Memory pool (fixed-size block allocator) in Embedded C

Introduction

A memory pool helps manage memory efficiently by dividing a big chunk into small fixed-size blocks. It makes allocating and freeing memory faster and predictable.

When you need to allocate many small objects of the same size repeatedly.

In embedded systems where memory is limited and fragmentation must be avoided.

When you want fast and simple memory allocation without using the general heap.

For real-time systems where predictable allocation time is important.

Syntax

Embedded C

typedef struct MemoryPool {
    void *pool_start;
    size_t block_size;
    size_t total_blocks;
    void *free_list;
} MemoryPool;

void MemoryPool_init(MemoryPool *mp, void *buffer, size_t block_size, size_t total_blocks);
void *MemoryPool_alloc(MemoryPool *mp);
void MemoryPool_free(MemoryPool *mp, void *block);

The pool is created from a buffer you provide.

Blocks are all the same size for simplicity and speed.

Examples

Initialize a memory pool with 32 blocks of 32 bytes each inside a 1024-byte buffer.

Embedded C

MemoryPool mp;
char buffer[1024];
MemoryPool_init(&mp, buffer, 32, 32);

Allocate one block from the pool. Check if allocation succeeded.

Embedded C

void *block = MemoryPool_alloc(&mp);
if (block != NULL) {
    // use the block
}

Return the block back to the pool when done.

Embedded C

MemoryPool_free(&mp, block);

Sample Program

This program creates a memory pool with 8 blocks of 16 bytes each. It allocates all blocks, tries to allocate one more (which fails), frees one block, and then allocates again successfully.

Embedded C

#include <stdio.h>
#include <stddef.h>

typedef struct MemoryPool {
    void *pool_start;
    size_t block_size;
    size_t total_blocks;
    void *free_list;
} MemoryPool;

void MemoryPool_init(MemoryPool *mp, void *buffer, size_t block_size, size_t total_blocks) {
    mp->pool_start = buffer;
    mp->block_size = block_size;
    mp->total_blocks = total_blocks;
    mp->free_list = buffer;

    // Link all blocks in the free list
    char *p = (char *)buffer;
    for (size_t i = 0; i < total_blocks - 1; i++) {
        *(void **)p = p + block_size; // next pointer
        p += block_size;
    }
    *(void **)p = NULL; // last block points to NULL
}

void *MemoryPool_alloc(MemoryPool *mp) {
    if (mp->free_list == NULL) {
        return NULL; // no free blocks
    }
    void *block = mp->free_list;
    mp->free_list = *(void **)block; // move free_list to next
    return block;
}

void MemoryPool_free(MemoryPool *mp, void *block) {
    *(void **)block = mp->free_list;
    mp->free_list = block;
}

int main() {
    char buffer[128];
    MemoryPool mp;
    MemoryPool_init(&mp, buffer, 16, 8); // 8 blocks of 16 bytes

    void *blocks[8];

    // Allocate all blocks
    for (int i = 0; i < 8; i++) {
        blocks[i] = MemoryPool_alloc(&mp);
        if (blocks[i] != NULL) {
            printf("Allocated block %d at %p\n", i, blocks[i]);
        } else {
            printf("Failed to allocate block %d\n", i);
        }
    }

    // Try to allocate one more block (should fail)
    void *extra = MemoryPool_alloc(&mp);
    if (extra == NULL) {
        printf("No more blocks available\n");
    }

    // Free one block
    MemoryPool_free(&mp, blocks[3]);
    printf("Freed block 3\n");

    // Allocate again (should succeed)
    void *new_block = MemoryPool_alloc(&mp);
    if (new_block != NULL) {
        printf("Allocated new block at %p\n", new_block);
    }

    return 0;
}

Important Notes

Memory pools avoid fragmentation by using fixed-size blocks.

Always free blocks back to the pool to reuse memory.

Do not free blocks not allocated from the pool or double free.

Summary

A memory pool divides memory into fixed-size blocks for fast allocation.

It is useful in embedded and real-time systems to manage memory predictably.

Initialize the pool, allocate blocks, and free them when done.

Practice

1.

What is the main advantage of using a memory pool with fixed-size blocks in embedded systems?

easy

A. Unlimited memory usage

B. Automatic garbage collection

C. Faster and predictable memory allocation

D. Dynamic resizing of blocks

Solution

1. Step 1: Understand memory pool purpose

   A memory pool divides memory into fixed-size blocks to allocate memory quickly without fragmentation.

2. Step 2: Identify advantage in embedded systems

   Embedded systems need fast and predictable memory allocation, which memory pools provide.

3. Final Answer:

   Faster and predictable memory allocation -> Option C

4. Quick Check:

   Memory pool advantage = Faster and predictable memory allocation [OK]

Hint: Memory pools speed up allocation by fixed block sizes [OK]

Common Mistakes:

• Confusing memory pools with dynamic resizing
• Thinking memory pools provide garbage collection
• Assuming memory pools allow unlimited memory

2.

Which of the following is the correct way to declare a fixed-size memory pool array for 10 blocks of 32 bytes each in C?

?

easy

A. char pool[10][32];

B. char pool[32][10];

C. char *pool[10];

D. char pool[10*32];

Solution

1. Step 1: Understand array dimensions

   We want 10 blocks, each block is 32 bytes, so array should be 10 rows and 32 columns.

2. Step 2: Match declaration syntax

   char pool[10][32]; declares 10 blocks of 32 bytes each correctly.

3. Final Answer:

   char pool[10][32]; -> Option A

4. Quick Check:

   Array size = 10 blocks x 32 bytes [OK]

Hint: Rows = blocks, columns = block size in bytes [OK]

Common Mistakes:

• Swapping dimensions (32 and 10)
• Using pointer array instead of fixed blocks
• Declaring a single large array without blocks

3.

Given this simplified memory pool allocator code snippet, what will be the output?

char pool[3][16];
int free_blocks = 3;

char* allocate() {
    if (free_blocks > 0) {
        return pool[--free_blocks];
    }
    return NULL;
}

int main() {
    char* p1 = allocate();
    char* p2 = allocate();
    char* p3 = allocate();
    char* p4 = allocate();
    printf("%p %p %p %p\n", p1, p2, p3, p4);
    return 0;
}

medium

A. Three valid pointers and one NULL pointer

B. All pointers are NULL

C. All pointers point to the same block

D. Compilation error due to missing headers

Solution

1. Step 1: Trace allocation calls

   Initially free_blocks=3. Each allocate() call decrements free_blocks and returns pool[free_blocks]. So p1=pool[2], p2=pool[1], p3=pool[0], p4=NULL because free_blocks=0.

2. Step 2: Understand output

   First three pointers are valid addresses, last pointer is NULL.

3. Final Answer:

   Three valid pointers and one NULL pointer -> Option A

4. Quick Check:

   3 blocks allocated, 4th returns NULL [OK]

Hint: Allocation decreases free_blocks; NULL when none left [OK]

Common Mistakes:

• Assuming pointers are all NULL
• Thinking pointers point to same block
• Forgetting to decrement free_blocks

4.

Identify the bug in this memory pool free function and select the correct fix.

void free_block(char* block) {
    if (free_blocks < MAX_BLOCKS) {
        pool[free_blocks++] = block;
    }
}

medium

A. No bug; function works correctly

B. Incrementing free_blocks before storing pointer causes overwrite

C. Should check if block is NULL before freeing

D. Store pointer in pool array is wrong; should mark block as free differently

Solution

1. Step 1: Analyze free_block logic

   The pool array holds actual memory blocks, not pointers. Storing a pointer in pool overwrites data.

2. Step 2: Correct way to free block

   Freeing should mark the block as available in a separate structure or use a free list, not overwrite pool data.

3. Final Answer:

   Store pointer in pool array is wrong; should mark block as free differently -> Option D

4. Quick Check:

   Freeing must not overwrite pool memory [OK]

Hint: Freeing must not overwrite pool data blocks [OK]

Common Mistakes:

• Overwriting pool memory with pointers
• Not checking block validity
• Assuming pool holds pointers, not blocks

5.

You want to implement a memory pool that supports 5 blocks of 64 bytes each. You also want to track which blocks are free using a bitmap (an integer where each bit represents a block's status: 0 = free, 1 = used). Which of the following code snippets correctly initializes the bitmap and allocates a block?

?

hard

A. int bitmap = 0xFF; // all used
char pool[5][64];
char* allocate() {
for (int i = 0; i < 5; i++) {
if ((bitmap & (1 << i)) == 0) {
bitmap &= ~(1 << i);
return pool[i];
}
}
return NULL;
}

B. int bitmap = 0; // all free
char pool[5][64];
char* allocate() {
for (int i = 0; i < 5; i++) {
if ((bitmap & (1 << i)) == 0) {
bitmap |= (1 << i);
return pool[i];
}
}
return NULL;
}

C. int bitmap = 0; // all used
char pool[5][64];
char* allocate() {
for (int i = 0; i < 5; i++) {
if ((bitmap & (1 << i)) != 0) {
bitmap |= (1 << i);
return pool[i];
}
}
return NULL;
}

D. int bitmap = 0xFF; // all free
char pool[5][64];
char* allocate() {
for (int i = 0; i < 5; i++) {
if ((bitmap & (1 << i)) != 0) {
bitmap &= ~(1 << i);
return pool[i];
}
}
return NULL;
}

Solution

1. Step 1: Understand bitmap meaning

   Bitmap bit 0 means free, 1 means used. So initializing bitmap=0 means all blocks free.

2. Step 2: Check allocation logic

   Allocation finds first free bit (0), sets it to 1 (used), and returns that block.

3. Final Answer:

   Bitmap initialized to 0; allocate sets bit to 1 on allocation -> Option B

4. Quick Check:

   0 = free, set bit to 1 when used [OK]

Hint: 0 bits free, set bit to 1 when allocated [OK]

Common Mistakes:

• Initializing bitmap with all bits set (all used)
• Checking wrong bit condition for free blocks
• Setting bits incorrectly on allocation