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

[mpls-linux-general] Proble in patching the mpls-linux to the latest kernel(2.4.10)

[nitin p. <tac...@ya...>]

Hi James,

I tried to patch the mpls-linux to the new kernel
avaiable at kernel.org(2.4.10).  I got rejection in
if_arp.h.rej; in thie file I am not able to patch MPLS
tunnel Interafce(ARHDR_MPLS_TUNEL).

can you please tell me what should I do to overcome
this problem.

Thanks in advance,

nitin

__________________________________________________
Do You Yahoo!?
NEW from Yahoo! GeoCities - quick and easy web site hosting, just $8.95/month.
http://geocities.yahoo.com/ps/info1

Re: [mpls-linux-general] IPV6 over MPLS

[Vino T. <jv...@te...>]

Hello James,

On Fri, 5 Oct 2001, James R. Leu wrote:
> > On Tue, 2 Oct 2001, James R. Leu wrote:
> > > > My requirement is to connect to isolated IPV6 network over an existing
> > > > IPV4/MPLS network. 
> > > > IPV6 Cloud ---- Linux --- MPLS/IPV4 cloud ----- Linux ---- IPV6 Cloud
> > > > What is the best way out? What type of translation is suitable for this
> > > > scenario (6to4/6over4/6in4 etc)?
> > > ow did you plan to create the LSP across the ipv4 cloud?  Do you need to run
> > > routing across this link?
> > 
> > The two V6 sites can be having their own Site local addresses. So a static
> > route via the Edge Linux box would be good enough.
> > 
> > > For this application I think an MPLS tunnel interface is would do what you need.
> > 
> > Is this kind of support already available in the MPLS code? If so does it
> > require a V6 to V4 translatore as well on the Edge Linux box? If so what
> > translator is best suited for this task.
> 
> Their are two things that are not in place for the egress LER.  Here the
> packets need to come into the MPLS stack and go out via the IPv6 stack.  So
> the layer 3 protocol associated with the incoming label needs to be
> IPv6.  Currently mplsadm doesn't provide a means to say which layer 3
> protocol should be associated with a incoming label (I've assumed IPv4 uptil
> now).  In addition I have not tested the code that rebuild the IPv6 header
> and sends into IPv6 land.  Both are minor fixes and given some time, I
> could complete them in a night.  THe problem is I have no idea how to test
> it :-)
> 
> There are two options: you could send me info as to how to setup ipv6
> and I could try to reproduce your senerio, or I could provide you versions
> to test for me.  Which could you prefer?
> 
> Actually there is a 3rd option which is I tell you where in the code the
> work needs to be done and you fix/test it yourself :-)

Option 2 and Option 3 are ok with me. The 3rd Option looks more juicy
though :-)
I got a setup similar to what I had mentioned in my earlier mail. 

IPV6 Network ---- Linux ----IPV4/MPLS Cloud ---- Linux -----IPV6 Network

The IPV6 Network constitutes of Linux boxes running IPV6 stack.

Regards,
Vino.

> > > > On Sun, 30 Sep 2001, James R. Leu wrote:
> > > > 
> > > > > On Sat, Sep 29, 2001 at 07:14:42PM +0000, Vino Thomas wrote:
> > > > > > Hello,
> > > > > > 
> > > > > > Does Linux has support for IPV6 over MPLS?
> > > > > 
> > > > > Glad you asked (sorry for the delay in responding to you) :-)
> > > > > 
> > > > > #1 If you are trying to LSR activity (only labl swap) the yes :-)
> > > > > #2 If you want to use MPLS tunnel interface, then yes.
> > > > > #3 If you do not want to use MPLS tunnel interfaces and want to have specific
> > > > >   entries in the IPv6 routing table sent traffic on a particular LSP then no,
> > > > >   but it would be too hard to add.
> > > > > 
> > > > > #1 is obvious, label swapping is layer 3 agnostic
> > > > > #2 is a little obscure, and I haven't tested it, but it should work.
> > > > >    -create a MPLS tunnel interface
> > > > >    -ifconfig it up (without an address)
> > > > >    -add a ipv6 route that has a nexthop of the mpls tunnel
> > > > >     (just an outgoing interface no next hop ip)
> > > > > #3 requires that some how ipv6 route entries are marked with which LSP
> > > > >    they should sent traffic on.  Everywhere a ipv6 route lookup is done
> > > > >    then needs to  check for this MPLS data.
> > > > > 
> > > > > #2 should get you up and sending ipv6 data on LSPs
> > > > > #3 will take more work, but needs to be done at some point
> > > > > 
> > > > > I hope this gets the gears moving.  We can go into more details as you
> > > > > ask more questions.
> > > > > 
> > > > > Jim
> > > > > 
> > > > > 
> > > > > > 
> > > > > > Regards,
> > > > > > Vino
> > > > > > 
> > > > > > 
> > > > > > 
> > > > > > 
> > > > > > 
> > > > > > _______________________________________________
> > > > > > mpls-linux-general mailing list
> > > > > > mpl...@li...
> > > > > > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > > > > 
> > > > > -- 
> > > > > James R. Leu
> > > > > 
> > > > > _______________________________________________
> > > > > mpls-linux-general mailing list
> > > > > mpl...@li...
> > > > > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > > > > 
> > > > 
> > > > 
> > > > _______________________________________________
> > > > mpls-linux-general mailing list
> > > > mpl...@li...
> > > > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > > 
> > > -- 
> > > James R. Leu
> > > 
> > > _______________________________________________
> > > mpls-linux-general mailing list
> > > mpl...@li...
> > > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > > 
> > 
> > 
> > _______________________________________________
> > mpls-linux-general mailing list
> > mpl...@li...
> > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> 
> -- 
> James R. Leu
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> 

Re: [mpls-linux-general] IPV6 over MPLS

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

On Fri, Oct 05, 2001 at 10:12:07AM +0000, Vino Thomas wrote:
> Hello,
> 
> On Tue, 2 Oct 2001, James R. Leu wrote:
> > > My requirement is to connect to isolated IPV6 network over an existing
> > > IPV4/MPLS network. 
> > > IPV6 Cloud ---- Linux --- MPLS/IPV4 cloud ----- Linux ---- IPV6 Cloud
> > > What is the best way out? What type of translation is suitable for this
> > > scenario (6to4/6over4/6in4 etc)?
> > ow did you plan to create the LSP across the ipv4 cloud?  Do you need to run
> > routing across this link?
> 
> The two V6 sites can be having their own Site local addresses. So a static
> route via the Edge Linux box would be good enough.
> 
> > For this application I think an MPLS tunnel interface is would do what you need.
> 
> Is this kind of support already available in the MPLS code? If so does it
> require a V6 to V4 translatore as well on the Edge Linux box? If so what
> translator is best suited for this task.

Their are two things that are not in place for the egress LER.  Here the
packets need to come into the MPLS stack and go out via the IPv6 stack.  So
the layer 3 protocol associated with the incoming label needs to be
IPv6.  Currently mplsadm doesn't provide a means to say which layer 3
protocol should be associated with a incoming label (I've assumed IPv4 uptil
now).  In addition I have not tested the code that rebuild the IPv6 header
and sends into IPv6 land.  Both are minor fixes and given some time, I
could complete them in a night.  THe problem is I have no idea how to test
it :-)

There are two options: you could send me info as to how to setup ipv6
and I could try to reproduce your senerio, or I could provide you versions
to test for me.  Which could you prefer?

Actually there is a 3rd option which is I tell you where in the code the
work needs to be done and you fix/test it yourself :-)

Jim

> Regards,
> Vino.
> 
> > > On Sun, 30 Sep 2001, James R. Leu wrote:
> > > 
> > > > On Sat, Sep 29, 2001 at 07:14:42PM +0000, Vino Thomas wrote:
> > > > > Hello,
> > > > > 
> > > > > Does Linux has support for IPV6 over MPLS?
> > > > 
> > > > Glad you asked (sorry for the delay in responding to you) :-)
> > > > 
> > > > #1 If you are trying to LSR activity (only labl swap) the yes :-)
> > > > #2 If you want to use MPLS tunnel interface, then yes.
> > > > #3 If you do not want to use MPLS tunnel interfaces and want to have specific
> > > >   entries in the IPv6 routing table sent traffic on a particular LSP then no,
> > > >   but it would be too hard to add.
> > > > 
> > > > #1 is obvious, label swapping is layer 3 agnostic
> > > > #2 is a little obscure, and I haven't tested it, but it should work.
> > > >    -create a MPLS tunnel interface
> > > >    -ifconfig it up (without an address)
> > > >    -add a ipv6 route that has a nexthop of the mpls tunnel
> > > >     (just an outgoing interface no next hop ip)
> > > > #3 requires that some how ipv6 route entries are marked with which LSP
> > > >    they should sent traffic on.  Everywhere a ipv6 route lookup is done
> > > >    then needs to  check for this MPLS data.
> > > > 
> > > > #2 should get you up and sending ipv6 data on LSPs
> > > > #3 will take more work, but needs to be done at some point
> > > > 
> > > > I hope this gets the gears moving.  We can go into more details as you
> > > > ask more questions.
> > > > 
> > > > Jim
> > > > 
> > > > 
> > > > > 
> > > > > Regards,
> > > > > Vino
> > > > > 
> > > > > 
> > > > > 
> > > > > 
> > > > > 
> > > > > _______________________________________________
> > > > > mpls-linux-general mailing list
> > > > > mpl...@li...
> > > > > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > > > 
> > > > -- 
> > > > James R. Leu
> > > > 
> > > > _______________________________________________
> > > > mpls-linux-general mailing list
> > > > mpl...@li...
> > > > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > > > 
> > > 
> > > 
> > > _______________________________________________
> > > mpls-linux-general mailing list
> > > mpl...@li...
> > > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > 
> > -- 
> > James R. Leu
> > 
> > _______________________________________________
> > mpls-linux-general mailing list
> > mpl...@li...
> > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > 
> 
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu

Re: [mpls-linux-general] IPV6 over MPLS

[Vino T. <jv...@te...>]

Hello,

On Tue, 2 Oct 2001, James R. Leu wrote:
> > My requirement is to connect to isolated IPV6 network over an existing
> > IPV4/MPLS network. 
> > IPV6 Cloud ---- Linux --- MPLS/IPV4 cloud ----- Linux ---- IPV6 Cloud
> > What is the best way out? What type of translation is suitable for this
> > scenario (6to4/6over4/6in4 etc)?
> ow did you plan to create the LSP across the ipv4 cloud?  Do you need to run
> routing across this link?

The two V6 sites can be having their own Site local addresses. So a static
route via the Edge Linux box would be good enough.

> For this application I think an MPLS tunnel interface is would do what you need.

Is this kind of support already available in the MPLS code? If so does it
require a V6 to V4 translatore as well on the Edge Linux box? If so what
translator is best suited for this task.

Regards,
Vino.

> > On Sun, 30 Sep 2001, James R. Leu wrote:
> > 
> > > On Sat, Sep 29, 2001 at 07:14:42PM +0000, Vino Thomas wrote:
> > > > Hello,
> > > > 
> > > > Does Linux has support for IPV6 over MPLS?
> > > 
> > > Glad you asked (sorry for the delay in responding to you) :-)
> > > 
> > > #1 If you are trying to LSR activity (only labl swap) the yes :-)
> > > #2 If you want to use MPLS tunnel interface, then yes.
> > > #3 If you do not want to use MPLS tunnel interfaces and want to have specific
> > >   entries in the IPv6 routing table sent traffic on a particular LSP then no,
> > >   but it would be too hard to add.
> > > 
> > > #1 is obvious, label swapping is layer 3 agnostic
> > > #2 is a little obscure, and I haven't tested it, but it should work.
> > >    -create a MPLS tunnel interface
> > >    -ifconfig it up (without an address)
> > >    -add a ipv6 route that has a nexthop of the mpls tunnel
> > >     (just an outgoing interface no next hop ip)
> > > #3 requires that some how ipv6 route entries are marked with which LSP
> > >    they should sent traffic on.  Everywhere a ipv6 route lookup is done
> > >    then needs to  check for this MPLS data.
> > > 
> > > #2 should get you up and sending ipv6 data on LSPs
> > > #3 will take more work, but needs to be done at some point
> > > 
> > > I hope this gets the gears moving.  We can go into more details as you
> > > ask more questions.
> > > 
> > > Jim
> > > 
> > > 
> > > > 
> > > > Regards,
> > > > Vino
> > > > 
> > > > 
> > > > 
> > > > 
> > > > 
> > > > _______________________________________________
> > > > mpls-linux-general mailing list
> > > > mpl...@li...
> > > > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > > 
> > > -- 
> > > James R. Leu
> > > 
> > > _______________________________________________
> > > mpls-linux-general mailing list
> > > mpl...@li...
> > > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > > 
> > 
> > 
> > _______________________________________________
> > mpls-linux-general mailing list
> > mpl...@li...
> > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> 
> -- 
> James R. Leu
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> 

Re: [mpls-linux-general] IPV6 over MPLS

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

On Mon, Oct 01, 2001 at 05:17:47PM +0000, Vino Thomas wrote:
> Hello,
> 
> My requirement is to connect to isolated IPV6 network over an existing
> IPV4/MPLS network. 
> 
> IPV6 Cloud ---- Linux --- MPLS/IPV4 cloud ----- Linux ---- IPV6 Cloud
> 
> What is the best way out? What type of translation is suitable for this
> scenario (6to4/6over4/6in4 etc)?

ow did you plan to create the LSP across the ipv4 cloud?  Do you need to run
routing across this link?

For this application I think an MPLS tunnel interface is would do what you need.

> Regards,
> Vino.
> 
> On Sun, 30 Sep 2001, James R. Leu wrote:
> 
> > On Sat, Sep 29, 2001 at 07:14:42PM +0000, Vino Thomas wrote:
> > > Hello,
> > > 
> > > Does Linux has support for IPV6 over MPLS?
> > 
> > Glad you asked (sorry for the delay in responding to you) :-)
> > 
> > #1 If you are trying to LSR activity (only labl swap) the yes :-)
> > #2 If you want to use MPLS tunnel interface, then yes.
> > #3 If you do not want to use MPLS tunnel interfaces and want to have specific
> >   entries in the IPv6 routing table sent traffic on a particular LSP then no,
> >   but it would be too hard to add.
> > 
> > #1 is obvious, label swapping is layer 3 agnostic
> > #2 is a little obscure, and I haven't tested it, but it should work.
> >    -create a MPLS tunnel interface
> >    -ifconfig it up (without an address)
> >    -add a ipv6 route that has a nexthop of the mpls tunnel
> >     (just an outgoing interface no next hop ip)
> > #3 requires that some how ipv6 route entries are marked with which LSP
> >    they should sent traffic on.  Everywhere a ipv6 route lookup is done
> >    then needs to  check for this MPLS data.
> > 
> > #2 should get you up and sending ipv6 data on LSPs
> > #3 will take more work, but needs to be done at some point
> > 
> > I hope this gets the gears moving.  We can go into more details as you
> > ask more questions.
> > 
> > Jim
> > 
> > 
> > > 
> > > Regards,
> > > Vino
> > > 
> > > 
> > > 
> > > 
> > > 
> > > _______________________________________________
> > > mpls-linux-general mailing list
> > > mpl...@li...
> > > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > 
> > -- 
> > James R. Leu
> > 
> > _______________________________________________
> > mpls-linux-general mailing list
> > mpl...@li...
> > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > 
> 
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu

RE: [mpls-linux-general] (no subject)

[Sanjeev C. <Sa...@co...>]

Hi Jens,

Also check your kernel version. It should be at least 2.4.5

Sanjeev

-----Original Message-----
From: Jens Kuehlberg [mailto:jen...@gm...]
Sent: Monday, October 01, 2001 5:58 AM
To: mpl...@li...
Subject: [mpls-linux-general] (no subject)


Hallo,

I have Red Hat Linux 7.1 and the linux-mpls (v0.993) Patch.

Did  Linux-mpls Patch support dynamic label distribution between PC? Or I
must install fixed label for MPLS?

Bye
Jens

-- 
GMX - Die Kommunikationsplattform im Internet.
http://www.gmx.net


_______________________________________________
mpls-linux-general mailing list
mpl...@li...
https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

Re: [mpls-linux-general] (no subject)

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

On Mon, Oct 01, 2001 at 02:58:04PM +0200, Jens Kuehlberg wrote:
> Hallo,
> 
> I have Red Hat Linux 7.1 and the linux-mpls (v0.993) Patch.
> 
> Did  Linux-mpls Patch support dynamic label distribution between PC? Or I
> must install fixed label for MPLS?

You can use either staic (mplsadm) or dynamic (via ldp-portable).

If you are going to use ldp-portable, make sure to get that latest from
CVS.

http://sourceforge.net/cvs/?group_id=15443

> 
> Bye
> Jens
> 
> -- 
> GMX - Die Kommunikationsplattform im Internet.
> http://www.gmx.net
> 
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu

[mpls-linux-general] (no subject)

[Jens K. <jen...@gm...>]

Hallo,

I have Red Hat Linux 7.1 and the linux-mpls (v0.993) Patch.

Did  Linux-mpls Patch support dynamic label distribution between PC? Or I
must install fixed label for MPLS?

Bye
Jens

-- 
GMX - Die Kommunikationsplattform im Internet.
http://www.gmx.net

Re: [mpls-linux-general] IPV6 over MPLS

[Vino T. <jv...@te...>]

Hello,

My requirement is to connect to isolated IPV6 network over an existing
IPV4/MPLS network. 

IPV6 Cloud ---- Linux --- MPLS/IPV4 cloud ----- Linux ---- IPV6 Cloud

What is the best way out? What type of translation is suitable for this
scenario (6to4/6over4/6in4 etc)?

Regards,
Vino.

On Sun, 30 Sep 2001, James R. Leu wrote:

> On Sat, Sep 29, 2001 at 07:14:42PM +0000, Vino Thomas wrote:
> > Hello,
> > 
> > Does Linux has support for IPV6 over MPLS?
> 
> Glad you asked (sorry for the delay in responding to you) :-)
> 
> #1 If you are trying to LSR activity (only labl swap) the yes :-)
> #2 If you want to use MPLS tunnel interface, then yes.
> #3 If you do not want to use MPLS tunnel interfaces and want to have specific
>   entries in the IPv6 routing table sent traffic on a particular LSP then no,
>   but it would be too hard to add.
> 
> #1 is obvious, label swapping is layer 3 agnostic
> #2 is a little obscure, and I haven't tested it, but it should work.
>    -create a MPLS tunnel interface
>    -ifconfig it up (without an address)
>    -add a ipv6 route that has a nexthop of the mpls tunnel
>     (just an outgoing interface no next hop ip)
> #3 requires that some how ipv6 route entries are marked with which LSP
>    they should sent traffic on.  Everywhere a ipv6 route lookup is done
>    then needs to  check for this MPLS data.
> 
> #2 should get you up and sending ipv6 data on LSPs
> #3 will take more work, but needs to be done at some point
> 
> I hope this gets the gears moving.  We can go into more details as you
> ask more questions.
> 
> Jim
> 
> 
> > 
> > Regards,
> > Vino
> > 
> > 
> > 
> > 
> > 
> > _______________________________________________
> > mpls-linux-general mailing list
> > mpl...@li...
> > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> 
> -- 
> James R. Leu
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> 

Re: [mpls-linux-general] IPV6 over MPLS

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

On Sat, Sep 29, 2001 at 07:14:42PM +0000, Vino Thomas wrote:
> Hello,
> 
> Does Linux has support for IPV6 over MPLS?

Glad you asked (sorry for the delay in responding to you) :-)

#1 If you are trying to LSR activity (only labl swap) the yes :-)
#2 If you want to use MPLS tunnel interface, then yes.
#3 If you do not want to use MPLS tunnel interfaces and want to have specific
  entries in the IPv6 routing table sent traffic on a particular LSP then no,
  but it would be too hard to add.

#1 is obvious, label swapping is layer 3 agnostic
#2 is a little obscure, and I haven't tested it, but it should work.
   -create a MPLS tunnel interface
   -ifconfig it up (without an address)
   -add a ipv6 route that has a nexthop of the mpls tunnel
    (just an outgoing interface no next hop ip)
#3 requires that some how ipv6 route entries are marked with which LSP
   they should sent traffic on.  Everywhere a ipv6 route lookup is done
   then needs to  check for this MPLS data.

#2 should get you up and sending ipv6 data on LSPs
#3 will take more work, but needs to be done at some point

I hope this gets the gears moving.  We can go into more details as you
ask more questions.

Jim


> 
> Regards,
> Vino
> 
> 
> 
> 
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu

[mpls-linux-general] IPV6 over MPLS

[Vino T. <jv...@te...>]

Hello,

Does Linux has support for IPV6 over MPLS?

Regards,
Vino

[mpls-linux-general] New version of MPLS-linux+iptables

[Olivier D. - FT.BD/FTRDugeonD/DAC/C. <Oli...@rd...>]

Hi all,

We just release a new version including full TC support :-) and more 
userfriendly packet mark. Here it is the ChangeLog

CahngeLog:
----------
09-28-2001 : Olivier Dugeon <Oli...@rd...>
              Alexandre Dagan <Ale...@wa...>

- fwmark and fwmpls maark must be set to the label and not the entire
key comming from /proc/net/mpls_out. More user friendly :-)
- Clean up mpls_output code
- Compute the RadiX hash key inside rt_set_nexthop from the label
- Add tc mpls classifier like fw classifier to assign class based on
the skb->fwmpls value
- Add the mpls protocol to tc from Steven Van den Berghe mpls-tc patch
- Reorganize the source with full and small directory

The patch is in attachment and i put it into sourceforge under the 
patches page.

Let us know if you have some pb. with it.

Olivier & Alexandre
-- 
  FTR&D/DAC/CPN
  Technopole Anticipa     | mailto:Oli...@fr...
  2, Avenue Pierre Marzin | Phone:  +(33) 2 96 05 28 80
  F-22307 LANNION         | Fax:    +(33) 2 96 05 18 52

[mpls-linux-general] IPV6 over MPLS (fwd)

[Vino T. <jv...@te...>]

Hello,

Forwarding to the MPLS list.

---------- Forwarded message ----------
Date: Thu, 27 Sep 2001 16:17:31 -0500
From: James R. Leu <jl...@mi...>
To: Vino Thomas <jv...@te...>
Subject: Re: IPV6 over MPLS

I suggest posting this to the mpls-linux mailing list, and I will respond
to it there, so that others can help out if they wish and also so the question
and it's response are archived.

http://lists.sourceforge.net/lists/listinfo/mpls-linux-general

On Thu, Sep 27, 2001 at 10:52:44AM +0000, Vino Thomas wrote:
> Hello James,
> 
> 
> On Thu, 20 Sep 2001, James R. Leu wrote:
> > > Does Linux has support for IPV6 over MPLS?
> > Right now there are only minimal hooks from the IPv6 stack to the MPLS
> > stack.  No testing has been done on it.  If you are interested in working on 
> > it I could point you in the right direction.
> 
> Yes I am.
> That would be great. Could you pass me the pointer?
> 
> Regards,
> Vino

-- 
James R. Leu

Re: [mpls-linux-general] VRF for Linux

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

On Thu, Sep 27, 2001 at 11:30:43AM -0500, Nick Eggleston wrote:
> As part of Cisco's MPLS implementation, they introduced a feature called
> VPN Routing/Forwarding (or Virtual Route Forwarding) referred to as VRFs
> [or maybe this is an RFC definition].
> 
> What work is being done for Linux to support this concept?
> In particular, making sure that ARP caches and FIBs can be VRFed.

Linux has the notion of multiple routing tables.  The table identifier can
be used as part of a route lookup.  Once a route has been choosen and
a next hop it determined, it will use the neighbor cache to to find
the next hop on that outgoing interface.

Currently, there is not the notion of an interface belonging to a table AFAIK.
Due to the way the arp cache works I do not think that you would need
an arp cache per table.

In addition I don't think there is a way for a socket to say which table
it wants to send it's packet to for a route lookup nor does it have a way
to limit which incoming interfaces can connect to a listening socket.

> In addition, is the VRF-aware versions of gated or zebra?

I do not believe so.

The details of how to add his support to linux are very intriging.
I wonder if anyone is working on this type of support yet ... maybe I should
start a VRF for Linux project ;-)

Jim

> 
> How about VRF-tagging of sockets or processes, to perform the same
> functions as the cisco commands:
> 
> telnet x.x.x.x /vrf name
> ping vrf name x.x.x.x
> traceroute vrf name x.x.x.x
> 
> Per-VRF nat, filtering, etc??
> 
> Can anyone help?
> 
> Thanks!
> 
> --Nick
> 
> ---
> 
> 
> VRF- Virtual Route Forwarding. Initially, a router has only one global
> default routing/forwarding table. VRFs can be viewed as multiple disjoined
> routing/forwarding tables, where the routes of a user have no correlation
> with the routes of another user.
> Source: http://www.cisco.com/univercd/cc/td/doc/product/software/ios122/122newft/122limit/122x/122dx/122dx1/dd_vrfaa.htm
> 
> 
> A VRF defines the VPN membership of a customer site attached to a PE
> router. A VRF consists of an IP routing table, a derived Cisco Express
> Forwarding (CEF) table, a set of interfaces that use the forwarding table,
> and a set of rules and routing protocol parameters that control the
> information that is included into the routing table
> Source: http://www.cisco.com/univercd/cc/td/doc/product/software/ios120/120newft/120t/120t5/vpn.htm
> 
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu

[mpls-linux-general] VRF for Linux

[Nick E. <ni...@dc...>]

As part of Cisco's MPLS implementation, they introduced a feature called
VPN Routing/Forwarding (or Virtual Route Forwarding) referred to as VRFs
[or maybe this is an RFC definition].

What work is being done for Linux to support this concept?
In particular, making sure that ARP caches and FIBs can be VRFed.

In addition, is the VRF-aware versions of gated or zebra?

How about VRF-tagging of sockets or processes, to perform the same
functions as the cisco commands:

telnet x.x.x.x /vrf name
ping vrf name x.x.x.x
traceroute vrf name x.x.x.x

Per-VRF nat, filtering, etc??

Can anyone help?

Thanks!

--Nick

---


VRF- Virtual Route Forwarding. Initially, a router has only one global
default routing/forwarding table. VRFs can be viewed as multiple disjoined
routing/forwarding tables, where the routes of a user have no correlation
with the routes of another user.
Source: http://www.cisco.com/univercd/cc/td/doc/product/software/ios122/122newft/122limit/122x/122dx/122dx1/dd_vrfaa.htm


A VRF defines the VPN membership of a customer site attached to a PE
router. A VRF consists of an IP routing table, a derived Cisco Express
Forwarding (CEF) table, a set of interfaces that use the forwarding table,
and a set of rules and routing protocol parameters that control the
information that is included into the routing table
Source: http://www.cisco.com/univercd/cc/td/doc/product/software/ios120/120newft/120t/120t5/vpn.htm

[mpls-linux-general] Re: Route Alert Option

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

As far as I know, no one uses the "router alert" label.  It was originally
added to the MPLS encoding spec as a "just in case" type of measure.

What is the correct behavior for a IP packet with the router alert bit
set?  How does the IP stack know "who" to locally deliver it to?  If you
can answer these questions then your further along the path to understanding
what the router alert label should do  then I am :-)

On Tue, Sep 25, 2001 at 06:29:18PM +0530, Rajesh P Jain wrote:
> "Hi",
>                 I have 2 questions on IP Route Alert Options (and , Of
> course corresponding to the MPLS Route Alert Label (LABEL Value 1)).
> 
> 1.    When a  LSR receives a Labelled packet with a value of 1, what is
> the treatment given to that labelled packet. (In the diagram LSR).
> 
> 2.     When an Ingress Node (In the Diagram LER A), when an IP Packet
> comes with IP Route Alert Option, How is the MPLS labels that packet ??
> I mean, does it label it with a value of 1 or does it look into the LIB
> for the corresponding destination address of the packet ???
> 
>                         Please throw light on these questions and
> clarify how does source-forge linux MPLS forwarding Engine (JRL's
> Implementation) treats the above 2 cases.
> 
> LER A --------- LSR ------------- LER B
> 
> TIA
>         Rajesh P Jain

-- 
James R. Leu

Re: [mpls-linux-general] FTP Core Dumps

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

On Tue, Sep 25, 2001 at 02:50:34PM +0530, Rajesh P Jain wrote:
> "Hi",
>                         I setup an LSP between LER A and LER B
> successfully and vice versa.
> 
>                         I ping from LER A to LER B. The ping works
> successfully and I get the Ping reply successfully at LER A.
> 
> 
>                         Now, with the same setup, if I do a FTP from LER
> A to LER B, the linux kernel core dumps.

Do you get an "opps trace" in the log file or any trace messages sent to
the console?

I've been trying to track down this path MTU related failure for
some time now ....

Jim
-- 
James R. Leu

[mpls-linux-general] Route Alert Option

[Rajesh P J. <rp...@sa...>]

"Hi",
                I have 2 questions on IP Route Alert Options (and , Of
course corresponding to the MPLS Route Alert Label (LABEL Value 1)).

1.    When a  LSR receives a Labelled packet with a value of 1, what is
the treatment given to that labelled packet. (In the diagram LSR).

2.     When an Ingress Node (In the Diagram LER A), when an IP Packet
comes with IP Route Alert Option, How is the MPLS labels that packet ??
I mean, does it label it with a value of 1 or does it look into the LIB
for the corresponding destination address of the packet ???

                        Please throw light on these questions and
clarify how does source-forge linux MPLS forwarding Engine (JRL's
Implementation) treats the above 2 cases.

LER A --------- LSR ------------- LER B

TIA
        Rajesh P Jain

[mpls-linux-general] FTP Core Dumps

[Rajesh P J. <rp...@sa...>]

"Hi",
                        I setup an LSP between LER A and LER B
successfully and vice versa.

                        I ping from LER A to LER B. The ping works
successfully and I get the Ping reply successfully at LER A.


                        Now, with the same setup, if I do a FTP from LER
A to LER B, the linux kernel core dumps.

                        Any help on this will be appreciated.

LER A----------LSR------------LERB

With Regards
                    Rajesh P Jain

Re: [mpls-linux-general] Ping Problem over MPLS

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

I need more information before I can help you.

Give me a diagram that has the IP addresses and then give me dumps from:
'route'
'more /proc/net/mpls_*'
'ifconfig'

for each box.

Jim

On Fri, Sep 21, 2001 at 12:04:49PM +0530, Adithya Bhat wrote:
> Hi All,
> 
> After setting up the network in the following manner,
> 
> HOSTA-----LERA-------LSR------LERB------HOSTB
> 
> and setting the LSP in both the directions (i.e. from HostA to HostB
> and Vice Versa ) , I am able to ping from Host A to Host B tru  the LSP.
> 
> (Host A & Host B are not MPLS Enabled)
> 
> But when try to ping from LERA to LERB, ping fails....
> 
> Please Note that I have :
> A route to LER-B with the Gateway as the LSR at LER-A.
> A route to LER-A with the Gateway as the LSR at LER-B.
> And no routes at the LSR, as the Forwarding should be done by MPLS (
> Label Swapping ).
> 
> Could any body tell me what am I missing here ??
> 
> Thanx in Advance
> 
> -Adithya
> 
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu

[mpls-linux-general] Ping Problem over MPLS

[Adithya B. <ad...@sa...>]

Hi All,

After setting up the network in the following manner,

HOSTA-----LERA-------LSR------LERB------HOSTB

and setting the LSP in both the directions (i.e. from HostA to HostB
and Vice Versa ) , I am able to ping from Host A to Host B tru  the LSP.

(Host A & Host B are not MPLS Enabled)

But when try to ping from LERA to LERB, ping fails....

Please Note that I have :
A route to LER-B with the Gateway as the LSR at LER-A.
A route to LER-A with the Gateway as the LSR at LER-B.
And no routes at the LSR, as the Forwarding should be done by MPLS (
Label Swapping ).

Could any body tell me what am I missing here ??

Thanx in Advance

-Adithya

Re: [mpls-linux-general] Difference between ISI RSVP and KOM RSVP

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

nitin panjwani wrote:

> HI James and ISBN guys ,

You can call me Pim :)
(it is IBCN by the way)

> I have two questions:
>
> 1)What is the difference between ISI RSVP and KOM
>   RSVP in : functionality  , implementation and other
>   features that affects the TE enhancement .

(please note that there is more involved then just the original
ISI work, see our web site for a brief explanation
http://dsmpls.atlantis.rug.ac.be)
I will only talk about our work:

* Support for DiffServ over MPLS
    * signaling through RSVP daemon (E-LSP only at the moment)
    * kernel support thanks to James and modified by Steven en Tom to
support DiffServ
* Explicitly routed LSPs
* DiffServ configuration scripts
* Mapping traffic based on protocol, destination, DSCP on LSP/EXP (based
on iptables, ip rule, multiple routing tables = extendable)

Not supported but will be very soon:
re-routing (already fixed a lot of problems in my tree)

Not supported, needs some work :
Intserv reservations for LSPs, L-LSP setup

TE signaling requiring more work:
resource affinity procedures, HELLO procedures

> 2)ISBN guys how many days did it take for you to
>   enhance the RSVP fetures in ISI RSVP ?

What we did is:
* merging BSD and Linux versions (NIST)
* add DS/MPLS support to James' MPLS (E and L LSPs)
* introduce empty TSPEC and FLOWSPEC (signaling, API, CLI) to support
E-LSPs
* introduced CoS TSPEC and FLOWSPEC (needs some more work to reflect to
current IETF draft, needed to support L-LSPs)
* integration, testing, bug fixing, documentation, web site ...
(I again refer to the web site for more information)

All in all we have been working on this for more then a year I guess.

> What kind of
>   network did you use for the implementatin?

12 node Linux network with 4 ifaces, running zebra OSPF

> and what
>   are the difficulties you faced initially for this
>   enhancement?

RSVP daemon: huge code base, some strange and legacy code (but it has a
lot features which makes implementing some non-trivial stuff almost
trivial and it is pretty well documented)
Linux integration: lot's of different grounds to cover (iptables,
rt-netlink, MPLS kernel code, ip rules, tc ...)

Pim.

>  Hope to hear some nice replies ,
>
>  Thanks in advance,
>  Nitin
>
> __________________________________________________
> Terrorist Attacks on U.S. - How can you help?
> Donate cash, emergency relief information
> http://dailynews.yahoo.com/fc/US/Emergency_Information/
>
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

--
Pim Van Heuven
Ghent University - IMEC
Department of Information Technology (INTEC)
Sint-Pietersnieuwstraat, 41
B-9000, Gent, Belgium
tel.: +32-(0)9267 3592, secr.: +32-(0)9267 3570
fax : +32-(0)9267 3599
e-mail : pim...@in...
WWW :  http://www.intec.rug.ac.be

[mpls-linux-general] regarding configuring the peer

[L S. R. <ls...@ya...>]

Dear all,
	we have setup of three LSRs
lsr1(eth1)---------(eth0)lsr2(eth1)---------(eth1)lsr3 we want to send targeted
hello from lsr1 to lsr3 ,so we are adding peer at lsr1 as 

******************************************
{	lsr1-eth1- 	172.16.2.1
	lsr2-eth0-	172.16.2.2
	lsr2-eth1-	172.16.3.2
	lsr3-eth1-	172.16.3.1
}
*****************************************
at lsr1:

>>add peer 172.16.3.1

then we are getting Unknown Entity in lsr3,one thing i want to make sure that
whether this adding of peer is meant for not directly connected (different
networks) hosts or not? if yes then how can lsr3 compares the 172.16.2.x with
172.16.3.x because both are different networks it is giving Unknown entity
problem ,if we connect the  lsr1 and lsr2 they are working properly.(without
giving Unknown entity problem).

please give me some suggestion
thanks in adv..
regards
lsr

[mpls-linux-general] Difference between ISI RSVP and KOM RSVP

[nitin p. <tac...@ya...>]

HI James and ISBN guys ,

I have two questions:

1)What is the difference between ISI RSVP and KOM 
  RSVP in : functionality  , implementation and other 
  features that affects the TE enhancement .

2)ISBN guys how many days did it take for you to 
  enhance the RSVP fetures in ISI RSVP ? What kind of 
  network did you use for the implementatin? and what 
  are the difficulties you faced initially for this 
  enhancement?

 Hope to hear some nice replies ,

 Thanks in advance,
 Nitin

__________________________________________________
Terrorist Attacks on U.S. - How can you help?
Donate cash, emergency relief information
http://dailynews.yahoo.com/fc/US/Emergency_Information/

Re: [mpls-linux-general] Help

[Abhijit G. <gab...@ee...>]

>        How should I extract routing table information from kernel
>through the net link socket.
>Please help me in this.

Look at the iproute2 sources.

--a