Snipped
>
> > +struct topology_config {
> > +#ifdef RTE_EAL_HWLOC_TOPOLOGY_PROBE
> > + hwloc_topology_t topology;
> > +#endif
> > +
> > + /* domain count */
> > + uint16_t l1_count;
> > + uint16_t l2_count;
> > + uint8_t l3_count;
> > + uint8_t io_count;
> > +
> > + /* total cores under all domain */
> > + uint16_t l1_core_count;
> > + uint16_t l2_core_count;
> > + uint16_t l3_core_count;
> > + uint16_t io_core_count;
> > +
> > + /* two dimensional array for each domain */
> > + struct core_domain_mapping **l1;
> > + struct core_domain_mapping **l2;
> > + struct core_domain_mapping **l3;
> > + struct core_domain_mapping **io; }; extern struct
> > +topology_config topo_cnfg;
> > +
>
> To work with primary/secondary process model, it might be better to keep this
> info
> in hugpage/shared memory. I.e put topology_config into space allocated with
> rte_malloc() and populated by primary process.
Initially we did use `rte_memzone_reserve` and moved it under `rte_cnfg
structure` for primary and secondary.
Exploring on the use cases we noticed in multiprocessor scenarios, lcore of the
primary did not share. Hence we moved to using `malloc` as local allocator.
We can easily move to rte_malloc, but do we need to share the HW topology based
lcore grouping to multi-process?
