#include <kern/machine.h>
#include <kern/processor.h>
#include <kern/sched_prim.h>
+#if CONFIG_SCHED_SFI
#include <kern/sfi.h>
+#endif
#include <kern/startup.h>
#include <kern/task.h>
#include <kern/thread.h>
#if CONFIG_TELEMETRY
#include <kern/telemetry.h>
#endif
-#include <kern/wait_queue.h>
#include <kern/xpr.h>
#include <kern/zalloc.h>
#include <kern/locks.h>
+#include <kern/debug.h>
+#include <corpses/task_corpse.h>
#include <prng/random.h>
#include <console/serial_protos.h>
#include <vm/vm_kern.h>
#include <sys/codesign.h>
#include <sys/kdebug.h>
#include <sys/random.h>
+#include <sys/ktrace.h>
+
+#include <kern/ltable.h>
+#include <kern/waitq.h>
+#include <ipc/ipc_voucher.h>
+#include <voucher/ipc_pthread_priority_internal.h>
+#include <mach/host_info.h>
+
#if CONFIG_ATM
#include <atm/atm_internal.h>
#include <sys/csr.h>
#endif
-#if CONFIG_BANK
#include <bank/bank_internal.h>
-#endif
#if ALTERNATE_DEBUGGER
#include <arm64/alternate_debugger.h>
#if CONFIG_MACF
#include <security/mac_mach_internal.h>
+#if CONFIG_VNGUARD
+extern void vnguard_policy_init(void);
#endif
-
-#if CONFIG_COUNTERS
-#include <pmc/pmc.h>
#endif
#if KPC
#include <kern/kpc.h>
#endif
-#if KPERF
-#include <kperf/kperf.h>
-#endif
-
#if HYPERVISOR
#include <kern/hv_support.h>
#endif
+#include <san/kasan.h>
+
+#if defined(__arm__) || defined(__arm64__)
+#include <arm/misc_protos.h> // for arm_vm_prot_finalize
+#endif
#include <i386/pmCPU.h>
static void kernel_bootstrap_thread(void);
void scale_setup(void);
extern void bsd_scale_setup(int);
extern unsigned int semaphore_max;
-extern void stackshot_lock_init(void);
+extern void stackshot_init(void);
+extern void ktrace_init(void);
+extern void oslog_init(void);
/*
* Running in virtual memory, on the interrupt stack.
unsigned int new_nkdbufs = 0;
unsigned int wake_nkdbufs = 0;
unsigned int write_trace_on_panic = 0;
-unsigned int trace_typefilter = 0;
+static char trace_typefilter[64] = { 0 };
+unsigned int trace_wrap = 0;
+boolean_t trace_serial = FALSE;
+boolean_t early_boot_complete = FALSE;
/* mach leak logging */
int log_leaks = 0;
-int turn_on_log_leaks = 0;
static inline void
kernel_bootstrap_log(const char *message)
{
// kprintf("kernel_bootstrap: %s\n", message);
- kernel_debug_string(message);
+ kernel_debug_string_early(message);
}
static inline void
kernel_bootstrap_thread_log(const char *message)
{
// kprintf("kernel_bootstrap_thread: %s\n", message);
- kernel_debug_string(message);
+ kernel_debug_string_early(message);
}
void
*/
timer_call_init();
+#if CONFIG_SCHED_SFI
/*
* Configure SFI classes
*/
sfi_early_init();
+#endif
}
+extern boolean_t IORamDiskBSDRoot(void);
+extern kern_return_t cpm_preallocate_early(void);
void
kernel_bootstrap(void)
printf("%s\n", version); /* log kernel version */
if (PE_parse_boot_argn("-l", namep, sizeof (namep))) /* leaks logging */
- turn_on_log_leaks = 1;
+ log_leaks = 1;
PE_parse_boot_argn("trace", &new_nkdbufs, sizeof (new_nkdbufs));
PE_parse_boot_argn("trace_wake", &wake_nkdbufs, sizeof (wake_nkdbufs));
PE_parse_boot_argn("trace_panic", &write_trace_on_panic, sizeof(write_trace_on_panic));
- PE_parse_boot_argn("trace_typefilter", &trace_typefilter, sizeof(trace_typefilter));
+ PE_parse_boot_arg_str("trace_typefilter", trace_typefilter, sizeof(trace_typefilter));
+ PE_parse_boot_argn("trace_wrap", &trace_wrap, sizeof(trace_wrap));
scale_setup();
machine_info.major_version = version_major;
machine_info.minor_version = version_minor;
+ oslog_init();
+
+#if KASAN
+ kernel_bootstrap_log("kasan_late_init");
+ kasan_late_init();
+#endif
+
#if CONFIG_TELEMETRY
kernel_bootstrap_log("telemetry_init");
telemetry_init();
#endif
- kernel_bootstrap_log("stackshot_lock_init");
- stackshot_lock_init();
+#if CONFIG_CSR
+ kernel_bootstrap_log("csr_init");
+ csr_init();
+#endif
+
+ if (PE_i_can_has_debugger(NULL) &&
+ PE_parse_boot_argn("-show_pointers", &namep, sizeof (namep))) {
+ doprnt_hide_pointers = FALSE;
+ }
+
+ kernel_bootstrap_log("console_init");
+ console_init();
+
+ kernel_bootstrap_log("stackshot_init");
+ stackshot_init();
kernel_bootstrap_log("sched_init");
sched_init();
- kernel_bootstrap_log("wait_queue_bootstrap");
- wait_queue_bootstrap();
+ kernel_bootstrap_log("ltable_bootstrap");
+ ltable_bootstrap();
+
+ kernel_bootstrap_log("waitq_bootstrap");
+ waitq_bootstrap();
kernel_bootstrap_log("ipc_bootstrap");
ipc_bootstrap();
#if CONFIG_MACF
+ kernel_bootstrap_log("mac_policy_init");
mac_policy_init();
#endif
+
kernel_bootstrap_log("ipc_init");
ipc_init();
/*
* Initialize the IPC, task, and thread subsystems.
*/
+
#if CONFIG_COALITIONS
- kernel_bootstrap_log("coalition_init");
- coalition_init();
+ kernel_bootstrap_log("coalitions_init");
+ coalitions_init();
#endif
kernel_bootstrap_log("task_init");
kernel_bootstrap_log("atm_init");
atm_init();
#endif
+ kernel_bootstrap_log("mach_init_activity_id");
+ mach_init_activity_id();
-#if CONFIG_CSR
- kernel_bootstrap_log("csr_init");
- csr_init();
-#endif
-
-#if CONFIG_BANK
/* Initialize the BANK Manager. */
kernel_bootstrap_log("bank_init");
bank_init();
-#endif
-
+
+ kernel_bootstrap_log("ipc_pthread_priority_init");
+ ipc_pthread_priority_init();
+
+ /* initialize the corpse config based on boot-args */
+ corpses_init();
+
+ /* initialize host_statistics */
+ host_statistics_init();
+
/*
* Create a kernel thread to execute the kernel bootstrap.
*/
if (result != KERN_SUCCESS) panic("kernel_bootstrap: result = %08X\n", result);
thread->state = TH_RUN;
+ thread->last_made_runnable_time = mach_absolute_time();
thread_deallocate(thread);
kernel_bootstrap_log("load_context - done");
vm_offset_t vm_kernel_addrperm;
vm_offset_t buf_kernel_addrperm;
+vm_offset_t vm_kernel_addrperm_ext;
+uint64_t vm_kernel_addrhash_salt;
+uint64_t vm_kernel_addrhash_salt_ext;
/*
* Now running in a thread. Kick off other services,
kernel_bootstrap_thread_log("thread_bind");
thread_bind(processor);
+#if __arm64__
+ if (IORamDiskBSDRoot()) {
+ cpm_preallocate_early();
+ }
+#endif /* __arm64__ */
+
/*
* Initialize ipc thread call support.
*/
device_service_create();
kth_started = 1;
-
+
#if (defined(__i386__) || defined(__x86_64__)) && NCOPY_WINDOWS > 0
/*
* Create and initialize the physical copy window for processor 0
alternate_debugger_init();
#endif
-#if CONFIG_COUNTERS
- pmc_bootstrap();
-#endif
-
#if KPC
kpc_init();
#endif
ecc_log_init();
#endif
-#if KPERF
- kperf_bootstrap();
-#endif
-
#if HYPERVISOR
hv_support_init();
#endif
vmx_init();
#endif
-#if (defined(__i386__) || defined(__x86_64__))
- if (turn_on_log_leaks && !new_nkdbufs)
- new_nkdbufs = 200000;
- if (trace_typefilter)
- start_kern_tracing_with_typefilter(new_nkdbufs,
- FALSE,
- trace_typefilter);
- else
- start_kern_tracing(new_nkdbufs, FALSE);
- if (turn_on_log_leaks)
- log_leaks = 1;
+ kernel_bootstrap_thread_log("ktrace_init");
+ ktrace_init();
-#endif
+ kdebug_init(new_nkdbufs, trace_typefilter, trace_wrap);
kernel_bootstrap_log("prng_init");
prng_cpu_init(master_cpu);
+#ifdef MACH_BSD
+ kernel_bootstrap_log("bsd_early_init");
+ bsd_early_init();
+#endif
+
+#if defined(__arm64__)
+ ml_lockdown_init();
+#endif
+
#ifdef IOKIT
+ kernel_bootstrap_log("PE_init_iokit");
PE_init_iokit();
#endif
assert(ml_get_interrupts_enabled() == FALSE);
+
+ /*
+ * Past this point, kernel subsystems that expect to operate with
+ * interrupts or preemption enabled may begin enforcement.
+ */
+ early_boot_complete = TRUE;
+
+#if INTERRUPT_MASKED_DEBUG
+ // Reset interrupts masked timeout before we enable interrupts
+ ml_spin_debug_clear_self();
+#endif
(void) spllo(); /* Allow interruptions */
#if (defined(__i386__) || defined(__x86_64__)) && NCOPY_WINDOWS > 0
cpu_userwindow_init(0);
#endif
-#if (!defined(__i386__) && !defined(__x86_64__))
- if (turn_on_log_leaks && !new_nkdbufs)
- new_nkdbufs = 200000;
- if (trace_typefilter)
- start_kern_tracing_with_typefilter(new_nkdbufs, FALSE, trace_typefilter);
- else
- start_kern_tracing(new_nkdbufs, FALSE);
- if (turn_on_log_leaks)
- log_leaks = 1;
-#endif
-
/*
* Initialize the shared region module.
*/
vm_commpage_init();
vm_commpage_text_init();
-
#if CONFIG_MACF
kernel_bootstrap_log("mac_policy_initmach");
mac_policy_initmach();
+#if CONFIG_VNGUARD
+ vnguard_policy_init();
+#endif
+#endif
+
+#if defined(__arm__) || defined(__arm64__)
+#if CONFIG_KERNEL_INTEGRITY
+ machine_lockdown_preflight();
+#endif
+ /*
+ * Finalize protections on statically mapped pages now that comm page mapping is established.
+ */
+ arm_vm_prot_finalize(PE_state.bootArgs);
#endif
+#if CONFIG_SCHED_SFI
kernel_bootstrap_log("sfi_init");
sfi_init();
+#endif
/*
- * Initialize the global used for permuting kernel
+ * Initialize the globals used for permuting kernel
* addresses that may be exported to userland as tokens
- * using VM_KERNEL_ADDRPERM(). Force the random number
- * to be odd to avoid mapping a non-zero
+ * using VM_KERNEL_ADDRPERM()/VM_KERNEL_ADDRPERM_EXTERNAL().
+ * Force the random number to be odd to avoid mapping a non-zero
* word-aligned address to zero via addition.
* Note: at this stage we can use the cryptographically secure PRNG
* rather than early_random().
vm_kernel_addrperm |= 1;
read_random(&buf_kernel_addrperm, sizeof(buf_kernel_addrperm));
buf_kernel_addrperm |= 1;
+ read_random(&vm_kernel_addrperm_ext, sizeof(vm_kernel_addrperm_ext));
+ vm_kernel_addrperm_ext |= 1;
+ read_random(&vm_kernel_addrhash_salt, sizeof(vm_kernel_addrhash_salt));
+ read_random(&vm_kernel_addrhash_salt_ext, sizeof(vm_kernel_addrhash_salt_ext));
+
+ vm_set_restrictions();
+
+
+
/*
* Start the user bootstrap.
*/
OSKextRemoveKextBootstrap();
+ /*
+ * Get rid of pages used for early boot tracing.
+ */
+ kdebug_free_early_buf();
+
serial_keyboard_init(); /* Start serial keyboard if wanted */
vm_page_init_local_q();
-
+
thread_bind(PROCESSOR_NULL);
/*
*
* Start the first thread on a processor.
*/
-static void
+static void __attribute__((noreturn))
load_context(
thread_t thread)
{
sched_run_incr(thread);
processor->active_thread = thread;
- processor->current_pri = thread->sched_pri;
- processor->current_thmode = thread->sched_mode;
+ processor_state_update_explicit(processor, thread->sched_pri,
+ SFI_CLASS_KERNEL, PSET_SMP, thread_get_perfcontrol_class(thread));
+ processor->starting_pri = thread->sched_pri;
processor->deadline = UINT64_MAX;
thread->last_processor = processor;
timer_start(&PROCESSOR_DATA(processor, system_state), processor->last_dispatch);
PROCESSOR_DATA(processor, current_state) = &PROCESSOR_DATA(processor, system_state);
+
PMAP_ACTIVATE_USER(thread, processor->cpu_id);
load_context_kprintf("machine_load_context\n");
if (scale > 16)
scale = 16;
task_max_base = 2500;
- } else if ((uint64_t)sane_size >= (uint64_t)(3 * 1024 * 1024 *1024ULL))
- scale = 2;
+ /* Raise limits for machines with >= 3GB */
+ } else if ((uint64_t)sane_size >= (uint64_t)(3 * 1024 * 1024 *1024ULL)) {
+ if ((uint64_t)sane_size < (uint64_t)(8 * 1024 * 1024 *1024ULL)) {
+ scale = 2;
+ } else {
+ /* limit to 64GB */
+ scale = MIN(16, (int)((uint64_t)sane_size / (uint64_t)(4 * 1024 * 1024 *1024ULL)));
+ }
+ }
task_max = MAX(task_max, task_max_base * scale);