On 7/13/2018 10:40 AM, Adrien Mazarguil wrote: > With mlx5, unlike normal flow rules implemented through Verbs for traffic > emitted and received by the application, those targeting different logical > ports of the device (VF representors for instance) are offloaded at the > switch level and must be configured through Netlink (TC interface). > > This patch adds preliminary support to manage such flow rules through the > flow API (rte_flow). > > Instead of rewriting tons of Netlink helpers and as previously suggested by > Stephen [1], this patch introduces a new dependency to libmnl [2] > (LGPL-2.1) when compiling mlx5. > > [1] https://mails.dpdk.org/archives/dev/2018-March/092676.html > [2] https://netfilter.org/projects/libmnl/
Just to highlight this new PMD level dependency to libmnl. tap pmd also uses netlink and vdev_netvsc also does nl communication, perhaps we can discuss unifying netlink usage around this new library. > > Signed-off-by: Adrien Mazarguil <adrien.mazarg...@6wind.com> > Acked-by: Nelio Laranjeiro <nelio.laranje...@6wind.com> > Cc: Yongseok Koh <ys...@mellanox.com> > -- > v2 changes: > > - Added NETLINK_CAP_ACK definition if missing from the host system. This > parameter is also not mandatory anymore and won't prevent creation of > NL sockets when not supported. > - Modified mlx5_nl_flow_nl_ack() and mlx5_nl_flow_init() to consume the > least amount of stack space based on message size, instead of the fixed > MNL_SOCKET_BUFFER_SIZE which is quite large. <...>
