Re: [PATCH 2/2] drm/i915: use semaphores for the display plane

[Chris Wilson <chris@chris-wilson.co.uk>]

On Thu, 22 Mar 2012 09:07:09 -0700, Ben Widawsky <ben@bwidawsk.net> wrote:
> On Thu, 22 Mar 2012 09:56:07 +0000
> Chris Wilson <chris@chris-wilson.co.uk> wrote:
> 
> > On Wed, 21 Mar 2012 17:19:13 -0700, Ben Widawsky <ben@bwidawsk.net>
> > wrote:
> > >   * For the display plane, we want to be in the GTT but out of any
> > > write
> > > - * domains. So in many ways this looks like set_to_gtt_domain()
> > > apart from the
> > > - * ability to pipeline the waits, pinning and any additional
> > > subtleties
> > > - * that may differentiate the display plane from ordinary buffers.
> > > + * domains. So in many ways this looks like set_to_gtt_domain().
> > ...apart from the whole pinning and pipelining. It looks less like
> > set-to-gtt-domain over time.
> > 
> > Not an improvement. I'll be the first to admit that it is not a great
> > comment, but at least it does try to capture why we don't just treat
> > the display plane as GTT. A better comment would explain our concept
> > of display plane and pipelining.
> 
> Separate patch though, don't you think? And it is now pipe-lining the
> waits, so I'm confused why you don't think it's an improvement.

Right. So how does removing the hint of why pin-to-display is different
from set-to-gtt when it now performs said function help improve the
comment? I completely agree that changing that comment is outside of the
scope of this patch.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre