JPH01175050A - Memory controller - Google Patents

Abstract

PURPOSE:To allow a task with the highest priority to always acquire a memory by permitting the acquisition of a storage area only when the size of an unused area of the memory is larger than acquisition reference data previously determined in each processing. CONSTITUTION:Reference residual data (data indicating the size of an unused area to be secured in a common memory 81 at the time of acquiring the common memory 81) previously determined in each processing are stored in a reference residual table 82 of a RAM 8. When the acquisition of the storage area of the common memory 8 required for processing is requested, the reference residual data stored in the table 82 and corresponding to the processing concerned are compared with the size of the unused area of the common memory 81, and only when the size of the unused area of the memory is larger, the acquisition of the storage area is permitted. When the minimum reference residual data are set up on the processing with highest priority, a task with the highest priority can always acquire the memory.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a memory management device, and particularly to a memory management device that manages memory shared by a plurality of processes.

[Prior art and problems to be solved by the invention] In the conventional memory management device, the following method is used as a memory acquisition control method.

(1) For every memory acquisition request from a task or module, it is checked whether the remaining amount of memory is larger than the requested amount, and if it is not, the memory is always acquired.

(2) Store memory acquisition requests in a queue, check the remaining memory capacity for the request with the highest priority among the requests in the queue, and if there is a remaining capacity greater than the request, acquire it.

However, with the methods (1) and (2) above, a situation occurs in which memory cannot be provided to a task or module that requires memory acquisition within a predetermined time. Next, this situation will be explained with respect to a conventional transaction processing system shown in FIG.

In FIG. 8, the cluster controller 11 has a plurality of P
It is connected to both the transmission line 2 to which the O8 terminal device is connected and the telephone line network 4 connected to the host computer 5, and mediates data transmission between them. Here, the host computer 5 side and the PO8 terminal device 3
When data is buffered from both sides to the cluster controller 11, the memory of the cluster controller 11 becomes full with messages from the PO8 terminal device 3 side, which is the lower level, and the message from the host computer 5, which is the higher level, becomes full. A situation may arise where the request cannot be accepted. In such a case, the communication procedure of the host computer 5 must be restarted from the beginning, so the communication efficiency of the cluster controller 11 is significantly reduced.

Therefore, the main object of the present invention is to provide a memory management device that allows tasks and modules with high priority to always obtain memory.

[Means for Solving the Problems] A memory management device according to the present invention includes a storage means that stores acquisition standard data for each process, which is a criterion for determining whether or not to acquire memory, and a storage unit that requires acquisition of a storage area necessary for the process. When it is done,
a determining means that compares the acquisition standard data corresponding to the process stored in the storage means with an unused area of the memory, and determines whether an unused area that is equal to or greater than the acquisition standard is secured in the memory; , in response to determining whether an unused area exceeding the acquisition standard is not secured in the memory by the determination means, acquisition is permitted even if the unused area of the memory is larger than the size of the area requested for acquisition. control means to prevent the

[Operation] When the memory management device according to the present invention requests the acquisition of a storage area necessary for a process, the size of the unused area of the memory is acquired as acquisition standard data that is predetermined for each process and serves as a criterion for permission or disapproval of acquisition. If the unused area larger than the acquisition criterion is not secured in memory, even if the size of the unused area of memory is larger than the size of the storage area requested for acquisition, Prohibit its acquisition.

Embodiment of the Invention FIG. 1 is a block diagram of a transaction processing system to which an embodiment of the invention is applied. In FIG. 1, a cluster controller 1 is connected to a plurality of PO8 terminal devices 3 via a transmission line 2, and is also connected to a host computer 5 via a telephone line network 4. The cluster controller 1 includes a CPU 6, a ROM 7, a RAM 8, and a transmission control section (not shown). The ROM 7 stores an operating program for the CPU 6 based on the flowcharts shown in FIGS. 6 and 7 later. The RAM 8 includes a shared memory 81 and a reference remaining amount table 82. The coexisting memory 81 is used when a task receives a message from the PO8 terminal device 3 or a message from the host computer 5. The base wall remaining quantity 82 stores reference remaining amount data predetermined for each process.

Here, the reference remaining amount data is data indicating the size of an unused area that should be secured in the shared memory when acquiring the shared memory. When a task requests to acquire the shared memory 81, it determines whether or not a predetermined amount or more of unused space is secured in the shared memory 81 based on the standard remaining capacity table 82, and if it is not secured, it Even if there is an unused area in the shared memory 81 that is larger than the storage area requested for acquisition, acquisition is prohibited.

FIG. 2 is a configuration diagram of the shared memory shown in FIG. 1. Second
In the figure, shared memory 81 includes an area control block 811 and a shared memory area 812. The area control block 811 stores a tag indicating whether the shared memory 81 is in use, size data indicating the size of the shared memory 81, management number data, a next area pointer, and a previous area pointer. Shared memory area 8
12 is a storage area shared by each process, and an area 813 for each process is provided. Each area has a header section 8.
14 are provided. The header part 814 is the area 813
A tag indicating whether or not the area 813 is in use, size data indicating the size of the area 813 (including the front and rear parts), management number data, a next area pointer indicating the address of the next area, and a previous area indicating the address of the previous area. area pointer. Here, if there is no area before the area in question, O is used as the previous area pointer. If there is no area following the area, 0 is used as the next area pointer.

FIG. 3 is a configuration diagram of the reference remaining amount table shown in FIG. 1. In FIG. 3, the reference remaining amount table 82 has a processing number (
ID), reference maximum remaining proc length data is set in advance and stored. As the standard maximum remaining block length data, data indicating the size of an unused area to be secured in the shared memory area 82 after memory is acquired may be used. Area 8
Data indicating the size of the unused storage area that should exist in 2 may also be used. Note that by making the base maximum remaining block length updatable and changing it in accordance with the priority of the process, the priority of the process and the degree of competition can be changed.

4A to 4C are diagrams for explaining the acquisition of the shared memory area 812 shown in FIG. 2, and FIGS. 5A to 5H are diagrams for explaining the release of the shared memory area 812 shown in FIG. 2. This is a diagram for Next, acquisition and release of memory will be explained with reference to FIGS. 48 to 5H.

When acquiring memory, the size and management number are specified and a predetermined routine is called. Then, when the area is acquired, the start address of the allocated area is given from this routine.

For example, assume that the shared memory area 8]2 is in the state shown in FIG. 4A. Here, the shared memory area 812
is divided into areas A, B, C, and D, and each area is provided with a header section 814. For example, when acquiring an area whose size is 15 and whose management number is 5, it is first checked whether the area whose management number is 5 has already been allocated. That is, starting from area A, the management numbers of the headers with tags numbered 10 are checked. Here, if there is no area with the management number 5, then it is determined which area to acquire. The acquisition conditions are that the tag is -, the size is 15+8 (8 is the length of the header) or more, and the area with the smallest size margin is searched. In this case, area p having an area size of 24 is optimal for the conditions. The size of the area is 24, and the remainder is 1, but since this cannot be left as an unused area, the remaining area is discarded. Next, change the tag of area D to 10,
The management number is set to 5, and the start address and end information of area D are returned as a return code. In addition, in the above case, the unit of acquisition is set to 8 bytes Il1,
If the lower 3 bits are rounded up and obtained as 24, the remainder will be 0, so no fraction will remain.

Next, it is assumed that the shared memory area 812 is in the state shown in FIG. 4B. Here, consider a case where an attempt is made to acquire an area whose size is 20 and whose management number is 7.

Since the area with management number 7 is not allocated here, area C is selected, which has a tag of -, a size of 20+8 or more, and a minimum size margin. Since the size of area C is 56, area C is divided to obtain area C1 of size 28 and leave an unused area C2 of size 28. This state is shown in FIG. 4C. Note that a header section is added to area C2.

Next, memory release will be explained. When releasing an unnecessary area, a management number is specified and a predetermined routine is called. For example, if the shared memory area 812 is
Assume that it is in the state shown in the figure. Here, the management number is 2.
Assume that area B is to be released. Since the tags of areas A and C before and after area B are 10, the tag of area B to be released is changed to - and the process ends. The released state is shown in FIG. 5B.

Next, it is assumed that the shared memory area 812 is in the state shown in FIG. 5C. Here, if area B is to be released, the tag of area A before area B is -, so
Change the tag of area B to - and combine area A and area B. As a result, area B1 is obtained. The state after combination is shown in FIG. 5D. Then, the size of area A and the size of area B are added to obtain the size of area B1 and the size data is updated, and the next area pointer of the header part of area B1 and the previous area pointer of the header part of area C are and update.

Next, assume that the shared memory area 812 is in the state shown in FIG. 5E. Here, if area B is to be released, the tag in the header of area C, which follows area B, is -, so change the tag in the header of area B to - and release area B and area C. Combine. As a result, area B2 is obtained.

The state after combination is shown in Fig. 5F. Then, the size of area B and the size of area C are added to obtain the size of area B2 and the size data is updated, and the next area pointer of the header part of area B2 and the previous area pointer of the header part of area D are and update.

Next, assume that the shared memory area 812 is in the state shown in FIG. 5G. Here, if area B is to be released, the tag of the header section of area A before area B is -, and the tag of the header section of area C next to area B is -, so area B Change the tag to - to area A,
Combine B and C. As a result, area B3 is obtained. The state after combination is shown in Figure 5H. And area A
.

The size of area B3 is determined by adding the sizes of B and C to update the size data, and the next area pointer of the header part of area B3 and the previous area pointer of the header part of area D are updated. After updating, if the update ends normally, a return code to that effect is returned, and if it does not end normally, a return code indicating that there is no area with the specified management number is returned.

Note that when checking whether or not the area has been allocated, a management number is designated and a predetermined routine is called. In this case, the header sections with the tag 10 are sequentially examined to find one that matches the given management number. and,
Returns a code indicating whether the allocation has been completed or not as a return code.

6 and 7 are flowcharts for explaining the operation of one embodiment of the present invention, in particular, FIG. 6 shows the memory initialization operation, and FIG. 7 shows the memory acquisition operation. . Next, the memory initialization operation will be described with reference to FIGS. 1 to 6.

The side that uses the shared memory on the program passes the start address of the shared memory, the size of the shared memory, and a pointer indicating the start address of the reference remaining capacity table to the memory management side on the program. The memory management side creates an area control block 8]1 shown in FIG. 2 in response to receiving these data. In other words, set the tag to -,
The size is set to N, and the management number primary area pointer and previous area pointer are each set to O. Subsequently, the shared memory area 8]2 is brightened, that is, cleared to 0.

Next, the memory acquisition operation will be explained with reference to FIGS. 1 to 7. First, the memory usage side that requests the use of the memory gives the memory management side the acquired size n, the management number, and the ID number m. In response, the memory management side performs the memory acquisition operation described in FIGS. 48 to 4C. If the memory cannot be acquired, other processing is performed, but if the memory can be acquired, the next area pointer is searched for the one with the tag of - and the size or maximum. Then, based on the D number, read out the corresponding standard maximum remaining block length from the standard remaining m table 82;
The size of the largest area searched (SMAX) and the corresponding i! The size is compared with the semi-maximum remaining block length (S). If SMAX>S, the acquisition is approved, a pointer to the acquisition area and a return code indicating that the acquisition is OK are returned, and the operation ends. On the other hand, if SMAX≦S, the memory release operation described in FIGS. 5A to 5H is performed without permitting memory acquisition. Next, a pointer of 0 indicating that the memory could not be acquired and a return code indicating NG are returned. In this case, if the shared memory acquisition operation is performed from the beginning, the area released here can be acquired.

[Effects of the Invention] As described above, according to the present invention, when acquisition of a storage area necessary for processing is requested, the size of the unused area of the memory is determined by the size of the acquisition standard data predetermined for each process. It is determined whether or not an unused area greater than the acquisition criterion is secured in memory by comparing whether or not it is larger than the acquisition criterion. Even if the area is larger than the size of the requested area, the acquisition is not allowed, so if you set the minimum acquisition standard data for high priority processing,
High priority tasks and modules can acquire memory at any time.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a transaction processing system to which an embodiment of the present invention is applied. FIG. 2 is a configuration diagram of the shared memory shown in FIG. 1. FIG. 3 is a configuration diagram of the reference remaining m table shown in FIG. 1. FIGS. 48 to 4C are diagrams for explaining acquisition of the memory areas shown in FIG. 2. 5A to mSH are diagrams for explaining release of the shared memory area shown in FIG. 2. FIG. FIG. 6 is a flow diagram showing the memory initialization operation. FIG. 7 is a flow diagram showing the memory acquisition operation. FIG. 8 is a block diagram of a conventional transaction processing system. In the figure, 1 is a cluster controller, 2 is a transmission line, 3 is a POS @ terminal device, 4 is a telephone line network, 5 is a host computer, 6 is a CPU, 7 is a ROM, 8 is a RAM,
81 is a shared memory, 82 is a reference remaining capacity table, 811 is an area control block, 812 is a shared memory area, and 814 is a header section.躬4A0 6th 'fD (Memotsu 1 jump 1 艷j) iゞ') The low p period and the 8th concave flame comes 1 (31, Sunrishi enters deniuden heheri
.

Claims (2)

[Claims] (1) A memory management device that manages memory shared by a plurality of processes, comprising a storage means that stores acquisition standard data for each process, which is a criterion for whether or not to acquire the memory, and a storage area necessary for the process. When acquisition is requested, the acquisition standard data corresponding to the process stored in the storage means is compared in size with the unused area of the memory, and an unused area equal to or greater than the acquisition standard is secured in the memory. a determination unit for determining whether or not an unused area of the memory is requested to be acquired in response to the determination by the determination unit that an unused area exceeding an acquisition criterion is not secured in the memory; A memory management device comprising a control means for controlling not to permit acquisition even if the size is larger than the size of the memory management device. (2) The memory management device according to claim 1, further comprising a setting change means for changing the settings of acquisition reference data for each process stored in the storage means.