Subject: Re: UVM/other problems for desktop users in current?
To: None <current-users@netbsd.org>
From: Ross Patterson <Ross.Patterson@CatchFS.Com>
List: current-users
Date: 12/19/2002 12:12:25
On Wednesday 18 December 2002 02:31 am, Jukka Marin wrote:
> On Wed, Dec 18, 2002 at 03:49:40PM +1100, Daniel Carosone wrote:
> > For those experiencing this problem - are you running softdep?
>
> No.  Is it safe?  Does it help with NFS (server, client)?  I use it in
> /tmp, but I haven't been brave enough to enable it elsewhere.

I've been following this discussion with great interest, at least since the 
question of softdeps arose and caused me to go root through my trash folder 
for the rest of the thread. :-)  This question ("is softdeps safe?") seems to 
come back on one or another of the NetBSD lists every few months.  Based on 
previous observations about reliability etc., I've assumed that the core 
developers wouldn't have enabled "option SOFTDEP" in the GENERIC 
configuration in a stable release if it wasn't safe.  But as some high school 
teacher always said, "When you assume you make an ass out of u and me."  

I did a lot of digging in the archives, and it seems like several 
authoritative folks claim softdeps are just fine.  Kirk McKusick (of course 
*he'd* say so :-) ) wrote affirmatively back in June 2000 
(http://mail-index.netbsd.org/current-users/2000/06/17/0024.html) on a thread 
about fsync() and softdeps, as did Don Lewis 
(http://mail-index.netbsd.org/current-users/2000/06/19/0011.html, 
http://mail-index.netbsd.org/current-users/2000/06/17/0003.html).

On the other hand, the NetBSD doc web pages 
(http://www.netbsd.org/Documentation/misc/ and 
http://www.netbsd.org/Documentation/tune/5.html#a3) indicate that softdeps 
are not enabled by default in the kernel (not true since June 2000 for i386 
at least) due to some licensing problem (which McKusick's web page says 
hasn't been the case since June 2001), and that they're considered 
experimental (and by implication unsafe).  The mount(8) man page agrees with 
that softdeps are experimental.

To sum up, has it been definitively established that softdeps are safe?  Or 
the converse?  I'm not qualified to answer, but I'd really love to know.

Thanks,
-- 
Ross A. Patterson
CatchFIRE Systems, Inc.
5885 Trinity Parkway, Suite 220
Centreville, VA  20120
(703) 563-4164