Browse Prior Art Database

Method for monitoring multiple systems in a finite time frame

IP.com Disclosure Number: IPCOM000020134D
Original Publication Date: 2003-Oct-28
Included in the Prior Art Database: 2003-Oct-28
Document File: 1 page(s) / 38K

Publishing Venue

IBM

Abstract

Computer systems executing firmware have no control of the processor while the firmware is running. If the OS was suppose to handle an interrupt or some time cirtical procedure while the processor was running firmware code it would be unable to do so. A method to exit firmware code based on a timer expiring needed to be invented. In the pSeries enterprise server the RTAS environment quickly executes hardware specific code. Some code, in particular hardware monitoring code needs to scan multiple sub-systems. Scaning multiple system can take a lengthly amount of time. Since the same processor that runs firmware code also handles interrupts the RTAS code must execute quickly enough so the interurpt can be handled before any detrimental effect is encountered. By RTAS monitoring the length of time it has currently used, RTAS can decide whether to continue monitoring more systems or simply remember where it left off and return to the OS

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 100% of the total text.

Page 1 of 1

Method for monitoring multiple systems in a finite time frame

For starters code must be organized in sections or chapters. Each chapter is a block of code that can be executed without a need for another chapter. A list of chapters is kept so RTAS (run time abstraction service) will know which chapter it left off with. RTAS does this so it executes all the chapters before starting on the first chapter again.

1. RTAS enters the monitoring code and records the time
2. RTAS calls the a function / chapter
3. The chapter returns to the RTAS call
4. RTAS checks if has completed all the chapters.
5. If all the chapters have been executed then RTAS simply returns to the OS. Otherwise go to step
6. RTAS determines if the time has expired
7. If the time has not expired RTAS repeats steps 2 - 6 otherwise go to step 8.
8. RTAS records the last chapter it has completed and returns to the OS.

The primary advantage here is that RTAS is limited to execution time. Set the bar. RTAS periodically checks to see if it has exceeded a prescribed amount of time. If it has, it returns and picks up where it left off next time. This could be used to help optimize performace numbers (although RTAS might be turned off for those cases).

This information can be used to target poorly coded sections of the firmware. That is, one would know where the worst offenders were and could rewrite, chapterize those.

Disclosed by International Business Machines Corporation

1