十年网站开发经验 + 多家企业客户 + 靠谱的建站团队
量身定制 + 运营维护+专业推广+无忧售后,网站问题一站解决
malloc实现的方式有哪些?这篇文章运用了实例代码展示,代码非常详细,可供感兴趣的小伙伴们参考借鉴,希望对大家有所帮助。
创新互联建站2013年开创至今,是专业互联网技术服务公司,拥有项目做网站、网站设计网站策划,项目实施与项目整合能力。我们以让每一个梦想脱颖而出为使命,1280元上虞做网站,已为上家服务,为上虞各地企业和个人服务,联系电话:18982081108
方式1:K&R malloc
又叫做first-fit规则, 即查找第一个可用的匹配块。与之相对应的是查找第一个最符合(best-fit)的可用块。
K&R malloc的实现来自书籍 the C programming language by Kernighan and Ritchie (K&R) Section 8.7
维持的free list是一个环。 第一个元素是base。
#define NALLOC 1024 /* minimum #units to request */ struct header { struct header *ptr; size_t size; }; typedef struct header Header; static Header base; static Header *freep = NULL;
指定分配的内存大小为sizeof(Header)的倍数,且一定大于nbytes。nunits就是这个倍数。
循环free list。 找到第一个符合即大于等于nbytes的块。
— 如果刚好合适,则链表删除此块并返回此块。
— 如果大于,则截断此元素
— 如果没有找到合适的块,则调用moreheap新分配一个。
/* malloc: general-purpose storage allocator */ void * kr_malloc(size_t nbytes) { Header *p, *prevp; unsigned nunits; nunits = (nbytes + sizeof(Header) - 1) / sizeof(Header) + 1; // base作为第一个元素。 if ((prevp = freep) == NULL) { /* no free list yet */ base.ptr = freep = prevp = &base; base.size = 0; } for (p = prevp->ptr; ; prevp = p, p = p->ptr) { if (p->size >= nunits) { /* big enough */ if (p->size == nunits) /* exactly */ prevp->ptr = p->ptr; else { /* allocate tail end */ p->size -= nunits; p += p->size; p->size = nunits; } freep = prevp; return (void *) (p + 1); } if (p == freep) { /* wrapped around free list */ if ((p = (Header *) moreheap(nunits)) == NULL) { return NULL; /* none left */ } } } }
sbrk是unix增加内存的操作系统调用。
wiki的解释是:
The brk and sbrk calls dynamically change the amount of space allocated for the data segment of the calling process. The change is made by resetting the program break of the process, which determines the maximum space that can be allocated. The program break is the address of the first location beyond the current end of the data region. The amount of available space increases as the break value increases. The available space is initialized to a value of zero, unless the break is lowered and then increased, as it may reuse the same pages in some unspecified way. The break value can be automatically rounded up to a size appropriate for the memory management architecture.[4] Upon successful completion, the brk subroutine returns a value of 0, and the sbrk subroutine returns the prior value of the program break (if the available space is increased then this prior value also points to the start of the new area). If either subroutine is unsuccessful, a value of −1 is returned and the errno global variable is set to indicate the error.
由于这里是增加内存,sbrk成功时会返回新增加区域的开始地址,如果失败则会返回-1。
新生成一个块之后,调用free函数将其加入到freelist当中。
注意这里的up+1 是什么意思。其代表的是新分配的区域的开头。以为新分配的区域之前有Header大小,用于标识大小和下一个区域。
static Header *moreheap(size_t nu) { char *cp; Header *up; if (nu < NALLOC) nu = NALLOC; cp = sbrk(nu * sizeof(Header)); if (cp == (char *) -1) return NULL; up = (Header *) cp; up->size = nu; kr_free((void *)(up + 1)); return freep; }
ap是要释放的区域,其前面还有Header大小。bp 指向了块的开头。
遍历free list,找到要插入的中间区域。
如果前后的区域正好是连在一起的,则进行合并。
/* free: put block ap in free list */ void kr_free(void *ap) { Header *bp, *p; if (ap == NULL) return; bp = (Header *) ap - 1; /* point to block header */ for (p = freep; !(bp > p && bp < p->ptr); p = p->ptr) if (p >= p->ptr && (bp > p || bp < p->ptr)) break; /* freed block at start or end of arena */ if (bp + bp->sizfe == p->ptr) { /* join to upper nbr */ bp->size += p->ptr->size; bp->ptr = p->ptr->ptr; } else { bp->ptr = p->ptr; } if (p + p->size == bp) { /* join to lower nbr */ p->size += bp->size; p->ptr = bp->ptr; } else { p->ptr = bp; } freep = p; }
malloc 与free 快速
内存开销低
内存碎片严重
不通用,用于特定应用程序。
struct region { void *start; void *cur; void *end; }; typedef struct region Region; static Region rg_base; Region * rg_create(size_t nbytes) { rg_base.start = sbrk(nbytes); rg_base.cur = rg_base.start; rg_base.end = rg_base.start + nbytes; return &rg_base; } void * rg_malloc(Region *r, size_t nbytes) { assert (r->cur + nbytes <= r->end); void *p = r->cur; r->cur += nbytes; return p; } // free all memory in region resetting cur to start void rg_free(Region *r) { r->cur = r->start; }
我觉得wiki的解释挺好的:
wiki解析
提示:
对于2^k 大小的空间,我们可以将其分割为大小为2^0, 2^1, 2^2, … 2^k的多种可能。
malloc(17) 会分配 32 bytes,因此其会一定程度上浪费空间。
数据结构带来的格外内存开销
malloc 和free快速。
下面介绍一种代码实现。
基本参数
ROUNDUP(n,sz) 求出要分配n哥字节时,希望分配的实际内存是大于等于n 并且是sz的倍数。
#define LEAF_SIZE 16 // The smallest allocation size (in bytes) #define NSIZES 15 // Number of entries in bd_sizes array #define MAXSIZE (NSIZES-1) // Largest index in bd_sizes array #define BLK_SIZE(k) ((1L << (k)) * LEAF_SIZE) // Size in bytes for size k #define HEAP_SIZE BLK_SIZE(MAXSIZE) #define NBLK(k) (1 << (MAXSIZE-k)) // Number of block at size k #define ROUNDUP(n,sz) (((((n)-1)/(sz))+1)*(sz)) // Round up to the next multiple of sz
NBLK求出在第k位置有多少块。
每一个大小k维护一个sz_info
, sz_info
中都有一个free list, alloc 是一个char数组用于记录块是否分配。split 是一个char数组用于块是否割裂。
这里要注意,使用的是bit数组来记录。 char有8位,如第n位代表的是当前k大小的第5个区块。
// The allocator has sz_info for each size k. Each sz_info has a free // list, an array alloc to keep track which blocks have been // allocated, and an split array to to keep track which blocks have // been split. The arrays are of type char (which is 1 byte), but the // allocator uses 1 bit per block (thus, one char records the info of // 8 blocks). struct sz_info { struct bd_list free; char *alloc; char *split; }; typedef struct sz_info Sz_info;
每一个级别的双链表 for free list
// A double-linked list for the free list of each level struct bd_list { struct bd_list *next; struct bd_list *prev; };
// Return 1 if bit at position index in array is set to 1 int bit_isset(char *array, int index) { char b = array[index/8]; char m = (1 << (index % 8)); return (b & m) == m; } // Set bit at position index in array to 1 void bit_set(char *array, int index) { char b = array[index/8]; char m = (1 << (index % 8)); array[index/8] = (b | m); } // Clear bit at position index in array void bit_clear(char *array, int index) { char b = array[index/8]; char m = (1 << (index % 8)); array[index/8] = (b & ~m); }
首先调用mmap分配一个非常大的区域。此大小为2 ^ MAXSIZE * 16,16为分配的最小块。
// Allocate memory for the heap managed by the allocator, and allocate // memory for the data structures of the allocator. void bd_init() { bd_base = mmap(NULL, HEAP_SIZE, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANONYMOUS, -1, 0); if (bd_base == MAP_FAILED) { fprintf(stderr, "couldn't map heap; %s\n", strerror(errno)); assert(bd_base); } // printf("bd: heap size %d\n", HEAP_SIZE); for (int k = 0; k < NSIZES; k++) { lst_init(&bd_sizes[k].free); int sz = sizeof(char)*ROUNDUP(NBLK(k), 8)/8; bd_sizes[k].alloc = malloc(sz); memset(bd_sizes[k].alloc, 0, sz); } for (int k = 1; k < NSIZES; k++) { int sz = sizeof(char)*ROUNDUP(NBLK(k), 8)/8; bd_sizes[k].split = malloc(sz); memset(bd_sizes[k].split, 0, sz); } lst_push(&bd_sizes[MAXSIZE].free, bd_base); }
// What is the first k such that 2^k >= n? int firstk(size_t n) { int k = 0; size_t size = LEAF_SIZE; while (size < n) { k++; size *= 2; } return k; }
// Compute the block index for address p at size k int blk_index(int k, char *p) { int n = p - (char *) bd_base; return n / BLK_SIZE(k); }
// Convert a block index at size k back into an address void *addr(int k, int bi) { int n = bi * BLK_SIZE(k); return (char *) bd_base + n; }
找到一个大小k,块2^k是大于等于要分配的大小。
如果db_size[k].free 不为空,说明当前有大小为k的空闲空间。
如果没有找到,则让k+1,继续找到更大的空间有无空闲。
如果找到,则lst_pop(&bd_sizes[k].free)获取第一个块。 blk_index(k, p)获取以k为衡量指标,p位于的第n个块。 并通过bit_set将bd_sizes[k].alloc 在第n号位置设置为1。
如果找到的k是比较大的空间,这时候需要对此空间进行分割,一分为2。 一直到找到一个比较符合的空间。
void * bd_malloc(size_t nbytes) { int fk, k; assert(bd_base != NULL); // Find a free block >= nbytes, starting with smallest k possible fk = firstk(nbytes); for (k = fk; k < NSIZES; k++) { if(!lst_empty(&bd_sizes[k].free)) break; } if(k >= NSIZES) // No free blocks? return NULL; // Found one; pop it and potentially split it. char *p = lst_pop(&bd_sizes[k].free); bit_set(bd_sizes[k].alloc, blk_index(k, p)); for(; k > fk; k--) { // 第2半的空间 char *q = p + BLK_SIZE(k-1); // 对于大小k来说,其在位置p处是分割的。 bit_set(bd_sizes[k].split, blk_index(k, p)); // 对于大小k-1来说,其在位置p处是分配的。 bit_set(bd_sizes[k-1].alloc, blk_index(k-1, p)); // 对于大小k-1来说,其在位置q处是空闲的。 lst_push(&bd_sizes[k-1].free, q); } // printf("malloc: %p size class %d\n", p, fk); return p; }
当要free位置p时,size找到第一个k,当k+1在p位置是分割的,则返回k。这时候说明在k处区域是可以合并的。这是一种优化。// Find the size of the block that p points to. int size(char *p) { for (int k = 0; k < NSIZES; k++) { if(bit_isset(bd_sizes[k+1].split, blk_index(k+1, p))) { return k; } } return 0; }
void bd_free(void *p) { void *q; int k; for (k = size(p); k < MAXSIZE; k++) { int bi = blk_index(k, p); bit_clear(bd_sizes[k].alloc, bi); int buddy = (bi % 2 == 0) ? bi+1 : bi-1; if (bit_isset(bd_sizes[k].alloc, buddy)) { break; } // budy is free; merge with buddy q = addr(k, buddy); lst_remove(q); if(buddy % 2 == 0) { p = q; } bit_clear(bd_sizes[k+1].split, blk_index(k+1, p)); } // 放入freelist当中。 // printf("free %p @ %d\n", p, k); lst_push(&bd_sizes[k].free, p); }
// Implementation of lists: double-linked and circular. Double-linked // makes remove fast. Circular simplifies code, because don't have to // check for empty list in insert and remove. void lst_init(Bd_list *lst) { lst->next = lst; lst->prev = lst; } int lst_empty(Bd_list *lst) { return lst->next == lst; } void lst_remove(Bd_list *e) { e->prev->next = e->next; e->next->prev = e->prev; } void* lst_pop(Bd_list *lst) { assert(lst->next != lst); Bd_list *p = lst->next; lst_remove(p); return (void *)p; } void lst_push(Bd_list *lst, void *p) { Bd_list *e = (Bd_list *) p; e->next = lst->next; e->prev = lst; lst->next->prev = p; lst->next = e; } void lst_print(Bd_list *lst) { for (Bd_list *p = lst->next; p != lst; p = p->next) { printf(" %p", p); } printf("\n"); }
* dlmalloc * slab allocator
以上就是malloc实现的几种方式的详细介绍内容了,看完之后是否有所收获呢?如果想了解更多相关内容,欢迎关注创新互联行业资讯!