[ovs-dev] [PATCH 6/8] ofproto-dpif-xlate: Do not add more MPLS LSEs than the datapath can handle

[Simon Horman]

On Thu, Jan 16, 2014 at 04:10:43PM -0800, Ben Pfaff wrote:
> On Wed, Jan 15, 2014 at 04:13:23PM +0900, Simon Horman wrote:
> > This is an proposed enhancement to
> > "Implement OpenFlow support for MPLS, for up to 3 labels."
> > 
> > This makes the conservative assumption that the datapath should
> > not be asked to apply MPLS push actions such that the resulting
> > packet will have an MPLS label stack depth greater than the
> > datapath can accept in a match.
> > 
> > In future it may be possible to use recirculation to handle
> > such cases.
> > 
> > Signed-off-by: Simon Horman <horms at verge.net.au>
> 
> I think that we can handle these in userspace with SLOW_ACTION, so I
> changed this patch to just the following:

Does this approach work in the case where
n < ctx->xbridge->max_mpls_depth && n >= ARRAY_SIZE(flow->mpls_lse) ?

> 
> diff --git a/ofproto/ofproto-dpif-xlate.c b/ofproto/ofproto-dpif-xlate.c
> index 28079be..017074e 100644
> --- a/ofproto/ofproto-dpif-xlate.c
> +++ b/ofproto/ofproto-dpif-xlate.c
> @@ -51,6 +51,7 @@
>  
>  COVERAGE_DEFINE(xlate_actions);
>  COVERAGE_DEFINE(xlate_actions_oversize);
> +COVERAGE_DEFINE(xlate_actions_mpls_overflow);
>  
>  VLOG_DEFINE_THIS_MODULE(ofproto_dpif_xlate);
>  
> @@ -2127,17 +2128,9 @@ compose_mpls_push_action(struct xlate_ctx *ctx, struct ofpact_push_mpls *mpls)
>          ctx->xout->slow |= commit_odp_actions(flow, &ctx->base_flow,
>                                                &ctx->xout->odp_actions,
>                                                &ctx->xout->wc);
> -    } else if (n >= ARRAY_SIZE(flow->mpls_lse)) {
> -        if (ctx->xin->packet != NULL) {
> -            static struct vlog_rate_limit rl = VLOG_RATE_LIMIT_INIT(1, 5);
> -            VLOG_WARN_RL(&rl, "bridge %s: dropping packet which after an "
> -                         "MPLS push action will have more MPLS LSEs than "
> -                         "the %"PRIuSIZE" that can be handled.",
> -                         ctx->xbridge->name, ARRAY_SIZE(flow->mpls_lse));
> -        }
> -        ctx->exit = true;
> -        ofpbuf_clear(&ctx->xout->odp_actions);
> -        return;
> +    } else if (n >= ctx->xbridge->max_mpls_depth) {
> +        COVERAGE_INC(xlate_actions_mpls_overflow);
> +        ctx->xout->slow |= SLOW_ACTION;
>      }
>  
>      flow_push_mpls(flow, mpls->ethertype, wc);
>