[Intel-gfx] [PATCH 50/70] drm/i915: The argument for postfix is redundant

Daniel Vetter daniel at ffwll.ch
Fri Apr 10 01:53:02 PDT 2015


On Tue, Apr 07, 2015 at 04:21:14PM +0100, Chris Wilson wrote:
> We are conservative on the amount of free space available in the ring to
> avoid overruning the potential MI_INTERRUPT after the seqno write.
> Further undermining the justification for the change was that it was
> applied incorrectly.
> 
> Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>

Hm, where are we conservative with our estimates? Could we just wait for
req->head instead? And I don't see what's been implemented wrongly with
postfix.

Looking at

commit a71d8d94525e8fd855c0466fb586ae1cb008f3a2
Author: Chris Wilson <chris at chris-wilson.co.uk>
Date:   Wed Feb 15 11:25:36 2012 +0000

    drm/i915: Record the tail at each request and use it to estimate the head

->postfix does get updated where ->tail was, Nick just renamed it from
->tail to ->postfix since execlist used ->tail with a different meaning.

Or do I miss something?

> diff --git a/drivers/gpu/drm/i915/intel_dvo.c b/drivers/gpu/drm/i915/intel_dvo.c
> index 9a27ec7100ef..f45caa6af7d2 100644
> --- a/drivers/gpu/drm/i915/intel_dvo.c
> +++ b/drivers/gpu/drm/i915/intel_dvo.c
> @@ -496,7 +496,7 @@ void intel_dvo_init(struct drm_device *dev)
>  		int gpio;
>  		bool dvoinit;
>  		enum pipe pipe;
> -		uint32_t dpll[2];
> +		uint32_t dpll[I915_MAX_PIPES];

Unrelated change and there's indeed only 2 dvo plls ever on gen2.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch


More information about the Intel-gfx mailing list