From owner-freebsd-bugs  Sun Jun 15 04:00:02 1997
Return-Path: <owner-bugs>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.5/8.8.5) id EAA05828
          for bugs-outgoing; Sun, 15 Jun 1997 04:00:02 -0700 (PDT)
Received: (from gnats@localhost)
          by hub.freebsd.org (8.8.5/8.8.5) id EAA05820;
          Sun, 15 Jun 1997 04:00:01 -0700 (PDT)
Date: Sun, 15 Jun 1997 04:00:01 -0700 (PDT)
Message-Id: <199706151100.EAA05820@hub.freebsd.org>
To: freebsd-bugs
Cc: 
From: Doug Rabson <dfr@nlsystems.com>
Subject: Re: kern/3838: fifos on nfs-mounted fs no longer permitted
Reply-To: Doug Rabson <dfr@nlsystems.com>
Sender: owner-bugs@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

The following reply was made to PR kern/3838; it has been noted by GNATS.

From: Doug Rabson <dfr@nlsystems.com>
To: Bruce Evans <bde@zeta.org.au>
Cc: admin@citylink.dinoex.sub.org, FreeBSD-gnats-submit@freebsd.org,
        dfr@freebsd.org
Subject: Re: kern/3838: fifos on nfs-mounted fs no longer permitted
Date: Sun, 15 Jun 1997 11:57:50 +0100 (BST)

 On Wed, 11 Jun 1997, Bruce Evans wrote:
 
 > >mkfifo on a nfs-mounted filesystem will reply: "operation not permitted".
 > >This didn't happen with release 2.1.0. (No, the fs isn't mounted with
 > >nodev!)
 > 
 > This was broken in the initial v3 import (rev.1.16), which is essentially
 > the same as Lite2.  There are several bugs:
 > 
 > 1. In the v2 case, nfsrv_create() lost some special handling for fifos.
 >    suser() is always called, although mkfifo doesn't require special
 >    privileges.
 > 2. In the v3 case, the new nfsrv_mknod() never had enough special handling
 >    for fifos.  suser() is always called, except for sockets.
 > 3. In the v3 case, after suser() fails, the error code is eventually
 >    thrown away.  nfsrv_mknod() returns 0, and mkfifo() returns the bogus
 >    errno EIO.
 > 
 > Bruce
 > 
 > Untested fixes for (1)-(2):
 > 
 > ! 			if (vap->va_type == VFIFO &&
 > ! 			    (error = suser(cred, (u_short *)0))) {
 
 I think this should be:
 
 			if (vap->va_type != VFIFO &&
 			    (error = suser(cred, (u_short *)0))) {
 
 Apart from that, your fixes seem to work fine.
 
 Oh.  I just noticed that your fixes have already been committed.  Sorry
 for not giving feedback earlier; I have been extremely busy on other stuff
 for the last few days.
 
 --
 Doug Rabson				Mail:  dfr@nlsystems.com
 Nonlinear Systems Ltd.			Phone: +44 181 951 1891
 					Fax:   +44 181 381 1039