[Bug optimization/15381] [3.5 Regression] Bootstrap comparison failure!

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

------- Additional Comments From danglin at gcc dot gnu dot org  2004-05-13 01:59 -------
This appears to be a bug in reorg, possibly related to the recent fix for
PR optimization/15296.  The miscompilation of the test program occurs
because cpp_classify_number in cppexp.c has been miscompiled:

        .stabn 68,0,98,L$M31-cpp_classify_number
L$M31:
        addl %r24,%r25,%r19
        comib,<< 1,%r19,L$0142
        ldil LR'L$C0005,%r24
        comib,<< 1,%r23,L$0142
        ldi 0,%r19
        comiclr,= 0,%r23,%r0
        ldil L'8192,%r19
        copy %r19,%r6
        comib,<> 0,%r24,L$0051
        depi -1,27,1,%r6

The "ldil" in the delay slot of the first "comib" incorrectly clobbers
%r24 in the fallthru path.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=15381