init_main.c revision 1.152
1 /* $NetBSD: init_main.c,v 1.152 1999/06/07 20:16:09 thorpej Exp $ */ 2 3 /* 4 * Copyright (c) 1995 Christopher G. Demetriou. All rights reserved. 5 * Copyright (c) 1982, 1986, 1989, 1991, 1992, 1993 6 * The Regents of the University of California. All rights reserved. 7 * (c) UNIX System Laboratories, Inc. 8 * All or some portions of this file are derived from material licensed 9 * to the University of California by American Telephone and Telegraph 10 * Co. or Unix System Laboratories, Inc. and are reproduced herein with 11 * the permission of UNIX System Laboratories, Inc. 12 * 13 * Redistribution and use in source and binary forms, with or without 14 * modification, are permitted provided that the following conditions 15 * are met: 16 * 1. Redistributions of source code must retain the above copyright 17 * notice, this list of conditions and the following disclaimer. 18 * 2. Redistributions in binary form must reproduce the above copyright 19 * notice, this list of conditions and the following disclaimer in the 20 * documentation and/or other materials provided with the distribution. 21 * 3. All advertising materials mentioning features or use of this software 22 * must display the following acknowledgement: 23 * This product includes software developed by the University of 24 * California, Berkeley and its contributors. 25 * 4. Neither the name of the University nor the names of its contributors 26 * may be used to endorse or promote products derived from this software 27 * without specific prior written permission. 28 * 29 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 30 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 31 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 32 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 33 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 34 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 35 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 36 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 37 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 38 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 39 * SUCH DAMAGE. 40 * 41 * @(#)init_main.c 8.16 (Berkeley) 5/14/95 42 */ 43 44 #include "fs_nfs.h" 45 #include "opt_nfsserver.h" 46 #include "opt_sysv.h" 47 48 #include "rnd.h" 49 50 #include <sys/param.h> 51 #include <sys/filedesc.h> 52 #include <sys/file.h> 53 #include <sys/errno.h> 54 #include <sys/exec.h> 55 #include <sys/kernel.h> 56 #include <sys/mount.h> 57 #include <sys/map.h> 58 #include <sys/proc.h> 59 #include <sys/kthread.h> 60 #include <sys/resourcevar.h> 61 #include <sys/signalvar.h> 62 #include <sys/systm.h> 63 #include <sys/vnode.h> 64 #include <sys/tty.h> 65 #include <sys/conf.h> 66 #include <sys/disklabel.h> 67 #include <sys/buf.h> 68 #include <sys/device.h> 69 #include <sys/socketvar.h> 70 #include <sys/protosw.h> 71 #include <sys/reboot.h> 72 #include <sys/user.h> 73 #ifdef SYSVSHM 74 #include <sys/shm.h> 75 #endif 76 #ifdef SYSVSEM 77 #include <sys/sem.h> 78 #endif 79 #ifdef SYSVMSG 80 #include <sys/msg.h> 81 #endif 82 #include <sys/domain.h> 83 #include <sys/mbuf.h> 84 #include <sys/namei.h> 85 #if NRND > 0 86 #include <sys/rnd.h> 87 #endif 88 89 #include <sys/syscall.h> 90 #include <sys/syscallargs.h> 91 92 #include <ufs/ufs/quota.h> 93 94 #include <machine/cpu.h> 95 96 #include <vm/vm.h> 97 #include <vm/vm_pageout.h> 98 99 #include <uvm/uvm.h> 100 101 #include <net/if.h> 102 #include <net/raw_cb.h> 103 104 char copyright[] = "\ 105 Copyright (c) 1996, 1997, 1998, 1999 106 The NetBSD Foundation, Inc. All rights reserved. 107 Copyright (c) 1982, 1986, 1989, 1991, 1993 108 The Regents of the University of California. All rights reserved. 109 110 "; 111 112 /* Components of the first process -- never freed. */ 113 struct session session0; 114 struct pgrp pgrp0; 115 struct proc proc0; 116 struct pcred cred0; 117 struct filedesc0 filedesc0; 118 struct cwdinfo cwdi0; 119 struct plimit limit0; 120 struct vmspace vmspace0; 121 struct sigacts sigacts0; 122 #ifndef curproc 123 struct proc *curproc = &proc0; 124 #endif 125 struct proc *initproc; 126 127 int cmask = CMASK; 128 extern struct user *proc0paddr; 129 130 struct vnode *rootvp, *swapdev_vp; 131 int boothowto; 132 struct timeval boottime; 133 struct timeval runtime; 134 135 static void check_console __P((struct proc *p)); 136 static void start_init __P((void *)); 137 static void start_pagedaemon __P((void *)); 138 static void start_reaper __P((void *)); 139 void main __P((void)); 140 141 extern char sigcode[], esigcode[]; 142 #ifdef SYSCALL_DEBUG 143 extern char *syscallnames[]; 144 #endif 145 146 struct emul emul_netbsd = { 147 "netbsd", 148 NULL, 149 sendsig, 150 SYS_syscall, 151 SYS_MAXSYSCALL, 152 sysent, 153 #ifdef SYSCALL_DEBUG 154 syscallnames, 155 #else 156 NULL, 157 #endif 158 0, 159 copyargs, 160 setregs, 161 sigcode, 162 esigcode, 163 }; 164 165 /* 166 * System startup; initialize the world, create process 0, mount root 167 * filesystem, and fork to create init and pagedaemon. Most of the 168 * hard work is done in the lower-level initialization routines including 169 * startup(), which does memory initialization and autoconfiguration. 170 */ 171 void 172 main() 173 { 174 struct proc *p; 175 struct pdevinit *pdev; 176 int i, s, error; 177 extern struct pdevinit pdevinit[]; 178 extern void roundrobin __P((void *)); 179 extern void schedcpu __P((void *)); 180 extern void disk_init __P((void)); 181 #if defined(NFSSERVER) || defined(NFS) 182 extern void nfs_init __P((void)); 183 #endif 184 185 /* 186 * Initialize the current process pointer (curproc) before 187 * any possible traps/probes to simplify trap processing. 188 */ 189 p = &proc0; 190 curproc = p; 191 /* 192 * Attempt to find console and initialize 193 * in case of early panic or other messages. 194 */ 195 consinit(); 196 printf("%s", copyright); 197 198 uvm_init(); 199 200 /* Do machine-dependent initialization. */ 201 cpu_startup(); 202 203 /* 204 * Initialize mbuf's. Do this now because we might attempt to 205 * allocate mbufs or mbuf clusters during autoconfiguration. 206 */ 207 mbinit(); 208 209 /* Initialize sockets. */ 210 soinit(); 211 212 disk_init(); /* must come before autoconfiguration */ 213 tty_init(); /* initialise tty list */ 214 #if NRND > 0 215 rnd_init(); 216 #endif 217 config_init(); /* init autoconfiguration data structures */ 218 configure(); /* ...and configure the hardware */ 219 220 /* 221 * Initialize process and pgrp structures. 222 */ 223 procinit(); 224 225 /* 226 * Create process 0 (the swapper). 227 */ 228 LIST_INSERT_HEAD(&allproc, p, p_list); 229 p->p_pgrp = &pgrp0; 230 LIST_INSERT_HEAD(PGRPHASH(0), &pgrp0, pg_hash); 231 LIST_INIT(&pgrp0.pg_members); 232 LIST_INSERT_HEAD(&pgrp0.pg_members, p, p_pglist); 233 234 pgrp0.pg_session = &session0; 235 session0.s_count = 1; 236 session0.s_sid = p->p_pid; 237 session0.s_leader = p; 238 239 /* 240 * Set P_NOCLDWAIT so that kernel threads are reparented to 241 * init(8) when they exit. init(8) can easily wait them out 242 * for us. 243 */ 244 p->p_flag = P_INMEM | P_SYSTEM | P_NOCLDWAIT; 245 p->p_stat = SRUN; 246 p->p_nice = NZERO; 247 p->p_emul = &emul_netbsd; 248 strncpy(p->p_comm, "swapper", MAXCOMLEN); 249 250 /* Create credentials. */ 251 cred0.p_refcnt = 1; 252 p->p_cred = &cred0; 253 p->p_ucred = crget(); 254 p->p_ucred->cr_ngroups = 1; /* group 0 */ 255 256 /* Create the file descriptor table. */ 257 finit(); 258 p->p_fd = &filedesc0.fd_fd; 259 fdinit1(&filedesc0); 260 261 /* Create the CWD info. */ 262 p->p_cwdi = &cwdi0; 263 cwdi0.cwdi_cmask = cmask; 264 cwdi0.cwdi_refcnt = 1; 265 266 /* Create the limits structures. */ 267 p->p_limit = &limit0; 268 for (i = 0; i < sizeof(p->p_rlimit)/sizeof(p->p_rlimit[0]); i++) 269 limit0.pl_rlimit[i].rlim_cur = 270 limit0.pl_rlimit[i].rlim_max = RLIM_INFINITY; 271 272 limit0.pl_rlimit[RLIMIT_NOFILE].rlim_max = maxfiles; 273 limit0.pl_rlimit[RLIMIT_NOFILE].rlim_cur = 274 maxfiles < NOFILE ? maxfiles : NOFILE; 275 276 limit0.pl_rlimit[RLIMIT_NPROC].rlim_max = maxproc; 277 limit0.pl_rlimit[RLIMIT_NPROC].rlim_cur = 278 maxproc < MAXUPRC ? maxproc : MAXUPRC; 279 280 i = ptoa(uvmexp.free); 281 limit0.pl_rlimit[RLIMIT_RSS].rlim_max = i; 282 limit0.pl_rlimit[RLIMIT_MEMLOCK].rlim_max = i; 283 limit0.pl_rlimit[RLIMIT_MEMLOCK].rlim_cur = i / 3; 284 limit0.p_refcnt = 1; 285 286 /* 287 * Initialize proc0's vmspace, which uses the kernel pmap. 288 * All kernel processes (which never have user space mappings) 289 * share proc0's vmspace, and thus, the kernel pmap. 290 */ 291 uvmspace_init(&vmspace0, pmap_kernel(), round_page(VM_MIN_ADDRESS), 292 trunc_page(VM_MAX_ADDRESS), TRUE); 293 p->p_vmspace = &vmspace0; 294 295 p->p_addr = proc0paddr; /* XXX */ 296 297 /* 298 * We continue to place resource usage info in the 299 * user struct so they're pageable. 300 */ 301 p->p_stats = &p->p_addr->u_stats; 302 303 /* 304 * Charge root for one process. 305 */ 306 (void)chgproccnt(0, 1); 307 308 rqinit(); 309 310 /* Configure virtual memory system, set vm rlimits. */ 311 uvm_init_limits(p); 312 313 /* Initialize the file systems. */ 314 #if defined(NFSSERVER) || defined(NFS) 315 nfs_init(); /* initialize server/shared data */ 316 #endif 317 vfsinit(); 318 319 /* Start real time and statistics clocks. */ 320 initclocks(); 321 322 #ifdef SYSVSHM 323 /* Initialize System V style shared memory. */ 324 shminit(); 325 #endif 326 327 #ifdef SYSVSEM 328 /* Initialize System V style semaphores. */ 329 seminit(); 330 #endif 331 332 #ifdef SYSVMSG 333 /* Initialize System V style message queues. */ 334 msginit(); 335 #endif 336 337 /* Attach pseudo-devices. */ 338 for (pdev = pdevinit; pdev->pdev_attach != NULL; pdev++) 339 (*pdev->pdev_attach)(pdev->pdev_count); 340 341 /* 342 * Initialize protocols. Block reception of incoming packets 343 * until everything is ready. 344 */ 345 s = splimp(); 346 ifinit(); 347 domaininit(); 348 splx(s); 349 350 #ifdef GPROF 351 /* Initialize kernel profiling. */ 352 kmstartup(); 353 #endif 354 355 /* Kick off timeout driven events by calling first time. */ 356 roundrobin(NULL); 357 schedcpu(NULL); 358 359 /* Determine the root and dump devices. */ 360 cpu_rootconf(); 361 cpu_dumpconf(); 362 363 /* Mount the root file system. */ 364 do { 365 domountroothook(); 366 if ((error = vfs_mountroot())) { 367 printf("cannot mount root, error = %d\n", error); 368 boothowto |= RB_ASKNAME; 369 setroot(root_device, 370 (rootdev != NODEV) ? DISKPART(rootdev) : 0); 371 } 372 } while (error != 0); 373 mountroothook_destroy(); 374 375 mountlist.cqh_first->mnt_flag |= MNT_ROOTFS; 376 mountlist.cqh_first->mnt_op->vfs_refcount++; 377 378 /* 379 * Get the vnode for '/'. Set filedesc0.fd_fd.fd_cdir to 380 * reference it. 381 */ 382 if (VFS_ROOT(mountlist.cqh_first, &rootvnode)) 383 panic("cannot find root vnode"); 384 cwdi0.cwdi_cdir = rootvnode; 385 VREF(cwdi0.cwdi_cdir); 386 VOP_UNLOCK(rootvnode, 0); 387 cwdi0.cwdi_rdir = NULL; 388 uvm_swap_init(); 389 390 /* 391 * Now can look at time, having had a chance to verify the time 392 * from the file system. Reset p->p_rtime as it may have been 393 * munched in mi_switch() after the time got set. 394 */ 395 p->p_stats->p_start = runtime = mono_time = boottime = time; 396 p->p_rtime.tv_sec = p->p_rtime.tv_usec = 0; 397 398 /* 399 * Initialize signal-related data structures, and signal state 400 * for proc0. 401 */ 402 signal_init(); 403 p->p_sigacts = &sigacts0; 404 siginit(p); 405 406 /* Create process 1 (init(8)). */ 407 if (fork1(p, 0, SIGCHLD, NULL, 0, NULL, &initproc)) 408 panic("fork init"); 409 cpu_set_kpc(initproc, start_init, initproc); 410 411 /* Create process 2, the pageout daemon kernel thread. */ 412 if (kthread_create(start_pagedaemon, NULL, NULL, "pagedaemon")) 413 panic("fork pagedaemon"); 414 415 /* Create process 3, the process reaper kernel thread. */ 416 if (kthread_create(start_reaper, NULL, NULL, "reaper")) 417 panic("fork reaper"); 418 419 /* Create any other deferred kernel threads. */ 420 kthread_run_deferred_queue(); 421 422 /* The scheduler is an infinite loop. */ 423 uvm_scheduler(); 424 /* NOTREACHED */ 425 } 426 427 static void 428 check_console(p) 429 struct proc *p; 430 { 431 struct nameidata nd; 432 int error; 433 434 NDINIT(&nd, LOOKUP, FOLLOW, UIO_SYSSPACE, "/dev/console", p); 435 error = namei(&nd); 436 if (error == 0) 437 vrele(nd.ni_vp); 438 else if (error == ENOENT) 439 printf("warning: no /dev/console\n"); 440 else 441 printf("warning: lookup /dev/console: error %d\n", error); 442 } 443 444 /* 445 * List of paths to try when searching for "init". 446 */ 447 static char *initpaths[] = { 448 "/sbin/init", 449 "/sbin/oinit", 450 "/sbin/init.bak", 451 NULL, 452 }; 453 454 /* 455 * Start the initial user process; try exec'ing each pathname in "initpaths". 456 * The program is invoked with one argument containing the boot flags. 457 */ 458 static void 459 start_init(arg) 460 void *arg; 461 { 462 struct proc *p = arg; 463 vaddr_t addr; 464 struct sys_execve_args /* { 465 syscallarg(const char *) path; 466 syscallarg(char * const *) argp; 467 syscallarg(char * const *) envp; 468 } */ args; 469 int options, i, error; 470 register_t retval[2]; 471 char flags[4], *flagsp; 472 char **pathp, *path, *slash, *ucp, **uap, *arg0, *arg1 = NULL; 473 474 /* 475 * Now in process 1. 476 */ 477 strncpy(p->p_comm, "init", MAXCOMLEN); 478 479 /* 480 * This is not the right way to do this. We really should 481 * hand-craft a descriptor onto /dev/console to hand to init, 482 * but that's a _lot_ more work, and the benefit from this easy 483 * hack makes up for the "good is the enemy of the best" effect. 484 */ 485 check_console(p); 486 487 /* 488 * Need just enough stack to hold the faked-up "execve()" arguments. 489 */ 490 addr = USRSTACK - PAGE_SIZE; 491 if (uvm_map(&p->p_vmspace->vm_map, &addr, PAGE_SIZE, 492 NULL, UVM_UNKNOWN_OFFSET, 493 UVM_MAPFLAG(UVM_PROT_ALL, UVM_PROT_ALL, UVM_INH_COPY, 494 UVM_ADV_NORMAL, 495 UVM_FLAG_FIXED|UVM_FLAG_OVERLAY|UVM_FLAG_COPYONW)) 496 != KERN_SUCCESS) 497 panic("init: couldn't allocate argument space"); 498 p->p_vmspace->vm_maxsaddr = (caddr_t)addr; 499 500 for (pathp = &initpaths[0]; (path = *pathp) != NULL; pathp++) { 501 ucp = (char *)(addr + PAGE_SIZE); 502 503 /* 504 * Construct the boot flag argument. 505 */ 506 flagsp = flags; 507 *flagsp++ = '-'; 508 options = 0; 509 510 if (boothowto & RB_SINGLE) { 511 *flagsp++ = 's'; 512 options = 1; 513 } 514 #ifdef notyet 515 if (boothowto & RB_FASTBOOT) { 516 *flagsp++ = 'f'; 517 options = 1; 518 } 519 #endif 520 521 /* 522 * Move out the flags (arg 1), if necessary. 523 */ 524 if (options != 0) { 525 *flagsp++ = '\0'; 526 i = flagsp - flags; 527 #ifdef DEBUG 528 printf("init: copying out flags `%s' %d\n", flags, i); 529 #endif 530 (void)copyout((caddr_t)flags, (caddr_t)(ucp -= i), i); 531 arg1 = ucp; 532 } 533 534 /* 535 * Move out the file name (also arg 0). 536 */ 537 i = strlen(path) + 1; 538 #ifdef DEBUG 539 printf("init: copying out path `%s' %d\n", path, i); 540 #endif 541 (void)copyout((caddr_t)path, (caddr_t)(ucp -= i), i); 542 arg0 = ucp; 543 544 /* 545 * Move out the arg pointers. 546 */ 547 uap = (char **)((long)ucp & ~ALIGNBYTES); 548 (void)suword((caddr_t)--uap, 0); /* terminator */ 549 if (options != 0) 550 (void)suword((caddr_t)--uap, (long)arg1); 551 slash = strrchr(path, '/'); 552 if (slash) 553 (void)suword((caddr_t)--uap, 554 (long)arg0 + (slash + 1 - path)); 555 else 556 (void)suword((caddr_t)--uap, (long)arg0); 557 558 /* 559 * Point at the arguments. 560 */ 561 SCARG(&args, path) = arg0; 562 SCARG(&args, argp) = uap; 563 SCARG(&args, envp) = NULL; 564 565 /* 566 * Now try to exec the program. If can't for any reason 567 * other than it doesn't exist, complain. 568 */ 569 error = sys_execve(p, &args, retval); 570 if (error == 0 || error == EJUSTRETURN) 571 return; 572 if (error != ENOENT) 573 printf("exec %s: error %d\n", path, error); 574 } 575 printf("init: not found\n"); 576 panic("no init"); 577 } 578 579 /* ARGSUSED */ 580 static void 581 start_pagedaemon(arg) 582 void *arg; 583 { 584 585 uvm_pageout(); 586 /* NOTREACHED */ 587 } 588 589 /* ARGSUSED */ 590 static void 591 start_reaper(arg) 592 void *arg; 593 { 594 595 reaper(); 596 /* NOTREACHED */ 597 } 598
Indexes created Sun Sep 21 18:09:52 GMT 2025