From: Rob Clark <robdclark@xxxxxxxxxxxx>
[ Upstream commit 49e477610087a02c3604061b8f3ee3a25a493987 ]
In msm_devfreq_suspend() we cancel idle_work synchronously so that it
doesn't run after we power of the hw or in the resume path. But this
means that we want to ensure that idle_work is not scheduled *after* we
no longer hold a runpm ref. So switch the ordering of pm_runtime_put()
vs msm_devfreq_idle().
v2. Only move the runpm _put_autosuspend, and not the _mark_last_busy()
Fixes: 9bc95570175a ("drm/msm: Devfreq tuning")
Signed-off-by: Rob Clark <robdclark@xxxxxxxxxxxx>
Link: https://lore.kernel.org/r/20210927152928.831245-1-robdclark@xxxxxxxxx
Reviewed-by: Akhil P Oommen <quic_akhilpo@xxxxxxxxxxx>
Reviewed-by: Douglas Anderson <dianders@xxxxxxxxxxxx>
Link: https://lore.kernel.org/r/20220608161334.2140611-1-robdclark@xxxxxxxxx
Signed-off-by: Rob Clark <robdclark@xxxxxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
drivers/gpu/drm/msm/msm_gpu.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/msm/msm_gpu.c b/drivers/gpu/drm/msm/msm_gpu.c
index 58eb3e1662cb..7d27d7cee688 100644
--- a/drivers/gpu/drm/msm/msm_gpu.c
+++ b/drivers/gpu/drm/msm/msm_gpu.c
@@ -664,7 +664,6 @@ static void retire_submit(struct msm_gpu *gpu, struct msm_ringbuffer *ring,
msm_submit_retire(submit);
pm_runtime_mark_last_busy(&gpu->pdev->dev);
- pm_runtime_put_autosuspend(&gpu->pdev->dev);
spin_lock_irqsave(&ring->submit_lock, flags);
list_del(&submit->node);
@@ -678,6 +677,8 @@ static void retire_submit(struct msm_gpu *gpu, struct msm_ringbuffer *ring,
msm_devfreq_idle(gpu);
mutex_unlock(&gpu->active_lock);
+ pm_runtime_put_autosuspend(&gpu->pdev->dev);
+
msm_gem_submit_put(submit);
}
--
2.35.1
