[ovs-dev,v2] OVN: add CT_LB action to ovn-trace

Add CT_LB action to ovn-trace utility in order to fix the
following ovn-trace error if a load balancer rule is added to
OVN configuration

ct_next(ct_state=est|trk /* default (use --ct to customize) */) {
	    *** ct_lb action not implemented;
};

Add '--lb_dst' option in order to specify the ip address to use
in VIP pool. If --lb_dst is not provided the destination ip will be
randomly choosen

Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
---
Changes since v1:
- update ovn-trace manual
- reset to 0 lb_dst when it is consumed by ct_lb action
---
 ovn/utilities/ovn-trace.8.xml | 12 ++++++-
 ovn/utilities/ovn-trace.c     | 64 +++++++++++++++++++++++++++++++++--
 2 files changed, 72 insertions(+), 4 deletions(-)

LGTM

Acked-by: Mark Michelson <mmichels@redhat.com>

On 09/20/2018 10:46 AM, Lorenzo Bianconi wrote:
> Add CT_LB action to ovn-trace utility in order to fix the
> following ovn-trace error if a load balancer rule is added to
> OVN configuration
> 
> ct_next(ct_state=est|trk /* default (use --ct to customize) */) {
> 	    *** ct_lb action not implemented;
> };
> 
> Add '--lb_dst' option in order to specify the ip address to use
> in VIP pool. If --lb_dst is not provided the destination ip will be
> randomly choosen
> 
> Signed-off-by: Lorenzo Bianconi <lorenzo.bianconi@redhat.com>
> ---
> Changes since v1:
> - update ovn-trace manual
> - reset to 0 lb_dst when it is consumed by ct_lb action
> ---
>   ovn/utilities/ovn-trace.8.xml | 12 ++++++-
>   ovn/utilities/ovn-trace.c     | 64 +++++++++++++++++++++++++++++++++--
>   2 files changed, 72 insertions(+), 4 deletions(-)
> 
> diff --git a/ovn/utilities/ovn-trace.8.xml b/ovn/utilities/ovn-trace.8.xml
> index 83b397f3a..3d75c7b86 100644
> --- a/ovn/utilities/ovn-trace.8.xml
> +++ b/ovn/utilities/ovn-trace.8.xml
> @@ -255,7 +255,11 @@
>   
>       <dt><code>ct_lb</code></dt>
>       <dd>
> -      Not yet implemented; currently implemented as a no-op.
> +      Forks the pipeline. In one fork, sets <code>ip4.dst</code> (or
> +      <code>ip6.dst</code>) to the ip provided with <code>--lb_dst</code>
> +      option (or to an ip randomly chosen from VIP pool). In the other fork,
> +      the pipeline continues without change after the <code>ct_lb</code>
> +      action
>       </dd>
>   
>       <dt><code>ct_commit</code></dt>
> @@ -424,6 +428,12 @@
>         </p>
>       </dd>
>   
> +    <dt><code>--lb_dst</code>[<code>=</code><var>ip</var>]</dt>
> +    <dd>
> +      Selects an ip from VIP pool to use as destination of the packet.
> +      lb_dst is not available in daemon mode
> +    </dd>
> +
>       <dt><code>--friendly-names</code></dt>
>       <dt><code>--no-friendly-names</code></dt>
>       <dd>
> diff --git a/ovn/utilities/ovn-trace.c b/ovn/utilities/ovn-trace.c
> index 7ca3d97aa..ee4583deb 100644
> --- a/ovn/utilities/ovn-trace.c
> +++ b/ovn/utilities/ovn-trace.c
> @@ -46,6 +46,7 @@
>   #include "stream.h"
>   #include "unixctl.h"
>   #include "util.h"
> +#include "random.h"
>   
>   VLOG_DEFINE_THIS_MODULE(ovntrace);
>   
> @@ -77,6 +78,9 @@ static uint32_t *ct_states;
>   static size_t n_ct_states;
>   static size_t ct_state_idx;
>   
> +/* --lb_dst: load balancer destination info */
> +static struct ovnact_ct_lb_dst lb_dst;
> +
>   /* --friendly-names, --no-friendly-names: Whether to substitute human-friendly
>    * port and datapath names for the awkward UUIDs typically used in the actual
>    * logical flows. */
> @@ -186,6 +190,16 @@ parse_ct_option(const char *state_s_)
>       ct_states[n_ct_states++] = state;
>   }
>   
> +static void
> +parse_lb_option(const char *s)
> +{
> +    if (ip_parse(s, &lb_dst.ipv4)) {
> +        lb_dst.family = AF_INET;
> +    } else if (ipv6_parse(s, &lb_dst.ipv6)) {
> +        lb_dst.family = AF_INET6;
> +    }
> +}
> +
>   static void
>   parse_options(int argc, char *argv[])
>   {
> @@ -202,7 +216,8 @@ parse_options(int argc, char *argv[])
>           OPT_NO_FRIENDLY_NAMES,
>           DAEMON_OPTION_ENUMS,
>           SSL_OPTION_ENUMS,
> -        VLOG_OPTION_ENUMS
> +        VLOG_OPTION_ENUMS,
> +        OPT_LB_DST
>       };
>       static const struct option long_options[] = {
>           {"db", required_argument, NULL, OPT_DB},
> @@ -217,6 +232,7 @@ parse_options(int argc, char *argv[])
>           {"no-friendly-names", no_argument, NULL, OPT_NO_FRIENDLY_NAMES},
>           {"help", no_argument, NULL, 'h'},
>           {"version", no_argument, NULL, 'V'},
> +        {"lb_dst", required_argument, NULL, OPT_LB_DST},
>           DAEMON_LONG_OPTIONS,
>           VLOG_LONG_OPTIONS,
>           STREAM_SSL_LONG_OPTIONS,
> @@ -274,6 +290,10 @@ parse_options(int argc, char *argv[])
>               use_friendly_names = false;
>               break;
>   
> +        case OPT_LB_DST:
> +            parse_lb_option(optarg);
> +            break;
> +
>           case 'h':
>               usage();
>   
> @@ -1822,6 +1842,45 @@ execute_ct_nat(const struct ovnact_ct_nat *ct_nat,
>        * flow, not ct_flow. */
>   }
>   
> +static void
> +execute_ct_lb(const struct ovnact_ct_lb *ct_lb,
> +              const struct ovntrace_datapath *dp, struct flow *uflow,
> +              enum ovnact_pipeline pipeline, struct ovs_list *super)
> +{
> +    struct flow ct_lb_flow = *uflow;
> +
> +    if (ct_lb->n_dsts) {
> +        int i, dest = random_range(ct_lb->n_dsts);
> +
> +        for (i = 0; i < ct_lb->n_dsts; i++) {
> +            if ((lb_dst.family == AF_INET &&
> +                 ct_lb->dsts[i].ipv4 == lb_dst.ipv4) ||
> +                (lb_dst.family == AF_INET6 &&
> +                 ipv6_addr_equals(&ct_lb->dsts[i].ipv6,
> +                                  &lb_dst.ipv6))) {
> +                dest = i;
> +                break;
> +            }
> +        }
> +        if (i < ct_lb->n_dsts) {
> +            memset(&lb_dst, 0, sizeof(struct ovnact_ct_lb_dst));
> +        }
> +
> +        if (ct_lb->dsts->family == AF_INET6) {
> +            ct_lb_flow.ipv6_dst = ct_lb->dsts[dest].ipv6;
> +        } else {
> +            ct_lb_flow.nw_dst = ct_lb->dsts[dest].ipv4;
> +        }
> +        if (ct_lb->dsts->port > 0) {
> +            ct_lb_flow.tp_dst = ct_lb->dsts->port;
> +        }
> +    }
> +
> +    struct ovntrace_node *node = ovntrace_node_append(
> +        super, OVNTRACE_NODE_TRANSFORMATION, "ct_lb");
> +    trace__(dp, &ct_lb_flow, ct_lb->ltable, pipeline, &node->subs);
> +}
> +
>   static void
>   execute_log(const struct ovnact_log *log, struct flow *uflow,
>               struct ovs_list *super)
> @@ -1910,8 +1969,7 @@ trace_actions(const struct ovnact *ovnacts, size_t ovnacts_len,
>               break;
>   
>           case OVNACT_CT_LB:
> -            ovntrace_node_append(super, OVNTRACE_NODE_ERROR,
> -                                 "*** ct_lb action not implemented");
> +            execute_ct_lb(ovnact_get_CT_LB(a), dp, uflow, pipeline, super);
>               break;
>   
>           case OVNACT_CT_CLEAR:
>

On Fri, Sep 21, 2018 at 03:45:00PM +0200, Lorenzo Bianconi wrote:
> > Here's a version that improves on these dimensions.  I have not tested
> > it.  What do you think?
> 
> thx for fixing these bits. The patch works properly, just one comment inline

...

> > +        } else {
> > +            ovntrace_node_append(super, OVNTRACE_NODE_ERROR,
> > +                                 "*** no load balancing destination "
> > +                                 "(use --lb-dst)");
> 
> This codepath is hit when you provide --ct option set to 'new' since in the
> switch egress pipeline ct_lb->n_dsts is 0 and lb_dst.family is set to
> AF_UNSPEC (but the processing continues properly). Is it ok to remove it or
> maybe check if pipeline is set to OVNACT_P_INGRESS?

OK, I moved this so that it only triggers if ct_lb->n_dsts != 0, which I
think should avoid the problem.  With that change, I applied it to
master.

Thanks again!