[Acked/cmt] Ack: [Precise][PATCH 0/5] GPU hangs regression on Precise

[Andy Whitcroft]

On Tue, Apr 09, 2013 at 09:21:01AM -0700, Brad Figg wrote:
> On 04/09/2013 09:06 AM, Luis Henriques wrote:
> > BugLink: http://bugs.launchpad.net/bugs/1140716
> > 
> > As per a comment by Chris Wilson in the this bug report, the following
> > commits should fix this issue:
> > 
> > 3ac7831314eba873d60b58718123c503f6961337 drm/i915: PIPE_CONTROL TLB invalidate requires CS stall
> > 9a28977181724ebbd9bdc45291cf29da55a729ee drm/i915: TLB invalidation with MI_FLUSH_DW requires a post-sync op v3
> > 7d54a904285b6e780291b91a518267bec5591913 drm/i915: Apply post-sync write for pipe control invalidates
> > 
> > Commit 9a28977181724ebbd9bdc45291cf29da55a729ee had to be backported to
> > Precise, but the other two are clean cherry-picks if the following two
> > extra commits are picked as well:
> > 
> > 97f209bcfc0c5db08d9badf8cbafd489f22a6e44 drm/i915: "Flush Me Harder" required on gen6+
> > cc0f6398225ffd2b890ff83eafe212b1ae863cad drm/i915: PIPE_CONTROL_TLB_INVALIDATE
> > 
> > Ben Widawsky (1):
> >   drm/i915: PIPE_CONTROL_TLB_INVALIDATE
> > 
> > Chris Wilson (1):
> >   drm/i915: Apply post-sync write for pipe control invalidates
> > 
> > Daniel Vetter (1):
> >   drm/i915: "Flush Me Harder" required on gen6+
> > 
> > Jesse Barnes (2):
> >   drm/i915: TLB invalidation with MI_FLUSH_DW requires a post-sync op v3
> >   drm/i915: PIPE_CONTROL TLB invalidate requires CS stall
> > 
> >  drivers/gpu/drm/i915/i915_reg.h         |  9 +++--
> >  drivers/gpu/drm/i915/intel_ringbuffer.c | 62 ++++++++++++++++++++++++---------
> >  drivers/gpu/drm/i915/intel_ringbuffer.h |  2 ++
> >  3 files changed, 54 insertions(+), 19 deletions(-)
> > 
> 
> These patches make me a little nervous. However, they are the recommended
> solution by one of the upstream devs for a regression. I think we should
> take these but keep a close eye on possible regressions that these might
> produce.
> 
> If you don't feel I'm being conservative enough feel free to disagree.

Same comments as per quantal.

Acked-by: Andy Whitcroft <apw at canonical.com>

-apw