On Fri, 17 Jan 2003, John Klos wrote:

> > I've had softdeps panics on 1.5.1/vax when copying larger (for a vax,
> > anyways) data sets from one disk to another (a slightly >4GB disk was
> > involved, I don't know if that is really an issue); I think I mentioned
> > that in the mailing list then (sorry, no pointers) but I generally
> > didn't consider softdeps too stable with that release so I didn't bother
> > writing a PR, especially since I've seen reports from others with
> > corruption and panics on 1.5.x/vax and non-vax with softdeps enabled at
> > that time.  So, I'd say, the bug(s) are there at least since 1.5.1.
>
> That, I am guessing, is a different problem. Softdeps yes/no was asked
> about corruption of > 4 gig slices in 1.6; it happened the same no matter.
> Softdeps are fixed in 1.6, apparently, as many people get heavy use out of
> systems with softdeps without corruption.
>
> The issue is whether 1.5.2 or, perhaps, 1.5.3 even, can handle > 4 gig
> slices without eventual corruption.

Since 1.5 and 1.5.1 worked for you, I'd expect that 1.5.2, and 1.5.3
will work as well.  That is, unless the offending change was backported
into the 1.5 release branch.  Given the flow of NetBSD releases, the
change would've been introduced during the development releases that led
up to the 1.6 release.

See:

   |  --bugfixes--->
  1.5--1.5.1--1.5.2--1.5.3->
   |
  1.5A   \
   |      \
  1.5B     } evil change likely happened somewhere in here.
   |      /
  ...    /
   |
  1.6--[eventually 1.6.1]->
   |
  1.6A
   |
   v


-brian.