kernel - Fix panic related to USB cd
[dragonfly.git] / sys / kern / init_main.c
CommitLineData
984263bc
MD
1/*
2 * Copyright (c) 1995 Terrence R. Lambert
3 * All rights reserved.
4 *
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.
dc71b7ab 21 * 3. Neither the name of the University nor the names of its contributors
984263bc
MD
22 * may be used to endorse or promote products derived from this software
23 * without specific prior written permission.
24 *
25 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
26 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
27 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
28 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
29 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
30 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
31 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
32 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
33 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
34 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
35 * SUCH DAMAGE.
36 *
37 * @(#)init_main.c 8.9 (Berkeley) 1/21/94
38 * $FreeBSD: src/sys/kern/init_main.c,v 1.134.2.8 2003/06/06 20:21:32 tegge Exp $
39 */
40
41#include "opt_init_path.h"
42
43#include <sys/param.h>
44#include <sys/file.h>
45#include <sys/filedesc.h>
46#include <sys/kernel.h>
47#include <sys/mount.h>
48#include <sys/sysctl.h>
49#include <sys/proc.h>
50#include <sys/resourcevar.h>
51#include <sys/signalvar.h>
52#include <sys/systm.h>
53#include <sys/vnode.h>
54#include <sys/sysent.h>
55#include <sys/reboot.h>
56#include <sys/sysproto.h>
57#include <sys/vmmeter.h>
58#include <sys/unistd.h>
59#include <sys/malloc.h>
0b692e79 60#include <sys/machintr.h>
684a93c4 61
58c2553a 62#include <sys/refcount.h>
dadab5e9 63#include <sys/file2.h>
f8c3996b 64#include <sys/thread2.h>
8f1f6170 65#include <sys/spinlock2.h>
984263bc
MD
66
67#include <machine/cpu.h>
68
69#include <vm/vm.h>
70#include <vm/vm_param.h>
71#include <sys/lock.h>
72#include <vm/pmap.h>
73#include <vm/vm_map.h>
ba39e2e0 74#include <vm/vm_extern.h>
984263bc
MD
75#include <sys/copyright.h>
76
cd29885a
MD
77int vfs_mountroot_devfs(void);
78
984263bc
MD
79/* Components of the first process -- never freed. */
80static struct session session0;
81static struct pgrp pgrp0;
b1b4e5a6 82static struct sigacts sigacts0;
0679adc4 83static struct filedesc filedesc0;
984263bc
MD
84static struct plimit limit0;
85static struct vmspace vmspace0;
b7c628e4
MD
86struct proc *initproc;
87struct proc proc0;
08f2f1bb 88struct lwp lwp0;
b7c628e4 89struct thread thread0;
0adbcbd6 90struct sys_kpmap *kpmap;
6e2a912c 91struct sysreaper initreaper;
984263bc
MD
92
93int cmask = CMASK;
d87e79b9 94u_int cpu_mi_feature;
916e604f 95cpumask_t usched_global_cpumask;
984263bc 96extern struct user *proc0paddr;
984263bc 97
984263bc 98int boothowto = 0; /* initialized so that it can be patched */
0c52fa62
SG
99SYSCTL_INT(_debug, OID_AUTO, boothowto, CTLFLAG_RD, &boothowto, 0,
100 "Reboot flags, from console subsystem");
1fe61e9c
SZ
101SYSCTL_OPAQUE(_kern, OID_AUTO, usched_global_cpumask, CTLFLAG_RW,
102 &usched_global_cpumask, sizeof(usched_global_cpumask), "LU",
103 "global user scheduler cpumask");
984263bc
MD
104
105/*
106 * This ensures that there is at least one entry so that the sysinit_set
7bd34050 107 * symbol is not undefined. A subsystem ID of SI_SPECIAL_DUMMY is never
984263bc
MD
108 * executed.
109 */
f3f3eadb 110SYSINIT(placeholder, SI_SPECIAL_DUMMY, SI_ORDER_ANY, NULL, NULL);
984263bc
MD
111
112/*
113 * The sysinit table itself. Items are checked off as the are run.
114 * If we want to register new sysinit types, add them to newsysinit.
115 */
dc62b251
MD
116SET_DECLARE(sysinit_set, struct sysinit);
117struct sysinit **sysinit, **sysinit_end;
118struct sysinit **newsysinit, **newsysinit_end;
119
984263bc
MD
120
121/*
122 * Merge a new sysinit set into the current set, reallocating it if
123 * necessary. This can only be called after malloc is running.
124 */
125void
dc62b251 126sysinit_add(struct sysinit **set, struct sysinit **set_end)
984263bc
MD
127{
128 struct sysinit **newset;
129 struct sysinit **sipp;
130 struct sysinit **xipp;
dc62b251 131 int count;
984263bc 132
dc62b251 133 count = set_end - set;
984263bc 134 if (newsysinit)
dc62b251 135 count += newsysinit_end - newsysinit;
984263bc 136 else
dc62b251 137 count += sysinit_end - sysinit;
efda3bd0 138 newset = kmalloc(count * sizeof(*sipp), M_TEMP, M_WAITOK);
984263bc 139 xipp = newset;
dc62b251
MD
140 if (newsysinit) {
141 for (sipp = newsysinit; sipp < newsysinit_end; sipp++)
984263bc 142 *xipp++ = *sipp;
dc62b251
MD
143 } else {
144 for (sipp = sysinit; sipp < sysinit_end; sipp++)
984263bc 145 *xipp++ = *sipp;
dc62b251
MD
146 }
147 for (sipp = set; sipp < set_end; sipp++)
984263bc 148 *xipp++ = *sipp;
984263bc 149 if (newsysinit)
efda3bd0 150 kfree(newsysinit, M_TEMP);
984263bc 151 newsysinit = newset;
dc62b251 152 newsysinit_end = newset + count;
984263bc
MD
153}
154
6bf59cd2
MD
155/*
156 * Callbacks from machine-dependant startup code (e.g. init386) to set
157 * up low level entities related to cpu #0's globaldata.
158 *
159 * Called from very low level boot code.
160 */
161void
162mi_proc0init(struct globaldata *gd, struct user *proc0paddr)
163{
fdce8919 164 lwkt_init_thread(&thread0, proc0paddr, LWKT_THREAD_STACK, 0, gd);
6bf59cd2 165 lwkt_set_comm(&thread0, "thread0");
3e291793 166 RB_INIT(&proc0.p_lwp_tree);
ba87a4ab 167 spin_init(&proc0.p_spin, "iproc_proc0");
5686ec5a 168 lwkt_token_init(&proc0.p_token, "iproc");
526c5c2b
SW
169 lwp0.lwp_tid = 1;
170 proc0.p_lasttid = lwp0.lwp_tid; /* +1 = next TID */
3e291793 171 lwp_rb_tree_RB_INSERT(&proc0.p_lwp_tree, &lwp0);
08f2f1bb
SS
172 lwp0.lwp_thread = &thread0;
173 lwp0.lwp_proc = &proc0;
6bf59cd2 174 proc0.p_usched = usched_init();
c07315c4 175 CPUMASK_ASSALLONES(lwp0.lwp_cpumask);
e2b148c6 176 lwkt_token_init(&lwp0.lwp_token, "lwp_token");
ba87a4ab 177 spin_init(&lwp0.lwp_spin, "iproc_lwp0");
6bf59cd2
MD
178 varsymset_init(&proc0.p_varsymset, NULL);
179 thread0.td_flags |= TDF_RUNNING;
180 thread0.td_proc = &proc0;
08f2f1bb 181 thread0.td_lwp = &lwp0;
6179d4f3 182 thread0.td_switch = cpu_lwkt_switch;
6b7dca13 183 lwkt_schedule_self(curthread);
6bf59cd2
MD
184}
185
984263bc
MD
186/*
187 * System startup; initialize the world, create process 0, mount root
188 * filesystem, and fork to create init and pagedaemon. Most of the
189 * hard work is done in the lower-level initialization routines including
190 * startup(), which does memory initialization and autoconfiguration.
191 *
192 * This allows simple addition of new kernel subsystems that require
193 * boot time initialization. It also allows substitution of subsystem
194 * (for instance, a scheduler, kernel profiler, or VM system) by object
195 * module. Finally, it allows for optional "kernel threads".
196 */
197void
198mi_startup(void)
199{
8a8d5d85 200 struct sysinit *sip; /* system initialization*/
17a9f566
MD
201 struct sysinit **sipp; /* system initialization*/
202 struct sysinit **xipp; /* interior loop of sort*/
203 struct sysinit *save; /* bubble*/
984263bc 204
dc62b251
MD
205 if (sysinit == NULL) {
206 sysinit = SET_BEGIN(sysinit_set);
dbefba87 207#if defined(__x86_64__) && defined(_KERNEL_VIRTUAL)
ca97d438
JG
208 /*
209 * XXX For whatever reason, on 64-bit vkernels
210 * the value of sysinit obtained from the
211 * linker set is wrong.
212 */
213 if ((long)sysinit % 8 != 0) {
8608b858 214 kprintf("Fixing sysinit value...\n");
b12defdc 215 sysinit = (void *)((long)(intptr_t)sysinit + 4);
ca97d438
JG
216 }
217#endif
dc62b251
MD
218 sysinit_end = SET_LIMIT(sysinit_set);
219 }
dbefba87 220#if defined(__x86_64__) && defined(_KERNEL_VIRTUAL)
ca97d438
JG
221 KKASSERT((long)sysinit % 8 == 0);
222#endif
dc62b251 223
984263bc
MD
224restart:
225 /*
226 * Perform a bubble sort of the system initialization objects by
227 * their subsystem (primary key) and order (secondary key).
228 */
dc62b251
MD
229 for (sipp = sysinit; sipp < sysinit_end; sipp++) {
230 for (xipp = sipp + 1; xipp < sysinit_end; xipp++) {
984263bc
MD
231 if ((*sipp)->subsystem < (*xipp)->subsystem ||
232 ((*sipp)->subsystem == (*xipp)->subsystem &&
233 (*sipp)->order <= (*xipp)->order))
234 continue; /* skip*/
235 save = *sipp;
236 *sipp = *xipp;
237 *xipp = save;
238 }
239 }
240
241 /*
242 * Traverse the (now) ordered list of system initialization tasks.
243 * Perform each task, and continue on to the next task.
244 *
245 * The last item on the list is expected to be the scheduler,
246 * which will not return.
247 */
dc62b251
MD
248 for (sipp = sysinit; sipp < sysinit_end; sipp++) {
249 sip = *sipp;
ba39e2e0 250 if (sip->subsystem == SI_SPECIAL_DUMMY)
984263bc
MD
251 continue; /* skip dummy task(s)*/
252
ba39e2e0 253 if (sip->subsystem == SI_SPECIAL_DONE)
984263bc
MD
254 continue;
255
20e194b6 256#if 0
06c66eb2
MD
257 if (bootverbose)
258 kprintf("(%08x-%p)\n", sip->subsystem, sip->func);
20e194b6 259#endif
06c66eb2 260
984263bc 261 /* Call function */
8a8d5d85 262 (*(sip->func))(sip->udata);
984263bc
MD
263
264 /* Check off the one we're just done */
ba39e2e0 265 sip->subsystem = SI_SPECIAL_DONE;
984263bc
MD
266
267 /* Check if we've installed more sysinit items via KLD */
268 if (newsysinit != NULL) {
dc62b251 269 if (sysinit != SET_BEGIN(sysinit_set))
efda3bd0 270 kfree(sysinit, M_TEMP);
984263bc 271 sysinit = newsysinit;
dc62b251 272 sysinit_end = newsysinit_end;
984263bc 273 newsysinit = NULL;
dc62b251 274 newsysinit_end = NULL;
984263bc
MD
275 goto restart;
276 }
277 }
278
279 panic("Shouldn't get here!");
280 /* NOTREACHED*/
281}
282
283
284/*
285 ***************************************************************************
286 ****
287 **** The following SYSINIT's belong elsewhere, but have not yet
288 **** been moved.
289 ****
290 ***************************************************************************
291 */
292static void
b0dfaa52 293print_caddr_t(void *data)
984263bc 294{
6ea70f76 295 kprintf("%s", (char *)data);
984263bc 296}
f3f3eadb 297SYSINIT(announce, SI_BOOT1_COPYRIGHT, SI_ORDER_FIRST, print_caddr_t, copyright);
984263bc 298
f8c3996b
MD
299/*
300 * Leave the critical section that protected us from spurious interrupts
301 * so device probes work.
302 */
303static void
304leavecrit(void *dummy __unused)
305{
7bf5fa56 306 MachIntrABI.stabilize();
0b692e79
MD
307 cpu_enable_intr();
308 MachIntrABI.cleanup();
f8c3996b
MD
309 crit_exit();
310 KKASSERT(!IN_CRITICAL_SECT(curthread));
dbcd0c9b 311
f8c3996b 312 if (bootverbose)
6ea70f76 313 kprintf("Leaving critical section, allowing interrupts\n");
f8c3996b 314}
f3f3eadb 315SYSINIT(leavecrit, SI_BOOT2_LEAVE_CRIT, SI_ORDER_ANY, leavecrit, NULL);
984263bc 316
dbcd0c9b
MD
317/*
318 * This is called after the threading system is up and running,
319 * including the softclock, clock interrupts, and SMP.
320 */
321static void
322tsleepworks(void *dummy __unused)
323{
324 tsleep_now_works = 1;
325}
f3f3eadb 326SYSINIT(tsleepworks, SI_BOOT2_FINISH_SMP, SI_ORDER_SECOND, tsleepworks, NULL);
dbcd0c9b
MD
327
328/*
329 * This is called after devices have configured. Tell the kernel we are
330 * no longer in cold boot.
331 */
332static void
333endofcoldboot(void *dummy __unused)
334{
335 cold = 0;
336}
f3f3eadb 337SYSINIT(endofcoldboot, SI_SUB_ISWARM, SI_ORDER_ANY, endofcoldboot, NULL);
dbcd0c9b 338
984263bc
MD
339/*
340 ***************************************************************************
341 ****
342 **** The two following SYSINT's are proc0 specific glue code. I am not
343 **** convinced that they can not be safely combined, but their order of
344 **** operation has been maintained as the same as the original init_main.c
345 **** for right now.
346 ****
347 **** These probably belong in init_proc.c or kern_proc.c, since they
348 **** deal with proc0 (the fork template process).
349 ****
350 ***************************************************************************
351 */
352/* ARGSUSED*/
353static void
354proc0_init(void *dummy __unused)
355{
0679adc4 356 struct proc *p;
81eea9f4 357 struct lwp *lp;
b3c6f0e5 358 struct uidinfo *uip;
984263bc
MD
359
360 p = &proc0;
08f2f1bb 361 lp = &lwp0;
984263bc 362
315b8b8b
JM
363 /*
364 * Initialize osrel
365 */
366 p->p_osrel = osreldate;
367
984263bc
MD
368 /*
369 * Initialize process and pgrp structures.
370 */
371 procinit();
372
984263bc
MD
373 /*
374 * additional VM structures
375 */
376 vm_init2();
377
378 /*
379 * Create process 0 (the swapper).
380 */
a8d3ab53 381 procinsertinit(p);
b22b6ecd 382 pgrpinsertinit(&pgrp0);
984263bc 383 LIST_INIT(&pgrp0.pg_members);
58c2553a
MD
384 lwkt_token_init(&pgrp0.pg_token, "pgrp0");
385 refcount_init(&pgrp0.pg_refs, 1);
386 lockinit(&pgrp0.pg_lock, "pgwt0", 0, 0);
984263bc
MD
387 LIST_INSERT_HEAD(&pgrp0.pg_members, p, p_pglist);
388
389 pgrp0.pg_session = &session0;
390 session0.s_count = 1;
391 session0.s_leader = p;
b22b6ecd 392 sessinsertinit(&session0);
984263bc 393
58c2553a
MD
394 pgref(&pgrp0);
395 p->p_pgrp = &pgrp0;
396
984263bc
MD
397 p->p_sysent = &aout_sysvec;
398
4643740a 399 p->p_flags = P_SYSTEM;
164b8401
SS
400 p->p_stat = SACTIVE;
401 lp->lwp_stat = LSRUN;
984263bc
MD
402 p->p_nice = NZERO;
403 p->p_rtprio.type = RTP_PRIO_NORMAL;
404 p->p_rtprio.prio = 0;
08f2f1bb 405 lp->lwp_rtprio = p->p_rtprio;
984263bc 406
d8061892 407 p->p_peers = NULL;
984263bc
MD
408 p->p_leader = p;
409
410 bcopy("swapper", p->p_comm, sizeof ("swapper"));
81eea9f4 411 bcopy("swapper", thread0.td_comm, sizeof ("swapper"));
984263bc
MD
412
413 /* Create credentials. */
b3c6f0e5
MD
414 uip = uicreate(0); /* for cr_ruidinfo */
415 uihold(uip); /* for cr_uidinfo */
984263bc 416 p->p_ucred = crget();
b3c6f0e5
MD
417 p->p_ucred->cr_ruidinfo = uip;
418 p->p_ucred->cr_uidinfo = uip;
984263bc 419 p->p_ucred->cr_ngroups = 1; /* group 0 */
9910d07b 420 thread0.td_ucred = crhold(p->p_ucred); /* bootstrap fork1() */
984263bc
MD
421
422 /* Don't jail it */
41c20dac 423 p->p_ucred->cr_prison = NULL;
984263bc 424
b1b4e5a6
SS
425 /* Create sigacts. */
426 p->p_sigacts = &sigacts0;
6fa9e71a 427 refcount_init(&p->p_sigacts->ps_refcnt, 1);
984263bc
MD
428
429 /* Initialize signal state for process 0. */
c0b8a06d 430 siginit(p);
984263bc
MD
431
432 /* Create the file descriptor table. */
228b401d 433 fdinit_bootstrap(p, &filedesc0, cmask);
984263bc
MD
434
435 /* Create the limits structures. */
c0b8a06d 436 plimit_init0(&limit0);
984263bc 437 p->p_limit = &limit0;
984263bc
MD
438
439 /* Allocate a prototype map so we have something to fork. */
440 pmap_pinit0(vmspace_pmap(&vmspace0));
441 p->p_vmspace = &vmspace0;
287ebb09 442 lp->lwp_vmspace = p->p_vmspace;
93f86408 443 vmspace_initrefs(&vmspace0);
e4846942
MD
444 vm_map_init(&vmspace0.vm_map,
445 round_page(VM_MIN_USER_ADDRESS),
446 trunc_page(VM_MAX_USER_ADDRESS),
447 vmspace_pmap(&vmspace0));
984263bc 448
a591f597
MD
449 kqueue_init(&lwp0.lwp_kqueue, &filedesc0);
450
984263bc
MD
451 /*
452 * Charge root for one process.
453 */
41c20dac 454 (void)chgproccnt(p->p_ucred->cr_uidinfo, 1, 0);
ba39e2e0 455 vm_init_limits(p);
984263bc 456}
f3f3eadb 457SYSINIT(p0init, SI_BOOT2_PROC0, SI_ORDER_FIRST, proc0_init, NULL);
984263bc 458
8fa76237
MD
459static int proc0_post_callback(struct proc *p, void *data __unused);
460
984263bc
MD
461/* ARGSUSED*/
462static void
463proc0_post(void *dummy __unused)
464{
465 struct timespec ts;
984263bc
MD
466
467 /*
468 * Now we can look at the time, having had a chance to verify the
469 * time from the file system. Pretend that proc0 started now.
470 */
586c4308 471 allproc_scan(proc0_post_callback, NULL, 0);
984263bc
MD
472
473 /*
474 * Give the ``random'' number generator a thump.
475 * XXX: Does read_random() contain enough bits to be used here ?
476 */
477 nanotime(&ts);
cddfb7bb 478 skrandom(ts.tv_sec ^ ts.tv_nsec);
984263bc 479}
8fa76237
MD
480
481static int
482proc0_post_callback(struct proc *p, void *data __unused)
483{
484 microtime(&p->p_start);
485 return(0);
486}
487
f3f3eadb 488SYSINIT(p0post, SI_SUB_PROC0_POST, SI_ORDER_FIRST, proc0_post, NULL);
984263bc
MD
489
490/*
491 ***************************************************************************
492 ****
493 **** The following SYSINIT's and glue code should be moved to the
494 **** respective files on a per subsystem basis.
495 ****
496 ***************************************************************************
497 */
498
499
500/*
501 ***************************************************************************
502 ****
503 **** The following code probably belongs in another file, like
504 **** kern/init_init.c.
505 ****
506 ***************************************************************************
507 */
508
509/*
510 * List of paths to try when searching for "init".
511 */
512static char init_path[MAXPATHLEN] =
513#ifdef INIT_PATH
514 __XSTRING(INIT_PATH);
515#else
22628b14 516 "/sbin/init:/sbin/oinit:/sbin/init.bak";
984263bc
MD
517#endif
518SYSCTL_STRING(_kern, OID_AUTO, init_path, CTLFLAG_RD, init_path, 0, "");
519
699c19e3
SW
520/*
521 * Shutdown timeout of init(8).
522 * Unused within kernel, but used to control init(8), hence do not remove.
523 */
524#ifndef INIT_SHUTDOWN_TIMEOUT
525#define INIT_SHUTDOWN_TIMEOUT 120
526#endif
527static int init_shutdown_timeout = INIT_SHUTDOWN_TIMEOUT;
528SYSCTL_INT(_kern, OID_AUTO, init_shutdown_timeout,
529 CTLFLAG_RW, &init_shutdown_timeout, 0, "Shutdown timeout of init(8). "
530 "Unused within kernel, but used to control init(8)");
531
984263bc
MD
532/*
533 * Start the initial user process; try exec'ing each pathname in init_path.
534 * The program is invoked with one argument containing the boot flags.
535 */
536static void
91bd9c1e 537start_init(void *dummy, struct trapframe *frame)
984263bc
MD
538{
539 vm_offset_t addr;
540 struct execve_args args;
541 int options, error;
542 char *var, *path, *next, *s;
543 char *ucp, **uap, *arg0, *arg1;
544 struct proc *p;
553ea3c8 545 struct lwp *lp;
21739618
MD
546 struct mount *mp;
547 struct vnode *vp;
fc350ba1
MD
548 char *env;
549
550 /*
551 * This is passed in by the bootloader
552 */
553 env = kgetenv("kernelname");
554 if (env != NULL)
555 strlcpy(kernelname, env, sizeof(kernelname));
984263bc
MD
556
557 p = curproc;
558
08f2f1bb 559 lp = ONLY_LWP_IN_PROC(p);
553ea3c8 560
984263bc 561 /* Get the vnode for '/'. Set p->p_fd->fd_cdir to reference it. */
861905fb 562 mp = mountlist_boot_getfirst();
21739618 563 if (VFS_ROOT(mp, &vp))
984263bc 564 panic("cannot find root vnode");
28623bf9
MD
565 if (mp->mnt_ncmountpt.ncp == NULL) {
566 cache_allocroot(&mp->mnt_ncmountpt, mp, vp);
567 cache_unlock(&mp->mnt_ncmountpt); /* leave ref intact */
8c361dda 568 }
21739618 569 p->p_fd->fd_cdir = vp;
597aea93 570 vref(p->p_fd->fd_cdir);
21739618 571 p->p_fd->fd_rdir = vp;
597aea93 572 vref(p->p_fd->fd_rdir);
28623bf9 573 vfs_cache_setroot(vp, cache_hold(&mp->mnt_ncmountpt));
a11aaa81 574 vn_unlock(vp); /* leave ref intact */
28623bf9
MD
575 cache_copy(&mp->mnt_ncmountpt, &p->p_fd->fd_ncdir);
576 cache_copy(&mp->mnt_ncmountpt, &p->p_fd->fd_nrdir);
984263bc 577
cd29885a
MD
578 kprintf("Mounting devfs\n");
579 vfs_mountroot_devfs();
580
984263bc
MD
581 /*
582 * Need just enough stack to hold the faked-up "execve()" arguments.
583 */
584 addr = trunc_page(USRSTACK - PAGE_SIZE);
0adbcbd6
MD
585 error = vm_map_find(&p->p_vmspace->vm_map, NULL, NULL,
586 0, &addr, PAGE_SIZE,
3091de50
MD
587 PAGE_SIZE, FALSE,
588 VM_MAPTYPE_NORMAL, VM_SUBSYS_INIT,
0adbcbd6 589 VM_PROT_ALL, VM_PROT_ALL, 0);
1b874851 590 if (error)
984263bc
MD
591 panic("init: couldn't allocate argument space");
592 p->p_vmspace->vm_maxsaddr = (caddr_t)addr;
4b566556 593 p->p_vmspace->vm_ssize = PAGE_SIZE;
984263bc 594
bc01a404 595 if ((var = kgetenv("init_path")) != NULL) {
984263bc
MD
596 strncpy(init_path, var, sizeof init_path);
597 init_path[sizeof init_path - 1] = 0;
598 }
315b8b8b 599
984263bc
MD
600 for (path = init_path; *path != '\0'; path = next) {
601 while (*path == ':')
602 path++;
603 if (*path == '\0')
604 break;
605 for (next = path; *next != '\0' && *next != ':'; next++)
606 /* nothing */ ;
607 if (bootverbose)
6ea70f76 608 kprintf("start_init: trying %.*s\n", (int)(next - path),
984263bc
MD
609 path);
610
611 /*
612 * Move out the boot flag argument.
613 */
614 options = 0;
615 ucp = (char *)USRSTACK;
616 (void)subyte(--ucp, 0); /* trailing zero */
617 if (boothowto & RB_SINGLE) {
618 (void)subyte(--ucp, 's');
619 options = 1;
620 }
621#ifdef notyet
622 if (boothowto & RB_FASTBOOT) {
623 (void)subyte(--ucp, 'f');
624 options = 1;
625 }
626#endif
627
628#ifdef BOOTCDROM
629 (void)subyte(--ucp, 'C');
630 options = 1;
631#endif
632 if (options == 0)
633 (void)subyte(--ucp, '-');
634 (void)subyte(--ucp, '-'); /* leading hyphen */
635 arg1 = ucp;
636
637 /*
638 * Move out the file name (also arg 0).
639 */
640 (void)subyte(--ucp, 0);
641 for (s = next - 1; s >= path; s--)
642 (void)subyte(--ucp, *s);
643 arg0 = ucp;
644
645 /*
646 * Move out the arg pointers.
647 */
648 uap = (char **)((intptr_t)ucp & ~(sizeof(intptr_t)-1));
5947157e
MD
649
650 /* terminator */
651 suword64((uint64_t *)(caddr_t)--uap, (uint64_t)0);
652
653 suword64((uint64_t *)(caddr_t)--uap, (uint64_t)(intptr_t)arg1);
654 suword64((uint64_t *)(caddr_t)--uap, (uint64_t)(intptr_t)arg0);
984263bc
MD
655
656 /*
657 * Point at the arguments.
658 */
659 args.fname = arg0;
660 args.argv = uap;
661 args.envv = NULL;
662
663 /*
664 * Now try to exec the program. If can't for any reason
665 * other than it doesn't exist, complain.
666 *
667 * Otherwise, return via fork_trampoline() all the way
668 * to user mode as init!
a2a5ad0d
MD
669 *
670 * WARNING! We may have been moved to another cpu after
0a3f9b47
MD
671 * acquiring the current user process designation. The
672 * MP lock will migrate with us though so we still have to
673 * release it.
984263bc 674 */
753fd850 675 if ((error = sys_execve(&args)) == 0) {
52eedfb5 676 lp->lwp_proc->p_usched->acquire_curproc(lp);
984263bc 677 return;
8a8d5d85 678 }
984263bc 679 if (error != ENOENT)
6ea70f76 680 kprintf("exec %.*s: error %d\n", (int)(next - path),
984263bc
MD
681 path, error);
682 }
6ea70f76 683 kprintf("init: not found in path %s\n", init_path);
984263bc
MD
684 panic("no init");
685}
686
687/*
688 * Like kthread_create(), but runs in it's own address space.
689 * We do this early to reserve pid 1.
690 *
691 * Note special case - do not make it runnable yet. Other work
692 * in progress will change this more.
693 */
694static void
695create_init(const void *udata __unused)
696{
697 int error;
bb3cd951 698 struct lwp *lp;
984263bc 699
e43a034f 700 crit_enter();
08f2f1bb 701 error = fork1(&lwp0, RFFDG | RFPROC, &initproc);
984263bc 702 if (error)
fc92d4aa 703 panic("cannot fork init: %d", error);
4643740a 704 initproc->p_flags |= P_SYSTEM;
6e2a912c 705 reaper_init(initproc, &initreaper);
08f2f1bb 706 lp = ONLY_LWP_IN_PROC(initproc);
bb3cd951 707 cpu_set_fork_handler(lp, start_init, NULL);
e43a034f 708 crit_exit();
984263bc 709}
f3f3eadb 710SYSINIT(init, SI_SUB_CREATE_INIT, SI_ORDER_FIRST, create_init, NULL);
984263bc
MD
711
712/*
713 * Make it runnable now.
714 */
715static void
716kick_init(const void *udata __unused)
717{
08f2f1bb 718 start_forked_proc(&lwp0, initproc);
984263bc 719}
f3f3eadb 720SYSINIT(kickinit, SI_SUB_KTHREAD_INIT, SI_ORDER_FIRST, kick_init, NULL);
8ad65e08 721
0adbcbd6
MD
722static void
723kpmap_init(const void *udata __unused)
724{
725 kpmap = kmalloc(roundup2(sizeof(*kpmap), PAGE_SIZE),
726 M_TEMP, M_ZERO | M_WAITOK);
727
728 kpmap->header[0].type = UKPTYPE_VERSION;
729 kpmap->header[0].offset = offsetof(struct sys_kpmap, version);
730 kpmap->header[1].type = KPTYPE_UPTICKS;
731 kpmap->header[1].offset = offsetof(struct sys_kpmap, upticks);
732 kpmap->header[2].type = KPTYPE_TS_UPTIME;
733 kpmap->header[2].offset = offsetof(struct sys_kpmap, ts_uptime);
734 kpmap->header[3].type = KPTYPE_TS_REALTIME;
735 kpmap->header[3].offset = offsetof(struct sys_kpmap, ts_realtime);
736 kpmap->header[4].type = KPTYPE_TSC_FREQ;
737 kpmap->header[4].offset = offsetof(struct sys_kpmap, tsc_freq);
738 kpmap->header[5].type = KPTYPE_TICK_FREQ;
739 kpmap->header[5].offset = offsetof(struct sys_kpmap, tick_freq);
6b0da239
IV
740 kpmap->header[6].type = KPTYPE_FAST_GTOD;
741 kpmap->header[6].offset = offsetof(struct sys_kpmap, fast_gtod);
0adbcbd6
MD
742 kpmap->version = KPMAP_VERSION;
743}
f3f3eadb 744SYSINIT(kpmapinit, SI_BOOT1_POST, SI_ORDER_FIRST, kpmap_init, NULL);
0adbcbd6 745
8ad65e08
MD
746/*
747 * Machine independant globaldata initialization
73e4f7b9
MD
748 *
749 * WARNING! Called from early boot, 'mycpu' may not work yet.
8ad65e08
MD
750 */
751void
d0e06f83 752mi_gdinit(struct globaldata *gd, int cpuid)
8ad65e08 753{
88c4d2f6 754 TAILQ_INIT(&gd->gd_systimerq);
e3161323 755 gd->gd_sysid_alloc = cpuid; /* prime low bits for cpu lookup */
d0e06f83 756 gd->gd_cpuid = cpuid;
c07315c4 757 CPUMASK_ASSBIT(gd->gd_cpumask, cpuid);
ee89e80b
MD
758 gd->gd_cpumask_simple = CPUMASK_SIMPLE(cpuid);
759 gd->gd_cpumask_offset = (uintptr_t)CPUMASK_ADDR(*(cpumask_t *)0, cpuid);
8ad65e08 760 lwkt_gdinit(gd);
41a01a4d 761 vm_map_entry_reserve_cpu_init(gd);
666ff13c
MD
762 if (gd->gd_cpuid == 0)
763 sleep_early_gdinit(gd);
764 else
765 sleep_gdinit(gd);
c1b91053 766 slab_gdinit(gd);
c07315c4 767 ATOMIC_CPUMASK_ORBIT(usched_global_cpumask, cpuid);
75979118 768 gd->gd_vmstats = vmstats;
8ad65e08
MD
769}
770