[1702] in linux-scsi channel archive
Adaptec 2940UW and aic7xxx, help needed
daemon@ATHENA.MIT.EDU (Sonja Hammel)
Sat Apr 12 08:54:55 1997
Date: Sat, 12 Apr 1997 14:51:05 +0200
From: Sonja Hammel <sh@wsh.mk.eunet.de>
To: linux-scsi@vger.rutgers.edu
CC: sh@wsh.mk.eunet.de
Hello,
my appologies if this has been discussed before. I'm new to this
mailing list.
I urgently need some expert help on aic7xxx and Adaptec's 2940UW.
Here comes my configuration:
/proc/version:
Linux version 2.0.29 (root@uli) (gcc version 2.7.2.1) #12 Sun Apr 6
17:32:25 CEST 1997
/proc/scsi/aic7xxx/0:
Adaptec AIC7xxx driver version: 4.0/3.2/4.0
Compile Options:
AIC7XXX_RESET_DELAY : 15
AIC7XXX_TWIN_SUPPORT : Enabled
AIC7XXX_TAGGED_QUEUEING: Disabled
AIC7XXX_PAGE_ENABLE : Disabled
AIC7XXX_PROC_STATS : Enabled
Adapter Configuration:
SCSI Adapter: AHA-2940 Ultra
(AIC-788x chipset)
Host Bus: Wide
Base IO: 0xe400
IRQ: 11
SCBs: Used 4, HW 16, Page 16
Interrupts: 4696
Serial EEPROM: True
Extended Translation: Enabled
SCSI Bus Reset: Enabled
Ultra SCSI: Disabled
Target Disconnect: Enabled
I have two 4 GB IBM DFRS drives connected to the bus. The last drive
has active termination enabled. It is connected to the last plug on the
chain.
From time to time the system locks up completely. Sometimes I see a
load of 10 and more. I always see a driver message of the kind
Kernel: SCSIO: Target 1 underflow - Wanted at least 53248,
got 49152, residual SG count 9.
One time I was able to track the error down to a corrupted file "xx". A
simple
"cp xx yy" always resulted in a lockup. This was probably due to some
bad blocks on the disk. Thus, I performed a low level format and
verified the disk several times. Every second or third verify still
delivers
one or two additional bad blocks.
I could not reproduce a similar effect with DOS.
These are the facts. I'm not familiar with SCSI. So here are my stupid
questions:
1. Is it possible that the system is that much sensitive to
bad blocks?
2. Is it more likely that there are still hardware problems
(cabling, termination) or do you know of any bugs in the
scsi-driver (aic7xxx) or Adaptec's scsi-bios.
3. There was some romour that the 2940 doesn't like IBM's DFRS.
Can you comment on this?
4. Do you have any suggestions how to proceed in testing?
Any help ist appreciated.
Thanks, Uli
