calculate per-thread time offset

The main thread squirrels away vlib_time_now (&vlib_global_main),
worker threads use it to calculate an offset in f64 seconds from their
own vlib_time_now(vm) value. We use that offset until the next barrier
sync.

Thanks to Damjan for the suggestion.

Change-Id: If56cdfe68e5ad8ac3b0d0fc885dc3ba556cd1215
Signed-off-by: Dave Barach <dave@barachs.net>
diff --git a/src/vlib/main.h b/src/vlib/main.h
index f89ecd3..4192c3f 100644
--- a/src/vlib/main.h
+++ b/src/vlib/main.h
@@ -63,6 +63,9 @@
   CLIB_CACHE_LINE_ALIGN_MARK (cacheline0);
   /* Instruction level timing state. */
   clib_time_t clib_time;
+  /* Offset from main thread time */
+  f64 time_offset;
+  f64 time_last_barrier_release;
 
   /* Time stamp of last node dispatch. */
   u64 cpu_time_last_node_dispatch;
@@ -232,7 +235,7 @@
 always_inline f64
 vlib_time_now (vlib_main_t * vm)
 {
-  return clib_time_now (&vm->clib_time);
+  return clib_time_now (&vm->clib_time) + vm->time_offset;
 }
 
 always_inline f64