[Bug optimization/11457] New:

[Bug optimization/11457] New:

[gcc-bugzilla at gcc dot gnu dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11457

           Product: gcc
           Version: 3.3.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: optimization
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: user42 at zip dot com dot au
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: i386-pc-linux-gnu
  GCC host triplet: i386-pc-linux-gnu
GCC target triplet: i386-pc-linux-gnu

When presented with tests like "n>=0" followed by "n==0", two i386
"test" instructions are generated, where one would suffice.  This is
apparently a regression from some point, since for instance gcc 2.95.4
generates only one.

Environment:
System: Linux blah 2.2.15 #1 Tue Apr 25 17:13:48 EST 2000 i586 unknown unknown GNU/Linux
Architecture: i586
	<machine, os, target, libraries (multiple lines)>
host: i386-pc-linux-gnu
build: i386-pc-linux-gnu
target: i386-pc-linux-gnu
configured with: ../src/configure -v --enable-languages=c,c++,java,f77,pascal,objc,ada,treelang --prefix=/usr --mandir=/usr/share/man --infodir=/usr/share/info --with-gxx-include-dir=/usr/include/c++/3.3 --enable-shared --with-system-zlib --enable-nls --without-included-gettext --enable-__cxa_atexit --enable-clocale=gnu --enable-debug --enable-java-gc=boehm --enable-java-awt=xlib --enable-objc-gc i386-linux

How-To-Repeat:
A file foo.c containing

	void
	foo (int n)
	{
	  if (n >= 0)
	    {
	      if (n == 0)
	        zero ();
	      else
	        non_negative ();
	    }
	  else
	    negative ();
	}

compiled with

	gcc -O2 -S foo.c

produces foo.s containing the sequence

        testl   %eax, %eax
        js      .L2
        testl   %eax, %eax
        jne     .L3

I think the second "testl" is unnecessary, and that the code would be
smaller and probably faster if it was omitted.

It seems this was done in the past, for instance a recent Debian
packaged gcc 2.95.4 does it,

        testl %eax,%eax
        jl .L3
        jne .L4

[Bug target/11457] [3.3/3.4 Regression] redundant compare

[pinskia at physics dot uc dot edu]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11457


pinskia at physics dot uc dot edu changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
          Component|optimization                |target
     Ever Confirmed|                            |1
           Keywords|                            |pessimizes-code
           Priority|P3                          |P2
   Last reconfirmed|0000-00-00 00:00:00         |2003-07-07 22:34:41
               date|                            |
            Summary|                            |[3.3/3.4 Regression]
                   |                            |redundant compare
   Target Milestone|---                         |3.4


------- Additional Comments From pinskia at physics dot uc dot edu  2003-07-07 22:34 -------
I can confirm this on 3.3.1 (20030616) and the mainline (20030707), 3.2.3, and 3.0.4 but 2.95.3 
did not produce the redundant compare so this is a regression.
On PPC, I do not get a redundant compare though:
_foo:
        cmpwi cr7,r3,0
        blt- cr7,L2
        bne- cr7,L3
        b L_zero$stub
L3:
        b L_non_negative$stub
L2:
        b L_negative$stub
Or on Vax:
foo:
        .word 0x0
        subl2 $4,%sp
        movl 4(%ap),%r0
        jlss .L2
        jneq .L3
        calls $0,zero
        ret
.L3:
        calls $0,non_negative
        ret
.L2:
        calls $0,negative
        ret
Or arm:
foo:
        @ args = 0, pretend = 0, frame = 0
        @ frame_needed = 0, uses_anonymous_args = 0
        @ link register save eliminated.
        cmp     r0, #0
        @ lr needed for prologue
        blt     .L2
        bne     .L3
        b       zero
.L3:
        b       non_negative
.L2:
        b       negative
So it looks very ia32 specific.

[Bug target/11457] [3.3/3.4 Regression] redundant compare

[pinskia at physics dot uc dot edu]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11457


pinskia at physics dot uc dot edu changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|3.4                         |3.3.2


------- Additional Comments From pinskia at physics dot uc dot edu  2003-07-19 20:55 -------
Moving target to 3.3.2 since this is a regression.

[Bug target/11457] [3.3/3.4 Regression] redundant compare

[pinskia at gcc dot gnu dot org]

PLEASE REPLY TO gcc-bugzilla@gcc.gnu.org ONLY, *NOT* gcc-bugs@gcc.gnu.org.

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=11457


pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |DUPLICATE


------- Additional Comments From pinskia at gcc dot gnu dot org  2003-08-24 19:43 -------
This is a dup of bug 1532 which has a different testcase but it shows the same bug.

*** This bug has been marked as a duplicate of 1532 ***