基于 i386 体系结构的 Linux 启动代码分析

Embedded Operating Systems2 启动代码文件 linux/arch/i386/boot/bootsect.S linux/arch/i386/boot/setup.S linux/arch/i386/boot/compressed/head.S linux/arch/i386/kernel/head.S linux/arch/i386/kernel/init_task.c linux/arch/init/main.c

Operating Systems3 实模式 实模式是为了兼容早期的 CPU 而设置的 系统总是始于实模式 实模式下 地址总线： 20 位 内存范围： 0~1MB 逻辑地址 = 段地址 + 段内偏移 段地址 = 段寄存器中的值 *16 （或左移 4 位） 段寄存器长度： 16bit

Operating Systems4 I386 实模式下的内存布局图 RAM 1-MB ROM-BIOS VIDEO-BIOS VRAM 0xA0000 0xC0000 0xF0000 0x00000

Operating Systems5 软盘和硬盘启动 软盘启动， bootsect.S 0x7c00  0x x7c00, BIOS 0x90000, lilo 堆栈， 0x3ff4(0x ), 向下增长 磁盘参数表， 12Bytes ， 0x3ff4~0x4000 显示 “Loading” Setup  0x90200 系统 小内核， 0x10000 （ 64KB 处），低 装载 大内核， 0x （ 1MB 处），高 装载 setup 硬盘启动，两阶段引导 装载 LILO （ LInuxLOader ） 第一个扇区 … 装载 LINUX Bootsect.S  0x90000 Setup.S  0x90200 系统  0x10000  0x 跳转到 setup

Operating Systems6 Setup 初始化硬件设备并为内核程 序的执行建立环境 内存检测 键盘 视频 磁盘控制器 IBM 微通道总线 MCA PS/2 设备（总线鼠标） APM BIOS 若低装载，将系统移动到 0x1000 处（ 4KB 处） 否则，不必 临时 IDT 和临时 GDT FPU PIC, 16 个硬件中断  中断向 量 32~47 实模式  保护模式 Startup_32

Operating Systems7 Compressed/head.S Startup_32 初始化段寄存器和一个临时堆栈 初始化 BSS 段 解压缩 高装载或低装载  解压缩  0x （ 1MB ） 跳转到 0x 处

Operating Systems8 Head.S Startup_32 初始化段寄存器 建立进程 0 的内核堆栈 Setup_idt 拷贝系统参数 识别处理器 GDT 、 IDT Start_kernel

Operating Systems9 The system-segment registers In protected-mode the CPU needs quick access to some important data-structures, such as: Memory-Segment Descriptors Interrupt-Gate Descriptors Call-Gate Descriptors Task-State Descriptors Page-Directory and Page-Table Descriptors Special CPU registers locate these items

Operating Systems10 GDT and IDT Two most vital system registers for protected-mode execution are: GDTR (Global Descriptor Table Register) IDTR (Interrupt Descriptor Table Register) 48-bits wide base-address + segment-limit (an array of descriptors, the GDT & the IDT) Special instructions SGDT/LGDT, SIDT/LIDT Segment Base-Address Segment Limit bits 32 bits

Operating Systems11 System Relationships descriptor Interrupt Descriptor Table Global Descriptor Table GDTR IDTR

Operating Systems12 LDT and TSS For protected-mode multitasking, the CPU needs to access two other data-structures: The current Local Descriptor Table (LDT) The current Task-State Segment (TSS) Again, special registers tell the CPU where to find these data-structures in memory (assuming protected-mode is enabled) Instructions: SLDT/LLDT, STR/LTR

Operating Systems13 Control Registers CR0 CR1 CR2 CR3

Operating Systems14 CR0 CR0, MSW register (Machine Status Word, 32-bit version) Contains system control flags that control operating mode and states of the processor PE-bit (Protection Enabled) 0  CPU is in real-mode, 1  CPU is in protected-mode Instruction: lmsw LINUX’ setup.S: movw $1, %ax lmsw %ax jmp flush_instr // why? flush_instr: To turn on the PE-bit (enables protected-mode),

Operating Systems15 CR1 、 CR2 、 CR3 CR1 Reserved CR2 Contains the page-fault linear address (the linear address that caused a page fault) CR3 Contains the physical address of the base of the page directory and two flags (PCD & PWT)

Operating Systems16 Segment Descriptor Format Base[31..24]GD RSVRSV AVLAVL Limit [19..16] P DPLDPL SX C/DC/D R/WR/W ABase[23..16] Base[15..0]Limit[15..0]

Operating Systems17 “Hidden” part of Segment Registers selectorSegment baseSegment limit Access rights The programmer-visible part of a segment-register The “invisible” parts of a segment-register indexTIRPL Segment selector

Operating Systems18 Three VRAM zones GRAPHICS MONOCHROME TEXT COLOR TEXT 64-KB 32-KB 0xA0000 0xB0000 0xB8000

Operating Systems19 Array of picture-elements Text-mode VRAM is organized as an array Each array-element occupies one word Word’s LSB holds ascii character-code Word’s MSB holds a color-number pair bgcolorfgcolorASCII character-code byte nybble

Operating Systems20 Color-Attribute Byte Blink RGB Intense GB R foreground color attribute background color attribute

Operating Systems21 Screen-element locations

Operating Systems22 逻辑地址 段：偏移量 线性地址 32 位无符号整数， 0~4G-1, 即 0x ~0xFFFFFFFF 物理地址 芯片级内存单元寻址 地址转换过程 内存寻址 物理地址线性地址 分段单元分页单元 逻辑地址

Operating Systems23 实模式？保护模式？分页？ Real mode Exists mostly to maintain processor compatibility with older models, & To allow the OS to bootstrap

Operating Systems24 段选择子 indexTIRPL Segment selector

Operating Systems25 __KERNEL_CS 0x10 = b Index(0x2), TI(0), RPL(0, highest) limit(0xFFFFF), base(0), G(1, =4096), D(1,=32 位 ), P(1, in main memory), DPL(0, hignest), S(1), type(0xA) E(1), executable, code C(0), not current R(1), readable A(0), unaccessed gdt:.word 0, 0, 0, 0 # dummy.word 0, 0, 0, 0 # unused.word 0xFFFF # 4Gb.word 0 # base address = 0.word 0x9A00 # code read/exec.word 0x00CF # granularity(4096).word 0xFFFF # 4Gb.word 0 # base address = 0.word 0x9200 # data read/write.word 0x00CF # granularity(4096)

Operating Systems26 段描述符

Operating Systems27

Operating Systems28 逻辑地址的转换 Segment selector Logical Address: Segment-offset Operand’s effective address Physical Address: descriptor Segment Descriptor Table + Segment Base-address (also Segment-Limit and Access Rights) Validity is checked by CPU

Operating Systems29 分页 PG Paging (bit 31 of CR0). 1  Enables paging with PE=1,  Linear address space is divided into fixed-size pages (4KB, 2MB, or 4MB) that can be mapped into physical memory and/or disk storage using paging mechanism. 0  Disables paging, linear addresses = physical addresses.

Operating Systems30 lss lss a, %esp ; ds  es, a  esp