drm/vmwgfx: Kill some lockdep warnings

Some global KMS state that is elsewhere protected by the mode_config
mutex here needs to be protected with a local mutex. Remove corresponding
lockdep checks and introduce a new driver-private global_kms_state_mutex,
and make sure its locking order is *after* the crtc locks in order to
avoid having to release those when the new mutex is taken.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Sinclair Yeh <syeh@vmware.com>
Cc: <stable@vger.kernel.org> # 4.6
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c b/drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c
index b949102..9ca818f 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c
@@ -553,12 +553,15 @@
 	}
 
 	/* Only one active implicit frame-buffer at a time. */
+	mutex_lock(&dev_priv->global_kms_state_mutex);
 	if (!turning_off && stdu->base.is_implicit && dev_priv->implicit_fb &&
 	    !(dev_priv->num_implicit == 1 && stdu->base.active_implicit)
 	    && dev_priv->implicit_fb != vfb) {
+		mutex_unlock(&dev_priv->global_kms_state_mutex);
 		DRM_ERROR("Multiple implicit framebuffers not supported.\n");
 		return -EINVAL;
 	}
+	mutex_unlock(&dev_priv->global_kms_state_mutex);
 
 	/* Since they always map one to one these are safe */
 	connector = &stdu->base.connector;