[Intel-gfx] [PATCH 14/14] drm/i915: Reinit display irqs and hpd from chv pipe-a power well

[Ville Syrjälä]

On Fri, Nov 14, 2014 at 03:49:25PM -0200, Paulo Zanoni wrote:
> 2014-10-30 15:43 GMT-02:00  <ville.syrjala at linux.intel.com>:
> > From: Ville Syrjälä <ville.syrjala at linux.intel.com>
> >
> > On chv the pipe-a power well is the new disp2d well, and it kills pretty
> > much everything in the display block. So we need to do the the same
> > dance that vlv does wrt. display irqs and hpd when the power well goes
> > up or down.
> 
> I don't have the docs for this, so I have to ask: does it kill *all*
> the interrupt bits (like VLV), or does it kill only the pipe A
> interrupt bits (like BDW on pipes B/C)?
> 
> If it kills everything: Reviewed-by: Paulo Zanoni <paulo.r.zanoni at intel.com>

Yeah, everything goes even though it's called the "pipe A" power well.
I think the plan was to have per-pipe wells, but then someone decided that
it's not worth it. For some reason they still hooked it up to the pipe A
power well bits instead of doing the more sensible thing and reusing the
same bits that VLV used for this stuff.

> 
> >
> > Signed-off-by: Ville Syrjälä <ville.syrjala at linux.intel.com>
> > ---
> >  drivers/gpu/drm/i915/intel_runtime_pm.c | 23 +++++++++++++++++++++++
> >  1 file changed, 23 insertions(+)
> >
> > diff --git a/drivers/gpu/drm/i915/intel_runtime_pm.c b/drivers/gpu/drm/i915/intel_runtime_pm.c
> > index dcbecff..f5a78d5 100644
> > --- a/drivers/gpu/drm/i915/intel_runtime_pm.c
> > +++ b/drivers/gpu/drm/i915/intel_runtime_pm.c
> > @@ -577,6 +577,23 @@ static void chv_pipe_power_well_enable(struct drm_i915_private *dev_priv,
> >                      power_well->data != PIPE_C);
> >
> >         chv_set_pipe_power_well(dev_priv, power_well, true);
> > +
> > +       if (power_well->data == PIPE_A) {
> > +               spin_lock_irq(&dev_priv->irq_lock);
> > +               valleyview_enable_display_irqs(dev_priv);
> > +               spin_unlock_irq(&dev_priv->irq_lock);
> > +
> > +               /*
> > +                * During driver initialization/resume we can avoid restoring the
> > +                * part of the HW/SW state that will be inited anyway explicitly.
> > +                */
> > +               if (dev_priv->power_domains.initializing)
> > +                       return;
> > +
> > +               intel_hpd_init(dev_priv);
> > +
> > +               i915_redisable_vga_power_on(dev_priv->dev);
> > +       }
> >  }
> >
> >  static void chv_pipe_power_well_disable(struct drm_i915_private *dev_priv,
> > @@ -586,6 +603,12 @@ static void chv_pipe_power_well_disable(struct drm_i915_private *dev_priv,
> >                      power_well->data != PIPE_B &&
> >                      power_well->data != PIPE_C);
> >
> > +       if (power_well->data == PIPE_A) {
> > +               spin_lock_irq(&dev_priv->irq_lock);
> > +               valleyview_disable_display_irqs(dev_priv);
> > +               spin_unlock_irq(&dev_priv->irq_lock);
> > +       }
> > +
> >         chv_set_pipe_power_well(dev_priv, power_well, false);
> >
> >         if (power_well->data == PIPE_A)
> > --
> > 2.0.4
> >
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx at lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/intel-gfx
> 
> 
> 
> -- 
> Paulo Zanoni

-- 
Ville Syrjälä
Intel OTC