X-Git-Url: https://git.saurik.com/apple/xnu.git/blobdiff_plain/c0fea4742e91338fffdcf79f86a7c1d5e2b97eb1..060df5ea7c632b1ac8cc8aac1fb59758165c2084:/bsd/kern/bsd_init.c diff --git a/bsd/kern/bsd_init.c b/bsd/kern/bsd_init.c index 05b66d530..e2868a40c 100644 --- a/bsd/kern/bsd_init.c +++ b/bsd/kern/bsd_init.c @@ -1,25 +1,31 @@ /* - * Copyright (c) 2000-2004 Apple Computer, Inc. All rights reserved. + * Copyright (c) 2000-2009 Apple Inc. All rights reserved. * - * @APPLE_LICENSE_HEADER_START@ + * @APPLE_OSREFERENCE_LICENSE_HEADER_START@ * - * The contents of this file constitute Original Code as defined in and - * are subject to the Apple Public Source License Version 1.1 (the - * "License"). You may not use this file except in compliance with the - * License. Please obtain a copy of the License at - * http://www.apple.com/publicsource and read it before using this file. + * This file contains Original Code and/or Modifications of Original Code + * as defined in and that are subject to the Apple Public Source License + * Version 2.0 (the 'License'). You may not use this file except in + * compliance with the License. The rights granted to you under the License + * may not be used to create, or enable the creation or redistribution of, + * unlawful or unlicensed copies of an Apple operating system, or to + * circumvent, violate, or enable the circumvention or violation of, any + * terms of an Apple operating system software license agreement. * - * This Original Code and all software distributed under the License are - * distributed on an "AS IS" basis, WITHOUT WARRANTY OF ANY KIND, EITHER + * Please obtain a copy of the License at + * http://www.opensource.apple.com/apsl/ and read it before using this file. + * + * The Original Code and all software distributed under the License are + * distributed on an 'AS IS' basis, WITHOUT WARRANTY OF ANY KIND, EITHER * EXPRESS OR IMPLIED, AND APPLE HEREBY DISCLAIMS ALL SUCH WARRANTIES, * INCLUDING WITHOUT LIMITATION, ANY WARRANTIES OF MERCHANTABILITY, - * FITNESS FOR A PARTICULAR PURPOSE OR NON-INFRINGEMENT. Please see the - * License for the specific language governing rights and limitations - * under the License. + * FITNESS FOR A PARTICULAR PURPOSE, QUIET ENJOYMENT OR NON-INFRINGEMENT. + * Please see the License for the specific language governing rights and + * limitations under the License. * - * @APPLE_LICENSE_HEADER_END@ - */ -/* + * @APPLE_OSREFERENCE_LICENSE_HEADER_END@ + * + * * Copyright (c) 1982, 1986, 1989, 1991, 1992, 1993 * The Regents of the University of California. All rights reserved. * (c) UNIX System Laboratories, Inc. @@ -66,6 +72,12 @@ * All rights reserved. The CMU software License Agreement specifies * the terms and conditions for use and redistribution. */ +/* + * NOTICE: This file was modified by McAfee Research in 2004 to introduce + * support for mandatory and extensible security protections. This notice + * is included in support of clause 2.2 (b) of the Apple Public License, + * Version 2.0. + */ #include #include @@ -83,15 +95,17 @@ #include #include -#include +#include #include #include +#include #include #include #include #include +#include #include @@ -111,23 +125,64 @@ #include #include #include - -#include -#include +#include /* for thread_resume() */ +#include /* for task_set_exception_ports() */ +#include /* for ux_handler() */ +#include /* for ubc_init() */ +#include /* for mcache_init() */ +#include /* for mbinit() */ +#include /* for knote_init() */ +#include /* for aio_init() */ +#include /* for psem_cache_init() */ +#include /* for dlil_init() */ +#include /* for proto_kpi_init() */ +#include /* for pipeinit() */ +#include /* for socketinit() */ +#include /* for domaininit() */ +#include /* for thread_wakeup() */ +#include /* for ether_family_init() */ +#include /* for vnode_pager_bootstrap() */ +#include /* for devfs_kernel_mount() */ +#include /* for host_set_exception_ports() */ +#include /* for host_priv_self() */ +#include /* for kmem_suballoc() */ +#include /* for psem_lock_init() */ +#include /* for log_setsize() */ +#include /* for tty_init() */ +#include /* for utun_register_control() */ +#include /* for net_str_id_init() */ #include +#if CONFIG_MACF +#include +#include /* mac_init_bsd() */ +#include /* mac_update_task_label() */ +#endif + #include -extern int app_profile; /* on/off switch for pre-heat cache */ +#if CONFIG_IMAGEBOOT +#include +#endif + +#if PFLOG +#include +#endif + +#include + +void * get_user_regs(thread_t); /* XXX kludge for */ +void IOKitInitializeTime(void); /* XXX */ +void IOSleep(unsigned int); /* XXX */ +void loopattach(void); /* XXX */ +void vc_progress_set(boolean_t, uint32_t); /* XXX */ -char copyright[] = +const char copyright[] = "Copyright (c) 1982, 1986, 1989, 1991, 1993\n\t" "The Regents of the University of California. " "All rights reserved.\n\n"; -extern void ux_handler(); - /* Components of the first process -- never freed. */ struct proc proc0; struct session session0; @@ -136,7 +191,8 @@ struct filedesc filedesc0; struct plimit limit0; struct pstats pstats0; struct sigacts sigacts0; -struct proc *kernproc, *initproc; +proc_t kernproc; +proc_t initproc; long tk_cancc; long tk_nin; @@ -158,17 +214,18 @@ char hostname[MAXHOSTNAMELEN]; int hostnamelen; char domainname[MAXDOMNAMELEN]; int domainnamelen; -#if __i386__ +#if defined(__i386__) || defined(__x86_64__) struct exec_archhandler exec_archhandler_ppc = { .path = "/usr/libexec/oah/translate", }; +const char * const kRosettaStandIn_str = "/usr/libexec/oah/RosettaNonGrata"; #else /* __i386__ */ struct exec_archhandler exec_archhandler_ppc; #endif /* __i386__ */ char rootdevice[16]; /* hfs device names have at least 9 chars */ -#ifdef KMEMSTATS +#if KMEMSTATS struct kmemstats kmemstats[M_LAST]; #endif @@ -176,36 +233,82 @@ int lbolt; /* awoken once a second */ struct vnode *rootvp; int boothowto = RB_DEBUG; -#define BSD_PAGABLE_MAP_SIZE (16 * 512 * 1024) +void lightning_bolt(void *); +extern kern_return_t IOFindBSDRoot(char *, unsigned int, dev_t *, u_int32_t *); +extern void IOSecureBSDRoot(const char * rootName); +extern kern_return_t IOKitBSDInit(void ); +extern void kminit(void); +extern void klogwakeup(void); +extern void file_lock_init(void); +extern void kmeminit(void); +extern void bsd_bufferinit(void); + +extern int srv; +extern int ncl; + vm_map_t bsd_pageable_map; vm_map_t mb_map; -semaphore_t execve_semaphore; + +static int bsd_simul_execs = BSD_SIMUL_EXECS; +static int bsd_pageable_map_size = BSD_PAGABLE_MAP_SIZE; +__private_extern__ int execargs_cache_size = BSD_SIMUL_EXECS; +__private_extern__ int execargs_free_count = BSD_SIMUL_EXECS; +__private_extern__ vm_offset_t * execargs_cache = NULL; + +void bsd_exec_setup(int); + +/* + * Set to disable grading 64 bit Mach-o binaries as executable, for testing; + * Intel only. + */ +__private_extern__ int bootarg_no64exec = 0; int cmask = CMASK; extern int customnbuf; -int parse_bsd_args(void); -extern int bsd_hardclockinit; +void bsd_init(void) __attribute__((section("__TEXT, initcode"))); +kern_return_t bsd_autoconf(void) __attribute__((section("__TEXT, initcode"))); +void bsd_utaskbootstrap(void) __attribute__((section("__TEXT, initcode"))); + +static void parse_bsd_args(void); extern task_t bsd_init_task; extern char init_task_failure_data[]; extern void time_zone_slock_init(void); -static void process_name(char *, struct proc *); +static void process_name(const char *, proc_t); static void setconf(void); funnel_t *kernel_flock; +#if SYSV_SHM extern void sysv_shm_lock_init(void); +#endif +#if SYSV_SEM extern void sysv_sem_lock_init(void); +#endif +#if SYSV_MSG extern void sysv_msg_lock_init(void); -extern void pshm_lock_init(); -extern void psem_lock_init(); -extern int maxprocperuid; +#endif +extern void pthread_init(void); /* kmem access not enabled by default; can be changed with boot-args */ int setup_kmem = 0; -extern void stackshot_lock_init(); +/* size of kernel trace buffer, disabled by default */ +unsigned int new_nkdbufs = 0; + +/* mach leak logging */ +int log_leaks = 0; +int turn_on_log_leaks = 0; + +extern void stackshot_lock_init(void); + + +/* If we are using CONFIG_DTRACE */ +#if CONFIG_DTRACE + extern void dtrace_postinit(void); +#endif + /* * Initialization code. @@ -223,11 +326,9 @@ extern void stackshot_lock_init(); * Sets the name for the given task. */ static void -process_name(s, p) - char *s; - struct proc *p; +process_name(const char *s, proc_t p) { - int length = strlen(s); + size_t length = strlen(s); bcopy(s, p->p_comm, length >= sizeof(p->p_comm) ? sizeof(p->p_comm) : @@ -236,21 +337,29 @@ process_name(s, p) /* To allow these values to be patched, they're globals here */ #include -struct rlimit vm_initial_limit_stack = { DFLSSIZ, MAXSSIZ }; +struct rlimit vm_initial_limit_stack = { DFLSSIZ, MAXSSIZ - PAGE_SIZE }; struct rlimit vm_initial_limit_data = { DFLDSIZ, MAXDSIZ }; struct rlimit vm_initial_limit_core = { DFLCSIZ, MAXCSIZ }; -extern thread_t cloneproc(struct proc *, int); +extern thread_t cloneproc(task_t, proc_t, int); extern int (*mountroot)(void); -extern int netboot_mountroot(); /* netboot.c */ -extern int netboot_setup(struct proc * p); +extern int netboot_mountroot(void); /* netboot.c */ +extern int netboot_setup(void); lck_grp_t * proc_lck_grp; +lck_grp_t * proc_slock_grp; +lck_grp_t * proc_fdmlock_grp; +lck_grp_t * proc_mlock_grp; lck_grp_attr_t * proc_lck_grp_attr; lck_attr_t * proc_lck_attr; +lck_mtx_t * proc_list_mlock; +lck_mtx_t * proc_klist_mlock; + +extern lck_mtx_t * execargs_cache_lock; /* hook called after root is mounted XXX temporary hack */ void (*mountroot_post_hook)(void); +void (*unmountroot_pre_hook)(void); /* * This function is called very early on in the Mach startup, from the @@ -260,128 +369,199 @@ void (*mountroot_post_hook)(void); * Internally, kernel_thread_create() calls thread_create_internal(), * which calls uthread_alloc(). The function of uthread_alloc() is * normally to allocate a uthread structure, and fill out the uu_sigmask, - * uu_act, and uu_ucred fields. It skips filling these out in the case - * of the "task" being "kernel_task", because the order of operation is - * inverted. To account for that, we need to manually fill in at least - * the uu_cred field so that the uthread structure can be used like any - * other. + * uu_context fields. It skips filling these out in the case of the "task" + * being "kernel_task", because the order of operation is inverted. To + * account for that, we need to manually fill in at least the contents + * of the uu_context.vc_ucred field so that the uthread structure can be + * used like any other. */ void -bsd_init() +bsd_init(void) { - register struct proc *p; struct uthread *ut; - extern kauth_cred_t rootcred; - register int i; - int s; + unsigned int i; +#if __i386__ || __x86_64__ int error; - thread_t th; +#endif struct vfs_context context; - void lightning_bolt(void ); kern_return_t ret; - boolean_t funnel_state; struct ucred temp_cred; - extern void file_lock_init(void); + +#define bsd_init_kprintf(x...) /* kprintf("bsd_init: " x) */ kernel_flock = funnel_alloc(KERNEL_FUNNEL); if (kernel_flock == (funnel_t *)0 ) { panic("bsd_init: Failed to allocate kernel funnel"); } - funnel_state = thread_funnel_set(kernel_flock, TRUE); - printf(copyright); + bsd_init_kprintf("calling kmeminit\n"); kmeminit(); + bsd_init_kprintf("calling parse_bsd_args\n"); parse_bsd_args(); - /* Initialize the uthread zone */ - //uthread_zone_init(); /* XXX redundant: previous uthread_alloc() */ - /* Initialize kauth subsystem before instancing the first credential */ + bsd_init_kprintf("calling kauth_init\n"); kauth_init(); /* Initialize process and pgrp structures. */ + bsd_init_kprintf("calling procinit\n"); procinit(); - kernproc = &proc0; + /* Initialize the ttys (MUST be before kminit()/bsd_autoconf()!)*/ + tty_init(); - p = kernproc; + kernproc = &proc0; /* implicitly bzero'ed */ /* kernel_task->proc = kernproc; */ - set_bsdtask_info(kernel_task,(void *)p); - p->p_pid = 0; + set_bsdtask_info(kernel_task,(void *)kernproc); /* give kernproc a name */ - process_name("kernel_task", p); + bsd_init_kprintf("calling process_name\n"); + process_name("kernel_task", kernproc); /* allocate proc lock group attribute and group */ + bsd_init_kprintf("calling lck_grp_attr_alloc_init\n"); proc_lck_grp_attr= lck_grp_attr_alloc_init(); - - proc_lck_grp = lck_grp_alloc_init("proc", proc_lck_grp_attr); + proc_lck_grp = lck_grp_alloc_init("proc", proc_lck_grp_attr); +#ifndef CONFIG_EMBEDDED + proc_slock_grp = lck_grp_alloc_init("proc-slock", proc_lck_grp_attr); + proc_fdmlock_grp = lck_grp_alloc_init("proc-fdmlock", proc_lck_grp_attr); + proc_mlock_grp = lck_grp_alloc_init("proc-mlock", proc_lck_grp_attr); +#endif /* Allocate proc lock attribute */ proc_lck_attr = lck_attr_alloc_init(); +#if 0 +#if __PROC_INTERNAL_DEBUG + lck_attr_setdebug(proc_lck_attr); +#endif +#endif - lck_mtx_init(&p->p_mlock, proc_lck_grp, proc_lck_attr); - lck_mtx_init(&p->p_fdmlock, proc_lck_grp, proc_lck_attr); +#ifdef CONFIG_EMBEDDED + proc_list_mlock = lck_mtx_alloc_init(proc_lck_grp, proc_lck_attr); + proc_klist_mlock = lck_mtx_alloc_init(proc_lck_grp, proc_lck_attr); + lck_mtx_init(&kernproc->p_mlock, proc_lck_grp, proc_lck_attr); + lck_mtx_init(&kernproc->p_fdmlock, proc_lck_grp, proc_lck_attr); + lck_spin_init(&kernproc->p_slock, proc_lck_grp, proc_lck_attr); +#else + proc_list_mlock = lck_mtx_alloc_init(proc_mlock_grp, proc_lck_attr); + proc_klist_mlock = lck_mtx_alloc_init(proc_mlock_grp, proc_lck_attr); + lck_mtx_init(&kernproc->p_mlock, proc_mlock_grp, proc_lck_attr); + lck_mtx_init(&kernproc->p_fdmlock, proc_fdmlock_grp, proc_lck_attr); + lck_spin_init(&kernproc->p_slock, proc_slock_grp, proc_lck_attr); +#endif + execargs_cache_lock = lck_mtx_alloc_init(proc_lck_grp, proc_lck_attr); + execargs_cache_size = bsd_simul_execs; + execargs_free_count = bsd_simul_execs; + execargs_cache = (vm_offset_t *)kalloc(bsd_simul_execs * sizeof(vm_offset_t)); + bzero(execargs_cache, bsd_simul_execs * sizeof(vm_offset_t)); + if (current_task() != kernel_task) printf("bsd_init: We have a problem, " "current task is not kernel task\n"); + bsd_init_kprintf("calling get_bsdthread_info\n"); ut = (uthread_t)get_bsdthread_info(current_thread()); +#if CONFIG_MACF + /* + * Initialize the MAC Framework + */ + mac_policy_initbsd(); + kernproc->p_mac_enforce = 0; +#endif /* MAC */ + /* * Create process 0. */ - LIST_INSERT_HEAD(&allproc, p, p_list); - p->p_pgrp = &pgrp0; + proc_list_lock(); + LIST_INSERT_HEAD(&allproc, kernproc, p_list); + kernproc->p_pgrp = &pgrp0; LIST_INSERT_HEAD(PGRPHASH(0), &pgrp0, pg_hash); LIST_INIT(&pgrp0.pg_members); - LIST_INSERT_HEAD(&pgrp0.pg_members, p, p_pglist); +#ifdef CONFIG_EMBEDDED + lck_mtx_init(&pgrp0.pg_mlock, proc_lck_grp, proc_lck_attr); +#else + lck_mtx_init(&pgrp0.pg_mlock, proc_mlock_grp, proc_lck_attr); +#endif + /* There is no other bsd thread this point and is safe without pgrp lock */ + LIST_INSERT_HEAD(&pgrp0.pg_members, kernproc, p_pglist); + kernproc->p_listflag |= P_LIST_INPGRP; + kernproc->p_pgrpid = 0; pgrp0.pg_session = &session0; + pgrp0.pg_membercnt = 1; + session0.s_count = 1; - session0.s_leader = p; + session0.s_leader = kernproc; + session0.s_listflags = 0; +#ifdef CONFIG_EMBEDDED + lck_mtx_init(&session0.s_mlock, proc_lck_grp, proc_lck_attr); +#else + lck_mtx_init(&session0.s_mlock, proc_mlock_grp, proc_lck_attr); +#endif + LIST_INSERT_HEAD(SESSHASH(0), &session0, s_hash); + proc_list_unlock(); + +#if CONFIG_LCTX + kernproc->p_lctx = NULL; +#endif + + kernproc->task = kernel_task; + + kernproc->p_stat = SRUN; + kernproc->p_flag = P_SYSTEM; + kernproc->p_nice = NZERO; + kernproc->p_pptr = kernproc; - p->task = kernel_task; + TAILQ_INIT(&kernproc->p_uthlist); + TAILQ_INSERT_TAIL(&kernproc->p_uthlist, ut, uu_list); - p->p_stat = SRUN; - p->p_flag = P_SYSTEM; - p->p_nice = NZERO; - p->p_pptr = p; - lockinit(&p->signal_lock, PVM, "signal", 0, 0); - TAILQ_INIT(&p->p_uthlist); - p->sigwait = FALSE; - p->sigwait_thread = THREAD_NULL; - p->exit_thread = THREAD_NULL; + kernproc->sigwait = FALSE; + kernproc->sigwait_thread = THREAD_NULL; + kernproc->exit_thread = THREAD_NULL; + kernproc->p_csflags = CS_VALID; /* * Create credential. This also Initializes the audit information. - * XXX It is not clear what the initial values should be for audit ID, - * XXX session ID, etc.. */ + bsd_init_kprintf("calling bzero\n"); bzero(&temp_cred, sizeof(temp_cred)); temp_cred.cr_ngroups = 1; - p->p_ucred = kauth_cred_create(&temp_cred); + temp_cred.cr_audit.as_aia_p = &audit_default_aia; + /* XXX the following will go away with cr_au */ + temp_cred.cr_au.ai_auid = AU_DEFAUDITID; + + bsd_init_kprintf("calling kauth_cred_create\n"); + kernproc->p_ucred = kauth_cred_create(&temp_cred); /* give the (already exisiting) initial thread a reference on it */ - kauth_cred_ref(p->p_ucred); - ut->uu_ucred = p->p_ucred; - - TAILQ_INIT(&p->aio_activeq); - TAILQ_INIT(&p->aio_doneq); - p->aio_active_count = 0; - p->aio_done_count = 0; + bsd_init_kprintf("calling kauth_cred_ref\n"); + kauth_cred_ref(kernproc->p_ucred); + ut->uu_context.vc_ucred = kernproc->p_ucred; + ut->uu_context.vc_thread = current_thread(); + TAILQ_INIT(&kernproc->p_aio_activeq); + TAILQ_INIT(&kernproc->p_aio_doneq); + kernproc->p_aio_total_count = 0; + kernproc->p_aio_active_count = 0; + + bsd_init_kprintf("calling file_lock_init\n"); file_lock_init(); +#if CONFIG_MACF + mac_cred_label_associate_kernel(kernproc->p_ucred); + mac_task_label_update_cred (kernproc->p_ucred, (struct task *) kernproc->task); +#endif + /* Create the file descriptor table. */ filedesc0.fd_refcnt = 1+1; /* +1 so shutdown will not _FREE_ZONE */ - p->p_fd = &filedesc0; + kernproc->p_fd = &filedesc0; filedesc0.fd_cmask = cmask; filedesc0.fd_knlistsize = -1; filedesc0.fd_knlist = NULL; @@ -389,8 +569,8 @@ bsd_init() filedesc0.fd_knhashmask = 0; /* Create the limits structures. */ - p->p_limit = &limit0; - for (i = 0; i < sizeof(p->p_rlimit)/sizeof(p->p_rlimit[0]); i++) + kernproc->p_limit = &limit0; + for (i = 0; i < sizeof(kernproc->p_rlimit)/sizeof(kernproc->p_rlimit[0]); i++) limit0.pl_rlimit[i].rlim_cur = limit0.pl_rlimit[i].rlim_max = RLIM_INFINITY; limit0.pl_rlimit[RLIMIT_NOFILE].rlim_cur = NOFILE; @@ -399,14 +579,15 @@ bsd_init() limit0.pl_rlimit[RLIMIT_STACK] = vm_initial_limit_stack; limit0.pl_rlimit[RLIMIT_DATA] = vm_initial_limit_data; limit0.pl_rlimit[RLIMIT_CORE] = vm_initial_limit_core; - limit0.p_refcnt = 1; + limit0.pl_refcnt = 1; - p->p_stats = &pstats0; - p->p_sigacts = &sigacts0; + kernproc->p_stats = &pstats0; + kernproc->p_sigacts = &sigacts0; /* * Charge root for two processes: init and mach_init. */ + bsd_init_kprintf("calling chgproccnt\n"); (void)chgproccnt(0, 1); /* @@ -414,11 +595,12 @@ bsd_init() * for temporary copying (execve()). */ { - vm_offset_t min; + vm_offset_t minimum; + bsd_init_kprintf("calling kmem_suballoc\n"); ret = kmem_suballoc(kernel_map, - &min, - (vm_size_t)BSD_PAGABLE_MAP_SIZE, + &minimum, + (vm_size_t)bsd_pageable_map_size, TRUE, VM_FLAGS_ANYWHERE, &bsd_pageable_map); @@ -433,54 +615,92 @@ bsd_init() * happen after a credential has been associated with * the kernel task. */ + bsd_init_kprintf("calling bsd_bufferinit\n"); bsd_bufferinit(); /* Initialize the execve() semaphore */ - ret = semaphore_create(kernel_task, &execve_semaphore, - SYNC_POLICY_FIFO, (BSD_PAGABLE_MAP_SIZE / NCARGS)); + bsd_init_kprintf("calling semaphore_create\n"); + if (ret != KERN_SUCCESS) panic("bsd_init: Failed to create execve semaphore"); /* * Initialize the calendar. */ - IOKitResetTime(); + bsd_init_kprintf("calling IOKitInitializeTime\n"); + IOKitInitializeTime(); + if (turn_on_log_leaks && !new_nkdbufs) + new_nkdbufs = 200000; + start_kern_tracing(new_nkdbufs); + if (turn_on_log_leaks) + log_leaks = 1; + + bsd_init_kprintf("calling ubc_init\n"); ubc_init(); /* Initialize the file systems. */ + bsd_init_kprintf("calling vfsinit\n"); vfsinit(); +#if SOCKETS + /* Initialize per-CPU cache allocator */ + mcache_init(); + /* Initialize mbuf's. */ + bsd_init_kprintf("calling mbinit\n"); mbinit(); + net_str_id_init(); /* for mbuf tags */ +#endif /* SOCKETS */ /* * Initializes security event auditing. * XXX: Should/could this occur later? */ - audit_init(); +#if CONFIG_AUDIT + bsd_init_kprintf("calling audit_init\n"); + audit_init(); +#endif /* Initialize kqueues */ + bsd_init_kprintf("calling knote_init\n"); knote_init(); /* Initialize for async IO */ + bsd_init_kprintf("calling aio_init\n"); aio_init(); /* Initialize pipes */ + bsd_init_kprintf("calling pipeinit\n"); pipeinit(); /* Initialize SysV shm subsystem locks; the subsystem proper is * initialized through a sysctl. */ +#if SYSV_SHM + bsd_init_kprintf("calling sysv_shm_lock_init\n"); sysv_shm_lock_init(); +#endif +#if SYSV_SEM + bsd_init_kprintf("calling sysv_sem_lock_init\n"); sysv_sem_lock_init(); +#endif +#if SYSV_MSG + bsd_init_kprintf("sysv_msg_lock_init\n"); sysv_msg_lock_init(); +#endif + bsd_init_kprintf("calling pshm_lock_init\n"); pshm_lock_init(); + bsd_init_kprintf("calling psem_lock_init\n"); psem_lock_init(); + pthread_init(); /* POSIX Shm and Sem */ + bsd_init_kprintf("calling pshm_cache_init\n"); pshm_cache_init(); + bsd_init_kprintf("calling psem_cache_init\n"); psem_cache_init(); + bsd_init_kprintf("calling time_zone_slock_init\n"); time_zone_slock_init(); /* Stack snapshot facility lock */ @@ -489,15 +709,31 @@ bsd_init() * Initialize protocols. Block reception of incoming packets * until everything is ready. */ + bsd_init_kprintf("calling sysctl_register_fixed\n"); sysctl_register_fixed(); + bsd_init_kprintf("calling sysctl_mib_init\n"); sysctl_mib_init(); +#if NETWORKING + bsd_init_kprintf("calling dlil_init\n"); dlil_init(); + bsd_init_kprintf("calling proto_kpi_init\n"); proto_kpi_init(); +#endif /* NETWORKING */ +#if SOCKETS + bsd_init_kprintf("calling socketinit\n"); socketinit(); + bsd_init_kprintf("calling domaininit\n"); domaininit(); +#endif /* SOCKETS */ - p->p_fd->fd_cdir = NULL; - p->p_fd->fd_rdir = NULL; + kernproc->p_fd->fd_cdir = NULL; + kernproc->p_fd->fd_rdir = NULL; + +#if CONFIG_EMBEDDED + /* Initialize kernel memory status notifications */ + bsd_init_kprintf("calling kern_memorystatus_init\n"); + kern_memorystatus_init(); +#endif #ifdef GPROF /* Initialize kernel profiling. */ @@ -506,10 +742,15 @@ bsd_init() /* kick off timeout driven events by calling first time */ thread_wakeup(&lbolt); - timeout((void (*)(void *))lightning_bolt, 0, hz); + timeout(lightning_bolt, 0, hz); + bsd_init_kprintf("calling bsd_autoconf\n"); bsd_autoconf(); +#if CONFIG_DTRACE + dtrace_postinit(); +#endif + /* * We attach the loopback interface *way* down here to ensure * it happens after autoconf(), otherwise it becomes the @@ -517,15 +758,31 @@ bsd_init() */ #include #if NLOOP > 0 + bsd_init_kprintf("calling loopattach\n"); loopattach(); /* XXX */ #endif - - /* Register the built-in dlil ethernet interface family */ + +#if PFLOG + /* Initialize packet filter log interface */ + pfloginit(); +#endif /* PFLOG */ + +#if NETHER > 0 + /* Register the built-in dlil ethernet interface family */ + bsd_init_kprintf("calling ether_family_init\n"); ether_family_init(); +#endif /* ETHER */ +#if NETWORKING /* Call any kext code that wants to run just after network init */ + bsd_init_kprintf("calling net_init_run\n"); net_init_run(); + + /* register user tunnel kernel control handler */ + utun_register_control(); +#endif /* NETWORKING */ + bsd_init_kprintf("calling vnode_pager_bootstrap\n"); vnode_pager_bootstrap(); #if 0 /* XXX Hack for early debug stop */ @@ -534,36 +791,89 @@ bsd_init() /* Debugger("hello"); */ #endif + bsd_init_kprintf("calling inittodr\n"); inittodr(0); +#if CONFIG_EMBEDDED + { + /* print out early VM statistics */ + kern_return_t kr1; + vm_statistics_data_t stat; + mach_msg_type_number_t count; + + count = HOST_VM_INFO_COUNT; + kr1 = host_statistics(host_self(), + HOST_VM_INFO, + (host_info_t)&stat, + &count); + kprintf("Mach Virtual Memory Statistics (page size of 4096) bytes\n" + "Pages free:\t\t\t%u.\n" + "Pages active:\t\t\t%u.\n" + "Pages inactive:\t\t\t%u.\n" + "Pages wired down:\t\t%u.\n" + "\"Translation faults\":\t\t%u.\n" + "Pages copy-on-write:\t\t%u.\n" + "Pages zero filled:\t\t%u.\n" + "Pages reactivated:\t\t%u.\n" + "Pageins:\t\t\t%u.\n" + "Pageouts:\t\t\t%u.\n" + "Object cache: %u hits of %u lookups (%d%% hit rate)\n", + + stat.free_count, + stat.active_count, + stat.inactive_count, + stat.wire_count, + stat.faults, + stat.cow_faults, + stat.zero_fill_count, + stat.reactivations, + stat.pageins, + stat.pageouts, + stat.hits, + stat.lookups, + (stat.hits == 0) ? 100 : + ((stat.lookups * 100) / stat.hits)); + } +#endif /* CONFIG_EMBEDDED */ + /* Mount the root file system. */ while( TRUE) { int err; + bsd_init_kprintf("calling setconf\n"); setconf(); - bsd_hardclockinit = -1; /* start ticking */ + bsd_init_kprintf("vfs_mountroot\n"); if (0 == (err = vfs_mountroot())) break; + rootdevice[0] = '\0'; #if NFSCLIENT if (mountroot == netboot_mountroot) { - printf("cannot mount network root, errno = %d\n", err); - mountroot = NULL; - if (0 == (err = vfs_mountroot())) - break; + PE_display_icon( 0, "noroot"); /* XXX a netboot-specific icon would be nicer */ + vc_progress_set(FALSE, 0); + for (i=1; 1; i*=2) { + printf("bsd_init: failed to mount network root, error %d, %s\n", + err, PE_boot_args()); + printf("We are hanging here...\n"); + IOSleep(i*60*1000); + } + /*NOTREACHED*/ } #endif printf("cannot mount root, errno = %d\n", err); boothowto |= RB_ASKNAME; } - context.vc_proc = p; - context.vc_ucred = p->p_ucred; + IOSecureBSDRoot(rootdevice); + + context.vc_thread = current_thread(); + context.vc_ucred = kernproc->p_ucred; mountlist.tqh_first->mnt_flag |= MNT_ROOTFS; + bsd_init_kprintf("calling VFS_ROOT\n"); /* Get the vnode for '/'. Set fdp->fd_fd.fd_cdir to reference it. */ if (VFS_ROOT(mountlist.tqh_first, &rootvnode, &context)) - panic("bsd_init: cannot find root vnode"); + panic("bsd_init: cannot find root vnode: %s", PE_boot_args()); rootvnode->v_flag |= VROOT; (void)vnode_ref(rootvnode); (void)vnode_put(rootvnode); @@ -573,97 +883,128 @@ bsd_init() if (mountroot == netboot_mountroot) { int err; /* post mount setup */ - if (err = netboot_setup(p)) { - panic("bsd_init: NetBoot could not find root, %d", err); + if ((err = netboot_setup()) != 0) { + PE_display_icon( 0, "noroot"); /* XXX a netboot-specific icon would be nicer */ + vc_progress_set(FALSE, 0); + for (i=1; 1; i*=2) { + printf("bsd_init: NetBoot could not find root, error %d: %s\n", + err, PE_boot_args()); + printf("We are hanging here...\n"); + IOSleep(i*60*1000); + } + /*NOTREACHED*/ } } #endif - microtime(&p->p_stats->p_start); - p->p_rtime.tv_sec = p->p_rtime.tv_usec = 0; +#if CONFIG_IMAGEBOOT + /* + * See if a system disk image is present. If so, mount it and + * switch the root vnode to point to it + */ + + if(imageboot_needed()) { + int err; + + /* An image was found */ + if((err = imageboot_setup())) { + /* + * this is not fatal. Keep trying to root + * off the original media + */ + printf("%s: imageboot could not find root, %d\n", + __FUNCTION__, err); + } + } +#endif /* CONFIG_IMAGEBOOT */ + + /* set initial time; all other resource data is already zero'ed */ + microtime(&kernproc->p_start); + kernproc->p_stats->p_start = kernproc->p_start; /* for compat */ #if DEVFS { - extern void devfs_kernel_mount(char * str); - - devfs_kernel_mount("/dev"); + char mounthere[] = "/dev"; /* !const because of internal casting */ + + bsd_init_kprintf("calling devfs_kernel_mount\n"); + devfs_kernel_mount(mounthere); } #endif /* DEVFS */ /* Initialize signal state for process 0. */ - siginit(p); + bsd_init_kprintf("calling siginit\n"); + siginit(kernproc); + bsd_init_kprintf("calling bsd_utaskbootstrap\n"); bsd_utaskbootstrap(); -#if __i386__ - // this should be done after the root filesystem is mounted - error = set_archhandler(p, CPU_TYPE_POWERPC); - if (error) +#if defined(__LP64__) + kernproc->p_flag |= P_LP64; + printf("Kernel is LP64\n"); +#endif +#if __i386__ || __x86_64__ + /* this should be done after the root filesystem is mounted */ + error = set_archhandler(kernproc, CPU_TYPE_POWERPC); + // 10/30/08 - gab: + // if default 'translate' can't be found, see if the understudy is available + if (ENOENT == error) { + strlcpy(exec_archhandler_ppc.path, kRosettaStandIn_str, MAXPATHLEN); + error = set_archhandler(kernproc, CPU_TYPE_POWERPC); + } + if (error) /* XXX make more generic */ exec_archhandler_ppc.path[0] = 0; #endif - + + bsd_init_kprintf("calling mountroot_post_hook\n"); + /* invoke post-root-mount hook */ if (mountroot_post_hook != NULL) mountroot_post_hook(); - - (void) thread_funnel_set(kernel_flock, funnel_state); + +#if 0 /* not yet */ + consider_zone_gc(FALSE); +#endif + + bsd_init_kprintf("done\n"); } /* Called with kernel funnel held */ void bsdinit_task(void) { - struct proc *p = current_proc(); + proc_t p = current_proc(); struct uthread *ut; - kern_return_t kr; - thread_t th_act; - shared_region_mapping_t system_region; + thread_t thread; process_name("init", p); ux_handler_init(); - th_act = current_thread(); + thread = current_thread(); (void) host_set_exception_ports(host_priv_self(), - EXC_MASK_ALL & ~(EXC_MASK_SYSCALL | - EXC_MASK_MACH_SYSCALL | - EXC_MASK_RPC_ALERT), - ux_exception_port, - EXCEPTION_DEFAULT, 0); - - (void) task_set_exception_ports(get_threadtask(th_act), - EXC_MASK_ALL & ~(EXC_MASK_SYSCALL | - EXC_MASK_MACH_SYSCALL | - EXC_MASK_RPC_ALERT), - ux_exception_port, - EXCEPTION_DEFAULT, 0); - + EXC_MASK_ALL & ~(EXC_MASK_RPC_ALERT),//pilotfish (shark) needs this port + (mach_port_t) ux_exception_port, + EXCEPTION_DEFAULT| MACH_EXCEPTION_CODES, + 0); + ut = (uthread_t)get_bsdthread_info(thread); - - ut = (uthread_t)get_bsdthread_info(th_act); - - bsd_hardclockinit = 1; /* Start bsd hardclock */ - bsd_init_task = get_threadtask(th_act); + bsd_init_task = get_threadtask(thread); init_task_failure_data[0] = 0; - system_region = lookup_default_shared_region(ENV_DEFAULT_ROOT, cpu_type()); - if (system_region == NULL) { - shared_file_boot_time_init(ENV_DEFAULT_ROOT, cpu_type()); - } else { - vm_set_shared_region(get_threadtask(th_act), system_region); - } + +#if CONFIG_MACF + mac_cred_label_associate_user(p->p_ucred); + mac_task_label_update_cred (p->p_ucred, (struct task *) p->task); +#endif load_init_program(p); - /* turn on app-profiling i.e. pre-heating */ - app_profile = 1; lock_trace = 1; } void -lightning_bolt() +lightning_bolt(__unused void *dummy) { boolean_t funnel_state; - extern void klogwakeup(void); funnel_state = thread_funnel_set(kernel_flock, TRUE); @@ -674,10 +1015,10 @@ lightning_bolt() (void) thread_funnel_set(kernel_flock, FALSE); } -bsd_autoconf() +kern_return_t +bsd_autoconf(void) { - extern kern_return_t IOKitBSDInit( void ); - + kprintf("bsd_autoconf: calling kminit\n"); kminit(); /* @@ -699,8 +1040,6 @@ bsd_autoconf() static void setconf(void) { - extern kern_return_t IOFindBSDRoot( char * rootName, - dev_t * root, u_int32_t * flags ); u_int32_t flags; kern_return_t err; @@ -709,14 +1048,12 @@ setconf(void) * which needs to be under network funnel. Right thing to do * here is to drop the funnel alltogether and regrab it afterwards */ - thread_funnel_set(kernel_flock, FALSE); - err = IOFindBSDRoot( rootdevice, &rootdev, &flags ); - thread_funnel_set(kernel_flock, TRUE); + err = IOFindBSDRoot(rootdevice, sizeof(rootdevice), &rootdev, &flags); if( err) { printf("setconf: IOFindBSDRoot returned an error (%d);" "setting rootdevice to 'sd0a'.\n", err); /* XXX DEBUG TEMP */ rootdev = makedev( 6, 0 ); - strcpy( rootdevice, "sd0a" ); + strlcpy(rootdevice, "sd0a", sizeof(rootdevice)); flags = 0; } @@ -734,89 +1071,101 @@ setconf(void) } -bsd_utaskbootstrap() +void +bsd_utaskbootstrap(void) { - thread_t th_act; + thread_t thread; struct uthread *ut; - th_act = cloneproc(kernproc, 0); - initproc = pfind(1); - /* Set the launch time for init */ - microtime(&initproc->p_stats->p_start); + /* + * Clone the bootstrap process from the kernel process, without + * inheriting either task characteristics or memory from the kernel; + */ + thread = cloneproc(TASK_NULL, kernproc, FALSE); + + /* Hold the reference as it will be dropped during shutdown */ + initproc = proc_find(1); +#if __PROC_INTERNAL_DEBUG + if (initproc == PROC_NULL) + panic("bsd_utaskbootstrap: initproc not set\n"); +#endif + /* + * Since we aren't going back out the normal way to our parent, + * we have to drop the transition locks explicitly. + */ + proc_signalend(initproc, 0); + proc_transend(initproc, 0); - ut = (struct uthread *)get_bsdthread_info(th_act); + ut = (struct uthread *)get_bsdthread_info(thread); ut->uu_sigmask = 0; - act_set_astbsd(th_act); - (void) thread_resume(th_act); + act_set_astbsd(thread); + (void) thread_resume(thread); } -parse_bsd_args() +static void +parse_bsd_args(void) { - extern char init_args[]; - char namep[16]; - extern int boothowto; - extern int srv; - extern int ncl; + char namep[16]; + int msgbuf; - int len; - - if (PE_parse_boot_arg("-s", namep)) { + if (PE_parse_boot_argn("-s", namep, sizeof (namep))) boothowto |= RB_SINGLE; - len = strlen(init_args); - if(len != 0) - strcat(init_args," -s"); - else - strcat(init_args,"-s"); - } - if (PE_parse_boot_arg("-b", namep)) { + if (PE_parse_boot_argn("-b", namep, sizeof (namep))) boothowto |= RB_NOBOOTRC; - len = strlen(init_args); - if(len != 0) - strcat(init_args," -b"); - else - strcat(init_args,"-b"); - } - if (PE_parse_boot_arg("-F", namep)) { - len = strlen(init_args); - if(len != 0) - strcat(init_args," -F"); - else - strcat(init_args,"-F"); - } + if (PE_parse_boot_argn("-x", namep, sizeof (namep))) /* safe boot */ + boothowto |= RB_SAFEBOOT; - if (PE_parse_boot_arg("-v", namep)) { - len = strlen(init_args); - if(len != 0) - strcat(init_args," -v"); - else - strcat(init_args,"-v"); - } + if (PE_parse_boot_argn("-l", namep, sizeof (namep))) /* leaks logging */ + turn_on_log_leaks = 1; + + /* disable 64 bit grading */ + if (PE_parse_boot_argn("-no64exec", namep, sizeof (namep))) + bootarg_no64exec = 1; - if (PE_parse_boot_arg("-x", namep)) { /* safe boot */ - len = strlen(init_args); - if(len != 0) - strcat(init_args," -x"); - else - strcat(init_args,"-x"); + PE_parse_boot_argn("ncl", &ncl, sizeof (ncl)); + if (PE_parse_boot_argn("nbuf", &max_nbuf_headers, + sizeof (max_nbuf_headers))) { + customnbuf = 1; } +#if !defined(SECURE_KERNEL) + PE_parse_boot_argn("kmem", &setup_kmem, sizeof (setup_kmem)); +#endif + PE_parse_boot_argn("trace", &new_nkdbufs, sizeof (new_nkdbufs)); - if (PE_parse_boot_arg("-d", namep)) { - len = strlen(init_args); - if(len != 0) - strcat(init_args," -d"); - else - strcat(init_args,"-d"); + if (PE_parse_boot_argn("msgbuf", &msgbuf, sizeof (msgbuf))) { + log_setsize(msgbuf); } +} - PE_parse_boot_arg("srv", &srv); - PE_parse_boot_arg("ncl", &ncl); - if (PE_parse_boot_arg("nbuf", &max_nbuf_headers)) - customnbuf = 1; - PE_parse_boot_arg("kmem", &setup_kmem); +void +bsd_exec_setup(int scale) +{ - return 0; + switch (scale) { + case 0: + case 1: + bsd_simul_execs = BSD_SIMUL_EXECS; + break; + case 2: + case 3: + bsd_simul_execs = 65; + break; + case 4: + case 5: + bsd_simul_execs = 129; + break; + case 6: + case 7: + bsd_simul_execs = 257; + break; + default: + bsd_simul_execs = 513; + break; + + } + bsd_pageable_map_size = (bsd_simul_execs * (NCARGS + PAGE_SIZE)); } #if !NFSCLIENT