[Bug middle-end/51439] New: [4.7 Regression] ICE(SIGFPE) in good_cloning_opportunity_p

[Bug middle-end/51439] New: [4.7 Regression] ICE(SIGFPE) in good_cloning_opportunity_p

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51439

             Bug #: 51439
           Summary: [4.7 Regression] ICE(SIGFPE) in
                    good_cloning_opportunity_p
    Classification: Unclassified
           Product: gcc
           Version: 4.7.0
            Status: UNCONFIRMED
          Keywords: ice-on-valid-code
          Severity: normal
          Priority: P3
         Component: middle-end
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: burnus@gcc.gnu.org
              Host: x86_64-unknown-linux-gnu


Created attachment 26007
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=26007
Reduced test case (t108.ii, 6.8 KiB)

The attached test case compiles with GCC 4.6 g++ -O3 and with today's GCC 4.7
up to -O2; with -O3 one gets an ICE:

$ g++-4.6 -c -O3 t108.ii
$ g++ -c t108.ii
$ g++ -c -O3 t108.ii
t108.ii:118:38: internal compiler error: Floating point exception


Program received signal SIGFPE, Arithmetic exception.
0x0000000000ef75f1 in good_cloning_opportunity_p (size_cost=0, count_sum=0,
freq_sum=1000, time_benefit=70, node=<optimized out>)
    at gcc/ipa-cp.c:1233
1233          int evaluation = (time_benefit * freq_sum) / size_cost;

(gdb) bt
#0  0x0000000000ef75f1 in good_cloning_opportunity_p (size_cost=0, count_sum=0,
freq_sum=1000, time_benefit=70, node=<optimized out>)
    at gcc/ipa-cp.c:1233
#1  good_cloning_opportunity_p (time_benefit=<optimized out>, freq_sum=1000,
count_sum=<optimized out>, size_cost=0, node=Unhandled dwarf expression opcode
0xfa
)
    at gcc/ipa-cp.c:1205
#2  0x0000000000efce81 in decide_whether_version_node (node=0x2aaaacde8c60) at
gcc/ipa-cp.c:2258
#3  ipcp_decision_stage (topo=<optimized out>) atgcc/ipa-cp.c:2401
#4  ipcp_driver () at gcc/ipa-cp.c:2441
#5  0x00000000009aedf5 in execute_one_pass (pass=0x15bf200) at
gcc/passes.c:2074
#6  0x00000000009af5ba in execute_ipa_pass_list (pass=0x15bf200) at
gcc/passes.c:2441
#7  0x00000000007a8700 in ipa_passes () at gcc/cgraphunit.c:2075

[Bug middle-end/51439] [4.7 Regression] ICE(SIGFPE) in good_cloning_opportunity_p

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51439

Dominique d'Humieres <dominiq at lps dot ens.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2011-12-06
     Ever Confirmed|0                           |1

--- Comment #1 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2011-12-06 17:19:37 UTC ---
r180989 is OK.

[Bug middle-end/51439] [4.7 Regression] ICE(SIGFPE) in good_cloning_opportunity_p

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51439

Tobias Burnus <burnus at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |4.7.0

[Bug middle-end/51439] [4.7 Regression] ICE(SIGFPE) in good_cloning_opportunity_p

[steven at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51439

Steven Bosscher <steven at gcc dot gnu.org> changed:

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

--- Comment #2 from Steven Bosscher <steven at gcc dot gnu.org> 2011-12-07 00:56:44 UTC ---
Did you build with checking enabled? There is an assert to make sure size_cost
is greater than zero:

  gcc_checking_assert (size_cost >= 0);


CC += jamborm

[Bug middle-end/51439] [4.7 Regression] ICE(SIGFPE) in good_cloning_opportunity_p

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51439

--- Comment #3 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-12-07 06:49:57 UTC ---
(In reply to comment #2)
> Did you build with checking enabled?

Yes, I am bootstrap with --enable-gold --with-plugin-ld=/usr/bin/gold
--enable-languages=c,c++,fortran,lto - which matches to my knowledge
checking=yes.

[Bug middle-end/51439] [4.7 Regression] ICE(SIGFPE) in good_cloning_opportunity_p

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51439

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

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

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-12-07 08:27:26 UTC ---
IMNSHO dup of PR51362.

*** This bug has been marked as a duplicate of bug 51362 ***

[Bug middle-end/51439] [4.7 Regression] ICE(SIGFPE) in good_cloning_opportunity_p

[jamborm at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51439

--- Comment #5 from Martin Jambor <jamborm at gcc dot gnu.org> 2011-12-13 15:58:17 UTC ---
Indeed it is and I have just verified that it is also fixed.