Memory Slices

Memory Slices — efficient way to allocate groups of equal-sized chunks of memory

Synopsis

#include <glib.h>

gpointer            g_slice_alloc                       (gsize block_size);
gpointer            g_slice_alloc0                      (gsize block_size);
gpointer            g_slice_copy                        (gsize block_size,
                                                         gconstpointer mem_block);
void                g_slice_free1                       (gsize block_size,
                                                         gpointer mem_block);
void                g_slice_free_chain_with_offset      (gsize block_size,
                                                         gpointer mem_chain,
                                                         gsize next_offset);

#define             g_slice_new                         (type)
#define             g_slice_new0                        (type)
#define             g_slice_dup                         (type,
                                                         mem)
#define             g_slice_free                        (type,
                                                         mem)
#define             g_slice_free_chain                  (type,
                                                         mem_chain,
                                                         next)

Description

Memory slices provide a space-efficient and multi-processing scalable way to allocate equal-sized pieces of memory, just like the original GMemChunks (from GLib <= 2.8), while avoiding their excessive memory-waste, scalability and performance problems.

To achieve these goals, the slice allocator uses a sophisticated, layered design that has been inspired by Bonwick's slab allocator [6]. It uses posix_memalign() to optimize allocations of many equally-sized chunks, and has per-thread free lists (the so-called magazine layer) to quickly satisfy allocation requests of already known structure sizes. This is accompanied by extra caching logic to keep freed memory around for some time before returning it to the system. Memory that is unused due to alignment constraints is used for cache colorization (random distribution of chunk addresses) to improve CPU cache utilization. The caching layer of the slice allocator adapts itself to high lock contention to improve scalability.

The slice allocator can allocate blocks as small as two pointers, and unlike malloc(), it does not reserve extra space per block. For large block sizes, g_slice_new() and g_slice_alloc() will automatically delegate to the system malloc() implementation. For newly written code it is recommended to use the new g_slice API instead of g_malloc() and friends, as long as objects are not resized during their lifetime and the object size used at allocation time is still available when freeing.

Example 15. Using the slice allocator

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
gchar *mem[10000];
gint i;
/* Allocate 10000 blocks. */
for (i = 0; i < 10000; i++)
  {
    mem[i] = g_slice_alloc (50);
    /* Fill in the memory with some junk. */
    for (j = 0; j < 50; j++)
	mem[i][j] = i * j;
  }
/* Now free all of the blocks. */
for (i = 0; i < 10000; i++)
  {
    g_slice_free1 (50, mem[i]);
  }

Example 16. Using the slice allocator with data structures

1
2
3
4
5
6
7
8
9
10
11
12
GRealArray *array;
/* Allocate one block, using the g_slice_new() macro. */
array = g_slice_new (GRealArray);
/* We can now use array just like a normal pointer to a structure. */
array->data            = NULL;
array->len             = 0;
array->alloc           = 0;
array->zero_terminated = (zero_terminated ? 1 : 0);
array->clear           = (clear ? 1 : 0);
array->elt_size        = elt_size;
/* We can free the block, so it can be reused. */
g_slice_free (GRealArray, array);

Details

g_slice_alloc ()

gpointer            g_slice_alloc                       (gsize block_size);

Allocates a block of memory from the slice allocator. The block adress handed out can be expected to be aligned to at least 1 * sizeof (void*), though in general slices are 2 * sizeof (void*) bytes aligned, if a malloc() fallback implementation is used instead, the alignment may be reduced in a libc dependent fashion. Note that the underlying slice allocation mechanism can be changed with the G_SLICE=always-malloc environment variable.

block_size :

the number of bytes to allocate

Returns :

a pointer to the allocated memory block

Since 2.10


g_slice_alloc0 ()

gpointer            g_slice_alloc0                      (gsize block_size);

Allocates a block of memory via g_slice_alloc() and initialize the returned memory to 0. Note that the underlying slice allocation mechanism can be changed with the G_SLICE=always-malloc environment variable.

block_size :

the number of bytes to allocate

Returns :

a pointer to the allocated block

Since 2.10


g_slice_copy ()

gpointer            g_slice_copy                        (gsize block_size,
                                                         gconstpointer mem_block);

Allocates a block of memory from the slice allocator and copies block_size bytes into it from mem_block.

block_size :

the number of bytes to allocate

mem_block :

the memory to copy

Returns :

a pointer to the allocated memory block

Since 2.14


g_slice_free1 ()

void                g_slice_free1                       (gsize block_size,
                                                         gpointer mem_block);

Frees a block of memory. The memory must have been allocated via g_slice_alloc() or g_slice_alloc0() and the block_size has to match the size specified upon allocation. Note that the exact release behaviour can be changed with the G_DEBUG=gc-friendly environment variable, also see G_SLICE for related debugging options.

block_size :

the size of the block

mem_block :

a pointer to the block to free

Since 2.10


g_slice_free_chain_with_offset ()

void                g_slice_free_chain_with_offset      (gsize block_size,
                                                         gpointer mem_chain,
                                                         gsize next_offset);

Frees a linked list of memory blocks of structure type type. The memory blocks must be equal-sized, allocated via g_slice_alloc() or g_slice_alloc0() and linked together by a next pointer (similar to GSList). The offset of the next field in each block is passed as third argument. Note that the exact release behaviour can be changed with the G_DEBUG=gc-friendly environment variable, also see G_SLICE for related debugging options.

block_size :

the size of the blocks

mem_chain :

a pointer to the first block of the chain

next_offset :

the offset of the next field in the blocks

Since 2.10


g_slice_new()

#define             g_slice_new(type)

A convenience macro to allocate a block of memory from the slice allocator. It calls g_slice_alloc() with sizeof (type) and casts the returned pointer to a pointer of the given type, avoiding a type cast in the source code. Note that the underlying slice allocation mechanism can be changed with the G_SLICE=always-malloc environment variable.

type :

the type to allocate, typically a structure name

Returns :

a pointer to the allocated block, cast to a pointer to type.

Since 2.10


g_slice_new0()

#define             g_slice_new0(type)

A convenience macro to allocate a block of memory from the slice allocator and set the memory to 0. It calls g_slice_alloc0() with sizeof (type) and casts the returned pointer to a pointer of the given type, avoiding a type cast in the source code. Note that the underlying slice allocation mechanism can be changed with the G_SLICE=always-malloc environment variable.

type :

the type to allocate, typically a structure name

Returns :

a pointer to the allocated block, cast to a pointer to type.

Since 2.10


g_slice_dup()

#define             g_slice_dup(type, mem)

A convenience macro to duplicate a block of memory using the slice allocator. It calls g_slice_copy() with sizeof (type) and casts the returned pointer to a pointer of the given type, avoiding a type cast in the source code. Note that the underlying slice allocation mechanism can be changed with the G_SLICE=always-malloc environment variable.

type :

the type to duplicate, typically a structure name

mem :

the memory to copy into the allocated block

Returns :

a pointer to the allocated block, cast to a pointer to type.

Since 2.14


g_slice_free()

#define             g_slice_free(type, mem)

A convenience macro to free a block of memory that has been allocated from the slice allocator. It calls g_slice_free1() using sizeof (type) as the block size. Note that the exact release behaviour can be changed with the G_DEBUG=gc-friendly environment variable, also see G_SLICE for related debugging options.

type :

the type of the block to free, typically a structure name

mem :

a pointer to the block to free

Since 2.10


g_slice_free_chain()

#define             g_slice_free_chain(type, mem_chain, next)

Frees a linked list of memory blocks of structure type type. The memory blocks must be equal-sized, allocated via g_slice_alloc() or g_slice_alloc0() and linked together by a next pointer (similar to GSList). The name of the next field in type is passed as third argument. Note that the exact release behaviour can be changed with the G_DEBUG=gc-friendly environment variable, also see G_SLICE for related debugging options.

type :

the type of the mem_chain blocks

mem_chain :

a pointer to the first block of the chain

next :

the field name of the next pointer in type

Since 2.10



[6] [Bonwick94] Jeff Bonwick, The slab allocator: An object-caching kernel memory allocator. USENIX 1994, and [Bonwick01] Bonwick and Jonathan Adams, Magazines and vmem: Extending the slab allocator to many cpu's and arbitrary resources. USENIX 2001