Hello! Sorry for being late with answering.
On Wed, Nov 25, 2009 at 01:30:56PM +0200, Niko Tyni wrote:
> On Wed, Nov 25, 2009 at 09:45:52AM +0100, Thomas Schwinge wrote:
> > Here is a more useful backtrace, thanks to installing the perl-debug
> > package:
>
> Thanks. Is it reproducible with #!/usr/bin/debugperl ?
Yes. Here is the backtrace, from ``break abort'':
#0 abort () at abort.c:55
act = {__sigaction_handler = {sa_handler = 0x3f, sa_sigaction = 0x3f},
sa_mask = 23067412, sa_flags = 85024}
sigs = <value optimized out>
#1 0x0112c71c in __libc_message (do_abort=2, fmt=0x1214956 "*** %s ***: %s
terminated\n") at ../sysdeps/posix/libc_fatal.c:138
ap = 0x15ffb2c "\364\177\"\001\003"
fd = 4
on_2 = <value optimized out>
list = <value optimized out>
nlist = 5
cp = <value optimized out>
#2 0x011dc7db in __fortify_fail (msg=0x121493e "stack smashing detected") at
fortify_fail.c:32
No locals.
#3 0x011dc790 in __stack_chk_fail () at stack_chk_fail.c:29
No locals.
#4 0x08073e11 in Perl_newATTRSUB (my_perl=0x82b7fe8, floor=322, o=0x8aa56c8,
proto=0x0, attrs=0x0, block=0x8aa7680) at op.c:5843
aname = 0x0
gv = 0x872d498
ps = 0x0
ps_len = 153893344
cv = 0x91a6d98
const_sv = <value optimized out>
gv_fetch_flags = 2
name = 0x82ff9dc "BEGIN"
__PRETTY_FUNCTION__ = "Perl_newATTRSUB"
#5 0x08072e90 in Perl_utilize (my_perl=0x82b7fe8, aver=1, floor=322,
version=0x0, idop=0x8aade80, arg=0x0) at op.c:3878
pack = 0x8a9d198
imop = <value optimized out>
veop = 0x0
#6 0x080c0ed4 in Perl_yyparse (my_perl=0x82b7fe8) at perly.y:659
yystate = 375
yyn = 80
yyresult = <value optimized out>
yytoken = 14
parser = 0x9296608
ps = 0x91a58bc
yyval = {ival = 1, pval = 0x1 <Address 0x1 out of bounds>, opval = 0x1,
gvval = 0x1, p_tkval = 0x1 <Address 0x1 out of bounds>,
i_tkval = 1}
#7 0x0819ba68 in S_doeval (my_perl=0x82b7fe8, gimme=128, startop=0x0,
outside=0x872d8a8, seq=4264) at pp_ctl.c:2981
sp = <value optimized out>
saveop = 0x8744538
__PRETTY_FUNCTION__ = "S_doeval"
#8 0x0819cba4 in Perl_pp_entereval (my_perl=0x82b7fe8) at pp_ctl.c:3671
sp = 0x8e6e048
cx = 0x82dee08
sv = <value optimized out>
gimme = <value optimized out>
was = 2712
tbuf = "_<(eval 1175)\000\350\177+\b\340,\002"
tmpbuf = <value optimized out>
safestr = 0x8ebaec8 "_<(eval 1175)"
len = 13
ok = <value optimized out>
runcv = 0x872d8a8
seq = 4264
saved_hh = 0x0
__PRETTY_FUNCTION__ = "Perl_pp_entereval"
#9 0x080ed3a7 in Perl_runops_debug (my_perl=0x82b7fe8) at dump.c:1968
No locals.
#10 0x0807ea20 in S_run_body (my_perl=0x82b7fe8) at perl.c:2431
__PRETTY_FUNCTION__ = "S_run_body"
#11 perl_run (my_perl=0x82b7fe8) at perl.c:2349
oldscope = 1
ret = <value optimized out>
cur_env = {je_prev = 0x82b819c, je_buf = {{__jmpbuf = {19038196,
23068240, 142560, 23068168, 23068080, 134735376},
__mask_was_saved = 0, __saved_mask = 0}}, je_ret = 0,
je_mustcatch = 0 '\000'}
__PRETTY_FUNCTION__ = "perl_run"
#12 0x08060a85 in main (argc=2, argv=0x15ffea8, env=0x15ffeb4) at perlmain.c:117
exitstatus = <value optimized out>
> Could you put together a testcase so others can reproduce it?
I will try to reduce this to a suitable testcase.
> Is this specific to the hurd port?
A quick attempt to reproduce this on GNU/Linux with mostly similar
packages installed didn't reveal this Perl bug, but I'll try some more to
make this reproducible on Linux-based systems. Of course, you (or
everyone else interested) could also get an account on this Hurd machine.
> I wasn't aware libc is
> compiled with fortifying options, is that the case elsewhere too?
Isn't it rather that Perl is being compiled with fortifying options (I
didn't check)? glibc / GCC only provide the infrastructure and print out
the error message.
Regards,
Thomas
signature.asc
Description: Digital signature
