public inbox for binutils@sourceware.org
 help / color / mirror / Atom feed
* m68k fix request for 2.11.x release branch
@ 2001-08-28  8:08 mike stump
  2001-08-28 22:26 ` Alan Modra
  0 siblings, 1 reply; 3+ messages in thread
From: mike stump @ 2001-08-28  8:08 UTC (permalink / raw)
  To: binutils

Below is a patch that is necessary for testcase just below.  Could
this be checked in to the 2.11.x release branch?  It used to work, but
someone broke it, and then Nick fixed it.  The testcase works in the
top of tree.  I tested the testcase with the patch in the 2.11.x
release branch, and it seems to work with it.

Thanks.

#NO_APP
.text
Ltext0:
	.balign 2
.globl _foobar
_foobar:
	.stabd 68,0,242
	link a6,#-24
	movel d0,d1
	movel d1,d0
	lsll #2,d0
	lea a6@(-44),a0
	moveq #110,d1
	moveq #110,d1
	moveq #110,d1
	cmpl a0@(d0:l),d1
	jeq L38
	jra L39
	.balign 2
L39:
	jra L37
	.balign 2
L38:
	.stabd 68,0,362
	addql #1,a6@(-4)
	jra L36
	.balign 2


2001-07-06  Nick Clifton  <nickc@cambridge.redhat.com>

	* config/tc-m68k.c (md_estimate_size_before_relax): Catch and
	ignore empty, ineffectual alignment frags when deciding if a
	branch can be short.

Index: tc-m68k.c
===================================================================
RCS file: /cvs/src/src/gas/config/tc-m68k.c,v
retrieving revision 1.17.2.2
diff -p -c -r1.17.2.2 tc-m68k.c
*** tc-m68k.c	2001/06/11 10:04:48	1.17.2.2
--- tc-m68k.c	2001/08/28 15:02:33
*************** md_estimate_size_before_relax (fragP, se
*** 4692,4704 ****
  	  fragS *l;
  
  	  stop = symbol_get_frag (fragP->fr_symbol);
  	  for (l = fragP->fr_next; l != stop; l = l->fr_next)
- 	    if (l->fr_fix + l->fr_var != 0)
- 	      break;
- 	  if (l == stop)
  	    {
! 	      fragP->fr_subtype = TAB (TABTYPE (fragP->fr_subtype), SHORT);
  	    }
  	}
        break;
      default:
--- 4692,4716 ----
  	  fragS *l;
  
  	  stop = symbol_get_frag (fragP->fr_symbol);
+ 
  	  for (l = fragP->fr_next; l != stop; l = l->fr_next)
  	    {
! 	      /* Catch empty alignment frags whoes fr_offset field
! 		 is an alignment requirement of 2 bytes.  The check
! 		 below will misinterpret this as evidence that real
! 		 code exists between the symbol and the instruction
! 		 and so will not convert the short jump into a word
! 		 jump.  */
! 	      if (l->fr_fix == 0
! 		  && l->fr_var == 1
! 		  && (l->fr_type == rs_align || l->fr_type == rs_align_code))
! 		continue;
!     
! 	      if (l->fr_fix + l->fr_var != 0)
! 		break;
  	    }
+ 	  if (l == stop)
+ 	    fragP->fr_subtype = TAB (TABTYPE (fragP->fr_subtype), SHORT);
  	}
        break;
      default:
--------------------

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: m68k fix request for 2.11.x release branch
  2001-08-28  8:08 m68k fix request for 2.11.x release branch mike stump
@ 2001-08-28 22:26 ` Alan Modra
  2001-08-29 11:27   ` Philip Blundell
  0 siblings, 1 reply; 3+ messages in thread
From: Alan Modra @ 2001-08-28 22:26 UTC (permalink / raw)
  To: mike stump; +Cc: binutils

On Tue, Aug 28, 2001 at 08:08:45AM -0700, mike stump wrote:

> It used to work, but someone broke it,

I think I probably broke it on mainline.

> and then Nick fixed it.

In case anyone is wondering, I took Nick's fix out of mainline when some
other changes made it no longer necessary.

I think the fix is OK for the branch, but how long do we want to continue
with 2.11.x?  Isn't it time for a 2.12 branch off mainline?

Alan

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: m68k fix request for 2.11.x release branch
  2001-08-28 22:26 ` Alan Modra
@ 2001-08-29 11:27   ` Philip Blundell
  0 siblings, 0 replies; 3+ messages in thread
From: Philip Blundell @ 2001-08-29 11:27 UTC (permalink / raw)
  To: binutils

>I think the fix is OK for the branch, but how long do we want to continue
>with 2.11.x?  Isn't it time for a 2.12 branch off mainline?

Soon, yeah.  I was planning to make one more release from the 2.11 branch 
within the next few weeks, then declare that branch dead and start thinking 
about a 2.12.

p.

-- 
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.5 (GNU/Linux)
Comment: Exmh version 2.1.1 10/15/1999 (debian)

iD8DBQE7jTQIVTLPJe9CT30RAi4zAJ4xAMlJzjnWb2p/vsj9sUGju4/tAwCfZxed
TCfW8Vse7YCZ9MVLhDRSGlc=
=I740
-----END PGP SIGNATURE-----

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2001-08-29 11:27 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2001-08-28  8:08 m68k fix request for 2.11.x release branch mike stump
2001-08-28 22:26 ` Alan Modra
2001-08-29 11:27   ` Philip Blundell

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).