mpls-linux-general — Mailing list for general discussion about MPLS for Linux

[mpls-linux-general] mplsadm

[farahmand z. <z_c...@ya...>]

Hi,

I compiled mplsadm succesfully,but I dont know how it works?!

in the other words, when I run it (./mplsadm) ,ther is no reflexion and for example a command like 

this: mplsadm -L eth0:0 is failed.

Hope somebody can help me .

With best regards.

yasi jaberi.



---------------------------------
Do You Yahoo!?
Yahoo! Tax Center - online filing with TurboTax

[mpls-linux-general] About the Linux-MPLS tcpdump bug

[Daniel G. C. <da...@on...>]

Hi, I was wondering about the bug in Linux-MPLS that says that, when
you capture MPLS packets with Ethereal (tcpdump), if you run ethereal
on an LER the packets seem to be corrupted.

When I try with a 3 host network, (3---2---1) and I make ping
from one side to the other side of the net, packets (in one way) seem
to be corrupted, but in the other way don't. This is, for example,
from 1 to 3 they are ok, but from 3 to 1 they aren't MPLS packets.

And I am running ethereal in the LSR that is in the core of the
network (number 2) in spite of running it on an LER (3 or 1) so I
want to be sure that my problem is this bug and not a different one.
The packets captured in 2 don't seem to be corrupted, they are MPLS
packets in one way but in the other way they are normal IP packets
(non MPLS-packets).

Also, I have test it in a network like this 4--3--2--1 (four hosts) running
ethereal on 3 and 2, and the same problem happens. From 4 to 1 the sendings
are MPLS labelled IP packets, and from 1 to 4 the packets are normal IP
packets (non MPLS).

If you can confirm that my problem is the bug and not a different one I will
be so happy.

Thanks,Dani.

[mpls-linux-general] problem in DOD

[anish v. <ani...@re...>]

hi jim,
  as u have suggested i changed the label_distribution_mode to 
DOD
in ldp_deafults.h.
after that i checked in initialization msg the 'A' bit it's
correct..
as i start ldp_linux
after adding global n interface it gives errors

ldp_label_mapping_send
ldp_linux:ldp_attr.c:515 ldp_attr_insert_upstream : Assertion '
a->in_tree == LDP_FALSE ' failed
Aborted (core dumped)

n after commenting above assert it gives following error
ldp_label_mapping_send
ldp_linux:ldp_attr.c:899 ldp_fs_add_attr : Assertion
' ptr == ((void *)0)' failed
Aborted (core dumped)

looking forward for ur reply
anish

Re: [mpls-linux-general] Address Mesg

[clz <cui...@ya...>]
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_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com

Re: [mpls-linux-general] RSVP-TE implimentaton for Linux

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

On Wed, Apr 10, 2002 at 02:58:18PM -0700, Scott Lee wrote:
> Is there any RSVP-TE implementation for linux?
> I checked out nist version of it, but it is
> for BSD not LINUX.

Check out:

http://ds_mpls.atlantis.rug.ac.be


> 
> Thanks!
> Scott
> 
> __________________________________________________
> Do You Yahoo!?
> Yahoo! Tax Center - online filing with TurboTax
> http://taxes.yahoo.com/
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu

[mpls-linux-general] RSVP-TE implimentaton for Linux

[Scott L. <nat...@ya...>]

Is there any RSVP-TE implementation for linux?
I checked out nist version of it, but it is
for BSD not LINUX.

Thanks!
Scott

__________________________________________________
Do You Yahoo!?
Yahoo! Tax Center - online filing with TurboTax
http://taxes.yahoo.com/

RE: [mpls-linux-general] Problems with MPLS packets

[Daniel G. C. <da...@on...>]

Thank you Jim, I saw the 'NOTE' file, but I didn't think that it was =
this, because it runs properly in one way but in the other don't (and is =
always in the same way).
So I thought that it was another diferent problem (perhaps my =
configuration, the ethernet cards,...) but if you say that is this, now =
I'm good. Thank you.

Dani
  ----- Original Message -----=20
  From: James R. Leu=20
  To: Daniel Gregorio Catalan=20
  Cc: Lista Mpls (envio)=20
  Sent: Wednesday, April 10, 2002 4:24 PM
  Subject: Re: [mpls-linux-general] Problems with MPLS packets


  This is a know problem.  It is not really a corrupted packet it just =
looks
  that way from ethereal/tcpudmp etc.

  From the "NOTE" file:
  -mpls packets appear to be corrupted when running tcpdump or the like =
on
   an LER (the packets are not really corrupted, tcpdump is getting an =
invalid
   version of the packets)=20

  Jim


  On Wed, Apr 10, 2002 at 02:22:00PM +0200, Daniel Gregorio Catalan =
wrote:
  > (sorry for my english)
  > I'm making a MPLS network with PCs with Linux-MPLS 1.0 over kernel =
2.4.13 in
  > a Red Hat Linux 7.1 system. I have configured a MPLS network with =
three PCs
  > and it seems to work properly, but when I use ethereal to capture =
some
  > packets I watch some problems.
  > I'm testing with ping command, and when I make ping in one way it =
works
  > properly, it puts the MPLS labels over the IP packets, but in the =
other way
  > it doesn't put the MPLS labels it sends it as normal IP packets.
  > Also, I have test it with the 2 host network that is in the example =
of the
  > Linux-MPLS packet, but it continues happenning. =BFIs this a normal =
behavior
  > of Linux-MPLS?
  >=20
  > The scenario is:
  >=20
  >      160.20.20.1              160.20.20.2
  >      A   -------------------------------- B
  >             eth0                        eth1
  >=20
  >         160.20.20.0 /255.255.252.0
  >=20
  > And the mpls configuration is:
  >=20
  > On A
  > -------
  > ifconfig eth0 160.20.20.1 netmask 255.255.252.0 up
  > route add -host 160.20.20.2 gw 160.20.20.2
  > mplsadm -A -B -O gen:10:eth0:ipv4:160.20.20.2 -f 160.20.20.2/32
  > mplsadm -A -I gen:20:0
  > mplsadm -L eth0:0
  >=20
  > On B
  > -------
  > ifconfig eth1 160.20.20.2 netmask 255.255.252.0 up
  > route add -host 160.20.20.1 gw 160.20.20.1
  > mplsadm -A -B -O gen:20:eth0:ipv4:160.20.20.1 -f 160.20.20.1/32
  > mplsadm -A -I gen:10:0
  > mplsadm -L eth1:0
  >=20
  >=20
  > The packets from B to A are MPLS, but the replys of ping (or the =
request of
  > ping) from A to B are not MPLS labeled packets.
  >=20
  > And when I make ping from B to A, in ethereal (in A) appear Bogus IP =
headers
  > (I don't know what exactly are they) but they are IP packets with =
some MPLS
  > labels inside, and after the bogus IP, in ethereal capture, appear =
the
  > normal IP reply without MPLS label.
  >=20
  > When I make ping from A to B, in ethereal (in B) appear that the =
reply from
  > B are MPLS-IP labeled packets, but the requests from A are normal IP
  > packets. And in ethereal (in A) capture list appear that are =
malformated IP
  > headers (with unknown MPLS labels inside), and after this it sends =
normal IP
  > packets (the ping reply).
  >=20
  > The installation in A and B are the same (they are images maked with =
Norton
  > Ghost), the PCs are exactly the same (the hardware is the same). In =
A eth0
  > is a 3Com card, and in B eth1 is a Realtek card but it isn't the =
problem
  > because I have test it with B and other PC (same hardware and SO
  > installation) that have the same ethernet card in eth1, too, and the =
problem
  > is the same.
  >=20
  > Perhaps, Ethereal don't work properly? or, Is this the normal =
behavior of
  > ping? I don't know why it happens, =BFcan someone help me? thanks.
  >=20
  > Best regards, Dani.
  >=20
  >=20
  > _______________________________________________
  > mpls-linux-general mailing list
  > mpl...@li...
  > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

  --=20
  James R. Leu

Re: [mpls-linux-general] Problems with MPLS packets

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

This is a know problem.  It is not really a corrupted packet it just look=
s
that way from ethereal/tcpudmp etc.

From the "NOTE" file:
-mpls packets appear to be corrupted when running tcpdump or the like on
 an LER (the packets are not really corrupted, tcpdump is getting an inva=
lid
 version of the packets)=20

Jim


On Wed, Apr 10, 2002 at 02:22:00PM +0200, Daniel Gregorio Catalan wrote:
> (sorry for my english)
> I'm making a MPLS network with PCs with Linux-MPLS 1.0 over kernel 2.4.=
13 in
> a Red Hat Linux 7.1 system. I have configured a MPLS network with three=
 PCs
> and it seems to work properly, but when I use ethereal to capture some
> packets I watch some problems.
> I'm testing with ping command, and when I make ping in one way it works
> properly, it puts the MPLS labels over the IP packets, but in the other=
 way
> it doesn't put the MPLS labels it sends it as normal IP packets.
> Also, I have test it with the 2 host network that is in the example of =
the
> Linux-MPLS packet, but it continues happenning. =BFIs this a normal beh=
avior
> of Linux-MPLS?
>=20
> The scenario is:
>=20
>      160.20.20.1              160.20.20.2
>      A   -------------------------------- B
>             eth0                        eth1
>=20
>         160.20.20.0 /255.255.252.0
>=20
> And the mpls configuration is:
>=20
> On A
> -------
> ifconfig eth0 160.20.20.1 netmask 255.255.252.0 up
> route add -host 160.20.20.2 gw 160.20.20.2
> mplsadm -A -B -O gen:10:eth0:ipv4:160.20.20.2 -f 160.20.20.2/32
> mplsadm -A -I gen:20:0
> mplsadm -L eth0:0
>=20
> On B
> -------
> ifconfig eth1 160.20.20.2 netmask 255.255.252.0 up
> route add -host 160.20.20.1 gw 160.20.20.1
> mplsadm -A -B -O gen:20:eth0:ipv4:160.20.20.1 -f 160.20.20.1/32
> mplsadm -A -I gen:10:0
> mplsadm -L eth1:0
>=20
>=20
> The packets from B to A are MPLS, but the replys of ping (or the reques=
t of
> ping) from A to B are not MPLS labeled packets.
>=20
> And when I make ping from B to A, in ethereal (in A) appear Bogus IP he=
aders
> (I don't know what exactly are they) but they are IP packets with some =
MPLS
> labels inside, and after the bogus IP, in ethereal capture, appear the
> normal IP reply without MPLS label.
>=20
> When I make ping from A to B, in ethereal (in B) appear that the reply =
from
> B are MPLS-IP labeled packets, but the requests from A are normal IP
> packets. And in ethereal (in A) capture list appear that are malformate=
d IP
> headers (with unknown MPLS labels inside), and after this it sends norm=
al IP
> packets (the ping reply).
>=20
> The installation in A and B are the same (they are images maked with No=
rton
> Ghost), the PCs are exactly the same (the hardware is the same). In A e=
th0
> is a 3Com card, and in B eth1 is a Realtek card but it isn't the proble=
m
> because I have test it with B and other PC (same hardware and SO
> installation) that have the same ethernet card in eth1, too, and the pr=
oblem
> is the same.
>=20
> Perhaps, Ethereal don't work properly? or, Is this the normal behavior =
of
> ping? I don't know why it happens, =BFcan someone help me? thanks.
>=20
> Best regards, Dani.
>=20
>=20
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

--=20
James R. Leu

[mpls-linux-general] Problems with MPLS packets

[Daniel G. C. <da...@on...>]

(sorry for my english)
I'm making a MPLS network with PCs with Linux-MPLS 1.0 over kernel 2.4.13 in
a Red Hat Linux 7.1 system. I have configured a MPLS network with three PCs
and it seems to work properly, but when I use ethereal to capture some
packets I watch some problems.
I'm testing with ping command, and when I make ping in one way it works
properly, it puts the MPLS labels over the IP packets, but in the other way
it doesn't put the MPLS labels it sends it as normal IP packets.
Also, I have test it with the 2 host network that is in the example of the
Linux-MPLS packet, but it continues happenning. ¿Is this a normal behavior
of Linux-MPLS?

The scenario is:

     160.20.20.1              160.20.20.2
     A   -------------------------------- B
            eth0                        eth1

        160.20.20.0 /255.255.252.0

And the mpls configuration is:

On A
-------
ifconfig eth0 160.20.20.1 netmask 255.255.252.0 up
route add -host 160.20.20.2 gw 160.20.20.2
mplsadm -A -B -O gen:10:eth0:ipv4:160.20.20.2 -f 160.20.20.2/32
mplsadm -A -I gen:20:0
mplsadm -L eth0:0

On B
-------
ifconfig eth1 160.20.20.2 netmask 255.255.252.0 up
route add -host 160.20.20.1 gw 160.20.20.1
mplsadm -A -B -O gen:20:eth0:ipv4:160.20.20.1 -f 160.20.20.1/32
mplsadm -A -I gen:10:0
mplsadm -L eth1:0


The packets from B to A are MPLS, but the replys of ping (or the request of
ping) from A to B are not MPLS labeled packets.

And when I make ping from B to A, in ethereal (in A) appear Bogus IP headers
(I don't know what exactly are they) but they are IP packets with some MPLS
labels inside, and after the bogus IP, in ethereal capture, appear the
normal IP reply without MPLS label.

When I make ping from A to B, in ethereal (in B) appear that the reply from
B are MPLS-IP labeled packets, but the requests from A are normal IP
packets. And in ethereal (in A) capture list appear that are malformated IP
headers (with unknown MPLS labels inside), and after this it sends normal IP
packets (the ping reply).

The installation in A and B are the same (they are images maked with Norton
Ghost), the PCs are exactly the same (the hardware is the same). In A eth0
is a 3Com card, and in B eth1 is a Realtek card but it isn't the problem
because I have test it with B and other PC (same hardware and SO
installation) that have the same ethernet card in eth1, too, and the problem
is the same.

Perhaps, Ethereal don't work properly? or, Is this the normal behavior of
ping? I don't know why it happens, ¿can someone help me? thanks.

Best regards, Dani.

Re: [mpls-linux-general] Address Mesg

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

On Wed, Apr 10, 2002 at 03:52:14PM +0800, clz wrote:
> Hi,Jim
>      When I read the RFC3036,i have some questions:
> 1) When there is a topology  change ,how does the LSR find a new next hop for an LSP? Does the Address Mesg give some help for this?

Correct.

> 2) In the Hello Message, if the Transport Address TLV is not present, when opening the LDP session TCP connection, we should use the IPv4 source address for the UDP packet or the LDP Identifier's IP addr for the LDP PDU?

Source address of the LDP hello packet.

> 3) In which case, does the LSR advertise its interface addresses with more Address Message? I think, one Address Message with one interface address is enough. Is it right?

When interfaces go up or down. (think link failure or a newly
provisioned interface).

Jim
-- 
James R. Leu

[mpls-linux-general] Address Mesg

[clz <cui...@ya...>]
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==


_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com

[mpls-linux-general] About LDP's program

[Lauren H <hly...@ya...>]

SGksSmltDQogICAgICAgICAgaSB3YW50IHRvIHVuZGVyc3RhbmQgdGhlIG1wbHMvbGRwIHByb2dy
YW0gaW4gZGVwdGguIHdvdWxkIHUgbGlrZSB0ZWxsIG1lIHRoZSBzdHJ1Y3R1cmUgb2YgdGhlIHBy
b2dyYW0/IHdoZXJlIGlzIHRoZSBtYWluKCkgcHJvZ3JhbT8NCg0KdGhhbnggaW4gYWR2YW5jZS4N
CkJFU1QgUkVHQVJEUw0KDQogTGF1cmVuIEgNCiAgICAgICAgIA0K


_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com

Re: [mpls-linux-general] Network adapter(xircom) problem !

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

On Wed, Apr 10, 2002 at 01:37:46AM +0500, nit...@si... wrote:
> Hi all,
> This is what we did:
> 1.    Installed redhat linux 7.2 with kernel source 2.4.7-10
> 2.    Downloaded and patched the kernel source with mpls-linux-1.0.tar.gz
> 3.    Then I followed the quickstart,
>         a)make menuconfig,
>           turned the MPLS support on,Didnt need to touch anything else
> 
>          b)make dep.
>          c)make clean.
>          d)make bzlilo

Close ...

make modules
make modules_install

Should fix your problem (remove your symbolic link first).

> 4.    Now, landed up with system.map and vmlinuz in "/"  dir
> There was no entry for this newly compiled kernel; so added an entry pointing to /vmlinuz and ran lilo.
> 5.    Then rebooted the new kernel
> 6.    Upon rebooting, it showed some modules were not in 
>       /lib/modules/2.4.7-10custom. So created a symbolic link with that name to /lib/modules/2.4.7-10.
> 7.    Now, all modules got inserted except xirc2ps_cs.o. It's failing modprobe with  loads of unresolved symbols.
> The other needed pcmcia modules get inserted without any problem.
> 
> There was a similar posting on the mailing list which was pointed out to config files. Now, I'm assuming that the config files installed by kernel-source-2.4.7-10 are same as those of kenel-2.4.7-10 (my older kernel)
> 
> If its not so, how can I get the older config.
> Any help in this regards will be highly appreciated.
> Thanx in advance. with regards. Nitinc 
> 
> -------------------------------------------------
> This mail helped a tree grow. Know more at http://green.sify.com
> 
> Take the shortest route to success! 
> Click here to know how http://education.sify.com
> 
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu

[mpls-linux-general] Network adapter(xircom) problem !

[<nit...@si...>]

Hi all,
This is what we did:
1.    Installed redhat linux 7.2 with kernel source 2.4.7-10
2.    Downloaded and patched the kernel source with mpls-linux-1.0.tar.gz
3.    Then I followed the quickstart,
        a)make menuconfig,
          turned the MPLS support on,Didnt need to touch anything else

         b)make dep.
         c)make clean.
         d)make bzlilo
4.    Now, landed up with system.map and vmlinuz in "/"  dir
There was no entry for this newly compiled kernel; so added an entry pointing to /vmlinuz and ran lilo.
5.    Then rebooted the new kernel
6.    Upon rebooting, it showed some modules were not in 
      /lib/modules/2.4.7-10custom. So created a symbolic link with that name to /lib/modules/2.4.7-10.
7.    Now, all modules got inserted except xirc2ps_cs.o. It's failing modprobe with  loads of unresolved symbols.
The other needed pcmcia modules get inserted without any problem.

There was a similar posting on the mailing list which was pointed out to config files. Now, I'm assuming that the config files installed by kernel-source-2.4.7-10 are same as those of kenel-2.4.7-10 (my older kernel)

If its not so, how can I get the older config.
Any help in this regards will be highly appreciated.
Thanx in advance. with regards. Nitinc 

-------------------------------------------------
This mail helped a tree grow. Know more at http://green.sify.com

Take the shortest route to success! 
Click here to know how http://education.sify.com

[mpls-linux-general] Problems with receiving packets on MPLS

[zeljan s. <zs...@pi...>]

Hi again,

Background:
Installed MPLS on 2 computers under Linux Debian release (with the help of CVS...got the newest version). 
The Link Layer is Ethernet.

Problem:
When I set up the tunnel from host A to host B using the mplsadm there are
no errors. But when I try to ping host B from host A I get no reply. With
the help of Ethereal program I can see that the packets sent from host A
have a "shim" (MPLS)header and are being sent with the correct outgoing label, but on the host B those packets never arrive (I don't see them on Ethereal). Or my thought is that they arrive, but when the host B checks the MPLS header he rejects them. Maybe something is wrong when he's trying to match the incoming label to the label that is set for the incoming packets. The same thing happens the other way around, if host A is pinged from B. 
I also checked the /proc/net/mpls_in and mpls_out and found that the labels are set.
One more thing, when trying to delete an outgoing label,
with mplsadm -D, I get error message that there is no such process.


Any help appreciated,

Zeljan



    

[mpls-linux-general] problem DOD

[anish v. <ani...@re...>]

hi jim,
  as u have suggested i changed the label_distribution_mode to DOD 
in ldp_deafults.h.
after that i checked in initialization msg the 'A' bit it's 
correct..
as i start ldp_linux
after adding global n interface it gives errors

ldp_label_mapping_send
ldp_linux:ldp_attr.c:515 ldp_attr_insert_upstream : Assertion ' 
a->in_tree == LDP_FALSE ' failed
Aborted (core dumped)

n after commenting above assert it gives following error
ldp_label_mapping_send
ldp_linux:ldp_attr.c:899 ldp_fs_add_attr : Assertion
' ptr == ((void *)0)' failed
Aborted (core dumped)

help me in this matter
bye
anish

Re: [mpls-linux-general] Problems with MPLS 128, kernel crash

[<pim...@in...>]

"James R. Leu" wrote:
> 
> *sigh*
> 
> I have seen this before.  I must be doing something wrong with the way I
> handle skbs.  I haven't been able to figure out what, but I must be doing
> something wrong.
Probably not saying something new but: 
sending packets slow: works
sending packets "fast" with debugging on: works
sending fast without debugging on a rate limited queue: *bang*	

My guess would be that something goes wrong when dropping packets.

Pim.
> 
> Jim
> 
> BTW ... impressive instructions :-)
> 
> On Tue, Apr 02, 2002 at 04:09:06PM +0200, pim...@in... wrote:
> > Hi James,
> >
> > I am able to crash the kernel (2.4.17+MPLS128+(ftos&dscp) patches from
> > iptables-1.2.6a) if debugging is off.
> >
> > Attached is the script that sets up the DS tree.
> >
> > cat /proc/net/mpls* (only relevant bits)
> > 0x00000003 0/0/0 2 SET_EXP(01) EXP2TC( EXP(0)->TC(ffff) EXP(1)->TC(00b8)
> > EXP(2)->TC(0028) EXP(3)->TC(0030) EXP(4)->TC(0048) EXP(5)->TC(0050)
> > EXP(6)->TC(0068) EXP(7)->TC(0070) ) PUSH(gen 21650) SET(eth2,10.0.4.1)
> > T21650/1        0x00000003
> >
> > traffic is mapped on T21650/1 via iptables (default gateway on table 2)
> >
> > ping -f 10.0.2.2
> > .... (couple packets are sent)
> > *crash*
> >
> > [console]
> > length:4
> > warning: kfree pased on skb still on list from 0xc01f9e18
> > kernel BUG: at skbuff.c:321
> > ...
> > kernel panic: Aiee killing the interrupt handler...
> >
> > I want to stress the fact that I am only able to reproduce it when
> > debugging is off.
> > Let me know if you need additional information/want me to test
> > something.
> >
> > Pim.
> >
> > --
> > Pim Van Heuven            Ghent University - IMEC
> > Department  of  Information  Technology   (INTEC)
> > Sint-Pietersnieuwstraat, 41 B-9000, Gent, Belgium
> > pim...@in...
> >         DiffServ/MPLS(RSVP-TE) for Linux:
> >         http://dsmpls.atlantis.rug.ac.be
> > #!/bin/bash
> > # This script will set-up an example DS tree
> >
> > IFACES="eth1 eth2 eth3"
> > TC=/usr/local/bin/tc
> >
> > if test $# -gt 0
> > then
> >   case $1 in
> >   s|-s)
> >       if test $# -gt 1
> >       then
> >               $TC -s qdisc ls dev $2
> >       else
> >               for i in $IFACES; do echo "***"$i"***"
> >               $TC -s qdisc ls dev eth2; done
> >       fi;;
> >   d|-d)
> >       if test $# -gt 1
> >       then
> >               $TC qdisc del root dev $2
> >               $TC filter del root dev $2
> >               $TC class del root dev $2
> >       else
> >               for i in $IFACES; do
> >               $TC qdisc del root dev $i
> >               $TC filter del root dev $i
> >               $TC class del root dev $i
> >               done
> >       fi;;
> >   h|-h|--help|-help)
> >         echo $0
> >       echo "     -s show statistics"
> >       echo "     -d delete filters and qdiscs"
> >       echo "     install qdiscs and filers"
> >   esac
> >   exit
> > else
> >
> > #iptables -A OUTPUT -t mangle -p rsvp -o eth0 -j DROP
> >
> > for i in $IFACES; do
> >
> >   echo Configuring interface $i
> >
> >   echo DSMARK on $i
> >   $TC qdisc add dev $i root handle 1:0 dsmark indices 64 default_index 0 #tcindex classid 2.2 #set_tc_index
> >
> >   echo Main Filter on $i
> >   $TC filter add dev $i parent 1:0 protocol mpls tcindex mask 0xfc shift 2
> >   echo prio
> >
> > ##### Set up of the proper qdisc on $i
> >
> > # First a DSMARK qdisc is introduced in order to retrieve TOS from IP header
> > #$TC qdisc add dev $i handle 1:0 root dsmark indices 64 #set_tc_index
> > $TC filter add dev $i parent 1:0 protocol mpls prio 1 tcindex mask 0xfc shift 2 pass_on
> >
> > # Second a CBQ qdisc is used in order to support EF, AF and BE classes
> > $TC qdisc add dev $i parent 1:0 handle 2:0 cbq bandwidth 100Mbit cell 8 avpkt 1000 mpu 64
> > $TC filter add dev $i parent 2:0 protocol mpls prio 1 tcindex mask 0xf0 shift 4 pass_on
> >
> > ## Definition of the CBQ leaf classes to support AF, EF and BE
> > echo AF1
> > ## AF Class 1 specific setup
> > $TC class add dev $i parent 2:0 classid 2:1 cbq bandwidth 100Mbit rate 15Mbit \
> >     avpkt 1000 prio 5 allot 1514 weight 1.5Mbit maxburst 21 defmap 0 borrow
> > $TC filter add dev $i parent 2:0 protocol mpls prio 1 handle 1 tcindex classid 2:1
> > $TC qdisc add dev $i parent 2:1 gred setup DPs 3 default 2 grio
> > # --- AF Class 1 DP 1---
> >    $TC filter add dev $i parent 1:0 protocol mpls prio 1 handle 0x0a tcindex classid 1:111
> >    $TC qdisc change dev $i parent 2:1 gred limit 600KB min 150KB max 450KB burst 200 \
> >        avpkt 1000 bandwidth 100Mbit DP 1 probability 0.02 prio 2
> > # --- AF Class 1 DP 2---
> >    $TC filter add dev $i parent 1:0 protocol mpls prio 1 handle 0x0c tcindex classid 1:112
> >    $TC qdisc change dev $i parent 2:1 gred limit 600KB min 150KB max 450KB burst 200 \
> >        avpkt 1000 bandwidth 100Mbit DP 2 probability 0.04 prio 3
> >
> > ## AF Class 2 specific setup
> > echo AF2
> > $TC class add dev $i parent 2:0 classid 2:2 cbq bandwidth 100Mbit rate 15Mbit \
> >     avpkt 1000 prio 4 allot 1514 weight 1.5Mbit maxburst 21 defmap 0 borrow
> > $TC filter add dev $i parent 2:0 protocol mpls prio 1 handle 2 tcindex classid 2:2
> > $TC qdisc add dev $i parent 2:2 gred setup DPs 3 default 2 grio
> > # --- AF Class 2 DP 1---
> >    $TC filter add dev $i parent 1:0 protocol mpls prio 1 handle 0x12 tcindex classid 1:121
> >    $TC qdisc change dev $i parent 2:2 gred limit 600KB min 150KB max 450KB burst 200 \
> >        avpkt 1000 bandwidth 100Mbit DP 1 probability 0.02 prio 2
> > # --- AF Class 2 DP 2---
> >   $TC filter add dev $i parent 1:0 protocol mpls prio 1 handle 0x14 tcindex classid 1:122
> >    $TC qdisc change dev $i parent 2:2 gred limit 600KB min 150KB max 450KB burst 200 \
> >        avpkt 1000 bandwidth 100Mbit DP 2 probability 0.04 prio 3
> >
> > ## AF Class 3 specific setup
> > echo AF3
> > $TC class add dev $i parent 2:0 classid 2:3 cbq bandwidth 100Mbit rate 15Mbit \
> >     avpkt 1000 prio 3 allot 1514 weight 1.5Mbit maxburst 21 defmap 0 borrow
> > $TC filter add dev $i parent 2:0 protocol mpls prio 1 handle 3 tcindex classid 2:3
> > $TC qdisc add dev $i parent 2:3 gred setup DPs 3 default 2 grio
> > # --- AF Class 3 DP 1---
> >    $TC filter add dev $i parent 1:0 protocol mpls prio 1 handle 0x1a tcindex classid 1:131
> >    $TC qdisc change dev $i parent 2:3 gred limit 600KB min 150KB max 450KB burst 200 \
> >        avpkt 1000 bandwidth 100Mbit DP 1 probability 0.02 prio 2
> > # --- AF Class 3 DP 2---
> >    $TC filter add dev $i parent 1:0 protocol mpls prio 1 handle 0x1c tcindex classid 1:132
> >    $TC qdisc change dev $i parent 2:3 gred limit 600KB min 150KB max 450KB burst 200 \
> >        avpkt 1000 bandwidth 100Mbit DP 2 probability 0.04 prio 3
> >
> > ## EF class specific setup
> > echo EF
> > $TC class add dev $i parent 2:0 classid 2:5 cbq bandwidth 100Mbit rate 10Mbit \
> >     avpkt 1000 prio 1 bounded allot 1514 weight 1Mbit maxburst 10 defmap 0
> > $TC qdisc add dev $i parent 2:5 pfifo limit 10
> > $TC filter add dev $i parent 1:0 protocol mpls prio 1 handle 0x2e tcindex classid 1:151
> > $TC filter add dev $i parent 2:0 protocol mpls prio 1 handle 5 tcindex classid 2:5
> >
> > ## BE class specific setup
> > echo BE
> > $TC class add dev $i parent 2:0 classid 2:6 cbq bandwidth 100Mbit rate 30Mbit \
> >     avpkt 1000 prio 7 allot 1514 weight 3Mbit maxburst 21 borrow
> > $TC qdisc add dev $i parent 2:6 red limit 600KB min 150KB max 450KB burst 200 \
> >     avpkt 1000 bandwidth 100Mbit probability 0.4
> > $TC filter add dev $i parent 1:0 protocol mpls prio 1 handle 0x0 tcindex classid 1:161
> > $TC filter add dev $i parent 2:0 protocol mpls prio 1 handle 6 tcindex classid 2:6
> >
> > ## OSPF - RSVP Pipe class specific setup
> > echo OSPF \& RSVP filter
> > $TC class add dev $i parent 2:0 classid 2:7 cbq bandwidth 100Mbit rate 1Mbit \
> >     avpkt 1000 prio 1 bounded allot 1514 weight 1Mbit maxburst 10 defmap 0
> > $TC qdisc add dev $i parent 2:5 pfifo limit 10
> > $TC filter add dev $i parent 1:0 protocol mpls prio 1 handle 0x30 tcindex classid 1:171
> > $TC filter add dev $i parent 2:0 protocol mpls prio 1 handle 7 tcindex classid 2:7
> >
> > done
> >
> > fi
> >
> >
> >
> >
> >
> 
> --
> James R. Leu

-- 
Pim Van Heuven            Ghent University - IMEC
Department  of  Information  Technology   (INTEC)
Sint-Pietersnieuwstraat, 41 B-9000, Gent, Belgium
pim...@in...
        DiffServ/MPLS(RSVP-TE) for Linux:
        http://dsmpls.atlantis.rug.ac.be

Re: [mpls-linux-general] problem with ldp

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

On Mon, Apr 08, 2002 at 11:54:15AM -0000, anish  verma wrote:
> hi jim,
>   as u have told me that ldp is not dynamic.
> i tried a below combination
> 
> ------- eth0            eth0  ---------
> :  1  :-----------------------:    2    :
> ------- 10                60   --------
> 
> all ip 10.129.152.*
> 
> on node 1
> ./ldp_linux
> >add global 10.129.152.10
> >add interface eth0
> than added routing entry of 10.129.152.60 to kernel routing table 
> of node 1 by
> add route 10.129.152.60/32 gw 10.129.152.60 dev eth0
> and also to fec tree of ldp..
> >add route 10.129.152.60/32 gw 10.129.152.60 dev eth0
> after that node 1 sends a mapping msg regarding new fec to node2 
> but no entry is added to mpls_out..
> i think it should work in this way..
> am i right..
> 
> i have one more query.. why there is no label request in ldp_linux 

ldp_linux defaults to DU mode.  Modify ldp_defaults.h to turn on DOD mode.

Jim
-- 
James R. Leu

[mpls-linux-general] problem with ldp

[anish v. <ani...@re...>]

hi jim,
  as u have told me that ldp is not dynamic.
i tried a below combination

------- eth0            eth0  ---------
:  1  :-----------------------:    2    :
------- 10                60   --------

all ip 10.129.152.*

on node 1
./ldp_linux
>add global 10.129.152.10
>add interface eth0
than added routing entry of 10.129.152.60 to kernel routing table 
of node 1 by
add route 10.129.152.60/32 gw 10.129.152.60 dev eth0
and also to fec tree of ldp..
>add route 10.129.152.60/32 gw 10.129.152.60 dev eth0
after that node 1 sends a mapping msg regarding new fec to node2 
but no entry is added to mpls_out..
i think it should work in this way..
am i right..

i have one more query.. why there is no label request in ldp_linux 
??
thanx in advance
anish

Re: [mpls-linux-general] Problems with MPLS 128, kernel crash

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

*sigh*

I have seen this before.  I must be doing something wrong with the way I
handle skbs.  I haven't been able to figure out what, but I must be doing
something wrong.

Jim

BTW ... impressive instructions :-)

On Tue, Apr 02, 2002 at 04:09:06PM +0200, pim...@in... wrote:
> Hi James,
> 
> I am able to crash the kernel (2.4.17+MPLS128+(ftos&dscp) patches from
> iptables-1.2.6a) if debugging is off.
> 
> Attached is the script that sets up the DS tree.
> 
> cat /proc/net/mpls* (only relevant bits)
> 0x00000003 0/0/0 2 SET_EXP(01) EXP2TC( EXP(0)->TC(ffff) EXP(1)->TC(00b8)
> EXP(2)->TC(0028) EXP(3)->TC(0030) EXP(4)->TC(0048) EXP(5)->TC(0050)
> EXP(6)->TC(0068) EXP(7)->TC(0070) ) PUSH(gen 21650) SET(eth2,10.0.4.1)
> T21650/1        0x00000003
> 
> traffic is mapped on T21650/1 via iptables (default gateway on table 2)
> 
> ping -f 10.0.2.2
> .... (couple packets are sent)
> *crash*
> 
> [console]
> length:4
> warning: kfree pased on skb still on list from 0xc01f9e18
> kernel BUG: at skbuff.c:321
> ...
> kernel panic: Aiee killing the interrupt handler...
> 
> I want to stress the fact that I am only able to reproduce it when
> debugging is off.
> Let me know if you need additional information/want me to test
> something.
> 
> Pim.
> 
> -- 
> Pim Van Heuven            Ghent University - IMEC
> Department  of  Information  Technology   (INTEC)
> Sint-Pietersnieuwstraat, 41 B-9000, Gent, Belgium
> pim...@in...
>         DiffServ/MPLS(RSVP-TE) for Linux:
>         http://dsmpls.atlantis.rug.ac.be
> #!/bin/bash
> # This script will set-up an example DS tree
> 
> IFACES="eth1 eth2 eth3" 
> TC=/usr/local/bin/tc
> 
> if test $# -gt 0
> then
>   case $1 in
>   s|-s) 
> 	if test $# -gt 1 
> 	then 
> 		$TC -s qdisc ls dev $2
> 	else
> 		for i in $IFACES; do echo "***"$i"***"	
> 		$TC -s qdisc ls dev eth2; done
> 	fi;;
>   d|-d) 
> 	if test $# -gt 1 
> 	then 
> 		$TC qdisc del root dev $2
> 		$TC filter del root dev $2
> 		$TC class del root dev $2
> 	else	
> 		for i in $IFACES; do 
> 		$TC qdisc del root dev $i
> 		$TC filter del root dev $i
> 		$TC class del root dev $i
> 		done
> 	fi;;
>   h|-h|--help|-help)
>         echo $0 
> 	echo "     -s show statistics"
> 	echo "     -d delete filters and qdiscs"
> 	echo "     install qdiscs and filers"
>   esac
>   exit
> else
> 
> #iptables -A OUTPUT -t mangle -p rsvp -o eth0 -j DROP
> 
> for i in $IFACES; do
> 
>   echo Configuring interface $i
> 
>   echo DSMARK on $i
>   $TC qdisc add dev $i root handle 1:0 dsmark indices 64 default_index 0 #tcindex classid 2.2 #set_tc_index 
> 
>   echo Main Filter on $i
>   $TC filter add dev $i parent 1:0 protocol mpls tcindex mask 0xfc shift 2
>   echo prio
> 
> ##### Set up of the proper qdisc on $i
> 
> # First a DSMARK qdisc is introduced in order to retrieve TOS from IP header
> #$TC qdisc add dev $i handle 1:0 root dsmark indices 64 #set_tc_index
> $TC filter add dev $i parent 1:0 protocol mpls prio 1 tcindex mask 0xfc shift 2 pass_on
> 
> # Second a CBQ qdisc is used in order to support EF, AF and BE classes
> $TC qdisc add dev $i parent 1:0 handle 2:0 cbq bandwidth 100Mbit cell 8 avpkt 1000 mpu 64
> $TC filter add dev $i parent 2:0 protocol mpls prio 1 tcindex mask 0xf0 shift 4 pass_on
> 
> ## Definition of the CBQ leaf classes to support AF, EF and BE
> echo AF1
> ## AF Class 1 specific setup
> $TC class add dev $i parent 2:0 classid 2:1 cbq bandwidth 100Mbit rate 15Mbit \
>     avpkt 1000 prio 5 allot 1514 weight 1.5Mbit maxburst 21 defmap 0 borrow
> $TC filter add dev $i parent 2:0 protocol mpls prio 1 handle 1 tcindex classid 2:1
> $TC qdisc add dev $i parent 2:1 gred setup DPs 3 default 2 grio
> # --- AF Class 1 DP 1---
>    $TC filter add dev $i parent 1:0 protocol mpls prio 1 handle 0x0a tcindex classid 1:111
>    $TC qdisc change dev $i parent 2:1 gred limit 600KB min 150KB max 450KB burst 200 \
>        avpkt 1000 bandwidth 100Mbit DP 1 probability 0.02 prio 2
> # --- AF Class 1 DP 2---
>    $TC filter add dev $i parent 1:0 protocol mpls prio 1 handle 0x0c tcindex classid 1:112
>    $TC qdisc change dev $i parent 2:1 gred limit 600KB min 150KB max 450KB burst 200 \
>        avpkt 1000 bandwidth 100Mbit DP 2 probability 0.04 prio 3
> 
> ## AF Class 2 specific setup
> echo AF2
> $TC class add dev $i parent 2:0 classid 2:2 cbq bandwidth 100Mbit rate 15Mbit \
>     avpkt 1000 prio 4 allot 1514 weight 1.5Mbit maxburst 21 defmap 0 borrow
> $TC filter add dev $i parent 2:0 protocol mpls prio 1 handle 2 tcindex classid 2:2
> $TC qdisc add dev $i parent 2:2 gred setup DPs 3 default 2 grio
> # --- AF Class 2 DP 1---
>    $TC filter add dev $i parent 1:0 protocol mpls prio 1 handle 0x12 tcindex classid 1:121
>    $TC qdisc change dev $i parent 2:2 gred limit 600KB min 150KB max 450KB burst 200 \
>        avpkt 1000 bandwidth 100Mbit DP 1 probability 0.02 prio 2
> # --- AF Class 2 DP 2---
>   $TC filter add dev $i parent 1:0 protocol mpls prio 1 handle 0x14 tcindex classid 1:122
>    $TC qdisc change dev $i parent 2:2 gred limit 600KB min 150KB max 450KB burst 200 \
>        avpkt 1000 bandwidth 100Mbit DP 2 probability 0.04 prio 3
> 
> ## AF Class 3 specific setup
> echo AF3
> $TC class add dev $i parent 2:0 classid 2:3 cbq bandwidth 100Mbit rate 15Mbit \
>     avpkt 1000 prio 3 allot 1514 weight 1.5Mbit maxburst 21 defmap 0 borrow
> $TC filter add dev $i parent 2:0 protocol mpls prio 1 handle 3 tcindex classid 2:3
> $TC qdisc add dev $i parent 2:3 gred setup DPs 3 default 2 grio
> # --- AF Class 3 DP 1---
>    $TC filter add dev $i parent 1:0 protocol mpls prio 1 handle 0x1a tcindex classid 1:131
>    $TC qdisc change dev $i parent 2:3 gred limit 600KB min 150KB max 450KB burst 200 \
>        avpkt 1000 bandwidth 100Mbit DP 1 probability 0.02 prio 2
> # --- AF Class 3 DP 2---
>    $TC filter add dev $i parent 1:0 protocol mpls prio 1 handle 0x1c tcindex classid 1:132
>    $TC qdisc change dev $i parent 2:3 gred limit 600KB min 150KB max 450KB burst 200 \
>        avpkt 1000 bandwidth 100Mbit DP 2 probability 0.04 prio 3
> 
> ## EF class specific setup
> echo EF
> $TC class add dev $i parent 2:0 classid 2:5 cbq bandwidth 100Mbit rate 10Mbit \
>     avpkt 1000 prio 1 bounded allot 1514 weight 1Mbit maxburst 10 defmap 0
> $TC qdisc add dev $i parent 2:5 pfifo limit 10
> $TC filter add dev $i parent 1:0 protocol mpls prio 1 handle 0x2e tcindex classid 1:151
> $TC filter add dev $i parent 2:0 protocol mpls prio 1 handle 5 tcindex classid 2:5
> 
> ## BE class specific setup
> echo BE
> $TC class add dev $i parent 2:0 classid 2:6 cbq bandwidth 100Mbit rate 30Mbit \
>     avpkt 1000 prio 7 allot 1514 weight 3Mbit maxburst 21 borrow
> $TC qdisc add dev $i parent 2:6 red limit 600KB min 150KB max 450KB burst 200 \
>     avpkt 1000 bandwidth 100Mbit probability 0.4
> $TC filter add dev $i parent 1:0 protocol mpls prio 1 handle 0x0 tcindex classid 1:161    
> $TC filter add dev $i parent 2:0 protocol mpls prio 1 handle 6 tcindex classid 2:6
> 
> ## OSPF - RSVP Pipe class specific setup
> echo OSPF \& RSVP filter
> $TC class add dev $i parent 2:0 classid 2:7 cbq bandwidth 100Mbit rate 1Mbit \
>     avpkt 1000 prio 1 bounded allot 1514 weight 1Mbit maxburst 10 defmap 0
> $TC qdisc add dev $i parent 2:5 pfifo limit 10
> $TC filter add dev $i parent 1:0 protocol mpls prio 1 handle 0x30 tcindex classid 1:171
> $TC filter add dev $i parent 2:0 protocol mpls prio 1 handle 7 tcindex classid 2:7
> 
> done
> 
> fi
> 
> 
> 
> 
> 


-- 
James R. Leu

Re: [mpls-linux-general] some more ques reg LDP

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

On Sat, Apr 06, 2002 at 04:21:36PM -0500, Anil.H wrote:
> Hi,
> 
> I have the following questions:-
> 
> I would like to know, how do i initialize the value of Message ID in the
> LDP header and what its significance is?

No, the LDP stack generates and track the mesg ids.
It can be used to help specify a specific label mapping or release.


> Am I right in saying that Address and Label advertisement messages occur
> in the Operational mode.

Yes.

Jim
-- 
James R. Leu

Re: [mpls-linux-general] ldp-portable compilation errors

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

Try adding -I /usr/src/linux/include to CFLAGS the Makefile.

Jim

On Fri, Apr 05, 2002 at 03:09:45AM -0500, eli...@ne... wrote:
> Hi all,
> 
> I am started to play with MPLS on linux.
> I am using 2.4.17 patched with mpls-linux_1_128.diff.
> compilation was OK. 
> Then tring to compile ldp-portable (CVS version) failed with the
> following message:
> ldp_netlink.h:16: warning: `struct mpls_bind_fec_req` declared inside parameter list
> 
> Than I removed the -Werror compilation failed again.
> ldp_mpls_impl.c: In function `ldp_mpls_outlabel_add':
> ldp_mpls_impl.c:47: structure has no member named `sa_family'
> ldp_mpls_impl.c:51: `RTM_NEWNHLFE' undeclared (first use in this function)
> ldp_mpls_impl.c:51: (Each undeclared identifier is reported only once
> ldp_mpls_impl.c:51: for each function it appears in.)
> ldp_mpls_impl.c: In function `ldp_mpls_outlabel_del':
> ldp_mpls_impl.c:64: structure has no member named `sa_family'
> ldp_mpls_impl.c:68: `RTM_DELNHLFE' undeclared (first use in this function)
> ldp_mpls_impl.c: In function `ldp_mpls_inlabel_add':
> ldp_mpls_impl.c:87: `RTM_NEWILM' undeclared (first use in this function)
> ldp_mpls_impl.c: In function `ldp_mpls_inlabel_del':
> ldp_mpls_impl.c:101: `RTM_DELILM' undeclared (first use in this function)
> ldp_mpls_impl.c: In function `ldp_mpls_in2out_add':
> ldp_mpls_impl.c:122: structure has no member named `sa_family'
> ldp_mpls_impl.c:132: `RTM_NEWXC' undeclared (first use in this function)
> ldp_mpls_impl.c: In function `ldp_mpls_in2out_del':
> ldp_mpls_impl.c:154: structure has no member named `sa_family'
> ldp_mpls_impl.c:164: `RTM_DELXC' undeclared (first use in this function)
> ldp_mpls_impl.c: In function `ldp_mpls_fec2out_add':
> ldp_mpls_impl.c:172: storage size of `mbf_req' isn't known
> ldp_mpls_impl.c:173: sizeof applied to an incomplete type
> ldp_mpls_impl.c:184: `RTM_NEWFTN' undeclared (first use in this function)
> ldp_mpls_impl.c:172: warning: unused variable `mbf_req'
> ldp_mpls_impl.c: In function `ldp_mpls_fec2out_del':
> ldp_mpls_impl.c:192: storage size of `mbf_req' isn't known
> ldp_mpls_impl.c:193: sizeof applied to an incomplete type
> ldp_mpls_impl.c:201: `RTM_DELFTN' undeclared (first use in this function)
> ldp_mpls_impl.c:192: warning: unused variable `mbf_req'
> make[1]: *** [ldp_mpls_impl.o] Error 1
> make[1]: Leaving directory `/root/dev/cvs/ldp-portable/linux-port'
> 
> Also is zebra-port is maintained, the files are very old.
> 
> Any help will be appreciated  
> Thanks,
> Eli
> 
> 
> 
> __________________________________________________________________
> Your favorite stores, helpful shopping tools and great gift ideas. Experience the convenience of buying online with Shop@Netscape! http://shopnow.netscape.com/
> 
> Get your own FREE, personal Netscape Mail account today at http://webmail.netscape.com/
> 
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu

Re: [mpls-linux-general] ldp wont compile

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

Make sure to do a make clean before compiling.  If that doesn't
work then I think there are some system library or compiler issues
on your system.

Jim

On Sun, Mar 31, 2002 at 02:46:01PM +0100, Hotlinks Internet Services wrote:
> Hello again,
> 
> I decided to "have a play" with mpls to see exactly what its capabilities are at the moment, mpls-linux compiled no problems,
> however I seem to be having a strange problem compiling ldp-portable
> 
> Here is what happens when I type make....
> 
> root@test:/usr/src/ldp-portable-0.060# make
> cd port ; make ldp_linux.o
> make[1]: Entering directory `/usr/src/ldp-portable-0.060/port'
> make[1]: `ldp_linux.o' is up to date.
> make[1]: Leaving directory `/usr/src/ldp-portable-0.060/port'
> cd port ; make libldpport.a
> make[1]: Entering directory `/usr/src/ldp-portable-0.060/port'
> gcc -I../lib -I. -Wall -g -O2 -D__LDP_LINUX__ -DMPLS_LINUX   -c -o ldp_mpls_impl.o ldp_mpls_impl.c
> ldp_mpls_impl.c: In function `ldp_mpls_outlabel_add':
> ldp_mpls_impl.c:47: warning: implicit declaration of function `send_nhlfe'
> ldp_mpls_impl.c: In function `ldp_mpls_inlabel_add':
> ldp_mpls_impl.c:82: warning: implicit declaration of function `send_ilm'
> ldp_mpls_impl.c: In function `ldp_mpls_in2out_add':
> ldp_mpls_impl.c:125: warning: implicit declaration of function `send_xc'
> ldp_mpls_impl.c: In function `ldp_mpls_fec2out_add':
> ldp_mpls_impl.c:174: warning: implicit declaration of function `send_ftn'
> gcc -I../lib -I. -Wall -g -O2 -D__LDP_LINUX__ -DMPLS_LINUX   -c -o ldp_socket_impl.o ldp_socket_impl.c
> gcc -I../lib -I. -Wall -g -O2 -D__LDP_LINUX__ -DMPLS_LINUX   -c -o ldp_timer_impl.o ldp_timer_impl.c
> gcc -I../lib -I. -Wall -g -O2 -D__LDP_LINUX__ -DMPLS_LINUX   -c -o ldp_tree_impl.o ldp_tree_impl.c
> gcc -I../lib -I. -Wall -g -O2 -D__LDP_LINUX__ -DMPLS_LINUX   -c -o ldp_netlink.o ldp_netlink.c
> ldp_netlink.c: In function `rtnl_send':
> ldp_netlink.c:103: warning: unused variable `i'
> gcc -I../lib -I. -Wall -g -O2 -D__LDP_LINUX__ -DMPLS_LINUX   -c -o ldp_policy_impl.o ldp_policy_impl.c
> ar rc libldpport.a ldp_fib_impl.o ldp_ifmgr_impl.o ldp_lock_impl.o ldp_mm_impl.o ldp_mpls_impl.o ldp_socket_impl.o ldp_timer_impl.o
> ldp_tree_impl.o ldp_netlink.o ldp_policy_impl.o
> ranlib libldpport.a
> make[1]: Leaving directory `/usr/src/ldp-portable-0.060/port'
> cd lib ; make libldp.a
> make[1]: Entering directory `/usr/src/ldp-portable-0.060/lib'
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_prefix.o ldp_prefix.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_global.o ldp_global.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_entity.o ldp_entity.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_session.o ldp_session.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_if.o ldp_if.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_peer.o ldp_peer.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_adj.o ldp_adj.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_inlabel.o ldp_inlabel.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_outlabel.o ldp_outlabel.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_addr.o ldp_addr.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_cfg.o ldp_cfg.c
> ldp_cfg.c: In function `ldp_cfg_close':
> ldp_cfg.c:31: warning: unused variable `data'
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_range.o ldp_range.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_hello.o ldp_hello.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_pdu.o ldp_pdu.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_pdu_setup.o ldp_pdu_setup.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_msg.o ldp_msg.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_msg_alloc.o ldp_msg_alloc.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_nortel.o ldp_nortel.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_state_machine.o ldp_state_machine.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_inet_addr.o ldp_inet_addr.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_state_funcs.o ldp_state_funcs.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_init.o ldp_init.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_keepalive.o ldp_keepalive.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_fec.o ldp_fec.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_label_request.o ldp_label_request.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_attr.o ldp_attr.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_label_mapping.o ldp_label_mapping.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_label_rel_with.o ldp_label_rel_with.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_notif.o ldp_notif.c
> gcc -I../port -I. -Wall -O2 -g -D__LDP_LINUX__   -c -o ldp_label_abort.o ldp_label_abort.c
> ar src libldp.a ldp_prefix.o ldp_global.o ldp_entity.o ldp_session.o ldp_if.o ldp_peer.o ldp_adj.o ldp_inlabel.o ldp_outlabel.o
> ldp_addr.o ldp_cfg.o ldp_range.o ldp_hello.o ldp_pdu.o ldp_pdu_setup.o ldp_msg.o ldp_msg_alloc.o ldp_nortel.o ldp_state_machine.o
> ldp_inet_addr.o ldp_state_funcs.o ldp_init.o ldp_keepalive.o ldp_fec.o ldp_label_request.o ldp_attr.o ldp_label_mapping.o
> ldp_label_rel_with.o ldp_notif.o ldp_label_abort.o
> make[1]: Leaving directory `/usr/src/ldp-portable-0.060/lib'
> gcc -o ldp_linux port/ldp_linux.o lib/libldp.a port/libldpport.a lib/libldp.a # -lefence
> /usr/i386-slackware-linux/bin/ld: port/ldp_linux.o: Relocations in generic ELF (EM: 2)
> port/ldp_linux.o: could not read symbols: File in wrong format
> collect2: ld returned 1 exit status
> make: *** [ldp_linux] Error 1
> 
> It's a freshly installed slackware 8 with kernel 2.4.12.
> 
> Any suggestions would be most appreciated.
> 
> Also, is there a working tar version of the zebra port?
> 
> Thanks
> 
> Panny Malialis
> 
> 
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu

Re: [mpls-linux-general] iptables

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

Make sure you've turned on all of the netfilter options in the kernel
config.

Jim

On Fri, Mar 29, 2002 at 05:37:40PM -0500, William Crotty wrote:
> 
> I've followed the mini-how-to guide that is in the archive.
> for linux 2.4.17 and mpls-linux 1.128.   Everything
> works fine but when i try and use the iptables command.
> I get this error
> 
> 
> [root@inigo utils]# iptables -A OUTPUT -d 10.10.3.230/32 -j MPLS --set-key
> 0x00000002
> modprobe: Can't locate module ip_tables
> iptables v1.2.4: can't initialize iptables table `filter': Table does not
> exist (do you need to insmod?)
> Perhaps iptables or your kernel needs to be upgraded.
> 
> any ideas.
> 
> 
> thanks
> 
> Bill
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu

[mpls-linux-general] Sorry for the silence ...

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

Hello everyone,

Sorry for the lack of responses recently.  Combination of my day job, vacation,
and preparing for the MPLS Forum interop have been occuping my time.

I will try and work through the bulk of the questions that have posted.  If
I don't respond to you specific e-mail feel free to repost.

Jim
-- 
James R. Leu