JP2693916B2 - Task scheduling method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a task scheduling method, and more particularly, to a real-time operating system (hereinafter referred to as RTO) which manages task operations according to priority.
S)).

[0002]

2. Description of the Related Art First, an RTOS will be described. One of the major significances of introducing a multitask system on a control system is the division of work. When a control system is constructed with a single task, the result is that the performance may not improve or a difficult program with poor maintainability may be formed, depending on the size of the system function or the complexity of input / output. Becomes As a measure to deal with this,
If the functions of the entire system can be subdivided into tasks that are small units of work and easily realized without impairing the real-time property of the control system, it is possible to expect improvement in productivity and maintainability. A system in which a plurality of tasks are operating simultaneously (while distributing the CPU execution right or system resources at a certain timing) is called a multitask system, and its control mechanism is called a multitask OS. Each task in the multi-task system can be considered as an independent program. A system in which the multi-task system is provided with real-time capability and can synchronize tasks or with the outside of the system in real time is a real-time system, and its control mechanism itself is an RTOS.

The relationship between the RTOS and the user program will be described below with reference to the relationship diagram between the RTOS and the application shown in FIG. FIG. 6 shows data in the storage unit in the conventional task scheduling method. In FIG. 6, in the user program group 601, from the system call group 604 provided,
By selecting the required system call and calling the accompanying interface library 602, the RTOS
It becomes possible to use the function of 603. Depending on the RTOS function, the contents of the RTOS management data 606 managed by the RTOS may be changed,
Further, the task scheduler 605 may be called in order to change the task execution right to another task. The system call group 604 includes system calls related to task management, inter-task synchronous communication, and interrupt management. The RTOS management data includes a management area (TCB: task control block) 607 that corresponds to each task in order to manage each task, and a task identifier exists to identify this. When a system call is used to operate a task, a task identifier is specified to specify which task to operate.

Next, states and state transitions that tasks can take in RTOS will be described with reference to the task transition diagram of FIG. In this case, the possible states of the task can be classified into the following four states. (1) Execution (RUN) state 701 The CPU utilization right has been given and is currently in execution,
Only one task can exist in the system at any given time. (2) Executable (READY) state 702 Although all the elements required to enter the execution state are available, it is possible to execute because a task with a higher priority (may be the same) than that task is being executed. The state that is not possible. That is, it can be executed any time when the CPU becomes available. (3) WAIT state 703 A state in which the task cannot be executed because conditions for executing the task are not satisfied. For example, the state of waiting for completion of I / O operation or establishment of synchronization between tasks. (4) Dormant state 704 A state in which the task has not been started yet, or a state after the task ends. When a new task is created, this state is entered first.

Each of these state transitions has the following meaning. (1) Activation [S71] A task in the DORMANT state is set to the READY state. (2) Termination [S72] The task in the RUN state is set to the DORMANT state. (3) DISPATCH [S73] Among the tasks in the READY state, it is called the task which should be shifted to the RUN state next and puts it in the RUN state. The task in the RUN state has the highest priority among the tasks in the READY state. Usually, this processing is called scheduling (or dispatch), and this processing unit is called scheduler (or dispatcher). (4) PREEMPT [S74] The task with a higher priority than the task currently in progress is RU.
When transiting to the N state, the currently executing task is once suspended and returns to the READY state. A task that has once moved to the READY state cannot move to the RUN state until it is selected again by the scheduler. (5) Wait condition [S75] When it is necessary to wait for an event occurrence, the task shifts from the RUN state to the WAIT state. (6) Wait release [S76] It means that a task in the WAIT state is set to the READY state due to the occurrence of an event from the system that releases the cause of the wait. (7) Forced termination [S77] It means that a task other than the RUN state is placed in the DORMANT state (forced termination). Immediateness is one of the important performance indicators of RTOS, and the interrupt prohibition time is the bottleneck of the immediacy in conventional RTOS.

If an interrupt occurs during the task scheduling of the RTOS and the RTOS function is called from the interrupt processing, the task with a higher priority than the task selected by the scheduler may be ready to be executed. That is, a task having a higher priority than the task in the running state exists in the runnable state. This causes a fatal inconvenience that a task having a high priority (high urgency) is not executed for the RTOS. To prevent this situation,
Processing is performed by the task scheduler in the interrupt disabled state.

Here, a series of flow of the task schedule will be described with reference to the conventional system configuration diagram of FIG. 5, the data configuration diagram in the storage section of FIG. 6, and the flowchart of the task scheduler of the conventional example of FIG. To do.

FIG. 5 is a system configuration diagram of a conventional example,
Data writing means 501, data reading means 502
And a control unit 500 including the data comparison unit 503, and a storage unit 504. 5, the storage unit 504 stores the user program group 60 shown in FIG.
1, RTO including interface library 602, system call group 604 and task scheduler 605
S603, RT including TCB 607 including task priority storage area 608 and task status storage area 609
OS management data 606 is stored. The tasks here are assigned priorities due to the differences in processing, and the fact that there are the states in which the transition is possible are paused, waited, executable, and executed are as described above. In the discussion herein, it is assumed that there are eight levels of priority. The priority of each task and the task status are stored in the task priority storage area 608 and the task status storage area 6 in the TCB 607 corresponding to the task.
09, respectively.

In the task scheduler, as its processing procedure is shown in the flowchart of FIG. 8, in the processing step 801 for selecting a task, the task to be put into the next execution state is selected from the executable tasks. . At that time, processing step 804 and processing step 80
5, processing step 806 and processing step 807, priority levels 1, 2, ...
And whether the tasks of 8 can be executed,
If it can be executed, the process proceeds to step 802,
When it is not possible, the process finally moves to the processing step 808, the interrupt permission processing is performed, and the processing step 8 is performed.
In step 09, HALT processing is performed, and the process returns to the beginning of processing step 801 again. That is, it is determined from the tasks with the highest priority whether or not they can be sequentially executed, and the task first found in the process of searching for this task becomes the selected task. When the execution task is selected, in the selected task state change processing step 802, the storage unit 504 is placed in order to put the selected task into the execution state.
In the TCB 607, the task priority order, task state, etc. are changed by the data writing means 501 of the control unit 500, and then, in the selected task register restoration processing step 803, register information necessary for executing the selected task. Etc., the data read means 502 from the storage unit 504 is used to restore the registers existing in the control unit 500, and control is passed to the selected task.

In the above schedule processing, since the processing is performed in the interrupt disabled state, no interrupt processing is accepted. Even if this interrupt processing request comes in during the schedule processing, the processing is performed when the interrupt disabled state is released after the end of the schedule processing. As described above, by making the interrupt disabled state, the immediacy for an interrupt processing request is inferior, but an interrupt occurs at the RTOS task schedule described above, and a task with a higher priority than the task in the execution state can be executed. A fatal inconvenience that a task having a high priority is not executed for the RTOS is prevented.

Although the RTOS has been described above, an example of this RTOS is described in Japanese Patent Laid-Open No. 1-316831. The RTOS proposed in Japanese Patent Laid-Open No. 1-316831 includes a central processing unit (CPU), an input / output processor (IOP),
This is a method in which an arithmetic processing unit (APU) is integrated and the priority order of function execution is selected based on the priority order. Although there are slight differences in the system call or interface library and the task status, the basic idea of the RTOS is the same as that of the RTOS described in this case. However, Japanese Patent Laid-Open No. 1-316831 does not mention the interrupt disabled state of the task scheduler, and does not consider the improvement of the immediacy of the system focusing on the task scheduler.

[0012]

In the above-described conventional task scheduling method, processing is performed in the interrupt disabled state from when the task is selected by the task scheduler until control is passed to the task. Since this is done, the longer this process takes, the more the system immediacy is lost.

It is an object of the present invention to reduce the interrupt prohibition time in response to the situation where the interrupt prohibition time in the above task schedule becomes the maximum interrupt prohibition time of the RTOS. To provide a scheduling method.

[0014]

Task scheduling method of the present invention SUMMARY OF] is a task schedule method applied to operation of a plurality of tasks in real-time operating system that manages the priority information on priority of task task priority storage area for holding the task of holding the information about the wait release task priority storage area and task scheduler processing status holding information task identifier storage area for holding information about the task identifier relates wait release task priority A storage unit including at least a scheduler processing status display storage area, a means for changing the information held in the task priority storage area and the waiting release task priority storage area, and a means for changing the information held in the task identifier storage area , task scheduler processing Corresponding to the unit and the task priority comparing means to change the information held in the status display memory area in the system including at least includes a control unit, at the time the task scheduler processing, the task scheduler processing status storage area varying means Accordingly, the task scheduler processing state display memory area stores information indicating that the task scheduler is being processed, means for changing the task priority storage area as well as the task identifier storage area
The task priority storage area and task
In the disk identifier storage area, the task
Holding the priority and identifier, and the interrupt enable state after the interrupt request in the interrupt enabled state
Occurs and the system call that releases the wait for the task is called.
If this is the case during the processing of this system call,
Waiting for the priority of the task to release the wait
The task scheduler processing status is retained in the priority storage area.
Status display storage area information indicates that it is being processed
If the task priority storage area priority and wait
Disk priority storage area priority and task priority ratio
The comparison means is used to compare, and the waiting task priority storage area
If the held priority is higher, the task knowledge
It is retained in the task identifier storage area using
The existing identifier to the identifier of the task to release the wait
Then, return from the interrupt process to the task scheduler process.
When this happens , the interrupt prohibition state is set, and the processing execution right is transferred to the task held in the task identifier storage area.

[0015] It should be noted that the Task Scheduler process described above,
Check the priority and status of multiple tasks and set them to the running status
Process to select the task to be transitioned and among multiple tasks
If there is no task to transition to the running state in
Includes processing to enter the permission state .

[0016]

Next, the present invention will be described with reference to the drawings.

FIG. 1 is a system configuration diagram applied to one embodiment of the present invention. As shown in FIG. 1, in this embodiment, at least a task priority storage area 101, a task identifier storage area 102, a wait release task priority storage area 103, and a task scheduler processing state display area 1
Storage unit 11 including 04, and at least data comparison unit 1
05, a data writing unit 106, and a control unit 12 including a data reading unit 107. The data comparing unit 105 includes at least a task priority order comparing unit 108. Is at least task priority storage area changing means 109, task identifier storage area changing means 110, and task scheduler processing state display area changing means 111.
It is included.

2 is a flow chart showing the processing procedure of the task scheduler of this embodiment, and FIG.
9 is a flowchart showing a processing procedure of a system call which is executed by calling a new interrupt request when the interrupt is enabled in the task scheduler process. FIG. 4 is a diagram showing the task ready queue. Hereinafter, with reference to FIG. 1, FIG. 2, FIG. 3 and FIG. 4, a concrete processing procedure of the task scheduler and the system call of this embodiment will be described. The storage unit 11 shown in FIG. 1 includes the elements shown in FIG. In addition, in this embodiment, three cases are considered as tasks. The states of the respective tasks are as shown in FIG. 4, and the task A 402 having the priority 1 is connected to the wait queue 401 to wait for some event, and the task B 412 having the priority 2 and the task B 412 C413 is executed
Since it is in the enabled state, it is assumed that the queue is connected to the queue of priority 2.

In FIG. 2, when the task scheduler processing is started, first, in processing step 201,
The task scheduler processing state display area changing unit 111 changes the information in the task scheduler processing state display area 104 existing in the storage unit 11 to information indicating that the task scheduler processing is being performed. Next, in process step 202, the task with the highest priority is selected from the tasks in the READY state. In processing step 202 for selecting this task, in processing step 210, processing step 211, ..., Processing step 212 and processing step 213,
It is determined whether or not the tasks having the priorities 1, 2, ..., 7 and 8 are executable, and if they are executable, the process proceeds to step 203. Further, if it is not feasible, it proceeds to the final process step 214, the processing of the interrupt enable is in the process step 214 is performed, the HALT processing is performed in processing step 21 5, again the process step 201 Return to. That is,
It is determined from the tasks with higher priority whether or not they can be sequentially executed, and the task first found in the process of searching for this task becomes the selected task. In the present embodiment, as an example, the task scheduler process selects the task B having the highest priority among the tasks in the ready state.

When the execution task is selected in this way, processing step 2 is executed to put the selected task into the execution state.
03, the task priority storage area changing unit 109 included in the data writing unit 106 causes the storage unit 1
The task priority order stored in the task priority order storage area 101 within 1 is changed to the priority of the task selected in the processing step 202. At the same time,
By the task identifier storage area changing means 110, the task identifier stored in the task identifier storage area 102 is
It is changed to the identifier of the task selected in the processing step 202. In this embodiment, since the selected task is the task B, the priority and the identifier of the task B are the targets of change and are in the state of saving. Next, in processing step 204, interrupt processing is permitted. Ready interrupt permitted in this process step 204, the following process when an interrupt request is generated at the timing before the process proceeds to step 20 5, or the process interrupt occurs during the previously made to the interrupt enabled state processing in step 204 If so, the interrupt process is performed.

Here, a case will be described in which the interrupt request is generated, interrupt processing is performed, and in the interrupt processing, a system call for canceling the waiting state of the task currently in the waiting state is issued.

In the interrupt processing, the system call processing (see the flowchart of FIG. 3) is called,
The waiting of task A in the waiting state is released. Then, in the system call processing, the priority order of the task designated in the processing step 301 of FIG. 3 is stored and held in the waiting release task priority order storage area 103 by using the task priority order storage area changing means 109. Then, after the waiting of the task A designated in the processing step 302 is released, in the processing step 303,
It is judged by the task scheduler whether or not it is being processed before the interrupt processing occurs. In the present embodiment, the information stored in the task scheduler processing status display area indicates that the task scheduler was processing before the interrupt processing occurred. Therefore, the processing procedure is as shown in FIG. Go to step 304. It should be noted that in the process step 303, if the process is not in progress, the process ends. Next, at process step 304, the task priority comparison means 108 causes the task priority order stored in the task priority order storage area 101 existing in the storage unit 11 and the wait release task priority order storage area 10 to be executed.
The task priorities held in 3 are compared. In this case, the task priority order stored in the task priority order storage area 101 is the processing step 202 in FIG.
The priority of the task selected in (1), and the task priority held in the wait release task priority storage area 103 is the task priority designated in the system call processing. When the task priority order held in the wait release task priority order storage area 103 is higher than the task priority order held in the task priority order storage area 101, the processing shifts to the processing step 305, otherwise. In this case, all the processing ends.

As shown in FIG. 4, if the task whose waiting is released by the system call has a higher priority than the task already selected by the scheduler (in this embodiment, the task scheduler The task selected by is the task B having the priority 2 and the task released from the waiting in the system call called from the interrupt process is the task A having the priority 1). The task identifier storage area changing unit 110 changes the task identifier held in the task identifier storage area 102 existing in the storage unit 11 to the identifier of the task whose waiting has been released by being designated by the above system call ( In this embodiment, the task identifier of the task identifier storage area is task B
Changed to task A). Then, the processing of the system call ends. In this case, when the priority of the task released from the system call is lower than that of the task selected by the scheduler, the task identifier storage area 102 existing in the storage unit 11 is stored in the storage unit 11. The system call process ends without changing the stored contents.

When the above system call processing is completed,
After that, the interrupt process ends, and the process returns to the task scheduler process. In the task scheduler process, the interrupt process disabled state is set in the processing step 205 of FIG. 2, and then, in the processing step 206, the task identifier stored in the task identifier storage area (the identifier of the task A in this embodiment). The task represented by is set as the selected task, and the information necessary for putting the selected task into the execution state is set. Then, in processing step 207, the status of the selected task is changed from the READY status to the RUN status, and in processing step 208, the task scheduler processing status display area changing means 111 displays the task scheduler processing status displayed in the storage unit 11. The information in the area 104 is changed to information indicating that the task scheduler is not processing. Next, in processing step 209, the information of the register necessary for the selected task to operate is restored, the processing of the task scheduler ends, and the process proceeds to the selected task execution processing.

Note that the task schedule is not being processed,
When the same interrupt request as the above-mentioned interrupt request is generated during the normal task processing and the same interrupt processing is performed, the system call is also executed during the interrupt processing (see the flowchart of FIG. 3). Is called, and the priority order of the task released from the wait state is changed in the processing step 301 by the task priority order storage area changing means 109 as in the above description, and the wait state is released in the processing step 302. In processing step 303, it is determined whether or not the task scheduler was processing before the interrupt processing occurred. In this case, the task scheduler processing status display area 104
Is an information source indicating that the scheduler is not being processed except when the scheduler is being processed.
It can be seen that the task scheduler was not processing before the interrupt processing occurred. If the task scheduler is not processing before the interrupt processing occurs, the system call processing ends.

[0026]

As described above, according to the present invention, the process of selecting the task with the highest priority from the tasks in the ready state and the setting of information for the task to transition to the running state By separating the processing that performs the task scheduling, the interrupt processing can be temporarily permitted during the task scheduler processing, which makes task scheduling, which is an important performance index of RTOS, a bottleneck of immediacy. It is possible to shorten the interruption prohibition time and improve the immediacy in the user system.

[Brief description of the drawings]

FIG. 1 is a system configuration diagram showing an embodiment of the present invention.

FIG. 2 is a diagram showing a flowchart of task scheduler processing of the present embodiment.

FIG. 3 is a diagram showing a flowchart of a system call process of this embodiment.

FIG. 4 is a diagram showing a task ready queue.

FIG. 5 is a system configuration diagram showing a conventional example.

FIG. 6 is a diagram showing a configuration in a storage unit.

FIG. 7 is a task transition diagram.

FIG. 8 is a diagram showing a flowchart of a task scheduler process of a conventional example.

FIG. 9 is a diagram showing a flowchart of a system call process of a conventional example.

[Explanation of symbols]

11, 504 Storage unit 12, 500 Control unit 101 Task priority storage area 102 Task identifier storage area 103 Task release task priority storage area 104 Task scheduler processing status display area 105, 503 Data comparison means 106, 501 Data writing means 107, 502 data reading means 108 task priority comparison means 109 task priority storage area changing means 110 task identifier storage area changing means 111 task scheduler processing state display area changing means 201-215, 301-305, 801-809, 9
01 processing step 401 waiting queue 402 task A priority 1 411 priority 2 queue 412 task B priority 2 413 task C priority 2 601 user program group 602 interface library 603 RTOS 604 system call group 605 task scheduler 606 RT OS management Data 607 TCB 608 Task priority storage area 609 Task status storage area 701 Execution status 702 Executable status 703 Waiting status 704 Dormant status

Claims (2)

(57) [Claims] 1. A task schedule method applied to operation of a plurality of tasks in real-time operating system that manages the priority, task priority order storage area for holding information about the priority of the task, to the task identifier task identifier storage area for holding information, and wait release task priority storage area and the data <br/> task scheduler process status storage area for holding information about the disk scheduler process status holding information about the wait release task priority and including at least a storage unit, said task priority storage area and means for changing the information stored in the wait release task priority storage area, means for changing the information stored in the task identifier storage area, the task scheduler Stored in the processing status display storage area Corresponding to the means and <br/> task priority comparing means to change the distribution system including at least includes a control unit, at the time the task scheduler process, by changing means of the task scheduler processing status storage area, wherein task scheduler process status storage area stores information indicating that the task scheduler is being processed, the changing means and changing means of said task identifier storage area of the task priority storage area, the task priority
The execution status is stored in the storage area and the task identifier storage area.
Holds the priority and identifier of the task to be transitioned to the state
And, an interrupt enable state after them, the interrupt enable shape
State, an interrupt request occurs and the task wait is released
If a system call to
Task that releases the wait while processing the system call
Holds the priority of each in the waiting task priority storage area
Information in the task scheduler processing status display storage area.
If the message indicates that the
Priority priority of storage area and priority of waiting task
Means for comparing the priority of the storage area with the task priority
, The waiting task priority storage area is saved.
If the priority held is higher, the task
It is stored in the task identifier storage area by using the identifier changing means.
Identifying the held identifier of the task that releases the wait
Change to child, interrupt process from the task scheduler
A task scheduling method, characterized in that when the process returns to the process, the interrupt-disabled state is set and the process execution right is transferred to the task held in the task identifier storage area. 2. The task scheduler process includes the plurality of processes.
Check the task priority and status of
Process of selecting a task to be transitioned and the plurality of tasks
If there is no task to transition to the execution state in
It is characterized by including processing to enable interrupts when
Task scheduling method of claim 1, wherein that.