]> git.saurik.com Git - apple/xnu.git/blob - bsd/kern/bsd_init.c
xnu-3789.70.16.tar.gz
[apple/xnu.git] / bsd / kern / bsd_init.c
1 /*
2 * Copyright (c) 2000-2015 Apple Inc. All rights reserved.
3 *
4 * @APPLE_OSREFERENCE_LICENSE_HEADER_START@
5 *
6 * This file contains Original Code and/or Modifications of Original Code
7 * as defined in and that are subject to the Apple Public Source License
8 * Version 2.0 (the 'License'). You may not use this file except in
9 * compliance with the License. The rights granted to you under the License
10 * may not be used to create, or enable the creation or redistribution of,
11 * unlawful or unlicensed copies of an Apple operating system, or to
12 * circumvent, violate, or enable the circumvention or violation of, any
13 * terms of an Apple operating system software license agreement.
14 *
15 * Please obtain a copy of the License at
16 * http://www.opensource.apple.com/apsl/ and read it before using this file.
17 *
18 * The Original Code and all software distributed under the License are
19 * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER
20 * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES,
21 * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY,
22 * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT.
23 * Please see the License for the specific language governing rights and
24 * limitations under the License.
25 *
26 * @APPLE_OSREFERENCE_LICENSE_HEADER_END@
27 *
28 *
29 * Copyright (c) 1982, 1986, 1989, 1991, 1992, 1993
30 * The Regents of the University of California. All rights reserved.
31 * (c) UNIX System Laboratories, Inc.
32 * All or some portions of this file are derived from material licensed
33 * to the University of California by American Telephone and Telegraph
34 * Co. or Unix System Laboratories, Inc. and are reproduced herein with
35 * the permission of UNIX System Laboratories, Inc.
36 *
37 * Redistribution and use in source and binary forms, with or without
38 * modification, are permitted provided that the following conditions
39 * are met:
40 * 1. Redistributions of source code must retain the above copyright
41 * notice, this list of conditions and the following disclaimer.
42 * 2. Redistributions in binary form must reproduce the above copyright
43 * notice, this list of conditions and the following disclaimer in the
44 * documentation and/or other materials provided with the distribution.
45 * 3. All advertising materials mentioning features or use of this software
46 * must display the following acknowledgement:
47 * This product includes software developed by the University of
48 * California, Berkeley and its contributors.
49 * 4. Neither the name of the University nor the names of its contributors
50 * may be used to endorse or promote products derived from this software
51 * without specific prior written permission.
52 *
53 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
54 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
55 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
56 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
57 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
58 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
59 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
60 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
61 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
62 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
63 * SUCH DAMAGE.
64 *
65 * @(#)init_main.c 8.16 (Berkeley) 5/14/95
66 */
67
68 /*
69 *
70 * Mach Operating System
71 * Copyright (c) 1987 Carnegie-Mellon University
72 * All rights reserved. The CMU software License Agreement specifies
73 * the terms and conditions for use and redistribution.
74 */
75 /*
76 * NOTICE: This file was modified by McAfee Research in 2004 to introduce
77 * support for mandatory and extensible security protections. This notice
78 * is included in support of clause 2.2 (b) of the Apple Public License,
79 * Version 2.0.
80 */
81
82 #include <sys/param.h>
83 #include <sys/filedesc.h>
84 #include <sys/kernel.h>
85 #include <sys/mount_internal.h>
86 #include <sys/proc_internal.h>
87 #include <sys/kauth.h>
88 #include <sys/systm.h>
89 #include <sys/vnode_internal.h>
90 #include <sys/conf.h>
91 #include <sys/buf_internal.h>
92 #include <sys/clist.h>
93 #include <sys/user.h>
94 #include <sys/time.h>
95 #include <sys/systm.h>
96 #include <sys/mman.h>
97
98 #include <security/audit/audit.h>
99
100 #include <sys/malloc.h>
101 #include <sys/dkstat.h>
102 #include <sys/codesign.h>
103
104 #include <kern/startup.h>
105 #include <kern/thread.h>
106 #include <kern/task.h>
107 #include <kern/ast.h>
108 #include <kern/kalloc.h>
109 #include <mach/mach_host.h>
110
111 #include <mach/vm_param.h>
112
113 #include <vm/vm_map.h>
114 #include <vm/vm_kern.h>
115
116 #include <sys/ux_exception.h> /* for ux_exception_port */
117
118 #include <sys/reboot.h>
119 #include <mach/exception_types.h>
120 #include <dev/busvar.h> /* for pseudo_inits */
121 #include <sys/kdebug.h>
122 #include <sys/reason.h>
123
124 #include <mach/mach_types.h>
125 #include <mach/vm_prot.h>
126 #include <mach/semaphore.h>
127 #include <mach/sync_policy.h>
128 #include <kern/clock.h>
129 #include <mach/kern_return.h>
130 #include <mach/thread_act.h> /* for thread_resume() */
131 #include <mach/task.h> /* for task_set_exception_ports() */
132 #include <sys/ux_exception.h> /* for ux_handler() */
133 #include <sys/ubc_internal.h> /* for ubc_init() */
134 #include <sys/mcache.h> /* for mcache_init() */
135 #include <sys/mbuf.h> /* for mbinit() */
136 #include <sys/event.h> /* for knote_init() */
137 #include <sys/kern_memorystatus.h> /* for memorystatus_init() */
138 #include <sys/aio_kern.h> /* for aio_init() */
139 #include <sys/semaphore.h> /* for psem_cache_init() */
140 #include <net/dlil.h> /* for dlil_init() */
141 #include <net/kpi_protocol.h> /* for proto_kpi_init() */
142 #include <net/iptap.h> /* for iptap_init() */
143 #include <sys/pipe.h> /* for pipeinit() */
144 #include <sys/socketvar.h> /* for socketinit() */
145 #include <sys/protosw.h> /* for domaininit() */
146 #include <kern/sched_prim.h> /* for thread_wakeup() */
147 #include <net/if_ether.h> /* for ether_family_init() */
148 #include <net/if_gif.h> /* for gif_init() */
149 #include <vm/vm_protos.h> /* for vnode_pager_bootstrap() */
150 #include <miscfs/devfs/devfsdefs.h> /* for devfs_kernel_mount() */
151 #include <mach/host_priv.h> /* for host_set_exception_ports() */
152 #include <kern/host.h> /* for host_priv_self() */
153 #include <vm/vm_kern.h> /* for kmem_suballoc() */
154 #include <sys/semaphore.h> /* for psem_lock_init() */
155 #include <sys/msgbuf.h> /* for log_setsize() */
156 #include <sys/tty.h> /* for tty_init() */
157 #include <sys/proc_uuid_policy.h> /* proc_uuid_policy_init() */
158 #include <netinet/flow_divert.h> /* flow_divert_init() */
159 #include <net/content_filter.h> /* for cfil_init() */
160 #include <net/necp.h> /* for necp_init() */
161 #include <net/network_agent.h> /* for netagent_init() */
162 #include <net/packet_mangler.h> /* for pkt_mnglr_init() */
163 #include <net/if_utun.h> /* for utun_register_control() */
164 #include <net/if_ipsec.h> /* for ipsec_register_control() */
165 #include <net/net_str_id.h> /* for net_str_id_init() */
166 #include <net/netsrc.h> /* for netsrc_init() */
167 #include <net/ntstat.h> /* for nstat_init() */
168 #include <netinet/tcp_cc.h> /* for tcp_cc_init() */
169 #include <netinet/mptcp_var.h> /* for mptcp_control_register() */
170 #include <kern/assert.h> /* for assert() */
171 #include <sys/kern_overrides.h> /* for init_system_override() */
172
173 #include <net/init.h>
174
175 #if CONFIG_MACF
176 #include <security/mac_framework.h>
177 #include <security/mac_internal.h> /* mac_init_bsd() */
178 #include <security/mac_mach_internal.h> /* mac_update_task_label() */
179 #endif
180
181 #include <machine/exec.h>
182
183 #if NFSCLIENT
184 #include <sys/netboot.h>
185 #endif
186
187 #if CONFIG_IMAGEBOOT
188 #include <sys/imageboot.h>
189 #endif
190
191 #if PFLOG
192 #include <net/if_pflog.h>
193 #endif
194
195
196 #include <pexpert/pexpert.h>
197 #include <machine/pal_routines.h>
198 #include <console/video_console.h>
199
200
201 void * get_user_regs(thread_t); /* XXX kludge for <machine/thread.h> */
202 void IOKitInitializeTime(void); /* XXX */
203 void IOSleep(unsigned int); /* XXX */
204 void loopattach(void); /* XXX */
205
206 const char copyright[] =
207 "Copyright (c) 1982, 1986, 1989, 1991, 1993\n\t"
208 "The Regents of the University of California. "
209 "All rights reserved.\n\n";
210
211 /* Components of the first process -- never freed. */
212 struct proc proc0;
213 struct session session0;
214 struct pgrp pgrp0;
215 struct filedesc filedesc0;
216 struct plimit limit0;
217 struct pstats pstats0;
218 struct sigacts sigacts0;
219 proc_t kernproc;
220 proc_t initproc;
221
222 long tk_cancc;
223 long tk_nin;
224 long tk_nout;
225 long tk_rawcc;
226
227 int lock_trace = 0;
228 /* Global variables to make pstat happy. We do swapping differently */
229 int nswdev, nswap;
230 int nswapmap;
231 void *swapmap;
232 struct swdevt swdevt[1];
233
234 dev_t rootdev; /* device of the root */
235 dev_t dumpdev; /* device to take dumps on */
236 long dumplo; /* offset into dumpdev */
237 long hostid;
238 char hostname[MAXHOSTNAMELEN];
239 int hostnamelen;
240 char domainname[MAXDOMNAMELEN];
241 int domainnamelen;
242
243 char rootdevice[16]; /* device names have at least 9 chars */
244
245 #if KMEMSTATS
246 struct kmemstats kmemstats[M_LAST];
247 #endif
248
249 struct vnode *rootvp;
250 int boothowto = RB_DEBUG;
251 int minimalboot = 0;
252
253 #if PROC_REF_DEBUG
254 __private_extern__ int proc_ref_tracking_disabled = 0; /* disable panics on leaked proc refs across syscall boundary */
255 #endif
256
257 #if OS_REASON_DEBUG
258 __private_extern__ int os_reason_debug_disabled = 0; /* disable asserts for when we fail to allocate OS reasons */
259 #endif
260
261 extern kern_return_t IOFindBSDRoot(char *, unsigned int, dev_t *, u_int32_t *);
262 extern void IOSecureBSDRoot(const char * rootName);
263 extern kern_return_t IOKitBSDInit(void );
264 extern void kminit(void);
265 extern void file_lock_init(void);
266 extern void kmeminit(void);
267 extern void bsd_bufferinit(void);
268 extern void oslog_setsize(int size);
269 extern void throttle_init(void);
270 extern void acct_init(void);
271
272 extern int serverperfmode;
273 extern int ncl;
274
275 vm_map_t bsd_pageable_map;
276 vm_map_t mb_map;
277
278 static int bsd_simul_execs;
279 static int bsd_pageable_map_size;
280 __private_extern__ int execargs_cache_size = 0;
281 __private_extern__ int execargs_free_count = 0;
282 __private_extern__ vm_offset_t * execargs_cache = NULL;
283
284 void bsd_exec_setup(int);
285
286 __private_extern__ int bootarg_vnode_cache_defeat = 0;
287
288 #if CONFIG_JETSAM && (DEVELOPMENT || DEBUG)
289 __private_extern__ int bootarg_no_vnode_jetsam = 0;
290 #endif /* CONFIG_JETSAM && (DEVELOPMENT || DEBUG) */
291
292 /*
293 * Prevent kernel-based ASLR from being used, for testing.
294 */
295 #if DEVELOPMENT || DEBUG
296 __private_extern__ int bootarg_disable_aslr = 0;
297 #endif
298
299 /*
300 * Allow an alternate dyld to be used for testing.
301 */
302
303 #if DEVELOPMENT || DEBUG
304 char dyld_alt_path[MAXPATHLEN];
305 int use_alt_dyld = 0;
306 #endif
307
308 int cmask = CMASK;
309 extern int customnbuf;
310
311 kern_return_t bsd_autoconf(void);
312 void bsd_utaskbootstrap(void);
313
314 static void parse_bsd_args(void);
315 #if CONFIG_DEV_KMEM
316 extern void dev_kmem_init(void);
317 #endif
318 extern void time_zone_slock_init(void);
319 extern void select_waitq_init(void);
320 static void process_name(const char *, proc_t);
321
322 static void setconf(void);
323
324 #if SYSV_SHM
325 extern void sysv_shm_lock_init(void);
326 #endif
327 #if SYSV_SEM
328 extern void sysv_sem_lock_init(void);
329 #endif
330 #if SYSV_MSG
331 extern void sysv_msg_lock_init(void);
332 #endif
333
334 extern void ulock_initialize(void);
335
336 #if CONFIG_MACF
337 #if defined (__i386__) || defined (__x86_64__)
338 /* MACF policy_check configuration flags; see policy_check.c for details */
339 int policy_check_flags = 0;
340
341 extern int check_policy_init(int);
342 #endif
343 #endif /* CONFIG_MACF */
344
345 /* If we are using CONFIG_DTRACE */
346 #if CONFIG_DTRACE
347 extern void dtrace_postinit(void);
348 #endif
349
350 /*
351 * Initialization code.
352 * Called from cold start routine as
353 * soon as a stack and segmentation
354 * have been established.
355 * Functions:
356 * turn on clock
357 * hand craft 0th process
358 * call all initialization routines
359 * hand craft 1st user process
360 */
361
362 /*
363 * Sets the name for the given task.
364 */
365 static void
366 process_name(const char *s, proc_t p)
367 {
368 strlcpy(p->p_comm, s, sizeof(p->p_comm));
369 strlcpy(p->p_name, s, sizeof(p->p_name));
370 }
371
372 /* To allow these values to be patched, they're globals here */
373 #include <machine/vmparam.h>
374 struct rlimit vm_initial_limit_stack = { DFLSSIZ, MAXSSIZ - PAGE_MAX_SIZE };
375 struct rlimit vm_initial_limit_data = { DFLDSIZ, MAXDSIZ };
376 struct rlimit vm_initial_limit_core = { DFLCSIZ, MAXCSIZ };
377
378 extern thread_t cloneproc(task_t, coalition_t, proc_t, int, int);
379 extern int (*mountroot)(void);
380
381 lck_grp_t * proc_lck_grp;
382 lck_grp_t * proc_slock_grp;
383 lck_grp_t * proc_fdmlock_grp;
384 lck_grp_t * proc_ucred_mlock_grp;
385 lck_grp_t * proc_mlock_grp;
386 lck_grp_attr_t * proc_lck_grp_attr;
387 lck_attr_t * proc_lck_attr;
388 lck_mtx_t * proc_list_mlock;
389 lck_mtx_t * proc_klist_mlock;
390
391
392 extern lck_mtx_t * execargs_cache_lock;
393
394 /* hook called after root is mounted XXX temporary hack */
395 void (*mountroot_post_hook)(void);
396 void (*unmountroot_pre_hook)(void);
397
398 /*
399 * This function is called before IOKit initialization, so that globals
400 * like the sysctl tree are initialized before kernel extensions
401 * are started (since they may want to register sysctls
402 */
403 void
404 bsd_early_init(void)
405 {
406 sysctl_early_init();
407 }
408
409 /*
410 * This function is called very early on in the Mach startup, from the
411 * function start_kernel_threads() in osfmk/kern/startup.c. It's called
412 * in the context of the current (startup) task using a call to the
413 * function kernel_thread_create() to jump into start_kernel_threads().
414 * Internally, kernel_thread_create() calls thread_create_internal(),
415 * which calls uthread_alloc(). The function of uthread_alloc() is
416 * normally to allocate a uthread structure, and fill out the uu_sigmask,
417 * uu_context fields. It skips filling these out in the case of the "task"
418 * being "kernel_task", because the order of operation is inverted. To
419 * account for that, we need to manually fill in at least the contents
420 * of the uu_context.vc_ucred field so that the uthread structure can be
421 * used like any other.
422 */
423
424 void
425 bsd_init(void)
426 {
427 struct uthread *ut;
428 unsigned int i;
429 struct vfs_context context;
430 kern_return_t ret;
431 struct ucred temp_cred;
432 struct posix_cred temp_pcred;
433 #if NFSCLIENT || CONFIG_IMAGEBOOT
434 boolean_t netboot = FALSE;
435 #endif
436
437 #define bsd_init_kprintf(x...) /* kprintf("bsd_init: " x) */
438
439 throttle_init();
440
441 printf(copyright);
442
443 bsd_init_kprintf("calling kmeminit\n");
444 kmeminit();
445
446 bsd_init_kprintf("calling parse_bsd_args\n");
447 parse_bsd_args();
448
449 #if CONFIG_DEV_KMEM
450 bsd_init_kprintf("calling dev_kmem_init\n");
451 dev_kmem_init();
452 #endif
453
454 /* Initialize kauth subsystem before instancing the first credential */
455 bsd_init_kprintf("calling kauth_init\n");
456 kauth_init();
457
458 /* Initialize process and pgrp structures. */
459 bsd_init_kprintf("calling procinit\n");
460 procinit();
461
462 /* Initialize the ttys (MUST be before kminit()/bsd_autoconf()!)*/
463 tty_init();
464
465 kernproc = &proc0; /* implicitly bzero'ed */
466
467 /* kernel_task->proc = kernproc; */
468 set_bsdtask_info(kernel_task,(void *)kernproc);
469
470 /* give kernproc a name */
471 bsd_init_kprintf("calling process_name\n");
472 process_name("kernel_task", kernproc);
473
474 /* allocate proc lock group attribute and group */
475 bsd_init_kprintf("calling lck_grp_attr_alloc_init\n");
476 proc_lck_grp_attr= lck_grp_attr_alloc_init();
477
478 proc_lck_grp = lck_grp_alloc_init("proc", proc_lck_grp_attr);
479 #if CONFIG_FINE_LOCK_GROUPS
480 proc_slock_grp = lck_grp_alloc_init("proc-slock", proc_lck_grp_attr);
481 proc_fdmlock_grp = lck_grp_alloc_init("proc-fdmlock", proc_lck_grp_attr);
482 proc_ucred_mlock_grp = lck_grp_alloc_init("proc-ucred-mlock", proc_lck_grp_attr);
483 proc_mlock_grp = lck_grp_alloc_init("proc-mlock", proc_lck_grp_attr);
484 #endif
485
486 /* Allocate proc lock attribute */
487 proc_lck_attr = lck_attr_alloc_init();
488 #if 0
489 #if __PROC_INTERNAL_DEBUG
490 lck_attr_setdebug(proc_lck_attr);
491 #endif
492 #endif
493
494 #if CONFIG_FINE_LOCK_GROUPS
495 proc_list_mlock = lck_mtx_alloc_init(proc_mlock_grp, proc_lck_attr);
496 proc_klist_mlock = lck_mtx_alloc_init(proc_mlock_grp, proc_lck_attr);
497 lck_mtx_init(&kernproc->p_mlock, proc_mlock_grp, proc_lck_attr);
498 lck_mtx_init(&kernproc->p_fdmlock, proc_fdmlock_grp, proc_lck_attr);
499 lck_mtx_init(&kernproc->p_ucred_mlock, proc_ucred_mlock_grp, proc_lck_attr);
500 lck_spin_init(&kernproc->p_slock, proc_slock_grp, proc_lck_attr);
501 #else
502 proc_list_mlock = lck_mtx_alloc_init(proc_lck_grp, proc_lck_attr);
503 proc_klist_mlock = lck_mtx_alloc_init(proc_lck_grp, proc_lck_attr);
504 lck_mtx_init(&kernproc->p_mlock, proc_lck_grp, proc_lck_attr);
505 lck_mtx_init(&kernproc->p_fdmlock, proc_lck_grp, proc_lck_attr);
506 lck_mtx_init(&kernproc->p_ucred_mlock, proc_lck_grp, proc_lck_attr);
507 lck_spin_init(&kernproc->p_slock, proc_lck_grp, proc_lck_attr);
508 #endif
509
510 assert(bsd_simul_execs != 0);
511 execargs_cache_lock = lck_mtx_alloc_init(proc_lck_grp, proc_lck_attr);
512 execargs_cache_size = bsd_simul_execs;
513 execargs_free_count = bsd_simul_execs;
514 execargs_cache = (vm_offset_t *)kalloc(bsd_simul_execs * sizeof(vm_offset_t));
515 bzero(execargs_cache, bsd_simul_execs * sizeof(vm_offset_t));
516
517 if (current_task() != kernel_task)
518 printf("bsd_init: We have a problem, "
519 "current task is not kernel task\n");
520
521 bsd_init_kprintf("calling get_bsdthread_info\n");
522 ut = (uthread_t)get_bsdthread_info(current_thread());
523
524 #if CONFIG_MACF
525 /*
526 * Initialize the MAC Framework
527 */
528 mac_policy_initbsd();
529 kernproc->p_mac_enforce = 0;
530
531 #if defined (__i386__) || defined (__x86_64__)
532 /*
533 * We currently only support this on i386/x86_64, as that is the
534 * only lock code we have instrumented so far.
535 */
536 check_policy_init(policy_check_flags);
537 #endif
538 #endif /* MAC */
539
540 /* Initialize System Override call */
541 init_system_override();
542
543 ulock_initialize();
544
545 /*
546 * Create process 0.
547 */
548 proc_list_lock();
549 LIST_INSERT_HEAD(&allproc, kernproc, p_list);
550 kernproc->p_pgrp = &pgrp0;
551 LIST_INSERT_HEAD(PGRPHASH(0), &pgrp0, pg_hash);
552 LIST_INIT(&pgrp0.pg_members);
553 #ifdef CONFIG_FINE_LOCK_GROUPS
554 lck_mtx_init(&pgrp0.pg_mlock, proc_mlock_grp, proc_lck_attr);
555 #else
556 lck_mtx_init(&pgrp0.pg_mlock, proc_lck_grp, proc_lck_attr);
557 #endif
558 /* There is no other bsd thread this point and is safe without pgrp lock */
559 LIST_INSERT_HEAD(&pgrp0.pg_members, kernproc, p_pglist);
560 kernproc->p_listflag |= P_LIST_INPGRP;
561 kernproc->p_pgrpid = 0;
562 kernproc->p_uniqueid = 0;
563
564 pgrp0.pg_session = &session0;
565 pgrp0.pg_membercnt = 1;
566
567 session0.s_count = 1;
568 session0.s_leader = kernproc;
569 session0.s_listflags = 0;
570 #ifdef CONFIG_FINE_LOCK_GROUPS
571 lck_mtx_init(&session0.s_mlock, proc_mlock_grp, proc_lck_attr);
572 #else
573 lck_mtx_init(&session0.s_mlock, proc_lck_grp, proc_lck_attr);
574 #endif
575 LIST_INSERT_HEAD(SESSHASH(0), &session0, s_hash);
576 proc_list_unlock();
577
578 #if CONFIG_PERSONAS
579 kernproc->p_persona = NULL;
580 #endif
581
582 kernproc->task = kernel_task;
583
584 kernproc->p_stat = SRUN;
585 kernproc->p_flag = P_SYSTEM;
586 kernproc->p_lflag = 0;
587 kernproc->p_ladvflag = 0;
588
589 #if defined(__LP64__)
590 kernproc->p_flag |= P_LP64;
591 #endif
592
593 #if DEVELOPMENT || DEBUG
594 if (bootarg_disable_aslr)
595 kernproc->p_flag |= P_DISABLE_ASLR;
596 #endif
597
598 kernproc->p_nice = NZERO;
599 kernproc->p_pptr = kernproc;
600
601 TAILQ_INIT(&kernproc->p_uthlist);
602 TAILQ_INSERT_TAIL(&kernproc->p_uthlist, ut, uu_list);
603
604 kernproc->sigwait = FALSE;
605 kernproc->sigwait_thread = THREAD_NULL;
606 kernproc->exit_thread = THREAD_NULL;
607 kernproc->p_csflags = CS_VALID;
608
609 /*
610 * Create credential. This also Initializes the audit information.
611 */
612 bsd_init_kprintf("calling bzero\n");
613 bzero(&temp_cred, sizeof(temp_cred));
614 bzero(&temp_pcred, sizeof(temp_pcred));
615 temp_pcred.cr_ngroups = 1;
616 /* kern_proc, shouldn't call up to DS for group membership */
617 temp_pcred.cr_flags = CRF_NOMEMBERD;
618 temp_cred.cr_audit.as_aia_p = audit_default_aia_p;
619
620 bsd_init_kprintf("calling kauth_cred_create\n");
621 /*
622 * We have to label the temp cred before we create from it to
623 * properly set cr_ngroups, or the create will fail.
624 */
625 posix_cred_label(&temp_cred, &temp_pcred);
626 kernproc->p_ucred = kauth_cred_create(&temp_cred);
627
628 /* update cred on proc */
629 PROC_UPDATE_CREDS_ONPROC(kernproc);
630
631 /* give the (already exisiting) initial thread a reference on it */
632 bsd_init_kprintf("calling kauth_cred_ref\n");
633 kauth_cred_ref(kernproc->p_ucred);
634 ut->uu_context.vc_ucred = kernproc->p_ucred;
635 ut->uu_context.vc_thread = current_thread();
636
637 TAILQ_INIT(&kernproc->p_aio_activeq);
638 TAILQ_INIT(&kernproc->p_aio_doneq);
639 kernproc->p_aio_total_count = 0;
640 kernproc->p_aio_active_count = 0;
641
642 bsd_init_kprintf("calling file_lock_init\n");
643 file_lock_init();
644
645 #if CONFIG_MACF
646 mac_cred_label_associate_kernel(kernproc->p_ucred);
647 #endif
648
649 /* Create the file descriptor table. */
650 kernproc->p_fd = &filedesc0;
651 filedesc0.fd_cmask = cmask;
652 filedesc0.fd_knlistsize = -1;
653 filedesc0.fd_knlist = NULL;
654 filedesc0.fd_knhash = NULL;
655 filedesc0.fd_knhashmask = 0;
656
657 /* Create the limits structures. */
658 kernproc->p_limit = &limit0;
659 for (i = 0; i < sizeof(kernproc->p_rlimit)/sizeof(kernproc->p_rlimit[0]); i++)
660 limit0.pl_rlimit[i].rlim_cur =
661 limit0.pl_rlimit[i].rlim_max = RLIM_INFINITY;
662 limit0.pl_rlimit[RLIMIT_NOFILE].rlim_cur = NOFILE;
663 limit0.pl_rlimit[RLIMIT_NPROC].rlim_cur = maxprocperuid;
664 limit0.pl_rlimit[RLIMIT_NPROC].rlim_max = maxproc;
665 limit0.pl_rlimit[RLIMIT_STACK] = vm_initial_limit_stack;
666 limit0.pl_rlimit[RLIMIT_DATA] = vm_initial_limit_data;
667 limit0.pl_rlimit[RLIMIT_CORE] = vm_initial_limit_core;
668 limit0.pl_refcnt = 1;
669
670 kernproc->p_stats = &pstats0;
671 kernproc->p_sigacts = &sigacts0;
672
673 /*
674 * Charge root for one process: launchd.
675 */
676 bsd_init_kprintf("calling chgproccnt\n");
677 (void)chgproccnt(0, 1);
678
679 /*
680 * Allocate a kernel submap for pageable memory
681 * for temporary copying (execve()).
682 */
683 {
684 vm_offset_t minimum;
685
686 bsd_init_kprintf("calling kmem_suballoc\n");
687 assert(bsd_pageable_map_size != 0);
688 ret = kmem_suballoc(kernel_map,
689 &minimum,
690 (vm_size_t)bsd_pageable_map_size,
691 TRUE,
692 VM_FLAGS_ANYWHERE | VM_MAKE_TAG(VM_KERN_MEMORY_BSD),
693 &bsd_pageable_map);
694 if (ret != KERN_SUCCESS)
695 panic("bsd_init: Failed to allocate bsd pageable map");
696 }
697
698 /*
699 * Initialize buffers and hash links for buffers
700 *
701 * SIDE EFFECT: Starts a thread for bcleanbuf_thread(), so must
702 * happen after a credential has been associated with
703 * the kernel task.
704 */
705 bsd_init_kprintf("calling bsd_bufferinit\n");
706 bsd_bufferinit();
707
708 /* Initialize the execve() semaphore */
709 bsd_init_kprintf("calling semaphore_create\n");
710
711 if (ret != KERN_SUCCESS)
712 panic("bsd_init: Failed to create execve semaphore");
713
714 /*
715 * Initialize the calendar.
716 */
717 bsd_init_kprintf("calling IOKitInitializeTime\n");
718 IOKitInitializeTime();
719
720 bsd_init_kprintf("calling ubc_init\n");
721 ubc_init();
722
723 /* Initialize the file systems. */
724 bsd_init_kprintf("calling vfsinit\n");
725 vfsinit();
726
727 #if CONFIG_PROC_UUID_POLICY
728 /* Initial proc_uuid_policy subsystem */
729 bsd_init_kprintf("calling proc_uuid_policy_init()\n");
730 proc_uuid_policy_init();
731 #endif
732
733 #if SOCKETS
734 /* Initialize per-CPU cache allocator */
735 mcache_init();
736
737 /* Initialize mbuf's. */
738 bsd_init_kprintf("calling mbinit\n");
739 mbinit();
740 net_str_id_init(); /* for mbuf tags */
741 #endif /* SOCKETS */
742
743 /*
744 * Initializes security event auditing.
745 * XXX: Should/could this occur later?
746 */
747 #if CONFIG_AUDIT
748 bsd_init_kprintf("calling audit_init\n");
749 audit_init();
750 #endif
751
752 /* Initialize kqueues */
753 bsd_init_kprintf("calling knote_init\n");
754 knote_init();
755
756 /* Initialize for async IO */
757 bsd_init_kprintf("calling aio_init\n");
758 aio_init();
759
760 /* Initialize pipes */
761 bsd_init_kprintf("calling pipeinit\n");
762 pipeinit();
763
764 /* Initialize SysV shm subsystem locks; the subsystem proper is
765 * initialized through a sysctl.
766 */
767 #if SYSV_SHM
768 bsd_init_kprintf("calling sysv_shm_lock_init\n");
769 sysv_shm_lock_init();
770 #endif
771 #if SYSV_SEM
772 bsd_init_kprintf("calling sysv_sem_lock_init\n");
773 sysv_sem_lock_init();
774 #endif
775 #if SYSV_MSG
776 bsd_init_kprintf("sysv_msg_lock_init\n");
777 sysv_msg_lock_init();
778 #endif
779 bsd_init_kprintf("calling pshm_lock_init\n");
780 pshm_lock_init();
781 bsd_init_kprintf("calling psem_lock_init\n");
782 psem_lock_init();
783
784 pthread_init();
785 /* POSIX Shm and Sem */
786 bsd_init_kprintf("calling pshm_cache_init\n");
787 pshm_cache_init();
788 bsd_init_kprintf("calling psem_cache_init\n");
789 psem_cache_init();
790 bsd_init_kprintf("calling time_zone_slock_init\n");
791 time_zone_slock_init();
792 bsd_init_kprintf("calling select_waitq_init\n");
793 select_waitq_init();
794
795 /*
796 * Initialize protocols. Block reception of incoming packets
797 * until everything is ready.
798 */
799 #if NETWORKING
800 bsd_init_kprintf("calling dlil_init\n");
801 dlil_init();
802 bsd_init_kprintf("calling proto_kpi_init\n");
803 proto_kpi_init();
804 #endif /* NETWORKING */
805 #if SOCKETS
806 bsd_init_kprintf("calling socketinit\n");
807 socketinit();
808 bsd_init_kprintf("calling domaininit\n");
809 domaininit();
810 iptap_init();
811 #if FLOW_DIVERT
812 flow_divert_init();
813 #endif /* FLOW_DIVERT */
814 #endif /* SOCKETS */
815
816 kernproc->p_fd->fd_cdir = NULL;
817 kernproc->p_fd->fd_rdir = NULL;
818
819 #if CONFIG_FREEZE
820 #ifndef CONFIG_MEMORYSTATUS
821 #error "CONFIG_FREEZE defined without matching CONFIG_MEMORYSTATUS"
822 #endif
823 /* Initialise background freezing */
824 bsd_init_kprintf("calling memorystatus_freeze_init\n");
825 memorystatus_freeze_init();
826 #endif
827
828 #if CONFIG_MEMORYSTATUS
829 /* Initialize kernel memory status notifications */
830 bsd_init_kprintf("calling memorystatus_init\n");
831 memorystatus_init();
832 #endif /* CONFIG_MEMORYSTATUS */
833
834 bsd_init_kprintf("calling acct_init\n");
835 acct_init();
836
837 #ifdef GPROF
838 /* Initialize kernel profiling. */
839 kmstartup();
840 #endif
841
842 bsd_init_kprintf("calling sysctl_mib_init\n");
843 sysctl_mib_init()
844
845 bsd_init_kprintf("calling bsd_autoconf\n");
846 bsd_autoconf();
847
848 bsd_init_kprintf("calling os_reason_init\n");
849 os_reason_init();
850
851 #if CONFIG_DTRACE
852 dtrace_postinit();
853 #endif
854
855 /*
856 * We attach the loopback interface *way* down here to ensure
857 * it happens after autoconf(), otherwise it becomes the
858 * "primary" interface.
859 */
860 #include <loop.h>
861 #if NLOOP > 0
862 bsd_init_kprintf("calling loopattach\n");
863 loopattach(); /* XXX */
864 #endif
865 #if NGIF
866 /* Initialize gif interface (after lo0) */
867 gif_init();
868 #endif
869
870 #if PFLOG
871 /* Initialize packet filter log interface */
872 pfloginit();
873 #endif /* PFLOG */
874
875 #if NETHER > 0
876 /* Register the built-in dlil ethernet interface family */
877 bsd_init_kprintf("calling ether_family_init\n");
878 ether_family_init();
879 #endif /* ETHER */
880
881 #if NETWORKING
882 /* Call any kext code that wants to run just after network init */
883 bsd_init_kprintf("calling net_init_run\n");
884 net_init_run();
885
886 #if CONTENT_FILTER
887 cfil_init();
888 #endif
889
890 #if PACKET_MANGLER
891 pkt_mnglr_init();
892 #endif
893
894 #if NECP
895 /* Initialize Network Extension Control Policies */
896 necp_init();
897 #endif
898
899 netagent_init();
900
901 /* register user tunnel kernel control handler */
902 utun_register_control();
903 #if IPSEC
904 ipsec_register_control();
905 #endif /* IPSEC */
906 netsrc_init();
907 nstat_init();
908 tcp_cc_init();
909 #if MPTCP
910 mptcp_control_register();
911 #endif /* MPTCP */
912 #endif /* NETWORKING */
913
914 bsd_init_kprintf("calling vnode_pager_bootstrap\n");
915 vnode_pager_bootstrap();
916
917 bsd_init_kprintf("calling inittodr\n");
918 inittodr(0);
919
920 /* Mount the root file system. */
921 while( TRUE) {
922 int err;
923
924 bsd_init_kprintf("calling setconf\n");
925 setconf();
926 #if NFSCLIENT
927 netboot = (mountroot == netboot_mountroot);
928 #endif
929
930 bsd_init_kprintf("vfs_mountroot\n");
931 if (0 == (err = vfs_mountroot()))
932 break;
933 rootdevice[0] = '\0';
934 #if NFSCLIENT
935 if (netboot) {
936 PE_display_icon( 0, "noroot"); /* XXX a netboot-specific icon would be nicer */
937 vc_progress_set(FALSE, 0);
938 for (i=1; 1; i*=2) {
939 printf("bsd_init: failed to mount network root, error %d, %s\n",
940 err, PE_boot_args());
941 printf("We are hanging here...\n");
942 IOSleep(i*60*1000);
943 }
944 /*NOTREACHED*/
945 }
946 #endif
947 printf("cannot mount root, errno = %d\n", err);
948 boothowto |= RB_ASKNAME;
949 }
950
951 IOSecureBSDRoot(rootdevice);
952
953 context.vc_thread = current_thread();
954 context.vc_ucred = kernproc->p_ucred;
955 mountlist.tqh_first->mnt_flag |= MNT_ROOTFS;
956
957 bsd_init_kprintf("calling VFS_ROOT\n");
958 /* Get the vnode for '/'. Set fdp->fd_fd.fd_cdir to reference it. */
959 if (VFS_ROOT(mountlist.tqh_first, &rootvnode, &context))
960 panic("bsd_init: cannot find root vnode: %s", PE_boot_args());
961 rootvnode->v_flag |= VROOT;
962 (void)vnode_ref(rootvnode);
963 (void)vnode_put(rootvnode);
964 filedesc0.fd_cdir = rootvnode;
965
966 #if NFSCLIENT
967 if (netboot) {
968 int err;
969
970 netboot = TRUE;
971 /* post mount setup */
972 if ((err = netboot_setup()) != 0) {
973 PE_display_icon( 0, "noroot"); /* XXX a netboot-specific icon would be nicer */
974 vc_progress_set(FALSE, 0);
975 for (i=1; 1; i*=2) {
976 printf("bsd_init: NetBoot could not find root, error %d: %s\n",
977 err, PE_boot_args());
978 printf("We are hanging here...\n");
979 IOSleep(i*60*1000);
980 }
981 /*NOTREACHED*/
982 }
983 }
984 #endif
985
986
987 #if CONFIG_IMAGEBOOT
988 /*
989 * See if a system disk image is present. If so, mount it and
990 * switch the root vnode to point to it
991 */
992 if (netboot == FALSE && imageboot_needed()) {
993 /*
994 * An image was found. No turning back: we're booted
995 * with a kernel from the disk image.
996 */
997 imageboot_setup();
998 }
999 #endif /* CONFIG_IMAGEBOOT */
1000
1001 /* set initial time; all other resource data is already zero'ed */
1002 microtime_with_abstime(&kernproc->p_start, &kernproc->p_stats->ps_start);
1003
1004 #if DEVFS
1005 {
1006 char mounthere[] = "/dev"; /* !const because of internal casting */
1007
1008 bsd_init_kprintf("calling devfs_kernel_mount\n");
1009 devfs_kernel_mount(mounthere);
1010 }
1011 #endif /* DEVFS */
1012
1013 /* Initialize signal state for process 0. */
1014 bsd_init_kprintf("calling siginit\n");
1015 siginit(kernproc);
1016
1017 bsd_init_kprintf("calling bsd_utaskbootstrap\n");
1018 bsd_utaskbootstrap();
1019
1020 pal_kernel_announce();
1021
1022 bsd_init_kprintf("calling mountroot_post_hook\n");
1023
1024 /* invoke post-root-mount hook */
1025 if (mountroot_post_hook != NULL)
1026 mountroot_post_hook();
1027
1028 #if 0 /* not yet */
1029 consider_zone_gc();
1030 #endif
1031
1032 bsd_init_kprintf("done\n");
1033 }
1034
1035 void
1036 bsdinit_task(void)
1037 {
1038 proc_t p = current_proc();
1039 struct uthread *ut;
1040 thread_t thread;
1041
1042 process_name("init", p);
1043
1044 ux_handler_init();
1045
1046 thread = current_thread();
1047 (void) host_set_exception_ports(host_priv_self(),
1048 EXC_MASK_ALL & ~(EXC_MASK_RPC_ALERT),//pilotfish (shark) needs this port
1049 (mach_port_t) ux_exception_port,
1050 EXCEPTION_DEFAULT| MACH_EXCEPTION_CODES,
1051 0);
1052
1053 ut = (uthread_t)get_bsdthread_info(thread);
1054
1055 #if CONFIG_MACF
1056 mac_cred_label_associate_user(p->p_ucred);
1057 #endif
1058
1059
1060 load_init_program(p);
1061 lock_trace = 1;
1062 }
1063
1064 kern_return_t
1065 bsd_autoconf(void)
1066 {
1067 kprintf("bsd_autoconf: calling kminit\n");
1068 kminit();
1069
1070 /*
1071 * Early startup for bsd pseudodevices.
1072 */
1073 {
1074 struct pseudo_init *pi;
1075
1076 for (pi = pseudo_inits; pi->ps_func; pi++)
1077 (*pi->ps_func) (pi->ps_count);
1078 }
1079
1080 return( IOKitBSDInit());
1081 }
1082
1083
1084 #include <sys/disklabel.h> /* for MAXPARTITIONS */
1085
1086 static void
1087 setconf(void)
1088 {
1089 u_int32_t flags;
1090 kern_return_t err;
1091
1092 err = IOFindBSDRoot(rootdevice, sizeof(rootdevice), &rootdev, &flags);
1093 if( err) {
1094 printf("setconf: IOFindBSDRoot returned an error (%d);"
1095 "setting rootdevice to 'sd0a'.\n", err); /* XXX DEBUG TEMP */
1096 rootdev = makedev( 6, 0 );
1097 strlcpy(rootdevice, "sd0a", sizeof(rootdevice));
1098 flags = 0;
1099 }
1100
1101 #if NFSCLIENT
1102 if( flags & 1 ) {
1103 /* network device */
1104 mountroot = netboot_mountroot;
1105 } else {
1106 #endif
1107 /* otherwise have vfs determine root filesystem */
1108 mountroot = NULL;
1109 #if NFSCLIENT
1110 }
1111 #endif
1112
1113 }
1114
1115 void
1116 bsd_utaskbootstrap(void)
1117 {
1118 thread_t thread;
1119 struct uthread *ut;
1120
1121 /*
1122 * Clone the bootstrap process from the kernel process, without
1123 * inheriting either task characteristics or memory from the kernel;
1124 */
1125 thread = cloneproc(TASK_NULL, COALITION_NULL, kernproc, FALSE, TRUE);
1126
1127 /* Hold the reference as it will be dropped during shutdown */
1128 initproc = proc_find(1);
1129 #if __PROC_INTERNAL_DEBUG
1130 if (initproc == PROC_NULL)
1131 panic("bsd_utaskbootstrap: initproc not set\n");
1132 #endif
1133 /*
1134 * Since we aren't going back out the normal way to our parent,
1135 * we have to drop the transition locks explicitly.
1136 */
1137 proc_signalend(initproc, 0);
1138 proc_transend(initproc, 0);
1139
1140 ut = (struct uthread *)get_bsdthread_info(thread);
1141 ut->uu_sigmask = 0;
1142 act_set_astbsd(thread);
1143 task_clear_return_wait(get_threadtask(thread));
1144 }
1145
1146 static void
1147 parse_bsd_args(void)
1148 {
1149 char namep[16];
1150 int msgbuf;
1151
1152 if ( PE_parse_boot_argn("-s", namep, sizeof (namep)))
1153 boothowto |= RB_SINGLE;
1154
1155 if (PE_parse_boot_argn("-b", namep, sizeof (namep)))
1156 boothowto |= RB_NOBOOTRC;
1157
1158 if (PE_parse_boot_argn("-x", namep, sizeof (namep))) /* safe boot */
1159 boothowto |= RB_SAFEBOOT;
1160
1161 if (PE_parse_boot_argn("-minimalboot", namep, sizeof(namep))) {
1162 /*
1163 * -minimalboot indicates that we want userspace to be bootstrapped to a
1164 * minimal environment. What constitutes minimal is up to the bootstrap
1165 * process.
1166 */
1167 minimalboot = 1;
1168 }
1169
1170
1171 /* disable vnode_cache_is_authorized() by setting vnode_cache_defeat */
1172 if (PE_parse_boot_argn("-vnode_cache_defeat", namep, sizeof (namep)))
1173 bootarg_vnode_cache_defeat = 1;
1174
1175 #if DEVELOPMENT || DEBUG
1176 if (PE_parse_boot_argn("-disable_aslr", namep, sizeof (namep)))
1177 bootarg_disable_aslr = 1;
1178 #endif
1179
1180 PE_parse_boot_argn("ncl", &ncl, sizeof (ncl));
1181 if (PE_parse_boot_argn("nbuf", &max_nbuf_headers,
1182 sizeof (max_nbuf_headers))) {
1183 customnbuf = 1;
1184 }
1185
1186 #if CONFIG_MACF
1187 #if defined (__i386__) || defined (__x86_64__)
1188 PE_parse_boot_argn("policy_check", &policy_check_flags, sizeof (policy_check_flags));
1189 #endif
1190 #endif /* CONFIG_MACF */
1191
1192 if (PE_parse_boot_argn("msgbuf", &msgbuf, sizeof (msgbuf))) {
1193 log_setsize(msgbuf);
1194 oslog_setsize(msgbuf);
1195 }
1196
1197 if (PE_parse_boot_argn("-novfscache", namep, sizeof(namep))) {
1198 nc_disabled = 1;
1199 }
1200
1201 #if CONFIG_JETSAM && (DEVELOPMENT || DEBUG)
1202 if (PE_parse_boot_argn("-no_vnode_jetsam", namep, sizeof(namep)))
1203 bootarg_no_vnode_jetsam = 1;
1204 #endif /* CONFIG_JETSAM && (DEVELOPMENT || DEBUG) */
1205
1206
1207
1208 #if PROC_REF_DEBUG
1209 if (PE_parse_boot_argn("-disable_procref_tracking", namep, sizeof(namep))) {
1210 proc_ref_tracking_disabled = 1;
1211 }
1212 #endif
1213
1214 #if OS_REASON_DEBUG
1215 if (PE_parse_boot_argn("-disable_osreason_debug", namep, sizeof(namep))) {
1216 os_reason_debug_disabled = 1;
1217 }
1218 #endif
1219
1220 PE_parse_boot_argn("sigrestrict", &sigrestrict_arg, sizeof(sigrestrict_arg));
1221
1222 #if DEVELOPMENT|| DEBUG
1223 if (PE_parse_boot_argn("-no_sigsys", namep, sizeof(namep))) {
1224 send_sigsys = false;
1225 }
1226 #endif
1227
1228 #if (DEVELOPMENT|| DEBUG)
1229 if (PE_parse_boot_argn("alt-dyld", dyld_alt_path, sizeof(dyld_alt_path))) {
1230 if (strlen(dyld_alt_path) > 0) {
1231 use_alt_dyld = 1;
1232 }
1233 }
1234 #endif
1235 }
1236
1237 void
1238 bsd_exec_setup(int scale)
1239 {
1240
1241 switch (scale) {
1242 case 0:
1243 case 1:
1244 bsd_simul_execs = BSD_SIMUL_EXECS;
1245 break;
1246 case 2:
1247 case 3:
1248 bsd_simul_execs = 65;
1249 break;
1250 case 4:
1251 case 5:
1252 bsd_simul_execs = 129;
1253 break;
1254 case 6:
1255 case 7:
1256 bsd_simul_execs = 257;
1257 break;
1258 default:
1259 bsd_simul_execs = 513;
1260 break;
1261
1262 }
1263 bsd_pageable_map_size = (bsd_simul_execs * BSD_PAGEABLE_SIZE_PER_EXEC);
1264 }
1265
1266 #if !NFSCLIENT
1267 int
1268 netboot_root(void);
1269
1270 int
1271 netboot_root(void)
1272 {
1273 return(0);
1274 }
1275 #endif