JPH0581051A - Resource destruction supervisory system - Google Patents

Abstract

PURPOSE:To speedily detect the destruction of a resource such as a memory and to easily specify a suspected process. CONSTITUTION:Whenever a clock is interrupted by a system clock the kernel part 1 of OS checks the presence or absence of the destruction of the respective resources 3 and 5 in a resource group 2 by inspecting the content of resource management information 6 added to it (S2). When it is destroyed, a system is stopped (S3). When it is not destroyed, programs for respective levels are executed and controlled (S5-S7), an execution process is selected by an application process scheduler (S9) and the application process is execution-controlled (S11). Register information on the process and data on the process control table in OS are accumulated in a collection memory 7 immediately before the execution control of an application process.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a resource destruction monitoring system in an electronic computer.

[0002]

2. Description of the Related Art Generally, in an electronic computer, resources such as memory are managed by being divided into, for example, predetermined sizes, and unused resources are connected in a chain so that, for example, at the beginning of the A method is used in which unused resources are used in order.

By the way, when an application program or the like is executed as a process on an electronic computer, a chain structure of unused resources is destroyed or a used resource is changed to an unused resource due to a bug of the application program. Resource destruction such as may occur.

Since resource destruction is not only a resource destruction but also a cause of other failures, when a resource destruction occurs, it is promptly detected, the cause is investigated, and necessary measures are taken. It is important to take.

Therefore, conventionally, the following measures (A) and (B) have been taken.

(A) When resources cannot be captured even after a certain period of time, it is determined that resource destruction has occurred and the system is stopped.

(B) Resource management information for determining resource destruction is added to the resource, and when the resource is captured and released, the information is collated to detect the resource destruction and the system is stopped.

[0008]

However, each of the above countermeasures has the following problems.

In measure A, the system is stopped not only when resources are destroyed by a program bug but also when the absolute number of resources required for program operation is insufficient. Therefore, it is not known whether the resource destruction has occurred until the various information is analyzed in detail when the system is stopped.

In the countermeasure B, the resource destruction is detected at the time of capturing and releasing the resource on the basis of the resource management information for determining the resource destruction added to the resource. Therefore, the problem of the countermeasure A is solved. However, there is a problem in identifying the suspicious process that caused the resource destruction. That is, if resource destruction is detected at the time of resource release, the suspicious process can be identified by tracing the program backward from the process that released the resource according to the process information managed by the OS. When the resource destruction is detected at the time of capturing the resources, all the processes executed in the past can be suspected processes. Therefore, it is difficult to actually trace the program, and the suspected processes cannot be specified.

The present invention has been made in view of such circumstances, and an object thereof is a resource destruction monitoring system capable of promptly detecting resource destruction and easily identifying a suspicious process. To provide.

[0012]

In order to achieve the above-mentioned object, the resource destruction monitoring system of the present invention is an electronic computer equipped with an OS for controlling execution of a process each time a clock interrupt by a system clock is received. The OS is provided with a resource check means for checking whether or not resources are destroyed at the time of the clock interrupt, and a process information save means for collecting information about the process related to the execution control and storing it in a memory. ing.

In a preferred embodiment of the present invention, the process information saving means collects information about application processes related to execution control and stores it in the memory.

The process information to be collected includes register information of the execution process and data of the process control table.

[0015]

In the resource destruction monitoring system of the present invention, the resource check means provided in the OS for controlling the process execution each time a clock interrupt by the system clock is received, the resource check means searches the chain structure of the resource at the time of the clock interrupt. The resource management information is checked to see if the resource is destroyed, and if so, the system is stopped.

Further, the process information saving means provided in the OS accumulates register information of the execution process and data of the process control table in the memory as information about processes such as application processes related to execution control.

Therefore, when resource destruction is detected by the resource checking means and the system is stopped, the suspicious process can be easily identified by checking the process information stored in the memory, and the program can be identified based on the information. You can easily find a program bug by tracing.

[0018]

Embodiments of the present invention will now be described in detail with reference to the drawings.

FIG. 1 is a block diagram of an embodiment of the present invention.
1 is a kernel part of the real-time OS, 2 is a resource group that is a memory, 3 is an unused resource, 4 is a chain connecting unused resources, 5 is a resource in use, 6 is resource management information, and 7 is a collection memory.

The resource group 2 is a collection of unused resources 3 and in-use resources 5. Unused resources 3 that are not yet used by any process are connected in a chain 4 in a chain, and resources are required. At any time, an arbitrary number of unused resources 3 among them are taken out, and the chains 4 are reconnected.

The resource management information 6 added to each resource 3 and 5 is information used for detection of resource destruction and the like, and includes, for example, a flag indicating use / unuse, data indicating the size of the resource, and the like. Has been.

The collection memory 7 is a memory for storing detailed information of the execution process collected by the kernel unit 1. In this embodiment, the detailed information includes register information of the execution process and data of the OS process control table. To store.

The kernel section 1 is the core of the real-time OS that operates as the basic software of the electronic computer, and when the system clock is interrupted by hardware (not shown) (S1), steps S2 to S1.
The process of 1 is performed, and the CC (CPU) becomes idle (S1
2).

Here, steps S5 to S9 and S11 are processes related to basic operations similar to the conventional ones (program execution control of each level,
Steps S2 to S4 surrounded by broken lines are steps for executing application process schedules, etc.).
S10 is a step of executing a process characteristic of the resource destruction monitoring method of this embodiment.

FIG. 2 is a time chart showing an example of processing performed when the kernel unit 1 receives a clock interrupt by the system clock. The operation of this embodiment will be described below with reference to the drawings.

When the clock interrupt 21 by the system clock is input (S1), the kernel section 1 first performs a resource check (S2, 22).

In this resource check, for example, the chain 4 of the unused resource 3 in the resource group 2 is traced, and whether the chain 4 is not broken or whether the unused / unused flag of the resource management information 6 of each unused resource 3 is used, for example. Checks whether unused resources are destroyed by checking whether all are unused. In addition to the above, the resource management information 6 of all resources including the used resource 5 is also checked to see if other types of resources are destroyed.

Assuming that the resources are not destroyed, the kernel unit 1 judges that the resources are not destroyed,
The save pointer is initialized to 0 (S4), then the highest priority program execution control (S5, 23), the level 1 program execution control (S6, 24), the level 2 program execution control (S7, 25) are performed, and further When is finished,
It is determined whether or not there is an interrupted application process (S8), and if there is not, an application process to be executed is selected by the application process scheduler (S9), which is the same operation as in the conventional case.

Immediately before the execution control (S11, 27) of the application process selected in step S9 or the suspended application process determined to exist in step S8 is performed, the register information of the process and the OS of the process are displayed. The data of the process control table (for example, the address during execution of the process, the suspend / resume address, the start address of the work area in use, etc.) is collected and stored in the area indicated by the save pointer of the collection memory 7 (S1).
0,26).

After the execution of step S11, that is, after the execution control of the application process, the kernel unit 1
Enters the CC idle state (S12) and waits for a clock interrupt by the next system clock.

Here, it is assumed that there is a program bug in the application process whose execution is controlled in step S11, and the resource in the resource group 2 is destroyed due to this.
However, this is not determined at this point, and the kernel unit 1 enters the CC idle state as described above, and waits for the clock interrupt by the next system clock.

When the next clock interrupt 29 by the system clock is input (S1), the kernel section 1 performs the resource check again (S2, 30). Then, at this point, the resource destruction is detected and the kernel unit 1 stops the operation of the OS. This stops the system (S3).

At the time of this system stop, the register information of the execution process collected in the previous cycle, that is, the application process in this case, and the O
Since the data of the process control table of S is accumulated, it is possible to easily identify the program causing the resource destruction, that is, the program having the bug by examining them, and to trace the program based on the above information. Bugs can be found easily.

In the above embodiment, the process information is saved only for the application process among the processes whose execution is controlled at the time of the clock interrupt by the system clock. This is generally level 1 or level 2 of kernel part 1.
The program is based on the assumption that there are no bugs. However, if a program of level 1, 2 or the like may also become a suspicious process, not only the application process but also process information for each process including level 1, 2, etc. may be saved.

[0035]

According to the resource destruction monitoring system of the present invention described above, the following effects can be obtained.

Since the resource check is performed every time a clock interrupt is generated by the system clock, whether or not the resource is destroyed, the resource destruction can be detected promptly.

Since the register information of the execution process and the data of the process control table of the OS are stored in the memory as process information each time a clock interrupt occurs, complicated processing such that many processes are executed in parallel is executed. Even under the existing environment, the suspicious process can be easily identified, and if the program is traced based on the information, the program bug can be easily found.

[Brief description of drawings]

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

FIG. 2 is a time chart showing an example of processing performed when the kernel unit 1 receives a clock interrupt by a system clock.

[Explanation of symbols]

 1 ... Kernel part 2 ... Memory resource group 3 ... Unused resource 4 ... Chain 5 ... In-use resource 6 ... Resource management information 7 ... Collection memory

Claims (3)

[Claims] 1. An electronic computer having an OS for controlling execution of a process each time a system clock receives a clock interrupt, and a resource check unit for checking whether or not resources are destroyed at the time of the clock interrupt, A resource destruction monitoring system, characterized in that the OS is provided with a process information saving means for collecting information on processes related to execution control and accumulating it in memory. 2. The resource destruction monitoring system according to claim 1, wherein the process information saving means collects information about an application process related to execution control and accumulates the information in the memory. 3. The resource destruction monitoring method according to claim 1, wherein the process information to be collected includes register information of an execution process and data of a process control table.