[PATCH] sched/fair: Invoke cpufreq hooks for CONFIG_SMP unset

[PATCH] sched/fair: Invoke cpufreq hooks for CONFIG_SMP unset

[Rafael J. Wysocki]

From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

Commit 34e2c555f3e1 (cpufreq: Add mechanism for registering utilization
update callbacks) overlooked the fact that update_load_avg(), where
CFS invokes cpufreq utilization update hooks, becomes an empty stub for
CONFIG_SMP unset.  In consequence, if CONFIG_SMP is not set, cpufreq
governors are never invoked from CFS and they do not have a chance to
evaluate CPU performace levels and update them often enough.  Needless
to say, things don't work as expected then.

Fix the problem by making the !CONFIG_SMP stub of update_load_avg()
invoke cpufreq update hooks too.

Fixes: 34e2c555f3e1 (cpufreq: Add mechanism for registering utilization update callbacks)
Reported-by: Steve Muckle <steve.muckle@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
---

This needs to go into v4.6.

---
 kernel/sched/fair.c |    9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

Index: linux-pm/kernel/sched/fair.c
===================================================================
--- linux-pm.orig/kernel/sched/fair.c
+++ linux-pm/kernel/sched/fair.c
@@ -3030,7 +3030,14 @@ static int idle_balance(struct rq *this_
 
 #else /* CONFIG_SMP */
 
-static inline void update_load_avg(struct sched_entity *se, int update_tg) {}
+static inline void update_load_avg(struct sched_entity *se, int not_used)
+{
+	struct cfs_rq *cfs_rq = cfs_rq_of(se);
+	struct rq *rq = rq_of(cfs_rq);
+
+	cpufreq_trigger_update(rq_clock(rq));
+}
+
 static inline void
 enqueue_entity_load_avg(struct cfs_rq *cfs_rq, struct sched_entity *se) {}
 static inline void

Re: [PATCH] sched/fair: Invoke cpufreq hooks for CONFIG_SMP unset

[Steve Muckle]

Looks good to me.

Also re-tested with intel_pstate on i7-3630QM !SMP, confirmed issue
is resolved. I didn't retest with ondemand because for some reason
that wasn't showing the problem before.

Re: [PATCH] sched/fair: Invoke cpufreq hooks for CONFIG_SMP unset

[Rafael J. Wysocki]

On Friday, May 06, 2016 12:08:25 PM Steve Muckle wrote:
> Looks good to me.
> 
> Also re-tested with intel_pstate on i7-3630QM !SMP, confirmed issue
> is resolved.

Thanks!

Re: [PATCH] sched/fair: Invoke cpufreq hooks for CONFIG_SMP unset

[Peter Zijlstra]

On Fri, May 06, 2016 at 02:58:43PM +0200, Rafael J. Wysocki wrote:
> From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> 
> Commit 34e2c555f3e1 (cpufreq: Add mechanism for registering utilization
> update callbacks) overlooked the fact that update_load_avg(), where
> CFS invokes cpufreq utilization update hooks, becomes an empty stub for
> CONFIG_SMP unset.  In consequence, if CONFIG_SMP is not set, cpufreq
> governors are never invoked from CFS and they do not have a chance to
> evaluate CPU performace levels and update them often enough.  Needless
> to say, things don't work as expected then.
> 
> Fix the problem by making the !CONFIG_SMP stub of update_load_avg()
> invoke cpufreq update hooks too.
> 
> Fixes: 34e2c555f3e1 (cpufreq: Add mechanism for registering utilization update callbacks)
> Reported-by: Steve Muckle <steve.muckle@linaro.org>
> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

Rafael; feel free to push this through the pm tree.

Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>

Re: [PATCH] sched/fair: Invoke cpufreq hooks for CONFIG_SMP unset

[Ingo Molnar]

* Peter Zijlstra <peterz@infradead.org> wrote:

> On Fri, May 06, 2016 at 02:58:43PM +0200, Rafael J. Wysocki wrote:
> > From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> > 
> > Commit 34e2c555f3e1 (cpufreq: Add mechanism for registering utilization
> > update callbacks) overlooked the fact that update_load_avg(), where
> > CFS invokes cpufreq utilization update hooks, becomes an empty stub for
> > CONFIG_SMP unset.  In consequence, if CONFIG_SMP is not set, cpufreq
> > governors are never invoked from CFS and they do not have a chance to
> > evaluate CPU performace levels and update them often enough.  Needless
> > to say, things don't work as expected then.
> > 
> > Fix the problem by making the !CONFIG_SMP stub of update_load_avg()
> > invoke cpufreq update hooks too.
> > 
> > Fixes: 34e2c555f3e1 (cpufreq: Add mechanism for registering utilization update callbacks)
> > Reported-by: Steve Muckle <steve.muckle@linaro.org>
> > Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> 
> Rafael; feel free to push this through the pm tree.
> 
> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>

I already have it in sched/urgent.

Thanks,

	Ingo

[tip:sched/urgent] sched/fair: Fix !CONFIG_SMP kernel cpufreq governor breakage

[tip-bot for Rafael J. Wysocki]

Commit-ID:  536bd00cdbb7b908573e5a93bae67b64cbae60d8
Gitweb:     http://git.kernel.org/tip/536bd00cdbb7b908573e5a93bae67b64cbae60d8
Author:     Rafael J. Wysocki <rafael.j.wysocki@intel.com>
AuthorDate: Fri, 6 May 2016 14:58:43 +0200
Committer:  Ingo Molnar <mingo@kernel.org>
CommitDate: Sat, 7 May 2016 07:45:34 +0200

sched/fair: Fix !CONFIG_SMP kernel cpufreq governor breakage

The following commit:

  34e2c555f3e1 ("cpufreq: Add mechanism for registering utilization update callbacks")

overlooked the fact that update_load_avg(), where CFS invokes cpufreq
utilization update callbacks, becomes an empty stub on UP kernels.

In consequence, if !CONFIG_SMP, cpufreq governors are never invoked
from CFS and they do not have a chance to evaluate CPU performace
levels and update them often enough.

Needless to say, things don't work as expected then.

Fix the problem by making the !CONFIG_SMP stub of update_load_avg()
invoke cpufreq update callbacks too.

Reported-by: Steve Muckle <steve.muckle@linaro.org>
Tested-by: Steve Muckle <steve.muckle@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Acked-by: Steve Muckle <steve.muckle@linaro.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Linux PM list <linux-pm@vger.kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Fixes: 34e2c555f3e1 (cpufreq: Add mechanism for registering utilization update callbacks)
Link: http://lkml.kernel.org/r/6282396.VVEdgVYxO3@vostro.rjw.lan
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
 kernel/sched/fair.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
index 0fe30e6..40748dc 100644
--- a/kernel/sched/fair.c
+++ b/kernel/sched/fair.c
@@ -3030,7 +3030,14 @@ static int idle_balance(struct rq *this_rq);
 
 #else /* CONFIG_SMP */
 
-static inline void update_load_avg(struct sched_entity *se, int update_tg) {}
+static inline void update_load_avg(struct sched_entity *se, int not_used)
+{
+	struct cfs_rq *cfs_rq = cfs_rq_of(se);
+	struct rq *rq = rq_of(cfs_rq);
+
+	cpufreq_trigger_update(rq_clock(rq));
+}
+
 static inline void
 enqueue_entity_load_avg(struct cfs_rq *cfs_rq, struct sched_entity *se) {}
 static inline void

Re: [PATCH] sched/fair: Invoke cpufreq hooks for CONFIG_SMP unset

[Peter Zijlstra]

On Sat, May 07, 2016 at 08:30:51AM +0200, Ingo Molnar wrote:
> * Peter Zijlstra <peterz@infradead.org> wrote:

> > Rafael; feel free to push this through the pm tree.
> > 
> > Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
> 
> I already have it in sched/urgent.

Ah, indeed you did, I must've missed that, thanks!

Re: [PATCH] sched/fair: Invoke cpufreq hooks for CONFIG_SMP unset

[Steve Muckle]

On Sat, May 07, 2016 at 08:30:51AM +0200, Ingo Molnar wrote:
> 
> * Peter Zijlstra <peterz@infradead.org> wrote:
> 
> > On Fri, May 06, 2016 at 02:58:43PM +0200, Rafael J. Wysocki wrote:
> > > From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> > > 
> > > Commit 34e2c555f3e1 (cpufreq: Add mechanism for registering utilization
> > > update callbacks) overlooked the fact that update_load_avg(), where
> > > CFS invokes cpufreq utilization update hooks, becomes an empty stub for
> > > CONFIG_SMP unset.  In consequence, if CONFIG_SMP is not set, cpufreq
> > > governors are never invoked from CFS and they do not have a chance to
> > > evaluate CPU performace levels and update them often enough.  Needless
> > > to say, things don't work as expected then.
> > > 
> > > Fix the problem by making the !CONFIG_SMP stub of update_load_avg()
> > > invoke cpufreq update hooks too.
> > > 
> > > Fixes: 34e2c555f3e1 (cpufreq: Add mechanism for registering utilization update callbacks)
> > > Reported-by: Steve Muckle <steve.muckle@linaro.org>
> > > Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> > 
> > Rafael; feel free to push this through the pm tree.
> > 
> > Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
> 
> I already have it in sched/urgent.

Looks like this just missed 4.6-rc7 and the release email for
that doesn't suggest an rc8, as it stands...

Re: [PATCH] sched/fair: Invoke cpufreq hooks for CONFIG_SMP unset

[Rafael J. Wysocki]

On Monday, May 09, 2016 04:53:16 PM Steve Muckle wrote:
> On Sat, May 07, 2016 at 08:30:51AM +0200, Ingo Molnar wrote:
> > 
> > * Peter Zijlstra <peterz@infradead.org> wrote:
> > 
> > > On Fri, May 06, 2016 at 02:58:43PM +0200, Rafael J. Wysocki wrote:
> > > > From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> > > > 
> > > > Commit 34e2c555f3e1 (cpufreq: Add mechanism for registering utilization
> > > > update callbacks) overlooked the fact that update_load_avg(), where
> > > > CFS invokes cpufreq utilization update hooks, becomes an empty stub for
> > > > CONFIG_SMP unset.  In consequence, if CONFIG_SMP is not set, cpufreq
> > > > governors are never invoked from CFS and they do not have a chance to
> > > > evaluate CPU performace levels and update them often enough.  Needless
> > > > to say, things don't work as expected then.
> > > > 
> > > > Fix the problem by making the !CONFIG_SMP stub of update_load_avg()
> > > > invoke cpufreq update hooks too.
> > > > 
> > > > Fixes: 34e2c555f3e1 (cpufreq: Add mechanism for registering utilization update callbacks)
> > > > Reported-by: Steve Muckle <steve.muckle@linaro.org>
> > > > Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> > > 
> > > Rafael; feel free to push this through the pm tree.
> > > 
> > > Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
> > 
> > I already have it in sched/urgent.
> 
> Looks like this just missed 4.6-rc7 and the release email for
> that doesn't suggest an rc8, as it stands...

But the final 4.6 will be released on Sunday or so and there are a few days
left still.

Re: [PATCH] sched/fair: Invoke cpufreq hooks for CONFIG_SMP unset

[Ingo Molnar]

* Rafael J. Wysocki <rjw@rjwysocki.net> wrote:

> On Monday, May 09, 2016 04:53:16 PM Steve Muckle wrote:
> > On Sat, May 07, 2016 at 08:30:51AM +0200, Ingo Molnar wrote:
> > > 
> > > * Peter Zijlstra <peterz@infradead.org> wrote:
> > > 
> > > > On Fri, May 06, 2016 at 02:58:43PM +0200, Rafael J. Wysocki wrote:
> > > > > From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> > > > > 
> > > > > Commit 34e2c555f3e1 (cpufreq: Add mechanism for registering utilization
> > > > > update callbacks) overlooked the fact that update_load_avg(), where
> > > > > CFS invokes cpufreq utilization update hooks, becomes an empty stub for
> > > > > CONFIG_SMP unset.  In consequence, if CONFIG_SMP is not set, cpufreq
> > > > > governors are never invoked from CFS and they do not have a chance to
> > > > > evaluate CPU performace levels and update them often enough.  Needless
> > > > > to say, things don't work as expected then.
> > > > > 
> > > > > Fix the problem by making the !CONFIG_SMP stub of update_load_avg()
> > > > > invoke cpufreq update hooks too.
> > > > > 
> > > > > Fixes: 34e2c555f3e1 (cpufreq: Add mechanism for registering utilization update callbacks)
> > > > > Reported-by: Steve Muckle <steve.muckle@linaro.org>
> > > > > Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> > > > 
> > > > Rafael; feel free to push this through the pm tree.
> > > > 
> > > > Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
> > > 
> > > I already have it in sched/urgent.
> > 
> > Looks like this just missed 4.6-rc7 and the release email for
> > that doesn't suggest an rc8, as it stands...
> 
> But the final 4.6 will be released on Sunday or so and there are a few days left 
> still.

Yeah, I'll send the sched/urgent bits to Linus later today.

Thanks,

	Ingo