[Bug lto/53604] New: ld reports errors using lto after upgrading from gcc-4.6.2 to gcc-4.7.0

[Bug lto/53604] New: ld reports errors using lto after upgrading from gcc-4.6.2 to gcc-4.7.0

[paul.scruby at ghco dot co.uk]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53604

             Bug #: 53604
           Summary: ld reports errors using lto after upgrading from
                    gcc-4.6.2 to gcc-4.7.0
    Classification: Unclassified
           Product: gcc
           Version: 4.7.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: lto
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: paul.scruby@ghco.co.uk


I recently tried to upgrade from gcc-4.6.2 to gcc-4.7.0.  I have been using the
lto with gcc-4.6.2 and binutils-2.22 for a while with no problems.

However, since I upgraded to gcc-4.7.0 ld is getting "hidden symbol isn't
defined" errors, something like this when I try to link with my static
libraries and a boost shared object:

  $  /opt/gcc-4.7.0/bin/g++ -o /home/pauls/repos/build/bin/my_app
-static-libstdc++ -m64 -march=corei7 -Wl,-rpath,/opt/boost-1.44/lib
-L/opt/boost-1.44/lib -flto -O3 -fuse-linker-plugin
-L/home/pauls/repos/build/lib /home/pauls/repos/build/obj/my_app.o -lmy_lib
-lboost_regex -ldl -lrt
 
`_ZThn16_N5boost16exception_detail10clone_implINS0_19error_info_injectorISt11logic_errorEEED1Ev'
referenced in section `.data.rel.ro' of /tmp/cc0n7i0I.ltrans12.ltrans.o:
defined in discarded section
`.gnu.linkonce.t._ZN5boost16exception_detail10clone_implINS0_19error_info_injectorISt11logic_errorEEED5Ev'
of /home/pauls/repos/build/obj/my_app.o (symbol from plugin)
 
/opt/gcc-4.7.0/lib/gcc/x86_64-fedora-linux/4.7.0/../../../../x86_64-fedora-linux/bin/ld:
/home/pauls/repos/build/bin/my_app: hidden symbol `.LTHUNK11.6239' isn't
defined

binutils 2.22 was pre-installed into:
  $ /opt/gcc-4.7.0

gcc4.7.0 was build with:
  $ export PATH=/opt/gcc-4.7.0/bin/:$PATH
  $ ../configure --prefix=/opt/gcc-4.7.0 --build=x86_64-fedora-linux
--with-system-zlib --enable-plugin --enable-bootstrap --enable-shared
--enable-threads=posix --enable-languages=c,c++ --disable-checking
--disable-libunwind-exceptions

I tried upgrading to binutils-2.22.52 to see if was an issue with ld that had
recently been patched, but I got the same error.  

I also tried upgrading from boost-1.44 to boost-1.49 to see if it was an issue
with boost exception code, but I also got the same error.

Thanks,

Paul

[Bug lto/53604] ld reports errors using lto after upgrading from gcc-4.6.2 to gcc-4.7.0

[hjl.tools at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53604

--- Comment #1 from H.J. Lu <hjl.tools at gmail dot com> 2012-06-07 15:50:56 UTC ---
I think it is a dup for PR 53572.  Can you try patch at

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53572#c3

[Bug lto/53604] ld reports errors using lto after upgrading from gcc-4.6.2 to gcc-4.7.0

[paul.scruby at ghco dot co.uk]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53604

--- Comment #2 from Paul Scruby <paul.scruby at ghco dot co.uk> 2012-06-07 17:14:13 UTC ---
The 4.7.0 code is different so the patch didn't work:
  'struct varpool_node' has no member named 'symbol'

I'm not familiar with the gcc internals so I'm not sure what I need to add to
4.7.0 to fix this?

If this is the function in question, what do I need to add to fix this problem?

static inline bool
varpool_can_remove_if_no_refs (struct varpool_node *node)
{
  return (!node->force_output && !node->used_from_other_partition
      && (flag_toplevel_reorder || DECL_COMDAT (node->decl)
          || DECL_ARTIFICIAL (node->decl))
        && (DECL_COMDAT (node->decl) || !node->externally_visible));
}

Thanks,

Paul

[Bug lto/53604] ld reports errors using lto after upgrading from gcc-4.6.2 to gcc-4.7.0

[tetra2005 at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53604

Yuri Gribov <tetra2005 at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |tetra2005 at gmail dot com

--- Comment #3 from Yuri Gribov <tetra2005 at gmail dot com> 2012-06-11 17:04:42 UTC ---
Paul,

I changed to 

  static inline bool
  varpool_can_remove_if_no_refs (struct varpool_node *node)
  {
    return (!node->force_output && !node->used_from_other_partition
            && (flag_toplevel_reorder || DECL_COMDAT (node->decl)
                || DECL_ARTIFICIAL (node->decl))
            && ((DECL_COMDAT (node->decl) && !varpool_used_from_object_file_p
(node)) || !node->externally_visible));
  }

and it helped.

[Bug lto/53604] ld reports errors using lto after upgrading from gcc-4.6.2 to gcc-4.7.0

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53604

Richard Guenther <rguenth at gcc dot gnu.org> changed:

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

--- Comment #4 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-06-12 10:42:15 UTC ---
dup

*** This bug has been marked as a duplicate of bug 53572 ***

[Bug lto/53604] ld reports errors using lto after upgrading from gcc-4.6.2 to gcc-4.7.0

[paul.scruby at ghco dot co.uk]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53604

Paul Scruby <paul.scruby at ghco dot co.uk> changed:

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

--- Comment #5 from Paul Scruby <paul.scruby at ghco dot co.uk> 2012-06-13 11:56:53 UTC ---
Thanks Yuri.  I swapped out this function with yours but ld was still unable to
link.  I'm trying linking to my own libraries statically using the lto and the
boost shared libraries.  Linking to one of the other seems to be fine.

I'm not sure if this is the same bug as bug 53572 (gcc4.7.1) or a different
one.  If you can point me in the direction of a gcc4.7.1 release, I'm happy to
test my build to see if this problem has now been resolved.

Thanks again,

Paul

[Bug lto/53604] ld reports errors using lto after upgrading from gcc-4.6.2 to gcc-4.7.0

[tetra2005 at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53604

--- Comment #6 from Yuri Gribov <tetra2005 at gmail dot com> 2012-06-13 12:30:27 UTC ---
Paul,

I used April version of Linaro gcc toolchain so you may want to give it a try.
If it still doesn't work you'll probably need to attach a repro code here.

[Bug lto/53604] ld reports errors using lto after upgrading from gcc-4.6.2 to gcc-4.7.0

[redi at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53604

--- Comment #7 from Jonathan Wakely <redi at gcc dot gnu.org> 2012-06-13 14:28:54 UTC ---
4.7.1 is due out Real Soon Now, the release candidate can be found on any GNU
mirror:
ftp://ftp.mirrorservice.org/sites/sourceware.org/pub/gcc/snapshots/4.7.1-RC-20120606/

[Bug lto/53604] ld reports errors using lto after upgrading from gcc-4.6.2 to gcc-4.7.0

[paul.scruby at ghco dot co.uk]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53604

--- Comment #8 from Paul Scruby <paul.scruby at ghco dot co.uk> 2012-06-13 14:34:38 UTC ---
Cheers Jon, I'll try this afternoon...

[Bug lto/53604] ld reports errors using lto after upgrading from gcc-4.6.2 to gcc-4.7.0

[paul.scruby at ghco dot co.uk]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53604

--- Comment #9 from Paul Scruby <paul.scruby at ghco dot co.uk> 2012-06-13 16:44:12 UTC ---
Unfortunately I still have this linking issue with gcc-4.7.1-RC-20120606 and
binutils-2.22 so I don't think it's a dup for PR 53572.

I'll try to make a standalone project tomorrow which demonstrates this problem
and attachment it to this report to see if someone else can reproduce it.  It
might not be a bug, I might just need to change the arguments that I'm passing
to link my app.

Thanks again,

Paul

[Bug lto/53604] ld reports errors using lto after upgrading from gcc-4.6.2 to gcc-4.7.0

[paul.scruby at ghco dot co.uk]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53604

--- Comment #10 from Paul Scruby <paul.scruby at ghco dot co.uk> 2012-06-14 08:16:28 UTC ---
I released that I needed to add Yuri's patch to gcc-4.7.1.RC-20120606 as it had
not made it into the trunk.  However, this still made no difference and I still
had the same linking problem.  I'll try to hack together a standalone example
today that I can attach to the report to see if someone can reproduce it.

[Bug lto/53604] ld reports errors using lto after upgrading from gcc-4.6.2 to gcc-4.7.0

[paul.scruby at ghco dot co.uk]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53604

--- Comment #11 from Paul Scruby <paul.scruby at ghco dot co.uk> 2012-06-14 11:46:32 UTC ---
I've written half a dozen test harnesses this morning to tried to recreate this
bug unsuccessfully.  My team writes very low-latency code so they do quite a
lot of 'abusive' (but legal) partial specialization tricks force some
optimizations and branch predications on certain paths.  I suspect that some of
this nastiness is creating bloat which is not playing ball in LTO land.

Unfortunately, these lto bugs are really hard to recreate and I have to finish
some work on other projects, but I will try to pick this up and get implement
this test harness when I get a chance.  For the time being we'll stick with
gcc-4.6.2 which is working really nicely for us...

Thanks for all your help,

Paul

[Bug lto/53604] ld reports errors using lto after upgrading from gcc-4.6.2 to gcc-4.7.0

[paul.scruby at ghco dot co.uk]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=53604

--- Comment #12 from Paul Scruby <paul.scruby at ghco dot co.uk> 2012-09-06 13:03:17 UTC ---
I just tried Matt Hargett's patch from Bug53572 with gcc4.7.1 and it still has
not fixed the problem we're having.  I'll continue test these issues in our
codebase with new versions of gcc as they're released.  If I get some more time
which I don't have at the moment, I'll try to replicate it with a standalone
example.

Thanks,

Paul