gpu: nvgpu: hold mutex to get cde context only

In gk20a_cde_convert(), we hold cde_app.mutex almost
for everything which is unnecessary

Also, this causes a deadlock scenario when gk20a_do_idle()
is called
In some cases it is possible that Thread 1
holds the lock and is trying to power on GPU, and
Thread 2 is trying to power off GPU and then grab
cde_app.mutex to cleanup GPU state

To fix this, grab the mutex only for gk20a_cde_get_context()
and then release it

Bug 200287073

Change-Id: Ic4856604652d0d4024abdeb5c2f8f03910c601a5
Signed-off-by: Deepak Nibade <dnibade@nvidia.com>
Reviewed-on: http://git-master/r/1324328
(cherry picked from commit 3c8eef0d0e9ef3307e2792e3e127c10f3ff118a7)
Reviewed-on: http://git-master/r/1322030
GVS: Gerrit_Virtual_Submit
Reviewed-by: Bharat Nihalani <bnihalani@nvidia.com>

1 files changed, 1 insertions, 2 deletions

diff --git a/drivers/gpu/nvgpu/gk20a/cde_gk20a.c b/drivers/gpu/nvgpu/gk20a/cde_gk20a.c
index c8434f3d..8f9d7831 100644
--- a/drivers/gpu/nvgpu/gk20a/cde_gk20a.c
+++ b/drivers/gpu/nvgpu/gk20a/cde_gk20a.c
@@ -950,8 +950,8 @@ __releases(&cde_app->mutex)
                 return -EINVAL;
 
         nvgpu_mutex_acquire(&g->cde_app.mutex);
-
         cde_ctx = gk20a_cde_get_context(g);
+        nvgpu_mutex_release(&g->cde_app.mutex);
         if (IS_ERR(cde_ctx)) {
                 err = PTR_ERR(cde_ctx);
                 goto exit_unlock;
@@ -1121,7 +1121,6 @@ exit_unlock:
         if (surface)
                 dma_buf_vunmap(compbits_scatter_buf, surface);
 
-        nvgpu_mutex_release(&g->cde_app.mutex);
         return err;
 }