Re: Strange behaviour for kernels > 6.12.x

On 2026-06-06, Greg Wooledge wrote:

> I keep thinking it's probably something in the networking code.  I'm
> reminded of TCP Window Scaling, which caused a bunch of problems when
> it was newly introduced.  Devices which were not properly configured
> sometimes could not communicate properly with Linux systems that had
> TCP Window Scaling enabled.  (This was not a bug in Linux.  It was a
> common bug or misconfiguration in other systems.)

Yes it could also be a security patch enforcing a new control which
fails. It could be apparmor denying access, etc. This is why I suggest to
look in the system logs. Looking in inn2 logs and dovecot logs could be
useful too.