Re: [PATCH 12/25] aio: Refcounting cleanup

[Kent Overstreet <koverstreet@google.com>]

On Wed, Nov 28, 2012 at 04:17:59PM -0800, Zach Brown wrote:
> 
> >  struct kioctx {
> >  	atomic_t		users;
> > -	int			dead;
> > +	atomic_t		dead;
> 
> Do we want to be paranoid and atomic_set() that to 0 when the ioctx is
> allocated?

I suppose we should, yeah.

> > +	while (!list_empty(&ctx->active_reqs)) {
> > +		struct list_head *pos = ctx->active_reqs.next;
> > +		struct kiocb *iocb = list_kiocb(pos);
> 
> I'd use list_first_entry() and ignore the list_kiocb() wrapper, I think.

Fine by me.

> 
> > +	if (!atomic_xchg(&ctx->dead, 1)) {
> > +		hlist_del_rcu(&ctx->list);
> > +		synchronize_rcu();
> 
> > +	hlist_for_each_entry_rcu(ctx, n, &mm->ioctx_list, list)
> > +		if (ctx->user_id == ctx_id){
> > +			BUG_ON(atomic_read(&ctx->dead));
> 
> Hmm.  Won't it be possible to race lookup and io_destroy() to hit this
> BUG?

Ouch, yes. Dunno what I was thinking when I added that.

--
To unsubscribe, send a message with 'unsubscribe linux-aio' in
the body to majordomo@kvack.org.  For more info on Linux AIO,
see: http://www.kvack.org/aio/
Don't email: <a href=mailto:"aart@kvack.org">aart@kvack.org</a>