Re: [PATCH] thermal: remove power allocator from list of default governors

[Javi Merino]

On Wed, Aug 05, 2015 at 07:49:41PM +0100, Eduardo Valentin wrote:
> On Wed, Aug 05, 2015 at 09:35:39AM -0700, Dmitry Torokhov wrote:
> > On Wed, Aug 05, 2015 at 09:37:55AM +0100, Javi Merino wrote:
> > > On Tue, Aug 04, 2015 at 05:39:21PM +0100, Dmitry Torokhov wrote:
> > > > As it currently stands the power allocator governor can not handle
> > > > thermal zones that are not specifically crafted and therefore can not be
> > > > used as a default governor.
> > > > 
> > > > Users need to explicitly enable this governor for thermal zones that do
> > > > have enough information for its operation.
> > > > 
> > > > Signed-off-by: Dmitry Torokhov <dmitry.torokhov@xxxxxxxxx>
> > > > ---
> > > >  drivers/thermal/Kconfig | 8 --------
> > > >  1 file changed, 8 deletions(-)
> > > > 
> > > > diff --git a/drivers/thermal/Kconfig b/drivers/thermal/Kconfig
> > > > index 0390044..34d05d3 100644
> > > > --- a/drivers/thermal/Kconfig
> > > > +++ b/drivers/thermal/Kconfig
> > > > @@ -82,14 +82,6 @@ config THERMAL_DEFAULT_GOV_USER_SPACE
> > > >  	  Select this if you want to let the user space manage the
> > > >  	  platform thermals.
> > > >  
> > > > -config THERMAL_DEFAULT_GOV_POWER_ALLOCATOR
> > > > -	bool "power_allocator"
> > > > -	select THERMAL_GOV_POWER_ALLOCATOR
> > > > -	help
> > > > -	  Select this if you want to control temperature based on
> > > > -	  system and device power allocation. This governor can only
> > > > -	  operate on cooling devices that implement the power API.
> > > > -
> > > 
> > > Currently the only way we have for a thermal zone configured from
> > > device tree to use a governor from the kernel boot is by using
> > > THERMAL_DEFAULT_GOV_*.  If we remove this option some devices won't
> > > have a workable thermal framework until userspace is up and running.
> > 
> > Why would step wise, or fair share be not workable (even if not optimal)
> > thermal frameworks? It doe snot take that long to get to [early]
> > userspace. Half of the boot time the thermal framework is not working
> > anyway because half of the devices that can act as colling devices are
> > not yet logically there.
> > 
> > > 
> > > Would you rather have the power allocator governor accept every
> > > thermal zone?
> > 
> > If it is to be one of default governors then yes, it needs to be able to
> > manage all thermal zones, the same way as the other 3 governors can, as
> > far as I know.
> 
> I believe fairshare also relies on platform data to be properly set,
> otherwise, it would not be functional. That's probably why I overlooked
> this point. That said, I would say, if we follow this logic, a similar
> patch is needed for fairshare.
> 
> Javi, ideally, the governor would need to have a workable
> default settings for any thermal zone that miss the platform settings.
> Do you think it would be doable for power allocator?

Yes, we can come up with something that accepts any thermal zone.
I'll send a patch soon.

Cheers,
Javi
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/