From: "Kevin Coffman" <kwc@citi.umich.edu>
Subject: Re: Fusing Debian patches back into nfs-utils
Date: Thu, 1 Jun 2006 12:42:47 -0400
Message-ID: <4d569c330606010942u6900ca96w1795b5a624ea21ee@mail.gmail.com>
References: <20060601155929.GA890@uio.no>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Cc: nfs@lists.sourceforge.net
To: "Steinar H. Gunderson" <sesse@debian.org>
In-Reply-To: <20060601155929.GA890@uio.no>
Sender: nfs-bounces@lists.sourceforge.net
Errors-To: nfs-bounces@lists.sourceforge.net

On 6/1/06, Steinar H. Gunderson <sesse@debian.org> wrote:
> (Please Cc me on any followups, I'm not subscribed to the list)
>
> Hi,

>  - fix_anonuid_and_anongid.diff: The current code uses -2 for default
>    anonymous uid and gid, in the hope that it will be equivalent to 65534.
>    However, on systems with 32-bit uids and gids, this returns another value
>    entirely, which gives rise to an unpredictable default uid and gid (which
>    probably doesn't match what's in passwd and group). This patch fixes the
>    code and documentation to use 65534 explicitly instead.

There is a similar issue in the svcgssd code.  However, while I was
looking into this, it appeared that the kernel code uses "-2" in
places as well.  It seems like a similar change may be needed in the
kernel code as well?

The svcgssd code has the added problem that it really wants to use the
per-export anon values if they were specified differently than the
default.  Those probably need to be passed up in the upcall, or we
need a way of indicating that the default values should be used in the
downcall.

K.C.


_______________________________________________
NFS maillist  -  NFS@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nfs