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

Re: "Hang" on boot



Seems like you have a problem with interrupt sharing... Even if your board
uses INTA-INTC, it's possible that all of these INT's are routed to the same
ISA IRQ. That's legal.

Another issue to consider is that you have comply to the Pci2PCI Brige
specification when designing expansion boards containing bridges. In your
case this means that there is a rule for combining Device Numbers and INTx
Pin useage. I.e. a device that uses Ad31 for IDSEL assignment, has to use
INTC on the primary bus connector. There's a table in the P2P Bridge spec
defining this relöationship.

regards,

.
Peter Marek
General Director
MarekMicro GmbH
Kropfersrichter Str. 6-8
D-92237 Sulzbach-Rosenberg
Germany
Phone: 049 - 9661 - 908 - 210
Fax:      049 - 9661 - 908 - 100
----- Original Message -----
From: David N Hicks <dave@prescoinc.com>
To: <pci-sig@znyx.com>
Sent: Friday, June 23, 2000 6:41 PM
Subject: "Hang" on boot


> Dear all,
>
> I have aproblem with a PCI plug-in card I designed in '96 that has an
bridge
> chip and three "local" PCI devices (on the embedded bus) and I was
wondering
> if anyone could provide some insight.
>
> Background:
> The card was designed for a custom application (low volume, installed by
> OEM) and uses the embedded chips default vendor and device IDs.  The card
> does not support subsystem vendor IDs.  The three embedded devices
attached
> to the "PCI to PCI Bridge" are a "Bridge Device", a "Memory Adapter", and
a
> "SCSI Bus Device".  These devices are assigned to Device IDs 4, 5, and 6
> (respectively).  Each of the three embedded PCI devices uses 1 interrupt
> which is assigned to INTA for the device.  The interrupts from the
secondary
> bus are routed to INTA, INTB, and INTC on the primary bus PCI connector.
>
> Problem:
> The card is not working when installed in "modern" motherboards.  It
doesn't
> "crash" the system, but keyboards and network interfaces stop working.
> Testing to date is minimal, but the problem seems like it may be an
> interrupt assignment issue.  In a failing system, the Memory Adapter is
> assigned the same interrupt as a network card and a USB device (Interrupt
> Line B).
>
> Any insights or recommendations on where I can look for answers?
>
> Thanks,
> Dave Hicks