Browse Prior Art Database

Memory Subsystem Failure Analysis and Isolation

IP.com Disclosure Number: IPCOM000042808D
Original Publication Date: 1984-Jun-01
Included in the Prior Art Database: 2005-Feb-04
Document File: 1 page(s) / 12K

Publishing Venue

IBM

Related People

Elliott, JM: AUTHOR [+3]

Abstract

Event-independent memory subsystem analysis and isolation t a represent a method of diagnosing CPU-to-storage failures and identifying the failing subsystem card. It is event independent because all isolation in the CPU-to-storage area is done independent of the number and types of failures. Systems with multiple storage cards can use this technique to greatest advantage. As in prior machines, IPL diagnostics perform a number of tests to determine if there are defects in storage cells, cell addressability, and error detection/correction. Unlike prior machines, when main-storage (MS) errors are detected, error information is saved and testing continues. All MS tests are run in this fashion to gather as much information about the failure as possible. Two tables are used for the collection of error information.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 73% of the total text.

Page 1 of 1

Memory Subsystem Failure Analysis and Isolation

Event-independent memory subsystem analysis and isolation t a represent a method of diagnosing CPU-to-storage failures and identifying the failing subsystem card. It is event independent because all isolation in the CPU-to- storage area is done independent of the number and types of failures. Systems with multiple storage cards can use this technique to greatest advantage. As in prior machines, IPL diagnostics perform a number of tests to determine if there are defects in storage cells, cell addressability, and error detection/correction. Unlike prior machines, when main-storage (MS) errors are detected, error information is saved and testing continues. All MS tests are run in this fashion to gather as much information about the failure as possible. Two tables are used for the collection of error information. The "Bad 2K Page" table contains the 2K- byte-page addresses of MS test failures. The "card-by-card" (CBC) table contains failing test IDs and error syndromes per card. Information is entered into the tables by the MS tests. Entries are made into the "Bad 2K page table" if the table is not full and if the failing 2K page address is not already in the table. Before returning to the MS test, a calculation is performed to determine which card has the failing 2K page address. The CBC table contains one record for every card. The test ID and error syndrome are logically ORed into the record corresponding to the calc...