IP wierdness in 1.3.66

Nick Holloway (Nick.Holloway@alfie.demon.co.uk)
18 Feb 1996 19:50:40 -0000

I decided to give 1.3.66 a go, as I wanted to give path MTU discovery
a trial. I used tcpdump as I thought I would be able to see it in action.
This is the start of a ftp transfer from the Linux machine (me) connected
via PPP to demon.co.uk, talking to a Solaris machine at warwick.ac.uk
(i.e. lots of routers).

18:33:30.403232 it.20 > me.1380: S 1528951013:1528951013(0) win 24820
<mss 1460> (DF)
18:33:30.403232 me.1380 > it.20: S 329833497:329833497(0) ack 1528951014
win 13140 <mss 1460>
18:33:32.432623 it.20 > me.1380: . ack 1 win 24820 (DF)
18:33:32.462614 me.1380 > it.20: P 1:1025(1024) ack 1 win 13140 (DF) [tos 0x8]
18:33:33.732233 me.1380 > it.20: P 1:1025(1024) ack 1 win 13140 (DF) [tos 0x8]
18:33:35.851597 it.20 > me.1380: . ack 1025 win 24820 (DF)
18:33:35.851597 me.1380 > it.20: P 1025:2485(1460) ack 1 win 13140 [tos 0x8]
18:33:37.771021 it.20 > me.1380: . ack 1025 win 24820 (DF)

I notice that the initial packets (and retry) go out with DF set, but
then later packets don't have that set. This seems contrary to what I
expect -- I would have thought all packets would have had the DF bit set.

The another anomoly is multiple acks. I thought the double-acks were
being stamped on, along with implementing delayed acks. I've actually
managed to have a triple-ack!

18:32:55.743633 158.152.1.107.4671 > 158.152.44.128.25:
P 6467:6473(6) ack 734 win 4096
18:32:55.743633 158.152.44.128.25 > 158.152.1.107.4671:
. ack 6473 win 12508
18:32:55.743633 158.152.44.128.25 > 158.152.1.107.4671:
. ack 6473 win 12508
18:32:55.753630 158.152.44.128.25 > 158.152.1.107.4671:
P 734:784(50) ack 6473 win 12508 (DF)

-- 
 `O O'  | Home: Nick.Holloway@alfie.demon.co.uk
// ^ \\ | Work: Nick.Holloway@parallax.co.uk  http://www.parallax.co.uk/~alfie/