Re: [mpls-linux-general] LDP :(

[James R. L. <jl...@mi...>]

If you want to debug further, I would add a fprintf to the functions in
impl_mm.c.  It will make ldpd very chatty, but it will help debug a
"double free" which is kind of what it is looking like.

On Thu, Apr 28, 2005 at 04:03:33PM +0200, Bertrand Lemaitre wrote:
> I have the following :
>=20
> (gdb) where
> #0  0xb7eae03e in malloc_consolidate (av=3D0xb7f5e480) at malloc.c:4391
> #1  0xb7eaf02d in _int_malloc (av=3D0xb7f5e480, bytes=3D24) at malloc.c:3=
849
> #2  0xb7eb05a1 in *__GI___libc_malloc (bytes=3D4096) at malloc.c:3308
> #3  0xb7fce60b in zmalloc (type=3D115, size=3D4096) at memory.c:56
> #4  0x0804a915 in mpls_malloc (size=3D4096) at impl_mm.c:12
> #5  0x080525c1 in ldp_buf_create (size=3D4096) at ldp_buf.c:31
> #6  0x0805c99c in ldp_if_create (g=3D0x809a050) at ldp_if.c:60
> #7  0x08054cde in ldp_cfg_if_set (handle=3D0x809a050, i=3D0x80987c8, flag=
=3D3)
>     at ldp_cfg.c:860
> #8  0x0804c77a in ldp_interface_create (li=3D0x8098738) at ldp_interface.=
c:51
> #9  0x0804c412 in ldp_new () at ldp.c:116
> #10 0x0804c962 in mpls_ldp (self=3D0x80819e0, vty=3D0x8099818, argc=3D0,
>     argv=3D0xbffff868) at ldp_vty.c:50
> #11 0xb7fd3ae4 in cmd_execute_command_real (vline=3D0x8099fe0, vty=3D0x80=
99818,
>     cmd=3D0x0) at command.c:2053
> #12 0xb7fd3bd0 in cmd_execute_command (vline=3D0x8099fe0, vty=3D0x8099818=
, cmd=3D0x0)
>     at command.c:2087
> #13 0xb7fd59e8 in vty_command (vty=3D0x8099818, buf=3D0x80998f8 "mpls ldp=
 ")
>     at vty.c:357
> #14 0xb7fd7247 in vty_read (thread=3D0x18) at vty.c:1209
> #15 0xb7fd0b8b in thread_call (thread=3D0xbffffc4c) at thread.c:877
> #16 0x0804c0e6 in main (argc=3D7, argv=3D0xbffffd54) at ldp_main.c:236
>=20
>=20
> I'm searching too...
>=20
>=20
>=20
> Selon "James R. Leu" <jl...@mi...>:
>=20
> > You could try running ldpd in gdb to figure out where it is dieing.
> >
> > gdb /usr/sbin/ldpd
> > run -A 127.0.0.1 -f /etc/quagga/ldpd.conf
> > <use telnet to issue ldpd commands>
> > <after ldpd dies>
> > where
> >
> > Send the output to the list and we can start debugging it.
> >
> > On Thu, Apr 28, 2005 at 03:18:25PM +0200, Bertrand Lemaitre wrote:
> > > Hi everybody,
> > >
> > > First of all, I can tell you that mpls between Quagga / Cisco is good.
> > > But, Cisco -> Quagga takes more problems.
> > >
> > > I think it's linked to LDP.
> > >
> > > So, when I try to type 'mpls ldp' in the ldpd administration console =
(via
> > > telnet), my console is over and I must kill the process ldpd to go ba=
ck
> > again.
> > >
> > > In the log, I have the follow message :
> > >
> > > Apr 28 14:53:45 NX7H7 kernel: MPLS mpls_skb_dump: from eth1 with len =
88
> > > (296)headroom=3D12 tailroom=3D28
> > >
> > >
> > > Any idea ?
> > >
> > > With TcpDump, I don't see any ldp packets...
> > >
> > > Thanks for your help
> > >
> > > Bertrand Lemaitre
> > >
> > > ----------------------------------------------------------------
> > > This message was sent using IMP, the Internet Messaging Program.
> > >
> > >
> > >
> > > -------------------------------------------------------
> > > SF.Net email is sponsored by: Tell us your software development plans!
> > > Take this survey and enter to win a one-year sub to SourceForge.net
> > > Plus IDC's 2005 look-ahead and a copy of this survey
> > > Click here to start!  http://www.idcswdc.com/cgi-bin/survey?id=3D105h=
ix
> > > _______________________________________________
> > > mpls-linux-general mailing list
> > > mpl...@li...
> > > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> >
> > --
> > James R. Leu
> > jl...@mi...
> >
>=20
>=20
> Bertrand Lemaitre
>=20
> ----------------------------------------------------------------
> This message was sent using IMP, the Internet Messaging Program.

--=20
James R. Leu
jl...@mi...