re PR target/77924 (-mfloat128-type change broke AIX)

2016-10-11  Michael Meissner  <meissner@linux.vnet.ibm.com>

	PR target/77924
	* config/rs6000/rs6000.c (rs6000_init_builtins): Only create the
	distinct __ibm128 IBM extended double type if long doubles are
	128-bits and the default format for long double is IEEE 128-bit.

From-SVN: r240994

gcc/ChangeLog

+2016-10-11  Michael Meissner  <meissner@linux.vnet.ibm.com>
+
+	PR target/77924
+	* config/rs6000/rs6000.c (rs6000_init_builtins): Only create the
+	distinct __ibm128 IBM extended double type if long doubles are
+	128-bits and the default format for long double is IEEE 128-bit.
+
 2016-10-11  Richard Biener  <rguenther@suse.de>
 
 	* dwarf2out.c (DEBUG_STR_OFFSETS_SECTION): Remove conditional.

gcc/config/rs6000/rs6000.c

@@ -16572,10 +16572,10 @@ rs6000_init_builtins (void)
      floating point, we need make sure the type is non-zero or else self-test
      fails during bootstrap.
 
-     We don't register a built-in type for __ibm128 or __float128 if the type
-     is the same as long double.  Instead we add a #define for __ibm128 or
-     __float128 in rs6000_cpu_cpp_builtins to long double.  */
-  if (TARGET_IEEEQUAD || !TARGET_LONG_DOUBLE_128)
+     We don't register a built-in type for __ibm128 if the type is the same as
+     long double.  Instead we add a #define for __ibm128 in
+     rs6000_cpu_cpp_builtins to long double.  */
+  if (TARGET_LONG_DOUBLE_128 && FLOAT128_IEEE_P (TFmode))
     {
       ibm128_float_type_node = make_node (REAL_TYPE);
       TYPE_PRECISION (ibm128_float_type_node) = 128;