[Click] FromHost initialization(uselevel)

[Eddie Kohler]

The output port out of range is referring to the IPRewriter's output port -- 
i.e. the Click element level.

Eddie


Lesedi wrote:
> Hello,
> 
> Sorry to be writing this again its just that i think Eddie when replying to
> my previous mail,didnt really get the problem that i have.
> 
> My situation is I am trying to intercept FTP application and make it run
> over a click code that I have.FTP server is in the same machine as the click
> code.So i tried many elements to solve the problem:
> 
> 1.I tried using FromSocket for the code to listen to the FTP port so that
> everytime the port is active all packets go through the code.Using this
> seems not to work,and now i thinnnnkk I understand that FromSocket does not
> listen to the port but tries to connect,so i tried using another way.But it
> would have been desirable to work it this way because connect to FTP client
> would have been easy.
> 
> 2.Now I tried using 3 machines.And tried the idea of transparent traffic
> diverter.One machine runs the click code,one machine is a click server and
> the other is a click client.With this I was able to listen to the FTP port
> (inpackets arriving at FromDevice(eth0),)by using another code to intercept
> the FTP sessions, using IPClassifier(dst tcp port  21).Then I am suppose to
> use IPRewriter (pattern fake_address 1024-65535 mymachine_address  8000 0 1)
> to  send them to a fake  address,then  use ToHost(devname). to store them in
> my_address machine.Then FromHost to run them through IPREwrite again and my
> click code i guess didnt go this far just an idea..
> 
> The problem with this is :
> a)says IPRewriter output port is out of range,but then thats the range that
> i can have(1024-65535)
> 
> b)ToHost needs a FromHost initialization.
> 
> I think both problems are because of the FromHost initialization.I tried
> searching for how to initialize FromHost in userlevel,but then I couldn't.So
> if you could help   me with how to initialize FromHost mayb my problem wil
> be solved.
> 
> thanking you in advance
> 
> sedi.
> _______________________________________________
> click mailing list
> click at amsterdam.lcs.mit.edu
> https://amsterdam.lcs.mit.edu/mailman/listinfo/click