JP2008033577A - Device with multi-task scheduling function and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device with a multi-task scheduling function that is applied to an embedded device of which system resource is limited, and can reduce the load of the system resource. <P>SOLUTION: A task 2 is divided into a plurality of segment routines according to the execution break point described as an executable instruction. The device comprises a task reference information accumulation part 104 for accumulating task reference information comprising the time that can be occupied and is prescribed for each task, and the segment execution time associated with each segment routine, and a task switching part 103 for executing a task so that the total of the segment execution time of a plurality of segment routines to be executed is within the time that can be occupied by the task using the task reference information. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a multitask scheduling function loading apparatus and program. In particular, the present invention relates to a multitask scheduling technique in an embedded device equipped with a CPU (Central Processing Unit).

  2. Description of the Related Art Conventionally, an operating system (OS) executed by a CPU has a multitask scheduling function that schedules the execution order of a plurality of tasks and apparently processes the plurality of tasks simultaneously. A task is a unit of processing viewed from the operating system (viewed by the user). That is, the entire processing of one application is handled as one task. For example, it is a processing unit such as mail transmission / reception or file download.

  The multitask scheduling function divides the CPU processing time into very short units and sequentially assigns them to a plurality of applications. That is, tasks are switched at regular time intervals (time slices). Thereby, a plurality of applications are apparently processed simultaneously. There are mainly the following two approaches to the multitask scheduling function.

  As a first approach, for round-robin scheduling, there is a technique in which execution priority is set for each task, execution permission time of each task is determined based on the priority, and each task is executed in order (see, for example, Patent Document 1). .

  As a second approach, there is a technique for determining the priority of round robin scheduling based on the number of executions of tasks executed in the past (see, for example, Patent Document 2).

JP 2002-163117 A JP 2000-21068A

  According to conventional multitask scheduling, a series of processing of an application is forcibly switched regardless of processing contents. Therefore, it is necessary to perform processing for temporarily canceling the task, saving the execution state of the task in the memory, and recovering from the memory. In addition, since a plurality of tasks are switched at high speed and frequently, a lot of load is applied to system resources such as a CPU or a memory. Such processing is hardly a problem in a personal computer having a system resource with a high processing capability, but becomes a burden in an embedded device in which only a system resource with a low processing capability can be mounted.

  On the other hand, there is a technique in which each running application itself manages the processing time without the OS managing the CPU. That is, each application voluntarily releases the system resources occupied by itself in its own free time so that other applications can use it. In this case, the OS does not need to have a multitask scheduling function. However, since the application developer needs to always develop an application in consideration of the multitask scheduling function, the development cost and complexity are very high.

  Therefore, an object of the present invention is to provide a multitask scheduling function-equipped device and a program that can be applied to an embedded device with limited system resources and can reduce the load of the system resources.

According to the present invention, in a multitask scheduling function-equipped device that schedules the execution order of a plurality of tasks,
The task is divided into a plurality of segment routines according to execution breakpoints described as execution instructions.
Task reference information accumulating means for accumulating task reference information having an occupable time defined for each task and a segment execution time associated with each segment routine;
And a task switching means for executing the task so that the total of the segment execution times of the plurality of segment routines to be executed is within the occupying time of the task using the task reference information. .

  According to another embodiment of the multi-task scheduling function-equipped device of the present invention, the task switching means includes the predicted execution time obtained by adding the segment execution time of the next segment routine to the total execution time up to the current time in the task. It is also preferable to continuously execute the next segment routine when it is within the task occupying time.

  According to another embodiment of the multitask scheduling function-equipped device of the present invention, the task is divided into execution breakpoints by an object-oriented base class, and a segment routine that inherits the base class is realized by a derived class It is also preferred that

  According to another embodiment of the apparatus with a multitask scheduling function of the present invention, the segment execution time is preferably an average execution time of the task in the segment routine.

  According to another embodiment of the multitask scheduling function-equipped device of the present invention, it is preferable that the task reference information further includes a priority for each task, and the occupable time is determined according to the priority.

  According to another embodiment of the apparatus with a multitask scheduling function of the present invention, the task reference information is preferably managed based on the task identifier and the segment identifier.

According to the present invention, in a multitask scheduling program for causing a computer to function to schedule the execution order of a plurality of tasks,
The task is divided into a plurality of segment routines according to execution breakpoints described as execution instructions.
Task reference information accumulating means for accumulating task reference information having an occupable time defined for each task and a segment execution time associated with each segment routine;
Using the task reference information, the computer functions as task switching means for executing the task so that the total segment execution time of the plurality of segment routines to be executed is within the occupying time of the task. And

  According to another embodiment of the multitask scheduling program of the present invention, the task switching means is configured such that the predicted execution time obtained by adding the segment execution time of the next segment routine to the total execution time up to the current time in the task is It is also preferable to make the computer function so that the next segment routine is continuously executed when it becomes within the occupying time.

  According to another embodiment of the multitask scheduling program of the present invention, a task is divided into execution breakpoints by an object-oriented base class, and a segment routine that inherits the base class is realized by a derived class. It is also preferable.

  According to the apparatus and program with the multitask scheduling function of the present invention, it is possible to reduce the load of system resources even in an embedded apparatus with limited system resources. A task is divided into a plurality of segment routines, and a segment routine that is a series of processing units is processed without stopping task switching, so that the load of system resources at the time of task switching can be reduced.

  Hereinafter, the best mode for carrying out the present invention will be described in detail with reference to the drawings.

  FIG. 1 is a software function configuration diagram of a multitask scheduling function loading apparatus according to the present invention.

  The apparatus of FIG. 1 has an operating system 1 and a plurality of tasks 2. These functional units are usually realized by a program executed by a computer.

  Task 2 basically includes an initialization unit, an execution unit, and an end unit. The initialization unit performs processing for recovering the execution state of the task from the memory. On the other hand, the end unit performs processing for storing the execution state of the task in the memory.

  According to the present invention, the execution unit is composed of a plurality of segment routines. The plurality of segment routines are divided according to execution breakpoints described as execution instructions.

  As a first method of designating an execution breakpoint, a code “Do Something;” defined in advance as an execution breakpoint is described in the application program. Normally, the application developer is not conscious of the OS scheduling function at all, but according to the present invention, it is necessary to specify task switching permission.

  Object-oriented inheritance is used as a second method for specifying execution breakpoints.

  FIG. 2 is a description of classes used in the tasks in the present invention.

  According to FIG. 2, the execution breakpoint is divided by the base class of the task, and a segment routine that inherits the base class is realized by the derived class. According to FIG. 2, the segment routine is defined by “OnSegmentRoute1 ()”, “OnSegmentRoute2 ()”... “OnSegmentRouteN ()”.

When the task is a mail transmission application (SMTP transmission), it is divided into, for example, the following five segment routines.
(1) Socket connection (2) Authentication (3) Encoding of transmission / reception character string (4) Transmission (5) Socket disconnection

  Application developers need not be aware of execution breakpoints by implementing a derived class that inherits the base class in the present invention as one task. The multitask scheduling unit 10 can switch tasks for each segment routine implemented by an application developer.

  A task is divided into segment routines which are a series of processing units. According to the present invention, task switching does not occur during segment routine processing. In other words, task switching does not occur when a large amount of memory is consumed, for example, as in loop processing. This leads to a reduction in the load of system resources for saving the execution state in the memory at the time of task switching and for recovering the execution state from the memory. In addition, unlike the round robin scheduling, the OS does not need to control the task switching timing, and the processing amount of the entire system is reduced. Of course, the division into segment routines does not depend on specific hardware or OS.

  The multitask scheduling unit 10 in FIG. 1 includes a task activation unit 101, a task switching unit 102, a timer 103, and a task reference information storage unit 104. These functional units are usually realized as part of an operating system program executed by a computer.

  The task activation unit 101 activates a plurality of tasks as necessary. According to FIG. 1, a plurality of tasks 2A to 2C are activated. The task activation unit 101 notifies the task switching unit 102 of the activated task ID (IDentifier: identifier) and other task information.

  The task switching unit 102 manages time using an interrupt signal notified from the timer 103 at a constant period, and switches a plurality of activated tasks. Here, the task switching unit 102 executes the task by using the task reference information so that the sum of the segment execution times of the plurality of segment routines to be executed is within the occupying time of the task. Specifically, the task switching unit 102 determines the next execution time when the predicted execution time obtained by adding the segment execution time of the next segment routine to the total execution time up to the current time in the task is within the occupying time of the task. Continue the segment routine.

  The timer 103 outputs an interrupt signal with a fixed period to the task switching unit 102.

  The task reference information storage unit 104 stores task reference information. The task reference information has an occupable time (time slice) defined for each task and a segment execution time associated with each segment routine. The task reference information further includes a priority for each task.

The occupable time is defined according to the priority of the task (5 levels from 1 to 5). According to FIG. 1, 50 ms is defined as a time slice unit. The occupable time is determined by the following equation.
Occupiable time = Time slice unit (50 ms) x Priority (1-5)
As a result, task A with priority 3 has an occupancy time of 150 ms, task B with priority 1 has an occupancy time of 50 ms, and task C with priority 2 has an occupancy time of 100 ms.

  The segment execution time is the actual execution time of each segment routine and is average. According to FIG. 2, the execution time of segment routines 1, 2 and 3 of task A are 55 ms, 50 ms and 60 ms, respectively. The execution times of segment routines 1 and 2 for task B are 50 ms and 40 ms, respectively. The execution times of segment routines 1 and 2 for task C are 60 ms and 40 ms, respectively.

  FIG. 3 is an operation explanatory diagram of the task switching unit.

  A description will be given using the task reference information shown in FIG. 1 along the sequence of FIG. Note that the total execution time is cleared (= 0 ms) each time the task is switched.

(S301) First, assume that task A is selected. The task switching unit executes the segment routine 1 of task A.

  At this time, the task switching unit measures the current execution time of the segment routine 1 of task A. Assume that the current execution time is 50 ms. Then, the total execution time of the currently executing task A is 50 ms (= 0 ms + 50 ms).

(S302) Next, an attempt is made to execute the segment routine 2 of task A. The average execution time of the segment routine 2 of task A is 50 ms. Since the total execution time is 50 ms, the predicted execution time is 100 ms (= 50 ms + 50 ms). Since the predicted execution time 100 ms is within 150 ms of the occupying time of task A, the task switching unit executes the segment routine 2 of task A.

  At this time, the task switching unit measures the current execution time of the segment routine 2 of task A. Assume that the current execution time is 40 ms. Then, the total execution time of the currently executing task A is 90 ms (= 50 ms + 40 ms).

(S303) Next, an attempt is made to execute segment routine 3 of task A. The average execution time of the segment routine 3 of task A is 60 ms. Since the total execution time is 90 ms, the predicted execution time is 150 ms (= 90 ms + 60 ms). Since the predicted execution time 150 ms is within 150 ms of the occupying time of task A, the task switching unit executes the segment routine 3 of task A.

  At this time, the task switching unit measures the current execution time of the segment routine 3 of task A. Assume that the current execution time is 70 ms. Then, the total execution time of the currently executing task A is 160 ms (= 90 ms + 70 ms). Here, since the occupation time of task A exceeds 150 ms, the task switching unit switches to a task other than task A.

(S304) Next, assume that task B is selected. The task switching unit executes the segment routine 1 of task B.

  At this time, the task switching unit measures the current execution time of the segment routine 1 of task B. Assume that the current execution time is 30 ms. Then, the total execution time of the currently executing task B is 30 ms (= 0 ms + 30 ms).

  Next, execution of segment routine 2 of task B is attempted. The average execution time of the segment routine 2 of task B is 40 ms. Since the total execution time is 30 ms, the predicted execution time is 70 ms (= 30 ms + 40 ms). Since the estimated execution time of 70 ms exceeds the occupying time of task B, which is 50 ms, the task switching unit switches to a task other than task B.

(S305) Next, assume that task C is selected. The task switching unit executes the segment routine 1 of task C.

  At this time, the task switching unit measures the current execution time of the segment routine 1 of task C. Assume that the current execution time is 50 ms. Then, the total execution time of the currently executing task C is 50 ms (= 0 ms + 50 ms).

(S306) Next, an attempt is made to execute the segment routine 2 of task C. The average execution time of the segment routine 2 of task C is 40 ms. Since the total execution time is 50 ms, the predicted execution time is 90 ms (= 50 ms + 40 ms). Since the predicted execution time 90 ms is within 100 ms of the occupying time of task B, the task switching unit executes the segment routine 2 of task C.

  At this time, the task switching unit measures the current execution time of the segment routine 2 of task C. Assume that the current execution time is 50 ms. Then, the total execution time of the currently executing task C is 100 ms (= 50 ms + 50 ms).

  FIG. 4 is a flowchart of the task switching unit in the present invention.

(S401) It is determined whether there is a task waiting to be executed. If it does not exist, the process ends.
(S402) When there is a task waiting for execution, the task to be executed is selected.
(S403) The total execution time of the currently executing task is cleared (= 0 ms).

(S404) One segment routine is executed for the currently executing task. At this time, the execution time of the segment routine is measured. For example, measurement can be performed by counting the number of times by an interrupt signal having a fixed period output from a timer.
(S405) The average execution time of the segment routine of the task reference information is updated. For the segment routine executed for the first time, its execution time is registered as it is in the task reference information. For a segment routine that has been executed twice or more, for example, the average execution time is calculated by the following equation by including the number of executions in the task reference information.
Updated average execution time =
{(Existing average execution time × execution count) + measured execution time} / (execution count + 1)
The calculated average execution time is updated to the task reference information.
(S406) The execution time of the executed segment routine is added to the total execution time of the currently executing task to update the total execution time.

(S407) It is determined whether the total execution time is within the occupying time of the currently executing task. If the total execution time exceeds the occupying time of the currently executing task, the process proceeds to S401 to switch to another task.
(S408) If the total execution time is within the occupying time of the currently executing task, it is determined whether or not the predicted execution time is within the occupying time of the currently executing task. The predicted execution time is a time obtained by adding the next segment execution time to the total execution time. If the total execution time exceeds the occupying time of the currently executing task, the process proceeds to S401 to switch to another task.
(S409) If the predicted execution time is within the occupying time of the currently executing task, the next segment routine is designated. Then, the process proceeds to S404, and the next segment routine is executed.

  As described above in detail, according to the multitask scheduling function-equipped device and program of the present invention, it is possible to reduce the load of system resources even in an embedded device with limited system resources.

  A task is divided into segment routines which are a series of processing units. According to the present invention, task switching does not occur during segment routine processing. In other words, task switching does not occur when a large amount of memory is consumed, for example, as in loop processing. This leads to a reduction in the load of system resources for saving the execution state in the memory at the time of task switching and for recovering the execution state from the memory. In addition, unlike the round robin scheduling, the OS does not need to control the task switching timing, and the processing amount of the entire system is reduced. Of course, the division into segment routines does not depend on specific hardware or OS.

  Further, according to the present invention, since each task collectively executes a plurality of segment routines that can be executed within the occupiable time, it is possible to reduce the use of system resources associated with unnecessary task switching.

  Furthermore, according to the present invention, since the execution breakpoint is divided by the base class in the object orientation, the application developer is aware of the execution breakpoint by implementing the process in the segment routine of the derived class. There is no need to do.

  According to the various embodiments of the present invention described above, those skilled in the art can easily make various changes, modifications and omissions within the scope of the technical idea and the viewpoint of the present invention. The above description is merely an example, and is not intended to be restrictive. The invention is limited only as defined in the following claims and the equivalents thereto.

It is a software function block diagram of the multitask scheduling function loading apparatus in this invention. It is a description of a class used in a task in the present invention. It is operation | movement explanatory drawing of a task switching part. It is a flowchart of the task switching part in this invention.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Operating system 10 Multitask scheduling part 101 Task starting part 102 Task switching part 103 Timer 104 Task reference information storage part 2 Task

Claims (9)

In the multitask scheduling function-equipped device that schedules the execution order of multiple tasks,
The task is divided into a plurality of segment routines according to execution breakpoints described as execution instructions,
Task reference information storage means for storing task reference information having occupable time defined for each task and segment execution time associated with each segment routine;
Using the task reference information, the task switching means for executing the task so that the sum of the segment execution times of the plurality of segment routines to be executed is within the occupying time of the task. A device with a multitask scheduling function.   The task switching means continues the next segment routine when the predicted execution time obtained by adding the segment execution time of the next segment routine to the total execution time of the task up to the present time is within the occupying time of the task. The multitask scheduling function-equipped device according to claim 1, wherein the multitask scheduling function-equipped device is executed.   3. The task according to claim 1, wherein the execution breakpoint is divided by a base class in an object orientation, and a segment routine that inherits the base class is realized by a derived class. Multitask scheduling device.   4. The multitask scheduling function-equipped device according to claim 1, wherein the segment execution time is an average execution time of the task in the segment routine. The task reference information further includes a priority for each task,
The multitask scheduling function-equipped device according to any one of claims 1 to 4, wherein the occupable time is determined according to the priority.   The multitask scheduling function-equipped device according to any one of claims 1 to 5, wherein the task reference information is managed based on a task identifier and a segment identifier. In a multitask scheduling program that causes a computer to function to schedule the execution order of a plurality of tasks,
The task is divided into a plurality of segment routines according to execution breakpoints described as execution instructions,
Task reference information storage means for storing task reference information having occupable time defined for each task and segment execution time associated with each segment routine;
Using the task reference information, cause the computer to function as the task switching means for executing the task so that the sum of the segment execution times of the plurality of segment routines to be executed is within the occupying time of the task. A multitask scheduling program characterized by that.   The task switching means continues the next segment routine when the predicted execution time obtained by adding the segment execution time of the next segment routine to the total execution time of the task up to the present time is within the occupying time of the task. The multitask scheduling program according to claim 7, wherein the computer is caused to function so as to be executed. 9. The task according to claim 7, wherein the execution breakpoint is divided by a base class in an object orientation, and a segment routine that inherits the base class is realized by a derived class. Multitask scheduling program.

Images