Re: egcs debugging (on alpha)

[Jason Duerstock <jason at sdi dot cluephone dot com>]

On Tue, 9 Feb 1999, Jeffrey A Law wrote:

> 
> 
>   In message <Pine.LNX.4.04.9902092327230.11735-100000@sdi.aatech.com>you write:
>   > > gdb faulted? or did cc1 fault?
>   > 
>   > cc1 faulted.  I figured you didn't need to see the same info as before.
> Err, the finish should have quit before we got into reload_cse_regs and hit
> the problem you're running into.  Could be a debugger problem though.  I don't
> regularly work on alpha, so I don't know how solid "finish" and its related
> commands are.

I have no idea either.  Any idea would?

> Hmmm, is cc1 itself optimized?  The bogus looking "chain" argument might be
> a symptom of trying to debug optimized code.

I tried rebuilding egcs with CFLAGS='-g -O0' and then used gdb to poking
around again and this is what I got:

[root@clue stdlib]# gdb
/usr/local/lib/gcc-lib/alpha-unknown-linux-gnu/egcs-2.93.06/cc1
GNU gdb 4.17.0.4 with Linux/x86 hardware watchpoint and FPU support
Copyright 1998 Free Software Foundation, Inc.
GDB is free software, covered by the GNU General Public License, and you
are
welcome to change it and/or distribute copies of it under certain
conditions.
Type "show copying" to see the conditions.
There is absolutely no warranty for GDB.  Type "show warranty" for
details.
This GDB was configured as "alpha-redhat-linux"...
(gdb) set args strtof.i -o strtof.s -O6 
(gdb) b reload
Breakpoint 1 at 0x1201c76a8: file reload1.c, line 589.
(gdb) b emit_reload_insns if chain->insn->fld[0].rtint == 11940
Breakpoint 2 at 0x1201da0ec: file reload1.c, line 6451.
(gdb) run
Starting program:
/usr/local/lib/gcc-lib/alpha-unknown-linux-gnu/egcs-2.93.06/cc1 strtof.i
-o strtof.s -O6
 __set_errno __copysign fabs atan strtod strtol strtoul strtof strtold
strtoq strtouq strtoll strtoull atof atoi atol atoll strnlen __mpn_add_1
__mpn_add __mpn_sub_1 __mpn_sub round_and_return str_to_mpn __mpn_lshift_1
correctly_grouped_prefix __strtof_internal
Breakpoint 1, reload (first=0x2, global=1, dumpfile=0x120278254)
    at reload1.c:589
589     {
(gdb) bt
#0  reload (first=0x2, global=1, dumpfile=0x120278254) at reload1.c:589
#1  0x120008c50 in rest_of_compilation (decl=
Cannot access memory at address 0x20.
) at toplev.c:4043
Cannot access memory at address 0x18.
(gdb) finish
Run till exit from #0  reload (first=0x2, global=1, dumpfile=0x120278254)
    at reload1.c:589

strtod.c: In function `__strtof_internal':
In file included from strtof.c:10:
strtod.c:1185: internal error--insn does not satisfy its constraints:
(insn 10799 10796 1574 (set (reg:SF 42 $f10)
        (reg:SF 597)) 279 {imb+1} (nil)
    (nil))
toplev.c:1447: Internal compiler error in function fatal_insn
Please submit a full bug report to `egcs-bugs@egcs.cygnus.com'.
See <URL:http://egcs.cygnus.com/faq.html#bugreport> for details.

Program exited with code 041.
(gdb)