Thanks for the feedback...
there is another (performance) problem in my code, actually:
with TCP, or non-datagram sockets, the socket buffer usually
contains a single record, so that sbappend would still have to
scan the chain m->m_next . An additional tail pointer would
be needed to avoid that o
In article [EMAIL PROTECTED]> you
write:
>
>I would be grateful if someone could test the attached patch which
>deals with the following problem:
>
> on all *BSD version, socket buffers contain a list of
> incoming and/or outgoing mbufs. Unfortunately the list only
> has a point
