Rusty Russell wrote:
>
> In message <20020912030933.A13608@kushida.apsleyroad.org> you write:
> > I don't see the point in this at all.
>
> Yes, I'm starting to realize that.
>
> Frankly, I'm sick of pointing out the same problems individually to
> every shallow thinker who thinks that "it's easy".
>
> The fundamental problems with modules are as follows:
> A) Many places in the kernel do not increment module reference counts
> for you, and it is difficult currently write a module which safely
> looks after its own reference counts (see
> net/ipv4/netfilter/ip_conntrack_core.c's ip_conntrack_cleanup())
This sort of thing makes correct unloading hard, but in my eyes it
is an argument for changing the module interface to something better
and say "that thing CAN'T be a module of its own!"
What if we *require* modules to use some open/close reference count that
don't change so often as to be a performance problem?
It makes the unloading/reloading problem easier, similiar to
mount/umount.
It'll work for lots of modules, such as:
* drivers for hardware devices, char, block, and NIC.
* filesystems
Things with no easy refcounting (and it cannot even be grafted on)
will have to be non-modular, or folded into some parent module.
Maybe IP connection tracking can't be modular with these rules - so
what? Make it compiled-in, or a part of the IPV4-module. So
if you really need to load and unload that you unload the
ipv4_with_conntrack module and then load a ipv4_without_conntrack
module.
To me, it seems like the current module interface is too fine-grained,
and that cause trouble. I.e. the overhead of correct refcounting
is so high in some cases that it isn't used - causing trouble
with safe unloading. The solution is to say no to such modules.
Making them work isn't easy - but why try?
Helge Hafting
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sun Sep 15 2002 - 22:00:32 EST