Re: [PATCH 2/4] drm/i915: Only apply the mb() when flushing the GTT domain during a finish

[Jesse Barnes <jbarnes@virtuousgeek.org>]

On Tue,  9 Oct 2012 19:24:38 +0100
Chris Wilson <chris@chris-wilson.co.uk> wrote:

> Now that we seem to have brought order to the GTT barriers, the last one
> to review is the terminal barrier before we unbind the buffer from the
> GTT. This needs to only be performed if the buffer still resides in the
> GTT domain, and so we can skip some needless barriers otherwise.
> 
> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> ---
>  drivers/gpu/drm/i915/i915_gem.c |    6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
> index 3c4577b..ed8d21a 100644
> --- a/drivers/gpu/drm/i915/i915_gem.c
> +++ b/drivers/gpu/drm/i915/i915_gem.c
> @@ -2526,15 +2526,15 @@ static void i915_gem_object_finish_gtt(struct drm_i915_gem_object *obj)
>  {
>  	u32 old_write_domain, old_read_domains;
>  
> -	/* Act a barrier for all accesses through the GTT */
> -	mb();
> -
>  	/* Force a pagefault for domain tracking on next user access */
>  	i915_gem_release_mmap(obj);
>  
>  	if ((obj->base.read_domains & I915_GEM_DOMAIN_GTT) == 0)
>  		return;
>  
> +	/* Wait for any direct GTT access to complete */
> +	mb();
> +
>  	old_read_domains = obj->base.read_domains;
>  	old_write_domain = obj->base.write_domain;
>  

Yeah this looks like a better place to put it.  You're trying to
serialize this against a subsequent fence or map operation?

Reviewed-by: Jesse Barnes <jbarnes@virtuousgeek.org>

-- 
Jesse Barnes, Intel Open Source Technology Center