* kmalloc
Prototype:
#include <linux/slab.h>
void *kmalloc(size_t size, int flags);
Kmalloc分配一段未清0的連續物理內存頁,並返回虛存地址。有點是快,並且可指定flag,如DMA內存,高地址區域內存等。缺點是不能分配大於128KB(處於跨平台考慮),幾個重要的flag:
GFP_ATOMIC
Used to allocate memory from interrupt handlers and other code outside of a process context. Never sleeps.
GFP_KERNEL
Normal allocation of kernel memory. May sleep.
GFP_USER
Used to allocate memory for user-space pages; it may sleep.
GFP_HIGHUSER
Like GFP_USER, but allocates from high memory, if any. High memory is described in the next subsection.
* slab allocator(lookaside cache)
從Memcached的實現知道有這麼一個內存管理策略,其顯著特點是分配一組相同大小的內存塊作為內存池,其實現對應於源代碼中的<linux/slab.h>和mm/slab.c。
Prototype:
#include <linux/malloc.h>
kmem_cache_t *kmem_cache_create(char *name, size_t size, size_t offset,
unsigned long flags, constructor( ), destructor( ));
int kmem_cache_destroy(kmem_cache_t *cache);
/proc/slabinfo
A virtual file containing statistics on slab cache usage.
*__get_free_pages
Prototype:
_ _get_free_pages(unsigned int flags, unsigned int order);
返回2^order個未清0連續物理頁面,flags與kmalloc中flags一致,允許的最大order值為10或者11(根據體系結構不同)
*alloc_pages
Prototype:
struct page *alloc_pages_node(int nid, unsigned int flags,
unsigned int order);
Kernel中頁分配器實現,__get_free_pages即調用alloc_pages實現的
The real core of the Linux page allocator is a function called alloc_pages_node:
*vmalloc
分配地址連續虛存,而不保證物理地址連續,大部分情況下適合“軟件”,而不是驅動程序。相對而言,kmalloc和__get_free_pages虛存map到物理內存只需要增減一個偏移,而使用vmalloc分配需要修改頁表,故vmalloc的開銷較大,分配少數幾個頁面的效率太低。
*per-cpu variables
Each cpu hold an independant copy in their respective processor's caches, so there is no lock required and improve better performance, implemented as a linux 2.6 feature. Defined in <linux/percpu.h>.
DEFINE_PER_CPU(type, name);
get_cpu_var(sockets_in_use)++;
put_cpu_var(sockets_in_use);