[Bug lto/48259] New: Internal compiler errors in lto1

[Bug lto/48259] New: Internal compiler errors in lto1

[d.g.gorbachev at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48259

           Summary: Internal compiler errors in lto1
           Product: gcc
           Version: 4.6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: lto
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: d.g.gorbachev@gmail.com


I just experimented a bit with static libstdc++ compiled with -flto. GCC fails
with internal compiler errors when using it. (Linking with static libgcc
compiled with -flto also reveals a bug in the linker.)

Does it make sense to report such bugs? Or maybe these libraries are too
'special' and can't be expected to work this way?

[Bug lto/48259] Internal compiler errors in lto1

[d.g.gorbachev at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48259

--- Comment #1 from Dmitry Gorbachev <d.g.gorbachev at gmail dot com> 2011-03-23 19:45:13 UTC ---
Created attachment 23761
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23761
testcase, based on libstdc++

GCC 4.6.0 and 4.7.0 with -flto -O2:
lto1: internal compiler error: in ipcp_iterate_stage, at ipa-cp.c:891

4.6.0 with -flto:
lto1: internal compiler error: in cgraph_mark_functions_to_output, at
cgraphunit.c:11

4.7.0 with -flto:
[...]
lto1: internal compiler error: failed to reclaim unneeded function

[Bug lto/48259] Internal compiler errors in lto1

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48259

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |WAITING
   Last reconfirmed|                            |2011.03.24 12:24:22
     Ever Confirmed|0                           |1

--- Comment #2 from Richard Guenther <rguenth at gcc dot gnu.org> 2011-03-24 12:24:22 UTC ---
At least LTOing libstdc++ should work, I'm not sure about libgcc, but if
no functions of libgcc end up being used it shouldn't break.

Hm, I can't reproduce it with 4.6.

rguenther@murzim:/tmp> g++-4.6 -flto -O2 1.C 2.C 3.s
rguenther@murzim:/tmp> g++-4.6 -flto -O2 1.C 2.C 3.s -fno-use-linker-plugin
rguenther@murzim:/tmp> g++-4.6 -flto -O2 1.C 2.C 3.s -fuse-linker-plugin
rguenther@murzim:/tmp> g++-4.6 -flto -O2 1.C 2.C 3.s -flto-partition=none
rguenther@murzim:/tmp> 

nor with trunk.  Can you be more specific about what linker you use?  Maybe
show -v output of a failing compile?

[Bug lto/48259] Internal compiler errors in lto1

[d.g.gorbachev at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48259

--- Comment #3 from Dmitry Gorbachev <d.g.gorbachev at gmail dot com> 2011-03-26 08:24:27 UTC ---
Created attachment 23775
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23775
Backtrace #1

Reproduced with i686-pc-linux-gnu-gcc-4.6.1 (GCC) 4.6.1 20110325 (prerelease)

> Can you be more specific about what linker you use?

GNU ld (GNU Binutils) 2.21.51.20110322
GNU gold (GNU Binutils 2.21.51.20110322) 1.11

> g++-4.6 -flto -O2 1.C 2.C 3.s -fno-use-linker-plugin

It works.

> g++-4.6 -flto -O2 1.C 2.C 3.s -flto-partition=none

It does not work. Works with `-flto 1.C 2.C 3.s -flto-partition=1to1'.

[Bug lto/48259] Internal compiler errors in lto1

[d.g.gorbachev at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48259

--- Comment #4 from Dmitry Gorbachev <d.g.gorbachev at gmail dot com> 2011-03-26 08:25:10 UTC ---
Created attachment 23776
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23776
Backtrace #2

[Bug lto/48259] Internal compiler errors in lto1

[d.g.gorbachev at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48259

--- Comment #5 from Dmitry Gorbachev <d.g.gorbachev at gmail dot com> 2011-03-26 08:28:41 UTC ---
Created attachment 23777
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23777
Output of gcc -v -O2 -flto

> Maybe show -v output of a failing compile?

[Bug lto/48259] Internal compiler errors in lto1

[d.g.gorbachev at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48259

--- Comment #6 from Dmitry Gorbachev <d.g.gorbachev at gmail dot com> 2011-03-26 08:29:23 UTC ---
Created attachment 23778
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23778
Output of gcc -v -flto

[Bug lto/48259] Internal compiler errors in lto1

[d.g.gorbachev at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48259

--- Comment #7 from Dmitry Gorbachev <d.g.gorbachev at gmail dot com> 2011-03-27 17:33:51 UTC ---
Created attachment 23782
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23782
Backtrace from GCC 4.7

[Bug lto/48259] Internal compiler errors in lto1

[d.g.gorbachev at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48259

Dmitry Gorbachev <d.g.gorbachev at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to work|                            |4.7.0
      Known to fail|4.7.0                       |

--- Comment #8 from Dmitry Gorbachev <d.g.gorbachev at gmail dot com> 2011-08-06 23:56:45 UTC ---
I can't reproduce it in 4.7.0.

[Bug lto/48259] Internal compiler errors in lto1

[d.g.gorbachev at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48259

Dmitry Gorbachev <d.g.gorbachev at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #23782|0                           |1
        is obsolete|                            |

--- Comment #9 from Dmitry Gorbachev <d.g.gorbachev at gmail dot com> 2011-08-21 20:32:06 UTC ---
Created attachment 25067
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=25067
Another testcase (extracted from libstdc++)

[Bug lto/48259] Internal compiler errors in lto1

[d.g.gorbachev at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48259

--- Comment #10 from Dmitry Gorbachev <d.g.gorbachev at gmail dot com> 2011-12-12 21:43:15 UTC ---
Recent GCC 4.7.0 works with both testcases. GCC 4.6.3 still fails on the 1st.

[Bug lto/48259] Internal compiler errors in lto1

[d.g.gorbachev at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48259

Dmitry Gorbachev <d.g.gorbachev at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |RESOLVED
         Resolution|                            |FIXED

--- Comment #11 from Dmitry Gorbachev <d.g.gorbachev at gmail dot com> 2013-04-12 20:45:54 UTC ---
Does not fail with 4.7.3 and 4.8.1. The 4.6.4 branch has been closed.