[Intel-gfx] [PATCH 1/2] drm/i915: Tidy workaround batch buffer emission

[Chris Wilson]

On Thu, Feb 16, 2017 at 07:59:30AM +0000, Tvrtko Ursulin wrote:
> 
> On 15/02/2017 21:18, Chris Wilson wrote:
> >Reviewed-by: Chris Wilson <chris at chris-wilson.co.uk>
> 
> Thanks. So you think it is worth it?

Yes. As you say, it brings it into line with the rest of the command
emission sequences - and using a consistent pattern is worth the
investment.

> I am just annoyed with the
> BUG_ON. I don't like to add them, but GEM_BUG_ON seemed to weak.

That came across in your choice. At the point we hit the BUG_ON, the
system is probably/should be dead already.

> On
> balance it felt that it is OK. Overall I think it is easier to
> follow the flow from the code with this organisation and it is more
> compact in all respects.

gen8_emit_pipe_control() is all the justification I need. That makes
the benefits very clear.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre