[RFA][PATCH] Fix tree-optimization/59919

Jeff Law law@redhat.com
Fri Jan 24 20:22:00 GMT 2014


On 01/24/14 03:52, Jakub Jelinek wrote:
>
> I'd say this belongs into infer_value_range instead.
Here we go with that variant:

OK for the trunk?

jeff
-------------- next part --------------
diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index 2998c72..3b1abbc 100644
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -1,3 +1,9 @@
+2014-01-23  Jeff Law  <law@redhat.com>
+
+	PR tree-optimization/59919
+	* tree-vrp.c (find_assert_locations_1): Do not register asserts
+	for non-returning calls.
+
 2014-01-23  Pat Haugen  <pthaugen@us.ibm.com>
 
 	* config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
diff --git a/gcc/testsuite/ChangeLog b/gcc/testsuite/ChangeLog
index 564d425..0d05178 100644
--- a/gcc/testsuite/ChangeLog
+++ b/gcc/testsuite/ChangeLog
@@ -1,3 +1,8 @@
+2014-01-23  Jeff Law  <law@redhat.com>
+
+	PR tree-optimization/59919
+	* gcc.c-torture/compile/pr59919.c: New test.
+
 2014-01-23  Paolo Carlini  <paolo.carlini@oracle.com>
 
 	PR c++/58980
diff --git a/gcc/testsuite/gcc.c-torture/compile/pr59919.c b/gcc/testsuite/gcc.c-torture/compile/pr59919.c
new file mode 100644
index 0000000..6809caa
--- /dev/null
+++ b/gcc/testsuite/gcc.c-torture/compile/pr59919.c
@@ -0,0 +1,18 @@
+typedef int jmp_buf[10];
+struct S
+{
+  int i;
+  jmp_buf buf;
+};
+
+void setjmp (jmp_buf);
+void foo (int *);
+__attribute__ ((__noreturn__, __nonnull__)) void bar (struct S *);
+
+void
+baz (struct S *p)
+{
+  bar (p);
+  setjmp (p->buf);
+  foo (&p->i);
+}
diff --git a/gcc/tree-vrp.c b/gcc/tree-vrp.c
index f6da192..7aa732d 100644
--- a/gcc/tree-vrp.c
+++ b/gcc/tree-vrp.c
@@ -4534,12 +4534,21 @@ infer_value_range (gimple stmt, tree op, enum tree_code *comp_code_p, tree *val_
   if (stmt_could_throw_p (stmt))
     return false;
 
-  /* If STMT is the last statement of a basic block with no
+  /* If STMT is the last statement of a basic block with no normal
      successors, there is no point inferring anything about any of its
      operands.  We would not be able to find a proper insertion point
      for the assertion, anyway.  */
-  if (stmt_ends_bb_p (stmt) && EDGE_COUNT (gimple_bb (stmt)->succs) == 0)
-    return false;
+  if (stmt_ends_bb_p (stmt))
+    {
+      edge_iterator ei;
+      edge e;
+
+      FOR_EACH_EDGE (e, ei, gimple_bb (stmt)->succs)
+	if (!(e->flags & EDGE_ABNORMAL))
+	  break;
+      if (e == NULL)
+	return false;
+    }
 
   if (infer_nonnull_range (stmt, op, true, true))
     {


More information about the Gcc-patches mailing list