JP2603013B2 - Ring buffer message management processing method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a ring buffer message management system for reading a message from a ring buffer for storing a message sent from a network.

[0002]

2. Description of the Related Art FIG. 4 is a system configuration diagram. Calculator 5
0 is connected to the network 10, and a message sent from the network 10 to the computer 50 is temporarily stored in the ring buffer 20.

FIG. 5 is an explanatory diagram of a conventional reading process.
The messages received by the computer 50 are stored in the ring buffer in the order received. For example, when a message A and a message B are received in this order, they are stored in the order of the message A and the message B as shown in FIG. The ring buffer management unit 25 stores the head position of the free space after the last message among the received messages as “free space head position” (hereinafter referred to as WP). Stores the head position of the free area after message B. When message C is received after message B, message C is written from WP, and
Is updated to the head position of the empty area after the message C. Further, the ring buffer 20 stores the head position of the foremost message among the unread messages as a “message head position” (hereinafter referred to as RP), and in the case of FIG. Is stored as the RP. When a message reading instruction is issued, reading is started from the RP, the message A is read, and when the reading of the message A is completed, the RP is updated to the head position of the message B. The same processing is performed for message B.

[0004]

In the conventional system, messages can be read only in the order of reception.

[0005] It is an object of the present invention to enable messages to be read in a desired order.

[0006]

FIG. 1 is a block diagram showing the principle of the present invention. In FIG. 1, 10 is a network, 20 is a ring buffer, 25 is a ring buffer management unit, 26 is a free area head storage unit, 27 is a tag change processing unit, 28 is a message head storage unit, 30 is a CPU, 40 is a memory, 5
0 represents a computer.

The ring buffer 20 is connected to the network 10
Are stored in the order in which they were sent. The message Mi is sent with the tag Ti attached.

The ring buffer management unit 25 includes a free area head storage unit 26, a tag change processing unit 27, and a message head storage unit 28, and manages the ring buffer 20.

The free area head storage unit 26 stores the head position of the free area after the received message as "free area head position" (hereinafter referred to as WP). The tag change processing unit 27 replaces the tag Ti of the read message Mi with
Change to a predetermined tag indicating that reading has been completed.

[0010] The message head storage unit 28 stores the head position of the first message among the unread messages Mi. The CPU 30 reads the message stored in the ring buffer 20 and stores the content and the processing result in the memory 40 as necessary.

[0011]

When a message reading instruction is issued from the CPU 30 with the designation of a tag, the ring buffer management unit 2
5 retrieves the message by the tag starting from the message head position, and transmits the content to the CPU 30. Further, the tag change processing unit 27 changes the tag of the read message to a tag indicating that the message has been read. Therefore, when the tag of the message at the forefront becomes a tag indicating that the reading has been completed, the message head storage unit 28 updates the RP to the head position of the 2R message. As a result, the area of the message outside the range from the RP to the WP becomes a free area.

When a message is sent from the network 10, the ring buffer management unit 25 writes the message from the WP, and after the writing is completed, sets the WP at the beginning of the free area after the sent message. Update to position.

[0013]

FIG. 2 is an explanatory view of an embodiment of the present invention. In the figure,
Reference numeral 20 denotes a ring buffer, MSi denotes a size of each message, Ti denotes a tag of each message, WP denotes a free area start position, and RP denotes a message start position. Ring buffer 20 stores message 4 in the order of arrival from message 1
I remember until. Here, the tag Ti indicates the identification ID of the sender or the terminal of the message i in the network. If the tag is XX, it indicates that the message has been read. Therefore, for example, it is understood that the messages 1 and 3 have been read, and the message 2 is a message from the sender or the terminal having the identification ID of AA.

FIG. 3 shows a processing flow at the time of reading. In the figure, WP and RP correspond to FIG. MP indicates the head position of the message to be processed, and MS indicates the size of the message to be processed.

In the process (10), a read command is issued with designation of a tag. In processing (20), RP is set to MP
Substitute for In process (30), if WP and MP are equal, wait until WP and MP are not equal.
In the process (40), the tag of the message at the position of the MP is read. In the process (50), it is determined whether or not the tag is XX. If it is XX, the process proceeds to processing (70). In processing (70), if MP and RP are equal, RP is set to RP +
Update to MP, and when MP and RP are not equal, RP
Without updating the process (90). In the process (90), the MP is updated to MP + MS, and the process proceeds to the process (30).
When it is not XX in the process (50), in the process (60),
Wait for the message to arrive completely. Processing (8
In 0), it is determined whether or not the tag of the message is the specified tag. If the tag is the specified tag, the message is read, the tag of the message is set to XX, and the process ends. When the tag is not the above specified tag,
In processing (90), MP is updated to MP + MS and processing (3)
Go to 0).

Next, in the case shown in FIG.
The processing for reading the message B will be described with reference to FIG. In the process (10), a read command is issued with the designated tag being BB. In process (20), first, message 1
Is a processing message.

In process (50), the tag of message 1 is XX
Is determined, and since MP and RP are equal in the process (70), the RP is updated to the beginning of the message 2 and the process (90)
To update the MP to the beginning of message 2. That is, the processing message is updated to the message 2. Since the tag of message 2 is AA, processing (30), (40), (5)
The MP is updated in process (90) via (0), (60), and (80), and the processed message is updated to message 3. Since the tag of message 3 is XX, processing (3
0) and (50), the process proceeds to the process (70). At this time, since the MP and the RP are not equal, the MP is updated in the process (90) without updating the RP. Update. Since the tag of the message 4 is the designated tag BB, the reading is executed in the process (100) and the tag is set to XX.

In this embodiment, the tag of the read message is XX, and the message is held in the ring buffer as it is until the next read request. That is, the release of the message (that the message goes out of the range from RP to WP and belongs to a free area) is after the next read request. By the way, when a computer having a copy-back cache memory uses a ring buffer, invalidate the area before the message arrives (if the data exists in that area of the cache memory, the data is invalidated). It is necessary to make it. However, in this embodiment,
Until the message is released, no new message arrives in the message area. Therefore, if an area of the ring buffer that is not released is invalidated immediately before the area is released, the cache memory can be used efficiently if the area of the cache memory corresponding to the area is invalidated. This is because the area of the cache memory corresponding to the message area where the ring buffer is not released can be used for processing other than the ring buffer.

[0019]

As described above, according to the present invention, messages can be read in a desired order.

[Brief description of the drawings]

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

FIG. 2 is an explanatory view of an embodiment of the present invention.

FIG. 3 shows a processing flow at the time of reading.

FIG. 4 is a system configuration diagram.

FIG. 5 is an explanatory diagram of a conventional reading process.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 10 Network 20 Ring buffer 25 Ring buffer management part 26 Free area head storage part 27 Tag change processing part 28 Message head storage part 30 CPU 40 Memory 50 Computer

Claims (1)

(57) [Claims] 1. A ring buffer (20) for storing a message (Mi) sent from a network (10).
In the message management processing method of the ring buffer when the CPU (30) reads the message (Mi), each message (Mi) is given a tag (Ti) when transmitted on the network, Among the messages stored in the buffer (20), the tag (T) of the read message (Mi) is read.
a tag change processing unit (27) for changing i) to a predetermined tag indicating that the message has been read, and a first message out of the unread messages stored in the ring buffer (20) A ring buffer management unit (25) including a message head storage unit (28) for storing a head position of a certain message; and the ring buffer management unit (25) includes a CPU (30).
When a command for reading a desired message with a tag designation is received from the CPU, the desired message is searched by the designated tag starting from the message head position, and the content is transmitted to the CPU (30). A ring buffer message management processing method.