The fence lock is part of the queue, therefore in the current design anything locking the fence should then also hold a ref to the queue to prevent the queue from being freed.
However, currently it looks like we signal the fence and then drop the queue ref, but if something is waiting on the fence, the waiter is kicked to wake up at some later point, where upon waking up it first grabs the lock before checking the fence state. But if we have already dropped the queue ref, then the lock might already be freed as part of the queue, leading to uaf.
To prevent this, move the fence lock into the fence itself so we don't run into lifetime issues. Alternative might be to have device level lock, or only release the queue in the fence release callback, however that might require pushing to another worker to avoid locking issues.
Fixes: dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs") References: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/2454 References: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/2342 References: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/2020 Signed-off-by: Matthew Auld matthew.auld@intel.com Cc: Matthew Brost matthew.brost@intel.com Cc: stable@vger.kernel.org # v6.8+ --- drivers/gpu/drm/xe/xe_exec_queue.c | 1 - drivers/gpu/drm/xe/xe_exec_queue_types.h | 2 -- drivers/gpu/drm/xe/xe_preempt_fence.c | 3 ++- drivers/gpu/drm/xe/xe_preempt_fence_types.h | 2 ++ 4 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/gpu/drm/xe/xe_exec_queue.c b/drivers/gpu/drm/xe/xe_exec_queue.c index 971e1234b8ea..0f610d273fb6 100644 --- a/drivers/gpu/drm/xe/xe_exec_queue.c +++ b/drivers/gpu/drm/xe/xe_exec_queue.c @@ -614,7 +614,6 @@ int xe_exec_queue_create_ioctl(struct drm_device *dev, void *data,
if (xe_vm_in_preempt_fence_mode(vm)) { q->lr.context = dma_fence_context_alloc(1); - spin_lock_init(&q->lr.lock);
err = xe_vm_add_compute_exec_queue(vm, q); if (XE_IOCTL_DBG(xe, err)) diff --git a/drivers/gpu/drm/xe/xe_exec_queue_types.h b/drivers/gpu/drm/xe/xe_exec_queue_types.h index 1408b02eea53..fc2a1a20b7e4 100644 --- a/drivers/gpu/drm/xe/xe_exec_queue_types.h +++ b/drivers/gpu/drm/xe/xe_exec_queue_types.h @@ -126,8 +126,6 @@ struct xe_exec_queue { u32 seqno; /** @lr.link: link into VM's list of exec queues */ struct list_head link; - /** @lr.lock: preemption fences lock */ - spinlock_t lock; } lr;
/** @ops: submission backend exec queue operations */ diff --git a/drivers/gpu/drm/xe/xe_preempt_fence.c b/drivers/gpu/drm/xe/xe_preempt_fence.c index 56e709d2fb30..83fbeea5aa20 100644 --- a/drivers/gpu/drm/xe/xe_preempt_fence.c +++ b/drivers/gpu/drm/xe/xe_preempt_fence.c @@ -134,8 +134,9 @@ xe_preempt_fence_arm(struct xe_preempt_fence *pfence, struct xe_exec_queue *q, { list_del_init(&pfence->link); pfence->q = xe_exec_queue_get(q); + spin_lock_init(&pfence->lock); dma_fence_init(&pfence->base, &preempt_fence_ops, - &q->lr.lock, context, seqno); + &pfence->lock, context, seqno);
return &pfence->base; } diff --git a/drivers/gpu/drm/xe/xe_preempt_fence_types.h b/drivers/gpu/drm/xe/xe_preempt_fence_types.h index b54b5c29b533..312c3372a49f 100644 --- a/drivers/gpu/drm/xe/xe_preempt_fence_types.h +++ b/drivers/gpu/drm/xe/xe_preempt_fence_types.h @@ -25,6 +25,8 @@ struct xe_preempt_fence { struct xe_exec_queue *q; /** @preempt_work: work struct which issues preemption */ struct work_struct preempt_work; + /** @lock: dma-fence fence lock */ + spinlock_t lock; /** @error: preempt fence is in error state */ int error; };
On Wed, Aug 14, 2024 at 12:01:30PM +0100, Matthew Auld wrote:
The fence lock is part of the queue, therefore in the current design anything locking the fence should then also hold a ref to the queue to prevent the queue from being freed.
However, currently it looks like we signal the fence and then drop the queue ref, but if something is waiting on the fence, the waiter is kicked to wake up at some later point, where upon waking up it first grabs the lock before checking the fence state. But if we have already dropped the queue ref, then the lock might already be freed as part of the queue, leading to uaf.
To prevent this, move the fence lock into the fence itself so we don't run into lifetime issues. Alternative might be to have device level lock, or only release the queue in the fence release callback, however that might require pushing to another worker to avoid locking issues.
Fixes: dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs") References: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/2454 References: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/2342 References: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/2020 Signed-off-by: Matthew Auld matthew.auld@intel.com Cc: Matthew Brost matthew.brost@intel.com
Good catch. This is indeed a problem, and you have it coded probably in the safest way possible.
With that: Reviewed-by: Matthew Brost matthew.brost@intel.com
Cc: stable@vger.kernel.org # v6.8+
drivers/gpu/drm/xe/xe_exec_queue.c | 1 - drivers/gpu/drm/xe/xe_exec_queue_types.h | 2 -- drivers/gpu/drm/xe/xe_preempt_fence.c | 3 ++- drivers/gpu/drm/xe/xe_preempt_fence_types.h | 2 ++ 4 files changed, 4 insertions(+), 4 deletions(-)
diff --git a/drivers/gpu/drm/xe/xe_exec_queue.c b/drivers/gpu/drm/xe/xe_exec_queue.c index 971e1234b8ea..0f610d273fb6 100644 --- a/drivers/gpu/drm/xe/xe_exec_queue.c +++ b/drivers/gpu/drm/xe/xe_exec_queue.c @@ -614,7 +614,6 @@ int xe_exec_queue_create_ioctl(struct drm_device *dev, void *data, if (xe_vm_in_preempt_fence_mode(vm)) { q->lr.context = dma_fence_context_alloc(1);
spin_lock_init(&q->lr.lock);err = xe_vm_add_compute_exec_queue(vm, q); if (XE_IOCTL_DBG(xe, err)) diff --git a/drivers/gpu/drm/xe/xe_exec_queue_types.h b/drivers/gpu/drm/xe/xe_exec_queue_types.h index 1408b02eea53..fc2a1a20b7e4 100644 --- a/drivers/gpu/drm/xe/xe_exec_queue_types.h +++ b/drivers/gpu/drm/xe/xe_exec_queue_types.h @@ -126,8 +126,6 @@ struct xe_exec_queue { u32 seqno; /** @lr.link: link into VM's list of exec queues */ struct list_head link;
/** @lr.lock: preemption fences lock */spinlock_t lock;/** @ops: submission backend exec queue operations */ diff --git a/drivers/gpu/drm/xe/xe_preempt_fence.c b/drivers/gpu/drm/xe/xe_preempt_fence.c index 56e709d2fb30..83fbeea5aa20 100644 --- a/drivers/gpu/drm/xe/xe_preempt_fence.c +++ b/drivers/gpu/drm/xe/xe_preempt_fence.c @@ -134,8 +134,9 @@ xe_preempt_fence_arm(struct xe_preempt_fence *pfence, struct xe_exec_queue *q, { list_del_init(&pfence->link); pfence->q = xe_exec_queue_get(q);
- spin_lock_init(&pfence->lock); dma_fence_init(&pfence->base, &preempt_fence_ops,
&q->lr.lock, context, seqno);
&pfence->lock, context, seqno);return &pfence->base; } diff --git a/drivers/gpu/drm/xe/xe_preempt_fence_types.h b/drivers/gpu/drm/xe/xe_preempt_fence_types.h index b54b5c29b533..312c3372a49f 100644 --- a/drivers/gpu/drm/xe/xe_preempt_fence_types.h +++ b/drivers/gpu/drm/xe/xe_preempt_fence_types.h @@ -25,6 +25,8 @@ struct xe_preempt_fence { struct xe_exec_queue *q; /** @preempt_work: work struct which issues preemption */ struct work_struct preempt_work;
- /** @lock: dma-fence fence lock */
- spinlock_t lock; /** @error: preempt fence is in error state */ int error;
};
2.46.0
linux-stable-mirror@lists.linaro.org
-
Matthew Auld -
Matthew Brost