[Intel-gfx] [PATCH 1/3] drm/i915: Dont enable hpd for eDP

[Daniel Vetter]

On Mon, Aug 10, 2015 at 05:51:48PM +0530, Sivakumar Thulasimani wrote:
> 
> 
> On 8/10/2015 5:44 PM, Jani Nikula wrote:
> >On Mon, 10 Aug 2015, Sivakumar Thulasimani <sivakumar.thulasimani at intel.com> wrote:
> >>On 8/10/2015 5:07 PM, Jani Nikula wrote:
> >>>On Mon, 10 Aug 2015, Sivakumar Thulasimani <sivakumar.thulasimani at intel.com> wrote:
> >>>>Reviewed-by: Sivakumar Thulasimani <sivakumar.thulasimani at intel.com>
> >>>>
> >>>>On 8/10/2015 10:35 AM, Sonika Jindal wrote:
> >>>>>With HPD support added for all ports including PORT_A, setting hpd_pin will
> >>>>>result in enabling of hpd to edp as well. There is no need to enable HPD on
> >>>>>PORT_A hence this patch removes hpd_pin update for PORT_A, where edp will
> >>>>>be connected. it can be added back when required
> >>>What? You can't just go ahead and remove HPD from eDP sinks.
> >>>
> >>>BR,
> >>>Jani.
> >>Nope, we are not removing HPD for edp sinks, it was never there in the
> >>first place. It was
> >>enabled for CHV (even there by mistake since PORT B/C was both DP and
> >>eDP) but it was
> >>never there for any other plaforms nor is it used for any purpose (PSR
> >>must use it, but i
> >>dont see code for it as well).
> >Are you saying there's no HPD enabled in our *hardware* for eDP? Or
> >driver?
> >
> >My point is, is this patch making it harder to enable eDP hpd handling
> >(e.g. for PSR or DP link re-training) in the future? We currently take
> >it into account in a few places, and if we start removing that, it will
> >be a loss of effort to first remove and then add it back.
> >
> >BR,
> >Jani.
> i was referring to our driver only.
> 
> Our VLV/CHV code already receives hpd for every pps on and off which is
> later ignored. if we dont disable HPD on eDPs this behavior will be extended
> for all platforms which i feel is too costly to keep enabled when there is
> no
> purpose for it right now.

don't optimize code because you "feel it's costly", only do it when you
have hard numbers. One interrupt per pps on or off transition won't be
measurable at all.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch