drm/amdgpu: fix a memory leak in fence cleanup when unloading

[ Upstream commit 7838fb5 ]

Commit b61badd ("drm/amdgpu: fix usage slab after free")
reordered when amdgpu_fence_driver_sw_fini() was called after
that patch, amdgpu_fence_driver_sw_fini() effectively became
a no-op as the sched entities we never freed because the
ring pointers were already set to NULL.  Remove the NULL
setting.

Reported-by: Lin.Cao <lincao12@amd.com>
Cc: Vitaly Prosyak <vitaly.prosyak@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Fixes: b61badd ("drm/amdgpu: fix usage slab after free")
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit a525fa3)
Cc: stable@vger.kernel.org
[ Adapt to conditional check ]
Signed-off-by: Sasha Levin <sashal@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Author: alexdeucher

drivers/gpu/drm/amd/amdgpu/amdgpu_ring.c

@@ -396,9 +396,6 @@ void amdgpu_ring_fini(struct amdgpu_ring *ring)
 	dma_fence_put(ring->vmid_wait);
 	ring->vmid_wait = NULL;
 	ring->me = 0;
-
-	if (!ring->is_mes_queue)
-		ring->adev->rings[ring->idx] = NULL;
 }
 
 /**