[PATCH 4.4 08/15] revert "net: account for current skb length when deciding about UFO"

From: Greg Kroah-Hartman
Date: Fri Aug 11 2017 - 18:07:48 EST


4.4-stable review patch. If anyone has any objections, please let me know.

------------------

This reverts commit ef09c9ff343122a0b245416066992d096416ff19 which is
commit a5cb659bbc1c8644efa0c3138a757a1e432a4880 upstream as it causes
merge issues with later patches that are much more important...

Cc: Michal Kubecek <mkubecek@xxxxxxx>
Cc: Vlad Yasevich <vyasevic@xxxxxxxxxx>
Cc: David S. Miller <davem@xxxxxxxxxxxxx>
Cc: Sasha Levin <alexander.levin@xxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>

---
net/ipv4/ip_output.c | 3 +--
net/ipv6/ip6_output.c | 2 +-
2 files changed, 2 insertions(+), 3 deletions(-)

--- a/net/ipv4/ip_output.c
+++ b/net/ipv4/ip_output.c
@@ -922,8 +922,7 @@ static int __ip_append_data(struct sock
csummode = CHECKSUM_PARTIAL;

cork->length += length;
- if ((((length + (skb ? skb->len : fragheaderlen)) > mtu) ||
- (skb && skb_is_gso(skb))) &&
+ if ((((length + fragheaderlen) > mtu) || (skb && skb_is_gso(skb))) &&
(sk->sk_protocol == IPPROTO_UDP) &&
(rt->dst.dev->features & NETIF_F_UFO) && !rt->dst.header_len &&
(sk->sk_type == SOCK_DGRAM) && !sk->sk_no_check_tx) {
--- a/net/ipv6/ip6_output.c
+++ b/net/ipv6/ip6_output.c
@@ -1357,7 +1357,7 @@ emsgsize:
*/

cork->length += length;
- if ((((length + (skb ? skb->len : headersize)) > mtu) ||
+ if ((((length + fragheaderlen) > mtu) ||
(skb && skb_is_gso(skb))) &&
(sk->sk_protocol == IPPROTO_UDP) &&
(rt->dst.dev->features & NETIF_F_UFO) &&