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

Re: [mpls-linux-general] James your help required

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

It is clearly a configuration issue.  I do not have time to
sanity check your commands.  All I can recommend is that you use the
kernel debugging (mplsadm2 -d) and the counters provided in
/proc/mpls/mpls_(in/out) to help you find your error.

On Fri, Jul 02, 2004 at 07:45:53PM +0300, iq...@cc... wrote:
> Hello,
> 
> Kindly check my last mail and give me your feedback.
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email sponsored by Black Hat Briefings & Training.
> Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
> digital self defense, top technical experts, no vendor pitches, 
> unmatched networking opportunities. Visit www.blackhat.com
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

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

[mpls-linux-general] James your help required

[<iq...@cc...>]

Hello,

Kindly check my last mail and give me your feedback.

Re: [mpls-linux-general] Question on outgoing_label KEY

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

On Mon, Jun 28, 2004 at 11:16:42AM -0700, Helen (Wenjing) Zhang wrote:
> In /proc/net/mpls_out I got:
> 
> 0x00000059 0/0/0 2 DS2EXP( DS(00)->EXP(00) DS(0a)->EXP(02) DS(0c)->EXP(03)
> DS(12)->EXP(04) DS(14)->EXP(05) DS(1a)->EXP(06) DS(1c)->EXP(07)
> DS(2e)->EXP(01) ) EXP2TC( EXP(0)->TC(ffff) EXP(1)->TC(00b8) EXP(2)->TC(0028)
> EXP(3)->TC(0030) EXP(4)->TC(0048) EXP(5)->TC(0050) EXP(6)->TC(0068)
> EXP(7)->TC(0070) ) PUSH(gen 11680) SET(eth0,192.168.2.2)
> 
> >From the reading material, "0x00000059" is an outgoing _label KEY.
> 
> What is the relationship with Label, LSP ID and the outgoing_label KEY?
> How to assign this KEY?
> How and where can I find the value for this KEY except the outgoing label
> table?

The outgoing key is generated by the kernel.  It must be unique per
outgoing label.  Are you looking to write a script of program which
creates outgoing labels and modifies the instructions and thus needs the
outgoing label key?  If you are using mplsadm2, any newly created key is
sent to standard out.  If you want to create outgoing labels in C use the
ioctl() command.  You can use the source for mplsadm2 as an example.

> 
> Thanks.
> 
> Helen
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email sponsored by Black Hat Briefings & Training.
> Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
> digital self defense, top technical experts, no vendor pitches, 
> unmatched networking opportunities. Visit www.blackhat.com
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

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

Re: [mpls-linux-general] Reserve Bandwidth Problem.. ?

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

I think you mis-understand what is meant by "reserving bandwidth"
when refering to RSVP-TE.

Most RSVP-TE implementation do not propogate bandwidth reservations to the
forwarding plan.  In otherwords, whether the LSP reserves 100Mbps or 10Mbps
I can still send 1000Mbps worth of traffic on the LSP.  There is no shaping
or policing occuring.

The "reserved bandwith" of established LSPs is "score carded" or tallied. The
sum of them is then used as part of RSVP-TEs "admission control", which
will permit or deny a LSPs request for "reserved bandwidth".

Pim's RSVP-TE implementation allows for setting PHB or queuing treatment
for packets that traverse a LSP.  Note that the PHB is only relavant when
there exists multiple LSPs with differing DSCPs.  In otherwords you
will not see any difference in performance between a single statically
configured LSP vesus a single RSVP-TE configured LSP.

On Fri, Jul 02, 2004 at 02:16:48PM +0700, baim wrote:
> Hi.. Pim, Jan, all...
> 
> I had read the documentation of rsvp-te installation from 
> dsmpls.rug.ac.be and successfull installing the packet. 
> But when i run The rsvp like the manual told me :
> 
> ./rsvpd -D at all node and router
> rapirecv_auto at egress
> rtest2 at ingress
> The process seems OK and i got no error message. But since 
> RSVP is used for reserving a bandwidth i wonder how to di 
> that ?
> 
> But i wonder how to reserve a bandwidth... ? 
> Is there any command that i had run... ? or may be i have 
> to install more packet for reserving the bandwidth.. ?
> 
> I hope you not scruple with my question since i am a 
> newbie at this such things.
> ===========================================================================================
> Netkuis Instan untuk wilayah Bandung (kode area 022) - SD,SMP,SMA berhadiah 
> total puluhan juta rupiah... periode I dimulai 1 April 2004
> ===========================================================================================
> 
> 
> -------------------------------------------------------
> This SF.Net email sponsored by Black Hat Briefings & Training.
> Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
> digital self defense, top technical experts, no vendor pitches, 
> unmatched networking opportunities. Visit www.blackhat.com
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

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

Re: [mpls-linux-general] Problems using libipq/QUEUE with MPLS

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

I have not worked with libipq.  Try sending a mesg to the libipq mailing
list.

On Fri, Jul 02, 2004 at 02:53:53AM +0300, Mohammad Rehan Sami wrote:
> I am having problems running libipq/ QUEUE with ur MPLS implementation. My 
> libipq program successfully
> compiles. I add the following rule to the iptables alongwith other iptables 
> rule necessary for the MPLS setup:
> 
> iptables -A OUTPUT -p icmp -j QUEUE
> 
> Now this should queue all originating ping packets and then depending on the
> NF mark in my libipq program, action should be taken. But it seems that 
> regardless of the NF mark(NF_DROP or NF_ACCEPT), nothing seems to happen. If i 
> say NF_DROP in my libipq program, still icmp packets traverse through the 
> network from my ingress router. Basically I want to queue specific packets and 
> then let my libipq program decide what to do. If accepted then send them to 
> the outgoing LSP using spec_nh instruction of MPLS. 
> 
> Furthermore, when I run the modprobe commands to check for ipatble_filter and 
> ip_queue, I get a fatal message stating "module not found". Could you please 
> tell me where I am wrong. I am using the libipq example program provided in 
> the manpage.
> 
> Regards,
> 
> 
> -------------------------------------------------------
> This SF.Net email sponsored by Black Hat Briefings & Training.
> Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
> digital self defense, top technical experts, no vendor pitches, 
> unmatched networking opportunities. Visit www.blackhat.com
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

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

Re: [mpls-linux-general] Two LSPs on the same physical path:Correction

[<iq...@cc...>]

Quoting Pierre ANSEL <pie...@po...>:

> Hello,
> Does this configuration work for LSR1 ?
> If not, did you enable forwarding on LSR 2 ? I don't exactly remember
> the
> command, but it is something like
> echo "1">/proc/net/sys/ipv4/all/forwarding
> 
> Moreover, do you bind the label 17 removing with the label 18 pushing ?
> You have to add
> mplsadm2 -B -I gen:17:0 -O 0x2
> and
> mplsadm2 -B -I gen:19:0 -O 0x2
> 
> Two points I don't understand clearly in your commands :
> - Why do you have the same 0x2 key for the different label pushings ?
> - Why do you use gen:19:0 whereas you use a label space of 1 ?
> 
> I hope this will help you.
> 
> Pierre
 
Hello,

This is the scenario, I want to be able to receive the response of a ping using 
the same physical path but a different LSP. For that I did the following on my 
LSR2

My configuration is something like this

 PC1          Ingress          LSR1        LSR2     Egress             PC2

  |-----------|    |-----------|  |--------|  |------|  |---------------|
192.168.10.3  .2 172.16.70.1   .10 .11    .30 .31   .50 192.168.11.1   .2
              e0   e1          e0 e1      e0  e1    e0  e1

I can see traffic on e1 of LSR2 and e0 of Egress with labels 18 and 19 but I am 
unable to forward traffic with label 19 on to e0 of LSR2 with label 20 using 
LSP1. I want to use LSP0 for icmp echo request and LSP1 for icmp echo response. 
Kindly tell me if something is wrong with my configuration on the above 
mentioned LSR.

Sorry for the confusion, actually the configuration on LSR2 is like this

./mplsadm2 -L eth0:0
./mplsadm2 -A -I gen:17:0
./mplsadm2 -A -O 0
./mplsadm2 -O 0x2 -o push:gen:18:set:eth1:ipv4:172.16.70.50
./mplsadm2 -B -I gen:17:0 -O 0x2 

./mplsadm2 -L eth1:1
./mplsadm2 -A -I gen:19:1
./mplsadm2 -A -O 0
./mplsadm2 -O 0x3 -o push:gen:20:set:eth0:ipv4:172.16.70.11
./mplsadm2 -B -I gen:19:1 -O 0x3

In my quest to send the mail quickly I typed the wrong configuration. Anyway, I 
am able to send packets from PC1 to PC2 using LSP0 but I am unable to forward 
the response from PC2 on LSP1 beyond LSR2. To clearify a bit further I have the 
following configuration at the Egress

./mplsadm2 -L eth0:0
./mplsadm2 -A -I gen:18:0 -i pop:peek
./mplsadm2 -A -O 0
./mplsadm2 -O 0x2 -o push:gen:19:set:eth0:ipv4:172.16.70.31

Looking forward to your useful input once again.

Re: [mpls-linux-general] RSVP-TE Question

[baim <j0...@te...>]

>First of all, why would you expect there to be a 
>difference in performance?
>Both senerios use MPLS forwarding.  The only difference 
>maybe the presence
>of EXP to queue mappings being done in the MPLS + RSVP-TE 
>case, which may or
>may not result in a performace difference.
I think there is a difference is performance since MPLS 
provide layer2 forwarding and RSVP providing QoS for 
trafic flow.

>Second, read mpls-linux/utils/README for info about 
>binding FECs to LSPs.
>Or take a look at my example on http://mpls-linux.sf.net/
Thanx for the answer i have buld the lsp using rsvpd 
daemon from the dsmpls site.
===========================================================================================
Netkuis Instan untuk wilayah Bandung (kode area 022) - SD,SMP,SMA berhadiah total puluhan juta rupiah... periode I dimulai 1 April 2004
===========================================================================================

[mpls-linux-general] Reserve Bandwidth Problem.. ?

[baim <j0...@te...>]

Hi.. Pim, Jan, all...

I had read the documentation of rsvp-te installation from 
dsmpls.rug.ac.be and successfull installing the packet. 
But when i run The rsvp like the manual told me :

./rsvpd -D at all node and router
rapirecv_auto at egress
rtest2 at ingress
The process seems OK and i got no error message. But since 
RSVP is used for reserving a bandwidth i wonder how to di 
that ?

But i wonder how to reserve a bandwidth... ? 
Is there any command that i had run... ? or may be i have 
to install more packet for reserving the bandwidth.. ?

I hope you not scruple with my question since i am a 
newbie at this such things.
===========================================================================================
Netkuis Instan untuk wilayah Bandung (kode area 022) - SD,SMP,SMA berhadiah total puluhan juta rupiah... periode I dimulai 1 April 2004
===========================================================================================

[mpls-linux-general] Problems using libipq/QUEUE with MPLS

[Mohammad R. S. <sa...@cc...>]

I am having problems running libipq/ QUEUE with ur MPLS implementation. My 
libipq program successfully
compiles. I add the following rule to the iptables alongwith other iptables 
rule necessary for the MPLS setup:

iptables -A OUTPUT -p icmp -j QUEUE

Now this should queue all originating ping packets and then depending on the
NF mark in my libipq program, action should be taken. But it seems that 
regardless of the NF mark(NF_DROP or NF_ACCEPT), nothing seems to happen. If i 
say NF_DROP in my libipq program, still icmp packets traverse through the 
network from my ingress router. Basically I want to queue specific packets and 
then let my libipq program decide what to do. If accepted then send them to 
the outgoing LSP using spec_nh instruction of MPLS. 

Furthermore, when I run the modprobe commands to check for ipatble_filter and 
ip_queue, I get a fatal message stating "module not found". Could you please 
tell me where I am wrong. I am using the libipq example program provided in 
the manpage.

Regards,

[mpls-linux-general] Two LSPs on the same physical path

[Itrat R. Q. <iq...@cc...>]

Hello,

This is the scenario, I want to be able to receive the response of a ping 
using the same physical path but a different LSP. For that I did the following 
on my LSR2

./mplsadm2 -L eth0:0
./mplsadm2 -A -I gen:17:0
./mplsadm2 -A -O 0
./mplsadm2 -O 0x2 -o push:gen:18:set:eth1:ipv4:172.16.70.50

./mplsadm2 -L eth1:1
./mplsadm2 -A -I gen:19:0
./mplsadm2 -A -O 0
./mplsadm2 -O 0x2 -o push:gen:20:set:eth0:ipv4:172.16.70.11

My configuration is something like this

  PC          Ingress           LSR1        LSR2    Egress              PC
   
   |-----------|    |-----------|  |--------|  |------|  |---------------|
192.168.10.3  .2 172.16.70.1   .10 .11    .30 .31    .50 192.168.11.1   .2 
              e0    e1          e0 e1      e0  e1    e0  e1 

I can see traffic on e1 of LSR2 and e0 of Egress with labels 18 and 19 but I 
am unable to forward traffic with label 19 on to e0 of LSR2 with label 20 
using LSP1. I want to use LSP0 for icmp echo request and LSP1 for icmp echo 
response. Kindly tell me if something is wrong with my configuration on the 
above mentioned LSR.

Looking forward to your help.

Re: [mpls-linux-general] Forwarding MPLS traffic on to the IP interface

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

On Tue, Jun 29, 2004 at 05:31:16PM +0300, Itrat Rasod Quadri wrote:
> Hello,
> 
> I want to know what iptable commands to use in order to route traffic from an 
> MPLS interface to an IP interface.

Do you want the packet to under go a IPv4 route lookup or do you want to
force the traffic to a specific nexthop?

> Thanks in advance.
> 
> 
> -------------------------------------------------------
> This SF.Net email sponsored by Black Hat Briefings & Training.
> Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
> digital self defense, top technical experts, no vendor pitches, 
> unmatched networking opportunities. Visit www.blackhat.com
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

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

[mpls-linux-general] Forwarding MPLS traffic on to the IP interface

[Itrat R. Q. <iq...@cc...>]

Hello,

I want to know what iptable commands to use in order to route traffic from an 
MPLS interface to an IP interface.

Thanks in advance.

[mpls-linux-general] Forwarding MPLS traffic onto IP interface

[Itrat R. Q. <iq...@cc...>]

Hello,

I want to know what iptable commands to use in order to route traffic from an 
MPLS interface to an IP interface.

Thanks in advance.

Re: [mpls-linux-general] RSVP-TE Question

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

On Tue, Jun 29, 2004 at 10:35:27AM +0700, baim wrote:
> Hi... i had installing rsvp-te for diffserv over mpls 
> version 0.70-rc2 successfully, and now i want to comparing 
> the performance of a network that only using MPLS vs MPLS 
> + RSVP.
> This is my network configuration :

First of all, why would you expect there to be a difference in performance?
Both senerios use MPLS forwarding.  The only difference maybe the presence
of EXP to queue mappings being done in the MPLS + RSVP-TE case, which may or
may not result in a performace difference.

Second, read mpls-linux/utils/README for info about binding FECs to LSPs.
Or take a look at my example on http://mpls-linux.sf.net/

> host1|--|eth0=R1=eth1|--|eth0=R2=eth1|--|R3|--|R4|--|host2
> at host1 :
> eth0 = 10.10.10.1
> at R1 :
> eth0 = 10.10.10.2
> eth1 = 11.11.11.1
> at R2 :
> eth0 = 11.11.11.2
> eth1 = 12.12.12.1
> 
> But a got a problem when i try to activate only MPLS at R1 
> and R2. Here what i done :
> At host1 :
> #route add -net 11.11.11.0 gw 10.10.10.2 netmask 
> 255.255.255.0
> #route add -net 12.12.12.0 gw 10.10.10.2 netmask 
> 255.255.255.0
> At R1 :
> #mplsadm -A -O gen:100:eth1:ipv4:11.11.11.2
> #mplsadm -A -I gen:110:0
> #mplsadm -L eth1:0
> #route add -host 11.11.11.2 gw 11.11.11.2
> #route add -host 12.12.12.1 gw 11.11.11.2
> #route add -host 10.10.10.1 gw 10.10.10.1
> At R2 :
> #mplsadm -A -O gen:110:eth0:ipv4:11.11.11.1
> #mplsadm -A -I gen:100:0
> #mplsadm -L eth0:0
> #route add -host 11.11.11.1 gw 11.11.11.1
> #route add -host 10.10.10.2 gw 11.11.11.1
> #route add -host 10.10.10.1 gw 11.11.11.1
> 
> there nothing error message with the configuration but 
> when i try to ping from host1 to the 11.11.11.2 (R2) there 
> is only a message : 
> #ping 11.11.11.2
> PING 11.11.11.2 (11.11.11.2) from 10.10.10.1 : 56(84) 
> bytes of data.
> I am using Redhat 7.1, installing RSVP-TE with the package 
> and guide from dsmpls.atlantis.rug.ac.be.
> 
> My question is:
> - What's wrong with my configuration ?
> - The ip_forward is 0. Should i change it to 1 ?
> - MPLS using layer 2 forwarding, so i do not need to make 
> the ip_forward = 1. Am I wrong.. ?
> - Can you give me the right mpls configuration with that 
> network configuration. ?
> ===========================================================================================
> Netkuis Instan untuk wilayah Bandung (kode area 022) - SD,SMP,SMA berhadiah 
> total puluhan juta rupiah... periode I dimulai 1 April 2004
> ===========================================================================================
> 
> 
> -------------------------------------------------------
> This SF.Net email sponsored by Black Hat Briefings & Training.
> Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
> digital self defense, top technical experts, no vendor pitches, 
> unmatched networking opportunities. Visit www.blackhat.com
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

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

[mpls-linux-general] RSVP-TE Question

[baim <j0...@te...>]

Hi... i had installing rsvp-te for diffserv over mpls 
version 0.70-rc2 successfully, and now i want to comparing 
the performance of a network that only using MPLS vs MPLS 
+ RSVP.
This is my network configuration :

host1|--|eth0=R1=eth1|--|eth0=R2=eth1|--|R3|--|R4|--|host2
at host1 :
eth0 = 10.10.10.1
at R1 :
eth0 = 10.10.10.2
eth1 = 11.11.11.1
at R2 :
eth0 = 11.11.11.2
eth1 = 12.12.12.1

But a got a problem when i try to activate only MPLS at R1 
and R2. Here what i done :
At host1 :
#route add -net 11.11.11.0 gw 10.10.10.2 netmask 
255.255.255.0
#route add -net 12.12.12.0 gw 10.10.10.2 netmask 
255.255.255.0
At R1 :
#mplsadm -A -O gen:100:eth1:ipv4:11.11.11.2
#mplsadm -A -I gen:110:0
#mplsadm -L eth1:0
#route add -host 11.11.11.2 gw 11.11.11.2
#route add -host 12.12.12.1 gw 11.11.11.2
#route add -host 10.10.10.1 gw 10.10.10.1
At R2 :
#mplsadm -A -O gen:110:eth0:ipv4:11.11.11.1
#mplsadm -A -I gen:100:0
#mplsadm -L eth0:0
#route add -host 11.11.11.1 gw 11.11.11.1
#route add -host 10.10.10.2 gw 11.11.11.1
#route add -host 10.10.10.1 gw 11.11.11.1

there nothing error message with the configuration but 
when i try to ping from host1 to the 11.11.11.2 (R2) there 
is only a message : 
#ping 11.11.11.2
PING 11.11.11.2 (11.11.11.2) from 10.10.10.1 : 56(84) 
bytes of data.
I am using Redhat 7.1, installing RSVP-TE with the package 
and guide from dsmpls.atlantis.rug.ac.be.

My question is:
- What's wrong with my configuration ?
- The ip_forward is 0. Should i change it to 1 ?
- MPLS using layer 2 forwarding, so i do not need to make 
the ip_forward = 1. Am I wrong.. ?
- Can you give me the right mpls configuration with that 
network configuration. ?
===========================================================================================
Netkuis Instan untuk wilayah Bandung (kode area 022) - SD,SMP,SMA berhadiah total puluhan juta rupiah... periode I dimulai 1 April 2004
===========================================================================================

[mpls-linux-general] Question on outgoing_label table

[Helen \(Wenjing\) Z. <wzh...@si...>]

I am learning RSVP-TE. From one of the intermediate router, I got the
outgoing_label table(/proc/net/mpls_out) like this:

0x00000005 0/0/0 1 EXP2TC( EXP(0)->TC(ffff) EXP(1)->TC(00b8)
EXP(2)->TC(0028) EXP(3)->TC(0030) EXP(4)->TC(0048) EXP(5)->TC(0050)
EXP(6)->TC(0068) EXP(7)->TC(0070) ) PUSH(gen 11680) SET(eth1,192.168.4.2)

0x0000000c 0/0/0 1 EXP2TC( EXP(0)->TC(ffff) EXP(1)->TC(00b8)
EXP(2)->TC(0028) EXP(3)->TC(0030) EXP(4)->TC(0048) EXP(5)->TC(0050)
EXP(6)->TC(0068) EXP(7)->TC(0070) ) PUSH(gen 11680) SET(eth2,192.168.5.2)

0x00000018 0/0/0 1 EXP2TC( EXP(0)->TC(ffff) EXP(1)->TC(00b8)
EXP(2)->TC(0028) EXP(3)->TC(0030) EXP(4)->TC(0048) EXP(5)->TC(0050)
EXP(6)->TC(0068) EXP(7)->TC(0070) ) PUSH(gen 11680) SET(eth1,192.168.4.2)

0x0000004b 0/0/0 1 EXP2TC( EXP(0)->TC(ffff) EXP(1)->TC(00b8)
EXP(2)->TC(0028) EXP(3)->TC(0030) EXP(4)->TC(0048) EXP(5)->TC(0050)
EXP(6)->TC(0068) EXP(7)->TC(0070) ) PUSH(gen 11680) SET(eth1,192.168.4.2)

0x00000067 0/0/0 2 EXP2TC( EXP(0)->TC(ffff) EXP(1)->TC(00b8)
EXP(2)->TC(0028) EXP(3)->TC(0030) EXP(4)->TC(0048) EXP(5)->TC(0050)
EXP(6)->TC(0068) EXP(7)->TC(0070) ) PUSH(gen 11680) SET(eth2,192.168.5.2)

There are 5 entries in this table, but only the last one with the KEY
"0x00000067" was working for the current process. I checked the
outgoing_label table again while I restarted that computer but didn't run
the RSVP-TE daemon, the first 4 entries were still there. When I started the
daemon, a new label entry came to the 5th with a fresh new KEY.

My questions are: Why there are some entries always reside in that label
table even without running RSVP-TE daemon? How can I delete those entries?

Thanks.

Helen

[mpls-linux-general] Question on outgoing_label KEY

[Helen \(Wenjing\) Z. <wzh...@si...>]

In /proc/net/mpls_out I got:

0x00000059 0/0/0 2 DS2EXP( DS(00)->EXP(00) DS(0a)->EXP(02) DS(0c)->EXP(03)
DS(12)->EXP(04) DS(14)->EXP(05) DS(1a)->EXP(06) DS(1c)->EXP(07)
DS(2e)->EXP(01) ) EXP2TC( EXP(0)->TC(ffff) EXP(1)->TC(00b8) EXP(2)->TC(0028)
EXP(3)->TC(0030) EXP(4)->TC(0048) EXP(5)->TC(0050) EXP(6)->TC(0068)
EXP(7)->TC(0070) ) PUSH(gen 11680) SET(eth0,192.168.2.2)

From the reading material, "0x00000059" is an outgoing _label KEY.

What is the relationship with Label, LSP ID and the outgoing_label KEY?
How to assign this KEY?
How and where can I find the value for this KEY except the outgoing label
table?

Thanks.

Helen

Re: [mpls-linux-general] configuring mpls through ioctl()

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

On Wed, Jun 23, 2004 at 06:09:58PM +0300, Muhammad R. Sami wrote:
> Can I issue ioctl system calls from within my libipq application program to
> configure MPLS routing from my application, if yes, can anyone please
> provide me with some example code on how to perform this task.
> Regards,

Yes you can.  All the example code you should need is in
mpls-linux/utils/mplsadm.c

> Muhammad R. Sami 
> Research Assistant, 
> Computer Engineering Department 
> P.O.Box 354 
> King Fahd University of Petroleum & Minerals 
> Dhahran 31261 
> Saudi Arabia. 
> Tel: +96638601423 
> Cell: +96657982951 
> www.ccse.kfupm.edu.sa/sami
> 
> -----Original Message-----
> From: James R. Leu [mailto:jl...@mi...] 
> Sent: Tuesday, June 22, 2004 7:06 PM
> To: Muhammad R. Sami
> Subject: Re: [mpls-linux-general] packet identification
> 
> All of the mplsadm features can be called via ioctl().  Looking at
> the code for mplsadm will give you an example of how to do it.
> 
> On Tue, Jun 22, 2004 at 06:58:27PM +0300, Muhammad R. Sami wrote:
> > I want to write an application that can issue mplsadm commands from
> within,
> > such that I do not have to configure MPLS (labels, LSP, EXP) manually.
> Here
> > are the steps:
> > 
> > 1- Receive packet from kernel to my application through libipq
> > 2- Check packet payload and then decide on NF_ACCEPT/NF_DROP.
> > 3- If NF_ACCEPT, then configure MPLS by issuing the mplsadm instructions
> > from within the application for all accepted packets.
> > 
> > I might also need to set queuing priority here on the basis of EXP value.
> > 
> > Regards,
> > 
> > Muhammad R. Sami 
> > Research Assistant, 
> > Computer Engineering Department 
> > P.O.Box 354 
> > King Fahd University of Petroleum & Minerals 
> > Dhahran 31261 
> > Saudi Arabia. 
> > Tel: +96638601423 
> > Cell: +96657982951 
> > www.ccse.kfupm.edu.sa/sami
> > -----Original Message-----
> > From: James R. Leu [mailto:jl...@mi...] 
> > Sent: Tuesday, June 22, 2004 6:37 PM
> > To: Muhammad R. Sami
> > Subject: Re: [mpls-linux-general] packet identification
> > 
> > On Tue, Jun 22, 2004 at 06:18:11PM +0300, Muhammad R. Sami wrote:
> > > I am using libipq to manipulate packets at userspace. I want to know
> > whether
> > > I can issue an mplsadm configuration command from within my application.
> > For
> > > example, after receiving the packet from the kernel through libipq, I
> > would
> > > like to set EXP value for all such packets by issuing the set_exp
> command
> > > from within that application. If this is possible, please let me know
> > soon.
> > > regards
> > 
> > I'm not sure what you're asking.  Are you trying to manipulate the
> > packet buffer in userland then send it?  Or are you trying to configure
> the
> > forwarding plan to do the manipulation?
> > 
> > 
> > > Muhammad R. Sami 
> > > Research Assistant, 
> > > Computer Engineering Department 
> > > P.O.Box 354 
> > > King Fahd University of Petroleum & Minerals 
> > > Dhahran 31261 
> > > Saudi Arabia. 
> > > Tel: +96638601423 
> > > Cell: +96657982951 
> > > www.ccse.kfupm.edu.sa/sami
> > > -----Original Message-----
> > > From: James R. Leu [mailto:jl...@mi...] 
> > > Sent: Monday, June 21, 2004 5:11 PM
> > > To: Muhammad R. Sami
> > > Subject: Re: [mpls-linux-general] packet identification
> > > 
> > > On Sun, Jun 20, 2004 at 12:48:24PM +0300, Muhammad R. Sami wrote:
> > > > You outlined the following steps, require assistance for the
> following.
> > > > 
> > > >  
> > > > 
> > > > A quick overview of what you will need to setup:
> > > > 
> > > > -bind a IPv4 route to a LSP-----DONE
> > > > 
> > > > -use iptables to identify and mark packets----DONE
> > > > 
> > > > -use the markings to set EXP values (or choose differnt LSPs)----DONE
> > > > 
> > > > -use EXP value to determin queuing priority--------How to do this
> part?
> > > How
> > > > can I assign priority on the basis of EXP bits value?
> > > 
> > > Use the exp2tc or exp2ds instructions then use 'tc' to setup the queues
> > and
> > > priorities.
> > > 
> > > Look at:
> > >
> >
> http://www.linuxsecurity.com/resource_files/firewalls/Adv-Routing-HOWTO/2.4r
> > > outing.html
> > > 
> > > In particular read how to use the TCINDEX.
> > > 
> > > > 
> > > >  
> > > > 
> > > 
> > > -- 
> > > James R. Leu
> > > jl...@mi...
> > > 
> > 
> > -- 
> > James R. Leu
> > jl...@mi...
> > 
> 
> -- 
> James R. Leu
> jl...@mi...
> 

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

[mpls-linux-general] Problem with MPLS under Linux :(

[Gomez A. <rod...@po...>]

Hello,
I've been trying MPLS for about 2 weeks and always unseccessfull.

I was using patches 1.172, 1.992 with kernels:
2.4.20, 2.4.2, 2.4.18, 2.6.1, 2.6.7, 2.6.4 under RH 9 and MDK 10.

I allways get error during compilation kernel or modules.
Ususlly it is:
../net/ip4/fib_semantics.c
unknown: RTA_SPEC_DATA, RTA_SPEC_PROTO

Does anyone did this stuff?
I was trying everybcombination and it ddoes not work :(

Robert

[mpls-linux-general] configuring mpls through ioctl()

[Muhammad R. S. <sa...@cc...>]

Can I issue ioctl system calls from within my libipq application program to
configure MPLS routing from my application, if yes, can anyone please
provide me with some example code on how to perform this task.
Regards,


Muhammad R. Sami 
Research Assistant, 
Computer Engineering Department 
P.O.Box 354 
King Fahd University of Petroleum & Minerals 
Dhahran 31261 
Saudi Arabia. 
Tel: +96638601423 
Cell: +96657982951 
www.ccse.kfupm.edu.sa/sami

-----Original Message-----
From: James R. Leu [mailto:jl...@mi...] 
Sent: Tuesday, June 22, 2004 7:06 PM
To: Muhammad R. Sami
Subject: Re: [mpls-linux-general] packet identification

All of the mplsadm features can be called via ioctl().  Looking at
the code for mplsadm will give you an example of how to do it.

On Tue, Jun 22, 2004 at 06:58:27PM +0300, Muhammad R. Sami wrote:
> I want to write an application that can issue mplsadm commands from
within,
> such that I do not have to configure MPLS (labels, LSP, EXP) manually.
Here
> are the steps:
> 
> 1- Receive packet from kernel to my application through libipq
> 2- Check packet payload and then decide on NF_ACCEPT/NF_DROP.
> 3- If NF_ACCEPT, then configure MPLS by issuing the mplsadm instructions
> from within the application for all accepted packets.
> 
> I might also need to set queuing priority here on the basis of EXP value.
> 
> Regards,
> 
> Muhammad R. Sami 
> Research Assistant, 
> Computer Engineering Department 
> P.O.Box 354 
> King Fahd University of Petroleum & Minerals 
> Dhahran 31261 
> Saudi Arabia. 
> Tel: +96638601423 
> Cell: +96657982951 
> www.ccse.kfupm.edu.sa/sami
> -----Original Message-----
> From: James R. Leu [mailto:jl...@mi...] 
> Sent: Tuesday, June 22, 2004 6:37 PM
> To: Muhammad R. Sami
> Subject: Re: [mpls-linux-general] packet identification
> 
> On Tue, Jun 22, 2004 at 06:18:11PM +0300, Muhammad R. Sami wrote:
> > I am using libipq to manipulate packets at userspace. I want to know
> whether
> > I can issue an mplsadm configuration command from within my application.
> For
> > example, after receiving the packet from the kernel through libipq, I
> would
> > like to set EXP value for all such packets by issuing the set_exp
command
> > from within that application. If this is possible, please let me know
> soon.
> > regards
> 
> I'm not sure what you're asking.  Are you trying to manipulate the
> packet buffer in userland then send it?  Or are you trying to configure
the
> forwarding plan to do the manipulation?
> 
> 
> > Muhammad R. Sami 
> > Research Assistant, 
> > Computer Engineering Department 
> > P.O.Box 354 
> > King Fahd University of Petroleum & Minerals 
> > Dhahran 31261 
> > Saudi Arabia. 
> > Tel: +96638601423 
> > Cell: +96657982951 
> > www.ccse.kfupm.edu.sa/sami
> > -----Original Message-----
> > From: James R. Leu [mailto:jl...@mi...] 
> > Sent: Monday, June 21, 2004 5:11 PM
> > To: Muhammad R. Sami
> > Subject: Re: [mpls-linux-general] packet identification
> > 
> > On Sun, Jun 20, 2004 at 12:48:24PM +0300, Muhammad R. Sami wrote:
> > > You outlined the following steps, require assistance for the
following.
> > > 
> > >  
> > > 
> > > A quick overview of what you will need to setup:
> > > 
> > > -bind a IPv4 route to a LSP-----DONE
> > > 
> > > -use iptables to identify and mark packets----DONE
> > > 
> > > -use the markings to set EXP values (or choose differnt LSPs)----DONE
> > > 
> > > -use EXP value to determin queuing priority--------How to do this
part?
> > How
> > > can I assign priority on the basis of EXP bits value?
> > 
> > Use the exp2tc or exp2ds instructions then use 'tc' to setup the queues
> and
> > priorities.
> > 
> > Look at:
> >
>
http://www.linuxsecurity.com/resource_files/firewalls/Adv-Routing-HOWTO/2.4r
> > outing.html
> > 
> > In particular read how to use the TCINDEX.
> > 
> > > 
> > >  
> > > 
> > 
> > -- 
> > James R. Leu
> > jl...@mi...
> > 
> 
> -- 
> James R. Leu
> jl...@mi...
> 

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

Re: [mpls-linux-general] what kernel for Implementating IPv6 over MPLS

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

I have code for ipv6 over MPLS in the patch based on the 2.6 linux
kernel.  It is still untested.  If you decide to go this route you will
have to help me test it.  I will do my best to provide you fixes.

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

On Tue, Jun 22, 2004 at 08:13:16AM +0700, hermansyah wrote:
> hello,
> 
> I need your help for my final project regarding to MPLS 
> while the title is IPv6 over MPLS. I have been installed 
> MPLS on Redhat Kernel 2.4.20-8 is it already supported for 
> Implementating IPv6 over MPLS? or should I up grade its 
> kernel? and what module still I need for implementating 
> IPv6 over MPLS? 
> ===========================================================================================
> "Gabung INSTANIA, dapatkan XENIA. Daftar di www.telkomnetinstan.com, 
> langsung dapat akses Internet Gratis..
> Dan ..ikuti "Instan Smile" berhadiah Xenia,Tour S'pore, Komputer,dll, info 
> hub : TELKOM Jatim 0-800-1-467826 "
> ===========================================================================================
> 
> 
> -------------------------------------------------------
> This SF.Net email sponsored by Black Hat Briefings & Training.
> Attend Black Hat Briefings & Training, Las Vegas July 24-29 - 
> digital self defense, top technical experts, no vendor pitches, 
> unmatched networking opportunities. Visit www.blackhat.com
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

[mpls-linux-general] what kernel for Implementating IPv6 over MPLS

[hermansyah <her...@pl...>]

hello,

I need your help for my final project regarding to MPLS 
while the title is IPv6 over MPLS. I have been installed 
MPLS on Redhat Kernel 2.4.20-8 is it already supported for 
Implementating IPv6 over MPLS? or should I up grade its 
kernel? and what module still I need for implementating 
 IPv6 over MPLS? 
===========================================================================================
"Gabung INSTANIA, dapatkan XENIA. Daftar di www.telkomnetinstan.com, langsung dapat akses Internet Gratis..
Dan ..ikuti "Instan Smile" berhadiah Xenia,Tour S'pore, Komputer,dll, info hub : TELKOM Jatim 0-800-1-467826 "
===========================================================================================

Re: [mpls-linux-general] NF mark and exp bits

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

On Mon, Jun 21, 2004 at 03:08:20PM +0300, Itrat Rasod Quadri wrote:
> Hello,
> 
> I would like to ask you if its possible to assign exp bits based on NF mark 
> and once exp bits are set can we assign priorities based on those exp bit 
> settings. If so then kindly provide some information on how it can be done.

You can use the nf2exp instruction or the nffwd instruction to set the EXP
bits based on the NF mark.

On the next hop you can use the exp2tc to set the TC index based on the EXP
bits.

Read:

http://www.linuxsecurity.com/resource_files/firewalls/Adv-Routing-HOWTO/2.4routing.html

To learn how to use TCINDEX to affect queuing.


> -------------------------------------------------------
> This SF.Net email is sponsored by The 2004 JavaOne(SM) Conference
> Learn from the experts at JavaOne(SM), Sun's Worldwide Java Developer
> Conference, June 28 - July 1 at the Moscone Center in San Francisco, CA
> REGISTER AND SAVE! http://java.sun.com/javaone/sf Priority Code NWMGYKND
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

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

[mpls-linux-general] NF mark and exp bits

[Itrat R. Q. <iq...@cc...>]

Hello,

I would like to ask you if its possible to assign exp bits based on NF mark 
and once exp bits are set can we assign priorities based on those exp bit 
settings. If so then kindly provide some information on how it can be done.

Re: [mpls-linux-general] TCP checksum problem with the MPLS linux 2.6 patch

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

No known issue.  Make sure that you are not using the chksum offloading
on the NICs.

On Mon, Jun 21, 2004 at 12:57:59PM +0200, Pierre ANSEL wrote:
> Hello !
> 
> We set up an MPLS path between two linux boxes. The patch was applied on a 
> 2.6.3 kernel. We're able to ping a host from the other. But our attemps to 
> establish a TCP connection (for FTP or SSH) don't work : the connection is 
> set up correctly (SYN / SYN&ACK / ACK), but the following packets (data 
> packets) have a wrong TCP checksum ! Ethereal, both on the server and on the 
> client, shows that the checksum is bad on both hosts (emission and 
> reception). We use 3COM 905C NICs...
> 
> Is there a known issue ? Thanks a lot for your help !
> 
> Regards,
> Pierre Ansel.
> 
> 
> -------------------------------------------------------
> This SF.Net email is sponsored by The 2004 JavaOne(SM) Conference
> Learn from the experts at JavaOne(SM), Sun's Worldwide Java Developer
> Conference, June 28 - July 1 at the Moscone Center in San Francisco, CA
> REGISTER AND SAVE! http://java.sun.com/javaone/sf Priority Code NWMGYKND
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

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