[Intel-gfx] [PATCH 13/14] drm/i915: Use vlv display irq setup code for chv

[Paulo Zanoni]

2014-10-31 8:04 GMT-02:00 Ville Syrjälä <ville.syrjala at linux.intel.com>:
> On Thu, Oct 30, 2014 at 06:41:11PM -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>
>> >
>> > Throw away the hand rolled display irq setup code on chv, and instead
>> > just call vlv_display_irq_postinstall() and vlv_display_irq_uninstall().
>> >
>> > Signed-off-by: Ville Syrjälä <ville.syrjala at linux.intel.com>
>> > ---
>> >  drivers/gpu/drm/i915/i915_irq.c | 37 ++-----------------------------------
>> >  1 file changed, 2 insertions(+), 35 deletions(-)
>> >
>> > diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
>> > index 628a129..722f73c 100644
>> > --- a/drivers/gpu/drm/i915/i915_irq.c
>> > +++ b/drivers/gpu/drm/i915/i915_irq.c
>> > @@ -3540,34 +3540,8 @@ static int gen8_irq_postinstall(struct drm_device *dev)
>> >  static int cherryview_irq_postinstall(struct drm_device *dev)
>> >  {
>> >         struct drm_i915_private *dev_priv = dev->dev_private;
>> > -       u32 enable_mask = I915_DISPLAY_PORT_INTERRUPT |
>> > -               I915_DISPLAY_PIPE_A_EVENT_INTERRUPT |
>> > -               I915_DISPLAY_PIPE_B_EVENT_INTERRUPT |
>> > -               I915_DISPLAY_PIPE_C_EVENT_INTERRUPT;
>> > -       u32 pipestat_enable = PLANE_FLIP_DONE_INT_STATUS_VLV |
>> > -               PIPE_CRC_DONE_INTERRUPT_STATUS;
>> > -       int pipe;
>> > -
>> > -       /*
>> > -        * Leave vblank interrupts masked initially.  enable/disable will
>> > -        * toggle them based on usage.
>> > -        */
>> > -       dev_priv->irq_mask = ~enable_mask;
>> > -
>> > -       for_each_pipe(dev_priv, pipe)
>> > -               I915_WRITE(PIPESTAT(pipe), 0xffff);
>> > -
>> > -       spin_lock_irq(&dev_priv->irq_lock);
>> > -       i915_enable_pipestat(dev_priv, PIPE_A, PIPE_GMBUS_INTERRUPT_STATUS);
>> > -       for_each_pipe(dev_priv, pipe)
>> > -               i915_enable_pipestat(dev_priv, pipe, pipestat_enable);
>> > -       spin_unlock_irq(&dev_priv->irq_lock);
>> >
>> > -       I915_WRITE(VLV_IIR, 0xffffffff);
>> > -       I915_WRITE(VLV_IIR, 0xffffffff);
>> > -       I915_WRITE(VLV_IER, enable_mask);
>> > -       I915_WRITE(VLV_IMR, dev_priv->irq_mask);
>> > -       POSTING_READ(VLV_IMR);
>> > +       vlv_display_irq_postinstall(dev_priv);
>>
>> This chunk changes the order of stuff a little bit, but seems mostly
>> equivalent. I'll consider it's ok.
>>
>> >
>> >         gen8_gt_irq_postinstall(dev_priv);
>> >
>> > @@ -3620,7 +3594,6 @@ static void valleyview_irq_uninstall(struct drm_device *dev)
>> >  static void cherryview_irq_uninstall(struct drm_device *dev)
>> >  {
>> >         struct drm_i915_private *dev_priv = dev->dev_private;
>> > -       int pipe;
>> >
>> >         if (!dev_priv)
>> >                 return;
>> > @@ -3632,13 +3605,7 @@ static void cherryview_irq_uninstall(struct drm_device *dev)
>> >
>> >         GEN5_IRQ_RESET(GEN8_PCU_);
>> >
>> > -       I915_WRITE(PORT_HOTPLUG_EN, 0);
>> > -       I915_WRITE(PORT_HOTPLUG_STAT, I915_READ(PORT_HOTPLUG_STAT));
>> > -
>> > -       for_each_pipe(dev_priv, pipe)
>> > -               I915_WRITE(PIPESTAT(pipe), 0xffff);
>> > -
>> > -       GEN5_IRQ_RESET(VLV_);
>> > +       vlv_display_irq_uninstall(dev_priv);
>>
>> The perfect match for the code you removed seems to be
>> vlv_display_irq_reset(). Why use vlv_display_irq_uninstall() instead?
>
> Becasue vlv uses it too :) I suppose the idea is to disable the
> interrupts the same way they got enabled for symmetry. So if the display
> interrupts got enabled by valleyview_display_irqs_install() we should
> uninstall them using valleyview_display_irqs_uninstall(). And also
> irq_mask gets zeroed there to accurately reflect the state of the
> VLV_IER/IMR registers. But yeah we could just call
> vlv_display_irq_reset() and get the same hardware effect.
>
> I'm actually thinking we should drop the .irq_preinstall() call from
> intel_runtime_pm_enable_interrupts() since IIRC we don't even
> unregister the irq handler itself there, and then add a bit of
> irq_mask and whatnot sanity checking to .irq_postinstall().

This whole VLV/CHV IRQ enabling is really different form the other
Gens and confusing... Since we're unifying things with the patch, I
guess fixing the confusion will be easier with this one applied...

Needs patch 10 applied. With that: Reviewed-by: Paulo Zanoni
<paulo.r.zanoni at intel.com>

>
>>
>> >  }
>> >
>> >  static void ironlake_irq_uninstall(struct drm_device *dev)
>> > --
>> > 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



-- 
Paulo Zanoni