Predictable Interrupt Scheduling with Low Overhead for Real-Time Kernels

Leyva-del-Foyo, Luis Eduardo; Mejía-Álvarez, Pedro; Niz, Dionisio de

doi:10.1109/rtcsa.2006.51

Cited by 14 publications

(12 citation statements)

References 13 publications

(25 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The other four columns of the figure show the different executable entities in the system for four different interrupt management schemes. The rightmost column represents our integrated model [Leyva-del-Foyo et al 2006b, 2006c. The other three middle columns represent different classes of interrupt model.…”

Section: Related Workmentioning

confidence: 99%

Integrated Task and Interrupt Management for Real-Time Systems

Leyva-del-Foyo

Mejía-Álvarez

Niz

2012

ACM Trans. Embed. Comput. Syst.

View full text Add to dashboard Cite

Real-time scheduling algorithms like RMA or EDF and their corresponding schedulability test have proven to be powerful tools for developing predictable real-time systems. However, the traditional interrupt management model presents multiple inconsistencies that break the assumptions of many of the real-time scheduling tests, diminishing its utility. In this article, we analyze these inconsistencies and present a model that resolves them by integrating interrupts and tasks in a single scheduling model. We then use the RMA theory to calculate the cost of the model and analyze the circumstances under which it can provide the most value. This model was implemented in a kernel module. The portability of the design of our module is discussed in terms of its independence from both the hardware and the kernel. We also discuss the implementation issues of the model over conventional PC hardware, along with its cost and novel optimizations for reducing the overhead. Finally, we present our experimental evaluation to show evidence of its temporal determinism and overhead.

show abstract

Section: Related Workmentioning

confidence: 99%

Integrated Task and Interrupt Management for Real-Time Systems

Leyva-del-Foyo

Mejía-Álvarez

Niz

2012

ACM Trans. Embed. Comput. Syst.

View full text Add to dashboard Cite

show abstract

“…To tackle this problem, Leyva-del-Foyo et al suggested to implement all tasks as threads, and to trigger them from very short interrupt handlers by posting a semaphore, for instance [5,6]. Furthermore, an interrupt-leveling mechanism implemented in software prevents any disturbance by lower-priority interrupt handlers by masking all corresponding interrupt sources.…”

Section: Rate-monotonic Priority Inversionmentioning

confidence: 99%

“…However, softwarebased approaches like the one presented in [5,6] suffer from the same problem. In those approaches, a small ISR is needed to notify the task related to the particular event-this ISR could completely utilize the CPU.…”

Section: Interrupt Overloadmentioning

confidence: 99%

Parallel, hardware-supported interrupt handling in an event-triggered real-time operating system

Scheler

Hofer

Oechslein

et al. 2009

Proceedings of the 2009 International Conference on Compilers, Architecture, and Synthesis for Embedded Systems

View full text Add to dashboard Cite

A common problem in event-triggered real-time systems is caused by low-priority tasks that are implemented as interrupt handlers interrupting and disturbing high-priority tasks that are implemented as threads. This problem is termed rate-monotonic priority inversion, and current software-based solutions are restricted in terms of more sophisticated scheduler features as demanded for instance by the AUTOSAR embedded-operating-system specification.We propose a hardware-based approach that makes use of a coprocessor to eliminate the potential priority inversion. By evaluating a prototypical implementation, we show that our approach both overcomes the restrictions of software approaches and introduces only a slight processing overhead in exchange for increased predictability.

show abstract

“…Katcher et al emphasize the importance of integrating timer interrupts into the scheduling scheme in order to avoid unnecessary interrupts, i.e., timer interrupt that will wake up a low priority task while a higher priority task is in execution. This integrated interrupt model has been further extended by Leyva-del-Foyo et al in [6][7][8] to include all hardware interrupts that can be produced in a real-time kernel. These authors introduce an integrated model for task and interrupt management and propose different hardware and software implementation approaches.…”

Section: Introductionmentioning

confidence: 99%

“…The integration of hardware interrupts and real-time tasks gives rise to a real-time task set composed by Hardware Activated Tasks and Software Activated Tasks respectively [8]. The release of Harware Activated Task is strictly controlled by an Interrupt Hardware Abstraction Layer in [7] or more loosely controlled using an optimistic approach in [8].…”

Section: Introductionmentioning

confidence: 99%

Integrated Schedulers for a Predictable Interrupt Management on Real-Time Kernels

Sáez

Crespo

2014

Reliable Software Technologies – Ada-Europe 2014

View full text Add to dashboard Cite

Abstract. To analyse the timeliness behaviour of a real-time system is one its key aspects. A big effort has been performed by the real-time community to develop accurate and more general schedulability analysis that can ensure the correct execution of the system. However, few works have analysed the side effects introduced by the scheduler and undesired execution of Interrupt Service Routines. Previous works addressed the interrupt interference by proposing an Integrated Interrupt Model that avoids unpredictable disturbance from external interrupts. Even so, the scheduling overhead due to the unnecessary activation of low priority tasks remains still unaddressed in this model. This work proposes a new Virtual implementation of an Integrated Interrupt Event-Driven Scheduler that copes with this pending issue. It also analyses the behaviour of the commonly used dual queue scheme under this kind of schedulers and proposes a more appropriated data structure to avoid unnecessary overheads.

show abstract

Predictable Interrupt Scheduling with Low Overhead for Real-Time Kernels

Abstract: In

Cited by 14 publications

References 13 publications

Integrated Task and Interrupt Management for Real-Time Systems

Integrated Task and Interrupt Management for Real-Time Systems

Parallel, hardware-supported interrupt handling in an event-triggered real-time operating system

Integrated Schedulers for a Predictable Interrupt Management on Real-Time Kernels

Contact Info

Product

Resources

About