Re: [RFC] Migrating net/sched to new module interface

[Roman Zippel]

Hi,

On Fri, 14 Feb 2003, Rusty Russell wrote:

> > It's not the same, please see: 
> > http://marc.theaimsgroup.com/?l=linux-kernel&m=104284223130775&w=2
> > I explained why the current module locking is more complex and why it's 
> > actually a three stage delete.
> 
> No, here is where you show *your* ignorance of kernel locking idioms,
> and that your axiom is that "the new system is more complex".

Another point you probably misunderstood: I said that the complexity of 
the new and the old system is exactly the same, please read more carefully 
before flaming, it might backfire.

> > Rusty, above are real problems, the module locking fixes these problems 
> > during module_init/module_exit, but how can these problems fixed in the 
> > other cases and how does the module locking help?
> 
> This isn't even a sensible question: "This is not a module problem.
> How does module locking help?"

I hate to drag people into a discussion, but maybe you're more inclined 
to believe Al:

http://marc.theaimsgroup.com/?l=linux-kernel&m=103761331525509&w=2
http://marc.theaimsgroup.com/?l=linux-kernel&m=103769023500378&w=2

Please read this very careful and think about it.

bye, Roman