tunnels: fix kasan splat when generating ipv4 pmtu error

If we try to emit an icmp error in response to a nonliner skb, we get

BUG: KASAN: slab-out-of-bounds in ip_compute_csum+0x134/0x220
Read of size 4 at addr ffff88811c50db00 by task iperf3/1691
CPU: 2 PID: 1691 Comm: iperf3 Not tainted 6.5.0-rc3+ Freescale#309
[..]
 kasan_report+0x105/0x140
 ip_compute_csum+0x134/0x220
 iptunnel_pmtud_build_icmp+0x554/0x1020
 skb_tunnel_check_pmtu+0x513/0xb80
 vxlan_xmit_one+0x139e/0x2ef0
 vxlan_xmit+0x1867/0x2760
 dev_hard_start_xmit+0x1ee/0x4f0
 br_dev_queue_push_xmit+0x4d1/0x660
 [..]

ip_compute_csum() cannot deal with nonlinear skbs, so avoid it.
After this change, splat is gone and iperf3 is no longer stuck.

Fixes: 4cb47a8 ("tunnels: PMTU discovery support for directly bridged IP packets")
Signed-off-by: Florian Westphal <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>

net/ipv4/ip_tunnel_core.c

@@ -224,7 +224,7 @@ static int iptunnel_pmtud_build_icmp(struct sk_buff *skb, int mtu)
 		.un.frag.__unused	= 0,
 		.un.frag.mtu		= htons(mtu),
 	};
-	icmph->checksum = ip_compute_csum(icmph, len);
+	icmph->checksum = csum_fold(skb_checksum(skb, 0, len, 0));
 	skb_reset_transport_header(skb);
 
 	niph = skb_push(skb, sizeof(*niph));