On Mon, Nov 8, 2021 at 12:32 PM Pavan Nikhilesh Bhagavatula <pbhagavatula at marvell.com> wrote: > We need to further split Rx/Tx functions in both net/cnxk and event/cnxk to enable more parallelism. > This is planned for the next release. Looking fwd to this change. -- David Marchand