[Click] Socket delayed read

[Christine Jones]

I've tried out the latest git repository version of Click, and the same problem occurs.

Christine

On Jul 19, 2012, at 7:30 AM, Christine Jones wrote:

> I am running the head of the main Click git repository as of 9-22-2011. That is right before Click 2.0.1 was released.
> 
> So, it has been awhile since I have updated. Have the bug fixes occurred more recently?
> 
> Thanks!
> 
> Christine
> 
> On Jul 19, 2012, at 6:54 AM, Beyers Cronje wrote:
> 
>> Hi,
>> 
>> What version of Click are you running? There has been some bug fixes to
>> Socket a while back that addressed some notifier and scheduling bugs.
>> 
>> Beyers
>> 
>> On Thu, Jul 19, 2012 at 3:25 PM, Christine Jones <cej at bbn.com> wrote:
>> 
>>> Hello!
>>> 
>>> I am using Click to set up a router that installs a proxy service for a
>>> subset of traffic. Much of the traffic flowing through the router is
>>> forwarded as normal. But some traffic that meets certain criteria is sent
>>> to the proxy (a user-space process on the local machine). The proxy also
>>> sends packets and other state back to the Click router.
>>> 
>>> There are two versions of this router, kernel and user-level. I am
>>> encountering a problem with the user-level version. The communication
>>> between router and proxy is via the Socket element configured to be a unix
>>> socket, as such...
>>> 
>>> proxy :: Socket(UNIX, 'pathname', CLIENT true)
>>> 
>>> FromDevice
>>>       ...
>>>       -> proxy;
>>> 
>>> proxy
>>>       ...
>>>       -> Queue
>>>       -> RawSocket(TCP);
>>> 
>>> The problem is that when the router has a period in which no packets are
>>> forwarded to the proxy the socket may not be read (although the socket is
>>> configured for reading in the select loop and there is data waiting from
>>> the proxy). If I force the router to write to that socket (with a periodic
>>> keep-alive) then the socket will be read. However, I would prefer to not
>>> have to rely on keep-alives. The socket should be selected when there is
>>> data to read.
>>> 
>>> Any thoughts?
>>> 
>>> Thanks!
>>> 
>>> Christine
>>> _______________________________________________
>>> click mailing list
>>> click at amsterdam.lcs.mit.edu
>>> https://amsterdam.lcs.mit.edu/mailman/listinfo/click
>>> 
>> _______________________________________________
>> click mailing list
>> click at amsterdam.lcs.mit.edu
>> https://amsterdam.lcs.mit.edu/mailman/listinfo/click
> 
> 
> _______________________________________________
> click mailing list
> click at amsterdam.lcs.mit.edu
> https://amsterdam.lcs.mit.edu/mailman/listinfo/click