* Added new ``RTE_ACL_CLASSIFY_AVX512X32`` vector implementation,
which can process up to 32 flows in parallel. Requires AVX512 support.
+* **Added support to update subport bandwidth dynamically.**
+
+ * Added new API ``rte_sched_port_subport_profile_add`` to add new
+ subport bandwidth profile to subport porfile table at runtime.
+
+ * Added support to update subport rate dynamically.
+
Removed Items
-------------
This enum value was not used inside DPDK, while it prevented to add new
classify algorithms without causing an ABI breakage.
+* sched: Added ``subport_profile_id`` as argument
+ to function ``rte_sched_subport_config``.
+
ABI Changes
-----------