JPH01216432A - Multitask system - Google Patents

Abstract

PURPOSE:To enable a processing to be performed at least in a timer interval time in one time of execution of a task by providing a timer device dedicated for task scheduling and a timer reset device to reset the count value of the timer device. CONSTITUTION:In the title system, a task scheduling dedicated timer device 19 used for the task scheduling only and the timer reset device 20 to reset the count value of the timer device 19 are provided. And at the time of completing the processing of the task under execution and setting the next task at a running state, the timer reset device 20 is started up, and the count value of the timer device 19 dedicated for the task scheduling is reset. In such a way, since the timer interval time is always supplied as a processing time in one time of execution of the task, it is possible to remove a waiting time for the completion of another task in the task whose processing time is short, thereby, the task scheduling can be performed efficiently.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention is directed to one central processing unit (hereinafter referred to as rcPUJ).
The present invention relates to so-called multitasking systems in which multiple tasks are executed, and in particular to allocation of execution time to each task.

[Conventional technology]

FIG. 3 is a state diagram showing the states of tasks (programs with unique purposes) in a multitasking system. In the same figure, 1 is an unregistered state where the task does not exist;
2 is a registration state in which the task exists but has not been started yet; 3 is a registration state in which the task has been started and is ready to be executed at any time, but is kept waiting because other tasks are being executed. 4 is the state in which the task is currently running, 4 is the run state, 5 is the wait state in which the task is waiting for an event to occur, and 6 is the suspend state in which the task cannot be executed because it has received an interrupt request. .

Also, arrow Ll gives up the CPU, arrow L2 gets the CPU,
Arrow L3 indicates termination/deletion, arrow L4 indicates termination, arrow L5 indicates start, arrow L6 indicates generation, arrow L7 indicates continue, arrow L8 indicates suspend, arrow L9 indicates wake up, and arrow LIO indicates sleep.

In a multitasking system, a special program called a monitor is generally provided separately from the tasks in order to control the execution of multiple tasks.

FIG. 4 is a block diagram showing the structure of a conventional multitasking system in which a monitor program stored in a monitor and tasks in a memory section are shown in association with a general computer system. In the figure, 11 is a CPU, 12 is a memory section, 13 is a timer device t, 14 is a monitor, 15 is a system call processing section, 16 is a timer interrupt processing section, 1
7 is a task scheduler that performs a process of selecting tasks to be executed; 18 is an initialization program section that stores an initialization program 18a; A1 to An are n tasks alx
This is a memory corresponding to an.

Next, the operation will be explained mainly focusing on the operation of the monitor.

It should be noted that only those particularly related to the present invention will be explained, and explanations of the wait state 5 and suspension state 6 in FIG. 3 will be omitted. First, it is assumed that all n tasks al-an are in unregistered status B1. In addition, the following is defined as a system call.

i) [N5TL: Change task from unregistered state to registered state 2
Move to.

1i) RQST: Register task from state 2 to ready state a3
Move to.

Initially, the initialization program 18a operates. The initialization program 18a performs lN for n tasks a1 to an.
5TL system call.

In other words, all n tasks a1 to an are in the registration state 2.
becomes.

Next, the initialization program 18a makes an RQST system call to temporarily start task a1 (the monitor 14 usually starts one task, also called an initial task).

As a result, one task al is in the ready state 3, and n-1 tasks a2 to an are in the registered state 2.

Next, when the initialization program 18a ends, the task scheduler 17 is called. Since there is only one task in the ready state B3, the task scheduler 17 shifts it to the run state 4 and starts executing task a1.

Next, assume that an RQST system call is issued to the second task a2. Note that this system call is issued by the memory A1 storing the task al or by the task interrupt processing unit 16.

Here, task a1 is in run state 4 and task a2 is in ready state 3, but which task is to be executed depends on the processing of the task scheduler 17.

The above is the process from task registration to execution.
There are two types of scheduling methods of the task scheduler 17 described below: (1) and (2).

■Compare the priorities of ready tasks and execute the task with the highest priority.

■Execute a task until the next timer interrupt or until the processing of that task is completed, and then execute the next task.

■ is called the priority method, and ■ is called the round tropin method.

Next, task scheduling using the round-tropin method will be explained using FIGS. 5 to 7.

Figure 5 shows the tasks to be executed, and shows the tasks (
Assume that five tasks a) to (e) are prepared.

Figure 6 shows the processing time of each task shown in Figure 5. For each task, the time taken to complete execution of the task is determined by how many times the timer interrupt interval (hereinafter referred to as "timer interval"). represent.

For example, task (a) has a timer interval of 1.6
Process in twice the time.

The prerequisites for task scheduling a) 1) are shown below.

b) Tasks (a) to (e) are all ready B3.

口) The execution order of tasks is task (al→(e)).

FIG. 7 is a state diagram showing the execution state of a task, and the solid line (
The portion (horizontal line) indicates the task currently being executed, and the numerical value on the solid line indicates the time for which the task was executed (time when the timer interval is 1.0). Next, the execution state of a task will be explained using FIG. 7. First, the timer scheduler 17 puts task (a) in run state 4 (see FIG. 3). Since the processing time of task (a) is 1.6, a timer interrupt is required until the processing is completed. Therefore, the task scheduler 17 suspends the execution of task (a) and puts task (b), which is the next scheduled task, into run state 4 (state 21). Task (b
) is 0.8, so the processing ends before the next timer interrupt arrives. Next, the task scheduler 17 puts task (C) into a run state B4 (state 22
), even though the processing time of task (C) is 0.8, most of the timer interval was used in the execution of task (b), so a timer interrupt occurred during the execution of task (C1), and the task scheduler 17 task(
d) is placed in a run state (state 623). By continuing such operations, tasks (a) to tel are executed alternately, making it possible to execute multitask processing.

[Problem to be solved by the invention]

Since the conventional multitasking system is configured as described above, there are cases where less time than the timer interval value is allocated to one execution. For this reason, there is a problem in that even a task whose processing time is shorter than the timer interval may require several times the timer interval to complete the processing.

The present invention has been made in view of these points, and an object of the present invention is to provide a multitasking system that allows processing to continue for at least the timer interval during execution of one task.

[Means to solve the problem]

In order to achieve this purpose, the multitasking system according to the present invention includes, in addition to a normal timer device, a timer device dedicated to task scheduling, and a timer device that is used when the processing of the current task is completed and the next task is executed. The present invention is provided with a timer reset device that resets the count value of a timer device dedicated to task scheduling.

[Effect]

In the multitasking system according to the present invention, when the processing of the task being executed is completed and the next task is placed in a running state, the timer reset device is activated to reset the count value of the timer device dedicated to task scheduling.

〔Example〕

FIG. 1 is a block diagram showing an embodiment of a multitasking system according to the present invention. In the figure, 19 is a task scheduling timer device used exclusively for task scheduling, and 20 is a timer reset device for resetting the count value of the task scheduling timer device 19. In FIG. 1, the same or equivalent parts as in FIG. 4 are given the same reference numerals.

FIG. 2 is a state diagram showing the execution state of a task according to this embodiment, where the real 'ftIA (horizontal line) part indicates the task currently being executed, and the numerical value on the solid line indicates the time that the task was executed (timer interval). (time when 1.0)
represents.

Next, the operation will be explained using FIG. 2.

First, as a task to be executed, the task (a
) to task (e) shall be prepared,
The prerequisites for task scheduling are shown below.

■Tasks (a) to (e) are all ready B3.

(2) The execution order of tasks is task (a) → task (e).

First, the giggle providing scheduler 17 converts task (a) into a run B4. Since the processing time of task (a) is 1.6, a timer interrupt occurs before the processing is completed, and the task scheduler 17 interrupts the execution of task (a).
The next task to be executed, task (bl), is set to run state 4 (state 831), and the processing time of task (bl is 0.8, so the processing ends before the timer interrupt occurs. At this time, the task scheduler l7 puts the next execution task (task TC) into run state 4, and activates the timer reset device 2o to clear the count value of the task scheduling dedicated timer device 19 (at time T
I, state 32) 0 Task (C) will be executed next, but the task scheduling dedicated timer device 19
Since the count value of has been reset, it remains 0 for 18 executions.
．． 8 task (C1 also has timer interval time 1
．． Since 0 is given as the processing time, the processing of task (C) can be completed before the next timer interrupt. In addition, in FIG. 2, times T2 to T5 indicate times at which the reset is required.

In the manner described above, a multitasking system is realized.

In the above embodiment, the timer reset device 20 is used to clear the count value of the task scheduling dedicated timer device 19, but this process may be performed by the task scheduler 17.

〔Effect of the invention〕

As explained above, the multitasking system according to the present invention always gives the timer interval time as the processing time when executing a single task. Tasks are less likely to have to wait for other tasks to finish, resulting in more efficient task scheduling.

[Brief explanation of the drawing]

FIG. 1 is a configuration diagram showing an embodiment of a multitasking system according to the present invention, FIG. 2 is a state diagram showing the execution state of a task in the embodiment of FIG. 1, and FIG. 3 is a state diagram showing the state transition of a task. FIG. 4 is a configuration diagram showing a conventional multitasking system, FIG. 5 is a configuration diagram showing a general task configuration, FIG. 6 is an explanatory diagram showing the processing time of each task, and FIG. 7 is a state diagram. FIG. 2 is a state diagram showing the execution state of tasks in a conventional multitasking system. 11...CPU, 12...Memory part, 13...Timer device! , 14... Monitor, 15... System call processing section, 16... Timer interrupt processing section, 17.
...Task scheduler, 18...Initialization program section, 19...Task scheduling dedicated timer device, 20...Timer reset device, A1-An...
·memory.

Claims (1)

[Claims] It has a central processing unit that executes programs such as a plurality of tasks, a memory unit that stores programs such as a plurality of tasks, and a timer device that generates an interrupt to the central processing unit at fixed time intervals. A computer system comprising: a task scheduling dedicated timer device used exclusively for task scheduling; and a timer reset device resetting a count value of the task scheduling dedicated timer device, wherein the central processing unit receives an interrupt from the task scheduling dedicated timer device. If this occurs, one of the plurality of tasks is selected and the selected task is executed, and the timer reset device resets the count value of the task scheduling dedicated timer device upon execution of each task. A multitasking system that allocates the minimum timer interrupt interval time to each task.