[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
PCI exclusive access question
- To: Mailing List Recipients <pci-sig-request@znyx.com>
- Subject: PCI exclusive access question
- From: Steven Frechette <stevenf@hw.stratus.com>
- Date: Mon, 4 Nov 1996 09:55:46 -0500 (EST)
- Resent-Date: Mon, 4 Nov 1996 09:55:46 -0500 (EST)
- Resent-From: pci-sig-request@znyx.com
- Resent-Message-Id: <"pSDb02.0.E4.ZhYVo"@dart>
- Resent-Sender: pci-sig-request@znyx.com
Hi everyone,
I am designing a host bridge in an intel processor based
system which uses the piix3 chip as the legacy expansion bus bridge. No devices
with cacheable memory will exist on any of the PCI busses.
On page 73 of the PCI specification the following statement is made: "In
general, LOCK# is to be used by bridges to provide backward compatibility
for existing devices and to prevent deadlocks. A host bus bridge may need
to support LOCK# as a target to provide backward compatibility with some
existing add-in cards that reside behind standard expansion bus bridges."
Question: The piix3 specification makes no references to the use
of the PCI LOCK# signal. Since this is my expansion bus bridge, is it
reasonable to assume that my host bridge should not have to support the
LOCK# mechanism when it is the target of a transaction. Also, does anyone
know of any other PCI devices that would want to master a locked
transaction upstream? Possibly graphics cards?
Steven E. Frechette E-Mail: stevenf@hw.stratus.com
Hardware Engineer Phone: 508-490-6409
Stratus Computer, Inc.
å t b