Check how wrapping is done in P4

doc/plan.org

@@ -5274,6 +5274,40 @@ Still using scapy + adjusted diff:
 gets
 
 sum 0xc5f2 (incorrect -> 0xe7cf)
+*** 2019-07-16: pcap for debugging tcp
+    pcap/tcp-udp-delta-2019-07-15-1454-h3.pcap
+    pcap/tcp-udp-delta-2019-07-15-1454-h1.pcap
+*** 2019-07-16: pcap for debugging udp
+    pcap/tcp-udp-delta-2019-07-16-0856-h3.pcap
+    pcap/tcp-udp-delta-2019-07-16-0856-h1.pcap
+*** 2019-07-16: testing with udp
+    - suspicion: missing some cases in tcp6
+
+#+BEGIN_CENTER
+p4@ubuntu:~$ mx h1 "echo V6-OK | socat - UDP6-LISTEN:2342"
+p4@ubuntu:~/master-thesis/bin$ mx h3 "echo V4-OK | socat - UDP:10.1.1.1:2342"
+#+END_CENTER
+
+08:57:19.998299 IP6 (hlim 64, next-header UDP (17) payload length: 14)
+2001:db8:1::a00:1.59997 > 2001:db8::1.2342: [bad udp cksum 0xd84c ->
+0xd84b!] UDP, length 6
+
+**** DONE Result: arp (partly?) missing?
+     CLOSED: [2019-07-16 Tue 11:00]
+
+**** TODO Result: udp checksum wrong
+     - should be: 0xd84b
+     - is: 0xd84c
+
+My code is +1 too high
+
+*** DONE 2019-07-16: understood scapy code
+    CLOSED: [2019-07-16 Tue 11:25]
+meta.v6sum = (bit<16>) ((((tmp>>8) & 0xff)|tmp<<8) & 0xffff) ;
+
+This is used to correct endianness!
+
+-> not needed in our case
 
 ** The NetPFGA saga
 Problems encountered: