Independent Audit Task for Operating System Fault Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Complex operating systems like LINUX in real-time computer applications often have undetected software errors and transient hardware faults, requiring rapid fault detection to ensure system safety, but existing methods are inadequate for timely and accurate identification during runtime.
Innovation Solution
Implementing an independent audit task that has read access to the control register of the computer before an application task starts, allowing for validation of the results from operating system calculation tasks to detect design errors or hardware faults, and enabling immediate emergency actions or delayed responses based on the detected faults.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If an independent audit task is implemented to validate control register contents, then fault detection capability is improved, but administrative overhead in task changes increases
Solution Approach 1:
The audit task is executed before the application task starts, validating control register contents in advance. This preliminary validation ensures that fault detection occurs before the main task execution, improving reliability without blocking the overall system flow. The audit task checks critical registers (PC, SP, status registers) before the application task begins, allowing early detection of potential faults.
Solution Approach 2:
The independent audit task acts as an intermediary between the operating system and the application task. It validates the control register contents that were set by the OS calculation task, serving as a buffer that separates the OS from the application while ensuring data integrity. This intermediary layer detects faults without requiring direct modification of the application task structure.
2Reliability
If validation of control register contents is performed, then system safety is improved, but implementation time within a frame increases
Solution Approach 1:
The audit task focuses validation on specific critical control registers (program counter, stack pointer, status registers) rather than performing a comprehensive check of all system state. This localized validation approach ensures system safety for critical operations while minimizing the time required compared to a full system validation. Only the most essential registers that could cause system failure are validated.
Solution Approach 2:
The audit task performs a partial validation of control register contents, checking only the essential elements needed for safe operation. Rather than validating every possible system state, the implementation focuses on the minimum necessary checks (PC, SP, status bits) that provide adequate safety assurance while keeping execution time within acceptable bounds for real-time operation.
3Measurement precision
If an independent audit task is carried out, then detection precision of faults is improved, but device complexity increases
Solution Approach 1:
The audit functionality is extracted as a separate independent task from the main application task and OS calculation task. This extraction allows the audit logic to be implemented as a distinct, focused validation routine that can precisely detect faults in control register contents without being entangled with the complexity of the main system. The separated audit task can use specialized validation logic tailored specifically for fault detection.
Data Source
AI summary
The invention relates to a method for detecting faults that occur or are present in an operating system of a computer, wherein an in particular independent audit task (106) is carried out during the run time before a starting time (102, 112) of the requested application task (107), wherein the control registers define the properties of the run time environment of the requested application task (107) and have reading access to the contents and validate these contents. Furthermore, the invention relates to a computer, on which such a method is carried out.