Subject: Re: CVS commit: basesrc
To: None <>
From: der Mouse <mouse@Rodents.Montreal.QC.CA>
List: tech-userlevel
Date: 02/12/2000 18:31:21
>>>  newfs_lfs refuses to create a filesystem on a partition not
>>>  labeled "4.4LFS".
> Yuck.  How do I newfs a partition not labeled this way now?

Ummm...UTSL? :-(

>> Shouldn't we try to make the various newfs_* programs behave
>> similarily?
> NO, unless there is an option to turn this bogosity off.

I entirely agree.

> I have, more than once (actually, in three places in my current
> setup) used completely _unlabeled_ disks: that is, newfs'ed the raw
> partition and used that.

Wanting to newfs the raw partition isn't a reason to want to newfs a
partition whose label type is different unless your disklabels can't
live in the first 8K of the disk.

> There is real utility to this, particularly for the disk-swapping
> case, now that FFS_EI exists.

Right.  (Though does FFS_EI also give you other-endian LFS support?)

Furthermore, this sort of behavior (refusing to newfs_xxx a partition
not of type xxx) renders filesystem xxx completely broken for disks
with legacy disklabels that can't represent partitions of type xxx.
And sometimes, to keep boot ROMs happy, you have to use a legacy

newfs refusing, I could see that, a la fsck.  But newfs_lfs should make
an LFS regardless, and I would argue shouldn't even check that it's a
device; it, like the rest of the newfs_xxx programs, should be
perfectly happy to make a filesystem on anything that supports lseek
and write.

Of course, that's just my opinion....

					der Mouse

		     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B