转自:https://www.cnblogs.com/yibuyibu/p/14806878.html
什么是CMA 参考这两篇博文,写得很好: http://www.wowotech.net/memory_management/cma.html https://www.cnblogs.com/LoyenWang/p/12182594.html https://biscuitos.github.io/blog/CMA/ CMA的初始化创建 * 默认cma创建(dma_contiguous_default_area),两种方式: 通过cmdline传递的参数"cma=",然后在kernel初始化阶段解析参数,并调用start_kernel()->setup_arch()->arm64_memblock_init()->dma_contiguous_reserve()完成创建(android中一般不通过cmdline传递): static phys_addr_t size_cmdline = -1; static phys_addr_t base_cmdline; static phys_addr_t limit_cmdline; //解析cmdline传递的cma参数 static int __init early_cma(char *p) { pr_debug("%s(%s)\n", __func__, p); size_cmdline = memparse(p, &p); if (*p != '@') return 0; base_cmdline = memparse(p + 1, &p); if (*p != '-') { limit_cmdline = base_cmdline + size_cmdline; return 0; } limit_cmdline = memparse(p + 1, &p); return 0; } early_param("cma", early_cma); 通过dts中配置cma节点,属性中包含"shared-dma-pool"以及"linux,cma-default",在kernel初始化阶段,通过调用start_kernel()->setup_arch()->arm64_memblock_init()->early_init_fdt_scan_reserved_mem()->fdt_init_reserved_mem()->__reserved_mem_init_node()完成对默认cma的创建和初始化: static int __init __reserved_mem_init_node(struct reserved_mem *rmem) { extern const struct of_device_id __reservedmem_of_table[]; const struct of_device_id *i; //__reservedmem_of_table是初始化中的一个section段,通过RESERVEDMEM_OF_DECLARE定义的都会被链接到这个段中 //参考:https://blog.csdn.net/rikeyone/article/details/79975138 for (i = __reservedmem_of_table; i < &__rmem_of_table_sentinel; i++) { reservedmem_of_init_fn initfn = i->data; const char *compat = i->compatible; if (!of_flat_dt_is_compatible(rmem->fdt_node, compat)) continue; if (initfn(rmem) == 0) { pr_info("initialized node %s, compatible id %s\n", rmem->name, compat); return 0; } } return -ENOENT; } //dma-contiguous.c文件中定义了该默认cma的创建回调。 //如果dts中没有配置,那该回调也不会执行。 //参考:https://blog.csdn.net/rikeyone/article/details/79975138 RESERVEDMEM_OF_DECLARE(cma, "shared-dma-pool", rmem_cma_setup); 默认cma似乎在好些android平台上都没有创建。 *其他CMA区创建 其他CMA区域创建都应该类似默认cma一样,通过RESERVEDMEM_OF_DECLARE接口定义一个结构体变量在__reservedmem_of_table段中,开机启动时就会调用对应的initfn完成初始化,同时还需要在dts中配置对应的属性节点。 所有CMA的创建最终都会调用cma_init_reserved_mem()函数: 主要从cma全局数组cma_areas中分配一个cma实体并将传递过来的参数用于初始化该cam实体。 初始化参数包括,cma的name、起始页框号base_pfn,总共页数count,以及每个bit代表多少个页2^(order_per_bit)。 更新全局变量totalcma_pages,记录总的cma页面数量,在meminfo中CmaTotal就是这个值。 int __init cma_init_reserved_mem(phys_addr_t base, phys_addr_t size, unsigned int order_per_bit, const char *name, struct cma **res_cma) { struct cma *cma; phys_addr_t alignment; /* Sanity checks */ //判断cma数量是否已经满了,因为cma_areas数组指定了系统中总的cma数量,通过内核宏控制 if (cma_area_count == ARRAY_SIZE(cma_areas)) { pr_err("Not enough slots for CMA reserved regions!\n"); return -ENOSPC; } //判断该cma内存区间是否与reversed中的某个区间是交叉的?为什么要这样判断? if (!size || !memblock_is_region_reserved(base, size)) return -EINVAL; /* ensure minimal alignment required by mm core */ //对齐方式按pageblock,也就是1024页(4M) alignment = PAGE_SIZE << max_t(unsigned long, MAX_ORDER - 1, pageblock_order); /* alignment should be aligned with order_per_bit */ //判断对齐方式alignment本身的大小与单个bit表示的内存大小,是否对齐 if (!IS_ALIGNED(alignment >> PAGE_SHIFT, 1 << order_per_bit)) return -EINVAL; //判断base和size以aligment方式对齐后,得到的值是否还是原来的值,也就是判断base和size是否基于alignment对齐 if (ALIGN(base, alignment) != base || ALIGN(size, alignment) != size) return -EINVAL; /* * Each reserved area must be initialised later, when more kernel * subsystems (like slab allocator) are available. */ //1. memblock是系统最初的内存管理器,分为memory type和reserved type,CMA最开始就属于reserved type //2. 运行到这里,就表示memblock已经建立,并且buddy还没建立,CMA在buddy前建立OK //3. CMA建立OK后,接着memblock中的memory type会释放给buddy,reserved type则不会 //4. CMA作为特殊的reserved type,最终通过系统初始化调用cma_init_reserved_areas,将内存归还给buddy //从cma_areas数组中分配一个cma对象 cma = &cma_areas[cma_area_count]; if (name) { cma->name = name; } else { cma->name = kasprintf(GFP_KERNEL, "cma%d\n", cma_area_count); if (!cma->name) return -ENOMEM; } cma->base_pfn = PFN_DOWN(base); //起始页号 cma->count = size >> PAGE_SHIFT; //总共页面数 cma->order_per_bit = order_per_bit; //一个bit代表的阶数 *res_cma = cma; cma_area_count++; totalcma_pages += (size / PAGE_SIZE); //totalcma_pages记录总的cma页面数量,在meminfo中CmaTotal就是这个值 return 0; } 到这里,只是完成对cma内存的保留和初始化,cma区最终还需要释放给buddy。 CMA区域释放给buddy 释放也是在kernel初始化过程中,会比cma的创建稍晚一些,是通过cma_init_reserved_areas接口完成的所有cma的初始化并将内存返还给buddy。 core_initcall(cma_init_reserved_areas)定义在kernel的init段中,通过start_kernel()->rest_init()创建内核线程kernel_init->kernel_init_freeable()->do_basic_setup()->do_initcalls()完成对各个init level的初始化。core init属于level1。 cma_init_reserved_areas()函数,遍历当前cma全局数组中已经分配的cma实体,通过调用cma_activate_area函数完成激活初始化,同时将内存释放给buddy: static int __init cma_init_reserved_areas(void) { int i; for (i = 0; i < cma_area_count; i++) { int ret = cma_activate_area(&cma_areas[i]); if (ret) return ret; } return 0; } core_initcall(cma_init_reserved_areas); cma_activate_area()函数: 以pageblock为单位,设置migrate type为MIGRATE_CMA,然后将其整个pageblock包含的页全部释放给buddy,并更新整个系统的可用内存总数 static int __init cma_activate_area(struct cma *cma) { int bitmap_size = BITS_TO_LONGS(cma_bitmap_maxno(cma)) * sizeof(long); unsigned long base_pfn = cma->base_pfn, pfn = base_pfn; //i代表有多少个page block,一般一个pageblock是1024页 unsigned i = cma->count >> pageblock_order; struct zone *zone; //cma也是通过bitmap来管理,每个bit代表多大,由order_per_bit决定。 //默认的cma的order_per_bit为0,一个bit代表2^0个page。 //分配bitmap cma->bitmap = kzalloc(bitmap_size, GFP_KERNEL); if (!cma->bitmap) return -ENOMEM; WARN_ON_ONCE(!pfn_valid(pfn)); zone = page_zone(pfn_to_page(pfn)); //以pageblock遍历, do { unsigned j; //记录当前pageblock的起始页 base_pfn = pfn; //判断当前pageblock中的所有页面是否满足要求:合法的页号、都在同一个zone中 for (j = pageblock_nr_pages; j; --j, pfn++) { WARN_ON_ONCE(!pfn_valid(pfn)); /* * alloc_contig_range requires the pfn range * specified to be in the same zone. Make this * simple by forcing the entire CMA resv range * to be in the same zone. */ if (page_zone(pfn_to_page(pfn)) != zone) goto not_in_zone; } //将当前pageblock初始化并释放给buddy init_cma_reserved_pageblock(pfn_to_page(base_pfn)); } while (--i); mutex_init(&cma->lock); #ifdef CONFIG_CMA_DEBUGFS INIT_HLIST_HEAD(&cma->mem_head); spin_lock_init(&cma->mem_head_lock); #endif return 0; not_in_zone: pr_err("CMA area %s could not be activated\n", cma->name); kfree(cma->bitmap); cma->count = 0; return -EINVAL; } cma_activate_area()->init_cma_reserved_pageblock()函数设置pageblock类型并释放内存给buddy: void __init init_cma_reserved_pageblock(struct page *page) { unsigned i = pageblock_nr_pages; struct page *p = page; do { //清除页描述flag中的PG_Reserved标志位 __ClearPageReserved(p); //设置page->_refcount = 0 set_page_count(p, 0); } while (++p, --i); //设置pageblock的迁移类型为MIGRATE_CMA set_pageblock_migratetype(page, MIGRATE_CMA); if (pageblock_order >= MAX_ORDER) { i = pageblock_nr_pages; p = page; do { set_page_refcounted(p); __free_pages(p, MAX_ORDER - 1); p += MAX_ORDER_NR_PAGES; } while (i -= MAX_ORDER_NR_PAGES); } else { //设置page->_refcount = 1 set_page_refcounted(page); //释放pages到buddy中,以pageblock释放,order为10 __free_pages(page, pageblock_order); } //调整对应zone中的managed_pages可管理页面数,即加上一个pageblock数量 //调整总的内存数量totalram_pages,即加上一个pageblock数量 adjust_managed_page_count(page, pageblock_nr_pages); } CMA的分配 CMA分配通过统一接口cma_alloc函数,会从bitmap中先查找满足要求的连续bit,然后通过alloc_contig_range实现分配,成功后的页面会从buddy总摘出来: struct page *cma_alloc(struct cma *cma, size_t count, unsigned int align, gfp_t gfp_mask) { unsigned long mask, offset; unsigned long pfn = -1; unsigned long start = 0; unsigned long bitmap_maxno, bitmap_no, bitmap_count; struct page *page = NULL; int ret = -ENOMEM; if (!cma || !cma->count) return NULL; pr_debug("%s(cma %p, count %zu, align %d)\n", __func__, (void *)cma, count, align); if (!count) return NULL; mask = cma_bitmap_aligned_mask(cma, align); offset = cma_bitmap_aligned_offset(cma, align); bitmap_maxno = cma_bitmap_maxno(cma); bitmap_count = cma_bitmap_pages_to_bits(cma, count); if (bitmap_count > bitmap_maxno) return NULL; for (;;) { mutex_lock(&cma->lock); //1. 从cma->bitmap中查找连续bitmap_count个为0的bit bitmap_no = bitmap_find_next_zero_area_off(cma->bitmap, bitmap_maxno, start, bitmap_count, mask, offset); if (bitmap_no >= bitmap_maxno) { mutex_unlock(&cma->lock); break; } //2. 将查找到的连续bit设置为1,表示内存被分配占用 bitmap_set(cma->bitmap, bitmap_no, bitmap_count); /* * It's safe to drop the lock here. We've marked this region for * our exclusive use. If the migration fails we will take the * lock again and unmark it. */ mutex_unlock(&cma->lock); //3. 计算分配的起始页的页号 pfn = cma->base_pfn + (bitmap_no << cma->order_per_bit); mutex_lock(&cma_mutex); //4. 分配从起始页开始的连续count个页,分配的migrate type为CMA类型 ret = alloc_contig_range(pfn, pfn + count, MIGRATE_CMA, gfp_mask); mutex_unlock(&cma_mutex); //5. 分配成功,就返回起始page if (ret == 0) { page = pfn_to_page(pfn); break; } cma_clear_bitmap(cma, pfn, count); if (ret != -EBUSY) break; pr_debug("%s(): memory range at %p is busy, retrying\n", __func__, pfn_to_page(pfn)); /* try again with a bit different memory target */ start = bitmap_no + mask + 1; } trace_cma_alloc(pfn, page, count, align); if (ret && !(gfp_mask & __GFP_NOWARN)) { pr_info("%s: alloc failed, req-size: %zu pages, ret: %d\n", __func__, count, ret); cma_debug_show_areas(cma); } pr_debug("%s(): returned %p\n", __func__, page); return page; } CMA的释放 释放操作也很清晰,通过cma_release函数实现,会将页面释放回buddy系统,并将cma的bitmap相应bit清零: bool cma_release(struct cma *cma, const struct page *pages, unsigned int count) { unsigned long pfn; if (!cma || !pages) return false; pr_debug("%s(page %p)\n", __func__, (void *)pages); pfn = page_to_pfn(pages); if (pfn < cma->base_pfn || pfn >= cma->base_pfn + cma->count) return false; VM_BUG_ON(pfn + count > cma->base_pfn + cma->count); //释放回buddy free_contig_range(pfn, count); //清零bit位,表示对应cma内存可用 cma_clear_bitmap(cma, pfn, count); trace_cma_release(pfn, pages, count); return true; } CMA与buddy 后续补充