On Thu, 2015-10-29 at 16:00 +0100, Paolo Abeni wrote:
> This patch do not add dissection code: it use the information provided
> by the available flowi4 structure. Moreover the skb is not available on
> the calling site (in __ip_route_output_key_hash) and pushing it all the
> way will require a lot of intrusive changes. Do you think it's the
> better option ?
If skb is provided, then we could use its information.
Local TCP flows now provide l4 hash in skb, and automatically shuffles
it in case of dst failures.
This provides resilience over dead paths, over variety of
aggregation-like devices (bonding being an example)
# git grep -n4 sk_rethink_txhash
include/net/sock.h-1695-{
include/net/sock.h-1696- sk->sk_txhash = net_tx_rndhash();
include/net/sock.h-1697-}
include/net/sock.h-1698-
include/net/sock.h:1699:static inline void sk_rethink_txhash(struct sock *sk)
include/net/sock.h-1700-{
include/net/sock.h-1701- if (sk->sk_txhash)
include/net/sock.h-1702- sk_set_txhash(sk);
include/net/sock.h-1703-}
--
include/net/sock.h-1725-static inline void dst_negative_advice(struct sock *sk)
include/net/sock.h-1726-{
include/net/sock.h-1727- struct dst_entry *ndst, *dst = __sk_dst_get(sk);
include/net/sock.h-1728-
include/net/sock.h:1729: sk_rethink_txhash(sk);
include/net/sock.h-1730-
include/net/sock.h-1731- if (dst && dst->ops->negative_advice) {
include/net/sock.h-1732- ndst = dst->ops->negative_advice(dst);
include/net/sock.h-1733-
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
