Re: Enabling built-in modules earlier in init

[Andrew Doran <ad%NetBSD.org@localhost>]

On Wed, Jun 16, 2010 at 09:47:54AM +0300, Antti Kantee wrote:
> On Wed Jun 16 2010 at 12:13:38 +1000, matthew green wrote:
> > i think having a class of builtin modules that are available during
> > autoconfig isn't a bad thing.  perhaps the right answer is infact to
> > convert MODULE_CLASS_SECMODEL into say MODULE_CLASS_EARLY, and remove
> > the special case handling for MODULE_CLASS_SECMODEL to deal with
> > calling secmodel_register()/secmodel_deregister() by pushing these
> > calls into the mi_modcmd() calls inside the secmodules themselves.
> 
> With the current ways of secmodel register, I'd be damn careful to not
> push it around.  The effect is that if it's called 0 times, you have a
> system which allows everything.  So if your suggestion is implemented
> and you're testing a new secmodel which buggily omits register alongside
> another correctly registering secmodel, things will appear to work fine,
> But if in some scenario the buggy one is loaded alone, well ... welcome
> to the wishing well.

That would be a really serious bug.  Don't play with fire unless you're
sure you won't get burned.

The secmodel registration stuff belongs in the module entry point.