> > NEW_BUFQ_STRATEGY seems to radically improve performance on my laptop,
> > but I've been cautioned by a couple of people that it is unstable and
> > has been known to ruin disks. Is anyone working on fixing it?
>
> I'm running NEW_BUFQ_STRATEGY on three machines:
>
>   - Sparc (SPARCstation-20L) with one SCSI disk
>   - I386 (AMD Athlon 1) with two SCSI disks, part as CCD
>   - I386 AMD Athlon 4) with one SCSI disk, one (unused) IDE disk
>
> I didn't see any problems so far.
>
> There was a thread on current-users starting with a mail from
> Urban Boquist <urban@boquist.net> on Thu, 31 Oct 2002 but to
> my knowledge these problems have gone.
>
> The QUEUEDEBUG problem is fixed (thanks enami).
>
> Bang Jun-Young <junyoung@mogua.com> had a problem with softdep
> enabled (pending blocks/files < 0). As I have not heard
> anything I suppose this problem is also fixed.
> --

just to be sure - does that code fixes problem with huge stalls under
large disk I/O present in 1.6?

if so - do i need to download tarballs from NetBSD-current to get version
with that, or i need to use cvs?