From owner-freebsd-hackers  Tue Aug  8 19:45:14 1995
Return-Path: hackers-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.11/8.6.6) id TAA05032
          for hackers-outgoing; Tue, 8 Aug 1995 19:45:14 -0700
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.2.228.19])
          by freefall.cdrom.com (8.6.11/8.6.6) with ESMTP id TAA04967
          ; Tue, 8 Aug 1995 19:44:36 -0700
Received: (from bde@localhost) by godzilla.zeta.org.au (8.6.9/8.6.9) id MAA01237; Wed, 9 Aug 1995 12:39:43 +1000
Date: Wed, 9 Aug 1995 12:39:43 +1000
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199508090239.MAA01237@godzilla.zeta.org.au>
To: davidg@Root.COM, gclarkii@freefall.cdrom.com
Subject: Re: FoxPro II.1
Cc: freebsd-hackers@freefall.cdrom.com
Sender: hackers-owner@FreeBSD.org
Precedence: bulk

>>How does the following look?
>>
>>CALL open(0x40524f,0,0)
>>NAMI "/dev/null":
>>RET  open 4
>>CALL fcntl(0x4,0x6,0xefbfbf34)
>>RET  fcntl -1 errno 99 Unknown error: 99
>>CALL close (0x4)
>>RET  close 0
>>
>>Then it goes on to load error messages and then it exits.

>   The above fcntl() is highly bogus. I suspect that the main problem is that
>the process group/PID (third argument to fcntl) wasn't specified, or a pointer
>to it was specified rather than the actual value (i.e. it wasn't dereferenced).
>...but why you would want to do an F_SETOWN on /dev/null is also a little
>strange.

I think 99 is a translation of errno 45 (EOPNOTSUPP).  See /sys/i386/ibcs2/
ibcs2_sysent.c#bsd_to_sysv_errno[NERR] (this table is apparently out of
date - NERR is 80 but FreeBSD-current has 82 errnos).

This is consistent with a valid process group/PID.  F_SETOWN on /dev/null
is just not supported.  F_SETOWN is only supported for sockets and ttys and
doesn't work right for ttys (it clashes with POSIX controlling terminal
semantics).

Bruce