Browse Prior Art Database

Exclusion Circuit - Identifying First Unrecoverable Event in a Series of Unrecoverable Events

IP.com Disclosure Number: IPCOM000013629D
Original Publication Date: 2001-May-17
Included in the Prior Art Database: 2003-Jun-18
Document File: 4 page(s) / 90K

Publishing Venue

IBM

Abstract

Exclusion Circuit Identifying First Unrecoverable Event in a Series of Unrecoverable Events Disclosed is a device that identifies the first occurrence in a series of unrecoverable events.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 53% of the total text.

Page 1 of 4

  Exclusion Circuit - Identifying First Unrecoverable Event in a Series of Unrecoverable Events

Disclosed is a device that identifies the first occurrence in a series of unrecoverable events.

In the system that is being developing error detection is critical. Some of the errors that are detected can be recovered from. Some recoverable errors are; correctable memory ECC and PCI device errors. Other errors are unrecoverable; for example: parity error, some bus protocol errors, bus time-out errors and internal chip errors. If these unrecoverable errors where to propagate throughout the system, corruption of the customers data would occur. Therefore the system will halted immediately upon the system detecting these unrecoverable errors. An Exclusion Circuit is needed to identify the first unrecoverable event.

Ports

SPINT_OUT

Recoverable Error Path

SPINT_OUT

Recoverable Error Path

P

P

P

P

Recoverable Error Path

Chip 1

Chip 2

Memory

Service Processor I2C Handler

Chip 3

PCI-X

1

Page 2 of 4

possible in order to keep the processors from accessing the Front Side Bus (i.e. the processors are locked up and are no longer executing code). The chip (Chip1, Chip2 or Chip3) that detects the error asserts its respective SPINT_OUT (Service Processor Interrupt) pin so that the service processor invokes its error handling routine. SPINT_OUT is only asserted on unrecoverable errors. The SPINT_OUT is the interrupt net of the I2C bus of clock, data and interrupt bus. An exclusion register captures the first occurrence of the SPINT_OUT assertion. (This is done so that the Service Processor facility will read only the error registers of the chip that asserted SPINT_OUT first.) The Service Processor interrupt handler is invoked. Through the I2C structure, the Service Processor facility reads the I2C register to determine which chip asserted the I2C interrupt. The Service Processor facility then stores the error registers into its memory for later Error Log Analysis.

Following steps show sequence for unrecoverable errors:

Unrecoverable error detected

Chip1 or Chip2 assert BPRI (Bus Priority Request)

Chip1, Chip2 or Chip3 assert SPINT_OUT (I2C Interrupt)

System logic captures first occurrence of SPINT_OUT

SP reads I2C registers, determines which chip first interrupted

SP reads appropriate registers and stores it

SP posts...