Linux内核启动流程（汇编+C源码分析），超级详细！

Linux内核构成

本文讲解的内核版本是2.6，arm平台的启动过程。因为arm在嵌入式中确实具有统治地位的。闲话少说开始吧，因为真的很多内容，但是童鞋们一定要耐心看完。

1 arch/arm/boot/compressed/Makefile arch/arm/boot/compressed

2. arch/arm/kernel

Linux内核启动流程

arch/arm/boot/compressed

Start:

.type start,#function

.rept 8

mov r0, r0

.endr

b 1f

.word 0x016f2818 @ Magic numbers to help the loader

.word start @ absolute load/run zImage address

.word _edata @ zImage end address

1: mov r7, r1 @ save architecture ID

mov r8, r2 @ save atags pointer

这也标志着u-boot将系统完全的交给了OS，bootloader生命终止。之后代码在133行会读取cpsr并判断是否处理器处于supervisor模式——从u-boot进入kernel，系统已经处于SVC32模式；而利用angel进入则处于user模式，还需要额外两条指令。之后是再次确认中断关闭，并完成cpsr写入

mrs r2, cpsr @ get current mode

tst r2, #3 @ not user?

bne not_angel

mov r0, #0x17 @ angel_SWIreason_EnterSVC

swi 0x123456 @ angel_SWI_ARM

not_angel:

mrs r2, cpsr @ turn off interrupts to

orr r2, r2, #0xc0 @ prevent angel from running

msr cpsr_c, r2

然后在LC0地址处将分段信息导入r0-r6、ip、sp等寄存器，并检查代码是否运行在与链接时相同的目标地址，以决定是否进行处理。由于现在很少有人不使用loader和tags，将zImage烧写到rom直接从0x0位置执行，所以这个处理是必须的（但是zImage的头现在也保留了不用loader也可启动的能力）。arm架构下自解压头一般是链接在0x0地址而被加载到0x30008000运行，所以要修正这个变化。涉及到

• r5寄存器存放的zImage基地址
• r6和r12（即ip寄存器）存放的got（global offset table）
• r2和r3存放的bss段起止地址
• sp栈指针地址

很简单，这些寄存器统统被加上一个你也能猜到的偏移地址 0x30008000。该地址是s3c2410相关的，其他的ARM处理器可以参考下表

PXA2xx是0xa0008000

IXP2x00和IXP4xx是0x00008000

Freescale i.MX31/37是0x80008000

TI davinci DM64xx是0x80008000

TI omap系列是0x80008000

AT91RM/SAM92xx系列是0x20008000

Cirrus EP93xx是0x00008000

这些操作发生在代码172行开始的地方，下面只粘贴一部分

add r5, r5, r0

add r6, r6, r0

add ip, ip, r0

后面在211行进行bss段的清零工作

not_relocated: mov r0, #0

1: str r0, [r2], #4 @ clear bss

str r0, [r2], #4

str r0, [r2], #4

str r0, [r2], #4

cmp r2, r3

blo 1b

然后224行，打开cache，并为后面解压缩设置64KB的临时malloc空间

bl cache_on

mov r1, sp @ malloc space above stack

add r2, sp, #0x10000 @ 64k max 接下来238行进行检查，确定内核解压缩后的Image目标地址是否会覆盖到zImage头，如果是则准备将zImage头转移到解压出来的内核后面

cmp r4, r2

bhs wont_overwrite

sub r3, sp, r5 @ > compressed kernel size

add r0, r4, r3, lsl #2 @ allow for 4x expansion

cmp r0, r5

bls wont_overwrite

mov r5, r2 @ decompress after malloc space

mov r0, r5

mov r3, r7

bl decompress_kernel

真实情况——在大多数的应用中，内核编译都会把压缩的zImage和非压缩的Image链接到同样的地址。这样做的好处是，人们不用关心内核是Image还是zImage，放到这个位置执行就OK，所以在解压缩后zImage头必须为真正的内核让路。

在250行解压完毕，内核长度返回值存放在r0寄存器里。在内核末尾空出128字节的栈空间用，并且使其长度128字节对齐。

add r0, r0, #127 + 128 @ alignment + stack

bic r0, r0, #127 @ align the kernel length

算出搬移代码的参数：计算内核末尾地址并存放于r1寄存器，需要搬移代码原来地址放在r2，需要搬移的长度放在r3。然后执行搬移，并设置好sp指针指向新的栈（原来的栈也会被内核覆盖掉）

add r1, r5, r0 @ end of decompressed kernel

adr r2, reloc_start

ldr r3, LC1

add r3, r2, r3

1: ldmia r2!, {r9 - r14} @ copy relocation code

stmia r1!, {r9 - r14}

ldmia r2!, {r9 - r14}

stmia r1!, {r9 - r14}

cmp r2, r3

blo 1b

add sp, r1, #128 @ relocate the stack

搬移完成后刷新cache，因为代码地址变化了不能让cache再命中被内核覆盖的老地址。然后跳转到新的地址继续执行

bl cache_clean_flush

add pc, r5, r0 @ call relocation code

注意——zImage在解压后的搬移和跳转会给gdb调试内核带来麻烦。因为用来调试的符号表是在编译是生成的，并不知道以后会被搬移到何处去，只有在内核解压缩完成之后，根据计算出来的参数“告诉”调试器这个变化。以撰写本文时使用的zImage为例，内核自解压头重定向后，reloc_start地址由0x30008360变为0x30533e60。故我们要把vmlinux的符号表也相应的从0x30008000后移到0x30533b00开始，这样gdb就可以正确的对应源代码和机器指令。

随着头部代码移动到新的位置，不会再和内核的目标地址冲突，可以开始内核自身的搬移了。此时r0寄存器存放的是内核长度（严格的说是长度外加128Byte的栈），r4存放的是内核的目的地址0x30008000，r5是目前内核存放地址，r6是CPU ID，r7是machine ID，r8是atags地址。代码从501行开始

reloc_start: add r9, r5, r0

sub r9, r9, #128 @ do not copy the stack

debug_reloc_start

mov r1, r4

1:

.rept 4

ldmia r5!, {r0, r2, r3, r10 - r14} @ relocate kernel

stmia r1!, {r0, r2, r3, r10 - r14}

.endr

cmp r5, r9

blo 1b

add sp, r1, #128 @ relocate the stack

接下来在516行清除并关闭cache，清零r0，将machine ID存入r1，atags指针存入r2，再跳入0x30008000执行真正的内核Image

call_kernel: bl cache_clean_flush

bl cache_off

mov r0, #0 @ must be zero

mov r1, r7 @ restore architecture number

mov r2, r8 @ restore atags pointer

mov pc, r4 @ call kernel

内核代码入口在arch/arm/kernel文件的83行。首先进入SVC32模式，并查询CPU ID，检查合法性

msr cpsr_c, #PSR_F_BIT | PSR_I_BIT | SVC_MODE @ ensure svc mode

@ and irqs disabled

mrc p15, 0, r9, c0, c0 @ get processor id

bl __lookup_processor_type @ r5=procinfo r9=cpuid

movs r10, r5 @ invalid processor (r5=0)?

beq __error_p @ yes, error 'p'

接着在87行进一步查询machine ID并检查合法性

bl __lookup_machine_type @ r5=machinfo

movs r8, r5 @ invalid machine (r5=0)?

beq __error_a @ yes, error 'a'

其中__lookup_processor_type在linux-2.6.24-moko-linuxbj/arch/arm/kernel文件的149行，该函数首将标号3的实际地址加载到r3，然后将编译时生成的__proc_info_begin虚拟地址载入到r5，__proc_info_end虚拟地址载入到r6，标号3的虚拟地址载入到r7。由于adr伪指令和标号3的使用，以及__proc_info_begin等符号在linux-2.6.24-moko-linuxbj/arch/arm/kernel而不是代码中被定义，此处代码不是非常直观，想弄清楚代码缘由的读者请耐心阅读这两个文件和adr伪指令的说明。

r3和r7分别存储的是同一位置标号3的物理地址（由于没有启用mmu，所以当前肯定是物理地址）和虚拟地址，所以儿者相减即得到虚拟地址和物理地址之间的offset。利用此offset，将r5和r6中保存的虚拟地址转变为物理地址

__lookup_processor_type:

adr r3, 3f

ldmda r3, {r5 - r7}

sub r3, r3, r7 @ get offset between virt&phys

add r5, r5, r3 @ convert virt addresses to

add r6, r6, r3 @ physical address space

然后从proc_info中读出内核编译时写入的processor ID和之前从cpsr中读到的processor ID对比，查看代码和CPU硬件是否匹配（想在arm920t上运行为cortex-a8编译的内核？不让！）。如果编译了多种处理器支持，如versatile板，则会循环每种type依次检验，如果硬件读出的ID在内核中找不到匹配，则r5置0返回

1: ldmia r5, {r3, r4} @ value, mask

and r4, r4, r9 @ mask wanted bits

teq r3, r4

beq 2f

add r5, r5, #PROC_INFO_SZ @ sizeof(proc_info_list)

cmp r5, r6

blo 1b

mov r5, #0 @ unknown processor

2: mov pc, lr

__lookup_machine_type在linux-2.6.24-moko-linuxbj/arch/arm/kernel文件的197行，编码方法与检查processor ID完全一样，请参考前段

__lookup_machine_type:

adr r3, 3b

ldmia r3, {r4, r5, r6}

sub r3, r3, r4 @ get offset between virt&phys

add r5, r5, r3 @ convert virt addresses to

add r6, r6, r3 @ physical address space

1: ldr r3, [r5, #MACHINFO_TYPE] @ get machine type

teq r3, r1 @ matches loader number?

beq 2f @ found

add r5, r5, #SIZEOF_MACHINE_DESC @ next machine_desc

cmp r5, r6

blo 1b

mov r5, #0 @ unknown machine

2: mov pc, lr

代码回到第92行，检查atags合法性，然后创建初始页表

bl __vet_atags

bl __create_page_tables

创建页表的代码在218行，首先将内核起始地址-0x4000到内核起始地址之间的16K存储器清0

__create_page_tables:

pgtbl r4 @ page table address

/*

* Clear the 16K level 1 swapper page table

*/

mov r0, r4

mov r3, #0

add r6, r0, #0x4000

1: str r3, [r0], #4

str r3, [r0], #4

str r3, [r0], #4

str r3, [r0], #4

teq r0, r6

bne 1b

然后在234行将proc_info中的mmu_flags加载到r7

ldr r7, [r10, #PROCINFO_MM_MMUFLAGS] @ mm_mmuflags在242行将PC指针右移20位，得到内核第一个1MB空间的段地址存入r6，在s3c2410平台该值是0x300。接着根据此值存入映射标识

mov r6, pc, lsr #20 @ start of kernel section

orr r3, r7, r6, lsl #20 @ flags + kernel base

str r3, [r4, r6, lsl #2] @ identity mapping

完成页表设置后回到102行，为打开虚拟地址映射作准备。设置sp指针，函数返回地址lr指向__enable_mmu，并跳转到linux-2.6.24-moko-linuxbj/arch/arm/mm的386行，清除I-cache、D-cache、write buffer和TLB

__arm920_setup:

mov r0, #0

mcr p15, 0, r0, c7, c7 @ invalidate I,D caches on v4

mcr p15, 0, r0, c7, c10, 4 @ drain write buffer on v4

#ifdef CONFIG_MMU

mcr p15, 0, r0, c8, c7 @ invalidate I,D TLBs on v4

#endif然后返回的158行，加载domain和页表，跳转到__turn_mmu_on

__enable_mmu:

#ifdef CONFIG_ALIGNMENT_TRAP

orr r0, r0, #CR_A

#else

bic r0, r0, #CR_A

#endif

#ifdef CONFIG_CPU_DCACHE_DISABLE

bic r0, r0, #CR_C

#endif

#ifdef CONFIG_CPU_BPREDICT_DISABLE

bic r0, r0, #CR_Z

#endif

#ifdef CONFIG_CPU_ICACHE_DISABLE

bic r0, r0, #CR_I

#endif

mov r5, #(domain_val(DOMAIN_USER, DOMAIN_MANAGER) | \

domain_val(DOMAIN_KERNEL, DOMAIN_MANAGER) | \

domain_val(DOMAIN_TABLE, DOMAIN_MANAGER) | \

domain_val(DOMAIN_IO, DOMAIN_CLIENT))

mcr p15, 0, r5, c3, c0, 0 @ load domain access register

mcr p15, 0, r4, c2, c0, 0 @ load page table pointer

b __turn_mmu_on在194行把mmu使能位写入mmu，激活虚拟地址。然后将原来保存在sp中的地址载入pc，跳转到的__mmap_switched，至此代码进入虚拟地址的世界

mov r0, r0

mcr p15, 0, r0, c1, c0, 0 @ write control reg

mrc p15, 0, r3, c0, c0, 0 @ read id reg

mov r3, r3

mov r3, r3

mov pc, r13

在的37行开始清除内核bss段，processor ID保存在r9，machine ID报存在r1，atags地址保存在r2，并将控制寄存器保存到r7定义的内存地址。接下来跳入linux-2.6.24-moko-linuxbj/ini的507行，start_kernel函数。这里只粘贴部分代码

__mmap_switched:

adr r3, __switch_data + 4

ldmia r3!, {r4, r5, r6, r7}

cmp r4, r5 @ Copy data segment if needed

1: cmpne r5, r6

ldrne fp, [r4], #4

strne fp, [r5], #4

bne 1b

asmlinkage void __init start_kernel(void)

{

char * command_line;

extern struct kernel_param __start___param[], __stop___param[];

smp_setup_processor_id();

/*

* Need to run as early as possible, to initialize the

* lockdep hash:

*/

lockdep_init();

debug_objects_early_init();

cgroup_init_early();

local_irq_disable();

early_boot_irqs_off();

early_init_irq_lock_class();

/*

* Interrupts are still disabled. Do necessary setups, then

* enable them

*/

lock_kernel();

tick_init();

boot_cpu_init();

page_address_init();

printk(KERN_NOTICE);

printk(linux_banner);

setup_arch(&command_line);

mm_init_owner(&init_mm, &init_task);

setup_command_line(command_line);

setup_per_cpu_areas();

setup_nr_cpu_ids();

smp_prepare_boot_cpu(); /* arch-specific boot-cpu hooks */

/*

* Set up the scheduler prior starting any interrupts (such as the

* timer interrupt). Full topology setup happens at smp_init()

* time - but meanwhile we still have a functioning scheduler.

*/

sched_init();

/*

* Disable preemption - early bootup scheduling is extremely

* fragile until we cpu_idle() for the first time.

*/

preempt_disable();

build_all_zonelists();

page_alloc_init();

printk(KERN_NOTICE "Kernel command line: %s\n", boot_command_line);

parse_early_param();

parse_args("Booting kernel", static_command_line, __start___param,

__stop___param - __start___param,

&unknown_bootoption);

if (!irqs_disabled()) {

printk(KERN_WARNING "start_kernel(): bug: interrupts were "

"enabled *very* early, fixing it\n");

local_irq_disable();

}

sort_main_extable();

trap_init();

rcu_init();

/* init some links before init_ISA_irqs() */

early_irq_init();

init_IRQ();

pidhash_init();

init_timers();

hrtimers_init();

softirq_init();

timekeeping_init();

time_init();

sched_clock_init();

profile_init();

if (!irqs_disabled())

printk(KERN_CRIT "start_kernel(): bug: interrupts were "

"enabled early\n");

early_boot_irqs_on();

local_irq_enable();

/*

* HACK ALERT! This is early. We're enabling the console before

* we've done PCI setups etc, and console_init() must be aware of

* this. But we do want output early, in case something goes wrong.

*/

console_init();

if (panic_later)

panic(panic_later, panic_param);

lockdep_info();

/*

* Need to run this when irqs are enabled, because it wants

* to self-test [hard/soft]-irqs on/off lock inversion bugs

* too:

*/

locking_selftest();

#ifdef CONFIG_BLK_DEV_INITRD

if (initrd_start && !initrd_below_start_ok &&

page_to_pfn(virt_to_page((void *)initrd_start)) < min_low_pfn) {

printk(KERN_CRIT "initrd overwritten (0x%08lx < 0x%08lx) - "

"disabling it.\n",

page_to_pfn(virt_to_page((void *)initrd_start)),

min_low_pfn);

initrd_start = 0;

}

#endif

vmalloc_init();

vfs_caches_init_early();

cpuset_init_early();

page_cgroup_init();

mem_init();

enable_debug_pagealloc();

cpu_hotplug_init();

kmem_cache_init();

debug_objects_mem_init();

idr_init_cache();

setup_per_cpu_pageset();

numa_policy_init();

if (late_time_init)

late_time_init();

calibrate_delay();

pidmap_init();

pgtable_cache_init();

prio_tree_init();

anon_vma_init();

#ifdef CONFIG_X86

if (efi_enabled)

efi_enter_virtual_mode();

#endif

thread_info_cache_init();

cred_init();

fork_init(num_physpages);

proc_caches_init();

buffer_init();

key_init();

security_init();

vfs_caches_init(num_physpages);

radix_tree_init();

signals_init();

/* rootfs populating might need page-writeback */

page_writeback_init();

#ifdef CONFIG_PROC_FS

proc_root_init();

#endif

cgroup_init();

cpuset_init();

taskstats_init_early();

delayacct_init();

check_bugs();

acpi_early_init(); /* before LAPIC and SMP init */

ftrace_init();

/* Do the rest non-__init'ed, we're now alive */

rest_init();

}

tatic noinline void __init_refok rest_init(void)

__releases(kernel_lock)

{

int pid;

kernel_thread(kernel_init, NULL, CLONE_FS | CLONE_SIGHAND);

numa_default_policy();

pid = kernel_thread(kthreadd, NULL, CLONE_FS | CLONE_FILES);

kthreadd_task = find_task_by_pid_ns(pid, &init_pid_ns);

unlock_kernel();

/*

* The boot idle thread must execute schedule()

* at least once to get things moving:

*/

init_idle_bootup_task(current);

rcu_scheduler_starting();

preempt_enable_no_resched();

schedule();

preempt_disable();

/* Call into cpu_idle with preempt disabled */

cpu_idle();

}

static noinline int init_post(void)

{

/* need to finish all async __init code before freeing the memory */

async_synchronize_full();

free_initmem();

unlock_kernel();

mark_rodata_ro();

system_state = SYSTEM_RUNNING;

numa_default_policy();

if (sys_open((const char __user *) "/dev/console", O_RDWR, 0) < 0)

printk(KERN_WARNING "Warning: unable to open an initial console.\n");

(void) sys_dup(0);

(void) sys_dup(0);

current->signal->flags |= SIGNAL_UNKILLABLE;

if (ramdisk_execute_command) {

run_init_process(ramdisk_execute_command);

printk(KERN_WARNING "Failed to execute %s\n",

ramdisk_execute_command);

}

/*

* We try each of these until one succeeds.

*

* The Bourne shell can be used instead of init if we are

* trying to recover a really broken machine.

*/

if (execute_command) {

run_init_process(execute_command);

printk(KERN_WARNING "Failed to execute %s. Attempting "

"defaults...\n", execute_command);

}

run_init_process("/sbin/init");

run_init_process("/etc/init");

run_init_process("/bin/init");

run_init_process("/bin/sh");

panic("No init found. Try passing init= option to kernel.");

}

好了，整个流程结束了。看不懂也没关系。因为在头条上代码排版真的难看懂。

喜欢我的文章的话，就关注我吧！在本头条号的置顶文章中有【文章分类】包含:

[C++进阶篇系列]

[高级网络编程篇系列]

[Linux系统篇系列]

[C++基础知识篇]

[协议篇系列]

[数据结构和算法系列]

[设计模式系列]

不要只收藏和转发哦，写文章其实很累的。