From: "Jason A. Donenfeld" <Jason@xxxxxxxxx> Date: Fri, 29 Nov 2019 04:32:05 +0100 > I'm not a huge fan of doing manual skb surgery either. The annoying > thing here is that skb_gso_segment returns a list of skbs that's > terminated by the last one's next pointer being NULL. I assume it's this > way so that the GSO code doesn't have to pass a head around. Sorry, I missed that this was processing a GSO list which doesn't use double linked list semantics. So ignore my feedback on this one :-)