]> git.proxmox.com Git - mirror_ubuntu-kernels.git/commitdiff
net/mlx5e: IPsec: Fix Software parser inner l3 type setting in case of encapsulation
authorRaed Salem <raeds@nvidia.com>
Wed, 17 Nov 2021 11:33:57 +0000 (13:33 +0200)
committerSaeed Mahameed <saeedm@nvidia.com>
Wed, 1 Dec 2021 06:35:02 +0000 (22:35 -0800)
Current code wrongly uses the skb->protocol field which reflects the
outer l3 protocol to set the inner l3 type in Software Parser (SWP)
fields settings in the ethernet segment (eseg) in flows where inner
l3 exists like in Vxlan over ESP flow, the above method wrongly use
the outer protocol type instead of the inner one. thus breaking cases
where inner and outer headers have different protocols.

Fix by setting the inner l3 type in SWP according to the inner l3 ip
header version.

Fixes: 2ac9cfe78223 ("net/mlx5e: IPSec, Add Innova IPSec offload TX data path")
Signed-off-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Maor Dickman <maord@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_rxtx.c

index fb5397324aa4f2b597a1f1b0bcd355836e0c382f..2db9573a3fe69d9f175c663334cc6c6c14ad5ec0 100644 (file)
@@ -191,7 +191,7 @@ static void mlx5e_ipsec_set_swp(struct sk_buff *skb,
                        eseg->swp_inner_l3_offset = skb_inner_network_offset(skb) / 2;
                        eseg->swp_inner_l4_offset =
                                (skb->csum_start + skb->head - skb->data) / 2;
-                       if (skb->protocol == htons(ETH_P_IPV6))
+                       if (inner_ip_hdr(skb)->version == 6)
                                eseg->swp_flags |= MLX5_ETH_WQE_SWP_INNER_L3_IPV6;
                        break;
                default: