[Click] Re: two configuration tests
Eddie Kohler
kohler at cs.ucla.edu
Wed Jun 30 16:39:19 EDT 2004
Yihong Zhou wrote:
> Sorry bother you guys. I found my stupid problem. I worked remotely on the click machine. That is, the machine was receving packets constantly from my local machine.
>
> It seems that if click runs in the user level, it won't interrupt other network services. The received packets are sent to both click and the server programs. But if the click runs in the linux kernel module, that would be different. Is that right?
Yes, that's right.
>
> By the way, can I write an element that has several pull input port and several push input port? What is the return value for "processing" then? AGNOSTIC? Can the agnostic port connect to the output port of a queue?
Yes, you can! See: http://www.pdos.lcs.mit.edu/click/doc/progman.html#processing
Eddie
>
> ----- Original Message -----
> From: Yihong Zhou
> To: click at amsterdam.lcs.mit.edu
> Sent: Wednesday, June 30, 2004 12:14 PM
> Subject: two configuration tests
>
>
> I run two configuration files in the userlevel:
>
> test-device.click : FromDevice(eth0)->print(ok)->discard
>
> and print-ping.click : FromDevice(eth0)->Classifier(12/0800)->Strip(14)->CheckIPHeader->IPFilter(allow icmp && icmp type echo) -> IPPrint -> Discard.
>
> I sent one ping packet to the machin. The second one works perfect. That is, it printed exactly one ping packet. But the first one printed infinitely with almost the same information: ok: 66 | 000d56be 78cb00b0 d0793d61 .....
>
> Can anybody tell me the reason?
> _______________________________________________
> click mailing list
> click at amsterdam.lcs.mit.edu
> https://amsterdam.lcs.mit.edu/mailman/listinfo/click
More information about the click
mailing list