Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1432271

[RFC v2 2/3] dma-buf/fence-array: add fence_array_teardown()

Show all headers | View raw

From: Gustavo Padovan <gustavo.padovan@collabora.co.uk>

As the array of fence callbacks held by an active struct fence_array
each has a reference to the struct fence_array, when the owner of the
fence_array is freed it must dispose of the callback references before
it can free the fence_array. This can not happen simply during
fence_release() because of the extra references and so we need a new
function to run before the final fence_put().

Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Christian König <christian.koenig@amd.com>
Signed-off-by: Gustavo Padovan <gustavo.padovan@collabora.co.uk>
---
 drivers/dma-buf/fence-array.c | 25 +++++++++++++++++++++++++
 include/linux/fence-array.h   |  1 +
 2 files changed, 26 insertions(+)

diff --git a/drivers/dma-buf/fence-array.c b/drivers/dma-buf/fence-array.c
index a8731c8..8891357 100644
--- a/drivers/dma-buf/fence-array.c
+++ b/drivers/dma-buf/fence-array.c
@@ -142,3 +142,28 @@ struct fence_array *fence_array_create(int num_fences, struct fence **fences,
 	return array;
 }
 EXPORT_SYMBOL(fence_array_create);
+
+/**
+ * fence_array_teardown - Teardown and clean up a fence array
+ * @array:	[in]	the fence array to teardown
+ *
+ * This function removes callbacks and extra references to the base fence on
+ * the fence_array for each unsignalled fence in the array. It should be called
+ * before calling fence_put() to remove the last reference on a fence_array,
+ * otherwise the fence won't be released due to extra references holded by the
+ * the fences that still have signalling enabled.
+ */
+void fence_array_teardown(struct fence_array *array)
+{
+	struct fence_array_cb *cb = (void *)(&array[1]);
+	int i;
+
+	for (i = 0; i < array->num_fences; i++) {
+		if (fence_is_signaled(array->fences[i]))
+		    continue;
+
+		fence_remove_callback(array->fences[i], &cb[i].cb);
+		fence_put(&array->base);
+	}
+}
+EXPORT_SYMBOL(fence_array_teardown);
diff --git a/include/linux/fence-array.h b/include/linux/fence-array.h
index d2e9f40..9f1b923 100644
--- a/include/linux/fence-array.h
+++ b/include/linux/fence-array.h
@@ -79,5 +79,6 @@ static inline struct fence_array *to_fence_array(struct fence *fence)
 struct fence_array *fence_array_create(int num_fences, struct fence **fences,
 				       u64 context, unsigned seqno,
 				       bool signal_on_any);
+void fence_array_teardown(struct fence_array *array);
 
 #endif /* __LINUX_FENCE_ARRAY_H */
-- 
2.5.5

Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

[RFC v2 0/3] dma-buf/sync_file: rework fences on struct sync_file Gustavo Padovan <gustavo@padovan.org> - 2016-06-27 21:30 +0200
  [RFC v2 1/3] dma-buf/fence-array: add fence_is_array() Gustavo Padovan <gustavo@padovan.org> - 2016-06-27 21:40 +0200
  [RFC v2 2/3] dma-buf/fence-array: add fence_array_teardown() Gustavo Padovan <gustavo@padovan.org> - 2016-06-27 21:40 +0200
    Re: [RFC v2 2/3] dma-buf/fence-array: add fence_array_teardown() Gustavo Padovan <gustavo.padovan@collabora.com> - 2016-06-28 16:20 +0200
      Re: [RFC v2 2/3] dma-buf/fence-array: add fence_array_teardown() Gustavo Padovan <gustavo.padovan@collabora.com> - 2016-06-28 17:20 +0200
  [RFC v2 3/3] dma-buf/sync_file: rework fence storage in struct file Gustavo Padovan <gustavo@padovan.org> - 2016-06-27 21:40 +0200
    Re: [RFC v2 3/3] dma-buf/sync_file: rework fence storage in struct  file Chris Wilson <chris@chris-wilson.co.uk> - 2016-06-28 10:10 +0200
      Re: [RFC v2 3/3] dma-buf/sync_file: rework fence storage in struct  file Gustavo Padovan <gustavo.padovan@collabora.com> - 2016-06-28 16:30 +0200
        Re: [RFC v2 3/3] dma-buf/sync_file: rework fence storage in struct  file Gustavo Padovan <gustavo.padovan@collabora.com> - 2016-06-28 17:40 +0200
        Re: [RFC v2 3/3] dma-buf/sync_file: rework fence storage in struct  file Chris Wilson <chris@chris-wilson.co.uk> - 2016-06-28 18:50 +0200

csiph-web