Re: [RFC] Is an alternative module interface needed/possible?

[Werner Almesberger <wa@almesberger.net>]

Roman Zippel wrote:
> Anyway, this alone would be not reason enough to change the module 
> interface, but another module interface would give us more flexibility and 
> reduce the locking complexity.

Wait, wait ! :-) There's one step you've left out: what we actually
expect the module interface to do. We have:

 - what it currently does, or what it did in the past
 - what users think it does
 - what users want it to do
 - what we think the users should want
 - what we think is a comfortable compromise

With "users", I mean primarily the guy who invokes "rmmod", or such.

Anyway, I'm afraid I can't offer much wisdom from experience for this
part, for I'm not much of a module user myself. I'll have more to say
on service interfaces, though.

Sorry for slowing down, but I'm currently quite busy absorbing all
the cool stuff that's recently been happening with UML. (So, blame
Jeff ;-))

- Werner

-- 
  _________________________________________________________________________
 / Werner Almesberger, Buenos Aires, Argentina         wa@almesberger.net /
/_http://www.almesberger.net/____________________________________________/