From nobody  Tue Jan 26 20:15:59 1999
Received: (from nobody@localhost)
          by hub.freebsd.org (8.8.8/8.8.8) id UAA21696;
          Tue, 26 Jan 1999 20:15:59 -0800 (PST)
          (envelope-from nobody)
Message-Id: <199901270415.UAA21696@hub.freebsd.org>
Date: Tue, 26 Jan 1999 20:15:59 -0800 (PST)
From: satz@sprynet.com
To: freebsd-gnats-submit@freebsd.org
Subject: SCSI timeout sometimes crashes 3.0-current
X-Send-Pr-Version: www-1.0

>Number:         9718
>Category:       kern
>Synopsis:       SCSI timeout sometimes crashes 3.0-current
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    gibbs
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:  
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Tue Jan 26 20:20:01 PST 1999
>Closed-Date:    Thu Feb 18 07:57:35 PST 1999
>Last-Modified:  Thu Feb 18 07:58:42 PST 1999
>Originator:     Greg Satz
>Release:        3.0-current
>Organization:
>Environment:
FreeBSD fnu.cisco.com 3.0-STABLE FreeBSD 3.0-STABLE #11: Sat Jan 23 22:01:28 MST 1999     satz@fnu.cisco.com:/usr/src/sys/compile/FNU  i386

>Description:
> (da1:ahb0:0:1:0): ECB 0xf40b3600 - timed out
> (da1:ahb0:0:1:0): Queuing BDR
> ahb0: No longer in timeout
> (da1:ahb0:0:1:0): ECB 0xf40b6900 - timed out

I often get timeouts from the following disk:
Jan 26 02:13:18 fnu /kernel: da1 at ahb0 bus 0 target 1 lun 0
Jan 26 02:13:18 fnu /kernel: da1: <SEAGATE ST43400N 0105> Fixed Direct Access SCSI-2 device

I never had this problem under 2.2.X and I have used both disk and controller:
Jan 26 02:13:17 fnu /kernel: ahb0: AHA1740A Single Ended SCSI Adapter, FW Rev. E   , ID=7, 64 ECBs

for quite a few years now. The motherboard is a dual processor Tynan S1462 with a single 166mhz Pentium
processor.

Let me know if there is anything I can do to help track this down.
>How-To-Repeat:
Seems the daily periodic script creates enough disk activity to cause
the problem almost nightly.
>Fix:

>Release-Note:
>Audit-Trail:
State-Changed-From-To: open->feedback 
State-Changed-By: gibbs 
State-Changed-When: Wed Jan 27 19:37:19 PST 1999 
State-Changed-Why:  
Your problem is likely caused by one of two things: 
1) A bug in the driver that was corrected in rev 1.6 of i386/eisa/ahb.c. 
2) The fact that the new CAM 174X driver takes advantage of tagged 
queuing and the additional load this causes on the card or bus 
is exposing a firmware or cabling problem. 

To test out theory 1, pick up rev 1.6 of ahb.c from a local FreeBSD 
mirror, recompile your kernel with the new version of the driver, 
and see if the hangs persist. 

To test out theory 2, disable tagged queuing for this driver by changing 
line 1217 of sys/i386/eisa/ahb.c from 

cpi->hba_inquiry = PI_SDTR_ABLE|PI_TAG_ABLE; 

to: 

cpi->hba_inquiry = PI_SDTR_ABLE; 

Please let me know if either of these suggestions address your problem. 
Of course, #1 is the solution I hope addresses your hang. 


Responsible-Changed-From-To: freebsd-bugs->gibbs 
Responsible-Changed-By: gibbs 
Responsible-Changed-When: Wed Jan 27 19:37:19 PST 1999 
Responsible-Changed-Why:  
My driver. 
State-Changed-From-To: feedback->closed 
State-Changed-By: steve 
State-Changed-When: Thu Feb 18 07:57:35 PST 1999 
State-Changed-Why:  
Originator requested that this be closed. 
>Unformatted:
