From c8a5d5ea3ba6a18958f8d76430e4cd68eea33943 Mon Sep 17 00:00:00 2001 From: Dave Airlie Date: Thu, 15 Jun 2023 12:22:11 +1000 Subject: [PATCH] nouveau: fix client work fence deletion race This seems to have existed for ever but is now more apparant after commit 9bff18d13473 ("drm/ttm: use per BO cleanup workers") My analysis: two threads are running, one in the irq signalling the fence, in dma_fence_signal_timestamp_locked, it has done the DMA_FENCE_FLAG_SIGNALLED_BIT setting, but hasn't yet reached the callbacks. The second thread in nouveau_cli_work_ready, where it sees the fence is signalled, so then puts the fence, cleanups the object and frees the work item, which contains the callback. Thread one goes again and tries to call the callback and causes the use-after-free. Proposed fix: lock the fence signalled check in nouveau_cli_work_ready, so either the callbacks are done or the memory is freed. Reviewed-by: Karol Herbst Fixes: 11e451e74050 ("drm/nouveau: remove fence wait code from deferred client work handler") Cc: stable@vger.kernel.org Signed-off-by: Dave Airlie Link: https://lore.kernel.org/dri-devel/20230615024008.1600281-1-airlied@gmail.com/ --- drivers/gpu/drm/nouveau/nouveau_drm.c | 14 ++++++++++---- 1 file changed, 10 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/nouveau/nouveau_drm.c b/drivers/gpu/drm/nouveau/nouveau_drm.c index cc7c5b4a05fd87..7aac9384600ed4 100644 --- a/drivers/gpu/drm/nouveau/nouveau_drm.c +++ b/drivers/gpu/drm/nouveau/nouveau_drm.c @@ -137,10 +137,16 @@ nouveau_name(struct drm_device *dev) static inline bool nouveau_cli_work_ready(struct dma_fence *fence) { - if (!dma_fence_is_signaled(fence)) - return false; - dma_fence_put(fence); - return true; + bool ret = true; + + spin_lock_irq(fence->lock); + if (!dma_fence_is_signaled_locked(fence)) + ret = false; + spin_unlock_irq(fence->lock); + + if (ret == true) + dma_fence_put(fence); + return ret; } static void