On Fri, 27 Apr 2018 01:03:30 +0300
Medvedkin Vladimir <medvedk...@gmail.com> wrote:
> This patch series introduces new library librte_rib which potentially could
> replace librte_lpm.
>
> RIB is an alternative to current LPM library.
> It solves the following problems
> - Increases the speed of control plane operations against lpm such as
> adding/deleting routes
> - Adds abstraction from dataplane algorithms, so it is possible to add
> different ip route lookup algorythms such as DXR/poptrie/lpc-trie/etc
> in addition to current dir24_8
> - It is possible to keep user defined application specific additional
> information in struct rte_rib_node which represents route entry.
> It can be next hop/set of next hops (i.e. active and feasible),
> pointers to link rte_rib_node based on some criteria (i.e. next_hop),
> plenty of additional control plane information.
>
> v4:
> fix various bugs
> make struct rte_rib opaque
> make inline functions instead of macro
> remove RTE_RIB_MALLOC node allocation type
> add new lookup functions
> remove rte_dir24_8_lookup()
> add rte_dir24_8_get_lookup()
> add meson support
> add fib configuration
>
>
> Medvedkin Vladimir (4):
> Add RIB library
> Add dir24_8 implementation for rib library
> Add autotests for RIB library
The existing DPDK LPM code does need more work it does trade space for time.
It does have some advantages though:
* LPM lookup table is independent of number of routes.
* LPM lookup table can be lock free reader safe using RCU.
The existing slowness of add and delete was fixed at Vyatta/Brocade by
replacing list with red-black tree. Patches were submitted but never merged.
