Linux应用程序的启动流程

目录

代码

用strace查看调用流程

分析

总结

---

看到一篇好文：摘抄记录--Linux应用程序 启动流程-BugMan-ChinaUnix博客

• 代码

#include <stdio.h>

int main (int argc, char *argv[])
{
    printf ("Hello World\n");

    return 0;
}

• 用strace查看调用流程

pc123@ubuntu:~/Public$ strace ./main
execve("./main", ["./main"], 0x7fff3906f3d0 /* 56 vars */) = 0
brk(NULL)                               = 0x55d0fc755000
access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or directory)
access("/etc/ld.so.preload", R_OK)      = -1 ENOENT (No such file or directory)
openat(AT_FDCWD, "/etc/ld.so.cache", O_RDONLY|O_CLOEXEC) = 3
fstat(3, {st_mode=S_IFREG|0644, st_size=81045, ...}) = 0
mmap(NULL, 81045, PROT_READ, MAP_PRIVATE, 3, 0) = 0x7fe65b878000
close(3)                                = 0
access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or directory)
openat(AT_FDCWD, "/lib/x86_64-linux-gnu/libc.so.6", O_RDONLY|O_CLOEXEC) = 3
read(3, "\177ELF\2\1\1\3\0\0\0\0\0\0\0\0\3\0>\0\1\0\0\0\240\35\2\0\0\0\0\0"..., 832) = 832
fstat(3, {st_mode=S_IFREG|0755, st_size=2030928, ...}) = 0
mmap(NULL, 8192, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x7fe65b876000
mmap(NULL, 4131552, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x7fe65b272000
mprotect(0x7fe65b459000, 2097152, PROT_NONE) = 0
mmap(0x7fe65b659000, 24576, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x1e7000) = 0x7fe65b659000
mmap(0x7fe65b65f000, 15072, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x7fe65b65f000
close(3)                                = 0
arch_prctl(ARCH_SET_FS, 0x7fe65b8774c0) = 0
mprotect(0x7fe65b659000, 16384, PROT_READ) = 0
mprotect(0x55d0fa83f000, 4096, PROT_READ) = 0
mprotect(0x7fe65b88c000, 4096, PROT_READ) = 0
munmap(0x7fe65b878000, 81045)           = 0
fstat(1, {st_mode=S_IFCHR|0620, st_rdev=makedev(136, 14), ...}) = 0
brk(NULL)                               = 0x55d0fc755000
brk(0x55d0fc776000)                     = 0x55d0fc776000
write(1, "hello world", 11hello world)             = 11
exit_group(0)                           = ?
+++ exited with 0 +++
pc123@ubuntu:~/Public$

• 分析

1. 毫无疑问，在shell内执行一个程序main，本质上是shell去调用execve函数执行main程序
2. execve是一个系统调用，Linux内核会在这个系统调用里面为main程序映射必要的内存
3. 动态解释器【动态解释器】：当execve将控制权给/lib64/ld-linux-x86-64.so.2的时候，这个文件进行动态库代码重定向等功能

   $ readelf -l hello | grep interpreter
         [Requesting program interpreter: /lib64/ld-linux-x86-64.so.2]

4. 通过读取elf头部信息中的entry部分，即可得到程序入口地址为0x540

   pc123@ubuntu:~/Public$ readelf -h main
   ELF Header:
     Magic:   7f 45 4c 46 02 01 01 00 00 00 00 00 00 00 00 00 
     Class:                             ELF64
     Data:                              2's complement, little endian
     Version:                           1 (current)
     OS/ABI:                            UNIX - System V
     ABI Version:                       0
     Type:                              DYN (Shared object file)
     Machine:                           Advanced Micro Devices X86-64
     Version:                           0x1
     Entry point address:               0x540
     Start of program headers:          64 (bytes into file)
     Start of section headers:          6448 (bytes into file)
     Flags:                             0x0
     Size of this header:               64 (bytes)
     Size of program headers:           56 (bytes)
     Number of program headers:         9
     Size of section headers:           64 (bytes)
     Number of section headers:         29
     Section header string table index: 28
   pc123@ubuntu:~/Public$ 
   

5. 通过objdump找到程序的入口地址：

   pc123@ubuntu:~/Public$ objdump -axd main > main.s

   在main.s文件找到地址为0x540的代码

   Disassembly of section .text:       段名：.text，表示代码段
   
   0000000000000540 <_start>:        函数名_start,存放地址0x540
    540:	31 ed                	xor    %ebp,%ebp
    542:	49 89 d1             	mov    %rdx,%r9
    545:	5e                   	pop    %rsi
    546:	48 89 e2             	mov    %rsp,%rdx
    549:	48 83 e4 f0          	and    $0xfffffffffffffff0,%rsp
    54d:	50                   	push   %rax
    54e:	54                   	push   %rsp
    54f:	4c 8d 05 8a 01 00 00 	lea    0x18a(%rip),%r8        # 6e0 <__libc_csu_fini>
    556:	48 8d 0d 13 01 00 00 	lea    0x113(%rip),%rcx        # 670 <__libc_csu_init>
    55d:	48 8d 3d e6 00 00 00 	lea    0xe6(%rip),%rdi        # 64a <main>
    564:	ff 15 76 0a 20 00    	callq  *0x200a76(%rip)        # 200fe0 <__libc_start_main@GLIBC_2.2.5>
    56a:	f4                   	hlt    
    56b:	0f 1f 44 00 00       	nopl   0x0(%rax,%rax,1)
   

   可见在shell执行main程序，程序的入口是_start函数，这个_start函数实际上是由C库实现的，

6. _start函数最后调用了__libc_start_main函数【汇编的callq *0x200a76这一行】，而调用__libc_start_main函数传入了0x54f, 0x556, 0x55d对应的这三个参数，其中对应的就是__libc_csu_fini， __libc_csu_initial， main函数【也就是我们写的代码】 

7. 【__libc_start_main@plt表示这是一个延迟加载函数，什么是延迟加载函数呢？延迟加载函数就是指在动态解释阶段不进行代码重定位，只有在真正使用该函数的时候，才去定位该函数的地址， 这样做的目的是加快程序启动】

8. __libc_start_main函数的运行的顺序为:__libc_csu_init->main->__libc_csu_fini，__libc_csu_init叫构造函数，__libc_csu_fini叫析构函数，在代码中可通过__attribute__ ((constructor))标记构造函数，__attribute__ ((destructor))来标记析构函数                      有这样的例子

   #include <stdio.h>
   
   static void hello_after() __attribute__ ((destructor));
   static void hello_before() __attribute__ ((constructor));
   
   static void hello_before(void)
   {
       printf("Before main\n");
   }
   
   static void hello_after(void)
   {
       printf("After main\n");
   }
   
   int main (int argc, char *argv[])
   {
       printf ("Hello World\n");
   
       return 0;
   }

   pc123@ubuntu:~/Public$ ./main 
   Before main
   Hello World
   After main
   pc123@ubuntu:~/Public$ 

• 总结

shell通过execve启动main程序------->main函数被C 库作为一个函数----------->整个程序的入口地址是_start【通过objdump对应的0x540确定】----->start函数调用了__libc_start_main，传入了构造，main, 析构函数。