[Bug target/65581] New: testsuite lto issue: multiple definition of `main', undefined reference to `WinMain'

[Bug target/65581] New: testsuite lto issue: multiple definition of `main', undefined reference to `WinMain'

[rainer@emrich-ebersheim.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65581

            Bug ID: 65581
           Summary: testsuite lto issue: multiple definition of `main',
                    undefined reference to `WinMain'
           Product: gcc
           Version: 5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: rainer@emrich-ebersheim.de

Runnning the testsuites I see a real lot of FAILs like the following.

Executing on host:
/opt/devel/SCRATCH/tmp.4bOnj9XWJa/gcc-5.0.0/gcc-5.0.0/gcc/xgcc
-B/opt/devel/SCRATCH/tmp.4bOnj9XWJa/gcc-5.0.0/gcc-5.0.0/gcc/
/opt/devel/gnu/src/gcc-mingw-w64/gcc-5.0.0/gcc/testsuite/gcc.dg/torture/20090618-1.c
 -fno-diagnostics-show-caret -fdiagnostics-color=never    -O2 -flto
-flto-partition=none    -lm    -o ./20090618-1.exe    (timeout = 300)
spawn /opt/devel/SCRATCH/tmp.4bOnj9XWJa/gcc-5.0.0/gcc-5.0.0/gcc/xgcc
-B/opt/devel/SCRATCH/tmp.4bOnj9XWJa/gcc-5.0.0/gcc-5.0.0/gcc/
/opt/devel/gnu/src/gcc-mingw-w64/gcc-5.0.0/gcc/testsuite/gcc.dg/torture/20090618-1.c
-fno-diagnostics-show-caret -fdiagnostics-color=never -O2 -flto
-flto-partition=none -lm -o ./20090618-1.exe
D:\Users\RAINER~1.EMR\AppData\Local\Temp\ccwwwiFD.lto.o:<artificial>:(.text.startup+0x0):
multiple definition of `main'
D:/x86_64-w64-trunk/mingw/lib/../lib/libmingw32.a(lib64_libmingw32_a-crt0_c.o):D:/opt/devel/gnu/src/gcc-mingw-w64/mingw-w64-runtime.git/mingw-w64-crt/crt/crt0_c.c:17:
first defined here
D:/x86_64-w64-trunk/mingw/lib/../lib/libmingw32.a(lib64_libmingw32_a-crt0_c.o):
In function `main':
D:/opt/devel/gnu/src/gcc-mingw-w64/mingw-w64-runtime.git/mingw-w64-crt/crt/crt0_c.c:18:
undefined reference to `WinMain'
collect2.exe: error: ld returned 1 exit status
compiler exited with status 1
output is:
D:\Users\RAINER~1.EMR\AppData\Local\Temp\ccwwwiFD.lto.o:<artificial>:(.text.startup+0x0):
multiple definition of `main'
D:/x86_64-w64-trunk/mingw/lib/../lib/libmingw32.a(lib64_libmingw32_a-crt0_c.o):D:/opt/devel/gnu/src/gcc-mingw-w64/mingw-w64-runtime.git/mingw-w64-crt/crt/crt0_c.c:17:
first defined here
D:/x86_64-w64-trunk/mingw/lib/../lib/libmingw32.a(lib64_libmingw32_a-crt0_c.o):
In function `main':
D:/opt/devel/gnu/src/gcc-mingw-w64/mingw-w64-runtime.git/mingw-w64-crt/crt/crt0_c.c:18:
undefined reference to `WinMain'
collect2.exe: error: ld returned 1 exit status

I don't know if this is a w64-mingw runtime issue or a gcc issue.

[Bug target/65581] [5 Regression] testsuite lto issue: multiple definition of `main', undefined reference to `WinMain'

[rainer@emrich-ebersheim.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65581

Rainer Emrich <rainer@emrich-ebersheim.de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|testsuite lto issue:        |[5 Regression] testsuite
                   |multiple definition of      |lto issue: multiple
                   |`main', undefined reference |definition of `main',
                   |to `WinMain'                |undefined reference to
                   |                            |`WinMain'

--- Comment #1 from Rainer Emrich <rainer@emrich-ebersheim.de> ---
New issue, gcc 5 regression.

[Bug target/65581] [5 Regression] testsuite lto issue: multiple definition of `main', undefined reference to `WinMain'

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65581

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
Can you try reducing to a minimal example?  it would seem that  int main() {
return 0; } should already fail this way?

[Bug target/65581] [5 Regression] testsuite lto issue: multiple definition of `main', undefined reference to `WinMain'

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65581

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |5.0

[Bug target/65581] [5 Regression] testsuite lto issue: multiple definition of `main', undefined reference to `WinMain'

[rainer@emrich-ebersheim.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65581

--- Comment #3 from Rainer Emrich <rainer@emrich-ebersheim.de> ---
(In reply to Richard Biener from comment #2)
> Can you try reducing to a minimal example?  it would seem that  int main() {
> return 0; } should already fail this way?
test.c:
int main()
{
  return  0;
}

Installed gcc:
gcc test.c -fno-diagnostics-show-caret -fdiagnostics-color=never -O2 -flto -lm
-o test.exe
D:\msys64\tmp\ccNrn9pc.lto.o:<artificial>:(.text.startup+0x0): multiple
definition of `main'
D:/x86_64-w64-trunk/mingw/lib/../lib/libmingw32.a(lib64_libmingw32_a-crt0_c.o):D:/opt/devel/gnu/src/gcc-mingw-w64/mingw-w64-runtime.git/mingw-w64-crt/crt/crt0_c.c:17:
first defined here
D:/x86_64-w64-trunk/mingw/lib/../lib/libmingw32.a(lib64_libmingw32_a-crt0_c.o):
In function `main':
D:/opt/devel/gnu/src/gcc-mingw-w64/mingw-w64-runtime.git/mingw-w64-crt/crt/crt0_c.c:18:
undefined reference to `WinMain'
collect2.exe: error: ld returned 1 exit status

Yeah, there's a real problem!

[Bug target/65581] [5 Regression] testsuite lto issue: multiple definition of `main', undefined reference to `WinMain'

[ktietz at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65581

Kai Tietz <ktietz at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |hubicka at ucw dot cz

--- Comment #6 from Kai Tietz <ktietz at gcc dot gnu.org> ---
As far as I understand this issue does LTO now handle stuff used from object
file different to prior versions. I add Jan.  He might be able to give us some
more points

[Bug target/65581] [5 Regression] testsuite lto issue: multiple definition of `main', undefined reference to `WinMain'

[rainer@emrich-ebersheim.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65581

--- Comment #7 from Rainer Emrich <rainer@emrich-ebersheim.de> ---
(In reply to Kai Tietz from comment #6)
> As far as I understand this issue does LTO now handle stuff used from object
> file different to prior versions. I add Jan.  He might be able to give us
> some more points

Any news on this issue. IMHO it would be a shame to not have lto working on
mingw-w64 targets for gcc 5.0, even mingw-w64 isn't primary/secondary platform.

[Bug target/65581] [5 Regression] testsuite lto issue: multiple definition of `main', undefined reference to `WinMain'

[hubicka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65581

Jan Hubicka <hubicka at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |hubicka at gcc dot gnu.org

--- Comment #8 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
Hmm, is it still a problem and if so, why it is marked as resolved/invalid?

crtbegin/crtend should be compiled without LTO even with LTO bootstrap.
Can you post the callgraph and resolution file created by compiling the
testcase with --save-temps and -fdump-ipa-cgraph?

[Bug target/65581] [5 Regression] testsuite lto issue: multiple definition of `main', undefined reference to `WinMain'

[rainer@emrich-ebersheim.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65581

--- Comment #9 from Rainer Emrich <rainer@emrich-ebersheim.de> ---
Created attachment 35210
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=35210&action=edit
Test case including temporaries

[Bug target/65581] [5 Regression] testsuite lto issue: multiple definition of `main', undefined reference to `WinMain'

[rainer@emrich-ebersheim.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65581

--- Comment #10 from Rainer Emrich <rainer@emrich-ebersheim.de> ---
(In reply to Jan Hubicka from comment #8)
> Hmm, is it still a problem and if so, why it is marked as resolved/invalid?
> 
> crtbegin/crtend should be compiled without LTO even with LTO bootstrap.
> Can you post the callgraph and resolution file created by compiling the
> testcase with --save-temps and -fdump-ipa-cgraph?

The attached test case is compiled with gcc version 5.0.0 20150326
(experimental) [trunk revision 221697] (GCC).

[Bug target/65581] [5 Regression] testsuite lto issue: multiple definition of `main', undefined reference to `WinMain'

[hubicka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65581

--- Comment #11 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
Thanks!
The resolution file is wrong:
1
test.o 1
172 795d441a PREEMPTED_REG main

the resolution of main() should be PREVAILING_DEF_IRONLY. PREEMTED_REG is
defined as follows:
PREEMPTED_REG (this definition was pre-empted by a definition in a regular
object file) 

This is linker bug. It basicaly instructs GCC to drop the definition of main
because it is defined elsewhere.

However GCC is not ready for non-weak symbol to be preemted and produces the
function into final binary anyway:

        .file   "<artificial>"
        .def    __main; .scl    2;      .type   32;     .endef
        .text
        .globl  main
        .def    main;   .scl    2;      .type   32;     .endef
        .seh_proc       main
main:
        subq    $40, %rsp
        .seh_stackalloc 40
        .seh_endprologue
        call    __main
        movl    $0, %eax
        addq    $40, %rsp
        ret
        .seh_endproc
        .ident  "GCC: (GNU) 5.0.0 20150326 (experimental) [trunk revision
221697]"

Is something wrong with this assembly? How does the testecase comile w/o LTO?

[Bug target/65581] [5 Regression] testsuite lto issue: multiple definition of `main', undefined reference to `WinMain'

[rainer@emrich-ebersheim.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65581

--- Comment #12 from Rainer Emrich <rainer@emrich-ebersheim.de> ---
Created attachment 35218
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=35218&action=edit
Test case compiled without -flto.

This without -flto.

By the way ld is:
GNU ld (GNU Binutils) 2.25.51.20150326

[Bug target/65581] [5 Regression] testsuite lto issue: multiple definition of `main', undefined reference to `WinMain'

[hubicka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65581

--- Comment #13 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
Rainer,
the compiled code (test.s) is identical to what LTO path produces, so I am
convinced this is a bug at binutils side.  Would you please mind filling up the
PR?

There are two issues at least - first is that the resolution of main should be
PREVAILING_DEF and second that it should not fail with undefined symbol.

[Bug target/65581] [5 Regression] testsuite lto issue: multiple definition of `main', undefined reference to `WinMain'

[rainer@emrich-ebersheim.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65581

--- Comment #14 from Rainer Emrich <rainer@emrich-ebersheim.de> ---
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Am 03.04.2015 01:31, schrieb hubicka at gcc dot gnu.org:
> --- Comment #13 from Jan Hubicka <hubicka at gcc dot gnu.org> --- the
> compiled code (test.s) is identical to what LTO path produces, so I am 
> convinced this is a bug at binutils side.  Would you please mind filling up
> the PR?
Yes, of course. I will do this later today or on the weekend. I'm trying
binutils 2.25 at the moment, to check if this version is ok.

> 
> There are two issues at least - first is that the resolution of main should
> be PREVAILING_DEF and second that it should not fail with undefined
> symbol.
> 

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.22 (MingW32)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iQEcBAEBAgAGBQJVHlviAAoJEB3HOsWs+KJbaWIH/A4Ocs13Reo+XxB955m8jBuX
46Eb8Uy23ObrJEIYN6i6pyXuptct2MJyj0KvGS3+M8h5WGHuhd9DKCB2kaRlejAR
1gCZdg1t8rF4iyghEfggoUCVDg7qn7bv041fnFhqYwzgx7vONTWEmFWp2PKxmYh5
hvsA18KC9wGEWYT1krY/ovc8VSejc92A9CVd2aGf50O/nRw4g7Y6ciW8FZlOdGlz
yJVca4ognWqeZDDc0pD2J54FthBQtymJ0yetmbhit+5xspS4PfbH/mGvtyr3xFDZ
W6YtXPkJxRBk2OcIJXqpV1eLCYv4NYir+cFbeRbdFWJBEt7wMKIYdG+aiL2Ltto=
=5aW4
-----END PGP SIGNATURE-----

[Bug target/65581] [5 Regression] testsuite lto issue: multiple definition of `main', undefined reference to `WinMain'

[rainer@emrich-ebersheim.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65581

--- Comment #15 from Rainer Emrich <rainer@emrich-ebersheim.de> ---
Bug report opened for ld.
https://sourceware.org/bugzilla/show_bug.cgi?id=18199

[Bug target/65581] [5 Regression] testsuite lto issue: multiple definition of `main', undefined reference to `WinMain'

[rainer@emrich-ebersheim.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65581

--- Comment #16 from Rainer Emrich <rainer@emrich-ebersheim.de> ---
Indeed this seems to be a linker bug in ld on binutils head.
Doesn't reproduce with binutils 2.25.