On 8/7/24 22:53, Mike Pattrick wrote:
> In deployments with multiple tunnels, it can be possible to enter an
> infinite loop where traffic generates an arp/nd lookup which is
> forwarded to a different tunnel, generating a new arp/nd packet that is
> send back to the first tunnel. Now the depth counter is incremented for
> new arp/nd packets, just as happens in xlate_table_action().
> 
> Signed-off-by: Mike Pattrick <[email protected]>
> ---
>  ofproto/ofproto-dpif-xlate.c | 8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
> 
> diff --git a/ofproto/ofproto-dpif-xlate.c b/ofproto/ofproto-dpif-xlate.c
> index 850597b3a..887893417 100644
> --- a/ofproto/ofproto-dpif-xlate.c
> +++ b/ofproto/ofproto-dpif-xlate.c
> @@ -679,6 +679,7 @@ static size_t count_skb_priorities(const struct xport *);
>  static bool dscp_from_skb_priority(const struct xport *, uint32_t 
> skb_priority,
>                                     uint8_t *dscp);
>  
> +static bool xlate_resubmit_resource_check(struct xlate_ctx *ctx);

Thanks, Mike!  I removed the unnecessary argument name here and applied
the change.  It also seemed reasonable to backport the fix to the to-be-lts
branch-3.3, so I did that.

Best regards, Ilya Maximets.
_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

Reply via email to