Darren Reed wrote:
>>> Of these, the first (protocol addition/removal) is implied by
>>> the creation of a new IP instance. The other two require
>>> separate notification.
>>
>> That assumes all protocols are up and running in the kernel before the
>> first interested consumer. Do we know that is the case?
>
> What I think you're missing is that all instances of a
> protocol are to be announced via the net_instance_*
> mechanism, not just those that are created at run time
> as zones with exclusive instances are added.
Ah - I hadn't realized that.
Erik
