[PATCH] cpufreq: Fix wrong time unit conversion

[PATCH] cpufreq: Fix wrong time unit conversion

[Frederic Weisbecker]

The time spent by a CPU under a given frequency is stored in jiffies unit
in the cpu var cpufreq_stats_table->time_in_state[i], i being the index of
the frequency.

This is what is displayed in the following file on the right column:

     cat /sys/devices/system/cpu/cpuX/cpufreq/stats/time_in_state
     2301000 19835820
     2300000 3172
     [...]

Now cpufreq converts this jiffies unit delta to clock_t before returning it
to the user as in the above file. And that conversion is achieved using the API
cputime64_to_clock_t().

Although it accidentally works on traditional tick based cputime accounting, where
cputime_t maps directly to jiffies, it doesn't work with other types of cputime
accounting such as CONFIG_VIRT_CPU_ACCOUNTING_* where cputime_t can map to nsecs
or any granularity preffered by the architecture.

For example we get a buggy zero delta on full dyntick configurations:

     cat /sys/devices/system/cpu/cpuX/cpufreq/stats/time_in_state
     2301000 0
     2300000 0
     [...]

Fix this with using the proper jiffies_64_t to clock_t conversion.

Reported-by: Carsten Emde <C.Emde@osadl.org>
Tested-by: Carsten Emde <C.Emde@osadl.org>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Carsten Emde <C.Emde@osadl.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Clark Williams <williams@redhat.com>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
---
 drivers/cpufreq/cpufreq_stats.c |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/drivers/cpufreq/cpufreq_stats.c b/drivers/cpufreq/cpufreq_stats.c
index d37568c..10e6138 100644
--- a/drivers/cpufreq/cpufreq_stats.c
+++ b/drivers/cpufreq/cpufreq_stats.c
@@ -81,7 +81,7 @@ static ssize_t show_time_in_state(struct cpufreq_policy *policy, char *buf)
 	for (i = 0; i < stat->state_num; i++) {
 		len += sprintf(buf + len, "%u %llu\n", stat->freq_table[i],
 			(unsigned long long)
-			cputime64_to_clock_t(stat->time_in_state[i]));
+			jiffies_64_to_clock_t(stat->time_in_state[i]));
 	}
 	return len;
 }
-- 
1.7.5.4

Re: [PATCH] cpufreq: Fix wrong time unit conversion

[Paul E. McKenney]

On Sat, Sep 07, 2013 at 06:35:08PM +0200, Frederic Weisbecker wrote:
> The time spent by a CPU under a given frequency is stored in jiffies unit
> in the cpu var cpufreq_stats_table->time_in_state[i], i being the index of
> the frequency.
> 
> This is what is displayed in the following file on the right column:
> 
>      cat /sys/devices/system/cpu/cpuX/cpufreq/stats/time_in_state
>      2301000 19835820
>      2300000 3172
>      [...]
> 
> Now cpufreq converts this jiffies unit delta to clock_t before returning it
> to the user as in the above file. And that conversion is achieved using the API
> cputime64_to_clock_t().
> 
> Although it accidentally works on traditional tick based cputime accounting, where
> cputime_t maps directly to jiffies, it doesn't work with other types of cputime
> accounting such as CONFIG_VIRT_CPU_ACCOUNTING_* where cputime_t can map to nsecs
> or any granularity preffered by the architecture.
> 
> For example we get a buggy zero delta on full dyntick configurations:
> 
>      cat /sys/devices/system/cpu/cpuX/cpufreq/stats/time_in_state
>      2301000 0
>      2300000 0
>      [...]
> 
> Fix this with using the proper jiffies_64_t to clock_t conversion.
> 
> Reported-by: Carsten Emde <C.Emde@osadl.org>
> Tested-by: Carsten Emde <C.Emde@osadl.org>
> Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
> Cc: Carsten Emde <C.Emde@osadl.org>
> Cc: Thomas Gleixner <tglx@linutronix.de>
> Cc: Clark Williams <williams@redhat.com>
> Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>

Good catch!

Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>

> ---
>  drivers/cpufreq/cpufreq_stats.c |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
> 
> diff --git a/drivers/cpufreq/cpufreq_stats.c b/drivers/cpufreq/cpufreq_stats.c
> index d37568c..10e6138 100644
> --- a/drivers/cpufreq/cpufreq_stats.c
> +++ b/drivers/cpufreq/cpufreq_stats.c
> @@ -81,7 +81,7 @@ static ssize_t show_time_in_state(struct cpufreq_policy *policy, char *buf)
>  	for (i = 0; i < stat->state_num; i++) {
>  		len += sprintf(buf + len, "%u %llu\n", stat->freq_table[i],
>  			(unsigned long long)
> -			cputime64_to_clock_t(stat->time_in_state[i]));
> +			jiffies_64_to_clock_t(stat->time_in_state[i]));
>  	}
>  	return len;
>  }
> -- 
> 1.7.5.4
> 

Re: [PATCH] cpufreq: Fix wrong time unit conversion

[Andreas Schwab]

See also <http://permalink.gmane.org/gmane.linux.kernel.cpufreq/11672>.

Andreas.

-- 
Andreas Schwab, schwab@linux-m68k.org
GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
"And now for something completely different."

Re: [PATCH] cpufreq: Fix wrong time unit conversion

[Rafael J. Wysocki]

On Saturday, September 07, 2013 12:02:26 PM Paul E. McKenney wrote:
> On Sat, Sep 07, 2013 at 06:35:08PM +0200, Frederic Weisbecker wrote:
> > The time spent by a CPU under a given frequency is stored in jiffies unit
> > in the cpu var cpufreq_stats_table->time_in_state[i], i being the index of
> > the frequency.
> > 
> > This is what is displayed in the following file on the right column:
> > 
> >      cat /sys/devices/system/cpu/cpuX/cpufreq/stats/time_in_state
> >      2301000 19835820
> >      2300000 3172
> >      [...]
> > 
> > Now cpufreq converts this jiffies unit delta to clock_t before returning it
> > to the user as in the above file. And that conversion is achieved using the API
> > cputime64_to_clock_t().
> > 
> > Although it accidentally works on traditional tick based cputime accounting, where
> > cputime_t maps directly to jiffies, it doesn't work with other types of cputime
> > accounting such as CONFIG_VIRT_CPU_ACCOUNTING_* where cputime_t can map to nsecs
> > or any granularity preffered by the architecture.
> > 
> > For example we get a buggy zero delta on full dyntick configurations:
> > 
> >      cat /sys/devices/system/cpu/cpuX/cpufreq/stats/time_in_state
> >      2301000 0
> >      2300000 0
> >      [...]
> > 
> > Fix this with using the proper jiffies_64_t to clock_t conversion.
> > 
> > Reported-by: Carsten Emde <C.Emde@osadl.org>
> > Tested-by: Carsten Emde <C.Emde@osadl.org>
> > Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
> > Cc: Carsten Emde <C.Emde@osadl.org>
> > Cc: Thomas Gleixner <tglx@linutronix.de>
> > Cc: Clark Williams <williams@redhat.com>
> > Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> > Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
> 
> Good catch!
> 
> Acked-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com>

Applied, will push for 3.12-rc1.

Thanks Frederic!

> 
> > ---
> >  drivers/cpufreq/cpufreq_stats.c |    2 +-
> >  1 files changed, 1 insertions(+), 1 deletions(-)
> > 
> > diff --git a/drivers/cpufreq/cpufreq_stats.c b/drivers/cpufreq/cpufreq_stats.c
> > index d37568c..10e6138 100644
> > --- a/drivers/cpufreq/cpufreq_stats.c
> > +++ b/drivers/cpufreq/cpufreq_stats.c
> > @@ -81,7 +81,7 @@ static ssize_t show_time_in_state(struct cpufreq_policy *policy, char *buf)
> >  	for (i = 0; i < stat->state_num; i++) {
> >  		len += sprintf(buf + len, "%u %llu\n", stat->freq_table[i],
> >  			(unsigned long long)
> > -			cputime64_to_clock_t(stat->time_in_state[i]));
> > +			jiffies_64_to_clock_t(stat->time_in_state[i]));
> >  	}
> >  	return len;
> >  }
> 
-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.

Re: [PATCH] cpufreq: Fix wrong time unit conversion

[Paul E. McKenney]

On Sat, Sep 07, 2013 at 09:59:38PM +0200, Andreas Schwab wrote:
> See also <http://permalink.gmane.org/gmane.linux.kernel.cpufreq/11672>.

Does look quite similar, now that you mention it.

Perhaps if Frederic has not already sent his patch he could add your
Signed-off-by to his patch?

							Thanx, Paul

Re: [PATCH] cpufreq: Fix wrong time unit conversion

[Frederic Weisbecker]

On Sat, Sep 07, 2013 at 09:59:38PM +0200, Andreas Schwab wrote:
> See also <http://permalink.gmane.org/gmane.linux.kernel.cpufreq/11672>.
> 
> Andreas.

Ah I missed it.

Raphael, if it is not too late, may be you could still set Andreas as the original author?

Thanks.

> 
> -- 
> Andreas Schwab, schwab@linux-m68k.org
> GPG Key fingerprint = 58CA 54C7 6D53 942B 1756  01D3 44D5 214B 8276 4ED5
> "And now for something completely different."

Re: [PATCH] cpufreq: Fix wrong time unit conversion

[Rafael J. Wysocki]

On Sunday, September 08, 2013 12:25:57 PM Frederic Weisbecker wrote:
> On Sat, Sep 07, 2013 at 09:59:38PM +0200, Andreas Schwab wrote:
> > See also <http://permalink.gmane.org/gmane.linux.kernel.cpufreq/11672>.
> > 
> > Andreas.
> 
> Ah I missed it.
> 
> Raphael, if it is not too late, may be you could still set Andreas as the original author?

Done.

Thanks,
Rafael

Re: [PATCH] cpufreq: Fix wrong time unit conversion

[Rafael J. Wysocki]

On Sunday, September 08, 2013 01:49:42 PM Rafael J. Wysocki wrote:
> On Sunday, September 08, 2013 12:25:57 PM Frederic Weisbecker wrote:
> > On Sat, Sep 07, 2013 at 09:59:38PM +0200, Andreas Schwab wrote:
> > > See also <http://permalink.gmane.org/gmane.linux.kernel.cpufreq/11672>.
> > > 
> > > Andreas.
> > 
> > Ah I missed it.
> > 
> > Raphael, if it is not too late, may be you could still set Andreas as the original author?
> 
> Done.

BTW, CCing power management patches to linux-pm@vger.kernel.org helps a bit,
because I can see them in patchwork then and that makes overlooking them or
forgetting about them much less likely.

Thanks,
Rafael

Re: [PATCH] cpufreq: Fix wrong time unit conversion

[Frederic Weisbecker]

On Sun, Sep 08, 2013 at 01:53:13PM +0200, Rafael J. Wysocki wrote:
> On Sunday, September 08, 2013 01:49:42 PM Rafael J. Wysocki wrote:
> > On Sunday, September 08, 2013 12:25:57 PM Frederic Weisbecker wrote:
> > > On Sat, Sep 07, 2013 at 09:59:38PM +0200, Andreas Schwab wrote:
> > > > See also <http://permalink.gmane.org/gmane.linux.kernel.cpufreq/11672>.
> > > > 
> > > > Andreas.
> > > 
> > > Ah I missed it.
> > > 
> > > Raphael, if it is not too late, may be you could still set Andreas as the original author?
> > 
> > Done.
> 
> BTW, CCing power management patches to linux-pm@vger.kernel.org helps a bit,
> because I can see them in patchwork then and that makes overlooking them or
> forgetting about them much less likely.

Ok, I'll do that next time.

Thanks.