[gcc r12-4398] tree-optimization/102659 - really avoid undef overflow in if-conversion

[Richard Biener]

https://gcc.gnu.org/g:9b2ad21ab3ebc21a3408108327fa1a7cbedaf217

commit r12-4398-g9b2ad21ab3ebc21a3408108327fa1a7cbedaf217
Author: Richard Biener <rguenther@suse.de>
Date:   Thu Oct 14 09:00:25 2021 +0200

    tree-optimization/102659 - really avoid undef overflow in if-conversion
    
    This plugs the remaining hole of POINTER_PLUS_EXPR with undefined
    overflow.  Unfortunately we have to go through some lengths to
    not put invariant conversions into the loop body since that confuses
    the vectorizers gather/scatter discovery which relies on identifying
    an invariant component of plus and minus expressions.  We can
    emit those in the loop preheader but then we have to accept that
    being non-empty when looking for the LOOP_VECTORIZED internal
    function call in the vectorizer.
    
    2021-10-14  Richard Biener  <rguenther@suse.de>
    
            PR tree-optimization/102659
            * tree-if-conv.c (if_convertible_gimple_assign_stmt_p): Also
            rewrite pointer typed undefined overflow operations.
            (predicate_statements): Likewise.  Make sure to emit invariant
            conversions in the preheader.
            * tree-vectorizer.c (vect_loop_vectorized_call): Look through
            non-empty preheaders.
            * tree-data-ref.c (dr_analyze_indices): Strip useless
            conversions to the MEM_REF base type.

Diff:
---
 gcc/tree-data-ref.c   |  1 +
 gcc/tree-if-conv.c    | 37 ++++++++++++++++++++++++++++++-------
 gcc/tree-vectorizer.c |  3 ++-
 3 files changed, 33 insertions(+), 8 deletions(-)

diff --git a/gcc/tree-data-ref.c b/gcc/tree-data-ref.c
index 18307a554fc..57bac06242f 100644
--- a/gcc/tree-data-ref.c
+++ b/gcc/tree-data-ref.c
@@ -1370,6 +1370,7 @@ dr_analyze_indices (struct indices *dri, tree ref, edge nest, loop_p loop)
       tree op = TREE_OPERAND (ref, 0);
       tree access_fn = analyze_scalar_evolution (loop, op);
       access_fn = instantiate_scev (nest, loop, access_fn);
+      STRIP_NOPS (access_fn);
       if (TREE_CODE (access_fn) == POLYNOMIAL_CHREC)
 	{
 	  tree memoff = TREE_OPERAND (ref, 1);
diff --git a/gcc/tree-if-conv.c b/gcc/tree-if-conv.c
index 0b6b07cfac6..15dcc1e2b94 100644
--- a/gcc/tree-if-conv.c
+++ b/gcc/tree-if-conv.c
@@ -1047,7 +1047,8 @@ if_convertible_gimple_assign_stmt_p (gimple *stmt,
 	fprintf (dump_file, "tree could trap...\n");
       return false;
     }
-  else if (INTEGRAL_TYPE_P (TREE_TYPE (lhs))
+  else if ((INTEGRAL_TYPE_P (TREE_TYPE (lhs))
+	    || POINTER_TYPE_P (TREE_TYPE (lhs)))
 	   && TYPE_OVERFLOW_UNDEFINED (TREE_TYPE (lhs))
 	   && arith_code_with_undefined_signed_overflow
 				(gimple_assign_rhs_code (stmt)))
@@ -2520,6 +2521,7 @@ predicate_statements (loop_p loop)
       for (gsi = gsi_start_bb (bb); !gsi_end_p (gsi);)
 	{
 	  gassign *stmt = dyn_cast <gassign *> (gsi_stmt (gsi));
+	  tree lhs;
 	  if (!stmt)
 	    ;
 	  else if (is_false_predicate (cond)
@@ -2574,15 +2576,36 @@ predicate_statements (loop_p loop)
 
 	      gsi_replace (&gsi, new_stmt, true);
 	    }
-	  else if (INTEGRAL_TYPE_P (TREE_TYPE (gimple_assign_lhs (stmt)))
-		   && TYPE_OVERFLOW_UNDEFINED
-			(TREE_TYPE (gimple_assign_lhs (stmt)))
+	  else if (((lhs = gimple_assign_lhs (stmt)), true)
+		   && (INTEGRAL_TYPE_P (TREE_TYPE (lhs))
+		       || POINTER_TYPE_P (TREE_TYPE (lhs)))
+		   && TYPE_OVERFLOW_UNDEFINED (TREE_TYPE (lhs))
 		   && arith_code_with_undefined_signed_overflow
 						(gimple_assign_rhs_code (stmt)))
 	    {
 	      gsi_remove (&gsi, true);
-	      gsi_insert_seq_before (&gsi, rewrite_to_defined_overflow (stmt),
-				     GSI_LAST_NEW_STMT);
+	      gimple_seq stmts = rewrite_to_defined_overflow (stmt);
+	      bool first = true;
+	      for (gimple_stmt_iterator gsi2 = gsi_start (stmts);
+		   !gsi_end_p (gsi2);)
+		{
+		  gassign *stmt2 = as_a <gassign *> (gsi_stmt (gsi2));
+		  gsi_remove (&gsi2, false);
+		  /* Make sure to move invariant conversions out of the
+		     loop.  */
+		  if (CONVERT_EXPR_CODE_P (gimple_assign_rhs_code (stmt2))
+		      && expr_invariant_in_loop_p (loop,
+						   gimple_assign_rhs1 (stmt2)))
+		    gsi_insert_on_edge_immediate (loop_preheader_edge (loop),
+						  stmt2);
+		  else if (first)
+		    {
+		      gsi_insert_before (&gsi, stmt2, GSI_NEW_STMT);
+		      first = false;
+		    }
+		  else
+		    gsi_insert_after (&gsi, stmt2, GSI_NEW_STMT);
+		}
 	    }
 	  else if (gimple_vdef (stmt))
 	    {
@@ -2601,7 +2624,7 @@ predicate_statements (loop_p loop)
 	      gimple_assign_set_rhs1 (stmt, ifc_temp_var (type, rhs, &gsi));
 	      update_stmt (stmt);
 	    }
-	  tree lhs = gimple_get_lhs (gsi_stmt (gsi));
+	  lhs = gimple_get_lhs (gsi_stmt (gsi));
 	  if (lhs && TREE_CODE (lhs) == SSA_NAME)
 	    ssa_names.add (lhs);
 	  gsi_next (&gsi);
diff --git a/gcc/tree-vectorizer.c b/gcc/tree-vectorizer.c
index 20daa31187d..4712dc6e7f9 100644
--- a/gcc/tree-vectorizer.c
+++ b/gcc/tree-vectorizer.c
@@ -852,7 +852,8 @@ vect_loop_vectorized_call (class loop *loop, gcond **cond)
   do
     {
       g = last_stmt (bb);
-      if (g)
+      if ((g && gimple_code (g) == GIMPLE_COND)
+	  || !single_succ_p (bb))
 	break;
       if (!single_pred_p (bb))
 	break;