Subject: Re: as long as we're hitting FFS...
To: None <>
From: Bill Studenmund <>
List: tech-kern
Date: 03/23/1999 19:33:15
On Tue, 23 Mar 1999, Eduardo E. Horvath wrote:

> On Tue, 23 Mar 1999, Bill Studenmund wrote:
> > One thing I'd like to make clear is that the current proposal was for
> > changes to sys/ufs/ffs, not for the creation of sys/ufs/liffs. As such, we
> I for one would feel much more comfortable if it were in sys/ufs/liffs.
> FFS is the single most important filesystem and breakage there is
> potentially catastrophic.  Also, most ports have support for FFS
> filesystems in the bootloaders and these changes would have to go there as
> well.

Then we have to keep two copies of the files options LIFFS has changed,
and make sure that changes in one go into the other.

That's a mess.

> Consider how long LFS has been broken without any really ill effects.
> This `LIFFS' is a different on-disk format that is incompatible with all
> existing FFS-derived filesystems including filesystems used by Ultrix, 
> Sunos, and SVR4 derivatives.  Because of that I assert that it is a
> different filesystem and should be separate.

But it was not implimented as a seperate filesystem. It was implimented as
ffs with larger inodes. The first 128 bytes of a large inode are treated
the same as the first 128 bytes of a small inode.

If you really don't like it, you don't have to define options LIFFS. :-)

Take care,