PATCH: add missing `return' to cgen_get_insn_value

PATCH: add missing `return' to cgen_get_insn_value

[Greg McGary]

Here's a fun one that worked fine on ix86 hosts but
returned bad data on SPARC.

OK to commit?

2001-02-21  Greg McGary  <greg@mcgary.org>

	* cgen-opc.c (cgen_get_insn_value): Add missing `return'.

Index: cgen-opc.c
===================================================================
RCS file: /cvs/src/src/opcodes/cgen-opc.c,v
retrieving revision 1.5
diff -u -p -r1.5 cgen-opc.c
--- cgen-opc.c	2001/01/02 16:34:07	1.5
+++ cgen-opc.c	2001/02/21 20:27:13
@@ -1,6 +1,6 @@
 /* CGEN generic opcode support.
 
-   Copyright (C) 1996, 1997, 1998, 1999, 2000 Free Software Foundation, Inc.
+   Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001 Free Software Foundation, Inc.
 
    This file is part of the GNU Binutils and GDB, the GNU debugger.
 
@@ -374,7 +374,7 @@ cgen_get_insn_value (cd, buf, length)
      unsigned char *buf;
      int length;
 {
-  bfd_get_bits (buf, length, cd->insn_endian == CGEN_ENDIAN_BIG);
+  return bfd_get_bits (buf, length, cd->insn_endian == CGEN_ENDIAN_BIG);
 }
 
 /* Cover function to store an insn value properly byteswapped.  */

Re: PATCH: add missing `return' to cgen_get_insn_value

[Ben Elliston]

>>>>> "Greg" == Greg McGary <greg@mcgary.org> writes:

  Greg> Here's a fun one that worked fine on ix86 hosts but
  Greg> returned bad data on SPARC.

  Greg> OK to commit?

Please do.

Re: PATCH: add missing `return' to cgen_get_insn_value

[Frank Ch. Eigler]

Hi -

On Wed, Feb 21, 2001 at 01:29:59PM -0700, Greg McGary wrote:
: Here's a fun one that worked fine on ix86 hosts but
: returned bad data on SPARC.
: 
: OK to commit?
: 
: 2001-02-21  Greg McGary  <greg@mcgary.org>
: 
: 	* cgen-opc.c (cgen_get_insn_value): Add missing `return'.

Yes, please!  As it turns out, our very own Jeff Johnston found
this same bug, a few minutes ago.


- FChE
-- 
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.4 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE6lD0AVZbdDOm/ZT0RAtYxAJ4uSqIucq9a4vbQFvNWqeh45yaBCACfTnVy
s1hwhFDZ8idPwLR9bwYw/Sc=
=mNKw
-----END PGP SIGNATURE-----

Re: PATCH: add missing `return' to cgen_get_insn_value

[Greg McGary]

"Frank Ch. Eigler" <fche@redhat.com> writes:

> Yes, please!  As it turns out, our very own Jeff Johnston found
> this same bug, a few minutes ago.

I committed it approx 20 minutes ago.

I guess people don't use machines with register windows very often for
hosting CGEN tools, eh?

I had been using SPARC for hosting my port for months, but only
recently started using & debugging relocations.  The bug surfaced in a
fixup that was supposed to write only the low-order 16 bits of an
insn.  It clobbered the entire insn because it failed to properly read
during the read-fixup-write cycle.

Greg

Re: PATCH: add missing `return' to cgen_get_insn_value

[Greg McGary]

"Frank Ch. Eigler" <fche@redhat.com> writes:

> Yes, please!  As it turns out, our very own Jeff Johnston found
> this same bug, a few minutes ago.

The puzzling thing is that no warnings surfaced when compiled with
`-W -Wall' from
gcc version 2.96 20000731 (Red Hat Linux 7.0)

I did get a warning from
gcc version 2.95.2 19991024 (release)
gcc version egcs-2.91.66 19990314/Linux (egcs-1.1.2 release)

I haven't tried a pre-3.0 GCC, so I don't know if it's still a
regression.

Re: PATCH: add missing `return' to cgen_get_insn_value

[Ben Elliston]

>>>>> "Greg" == Greg McGary <greg@mcgary.org> writes:

  Greg> The puzzling thing is that no warnings surfaced when compiled with
  Greg> `-W -Wall' from
  Greg> gcc version 2.96 20000731 (Red Hat Linux 7.0)

That is surprising.  A function with a return type failing to return a
value is a pretty common warning!

Ben

re: PATCH: add missing `return' to cgen_get_insn_value

[matthew green]

   
   I guess people don't use machines with register windows very often for
   hosting CGEN tools, eh?
   
   I had been using SPARC for hosting my port for months, but only
   recently started using & debugging relocations.  The bug surfaced in a
   fixup that was supposed to write only the low-order 16 bits of an
   insn.  It clobbered the entire insn because it failed to properly read
   during the read-fixup-write cycle.


actually, 95% of my work with cgen is under solaris 2.6/sparc.


.mrg.