On Fri, Dec 12, 2014 at 03:30:14PM +0530, Deepak S wrote: > > On Monday 08 December 2014 11:57 PM, Mika Kuoppala wrote: > >From: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx> > > > >Calling intel_runtime_pm_put() is illegal from a soft-irq context, so > >revert the crude hack > > > >commit aa0b3b5bb8768c1a6a6788869d9c7015eae7e80c > >Author: Paulo Zanoni <paulo.r.zanoni@xxxxxxxxx> > >Date: Tue Apr 1 14:55:07 2014 -0300 > > > > drm/i915: don't schedule force_wake_timer at gen6_read > > > >and apply the single line corrective instead. > > > >References: https://bugs.freedesktop.org/show_bug.cgi?id=80913 > >Cc: Paulo Zanoni <paulo.r.zanoni@xxxxxxxxx> > >Cc: Daniel Vetter <daniel.vetter@xxxxxxxx> > >Signed-off-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx> > >--- > >@@ -777,12 +772,11 @@ gen6_read##x(struct drm_i915_private *dev_priv, off_t reg, bool trace) { \ > > NEEDS_FORCE_WAKE((dev_priv), (reg))) { \ > > dev_priv->uncore.funcs.force_wake_get(dev_priv, \ > > FORCEWAKE_ALL); \ > >- val = __raw_i915_read##x(dev_priv, reg); \ > >- dev_priv->uncore.funcs.force_wake_put(dev_priv, \ > >- FORCEWAKE_ALL); \ > >- } else { \ > >- val = __raw_i915_read##x(dev_priv, reg); \ > >+ dev_priv->uncore.forcewake_count++; \ > >+ mod_timer_pinned(&dev_priv->uncore. , \ > >+ jiffies + 1); \ > > why timer, we can do a put after register read right? The presumption is that we will do another mmio access requiring the forcewake very shortly, and we want to avoid the forcewake clear/ack cycle. So we defer dropping the forcewake until the end of the kernel context on this cpu (the goal being that as the scheduler switches back to the userspace context, we release the wakelock, it would be great if there was an explicit callback for that...). -Chris -- Chris Wilson, Intel Open Source Technology Centre _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx http://lists.freedesktop.org/mailman/listinfo/intel-gfx
