bsd/kern/bsd_init.c

   1 /*
   2  * Copyright (c) 2000-2009 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
 123 #include <mach/mach_types.h>
 124 #include <mach/vm_prot.h>
 125 #include <mach/semaphore.h>
 126 #include <mach/sync_policy.h>
 127 #include <kern/clock.h>
 128 #include <mach/kern_return.h>
 129 #include <mach/thread_act.h>            /* for thread_resume() */
 130 #include <mach/task.h>                  /* for task_set_exception_ports() */
 131 #include <sys/ux_exception.h>           /* for ux_handler() */
 132 #include <sys/ubc_internal.h>           /* for ubc_init() */
 133 #include <sys/mcache.h>                 /* for mcache_init() */
 134 #include <sys/mbuf.h>                   /* for mbinit() */
 135 #include <sys/event.h>                  /* for knote_init() */
 136 #include <sys/kern_memorystatus.h>      /* for kern_memorystatus_init() */
 137 #include <sys/aio_kern.h>               /* for aio_init() */
 138 #include <sys/semaphore.h>              /* for psem_cache_init() */
 139 #include <net/dlil.h>                   /* for dlil_init() */
 140 #include <net/kpi_protocol.h>           /* for proto_kpi_init() */
 141 #include <sys/pipe.h>                   /* for pipeinit() */
 142 #include <sys/socketvar.h>              /* for socketinit() */
 143 #include <sys/protosw.h>                /* for domaininit() */
 144 #include <kern/sched_prim.h>            /* for thread_wakeup() */
 145 #include <net/if_ether.h>               /* for ether_family_init() */
 146 #include <vm/vm_protos.h>               /* for vnode_pager_bootstrap() */
 147 #include <miscfs/devfs/devfsdefs.h>     /* for devfs_kernel_mount() */
 148 #include <mach/host_priv.h>             /* for host_set_exception_ports() */
 149 #include <kern/host.h>                  /* for host_priv_self() */
 150 #include <vm/vm_kern.h>                 /* for kmem_suballoc() */
 151 #include <sys/semaphore.h>              /* for psem_lock_init() */
 152 #include <sys/msgbuf.h>                 /* for log_setsize() */
 153 #include <sys/tty.h>                    /* for tty_init() */
 154 #include <net/if_utun.h>                /* for utun_register_control() */
 155 #include <net/net_str_id.h>             /* for net_str_id_init() */
 156 #include <net/netsrc.h>                 /* for netsrc_init() */
 157 #include <kern/assert.h>                /* for assert() */
 158
 159 #include <net/init.h>
 160
 161 #if CONFIG_MACF
 162 #include <security/mac_framework.h>
 163 #include <security/mac_internal.h>      /* mac_init_bsd() */
 164 #include <security/mac_mach_internal.h> /* mac_update_task_label() */
 165 #endif
 166
 167 #include <machine/exec.h>
 168
 169 #if NFSCLIENT
 170 #include <sys/netboot.h>
 171 #endif
 172
 173 #if CONFIG_IMAGEBOOT
 174 #include <sys/imageboot.h>
 175 #endif
 176
 177 #if PFLOG
 178 #include <net/if_pflog.h>
 179 #endif
 180
 181 #include <pexpert/pexpert.h>
 182 #include <machine/pal_routines.h>
 183
 184 void * get_user_regs(thread_t);         /* XXX kludge for <machine/thread.h> */
 185 void IOKitInitializeTime(void);         /* XXX */
 186 void IOSleep(unsigned int);             /* XXX */
 187 void loopattach(void);                  /* XXX */
 188 void vc_progress_set(boolean_t, uint32_t); /* XXX */
 189
 190 const char    copyright[] =
 191 "Copyright (c) 1982, 1986, 1989, 1991, 1993\n\t"
 192 "The Regents of the University of California. "
 193 "All rights reserved.\n\n";
 194
 195 /* Components of the first process -- never freed. */
 196 struct  proc proc0;
 197 struct  session session0;
 198 struct  pgrp pgrp0;
 199 struct  filedesc filedesc0;
 200 struct  plimit limit0;
 201 struct  pstats pstats0;
 202 struct  sigacts sigacts0;
 203 proc_t kernproc;
 204 proc_t initproc;
 205
 206 long tk_cancc;
 207 long tk_nin;
 208 long tk_nout;
 209 long tk_rawcc;
 210
 211 int lock_trace = 0;
 212 /* Global variables to make pstat happy. We do swapping differently */
 213 int nswdev, nswap;
 214 int nswapmap;
 215 void *swapmap;
 216 struct swdevt swdevt[1];
 217
 218 dev_t   rootdev;                /* device of the root */
 219 dev_t   dumpdev;                /* device to take dumps on */
 220 long    dumplo;                 /* offset into dumpdev */
 221 long    hostid;
 222 char    hostname[MAXHOSTNAMELEN];
 223 int             hostnamelen;
 224 char    domainname[MAXDOMNAMELEN];
 225 int             domainnamelen;
 226 #if defined(__i386__) || defined(__x86_64__)
 227 struct exec_archhandler exec_archhandler_ppc = {
 228         .path = "/usr/libexec/oah/RosettaNonGrata",
 229 };
 230 #else /* __i386__ */
 231 struct exec_archhandler exec_archhandler_ppc;
 232 #endif /* __i386__ */
 233
 234 char rootdevice[16];    /* hfs device names have at least 9 chars */
 235
 236 #if  KMEMSTATS
 237 struct  kmemstats kmemstats[M_LAST];
 238 #endif
 239
 240 int     lbolt;                          /* awoken once a second */
 241 struct  vnode *rootvp;
 242 int boothowto = RB_DEBUG;
 243
 244 void lightning_bolt(void *);
 245 extern kern_return_t IOFindBSDRoot(char *, unsigned int, dev_t *, u_int32_t *);
 246 extern void IOSecureBSDRoot(const char * rootName);
 247 extern kern_return_t IOKitBSDInit(void );
 248 extern void kminit(void);
 249 extern void klogwakeup(void);
 250 extern void file_lock_init(void);
 251 extern void kmeminit(void);
 252 extern void bsd_bufferinit(void);
 253
 254 extern int serverperfmode;
 255 extern int ncl;
 256
 257 vm_map_t        bsd_pageable_map;
 258 vm_map_t        mb_map;
 259
 260 static  int bsd_simul_execs;
 261 static int bsd_pageable_map_size;
 262 __private_extern__ int execargs_cache_size = 0;
 263 __private_extern__ int execargs_free_count = 0;
 264 __private_extern__ vm_offset_t * execargs_cache = NULL;
 265
 266 void bsd_exec_setup(int);
 267
 268 /*
 269  * Set to disable grading 64 bit Mach-o binaries as executable, for testing;
 270  * Intel only.
 271  */
 272 __private_extern__ int bootarg_no64exec = 0;
 273 __private_extern__ int bootarg_vnode_cache_defeat = 0;
 274
 275 /*
 276  * Prevent kernel-based ASLR from being used, for testing.
 277  */
 278 #if DEVELOPMENT || DEBUG
 279 __private_extern__ int bootarg_disable_aslr = 0;
 280 #endif
 281
 282 int     cmask = CMASK;
 283 extern int customnbuf;
 284
 285 void bsd_init(void) __attribute__((section("__TEXT, initcode")));
 286 kern_return_t bsd_autoconf(void) __attribute__((section("__TEXT, initcode")));
 287 void bsd_utaskbootstrap(void) __attribute__((section("__TEXT, initcode")));
 288
 289 static void parse_bsd_args(void);
 290 extern task_t bsd_init_task;
 291 extern char    init_task_failure_data[];
 292 extern void time_zone_slock_init(void);
 293 extern void select_wait_queue_init(void);
 294 static void process_name(const char *, proc_t);
 295
 296 static void setconf(void);
 297
 298 funnel_t *kernel_flock;
 299
 300 #if SYSV_SHM
 301 extern void sysv_shm_lock_init(void);
 302 #endif
 303 #if SYSV_SEM
 304 extern void sysv_sem_lock_init(void);
 305 #endif
 306 #if SYSV_MSG
 307 extern void sysv_msg_lock_init(void);
 308 #endif
 309
 310 #if !defined(SECURE_KERNEL)
 311 /* kmem access not enabled by default; can be changed with boot-args */
 312 /* We don't need to keep this symbol around in RELEASE kernel */
 313 int setup_kmem = 0;
 314 #endif
 315
 316 #if CONFIG_MACF
 317 #if defined (__i386__) || defined (__x86_64__)
 318 /* MACF policy_check configuration flags; see policy_check.c for details */
 319 int policy_check_flags = 0;
 320
 321 extern int check_policy_init(int);
 322 #endif
 323 #endif  /* CONFIG_MACF */
 324
 325 extern void stackshot_lock_init(void);
 326
 327
 328 /* If we are using CONFIG_DTRACE */
 329 #if CONFIG_DTRACE
 330         extern void dtrace_postinit(void);
 331 #endif
 332
 333
 334 /*
 335  * Initialization code.
 336  * Called from cold start routine as
 337  * soon as a stack and segmentation
 338  * have been established.
 339  * Functions:
 340  *      turn on clock
 341  *      hand craft 0th process
 342  *      call all initialization routines
 343  *  hand craft 1st user process
 344  */
 345
 346 /*
 347  *      Sets the name for the given task.
 348  */
 349 static void
 350 process_name(const char *s, proc_t p)
 351 {
 352         size_t length = strlen(s);
 353
 354         bcopy(s, p->p_comm,
 355                 length >= sizeof(p->p_comm) ? sizeof(p->p_comm) :
 356                         length + 1);
 357 }
 358
 359 /* To allow these values to be patched, they're globals here */
 360 #include <machine/vmparam.h>
 361 struct rlimit vm_initial_limit_stack = { DFLSSIZ, MAXSSIZ - PAGE_SIZE };
 362 struct rlimit vm_initial_limit_data = { DFLDSIZ, MAXDSIZ };
 363 struct rlimit vm_initial_limit_core = { DFLCSIZ, MAXCSIZ };
 364
 365 extern thread_t cloneproc(task_t, proc_t, int);
 366 extern int      (*mountroot)(void);
 367
 368 lck_grp_t * proc_lck_grp;
 369 lck_grp_t * proc_slock_grp;
 370 lck_grp_t * proc_fdmlock_grp;
 371 lck_grp_t * proc_mlock_grp;
 372 lck_grp_attr_t * proc_lck_grp_attr;
 373 lck_attr_t * proc_lck_attr;
 374 lck_mtx_t * proc_list_mlock;
 375 lck_mtx_t * proc_klist_mlock;
 376
 377 extern lck_mtx_t * execargs_cache_lock;
 378
 379 /* hook called after root is mounted XXX temporary hack */
 380 void (*mountroot_post_hook)(void);
 381 void (*unmountroot_pre_hook)(void);
 382
 383 /*
 384  * This function is called very early on in the Mach startup, from the
 385  * function start_kernel_threads() in osfmk/kern/startup.c.  It's called
 386  * in the context of the current (startup) task using a call to the
 387  * function kernel_thread_create() to jump into start_kernel_threads().
 388  * Internally, kernel_thread_create() calls thread_create_internal(),
 389  * which calls uthread_alloc().  The function of uthread_alloc() is
 390  * normally to allocate a uthread structure, and fill out the uu_sigmask,
 391  * uu_context fields.  It skips filling these out in the case of the "task"
 392  * being "kernel_task", because the order of operation is inverted.  To
 393  * account for that, we need to manually fill in at least the contents
 394  * of the uu_context.vc_ucred field so that the uthread structure can be
 395  * used like any other.
 396  */
 397 void
 398 bsd_init(void)
 399 {
 400         struct uthread *ut;
 401         unsigned int i;
 402 #if __i386__ || __x86_64__
 403         int error;
 404 #endif
 405         struct vfs_context context;
 406         kern_return_t   ret;
 407         struct ucred temp_cred;
 408         struct posix_cred temp_pcred;
 409 #if NFSCLIENT || CONFIG_IMAGEBOOT
 410         boolean_t       netboot = FALSE;
 411 #endif
 412
 413 #define bsd_init_kprintf(x...) /* kprintf("bsd_init: " x) */
 414
 415         kernel_flock = funnel_alloc(KERNEL_FUNNEL);
 416         if (kernel_flock == (funnel_t *)0 ) {
 417                 panic("bsd_init: Failed to allocate kernel funnel");
 418         }
 419
 420         printf(copyright);
 421
 422         bsd_init_kprintf("calling kmeminit\n");
 423         kmeminit();
 424
 425         bsd_init_kprintf("calling parse_bsd_args\n");
 426         parse_bsd_args();
 427
 428         /* Initialize kauth subsystem before instancing the first credential */
 429         bsd_init_kprintf("calling kauth_init\n");
 430         kauth_init();
 431
 432         /* Initialize process and pgrp structures. */
 433         bsd_init_kprintf("calling procinit\n");
 434         procinit();
 435
 436         /* Initialize the ttys (MUST be before kminit()/bsd_autoconf()!)*/
 437         tty_init();
 438
 439         kernproc = &proc0;      /* implicitly bzero'ed */
 440
 441         /* kernel_task->proc = kernproc; */
 442         set_bsdtask_info(kernel_task,(void *)kernproc);
 443
 444         /* give kernproc a name */
 445         bsd_init_kprintf("calling process_name\n");
 446         process_name("kernel_task", kernproc);
 447
 448         /* allocate proc lock group attribute and group */
 449         bsd_init_kprintf("calling lck_grp_attr_alloc_init\n");
 450         proc_lck_grp_attr= lck_grp_attr_alloc_init();
 451
 452         proc_lck_grp = lck_grp_alloc_init("proc",  proc_lck_grp_attr);
 453 #if CONFIG_FINE_LOCK_GROUPS
 454         proc_slock_grp = lck_grp_alloc_init("proc-slock",  proc_lck_grp_attr);
 455         proc_fdmlock_grp = lck_grp_alloc_init("proc-fdmlock",  proc_lck_grp_attr);
 456         proc_mlock_grp = lck_grp_alloc_init("proc-mlock",  proc_lck_grp_attr);
 457 #endif
 458         /* Allocate proc lock attribute */
 459         proc_lck_attr = lck_attr_alloc_init();
 460 #if 0
 461 #if __PROC_INTERNAL_DEBUG
 462         lck_attr_setdebug(proc_lck_attr);
 463 #endif
 464 #endif
 465
 466 #if CONFIG_FINE_LOCK_GROUPS
 467         proc_list_mlock = lck_mtx_alloc_init(proc_mlock_grp, proc_lck_attr);
 468         proc_klist_mlock = lck_mtx_alloc_init(proc_mlock_grp, proc_lck_attr);
 469         lck_mtx_init(&kernproc->p_mlock, proc_mlock_grp, proc_lck_attr);
 470         lck_mtx_init(&kernproc->p_fdmlock, proc_fdmlock_grp, proc_lck_attr);
 471         lck_spin_init(&kernproc->p_slock, proc_slock_grp, proc_lck_attr);
 472 #else
 473         proc_list_mlock = lck_mtx_alloc_init(proc_lck_grp, proc_lck_attr);
 474         proc_klist_mlock = lck_mtx_alloc_init(proc_lck_grp, proc_lck_attr);
 475         lck_mtx_init(&kernproc->p_mlock, proc_lck_grp, proc_lck_attr);
 476         lck_mtx_init(&kernproc->p_fdmlock, proc_lck_grp, proc_lck_attr);
 477         lck_spin_init(&kernproc->p_slock, proc_lck_grp, proc_lck_attr);
 478 #endif
 479
 480         assert(bsd_simul_execs != 0);
 481         execargs_cache_lock = lck_mtx_alloc_init(proc_lck_grp, proc_lck_attr);
 482         execargs_cache_size = bsd_simul_execs;
 483         execargs_free_count = bsd_simul_execs;
 484         execargs_cache = (vm_offset_t *)kalloc(bsd_simul_execs * sizeof(vm_offset_t));
 485         bzero(execargs_cache, bsd_simul_execs * sizeof(vm_offset_t));
 486
 487         if (current_task() != kernel_task)
 488                 printf("bsd_init: We have a problem, "
 489                                 "current task is not kernel task\n");
 490
 491         bsd_init_kprintf("calling get_bsdthread_info\n");
 492         ut = (uthread_t)get_bsdthread_info(current_thread());
 493
 494 #if CONFIG_MACF
 495         /*
 496          * Initialize the MAC Framework
 497          */
 498         mac_policy_initbsd();
 499         kernproc->p_mac_enforce = 0;
 500
 501 #if defined (__i386__) || defined (__x86_64__)
 502         /*
 503          * We currently only support this on i386/x86_64, as that is the
 504          * only lock code we have instrumented so far.
 505          */
 506         check_policy_init(policy_check_flags);
 507 #endif
 508 #endif /* MAC */
 509
 510         /*
 511          * Create process 0.
 512          */
 513         proc_list_lock();
 514         LIST_INSERT_HEAD(&allproc, kernproc, p_list);
 515         kernproc->p_pgrp = &pgrp0;
 516         LIST_INSERT_HEAD(PGRPHASH(0), &pgrp0, pg_hash);
 517         LIST_INIT(&pgrp0.pg_members);
 518 #ifdef CONFIG_FINE_LOCK_GROUPS
 519         lck_mtx_init(&pgrp0.pg_mlock, proc_mlock_grp, proc_lck_attr);
 520 #else
 521         lck_mtx_init(&pgrp0.pg_mlock, proc_lck_grp, proc_lck_attr);
 522 #endif
 523         /* There is no other bsd thread this point and is safe without pgrp lock */
 524         LIST_INSERT_HEAD(&pgrp0.pg_members, kernproc, p_pglist);
 525         kernproc->p_listflag |= P_LIST_INPGRP;
 526         kernproc->p_pgrpid = 0;
 527         kernproc->p_uniqueid = 0;
 528
 529         pgrp0.pg_session = &session0;
 530         pgrp0.pg_membercnt = 1;
 531
 532         session0.s_count = 1;
 533         session0.s_leader = kernproc;
 534         session0.s_listflags = 0;
 535 #ifdef CONFIG_FINE_LOCK_GROUPS
 536         lck_mtx_init(&session0.s_mlock, proc_mlock_grp, proc_lck_attr);
 537 #else
 538         lck_mtx_init(&session0.s_mlock, proc_lck_grp, proc_lck_attr);
 539 #endif
 540         LIST_INSERT_HEAD(SESSHASH(0), &session0, s_hash);
 541         proc_list_unlock();
 542
 543 #if CONFIG_LCTX
 544         kernproc->p_lctx = NULL;
 545 #endif
 546
 547         kernproc->task = kernel_task;
 548
 549         kernproc->p_stat = SRUN;
 550         kernproc->p_flag = P_SYSTEM;
 551         kernproc->p_lflag = 0;
 552         kernproc->p_ladvflag = 0;
 553
 554 #if DEVELOPMENT || DEBUG
 555         if (bootarg_disable_aslr)
 556                 kernproc->p_flag |= P_DISABLE_ASLR;
 557 #endif
 558
 559         kernproc->p_nice = NZERO;
 560         kernproc->p_pptr = kernproc;
 561
 562         TAILQ_INIT(&kernproc->p_uthlist);
 563         TAILQ_INSERT_TAIL(&kernproc->p_uthlist, ut, uu_list);
 564
 565         kernproc->sigwait = FALSE;
 566         kernproc->sigwait_thread = THREAD_NULL;
 567         kernproc->exit_thread = THREAD_NULL;
 568         kernproc->p_csflags = CS_VALID;
 569
 570         /*
 571          * Create credential.  This also Initializes the audit information.
 572          */
 573         bsd_init_kprintf("calling bzero\n");
 574         bzero(&temp_cred, sizeof(temp_cred));
 575         bzero(&temp_pcred, sizeof(temp_pcred));
 576         temp_pcred.cr_ngroups = 1;
 577
 578         temp_cred.cr_audit.as_aia_p = audit_default_aia_p;
 579
 580         bsd_init_kprintf("calling kauth_cred_create\n");
 581         /*
 582          * We have to label the temp cred before we create from it to
 583          * properly set cr_ngroups, or the create will fail.
 584          */
 585         posix_cred_label(&temp_cred, &temp_pcred);
 586         kernproc->p_ucred = kauth_cred_create(&temp_cred);
 587
 588         /* update cred on proc */
 589         PROC_UPDATE_CREDS_ONPROC(kernproc);
 590
 591         /* give the (already exisiting) initial thread a reference on it */
 592         bsd_init_kprintf("calling kauth_cred_ref\n");
 593         kauth_cred_ref(kernproc->p_ucred);
 594         ut->uu_context.vc_ucred = kernproc->p_ucred;
 595         ut->uu_context.vc_thread = current_thread();
 596
 597         TAILQ_INIT(&kernproc->p_aio_activeq);
 598         TAILQ_INIT(&kernproc->p_aio_doneq);
 599         kernproc->p_aio_total_count = 0;
 600         kernproc->p_aio_active_count = 0;
 601
 602         bsd_init_kprintf("calling file_lock_init\n");
 603         file_lock_init();
 604
 605 #if CONFIG_MACF
 606         mac_cred_label_associate_kernel(kernproc->p_ucred);
 607         mac_task_label_update_cred (kernproc->p_ucred, (struct task *) kernproc->task);
 608 #endif
 609
 610         /* Create the file descriptor table. */
 611         filedesc0.fd_refcnt = 1+1;      /* +1 so shutdown will not _FREE_ZONE */
 612         kernproc->p_fd = &filedesc0;
 613         filedesc0.fd_cmask = cmask;
 614         filedesc0.fd_knlistsize = -1;
 615         filedesc0.fd_knlist = NULL;
 616         filedesc0.fd_knhash = NULL;
 617         filedesc0.fd_knhashmask = 0;
 618
 619         /* Create the limits structures. */
 620         kernproc->p_limit = &limit0;
 621         for (i = 0; i < sizeof(kernproc->p_rlimit)/sizeof(kernproc->p_rlimit[0]); i++)
 622                 limit0.pl_rlimit[i].rlim_cur =
 623                         limit0.pl_rlimit[i].rlim_max = RLIM_INFINITY;
 624         limit0.pl_rlimit[RLIMIT_NOFILE].rlim_cur = NOFILE;
 625         limit0.pl_rlimit[RLIMIT_NPROC].rlim_cur = maxprocperuid;
 626         limit0.pl_rlimit[RLIMIT_NPROC].rlim_max = maxproc;
 627         limit0.pl_rlimit[RLIMIT_STACK] = vm_initial_limit_stack;
 628         limit0.pl_rlimit[RLIMIT_DATA] = vm_initial_limit_data;
 629         limit0.pl_rlimit[RLIMIT_CORE] = vm_initial_limit_core;
 630         limit0.pl_refcnt = 1;
 631
 632         kernproc->p_stats = &pstats0;
 633         kernproc->p_sigacts = &sigacts0;
 634
 635         /*
 636          * Charge root for two  processes: init and mach_init.
 637          */
 638         bsd_init_kprintf("calling chgproccnt\n");
 639         (void)chgproccnt(0, 1);
 640
 641         /*
 642          *      Allocate a kernel submap for pageable memory
 643          *      for temporary copying (execve()).
 644          */
 645         {
 646                 vm_offset_t     minimum;
 647
 648                 bsd_init_kprintf("calling kmem_suballoc\n");
 649                 assert(bsd_pageable_map_size != 0);
 650                 ret = kmem_suballoc(kernel_map,
 651                                 &minimum,
 652                                 (vm_size_t)bsd_pageable_map_size,
 653                                 TRUE,
 654                                 VM_FLAGS_ANYWHERE,
 655                                 &bsd_pageable_map);
 656                 if (ret != KERN_SUCCESS)
 657                         panic("bsd_init: Failed to allocate bsd pageable map");
 658         }
 659
 660         /*
 661          * Initialize buffers and hash links for buffers
 662          *
 663          * SIDE EFFECT: Starts a thread for bcleanbuf_thread(), so must
 664          *              happen after a credential has been associated with
 665          *              the kernel task.
 666          */
 667         bsd_init_kprintf("calling bsd_bufferinit\n");
 668         bsd_bufferinit();
 669
 670         /* Initialize the execve() semaphore */
 671         bsd_init_kprintf("calling semaphore_create\n");
 672
 673         if (ret != KERN_SUCCESS)
 674                 panic("bsd_init: Failed to create execve semaphore");
 675
 676         /*
 677          * Initialize the calendar.
 678          */
 679         bsd_init_kprintf("calling IOKitInitializeTime\n");
 680         IOKitInitializeTime();
 681
 682         bsd_init_kprintf("calling ubc_init\n");
 683         ubc_init();
 684
 685         /*
 686          * Initialize device-switches.
 687          */
 688         bsd_init_kprintf("calling devsw_init() \n");
 689         devsw_init();
 690
 691         /* Initialize the file systems. */
 692         bsd_init_kprintf("calling vfsinit\n");
 693         vfsinit();
 694
 695 #if SOCKETS
 696         /* Initialize per-CPU cache allocator */
 697         mcache_init();
 698
 699         /* Initialize mbuf's. */
 700         bsd_init_kprintf("calling mbinit\n");
 701         mbinit();
 702         net_str_id_init(); /* for mbuf tags */
 703 #endif /* SOCKETS */
 704
 705         /*
 706          * Initializes security event auditing.
 707          * XXX: Should/could this occur later?
 708          */
 709 #if CONFIG_AUDIT
 710         bsd_init_kprintf("calling audit_init\n");
 711         audit_init();
 712 #endif
 713
 714         /* Initialize kqueues */
 715         bsd_init_kprintf("calling knote_init\n");
 716         knote_init();
 717
 718         /* Initialize for async IO */
 719         bsd_init_kprintf("calling aio_init\n");
 720         aio_init();
 721
 722         /* Initialize pipes */
 723         bsd_init_kprintf("calling pipeinit\n");
 724         pipeinit();
 725
 726         /* Initialize SysV shm subsystem locks; the subsystem proper is
 727          * initialized through a sysctl.
 728          */
 729 #if SYSV_SHM
 730         bsd_init_kprintf("calling sysv_shm_lock_init\n");
 731         sysv_shm_lock_init();
 732 #endif
 733 #if SYSV_SEM
 734         bsd_init_kprintf("calling sysv_sem_lock_init\n");
 735         sysv_sem_lock_init();
 736 #endif
 737 #if SYSV_MSG
 738         bsd_init_kprintf("sysv_msg_lock_init\n");
 739         sysv_msg_lock_init();
 740 #endif
 741         bsd_init_kprintf("calling pshm_lock_init\n");
 742         pshm_lock_init();
 743         bsd_init_kprintf("calling psem_lock_init\n");
 744         psem_lock_init();
 745
 746         pthread_init();
 747         /* POSIX Shm and Sem */
 748         bsd_init_kprintf("calling pshm_cache_init\n");
 749         pshm_cache_init();
 750         bsd_init_kprintf("calling psem_cache_init\n");
 751         psem_cache_init();
 752         bsd_init_kprintf("calling time_zone_slock_init\n");
 753         time_zone_slock_init();
 754         bsd_init_kprintf("calling select_wait_queue_init\n");
 755         select_wait_queue_init();
 756
 757         /* Stack snapshot facility lock */
 758         stackshot_lock_init();
 759         /*
 760          * Initialize protocols.  Block reception of incoming packets
 761          * until everything is ready.
 762          */
 763         bsd_init_kprintf("calling sysctl_register_fixed\n");
 764         sysctl_register_fixed();
 765         bsd_init_kprintf("calling sysctl_mib_init\n");
 766         sysctl_mib_init();
 767 #if NETWORKING
 768         bsd_init_kprintf("calling dlil_init\n");
 769         dlil_init();
 770         bsd_init_kprintf("calling proto_kpi_init\n");
 771         proto_kpi_init();
 772 #endif /* NETWORKING */
 773 #if SOCKETS
 774         bsd_init_kprintf("calling socketinit\n");
 775         socketinit();
 776         bsd_init_kprintf("calling domaininit\n");
 777         domaininit();
 778 #endif /* SOCKETS */
 779
 780         kernproc->p_fd->fd_cdir = NULL;
 781         kernproc->p_fd->fd_rdir = NULL;
 782
 783 #if CONFIG_FREEZE
 784         /* Initialise background hibernation */
 785         bsd_init_kprintf("calling kern_hibernation_init\n");
 786         kern_hibernation_init();
 787 #endif
 788
 789 #if CONFIG_EMBEDDED
 790         /* Initialize kernel memory status notifications */
 791         bsd_init_kprintf("calling kern_memorystatus_init\n");
 792         kern_memorystatus_init();
 793 #endif
 794
 795 #ifdef GPROF
 796         /* Initialize kernel profiling. */
 797         kmstartup();
 798 #endif
 799
 800         /* kick off timeout driven events by calling first time */
 801         thread_wakeup(&lbolt);
 802         timeout(lightning_bolt, 0, hz);
 803
 804         bsd_init_kprintf("calling bsd_autoconf\n");
 805         bsd_autoconf();
 806
 807 #if CONFIG_DTRACE
 808         dtrace_postinit();
 809 #endif
 810
 811         /*
 812          * We attach the loopback interface *way* down here to ensure
 813          * it happens after autoconf(), otherwise it becomes the
 814          * "primary" interface.
 815          */
 816 #include <loop.h>
 817 #if NLOOP > 0
 818         bsd_init_kprintf("calling loopattach\n");
 819         loopattach();                   /* XXX */
 820 #endif
 821
 822 #if PFLOG
 823         /* Initialize packet filter log interface */
 824         pfloginit();
 825 #endif /* PFLOG */
 826
 827 #if NETHER > 0
 828         /* Register the built-in dlil ethernet interface family */
 829         bsd_init_kprintf("calling ether_family_init\n");
 830         ether_family_init();
 831 #endif /* ETHER */
 832
 833 #if NETWORKING
 834         /* Call any kext code that wants to run just after network init */
 835         bsd_init_kprintf("calling net_init_run\n");
 836         net_init_run();
 837
 838         /* register user tunnel kernel control handler */
 839         utun_register_control();
 840     netsrc_init();
 841
 842         /* wait for network domain to finish */
 843         domainfin();
 844 #endif /* NETWORKING */
 845
 846         bsd_init_kprintf("calling vnode_pager_bootstrap\n");
 847         vnode_pager_bootstrap();
 848 #if 0
 849         /* XXX Hack for early debug stop */
 850         printf("\nabout to sleep for 10 seconds\n");
 851         IOSleep( 10 * 1000 );
 852         /* Debugger("hello"); */
 853 #endif
 854
 855         bsd_init_kprintf("calling inittodr\n");
 856         inittodr(0);
 857
 858         /* Mount the root file system. */
 859         while( TRUE) {
 860                 int err;
 861
 862                 bsd_init_kprintf("calling setconf\n");
 863                 setconf();
 864 #if NFSCLIENT
 865                 netboot = (mountroot == netboot_mountroot);
 866 #endif
 867
 868                 bsd_init_kprintf("vfs_mountroot\n");
 869                 if (0 == (err = vfs_mountroot()))
 870                         break;
 871                 rootdevice[0] = '\0';
 872 #if NFSCLIENT
 873                 if (netboot) {
 874                         PE_display_icon( 0, "noroot");  /* XXX a netboot-specific icon would be nicer */
 875                         vc_progress_set(FALSE, 0);
 876                         for (i=1; 1; i*=2) {
 877                                 printf("bsd_init: failed to mount network root, error %d, %s\n",
 878                                         err, PE_boot_args());
 879                                 printf("We are hanging here...\n");
 880                                 IOSleep(i*60*1000);
 881                         }
 882                         /*NOTREACHED*/
 883                 }
 884 #endif
 885                 printf("cannot mount root, errno = %d\n", err);
 886                 boothowto |= RB_ASKNAME;
 887         }
 888
 889         IOSecureBSDRoot(rootdevice);
 890
 891         context.vc_thread = current_thread();
 892         context.vc_ucred = kernproc->p_ucred;
 893         mountlist.tqh_first->mnt_flag |= MNT_ROOTFS;
 894
 895         bsd_init_kprintf("calling VFS_ROOT\n");
 896         /* Get the vnode for '/'.  Set fdp->fd_fd.fd_cdir to reference it. */
 897         if (VFS_ROOT(mountlist.tqh_first, &rootvnode, &context))
 898                 panic("bsd_init: cannot find root vnode: %s", PE_boot_args());
 899         rootvnode->v_flag |= VROOT;
 900         (void)vnode_ref(rootvnode);
 901         (void)vnode_put(rootvnode);
 902         filedesc0.fd_cdir = rootvnode;
 903
 904 #if NFSCLIENT
 905         if (netboot) {
 906                 int err;
 907
 908                 netboot = TRUE;
 909                 /* post mount setup */
 910                 if ((err = netboot_setup()) != 0) {
 911                         PE_display_icon( 0, "noroot");  /* XXX a netboot-specific icon would be nicer */
 912                         vc_progress_set(FALSE, 0);
 913                         for (i=1; 1; i*=2) {
 914                                 printf("bsd_init: NetBoot could not find root, error %d: %s\n",
 915                                         err, PE_boot_args());
 916                                 printf("We are hanging here...\n");
 917                                 IOSleep(i*60*1000);
 918                         }
 919                         /*NOTREACHED*/
 920                 }
 921         }
 922 #endif
 923
 924
 925 #if CONFIG_IMAGEBOOT
 926         /*
 927          * See if a system disk image is present. If so, mount it and
 928          * switch the root vnode to point to it
 929          */
 930         if (netboot == FALSE && imageboot_needed()) {
 931                 /*
 932                  * An image was found.  No turning back: we're booted
 933                  * with a kernel from the disk image.
 934                  */
 935                 imageboot_setup();
 936         }
 937 #endif /* CONFIG_IMAGEBOOT */
 938
 939         /* set initial time; all other resource data is  already zero'ed */
 940         microtime(&kernproc->p_start);
 941         kernproc->p_stats->p_start = kernproc->p_start; /* for compat */
 942
 943 #if DEVFS
 944         {
 945             char mounthere[] = "/dev";  /* !const because of internal casting */
 946
 947             bsd_init_kprintf("calling devfs_kernel_mount\n");
 948             devfs_kernel_mount(mounthere);
 949         }
 950 #endif /* DEVFS */
 951
 952         /* Initialize signal state for process 0. */
 953         bsd_init_kprintf("calling siginit\n");
 954         siginit(kernproc);
 955
 956         bsd_init_kprintf("calling bsd_utaskbootstrap\n");
 957         bsd_utaskbootstrap();
 958
 959 #if defined(__LP64__)
 960         kernproc->p_flag |= P_LP64;
 961         printf("Kernel is LP64\n");
 962 #endif
 963
 964         pal_kernel_announce();
 965
 966 #if __i386__ || __x86_64__
 967         /* this should be done after the root filesystem is mounted */
 968         error = set_archhandler(kernproc, CPU_TYPE_POWERPC);
 969         if (error) /* XXX make more generic */
 970                 exec_archhandler_ppc.path[0] = 0;
 971 #endif
 972
 973         bsd_init_kprintf("calling mountroot_post_hook\n");
 974
 975         /* invoke post-root-mount hook */
 976         if (mountroot_post_hook != NULL)
 977                 mountroot_post_hook();
 978
 979 #if 0 /* not yet */
 980         consider_zone_gc(FALSE);
 981 #endif
 982
 983         bsd_init_kprintf("done\n");
 984 }
 985
 986 /* Called with kernel funnel held */
 987 void
 988 bsdinit_task(void)
 989 {
 990         proc_t p = current_proc();
 991         struct uthread *ut;
 992         thread_t thread;
 993
 994         process_name("init", p);
 995
 996         ux_handler_init();
 997
 998         thread = current_thread();
 999         (void) host_set_exception_ports(host_priv_self(),
1000                                         EXC_MASK_ALL & ~(EXC_MASK_RPC_ALERT),//pilotfish (shark) needs this port
1001                                         (mach_port_t) ux_exception_port,
1002                                         EXCEPTION_DEFAULT| MACH_EXCEPTION_CODES,
1003                                         0);
1004
1005         ut = (uthread_t)get_bsdthread_info(thread);
1006
1007         bsd_init_task = get_threadtask(thread);
1008         init_task_failure_data[0] = 0;
1009
1010 #if CONFIG_MACF
1011         mac_cred_label_associate_user(p->p_ucred);
1012         mac_task_label_update_cred (p->p_ucred, (struct task *) p->task);
1013 #endif
1014         load_init_program(p);
1015         lock_trace = 1;
1016 }
1017
1018 void
1019 lightning_bolt(__unused void *dummy)
1020 {
1021         boolean_t       funnel_state;
1022
1023         funnel_state = thread_funnel_set(kernel_flock, TRUE);
1024
1025         thread_wakeup(&lbolt);
1026         timeout(lightning_bolt,0,hz);
1027         klogwakeup();
1028
1029         (void) thread_funnel_set(kernel_flock, FALSE);
1030 }
1031
1032 kern_return_t
1033 bsd_autoconf(void)
1034 {
1035         kprintf("bsd_autoconf: calling kminit\n");
1036         kminit();
1037
1038         /*
1039          * Early startup for bsd pseudodevices.
1040          */
1041         {
1042             struct pseudo_init *pi;
1043
1044             for (pi = pseudo_inits; pi->ps_func; pi++)
1045                 (*pi->ps_func) (pi->ps_count);
1046         }
1047
1048         return( IOKitBSDInit());
1049 }
1050
1051
1052 #include <sys/disklabel.h>  /* for MAXPARTITIONS */
1053
1054 static void
1055 setconf(void)
1056 {
1057         u_int32_t       flags;
1058         kern_return_t   err;
1059
1060         /*
1061          * calls into IOKit can generate networking registrations
1062          * which needs to be under network funnel. Right thing to do
1063          * here is to drop the funnel alltogether and regrab it afterwards
1064          */
1065         err = IOFindBSDRoot(rootdevice, sizeof(rootdevice), &rootdev, &flags);
1066         if( err) {
1067                 printf("setconf: IOFindBSDRoot returned an error (%d);"
1068                         "setting rootdevice to 'sd0a'.\n", err); /* XXX DEBUG TEMP */
1069                 rootdev = makedev( 6, 0 );
1070                 strlcpy(rootdevice, "sd0a", sizeof(rootdevice));
1071                 flags = 0;
1072         }
1073
1074 #if NFSCLIENT
1075         if( flags & 1 ) {
1076                 /* network device */
1077                 mountroot = netboot_mountroot;
1078         } else {
1079 #endif
1080                 /* otherwise have vfs determine root filesystem */
1081                 mountroot = NULL;
1082 #if NFSCLIENT
1083         }
1084 #endif
1085
1086 }
1087
1088 void
1089 bsd_utaskbootstrap(void)
1090 {
1091         thread_t thread;
1092         struct uthread *ut;
1093
1094         /*
1095          * Clone the bootstrap process from the kernel process, without
1096          * inheriting either task characteristics or memory from the kernel;
1097          */
1098         thread = cloneproc(TASK_NULL, kernproc, FALSE);
1099
1100         /* Hold the reference as it will be dropped during shutdown */
1101         initproc = proc_find(1);
1102 #if __PROC_INTERNAL_DEBUG
1103         if (initproc == PROC_NULL)
1104                 panic("bsd_utaskbootstrap: initproc not set\n");
1105 #endif
1106         /*
1107          * Since we aren't going back out the normal way to our parent,
1108          * we have to drop the transition locks explicitly.
1109          */
1110         proc_signalend(initproc, 0);
1111         proc_transend(initproc, 0);
1112
1113         ut = (struct uthread *)get_bsdthread_info(thread);
1114         ut->uu_sigmask = 0;
1115         act_set_astbsd(thread);
1116         (void) thread_resume(thread);
1117 }
1118
1119 static void
1120 parse_bsd_args(void)
1121 {
1122         char namep[16];
1123         int msgbuf;
1124
1125         if (PE_parse_boot_argn("-s", namep, sizeof (namep)))
1126                 boothowto |= RB_SINGLE;
1127
1128         if (PE_parse_boot_argn("-b", namep, sizeof (namep)))
1129                 boothowto |= RB_NOBOOTRC;
1130
1131         if (PE_parse_boot_argn("-x", namep, sizeof (namep))) /* safe boot */
1132                 boothowto |= RB_SAFEBOOT;
1133
1134         /* disable 64 bit grading */
1135         if (PE_parse_boot_argn("-no64exec", namep, sizeof (namep)))
1136                 bootarg_no64exec = 1;
1137
1138         /* disable vnode_cache_is_authorized() by setting vnode_cache_defeat */
1139         if (PE_parse_boot_argn("-vnode_cache_defeat", namep, sizeof (namep)))
1140                 bootarg_vnode_cache_defeat = 1;
1141
1142 #if DEVELOPMENT || DEBUG
1143         if (PE_parse_boot_argn("-disable_aslr", namep, sizeof (namep)))
1144                 bootarg_disable_aslr = 1;
1145 #endif
1146
1147         PE_parse_boot_argn("ncl", &ncl, sizeof (ncl));
1148         if (PE_parse_boot_argn("nbuf", &max_nbuf_headers,
1149                                 sizeof (max_nbuf_headers))) {
1150                 customnbuf = 1;
1151         }
1152 #if !defined(SECURE_KERNEL)
1153         PE_parse_boot_argn("kmem", &setup_kmem, sizeof (setup_kmem));
1154 #endif
1155
1156 #if CONFIG_MACF
1157 #if defined (__i386__) || defined (__x86_64__)
1158         PE_parse_boot_argn("policy_check", &policy_check_flags, sizeof (policy_check_flags));
1159 #endif
1160 #endif  /* CONFIG_MACF */
1161
1162         if (PE_parse_boot_argn("msgbuf", &msgbuf, sizeof (msgbuf))) {
1163                 log_setsize(msgbuf);
1164         }
1165
1166         if (PE_parse_boot_argn("-novfscache", namep, sizeof(namep))) {
1167                 nc_disabled = 1;
1168         }
1169 }
1170
1171 void
1172 bsd_exec_setup(int scale)
1173 {
1174
1175         switch (scale) {
1176                 case 0:
1177                 case 1:
1178                         bsd_simul_execs = BSD_SIMUL_EXECS;
1179                         break;
1180                 case 2:
1181                 case 3:
1182                         bsd_simul_execs = 65;
1183                         break;
1184                 case 4:
1185                 case 5:
1186                         bsd_simul_execs = 129;
1187                         break;
1188                 case 6:
1189                 case 7:
1190                         bsd_simul_execs = 257;
1191                         break;
1192                 default:
1193                         bsd_simul_execs = 513;
1194                         break;
1195
1196         }
1197         bsd_pageable_map_size = (bsd_simul_execs * BSD_PAGEABLE_SIZE_PER_EXEC);
1198 }
1199
1200 #if !NFSCLIENT
1201 int
1202 netboot_root(void);
1203
1204 int
1205 netboot_root(void)
1206 {
1207         return(0);
1208 }
1209 #endif