Re: [lwip-users] Strategy Question for DHCP

[Zach Smith]

Here is what I can tell you. The targets are auto configuring themselves in the range you define. In the default (RFC compliant) range there are 254*256 possibilities. So it is highly unlikely that 2 nodes will choose the same one. But if so, then yes they discover each other by use of ARP. If a conflict is found, the host increments its address and tries again. I think you are right that there is no lease time concept. Once the host chooses and auto ip address and determines it is available then it keeps it.

 

DHCP will continue to run sending DHCP discover messages periodically. So, if a dhcp server comes on line then eventually the host will get a new IP address assigned by the DHCP server and abandon the autoip self-configured address.

 

Regarding the Ethernet link, every time the Ethernet link changes you should call netif_set_link_down() or up() which I believe restarts the whole thing. I think when the network link comes back up, autoip starts at the next incremented address (It might be more desirable to restart at the original address. This may be a source of a possible improvement that could be made to source code?)

 

Have a look at the struct autoip member tried_llipaddr in source code. This is what increments every time an autoip address is tried. On boot up it will be 0 and therefore the initial address tried should be the same every time. Not sure I get exactly what you are describing in number 2 there but regardless, have a look at netif->autoip->tried_llipaddr throughout the source code and you should be able to figure out what is causing it.

 

From: lwip-users-bounces+address@hidden [mailto:lwip-users-bounces+address@hidden On Behalf Of Robert Deschambault
Sent: Friday, June 19, 2015 9:52 AM
To: Mailing list for lwIP users
Subject: Re: [lwip-users] Strategy Question for DHCP

 

Ok, now that I have autoip working, I need some advice on how to use this feature properly.  Right now, I can see my target self assign a local link address.  I tried some experiments to see how autoip behaves.

 

1. I tried unplugging my Ethernet cable from the switch and I see the target self assign a new IP address.  I assume that is the expected response as there doesn't seem to be any concept of lease time like DHCP.  Is this expected behavior?  Is there any interaction with my Windows Ethernet connection?  I see ARP requests in my wireshark record between the target and the Windows host which essentially tells the target that the IP address has already been assigned for that IP address.

 

2. I have two programs in my flash, a boot loader and a main application.  The way we update is to send a command in the main application, which restarts the target and runs the boot loader.  So the Ethernet link is dropped when we jump to boot.  I saw that when the software jumped to boot it still acquired the same IP, but when the main restarted it renegotiated and came up with a new IP.  On the first jump it is imperative that we jump to the same IP address from main to boot, otherwise the load will fail.  Right now I only have one target connected, but in a real system there are up to 12 targets.  We only update one target at a time.  Will our approach work or will we have to do something drastic with the Ethernet links and reset them or recycle power?

 

Any comments or tips would be greatly appreciated.

 

Thanks.

 

-- 

Bob Deschambault

---

Confidentiality Notice: This e-mail may contain confidential and privileged material for the sole use of the intended recipient(s). Any review, use, distribution or disclosure by others is strictly prohibited. If you are not the intended recipient (or authorized to receive from the recipient), please contact the sender by reply e-mail and delete all copies of the message.