--
Erik B. Andersen Web: http://www.inconnect.com/~andersen/
email: andersee@debian.org
--This message was written using 73% post-consumer electrons--
On Tue, 22 Jul 1997, Rogier Wolff wrote:
> Colin Plumb wrote:
> >
> > Looking at reiserfs, I found a University of Michigan tech report
> > on "soft updates" a technique to remove all synchronous writes
> > from a file system yet maintain fsck-less crashes.
> >
> > (By always marking data in use before it's used and ensuring that
> > data is not used before it's marked free, a crash can leave some
>
> If you first mark a block as used, and only then write the block,
> the filesystem won't be corrupt when the system crashes inbetween.
> However, your DATA will be corrupt.
>
> I think it's better to have fsck detect a broken filesystem, than to
> have a file silently contain bad data. As far as I know, Linus also
> thinks this way.
>
> You also eliminate the possibility of reordering writes at the
> driver or drive level. (e.g. "tagged queuing" is forbidden...)
>
> All this would be very hard to debug: You'd have to try throwing
> the switch on your system quite a few times to be sure it works.
>
> Roger.
>
>
>