>On Fri, 30 Apr 1999, Michael G. Schabert wrote:
>
>> If taking care with cables & terminations fixes instability, then it's not
>> the driver that's buggy.
>
>	I beg to differ on that one. That is akinb to saying that if
>	you consume all available VM and the machine hangs, its not the
>	VM system that is buggy.
>
>	The hardware may be doing something 'incorrect', but its not
>	entirely unexpected, and the driver _should_ be able to handle it.

My point was not to say that the driver isn't incidentally buggy, but if
there are termination or cabling problems, then it's not the bugginess of
the driver that's the problem. If fixing the term & cabling fixes the
errors, then there's no way of even knowing that the original problem WAS
the error that was reported (scsi protocol errors in manuel's case), since
it was likely spurious emissions & could've been anything, & didn't have to
be something "not entirely unexpected".

I've heard enough complaints of ncr being buggy that I certainly don't
doubt that it is...I just wouldn't have blamed the driver for the problem
that manuel was complaining about. There's not a driver on the planet that
can't be broken with bad-enough cables & term...that's one of the fun
things with SCSI voodoo :-).


Mike
Bikers don't *DO* taglines.