[PATCH] Skip ubsan internal fns in tail-merge

[Tom de Vries]

[ was: Re: [PATCH 3/3] Fix ubsan tests by disabling of an optimization. ]
On 10/07/15 22:11, Jeff Law wrote:
On 07/10/2015 02:19 AM, Richard Biener wrote:

But the warning on the "bogus" line will still be warranted, so user
goes and
fixes it.
But when the user gets the "bogus" line, he may look at the code and
determine that the reported line can't possibly be executed -- so they
get confused, assume the warning is bogus and ignore it.



  Then tail-merge no longer applies and he gets the warning on the
other warranted line.
That assumes that we'd get to both paths.  We may not.  The paths may be
totally independent.


IIUC, the conclusion of this discussion (starting at 
https://gcc.gnu.org/ml/gcc-patches/2015-07/msg00758.html ) is that we 
want to disable tail-merge for ubsan internal functions?

This patch implements that.

Should the asan internal function (ASAN_CHECK) be handled in the same 
way, as suggested here ( 
https://gcc.gnu.org/ml/gcc-patches/2015-07/msg00776.html )?

OK for stage4 trunk if bootstrap and reg-test succeeds?

Thanks,
- Tom

Skip ubsan internal fns in tail-merge

2016-03-07  Tom de Vries  <t...@codesourcery.com>

	* tree-ssa-tail-merge.c (merge_stmt_p): New function, factored out
	of ...
	(find_duplicate): ... here.
	(merge_stmt_p): Return false for ubsan functions.

---
 gcc/tree-ssa-tail-merge.c | 31 +++++++++++++++++++++++++++++--
 1 file changed, 29 insertions(+), 2 deletions(-)

diff --git a/gcc/tree-ssa-tail-merge.c b/gcc/tree-ssa-tail-merge.c
index 5d32790..7f6e922 100644
--- a/gcc/tree-ssa-tail-merge.c
+++ b/gcc/tree-ssa-tail-merge.c
@@ -1207,6 +1207,33 @@ gsi_advance_bw_nondebug_nonlocal (gimple_stmt_iterator *gsi, tree *vuse,
     }
 }
 
+/* Return true if STMT is allowed to be merged.  */
+
+static bool
+merge_stmt_p (gimple *stmt)
+{
+  if (is_tm_ending (stmt))
+    return false;
+
+  if (is_gimple_call (stmt)
+      && gimple_call_internal_p (stmt))
+    switch (gimple_call_internal_fn (stmt))
+      {
+      case IFN_UBSAN_NULL:
+      case IFN_UBSAN_BOUNDS:
+      case IFN_UBSAN_VPTR:
+      case IFN_UBSAN_CHECK_ADD:
+      case IFN_UBSAN_CHECK_SUB:
+      case IFN_UBSAN_CHECK_MUL:
+      case IFN_UBSAN_OBJECT_SIZE:
+	return false;
+      default:
+	break;
+      }
+
+  return true;
+}
+
 /* Determines whether BB1 and BB2 (members of same_succ) are duplicates.  If so,
    clusters them.  */
 
@@ -1229,8 +1256,8 @@ find_duplicate (same_succ *same_succ, basic_block bb1, basic_block bb2)
       /* What could be better than this here is to blacklist the bb
 	 containing the stmt, when encountering the stmt f.i. in
 	 same_succ_hash.  */
-      if (is_tm_ending (stmt1)
-	  || is_tm_ending (stmt2))
+      if (!merge_stmt_p (stmt1)
+	  || !merge_stmt_p (stmt2))
 	return;
 
       if (!gimple_equal_p (same_succ, stmt1, stmt2))