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

Re: [mpls-linux-devel] Status of MPLS in the p4 repository

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

On Tue, Jun 07, 2005 at 12:01:18PM +0200, Vincent Untz wrote:
> Hi,
>=20
> I'm trying to use the 2.6.11 version of the MPLS-enabled kernel,
> available in James' p4 repository. However, enabling IP_MPLS makes UML
> crash. It seems it's because of the MPLS_IPV4_EXPLICIT_NULL label
> addition when MPLS is initialized. The other option is to use IP_MPLS as
> a module (I haven't tried yet).

Stay away from compiling any part of MPLS linux as a module, not much
work has been done testing that.

> Does anyone have an idea why?

Honestly I haven't built a 2.6.11 MPLS enabled kernel.  I integrated
the 2.6.11 changes on top of a working tree, so I just assumed that
it would result in a working 2.6.11 kernel ... I guess I was wrong.
I will try to build a 2.6.11 MPLS enabled RPM and do some testing
myself.


>=20
> Vincent
>=20
> --=20
> Les gens heureux ne sont pas press??s.



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

[mpls-linux-devel] Status of MPLS in the p4 repository

[Vincent U. <vin...@im...>]

Hi,

I'm trying to use the 2.6.11 version of the MPLS-enabled kernel,
available in James' p4 repository. However, enabling IP_MPLS makes UML
crash. It seems it's because of the MPLS_IPV4_EXPLICIT_NULL label
addition when MPLS is initialized. The other option is to use IP_MPLS as
a module (I haven't tried yet).

Does anyone have an idea why?

Vincent

--=20
Les gens heureux ne sont pas press=C3=A9s.

Re: [mpls-linux-devel] [patch] Fix wrong error code when creating an ILM/NHLFE that already exists

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

Applied, thank you.

On Tue, May 24, 2005 at 04:56:49PM +0200, Vincent Untz wrote:
> Hi,
>=20
> Currently, the kernel returns -ESRCH when the user tries to create an
> ILM that already exists. It should really be -EEXIST :-)
>=20
> See attached patch.
>=20
> Vincent
>=20
> --=20
> Les gens heureux ne sont pas press??s.





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

Re: [mpls-linux-devel] [patch] Fix crash in mpls command

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

Applied, thank you.

On Tue, May 24, 2005 at 04:40:36PM +0200, Vincent Untz wrote:
> Hi,
>=20
> The attached patch fix a crash when you do "mpls ilm" (or similar). I
> chose to make it list the created ILM in this case, just like "ip route"
> shows the routes.
>=20
> Cheers,
>=20
> Vincent
>=20
> --=20
> Les gens heureux ne sont pas press??s.





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

[mpls-linux-devel] [patch] Fix wrong error code when creating an ILM/NHLFE that already exists

[Vincent U. <vin...@im...>]

Hi,

Currently, the kernel returns -ESRCH when the user tries to create an
ILM that already exists. It should really be -EEXIST :-)

See attached patch.

Vincent

-- 
Les gens heureux ne sont pas pressés.

[mpls-linux-devel] [patch] Fix crash in mpls command

[Vincent U. <vin...@im...>]

Hi,

The attached patch fix a crash when you do "mpls ilm" (or similar). I
chose to make it list the created ILM in this case, just like "ip route"
shows the routes.

Cheers,

Vincent

-- 
Les gens heureux ne sont pas pressés.

[mpls-linux-devel] Updates to ldp-portable

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

I have finished submitting some significant changes to ldp-portable and
quagga-mpls.  It still is not ready for users, but it handles interface
and route events without crashing... which is good.  So far all my work
has concentrated on stabalizing ldpd when no sessions have been formed.
My next step will be to fixup session establishment.

If you are working with quagga-mpls or ldp-portable I suggest getting
an updated version from my P4 tree.  I will work on making a proper
release of quagga-mpls and ldp-portable in the coming days.

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

[mpls-linux-devel] Diff from 1.935 to 1.938

[jrussi <jr...@uo...>]

Hi,

I would like to know why from version 1.935 to 1.938 (and
1.946 now) you started to use a "htons" in function "parse"
of lipipt_spec_nh.c, at line 62:

spec_nh_info->proto =3D htons((unsigned short)temp);


I tried to find it at P4 server, but I'm not familiarized
with it.

Could you please answer me? It has to be with the
iptables-save bug.

Thank you,

Josdeyvi=0A =0A__________________________________________________________=
________________=0AAcabe com aquelas janelinhas que pulam na sua tela.=0A=
AntiPop-up UOL - =C9 gr=E1tis!=0Ahttp://antipopup.uol.com.br/=0A

[mpls-linux-devel] Backup path mpls

[Miquel C. P. <u10...@co...>]

Hello, I am Student of ETIS from Spain

I create a experimental plataform of routing with 6 computers. In my
experiment I have one path under mpls and run correctly. But, can I create
the alternative path for when the primary path failed.


Your see my esqueme:

		(primary path is host_2, host_4)

              host_2 -------- host_4
	      |			   |
  (source)    |      	           |        (destin)
   host_1 ----|			   |----host_6
	      |			   |
	      |			   |
	      host_3 -------- host_5

			(the alternetive path is host_3, and host_4)


How to detect that path is failed? There is something program I can use
for monitoring the link between hosts and that soported mpls?

And, when the link failed, there is any signal that indicate this state.
Is implemented the CR_LDP and RSVP_TE signal in mpls_linux?

I can find information envelope mpls and backup path?


thank you!!!!!

Re: [mpls-linux-devel] Nexthop eth ip

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

On Tue, Apr 19, 2005 at 08:48:50AM +0200, jul...@aq... wrote:
> Hi,
>=20
> When you add a route
>=20
> ip route ********* nexthop eth0 10.0.2.1
>=20
> you indicate the ip of the nexhop so that we can get the mac adresse.
>=20
> the ip of the nexthop is newer used ?
>=20
> is it possible to indicate only the mac adresse ?

Are you refering to the case where you are adding a route that points
to an NHLFE?

    ip route add 1.1.1.0/24 via 10.0.2.1 spec_nh 0x8847 0x2

You are correct that the nexthop ip is never using by MPLS, but it is used
by IPv4.  IPv4 needs to know what interface this traffic will leave
so that locally originated traffic can choose an appropriate source address.
You could also specify just the outgoing interface:

    ip route add 1.1.1.0/24 dev eth1 spec_nh 0x8847 0x2

But you will still need to have an IP address attached to eth1.  This
is a requirement of IPv4, not MPLS.

>=20
>=20
> Thanks,
>=20
>=20
>=20
>=20
>=20
> --=20
> Julien CAYSSOL
> INSA de LYON                  -=3D-      D?partement Telecom Services et =
Usages
> http://www.aqwz.com/                                    T?l : +33 664 973=
 067
>=20
>=20
> -------------------------------------------------------
> This SF.Net email is sponsored by: New Crystal Reports XI.
> Version 11 adds new functionality designed to reduce time involved in
> creating, integrating, and deploying reporting solutions. Free runtime in=
fo,
> new features, or free trial, at: http://www.businessobjects.com/devxi/728
> _______________________________________________
> mpls-linux-devel mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-devel

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

Re: [mpls-linux-devel] Problem with netlink use by MPLS

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

Good job finding the root cause. I will gladly accept a patch that fixes th=
is.
When implementing you patch you may want to look at the some of the
discussions on netdev about netlink handling other address families.  I'm
not sure if code was submitted for this, but if so it would be in the
2.6.11+ time frame.

Let me know if I can answer any questions.

On Tue, Apr 19, 2005 at 11:38:59AM +0200, Vincent Untz wrote:
> Hi,
>=20
> I reported a crash in this thread:
> http://sourceforge.net/mailarchive/forum.php?thread_id=3D7071062&forum_id=
=3D5051
>=20
> After some debugging, we think we found the problem. Here's what's
> happening:
>=20
>  * tcpdump/ethereal/etc. call getifaddrs() to get the the list of
>    interfaces
>  * in glibc >=3D 2.3.3, getifaddrs() uses netlink to find the
>    informations. It opens a netlink socket with the NETLINK_ROUTE family
>    and send RTM_GETLINK and RTM_GETADDR requests for every protocol
>    (AF_UNSPEC)
>  * MPLS uses the NETLINK_ROUTE family and overrides the RTM_* types with
>    its own types. For example, RTM_GETADDR is overrided by
>    MPLS_RTM_GETILM.
>  * when the kernel receives the RTM_GETADDR requests for every protocol,
>    it uses the rtnelink_links[] table to find the table for each
>    protocol that determines what function is called. For MPLS, this
>    table is mpls_rtnetlink_table. Thus, when a RTM_GETADDR request for
>    every protocol is sent, the MPLS_RTM_GETILM callback is called and
>    returns some data.
>  * as this data is not expected (it's not a link and it's not an
>    adresse), the libc abort().
>=20
> There are two problems here :
>  * MPLS overrides the meaning of RTM_* when it shouldn't. I don't think
>    this is easily fixable since the number of RTM_* types is limited and
>    there is not enough space left for MPLS.
>  * MPLS uses NETLINK_ROUTE. It should really use a new netlink family,
>    for example NETLINK_MPLS. I don't think this will be hard to do. It
>    probably involves copying some code from net/core/rtnetlink.c in
>    net/mpls/mpls_netlink.c
>=20
> We'll try to make a patch to add NETLINK_MPLS.
>=20
> Vincent
>=20
> --=20
> Les gens heureux ne sont pas press??s.



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

[mpls-linux-devel] Problem with netlink use by MPLS

[Vincent U. <vin...@im...>]

Hi,

I reported a crash in this thread:
http://sourceforge.net/mailarchive/forum.php?thread_id=3D7071062&forum_id=
=3D5051

After some debugging, we think we found the problem. Here's what's
happening:

 * tcpdump/ethereal/etc. call getifaddrs() to get the the list of
   interfaces
 * in glibc >=3D 2.3.3, getifaddrs() uses netlink to find the
   informations. It opens a netlink socket with the NETLINK_ROUTE family
   and send RTM_GETLINK and RTM_GETADDR requests for every protocol
   (AF_UNSPEC)
 * MPLS uses the NETLINK_ROUTE family and overrides the RTM_* types with
   its own types. For example, RTM_GETADDR is overrided by
   MPLS_RTM_GETILM.
 * when the kernel receives the RTM_GETADDR requests for every protocol,
   it uses the rtnelink_links[] table to find the table for each
   protocol that determines what function is called. For MPLS, this
   table is mpls_rtnetlink_table. Thus, when a RTM_GETADDR request for
   every protocol is sent, the MPLS_RTM_GETILM callback is called and
   returns some data.
 * as this data is not expected (it's not a link and it's not an
   adresse), the libc abort().

There are two problems here :
 * MPLS overrides the meaning of RTM_* when it shouldn't. I don't think
   this is easily fixable since the number of RTM_* types is limited and
   there is not enough space left for MPLS.
 * MPLS uses NETLINK_ROUTE. It should really use a new netlink family,
   for example NETLINK_MPLS. I don't think this will be hard to do. It
   probably involves copying some code from net/core/rtnetlink.c in
   net/mpls/mpls_netlink.c

We'll try to make a patch to add NETLINK_MPLS.

Vincent

--=20
Les gens heureux ne sont pas press=C3=A9s.

[mpls-linux-devel] Nexthop eth ip

[<jul...@aq...>]

Hi,

When you add a route

ip route ********* nexthop eth0 10.0.2.1

you indicate the ip of the nexhop so that we can get the mac adresse.

the ip of the nexthop is newer used ?

is it possible to indicate only the mac adresse ?


Thanks,





-- 
Julien CAYSSOL
INSA de LYON                  -=-      Département Telecom Services et Usages
http://www.aqwz.com/                                    Tèl : +33 664 973 067

[mpls-linux-devel] compilation quagga : errors

[<ble...@in...>]

Hi,

I posted several others posts but I found the solutions.
Now, I'm trying to compile quagga 0.97-3 with mpls.
During the make, I have  :

In file included from /usr/include/linux/mpls.h:32,
                 from ../zebra/mpls_lib.h:27,
                 from ../zebra/mpls_vty.h:27,
                 from ../zebra/zserv.h:26,
                 from zclient.c:36:
/usr/include/linux/if.h:95: error: redefinition of `struct ifmap'
/usr/include/linux/if.h:131: error: redefinition of `struct ifreq'
/usr/include/linux/if.h:181: error: redefinition of `struct ifconf'
make[2]: *** [zclient.lo] Error 1
make[2]: Leaving directory `/usr/src/quagga-0.97.3/lib'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/usr/src/quagga-0.97.3'
make: *** [all] Error 2


Can somebody help me to solve this error ?

Thanks

BL



----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.

[mpls-linux-devel] Compilation of iproute2 problems

[<ble...@in...>]

Hi everybody,

I'm working in mpls with 2.6.9 this time !
I installed my kernel successful.
But, when I try to install iproute2, I have an error :

In file included from /usr/include/asm/byteorder.h:57,
                 from ss.c:36:
/usr/include/linux/byteorder/little_endian.h:43: error: parse error before
"__cpu_to_le64p"
/usr/include/linux/byteorder/little_endian.h: In function `__cpu_to_le64p':
/usr/include/linux/byteorder/little_endian.h:45: error: `__le64' undeclared
(first use in this function)
/usr/include/linux/byteorder/little_endian.h:45: error: (Each undeclared
identifier is reported only once
/usr/include/linux/byteorder/little_endian.h:45: error: for each function it
appears in.)
/usr/include/linux/byteorder/little_endian.h: At top level:
/usr/include/linux/byteorder/little_endian.h:47: error: parse error before '*'
token
/usr/include/linux/byteorder/little_endian.h: In function `__le64_to_cpup':
/usr/include/linux/byteorder/little_endian.h:49: error: `p' undeclared (first
use in this function)
/usr/include/linux/byteorder/little_endian.h: At top level:
/usr/include/linux/byteorder/little_endian.h:51: error: parse error before
"__cpu_to_le32p"
/usr/include/linux/byteorder/little_endian.h: In function `__cpu_to_le32p':
/usr/include/linux/byteorder/little_endian.h:53: error: `__le32' undeclared
(first use in this function)
/usr/include/linux/byteorder/little_endian.h: At top level:
/usr/include/linux/byteorder/little_endian.h:55: error: parse error before '*'
token


Thanks for you quick help

Bertrand



----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.

[mpls-linux-devel] iproute2 error compilation

[<ble...@in...>]

Hi everybody,

I'm working in mpls with 2.6.9 this time !
I installed my kernel successful.
But, when I try to install iproute2, I have an error :

In file included from /usr/include/asm/byteorder.h:57,
                 from ss.c:36:
/usr/include/linux/byteorder/little_endian.h:43: error: parse error before
"__cpu_to_le64p"
/usr/include/linux/byteorder/little_endian.h: In function `__cpu_to_le64p':
/usr/include/linux/byteorder/little_endian.h:45: error: `__le64' undeclared
(first use in this function)
/usr/include/linux/byteorder/little_endian.h:45: error: (Each undeclared
identifier is reported only once
/usr/include/linux/byteorder/little_endian.h:45: error: for each function it
appears in.)
/usr/include/linux/byteorder/little_endian.h: At top level:
/usr/include/linux/byteorder/little_endian.h:47: error: parse error before '*'
token
/usr/include/linux/byteorder/little_endian.h: In function `__le64_to_cpup':
/usr/include/linux/byteorder/little_endian.h:49: error: `p' undeclared (first
use in this function)
/usr/include/linux/byteorder/little_endian.h: At top level:
/usr/include/linux/byteorder/little_endian.h:51: error: parse error before
"__cpu_to_le32p"
/usr/include/linux/byteorder/little_endian.h: In function `__cpu_to_le32p':
/usr/include/linux/byteorder/little_endian.h:53: error: `__le32' undeclared
(first use in this function)
/usr/include/linux/byteorder/little_endian.h: At top level:
/usr/include/linux/byteorder/little_endian.h:55: error: parse error before '*'
token


Thanks for you quick help

Bertrand



----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.

[mpls-linux-devel] compilation problem : some ideas !! thanks

[<ble...@in...>]

Hi,

I'm compiling linux-mpls and I have some trouble :

CC      net/mpls/af_mpls.o
net/mpls/af_mpls.c:40: error: `sock_no_release' undeclared here (not in a
function)
net/mpls/af_mpls.c:40: error: initializer element is not constant
net/mpls/af_mpls.c:40: error: (near initialization for `mpls_sk_ops.release')
make[2]: *** [net/mpls/af_mpls.o] Error 1
make[1]: *** [net/mpls] Error 2
make: *** [net] Error 2


I saw that somebody already had such an error with 2.6.11.
I'm installing linux-mpls on a 2.6.10 because my firm obliged me to work with
this kind of kernel.
Can you tell me how to update mpls for kernel 2.6.10 ?

Moreover, have you an idea on how to test the mpls-linux distribution without
copmipiling the kernel ? (make is very long on my computer...)

Thanks for answers because it's very important for me and my firm to install
mpls.

Thanks

----------------------------------------------------------------
This message was sent using IMP, the Internet Messaging Program.

[mpls-linux-devel] help me to install mpls-linux in Fedora Core 2

[amit p. <am...@my...>]

hi all !
        Will you please help me , how to provide MPLS support in Linux ?
         I am using Fedora Core 2.6.5-1.358. I downloaded the latest files =
under project mpls-linux.
I also downloaded ldp-portable files since I want Linux kernel to support f=
or Tag distribution. 
 
   The following tar files I am am trying to install in Fedora Core 2 (2.6.=
5-1.335).
   I downloaded with src code since I need to know the MPLS code:
       1. mpls-linux-1.946a.tar.bz2
       2. iproute2-2.6.9-3.tar.bz2
       3. iptables-1.2.11-3.1.FC3.tar.bz2
       4. quagga-0.98.3.tar 

  I really donot know how to go to build MPLS enabled Linux Kernel ?

  I am a naive in this field.  Will you please suggest how to go for it ? A=
lso , have I downloaded the correct versions of the Tar file for Fedora C=
ore 2 .

Rgds
Amit 

 

[mpls-linux-devel] (no subject)

[vittorio_manetti <vit...@li...>]

Hi, 
I`m going to graduate in engineering at Federico II of Naples. I`m =
working on a project for my graduation thesis 
concerning mpls-linux. I`=
d like to know if within the last version of mpls-linux (1.946) there is =
a routine which 
allows an association among a FEC and several NHLFE acc=
es keys, when several route meet the request for QoS 
concerning a FEC s=
pecification. I have already looked at the mailing list, but it doesn't s=
eem to me to have found 
information concerning such problem. I wold be =
very pleased to be in contact with the person who is working with 
that,=
 whether this funtionally isn`t already active, but in the course of the =
realization. 
 
Thank you a lot. 
 =0A=0A=0A=0A_______________________=
_____________________________________=0ANavighi a 4 MEGA e i primi 3 mesi=
 sono GRATIS. =0AScegli Libero Adsl Flat senza limiti su http://www.liber=
o.it=0A

Re: [mpls-linux-devel] Quagga-mpls

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

On Fri, Mar 25, 2005 at 05:30:06PM +0200, Hasso Tepper wrote:
> Shame that my mail client doesn't have (yet?) feature to check whether I=
=20
> attached correct patch. Therefore I have to send now new one with tiny=20
> improvement - this one compiles.

:-)

Thank you!  Your patch has been applies as changes 1120 and 1121.

> with my best wishes,
>=20
> --=20
> Hasso Tepper
> Elion Enterprises Ltd.
> WAN administrator



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

[mpls-linux-devel] Re: your mail

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

Here is a step-by-step process on how to get unsubscribed from any
Sourceforge mailing list.

You will need the following:

o Rum
o Creme of Coconut
o Pineapple Juice
o Ice, crushed is best
o Pineapple wedge & cherry garnish
o a Glass
o a Straw
o a Blender

1) Pour 4 oz. of Creme of Coconut into the blender
2) Pour 4 ounces of Pineapple Juice into the blender
3) Add 3 ounces of Rum (leave off if your religion or personal
   preferences disagree with alchohol)
4) Add about 2 cups of crushed ice
5) Blend till smooth
6) Garnish with a cherry and wedge of pineapple
7) Pour into the glass, add a straw and the garnish
8) Drink the tasty beverage
9) Relax, and notice the text at the bottom of *every post* to the
   mailing list:

https://lists.sourceforge.net/lists/listinfo/mpls-linux-devel

10) Decide to visit that page
11) Notice the text at the bottom of that page:

  To change your subscription (... or unsubscribe from mpls-linux-general),
  enter your subscription email address:

12) Enter "its...@ya..." in the little box and press "Edit Options"
13) Drink more of the tasty beverage
14) Enter in your password and hit "unsubscribe", and then finish the
    tasty beverage.

If you don't remember or know your password, add the following steps

15) Press "Email My Password To Me"
16) Drink more of the tasty beverage
17) Read your e-mail and retrieve your password
18) Go to step 14.

And you're all done!

> please unsubscribe me from this list
>=20
> 	=09
> ---------------------------------
> Do you Yahoo!?
>  Make Yahoo! your home page  =20
--=20
James R. Leu
jl...@mi...

Re: [mpls-linux-devel] Quagga-mpls

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

Shame that my mail client doesn't have (yet?) feature to check whether I 
attached correct patch. Therefore I have to send now new one with tiny 
improvement - this one compiles.


with my best wishes,

-- 
Hasso Tepper
Elion Enterprises Ltd.
WAN administrator

[mpls-linux-devel] Quagga-mpls

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

I finally took time to take a closer look what's done with Quagga regarding 
mpls :). If I will find more time later, I will try to put up some Debian 
packages which should simplify testing mpls in Debian. I didn't try to do 
anything useful yet, I have no mpls kernel. Just looked whether ldpd does 
something (without success, but I didn't try hard) and what possibilities 
are there in vty.

My first findings are in attached patch. There are fix for isisd compile, 
fix garbage in vty in label-map command and make use of FLAG macros where 
possible (there were some fixes in this part as well). I didn't fix 
zread_mpls_out_segment_add and zread_mpls_out_segment_delete in zserv.c 
though, because I didn't understand what is attempted there. And using 
api.gate.u.prefix6 if (api.gate.family == AF_INET) seems to be wrong for me 
as well in these functions ;).


with my best wishes,

-- 
Hasso Tepper
Elion Enterprises Ltd.
WAN administrator

Re: [mpls-linux-devel] code

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

You can get all of the mpls-linux source code from:

http://mpls-linux.sourceforge.net/

If you want the source code only, download mpls-linux-1.946.tar.bz2

http://prdownloads.sourceforge.net/mpls-linux/mpls-linux-1.946.tar.bz2?down=
load

In there you will find patch files for all of the relative components.

On Fri, Mar 04, 2005 at 05:42:11AM +0000, gemini murthy wrote:
> Hai,
>  I am gemini,doing final year engineering project.I want the simple sourc=
e code for the implementation of MPLS.please send me the code which should =
be executed.send and reply me immediately
> =20
> bye,
> gemini
> =20
>=20
> Yahoo! India Matrimony: Find your life partneronline.
--=20
James R. Leu
jl...@mi...

[mpls-linux-devel] code

[gemini m. <ang...@ya...>]

Hai,
 I am gemini,doing final year engineering project.I want the simple source code for the implementation of MPLS.please send me the code which should be executed.send and reply me immediately
 
bye,
gemini
 

Yahoo! India Matrimony: Find your life partneronline.