JPH0964918A - Buffer management system in communication controller - Google Patents

Abstract

PROBLEM TO BE SOLVED: To optimize a main storage area on the occurrence of the shortage of areas for buffers by acquiring buffer areas requiring a capacity for a required data length from the main storage area and returning the acquired area capacity to the main storage area after the areas are used. SOLUTION: A dividing means divides a main storage area 107 into an M.BUF 108 used for usual data communication and an E.BUF 109 used for saving data tentatively on the occurrence of a fault at the start of a communication controller. A packet task 100 and a data link task 101 acquire areas for a buffer corresponding to each layer of a communication protocol at data communication from the M.BUF area in the case of data communication for the communication processing and when acquisition of areas for a buffer is failed, after the optimizing processing of the M.BUF area via an optimize-processing 103, areas for a buffer are acquired again from the M.BUF area to restart the communication processing. The optimize-processing 103 uses the E.BUF areas to apply optimize-processing to the M.BUF area by a predetermined method.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a communication control device, and more particularly to a communication control device that performs communication by a packet method and a buffer management method in the communication control device that performs communication by a high level data link control procedure.

[0002]

2. Description of the Related Art In a conventional buffer management method in a communication control device, a main storage area is divided into data lengths for transmission and reception (hereinafter referred to as a buffer), and one event is generated during communication or when an event is reported to a higher-level communication protocol device. On the other hand, one frame is allocated, and after use, the buffer is controlled by the shared buffer management method by returning it as an empty area again.

[0003]

In the conventional buffer management method described above, when performing communication of a long frame in which the length of one frame is 4 kilobytes or more, not only data transmission / reception but also an upper communication protocol apparatus having a short data length is transmitted. There was a problem that the communication of the event reports of (1) and (2) occurred and the buffer was exhausted immediately, making communication impossible. In order to avoid this, a method of adding resources by an external storage device can be considered, but there is a drawback that the external storage device is expensive and impairs economic efficiency.

[0004]

According to a first aspect of the present invention, in a buffer management system in a communication control device for performing data communication using a packet switching line, a main storage area is used for normal data communication when the communication control device is activated. A dividing unit that divides data into a first buffer area and a second buffer area that evacuates data when a failure occurs, and a buffer for the respective layers of a communication protocol during the data communication is acquired from the first buffer area for communication. When the buffer is acquired, the process is performed, and when the buffer is failed to be acquired, the buffer is acquired from the first buffer region again and the communication process is restarted. Task group corresponding to each layer,
It is characterized by comprising the optimizer for performing the optimize process of the first buffer area by a predetermined method using the second area.

The second invention is the same as the first invention.
The optimize process arranges a buffer in use in the first buffer area in the second buffer area continuously from the beginning of the area, and then sequentially arranges the buffer in the second buffer area. It is characterized in that the buffer is copied to the cleared first buffer area.

A third aspect of the present invention is a buffer management method in a communication control device for performing data communication using a high level data link control procedure, wherein the main storage area is used for normal data communication when the communication control device is activated. A first buffer area and a dividing means for dividing the data into a second buffer area for urgently saving data when a failure occurs; and a buffer corresponding to each layer of the communication protocol during the data communication.
If the acquisition of the buffer is unsuccessful, the optimization process is performed on the first buffer region via the optimizer, and then the buffer is re-created from the first buffer region. The task group corresponding to each layer for acquiring and restarting the communication processing, and the optimizer for performing the optimize processing of the first buffer area by a predetermined method using the second area, The utilization efficiency of the first buffer area in the main storage area is improved.

[0007]

Next, the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram showing an embodiment of a buffer management system in the communication control apparatus of the present invention.

According to the present invention, as shown in FIG. 1, the main storage area 107 is used for normal data transmission / reception when the communication control apparatus is activated, and a main buffer area (hereinafter referred to as M Called BUF) 108
And a dividing means (not shown) for dividing into a buffer area for emergency data transmission / reception (hereinafter referred to as E / BUF) 109 used when a buffer exhaustion occurs, and a packet layer for dynamically securing a buffer from the main storage area. The packet task 100 performs processing, the data link task 101 performs data link layer processing, a buffer header queue 105 and a free buffer header queue 106 located between the main storage area 107 and the task.

FIG. 2 shows the two queues 105 and 106.
5 is a diagram showing an example of the configuration of a buffer header record 200 registered in the NEXT. P201, task ID 202 indicating which task is using the buffer, buffer header record number 203 used to identify each buffer by assigning a unique number from 1 to n, and the length of data stored in the buffer A buffer length 204 indicating the length and a buffer address 205 indicating the address on the M.BUF where the actual data is stored.

FIG. 3 is a diagram showing an example of the structure of the buffer on the M.BUF. The buffer on the M.BUF is a buffer ID 301 for identifying the contents of the data, and 1 to n to correspond to the buffer header record. Up to a buffer number 302, which is a unique number, and data 303 in which actual data is stored.

Next, the buffer acquisition process 102, the buffer release process 104, and the buffer search process used in the packet task and the data link task will be described with reference to FIGS.

When each task acquires a buffer, each task calls the buffer acquisition process shown in FIG. 7 (a).

In the buffer acquisition process, the buffer header record is acquired from the empty buffer header queue 106, the task ID and the buffer length are set in the buffer header record,
A free area of M / BUF is searched from the information of all buffer header records registered in the buffer header queue 105, and the start address thereof is set to the buffer address of the buffer header record. Further, the buffer header record number is copied to the buffer number of the buffer pointed to by the buffer address. Finally, the success / failure of buffer acquisition is notified to each task (S51 to S54).

When each task releases the buffer, each task calls the buffer releasing process shown in FIG. 7 (b). In the buffer release process, after calling the buffer search process shown in FIG. 7C, all the information of the searched buffer header record except the buffer header record number is erased and the buffer header record is returned to the empty buffer header queue 106. To do. Finally, each task is notified of the success and failure of the buffer release (S61 to S63).

When searching the buffer, each task is executed as shown in FIG.
The buffer search process shown in (c) is called. In the buffer search process, the buffer header record number is searched from the buffer header queue 105 as a key to search for a buffer header record, and a pointer to the buffer header record is returned to each task (S71 to S73).

However, since the above buffer release processing only separates the buffer header record and the buffer address on the M.BUF 108, the buffer used area and the buffer unused area having different lengths in the M.BUF are complicated and the buffer is not used. Depletion can occur frequently.

In order to minimize this problem, MBU
F-optimize processing 103 is performed. Figure 4 shows M-BUF
It is explanatory drawing of the optimization process, FIG.6 (b) is M *.
It is an operation | movement flowchart of a BUF optimize process. The M.BUF optimization process checks all the buffer header records registered in the buffer header queue 105 to check the buffer currently in use, and then the M.BU.
The second buffer 402, the fifth buffer 405, and the seventh buffer 407 that are being used in F108 are all saved to E / BUF109, and M / BUF108 is temporarily emptied, and then one second buffer from E / BUF109. Take out, M
-Restore to BUF, write the buffer storage address at that time to the buffer address of the buffer header record having the same buffer header record number as the buffer number, and perform the series of these processes until E-BUF becomes empty (S40 to S43). ). By the above processing, the buffer used area and the buffer unused area of the M · BUF are completely divided into two, and the buffer can be dynamically secured for the buffer unused area.

Using the optimization processing described above, how each task operates when a buffer is exhausted, when the buffer is exhausted in the packet layer, and when the buffer is exhausted in the data link layer Will be described separately with reference to the flowcharts of FIGS.

First, a case where a buffer is exhausted in a task (hereinafter, referred to as a packet task) for processing a packet layer will be described. When the buffer acquisition process (S1) fails in the packet task process, the M.BUFF optimize process is called (S2). In the M / BUF optimization process, the memory control target of the task is M / BUF.
To E-BUF and control is returned to the task (S4
4, S45). The packet task notifies the task (hereinafter referred to as the data link task), which is a lower layer of the packet layer, of processing the data link layer, which is a lower layer of the packet layer, if the packet task is not in communication (S3). , S6), and if communication is in progress, an RNR packet transmission request is made (S5, S6). An E / BUF buffer is used during communication between tasks.

When the data link task receives the buffer busy notification from the packet task, it transmits an RNR packet (S33), and transmits an RNR frame to put the data link layer in a busy state (S34). After the line busy setting is completed, the packet task is notified of the line busy setting completion (S35, S36). When the packet task receives the notification that the line busy setting is completed (S12), the M.BU is restarted.
The F optimize process is executed (S14). M ・ BU
In the F optimization process, the buffer use area is checked for the buffer header record registered in the buffer header queue, and the E / BUF area is used for relocation (S40 to S42). After the relocation is completed, the packet task is notified of the completion of the M.BUF relocation. Packet task is M
When the BUF relocation completion notification is received, the buffer busy cancellation is notified to the data link task in order to cancel the circuit busy, and the RR packet is requested to be transmitted (S15, S1).
7). When the data link task receives the line busy cancellation notification from the packet task, it transmits an RR packet to cancel the busy state of the packet layer (S38 to S3).
9). Further, an RR frame is transmitted to release the busy state of the data link layer, and the data link task notifies the packet task of completion of releasing the line busy (S31).
0-S312). When the packet task is notified that the line busy is released, the packet task acquires the buffer again and continues the communication (S11).

A case where the buffer is exhausted in the data link task processing will be described. If the data link task fails to acquire the buffer, it calls the M.BUF optimization process (S21). In the M / BUF optimization process, the memory control target of the task is changed from M / BUF to E.
Switch to BUF and return control to task (S44-S4
5). The data link task notifies the packet task of buffer busy after making the line busy, and transmits an RNR frame if communication is in progress (S22 to S2).
3). An E / BUF buffer is used during communication between tasks. When the packet task receives the buffer busy notification from the data link task, it temporarily stops the packet data transmission. When the data link task completes the line busy setting, it again calls the M.BUF optimization process (S24). The M.BUF optimization process checks the buffer used area for all buffer header records registered in the buffer header queue, and E.
Rearrangement is performed using BUF (S40 to S43). If the buffer acquisition fails even after the optimization process is executed, the lack of buffer is notified to the upper protocol device and the line is disconnected (S27). After the relocation is completed, the data link task is notified of the completion of the M.BUF relocation. When the data link task receives the notification of the completion of the M / BUF relocation, it transmits an RR frame to release the line busy, and after confirming the response, notifies the packet task of the buffer busy release (S25 to S26). Data link task is M-BUF
The optimize process ends, and the buffer is acquired again. When the bucket task receives the buffer busy cancellation notification, it resumes data transmission and reception.

The M.BUF optimization process in the case of using the high level data link control procedure is the same as the process in which all the processes for the packet task are excluded, and will be omitted.

[0024]

As described above, according to the present invention, a buffer is dynamically acquired from the main storage area by the required data length and returned to the main storage area after use. Further, when the buffer is exhausted, the resources can be effectively used by optimizing the main storage area after temporarily setting the communication line in a busy state.

[Brief description of drawings]

FIG. 1 is a block diagram showing one embodiment of the present invention.

FIG. 2 is a buffer header record 20 in this embodiment.
FIG. 9 is a diagram illustrating a configuration example of a zero.

FIG. 3 is a diagram showing a configuration example of a buffer on an M · BUF in the present embodiment.

FIG. 4 is an explanatory diagram of M / BUF optimization processing of the present invention.

5A is an operation flow chart of packet task / buffer acquisition processing, FIG. 5B is an operation flow chart of data link / buffer acquisition processing, and FIG. 5C is an operation flow chart of packet task / packet reception processing. .

6A is an operation flow diagram of a data link task reception process, and FIG. 6B is an operation flow diagram of an M.BUF optimization process.

FIG. 7A is an operation flow chart of buffer acquisition processing,
(B) is an operation flow diagram of a buffer release process, and (c) is an operation flow diagram of a buffer search process.

[Explanation of symbols]

 100 packet task 101 data link task 102 buffer acquisition processing 103 M / BUF optimization processing 104 buffer release processing 105 buffer header queue 106 empty buffer header queue 107 main storage area 108 M / BUF 109 E / BUF 200 buffer header record 201 NEXT. P 202 task ID 203 buffer header record number 204 buffer length 205 buffer address 300 buffer 301 buffer ID 302 buffer number 303 data 401 first buffer 402 second buffer 403 third buffer 404 fourth buffer 405 fifth buffer 406 sixth buffer 407 7th buffer 408 8th buffer

Claims (3)

[Claims] 1. A buffer management method in a communication control device for performing data communication using a packet-switched line, comprising: a first buffer region that uses a main storage area for normal data communication when the communication control device is activated; and a data buffer when a failure occurs. And a dividing means for dividing into a second buffer area for emergency saving, and a buffer corresponding to each layer of the communication protocol during the data communication is acquired from the first buffer area for communication processing, and acquisition of the buffer fails. In that case, the task group corresponding to each layer for re-acquiring the buffer from the first buffer area and resuming the communication processing after performing the optimization processing of the first buffer area through the optimizer, Optimize the first buffer area by a predetermined method using the second area A buffer management method in a communication control device, comprising: the optimizer, wherein utilization efficiency of the first buffer area of the main storage area is improved. 2. The optimizing process allocates a buffer in use in the first buffer area to the second buffer area continuously from the beginning of the area and then continues to the second buffer area. 2. The buffer management method in the communication control device according to claim 1, wherein the buffer arranged as described above is copied to the cleared first buffer area. 3. A buffer management method in a communication control device for performing data communication using a high-level data link control procedure, wherein a main storage area is used as a first buffer area for normal data communication when the communication control device is activated, and a failure. When the data is generated, a dividing unit that divides the data into a second buffer area for emergency saving, and a buffer for the respective layers of the communication protocol during the data communication are acquired from the first buffer area to perform communication processing, When the acquisition fails, the optimization processing of the first buffer area is performed through the optimization means, and then the buffer is acquired again from the first buffer area and the communication processing is restarted. And an optimizer for the first buffer area according to a predetermined method using the second area. A buffer management system in a communication control device, comprising: the optimizer for performing an equalizing process to improve utilization efficiency of the first buffer area of the main storage area.