drm/syncobj: handle NULL fence in syncobj_eventfd_entry_func

[ Upstream commit 2aa6f5b ]

During syncobj_eventfd_entry_func, dma_fence_chain_find_seqno may set
the fence to NULL if the given seqno is signaled and a later seqno has
already been submitted. In that case, the eventfd should be signaled
immediately which currently does not happen.

This is a similar issue to the one addressed by commit b19926d
("drm/syncobj: Deal with signalled fences in drm_syncobj_find_fence.").

As a fix, if the return value of dma_fence_chain_find_seqno indicates
success but it sets the fence to NULL, we will assign a stub fence to
ensure the following code still signals the eventfd.

v1 -> v2: assign a stub fence instead of signaling the eventfd

Signed-off-by: Erik Kurzinger <ekurzinger@nvidia.com>
Fixes: c7a4722 ("drm/syncobj: add IOCTL to register an eventfd")
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patchwork.freedesktop.org/patch/msgid/20240221184527.37667-1-ekurzinger@nvidia.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

drivers/gpu/drm/drm_syncobj.c

@@ -1380,10 +1380,21 @@ syncobj_eventfd_entry_func(struct drm_syncobj *syncobj,
 
 	/* This happens inside the syncobj lock */
 	fence = dma_fence_get(rcu_dereference_protected(syncobj->fence, 1));
+	if (!fence)
+		return;
+
 	ret = dma_fence_chain_find_seqno(&fence, entry->point);
-	if (ret != 0 || !fence) {
+	if (ret != 0) {
+		/* The given seqno has not been submitted yet. */
 		dma_fence_put(fence);
 		return;
+	} else if (!fence) {
+		/* If dma_fence_chain_find_seqno returns 0 but sets the fence
+		 * to NULL, it implies that the given seqno is signaled and a
+		 * later seqno has already been submitted. Assign a stub fence
+		 * so that the eventfd still gets signaled below.
+		 */
+		fence = dma_fence_get_stub();
 	}
 
 	list_del_init(&entry->node);