Re: [lwip-users] dhcp_release_and_stop() clears the assigned IP address

[address@hidden]

Am 13.02.2019 um 21:48 schrieb R. Diez:
> > I can't really follow you there. But than again,
> > [...]
>
> Do you mean that lwIP's DHCP client never times out then?

Unless you tell it to stop, no. There is one exception that it starts 
AutoIP in parallel after some time without success...

It simply does not have status callbacks. You're welcome to implement 
that though...

> > I haven't used a setup like you (letting DHCP continue
> > and having a fallback to a static IP).
> > I haven't come accross communication standards
> > requiring this so far.> But then again, we have a tool to talk to devices that 
> > are set
> > to DHCP but fail to get an address...
>
> What tool is this? Is that perhaps some kind of small DHCP server?

No, a private scanning tool using broadcasts...

> My current approach has been working very well: If a user removes the device 
> from the network and connects it to a local laptop for reconfiguration 
> purposes, after a few seconds he can access the device at the fall-back IP 
> address. No external tools are necessary.

What fallback IP is that? How do you ensure it fits the subnet this 
device is connected to? How do you ensure this fallback address (I 
assume it's 192.168.0.1 *g*) is not used by other devices on that subnet?

> And there is no need to reset the device IP configuration with any external 
> button. Just reconnect it to the main network, and it works again as usual.
>
> I think I have seen other devices with DHCP but with a fall-back IP address too.

I do have, too, but to me this is kind of broken design... IP addresses 
must be assigned by someone. The devices I saw had the same fallback IP 
for all devices (factory-provided, e.g. 192.168.0.1). If you happen to 
connect that to a real subnet using this range, bad things can happen.

Regards,
Simon