Subject: Re: NCR PCI SCSI bugs/fixes
To: None <port-i386@NetBSD.ORG>
From: Thor Lancelot Simon <tls@cloud9.net>
List: port-i386
Date: 01/21/1995 18:56:26
Having just fought a desperate and losing battle against a NCR controller in
an ASUS 486/66 machine (which involved my wasting far too much of mycroft's
time to boot) I suspect that the NCR driver -- among other random lossage --
has only been extensively tested on 90Mhz Pentium machines, and loses
progressively worse on slower machines as some internal timing constraint is
met less and less often.

Personally, I bought a BusLogic.  I am not happy about this state of affairs,
but while (modulo my near-absolute lack of familarity with PC hardware) I
don't object to meddling with and trying to debug straightforward driver code,
the NCR driver is baroque and foul to say the least.

Anyone want to pitch in for a "support Charles in luxury while he rewrites the
NCR driver" fund?  It seems to need doing. :-)

I have clients who are dependant upon the correct functioning of the NCR
driver, because they have wide SCSI disks.  Now that I've seen a bit more of
the guts of it, I'm starting to seriously regret suggesting that they buy NCR
controllers. 

Does the bt742 driver work with the BT7whatever that's wide scsi?

Thor