Content-class: urn:content-classes:message
MIME-Version: 1.0
Content-Type: text/plain; 	charset="UTF-8"
Subject: RE: "tar --one-file-system" accesses remote file systems
Date: Wed, 13 Feb 2008 12:03:58 -0800
Message-ID: <70952A932255A2489522275A628B97C30620E123@xmb-sjc-233.amer.cisco.com>
In-Reply-To: <47B3413C.4040007@cygwin.com>
References: <47B3413C.4040007@cygwin.com>
From: "Matt Seitz (matseitz)" <matseitz@cisco.com>
To: <cygwin@cygwin.com>
Mailing-List: contact cygwin-help@cygwin.com; run by ezmlm
Precedence: bulk
Sender: cygwin-owner@cygwin.com
Mail-Followup-To: cygwin@cygwin.com
Content-Transfer-Encoding: 8bit

> From: Larry Hall (Cygwin) 
> 
> Matt Seitz (matseitz) wrote:
> > 
> >> Matt Seitz (matseitz) wrote:
> >>> This problem and a proposed solution was mentioned in an earlier 
> >>> e'mail 
> (http://www.cygwin.com/ml/cygwin/1998-09/msg00839.html). 
> > 
> > Ah, yes, the mounted CIFS share is reported as a FAT file 
> > system*. 
> 
> That's it I expect.  Going straight to the code, in 
> fhandler_disk_file.cc, here's some code from 
> fhandler_base::fstat_helper():
> 
>    /* Enforce namehash as inode number on untrusted file systems. */
>    if (pc.isgood_inode (nFileIndex))
>      buf->st_ino = (__ino64_t) nFileIndex;
>    else
>      buf->st_ino = get_namehash ();
> 
> One of the things that isgood_inode() checks for is that it's 
> not a FAT drive.  In case it is, you end up with a faked hash inode.

Thanks for the diagnosis.  I'm curious about something.  The message I reference above also mentioned an issue with "st_dev".  It seems to imply that correcting the "st_dev" to use the volume serial number could resolve this issue.  What is your opinion on that theory? 


--
Matt Seitz
Manager, File System Virtualization
Cisco Systems, Inc.
.:|:.:|:.