On Tue, 2015-12-15 at 22:57 +0200, Ville Syrjälä wrote:
> On Tue, Dec 15, 2015 at 08:10:30PM +0200, Imre Deak wrote:
> > All platforms with power well support have runtime PM support, so
> > simplify things by explicitly disabling power well support on
> > platforms
> > without runtime PM support. This results in holding the init power
> > domain
> > reference whenever the driver is loaded in addition to an RPM
> > reference,
> > which reflects the reality better and makes it possible to simplify
> > things by removing the HAS_RUNTIME_PM special casing from more
> > places in
> > a follow-up patch.
> >
> > Signed-off-by: Imre Deak <imre.deak@xxxxxxxxx>
> > ---
> > drivers/gpu/drm/i915/intel_runtime_pm.c | 5 +++++
> > 1 file changed, 5 insertions(+)
> >
> > diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c
> > b/drivers/gpu/drm/i915/intel_runtime_pm.c
> > index 9945040..f4ff5f5 100644
> > --- a/drivers/gpu/drm/i915/intel_runtime_pm.c
> > +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c
> > @@ -1908,6 +1908,11 @@ static int
> > sanitize_disable_power_well_option(const struct drm_i915_private
> > *dev_priv,
> > int disable_power_well)
> > {
> > + if (!HAS_RUNTIME_PM(dev_priv)) {
> > + DRM_DEBUG_KMS("No runtime PM support, disabling
> > display power well support\n");
> > + return 0;
> > + }
>
> Feels a bit too magic to me. Needs a comment at least, otherwise
> someone
> is going to change disable_power_well back into something you can
> disable
> at runtime and then all the old stuff might break.
>
> Grabbing an extra rpm reference explicitly for this purpose might be
> less confusing.
Ok, agreed. But making power well support depend on RPM still ok with
you? I can add then the extra RPM get/put only for !HAS_RUNTIME_PM.
>
> > +
> > if (disable_power_well >= 0)
> > return !!disable_power_well;
> >
> > --
> > 2.5.0
>
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
