--- old/src/hotspot/cpu/sparc/templateTable_sparc.cpp	2017-10-24 16:40:16.488050413 +0200
+++ new/src/hotspot/cpu/sparc/templateTable_sparc.cpp	2017-10-24 16:40:16.241042296 +0200
@@ -1499,7 +1499,7 @@
     // Push returnAddress for "ret" on stack
     __ push_ptr(Otos_i);
     // And away we go!
-    __ dispatch_next(vtos);
+    __ dispatch_next(vtos, 0, true);
     return;
   }
 
@@ -1607,7 +1607,7 @@
   // continue with bytecode @ target
   // %%%%% Like Intel, could speed things up by moving bytecode fetch to code above,
   // %%%%% and changing dispatch_next to dispatch_only
-  __ dispatch_next(vtos);
+  __ dispatch_next(vtos, 0, true);
 }
 
 
@@ -1676,7 +1676,7 @@
   __ ld_ptr(Lmethod, Method::const_offset(), G3_scratch);
   __ add(G3_scratch, Otos_i, G3_scratch);
   __ add(G3_scratch, in_bytes(ConstMethod::codes_offset()), Lbcp);
-  __ dispatch_next(vtos);
+  __ dispatch_next(vtos, 0, true);
 }