On Wed, Oct 16, 2002 at 01:50:43PM -0800, Ethan Benson wrote:
> On Wed, Oct 16, 2002 at 11:26:52PM +0200, Andi Kleen wrote:
> > On Wed, Oct 16, 2002 at 12:38:53PM -0800, Ethan Benson wrote:
> > > On Wed, Oct 16, 2002 at 05:50:47AM -0400, TJ Easter wrote:
> > > > Anyone know if there are plans to impliment the immutable/append-only
> > > > options?
> > >
> > > a couple monthes ago Steve said it looked like it would be pretty easy
> > > to do, and without breaking backward compatibilty (since older
> > > implementations would simply ignore the bits). he just doesn't really
> > > have time, someone with a little experience in the XFS code could
> > > probably do it though.
> > >
> > > immutable, append-only, and the S (sync) bits would be the useful ones.
> >
> > 'D' for directories is quite useful too, because some mail servers
> > assume that everybody has synchronous rename() like traditional BSD FFS.
>
> your thinking S which can be applied to both files and directories.
>
> there is no chattr +D at least in my version of it.
Then your version is outdated:
% chattr
usage: chattr [-RV] [-+=AacDdijsSu] [-v version] files...
>
> > 'd' (don't dump) should be also simple and useful.
>
> xfs already honors an extended attribute for files only, the
> maintainers believe that honoring such a thing for directories is evil.
If 'the maintainers' really said such a thing then I disagree with
the maintainers. Honoring it for directories should be fine and I can
think of applications for it.
>
> > and of course 'u' if it was implemented (probably a bit more work)
> > would clear a major FAQ item on this list.
>
> not even ext2 supports u.
And? Is XFS only limited to things ext2 supports?
-Andi
|