[Intel-gfx] [PATCH] drm/i915: edp resume/On time optimization.

[Ville Syrjälä]

On Thu, Dec 17, 2015 at 01:04:25AM +0000, Kumar, Abhay wrote:
> Sure. In next patch set will get rid of jiffies altogether and will use getboottime() instead of do_gettimeofday() for panel_power_cycle_delay.
> 
> Does this make sense?

I can't say for sure until I see a patch, but that's my current thinking
yes.

> 
> 
> -----Original Message-----
> From: Ville Syrjälä [mailto:ville.syrjala at linux.intel.com] 
> Sent: Wednesday, December 16, 2015 2:56 AM
> To: Kumar, Abhay
> Cc: Intel-gfx at lists.freedesktop.org
> Subject: Re: [Intel-gfx] [PATCH] drm/i915: edp resume/On time optimization.
> 
> On Tue, Dec 15, 2015 at 02:16:38PM -0800, abhay.kumar at intel.com wrote:
> > From: Abhay Kumar <abhay.kumar at intel.com>
> > 
> > Make resume codepath not to wait for panel_power_cycle_delay(t11_t12) 
> > if this time is already spent in suspend/poweron time.
> > 
> > Signed-off-by: Abhay Kumar <abhay.kumar at intel.com>
> > ---
> >  drivers/gpu/drm/i915/intel_ddi.c |  3 +++  
> > drivers/gpu/drm/i915/intel_dp.c  | 18 ++++++++++++++++++  
> > drivers/gpu/drm/i915/intel_drv.h |  2 ++
> >  3 files changed, 23 insertions(+)
> > 
> > diff --git a/drivers/gpu/drm/i915/intel_ddi.c 
> > b/drivers/gpu/drm/i915/intel_ddi.c
> > index f00a3c9..d2a5a89 100644
> > --- a/drivers/gpu/drm/i915/intel_ddi.c
> > +++ b/drivers/gpu/drm/i915/intel_ddi.c
> > @@ -2395,6 +2395,9 @@ static void intel_ddi_post_disable(struct intel_encoder *intel_encoder)
> >  		intel_dp_sink_dpms(intel_dp, DRM_MODE_DPMS_OFF);
> >  		intel_edp_panel_vdd_on(intel_dp);
> >  		intel_edp_panel_off(intel_dp);
> > +
> > +		/* storing panel power off time */
> > +		do_gettimeofday(&intel_dp->panel_power_off_timestamp);
> 
> I think what we want to use is CLOCK_BOOTTIME. It's like MONOTONIC, except it counts the suspend time too.
> 
> Initially I figured we'd use REALTIME, and only do the adjustment around suspend/resume. But actually BOOTTIME should be perfectly safe to use all the time (changing the system time doesn't affect it). So maybe we just want to convert the power cycle delay handling entirely over to using the BOOTTIME clock instead of jiffies?
> 
> >  	}
> >  
> >  	if (IS_SKYLAKE(dev) || IS_KABYLAKE(dev)) diff --git 
> > a/drivers/gpu/drm/i915/intel_dp.c b/drivers/gpu/drm/i915/intel_dp.c 
> > index 0f1eb96..1ca01b1 100644
> > --- a/drivers/gpu/drm/i915/intel_dp.c
> > +++ b/drivers/gpu/drm/i915/intel_dp.c
> > @@ -2032,6 +2032,9 @@ static void edp_panel_on(struct intel_dp *intel_dp)
> >  	struct drm_i915_private *dev_priv = dev->dev_private;
> >  	u32 pp;
> >  	i915_reg_t pp_ctrl_reg;
> > +	u32 panel_power_off_duration;
> > +	u32 temp_power_cycle_delay;
> > +
> >  
> >  	lockdep_assert_held(&dev_priv->pps_mutex);
> >  
> > @@ -2045,8 +2048,22 @@ static void edp_panel_on(struct intel_dp *intel_dp)
> >  		 "eDP port %c panel power already on\n",
> >  		 port_name(dp_to_dig_port(intel_dp)->port)))
> >  		return;
> > +	/* taking the diffrence of currrent time and panel power off time
> > +	   and then make panel to wait for T12 if needed */
> > +	do_gettimeofday(&intel_dp->panel_power_on_timestamp);
> > +
> > +	panel_power_off_duration  = (intel_dp->panel_power_on_timestamp.tv_sec-intel_dp->panel_power_off_timestamp.tv_sec) * 1000000 +  intel_dp->panel_power_on_timestamp.tv_usec-intel_dp->panel_power_off_timestamp.tv_usec;
> > +	panel_power_off_duration = panel_power_off_duration / 1000 ;
> > +	temp_power_cycle_delay = intel_dp->panel_power_cycle_delay;
> > +
> > +	if(panel_power_off_duration >= intel_dp->panel_power_cycle_delay) {
> > +		intel_dp->panel_power_cycle_delay = 0;
> > +	} else {
> > +		intel_dp->panel_power_cycle_delay = intel_dp->panel_power_cycle_delay - panel_power_off_duration;
> > +	}
> >  
> >  	wait_panel_power_cycle(intel_dp);
> > +	intel_dp->panel_power_cycle_delay = temp_power_cycle_delay;
> >  
> >  	pp_ctrl_reg = _pp_ctrl_reg(intel_dp);
> >  	pp = ironlake_get_pp_control(intel_dp);
> > @@ -5127,6 +5144,7 @@ static void intel_dp_init_panel_power_timestamps(struct intel_dp *intel_dp)
> >  	intel_dp->last_power_cycle = jiffies;
> >  	intel_dp->last_power_on = jiffies;
> >  	intel_dp->last_backlight_off = jiffies;
> > +	do_gettimeofday(&intel_dp->panel_power_off_timestamp);
> >  }
> >  
> >  static void
> > diff --git a/drivers/gpu/drm/i915/intel_drv.h 
> > b/drivers/gpu/drm/i915/intel_drv.h
> > index 76dfa28..66ed2cb 100644
> > --- a/drivers/gpu/drm/i915/intel_drv.h
> > +++ b/drivers/gpu/drm/i915/intel_drv.h
> > @@ -769,6 +769,8 @@ struct intel_dp {
> >  	unsigned long last_power_cycle;
> >  	unsigned long last_power_on;
> >  	unsigned long last_backlight_off;
> > +	struct timeval panel_power_off_timestamp;
> > +	struct timeval panel_power_on_timestamp;
> >  
> >  	struct notifier_block edp_notifier;
> >  
> > --
> > 1.9.1
> > 
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx at lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/intel-gfx
> 
> --
> Ville Syrjälä
> Intel OTC

-- 
Ville Syrjälä
Intel OTC