 also occurs quite often. Condition 3 never happens.
>>
>> I could not test the wake_up() in force_quiescent_state() as that is not
>> triggered trivially, but I am assuming we can replace wake_up() there too.
>>
>> Hence this commit tries to avoid calling wake_up() whenever we can by using
>> rcu_gp_kthread_wake() function.
>
> This one does sound much more plausible than the earlier one.  I have
> a few more questions that I will ask in your follow-up message.
>
>> One concern is the comment which states that we need a memory barrier at that
>> location which is being implied by the wake_up(). Should we put an smp_mb() and
>> just not rely on the barrier provided by wake_up()? Thoughts?
>
> Let's see...  The memory barriers are unnecessary for your case 1
> and case 3.  That leaves your case 2, which is all about ->gp_flags.
> It is quite possible that this case is now fully covered by locking,
> so that the comment is obsolete.  But why don't you check?
>

I checked all the locations where gp_flags is being updated and the
root node lock is held in all the cases.
So I guess we can remove the comment too.


>> Signed-off-by: Pranith Kumar <bobby.prani@xxxxxxxxx>
>> ---
>>  kernel/rcu/tree.c | 6 ++++--
>>  1 file changed, 4 insertions(+), 2 deletions(-)
>>
>> diff --git a/kernel/rcu/tree.c b/kernel/rcu/tree.c
>> index 72e0b1f..d0e0d6e 100644
>> --- a/kernel/rcu/tree.c
>> +++ b/kernel/rcu/tree.c
>> @@ -1938,7 +1938,8 @@ static void rcu_report_qs_rsp(struct rcu_state *rsp, unsigned long flags)
>>  {
>>       WARN_ON_ONCE(!rcu_gp_in_progress(rsp));
>>       raw_spin_unlock_irqrestore(&rcu_get_root(rsp)->lock, flags);
>> -     wake_up(&rsp->gp_wq);  /* Memory barrier implied by wake_up() path. */
>> +     /* Memory barrier implied by wake_up() path. */
>> +     rcu_gp_kthread_wake(rsp);
>>  }
>>
>>  /*
>> @@ -2516,7 +2517,8 @@ static void force_quiescent_state(struct rcu_state *rsp)
>>       ACCESS_ONCE(rsp->gp_flags) =
>>               ACCESS_ONCE(rsp->gp_flags) | RCU_GP_FLAG_FQS;
>>       raw_spin_unlock_irqrestore(&rnp_old->lock, flags);
>> -     wake_up(&rsp->gp_wq);  /* Memory barrier implied by wake_up() path. */
>> +     /* Memory barrier implied by wake_up() path. */
>> +     rcu_gp_kthread_wake(rsp);
>>  }
>>
>>  /*
>> --
>> 2.0.1
>>
>



-- 
Pranith
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/