[PULL,10/11] cpus: call cpu_update_icount on read

Message ID 20170410125524.21008-11-alex.bennee@linaro.org
State New
Headers show
  • Final icount and misc MTTCG fixes for 2.9
Related show

Commit Message

Alex Bennée April 10, 2017, 12:55 p.m.
This ensures each time the vCPU thread reads the icount we update the
master timer_state.qemu_icount field. This way as long as updates are
in BQL protected sections (which they should be) the main-loop can
never come to update the log and find time has gone backwards.

Signed-off-by: Alex Bennée <alex.bennee@linaro.org>



diff --git a/cpus.c b/cpus.c
index 9c8bd2c991..740b8dc3f8 100644
--- a/cpus.c
+++ b/cpus.c
@@ -253,19 +253,21 @@  void cpu_update_icount(CPUState *cpu)
 int64_t cpu_get_icount_raw(void)
-    int64_t icount;
     CPUState *cpu = current_cpu;
-    icount = atomic_read(&timers_state.qemu_icount);
     if (cpu && cpu->running) {
         if (!cpu->can_do_io) {
             fprintf(stderr, "Bad icount read\n");
         /* Take into account what has run */
-        icount += cpu_get_icount_executed(cpu);
+        cpu_update_icount(cpu);
-    return icount;
+    return atomic_read__nocheck(&timers_state.qemu_icount);
+#else /* FIXME: we need 64bit atomics to do this safely */
+    return timers_state.qemu_icount;
 /* Return the virtual CPU time, based on the instruction counter.  */