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 的第一条指令,也就是下面这个函数
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:
# psPID USER COMMAND1 root init2 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_INITRDif (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;}#endifpage_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_X86if (efi_enabled(EFI_RUNTIME_SERVICES))efi_enter_virtual_mode();#endifthread_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。
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。这点以后再研究吧。先插个眼👁。
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: OKStarting klogd: OKRunning sysctl: OKInitializing random number generator: OKSaving random seed: [ 5.377468] random: dd: uninitialized urandom read (512 bytes read)OKStarting network: OKWelcome to Bananapi M1buildroot 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();#endifarch_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
# psPID USER COMMAND1 root init2 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]...
转自:
https://liyongjun.blog.csdn.net/article/details/121529063
这个大佬的其他文章写的也非常不错,对初学者有很大的借鉴意义,虽然说csdn有些地方做得不是很好,但是确实是国内内容最丰富的技术平台。