[NET]: IPV6 checksum offloading in network devices

The existing model for checksum offload does not correctly handle
devices that can offload IPV4 and IPV6 only. The NETIF_F_HW_CSUM flag
implies device can do any arbitrary protocol.

This patch:
 * adds NETIF_F_IPV6_CSUM for those devices
 * fixes bnx2 and tg3 devices that need it
 * add NETIF_F_IPV6_CSUM to ipv6 output (incl GSO)
 * fixes assumptions about NETIF_F_ALL_CSUM in nat
 * adjusts bridge union of checksumming computation

Signed-off-by: David S. Miller <davem@davemloft.net>

1 files changed, 9 insertions, 1 deletions

diff --git a/net/bridge/br_if.c b/net/bridge/br_if.c
index 849deaf14108..7b4ce9113be2 100644
--- a/net/bridge/br_if.c
+++ b/net/bridge/br_if.c
@@ -368,10 +368,18 @@ void br_features_recompute(struct net_bridge *br)
         list_for_each_entry(p, &br->port_list, list) {
                 unsigned long feature = p->dev->features;
 
+                /* if device needs checksumming, downgrade to hw checksumming */
                 if (checksum & NETIF_F_NO_CSUM && !(feature & NETIF_F_NO_CSUM))
                         checksum ^= NETIF_F_NO_CSUM | NETIF_F_HW_CSUM;
+
+                /* if device can't do all checksum, downgrade to ipv4/ipv6 */
                 if (checksum & NETIF_F_HW_CSUM && !(feature & NETIF_F_HW_CSUM))
-                        checksum ^= NETIF_F_HW_CSUM | NETIF_F_IP_CSUM;
+                        checksum ^= NETIF_F_HW_CSUM
+                                | NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM;
+
+                if (checksum & NETIF_F_IPV6_CSUM && !(feature & NETIF_F_IPV6_CSUM))
+                        checksum &= ~NETIF_F_IPV6_CSUM;
+
                 if (!(feature & NETIF_F_IP_CSUM))
                         checksum = 0;