The Linux Kernel/System/Booting

Booting and halting Boot sequence Power-on self-test Firmware/POST – hardware initialization, device detection GRUB Bootloader (GRUB, systemd-boot, EFI stub admin-guide/efi-stub.html) – loads kernel image into memory Kernel – decompresses, initializes, starts 1 init Kernel booting This is loaded in two stages - in the first stage the kernel (as a compressed image file) is loaded into memory and decompressed, and a few fundamental functions such as essential hardware and basic memory management (memory paging) are set up. Control is then switched one final time to the main kernel start process calling start_kernel, which then performs the majority of system setup (interrupts, the rest of memory management, device and driver initialization, etc.) before spawning separately, the idle process and scheduler, and the init process (which is executed in user space). Kernel loading stage The kernel as loaded is typically an image file, compressed into either zImage or bzImage formats with zlib. A routine at the head of it does a minimal amount of hardware setup, decompresses the image fully into high memory, and takes note of any RAM disk if configured. It then executes kernel startup via startup_64 (for x86_64 architecture). arch/x86/boot – boot protocol, CPU feature detection (x86 Feature Flags arch/x86/cpuinfo.html) arch/x86/boot/compressed/vmlinux.lds.S - linker script defines entry startup_64 in arch/x86/boot/compressed/head_64.S - assembly of extractor extract_kernel - extractor in language C prints Decompressing Linux... done. Booting the kernel. Kernel startup stage The startup function for the kernel (also called the swapper or process 0) establishes memory management (paging tables and memory paging), detects the type of CPU and any additional functionality such as floating point capabilities, and then switches to non-architecture specific Linux kernel functionality via a call to start_kernel. ↯ Startup call hierarchy: arch/x86/kernel/vmlinux.lds.S – linker script arch/x86/kernel/head_64.S – assembly of uncompressed startup code arch/x86/kernel/head64.c – platform dependent startup: x86_64_start_kernel x86_64_start_reservations init/main.c – main initialization code init/init_task.c – initial task (PID 0, swapper), init_task init/do_mounts.c – root device mounting (mount_root, prepare_namespace) init/do_mounts_initrd.c – initrd (initial ramdisk) unpacking and pivot to real root init/do_mounts_rd.c – RAM disk image mounting (legacy rd= parameter) init/initramfs.c – built-in initramfs (cpio) unpacking (populate_rootfs) init/calibrate.c – BogoMIPS delay loop calibration (calibrate_delay) init/version.c – kernel version string and build timestamp (linux_banner) init/noinitramfs.c – default root mount when no initramfs is configured start_kernel 200 SLOC mm_init mem_init vmalloc_init sched_init rcu_init – Read-copy-update rest_init kernel_init - deferred kernel thread #1 kernel_init_freeable This and following functions are defined with attribute __init prepare_namespace initrd_load mount_root run_init_process obviously runs the first process 1 init kthreadd – deferred kernel thread #2 cpu_startup_entry do_idle 🛠️ Utilities 8 dracut – generate initramfs images; dracut --list-modules shows available modules 1 lsinitrd – list contents of an initramfs image start_kernel executes a wide range of initialization functions. It sets up interrupt handling (IRQs), further configures memory, starts the 1 init process (the first user-space process), and then starts the idle task via cpu_startup_entry. Notably, the kernel startup process also mounts the initial ramdisk (initrd) that was loaded previously as the temporary root file system during the boot phase. The initrd allows driver modules to be loaded directly from memory, without reliance upon other devices (e.g. a hard disk) and the drivers that are needed to access them (e.g. a SATA driver). This split of some drivers statically compiled into the kernel and other drivers loaded from initrd allows for a smaller kernel. The root file system is later switched via a call to 8 pivot_root / 2 pivot_root which unmounts the temporary root file system and replaces it with the use of the real one, once the latter is accessible. The memory used by the temporary root file system is then reclaimed. ⚙️ Internals arch/x86/Kconfig.debug – x86 debugging Kconfig options linux/smpboot.h kernel/smpboot.c – common SMP CPU bringup/teardown arch/x86/kernel/smpboot.c – x86 SMP booting linux/kallsyms.h, kernel/kallsyms.c – the /proc/kallsyms kernel symbol table ... 📖 References Article about booting of the kernel Initial RAM disk admin-guide/initrd.html initcall_debug, boot argument Linux startup process init Linux (U)EFI boot process The kernel’s command-line parameters admin-guide/kernel-parameters.html The EFI Boot Stub admin-guide/efi-stub.html Boot Configuration admin-guide/bootconfig.html Boot time memory management core-api/boot-time-mm.html Kernel booting process Kernel initialization process 📚 Further reading Boot-time tracing trace/boottime-trace.html E820 💾 Historical http://tldp.org/HOWTO/Linux-i386-Boot-Code-HOWTO/ http://www.tldp.org/LDP/lki/lki-1.html http://www.tldp.org/HOWTO/KernelAnalysis-HOWTO-4.html Halting or rebooting 🔧 TODO ⚲ API linux/reboot.h linux/stop_machine.h print_stop_info stop_machine stop_core_cpuslocked 2 reboot ↪ sys_reboot machine_restart or machine_halt or machine_power_off 2 kexec_load ↪ kernel/kexec.c – load a new kernel for later execution (kdump (Linux) kdump) reboot_mode linux/reboot-mode.h reboot_mode_driver devm_reboot_mode_register ⚙️ Internals kernel/reboot.c – reboot, halt and power-off kernel/stop_machine.c – stop all CPUs for safe code patching cpu_stopper cpu_stop_init cpu_stopper_thread – "migration" tasks arch/x86/kernel/reboot.c – x86 reboot methods The_Linux_Kernel/Softdog Driver Power management Keyword: suspend, alarm, hibernation. ⚲ API /sys/power/ /sys/kernel/debug/wakeup_sources ⌨️ hands-on: sudo awk '{gsub("^ ","?")} NR>1 {if ($6) {print $1}}' /sys/kernel/debug/wakeup_sources linux/pm.h linux dev_pm_ops linux/pm_qos.h linux/pm_clock.h linux/pm_domain.h linux/pm_wakeirq.h linux/pm_wakeup.h wakeup_source wakeup_source_register linux/suspend.h pm_suspend suspends the system linux/freezer.h – process freezer for suspend/hibernate Suspend and wakeup depend on 2 timer_create and 2 timerfd_create with clock ids CLOCK_REALTIME_ALARM or CLOCK_BOOTTIME_ALARM will wake the system if it is suspended. 2 epoll_ctl with flag EPOLLWAKEUP blocks suspend See also 7 capabilities CAP_WAKE_ALARM, CAP_BLOCK_SUSPEND ⚙️ Internals CONFIG_PM CONFIG_SUSPEND kernel/power – suspend, hibernate and PM core alarm_init kernel/time/alarmtimer.c – alarm timer for wakeup events drivers/base/power: wakeup_sources 📖 References PM administration admin-guide/pm CPU and Device PM driver-api/pm Power Management power sysfs power testing ABI admin-guide/abi-testing.html#file-testing-sysfs-power https://lwn.net/Kernel/Index/#Power_management PowerTOP cpupower tlp – apply laptop power management settings ACPI – Advanced Configuration and Power Interface Runtime PM Keywords: runtime power management, devices power management opportunistic suspend, autosuspend, autosleep. ⚲ API /sys/devices/.../power/: async autosuspend_delay_ms control runtime_active_kids runtime_active_time runtime_enabled runtime_status runtime_suspended_time runtime_usage linux/pm_runtime.h pm_runtime_mark_last_busy devm_pm_runtime_enable pm_runtime_enable pm_runtime_disable pm_runtime_get – asynchronous get pm_runtime_get_sync pm_runtime_resume_and_get – preferable synchronous get pm_runtime_put pm_runtime_put_noidle – just decrement usage counter pm_runtime_put_sync pm_runtime_put_autosuspend SET_RUNTIME_PM_OPS 👁 Example: ac97_pm ⚙️ Internals CONFIG_PM_AUTOSLEEP ... 📖 References Runtime Power Management Framework for I/O Devices power/runtime_pm.html CPU idle power saving methods for real-time workloads Sysfs devices PM API Power Management for USB Opportunistic suspend 📚 Further reading https://deepwiki.com/torvalds/linux/5-hardware-drivers