re PR rtl-optimization/50107 ([IRA, i386] allocates registers in very non-optimal way)

2011-08-17  Vladimir Makarov  <vmakarov@redhat.com>

	PR rtl-optimization/50107
	* ira-int.h (ira_hard_reg_not_in_set_p): Remove.
	(ira_hard_reg_in_set_p): New.

	* ira-color.c (calculate_saved_nregs): New.
	(assign_hard_reg): Use it.  Set up allocated_hard_reg_p for all
	hard regs.
	(allocno_reload_assign, fast_allocation): Use
	ira_hard_reg_set_intersection_p instead of
	ira_hard_reg_not_in_set_p.

	* ira.c (setup_reg_renumber): Use
	ira_hard_reg_set_intersection_p instead of
	ira_hard_reg_not_in_set_p.
	(setup_allocno_assignment_flags, calculate_allocation_cost): Use
	ira_hard_reg_in_set_p instead of ira_hard_reg_not_in_set_p.

	* ira-costs.c (ira_tune_allocno_costs): Use
	ira_hard_reg_set_intersection_p instead of
	ira_hard_reg_not_in_set_p.

From-SVN: r177865

gcc/ChangeLog

+2011-08-17  Vladimir Makarov  <vmakarov@redhat.com>
+
+	PR rtl-optimization/50107
+	* ira-int.h (ira_hard_reg_not_in_set_p): Remove.
+	(ira_hard_reg_in_set_p): New.
+
+	* ira-color.c (calculate_saved_nregs): New.
+	(assign_hard_reg): Use it.  Set up allocated_hard_reg_p for all
+	hard regs.
+	(allocno_reload_assign, fast_allocation): Use
+	ira_hard_reg_set_intersection_p instead of
+	ira_hard_reg_not_in_set_p.
+
+	* ira.c (setup_reg_renumber): Use
+	ira_hard_reg_set_intersection_p instead of
+	ira_hard_reg_not_in_set_p.
+	(setup_allocno_assignment_flags, calculate_allocation_cost): Use
+	ira_hard_reg_in_set_p instead of ira_hard_reg_not_in_set_p.
+
+	* ira-costs.c (ira_tune_allocno_costs): Use
+	ira_hard_reg_set_intersection_p instead of
+	ira_hard_reg_not_in_set_p.
+
 2011-08-18  H.J. Lu  <hongjiu.lu@intel.com>
 	    Igor Zamyatin <igor.zamyatin@intel.com>
 

gcc/ira-color.c

@@ -1519,6 +1519,26 @@ check_hard_reg_p (ira_allocno_t a, int hard_regno,
     }
   return j == nregs;
 }
+#ifndef HONOR_REG_ALLOC_ORDER
+
+/* Return number of registers needed to be saved and restored at
+   function prologue/epilogue if we allocate HARD_REGNO to hold value
+   of MODE.  */
+static int
+calculate_saved_nregs (int hard_regno, enum machine_mode mode)
+{
+  int i;
+  int nregs = 0;
+
+  ira_assert (hard_regno >= 0);
+  for (i = hard_regno_nregs[hard_regno][mode] - 1; i >= 0; i--)
+    if (!allocated_hardreg_p[hard_regno + i]
+	&& !TEST_HARD_REG_BIT (call_used_reg_set, hard_regno + i)
+	&& !LOCAL_REGNO (hard_regno + i))
+      nregs++;
+  return nregs;
+}
+#endif
 
 /* Choose a hard register for allocno A.  If RETRY_P is TRUE, it means
    that the function called from function
@@ -1547,7 +1567,7 @@ static bool
 assign_hard_reg (ira_allocno_t a, bool retry_p)
 {
   HARD_REG_SET conflicting_regs[2], profitable_hard_regs[2];
-  int i, j, hard_regno, best_hard_regno, class_size;
+  int i, j, hard_regno, best_hard_regno, class_size, saved_nregs;
   int cost, mem_cost, min_cost, full_cost, min_full_cost, nwords, word;
   int *a_costs;
   enum reg_class aclass;
@@ -1716,16 +1736,14 @@ assign_hard_reg (ira_allocno_t a, bool retry_p)
       cost = costs[i];
       full_cost = full_costs[i];
 #ifndef HONOR_REG_ALLOC_ORDER
-      if (! allocated_hardreg_p[hard_regno]
-	  && ira_hard_reg_not_in_set_p (hard_regno, mode, call_used_reg_set)
-	  && !LOCAL_REGNO (hard_regno))
+      if ((saved_nregs = calculate_saved_nregs (hard_regno, mode)) != 0)
 	/* We need to save/restore the hard register in
 	   epilogue/prologue.  Therefore we increase the cost.  */
 	{
-	  /* ??? If only part is call clobbered.  */
 	  rclass = REGNO_REG_CLASS (hard_regno);
-	  add_cost = (ira_memory_move_cost[mode][rclass][0]
-		      + ira_memory_move_cost[mode][rclass][1] - 1);
+	  add_cost = ((ira_memory_move_cost[mode][rclass][0]
+		       + ira_memory_move_cost[mode][rclass][1])
+		      * saved_nregs / hard_regno_nregs[hard_regno][mode] - 1);
 	  cost += add_cost;
 	  full_cost += add_cost;
 	}
@@ -1748,7 +1766,10 @@ assign_hard_reg (ira_allocno_t a, bool retry_p)
     }
  fail:
   if (best_hard_regno >= 0)
-    allocated_hardreg_p[best_hard_regno] = true;
+    {
+      for (i = hard_regno_nregs[best_hard_regno][mode] - 1; i >= 0; i--)
+	allocated_hardreg_p[best_hard_regno + 1] = true;
+    }
   ALLOCNO_HARD_REGNO (a) = best_hard_regno;
   ALLOCNO_ASSIGNED_P (a) = true;
   if (best_hard_regno >= 0)
@@ -3975,8 +3996,8 @@ allocno_reload_assign (ira_allocno_t a, HARD_REG_SET forbidden_regs)
 	       : ALLOCNO_HARD_REG_COSTS (a)[ira_class_hard_reg_index
 					    [aclass][hard_regno]]));
       if (ALLOCNO_CALLS_CROSSED_NUM (a) != 0
-	  && ! ira_hard_reg_not_in_set_p (hard_regno, ALLOCNO_MODE (a),
-					  call_used_reg_set))
+	  && ira_hard_reg_set_intersection_p (hard_regno, ALLOCNO_MODE (a),
+					      call_used_reg_set))
 	{
 	  ira_assert (flag_caller_saves);
 	  caller_save_needed = 1;
@@ -4467,7 +4488,7 @@ fast_allocation (void)
 	      && hard_regno <= LAST_STACK_REG)
 	    continue;
 #endif
-	  if (!ira_hard_reg_not_in_set_p (hard_regno, mode, conflict_hard_regs)
+	  if (ira_hard_reg_set_intersection_p (hard_regno, mode, conflict_hard_regs)
 	      || (TEST_HARD_REG_BIT
 		  (ira_prohibited_class_mode_regs[aclass][mode], hard_regno)))
 	    continue;

gcc/ira-costs.c

@@ -2072,9 +2072,9 @@ ira_tune_allocno_costs (void)
 	      skip_p = false;
 	      FOR_EACH_ALLOCNO_OBJECT (a, obj, oi)
 		{
-		  if (! ira_hard_reg_not_in_set_p (regno, mode,
-						   OBJECT_CONFLICT_HARD_REGS
-						   (obj)))
+		  if (ira_hard_reg_set_intersection_p (regno, mode,
+						       OBJECT_CONFLICT_HARD_REGS
+						       (obj)))
 		    {
 		      skip_p = true;
 		      break;
@@ -2084,7 +2084,7 @@ ira_tune_allocno_costs (void)
 		continue;
 	      rclass = REGNO_REG_CLASS (regno);
 	      cost = 0;
-	      if (! ira_hard_reg_not_in_set_p (regno, mode, call_used_reg_set)
+	      if (ira_hard_reg_set_intersection_p (regno, mode, call_used_reg_set)
 		  || HARD_REGNO_CALL_PART_CLOBBERED (regno, mode))
 		cost += (ALLOCNO_CALL_FREQ (a)
 			 * (ira_memory_move_cost[mode][rclass][0]

gcc/ira-int.h

@@ -1323,17 +1323,17 @@ hard_reg_set_size (HARD_REG_SET set)
 }
 
 /* The function returns TRUE if hard registers starting with
-   HARD_REGNO and containing value of MODE are not in set
+   HARD_REGNO and containing value of MODE are fully in set
    HARD_REGSET.  */
 static inline bool
-ira_hard_reg_not_in_set_p (int hard_regno, enum machine_mode mode,
-			   HARD_REG_SET hard_regset)
+ira_hard_reg_in_set_p (int hard_regno, enum machine_mode mode,
+		       HARD_REG_SET hard_regset)
 {
   int i;
 
   ira_assert (hard_regno >= 0);
   for (i = hard_regno_nregs[hard_regno][mode] - 1; i >= 0; i--)
-    if (TEST_HARD_REG_BIT (hard_regset, hard_regno + i))
+    if (!TEST_HARD_REG_BIT (hard_regset, hard_regno + i))
       return false;
   return true;
 }

gcc/ira.c

@@ -1953,8 +1953,8 @@ setup_reg_renumber (void)
 				      reg_class_contents[pclass]);
 	    }
 	  if (ALLOCNO_CALLS_CROSSED_NUM (a) != 0
-	      && ! ira_hard_reg_not_in_set_p (hard_regno, ALLOCNO_MODE (a),
-					      call_used_reg_set))
+	      && ira_hard_reg_set_intersection_p (hard_regno, ALLOCNO_MODE (a),
+						  call_used_reg_set))
 	    {
 	      ira_assert (!optimize || flag_caller_saves
 			  || regno >= ira_reg_equiv_len
@@ -1992,10 +1992,10 @@ setup_allocno_assignment_flags (void)
 				|| ALLOCNO_EMIT_DATA (a)->mem_optimized_dest_p
 				|| (ALLOCNO_MEMORY_COST (a)
 				    - ALLOCNO_CLASS_COST (a)) < 0);
-      ira_assert (hard_regno < 0
-		  || ! ira_hard_reg_not_in_set_p (hard_regno, ALLOCNO_MODE (a),
-						  reg_class_contents
-						  [ALLOCNO_CLASS (a)]));
+      ira_assert
+	(hard_regno < 0
+	 || ira_hard_reg_in_set_p (hard_regno, ALLOCNO_MODE (a),
+				   reg_class_contents[ALLOCNO_CLASS (a)]));
     }
 }
 
@@ -2013,9 +2013,9 @@ calculate_allocation_cost (void)
     {
       hard_regno = ALLOCNO_HARD_REGNO (a);
       ira_assert (hard_regno < 0
-		  || ! ira_hard_reg_not_in_set_p
-		       (hard_regno, ALLOCNO_MODE (a),
-			reg_class_contents[ALLOCNO_CLASS (a)]));
+		  || (ira_hard_reg_in_set_p
+		      (hard_regno, ALLOCNO_MODE (a),
+		       reg_class_contents[ALLOCNO_CLASS (a)])));
       if (hard_regno < 0)
 	{
 	  cost = ALLOCNO_MEMORY_COST (a);