JP3356090B2 - Memory management method, memory management device, and recording medium - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a memory management method, a memory management device, and a recording medium, and more particularly, to an operating system (OS).
The present invention relates to a memory management method, a memory management device, and a recording medium that enable allocation and release of a memory of an arbitrary size in a real-time OS intended to guarantee processing time in the memory.

[0002]

2. Description of the Related Art In general, a memory management method of a real-time OS is such that a memory is divided into a certain fixed size, and a memory is dynamically allocated in response to a memory acquisition / release request from a task. They are managed by block.

Conventionally, such a memory block management method is as follows, for example, as disclosed in Japanese Patent Laid-Open No. Hei 6-309197. That is, when a memory acquisition request from a task occurs, a free memory block satisfying a required size is detected and assigned to the task. On the other hand, when a memory release request is issued, a release position in all empty memory blocks is detected in order to prevent fragmentation of the memory blocks. If there are consecutive free memory blocks before and after the memory block requested to be released, the free memory blocks are linked to these free memory blocks.
It is managed as two free memory blocks.

[0004]

However, in the conventional memory management system, it is necessary to connect the memory blocks before and after the empty memory block in order to prevent the memory from being fragmented when the memory is released. Since the release position is detected, the occupation time of the CPU becomes long.

[0005] The reason is that when a desired free memory block is connected near the beginning of a free memory block, processing is performed at high speed. However, when a desired free memory block is detected near the end of the free memory block. Is because the detection time becomes longer.

SUMMARY OF THE INVENTION In view of the above, it is an object of the present invention to reduce the number of searches for detecting a release position from the entire memory area when a memory is released by a task, thereby shortening the occupation time of the CPU. It is intended to realize high-speed memory release processing and to improve the performance of a real-time OS.

[0007]

According to a first aspect of the present invention, there is provided a memory management method for allocating and releasing a memory to a task in response to a request from the task. The memory is divided into used memory that has been used and unused memory that has been released, and is managed in block units. The used memory is connected in the order in which the memory was acquired, the unused memory is connected in the order of the addresses, and the used memory is determined. When releasing a block, save the start address of the block to be released and, based on the start address of the stored block, detect the release position of the next block. If the start address is
It is determined whether the address is
The start address of the next block to be released and released is stored.
Is determined not to be greater than the starting address
Memory release position in the direction of smaller addresses.
The address of the next block to be searched and released is saved.
Is determined to be larger than the starting address
Search for a memory release position in the direction of larger address.
Do, each block is at least connected to that block
Start address of the next block that is
The start address of the block before
Start address of the user-usable area and the user
Table that contains the end address of the
Memory is used, and the used memory is the first block of the used memory.
Start address of the memory and the end of the last block of the used memory
It is managed by a table consisting of the
Memory is the start address of the first block of unused memory
And from the start address of the last block of unused memory
It is characterized by being managed by a table . Claim
Item 2 is a memory management device for allocating and releasing memory to a task in response to a request from the task, wherein the memory is allocated to a used memory that has been allocated, and a memory to be released. is managed in block units is divided into unused memory is a memory used, joint technique in order that won memory, the unused memory, and tether memory management unit in address order, a given block of memory used when releasing the, based on a storage means for storing the start address of the block to be released, the head address of a block stored by the storage means, when performing detection of the release position of the next block, is released Of the next block
The start address is the start address stored by the storage unit
The next address to be released
The start address of the block is stored by the storage means.
If it is determined that the address is not larger than
Search for a memory release position in the direction of smaller addresses.
The address of the next block to be released
Is determined to be larger than the first address stored by
When the address is released, the memory is released in the larger direction
E Bei and detection means for searching for positions, each block is small
At least, the next block connected to that block
Start address and start address of the block before that block
Dress and the first address of the user-usable area of that block.
Dress and the end address of the user usable area of that block.
Is managed by a table consisting of dresses and used memory
Is the start address of the first block of the used memory and the
Data consisting of the start address of the last block of the
Unused memory is managed by the
Start address of the first block and end of unused memory
Managed by a table consisting of the start addresses of the blocks
It is characterized by being performed. The recording medium according to claim 3.
Memory for the task, as requested by the task
Memory management methods for allocating and releasing
Computer that stores programs to be executed by
Data readable recording medium, the memory of which is
Used memory that has been allocated, and unused memory that has been released.
Memory and manage it in block units.
Memory in the order in which they acquired memory, and add unused memory.
And release the specified block of the used memory
When the start address of the released block is saved,
Based on the start address of the existing block,
When detecting the release position of a block, the next
The start address of the block is
The next block to be released.
The start address of the
When it is determined that the address is not large, the one with the smaller address
Search for the memory release location in the
Lock address is higher than the stored first address
When it is determined that the address is larger, the direction of the address is larger
Then, a search for the memory release location is performed, and each block is
At least, the end of the next block connected to that block
Start address and start address of the block before that block
Address and the start address of the user-usable area of the block.
Address and the end address of the user-usable area for that block.
Is managed by a table consisting of
Is the start address of the first block of the used memory and the
Data consisting of the start address of the last block of the
Unused memory is managed by the
Start address of the first block and end of unused memory
Managed by a table consisting of the start addresses of the blocks
The computer to perform the required memory management method
Is recorded. In the memory management method, the memory management device, and the recording medium according to the present invention, the memory is divided into a used memory that has been allocated and an unused memory that has been released, and is managed in units of blocks. When the memory is connected in the order in which the memory was acquired, the unused memory is connected in the order of the addresses, and when the predetermined block of the used memory is released, the start address of the released block is stored, and the start address of the stored block is used as a reference. When detecting the release position of the next block, the start address of the next block to be released
Is greater than the stored start address.
The start address of the next block that is
Is determined not to be greater than the existing start address
When the address becomes smaller,
The address of the next block to be searched and released is stored.
Is determined to be larger than the existing start address
Memory release position in the direction of larger addresses.
Do a search and each block at least
Start address of the next connected block and its block
Address of the block before the block and the
The start address of the user usable area and the block
Table consisting of the end address of the user usable area
Is used, and the used memory is the first block of the used memory.
Lock start address and last block of used memory
Is managed by a table consisting of
Memory is the first address of the first block of unused memory.
Address and the start address of the last block of unused memory
Is managed by a table consisting of

[0008]

Embodiments of the present invention will be described below in detail with reference to the drawings. FIG. 1 is a block diagram showing an example of a hardware (HW) configuration of a microcomputer to which a memory management method according to the present invention is applied. As shown in FIG. 1, the microcomputer includes a CPU (central processing).
1 includes a memory controller 5, a timer controller 6, and an interrupt controller 7 connected to the CPU 1 via an A / D (address / data) bus 9.

Further, the microcomputer has a CPU 1
(Read) that stores the instruction code for operating
only memory 2 and a main memory (RAM (random memory) for temporarily storing various data necessary for the operation of the CPU 1 and loading a program.
(access memory)) (hereinafter, simply referred to as a memory) 3 and a storage device 4 for storing data.

Next, an example of the configuration of software (SW) operating in the HW configuration will be described with reference to the functional block diagram of FIG. SW operating in the HW configuration of FIG.
An interrupt handler 10 for processing an interrupt generated from the interrupt controller 7 of FIG. 1 and an exception handler 11 for processing an exception generated in the CPU 1 are controlled by a plurality of users 14 in which necessary processing is described. OS (operating s) that controls
system 12). Further, the real-time OS 12 is provided with a memory release position management mechanism 13.

The real-time OS 12 determines a task to be operated in real time in response to an external event that has occurred, and transfers an operation right to the task, thereby controlling the smooth operation of the system. Further, as described above, the real-time OS 12 includes the memory release position management mechanism 13, and manages the memory block released by the task.

Next, the operation of the embodiment shown in FIG. 1 will be described. In the above embodiment, the CPU 1
Interruption by the timer controller 6 at fixed time intervals,
Interrupt controller 7 connected to external interrupt factor
It is necessary to perform appropriate processing for various external factors such as interrupts from the Internet.

In order to realize this, software operating on the CPU 1 is controlled by a real-time OS 12 as shown in FIG. Then, the real-time OS 12 manages a plurality of tasks for processing the generated event, and determines and executes a necessary task 14 for various generated events.

The task 14 statically secures a minimum required work area, but dynamically acquires and releases a temporarily required memory dynamically when required. Task 1 when you need it
By using the memory 3, the memory capacity can be saved.

As shown in FIG. 3, the memory 3 is managed by the real-time OS 12, and the task 14 acquires and releases the memory 3 via the real-time OS 12. When a memory acquisition request is issued from task 14, real-time O
The memory control unit 23 in S12 searches for a free memory block to detect a block that satisfies the request size, and gives the detected memory block address to the task 14 that has issued the request, thereby giving the memory block an address.

The head address of the memory block released most recently (most recently) is held in the memory release position management mechanism 13.
Thus, each time a memory block is released, the memory block address is updated, and the address of the most recently released memory block is held.

When a memory release request is issued from the task 14, the real-time OS 12 needs to search for the release position of the memory block 14, but the memory release position management mechanism 13 sends the start address of the previously saved memory block. And compares it with the start address of the memory block to be released. By this comparison, if the head address of the memory block to be released and the previously released memory block are continuous, the release position can be detected at high speed.

If the memory release processing is not continuous,
For example, when the memory acquisition process is performed during the memory release process, in the memory acquisition process, the head address of the memory block stored in the memory release position management mechanism 13 is returned to a predetermined initial value, and the next When a memory release request is issued, the conventional search is performed.

FIG. 4 is a diagram showing a configuration example of the memory block management table. Memory 3 has already been assigned to task 14
Is divided into the size of the requested memory block. Each memory block is provided with a memory block management table. The memory block management table includes a next block 31 indicating the start address of the next block, a previous block 32 indicating the start address of the previous block, and a task. Address 33 and last address 3 indicating the range of the memory area usable by
4 and a block use status flag (hereinafter simply referred to as a flag) 35 indicating the state of the memory block. When "F (Free)" is stored in the flag 35, it indicates that the memory is free and "U (Free)" is set.
When (Used) "is stored, it indicates that the memory is in use.

FIG. 5 is a diagram showing the relationship between the memory management table and the memory blocks. Empty blocks are connected in order of address, and blocks in use are connected in the order in which the blocks were acquired. The free memory management table 41 for managing the entire free memory includes a free memory first block 42 which is a field for holding the address of the first block of the free memory, and a free memory last block which is a field for holding the address of the last block of the free memory. 4
3, the memory head block 42 indicates the head block of the free memory, and the free memory last block 43 indicates the last block of the free memory.

The used memory management table 44 for managing the entire used memory includes a used memory first block 45 which is a field for holding the address of the first block of the used memory, and a field for holding the address of the last block of the used memory. A certain used memory last block 46 is provided. The used memory head block 45 indicates the head block of the used memory, and the used memory last block 46 indicates the last block of the used memory.

FIG. 6 is a diagram showing the relationship between the memory 3 and the memory release position management mechanism 13. As shown in the figure, the most recently released free memory block in the memory 3 is indicated by a memory release position pointer in the memory release position management mechanism 13.

The task 14 dynamically acquires and releases the memory 3 when necessary. Since the real-time OS 12 manages the memory 3 as shown in FIG.
Performs acquisition and release of the memory 3 via the real-time OS 12. When a memory acquisition request is issued from task 14,
The real-time OS 12 detects a free memory block that satisfies the memory size (request size) requested by the task 14 by tracing the memory block from the free memory head block 42 of the free memory management table 41 as shown in FIG.

The size of the memory block is obtained from the following equation (1).

Memory block size = (last address 34) − (start address 33) (1)

If the free memory block satisfies the required size, the memory block is given to the task 14 and the process is terminated. On the other hand, if the required size is not satisfied, the next block 31 is sequentially traced, the size of the memory block is calculated by the equation (1), and the search is performed until a memory block satisfying the required size is detected.

The memory release position management mechanism 13 holds the head address of a memory block. Every time a memory block is released by the memory release position management mechanism 13, the memory block address is updated and the memory block address is updated most recently. The address of the memory block is held.

When a memory release request is issued from the task 14, the real-time OS 12 needs to search for a release position of a memory block in order to prevent block fragmentation. At this time, the head address of the previously saved memory block is extracted from the memory release position management mechanism 13 and compared with the head address of the memory block to be released. By this comparison, if the memory block to be released is continuous with the previously released memory block, it is not necessary to search for the release position, so that the release position can be detected at high speed.

In the memory-related processing, when the memory release processing is not continuous, for example, when the memory acquisition processing is executed after the memory release processing, the memory release position management mechanism 13 saves the data in the memory acquisition processing. The memory block start address is returned to a predetermined initial value NULL. Then, when a memory release request is generated next, the free memory first block 42
Search from.

This is because when a memory acquisition request occurs,
The possibility that a memory block of a continuous address is released is reduced, and the memory 3 may be acquired from an address managed by the memory release position management mechanism 13, which complicates the processing. , Real-time O
This is to prevent the performance of S12 from deteriorating.

When a memory release request is generated, the real-time OS 12 checks in step 61 whether the address held in the memory release position management mechanism 13 is the initial value NULL, as shown in the flowchart of FIG. Is performed. When the address held in the memory release position management mechanism 13 is NULL,
Proceed to step 65.

At step 65, the free memory start block 42 of the free memory management table 41 is
The release position is searched in order from. If the following expression (2) is satisfied, the start address of the released free memory block is
Registered in the free memory head block 42 of the free memory management table 41, the next block 3 of the free memory block
1, the address registered in the free memory head block 42 of the free memory management table 41 is registered. Until that time, the head address of the currently released memory block is registered in the block 32 before the free memory block specified by the address of the free memory head block 42 in the free memory management table 41.

(Address stored in free memory first block 42)> (first address of memory block to be released) (2)

On the other hand, in step 61, if the address held by the memory release position management mechanism 13 is the initial value NULL and does not satisfy the expression (2), the free memory is held until the following expression (3) is satisfied. From the first block,
Find the release position. That is, the head address stored in the next block 31 of each block is sequentially compared with the head address of the memory block to be released from the head block of the free memory, and when the following expression (4) is satisfied for the first time,
The release position is immediately before the block. When the processing in step 65 is completed, the process proceeds to step 66, where the released memory block is inserted into the release position.

(Start address of memory block to be released)> (Start address of next block 31) (3)

(Start address of next block 31)> (Start address of memory block to be released) (4)

On the other hand, when the address held by the memory release position management mechanism 13 is not the initial value NULL, the released memory block may exist near the address held by the memory release position management mechanism 13. Is high.
Therefore, based on the address indicated by the pointer in the memory release position management mechanism 13, the release position of the memory block to be released is determined in advance by the following equation (5) or (6). To detect the release position.

(Memory release position management mechanism storage address)> (first address of memory block to be released) (5)

(Start address of memory block to be released)> (Address stored in memory release position management mechanism) (6)

That is, in step 62, it is determined whether or not the address of the released memory block is larger than the address stored in the memory release position management mechanism 13. If it is determined that the address of the released memory block is not larger (smaller) than the address stored in the memory release position management mechanism 13 (satisfies the above expression (5)), the process proceeds to step 63, where the address becomes higher. Searches the memory release position in the smaller direction.

On the other hand, if it is determined that the address of the released memory block is larger than the address stored in the memory release position management mechanism 13 (satisfies the above expression (6)), the process proceeds to step 64, where the address is determined. Search for a memory release position in the larger direction.

After detecting the release position, in step 66, the released memory block is inserted into the detected release position. Next, in step 67,
Whether the end address 34 of the previous memory block indicated by the address stored in the block 32 before the release position memory block matches the start address of the released memory block, that is, whether the inserted memory block It is determined whether there is an adjacent free memory block.

As a result, the end address 34 of the previous memory block indicated by the address stored in the block 32 before the memory block at the release position matches the start address of the released memory block, that is, the insertion address. If it is determined that there is a free memory block adjacent to the selected memory block, the process proceeds to step 68, where the free memory block indicated by the address stored in the previous block 32 is combined.

If the start address of the next memory block indicated by the address stored in the block 31 next to the memory block at the release position matches the end address 34 of the released memory block, the next block It is linked with the next memory block indicated by the address stored in 31. Then, a new free memory block is set.

Then, in step 69, the head address of the free memory block formed as described above is held by the memory release position management mechanism 13.

If it is determined in step 67 that there is no free memory block adjacent to the inserted memory block, the head address of the inserted memory block is stored in the memory release position management mechanism 13.

As described above, the following effects can be obtained by the above embodiment. That is, when memory blocks are successively released, the number of times of searching for the memory block release position is reduced, so that the time of the CPU 1 occupied by the memory release processing can be reduced, and the processing capability of the real-time OS 12 can be increased. Can be.
The reason is that, since the head address of the latest released memory block is stored, the number of memory block searches required to detect the release position of the memory block can be reduced.

The memory block size may be fixed and divided by a memory block management method, and a memory release position management mechanism may be provided.

Further, in the above embodiment, the case where the memory space is divided into the used memory and the free memory has been described, but the memory may be used for another purpose.

Further, in the above embodiment, the case where the present invention is followed by a microcomputer has been described, but the present invention is not limited to this.

[0051]

As described above, according to the memory management method, the memory management device, and the recording medium of the present invention, the memory is divided into the used memory allocated and the unused memory released. And manage them in blocks,
When the used memory is connected in the order in which the memory was acquired, the unused memory is connected in the order of the addresses, and when releasing a predetermined block of the used memory, the leading address of the released block is stored, and the leading address of the saved block is stored. as a reference, the case of detecting the release position of the next block, is released
The start address of the next block is
It is determined whether the address is greater than the head address and the next
The start address of the block is
Address is not greater than the address,
In the smaller direction, search for the memory release location and release
The address of the next block is
Address is larger than the address
Search for the memory release position in the direction
Is at least the next block connected to that block.
Block start address and the block before that block
The start address and the user-usable area of the block
The start address and the user-usable area of the block
Used and managed by a table consisting of end addresses
Memory is the start address of the first block of the used memory
From the start address of the last block of the used memory.
Unused memory is managed by a table
Start address of the first block of memory and unused memory
Table consisting of the start address of the last block of
The number of searches for detecting a release position from all memory areas can be reduced, the time occupied by the CPU can be shortened, the memory release processing can be speeded up, and the real-time OS can be managed. Performance can be improved.

[Brief description of the drawings]

FIG. 1 is a diagram showing a hardware configuration of an embodiment of a microcomputer to which a memory management method of the present invention is applied.

FIG. 2 is a diagram illustrating a software configuration of an embodiment of a microcomputer to which a memory management method according to the present invention is applied;

FIG. 3 shows a task and memory acquisition of the real-time OS 12,
It is a figure explaining release processing.

FIG. 4 is a diagram illustrating a configuration example of a memory block management table.

FIG. 5 is a diagram showing a relationship between memory management tables 41 and 44 and memory blocks.

FIG. 6 is a diagram showing a relationship between a memory 3 and a memory release position management mechanism 13.

FIG. 7 is a flowchart illustrating a processing procedure of a memory release processing.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 CPU 2 ROM 3 Main memory 4 Storage device 5 Memory controller 6 Timer controller 7 Interrupt controller 9 A / D bus 10 Interrupt handler 11 Exception handler 12 Real-time OS 13 Memory release position management mechanism 14 Task 23 Memory control unit 31 Next block 32 Previous block 33 Start address 34 End address 35 Block use status flag 41 Free memory management table 42 Free memory first block 43 Free memory last block 44 Used memory management table 45 Used memory first block 46 Used memory last block

──────────────────────────────────────────────────続 き Continued on front page (58) Field surveyed (Int. Cl. ⁷ , DB name) G06F 12/02 540

Claims (3)

(57) [Claims] 1. A memory management method for allocating and releasing memory to a task in response to a request from the task, comprising:
When the memory is divided into free unused memories and managed in block units, the used memories are connected in the order in which the memories are acquired, the unused memories are connected in the order of addresses, and a predetermined block of the used memory is released. If the save the start address of the block to be released, based on the head address of the block that stores, to detect the release position of the next block, is released
Before the next block is saved.
It is determined whether the address is greater than the
The start address of the next block is
If it is determined that the address is not larger than the start address,
Search for the memory release position in the direction where the
The address of the next block to be freed is saved
If it is determined that the address is larger than the start address,
A search for a memory release position is performed in a direction in which the blocks are larger, and each block is connected to at least the block.
Address of the next block that is
Start address of previous block and user of that block
Start address of usable area and user of that block
According to the table consisting of the end address of the usable area
Managed, and the used memory is the first block of the used memory.
Start address and the end of the last block of the used memory
The unused memory is managed by a table including a head address, and the unused memory is a block at the head of the unused memory.
Block start address and the last block of the unused memory.
A memory management method characterized in that the memory management method is managed by a table including a head address of the memory. 2. A memory management device for allocating and releasing memory for a task in response to a request from a task, comprising: a memory used for the allocation;
Divided into the unused memory being freed managed in block units, the use memory, <br/> sequentially acquiring the memory tied technique, the unused memory, tether memory tube in address order
Management means; storage means for storing a start address of a released block when releasing a predetermined block of the used memory; and, based on a start address of the block stored by the storage means, When detecting the release position of a block, the start address of the next block to be released
Is the first address stored by the storage means.
The next block to be released.
The start address of the lock is stored by the storage means.
Is determined not to be greater than the starting address
Memory release position in the direction of smaller addresses.
The address of the next block to be searched and released is
From the start address stored by the storage means
When it is determined that the address is larger, the direction of the address is larger
In, e Bei and detection means for searching the memory release position, wherein each block, at least, is connected to the block
Address of the next block that is
Start address of previous block and user of that block
Start address of usable area and user of that block
According to the table consisting of the end address of the usable area
Managed, and the used memory is the first block of the used memory.
Start address and the end of the last block of the used memory
The unused memory is managed by a table including a head address, and the unused memory is a block at the head of the unused memory.
Block start address and the last block of the unused memory.
A memory management device characterized in that the memory management device is managed by a table including a head address of the memory. 3. The task according to a request from the task.
Management that allocates and frees memory for memory
Write a program that causes a computer to perform the method.
A recorded computer-readable recording medium, wherein the memory is used memory to which allocation is performed,
Divided into free unused memory and block units
And connect the used memories in the order in which the memories were acquired.
When the used memory is connected in the order of addresses and a predetermined block of the used memory is released, the release is performed.
The start address of the block to be stored is stored, and based on the start address of the stored block.
To detect the release position of the next block,
Before the next block is saved.
It is determined whether the address is greater than the
The start address of the next block is
If it is determined that the address is not larger than the start address,
Search for the memory release position in the direction where the
The address of the next block to be freed is saved
If it is determined that the address is larger than the start address,
A search for a memory release position is performed in the direction in which the
Address of the next block that is
Start address of previous block and user of that block
Start address of usable area and user of that block
According to the table consisting of the end address of the usable area
Managed, and the used memory is the first block of the used memory.
Start address and the end of the last block of the used memory
The unused memory is managed by a table including a head address, and the unused memory is a block at the head of the unused memory.
Block start address and the last block of the unused memory.
Messages managed by a table consisting of
A program for causing the computer to execute the memory management method.
Computer-readable recording medium on which the program was recorded
body.