JP4627275B2 - Monitoring program, monitoring method, and monitoring apparatus - Google Patents

Description

  The present invention relates to a monitoring program, a monitoring method, and a monitoring apparatus, and more particularly to a monitoring program, a monitoring method, and a monitoring apparatus that can detect a loop that occurs in a monitored program.

  During execution of a program, there is a state (loop state) in which an address executed by the program changes periodically. A fault such as a loop never ending (infinite loop) may occur during this loop state. Such a failure occurring in the loop is hereinafter referred to as a loop failure.

As a technique for detecting a loop fault, for example, techniques of Patent Documents 1 and 2 are known.
In Patent Document 1, the value of the program counter is acquired at a constant time interval, and when the acquired program counter value takes only one value, it is determined that a loop failure has occurred.

  Further, in Patent Document 2, instead of using the value of the program counter itself, the conversion value obtained from the program counter value acquired at regular intervals using a conversion value obtained by dividing the value of the program counter by, for example, 1000. It is determined that a loop fault has occurred when only one value is taken.

  However, in the technique of Patent Document 1, it can be detected that a loop in which the program counter value takes only one value is generated, but for example, a loop is generated by taking two or more values. Cannot be detected.

  In the technique of Patent Document 2, if the address (program counter value) changes periodically within a predetermined range, the occurrence of a loop can be detected. However, the occurrence of a loop extending over a wider range than the predetermined range (for example, extending over a plurality of functions) cannot be detected.

Furthermore, in I / O processing or the like, there are cases where the address changes within a fairly narrow range and a loop is generated (repetition of one character input, etc.). There is a risk of erroneously detecting / O processing or the like as a problematic loop.
Japanese Patent Application Laid-Open No. 10-161908 “A microcomputer runaway detection method” Japanese Patent Laid-Open No. 3-292545 “Program Loop Detection and Recovery Device”

  An object of the present invention is to provide a monitoring program, a monitoring method, and a monitoring apparatus that can detect a loop fault occurring in a monitored program with higher accuracy.

  The monitoring program according to the first aspect of the present invention is a monitoring program for causing a computer to execute a process for monitoring a monitored program, obtains an address of the monitored program being executed at a predetermined timing, and stores the acquired address in a buffer. And the number of changes indicating how many different values the values of the plurality of addresses stored in the buffer or the values calculated based on the values of the addresses are within a predetermined threshold. A first determination step of determining whether the number of changes is within the threshold, a step of acquiring a system call issue status of the monitored program, and an issuance of the acquired system call of the monitored program Based on the situation, the monitored program is not currently issuing a system call and the last call A monitoring program that causes the computer to execute a second determination step for determining that a loop failure has occurred in the monitored program when the time does not exist within the time range stored in the buffer. .

  Here, when it is determined in the first determination step that the number of changes is within the threshold value, that is, when a loop state is detected, the last issue time of the system call is further within the range of the time stored in the buffer. Is determined not to exist. As a result, the system call, which is one of the evidences that the monitored program is operating normally, has not been issued even after the time (in the range of the time stored in the buffer) has passed. Thus, it is possible to detect a case in which a loop failure is suspected (a loop failure can be detected).

  In addition, when a loop state is detected, by further determining that the monitored program is not currently issuing a system call, the address changes within a fairly narrow range by I / O processing, etc. It is possible to avoid erroneously detecting a loop occurrence as a loop failure. In addition, by looking at the number of changes in the first determination step, for example, with respect to the higher-order bits of the acquired address, it is possible to detect a loop even over a wide range (for example, spanning multiple functions). In combination with the determination regarding the system call, a loop fault can be detected for a loop extending over the wide range. From the above, it becomes possible to detect a loop fault with higher accuracy.

  The monitoring program according to the second aspect of the present invention is a monitoring program for causing a computer to execute a process for monitoring a monitored program, obtains the address of the monitored program being executed at a predetermined timing, and stores it in a buffer. And the number of changes indicating how many different values the values of the plurality of addresses stored in the buffer or the values calculated based on the values of the addresses are within a predetermined threshold. A first determination step of determining whether the number of changes is within the threshold, a step of acquiring a system call issue status of the monitored program, and an issuance of the acquired system call of the monitored program Based on the situation, the monitored program is not currently issuing a system call and the A second determination step for determining whether an issue time does not exist within a time range stored in the buffer; and the second determination step, wherein the monitored program is not currently issuing a system call; and A loop failure provisional determination step that temporarily determines that a loop failure has occurred when it is determined that the last issue time of the system call is not within the time range stored in the buffer; After the determination is made, a predetermined monitoring duration, a state in which the number of changes is within the threshold value continues, a state in which a system call is not issued continues, and the last issue time of the system call is the buffer If a state that does not exist within the time range stored in the file continues, it is determined that a loop failure has occurred, and a temporary determination of the loop failure is made. From any point in time within a predetermined monitoring duration, the number of changes exceeds the threshold, a system call is being issued, or the last issue time of a system call is the buffer If it is within the range of the time stored in, the monitoring program is characterized by causing the computer to execute a loop failure determination step for determining to cancel the provisional determination of the loop failure.

  Here, when it is determined in the first determination step that the number of changes is within the threshold value, that is, when a loop state is detected, the last issue time of the system call is further within the range of the time stored in the buffer. Is determined not to exist. As a result, the system call, which is one of the evidences that the monitored program is operating normally, has not been issued even after the time (in the range of the time stored in the buffer) has passed. Thus, it is possible to detect a case in which a loop failure is suspected (a loop failure can be detected).

  In addition, when a loop state is detected, by further determining that the monitored program is not currently issuing a system call, the address changes within a fairly narrow range by I / O processing, etc. It is possible to avoid erroneously detecting a loop occurrence as a loop failure. In addition, by looking at the number of changes in the first determination step, for example, with respect to the higher-order bits of the acquired address, it is possible to detect a loop even over a wide range (for example, spanning multiple functions). In combination with the determination regarding the system call, a loop fault can be detected for a loop extending over the wide range.

  Furthermore, since the determination result of the first determination step and the second determination step lasts for a predetermined monitoring duration time, it is determined that a loop failure has been detected. Thus, it is possible to avoid erroneously detecting a normal operation as a loop failure. From the above, it becomes possible to detect a loop fault with higher accuracy.

  The monitoring method according to the third aspect of the present invention is the monitoring method in which the computer executes the process of monitoring the monitored program, and acquires the address of the monitored program being executed at a predetermined timing and stores it in a buffer. And the number of changes indicating how many different values the values of the plurality of addresses stored in the buffer or the values calculated based on the values of the addresses are within a predetermined threshold. A first determination step of determining whether the number of changes is within the threshold, a step of acquiring a system call issue status of the monitored program, and an issuance of the acquired system call of the monitored program Based on the situation, the monitored program is not currently issuing a system call and the last issue time of the system call is If not in the range of times to be stored in Ffa a monitoring method, characterized in that it comprises a second determination step of determining that a loop fault occurs in the monitored program.

  The monitoring method according to the fourth aspect of the present invention is the monitoring method in which the computer executes processing for monitoring the monitored program, and acquires the address of the monitored program being executed at a predetermined timing and stores it in a buffer. And the number of changes indicating how many different values the values of the plurality of addresses stored in the buffer or the values calculated based on the values of the addresses are within a predetermined threshold. A first determination step of determining whether the number of changes is within the threshold, a step of acquiring a system call issue status of the monitored program, and an issuance of the acquired system call of the monitored program Based on the situation, the monitored program is not currently issuing a system call, and the last issue time of the system call is A second determination step for determining whether or not the current time is stored within the time range stored in the buffer; and the second determination step, wherein the monitored program is not currently issuing a system call, and the system call When it is determined that the last issue time is not within the time range stored in the buffer, a loop failure provisional determination step for temporarily determining that a loop failure has occurred and a provisional determination of the loop failure are performed. After that, a state in which the number of changes continues within a predetermined monitoring duration, a state in which a system call is not being issued continues, and the last issue time of the system call is stored in the buffer. If a state that does not exist within the time range continues, it is determined that a loop failure has occurred, and after the provisional determination of the loop failure has been made, the predetermined time is determined in advance. At any point in time within the monitored duration, the number of changes exceeds the threshold value, a system call is being issued, or the last issue time of a system call is stored in the buffer A loop failure determination step for performing a determination to cancel the provisional determination of the loop failure if it exists within a time range.

  The monitoring apparatus according to the fifth aspect of the present invention is the monitoring apparatus that executes the process of monitoring the monitored program, wherein the address of the monitored program being executed is acquired at a predetermined timing and stored in a buffer. The number of changes indicating how many different values the acquisition unit and the values of a plurality of addresses stored in the buffer or values calculated based on the values of these addresses are within a predetermined threshold value A first determination step for determining whether or not the system call issuance status acquisition unit acquires a system call issuance status of the monitored program when the number of changes is within the threshold; and the acquired monitored program Based on the system call issuance status, the monitored program is not currently issuing a system call, and the system call is issued Time is when it is not in a range of times that are stored in the buffer, a monitoring device, characterized in that it comprises a second judging unit judges that the loop fault occurs in the monitored program.

According to the present invention, it is possible to detect the occurrence of a loop fault even for a loop that spans a wide range (for example, spans a plurality of functions).
Further, according to the present invention, it is possible to avoid erroneously detecting a loop fault when an address is changed in a considerably narrow range and a loop is generated by I / O processing or the like.

  Further, according to the present invention, it is possible to avoid erroneously detecting a normal operation as a loop failure only by a long loop state time.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.
FIG. 1 is an overall configuration diagram of a monitoring system common to each embodiment of the present invention.
In FIG. 1, a monitored program (monitored process) 11 is a program whose operation is monitored by a monitoring processing unit (monitoring process) 12 of the present invention. This monitored program is executed on the operating system (OS) 13.

  The OS 13 includes a function (system call issue status monitoring function) 14 for monitoring the issue status of a system call issued by a program executed on the OS 13. The monitoring function 14 stores history information of system calls issued by programs executed on itself. This history includes the ID of the program (process), the issuance time of the system call (time when the OS receives the system call from the program), the status of the system call (terminated (normal end, error end)), or issuing (from the OS ) Or the like).

  The OS 13 also includes a function (program counter value acquisition function) 15 that acquires a program counter value indicating which address on the program executed on itself is currently being executed by the CPU. Upon receiving the program counter value acquisition request from the monitoring processing unit 12, the acquisition function 15 outputs the acquisition request to the monitored program (monitored process) 11. In response to the received acquisition request, the monitored program 11 transmits the address value (program counter value) being executed (fetched) by the CPU at that time to the acquisition function 15. The acquisition function 15 transmits the value of the program counter to the monitoring processing unit 12.

FIG. 2 is a block diagram illustrating a configuration of the monitoring processing unit according to the first embodiment of this invention.
As shown in FIG. 2, the monitoring processing unit 20 acquires the address (program counter value) of the monitored program being executed at a predetermined timing and stores it in the buffer unit 23. Based on the acquired program counter value, a higher-order bit (conversion value) is extracted from the program counter value and written to the corresponding item in the buffer unit 23, and a plurality of addresses stored in the buffer unit 23 Based on the higher-order bits of the address, it is determined whether the number of changes indicating the value of those addresses or the value calculated based on the values of these addresses is within a predetermined threshold The loop detection unit 25 that detects a loop, and when it is determined that the number of changes is within the threshold value, obtains the system call issue status The system call issuance status of the monitored program is obtained via the system 22, and based on the issuance status, the monitored program is not currently issuing a system call, and the last issue time of the system call is the buffer If the loop failure does not exist within the time range stored in the loop program, the loop failure determination unit 26 determines that a loop failure has occurred in the monitored program, and issues an acquisition request for the system call issuance status of the monitored program. It is determined that a loop failure has occurred by the system call issue status acquisition unit 22 and the loop failure determination unit 26 that output to the issue status monitoring function 14 and acquire the system call issue status of the monitored program. An output unit 27 that outputs a message indicating that a loop failure has occurred.

FIG. 3 is a diagram showing an example (part 1) of data stored in the buffer unit 23 of FIG.
As shown in FIG. 3, the buffer unit 23 is a first-in first-out FIFO (First-In First-Out). The buffer unit 23 is a buffer capable of storing a predetermined number of data (9 data in FIG. 3).

  The data stored in the buffer includes time (time when the program counter value is stored in the buffer unit 23), program counter value (PC value), conversion value (upper bit obtained by truncating the lower bit from the program counter value) Value), and each item.

The converted value is obtained, for example, by substituting a value obtained by dividing the program counter value by 1000 into an integer type variable.
Although not certain from FIG. 2, the user designates the following items via a user interface unit (not shown in FIG. 2) attached to the monitoring processing unit 20.
(1) Program counter value acquisition interval (time interval at which the monitoring processing unit 20 acquires the program counter value via the program counter value acquisition function 15 of the OS 13)
(2) Monitor table size (number of data stored in the buffer unit 23)
(3) Number of changes in the program counter value (if the number of different values of the higher-order bits of the program counter value held in the buffer unit 23 is less than this number of changes, the monitored program will loop Threshold to be considered as being in a state)
(4) Monitored program (monitored process)
4A and 4B are flowcharts of monitoring processing executed by the monitoring processing unit according to the first embodiment of this invention.

  First, in FIG. 4A, an initialization process is executed in step S101. In this initialization process, the current time is set as a reference variable that is referred to by the timer monitoring processing unit (not shown in FIG. 2) during timer monitoring, and the area for the monitoring table size described above is used for the monitoring table (buffer). An area is secured on the memory.

  In step S102, timer monitoring processing by the timer monitoring processing unit is executed. In this timer monitoring process, the difference between the value set in the reference variable in step S101 and the current time at the time of step S102 is calculated, and while the difference is shorter than the above-described program counter value acquisition interval, the difference is calculated. Is repeated, and when the difference coincides with the program counter value acquisition interval, the current time is set in the reference variable, and the process proceeds to step S103.

  In step S103, the program counter value acquisition unit 21 acquires the address (program counter value) of the monitored program being executed via the existing function (program counter value acquisition function) 15 of the OS 13, and the buffer unit (monitoring). Table) 23 is added to the end.

  That is, as shown in FIG. 3, the data added to the monitoring table includes items for writing time to the monitoring table (current time set in the reference variable in step S102), program counter value, and conversion value. .

  In step S104, it is determined whether or not the number of data specified for the above-described monitoring table size is stored in the monitoring table. Note that, considering that data is written to the monitoring table at regular intervals (program counter value acquisition interval), it is determined in step S104 whether there is data for a certain time in the monitoring table. You can also.

If it is determined in step S104 that the number of data designated by the monitoring table size is not stored in the monitoring table, the process returns to step S102.
On the other hand, if the number of data specified in the monitoring table size is stored in the monitoring table in step S104, the number of different values of the conversion value on the monitoring table in step S105 is It is determined whether the number of changes (threshold value) of the program counter value is equal to or greater than the above.

  5A, 5B, and 5C are diagrams illustrating examples of data in the monitoring table when the number of changes in the program counter value is set to 4. FIG. In the monitoring table of FIG. 5A, since the conversion value takes four or more different values, it is determined that the monitored program is not in the loop state. In the monitoring table of FIG. 5B, since the conversion value takes three different values, it is determined that the monitored program is in a loop state across several functions, for example. In the monitoring table of FIG. 5C, since the conversion value takes only one value, for example, the monitored program determines that the loop state is in a narrower range than the loop state across several functions in FIG. 5B. Is done.

  If it is determined in step S105 that the number indicating how many different values the conversion value on the monitoring table takes is greater than or equal to the number of changes (threshold value) in the program counter value, the process returns to step S102.

  On the other hand, if it is determined that the number indicating how many different values the conversion value on the monitoring table takes is smaller than the number of changes (threshold value) of the program counter value, in step S106, the system call issue status acquisition unit 22 Thus, a system call issue status acquisition request for the monitored program is output to the system call issue status monitoring function 14 of the OS 13, and the system call issue status of the monitored program is acquired from the monitoring function 14.

  Subsequently, the process proceeds to FIG. 4B, and the loop failure determination unit 26 performs the determination in steps S107, S108, and S109 based on the acquired system call issuance status of the monitored program. First, in step S107, the loop failure determination unit 26 determines whether the monitored program is issuing a system call, that is, waiting for a return instruction from the OS.

  If it is determined in step S107 that the monitored program is issuing a system call, the process returns to step S102. A case where a loop is formed by I / O processing or the like is included in this case.

  On the other hand, if it is determined in step S107 that the monitored program is not issuing a system call, the loop failure determination unit 26 determines in step S108 whether the system call has been issued within the loop time. Note that “within the loop time” means within a time range when data is written to the monitoring table.

  If it is determined in step S108 that a system call has not been issued within the loop time, that is, if it is determined that a system call has been issued before the loop time, the process proceeds to step S110.

  On the other hand, if it is determined in step S108 that the system call has been issued within the loop time, it is determined in step S109 whether or not the system call issued within the loop time has ended normally.

If it is determined in step S109 that the system call issued within the loop time has ended normally, the process returns to step S102.
If it is determined in step S109 that the system call issued within the loop time has ended in error, the process proceeds to step S110.

  Note that the user may be able to specify whether or not to perform the determination in step S109 via the above-described user interface. When the determination in step S109 is not performed, if it is determined in step S108 that a system call has been issued within the loop time, the process immediately returns to step S102.

In step S110, material collection is performed. In collecting data, the following information is output via the output unit 27.
1. A plurality of data held in the buffer unit 33.
2. Information about recently acquired system calls.

  Here, the above 1. Is useful information for specifying the loop range. In addition, the above 2. Is information useful for identifying what the program was processing immediately before the loop processing.

In step S111, a message indicating that a loop failure has occurred is output by the output unit 27, and the process returns to step S102.
Subsequently, a second embodiment of the present invention will be described.

  Among the cases determined as loop faults in the first embodiment, there may be cases in which the loop state (a state in which a number of addresses smaller than the threshold is periodically repeated) is only long and operates normally. . In the second embodiment, such a case is excluded from the determination of the loop fault.

FIG. 6 is a block diagram illustrating a configuration of the monitoring processing unit according to the second embodiment of this invention.
As shown in FIG. 6, the monitoring processing unit 30 acquires a program counter value acquisition unit 31 that acquires an address (program counter value) of a monitored program being executed at a predetermined timing and stores it in the buffer unit 33. Based on the acquired program counter value, a higher-order bit (conversion value) is extracted from the program counter value and written in the corresponding item of the buffer unit 33, and a plurality of addresses stored in the buffer unit 33 Based on the higher-order bits of the address, it is determined whether the number of changes indicating the value of those addresses or the value calculated based on the values of these addresses is within a predetermined threshold Thus, when it is determined that the number of changes is within the threshold value, the loop detection unit 36 that detects a loop detects the system call issue status. The system call issuance status of the monitored program is acquired via the unit 32, and based on the issuance status, the monitored program is not currently issuing a system call, and the last issue time of the system call is The temporary failure determination unit 37 for determining whether or not the time is stored within the time range stored in the buffer unit 33, and the system call issue status monitoring function 14 of the OS 13 sends an acquisition request for the system call issue status of the monitored program. A loop failure determination unit 35 including a system call issue status acquisition unit 32, a loop detection unit 36, and a loop failure temporary determination unit 37 that outputs and acquires the system call issue status of the monitored program from the monitoring function 14. When it is determined that a loop failure has occurred, a message indicating that a loop failure has occurred Output unit 38 which force, when the loop fault by the loop fault determining unit 35 is determined to have occurred, a buffer size changing unit 39 for changing the buffer size of the buffer portion 33 to a proper value.

  Here, the loop failure determination unit 35 determines whether the monitored program is not currently issuing a system call and the last issue time of the system call is stored in the buffer unit 33 by the temporary loop failure determination unit 37. When it is determined that it does not exist within the range, a provisional determination is made that a loop failure has occurred, a timer for continuing monitoring is started, and a provisional determination is made after the provisional determination of the loop failure is made. A state in which the monitoring duration and the number of changes are within a threshold value continue, a state in which a system call is not being issued continues, and a system call last issue time does not exist within the time range stored in the buffer unit 33 If it continues, it is determined that a loop failure has occurred, and after the provisional determination of the loop failure has been made, the change is made at any point within a predetermined monitoring duration. If the number exceeds the threshold, the system call is being issued, or the last issue time of the system call is within the time range stored in the buffer unit 33, the loop failure The determination that cancels the provisional determination is performed.

  When the loop failure tentative determination is canceled by the loop failure determination unit 35, it is considered that the buffer size of the buffer unit 33 is not large enough to perform the loop failure determination. In other words, the loop failure determination unit 35 determines whether the number of changes exceeds the threshold value, the system call is being issued, or the last issue time of the system call is within the time range stored in the buffer unit 33. The buffer size changing unit 39 is notified of the time that was present. The buffer size changing unit 39 calculates the increment of the number of stages of the buffer unit 33 by dividing the notified time value by the timing (time interval) for acquiring the address, and notifies the buffer unit 33 of it. The buffer size of the buffer unit 33 is changed.

Although not certain from FIG. 6, the user designates the following items via a user interface unit (not shown in FIG. 6) attached to the monitoring processing unit 30.
(1) Program counter value acquisition interval (time interval at which the monitoring processor 30 acquires the program counter value via the program counter value acquisition function 15 of the OS 13)
(2) Monitor table size (number of data stored in buffer 33)
(3) Number of changes in program counter value (If the number of different values of the higher-order bits of the program counter value held in the buffer unit 33 is less than this number of changes, the monitored program will loop Threshold to be considered as being in a state)
(4) Monitored program (monitored process)
(5) Monitoring continuation time (If it is determined that a loop failure has occurred and the situation at the time of occurrence continues for the monitoring continuation time, it is determined that a loop failure has occurred.)
7A, 7B, and 7C are flowcharts of monitoring processing executed by the monitoring processing unit according to the second embodiment of this invention.

  First, in FIG. 7A, an initialization process is executed in step S201. In this initialization processing, the current time is set as a reference variable that is referred to by the timer monitoring processing unit (not shown in FIG. 6) during timer monitoring, and the monitoring continuation flag indicating whether the monitoring duration is in progress is set to “monitoring continuation”. An area corresponding to the above-described monitoring table size is secured on the memory as a monitoring table (buffer) area.

  In step S202, timer monitoring processing by the timer monitoring processing unit is executed. In this timer monitoring process, the difference between the value set in the reference variable in step S201 and the current time at the time of step S202 is calculated, and while the difference is shorter than the above-described program counter value acquisition interval, the difference is calculated. When the difference is equal to the program counter value acquisition interval, the current time is set in the reference variable, and the process proceeds to step S203.

  In step S203, the program counter value acquisition unit 21 acquires the address (program counter value) of the monitored program being executed via the existing function (program counter value acquisition function) 15 of the OS 13, and the buffer unit (monitoring). Table) 23 is added to the end.

  In other words, the data added to the monitoring table includes items for writing time to the monitoring table (current time set in the reference variable in step S202), program counter value, and conversion value.

  In step S204, it is determined whether or not the number of data specified for the above-described monitoring table size is stored in the monitoring table. Note that, considering that data is written to the monitoring table at regular intervals (program counter value acquisition interval), it is determined in step S104 whether there is data for a certain time in the monitoring table. You can also.

If it is determined in step S204 that the number of data specified in the monitoring table size is not stored in the monitoring table, the process returns to step S102.
On the other hand, if the number of data specified in the monitoring table size is stored in the monitoring table in step S204, the number indicating how many values the conversion value on the monitoring table takes in step S205 is as follows. It is determined whether the number of changes (threshold value) of the program counter value is equal to or greater than the above.

  If it is determined in step S205 that the number indicating the number of conversion values on the monitoring table is greater than or equal to the number of changes (threshold value) in the program counter value, the process proceeds to step S214.

  On the other hand, if it is determined that the number indicating the conversion value on the monitoring table is smaller than the number of changes (threshold value) of the program counter value, the system call issuance status acquisition unit 22 in step S206. The system call issue status acquisition request of the monitored program is output to the system call issue status monitoring function 14 of the OS 13, and the system call issue status of the monitored program is acquired from the monitoring function 14.

  Subsequently, the process proceeds to FIG. 7B, and the loop failure determination unit 26 performs determinations in steps S207, S208, and S209 based on the acquired system call issuance status of the monitored program. First, in step S207, the loop failure determination unit 26 determines whether the monitored program is issuing a system call, that is, waiting for a return instruction from the OS.

  If it is determined in step S207 that the monitored program is issuing a system call, the process proceeds to step S214. A case where a loop is formed by I / O processing or the like is included in this case.

  On the other hand, if it is determined in step S207 that the monitored program is not issuing a system call, the loop failure determination unit 26 determines in step S208 whether the system call has been issued within the loop time. Note that “within the loop time” means within a time range when data is written to the monitoring table.

  If it is determined in step S208 that a system call has not been issued within the loop time, that is, if it is determined that a system call has been issued before the loop time, the process proceeds to step S210.

  On the other hand, if it is determined in step S208 that the system call has been issued within the loop time, in step S209, the value of the monitoring continuation flag is a value of “continuing monitoring” or monitoring continuation. It is determined whether the system call issued within the loop time has ended normally.

  In step S209, if the system call issued within the loop time has ended normally, or if the value of the monitoring continuation flag is “continuing monitoring”, the process proceeds to step S214. .

  If it is determined in step S209 that the system call issued within the loop time has ended in error and the value of the monitoring continuation flag is not “monitoring is continuing”, the process proceeds to step S210.

  Note that the user may be able to specify whether or not to perform the determination in step S209 via the above-described user interface. If the determination in step S209 is not performed, if it is determined in step S208 that a system call has been issued within the loop time, the process immediately proceeds to step S214.

  In step S <b> 210, it is determined whether the value of the monitoring continuation flag is a value “monitoring is ongoing”. When the control of step S210 is entered for the first time, the value of the monitoring continuation flag is initialized to the value “not monitoring continuation”. Therefore, in step S210, the value of the monitoring continuation flag is set to “monitoring continuation”. It is determined that the value is not “continuing”, and the process proceeds to the monitoring continuation state in step S211. That is, the loop failure determination unit 35 starts a timer for continuing monitoring, and changes the value of the monitoring continuation flag to a value of “continuing monitoring”.

In step S212, material collection is performed. In data collection, the following information is output via the output unit 38.
1. A plurality of data held in the buffer unit 33.
2. Information about recently acquired system calls.

  Here, the above 1. Is useful information for specifying the loop range. In addition, the above 2. Is information useful for identifying what the program was processing immediately before the loop processing.

In step S213, a message indicating that a loop failure has occurred is output by the output unit 38, and the process returns to step S202.
On the other hand, if it is determined in step S210 that the value of the monitoring continuation flag is the value “continuing monitoring”, the process immediately returns to step S202. This is to avoid repeatedly outputting the message “A loop failure has occurred” in the monitoring continuation state.

In FIG. 7C, first, in step S214, it is determined whether or not the value of the monitoring continuation flag is “monitoring is ongoing”.
Step S214 is a step in which control is passed from steps S205, S207, and S209. Therefore, here, if it is determined that the value of the monitoring continuation flag is the value of “continuing monitoring”, the condition at the time when the loop fault is tentatively determined in the monitoring continuation state is It is no longer satisfied, that is, the number of changes exceeds the threshold value, the system call is being issued, or the last issue time of the system call is within the time range stored in the buffer unit 33 I mean.

  In this case, in step S215, a transition from the monitoring continuation state to the non-monitoring continuation state is performed. That is, the loop failure determination unit 35 outputs the value of the timer for continuing monitoring at that time to the buffer size changing unit 39, and the value of the monitoring continuation flag is “not continuously monitoring”. Changed to a value.

  Subsequently, in step S216, the buffer size changing unit 39 divides the notified time value by the address acquisition timing (program counter value acquisition interval), thereby calculating the number of stages of the buffer unit 33. Thus, the buffer size of the buffer unit 33 is changed by notifying the buffer unit 33.

If it is determined in step S214 that the value of the monitoring continuation flag is not the value of “continuing monitoring”, the process returns to step S202.
FIG. 8 is a diagram illustrating an example of a storage medium.

  The monitoring process of the present invention can be realized by the information processing apparatus 61. The program and data for the processing of the present invention can be loaded from the storage device 65 of the information processing device 61 to the memory of the information processing device 61 and executed, or loaded from the portable storage medium 63 to the memory of the information processing device 61. It is also possible to execute it by loading it into the memory of the information processing device 61 from the external storage device 62 via the network 66.

The present invention may have the following configuration.
(Supplementary Note 1) In a monitoring program that causes a computer to execute processing for monitoring a monitored program,
Acquiring the address of the monitored program being executed at a predetermined timing and storing it in a buffer;
It is determined whether the number of changes indicating how many different values the values of a plurality of addresses stored in the buffer or values calculated based on the values of these addresses are within a predetermined threshold value. A first determination step;
Obtaining a system call issue status of the monitored program when the conversion value is within the threshold; and
Based on the acquired system call issuance status of the monitored program, the monitored program is not currently issuing a system call, and the last issue time of the system call is within the time range stored in the buffer. A monitoring program that causes the computer to execute a second determination step for determining that a loop failure has occurred in the monitored program when the program is not present in the monitoring program.
(Supplementary Note 2) In the second determination step, even if the monitored program is not currently issuing a system call and the last issue time of the system call is not within the time range stored in the buffer, The monitoring program according to claim 1, wherein if the system call issued by the monitored program has been normally terminated within the time range stored in the buffer, it is determined that no loop failure has occurred. .
(Additional remark 3) In the monitoring program which makes a computer perform the process which monitors a to-be-monitored program,
Obtaining an address of the monitored program being executed at a predetermined timing (time interval) and storing it in a buffer;
It is determined whether the number of changes indicating how many different values the values of a plurality of addresses stored in the buffer or values calculated based on the values of these addresses are within a predetermined threshold value. A first determination step;
Obtaining a system call issue status of the monitored program when the conversion value is within the threshold; and
Based on the obtained system call issuance status of the monitored program, the monitored program is not currently issuing a system call, and the last issue time of the system call is stored in the buffer. A second determination step of determining whether or not it exists in
When it is determined in the second determination step that the monitored program is not currently issuing a system call and the last issue time of the system call does not exist within the time range stored in the buffer A loop failure provisional determination step that temporarily determines that a loop failure has occurred;
After the provisional determination of the loop failure, a state in which the conversion value is within the threshold continues for a predetermined monitoring duration, a state in which a system call is not being issued continues, and a system call is last issued If the time does not exist within the time range stored in the buffer, it is determined that a loop failure has occurred,
The conversion value exceeds the threshold value, or a system call is being issued at any point in time within a predetermined monitoring duration after the provisional determination of the loop failure, or If the last issue time of the system call is within the range of the time stored in the buffer, the computer is caused to execute a loop failure determination step for determining to cancel the provisional determination of the loop failure. Feature monitoring program.
(Additional remark 4) In the said loop failure determination step, when the determination which cancels the said temporary determination is performed, the number of changes exceeds a threshold value, it will be in the state which is issuing the system call, or the last issue of a system call Stage time increment calculation for calculating the increment of the stage number of the buffer by dividing the time value by the timing (time interval) for acquiring the address, whether the time exists within the time range stored in the buffer Steps,
The monitoring program according to claim 3, further comprising a buffer size changing step of increasing the number of stages of the buffer by the calculated increment.
(Supplementary Note 5) In a monitoring method in which a computer executes processing for monitoring a monitored program,
Acquiring the address of the monitored program being executed at a predetermined timing and storing it in a buffer;
It is determined whether the number of changes indicating how many different values the values of a plurality of addresses stored in the buffer or values calculated based on the values of these addresses are within a predetermined threshold value. A first determination step;
Obtaining a system call issue status of the monitored program when the conversion value is within the threshold; and
Based on the acquired system call issuance status of the monitored program, the monitored program is not currently issuing a system call, and the last issue time of the system call is within the time range stored in the buffer. A monitoring method comprising: a second determination step for determining that a loop failure has occurred in the monitored program when the program is not present in the monitoring program.
(Supplementary Note 6) In the second determination step, even when the monitored program is not currently issuing a system call and the last issue time of the system call is not within the time range stored in the buffer The monitoring according to appendix 5, wherein when the system call issued by the monitored program is normally terminated within the time range stored in the buffer, it is determined that no loop failure has occurred. Method.
(Supplementary Note 7) In a monitoring method in which a computer executes processing for monitoring a monitored program,
Acquiring the address of the monitored program being executed at a predetermined timing and storing it in a buffer;
It is determined whether the number of changes indicating how many different values the values of a plurality of addresses stored in the buffer or values calculated based on the values of these addresses are within a predetermined threshold value. A first determination step;
Obtaining a system call issue status of the monitored program when the conversion value is within the threshold; and
Based on the obtained system call issuance status of the monitored program, the monitored program is not currently issuing a system call, and the last issue time of the system call is stored in the buffer. A second determination step of determining whether or not it exists in
When it is determined in the second determination step that the monitored program is not currently issuing a system call and the last issue time of the system call does not exist within the time range stored in the buffer A loop failure provisional determination step that temporarily determines that a loop failure has occurred;
After the provisional determination of the loop failure, a state in which the conversion value is within the threshold continues for a predetermined monitoring duration, a state in which a system call is not being issued continues, and a system call is last issued If the time does not exist within the time range stored in the buffer, it is determined that a loop failure has occurred,
The conversion value exceeds the threshold value, or a system call is being issued at any point in time within a predetermined monitoring duration after the provisional determination of the loop failure, or And a loop failure determination step for determining whether to cancel the provisional determination of the loop failure when the last issue time of the system call is within the range of the time stored in the buffer. Method.
(Supplementary Note 8) In the loop failure determination step, when the determination to cancel the provisional decision is made, whether the number of changes exceeds a threshold value, a system call is being issued, or a system call is last issued Stage time increment calculation for calculating the increment of the stage number of the buffer by dividing the time value by the timing (time interval) for acquiring the address, whether the time exists within the time range stored in the buffer Steps,
The monitoring method according to claim 7, further comprising a buffer size changing step of increasing the number of stages of the buffer by the calculated increment.
(Additional remark 9) In the monitoring apparatus which performs the process which monitors a to-be-monitored program,
An address acquisition unit that acquires an address of the monitored program being executed at a predetermined timing and stores the acquired address in a buffer;
It is determined whether the number of changes indicating how many different values the values of a plurality of addresses stored in the buffer or values calculated based on the values of these addresses are within a predetermined threshold value. A first determination step;
A system call issuance status acquisition unit for acquiring a system call issuance status of the monitored program when the conversion value is within the threshold;
Based on the acquired system call issuance status of the monitored program, the monitored program is not currently issuing a system call, and the last issue time of the system call is within the time range stored in the buffer. A monitoring apparatus comprising: a second determination unit that determines that a loop failure has occurred in the monitored program when it does not exist in the monitoring program.
(Supplementary Note 10) Even if the monitored program is not currently issuing a system call and the last issue time of the system call does not exist within the time range stored in the buffer, the second determination unit The monitoring according to appendix 9, wherein if the system call issued by the monitored program has been normally terminated within the time range stored in the buffer, it is determined that no loop failure has occurred. apparatus.
(Additional remark 11) In the monitoring apparatus which performs the process which monitors a to-be-monitored program,
An address acquisition unit for acquiring an address of the monitored program being executed at a predetermined timing and storing it in a buffer;
It is determined whether the number of changes indicating how many different values the values of a plurality of addresses stored in the buffer or values calculated based on the values of these addresses are within a predetermined threshold value. A first determination unit;
A system call issuance status acquisition unit for acquiring a system call issuance status of the monitored program when the conversion value is within the threshold;
Based on the obtained system call issuance status of the monitored program, the monitored program is not currently issuing a system call, and the last issue time of the system call is stored in the buffer. A second determination unit for determining whether or not it exists,
When the second determination unit determines that the monitored program is not currently issuing a system call and the last issue time of the system call does not exist within the time range stored in the buffer A loop failure provisional determination unit that temporarily determines that a loop failure has occurred;
After the provisional determination of the loop failure, a state in which the conversion value is within the threshold continues for a predetermined monitoring duration, a state in which a system call is not being issued continues, and a system call is last issued If the time does not exist within the time range stored in the buffer, it is determined that a loop failure has occurred,
The conversion value exceeds the threshold value, or a system call is being issued at any point in time within a predetermined monitoring duration after the provisional determination of the loop failure, or A monitoring system comprising: a loop failure determination unit that performs a determination to cancel the provisional determination of the loop failure when the last issue time of the system call is within the range of the time stored in the buffer. apparatus.
(Additional remark 12) When the determination which cancels the said tentative decision is performed by the said loop failure determination part, the number of changes exceeds a threshold value, it will be in the state which is issuing the system call, or the last issue of a system call Stage time increment calculation for calculating the increment of the stage number of the buffer by dividing the time value by the timing (time interval) for acquiring the address, whether the time exists within the time range stored in the buffer And
The monitoring apparatus according to claim 11, further comprising a buffer size changing unit that increases the number of stages of the buffer by the calculated increment.

1 is an overall configuration diagram of a monitoring system common to each embodiment of the present invention. It is a block diagram which shows the structure of the monitoring process part of 1st Embodiment of this invention. It is a figure which shows the example of data (the 1) stored in the buffer part 23 of FIG. It is a flowchart (the 1) of the monitoring process performed by the monitoring process part of 1st Embodiment of this invention. It is a flowchart (the 2) of the monitoring process performed by the monitoring process part of 1st Embodiment of this invention. It is a figure which shows the example of data when it determines with the group of the address stored in the buffer part of FIG. 2 not being in a loop state. FIG. 6 is a diagram (part 1) illustrating an example of data when it is determined that the set of addresses stored in the buffer unit in FIG. 2 is in a loop state; FIG. 6 is a diagram (part 2) illustrating an example of data when it is determined that the group of addresses stored in the buffer unit in FIG. 2 is in a loop state. It is a block diagram which shows the structure of the monitoring process part of 2nd Embodiment of this invention. It is a flowchart (the 1) of the monitoring process performed by the monitoring process part of 2nd Embodiment of this invention. It is a flowchart (the 2) of the monitoring process performed by the monitoring process part of 2nd Embodiment of this invention. It is a flowchart (the 3) of the monitoring process performed by the monitoring process part of 2nd Embodiment of this invention. It is a figure which shows the example of a storage medium.

Explanation of symbols

11 Monitored program (monitored process)
12 Monitoring processor (monitoring program)
13 Operating System 14 System Call Issuance Status Monitoring Function 15 Program Counter Value Acquisition Function 20, 30 Monitor Processing Unit 21, 31 Program Counter Value Acquisition Unit 22, 32 System Call Issuance Status Acquisition Unit 23, 33 Buffer Unit (FIFO)
24, 34 Conversion value calculation unit 25, 36 Loop detection unit 26, 35 Loop failure determination unit 27, 38 Output unit 37 Loop failure temporary determination unit 39 Buffer size change unit

Claims (5)

In a monitoring program for causing a computer to execute processing for monitoring a monitored program,
An acquisition request for the address of the monitored program being executed is transmitted to the monitored program being executed on the operating system on which the monitoring program is executed at a predetermined timing, and based on the acquisition request Acquiring the address transmitted by the monitored program and storing it in a buffer, and storing the storage time when the address is stored in the buffer ;
It is determined whether the number of changes indicating how many different values the values of a plurality of addresses stored in the buffer or values calculated based on the values of these addresses are within a predetermined threshold value. A first determination step;
Obtaining a system call issue status of the monitored program via the operating system when the number of changes is within the threshold;
Based on the system call issuance status of the monitored program acquired, the monitored program is not currently issuing a system call, and the last issue time of the system call is the storage time stored in the buffer. A monitoring program that causes the computer to execute a second determination step for determining that a loop failure has occurred in the monitored program when the value is not within the range. In a monitoring program for causing a computer to execute processing for monitoring a monitored program,
An acquisition request for the address of the monitored program being executed is transmitted to the monitored program being executed on the operating system on which the monitoring program is executed at a predetermined timing, and based on the acquisition request Acquiring the address transmitted by the monitored program and storing it in a buffer, and storing the storage time when the address is stored in the buffer ;
It is determined whether the number of changes indicating how many different values the values of a plurality of addresses stored in the buffer or values calculated based on the values of these addresses are within a predetermined threshold value. A first determination step;
Obtaining a system call issue status of the monitored program via the operating system when the number of changes is within the threshold;
The storage in which the monitored program is not currently issuing a system call and the last issue time of the system call is stored in the buffer based on the acquired system call issue status of the monitored program a second judgment step of judging whether the range of time,
In the second determination step, it is determined that the monitored program is not currently issuing a system call and that the last issue time of the system call is not within the range of the storage time stored in the buffer. A loop fault provisional determination step that temporarily determines that a loop fault has occurred,
After the provisional determination of the loop fault, a state in which the predetermined monitoring duration, the number of changes is within the threshold continues, a state in which a system call is not issued continues, and a system call is last issued If the time is not within the range of the storage time stored in the buffer, it is determined that a loop failure has occurred,
After the provisional determination of the loop failure, at any point within a predetermined monitoring duration, the number of changes exceeds the threshold, a system call is being issued, or when it becomes in the range of the storage time the last issue time of the system call are stored before Symbol buffer, it is run loop fault determining step for determining to cancel the provisional determination of the loop failure, to the computer A monitoring program characterized by In a monitoring method in which a computer executes processing for monitoring a monitored program,
An acquisition request for the address of the monitored program being executed is transmitted to the monitored program being executed on the operating system on which the monitoring program is executed at a predetermined timing, and based on the acquisition request Acquiring the address transmitted by the monitored program and storing it in a buffer, and storing the storage time when the address is stored in the buffer ;
It is determined whether the number of changes indicating how many different values the values of a plurality of addresses stored in the buffer or the values calculated based on the values of the addresses are within a predetermined threshold. A first determination step;
Obtaining a system call issue status of the monitored program via the operating system when the number of changes is within the threshold;
Based on the acquired system call issue status of the monitored program, the monitored program is not currently issuing a system call, and the last issue time of the system call is the storage time stored in the buffer. A monitoring method comprising: a second determination step for determining that a loop failure has occurred in a monitored program when it is not within the range. In a monitoring method in which a computer executes processing for monitoring a monitored program,
An acquisition request for the address of the monitored program being executed is transmitted to the monitored program being executed on the operating system on which the monitoring program is executed at a predetermined timing, and the acquisition request is transmitted based on the acquisition request. Acquiring the address transmitted by the monitored program and storing it in a buffer, and storing the storage time when the address is stored in the buffer in the buffer ;
It is determined whether the number of changes indicating how many different values the values of a plurality of addresses stored in the buffer or values calculated based on the values of these addresses are within a predetermined threshold value. A first determination step;
Obtaining a system call issue status of the monitored program via the operating system when the number of changes is within the threshold;
The storage in which the monitored program is not currently issuing a system call and the last issue time of the system call is stored in the buffer based on the acquired system call issue status of the monitored program a second judgment step of judging whether the range of time,
In the second determination step, it is determined that the monitored program is not currently issuing a system call and that the last issue time of the system call is not within the range of the storage time stored in the buffer. A loop fault provisional determination step that temporarily determines that a loop fault has occurred,
After the provisional determination of the loop fault, a state in which the predetermined monitoring duration, the number of changes is within the threshold continues, a state in which a system call is not issued continues, and a system call is last issued If the time is not within the range of the storage time stored in the buffer, it is determined that a loop failure has occurred,
After the provisional determination of the loop failure, at any point within a predetermined monitoring duration, the number of changes exceeds the threshold, a system call is being issued, or If the last issue time of the system call is made within the range of the storage time stored in the buffer, monitored, characterized in that it comprises a loop fault determination step of determining to cancel the provisional determination of the loop fault Method. In a monitoring device that executes processing for monitoring a monitored program,
A buffer for storing a predetermined number of addresses of the monitored program being executed on an operating system on which the monitoring program is executed, together with a storage time at which the address is stored;
The running transmitted at a predetermined timing a request for obtaining the address of the monitoring program, the acquisition request to the address acquisition unit which the monitoring program is stored in the buffer by obtaining the address transmitted on the basis of When,
It is determined whether the number of changes indicating how many different values the values of a plurality of addresses stored in the buffer or values calculated based on the values of these addresses are within a predetermined threshold value. A first determination unit;
A system call issuance status acquisition unit that acquires a system call issuance status of the monitored program via the operating system when the number of changes is within the threshold;
The storage time at which the monitored program is not currently issuing a system call and the last issue time of the system call is stored in the buffer based on the acquired system call issuance status of the monitored program A monitoring apparatus comprising: a second determination unit that determines that a loop failure has occurred in the monitored program when the value is not within the range.

Images