tcg: distribute tcg_time into TCG contexts

When we implemented per-vCPU TCG contexts, we forgot to also
distribute the tcg_time counter, which has remained as a global
accessed without any serialization, leading to potentially missed
counts.

Fix it by distributing the field over the TCG contexts, embedding
it into TCGProfile with a field called "cpu_exec_time", which is more
descriptive than "tcg_time". Add a function to query this value
directly, and for completeness, fill in the field in
tcg_profile_snapshot, even though its callers do not use it.

Signed-off-by: Emilio G. Cota <cota@braap.org>
Message-Id: <20181010144853.13005-5-cota@braap.org>
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
diff --git a/monitor.c b/monitor.c
index b9258a7..823b5a1 100644
--- a/monitor.c
+++ b/monitor.c
@@ -83,6 +83,7 @@
 #include "sysemu/cpus.h"
 #include "sysemu/iothread.h"
 #include "qemu/cutils.h"
+#include "tcg/tcg.h"
 
 #if defined(TARGET_S390X)
 #include "hw/s390x/storage-keys.h"
@@ -1966,16 +1967,22 @@
 
 #ifdef CONFIG_PROFILER
 
-int64_t tcg_time;
 int64_t dev_time;
 
 static void hmp_info_profile(Monitor *mon, const QDict *qdict)
 {
+    static int64_t last_cpu_exec_time;
+    int64_t cpu_exec_time;
+    int64_t delta;
+
+    cpu_exec_time = tcg_cpu_exec_time();
+    delta = cpu_exec_time - last_cpu_exec_time;
+
     monitor_printf(mon, "async time  %" PRId64 " (%0.3f)\n",
                    dev_time, dev_time / (double)NANOSECONDS_PER_SECOND);
     monitor_printf(mon, "qemu time   %" PRId64 " (%0.3f)\n",
-                   tcg_time, tcg_time / (double)NANOSECONDS_PER_SECOND);
-    tcg_time = 0;
+                   delta, delta / (double)NANOSECONDS_PER_SECOND);
+    last_cpu_exec_time = cpu_exec_time;
     dev_time = 0;
 }
 #else