[COMMITTED] tests: backtrace.c only check we caught the last instruction on x86_64.

Mark Wielaard mjw at redhat.com
Sat Dec 21 18:44:28 UTC 2013


On some architectures gcc might introduce some "padding instructions"
at the end of the function (like on ppc64). So only assert we are at the
last instruction of backtracegen if on x86_64 native. In theory the assert
could even fail on that architectures, but in practice it doesn't and it
is a nice test to have.

Signed-off-by: Mark Wielaard <mjw at redhat.com>
---
 tests/ChangeLog   | 5 +++++
 tests/backtrace.c | 8 +++++++-
 2 files changed, 12 insertions(+), 1 deletion(-)

diff --git a/tests/ChangeLog b/tests/ChangeLog
index 0d91c3b..23c5051 100644
--- a/tests/ChangeLog
+++ b/tests/ChangeLog
@@ -1,3 +1,8 @@
+2013-12-21  Mark Wielaard  <mjw at redhat.com>
+
+	* backtrace.c (callback_verify): Only assert that case 5 is the last
+	instruction of backtracegen on x86_64 native.
+
 2013-12-18  Jan Kratochvil  <jan.kratochvil at redhat.com>
 	    Mark Wielaard  <mjw at redhat.com>
 
diff --git a/tests/backtrace.c b/tests/backtrace.c
index 06d3878..8a7d6df 100644
--- a/tests/backtrace.c
+++ b/tests/backtrace.c
@@ -118,7 +118,13 @@ callback_verify (pid_t tid, unsigned frameno, Dwarf_Addr pc,
       mod = dwfl_addrmodule (dwfl, pc);
       if (mod)
 	symname2 = dwfl_module_addrname (mod, pc);
-      assert (symname2 == NULL || strcmp (symname2, "backtracegen") != 0);
+
+      // Note that the following assert might in theory even fail on x86_64,
+      // there is no guarantee that the compiler doesn't reorder the
+      // instructions or even inserts some padding instructions at the end
+      // (which apparently happens on ppc64).
+      if (is_x86_64_native)
+        assert (symname2 == NULL || strcmp (symname2, "backtracegen") != 0);
       break;
   }
 }
-- 
1.8.4.2



More information about the elfutils-devel mailing list