public inbox for binutils@sourceware.org
 help / color / mirror / Atom feed
* [MIPS PATCH RFA] Assume that macros won't be generating MIPS-3D insns.
@ 2002-04-24 14:36 cgd
  2002-04-24 14:46 ` Eric Christopher
  0 siblings, 1 reply; 2+ messages in thread
From: cgd @ 2002-04-24 14:36 UTC (permalink / raw)
  To: binutils

[ sorry if this comes through as a duplicate; tried to send it earlier
  and I think it got eaten by some DNS lossage... ]

The patch below undoes one of the changes that I added as part of the
MIPS-3D support.  I enabled support in macro_build() for inclusion of
MIPS-3D instructions in the macro expansions.

On further reflection, I'd like to undo that because:

* the likelyhood of macro instructions which generate expansions
  including MIPS-3D instructions is currently nil, and hopefully we
  won't be adding (m)any new macro instructions, let alone ones which
  use MIPS-3D opcodes.

* it sets bad precedent: future ASE code would have to be put there as
  well, which leads to much duplication of the opcode operand
  handling, etc., when in reality it really isn't necessary.


Tested only by gmake && gmake check on binutils, host x86-linux,
targets mips-linux and mips64-elf.  No failures.


chris
===================================================================
2002-04-24  Chris G. Demetriou  <cgd@broadcom.com>

	* config/tc-mips.c (macro_build): Do _not_ allow MIPS-3D
	instructions to be generated by macros.

Index: config/tc-mips.c
===================================================================
RCS file: /cvs/src/src/gas/config/tc-mips.c,v
retrieving revision 1.117
diff -u -p -r1.117 tc-mips.c
--- config/tc-mips.c	22 Apr 2002 21:50:44 -0000	1.117
+++ config/tc-mips.c	24 Apr 2002 18:28:21 -0000
@@ -2913,12 +2913,11 @@ macro_build (place, counter, ep, name, f
   /* Search until we get a match for NAME.  */
   while (1)
     {
+      /* It is assumed here that macros will never generate 
+         MIPS-3D instructions.  */
       if (strcmp (fmt, insn.insn_mo->args) == 0
 	  && insn.insn_mo->pinfo != INSN_MACRO
-	  && OPCODE_IS_MEMBER (insn.insn_mo,
-			       (mips_opts.isa
-	      		        | (mips_opts.ase_mips3d ? INSN_MIPS3D : 0)),
-			       mips_arch)
+	  && OPCODE_IS_MEMBER (insn.insn_mo, mips_opts.isa, mips_arch)
 	  && (mips_arch != CPU_R4650 || (insn.insn_mo->pinfo & FP_D) == 0))
 	break;
 

^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [MIPS PATCH RFA] Assume that macros won't be generating MIPS-3D insns.
  2002-04-24 14:36 [MIPS PATCH RFA] Assume that macros won't be generating MIPS-3D insns cgd
@ 2002-04-24 14:46 ` Eric Christopher
  0 siblings, 0 replies; 2+ messages in thread
From: Eric Christopher @ 2002-04-24 14:46 UTC (permalink / raw)
  To: cgd; +Cc: binutils


> ===================================================================
> 2002-04-24  Chris G. Demetriou  <cgd@broadcom.com>
> 
> 	* config/tc-mips.c (macro_build): Do _not_ allow MIPS-3D
> 	instructions to be generated by macros.
> 

OK.

-eric

-- 
A fire drill does not demand
a fire.

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2002-04-24 21:46 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2002-04-24 14:36 [MIPS PATCH RFA] Assume that macros won't be generating MIPS-3D insns cgd
2002-04-24 14:46 ` Eric Christopher

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).