[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: how do we write a single byte in the config. registers?
- To: "John R Pierce" <pierce@scruznet.com>
- Subject: RE: how do we write a single byte in the config. registers?
- From: "Taylor, Mike" <mike@buntypost.dundee.ncr.com>
- Date: Thu Nov 21 08:44 GMT 1996
- Encoding: 38 TEXT
- Return-Receipt-To: <mike@buntypost.dundee.ncr.com>
John,
Thanks for the reply, but the Config #2 accesses are not trapped by the
motherboard. Method #2 uses IO space C000-CFFF. Bits 11-8 select the PCI
device (1 of 16), bits 7-2 select a DWORD in the device's configuration
space. IO accesses are therefore to C0xx, C1xx, thru' CFxx. This problem
exists, a well known graphics card BIOS attempts Config #2 cycles and
interferes with a proprietary card in our system.
[ The card is actually the MATROX Millenium, MATROX are now aware of the
problem and their next BIOS, 2.3, will correct the problem.]
Regards,
Mike
----------
From: John R Pierce[SMTP:pierce@scruznet.com]
Sent: 20 November 1996 08:23
To: Taylor, Mike; pci-sig
Subject: Re: how do we write a single byte in the config. registers?
> Beware of config. method #2. If your host chipset does not support
method
> #2, Triton family for example, then the accesses appear on the ISA bus
> and ISA cards can react to the aliases of the config addresses. e.g.
C200
> can be seen as x200.
Ah, but isn't the port in question always X0XX? [I don't have my PCI
references handy to recall the details here...] If so, that shouldn't be
a
problem, as addresses with 00 in bits 8,9 are reserved for the
motherboard and
a PCI motherboard will decode bits 10-15=0 before enabling motherboard
IO...
-jrp
0 L :