Subject: Re: hfstools
To: David Bushong <dbushong@volga.EECS.Berkeley.EDU>
From: Bernard Gardner <B.Gardner@eng.usyd.edu.au>
Date: 08/09/1996 20:48:05
On Aug 6, 17:51, Joel Klecker wrote:
> Subject: Re: hfstools
> 8/5/96 at approximately 2:27 AM -0700, David Bushong wrote:
> > drive and partition w/ the hfs volume on it), I get "hmount: /dev/sd2e:
> > unable to obtain lock for device (Operation not supported)". If I try
> > "hmount /dev/sd2a" (the drive with the hfs volume, but just the first
> > partition), I get "hmount: /dev/sd2a: error opening device (Device busy)"
> Hmm, works fine for me. All I can think of is make sure you have the newest
> version from <ftp://ftp.ccs.neu.edu/pub/people/rob/hfs/>, and also check
> out the web page at <http://www.mars.org/~rob/proj/hfs/>.
I made hfs work on my machine by commenting out the "meat" of the locking
function and just letting it return 0 to indicate that there was no error,
but I don't recommend this as a solution. I won't be responsible if you
trash your disks by doing this.
The problem is actually in the latest version of hfs, the new version
introduces locking, using fcntl, and NetBSD doesn't seem to support
fcntl locks on devices. Does anyone know why this is?
There is code at http://marvin.eng.usyd.edu.au/~bernardg/ltest.c that
demonstrates the problem. It works correctly on a BSDI box that I just
tested it on, but on NetBSD (mac68k and pmax) it dies with ENOTSUPP.
I suspect it would work correctly on other machines, but I don't have
any around whose disks I can unmount to test this.