[v3,2/2] Documentation/sysctl: Document uclamp sysctl knobs

Commit Message

Qais Yousef April 28, 2020, 4:41 p.m. UTC

Uclamp exposes 3 sysctl knobs:

	* sched_util_clamp_min
	* sched_util_clamp_max
	* sched_util_clamp_min_rt_default

Document them in sysctl/kernel.rst.

Signed-off-by: Qais Yousef <[e-mail]>
CC: Jonathan Corbet <[e-mail]>
CC: Juri Lelli <[e-mail]>
CC: Vincent Guittot <[e-mail]>
CC: Dietmar Eggemann <[e-mail]>
CC: Steven Rostedt <[e-mail]>
CC: Ben Segall <[e-mail]>
CC: Mel Gorman <[e-mail]>
CC: Luis Chamberlain <[e-mail]>
CC: Kees Cook <[e-mail]>
CC: Iurii Zaikin <[e-mail]>
CC: Quentin Perret <[e-mail]>
CC: Valentin Schneider <[e-mail]>
CC: Patrick Bellasi <[e-mail]>
CC: Pavan Kondeti <[e-mail]>
CC: [e-mail]
CC: [e-mail]
CC: [e-mail]
---
 Documentation/admin-guide/sysctl/kernel.rst | 48 +++++++++++++++++++++
 1 file changed, 48 insertions(+)

Comments

Randy Dunlap April 28, 2020, 5:43 p.m. UTC | #1

Hi--

I have a few corrections for you below:

On 4/28/20 9:41 AM, Qais Yousef wrote:
> Uclamp exposes 3 sysctl knobs:
> 
> 	* sched_util_clamp_min
> 	* sched_util_clamp_max
> 	* sched_util_clamp_min_rt_default
> 
> Document them in sysctl/kernel.rst.
> 
> Signed-off-by: Qais Yousef <[e-mail]>
> CC: Jonathan Corbet <[e-mail]>
> CC: Juri Lelli <[e-mail]>
> CC: Vincent Guittot <[e-mail]>
> CC: Dietmar Eggemann <[e-mail]>
> CC: Steven Rostedt <[e-mail]>
> CC: Ben Segall <[e-mail]>
> CC: Mel Gorman <[e-mail]>
> CC: Luis Chamberlain <[e-mail]>
> CC: Kees Cook <[e-mail]>
> CC: Iurii Zaikin <[e-mail]>
> CC: Quentin Perret <[e-mail]>
> CC: Valentin Schneider <[e-mail]>
> CC: Patrick Bellasi <[e-mail]>
> CC: Pavan Kondeti <[e-mail]>
> CC: [e-mail]
> CC: [e-mail]
> CC: [e-mail]
> ---
>  Documentation/admin-guide/sysctl/kernel.rst | 48 +++++++++++++++++++++
>  1 file changed, 48 insertions(+)
> 
> diff --git a/Documentation/admin-guide/sysctl/kernel.rst b/Documentation/admin-guide/sysctl/kernel.rst
> index 0d427fd10941..e7255f71493c 100644
> --- a/Documentation/admin-guide/sysctl/kernel.rst
> +++ b/Documentation/admin-guide/sysctl/kernel.rst
> @@ -940,6 +940,54 @@ Enables/disables scheduler statistics. Enabling this feature
>  incurs a small amount of overhead in the scheduler but is
>  useful for debugging and performance tuning.
>  
> +sched_util_clamp_min:
> +=====================
> +
> +Max allowed *minimum* utilization.
> +
> +Default value is SCHED_CAPACITY_SCALE (1024), which is the maximum possible
> +value.
> +
> +It means that any requested uclamp.min value cannot be greater than
> +sched_util_clamp_min, ie: it is restricted to the range

                         i.e., it is

> +[0:sched_util_clamp_min].
> +
> +sched_util_clamp_max:
> +=====================
> +
> +Max allowed *maximum* utilization.
> +
> +Default value is SCHED_CAPACITY_SCALE (1024), which is the maximum possible
> +value.
> +
> +It means that any requested uclamp.max value cannot be greater than
> +sched_util_clamp_max, ie: it is restricted to the range

                         i.e., it is

> +[0:sched_util_clamp_max].
> +
> +sched_util_clamp_min_rt_default:
> +================================
> +
> +By default Linux is tuned for performance. Which means that RT tasks always run
> +at the highest frequency and most capable (highest capacity) CPU (in
> +heterogeneous systems).
> +
> +Uclamp achieves this by setting the requested uclamp.min of all RT tasks to
> +SCHED_CAPACITY_SCALE (1024) by default. Which effectively boosts the tasks to

                               by default, which

> +run at the highest frequency and bias them to run on the biggest CPU.

                                    biases them

> +
> +This knob allows admins to change the default behavior when uclamp is being
> +used. In battery powered devices particularly, running at the maximum
> +capacity and frequency will increase energy consumption and shorten the battery
> +life.
> +
> +This knob is only effective for RT tasks which the user hasn't modified their
> +requested uclamp.min value via sched_setattr() syscall.
> +
> +This knob will not escape the constraint imposed by sched_util_clamp_min
> +defined above.
> +
> +Any modification is applied lazily on the next opportunity the scheduler needs
> +to calculate the effective value of uclamp.min of the task.
>  
>  seccomp
>  =======
> 

thanks.

Qais Yousef April 29, 2020, 9:54 a.m. UTC | #2

On 04/28/20 10:43, Randy Dunlap wrote:
> Hi--
> 
> I have a few corrections for you below:

Thanks Randy. I applied all your suggestions for the next version.

Thanks

--
Qais Yousef

> 
> On 4/28/20 9:41 AM, Qais Yousef wrote:
> > Uclamp exposes 3 sysctl knobs:
> > 
> > 	* sched_util_clamp_min
> > 	* sched_util_clamp_max
> > 	* sched_util_clamp_min_rt_default
> > 
> > Document them in sysctl/kernel.rst.
> > 
> > Signed-off-by: Qais Yousef <[e-mail]>
> > CC: Jonathan Corbet <[e-mail]>
> > CC: Juri Lelli <[e-mail]>
> > CC: Vincent Guittot <[e-mail]>
> > CC: Dietmar Eggemann <[e-mail]>
> > CC: Steven Rostedt <[e-mail]>
> > CC: Ben Segall <[e-mail]>
> > CC: Mel Gorman <[e-mail]>
> > CC: Luis Chamberlain <[e-mail]>
> > CC: Kees Cook <[e-mail]>
> > CC: Iurii Zaikin <[e-mail]>
> > CC: Quentin Perret <[e-mail]>
> > CC: Valentin Schneider <[e-mail]>
> > CC: Patrick Bellasi <[e-mail]>
> > CC: Pavan Kondeti <[e-mail]>
> > CC: [e-mail]
> > CC: [e-mail]
> > CC: [e-mail]
> > ---
> >  Documentation/admin-guide/sysctl/kernel.rst | 48 +++++++++++++++++++++
> >  1 file changed, 48 insertions(+)
> > 
> > diff --git a/Documentation/admin-guide/sysctl/kernel.rst b/Documentation/admin-guide/sysctl/kernel.rst
> > index 0d427fd10941..e7255f71493c 100644
> > --- a/Documentation/admin-guide/sysctl/kernel.rst
> > +++ b/Documentation/admin-guide/sysctl/kernel.rst
> > @@ -940,6 +940,54 @@ Enables/disables scheduler statistics. Enabling this feature
> >  incurs a small amount of overhead in the scheduler but is
> >  useful for debugging and performance tuning.
> >  
> > +sched_util_clamp_min:
> > +=====================
> > +
> > +Max allowed *minimum* utilization.
> > +
> > +Default value is SCHED_CAPACITY_SCALE (1024), which is the maximum possible
> > +value.
> > +
> > +It means that any requested uclamp.min value cannot be greater than
> > +sched_util_clamp_min, ie: it is restricted to the range
> 
>                          i.e., it is
> 
> > +[0:sched_util_clamp_min].
> > +
> > +sched_util_clamp_max:
> > +=====================
> > +
> > +Max allowed *maximum* utilization.
> > +
> > +Default value is SCHED_CAPACITY_SCALE (1024), which is the maximum possible
> > +value.
> > +
> > +It means that any requested uclamp.max value cannot be greater than
> > +sched_util_clamp_max, ie: it is restricted to the range
> 
>                          i.e., it is
> 
> > +[0:sched_util_clamp_max].
> > +
> > +sched_util_clamp_min_rt_default:
> > +================================
> > +
> > +By default Linux is tuned for performance. Which means that RT tasks always run
> > +at the highest frequency and most capable (highest capacity) CPU (in
> > +heterogeneous systems).
> > +
> > +Uclamp achieves this by setting the requested uclamp.min of all RT tasks to
> > +SCHED_CAPACITY_SCALE (1024) by default. Which effectively boosts the tasks to
> 
>                                by default, which
> 
> > +run at the highest frequency and bias them to run on the biggest CPU.
> 
>                                     biases them
> 
> > +
> > +This knob allows admins to change the default behavior when uclamp is being
> > +used. In battery powered devices particularly, running at the maximum
> > +capacity and frequency will increase energy consumption and shorten the battery
> > +life.
> > +
> > +This knob is only effective for RT tasks which the user hasn't modified their
> > +requested uclamp.min value via sched_setattr() syscall.
> > +
> > +This knob will not escape the constraint imposed by sched_util_clamp_min
> > +defined above.
> > +
> > +Any modification is applied lazily on the next opportunity the scheduler needs
> > +to calculate the effective value of uclamp.min of the task.
> >  
> >  seccomp
> >  =======
> > 
> 
> thanks.
> -- 
> ~Randy
>

Qais Yousef April 30, 2020, 10:03 a.m. UTC | #3

On 04/30/20 13:06, Tao Zhou wrote:
> Hi,
> 
> On Tue, Apr 28, 2020 at 05:41:34PM +0100, Qais Yousef wrote:
> > Uclamp exposes 3 sysctl knobs:
> > 
> > 	* sched_util_clamp_min
> > 	* sched_util_clamp_max
> > 	* sched_util_clamp_min_rt_default
>  
> > Document them in sysctl/kernel.rst.
> > 
> > Signed-off-by: Qais Yousef <[e-mail]>
> > CC: Jonathan Corbet <[e-mail]>
> > CC: Juri Lelli <[e-mail]>
> > CC: Vincent Guittot <[e-mail]>
> > CC: Dietmar Eggemann <[e-mail]>
> > CC: Steven Rostedt <[e-mail]>
> > CC: Ben Segall <[e-mail]>
> > CC: Mel Gorman <[e-mail]>
> > CC: Luis Chamberlain <[e-mail]>
> > CC: Kees Cook <[e-mail]>
> > CC: Iurii Zaikin <[e-mail]>
> > CC: Quentin Perret <[e-mail]>
> > CC: Valentin Schneider <[e-mail]>
> > CC: Patrick Bellasi <[e-mail]>
> > CC: Pavan Kondeti <[e-mail]>
> > CC: [e-mail]
> > CC: [e-mail]
> > CC: [e-mail]
> > ---
> >  Documentation/admin-guide/sysctl/kernel.rst | 48 +++++++++++++++++++++
> >  1 file changed, 48 insertions(+)
> > 
> > diff --git a/Documentation/admin-guide/sysctl/kernel.rst b/Documentation/admin-guide/sysctl/kernel.rst
> > index 0d427fd10941..e7255f71493c 100644
> > --- a/Documentation/admin-guide/sysctl/kernel.rst
> > +++ b/Documentation/admin-guide/sysctl/kernel.rst
> > @@ -940,6 +940,54 @@ Enables/disables scheduler statistics. Enabling this feature
> >  incurs a small amount of overhead in the scheduler but is
> >  useful for debugging and performance tuning.
> >  
> > +sched_util_clamp_min:
> > +=====================
> > +
> > +Max allowed *minimum* utilization.
> > +
> > +Default value is SCHED_CAPACITY_SCALE (1024), which is the maximum possible
> > +value.
> > +
> > +It means that any requested uclamp.min value cannot be greater than
>                                                           ^^^^^^^
> 
> Seems that 'greater' should be 'smaller'.
> And the range is [sched_util_clamp_min:SCHED_CAPACITY_SCALE]
> Uclamp request should not below this system value.
> Or I am totally wrong for memory leak.

No the range is [0:sched_util_clamp_min].

It is what it is implemented by this condition:

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/kernel/sched/core.c?h=v5.7-rc3#n913

So if the requested uclamp.min value is greater than sched_util_clamp_min, we
return sched_util_clamp_min.

Thanks

--
Qais Yousef

> 
> Thank you
> 
> > +sched_util_clamp_min, ie: it is restricted to the range
> > +[0:sched_util_clamp_min].
> > +
> > +sched_util_clamp_max:
> > +=====================
> > +
> > +Max allowed *maximum* utilization.
> > +
> > +Default value is SCHED_CAPACITY_SCALE (1024), which is the maximum possible
> > +value.
> > +
> > +It means that any requested uclamp.max value cannot be greater than
> > +sched_util_clamp_max, ie: it is restricted to the range
> > +[0:sched_util_clamp_max].
> > +
> > +sched_util_clamp_min_rt_default:
> > +================================
> > +
> > +By default Linux is tuned for performance. Which means that RT tasks always run
> > +at the highest frequency and most capable (highest capacity) CPU (in
> > +heterogeneous systems).
> > +
> > +Uclamp achieves this by setting the requested uclamp.min of all RT tasks to
> > +SCHED_CAPACITY_SCALE (1024) by default. Which effectively boosts the tasks to
> > +run at the highest frequency and bias them to run on the biggest CPU.
> > +
> > +This knob allows admins to change the default behavior when uclamp is being
> > +used. In battery powered devices particularly, running at the maximum
> > +capacity and frequency will increase energy consumption and shorten the battery
> > +life.
> > +
> > +This knob is only effective for RT tasks which the user hasn't modified their
> > +requested uclamp.min value via sched_setattr() syscall.
> > +
> > +This knob will not escape the constraint imposed by sched_util_clamp_min
> > +defined above.
> > +
> > +Any modification is applied lazily on the next opportunity the scheduler needs
> > +to calculate the effective value of uclamp.min of the task.
> >  
> >  seccomp
> >  =======
> > -- 
> > 2.17.1
> >

Patch

diff --git a/Documentation/admin-guide/sysctl/kernel.rst b/Documentation/admin-guide/sysctl/kernel.rst
index 0d427fd10941..e7255f71493c 100644
--- a/Documentation/admin-guide/sysctl/kernel.rst
+++ b/Documentation/admin-guide/sysctl/kernel.rst
@@ -940,6 +940,54 @@  Enables/disables scheduler statistics. Enabling this feature
 incurs a small amount of overhead in the scheduler but is
 useful for debugging and performance tuning.
 
+sched_util_clamp_min:
+=====================
+
+Max allowed *minimum* utilization.
+
+Default value is SCHED_CAPACITY_SCALE (1024), which is the maximum possible
+value.
+
+It means that any requested uclamp.min value cannot be greater than
+sched_util_clamp_min, ie: it is restricted to the range
+[0:sched_util_clamp_min].
+
+sched_util_clamp_max:
+=====================
+
+Max allowed *maximum* utilization.
+
+Default value is SCHED_CAPACITY_SCALE (1024), which is the maximum possible
+value.
+
+It means that any requested uclamp.max value cannot be greater than
+sched_util_clamp_max, ie: it is restricted to the range
+[0:sched_util_clamp_max].
+
+sched_util_clamp_min_rt_default:
+================================
+
+By default Linux is tuned for performance. Which means that RT tasks always run
+at the highest frequency and most capable (highest capacity) CPU (in
+heterogeneous systems).
+
+Uclamp achieves this by setting the requested uclamp.min of all RT tasks to
+SCHED_CAPACITY_SCALE (1024) by default. Which effectively boosts the tasks to
+run at the highest frequency and bias them to run on the biggest CPU.
+
+This knob allows admins to change the default behavior when uclamp is being
+used. In battery powered devices particularly, running at the maximum
+capacity and frequency will increase energy consumption and shorten the battery
+life.
+
+This knob is only effective for RT tasks which the user hasn't modified their
+requested uclamp.min value via sched_setattr() syscall.
+
+This knob will not escape the constraint imposed by sched_util_clamp_min
+defined above.
+
+Any modification is applied lazily on the next opportunity the scheduler needs
+to calculate the effective value of uclamp.min of the task.
 
 seccomp
 =======