[Intel-gfx] [PATCH] drm/i915: Rebalance runtime pm vs forcewake

[Chris Wilson]

On Fri, Mar 14, 2014 at 04:51:16PM +0100, Daniel Vetter wrote:
> On Fri, Mar 14, 2014 at 08:37:15AM +0000, Chris Wilson wrote:
> > ---
> >  drivers/gpu/drm/i915/i915_drv.c     | 2 +-
> >  drivers/gpu/drm/i915/intel_uncore.c | 9 ++-------
> >  2 files changed, 3 insertions(+), 8 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
> > index 5a0d34c47885..3fbf8aa8d119 100644
> > --- a/drivers/gpu/drm/i915/i915_drv.c
> > +++ b/drivers/gpu/drm/i915/i915_drv.c
> > @@ -845,11 +845,11 @@ static int i915_runtime_suspend(struct device *device)
> >  	struct drm_i915_private *dev_priv = dev->dev_private;
> >  
> >  	WARN_ON(!HAS_RUNTIME_PM(dev));
> > -	assert_force_wake_inactive(dev_priv);
> 
> Why is this necessary? Also I've already pushed a pile of other patches on
> top of all this, so I think a full commit is better. Also gives us an
> excuse to document our flailing here a bit better in a neat commit message
> ... Imo we should also mention that the forcewake_put here isn't really
> perf critical any more (if this is really the case).

I was continuing the conversation with example code... This is, I think,
the simplest method for removing the pm_put from the forcewake timer,
and just wanted to make sure that we were in agreement before writing a
paragraph to explain the problem.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre