[3339] in linux-scsi channel archive
Re: scsi : aborting command due to timeout
daemon@ATHENA.MIT.EDU (Umesh Soni)
Sun Feb 15 17:55:02 1998
To: "Daniel M. Eischen" <deischen@iworks.InterWorks.org>
cc: U.Soni@cs.ucl.ac.uk, linux-scsi@vger.rutgers.edu
In-reply-to: Your message of "Thu, 12 Feb 1998 22:45:46 CST." <199802130445.WAA04131@iworks.InterWorks.org>
Date: Sun, 15 Feb 1998 22:50:25 +0000
From: Umesh Soni <U.Soni@cs.ucl.ac.uk>
>
>> The problem with the linux aic7xxx driver is that regardless of what I
>> do with the motherboard (high-byte) jumper, the driver tells me that
>> high-byte termination is set to exactly the same state as the low-byte
>> termination.
>>
>> So, the only settings that I can arrive at are;
>> Low-byte(on), High-byte(on) Both on.
>>
>> or, Low-byte(off), High-byte(off). Both off.
>
>Unlesss a BIOS upgrade will let you use auto-termination or give
>you access to high byte termination, there is nothing you can do
>other than hardcode the driver to force the proper termination
>setting. This sux :-( Probably another config option is in
>order here (another unhappy face :( ).
>
Thinking about this (and taking clues from the source -luke :-) ), it
might be easier and safer to add a runtime option so that in the
infrequent case where termination needs to be forced, it can be
asserted on the command line -or via lilo.
For example using the following;
high_byte:[0,1] # Force high byte termination; 0 off, 1 on
low_byte:[0,1] # Force low byte termination; 0 off, 1 on
We might say something like (in /etc/lilo.conf);
append="aic7xxx=low_byte:0,high_byte:1,verbose,extended"
Umesh
