more AIC problems

Doug Ledford dledford at dialnet.net
Wed Mar 11 09:59:36 PST 1998


Derek Seaman wrote:
> 
> I have some more information on the problems with dual 2940A v1.30 cards.
> Even with the BIOS disabled on an empty card, but card termination
> disabled, the system refuses to boot. I get a spurious interrupt error. If
> I enable card termination the system is OK.

This is exactly what happens when there is no termination what so ever on
the SCSI bus.  It's not really a driver problem, that is a problem with the
SCSI bus in general.  There are several lines that need to be pulled up to
at least a +2V reference signal for the bus to work, and the termination is
responsible for providing that pull up.  If there is no termination, that
pull up doesn't get done, and the reset pin for one is *always* active.

> 
> Also, the problem with data over runs happens even with the 2nd card's BIOS
> disabled. I had previously thought disabling the BIOS corrected the
> problem. But this morning the data over runs started happening again.
> Moving all the devices to one card _always_ corrects the problems.

Data overruns are usually a sporadic problem specific to certain devices.  I
would suspect that if you had all of the devices on one card long enough, it
would generate overruns as well because frmo the sounds of it, you have one
of the hard drives that we see this problem with.

> So it looks like there might be a problem handling two SCSI cards in one
> system. Maybe because of the 1.30 BIOS?

Doubtful, the driver has been stable and running with multiple controllers
in one system for a *long* time (including the ftp server where people get
my patches is one such system).  The problem is more likely related to the
1.30 BIOS specifically on the 2940AU cards you have.

> Any help would be GREATLY appreciated, because using one SCSI card in my
> system is not a viable option for a production quality system.

Well, to be quite honest, the 2940AU card isn't really a viable option for a
heavily loaded production system.  The overhead cause by the extra SCB
paging on the PCI bus could become very significant when you consider that
card only has three hardware SCBs compared to 16 on the 2940 series and 32
on the 3940UW series (newer models with the 7895 chipset).

-- 

 Doug Ledford  <dledford at dialnet.net>
  Opinions expressed are my own, but
     they should be everybody's.

To Unsubscribe: send mail to majordomo at FreeBSD.org
with "unsubscribe aic7xxx" in the body of the message



More information about the aic7xxx mailing list