Acked-by: Ethan Jackson <et...@nicira.com>
On Wed, Sep 4, 2013 at 12:39 PM, Ben Pfaff <b...@nicira.com> wrote:
> From: Ethan Jackson <et...@nicira.com>
>
> Facets maintain a pointer to their owning ofproto-dpif, and therefore
> when an ofproto-dpif is destroyed all of their child facets should be
> destroyed along with it. If they aren't, it's possible a subfacet
> looked up in the dpif-backer could access it's parent facet and
> therefore a defunct parent ofproto causing a segmentation fault.
>
> Bug #19421.
> Bug #19423.
> Diagnosed-by: Ben Pfaff <b...@nicira.com>
> Signed-off-by: Ethan Jackson <et...@nicira.com>
> Signed-off-by: Ben Pfaff <b...@nicira.com>
> ---
> ofproto/ofproto-dpif.c | 9 +++++++++
> 1 file changed, 9 insertions(+)
>
> diff --git a/ofproto/ofproto-dpif.c b/ofproto/ofproto-dpif.c
> index 70a226c..336453c 100644
> --- a/ofproto/ofproto-dpif.c
> +++ b/ofproto/ofproto-dpif.c
> @@ -1421,8 +1421,17 @@ destruct(struct ofproto *ofproto_)
> struct rule_dpif *rule, *next_rule;
> struct ofputil_packet_in *pin, *next_pin;
> struct ofputil_flow_mod *fm, *next_fm;
> + struct facet *facet, *next_facet;
> + struct cls_cursor cursor;
> struct oftable *table;
>
> + ovs_rwlock_rdlock(&ofproto->facets.rwlock);
> + cls_cursor_init(&cursor, &ofproto->facets, NULL);
> + ovs_rwlock_unlock(&ofproto->facets.rwlock);
> + CLS_CURSOR_FOR_EACH_SAFE (facet, next_facet, cr, &cursor) {
> + facet_remove(facet);
> + }
> +
> ofproto->backer->need_revalidate = REV_RECONFIGURE;
> ovs_rwlock_wrlock(&xlate_rwlock);
> xlate_remove_ofproto(ofproto);
> --
> 1.7.10.4
>
_______________________________________________
dev mailing list
dev@openvswitch.org
http://openvswitch.org/mailman/listinfo/dev
