[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Master Wait States



Yes, I know that Socket 7 era SiS chipsets (http://www.sis.com.tw) do keep 
IRDY# deasserted for a few cycles when starting a transaction before finally 
asserting it.
I don't know about the recent ones (i.e., Pentium 4 or Athlon generation), 
but it is likely that they still do that.
You can always check that by going to a computer dealer, buy one SiS 
chipset-based motherboard, and hook up a logic analyzer to see how the PCI 
bus behaves.
The PCI specification says that initiator has the right to insert wait 
states when starting a transaction, as long as IRDY# is asserted within 8 
clock cycles of the assertion of FRAME# (PCI 2.2 specification  3.5.2. 
Master Data Latency).
        Speaking of corner cases, you should also make sure that the PCI 
interface you sound like you are working on can handle a burst configuration 
transaction properly.
The specification says a burst configuration transaction is permitted, and I 
heard that Intel 430FX chipset (Released in 1995) does actually perform it.
Always signal Disconnect with Data when responding to any configuration 
transaction.


Kevin Brace



>From: "Charles Neumann" <c.neumann@aristoslogic.com>
>To: PCI SIG Developers Forum <pci-sig@znyx.com>
>Subject: Master Wait States
>Date: Fri, 20 Dec 2002 13:56:50 -0800
>
>Is anybody aware of PCI Masters that will actually insert waitstates at
>the begining of a transaction, by delaying assertion of irdy?  I was
>wondering if this is a reasonable corner case for testing?  Many thanks,
>
>     - Charlie Neumann

_________________________________________________________________
MSN 8: advanced junk mail protection and 3 months FREE*. 
http://join.msn.com/?page=features/junkmail&xAPID=42&PS=47575&PI=7324&DI=7474&SU= 
http://www.hotmail.msn.com/cgi-bin/getmsg&HL=1216hotmailtaglines_advancedjmf_3mf