> Any time you get medium errors, that's a definite hint to switch to an sbc
> kernel.  Yeah.  It sounds like the infamous ncrscsi bug rearing its ugly
> head.  BTW, has anybody figured out what's wrong with that thing, or is
> that driver basically a lost cause?

The interrupt-driven PDMA mode on that driver does have some problems.
I don't think it's a lost cause, but it does need work (we _know_ that
the MacOS (esp. w/ HDT or similar) gets much better performance and
reliability).  No one has volunteered to do any work on it or on the
interrupt-driven part of the SBC driver (which suffers from the same
problems, I believe).  That would certainly be a place where folks could
contribute.  I'd be happy to help anyone who wants to work on it.
Me and my Fireballs are quite happy with the current driver, though.
No fs damage that I can recall and a current uptime of 55 days.  I am
exercising the network more than the disks, though:

Name  Mtu   Network     Address              Ipkts Ierrs    Opkts Oerrs  Coll
lo0   32768 <Link>                          387726     0   387726     0     0
lo0   32768 127         localhost           387726     0   387726     0     0
ae0   1500  <Link>      00:00:94:20:ce:32 139346535 15770 71214348  1191 1307792
ae0   1500  198.82.200  puma.macbsd.com   139346535 15770 71214348  1191 1307792
ae0   1500  198.82.200  acmerocket.com    139346535 15770 71214348  1191 1307792
ae1   1500  <Link>      00:00:94:01:43:d1 142626617   152 56627451     0 31302
ae1   1500  198.82.155. puma2             142626617   152 56627451     0 31302
ppp0* 1500  <Link>                               0     0        0     0     0
sl0*  296   <Link>                               0     0        0     0     0

-allen

-- 
              Allen Briggs - end killing - briggs@macbsd.com