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

Re: [mpls-linux-general] mpls beta+usage

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

I figured it out.  It's not implemented yet :-)  Well kind of.
I have implemented the code to build in the internal instruction, but
the instruction isn't handling in the forwarind plane.

I will implement it and release a new patch.

Thanks for catching this :-)

Jim

On Wed, Jan 16, 2002 at 04:27:54PM -0500, ncc49a wrote:
> HI James,
> 
>   I have actually tried the below already:
> 
> >>
>  $MPLSADM -v -A -B -O gen:100:eth0 -o 
> > ds2exp:0x0:0x1a:0x1:push:gen:100:set:eth0:ipv4:$PC_C_IP -f 
>          ^^^
> Change 0x0 to 0xFF and see if that helps.
> <<
> 
>  But, I still can't get it to work. Below is the console output when mplsadm2 
> is being executed in verbose mode.
> 
> Label Space input: eth0:1
> If: eth0 LS: 1
> Out label input: gen:16:eth0
> Out instr input: ds2exp:0xff:0x1a:0x01:push:gen:16:set:eth0:ipv4:192.168.0.1
> FEC input: 192.168.0.1/32
> Label type: gen
> Instruction: ds2exp
> DS: 0xff 255
> EXP: 0x1a 26
> DS: 0x01 1
> Instruction: push
> Label type: gen
> Instruction: set
> Nexthop protocol: ipv4
> Length: 3
> 
>   The EXP bits in the mpls shim header are still not set when packets with DS 
> field 0x68 and DSCP bits 0x1a are sent through the mpls ingress.
> 
>    I will also like to ask if the mask 0xff is applied to the DS field or the 
> DSCP bits....
> 
>    Thanks a lot!
> 
> Paul
> 
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu

[mpls-linux-general] =?gb2312?B?TGFjayBsaWJsZHBsaW51eC5hPw==?=

[ÕÅΰһ <zh...@16...>]

Hi,All,
  When I complied in the direction "ldp-portable-0.060",the compiler always core dump.I found it seems that the compiler could not find the libldplinux.a.Did it result in the failure of compile?

Waiting for your help.

Yours Sincerely 

Zhang,Weiyi




=============================================================
http://dating.163.com       ´ºÌ컨»á¿ª£¬Äк¢Å®º¢ÒªÁµ°®¡«
http://im.163.com           ÍøÒ׶ÌÐÅͨ£ºÃâ·Ñ·¢ËÍÊÖ»ú¶ÌÐŵÄ×ÀÃæÈí¼þ
http://vip.163.com          ÍøÒ×VIPÓÊÏä 30ÃëÁ¢µÃ

[mpls-linux-general] mpls beta+usage

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

HI James,

  I have actually tried the below already:

>>
 $MPLSADM -v -A -B -O gen:100:eth0 -o 
> ds2exp:0x0:0x1a:0x1:push:gen:100:set:eth0:ipv4:$PC_C_IP -f 
         ^^^
Change 0x0 to 0xFF and see if that helps.
<<

 But, I still can't get it to work. Below is the console output when mplsadm2 
is being executed in verbose mode.

Label Space input: eth0:1
If: eth0 LS: 1
Out label input: gen:16:eth0
Out instr input: ds2exp:0xff:0x1a:0x01:push:gen:16:set:eth0:ipv4:192.168.0.1
FEC input: 192.168.0.1/32
Label type: gen
Instruction: ds2exp
DS: 0xff 255
EXP: 0x1a 26
DS: 0x01 1
Instruction: push
Label type: gen
Instruction: set
Nexthop protocol: ipv4
Length: 3

  The EXP bits in the mpls shim header are still not set when packets with DS 
field 0x68 and DSCP bits 0x1a are sent through the mpls ingress.

   I will also like to ask if the mask 0xff is applied to the DS field or the 
DSCP bits....

   Thanks a lot!

Paul

Re: [mpls-linux-general] Mpls-beta + usage

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

> $MPLSADM -v -A -B -O gen:100:eth0 -o 
> ds2exp:0x0:0x1a:0x1:push:gen:100:set:eth0:ipv4:$PC_C_IP -f 
         ^^^
Change 0x0 to 0xFF and see if that helps.

> $PC_C_IP/32
> 
>   I sniff at PC C using Ethereal.
> 
>   From the above, I think I should get all outgoing packets 
> from PC B to C having EXP field marked with 1, right? But, 
> I don't.
> 
>   Will appreciate any advice provided.
> 
>   Thanks!
> 
> Paul
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu

[mpls-linux-general] Mpls-beta + usage

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

Hi,

  I am trying out the beta version of mpls extracted from 
the cvs. I haev 3 PCs linked together in the following 
config.

     eth0   eth1 eth0   eth0
  PC A ------ PC B ------ PC C

  PC A marks all outgoing packets with a Diffserv field of 
0x68 and a DSCP of 0x1a.

  PC B acts as an mpls ingress. The configuration script is 
as such:

$MPLSADM -v -L eth0:

$MPLSADM -v -A -B -O gen:100:eth0 -o 
ds2exp:0x0:0x1a:0x1:push:gen:100:set:eth0:ipv4:$PC_C_IP -f 
$PC_C_IP/32

  I sniff at PC C using Ethereal.

  From the above, I think I should get all outgoing packets 
from PC B to C having EXP field marked with 1, right? But, 
I don't.

  Will appreciate any advice provided.

  Thanks!

Paul

[mpls-linux-general] regarding traffic engg...

[<fr...@hs...>]

Hi James,
              first of all let me take the opportunity to appreciate the efforts
 put in implementing
              mpls.

     Actually we want to simulate the traffic Engineering by setting up
   two LSP's between same HOST and DESTINATION. We want to do with transport
   layer port.

     For same DESTINATION , based on Tranport Layer port, we are trying
   to set up two LSP's.

    Actually here the problem is , for same fec, whichever binding info
   is given last is remaining in the
     rt->u.dst.dst_proto_data[AUX_PROTO_DATA_MPLS] =
     res->proto_data[AUX_PROTO_DATA_MPLS];
     /* from route.c file : function :rt_set_nexthop    */
     I think due to this limitation, it is getting the recently configured
   NHLFE info , for same FEC's.

mplsadm -v -A -B -O gen:20:eth1:ipv4:192.168.210.1 -f 192.168.206.12/32
mplsadm -v -A -B -O gen:21:ppp0:ipv4:192.168.210.1 -f 192.168.206.12/32

    In the above case , for FEC 192.168.206.12 , it is always going through
 ppp0 interface.

    I have a doubt that information about "eth1" for 192.168.206.12/32 might
  be overlapping with "ppp0". as i noticed so many times , it is always going
  through  "ppp0".

    If i do it in reverse order , it is doing vice versa.
    You are using platform based labelspace, for that can't we bind the same FEC
  to two different interfaces?

    can't we have  more than one NHLFE for same FEC ?
Please guide in this regard.

   Thanks in advance......

Best Regards,
Faisal Rehman.


********************************************************************************
"This message is proprietary to Hughes Software Systmes Limited (HSS) and/or its
customers and intended solely for the use of the individual or organisation to
whom it is addressed. It may contain  privileged or confidential information.
If you have received this message in error, please notify the originator
immediately. If you are not the intended recipient, you are notified that you
are strictly prohibited from using, copying, altering, or disclosing the
contents of this message. HSS accepts no responsibility for loss or damage
arising from the use of the information transimitted by this email including
damage from virus."
********************************************************************************

[mpls-linux-general] Problems with ethernet card

[Daniel G. C. <da...@on...>]

Hi! I'm Daniel Gregorio a student of computer science at University of =
Valencia (Spain). I'm doing a test of MPLS functionalities so I'm using =
the MPLS-Linux 0.996 in 4 computers. I don't know a lot about Linux, so =
I will be very pleased if someone can help me.

I have a problem with the ethernet cards. When I boot the MPLS-Linux =
kernel in the boot messages doesn't appear any reference to eth0 or =
similar (the only reference is '0 3c209 cards found'). Then when I try =
to start the ethernet card with ifconfig it says that the eth0 interface =
is unknown, however when I start with the original kernel it runs ok. =
How can I fix this problem? Is there any special command in mplsadm that =
helps with these? I have to put any special options when I compile the =
kernel (different from the normal ones)?
The card is a 3com 509 etherlink ethernet card, and the driver is =
built-in in the kernel (I haven't put it as a module, perhaps this is =
the problem?), I'm running the MPLS-Linux over the linux kernel 2.4.13 =
in a Pentium classic 133Mhz processor with 32Mb RAM.

Thank you for help me! Bye

Dani

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

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

I'd just like at add one more point, if you do have an modules configured in
the kernel, make sure to do a 'make modules' and 'make modules_install'

I agree with Abhijit though, you need to figure out wht driver your ethernet
card uses and compile it into the kernel.

On Mon, Jan 14, 2002 at 11:49:08AM +0530, Abhijit Gadgil wrote:
> 
> Problem is there with the network card drivers, if you have compiled them
> as modules look for the modules path. Else u shd compile the sources for
> the network card drivers inside the kernel
> 
> (to check which network card you are using do a 'lsmod' and locate the
> installed module.) And compile its sources with the kernel sources of the
> patched-mpls sources.
> 
> -abhijit
> 
> 
> Jagdish Garg wrote :
> 
> >Hi,
> >I have downloaded the mpls-1.0 patch and applied it on Red hat 7.1 ,
> >kernel -2.4.7.
> >It got compiled successfully but when I booted from the mpls image , it
> >does not start the interface "eth0".
> >
> >While if again I try to boot it from the original image , I can see the
> >interface "etho" working.Please clarify.
> >
> >
> >Thanks and Regards
> >
> >Jagdish
> >
> >
> >_______________________________________________
> >mpls-linux-general mailing list
> >mpl...@li...
> >https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
> >
> 
> -- 
> How do u delete a letter in vi and paste it after the next one?
> 
> "xp"
> 
> --
> 
> 
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
James R. Leu

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

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

Problem is there with the network card drivers, if you have compiled them
as modules look for the modules path. Else u shd compile the sources for
the network card drivers inside the kernel

(to check which network card you are using do a 'lsmod' and locate the
installed module.) And compile its sources with the kernel sources of the
patched-mpls sources.

-abhijit


Jagdish Garg wrote :

>Hi,
>I have downloaded the mpls-1.0 patch and applied it on Red hat 7.1 ,
>kernel -2.4.7.
>It got compiled successfully but when I booted from the mpls image , it
>does not start the interface "eth0".
>
>While if again I try to boot it from the original image , I can see the
>interface "etho" working.Please clarify.
>
>
>Thanks and Regards
>
>Jagdish
>
>
>_______________________________________________
>mpls-linux-general mailing list
>mpl...@li...
>https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
>

-- 
How do u delete a letter in vi and paste it after the next one?

"xp"

--

[mpls-linux-general] (no subject)

[Jagdish G. <jag...@ie...>]

Hi,
I have downloaded the mpls-1.0 patch and applied it on Red hat 7.1 ,
kernel -2.4.7.
It got compiled successfully but when I booted from the mpls image , it
does not start the interface "eth0".

While if again I try to boot it from the original image , I can see the
interface "etho" working.Please clarify.


Thanks and Regards

Jagdish

[mpls-linux-general] Mpls for Linux Beta Version - Install + usage

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

Hi,

  To install the new beta version, do we need to enable any
other options while recompiling the kernel (other than the
ones already provided in the current docs?).

  Also, while installing iptables, are there any additional
patches which we need to apply (if when using patch-o-matic..)?

   Will also like to know if anyone has tried out the beta
version, and will like to post a simple example illustrating
some of its new features.

   Thanks a  lot!

Paul

Re: [mpls-linux-general] ldp_linux questions

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

Hello,

First thing to remember is that ldp_linux is just an example porting layer
for the ldp-portable library.  With that in mind ...

On Thu, Jan 10, 2002 at 05:53:49PM +0200, Asviyan, Anatoly wrote:
> 	hello All
> I have some questions about ldp_linux:
> Q1: does it support multiple routing tables? 

ldp-portable is routing table agnostic, ldp_linux only looks at the main
routing table (only at startup too)

> Q2: can I restrict its activity to specific table(s)?

The porting layer could be implemented in such away as to allow this.

> Q3: can I manually feed ldp_linux with destinations that I need labels for?

You can feed ldp-portable individula routes.  ldp_linux chooses to send all
routes, ldp-portable then runs them through ldp_linux's route filters, which
may or maynot reject them.

> Q4: is there some documentation for ldp_linux (besides source) ?

Short answer, no.  If your looking for documentation on ldp-portable
and it's implementation of RFC3036, you'll see it pretty strainght forward,
if you looking for information about the porting layer, look at the examples.
Also, feel free to ask me questions (via the list) and I'll do my best
to answer.

Jim

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

-- 
James R. Leu

[mpls-linux-general] ldp_linux questions

[Asviyan, A. <ana...@te...>]

	hello All
I have some questions about ldp_linux:
Q1: does it support multiple routing tables? 
Q2: can I restrict its activity to specific table(s)?
Q3: can I manually feed ldp_linux with destinations that I need labels for?
Q4: is there some documentation for ldp_linux (besides source) ?

Thanks a lot
--
Regards, Anatoly 

Re: [mpls-linux-general] mpls-linux beta version

[Steven V. d. B. <ste...@in...>]

Hi Paul,

You have to add a "-r mpls-linux_1_108" to the cvs checkout commandline to 
get the correct revision.

Best regards,
Steven

On Thursday 10 January 2002 07:37, Ng Tat-Peng Paul wrote:
> Hi,
>
>   I have been unable to find the new kernel patch + iptables patch in the
> cvs.
>
>    Also, is the new kernel patch for kernel 2.4.13 too?
>
>    Thanks a lot!
>
> Paul
>
>
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

-- 
Steven Van den Berghe             
ste...@in...
Workgroup Broadband Communication Networks
Department Information Technology 
Ghent University - Belgium
Phone:  +32 (0)9 267 35 86 | Fax  :  +32 (0)9 267 35 99
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
 "If the Start Windows Restart when Windows starts check
 box is checked Windows Restart will start automatically
 every time Windows is started." - Actual excerpt from a
 windows program help file!  
*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*
 

[mpls-linux-general] mpls-linux beta version

[Ng Tat-P. P. <eng...@nu...>]

Hi,

  I have been unable to find the new kernel patch + iptables patch in the
cvs. 

   Also, is the new kernel patch for kernel 2.4.13 too?

   Thanks a lot!

Paul

Re: [mpls-linux-general] How to combine label with IP packet?

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

On Wed, Jan 09, 2002 at 03:42:12PM +0800, =D5=C5=CE=B0=D2=BB wrote:
> Dear All,
>   I know that we can establish a LSP by extended RSVP,but the label inf=
ormation is kept in a kernel table.I want to know:
> How to "put" the label into the IP packet?
> How to test the function of MPLS?

There needs to be a mechanism to redirect IP packets into a MPLS layer.
This can be done via the routing table or via a MPLS tunnel interface.

Once tha packet is in the MPLS layer a lable shim can be added.

Jim

> =20
>  Waiting for your help,Thanks!
>=20
> Zhang,Weiyi
>=20
>=20
>=20
>=20
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D
> http://vip.163.com          =D7=A2=B2=E1VIP=D3=CA=CF=E4  =B4=F3=BD=B1=C8=
=CE=C4=E3=C4=C3
> http://im.163.com           =CD=F8=D2=D7=B6=CC=D0=C5=CD=A8=A3=BA=C3=E2=B7=
=D1=B7=A2=CB=CD=CA=D6=BB=FA=B6=CC=D0=C5=B5=C4=D7=C0=C3=E6=C8=ED=BC=FE
> http://dating.163.com       =D4=BC=BB=E1=D6=AE=B3=C7=A3=AC=B4=B9=C1=AF=C4=
=E3=D1=DB=D6=D0=CE=DE=C9=F9=C4=C5=BA=B0=A3=A1
>=20
> _______________________________________________
> mpls-linux-general mailing list
> mpl...@li...
> https://lists.sourceforge.net/lists/listinfo/mpls-linux-general

--=20
James R. Leu

[mpls-linux-general] =?gb2312?B?SG93IHRvIGNvbWJpbmUgbGFiZWwgd2l0aCBJUCBwYWNrZXQ/?=

[ÕÅΰһ <zh...@16...>]

Dear All,
  I know that we can establish a LSP by extended RSVP,but the label information is kept in a kernel table.I want to know:
How to "put" the label into the IP packet?
How to test the function of MPLS?
 
 Waiting for your help,Thanks!

Zhang,Weiyi




=============================================================
http://vip.163.com          ×¢²áVIPÓÊÏä  ´ó½±ÈÎÄãÄÃ
http://im.163.com           ÍøÒ׶ÌÐÅͨ£ºÃâ·Ñ·¢ËÍÊÖ»ú¶ÌÐŵÄ×ÀÃæÈí¼þ
http://dating.163.com       Ô¼»áÖ®³Ç£¬´¹Á¯ÄãÑÛÖÐÎÞÉùÄź°£¡

[mpls-linux-general] Regarding mpls_Finish method

[Jagdish G. <jag...@ie...>]

Hi
In mpls_finish (mpls_input.c) method the data is move to fill up the
mpls_gap. But the skb->tail field is not modified, Please clarify.


Jagdish

[mpls-linux-general] Regarding mpls 1.0

[<um...@de...>]

Hi all,
I have one doubt in mpls 1.0 version regarding mpls patch in
rt_set_nexthop in route.c.

1)In this we are putting the patch only if we have the fib info
available.that means we are sure fib info is available in case of mpls ,
but how ?
2)Under if (fi) , there is one more checking for mtu ==0 , then we are
assiging pmtu.What if mtu is not zero ??

Thanks and Regards
uma

[mpls-linux-general] regarding sk_buff structure....

[<vvk...@hs...>]

Hi  Abhijit !
     Thank you very much for ur valuable info .................
          Here inside "mpls_sendmsg" , i am not doing anything related to device
........

Just i am maintaining max number of bufs: 5 , in the pool at this
module.........

when application comes up it will execute read function...............

inside by using /proc filesystem , i implemented read function .. based on that
it will fetch the contents from buffer

and give it to application.............

that is the actual story............

Just i ma sending the files .....If you find bandwidth , please look at into
that files.................

(See attached file: mpls_buff.c)(See attached file: mpls_pcap.c)(See attached
file: mpls_pcap_ops.c)(See attached file: mpls_proc.h)

these are very smll files...........


the " mpls_get_packet_data" function is in the mpls_buff.c file...............

With thanx and regards,
Ramki kvv.





Abhijit Gadgil <gab...@ee...> on 01/09/2002 09:17:34 AM

To:   Veera R Koduri/HSS@HSS
cc:   mpl...@li...

Subject:  Re: [mpls-linux-general] regarding sk_buff structure....






If you want to send the packet to user-space application, mpls_sendmsg is
perhaps not the way you shd to it. Well what i think is happening is (its
most likely like that, i m not sure)

mpls_sendmsg or subsequent calls from it (if any) are refering to skb->dev
or skb->dst somewhere. And its likely that these are not set properly.

Also you can have a look at the sources of tun-tap device in drivers/net/
to check how to send packets to userspace. (Excellent stuff! I have seen
so far).

hth

-abhijit

vvk...@hs... wrote :

>
>
>Hi  Experts !
>
>     I have come across with one problem that , i want to get the packet  and
>send it to an application ,
>running in user space ..........  My objective is to get the  SHIM ,src,dest
>address and packet data ........
>
>  I am following  style.......................of code.
>
>#define MPLS_MAX_BUFF_SIZE   1500
>#define MPLS_MAX_DATA_SIZE   1500 - 4* sizeof ( unsigned long ) -1
>
>int  mpls_get_packet_data( struct sk_buff *skb,    unsigned long shim,
>                  int type )
>{
>   char        *fname="mpls_get_packet_data";
>   mpls_info_t *buff=NULL;
>
>      printk(KERN_ALERT"%s : entering .....\n",fname);
>
>      buff=(mpls_info_t *)kmalloc(sizeof(mpls_info_t), GFP_KERNEL);
>      if ( buff == NULL )
>      {
>         printk(KERN_ALERT"%s: kmalloc error",fname);
>         return -ENOMEM;
>      }
>
>      if ( skb == NULL )
>      {
>         printk(KERN_ALERT"%s: skb NULL argument error",fname);
>         kfree(buff);
>         return -EINVAL;
>      }
>
>      /* Now get the  "src", "fec ..i.e dest address",
>       * "data" from sk_buff structure .........
>       * Fill it in the "mpls_info_t" structure...
>       * and place the msg in the queue .......
>       */
>
>
>       buff->type = type;  /* Label Type ...IN or OUT Label */
>       buff->shim = shim;  /* fill the SHIM info .......*/
>       buff->src  = skb->nh.iph->saddr;  /* src address from sk_buff */
>       buff->fec  = skb->nh.iph->daddr;  /* dest address from sk_buff */
>       memset(buff->data, 0, MPLS_MAX_DATA_SIZE-1);
>
>       if ( skb->data != NULL )
>       {
>             printk(KERN_ALERT"%s: skb data length [%d]\n",fname,skb->len);
>             memcpy(buff->data, skb->data, skb->len);
>       }
>
>       return mpls_sendmsg(buff, MPLS_MAX_BUFF_SIZE);
> }
>
>This function i am calling in the "mpls_output2"  function in the file :  "
>mpls_output.c" file of mpls
>src for kinux............
>
>here problem is .........  after some time  kernel  panic msg arriving and
>system is hanging up ....
>no response ............    Error is : invalid operand : 00000
>
>Is the way i am following is correct or not ?
>Can  you  please guide me , how to retrieve the data............from  "sk_buff"
>structure ...............?
>
>Please guide me in this ..............
>
>I am following the Linux Device Drivers book by oerilly .. by alessandro
>rubini............
>
>Thanks in advance........
>Regards,Ramki kvv.
>
>Hughes Software Systems


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

Re: [mpls-linux-general] regarding sk_buff structure....

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

If you want to send the packet to user-space application, mpls_sendmsg is
perhaps not the way you shd to it. Well what i think is happening is (its
most likely like that, i m not sure)

mpls_sendmsg or subsequent calls from it (if any) are refering to skb->dev
or skb->dst somewhere. And its likely that these are not set properly.

Also you can have a look at the sources of tun-tap device in drivers/net/
to check how to send packets to userspace. (Excellent stuff! I have seen
so far).

hth

-abhijit

vvk...@hs... wrote :

>
>
>Hi  Experts !
>
>     I have come across with one problem that , i want to get the packet  and
>send it to an application ,
>running in user space ..........  My objective is to get the  SHIM ,src,dest
>address and packet data ........
>
>  I am following  style.......................of code.
>
>#define MPLS_MAX_BUFF_SIZE   1500
>#define MPLS_MAX_DATA_SIZE   1500 - 4* sizeof ( unsigned long ) -1
>
>int  mpls_get_packet_data( struct sk_buff *skb,    unsigned long shim,
>                  int type )
>{
>   char        *fname="mpls_get_packet_data";
>   mpls_info_t *buff=NULL;
>
>      printk(KERN_ALERT"%s : entering .....\n",fname);
>
>      buff=(mpls_info_t *)kmalloc(sizeof(mpls_info_t), GFP_KERNEL);
>      if ( buff == NULL )
>      {
>         printk(KERN_ALERT"%s: kmalloc error",fname);
>         return -ENOMEM;
>      }
>
>      if ( skb == NULL )
>      {
>         printk(KERN_ALERT"%s: skb NULL argument error",fname);
>         kfree(buff);
>         return -EINVAL;
>      }
>
>      /* Now get the  "src", "fec ..i.e dest address",
>       * "data" from sk_buff structure .........
>       * Fill it in the "mpls_info_t" structure...
>       * and place the msg in the queue .......
>       */
>
>
>       buff->type = type;  /* Label Type ...IN or OUT Label */
>       buff->shim = shim;  /* fill the SHIM info .......*/
>       buff->src  = skb->nh.iph->saddr;  /* src address from sk_buff */
>       buff->fec  = skb->nh.iph->daddr;  /* dest address from sk_buff */
>       memset(buff->data, 0, MPLS_MAX_DATA_SIZE-1);
>
>       if ( skb->data != NULL )
>       {
>             printk(KERN_ALERT"%s: skb data length [%d]\n",fname,skb->len);
>             memcpy(buff->data, skb->data, skb->len);
>       }
>
>       return mpls_sendmsg(buff, MPLS_MAX_BUFF_SIZE);
> }
>
>This function i am calling in the "mpls_output2"  function in the file :  "
>mpls_output.c" file of mpls
>src for kinux............
>
>here problem is .........  after some time  kernel  panic msg arriving and
>system is hanging up ....
>no response ............    Error is : invalid operand : 00000
>
>Is the way i am following is correct or not ?
>Can  you  please guide me , how to retrieve the data............from  "sk_buff"
>structure ...............?
>
>Please guide me in this ..............
>
>I am following the Linux Device Drivers book by oerilly .. by alessandro
>rubini............
>
>Thanks in advance........
>Regards,Ramki kvv.
>
>Hughes Software Systems

Re: [mpls-linux-general] Complinig modified Routing table

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

If you are making any changes to fib_frontend.c make sure you are doing
right things, 'cos as u urself would have observed, FIB implementation in
kernel is not straight forward, you will have to handle all the files
fib_rules, fib_semantics, fib_hash (I might have forgotten a couple of
flies.)

Again I would like to ask is it really needed to add anything to routing
table implementation? If you believe there is no other way go ahead with
what u think needs to be implemented. I suggest to spend some time with
the complete fib implementation (its not just fib_forntend).

I started with writting abt fib_commentry some time back but didnt finish
it. (whatever i understood)

And the answer to the question you have asked is "YES". You will have to
make the whole thing again. And boot into newly compiled kernel.

-abhijit

Kanwar Deep Singh wrote :

>Hi all,
> Suppose if I make some changes in the way in which
>kernel stores routing table (adding one mored field in
>existing routing table). by adding code in files which
>are used to calculate & print routing table (eg. file
>../net/ipv4/fib_frontend.c is used to fill up file
>/proc/route which is read to print routing table by
>#route -ee )
> No if I want to reflect the changes after i've made
>changes to (lets say fib_frontend.c file) by giving
>#route -ee command
>do I need to "make" complete kernel again... (as
>compiling of kernel takes a lot of time..)or is their
>any easy way of compiling the file I've just changed.
>we need to know this because in our project we need to
>implement mpls ( through package linux-mpls)over the
>modified routing table
>thanks in advance,
>Kanwar.
>
>__________________________________________________
>Do You Yahoo!?
>Send FREE video emails in Yahoo! Mail!
>http://promo.yahoo.com/videomail/
>
>_______________________________________________
>mpls-linux-general mailing list
>mpl...@li...
>https://lists.sourceforge.net/lists/listinfo/mpls-linux-general
>

-- 
How do u delete a letter in vi and paste it after the next one?

"xp"

--

[mpls-linux-general] regarding tranport Layer port....

[<vvk...@hs...>]

Hi  Experts !

     I  want to clarify my doubt, whether  IP uses some encoding
technique...............while tranmitting the
packet........

Assume sk_buff  pointer   "skb"    , i am trying to get the port number of the
application from the TCP header in the following way.........

skb->h.th->dest  ........./* Destination port ........*/

We are trying to transfer  certain traffic based on Transport layer port, as
part of the application of
Traffic Engineering .......

For telnet application : port : 23 ,   but when i get the dest value in
"mpls_output2" function in  "mpls_output.c" file
it is showing some other values like : 319, 510 ................   As everybody
knows Telnet port : 23

I tried to print it normally  and then   __const_htons,  __const_htons
..............

i didn't get the dest port : 23 in any of  the cases..........

Please guide me in this ..............

Thanks in advance........
Regards,Ramki kvv.

Hughes Software Systems

[mpls-linux-general] regarding sk_buff structure....

[<vvk...@hs...>]

Hi  Experts !

     I have come across with one problem that , i want to get the packet  and
send it to an application ,
running in user space ..........  My objective is to get the  SHIM ,src,dest
address and packet data ........

  I am following  style.......................of code.

#define MPLS_MAX_BUFF_SIZE   1500
#define MPLS_MAX_DATA_SIZE   1500 - 4* sizeof ( unsigned long ) -1

int  mpls_get_packet_data( struct sk_buff *skb,    unsigned long shim,
                  int type )
{
   char        *fname="mpls_get_packet_data";
   mpls_info_t *buff=NULL;

      printk(KERN_ALERT"%s : entering .....\n",fname);

      buff=(mpls_info_t *)kmalloc(sizeof(mpls_info_t), GFP_KERNEL);
      if ( buff == NULL )
      {
         printk(KERN_ALERT"%s: kmalloc error",fname);
         return -ENOMEM;
      }

      if ( skb == NULL )
      {
         printk(KERN_ALERT"%s: skb NULL argument error",fname);
         kfree(buff);
         return -EINVAL;
      }

      /* Now get the  "src", "fec ..i.e dest address",
       * "data" from sk_buff structure .........
       * Fill it in the "mpls_info_t" structure...
       * and place the msg in the queue .......
       */


       buff->type = type;  /* Label Type ...IN or OUT Label */
       buff->shim = shim;  /* fill the SHIM info .......*/
       buff->src  = skb->nh.iph->saddr;  /* src address from sk_buff */
       buff->fec  = skb->nh.iph->daddr;  /* dest address from sk_buff */
       memset(buff->data, 0, MPLS_MAX_DATA_SIZE-1);

       if ( skb->data != NULL )
       {
             printk(KERN_ALERT"%s: skb data length [%d]\n",fname,skb->len);
             memcpy(buff->data, skb->data, skb->len);
       }

       return mpls_sendmsg(buff, MPLS_MAX_BUFF_SIZE);
 }

This function i am calling in the "mpls_output2"  function in the file :  "
mpls_output.c" file of mpls
src for kinux............

here problem is .........  after some time  kernel  panic msg arriving and
system is hanging up ....
no response ............    Error is : invalid operand : 00000

Is the way i am following is correct or not ?
Can  you  please guide me , how to retrieve the data............from  "sk_buff"
structure ...............?

Please guide me in this ..............

I am following the Linux Device Drivers book by oerilly .. by alessandro
rubini............

Thanks in advance........
Regards,Ramki kvv.

Hughes Software Systems

[mpls-linux-general] Complinig modified Routing table

[Kanwar D. S. <kan...@ya...>]

Hi all,
 Suppose if I make some changes in the way in which
kernel stores routing table (adding one mored field in
existing routing table). by adding code in files which
are used to calculate & print routing table (eg. file
../net/ipv4/fib_frontend.c is used to fill up file
/proc/route which is read to print routing table by
#route -ee )
 No if I want to reflect the changes after i've made
changes to (lets say fib_frontend.c file) by giving
#route -ee command 
do I need to "make" complete kernel again... (as
compiling of kernel takes a lot of time..)or is their
any easy way of compiling the file I've just changed.
we need to know this because in our project we need to
implement mpls ( through package linux-mpls)over the
modified routing table
thanks in advance,
Kanwar.

__________________________________________________
Do You Yahoo!?
Send FREE video emails in Yahoo! Mail!
http://promo.yahoo.com/videomail/