msm_devfreq_init only initializes the idle_work hrtimer when it succeeds
to create a devfreq instance (devfreq support is optional), yet
msm_devfreq_idle is called unconditionally from retire_submit and queues
work on it. We're seeing:
[ 2.005265] adreno 1c00000.gpu: [drm:msm_devfreq_init] *ERROR* Couldn't initialize GPU devfreq
Followed by a pagefault in:
[ 16.650316] pc : hrtimer_start_range_ns+0x64/0x360
[ 16.650336] lr : msm_hrtimer_queue_work+0x18/0x2c
Moments later. Just like msm_devfreq_active, check if the devfreq
instance is not NULL before proceeding.
Fixes: 658f4c829688 ("drm/msm/devfreq: Add 1ms delay before clamping freq")
Reviewed-by: Konrad Dybcio <konrad.dybcio@xxxxxxxxxxxxxx>
Signed-off-by: Marijn Suijten <marijn.suijten@xxxxxxxxxxxxxx>
---
drivers/gpu/drm/msm/msm_gpu_devfreq.c | 3 +++
1 file changed, 3 insertions(+)
diff --git a/drivers/gpu/drm/msm/msm_gpu_devfreq.c b/drivers/gpu/drm/msm/msm_gpu_devfreq.c
index 8b7473f69cb8..ae3309ff96e4 100644
--- a/drivers/gpu/drm/msm/msm_gpu_devfreq.c
+++ b/drivers/gpu/drm/msm/msm_gpu_devfreq.c
@@ -227,6 +227,9 @@ void msm_devfreq_idle(struct msm_gpu *gpu)
{
struct msm_gpu_devfreq *df = &gpu->devfreq;
+ if (!df->devfreq)
+ return;
+
msm_hrtimer_queue_work(&df->idle_work, ms_to_ktime(1),
HRTIMER_MODE_ABS);
}
--
2.34.0
