From owner-freebsd-hackers  Thu Oct 29 08:59:05 1998
Return-Path: <owner-freebsd-hackers@FreeBSD.ORG>
Received: (from majordom@localhost)
          by hub.freebsd.org (8.8.8/8.8.8) id IAA16407
          for freebsd-hackers-outgoing; Thu, 29 Oct 1998 08:59:05 -0800 (PST)
          (envelope-from owner-freebsd-hackers@FreeBSD.ORG)
Received: from godzilla.zeta.org.au (godzilla.zeta.org.au [203.15.68.22])
          by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id IAA16401;
          Thu, 29 Oct 1998 08:59:02 -0800 (PST)
          (envelope-from bde@godzilla.zeta.org.au)
Received: (from bde@localhost)
	by godzilla.zeta.org.au (8.8.7/8.8.7) id DAA22379;
	Fri, 30 Oct 1998 03:58:58 +1100
Date: Fri, 30 Oct 1998 03:58:58 +1100
From: Bruce Evans <bde@zeta.org.au>
Message-Id: <199810291658.DAA22379@godzilla.zeta.org.au>
To: committers@FreeBSD.ORG, finrod@ewox.org, hackers@FreeBSD.ORG
Subject: Re: sysctl
Sender: owner-freebsd-hackers@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.ORG

>While fixing a sysctl-related bug in arp, I noticed that sysctl(3) is
>kinda fubar. According to the man page, it either returns 0, or
>returns -1 and sets errno. However, reading src/lib/libc/gen/sysctl.c
>and src/sys/kern/kern_sysctl.c reveals that the sysctl syscall returns
>the error code directly, and the sysctl library function does a little
>of both. The "fix" I did to src/lib/libkvm/kvm_proc.c depends on

I think you need to read the code again.  All syscall functions in
in the kernel (at the level of sysent[]) return the error code directly.  
Normal return values are returned indirectly in p->p_retval[].

>sysctl(3) acting in accordance to the man page, which it doesn't (it
>returns ENOMEM instead of setting errno to ENOMEM and returning -1).
>
>Should sysctl(3) be fixed to act as documented, or should the man page

It already does, at least for the call near your fix in kvm_proc.c.

Bruce

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message