The only things this protects is reading ->flags and ->size, both of
which are invariant over the lifetime of an exynos gem bo. So no
locking needed at all (besides that, nothing protects the writers
anyway).

Aside: exynos_gem_obj->size is redundant with
exynos_gem_obj->base.size and probably should be removed.

v2: Use _unlocked unreference (Daniel Stone).

Cc: Daniel Stone <daniel at fooishbar.org>
Cc: Inki Dae <inki.dae at samsung.com>
Signed-off-by: Daniel Vetter <daniel.vetter at intel.com>
---
 drivers/gpu/drm/exynos/exynos_drm_gem.c | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)

diff --git a/drivers/gpu/drm/exynos/exynos_drm_gem.c 
b/drivers/gpu/drm/exynos/exynos_drm_gem.c
index 60b9975bb0b1..6fb98f4c3544 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_gem.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_gem.c
@@ -362,12 +362,9 @@ int exynos_drm_gem_get_ioctl(struct drm_device *dev, void 
*data,
        struct drm_exynos_gem_info *args = data;
        struct drm_gem_object *obj;

-       mutex_lock(&dev->struct_mutex);
-
        obj = drm_gem_object_lookup(dev, file_priv, args->handle);
        if (!obj) {
                DRM_ERROR("failed to lookup gem object.\n");
-               mutex_unlock(&dev->struct_mutex);
                return -EINVAL;
        }

@@ -376,8 +373,7 @@ int exynos_drm_gem_get_ioctl(struct drm_device *dev, void 
*data,
        args->flags = exynos_gem->flags;
        args->size = exynos_gem->size;

-       drm_gem_object_unreference(obj);
-       mutex_unlock(&dev->struct_mutex);
+       drm_gem_object_unreference_unlocked(obj);

        return 0;
 }
-- 
2.8.0.rc3

Reply via email to