Re: bootstrap fails with current mainline

[Andreas Jaeger <aj at suse dot de>]

"David S. Miller" <davem@redhat.com> writes:

>    From: Andreas Jaeger <aj@suse.de>
>    Date: 11 Jun 2002 15:21:33 +0200
> 
>    /cvs/gcc/gcc/hashtable.c:326: internal compiler error: RTL check: expected elt 5 type 'e' or 'u', have '0' (rtx code_label) in try_split, at emit-rtl.c:3121
>    Please submit a full bug report,
>    with preprocessed source if appropriate.
>  ...   
>    I can sent more information if needed, just ask what you need...
>  ...  
>    Dave, since between my last two updates the only major change is your
>    SEQUENCE rtl change, I really suspect your patch.  Can you have a
>    look, please?
> 
> Can you dump the RTX of "insn_last" from the debugging when this
> triggers?  Thanks.  We should only see INSN's there.... Hmmm...


GNU C version 3.2 20020611 (experimental) (i686-pc-linux-gnu)
        compiled by GNU C version 3.1.1 20020521 (prerelease).
/cvs/gcc/gcc/hashtable.c: In function `approx_sqrt':
/cvs/gcc/gcc/hashtable.c:326: internal compiler error: RTL check: expected elt 5 type 'e' or 'u', have '0' (rtx code_label) in try_split, at emit-rtl.c:3121
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://www.gnu.org/software/gcc/bugs.html> for instructions.

Breakpoint 2, 0x4005f1dd in exit () from /lib/libc.so.6
(gdb) bt
#0  0x4005f1dd in exit () from /lib/libc.so.6
#1  0x081538ae in warning_with_file_and_line (
    file=0x8b4a300 "RTL check: expected elt %d type '%c' or '%c', have '%c' (rtx %s) in %s, at %s:%d", line=5, 
    msgid=0x65 <Address 0x65 out of bounds>) at /cvs/gcc/gcc/diagnostic.c:1144
#2  0x087f4c64 in rtl_check_failed_code1 (r=0x4046cd00, code=MATCH_OPERAND, file=0x65 <Address 0x65 out of bounds>, line=117, 
    func=0x8af8b40 "/cvs/gcc/gcc/emit-rtl.c") at /cvs/gcc/gcc/rtl.c:594
#3  0x08184cff in try_split (pat=0x40455ea0, trial=0x4045ccc0, last=1) at /cvs/gcc/gcc/emit-rtl.c:3121
#4  0x08752e6b in split_insn (insn=0x4045ccc0) at /cvs/gcc/gcc/recog.c:2690
#5  0x08753254 in split_all_insns (upd_life=0) at /cvs/gcc/gcc/recog.c:2746
#6  0x08859164 in rest_of_compilation (decl=0x403a4000) at /cvs/gcc/gcc/toplev.c:3278
#7  0x08071c6d in c_expand_body (fndecl=0x403a4000, nested_p=0, can_defer_p=1) at /cvs/gcc/gcc/c-decl.c:6869
#8  0x080717a1 in finish_function (nested=0, can_defer_p=1) at /cvs/gcc/gcc/c-decl.c:6736
#9  0x0804a654 in yyparse () at c-parse.y:394
#10 0x08056400 in c_common_parse_file (set_yydebug=0) at /cvs/gcc/gcc/c-lex.c:165
#11 0x088565a9 in compile_file () at /cvs/gcc/gcc/toplev.c:2094
#12 0x0885c7b5 in do_compile () at /cvs/gcc/gcc/toplev.c:5192
#13 0x0885c81c in toplev_main (argc=19, argv=0xbffff344) at /cvs/gcc/gcc/toplev.c:5224
#14 0x080c2d9a in main (argc=19, argv=0xbffff344) at /cvs/gcc/gcc/main.c:35
#15 0x4004b9ed in __libc_start_main () from /lib/libc.so.6
#3  0x08184cff in try_split (pat=0x40455ea0, trial=0x4045ccc0, last=1) at /cvs/gcc/gcc/emit-rtl.c:3121

(gdb) up
3121                  if (rtx_equal_p (PATTERN (insn_last), pat))
(gdb) p insn_last
$1 = (struct rtx_def *) 0x4046cd00
(gdb) pr
(code_label 106 105 0 95 "" "" [0 uses])
(gdb) p *insn_last
$2 = {code = CODE_LABEL, mode = VOIDmode, jump = 0, call = 0, unchanging = 0, volatil = 0, in_struct = 0, used = 0, 
  integrated = 0, frame_related = 0, fld = {{rtwint = 106, rtint = 106, rtuint = 106, rtstr = 0x6a <Address 0x6a out of bounds>, 
      rtx = 0x6a, rtvec = 0x6a, rttype = 106, rt_addr_diff_vec_flags = {min_align = 106, base_after_vec = 0, min_after_vec = 0, 
        max_after_vec = 0, min_after_base = 0, max_after_base = 0, offset_unsigned = 0, scale = 0}, rt_cselib = 0x6a, 
      rtbit = 0x6a, rttree = 0x6a, bb = 0x6a, rtmem = 0x6a}}}

Andreas
-- 
 Andreas Jaeger
  SuSE Labs aj@suse.de
   private aj@arthur.inka.de
    http://www.suse.de/~aj