[PATCH][PR tree-optimization/pr67755] Fix profile insanity adjustments

	PR tree-optimization/pr67755
	* tree-ssa-threadupdate.c (struct ssa_local_info_t): Add new field
	"need_profile_correction".
	(thread_block_1): Initialize new field to false by default.  If we
	have multiple thread paths through a common joiner to different
	final targets, then set new field to true.
	(compute_path_counts): Only do count adjustment when it's really
	needed.

	PR tree-optimization/67755
	* gcc.dg/tree-ssa/pr67755.c: New test.

From-SVN: r232313

gcc/ChangeLog

+2016-01-12  Jeff Law  <law@redhat.com>
+
+	PR tree-optimization/pr67755
+	* tree-ssa-threadupdate.c (struct ssa_local_info_t): Add new field
+	"need_profile_correction".
+	(thread_block_1): Initialize new field to false by default.  If we
+	have multiple thread paths through a common joiner to different
+	final targets, then set new field to true.
+	(compute_path_counts): Only do count adjustment when it's really
+	needed.
+
 2016-01-12  Sandra Loosemore <sandra@codesourcery.com>
 
 	* doc/invoke.texi (Spec Files): Move section down in file, past

gcc/testsuite/ChangeLog

-2015-01-13  Thomas Preud'homme  <thomas.preudhomme@arm.com>
+2016-01-13  Jeff Law  <law@redhat.com>
+
+	PR tree-optimization/67755
+	* gcc.dg/tree-ssa/pr67755.c: New test.
+
+2016-01-13  Thomas Preud'homme  <thomas.preudhomme@arm.com>
 
 	* gcc.c-torture/unsorted/dump-noaddr.x (dump_compare): Replace static
 	pass number in output by a star.

gcc/testsuite/gcc.dg/tree-ssa/pr67755.c

+/* { dg-do compile } */
+/* { dg-options "-O1 -fdump-tree-dom2-details-blocks" } */
+/* We want to verify no outgoing edge from a conditional
+   has a probability of 100%.  */
+/* { dg-final { scan-tree-dump-not "succ:\[ \]+. .100.0%.  .\(TRUE|FALSE\)_VALUE" "dom2"} } */
+
+
+void (*zend_block_interruptions) (void);
+
+int * _zend_mm_alloc_int (int * heap, long int size)
+{
+  int *best_fit;
+  long int true_size = (size < 15 ? 32 : size);
+
+  if (zend_block_interruptions)
+    zend_block_interruptions ();
+
+  if (__builtin_expect ((true_size < 543), 1))
+    best_fit = heap + 2;
+  else
+    best_fit = heap;
+
+  return best_fit;
+}
+

gcc/tree-ssa-threadupdate.c

@@ -239,6 +239,11 @@ struct ssa_local_info_t
 
   /* Blocks duplicated for the thread.  */
   bitmap duplicate_blocks;
+
+  /* When we have multiple paths through a joiner which reach different
+     final destinations, then we may need to correct for potential
+     profile insanities.  */
+  bool need_profile_correction;
 };
 
 /* Passes which use the jump threading code register jump threading
@@ -826,7 +831,8 @@ compute_path_counts (struct redirection_data *rd,
      So ensure that this path's path_out_count is at least the
      difference between elast->count and nonpath_count.  Otherwise the edge
      counts after threading will not be sane.  */
-  if (has_joiner && path_out_count < elast->count - nonpath_count)
+  if (local_info->need_profile_correction
+      && has_joiner && path_out_count < elast->count - nonpath_count)
     {
       path_out_count = elast->count - nonpath_count;
       /* But neither can we go above the minimum count along the path
@@ -1492,6 +1498,7 @@ thread_block_1 (basic_block bb, bool noloop_only, bool joiners)
   ssa_local_info_t local_info;
 
   local_info.duplicate_blocks = BITMAP_ALLOC (NULL);
+  local_info.need_profile_correction = false;
 
   /* To avoid scanning a linear array for the element we need we instead
      use a hash table.  For normal code there should be no noticeable
@@ -1502,6 +1509,7 @@ thread_block_1 (basic_block bb, bool noloop_only, bool joiners)
 
   /* Record each unique threaded destination into a hash table for
      efficient lookups.  */
+  edge last = NULL;
   FOR_EACH_EDGE (e, ei, bb->preds)
     {
       if (e->aux == NULL)
@@ -1555,6 +1563,17 @@ thread_block_1 (basic_block bb, bool noloop_only, bool joiners)
       /* Insert the outgoing edge into the hash table if it is not
 	 already in the hash table.  */
       lookup_redirection_data (e, INSERT);
+
+      /* When we have thread paths through a common joiner with different
+	 final destinations, then we may need corrections to deal with
+	 profile insanities.  See the big comment before compute_path_counts.  */
+      if ((*path)[1]->type == EDGE_COPY_SRC_JOINER_BLOCK)
+	{
+	  if (!last)
+	    last = e2;
+	  else if (e2 != last)
+	    local_info.need_profile_correction = true;
+	}
     }
 
   /* We do not update dominance info.  */