From owner-freebsd-commit  Wed Sep  6 05:19:15 1995
Return-Path: commit-owner
Received: (from majordom@localhost)
          by freefall.freebsd.org (8.6.11/8.6.6) id FAA21490
          for freebsd-commit-outgoing; Wed, 6 Sep 1995 05:19:15 -0700
Received: (from majordom@localhost)
          by freefall.freebsd.org (8.6.11/8.6.6) id FAA21462
          for cvs-all-outgoing; Wed, 6 Sep 1995 05:19:12 -0700
Received: (from majordom@localhost)
          by freefall.freebsd.org (8.6.11/8.6.6) id FAA21449
          for cvs-sys-outgoing; Wed, 6 Sep 1995 05:19:10 -0700
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.34])
          by freefall.freebsd.org (8.6.11/8.6.6) with ESMTP id FAA21328
          ; Wed, 6 Sep 1995 05:18:45 -0700
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.9/8.6.9) id WAA20880; Wed, 6 Sep 1995 22:17:09 +1000
Date: Wed, 6 Sep 1995 22:17:09 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199509061217.WAA20880@godzilla.zeta.org.au>
To: bde@zeta.org.au, wollman@lcs.mit.edu
Subject: Re: cvs commit: src/sys/vm vm_map.h vm_map.c
Cc: CVS-commiters@freefall.FreeBSD.org, bde@freefall.FreeBSD.org,
        cvs-sys@freefall.FreeBSD.org
Sender: commit-owner@FreeBSD.org
Precedence: bulk

>> 2) for protoswitch functions.  The number and type of args sometimes
>>    depends on the protocol.

>For pr_input routines, it always depends on the protocol.  The
>solution that I prefer for these is to declare pr_input as a `void
>(*)(void)' and then define some protocol-specific macros to cast it to
>the correct value.  If there were a way I could think of to use unions
>the same way for increased type-safety, I'd prefer that.

I'm still thinking about this.  I've found so many gratuitous function
type mismatches in places where prototypes were supposedly being used
that I disklike casts even more than before.

What about the pr_output routines?  pr_output seems to be called only
once, in raw_usrreq.c, and one callee, ip_output, seems to be completely
unprepared for the call.  This is detected as a bogus initialization in
inetsw in in_proto.c.

Bruce