JPH0218637A - Memory control method - Google Patents

Abstract

PURPOSE:To shorten the processing time for control of a memory by dividing a memory space into page units and allocating only the memory blocks having the same size and the same structure into one page. CONSTITUTION:The control information on each page unit of pages 41-43 allocated with the memory blocks A1-A6 or B1 and B2 having the same size and the same structure are stored in a page control block 1. While a memory block control block 2 stores the control information on each block unit of the memory blocks A1-A6 and B1 and B2 which are allocated to the pages 41-43. At the same time, the idle memory block information obtained when the blocks A1-A6, B1 and B2 are deallocated is also stored in the block 2. Then an idle page control block 3 stores the idle page information obtained when the blocks A1-A6 or B1 and B2 of the same kind are all deallocated so as to ensure the reuse of undesired pages 41-43. In such a way, the processing time is shortened.

Description

Detailed Description of the Invention [Field of Industrial Application] The present invention relates to a method of dynamically allocating memory and managing each allocated memory block, and particularly to a method of dynamically allocating memory and managing each allocated memory block. Concerning memory management methods in an environment where sufficient memory can be secured.

[Conventional technology 1] Conventionally, in this type of memory management method, memory blocks allocated for the same purpose are managed in a unidirectional list structure starting from a directory, and various types of memory blocks are mixed in the memory. Therefore, allocation of a new memory block, deallocation of a specific memory block, and garbage collection after deallocation were all performed while following the chain between each memory block.

[Problem to be solved by the invention] The conventional memory management method described above sequentially allocates memory blocks in the memory space in order to improve memory efficiency, and various types of memory blocks are allocated in a mixed manner. When deleting all memory blocks, it is necessary to deallocate each memory block one by one while tracing the chain between memory blocks, and it is necessary to perform garbage collection while being aware of the continuity of the deallocated memory blocks. Yes, and when allocating a new memory block, if there is a memory block that was previously deallocated, check to see if there is an area where it is possible to allocate the memory block without allocating it while tracing the chain between the freed memory blocks. , when there is no available area for allocation, it is allocated to unused memory, which has the disadvantage of increasing processing time.

[Means for Solving the Problems] The memory management method of the present invention divides the memory space into pages, and allocates only memory blocks of the same size and structure within one page.

[Effect] Only memory blocks of the same type are allocated to the memory space divided into pages, so when deleting all memory blocks of a certain type, each memory block is deleted without tracing the chain between memory blocks. There is no need to deallocate each memory block one by one, there is no need for garbage collection when memory blocks are deallocated, and when new memory blocks are allocated,
Since there is no need to allocate a chain between freed memory blocks from scratch, processing time is shortened.

[Example 1 Next, an example of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram of an embodiment of the memory management method of the present invention, and FIG. 2 shows each information stored in the page management block 1 and memory block management block 2 of FIG. 1 and each page 41, 42 of the memory. .43 is a diagram showing the relationship between memory blocks A1 to Aa, B+, and B2 allocated to memory blocks A1 to Aa, B+, and B2.

In FIGS. 1 and 2, "Prison" indicates an empty area, and "Second" indicates a deallocated area.

Page management block 1 includes memory blocks A1 to A6 or B of the same type (same size and structure),
Management information for each page of B2 or allocated pages 41, 42, and 43 is stored. The memory block management block 2 includes memory blocks A1 to As, B1 . Management information for each block of B2 is stored, and memory block A
Free memory block information when 1 to As, Bl, and B2 are deallocated is stored.

Free page management block 3 contains pages 4 that are no longer needed.
In order to reuse memory blocks A1 to A6 or B1 and B2 of the same type, free page information is stored when all of the same type of memory blocks A1 to A6 or B1 and B2 are deallocated.

In FIGS. 1 and 2, a first page pointer 11.
Last page pointer 12. First memory block pointer 21. Last memory block pointer 22 Other pointers and free memory block chain 23. A free page chain 31 is shown. Also, memory block A
The suffixes in B1 to A6 and B1 and B2 respectively indicate the order in which the memory blocks were allocated.

When allocating a new memory block, as shown in FIG. 2, there is a free memory block chain 23 of the same type of memory block in the memory block management block 2, so the free memory block A2 of page 41 is reused to allocate memory. Allocate blocks. At this time, since pages are allocated for each of the same type of memory blocks A, ~A6 and B, , B2, there is no need to search for free memory blocks, and the first free memory block A2
can be reused. If there is no free memory block A2 or A5, a memory block is allocated to an unused area in page 42, and if there is no area to allocate a memory block to an unused area of page 42, as shown in FIG. If there is a free page chain 31 of the free page management block 3, the first free page 44 is allocated, and if there is no free area on pages 44 and 45, one new page of memory is allocated, and the allocated page (not shown) is allocated. Allocate a new memory block.

One memory block A1-A6. B1. When deallocating B2, memory blocks A1 to As B
Delete it from the chain between I and B2 and add it to the chain of free memory blocks. Deallocated memory blocks A1-A6. Page 41 including B+ and B2.
Since only memory blocks of the same type are allocated to 42 and 43, garbage collection is not necessary.

All memory blocks A, ~8 of the same type. or B
When dealinglocating B2, all the pages 41 to 45 including memory blocks for deallocation are connected to the free page chain 31 of the free page management block 3, and the corresponding page management program Nk1 is initialized. and each memory block A1 to A6 . B1.
There is no need to process each B2.

As mentioned above, the memory space is managed by dividing it into pages 41, 42 or 43, and only 92 memory blocks of the same type, A1 to A6 or B, are allocated to one page. , memory management processing is sped up.

[Effects of the Invention] As explained above, the present invention manages memory space by dividing it into pages, and allocates only memory blocks of the same size and structure within one page. Since blocks are not mixed, when deleting a memory block, one
There is no need to deallocate each memory block one by one, and there is no need to perform garbage collection when deallocated memory blocks are consecutive, and when allocating a new memory block, it is allocated while following the chain between free memory blocks. Since there is no need to check the target area, there is an effect that the speed of memory management processing can be significantly increased.

[Brief explanation of drawings]

FIG. 1 is a block diagram of an embodiment of the memory management method of the present invention, and FIG. 2 shows each information stored in the page management block 1 and memory block management block 2 of FIG. 1 and each page 41, 42 of the memory. .43 allocated memory blocks A1 to Aa, B1. It is a figure which shows the relationship with B2. l...Page management block, 2...Memory block management block, 3...Empty page management block, 11...First page pointer, 12...Last page pointer, 21... ...Start memory block pointer, 22...
- Final memory block pointer, 23... Free memory block chain, 31... Free page chain, 4], 42.43.44,45... Page, A1. A2. A3. A4. As, A6・
...Memory block, B1. B2・・・・・・・・・
・・・・・・・・・・・・Memory block.

Claims (1)

[Claims] 1. When allocating a new memory block in the memory space, if there is a deallocated memory block, the memory block is reused, and if there is no deallocated memory block, a new memory is dynamically allocated. In a memory management method that allocates memory blocks by allocating memory blocks and then deallocates and reuses unnecessary memory blocks, the memory space is divided into pages, and only memory blocks of the same size and structure are allocated within one page. A memory management method characterized by allocation.