JPH09212373A - Process monitor device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To make it possible to speedily set an optimum sleep time and dynamically and suitably vary it by providing a process sleep table. SOLUTION: A process 1 informs a monitor process 2 of the sleep time. The monitor process 2 sets the reported sleep time of the process 1 in the process sleep table 3. Then, corresponding sleep, swap-out, swap-in, and wake-up processes of the process 1 are performed in order according to the set sleep time. At this time, the monitor process 2 dynamically varies the sleep time in the table 3 in response to the information from the process 1. Further, the log before the process 1 or monitor process 2 enters a sleep state, process by process, and the log after waking up are gathered and the monitor process 2 totalizes statistical information on the sleep time by the processes according to the logs to display the sleep and execution states.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a process monitoring device for monitoring a sleep state of a process.

[0002]

2. Description of the Related Art Conventionally, when testing or debugging a device having a built-in computer and a plurality of processes collaborating to perform processing, a sleep time of a certain process is set and the processing is executed so that the sleep time is appropriate. When I tried to change it instead, I changed the sleep time, restarted and changed the sleep time, and repeated device testing and debugging.

[0003]

As described above, conventionally, in the case of an apparatus in which a plurality of processes cooperate to perform processing,
When I tried to change the sleep time of the process, I changed the sleep time and then restarted and changed the sleep time for testing and debugging to set the optimum sleep time. There was a problem that it was difficult to ask for.

There is also a problem that the sleep time of the process cannot be dynamically changed at the time of testing or debugging. In addition, even if a process that was once in the sleep state is in the active state after the sleep time elapses, the process may be swapped out, so it cannot be executed immediately, and execution start is delayed until swapped in. There was a problem of being lost.

There is also a problem that it is difficult to collect statistical information of state transitions and set an optimal sleep time because logs of various events when a process enters a sleep state are not collected.

[0006] The present invention solves these problems,
Providing a process sleep table to set or dynamically change the sleep time of the process and collect logs to set the optimum sleep time quickly or dynamically change the optimum sleep time. The purpose is to realize the presentation of statistical information for setting.

[0007]

Means for solving the problem will be described with reference to FIG. In FIG. 1, a process 1 performs various kinds of processing, and here, a sleep time is set, and the monitoring process 2 sleeps or wakes up.

The monitoring process 2 monitors the process 1 and puts it to sleep or wakes up. The process sleep table 3 sets a sleep time for each process.

The process end table 4 is for setting that the process 1 has completed the processing. Log data 5
Is a collection of various log data of process 1.

Next, the operation will be described. In response to process 1 notifying monitor process 2 of the sleep time,
The monitoring process 2 sets the notified sleep time of the process 1 in the process sleep table 3, and the process 1 executes sleep, swap-out, swap-in, or wake-up corresponding processing sequentially according to the set sleep time. I have to.

At this time, the monitoring process 2 dynamically changes the sleep time in the process sleep table 3 in response to the notification from the process 1. Further, the monitoring process 2 dynamically changes the sleep time of the sleeping process 1 set in the process sleep table 3 in response to the notification of the completion time or the completion rate of the process of the other process 1. There is.

Also, the process 1 or the monitoring process 2
Collects the log before going into sleep state and the log after waking up for each process, and the monitoring process 2 collects the statistical information of the sleep time for each process based on the log and monitors the sleep and execution states. I try to make a presentation.

Therefore, the process sleep table 3 is provided, and the sleep time from the process 1 is set or dynamically changed in the process sleep table 3 and the log is collected to quickly set the optimum sleep time or dynamically change the optimum sleep time. In addition, it is possible to present statistical information for setting the optimum sleep time.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, embodiments and operations of the present invention will be sequentially described in detail with reference to FIGS.

FIG. 2 shows a flowchart for explaining the operation of the present invention. This is a detailed description of the operation of the configuration of FIG. Here, the process P1 is an example of the illustrated process, and the monitoring process 2 is the monitoring process 2 of FIG.

In FIG. 2, S1 starts a monitoring process. S2 is initialized. This creates various initial settings, here, the process sleep table 3, for example, the framework of the process sleep table 3 of FIG. 5E described later (a table in which various values are not set).

On the other hand, on the process P1 side, S21 starts the process P1. In S22, the sleep is notified to the monitoring process 2. This is because the process P1 has finished processing in S21, so the monitoring process 2 is notified in S22 that the own process is in a sleep state.

In S3, the monitoring process 2 sets the data notified in S22 in the process sleep table 3, for example, as shown in (e) of FIG. 5 to be described later: -Process name: P1-Sleep start time: 00 Time 00 minutes-Sleep end time: 00: 04-Deep sleep start time: 00: 01-Set deep sleep end time: 00:03. Here, the sleep start time and the sleep end time are the time when the process enters the sleep state and the time when the process enters the operation state from the sleep state. The deep sleep start time is a time at which a process swap-out (a program in the main memory is saved to the external storage device) is started. The deep sleep end time is the time to start swap-in (restore the program in the external storage device to the main storage) of the process.

In step S4, it is determined whether it is the start time of deep sleep.
If YES, the process P1 is started to be swapped out in S5 (the program of the process P1 in the main memory is saved in the external storage device), and the process proceeds to S6. In the case of NO,
Proceed to S6.

In step S6, it is determined whether it is the end time of deep sleep.
In the case of YES, the process P1 is started to be restored in the memory in S7 (swap-in is started and restored in the main memory of the process P1 in the external storage device), and the process proceeds to S8. In the case of NO, the process proceeds to S8.

In step S8, it is determined whether it is the sleep end time. Y
In the case of ES, the process P1 is notified of wake-up in S9. Then, the data is deleted in S10, and the process returns to S4. On the other hand, in step S23, the process P1 that receives the wakeup notification from the monitoring process 2 in step S9 wakes up (starts up).

At step S24, the operation is started. As described above, the sleep time (sleep start time, sleep end time, deep sleep start time, deep sleep end time shown in (e) of FIG. 5) is set in the process sleep table 3, and the set sleep start time is obtained. When the deep sleep start time is reached, the relevant process is started to sleep, and when the deep sleep start time is reached, the process is swapped out, when the deep sleep end time is reached, the process is swapped in, and when the sleep end time is reached, the wakeup (start) is performed It becomes possible.

When the process P1 is put to sleep, the process P1 is notified of sleep in S11 of FIG. Then, in step S25, the process P1 sleeps, and in step S4
Return to and repeat. As a result, the monitoring process 2 puts the process P1 in the sleep state, and according to the process sleep table 3, the processes as described above can be performed according to the steps S4 to S10.

FIG. 3 is an explanatory diagram of the state transition of the process of the present invention. Here, the thick line represents the process execution state, and the dotted line represents the process sleep state.

In FIG. 3, (a) makes a state transition from execution to sleep. (A) starts a light sleep.
(C) ends a light sleep.

In step (d), a deep sleep is started, and swapping out of the process (saving the program of the process on the main memory to the external storage device and increasing the free area on the main memory) is started.

(E) ends the swap-out.
(F) ends a deep sleep. (K) starts a light sleep and starts restoration of the process to the memory (swap-in, that is, restoration of the program of the process from the external storage device to the main storage).

(H) ends the restoration to the memory. (K)
Exits sleep, wakes up, and the process begins running. As described above, the process sleeps and enters a deep sleep after a lapse of a predetermined time, swaps out to increase the free space in the main memory, and swaps in before the wake-up a predetermined time to execute the process program. A series of processes of immediately starting the program of the process in the main memory and starting the execution when the program is restored to the memory and awakened to a light sleep are automatically executed according to the process sleep table 3. Become.

FIG. 4 shows a flowchart for changing the sleep time according to the present invention. Here, the process P2 is sleeping according to the sleep time (sleep start time, sleep end time, deep sleep start time, deep sleep end time) set in the process sleep table 3 of the process P1, and the process P2 is in the sleep time of the process P1. Is a procedure for changing.

In FIG. 4, in step S31, the process P1 is set to the process sleep table 3 at the time "00:00".
Follow to sleep. In S32, the process P2 notifies the monitoring process 2 that the process P1 is to be waked up after 4 minutes at the time “00:00”. Correspondingly, the monitoring process 2 is set to wake up the process P1 of the process sleep table 3 in S33 after 4 minutes, that is, • sleep end time “00:04” • deep sleep start time “00:01” Minutes "・ Deep sleep end time" 00:03 "Set as (1). Then, the process proceeds to (A) of FIG.

In step S34, the process P2 waits for the time "00: 0".
1 minute "notifies the monitoring process 2 that the process P1 is to be woken up in 2 minutes. Correspondingly, the monitoring process 2 is set to wake up the process P1 in the process sleep table 3 in 2 minutes after 2 minutes. ,
That is, the sleep end time is "00:03", the deep sleep start time is "00:01", and the deep sleep end time is "00:02" (2). Then, the process proceeds to (A) of FIG. here,
According to (2) of the final process sleep table 3,
When it becomes "00:02", the deep sleep ends, the swap-in is performed, and the process P1 is started by wake up at "00:03" to start the operation. As a result, “00” in the process sleep table 3 of (1)
Although the process P1 is scheduled to wake up at “04:”, the process P1 can be waked up at the time dynamically changed to “00:03” by S34 and S35.

FIG. 5 shows an execution status diagram of the process P2 of the present invention. This is a schematic representation of the state when the sleep time in FIG. 4 described above is dynamically changed. FIG. 5A shows the situation of 00:00. This schematically shows the situation when the process sleep table 3 is set as shown in (1) by S32 and S33 of FIG. 4 described above. Here, a thick black line represents execution, and a white double line represents sleep.

00:00: Sleep start 00:04: Wakeup start FIG. 5 (b) shows the situation of 00:00. this is,
FIG. 6 is a schematic diagram showing the situation when the process sleep table 3 is set as shown in (2) by S34 and S35 of FIG. 4 described above. Here, a thick black line represents execution, and a white double line represents sleep.

00:01: sleep start 00:03: wakeup start FIG. 5C shows the execution status of the process P2. This is the execution status when the process P2 executes and changes so that the process is just handed over to the process P1 at the end of execution. That is, when the process P2 starts executing from 00:00 and reaches 00:00, it is determined that the expected execution end time is 00:03. Therefore, the process P1 that takes over the process at 00:00 is awakened. It is set in the process sleep table 3 so as to perform.

FIG. 5D shows the execution status of the processes P1 and P2. (A) shows how the process P2 is set in the process sleep table 3 to wake up the process P1 after 4 minutes when the process P2 is 00:00 (see S32 and S33 in FIG. 4).

(A) shows how the process sleep table 3 is set to wake up the process P1 after 2 minutes when the process P2 is 00:01 (see S34 and S35 in FIG. 4).

(C) shows how the process P1 is wake-up when the monitoring process is 00:03. FIG. 5E shows an example of the process sleep table. This process sleep table 3 is the same as that shown in FIG.
It is after being dynamically changed in S33, and is set as shown.

As shown in (a) to (e) of FIG. 5 above, the process P2 is set to the process P1 so that the process P1 wakes up at the expected execution completion time in accordance with the execution status of the process P2. The process sleep table 3 can be dynamically notified and changed.

FIG. 6 shows a process queuing flowchart of the present invention. In FIG. 6, S41 is a process P2.
Is running. In S42, it is determined whether or not it has ended. In the case of YES, the process P2 that sleeps is set in the process end table 4 in S43, the log is collected in S44, and S
It sleeps or ends at 45. On the other hand, NO in S42
In the case of, it returns to S41.

Through the above steps S41 to S45, the process P2 is terminated while it is being executed, and the process P2 is set in the process termination table 4 to notify the monitoring process 2 that the process P2 has been terminated.

In step S51, a log of various control data of the process P1 immediately before the process P1 sleeps is collected. In S52, the process P1 sleeps.

In step S53, the process P1 wakes up in response to the wakeup notification shown in FIG. 2B. In S54, logs such as various control data immediately after the process P1 wakes up are collected.

In step S55, it is determined whether P2 is set in the process end table. In the case of YES, since it is set that the waiting process P2 has ended, S5 is executed.
At 6, the process P1 starts operation. On the other hand, N in S55
In the case of O, it is found that the waiting process P2 has not ended yet. Therefore, in step S57, a dynamic change notification is sent to the process sleep table 3, and in step S58, the monitoring process responds to this notification. Three
Is dynamically changed, and the process returns to S53.

As described above, the process P1 repeatedly wakes up every time the sleep time elapses from the sleep state and dynamically changes the sleep time of its own process P1 until the process P2 is terminated in the process termination table 4. It becomes possible to wait until the process P2 ends.

FIG. 7 shows an example of the process end table of the present invention. This is the process P2 in S43 of FIG.
Is a table for setting when the process is finished and notifying the monitoring process 2.

FIG. 8 is an explanatory diagram of process waiting according to the present invention. This is a schematic representation of the waiting situation in FIG. 6 described above. In FIG. 8A, the process P1 starts.

In (a), the process P1 starts sleep to wait for the end of the process P2. (U), (F),
(K) is a process P for detecting the end of the process P2.
1 starts executing.

(D) and (v) detect the end of the process P2. In (e) and (h), since the process P2 is being executed, the process P1 starts sleep again.

(K) detects the end of the process P2 and executes the process P1 for the end of the process P2.
(H) ends the process P1.

As described above, as shown in the upper part of the figure, the process P1 has a long first sleep time as shown in (a)-(c), and the second and subsequent (o)-(f), (h). −
The sleep time of (X) is short, and it is possible to dynamically change the sleep time to the process sleep table 3 and wait for the end of the process P2.

FIG. 9 is an explanatory diagram for calculating the total process count and the like from the log of the present invention. In FIG. 9, S61 is
Capture logs. This captures all the logs of the state (control data and various information) immediately before sleep and the state immediately after wakeup in FIG. 6 described above.

In step S62, sorting is performed. This sorts by process. S63 creates a state transition for each process. S64 is during sleep every designated time /
The total number of running processes is calculated, the average value and the median of the frequencies are calculated, and tables and graphs are created. For example, for the log data of (a) of FIG. 10 described later, a frequency calculation result table of the sleep time of (c) of FIG. 10 is created,
FIG. 11A is a process frequency distribution diagram of the sleep time,
It is calculated and displayed as shown in the screen display example of the process status transition of FIG.

As described above, logs in a state immediately before sleep and a state immediately after wakeup are collected, and these logs are collected to calculate the total number of sleeping / running processes for each designated time for each process. , It is possible to present an index for obtaining the optimum value of the sleep time.

FIG. 10 is an explanatory view (1) of the log of the present invention.
Is shown. FIG. 10A shows an example of log data. Here, as shown in the figure, the log data collects the event code, the event name, and the event occurrence time in association with the process name.

FIG. 10B shows an example of the event code table. In this event code table, as shown in the figure,
The event name is registered in association with the event code.
FIG. 10C shows an example of a sleep frequency calculation result table. This is an example of the result of calculating the frequency of the sleep time from the log data, and as shown in the figure, the process frequency (number) is aggregated in association with the sleep time.

FIG. 11 is an explanatory view of the log of the present invention (part 2).
Is shown. FIG. 11A shows a relationship curve between the sleep time and the process frequency. This is (c) of FIG. 10 described above.
The sleep time is shown on the horizontal axis and the process frequency is shown on the vertical axis. Here, the process frequency is “1” during the sleep time “02 minutes” to “03 minutes”, and the process frequency is 0 at all other times.

FIG. 11B shows a screen display example of the process status transition. This is process P1, process P2
Is a horizontal bar graph showing the process state transition (running, light sleep, deep sleep, light sleep, running) with the passage of time.

[0058]

As described above, according to the present invention,
The process sleep table 3 is provided and the process 1
The sleep time is set, dynamically changed, and waited, and the configuration that logs are collected, so the optimal sleep time can be set quickly or dynamically and optimally. It is possible to meet and present statistical information for setting the optimal sleep time.

[Brief description of drawings]

FIG. 1 is a system configuration diagram of the present invention.

FIG. 2 is a flowchart illustrating the operation of the present invention.

FIG. 3 is an explanatory diagram of state transitions of the process of the present invention.

FIG. 4 is a flowchart for changing a sleep time according to the present invention.

FIG. 5 is an execution status diagram of process P2 of the present invention.

FIG. 6 is a queuing flowchart of the process of the present invention.

FIG. 7 is an example of a process end table of the present invention.

FIG. 8 is an explanatory diagram of queuing of the process of the present invention.

FIG. 9 is an explanatory diagram for calculating the total process count and the like from the log of the present invention.

FIG. 10 is an explanatory diagram (1) of a log of the present invention.

FIG. 11 is a log explanatory view (2) of the present invention.

[Explanation of symbols]

 1: Process 2: Monitoring process 3: Process sleep table 4: Process end table 5: Log data

Claims (4)

[Claims] 1. A process monitoring device for monitoring a sleep state of a process, wherein a process sleep table for setting a sleep time of the process, a sleep time of the process is set in the process sleep table, and the process is set according to the set sleep time. A process monitoring device comprising: a monitoring process that sequentially executes corresponding processes of sleep, swap-out, swap-in, and wake-up. 2. The process monitoring apparatus according to claim 1, wherein the sleep time in the process sleep table is dynamically changed in response to a notification from a process. 3. The sleep time of a sleeping process set in the process sleep table is dynamically changed in response to notification of completion time or completion rate of processing of another process. The process monitoring device according to claim 1 or 2. 4. A log before entering the sleep state and a log after the wakeup are collected for each process, and the sleep and execution status of each process is aggregated and presented based on the log. The process monitoring device according to any one of claims 1 to 3, which is characterized.