[Bug target/65414] [5 Regression] ICE while building libgcc after stage 2, bootstrap failure on aarch64-linux-gnu

pinskia at gcc dot gnu.org gcc-bugzilla@gcc.gnu.org
Sat Mar 14 20:42:00 GMT 2015


https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65414

--- Comment #7 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
There is some memory issue going on here.
If I rebuild libcpp/expr.o manually and not from full build we get:
        cmp     x14, xzr
        cset    w22, ne
        cmp     x19, xzr
        cset    w0, ne
        and     w22, w22, w0
        .loc 1 1956 0
        ldr     x8, [x29, 152]
        mov     x0, x23
        mov     x1, x19
        bl      _ZL12num_part_mulmm

But the built one gets:
   0x0000000000f712b8 <_cpp_parse_expr+1600>:   cmp     x14, xzr
   0x0000000000f712bc <_cpp_parse_expr+1604>:   cset    w22, ne
   0x0000000000f712c0 <_cpp_parse_expr+1608>:   cmp     x19, xzr
   0x0000000000f712c4 <_cpp_parse_expr+1612>:   cset    w0, ne
   0x0000000000f712c8 <_cpp_parse_expr+1616>:   and     w22, w22, w0
   0x0000000000f712cc <_cpp_parse_expr+1620>:   mov     x0, x23
   0x0000000000f712d0 <_cpp_parse_expr+1624>:   mov     x8, x16
   0x0000000000f712d4 <_cpp_parse_expr+1628>:   mov     x1, x19
   0x0000000000f712d8 <_cpp_parse_expr+1632>:   bl      0xf6eee0
<_ZL12num_part_mulmm>



More information about the Gcc-bugs mailing list