init function invocation of drivers compiled into kernel

by

The init routine of a built-in driver can still use the module_init() macro to declare that entry point. Or the driver can use device_initcall() when the driver would never be compiled as a loadable module. Or to move its initialization very early in the boot sequence, the driver could use subsys_initcall().

In include/linux/init.h the sequence for invoking these init routines is described as:

/* initcalls are now grouped by functionality into separate 
 * subsections. Ordering inside the subsections is determined
 * by link order. 
 * For backwards compatibility, initcall() puts the call in 
 * the device init subsection.
 *
 * The `id' arg to __define_initcall() is needed so that multiple initcalls
 * can point at the same handler without causing duplicate-symbol build errors.
 */

I assume that these subsections for device drivers correspond to the subdirectories within the drivers directory of the Linux kernel source tree, and that the link order is recorded in the built-in.o file of each subdirectory in drivers. So during kernel boot the init routine of each built-in driver is eventually executed by do_initcalls() in init/main.c.

The init routine of the device driver is responsible for probing the system to verify that the HW device actually exists. The driver should not allocate any resources or register any devices when the probe fails.

UPDATE:
Passing the option “initcall_debug” on the kernel command line will cause timing information to be printed to the console for each initcall. initcalls are used to initialize statically linked kernel drivers and subsystems and contribute a significant amount of time to the Linux boot process. The output looks like:

calling  tty_class_init+0x0/0x44 @ 1
initcall tty_class_init+0x0/0x44 returned 0 after 9765 usecs
calling  spi_init+0x0/0x90 @ 1
initcall spi_init+0x0/0x90 returned 0 after 9765 usecs

Reference: http://elinux.org/Initcall_Debug

More Related Contents:

  1. kernel stack and user space stack
  2. How to mmap a Linux kernel buffer to user space?
  3. Do Kernel pages get swapped out?
  4. Just black screen after running Qemu
  5. Linux Stack Sizes
  6. How can I monitor the NIC status(up/down) in a C program without polling the kernel?
  7. How do the likely/unlikely macros in the Linux kernel work and what is their benefit?
  8. Increasing the maximum number of TCP/IP connections in Linux
  9. Is there any API for determining the physical address from virtual address in Linux?
  10. What happens if there is no exit system call in an assembly program?
  11. Use of floating point in the Linux kernel
  12. Maximum number of processes in linux [closed]
  13. How to continuously monitor the directory using dnotify /inotify command
  14. Why can’t I mmap(MAP_FIXED) the highest virtual page in a 32-bit Linux process on a 64-bit kernel?
  15. Accurate calculation of CPU usage given in percentage in Linux?
  16. How to “hibernate” a process in Linux by storing its memory to disk and restoring it later?
  17. What is the overhead of a context-switch?
  18. What is the interface for ARM system calls and where is it defined in the Linux kernel?
  19. How does Linux determine the next PID?
  20. What is the difference between a Linux platform driver and normal device driver?
  21. Programmatically get parent pid of another process?
  22. How is the Linux kernel tested?
  23. How to avoid transparent_hugepage/defrag warning from mongodb?
  24. Converting jiffies to milli seconds
  25. “zero copy networking” vs “kernel bypass”?
  26. Direct Memory Access in Linux
  27. DNS caching in Linux [closed]
  28. Check the open FD limit for a given process in Linux
  29. Signed executables under Linux
  30. How to discover what Linux distribution is in use

Leave a Comment