JPH06119190A - Task switching processing system - Google Patents

Abstract

PURPOSE:To provide task switching processing to perform customization. CONSTITUTION:When processing is transferred from system call processing to task switching processing, it is decided whether or not the processing is the system call processing nested at present (step 11), thence, it is decided whether or not the system call processing issued during interruption processing is being performed (step 12). In such operation explanation, since it is the system call processing in which a task is issued as usual, the processing is transferred to the one in step 13. Firstly, the task with the highest priority out of the tasks set in ready states is searched (step 13). Thence, it is decided whether or not the task is equal to the task from which a system call is issued (step 14). In the operation explanation, since the task is switched from a task 1 to a task 2, the processing is transferred to step 15. The address of a table and an initial pointer in a corresponding table can be obtained in step 15.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a task switching processing system, and more particularly to a task switching processing system used in a real-time operating system for an interrupt system.

[0002]

2. Description of the Related Art Conventionally, a real-time operating system for a microprocessor (hereinafter abbreviated as RTOS) is generally sold only in an executable form. This is because the user uses the RTOS as a component of software like a black box. However, in recent RTOS, with the expansion of the function and the code size, it has become possible for the user to select and incorporate the function.

The functions of the RTOS are roughly classified into the following two types.

(1) System call processing (2) Task switching processing Since system call processing is independent for each system call, incorporation can be easily selected. However, the task management / switching process is a core function of RTOS, and its performance is also required. For this,
The structure is such that the user cannot select the function.
This task switching process will be described with reference to FIGS. 9 and 10.

FIG. 9 is a flow chart showing the procedure of task switching processing, and FIG. 10 is a block diagram showing the operating environment of this software. For simplicity, the tasks are
It is assumed that there are two tasks, task 1 and task 2. The task process is a process performed using a coprocessor for both task 1 and task 2. Now, task 1 is operating, and it is switched to task 2 by the system call issued by itself. This switching process is performed according to the flowchart shown in FIG.

In the task switching process, a determination process is performed as to whether or not the current system call process is a nested system call process (step 81), and then a determination process as to whether or not the system call process is issued during the interrupt process. (Step 82). In this explanation of the operation, since the task is the processing of the system call issued as usual, the processing shifts to step 83 in FIG. First, the task with the highest priority is searched for from among the tasks waiting for CPU allocation (hereinafter referred to as "ready state") (step 83). Next, it is determined whether this task and the task that issued the system call are the same (step 8).
4). In this operation explanation, task 1 to task 2
To switch to, the processing moves to step 85.

In step 85, the coprocessor usage decision is made, and task 2, which is switched, saves the coprocessor context of task 1 currently in use to use the coprocessor (step 2). 86). Subsequently, the coprocessor context of task 2 is loaded (step 87) and finally task 1
Is saved (step 88), the context of task 2 is saved (step 89), and the context of task 2 is loaded (step 9).
0).

As described above, task 1 to task 2
The task switching process is performed.

[0009]

In the conventional task switching processing method described above, the task switching processing of RTOS is provided to the user as a black box. For this reason, there are drawbacks shown below.

The first drawback is that even if the user has an unnecessary function, it is directly incorporated in the system.

In the task switching process, the resources assigned by the RTOS to each task must be switched. Taking a coprocessor as an example, if the coprocessor is not used at all, or if it is used by only one task, the support in the task switching processing is unnecessary. In embedded systems, the software memory size is limited to reduce device cost. For this reason, the installed software has to eliminate waste even as much as one byte for unnecessary parts. In the conventional task switching processing method, it is impossible to delete this unnecessary part. This results in unnecessarily large size restrictions on the user application.

The second drawback is the lack of scalability.

In the user application, RTO
In addition to the resource for each task that S originally supports, a resource to be switched may occur. Consider, for example, the CPU “μPD70136” manufactured by NEC.
Unlike "page registers" for memory management
Is provided. When switching tasks, this page
Attempting to switch the register is impossible because the conventional processing method does not have expandability. This impedes flexibility in building an application and imposes great restrictions.

[0014]

A task switching processing method of the present invention is a processing procedure for switching a task currently assigned to a predetermined CPU to another task in an operating system for an embedded system. As the first processing routine registered in the address table of the operating system.
Registered next to the first processing step of obtaining an address, the second processing step of executing the processing routine obtained in the first processing step, and the processing routine executed in the second processing step. And a third processing step for transferring control to the first processing step, and sequentially calling processing routines registered in advance in the address table. It is characterized in that task individual information to be switched is sequentially switched.

The address table may be held for each of the tasks, or may be held in common for the tasks and may be held for each of the tasks together.

[0016]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, the present invention will be described with reference to the drawings.

FIG. 1 is a flowchart of a task switching pre-processing according to a first embodiment of the present invention, FIG. 2 is a flowchart of a coprocessor switching processing of this embodiment, and FIG. 3 is a task-switching processing flowchart. . Moreover, in FIG.
It is a figure which shows the structure of the address table in T. It is assumed that the task is in the same state as in the case of the above-mentioned conventional example.

When the system call processing shifts to the task switching processing, it is judged whether or not the current system call processing is a nested system call processing (step 11), and then whether or not the system call processing issued during the interrupt processing is in progress. Is determined (step 12). In this explanation of the operation, since the task is the processing of the system call issued as usual, the processing shifts to step 13. First, the task with the highest priority is searched from the tasks in the ready state (step 13). Next, it is determined whether this task is the same as the task that issued the system call (step 14). In this explanation of operation, Task 1
The process proceeds to step 15 to switch from task 2 to task 2.

In step 15, the address of the table shown in FIG. 4 and the pointer of the first entry in the table are obtained. The address of the processing routine in the first entry is the address of the switching processing routine of the coprocessor. Therefore, the process moves to step 31 shown in FIG. In step 31, the coprocessor use determination processing is performed, and the task 2 to be switched is saved in the coprocessor context of the task 1 currently in use in order to use the coprocessor (step 32). . Subsequently, the coprocessor context for task 2 is loaded (step 33). In this way, the coprocessor switching process ends, and the process returns to step 17 in FIG.

In step 17 in FIG. 1, the pointers to the above-mentioned table entries are updated,
The address of the processing routine in the next entry is the address of the task switching processing routine. Therefore, the process moves to step 21 shown in FIG. In the post-task-switching processing routine shown in FIG. 2, the context of task 1 is first saved (step 21), and then the context of task 2 is loaded (step 22).

As described above, the processing of the task switching post-processing routine is completed, and the processing is switched from the task 1 to the task 2.

Next, a second embodiment of the present invention will be described. 5 is a flowchart of task switching processing of this embodiment, FIG. 6 is a flowchart of coprocessor processing, FIG. 7 is a flowchart of register processing, and FIG.
It is a figure which shows the structure of the address table in a task control block (abbreviated to TCB hereafter).
It is assumed that the task is in the same state as in the case of the above-mentioned conventional example.

When the system call processing shifts to the task switching processing, it is judged whether or not the current system call processing is a nested system call processing (step 41 in FIG. 5). It is determined whether or not (step 42). In this explanation of the operation, since the task is the processing of the system call issued as usual, the processing shifts to step 43.
First, the task with the highest priority is searched from the tasks in the ready state (step 43). Next, it is determined whether this task and the task that issued the system call are the same (step 44). In this explanation of operation,
The process proceeds to step 45 to switch from task 1 to task 2.

In step 45, a flag indicating that switching is performed is set in the switching process identification flag area of the TCB shown in FIG.
A pointer to the first entry in the table in is obtained (step 46). The address of the processing routine in the first entry is the address of the coprocessor processing routine. Therefore, the processing shifts to step 61 of the flowchart shown in FIG. Step 61 in FIG.
At, it is determined whether or not the above-mentioned identification flag is set. Since the flag of task 1 is set in step 45 of FIG. 5, the process proceeds to step 62 of FIG. Task 2 that can be switched
Saves the coprocessor context of task 1 currently in use to use the coprocessor (step 63). In this way, the coprocessor processing ends, and the processing returns to step 48 in the processing flow of FIG.

In step 48 in FIG. 5, the pointer to the above-mentioned entry is updated, and the address of the processing routine in the next entry is the address of the register processing routine. Therefore, the processing shifts to step 71 of the processing flow shown in FIG. Step 71
In the same manner as in step 61 in FIG. 6, it is determined whether or not the identification flag is set. Since the flag of task 1 has been set, the context of task 1 is saved (step 72) and the register processing ends. Then, the processing procedure shifts to step 48 in the processing flow of FIG. Again in step 48, the pointers to the previously mentioned entries are updated. The address of the processing routine in the next entry is "0". Therefore, the process is step 5 in FIG.
Move to 0. In step 50, the above-mentioned identification flag of task 1 is cleared, and then TC of task 2 is cleared.
The pointer to the first entry in the table in B is obtained (step 51).

The address of the processing routine in the first entry is the address of the coprocessor processing routine.
Therefore, the process is step 6 in the process flow of FIG.
Then, in step 61, it is determined whether or not the identification flag is set. Since the flag of task 2 has been cleared, step 64 in FIG.
, The coprocessor context of task 2 is loaded in step 64, and the coprocessor processing ends. Then, the process flow returns to step 53 shown in FIG. In step 53, the pointer to the aforementioned entry is updated, and the address of the processing routine in the next entry is the address of the register processing routine. Therefore, the processing flow shifts to the processing of step 71 shown in FIG. 7. Similar to the processing of step 61 in the processing flow of FIG. 6, step 71
At, it is determined whether the identification flag is set. Since the flag of task 2 has not been cleared, the context of task 2 is loaded (step 73) and the register processing ends. As a result, the processing flow moves to the processing of step 53 shown in FIG. In step 53, the pointer to the above-mentioned entry is updated again, and the address of the processing routine in the next entry is "0". Therefore, the task switching process ends.

The task 1 is processed through the above processing procedure.
The task switching from the processing of 1 to the processing of task 2 is performed.

[0028]

As described above, according to the present invention, the task switching processing is divided into modules according to functions and the addresses thereof are managed by a table, so that the function selection in the task switching processing is simply rewritten in the table. Can be easily realized by
There is an effect that the function addition to the RTOS is easily realized.

Further, by holding the table for each task, it is possible to avoid uniform processing at the time of task switching, and thus it is possible to greatly reduce the processing time at the time of context switching. There is an effect.

[Brief description of drawings]

FIG. 1 is a diagram showing a flowchart of task switching preprocessing according to a first embodiment of this invention.

FIG. 2 is a diagram showing a flowchart of a coprocessor switching process of the first embodiment.

FIG. 3 is a diagram showing a flowchart of post-task switching processing of the first embodiment.

FIG. 4 is a diagram showing the structure of an address table in the SBT.

FIG. 5 is a diagram showing a flowchart of task switching processing according to the second embodiment of this invention.

FIG. 6 is a diagram showing a flowchart of a coprocessor switching process of the second embodiment.

FIG. 7 is a diagram showing a flowchart of a register switching process of the second embodiment.

FIG. 8 is a diagram showing the structure of an address table in the TCB.

FIG. 9 is a diagram showing a flowchart of a task switching process of a conventional example.

[Explanation of symbols]

11, 41, 81 System call processing determination step 12, 42, 82 Interrupt processing determination step 13, 43, 83 Ready task search step 14, 44, 84 Current task identity determination step 15, 17, 46, 48, 51, 54 entry acquisition step 16, 47, 52 various processing steps 21, 88 task context saving step 22, 55, 73, 89 task context loading step 31, 62, 85 processor usage determining step 32, 63, 72,86 save coprocessor context 33,64,87 load coprocessor context 45 set flag 49 "0" decision step 50 clear flag 61,71 set flag Determining step

Claims (3)

[Claims] 1. An operating system for embedded systems
In the system, as a processing procedure for switching a task currently assigned to a predetermined CPU to another task, the first processing routine registered in the address table of the operating system Registered next to the first processing step for obtaining an address, the second processing step for executing the processing routine obtained in the first processing step, and the processing routine executed in the second processing step. A processing routine address that is present and transfers control to the first processing step
The task switching processing method comprising the steps of: (1) and sequentially calling the processing routines registered in advance in the address table to sequentially switch the task individual information to be switched. 2. The task switching processing method according to claim 1, wherein the address table is held for each task. 3. The task switching processing method according to claim 1, wherein the address table is held in common for the tasks and also in the individual tasks.