Evaluation of Automatic Flow Analysis for WCET Calculation on Industrial Real-Time System Code

Abstract

A static Worst-Case Execution Time (WCET) analysis derives upper bounds for the execution times of programs. Such analysis requires information about the possible program flows. The current practice is to provide this information manually, which can be laborious and error-prone. An alternative is to derive this information through an automated flow analysis.In this article, we present a case study where an automatic flow analysis method was tested on industrial real-time system code. The same code was the subject of an earlier WCET case study, where it was analysed using manual annotations for the flow information. The purpose of the current study was to see to which extent the same flow information could be found automatically. The results show that for the most part this is indeed possible, and we could derive comparable WCET estimates using the automatically generated flow information. In addition, valuable insights were gained on what is needed to make flow analysis methods work on real production code.

Experiences from Adjusting Industrial Software for Worst-Case Execution Time Analysis

“…Specifically, for the OPC UA publisher, there is nearly no dependency on global data structures, contrarily as reported in [24] for embedded systems. The reason is that there are no global values that external program parts can change, and loops do not contain any break statements, nor are variables involved in the loop condition that are manipulated within the loop.…”

“…Also, the effort increases when adding annotations to create better estimates, especially for complex code. In [24], the authors tested an automatic flow analysis method on industrial real-time system code and achieved high accuracy results. However, all these case studies concern the analysis of binary code only.…”

Static Timing Analysis of OPC UA PubSub

“…A noteworthy fact is that neither the publisher nor the subscriber has dependencies on global data structures, especially as [30] reports such dependencies as quite common in industrial code. This circumstance occurs due to the absence of global values changeable by external program parts.…”

“…The authors achieved very high accuracy but also indicated a high complexity due to the required annotation work. In [30], the authors used automatic flow analysis on industrial real-time system code to reduce manual work. Most case studies analyze binary code, which makes the annotation more difficult.…”

Performance Evaluation of Real-Time Systems

“…The ALF tools like AlfBackend or Melmac are the examples. SWEET can handle complex 'C' features like pointers, unstructured code as well as http://journals.uob.edu.bh recursion [27]. It works by collecting the information about loop bounds and infeasible paths to find the safe and tight WCET for the program in consideration.…”