Trace analysis of the boot process of the Linux kernel

Trace Analysis The boot process of the Linux kernel uses GDB to trace the debug kernel

• Using QEMU

  qemu -kernel linux-3.18.6 /arch/x86/boot/bzImage -initrd rootfs.img -s -S

  Parameters:

  • -S: Freezes the CPU at initialization
  • -S: assigning 1234 ports to GDB

• GDB debugging

  Open the shell.

  gdb(gdb) file linux-3.18.6/vmlinux  #在 gdb 界面中 target remote之前加载符号表(gdb) target remote :1234   #建立连接(gdb) break start_kernel    #设置断点

Starting from Start_kernel to Init process

• Set_task_stack_end_magic ()

  To detect Stack Overflow

• SMP_SETUP_PROCESSOR_ID ()

  Setting up a symmetric multi-processor

• Cgroup_init_early ()

  Initialize Control Groups

  *control Groups provide a mechanism for aggregating/partitioning sets of
  Tasks, and all their The future children, to hierarchical groups with
  Specialized behaviour.*

• Page_address_init ()

  Page address initialization (part of memory management)

• Setup_arch ()

  Setup_arch-architecture-specific boot-time initializations

• Build_all_zonelists ()

  Called with Zonelists_mutex held always unless system_state = = system_booting.

• Page_alloc_init ()

• Setup_log_buf ()

  Initialize the log buffer (KERNEL/PRINTK/PRINTK.C)

• Pidhash_init ()

  Initializing a PID hash table

  The PID hash table is scaled according to the amount of memory in the machine. From a minimum of slots up to 4096 slots at one gigabyte or more.

• Vfs_caches_init_early ()

• Sort_main_extable ()

  Sort the kernel ' s built-in exception table

• Trap_init ()

  Initializing interrupt vectors

• Mm_init ()

  Memory Management Initialization

• Sched_init ()

  Scheduling Service Initialization

• ......

• Rest_init ()

  Remaining initialization

  • Kernel_init:init process
  • Kthreadd: Kernel thread
  • Cpu_idle process: The code keeps looping, and if there are no executable processes in the system, the idle process is executed

Summarize

• When Start_kernel executes to the last part, a new kernel_init process is created in Rest_init, the kernel_thread(kernel_init, NULL, CLONE_FS); init process is the number 1th process in the system, the ancestor of all subsequent user-state processes, and the new Kthreadd process, pid = kernel_thread(kthreadd, NULL, CLONE_FS | CLONE_FILES); Kthreadd as the number 2nd process, is the ancestor of all kernel threads, in cpu_startup_entry(CPUHP_ONLINE) , is a while（1） loop that always has an idle process executing, and if the system does not always have any processes that can be executed, the idle process executes.

Last quote from the Monensin teacher's passage:

Daosh One: (Start_kernel)

Life Two: (Kernel_init and Kthreadd)

Two students three: (that is, the 0,1,2 process)

Sansheng: (Process 1th is all user-state process ancestor, process 2nd is all kernel thread ancestor)

Copyright NOTICE: This article for Bo Master original article, without Bo Master permission not reproduced.

Trace analysis of the boot process of the Linux kernel