Re: [PATCH net-next 2/3] udp: avoid a cache miss on dequeue

[Paolo Abeni <pabeni@redhat.com>]

On Thu, 2017-06-01 at 09:40 -0700, Eric Dumazet wrote:
> On Thu, Jun 1, 2017 at 9:21 AM, Paolo Abeni <pabeni@redhat.com> wrote:
> 
> > I'm sorry, I do not follow. I'm concerned about the secpath field (skb-
> > > sp), which is the only one that can be not NULL in
> > 
> > __udp_queue_rcv_skb().
> > 
> > If the secpath is not NULL, calling there secpath_reset() (or the to-
> > be-introduced skb_reset_head_state()), we will properly release it and
> > we will clear the field, too.
> > 
> > Calling skb_release_head_state() in the same scenario, we release the
> > secpath, but we don't clear it. So if the packet is later dropped we
> > will get a double free, unless we add and use a specialized a
> > free_stateless_skb(), too.
> 
> Then simply use secpath_reset() instead of secpath_put() from
> skb_release_head_state()
> 
> Clearly having these subtle differences bring confusion, for very little gain.
> 
> secpath_put() should be removed. Most of its callers simply set
> skb->sp back to NULL anyway.

To make the code robust we would have to NULL all the other fields
(nfct, nf_bridge, destructor, sk) that are currently not cleared in
skb_release_head_state(), elsewhere if one day, after some change, any
that fields become non-NULL in this code path we risk a double-free
after skb_release_head_state(), even if the code looks safe.

Will that be a little too invasive for this small use-case? Can't we
prefer a new helper or simply a secpath_reset() plus some appropriate
comments?

Thanks,

Paolo