vdpa/mlx5: Fix rule forwarding VLAN to TIR
authorEli Cohen <elic@nvidia.com>
Mon, 14 Nov 2022 13:17:52 +0000 (15:17 +0200)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Thu, 12 Jan 2023 11:02:05 +0000 (12:02 +0100)
[ Upstream commit a6ce72c0fb6041f9871f880b2d02b294f7f49cb4 ]

Set the VLAN id to the header values field instead of overwriting the
headers criteria field.

Before this fix, VLAN filtering would not really work and tagged packets
would be forwarded unfiltered to the TIR.

Fixes: baf2ad3f6a98 ("vdpa/mlx5: Add RX MAC VLAN filter support")
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Eli Cohen <elic@nvidia.com>
Message-Id: <20221114131759.57883-2-elic@nvidia.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
drivers/vdpa/mlx5/net/mlx5_vnet.c

index 9091336..3fb06dc 100644 (file)
@@ -1468,11 +1468,13 @@ static int mlx5_vdpa_add_mac_vlan_rules(struct mlx5_vdpa_net *ndev, u8 *mac,
        dmac_v = MLX5_ADDR_OF(fte_match_param, headers_v, outer_headers.dmac_47_16);
        eth_broadcast_addr(dmac_c);
        ether_addr_copy(dmac_v, mac);
-       MLX5_SET(fte_match_set_lyr_2_4, headers_c, cvlan_tag, 1);
+       if (ndev->mvdev.actual_features & BIT_ULL(VIRTIO_NET_F_CTRL_VLAN)) {
+               MLX5_SET(fte_match_set_lyr_2_4, headers_c, cvlan_tag, 1);
+               MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, headers_c, first_vid);
+       }
        if (tagged) {
                MLX5_SET(fte_match_set_lyr_2_4, headers_v, cvlan_tag, 1);
-               MLX5_SET_TO_ONES(fte_match_set_lyr_2_4, headers_c, first_vid);
-               MLX5_SET(fte_match_set_lyr_2_4, headers_c, first_vid, vid);
+               MLX5_SET(fte_match_set_lyr_2_4, headers_v, first_vid, vid);
        }
        flow_act.action = MLX5_FLOW_CONTEXT_ACTION_FWD_DEST;
        dest.type = MLX5_FLOW_DESTINATION_TYPE_TIR;