From owner-cvs-sys  Tue Dec  5 14:12:59 1995
Return-Path: owner-cvs-sys
Received: (from root@localhost)
          by freefall.freebsd.org (8.6.12/8.6.6) id OAA10249
          for cvs-sys-outgoing; Tue, 5 Dec 1995 14:12:59 -0800
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.freebsd.org (8.6.12/8.6.6) with ESMTP id OAA10222
          ; Tue, 5 Dec 1995 14:12:31 -0800
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.9/8.6.9) id JAA17085; Wed, 6 Dec 1995 09:07:26 +1100
Date: Wed, 6 Dec 1995 09:07:26 +1100
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199512052207.JAA17085@godzilla.zeta.org.au>
To: bde@freefall.freebsd.org, davidg@Root.COM
Subject: Re: cvs commit: src/sys/netinet in_pcb.h
Cc: CVS-committers@freefall.freebsd.org, cvs-sys@freefall.freebsd.org
Sender: owner-cvs-sys@FreeBSD.ORG
Precedence: bulk

>>  Modified:    sys/netinet  in_pcb.h
>>  Log:
>>  Added explicit include of <sys/queue.h>.  Currently, some things only
>>  compile because <vm/vm.h> happens to be gratuitously included before
>>  <netinet/in_pcb.h> and <vm/vm.h> happens to include <sys/queue.h>.

>   I'm happy to see that you're untangling this mess...it's been sorely needed
>for a long time.

I don't plan to do anything about the deeper nesting soon.  I've
eliminated about 66 #includes of <sys/user.h> (mostly in isa/sound and
isa/pcvt which have a general header that #includes all the OS-dependent
headers so that everything in the driver(s) depends on most of the OS
headers :-(), and about 67 #includes of <sys/vm.h>.

user.h isn't used in drivers any more.  About 20 drivers still need vm.h
just to get the declaration for vtophys().  This is implemented as a
macro in <machine/pmap.h> so it won't be easy to fix.

Bruce