These patches had merge conflicts that needed to be resolved when applying:
* selftests: fib_rule_tests: Fix icmp proto with ipv6
- minor offset adjustment
* udp: fix GRO reception in case of length mismatch
- Our tree has had some touch-up work to the variables here. The stable tree was using something like this: "uh->len > uh2->len", but our tree uses "ulen > ntohs(uh2->len)" -- where ulen is declared further up in the function (from uh->len) and we're using "ntohs" on the other value.
These patches had merge conflicts that needed to be resolved when applying:
* selftests: fib_rule_tests: Fix icmp proto with ipv6
- minor offset adjustment
* udp: fix GRO reception in case of length mismatch
- Our tree has had some touch-up work to the variables here. The stable tree was using something like this: "uh->len > uh2->len", but our tree uses "ulen > ntohs(uh2->len)" -- where ulen is declared further up in the function (from uh->len) and we're using "ntohs" on the other value.