Re: [6/9] [RFC] Steps to fixing the driver model locking

[Patrick Mochel <mochel@digitalimplant.org>]

On Wed, 23 Mar 2005, OGAWA Hirofumi wrote:

> Patrick Mochel <mochel@digitalimplant.org> writes:
>
> > +void klist_del(struct klist_node * n)
> > +{
> > +	struct klist * k = n->n_klist;
> > +
> > +	spin_lock(&k->k_lock);
> > +	klist_dec_and_del(n);
> > +	spin_unlock(&k->k_lock);
> > +}
>
> Can't we use atomic_dec_and_lock()?

No. It uses the kref_inc() and kref_dec(), which do not have the
equivalent atomic_dec_and_lock() primitives.

> [...]
>
> > +void klist_remove(struct klist_node * n)
> > +{
> > +	spin_lock(&n->n_klist->k_lock);
> > +	klist_dec_and_del(n);
> > +	spin_unlock(&n->n_klist->k_lock);
> > +	wait_for_completion(&n->n_removed);
> > +}
>
> Why isn't those going into drivers/base/?  Personally, klist seems
> drivers/base stuff rather than generic stuff...

Could be for now. I'd like to convert more of the kobject/kest mess to use
a cleaner, and rwsem-free locking mechanism, which would nearly justify it
in lib/. All in all, that change is purely cosmetic.


	Pat
-
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/