AHA-2930 with Linux?

[Doug Ledford]

On 04-Dec-97 Daniel M. Eischen wrote:
>> This, again, is a termination issue.  I would try to find out what your
>> termination settings currently are for both the internal and external
>> devices to make sure it's correct.
>
>It may not be a termination issue.  I don't think the older drivers
>caught a SCSI bus reset (ENSCSIRST?) whereas the new driver does.
>These scanners seem to cause the controller to think the bus was
>reset (unless the scanners are actually resetting the bus!).

The SCSI controller itself signals that a reset has occured if the voltage
on the RST pin drops to low.  As it turns out, that voltage is driven high
by the termination.  When termination is bad (aka, the card disables
termination because of both an external and internal cable) and maybe an
internal device isn't terminated, you can run into problems.  Most scanners
use an external/passive terminator.  A single end of the bus, terminated
with a passive terminator, runs a decent risk of triggering false SCSI bus
resets simply because there isn't enough power to drive the RST pin high. 
So, I would expect this behaviour if something were wrong, such as the
internal device that is suppossed to supply termination isn't doing so (for
whatever reason, either not jumpered to do so or has a defective
terminator).  The same configuration, with the external cable removed, could
be very reliable as the card would then provide termination and the Adaptec
cards actually are half-way decent with only a single terminator (although
there is a small, but noticeable improvement in reliability and bus
operation if you do have both terminators installed).

>  A
>SCSI bus reset at device probe time gives the driver all sorts
>of problems.
>
>I don't know what causes the reset to occur, but take good look
>at the scanner, termination, and the cables.

----------------------------------
E-Mail: Doug Ledford <dledford at dialnet.net>
Date: 05-Dec-97
Time: 00:08:15
----------------------------------