Sam <[EMAIL PROTECTED]> wrote:
> On Thu, 10 Feb 2000, Len Budney wrote:
> >
> > Which brings us back to link() or rename()...
>
> And, if close() does not update the metadata, there's no reason why link
> or rename should either.
>
> What this REALLY brings us back to is the fact that the only thing that
> journaling guarantees you is that you won't have to refsck everything
> after a reboot...
Journalling is absolutely orthogonal to the reliability issue. The
reliability issue is: What are the semantics of fsync(), link() and
rename()? If they return after the requested operation completes to disk,
we can guarantee reliability. If not, we can't.
Which brings us back to your mistake; you make a claim about ``journalling
filesystems'' which is true for some, and false for others. FFS had
those semantics, for example, but happened to break them when soft
updates were added.
Len.
--
256-bit keys will forever be immune from brute-force attacks until
computers are made up of something other than matter and occupy something
other than space.
-- Bruce Schneier
