为什么要使用bootmem分配器，内存管理不是有buddy系统和slab分配器吗？由于在系统初始化的时候需要执行一些内存管理，内存分配的任务，这个时候buddy系统，slab分配器等并没有被初始化好，此时就引入了一种内存管理器bootmem分配器在系统初始化的时候进行内存管理与分配，当buddy系统和slab分配器初始化好后，在mem_init()中对bootmem分配器进行释放，内存管理与分配由buddy系统，slab分配器等进行接管。
bootmem分配器使用一个bitmap来标记物理页是否被占用，分配的时候按照第一适应的原则，从bitmap中进行查找，如果这位为1，表示已经被占用，否则表示未被占用。为什么系统运行的时候不使用bootmem分配器呢？bootmem分配器每次在bitmap中进行线性搜索，效率非常低，而且在内存的起始端留下许多小的空闲碎片，在需要非常大的内存块的时候，检查位图这一过程就显得代价很高。bootmem分配器是用于在启动阶段分配内存的，对该分配器的需求集中于简单性方面，而不是性能和通用性。
memblock算法是linux内核初始化阶段的一个内存分配器,本质上是取代了原来的bootmem算法. memblock实现比较简单,而它的作用就是在page allocator初始化之前来管理内存,完成分配和释放请求.
为了保证系统的兼容性, 内核为bootmem和memblock提供了相同的API接口.
这样在编译Kernel的时候可以选择nobootmem或者bootmem 来在buddy system起来之前管理memory. 这两种机制对提供的API是一致的，因此对用户是透明的
参见
mm/Makefile
ifdef CONFIG_NO_BOOTMEM
obj-y += nobootmem.o
else
obj-y += bootmem.o
endif
由于接口是一致的, 那么他们共同使用一份
| 头文件 | bootmem接口 | nobootmem接口 |
| include/linux/bootmem.h | mm/bootmem.c | mm/nobootmem.c |
Memblock是在早期引导过程中管理内存的方法之一，此时内核内存分配器还没运行. Memblock以前被定义为Logical Memory Block( 逻辑内存块), 但根据
Yinghai Lu的补丁
, 它被重命名为memblock.
+-------------------------------------------------------+
| 外部模块申请内存
|
+-------------------------------------------------------+
| |
| |
↓ ↓
+------------------------+ +------------------------+
|
bootmem.c
| | nobootmem.c |
| __alloc_bootmem（） | | __alloc_bootmem（） |
+------------------------+ +------------------------+
|
|
↓
+-----------------------------------+
| memblock.c |
| memblock_find_in_range_node（） |
+-----------------------------------+
这里仅仅介绍bootmem。
前面一节《
linux 内存管理---物理内存探测（二）》记录了物理内存的分布，那么之后就交由bootmem来管理了。
static void __init bootmem_init(void)
{
unsigned long reserved_end;
unsigned long mapstart = ~0UL;
unsigned long bootmap_size;
int i;
/*
* Init any data related to initrd. It's a nop if INITRD is
* not selected. Once that done we can determine the low bound
* of usable memory.
*/
reserved_end = max(init_initrd(),
(unsigned long) PFN_UP(__pa_symbol(&_end)));
//得到内核映像或者initrd占用的最后一个页框
/*
* max_low_pfn is not a number of pages. The number of pages
* of the system is given by 'max_low_pfn - min_low_pfn'.
*/
min_low_pfn = ~0UL;
max_low_pfn = 0;
/*
* Find the highest page frame number we have available.
*/
for (i = 0; i < boot_mem_map.nr_map; i++) {
unsigned long start, end;
if (boot_mem_map.map[i].type != BOOT_MEM_RAM)
continue;
start = PFN_UP(boot_mem_map.map[i].addr);
end = PFN_DOWN(boot_mem_map.map[i].addr
+ boot_mem_map.map[i].size);
if (end > max_low_pfn)
max_low_pfn = end;
if (start < min_low_pfn)
min_low_pfn = start;
if (end <= reserved_end)
continue;
if (start >= mapstart)
continue;
mapstart = max(reserved_end, start);
//得到mapstart的页框，用于bootmem记录分配的情况，mapstart就在内核映像后面的一个页框
}
if (min_low_pfn >= max_low_pfn)
panic("Incorrect memory mapping !!!");
if (min_low_pfn > ARCH_PFN_OFFSET) {
pr_info("Wasting %lu bytes for tracking %lu unused pages\n",
(min_low_pfn - ARCH_PFN_OFFSET) * sizeof(struct page),
min_low_pfn - ARCH_PFN_OFFSET);
} else if (min_low_pfn < ARCH_PFN_OFFSET) {
pr_info("%lu free pages won't be used\n",
ARCH_PFN_OFFSET - min_low_pfn);
}
min_low_pfn = ARCH_PFN_OFFSET;
//#define ARCH_PFN_OFFSET PFN_UP(PHYS_OFFSET)
/*
* Determine low and high memory ranges
*/
max_pfn = max_low_pfn;
if (max_low_pfn > PFN_DOWN(HIGHMEM_START)) {
//最大不超过0x20000000+768M
#ifdef CONFIG_HIGHMEM
highstart_pfn = PFN_DOWN(HIGHMEM_START);
highend_pfn = max_low_pfn;
#endif
max_low_pfn = PFN_DOWN(HIGHMEM_START);
}
/*
* Initialize the boot-time allocator with low memory only.
*/
bootmap_size =
init_bootmem_node(NODE_DATA(0),
mapstart,
min_low_pfn,
max_low_pfn);
//初始化bootmem, 最小页框，最大页框，包括中间的空洞
...
/*
* Register fully available low RAM pages with the bootmem allocator.
*/
for (i = 0; i < boot_mem_map.nr_map; i++) {
unsigned long start, end, size;
start = PFN_UP(boot_mem_map.map[i].addr);
end = PFN_DOWN(boot_mem_map.map[i].addr
+ boot_mem_map.map[i].size);
/*
* Reserve usable memory.
*/
switch (boot_mem_map.map[i].type) {
case BOOT_MEM_RAM:
break;
case BOOT_MEM_INIT_RAM:
memory_present(0, start, end);
continue;
default:
/* Not usable memory */
continue;
}
/*
* We are rounding up the start address of usable memory
* and at the end of the usable range downwards.
*/
if (start >= max_low_pfn)
continue;
if (start < reserved_end)
//从内核映像最后一个页框开始标记为可用
start = reserved_end;
if (end > max_low_pfn)
end = max_low_pfn;
/*
* ... finally, is the area going away?
*/
if (end <= start)
continue;
size = end - start;
/* Register lowmem ranges */
#ifdef CONFIG_BRCMSTB
/* carve out space for bmem */
brcm_free_bootmem(PFN_PHYS(start), size << PAGE_SHIFT);
//剔除bmem内存，bmem内存是保留给设备DMA用的
#else
free_bootmem(PFN_PHYS(start), size << PAGE_SHIFT);
//标记内核映像结束的页框到连续页框最后一个页框之间的页框为free可用页框
#endif
}
/*
* Reserve the bootmap memory.
*/
reserve_bootmem(PFN_PHYS(mapstart), bootmap_size, BOOTMEM_DEFAULT);
//标记bootmem用于分配标记占用的页为保留
/*
* Reserve initrd memory if needed.
*/
finalize_initrd();
//标记initrd占用的页为保留
/*
* Call memory_present() on all valid ranges, for
SPARSEMEM.
* This must be done after setting up bootmem, since memory_present()
* may allocate bootmem.
*/
for (i = 0; i < boot_mem_map.nr_map; i++) {
unsigned long start, end;
if (boot_mem_map.map[i].type != BOOT_MEM_RAM)
continue;
start = PFN_UP(boot_mem_map.map[i].addr);
end = PFN_DOWN(boot_mem_map.map[i].addr
+ boot_mem_map.map[i].size);
memory_present(0, start, end);
//主要是物理内存空洞，对于mips，低256M为DRAM，接着256M为register，接着768M为DRAM，所以对于系统内存大于256M，就肯定有内存空洞了
}
}
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);
}
在include/asm-mips/mach-generic/spaces.h中：
#ifndef PHYS_OFFSET
#define PHYS_OFFSET _AC(0, UL)
#endif
#ifdef CONFIG_32BIT
#define CAC_BASE _AC(0x80000000, UL)
#endif
#define BRCM_MAX_UPPER_MB _AC(768, UL)
#define UPPERMEM_START _AC(0x20000000, UL)
#define HIGHMEM_START (UPPERMEM_START + (BRCM_MAX_UPPER_MB << 20))
#ifndef PAGE_OFFSET
#define PAGE_OFFSET (CAC_BASE + PHYS_OFFSET)
#endif
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);
bdata->node_bootmem_map =
phys_to_virt(PFN_PHYS(mapstart));
//记录分配标志
bdata->node_min_pfn = start;
bdata->node_low_pfn = end;
link_bootmem(bdata);
/*
* Initially all pages are reserved - setup_arch() has to
* register free RAM areas explicitly.
*/
mapsize =
bootmap_bytes(end - start); //需要多少个byte来记录
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;
}
static unsigned long __init bootmap_bytes(unsigned long pages)
{
unsigned long bytes = DIV_ROUND_UP(pages, 8);
return ALIGN(bytes, sizeof(long));
}
一个byte有8bit，每个bit可用来记录一个页是否分配或释放，非0表示页可用，因此一个byte可用记录8个页。
void __init free_bootmem(unsigned long addr, unsigned long size)
{
unsigned long start, end;
kmemleak_free_part(__va(addr), size);
start = PFN_UP(addr);
end = PFN_DOWN(addr + size);
mark_bootmem(start, end, 0, 0);
}
static int __init mark_bootmem(unsigned long start, unsigned long end,
int reserve, int flags)
{
unsigned long pos;
bootmem_data_t *bdata;
pos = start;
list_for_each_entry(bdata, &bdata_list, list) {
int err;
unsigned long max;
if (pos < bdata->node_min_pfn ||
pos >= bdata->node_low_pfn) {
BUG_ON(pos != start);
continue;
}
max = min(bdata->node_low_pfn, end);
err = mark_bootmem_node(bdata, pos, max, reserve, flags);
if (reserve && err) {
mark_bootmem(start, pos, 0, 0);
return err;
}
if (max == end)
return 0;
pos = bdata->node_low_pfn;
}
BUG();
}
static int __init mark_bootmem_node(bootmem_data_t *bdata,
unsigned long start, unsigned long end,
int reserve, int flags)
{
unsigned long sidx, eidx;
sidx = start - bdata->node_min_pfn;
eidx = end - bdata->node_min_pfn;
if (reserve)
return __reserve(bdata, sidx, eidx, flags);
else
__free(bdata, sidx, eidx);
return 0;
}
static void __init __free(bootmem_data_t *bdata,
unsigned long sidx, unsigned long eidx)
{
unsigned long idx;
...
if (bdata->hint_idx > sidx)
bdata->hint_idx = sidx;
for (idx = sidx; idx < eidx; idx++)
if (!test_and_clear_bit(idx, bdata->node_bootmem_map))
BUG();
}
static int __init __reserve(bootmem_data_t *bdata, unsigned long sidx,
unsigned long eidx, int flags)
{
unsigned long idx;
int exclusive = flags & BOOTMEM_EXCLUSIVE;
for (idx = sidx; idx < eidx; idx++)
if (test_and_set_bit(idx, bdata->node_bootmem_map)) {
if (exclusive) {
//如果是互斥的，页框已经为1，再设置为reserve
__free(bdata, sidx, idx);
return -EBUSY;
}
bdebug("silent double reserve of PFN %lx\n",
idx + bdata->node_min_pfn);
}
return 0;
}
调用bootmem_init（）函数之后bootmem就初始化完成了，当然可能有人会问在bootmem初始化之前内核要分配内存怎么办，而且在bootmem初始化过程中要用到内存哪里来？这就是一个先有鸡还是先有蛋的问题，内核采取的办法是在bootmem可用之前包括bootmem的初始化，内核的一切内存需要都采用静态内存，即全局变量的形式，比如bootmem的初始化过程中：
NODE_DATA(0) 宏展开为：
#define NODE_DATA(nid) (&contig_page_data)
struct pglist_data __refdata contig_page_data = {
.bdata = &bootmem_node_data[0]
};
contig_page_data就是定义为一个全局结构体变量，其中bdata为它的成员变量指针，直接指向另外一个全局变量：
bootmem_data_t bootmem_node_data[MAX_NUMNODES] __initdata;
bootmem初始化完成后就可以通过下列函数分配内存了：
#define alloc_bootmem(x) \
__alloc_bootmem(x, SMP_CACHE_BYTES, BOOTMEM_LOW_LIMIT)
#define alloc_bootmem_align(x, align) \
__alloc_bootmem(x, align, BOOTMEM_LOW_LIMIT)
#define alloc_bootmem_nopanic(x) \
__alloc_bootmem_nopanic(x, SMP_CACHE_BYTES, BOOTMEM_LOW_LIMIT)
#define alloc_bootmem_pages(x) \
__alloc_bootmem(x, PAGE_SIZE, BOOTMEM_LOW_LIMIT)
#define alloc_bootmem_pages_nopanic(x) \
__alloc_bootmem_nopanic(x, PAGE_SIZE, BOOTMEM_LOW_LIMIT)
#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)
下面简单进行说明：
static void * __init alloc_bootmem_core(struct bootmem_data *bdata,
unsigned long size, unsigned long align,
unsigned long goal, unsigned long limit)
{
unsigned long fallback = 0;
unsigned long min, max, start, sidx, midx, step;
...
min = bdata->node_min_pfn;
max = bdata->node_low_pfn;
goal >>= PAGE_SHIFT;
limit >>= PAGE_SHIFT;
if (limit && max > limit)
max = limit;
if (max <= min)
return NULL;
step = max(align >> PAGE_SHIFT, 1UL);
if (goal && min < goal && goal < max)
start = ALIGN(goal, step);
else
start = ALIGN(min, step);
sidx = start - bdata->node_min_pfn;
midx = max - bdata->node_min_pfn;
if (bdata->hint_idx > sidx) {
/*
* Handle the valid case of sidx being zero and still
* catch the fallback below.
*/
fallback = sidx + 1;
sidx = align_idx(bdata, bdata->hint_idx, step);
}
while (1) {
int merge;
void *region;
unsigned long eidx, i, start_off, end_off;
find_block:
sidx = find_next_zero_bit(bdata->node_bootmem_map, midx, sidx);
//查找满足要求的起始页框
sidx = align_idx(bdata, sidx, step);
eidx = sidx + PFN_UP(size);
if (sidx >= midx || eidx > midx)
break;
for (i = sidx; i < eidx; i++)
if (test_bit(i, bdata->node_bootmem_map)) {
sidx = align_idx(bdata, i, step);
if (sidx == i)
sidx += step;
goto find_block;
}
if (bdata->last_end_off & (PAGE_SIZE - 1) &&
PFN_DOWN(bdata->last_end_off) + 1 == sidx)
start_off = align_off(bdata, bdata->last_end_off, align);
else
start_off = PFN_PHYS(sidx);
merge = PFN_DOWN(start_off) < sidx;
end_off = start_off + size;
bdata->last_end_off = end_off;
bdata->hint_idx = PFN_UP(end_off);
/*
* Reserve the area now:
*/
if (__reserve(bdata, PFN_DOWN(start_off) + merge,
PFN_UP(end_off), BOOTMEM_EXCLUSIVE))
//将分配后的页框设置为保留
BUG();
region = phys_to_virt(PFN_PHYS(bdata->node_min_pfn) +
start_off);
memset(region, 0, size);
/*
* The min_count is set to 0 so that bootmem allocated blocks
* are never reported as leaks.
*/
kmemleak_alloc(region, size, 0, 0);
return region;
}
if (fallback) {
sidx = align_idx(bdata, fallback - 1, step);
fallback = 0;
goto find_block;
}
return NULL;
}
参考文档：
860
被折叠的 条评论
为什么被折叠?
到【灌水乐园】发言
