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: sta...@vger.kernel.org # 5.4
Signed-off-by: Lucas Stach <l.st...@pengutronix.de>
Tested-by: Michael Walle <mich...@walle.cc>
---
 drivers/gpu/drm/etnaviv/etnaviv_gpu.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/gpu/drm/etnaviv/etnaviv_gpu.c 
b/drivers/gpu/drm/etnaviv/etnaviv_gpu.c
index c8b9b0cc4442..c1b9c5cbed11 100644
--- a/drivers/gpu/drm/etnaviv/etnaviv_gpu.c
+++ b/drivers/gpu/drm/etnaviv/etnaviv_gpu.c
@@ -1368,6 +1368,8 @@ struct dma_fence *etnaviv_gpu_submit(struct 
etnaviv_gem_submit *submit)
                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);
        }
 
-- 
2.30.2

Reply via email to