init_main.c revision 1.83 1 /* $NetBSD: init_main.c,v 1.83 1996/04/04 00:26:28 cgd 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 8.9 (Berkeley) 1/21/94
42 */
43
44 #include <sys/param.h>
45 #include <sys/filedesc.h>
46 #include <sys/errno.h>
47 #include <sys/exec.h>
48 #include <sys/kernel.h>
49 #include <sys/mount.h>
50 #include <sys/map.h>
51 #include <sys/proc.h>
52 #include <sys/resourcevar.h>
53 #include <sys/signalvar.h>
54 #include <sys/systm.h>
55 #include <sys/vnode.h>
56 #include <sys/conf.h>
57 #include <sys/buf.h>
58 #ifdef REAL_CLISTS
59 #include <sys/clist.h>
60 #endif
61 #include <sys/device.h>
62 #include <sys/protosw.h>
63 #include <sys/reboot.h>
64 #include <sys/user.h>
65 #include <sys/cpu.h>
66 #ifdef SYSVSHM
67 #include <sys/shm.h>
68 #endif
69 #ifdef SYSVSEM
70 #include <sys/sem.h>
71 #endif
72 #ifdef SYSVMSG
73 #include <sys/msg.h>
74 #endif
75 #include <sys/domain.h>
76 #include <sys/mbuf.h>
77
78 #include <sys/syscall.h>
79 #include <sys/syscallargs.h>
80
81 #include <ufs/ufs/quota.h>
82
83 #include <machine/cpu.h>
84
85 #include <vm/vm.h>
86 #include <vm/vm_pageout.h>
87
88 #include <net/if.h>
89 #include <net/raw_cb.h>
90
91 char copyright[] =
92 "Copyright (c) 1982, 1986, 1989, 1991, 1993\n\tThe Regents of the University of California. All rights reserved.\n\n";
93
94 /* Components of the first process -- never freed. */
95 struct session session0;
96 struct pgrp pgrp0;
97 struct proc proc0;
98 struct pcred cred0;
99 struct filedesc0 filedesc0;
100 struct plimit limit0;
101 struct vmspace vmspace0;
102 struct proc *curproc = &proc0;
103 struct proc *initproc, *pageproc;
104
105 int cmask = CMASK;
106 extern struct user *proc0paddr;
107
108 struct vnode *rootvp, *swapdev_vp;
109 int boothowto;
110 struct timeval boottime;
111 struct timeval runtime;
112
113 static void start_init __P((struct proc *));
114 static void start_pagedaemon __P((struct proc *));
115 void main __P((void *));
116
117 #ifdef cpu_set_init_frame
118 void *initframep; /* XXX should go away */
119 #endif
120
121 extern char sigcode[], esigcode[];
122 #ifdef SYSCALL_DEBUG
123 extern char *syscallnames[];
124 #endif
125
126 struct emul emul_netbsd = {
127 "netbsd",
128 NULL,
129 sendsig,
130 SYS_syscall,
131 SYS_MAXSYSCALL,
132 sysent,
133 #ifdef SYSCALL_DEBUG
134 syscallnames,
135 #else
136 NULL,
137 #endif
138 0,
139 copyargs,
140 setregs,
141 sigcode,
142 esigcode,
143 };
144
145 /*
146 * System startup; initialize the world, create process 0, mount root
147 * filesystem, and fork to create init and pagedaemon. Most of the
148 * hard work is done in the lower-level initialization routines including
149 * startup(), which does memory initialization and autoconfiguration.
150 */
151 void
152 main(framep)
153 void *framep; /* XXX should go away */
154 {
155 register struct proc *p;
156 register struct pdevinit *pdev;
157 register int i;
158 int s;
159 register_t rval[2];
160 extern int (*mountroot) __P((void));
161 extern struct pdevinit pdevinit[];
162 extern void roundrobin __P((void *));
163 extern void schedcpu __P((void *));
164 extern void disk_init __P((void));
165
166 /*
167 * Initialize the current process pointer (curproc) before
168 * any possible traps/probes to simplify trap processing.
169 */
170 p = &proc0;
171 curproc = p;
172 /*
173 * Attempt to find console and initialize
174 * in case of early panic or other messages.
175 */
176 consinit();
177 printf(copyright);
178
179 vm_mem_init();
180 kmeminit();
181 disk_init(); /* must come before autoconfiguration */
182 config_init(); /* init autoconfiguration data structures */
183 cpu_startup();
184
185 /*
186 * Initialize process and pgrp structures.
187 */
188 procinit();
189
190 /*
191 * Create process 0 (the swapper).
192 */
193 LIST_INSERT_HEAD(&allproc, p, p_list);
194 p->p_pgrp = &pgrp0;
195 LIST_INSERT_HEAD(PGRPHASH(0), &pgrp0, pg_hash);
196 LIST_INIT(&pgrp0.pg_members);
197 LIST_INSERT_HEAD(&pgrp0.pg_members, p, p_pglist);
198
199 pgrp0.pg_session = &session0;
200 session0.s_count = 1;
201 session0.s_leader = p;
202
203 p->p_flag = P_INMEM | P_SYSTEM;
204 p->p_stat = SRUN;
205 p->p_nice = NZERO;
206 p->p_emul = &emul_netbsd;
207 bcopy("swapper", p->p_comm, sizeof ("swapper"));
208
209 /* Create credentials. */
210 cred0.p_refcnt = 1;
211 p->p_cred = &cred0;
212 p->p_ucred = crget();
213 p->p_ucred->cr_ngroups = 1; /* group 0 */
214
215 /* Create the file descriptor table. */
216 p->p_fd = &filedesc0.fd_fd;
217 filedesc0.fd_fd.fd_refcnt = 1;
218 filedesc0.fd_fd.fd_cmask = cmask;
219 filedesc0.fd_fd.fd_ofiles = filedesc0.fd_dfiles;
220 filedesc0.fd_fd.fd_ofileflags = filedesc0.fd_dfileflags;
221 filedesc0.fd_fd.fd_nfiles = NDFILE;
222
223 /* Create the limits structures. */
224 p->p_limit = &limit0;
225 for (i = 0; i < sizeof(p->p_rlimit)/sizeof(p->p_rlimit[0]); i++)
226 limit0.pl_rlimit[i].rlim_cur =
227 limit0.pl_rlimit[i].rlim_max = RLIM_INFINITY;
228 limit0.pl_rlimit[RLIMIT_NOFILE].rlim_cur = NOFILE;
229 limit0.pl_rlimit[RLIMIT_NPROC].rlim_cur = MAXUPRC;
230 i = ptoa(cnt.v_free_count);
231 limit0.pl_rlimit[RLIMIT_RSS].rlim_max = i;
232 limit0.pl_rlimit[RLIMIT_MEMLOCK].rlim_max = i;
233 limit0.pl_rlimit[RLIMIT_MEMLOCK].rlim_cur = i / 3;
234 limit0.p_refcnt = 1;
235
236 /* Allocate a prototype map so we have something to fork. */
237 p->p_vmspace = &vmspace0;
238 vmspace0.vm_refcnt = 1;
239 pmap_pinit(&vmspace0.vm_pmap);
240 vm_map_init(&p->p_vmspace->vm_map, round_page(VM_MIN_ADDRESS),
241 trunc_page(VM_MAX_ADDRESS), TRUE);
242 vmspace0.vm_map.pmap = &vmspace0.vm_pmap;
243 p->p_addr = proc0paddr; /* XXX */
244
245 /*
246 * We continue to place resource usage info and signal
247 * actions in the user struct so they're pageable.
248 */
249 p->p_stats = &p->p_addr->u_stats;
250 p->p_sigacts = &p->p_addr->u_sigacts;
251
252 /*
253 * Charge root for one process.
254 */
255 (void)chgproccnt(0, 1);
256
257 rqinit();
258
259 /* Configure virtual memory system, set vm rlimits. */
260 vm_init_limits(p);
261
262 /* Initialize the file systems. */
263 vfsinit();
264
265 /* Start real time and statistics clocks. */
266 initclocks();
267
268 /* Initialize mbuf's. */
269 mbinit();
270
271 #ifdef REAL_CLISTS
272 /* Initialize clists. */
273 clist_init();
274 #endif
275
276 #ifdef SYSVSHM
277 /* Initialize System V style shared memory. */
278 shminit();
279 #endif
280
281 #ifdef SYSVSEM
282 /* Initialize System V style semaphores. */
283 seminit();
284 #endif
285
286 #ifdef SYSVMSG
287 /* Initialize System V style message queues. */
288 msginit();
289 #endif
290
291 /* Attach pseudo-devices. */
292 for (pdev = pdevinit; pdev->pdev_attach != NULL; pdev++)
293 (*pdev->pdev_attach)(pdev->pdev_count);
294
295 /*
296 * Initialize protocols. Block reception of incoming packets
297 * until everything is ready.
298 */
299 s = splimp();
300 ifinit();
301 domaininit();
302 splx(s);
303
304 #ifdef GPROF
305 /* Initialize kernel profiling. */
306 kmstartup();
307 #endif
308
309 /* Kick off timeout driven events by calling first time. */
310 roundrobin(NULL);
311 schedcpu(NULL);
312
313 /* Mount the root file system. */
314 if ((*mountroot)())
315 panic("cannot mount root");
316 mountlist.cqh_first->mnt_flag |= MNT_ROOTFS;
317 mountlist.cqh_first->mnt_op->vfs_refcount++;
318
319 /* Get the vnode for '/'. Set filedesc0.fd_fd.fd_cdir to reference it. */
320 if (VFS_ROOT(mountlist.cqh_first, &rootvnode))
321 panic("cannot find root vnode");
322 filedesc0.fd_fd.fd_cdir = rootvnode;
323 VREF(filedesc0.fd_fd.fd_cdir);
324 VOP_UNLOCK(rootvnode);
325 filedesc0.fd_fd.fd_rdir = NULL;
326 swapinit();
327
328 /*
329 * Now can look at time, having had a chance to verify the time
330 * from the file system. Reset p->p_rtime as it may have been
331 * munched in mi_switch() after the time got set.
332 */
333 p->p_stats->p_start = runtime = mono_time = boottime = time;
334 p->p_rtime.tv_sec = p->p_rtime.tv_usec = 0;
335
336 /* Initialize signal state for process 0. */
337 siginit(p);
338
339 /* Create process 1 (init(8)). */
340 if (sys_fork(p, NULL, rval))
341 panic("fork init");
342 #ifdef cpu_set_init_frame /* XXX should go away */
343 if (rval[1]) {
344 /*
345 * Now in process 1.
346 */
347 initframep = framep;
348 start_init(curproc);
349 return;
350 }
351 #else
352 cpu_set_kpc(pfind(1), start_init);
353 #endif
354
355 /* Create process 2 (the pageout daemon). */
356 if (sys_fork(p, NULL, rval))
357 panic("fork pager");
358 #ifdef cpu_set_init_frame /* XXX should go away */
359 if (rval[1]) {
360 /*
361 * Now in process 2.
362 */
363 start_pagedaemon(curproc);
364 }
365 #else
366 cpu_set_kpc(pfind(2), start_pagedaemon);
367 #endif
368
369 /* The scheduler is an infinite loop. */
370 scheduler();
371 /* NOTREACHED */
372 }
373
374 /*
375 * List of paths to try when searching for "init".
376 */
377 static char *initpaths[] = {
378 "/sbin/init",
379 "/sbin/oinit",
380 "/sbin/init.bak",
381 NULL,
382 };
383
384 /*
385 * Start the initial user process; try exec'ing each pathname in "initpaths".
386 * The program is invoked with one argument containing the boot flags.
387 */
388 static void
389 start_init(p)
390 struct proc *p;
391 {
392 vm_offset_t addr;
393 struct sys_execve_args /* {
394 syscallarg(char *) path;
395 syscallarg(char **) argp;
396 syscallarg(char **) envp;
397 } */ args;
398 int options, i, error;
399 register_t retval[2];
400 char flags[4], *flagsp;
401 char **pathp, *path, *ucp, **uap, *arg0, *arg1 = NULL;
402
403 /*
404 * Now in process 1.
405 */
406 initproc = p;
407
408 #ifdef cpu_set_init_frame /* XXX should go away */
409 /*
410 * We need to set the system call frame as if we were entered through
411 * a syscall() so that when we call sys_execve() below, it will be able
412 * to set the entry point (see setregs) when it tries to exec. The
413 * startup code in "locore.s" has allocated space for the frame and
414 * passed a pointer to that space as main's argument.
415 */
416 cpu_set_init_frame(p, initframep);
417 #endif
418
419 /*
420 * Need just enough stack to hold the faked-up "execve()" arguments.
421 */
422 addr = USRSTACK - PAGE_SIZE;
423 if (vm_allocate(&p->p_vmspace->vm_map, &addr, (vm_size_t)PAGE_SIZE,
424 FALSE) != 0)
425 panic("init: couldn't allocate argument space");
426 p->p_vmspace->vm_maxsaddr = (caddr_t)addr;
427
428 for (pathp = &initpaths[0]; (path = *pathp) != NULL; pathp++) {
429 ucp = (char *)(addr + PAGE_SIZE);
430
431 /*
432 * Construct the boot flag argument.
433 */
434 flagsp = flags;
435 *flagsp++ = '-';
436 options = 0;
437
438 if (boothowto & RB_SINGLE) {
439 *flagsp++ = 's';
440 options = 1;
441 }
442 #ifdef notyet
443 if (boothowto & RB_FASTBOOT) {
444 *flagsp++ = 'f';
445 options = 1;
446 }
447 #endif
448
449 /*
450 * Move out the flags (arg 1), if necessary.
451 */
452 if (options != 0) {
453 *flagsp++ = '\0';
454 i = flagsp - flags;
455 #ifdef DEBUG
456 printf("init: copying out flags `%s' %d\n", flags, i);
457 #endif
458 (void)copyout((caddr_t)flags, (caddr_t)(ucp -= i), i);
459 arg1 = ucp;
460 }
461
462 /*
463 * Move out the file name (also arg 0).
464 */
465 i = strlen(path) + 1;
466 #ifdef DEBUG
467 printf("init: copying out path `%s' %d\n", path, i);
468 #endif
469 (void)copyout((caddr_t)path, (caddr_t)(ucp -= i), i);
470 arg0 = ucp;
471
472 /*
473 * Move out the arg pointers.
474 */
475 uap = (char **)((long)ucp & ~ALIGNBYTES);
476 (void)suword((caddr_t)--uap, 0); /* terminator */
477 if (options != 0)
478 (void)suword((caddr_t)--uap, (long)arg1);
479 (void)suword((caddr_t)--uap, (long)arg0);
480
481 /*
482 * Point at the arguments.
483 */
484 SCARG(&args, path) = arg0;
485 SCARG(&args, argp) = uap;
486 SCARG(&args, envp) = NULL;
487
488 /*
489 * Now try to exec the program. If can't for any reason
490 * other than it doesn't exist, complain.
491 */
492 if ((error = sys_execve(p, &args, retval)) == 0)
493 return;
494 if (error != ENOENT)
495 printf("exec %s: error %d\n", path, error);
496 }
497 printf("init: not found\n");
498 panic("no init");
499 }
500
501 static void
502 start_pagedaemon(p)
503 struct proc *p;
504 {
505
506 /*
507 * Now in process 2.
508 */
509 pageproc = p;
510 p->p_flag |= P_INMEM | P_SYSTEM; /* XXX */
511 bcopy("pagedaemon", curproc->p_comm, sizeof ("pagedaemon"));
512 vm_pageout();
513 /* NOTREACHED */
514 }
515