From: Lucas Stach <l.stach@xxxxxxxxxxxxxx> commit cda7532916f7bc860b36a1806cb8352e6f63dacb upstream. The prev context is the MMU context at the time of the job queueing in hardware. As a job might be queued multiple times due to recovery after a GPU hang, we need to make sure to put the stale prev MMU context from a prior queuing, to avoid the reference and thus the MMU context leaking. Cc: stable@xxxxxxxxxxxxxxx # 5.4 Signed-off-by: Lucas Stach <l.stach@xxxxxxxxxxxxxx> Tested-by: Michael Walle <michael@xxxxxxxx> Tested-by: Marek Vasut <marex@xxxxxxx> Reviewed-by: Christian Gmeiner <christian.gmeiner@xxxxxxxxx> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- drivers/gpu/drm/etnaviv/etnaviv_gpu.c | 2 ++ 1 file changed, 2 insertions(+) --- a/drivers/gpu/drm/etnaviv/etnaviv_gpu.c +++ b/drivers/gpu/drm/etnaviv/etnaviv_gpu.c @@ -1356,6 +1356,8 @@ struct dma_fence *etnaviv_gpu_submit(str gpu->mmu_context = etnaviv_iommu_context_get(submit->mmu_context); etnaviv_gpu_start_fe_idleloop(gpu); } else { + if (submit->prev_mmu_context) + etnaviv_iommu_context_put(submit->prev_mmu_context); submit->prev_mmu_context = etnaviv_iommu_context_get(gpu->mmu_context); }