drm/amdgpu: Don't leak runtime pm ref on driver unload
authorLukas Wunner <lukas@wunner.de>
Wed, 8 Jun 2016 16:47:27 +0000 (18:47 +0200)
committerDaniel Vetter <daniel.vetter@ffwll.ch>
Fri, 24 Jun 2016 06:49:30 +0000 (08:49 +0200)
amdgpu_driver_load_kms() calls pm_runtime_put_autosuspend() if
amdgpu_device_is_px(dev), but amdgpu_driver_unload_kms() calls
pm_runtime_get_sync() unconditionally. We therefore leak a runtime pm
ref whenever amdgpu is unloaded on a non-PX machine or if runpm=0. The
GPU will subsequently never runtime suspend after loading amdgpu again.

Fix by taking the runtime pm ref under the same condition that it was
released on driver load.

Fixes: d38ceaf99ed0 ("drm/amdgpu: add core driver (v4)")
Cc: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/86364eeb5711323608930c4f0f69046792ff6d3c.1465392124.git.lukas@wunner.de
drivers/gpu/drm/amd/amdgpu/amdgpu_kms.c

index 40a2370..9b1f979 100644 (file)
@@ -60,7 +60,9 @@ int amdgpu_driver_unload_kms(struct drm_device *dev)
        if (adev->rmmio == NULL)
                goto done_free;
 
-       pm_runtime_get_sync(dev->dev);
+       if (amdgpu_device_is_px(dev)) {
+               pm_runtime_get_sync(dev->dev);
+       }
 
        amdgpu_amdkfd_device_fini(adev);