RE: [lwip-devel] Possible DHCP problem or bug?

[Bill Auerbach]

Could it be that routers are smart about this?  Or mine is? It seems that it has made the request to change only once and lwIP does make the change.  But if I cycle power with it coming back up at the previous duplicate IP address the router does not request the change again.  We saw it here 3 times on 3 different lwIP systems and now on the 3^rd one I can’t get it to duplicate.  I know it changes because I was specifically testing for this case and saw it change because we were suspecting it’s a problem a user is having with a statically configured device colliding on his network.

 

I’m really saying if there is a problem,  I am having trouble repeating it.

 

Bill

 

From: address@hidden [mailto:address@hidden On Behalf Of address@hidden
Sent: Monday, February 22, 2010 4:02 PM
To: address@hidden
Subject: Re: [lwip-devel] Possible DHCP problem or bug?

 

I'd consider that a bug worth of its own bugtracker entry. The address shouldn't change. I guess there's something wrong in the DHCP code that (receiving an ARP response,) it doesn't check whether it is started or not. Could you please file a bug report?

Thanks,
Simon


Bill Auerbach wrote:

I don’t want to submit this as a bug until others voice their ideas  based on their knowledge of DHCP and agree that it is.

I have built lwIP with DHCP enabled.  I’m running in a static IP mode – meaning dhcp_start wasn’t called and netif_set_ipaddr is called with an address.  I have done this and connected on a network (with a DHCP server) that had a PC already at the same DHCPP assigned address as the static lwIP address.  I get the duplicate IP address bubble from Windows XP.  A few seconds later when I query my lwIP device to get its IP address, it has been changed to another IP address.

I would expect without calling dhcp_start that my device’s IP address would not change.  It won’t communicate, sure, but I don’t expect it to be changed so that it can communicate unless I had called dhcp_start.  I do not know what happens if the PC is turned on second.  I can check that as well, but I think already the current behavior is not what it should be.

Thanks,

Bill

_______________________________________________

lwip-devel mailing list

address@hidden

http://lists.nongnu.org/mailman/listinfo/lwip-devel