[Bug optimization/14888] New: [3.5 regression] ICE with -O2 -ffast-math in final_scan_insn()

[martin at mpa-garching dot mpg dot de]

The following testcase causes an ICE in gcc 3.5 when compiled with -O2
-ffast-math, but is compiled properly by earlier versions of gcc and/or without
optimization.

--cut here

class xcomplex
  {
  public:
    float re, im;

    xcomplex &operator*= (const float &fact)
      { re*=fact; im*=fact; return *this; }
  };

void foo (xcomplex &almT, xcomplex &almG)
  {
  double gb;
  almT*=gb;
  almG*=gb*42;
  }

--cut here

~/tmp>g++ -v -O2 -ffast-math -c ./bugrep.cc
Reading specs from /afs/mpa/data/martin/ugcc/lib/gcc/i686-pc-linux-gnu/3.5.0/specs
Configured with: /scratch/gcc/configure --quiet
--prefix=/afs/mpa/data/martin/ugcc --enable-languages=c++
Thread model: posix
gcc version 3.5.0 20040408 (experimental)
 /afs/mpa/data/martin/ugcc/libexec/gcc/i686-pc-linux-gnu/3.5.0/cc1plus -quiet -v
-D_GNU_SOURCE ./bugrep.cc -quiet -dumpbase bugrep.cc -mtune=pentiumpro -auxbase
bugrep -O2 -version -ffast-math -o /tmp/cckVCfL2.s
ignoring nonexistent directory
"/afs/mpa/data/martin/ugcc/lib/gcc/i686-pc-linux-gnu/3.5.0/../../../../i686-pc-linux-gnu/include"
#include "..." search starts here:
#include <...> search starts here:
 /afs/mpa/data/martin/ugcc/lib/gcc/i686-pc-linux-gnu/3.5.0/../../../../include/c++/3.5.0
 /afs/mpa/data/martin/ugcc/lib/gcc/i686-pc-linux-gnu/3.5.0/../../../../include/c++/3.5.0/i686-pc-linux-gnu
 /afs/mpa/data/martin/ugcc/lib/gcc/i686-pc-linux-gnu/3.5.0/../../../../include/c++/3.5.0/backward
 /usr/local/include
 /afs/mpa/data/martin/ugcc/include
 /afs/mpa/data/martin/ugcc/lib/gcc/i686-pc-linux-gnu/3.5.0/include
 /usr/include
End of search list.
GNU C++ version 3.5.0 20040408 (experimental) (i686-pc-linux-gnu)
        compiled by GNU C version 3.5.0 20040408 (experimental).
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096
./bugrep.cc: In function `void foo(xcomplex&, xcomplex&)':
./bugrep.cc:15: error: could not split insn
(insn:TI 18 95 3 (set (reg:SF 8 st)
        (float_truncate:SF (reg:DF 8 st))) 91 {truncdfsf2_noop} (nil)
    (nil))
./bugrep.cc:15: internal compiler error: in final_scan_insn, at final.c:2433
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.

-- 
           Summary: [3.5 regression] ICE with -O2 -ffast-math in
                    final_scan_insn()
           Product: gcc
           Version: 3.5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: optimization
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: martin at mpa-garching dot mpg dot de
                CC: gcc-bugs at gcc dot gnu dot org,martin at mpa-garching
                    dot mpg dot de
 GCC build triplet: i686-pc-linux-gnu
  GCC host triplet: i686-pc-linux-gnu
GCC target triplet: i686-pc-linux-gnu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14888