JPH0991154A - Stack allocating method and control equipment - Google Patents

Abstract

PROBLEM TO BE SOLVED: To always preferentially starts a high-order task without the exclusive control of a stack by providing one stack for every priority of the task and sharing one stack between the tasks having the equal priority. SOLUTION: A task managing table 10 is provided for each task and has a flag 104 for distinguishing whether the task is a newly activated task or an interrupted task in addition to priority 101 for tasks, start address showing the execution program head of the task and stack pointer 103 instructing the final part of the stack allocated to the task as contents. In this case, one stack is allocated for every priority of the task and one stack is shared between the tasks having the equal priority. An executing task selection program 12 investigates queues Q6 and Q7 provided for every priority so that the tasks can be extracted and executed successively from the task which has the highest priority and early activated.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a stack allocation method used in a multitasking operating system and a control device using the method, and particularly to a stack suitable for a multitasking operating system for a single-chip microcomputer having a small RAM capacity. The present invention relates to an allocation method and a control device using the method.

[0002]

2. Description of the Related Art FIG. 2 shows an example of a control system using a single-chip microcomputer and is an explanatory view of a control system of a hydraulic excavator. The single-chip microcomputer 200 installed in the hydraulic excavator 100 receives as input the pressure sensor of the cylinder attached to the hydraulic excavator, the tilt angle sensor of the swash plate pump, and the switch of the operation panel,
A target value for controlling the hydraulic control system or the machine control system is calculated, and the result is output to the pump or engine. These processes are performed by software, and a multitasking operating system (hereinafter referred to as a multitasking OS) is used to execute this software.

The tasks of this multitasking OS are configured as shown in FIG. That is, the task T1 with the highest priority P1
Reference numeral 1 denotes an error processing task, which performs processing for backing up an appropriate value when the various sensors malfunction or fail. For example, if there is an abnormality in the pressure sensor of the boom cylinder and the correct value cannot be input, set the input value to 0.
If it is a ~ 5V sensor, it is fixed to a value required for stable operation of the boom cylinder, for example, 2.5V which is an intermediate value, and the subsequent calculations are continued. Since this task is an urgent task, it has the highest priority.

The task of medium priority P2 is a task of inputting data from various sensors and issuing a command to each actuator. For example, task T21 is a task of inputting information and hydraulic oil temperature from a pressure sensor, T22 is a task of inputting a pump tilt angle and issuing a command to a governor for controlling the engine speed, and task T23 is a task of pressure sensor. This is a task that inputs information and issues a command to the regulator that controls the discharge flow rate of the pump.

The task of lower priority P3 is only a task of calculating a target value for controlling the hydraulic control system or the mechanical control system. For example, task T31 is a task for performing a calculation for controlling the discharge flow rate of the pump, T32 is a task for performing a calculation for controlling the traveling motor, and T33 is a task for performing a calculation for controlling the engine speed. is there. These tasks have lower priorities because they can be processed when tasks with higher priorities P1 and P2 are not being executed. The task group with the medium priority P2 is a task group that outputs the result calculated by the task group with the priority P3. Furthermore, in the case where each priority has a plurality of tasks (for example, P3 has three tasks and P2 also has three tasks), the tasks within that priority are T21 → T2.
It is executed sequentially as 2 → T23.

In such a multi-task OS, when an internal variable or task is suspended, a return address or register information necessary for restarting it, a stack area for storing a return address when a subroutine is called, etc. R
Prepare for AM. The present invention relates to such a stack area configuration method, but in a system having a sufficient memory, a stack area is fixedly assigned to each task to facilitate parallel processing of a plurality of tasks. ing. However, in a system such as a single-chip microcomputer that can use only a small memory, it is impossible to fixedly allocate the stack area to all of a large number of tasks.

In order to solve this problem, it is known that a recent multi-task OS for a single chip microcomputer shares a data area and a stack area in a RAM area. In a system in which this one area is shared by multiple tasks, when a task activation request is made, in the case of a single-chip microcomputer, the stack area is extremely small, so it is confirmed whether the stack area is in use. Exclusive control is performed to determine whether to execute the task for which the activation request was issued or place it in the waiting state.

As a method of sharing a stack between a plurality of tasks without performing exclusive control of the stack area, the stack area may include a portion for storing information for temporary suspension / resumption and a subroutine / interruption. A subroutine / interrupt information storage stack is shared among a plurality of tasks by dividing it into sections for storing return information.
There is a method as disclosed in No. 810.

[0009]

With the above-mentioned conventional technique,
The method of allocating a stack for each task causes an increase in space and cost associated with an increase in the size of the substrate of the device. In addition, construction machines and the like have a problem that they become vulnerable to vibration as the circuit becomes larger. Therefore, a single-chip microcomputer is used. However, if the stack area is shared by a plurality of tasks, it is necessary to provide the multitask OS with a function for managing the shared stack area, which leads to an increase in processing time at the time of task switching. Further, in a multi-task OS in which a task is provided with an execution priority and task switching is performed by this, when a stack area is shared by tasks having different priorities, a task with a higher priority is
It may happen that a task with a low priority is kept waiting because of waiting for resources.

This latter problem occurs as follows.
Now, for example, when the task T11 of higher priority is activated during execution of the task T22 of FIG.
The execution of task T11 is interrupted, the return address of the task T22 and the register information are stored in the stack area, and the task T11 is executed. Here, if the tasks T11 and T22 share the stack area, the execution of the high-order task cannot be started, the processing of the task T22 is restarted, and the task T11 waits for the end of the task T22. However, if there is an abnormality in the data input by the medium priority task or if there is an abnormality in the calculation result of the lower priority task, the highest priority task T11 is immediately executed and there is an abnormality. It is necessary to back up the sensor or correct the calculation result, but if the task T11 is executed later, the calculation result using the abnormal data is output to the actuator as it is, and the machine malfunctions. There is a risk of being connected. For example, the front does not operate according to the operator's operation.

In order to prevent such a problem, it is possible to allocate a task stack area to each task which is apparently not processed in parallel during software development. However, in order to allocate such a stack area, knowledge of the hardware of the microcomputer is required in addition to knowledge of the software, which increases the burden on the person in charge of software development. This negates the merit of improving the software development efficiency by introducing the multitasking OS. Further, the method of dividing the stack area into the pause / resume information storage portion and the subroutine / interrupt information storage portion described in Japanese Patent Laid-Open No. 5-173810 cannot share all the portions of the stack. Depending on the configuration, waste of the RAM area still remains.

An object of the present invention is to provide a multi-task OS which uses a small-capacity stack area and performs task switching according to priority so that a high-order task can always be activated with priority without stack exclusive control. An object is to provide an area allocation method and a control device using the method.

[0013]

The present invention is a stack allocation method in a multitasking operating system, in which one stack is provided for each priority of tasks, and tasks having the same priority share one stack. Disclosed is a stack allocating method.

Further, according to the present invention, a first task group having the same priority and a second task group having a different priority from the task group.
When the arbitrary task of the first task group and the arbitrary task of the second task group do not compete for the task group of No. 1, the stack corresponding to the first task group and the second task group
Disclosed is a stack allocating method, which is characterized in that a stack corresponding to the task group of 1 is shared to be one.

Further, according to the present invention, a queue is provided for each priority, a flag is provided for each task, and for a task interrupted by an interrupt, information necessary for restarting the stack corresponding to the priority of the task is provided. Store the task, connect the task to the head of the queue corresponding to its priority with the flag as a value indicating suspension, and for the task that has been activated, set the flag as a value indicating new activation and give priority to the task. When one task is selected as an executing task by checking the queue by connecting to the end of the corresponding queue, and when the flag of the executing task indicates interruption, the restart stored in the corresponding stack is resumed. When the flag of the selected execution task indicates new activation, the priority of the execution task is restarted by using the information necessary for Starts the processing from the start address the set by setting the start address of the execution task to the stack of response, to perform multitasking by assigning stack thus dynamically discloses stack allocation method comprising.

Furthermore, the present invention provides a task for fetching a state quantity of a control target, a task for performing monitoring and control signal generation processing of a control target using the state quantity fetched by the task, and an abnormality processing. A task including at least one of the tasks described above is configured to execute processing using a stack allocated by the stack allocation method according to any one of claims 1 to 3, and to control a control target. Disclosed is a control device.

Further, the present invention is characterized in that the controlled object is a construction machine, the task executing means is a single-chip microcomputer, and the stack is set in a memory of the single-chip microcomputer. A control device is disclosed.

[0018]

BEST MODE FOR CARRYING OUT THE INVENTION The present invention will be described below with reference to its embodiments. FIG. 1 shows an example of the configuration of a multi-task OS using the stack allocation method according to the present invention, and shows only the part related to task switching. The task management table 10 is provided for each task, and the priority 101 of the task, the start address 102 indicating the head of the execution program of the task, and the stack pointer 103 indicating the end of the stack allocated to the task.
In addition to this, a flag 1 for distinguishing whether the task is a newly started task or a suspended task
It has 04 as its contents.

Here, in this embodiment, one stack is allocated for each priority of tasks, and tasks having the same priority share one stack. In FIG. 1, the stacks ST6 and ST7 for the priorities “6” and “7” are illustrated in the RAM 20. The stack pointer setting table 11 stores the top address of the stack for each priority, and the priority is referred to when the task management table 10 is created, and this top address is written in the task management table 10 as a stack pointer. . Also, priority 101 and start address 10
2 is given by the user designating each task and is written in the task management table 10.

Queues (queues) Q6, Q7 are list control blocks for arranging and managing the executable tasks of each priority in the order in which they are started. In FIG. 1, the priorities "6" and "7" are set. Although only the queue is shown, it is actually provided for all priorities. Then, each queue manages the task by the head address of the task management table 10 corresponding to the activated task. The execution task selection program 12 examines the queue provided for each of the above priorities, and has the highest priority and is activated earlier so that the tasks are taken out in order and executed.

FIG. 4 is a flow chart showing the task switching processing by the multitask OS as described above.
If another task A activation request is issued during execution of task B (step 401), the execution task B is interrupted, and a return address for restarting it and register information at that time are given. The task B is stored in the stack B provided in correspondence with the priority (step 402). Further task B
The value of the stack pointer 103 of the task management table for the task is set as the last address of the information stored in the stack B in step 402 (step 403), and the task B is connected to the head of the queue corresponding to the priority (step 40).
4). Further, the flag 104 is set to "1" (step 40).
5). However, the flag value "1" indicates that the task is suspended.

Since the suspension of the task B being executed and the processing associated therewith are completed by the above processing, the task A for which the activation request has been made next is connected to the queue corresponding to the priority of the task.
For example, if the priority of task A is “6”, the queue Q
The start address of the management table of task A is input to the last position within 6 (step 406). Then, the flag in the management table corresponding to the task is set to "0" (step 40).
7). This "0" indicates that the task has been newly requested to be activated.

Next, the execution task selection program 12 examines each queue, and among the tasks connected to each queue, the task C with the highest priority and at the head of the queue corresponding to that priority. (Step 40
8) The task C flag is checked to see if it has been newly activated (step 409). If flag =
If the task is "1", that is, the suspended task, the stack pointer 103 for resuming is written in the stack pointer 103 of the task management table 10. Therefore, an interrupt return instruction for this address is issued and the suspended task is suspended. The execution of task C that has been restarted is restarted (step 410).
If the flag of the task C is "1", that is, if the task is newly started, the start address 102 of the task management table is set in the stack corresponding to the task priority (step 411), and the stack pointer is set. Issue an interrupt return instruction to start execution of the task C (step 412).

FIG. 5 is a block diagram of the system described above.
6 illustrates an example of stack allocation of a multi-task OS when switching tasks. In a state where the task 1 with the priority “7” and the task 3 with the priority “6” are operating in parallel, the task 1 is the stack ST for the task with the priority “7”.
7, task 3 is stack ST6 for task with priority "6"
To use. When the execution of the task 3 is finished and the task 4 having the same priority “6” is activated, the task 4 uses the stack ST6 for the task of the priority “6”. Similarly, when the executions of task 1 and task 4 are completed and task 2 and task 5 are activated, task 2 uses task stack ST7 for priority "7" and task 5 uses task stack ST7 for priority "6". The stack ST6 is used. If task 2 is activated during execution of task 1, the stack is allocated to task 2 only after task 1 ends, so task 1 and task 2 do not access the same stack.

As described above, according to the present invention, since the stack is prepared for each priority, the execution of the task with the higher priority does not wait for the completion of the execution of the task with the lower priority, and the urgent task is required. The process that requires the process can be preferentially executed. Also, when multiple tasks with the same priority are started, the stack is not used for new tasks that are started later, and when that task is read from the queue and enters the execution stage. By using the stack for work to start execution, it is possible to prevent the contents of the stack from being accidentally rewritten without complicated exclusive control. Moreover, since one stack may be prepared for each priority, the stack area can be reduced as a whole, and multitask processing can be easily performed even when the memory is small like a single-chip microcomputer.

In the configuration of FIG. 1, one stack is assigned for each priority of the task, but there is a series of task groups that are sequentially activated, and those tasks have a plurality of priorities. However, if there is no conflict between tasks with different priorities, stacking these tasks together and allocating one stack saves more stack. In order to perform such allocation, the top addresses of stacks having a plurality of priorities corresponding to the series of task groups may be written in the stack pointer setting table 11 as the same value.

Next, an example of a construction machine control apparatus using the stack allocation method of the present invention will be described. A single-chip microcomputer is used as the control device for the construction machine, and FIG. 6 is a block diagram showing a configuration example thereof, which is a control device for forming the control system described in FIG. In the figure, information from various sensors attached to the construction machine, for example, a sensor that measures the pump discharge pressure, a sensor that measures the tilt angle of the pump, a sensor that measures the hydraulic oil temperature, etc.
RAM 63 via analog input interface 65
It is saved in the data storage area above. Further, in measuring the engine speed, a number of pulses corresponding to the engine speed are input to the digital input interface 67, the engine speed is measured using the built-in timer 64, and the data is stored in the RAM 63. CPU61 is ROM62
This information is processed by using the program stored in, the control target value of the hydraulic control system or the machine control system is calculated, and the command is output to the pump and the engine through the analog output interface 66. Input and output of these data and commands to the outside are performed by a plurality of tasks operating in multitask, but the number of tasks increases as the degree of sophisticated control is increased.

Therefore, priority is given to tasks according to the importance of control of the construction machine, and one stack is provided on the RAM 63 for each priority, thereby performing multitask processing using a relatively small stack area. To be realized. For example, the control system of FIG. 2 shows three priorities P1, P2, and P3, but in the present invention, one stack is allocated for each of these priorities. As a result, in the case of FIG. 2, compared with the conventional method of allocating one stack for each task,
Since only one stack is required, the stack area may be 3/7.
Moreover, for example, when the task T11 of the highest priority P1 is requested to be activated during the execution of the task of the lower priority P2 or P3,
Execution of the task T11 can be started without waiting for the end of the lower task, and the safety and controllability of the construction machine can be improved. Further, since the processing apparatus can be small in size, cost reduction, space saving, and improvement in reliability due to improvement in vibration resistance can be achieved.

Although an example of application to a construction machine has been described here, it is obvious that the present invention is also suitable for a control system that processes a large number of tasks with a single-chip microcomputer, such as a control device for an automobile. .

[0030]

According to the present invention, by allocating a common stack area to tasks having the same priority, it is possible to reduce the RAM area usage of the multitask OS. In addition, task priority switching due to complicated resource exclusion control is eliminated, and task switching time is shortened. In addition, since the multi-task OS can be installed in a single-chip microcomputer with a small RAM capacity, which is embedded in a device, programs can be easily created and management can be facilitated. Further, it is possible to reduce the cost of a control device such as a construction machine, save space, and improve vibration resistance.

[Brief description of drawings]

FIG. 1 is an explanatory diagram of a stack allocation method according to the present invention.

FIG. 2 is an explanatory diagram of a hydraulic excavator control device using a single-chip microcomputer.

FIG. 3 is a configuration example of hydraulic excavator control software that runs on a multitasking OS.

FIG. 4 is a flowchart of task switching processing when the stack allocation method of the present invention is used.

FIG. 5 is a diagram showing an example of a stack allocation method of the present invention.

FIG. 6 is a block diagram showing a configuration example of a control device.

[Explanation of symbols]

 10 task management table 11 stack pointer setting table 12 execution task selection program 13 RAM 61 CPU 62 ROM 63 RAM 64 timer 65 analog input interface 66 analog output interface 67 digital input interface 104 flag ST6, ST7 stack Q6, Q7 queue

Claims (5)

[Claims] 1. A stack allocation method in a multitasking operating system, wherein one stack is provided for each priority of tasks, and tasks having the same priority share one stack. Stack allocation method. 2. A first task group having the same priority,
If there is no conflict between an arbitrary task of the first task group and an arbitrary task of the second task group with respect to a second task group having a priority different from that of the task group, 2. The stack allocation method according to claim 1, wherein the stack corresponding to the task group and the stack corresponding to the second task group are shared to be one. 3. A queue is provided for each priority, a flag is provided for each task, and for a task interrupted by an interrupt, information necessary for restart is stored in a stack corresponding to the priority of the task,
The task is connected to the head of the queue corresponding to its priority with that flag as a value indicating suspension, and for the task that has been activated, the flag is set as a value indicating new activation and that task is associated with that priority. When one task is selected as an execution task by connecting to the end of the queue and examining the queue, and the flag of the execution task indicates suspension,
When the process is restarted using the information necessary for resumption stored in the corresponding stack, and the flag of the selected execution task indicates a new start, the stack of the execution task with the priority corresponding to the execution task 3. The stack allocating method according to claim 1, wherein a start address is set, processing is started from the set start address, and stacks are dynamically allocated in this way to perform multitask processing. 4. A task for fetching a state quantity of a control target, a task for monitoring a control target and generating a control signal using the state quantity fetched by the task,
And a task including at least one of the tasks for abnormal processing is executed by using a stack allocated by the stack allocation method according to claim 1, and a control target is controlled. A control device characterized by being configured to perform. 5. The control target is a construction machine, and the task executing means is a single-chip microcomputer.
The control device according to claim 4, wherein the stack is set in a memory of the single-chip microcomputer.