[Intel-gfx] [PATCH] drm/i915: Skip force-wake for uncached mmio flush of GGTT writes

[Chris Wilson]

On Mon, Mar 20, 2017 at 12:02:02PM +0200, Mika Kuoppala wrote:
> Chris Wilson <chris at chris-wilson.co.uk> writes:
> 
> > The trick of using an uncached mmio read to ensure that the GGTT writes
> > are flushed does not require us to do the forcewake dance, so avoid it
> > in the hope of reducing the frequency that we do keep the device forced
> > awake.
> >
> > Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> > ---
> >  drivers/gpu/drm/i915/i915_gem.c | 7 +++++--
> >  1 file changed, 5 insertions(+), 2 deletions(-)
> >
> > diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
> > index 5e280e3bfd86..d468300e2f05 100644
> > --- a/drivers/gpu/drm/i915/i915_gem.c
> > +++ b/drivers/gpu/drm/i915/i915_gem.c
> > @@ -3094,8 +3094,11 @@ i915_gem_object_flush_gtt_write_domain(struct drm_i915_gem_object *obj)
> >  	 * system agents we cannot reproduce this behaviour).
> >  	 */
> >  	wmb();
> > -	if (INTEL_GEN(dev_priv) >= 6 && !HAS_LLC(dev_priv))
> > -		POSTING_READ(RING_ACTHD(dev_priv->engine[RCS]->mmio_base));
> > +	if (INTEL_GEN(dev_priv) >= 6 && !HAS_LLC(dev_priv)) {
> > +		spin_lock_irq(&dev_priv->uncore.lock);
> > +		POSTING_READ_FW(RING_ACTHD(dev_priv->engine[RCS]->mmio_base));
> > +		spin_unlock_irq(&dev_priv->uncore.lock);
> > +	}
> >
> 
> Why is it so that the flushing (from gpu side) doesn't need
> the rc6?

It's not GPU, it's GTT. Just the usual thing with our fake pci not
following pci ordering rules.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre