Skip site navigation (1)Skip section navigation (2)
Date:      Sun, 3 Dec 2000 23:21:00 -0600
From:      "Zach N. Heilig" <zach@uffdaonline.net>
To:        "Justin T. Gibbs" <gibbs@scsiguy.com>
Cc:        Co-op <sparqy2@home.com>, freebsd-stable@FreeBSD.ORG, freebsd-bugs@FreeBSD.ORG
Subject:   Re: Old AHA-2940 problem [was: kernel panics in 4.2]
Message-ID:  <20001203232100.A11302@murkwood.znh.org>
In-Reply-To: <28155B5C7DCC3B0451AD05D2@mail.uffdaonline.net>; from gibbs@scsiguy.com on Sun, Dec 03, 2000 at 09:15:15AM -0700
References:  <3A2A6728.7D6698BA@home.com> <28155B5C7DCC3B0451AD05D2@mail.uffdaonline.net>

next in thread | previous in thread | raw e-mail | index | archive | help
On Sun, Dec 03, 2000 at 09:15:15AM -0700, Justin T. Gibbs wrote:
> >Thanks Antonio,
> >Looking at the 4.2 release notes, the ahc driver was updated.  Wonder if the
> >kernel guys made changes that made supporting the old bios's impossible.
> 
> At the point that the kernel takes over control of the 2940, the BIOS
> should be irrelevant.  I'll see if I can scare up a 2940 of that
> vintage and reproduce your problem.  The probable cause is that we,
> at some point during initialization, reference a piece of SCB ram
> that has never been written to.

I have one that seems to have this problem.
ahc0: <Adaptec 2940 SCSI adapter> port 0xe400-0xe4ff mem 0xea101000-0xea101fff irq 10 at device 11.0 on pci0
aic7870: Single Channel A, SCSI Id=7, 16/255 SCBs

The exact symptons I see:

power-on boot: 4.2-stable: infinite loop (right before what would
normally be 'Waiting 2 seconds for SCSI devices to settle' -- it's a
multiple line panic message, so I can't read it (it scrolls far too
fast).  In -current: a regular page-not-present panic.  This happens
100% of the time.

in 4.2-stable, hitting reset (the only way to cause a reboot, without
turning off the power) allows the next (and any other reboot, until
power off) to succeed properly. in -current, I just need to wait for the
automatic reboot.

after a successful boot, devices on the ahc bus seem to be usable (I
kept a FreeBSD cvs repository, /usr/src, /usr/ports, /usr/obj on drives
attached to that controller, and didn't have problems running cvsup,
cvs update, or make world. (past tense because I moved the entire chain
to a symbios based card -- now I just need to find an inexpensive PCI
SCSI card with a narrow external connector and I can ditch the ancient
2940 for good).

-- 
Zach Heilig <zach@uffdaonline.net>


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




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20001203232100.A11302>