Re: c/693

Re: c/693

[jsm28]

The following reply was made to PR c/693; it has been noted by GNATS.

From: jsm28@gcc.gnu.org
To: gcc-gnats@gcc.gnu.org, jsm28@cam.ac.uk, nobody@gcc.gnu.org,
  travis.moulton@motorola.com
Cc:  
Subject: Re: c/693
Date: 21 May 2001 17:07:49 -0000

 Synopsis: Optimizer ignores possible zero result of an unsigned multiply
 
 State-Changed-From-To: feedback->closed
 State-Changed-By: jsm28
 State-Changed-When: Mon May 21 10:07:49 2001
 State-Changed-Why:
     Submitter reports bug was in binutils not GCC.
 
 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=693&database=gcc

Re: c/693

[Travis Moulton]

The following reply was made to PR c/693; it has been noted by GNATS.

From: Travis Moulton <travis@mail.wm.sps.mot.com>
To: jsm28@gcc.gnu.org
Cc: gcc-gnats@gcc.gnu.org, jsm28@cam.ac.uk
Subject: Re: c/693
Date: Mon, 21 May 2001 12:58:14 -0400 (EDT)

 My bad, this was a binutils bug.  In Version 2.9.1, opcodes/m68k-opc.c
 listed the opcodes of mulsl as being:
 
 {"mulsw",	one(0140700),		one(0170700), ";wDd", m68000up|mcf5200 },
 {"mulsl",	two(0046000,004000), two(0177700,0107770), ";lD1", m68020up|cpu32 },
 {"mulsl",	two(0046000,004000), two(0177700,0107770), "DsD1", mcf5200 },
 {"mulsl",	two(0046000,004000), two(0177700,0107770), "asD1", mcf5200 },
 {"mulsl",	two(0046000,004000), two(0177700,0107770), "+sD1", mcf5200 },
 {"mulsl",	two(0046000,004000), two(0177700,0107770), "-sD1", mcf5200 },
 {"mulsl",	two(0046000,004000), two(0177700,0107770), "dsD1", mcf5200 },
 {"mulsl",	two(0046000,006000), two(0177700,0107770), ";lD3D1",m68020up|cpu32 },
 
 When they should be 
 
 {"mulsw",	one(0140700),		one(0170700), ";wDd",  m68000up|mcf5200 },
 {"mulsl",	two(0046000,004000), two(0177700,0107770), ";lD1", m68020up|cpu32 },
 {"mulsl",	two(0046000,004000), two(0177700,0107770), "DsD1", mcf5200 },
 {"mulsl",	two(0046020,004000), two(0177700,0107770), "asD1", mcf5200 },
 {"mulsl",	two(0046030,004000), two(0177700,0107770), "+sD1", mcf5200 },
 {"mulsl",	two(0046040,004000), two(0177700,0107770), "-sD1", mcf5200 },
 {"mulsl",	two(0046050,004000), two(0177700,0107770), "dsD1", mcf5200 },
 {"mulsl",	two(0046000,006000), two(0177700,0107770), ";lD3D1",m68020up|cpu32 }, 
                     ^^^^^^^
 ...Actually, I'd guess the last one is wrong too, but it doesn't effect my
 platform.  I leave it to see if it has been fixed, pass along a patch,
 etc.
 
 	-Travis
 
 On 20 May 2001 jsm28@gcc.gnu.org wrote:
 
 > Synopsis: Optimizer ignores possible zero result of an unsigned multiply
 > 
 > State-Changed-From-To: open->feedback
 > State-Changed-By: jsm28
 > State-Changed-When: Sun May 20 07:31:49 2001
 > State-Changed-Why:
 >     The attachment seems to be missing from this PR; could
 >     you please resubmit it?
 >     
 >     Does this problem still occur with current GCC 3.0 snapshots?
 > 
 > http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=693&database=gcc
 > 
 

Re: c/693

[jsm28]

The following reply was made to PR c/693; it has been noted by GNATS.

From: jsm28@gcc.gnu.org
To: gcc-gnats@gcc.gnu.org, jsm28@cam.ac.uk, nobody@gcc.gnu.org,
  travis.moulton@motorola.com
Cc:  
Subject: Re: c/693
Date: 20 May 2001 14:31:49 -0000

 Synopsis: Optimizer ignores possible zero result of an unsigned multiply
 
 State-Changed-From-To: open->feedback
 State-Changed-By: jsm28
 State-Changed-When: Sun May 20 07:31:49 2001
 State-Changed-Why:
     The attachment seems to be missing from this PR; could
     you please resubmit it?
     
     Does this problem still occur with current GCC 3.0 snapshots?
 
 http://gcc.gnu.org/cgi-bin/gnatsweb.pl?cmd=view&pr=693&database=gcc