Commit | Line | Data |
---|---|---|
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 |
84 | int vfs_mountroot_devfs(void); |
85 | ||
984263bc MD |
86 | /* Components of the first process -- never freed. */ |
87 | static struct session session0; | |
88 | static struct pgrp pgrp0; | |
b1b4e5a6 | 89 | static struct sigacts sigacts0; |
0679adc4 | 90 | static struct filedesc filedesc0; |
984263bc MD |
91 | static struct plimit limit0; |
92 | static struct vmspace vmspace0; | |
b7c628e4 MD |
93 | struct proc *initproc; |
94 | struct proc proc0; | |
08f2f1bb | 95 | struct lwp lwp0; |
b7c628e4 | 96 | struct thread thread0; |
984263bc MD |
97 | |
98 | int cmask = CMASK; | |
d87e79b9 | 99 | u_int cpu_mi_feature; |
916e604f | 100 | cpumask_t usched_global_cpumask; |
984263bc MD |
101 | extern struct user *proc0paddr; |
102 | extern int fallback_elf_brand; | |
103 | ||
984263bc | 104 | int boothowto = 0; /* initialized so that it can be patched */ |
0c52fa62 SG |
105 | SYSCTL_INT(_debug, OID_AUTO, boothowto, CTLFLAG_RD, &boothowto, 0, |
106 | "Reboot flags, from console subsystem"); | |
916e604f MD |
107 | SYSCTL_ULONG(_kern, OID_AUTO, usched_global_cpumask, CTLFLAG_RW, |
108 | &usched_global_cpumask, 0, "global user scheduler cpumask"); | |
984263bc MD |
109 | |
110 | /* | |
111 | * This ensures that there is at least one entry so that the sysinit_set | |
7bd34050 | 112 | * symbol is not undefined. A subsystem ID of SI_SPECIAL_DUMMY is never |
984263bc MD |
113 | * executed. |
114 | */ | |
ba39e2e0 | 115 | SYSINIT(placeholder, SI_SPECIAL_DUMMY, SI_ORDER_ANY, NULL, NULL) |
984263bc MD |
116 | |
117 | /* | |
118 | * The sysinit table itself. Items are checked off as the are run. | |
119 | * If we want to register new sysinit types, add them to newsysinit. | |
120 | */ | |
dc62b251 MD |
121 | SET_DECLARE(sysinit_set, struct sysinit); |
122 | struct sysinit **sysinit, **sysinit_end; | |
123 | struct sysinit **newsysinit, **newsysinit_end; | |
124 | ||
984263bc MD |
125 | |
126 | /* | |
127 | * Merge a new sysinit set into the current set, reallocating it if | |
128 | * necessary. This can only be called after malloc is running. | |
129 | */ | |
130 | void | |
dc62b251 | 131 | sysinit_add(struct sysinit **set, struct sysinit **set_end) |
984263bc MD |
132 | { |
133 | struct sysinit **newset; | |
134 | struct sysinit **sipp; | |
135 | struct sysinit **xipp; | |
dc62b251 | 136 | int count; |
984263bc | 137 | |
dc62b251 | 138 | count = set_end - set; |
984263bc | 139 | if (newsysinit) |
dc62b251 | 140 | count += newsysinit_end - newsysinit; |
984263bc | 141 | else |
dc62b251 | 142 | count += sysinit_end - sysinit; |
efda3bd0 | 143 | newset = kmalloc(count * sizeof(*sipp), M_TEMP, M_WAITOK); |
984263bc | 144 | xipp = newset; |
dc62b251 MD |
145 | if (newsysinit) { |
146 | for (sipp = newsysinit; sipp < newsysinit_end; sipp++) | |
984263bc | 147 | *xipp++ = *sipp; |
dc62b251 MD |
148 | } else { |
149 | for (sipp = sysinit; sipp < sysinit_end; sipp++) | |
984263bc | 150 | *xipp++ = *sipp; |
dc62b251 MD |
151 | } |
152 | for (sipp = set; sipp < set_end; sipp++) | |
984263bc | 153 | *xipp++ = *sipp; |
984263bc | 154 | if (newsysinit) |
efda3bd0 | 155 | kfree(newsysinit, M_TEMP); |
984263bc | 156 | newsysinit = newset; |
dc62b251 | 157 | newsysinit_end = newset + count; |
984263bc MD |
158 | } |
159 | ||
6bf59cd2 MD |
160 | /* |
161 | * Callbacks from machine-dependant startup code (e.g. init386) to set | |
162 | * up low level entities related to cpu #0's globaldata. | |
163 | * | |
164 | * Called from very low level boot code. | |
165 | */ | |
166 | void | |
167 | mi_proc0init(struct globaldata *gd, struct user *proc0paddr) | |
168 | { | |
fdce8919 | 169 | lwkt_init_thread(&thread0, proc0paddr, LWKT_THREAD_STACK, 0, gd); |
6bf59cd2 | 170 | lwkt_set_comm(&thread0, "thread0"); |
3e291793 | 171 | RB_INIT(&proc0.p_lwp_tree); |
8f1f6170 | 172 | spin_init(&proc0.p_spin); |
5686ec5a | 173 | lwkt_token_init(&proc0.p_token, "iproc"); |
3e291793 MD |
174 | proc0.p_lasttid = 0; /* +1 = next TID */ |
175 | lwp_rb_tree_RB_INSERT(&proc0.p_lwp_tree, &lwp0); | |
08f2f1bb SS |
176 | lwp0.lwp_thread = &thread0; |
177 | lwp0.lwp_proc = &proc0; | |
6bf59cd2 | 178 | proc0.p_usched = usched_init(); |
da23a592 | 179 | lwp0.lwp_cpumask = (cpumask_t)-1; |
6bf59cd2 MD |
180 | varsymset_init(&proc0.p_varsymset, NULL); |
181 | thread0.td_flags |= TDF_RUNNING; | |
182 | thread0.td_proc = &proc0; | |
08f2f1bb | 183 | thread0.td_lwp = &lwp0; |
6179d4f3 | 184 | thread0.td_switch = cpu_lwkt_switch; |
6b7dca13 | 185 | lwkt_schedule_self(curthread); |
6bf59cd2 MD |
186 | } |
187 | ||
984263bc MD |
188 | /* |
189 | * System startup; initialize the world, create process 0, mount root | |
190 | * filesystem, and fork to create init and pagedaemon. Most of the | |
191 | * hard work is done in the lower-level initialization routines including | |
192 | * startup(), which does memory initialization and autoconfiguration. | |
193 | * | |
194 | * This allows simple addition of new kernel subsystems that require | |
195 | * boot time initialization. It also allows substitution of subsystem | |
196 | * (for instance, a scheduler, kernel profiler, or VM system) by object | |
197 | * module. Finally, it allows for optional "kernel threads". | |
198 | */ | |
199 | void | |
200 | mi_startup(void) | |
201 | { | |
8a8d5d85 | 202 | struct sysinit *sip; /* system initialization*/ |
17a9f566 MD |
203 | struct sysinit **sipp; /* system initialization*/ |
204 | struct sysinit **xipp; /* interior loop of sort*/ | |
205 | struct sysinit *save; /* bubble*/ | |
984263bc | 206 | |
dc62b251 MD |
207 | if (sysinit == NULL) { |
208 | sysinit = SET_BEGIN(sysinit_set); | |
ca97d438 JG |
209 | #if defined(__amd64__) && defined(_KERNEL_VIRTUAL) |
210 | /* | |
211 | * XXX For whatever reason, on 64-bit vkernels | |
212 | * the value of sysinit obtained from the | |
213 | * linker set is wrong. | |
214 | */ | |
215 | if ((long)sysinit % 8 != 0) { | |
8608b858 | 216 | kprintf("Fixing sysinit value...\n"); |
ca97d438 JG |
217 | sysinit = (long)sysinit + 4; |
218 | } | |
219 | #endif | |
dc62b251 MD |
220 | sysinit_end = SET_LIMIT(sysinit_set); |
221 | } | |
ca97d438 JG |
222 | #if defined(__amd64__) && defined(_KERNEL_VIRTUAL) |
223 | KKASSERT((long)sysinit % 8 == 0); | |
224 | #endif | |
dc62b251 | 225 | |
984263bc MD |
226 | restart: |
227 | /* | |
228 | * Perform a bubble sort of the system initialization objects by | |
229 | * their subsystem (primary key) and order (secondary key). | |
230 | */ | |
dc62b251 MD |
231 | for (sipp = sysinit; sipp < sysinit_end; sipp++) { |
232 | for (xipp = sipp + 1; xipp < sysinit_end; xipp++) { | |
984263bc MD |
233 | if ((*sipp)->subsystem < (*xipp)->subsystem || |
234 | ((*sipp)->subsystem == (*xipp)->subsystem && | |
235 | (*sipp)->order <= (*xipp)->order)) | |
236 | continue; /* skip*/ | |
237 | save = *sipp; | |
238 | *sipp = *xipp; | |
239 | *xipp = save; | |
240 | } | |
241 | } | |
242 | ||
243 | /* | |
244 | * Traverse the (now) ordered list of system initialization tasks. | |
245 | * Perform each task, and continue on to the next task. | |
246 | * | |
247 | * The last item on the list is expected to be the scheduler, | |
248 | * which will not return. | |
249 | */ | |
dc62b251 MD |
250 | for (sipp = sysinit; sipp < sysinit_end; sipp++) { |
251 | sip = *sipp; | |
ba39e2e0 | 252 | if (sip->subsystem == SI_SPECIAL_DUMMY) |
984263bc MD |
253 | continue; /* skip dummy task(s)*/ |
254 | ||
ba39e2e0 | 255 | if (sip->subsystem == SI_SPECIAL_DONE) |
984263bc MD |
256 | continue; |
257 | ||
258 | /* Call function */ | |
8a8d5d85 | 259 | (*(sip->func))(sip->udata); |
984263bc MD |
260 | |
261 | /* Check off the one we're just done */ | |
ba39e2e0 | 262 | sip->subsystem = SI_SPECIAL_DONE; |
984263bc MD |
263 | |
264 | /* Check if we've installed more sysinit items via KLD */ | |
265 | if (newsysinit != NULL) { | |
dc62b251 | 266 | if (sysinit != SET_BEGIN(sysinit_set)) |
efda3bd0 | 267 | kfree(sysinit, M_TEMP); |
984263bc | 268 | sysinit = newsysinit; |
dc62b251 | 269 | sysinit_end = newsysinit_end; |
984263bc | 270 | newsysinit = NULL; |
dc62b251 | 271 | newsysinit_end = NULL; |
984263bc MD |
272 | goto restart; |
273 | } | |
274 | } | |
275 | ||
276 | panic("Shouldn't get here!"); | |
277 | /* NOTREACHED*/ | |
278 | } | |
279 | ||
280 | ||
281 | /* | |
282 | *************************************************************************** | |
283 | **** | |
284 | **** The following SYSINIT's belong elsewhere, but have not yet | |
285 | **** been moved. | |
286 | **** | |
287 | *************************************************************************** | |
288 | */ | |
289 | static void | |
290 | print_caddr_t(void *data __unused) | |
291 | { | |
6ea70f76 | 292 | kprintf("%s", (char *)data); |
984263bc | 293 | } |
ba39e2e0 | 294 | SYSINIT(announce, SI_BOOT1_COPYRIGHT, SI_ORDER_FIRST, print_caddr_t, copyright) |
984263bc | 295 | |
f8c3996b MD |
296 | /* |
297 | * Leave the critical section that protected us from spurious interrupts | |
298 | * so device probes work. | |
299 | */ | |
300 | static void | |
301 | leavecrit(void *dummy __unused) | |
302 | { | |
7bf5fa56 | 303 | MachIntrABI.stabilize(); |
0b692e79 MD |
304 | cpu_enable_intr(); |
305 | MachIntrABI.cleanup(); | |
f8c3996b MD |
306 | crit_exit(); |
307 | KKASSERT(!IN_CRITICAL_SECT(curthread)); | |
dbcd0c9b | 308 | |
f8c3996b | 309 | if (bootverbose) |
6ea70f76 | 310 | kprintf("Leaving critical section, allowing interrupts\n"); |
f8c3996b | 311 | } |
ba39e2e0 | 312 | SYSINIT(leavecrit, SI_BOOT2_LEAVE_CRIT, SI_ORDER_ANY, leavecrit, NULL) |
984263bc | 313 | |
dbcd0c9b MD |
314 | /* |
315 | * This is called after the threading system is up and running, | |
316 | * including the softclock, clock interrupts, and SMP. | |
317 | */ | |
318 | static void | |
319 | tsleepworks(void *dummy __unused) | |
320 | { | |
321 | tsleep_now_works = 1; | |
322 | } | |
323 | SYSINIT(tsleepworks, SI_BOOT2_FINISH_SMP, SI_ORDER_SECOND, tsleepworks, NULL) | |
324 | ||
325 | /* | |
326 | * This is called after devices have configured. Tell the kernel we are | |
327 | * no longer in cold boot. | |
328 | */ | |
329 | static void | |
330 | endofcoldboot(void *dummy __unused) | |
331 | { | |
332 | cold = 0; | |
333 | } | |
334 | SYSINIT(endofcoldboot, SI_SUB_ISWARM, SI_ORDER_ANY, endofcoldboot, NULL) | |
335 | ||
984263bc MD |
336 | /* |
337 | *************************************************************************** | |
338 | **** | |
339 | **** The two following SYSINT's are proc0 specific glue code. I am not | |
340 | **** convinced that they can not be safely combined, but their order of | |
341 | **** operation has been maintained as the same as the original init_main.c | |
342 | **** for right now. | |
343 | **** | |
344 | **** These probably belong in init_proc.c or kern_proc.c, since they | |
345 | **** deal with proc0 (the fork template process). | |
346 | **** | |
347 | *************************************************************************** | |
348 | */ | |
349 | /* ARGSUSED*/ | |
350 | static void | |
351 | proc0_init(void *dummy __unused) | |
352 | { | |
0679adc4 | 353 | struct proc *p; |
81eea9f4 | 354 | struct lwp *lp; |
984263bc MD |
355 | |
356 | p = &proc0; | |
08f2f1bb | 357 | lp = &lwp0; |
984263bc MD |
358 | |
359 | /* | |
360 | * Initialize process and pgrp structures. | |
361 | */ | |
362 | procinit(); | |
363 | ||
984263bc MD |
364 | /* |
365 | * additional VM structures | |
366 | */ | |
367 | vm_init2(); | |
368 | ||
369 | /* | |
370 | * Create process 0 (the swapper). | |
371 | */ | |
372 | LIST_INSERT_HEAD(&allproc, p, p_list); | |
373 | p->p_pgrp = &pgrp0; | |
374 | LIST_INSERT_HEAD(PGRPHASH(0), &pgrp0, pg_hash); | |
375 | LIST_INIT(&pgrp0.pg_members); | |
376 | LIST_INSERT_HEAD(&pgrp0.pg_members, p, p_pglist); | |
377 | ||
378 | pgrp0.pg_session = &session0; | |
379 | session0.s_count = 1; | |
380 | session0.s_leader = p; | |
381 | ||
382 | p->p_sysent = &aout_sysvec; | |
383 | ||
344ad853 | 384 | p->p_flag = P_SYSTEM; |
164b8401 SS |
385 | p->p_stat = SACTIVE; |
386 | lp->lwp_stat = LSRUN; | |
984263bc MD |
387 | p->p_nice = NZERO; |
388 | p->p_rtprio.type = RTP_PRIO_NORMAL; | |
389 | p->p_rtprio.prio = 0; | |
08f2f1bb | 390 | lp->lwp_rtprio = p->p_rtprio; |
984263bc MD |
391 | |
392 | p->p_peers = 0; | |
393 | p->p_leader = p; | |
394 | ||
395 | bcopy("swapper", p->p_comm, sizeof ("swapper")); | |
81eea9f4 | 396 | bcopy("swapper", thread0.td_comm, sizeof ("swapper")); |
984263bc MD |
397 | |
398 | /* Create credentials. */ | |
984263bc | 399 | p->p_ucred = crget(); |
41c20dac | 400 | p->p_ucred->cr_ruidinfo = uifind(0); |
984263bc MD |
401 | p->p_ucred->cr_ngroups = 1; /* group 0 */ |
402 | p->p_ucred->cr_uidinfo = uifind(0); | |
9910d07b | 403 | thread0.td_ucred = crhold(p->p_ucred); /* bootstrap fork1() */ |
984263bc MD |
404 | |
405 | /* Don't jail it */ | |
41c20dac | 406 | p->p_ucred->cr_prison = NULL; |
984263bc | 407 | |
b1b4e5a6 SS |
408 | /* Create sigacts. */ |
409 | p->p_sigacts = &sigacts0; | |
410 | p->p_sigacts->ps_refcnt = 1; | |
984263bc MD |
411 | |
412 | /* Initialize signal state for process 0. */ | |
c0b8a06d | 413 | siginit(p); |
984263bc MD |
414 | |
415 | /* Create the file descriptor table. */ | |
228b401d | 416 | fdinit_bootstrap(p, &filedesc0, cmask); |
984263bc MD |
417 | |
418 | /* Create the limits structures. */ | |
c0b8a06d | 419 | plimit_init0(&limit0); |
984263bc | 420 | p->p_limit = &limit0; |
984263bc MD |
421 | |
422 | /* Allocate a prototype map so we have something to fork. */ | |
423 | pmap_pinit0(vmspace_pmap(&vmspace0)); | |
424 | p->p_vmspace = &vmspace0; | |
287ebb09 | 425 | lp->lwp_vmspace = p->p_vmspace; |
e3161323 | 426 | sysref_init(&vmspace0.vm_sysref, &vmspace_sysref_class); |
e4846942 MD |
427 | vm_map_init(&vmspace0.vm_map, |
428 | round_page(VM_MIN_USER_ADDRESS), | |
429 | trunc_page(VM_MAX_USER_ADDRESS), | |
430 | vmspace_pmap(&vmspace0)); | |
e3161323 | 431 | sysref_activate(&vmspace0.vm_sysref); |
984263bc | 432 | |
a591f597 MD |
433 | kqueue_init(&lwp0.lwp_kqueue, &filedesc0); |
434 | ||
984263bc MD |
435 | /* |
436 | * Charge root for one process. | |
437 | */ | |
41c20dac | 438 | (void)chgproccnt(p->p_ucred->cr_uidinfo, 1, 0); |
ba39e2e0 | 439 | vm_init_limits(p); |
984263bc | 440 | } |
ba39e2e0 | 441 | SYSINIT(p0init, SI_BOOT2_PROC0, SI_ORDER_FIRST, proc0_init, NULL) |
984263bc | 442 | |
8fa76237 MD |
443 | static int proc0_post_callback(struct proc *p, void *data __unused); |
444 | ||
984263bc MD |
445 | /* ARGSUSED*/ |
446 | static void | |
447 | proc0_post(void *dummy __unused) | |
448 | { | |
449 | struct timespec ts; | |
984263bc MD |
450 | |
451 | /* | |
452 | * Now we can look at the time, having had a chance to verify the | |
453 | * time from the file system. Pretend that proc0 started now. | |
454 | */ | |
8fa76237 | 455 | allproc_scan(proc0_post_callback, NULL); |
984263bc MD |
456 | |
457 | /* | |
458 | * Give the ``random'' number generator a thump. | |
459 | * XXX: Does read_random() contain enough bits to be used here ? | |
460 | */ | |
461 | nanotime(&ts); | |
cddfb7bb | 462 | skrandom(ts.tv_sec ^ ts.tv_nsec); |
984263bc | 463 | } |
8fa76237 MD |
464 | |
465 | static int | |
466 | proc0_post_callback(struct proc *p, void *data __unused) | |
467 | { | |
468 | microtime(&p->p_start); | |
469 | return(0); | |
470 | } | |
471 | ||
ba39e2e0 | 472 | SYSINIT(p0post, SI_SUB_PROC0_POST, SI_ORDER_FIRST, proc0_post, NULL) |
984263bc MD |
473 | |
474 | /* | |
475 | *************************************************************************** | |
476 | **** | |
477 | **** The following SYSINIT's and glue code should be moved to the | |
478 | **** respective files on a per subsystem basis. | |
479 | **** | |
480 | *************************************************************************** | |
481 | */ | |
482 | ||
483 | ||
484 | /* | |
485 | *************************************************************************** | |
486 | **** | |
487 | **** The following code probably belongs in another file, like | |
488 | **** kern/init_init.c. | |
489 | **** | |
490 | *************************************************************************** | |
491 | */ | |
492 | ||
493 | /* | |
494 | * List of paths to try when searching for "init". | |
495 | */ | |
496 | static char init_path[MAXPATHLEN] = | |
497 | #ifdef INIT_PATH | |
498 | __XSTRING(INIT_PATH); | |
499 | #else | |
22628b14 | 500 | "/sbin/init:/sbin/oinit:/sbin/init.bak"; |
984263bc MD |
501 | #endif |
502 | SYSCTL_STRING(_kern, OID_AUTO, init_path, CTLFLAG_RD, init_path, 0, ""); | |
503 | ||
504 | /* | |
505 | * Start the initial user process; try exec'ing each pathname in init_path. | |
506 | * The program is invoked with one argument containing the boot flags. | |
507 | */ | |
508 | static void | |
91bd9c1e | 509 | start_init(void *dummy, struct trapframe *frame) |
984263bc MD |
510 | { |
511 | vm_offset_t addr; | |
512 | struct execve_args args; | |
513 | int options, error; | |
514 | char *var, *path, *next, *s; | |
515 | char *ucp, **uap, *arg0, *arg1; | |
516 | struct proc *p; | |
553ea3c8 | 517 | struct lwp *lp; |
21739618 MD |
518 | struct mount *mp; |
519 | struct vnode *vp; | |
fc350ba1 MD |
520 | char *env; |
521 | ||
522 | /* | |
523 | * This is passed in by the bootloader | |
524 | */ | |
525 | env = kgetenv("kernelname"); | |
526 | if (env != NULL) | |
527 | strlcpy(kernelname, env, sizeof(kernelname)); | |
984263bc | 528 | |
2b0bd8aa MD |
529 | /* |
530 | * The MP lock is not held on entry. We release it before | |
531 | * returning to userland. | |
532 | */ | |
533 | get_mplock(); | |
984263bc MD |
534 | p = curproc; |
535 | ||
08f2f1bb | 536 | lp = ONLY_LWP_IN_PROC(p); |
553ea3c8 | 537 | |
984263bc | 538 | /* Get the vnode for '/'. Set p->p_fd->fd_cdir to reference it. */ |
861905fb | 539 | mp = mountlist_boot_getfirst(); |
21739618 | 540 | if (VFS_ROOT(mp, &vp)) |
984263bc | 541 | panic("cannot find root vnode"); |
28623bf9 MD |
542 | if (mp->mnt_ncmountpt.ncp == NULL) { |
543 | cache_allocroot(&mp->mnt_ncmountpt, mp, vp); | |
544 | cache_unlock(&mp->mnt_ncmountpt); /* leave ref intact */ | |
8c361dda | 545 | } |
21739618 | 546 | p->p_fd->fd_cdir = vp; |
597aea93 | 547 | vref(p->p_fd->fd_cdir); |
21739618 | 548 | p->p_fd->fd_rdir = vp; |
597aea93 | 549 | vref(p->p_fd->fd_rdir); |
28623bf9 | 550 | vfs_cache_setroot(vp, cache_hold(&mp->mnt_ncmountpt)); |
a11aaa81 | 551 | vn_unlock(vp); /* leave ref intact */ |
28623bf9 MD |
552 | cache_copy(&mp->mnt_ncmountpt, &p->p_fd->fd_ncdir); |
553 | cache_copy(&mp->mnt_ncmountpt, &p->p_fd->fd_nrdir); | |
984263bc | 554 | |
cd29885a MD |
555 | kprintf("Mounting devfs\n"); |
556 | vfs_mountroot_devfs(); | |
557 | ||
984263bc MD |
558 | /* |
559 | * Need just enough stack to hold the faked-up "execve()" arguments. | |
560 | */ | |
561 | addr = trunc_page(USRSTACK - PAGE_SIZE); | |
9388fcaa MD |
562 | error = vm_map_find(&p->p_vmspace->vm_map, NULL, 0, &addr, |
563 | PAGE_SIZE, PAGE_SIZE, | |
564 | FALSE, VM_MAPTYPE_NORMAL, | |
1b874851 MD |
565 | VM_PROT_ALL, VM_PROT_ALL, |
566 | 0); | |
567 | if (error) | |
984263bc MD |
568 | panic("init: couldn't allocate argument space"); |
569 | p->p_vmspace->vm_maxsaddr = (caddr_t)addr; | |
570 | p->p_vmspace->vm_ssize = 1; | |
571 | ||
bc01a404 | 572 | if ((var = kgetenv("init_path")) != NULL) { |
984263bc MD |
573 | strncpy(init_path, var, sizeof init_path); |
574 | init_path[sizeof init_path - 1] = 0; | |
575 | } | |
bc01a404 | 576 | if ((var = kgetenv("kern.fallback_elf_brand")) != NULL) |
984263bc MD |
577 | fallback_elf_brand = strtol(var, NULL, 0); |
578 | ||
579 | for (path = init_path; *path != '\0'; path = next) { | |
580 | while (*path == ':') | |
581 | path++; | |
582 | if (*path == '\0') | |
583 | break; | |
584 | for (next = path; *next != '\0' && *next != ':'; next++) | |
585 | /* nothing */ ; | |
586 | if (bootverbose) | |
6ea70f76 | 587 | kprintf("start_init: trying %.*s\n", (int)(next - path), |
984263bc MD |
588 | path); |
589 | ||
590 | /* | |
591 | * Move out the boot flag argument. | |
592 | */ | |
593 | options = 0; | |
594 | ucp = (char *)USRSTACK; | |
595 | (void)subyte(--ucp, 0); /* trailing zero */ | |
596 | if (boothowto & RB_SINGLE) { | |
597 | (void)subyte(--ucp, 's'); | |
598 | options = 1; | |
599 | } | |
600 | #ifdef notyet | |
601 | if (boothowto & RB_FASTBOOT) { | |
602 | (void)subyte(--ucp, 'f'); | |
603 | options = 1; | |
604 | } | |
605 | #endif | |
606 | ||
607 | #ifdef BOOTCDROM | |
608 | (void)subyte(--ucp, 'C'); | |
609 | options = 1; | |
610 | #endif | |
611 | if (options == 0) | |
612 | (void)subyte(--ucp, '-'); | |
613 | (void)subyte(--ucp, '-'); /* leading hyphen */ | |
614 | arg1 = ucp; | |
615 | ||
616 | /* | |
617 | * Move out the file name (also arg 0). | |
618 | */ | |
619 | (void)subyte(--ucp, 0); | |
620 | for (s = next - 1; s >= path; s--) | |
621 | (void)subyte(--ucp, *s); | |
622 | arg0 = ucp; | |
623 | ||
624 | /* | |
625 | * Move out the arg pointers. | |
626 | */ | |
627 | uap = (char **)((intptr_t)ucp & ~(sizeof(intptr_t)-1)); | |
628 | (void)suword((caddr_t)--uap, (long)0); /* terminator */ | |
629 | (void)suword((caddr_t)--uap, (long)(intptr_t)arg1); | |
630 | (void)suword((caddr_t)--uap, (long)(intptr_t)arg0); | |
631 | ||
632 | /* | |
633 | * Point at the arguments. | |
634 | */ | |
635 | args.fname = arg0; | |
636 | args.argv = uap; | |
637 | args.envv = NULL; | |
638 | ||
639 | /* | |
640 | * Now try to exec the program. If can't for any reason | |
641 | * other than it doesn't exist, complain. | |
642 | * | |
643 | * Otherwise, return via fork_trampoline() all the way | |
644 | * to user mode as init! | |
a2a5ad0d MD |
645 | * |
646 | * WARNING! We may have been moved to another cpu after | |
0a3f9b47 MD |
647 | * acquiring the current user process designation. The |
648 | * MP lock will migrate with us though so we still have to | |
649 | * release it. | |
984263bc | 650 | */ |
753fd850 | 651 | if ((error = sys_execve(&args)) == 0) { |
8a8d5d85 | 652 | rel_mplock(); |
52eedfb5 | 653 | lp->lwp_proc->p_usched->acquire_curproc(lp); |
984263bc | 654 | return; |
8a8d5d85 | 655 | } |
984263bc | 656 | if (error != ENOENT) |
6ea70f76 | 657 | kprintf("exec %.*s: error %d\n", (int)(next - path), |
984263bc MD |
658 | path, error); |
659 | } | |
6ea70f76 | 660 | kprintf("init: not found in path %s\n", init_path); |
984263bc MD |
661 | panic("no init"); |
662 | } | |
663 | ||
664 | /* | |
665 | * Like kthread_create(), but runs in it's own address space. | |
666 | * We do this early to reserve pid 1. | |
667 | * | |
668 | * Note special case - do not make it runnable yet. Other work | |
669 | * in progress will change this more. | |
670 | */ | |
671 | static void | |
672 | create_init(const void *udata __unused) | |
673 | { | |
674 | int error; | |
bb3cd951 | 675 | struct lwp *lp; |
984263bc | 676 | |
e43a034f | 677 | crit_enter(); |
08f2f1bb | 678 | error = fork1(&lwp0, RFFDG | RFPROC, &initproc); |
984263bc | 679 | if (error) |
fc92d4aa | 680 | panic("cannot fork init: %d", error); |
344ad853 | 681 | initproc->p_flag |= P_SYSTEM; |
08f2f1bb | 682 | lp = ONLY_LWP_IN_PROC(initproc); |
bb3cd951 | 683 | cpu_set_fork_handler(lp, start_init, NULL); |
e43a034f | 684 | crit_exit(); |
984263bc | 685 | } |
ba39e2e0 | 686 | SYSINIT(init, SI_SUB_CREATE_INIT, SI_ORDER_FIRST, create_init, NULL) |
984263bc MD |
687 | |
688 | /* | |
689 | * Make it runnable now. | |
690 | */ | |
691 | static void | |
692 | kick_init(const void *udata __unused) | |
693 | { | |
08f2f1bb | 694 | start_forked_proc(&lwp0, initproc); |
984263bc | 695 | } |
ba39e2e0 | 696 | SYSINIT(kickinit, SI_SUB_KTHREAD_INIT, SI_ORDER_FIRST, kick_init, NULL) |
8ad65e08 MD |
697 | |
698 | /* | |
699 | * Machine independant globaldata initialization | |
73e4f7b9 MD |
700 | * |
701 | * WARNING! Called from early boot, 'mycpu' may not work yet. | |
8ad65e08 MD |
702 | */ |
703 | void | |
d0e06f83 | 704 | mi_gdinit(struct globaldata *gd, int cpuid) |
8ad65e08 | 705 | { |
88c4d2f6 | 706 | TAILQ_INIT(&gd->gd_systimerq); |
e3161323 | 707 | gd->gd_sysid_alloc = cpuid; /* prime low bits for cpu lookup */ |
d0e06f83 | 708 | gd->gd_cpuid = cpuid; |
da23a592 | 709 | gd->gd_cpumask = CPUMASK(cpuid); |
8ad65e08 | 710 | lwkt_gdinit(gd); |
41a01a4d | 711 | vm_map_entry_reserve_cpu_init(gd); |
fc17ad60 | 712 | sleep_gdinit(gd); |
916e604f | 713 | usched_global_cpumask |= CPUMASK(cpuid); |
8ad65e08 MD |
714 | } |
715 |