Linux中setup_arch和setup_memory相关函数的实现
初始化页面地址映射page_address_init
void __init page_address_init(void)
{int i;INIT_LIST_HEAD(&page_address_pool);for (i = 0; i < ARRAY_SIZE(page_address_maps); i++)list_add(&page_address_maps[i].list, &page_address_pool);for (i = 0; i < ARRAY_SIZE(page_address_htable); i++) {INIT_LIST_HEAD(&page_address_htable[i].lh);spin_lock_init(&page_address_htable[i].lock);}spin_lock_init(&pool_lock);
}
1. 初始化空闲链表池
INIT_LIST_HEAD(&page_address_pool)初始化一个全局的空闲链表头page_address_pool，用于管理可用的page_address_map结构体- 通过循环将
page_address_maps数组中的所有元素（预分配的映射结构体）添加到page_address_pool链表中，作为初始空闲资源
2. 初始化哈希表
page_address_htable是一个哈希表数组，用于快速查找页面地址映射- 对每个哈希桶
page_address_htable[i]INIT_LIST_HEAD(&page_address_htable[i].lh)初始化链表头，用于存储映射条目spin_lock_init(&page_address_htable[i].lock)初始化自旋锁，保证多核环境下对哈希桶的并发访问安全
3. 初始化全局锁
spin_lock_init(&pool_lock)初始化pool_lock，用于保护空闲池page_address_pool的并发访问
4. 背景知识
- 用途：该函数为内核的 高端内存（
HighMem） 映射机制提供基础支持。在 32 位系统中，内核虚拟地址空间有限，无法直接映射所有物理内存，需要通过动态映射（如kmap()）访问高端内存页。page_address_map结构体用于记录页面虚拟地址与物理页面的映射关系 - 数据结构
page_address_pool：空闲的page_address_map结构体池，避免频繁内存分配page_address_htable：哈希表，通过页面指针快速查找对应的映射信息pool_lock：保护空闲池的自旋锁
架构相关的初始化setup_arch
void __init setup_arch(char **cmdline_p)
{unsigned long max_low_pfn;memcpy(&boot_cpu_data, &new_cpu_data, sizeof(new_cpu_data));pre_setup_arch_hook();early_cpu_init();/** FIXME: This isn't an official loader_type right* now but does currently work with elilo.* If we were configured as an EFI kernel, check to make* sure that we were loaded correctly from elilo and that* the system table is valid. If not, then initialize normally.*/
#ifdef CONFIG_EFIif ((LOADER_TYPE == 0x50) && EFI_SYSTAB)efi_enabled = 1;
#endifROOT_DEV = old_decode_dev(ORIG_ROOT_DEV);drive_info = DRIVE_INFO;screen_info = SCREEN_INFO;edid_info = EDID_INFO;apm_info.bios = APM_BIOS_INFO;ist_info = IST_INFO;saved_videomode = VIDEO_MODE;if( SYS_DESC_TABLE.length != 0 ) {MCA_bus = SYS_DESC_TABLE.table[3] &0x2;machine_id = SYS_DESC_TABLE.table[0];machine_submodel_id = SYS_DESC_TABLE.table[1];BIOS_revision = SYS_DESC_TABLE.table[2];}aux_device_present = AUX_DEVICE_INFO;#ifdef CONFIG_BLK_DEV_RAMrd_image_start = RAMDISK_FLAGS & RAMDISK_IMAGE_START_MASK;rd_prompt = ((RAMDISK_FLAGS & RAMDISK_PROMPT_FLAG) != 0);rd_doload = ((RAMDISK_FLAGS & RAMDISK_LOAD_FLAG) != 0);
#endifARCH_SETUPif (efi_enabled)efi_init();else {printk(KERN_INFO "BIOS-provided physical RAM map:\n");print_memory_map(machine_specific_memory_setup());}copy_edd();if (!MOUNT_ROOT_RDONLY)root_mountflags &= ~MS_RDONLY;init_mm.start_code = (unsigned long) _text;init_mm.end_code = (unsigned long) _etext;init_mm.end_data = (unsigned long) _edata;init_mm.brk = init_pg_tables_end + PAGE_OFFSET;code_resource.start = virt_to_phys(_text);code_resource.end = virt_to_phys(_etext)-1;data_resource.start = virt_to_phys(_etext);data_resource.end = virt_to_phys(_edata)-1;parse_cmdline_early(cmdline_p);max_low_pfn = setup_memory();/** NOTE: before this point _nobody_ is allowed to allocate* any memory using the bootmem allocator. Although the* alloctor is now initialised only the first 8Mb of the kernel* virtual address space has been mapped. All allocations before* paging_init() has completed must use the alloc_bootmem_low_pages()* variant (which allocates DMA'able memory) and care must be taken* not to exceed the 8Mb limit.*/#ifdef CONFIG_SMPsmp_alloc_memory(); /* AP processor realmode stacks in low memory*/
#endifpaging_init();/** NOTE: at this point the bootmem allocator is fully available.*/#ifdef CONFIG_EARLY_PRINTK{char *s = strstr(*cmdline_p, "earlyprintk=");if (s) {extern void setup_early_printk(char *);setup_early_printk(s);printk("early console enabled\n");}}
#endifdmi_scan_machine();#ifdef CONFIG_X86_GENERICARCHgeneric_apic_probe(*cmdline_p);
#endifif (efi_enabled)efi_map_memmap();/** Parse the ACPI tables for possible boot-time SMP configuration.*/acpi_boot_init();#ifdef CONFIG_X86_LOCAL_APICif (smp_found_config)get_smp_config();
#endifregister_memory(max_low_pfn);#ifdef CONFIG_VT
#if defined(CONFIG_VGA_CONSOLE)if (!efi_enabled || (efi_mem_type(0xa0000) != EFI_CONVENTIONAL_MEMORY))conswitchp = &vga_con;
#elif defined(CONFIG_DUMMY_CONSOLE)conswitchp = &dummy_con;
#endif
#endif
}
1. 函数功能概述
setup_arch() 是 Linux 内核架构相关的初始化函数，负责在系统启动时初始化与体系结构（如 x86、ARM）相关的硬件和内存布局。它的主要任务包括：
- CPU 和硬件信息初始化（如 CPU 型号、BIOS 数据、EFI 检测）
- 内存管理初始化（如物理内存映射、内存分配器设置）
- 核心数据结构设置（如内核代码/数据段的地址范围）
- ACPI/SMP 初始化（如多核支持）
- 设备信息收集（如 DMI、EDD 信息）
- 控制台初始化（如 VGA 终端）
2. 代码分段解析
2.1. CPU 和硬件信息初始化
memcpy(&boot_cpu_data, &new_cpu_data, sizeof(new_cpu_data));
pre_setup_arch_hook();
early_cpu_init();
- 初始化 CPU 相关数据
memcpy将new_cpu_data复制到boot_cpu_data，保存当前 CPU 的信息pre_setup_arch_hook()是架构相关的钩子函数，暂为空early_cpu_init()进一步初始化 CPU 特性
2.2. EFI 和 BIOS 信息检测
#ifdef CONFIG_EFIif ((LOADER_TYPE == 0x50) && EFI_SYSTAB)efi_enabled = 1;
#endif
- 检测是否通过 EFI 启动
- 如果启动加载器类型（
LOADER_TYPE）是0x50且存在 EFI 系统表（EFI_SYSTAB），则启用 EFI 支持（efi_enabled = 1）
- 如果启动加载器类型（
2.3. 硬件参数传递
ROOT_DEV = old_decode_dev(ORIG_ROOT_DEV);
drive_info = DRIVE_INFO;
screen_info = SCREEN_INFO;
edid_info = EDID_INFO;
apm_info.bios = APM_BIOS_INFO;
ist_info = IST_INFO;
saved_videomode = VIDEO_MODE;
- 从启动加载器（如 GRUB）传递硬件参数到内核。
ORIG_ROOT_DEV是启动时指定的根文件系统设备，old_decode_dev将其转换为内核可用的格式- 其他变量（如
drive_info、screen_info）保存磁盘、屏幕、电源管理等硬件信息
2.4. 系统描述表解析
if (SYS_DESC_TABLE.length != 0) {MCA_bus = SYS_DESC_TABLE.table[3] & 0x2;machine_id = SYS_DESC_TABLE.table[0];machine_submodel_id = SYS_DESC_TABLE.table[1];BIOS_revision = SYS_DESC_TABLE.table[2];
}
- 解析 BIOS 提供的系统描述表
SYS_DESC_TABLE- 检测是否使用 MCA总线
- 保存机器 ID、子型号和 BIOS 版本信息
2.5. 内存初始化准备
#ifdef CONFIG_BLK_DEV_RAMrd_image_start = RAMDISK_FLAGS & RAMDISK_IMAGE_START_MASK;rd_prompt = ((RAMDISK_FLAGS & RAMDISK_PROMPT_FLAG) != 0);rd_doload = ((RAMDISK_FLAGS & RAMDISK_LOAD_FLAG) != 0);
#endif
- 初始化 RAMDISK（内存磁盘）参数
- 确定
initramfs镜像在内存中的起始地址，内核后续会使用此地址找到initramfs镜像并解压到根文件系统 - 控制是否在加载
initramfs时显示交互式提示，内核在加载initramfs前等待用户确认 - 控制是否实际加载
initramfs，嵌入式设备可能禁用此标志以节省内存
- 确定
2.6. 架构特定初始化
ARCH_SETUP
if (efi_enabled)efi_init();
else {printk(KERN_INFO "BIOS-provided physical RAM map:\n");print_memory_map(machine_specific_memory_setup());
}
- 调用架构相关的初始化代码
- 如果是 EFI 启动，调用
efi_init() - 否则，打印 BIOS 提供的物理内存布局（通过
machine_specific_memory_setup()获取）
- 如果是 EFI 启动，调用
2.7. EDD 和根文件系统设置
copy_edd();
if (!MOUNT_ROOT_RDONLY)root_mountflags &= ~MS_RDONLY;
- 作用
copy_edd()复制 BIOS 的 EDD（Enhanced Disk Drive）信息，帮助操作系统在启动时准确识别和访问磁盘设备- 根据
MOUNT_ROOT_RDONLY决定根文件系统是否以只读方式挂载
2.8. 内核代码/数据段地址设置
init_mm.start_code = (unsigned long) _text;
init_mm.end_code = (unsigned long) _etext;
init_mm.end_data = (unsigned long) _edata;
init_mm.brk = init_pg_tables_end + PAGE_OFFSET;code_resource.start = virt_to_phys(_text);
code_resource.end = virt_to_phys(_etext)-1;
data_resource.start = virt_to_phys(_etext);
data_resource.end = virt_to_phys(_edata)-1;
- 初始化内核内存管理结构
init_mm和资源描述符code_resource、data_resource_text、_etext、_edata是链接脚本定义的符号，标记代码段和数据段的虚拟地址virt_to_phys将虚拟地址转换为物理地址
2.9. 命令行解析
parse_cmdline_early(cmdline_p);
- 作用：解析内核启动命令行参数，可能影响后续初始化（如内存大小、调试选项）
2.10. 物理内存初始化
max_low_pfn = setup_memory();
- 作用：初始化物理内存管理器（如
bootmem或memblock），返回可用内存的最大页帧号（max_low_pfn）
2.11. SMP 和分页初始化
#ifdef CONFIG_SMPsmp_alloc_memory(); /* AP processor realmode stacks in low memory */
#endif
paging_init();
- 作用
- 如果是多核系统（
CONFIG_SMP），为 AP（Application Processor）分配启动栈。 paging_init()初始化分页机制，设置内核页表
- 如果是多核系统（
2.12. 早期控制台和 DMI 扫描
#ifdef CONFIG_EARLY_PRINTK{char *s = strstr(*cmdline_p, "earlyprintk=");if (s) {extern void setup_early_printk(char *);setup_early_printk(s);printk("early console enabled\n");}}
#endif
dmi_scan_machine();
- 作用
- 如果内核命令行指定
earlyprintk，启用早期调试控制台 dmi_scan_machine()扫描 DMI（SMBIOS）信息（如厂商、型号）
- 如果内核命令行指定
2.13. ACPI 和 SMP 配置
#ifdef CONFIG_X86_GENERICARCHgeneric_apic_probe(*cmdline_p);
#endif
if (efi_enabled)efi_map_memmap();
acpi_boot_init();
#ifdef CONFIG_X86_LOCAL_APICif (smp_found_config)get_smp_config();
#endif
- 作用
- 探测 APIC（高级可编程中断控制器）
- 如果是 EFI 系统，映射 EFI 内存布局
acpi_boot_init()解析 ACPI 表（如 MADT 表）初始化多核- 如果发现 SMP 配置（
smp_found_config），加载多核配置（get_smp_config）
2.14. 内存注册和控制台初始化
register_memory(max_low_pfn);
#ifdef CONFIG_VT
#if defined(CONFIG_VGA_CONSOLE)if (!efi_enabled || (efi_mem_type(0xa0000) != EFI_CONVENTIONAL_MEMORY))conswitchp = &vga_con;
#elif defined(CONFIG_DUMMY_CONSOLE)conswitchp = &dummy_con;
#endif
#endif
- 作用
register_memory注册内存资源（如/proc/iomem）- 初始化控制台：
- 如果是 VGA 控制台且非 EFI 或 EFI 未占用
0xA0000地址，使用 VGA 终端 - 否则使用虚拟控制台（
dummy_con）
- 如果是 VGA 控制台且非 EFI 或 EFI 未占用
早期内存初始化setup_memory
static unsigned long __init setup_memory(void)
{unsigned long bootmap_size, start_pfn, max_low_pfn;/** partially used pages are not usable - thus* we are rounding upwards:*/start_pfn = PFN_UP(init_pg_tables_end);find_max_pfn();max_low_pfn = find_max_low_pfn();#ifdef CONFIG_HIGHMEMhighstart_pfn = highend_pfn = max_pfn;if (max_pfn > max_low_pfn) {highstart_pfn = max_low_pfn;}printk(KERN_NOTICE "%ldMB HIGHMEM available.\n",pages_to_mb(highend_pfn - highstart_pfn));
#endifprintk(KERN_NOTICE "%ldMB LOWMEM available.\n",pages_to_mb(max_low_pfn));/** Initialize the boot-time allocator (with low memory only):*/bootmap_size = init_bootmem(start_pfn, max_low_pfn);register_bootmem_low_pages(max_low_pfn);/** Reserve the bootmem bitmap itself as well. We do this in two* steps (first step was init_bootmem()) because this catches* the (very unlikely) case of us accidentally initializing the* bootmem allocator with an invalid RAM area.*/reserve_bootmem(HIGH_MEMORY, (PFN_PHYS(start_pfn) +bootmap_size + PAGE_SIZE-1) - (HIGH_MEMORY));/** reserve physical page 0 - it's a special BIOS page on many boxes,* enabling clean reboots, SMP operation, laptop functions.*/reserve_bootmem(0, PAGE_SIZE);/* reserve EBDA region, it's a 4K region */reserve_ebda_region();/* could be an AMD 768MPX chipset. Reserve a page before VGA to preventPCI prefetch into it (errata #56). Usually the page is reserved anyways,unless you have no PS/2 mouse plugged in. */if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD &&boot_cpu_data.x86 == 6)reserve_bootmem(0xa0000 - 4096, 4096);#ifdef CONFIG_SMP/** But first pinch a few for the stack/trampoline stuff* FIXME: Don't need the extra page at 4K, but need to fix* trampoline before removing it. (see the GDT stuff)*/reserve_bootmem(PAGE_SIZE, PAGE_SIZE);
#endif
#ifdef CONFIG_ACPI_SLEEP/** Reserve low memory region for sleep support.*/acpi_reserve_bootmem();
#endif
#ifdef CONFIG_X86_FIND_SMP_CONFIG/** Find and reserve possible boot-time SMP configuration:*/find_smp_config();
#endif#ifdef CONFIG_BLK_DEV_INITRDif (LOADER_TYPE && INITRD_START) {if (INITRD_START + INITRD_SIZE <= (max_low_pfn << PAGE_SHIFT)) {reserve_bootmem(INITRD_START, INITRD_SIZE);initrd_start =INITRD_START ? INITRD_START + PAGE_OFFSET : 0;initrd_end = initrd_start+INITRD_SIZE;}else {printk(KERN_ERR "initrd extends beyond end of memory ""(0x%08lx > 0x%08lx)\ndisabling initrd\n",INITRD_START + INITRD_SIZE,max_low_pfn << PAGE_SHIFT);initrd_start = 0;}}
#endifreturn max_low_pfn;
}
1. 代码详细解析
1.1. 变量声明和起始页框计算
unsigned long bootmap_size, start_pfn, max_low_pfn;/** partially used pages are not usable - thus* we are rounding upwards:*/
start_pfn = PFN_UP(init_pg_tables_end);
-
变量说明：
bootmap_size：bootmem分配器位图大小start_pfn：可用内存的起始页框号max_low_pfn：低端内存的最大页框号
-
PFN_UP(init_pg_tables_end)：init_pg_tables_end：页表初始化结束地址PFN_UP()：将地址向上取整到页边界（跳过部分使用的页）- 确保内存分配从完整的页开始
1.2. 内存范围探测
find_max_pfn();
max_low_pfn = find_max_low_pfn();
find_max_pfn()：探测系统最大物理内存大小，设置全局变量max_pfnfind_max_low_pfn()：探测低端内存（直接映射区域）的最大页框号- 低端内存通常是物理内存的前896MB，可以直接由内核线性映射
1.3. 高端内存配置
#ifdef CONFIG_HIGHMEM
highstart_pfn = highend_pfn = max_pfn;
if (max_pfn > max_low_pfn) {highstart_pfn = max_low_pfn;
}
printk(KERN_NOTICE "%ldMB HIGHMEM available.\n",pages_to_mb(highend_pfn - highstart_pfn));
#endif
printk(KERN_NOTICE "%ldMB LOWMEM available.\n",pages_to_mb(max_low_pfn));
- HIGHMEM配置：
highstart_pfn：高端内存起始页框号highend_pfn：高端内存结束页框号- 如果存在高端内存（
max_pfn > max_low_pfn），则设置高端内存范围
- 信息输出：打印可用低端内存和高端内存大小
1.4. Bootmem分配器初始化
bootmap_size = init_bootmem(start_pfn, max_low_pfn);
register_bootmem_low_pages(max_low_pfn);
-
init_bootmem(start_pfn, max_low_pfn)：- 初始化
bootmem分配器（早期内存分配器） - 参数：起始页框号，低端内存结束页框号
- 返回值：
bootmem位图所需大小
- 初始化
-
register_bootmem_low_pages(max_low_pfn)：- 将所有低端内存页面注册为可用状态
- 遍历所有物理页，标记为空闲
- 最终会调用函数
free_bootmem_core将bootmem的位图清0
1.5. 关键内存区域保留
1.5.1. 保留bootmem位图自身
reserve_bootmem(HIGH_MEMORY, (PFN_PHYS(start_pfn) +bootmap_size + PAGE_SIZE-1) - (HIGH_MEMORY));
- 目的：保护
bootmem分配器使用的位图不被分配 - 计算：从
HIGH_MEMORY到start_pfn + bootmap_size的区域 PFN_PHYS()：将页框号转换为物理地址
1.5.2. 保留物理页0
reserve_bootmem(0, PAGE_SIZE);
- 重要性：物理地址0是特殊BIOS页面
- 必须保留以防止内核使用这个关键区域
1.5.3. 保留EBDA区域
reserve_ebda_region();
- EBDA：扩展BIOS数据区域
- 大小：4KB区域
- 作用：存储BIOS相关数据
1.6. 硬件特定保留
AMD芯片组特殊处理
if (boot_cpu_data.x86_vendor == X86_VENDOR_AMD &&boot_cpu_data.x86 == 6)reserve_bootmem(0xa0000 - 4096, 4096);
- 目标：AMD 768MPX芯片组
- 问题：存在PCI预取错误（errata #56）
- 解决方案：在VGA内存前保留4KB页面
- 位置：0xa0000 - 4096（即0x9F000-0xA0000之间）
1.7. 系统功能内存保留
1.7.1. SMP支持保留
#ifdef CONFIG_SMP
reserve_bootmem(PAGE_SIZE, PAGE_SIZE);
#endif
- 目的：为SMP启动代码保留内存，保留物理地址的第1个页面
- 用途：
- SMP处理器启动时的临时栈
- 蹦床代码，从实模式切换到保护模式的代码
- GDT相关结构
1.7.2. ACPI睡眠支持
#ifdef CONFIG_ACPI_SLEEP
acpi_reserve_bootmem();
#endif
- 功能：为ACPI睡眠功能保留低内存区域
- 用途：系统休眠/唤醒时保存状态信息
1.7.3. SMP配置查找
#ifdef CONFIG_X86_FIND_SMP_CONFIG
find_smp_config();
#endif
- 目的：查找并保留可能的启动时SMP配置信息
- 方法：扫描已知的SMP配置表位置
1.8. 初始RAM磁盘处理
#ifdef CONFIG_BLK_DEV_INITRD
if (LOADER_TYPE && INITRD_START) {if (INITRD_START + INITRD_SIZE <= (max_low_pfn << PAGE_SHIFT)) {reserve_bootmem(INITRD_START, INITRD_SIZE);initrd_start = INITRD_START ? INITRD_START + PAGE_OFFSET : 0;initrd_end = initrd_start+INITRD_SIZE;}else {printk(KERN_ERR "initrd extends beyond end of memory ""(0x%08lx > 0x%08lx)\ndisabling initrd\n",INITRD_START + INITRD_SIZE,max_low_pfn << PAGE_SHIFT);initrd_start = 0;}
}
#endif
- 条件检查：
- 存在引导加载器类型
- INITRD起始地址有效
- 有效性验证：确保
initrd完全在低端内存内 - 成功情况：
- 保留
initrd内存区域 - 设置
initrd_start和initrd_end（转换为虚拟地址） - 失败情况：
initrd超出内存范围，禁用initrd
- 保留
1.9. 函数返回
return max_low_pfn;
- 返回值：低端内存的最大页框号
- 用途：为后续内存初始化提供关键信息
2. 函数功能总结
setup_memory() 是x86架构Linux内核启动过程中关键的早期内存初始化函数，主要完成：
- 内存探测 - 确定物理内存范围和布局
- 分配器初始化 - 建立
bootmem早期内存分配器 - 关键区域保护 - 保留系统正常运行必需的内存区域
- 硬件兼容处理 - 处理特定硬件的内存需求
- 功能模块准备 - 为SMP、ACPI、
initrd等预留内存
确定低端内存的最大页框号find_max_low_pfn
unsigned long __init find_max_low_pfn(void)
{unsigned long max_low_pfn;max_low_pfn = max_pfn;if (max_low_pfn > MAXMEM_PFN) {if (highmem_pages == -1)highmem_pages = max_pfn - MAXMEM_PFN;if (highmem_pages + MAXMEM_PFN < max_pfn)max_pfn = MAXMEM_PFN + highmem_pages;if (highmem_pages + MAXMEM_PFN > max_pfn) {printk("only %luMB highmem pages available, ignoring highmem size of %uMB.\n", pages_to_mb(max_pfn - MAXMEM_PFN), pages_to_mb(highmem_pages));highmem_pages = 0;}max_low_pfn = MAXMEM_PFN;
#ifndef CONFIG_HIGHMEM/* Maximum memory usable is what is directly addressable */printk(KERN_WARNING "Warning only %ldMB will be used.\n",MAXMEM>>20);if (max_pfn > MAX_NONPAE_PFN)printk(KERN_WARNING "Use a PAE enabled kernel.\n");elseprintk(KERN_WARNING "Use a HIGHMEM enabled kernel.\n");max_pfn = MAXMEM_PFN;
#else /* !CONFIG_HIGHMEM */
#ifndef CONFIG_X86_PAEif (max_pfn > MAX_NONPAE_PFN) {max_pfn = MAX_NONPAE_PFN;printk(KERN_WARNING "Warning only 4GB will be used.\n");printk(KERN_WARNING "Use a PAE enabled kernel.\n");}
#endif /* !CONFIG_X86_PAE */
#endif /* !CONFIG_HIGHMEM */} else {if (highmem_pages == -1)highmem_pages = 0;
#ifdef CONFIG_HIGHMEMif (highmem_pages >= max_pfn) {printk(KERN_ERR "highmem size specified (%uMB) is bigger than pages available (%luMB)!.\n", pages_to_mb(highmem_pages), pages_to_mb(max_pfn));highmem_pages = 0;}if (highmem_pages) {if (max_low_pfn-highmem_pages < 64*1024*1024/PAGE_SIZE){printk(KERN_ERR "highmem size %uMB results in smaller than 64MB lowmem, ignoring it.\n", pages_to_mb(highmem_pages)); highmem_pages = 0;}max_low_pfn -= highmem_pages;}
#elseif (highmem_pages)printk(KERN_ERR "ignoring highmem size on non-highmem kernel!\n");
#endif}return max_low_pfn;
}
1. 代码详细解析
1.1. 变量初始化和基础检查
unsigned long max_low_pfn;
max_low_pfn = max_pfn;
if (max_low_pfn > MAXMEM_PFN) {
max_low_pfn = max_pfn：初始假设所有内存都是低端内存MAXMEM_PFN：低端内存的最大页框号- 条件判断：检查是否物理内存超过了低端内存限制
1.2. 高端内存情况处理（物理内存 > 低端内存限制）
1.2.1 高端内存页数计算
if (highmem_pages == -1)highmem_pages = max_pfn - MAXMEM_PFN;
if (highmem_pages + MAXMEM_PFN < max_pfn)max_pfn = MAXMEM_PFN + highmem_pages;
if (highmem_pages + MAXMEM_PFN > max_pfn) {printk("only %luMB highmem pages available, ignoring highmem size of %uMB.\n", pages_to_mb(max_pfn - MAXMEM_PFN), pages_to_mb(highmem_pages));highmem_pages = 0;
}
max_low_pfn = MAXMEM_PFN;
highmem_pages == -1：如果未指定高端内存大小，自动计算- 自动计算：
max_pfn - MAXMEM_PFN= 总内存 - 低端内存 = 高端内存 - 边界检查：确保计算的高端内存不超过实际物理内存，如果超过，说明用户指定的
highmem_pages过大，忽略用户配置并置0 - 强制设置：
max_low_pfn = MAXMEM_PFN，低端内存固定为最大值
1.2.2 未配置HIGHMEM的情况
#ifndef CONFIG_HIGHMEM
/* Maximum memory usable is what is directly addressable */
printk(KERN_WARNING "Warning only %ldMB will be used.\n", MAXMEM>>20);
if (max_pfn > MAX_NONPAE_PFN)printk(KERN_WARNING "Use a PAE enabled kernel.\n");
elseprintk(KERN_WARNING "Use a HIGHMEM enabled kernel.\n");
max_pfn = MAXMEM_PFN;
- 内核编译时未启用HIGHMEM支持
- 只能使用直接映射的低端内存
- 限制内存：
max_pfn = MAXMEM_PFN
1.2.3 配置了HIGHMEM但未配置PAE的情况
#else /* !CONFIG_HIGHMEM */
#ifndef CONFIG_X86_PAE
if (max_pfn > MAX_NONPAE_PFN) {max_pfn = MAX_NONPAE_PFN;printk(KERN_WARNING "Warning only 4GB will be used.\n");printk(KERN_WARNING "Use a PAE enabled kernel.\n");
}
#endif /* !CONFIG_X86_PAE */
#endif /* !CONFIG_HIGHMEM */
MAX_NONPAE_PFN：非PAE系统最大支持4GB内存对应的页框号- 如果物理内存超过4GB，需要PAE（物理地址扩展）支持
- 否则限制为4GB
1.3. 纯低端内存情况处理（物理内存 ≤ 低端内存限制）
1.3.1 基础初始化
} else {if (highmem_pages == -1)highmem_pages = 0;
- 所有内存都在低端内存范围内
- 设置
highmem_pages = 0（没有高端内存）
1.3.2 配置了HIGHMEM的验证
#ifdef CONFIG_HIGHMEM
if (highmem_pages >= max_pfn) {printk(KERN_ERR "highmem size specified (%uMB) is bigger than pages available (%luMB)!.\n", pages_to_mb(highmem_pages), pages_to_mb(max_pfn));highmem_pages = 0;
}
if (highmem_pages) {if (max_low_pfn-highmem_pages < 64*1024*1024/PAGE_SIZE){printk(KERN_ERR "highmem size %uMB results in smaller than 64MB lowmem, ignoring it.\n", pages_to_mb(highmem_pages));highmem_pages = 0;}max_low_pfn -= highmem_pages;
}
合理性检查：
- 大小检查：高端内存不能超过总内存
- 低端内存保护：必须保留至少64MB低端内存
- 调整计算：
max_low_pfn -= highmem_pages
1.3.3 未配置HIGHMEM的警告
#else
if (highmem_pages)printk(KERN_ERR "ignoring highmem size on non-highmem kernel!\n");
#endif
- 内核不支持HIGHMEM但用户指定了
highmem参数 - 忽略设置并警告用户
1.4. 函数返回
return max_low_pfn;
返回最终计算的低端内存最大页框号