From owner-freebsd-hardware  Tue Dec 17 18:24:45 1996
Return-Path: <owner-hardware>
Received: (from root@localhost)
          by freefall.freebsd.org (8.8.4/8.8.4) id SAA10080
          for hardware-outgoing; Tue, 17 Dec 1996 18:24:45 -0800 (PST)
Received: from Octopussy.MI.Uni-Koeln.DE (Octopussy.MI.Uni-Koeln.DE [134.95.166.20])
          by freefall.freebsd.org (8.8.4/8.8.4) with SMTP id SAA10073
          for <freebsd-hardware@FreeBSD.ORG>; Tue, 17 Dec 1996 18:24:40 -0800 (PST)
Received: from x14.mi.uni-koeln.de (annexr3-15.slip.Uni-Koeln.DE) by Octopussy.MI.Uni-Koeln.DE with SMTP id AA03693
  (5.67b/IDA-1.5 for <freebsd-hardware@FreeBSD.ORG>); Wed, 18 Dec 1996 03:24:36 +0100
Received: (from se@localhost) by x14.mi.uni-koeln.de (8.8.4/8.6.9) id DAA15824; Wed, 18 Dec 1996 03:10:11 +0100 (CET)
Message-Id: <Mutt.19961218030851.se@x14.mi.uni-koeln.de>
Date: Wed, 18 Dec 1996 03:08:51 +0100
From: se@FreeBSD.ORG (Stefan Esser)
To: ruey@tpts1.seed.net.tw (PC_user)
Cc: freebsd-hardware@FreeBSD.ORG
Subject: Re: TekrAm DC-390F scsi card
References: <199612171728.BAA24094@tpts1.seed.net.tw>
X-Mailer: Mutt 0.53
Mime-Version: 1.0
In-Reply-To: <199612171728.BAA24094@tpts1.seed.net.tw>; from PC_user on Dec 18, 1996 01:28:45 +1419
Sender: owner-hardware@FreeBSD.ORG
X-Loop: FreeBSD.org
Precedence: bulk

On Dec 18, ruey@tpts1.seed.net.tw (PC_user) wrote:
> Hi,
> 
>  I'm trying to get my TekrAm DC-390F (ncr53c875) work under FreeBSD. But 
> I always get the following message while booting 
> 
> Probing for devices on PCI bus 0:
> chip0 <SiS 85c496> rev 1 on pci0:5
> de0 <Digital 21041 Ethernet> rev 17 int a irq 9 on pci0:10
> de0: 21041 [10Mb/s] pass 1.1
> de0: address 00:80:c8:34:93:bd
> vga0 <VGA-compatible display device> rev 6 int a irq 9 on pci0:11
> ncr0 <ncr 53c875 wide scsi> rev 1 int a irq 10 on pci0:12
> CACHE TEST FAILED: reg dstat-sstat2 readback ffffffff.
> CACHE INCORRECTLY CONFIGURED.
> Probing for devices on the ISA bus:
> sc0 at 0x60-0x6f irq 1 on motherboard
> ...
> 
> and the system cannot detect any scsi device.
>  Please tell me how to fix this problem.

Could you please send VERBOSE boot messages
(ie. enter "-v"- at the "Boot: " prompt).

The NCR driver does some basic tests before it 
starts to actually initialize the controller.
And the first one already failed on your system.

The test writes "0xffffffff" into some NCR
register, which are guaranteed to return 0 in 
some of their bits. If the value read back does
not contain those 0 bits, then the CPU read it
from its cache, not from the NCR chip (which 
must be memory mapped into a non-cacheable 
memory region).

Seems that your CPU read back the 0xffff with
no single bit cleared. This prevents operation
of the NCR (and any other memory mapped PCI 
card) on your motherboard.

Please send the VERBOSE boot message log. You 
will find lines like the following in it:

ncr1 <ncr 53c875 wide scsi> rev 1 int a irq 12 on pci0:4
	mapreg[10] type=1 addr=0000e400 size=0100.
	mapreg[14] type=0 addr=fbfde000 size=0100.
	mapreg[18] type=0 addr=fbfdd000 size=1000.
	reg20: virtual=0xf54f8000 physical=0xfbfde000 size=0x100

On my system the Tekram is mapped to address
0xfbfde000 (mapreg[14]).

I've once before seen a system, whose PCI BIOS
mapped some PCI chip into memory space below 1MB.

It is often possible to remap any PCI device to 
some other unused memory address.
You may try to build a kernel with NCR_IOMAPPED
defined. This will make the driver use port I/O
instead of memory accesses.

Regards, STefan