Hi Jacob, Xavier, On 10/23/26 4:38 PM, Xavier Simonart wrote: >> +/* Number of local ports per querier PACKET_OUT. >> + * >> + * (MCAST_QUERIER_PORT_BATCH) * (resubmits in egress pipeline) must be >> + * less than MAX_RESUBMITS. */ >> +#define MCAST_QUERIER_PORT_BATCH 100 >> > I did not point this out in v1 - should we add an assert to check that this > remains correct even if > new egress tables are added? e.g. something like > BUILD_ASSERT_DECL(MCAST_QUERIER_PORT_BATCH * (LOG_PIPELINE_EGRESS_LEN + 6) > < 4096); > where e.g. 6 accounts for tables starting from OFTABLE_SAVE_INPORT which > can be used by egress datapath + some margin. > It is not bullet proof (we would hit an issue if many tables are added > after OFTABLE_EVPN_ARP_LOOKUP and used by egress), but at least we > would cover LOG_PIPELINE_EGRESS_LEN increase ?
Unfortunately, it's even worse. E.g., if one of the egress LSPs is a localnet port, when the packet is sent out that port there's an unknown number of additional resubmits that may happen. I think I'd keep the assert out of v3, TBH. Regards, Dumitru _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
