Skip Menu |
 

Date: Sat, 10 Jul 2004 19:32:35 -0700 (PDT)
From: Carl Gold <carlg@caltech.edu>
To: krb5-bugs@mit.edu
Subject: Hang in malloc_consolidate()
I don't know who this email goes to, but is there any resolution on this
malloc_conslidate() hang issue? I seem to be having a simliar problem
and I googled to find the web page refering to the problem

http://mailman.mit.edu/pipermail/krb5-bugs/2004-February/002279.html

This is what my stack trace looks like when I interrupt in the hang:

#0 0x42074c47 in malloc_consolidate () from /lib/tls/libc.so.6
#1 0x42074ae8 in _int_free () from /lib/tls/libc.so.6
#2 0x42073786 in free () from /lib/tls/libc.so.6
#3 0x400be983 in operator delete(void*) () from /usr/lib/libstdc++.so.5
#4 0x400ac4dc in std::__default_alloc_template<true,
Show quoted text
0>::deallocate(void*, unsigned) () from /usr/lib/libstdc++.so.5
#5 0x0805222e in std::_Deque_base<double, std::allocator<double>
Show quoted text
>::_M_destroy_nodes(double**, double**) ()
#6 0x08050b3c in std::_Deque_base<double, std::allocator<double>
Show quoted text
>::~_Deque_base() ()
#7 0x08067089 in std::deque<double, std::allocator<double> >::~deque() ()
...

-Carl
Download (untitled) / with headers
text/plain 1.1KiB
[carlg@caltech.edu - Sun Jul 11 13:05:20 2004]:

Show quoted text
> I don't know who this email goes to, but is there any resolution on this
> malloc_conslidate() hang issue? I seem to be having a simliar problem
> and I googled to find the web page refering to the problem
>
> http://mailman.mit.edu/pipermail/krb5-bugs/2004-February/002279.html
>
> This is what my stack trace looks like when I interrupt in the hang:
>
> #0 0x42074c47 in malloc_consolidate () from /lib/tls/libc.so.6
> #1 0x42074ae8 in _int_free () from /lib/tls/libc.so.6
> #2 0x42073786 in free () from /lib/tls/libc.so.6
> #3 0x400be983 in operator delete(void*) () from /usr/lib/libstdc++.so.5
> #4 0x400ac4dc in std::__default_alloc_template<true,
> 0>::deallocate(void*, unsigned) () from /usr/lib/libstdc++.so.5
> #5 0x0805222e in std::_Deque_base<double, std::allocator<double>
> >::_M_destroy_nodes(double**, double**) ()
> #6 0x08050b3c in std::_Deque_base<double, std::allocator<double>
> >::~_Deque_base() ()
> #7 0x08067089 in std::deque<double, std::allocator<double> >::~deque() ()
> ...

Is this program using krb5 at all? There isn't enough of the backtrace
for me to tell.
To: rt@krbdev.mit.edu
Cc: carlg@caltech.edu
Subject: Re: [krbdev.mit.edu #2633] Hang in malloc_consolidate()
From: Sam Hartman <hartmans@mit.edu>
Date: Thu, 19 Aug 2004 20:59:26 -0400
RT-Send-Cc:


I'd recommend upgrading to Kerberos 1.3.x. There seem to be several
memory management problems in krb5 particularly related to
krb5_locate_kdc in the 1.2.x code train that cause problems for the
new malloc.
Probably irrelevant since recent fixes to memory management.