In D20948#454139, @markj wrote:

We also have the (unused) so_linger_set(), which should possibly assert that the input value is valid.

Use KASSERT in so_linger_set() to ensure that the value is valid as suggested by markj@.

I tested this on a slow system with and without VIMAGE. It works now also on these platforms.

tcp_input_data needs some tweaks to work in kernels with options VIMAGE in its configuration file.

The already removed #include "opt_kern_tls.h" found its way back and breaks compilation.

D20834 fixes the issue for me.
@shurd: Could you test if D20834 fixes your issue?

You might want to address the nits. All of them are whitespace issue.

With removing these four header files, I can compile and use RACK on a head system.

I observed also a lockup of a bhyve VM on a slow system, where the VM used the config file:

tuexen@syzkaller:~ % cat head/sys/amd64/conf/SYZKALLER 
include         GENERIC
ident           SYZKALLER

In D19622#440454, @kbowling wrote:

I'll register strong desire to discard fragments immediately on interface vanishing. I don't expect that to come with any significant weight, my rational is whatever makes state management easier within current codebase and usage patterns of fbsd.

I tested this with syzkaller (if_tap) and packetdrill (if_tun) and it works as intended.