[Bug c/30075] New: Missed optimizations with -fwhole-program -combine

[Bug c/30075] New: Missed optimizations with -fwhole-program -combine

[ajax at redhat dot com]

---
atropine:~/combine-bug% cat foo.c
int foo(void) {
    return 0;
}

atropine:~/combine-bug% cat bar.c 
extern int foo(void);

void *array[] = {
    foo
};

atropine:~/combine-bug% gcc -shared -fPIC -combine -fwhole-program -o libfoo.so
foo.c bar.c
atropine:~/combine-bug% nm libfoo.so | egrep '(foo)|(array)'
000000000000048c t foo
atropine:~/combine-bug% rpm -q gcc
gcc-4.1.1-30
---

foo is still emitted, when you'd think you could garbage-collect it, since the
array that references it has fallen out of existance.  If you use
-fvisibility=hidden instead of -fwhole-program the result is even worse, both
foo and array are emitted even though the resulting DSO does not give any
access to them.


-- 
           Summary: Missed optimizations with -fwhole-program -combine
           Product: gcc
           Version: 4.1.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: ajax at redhat dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30075

[Bug middle-end/30075] Missed optimizations with static variable and static function

[pinskia at gcc dot gnu dot org]

------- Comment #1 from pinskia at gcc dot gnu dot org  2006-12-05 18:46 -------
> If you use -fvisibility=hidden instead of -fwhole-program the result is even
> worse, both foo and array are emitted even though the resulting DSO does not
> give any access to them.

Well that is not GCC's fault really, you should use -ffunction-sections
-Wl,--gc-sections for that


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|Missed optimizations with - |Missed optimizations with
                   |fwhole-program -combine     |static variable and static
                   |                            |function


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30075

[Bug c/30075] Missed optimizations with -fwhole-program -combine

[ajax at redhat dot com]

------- Comment #2 from ajax at redhat dot com  2006-12-05 19:37 -------
Just to clarify, I neglected to use -O in the example above, but this behaviour
is still seen even with -O.


-- 

ajax at redhat dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|middle-end                  |c


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30075

[Bug middle-end/30075] Missed optimizations with -fwhole-program -combine

[pinskia at gcc dot gnu dot org]

-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |enhancement
          Component|c                           |middle-end


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30075

Re: [Bug middle-end/30075] Missed optimizations with -fwhole-program -combine

[Daniel Berlin]

I would not expect this to be fixed anytime soon.  I have yet to find
any real people who use either combine or -fwhole-program.  They use
*way* too much memory on real programs.  As a result, no real people
involved in optimization work on optimizers for them.

On 5 Dec 2006 19:38:51 -0000, pinskia at gcc dot gnu dot org
<gcc-bugzilla@gcc.gnu.org> wrote:
>
>
> --
>

[Bug middle-end/30075] Missed optimizations with -fwhole-program -combine

[dberlin at dberlin dot org]

------- Comment #3 from dberlin at gcc dot gnu dot org  2006-12-06 02:55 -------
Subject: Re:  Missed optimizations with -fwhole-program -combine

I would not expect this to be fixed anytime soon.  I have yet to find
any real people who use either combine or -fwhole-program.  They use
*way* too much memory on real programs.  As a result, no real people
involved in optimization work on optimizers for them.

On 5 Dec 2006 19:38:51 -0000, pinskia at gcc dot gnu dot org
<gcc-bugzilla@gcc.gnu.org> wrote:
>
>
> --
>


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30075

[Bug middle-end/30075] Missed optimizations with -fwhole-program -combine

[acahalan at gmail dot com]

------- Comment #4 from acahalan at gmail dot com  2007-06-26 03:10 -------
(In reply to comment #3)
> Subject: Re:  Missed optimizations with -fwhole-program -combine
> 
> I would not expect this to be fixed anytime soon.  I have yet to find
> any real people who use either combine or -fwhole-program.  They use
> *way* too much memory on real programs.  As a result, no real people
> involved in optimization work on optimizers for them.

I'm real, and I want to use those.

I can't use them because of bug #29171. They appear to be 100% unusable for
non-trivial programs because gcc WILL emit calls to memcpy and there isn't any
possible (documented at least) way to provide a memcpy. No matter how I
__attribute__ my memcpy, and no matter if I use -ffreestanding or not, gcc
wants to call an extern memcpy. Linking against a libc is against the
documented requirements for using these options.


-- 

acahalan at gmail dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |acahalan at gmail dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30075

[Bug middle-end/30075] Missed optimizations with -fwhole-program -combine

[dberlin at dberlin dot org]

------- Comment #5 from dberlin at gcc dot gnu dot org  2007-06-26 15:29 -------
Subject: Re:  Missed optimizations with -fwhole-program -combine

On 26 Jun 2007 03:10:26 -0000, acahalan at gmail dot com
<gcc-bugzilla@gcc.gnu.org> wrote:
>
>
> ------- Comment #4 from acahalan at gmail dot com  2007-06-26 03:10 -------
> (In reply to comment #3)
> > Subject: Re:  Missed optimizations with -fwhole-program -combine
> >
> > I would not expect this to be fixed anytime soon.  I have yet to find
> > any real people who use either combine or -fwhole-program.  They use
> > *way* too much memory on real programs.  As a result, no real people
> > involved in optimization work on optimizers for them.
>
> I'm real, and I want to use those.
>

That's nice and all, but i still wouldn't expect any work on them
until LTO is finished.
They are useless options right now.
I'd vote to remove them.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30075

Re: [Bug middle-end/30075] Missed optimizations with -fwhole-program -combine

[Daniel Berlin]

On 26 Jun 2007 03:10:26 -0000, acahalan at gmail dot com
<gcc-bugzilla@gcc.gnu.org> wrote:
>
>
> ------- Comment #4 from acahalan at gmail dot com  2007-06-26 03:10 -------
> (In reply to comment #3)
> > Subject: Re:  Missed optimizations with -fwhole-program -combine
> >
> > I would not expect this to be fixed anytime soon.  I have yet to find
> > any real people who use either combine or -fwhole-program.  They use
> > *way* too much memory on real programs.  As a result, no real people
> > involved in optimization work on optimizers for them.
>
> I'm real, and I want to use those.
>

That's nice and all, but i still wouldn't expect any work on them
until LTO is finished.
They are useless options right now.
I'd vote to remove them.

[Bug middle-end/30075] Missed optimizations with -fwhole-program -combine

[rguenth at gcc dot gnu dot org]

------- Comment #6 from rguenth at gcc dot gnu dot org  2010-09-17 09:06 -------
-combine has been removed from GCC 4.6 in favor of LTO, closing as WONTFIX.


-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |WONTFIX


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=30075