| Message ID | 1578887809-29492-1-git-send-email-vishal.deep.ajmera@ericsson.com |
|---|---|
| State | Superseded |
| Headers | show |
| Series | [ovs-dev] ofproto: Add support to watch controller port liveness in fast-failover group | expand |
On Mon, Jan 13, 2020 at 09:26:49AM +0530, Vishal Deep Ajmera via dev wrote: > Currently fast-failover group does not support checking liveness of controller > port (OFPP_CONTROLLER). However this feature can be useful for selecting > alternate pipeline when controller connection itself is down for e.g. > by using local DHCP server to reply for any DHCP request originating from VMs. > > This patch adds the support for watching controller port liveness in fast- > failover group. Controller port is considered live when atleast one > of-connection is alive. > > Example usage: > > ovs-ofctl add-group br-int 'group_id=1234,type=ff, > bucket=watch_port:CONTROLLER,actions:<A>, > bucket=watch_port:1,actions:<B> > > Signed-off-by: Vishal Deep Ajmera <vishal.deep.ajmera@ericsson.com> Thanks for the new feature. Please add an item to NEWS. OVS is currently in "soft freeze" for the next release. We should fork for the next release at the end of the week. Therefore, please resubmit your patch next week.
> > Thanks for the new feature. > > Please add an item to NEWS. > > OVS is currently in "soft freeze" for the next release. We should fork > for the next release at the end of the week. Therefore, please resubmit > your patch next week. Thanks Ben. Sure I will post new patch after fork is done. Warm Regards, Vishal Ajmera
> > > > Please add an item to NEWS. > > > > OVS is currently in "soft freeze" for the next release. We should fork > > for the next release at the end of the week. Therefore, please resubmit > > your patch next week. > > Thanks Ben. Sure I will post new patch after fork is done. > Hi Ben, I have reposted the patch in mailing list. Warm Regards, Vishal Ajmera
diff --git a/lib/ofp-group.c b/lib/ofp-group.c index b675e80..bf0f8af 100644 --- a/lib/ofp-group.c +++ b/lib/ofp-group.c @@ -660,7 +660,8 @@ parse_bucket_str(struct ofputil_bucket *bucket, char *str_, } else if (!strcasecmp(key, "watch_port")) { if (!ofputil_port_from_string(value, port_map, &bucket->watch_port) || (ofp_to_u16(bucket->watch_port) >= ofp_to_u16(OFPP_MAX) - && bucket->watch_port != OFPP_ANY)) { + && bucket->watch_port != OFPP_ANY + && bucket->watch_port != OFPP_CONTROLLER)) { error = xasprintf("%s: invalid watch_port", value); } } else if (!strcasecmp(key, "watch_group")) { diff --git a/ofproto/ofproto-dpif-xlate.c b/ofproto/ofproto-dpif-xlate.c index 4407f9c..461e8aa 100644 --- a/ofproto/ofproto-dpif-xlate.c +++ b/ofproto/ofproto-dpif-xlate.c @@ -1884,9 +1884,12 @@ bucket_is_alive(const struct xlate_ctx *ctx, return (!ofputil_bucket_has_liveness(bucket) || (bucket->watch_port != OFPP_ANY + && bucket->watch_port != OFPP_CONTROLLER && odp_port_is_alive(ctx, bucket->watch_port)) || (bucket->watch_group != OFPG_ANY - && group_is_alive(ctx, bucket->watch_group, depth + 1))); + && group_is_alive(ctx, bucket->watch_group, depth + 1)) + || (bucket->watch_port == OFPP_CONTROLLER + && ofproto_is_alive(&ctx->xbridge->ofproto->up))); } static void diff --git a/ofproto/ofproto-dpif.c b/ofproto/ofproto-dpif.c index d3cb392..818a077 100644 --- a/ofproto/ofproto-dpif.c +++ b/ofproto/ofproto-dpif.c @@ -1802,6 +1802,7 @@ run(struct ofproto *ofproto_) { struct ofproto_dpif *ofproto = ofproto_dpif_cast(ofproto_); uint64_t new_seq, new_dump_seq; + bool is_connected; if (mbridge_need_revalidate(ofproto->mbridge)) { ofproto->backer->need_revalidate = REV_RECONFIGURE; @@ -1870,6 +1871,15 @@ run(struct ofproto *ofproto_) ofproto->backer->need_revalidate = REV_MCAST_SNOOPING; } + /* Check if controller connection is toggled. */ + is_connected = ofproto_is_alive(&ofproto->up); + if (ofproto->is_controller_connected != is_connected) { + ofproto->is_controller_connected = is_connected; + /* Trigger revalidation as fast failover group monitoring + * controller port may need to check liveness again. */ + ofproto->backer->need_revalidate = REV_RECONFIGURE; + } + new_dump_seq = seq_read(udpif_dump_seq(ofproto->backer->udpif)); if (ofproto->dump_seq != new_dump_seq) { struct rule *rule, *next_rule; diff --git a/ofproto/ofproto-dpif.h b/ofproto/ofproto-dpif.h index c9d5df3..aee61d6 100644 --- a/ofproto/ofproto-dpif.h +++ b/ofproto/ofproto-dpif.h @@ -342,6 +342,9 @@ struct ofproto_dpif { struct guarded_list ams; /* Contains "struct ofproto_async_msgs"s. */ struct seq *ams_seq; /* For notifying 'ams' reception. */ uint64_t ams_seqno; + + bool is_controller_connected; /* True if any controller admitted this + * switch connection. */ }; struct ofproto_dpif *ofproto_dpif_lookup_by_name(const char *name); diff --git a/ofproto/ofproto.c b/ofproto/ofproto.c index 08830d8..a155fad 100644 --- a/ofproto/ofproto.c +++ b/ofproto/ofproto.c @@ -1899,7 +1899,8 @@ ofproto_wait(struct ofproto *p) bool ofproto_is_alive(const struct ofproto *p) { - return connmgr_has_controllers(p->connmgr); + return (connmgr_has_controllers(p->connmgr) + && connmgr_is_any_controller_admitted(p->connmgr)); } /* Adds some memory usage statistics for 'ofproto' into 'usage', for use with
Currently fast-failover group does not support checking liveness of controller port (OFPP_CONTROLLER). However this feature can be useful for selecting alternate pipeline when controller connection itself is down for e.g. by using local DHCP server to reply for any DHCP request originating from VMs. This patch adds the support for watching controller port liveness in fast- failover group. Controller port is considered live when atleast one of-connection is alive. Example usage: ovs-ofctl add-group br-int 'group_id=1234,type=ff, bucket=watch_port:CONTROLLER,actions:<A>, bucket=watch_port:1,actions:<B> Signed-off-by: Vishal Deep Ajmera <vishal.deep.ajmera@ericsson.com> --- lib/ofp-group.c | 3 ++- ofproto/ofproto-dpif-xlate.c | 5 ++++- ofproto/ofproto-dpif.c | 10 ++++++++++ ofproto/ofproto-dpif.h | 3 +++ ofproto/ofproto.c | 3 ++- 5 files changed, 21 insertions(+), 3 deletions(-)