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/tcg/tcg.c b/tcg/tcg.c
index 8f26916..e85133e 100644
--- a/tcg/tcg.c
+++ b/tcg/tcg.c
@@ -30,6 +30,7 @@
 /* Define to jump the ELF file used to communicate with GDB.  */
 #undef DEBUG_JIT
 
+#include "qemu/error-report.h"
 #include "qemu/cutils.h"
 #include "qemu/host-utils.h"
 #include "qemu/timer.h"
@@ -3361,6 +3362,7 @@
         const TCGProfile *orig = &s->prof;
 
         if (counters) {
+            PROF_ADD(prof, orig, cpu_exec_time);
             PROF_ADD(prof, orig, tb_count1);
             PROF_ADD(prof, orig, tb_count);
             PROF_ADD(prof, orig, op_count);
@@ -3412,11 +3414,32 @@
                     prof.table_op_count[i]);
     }
 }
+
+int64_t tcg_cpu_exec_time(void)
+{
+    unsigned int n_ctxs = atomic_read(&n_tcg_ctxs);
+    unsigned int i;
+    int64_t ret = 0;
+
+    for (i = 0; i < n_ctxs; i++) {
+        const TCGContext *s = atomic_read(&tcg_ctxs[i]);
+        const TCGProfile *prof = &s->prof;
+
+        ret += atomic_read(&prof->cpu_exec_time);
+    }
+    return ret;
+}
 #else
 void tcg_dump_op_count(FILE *f, fprintf_function cpu_fprintf)
 {
     cpu_fprintf(f, "[TCG profiler not compiled]\n");
 }
+
+int64_t tcg_cpu_exec_time(void)
+{
+    error_report("%s: TCG profiler not compiled", __func__);
+    exit(EXIT_FAILURE);
+}
 #endif
 
 
diff --git a/tcg/tcg.h b/tcg/tcg.h
index d80ef2a..c59f254 100644
--- a/tcg/tcg.h
+++ b/tcg/tcg.h
@@ -629,6 +629,7 @@
 QEMU_BUILD_BUG_ON(NB_OPS > (1 << 8));
 
 typedef struct TCGProfile {
+    int64_t cpu_exec_time;
     int64_t tb_count1;
     int64_t tb_count;
     int64_t op_count; /* total insn count */
@@ -1002,6 +1003,7 @@
 #define tcg_check_temp_count() 0
 #endif
 
+int64_t tcg_cpu_exec_time(void);
 void tcg_dump_info(FILE *f, fprintf_function cpu_fprintf);
 void tcg_dump_op_count(FILE *f, fprintf_function cpu_fprintf);