DE10000571A1 - Non-real time operating system extension for automation, metrology applications, involves assigning free high priority system clock source, to real-time program - Google Patents

Abstract

The existing non-real time operating system is separated from its high priority system clock source program, technically and is assigned to lower priority system clock source. The currently free high priority system clock source, is assigned for real-time program. An independent claim is also included for external clock generator.

Description

The invention describes a method for expanding a non-real-time operating system with a real-time capable program, based on 2 independent timers.

Real-time capable operating systems are characterized by a deterministic time behavior regarding the Execution times (e.g. precise activation of signals) and reaction times (latency times, e.g. B. from the arrival of a sensor signal to program processing). Furthermore have real-time capable multitasking operating systems a deterministic time behavior of their Program cycle time (task switchover time).

Operating systems that are not real-time capable have no, or only a weak, pronounced one deterministic time behavior. The program turnaround time, as well as latency times can be strong vary until completely stopped. However, they are usually characterized by good ones Visualization options and enable standardized operator guidance, as well as the Use of standardized development tools.

A combination of the properties is therefore useful, especially if the operating system for Control tasks in automation or measurement technology should be used.

Problem

The program turnaround time on a PC-based platform is determined by the system interrupt (IRQ0) which is generated by the system's own timer. Non-real-time capable multitasking Operating systems manage the round trip time of the individual programs with priority High-priority programs preferred over low-priority ones. In the worst case, therefore low-priority program no longer handled by the processor. A deterministic behavior can only the highest-priority program or the management algorithm of the operating system be attributed to yourself.

Known methods for implementing real-time capable programs in non-real-time capable Operating systems therefore call the real-time capable program immediately after the system Interrupts on (even before calling the management algorithm of the non-real-time capable Operating system) to ensure deterministic time behavior. One in the real time capable Program of additional integrated, distribution algorithm calls sequentially both the actual operating system management, as well as the real-time capable program. Through the sequential processing of the system interrupt remains the priority for the non-real-time capable Operating system and identical for the real-time capable program. This can have the consequence that the Operating system can block the deterministic process.

Approach

The approach according to the invention is based on the software-technical separation of the existing non-real-time operating system from its high-priority system clock source and its connection to a second, lower-priority clock source (here come as a second clock source both the internal real-time clock and an external clock source in question). The thus released high-priority system clock source is now used for the real-time capable program. Not that one- With this method, the real-time capable operating system is asynchronous from the real-time capable program decoupled. The clock sources are independent of one another in this method.

With this method, shorter cycle times can be achieved for the real-time program unlike other real-time operating system extensions, there is no additional Management algorithm within the real-time program is needed. Typical Circulation times of the real-time capable program below 10 us (micro seconds) can be achieved. The Circulation times of the real-time capable program depend on the performance characteristics of the Hardware platform freely adjustable.  

Another advantage of this method is the different hardware interrupt Priorities of the two clock sources. The high priority system management functions of the Operating system can now even by the highest priority hardware interrupt (clock source for the real-time capable program) and thus the deterministic time behavior of the real-time program can be maintained.

Use of the method in PC-based platforms

The process is particularly suitable for PC-based platforms. For PC-based platforms By default, the internal system interrupt (IRQ0, highest priority hardware interrupt) serves as a system Clock source for the existing non-real-time operating system. The not real-time capable Operating system is now decoupled from this interrupt and a new low-priority hardware Interrupt assigned. The interrupts IRQ3, IRQ4 (serial interface Interrupts) or the interrupt IRQ8 (real-time timer cascaded on IRQ2) in question. Will the Interrupts IRQ3 or IRQ4 are used as the system clock source for the non-real-time operating system this can be done by an external clock on the serial interface. However, will the interrupt IRQ8 used as a system clock source, there is no need for an external clock However, a management algorithm for the existing IRQ8 interrupt handling must be used be provided. The real-time capable program is now connected to the free system interrupt (IRQ0) attached and can run asynchronously to the non-real-time capable operating system.

System initialization process

The real-time capable program is installed at the runtime of the non-real-time operating system. The following sequence is followed:
(See Fig. 1 Operating system before installing the real-time extension)

• a) Deactivate all hardware interrupts
• b) Implementation of the non-real-time operating system on the low-priority interrupt (2nd clock Giver)
• c) coupling the real-time capable subsystem to the previous high-priority interrupt (system Clock generator)
• d) Setting the time base for the high-priority interrupt clock
• e) Setting the time base for the low-priority interrupt clock (if programmable)
• f) Activate all hardware interrupts

(Please refer

Fig.

2 operating system after installation of the real-time extension)

Claims (5)

1. A method for expanding a non-real-time capable operating system with a real-time capable program, characterized in that the existing non-real-time capable operating system is software-technically separated from its high-priority system clock source and a second, low-priority system clock -Source is assigned. The thus released high-priority system clock source is now used for the real-time capable program. 2. The method according to claim 1, characterized in that as the second clock source for the non- real-time program, both the internal real-time clock control and an external clock Source comes into question. 3. The method according to claim 1 and 2, characterized in that the clock sources are independent of each other. 4. The method according to claim 1, characterized in that the program cycle time of real-time capable program, depending on the performance characteristics of the hardware used Platform is freely adjustable. 5. Device for generating an external clock source, according to claim 2, characterized marked that interrupts can be triggered via the serial interface.