Subzero: Fixed deadlock with _start but no globals

tests_lit/asan_tests/no_globals.ll

Index: tests_lit/asan_tests/no_globals.ll
diff --git a/tests_lit/asan_tests/no_globals.ll b/tests_lit/asan_tests/no_globals.ll
new file mode 100644
index 0000000000000000000000000000000000000000..7e1c20d3111f196bd9f61cda359dd4ddb60872ff
--- /dev/null
+++ b/tests_lit/asan_tests/no_globals.ll
@@ -0,0 +1,17 @@
+; Check that Subzero can instrument _start when there are no globals

[Jim Stichnoth, 2016/07/19 14:04:13]

Can you add a more complete description of why this test is here?  Specifically
about the possibility of deadlock when there are no global initializers and
_start is the first function.

[tlively, 2016/07/19 22:27:04]

Done.

+
+; REQUIRES: no_minimal_build
+
+; RUN: %p2i -i %s --args -verbose=inst -fsanitize-address \
+; RUN:     | FileCheck --check-prefix=DUMP %s
+
+define void @_start(i32 %arg) {
+  ret void
+}
+
+; DUMP-LABEL: ================ Instrumented CFG ================

[Jim Stichnoth, 2016/07/19 14:04:13]

Can the tested output be made less specific?

I thought it was a deadlock issue, so would simply
  DUMP: __asan_init
be a sufficient test?

(But be sure to verify that it fails with the old code and succeeds with the CL,
before taking my suggestion...)

I say this because the dump output format isn't strongly guaranteed to be stable
over time.  For example, we might want to change the "Instrumented CFG" string,
or the number of '=' characters, etc.

[tlively, 2016/07/19 22:27:04]

Done.

+; DUMP-NEXT: define void @_start(i32 %arg) {
+; DUMP-NEXT: __0:
+; DUMP-NEXT: call void @__asan_init(i32 0, i32 @__$rz_array, i32 @__$rz_sizes)
+; DUMP-NEXT: ret void
+; DUMP-NEXT: }