[Bug target/14915] New: Testsuite failure on mips-sgi-irix6.5 in gcc.dg/builtins-32.c

["ghazi at gcc dot gnu dot org" <gcc-bugzilla at gcc dot gnu dot org>]

I'm getting a testsuite failure on mips-sgi-irix6.5 in gcc.dg/builtins-32.c.  Th
e test passes the "float" sized checks, but fails for the "double" size.  Specif
ically, I get an abort for test(-2.0) which incorrectly returns zero.  We're 
exercising the extract_bit_field() path in expand_builtin_signbit() in 
the "double" case but not in the "float" case.

The failure runs from here:
http://gcc.gnu.org/ml/gcc-testresults/2004-02/msg00173.html
through to a current report:
http://gcc.gnu.org/ml/gcc-testresults/2004-04/msg00359.html

Mips sets BITS_BIG_ENDIAN to zero, so it doesn't reverse the bitpos.  But if I 
force it to do so, the testcase passes.  Like so:

--- ../../orig/gcc-CVS20040409/gcc/builtins.c  Fri Apr  9 17:03:58 2004
+++ builtins.c  Sat Apr 10 11:26:17 2004
@@ -5036,7 +5036,7 @@ expand_builtin_signbit (tree exp, rtx ta

   if (GET_MODE_BITSIZE (imode) > GET_MODE_BITSIZE (rmode))
     {
-      if (BITS_BIG_ENDIAN)
+      if (BITS_BIG_ENDIAN || 1)
        bitpos = GET_MODE_BITSIZE (imode) - 1 - bitpos;
       temp = copy_to_mode_reg (imode, temp);
       temp = extract_bit_field (temp, 1, bitpos, 1,

Perhaps the logic to enter this path isn't completely correct?
E.g. do we need to test some other macro in addition to BITS_BIG_ENDIAN?

-- 
           Summary: Testsuite failure on mips-sgi-irix6.5 in
                    gcc.dg/builtins-32.c
           Product: gcc
           Version: 3.5.0
            Status: UNCONFIRMED
          Keywords: wrong-code
          Severity: normal
          Priority: P2
         Component: target
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: ghazi at gcc dot gnu dot org
                CC: gcc-bugs at gcc dot gnu dot org,roger at eyesopen dot
                    com
 GCC build triplet: mips-sgi-irix6.5
  GCC host triplet: mips-sgi-irix6.5
GCC target triplet: mips-sgi-irix6.5


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14915