init_main.c revision 1.403
1 /* $NetBSD: init_main.c,v 1.403 2009/10/02 18:50:14 elad 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 1.403 2009/10/02 18:50:14 elad 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 "drvctl.h" 117 #include "ksyms.h" 118 #include "rnd.h" 119 #include "sysmon_envsys.h" 120 #include "sysmon_power.h" 121 #include "sysmon_taskq.h" 122 #include "sysmon_wdog.h" 123 #include "veriexec.h" 124 125 #include <sys/param.h> 126 #include <sys/acct.h> 127 #include <sys/filedesc.h> 128 #include <sys/file.h> 129 #include <sys/errno.h> 130 #include <sys/callout.h> 131 #include <sys/cpu.h> 132 #include <sys/kernel.h> 133 #include <sys/mount.h> 134 #include <sys/proc.h> 135 #include <sys/kthread.h> 136 #include <sys/resourcevar.h> 137 #include <sys/signalvar.h> 138 #include <sys/systm.h> 139 #include <sys/vnode.h> 140 #include <sys/fstrans.h> 141 #include <sys/tty.h> 142 #include <sys/conf.h> 143 #include <sys/disklabel.h> 144 #include <sys/buf.h> 145 #include <sys/device.h> 146 #include <sys/exec.h> 147 #include <sys/socketvar.h> 148 #include <sys/protosw.h> 149 #include <sys/percpu.h> 150 #include <sys/pset.h> 151 #include <sys/sysctl.h> 152 #include <sys/reboot.h> 153 #include <sys/user.h> 154 #include <sys/sysctl.h> 155 #include <sys/event.h> 156 #include <sys/mbuf.h> 157 #include <sys/sched.h> 158 #include <sys/sleepq.h> 159 #include <sys/iostat.h> 160 #include <sys/vmem.h> 161 #include <sys/uuid.h> 162 #include <sys/extent.h> 163 #include <sys/disk.h> 164 #include <sys/msgbuf.h> 165 #include <sys/module.h> 166 #include <sys/event.h> 167 #include <sys/lockf.h> 168 #include <sys/once.h> 169 #include <sys/ksyms.h> 170 #include <sys/uidinfo.h> 171 #include <sys/kprintf.h> 172 #ifdef FAST_IPSEC 173 #include <netipsec/ipsec.h> 174 #endif 175 #ifdef SYSVSHM 176 #include <sys/shm.h> 177 #endif 178 #ifdef SYSVSEM 179 #include <sys/sem.h> 180 #endif 181 #ifdef SYSVMSG 182 #include <sys/msg.h> 183 #endif 184 #include <sys/domain.h> 185 #include <sys/namei.h> 186 #if NRND > 0 187 #include <sys/rnd.h> 188 #endif 189 #include <sys/pipe.h> 190 #if NVERIEXEC > 0 191 #include <sys/verified_exec.h> 192 #endif /* NVERIEXEC > 0 */ 193 #ifdef KTRACE 194 #include <sys/ktrace.h> 195 #endif 196 #include <sys/kauth.h> 197 #ifdef WAPBL 198 #include <sys/wapbl.h> 199 #endif 200 #ifdef KERN_SA 201 #include <sys/savar.h> 202 #endif 203 #include <net80211/ieee80211_netbsd.h> 204 205 #include <sys/syscall.h> 206 #include <sys/syscallargs.h> 207 208 #if defined(PAX_MPROTECT) || defined(PAX_SEGVGUARD) || defined(PAX_ASLR) 209 #include <sys/pax.h> 210 #endif /* PAX_MPROTECT || PAX_SEGVGUARD || PAX_ASLR */ 211 212 #include <ufs/ufs/quota.h> 213 214 #include <miscfs/genfs/genfs.h> 215 #include <miscfs/syncfs/syncfs.h> 216 217 #include <sys/cpu.h> 218 219 #include <uvm/uvm.h> 220 221 #if NSYSMON_TASKQ > 0 222 #include <dev/sysmon/sysmon_taskq.h> 223 #endif 224 225 #include <dev/cons.h> 226 227 #if NSYSMON_ENVSYS > 0 || NSYSMON_POWER > 0 || NSYSMON_WDOG > 0 228 #include <dev/sysmon/sysmonvar.h> 229 #endif 230 231 #include <net/if.h> 232 #include <net/raw_cb.h> 233 234 #include <prop/proplib.h> 235 236 #ifdef COMPAT_50 237 #include <compat/sys/time.h> 238 struct timeval50 boottime50; 239 #endif 240 241 #ifdef _KERNEL_OPT 242 #include "opt_userconf.h" 243 #endif 244 #ifdef USERCONF 245 #include <sys/userconf.h> 246 #endif 247 248 extern struct proc proc0; 249 extern struct lwp lwp0; 250 extern struct cwdinfo cwdi0; 251 extern time_t rootfstime; 252 253 #ifndef curlwp 254 struct lwp *curlwp = &lwp0; 255 #endif 256 struct proc *initproc; 257 258 struct vnode *rootvp, *swapdev_vp; 259 int boothowto; 260 int cold = 1; /* still working on startup */ 261 struct timespec boottime; /* time at system startup - will only follow settime deltas */ 262 263 int start_init_exec; /* semaphore for start_init() */ 264 265 static void check_console(struct lwp *l); 266 static void start_init(void *); 267 static void configure(void); 268 static void configure2(void); 269 void main(void); 270 271 /* 272 * System startup; initialize the world, create process 0, mount root 273 * filesystem, and fork to create init and pagedaemon. Most of the 274 * hard work is done in the lower-level initialization routines including 275 * startup(), which does memory initialization and autoconfiguration. 276 */ 277 void 278 main(void) 279 { 280 struct timespec time; 281 struct lwp *l; 282 struct proc *p; 283 int s, error; 284 #ifdef NVNODE_IMPLICIT 285 int usevnodes; 286 #endif 287 CPU_INFO_ITERATOR cii; 288 struct cpu_info *ci; 289 290 l = &lwp0; 291 #ifndef LWP0_CPU_INFO 292 l->l_cpu = curcpu(); 293 #endif 294 l->l_pflag |= LP_RUNNING; 295 296 /* 297 * Attempt to find console and initialize 298 * in case of early panic or other messages. 299 */ 300 consinit(); 301 302 kernel_lock_init(); 303 once_init(); 304 mutex_init(&cpu_lock, MUTEX_DEFAULT, IPL_NONE); 305 306 /* Initialize the device switch tables. */ 307 devsw_init(); 308 309 uvm_init(); 310 311 prop_kern_init(); 312 313 #if ((NKSYMS > 0) || (NDDB > 0) || (NMODULAR > 0)) 314 ksyms_init(); 315 #endif 316 kprintf_init(); 317 318 percpu_init(); 319 320 /* Initialize lock caches. */ 321 mutex_obj_init(); 322 rw_obj_init(); 323 324 /* Initialize the extent manager. */ 325 extent_init(); 326 327 /* Do machine-dependent initialization. */ 328 cpu_startup(); 329 330 /* Initialize the sysctl subsystem. */ 331 sysctl_init(); 332 333 /* Initialize callouts, part 1. */ 334 callout_startup(); 335 336 /* Start module system. */ 337 module_init(); 338 339 /* 340 * Initialize the kernel authorization subsystem and start the 341 * default security model, if any. We need to do this early 342 * enough so that subsystems relying on any of the aforementioned 343 * can work properly. Since the security model may dictate the 344 * credential inheritance policy, it is needed at least before 345 * any process is created, specifically proc0. 346 */ 347 kauth_init(); 348 module_init_class(MODULE_CLASS_SECMODEL); 349 350 /* Initialize the buffer cache */ 351 bufinit(); 352 353 /* Initialize sockets. */ 354 soinit(); 355 356 /* 357 * The following things must be done before autoconfiguration. 358 */ 359 evcnt_init(); /* initialize event counters */ 360 #if NRND > 0 361 rnd_init(); /* initialize random number generator */ 362 #endif 363 364 /* Initialize process and pgrp structures. */ 365 #ifdef KERN_SA 366 sa_init(); 367 #endif 368 procinit(); 369 lwpinit(); 370 371 /* Initialize signal-related data structures. */ 372 signal_init(); 373 374 /* Initialize resource management. */ 375 resource_init(); 376 377 /* Create process 0 (the swapper). */ 378 proc0_init(); 379 380 /* Disable preemption during boot. */ 381 kpreempt_disable(); 382 383 /* Initialize the UID hash table. */ 384 uid_init(); 385 386 /* Charge root for one process. */ 387 (void)chgproccnt(0, 1); 388 389 /* Initialize timekeeping. */ 390 time_init(); 391 392 /* Initialize the run queues, turnstiles and sleep queues. */ 393 sched_rqinit(); 394 turnstile_init(); 395 sleeptab_init(&sleeptab); 396 397 /* Initialize processor-sets */ 398 psets_init(); 399 400 /* MI initialization of the boot cpu */ 401 error = mi_cpu_attach(curcpu()); 402 KASSERT(error == 0); 403 404 /* Initialize timekeeping, part 2. */ 405 time_init2(); 406 407 /* 408 * Initialize mbuf's. Do this now because we might attempt to 409 * allocate mbufs or mbuf clusters during autoconfiguration. 410 */ 411 mbinit(); 412 413 /* Initialize I/O statistics. */ 414 iostat_init(); 415 416 /* Initialize the log device. */ 417 loginit(); 418 419 /* Second part of module system initialization. */ 420 module_init2(); 421 422 /* Initialize the file systems. */ 423 #ifdef NVNODE_IMPLICIT 424 /* 425 * If maximum number of vnodes in namei vnode cache is not explicitly 426 * defined in kernel config, adjust the number such as we use roughly 427 * 10% of memory for vnodes and associated data structures in the 428 * assumed worst case. Do not provide fewer than NVNODE vnodes. 429 */ 430 usevnodes = 431 calc_cache_size(kernel_map, 10, VNODE_VA_MAXPCT) / VNODE_COST; 432 if (usevnodes > desiredvnodes) 433 desiredvnodes = usevnodes; 434 #endif 435 vfsinit(); 436 lf_init(); 437 438 /* Initialize fstrans. */ 439 fstrans_init(); 440 441 /* Initialize the file descriptor system. */ 442 fd_sys_init(); 443 444 /* Initialize cwd structures */ 445 cwd_sys_init(); 446 447 /* Initialize kqueue. */ 448 kqueue_init(); 449 450 /* Initialize the system monitor subsystems. */ 451 #if NSYSMON_TASKQ > 0 452 sysmon_task_queue_preinit(); 453 #endif 454 455 #if NSYSMON_ENVSYS > 0 456 sysmon_envsys_init(); 457 #endif 458 459 #if NSYSMON_POWER > 0 460 sysmon_power_init(); 461 #endif 462 463 #if NSYSMON_WDOG > 0 464 sysmon_wdog_init(); 465 #endif 466 467 inittimecounter(); 468 ntp_init(); 469 470 /* Initialize tty subsystem. */ 471 tty_init(); 472 ttyldisc_init(); 473 474 /* Initialize the buffer cache, part 2. */ 475 bufinit2(); 476 477 /* Initialize the disk wedge subsystem. */ 478 dkwedge_init(); 479 480 /* Initialize interfaces. */ 481 ifinit1(); 482 483 /* Configure the system hardware. This will enable interrupts. */ 484 configure(); 485 486 ssp_init(); 487 488 configure2(); 489 /* Now timer is working. Enable preemption. */ 490 kpreempt_enable(); 491 492 ubc_init(); /* must be after autoconfig */ 493 494 #ifdef SYSVSHM 495 /* Initialize System V style shared memory. */ 496 shminit(); 497 #endif 498 499 #ifdef SYSVSEM 500 /* Initialize System V style semaphores. */ 501 seminit(); 502 #endif 503 504 #ifdef SYSVMSG 505 /* Initialize System V style message queues. */ 506 msginit(); 507 #endif 508 509 #if NVERIEXEC > 0 510 /* 511 * Initialise the Veriexec subsystem. 512 */ 513 veriexec_init(); 514 #endif /* NVERIEXEC > 0 */ 515 516 #if defined(PAX_MPROTECT) || defined(PAX_SEGVGUARD) || defined(PAX_ASLR) 517 pax_init(); 518 #endif /* PAX_MPROTECT || PAX_SEGVGUARD || PAX_ASLR */ 519 520 #ifdef FAST_IPSEC 521 /* Attach network crypto subsystem */ 522 ipsec_attach(); 523 #endif 524 525 /* 526 * Initialize protocols. Block reception of incoming packets 527 * until everything is ready. 528 */ 529 s = splnet(); 530 ifinit(); 531 domaininit(true); 532 if_attachdomain(); 533 splx(s); 534 535 #ifdef GPROF 536 /* Initialize kernel profiling. */ 537 kmstartup(); 538 #endif 539 540 /* Initialize system accounting. */ 541 acct_init(); 542 543 #ifndef PIPE_SOCKETPAIR 544 /* Initialize pipes. */ 545 pipe_init(); 546 #endif 547 548 #ifdef KTRACE 549 /* Initialize ktrace. */ 550 ktrinit(); 551 #endif 552 553 /* Initialize the UUID system calls. */ 554 uuid_init(); 555 556 #ifdef WAPBL 557 /* Initialize write-ahead physical block logging. */ 558 wapbl_init(); 559 #endif 560 561 /* 562 * Create process 1 (init(8)). We do this now, as Unix has 563 * historically had init be process 1, and changing this would 564 * probably upset a lot of people. 565 * 566 * Note that process 1 won't immediately exec init(8), but will 567 * wait for us to inform it that the root file system has been 568 * mounted. 569 */ 570 if (fork1(l, 0, SIGCHLD, NULL, 0, start_init, NULL, NULL, &initproc)) 571 panic("fork init"); 572 573 /* 574 * Load any remaining builtin modules, and hand back temporary 575 * storage to the VM system. 576 */ 577 module_init_class(MODULE_CLASS_ANY); 578 579 /* 580 * Finalize configuration now that all real devices have been 581 * found. This needs to be done before the root device is 582 * selected, since finalization may create the root device. 583 */ 584 config_finalize(); 585 586 sysctl_finalize(); 587 588 /* 589 * Now that autoconfiguration has completed, we can determine 590 * the root and dump devices. 591 */ 592 cpu_rootconf(); 593 cpu_dumpconf(); 594 595 /* Mount the root file system. */ 596 do { 597 domountroothook(); 598 if ((error = vfs_mountroot())) { 599 printf("cannot mount root, error = %d\n", error); 600 boothowto |= RB_ASKNAME; 601 setroot(root_device, 602 (rootdev != NODEV) ? DISKPART(rootdev) : 0); 603 } 604 } while (error != 0); 605 mountroothook_destroy(); 606 607 /* 608 * Initialise the time-of-day clock, passing the time recorded 609 * in the root filesystem (if any) for use by systems that 610 * don't have a non-volatile time-of-day device. 611 */ 612 inittodr(rootfstime); 613 614 CIRCLEQ_FIRST(&mountlist)->mnt_flag |= MNT_ROOTFS; 615 CIRCLEQ_FIRST(&mountlist)->mnt_op->vfs_refcount++; 616 617 /* 618 * Get the vnode for '/'. Set filedesc0.fd_fd.fd_cdir to 619 * reference it. 620 */ 621 error = VFS_ROOT(CIRCLEQ_FIRST(&mountlist), &rootvnode); 622 if (error) 623 panic("cannot find root vnode, error=%d", error); 624 cwdi0.cwdi_cdir = rootvnode; 625 VREF(cwdi0.cwdi_cdir); 626 VOP_UNLOCK(rootvnode, 0); 627 cwdi0.cwdi_rdir = NULL; 628 629 /* 630 * Now that root is mounted, we can fixup initproc's CWD 631 * info. All other processes are kthreads, which merely 632 * share proc0's CWD info. 633 */ 634 initproc->p_cwdi->cwdi_cdir = rootvnode; 635 VREF(initproc->p_cwdi->cwdi_cdir); 636 initproc->p_cwdi->cwdi_rdir = NULL; 637 638 /* 639 * Now can look at time, having had a chance to verify the time 640 * from the file system. Reset l->l_rtime as it may have been 641 * munched in mi_switch() after the time got set. 642 */ 643 getnanotime(&time); 644 boottime = time; 645 #ifdef COMPAT_50 646 { 647 struct timeval tv; 648 TIMESPEC_TO_TIMEVAL(&tv, &time); 649 timeval_to_timeval50(&tv, &boottime50); 650 } 651 #endif 652 mutex_enter(proc_lock); 653 LIST_FOREACH(p, &allproc, p_list) { 654 KASSERT((p->p_flag & PK_MARKER) == 0); 655 mutex_enter(p->p_lock); 656 TIMESPEC_TO_TIMEVAL(&p->p_stats->p_start, &time); 657 LIST_FOREACH(l, &p->p_lwps, l_sibling) { 658 lwp_lock(l); 659 memset(&l->l_rtime, 0, sizeof(l->l_rtime)); 660 lwp_unlock(l); 661 } 662 mutex_exit(p->p_lock); 663 } 664 mutex_exit(proc_lock); 665 binuptime(&curlwp->l_stime); 666 667 for (CPU_INFO_FOREACH(cii, ci)) { 668 ci->ci_schedstate.spc_lastmod = time_second; 669 } 670 671 /* Create the pageout daemon kernel thread. */ 672 uvm_swap_init(); 673 if (kthread_create(PRI_PGDAEMON, KTHREAD_MPSAFE, NULL, uvm_pageout, 674 NULL, NULL, "pgdaemon")) 675 panic("fork pagedaemon"); 676 677 /* Create the filesystem syncer kernel thread. */ 678 if (kthread_create(PRI_IOFLUSH, KTHREAD_MPSAFE, NULL, sched_sync, 679 NULL, NULL, "ioflush")) 680 panic("fork syncer"); 681 682 /* Create the aiodone daemon kernel thread. */ 683 if (workqueue_create(&uvm.aiodone_queue, "aiodoned", 684 uvm_aiodone_worker, NULL, PRI_VM, IPL_NONE, WQ_MPSAFE)) 685 panic("fork aiodoned"); 686 687 vmem_rehash_start(); 688 689 /* Initialize exec structures */ 690 exec_init(1); 691 692 /* 693 * Okay, now we can let init(8) exec! It's off to userland! 694 */ 695 mutex_enter(proc_lock); 696 start_init_exec = 1; 697 cv_broadcast(&lbolt); 698 mutex_exit(proc_lock); 699 700 /* The scheduler is an infinite loop. */ 701 uvm_scheduler(); 702 /* NOTREACHED */ 703 } 704 705 /* 706 * Configure the system's hardware. 707 */ 708 static void 709 configure(void) 710 { 711 712 /* Initialize autoconf data structures. */ 713 config_init_mi(); 714 /* 715 * XXX 716 * callout_setfunc() requires mutex(9) so it can't be in config_init() 717 * on amiga and atari which use config_init() and autoconf(9) fucntions 718 * to initialize console devices. 719 */ 720 config_twiddle_init(); 721 722 pmf_init(); 723 #if NDRVCTL > 0 724 drvctl_init(); 725 #endif 726 727 #ifdef USERCONF 728 if (boothowto & RB_USERCONF) 729 user_config(); 730 #endif 731 732 if ((boothowto & (AB_SILENT|AB_VERBOSE)) == AB_SILENT) { 733 printf_nolog("Detecting hardware..."); 734 } 735 736 /* 737 * Do the machine-dependent portion of autoconfiguration. This 738 * sets the configuration machinery here in motion by "finding" 739 * the root bus. When this function returns, we expect interrupts 740 * to be enabled. 741 */ 742 cpu_configure(); 743 } 744 745 static void 746 configure2(void) 747 { 748 CPU_INFO_ITERATOR cii; 749 struct cpu_info *ci; 750 int s; 751 752 /* 753 * Now that we've found all the hardware, start the real time 754 * and statistics clocks. 755 */ 756 initclocks(); 757 758 cold = 0; /* clocks are running, we're warm now! */ 759 s = splsched(); 760 curcpu()->ci_schedstate.spc_flags |= SPCF_RUNNING; 761 splx(s); 762 763 /* Boot the secondary processors. */ 764 for (CPU_INFO_FOREACH(cii, ci)) { 765 uvm_cpu_attach(ci); 766 } 767 mp_online = true; 768 #if defined(MULTIPROCESSOR) 769 cpu_boot_secondary_processors(); 770 #endif 771 772 /* Setup the runqueues and scheduler. */ 773 runq_init(); 774 sched_init(); 775 776 /* 777 * Bus scans can make it appear as if the system has paused, so 778 * twiddle constantly while config_interrupts() jobs are running. 779 */ 780 config_twiddle_fn(NULL); 781 782 /* 783 * Create threads to call back and finish configuration for 784 * devices that want interrupts enabled. 785 */ 786 config_create_interruptthreads(); 787 788 /* Get the threads going and into any sleeps before continuing. */ 789 yield(); 790 } 791 792 static void 793 check_console(struct lwp *l) 794 { 795 struct vnode *vp; 796 int error; 797 798 error = namei_simple_kernel("/dev/console", 799 NSM_FOLLOW_NOEMULROOT, &vp); 800 if (error == 0) 801 vrele(vp); 802 else if (error == ENOENT) 803 printf("warning: no /dev/console\n"); 804 else 805 printf("warning: lookup /dev/console: error %d\n", error); 806 } 807 808 /* 809 * List of paths to try when searching for "init". 810 */ 811 static const char * const initpaths[] = { 812 "/sbin/init", 813 "/sbin/oinit", 814 "/sbin/init.bak", 815 NULL, 816 }; 817 818 /* 819 * Start the initial user process; try exec'ing each pathname in "initpaths". 820 * The program is invoked with one argument containing the boot flags. 821 */ 822 static void 823 start_init(void *arg) 824 { 825 struct lwp *l = arg; 826 struct proc *p = l->l_proc; 827 vaddr_t addr; 828 struct sys_execve_args /* { 829 syscallarg(const char *) path; 830 syscallarg(char * const *) argp; 831 syscallarg(char * const *) envp; 832 } */ args; 833 int options, i, error; 834 register_t retval[2]; 835 char flags[4], *flagsp; 836 const char *path, *slash; 837 char *ucp, **uap, *arg0, *arg1 = NULL; 838 char ipath[129]; 839 int ipx, len; 840 841 /* 842 * Now in process 1. 843 */ 844 strncpy(p->p_comm, "init", MAXCOMLEN); 845 846 /* 847 * Wait for main() to tell us that it's safe to exec. 848 */ 849 mutex_enter(proc_lock); 850 while (start_init_exec == 0) 851 cv_wait(&lbolt, proc_lock); 852 mutex_exit(proc_lock); 853 854 /* 855 * This is not the right way to do this. We really should 856 * hand-craft a descriptor onto /dev/console to hand to init, 857 * but that's a _lot_ more work, and the benefit from this easy 858 * hack makes up for the "good is the enemy of the best" effect. 859 */ 860 check_console(l); 861 862 /* 863 * Need just enough stack to hold the faked-up "execve()" arguments. 864 */ 865 addr = (vaddr_t)STACK_ALLOC(USRSTACK, PAGE_SIZE); 866 if (uvm_map(&p->p_vmspace->vm_map, &addr, PAGE_SIZE, 867 NULL, UVM_UNKNOWN_OFFSET, 0, 868 UVM_MAPFLAG(UVM_PROT_ALL, UVM_PROT_ALL, UVM_INH_COPY, 869 UVM_ADV_NORMAL, 870 UVM_FLAG_FIXED|UVM_FLAG_OVERLAY|UVM_FLAG_COPYONW)) != 0) 871 panic("init: couldn't allocate argument space"); 872 p->p_vmspace->vm_maxsaddr = (void *)STACK_MAX(addr, PAGE_SIZE); 873 874 ipx = 0; 875 while (1) { 876 if (boothowto & RB_ASKNAME) { 877 printf("init path"); 878 if (initpaths[ipx]) 879 printf(" (default %s)", initpaths[ipx]); 880 printf(": "); 881 len = cngetsn(ipath, sizeof(ipath)-1); 882 if (len == 4 && strcmp(ipath, "halt") == 0) { 883 cpu_reboot(RB_HALT, NULL); 884 } else if (len == 6 && strcmp(ipath, "reboot") == 0) { 885 cpu_reboot(0, NULL); 886 #if defined(DDB) 887 } else if (len == 3 && strcmp(ipath, "ddb") == 0) { 888 console_debugger(); 889 continue; 890 #endif 891 } else if (len > 0 && ipath[0] == '/') { 892 ipath[len] = '\0'; 893 path = ipath; 894 } else if (len == 0 && initpaths[ipx] != NULL) { 895 path = initpaths[ipx++]; 896 } else { 897 printf("use absolute path, "); 898 #if defined(DDB) 899 printf("\"ddb\", "); 900 #endif 901 printf("\"halt\", or \"reboot\"\n"); 902 continue; 903 } 904 } else { 905 if ((path = initpaths[ipx++]) == NULL) { 906 ipx = 0; 907 boothowto |= RB_ASKNAME; 908 continue; 909 } 910 } 911 912 ucp = (char *)USRSTACK; 913 914 /* 915 * Construct the boot flag argument. 916 */ 917 flagsp = flags; 918 *flagsp++ = '-'; 919 options = 0; 920 921 if (boothowto & RB_SINGLE) { 922 *flagsp++ = 's'; 923 options = 1; 924 } 925 #ifdef notyet 926 if (boothowto & RB_FASTBOOT) { 927 *flagsp++ = 'f'; 928 options = 1; 929 } 930 #endif 931 932 /* 933 * Move out the flags (arg 1), if necessary. 934 */ 935 if (options != 0) { 936 *flagsp++ = '\0'; 937 i = flagsp - flags; 938 #ifdef DEBUG 939 printf("init: copying out flags `%s' %d\n", flags, i); 940 #endif 941 arg1 = STACK_ALLOC(ucp, i); 942 ucp = STACK_MAX(arg1, i); 943 (void)copyout((void *)flags, arg1, i); 944 } 945 946 /* 947 * Move out the file name (also arg 0). 948 */ 949 i = strlen(path) + 1; 950 #ifdef DEBUG 951 printf("init: copying out path `%s' %d\n", path, i); 952 #else 953 if (boothowto & RB_ASKNAME || path != initpaths[0]) 954 printf("init: trying %s\n", path); 955 #endif 956 arg0 = STACK_ALLOC(ucp, i); 957 ucp = STACK_MAX(arg0, i); 958 (void)copyout(path, arg0, i); 959 960 /* 961 * Move out the arg pointers. 962 */ 963 ucp = (void *)STACK_ALIGN(ucp, ALIGNBYTES); 964 uap = (char **)STACK_ALLOC(ucp, sizeof(char *) * 3); 965 SCARG(&args, path) = arg0; 966 SCARG(&args, argp) = uap; 967 SCARG(&args, envp) = NULL; 968 slash = strrchr(path, '/'); 969 if (slash) 970 (void)suword((void *)uap++, 971 (long)arg0 + (slash + 1 - path)); 972 else 973 (void)suword((void *)uap++, (long)arg0); 974 if (options != 0) 975 (void)suword((void *)uap++, (long)arg1); 976 (void)suword((void *)uap++, 0); /* terminator */ 977 978 /* 979 * Now try to exec the program. If can't for any reason 980 * other than it doesn't exist, complain. 981 */ 982 error = sys_execve(l, &args, retval); 983 if (error == 0 || error == EJUSTRETURN) { 984 KERNEL_UNLOCK_LAST(l); 985 return; 986 } 987 printf("exec %s: error %d\n", path, error); 988 } 989 printf("init: not found\n"); 990 panic("no init"); 991 } 992 993 /* 994 * calculate cache size from physmem and vm_map size. 995 */ 996 vaddr_t 997 calc_cache_size(struct vm_map *map, int pct, int va_pct) 998 { 999 paddr_t t; 1000 1001 /* XXX should consider competing cache if any */ 1002 /* XXX should consider submaps */ 1003 t = (uintmax_t)physmem * pct / 100 * PAGE_SIZE; 1004 if (map != NULL) { 1005 vsize_t vsize; 1006 1007 vsize = vm_map_max(map) - vm_map_min(map); 1008 vsize = (uintmax_t)vsize * va_pct / 100; 1009 if (t > vsize) { 1010 t = vsize; 1011 } 1012 } 1013 return t; 1014 } 1015 1016 /* 1017 * Print the system start up banner. 1018 * 1019 * - Print a limited banner if AB_SILENT. 1020 * - Always send normal banner to the log. 1021 */ 1022 #define MEM_PBUFSIZE sizeof("99999 MB") 1023 1024 void 1025 banner(void) 1026 { 1027 static char notice[] = " Notice: this software is " 1028 "protected by copyright"; 1029 char pbuf[81]; 1030 void (*pr)(const char *, ...); 1031 int i; 1032 1033 if ((boothowto & AB_SILENT) != 0) { 1034 snprintf(pbuf, sizeof(pbuf), "%s %s (%s)", 1035 ostype, osrelease, kernel_ident); 1036 printf_nolog("%s", pbuf); 1037 for (i = 80 - strlen(pbuf) - sizeof(notice); i > 0; i--) 1038 printf(" "); 1039 printf_nolog("%s\n", notice); 1040 pr = aprint_normal; 1041 } else { 1042 pr = printf; 1043 } 1044 1045 memset(pbuf, 0, sizeof(pbuf)); 1046 (*pr)("%s%s", copyright, version); 1047 format_bytes(pbuf, MEM_PBUFSIZE, ctob((uint64_t)physmem)); 1048 (*pr)("total memory = %s\n", pbuf); 1049 format_bytes(pbuf, MEM_PBUFSIZE, ctob((uint64_t)uvmexp.free)); 1050 (*pr)("avail memory = %s\n", pbuf); 1051 } 1052
Indexes created Sun Sep 21 20:09:37 GMT 2025