From owner-freebsd-bugs  Wed Mar 12 13:10:07 1997
Return-Path: <owner-bugs>
Received: (from root@localhost)
          by freefall.freebsd.org (8.8.5/8.8.5) id NAA11559
          for bugs-outgoing; Wed, 12 Mar 1997 13:10:07 -0800 (PST)
Received: (from gnats@localhost)
          by freefall.freebsd.org (8.8.5/8.8.5) id NAA11551;
          Wed, 12 Mar 1997 13:10:03 -0800 (PST)
Date: Wed, 12 Mar 1997 13:10:03 -0800 (PST)
Message-Id: <199703122110.NAA11551@freefall.freebsd.org>
To: freebsd-bugs
Cc: 
From: vanmaren@fast.cs.utah.edu (Kevin Van Maren)
Subject: Re: kern/2944: NCR/SYM875 card with Ultra drive
Reply-To: vanmaren@fast.cs.utah.edu (Kevin Van Maren)
Sender: owner-bugs@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

The following reply was made to PR kern/2944; it has been noted by GNATS.

From: vanmaren@fast.cs.utah.edu (Kevin Van Maren)
To: se@freebsd.org, vanmaren@fast.cs.utah.edu
Cc: FreeBSD-gnats-submit@freebsd.org
Subject: Re: kern/2944: NCR/SYM875 card with Ultra drive
Date: Wed, 12 Mar 1997 14:06:27 -0700

 > On Mar 10, Kevin Van Maren <vanmaren@marker.cs.utah.edu> wrote:
 > > >Synopsis:       NCR/SYM875 card with Ultra drive
 > > 
 > > When "changing root device to sd0a" comes up, the machine sits there
 > > for several minutes, until it continues (after printing an error
 > > message).  The machine appears to operate correctly after that.
 > > 
 > > changing root device to sd0a
 > > ncr0: aborting job ...
 > > ncr0:6: ERROR (90:0) (8-0-0) (0/13) @ (a54:50000000).
 > 
 > The error code (0x90) indicates a command ahs been aborted.
 > I can only guess what command that might be, but I assume a
 > incompatibility between the driver and your particular disk
 > drive.
 > 
 > Is it possible, that the driver tries to use WIDE transfers 
 > with the disk ?
 > 
 > This may happen, if both the controller and the drive are
 > capable of WIDE transfers, but the SCSI bus is only 8bit
 > wide. In such a case, the command would time out, and the
 > driver would continue in a failsafe mode ...
 
 The drive is narrow -- I don't believe they make a wide version.
 
 > In order to diagnose this problem, I need SCSI DEBUG logs.
 > You'll have to rebuild your kernel with SCSI debug enabled
 > (an option of the generic SCSI driver, not the NCR driver).
 > Use the kernel option line "options SCSIDEBUG" and edit 
 > /sys/scsi/scsi_debug.h to define DEBUGTARGET as 6 ...
 
 I'll try to get to it tonight.
 
 > > Also, note that Ultra negotiation is not enabled by the driver,
 > > although the BIOS does initially enable it.
 > 
 > The driver disables it, currently. I have been to busy to
 > cleanly add Ultra-20 support. It is easy as a quick hack,
 > but that's not what I'm after ...
 
 Okay.  I'm sure you announce when it is working right.
 
 > > (ncr0:6:0): "QUANTUM FIREBALL_TM3200S 300X" type 0 fixed SCSI 2
 > > sd0(ncr0:6:0): Direct-Access 
 > > sd0(ncr0:6:0): 10.0 MB/s (100 ns, offset 15)
 > 
 > Hmmm, no trace of an attempt to use WIDE transfers ...
 > 
 > > ncr0:6: ERROR (90:0) (8-0-0) (0/13) @ (a54:50000000).
 > >         script cmd = 740a8f00
 > >         reg:     de 00 00 13 47 00 0f 0f 35 08 86 00 90 00 0f 02.
 > > ncr0: restart (fatal error).
 > > sd0(ncr0:6:0): COMMAND FAILED (9 ff) @f19f9c00.
 > > sd0(ncr0:6:0): 10.0 MB/s (100 ns, offset 15)
 > 
 > I have just got another idea:
 > 
 > Please check whether a kernel with "options FAILSAFE" does
 > also show this behaviour ...
 
 This was with the install kernel.  I rebuilt a modified GENERIC
 kernel (changed root to sd0 from wd0, got rid of drivers, etc) 
 and I didn't get this message anymore.  (I think FAILSAFE is 
 enabled; I'll have to check when I get home).
 
 Should I try GENERIC unmodified except with the SCSIDEBUG?
 
 > Regards, STefan
 
 Thanks,
 Kevin