[Wifidog] Re: Fwd: WiFiDog trial with NYCwireless

[Alexandre Carmel-Veilleux]

On Tue, Dec 07, 2004 at 11:07:44PM -0500, Benoit Grégoire wrote:
> 
> Good to know.

	The main worry in the libhttpd code is a memory leak. There might
also be a segfault or two in code paths we don't cross. Not an issue for
us, but I wouldn't redistribute libhttpd separately just yet.

> Well, I suppose one way to do it would be to create a rule in the firewall 
> like we do once a user is authenticated, but only for accounting purposes.  
> We can then sent the subnet adress and mask as the username.  It may be 
> cleaner to implement if we actually set a flag to tell the auth server it's 
> goint ot a whitelisted subnet.  The auth server would then take care of 
> treating those statistics.  The big plus is that the stats and bandwidth 
> usage of the hotspots would remain consistent with actual bandwidth usage, 
> and we can still create statistics for the whitelisted subnets.

	I might do some experimentation with IPTables, I have an RH8 box
I'm supposed to move to Debian (it needs it and I want to play with FAI
before re-imaging my entire lab.) I can test it then, should be around
next week or so.

> Now since I didn't write much of the code for the gateway, I don't know if 
> that's a reasonably simple way to approach the problem.

	The logging code can be integrated into the fw_counter() and
related bits of code and piggy-back on an existing thread.

Alex

_______________________________________________
Wifidog mailing list
Wifidog at isf.waglo.com
http://isf.waglo.com/mailman/listinfo/wifidog_isf.waglo.com