[PATCH] fence: checking for NULL pointer after dereferencing it.

Xinpeng Liu danielliu861 at gmail.com
Tue Aug 6 18:04:24 UTC 2019


drivers/dma-buf/dma-fence.c:131 in function dma_fence_signal_locked,
if (WARN_ON(!fence)) should be before lockdep_assert_held(fence->lock);
otherwise there is not any meaning.

Signed-off-by: Xinpeng Liu <danielliu861 at gmail.com>
---
 drivers/dma-buf/dma-fence.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/drivers/dma-buf/dma-fence.c b/drivers/dma-buf/dma-fence.c
index 59ac96e..0b11063 100644
--- a/drivers/dma-buf/dma-fence.c
+++ b/drivers/dma-buf/dma-fence.c
@@ -131,10 +131,9 @@ int dma_fence_signal_locked(struct dma_fence *fence)
 	struct dma_fence_cb *cur, *tmp;
 	int ret = 0;
 
-	lockdep_assert_held(fence->lock);
-
 	if (WARN_ON(!fence))
 		return -EINVAL;
+	lockdep_assert_held(fence->lock);
 
 	if (test_and_set_bit(DMA_FENCE_FLAG_SIGNALED_BIT, &fence->flags)) {
 		ret = -EINVAL;
-- 
1.8.3.1



More information about the dri-devel mailing list