Re: [PATCH] pkt_sched: Destroy gen estimators under rtnl_lock().

[Jarek Poplawski <jarkao2@gmail.com>]

On Tue, Aug 19, 2008 at 06:05:57PM +1000, Herbert Xu wrote:
> On Tue, Aug 19, 2008 at 07:56:23AM +0000, Jarek Poplawski wrote:
> >
> > I'm not sure I can understand you: could you look at htb_destroy()
> > instead and think of this as a child qdisc of prio or another htb?
> > Having a top level "queue" lock guarantees there is no activity at
> > the whole tree at the moment.
> 
> htb_destroy can either be called by qdisc_destroy or when a brand
> new HTB qdisc fails construction.  The latter case is trivial since
> the qdisc has never been used.
> 
> In the first case, as you have seen from my previous email, the
> entire branch containing the HTB qdisc (that is, either the HTB
> qdisc itself if it's being deleted directly, or the branch stemming
> from its ancestor that's being deleted) must no longer have any
> references to it at all apart from this thread of execution.
> 
> As such we can do whatever we want with it, including freeing it.
> 

As I've written before I'm mainly concerned with things like
tcf_destroy_chain(), especially wrt. cls_u32, but I can be wrong with
this. So, if you don't have such concerns, let's forget it for now,
and after I look at this more maybe we'll get back to this discussion.

Jarek P.