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

[mpls-linux-general] (no subject)

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

hi,
i am working on MPLS regarding that i want to open a file from 
kernel to store ip address of the destination of forwarded pkt..
for which i am using the code below in mpls_ouput function ..

char *buffer = kmalloc(4096,GFP_KERNEL);
oldfs = get_fs();
set_fs(KERNEL_DS);
fd=open("/dev/test",O_RDWR,0);
if(fd){
   read(fd,buffer,BUFFSIZE);
}
close(fd);
set_fs(oldfs);

but as soon as i ping another m/c my PC hangs..
can u please tell what is wrong with the above code..
whether i should use some lock... or is there any other method of 
opening a file from kernel
thanx in advance
anish

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

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

Hello,

On Thu, Mar 21, 2002 at 08:09:23AM +0800, Lauren H wrote:
> Hi,Jim
>   =20
>     If the operation mode is Downstream on Demand/Ordered/Conservative, w=
hat will change to the following mesg ordering you have provided?
>=20
> This is my opinion:
> (1) Once a TCP connection is established the active LDP speaker instead o=
f each LDP speaker sends an  Init message first.

The draft says that the active should send an Init mesg first for both case=
s,
(DoD and DU).  The main point that I was trying to get across with the
Init message is that both sides sent them independent of each other (ie
the passive should not adjust its Init message after hearing what the
active sends, it is not a negotiation, it is a comparison).

> (2) When a route changes, label releases and requests are issued first, t=
hen label withdraws and mappings are issued later.

In DU when the session is finally established each side sends label mappings
for every FEC in it routing table.  In DoD mode they will send label reques=
ts
instead.

>  Is it right? Anything else?
>=20
> In addition, I have other questions:
> (1) If a LDP session is established between lsrs that are not directly co=
nnected at the link level, should the lsrs be configured to be able to send=
 and/or receive Targeted Hello? So the network operators should know which =
lsr can send/receive Targeted Hello?

Yes.  It is not necessary, but all implementations I have worked with
are implemented this way.

> (2) When operating in Downstream on Demand mode, is it necessary to send =
the Address message?

Yes.  Otherwise how would a LSR know who to send a label requets to?

> Best Regards,
>=20
> Lauren H
>=20
>=20
> ----- Original Message -----=20
> From: James R. Leu <jl...@mi...>
> To: Anil H <ani...@ya...>
> Cc: <mpl...@li...>
> Sent: Wednesday, March 20, 2002 5:56 AM
> Subject: Re: [mpls-linux-general] reg LDP
>=20
>=20
> > I can give you a quick overview of mesg odering, but you will want to
> > look at RFC3036 for the details.
> >=20
> > I'm only going to explain the direct/DU/Independent/Conservitive mode
> > of operation (look to the RFC as to what those mean)
> >=20
> > -Each LDP speaker is configured to talk LDP on specific interfaces.
> > -Each LDP speaker sends udp hello's on their interface to the all route=
rs
> >  multicast address, port 646.
> > -When an LDP speaker recieves a hellos it creates a logical adj for it.
> > -For each adj the LDP speaker decides whether it will play the
> >  active role or the passive role in the adj. (only 2 LDP speakers per a=
dj,
> >  on a broadcast medium an LDP speaker will form multiple adj).  The
> >  process of deciding active vs passive is outlined in the RFC.
> > -The active LDP speaker in the adj will attempt to create a TCP connect=
ion
> >  to the passive LDP speaker on port 646.
> > -Once a TCP connection is established each LDP speaker sends an Init
> >  message.
> > -When an LDP speaker recieves a Init message it looks at the parameters
> >  in the Init message and compares them with it configured parameters.
> >  If the parameters are acceptable a keepalive message is sent back.
> >  If the parameters are un-acceptable a notification is sent back and the
> >  TCP connection is closed
> > -When an LDP speaker hears the first keepalive for an adj, it then send=
s out
> >  one or multiple address messages which advertise the IP address attach=
ed
> >  to the LDP speaker.
> > -After sending address messages, the LDP speaker sends label advertisem=
ents
> >  for each FEC in it routing table.
> > -When a route changes, label withdraws and mappings are issued.
> > -When an interface changes, address withdraw and address messages are i=
ssued.
> >=20
> > I hope this atleast gives you a starting point.  The RFC is really the
> > best place to get the details though. (ofcourse you can alwyas read the
> > source to ldp-portable as well :-)
> >=20
> > Jim
> >=20
> > On Tue, Mar 19, 2002 at 11:20:34AM -0800, Anil H wrote:
> > > hi,
> > >=20
> > > Could ne1 give me the basic flow for LDP (for eg the
> > > sequence of messages) r some pointers to were i can
> > > find the information apart from the RFC.
> > >=20
> > > Thanks in advance,
> > > Anil
> > >=20
> > > __________________________________________________
> > > Do You Yahoo!?
> > > Yahoo! Sports - live college hoops coverage
> > > http://sports.yahoo.com/
> > >=20
> > > _______________________________________________
> > > mpls-linux-general mailing list
> > > mpl...@li...
> > > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> >=20
> > --=20
> > James R. Leu
> >=20
> > _______________________________________________
> > mpls-linux-general mailing list
> > mpl...@li...
> > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> =FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=
=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FC:=18=A2=E6=1A=86=
=8A=FF=FCg=AD=CA=8B=AB~=B7=9E=FF&=A1=A2=8F=DC=A2f=9Dv=B7=AC=B1=ABa=B6=DA=FF=
=FF=F9=9A=8A_=F2j=1A(=FD=CA&=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=
=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=E6=A6[=FF=96)=
=EE=C7=F8=1E=9D=EA=DA=96f=A2=96)=E0=96+-=9A=99l=FEX=A7=BB=1F=E0zw=ABj_=E5=
=8A=CBl=FE=CA.=AD=C7=9F=A2=B8=1E=FEw=AD=86=DBi=B3=FF=FF=96+-=B3=FB(=BA=B7=
=1E~=8A=E0{=F9=DE=B7=F9b=B2=DB?=96+-=8Aw=E8=FEje=B3=F9b=9E=EC=FF=81=E9=DE

--=20
James R. Leu

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

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

Re: [mpls-linux-general] Sending heavy traffic via MPLS

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

If I remember correctly 'scp' uses a TCP session.  There has been a long
standing issue with using TCP (path discovery) and MPLS.  You could try to
setting the MTU of the host interface to something less then 1496 and see
if that helps a all.

Jim

On Thu, Mar 21, 2002 at 04:00:25AM +0800, ncc49a wrote:
> Hi,
> 
>     I have set up an MPLS LSP consisting of Linux and Cisco routers. I have
> verified the LSP by doing a ping and checking the packet header's via
> Ethereal, and some Cisco debug commands. I then tried to send some "heavy
> traffic" over the LSP, by  doing a "scp" of a large file from one end host
> to another...
> 
>     However, the "scp" hangs... Has anyone faced such problems when sending
> heavy traffic? Could it be that my Linux routers are not "fast" enough to
> handle the MPLS packets?
> 
>     Thanks!
> 
> Regards,
> 
> Paul
> 
> 
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu

[mpls-linux-general] Sending heavy traffic via MPLS

[ncc49a <nc...@si...>]

Hi,

    I have set up an MPLS LSP consisting of Linux and Cisco routers. I have
verified the LSP by doing a ping and checking the packet header's via
Ethereal, and some Cisco debug commands. I then tried to send some "heavy
traffic" over the LSP, by  doing a "scp" of a large file from one end host
to another...

    However, the "scp" hangs... Has anyone faced such problems when sending
heavy traffic? Could it be that my Linux routers are not "fast" enough to
handle the MPLS packets?

    Thanks!

Regards,

Paul

Re: [mpls-linux-general] reg LDP

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

I can give you a quick overview of mesg odering, but you will want to
look at RFC3036 for the details.

I'm only going to explain the direct/DU/Independent/Conservitive mode
of operation (look to the RFC as to what those mean)

-Each LDP speaker is configured to talk LDP on specific interfaces.
-Each LDP speaker sends udp hello's on their interface to the all routers
 multicast address, port 646.
-When an LDP speaker recieves a hellos it creates a logical adj for it.
-For each adj the LDP speaker decides whether it will play the
 active role or the passive role in the adj. (only 2 LDP speakers per adj,
 on a broadcast medium an LDP speaker will form multiple adj).  The
 process of deciding active vs passive is outlined in the RFC.
-The active LDP speaker in the adj will attempt to create a TCP connection
 to the passive LDP speaker on port 646.
-Once a TCP connection is established each LDP speaker sends an Init
 message.
-When an LDP speaker recieves a Init message it looks at the parameters
 in the Init message and compares them with it configured parameters.
 If the parameters are acceptable a keepalive message is sent back.
 If the parameters are un-acceptable a notification is sent back and the
 TCP connection is closed
-When an LDP speaker hears the first keepalive for an adj, it then sends out
 one or multiple address messages which advertise the IP address attached
 to the LDP speaker.
-After sending address messages, the LDP speaker sends label advertisements
 for each FEC in it routing table.
-When a route changes, label withdraws and mappings are issued.
-When an interface changes, address withdraw and address messages are issued.

I hope this atleast gives you a starting point.  The RFC is really the
best place to get the details though. (ofcourse you can alwyas read the
source to ldp-portable as well :-)

Jim

On Tue, Mar 19, 2002 at 11:20:34AM -0800, Anil H wrote:
> hi,
> 
> Could ne1 give me the basic flow for LDP (for eg the
> sequence of messages) r some pointers to were i can
> find the information apart from the RFC.
> 
> Thanks in advance,
> Anil
> 
> __________________________________________________
> Do You Yahoo!?
> Yahoo! Sports - live college hoops coverage
> http://sports.yahoo.com/
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu

[mpls-linux-general] reg LDP

[Anil H <ani...@ya...>]

hi,

Could ne1 give me the basic flow for LDP (for eg the
sequence of messages) r some pointers to were i can
find the information apart from the RFC.

Thanks in advance,
Anil

__________________________________________________
Do You Yahoo!?
Yahoo! Sports - live college hoops coverage
http://sports.yahoo.com/

Re: Re: [mpls-linux-general] A question about quit( )

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

On Tue, Mar 19, 2002 at 09:34:07AM +0800, zjh...@so... wrote:
> hi,Jim
> I had read the ldp_cfg.c.I want to use the func( ) in ldp_cfg.c to finish the quit( ).
> I will broke the following function in order: 
>   ldp_cfg_range_set( )
>   ldp_cfg_if/peer_set( ) 
>   ldp_cfg_entity_set( )
>   ldp_cfg_global_set( )
>   ldp_cfg_global_close( )
>   ldp_global_shutdown( )
>   ldp_global_delete( )
> Is it practical? I also have seen ldp_global.c.There seem to be other function.Perhabps they can do the same work!
> Can you give me your advice?

I think you have it correct.

> 
> At last ,I want report a bug to you. 
> When the program run in VXWORKS.There is another problem.The message type is always wrong.Because the Ubit and the MESSAGE TYPE is 2 BYTE together.They work wrong in BIG_ENDIAN BYTE ORDER. Such as HELLO message TYPE is 0X0100;But it will be 0X0200 if running in vxworks.This is caused by the order of Ubit and MESSAGE TYPE FIELD.
> perhaps you can use the makefile option such as _D_BIG_ENDIAN to solve it.Now the makefile option _D_BIG_ENDIAN 
> doesn't work.

Check out the code in ldp_nortel.h

Jim

> -----  Original Message  -----
> From: James R. Leu 
> To: zjh...@so... 
> Cc: mpl...@li... 
> Subject: Re: [mpls-linux-general] A question about quit( )
> Sent: Tue Mar 19 07:50:58 CST 2002
> 
> > Hello,
> > 
> > See my response in-line.
> > 
> > On Fri, Mar 15, 2002 at 09:59:03PM +0800, zjh...@so... wrote:
> > > hi,Jim
> > > I have read your code about the command func :quit( ).In linux you use the system API "exit( )"to finish almost all the work. They work well in Linux.
> > > 
> > > I have also tried your code in vxworks .Using the quit command ofent make program down.
> > > So I want to write a quit( ) but not use the "exit( )",do everything by myself.Can you 
> > > give me some advice:such as the sequence to close"session,interface,adjacency,global "etc.
> > 
> > Removing the entities will tear down the sessions/adjacecncies.  Then removing
> > the global will cleanup the porting layer.
> > 
> > > I am busy in try to port your code to vxworks.Once I finish it ,I will share my idea and my
> > > experience and my code with all  in the mailing list.I will try my best.
> > 
> > Excellent!  I have just starting working on ldp-portabel again, so hopefully
> > I can help you fix any LDP related bugs that you find.
> > 
> > Jim
> > 
> > 
> > > 
> > > All thanks!!
> > > 
> > > Your sincerely :zhang jianhui
> > > 
> > > _______________________________________________
> > > mpls-linux-general mailing list
> > > mpl...@li...
> > > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> > 
> > -- 
> > James R. Leu

-- 
James R. Leu

Re: [mpls-linux-general] Re:About CR-LDP and LIB

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

On Tue, Mar 19, 2002 at 10:29:28AM +0800, clz wrote:
> Hi,Jim
>      Thank you very much.
>     =20
>     You have told me that ldp-portable store LFIB in the FEC tree and mpl=
s-linux store LIB in the IPv4 FIB. So would you like tell me in which .c/.h=
 document I will find them ? =20

The FIB is located in the kernel source linux/net/ipv4/fib_*
In LDP portable look at ldp-portable/lib/ldp_attr.[ch]

>     Another question:=20
>     In the network ,we can't estimate what traffic the router/switcher wi=
ll receive, so if    we use the TE-MIB as the user's interface to setting u=
p TE LSPS via CR-LDP, does it mean we can only identify  some TE traffic?  =
If the LER receive some DiffServ-traffic,how to trigger the LER to establis=
h a CR-LSP via CR-LDP promptly?  =20

At the operator of the router/switch it is your responsibility to categorize
all of the traffic going through your router/switch.  Most of the time
you do this by knowing where it will exit your network and then you create
TE-LSPs to those exist points.

Jim


>=20
> RG
>=20
> clz
>=20
>=20
>=20
>=20
>=20
>=20
>=20
> =FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=
=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FC:=18=A2=E6=1A=86=
=8A=FF=FCg=AD=CA=8B=AB~=B7=9E=FF&=A1=A2=8F=DC=A2f=9Dv=B7=AC=B1=ABa=B6=DA=FF=
=FF=F9=9A=8A_=F2j=1A(=FD=CA&=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=
=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=E6=A6[=FF=96)=
=EE=C7=F8=1E=9D=EA=DA=96f=A2=96)=E0=96+-=9A=99l=FEX=A7=BB=1F=E0zw=ABj_=E5=
=8A=CBl=FE=CA.=AD=C7=9F=A2=B8=1E=FEw=AD=86=DBi=B3=FF=FF=96+-=B3=FB(=BA=B7=
=1E~=8A=E0{=F9=DE=B7=F9b=B2=DB?=96+-=8Aw=E8=FEje=B3=F9b=9E=EC=FF=81=E9=DE

--=20
James R. Leu

[mpls-linux-general] Re:About CR-LDP and LIB

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

Re:Re: [mpls-linux-general] A question about quit( )

[<zjh...@so...>]

hi,Jim
I had read the ldp_cfg.c.I want to use the func( ) in ldp_cfg.c to finish the quit( ).
I will broke the following function in order: 
  ldp_cfg_range_set( )
  ldp_cfg_if/peer_set( ) 
  ldp_cfg_entity_set( )
  ldp_cfg_global_set( )
  ldp_cfg_global_close( )
  ldp_global_shutdown( )
  ldp_global_delete( )
Is it practical? I also have seen ldp_global.c.There seem to be other function.Perhabps they can do the same work!
Can you give me your advice?

At last ,I want report a bug to you. 
When the program run in VXWORKS.There is another problem.The message type is always wrong.Because the Ubit and the MESSAGE TYPE is 2 BYTE together.They work wrong in BIG_ENDIAN BYTE ORDER. Such as HELLO message TYPE is 0X0100;But it will be 0X0200 if running in vxworks.This is caused by the order of Ubit and MESSAGE TYPE FIELD.
perhaps you can use the makefile option such as _D_BIG_ENDIAN to solve it.Now the makefile option _D_BIG_ENDIAN 
doesn't work.

All thanks!!

Your sincerely :zhang jianhui




-----  Original Message  -----
From: James R. Leu 
To: zjh...@so... 
Cc: mpl...@li... 
Subject: Re: [mpls-linux-general] A question about quit( )
Sent: Tue Mar 19 07:50:58 CST 2002

> Hello,
> 
> See my response in-line.
> 
> On Fri, Mar 15, 2002 at 09:59:03PM +0800, zjh...@so... wrote:
> > hi,Jim
> > I have read your code about the command func :quit( ).In linux you use the system API "exit( )"to finish almost all the work. They work well in Linux.
> > 
> > I have also tried your code in vxworks .Using the quit command ofent make program down.
> > So I want to write a quit( ) but not use the "exit( )",do everything by myself.Can you 
> > give me some advice:such as the sequence to close"session,interface,adjacency,global "etc.
> 
> Removing the entities will tear down the sessions/adjacecncies.  Then removing
> the global will cleanup the porting layer.
> 
> > I am busy in try to port your code to vxworks.Once I finish it ,I will share my idea and my
> > experience and my code with all  in the mailing list.I will try my best.
> 
> Excellent!  I have just starting working on ldp-portabel again, so hopefully
> I can help you fix any LDP related bugs that you find.
> 
> Jim
> 
> 
> > 
> > All thanks!!
> > 
> > Your sincerely :zhang jianhui
> > 
> > _______________________________________________
> > mpls-linux-general mailing list
> > mpl...@li...
> > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> 
> -- 
> James R. Leu

Re: [mpls-linux-general] About CR-LDP and LIB

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

On Mon, Mar 18, 2002 at 04:08:27PM +0800, clz wrote:
> Hi,Jim
>=20
>    Now I have two questions as following:
>=20
> 1) If I use CR-LDP,how to trigger the LER to generate a label Request Mes=
sage? When to use Request Driven Label Assignment,how to identify that some=
 traffic is  QoS-guaranteed without RSVP?

Most on the time TE LSPs are triggered by user configuration.  I plan to
use the TE-MIB as the users interface to setting up TE LSPS (via CR-LDP).

> 2)What is the difference between LIB and LFIB?
>   How can I find the LIB's data structure in the  ldp-portable-0.060.tar =
and  mpls-linux-0.993.tar?

Hmmm.  I'm not sure.  I would think they are the same, which is the informa=
tion
about how IPv4 routes map to LDP generated LSPs.

Both actaully contain this info.  ldp-portable keeps a copy of it 'view' of
the LIB, and the kernel has a copy of what is actually active.
ldp-portable view is stored in the FEC tree, and mpls-linux store it in the
IPv4 FIB.

Jim


>=20
> RG
> Thanks in advance.
>=20
> clz
>=20
>=20
>=20
>=20
> =FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=
=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FC:=18=A2=E6=1A=86=
=8A=FF=FCg=AD=CA=8B=AB~=B7=9E=FF&=A1=A2=8F=DC=A2f=9Dv=B7=AC=B1=ABa=B6=DA=FF=
=FF=F9=9A=8A_=F2j=1A(=FD=CA&=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=
=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=FF=E6=A6[=FF=96)=
=EE=C7=F8=1E=9D=EA=DA=96f=A2=96)=E0=96+-=9A=99l=FEX=A7=BB=1F=E0zw=ABj_=E5=
=8A=CBl=FE=CA.=AD=C7=9F=A2=B8=1E=FEw=AD=86=DBi=B3=FF=FF=96+-=B3=FB(=BA=B7=
=1E~=8A=E0{=F9=DE=B7=F9b=B2=DB?=96+-=8Aw=E8=FEje=B3=F9b=9E=EC=FF=81=E9=DE

--=20
James R. Leu

Re: [mpls-linux-general] A question about quit( )

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

Hello,

See my response in-line.

On Fri, Mar 15, 2002 at 09:59:03PM +0800, zjh...@so... wrote:
> hi,Jim
> I have read your code about the command func :quit( ).In linux you use the system API "exit( )"to finish almost all the work. They work well in Linux.
> 
> I have also tried your code in vxworks .Using the quit command ofent make program down.
> So I want to write a quit( ) but not use the "exit( )",do everything by myself.Can you 
> give me some advice:such as the sequence to close"session,interface,adjacency,global "etc.

Removing the entities will tear down the sessions/adjacecncies.  Then removing
the global will cleanup the porting layer.

> I am busy in try to port your code to vxworks.Once I finish it ,I will share my idea and my
> experience and my code with all  in the mailing list.I will try my best.

Excellent!  I have just starting working on ldp-portabel again, so hopefully
I can help you fix any LDP related bugs that you find.

Jim


> 
> All thanks!!
> 
> Your sincerely :zhang jianhui
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu

Re: [mpls-linux-general] ARP and MPLS

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

This maynot be the problem but you also need to set the 'labelspace' on the
incoming interfaces:

mplsadm2 -L eth1:0

Another thing you can try is to turn on mpls kernel debugging:

mplsadm2 -d

Look at the ouptut from 'dmesg' or in /var/log/messages.

Jim

On Sun, Mar 17, 2002 at 05:15:47PM +0100, Linder Erik wrote:
> Hi,
> I have put up a very simple network with 2 pc . After doing all the
> configuration stuff, I try to ping B from A and there I get something
> strange: tha ARP request are sent but keep unanswered
> Tcpdump on the emitting eth gives:
> 17:26:34.873599 arp who-has 137.194.214.142 tell 137.194.214.141
> 
> and it'is all. I don't know what to do , or start looking where
> something has got wrong.
> 
> I hope someone can help me
> Erik Linder
> 
> Here is my config:
> 
>   ---------             ---------
>   |   A   |             |   B   |
>   ---------             ---------
>       |eth1                 |eth1
>       |.141                 |.142
>  -----------------------------------
> A
> ifconfig eth1 137.194.214.141 netmask 255.255.255.192 up
>  
> B
> ifconfig eth1 137.194.214.142 netmask 255.255.255.192 up
> 
> Sur A:
>    mplsadm2 -A -O gen:16:eth1:ipv4:137.194.214.142  
>    iptables -A OUTPUT -d 137.194.214.142/32 -j MPLS --set-key KEY1
>    mplsadm2 -A -I gen:17:0
>  
>    Ou KEY1 est le resultat renvoye par la commande precedante 'mplsadm2'
> 
>  
>  Sur B:
>    mplsadm2 -A -O gen:17:eth1:ipv4:137.194.214.141
>    iptables -A OUTPUT -d 137.194.214.141/32 -j MPLS --set-key KEY2
>    mplsadm2 -A -I gen:16:0
>  
>       Ou KEY1 est le resultat renvoye par la commande precedante
> 'mplsadm2' 
> 
> Sur les 2 host:
>    mplsadm2 -L eth1:0
> 
> 
> PS: I tried using route instead of  iptables and I got the same result
> 
> 
> 
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu

[mpls-linux-general] About CR-LDP and LIB

[clz <hly...@ya...>]

SGksSmltDQoNCiAgIE5vdyBJIGhhdmUgdHdvIHF1ZXN0aW9ucyBhcyBmb2xsb3dpbmc6DQoNCjEp
IElmIEkgdXNlIENSLUxEUCxob3cgdG8gdHJpZ2dlciB0aGUgTEVSIHRvIGdlbmVyYXRlIGEgbGFi
ZWwgUmVxdWVzdCBNZXNzYWdlPyBXaGVuIHRvIHVzZSBSZXF1ZXN0IERyaXZlbiBMYWJlbCBBc3Np
Z25tZW50LGhvdyB0byBpZGVudGlmeSB0aGF0IHNvbWUgdHJhZmZpYyBpcyAgUW9TLWd1YXJhbnRl
ZWQgd2l0aG91dCBSU1ZQPw0KDQoyKVdoYXQgaXMgdGhlIGRpZmZlcmVuY2UgYmV0d2VlbiBMSUIg
YW5kIExGSUI/DQogIEhvdyBjYW4gSSBmaW5kIHRoZSBMSUIncyBkYXRhIHN0cnVjdHVyZSBpbiB0
aGUgIGxkcC1wb3J0YWJsZS0wLjA2MC50YXIgYW5kICBtcGxzLWxpbnV4LTAuOTkzLnRhcj8NCg0K
DQpSRw0KVGhhbmtzIGluIGFkdmFuY2UuDQoNCmNseg0KDQoNCg0KDQo=


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

[mpls-linux-general] ARP and MPLS

[Linder E. <li...@en...>]

Hi,
I have put up a very simple network with 2 pc . After doing all the
configuration stuff, I try to ping B from A and there I get something
strange: tha ARP request are sent but keep unanswered
Tcpdump on the emitting eth gives:
17:26:34.873599 arp who-has 137.194.214.142 tell 137.194.214.141

and it'is all. I don't know what to do , or start looking where
something has got wrong.

I hope someone can help me
Erik Linder

Here is my config:

  ---------             ---------
  |   A   |             |   B   |
  ---------             ---------
      |eth1                 |eth1
      |.141                 |.142
 -----------------------------------
A
ifconfig eth1 137.194.214.141 netmask 255.255.255.192 up
 
B
ifconfig eth1 137.194.214.142 netmask 255.255.255.192 up

Sur A:
   mplsadm2 -A -O gen:16:eth1:ipv4:137.194.214.142  
   iptables -A OUTPUT -d 137.194.214.142/32 -j MPLS --set-key KEY1
   mplsadm2 -A -I gen:17:0
 
   Ou KEY1 est le resultat renvoye par la commande precedante 'mplsadm2'

 
 Sur B:
   mplsadm2 -A -O gen:17:eth1:ipv4:137.194.214.141
   iptables -A OUTPUT -d 137.194.214.141/32 -j MPLS --set-key KEY2
   mplsadm2 -A -I gen:16:0
 
      Ou KEY1 est le resultat renvoye par la commande precedante
'mplsadm2' 

Sur les 2 host:
   mplsadm2 -L eth1:0


PS: I tried using route instead of  iptables and I got the same result

[mpls-linux-general] A question about quit( )

[<zjh...@so...>]

hi,Jim
I have read your code about the command func :quit( ).In linux you use the system API "exit( )"to finish almost all the work. They work well in Linux.

I have also tried your code in vxworks .Using the quit command ofent make program down.
So I want to write a quit( ) but not use the "exit( )",do everything by myself.Can you 
give me some advice:such as the sequence to close"session,interface,adjacency,global "etc.

I am busy in try to port your code to vxworks.Once I finish it ,I will share my idea and my
experience and my code with all  in the mailing list.I will try my best.

All thanks!!

Your sincerely :zhang jianhui

Re: [mpls-linux-general] generic querry

[Radu D. <dr...@ee...>]

> and to see the packets urself, you can use ethereal. ;-)
>
> -abhijit
If the node is the source or an LSR,
but not if the node is the destination of the MPLS packet  :))


>
> Radu Dragos wrote :
>
> >Hi.
> >1. check the files within the /proc/net/mpls*
> ># watch -n1 cat /proc/net/mpls*
> >And you will see changes in the number of packets using an output label
or
> >having an input label
> >
> >2. turn on debug
> ># ./mplsadm2 -d
> >than if the router is "doing label switching" you will see a lot of
> >messages in /var/log/messages
> >#tail -f /var/log/messages
> >
> >Radu
> >
> >
> >----- Original Message -----
> >From: Brij Kishore Rai <bk...@ua...>
> >To: <mpl...@li...>
> >Sent: Friday, March 15, 2002 11:41 AM
> >Subject: [mpls-linux-general] generic querry
> >
> >
> >> Hi,
> >>      I am working on MPLS. I have a very basic question how I can make
> >> sure of that my routers are doing label switching???
> >>
> >> Best regards,
> >> --Brij Kishore
>
> --
> bitwise byte foolish
> --
>
>
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

Re: [mpls-linux-general] generic querry

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

and to see the packets urself, you can use ethereal. ;-)

-abhijit

Radu Dragos wrote :

>Hi.
>1. check the files within the /proc/net/mpls*
># watch -n1 cat /proc/net/mpls*
>And you will see changes in the number of packets using an output label or
>having an input label
>
>2. turn on debug
># ./mplsadm2 -d
>than if the router is "doing label switching" you will see a lot of
>messages in /var/log/messages
>#tail -f /var/log/messages
>
>Radu
>
>
>----- Original Message -----
>From: Brij Kishore Rai <bk...@ua...>
>To: <mpl...@li...>
>Sent: Friday, March 15, 2002 11:41 AM
>Subject: [mpls-linux-general] generic querry
>
>
>> Hi,
>>      I am working on MPLS. I have a very basic question how I can make
>> sure of that my routers are doing label switching???
>>
>> Best regards,
>> --Brij Kishore

-- 
bitwise byte foolish
--

Re: [mpls-linux-general] generic querry

[Radu D. <dr...@ee...>]

Hi.
1. check the files within the /proc/net/mpls*
# watch -n1 cat /proc/net/mpls*
And you will see changes in the number of packets using an output label or
having an input label

2. turn on debug
# ./mplsadm2 -d
than if the router is "doing label switching" you will see a lot of
messages in /var/log/messages
#tail -f /var/log/messages

Radu


----- Original Message -----
From: Brij Kishore Rai <bk...@ua...>
To: <mpl...@li...>
Sent: Friday, March 15, 2002 11:41 AM
Subject: [mpls-linux-general] generic querry


> Hi,
>      I am working on MPLS. I have a very basic question how I can make
> sure of that my routers are doing label switching???
>
> Best regards,
> --Brij Kishore
>
>
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

[mpls-linux-general] generic querry

[Brij K. R. <bk...@ua...>]

Hi,
     I am working on MPLS. I have a very basic question how I can make
sure of that my routers are doing label switching???

Best regards,
--Brij Kishore

[mpls-linux-general] Remove from mailing list

[Rekha R <rek...@wi...>]

**************************Disclaimer************************************


Information contained in this E-MAIL being proprietary to Wipro Limited
is 'privileged' and 'confidential' and intended for use only by the
individual or entity to which it is addressed. You are notified that any
use, copying or dissemination of the information contained in the E-MAIL
in any manner whatsoever is strictly prohibited.


*****************************************************************************

[mpls-linux-general] Off topic: udp traffic generator

[Stefano A. <avi...@li...>]

Hi all,

if someone needs a traffic generator, I have realized one which uses =
MGEN's utilities to analyze log files but presents some additional =
features with respect to MGEN .
It is available at http://www.grid.unina.it/grid/mtools/.

I would happy if someone uses it and tells me what he/she thinks about.

Regards,
Stefano

Re: [mpls-linux-general] Persistent FEC binding

[Radu D. <dr...@ee...>]

Here I attached the two requested files:
rt_cache2 - before flush
rt_cache3 - after flush

----- Original Message -----
From: James R. Leu <jl...@mi...>
To: Radu Dragos <dr...@ee...>
Cc: <mpl...@li...>
Sent: Thursday, March 14, 2002 5:31 PM
Subject: Re: [mpls-linux-general] Persistent FEC binding


> On Thu, Mar 14, 2002 at 03:23:46PM -0000, Radu Dragos wrote:
> > Hi Jim and all,
> > I encountered the following problem.
> >
> > I have 2 output labels:
> >
> > ./mplsadm2 -A -O gen:16:eth0:ipv4:192.168.100.1
> > ./mplsadm2 -A -O gen:17:eth0:ipv4:192.168.100.1
> >
> > (supose the 2 keys are 0x2 and 0x3)
> >
> > Firs I bind label 16 to a FEC:
> > iptables -A OTPUT -d ip_dst_1 -j MPLS --set-key 0x2
> >
> > Then if I want to unbind and I flush the rules :
> > iptables -F
> >
> > The packets will continue to use the binding.
> > Even if I create another binding like
> >   iptables -A OTPUT -d ip_dst_2 -j MPLS --set-key 0x3
> > the packets will still use the first label....
>
> It could be that the route cache is stale.  Dump /proc/net/rt_cache
before
> and after the flush and send me the data.
>
> Jim
>
> >
> > The only way I could delete the binding was by deleting the output
label
> > 16.
> >
> > Is there any other way to unbind the labels ???
> >
> > Thanks in advance,
> > Radu
> >
> >
> >
> >
> >
> >
> >
> >
> >
> >
> > _______________________________________________
> > mpls-linux-general mailing list
> > mpl...@li...
> > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
>
> --
> James R. Leu

Re: [mpls-linux-general] Persistent FEC binding

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

On Thu, Mar 14, 2002 at 03:23:46PM -0000, Radu Dragos wrote:
> Hi Jim and all,
> I encountered the following problem.
> 
> I have 2 output labels:
> 
> ./mplsadm2 -A -O gen:16:eth0:ipv4:192.168.100.1
> ./mplsadm2 -A -O gen:17:eth0:ipv4:192.168.100.1
> 
> (supose the 2 keys are 0x2 and 0x3)
> 
> Firs I bind label 16 to a FEC:
> iptables -A OTPUT -d ip_dst_1 -j MPLS --set-key 0x2
> 
> Then if I want to unbind and I flush the rules :
> iptables -F
> 
> The packets will continue to use the binding.
> Even if I create another binding like
>   iptables -A OTPUT -d ip_dst_2 -j MPLS --set-key 0x3
> the packets will still use the first label....

It could be that the route cache is stale.  Dump /proc/net/rt_cache before
and after the flush and send me the data.

Jim

> 
> The only way I could delete the binding was by deleting the output label
> 16.
> 
> Is there any other way to unbind the labels ???
> 
> Thanks in advance,
> Radu
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu