kernel - Provide descriptions for lwkt.* and debug.* sysctl's
[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.
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 * $FreeBSD: src/sys/kern/init_main.c,v 1.134.2.8 2003/06/06 20:21:32 tegge Exp $
7bd34050 43 * $DragonFly: src/sys/kern/init_main.c,v 1.87 2008/06/07 11:37:23 mneumann Exp $
984263bc
MD
44 */
45
46#include "opt_init_path.h"
47
48#include <sys/param.h>
49#include <sys/file.h>
50#include <sys/filedesc.h>
51#include <sys/kernel.h>
52#include <sys/mount.h>
53#include <sys/sysctl.h>
54#include <sys/proc.h>
55#include <sys/resourcevar.h>
56#include <sys/signalvar.h>
57#include <sys/systm.h>
58#include <sys/vnode.h>
59#include <sys/sysent.h>
60#include <sys/reboot.h>
61#include <sys/sysproto.h>
62#include <sys/vmmeter.h>
63#include <sys/unistd.h>
64#include <sys/malloc.h>
0b692e79 65#include <sys/machintr.h>
684a93c4 66
dadab5e9 67#include <sys/file2.h>
f8c3996b 68#include <sys/thread2.h>
e3161323 69#include <sys/sysref2.h>
8f1f6170 70#include <sys/spinlock2.h>
684a93c4 71#include <sys/mplock2.h>
984263bc
MD
72
73#include <machine/cpu.h>
74
75#include <vm/vm.h>
76#include <vm/vm_param.h>
77#include <sys/lock.h>
78#include <vm/pmap.h>
79#include <vm/vm_map.h>
ba39e2e0 80#include <vm/vm_extern.h>
984263bc
MD
81#include <sys/user.h>
82#include <sys/copyright.h>
83
cd29885a
MD
84int vfs_mountroot_devfs(void);
85
984263bc
MD
86/* Components of the first process -- never freed. */
87static struct session session0;
88static struct pgrp pgrp0;
b1b4e5a6 89static struct sigacts sigacts0;
0679adc4 90static struct filedesc filedesc0;
984263bc
MD
91static struct plimit limit0;
92static struct vmspace vmspace0;
b7c628e4
MD
93struct proc *initproc;
94struct proc proc0;
08f2f1bb 95struct lwp lwp0;
b7c628e4 96struct thread thread0;
984263bc
MD
97
98int cmask = CMASK;
99extern struct user *proc0paddr;
100extern int fallback_elf_brand;
101
984263bc 102int boothowto = 0; /* initialized so that it can be patched */
0c52fa62
SG
103SYSCTL_INT(_debug, OID_AUTO, boothowto, CTLFLAG_RD, &boothowto, 0,
104 "Reboot flags, from console subsystem");
984263bc
MD
105
106/*
107 * This ensures that there is at least one entry so that the sysinit_set
7bd34050 108 * symbol is not undefined. A subsystem ID of SI_SPECIAL_DUMMY is never
984263bc
MD
109 * executed.
110 */
ba39e2e0 111SYSINIT(placeholder, SI_SPECIAL_DUMMY, SI_ORDER_ANY, NULL, NULL)
984263bc
MD
112
113/*
114 * The sysinit table itself. Items are checked off as the are run.
115 * If we want to register new sysinit types, add them to newsysinit.
116 */
dc62b251
MD
117SET_DECLARE(sysinit_set, struct sysinit);
118struct sysinit **sysinit, **sysinit_end;
119struct sysinit **newsysinit, **newsysinit_end;
120
984263bc
MD
121
122/*
123 * Merge a new sysinit set into the current set, reallocating it if
124 * necessary. This can only be called after malloc is running.
125 */
126void
dc62b251 127sysinit_add(struct sysinit **set, struct sysinit **set_end)
984263bc
MD
128{
129 struct sysinit **newset;
130 struct sysinit **sipp;
131 struct sysinit **xipp;
dc62b251 132 int count;
984263bc 133
dc62b251 134 count = set_end - set;
984263bc 135 if (newsysinit)
dc62b251 136 count += newsysinit_end - newsysinit;
984263bc 137 else
dc62b251 138 count += sysinit_end - sysinit;
efda3bd0 139 newset = kmalloc(count * sizeof(*sipp), M_TEMP, M_WAITOK);
984263bc 140 xipp = newset;
dc62b251
MD
141 if (newsysinit) {
142 for (sipp = newsysinit; sipp < newsysinit_end; sipp++)
984263bc 143 *xipp++ = *sipp;
dc62b251
MD
144 } else {
145 for (sipp = sysinit; sipp < sysinit_end; sipp++)
984263bc 146 *xipp++ = *sipp;
dc62b251
MD
147 }
148 for (sipp = set; sipp < set_end; sipp++)
984263bc 149 *xipp++ = *sipp;
984263bc 150 if (newsysinit)
efda3bd0 151 kfree(newsysinit, M_TEMP);
984263bc 152 newsysinit = newset;
dc62b251 153 newsysinit_end = newset + count;
984263bc
MD
154}
155
156/*
6bf59cd2
MD
157 * Callbacks from machine-dependant startup code (e.g. init386) to set
158 * up low level entities related to cpu #0's globaldata.
159 *
160 * Called from very low level boot code.
161 */
162void
163mi_proc0init(struct globaldata *gd, struct user *proc0paddr)
164{
fdce8919 165 lwkt_init_thread(&thread0, proc0paddr, LWKT_THREAD_STACK, 0, gd);
6bf59cd2 166 lwkt_set_comm(&thread0, "thread0");
c9e9fb21
MD
167#ifdef SMP
168 thread0.td_mpcount = 1; /* will hold mplock initially */
169#endif
3e291793 170 RB_INIT(&proc0.p_lwp_tree);
8f1f6170 171 spin_init(&proc0.p_spin);
3e291793
MD
172 proc0.p_lasttid = 0; /* +1 = next TID */
173 lwp_rb_tree_RB_INSERT(&proc0.p_lwp_tree, &lwp0);
08f2f1bb
SS
174 lwp0.lwp_thread = &thread0;
175 lwp0.lwp_proc = &proc0;
6bf59cd2 176 proc0.p_usched = usched_init();
08f2f1bb 177 lwp0.lwp_cpumask = 0xFFFFFFFF;
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
186/*
984263bc
MD
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);
ca97d438
JG
207#if defined(__amd64__) && defined(_KERNEL_VIRTUAL)
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");
ca97d438
JG
215 sysinit = (long)sysinit + 4;
216 }
217#endif
dc62b251
MD
218 sysinit_end = SET_LIMIT(sysinit_set);
219 }
ca97d438
JG
220#if defined(__amd64__) && defined(_KERNEL_VIRTUAL)
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
256 /* Call function */
8a8d5d85 257 (*(sip->func))(sip->udata);
984263bc
MD
258
259 /* Check off the one we're just done */
ba39e2e0 260 sip->subsystem = SI_SPECIAL_DONE;
984263bc
MD
261
262 /* Check if we've installed more sysinit items via KLD */
263 if (newsysinit != NULL) {
dc62b251 264 if (sysinit != SET_BEGIN(sysinit_set))
efda3bd0 265 kfree(sysinit, M_TEMP);
984263bc 266 sysinit = newsysinit;
dc62b251 267 sysinit_end = newsysinit_end;
984263bc 268 newsysinit = NULL;
dc62b251 269 newsysinit_end = NULL;
984263bc
MD
270 goto restart;
271 }
272 }
273
274 panic("Shouldn't get here!");
275 /* NOTREACHED*/
276}
277
278
279/*
280 ***************************************************************************
281 ****
282 **** The following SYSINIT's belong elsewhere, but have not yet
283 **** been moved.
284 ****
285 ***************************************************************************
286 */
287static void
288print_caddr_t(void *data __unused)
289{
6ea70f76 290 kprintf("%s", (char *)data);
984263bc 291}
ba39e2e0 292SYSINIT(announce, SI_BOOT1_COPYRIGHT, SI_ORDER_FIRST, print_caddr_t, copyright)
984263bc 293
f8c3996b
MD
294/*
295 * Leave the critical section that protected us from spurious interrupts
296 * so device probes work.
297 */
298static void
299leavecrit(void *dummy __unused)
300{
0b692e79
MD
301 MachIntrABI.finalize();
302 cpu_enable_intr();
303 MachIntrABI.cleanup();
f8c3996b
MD
304 crit_exit();
305 KKASSERT(!IN_CRITICAL_SECT(curthread));
dbcd0c9b 306
f8c3996b 307 if (bootverbose)
6ea70f76 308 kprintf("Leaving critical section, allowing interrupts\n");
f8c3996b 309}
ba39e2e0 310SYSINIT(leavecrit, SI_BOOT2_LEAVE_CRIT, SI_ORDER_ANY, leavecrit, NULL)
984263bc
MD
311
312/*
dbcd0c9b
MD
313 * This is called after the threading system is up and running,
314 * including the softclock, clock interrupts, and SMP.
315 */
316static void
317tsleepworks(void *dummy __unused)
318{
319 tsleep_now_works = 1;
320}
321SYSINIT(tsleepworks, SI_BOOT2_FINISH_SMP, SI_ORDER_SECOND, tsleepworks, NULL)
322
323/*
324 * This is called after devices have configured. Tell the kernel we are
325 * no longer in cold boot.
326 */
327static void
328endofcoldboot(void *dummy __unused)
329{
330 cold = 0;
331}
332SYSINIT(endofcoldboot, SI_SUB_ISWARM, SI_ORDER_ANY, endofcoldboot, NULL)
333
334/*
984263bc
MD
335 ***************************************************************************
336 ****
337 **** The two following SYSINT's are proc0 specific glue code. I am not
338 **** convinced that they can not be safely combined, but their order of
339 **** operation has been maintained as the same as the original init_main.c
340 **** for right now.
341 ****
342 **** These probably belong in init_proc.c or kern_proc.c, since they
343 **** deal with proc0 (the fork template process).
344 ****
345 ***************************************************************************
346 */
347/* ARGSUSED*/
348static void
349proc0_init(void *dummy __unused)
350{
0679adc4 351 struct proc *p;
81eea9f4 352 struct lwp *lp;
984263bc
MD
353
354 p = &proc0;
08f2f1bb 355 lp = &lwp0;
984263bc
MD
356
357 /*
358 * Initialize process and pgrp structures.
359 */
360 procinit();
361
362 /*
984263bc
MD
363 * additional VM structures
364 */
365 vm_init2();
366
367 /*
368 * Create process 0 (the swapper).
369 */
370 LIST_INSERT_HEAD(&allproc, p, p_list);
371 p->p_pgrp = &pgrp0;
372 LIST_INSERT_HEAD(PGRPHASH(0), &pgrp0, pg_hash);
373 LIST_INIT(&pgrp0.pg_members);
374 LIST_INSERT_HEAD(&pgrp0.pg_members, p, p_pglist);
375
376 pgrp0.pg_session = &session0;
377 session0.s_count = 1;
378 session0.s_leader = p;
379
380 p->p_sysent = &aout_sysvec;
381
344ad853 382 p->p_flag = P_SYSTEM;
164b8401
SS
383 p->p_stat = SACTIVE;
384 lp->lwp_stat = LSRUN;
984263bc
MD
385 p->p_nice = NZERO;
386 p->p_rtprio.type = RTP_PRIO_NORMAL;
387 p->p_rtprio.prio = 0;
08f2f1bb 388 lp->lwp_rtprio = p->p_rtprio;
984263bc
MD
389
390 p->p_peers = 0;
391 p->p_leader = p;
392
393 bcopy("swapper", p->p_comm, sizeof ("swapper"));
81eea9f4 394 bcopy("swapper", thread0.td_comm, sizeof ("swapper"));
984263bc
MD
395
396 /* Create credentials. */
984263bc 397 p->p_ucred = crget();
41c20dac 398 p->p_ucred->cr_ruidinfo = uifind(0);
984263bc
MD
399 p->p_ucred->cr_ngroups = 1; /* group 0 */
400 p->p_ucred->cr_uidinfo = uifind(0);
9910d07b 401 thread0.td_ucred = crhold(p->p_ucred); /* bootstrap fork1() */
984263bc
MD
402
403 /* Don't jail it */
41c20dac 404 p->p_ucred->cr_prison = NULL;
984263bc 405
b1b4e5a6
SS
406 /* Create sigacts. */
407 p->p_sigacts = &sigacts0;
408 p->p_sigacts->ps_refcnt = 1;
984263bc
MD
409
410 /* Initialize signal state for process 0. */
c0b8a06d 411 siginit(p);
984263bc
MD
412
413 /* Create the file descriptor table. */
228b401d 414 fdinit_bootstrap(p, &filedesc0, cmask);
984263bc
MD
415
416 /* Create the limits structures. */
c0b8a06d 417 plimit_init0(&limit0);
984263bc 418 p->p_limit = &limit0;
984263bc
MD
419
420 /* Allocate a prototype map so we have something to fork. */
421 pmap_pinit0(vmspace_pmap(&vmspace0));
422 p->p_vmspace = &vmspace0;
287ebb09 423 lp->lwp_vmspace = p->p_vmspace;
e3161323 424 sysref_init(&vmspace0.vm_sysref, &vmspace_sysref_class);
e4846942
MD
425 vm_map_init(&vmspace0.vm_map,
426 round_page(VM_MIN_USER_ADDRESS),
427 trunc_page(VM_MAX_USER_ADDRESS),
428 vmspace_pmap(&vmspace0));
e3161323 429 sysref_activate(&vmspace0.vm_sysref);
984263bc 430
a591f597
MD
431 kqueue_init(&lwp0.lwp_kqueue, &filedesc0);
432
984263bc 433 /*
984263bc
MD
434 * Charge root for one process.
435 */
41c20dac 436 (void)chgproccnt(p->p_ucred->cr_uidinfo, 1, 0);
ba39e2e0 437 vm_init_limits(p);
984263bc 438}
ba39e2e0 439SYSINIT(p0init, SI_BOOT2_PROC0, SI_ORDER_FIRST, proc0_init, NULL)
984263bc 440
8fa76237
MD
441static int proc0_post_callback(struct proc *p, void *data __unused);
442
984263bc
MD
443/* ARGSUSED*/
444static void
445proc0_post(void *dummy __unused)
446{
447 struct timespec ts;
984263bc
MD
448
449 /*
450 * Now we can look at the time, having had a chance to verify the
451 * time from the file system. Pretend that proc0 started now.
452 */
8fa76237 453 allproc_scan(proc0_post_callback, NULL);
984263bc
MD
454
455 /*
456 * Give the ``random'' number generator a thump.
457 * XXX: Does read_random() contain enough bits to be used here ?
458 */
459 nanotime(&ts);
cddfb7bb 460 skrandom(ts.tv_sec ^ ts.tv_nsec);
984263bc 461}
8fa76237
MD
462
463static int
464proc0_post_callback(struct proc *p, void *data __unused)
465{
466 microtime(&p->p_start);
467 return(0);
468}
469
ba39e2e0 470SYSINIT(p0post, SI_SUB_PROC0_POST, SI_ORDER_FIRST, proc0_post, NULL)
984263bc
MD
471
472/*
473 ***************************************************************************
474 ****
475 **** The following SYSINIT's and glue code should be moved to the
476 **** respective files on a per subsystem basis.
477 ****
478 ***************************************************************************
479 */
480
481
482/*
483 ***************************************************************************
484 ****
485 **** The following code probably belongs in another file, like
486 **** kern/init_init.c.
487 ****
488 ***************************************************************************
489 */
490
491/*
492 * List of paths to try when searching for "init".
493 */
494static char init_path[MAXPATHLEN] =
495#ifdef INIT_PATH
496 __XSTRING(INIT_PATH);
497#else
22628b14 498 "/sbin/init:/sbin/oinit:/sbin/init.bak";
984263bc
MD
499#endif
500SYSCTL_STRING(_kern, OID_AUTO, init_path, CTLFLAG_RD, init_path, 0, "");
501
502/*
503 * Start the initial user process; try exec'ing each pathname in init_path.
504 * The program is invoked with one argument containing the boot flags.
505 */
506static void
91bd9c1e 507start_init(void *dummy, struct trapframe *frame)
984263bc
MD
508{
509 vm_offset_t addr;
510 struct execve_args args;
511 int options, error;
512 char *var, *path, *next, *s;
513 char *ucp, **uap, *arg0, *arg1;
514 struct proc *p;
553ea3c8 515 struct lwp *lp;
21739618
MD
516 struct mount *mp;
517 struct vnode *vp;
fc350ba1
MD
518 char *env;
519
520 /*
521 * This is passed in by the bootloader
522 */
523 env = kgetenv("kernelname");
524 if (env != NULL)
525 strlcpy(kernelname, env, sizeof(kernelname));
984263bc 526
2b0bd8aa
MD
527 /*
528 * The MP lock is not held on entry. We release it before
529 * returning to userland.
530 */
531 get_mplock();
984263bc
MD
532 p = curproc;
533
08f2f1bb 534 lp = ONLY_LWP_IN_PROC(p);
553ea3c8 535
984263bc 536 /* Get the vnode for '/'. Set p->p_fd->fd_cdir to reference it. */
861905fb 537 mp = mountlist_boot_getfirst();
21739618 538 if (VFS_ROOT(mp, &vp))
984263bc 539 panic("cannot find root vnode");
28623bf9
MD
540 if (mp->mnt_ncmountpt.ncp == NULL) {
541 cache_allocroot(&mp->mnt_ncmountpt, mp, vp);
542 cache_unlock(&mp->mnt_ncmountpt); /* leave ref intact */
8c361dda 543 }
21739618 544 p->p_fd->fd_cdir = vp;
597aea93 545 vref(p->p_fd->fd_cdir);
21739618 546 p->p_fd->fd_rdir = vp;
597aea93 547 vref(p->p_fd->fd_rdir);
28623bf9 548 vfs_cache_setroot(vp, cache_hold(&mp->mnt_ncmountpt));
a11aaa81 549 vn_unlock(vp); /* leave ref intact */
28623bf9
MD
550 cache_copy(&mp->mnt_ncmountpt, &p->p_fd->fd_ncdir);
551 cache_copy(&mp->mnt_ncmountpt, &p->p_fd->fd_nrdir);
984263bc 552
cd29885a
MD
553 kprintf("Mounting devfs\n");
554 vfs_mountroot_devfs();
555
984263bc
MD
556 /*
557 * Need just enough stack to hold the faked-up "execve()" arguments.
558 */
559 addr = trunc_page(USRSTACK - PAGE_SIZE);
9388fcaa
MD
560 error = vm_map_find(&p->p_vmspace->vm_map, NULL, 0, &addr,
561 PAGE_SIZE, PAGE_SIZE,
562 FALSE, VM_MAPTYPE_NORMAL,
1b874851
MD
563 VM_PROT_ALL, VM_PROT_ALL,
564 0);
565 if (error)
984263bc
MD
566 panic("init: couldn't allocate argument space");
567 p->p_vmspace->vm_maxsaddr = (caddr_t)addr;
568 p->p_vmspace->vm_ssize = 1;
569
bc01a404 570 if ((var = kgetenv("init_path")) != NULL) {
984263bc
MD
571 strncpy(init_path, var, sizeof init_path);
572 init_path[sizeof init_path - 1] = 0;
573 }
bc01a404 574 if ((var = kgetenv("kern.fallback_elf_brand")) != NULL)
984263bc
MD
575 fallback_elf_brand = strtol(var, NULL, 0);
576
577 for (path = init_path; *path != '\0'; path = next) {
578 while (*path == ':')
579 path++;
580 if (*path == '\0')
581 break;
582 for (next = path; *next != '\0' && *next != ':'; next++)
583 /* nothing */ ;
584 if (bootverbose)
6ea70f76 585 kprintf("start_init: trying %.*s\n", (int)(next - path),
984263bc
MD
586 path);
587
588 /*
589 * Move out the boot flag argument.
590 */
591 options = 0;
592 ucp = (char *)USRSTACK;
593 (void)subyte(--ucp, 0); /* trailing zero */
594 if (boothowto & RB_SINGLE) {
595 (void)subyte(--ucp, 's');
596 options = 1;
597 }
598#ifdef notyet
599 if (boothowto & RB_FASTBOOT) {
600 (void)subyte(--ucp, 'f');
601 options = 1;
602 }
603#endif
604
605#ifdef BOOTCDROM
606 (void)subyte(--ucp, 'C');
607 options = 1;
608#endif
609 if (options == 0)
610 (void)subyte(--ucp, '-');
611 (void)subyte(--ucp, '-'); /* leading hyphen */
612 arg1 = ucp;
613
614 /*
615 * Move out the file name (also arg 0).
616 */
617 (void)subyte(--ucp, 0);
618 for (s = next - 1; s >= path; s--)
619 (void)subyte(--ucp, *s);
620 arg0 = ucp;
621
622 /*
623 * Move out the arg pointers.
624 */
625 uap = (char **)((intptr_t)ucp & ~(sizeof(intptr_t)-1));
626 (void)suword((caddr_t)--uap, (long)0); /* terminator */
627 (void)suword((caddr_t)--uap, (long)(intptr_t)arg1);
628 (void)suword((caddr_t)--uap, (long)(intptr_t)arg0);
629
630 /*
631 * Point at the arguments.
632 */
633 args.fname = arg0;
634 args.argv = uap;
635 args.envv = NULL;
636
637 /*
638 * Now try to exec the program. If can't for any reason
639 * other than it doesn't exist, complain.
640 *
641 * Otherwise, return via fork_trampoline() all the way
642 * to user mode as init!
a2a5ad0d
MD
643 *
644 * WARNING! We may have been moved to another cpu after
0a3f9b47
MD
645 * acquiring the current user process designation. The
646 * MP lock will migrate with us though so we still have to
647 * release it.
984263bc 648 */
753fd850 649 if ((error = sys_execve(&args)) == 0) {
8a8d5d85 650 rel_mplock();
52eedfb5 651 lp->lwp_proc->p_usched->acquire_curproc(lp);
984263bc 652 return;
8a8d5d85 653 }
984263bc 654 if (error != ENOENT)
6ea70f76 655 kprintf("exec %.*s: error %d\n", (int)(next - path),
984263bc
MD
656 path, error);
657 }
6ea70f76 658 kprintf("init: not found in path %s\n", init_path);
984263bc
MD
659 panic("no init");
660}
661
662/*
663 * Like kthread_create(), but runs in it's own address space.
664 * We do this early to reserve pid 1.
665 *
666 * Note special case - do not make it runnable yet. Other work
667 * in progress will change this more.
668 */
669static void
670create_init(const void *udata __unused)
671{
672 int error;
bb3cd951 673 struct lwp *lp;
984263bc 674
e43a034f 675 crit_enter();
08f2f1bb 676 error = fork1(&lwp0, RFFDG | RFPROC, &initproc);
984263bc 677 if (error)
fc92d4aa 678 panic("cannot fork init: %d", error);
344ad853 679 initproc->p_flag |= P_SYSTEM;
08f2f1bb 680 lp = ONLY_LWP_IN_PROC(initproc);
bb3cd951 681 cpu_set_fork_handler(lp, start_init, NULL);
e43a034f 682 crit_exit();
984263bc 683}
ba39e2e0 684SYSINIT(init, SI_SUB_CREATE_INIT, SI_ORDER_FIRST, create_init, NULL)
984263bc
MD
685
686/*
687 * Make it runnable now.
688 */
689static void
690kick_init(const void *udata __unused)
691{
08f2f1bb 692 start_forked_proc(&lwp0, initproc);
984263bc 693}
ba39e2e0 694SYSINIT(kickinit, SI_SUB_KTHREAD_INIT, SI_ORDER_FIRST, kick_init, NULL)
8ad65e08
MD
695
696/*
697 * Machine independant globaldata initialization
73e4f7b9
MD
698 *
699 * WARNING! Called from early boot, 'mycpu' may not work yet.
8ad65e08
MD
700 */
701void
d0e06f83 702mi_gdinit(struct globaldata *gd, int cpuid)
8ad65e08 703{
88c4d2f6 704 TAILQ_INIT(&gd->gd_systimerq);
e3161323 705 gd->gd_sysid_alloc = cpuid; /* prime low bits for cpu lookup */
d0e06f83 706 gd->gd_cpuid = cpuid;
fda1ad89 707 gd->gd_cpumask = (cpumask_t)1 << cpuid;
8ad65e08 708 lwkt_gdinit(gd);
41a01a4d 709 vm_map_entry_reserve_cpu_init(gd);
fc17ad60 710 sleep_gdinit(gd);
8ad65e08
MD
711}
712