From owner-freebsd-scsi  Fri Jun  6 16:38:00 1997
Return-Path: <owner-freebsd-scsi>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.5/8.8.5) id QAA19143
          for freebsd-scsi-outgoing; Fri, 6 Jun 1997 16:38:00 -0700 (PDT)
Received: from vader.cs.berkeley.edu (vader.CS.Berkeley.EDU [128.32.38.234])
          by hub.freebsd.org (8.8.5/8.8.5) with ESMTP id QAA19132
          for <scsi@freebsd.org>; Fri, 6 Jun 1997 16:37:55 -0700 (PDT)
Received: (from asami@localhost) by vader.cs.berkeley.edu (8.8.5/8.7.3) id QAA19597; Fri, 6 Jun 1997 16:37:54 -0700 (PDT)
Date: Fri, 6 Jun 1997 16:37:54 -0700 (PDT)
Message-Id: <199706062337.QAA19597@vader.cs.berkeley.edu>
To: scsi@freebsd.org
Subject: 3940uw with long cables
From: asami@cs.berkeley.edu (Satoshi Asami)
Sender: owner-freebsd-scsi@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

Howdy,

I've got a machine that has a bunch of disks -- unfortunately, some of 
the cables are quite long due to physical constraints (enclosures
stacked on top of each other, etc.)  I see the following kind of
errors when the probe reaches one of the longer cables:

===
Probing for devices on PCI bus 3:
ahc4 <Adaptec 3940 Ultra SCSI host adapter> rev 0 int a irq 11 on pci3:4
ahc4: aic7880 Wide Channel A, SCSI Id=6, 16 SCBs
ahc4 waiting for scsi devices to settle
ahc4: target 0 Tagged Queuing Device
(ahc4:0:0): "IBM OEM DCHS09Y 2424" type 0 fixed SCSI 2
sd28(ahc4:0:0): Direct-Access 8689MB (17796077 512 byte sectors)
ahc4: target 1 Tagged Queuing Device
(ahc4:1:0): "IBM OEM DCHS09Y 2424" type 0 fixed SCSI 2
sd29(ahc4:1:0): Direct-Access 8689MB (17796077 512 byte sectors)
ahc4: target 2 Tagged Queuing Device
(ahc4:2:0): "IBM OEM DCHS09Y 2424" type 0 fixed SCSI 2
sd30(ahc4:2:0): Direct-Access 8689MB (17796077 512 byte sectors)
ahc4: target 3 Tagged Queuing Device
(ahc4:3:0): "IBM OEM DCHS09Y 2424" type 0 fixed SCSI 2
sd31(ahc4:3:0): Direct-Access 8689MB (17796077 512 byte sectors)
ahc4: target 4 Tagged Queuing Device
(ahc4:4:0): "IBM OEM DCHS09Y 2424" type 0 fixed SCSI 2
sd32(ahc4:4:0): Direct-Access 8689MB (17796077 512 byte sectors)
ahc4: target 5 Tagged Queuing Device
(ahc4:5:0): "IBM OEM DCHS09Y 2424" type 0 fixed SCSI 2
sd33(ahc4:5:0): Direct-Access 8689MB (17796077 512 byte sectors)
ahc4: target 8 Tagged Queuing Device
(ahc4:8:0): "IBM OEM DCHS09Y 2424" type 0 fixed SCSI 2
sd34(ahc4:8:0): Direct-Access 8689MB (17796077 512 byte sectors)
ahc4: board is not responding
(ahc4:9:0): SCB 0x0 - timed out while idle, LASTPHASE == 0x1, SCSISIGI == 0xb4
SEQADDR = 0x42 SCSISEQ = 0x12 SSTAT0 = 0x5 SSTAT1 = 0x2
(ahc4:9:0): SCB 0: Immediate reset.  Flags = 0x801
(ahc4:9:0): no longer in timeout
ahc4: Issued Channel A Bus Reset. 1 SCBs aborted
ahc4: board is not responding
(ahc4:9:0): SCB 0x0 - timed out while idle, LASTPHASE == 0x1, SCSISIGI == 0xb4
SEQADDR = 0x42 SCSISEQ = 0x12 SSTAT0 = 0x5 SSTAT1 = 0x2
(ahc4:9:0): SCB 0: Immediate reset.  Flags = 0x801
(ahc4:9:0): no longer in timeout
ahc4: Issued Channel A Bus Reset. 1 SCBs aborted
===

Is this something that can be only cured by somehow shortening the
cables?  This is 2.2-stable, IBM U-W disks on Adaptec 3940UW (with
20MHz mode disabled).  We have 14-disk enclosures stacked on top of
each other.  The cables inside the enclosures are about 10 feet total
in my eyeball test.  The external cables are either 2, 2 1/2, 3 or 3
1/2 feet long.  It's only the 3 or 3 1/2 feet cables that give us
problems.  (Which is also very interesting, does 12 1/2 and 13 make
*that* much difference?!?)

Satoshi