3.18-stable review patch. If anyone has any objections, please let me know. ------------------ From: Jesse Gross <jesse@xxxxxxxxxx> [ Upstream commit 5f35227ea34bb616c436d9da47fc325866c428f3 ] GSO isn't the only offload feature with restrictions that potentially can't be expressed with the current features mechanism. Checksum is another although it's a general issue that could in theory apply to anything. Even if it may be possible to implement these restrictions in other ways, it can result in duplicate code or inefficient per-packet behavior. This generalizes ndo_gso_check so that drivers can remove any features that don't make sense for a given packet, similar to netif_skb_features(). It also converts existing driver restrictions to the new format, completing the work that was done to support tunnel protocols since the issues apply to checksums as well. By actually removing features from the set that are used to do offloading, it solves another problem with the existing interface. In these cases, GSO would run with the original set of features and not do anything because it appears that segmentation is not required. CC: Tom Herbert <therbert@xxxxxxxxxx> CC: Joe Stringer <joestringer@xxxxxxxxxx> CC: Eric Dumazet <edumazet@xxxxxxxxxx> CC: Hayes Wang <hayeswang@xxxxxxxxxxx> Signed-off-by: Jesse Gross <jesse@xxxxxxxxxx> Acked-by: Tom Herbert <therbert@xxxxxxxxxx> Fixes: 04ffcb255f22 ("net: Add ndo_gso_check") Tested-by: Hayes Wang <hayeswang@xxxxxxxxxxx> Signed-off-by: David S. Miller <davem@xxxxxxxxxxxxx> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- drivers/net/ethernet/emulex/benet/be_main.c | 8 ++++-- drivers/net/ethernet/mellanox/mlx4/en_netdev.c | 10 ++++---- drivers/net/ethernet/qlogic/qlcnic/qlcnic_main.c | 8 ++++-- include/linux/netdevice.h | 20 +++++++++------- include/net/vxlan.h | 28 +++++++++++++++++++---- net/core/dev.c | 23 +++++++++++------- 6 files changed, 65 insertions(+), 32 deletions(-) --- a/drivers/net/ethernet/emulex/benet/be_main.c +++ b/drivers/net/ethernet/emulex/benet/be_main.c @@ -4427,9 +4427,11 @@ static void be_del_vxlan_port(struct net be16_to_cpu(port)); } -static bool be_gso_check(struct sk_buff *skb, struct net_device *dev) +static netdev_features_t be_features_check(struct sk_buff *skb, + struct net_device *dev, + netdev_features_t features) { - return vxlan_gso_check(skb); + return vxlan_features_check(skb, features); } #endif @@ -4460,7 +4462,7 @@ static const struct net_device_ops be_ne #ifdef CONFIG_BE2NET_VXLAN .ndo_add_vxlan_port = be_add_vxlan_port, .ndo_del_vxlan_port = be_del_vxlan_port, - .ndo_gso_check = be_gso_check, + .ndo_features_check = be_features_check, #endif }; --- a/drivers/net/ethernet/mellanox/mlx4/en_netdev.c +++ b/drivers/net/ethernet/mellanox/mlx4/en_netdev.c @@ -2363,9 +2363,11 @@ static void mlx4_en_del_vxlan_port(struc queue_work(priv->mdev->workqueue, &priv->vxlan_del_task); } -static bool mlx4_en_gso_check(struct sk_buff *skb, struct net_device *dev) +static netdev_features_t mlx4_en_features_check(struct sk_buff *skb, + struct net_device *dev, + netdev_features_t features) { - return vxlan_gso_check(skb); + return vxlan_features_check(skb, features); } #endif @@ -2398,7 +2400,7 @@ static const struct net_device_ops mlx4_ #ifdef CONFIG_MLX4_EN_VXLAN .ndo_add_vxlan_port = mlx4_en_add_vxlan_port, .ndo_del_vxlan_port = mlx4_en_del_vxlan_port, - .ndo_gso_check = mlx4_en_gso_check, + .ndo_features_check = mlx4_en_features_check, #endif }; @@ -2432,7 +2434,7 @@ static const struct net_device_ops mlx4_ #ifdef CONFIG_MLX4_EN_VXLAN .ndo_add_vxlan_port = mlx4_en_add_vxlan_port, .ndo_del_vxlan_port = mlx4_en_del_vxlan_port, - .ndo_gso_check = mlx4_en_gso_check, + .ndo_features_check = mlx4_en_features_check, #endif }; --- a/drivers/net/ethernet/qlogic/qlcnic/qlcnic_main.c +++ b/drivers/net/ethernet/qlogic/qlcnic/qlcnic_main.c @@ -504,9 +504,11 @@ static void qlcnic_del_vxlan_port(struct adapter->flags |= QLCNIC_DEL_VXLAN_PORT; } -static bool qlcnic_gso_check(struct sk_buff *skb, struct net_device *dev) +static netdev_features_t qlcnic_features_check(struct sk_buff *skb, + struct net_device *dev, + netdev_features_t features) { - return vxlan_gso_check(skb); + return vxlan_features_check(skb, features); } #endif @@ -531,7 +533,7 @@ static const struct net_device_ops qlcni #ifdef CONFIG_QLCNIC_VXLAN .ndo_add_vxlan_port = qlcnic_add_vxlan_port, .ndo_del_vxlan_port = qlcnic_del_vxlan_port, - .ndo_gso_check = qlcnic_gso_check, + .ndo_features_check = qlcnic_features_check, #endif #ifdef CONFIG_NET_POLL_CONTROLLER .ndo_poll_controller = qlcnic_poll_controller, --- a/include/linux/netdevice.h +++ b/include/linux/netdevice.h @@ -998,12 +998,15 @@ typedef u16 (*select_queue_fallback_t)(s * Callback to use for xmit over the accelerated station. This * is used in place of ndo_start_xmit on accelerated net * devices. - * bool (*ndo_gso_check) (struct sk_buff *skb, - * struct net_device *dev); + * netdev_features_t (*ndo_features_check) (struct sk_buff *skb, + * struct net_device *dev + * netdev_features_t features); * Called by core transmit path to determine if device is capable of - * performing GSO on a packet. The device returns true if it is - * able to GSO the packet, false otherwise. If the return value is - * false the stack will do software GSO. + * performing offload operations on a given packet. This is to give + * the device an opportunity to implement any restrictions that cannot + * be otherwise expressed by feature flags. The check is called with + * the set of features that the stack has calculated and it returns + * those the driver believes to be appropriate. */ struct net_device_ops { int (*ndo_init)(struct net_device *dev); @@ -1153,8 +1156,9 @@ struct net_device_ops { struct net_device *dev, void *priv); int (*ndo_get_lock_subclass)(struct net_device *dev); - bool (*ndo_gso_check) (struct sk_buff *skb, - struct net_device *dev); + netdev_features_t (*ndo_features_check) (struct sk_buff *skb, + struct net_device *dev, + netdev_features_t features); }; /** @@ -3584,8 +3588,6 @@ static inline bool netif_needs_gso(struc netdev_features_t features) { return skb_is_gso(skb) && (!skb_gso_ok(skb, features) || - (dev->netdev_ops->ndo_gso_check && - !dev->netdev_ops->ndo_gso_check(skb, dev)) || unlikely((skb->ip_summed != CHECKSUM_PARTIAL) && (skb->ip_summed != CHECKSUM_UNNECESSARY))); } --- a/include/net/vxlan.h +++ b/include/net/vxlan.h @@ -1,6 +1,9 @@ #ifndef __NET_VXLAN_H #define __NET_VXLAN_H 1 +#include <linux/ip.h> +#include <linux/ipv6.h> +#include <linux/if_vlan.h> #include <linux/skbuff.h> #include <linux/netdevice.h> #include <linux/udp.h> @@ -51,16 +54,33 @@ int vxlan_xmit_skb(struct vxlan_sock *vs __be32 src, __be32 dst, __u8 tos, __u8 ttl, __be16 df, __be16 src_port, __be16 dst_port, __be32 vni, bool xnet); -static inline bool vxlan_gso_check(struct sk_buff *skb) +static inline netdev_features_t vxlan_features_check(struct sk_buff *skb, + netdev_features_t features) { - if ((skb_shinfo(skb)->gso_type & SKB_GSO_UDP_TUNNEL) && + u8 l4_hdr = 0; + + if (!skb->encapsulation) + return features; + + switch (vlan_get_protocol(skb)) { + case htons(ETH_P_IP): + l4_hdr = ip_hdr(skb)->protocol; + break; + case htons(ETH_P_IPV6): + l4_hdr = ipv6_hdr(skb)->nexthdr; + break; + default: + return features;; + } + + if ((l4_hdr == IPPROTO_UDP) && (skb->inner_protocol_type != ENCAP_TYPE_ETHER || skb->inner_protocol != htons(ETH_P_TEB) || (skb_inner_mac_header(skb) - skb_transport_header(skb) != sizeof(struct udphdr) + sizeof(struct vxlanhdr)))) - return false; + return features & ~(NETIF_F_ALL_CSUM | NETIF_F_GSO_MASK); - return true; + return features; } /* IP header + UDP + VXLAN + Ethernet header */ --- a/net/core/dev.c +++ b/net/core/dev.c @@ -2566,7 +2566,7 @@ static netdev_features_t harmonize_featu netdev_features_t netif_skb_features(struct sk_buff *skb) { - const struct net_device *dev = skb->dev; + struct net_device *dev = skb->dev; netdev_features_t features = dev->features; u16 gso_segs = skb_shinfo(skb)->gso_segs; __be16 protocol = skb->protocol; @@ -2574,13 +2574,20 @@ netdev_features_t netif_skb_features(str if (gso_segs > dev->gso_max_segs || gso_segs < dev->gso_min_segs) features &= ~NETIF_F_GSO_MASK; + /* If encapsulation offload request, verify we are testing + * hardware encapsulation features instead of standard + * features for the netdev + */ + if (skb->encapsulation) + features &= dev->hw_enc_features; + if (!vlan_tx_tag_present(skb)) { if (unlikely(protocol == htons(ETH_P_8021Q) || protocol == htons(ETH_P_8021AD))) { struct vlan_ethhdr *veh = (struct vlan_ethhdr *)skb->data; protocol = veh->h_vlan_encapsulated_proto; } else { - return harmonize_features(skb, features); + goto finalize; } } @@ -2598,6 +2605,11 @@ netdev_features_t netif_skb_features(str NETIF_F_HW_VLAN_CTAG_TX | NETIF_F_HW_VLAN_STAG_TX); +finalize: + if (dev->netdev_ops->ndo_features_check) + features &= dev->netdev_ops->ndo_features_check(skb, dev, + features); + return harmonize_features(skb, features); } EXPORT_SYMBOL(netif_skb_features); @@ -2672,13 +2684,6 @@ static struct sk_buff *validate_xmit_skb if (unlikely(!skb)) goto out_null; - /* If encapsulation offload request, verify we are testing - * hardware encapsulation features instead of standard - * features for the netdev - */ - if (skb->encapsulation) - features &= dev->hw_enc_features; - if (netif_needs_gso(dev, skb, features)) { struct sk_buff *segs; -- To unsubscribe from this list: send the line "unsubscribe stable" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html