init_main.c revision 1.400
1 /* $NetBSD: init_main.c,v 1.400 2009/09/16 15:23:04 pooka Exp $ */ 2 3 /*- 4 * Copyright (c) 2008, 2009 The NetBSD Foundation, Inc. 5 * All rights reserved. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 10 * 1. Redistributions of source code must retain the above copyright 11 * notice, this list of conditions and the following disclaimer. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice, this list of conditions and the following disclaimer in the 14 * documentation and/or other materials provided with the distribution. 15 * 16 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 17 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 18 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 19 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 20 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 21 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 22 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 23 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 24 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 25 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 26 * POSSIBILITY OF SUCH DAMAGE. 27 */ 28 29 /* 30 * Copyright (c) 1982, 1986, 1989, 1991, 1992, 1993 31 * The Regents of the University of California. All rights reserved. 32 * (c) UNIX System Laboratories, Inc. 33 * All or some portions of this file are derived from material licensed 34 * to the University of California by American Telephone and Telegraph 35 * Co. or Unix System Laboratories, Inc. and are reproduced herein with 36 * the permission of UNIX System Laboratories, Inc. 37 * 38 * Redistribution and use in source and binary forms, with or without 39 * modification, are permitted provided that the following conditions 40 * are met: 41 * 1. Redistributions of source code must retain the above copyright 42 * notice, this list of conditions and the following disclaimer. 43 * 2. Redistributions in binary form must reproduce the above copyright 44 * notice, this list of conditions and the following disclaimer in the 45 * documentation and/or other materials provided with the distribution. 46 * 3. Neither the name of the University nor the names of its contributors 47 * may be used to endorse or promote products derived from this software 48 * without specific prior written permission. 49 * 50 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 51 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 52 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 53 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 54 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 55 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 56 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 57 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 58 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 59 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 60 * SUCH DAMAGE. 61 * 62 * @(#)init_main.c 8.16 (Berkeley) 5/14/95 63 */ 64 65 /* 66 * Copyright (c) 1995 Christopher G. Demetriou. All rights reserved. 67 * 68 * Redistribution and use in source and binary forms, with or without 69 * modification, are permitted provided that the following conditions 70 * are met: 71 * 1. Redistributions of source code must retain the above copyright 72 * notice, this list of conditions and the following disclaimer. 73 * 2. Redistributions in binary form must reproduce the above copyright 74 * notice, this list of conditions and the following disclaimer in the 75 * documentation and/or other materials provided with the distribution. 76 * 3. All advertising materials mentioning features or use of this software 77 * must display the following acknowledgement: 78 * This product includes software developed by the University of 79 * California, Berkeley and its contributors. 80 * 4. Neither the name of the University nor the names of its contributors 81 * may be used to endorse or promote products derived from this software 82 * without specific prior written permission. 83 * 84 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 85 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 86 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 87 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 88 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 89 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 90 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 91 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 92 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 93 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 94 * SUCH DAMAGE. 95 * 96 * @(#)init_main.c 8.16 (Berkeley) 5/14/95 97 */ 98 99 #include <sys/cdefs.h> 100 __KERNEL_RCSID(0, "$NetBSD: init_main.c,v 1.400 2009/09/16 15:23:04 pooka Exp $"); 101 102 #include "opt_ddb.h" 103 #include "opt_ipsec.h" 104 #include "opt_modular.h" 105 #include "opt_ntp.h" 106 #include "opt_pipe.h" 107 #include "opt_sa.h" 108 #include "opt_syscall_debug.h" 109 #include "opt_sysv.h" 110 #include "opt_fileassoc.h" 111 #include "opt_ktrace.h" 112 #include "opt_pax.h" 113 #include "opt_compat_netbsd.h" 114 #include "opt_wapbl.h" 115 116 #include "ksyms.h" 117 #include "rnd.h" 118 #include "sysmon_envsys.h" 119 #include "sysmon_power.h" 120 #include "sysmon_taskq.h" 121 #include "sysmon_wdog.h" 122 #include "veriexec.h" 123 124 #include <sys/param.h> 125 #include <sys/acct.h> 126 #include <sys/filedesc.h> 127 #include <sys/file.h> 128 #include <sys/errno.h> 129 #include <sys/callout.h> 130 #include <sys/cpu.h> 131 #include <sys/kernel.h> 132 #include <sys/mount.h> 133 #include <sys/proc.h> 134 #include <sys/kthread.h> 135 #include <sys/resourcevar.h> 136 #include <sys/signalvar.h> 137 #include <sys/systm.h> 138 #include <sys/vnode.h> 139 #include <sys/fstrans.h> 140 #include <sys/tty.h> 141 #include <sys/conf.h> 142 #include <sys/disklabel.h> 143 #include <sys/buf.h> 144 #include <sys/device.h> 145 #include <sys/exec.h> 146 #include <sys/socketvar.h> 147 #include <sys/protosw.h> 148 #include <sys/percpu.h> 149 #include <sys/pset.h> 150 #include <sys/sysctl.h> 151 #include <sys/reboot.h> 152 #include <sys/user.h> 153 #include <sys/sysctl.h> 154 #include <sys/event.h> 155 #include <sys/mbuf.h> 156 #include <sys/sched.h> 157 #include <sys/sleepq.h> 158 #include <sys/iostat.h> 159 #include <sys/vmem.h> 160 #include <sys/uuid.h> 161 #include <sys/extent.h> 162 #include <sys/disk.h> 163 #include <sys/msgbuf.h> 164 #include <sys/module.h> 165 #include <sys/event.h> 166 #include <sys/lockf.h> 167 #include <sys/once.h> 168 #include <sys/ksyms.h> 169 #include <sys/uidinfo.h> 170 #include <sys/kprintf.h> 171 #ifdef FAST_IPSEC 172 #include <netipsec/ipsec.h> 173 #endif 174 #ifdef SYSVSHM 175 #include <sys/shm.h> 176 #endif 177 #ifdef SYSVSEM 178 #include <sys/sem.h> 179 #endif 180 #ifdef SYSVMSG 181 #include <sys/msg.h> 182 #endif 183 #include <sys/domain.h> 184 #include <sys/namei.h> 185 #if NRND > 0 186 #include <sys/rnd.h> 187 #endif 188 #include <sys/pipe.h> 189 #if NVERIEXEC > 0 190 #include <sys/verified_exec.h> 191 #endif /* NVERIEXEC > 0 */ 192 #ifdef KTRACE 193 #include <sys/ktrace.h> 194 #endif 195 #include <sys/kauth.h> 196 #ifdef WAPBL 197 #include <sys/wapbl.h> 198 #endif 199 #ifdef KERN_SA 200 #include <sys/savar.h> 201 #endif 202 #include <net80211/ieee80211_netbsd.h> 203 204 #include <sys/syscall.h> 205 #include <sys/syscallargs.h> 206 207 #if defined(PAX_MPROTECT) || defined(PAX_SEGVGUARD) || defined(PAX_ASLR) 208 #include <sys/pax.h> 209 #endif /* PAX_MPROTECT || PAX_SEGVGUARD || PAX_ASLR */ 210 211 #include <ufs/ufs/quota.h> 212 213 #include <miscfs/genfs/genfs.h> 214 #include <miscfs/syncfs/syncfs.h> 215 216 #include <sys/cpu.h> 217 218 #include <uvm/uvm.h> 219 220 #if NSYSMON_TASKQ > 0 221 #include <dev/sysmon/sysmon_taskq.h> 222 #endif 223 224 #include <dev/cons.h> 225 226 #if NSYSMON_ENVSYS > 0 || NSYSMON_POWER > 0 || NSYSMON_WDOG > 0 227 #include <dev/sysmon/sysmonvar.h> 228 #endif 229 230 #include <net/if.h> 231 #include <net/raw_cb.h> 232 233 #include <secmodel/secmodel.h> 234 235 #include <prop/proplib.h> 236 237 #ifdef COMPAT_50 238 #include <compat/sys/time.h> 239 struct timeval50 boottime50; 240 #endif 241 242 #ifdef _KERNEL_OPT 243 #include "opt_userconf.h" 244 #endif 245 #ifdef USERCONF 246 #include <sys/userconf.h> 247 #endif 248 249 extern struct proc proc0; 250 extern struct lwp lwp0; 251 extern struct cwdinfo cwdi0; 252 extern time_t rootfstime; 253 254 #ifndef curlwp 255 struct lwp *curlwp = &lwp0; 256 #endif 257 struct proc *initproc; 258 259 struct vnode *rootvp, *swapdev_vp; 260 int boothowto; 261 int cold = 1; /* still working on startup */ 262 struct timespec boottime; /* time at system startup - will only follow settime deltas */ 263 264 int start_init_exec; /* semaphore for start_init() */ 265 266 static void check_console(struct lwp *l); 267 static void start_init(void *); 268 static void configure(void); 269 static void configure2(void); 270 void main(void); 271 272 void __secmodel_none(void); 273 __weak_alias(secmodel_start,__secmodel_none); 274 void 275 __secmodel_none(void) 276 { 277 return; 278 } 279 280 /* 281 * System startup; initialize the world, create process 0, mount root 282 * filesystem, and fork to create init and pagedaemon. Most of the 283 * hard work is done in the lower-level initialization routines including 284 * startup(), which does memory initialization and autoconfiguration. 285 */ 286 void 287 main(void) 288 { 289 struct timespec time; 290 struct lwp *l; 291 struct proc *p; 292 int s, error; 293 #ifdef NVNODE_IMPLICIT 294 int usevnodes; 295 #endif 296 CPU_INFO_ITERATOR cii; 297 struct cpu_info *ci; 298 299 l = &lwp0; 300 #ifndef LWP0_CPU_INFO 301 l->l_cpu = curcpu(); 302 #endif 303 l->l_pflag |= LP_RUNNING; 304 305 /* 306 * Attempt to find console and initialize 307 * in case of early panic or other messages. 308 */ 309 consinit(); 310 311 kernel_lock_init(); 312 once_init(); 313 mutex_init(&cpu_lock, MUTEX_DEFAULT, IPL_NONE); 314 315 /* Initialize the device switch tables. */ 316 devsw_init(); 317 318 uvm_init(); 319 320 prop_kern_init(); 321 322 #if ((NKSYMS > 0) || (NDDB > 0) || (NMODULAR > 0)) 323 ksyms_init(); 324 #endif 325 kprintf_init(); 326 327 percpu_init(); 328 329 /* Initialize lock caches. */ 330 mutex_obj_init(); 331 rw_obj_init(); 332 333 /* Initialize the extent manager. */ 334 extent_init(); 335 336 /* Do machine-dependent initialization. */ 337 cpu_startup(); 338 339 /* Initialize the sysctl subsystem. */ 340 sysctl_init(); 341 342 /* Initialize callouts, part 1. */ 343 callout_startup(); 344 345 /* 346 * Initialize the kernel authorization subsystem and start the 347 * default security model, if any. We need to do this early 348 * enough so that subsystems relying on any of the aforementioned 349 * can work properly. Since the security model may dictate the 350 * credential inheritance policy, it is needed at least before 351 * any process is created, specifically proc0. 352 */ 353 kauth_init(); 354 secmodel_start(); 355 356 /* Initialize the buffer cache */ 357 bufinit(); 358 359 /* Initialize sockets. */ 360 soinit(); 361 362 /* 363 * The following things must be done before autoconfiguration. 364 */ 365 evcnt_init(); /* initialize event counters */ 366 #if NRND > 0 367 rnd_init(); /* initialize random number generator */ 368 #endif 369 370 /* Initialize process and pgrp structures. */ 371 #ifdef KERN_SA 372 sa_init(); 373 #endif 374 procinit(); 375 lwpinit(); 376 377 /* Initialize signal-related data structures. */ 378 signal_init(); 379 380 /* Initialize resource management. */ 381 resource_init(); 382 383 /* Create process 0 (the swapper). */ 384 proc0_init(); 385 386 /* Disable preemption during boot. */ 387 kpreempt_disable(); 388 389 /* Initialize the UID hash table. */ 390 uid_init(); 391 392 /* Charge root for one process. */ 393 (void)chgproccnt(0, 1); 394 395 /* Initialize timekeeping. */ 396 time_init(); 397 398 /* Initialize the run queues, turnstiles and sleep queues. */ 399 sched_rqinit(); 400 turnstile_init(); 401 sleeptab_init(&sleeptab); 402 403 /* Initialize processor-sets */ 404 psets_init(); 405 406 /* MI initialization of the boot cpu */ 407 error = mi_cpu_attach(curcpu()); 408 KASSERT(error == 0); 409 410 /* Initialize timekeeping, part 2. */ 411 time_init2(); 412 413 /* 414 * Initialize mbuf's. Do this now because we might attempt to 415 * allocate mbufs or mbuf clusters during autoconfiguration. 416 */ 417 mbinit(); 418 419 /* Initialize I/O statistics. */ 420 iostat_init(); 421 422 /* Initialize the log device. */ 423 loginit(); 424 425 /* Start module system. */ 426 module_init(); 427 428 /* Initialize the file systems. */ 429 #ifdef NVNODE_IMPLICIT 430 /* 431 * If maximum number of vnodes in namei vnode cache is not explicitly 432 * defined in kernel config, adjust the number such as we use roughly 433 * 10% of memory for vnodes and associated data structures in the 434 * assumed worst case. Do not provide fewer than NVNODE vnodes. 435 */ 436 usevnodes = 437 calc_cache_size(kernel_map, 10, VNODE_VA_MAXPCT) / VNODE_COST; 438 if (usevnodes > desiredvnodes) 439 desiredvnodes = usevnodes; 440 #endif 441 vfsinit(); 442 lf_init(); 443 444 /* Initialize fstrans. */ 445 fstrans_init(); 446 447 /* Initialize the file descriptor system. */ 448 fd_sys_init(); 449 450 /* Initialize cwd structures */ 451 cwd_sys_init(); 452 453 /* Initialize kqueue. */ 454 kqueue_init(); 455 456 /* Initialize the system monitor subsystems. */ 457 #if NSYSMON_TASKQ > 0 458 sysmon_task_queue_preinit(); 459 #endif 460 461 #if NSYSMON_ENVSYS > 0 462 sysmon_envsys_init(); 463 #endif 464 465 #if NSYSMON_POWER > 0 466 sysmon_power_init(); 467 #endif 468 469 #if NSYSMON_WDOG > 0 470 sysmon_wdog_init(); 471 #endif 472 473 inittimecounter(); 474 ntp_init(); 475 476 /* Initialize tty subsystem. */ 477 tty_init(); 478 ttyldisc_init(); 479 480 /* Initialize the buffer cache, part 2. */ 481 bufinit2(); 482 483 /* Initialize the disk wedge subsystem. */ 484 dkwedge_init(); 485 486 /* Initialize interfaces. */ 487 ifinit1(); 488 489 /* Configure the system hardware. This will enable interrupts. */ 490 configure(); 491 492 ssp_init(); 493 494 configure2(); 495 /* Now timer is working. Enable preemption. */ 496 kpreempt_enable(); 497 498 ubc_init(); /* must be after autoconfig */ 499 500 #ifdef SYSVSHM 501 /* Initialize System V style shared memory. */ 502 shminit(); 503 #endif 504 505 #ifdef SYSVSEM 506 /* Initialize System V style semaphores. */ 507 seminit(); 508 #endif 509 510 #ifdef SYSVMSG 511 /* Initialize System V style message queues. */ 512 msginit(); 513 #endif 514 515 #if NVERIEXEC > 0 516 /* 517 * Initialise the Veriexec subsystem. 518 */ 519 veriexec_init(); 520 #endif /* NVERIEXEC > 0 */ 521 522 #if defined(PAX_MPROTECT) || defined(PAX_SEGVGUARD) || defined(PAX_ASLR) 523 pax_init(); 524 #endif /* PAX_MPROTECT || PAX_SEGVGUARD || PAX_ASLR */ 525 526 #ifdef FAST_IPSEC 527 /* Attach network crypto subsystem */ 528 ipsec_attach(); 529 #endif 530 531 /* 532 * Initialize protocols. Block reception of incoming packets 533 * until everything is ready. 534 */ 535 s = splnet(); 536 ifinit(); 537 domaininit(true); 538 if_attachdomain(); 539 splx(s); 540 541 #ifdef GPROF 542 /* Initialize kernel profiling. */ 543 kmstartup(); 544 #endif 545 546 /* Initialize system accounting. */ 547 acct_init(); 548 549 #ifndef PIPE_SOCKETPAIR 550 /* Initialize pipes. */ 551 pipe_init(); 552 #endif 553 554 #ifdef KTRACE 555 /* Initialize ktrace. */ 556 ktrinit(); 557 #endif 558 559 /* Initialize the UUID system calls. */ 560 uuid_init(); 561 562 #ifdef WAPBL 563 /* Initialize write-ahead physical block logging. */ 564 wapbl_init(); 565 #endif 566 567 /* 568 * Create process 1 (init(8)). We do this now, as Unix has 569 * historically had init be process 1, and changing this would 570 * probably upset a lot of people. 571 * 572 * Note that process 1 won't immediately exec init(8), but will 573 * wait for us to inform it that the root file system has been 574 * mounted. 575 */ 576 if (fork1(l, 0, SIGCHLD, NULL, 0, start_init, NULL, NULL, &initproc)) 577 panic("fork init"); 578 579 /* 580 * Load any remaining builtin modules, and hand back temporary 581 * storage to the VM system. 582 */ 583 module_init_class(MODULE_CLASS_ANY); 584 585 /* 586 * Finalize configuration now that all real devices have been 587 * found. This needs to be done before the root device is 588 * selected, since finalization may create the root device. 589 */ 590 config_finalize(); 591 592 sysctl_finalize(); 593 594 /* 595 * Now that autoconfiguration has completed, we can determine 596 * the root and dump devices. 597 */ 598 cpu_rootconf(); 599 cpu_dumpconf(); 600 601 /* Mount the root file system. */ 602 do { 603 domountroothook(); 604 if ((error = vfs_mountroot())) { 605 printf("cannot mount root, error = %d\n", error); 606 boothowto |= RB_ASKNAME; 607 setroot(root_device, 608 (rootdev != NODEV) ? DISKPART(rootdev) : 0); 609 } 610 } while (error != 0); 611 mountroothook_destroy(); 612 613 /* 614 * Initialise the time-of-day clock, passing the time recorded 615 * in the root filesystem (if any) for use by systems that 616 * don't have a non-volatile time-of-day device. 617 */ 618 inittodr(rootfstime); 619 620 CIRCLEQ_FIRST(&mountlist)->mnt_flag |= MNT_ROOTFS; 621 CIRCLEQ_FIRST(&mountlist)->mnt_op->vfs_refcount++; 622 623 /* 624 * Get the vnode for '/'. Set filedesc0.fd_fd.fd_cdir to 625 * reference it. 626 */ 627 error = VFS_ROOT(CIRCLEQ_FIRST(&mountlist), &rootvnode); 628 if (error) 629 panic("cannot find root vnode, error=%d", error); 630 cwdi0.cwdi_cdir = rootvnode; 631 VREF(cwdi0.cwdi_cdir); 632 VOP_UNLOCK(rootvnode, 0); 633 cwdi0.cwdi_rdir = NULL; 634 635 /* 636 * Now that root is mounted, we can fixup initproc's CWD 637 * info. All other processes are kthreads, which merely 638 * share proc0's CWD info. 639 */ 640 initproc->p_cwdi->cwdi_cdir = rootvnode; 641 VREF(initproc->p_cwdi->cwdi_cdir); 642 initproc->p_cwdi->cwdi_rdir = NULL; 643 644 /* 645 * Now can look at time, having had a chance to verify the time 646 * from the file system. Reset l->l_rtime as it may have been 647 * munched in mi_switch() after the time got set. 648 */ 649 getnanotime(&time); 650 boottime = time; 651 #ifdef COMPAT_50 652 { 653 struct timeval tv; 654 TIMESPEC_TO_TIMEVAL(&tv, &time); 655 timeval_to_timeval50(&tv, &boottime50); 656 } 657 #endif 658 mutex_enter(proc_lock); 659 LIST_FOREACH(p, &allproc, p_list) { 660 KASSERT((p->p_flag & PK_MARKER) == 0); 661 mutex_enter(p->p_lock); 662 TIMESPEC_TO_TIMEVAL(&p->p_stats->p_start, &time); 663 LIST_FOREACH(l, &p->p_lwps, l_sibling) { 664 lwp_lock(l); 665 memset(&l->l_rtime, 0, sizeof(l->l_rtime)); 666 lwp_unlock(l); 667 } 668 mutex_exit(p->p_lock); 669 } 670 mutex_exit(proc_lock); 671 binuptime(&curlwp->l_stime); 672 673 for (CPU_INFO_FOREACH(cii, ci)) { 674 ci->ci_schedstate.spc_lastmod = time_second; 675 } 676 677 /* Create the pageout daemon kernel thread. */ 678 uvm_swap_init(); 679 if (kthread_create(PRI_PGDAEMON, KTHREAD_MPSAFE, NULL, uvm_pageout, 680 NULL, NULL, "pgdaemon")) 681 panic("fork pagedaemon"); 682 683 /* Create the filesystem syncer kernel thread. */ 684 if (kthread_create(PRI_IOFLUSH, KTHREAD_MPSAFE, NULL, sched_sync, 685 NULL, NULL, "ioflush")) 686 panic("fork syncer"); 687 688 /* Create the aiodone daemon kernel thread. */ 689 if (workqueue_create(&uvm.aiodone_queue, "aiodoned", 690 uvm_aiodone_worker, NULL, PRI_VM, IPL_NONE, WQ_MPSAFE)) 691 panic("fork aiodoned"); 692 693 vmem_rehash_start(); 694 695 /* Initialize exec structures */ 696 exec_init(1); 697 698 /* 699 * Okay, now we can let init(8) exec! It's off to userland! 700 */ 701 mutex_enter(proc_lock); 702 start_init_exec = 1; 703 cv_broadcast(&lbolt); 704 mutex_exit(proc_lock); 705 706 /* The scheduler is an infinite loop. */ 707 uvm_scheduler(); 708 /* NOTREACHED */ 709 } 710 711 /* 712 * Configure the system's hardware. 713 */ 714 static void 715 configure(void) 716 { 717 718 /* Initialize autoconf data structures. */ 719 config_init(); 720 /* 721 * XXX 722 * callout_setfunc() requires mutex(9) so it can't be in config_init() 723 * on amiga and atari which use config_init() and autoconf(9) fucntions 724 * to initialize console devices. 725 */ 726 config_twiddle_init(); 727 728 pmf_init(); 729 #if NDRVCTL > 0 730 drvctl_init(); 731 #endif 732 733 #ifdef USERCONF 734 if (boothowto & RB_USERCONF) 735 user_config(); 736 #endif 737 738 if ((boothowto & (AB_SILENT|AB_VERBOSE)) == AB_SILENT) { 739 printf_nolog("Detecting hardware..."); 740 } 741 742 /* 743 * Do the machine-dependent portion of autoconfiguration. This 744 * sets the configuration machinery here in motion by "finding" 745 * the root bus. When this function returns, we expect interrupts 746 * to be enabled. 747 */ 748 cpu_configure(); 749 } 750 751 static void 752 configure2(void) 753 { 754 CPU_INFO_ITERATOR cii; 755 struct cpu_info *ci; 756 int s; 757 758 /* 759 * Now that we've found all the hardware, start the real time 760 * and statistics clocks. 761 */ 762 initclocks(); 763 764 cold = 0; /* clocks are running, we're warm now! */ 765 s = splsched(); 766 curcpu()->ci_schedstate.spc_flags |= SPCF_RUNNING; 767 splx(s); 768 769 /* Boot the secondary processors. */ 770 for (CPU_INFO_FOREACH(cii, ci)) { 771 uvm_cpu_attach(ci); 772 } 773 mp_online = true; 774 #if defined(MULTIPROCESSOR) 775 cpu_boot_secondary_processors(); 776 #endif 777 778 /* Setup the runqueues and scheduler. */ 779 runq_init(); 780 sched_init(); 781 782 /* 783 * Bus scans can make it appear as if the system has paused, so 784 * twiddle constantly while config_interrupts() jobs are running. 785 */ 786 config_twiddle_fn(NULL); 787 788 /* 789 * Create threads to call back and finish configuration for 790 * devices that want interrupts enabled. 791 */ 792 config_create_interruptthreads(); 793 794 /* Get the threads going and into any sleeps before continuing. */ 795 yield(); 796 } 797 798 static void 799 check_console(struct lwp *l) 800 { 801 struct vnode *vp; 802 int error; 803 804 error = namei_simple_kernel("/dev/console", 805 NSM_FOLLOW_NOEMULROOT, &vp); 806 if (error == 0) 807 vrele(vp); 808 else if (error == ENOENT) 809 printf("warning: no /dev/console\n"); 810 else 811 printf("warning: lookup /dev/console: error %d\n", error); 812 } 813 814 /* 815 * List of paths to try when searching for "init". 816 */ 817 static const char * const initpaths[] = { 818 "/sbin/init", 819 "/sbin/oinit", 820 "/sbin/init.bak", 821 NULL, 822 }; 823 824 /* 825 * Start the initial user process; try exec'ing each pathname in "initpaths". 826 * The program is invoked with one argument containing the boot flags. 827 */ 828 static void 829 start_init(void *arg) 830 { 831 struct lwp *l = arg; 832 struct proc *p = l->l_proc; 833 vaddr_t addr; 834 struct sys_execve_args /* { 835 syscallarg(const char *) path; 836 syscallarg(char * const *) argp; 837 syscallarg(char * const *) envp; 838 } */ args; 839 int options, i, error; 840 register_t retval[2]; 841 char flags[4], *flagsp; 842 const char *path, *slash; 843 char *ucp, **uap, *arg0, *arg1 = NULL; 844 char ipath[129]; 845 int ipx, len; 846 847 /* 848 * Now in process 1. 849 */ 850 strncpy(p->p_comm, "init", MAXCOMLEN); 851 852 /* 853 * Wait for main() to tell us that it's safe to exec. 854 */ 855 mutex_enter(proc_lock); 856 while (start_init_exec == 0) 857 cv_wait(&lbolt, proc_lock); 858 mutex_exit(proc_lock); 859 860 /* 861 * This is not the right way to do this. We really should 862 * hand-craft a descriptor onto /dev/console to hand to init, 863 * but that's a _lot_ more work, and the benefit from this easy 864 * hack makes up for the "good is the enemy of the best" effect. 865 */ 866 check_console(l); 867 868 /* 869 * Need just enough stack to hold the faked-up "execve()" arguments. 870 */ 871 addr = (vaddr_t)STACK_ALLOC(USRSTACK, PAGE_SIZE); 872 if (uvm_map(&p->p_vmspace->vm_map, &addr, PAGE_SIZE, 873 NULL, UVM_UNKNOWN_OFFSET, 0, 874 UVM_MAPFLAG(UVM_PROT_ALL, UVM_PROT_ALL, UVM_INH_COPY, 875 UVM_ADV_NORMAL, 876 UVM_FLAG_FIXED|UVM_FLAG_OVERLAY|UVM_FLAG_COPYONW)) != 0) 877 panic("init: couldn't allocate argument space"); 878 p->p_vmspace->vm_maxsaddr = (void *)STACK_MAX(addr, PAGE_SIZE); 879 880 ipx = 0; 881 while (1) { 882 if (boothowto & RB_ASKNAME) { 883 printf("init path"); 884 if (initpaths[ipx]) 885 printf(" (default %s)", initpaths[ipx]); 886 printf(": "); 887 len = cngetsn(ipath, sizeof(ipath)-1); 888 if (len == 4 && strcmp(ipath, "halt") == 0) { 889 cpu_reboot(RB_HALT, NULL); 890 } else if (len == 6 && strcmp(ipath, "reboot") == 0) { 891 cpu_reboot(0, NULL); 892 #if defined(DDB) 893 } else if (len == 3 && strcmp(ipath, "ddb") == 0) { 894 console_debugger(); 895 continue; 896 #endif 897 } else if (len > 0 && ipath[0] == '/') { 898 ipath[len] = '\0'; 899 path = ipath; 900 } else if (len == 0 && initpaths[ipx] != NULL) { 901 path = initpaths[ipx++]; 902 } else { 903 printf("use absolute path, "); 904 #if defined(DDB) 905 printf("\"ddb\", "); 906 #endif 907 printf("\"halt\", or \"reboot\"\n"); 908 continue; 909 } 910 } else { 911 if ((path = initpaths[ipx++]) == NULL) { 912 ipx = 0; 913 boothowto |= RB_ASKNAME; 914 continue; 915 } 916 } 917 918 ucp = (char *)USRSTACK; 919 920 /* 921 * Construct the boot flag argument. 922 */ 923 flagsp = flags; 924 *flagsp++ = '-'; 925 options = 0; 926 927 if (boothowto & RB_SINGLE) { 928 *flagsp++ = 's'; 929 options = 1; 930 } 931 #ifdef notyet 932 if (boothowto & RB_FASTBOOT) { 933 *flagsp++ = 'f'; 934 options = 1; 935 } 936 #endif 937 938 /* 939 * Move out the flags (arg 1), if necessary. 940 */ 941 if (options != 0) { 942 *flagsp++ = '\0'; 943 i = flagsp - flags; 944 #ifdef DEBUG 945 printf("init: copying out flags `%s' %d\n", flags, i); 946 #endif 947 arg1 = STACK_ALLOC(ucp, i); 948 ucp = STACK_MAX(arg1, i); 949 (void)copyout((void *)flags, arg1, i); 950 } 951 952 /* 953 * Move out the file name (also arg 0). 954 */ 955 i = strlen(path) + 1; 956 #ifdef DEBUG 957 printf("init: copying out path `%s' %d\n", path, i); 958 #else 959 if (boothowto & RB_ASKNAME || path != initpaths[0]) 960 printf("init: trying %s\n", path); 961 #endif 962 arg0 = STACK_ALLOC(ucp, i); 963 ucp = STACK_MAX(arg0, i); 964 (void)copyout(path, arg0, i); 965 966 /* 967 * Move out the arg pointers. 968 */ 969 ucp = (void *)STACK_ALIGN(ucp, ALIGNBYTES); 970 uap = (char **)STACK_ALLOC(ucp, sizeof(char *) * 3); 971 SCARG(&args, path) = arg0; 972 SCARG(&args, argp) = uap; 973 SCARG(&args, envp) = NULL; 974 slash = strrchr(path, '/'); 975 if (slash) 976 (void)suword((void *)uap++, 977 (long)arg0 + (slash + 1 - path)); 978 else 979 (void)suword((void *)uap++, (long)arg0); 980 if (options != 0) 981 (void)suword((void *)uap++, (long)arg1); 982 (void)suword((void *)uap++, 0); /* terminator */ 983 984 /* 985 * Now try to exec the program. If can't for any reason 986 * other than it doesn't exist, complain. 987 */ 988 error = sys_execve(l, &args, retval); 989 if (error == 0 || error == EJUSTRETURN) { 990 KERNEL_UNLOCK_LAST(l); 991 return; 992 } 993 printf("exec %s: error %d\n", path, error); 994 } 995 printf("init: not found\n"); 996 panic("no init"); 997 } 998 999 /* 1000 * calculate cache size from physmem and vm_map size. 1001 */ 1002 vaddr_t 1003 calc_cache_size(struct vm_map *map, int pct, int va_pct) 1004 { 1005 paddr_t t; 1006 1007 /* XXX should consider competing cache if any */ 1008 /* XXX should consider submaps */ 1009 t = (uintmax_t)physmem * pct / 100 * PAGE_SIZE; 1010 if (map != NULL) { 1011 vsize_t vsize; 1012 1013 vsize = vm_map_max(map) - vm_map_min(map); 1014 vsize = (uintmax_t)vsize * va_pct / 100; 1015 if (t > vsize) { 1016 t = vsize; 1017 } 1018 } 1019 return t; 1020 } 1021 1022 /* 1023 * Print the system start up banner. 1024 * 1025 * - Print a limited banner if AB_SILENT. 1026 * - Always send normal banner to the log. 1027 */ 1028 #define MEM_PBUFSIZE sizeof("99999 MB") 1029 1030 void 1031 banner(void) 1032 { 1033 static char notice[] = " Notice: this software is " 1034 "protected by copyright"; 1035 char pbuf[81]; 1036 void (*pr)(const char *, ...); 1037 int i; 1038 1039 if ((boothowto & AB_SILENT) != 0) { 1040 snprintf(pbuf, sizeof(pbuf), "%s %s (%s)", 1041 ostype, osrelease, kernel_ident); 1042 printf_nolog("%s", pbuf); 1043 for (i = 80 - strlen(pbuf) - sizeof(notice); i > 0; i--) 1044 printf(" "); 1045 printf_nolog("%s\n", notice); 1046 pr = aprint_normal; 1047 } else { 1048 pr = printf; 1049 } 1050 1051 memset(pbuf, 0, sizeof(pbuf)); 1052 (*pr)("%s%s", copyright, version); 1053 format_bytes(pbuf, MEM_PBUFSIZE, ctob((uint64_t)physmem)); 1054 (*pr)("total memory = %s\n", pbuf); 1055 format_bytes(pbuf, MEM_PBUFSIZE, ctob((uint64_t)uvmexp.free)); 1056 (*pr)("avail memory = %s\n", pbuf); 1057 } 1058
Indexes created Sun Sep 21 20:09:37 GMT 2025