bootmem allocator分析
2013-08-13 16:37
204 查看
Codebase: android 4.1
Kernel: 3.4.0
Chipset: msm8x25q
在系统启动时，内存的伙伴系统/slab算法还没有初始化之前，系统也需要来作内存管理，分配一些核心数据结构，bootmem分配器就实现了该功能，它用于在启动阶段早期分配内存。
Bootmem分配器使用位图来管理页，位图数量和系统的物理内存也数量是相同的。当页被使用时，就标记为1，否则为0表示空闲页。
由于该分配器管理机制比较简单，并没有考虑性能和通用性，所以在伙伴系统完成初始化之后，bootmem分配器就要交出管理权，然后销毁掉。
在UMA系统上，只有一个bootmemallocator，名字叫bootmem_node_data, 位于kernel/mm/bootmem.c中，它通过成为struct pglist_data的一个元素与变量contig_page_data联系起来。
[html]
view plaincopyprint?
struct pglist_data __refdata contig_page_data = {
.bdata = &bootmem_node_data[0]
};
bootmem_data_t bootmem_node_data[MAX_NUMNODES] __initdata;
start_kernel -> setup_arch -> paging_init-> bootmem_init
[html]
view plaincopyprint?
void __init bootmem_init(void)
{
unsigned long min, max_low, max_high;
max_low = max_high = 0;
/*min: 物理内存起始地址pfn号
max_low: 低端内存结束地址pfn号
max_high: 高端内存结束地址pfn号
*/
find_limits(&min, &max_low, &max_high);
/*根据参数看是初始化Lowmem区域?*/
arm_bootmem_init(min, max_low);
/*
* Sparsemem tries to allocate bootmem in memory_present(),
* so must be done after the fixed reservations
*/
arm_memory_present();
/*
* sparse_init() needs the bootmem allocator up and running.
*/
sparse_init();
/*
* Now free the memory - free_area_init_node needs
* the sparse mem_map arrays initialized by sparse_init()
* for memmap_init_zone(), otherwise all PFNs are invalid.
*/
arm_bootmem_free(min, max_low, max_high);
/*保存lowmem和highmem对应的pfn numbers，
这并表示实际能操作的pfn number,因为start pfn不一定从0开始。*/
max_low_pfn =
max_low - PHYS_PFN_OFFSET;
max_pfn = max_high - PHYS_PFN_OFFSET;
}
[html]
view plaincopyprint?
static void __init find_limits(unsigned long *min, unsigned long *max_low,
unsigned long *max_high)
{
struct meminfo *mi = &meminfo;
int i;
/*循环直到是highmem的bank才停止。*/
/* This assumes the meminfo array is properly sorted */
*min = bank_pfn_start(&mi->bank[0]);
for_each_bank (i, mi)
if (mi->bank[i].highmem)
break;
/*获得lowmem和highmem结束地址的pfn.*/
*max_low =
bank_pfn_end(&mi->bank[i - 1]);
*max_high = bank_pfn_end(&mi->bank[mi->nr_banks - 1]);
}
函数比较简单，不过要注意的是，这里获得的内存是内核当前拥有的memory，当然也包含了已经被reserved的区域。arm_bootmem_init()会重新划分。
[html]
view plaincopyprint?
static void __init arm_bootmem_init(unsigned long start_pfn,
unsigned long end_pfn)
{
struct memblock_region *reg;
unsigned int boot_pages;
phys_addr_t bitmap;
pg_data_t *pgdat;
/*
* Allocate the bootmem bitmap page. This must be in a region
* of memory which has already been mapped.
*/
/*end_pfn – start_pfn为lowmem的pfn numbers。*/
boot_pages = bootmem_bootmap_pages(end_pfn - start_pfn);
/*根据pfn numbers来分配bitmap位图, 以L1 cache能操作的字节数作对齐，
是为了让L1 cache能操作? 最大能分配地址为end_pfn。*/
bitmap = memblock_alloc_base(boot_pages
<< PAGE_SHIFT, L1_CACHE_BYTES,
__pfn_to_phys(end_pfn));
/*
* Initialise the bootmem allocator, handing the
* memory banks over to bootmem.
*/
node_set_online(0);
pgdat = NODE_DATA(0);
/*初始化pgda也就是全局变量contig_page_data 中的bdata元素，也就是
bootmem_node_data 变量。*/
init_bootmem_node(pgdat, __phys_to_pfn(bitmap), start_pfn, end_pfn);
/*在前面meminfo介绍中有说到，struct memblock中的memroy 元素表示
空闲内存区域，而reseved表示要保留的区域。所以这里会将reserved对应的
Bitmap标为1，而free memory标志为0.*/
/* Free the lowmem regions from memblock into bootmem. */
for_each_memblock(memory, reg) {
unsigned long start =
memblock_region_memory_base_pfn(reg);
unsigned long end =
memblock_region_memory_end_pfn(reg);
if (end >= end_pfn)
end = end_pfn;
if (start >= end)
break;
free_bootmem(__pfn_to_phys(start), (end - start)
<< PAGE_SHIFT);
}
/* Reserve the lowmem memblock reserved regions in bootmem. */
for_each_memblock(reserved, reg) {
unsigned long start =
memblock_region_reserved_base_pfn(reg);
unsigned long end =
memblock_region_reserved_end_pfn(reg);
if (end >= end_pfn)
end =
end_pfn;
if (start >= end)
break;
reserve_bootmem(__pfn_to_phys(start),
(end - start) <<
PAGE_SHIFT, BOOTMEM_DEFAULT);
}
}
[html]
view plaincopyprint?
unsigned long __init bootmem_bootmap_pages(unsigned long pages)
{
unsigned long bytes =
bootmap_bytes(pages);
/*以4k作为一个单位分配*/
return PAGE_ALIGN(bytes) >> PAGE_SHIFT;
}
static unsigned long __init bootmap_bytes(unsigned long pages)
{
/*每个page作为一个bit保留在unsigned long变量中。 */
unsigned long bytes = DIV_ROUND_UP(pages, 8);
return ALIGN(bytes, sizeof(long));
}
view plaincopyprint?
phys_addr_t __init memblock_alloc_base(phys_addr_t size, phys_addr_t align, phys_addr_t max_addr)
{
phys_addr_t alloc;
alloc = __memblock_alloc_base(size, align, max_addr);
if (alloc == 0)
panic("ERROR: Failed to allocate 0x%llx bytes below 0x%llx.\n",
(unsigned long long) size, (unsigned long long) max_addr);
return alloc;
}
phys_addr_t __init __memblock_alloc_base(phys_addr_t size, phys_addr_t align, phys_addr_t max_addr)
{
return memblock_alloc_base_nid(size, align, max_addr, MAX_NUMNODES);
}
static phys_addr_t __init memblock_alloc_base_nid(phys_addr_t size,
phys_addr_t align, phys_addr_t max_addr,
int nid)
{
phys_addr_t found;
size = round_up(size, align);
found = memblock_find_in_range_node(0, max_addr, size, align, nid);
if (found && !memblock_reserve(found, size))
return found;
return 0;
}
phys_addr_t __init_memblock memblock_find_in_range_node(phys_addr_t start,
phys_addr_t end, phys_addr_t size,
phys_addr_t align, int nid)
{
phys_addr_t this_start, this_end, cand;
u64 i;
/* avoid allocating the first page */
/*保留第一页，用来干嘛？*/
start = max_t(phys_addr_t, start, PAGE_SIZE);
end = max(start, end);
/*从struct memblock的一块空闲区域的最高地址往下分配一块区域。*/
for_each_free_mem_range_reverse(i, nid, &this_start, &this_end, NULL) {
/*this_start 和this_end 在start 和end中间的话直接返回，否则
返回end。*/
this_start = clamp(this_start, start, end);
this_end =
clamp(this_end, start, end);
if (this_end < size)
continue;
/*得到分配内存地址，大小为size。*/
cand =
round_down(this_end - size, align);
if (cand >= this_start)
return cand;
}
return 0;
}
view plaincopyprint?
unsigned long __init init_bootmem_node(pg_data_t *pgdat, unsigned long freepfn,
unsigned long startpfn, unsigned long endpfn)
{
return init_bootmem_core(pgdat->bdata, freepfn, startpfn, endpfn);
}
static unsigned long __init init_bootmem_core(bootmem_data_t *bdata,
unsigned long mapstart, unsigned long start, unsigned long end)
{
unsigned long mapsize;
mminit_validate_memmodel_limits(&start, &end);
/*得到bitmap表的首地址以及最小和最大pfn*/
bdata->node_bootmem_map =
phys_to_virt(PFN_PHYS(mapstart));
bdata->node_min_pfn =
start;
bdata->node_low_pfn =
end;
/*加入到全局的bdata_list链表变量中，方便管理。*/
link_bootmem(bdata);
/*
* Initially all pages are reserved - setup_arch() has to
* register free RAM areas explicitly.
*/
/*将bitmap表中每个bit都设置成已经使用了。下一步
for_each_memblock()会重新设置。*/
mapsize = bootmap_bytes(end - start);
memset(bdata->node_bootmem_map, 0xff, mapsize);
bdebug("nid=%td start=%lx
map=%lx end=%lx
mapsize=%lx\n",
bdata - bootmem_node_data, start, mapstart, end, mapsize);
return mapsize;
}
到此，bootmem allocator已经初始化完成。
路径： kernel/kernel/include/linux/bootmem.h
[html]
view plaincopyprint?
/*按指定size从ZONE_NORMAL区域分配*/
#define alloc_bootmem(x) \
__alloc_bootmem(x, SMP_CACHE_BYTES, BOOTMEM_LOW_LIMIT)
/*按指定size从ZONE_NORMAL区域分配, 以align对齐*/
#define alloc_bootmem_align(x, align) \
__alloc_bootmem(x, align, BOOTMEM_LOW_LIMIT)
/*按指定size从ZONE_NORMAL区域分配, 以一页对齐*/
#define alloc_bootmem_pages(x) \
__alloc_bootmem(x, PAGE_SIZE, BOOTMEM_LOW_LIMIT)
/* SMP_CACHE_BYTES 是为了让数据能更好地在L1 cache中使用，虽然是SMP开头。*/
#define alloc_bootmem_nopanic(x) \
__alloc_bootmem_nopanic(x, SMP_CACHE_BYTES, BOOTMEM_LOW_LIMIT)
#define alloc_bootmem_node(pgdat, x) \
__alloc_bootmem_node(pgdat, x, SMP_CACHE_BYTES, BOOTMEM_LOW_LIMIT)
/*以_node后缀结尾表示只在NUMA系统上使用。*/
#define alloc_bootmem_node(pgdat, x) \
__alloc_bootmem_node(pgdat, x, SMP_CACHE_BYTES, BOOTMEM_LOW_LIMIT)
#define alloc_bootmem_node_nopanic(pgdat, x) \
__alloc_bootmem_node_nopanic(pgdat, x, SMP_CACHE_BYTES, BOOTMEM_LOW_LIMIT)
#define alloc_bootmem_pages_node(pgdat, x) \
__alloc_bootmem_node(pgdat, x, PAGE_SIZE, BOOTMEM_LOW_LIMIT)
#define alloc_bootmem_pages_node_nopanic(pgdat, x) \
__alloc_bootmem_node_nopanic(pgdat, x, PAGE_SIZE, BOOTMEM_LOW_LIMIT)
/*这几个和上面的区别是从ZONE_DMA区域分配。*/
#define alloc_bootmem_low(x) \
__alloc_bootmem_low(x, SMP_CACHE_BYTES, 0)
#define alloc_bootmem_low_pages(x) \
__alloc_bootmem_low(x, PAGE_SIZE, 0)
#define alloc_bootmem_low_pages_node(pgdat, x) \
__alloc_bootmem_low_node(pgdat, x, PAGE_SIZE, 0)
关于内存的分配，请允许我偷懒下，不对代码做详细分析了，有点费时间，流程如下：
__alloc_bootmem -> ___alloc_bootmem -> ___alloc_bootmem_nopanic -> alloc_bootmem_core -> find_next_zero_bit
大概的步骤就是：
1. 扫描bitmap位图，寻找空闲的位
2. 如果查找的页紧挨着上一次分配的页，就先检查这次要分配的内存是否能在上一页直接分配，因为bootmem allocator支持小于一页的分配。
3. 在新分配的页的bitmap对应的bit设置为1后，将当前的偏移保存，如果页没有完全分配，那么页里面的偏移量也保存。
这个接口没有分析，据资料记载说分配页可能会有风险。
start_kernel -> mm_init -> mem_init -> free_all_bootmem -> free_all_bootmem_core -> __free_pages_bootmem -> __free_pages
可以看到最终调用的是__free_pages()，这个函数会将这些Pages释放到伙伴系统中管理。
注意这里只是将空闲的页释放掉了。占据的页还存在。
由于bootmem分配的页里面的数据基本上都是用于内存基本结构，在系统运行期间会一直被用到，所以不会被释放。不过像__init这种类型的数据段只在系统开机的时候被使用，所以系统初始化完成之后就可以释放掉了。
系统使用的函数是free_initmem()， 调用流程如下：
start_kernel -> rest_init -> kernel_init -> init_post -> free_initmem
[html]
view plaincopyprint?
void free_initmem(void)
{
unsigned long reclaimed_initmem;
poison_init_mem(__init_begin, __init_end - __init_begin);
if (!machine_is_integrator() && !machine_is_cintegrator()) {
reclaimed_initmem =
free_area(__phys_to_pfn(__pa(__init_begin)),
__phys_to_pfn(__pa(__init_end)),
"init");
totalram_pages += reclaimed_initmem;
}
}
可以看到，释放的数据就是保存在__init_begin和__init_end那一段之间！当然，最后也是调用__free_pages()释放到伙伴系统中管理的。
使用bootmem分配内存的时候需要注意下一定要处于bootmem 分配器初始化和销毁之间！
2013/03/21
Kernel: 3.4.0
Chipset: msm8x25q
在系统启动时，内存的伙伴系统/slab算法还没有初始化之前，系统也需要来作内存管理，分配一些核心数据结构，bootmem分配器就实现了该功能，它用于在启动阶段早期分配内存。
Bootmem分配器使用位图来管理页，位图数量和系统的物理内存也数量是相同的。当页被使用时，就标记为1，否则为0表示空闲页。
由于该分配器管理机制比较简单，并没有考虑性能和通用性，所以在伙伴系统完成初始化之后，bootmem分配器就要交出管理权，然后销毁掉。
在UMA系统上，只有一个bootmemallocator，名字叫bootmem_node_data, 位于kernel/mm/bootmem.c中，它通过成为struct pglist_data的一个元素与变量contig_page_data联系起来。
[html]
view plaincopyprint?
struct pglist_data __refdata contig_page_data = {
.bdata = &bootmem_node_data[0]
};
bootmem_data_t bootmem_node_data[MAX_NUMNODES] __initdata;
struct pglist_data __refdata contig_page_data = {
.bdata = &bootmem_node_data[0]
};
bootmem_data_t bootmem_node_data[MAX_NUMNODES] __initdata;Bootmem 初始化：
系统启动的时候有如下流程：start_kernel -> setup_arch -> paging_init-> bootmem_init
[html]
view plaincopyprint?
void __init bootmem_init(void)
{
unsigned long min, max_low, max_high;
max_low = max_high = 0;
/*min: 物理内存起始地址pfn号
max_low: 低端内存结束地址pfn号
max_high: 高端内存结束地址pfn号
*/
find_limits(&min, &max_low, &max_high);
/*根据参数看是初始化Lowmem区域?*/
arm_bootmem_init(min, max_low);
/*
* Sparsemem tries to allocate bootmem in memory_present(),
* so must be done after the fixed reservations
*/
arm_memory_present();
/*
* sparse_init() needs the bootmem allocator up and running.
*/
sparse_init();
/*
* Now free the memory - free_area_init_node needs
* the sparse mem_map arrays initialized by sparse_init()
* for memmap_init_zone(), otherwise all PFNs are invalid.
*/
arm_bootmem_free(min, max_low, max_high);
/*保存lowmem和highmem对应的pfn numbers，
这并表示实际能操作的pfn number,因为start pfn不一定从0开始。*/
max_low_pfn =
max_low - PHYS_PFN_OFFSET;
max_pfn = max_high - PHYS_PFN_OFFSET;
}
void __init bootmem_init(void)
{
unsigned long min, max_low, max_high;
max_low = max_high = 0;
/*min: 物理内存起始地址pfn号
max_low: 低端内存结束地址pfn号
max_high: 高端内存结束地址pfn号
*/
find_limits(&min, &max_low, &max_high);
/*根据参数看是初始化Lowmem区域?*/
arm_bootmem_init(min, max_low);
/*
* Sparsemem tries to allocate bootmem in memory_present(),
* so must be done after the fixed reservations
*/
arm_memory_present();
/*
* sparse_init() needs the bootmem allocator up and running.
*/
sparse_init();
/*
* Now free the memory - free_area_init_node needs
* the sparse mem_map arrays initialized by sparse_init()
* for memmap_init_zone(), otherwise all PFNs are invalid.
*/
arm_bootmem_free(min, max_low, max_high);
/*保存lowmem和highmem对应的pfn numbers，
这并表示实际能操作的pfn number,因为start pfn不一定从0开始。*/
max_low_pfn = max_low - PHYS_PFN_OFFSET;
max_pfn = max_high - PHYS_PFN_OFFSET;
}find_limits():
里面有些函数需要展开来分析下，先看find_limits：[html]
view plaincopyprint?
static void __init find_limits(unsigned long *min, unsigned long *max_low,
unsigned long *max_high)
{
struct meminfo *mi = &meminfo;
int i;
/*循环直到是highmem的bank才停止。*/
/* This assumes the meminfo array is properly sorted */
*min = bank_pfn_start(&mi->bank[0]);
for_each_bank (i, mi)
if (mi->bank[i].highmem)
break;
/*获得lowmem和highmem结束地址的pfn.*/
*max_low =
bank_pfn_end(&mi->bank[i - 1]);
*max_high = bank_pfn_end(&mi->bank[mi->nr_banks - 1]);
}
static void __init find_limits(unsigned long *min, unsigned long *max_low,
unsigned long *max_high)
{
struct meminfo *mi = &meminfo;
int i;
/*循环直到是highmem的bank才停止。*/
/* This assumes the meminfo array is properly sorted */
*min = bank_pfn_start(&mi->bank[0]);
for_each_bank (i, mi)
if (mi->bank[i].highmem)
break;
/*获得lowmem和highmem结束地址的pfn.*/
*max_low = bank_pfn_end(&mi->bank[i - 1]);
*max_high = bank_pfn_end(&mi->bank[mi->nr_banks - 1]);
}函数比较简单，不过要注意的是，这里获得的内存是内核当前拥有的memory，当然也包含了已经被reserved的区域。arm_bootmem_init()会重新划分。
arm_bootmem_init():
arm_bootmem_init()是核心的函数。[html]
view plaincopyprint?
static void __init arm_bootmem_init(unsigned long start_pfn,
unsigned long end_pfn)
{
struct memblock_region *reg;
unsigned int boot_pages;
phys_addr_t bitmap;
pg_data_t *pgdat;
/*
* Allocate the bootmem bitmap page. This must be in a region
* of memory which has already been mapped.
*/
/*end_pfn – start_pfn为lowmem的pfn numbers。*/
boot_pages = bootmem_bootmap_pages(end_pfn - start_pfn);
/*根据pfn numbers来分配bitmap位图, 以L1 cache能操作的字节数作对齐，
是为了让L1 cache能操作? 最大能分配地址为end_pfn。*/
bitmap = memblock_alloc_base(boot_pages
<< PAGE_SHIFT, L1_CACHE_BYTES,
__pfn_to_phys(end_pfn));
/*
* Initialise the bootmem allocator, handing the
* memory banks over to bootmem.
*/
node_set_online(0);
pgdat = NODE_DATA(0);
/*初始化pgda也就是全局变量contig_page_data 中的bdata元素，也就是
bootmem_node_data 变量。*/
init_bootmem_node(pgdat, __phys_to_pfn(bitmap), start_pfn, end_pfn);
/*在前面meminfo介绍中有说到，struct memblock中的memroy 元素表示
空闲内存区域，而reseved表示要保留的区域。所以这里会将reserved对应的
Bitmap标为1，而free memory标志为0.*/
/* Free the lowmem regions from memblock into bootmem. */
for_each_memblock(memory, reg) {
unsigned long start =
memblock_region_memory_base_pfn(reg);
unsigned long end =
memblock_region_memory_end_pfn(reg);
if (end >= end_pfn)
end = end_pfn;
if (start >= end)
break;
free_bootmem(__pfn_to_phys(start), (end - start)
<< PAGE_SHIFT);
}
/* Reserve the lowmem memblock reserved regions in bootmem. */
for_each_memblock(reserved, reg) {
unsigned long start =
memblock_region_reserved_base_pfn(reg);
unsigned long end =
memblock_region_reserved_end_pfn(reg);
if (end >= end_pfn)
end =
end_pfn;
if (start >= end)
break;
reserve_bootmem(__pfn_to_phys(start),
(end - start) <<
PAGE_SHIFT, BOOTMEM_DEFAULT);
}
}
static void __init arm_bootmem_init(unsigned long start_pfn,
unsigned long end_pfn)
{
struct memblock_region *reg;
unsigned int boot_pages;
phys_addr_t bitmap;
pg_data_t *pgdat;
/*
* Allocate the bootmem bitmap page. This must be in a region
* of memory which has already been mapped.
*/
/*end_pfn – start_pfn为lowmem的pfn numbers。*/
boot_pages = bootmem_bootmap_pages(end_pfn - start_pfn);
/*根据pfn numbers来分配bitmap位图, 以L1 cache能操作的字节数作对齐，
是为了让L1 cache能操作? 最大能分配地址为end_pfn。*/
bitmap = memblock_alloc_base(boot_pages << PAGE_SHIFT, L1_CACHE_BYTES,
__pfn_to_phys(end_pfn));
/*
* Initialise the bootmem allocator, handing the
* memory banks over to bootmem.
*/
node_set_online(0);
pgdat = NODE_DATA(0);
/*初始化pgda也就是全局变量contig_page_data 中的bdata元素，也就是
bootmem_node_data 变量。*/
init_bootmem_node(pgdat, __phys_to_pfn(bitmap), start_pfn, end_pfn);
/*在前面meminfo介绍中有说到，struct memblock中的memroy 元素表示
空闲内存区域，而reseved表示要保留的区域。所以这里会将reserved对应的
Bitmap标为1，而free memory标志为0.*/
/* Free the lowmem regions from memblock into bootmem. */
for_each_memblock(memory, reg) {
unsigned long start = memblock_region_memory_base_pfn(reg);
unsigned long end = memblock_region_memory_end_pfn(reg);
if (end >= end_pfn)
end = end_pfn;
if (start >= end)
break;
free_bootmem(__pfn_to_phys(start), (end - start) << PAGE_SHIFT);
}
/* Reserve the lowmem memblock reserved regions in bootmem. */
for_each_memblock(reserved, reg) {
unsigned long start = memblock_region_reserved_base_pfn(reg);
unsigned long end = memblock_region_reserved_end_pfn(reg);
if (end >= end_pfn)
end = end_pfn;
if (start >= end)
break;
reserve_bootmem(__pfn_to_phys(start),
(end - start) << PAGE_SHIFT, BOOTMEM_DEFAULT);
}
}bootmem_bootmap_pages():
继续分解函数，先看bootmem_bootmap_pages().[html]
view plaincopyprint?
unsigned long __init bootmem_bootmap_pages(unsigned long pages)
{
unsigned long bytes =
bootmap_bytes(pages);
/*以4k作为一个单位分配*/
return PAGE_ALIGN(bytes) >> PAGE_SHIFT;
}
static unsigned long __init bootmap_bytes(unsigned long pages)
{
/*每个page作为一个bit保留在unsigned long变量中。 */
unsigned long bytes = DIV_ROUND_UP(pages, 8);
return ALIGN(bytes, sizeof(long));
}
unsigned long __init bootmem_bootmap_pages(unsigned long pages)
{
unsigned long bytes = bootmap_bytes(pages);
/*以4k作为一个单位分配*/
return PAGE_ALIGN(bytes) >> PAGE_SHIFT;
}
static unsigned long __init bootmap_bytes(unsigned long pages)
{
/*每个page作为一个bit保留在unsigned long变量中。 */
unsigned long bytes = DIV_ROUND_UP(pages, 8);
return ALIGN(bytes, sizeof(long));
}memblock_alloc_base():
[html]view plaincopyprint?
phys_addr_t __init memblock_alloc_base(phys_addr_t size, phys_addr_t align, phys_addr_t max_addr)
{
phys_addr_t alloc;
alloc = __memblock_alloc_base(size, align, max_addr);
if (alloc == 0)
panic("ERROR: Failed to allocate 0x%llx bytes below 0x%llx.\n",
(unsigned long long) size, (unsigned long long) max_addr);
return alloc;
}
phys_addr_t __init __memblock_alloc_base(phys_addr_t size, phys_addr_t align, phys_addr_t max_addr)
{
return memblock_alloc_base_nid(size, align, max_addr, MAX_NUMNODES);
}
static phys_addr_t __init memblock_alloc_base_nid(phys_addr_t size,
phys_addr_t align, phys_addr_t max_addr,
int nid)
{
phys_addr_t found;
size = round_up(size, align);
found = memblock_find_in_range_node(0, max_addr, size, align, nid);
if (found && !memblock_reserve(found, size))
return found;
return 0;
}
phys_addr_t __init_memblock memblock_find_in_range_node(phys_addr_t start,
phys_addr_t end, phys_addr_t size,
phys_addr_t align, int nid)
{
phys_addr_t this_start, this_end, cand;
u64 i;
/* avoid allocating the first page */
/*保留第一页，用来干嘛？*/
start = max_t(phys_addr_t, start, PAGE_SIZE);
end = max(start, end);
/*从struct memblock的一块空闲区域的最高地址往下分配一块区域。*/
for_each_free_mem_range_reverse(i, nid, &this_start, &this_end, NULL) {
/*this_start 和this_end 在start 和end中间的话直接返回，否则
返回end。*/
this_start = clamp(this_start, start, end);
this_end =
clamp(this_end, start, end);
if (this_end < size)
continue;
/*得到分配内存地址，大小为size。*/
cand =
round_down(this_end - size, align);
if (cand >= this_start)
return cand;
}
return 0;
}
phys_addr_t __init memblock_alloc_base(phys_addr_t size, phys_addr_t align, phys_addr_t max_addr)
{
phys_addr_t alloc;
alloc = __memblock_alloc_base(size, align, max_addr);
if (alloc == 0)
panic("ERROR: Failed to allocate 0x%llx bytes below 0x%llx.\n",
(unsigned long long) size, (unsigned long long) max_addr);
return alloc;
}
phys_addr_t __init __memblock_alloc_base(phys_addr_t size, phys_addr_t align, phys_addr_t max_addr)
{
return memblock_alloc_base_nid(size, align, max_addr, MAX_NUMNODES);
}
static phys_addr_t __init memblock_alloc_base_nid(phys_addr_t size,
phys_addr_t align, phys_addr_t max_addr,
int nid)
{
phys_addr_t found;
size = round_up(size, align);
found = memblock_find_in_range_node(0, max_addr, size, align, nid);
if (found && !memblock_reserve(found, size))
return found;
return 0;
}
phys_addr_t __init_memblock memblock_find_in_range_node(phys_addr_t start,
phys_addr_t end, phys_addr_t size,
phys_addr_t align, int nid)
{
phys_addr_t this_start, this_end, cand;
u64 i;
/* avoid allocating the first page */
/*保留第一页，用来干嘛？*/
start = max_t(phys_addr_t, start, PAGE_SIZE);
end = max(start, end);
/*从struct memblock的一块空闲区域的最高地址往下分配一块区域。*/
for_each_free_mem_range_reverse(i, nid, &this_start, &this_end, NULL) {
/*this_start 和this_end 在start 和end中间的话直接返回，否则
返回end。*/
this_start = clamp(this_start, start, end);
this_end = clamp(this_end, start, end);
if (this_end < size)
continue;
/*得到分配内存地址，大小为size。*/
cand = round_down(this_end - size, align);
if (cand >= this_start)
return cand;
}
return 0;
}init_bootmem_node():
[html]view plaincopyprint?
unsigned long __init init_bootmem_node(pg_data_t *pgdat, unsigned long freepfn,
unsigned long startpfn, unsigned long endpfn)
{
return init_bootmem_core(pgdat->bdata, freepfn, startpfn, endpfn);
}
static unsigned long __init init_bootmem_core(bootmem_data_t *bdata,
unsigned long mapstart, unsigned long start, unsigned long end)
{
unsigned long mapsize;
mminit_validate_memmodel_limits(&start, &end);
/*得到bitmap表的首地址以及最小和最大pfn*/
bdata->node_bootmem_map =
phys_to_virt(PFN_PHYS(mapstart));
bdata->node_min_pfn =
start;
bdata->node_low_pfn =
end;
/*加入到全局的bdata_list链表变量中，方便管理。*/
link_bootmem(bdata);
/*
* Initially all pages are reserved - setup_arch() has to
* register free RAM areas explicitly.
*/
/*将bitmap表中每个bit都设置成已经使用了。下一步
for_each_memblock()会重新设置。*/
mapsize = bootmap_bytes(end - start);
memset(bdata->node_bootmem_map, 0xff, mapsize);
bdebug("nid=%td start=%lx
map=%lx end=%lx
mapsize=%lx\n",
bdata - bootmem_node_data, start, mapstart, end, mapsize);
return mapsize;
}
unsigned long __init init_bootmem_node(pg_data_t *pgdat, unsigned long freepfn,
unsigned long startpfn, unsigned long endpfn)
{
return init_bootmem_core(pgdat->bdata, freepfn, startpfn, endpfn);
}
static unsigned long __init init_bootmem_core(bootmem_data_t *bdata,
unsigned long mapstart, unsigned long start, unsigned long end)
{
unsigned long mapsize;
mminit_validate_memmodel_limits(&start, &end);
/*得到bitmap表的首地址以及最小和最大pfn*/
bdata->node_bootmem_map = phys_to_virt(PFN_PHYS(mapstart));
bdata->node_min_pfn = start;
bdata->node_low_pfn = end;
/*加入到全局的bdata_list链表变量中，方便管理。*/
link_bootmem(bdata);
/*
* Initially all pages are reserved - setup_arch() has to
* register free RAM areas explicitly.
*/
/*将bitmap表中每个bit都设置成已经使用了。下一步
for_each_memblock()会重新设置。*/
mapsize = bootmap_bytes(end - start);
memset(bdata->node_bootmem_map, 0xff, mapsize);
bdebug("nid=%td start=%lx map=%lx end=%lx mapsize=%lx\n",
bdata - bootmem_node_data, start, mapstart, end, mapsize);
return mapsize;
}free_bootmem()/reserve_bootmem():
这两个函数比较简单了，表示将bootmem的页分别标记成空闲和使用中。到此，bootmem allocator已经初始化完成。
Bootmem内存分配：
Bootmem的分配有多种接口，不过最终调用的都是__alloc_bootmem()，而__alloc_bootmem()调用了___alloc_bootmem_nopanic()。路径： kernel/kernel/include/linux/bootmem.h
[html]
view plaincopyprint?
/*按指定size从ZONE_NORMAL区域分配*/
#define alloc_bootmem(x) \
__alloc_bootmem(x, SMP_CACHE_BYTES, BOOTMEM_LOW_LIMIT)
/*按指定size从ZONE_NORMAL区域分配, 以align对齐*/
#define alloc_bootmem_align(x, align) \
__alloc_bootmem(x, align, BOOTMEM_LOW_LIMIT)
/*按指定size从ZONE_NORMAL区域分配, 以一页对齐*/
#define alloc_bootmem_pages(x) \
__alloc_bootmem(x, PAGE_SIZE, BOOTMEM_LOW_LIMIT)
/* SMP_CACHE_BYTES 是为了让数据能更好地在L1 cache中使用，虽然是SMP开头。*/
#define alloc_bootmem_nopanic(x) \
__alloc_bootmem_nopanic(x, SMP_CACHE_BYTES, BOOTMEM_LOW_LIMIT)
#define alloc_bootmem_node(pgdat, x) \
__alloc_bootmem_node(pgdat, x, SMP_CACHE_BYTES, BOOTMEM_LOW_LIMIT)
/*以_node后缀结尾表示只在NUMA系统上使用。*/
#define alloc_bootmem_node(pgdat, x) \
__alloc_bootmem_node(pgdat, x, SMP_CACHE_BYTES, BOOTMEM_LOW_LIMIT)
#define alloc_bootmem_node_nopanic(pgdat, x) \
__alloc_bootmem_node_nopanic(pgdat, x, SMP_CACHE_BYTES, BOOTMEM_LOW_LIMIT)
#define alloc_bootmem_pages_node(pgdat, x) \
__alloc_bootmem_node(pgdat, x, PAGE_SIZE, BOOTMEM_LOW_LIMIT)
#define alloc_bootmem_pages_node_nopanic(pgdat, x) \
__alloc_bootmem_node_nopanic(pgdat, x, PAGE_SIZE, BOOTMEM_LOW_LIMIT)
/*这几个和上面的区别是从ZONE_DMA区域分配。*/
#define alloc_bootmem_low(x) \
__alloc_bootmem_low(x, SMP_CACHE_BYTES, 0)
#define alloc_bootmem_low_pages(x) \
__alloc_bootmem_low(x, PAGE_SIZE, 0)
#define alloc_bootmem_low_pages_node(pgdat, x) \
__alloc_bootmem_low_node(pgdat, x, PAGE_SIZE, 0)
/*按指定size从ZONE_NORMAL区域分配*/ #define alloc_bootmem(x) \ __alloc_bootmem(x, SMP_CACHE_BYTES, BOOTMEM_LOW_LIMIT) /*按指定size从ZONE_NORMAL区域分配, 以align对齐*/ #define alloc_bootmem_align(x, align) \ __alloc_bootmem(x, align, BOOTMEM_LOW_LIMIT) /*按指定size从ZONE_NORMAL区域分配, 以一页对齐*/ #define alloc_bootmem_pages(x) \ __alloc_bootmem(x, PAGE_SIZE, BOOTMEM_LOW_LIMIT) /* SMP_CACHE_BYTES 是为了让数据能更好地在L1 cache中使用，虽然是SMP开头。*/ #define alloc_bootmem_nopanic(x) \ __alloc_bootmem_nopanic(x, SMP_CACHE_BYTES, BOOTMEM_LOW_LIMIT) #define alloc_bootmem_node(pgdat, x) \ __alloc_bootmem_node(pgdat, x, SMP_CACHE_BYTES, BOOTMEM_LOW_LIMIT) /*以_node后缀结尾表示只在NUMA系统上使用。*/ #define alloc_bootmem_node(pgdat, x) \ __alloc_bootmem_node(pgdat, x, SMP_CACHE_BYTES, BOOTMEM_LOW_LIMIT) #define alloc_bootmem_node_nopanic(pgdat, x) \ __alloc_bootmem_node_nopanic(pgdat, x, SMP_CACHE_BYTES, BOOTMEM_LOW_LIMIT) #define alloc_bootmem_pages_node(pgdat, x) \ __alloc_bootmem_node(pgdat, x, PAGE_SIZE, BOOTMEM_LOW_LIMIT) #define alloc_bootmem_pages_node_nopanic(pgdat, x) \ __alloc_bootmem_node_nopanic(pgdat, x, PAGE_SIZE, BOOTMEM_LOW_LIMIT) /*这几个和上面的区别是从ZONE_DMA区域分配。*/ #define alloc_bootmem_low(x) \ __alloc_bootmem_low(x, SMP_CACHE_BYTES, 0) #define alloc_bootmem_low_pages(x) \ __alloc_bootmem_low(x, PAGE_SIZE, 0) #define alloc_bootmem_low_pages_node(pgdat, x) \ __alloc_bootmem_low_node(pgdat, x, PAGE_SIZE, 0)
关于内存的分配，请允许我偷懒下，不对代码做详细分析了，有点费时间，流程如下：
__alloc_bootmem -> ___alloc_bootmem -> ___alloc_bootmem_nopanic -> alloc_bootmem_core -> find_next_zero_bit
大概的步骤就是：
1. 扫描bitmap位图，寻找空闲的位
2. 如果查找的页紧挨着上一次分配的页，就先检查这次要分配的内存是否能在上一页直接分配，因为bootmem allocator支持小于一页的分配。
3. 在新分配的页的bitmap对应的bit设置为1后，将当前的偏移保存，如果页没有完全分配，那么页里面的偏移量也保存。
Bootmem内存释放：
内核提供的释放bootmem接口是free_bootmem(unsignedlong addr, unsigned long size), 还有一个是用于NUMA的。这个接口没有分析，据资料记载说分配页可能会有风险。
Bootmem停用：
一旦伙伴系统初始化完成之后，bootmemallocator就要停止使用了，系统是通过函数free_all_bootmem()来停止的，有如下调用流程：start_kernel -> mm_init -> mem_init -> free_all_bootmem -> free_all_bootmem_core -> __free_pages_bootmem -> __free_pages
可以看到最终调用的是__free_pages()，这个函数会将这些Pages释放到伙伴系统中管理。
注意这里只是将空闲的页释放掉了。占据的页还存在。
由于bootmem分配的页里面的数据基本上都是用于内存基本结构，在系统运行期间会一直被用到，所以不会被释放。不过像__init这种类型的数据段只在系统开机的时候被使用，所以系统初始化完成之后就可以释放掉了。
系统使用的函数是free_initmem()， 调用流程如下：
start_kernel -> rest_init -> kernel_init -> init_post -> free_initmem
[html]
view plaincopyprint?
void free_initmem(void)
{
unsigned long reclaimed_initmem;
poison_init_mem(__init_begin, __init_end - __init_begin);
if (!machine_is_integrator() && !machine_is_cintegrator()) {
reclaimed_initmem =
free_area(__phys_to_pfn(__pa(__init_begin)),
__phys_to_pfn(__pa(__init_end)),
"init");
totalram_pages += reclaimed_initmem;
}
}
void free_initmem(void)
{
unsigned long reclaimed_initmem;
poison_init_mem(__init_begin, __init_end - __init_begin);
if (!machine_is_integrator() && !machine_is_cintegrator()) {
reclaimed_initmem = free_area(__phys_to_pfn(__pa(__init_begin)),
__phys_to_pfn(__pa(__init_end)),
"init");
totalram_pages += reclaimed_initmem;
}
}可以看到，释放的数据就是保存在__init_begin和__init_end那一段之间！当然，最后也是调用__free_pages()释放到伙伴系统中管理的。
使用bootmem分配内存的时候需要注意下一定要处于bootmem 分配器初始化和销毁之间！
2013/03/21
内容来自用户分享和网络整理，不保证内容的准确性，如有侵权内容，可联系管理员处理 
相关文章推荐
- bootmem_init分析(Bootmem Allocator)
- bootmem allocator分析
- boot memory allocator——自举内存分配器(五：停用bootmem)
- uclinux2.6(bf561)中的bootmem分析(1)：猜测
- stl源码分析之allocator
- uclinux2.6(bf561)中的bootmem分析(2)：调用前的参数分析
- memcached源码分析之内存管理 -- slab allocator(1)
- bootmem allocator分析
- uclinux2.6(bf561)中的bootmem分析(3)：init_bootmem_node
- Contiguous Memory Allocator (CMA) 源码分析
- Linux中的内存分配和释放之free_bootmem()函数分析
- uclinux2.6(bf561)中的bootmem分析(4)：alloc_bootmem_pages
- [STL]源码分析：allocator
- stl分析之allocator
- STL源码分析读书笔记--第二章--空间配置器（allocator）
- memcached源码分析之内存管理 -- Slab Allocator(2)
- C++ STL源码解析：空间配置器allocator分析
- mm_init分析(Zoned Buddy Allocator)
- [转]&nbsp;bootmem_init分析(Bootm…
- 分析free_all_bootmem_core