Re: [PATCH 43/55] drm/i915: Refactor activity tracking for requests

[Joonas Lahtinen <joonas.lahtinen@linux.intel.com>]

On ma, 2016-07-25 at 18:32 +0100, Chris Wilson wrote:
>  static void
> -i915_gem_object_retire__write(struct drm_i915_gem_object *obj)
> +i915_gem_object_retire__fence(struct i915_gem_active *active,
> +			      struct drm_i915_gem_request *req)
>  {
> -	GEM_BUG_ON(!__i915_gem_active_is_busy(&obj->last_write));
> -	GEM_BUG_ON(!(obj->active &
> -		     intel_engine_flag(i915_gem_active_get_engine(&obj->last_write,
> -								  &obj->base.dev->struct_mutex))));
> +}
>  

An empty function? Could have at least a comment why currently empty.

> -	i915_gem_active_set(&obj->last_write, NULL);
> -	intel_fb_obj_flush(obj, true, ORIGIN_CS);
> +static void
> +i915_gem_object_retire__write(struct i915_gem_active *active,
> +			      struct drm_i915_gem_request *request)
> +{
> +	intel_fb_obj_flush(container_of(active,
> +					struct drm_i915_gem_object,
> +					last_write),

Add a function, manual container_of are horrible. And do it in the
beginning of a function as a separate line, too.

> +			   true,
> +			   ORIGIN_CS);
>  }
>  
>  static void
> -i915_gem_object_retire__read(struct drm_i915_gem_object *obj, int idx)
> +i915_gem_object_retire__read(struct i915_gem_active *active,
> +			     struct drm_i915_gem_request *request)
>  {
> -	struct intel_engine_cs *engine;
> +	int idx = request->engine->id;
> +	struct drm_i915_gem_object *obj =
> +		container_of(active, struct drm_i915_gem_object, last_read[idx]);

Ditto.

>  	struct i915_vma *vma;
>  
> -	GEM_BUG_ON(!__i915_gem_active_is_busy(&obj->last_read[idx]));
> -	GEM_BUG_ON(!(obj->active & (1 << idx)));
> -
> -	list_del_init(&obj->engine_list[idx]);
> -	i915_gem_active_set(&obj->last_read[idx], NULL);
> -
> -	engine = i915_gem_active_get_engine(&obj->last_write,
> -					    &obj->base.dev->struct_mutex);
> -	if (engine && engine->id == idx)
> -		i915_gem_object_retire__write(obj);
> +	GEM_BUG_ON((obj->active & (1 << idx)) == 0);

BIT() or maybe even ENGINE_MASK() when we have such a beauty. Or do you
intend to make this about something else but engines eventually?

>  
>  	obj->active &= ~(1 << idx);
>  	if (obj->active)
> @@ -2419,15 +2384,13 @@ i915_gem_object_retire__read(struct drm_i915_gem_object *obj, int idx)
>  	 * so that we don't steal from recently used but inactive objects
>  	 * (unless we are forced to ofc!)
>  	 */
> -	list_move_tail(&obj->global_list,
> -		       &to_i915(obj->base.dev)->mm.bound_list);
> +	list_move_tail(&obj->global_list, &request->i915->mm.bound_list);

As a follow-up s/global_list/global_link/?

>  void i915_gem_retire_requests(struct drm_i915_private *dev_priv)
> @@ -2818,8 +2742,7 @@ out:
>  }
>  
>  static int
> -__i915_gem_object_sync(struct drm_i915_gem_object *obj,
> -		       struct drm_i915_gem_request *to,
> +__i915_gem_object_sync(struct drm_i915_gem_request *to,
>  		       struct drm_i915_gem_request *from)
>  {
>  	int ret;
> @@ -2827,9 +2750,6 @@ __i915_gem_object_sync(struct drm_i915_gem_object *obj,
>  	if (to->engine == from->engine)
>  		return 0;
>  
> -	if (i915_gem_request_completed(from))
> -		return 0;
> -

Why remove the early exit?
 
> @@ -172,6 +176,24 @@ static void i915_gem_request_retire(struct drm_i915_gem_request *request)
>  	 */
>  	request->ring->last_retired_head = request->postfix;
>  
> +	/* Walk through the active list, calling retire on each. This allows
> +	 * objects to track their GPU activity and mark themselves as idle
> +	 * when their *last* active request is completed (updating state
> +	 * tracking lists for eviction, active references for GEM, etc).
> +	 *
> +	 * As the ->retire() may free the node, we decouple it first and
> +	 * pass along the auxiliary information (to avoid dereferencing
> +	 * the node after the callback).
> +	 */
> +	list_for_each_entry_safe(active, next, &request->active_list, link) {
> +		prefetchw(next);

Would not this be an improvement to go to list_for_each_entry{,_safe}
rather?

> +
> +		INIT_LIST_HEAD(&active->link);
> +		active->__request = NULL;
> +
> +		active->retire(active, request);
> +	}
> +
>  	i915_gem_request_remove_from_client(request);
>  
>  	if (request->previous_context) {
>  
> @@ -705,10 +723,13 @@ int i915_wait_request(struct drm_i915_gem_request *req)
>  {
>  	int ret;
>  
> -	GEM_BUG_ON(!req);
>  	lockdep_assert_held(&req->i915->drm.struct_mutex);
> +	GEM_BUG_ON(list_empty(&req->link));

Humm, why no waiting on requests without the tracker object? Or then
need to use __i915_wait_request? Kerneldoc might be useful.

>  i915_gem_active_peek(const struct i915_gem_active *active, struct mutex *mutex)
>  {
> -	return active->__request;
> +	struct drm_i915_gem_request *request;
> +
> +	request = active->__request;
> +	if (!request || i915_gem_request_completed(request))
> +		return NULL;

I see early exit was kinda migrated here.

> +
> +	return request;
>  }
>  
>  /**
> @@ -326,13 +360,7 @@ i915_gem_active_peek(const struct i915_gem_active *active, struct mutex *mutex)
>  static inline struct drm_i915_gem_request *
>  i915_gem_active_get(const struct i915_gem_active *active, struct mutex *mutex)
>  {
> -	struct drm_i915_gem_request *request;
> -
> -	request = i915_gem_active_peek(active, mutex);
> -	if (!request || i915_gem_request_completed(request))
> -		return NULL;
> -
> -	return i915_gem_request_get(request);
> +	return i915_gem_request_get(i915_gem_active_peek(active, mutex));

On average looks better with a variable in between and not all
functions chained.

Regards, Joonas
-- 
Joonas Lahtinen
Open Source Technology Center
Intel Corporation
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx