This is the mail archive of the gcc-patches@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[patch] m68k.md: Don't use constraint 'y'.


Hi,

Attached is a patch to remove uses of constraint 'y' as it is
undefined.

While poking the m68k port to use define_constraint, I noticed that
the following patch from Andreas

2003-05-17  Andreas Schwab  <schwab@suse.de>

	* config/m68k/coff.h: Remove support for Sun FPA and Sun SKY board.
	* config/m68k/linux.h: Likewise.
	* config/m68k/m68k-none.h: Likewise.
	* config/m68k/netbsd-elf.h: Likewise.
	* config/m68k/sgs.h: Likewise.
	* config/m68k/m68k.h: Likewise.
	* config/m68k/m68k.md: Likewise.
	* config/m68k/m68k.c: Likewise.
	* doc/md.texi (Machine Constraints): Remove Sun FPA specific
	constraints.
	* doc/invoke.texi (Option Summary): Remove -mfpa.
	(M680x0 Options): Likewise.

removed the 'y' constraint but did not completely remove uses of it.

This patch simply removes remaining uses of the 'y' constraint.

Tested by building gcc.  OK to apply?

Kazu Hirata

2007-02-05  Kazu Hirata  <kazu@codesourcery.com>

	* m68k.md (anonymous insn, pushdi, anonymous insn): Don't use
	the 'y' constraint.

Index: m68k.md
===================================================================
--- m68k.md	(revision 121581)
+++ m68k.md	(working copy)
@@ -147,7 +147,7 @@ (define_mode_attr const [(SF "F") (DF "G
 
 (define_insn ""
   [(set (match_operand:DF 0 "push_operand" "=m")
-	(match_operand:DF 1 "general_operand" "ro<>fyE"))]
+	(match_operand:DF 1 "general_operand" "ro<>fE"))]
   ""
 {
   if (FP_REG_P (operands[1]))
@@ -157,7 +157,7 @@ (define_insn ""
 
 (define_insn "pushdi"
   [(set (match_operand:DI 0 "push_operand" "=m")
-	(match_operand:DI 1 "general_operand" "ro<>Fyi"))]
+	(match_operand:DI 1 "general_operand" "ro<>Fi"))]
   ""
 {
   return output_move_double (operands);
@@ -660,7 +660,7 @@ (define_insn ""
   ;; Notes: make sure no alternative allows g vs g.
   ;; We don't allow f-regs since fixed point cannot go in them.
   [(set (match_operand:SI 0 "nonimmediate_operand" "=g,d,a<")
-        (match_operand:SI 1 "general_src_operand" "daymSKT,n,i"))]
+        (match_operand:SI 1 "general_src_operand" "damSKT,n,i"))]
 
   "!TARGET_COLDFIRE"
 {


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]