diff mbox series

drm/i915: Initialise the obj->rcu head

Message ID 20181109090311.15321-1-chris@chris-wilson.co.uk (mailing list archive)
State New, archived
Headers show
Series drm/i915: Initialise the obj->rcu head | expand

Commit Message

Chris Wilson Nov. 9, 2018, 9:03 a.m. UTC
Make the rcu_head known to the system, in particular for debugobjects.
And having declared it for debugobjects, we need to tidy up afterwards.

v2: mark the obj->rcu as being destroy when we reuse its location for
the freed list.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=108691
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
---
 drivers/gpu/drm/i915/i915_gem.c | 9 +++++++++
 1 file changed, 9 insertions(+)

Comments

Mika Kuoppala Nov. 9, 2018, 10:32 a.m. UTC | #1
Chris Wilson <chris@chris-wilson.co.uk> writes:

> Make the rcu_head known to the system, in particular for debugobjects.
> And having declared it for debugobjects, we need to tidy up afterwards.
>
> v2: mark the obj->rcu as being destroy when we reuse its location for
> the freed list.
>
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=108691
> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>

> ---
>  drivers/gpu/drm/i915/i915_gem.c | 9 +++++++++
>  1 file changed, 9 insertions(+)
>
> diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
> index dc120b5d8e05..1c9a41ea6834 100644
> --- a/drivers/gpu/drm/i915/i915_gem.c
> +++ b/drivers/gpu/drm/i915/i915_gem.c
> @@ -4739,6 +4739,8 @@ void i915_gem_object_init(struct drm_i915_gem_object *obj,
>  	INIT_LIST_HEAD(&obj->lut_list);
>  	INIT_LIST_HEAD(&obj->batch_pool_link);
>  
> +	init_rcu_head(&obj->rcu);
> +
>  	obj->ops = ops;
>  
>  	reservation_object_init(&obj->__builtin_resv);
> @@ -5005,6 +5007,13 @@ static void __i915_gem_free_object_rcu(struct rcu_head *head)
>  		container_of(head, typeof(*obj), rcu);
>  	struct drm_i915_private *i915 = to_i915(obj->base.dev);
>  
> +	/*
> +	 * We reuse obj->rcu for the freed list, so we had better not treat
> +	 * is like a rcu_head from this point forwards. And we expect all
> +	 * objects to be freed via this path.
> +	 */
> +	destroy_rcu_head(&obj->rcu);
> +
>  	/*
>  	 * Since we require blocking on struct_mutex to unbind the freed
>  	 * object from the GPU before releasing resources back to the
> -- 
> 2.19.1
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Chris Wilson Nov. 9, 2018, 10:49 a.m. UTC | #2
Quoting Mika Kuoppala (2018-11-09 10:32:30)
> Chris Wilson <chris@chris-wilson.co.uk> writes:
> 
> > Make the rcu_head known to the system, in particular for debugobjects.
> > And having declared it for debugobjects, we need to tidy up afterwards.
> >
> > v2: mark the obj->rcu as being destroy when we reuse its location for
> > the freed list.
> >
> > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=108691
> > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> 
> Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>

Merged the first as it will close a couple of bugs, care to check over
the second? :)
-Chris
diff mbox series

Patch

diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c
index dc120b5d8e05..1c9a41ea6834 100644
--- a/drivers/gpu/drm/i915/i915_gem.c
+++ b/drivers/gpu/drm/i915/i915_gem.c
@@ -4739,6 +4739,8 @@  void i915_gem_object_init(struct drm_i915_gem_object *obj,
 	INIT_LIST_HEAD(&obj->lut_list);
 	INIT_LIST_HEAD(&obj->batch_pool_link);
 
+	init_rcu_head(&obj->rcu);
+
 	obj->ops = ops;
 
 	reservation_object_init(&obj->__builtin_resv);
@@ -5005,6 +5007,13 @@  static void __i915_gem_free_object_rcu(struct rcu_head *head)
 		container_of(head, typeof(*obj), rcu);
 	struct drm_i915_private *i915 = to_i915(obj->base.dev);
 
+	/*
+	 * We reuse obj->rcu for the freed list, so we had better not treat
+	 * is like a rcu_head from this point forwards. And we expect all
+	 * objects to be freed via this path.
+	 */
+	destroy_rcu_head(&obj->rcu);
+
 	/*
 	 * Since we require blocking on struct_mutex to unbind the freed
 	 * object from the GPU before releasing resources back to the