Subject: Re: LFS frailty vs. datestamping [Was Re: /dev/clock pseudodevice]
To: Andrew Gillham <>
From: Greywolf <>
List: tech-kern
Date: 07/30/2001 12:07:06
On Mon, 30 Jul 2001, Andrew Gillham wrote:

# FWIW, I somewhat agree with Greywolf that LFS is fragile.  Perhaps not for
# the same reasons though.
# My reasons:
#     1. Filling up the partition (attempting to anyway) causes bad
#        things to happen.  Either kernel panics, or just going catatonic.
#        (a tar process can sit for hours without getting an error returned)
#     2. The current fsck_lfs can't fix anything, requiring me to newfs_lfs.
#     3. Not UBC'ified yet.  This means things that now expect mmap() to be
#        coherent (which it is on FFS) don't work when moved to LFS.
# Sort of unrelated:
#     4. I have occasionally typed "newfs" rather than "newfs_lfs" and newfs
#        happily changes the partition type to 4.2BSD and creates an FFS
#        file system.  This seems wrong, :)

You know, dammit, this is the second instance in quite a few years regarding
newfs checking the filesystem type!

I submitted a change for this I think in 1993 sometime, and it was
immediately regarded as "an egregious hack" and not committed.  Check
bin/4116 [it got closed] for the patch in question.

My coding style in that case might not have been the best, but honestly,
I didn't see a better way to handle it at the time.

I was *pissed* when it couldn't at least print a warning that I should
use a different filesystem type for newfs (and then it could allow me
to override it if I *really* wanted to) and it clobbered my MS-DOS partition
[that was unfortunately also my dual-boot to Windows at the time, which,
while it was not my choice of OS, was required for me to function in
that environment].

But I digress...

# I would still consider LFS experimental, and as such, some issues are to be
# expected.  It needs to be clearly marked as experimental also.

I've already spoken my piece on LFS for a while, and have subsequently
received several clue*4s (which I requested, and my thanks were in
earnest, BTW), so if it works for you guys, great.  If I can come up
with something a bit more palatable (assuming that I can code it up
correctly in the first place), I'll let y'all know.

# -Andrew

NetBSD: Twice the Bits-Clean of other Leading OSes.