2017-12-19 17:49:32

by Joel Fernandes

[permalink] [raw]
Subject: [PATCH RESEND] cpufreq: schedutil: Use idle_calls counter of the remote CPU

Since the recent remote cpufreq callback work, its possible that a cpufreq
update is triggered from a remote CPU. For single policies however, the current
code uses the local CPU when trying to determine if the remote sg_cpu entered
idle or is busy. This is incorrect. To remedy this, compare with the nohz tick
idle_calls counter of the remote CPU.

Acked-by: Viresh Kumar <[email protected]>
Signed-off-by: Joel Fernandes <[email protected]>
---
Just resending this which is cpufreq-related as requested by Rafael rebased
on linus/master.

The other 2 patches in my last series which can go in independent of this one are:
https://patchwork.kernel.org/patch/10115395/
https://patchwork.kernel.org/patch/10115401/
I'm still waiting on scheduler maintainers to comment on those. Unfortunately,
I haven't heard back anything yet since the last repost of those.

include/linux/tick.h | 1 +
kernel/sched/cpufreq_schedutil.c | 2 +-
kernel/time/tick-sched.c | 13 +++++++++++++
3 files changed, 15 insertions(+), 1 deletion(-)

diff --git a/include/linux/tick.h b/include/linux/tick.h
index f442d1a42025..7cc35921218e 100644
--- a/include/linux/tick.h
+++ b/include/linux/tick.h
@@ -119,6 +119,7 @@ extern void tick_nohz_idle_exit(void);
extern void tick_nohz_irq_exit(void);
extern ktime_t tick_nohz_get_sleep_length(void);
extern unsigned long tick_nohz_get_idle_calls(void);
+extern unsigned long tick_nohz_get_idle_calls_cpu(int cpu);
extern u64 get_cpu_idle_time_us(int cpu, u64 *last_update_time);
extern u64 get_cpu_iowait_time_us(int cpu, u64 *last_update_time);
#else /* !CONFIG_NO_HZ_COMMON */
diff --git a/kernel/sched/cpufreq_schedutil.c b/kernel/sched/cpufreq_schedutil.c
index 2f52ec0f1539..d6717a3331a1 100644
--- a/kernel/sched/cpufreq_schedutil.c
+++ b/kernel/sched/cpufreq_schedutil.c
@@ -244,7 +244,7 @@ static void sugov_iowait_boost(struct sugov_cpu *sg_cpu, unsigned long *util,
#ifdef CONFIG_NO_HZ_COMMON
static bool sugov_cpu_is_busy(struct sugov_cpu *sg_cpu)
{
- unsigned long idle_calls = tick_nohz_get_idle_calls();
+ unsigned long idle_calls = tick_nohz_get_idle_calls_cpu(sg_cpu->cpu);
bool ret = idle_calls == sg_cpu->saved_idle_calls;

sg_cpu->saved_idle_calls = idle_calls;
diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
index 99578f06c8d4..77555faf6fbc 100644
--- a/kernel/time/tick-sched.c
+++ b/kernel/time/tick-sched.c
@@ -985,6 +985,19 @@ ktime_t tick_nohz_get_sleep_length(void)
return ts->sleep_length;
}

+/**
+ * tick_nohz_get_idle_calls_cpu - return the current idle calls counter value
+ * for a particular CPU.
+ *
+ * Called from the schedutil frequency scaling governor in scheduler context.
+ */
+unsigned long tick_nohz_get_idle_calls_cpu(int cpu)
+{
+ struct tick_sched *ts = tick_get_tick_sched(cpu);
+
+ return ts->idle_calls;
+}
+
/**
* tick_nohz_get_idle_calls - return the current idle calls counter value
*
--
2.15.1.504.g5279b80103-goog


2017-12-19 18:48:31

by Peter Zijlstra

[permalink] [raw]
Subject: Re: [PATCH RESEND] cpufreq: schedutil: Use idle_calls counter of the remote CPU

On Tue, Dec 19, 2017 at 09:47:12AM -0800, Joel Fernandes wrote:
> Since the recent remote cpufreq callback work, its possible that a cpufreq
> update is triggered from a remote CPU. For single policies however, the current
> code uses the local CPU when trying to determine if the remote sg_cpu entered
> idle or is busy. This is incorrect. To remedy this, compare with the nohz tick
> idle_calls counter of the remote CPU.
>
> Acked-by: Viresh Kumar <[email protected]>
> Signed-off-by: Joel Fernandes <[email protected]>

Acked-by: Peter Zijlstra (Intel) <[email protected]>

> ---
> Just resending this which is cpufreq-related as requested by Rafael rebased
> on linus/master.
>
> The other 2 patches in my last series which can go in independent of this one are:
> https://patchwork.kernel.org/patch/10115395/
> https://patchwork.kernel.org/patch/10115401/
> I'm still waiting on scheduler maintainers to comment on those. Unfortunately,
> I haven't heard back anything yet since the last repost of those.

Both of us have been somewhat preoccupied with that whole kaiser/pti
thing the past few weeks.

I have an absolutely stupid backlog :/

2017-12-19 19:32:14

by Joel Fernandes

[permalink] [raw]
Subject: Re: [PATCH RESEND] cpufreq: schedutil: Use idle_calls counter of the remote CPU

On Tue, Dec 19, 2017 at 10:48 AM, Peter Zijlstra <[email protected]> wrote:
> On Tue, Dec 19, 2017 at 09:47:12AM -0800, Joel Fernandes wrote:
>> Since the recent remote cpufreq callback work, its possible that a cpufreq
>> update is triggered from a remote CPU. For single policies however, the current
>> code uses the local CPU when trying to determine if the remote sg_cpu entered
>> idle or is busy. This is incorrect. To remedy this, compare with the nohz tick
>> idle_calls counter of the remote CPU.
>>
>> Acked-by: Viresh Kumar <[email protected]>
>> Signed-off-by: Joel Fernandes <[email protected]>
>
> Acked-by: Peter Zijlstra (Intel) <[email protected]>

Sweet!

>
>> ---
>> Just resending this which is cpufreq-related as requested by Rafael rebased
>> on linus/master.
>>
>> The other 2 patches in my last series which can go in independent of this one are:
>> https://patchwork.kernel.org/patch/10115395/
>> https://patchwork.kernel.org/patch/10115401/
>> I'm still waiting on scheduler maintainers to comment on those. Unfortunately,
>> I haven't heard back anything yet since the last repost of those.
>
> Both of us have been somewhat preoccupied with that whole kaiser/pti
> thing the past few weeks.

I understand, thanks for taking time to look at it! Hopefully you're
Ok with the second one as well
(https://patchwork.kernel.org/patch/10115401). And this cap aware
one's been pretty beaten to death too:
https://patchwork.kernel.org/patch/10113337/ but let me know your
objections if any.

>
> I have an absolutely stupid backlog :/

I see. :/
I am thinking of spending more time reviewing fwiw and hopefully
helping relieve some of that burden. Happy to help in any other way as
well so let me/us know how we can help.

thanks,

- Joel

2017-12-20 14:46:05

by Rafael J. Wysocki

[permalink] [raw]
Subject: Re: [PATCH RESEND] cpufreq: schedutil: Use idle_calls counter of the remote CPU

On Tuesday, December 19, 2017 8:32:09 PM CET Joel Fernandes wrote:
> On Tue, Dec 19, 2017 at 10:48 AM, Peter Zijlstra <[email protected]> wrote:
> > On Tue, Dec 19, 2017 at 09:47:12AM -0800, Joel Fernandes wrote:
> >> Since the recent remote cpufreq callback work, its possible that a cpufreq
> >> update is triggered from a remote CPU. For single policies however, the current
> >> code uses the local CPU when trying to determine if the remote sg_cpu entered
> >> idle or is busy. This is incorrect. To remedy this, compare with the nohz tick
> >> idle_calls counter of the remote CPU.
> >>
> >> Acked-by: Viresh Kumar <[email protected]>
> >> Signed-off-by: Joel Fernandes <[email protected]>
> >
> > Acked-by: Peter Zijlstra (Intel) <[email protected]>
>
> Sweet!

OK, so any chance to get a Fixes: tag for the patch?

Thanks,
Rafael

2017-12-20 17:44:41

by Joel Fernandes

[permalink] [raw]
Subject: Re: [PATCH RESEND] cpufreq: schedutil: Use idle_calls counter of the remote CPU

Hi Rafael,

On 12/20/2017 06:45 AM, Rafael J. Wysocki wrote:
> On Tuesday, December 19, 2017 8:32:09 PM CET Joel Fernandes wrote:
>> On Tue, Dec 19, 2017 at 10:48 AM, Peter Zijlstra <[email protected]> wrote:
>>> On Tue, Dec 19, 2017 at 09:47:12AM -0800, Joel Fernandes wrote:
>>>> Since the recent remote cpufreq callback work, its possible that a cpufreq
>>>> update is triggered from a remote CPU. For single policies however, the current
>>>> code uses the local CPU when trying to determine if the remote sg_cpu entered
>>>> idle or is busy. This is incorrect. To remedy this, compare with the nohz tick
>>>> idle_calls counter of the remote CPU.
>>>>
>>>> Acked-by: Viresh Kumar <[email protected]>
>>>> Signed-off-by: Joel Fernandes <[email protected]>
>>>
>>> Acked-by: Peter Zijlstra (Intel) <[email protected]>
>>
>> Sweet!
>
> OK, so any chance to get a Fixes: tag for the patch?

Sure. Please find an inline patch below with all acks and fixes tag. thanks!

---8<---
From: Joel Fernandes <[email protected]>
Subject: [PATCH] cpufreq: schedutil: Use idle_calls counter of the remote CPU

Since the recent remote cpufreq callback work, its possible that a cpufreq
update is triggered from a remote CPU. For single policies however, the current
code uses the local CPU when trying to determine if the remote sg_cpu entered
idle or is busy. This is incorrect. To remedy this, compare with the nohz tick
idle_calls counter of the remote CPU.

Fixes: 674e75411fc2 ("sched: cpufreq: Allow remote cpufreq callbacks")
Acked-by: Viresh Kumar <[email protected]>
Acked-by: Peter Zijlstra (Intel) <[email protected]>
Signed-off-by: Joel Fernandes <[email protected]>
---
include/linux/tick.h | 1 +
kernel/sched/cpufreq_schedutil.c | 2 +-
kernel/time/tick-sched.c | 13 +++++++++++++
3 files changed, 15 insertions(+), 1 deletion(-)

diff --git a/include/linux/tick.h b/include/linux/tick.h
index f442d1a42025..7cc35921218e 100644
--- a/include/linux/tick.h
+++ b/include/linux/tick.h
@@ -119,6 +119,7 @@ extern void tick_nohz_idle_exit(void);
extern void tick_nohz_irq_exit(void);
extern ktime_t tick_nohz_get_sleep_length(void);
extern unsigned long tick_nohz_get_idle_calls(void);
+extern unsigned long tick_nohz_get_idle_calls_cpu(int cpu);
extern u64 get_cpu_idle_time_us(int cpu, u64 *last_update_time);
extern u64 get_cpu_iowait_time_us(int cpu, u64 *last_update_time);
#else /* !CONFIG_NO_HZ_COMMON */
diff --git a/kernel/sched/cpufreq_schedutil.c b/kernel/sched/cpufreq_schedutil.c
index 2f52ec0f1539..d6717a3331a1 100644
--- a/kernel/sched/cpufreq_schedutil.c
+++ b/kernel/sched/cpufreq_schedutil.c
@@ -244,7 +244,7 @@ static void sugov_iowait_boost(struct sugov_cpu *sg_cpu, unsigned long *util,
#ifdef CONFIG_NO_HZ_COMMON
static bool sugov_cpu_is_busy(struct sugov_cpu *sg_cpu)
{
- unsigned long idle_calls = tick_nohz_get_idle_calls();
+ unsigned long idle_calls = tick_nohz_get_idle_calls_cpu(sg_cpu->cpu);
bool ret = idle_calls == sg_cpu->saved_idle_calls;

sg_cpu->saved_idle_calls = idle_calls;
diff --git a/kernel/time/tick-sched.c b/kernel/time/tick-sched.c
index 99578f06c8d4..77555faf6fbc 100644
--- a/kernel/time/tick-sched.c
+++ b/kernel/time/tick-sched.c
@@ -985,6 +985,19 @@ ktime_t tick_nohz_get_sleep_length(void)
return ts->sleep_length;
}

+/**
+ * tick_nohz_get_idle_calls_cpu - return the current idle calls counter value
+ * for a particular CPU.
+ *
+ * Called from the schedutil frequency scaling governor in scheduler context.
+ */
+unsigned long tick_nohz_get_idle_calls_cpu(int cpu)
+{
+ struct tick_sched *ts = tick_get_tick_sched(cpu);
+
+ return ts->idle_calls;
+}
+
/**
* tick_nohz_get_idle_calls - return the current idle calls counter value
*