gpu: nvgpu: When powering down, abort if not idle

When trying to power down GPU the engine might be still busy. In this
case delay power down by returning -EBUSY from
gk20a_pm_runtime_suspend().

Bug 200224907

Change-Id: Ibad74c090add24a185bc1a7a02df367af9b95ced
Signed-off-by: Terje Bergstrom <tbergstrom@nvidia.com>
Reviewed-on: http://git-master/r/1213042
Reviewed-by: mobile promotions <svcmobile_promotions@nvidia.com>
Tested-by: mobile promotions <svcmobile_promotions@nvidia.com>

1 files changed, 9 insertions, 1 deletions

diff --git a/drivers/gpu/nvgpu/gk20a/gk20a.c b/drivers/gpu/nvgpu/gk20a/gk20a.c
index f6bb9445..10fa86a1 100644
--- a/drivers/gpu/nvgpu/gk20a/gk20a.c
+++ b/drivers/gpu/nvgpu/gk20a/gk20a.c
@@ -767,6 +767,10 @@ static int gk20a_pm_prepare_poweroff(struct device *dev)
         if (!g->power_on)
                 goto done;
 
+        if (gk20a_fifo_is_engine_busy(g)) {
+                mutex_unlock(&g->poweroff_lock);
+                return -EBUSY;
+        }
         gk20a_scale_suspend(dev);
 
         /* cancel any pending cde work */
@@ -1353,6 +1357,7 @@ static int gk20a_pm_runtime_suspend(struct device *dev)
 fail_railgate:
         gk20a_pm_finalize_poweron(dev);
 fail:
+        pm_runtime_mark_last_busy(dev);
         return err;
 }
 
@@ -1751,6 +1756,7 @@ int __gk20a_do_idle(struct device *dev, bool force_reset)
                 msecs_to_jiffies(GK20A_WAIT_FOR_IDLE_MS);
         int ref_cnt;
         bool is_railgated;
+        int err = 0;
 
         /* acquire busy lock to block other busy() calls */
         down_write(&g->busy_lock);
@@ -1825,7 +1831,9 @@ int __gk20a_do_idle(struct device *dev, bool force_reset)
                  */
 
                 /* Save the GPU state */
-                gk20a_pm_prepare_poweroff(dev);
+                err = gk20a_pm_prepare_poweroff(dev);
+                if (err)
+                        goto fail_drop_usage_count;
 
                 /* railgate GPU */
                 platform->railgate(dev);