Linux Kernel Architecture

Chapter3:MemoryManagement

The relationship betweenfree_pagesand__free_pagesis established by means of a function instead of

a macro because the virtual address must first be converted to a pointer tostruct page.

mm/page_alloc.c

void free_pages(unsigned long addr, unsigned int order)

{

if (addr != 0) {

__free_pages(virt_to_page(addr), order);

}

}

virt_to_pageconverts virtual memory addresses to pointers topageinstances. Basically, this is the

reverse of thepage_addresshelper function introduced above.

Figure 3-31 summarizes the relationships among the various memory-freeing functions in a graphical

overview.

free_page

free_pages

__free_pages

__free_page

Figure 3-31: Relationships

among the memory-freeing

functions of the buddy

system.

3.5.5 Reserving Pages

All API functions lead back toalloc_pages_node, which is a kind of ‘‘launch pad‘‘ for central implemen-

tation of the buddy system.

<gfp.h>

static inline struct page *alloc_pages_node(int nid, gfp_t gfp_mask,

unsigned int order)

{

if (unlikely(order >= MAX_ORDER))

return NULL;

/* Unknown node is current node */

if (nid < 0)

nid = numa_node_id();

return __alloc_pages(gfp_mask, order,

NODE_DATA(nid)->node_zonelists + gfp_zone(gfp_mask));

}

Just a simple check is carried out to ensure that no overly large memory chunk is allocated. If a neg-

ative node ID (which does not exist) is specified, the kernel automatically uses the ID that belongs to