JPH0793169A - Inter-process communication system - Google Patents

Abstract

PURPOSE:To perform the processing of requested data in real time and to quickly return a state to the one to process the data in real time by interrupting futile processing even when a large number of messages reside due to an unknown factor in a message queue. CONSTITUTION:This system is constituted in such a way that a buffer 2 provided at every client process 1 in which each client process 1 sets data and a consecutive number, and the message queue to connect the leading address of the data overwritten on the buffer 2 to the message 31 on which the consecutive number is set are provided at shared space, and processing to fetch the data from the buffer 2 is performed only when the consecutive number of the message 31 taken out from the forefront of the message queue 3 by a server process 4 is not equal to that of the data in the buffer 2 at the leading address, and the next processing is started by interrupting the processing when noncoincidence is obtained between them.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an interprocess communication system for communicating between processes.

[0002]

2. Description of the Related Art Conventionally, when data is asynchronously transferred from a plurality of processes to a single process and a processing request is made, as shown in FIG. 7, a message queue for connecting messages in a shared space and data to be processed are provided. A buffer for writing and passing is provided, and each process asynchronously writes data to the buffer and connects the message to the message queue to request processing. The process for performing the process repeatedly takes out the message from the head of the message queue, reads the data from the corresponding buffer, and performs the process. The configuration and operation of FIG. 7 will be briefly described below.

FIG. 7 shows an explanatory view of the prior art. In FIG. 7, processes A, B, and C are processes in the spaces A, B, and C that asynchronously request data processing.

The buffers A, B, C are buffers provided in the shared space and allocated by the processes A, B, C for writing data, respectively. Processes A, B, and C asynchronously overwrite the buffers A, B, and C with the latest data.

The process D reads data from each of the buffers A, B, and C for which processing has been requested by the processes A, B, and C, and performs processing (for example, display processing). The process D is provided in the space D. Is.

The message queue includes processes A, B, and C.
Is for providing a process request to the process D by connecting messages in which buffers A, B, and C for writing the contents and data to be processed are set, and are provided in the shared space. The process D sequentially takes out the messages from the head of the message queue and processes them.

Buffers A, B, and C are processes A, B, and
C is an area for overwriting data, which is provided in the common space. Next, the operation will be described.

(1) The data requested by the process A is written in the buffer A, and at the same time, the process request in which the start addresses of the own process A and the buffer A are set is written in a message and connected to the message queue asynchronously. Similarly, the other processes B and C also write data to the buffers B and C, and at the same time, write the processing request in which the start addresses of the own processes B and C and the buffers B and C are set to the message and connect them to the message queue asynchronously. To each.

(2) Process D removes one message from the head of the message queue, retrieves data from any of buffers A, B, and C at the head address designated by the message, and performs the process designated by the message (for example, The display process) is repeated.

As described above, the processes A, B, and C asynchronously write data in the buffers A, B, and C, and the process D.
In this case, the data is sequentially fetched from the buffers A, B and C and processed (for example, display processing).

[0011]

It is assumed that the processes A, B, and C having the above-described configuration of FIG. 7 asynchronously write data to the buffers A, B, and C, and the process D sequentially fetches and processes the data. , If the data processing of the process D is delayed for some reason (for example, the processing is delayed due to waiting for access by external I / O, etc.), it is delayed to take out the message connected to the message queue and perform the processing. There is a problem that the data in the buffers A, B, and C asynchronously sent from the processes A, B, and C cannot be processed in real time, for example, the data at the current time cannot be displayed on the screen in real time. Further, when the process D is delayed in processing and a large number of messages are accumulated in the message queue, the number of messages connected to the message queue does not decrease easily, and the load of the process D continues for a long time for a long time. There is also a problem that other processing cannot be performed and a lot of time is required until processing can be performed in real time.

The present invention solves these problems by
Each client process writes the data and serial number in the buffer, sets the serial number in the message, connects to the message queue and requests processing, and the server process fetches the message serial number from the message queue and the data serial number in the buffer is equal. Processing is performed only when it is not equal, processing is stopped and only the latest data is processed, the requested data is processed in real time, and even if many messages remain in the message queue for some reason. The purpose is to stop useless processing and quickly return to a state in which data is processed in real time.

[0013]

[Means for Solving the Problems] Means for solving the problems will be described with reference to FIG. In FIG. 1, a client process 1 requests data processing.

The buffer 2 is an area provided in the shared space and in which each client process 1 writes data. The message 31 sets the start address of the data to be processed on the buffer 2 and the serial number of the data.

The message queue 3 connects the messages 31 and requests the server process 4 for processing. The server process 4 sequentially takes out the message 31 from the head of the message queue 3 and performs processing only when the serial number set in the message 31 and the serial number of the data in the buffer 2 at the set start address are equal. Is.

[0016]

According to the present invention, as shown in FIG. 1, each client process 1 writes data and a serial number of the data into the buffer 2, and at the same time, a message 31 in which the start address of the buffer 2 and the serial number of the data are set is message queued. 3, the server process 4 sequentially fetches the message 31 from the head of the message queue 3, and the data from the buffer 2 is read only when the serial number in the fetched message 31 and the serial number of the data in the buffer 2 at the head address are equal. Is taken out and processed. If they are not equal, the process is stopped and the next process is performed.

When the start address of the buffer 2 is fixed, the client process 1 connects the message 31 for which a serial number is set to the message queue 3, and the server process 4 extracts the serial number from the message 31 and the fixed start address. Only when the serial number of the data in the buffer 2 is the same, the data is taken out from the buffer 2 and the processing is performed. When they are not the same, the processing is stopped and the next processing is performed.

A plurality of processes are provided as the client process 1, and each of these processes has its own process ID.
A message 31 having a serial number, a serial number, and a start address if necessary is connected to the message queue 3.

Therefore, each client process 1 writes the data and serial number in the buffer 2 and sends the message 31.
A serial number is also set to connect to the message queue 3 to request processing, and the process is performed only when the serial number of the message 31 fetched from the message queue 3 by the server process 4 and the serial number of the data in the buffer 2 are equal, and are equal. By stopping the processing when there is no data, and processing only the latest data, the requested data is processed in real time, and even if many messages 31 remain in the message queue 3 for some reason, it is a wasteful processing. It is possible to stop the operation and quickly return to the state of processing the data in real time.

[0020]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, the construction and operation of an embodiment of the present invention will be described in detail with reference to FIGS.

FIG. 1 shows a system configuration diagram of the present invention.
In FIG. 1, the client process 1 requests the server process 4 to process data asynchronously. Here, for example, as shown in the figure, the process A, the process B, the process C, etc. It exists in a space different from the space B, the space C, and the like. The process A, the process B, and the process C, which are the client processes 1, are each composed of the transmission process 11 and the like.

The transmission process 11 writes the data and the serial number of the data to the buffer 2, or writes a new serial number to the buffer 2 regardless of the data change, and sets the data serial number and the start address of the buffer 2. The message 31 is connected to the message queue 3 or the like.

The buffer 2 is provided in the shared space, and the processes A, B, and C which are the client processes 1 write data and serial numbers. Here, as the buffers 2 for the process A, the process B, and the process C, the buffer A, the buffer B, and the buffer C having a fixed start address are provided. As shown in the figure, the serial numbers and data are overwritten in the buffers A, B, and C, and the latest serial numbers and data are always written.

The message queue 3 manages the processing order by asynchronously connecting the messages 31 in which the start address and the serial number of the buffer 2 are set. The server process 4 receives the message 31 from the head of the message queue 3.
Are sequentially taken out and processed, and are composed of a reception process 41 and the like.

The receiving process 41 takes out the message 31 from the head of the message queue 3, and only when the serial number set in the extracted message 31 is equal to the serial number of the data of the buffer 2 at the set start address. Processing (for example, processing of displaying data on the screen) is performed. Therefore, regardless of whether the data in the buffer 2 has been overwritten, the latest data having the same serial number is taken out from the buffer 2 and processed.

The details will be sequentially described below. FIG. 2 shows a transmission processing flowchart of the client process of the present invention. In FIG. 2, S1 sets data in the buffer 2. This is the client process 1 of FIG. 1, for example, the transmission process 11 of the process A overwrites (sets) the data to be processed by the server process 4 in the buffer A for the process A in the buffer 2. ).

In S2, the serial number of data is set in the buffer 2. This is, for example, after the transmission processing 11 of the process A sets data in the buffer A in S1,
The serial number of the data is overwritten (set) in the buffer A.

In step S3, a message is created (data processing request). This is because, for example, the process A is shown in FIG.
Message 31 in which the data processing request (type), A (transmission source process ID), the buffer start address, and the data serial number are set.

In S4, the message is transmitted (connected to the message queue 3). This connects (enqueues) the message 31 created in S3 to the end of the message queue 3. As a result, the transmission of the message 31 is requested.

As described above, in the client process 1, after the data to be requested by the transmission processing 11 and the serial number of the data are overwritten in the buffer 2, the start address of the buffer 2 and the serial number of the data are set in the message 31. To the end of the message queue 3 to request transmission. As a result, the latest data and the serial number of the data are always overwritten in the buffer 2, and the message 3 at the time of overwriting is written in the message queue 3.
1 is sequentially connected to the end of the message queue 3 to request transmission.

FIG. 3 shows an example of the buffer contents of the present invention.
This is an example of the buffer 2 provided in the shared space in FIG. Here, the area of the predetermined size L from the start address a is the area of the buffer A (buffer 2 for the process A), and the serial number a of the data and the data are overwritten as shown in the figure. Similarly, from the start addresses b and c to a predetermined size L
Area is an area of buffers B and C (buffers 2 for processes B and C), and data serial numbers b and c as shown in the figure.
And overwrite the data respectively. Where size L
May be set individually.

FIG. 4 shows an example of message contents of the present invention. This is an example of the message 31 connected to the message queue 3 provided in the shared space in FIG. Here, the following information shown in the figure is set.

-Type: For example, "processing request", "Status notification" -Source process ID: For example, "A" for process A-Start address of buffer: For example, for process A, start address a ( (See FIG. 3).

Data serial number: For example, a sequential number for each process (1, 2, 3, ...) FIG. 5 shows a reception process flowchart of the server process of the present invention.

In FIG. 5, S11 receives the message. This is the reception process 4 of the server process 4 of FIG.
1 is message 31 from the top of message queue 3
Take out one.

In step S12, the serial number X of the data in the message 31 is acquired. This acquires the serial number X of the data in the message 31 extracted in S11 (see FIG. 4). S
13 acquires the serial number of the data in the buffer 2 designated by the message 31. This acquires the serial number Y of data from the position of the buffer 2 of the head address set in the message 31 extracted in S11.

In step S14, it is determined whether X = Y. In the case of YES, since the serial number X in the message 31 and the serial number Y of the data in the buffer 2 are equal and it is found that the message 31 is the latest data, the data in the buffer 2 is acquired in S15, and this acquisition is performed in S16. The processed data is processed (for example, displayed on the screen as shown in FIG. 6 described later). On the other hand, in the case of NO in S14, it is found that the message 31 is not the latest data, so the process returns to S11, the next message 31 is taken out from the message queue 3, and S12 and subsequent steps are repeated.

As described above, the reception process 41 in the server process 4 starts the message 3 from the head of the message queue 3.
1 is taken out, and the serial number set in this message 31 is compared with the serial number of the data of the buffer 2 at the set start address, and when they match, the latest data is found. On the other hand, when they do not match, it is determined that the data is old data, so the processing of the data is stopped and the process proceeds to the reception process of the next message 31. As a result, the processing of the old message 31 is stopped and the processing of the next message 31 is started, only the latest data is processed, and the real-time property of data processing can be greatly improved.

FIG. 6 shows an application example of the present invention. this is,
The client process 1 asynchronously buffers the data in which the number of objects, the number of people, etc. change as time passes.
Also, the process of requesting the server process 4 using the message queue 3 and displaying the number and the number of the server process 4 on the screen is schematically shown. The horizontal axis is the time axis.

FIG. 6A shows actual data. This indicates changes in the number of objects and the number of objects detected by the client process 1 in FIG. Here, since the data changed as: time t1: 10 time t2: 15 time t3: 20 is detected, when these data are detected, the buffer 2 is immediately overwritten and the serial number is set, and the serial number is set for each time. Then, the message 31 in which the top address of the buffer 2 is set is connected to the message queue 3 and transmission is sequentially requested.

FIG. 6B shows the real time. this is,
The actual time at which the actual data in (a) of FIG. 6 is detected is t1, t
2 and t3. FIG. 6C shows the conventional processing (processing by the process D in FIG. 7). In the conventional processing by the process D of FIG. 7, data "1" at time t1
Since 0 is the first time, it can be displayed as "10" at the same time as the real time. However, regarding the data "15" at time t2, the processing of another message 31 connected to the message queue 3 is performed. Since it has been performed, the relevant “15” can be displayed as shown in the figure with a delay of Δt 2. Further, regarding the data “20” at the time t3,
Other message 3 connected to message queue 3
Since the process 1 is performed, the "20" is displayed as shown in the figure with a large delay from Δt3, and in particular, the data "20" at time t3 cannot be said to be displayed in real time. .

FIG. 6D shows the processing of the present invention (processing by the server process 4). In the processing by the server process 4 of the present invention, the data “10” at the time t1 is the first time, so that “1” is set at the same time as the real time.
If the message 31 at the time t2 is taken out from the message queue 3 just after the time t3, the data and the serial number of the buffer 2 at the time t3 are displayed. The data "1" at the time t2 is overwritten by the data.
The display process of 5 "is stopped and the process proceeds to the next process. When the message 31 at the time t3 is taken out from the message queue 3, the data in the buffer 2 and the serial number are at the time t.
Since they are No. 3 and match, the display processing of the data “20” at the time t3 can be performed and the real-time processing (display) can be performed as shown in.

The objects include the number of visitors, traffic volume,
There are reservation status, temperature, etc. (things that can be monitored by numbers, etc.), and the quantity and number of these objects are displayed in real time.

[0044]

As described above, according to the present invention,
Each client process 1 writes the data and serial number in the buffer 2, sets a serial number also in the message 31, connects to the message queue 3 and requests processing, and the server process 4 retrieves the message 3 from the message queue 3.
Since the processing is performed only when the serial number of 1 and the serial number of the data in the buffer 2 are equal, and when they are not equal, the processing is stopped and only the latest data is processed.
The requested data can be processed in real time, and even if many messages 31 are retained in the message queue 3 for some reason, useless processing can be stopped and the data can be quickly returned to the real-time processing state. .

[Brief description of drawings]

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

FIG. 2 is a transmission process flowchart of a client process of the present invention.

FIG. 3 is an example of buffer contents of the present invention.

FIG. 4 is an example of message contents of the present invention.

FIG. 5 is a reception processing flowchart of a server process of the present invention.

FIG. 6 is an application example of the present invention.

FIG. 7 is an explanatory diagram of a conventional technique.

[Explanation of symbols]

 1: Client process 11: Transmission process 2: Buffer 3: Message queue 31: Message 4: Server process 41: Reception process

Claims (3)

[Claims] 1. In an inter-process communication method for communicating between processes, each client process (1) requesting processing overwrites data and each client process (1) sets a different serial number each time the data is overwritten. ), A buffer (2) provided for each of them, and a message queue (3) connecting the message (31) with the start address of the data overwritten in the buffer (2) and the serial number set in the shared space are requested. Only when the serial number of the message (31) fetched from the head of the message queue (3) by the server process (4) performing the processing is equal to the serial number of the data in the buffer (2) at the set head address. Data is taken out from the buffer (2) and processed,
An inter-process communication method characterized by being configured to stop processing and proceed to the next processing when they are not equal. 2. The process according to claim 1, wherein the message (31) having the serial number set therein is connected to the message queue (3) when the head address of the buffer (2) is fixed. Inter-communication method. 3. The client process (1) comprises a plurality of processes, and each of these processes has its own process ID.
The interprocess communication system according to claim 1 or 2, wherein a message (31) having a serial number, a serial number, and a start address set as necessary is connected to the message queue (3).