Duplicated packets cause read error?

[zouyonghao]

I tested f-stack with the following packetdrill script

// DO NOT delete or add line in this file

// Establish a connection.
0   socket(..., SOCK_STREAM, IPPROTO_TCP) = 3
+0  setsockopt(3, SOL_SOCKET, SO_REUSEADDR, [1], 4) = 0
+0 fcntl(3, F_SETFL, O_RDWR|O_NONBLOCK) = 0
+0  bind(3, ..., ...) = 0
+0  listen(3, 1) = 0

+0  < S 0:0(0) win 32792 <mss 1000,sackOK,TS val 0 ecr 0,nop,wscale 7>
+.5  > S. 0:0(0) ack 1 <...>

+.1 < . 1:1(0) ack 1 win 32792 // do not send TS when testing f-stack by default
+0.1 accept(3, ..., ...) = 4

+0 < . 1:2001(2000) ack 1 win 25726
+0 < . 1:4091(4090) ack 1 win 25198
+0.1 read(4, ..., 20290) =  4090

The read results contains some error, but same script is executed correct in FreeBSD.

I'm not sure whether my packetdrill implementation for f-stack is correct, does f-stack team ever test f-stack with packetdrill?

And maybe you could reproduce the same situation with other tools.

[jfb8856606]

Which version of FreeBSD you tested?

F-Stack use FreeBSD release 11.0, and will try to minimize or not change the behavior of the TCP/IP protocol stack, there may be some different in different versions.

[zouyonghao]

I test it on FreeBSD 12.1, but the read data contains error when I test my script with f-stack, and I think it may also do not happen on FreeBSD 11.0.
It may be a bug in f-stack, could you test this case?