init/main.c

   1 /*
   2  *  linux/init/main.c
   3  *
   4  *  Copyright (C) 1991, 1992  Linus Torvalds
   5  *
   6  *  GK 2/5/95  -  Changed to support mounting root fs via NFS
   7  *  Added initrd & change_root: Werner Almesberger & Hans Lermen, Feb '96
   8  *  Moan early if gcc is old, avoiding bogus kernels - Paul Gortmaker, May '96
   9  *  Simplified starting of init:  Michael A. Griffith <grif@acm.org>
  10  */
  11
  12 #include <linux/types.h>
  13 #include <linux/module.h>
  14 #include <linux/proc_fs.h>
  15 #include <linux/kernel.h>
  16 #include <linux/syscalls.h>
  17 #include <linux/stackprotector.h>
  18 #include <linux/string.h>
  19 #include <linux/ctype.h>
  20 #include <linux/delay.h>
  21 #include <linux/ioport.h>
  22 #include <linux/init.h>
  23 #include <linux/initrd.h>
  24 #include <linux/bootmem.h>
  25 #include <linux/acpi.h>
  26 #include <linux/tty.h>
  27 #include <linux/percpu.h>
  28 #include <linux/kmod.h>
  29 #include <linux/vmalloc.h>
  30 #include <linux/kernel_stat.h>
  31 #include <linux/start_kernel.h>
  32 #include <linux/security.h>
  33 #include <linux/smp.h>
  34 #include <linux/profile.h>
  35 #include <linux/rcupdate.h>
  36 #include <linux/moduleparam.h>
  37 #include <linux/kallsyms.h>
  38 #include <linux/writeback.h>
  39 #include <linux/cpu.h>
  40 #include <linux/cpuset.h>
  41 #include <linux/cgroup.h>
  42 #include <linux/efi.h>
  43 #include <linux/tick.h>
  44 #include <linux/interrupt.h>
  45 #include <linux/taskstats_kern.h>
  46 #include <linux/delayacct.h>
  47 #include <linux/unistd.h>
  48 #include <linux/rmap.h>
  49 #include <linux/mempolicy.h>
  50 #include <linux/key.h>
  51 #include <linux/buffer_head.h>
  52 #include <linux/page_cgroup.h>
  53 #include <linux/debug_locks.h>
  54 #include <linux/debugobjects.h>
  55 #include <linux/lockdep.h>
  56 #include <linux/kmemleak.h>
  57 #include <linux/pid_namespace.h>
  58 #include <linux/device.h>
  59 #include <linux/kthread.h>
  60 #include <linux/sched.h>
  61 #include <linux/signal.h>
  62 #include <linux/idr.h>
  63 #include <linux/kgdb.h>
  64 #include <linux/ftrace.h>
  65 #include <linux/async.h>
  66 #include <linux/kmemcheck.h>
  67 #include <linux/sfi.h>
  68 #include <linux/shmem_fs.h>
  69 #include <linux/slab.h>
  70 #include <linux/perf_event.h>
  71
  72 #include <asm/io.h>
  73 #include <asm/bugs.h>
  74 #include <asm/setup.h>
  75 #include <asm/sections.h>
  76 #include <asm/cacheflush.h>
  77
  78 #ifdef CONFIG_X86_LOCAL_APIC
  79 #include <asm/smp.h>
  80 #endif
  81
  82 static int kernel_init(void *);
  83
  84 extern void init_IRQ(void);
  85 extern void fork_init(unsigned long);
  86 extern void mca_init(void);
  87 extern void sbus_init(void);
  88 extern void prio_tree_init(void);
  89 extern void radix_tree_init(void);
  90 extern void free_initmem(void);
  91 #ifndef CONFIG_DEBUG_RODATA
  92 static inline void mark_rodata_ro(void) { }
  93 #endif
  94
  95 #ifdef CONFIG_TC
  96 extern void tc_init(void);
  97 #endif
  98
  99 /*
 100  * Debug helper: via this flag we know that we are in 'early bootup code'
 101  * where only the boot processor is running with IRQ disabled.  This means
 102  * two things - IRQ must not be enabled before the flag is cleared and some
 103  * operations which are not allowed with IRQ disabled are allowed while the
 104  * flag is set.
 105  */
 106 bool early_boot_irqs_disabled __read_mostly;
 107
 108 enum system_states system_state __read_mostly;
 109 EXPORT_SYMBOL(system_state);
 110
 111 /*
 112  * Boot command-line arguments
 113  */
 114 #define MAX_INIT_ARGS CONFIG_INIT_ENV_ARG_LIMIT
 115 #define MAX_INIT_ENVS CONFIG_INIT_ENV_ARG_LIMIT
 116
 117 extern void time_init(void);
 118 /* Default late time init is NULL. archs can override this later. */
 119 void (*__initdata late_time_init)(void);
 120 extern void softirq_init(void);
 121
 122 /* Untouched command line saved by arch-specific code. */
 123 char __initdata boot_command_line[COMMAND_LINE_SIZE];
 124 /* Untouched saved command line (eg. for /proc) */
 125 char *saved_command_line;
 126 /* Command line for parameter parsing */
 127 static char *static_command_line;
 128
 129 static char *execute_command;
 130 static char *ramdisk_execute_command;
 131
 132 /*
 133  * If set, this is an indication to the drivers that reset the underlying
 134  * device before going ahead with the initialization otherwise driver might
 135  * rely on the BIOS and skip the reset operation.
 136  *
 137  * This is useful if kernel is booting in an unreliable environment.
 138  * For ex. kdump situaiton where previous kernel has crashed, BIOS has been
 139  * skipped and devices will be in unknown state.
 140  */
 141 unsigned int reset_devices;
 142 EXPORT_SYMBOL(reset_devices);
 143
 144 static int __init set_reset_devices(char *str)
 145 {
 146         reset_devices = 1;
 147         return 1;
 148 }
 149
 150 __setup("reset_devices", set_reset_devices);
 151
 152 static const char * argv_init[MAX_INIT_ARGS+2] = { "init", NULL, };
 153 const char * envp_init[MAX_INIT_ENVS+2] = { "HOME=/", "TERM=linux", NULL, };
 154 static const char *panic_later, *panic_param;
 155
 156 extern const struct obs_kernel_param __setup_start[], __setup_end[];
 157
 158 static int __init obsolete_checksetup(char *line)
 159 {
 160         const struct obs_kernel_param *p;
 161         int had_early_param = 0;
 162
 163         p = __setup_start;
 164         do {
 165                 int n = strlen(p->str);
 166                 if (parameqn(line, p->str, n)) {
 167                         if (p->early) {
 168                                 /* Already done in parse_early_param?
 169                                  * (Needs exact match on param part).
 170                                  * Keep iterating, as we can have early
 171                                  * params and __setups of same names 8( */
 172                                 if (line[n] == '\0' || line[n] == '=')
 173                                         had_early_param = 1;
 174                         } else if (!p->setup_func) {
 175                                 printk(KERN_WARNING "Parameter %s is obsolete,"
 176                                        " ignored\n", p->str);
 177                                 return 1;
 178                         } else if (p->setup_func(line + n))
 179                                 return 1;
 180                 }
 181                 p++;
 182         } while (p < __setup_end);
 183
 184         return had_early_param;
 185 }
 186
 187 /*
 188  * This should be approx 2 Bo*oMips to start (note initial shift), and will
 189  * still work even if initially too large, it will just take slightly longer
 190  */
 191 unsigned long loops_per_jiffy = (1<<12);
 192
 193 EXPORT_SYMBOL(loops_per_jiffy);
 194
 195 static int __init debug_kernel(char *str)
 196 {
 197         console_loglevel = 10;
 198         return 0;
 199 }
 200
 201 static int __init quiet_kernel(char *str)
 202 {
 203         console_loglevel = 4;
 204         return 0;
 205 }
 206
 207 early_param("debug", debug_kernel);
 208 early_param("quiet", quiet_kernel);
 209
 210 static int __init loglevel(char *str)
 211 {
 212         int newlevel;
 213
 214         /*
 215          * Only update loglevel value when a correct setting was passed,
 216          * to prevent blind crashes (when loglevel being set to 0) that
 217          * are quite hard to debug
 218          */
 219         if (get_option(&str, &newlevel)) {
 220                 console_loglevel = newlevel;
 221                 return 0;
 222         }
 223
 224         return -EINVAL;
 225 }
 226
 227 early_param("loglevel", loglevel);
 228
 229 /*
 230  * Unknown boot options get handed to init, unless they look like
 231  * unused parameters (modprobe will find them in /proc/cmdline).
 232  */
 233 static int __init unknown_bootoption(char *param, char *val)
 234 {
 235         /* Change NUL term back to "=", to make "param" the whole string. */
 236         if (val) {
 237                 /* param=val or param="val"? */
 238                 if (val == param+strlen(param)+1)
 239                         val[-1] = '=';
 240                 else if (val == param+strlen(param)+2) {
 241                         val[-2] = '=';
 242                         memmove(val-1, val, strlen(val)+1);
 243                         val--;
 244                 } else
 245                         BUG();
 246         }
 247
 248         /* Handle obsolete-style parameters */
 249         if (obsolete_checksetup(param))
 250                 return 0;
 251
 252         /* Unused module parameter. */
 253         if (strchr(param, '.') && (!val || strchr(param, '.') < val))
 254                 return 0;
 255
 256         if (panic_later)
 257                 return 0;
 258
 259         if (val) {
 260                 /* Environment option */
 261                 unsigned int i;
 262                 for (i = 0; envp_init[i]; i++) {
 263                         if (i == MAX_INIT_ENVS) {
 264                                 panic_later = "Too many boot env vars at `%s'";
 265                                 panic_param = param;
 266                         }
 267                         if (!strncmp(param, envp_init[i], val - param))
 268                                 break;
 269                 }
 270                 envp_init[i] = param;
 271         } else {
 272                 /* Command line option */
 273                 unsigned int i;
 274                 for (i = 0; argv_init[i]; i++) {
 275                         if (i == MAX_INIT_ARGS) {
 276                                 panic_later = "Too many boot init vars at `%s'";
 277                                 panic_param = param;
 278                         }
 279                 }
 280                 argv_init[i] = param;
 281         }
 282         return 0;
 283 }
 284
 285 #ifdef CONFIG_DEBUG_PAGEALLOC
 286 int __read_mostly debug_pagealloc_enabled = 0;
 287 #endif
 288
 289 static int __init init_setup(char *str)
 290 {
 291         unsigned int i;
 292
 293         execute_command = str;
 294         /*
 295          * In case LILO is going to boot us with default command line,
 296          * it prepends "auto" before the whole cmdline which makes
 297          * the shell think it should execute a script with such name.
 298          * So we ignore all arguments entered _before_ init=... [MJ]
 299          */
 300         for (i = 1; i < MAX_INIT_ARGS; i++)
 301                 argv_init[i] = NULL;
 302         return 1;
 303 }
 304 __setup("init=", init_setup);
 305
 306 static int __init rdinit_setup(char *str)
 307 {
 308         unsigned int i;
 309
 310         ramdisk_execute_command = str;
 311         /* See "auto" comment in init_setup */
 312         for (i = 1; i < MAX_INIT_ARGS; i++)
 313                 argv_init[i] = NULL;
 314         return 1;
 315 }
 316 __setup("rdinit=", rdinit_setup);
 317
 318 #ifndef CONFIG_SMP
 319 static const unsigned int setup_max_cpus = NR_CPUS;
 320 #ifdef CONFIG_X86_LOCAL_APIC
 321 static void __init smp_init(void)
 322 {
 323         APIC_init_uniprocessor();
 324 }
 325 #else
 326 #define smp_init()      do { } while (0)
 327 #endif
 328
 329 static inline void setup_nr_cpu_ids(void) { }
 330 static inline void smp_prepare_cpus(unsigned int maxcpus) { }
 331 #endif
 332
 333 /*
 334  * We need to store the untouched command line for future reference.
 335  * We also need to store the touched command line since the parameter
 336  * parsing is performed in place, and we should allow a component to
 337  * store reference of name/value for future reference.
 338  */
 339 static void __init setup_command_line(char *command_line)
 340 {
 341         saved_command_line = alloc_bootmem(strlen (boot_command_line)+1);
 342         static_command_line = alloc_bootmem(strlen (command_line)+1);
 343         strcpy (saved_command_line, boot_command_line);
 344         strcpy (static_command_line, command_line);
 345 }
 346
 347 /*
 348  * We need to finalize in a non-__init function or else race conditions
 349  * between the root thread and the init thread may cause start_kernel to
 350  * be reaped by free_initmem before the root thread has proceeded to
 351  * cpu_idle.
 352  *
 353  * gcc-3.4 accidentally inlines this function, so use noinline.
 354  */
 355
 356 static __initdata DECLARE_COMPLETION(kthreadd_done);
 357
 358 static noinline void __init_refok rest_init(void)
 359 {
 360         int pid;
 361
 362         rcu_scheduler_starting();
 363         /*
 364          * We need to spawn init first so that it obtains pid 1, however
 365          * the init task will end up wanting to create kthreads, which, if
 366          * we schedule it before we create kthreadd, will OOPS.
 367          */
 368         kernel_thread(kernel_init, NULL, CLONE_FS | CLONE_SIGHAND);
 369         numa_default_policy();
 370         pid = kernel_thread(kthreadd, NULL, CLONE_FS | CLONE_FILES);
 371         rcu_read_lock();
 372         kthreadd_task = find_task_by_pid_ns(pid, &init_pid_ns);
 373         rcu_read_unlock();
 374         complete(&kthreadd_done);
 375
 376         /*
 377          * The boot idle thread must execute schedule()
 378          * at least once to get things moving:
 379          */
 380         init_idle_bootup_task(current);
 381         preempt_enable_no_resched();
 382         schedule();
 383
 384         /* Call into cpu_idle with preempt disabled */
 385         preempt_disable();
 386         cpu_idle();
 387 }
 388
 389 /* Check for early params. */
 390 static int __init do_early_param(char *param, char *val)
 391 {
 392         const struct obs_kernel_param *p;
 393
 394         for (p = __setup_start; p < __setup_end; p++) {
 395                 if ((p->early && parameq(param, p->str)) ||
 396                     (strcmp(param, "console") == 0 &&
 397                      strcmp(p->str, "earlycon") == 0)
 398                 ) {
 399                         if (p->setup_func(val) != 0)
 400                                 printk(KERN_WARNING
 401                                        "Malformed early option '%s'\n", param);
 402                 }
 403         }
 404         /* We accept everything at this stage. */
 405         return 0;
 406 }
 407
 408 void __init parse_early_options(char *cmdline)
 409 {
 410         parse_args("early options", cmdline, NULL, 0, do_early_param);
 411 }
 412
 413 /* Arch code calls this early on, or if not, just before other parsing. */
 414 void __init parse_early_param(void)
 415 {
 416         static __initdata int done = 0;
 417         static __initdata char tmp_cmdline[COMMAND_LINE_SIZE];
 418
 419         if (done)
 420                 return;
 421
 422         /* All fall through to do_early_param. */
 423         strlcpy(tmp_cmdline, boot_command_line, COMMAND_LINE_SIZE);
 424         parse_early_options(tmp_cmdline);
 425         done = 1;
 426 }
 427
 428 /*
 429  *      Activate the first processor.
 430  */
 431
 432 static void __init boot_cpu_init(void)
 433 {
 434         int cpu = smp_processor_id();
 435         /* Mark the boot cpu "present", "online" etc for SMP and UP case */
 436         set_cpu_online(cpu, true);
 437         set_cpu_active(cpu, true);
 438         set_cpu_present(cpu, true);
 439         set_cpu_possible(cpu, true);
 440 }
 441
 442 void __init __weak smp_setup_processor_id(void)
 443 {
 444 }
 445
 446 void __init __weak thread_info_cache_init(void)
 447 {
 448 }
 449
 450 /*
 451  * Set up kernel memory allocators
 452  */
 453 static void __init mm_init(void)
 454 {
 455         /*
 456          * page_cgroup requires countinous pages as memmap
 457          * and it's bigger than MAX_ORDER unless SPARSEMEM.
 458          */
 459         page_cgroup_init_flatmem();
 460         mem_init();
 461         kmem_cache_init();
 462         percpu_init_late();
 463         pgtable_cache_init();
 464         vmalloc_init();
 465 }
 466
 467 asmlinkage void __init start_kernel(void)
 468 {
 469         char * command_line;
 470         extern const struct kernel_param __start___param[], __stop___param[];
 471
 472         smp_setup_processor_id();
 473
 474         /*
 475          * Need to run as early as possible, to initialize the
 476          * lockdep hash:
 477          */
 478         lockdep_init();
 479         debug_objects_early_init();
 480
 481         /*
 482          * Set up the the initial canary ASAP:
 483          */
 484         boot_init_stack_canary();
 485
 486         cgroup_init_early();
 487
 488         local_irq_disable();
 489         early_boot_irqs_disabled = true;
 490
 491 /*
 492  * Interrupts are still disabled. Do necessary setups, then
 493  * enable them
 494  */
 495         tick_init();
 496         boot_cpu_init();
 497         page_address_init();
 498         printk(KERN_NOTICE "%s", linux_banner);
 499         setup_arch(&command_line);
 500         mm_init_owner(&init_mm, &init_task);
 501         mm_init_cpumask(&init_mm);
 502         setup_command_line(command_line);
 503         setup_nr_cpu_ids();
 504         setup_per_cpu_areas();
 505         smp_prepare_boot_cpu(); /* arch-specific boot-cpu hooks */
 506
 507         build_all_zonelists(NULL);
 508         page_alloc_init();
 509
 510         printk(KERN_NOTICE "Kernel command line: %s\n", boot_command_line);
 511         parse_early_param();
 512         parse_args("Booting kernel", static_command_line, __start___param,
 513                    __stop___param - __start___param,
 514                    &unknown_bootoption);
 515
 516         jump_label_init();
 517
 518         /*
 519          * These use large bootmem allocations and must precede
 520          * kmem_cache_init()
 521          */
 522         setup_log_buf(0);
 523         pidhash_init();
 524         vfs_caches_init_early();
 525         sort_main_extable();
 526         trap_init();
 527         mm_init();
 528
 529         /*
 530          * Set up the scheduler prior starting any interrupts (such as the
 531          * timer interrupt). Full topology setup happens at smp_init()
 532          * time - but meanwhile we still have a functioning scheduler.
 533          */
 534         sched_init();
 535         /*
 536          * Disable preemption - early bootup scheduling is extremely
 537          * fragile until we cpu_idle() for the first time.
 538          */
 539         preempt_disable();
 540         if (!irqs_disabled()) {
 541                 printk(KERN_WARNING "start_kernel(): bug: interrupts were "
 542                                 "enabled *very* early, fixing it\n");
 543                 local_irq_disable();
 544         }
 545         idr_init_cache();
 546         perf_event_init();
 547         rcu_init();
 548         radix_tree_init();
 549         /* init some links before init_ISA_irqs() */
 550         early_irq_init();
 551         init_IRQ();
 552         prio_tree_init();
 553         init_timers();
 554         hrtimers_init();
 555         softirq_init();
 556         timekeeping_init();
 557         time_init();
 558         profile_init();
 559         call_function_init();
 560         if (!irqs_disabled())
 561                 printk(KERN_CRIT "start_kernel(): bug: interrupts were "
 562                                  "enabled early\n");
 563         early_boot_irqs_disabled = false;
 564         local_irq_enable();
 565
 566         kmem_cache_init_late();
 567
 568         /*
 569          * HACK ALERT! This is early. We're enabling the console before
 570          * we've done PCI setups etc, and console_init() must be aware of
 571          * this. But we do want output early, in case something goes wrong.
 572          */
 573         console_init();
 574         if (panic_later)
 575                 panic(panic_later, panic_param);
 576
 577         lockdep_info();
 578
 579         /*
 580          * Need to run this when irqs are enabled, because it wants
 581          * to self-test [hard/soft]-irqs on/off lock inversion bugs
 582          * too:
 583          */
 584         locking_selftest();
 585
 586 #ifdef CONFIG_BLK_DEV_INITRD
 587         if (initrd_start && !initrd_below_start_ok &&
 588             page_to_pfn(virt_to_page((void *)initrd_start)) < min_low_pfn) {
 589                 printk(KERN_CRIT "initrd overwritten (0x%08lx < 0x%08lx) - "
 590                     "disabling it.\n",
 591                     page_to_pfn(virt_to_page((void *)initrd_start)),
 592                     min_low_pfn);
 593                 initrd_start = 0;
 594         }
 595 #endif
 596         page_cgroup_init();
 597         enable_debug_pagealloc();
 598         debug_objects_mem_init();
 599         kmemleak_init();
 600         setup_per_cpu_pageset();
 601         numa_policy_init();
 602         if (late_time_init)
 603                 late_time_init();
 604         sched_clock_init();
 605         calibrate_delay();
 606         pidmap_init();
 607         anon_vma_init();
 608 #ifdef CONFIG_X86
 609         if (efi_enabled)
 610                 efi_enter_virtual_mode();
 611 #endif
 612         thread_info_cache_init();
 613         cred_init();
 614         fork_init(totalram_pages);
 615         proc_caches_init();
 616         buffer_init();
 617         key_init();
 618         security_init();
 619         dbg_late_init();
 620         vfs_caches_init(totalram_pages);
 621         signals_init();
 622         /* rootfs populating might need page-writeback */
 623         page_writeback_init();
 624 #ifdef CONFIG_PROC_FS
 625         proc_root_init();
 626 #endif
 627         cgroup_init();
 628         cpuset_init();
 629         taskstats_init_early();
 630         delayacct_init();
 631
 632         check_bugs();
 633
 634         acpi_early_init(); /* before LAPIC and SMP init */
 635         sfi_init_late();
 636
 637         ftrace_init();
 638
 639         /* Do the rest non-__init'ed, we're now alive */
 640         rest_init();
 641 }
 642
 643 /* Call all constructor functions linked into the kernel. */
 644 static void __init do_ctors(void)
 645 {
 646 #ifdef CONFIG_CONSTRUCTORS
 647         ctor_fn_t *fn = (ctor_fn_t *) __ctors_start;
 648
 649         for (; fn < (ctor_fn_t *) __ctors_end; fn++)
 650                 (*fn)();
 651 #endif
 652 }
 653
 654 int initcall_debug;
 655 core_param(initcall_debug, initcall_debug, bool, 0644);
 656
 657 static char msgbuf[64];
 658
 659 static int __init_or_module do_one_initcall_debug(initcall_t fn)
 660 {
 661         ktime_t calltime, delta, rettime;
 662         unsigned long long duration;
 663         int ret;
 664
 665         printk(KERN_DEBUG "calling  %pF @ %i\n", fn, task_pid_nr(current));
 666         calltime = ktime_get();
 667         ret = fn();
 668         rettime = ktime_get();
 669         delta = ktime_sub(rettime, calltime);
 670         duration = (unsigned long long) ktime_to_ns(delta) >> 10;
 671         printk(KERN_DEBUG "initcall %pF returned %d after %lld usecs\n", fn,
 672                 ret, duration);
 673
 674         return ret;
 675 }
 676
 677 int __init_or_module do_one_initcall(initcall_t fn)
 678 {
 679         int count = preempt_count();
 680         int ret;
 681
 682         if (initcall_debug)
 683                 ret = do_one_initcall_debug(fn);
 684         else
 685                 ret = fn();
 686
 687         msgbuf[0] = 0;
 688
 689         if (ret && ret != -ENODEV && initcall_debug)
 690                 sprintf(msgbuf, "error code %d ", ret);
 691
 692         if (preempt_count() != count) {
 693                 strlcat(msgbuf, "preemption imbalance ", sizeof(msgbuf));
 694                 preempt_count() = count;
 695         }
 696         if (irqs_disabled()) {
 697                 strlcat(msgbuf, "disabled interrupts ", sizeof(msgbuf));
 698                 local_irq_enable();
 699         }
 700         if (msgbuf[0]) {
 701                 printk("initcall %pF returned with %s\n", fn, msgbuf);
 702         }
 703
 704         return ret;
 705 }
 706
 707
 708 extern initcall_t __initcall_start[], __initcall_end[], __early_initcall_end[];
 709
 710 static void __init do_initcalls(void)
 711 {
 712         initcall_t *fn;
 713
 714         for (fn = __early_initcall_end; fn < __initcall_end; fn++)
 715                 do_one_initcall(*fn);
 716 }
 717
 718 /*
 719  * Ok, the machine is now initialized. None of the devices
 720  * have been touched yet, but the CPU subsystem is up and
 721  * running, and memory and process management works.
 722  *
 723  * Now we can finally start doing some real work..
 724  */
 725 static void __init do_basic_setup(void)
 726 {
 727         cpuset_init_smp();
 728         usermodehelper_init();
 729         shmem_init();
 730         driver_init();
 731         init_irq_proc();
 732         do_ctors();
 733         usermodehelper_enable();
 734         do_initcalls();
 735 }
 736
 737 static void __init do_pre_smp_initcalls(void)
 738 {
 739         initcall_t *fn;
 740
 741         for (fn = __initcall_start; fn < __early_initcall_end; fn++)
 742                 do_one_initcall(*fn);
 743 }
 744
 745 static void run_init_process(const char *init_filename)
 746 {
 747         argv_init[0] = init_filename;
 748         kernel_execve(init_filename, argv_init, envp_init);
 749 }
 750
 751 /* This is a non __init function. Force it to be noinline otherwise gcc
 752  * makes it inline to init() and it becomes part of init.text section
 753  */
 754 static noinline int init_post(void)
 755 {
 756         /* need to finish all async __init code before freeing the memory */
 757         async_synchronize_full();
 758         free_initmem();
 759         mark_rodata_ro();
 760         system_state = SYSTEM_RUNNING;
 761         numa_default_policy();
 762
 763
 764         current->signal->flags |= SIGNAL_UNKILLABLE;
 765
 766         if (ramdisk_execute_command) {
 767                 run_init_process(ramdisk_execute_command);
 768                 printk(KERN_WARNING "Failed to execute %s\n",
 769                                 ramdisk_execute_command);
 770         }
 771
 772         /*
 773          * We try each of these until one succeeds.
 774          *
 775          * The Bourne shell can be used instead of init if we are
 776          * trying to recover a really broken machine.
 777          */
 778         if (execute_command) {
 779                 run_init_process(execute_command);
 780                 printk(KERN_WARNING "Failed to execute %s.  Attempting "
 781                                         "defaults...\n", execute_command);
 782         }
 783         run_init_process("/sbin/init");
 784         run_init_process("/etc/init");
 785         run_init_process("/bin/init");
 786         run_init_process("/bin/sh");
 787
 788         panic("No init found.  Try passing init= option to kernel. "
 789               "See Linux Documentation/init.txt for guidance.");
 790 }
 791
 792 static int __init kernel_init(void * unused)
 793 {
 794         /*
 795          * Wait until kthreadd is all set-up.
 796          */
 797         wait_for_completion(&kthreadd_done);
 798
 799         /* Now the scheduler is fully set up and can do blocking allocations */
 800         gfp_allowed_mask = __GFP_BITS_MASK;
 801
 802         /*
 803          * init can allocate pages on any node
 804          */
 805         set_mems_allowed(node_states[N_HIGH_MEMORY]);
 806         /*
 807          * init can run on any cpu.
 808          */
 809         set_cpus_allowed_ptr(current, cpu_all_mask);
 810
 811         cad_pid = task_pid(current);
 812
 813         smp_prepare_cpus(setup_max_cpus);
 814
 815         do_pre_smp_initcalls();
 816         lockup_detector_init();
 817
 818         smp_init();
 819         sched_init_smp();
 820
 821         do_basic_setup();
 822
 823         /* Open the /dev/console on the rootfs, this should never fail */
 824         if (sys_open((const char __user *) "/dev/console", O_RDWR, 0) < 0)
 825                 printk(KERN_WARNING "Warning: unable to open an initial console.\n");
 826
 827         (void) sys_dup(0);
 828         (void) sys_dup(0);
 829         /*
 830          * check if there is an early userspace init.  If yes, let it do all
 831          * the work
 832          */
 833
 834         if (!ramdisk_execute_command)
 835                 ramdisk_execute_command = "/init";
 836
 837         if (sys_access((const char __user *) ramdisk_execute_command, 0) != 0) {
 838                 ramdisk_execute_command = NULL;
 839                 prepare_namespace();
 840         }
 841
 842         /*
 843          * Ok, we have completed the initial bootup, and
 844          * we're essentially up and running. Get rid of the
 845          * initmem segments and start the user-mode stuff..
 846          */
 847
 848         init_post();
 849         return 0;
 850 }