【Linux内核】启动流程——Kernel 启动流程梳理
本文转载自：Kernel 启动流程梳理
原作者：Li-Yongjun
内核生命周期
uboot 打印完 Starting kernel . . .，就完成了自己的使命，控制权便交给了 kernel，kernel 一开始的代码也是汇编（入口为 stext），主要完成验证是否支持此 CPU、验证 uboot 传入的设备树（dtb）合法性、使能 MMU 等工作，最终会调用 C 函数 start_kernel()
init/main.c
asmlinkage __visible void __init start_kernel(void)
{
...
rest_init();
}
start_kernel 相当于内核的 main 函数，内核的生命周期就是从执行这个函数的第一条语句开始的，直到最后一个函数 reset_init()，内核将不再从这个函数中返回，而是陷入这个函数里面的一个 while(1) 死循环，这个死循环被作为 idle 进程，也就是 0 号进程。
所以，内核的生命周期，就是一个完整的 start_kernel 函数。始于 start_kernel 函数的第一条语句，停留在最后的死循环。
init 进程
kernel 会创建众多内核线程，来持续致力于内存、磁盘、CPU 的管理，其中有两个内核线程比较重要，需要我们重点讲解，那就是 1 号内核线程 kernel_init 和 2 号内核线程 kthreadd。1 号内核线程最终会被用户的第一个进程 init 代替，也就成了 1 号进程。如下：
# ps
PID USER COMMAND
1 root init
2 root [kthreadd]
3 root [rcu_gp]
4 root [rcu_par_gp]
7 root [kworker/u4:0-ev]
8 root [mm_percpu_wq]
9 root [ksoftirqd/0]
...
COMMAND 这一列，带中括号的是内核线程，不带中括号的是用户进程。从 PID 统一编址就可以看出，它俩地位是一样的。
下面我们深入分析一下从 start_kernel 到最终运行 init 进程，kernel 都经历了什么。
打印
添加打印，是分析流程的好方法。
asmlinkage __visible void __init start_kernel(void)
{
char *command_line;
char *after_dashes;
set_task_stack_end_magic(&init_task);
smp_setup_processor_id();
debug_objects_early_init();
cgroup_init_early();
local_irq_disable();
early_boot_irqs_disabled = true;
/*
* Interrupts are still disabled. Do necessary setups, then
* enable them.
*/
boot_cpu_init();
page_address_init();
pr_notice("%s", linux_banner);
setup_arch(&command_line);
/*
* Set up the the initial canary and entropy after arch
* and after adding latent and command line entropy.
*/
add_latent_entropy();
add_device_randomness(command_line, strlen(command_line));
boot_init_stack_canary();
mm_init_cpumask(&init_mm);
setup_command_line(command_line);
setup_nr_cpu_ids();
setup_per_cpu_areas();
smp_prepare_boot_cpu(); /* arch-specific boot-cpu hooks */
boot_cpu_hotplug_init();
build_all_zonelists(NULL);
page_alloc_init();
pr_notice("Kernel command line: %s\n", boot_command_line);
parse_early_param();
after_dashes = parse_args("Booting kernel",
static_command_line, __start___param,
__stop___param - __start___param,
-1, -1, NULL, &unknown_bootoption);
if (!IS_ERR_OR_NULL(after_dashes))
parse_args("Setting init args", after_dashes, NULL, 0, -1, -1,
NULL, set_init_arg);
jump_label_init();
/*
* These use large bootmem allocations and must precede
* kmem_cache_init()
*/
setup_log_buf(0);
vfs_caches_init_early();
sort_main_extable();
trap_init();
mm_init();
ftrace_init();
/* trace_printk can be enabled here */
early_trace_init();
/*
* 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();
if (WARN(!irqs_disabled(),
"Interrupts were enabled *very* early, fixing it\n"))
local_irq_disable();
radix_tree_init();
/*
* Set up housekeeping before setting up workqueues to allow the unbound
* workqueue to take non-housekeeping into account.
*/
housekeeping_init();
/*
* Allow workqueue creation and work item queueing/cancelling
* early. Work item execution depends on kthreads and starts after
* workqueue_init().
*/
workqueue_init_early();
rcu_init();
/* Trace events are available after this */
trace_init();
if (initcall_debug)
initcall_debug_enable();
context_tracking_init();
/* init some links before init_ISA_irqs() */
early_irq_init();
init_IRQ();
tick_init();
rcu_init_nohz();
init_timers();
hrtimers_init();
softirq_init();
timekeeping_init();
time_init();
sched_clock_postinit();
printk_safe_init();
perf_event_init();
profile_init();
call_function_init();
WARN(!irqs_disabled(), "Interrupts were enabled early\n");
early_boot_irqs_disabled = false;
local_irq_enable();
kmem_cache_init_late();
/*
* 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();
printk("## start_kernel() --> console_init()\n");
if (panic_later)
panic("Too many boot %s vars at `%s'", 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();
/*
* This needs to be called before any devices perform DMA
* operations that might use the SWIOTLB bounce buffers. It will
* mark the bounce buffers as decrypted so that their usage will
* not cause "plain-text" data to be decrypted when accessed.
*/
mem_encrypt_init();
#ifdef CONFIG_BLK_DEV_INITRD
if (initrd_start && !initrd_below_start_ok &&
page_to_pfn(virt_to_page((void *)initrd_start)) < min_low_pfn) {
pr_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
page_ext_init();
kmemleak_init();
debug_objects_mem_init();
setup_per_cpu_pageset();
numa_policy_init();
acpi_early_init();
if (late_time_init)
late_time_init();
calibrate_delay();
pid_idr_init();
anon_vma_init();
#ifdef CONFIG_X86
if (efi_enabled(EFI_RUNTIME_SERVICES))
efi_enter_virtual_mode();
#endif
thread_stack_cache_init();
cred_init();
fork_init();
proc_caches_init();
uts_ns_init();
buffer_init();
key_init();
security_init();
dbg_late_init();
vfs_caches_init();
pagecache_init();
signals_init();
seq_file_init();
proc_root_init();
nsfs_init();
cpuset_init();
cgroup_init();
taskstats_init_early();
delayacct_init();
check_bugs();
acpi_subsystem_init();
arch_post_acpi_subsys_init();
sfi_init_late();
if (efi_enabled(EFI_RUNTIME_SERVICES)) {
efi_free_boot_services();
}
printk("## run rest_init()\n");
/* Do the rest non-__init'ed, we're now alive */
rest_init();
printk("## after rest_init()\n");
}
一开始尝试在函数刚开始就添加 printk 打印，结果发现添加完 printk 后内核起不来，最后保守起见，在 131 行 console_init(); 后才开始添加打印。
输出log
Starting kernel ...
[ 0.000000] Booting Linux on physical CPU 0x0
[ 0.000000] Linux version 4.18.12 (liyongjun@Box) (gcc version 9.3.0 (Buildroot 2021.05)) #14 SMP Thu Nov 25 00:37:30 CST 2021
[ 0.000000] CPU: ARMv7 Processor [410fc074] revision 4 (ARMv7), cr=10c5387d
[ 0.000000] CPU: div instructions available: patching division code
[ 0.000000] CPU: PIPT / VIPT nonaliasing data cache, VIPT aliasing instruction cache
[ 0.000000] OF: fdt: Machine model: LeMaker Banana Pi
[ 0.000000] Memory policy: Data cache writealloc
[ 0.000000] cma: Reserved 16 MiB at 0x7ec00000
[ 0.000000] psci: probing for conduit method from DT.
[ 0.000000] psci: Using PSCI v0.1 Function IDs from DT
[ 0.000000] random: get_random_bytes called from start_kernel+0xa0/0x430 with crng_init=0
[ 0.000000] percpu: Embedded 16 pages/cpu @(ptrval) s34444 r8192 d22900 u65536
[ 0.000000] Built 1 zonelists, mobility grouping on. Total pages: 260202
[ 0.000000] Kernel command line: console=ttyS0,57600 earlyprintk root=/dev/mmcblk0p2 rootwait
[ 0.000000] Dentry cache hash table entries: 131072 (order: 7, 524288 bytes)
[ 0.000000] Inode-cache hash table entries: 65536 (order: 6, 262144 bytes)
[ 0.000000] Memory: 1011460K/1046952K available (6144K kernel code, 418K rwdata, 1524K rodata, 1024K init, 240K bss, 19108K reserved, 16384K cma-reserved, 244136K highmem)
[ 0.000000] Virtual kernel memory layout:
[ 0.000000] vector : 0xffff0000 - 0xffff1000 ( 4 kB)
[ 0.000000] fixmap : 0xffc00000 - 0xfff00000 (3072 kB)
[ 0.000000] vmalloc : 0xf0800000 - 0xff800000 ( 240 MB)
[ 0.000000] lowmem : 0xc0000000 - 0xf0000000 ( 768 MB)
[ 0.000000] pkmap : 0xbfe00000 - 0xc0000000 ( 2 MB)
[ 0.000000] modules : 0xbf000000 - 0xbfe00000 ( 14 MB)
[ 0.000000] .text : 0x(ptrval) - 0x(ptrval) (7136 kB)
[ 0.000000] .init : 0x(ptrval) - 0x(ptrval) (1024 kB)
[ 0.000000] .data : 0x(ptrval) - 0x(ptrval) ( 419 kB)
[ 0.000000] .bss : 0x(ptrval) - 0x(ptrval) ( 241 kB)
[ 0.000000] SLUB: HWalign=64, Order=0-3, MinObjects=0, CPUs=2, Nodes=1
[ 0.000000] Hierarchical RCU implementation.
[ 0.000000] RCU restricting CPUs from NR_CPUS=8 to nr_cpu_ids=2.
[ 0.000000] RCU: Adjusting geometry for rcu_fanout_leaf=16, nr_cpu_ids=2
[ 0.000000] NR_IRQS: 16, nr_irqs: 16, preallocated irqs: 16
[ 0.000000] GIC: Using split EOI/Deactivate mode
[ 0.000000] arch_timer: cp15 timer(s) running at 24.00MHz (phys).
[ 0.000000] clocksource: arch_sys_counter: mask: 0xffffffffffffff max_cycles: 0x588fe9dc0, max_idle_ns: 440795202592 ns
[ 0.000007] sched_clock: 56 bits at 24MHz, resolution 41ns, wraps every 4398046511097ns
[ 0.000021] Switching to timer-based delay loop, resolution 41ns
[ 0.000334] clocksource: timer: mask: 0xffffffff max_cycles: 0xffffffff, max_idle_ns: 79635851949 ns
[ 0.000586] clocksource: hstimer: mask: 0xffffffff max_cycles: 0xffffffff, max_idle_ns: 6370868154 ns
[ 0.000822] Console: colour dummy device 80x30
[ 0.000844] ## start_kernel() --> console_init() // 132 行的打印
可是，发现在 console_init() 之前就有不少打印了，这个地方还是有些不解。猜测要么是在 console_init() 之前就已经可以打印了；要么是在 console_init() 之前先将打印缓存着，等初始化之后再打印。
接着看打印
[ 0.000844] ## start_kernel() --> console_init()
[ 0.000872] Calibrating delay loop (skipped), value calculated using timer frequency.. 48.00 BogoMIPS (lpj=240000)
[ 0.000886] pid_max: default: 32768 minimum: 301
[ 0.001054] Mount-cache hash table entries: 2048 (order: 1, 8192 bytes)
[ 0.001070] Mountpoint-cache hash table entries: 2048 (order: 1, 8192 bytes)
[ 0.001727] CPU: Testing write buffer coherency: ok
[ 0.001779] ## run rest_init()
从 133 行到 206 行，代码不少，打印却只有寥寥 5 行。少就说明不重要，那就不分析了，哈哈。
看下面的代码，乖乖，就剩一句了：rest_init();，并且第 210 行的 printk 等到系统完全起来都没有打印，说明 rest_init() 就没返回。看来是个扛把子。
static noinline void __ref rest_init(void)
{
struct task_struct *tsk;
int pid;
printk("## rcu_scheduler_starting()\n");
rcu_scheduler_starting();
/*
* We need to spawn init first so that it obtains pid 1, however
* the init task will end up wanting to create kthreads, which, if
* we schedule it before we create kthreadd, will OOPS.
*/
pid = kernel_thread(kernel_init, NULL, CLONE_FS);
/*
* Pin init on the boot CPU. Task migration is not properly working
* until sched_init_smp() has been run. It will set the allowed
* CPUs for init to the non isolated CPUs.
*/
rcu_read_lock();
tsk = find_task_by_pid_ns(pid, &init_pid_ns);
set_cpus_allowed_ptr(tsk, cpumask_of(smp_processor_id()));
rcu_read_unlock();
printk("## bb\n");
numa_default_policy();
pid = kernel_thread(kthreadd, NULL, CLONE_FS | CLONE_FILES);
rcu_read_lock();
kthreadd_task = find_task_by_pid_ns(pid, &init_pid_ns);
rcu_read_unlock();
printk("## cc\n");
/*
* Enable might_sleep() and smp_processor_id() checks.
* They cannot be enabled earlier because with CONFIG_PREEMPT=y
* kernel_thread() would trigger might_sleep() splats. With
* CONFIG_PREEMPT_VOLUNTARY=y the init task might have scheduled
* already, but it's stuck on the kthreadd_done completion.
*/
system_state = SYSTEM_SCHEDULING;
printk("## dd\n");
complete(&kthreadd_done);
/*
* The boot idle thread must execute schedule()
* at least once to get things moving:
*/
printk("## ee\n");
schedule_preempt_disabled();
printk("## ff\n");
/* Call into cpu_idle with preempt disabled */
cpu_startup_entry(CPUHP_ONLINE);
// printk("## ff\n");
}
对应打印如下：
[ 0.001779] ## run rest_init()
[ 0.001786] ## rcu_scheduler_starting()
[ 0.002030] ## bb
[ 0.002089] ## cc
[ 0.002097] ## dd
[ 0.002104] ## ee
[ 0.002150] ## kernel_init()
[ 0.002199] /cpus/cpu@0 missing clock-frequency property
[ 0.002214] /cpus/cpu@1 missing clock-frequency property
[ 0.002228] CPU0: thread -1, cpu 0, socket 0, mpidr 80000000
[ 0.002799] Setting up static identity map for 0x40100000 - 0x40100060
[ 0.002972] Hierarchical SRCU implementation.
[ 0.003776] smp: Bringing up secondary CPUs ...
[ 0.014335] ## ff
打印了 kernel_init()，是因为调用了 schedule_preempt_disabled()，其内部调用了 schedule() 执行进程切换，进入睡眠。当然，在当前进程被唤醒时，程序也是从这个断点开始运行。
void __sched schedule_preempt_disabled(void)
{
sched_preempt_enable_no_resched();
schedule();
preempt_disable();
}
执行进程切换后，kernel_init() 才有机会运行
static int __ref kernel_init(void *unused)
{
int ret;
printk("## kernel_init()\n");
kernel_init_freeable();
/* need to finish all async __init code before freeing the memory */
async_synchronize_full();
ftrace_free_init_mem();
jump_label_invalidate_initmem();
free_initmem();
printk("## 2.\n");
mark_readonly();
printk("## 3.\n");
system_state = SYSTEM_RUNNING;
numa_default_policy();
rcu_end_inkernel_boot();
printk("## 4.\n");
if (ramdisk_execute_command) {
ret = run_init_process(ramdisk_execute_command);
if (!ret)
return 0;
pr_err("Failed to execute %s (error %d)\n",
ramdisk_execute_command, ret);
}
/*
* 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) {
ret = run_init_process(execute_command);
if (!ret)
return 0;
panic("Requested init %s failed (error %d).",
execute_command, ret);
}
if (!try_to_run_init_process("/sbin/init") ||
!try_to_run_init_process("/etc/init") ||
!try_to_run_init_process("/bin/init") ||
!try_to_run_init_process("/bin/sh")) {
printk("## kernel_init() --> /sbin/init\n");
return 0;
}
panic("No working init found. Try passing init= option to kernel. "
"See Linux Documentation/admin-guide/init.rst for guidance.");
}
对应打印如下：
[ 0.002150] ## kernel_init()
[ 0.002199] /cpus/cpu@0 missing clock-frequency property
[ 0.002214] /cpus/cpu@1 missing clock-frequency property
[ 0.002228] CPU0: thread -1, cpu 0, socket 0, mpidr 80000000
[ 0.002799] Setting up static identity map for 0x40100000 - 0x40100060
[ 0.002972] Hierarchical SRCU implementation.
[ 0.003776] smp: Bringing up secondary CPUs ...
[ 0.014335] ## ff
[ 0.014540] CPU1: thread -1, cpu 1, socket 0, mpidr 80000001
[ 0.014681] smp: Brought up 1 node, 2 CPUs
[ 0.014697] SMP: Total of 2 processors activated (96.00 BogoMIPS).
[ 0.014704] CPU: All CPU(s) started in HYP mode.
[ 0.014709] CPU: Virtualization extensions available.
[ 0.015730] devtmpfs: initialized
[ 0.015786] ## 11.
[ 0.015802] ## 22.
[ 0.015811] ## 33.
[ 0.015820] ## 44.
[ 0.015825] ## 55.
[ 0.015918] ## 66.
[ 0.015999] ## 77.
[ 0.016006] ## 88.
[ 0.016058] ## 99.
[ 0.021654] ## AA.
[ 0.021836] ### AAA.
[ 0.022450] VFP support v0.3: implementor 41 architecture 2 part 30 variant 7 rev 4
[ 0.022721] clocksource: jiffies: mask: 0xffffffff max_cycles: 0xffffffff, max_idle_ns: 19112604462750000 ns
[ 0.022747] futex hash table entries: 512 (order: 3, 32768 bytes)
[ 0.023540] pinctrl core: initialized pinctrl subsystem
[ 0.024726] NET: Registered protocol family 16
[ 0.026025] DMA: preallocated 256 KiB pool for atomic coherent allocations
[ 0.027158] hw-breakpoint: found 5 (+1 reserved) breakpoint and 4 watchpoint registers.
[ 0.027173] hw-breakpoint: maximum watchpoint size is 8 bytes.
[ 0.042711] reg-fixed-voltage gmac-3v3: could not find pctldev for node /soc@1c00000/pinctrl@1c20800/gmac_power_pin@0, deferring probe
[ 0.043233] SCSI subsystem initialized
[ 0.043811] usbcore: registered new interface driver usbfs
[ 0.043864] usbcore: registered new interface driver hub
[ 0.043933] usbcore: registered new device driver usb
[ 0.044199] pps_core: LinuxPPS API ver. 1 registered
[ 0.044209] pps_core: Software ver. 5.3.6 - Copyright 2005-2007 Rodolfo Giometti <giometti@linux.it>
[ 0.044230] PTP clock support registered
[ 0.044563] Advanced Linux Sound Architecture Driver Initialized.
[ 0.045501] clocksource: Switched to clocksource arch_sys_counter
[ 0.046238] simple-framebuffer 7fe79000.framebuffer: framebuffer at 0x7fe79000, 0x178e00 bytes, mapped to 0x(ptrval)
[ 0.046258] simple-framebuffer 7fe79000.framebuffer: format=x8r8g8b8, mode=656x536x32, linelength=2880
[ 0.052472] Console: switching to colour frame buffer device 82x33
[ 0.058369] simple-framebuffer 7fe79000.framebuffer: fb0: simplefb registered!
[ 0.066686] NET: Registered protocol family 2
[ 0.067281] tcp_listen_portaddr_hash hash table entries: 512 (order: 0, 6144 bytes)
[ 0.067316] TCP established hash table entries: 8192 (order: 3, 32768 bytes)
[ 0.067395] TCP bind hash table entries: 8192 (order: 4, 65536 bytes)
[ 0.067522] TCP: Hash tables configured (established 8192 bind 8192)
[ 0.067655] UDP hash table entries: 512 (order: 2, 16384 bytes)
[ 0.067722] UDP-Lite hash table entries: 512 (order: 2, 16384 bytes)
[ 0.067964] NET: Registered protocol family 1
[ 0.068567] RPC: Registered named UNIX socket transport module.
[ 0.068583] RPC: Registered udp transport module.
[ 0.068590] RPC: Registered tcp transport module.
[ 0.068596] RPC: Registered tcp NFSv4.1 backchannel transport module.
[ 0.069204] hw perfevents: no interrupt-affinity property for /pmu, guessing.
[ 0.069499] hw perfevents: enabled with armv7_cortex_a7 PMU driver, 5 counters available
[ 0.071072] workingset: timestamp_bits=30 max_order=18 bucket_order=0
[ 0.077754] NFS: Registering the id_resolver key type
[ 0.077803] Key type id_resolver registered
[ 0.077812] Key type id_legacy registered
[ 0.079006] bounce: pool size: 64 pages
[ 0.079106] Block layer SCSI generic (bsg) driver version 0.4 loaded (major 248)
[ 0.079119] io scheduler noop registered
[ 0.079127] io scheduler deadline registered
[ 0.079295] io scheduler cfq registered (default)
[ 0.079307] io scheduler mq-deadline registered
[ 0.079315] io scheduler kyber registered
[ 0.080074] sun4i-usb-phy 1c13400.phy: could not find pctldev for node /soc@1c00000/pinctrl@1c20800/usb0_id_detect_pin@0, deferring probe
[ 0.083501] sun4i-pinctrl 1c20800.pinctrl: initialized sunXi PIO driver
[ 0.140590] Serial: 8250/16550 driver, 8 ports, IRQ sharing disabled
[ 0.143240] console [ttyS0] disabled
[ 0.163431] 1c28000.serial: ttyS0 at MMIO 0x1c28000 (irq = 48, base_baud = 1500000) is a U6_16550A
[ 1.630962] console [ttyS0] enabled
[ 1.661099] 1c28c00.serial: ttyS1 at MMIO 0x1c28c00 (irq = 49, base_baud = 1500000) is a U6_16550A
[ 1.702024] 1c29c00.serial: ttyS2 at MMIO 0x1c29c00 (irq = 50, base_baud = 1500000) is a U6_16550A
[ 1.723737] sun4i-backend 1e60000.display-backend: Couldn't find matching frontend, frontend features disabled
[ 1.744080] sun4i-drm display-engine: bound 1e60000.display-backend (ops 0xc07482cc)
[ 1.759592] sun4i-backend 1e40000.display-backend: Couldn't find matching frontend, frontend features disabled
[ 1.779827] sun4i-drm display-engine: bound 1e40000.display-backend (ops 0xc07482cc)
[ 1.795785] sun4i-drm display-engine: No panel or bridge found... RGB output disabled
[ 1.811456] sun4i-drm display-engine: bound 1c0c000.lcd-controller (ops 0xc074724c)
[ 1.827130] sun4i-drm display-engine: No panel or bridge found... RGB output disabled
[ 1.842825] sun4i-drm display-engine: bound 1c0d000.lcd-controller (ops 0xc074724c)
[ 1.858540] sun4i-drm display-engine: bound 1c16000.hdmi (ops 0xc0748cd0)
[ 1.872123] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
[ 1.885344] [drm] No driver support for vblank timestamp query.
[ 1.897182] fb: switching to sun4i-drm-fb from simple
[ 1.907391] Console: switching to colour dummy device 80x30
[ 2.418984] [drm] Cannot find any crtc or sizes
[ 2.428711] [drm] Initialized sun4i-drm 1.0.0 20150629 for display-engine on minor 0
[ 2.545528] ahci-sunxi 1c18000.sata: controller can't do PMP, turning off CAP_PMP
[ 2.560532] ahci-sunxi 1c18000.sata: SSS flag set, parallel bus scan disabled
[ 2.574817] ahci-sunxi 1c18000.sata: AHCI 0001.0100 32 slots 1 ports 3 Gbps 0x1 impl platform mode
[ 2.592721] ahci-sunxi 1c18000.sata: flags: ncq sntf stag pm led clo only pio slum part ccc
[ 2.610642] scsi host0: ahci-sunxi
[ 2.617874] ata1: SATA max UDMA/133 mmio [mem 0x01c18000-0x01c18fff] port 0x100 irq 36
[ 2.635546] libphy: Fixed MDIO Bus: probed
[ 2.643735] CAN device driver interface
[ 2.652152] sun7i-dwmac 1c50000.ethernet: PTP uses main clock
[ 2.663680] sun7i-dwmac 1c50000.ethernet: no reset control found
[ 2.676345] ehci_hcd: USB 2.0 'Enhanced' Host Controller (EHCI) Driver
[ 2.689416] ehci-platform: EHCI generic platform driver
[ 2.700248] ohci_hcd: USB 1.1 'Open' Host Controller (OHCI) Driver
[ 2.712647] ohci-platform: OHCI generic platform driver
[ 2.725114] sunxi-rtc 1c20d00.rtc: registered as rtc0
[ 2.735252] sunxi-rtc 1c20d00.rtc: RTC enabled
[ 2.744274] i2c /dev entries driver
[ 2.752374] axp20x-i2c 1-0034: AXP20x variant AXP209 found
[ 2.778365] input: axp20x-pek as /devices/platform/soc@1c00000/1c2ac00.i2c/i2c-1/1-0034/axp20x-pek/input/input0
[ 2.799556] ldo1: supplied by regulator-dummy
[ 2.808582] ldo2: supplied by regulator-dummy
[ 2.818392] ldo3: supplied by regulator-dummy
[ 2.827795] ldo4: supplied by regulator-dummy
[ 2.836730] ldo5: supplied by regulator-dummy
[ 2.846544] dcdc2: supplied by regulator-dummy
[ 2.856093] dcdc3: supplied by regulator-dummy
[ 2.866705] axp20x-i2c 1-0034: AXP20X driver loaded
[ 2.877918] Registered IR keymap rc-empty
[ 2.886099] rc rc0: sunxi-ir as /devices/platform/soc@1c00000/1c21800.ir/rc/rc0
[ 2.900854] input: sunxi-ir as /devices/platform/soc@1c00000/1c21800.ir/rc/rc0/input1
[ 2.916966] sunxi-ir 1c21800.ir: initialized sunXi IR driver
[ 2.930370] sunxi-wdt 1c20c90.watchdog: Watchdog enabled (timeout=16 sec, nowayout=0)
[ 2.949908] sunxi-mmc 1c0f000.mmc: Got CD GPIO
[ 2.957174] ata1: SATA link down (SStatus 0 SControl 300)
[ 2.984443] sunxi-mmc 1c0f000.mmc: base:0x(ptrval) irq:30
[ 2.996403] sun4i-ss 1c15000.crypto-engine: Die ID 0
[ 3.008208] usbcore: registered new interface driver usbhid
[ 3.019411] usbhid: USB HID core driver
[ 3.028477] random: fast init done
[ 3.038082] sun4i-codec 1c22c00.codec: ASoC: Failed to create component debugfs directory
[ 3.055988] sun4i-codec 1c22c00.codec: Codec <-> 1c22c00.codec mapping ok
[ 3.070975] NET: Registered protocol family 17
[ 3.079931] can: controller area network core (rev 20170425 abi 9)
[ 3.092458] NET: Registered protocol family 29
[ 3.101361] can: raw protocol (rev 20170425)
[ 3.109908] can: broadcast manager protocol (rev 20170425 t)
[ 3.121240] can: netlink gateway (rev 20170425) max_hops=1
[ 3.132444] Key type dns_resolver registered
[ 3.141113] Registering SWP/SWPB emulation handler
[ 3.155150] mmc0: host does not support reading read-only switch, assuming write-enable
[ 3.162888] sun7i-dwmac 1c50000.ethernet: PTP uses main clock
[ 3.182777] sun7i-dwmac 1c50000.ethernet: no reset control found
[ 3.194864] mmc0: new high speed SDHC card at address 59b4
[ 3.206939] mmcblk0: mmc0:59b4 USD00 7.32 GiB
[ 3.217882] mmcblk0: p1 p2
[ 3.305735] sun7i-dwmac 1c50000.ethernet: Version ID not available
[ 3.318137] sun7i-dwmac 1c50000.ethernet: DWMAC1000
[ 3.328068] sun7i-dwmac 1c50000.ethernet: DMA HW capability register supported
[ 3.342498] sun7i-dwmac 1c50000.ethernet: Normal descriptors
[ 3.353804] sun7i-dwmac 1c50000.ethernet: Ring mode enabled
[ 3.372879] libphy: stmmac: probed
[ 3.379711] mdio_bus stmmac-0:00: attached PHY driver [unbound] (mii_bus:phy_addr=stmmac-0:00, irq=POLL)
[ 3.398646] mdio_bus stmmac-0:01: attached PHY driver [unbound] (mii_bus:phy_addr=stmmac-0:01, irq=POLL)
[ 3.437826] ehci-platform 1c14000.usb: EHCI Host Controller
[ 3.449048] ehci-platform 1c14000.usb: new USB bus registered, assigned bus number 1
[ 4.445622] [drm] Cannot find any crtc or sizes
[ 4.454972] ehci-platform 1c14000.usb: irq 32, io mem 0x01c14000
[ 4.495500] ehci-platform 1c14000.usb: USB 2.0 started, EHCI 1.00
[ 4.508602] hub 1-0:1.0: USB hub found
[ 4.516156] hub 1-0:1.0: 1 port detected
[ 4.524842] ehci-platform 1c1c000.usb: EHCI Host Controller
[ 4.536034] ehci-platform 1c1c000.usb: new USB bus registered, assigned bus number 2
[ 4.551813] ehci-platform 1c1c000.usb: irq 37, io mem 0x01c1c000
[ 4.585504] ehci-platform 1c1c000.usb: USB 2.0 started, EHCI 1.00
[ 4.598505] hub 2-0:1.0: USB hub found
[ 4.606055] hub 2-0:1.0: 1 port detected
[ 4.614695] ohci-platform 1c14400.usb: Generic Platform OHCI controller
[ 4.627974] ohci-platform 1c14400.usb: new USB bus registered, assigned bus number 3
[ 4.643687] ohci-platform 1c14400.usb: irq 33, io mem 0x01c14400
[ 4.730281] hub 3-0:1.0: USB hub found
[ 4.737832] hub 3-0:1.0: 1 port detected
[ 4.746524] ohci-platform 1c1c400.usb: Generic Platform OHCI controller
[ 4.759793] ohci-platform 1c1c400.usb: new USB bus registered, assigned bus number 4
[ 4.775542] ohci-platform 1c1c400.usb: irq 38, io mem 0x01c1c400
[ 4.860311] hub 4-0:1.0: USB hub found
[ 4.867876] hub 4-0:1.0: 1 port detected
[ 4.876658] usb_phy_generic usb_phy_generic.0.auto: usb_phy_generic.0.auto supply vcc not found, using dummy regulator
[ 4.898584] musb-hdrc musb-hdrc.1.auto: MUSB HDRC host driver
[ 4.910112] musb-hdrc musb-hdrc.1.auto: new USB bus registered, assigned bus number 5
[ 4.926629] hub 5-0:1.0: USB hub found
[ 4.934193] hub 5-0:1.0: 1 port detected
[ 4.942801] sunxi-rtc 1c20d00.rtc: setting system clock to 1970-01-01 00:00:11 UTC (11)
[ 4.959189] vcc3v0: disabling
[ 4.965121] vcc5v0: disabling
[ 4.971086] usb0-vbus: disabling
[ 4.977556] ALSA device list:
[ 4.983479] #0: sun4i-codec
[ 4.989417] ### BBB.
[ 4.997149] EXT4-fs (mmcblk0p2): INFO: recovery required on readonly filesystem
[ 5.011809] EXT4-fs (mmcblk0p2): write access will be enabled during recovery
[ 5.029755] EXT4-fs (mmcblk0p2): recovery complete
[ 5.090424] EXT4-fs (mmcblk0p2): mounted filesystem with ordered data mode. Opts: (null)
[ 5.106669] VFS: Mounted root (ext4 filesystem) readonly on device 179:2.
[ 5.121202] devtmpfs: mounted
[ 5.127186] ## 111.
[ 5.132783] Freeing unused kernel memory: 1024K
[ 5.141851] ## 2.
可以看到从第 5 行到第 12 行区区几个函数，竟然打印了那么多，而且还看到了众多驱动相关的打印，如 usb、Console、Key、io、pinctrl、Serial、CAN、RTC、input、mmc、mdio_bus，说明我们整天开发的驱动，是在这个阶段初始化的呀。这里我直接给出调用关系，可以看到，驱动相关的打印是在 do_initcalls() 中产生的
kernel_init()
kernel_init_freeable()
do_basic_setup()
printk("### AAA.\n");
do_initcalls();
printk("### BBB.\n");
static void __init do_initcalls(void)
{
int level;
for (level = 0; level < ARRAY_SIZE(initcall_levels) - 1; level++)
do_initcall_level(level);
}
do_initcalls() 函数循环将编译进内核的内核模块的初始化函数依次调用一遍。这些初始化函数也是有等级的，总共有 7 个级别 1-7，数字越大，级别越低，被调用的就越晚。其中我们最长用的 module_init() 为第 6 级，等级算是比较低的了，调用也会较晚，想想也是，好多外部模块都是使用的 module_init()，它们可都是在内核完全启动后，再手动安装的呀，更晚。
这个阶段打印了很多 log，说明内核中大部分的代码都是驱动，内核中写了大量的内核模块初始化的函数。
驱动初始化完，内核也就接近尾声了
[ 4.989417] ### BBB.
[ 4.997149] EXT4-fs (mmcblk0p2): INFO: recovery required on readonly filesystem
[ 5.011809] EXT4-fs (mmcblk0p2): write access will be enabled during recovery
[ 5.029755] EXT4-fs (mmcblk0p2): recovery complete
[ 5.090424] EXT4-fs (mmcblk0p2): mounted filesystem with ordered data mode. Opts: (null)
[ 5.106669] VFS: Mounted root (ext4 filesystem) readonly on device 179:2.
[ 5.121202] devtmpfs: mounted
[ 5.127186] ## 111.
[ 5.132783] Freeing unused kernel memory: 1024K
[ 5.141851] ## 2.
[ 5.145862] ## 3.
[ 5.149703] ## 4.
[ 5.162188] ## kernel_init() --> /sbin/init
## init_main()
[ 5.232554] EXT4-fs (mmcblk0p2): re-mounted. Opts: (null)
Starting syslogd: OK
Starting klogd: OK
Running sysctl: OK
Initializing random number generator: OK
Saving random seed: [ 5.377468] random: dd: uninitialized urandom read (512 bytes read)
OK
Starting network: OK
Welcome to Bananapi M1
buildroot login:
- 挂载根文件系统
- 内核收尾工作
- kernel_init 内核线程调用 do_execve 函数，去执行 init 程序，产生 init 进程（1 号进程）。该内核线程终结。
rest_init()函数的最后一句cpu_startup_entry(CPUHP_ONLINE);最终变成了 idle 进程（0 号进程）。
void cpu_startup_entry(enum cpuhp_state state)
{
/*
* This #ifdef needs to die, but it's too late in the cycle to
* make this generic (ARM and SH have never invoked the canary
* init for the non boot CPUs!). Will be fixed in 3.11
*/
#ifdef CONFIG_X86
/*
* If we're the non-boot CPU, nothing set the stack canary up
* for us. The boot CPU already has it initialized but no harm
* in doing it again. This is a good place for updating it, as
* we wont ever return from this function (so the invalid
* canaries already on the stack wont ever trigger).
*/
boot_init_stack_canary();
#endif
arch_cpu_idle_prepare();
cpuhp_online_idle(state);
while (1)
do_idle();
}
而在 kernel_init 同期稍晚创建的 kthreadd 内核线程，则一直存在，成为 2 号进程
# ps
PID USER COMMAND
1 root init
2 root [kthreadd]
3 root [rcu_gp]
4 root [rcu_par_gp]
7 root [kworker/u4:0-ev]
8 root [mm_percpu_wq]
9 root [ksoftirqd/0]
...
本文转载自：Kernel 启动流程梳理
原作者：Li-Yongjun
本文转载仅作学习交流使用，支持原创请访问原文章！