Add safety net against potential infinite loop

cpu_interrupt might be called while translating the TB, but before it
is linked into a potentially infinite loop and becomes env->current_tb.

Currently this can (and does) cause huge problems only when using
dyntick clock, with other (periodic) clocks host_alarm_handler will
eventually be executed resulting in a call to cpu_interrupt which will
reset the recursion of running TB and the damage is "only" latency.

git-svn-id: svn://svn.savannah.nongnu.org/qemu/trunk@5620 c046a42c-6fe2-441c-8c8c-71466251a162

diff --git a/cpu-exec.c b/cpu-exec.c
index d31a67aa4d611f839cf45be45487a4698848c688..62886a4f187103ca878e831ab5bf97f03b75712e 100644 (file)
--- a/cpu-exec.c
+++ b/cpu-exec.c
@@ -623,6 +623,14 @@ int cpu_exec(CPUState *env1)
                 }
                 spin_unlock(&tb_lock);
                 env->current_tb = tb;
+
+                /* cpu_interrupt might be called while translating the
+                   TB, but before it is linked into a potentially
+                   infinite loop and becomes env->current_tb. Avoid
+                   starting execution if there is a pending interrupt. */
+                if (unlikely (env->interrupt_request & CPU_INTERRUPT_EXIT))
+                    env->current_tb = NULL;
+
                 while (env->current_tb) {
                     tc_ptr = tb->tc_ptr;
                 /* execute the generated code */