Re: [RFC/PATCH 3/5] rt: plist_head_splice

[Peter Zijlstra]

On Tue, 2007-10-23 at 11:10 -0400, Steven Rostedt wrote:
> --
> On Tue, 23 Oct 2007, Peter Zijlstra wrote:
> > +
> > +void plist_head_splice(struct plist_head *src, struct plist_head *dst)
> > +{
> > +	struct plist_node *src_iter_first, *src_iter_last, *dst_iter;
> > +	struct plist_node *tail = container_of(dst, struct plist_node, plist);
> > +
> > +	dst_iter = next_prio(tail);
> > +
> > +	while (!plist_head_empty(src) && dst_iter != tail) {
> > +		src_iter_first = plist_first(src);
> > +
> > +		src_iter_last = next_prio(src_iter_first);
> > +		src_iter_last = prev_node(src_iter_last);
> > +
> > +		WARN_ON(src_iter_first->prio != src_iter_last->prio);
> > +		WARN_ON(list_empty(&src_iter_first->plist.prio_list));
> > +
> > +		while (src_iter_first->prio > dst_iter->prio) {
> > +			dst_iter = next_prio(dst_iter);
> > +			if (dst_iter == tail)
> > +				goto tail;
> > +		}
> > +
> > +		list_del_init(&src_iter_first->plist.prio_list);
> > +
> > +		if (src_iter_first->prio < dst_iter->prio) {
> 
> I may be confused here, but shouldn't we be linking the
> src_iter_first->prio_list somewhere here? Doesn't all different prios need
> to be in its separate prio_list? Otherwise two splices in a row can cause
> the above WARN_ON (prio != prio).
> 

> > +			list_add_tail(&src_iter_first->plist.node_list,
> > +					&dst_iter->plist.node_list);

Uhm, yeah, that was supposed the be prio_list indeed.

Thanks!

> > +		} else if (src_iter_first->prio == dst_iter->prio) {
> > +			dst_iter = next_prio(dst_iter);
> > +		} else BUG();
> > +
> > +		list_splice2_tail(&src_iter_first->plist.node_list,
> > +			       	  &src_iter_last->plist.node_list,
> > +				  &dst_iter->plist.node_list);
> > +	}
> > +
> > +tail:
> > +	list_splice_tail_init(&src->prio_list, &dst->prio_list);
> > +	list_splice_tail_init(&src->node_list, &dst->node_list);
> > +}
> >
> > --
> >
> >
> >

Attachment: signature.asc
Description: This is a digitally signed message part