Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 26 Jan 1996 15:29:10 +0100
From:      se@zpr.uni-koeln.de (Stefan Esser)
To:        Fred Cawthorne <fred@jjarray.umd.edu>
Cc:        freebsd-scsi@freebsd.org
Subject:   Re: worm device.
Message-ID:  <199601261429.AA08942@Sysiphos>
In-Reply-To: Fred Cawthorne <fred@jjarray.umd.edu> "Re: worm device." (Jan 25, 22:06)

next in thread | previous in thread | raw e-mail | index | archive | help
On Jan 25, 22:06, Fred Cawthorne wrote:
} Subject: Re: worm device.

} I am tracking down a problem now where I can't do alot of disk activity
} on my new DEC scsi drive when the HP CD-R drive is plugged in.  I unplugged
} it from the scsi bus and everything is fine.  Plugging in and unplugging
} my external cdrom and zip drive doesn't change anything.  Keep in mind that
} there is no activity on the HP CD-R side of things.  This is really 
} strange...  I think I am going to have to plug in an Adaptec 1542 and 
} use that for the HP CD-R drive.  
} Here's what happened when I tried to copy my entire /user directory (600M)
} into a partiton on the 2 gig DEC drive:
} 
} First, here's the stuff on my scsi bus.

} 	ncr0 <ncr 53c810 scsi> rev 1 int a irq 10 on pci0:5
} 	ncr0 waiting for scsi devices to settle
} 	(ncr0:0:0): "MICROP 2217-15MQ1001901 HQ30" type 0 fixed SCSI 2
} 	sd0(ncr0:0:0): Direct-Access 
} 	sd0(ncr0:0:0): FAST SCSI-2 100ns (10 Mb/sec) offset 8.
} 	1685MB (3450902 512 byte sectors)
} 	(ncr0:2:0): "DEC DSP3210S 436D" type 0 fixed SCSI 2
} 	sd1(ncr0:2:0): Direct-Access 
} 	sd1(ncr0:2:0): FAST SCSI-2 100ns (10 Mb/sec) offset 8.
} 	2047MB (4194303 512 byte sectors)
} 	(ncr0:6:0): "HP C4324/C4325 1.20" type 5 removable SCSI 2
} 	worm0(ncr0:6:0): Write-Once - UNTESTED worm0: oops not queued
} 	with 333227 2048 byte blocks.

} And the error message:  (After copying a few hundred megs over)

} 	ncr0:2: ERROR (20:40) (8-28-0) (8/13) @ (1150:180000a0).

The status 20:40 indicates a Bus Fault, i.e. a PCI bus transaction
did not succeed within a certain number of bus cycles. This may be 
a chip set problem (have seen it on the old rev. 2 of the Saturn,
if PCI Bursts are enabled) or may be caused by the NCR trying to 
access a memory address outside the actually installed DRAM range.

No other error bit is set (the 20 is BF, the 40 is CMP and indicates 
that the SCSI target is connected, i.e. Arbitration Complete).

} 		script cmd = 88030000
} 		reg:	 da 10 80 13 47 08 02 1f 01 08 82 28 00 00 08 00.
} 	ncr0: have to clear fifos.
} 	ncr0: restart (fatal error).
} 	sd1(ncr0:2:0): COMMAND FAILED (9 ff) @f0b5aa00.
} 	ncr0: restart (ncr dead ?).
} 	sd0(ncr0:0:0): FAST SCSI-2 100ns (10 Mb/sec) offset 8.
} 	sd1(ncr0:2:0): FAST SCSI-2 100ns (10 Mb/sec) offset 8.

The NCR was successfully reinitialized, but the ongoing SCSI 
transaction had been interrupted.

} I filled the whole 2.1 gig drive up fine when I unplugged the CD-R and rebooted...
} Any ideas?

I doubt that the BF is caused by a driver error. This has been observed
before on systems with chip sets that can't actually support bus-master
controllers under high load, if all PCI performance options have been
enabled. This doesn't seem to be an issue with current chip sets, AFAIK,
but please let me know, what you are using ...

Regards,


-- 
 Stefan Esser, Zentrum fuer Paralleles Rechnen		Tel:	+49 221 4706021
 Universitaet zu Koeln, Weyertal 80, 50931 Koeln	FAX:	+49 221 4705160
 ==============================================================================
 http://www.zpr.uni-koeln.de/~se			  <se@ZPR.Uni-Koeln.DE>



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199601261429.AA08942>