Back to 16 bit arithmetic

This commit is contained in:
Nico Schottelius 2019-07-16 12:06:26 +02:00
commit ecab110d7c
3 changed files with 54 additions and 56 deletions

View file

@ -5309,6 +5309,25 @@ This is used to correct endianness!
-> not needed in our case
*** TODO 2019-07-16: Wraptest in P4
#+BEGIN_CENTER
mx h3
arp -s 10.0.0.2 00:00:0a:00:00:02
echo V4-OK | socat - TCP:10.0.0.2:2342
#+END_CENTER
Result on the wire of the wraptest:
#+BEGIN_CENTER
10.0.0.3.60106 > 10.0.0.2.2342: Flags [S], cksum 0x0001 (incorrect ->
0xc7f9), seq 3677403557, win 28380, options [mss 9460
,sackOK,TS val 2328087128 ecr 0,nop,wscale 9], length 0
#+END_CENTER
0xffff + 2 = 1
-> as expected
Thus carryover is already implemented. Try to use 16 bit ints
** The NetPFGA saga
Problems encountered:
- The logfile for a compile run is 10k+ lines