JP2004318469A - Program tracing system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a program tracing system which does not take time in repairing destruction of a queue nor cause runaway. <P>SOLUTION: This system is provided with a trace buffer 20, a queue 30 and a resource check section 40, and detects a usage error of dynamic management of memory often used in the resource check section 40 when finishing interruption processing and the starting of task switch processing until execution of only self-task/self-interruption processing, and block the influence to other tasks. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
TECHNICAL FIELD OF THE INVENTION
The present invention relates to a program tracing device, for example, a program tracing device for a program controlled by a real-time monitor (OS: Operating System) that executes a plurality of events in an order of processing based on priorities and executes the events.
[0002]
[Prior art]
The program tracing device is used for analyzing the execution operation of a program in a computer system in which a plurality of execution units can operate in parallel, for example, a computer system that supports multi-process, multi-thread, and the like.
[0003]
As a conventional technique relating to such a program trace device, a process trace device that creates a process trace table and traces a process is disclosed (for example, see Patent Document 1). Referring to FIG. 2, this conventional process trace apparatus 100 will be briefly described. The program tracing device 100 includes a queue 110, a trace buffer 120, and a queue registration check unit 130. The queue 110 manages memory resources (resources) and timer resources. The trace buffer 120 stores a change history of the chain structure of the queue 110. When the resource is to be used (that is, when the resource is to be acquired or deleted), the queue registration checking unit 130 checks whether there is an abnormality in the change history of the chain structure of the queue 110.
[0004]
The operation of the program tracing device 100 is as follows. When the user wants to use the memory resource, he specifies the number of bytes he wants, and if he wants to use the timer resource, he specifies that he wants to use one timer. In the called side acquisition processing, the queue registration check unit 130 is executed, the change history of the chain structure of the queue 110 is checked from the saved trace buffer 120, and the history is compared with the current state of the queue 110. To check whether the queue 110 has been destroyed. If there is an abnormality, the queue registration check unit 130 stops the program execution by the HALT instruction of the CPU (Central Processing Unit) and executes an abnormality detection process such as an infinite loop or resetting.
[0005]
Also, a program trace device that detects an error in task processing in real time has been disclosed (for example, see Patent Document 2). As shown in FIG. 3, the program trace device 200 includes a queue 210, a trace buffer 220, and a resource check unit 230. Here, the queue 210 manages memory resources and timer resources. The trace buffer 220 acquires the resources such as the memory and the timer which are dynamically managed by the queue 210 (the resources such as the memory and the timer which are vacant only during use and returned when the use is completed). , And saves access information to resources when writing and releasing are performed. When exiting the interrupt process or at the beginning of the task switching process of the real-time monitor (OS), whether resources such as a memory or a timer have been accessed within a range where there is a current interrupt process or a task process that was in the RUN state immediately before. Check if.
[0006]
[Patent Document 1]
JP-A-10-301805 (page 8, FIG. 17)
[Patent Document 2]
JP-A-2002-251299 (pages 3-4, FIG. 1, FIG. 3)
[0007]
[Problems to be solved by the invention]
According to the above-described prior art, even if queue destruction is detected when a resource is desired to be used, it is not immediately known which task and when the queue was destructed, so that it takes time to correct the task. The reason is that in a real-time monitor that determines and executes a processing order based on the priority for each of a plurality of occurrence events, queue destruction may occur due to any interrupt processing as described below.
[0008]
The first is "forget to release resources". If there is a route that does not return the acquired resources such as memory and timer due to a program bug at any one place of all tasks and interrupt processing, resources are kept secured every time the route is passed. Resources run out. Thereafter, when it is determined that there is no resource, an abnormal process is executed. However, if a process without resources is not prepared, a route in which resources are secured is executed despite the lack of resources, and data is written to address 0, or resources have been secured in the past. Since the data at that time is used as it is, another task or interrupt processing overwrites the used memory or timer. As a result, they run away after this.
[0009]
Second, "resource not acquired". If there is a route that does not acquire any resource in the process of acquiring resources, or if more memory is used, the data of the task or interrupt process that acquired the resources correctly is overwritten, Runaway occurs because of destroying the data.
[0010]
[Object of the invention]
SUMMARY OF THE INVENTION The present invention has been made in view of the above-mentioned problems of the related art, and provides an improved program trace device that automatically tests and can cope with the above-mentioned "forget resource release" and "resource not obtained". The purpose is to do.
[0011]
[Means for Solving the Problems]
In order to solve the above-described problem, the program trace device according to the present invention employs the following characteristic configuration.
[0012]
(1) A trace buffer for storing access information to a resource to be obtained, written, or released, a queue for managing a start point to an end point of the resource, and a resource check unit for checking whether the resource has been accessed And a resource tracing unit for detecting the use error of the resource up to the execution of only its own task processing.
[0013]
(2) The above-mentioned (1) wherein the resource check unit executes the check when exiting the interrupt processing or at the beginning of the task switching processing, to the extent possible for the current interrupt processing or the task in the RUN state immediately before. Program trace device.
[0014]
(3) The program trace device according to the above (1) or (2), wherein the resource is a memory and performs dynamic management.
[0015]
(4) The resource check unit according to (1), (2) or (3), wherein the dynamically managed memory is checked separately in a state where the memory is acquired and in a state where the memory has been used and released. Program trace device.
[0016]
(5) The program tracing device according to (4), wherein the state in which the memory is acquired is further checked separately in a state in which the memory is acquired and a state in use.
[0017]
(6) The program tracing apparatus according to (4) or (5), wherein the state of use and release after use is checked separately in an initial state after power-on and in a released state.
[0018]
BEST MODE FOR CARRYING OUT THE INVENTION
Hereinafter, the configuration and operation of a preferred embodiment of a program trace device according to the present invention will be described in detail with reference to the accompanying drawings.
[0019]
FIG. 1 is a system configuration diagram of a preferred embodiment of a program trace device according to the present invention. The program tracing device 10 is intended for a computer capable of operating a plurality of tasks and interrupt processing. The program trace device 10 of FIG. 1 includes a trace buffer 20, a queue 30, and a resource check unit 40.
[0020]
The queue 30 manages memory resources. That is, when the queue 30 wants to use a memory resource, it designates a desired number of bytes and calls a memory acquisition process. In the called party's acquisition process, a search is made for a free resource from the starting point, and if free, the called party is notified of the obtained location. In many cases, in the acquisition process, the memory resources are managed by the queue 30 in a chain structure. The method is as follows.
[0021]
The address of one used memory resource is stored at the starting point of the queue 30 for managing the memory resource. The address of one memory resource used next is stored after the address indicated there. After the address, the address of one memory resource used next is further stored, and the address of the next memory resource is stored until the end point indicating that the used memory resource ends here. I do.
[0022]
Next, when acquiring, writing, or releasing a memory resource managed by the queue 30, the trace buffer 20 stores access information to the memory resource. The resource check unit 40 checks whether the memory resources have been accessed within a possible range when exiting the interrupt process and at the beginning of the task switching process.
[0023]
Next, the overall operation of the program trace device 10 shown in FIG. 1 will be described in detail. First, when each task / interrupt process acquires, writes, or releases a memory resource managed in the queue 30, information on access to the resource is stored in the trace buffer 20. Save the following values:
[0024]
(1) At the time of acquisition {circle around (1)} The resource state is set to “the state in which only acquisition is performed”.
(2) Store the address and range of the acquired resource.
(3) Store the acquired process number, the process number that can be written, and the process number that can be released.
[0025]
(2) At the time of writing (1) The state of the resource is set to “written state”.
(2) Store the written process number.
{Circle around (3)} Store the written value (only the amount that can be actually stored by resources).
[0026]
(3) At the time of release (1) The state of the resource is set to the "released state".
(2) Store the released process number.
[0027]
Next, at the time of exiting the interrupt processing or after executing the processing of the OS involving task switching such as inter-task notification and event occurrence waiting within the task, the resource check unit 40 executes (A), (B), and ( By making the determinations in C) and (D), it is detected whether or not the current interrupt processing or the task in the RUN state has performed abnormal processing on resources.
[0028]
That is, the present invention provides a dynamic management of frequently used memory in a program which is controlled by a real-time monitor which determines a processing order based on a priority for each occurrence of a large number of events and has a frequently used program structure. In other words, a memory error that is vacant only for the usage period and released after use) is detected until the execution of only the own task / self interrupt processing, and errors in the own task / self interrupt processing are not affected by other tasks. It is intended to provide an OS. Therefore, when exiting the interrupt processing and at the beginning of the task switching processing of the real-time monitor (OS), it is checked whether the current interrupt processing or the task which was in the RUN state immediately before accessed the memory within the previously declared range. Execute the processing to be performed.
[0029]
Further, when the state of each task transits, a process is executed to check whether the memory is released as designed in each state or there is no omission of release. The memory check is performed by the following method. In other words, the dynamic memory is in a state (first state) acquired by the task processing / interrupt processing because it wants to use it, an initial state after the power of the apparatus is turned on, or the memory acquired by the task processing / interrupt processing is finished. Therefore, it is roughly classified into two states of a released state (second state). The first state is divided into (A) and (B), and the second state is divided into (C) and (D).
[0030]
(A) State of Acquired If only the acquired state is not used, the content value of the memory should not be different from that before execution of the current task processing or interrupt processing. At the time of exiting the interrupt process and at the first timing of the task switching process of the OS, the acquired memory value is read, and the value has changed from the value of the previous interrupt process or the value of the task in the RUN state immediately before. In this case, the task that has been in the interrupt processing or the RUN state has performed abnormal processing on the acquired memory. At this time, the amount of memory used, for example, 80% of the designed amount is checked to determine whether there is an abnormality or whether the task is in a memory acquisition state designed in advance in the state of the task.
[0031]
(B) A state in which the memory has been acquired and the memory acquired by correct processing as designed has been rewritten, and the memory is in use. At this time, the memory value should be the same as the value rewritten by the correct processing as designed. Further, it can be checked that the acquired address has been rewritten with the correct process within the acquired range as designed, by confirming the information obtained at the time of acquisition only and the process number determined in advance by design. At the time of exiting the interrupt process and at the first timing of the task switching process of the OS, the same value, the address, the range, and the write process number are checked to detect an abnormality.
[0032]
(C) The contents of the initial state memory after power-on should be the initial values. In particular, if the memory does not set an initial value, there should be no change from the value at power-on. The same value is checked and detected at the first timing of the task switching process of the OS when exiting the interrupt process.
[0033]
(D) The content value of the released state memory should not change from the value at the time of release. The same value is checked and detected at the first timing of the task switching process of the OS when exiting the interrupt process.
[0034]
As described above, it is possible to detect an erroneous use of the resources of the memory that is dynamically managed up to the execution of only the own task / interrupt processing.
[0035]
The configuration and operation of the preferred embodiment of the program trace device of the present invention have been described above in detail. However, such an embodiment is merely an example of the present invention, and does not limit the present invention in any way. It will be readily apparent to those skilled in the art that various modifications and changes can be made in accordance with the particular application without departing from the spirit of the invention.
[0036]
【The invention's effect】
As is apparent from the above description, the program tracing device of the present invention has the following practically significant effects. That is, an error in the usage of the memory resources is detected up to the execution of only the own task / self interrupt processing, and the error of the own task / self interrupt processing does not affect other tasks.
[Brief description of the drawings]
FIG. 1 is a block diagram of a preferred embodiment of a program trace device according to the present invention.
FIG. 2 is a block diagram of a first conventional example of a program trace device.
FIG. 3 is a block diagram of a second conventional example of a program trace device.
[Explanation of symbols]
10 Program Trace Device 20 Trace Buffer 30 Queue 40 Resource Check Unit

Claims (6)

The system includes a trace buffer for storing access information to a resource to be obtained, written, or released, a queue for managing a start point to an end point of the resource, and a resource check unit for checking whether the resource has been accessed. A program tracing device, wherein the resource check unit detects the use error of the resource until execution of only the own task process. The resource check unit, when exiting an interrupt process or at the beginning of a task switching process, executes the check in the possible range of the current interrupt process or the task that was in the RUN state immediately before. 2. The program trace device according to 1. 3. The program trace apparatus according to claim 1, wherein the resource is a memory and performs dynamic management. The program according to claim 1, wherein the resource check unit checks separately the state in which the dynamically managed memory has been acquired and the state in which the memory has been used and released. Trace device. 5. The program tracing apparatus according to claim 4, wherein the state in which the memory is acquired is further checked separately into a state in which the memory is acquired and a state in use. 6. The program tracing apparatus according to claim 4, wherein the released state after use is further checked separately in an initial state after power-on and a released state.

Images