drm/dp_mst: Fix drm_dp_check_mstb_guid() return code

We actually expect this to return a 0 on success, or negative error code
on failure. In order to do that, we check whether or not we managed to
write the whole GUID and then return 0 if so, otherwise return a
negative error code. Also, let's add an error message here so it's a
little more obvious when this fails in the middle of a link address
probe.

This should fix issues with certain MST hubs seemingly stopping for no
reason in the middle of the link address probe process.

Fixes: cb897542 ("drm/dp_mst: Fix W=1 warnings")
Cc: Benjamin Gaignard <benjamin.gaignard@st.com>
Cc: Sean Paul <sean@poorly.run>
Cc: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Lyude Paul <lyude@redhat.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200306234923.547873-3-lyude@redhat.comReviewed-by: Alex Deucher <alexander.deucher@amd.com>

drivers/gpu/drm/drm_dp_mst_topology.c

@@ -2092,7 +2092,10 @@ static int drm_dp_check_mstb_guid(struct drm_dp_mst_branch *mstb, u8 *guid)
 		}
 	}
 
-	return ret;
+	if (ret < 16 && ret > 0)
+		return -EPROTO;
+
+	return ret == 16 ? 0 : ret;
 }
 
 static void build_mst_prop_path(const struct drm_dp_mst_branch *mstb,
@@ -2907,8 +2910,14 @@ static int drm_dp_send_link_address(struct drm_dp_mst_topology_mgr *mgr,
 	drm_dp_dump_link_address(reply);
 
 	ret = drm_dp_check_mstb_guid(mstb, reply->guid);
-	if (ret)
+	if (ret) {
+		char buf[64];
+
+		drm_dp_mst_rad_to_str(mstb->rad, mstb->lct, buf, sizeof(buf));
+		DRM_ERROR("GUID check on %s failed: %d\n",
+			  buf, ret);
 		goto out;
+	}
 
 	for (i = 0; i < reply->nports; i++) {
 		port_mask |= BIT(reply->ports[i].port_number);