[PARPORT] Re: SCSI Timeout on AIC7xxx when PPA is active.


David Campbell (campbell@gear.torque.net)
Thu, 28 May 1998 13:01:06 +0000


> Date: Wed, 27 May 1998 22:20:50 -0500
> From: Mark Garlanger <garlangr@cyberramp.net>
> To: linux-parport@torque.net, linux-scsi@vger.rutgers.edu
> Subject: [PARPORT] SCSI Timeout on AIC7xxx when PPA is active.

> Hi all,
>
> Is anyone successfully running the AIC7xxx and PPA drivers togather?
> When I try to use the PPA drive with the AIC7xxx driver, I am seeing
> SCSI Timeout on the AIC7xxx driver during the booting.

> I've tried both the stock ppa driver and the latest 1.42 version. The
> AIC7xxx is the stock 2.0.33 driver.

> Could there be some interaction in the mid-level scsi routines between
> the aic7xxx and ppa drivers?

Hmm...

Could you fire up your system in the following senarios (assuming root
partition on the HDD):

1) Zip drive unplugged but driver still in the kernel
2) Zip drive plugged in, turned on but no disk in drive
3) Zip drive plugged in, turned on but with a disk in drive

Expected results:
1) Everything OK - ppa driver does nothing
2) Everything OK - Minimal access to identify drive presence
3) Possible crash - A number of device accesses to read partition table

The output of the "dmesg" command will shed more light on the problem.
Also rebuild the kernel with ppa being a module. See if loading ppa as a
module causes a problem (but only after performing the test runs above).

If you can get the two drivers into the kernel together, try a bulk copy
between the two drives (ie: copy a very large file).

I have heard problems with ppa and other SCSI devices but I am not sure
where the problem lies, on its own the ppa driver works (almost - well I
haven't seen any problems lately) flawlessly. I have made large number of
modifications to ensure the driver "yields" to other processes due to the
low bandwith (0.6 MB/sec vs 10MB/sec). How this affects other low level
SCSI drivers is unknown.

My understanding of the mid-level SCSI protocol is that each device
processes the pending data independent of other low level drivers. I stand
to be corrected.

David Campbell
=======================================================
campbell@torque.net

"All parallel ports are equal - Some are more equal than others"

-- To unsubscribe, send mail to: linux-parport-request@torque.net --
-- with the single word "unsubscribe" in the body of the message. --



This archive was generated by hypermail 2.0b3 on Wed 30 Dec 1998 - 10:17:47 EST