Browse Prior Art Database

Asynchronous Error Queue Design for Error Recovery

IP.com Disclosure Number: IPCOM000116908D
Original Publication Date: 1995-Nov-01
Included in the Prior Art Database: 2005-Mar-31
Document File: 2 page(s) / 81K

Publishing Venue

IBM

Related People

Fry, SM: AUTHOR [+3]

Abstract

In multi-processor product applications with a shared resource for error handling, there is always the problem of interpreting concurrent errors. If errors are allowed to be queued, then the problem is compounded since the queues are most likely built asynchronously yet they represent a cause and effect relationship.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Asynchronous Error Queue Design for Error Recovery

      In multi-processor product applications with a shared resource
for error handling, there is always the problem of interpreting
concurrent errors.  If errors are allowed to be queued, then the
problem is compounded since the queues are most likely built
asynchronously yet they represent a cause and effect relationship.

      This article discusses a solution implemented in a tape product
application containing two microprocessors.  The lower processor is a

Digital Signal Processor primarily focused on control and monitoring
of basic drive motion and track following servo.  It interfaces with
an Intel I960 microprocessor which controls the passage of data
between a SCSI Host and tape via data buffers and the I960/DSP
interface.  These two processors maintain distinct error queues that
have a relative time relationship.  Together they detect and log
errors which occur:
  1.  While a cartridge is loading and unloading
  2.  While a cartridge is loaded and idle
  3.  During data transfer or other host-initiated task

      The I960 subsystem initiates and controls high level commands
requested by the SCSI host.  It is responsible for logging hardware
and firmware detected errors that occur during tape motion.  It will
shut down drive motion and initiate data error recovery through the
I960/DSP interface if appropriate.  The DSP Error Queue is maintained
within the I960/DSP interface, which provides clock commonality for
both error queues.  DSP-detected errors are logged asynchronously
with respect to I960-detected errors.  Since the I960 and DSP
function together, errors detected by either microprocessor subsystem
may introduce defects into the other.  It is important, therefore, to
maintain a time-relationship of errors in both subsystems.

      The asynchronous...