[Bug middle-end/58094] New: [4.9 Regression] IPA devirt testsuite errors

[Bug middle-end/58094] New: [4.9 Regression] IPA devirt testsuite errors

[dje at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58094

            Bug ID: 58094
           Summary: [4.9 Regression] IPA devirt testsuite errors
           Product: gcc
           Version: 4.9.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: middle-end
          Assignee: unassigned at gcc dot gnu.org
          Reporter: dje at gcc dot gnu.org

http://gcc.gnu.org/ml/gcc-patches/2013-08/msg00092.html

This patch caused new testsuite regressions:

FAIL: g++.dg/ipa/devirt-11.C -std=gnu++98  scan-ipa-dump-times inline
"Discovered a virtual call to a known target" 3
FAIL: g++.dg/ipa/devirt-11.C -std=gnu++98  scan-ipa-dump-times inline
"and turned into root of the clone tree" 1
FAIL: g++.dg/ipa/devirt-11.C -std=gnu++11  scan-ipa-dump-times inline
"Discovered a virtual call to a known target" 3
FAIL: g++.dg/ipa/devirt-11.C -std=gnu++11  scan-ipa-dump-times inline
"and turned into root of the clone tree" 1

FAIL: g++.dg/tree-ssa/pr42337.C -std=gnu++98 (test for excess errors)
FAIL: g++.dg/tree-ssa/pr42337.C -std=gnu++11 (test for excess errors)

[Bug middle-end/58094] [4.9 Regression] IPA devirt testsuite errors

[dje at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58094

David Edelsohn <dje at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|                            |powerpc*-*-*
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2013-08-06
                 CC|                            |hubicka at gcc dot gnu.org,
                   |                            |marxin.liska at gmail dot com
               Host|                            |powerpc*-*-*
     Ever confirmed|0                           |1
              Build|                            |powerpc*-*-*

--- Comment #1 from David Edelsohn <dje at gcc dot gnu.org> ---
Confirmed.

[Bug middle-end/58094] [4.9 Regression] IPA devirt testsuite errors

[hubicka at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58094

--- Comment #2 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
I am not able to reproduce those on gcc110.fsffrance.org.  Would be possible to
have -fdump-ipa-all -fdump-tree-all dumps of the devirt testcase?  I think both
are related to fast that ipa-prop is not propagating across the local aliases,
but I do not see why the local aliases are introduced here at first place.
Perhaps binds_local_p is somehow wrong for AIX?

[Bug middle-end/58094] [4.9 Regression] IPA devirt testsuite errors

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58094

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.9.0

[Bug middle-end/58094] [4.9 Regression] IPA devirt testsuite errors

[hubicka at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58094

--- Comment #4 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
Author: hubicka
Date: Fri Sep  6 10:13:37 2013
New Revision: 202316

URL: http://gcc.gnu.org/viewcvs?rev=202316&root=gcc&view=rev
Log:
    PR middle-end/58094
    * cgraph.h (symtab_semantically_equivalent_p): Declare.
    * tree-tailcall.c: Include ipa-utils.h.
    (find_tail_calls): Use it.
    * ipa-pure-const.c (check_call): Likewise.
    * ipa-utils.c (recursive_call_p): New function.
    * ipa-utils.h (recursive_call_p): Dclare.
    * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
    (symtab_semantically_equivalent_p): New function.
    * Makefile.in (tree-tailcall.o): Update dependencies.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/Makefile.in
    trunk/gcc/cgraph.h
    trunk/gcc/ipa-pure-const.c
    trunk/gcc/ipa-utils.c
    trunk/gcc/ipa-utils.h
    trunk/gcc/symtab.c
    trunk/gcc/tree-tailcall.c

[Bug middle-end/58094] [4.9 Regression] IPA devirt testsuite errors

[hubicka at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58094

--- Comment #5 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
Author: hubicka
Date: Fri Sep  6 15:02:03 2013
New Revision: 202337

URL: http://gcc.gnu.org/viewcvs?rev=202337&root=gcc&view=rev
Log:

    PR middle-end/58094
    * ipa-inline.c (has_caller_p): New function.
    (want_inline_function_to_all_callers_p): Use it.
    (sum_callers, inline_to_all_callers): Break out from ...
    (ipa_inline): ... here.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/ipa-inline.c

[Bug middle-end/58094] [4.9 Regression] IPA devirt testsuite errors

[hubicka at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58094

--- Comment #6 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
Author: hubicka
Date: Fri Sep 13 14:35:32 2013
New Revision: 202567

URL: http://gcc.gnu.org/viewcvs?rev=202567&root=gcc&view=rev
Log:
    PR middle-end/58094
    * ipa-inline.c (check_callers): New function.
    (check_caller_edge): Remove.
    (want_inline_function_to_all_callers_p): Also permit alises that are
    called dirrectly.
    (inline_to_all_callers): Terminate the walk when devirtualization
    introduce new calls.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/ipa-inline.c

[Bug middle-end/58094] [4.9 Regression] IPA devirt testsuite errors

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58094

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

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

--- Comment #7 from Richard Biener <rguenth at gcc dot gnu.org> ---
I assume fixed.

[Bug middle-end/58094] [4.9 Regression] IPA devirt testsuite errors

[doko at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58094

Matthias Klose <doko at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|RESOLVED                    |NEW
                 CC|                            |doko at gcc dot gnu.org
         Resolution|FIXED                       |---

--- Comment #8 from Matthias Klose <doko at gcc dot gnu.org> ---
I see these test cases failing with 4.9 20140512.  Looking at the testresults
ML, they are seen by others too.

http://gcc.gnu.org/ml/gcc-testresults/2014-05/msg00892.html

[Bug middle-end/58094] [4.9 Regression] IPA devirt testsuite errors

[doko at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58094

--- Comment #9 from Matthias Klose <doko at gcc dot gnu.org> ---
x86_64-linux-gnu and i586-linux-gnu too

[Bug middle-end/58094] [4.9 Regression] IPA devirt testsuite errors

[hp at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58094

Hans-Peter Nilsson <hp at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|powerpc*-*-*                |powerpc*-*-*, cris-*-*,
                   |                            |i686-pc-linux-gnu,
                   |                            |x86_64-linux-gnu
                 CC|                            |hp at gcc dot gnu.org

--- Comment #10 from Hans-Peter Nilsson <hp at gcc dot gnu.org> ---
cris-elf too, started failing in the window (210049:210079], so most likely
r210079.  Honza is already CC:ed.

[Bug middle-end/58094] [4.9 Regression] IPA devirt testsuite errors

[hubicka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58094

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |hubicka at gcc dot gnu.org

--- Comment #11 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
Mine, that testcase is fragile, perhaps we can just disable the template (it
was originally testing a scenario that no longer happens anyway)

[Bug middle-end/58094] [4.9 Regression] IPA devirt testsuite errors

[hubicka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58094

--- Comment #12 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
Author: hubicka
Date: Wed May 21 05:52:07 2014
New Revision: 210675

URL: http://gcc.gnu.org/viewcvs?rev=210675&root=gcc&view=rev
Log:

    PR middle-end/58094
    * g++.dg/ipa/devirt-11.C: Be lax about number of devirtualizations.

Modified:
    branches/gcc-4_9-branch/gcc/testsuite/ChangeLog
    branches/gcc-4_9-branch/gcc/testsuite/g++.dg/ipa/devirt-11.C

[Bug middle-end/58094] [4.9 Regression] IPA devirt testsuite errors

[hubicka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58094

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

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

--- Comment #13 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
Fixed.