Archive Python DAT tape and Adaptec

[Michael Meissner]

(this is being sent to both the linux-kernel and the Adaptec scsi
lists, so you might want to make sure followups are appropriate --
note for replies to me, I'm not yet on the AIC7xxx list, and am having
trouble following the linux-kernel list due to volume).

The Adaptec driver that's been shipping with the latest experimental
Linux kernels with a version # of 5.0.20 does not work reliably with
my 3 year old Archive DAT DDS-1 tape drive.  I had a similar problem
with another Archive DAT of the same vintage on my work computer
(about six months ago, the Adaptec was dealt with the tape drive
completely fine until I switched to TekRam 390{U,F} controllers).  The
Archive is an asynch-only device, and I have the jumpers set for
SCSI-2 and Parity support.

The home computer is a 233 Mhz Pentium-II on an Abit LX6 motherboard,
and it is currently cabled with the Adaptec having an internal 50 pin
ribbon cable connecting the tape drive (id #2), a 4x Toshiba CDROM (id
#4), and an active terminator on the end of the cable.  No connection
is made on either of the 68 pin ports.  The disks are on a separate
TekRam 390F controller.  I bumped up the SCSI reset time to 15 seconds
to give things time to settle down, but it didn't help.

About half the time when the driver resets the bus, I get something
like:

Sep  3 12:09:08 wogglebug kernel: aic7xxx: Warning - detected auto-termination on controller:
Sep  3 12:09:08 wogglebug kernel: aic7xxx: <Adaptec AHA-294X Ultra SCSI host adapter> at PCI 9/0
Sep  3 12:09:08 wogglebug kernel: aic7xxx: Please verify driver detected settings are correct.
Sep  3 12:09:08 wogglebug kernel: aic7xxx: If not, then please properly set the device termination
Sep  3 12:09:08 wogglebug kernel: aic7xxx: in the Adaptec SCSI BIOS by hitting CTRL-A when prompted
Sep  3 12:09:08 wogglebug kernel: aic7xxx: during machine bootup.
Sep  3 12:09:08 wogglebug kernel: aic7xxx: Cables present (Int-50 YES, Int-68 NO, Ext-68 NO)
Sep  3 12:09:08 wogglebug kernel: aic7xxx: Termination (Low ON, High ON)
Sep  3 12:09:08 wogglebug kernel: (scsi1) <Adaptec AHA-294X Ultra SCSI host adapter> found at PCI 9/0
Sep  3 12:09:08 wogglebug kernel: (scsi1) Wide Channel, SCSI ID=7, 16/255 SCBs
Sep  3 12:09:08 wogglebug kernel: (scsi1) BIOS enabled, IO Port 0x6800, IRQ 12
Sep  3 12:09:08 wogglebug kernel: (scsi1) IO Memory at 0xe4000000, MMAP Memory at 0xc4856000
Sep  3 12:09:08 wogglebug kernel: (scsi1) Downloading sequencer code... 413 instructions downloaded
Sep  3 12:09:08 wogglebug kernel: scsi1 : Adaptec AHA274x/284x/294x (EISA/VLB/PCI-Fast SCSI) 5.0.20/3.2.4
Sep  3 12:09:08 wogglebug kernel:        <Adaptec AHA-294X Ultra SCSI host adapter>
Sep  3 12:09:08 wogglebug kernel: scsi : 2 hosts.
Sep  3 12:09:08 wogglebug kernel: (scsi1:0:-1:-1) Scanning channel for devices.
Sep  3 12:09:10 wogglebug kernel: (scsi1:0:2:0) Parity error during phase Command.
Sep  3 12:09:16 wogglebug kernel: scsi : aborting command due to timeout : pid 140158, scsi1, channel 0, id 2, lun 0 Request Sense 00 00 00 10 00 
Sep  3 12:09:18 wogglebug kernel: SCSI host 1 abort (pid 140158) timed out - resetting
Sep  3 12:09:18 wogglebug kernel: SCSI bus is being reset for host 1 channel 0.

If I compile it as a module, the first time the driver is loaded, it
messes up the device name and type, for instance:

Sep  2 22:15:21 wogglebug kernel: (scsi1:0:2:0) Parity error during phase Command.
Sep  2 22:15:21 wogglebug kernel:   Vendor: MO @@/4@  Model: xY'@/             Rev:     
Sep  2 22:15:21 wogglebug kernel:   Type:   Medium Changer                     ANSI SCSI revision: 00
Sep  2 22:15:21 wogglebug kernel: scsi : aborting command due to timeout : pid 1175, scsi1, channel 0, id 3, lun 0 Test Unit Ready 00 00 00 00 00 
Sep  2 22:15:21 wogglebug kernel: SCSI host 1 abort (pid 1175) timed out - resetting

Subsequent unloads and loads of the driver will get the parity error,
but recognize the Adaptec correctly.

I downloaded Doug Ledford's beta driver from ftp.dialnet.net, and it
correctly recognized the tape with no error each of the times I've
loaded the driver.  Before linux 2.2 is released, can we please
upgrade the driver in the tree?

Here is the log output with the newest driver:

Sep  3 12:15:37 wogglebug kernel: aic7xxx: Warning - detected auto-termination on controller:
Sep  3 12:15:37 wogglebug kernel: aic7xxx: <Adaptec AHA-294X Ultra SCSI host adapter> at PCI 9/0
Sep  3 12:15:37 wogglebug kernel: aic7xxx: Please verify driver detected settings are correct.
Sep  3 12:15:37 wogglebug kernel: aic7xxx: If not, then please properly set the device termination
Sep  3 12:15:37 wogglebug kernel: aic7xxx: in the Adaptec SCSI BIOS by hitting CTRL-A when prompted
Sep  3 12:15:37 wogglebug kernel: aic7xxx: during machine bootup.
Sep  3 12:15:37 wogglebug kernel: aic7xxx: Cables present (Int-50 YES, Int-68 NO, Ext-68 NO)
Sep  3 12:15:37 wogglebug kernel: (scsi1) <Adaptec AHA-294X Ultra SCSI host adapter> found at PCI 9/0
Sep  3 12:15:37 wogglebug kernel: (scsi1) Wide Channel, SCSI ID=7, 16/255 SCBs
Sep  3 12:15:37 wogglebug kernel: (scsi1) Downloading sequencer code... 419 instructions downloaded
Sep  3 12:15:37 wogglebug kernel: scsi1 : Adaptec AHA274x/284x/294x (EISA/VLB/PCI-Fast SCSI) 5.1.0pre8/3.2.4
Sep  3 12:15:37 wogglebug kernel:        <Adaptec AHA-294X Ultra SCSI host adapter>
Sep  3 12:15:37 wogglebug kernel: scsi : 2 hosts.
Sep  3 12:15:37 wogglebug kernel:   Vendor: ARCHIVE   Model: Python 25588-XXX  Rev: 2.96
Sep  3 12:15:37 wogglebug kernel:   Type:   Sequential-Access                  ANSI SCSI revision: 02
Sep  3 12:15:37 wogglebug kernel: Detected scsi tape st0 at scsi1, channel 0, id 2, lun 0
Sep  3 12:15:38 wogglebug kernel:   Vendor: TOSHIBA   Model: CD-ROM XM-3501TA  Rev: 3054
Sep  3 12:15:38 wogglebug kernel:   Type:   CD-ROM                             ANSI SCSI revision: 02
Sep  3 12:15:38 wogglebug kernel: Detected scsi CD-ROM sr0 at scsi1, channel 0, id 4, lun 0
Sep  3 12:15:38 wogglebug kernel: (scsi1:0:4:0) Synchronous at 4.0 Mbyte/sec, offset 15.

Note, the newest NCR53C8XX drivers also have problems with the Archive
DATs, but I will post another message about them, once I can get the
approprate information for the aborts.

-- 
Michael Meissner, Cygnus Solutions (Massachusetts office)
4th floor, 955 Massachusetts Avenue, Cambridge, MA 02139, USA
meissner at cygnus.com,	617-354-5416 (office),	617-354-7161 (fax)

To Unsubscribe: send mail to majordomo at FreeBSD.org
with "unsubscribe freebsd-aic7xxx" in the body of the message