From owner-freebsd-mobile  Sat May 20  7:39:57 2000
Delivered-To: freebsd-mobile@freebsd.org
Received: from spike.porcupine.org (umbilical.porcupine.org [168.100.189.1])
	by hub.freebsd.org (Postfix) with ESMTP
	id E590737B6B4; Sat, 20 May 2000 07:39:51 -0700 (PDT)
	(envelope-from wietse@porcupine.org)
Received: by spike.porcupine.org (Postfix, from userid 100)
	id 0D9664563D; Sat, 20 May 2000 10:39:50 -0400 (EDT)
Subject: Data point: FreeBSD 4.0 + removable SCSI support
In-Reply-To: <20000519154650.2816B4563D@spike.porcupine.org> from Wietse Venema at "May 19, 0 11:46:50 am"
To: wietse@porcupine.org (Wietse Venema)
Cc: tbuswell@acadia.net, wietse@porcupine.org,
	freebsd-mobile@FreeBSD.ORG, freebsd-scsi@FreeBSD.ORG
X-Time-Zone:  USA EST, 6 hours behind central European time
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Content-Length: 2095      
Message-Id: <20000520143950.0D9664563D@spike.porcupine.org>
Date: Sat, 20 May 2000 10:39:50 -0400 (EDT)
From: wietse@porcupine.org (Wietse Venema)
Sender: owner-freebsd-mobile@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.org

Ted Buswell:
> Even though you don't care anymore, whether or not the Jaz is upsetting
> the card or driver by spewing garbage is easy to determine: with nothing
> plugged into your bus adapter, does the rescan still lockup your
> machine?

Without devices attached to the SlimSCSI 1460B, "camcontrol rescan 0"
reports success.

With random hardware connected to the SlimSCSI 1460B, "camcontrol
rescan 0" reports success, and the kernel says:

    May 20 14:04:01 bristle /kernel: sa0 at aic1 bus 0 target 5 lun 0
    May 20 14:04:01 bristle /kernel: sa0: <ARCHIVE VIPER 150  21531
	-004> Removable Sequential Access SCSI-CCS device
    May 20 14:04:01 bristle /kernel: sa0: 3.300MB/s transfers

For the sake of completeness, this is what the kernel logs when the
SlimSCSI 1460B itself is installed:

    May 20 13:56:31 bristle /kernel: aic1: <Adaptec 6260/6360 SCSI
	controller> at port 0x340-0x35f irq 3 slot 0 on pccard0
    May 20 13:56:31 bristle /kernel: aic1: aic6360, dma, disconnection,
	parity check

With the same IRQ setting and the same I/O port setting on the same
machine, FreeBSD3.4+PAO3 is able to access the same JAZ 2GB drive
successfully, just like the much older FreeBSD 2.2.8.

However, FreeBSD3.4+PAO3 occasionally drops me into the kernel
debugger after I execute /usr/bin/halt. That is not supposed to
happen.

> And while we're talking about it, is the lockup a real
> lockup, or is it simply that the camcontrol process hangs and you have 
> it in your startup sequence in such a way that the boot is not able to 
> continue?

The machine does not respond to keyboard events or network events.
The reset and power buttons still have effect, though.

Conclusion: the aic device driver has a problem with an unexpected
(lack of) response from the JAZ 2GB drive. FreeBSD3.4+PAO3 gripes,
too, but it does recover (except for sometimes dropping me into a
kernel debugger).

If you need to know more, there is a few more days before I have
to return the test disk. After that, I will not be able to do
further experiments with 4.0 on that laptop until July or so.

	Wietse


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