[ovs-dev] [PATCH ovn v5 01/16] ovn-northd: Move out Table 0 (ingress) operations to functions
Anton Ivanov
anton.ivanov at cambridgegreys.com
Tue Sep 15 15:28:24 UTC 2020
On 15/09/2020 12:12, Numan Siddique wrote:
>
>
>
> On Fri, Sep 11, 2020 at 3:11 PM <anton.ivanov at cambridgegreys.com
> <mailto:anton.ivanov at cambridgegreys.com>> wrote:
>
> From: Anton Ivanov <anton.ivanov at cambridgegreys.com
> <mailto:anton.ivanov at cambridgegreys.com>>
>
> Signed-off-by: Anton Ivanov <anton.ivanov at cambridgegreys.com
> <mailto:anton.ivanov at cambridgegreys.com>>
>
>
> Hi Anton,
>
> Thanks for this series.
>
> I found most of the patches to be splitting the code into functions
> into proper logical blocks.
> However, patches 3 to 7, split the code into functions which I think
> can be further reorganized properly.
> What I mean is that if we take ARP response flows for NAT entries as
> an example, the
> function build_lrouter_flows_ip_input_od() in the patch 3 adds ARP
> response flows for some scenarios
> and the function build_lrouter_flows_ipv4_input_table_3_op() in patch
> 4 adds ARP response flows
> for other scenarios.
>
> I think it's better to revisit patches 3 to 7 and move out the code
> into functions which separates
> the lflow generation properly.
>
> I also think the function names are a bit odd and the naming can be
> improved.
>
> I found other patches in the series to be fine except for the function
> naming.
> Hence I reworked a bit renaming the functions and moving the comments
> from the function
> declarations to near the function definitions. I think this will be
> more helpful.
> And I applied the patches 1,2, 8 to 16 to master.
Cool, thanks. I will rebase my branch off that.
>
> Please note I also removed the marker comment while committing the
> patch 16. I think we can remove it
> as it's a bit odd to keep the comment.
Cool.
I found that in its absence, diff generates stuff that is not readable
and impossible to rebase even after minimal changes.
In the meantime, I reworked the actual multi-threaded processing patch
so that it unifies lswitch and lrouter processing and runs the worker
trheadpool once per iteration. I am going to re-test that - it should
improve performance and scalability a bit further.
A.
>
> Thanks
> Numan
>
> ---
> northd/ovn-northd.c | 135
> +++++++++++++++++++++++++++-----------------
> 1 file changed, 82 insertions(+), 53 deletions(-)
>
> diff --git a/northd/ovn-northd.c b/northd/ovn-northd.c
> index b95d6cd8a..14e4a6939 100644
> --- a/northd/ovn-northd.c
> +++ b/northd/ovn-northd.c
> @@ -8482,6 +8482,26 @@ build_lrouter_force_snat_flows(struct hmap
> *lflows, struct ovn_datapath *od,
> ds_destroy(&actions);
> }
>
> +
> +/* Logical router ingress table 0: Admission control framework. */
> +static void
> +build_lrouter_flows_ingress_table_0_od(
> + struct ovn_datapath *od, struct hmap *lflows);
> +
> +/* Logical router ingress table 0: match (priority 50). */
> +static void
> +build_lrouter_flows_ingress_table_0_op(
> + struct ovn_port *op, struct hmap *lflows,
> + struct ds *match, struct ds *actions);
> +
> +/*
> + * Do not remove this comment - it is here on purpose
> + * It serves as a marker so that pulling operations out
> + * of build_lrouter_flows results in a clean diff with
> + * a separate new operations function and clean changes
> + * to build_lroute_flows
> + */
> +
> static void
> build_lrouter_flows(struct hmap *datapaths, struct hmap *ports,
> struct hmap *lflows, struct shash *meter_groups,
> @@ -8493,65 +8513,14 @@ build_lrouter_flows(struct hmap
> *datapaths, struct hmap *ports,
> struct ds match = DS_EMPTY_INITIALIZER;
> struct ds actions = DS_EMPTY_INITIALIZER;
>
> - /* Logical router ingress table 0: Admission control
> framework. */
> struct ovn_datapath *od;
> HMAP_FOR_EACH (od, key_node, datapaths) {
> - if (!od->nbr) {
> - continue;
> - }
> -
> - /* Logical VLANs not supported.
> - * Broadcast/multicast source address is invalid. */
> - ovn_lflow_add(lflows, od, S_ROUTER_IN_ADMISSION, 100,
> - "vlan.present || eth.src[40]", "drop;");
> + build_lrouter_flows_ingress_table_0_od(od, lflows);
> }
>
> - /* Logical router ingress table 0: match (priority 50). */
> struct ovn_port *op;
> HMAP_FOR_EACH (op, key_node, ports) {
> - if (!op->nbrp) {
> - continue;
> - }
> -
> - if (!lrport_is_enabled(op->nbrp)) {
> - /* Drop packets from disabled logical ports (since
> logical flow
> - * tables are default-drop). */
> - continue;
> - }
> -
> - if (op->derived) {
> - /* No ingress packets should be received on a
> chassisredirect
> - * port. */
> - continue;
> - }
> -
> - /* Store the ethernet address of the port receiving the
> packet.
> - * This will save us from having to match on inport
> further down in
> - * the pipeline.
> - */
> - ds_clear(&actions);
> - ds_put_format(&actions, REG_INPORT_ETH_ADDR " = %s; next;",
> - op->lrp_networks.ea_s);
> -
> - ds_clear(&match);
> - ds_put_format(&match, "eth.mcast && inport == %s",
> op->json_key);
> - ovn_lflow_add_with_hint(lflows, op->od,
> S_ROUTER_IN_ADMISSION, 50,
> - ds_cstr(&match), ds_cstr(&actions),
> - &op->nbrp->header_);
> -
> - ds_clear(&match);
> - ds_put_format(&match, "eth.dst == %s && inport == %s",
> - op->lrp_networks.ea_s, op->json_key);
> - if (op->od->l3dgw_port && op == op->od->l3dgw_port
> - && op->od->l3redirect_port) {
> - /* Traffic with eth.dst = l3dgw_port->lrp_networks.ea_s
> - * should only be received on the "redirect-chassis". */
> - ds_put_format(&match, " && is_chassis_resident(%s)",
> - op->od->l3redirect_port->json_key);
> - }
> - ovn_lflow_add_with_hint(lflows, op->od,
> S_ROUTER_IN_ADMISSION, 50,
> - ds_cstr(&match), ds_cstr(&actions),
> - &op->nbrp->header_);
> + build_lrouter_flows_ingress_table_0_op(op, lflows,
> &match, &actions);
> }
>
> /* Logical router ingress table 1: LOOKUP_NEIGHBOR and
> @@ -10956,6 +10925,66 @@ build_lrouter_flows(struct hmap
> *datapaths, struct hmap *ports,
> ds_destroy(&actions);
> }
>
> +static void
> +build_lrouter_flows_ingress_table_0_od(
> + struct ovn_datapath *od, struct hmap *lflows)
> +{
> + if (od->nbr) {
> + /* Logical VLANs not supported.
> + * Broadcast/multicast source address is invalid. */
> + ovn_lflow_add(lflows, od, S_ROUTER_IN_ADMISSION, 100,
> + "vlan.present || eth.src[40]", "drop;");
> + }
> +}
> +
> +static void
> +build_lrouter_flows_ingress_table_0_op(
> + struct ovn_port *op, struct hmap *lflows,
> + struct ds *match, struct ds *actions)
> +{
> + if (op->nbrp) {
> + if (!lrport_is_enabled(op->nbrp)) {
> + /* Drop packets from disabled logical ports (since
> logical flow
> + * tables are default-drop). */
> + return;
> + }
> +
> + if (op->derived) {
> + /* No ingress packets should be received on a
> chassisredirect
> + * port. */
> + return;
> + }
> +
> + /* Store the ethernet address of the port receiving the
> packet.
> + * This will save us from having to match on inport
> further down in
> + * the pipeline.
> + */
> + ds_clear(actions);
> + ds_put_format(actions, REG_INPORT_ETH_ADDR " = %s; next;",
> + op->lrp_networks.ea_s);
> +
> + ds_clear(match);
> + ds_put_format(match, "eth.mcast && inport == %s",
> op->json_key);
> + ovn_lflow_add_with_hint(lflows, op->od,
> S_ROUTER_IN_ADMISSION, 50,
> + ds_cstr(match), ds_cstr(actions),
> + &op->nbrp->header_);
> +
> + ds_clear(match);
> + ds_put_format(match, "eth.dst == %s && inport == %s",
> + op->lrp_networks.ea_s, op->json_key);
> + if (op->od->l3dgw_port && op == op->od->l3dgw_port
> + && op->od->l3redirect_port) {
> + /* Traffic with eth.dst = l3dgw_port->lrp_networks.ea_s
> + * should only be received on the "redirect-chassis". */
> + ds_put_format(match, " && is_chassis_resident(%s)",
> + op->od->l3redirect_port->json_key);
> + }
> + ovn_lflow_add_with_hint(lflows, op->od,
> S_ROUTER_IN_ADMISSION, 50,
> + ds_cstr(match), ds_cstr(actions),
> + &op->nbrp->header_);
> + }
> +}
> +
> /* Updates the Logical_Flow and Multicast_Group tables in the
> OVN_SB database,
> * constructing their contents based on the OVN_NB database. */
> static void
> --
> 2.20.1
>
> _______________________________________________
> dev mailing list
> dev at openvswitch.org <mailto:dev at openvswitch.org>
> https://mail.openvswitch.org/mailman/listinfo/ovs-dev
>
--
Anton R. Ivanov
Cambridgegreys Limited. Registered in England. Company Number 10273661
https://www.cambridgegreys.com/
More information about the dev
mailing list