Peter Memishian wrote:
> > But in that case the plugin would have to be invoked to validate the
> > plugin data specified by MAC drivers when they register, for example to
> > detect a mismatching version number. I don't remember seeing such hook
> > in your proposal.
>
> That's a good suggestion. I'll add a plugin operation to validate new
> plugin data.
I presume this will be an optional callback, since only plugins using
plugin data will need it?
Yes.
(And further, that if mac_register() is passed
plugin data, but the plugin does not support this callback, that
mac_register() will return an error?)
Yes, that sounds reasonable. The documentation for the plugin framework
can mention that if a plugin supports plugin data, then it must provide
the plugin data validation operation.
-Seb
_______________________________________________
networking-discuss mailing list
[email protected]
