RE: [lwip-users] Hardware - level for netif_remove()

[Bill Auerbach]

True.  Add to the netif more function pointers, like init, power_up,
power_down, shut_down, etc.  I would expect being an embedded systems
targeted TCP/IP stack that most applications have one netif.  I've never
been crazy about netif_add calling init.  I would expect the netif to be
added already initialized.  Or really why would it have to be?  In any case,
I worked on a stack with netif_power_up and netif_power_down and it was used
in embedded systems that needed only intermittent network activity or low
power designs.

Bill

>The problem is that we would have to pass a parameter to netif_remove,
>since the init function is not known at that time (not saved in struct
>netif). Being like that, we could also add another function for
>low_level_remove() instead of re-using the init function. The question
>is whether many people need it because it adds to the code size...
>> Impact on existing ports is minimal.
>The main impact would be that we would have to add a new parameter to
>netif_remove().