[Intel-gfx] [PATCH] Partially revert "drm/i915: tune the RC6 threshold for stability"

Chris Wilson chris at chris-wilson.co.uk
Sun Nov 17 11:49:57 CET 2013


On Sat, Nov 16, 2013 at 04:06:01PM +0100, Daniel Vetter wrote:
> This reverts commit 351aa5666d02062b52329bcfe4bcf9d1f882fba9.
> 
> It breaks rc6 on at least one snb machine. Since we don't yet have a
> report for ivb let's keep it there for now.
> 
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=71656
> Cc: Stéphane Marchesin <marcheu at chromium.org>
> Cc: erik at vontaene.de
> Signed-off-by: Daniel Vetter <daniel.vetter at ffwll.ch>

Acked-by: Chris Wilson <chris at chris-wilson.co.uk>

I think maybe the IDLE_HYSTERSIS is more likely to paper over the issues
Stéphane observed without affecting rc6 times as much. Worth
investigating if someone has the time and equipment to do so.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre



More information about the Intel-gfx mailing list