mpls-linux-devel — Mailig list for discussing the development of MPLS on Linux

Re: [mpls-linux-devel] linux-kernel-2.6.14

[Bob B. <bob...@gm...>]

Hi James,

Thanks for the p4 hints.

I have made some progress in your dev tree.
I have built and booted into an mpls enabled kernel.

But when I try to run through the MPLS-Linux example, here
 are the results:

bash-3.00# cat /sys/mpls/version
1.950
bash-3.00# cat /sys/mpls/debug
1
bash-3.00# /usr/sbin/mpls nhlfe add key 0
NHLFE entry key 0x00000002 mtu 0 propagate_ttl
         (0 bytes, 0 pkts)
bash-3.00# /usr/sbin/mpls nhlfe change key 0x2 instructions push gen
10000 nexthop eth0 ipv4 172.16.87.21
bash-3.00# /usr/sbin/ip route add 172.16.87.21/32 via 172.16.87.21
spec_nh 0x8847 0x2
Error: either "to" is duplicate, or "spec_nh" is a garbage.
bash-3.00#


I did recompile the iproute2 files after rebooting into my new kernel.
I did edit the Makefile, ran make and make install (as root).
What did I miss? How do I troubleshoot?

Thanks,

--
-Bob

Re: [mpls-linux-devel] linux-kernel-2.6.14

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

On Fri, Dec 02, 2005 at 09:05:50AM -0500, Bob Beers wrote:
<snip>
> BTW, is there a shortcut for 'p4 sync'?  When I do that, it gets
> _everything_ again,
>  doesn't seem to get just changed files -- (takes a _lo-o-o-o-ong_ time).

Your client view has a great impact on this.

What I do is have multiple clients each with a very specific view.
That way when sync'ing a client p4 only has to consider a small number
of files.  Also if you've been syncing recently there have been large
changes in the tree, merging that latest linux rev to all sorts of
branches.

BTW you can see all of the clients by doing a p4 clients (note 's' at the e=
nd).
Do 'p4 clients | grep jleu' and you'll see how many clients I have created.
Doing a 'p4 client -o <client-name>' and you can see what the details for
that client are.  For example doing a:

p4 client -o jleu:mpls-kernel-3

shows the details of my HOL mpls-kernel tree, and show the view I'm using:

        //depot/mpls-kernel-1.1/... //jleu:mpls-kernel-3/...

Which maps just the mpls-kernel-1.1 directory to my client.  If you
didn't modify your view, then your view looks like:

        //depot/... //jleu:mpls-kernel-3/...

Which means every directory in the dev tree is mapped into your client ...
which is a bad idea :-)

I see you have a client (actually 3, but they all look like they resuse
the same directory).  If I were you I would move all of the views execpt
for the mpls-kernel-1.1 to a new client.  Leave the mpls-kernel-1.1 line
so you don't have to re-sync it.  My guess is that the size of the
mpls-kernel-1.1 directory is what is making it so slow, that and the fact
my p4 server lives off of a DSL line with an upstream of 400Kbps.

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

Re: [mpls-linux-devel] linux-kernel-2.6.14

[Bob B. <bob...@gm...>]

On 12/1/05, James R. Leu <jl...@mi...> wrote:
> Since you're using the mpls-kernel from my dev tree, you will
> need to use the iproute-mpls from my dev tree as well.
> (ie the patch in mpls-linux is out-dated)

ok, let me see if I understand my options:

1) a)use your dev tree for kernel 2.6.15-rc1-mpls
              (all MPLS in kernel -- no modules)
    b)use dev tree for iproute2 and mpls is the binary tool

2) a)use 2.6.9 vanilla kernel, apply linux-kernel.diff from 1.946 tarball
              (all MPLS in kernel -- no modules)
    b)use iproute2 package from ???, apply iproute2.diff from 1.946 tarball

I'll willing to try both/either.

BTW, is there a shortcut for 'p4 sync'?  When I do that, it gets
_everything_ again,
 doesn't seem to get just changed files -- (takes a _lo-o-o-o-ong_ time).

--
Thanks,

-Bob

Re: [mpls-linux-devel] linux-kernel-2.6.14

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

Since you're using the mpls-kernel from my dev tree, you will
need to use the iproute-mpls from my dev tree as well.
(ie the patch in mpls-linux is out-dated)

mplsadm2 is the name of the old utility to add MPLS
entities.  That functionality is not handled by the 'mpls' utility in
iproute-mpls.

I'll update the README to remove references to mplsadm2

On Thu, Dec 01, 2005 at 03:47:36PM -0500, Bob Beers wrote:
> On 12/1/05, James R. Leu <jl...@mi...> wrote:
> > For now don't compile MPLS as a module.  I'm still working out module
> > loading issues.
> >
>=20
> OK, all MPLS built-in, no modules.
>=20
> but ... trying the example in mpls-linux-patches.README.iproute2
>=20
> iproute2/mpls# ./mpls nhlfe add key 0
> Cannot talk to rtnetlink: Connection refused
>=20
> Did I leave out a module?
>=20
> Also, I'm confused about the mplsadm2 command mentioned in
> patches/README.examples? Where do I find that tool?
>=20
> --
> -Bob

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

Re: [mpls-linux-devel] linux-kernel-2.6.14

[Bob B. <bob...@gm...>]

On 12/1/05, James R. Leu <jl...@mi...> wrote:
> For now don't compile MPLS as a module.  I'm still working out module
> loading issues.
>

OK, all MPLS built-in, no modules.

but ... trying the example in mpls-linux-patches.README.iproute2

iproute2/mpls# ./mpls nhlfe add key 0
Cannot talk to rtnetlink: Connection refused

Did I leave out a module?

Also, I'm confused about the mplsadm2 command mentioned in
patches/README.examples? Where do I find that tool?

--
-Bob

Re: [mpls-linux-devel] linux-kernel-2.6.14

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

For now don't compile MPLS as a module.  I'm still working out module
loading issues.

On Thu, Dec 01, 2005 at 12:29:43PM -0500, Bob Beers wrote:
> On 11/23/05, James R. Leu <jl...@mi...> wrote:
> > At this time your quickest path is to revert back to 2.6.9 and
> > apply the 1.946 patch.  Although I'm getting dangerously close to
> > making a new release ..  (probably next week after the US thanksgiving
> > holiday).
> >
>=20
> Ok, trying two paths at the same time.  I am trying this, but make fails =
with
>  a few warnings:
>=20
> net/mpls/mpls_netlink.o(.init.text+0x0): In function `init_module':
> : multiple definition of `init_module'
> net/mpls/mpls_init.o(.init.text+0x0): first defined here
> ...
>=20
> and similar complaint about `cleanup_module'.
>=20
> Clues for me?
>=20
> --
> -Bob

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

Re: [mpls-linux-devel] linux-kernel-2.6.14

[Bob B. <bob...@gm...>]

On 11/23/05, James R. Leu <jl...@mi...> wrote:
> At this time your quickest path is to revert back to 2.6.9 and
> apply the 1.946 patch.  Although I'm getting dangerously close to
> making a new release ..  (probably next week after the US thanksgiving
> holiday).
>

Ok, trying two paths at the same time.  I am trying this, but make fails wi=
th
 a few warnings:

net/mpls/mpls_netlink.o(.init.text+0x0): In function `init_module':
: multiple definition of `init_module'
net/mpls/mpls_init.o(.init.text+0x0): first defined here
...

and similar complaint about `cleanup_module'.

Clues for me?

--
-Bob

Re: [mpls-linux-devel] linux-kernel-2.6.14

[Bob B. <bob...@gm...>]

On 11/23/05, James R. Leu <jl...@mi...> wrote:
> At this time your quickest path is to revert back to 2.6.9 and
> apply the 1.946 patch.  Although I'm getting dangerously close to
> making a new release ..  (probably next week after the US thanksgiving
> holiday).
>

2.6.9 is just a bit too old for some of the other kernel things I need,
 but if I am completely stumped I may try it. :P

> If you want to forge ahead, update you linux-kernel tree and make sure
> to pull out a new iproute2-mpls tree
> (View: //depot/iproute2-mpls/... //<new-client-name>/...)

//depot/iproute2-mpls-1.1/ ??

Thanks,

-Bob

Re: [mpls-linux-devel] linux-kernel-2.6.14

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

At this time your quickest path is to revert back to 2.6.9 and
apply the 1.946 patch.  Although I'm getting dangerously close to=20
making a new release ..  (probably next week after the US thanksgiving
holiday).

If you want to forge ahead, update you linux-kernel tree and make sure
to pull out a new iproute2-mpls tree
(View: //depot/iproute2-mpls/... //<new-client-name>/...)

As of last night I was able to use the 'mpls' utility to create static
LSPs and bind routes to them.

If you're really looking for something bleeding edge, grab my ebtables-mpls
tree and you can do "ethernet over MPLS" (the support was there many
release ago, in a hacked up sort of way, now using ebtables I have a much
better implementation)

On Tue, Nov 22, 2005 at 04:26:37PM -0500, Bob Beers wrote:
> On 11/10/05, James R. Leu <jl...@mi...> wrote:
> > My development tree is currently patched against 2.6.13 so getting
> > mpls-linux to work on a 2.6.11 (and beyond) kernel is possible.
> >
> >
> I got your development tree as of 20051121, and applied
>  patch-2.6.15-rc1-rc2 since I noticed you were at -rc1.
>=20
> But I get a compile error:
>=20
>   CC [M]  net/ipv4/mpls4.o
> net/ipv4/mpls4.c: In function `mpls4_local_deliver':
> net/ipv4/mpls4.c:298: error: too few arguments to function `ip_rcv'
> make[2]: *** [net/ipv4/mpls4.o] Error 1
> make[1]: *** [net/ipv4] Error 2
> make: *** [net] Error 2
>=20
> and, of course it's true: ip_rvc() expects 4 arguments.
> net/ipv4/ip_input.c:
>      int ip_rcv(struct sk_buff *skb,
>                     struct net_device *dev,
>                     struct packet_type *pt,
>                     struct net_device *orig_dev)
>=20
> but mpls4_local_deliver() provides only 3:
> net/ipv4/mpls4.c:
>        return ip_rcv(skb, skb->dev, NULL);
>=20
> Should I go just drop back to the 1.946 tarball and apply to say, 2.6.10?
> What is likely to get me to a working MPLS enabled 2.6.x kernel most quic=
kly?
>=20
> Thanks,
>=20
> -Bob

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

Re: [mpls-linux-devel] linux-kernel-2.6.14

[Bob B. <bob...@gm...>]

On 11/10/05, James R. Leu <jl...@mi...> wrote:
> My development tree is currently patched against 2.6.13 so getting
> mpls-linux to work on a 2.6.11 (and beyond) kernel is possible.
>
>
I got your development tree as of 20051121, and applied
 patch-2.6.15-rc1-rc2 since I noticed you were at -rc1.

But I get a compile error:

  CC [M]  net/ipv4/mpls4.o
net/ipv4/mpls4.c: In function `mpls4_local_deliver':
net/ipv4/mpls4.c:298: error: too few arguments to function `ip_rcv'
make[2]: *** [net/ipv4/mpls4.o] Error 1
make[1]: *** [net/ipv4] Error 2
make: *** [net] Error 2

and, of course it's true: ip_rvc() expects 4 arguments.
net/ipv4/ip_input.c:
     int ip_rcv(struct sk_buff *skb,
                    struct net_device *dev,
                    struct packet_type *pt,
                    struct net_device *orig_dev)

but mpls4_local_deliver() provides only 3:
net/ipv4/mpls4.c:
       return ip_rcv(skb, skb->dev, NULL);

Should I go just drop back to the 1.946 tarball and apply to say, 2.6.10?
What is likely to get me to a working MPLS enabled 2.6.x kernel most quickl=
y?

Thanks,

-Bob

[mpls-linux-devel] the problem of create ldp

[yaya <ya...@ru...>]
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Re: [mpls-linux-devel] LSP Ping

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

As far as I know nothing can from those discussions.  I would be
happy to help someone who is willing to work on it.

On Thu, Nov 17, 2005 at 07:21:58AM -0800, Timothy Vann wrote:
> I saw an earlier post where someone was asking about how to do LSP ping.
> Just wondering if anyone knows if any work has been done toward that?  And
> if so, where can I find the code?
>=20
> =20
>=20
> Thanks,
>=20
> Tim
>=20

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

[mpls-linux-devel] LSP Ping

[Timothy V. <tv...@ex...>]

I saw an earlier post where someone was asking about how to do LSP ping.
Just wondering if anyone knows if any work has been done toward that?  And
if so, where can I find the code?

 

Thanks,

Tim

Re: [mpls-linux-devel] linux-kernel-2.6.14

[Bob B. <bob...@gm...>]

> mpls-linux to work on a 2.6.11 (and beyond) kernel is possible.
> My guess is you are seeing the issue with 'sock_no_release'
> being removed.  Someone has already posted patches for this, in fact
> they might be available from http://sf.net/projects/mpls-linux under the
> patch section.

great, I'll look there.


> that 1.950 is almost ready for release.  It adds some significant
> features, as well as fixes a race conditions with respect to IPv4 route c=
ache
> and NHLFE deletes.

also good, I'll keep an eye out for that release.

--
-Bob

Re: [mpls-linux-devel] linux-kernel-2.6.14

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

My development tree is currently patched against 2.6.13 so getting
mpls-linux to work on a 2.6.11 (and beyond) kernel is possible.
My guess is you are seeing the issue with 'sock_no_release'
being removed.  Someone has already posted patches for this, in fact
they might be available from http://sf.net/projects/mpls-linux under the
patch section.

May I ask what your working on?  Also you might be interested to know
that 1.950 is almost ready for release.  It adds some significant
features, as well as fixes a race conditions with respect to IPv4 route cac=
he
and NHLFE deletes.

On Thu, Nov 10, 2005 at 10:27:52AM -0500, Bob Beers wrote:
> Hi list,
>=20
> I want to intergrate mpls-linux into my device, and would like to
>  use kernel 2.6.14.  To that end I have modified the patches/linux-kernel=
.diff
>  to get a successful patch  (a diff for the diff file is attached).
>=20
> Is anyone else working on this?  I am getting compile errors, and can
>  probably work through them, but, if it is already done somewhere I could
>  save a lot of time.
>=20
> Are there any reasons not to move forward with my plans?
>=20
> Thanks,
>=20
> -Bob



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

[mpls-linux-devel] linux-kernel-2.6.14

[Bob B. <bob...@gm...>]

Hi list,

I want to intergrate mpls-linux into my device, and would like to
 use kernel 2.6.14.  To that end I have modified the patches/linux-kernel.d=
iff
 to get a successful patch  (a diff for the diff file is attached).

Is anyone else working on this?  I am getting compile errors, and can
 probably work through them, but, if it is already done somewhere I could
 save a lot of time.

Are there any reasons not to move forward with my plans?

Thanks,

-Bob

[mpls-linux-devel] [PATCH] RFC: updated net-shim

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

Those of you following along at home remember I requested feedback about
the viability of my last net-shim patch.  Well, I didn't hear a peep.
I've updated it anyway.  Here is a list of the changes:

-shim protocols no longer need to register a protocol type, just a name.
-shim protocols are stored in a a plain linked list, no need for the hashing
-shim data now has a datalen field and data is a zero len array. The user
 must allocate enought room.
-removed struct rtshim from struct rtnexthop, the shim info is added to
 multi-nexthop routes via RTA_PUT (just like RTA_GATEWAY).  This is
 significant, now an unchanged iproute2 will still be able to add
 ipv4 multi-nexthop routes
-minimal documentation for all of the shim functions.

Things that still need to be resolved
-move proc file system code to a seperate file (maybe discard it?)
-figure out how to add shim support to IPv6 without making changes to
 struct in6_rtmsg, especially since rtshim is variable size ....
-migrate to 2.6.13
-make sure net-shim works with fib_trie.c

I would still like to hear some feedback, Jamal are you out there?

Patch attached.

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

[mpls-linux-devel] RFC: Generic network shim infrastructure

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

In hopes to get this MPLS effort 'unstuck from the mud' I've decided to
start the path of getting some of the infrastructure needed for MPLS
reviewed by the powers that be.  I hope DaveM and Jamal are still reading
this list, because they are probably the first hurdle I need to cross.

This patch introduces the network shim infrastructure.  It is designed
for protocols that live below layer 3 but above layer 2 in the OSI model.
This is exactly where MPLS lives.  I believe this infrastructure could also
be used by IPSEC for route base VPNs without using tunnels.

The guts of the patch are the 'shim live cycle', but it also
adds a new 'family' to netlink, RTM_*SHIM, and adds some the necessary
hooks to ipv4 and ipv6.

<aside>
I have tried using the XFRM infrastructure, but the problem I continue
to run into is it's apparent lack of ties to routing entries.  For example
if I want traffic destined to 10.0.0.0/8 to go through a specific XFRM
I can do that, but if at the same time I don't want 10.0.1.0/24 to use 
any XFRM, this is where I am unable to get XFRM to do my bidding. 
I tried adding per-route XFRMs but it looks like the dst stacking is in
the wrong order for what I need .... I digress
</aside>

Please look this over and let me know if there are any glaring
kernel no-no's.  I still have commenting and documentation to do,
but the actual infrastructure is complete.  Of course I've implemented
MPLS to use this, but we'll tackle that after we get the needed infrastructure
into the kernel.

Patch attached

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

RE: [mpls-linux-devel] Quagga MPLS vtysh partially solved

[Joan R. <jr...@sa...>]

=20
I've succeeded compiling quagga-mpls with vtysh support with an uggly =
trick,
but at least it works:

The libreadline version installed by default on my Fedora Core 3 box was
libreadline.so.4.3.

I think that in order to compile newest versions of vtysh the =
libreadline5
version is needed. I've found it hard to upgrade libreadline to a newest
version so what I've finally done is to copy some files from a Fedora =
Core 4
box:

I've copied libreadline5* and libhistory* within the /usr/lib/ directory
and also the contents of /usr/include/readline/ to my FC3 box.

Then i've ran ./configure --enable-user=3Droot --enable-group=3Droot
--enable-ospf-te --enable-vtysh and this time i've succeeded.

I'm sure there must be cleanest ways but this is the only one i've found =
to
work.

Anybody knows how to do it easier?

Thanx in advance.

Joan Ruiz.

=20

-----Original Message-----
From: mpl...@li...
[mailto:mpl...@li...] On Behalf Of Joan =
Ruiz
Sent: martes, 19 de julio de 2005 13:49
To: 'mpls-linux-devel'
Subject: RE: [mpls-linux-devel] Quagga MPLS vtysh

 Hello everybody.

I'm trying to compile the latest Perforce version of quagga-mpls with =
vtysh
support and I find the following error when running the ./configure =
script :


./configure --enable-user=3Droot --enable-group=3Droot --enable-ospf-te
--enable-vtysh ...output ommited...
configure: error: vtysh needs libreadline but was not found and usable =
on
your system.

I've never tried to compile vtysh before, and i've been unable to find =
the
required libreadline library. Any idea?


I'm using a standard Fedora Core 3 environment with mpls support (Linux
version 2.6.9-1.6_FC2mpls_1_946a (jleu@jleu-laptop) (gcc version 3.3.3
20040412 (Red Hat Linux 3.3.3-7)) #1 Wed Jan 26 23:12:20 CST 2005)

Joan Ruiz
jr...@sa...


=20

-----Original Message-----
From: mpl...@li...
[mailto:mpl...@li...] On Behalf Of =
pramod
Sent: s=E1bado, 16 de julio de 2005 11:00
To: mpls-linux-devel
Subject: [mpls-linux-devel] Quagga MPLS

hi,
   I saw from the mpls-linux site that  BGP of Quagga supports mpls =
label
carrying information....
I have some doubts regarding ..
Does it carry stack of labels or only a single label How is the CE-PE =
and
PE-CE transition handled...( using Route-maps or any other way )

thanks
 pramod


-------------------------------------------------------
SF.Net email is sponsored by: Discover Easy Linux Migration Strategies =
from
IBM. Find simple to follow Roadmaps, straightforward articles, =
informative
Webcasts and more! Get everything you need to get up to speed, fast.
http://ads.osdn.com/?ad_id=3D7477&alloc_id=3D16492&op=3Dclick
_______________________________________________
mpls-linux-devel mailing list
mpl...@li...
https://lists.sourceforge.net/lists/listinfo/mpls-linux-devel



-------------------------------------------------------
SF.Net email is sponsored by: Discover Easy Linux Migration Strategies =
from
IBM. Find simple to follow Roadmaps, straightforward articles, =
informative
Webcasts and more! Get everything you need to get up to speed, fast.
http://ads.osdn.com/?ad_idt77&alloc_id=16492&op=CCk
_______________________________________________
mpls-linux-devel mailing list
mpl...@li...
https://lists.sourceforge.net/lists/listinfo/mpls-linux-devel

RE: [mpls-linux-devel] Quagga MPLS vtysh

[Joan R. <jr...@sa...>]

 Hello everybody.

I'm trying to compile the latest Perforce version of quagga-mpls with =
vtysh
support and I find the following error when running the ./configure =
script :


./configure --enable-user=3Droot --enable-group=3Droot --enable-ospf-te
--enable-vtysh
...output ommited...
configure: error: vtysh needs libreadline but was not found and usable =
on
your system.

I've never tried to compile vtysh before, and i've been unable to find =
the
required libreadline library. Any idea?


I'm using a standard Fedora Core 3 environment with mpls support (Linux
version 2.6.9-1.6_FC2mpls_1_946a (jleu@jleu-laptop) (gcc version 3.3.3
20040412 (Red Hat Linux 3.3.3-7)) #1 Wed Jan 26 23:12:20 CST 2005)

Joan Ruiz
jr...@sa...


=20

-----Original Message-----
From: mpl...@li...
[mailto:mpl...@li...] On Behalf Of =
pramod
Sent: s=E1bado, 16 de julio de 2005 11:00
To: mpls-linux-devel
Subject: [mpls-linux-devel] Quagga MPLS

hi,
   I saw from the mpls-linux site that  BGP of Quagga supports mpls =
label
carrying information....
I have some doubts regarding ..
Does it carry stack of labels or only a single label How is the CE-PE =
and
PE-CE transition handled...( using Route-maps or any other way )

thanks
 pramod


-------------------------------------------------------
SF.Net email is sponsored by: Discover Easy Linux Migration Strategies =
from
IBM. Find simple to follow Roadmaps, straightforward articles, =
informative
Webcasts and more! Get everything you need to get up to speed, fast.
http://ads.osdn.com/?ad_id=3D7477&alloc_id=3D16492&op=3Dclick
_______________________________________________
mpls-linux-devel mailing list
mpl...@li...
https://lists.sourceforge.net/lists/listinfo/mpls-linux-devel

Re: [mpls-linux-devel] Quagga MPLS

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

On Sat, Jul 16, 2005 at 02:29:49PM +0530, pramod wrote:
> hi,
>   I saw from the mpls-linux site that  BGP of Quagga supports mpls=20
> label carrying information....
> I have some doubts regarding ..
> Does it carry stack of labels or only a single label
> How is the CE-PE and PE-CE transition handled...( using Route-maps or=20
> any other way )

I do not know the details of quagga's BGP implementation of MPLS labels.
I have not gotten around to hooking BGP into my MPLS framework.

One thing to note, just because BGP carries MPLS labels does not mean
is has all of the components needed for RFC2547bis VPNs.

> thanks
> pramod
>=20
>=20
> -------------------------------------------------------
> SF.Net email is sponsored by: Discover Easy Linux Migration Strategies
> from IBM. Find simple to follow Roadmaps, straightforward articles,
> informative Webcasts and more! Get everything you need to get up to
> speed, fast. http://ads.osdn.com/?ad_id=3D7477&alloc_id=3D16492&op=3Dclick
> _______________________________________________
> mpls-linux-devel mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-devel

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

[mpls-linux-devel] Quagga MPLS

[pramod <pr...@at...>]

hi,
   I saw from the mpls-linux site that  BGP of Quagga supports mpls 
label carrying information....
I have some doubts regarding ..
Does it carry stack of labels or only a single label
How is the CE-PE and PE-CE transition handled...( using Route-maps or 
any other way )

thanks
 pramod

Re: [mpls-linux-devel] [PATCH] Vtysh and "mpls ip" command

[Hasso T. <ha...@es...>]

James R. Leu wrote:
> On Thu, Jun 23, 2005 at 08:56:05PM +0300, Hasso Tepper wrote:
> > As I prefer always to use vtysh, I made the "mpls ip" command work under
> > it. LDP_IF_NODE commands should work as well now. This rises the question
> > though.
> 
> Thank you very much for taking the time to fix this.  I have tried
> multiple time to figure it out why it didn't work.  This is a huge
> help for my testing.  I have applied that patch to my tree.

NP. This is part of Quagga I know really well. That's how I started hacking
on it ;).

> > Why is ldp_interface_config_write() if 0'd out? Not yet ready code? Or
> > commented out because you have some better ideas regarding cli structure
> > of this part, James? Can I help with something regarding this?
> 
> When I re-implemented ldpd using the MPLS framework I if 0'd out
> all of the code that I had some doubts on.  Basicly trimming the amount
> of stuff I had to get working.  Now that things are stabalizing
> with respect to ldpd, I will probably go back and start enabling all
> of the 'bells and whistles'.

OK. Then logic will stay. Then some further fixes are needed to handle
configuration in LDP_IF_NODE. I'm going to fix this, just one comment.
I looked at this commented out ldp_interface_config_write() and if I'm
not mistaken you indented LDP_IF_NODE further from INTERFACE_NODE. I
don't think that it's very good idea. Although I agree that making it
somewhat separate is very good idea. I'd recommend doing it like family
nodes in bgp configuration. Ie it will look like something like this in
vtysh:

interface eth0
 description ldp interface
 ip address 192.168.1.1/24
 ip ospf authentication message-digest
 ip ospf message-digest-key 1 md5 foo
!
 mpls ip
 distribution-mode dod
 keepalive-interval 10
!
interface eth1
 description ...
...

What do you think?

-- 
Hasso Tepper
Elion Enterprises Ltd.
WAN administrator

Re: [mpls-linux-devel] [PATCH] Vtysh and "mpls ip" command

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

On Thu, Jun 23, 2005 at 08:56:05PM +0300, Hasso Tepper wrote:
> As I prefer always to use vtysh, I made the "mpls ip" command work under
> it. LDP_IF_NODE commands should work as well now. This rises the question
> though.

Thank you very much for taking the time to fix this.  I have tried
multiple time to figure it out why it didn't work.  This is a huge
help for my testing.  I have applied that patch to my tree.

> Why is ldp_interface_config_write() if 0'd out? Not yet ready code? Or
> commented out because you have some better ideas regarding cli structure
> of this part, James? Can I help with something regarding this?

When I re-implemented ldpd using the MPLS framework I if 0'd out
all of the code that I had some doubts on.  Basicly trimming the amount
of stuff I had to get working.  Now that things are stabalizing
with respect to ldpd, I will probably go back and start enabling all
of the 'bells and whistles'.

> --=20
> Hasso Tepper
> Elion Enterprises Ltd.
> WAN administrator



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

[mpls-linux-devel] [PATCH] Vtysh and "mpls ip" command

[Hasso T. <ha...@es...>]

As I prefer always to use vtysh, I made the "mpls ip" command work under
it. LDP_IF_NODE commands should work as well now. This rises the question
though.

Why is ldp_interface_config_write() if 0'd out? Not yet ready code? Or
commented out because you have some better ideas regarding cli structure
of this part, James? Can I help with something regarding this?


-- 
Hasso Tepper
Elion Enterprises Ltd.
WAN administrator