> >Now that I've solved it the hard way, is there a list of what drives give
> >ncrscsi difficulties?

[...]

> I keep telling people, there is more to it than just the drives, I have had
> a Quantum Trailblazer 840 going under NetBSD 1.1 for 11 months or so with
> no drive problems what so ever, and now I am running a custom 1.2 (CHRIS)
> kernel that is using the ncrscsi driver, and still no problems.  We need to
> look at this much more closely.
> 
> Christopher R. Bowman

That's how I see it, too. ncrscsi kernels from before mid-May have never
given me much disk related trouble besides the usual unionfs lockups. I
have definitely not seen any _hard _ disk trouble (Quantum LPS 105S,
Empire 1080, TRB 850); the MO drive is a different tale. Heavy trouble
with ncrscsi began after that date (interrupt related code changes, if I
recall correctly). Most of the executables started from /etc/rc dumped
core; root is on the LPS 105, so I can't even say anything about the
Snailblazer: I didn't get that far. 

I am still playing with the idea of asking Allen for the pre-May ncrscsi
sources for my custom kernels. As it is, the only kernel that I can sup
with is ncrscsi, but has a broken serial console (sbc kernels eat my MOs
when writing to them). :(

        hauke  

-- 
"It's never straight up and down"  (DEVO)