[Intel-gfx] [PATCH 4/5] drm/i915: Increase initial busyspin limit

Chris Wilson chris at chris-wilson.co.uk
Thu Aug 2 14:54:30 UTC 2018


Quoting Tvrtko Ursulin (2018-08-02 15:47:37)
> 
> On 28/07/2018 17:46, Chris Wilson wrote:
> > Here we bump the busyspin for the current request to avoid sleeping and
> > the cost of both idling and downclocking the CPU.
> > 
> > Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> > Cc: Sagar Kamble <sagar.a.kamble at intel.com>
> > Cc: Eero Tamminen <eero.t.tamminen at intel.com>
> > Cc: Francisco Jerez <currojerez at riseup.net>
> > Cc: Tvrtko Ursulin <tvrtko.ursulin at intel.com>
> > Cc: Ben Widawsky <ben at bwidawsk.net>
> > Cc: Joonas Lahtinen <joonas.lahtinen at linux.intel.com>
> > Cc: MichaƂ Winiarski <michal.winiarski at intel.com>
> > ---
> >   drivers/gpu/drm/i915/Kconfig.profile | 2 +-
> >   1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/Kconfig.profile b/drivers/gpu/drm/i915/Kconfig.profile
> > index de394dea4a14..54e0ba4051e7 100644
> > --- a/drivers/gpu/drm/i915/Kconfig.profile
> > +++ b/drivers/gpu/drm/i915/Kconfig.profile
> > @@ -1,6 +1,6 @@
> >   config DRM_I915_SPIN_REQUEST_IRQ
> >       int
> > -     default 5 # microseconds
> > +     default 250 # microseconds
> >       help
> >         Before sleeping waiting for a request (GPU operation) to complete,
> >         we may spend some time polling for its completion. As the IRQ may
> > 
> 
> But the histograms from previous patch do not suggest 250us busy spin 
> gets us into interesting territory. And we have to know the distribution 
> between IRQ and CS spins.

This is for a different problem than presented in those histograms. This
is to hide the reclocking that occurred if we sleep.

Previous justification for the initial spin was to try and hide the irq
setup costs, this is to try and hide the sleep costs.

The purpose of this patch was just to dip the toe into the water of what
is possible and solicit feedback since we are still waiting for the
panacea patches to improve cpufreq. (Why it had such a lackluster
justification.)
-Chris


More information about the Intel-gfx mailing list