On Mon, Jun 29, 2015 at 08:08:59PM +0300, Imre Deak wrote:
> On Mon, 2015-06-29 at 17:59 +0100, Damien Lespiau wrote:
> > On Mon, Jun 29, 2015 at 07:54:42PM +0300, Imre Deak wrote:
> > > On Mon, 2015-06-29 at 17:44 +0100, Damien Lespiau wrote:
> > > > This code is all dead code since we want to go up to DC6, always.
> > >
> > > On BXT DC6 is not available, so we can only go to DC5. It's disabled on
> > > BXT atm, since runtime PM isn't enabled either.
> >
> > I knew there was a catch!
> >
> > That code is pretty convoluted btw, it would make sense to have proper
> > pre_disable/post_enabled hooks on power wells?
>
> Yes, it would make sense to simplify the code. I haven't thought about
> adding new hooks, not sure how useful they would be on other platforms.
> But to me it seems that the enable/disable parts in skl_set_power_well()
> could be split out to the existing skl_power_well_{enable,disable}, have
> you looked into that already?
Nop, I haven't, but that's indeed a nice middle ground.
--
Damien
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
