[PATCH] drm/i915: Do not leak pages when freeing userptr objects

[PATCH] drm/i915: Do not leak pages when freeing userptr objects

[Tvrtko Ursulin]

From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>

sg_alloc_table_from_pages() can build us a table with coalesced ranges which
means we need to iterate over pages and not sg table entries when releasing
page references.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: "Barbalho, Rafael" <rafael.barbalho@intel.com>
---
 drivers/gpu/drm/i915/i915_gem_userptr.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_gem_userptr.c b/drivers/gpu/drm/i915/i915_gem_userptr.c
index d384139..229694d 100644
--- a/drivers/gpu/drm/i915/i915_gem_userptr.c
+++ b/drivers/gpu/drm/i915/i915_gem_userptr.c
@@ -682,15 +682,15 @@ static void
 i915_gem_userptr_put_pages(struct drm_i915_gem_object *obj)
 {
 	struct scatterlist *sg;
-	int i;
+	struct sg_page_iter sg_iter;
 
 	BUG_ON(obj->userptr.work != NULL);
 
 	if (obj->madv != I915_MADV_WILLNEED)
 		obj->dirty = 0;
 
-	for_each_sg(obj->pages->sgl, sg, obj->pages->nents, i) {
-		struct page *page = sg_page(sg);
+	for_each_sg_page(obj->pages->sgl, &sg_iter, obj->pages->nents, 0) {
+		struct page *page = sg_page_iter_page(&sg_iter);
 
 		if (obj->dirty)
 			set_page_dirty(page);
-- 
2.1.0

Re: [PATCH] drm/i915: Do not leak pages when freeing userptr objects

[Barbalho, Rafael]

> -----Original Message-----
> From: Tvrtko Ursulin [mailto:tvrtko.ursulin@linux.intel.com]
> Sent: Friday, September 26, 2014 3:05 PM
> To: Intel-gfx@lists.freedesktop.org
> Cc: Ursulin, Tvrtko; Chris Wilson; Barbalho, Rafael
> Subject: [PATCH] drm/i915: Do not leak pages when freeing userptr objects
> 
> From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
> 
> sg_alloc_table_from_pages() can build us a table with coalesced ranges
> which
> means we need to iterate over pages and not sg table entries when releasing
> page references.
> 
> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
> Cc: Chris Wilson <chris@chris-wilson.co.uk>
> Cc: "Barbalho, Rafael" <rafael.barbalho@intel.com>

Tested-by: Rafael Barbalho <rafael.barbalho@intel.com>

> ---
>  drivers/gpu/drm/i915/i915_gem_userptr.c | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/i915_gem_userptr.c
> b/drivers/gpu/drm/i915/i915_gem_userptr.c
> index d384139..229694d 100644
> --- a/drivers/gpu/drm/i915/i915_gem_userptr.c
> +++ b/drivers/gpu/drm/i915/i915_gem_userptr.c
> @@ -682,15 +682,15 @@ static void
>  i915_gem_userptr_put_pages(struct drm_i915_gem_object *obj)
>  {
>  	struct scatterlist *sg;
> -	int i;
> +	struct sg_page_iter sg_iter;
> 
>  	BUG_ON(obj->userptr.work != NULL);
> 
>  	if (obj->madv != I915_MADV_WILLNEED)
>  		obj->dirty = 0;
> 
> -	for_each_sg(obj->pages->sgl, sg, obj->pages->nents, i) {
> -		struct page *page = sg_page(sg);
> +	for_each_sg_page(obj->pages->sgl, &sg_iter, obj->pages->nents, 0)
> {
> +		struct page *page = sg_page_iter_page(&sg_iter);
> 
>  		if (obj->dirty)
>  			set_page_dirty(page);
> --
> 2.1.0

Re: [PATCH] drm/i915: Do not leak pages when freeing userptr objects

[Chris Wilson]

On Fri, Sep 26, 2014 at 03:05:22PM +0100, Tvrtko Ursulin wrote:
> From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
> 
> sg_alloc_table_from_pages() can build us a table with coalesced ranges which
> means we need to iterate over pages and not sg table entries when releasing
> page references.
> 
> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
> Cc: Chris Wilson <chris@chris-wilson.co.uk>
> Cc: "Barbalho, Rafael" <rafael.barbalho@intel.com>

Oh that's fun. I blame Imre for the recent invention of for_each_sg_page()!
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: stable@vger.kernel.org
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre

Re: [PATCH] drm/i915: Do not leak pages when freeing userptr objects

[Jani Nikula]

On Fri, 26 Sep 2014, Chris Wilson <chris@chris-wilson.co.uk> wrote:
> On Fri, Sep 26, 2014 at 03:05:22PM +0100, Tvrtko Ursulin wrote:
>> From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
>> 
>> sg_alloc_table_from_pages() can build us a table with coalesced ranges which
>> means we need to iterate over pages and not sg table entries when releasing
>> page references.
>> 
>> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
>> Cc: Chris Wilson <chris@chris-wilson.co.uk>
>> Cc: "Barbalho, Rafael" <rafael.barbalho@intel.com>
>
> Oh that's fun. I blame Imre for the recent invention of for_each_sg_page()!
> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
> Cc: stable@vger.kernel.org

Pushed to drm-intel-fixes, with the unused struct scatterlist *sg
variable removed. Thanks for the patch, review, and testing.

BR,
Jani.


> -Chris
>
> -- 
> Chris Wilson, Intel Open Source Technology Centre
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx

-- 
Jani Nikula, Intel Open Source Technology Center

Re: [PATCH] drm/i915: Do not leak pages when freeing userptr objects

[Daniel Vetter]

On Fri, Sep 26, 2014 at 03:55:51PM +0100, Chris Wilson wrote:
> On Fri, Sep 26, 2014 at 03:05:22PM +0100, Tvrtko Ursulin wrote:
> > From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
> > 
> > sg_alloc_table_from_pages() can build us a table with coalesced ranges which
> > means we need to iterate over pages and not sg table entries when releasing
> > page references.
> > 
> > Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
> > Cc: Chris Wilson <chris@chris-wilson.co.uk>
> > Cc: "Barbalho, Rafael" <rafael.barbalho@intel.com>
> 
> Oh that's fun. I blame Imre for the recent invention of for_each_sg_page()!
> Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
> Cc: stable@vger.kernel.org

Again merged to dinf for vetting time, thanks for patch&review.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch