[dpdk-stable] patch 'net/mlx5: fix flow type for allmulti rules' has been queued to LTS release 17.11.1
Yuanhan Liu
yliu at fridaylinux.org
Tue Feb 13 12:56:48 CET 2018
On Wed, Jan 24, 2018 at 11:31:49PM +0800, Yuanhan Liu wrote:
> Hi,
>
> FYI, your patch has been queued to LTS release 17.11.1
>
> Note it hasn't been pushed to http://dpdk.org/browse/dpdk-stable yet.
> It will be pushed if I get no objections before 01/26/18. So please
> shout if anyone has objections.
Per asked Shahaf, this one is now reverted, due to there is a bug.
--yliu
>
> ---
> >From b187189e1371cba301f85e1f11c3c7d0211e4a0a Mon Sep 17 00:00:00 2001
> From: Raslan Darawsheh <rasland at mellanox.com>
> Date: Tue, 5 Dec 2017 11:37:50 +0200
> Subject: [PATCH] net/mlx5: fix flow type for allmulti rules
>
> [ upstream commit 0a40a1363a4df8043f9a7e93289eda2ea52815a7 ]
>
> Chnaged ibv_flow_attr type for allmulti rule to IBV_FLOW_ATTR_MC_DEFAULT
> instead of IBV_FLOW_ATTR_NORMAL, in case allmulti was enabled.
>
> Fixes: 272733b5 ("net/mlx5: use flow to enable unicast traffic")
>
> Signed-off-by: Raslan Darawsheh <rasland at mellanox.com>
> Acked-by: Nelio Laranjeiro <nelio.laranjeiro at 6wind.com>
> ---
> drivers/net/mlx5/mlx5_flow.c | 13 +++++++++++++
> 1 file changed, 13 insertions(+)
>
> diff --git a/drivers/net/mlx5/mlx5_flow.c b/drivers/net/mlx5/mlx5_flow.c
> index f32dfdd..75ea299 100644
> --- a/drivers/net/mlx5/mlx5_flow.c
> +++ b/drivers/net/mlx5/mlx5_flow.c
> @@ -432,6 +432,7 @@ static const struct mlx5_flow_items mlx5_flow_items[] = {
> /** Structure to pass to the conversion function. */
> struct mlx5_flow_parse {
> uint32_t inner; /**< Set once VXLAN is encountered. */
> + uint32_t allmulti:1; /**< Set once allmulti dst MAC is encountered. */
> uint32_t create:1;
> /**< Whether resources should remain after a validate. */
> uint32_t drop:1; /**< Target is a drop queue. */
> @@ -1206,6 +1207,17 @@ exit_free:
> }
> }
> }
> + if (parser->allmulti &&
> + parser->layer == HASH_RXQ_ETH) {
> + for (i = 0; i != hash_rxq_init_n; ++i) {
> + if (!parser->queue[i].ibv_attr)
> + continue;
> + if (parser->queue[i].ibv_attr->num_of_specs != 1)
> + break;
> + parser->queue[i].ibv_attr->type =
> + IBV_FLOW_ATTR_MC_DEFAULT;
> + }
> + }
> return ret;
> exit_enomem:
> for (i = 0; i != hash_rxq_init_n; ++i) {
> @@ -1311,6 +1323,7 @@ mlx5_flow_create_eth(const struct rte_flow_item *item,
> eth.val.ether_type &= eth.mask.ether_type;
> }
> mlx5_flow_create_copy(parser, ð, eth_size);
> + parser->allmulti = eth.val.dst_mac[0] & 1;
> return 0;
> }
>
> --
> 2.7.4
More information about the stable
mailing list