PanelSoft

User Interfaces and Usability for Embedded Systems

Robust Software for Embedded Systems

2 Day Course

Embedded software engineers employ design patterns to ensure that their software is suitable for industrial and consumer use. Failure detection and recovery is vital and the software must be designed so that it can run continuously without the need for a reboot.

These techniques include the use of watchdog hardware to detect a deadlocked system. Error handling mechanisms should be capable of reporting the exact location of a software bug, even after the system has been deployed. Memory management techniques need to ensure that that a memory leak or stack overflow will not lurk like a timebomb in a long-running system. Programmers need to be aware of real-time pitfalls when using interrupts and an RTOS.

This course is ideal for engineers who can write software, but are new to embedded systems, especially if those embedded systems have safety concerns or require high reliability.

The course is based on ten years of experience writing embedded software for life-critical medical equipment. The course has been developed by Niall Murphy, author of the popular Murphy's Law column in Embedded Systems Programming magazine. The course is available on-site and is also part of our public training schedule. We are always happy to discuss course customisation.

Related magazine articles, authored by the course presenter,can be found at Robust Programming Articles.

Audience

This course is ideal for engineers who can write software, but are new to embedded systems.

Content

Introduction

• Definitions for Safety, Availability, Reliability
• Graceful degradation
• Industry magic bullets

Data Integrity Checks

• Checksums
• CRCs

Persistent Storage

• Error Detection
  
• Double Buffering
  
• Versioning

Built-in Self-test

• RAM/ROM/address line tests
  
• Loopback tests
• Cable and subassembly tests
  

Interlocks

• Hardware limitations on software actions
  
• Software limitations on user actions

Exception Handling and Asserts

• Language Support
  
• Assert Macro
  
• Debug code issues

Memory Management

• Static Allocation
  
• Stack Measurement
  
• malloc and fragmentation
  
• Pools
  
• Detecting Leaks Interrupts

Interrupts

• Reentrancy
  
• Motivation
  
• Parallelism
  
• Timed Polling

Real Time Operating System

• Tasking
  
• Decomposition
  
• Preemptive Kernels
  
• Priority Inversion
  
• Queuing
• Timing Accuracy

Watchdogs

• Goals
  
• Response to failure
  
• Timeout Length
  
• Self Test
  
• Multi-tasking

Mathematical Dangers

• Fixed Point Math
  
• Using spreadsheets for testing
  
• Floating point equality

Control Loop

• Monitoring vs. Control
  
• Calibration

Static Analysis Techniques

• lint
  
• Hazard Analysis
  
• Failure Mode Effects Analysis
  
• Design and code inspections