Re: [PATCH 2/4] percpu-rwsem: introduce percpu_rwsem_release() and percpu_rwsem_acquire()

[Oleg Nesterov <oleg@redhat.com>]

On 07/31, Peter Zijlstra wrote:
>
> On Wed, Jul 22, 2015 at 11:15:35PM +0200, Oleg Nesterov wrote:
> > +
> > +static inline void percpu_rwsem_release(struct percpu_rw_semaphore *sem,
> > +					bool read, unsigned long ip)
> > +{
> > +	lock_release(&sem->rw_sem.dep_map, 1, ip);
> > +#ifdef CONFIG_RWSEM_SPIN_ON_OWNER
> > +	if (!read)
> > +		sem->rw_sem.owner = NULL;
> > +#endif
> > +}
> > +
> > +static inline void percpu_rwsem_acquire(struct percpu_rw_semaphore *sem,
> > +					bool read, unsigned long ip)
> > +{
> > +	lock_acquire(&sem->rw_sem.dep_map, 0, 1, read, 1, NULL, ip);
> > +}
>
> This is of course entirely vile.. Can't we open code that in the freezer
> code?

Move these helpers somewhere in fs/super.c ? I don't think we should.
All members in struct percpu_rw_semaphore are private, nobody outside
of percpu-rwsem.[ch] should touch ->rw_sem at least.

> Having helpers here might give some people the impression that
> this is a sane thing to do.

Yes, but this doesn't differ from other lockdep release/acquire helpers,
they should be used with care.

BTW, we also need these helpers to add the multi-writer support, but this
is almost off-topic right now.

> Also, when you do that in the freezer, put a big honking comment on it.

OK, I'll update the comment(s).

Oleg.