[PATCH] cpuidle: Default y for pseries

[PATCH] cpuidle: Default y for pseries

[Benjamin Herrenschmidt]

We just replaced the pseries platform idle loops with a cpuidle backend,
however that means that you won't get any power saving and won't return
any unused idle time to the hypervisor unless cpuidle is enabled.

Thus is should default to y when pseries is enabled. I prefer that to
a select so we can still make it modular if we want to.

Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---

Linus, do you want to just pick that up or should I put it into powerpc.git
and ask you to pull ? I will have 2 or 3 other fixes there later today,
but I wanted to make sure you were ok with the approach with this
specific one.

diff --git a/drivers/cpuidle/Kconfig b/drivers/cpuidle/Kconfig
index 7dbc4a8..62ca70d 100644
--- a/drivers/cpuidle/Kconfig
+++ b/drivers/cpuidle/Kconfig
@@ -1,7 +1,8 @@
 
 config CPU_IDLE
 	bool "CPU idle PM support"
-	default ACPI
+	default y if ACPI
+	default y if PPC_PSERIES
 	help
 	  CPU idle is a generic framework for supporting software-controlled
 	  idle processor power management.  It includes modular cross-platform

Re: [PATCH] cpuidle: Default y for pseries

[Linus Torvalds]

On Tue, Jan 10, 2012 at 5:05 PM, Benjamin Herrenschmidt
<benh@kernel.crashing.org> wrote:
>
> Linus, do you want to just pick that up or should I put it into powerpc.git
> and ask you to pull ? I will have 2 or 3 other fixes there later today,
> but I wanted to make sure you were ok with the approach with this
> specific one.

It doesn't seem to be all that different from the "default y if ACPI"
case, so I guess it works ok.

That said, I wonder if the right approach wouldn't be

   default y if SUPPORT_CPU_IDLE

or something along those lines. And then both ACPI and PPC_PSERIES
could just select that instead. Because I do hate having random
board-level knowledge in something like this. I dunno.

                   Linus

Re: [PATCH] cpuidle: Default y for pseries

[Benjamin Herrenschmidt]

On Tue, 2012-01-10 at 22:08 -0800, Linus Torvalds wrote:
> On Tue, Jan 10, 2012 at 5:05 PM, Benjamin Herrenschmidt
> <benh@kernel.crashing.org> wrote:
> >
> > Linus, do you want to just pick that up or should I put it into powerpc.git
> > and ask you to pull ? I will have 2 or 3 other fixes there later today,
> > but I wanted to make sure you were ok with the approach with this
> > specific one.
> 
> It doesn't seem to be all that different from the "default y if ACPI"
> case, so I guess it works ok.

It works for my case, that's tested, but ...

> That said, I wonder if the right approach wouldn't be
> 
>    default y if SUPPORT_CPU_IDLE
> 
> or something along those lines. And then both ACPI and PPC_PSERIES
> could just select that instead. Because I do hate having random
> board-level knowledge in something like this. I dunno.

I tend to agree, I wasn't too keen on touching ACPI related stuff I
suppose it shouldn't be hard :-) Btw, what about the change:

-	default ACPI
+	default y if ACPI

(To be honest I'm not sure what the first form does in details).

Oh, also, I just see that in drivers/acpi/Kconfig:

config ACPI_PROCESSOR
	tristate "Processor"
	select THERMAL
	select CPU_IDLE
	default y

Hrm... maybe I should just do the same in pseries and remove both the
"default" statements above, what do you think ?

On pSeries I'm keen to build that in rather than make it a module too
because you get no idle handling until it loads which can be
problematic. Built-in, it seems to be quite early in the link order (if
we can still trust that nowadays ...).

Cheers,
Ben.

Re: cpuidle: Default y for pseries

[Thadeu Lima de Souza Cascardo]

On Tue, Jan 10, 2012 at 03:05:35PM -0000, Benjamin Herrenschmidt wrote:
> We just replaced the pseries platform idle loops with a cpuidle backend,
> however that means that you won't get any power saving and won't return
> any unused idle time to the hypervisor unless cpuidle is enabled.
> 
> Thus is should default to y when pseries is enabled. I prefer that to
> a select so we can still make it modular if we want to.
> 
> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
> 
> ---
> Linus, do you want to just pick that up or should I put it into powerpc.git
> and ask you to pull ? I will have 2 or 3 other fixes there later today,
> but I wanted to make sure you were ok with the approach with this
> specific one.

Hi, Ben.

Note that building with CONFIG_PSERIES_IDLE=m fails.

  CC [M]  arch/powerpc/platforms/pseries/processor_idle.o
arch/powerpc/platforms/pseries/processor_idle.c:35: error: redefinition
of ‘update_smt_snooze_delay’
/root/linux/arch/powerpc/include/asm/system.h:230: note: previous
definition of ‘update_smt_snooze_delay’ was here
arch/powerpc/platforms/pseries/processor_idle.c:175: error: redefinition
of ‘pseries_notify_cpuidle_add_cpu’
/root/linux/arch/powerpc/include/asm/system.h:231: note: previous
definition of ‘pseries_notify_cpuidle_add_cpu’ was here
make[2]: *** [arch/powerpc/platforms/pseries/processor_idle.o] Error 1
make[1]: *** [arch/powerpc/platforms/pseries] Error 2
make: *** [arch/powerpc/platforms] Error 2

asm/system.h has empty inline implementations for
update_smt_snooze_delay and pseries_notify_cpuidle_add_cpu, which are
used when CONFIG_PSERIES_IDLE is undefined. Since those two functions
are used in core power architecture functions (store_smt_snooze_delay
at kernel/sysfs.c and smp_xics_setup_cpu at platforms/pseries/smp.c),
this requires some rework in these interactions or we should simply
disable PSERIES_IDLE to be built as a module for now.

Regards.
Cascardo.

> 
> diff --git a/drivers/cpuidle/Kconfig b/drivers/cpuidle/Kconfig
> index 7dbc4a8..62ca70d 100644
> --- a/drivers/cpuidle/Kconfig
> +++ b/drivers/cpuidle/Kconfig
> @@ -1,7 +1,8 @@
>  
>  config CPU_IDLE
>  	bool "CPU idle PM support"
> -	default ACPI
> +	default y if ACPI
> +	default y if PPC_PSERIES
>  	help
>  	  CPU idle is a generic framework for supporting software-controlled
>  	  idle processor power management.  It includes modular cross-platform

Re: cpuidle: Default y for pseries

[Benjamin Herrenschmidt]

On Wed, 2012-01-11 at 20:37 -0200, Thadeu Lima de Souza Cascardo wrote:
> On Tue, Jan 10, 2012 at 03:05:35PM -0000, Benjamin Herrenschmidt wrote:
> > We just replaced the pseries platform idle loops with a cpuidle backend,
> > however that means that you won't get any power saving and won't return
> > any unused idle time to the hypervisor unless cpuidle is enabled.
> > 
> > Thus is should default to y when pseries is enabled. I prefer that to
> > a select so we can still make it modular if we want to.
> > 
> > Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
> > 
> > ---
> > Linus, do you want to just pick that up or should I put it into powerpc.git
> > and ask you to pull ? I will have 2 or 3 other fixes there later today,
> > but I wanted to make sure you were ok with the approach with this
> > specific one.
> 
> Hi, Ben.
> 
> Note that building with CONFIG_PSERIES_IDLE=m fails.

Ah ok. Well, making it built-in only makes sense anyway as I said
separately, so I think I'll just select it.

Cheers,
Ben.

>   CC [M]  arch/powerpc/platforms/pseries/processor_idle.o
> arch/powerpc/platforms/pseries/processor_idle.c:35: error: redefinition
> of ‘update_smt_snooze_delay’
> /root/linux/arch/powerpc/include/asm/system.h:230: note: previous
> definition of ‘update_smt_snooze_delay’ was here
> arch/powerpc/platforms/pseries/processor_idle.c:175: error: redefinition
> of ‘pseries_notify_cpuidle_add_cpu’
> /root/linux/arch/powerpc/include/asm/system.h:231: note: previous
> definition of ‘pseries_notify_cpuidle_add_cpu’ was here
> make[2]: *** [arch/powerpc/platforms/pseries/processor_idle.o] Error 1
> make[1]: *** [arch/powerpc/platforms/pseries] Error 2
> make: *** [arch/powerpc/platforms] Error 2
> 
> asm/system.h has empty inline implementations for
> update_smt_snooze_delay and pseries_notify_cpuidle_add_cpu, which are
> used when CONFIG_PSERIES_IDLE is undefined. Since those two functions
> are used in core power architecture functions (store_smt_snooze_delay
> at kernel/sysfs.c and smp_xics_setup_cpu at platforms/pseries/smp.c),
> this requires some rework in these interactions or we should simply
> disable PSERIES_IDLE to be built as a module for now.
> 
> Regards.
> Cascardo.
> 
> > 
> > diff --git a/drivers/cpuidle/Kconfig b/drivers/cpuidle/Kconfig
> > index 7dbc4a8..62ca70d 100644
> > --- a/drivers/cpuidle/Kconfig
> > +++ b/drivers/cpuidle/Kconfig
> > @@ -1,7 +1,8 @@
> >  
> >  config CPU_IDLE
> >  	bool "CPU idle PM support"
> > -	default ACPI
> > +	default y if ACPI
> > +	default y if PPC_PSERIES
> >  	help
> >  	  CPU idle is a generic framework for supporting software-controlled
> >  	  idle processor power management.  It includes modular cross-platform

Re: cpuidle: Default y for pseries

[Deepthi Dharwar]

On 01/12/2012 04:36 AM, Benjamin Herrenschmidt wrote:

> On Wed, 2012-01-11 at 20:37 -0200, Thadeu Lima de Souza Cascardo wrote:
>> On Tue, Jan 10, 2012 at 03:05:35PM -0000, Benjamin Herrenschmidt wrote:
>>> We just replaced the pseries platform idle loops with a cpuidle backend,
>>> however that means that you won't get any power saving and won't return
>>> any unused idle time to the hypervisor unless cpuidle is enabled.
>>>
>>> Thus is should default to y when pseries is enabled. I prefer that to
>>> a select so we can still make it modular if we want to.
>>>
>>> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
>>>
>>> ---
>>> Linus, do you want to just pick that up or should I put it into powerpc.git
>>> and ask you to pull ? I will have 2 or 3 other fixes there later today,
>>> but I wanted to make sure you were ok with the approach with this
>>> specific one.
>>
>> Hi, Ben.
>>
>> Note that building with CONFIG_PSERIES_IDLE=m fails.
> 
> Ah ok. Well, making it built-in only makes sense anyway as I said
> separately, so I think I'll just select it.
> 
> Cheers,
> Ben.
> 
>>   CC [M]  arch/powerpc/platforms/pseries/processor_idle.o
>> arch/powerpc/platforms/pseries/processor_idle.c:35: error: redefinition
>> of ‘update_smt_snooze_delay’
>> /root/linux/arch/powerpc/include/asm/system.h:230: note: previous
>> definition of ‘update_smt_snooze_delay’ was here
>> arch/powerpc/platforms/pseries/processor_idle.c:175: error: redefinition
>> of ‘pseries_notify_cpuidle_add_cpu’
>> /root/linux/arch/powerpc/include/asm/system.h:231: note: previous
>> definition of ‘pseries_notify_cpuidle_add_cpu’ was here
>> make[2]: *** [arch/powerpc/platforms/pseries/processor_idle.o] Error 1
>> make[1]: *** [arch/powerpc/platforms/pseries] Error 2
>> make: *** [arch/powerpc/platforms] Error 2
>>
>> asm/system.h has empty inline implementations for
>> update_smt_snooze_delay and pseries_notify_cpuidle_add_cpu, which are
>> used when CONFIG_PSERIES_IDLE is undefined. Since those two functions
>> are used in core power architecture functions (store_smt_snooze_delay
>> at kernel/sysfs.c and smp_xics_setup_cpu at platforms/pseries/smp.c),
>> this requires some rework in these interactions or we should simply
>> disable PSERIES_IDLE to be built as a module for now.
>>
>> Regards.
>> Cascardo.
>>
>>>
>>> diff --git a/drivers/cpuidle/Kconfig b/drivers/cpuidle/Kconfig
>>> index 7dbc4a8..62ca70d 100644
>>> --- a/drivers/cpuidle/Kconfig
>>> +++ b/drivers/cpuidle/Kconfig
>>> @@ -1,7 +1,8 @@
>>>  
>>>  config CPU_IDLE
>>>  	bool "CPU idle PM support"
>>> -	default ACPI
>>> +	default y if ACPI
>>> +	default y if PPC_PSERIES
>>>  	help
>>>  	  CPU idle is a generic framework for supporting software-controlled
>>>  	  idle processor power management.  It includes modular cross-platform



The following patch disables pseries cpuidle driver to be loaded as a
module as there are build problems reported when one is trying to do so.
This is a work around for now until the problem is fixed.

arch/powerpc/platforms/pseries/Kconfig |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)

diff --git a/arch/powerpc/platforms/pseries/Kconfig
b/arch/powerpc/platforms/pseries/Kconfig
index ae7b6d4..31f22c1 100644
--- a/arch/powerpc/platforms/pseries/Kconfig
+++ b/arch/powerpc/platforms/pseries/Kconfig
@@ -122,7 +122,7 @@ config DTL
          Say N if you are unsure.

 config PSERIES_IDLE
-       tristate "Cpuidle driver for pSeries platforms"
+       bool "Cpuidle driver for pSeries platforms"
        depends on CPU_IDLE
        depends on PPC_PSERIES
        default y

As pointed out, asm/system.h has empty inline implementations for
update_smt_snooze_delay and pseries_notify_cpuidle_add_cpu, which are
used when CONFIG_PSERIES_IDLE is undefined. Since those two functions
are used in core power architecture functions (store_smt_snooze_delay
at kernel/sysfs.c and smp_xics_setup_cpu at platforms/pseries/smp.c),

Going forward, the aim is to remove the dependency of using these
functions from core power architecture functions. So I am proposing the
following changes:

a) Removing update_smt_snooze_delay () and allow the pseries
idle backend driver to query snooze entry on need basis
while initializing the driver fields by exporting an api to return
snooze value for a given cpu.

b) Replacing the pseries_notify_cpuidle_add_cpu() call
by registering a cpu_notifier to take actions accordingly.

Regards,
Deepthi