]> git.saurik.com Git - apple/xnu.git/blobdiff - osfmk/kern/task.c
xnu-7195.101.1.tar.gz
[apple/xnu.git] / osfmk / kern / task.c
index c5efca2a767b87bbd1537a534cb6203ad21371c2..1266acd83e8b3d820f7af5c96b0b1dc38e1790df 100644 (file)
@@ -1,8 +1,8 @@
 /*
- * Copyright (c) 2000-2009 Apple Inc. All rights reserved.
+ * Copyright (c) 2000-2020 Apple Inc. All rights reserved.
  *
  * @APPLE_OSREFERENCE_LICENSE_HEADER_START@
- * 
+ *
  * 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
  * 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.
- * 
+ *
  * 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,
  * 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_OSREFERENCE_LICENSE_HEADER_END@
  */
 /*
  * @OSF_FREE_COPYRIGHT@
  */
-/* 
+/*
  * Mach Operating System
  * Copyright (c) 1991,1990,1989,1988 Carnegie Mellon University
  * All Rights Reserved.
- * 
+ *
  * Permission to use, copy, modify and distribute this software and its
  * documentation is hereby granted, provided that both the copyright
  * notice and this permission notice appear in all copies of the
  * software, derivative works or modified versions, and any portions
  * thereof, and that both notices appear in supporting documentation.
- * 
+ *
  * CARNEGIE MELLON ALLOWS FREE USE OF THIS SOFTWARE IN ITS "AS IS"
  * CONDITION.  CARNEGIE MELLON DISCLAIMS ANY LIABILITY OF ANY KIND FOR
  * ANY DAMAGES WHATSOEVER RESULTING FROM THE USE OF THIS SOFTWARE.
- * 
+ *
  * Carnegie Mellon requests users of this software to return to
- * 
+ *
  *  Software Distribution Coordinator  or  Software.Distribution@CS.CMU.EDU
  *  School of Computer Science
  *  Carnegie Mellon University
  *  Pittsburgh PA 15213-3890
- * 
+ *
  * any improvements or extensions that they make and grant Carnegie Mellon
  * the rights to redistribute these changes.
  */
  * Copyright (c) 2005 SPARTA, Inc.
  */
 
-#include <mach_kdb.h>
-#include <fast_tas.h>
-#include <platforms.h>
-
 #include <mach/mach_types.h>
 #include <mach/boolean.h>
 #include <mach/host_priv.h>
 #include <mach/machine/vm_types.h>
 #include <mach/vm_param.h>
+#include <mach/mach_vm.h>
 #include <mach/semaphore.h>
 #include <mach/task_info.h>
+#include <mach/task_inspect.h>
 #include <mach/task_special_ports.h>
+#include <mach/sdt.h>
 
+#include <ipc/ipc_importance.h>
 #include <ipc/ipc_types.h>
 #include <ipc/ipc_space.h>
 #include <ipc/ipc_entry.h>
+#include <ipc/ipc_hash.h>
 
 #include <kern/kern_types.h>
 #include <kern/mach_param.h>
 #include <kern/misc_protos.h>
 #include <kern/task.h>
 #include <kern/thread.h>
+#include <kern/coalition.h>
 #include <kern/zalloc.h>
 #include <kern/kalloc.h>
+#include <kern/kern_cdata.h>
 #include <kern/processor.h>
-#include <kern/sched_prim.h>   /* for thread_wakeup */
+#include <kern/sched_prim.h>    /* for thread_wakeup */
 #include <kern/ipc_tt.h>
-#include <kern/ledger.h>
 #include <kern/host.h>
 #include <kern/clock.h>
 #include <kern/timer.h>
 #include <kern/assert.h>
 #include <kern/sync_lock.h>
 #include <kern/affinity.h>
+#include <kern/exc_resource.h>
+#include <kern/machine.h>
+#include <kern/policy_internal.h>
+#include <kern/restartable.h>
+
+#include <corpses/task_corpse.h>
+#if CONFIG_TELEMETRY
+#include <kern/telemetry.h>
+#endif
+
+#if MONOTONIC
+#include <kern/monotonic.h>
+#include <machine/monotonic.h>
+#endif /* MONOTONIC */
+
+#include <os/log.h>
 
 #include <vm/pmap.h>
 #include <vm/vm_map.h>
-#include <vm/vm_kern.h>                /* for kernel_map, ipc_kernel_map */
+#include <vm/vm_kern.h>         /* for kernel_map, ipc_kernel_map */
 #include <vm/vm_pageout.h>
 #include <vm/vm_protos.h>
+#include <vm/vm_purgeable_internal.h>
+#include <vm/vm_compressor_pager.h>
 
-#if    MACH_KDB
-#include <ddb/db_sym.h>
-#endif /* MACH_KDB */
-
-#ifdef __ppc__
-#include <ppc/exception.h>
-#include <ppc/hw_perfmon.h>
-#endif
-
-
+#include <sys/resource.h>
+#include <sys/signalvar.h> /* for coredump */
+#include <sys/bsdtask_info.h>
 /*
  * Exported interfaces
  */
 #include <mach/mach_host_server.h>
 #include <mach/host_security_server.h>
 #include <mach/mach_port_server.h>
-#include <mach/security_server.h>
 
 #include <vm/vm_shared_region.h>
 
-#if CONFIG_MACF_MACH
+#include <libkern/OSDebug.h>
+#include <libkern/OSAtomic.h>
+#include <libkern/section_keywords.h>
+
+#include <mach-o/loader.h>
+#include <kdp/kdp_dyld.h>
+
+#include <kern/sfi.h>           /* picks up ledger.h */
+
+#if CONFIG_MACF
 #include <security/mac_mach_internal.h>
 #endif
 
-#if CONFIG_COUNTERS
-#include <pmc/pmc.h>
-#endif /* CONFIG_COUNTERS */
+#include <IOKit/IOBSD.h>
+
+#if KPERF
+extern int kpc_force_all_ctrs(task_t, int);
+#endif
+
+SECURITY_READ_ONLY_LATE(task_t) kernel_task;
+
+static SECURITY_READ_ONLY_LATE(zone_t) task_zone;
+ZONE_INIT(&task_zone, "tasks", sizeof(struct task),
+    ZC_NOENCRYPT | ZC_ZFREE_CLEARMEM,
+    ZONE_ID_TASK, NULL);
+
+extern int exc_via_corpse_forking;
+extern int corpse_for_fatal_memkill;
+extern boolean_t proc_send_synchronous_EXC_RESOURCE(void *p);
+extern void task_disown_frozen_csegs(task_t owner_task);
+
+/* Flag set by core audio when audio is playing. Used to stifle EXC_RESOURCE generation when active. */
+int audio_active = 0;
+
+/*
+ *     structure for tracking zone usage
+ *     Used either one per task/thread for all zones or <per-task,per-zone>.
+ */
+typedef struct zinfo_usage_store_t {
+       /* These fields may be updated atomically, and so must be 8 byte aligned */
+       uint64_t        alloc __attribute__((aligned(8)));              /* allocation counter */
+       uint64_t        free __attribute__((aligned(8)));               /* free counter */
+} zinfo_usage_store_t;
+
+zinfo_usage_store_t tasks_tkm_private;
+zinfo_usage_store_t tasks_tkm_shared;
+
+/* A container to accumulate statistics for expired tasks */
+expired_task_statistics_t               dead_task_statistics;
+LCK_SPIN_DECLARE_ATTR(dead_task_statistics_lock, &task_lck_grp, &task_lck_attr);
+
+ledger_template_t task_ledger_template = NULL;
+
+/* global lock for task_dyld_process_info_notify_{register, deregister, get_trap} */
+LCK_GRP_DECLARE(g_dyldinfo_mtx_grp, "g_dyldinfo");
+LCK_MTX_DECLARE(g_dyldinfo_mtx, &g_dyldinfo_mtx_grp);
+
+SECURITY_READ_ONLY_LATE(struct _task_ledger_indices) task_ledgers __attribute__((used)) =
+{.cpu_time = -1,
+ .tkm_private = -1,
+ .tkm_shared = -1,
+ .phys_mem = -1,
+ .wired_mem = -1,
+ .internal = -1,
+ .iokit_mapped = -1,
+ .alternate_accounting = -1,
+ .alternate_accounting_compressed = -1,
+ .page_table = -1,
+ .phys_footprint = -1,
+ .internal_compressed = -1,
+ .purgeable_volatile = -1,
+ .purgeable_nonvolatile = -1,
+ .purgeable_volatile_compressed = -1,
+ .purgeable_nonvolatile_compressed = -1,
+ .tagged_nofootprint = -1,
+ .tagged_footprint = -1,
+ .tagged_nofootprint_compressed = -1,
+ .tagged_footprint_compressed = -1,
+ .network_volatile = -1,
+ .network_nonvolatile = -1,
+ .network_volatile_compressed = -1,
+ .network_nonvolatile_compressed = -1,
+ .media_nofootprint = -1,
+ .media_footprint = -1,
+ .media_nofootprint_compressed = -1,
+ .media_footprint_compressed = -1,
+ .graphics_nofootprint = -1,
+ .graphics_footprint = -1,
+ .graphics_nofootprint_compressed = -1,
+ .graphics_footprint_compressed = -1,
+ .neural_nofootprint = -1,
+ .neural_footprint = -1,
+ .neural_nofootprint_compressed = -1,
+ .neural_footprint_compressed = -1,
+ .platform_idle_wakeups = -1,
+ .interrupt_wakeups = -1,
+#if CONFIG_SCHED_SFI
+ .sfi_wait_times = { 0 /* initialized at runtime */},
+#endif /* CONFIG_SCHED_SFI */
+ .cpu_time_billed_to_me = -1,
+ .cpu_time_billed_to_others = -1,
+ .physical_writes = -1,
+ .logical_writes = -1,
+ .logical_writes_to_external = -1,
+#if DEBUG || DEVELOPMENT
+ .pages_grabbed = -1,
+ .pages_grabbed_kern = -1,
+ .pages_grabbed_iopl = -1,
+ .pages_grabbed_upl = -1,
+#endif
+#if CONFIG_FREEZE
+ .frozen_to_swap = -1,
+#endif /* CONFIG_FREEZE */
+ .energy_billed_to_me = -1,
+ .energy_billed_to_others = -1,
+#if CONFIG_PHYS_WRITE_ACCT
+ .fs_metadata_writes = -1,
+#endif /* CONFIG_PHYS_WRITE_ACCT */
+};
+
+/* System sleep state */
+boolean_t tasks_suspend_state;
 
-task_t                 kernel_task;
-zone_t                 task_zone;
-lck_attr_t      task_lck_attr;
-lck_grp_t       task_lck_grp;
-lck_grp_attr_t  task_lck_grp_attr;
+
+void init_task_ledgers(void);
+void task_footprint_exceeded(int warning, __unused const void *param0, __unused const void *param1);
+void task_wakeups_rate_exceeded(int warning, __unused const void *param0, __unused const void *param1);
+void task_io_rate_exceeded(int warning, const void *param0, __unused const void *param1);
+void __attribute__((noinline)) SENDING_NOTIFICATION__THIS_PROCESS_IS_CAUSING_TOO_MANY_WAKEUPS(void);
+void __attribute__((noinline)) PROC_CROSSED_HIGH_WATERMARK__SEND_EXC_RESOURCE_AND_SUSPEND(int max_footprint_mb, boolean_t is_fatal);
+void __attribute__((noinline)) SENDING_NOTIFICATION__THIS_PROCESS_IS_CAUSING_TOO_MUCH_IO(int flavor);
+
+kern_return_t task_suspend_internal(task_t);
+kern_return_t task_resume_internal(task_t);
+static kern_return_t task_start_halt_locked(task_t task, boolean_t should_mark_corpse);
+
+extern kern_return_t iokit_task_terminate(task_t task);
+extern void          iokit_task_app_suspended_changed(task_t task);
+
+extern kern_return_t exception_deliver(thread_t, exception_type_t, mach_exception_data_t, mach_msg_type_number_t, struct exception_action *, lck_mtx_t *);
+extern void bsd_copythreadname(void *dst_uth, void *src_uth);
+extern kern_return_t thread_resume(thread_t thread);
+
+// Warn tasks when they hit 80% of their memory limit.
+#define PHYS_FOOTPRINT_WARNING_LEVEL 80
+
+#define TASK_WAKEUPS_MONITOR_DEFAULT_LIMIT              150 /* wakeups per second */
+#define TASK_WAKEUPS_MONITOR_DEFAULT_INTERVAL   300 /* in seconds. */
+
+/*
+ * Level (in terms of percentage of the limit) at which the wakeups monitor triggers telemetry.
+ *
+ * (ie when the task's wakeups rate exceeds 70% of the limit, start taking user
+ *  stacktraces, aka micro-stackshots)
+ */
+#define TASK_WAKEUPS_MONITOR_DEFAULT_USTACKSHOTS_TRIGGER        70
+
+int task_wakeups_monitor_interval; /* In seconds. Time period over which wakeups rate is observed */
+int task_wakeups_monitor_rate;     /* In hz. Maximum allowable wakeups per task before EXC_RESOURCE is sent */
+
+unsigned int task_wakeups_monitor_ustackshots_trigger_pct; /* Percentage. Level at which we start gathering telemetry. */
+
+int disable_exc_resource; /* Global override to supress EXC_RESOURCE for resource monitor violations. */
+
+ledger_amount_t max_task_footprint = 0;  /* Per-task limit on physical memory consumption in bytes     */
+unsigned int max_task_footprint_warning_level = 0;  /* Per-task limit warning percentage */
+int max_task_footprint_mb = 0;  /* Per-task limit on physical memory consumption in megabytes */
+
+/* I/O Monitor Limits */
+#define IOMON_DEFAULT_LIMIT                     (20480ull)      /* MB of logical/physical I/O */
+#define IOMON_DEFAULT_INTERVAL                  (86400ull)      /* in seconds */
+
+uint64_t task_iomon_limit_mb;           /* Per-task I/O monitor limit in MBs */
+uint64_t task_iomon_interval_secs;      /* Per-task I/O monitor interval in secs */
+
+#define IO_TELEMETRY_DEFAULT_LIMIT              (10ll * 1024ll * 1024ll)
+int64_t io_telemetry_limit;                     /* Threshold to take a microstackshot (0 indicated I/O telemetry is turned off) */
+int64_t global_logical_writes_count = 0;        /* Global count for logical writes */
+int64_t global_logical_writes_to_external_count = 0;        /* Global count for logical writes to external storage*/
+static boolean_t global_update_logical_writes(int64_t, int64_t*);
+
+#define TASK_MAX_THREAD_LIMIT 256
+
+#if MACH_ASSERT
+int pmap_ledgers_panic = 1;
+int pmap_ledgers_panic_leeway = 3;
+#endif /* MACH_ASSERT */
 
 int task_max = CONFIG_TASK_MAX; /* Max number of tasks */
 
-/* Forwards */
+#if CONFIG_COREDUMP
+int hwm_user_cores = 0; /* high watermark violations generate user core files */
+#endif
+
+#ifdef MACH_BSD
+extern uint32_t proc_platform(const struct proc *);
+extern uint32_t proc_min_sdk(struct proc *);
+extern void     proc_getexecutableuuid(void *, unsigned char *, unsigned long);
+extern int      proc_pid(struct proc *p);
+extern int      proc_selfpid(void);
+extern struct proc *current_proc(void);
+extern char     *proc_name_address(struct proc *p);
+extern uint64_t get_dispatchqueue_offset_from_proc(void *);
+extern int kevent_proc_copy_uptrs(void *proc, uint64_t *buf, uint32_t bufsize);
+extern void workq_proc_suspended(struct proc *p);
+extern void workq_proc_resumed(struct proc *p);
+
+#if CONFIG_MEMORYSTATUS
+extern void     proc_memstat_terminated(struct proc* p, boolean_t set);
+extern void     memorystatus_on_ledger_footprint_exceeded(int warning, boolean_t memlimit_is_active, boolean_t memlimit_is_fatal);
+extern void     memorystatus_log_exception(const int max_footprint_mb, boolean_t memlimit_is_active, boolean_t memlimit_is_fatal);
+extern boolean_t memorystatus_allowed_vm_map_fork(task_t task);
+extern uint64_t  memorystatus_available_memory_internal(struct proc *p);
+
+#if DEVELOPMENT || DEBUG
+extern void memorystatus_abort_vm_map_fork(task_t);
+#endif
 
-void           task_hold_locked(
-                       task_t          task);
-void           task_wait_locked(
-                       task_t          task);
-void           task_release_locked(
-                       task_t          task);
-void           task_free(
-                       task_t          task );
-void           task_synchronizer_destroy_all(
-                       task_t          task);
+#endif /* CONFIG_MEMORYSTATUS */
 
-kern_return_t  task_set_ledger(
-                       task_t          task,
-                       ledger_t        wired,
-                       ledger_t        paged);
+#endif /* MACH_BSD */
 
-int check_for_tasksuspend(
-                       task_t task);
+#if DEVELOPMENT || DEBUG
+int exc_resource_threads_enabled;
+#endif /* DEVELOPMENT || DEBUG */
 
-void
-task_backing_store_privileged(
-                       task_t task)
-{
-       task_lock(task);
-       task->priv_flags |= VM_BACKING_STORE_PRIV;
-       task_unlock(task);
-       return;
-}
+#if (DEVELOPMENT || DEBUG)
+uint32_t task_exc_guard_default = TASK_EXC_GUARD_MP_DELIVER | TASK_EXC_GUARD_MP_ONCE | TASK_EXC_GUARD_MP_CORPSE |
+    TASK_EXC_GUARD_VM_DELIVER | TASK_EXC_GUARD_VM_ONCE | TASK_EXC_GUARD_VM_CORPSE;
+#else
+uint32_t task_exc_guard_default = 0;
+#endif
 
+/* Forwards */
+
+static void task_hold_locked(task_t task);
+static void task_wait_locked(task_t task, boolean_t until_not_runnable);
+static void task_release_locked(task_t task);
+
+static void task_synchronizer_destroy_all(task_t task);
+static os_ref_count_t
+task_add_turnstile_watchports_locked(
+       task_t                      task,
+       struct task_watchports      *watchports,
+       struct task_watchport_elem  **previous_elem_array,
+       ipc_port_t                  *portwatch_ports,
+       uint32_t                    portwatch_count);
+
+static os_ref_count_t
+task_remove_turnstile_watchports_locked(
+       task_t                 task,
+       struct task_watchports *watchports,
+       ipc_port_t             *port_freelist);
+
+static struct task_watchports *
+task_watchports_alloc_init(
+       task_t        task,
+       thread_t      thread,
+       uint32_t      count);
+
+static void
+task_watchports_deallocate(
+       struct task_watchports *watchports);
 
 void
 task_set_64bit(
-               task_t task,
-               boolean_t is64bit)
+       task_t task,
+       boolean_t is_64bit,
+       boolean_t is_64bit_data)
 {
-#if defined(__i386__) || defined(__x86_64__)
+#if defined(__i386__) || defined(__x86_64__) || defined(__arm64__)
        thread_t thread;
-#endif /* __i386__ */
-       int     vm_flags = 0;
+#endif /* defined(__i386__) || defined(__x86_64__) || defined(__arm64__) */
 
-       if (is64bit) {
-               if (task_has_64BitAddr(task))
-                       return;
+       task_lock(task);
 
-               task_set_64BitAddr(task);
+       /*
+        * Switching to/from 64-bit address spaces
+        */
+       if (is_64bit) {
+               if (!task_has_64Bit_addr(task)) {
+                       task_set_64Bit_addr(task);
+               }
        } else {
-               if ( !task_has_64BitAddr(task))
-                       return;
+               if (task_has_64Bit_addr(task)) {
+                       task_clear_64Bit_addr(task);
+               }
+       }
 
-               /*
-                * Deallocate all memory previously allocated
-                * above the 32-bit address space, since it won't
-                * be accessible anymore.
-                */
-               /* remove regular VM map entries & pmap mappings */
-               (void) vm_map_remove(task->map,
-                                    (vm_map_offset_t) VM_MAX_ADDRESS,
-                                    MACH_VM_MAX_ADDRESS,
-                                    0);
-#ifdef __ppc__
-               /*
-                * PPC51: ppc64 is limited to 51-bit addresses.
-                * Memory mapped above that limit is handled specially
-                * at the pmap level, so let pmap clean the commpage mapping
-                * explicitly...
-                */
-               pmap_unmap_sharedpage(task->map->pmap); /* Unmap commpage */
-               /* ... and avoid regular pmap cleanup */
-               vm_flags |= VM_MAP_REMOVE_NO_PMAP_CLEANUP;
-#endif /* __ppc__ */
-               /* remove the higher VM mappings */
-               (void) vm_map_remove(task->map,
-                                    MACH_VM_MAX_ADDRESS,
-                                    0xFFFFFFFFFFFFF000ULL,
-                                    vm_flags);
-               task_clear_64BitAddr(task);
+       /*
+        * Switching to/from 64-bit register state.
+        */
+       if (is_64bit_data) {
+               if (task_has_64Bit_data(task)) {
+                       goto out;
+               }
+
+               task_set_64Bit_data(task);
+       } else {
+               if (!task_has_64Bit_data(task)) {
+                       goto out;
+               }
+
+               task_clear_64Bit_data(task);
        }
+
        /* FIXME: On x86, the thread save state flavor can diverge from the
         * task's 64-bit feature flag due to the 32-bit/64-bit register save
         * state dichotomy. Since we can be pre-empted in this interval,
         * certain routines may observe the thread as being in an inconsistent
         * state with respect to its task's 64-bitness.
         */
-#if defined(__i386__) || defined(__x86_64__)
-       task_lock(task);
+
+#if defined(__x86_64__) || defined(__arm64__)
        queue_iterate(&task->threads, thread, thread_t, task_threads) {
                thread_mtx_lock(thread);
                machine_thread_switch_addrmode(thread);
                thread_mtx_unlock(thread);
        }
+#endif /* defined(__x86_64__) || defined(__arm64__) */
+
+out:
+       task_unlock(task);
+}
+
+boolean_t
+task_get_64bit_data(task_t task)
+{
+       return task_has_64Bit_data(task);
+}
+
+void
+task_set_platform_binary(
+       task_t task,
+       boolean_t is_platform)
+{
+       task_lock(task);
+       if (is_platform) {
+               task->t_flags |= TF_PLATFORM;
+               /* set exc guard default behavior for first-party code */
+               task->task_exc_guard = (task_exc_guard_default & TASK_EXC_GUARD_ALL);
+       } else {
+               task->t_flags &= ~(TF_PLATFORM);
+               /* set exc guard default behavior for third-party code */
+               task->task_exc_guard = ((task_exc_guard_default >> TASK_EXC_GUARD_THIRD_PARTY_DEFAULT_SHIFT) & TASK_EXC_GUARD_ALL);
+       }
        task_unlock(task);
-#endif /* __i386__ */
 }
 
+/*
+ * Set or clear per-task TF_CA_CLIENT_WI flag according to specified argument.
+ * Returns "false" if flag is already set, and "true" in other cases.
+ */
+bool
+task_set_ca_client_wi(
+       task_t task,
+       boolean_t set_or_clear)
+{
+       bool ret = true;
+       task_lock(task);
+       if (set_or_clear) {
+               /* Tasks can have only one CA_CLIENT work interval */
+               if (task->t_flags & TF_CA_CLIENT_WI) {
+                       ret = false;
+               } else {
+                       task->t_flags |= TF_CA_CLIENT_WI;
+               }
+       } else {
+               task->t_flags &= ~TF_CA_CLIENT_WI;
+       }
+       task_unlock(task);
+       return ret;
+}
 
 void
-task_set_dyld_info(task_t task, mach_vm_address_t addr, mach_vm_size_t size)
+task_set_dyld_info(
+       task_t task,
+       mach_vm_address_t addr,
+       mach_vm_size_t size)
 {
        task_lock(task);
        task->all_image_info_addr = addr;
@@ -272,1733 +548,7815 @@ task_set_dyld_info(task_t task, mach_vm_address_t addr, mach_vm_size_t size)
 }
 
 void
-task_init(void)
+task_set_mach_header_address(
+       task_t task,
+       mach_vm_address_t addr)
 {
+       task_lock(task);
+       task->mach_header_vm_address = addr;
+       task_unlock(task);
+}
 
-       lck_grp_attr_setdefault(&task_lck_grp_attr);
-       lck_grp_init(&task_lck_grp, "task", &task_lck_grp_attr);
-       lck_attr_setdefault(&task_lck_attr);
-       lck_mtx_init(&tasks_threads_lock, &task_lck_grp, &task_lck_attr);
-
-       task_zone = zinit(
-                       sizeof(struct task),
-                       task_max * sizeof(struct task),
-                       TASK_CHUNK * sizeof(struct task),
-                       "tasks");
-       zone_change(task_zone, Z_NOENCRYPT, TRUE);
-
-       /*
-        * Create the kernel task as the first task.
-        */
-#ifdef __LP64__
-       if (task_create_internal(TASK_NULL, FALSE, TRUE, &kernel_task) != KERN_SUCCESS)
-#else
-       if (task_create_internal(TASK_NULL, FALSE, FALSE, &kernel_task) != KERN_SUCCESS)
-#endif
-               panic("task_init\n");
-
-       vm_map_deallocate(kernel_task->map);
-       kernel_task->map = kernel_map;
+void
+task_bank_reset(__unused task_t task)
+{
+       if (task->bank_context != NULL) {
+               bank_task_destroy(task);
+       }
 }
 
 /*
- * Create a task running in the kernel address space.  It may
- * have its own map of size mem_size and may have ipc privileges.
+ * NOTE: This should only be called when the P_LINTRANSIT
+ *      flag is set (the proc_trans lock is held) on the
+ *      proc associated with the task.
  */
-kern_return_t
-kernel_task_create(
-       __unused task_t         parent_task,
-       __unused vm_offset_t            map_base,
-       __unused vm_size_t              map_size,
-       __unused task_t         *child_task)
+void
+task_bank_init(__unused task_t task)
 {
-       return (KERN_INVALID_ARGUMENT);
+       if (task->bank_context != NULL) {
+               panic("Task bank init called with non null bank context for task: %p and bank_context: %p", task, task->bank_context);
+       }
+       bank_task_initialize(task);
 }
 
-kern_return_t
-task_create(
-       task_t                          parent_task,
-       __unused ledger_port_array_t    ledger_ports,
-       __unused mach_msg_type_number_t num_ledger_ports,
-       __unused boolean_t              inherit_memory,
-       __unused task_t                 *child_task)    /* OUT */
+void
+task_set_did_exec_flag(task_t task)
 {
-       if (parent_task == TASK_NULL)
-               return(KERN_INVALID_ARGUMENT);
-
-       /*
-        * No longer supported: too many calls assume that a task has a valid
-        * process attached.
-        */
-       return(KERN_FAILURE);
+       task->t_procflags |= TPF_DID_EXEC;
 }
 
-kern_return_t
-host_security_create_task_token(
-       host_security_t                 host_security,
-       task_t                          parent_task,
-       __unused security_token_t       sec_token,
-       __unused audit_token_t          audit_token,
-       __unused host_priv_t            host_priv,
-       __unused ledger_port_array_t    ledger_ports,
-       __unused mach_msg_type_number_t num_ledger_ports,
-       __unused boolean_t              inherit_memory,
-       __unused task_t                 *child_task)    /* OUT */
-{
-       if (parent_task == TASK_NULL)
-               return(KERN_INVALID_ARGUMENT);
-
-       if (host_security == HOST_NULL)
-               return(KERN_INVALID_SECURITY);
-
-       /*
-        * No longer supported.
-        */
-       return(KERN_FAILURE);
+void
+task_clear_exec_copy_flag(task_t task)
+{
+       task->t_procflags &= ~TPF_EXEC_COPY;
 }
 
-kern_return_t
-task_create_internal(
-       task_t          parent_task,
-       boolean_t       inherit_memory,
-       boolean_t       is_64bit,
-       task_t          *child_task)            /* OUT */
+event_t
+task_get_return_wait_event(task_t task)
 {
-       task_t                  new_task;
-       vm_shared_region_t      shared_region;
-
-       new_task = (task_t) zalloc(task_zone);
-
-       if (new_task == TASK_NULL)
-               return(KERN_RESOURCE_SHORTAGE);
+       return (event_t)&task->returnwait_inheritor;
+}
 
-       /* one ref for just being alive; one for our caller */
-       new_task->ref_count = 2;
+void
+task_clear_return_wait(task_t task, uint32_t flags)
+{
+       if (flags & TCRW_CLEAR_INITIAL_WAIT) {
+               thread_wakeup(task_get_return_wait_event(task));
+       }
 
-       /* if inherit_memory is true, parent_task MUST not be NULL */
-       if (inherit_memory)
-               new_task->map = vm_map_fork(parent_task->map);
-       else
-               new_task->map = vm_map_create(pmap_create(0, is_64bit),
-                                       (vm_map_offset_t)(VM_MIN_ADDRESS),
-                                       (vm_map_offset_t)(VM_MAX_ADDRESS), TRUE);
+       if (flags & TCRW_CLEAR_FINAL_WAIT) {
+               is_write_lock(task->itk_space);
 
-       /* Inherit memlock limit from parent */
-       if (parent_task)
-               vm_map_set_user_wire_limit(new_task->map, (vm_size_t)parent_task->map->user_wire_limit);
+               task->t_returnwaitflags &= ~TRW_LRETURNWAIT;
+               task->returnwait_inheritor = NULL;
 
-       lck_mtx_init(&new_task->lock, &task_lck_grp, &task_lck_attr);
-       queue_init(&new_task->threads);
-       new_task->suspend_count = 0;
-       new_task->thread_count = 0;
-       new_task->active_thread_count = 0;
-       new_task->user_stop_count = 0;
-       new_task->role = TASK_UNSPECIFIED;
-       new_task->active = TRUE;
-       new_task->halting = FALSE;
-       new_task->user_data = NULL;
-       new_task->faults = 0;
-       new_task->cow_faults = 0;
-       new_task->pageins = 0;
-       new_task->messages_sent = 0;
-       new_task->messages_received = 0;
-       new_task->syscalls_mach = 0;
-       new_task->priv_flags = 0;
-       new_task->syscalls_unix=0;
-       new_task->c_switch = new_task->p_switch = new_task->ps_switch = 0;
-       new_task->taskFeatures[0] = 0;                          /* Init task features */
-       new_task->taskFeatures[1] = 0;                          /* Init task features */
+               if (task->t_returnwaitflags & TRW_LRETURNWAITER) {
+                       struct turnstile *turnstile = turnstile_prepare((uintptr_t) task_get_return_wait_event(task),
+                           NULL, TURNSTILE_NULL, TURNSTILE_ULOCK);
 
-#ifdef MACH_BSD
-       new_task->bsd_info = NULL;
-#endif /* MACH_BSD */
+                       waitq_wakeup64_all(&turnstile->ts_waitq,
+                           CAST_EVENT64_T(task_get_return_wait_event(task)),
+                           THREAD_AWAKENED, 0);
 
-#if defined(__i386__) || defined(__x86_64__)
-       new_task->i386_ldt = 0;
-       new_task->task_debug = NULL;
+                       turnstile_update_inheritor(turnstile, NULL,
+                           TURNSTILE_IMMEDIATE_UPDATE | TURNSTILE_INHERITOR_THREAD);
+                       turnstile_update_inheritor_complete(turnstile, TURNSTILE_INTERLOCK_HELD);
 
-#endif
+                       turnstile_complete((uintptr_t) task_get_return_wait_event(task), NULL, NULL, TURNSTILE_ULOCK);
+                       turnstile_cleanup();
+                       task->t_returnwaitflags &= ~TRW_LRETURNWAITER;
+               }
+               is_write_unlock(task->itk_space);
+       }
+}
 
-#ifdef __ppc__
-       if(BootProcInfo.pf.Available & pf64Bit) new_task->taskFeatures[0] |= tf64BitData;       /* If 64-bit machine, show we have 64-bit registers at least */
-#endif
+void __attribute__((noreturn))
+task_wait_to_return(void)
+{
+       task_t task = current_task();
 
-       queue_init(&new_task->semaphore_list);
-       queue_init(&new_task->lock_set_list);
-       new_task->semaphores_owned = 0;
-       new_task->lock_sets_owned = 0;
+       is_write_lock(task->itk_space);
 
-#if CONFIG_MACF_MACH
-       new_task->label = labelh_new(1);
-       mac_task_label_init (&new_task->maclabel);
-#endif
+       if (task->t_returnwaitflags & TRW_LRETURNWAIT) {
+               struct turnstile *turnstile = turnstile_prepare((uintptr_t) task_get_return_wait_event(task),
+                   NULL, TURNSTILE_NULL, TURNSTILE_ULOCK);
 
-       ipc_task_init(new_task, parent_task);
+               do {
+                       task->t_returnwaitflags |= TRW_LRETURNWAITER;
+                       turnstile_update_inheritor(turnstile, task->returnwait_inheritor,
+                           (TURNSTILE_DELAYED_UPDATE | TURNSTILE_INHERITOR_THREAD));
 
-       new_task->total_user_time = 0;
-       new_task->total_system_time = 0;
+                       waitq_assert_wait64(&turnstile->ts_waitq,
+                           CAST_EVENT64_T(task_get_return_wait_event(task)),
+                           THREAD_UNINT, TIMEOUT_WAIT_FOREVER);
 
-       new_task->vtimers = 0;
+                       is_write_unlock(task->itk_space);
 
-       new_task->shared_region = NULL;
+                       turnstile_update_inheritor_complete(turnstile, TURNSTILE_INTERLOCK_NOT_HELD);
 
-       new_task->affinity_space = NULL;
+                       thread_block(THREAD_CONTINUE_NULL);
 
-#if CONFIG_COUNTERS
-       new_task->t_chud = 0U;
-#endif
+                       is_write_lock(task->itk_space);
+               } while (task->t_returnwaitflags & TRW_LRETURNWAIT);
 
-       if (parent_task != TASK_NULL) {
-               new_task->sec_token = parent_task->sec_token;
-               new_task->audit_token = parent_task->audit_token;
+               turnstile_complete((uintptr_t) task_get_return_wait_event(task), NULL, NULL, TURNSTILE_ULOCK);
+       }
 
-               /* inherit the parent's shared region */
-               shared_region = vm_shared_region_get(parent_task);
-               vm_shared_region_set(new_task, shared_region);
+       is_write_unlock(task->itk_space);
+       turnstile_cleanup();
 
-               new_task->wired_ledger_port = ledger_copy(
-                       convert_port_to_ledger(parent_task->wired_ledger_port));
-               new_task->paged_ledger_port = ledger_copy(
-                       convert_port_to_ledger(parent_task->paged_ledger_port));
-               if(task_has_64BitAddr(parent_task))
-                       task_set_64BitAddr(new_task);
-               new_task->all_image_info_addr = parent_task->all_image_info_addr;
-               new_task->all_image_info_size = parent_task->all_image_info_size;
 
-#if defined(__i386__) || defined(__x86_64__)
-               if (inherit_memory && parent_task->i386_ldt)
-                       new_task->i386_ldt = user_ldt_copy(parent_task->i386_ldt);
+#if CONFIG_MACF
+       /*
+        * Before jumping to userspace and allowing this process to execute any code,
+        * notify any interested parties.
+        */
+       mac_proc_notify_exec_complete(current_proc());
 #endif
-               if (inherit_memory && parent_task->affinity_space)
-                       task_affinity_create(parent_task, new_task);
 
-               new_task->pset_hint = parent_task->pset_hint = task_choose_pset(parent_task);
-       }
-       else {
-               new_task->sec_token = KERNEL_SECURITY_TOKEN;
-               new_task->audit_token = KERNEL_AUDIT_TOKEN;
-               new_task->wired_ledger_port = ledger_copy(root_wired_ledger);
-               new_task->paged_ledger_port = ledger_copy(root_paged_ledger);
-#ifdef __LP64__
-               if(is_64bit)
-                       task_set_64BitAddr(new_task);
-#endif
+       thread_bootstrap_return();
+}
 
-               new_task->pset_hint = PROCESSOR_SET_NULL;
+#ifdef CONFIG_32BIT_TELEMETRY
+boolean_t
+task_consume_32bit_log_flag(task_t task)
+{
+       if ((task->t_procflags & TPF_LOG_32BIT_TELEMETRY) != 0) {
+               task->t_procflags &= ~TPF_LOG_32BIT_TELEMETRY;
+               return TRUE;
+       } else {
+               return FALSE;
        }
+}
 
-       if (kernel_task == TASK_NULL) {
-               new_task->priority = BASEPRI_KERNEL;
-               new_task->max_priority = MAXPRI_KERNEL;
-       }
-       else {
-               new_task->priority = BASEPRI_DEFAULT;
-               new_task->max_priority = MAXPRI_USER;
-       }
-       
-       lck_mtx_lock(&tasks_threads_lock);
-       queue_enter(&tasks, new_task, task_t, tasks);
-       tasks_count++;
-       lck_mtx_unlock(&tasks_threads_lock);
+void
+task_set_32bit_log_flag(task_t task)
+{
+       task->t_procflags |= TPF_LOG_32BIT_TELEMETRY;
+}
+#endif /* CONFIG_32BIT_TELEMETRY */
 
-       if (vm_backing_store_low && parent_task != NULL)
-               new_task->priv_flags |= (parent_task->priv_flags&VM_BACKING_STORE_PRIV);
+boolean_t
+task_is_exec_copy(task_t task)
+{
+       return task_is_exec_copy_internal(task);
+}
 
-       ipc_task_enable(new_task);
+boolean_t
+task_did_exec(task_t task)
+{
+       return task_did_exec_internal(task);
+}
 
-       *child_task = new_task;
-       return(KERN_SUCCESS);
+boolean_t
+task_is_active(task_t task)
+{
+       return task->active;
 }
 
-/*
- *     task_deallocate:
- *
- *     Drop a reference on a task.
- */
-void
-task_deallocate(
-       task_t          task)
+boolean_t
+task_is_halting(task_t task)
 {
-       if (task == TASK_NULL)
-           return;
+       return task->halting;
+}
 
-       if (task_deallocate_internal(task) > 0)
-               return;
+#if TASK_REFERENCE_LEAK_DEBUG
+#include <kern/btlog.h>
 
-       ipc_task_terminate(task);
+static btlog_t *task_ref_btlog;
+#define TASK_REF_OP_INCR        0x1
+#define TASK_REF_OP_DECR        0x2
 
-       if (task->affinity_space)
-               task_affinity_deallocate(task);
+#define TASK_REF_NUM_RECORDS    100000
+#define TASK_REF_BTDEPTH        7
 
-       vm_map_deallocate(task->map);
-       is_release(task->itk_space);
+void
+task_reference_internal(task_t task)
+{
+       void *       bt[TASK_REF_BTDEPTH];
+       int             numsaved = 0;
 
-       lck_mtx_destroy(&task->lock, &task_lck_grp);
+       task_require(task);
+       os_ref_retain(&task->ref_count);
 
-#if CONFIG_MACF_MACH
-       labelh_release(task->label);
-#endif
-       zfree(task_zone, task);
+       numsaved = OSBacktrace(bt, TASK_REF_BTDEPTH);
+       btlog_add_entry(task_ref_btlog, task, TASK_REF_OP_INCR,
+           bt, numsaved);
 }
 
-/*
- *     task_name_deallocate:
- *
- *     Drop a reference on a task name.
- */
-void
-task_name_deallocate(
-       task_name_t             task_name)
+os_ref_count_t
+task_deallocate_internal(task_t task)
 {
-       return(task_deallocate((task_t)task_name));
-}
+       void *       bt[TASK_REF_BTDEPTH];
+       int             numsaved = 0;
 
+       numsaved = OSBacktrace(bt, TASK_REF_BTDEPTH);
+       btlog_add_entry(task_ref_btlog, task, TASK_REF_OP_DECR,
+           bt, numsaved);
+
+       return os_ref_release(&task->ref_count);
+}
+
+#endif /* TASK_REFERENCE_LEAK_DEBUG */
+
+void
+task_init(void)
+{
+       /*
+        * Configure per-task memory limit.
+        * The boot-arg is interpreted as Megabytes,
+        * and takes precedence over the device tree.
+        * Setting the boot-arg to 0 disables task limits.
+        */
+       if (!PE_parse_boot_argn("max_task_pmem", &max_task_footprint_mb,
+           sizeof(max_task_footprint_mb))) {
+               /*
+                * No limit was found in boot-args, so go look in the device tree.
+                */
+               if (!PE_get_default("kern.max_task_pmem", &max_task_footprint_mb,
+                   sizeof(max_task_footprint_mb))) {
+                       /*
+                        * No limit was found in device tree.
+                        */
+                       max_task_footprint_mb = 0;
+               }
+       }
+
+       if (max_task_footprint_mb != 0) {
+#if CONFIG_MEMORYSTATUS
+               if (max_task_footprint_mb < 50) {
+                       printf("Warning: max_task_pmem %d below minimum.\n",
+                           max_task_footprint_mb);
+                       max_task_footprint_mb = 50;
+               }
+               printf("Limiting task physical memory footprint to %d MB\n",
+                   max_task_footprint_mb);
+
+               max_task_footprint = (ledger_amount_t)max_task_footprint_mb * 1024 * 1024; // Convert MB to bytes
+
+               /*
+                * Configure the per-task memory limit warning level.
+                * This is computed as a percentage.
+                */
+               max_task_footprint_warning_level = 0;
+
+               if (max_mem < 0x40000000) {
+                       /*
+                        * On devices with < 1GB of memory:
+                        *    -- set warnings to 50MB below the per-task limit.
+                        */
+                       if (max_task_footprint_mb > 50) {
+                               max_task_footprint_warning_level = ((max_task_footprint_mb - 50) * 100) / max_task_footprint_mb;
+                       }
+               } else {
+                       /*
+                        * On devices with >= 1GB of memory:
+                        *    -- set warnings to 100MB below the per-task limit.
+                        */
+                       if (max_task_footprint_mb > 100) {
+                               max_task_footprint_warning_level = ((max_task_footprint_mb - 100) * 100) / max_task_footprint_mb;
+                       }
+               }
+
+               /*
+                * Never allow warning level to land below the default.
+                */
+               if (max_task_footprint_warning_level < PHYS_FOOTPRINT_WARNING_LEVEL) {
+                       max_task_footprint_warning_level = PHYS_FOOTPRINT_WARNING_LEVEL;
+               }
+
+               printf("Limiting task physical memory warning to %d%%\n", max_task_footprint_warning_level);
+
+#else
+               printf("Warning: max_task_pmem specified, but jetsam not configured; ignoring.\n");
+#endif /* CONFIG_MEMORYSTATUS */
+       }
+
+#if DEVELOPMENT || DEBUG
+       if (!PE_parse_boot_argn("exc_resource_threads",
+           &exc_resource_threads_enabled,
+           sizeof(exc_resource_threads_enabled))) {
+               exc_resource_threads_enabled = 1;
+       }
+       PE_parse_boot_argn("task_exc_guard_default",
+           &task_exc_guard_default,
+           sizeof(task_exc_guard_default));
+#endif /* DEVELOPMENT || DEBUG */
+
+#if CONFIG_COREDUMP
+       if (!PE_parse_boot_argn("hwm_user_cores", &hwm_user_cores,
+           sizeof(hwm_user_cores))) {
+               hwm_user_cores = 0;
+       }
+#endif
+
+       proc_init_cpumon_params();
+
+       if (!PE_parse_boot_argn("task_wakeups_monitor_rate", &task_wakeups_monitor_rate, sizeof(task_wakeups_monitor_rate))) {
+               task_wakeups_monitor_rate = TASK_WAKEUPS_MONITOR_DEFAULT_LIMIT;
+       }
+
+       if (!PE_parse_boot_argn("task_wakeups_monitor_interval", &task_wakeups_monitor_interval, sizeof(task_wakeups_monitor_interval))) {
+               task_wakeups_monitor_interval = TASK_WAKEUPS_MONITOR_DEFAULT_INTERVAL;
+       }
+
+       if (!PE_parse_boot_argn("task_wakeups_monitor_ustackshots_trigger_pct", &task_wakeups_monitor_ustackshots_trigger_pct,
+           sizeof(task_wakeups_monitor_ustackshots_trigger_pct))) {
+               task_wakeups_monitor_ustackshots_trigger_pct = TASK_WAKEUPS_MONITOR_DEFAULT_USTACKSHOTS_TRIGGER;
+       }
+
+       if (!PE_parse_boot_argn("disable_exc_resource", &disable_exc_resource,
+           sizeof(disable_exc_resource))) {
+               disable_exc_resource = 0;
+       }
+
+       if (!PE_parse_boot_argn("task_iomon_limit_mb", &task_iomon_limit_mb, sizeof(task_iomon_limit_mb))) {
+               task_iomon_limit_mb = IOMON_DEFAULT_LIMIT;
+       }
+
+       if (!PE_parse_boot_argn("task_iomon_interval_secs", &task_iomon_interval_secs, sizeof(task_iomon_interval_secs))) {
+               task_iomon_interval_secs = IOMON_DEFAULT_INTERVAL;
+       }
+
+       if (!PE_parse_boot_argn("io_telemetry_limit", &io_telemetry_limit, sizeof(io_telemetry_limit))) {
+               io_telemetry_limit = IO_TELEMETRY_DEFAULT_LIMIT;
+       }
+
+/*
+ * If we have coalitions, coalition_init() will call init_task_ledgers() as it
+ * sets up the ledgers for the default coalition. If we don't have coalitions,
+ * then we have to call it now.
+ */
+#if CONFIG_COALITIONS
+       assert(task_ledger_template);
+#else /* CONFIG_COALITIONS */
+       init_task_ledgers();
+#endif /* CONFIG_COALITIONS */
+
+#if TASK_REFERENCE_LEAK_DEBUG
+       task_ref_btlog = btlog_create(TASK_REF_NUM_RECORDS, TASK_REF_BTDEPTH, TRUE /* caller_will_remove_entries_for_element? */);
+       assert(task_ref_btlog);
+#endif
+
+       /*
+        * Create the kernel task as the first task.
+        */
+#ifdef __LP64__
+       if (task_create_internal(TASK_NULL, NULL, FALSE, TRUE, TRUE, TF_NONE, TPF_NONE, TWF_NONE, &kernel_task) != KERN_SUCCESS)
+#else
+       if (task_create_internal(TASK_NULL, NULL, FALSE, FALSE, FALSE, TF_NONE, TPF_NONE, TWF_NONE, &kernel_task) != KERN_SUCCESS)
+#endif
+       { panic("task_init\n");}
+
+#if defined(HAS_APPLE_PAC)
+       kernel_task->rop_pid = ml_default_rop_pid();
+       kernel_task->jop_pid = ml_default_jop_pid();
+       // kernel_task never runs at EL0, but machine_thread_state_convert_from/to_user() relies on
+       // disable_user_jop to be false for kernel threads (e.g. in exception delivery on thread_exception_daemon)
+       ml_task_set_disable_user_jop(kernel_task, FALSE);
+#endif
+
+       vm_map_deallocate(kernel_task->map);
+       kernel_task->map = kernel_map;
+}
+
+/*
+ * Create a task running in the kernel address space.  It may
+ * have its own map of size mem_size and may have ipc privileges.
+ */
+kern_return_t
+kernel_task_create(
+       __unused task_t         parent_task,
+       __unused vm_offset_t            map_base,
+       __unused vm_size_t              map_size,
+       __unused task_t         *child_task)
+{
+       return KERN_INVALID_ARGUMENT;
+}
+
+kern_return_t
+task_create(
+       task_t                          parent_task,
+       __unused ledger_port_array_t    ledger_ports,
+       __unused mach_msg_type_number_t num_ledger_ports,
+       __unused boolean_t              inherit_memory,
+       __unused task_t                 *child_task)    /* OUT */
+{
+       if (parent_task == TASK_NULL) {
+               return KERN_INVALID_ARGUMENT;
+       }
+
+       /*
+        * No longer supported: too many calls assume that a task has a valid
+        * process attached.
+        */
+       return KERN_FAILURE;
+}
+
+kern_return_t
+host_security_create_task_token(
+       host_security_t                 host_security,
+       task_t                          parent_task,
+       __unused security_token_t       sec_token,
+       __unused audit_token_t          audit_token,
+       __unused host_priv_t            host_priv,
+       __unused ledger_port_array_t    ledger_ports,
+       __unused mach_msg_type_number_t num_ledger_ports,
+       __unused boolean_t              inherit_memory,
+       __unused task_t                 *child_task)    /* OUT */
+{
+       if (parent_task == TASK_NULL) {
+               return KERN_INVALID_ARGUMENT;
+       }
+
+       if (host_security == HOST_NULL) {
+               return KERN_INVALID_SECURITY;
+       }
+
+       /*
+        * No longer supported.
+        */
+       return KERN_FAILURE;
+}
+
+/*
+ * Task ledgers
+ * ------------
+ *
+ * phys_footprint
+ *   Physical footprint: This is the sum of:
+ *     + (internal - alternate_accounting)
+ *     + (internal_compressed - alternate_accounting_compressed)
+ *     + iokit_mapped
+ *     + purgeable_nonvolatile
+ *     + purgeable_nonvolatile_compressed
+ *     + page_table
+ *
+ * internal
+ *   The task's anonymous memory, which on iOS is always resident.
+ *
+ * internal_compressed
+ *   Amount of this task's internal memory which is held by the compressor.
+ *   Such memory is no longer actually resident for the task [i.e., resident in its pmap],
+ *   and could be either decompressed back into memory, or paged out to storage, depending
+ *   on our implementation.
+ *
+ * iokit_mapped
+ *   IOKit mappings: The total size of all IOKit mappings in this task, regardless of
+ *    clean/dirty or internal/external state].
+ *
+ * alternate_accounting
+ *   The number of internal dirty pages which are part of IOKit mappings. By definition, these pages
+ *   are counted in both internal *and* iokit_mapped, so we must subtract them from the total to avoid
+ *   double counting.
+ *
+ * pages_grabbed
+ *   pages_grabbed counts all page grabs in a task.  It is also broken out into three subtypes
+ *   which track UPL, IOPL and Kernel page grabs.
+ */
+void
+init_task_ledgers(void)
+{
+       ledger_template_t t;
+
+       assert(task_ledger_template == NULL);
+       assert(kernel_task == TASK_NULL);
+
+#if MACH_ASSERT
+       PE_parse_boot_argn("pmap_ledgers_panic",
+           &pmap_ledgers_panic,
+           sizeof(pmap_ledgers_panic));
+       PE_parse_boot_argn("pmap_ledgers_panic_leeway",
+           &pmap_ledgers_panic_leeway,
+           sizeof(pmap_ledgers_panic_leeway));
+#endif /* MACH_ASSERT */
+
+       if ((t = ledger_template_create("Per-task ledger")) == NULL) {
+               panic("couldn't create task ledger template");
+       }
+
+       task_ledgers.cpu_time = ledger_entry_add(t, "cpu_time", "sched", "ns");
+       task_ledgers.tkm_private = ledger_entry_add(t, "tkm_private",
+           "physmem", "bytes");
+       task_ledgers.tkm_shared = ledger_entry_add(t, "tkm_shared", "physmem",
+           "bytes");
+       task_ledgers.phys_mem = ledger_entry_add(t, "phys_mem", "physmem",
+           "bytes");
+       task_ledgers.wired_mem = ledger_entry_add(t, "wired_mem", "physmem",
+           "bytes");
+       task_ledgers.internal = ledger_entry_add(t, "internal", "physmem",
+           "bytes");
+       task_ledgers.iokit_mapped = ledger_entry_add(t, "iokit_mapped", "mappings",
+           "bytes");
+       task_ledgers.alternate_accounting = ledger_entry_add(t, "alternate_accounting", "physmem",
+           "bytes");
+       task_ledgers.alternate_accounting_compressed = ledger_entry_add(t, "alternate_accounting_compressed", "physmem",
+           "bytes");
+       task_ledgers.page_table = ledger_entry_add(t, "page_table", "physmem",
+           "bytes");
+       task_ledgers.phys_footprint = ledger_entry_add(t, "phys_footprint", "physmem",
+           "bytes");
+       task_ledgers.internal_compressed = ledger_entry_add(t, "internal_compressed", "physmem",
+           "bytes");
+       task_ledgers.purgeable_volatile = ledger_entry_add(t, "purgeable_volatile", "physmem", "bytes");
+       task_ledgers.purgeable_nonvolatile = ledger_entry_add(t, "purgeable_nonvolatile", "physmem", "bytes");
+       task_ledgers.purgeable_volatile_compressed = ledger_entry_add(t, "purgeable_volatile_compress", "physmem", "bytes");
+       task_ledgers.purgeable_nonvolatile_compressed = ledger_entry_add(t, "purgeable_nonvolatile_compress", "physmem", "bytes");
+#if DEBUG || DEVELOPMENT
+       task_ledgers.pages_grabbed = ledger_entry_add(t, "pages_grabbed", "physmem", "count");
+       task_ledgers.pages_grabbed_kern = ledger_entry_add(t, "pages_grabbed_kern", "physmem", "count");
+       task_ledgers.pages_grabbed_iopl = ledger_entry_add(t, "pages_grabbed_iopl", "physmem", "count");
+       task_ledgers.pages_grabbed_upl = ledger_entry_add(t, "pages_grabbed_upl", "physmem", "count");
+#endif
+       task_ledgers.tagged_nofootprint = ledger_entry_add(t, "tagged_nofootprint", "physmem", "bytes");
+       task_ledgers.tagged_footprint = ledger_entry_add(t, "tagged_footprint", "physmem", "bytes");
+       task_ledgers.tagged_nofootprint_compressed = ledger_entry_add(t, "tagged_nofootprint_compressed", "physmem", "bytes");
+       task_ledgers.tagged_footprint_compressed = ledger_entry_add(t, "tagged_footprint_compressed", "physmem", "bytes");
+       task_ledgers.network_volatile = ledger_entry_add(t, "network_volatile", "physmem", "bytes");
+       task_ledgers.network_nonvolatile = ledger_entry_add(t, "network_nonvolatile", "physmem", "bytes");
+       task_ledgers.network_volatile_compressed = ledger_entry_add(t, "network_volatile_compressed", "physmem", "bytes");
+       task_ledgers.network_nonvolatile_compressed = ledger_entry_add(t, "network_nonvolatile_compressed", "physmem", "bytes");
+       task_ledgers.media_nofootprint = ledger_entry_add(t, "media_nofootprint", "physmem", "bytes");
+       task_ledgers.media_footprint = ledger_entry_add(t, "media_footprint", "physmem", "bytes");
+       task_ledgers.media_nofootprint_compressed = ledger_entry_add(t, "media_nofootprint_compressed", "physmem", "bytes");
+       task_ledgers.media_footprint_compressed = ledger_entry_add(t, "media_footprint_compressed", "physmem", "bytes");
+       task_ledgers.graphics_nofootprint = ledger_entry_add(t, "graphics_nofootprint", "physmem", "bytes");
+       task_ledgers.graphics_footprint = ledger_entry_add(t, "graphics_footprint", "physmem", "bytes");
+       task_ledgers.graphics_nofootprint_compressed = ledger_entry_add(t, "graphics_nofootprint_compressed", "physmem", "bytes");
+       task_ledgers.graphics_footprint_compressed = ledger_entry_add(t, "graphics_footprint_compressed", "physmem", "bytes");
+       task_ledgers.neural_nofootprint = ledger_entry_add(t, "neural_nofootprint", "physmem", "bytes");
+       task_ledgers.neural_footprint = ledger_entry_add(t, "neural_footprint", "physmem", "bytes");
+       task_ledgers.neural_nofootprint_compressed = ledger_entry_add(t, "neural_nofootprint_compressed", "physmem", "bytes");
+       task_ledgers.neural_footprint_compressed = ledger_entry_add(t, "neural_footprint_compressed", "physmem", "bytes");
+
+#if CONFIG_FREEZE
+       task_ledgers.frozen_to_swap = ledger_entry_add(t, "frozen_to_swap", "physmem", "bytes");
+#endif /* CONFIG_FREEZE */
+
+       task_ledgers.platform_idle_wakeups = ledger_entry_add(t, "platform_idle_wakeups", "power",
+           "count");
+       task_ledgers.interrupt_wakeups = ledger_entry_add(t, "interrupt_wakeups", "power",
+           "count");
+
+#if CONFIG_SCHED_SFI
+       sfi_class_id_t class_id, ledger_alias;
+       for (class_id = SFI_CLASS_UNSPECIFIED; class_id < MAX_SFI_CLASS_ID; class_id++) {
+               task_ledgers.sfi_wait_times[class_id] = -1;
+       }
+
+       /* don't account for UNSPECIFIED */
+       for (class_id = SFI_CLASS_UNSPECIFIED + 1; class_id < MAX_SFI_CLASS_ID; class_id++) {
+               ledger_alias = sfi_get_ledger_alias_for_class(class_id);
+               if (ledger_alias != SFI_CLASS_UNSPECIFIED) {
+                       /* Check to see if alias has been registered yet */
+                       if (task_ledgers.sfi_wait_times[ledger_alias] != -1) {
+                               task_ledgers.sfi_wait_times[class_id] = task_ledgers.sfi_wait_times[ledger_alias];
+                       } else {
+                               /* Otherwise, initialize it first */
+                               task_ledgers.sfi_wait_times[class_id] = task_ledgers.sfi_wait_times[ledger_alias] = sfi_ledger_entry_add(t, ledger_alias);
+                       }
+               } else {
+                       task_ledgers.sfi_wait_times[class_id] = sfi_ledger_entry_add(t, class_id);
+               }
+
+               if (task_ledgers.sfi_wait_times[class_id] < 0) {
+                       panic("couldn't create entries for task ledger template for SFI class 0x%x", class_id);
+               }
+       }
+
+       assert(task_ledgers.sfi_wait_times[MAX_SFI_CLASS_ID - 1] != -1);
+#endif /* CONFIG_SCHED_SFI */
+
+       task_ledgers.cpu_time_billed_to_me = ledger_entry_add(t, "cpu_time_billed_to_me", "sched", "ns");
+       task_ledgers.cpu_time_billed_to_others = ledger_entry_add(t, "cpu_time_billed_to_others", "sched", "ns");
+       task_ledgers.physical_writes = ledger_entry_add(t, "physical_writes", "res", "bytes");
+       task_ledgers.logical_writes = ledger_entry_add(t, "logical_writes", "res", "bytes");
+       task_ledgers.logical_writes_to_external = ledger_entry_add(t, "logical_writes_to_external", "res", "bytes");
+#if CONFIG_PHYS_WRITE_ACCT
+       task_ledgers.fs_metadata_writes = ledger_entry_add(t, "fs_metadata_writes", "res", "bytes");
+#endif /* CONFIG_PHYS_WRITE_ACCT */
+       task_ledgers.energy_billed_to_me = ledger_entry_add(t, "energy_billed_to_me", "power", "nj");
+       task_ledgers.energy_billed_to_others = ledger_entry_add(t, "energy_billed_to_others", "power", "nj");
+
+       if ((task_ledgers.cpu_time < 0) ||
+           (task_ledgers.tkm_private < 0) ||
+           (task_ledgers.tkm_shared < 0) ||
+           (task_ledgers.phys_mem < 0) ||
+           (task_ledgers.wired_mem < 0) ||
+           (task_ledgers.internal < 0) ||
+           (task_ledgers.iokit_mapped < 0) ||
+           (task_ledgers.alternate_accounting < 0) ||
+           (task_ledgers.alternate_accounting_compressed < 0) ||
+           (task_ledgers.page_table < 0) ||
+           (task_ledgers.phys_footprint < 0) ||
+           (task_ledgers.internal_compressed < 0) ||
+           (task_ledgers.purgeable_volatile < 0) ||
+           (task_ledgers.purgeable_nonvolatile < 0) ||
+           (task_ledgers.purgeable_volatile_compressed < 0) ||
+           (task_ledgers.purgeable_nonvolatile_compressed < 0) ||
+           (task_ledgers.tagged_nofootprint < 0) ||
+           (task_ledgers.tagged_footprint < 0) ||
+           (task_ledgers.tagged_nofootprint_compressed < 0) ||
+           (task_ledgers.tagged_footprint_compressed < 0) ||
+#if CONFIG_FREEZE
+           (task_ledgers.frozen_to_swap < 0) ||
+#endif /* CONFIG_FREEZE */
+           (task_ledgers.network_volatile < 0) ||
+           (task_ledgers.network_nonvolatile < 0) ||
+           (task_ledgers.network_volatile_compressed < 0) ||
+           (task_ledgers.network_nonvolatile_compressed < 0) ||
+           (task_ledgers.media_nofootprint < 0) ||
+           (task_ledgers.media_footprint < 0) ||
+           (task_ledgers.media_nofootprint_compressed < 0) ||
+           (task_ledgers.media_footprint_compressed < 0) ||
+           (task_ledgers.graphics_nofootprint < 0) ||
+           (task_ledgers.graphics_footprint < 0) ||
+           (task_ledgers.graphics_nofootprint_compressed < 0) ||
+           (task_ledgers.graphics_footprint_compressed < 0) ||
+           (task_ledgers.neural_nofootprint < 0) ||
+           (task_ledgers.neural_footprint < 0) ||
+           (task_ledgers.neural_nofootprint_compressed < 0) ||
+           (task_ledgers.neural_footprint_compressed < 0) ||
+           (task_ledgers.platform_idle_wakeups < 0) ||
+           (task_ledgers.interrupt_wakeups < 0) ||
+           (task_ledgers.cpu_time_billed_to_me < 0) || (task_ledgers.cpu_time_billed_to_others < 0) ||
+           (task_ledgers.physical_writes < 0) ||
+           (task_ledgers.logical_writes < 0) ||
+           (task_ledgers.logical_writes_to_external < 0) ||
+#if CONFIG_PHYS_WRITE_ACCT
+           (task_ledgers.fs_metadata_writes < 0) ||
+#endif /* CONFIG_PHYS_WRITE_ACCT */
+           (task_ledgers.energy_billed_to_me < 0) ||
+           (task_ledgers.energy_billed_to_others < 0)
+           ) {
+               panic("couldn't create entries for task ledger template");
+       }
+
+       ledger_track_credit_only(t, task_ledgers.phys_footprint);
+       ledger_track_credit_only(t, task_ledgers.page_table);
+       ledger_track_credit_only(t, task_ledgers.internal);
+       ledger_track_credit_only(t, task_ledgers.internal_compressed);
+       ledger_track_credit_only(t, task_ledgers.iokit_mapped);
+       ledger_track_credit_only(t, task_ledgers.alternate_accounting);
+       ledger_track_credit_only(t, task_ledgers.alternate_accounting_compressed);
+       ledger_track_credit_only(t, task_ledgers.purgeable_volatile);
+       ledger_track_credit_only(t, task_ledgers.purgeable_nonvolatile);
+       ledger_track_credit_only(t, task_ledgers.purgeable_volatile_compressed);
+       ledger_track_credit_only(t, task_ledgers.purgeable_nonvolatile_compressed);
+#if DEBUG || DEVELOPMENT
+       ledger_track_credit_only(t, task_ledgers.pages_grabbed);
+       ledger_track_credit_only(t, task_ledgers.pages_grabbed_kern);
+       ledger_track_credit_only(t, task_ledgers.pages_grabbed_iopl);
+       ledger_track_credit_only(t, task_ledgers.pages_grabbed_upl);
+#endif
+
+       ledger_track_credit_only(t, task_ledgers.tagged_nofootprint);
+       ledger_track_credit_only(t, task_ledgers.tagged_footprint);
+       ledger_track_credit_only(t, task_ledgers.tagged_nofootprint_compressed);
+       ledger_track_credit_only(t, task_ledgers.tagged_footprint_compressed);
+       ledger_track_credit_only(t, task_ledgers.network_volatile);
+       ledger_track_credit_only(t, task_ledgers.network_nonvolatile);
+       ledger_track_credit_only(t, task_ledgers.network_volatile_compressed);
+       ledger_track_credit_only(t, task_ledgers.network_nonvolatile_compressed);
+       ledger_track_credit_only(t, task_ledgers.media_nofootprint);
+       ledger_track_credit_only(t, task_ledgers.media_footprint);
+       ledger_track_credit_only(t, task_ledgers.media_nofootprint_compressed);
+       ledger_track_credit_only(t, task_ledgers.media_footprint_compressed);
+       ledger_track_credit_only(t, task_ledgers.graphics_nofootprint);
+       ledger_track_credit_only(t, task_ledgers.graphics_footprint);
+       ledger_track_credit_only(t, task_ledgers.graphics_nofootprint_compressed);
+       ledger_track_credit_only(t, task_ledgers.graphics_footprint_compressed);
+       ledger_track_credit_only(t, task_ledgers.neural_nofootprint);
+       ledger_track_credit_only(t, task_ledgers.neural_footprint);
+       ledger_track_credit_only(t, task_ledgers.neural_nofootprint_compressed);
+       ledger_track_credit_only(t, task_ledgers.neural_footprint_compressed);
+
+       ledger_track_maximum(t, task_ledgers.phys_footprint, 60);
+#if MACH_ASSERT
+       if (pmap_ledgers_panic) {
+               ledger_panic_on_negative(t, task_ledgers.phys_footprint);
+               ledger_panic_on_negative(t, task_ledgers.page_table);
+               ledger_panic_on_negative(t, task_ledgers.internal);
+               ledger_panic_on_negative(t, task_ledgers.internal_compressed);
+               ledger_panic_on_negative(t, task_ledgers.iokit_mapped);
+               ledger_panic_on_negative(t, task_ledgers.alternate_accounting);
+               ledger_panic_on_negative(t, task_ledgers.alternate_accounting_compressed);
+               ledger_panic_on_negative(t, task_ledgers.purgeable_volatile);
+               ledger_panic_on_negative(t, task_ledgers.purgeable_nonvolatile);
+               ledger_panic_on_negative(t, task_ledgers.purgeable_volatile_compressed);
+               ledger_panic_on_negative(t, task_ledgers.purgeable_nonvolatile_compressed);
+#if CONFIG_PHYS_WRITE_ACCT
+               ledger_panic_on_negative(t, task_ledgers.fs_metadata_writes);
+#endif /* CONFIG_PHYS_WRITE_ACCT */
+
+               ledger_panic_on_negative(t, task_ledgers.tagged_nofootprint);
+               ledger_panic_on_negative(t, task_ledgers.tagged_footprint);
+               ledger_panic_on_negative(t, task_ledgers.tagged_nofootprint_compressed);
+               ledger_panic_on_negative(t, task_ledgers.tagged_footprint_compressed);
+               ledger_panic_on_negative(t, task_ledgers.network_volatile);
+               ledger_panic_on_negative(t, task_ledgers.network_nonvolatile);
+               ledger_panic_on_negative(t, task_ledgers.network_volatile_compressed);
+               ledger_panic_on_negative(t, task_ledgers.network_nonvolatile_compressed);
+               ledger_panic_on_negative(t, task_ledgers.media_nofootprint);
+               ledger_panic_on_negative(t, task_ledgers.media_footprint);
+               ledger_panic_on_negative(t, task_ledgers.media_nofootprint_compressed);
+               ledger_panic_on_negative(t, task_ledgers.media_footprint_compressed);
+               ledger_panic_on_negative(t, task_ledgers.graphics_nofootprint);
+               ledger_panic_on_negative(t, task_ledgers.graphics_footprint);
+               ledger_panic_on_negative(t, task_ledgers.graphics_nofootprint_compressed);
+               ledger_panic_on_negative(t, task_ledgers.graphics_footprint_compressed);
+               ledger_panic_on_negative(t, task_ledgers.neural_nofootprint);
+               ledger_panic_on_negative(t, task_ledgers.neural_footprint);
+               ledger_panic_on_negative(t, task_ledgers.neural_nofootprint_compressed);
+               ledger_panic_on_negative(t, task_ledgers.neural_footprint_compressed);
+       }
+#endif /* MACH_ASSERT */
+
+#if CONFIG_MEMORYSTATUS
+       ledger_set_callback(t, task_ledgers.phys_footprint, task_footprint_exceeded, NULL, NULL);
+#endif /* CONFIG_MEMORYSTATUS */
+
+       ledger_set_callback(t, task_ledgers.interrupt_wakeups,
+           task_wakeups_rate_exceeded, NULL, NULL);
+       ledger_set_callback(t, task_ledgers.physical_writes, task_io_rate_exceeded, (void *)FLAVOR_IO_PHYSICAL_WRITES, NULL);
+
+#if XNU_MONITOR
+       ledger_template_complete_secure_alloc(t);
+#else /* XNU_MONITOR */
+       ledger_template_complete(t);
+#endif /* XNU_MONITOR */
+       task_ledger_template = t;
+}
+
+os_refgrp_decl(static, task_refgrp, "task", NULL);
+
+kern_return_t
+task_create_internal(
+       task_t          parent_task,
+       coalition_t     *parent_coalitions __unused,
+       boolean_t       inherit_memory,
+       __unused boolean_t      is_64bit,
+       boolean_t is_64bit_data,
+       uint32_t        t_flags,
+       uint32_t        t_procflags,
+       uint8_t         t_returnwaitflags,
+       task_t          *child_task)            /* OUT */
+{
+       task_t                  new_task;
+       vm_shared_region_t      shared_region;
+       ledger_t                ledger = NULL;
+
+       new_task = (task_t) zalloc(task_zone);
+
+       if (new_task == TASK_NULL) {
+               return KERN_RESOURCE_SHORTAGE;
+       }
+
+       /* one ref for just being alive; one for our caller */
+       os_ref_init_count(&new_task->ref_count, &task_refgrp, 2);
+
+       /* allocate with active entries */
+       assert(task_ledger_template != NULL);
+       if ((ledger = ledger_instantiate(task_ledger_template,
+           LEDGER_CREATE_ACTIVE_ENTRIES)) == NULL) {
+               zfree(task_zone, new_task);
+               return KERN_RESOURCE_SHORTAGE;
+       }
+
+       counter_alloc(&(new_task->faults));
+
+#if defined(HAS_APPLE_PAC)
+       ml_task_set_rop_pid(new_task, parent_task, inherit_memory);
+       ml_task_set_jop_pid(new_task, parent_task, inherit_memory);
+       ml_task_set_disable_user_jop(new_task, inherit_memory ? parent_task->disable_user_jop : FALSE);
+#endif
+
+
+       new_task->ledger = ledger;
+
+#if defined(CONFIG_SCHED_MULTIQ)
+       new_task->sched_group = sched_group_create();
+#endif
+
+       /* if inherit_memory is true, parent_task MUST not be NULL */
+       if (!(t_flags & TF_CORPSE_FORK) && inherit_memory) {
+               new_task->map = vm_map_fork(ledger, parent_task->map, 0);
+       } else {
+               unsigned int pmap_flags = is_64bit ? PMAP_CREATE_64BIT : 0;
+               new_task->map = vm_map_create(pmap_create_options(ledger, 0, pmap_flags),
+                   (vm_map_offset_t)(VM_MIN_ADDRESS),
+                   (vm_map_offset_t)(VM_MAX_ADDRESS), TRUE);
+       }
+
+       /* Inherit memlock limit from parent */
+       if (parent_task) {
+               vm_map_set_user_wire_limit(new_task->map, (vm_size_t)parent_task->map->user_wire_limit);
+       }
+
+       lck_mtx_init(&new_task->lock, &task_lck_grp, &task_lck_attr);
+       queue_init(&new_task->threads);
+       new_task->suspend_count = 0;
+       new_task->thread_count = 0;
+       new_task->active_thread_count = 0;
+       new_task->user_stop_count = 0;
+       new_task->legacy_stop_count = 0;
+       new_task->active = TRUE;
+       new_task->halting = FALSE;
+       new_task->priv_flags = 0;
+       new_task->t_flags = t_flags;
+       new_task->t_procflags = t_procflags;
+       new_task->t_returnwaitflags = t_returnwaitflags;
+       new_task->returnwait_inheritor = current_thread();
+       new_task->importance = 0;
+       new_task->crashed_thread_id = 0;
+       new_task->exec_token = 0;
+       new_task->watchports = NULL;
+       new_task->restartable_ranges = NULL;
+       new_task->task_exc_guard = 0;
+
+       new_task->bank_context = NULL;
+
+#ifdef MACH_BSD
+       new_task->bsd_info = NULL;
+       new_task->corpse_info = NULL;
+#endif /* MACH_BSD */
+
+#if CONFIG_MACF
+       new_task->crash_label = NULL;
+
+       new_task->mach_trap_filter_mask = NULL;
+       new_task->mach_kobj_filter_mask = NULL;
+#endif
+
+#if CONFIG_MEMORYSTATUS
+       if (max_task_footprint != 0) {
+               ledger_set_limit(ledger, task_ledgers.phys_footprint, max_task_footprint, PHYS_FOOTPRINT_WARNING_LEVEL);
+       }
+#endif /* CONFIG_MEMORYSTATUS */
+
+       if (task_wakeups_monitor_rate != 0) {
+               uint32_t flags = WAKEMON_ENABLE | WAKEMON_SET_DEFAULTS;
+               int32_t  rate; // Ignored because of WAKEMON_SET_DEFAULTS
+               task_wakeups_monitor_ctl(new_task, &flags, &rate);
+       }
+
+#if CONFIG_IO_ACCOUNTING
+       uint32_t flags = IOMON_ENABLE;
+       task_io_monitor_ctl(new_task, &flags);
+#endif /* CONFIG_IO_ACCOUNTING */
+
+       machine_task_init(new_task, parent_task, inherit_memory);
+
+       new_task->task_debug = NULL;
+
+#if DEVELOPMENT || DEBUG
+       new_task->task_unnested = FALSE;
+       new_task->task_disconnected_count = 0;
+#endif
+       queue_init(&new_task->semaphore_list);
+       new_task->semaphores_owned = 0;
+
+       ipc_task_init(new_task, parent_task);
+
+       new_task->vtimers = 0;
+
+       new_task->shared_region = NULL;
+
+       new_task->affinity_space = NULL;
+
+       new_task->t_kpc = 0;
+
+       new_task->pidsuspended = FALSE;
+       new_task->frozen = FALSE;
+       new_task->changing_freeze_state = FALSE;
+       new_task->rusage_cpu_flags = 0;
+       new_task->rusage_cpu_percentage = 0;
+       new_task->rusage_cpu_interval = 0;
+       new_task->rusage_cpu_deadline = 0;
+       new_task->rusage_cpu_callt = NULL;
+#if MACH_ASSERT
+       new_task->suspends_outstanding = 0;
+#endif
+
+#if HYPERVISOR
+       new_task->hv_task_target = NULL;
+#endif /* HYPERVISOR */
+
+#if CONFIG_TASKWATCH
+       queue_init(&new_task->task_watchers);
+       new_task->num_taskwatchers  = 0;
+       new_task->watchapplying  = 0;
+#endif /* CONFIG_TASKWATCH */
+
+       new_task->mem_notify_reserved = 0;
+       new_task->memlimit_attrs_reserved = 0;
+
+       new_task->requested_policy = default_task_requested_policy;
+       new_task->effective_policy = default_task_effective_policy;
+
+       new_task->task_shared_region_slide = -1;
+
+       task_importance_init_from_parent(new_task, parent_task);
+
+       if (parent_task != TASK_NULL) {
+               new_task->sec_token = parent_task->sec_token;
+               new_task->audit_token = parent_task->audit_token;
+
+               /* inherit the parent's shared region */
+               shared_region = vm_shared_region_get(parent_task);
+               vm_shared_region_set(new_task, shared_region);
+
+#if __has_feature(ptrauth_calls)
+               /* use parent's shared_region_id */
+               char *shared_region_id = task_get_vm_shared_region_id_and_jop_pid(parent_task, NULL);
+               if (shared_region_id != NULL) {
+                       shared_region_key_alloc(shared_region_id, FALSE, 0);   /* get a reference */
+               }
+               task_set_shared_region_id(new_task, shared_region_id);
+#endif /* __has_feature(ptrauth_calls) */
+
+               if (task_has_64Bit_addr(parent_task)) {
+                       task_set_64Bit_addr(new_task);
+               }
+
+               if (task_has_64Bit_data(parent_task)) {
+                       task_set_64Bit_data(new_task);
+               }
+
+               new_task->all_image_info_addr = parent_task->all_image_info_addr;
+               new_task->all_image_info_size = parent_task->all_image_info_size;
+               new_task->mach_header_vm_address = 0;
+
+               if (inherit_memory && parent_task->affinity_space) {
+                       task_affinity_create(parent_task, new_task);
+               }
+
+               new_task->pset_hint = parent_task->pset_hint = task_choose_pset(parent_task);
+
+               if (parent_task->t_flags & TF_NO_SMT) {
+                       new_task->t_flags |= TF_NO_SMT;
+               }
+
+               if (parent_task->t_flags & TF_TECS) {
+                       new_task->t_flags |= TF_TECS;
+               }
+
+               if (parent_task->t_flags & TF_FILTER_MSG) {
+                       new_task->t_flags |= TF_FILTER_MSG;
+               }
+
+               new_task->priority = BASEPRI_DEFAULT;
+               new_task->max_priority = MAXPRI_USER;
+
+               task_policy_create(new_task, parent_task);
+       } else {
+               new_task->sec_token = KERNEL_SECURITY_TOKEN;
+               new_task->audit_token = KERNEL_AUDIT_TOKEN;
+#ifdef __LP64__
+               if (is_64bit) {
+                       task_set_64Bit_addr(new_task);
+               }
+#endif
+
+               if (is_64bit_data) {
+                       task_set_64Bit_data(new_task);
+               }
+
+               new_task->all_image_info_addr = (mach_vm_address_t)0;
+               new_task->all_image_info_size = (mach_vm_size_t)0;
+
+               new_task->pset_hint = PROCESSOR_SET_NULL;
+
+               if (kernel_task == TASK_NULL) {
+                       new_task->priority = BASEPRI_KERNEL;
+                       new_task->max_priority = MAXPRI_KERNEL;
+               } else {
+                       new_task->priority = BASEPRI_DEFAULT;
+                       new_task->max_priority = MAXPRI_USER;
+               }
+       }
+
+       bzero(new_task->coalition, sizeof(new_task->coalition));
+       for (int i = 0; i < COALITION_NUM_TYPES; i++) {
+               queue_chain_init(new_task->task_coalition[i]);
+       }
+
+       /* Allocate I/O Statistics */
+       new_task->task_io_stats = kheap_alloc(KHEAP_DATA_BUFFERS,
+           sizeof(struct io_stat_info), Z_WAITOK | Z_ZERO);
+       assert(new_task->task_io_stats != NULL);
+
+       bzero(&(new_task->cpu_time_eqos_stats), sizeof(new_task->cpu_time_eqos_stats));
+       bzero(&(new_task->cpu_time_rqos_stats), sizeof(new_task->cpu_time_rqos_stats));
+
+       bzero(&new_task->extmod_statistics, sizeof(new_task->extmod_statistics));
+
+       /* Copy resource acc. info from Parent for Corpe Forked task. */
+       if (parent_task != NULL && (t_flags & TF_CORPSE_FORK)) {
+               task_rollup_accounting_info(new_task, parent_task);
+       } else {
+               /* Initialize to zero for standard fork/spawn case */
+               new_task->total_user_time = 0;
+               new_task->total_system_time = 0;
+               new_task->total_ptime = 0;
+               new_task->total_runnable_time = 0;
+               new_task->pageins = 0;
+               new_task->cow_faults = 0;
+               new_task->messages_sent = 0;
+               new_task->messages_received = 0;
+               new_task->syscalls_mach = 0;
+               new_task->syscalls_unix = 0;
+               new_task->c_switch = 0;
+               new_task->p_switch = 0;
+               new_task->ps_switch = 0;
+               new_task->decompressions = 0;
+               new_task->low_mem_notified_warn = 0;
+               new_task->low_mem_notified_critical = 0;
+               new_task->purged_memory_warn = 0;
+               new_task->purged_memory_critical = 0;
+               new_task->low_mem_privileged_listener = 0;
+               new_task->memlimit_is_active = 0;
+               new_task->memlimit_is_fatal = 0;
+               new_task->memlimit_active_exc_resource = 0;
+               new_task->memlimit_inactive_exc_resource = 0;
+               new_task->task_timer_wakeups_bin_1 = 0;
+               new_task->task_timer_wakeups_bin_2 = 0;
+               new_task->task_gpu_ns = 0;
+               new_task->task_writes_counters_internal.task_immediate_writes = 0;
+               new_task->task_writes_counters_internal.task_deferred_writes = 0;
+               new_task->task_writes_counters_internal.task_invalidated_writes = 0;
+               new_task->task_writes_counters_internal.task_metadata_writes = 0;
+               new_task->task_writes_counters_external.task_immediate_writes = 0;
+               new_task->task_writes_counters_external.task_deferred_writes = 0;
+               new_task->task_writes_counters_external.task_invalidated_writes = 0;
+               new_task->task_writes_counters_external.task_metadata_writes = 0;
+#if CONFIG_PHYS_WRITE_ACCT
+               new_task->task_fs_metadata_writes = 0;
+#endif /* CONFIG_PHYS_WRITE_ACCT */
+
+               new_task->task_energy = 0;
+#if MONOTONIC
+               memset(&new_task->task_monotonic, 0, sizeof(new_task->task_monotonic));
+#endif /* MONOTONIC */
+       }
+
+
+#if CONFIG_COALITIONS
+       if (!(t_flags & TF_CORPSE_FORK)) {
+               /* TODO: there is no graceful failure path here... */
+               if (parent_coalitions && parent_coalitions[COALITION_TYPE_RESOURCE]) {
+                       coalitions_adopt_task(parent_coalitions, new_task);
+               } else if (parent_task && parent_task->coalition[COALITION_TYPE_RESOURCE]) {
+                       /*
+                        * all tasks at least have a resource coalition, so
+                        * if the parent has one then inherit all coalitions
+                        * the parent is a part of
+                        */
+                       coalitions_adopt_task(parent_task->coalition, new_task);
+               } else {
+                       /* TODO: assert that new_task will be PID 1 (launchd) */
+                       coalitions_adopt_init_task(new_task);
+               }
+               /*
+                * on exec, we need to transfer the coalition roles from the
+                * parent task to the exec copy task.
+                */
+               if (parent_task && (t_procflags & TPF_EXEC_COPY)) {
+                       int coal_roles[COALITION_NUM_TYPES];
+                       task_coalition_roles(parent_task, coal_roles);
+                       (void)coalitions_set_roles(new_task->coalition, new_task, coal_roles);
+               }
+       } else {
+               coalitions_adopt_corpse_task(new_task);
+       }
+
+       if (new_task->coalition[COALITION_TYPE_RESOURCE] == COALITION_NULL) {
+               panic("created task is not a member of a resource coalition");
+       }
+#endif /* CONFIG_COALITIONS */
+
+       new_task->dispatchqueue_offset = 0;
+       if (parent_task != NULL) {
+               new_task->dispatchqueue_offset = parent_task->dispatchqueue_offset;
+       }
+
+       new_task->task_can_transfer_memory_ownership = FALSE;
+       new_task->task_volatile_objects = 0;
+       new_task->task_nonvolatile_objects = 0;
+       new_task->task_objects_disowning = FALSE;
+       new_task->task_objects_disowned = FALSE;
+       new_task->task_owned_objects = 0;
+       queue_init(&new_task->task_objq);
+
+#if CONFIG_FREEZE
+       queue_init(&new_task->task_frozen_cseg_q);
+#endif /* CONFIG_FREEZE */
+
+       task_objq_lock_init(new_task);
+
+#if __arm64__
+       new_task->task_legacy_footprint = FALSE;
+       new_task->task_extra_footprint_limit = FALSE;
+       new_task->task_ios13extended_footprint_limit = FALSE;
+#endif /* __arm64__ */
+       new_task->task_region_footprint = FALSE;
+       new_task->task_has_crossed_thread_limit = FALSE;
+       new_task->task_thread_limit = 0;
+#if CONFIG_SECLUDED_MEMORY
+       new_task->task_can_use_secluded_mem = FALSE;
+       new_task->task_could_use_secluded_mem = FALSE;
+       new_task->task_could_also_use_secluded_mem = FALSE;
+       new_task->task_suppressed_secluded = FALSE;
+#endif /* CONFIG_SECLUDED_MEMORY */
+
+       /*
+        * t_flags is set up above. But since we don't
+        * support darkwake mode being set that way
+        * currently, we clear it out here explicitly.
+        */
+       new_task->t_flags &= ~(TF_DARKWAKE_MODE);
+
+       queue_init(&new_task->io_user_clients);
+       new_task->loadTag = 0;
+
+       ipc_task_enable(new_task);
+
+       lck_mtx_lock(&tasks_threads_lock);
+       queue_enter(&tasks, new_task, task_t, tasks);
+       tasks_count++;
+       if (tasks_suspend_state) {
+               task_suspend_internal(new_task);
+       }
+       lck_mtx_unlock(&tasks_threads_lock);
+
+       *child_task = new_task;
+       return KERN_SUCCESS;
+}
+
+/*
+ *     task_rollup_accounting_info
+ *
+ *     Roll up accounting stats. Used to rollup stats
+ *     for exec copy task and corpse fork.
+ */
+void
+task_rollup_accounting_info(task_t to_task, task_t from_task)
+{
+       assert(from_task != to_task);
+
+       to_task->total_user_time = from_task->total_user_time;
+       to_task->total_system_time = from_task->total_system_time;
+       to_task->total_ptime = from_task->total_ptime;
+       to_task->total_runnable_time = from_task->total_runnable_time;
+       counter_add(&to_task->faults, counter_load(&from_task->faults));
+       to_task->pageins = from_task->pageins;
+       to_task->cow_faults = from_task->cow_faults;
+       to_task->decompressions = from_task->decompressions;
+       to_task->messages_sent = from_task->messages_sent;
+       to_task->messages_received = from_task->messages_received;
+       to_task->syscalls_mach = from_task->syscalls_mach;
+       to_task->syscalls_unix = from_task->syscalls_unix;
+       to_task->c_switch = from_task->c_switch;
+       to_task->p_switch = from_task->p_switch;
+       to_task->ps_switch = from_task->ps_switch;
+       to_task->extmod_statistics = from_task->extmod_statistics;
+       to_task->low_mem_notified_warn = from_task->low_mem_notified_warn;
+       to_task->low_mem_notified_critical = from_task->low_mem_notified_critical;
+       to_task->purged_memory_warn = from_task->purged_memory_warn;
+       to_task->purged_memory_critical = from_task->purged_memory_critical;
+       to_task->low_mem_privileged_listener = from_task->low_mem_privileged_listener;
+       *to_task->task_io_stats = *from_task->task_io_stats;
+       to_task->cpu_time_eqos_stats = from_task->cpu_time_eqos_stats;
+       to_task->cpu_time_rqos_stats = from_task->cpu_time_rqos_stats;
+       to_task->task_timer_wakeups_bin_1 = from_task->task_timer_wakeups_bin_1;
+       to_task->task_timer_wakeups_bin_2 = from_task->task_timer_wakeups_bin_2;
+       to_task->task_gpu_ns = from_task->task_gpu_ns;
+       to_task->task_writes_counters_internal.task_immediate_writes = from_task->task_writes_counters_internal.task_immediate_writes;
+       to_task->task_writes_counters_internal.task_deferred_writes = from_task->task_writes_counters_internal.task_deferred_writes;
+       to_task->task_writes_counters_internal.task_invalidated_writes = from_task->task_writes_counters_internal.task_invalidated_writes;
+       to_task->task_writes_counters_internal.task_metadata_writes = from_task->task_writes_counters_internal.task_metadata_writes;
+       to_task->task_writes_counters_external.task_immediate_writes = from_task->task_writes_counters_external.task_immediate_writes;
+       to_task->task_writes_counters_external.task_deferred_writes = from_task->task_writes_counters_external.task_deferred_writes;
+       to_task->task_writes_counters_external.task_invalidated_writes = from_task->task_writes_counters_external.task_invalidated_writes;
+       to_task->task_writes_counters_external.task_metadata_writes = from_task->task_writes_counters_external.task_metadata_writes;
+#if CONFIG_PHYS_WRITE_ACCT
+       to_task->task_fs_metadata_writes = from_task->task_fs_metadata_writes;
+#endif /* CONFIG_PHYS_WRITE_ACCT */
+       to_task->task_energy = from_task->task_energy;
+
+       /* Skip ledger roll up for memory accounting entries */
+       ledger_rollup_entry(to_task->ledger, from_task->ledger, task_ledgers.cpu_time);
+       ledger_rollup_entry(to_task->ledger, from_task->ledger, task_ledgers.platform_idle_wakeups);
+       ledger_rollup_entry(to_task->ledger, from_task->ledger, task_ledgers.interrupt_wakeups);
+#if CONFIG_SCHED_SFI
+       for (sfi_class_id_t class_id = SFI_CLASS_UNSPECIFIED; class_id < MAX_SFI_CLASS_ID; class_id++) {
+               ledger_rollup_entry(to_task->ledger, from_task->ledger, task_ledgers.sfi_wait_times[class_id]);
+       }
+#endif
+       ledger_rollup_entry(to_task->ledger, from_task->ledger, task_ledgers.cpu_time_billed_to_me);
+       ledger_rollup_entry(to_task->ledger, from_task->ledger, task_ledgers.cpu_time_billed_to_others);
+       ledger_rollup_entry(to_task->ledger, from_task->ledger, task_ledgers.physical_writes);
+       ledger_rollup_entry(to_task->ledger, from_task->ledger, task_ledgers.logical_writes);
+       ledger_rollup_entry(to_task->ledger, from_task->ledger, task_ledgers.energy_billed_to_me);
+       ledger_rollup_entry(to_task->ledger, from_task->ledger, task_ledgers.energy_billed_to_others);
+}
+
+int task_dropped_imp_count = 0;
+
+/*
+ *     task_deallocate:
+ *
+ *     Drop a reference on a task.
+ */
+void
+task_deallocate(
+       task_t          task)
+{
+       ledger_amount_t credit, debit, interrupt_wakeups, platform_idle_wakeups;
+       os_ref_count_t refs;
+
+       if (task == TASK_NULL) {
+               return;
+       }
+
+       refs = task_deallocate_internal(task);
+
+#if IMPORTANCE_INHERITANCE
+       if (refs == 1) {
+               /*
+                * If last ref potentially comes from the task's importance,
+                * disconnect it.  But more task refs may be added before
+                * that completes, so wait for the reference to go to zero
+                * naturally (it may happen on a recursive task_deallocate()
+                * from the ipc_importance_disconnect_task() call).
+                */
+               if (IIT_NULL != task->task_imp_base) {
+                       ipc_importance_disconnect_task(task);
+               }
+               return;
+       }
+#endif /* IMPORTANCE_INHERITANCE */
+
+       if (refs > 0) {
+               return;
+       }
+
+       /*
+        * The task should be dead at this point. Ensure other resources
+        * like threads, are gone before we trash the world.
+        */
+       assert(queue_empty(&task->threads));
+       assert(task->bsd_info == NULL);
+       assert(!is_active(task->itk_space));
+       assert(!task->active);
+       assert(task->active_thread_count == 0);
+
+       lck_mtx_lock(&tasks_threads_lock);
+       assert(terminated_tasks_count > 0);
+       queue_remove(&terminated_tasks, task, task_t, tasks);
+       terminated_tasks_count--;
+       lck_mtx_unlock(&tasks_threads_lock);
+
+       /*
+        * remove the reference on bank context
+        */
+       task_bank_reset(task);
+
+       if (task->task_io_stats) {
+               kheap_free(KHEAP_DATA_BUFFERS, task->task_io_stats,
+                   sizeof(struct io_stat_info));
+       }
+
+       /*
+        *      Give the machine dependent code a chance
+        *      to perform cleanup before ripping apart
+        *      the task.
+        */
+       machine_task_terminate(task);
+
+       ipc_task_terminate(task);
+
+       /* let iokit know */
+       iokit_task_terminate(task);
+
+       if (task->affinity_space) {
+               task_affinity_deallocate(task);
+       }
+
+#if MACH_ASSERT
+       if (task->ledger != NULL &&
+           task->map != NULL &&
+           task->map->pmap != NULL &&
+           task->map->pmap->ledger != NULL) {
+               assert(task->ledger == task->map->pmap->ledger);
+       }
+#endif /* MACH_ASSERT */
+
+       vm_owned_objects_disown(task);
+       assert(task->task_objects_disowned);
+       if (task->task_volatile_objects != 0 ||
+           task->task_nonvolatile_objects != 0 ||
+           task->task_owned_objects != 0) {
+               panic("task_deallocate(%p): "
+                   "volatile_objects=%d nonvolatile_objects=%d owned=%d\n",
+                   task,
+                   task->task_volatile_objects,
+                   task->task_nonvolatile_objects,
+                   task->task_owned_objects);
+       }
+
+       vm_map_deallocate(task->map);
+       is_release(task->itk_space);
+       if (task->restartable_ranges) {
+               restartable_ranges_release(task->restartable_ranges);
+       }
+
+       ledger_get_entries(task->ledger, task_ledgers.interrupt_wakeups,
+           &interrupt_wakeups, &debit);
+       ledger_get_entries(task->ledger, task_ledgers.platform_idle_wakeups,
+           &platform_idle_wakeups, &debit);
+
+#if defined(CONFIG_SCHED_MULTIQ)
+       sched_group_destroy(task->sched_group);
+#endif
+
+       /* Accumulate statistics for dead tasks */
+       lck_spin_lock(&dead_task_statistics_lock);
+       dead_task_statistics.total_user_time += task->total_user_time;
+       dead_task_statistics.total_system_time += task->total_system_time;
+
+       dead_task_statistics.task_interrupt_wakeups += interrupt_wakeups;
+       dead_task_statistics.task_platform_idle_wakeups += platform_idle_wakeups;
+
+       dead_task_statistics.task_timer_wakeups_bin_1 += task->task_timer_wakeups_bin_1;
+       dead_task_statistics.task_timer_wakeups_bin_2 += task->task_timer_wakeups_bin_2;
+       dead_task_statistics.total_ptime += task->total_ptime;
+       dead_task_statistics.total_pset_switches += task->ps_switch;
+       dead_task_statistics.task_gpu_ns += task->task_gpu_ns;
+       dead_task_statistics.task_energy += task->task_energy;
+
+       lck_spin_unlock(&dead_task_statistics_lock);
+       lck_mtx_destroy(&task->lock, &task_lck_grp);
+
+       if (!ledger_get_entries(task->ledger, task_ledgers.tkm_private, &credit,
+           &debit)) {
+               OSAddAtomic64(credit, (int64_t *)&tasks_tkm_private.alloc);
+               OSAddAtomic64(debit, (int64_t *)&tasks_tkm_private.free);
+       }
+       if (!ledger_get_entries(task->ledger, task_ledgers.tkm_shared, &credit,
+           &debit)) {
+               OSAddAtomic64(credit, (int64_t *)&tasks_tkm_shared.alloc);
+               OSAddAtomic64(debit, (int64_t *)&tasks_tkm_shared.free);
+       }
+       ledger_dereference(task->ledger);
+
+#if TASK_REFERENCE_LEAK_DEBUG
+       btlog_remove_entries_for_element(task_ref_btlog, task);
+#endif
+
+       counter_free(&task->faults);
+
+#if CONFIG_COALITIONS
+       task_release_coalitions(task);
+#endif /* CONFIG_COALITIONS */
+
+       bzero(task->coalition, sizeof(task->coalition));
+
+#if MACH_BSD
+       /* clean up collected information since last reference to task is gone */
+       if (task->corpse_info) {
+               void *corpse_info_kernel = kcdata_memory_get_begin_addr(task->corpse_info);
+               task_crashinfo_destroy(task->corpse_info);
+               task->corpse_info = NULL;
+               if (corpse_info_kernel) {
+                       kheap_free(KHEAP_DATA_BUFFERS, corpse_info_kernel,
+                           CORPSEINFO_ALLOCATION_SIZE);
+               }
+       }
+#endif
+
+#if CONFIG_MACF
+       if (task->crash_label) {
+               mac_exc_free_label(task->crash_label);
+               task->crash_label = NULL;
+       }
+#endif
+
+       assert(queue_empty(&task->task_objq));
+       task_objq_lock_destroy(task);
+
+       zfree(task_zone, task);
+}
+
+/*
+ *     task_name_deallocate:
+ *
+ *     Drop a reference on a task name.
+ */
+void
+task_name_deallocate(
+       task_name_t             task_name)
+{
+       return task_deallocate((task_t)task_name);
+}
+
+/*
+ *     task_policy_set_deallocate:
+ *
+ *     Drop a reference on a task type.
+ */
+void
+task_policy_set_deallocate(task_policy_set_t task_policy_set)
+{
+       return task_deallocate((task_t)task_policy_set);
+}
+
+/*
+ *     task_policy_get_deallocate:
+ *
+ *     Drop a reference on a task type.
+ */
+void
+task_policy_get_deallocate(task_policy_get_t task_policy_get)
+{
+       return task_deallocate((task_t)task_policy_get);
+}
+
+/*
+ *     task_inspect_deallocate:
+ *
+ *     Drop a task inspection reference.
+ */
+void
+task_inspect_deallocate(
+       task_inspect_t          task_inspect)
+{
+       return task_deallocate((task_t)task_inspect);
+}
+
+/*
+ *     task_read_deallocate:
+ *
+ *     Drop a reference on task read port.
+ */
+void
+task_read_deallocate(
+       task_read_t          task_read)
+{
+       return task_deallocate((task_t)task_read);
+}
+
+/*
+ *     task_suspension_token_deallocate:
+ *
+ *     Drop a reference on a task suspension token.
+ */
+void
+task_suspension_token_deallocate(
+       task_suspension_token_t         token)
+{
+       return task_deallocate((task_t)token);
+}
+
+
+/*
+ * task_collect_crash_info:
+ *
+ * collect crash info from bsd and mach based data
+ */
+kern_return_t
+task_collect_crash_info(
+       task_t task,
+#ifdef CONFIG_MACF
+       struct label *crash_label,
+#endif
+       int is_corpse_fork)
+{
+       kern_return_t kr = KERN_SUCCESS;
+
+       kcdata_descriptor_t crash_data = NULL;
+       kcdata_descriptor_t crash_data_release = NULL;
+       mach_msg_type_number_t size = CORPSEINFO_ALLOCATION_SIZE;
+       mach_vm_offset_t crash_data_ptr = 0;
+       void *crash_data_kernel = NULL;
+       void *crash_data_kernel_release = NULL;
+#if CONFIG_MACF
+       struct label *label, *free_label;
+#endif
+
+       if (!corpses_enabled()) {
+               return KERN_NOT_SUPPORTED;
+       }
+
+#if CONFIG_MACF
+       free_label = label = mac_exc_create_label();
+#endif
+
+       task_lock(task);
+
+       assert(is_corpse_fork || task->bsd_info != NULL);
+       if (task->corpse_info == NULL && (is_corpse_fork || task->bsd_info != NULL)) {
+#if CONFIG_MACF
+               /* Set the crash label, used by the exception delivery mac hook */
+               free_label = task->crash_label; // Most likely NULL.
+               task->crash_label = label;
+               mac_exc_update_task_crash_label(task, crash_label);
+#endif
+               task_unlock(task);
+
+               crash_data_kernel = kheap_alloc(KHEAP_DATA_BUFFERS,
+                   CORPSEINFO_ALLOCATION_SIZE, Z_WAITOK | Z_ZERO);
+               if (crash_data_kernel == NULL) {
+                       kr = KERN_RESOURCE_SHORTAGE;
+                       goto out_no_lock;
+               }
+               crash_data_ptr = (mach_vm_offset_t) crash_data_kernel;
+
+               /* Do not get a corpse ref for corpse fork */
+               crash_data = task_crashinfo_alloc_init((mach_vm_address_t)crash_data_ptr, size,
+                   is_corpse_fork ? 0 : CORPSE_CRASHINFO_HAS_REF,
+                   KCFLAG_USE_MEMCOPY);
+               if (crash_data) {
+                       task_lock(task);
+                       crash_data_release = task->corpse_info;
+                       crash_data_kernel_release = kcdata_memory_get_begin_addr(crash_data_release);
+                       task->corpse_info = crash_data;
+
+                       task_unlock(task);
+                       kr = KERN_SUCCESS;
+               } else {
+                       kheap_free(KHEAP_DATA_BUFFERS, crash_data_kernel,
+                           CORPSEINFO_ALLOCATION_SIZE);
+                       kr = KERN_FAILURE;
+               }
+
+               if (crash_data_release != NULL) {
+                       task_crashinfo_destroy(crash_data_release);
+               }
+               if (crash_data_kernel_release != NULL) {
+                       kheap_free(KHEAP_DATA_BUFFERS, crash_data_kernel_release,
+                           CORPSEINFO_ALLOCATION_SIZE);
+               }
+       } else {
+               task_unlock(task);
+       }
+
+out_no_lock:
+#if CONFIG_MACF
+       if (free_label != NULL) {
+               mac_exc_free_label(free_label);
+       }
+#endif
+       return kr;
+}
+
+/*
+ * task_deliver_crash_notification:
+ *
+ * Makes outcall to registered host port for a corpse.
+ */
+kern_return_t
+task_deliver_crash_notification(
+       task_t task,
+       thread_t thread,
+       exception_type_t etype,
+       mach_exception_subcode_t subcode)
+{
+       kcdata_descriptor_t crash_info = task->corpse_info;
+       thread_t th_iter = NULL;
+       kern_return_t kr = KERN_SUCCESS;
+       wait_interrupt_t wsave;
+       mach_exception_data_type_t code[EXCEPTION_CODE_MAX];
+       ipc_port_t task_port, old_notify;
+
+       if (crash_info == NULL) {
+               return KERN_FAILURE;
+       }
+
+       task_lock(task);
+       if (task_is_a_corpse_fork(task)) {
+               /* Populate code with EXC_{RESOURCE,GUARD} for corpse fork */
+               code[0] = etype;
+               code[1] = subcode;
+       } else {
+               /* Populate code with EXC_CRASH for corpses */
+               code[0] = EXC_CRASH;
+               code[1] = 0;
+               /* Update the code[1] if the boot-arg corpse_for_fatal_memkill is set */
+               if (corpse_for_fatal_memkill) {
+                       code[1] = subcode;
+               }
+       }
+
+       queue_iterate(&task->threads, th_iter, thread_t, task_threads)
+       {
+               if (th_iter->corpse_dup == FALSE) {
+                       ipc_thread_reset(th_iter);
+               }
+       }
+       task_unlock(task);
+
+       /* Arm the no-sender notification for taskport */
+       task_reference(task);
+       task_port = convert_task_to_port(task);
+       ip_lock(task_port);
+       require_ip_active(task_port);
+       ipc_port_nsrequest(task_port, task_port->ip_mscount, ipc_port_make_sonce_locked(task_port), &old_notify);
+       /* port unlocked */
+       assert(IP_NULL == old_notify);
+
+       wsave = thread_interrupt_level(THREAD_UNINT);
+       kr = exception_triage_thread(EXC_CORPSE_NOTIFY, code, EXCEPTION_CODE_MAX, thread);
+       if (kr != KERN_SUCCESS) {
+               printf("Failed to send exception EXC_CORPSE_NOTIFY. error code: %d for pid %d\n", kr, task_pid(task));
+       }
+
+       (void)thread_interrupt_level(wsave);
+
+       /*
+        * Drop the send right on task port, will fire the
+        * no-sender notification if exception deliver failed.
+        */
+       ipc_port_release_send(task_port);
+       return kr;
+}
+
+/*
+ *     task_terminate:
+ *
+ *     Terminate the specified task.  See comments on thread_terminate
+ *     (kern/thread.c) about problems with terminating the "current task."
+ */
+
+kern_return_t
+task_terminate(
+       task_t          task)
+{
+       if (task == TASK_NULL) {
+               return KERN_INVALID_ARGUMENT;
+       }
+
+       if (task->bsd_info) {
+               return KERN_FAILURE;
+       }
+
+       return task_terminate_internal(task);
+}
+
+#if MACH_ASSERT
+extern int proc_pid(struct proc *);
+extern void proc_name_kdp(task_t t, char *buf, int size);
+#endif /* MACH_ASSERT */
+
+#define VM_MAP_PARTIAL_REAP 0x54  /* 0x150 */
+static void
+__unused task_partial_reap(task_t task, __unused int pid)
+{
+       unsigned int    reclaimed_resident = 0;
+       unsigned int    reclaimed_compressed = 0;
+       uint64_t        task_page_count;
+
+       task_page_count = (get_task_phys_footprint(task) / PAGE_SIZE_64);
+
+       KERNEL_DEBUG_CONSTANT((MACHDBG_CODE(DBG_MACH_VM, VM_MAP_PARTIAL_REAP) | DBG_FUNC_START),
+           pid, task_page_count, 0, 0, 0);
+
+       vm_map_partial_reap(task->map, &reclaimed_resident, &reclaimed_compressed);
+
+       KERNEL_DEBUG_CONSTANT((MACHDBG_CODE(DBG_MACH_VM, VM_MAP_PARTIAL_REAP) | DBG_FUNC_END),
+           pid, reclaimed_resident, reclaimed_compressed, 0, 0);
+}
+
+kern_return_t
+task_mark_corpse(task_t task)
+{
+       kern_return_t kr = KERN_SUCCESS;
+       thread_t self_thread;
+       (void) self_thread;
+       wait_interrupt_t wsave;
+#if CONFIG_MACF
+       struct label *crash_label = NULL;
+#endif
+
+       assert(task != kernel_task);
+       assert(task == current_task());
+       assert(!task_is_a_corpse(task));
+
+#if CONFIG_MACF
+       crash_label = mac_exc_create_label_for_proc((struct proc*)task->bsd_info);
+#endif
+
+       kr = task_collect_crash_info(task,
+#if CONFIG_MACF
+           crash_label,
+#endif
+           FALSE);
+       if (kr != KERN_SUCCESS) {
+               goto out;
+       }
+
+       self_thread = current_thread();
+
+       wsave = thread_interrupt_level(THREAD_UNINT);
+       task_lock(task);
+
+       task_set_corpse_pending_report(task);
+       task_set_corpse(task);
+       task->crashed_thread_id = thread_tid(self_thread);
+
+       kr = task_start_halt_locked(task, TRUE);
+       assert(kr == KERN_SUCCESS);
+
+       ipc_task_reset(task);
+       /* Remove the naked send right for task port, needed to arm no sender notification */
+       task_set_special_port_internal(task, TASK_KERNEL_PORT, IPC_PORT_NULL);
+       ipc_task_enable(task);
+
+       task_unlock(task);
+       /* terminate the ipc space */
+       ipc_space_terminate(task->itk_space);
+
+       /* Add it to global corpse task list */
+       task_add_to_corpse_task_list(task);
+
+       task_start_halt(task);
+       thread_terminate_internal(self_thread, TH_TERMINATE_OPTION_NONE);
+
+       (void) thread_interrupt_level(wsave);
+       assert(task->halting == TRUE);
+
+out:
+#if CONFIG_MACF
+       mac_exc_free_label(crash_label);
+#endif
+       return kr;
+}
+
+/*
+ *     task_clear_corpse
+ *
+ *     Clears the corpse pending bit on task.
+ *     Removes inspection bit on the threads.
+ */
+void
+task_clear_corpse(task_t task)
+{
+       thread_t th_iter = NULL;
+
+       task_lock(task);
+       queue_iterate(&task->threads, th_iter, thread_t, task_threads)
+       {
+               thread_mtx_lock(th_iter);
+               th_iter->inspection = FALSE;
+               ipc_thread_disable(th_iter);
+               thread_mtx_unlock(th_iter);
+       }
+
+       thread_terminate_crashed_threads();
+       /* remove the pending corpse report flag */
+       task_clear_corpse_pending_report(task);
+
+       task_unlock(task);
+}
+
+/*
+ *     task_port_notify
+ *
+ *     Called whenever the Mach port system detects no-senders on
+ *     the task port of a corpse.
+ *     Each notification that comes in should terminate the task (corpse).
+ */
+void
+task_port_notify(mach_msg_header_t *msg)
+{
+       mach_no_senders_notification_t *notification = (void *)msg;
+       ipc_port_t port = notification->not_header.msgh_remote_port;
+       task_t task;
+
+       require_ip_active(port);
+       assert(IKOT_TASK_CONTROL == ip_kotype(port));
+       task = (task_t) ip_get_kobject(port);
+
+       assert(task_is_a_corpse(task));
+
+       /* Remove the task from global corpse task list */
+       task_remove_from_corpse_task_list(task);
+
+       task_clear_corpse(task);
+       task_terminate_internal(task);
+}
+
+/*
+ *     task_port_with_flavor_notify
+ *
+ *     Called whenever the Mach port system detects no-senders on
+ *     the task inspect or read port. These ports are allocated lazily and
+ *     should be deallocated here when there are no senders remaining.
+ */
+void
+task_port_with_flavor_notify(mach_msg_header_t *msg)
+{
+       mach_no_senders_notification_t *notification = (void *)msg;
+       ipc_port_t port = notification->not_header.msgh_remote_port;
+       task_t task;
+       mach_task_flavor_t flavor;
+       ipc_kobject_type_t kotype;
+
+       ip_lock(port);
+       if (port->ip_srights > 0) {
+               ip_unlock(port);
+               return;
+       }
+       task = (task_t)ipc_kobject_get(port);
+       kotype = ip_kotype(port);
+       if (task != TASK_NULL) {
+               assert((IKOT_TASK_READ == kotype) || (IKOT_TASK_INSPECT == kotype));
+               task_reference_internal(task);
+       }
+       ip_unlock(port);
+
+       if (task == TASK_NULL) {
+               /* The task is exiting or disabled; it will eventually deallocate the port */
+               return;
+       }
+
+       if (kotype == IKOT_TASK_READ) {
+               flavor = TASK_FLAVOR_READ;
+       } else {
+               flavor = TASK_FLAVOR_INSPECT;
+       }
+
+       itk_lock(task);
+       ip_lock(port);
+       /*
+        * If the port is no longer active, then ipc_task_terminate() ran
+        * and destroyed the kobject already. Just deallocate the task
+        * ref we took and go away.
+        *
+        * It is also possible that several nsrequests are in flight,
+        * only one shall NULL-out the port entry, and this is the one
+        * that gets to dealloc the port.
+        *
+        * Check for a stale no-senders notification. A call to any function
+        * that vends out send rights to this port could resurrect it between
+        * this notification being generated and actually being handled here.
+        */
+       if (!ip_active(port) ||
+           task->itk_task_ports[flavor] != port ||
+           port->ip_srights > 0) {
+               ip_unlock(port);
+               itk_unlock(task);
+               task_deallocate(task);
+               return;
+       }
+
+       assert(task->itk_task_ports[flavor] == port);
+       task->itk_task_ports[flavor] = IP_NULL;
+
+       ipc_kobject_set_atomically(port, IKO_NULL, IKOT_NONE);
+       ip_unlock(port);
+       itk_unlock(task);
+       task_deallocate(task);
+
+       ipc_port_dealloc_kernel(port);
+}
+
+/*
+ *     task_wait_till_threads_terminate_locked
+ *
+ *     Wait till all the threads in the task are terminated.
+ *     Might release the task lock and re-acquire it.
+ */
+void
+task_wait_till_threads_terminate_locked(task_t task)
+{
+       /* wait for all the threads in the task to terminate */
+       while (task->active_thread_count != 0) {
+               assert_wait((event_t)&task->active_thread_count, THREAD_UNINT);
+               task_unlock(task);
+               thread_block(THREAD_CONTINUE_NULL);
+
+               task_lock(task);
+       }
+}
+
+/*
+ *     task_duplicate_map_and_threads
+ *
+ *     Copy vmmap of source task.
+ *     Copy active threads from source task to destination task.
+ *     Source task would be suspended during the copy.
+ */
+kern_return_t
+task_duplicate_map_and_threads(
+       task_t task,
+       void *p,
+       task_t new_task,
+       thread_t *thread_ret,
+       uint64_t **udata_buffer,
+       int *size,
+       int *num_udata)
+{
+       kern_return_t kr = KERN_SUCCESS;
+       int active;
+       thread_t thread, self, thread_return = THREAD_NULL;
+       thread_t new_thread = THREAD_NULL, first_thread = THREAD_NULL;
+       thread_t *thread_array;
+       uint32_t active_thread_count = 0, array_count = 0, i;
+       vm_map_t oldmap;
+       uint64_t *buffer = NULL;
+       int buf_size = 0;
+       int est_knotes = 0, num_knotes = 0;
+
+       self = current_thread();
+
+       /*
+        * Suspend the task to copy thread state, use the internal
+        * variant so that no user-space process can resume
+        * the task from under us
+        */
+       kr = task_suspend_internal(task);
+       if (kr != KERN_SUCCESS) {
+               return kr;
+       }
+
+       if (task->map->disable_vmentry_reuse == TRUE) {
+               /*
+                * Quite likely GuardMalloc (or some debugging tool)
+                * is being used on this task. And it has gone through
+                * its limit. Making a corpse will likely encounter
+                * a lot of VM entries that will need COW.
+                *
+                * Skip it.
+                */
+#if DEVELOPMENT || DEBUG
+               memorystatus_abort_vm_map_fork(task);
+#endif
+               task_resume_internal(task);
+               return KERN_FAILURE;
+       }
+
+       /* Check with VM if vm_map_fork is allowed for this task */
+       if (memorystatus_allowed_vm_map_fork(task)) {
+               /* Setup new task's vmmap, switch from parent task's map to it COW map */
+               oldmap = new_task->map;
+               new_task->map = vm_map_fork(new_task->ledger,
+                   task->map,
+                   (VM_MAP_FORK_SHARE_IF_INHERIT_NONE |
+                   VM_MAP_FORK_PRESERVE_PURGEABLE |
+                   VM_MAP_FORK_CORPSE_FOOTPRINT));
+               vm_map_deallocate(oldmap);
+
+               /* copy ledgers that impact the memory footprint */
+               vm_map_copy_footprint_ledgers(task, new_task);
+
+               /* Get all the udata pointers from kqueue */
+               est_knotes = kevent_proc_copy_uptrs(p, NULL, 0);
+               if (est_knotes > 0) {
+                       buf_size = (est_knotes + 32) * sizeof(uint64_t);
+                       buffer = kheap_alloc(KHEAP_DATA_BUFFERS, buf_size, Z_WAITOK);
+                       num_knotes = kevent_proc_copy_uptrs(p, buffer, buf_size);
+                       if (num_knotes > est_knotes + 32) {
+                               num_knotes = est_knotes + 32;
+                       }
+               }
+       }
+
+       active_thread_count = task->active_thread_count;
+       if (active_thread_count == 0) {
+               if (buffer != NULL) {
+                       kheap_free(KHEAP_DATA_BUFFERS, buffer, buf_size);
+               }
+               task_resume_internal(task);
+               return KERN_FAILURE;
+       }
+
+       thread_array = kheap_alloc(KHEAP_TEMP,
+           sizeof(thread_t) * active_thread_count, Z_WAITOK);
+
+       /* Iterate all the threads and drop the task lock before calling thread_create_with_continuation */
+       task_lock(task);
+       queue_iterate(&task->threads, thread, thread_t, task_threads) {
+               /* Skip inactive threads */
+               active = thread->active;
+               if (!active) {
+                       continue;
+               }
+
+               if (array_count >= active_thread_count) {
+                       break;
+               }
+
+               thread_array[array_count++] = thread;
+               thread_reference(thread);
+       }
+       task_unlock(task);
+
+       for (i = 0; i < array_count; i++) {
+               kr = thread_create_with_continuation(new_task, &new_thread, (thread_continue_t)thread_corpse_continue);
+               if (kr != KERN_SUCCESS) {
+                       break;
+               }
+
+               /* Equivalent of current thread in corpse */
+               if (thread_array[i] == self) {
+                       thread_return = new_thread;
+                       new_task->crashed_thread_id = thread_tid(new_thread);
+               } else if (first_thread == NULL) {
+                       first_thread = new_thread;
+               } else {
+                       /* drop the extra ref returned by thread_create_with_continuation */
+                       thread_deallocate(new_thread);
+               }
+
+               kr = thread_dup2(thread_array[i], new_thread);
+               if (kr != KERN_SUCCESS) {
+                       thread_mtx_lock(new_thread);
+                       new_thread->corpse_dup = TRUE;
+                       thread_mtx_unlock(new_thread);
+                       continue;
+               }
+
+               /* Copy thread name */
+               bsd_copythreadname(new_thread->uthread, thread_array[i]->uthread);
+               new_thread->thread_tag = thread_array[i]->thread_tag;
+               thread_copy_resource_info(new_thread, thread_array[i]);
+       }
+
+       /* return the first thread if we couldn't find the equivalent of current */
+       if (thread_return == THREAD_NULL) {
+               thread_return = first_thread;
+       } else if (first_thread != THREAD_NULL) {
+               /* drop the extra ref returned by thread_create_with_continuation */
+               thread_deallocate(first_thread);
+       }
+
+       task_resume_internal(task);
+
+       for (i = 0; i < array_count; i++) {
+               thread_deallocate(thread_array[i]);
+       }
+       kheap_free(KHEAP_TEMP, thread_array, sizeof(thread_t) * active_thread_count);
+
+       if (kr == KERN_SUCCESS) {
+               *thread_ret = thread_return;
+               *udata_buffer = buffer;
+               *size = buf_size;
+               *num_udata = num_knotes;
+       } else {
+               if (thread_return != THREAD_NULL) {
+                       thread_deallocate(thread_return);
+               }
+               if (buffer != NULL) {
+                       kheap_free(KHEAP_DATA_BUFFERS, buffer, buf_size);
+               }
+       }
+
+       return kr;
+}
+
+#if CONFIG_SECLUDED_MEMORY
+extern void task_set_can_use_secluded_mem_locked(
+       task_t          task,
+       boolean_t       can_use_secluded_mem);
+#endif /* CONFIG_SECLUDED_MEMORY */
+
+#if MACH_ASSERT
+int debug4k_panic_on_terminate = 0;
+#endif /* MACH_ASSERT */
+kern_return_t
+task_terminate_internal(
+       task_t                  task)
+{
+       thread_t                        thread, self;
+       task_t                          self_task;
+       boolean_t                       interrupt_save;
+       int                             pid = 0;
+
+       assert(task != kernel_task);
+
+       self = current_thread();
+       self_task = self->task;
+
+       /*
+        *      Get the task locked and make sure that we are not racing
+        *      with someone else trying to terminate us.
+        */
+       if (task == self_task) {
+               task_lock(task);
+       } else if (task < self_task) {
+               task_lock(task);
+               task_lock(self_task);
+       } else {
+               task_lock(self_task);
+               task_lock(task);
+       }
+
+#if CONFIG_SECLUDED_MEMORY
+       if (task->task_can_use_secluded_mem) {
+               task_set_can_use_secluded_mem_locked(task, FALSE);
+       }
+       task->task_could_use_secluded_mem = FALSE;
+       task->task_could_also_use_secluded_mem = FALSE;
+
+       if (task->task_suppressed_secluded) {
+               stop_secluded_suppression(task);
+       }
+#endif /* CONFIG_SECLUDED_MEMORY */
+
+       if (!task->active) {
+               /*
+                *      Task is already being terminated.
+                *      Just return an error. If we are dying, this will
+                *      just get us to our AST special handler and that
+                *      will get us to finalize the termination of ourselves.
+                */
+               task_unlock(task);
+               if (self_task != task) {
+                       task_unlock(self_task);
+               }
+
+               return KERN_FAILURE;
+       }
+
+       if (task_corpse_pending_report(task)) {
+               /*
+                *      Task is marked for reporting as corpse.
+                *      Just return an error. This will
+                *      just get us to our AST special handler and that
+                *      will get us to finish the path to death
+                */
+               task_unlock(task);
+               if (self_task != task) {
+                       task_unlock(self_task);
+               }
+
+               return KERN_FAILURE;
+       }
+
+       if (self_task != task) {
+               task_unlock(self_task);
+       }
+
+       /*
+        * Make sure the current thread does not get aborted out of
+        * the waits inside these operations.
+        */
+       interrupt_save = thread_interrupt_level(THREAD_UNINT);
+
+       /*
+        *      Indicate that we want all the threads to stop executing
+        *      at user space by holding the task (we would have held
+        *      each thread independently in thread_terminate_internal -
+        *      but this way we may be more likely to already find it
+        *      held there).  Mark the task inactive, and prevent
+        *      further task operations via the task port.
+        */
+       task_hold_locked(task);
+       task->active = FALSE;
+       ipc_task_disable(task);
+
+#if CONFIG_TELEMETRY
+       /*
+        * Notify telemetry that this task is going away.
+        */
+       telemetry_task_ctl_locked(task, TF_TELEMETRY, 0);
+#endif
+
+       /*
+        *      Terminate each thread in the task.
+        */
+       queue_iterate(&task->threads, thread, thread_t, task_threads) {
+               thread_terminate_internal(thread, TH_TERMINATE_OPTION_NONE);
+       }
+
+#ifdef MACH_BSD
+       if (task->bsd_info != NULL && !task_is_exec_copy(task)) {
+               pid = proc_pid(task->bsd_info);
+       }
+#endif /* MACH_BSD */
+
+       task_unlock(task);
+
+       proc_set_task_policy(task, TASK_POLICY_ATTRIBUTE,
+           TASK_POLICY_TERMINATED, TASK_POLICY_ENABLE);
+
+       /* Early object reap phase */
+
+// PR-17045188: Revisit implementation
+//        task_partial_reap(task, pid);
+
+#if CONFIG_TASKWATCH
+       /*
+        * remove all task watchers
+        */
+       task_removewatchers(task);
+
+#endif /* CONFIG_TASKWATCH */
+
+       /*
+        *      Destroy all synchronizers owned by the task.
+        */
+       task_synchronizer_destroy_all(task);
+
+       /*
+        *      Clear the watchport boost on the task.
+        */
+       task_remove_turnstile_watchports(task);
+
+       /*
+        *      Destroy the IPC space, leaving just a reference for it.
+        */
+       ipc_space_terminate(task->itk_space);
+
+#if 00
+       /* if some ledgers go negative on tear-down again... */
+       ledger_disable_panic_on_negative(task->map->pmap->ledger,
+           task_ledgers.phys_footprint);
+       ledger_disable_panic_on_negative(task->map->pmap->ledger,
+           task_ledgers.internal);
+       ledger_disable_panic_on_negative(task->map->pmap->ledger,
+           task_ledgers.internal_compressed);
+       ledger_disable_panic_on_negative(task->map->pmap->ledger,
+           task_ledgers.iokit_mapped);
+       ledger_disable_panic_on_negative(task->map->pmap->ledger,
+           task_ledgers.alternate_accounting);
+       ledger_disable_panic_on_negative(task->map->pmap->ledger,
+           task_ledgers.alternate_accounting_compressed);
+#endif
+
+       /*
+        * If the current thread is a member of the task
+        * being terminated, then the last reference to
+        * the task will not be dropped until the thread
+        * is finally reaped.  To avoid incurring the
+        * expense of removing the address space regions
+        * at reap time, we do it explictly here.
+        */
+
+       vm_map_lock(task->map);
+       vm_map_disable_hole_optimization(task->map);
+       vm_map_unlock(task->map);
+
+#if MACH_ASSERT
+       /*
+        * Identify the pmap's process, in case the pmap ledgers drift
+        * and we have to report it.
+        */
+       char procname[17];
+       if (task->bsd_info && !task_is_exec_copy(task)) {
+               pid = proc_pid(task->bsd_info);
+               proc_name_kdp(task, procname, sizeof(procname));
+       } else {
+               pid = 0;
+               strlcpy(procname, "<unknown>", sizeof(procname));
+       }
+       pmap_set_process(task->map->pmap, pid, procname);
+       if (vm_map_page_shift(task->map) < (int)PAGE_SHIFT) {
+               DEBUG4K_LIFE("map %p procname: %s\n", task->map, procname);
+               if (debug4k_panic_on_terminate) {
+                       panic("DEBUG4K: %s:%d %d[%s] map %p\n", __FUNCTION__, __LINE__, pid, procname, task->map);
+               }
+       }
+#endif /* MACH_ASSERT */
+
+       vm_map_terminate(task->map);
+
+       /* release our shared region */
+       vm_shared_region_set(task, NULL);
+
+#if __has_feature(ptrauth_calls)
+       task_set_shared_region_id(task, NULL);
+#endif /* __has_feature(ptrauth_calls) */
+
+       lck_mtx_lock(&tasks_threads_lock);
+       queue_remove(&tasks, task, task_t, tasks);
+       queue_enter(&terminated_tasks, task, task_t, tasks);
+       tasks_count--;
+       terminated_tasks_count++;
+       lck_mtx_unlock(&tasks_threads_lock);
+
+       /*
+        * We no longer need to guard against being aborted, so restore
+        * the previous interruptible state.
+        */
+       thread_interrupt_level(interrupt_save);
+
+#if KPC
+       /* force the task to release all ctrs */
+       if (task->t_kpc & TASK_KPC_FORCED_ALL_CTRS) {
+               kpc_force_all_ctrs(task, 0);
+       }
+#endif /* KPC */
+
+#if CONFIG_COALITIONS
+       /*
+        * Leave our coalitions. (drop activation but not reference)
+        */
+       coalitions_remove_task(task);
+#endif
+
+#if CONFIG_FREEZE
+       extern int      vm_compressor_available;
+       if (VM_CONFIG_FREEZER_SWAP_IS_ACTIVE && vm_compressor_available) {
+               task_disown_frozen_csegs(task);
+               assert(queue_empty(&task->task_frozen_cseg_q));
+       }
+#endif /* CONFIG_FREEZE */
+
+       /*
+        * Get rid of the task active reference on itself.
+        */
+       task_deallocate(task);
+
+       return KERN_SUCCESS;
+}
+
+void
+tasks_system_suspend(boolean_t suspend)
+{
+       task_t task;
+
+       lck_mtx_lock(&tasks_threads_lock);
+       assert(tasks_suspend_state != suspend);
+       tasks_suspend_state = suspend;
+       queue_iterate(&tasks, task, task_t, tasks) {
+               if (task == kernel_task) {
+                       continue;
+               }
+               suspend ? task_suspend_internal(task) : task_resume_internal(task);
+       }
+       lck_mtx_unlock(&tasks_threads_lock);
+}
+
+/*
+ * task_start_halt:
+ *
+ *      Shut the current task down (except for the current thread) in
+ *     preparation for dramatic changes to the task (probably exec).
+ *     We hold the task and mark all other threads in the task for
+ *     termination.
+ */
+kern_return_t
+task_start_halt(task_t task)
+{
+       kern_return_t kr = KERN_SUCCESS;
+       task_lock(task);
+       kr = task_start_halt_locked(task, FALSE);
+       task_unlock(task);
+       return kr;
+}
+
+static kern_return_t
+task_start_halt_locked(task_t task, boolean_t should_mark_corpse)
+{
+       thread_t thread, self;
+       uint64_t dispatchqueue_offset;
+
+       assert(task != kernel_task);
+
+       self = current_thread();
+
+       if (task != self->task && !task_is_a_corpse_fork(task)) {
+               return KERN_INVALID_ARGUMENT;
+       }
+
+       if (task->halting || !task->active || !self->active) {
+               /*
+                * Task or current thread is already being terminated.
+                * Hurry up and return out of the current kernel context
+                * so that we run our AST special handler to terminate
+                * ourselves.
+                */
+               return KERN_FAILURE;
+       }
+
+       task->halting = TRUE;
+
+       /*
+        * Mark all the threads to keep them from starting any more
+        * user-level execution.  The thread_terminate_internal code
+        * would do this on a thread by thread basis anyway, but this
+        * gives us a better chance of not having to wait there.
+        */
+       task_hold_locked(task);
+       dispatchqueue_offset = get_dispatchqueue_offset_from_proc(task->bsd_info);
+
+       /*
+        * Terminate all the other threads in the task.
+        */
+       queue_iterate(&task->threads, thread, thread_t, task_threads)
+       {
+               if (should_mark_corpse) {
+                       thread_mtx_lock(thread);
+                       thread->inspection = TRUE;
+                       thread_mtx_unlock(thread);
+               }
+               if (thread != self) {
+                       thread_terminate_internal(thread, TH_TERMINATE_OPTION_NONE);
+               }
+       }
+       task->dispatchqueue_offset = dispatchqueue_offset;
+
+       task_release_locked(task);
+
+       return KERN_SUCCESS;
+}
+
+
+/*
+ * task_complete_halt:
+ *
+ *     Complete task halt by waiting for threads to terminate, then clean
+ *     up task resources (VM, port namespace, etc...) and then let the
+ *     current thread go in the (practically empty) task context.
+ *
+ *     Note: task->halting flag is not cleared in order to avoid creation
+ *     of new thread in old exec'ed task.
+ */
+void
+task_complete_halt(task_t task)
+{
+       task_lock(task);
+       assert(task->halting);
+       assert(task == current_task());
+
+       /*
+        *      Wait for the other threads to get shut down.
+        *      When the last other thread is reaped, we'll be
+        *      woken up.
+        */
+       if (task->thread_count > 1) {
+               assert_wait((event_t)&task->halting, THREAD_UNINT);
+               task_unlock(task);
+               thread_block(THREAD_CONTINUE_NULL);
+       } else {
+               task_unlock(task);
+       }
+
+       /*
+        *      Give the machine dependent code a chance
+        *      to perform cleanup of task-level resources
+        *      associated with the current thread before
+        *      ripping apart the task.
+        */
+       machine_task_terminate(task);
+
+       /*
+        *      Destroy all synchronizers owned by the task.
+        */
+       task_synchronizer_destroy_all(task);
+
+       /*
+        *      Destroy the contents of the IPC space, leaving just
+        *      a reference for it.
+        */
+       ipc_space_clean(task->itk_space);
+
+       /*
+        * Clean out the address space, as we are going to be
+        * getting a new one.
+        */
+       vm_map_remove(task->map, task->map->min_offset,
+           task->map->max_offset,
+           /*
+            * Final cleanup:
+            * + no unnesting
+            * + remove immutable mappings
+            * + allow gaps in the range
+            */
+           (VM_MAP_REMOVE_NO_UNNESTING |
+           VM_MAP_REMOVE_IMMUTABLE |
+           VM_MAP_REMOVE_GAPS_OK));
+
+       /*
+        * Kick out any IOKitUser handles to the task. At best they're stale,
+        * at worst someone is racing a SUID exec.
+        */
+       iokit_task_terminate(task);
+}
+
+/*
+ *     task_hold_locked:
+ *
+ *     Suspend execution of the specified task.
+ *     This is a recursive-style suspension of the task, a count of
+ *     suspends is maintained.
+ *
+ *     CONDITIONS: the task is locked and active.
+ */
+void
+task_hold_locked(
+       task_t          task)
+{
+       thread_t        thread;
+
+       assert(task->active);
+
+       if (task->suspend_count++ > 0) {
+               return;
+       }
+
+       if (task->bsd_info) {
+               workq_proc_suspended(task->bsd_info);
+       }
+
+       /*
+        *      Iterate through all the threads and hold them.
+        */
+       queue_iterate(&task->threads, thread, thread_t, task_threads) {
+               thread_mtx_lock(thread);
+               thread_hold(thread);
+               thread_mtx_unlock(thread);
+       }
+}
+
+/*
+ *     task_hold:
+ *
+ *     Same as the internal routine above, except that is must lock
+ *     and verify that the task is active.  This differs from task_suspend
+ *     in that it places a kernel hold on the task rather than just a
+ *     user-level hold.  This keeps users from over resuming and setting
+ *     it running out from under the kernel.
+ *
+ *      CONDITIONS: the caller holds a reference on the task
+ */
+kern_return_t
+task_hold(
+       task_t          task)
+{
+       if (task == TASK_NULL) {
+               return KERN_INVALID_ARGUMENT;
+       }
+
+       task_lock(task);
+
+       if (!task->active) {
+               task_unlock(task);
+
+               return KERN_FAILURE;
+       }
+
+       task_hold_locked(task);
+       task_unlock(task);
+
+       return KERN_SUCCESS;
+}
+
+kern_return_t
+task_wait(
+       task_t          task,
+       boolean_t       until_not_runnable)
+{
+       if (task == TASK_NULL) {
+               return KERN_INVALID_ARGUMENT;
+       }
+
+       task_lock(task);
+
+       if (!task->active) {
+               task_unlock(task);
+
+               return KERN_FAILURE;
+       }
+
+       task_wait_locked(task, until_not_runnable);
+       task_unlock(task);
+
+       return KERN_SUCCESS;
+}
+
+/*
+ *     task_wait_locked:
+ *
+ *     Wait for all threads in task to stop.
+ *
+ * Conditions:
+ *     Called with task locked, active, and held.
+ */
+void
+task_wait_locked(
+       task_t          task,
+       boolean_t               until_not_runnable)
+{
+       thread_t        thread, self;
+
+       assert(task->active);
+       assert(task->suspend_count > 0);
+
+       self = current_thread();
+
+       /*
+        *      Iterate through all the threads and wait for them to
+        *      stop.  Do not wait for the current thread if it is within
+        *      the task.
+        */
+       queue_iterate(&task->threads, thread, thread_t, task_threads) {
+               if (thread != self) {
+                       thread_wait(thread, until_not_runnable);
+               }
+       }
+}
+
+boolean_t
+task_is_app_suspended(task_t task)
+{
+       return task->pidsuspended;
+}
+
+/*
+ *     task_release_locked:
+ *
+ *     Release a kernel hold on a task.
+ *
+ *      CONDITIONS: the task is locked and active
+ */
+void
+task_release_locked(
+       task_t          task)
+{
+       thread_t        thread;
+
+       assert(task->active);
+       assert(task->suspend_count > 0);
+
+       if (--task->suspend_count > 0) {
+               return;
+       }
+
+       if (task->bsd_info) {
+               workq_proc_resumed(task->bsd_info);
+       }
+
+       queue_iterate(&task->threads, thread, thread_t, task_threads) {
+               thread_mtx_lock(thread);
+               thread_release(thread);
+               thread_mtx_unlock(thread);
+       }
+}
+
+/*
+ *     task_release:
+ *
+ *     Same as the internal routine above, except that it must lock
+ *     and verify that the task is active.
+ *
+ *      CONDITIONS: The caller holds a reference to the task
+ */
+kern_return_t
+task_release(
+       task_t          task)
+{
+       if (task == TASK_NULL) {
+               return KERN_INVALID_ARGUMENT;
+       }
+
+       task_lock(task);
+
+       if (!task->active) {
+               task_unlock(task);
+
+               return KERN_FAILURE;
+       }
+
+       task_release_locked(task);
+       task_unlock(task);
+
+       return KERN_SUCCESS;
+}
+
+static kern_return_t
+task_threads_internal(
+       task_t                      task,
+       thread_act_array_t         *threads_out,
+       mach_msg_type_number_t     *count,
+       mach_thread_flavor_t        flavor)
+{
+       mach_msg_type_number_t  actual;
+       thread_t                                *thread_list;
+       thread_t                                thread;
+       vm_size_t                               size, size_needed;
+       void                                    *addr;
+       unsigned int                    i, j;
+
+       size = 0; addr = NULL;
+
+       if (task == TASK_NULL) {
+               return KERN_INVALID_ARGUMENT;
+       }
+
+       assert(flavor <= THREAD_FLAVOR_INSPECT);
+
+       for (;;) {
+               task_lock(task);
+               if (!task->active) {
+                       task_unlock(task);
+
+                       if (size != 0) {
+                               kfree(addr, size);
+                       }
+
+                       return KERN_FAILURE;
+               }
+
+               actual = task->thread_count;
+
+               /* do we have the memory we need? */
+               size_needed = actual * sizeof(mach_port_t);
+               if (size_needed <= size) {
+                       break;
+               }
+
+               /* unlock the task and allocate more memory */
+               task_unlock(task);
+
+               if (size != 0) {
+                       kfree(addr, size);
+               }
+
+               assert(size_needed > 0);
+               size = size_needed;
+
+               addr = kalloc(size);
+               if (addr == 0) {
+                       return KERN_RESOURCE_SHORTAGE;
+               }
+       }
+
+       /* OK, have memory and the task is locked & active */
+       thread_list = (thread_t *)addr;
+
+       i = j = 0;
+
+       for (thread = (thread_t)queue_first(&task->threads); i < actual;
+           ++i, thread = (thread_t)queue_next(&thread->task_threads)) {
+               thread_reference_internal(thread);
+               thread_list[j++] = thread;
+       }
+
+       assert(queue_end(&task->threads, (queue_entry_t)thread));
+
+       actual = j;
+       size_needed = actual * sizeof(mach_port_t);
+
+       /* can unlock task now that we've got the thread refs */
+       task_unlock(task);
+
+       if (actual == 0) {
+               /* no threads, so return null pointer and deallocate memory */
+
+               *threads_out = NULL;
+               *count = 0;
+
+               if (size != 0) {
+                       kfree(addr, size);
+               }
+       } else {
+               /* if we allocated too much, must copy */
+
+               if (size_needed < size) {
+                       void *newaddr;
+
+                       newaddr = kalloc(size_needed);
+                       if (newaddr == 0) {
+                               for (i = 0; i < actual; ++i) {
+                                       thread_deallocate(thread_list[i]);
+                               }
+                               kfree(addr, size);
+                               return KERN_RESOURCE_SHORTAGE;
+                       }
+
+                       bcopy(addr, newaddr, size_needed);
+                       kfree(addr, size);
+                       thread_list = (thread_t *)newaddr;
+               }
+
+               *threads_out = thread_list;
+               *count = actual;
+
+               /* do the conversion that Mig should handle */
+
+               switch (flavor) {
+               case THREAD_FLAVOR_CONTROL:
+                       if (task == current_task()) {
+                               for (i = 0; i < actual; ++i) {
+                                       ((ipc_port_t *) thread_list)[i] = convert_thread_to_port_pinned(thread_list[i]);
+                               }
+                       } else {
+                               for (i = 0; i < actual; ++i) {
+                                       ((ipc_port_t *) thread_list)[i] = convert_thread_to_port(thread_list[i]);
+                               }
+                       }
+                       break;
+               case THREAD_FLAVOR_READ:
+                       for (i = 0; i < actual; ++i) {
+                               ((ipc_port_t *) thread_list)[i] = convert_thread_read_to_port(thread_list[i]);
+                       }
+                       break;
+               case THREAD_FLAVOR_INSPECT:
+                       for (i = 0; i < actual; ++i) {
+                               ((ipc_port_t *) thread_list)[i] = convert_thread_inspect_to_port(thread_list[i]);
+                       }
+                       break;
+               }
+       }
+
+       return KERN_SUCCESS;
+}
+
+kern_return_t
+task_threads(
+       task_t                      task,
+       thread_act_array_t         *threads_out,
+       mach_msg_type_number_t     *count)
+{
+       return task_threads_internal(task, threads_out, count, THREAD_FLAVOR_CONTROL);
+}
+
+
+kern_return_t
+task_threads_from_user(
+       mach_port_t                 port,
+       thread_act_array_t         *threads_out,
+       mach_msg_type_number_t     *count)
+{
+       ipc_kobject_type_t kotype;
+       kern_return_t kr;
+
+       task_t task = convert_port_to_task_check_type(port, &kotype, TASK_FLAVOR_INSPECT, FALSE);
+
+       if (task == TASK_NULL) {
+               return KERN_INVALID_ARGUMENT;
+       }
+
+       switch (kotype) {
+       case IKOT_TASK_CONTROL:
+               kr = task_threads_internal(task, threads_out, count, THREAD_FLAVOR_CONTROL);
+               break;
+       case IKOT_TASK_READ:
+               kr = task_threads_internal(task, threads_out, count, THREAD_FLAVOR_READ);
+               break;
+       case IKOT_TASK_INSPECT:
+               kr = task_threads_internal(task, threads_out, count, THREAD_FLAVOR_INSPECT);
+               break;
+       default:
+               panic("strange kobject type");
+               break;
+       }
+
+       task_deallocate(task);
+       return kr;
+}
+
+#define TASK_HOLD_NORMAL        0
+#define TASK_HOLD_PIDSUSPEND    1
+#define TASK_HOLD_LEGACY        2
+#define TASK_HOLD_LEGACY_ALL    3
+
+static kern_return_t
+place_task_hold(
+       task_t task,
+       int mode)
+{
+       if (!task->active && !task_is_a_corpse(task)) {
+               return KERN_FAILURE;
+       }
+
+       /* Return success for corpse task */
+       if (task_is_a_corpse(task)) {
+               return KERN_SUCCESS;
+       }
+
+       KDBG_RELEASE(MACHDBG_CODE(DBG_MACH_IPC, MACH_TASK_SUSPEND),
+           task_pid(task),
+           task->thread_count > 0 ?((thread_t)queue_first(&task->threads))->thread_id : 0,
+           task->user_stop_count, task->user_stop_count + 1);
+
+#if MACH_ASSERT
+       current_task()->suspends_outstanding++;
+#endif
+
+       if (mode == TASK_HOLD_LEGACY) {
+               task->legacy_stop_count++;
+       }
+
+       if (task->user_stop_count++ > 0) {
+               /*
+                *      If the stop count was positive, the task is
+                *      already stopped and we can exit.
+                */
+               return KERN_SUCCESS;
+       }
+
+       /*
+        * Put a kernel-level hold on the threads in the task (all
+        * user-level task suspensions added together represent a
+        * single kernel-level hold).  We then wait for the threads
+        * to stop executing user code.
+        */
+       task_hold_locked(task);
+       task_wait_locked(task, FALSE);
+
+       return KERN_SUCCESS;
+}
+
+static kern_return_t
+release_task_hold(
+       task_t          task,
+       int                     mode)
+{
+       boolean_t release = FALSE;
+
+       if (!task->active && !task_is_a_corpse(task)) {
+               return KERN_FAILURE;
+       }
+
+       /* Return success for corpse task */
+       if (task_is_a_corpse(task)) {
+               return KERN_SUCCESS;
+       }
+
+       if (mode == TASK_HOLD_PIDSUSPEND) {
+               if (task->pidsuspended == FALSE) {
+                       return KERN_FAILURE;
+               }
+               task->pidsuspended = FALSE;
+       }
+
+       if (task->user_stop_count > (task->pidsuspended ? 1 : 0)) {
+               KERNEL_DEBUG_CONSTANT_IST(KDEBUG_TRACE,
+                   MACHDBG_CODE(DBG_MACH_IPC, MACH_TASK_RESUME) | DBG_FUNC_NONE,
+                   task_pid(task), ((thread_t)queue_first(&task->threads))->thread_id,
+                   task->user_stop_count, mode, task->legacy_stop_count);
+
+#if MACH_ASSERT
+               /*
+                * This is obviously not robust; if we suspend one task and then resume a different one,
+                * we'll fly under the radar. This is only meant to catch the common case of a crashed
+                * or buggy suspender.
+                */
+               current_task()->suspends_outstanding--;
+#endif
+
+               if (mode == TASK_HOLD_LEGACY_ALL) {
+                       if (task->legacy_stop_count >= task->user_stop_count) {
+                               task->user_stop_count = 0;
+                               release = TRUE;
+                       } else {
+                               task->user_stop_count -= task->legacy_stop_count;
+                       }
+                       task->legacy_stop_count = 0;
+               } else {
+                       if (mode == TASK_HOLD_LEGACY && task->legacy_stop_count > 0) {
+                               task->legacy_stop_count--;
+                       }
+                       if (--task->user_stop_count == 0) {
+                               release = TRUE;
+                       }
+               }
+       } else {
+               return KERN_FAILURE;
+       }
+
+       /*
+        *      Release the task if necessary.
+        */
+       if (release) {
+               task_release_locked(task);
+       }
+
+       return KERN_SUCCESS;
+}
+
+boolean_t
+get_task_suspended(task_t task)
+{
+       return 0 != task->user_stop_count;
+}
+
+/*
+ *     task_suspend:
+ *
+ *     Implement an (old-fashioned) user-level suspension on a task.
+ *
+ *     Because the user isn't expecting to have to manage a suspension
+ *     token, we'll track it for him in the kernel in the form of a naked
+ *     send right to the task's resume port.  All such send rights
+ *     account for a single suspension against the task (unlike task_suspend2()
+ *     where each caller gets a unique suspension count represented by a
+ *     unique send-once right).
+ *
+ * Conditions:
+ *      The caller holds a reference to the task
+ */
+kern_return_t
+task_suspend(
+       task_t          task)
+{
+       kern_return_t                   kr;
+       mach_port_t                     port;
+       mach_port_name_t                name;
+
+       if (task == TASK_NULL || task == kernel_task) {
+               return KERN_INVALID_ARGUMENT;
+       }
+
+       task_lock(task);
+
+       /*
+        * place a legacy hold on the task.
+        */
+       kr = place_task_hold(task, TASK_HOLD_LEGACY);
+       if (kr != KERN_SUCCESS) {
+               task_unlock(task);
+               return kr;
+       }
+
+       /*
+        * Claim a send right on the task resume port, and request a no-senders
+        * notification on that port (if none outstanding).
+        */
+       (void)ipc_kobject_make_send_lazy_alloc_port((ipc_port_t *) &task->itk_resume,
+           (ipc_kobject_t)task, IKOT_TASK_RESUME, IPC_KOBJECT_ALLOC_NONE, true,
+           OS_PTRAUTH_DISCRIMINATOR("task.itk_resume"));
+       port = task->itk_resume;
+       task_unlock(task);
+
+       /*
+        * Copyout the send right into the calling task's IPC space.  It won't know it is there,
+        * but we'll look it up when calling a traditional resume.  Any IPC operations that
+        * deallocate the send right will auto-release the suspension.
+        */
+       if (IP_VALID(port)) {
+               kr = ipc_object_copyout(current_space(), ip_to_object(port),
+                   MACH_MSG_TYPE_MOVE_SEND, IPC_OBJECT_COPYOUT_FLAGS_NONE,
+                   NULL, NULL, &name);
+       } else {
+               kr = KERN_SUCCESS;
+       }
+       if (kr != KERN_SUCCESS) {
+               printf("warning: %s(%d) failed to copyout suspension "
+                   "token for pid %d with error: %d\n",
+                   proc_name_address(current_task()->bsd_info),
+                   proc_pid(current_task()->bsd_info),
+                   task_pid(task), kr);
+       }
+
+       return kr;
+}
+
+/*
+ *     task_resume:
+ *             Release a user hold on a task.
+ *
+ * Conditions:
+ *             The caller holds a reference to the task
+ */
+kern_return_t
+task_resume(
+       task_t  task)
+{
+       kern_return_t    kr;
+       mach_port_name_t resume_port_name;
+       ipc_entry_t              resume_port_entry;
+       ipc_space_t              space = current_task()->itk_space;
+
+       if (task == TASK_NULL || task == kernel_task) {
+               return KERN_INVALID_ARGUMENT;
+       }
+
+       /* release a legacy task hold */
+       task_lock(task);
+       kr = release_task_hold(task, TASK_HOLD_LEGACY);
+       task_unlock(task);
+
+       is_write_lock(space);
+       if (is_active(space) && IP_VALID(task->itk_resume) &&
+           ipc_hash_lookup(space, ip_to_object(task->itk_resume), &resume_port_name, &resume_port_entry) == TRUE) {
+               /*
+                * We found a suspension token in the caller's IPC space. Release a send right to indicate that
+                * we are holding one less legacy hold on the task from this caller.  If the release failed,
+                * go ahead and drop all the rights, as someone either already released our holds or the task
+                * is gone.
+                */
+               if (kr == KERN_SUCCESS) {
+                       ipc_right_dealloc(space, resume_port_name, resume_port_entry);
+               } else {
+                       ipc_right_destroy(space, resume_port_name, resume_port_entry, FALSE, 0);
+               }
+               /* space unlocked */
+       } else {
+               is_write_unlock(space);
+               if (kr == KERN_SUCCESS) {
+                       printf("warning: %s(%d) performed out-of-band resume on pid %d\n",
+                           proc_name_address(current_task()->bsd_info), proc_pid(current_task()->bsd_info),
+                           task_pid(task));
+               }
+       }
+
+       return kr;
+}
+
+/*
+ * Suspend the target task.
+ * Making/holding a token/reference/port is the callers responsibility.
+ */
+kern_return_t
+task_suspend_internal(task_t task)
+{
+       kern_return_t    kr;
+
+       if (task == TASK_NULL || task == kernel_task) {
+               return KERN_INVALID_ARGUMENT;
+       }
+
+       task_lock(task);
+       kr = place_task_hold(task, TASK_HOLD_NORMAL);
+       task_unlock(task);
+       return kr;
+}
+
+/*
+ * Suspend the target task, and return a suspension token. The token
+ * represents a reference on the suspended task.
+ */
+kern_return_t
+task_suspend2(
+       task_t                  task,
+       task_suspension_token_t *suspend_token)
+{
+       kern_return_t    kr;
+
+       kr = task_suspend_internal(task);
+       if (kr != KERN_SUCCESS) {
+               *suspend_token = TASK_NULL;
+               return kr;
+       }
+
+       /*
+        * Take a reference on the target task and return that to the caller
+        * as a "suspension token," which can be converted into an SO right to
+        * the now-suspended task's resume port.
+        */
+       task_reference_internal(task);
+       *suspend_token = task;
+
+       return KERN_SUCCESS;
+}
+
+/*
+ * Resume the task
+ * (reference/token/port management is caller's responsibility).
+ */
+kern_return_t
+task_resume_internal(
+       task_suspension_token_t         task)
+{
+       kern_return_t kr;
+
+       if (task == TASK_NULL || task == kernel_task) {
+               return KERN_INVALID_ARGUMENT;
+       }
+
+       task_lock(task);
+       kr = release_task_hold(task, TASK_HOLD_NORMAL);
+       task_unlock(task);
+       return kr;
+}
+
+/*
+ * Resume the task using a suspension token. Consumes the token's ref.
+ */
+kern_return_t
+task_resume2(
+       task_suspension_token_t         task)
+{
+       kern_return_t kr;
+
+       kr = task_resume_internal(task);
+       task_suspension_token_deallocate(task);
+
+       return kr;
+}
+
+boolean_t
+task_suspension_notify(mach_msg_header_t *request_header)
+{
+       ipc_port_t port = request_header->msgh_remote_port;
+       task_t task = convert_port_to_task_suspension_token(port);
+       mach_msg_type_number_t not_count;
+
+       if (task == TASK_NULL || task == kernel_task) {
+               return TRUE;  /* nothing to do */
+       }
+       switch (request_header->msgh_id) {
+       case MACH_NOTIFY_SEND_ONCE:
+               /* release the hold held by this specific send-once right */
+               task_lock(task);
+               release_task_hold(task, TASK_HOLD_NORMAL);
+               task_unlock(task);
+               break;
+
+       case MACH_NOTIFY_NO_SENDERS:
+               not_count = ((mach_no_senders_notification_t *)request_header)->not_count;
+
+               task_lock(task);
+               ip_lock(port);
+               if (port->ip_mscount == not_count) {
+                       /* release all the [remaining] outstanding legacy holds */
+                       assert(port->ip_nsrequest == IP_NULL);
+                       ip_unlock(port);
+                       release_task_hold(task, TASK_HOLD_LEGACY_ALL);
+                       task_unlock(task);
+               } else if (port->ip_nsrequest == IP_NULL) {
+                       ipc_port_t old_notify;
+
+                       task_unlock(task);
+                       /* new send rights, re-arm notification at current make-send count */
+                       ipc_port_nsrequest(port, port->ip_mscount, ipc_port_make_sonce_locked(port), &old_notify);
+                       assert(old_notify == IP_NULL);
+                       /* port unlocked */
+               } else {
+                       ip_unlock(port);
+                       task_unlock(task);
+               }
+               break;
+
+       default:
+               break;
+       }
+
+       task_suspension_token_deallocate(task); /* drop token reference */
+       return TRUE;
+}
+
+static kern_return_t
+task_pidsuspend_locked(task_t task)
+{
+       kern_return_t kr;
+
+       if (task->pidsuspended) {
+               kr = KERN_FAILURE;
+               goto out;
+       }
+
+       task->pidsuspended = TRUE;
+
+       kr = place_task_hold(task, TASK_HOLD_PIDSUSPEND);
+       if (kr != KERN_SUCCESS) {
+               task->pidsuspended = FALSE;
+       }
+out:
+       return kr;
+}
+
+
+/*
+ *     task_pidsuspend:
+ *
+ *     Suspends a task by placing a hold on its threads.
+ *
+ * Conditions:
+ *      The caller holds a reference to the task
+ */
+kern_return_t
+task_pidsuspend(
+       task_t          task)
+{
+       kern_return_t    kr;
+
+       if (task == TASK_NULL || task == kernel_task) {
+               return KERN_INVALID_ARGUMENT;
+       }
+
+       task_lock(task);
+
+       kr = task_pidsuspend_locked(task);
+
+       task_unlock(task);
+
+       if ((KERN_SUCCESS == kr) && task->message_app_suspended) {
+               iokit_task_app_suspended_changed(task);
+       }
+
+       return kr;
+}
+
+/*
+ *     task_pidresume:
+ *             Resumes a previously suspended task.
+ *
+ * Conditions:
+ *             The caller holds a reference to the task
+ */
+kern_return_t
+task_pidresume(
+       task_t  task)
+{
+       kern_return_t    kr;
+
+       if (task == TASK_NULL || task == kernel_task) {
+               return KERN_INVALID_ARGUMENT;
+       }
+
+       task_lock(task);
+
+#if CONFIG_FREEZE
+
+       while (task->changing_freeze_state) {
+               assert_wait((event_t)&task->changing_freeze_state, THREAD_UNINT);
+               task_unlock(task);
+               thread_block(THREAD_CONTINUE_NULL);
+
+               task_lock(task);
+       }
+       task->changing_freeze_state = TRUE;
+#endif
+
+       kr = release_task_hold(task, TASK_HOLD_PIDSUSPEND);
+
+       task_unlock(task);
+
+       if ((KERN_SUCCESS == kr) && task->message_app_suspended) {
+               iokit_task_app_suspended_changed(task);
+       }
+
+#if CONFIG_FREEZE
+
+       task_lock(task);
+
+       if (kr == KERN_SUCCESS) {
+               task->frozen = FALSE;
+       }
+       task->changing_freeze_state = FALSE;
+       thread_wakeup(&task->changing_freeze_state);
+
+       task_unlock(task);
+#endif
+
+       return kr;
+}
+
+os_refgrp_decl(static, task_watchports_refgrp, "task_watchports", NULL);
+
+/*
+ *     task_add_turnstile_watchports:
+ *             Setup watchports to boost the main thread of the task.
+ *
+ *     Arguments:
+ *             task: task being spawned
+ *             thread: main thread of task
+ *             portwatch_ports: array of watchports
+ *             portwatch_count: number of watchports
+ *
+ *     Conditions:
+ *             Nothing locked.
+ */
+void
+task_add_turnstile_watchports(
+       task_t          task,
+       thread_t        thread,
+       ipc_port_t      *portwatch_ports,
+       uint32_t        portwatch_count)
+{
+       struct task_watchports *watchports = NULL;
+       struct task_watchport_elem *previous_elem_array[TASK_MAX_WATCHPORT_COUNT] = {};
+       os_ref_count_t refs;
+
+       /* Check if the task has terminated */
+       if (!task->active) {
+               return;
+       }
+
+       assert(portwatch_count <= TASK_MAX_WATCHPORT_COUNT);
+
+       watchports = task_watchports_alloc_init(task, thread, portwatch_count);
+
+       /* Lock the ipc space */
+       is_write_lock(task->itk_space);
+
+       /* Setup watchports to boost the main thread */
+       refs = task_add_turnstile_watchports_locked(task,
+           watchports, previous_elem_array, portwatch_ports,
+           portwatch_count);
+
+       /* Drop the space lock */
+       is_write_unlock(task->itk_space);
+
+       if (refs == 0) {
+               task_watchports_deallocate(watchports);
+       }
+
+       /* Drop the ref on previous_elem_array */
+       for (uint32_t i = 0; i < portwatch_count && previous_elem_array[i] != NULL; i++) {
+               task_watchport_elem_deallocate(previous_elem_array[i]);
+       }
+}
+
+/*
+ *     task_remove_turnstile_watchports:
+ *             Clear all turnstile boost on the task from watchports.
+ *
+ *     Arguments:
+ *             task: task being terminated
+ *
+ *     Conditions:
+ *             Nothing locked.
+ */
+void
+task_remove_turnstile_watchports(
+       task_t          task)
+{
+       os_ref_count_t refs = TASK_MAX_WATCHPORT_COUNT;
+       struct task_watchports *watchports = NULL;
+       ipc_port_t port_freelist[TASK_MAX_WATCHPORT_COUNT] = {};
+       uint32_t portwatch_count;
+
+       /* Lock the ipc space */
+       is_write_lock(task->itk_space);
+
+       /* Check if watchport boost exist */
+       if (task->watchports == NULL) {
+               is_write_unlock(task->itk_space);
+               return;
+       }
+       watchports = task->watchports;
+       portwatch_count = watchports->tw_elem_array_count;
+
+       refs = task_remove_turnstile_watchports_locked(task, watchports,
+           port_freelist);
+
+       is_write_unlock(task->itk_space);
+
+       /* Drop all the port references */
+       for (uint32_t i = 0; i < portwatch_count && port_freelist[i] != NULL; i++) {
+               ip_release(port_freelist[i]);
+       }
+
+       /* Clear the task and thread references for task_watchport */
+       if (refs == 0) {
+               task_watchports_deallocate(watchports);
+       }
+}
+
+/*
+ *     task_transfer_turnstile_watchports:
+ *             Transfer all watchport turnstile boost from old task to new task.
+ *
+ *     Arguments:
+ *             old_task: task calling exec
+ *             new_task: new exec'ed task
+ *             thread: main thread of new task
+ *
+ *     Conditions:
+ *             Nothing locked.
+ */
+void
+task_transfer_turnstile_watchports(
+       task_t   old_task,
+       task_t   new_task,
+       thread_t new_thread)
+{
+       struct task_watchports *old_watchports = NULL;
+       struct task_watchports *new_watchports = NULL;
+       os_ref_count_t old_refs = TASK_MAX_WATCHPORT_COUNT;
+       os_ref_count_t new_refs = TASK_MAX_WATCHPORT_COUNT;
+       uint32_t portwatch_count;
+
+       if (old_task->watchports == NULL || !new_task->active) {
+               return;
+       }
+
+       /* Get the watch port count from the old task */
+       is_write_lock(old_task->itk_space);
+       if (old_task->watchports == NULL) {
+               is_write_unlock(old_task->itk_space);
+               return;
+       }
+
+       portwatch_count = old_task->watchports->tw_elem_array_count;
+       is_write_unlock(old_task->itk_space);
+
+       new_watchports = task_watchports_alloc_init(new_task, new_thread, portwatch_count);
+
+       /* Lock the ipc space for old task */
+       is_write_lock(old_task->itk_space);
+
+       /* Lock the ipc space for new task */
+       is_write_lock(new_task->itk_space);
+
+       /* Check if watchport boost exist */
+       if (old_task->watchports == NULL || !new_task->active) {
+               is_write_unlock(new_task->itk_space);
+               is_write_unlock(old_task->itk_space);
+               (void)task_watchports_release(new_watchports);
+               task_watchports_deallocate(new_watchports);
+               return;
+       }
+
+       old_watchports = old_task->watchports;
+       assert(portwatch_count == old_task->watchports->tw_elem_array_count);
+
+       /* Setup new task watchports */
+       new_task->watchports = new_watchports;
+
+       for (uint32_t i = 0; i < portwatch_count; i++) {
+               ipc_port_t port = old_watchports->tw_elem[i].twe_port;
+
+               if (port == NULL) {
+                       task_watchport_elem_clear(&new_watchports->tw_elem[i]);
+                       continue;
+               }
+
+               /* Lock the port and check if it has the entry */
+               ip_lock(port);
+               imq_lock(&port->ip_messages);
+
+               task_watchport_elem_init(&new_watchports->tw_elem[i], new_task, port);
+
+               if (ipc_port_replace_watchport_elem_conditional_locked(port,
+                   &old_watchports->tw_elem[i], &new_watchports->tw_elem[i]) == KERN_SUCCESS) {
+                       task_watchport_elem_clear(&old_watchports->tw_elem[i]);
+
+                       task_watchports_retain(new_watchports);
+                       old_refs = task_watchports_release(old_watchports);
+
+                       /* Check if all ports are cleaned */
+                       if (old_refs == 0) {
+                               old_task->watchports = NULL;
+                       }
+               } else {
+                       task_watchport_elem_clear(&new_watchports->tw_elem[i]);
+               }
+               /* mqueue and port unlocked by ipc_port_replace_watchport_elem_conditional_locked */
+       }
+
+       /* Drop the reference on new task_watchports struct returned by task_watchports_alloc_init */
+       new_refs = task_watchports_release(new_watchports);
+       if (new_refs == 0) {
+               new_task->watchports = NULL;
+       }
+
+       is_write_unlock(new_task->itk_space);
+       is_write_unlock(old_task->itk_space);
+
+       /* Clear the task and thread references for old_watchport */
+       if (old_refs == 0) {
+               task_watchports_deallocate(old_watchports);
+       }
+
+       /* Clear the task and thread references for new_watchport */
+       if (new_refs == 0) {
+               task_watchports_deallocate(new_watchports);
+       }
+}
+
+/*
+ *     task_add_turnstile_watchports_locked:
+ *             Setup watchports to boost the main thread of the task.
+ *
+ *     Arguments:
+ *             task: task to boost
+ *             watchports: watchport structure to be attached to the task
+ *             previous_elem_array: an array of old watchport_elem to be returned to caller
+ *             portwatch_ports: array of watchports
+ *             portwatch_count: number of watchports
+ *
+ *     Conditions:
+ *             ipc space of the task locked.
+ *             returns array of old watchport_elem in previous_elem_array
+ */
+static os_ref_count_t
+task_add_turnstile_watchports_locked(
+       task_t                      task,
+       struct task_watchports      *watchports,
+       struct task_watchport_elem  **previous_elem_array,
+       ipc_port_t                  *portwatch_ports,
+       uint32_t                    portwatch_count)
+{
+       os_ref_count_t refs = TASK_MAX_WATCHPORT_COUNT;
+
+       /* Check if the task is still active */
+       if (!task->active) {
+               refs = task_watchports_release(watchports);
+               return refs;
+       }
+
+       assert(task->watchports == NULL);
+       task->watchports = watchports;
+
+       for (uint32_t i = 0, j = 0; i < portwatch_count; i++) {
+               ipc_port_t port = portwatch_ports[i];
+
+               task_watchport_elem_init(&watchports->tw_elem[i], task, port);
+               if (port == NULL) {
+                       task_watchport_elem_clear(&watchports->tw_elem[i]);
+                       continue;
+               }
+
+               ip_lock(port);
+               imq_lock(&port->ip_messages);
+
+               /* Check if port is in valid state to be setup as watchport */
+               if (ipc_port_add_watchport_elem_locked(port, &watchports->tw_elem[i],
+                   &previous_elem_array[j]) != KERN_SUCCESS) {
+                       task_watchport_elem_clear(&watchports->tw_elem[i]);
+                       continue;
+               }
+               /* port and mqueue unlocked on return */
+
+               ip_reference(port);
+               task_watchports_retain(watchports);
+               if (previous_elem_array[j] != NULL) {
+                       j++;
+               }
+       }
+
+       /* Drop the reference on task_watchport struct returned by os_ref_init */
+       refs = task_watchports_release(watchports);
+       if (refs == 0) {
+               task->watchports = NULL;
+       }
+
+       return refs;
+}
+
+/*
+ *     task_remove_turnstile_watchports_locked:
+ *             Clear all turnstile boost on the task from watchports.
+ *
+ *     Arguments:
+ *             task: task to remove watchports from
+ *             watchports: watchports structure for the task
+ *             port_freelist: array of ports returned with ref to caller
+ *
+ *
+ *     Conditions:
+ *             ipc space of the task locked.
+ *             array of ports with refs are returned in port_freelist
+ */
+static os_ref_count_t
+task_remove_turnstile_watchports_locked(
+       task_t                 task,
+       struct task_watchports *watchports,
+       ipc_port_t             *port_freelist)
+{
+       os_ref_count_t refs = TASK_MAX_WATCHPORT_COUNT;
+
+       for (uint32_t i = 0, j = 0; i < watchports->tw_elem_array_count; i++) {
+               ipc_port_t port = watchports->tw_elem[i].twe_port;
+               if (port == NULL) {
+                       continue;
+               }
+
+               /* Lock the port and check if it has the entry */
+               ip_lock(port);
+               imq_lock(&port->ip_messages);
+               if (ipc_port_clear_watchport_elem_internal_conditional_locked(port,
+                   &watchports->tw_elem[i]) == KERN_SUCCESS) {
+                       task_watchport_elem_clear(&watchports->tw_elem[i]);
+                       port_freelist[j++] = port;
+                       refs = task_watchports_release(watchports);
+
+                       /* Check if all ports are cleaned */
+                       if (refs == 0) {
+                               task->watchports = NULL;
+                               break;
+                       }
+               }
+               /* mqueue and port unlocked by ipc_port_clear_watchport_elem_internal_conditional_locked */
+       }
+       return refs;
+}
+
+/*
+ *     task_watchports_alloc_init:
+ *             Allocate and initialize task watchport struct.
+ *
+ *     Conditions:
+ *             Nothing locked.
+ */
+static struct task_watchports *
+task_watchports_alloc_init(
+       task_t        task,
+       thread_t      thread,
+       uint32_t      count)
+{
+       struct task_watchports *watchports = kalloc(sizeof(struct task_watchports) +
+           count * sizeof(struct task_watchport_elem));
+
+       task_reference(task);
+       thread_reference(thread);
+       watchports->tw_task = task;
+       watchports->tw_thread = thread;
+       watchports->tw_elem_array_count = count;
+       os_ref_init(&watchports->tw_refcount, &task_watchports_refgrp);
+
+       return watchports;
+}
+
+/*
+ *     task_watchports_deallocate:
+ *             Deallocate task watchport struct.
+ *
+ *     Conditions:
+ *             Nothing locked.
+ */
+static void
+task_watchports_deallocate(
+       struct task_watchports *watchports)
+{
+       uint32_t portwatch_count = watchports->tw_elem_array_count;
+
+       task_deallocate(watchports->tw_task);
+       thread_deallocate(watchports->tw_thread);
+       kfree(watchports, sizeof(struct task_watchports) + portwatch_count * sizeof(struct task_watchport_elem));
+}
+
+/*
+ *     task_watchport_elem_deallocate:
+ *             Deallocate task watchport element and release its ref on task_watchport.
+ *
+ *     Conditions:
+ *             Nothing locked.
+ */
+void
+task_watchport_elem_deallocate(
+       struct task_watchport_elem *watchport_elem)
+{
+       os_ref_count_t refs = TASK_MAX_WATCHPORT_COUNT;
+       task_t task = watchport_elem->twe_task;
+       struct task_watchports *watchports = NULL;
+       ipc_port_t port = NULL;
+
+       assert(task != NULL);
+
+       /* Take the space lock to modify the elememt */
+       is_write_lock(task->itk_space);
+
+       watchports = task->watchports;
+       assert(watchports != NULL);
+
+       port = watchport_elem->twe_port;
+       assert(port != NULL);
+
+       task_watchport_elem_clear(watchport_elem);
+       refs = task_watchports_release(watchports);
+
+       if (refs == 0) {
+               task->watchports = NULL;
+       }
+
+       is_write_unlock(task->itk_space);
+
+       ip_release(port);
+       if (refs == 0) {
+               task_watchports_deallocate(watchports);
+       }
+}
+
+/*
+ *     task_has_watchports:
+ *             Return TRUE if task has watchport boosts.
+ *
+ *     Conditions:
+ *             Nothing locked.
+ */
+boolean_t
+task_has_watchports(task_t task)
+{
+       return task->watchports != NULL;
+}
+
+#if DEVELOPMENT || DEBUG
+
+extern void IOSleep(int);
+
+kern_return_t
+task_disconnect_page_mappings(task_t task)
+{
+       int     n;
+
+       if (task == TASK_NULL || task == kernel_task) {
+               return KERN_INVALID_ARGUMENT;
+       }
+
+       /*
+        * this function is used to strip all of the mappings from
+        * the pmap for the specified task to force the task to
+        * re-fault all of the pages it is actively using... this
+        * allows us to approximate the true working set of the
+        * specified task.  We only engage if at least 1 of the
+        * threads in the task is runnable, but we want to continuously
+        * sweep (at least for a while - I've arbitrarily set the limit at
+        * 100 sweeps to be re-looked at as we gain experience) to get a better
+        * view into what areas within a page are being visited (as opposed to only
+        * seeing the first fault of a page after the task becomes
+        * runnable)...  in the future I may
+        * try to block until awakened by a thread in this task
+        * being made runnable, but for now we'll periodically poll from the
+        * user level debug tool driving the sysctl
+        */
+       for (n = 0; n < 100; n++) {
+               thread_t        thread;
+               boolean_t       runnable;
+               boolean_t       do_unnest;
+               int             page_count;
+
+               runnable = FALSE;
+               do_unnest = FALSE;
+
+               task_lock(task);
+
+               queue_iterate(&task->threads, thread, thread_t, task_threads) {
+                       if (thread->state & TH_RUN) {
+                               runnable = TRUE;
+                               break;
+                       }
+               }
+               if (n == 0) {
+                       task->task_disconnected_count++;
+               }
+
+               if (task->task_unnested == FALSE) {
+                       if (runnable == TRUE) {
+                               task->task_unnested = TRUE;
+                               do_unnest = TRUE;
+                       }
+               }
+               task_unlock(task);
+
+               if (runnable == FALSE) {
+                       break;
+               }
+
+               KERNEL_DEBUG_CONSTANT_IST(KDEBUG_TRACE, (MACHDBG_CODE(DBG_MACH_WORKINGSET, VM_DISCONNECT_TASK_PAGE_MAPPINGS)) | DBG_FUNC_START,
+                   task, do_unnest, task->task_disconnected_count, 0, 0);
+
+               page_count = vm_map_disconnect_page_mappings(task->map, do_unnest);
+
+               KERNEL_DEBUG_CONSTANT_IST(KDEBUG_TRACE, (MACHDBG_CODE(DBG_MACH_WORKINGSET, VM_DISCONNECT_TASK_PAGE_MAPPINGS)) | DBG_FUNC_END,
+                   task, page_count, 0, 0, 0);
+
+               if ((n % 5) == 4) {
+                       IOSleep(1);
+               }
+       }
+       return KERN_SUCCESS;
+}
+
+#endif
+
+
+#if CONFIG_FREEZE
+
+/*
+ *     task_freeze:
+ *
+ *     Freeze a task.
+ *
+ * Conditions:
+ *      The caller holds a reference to the task
+ */
+extern void             vm_wake_compactor_swapper(void);
+extern queue_head_t     c_swapout_list_head;
+extern struct freezer_context freezer_context_global;
+
+kern_return_t
+task_freeze(
+       task_t    task,
+       uint32_t           *purgeable_count,
+       uint32_t           *wired_count,
+       uint32_t           *clean_count,
+       uint32_t           *dirty_count,
+       uint32_t           dirty_budget,
+       uint32_t           *shared_count,
+       int                *freezer_error_code,
+       boolean_t          eval_only)
+{
+       kern_return_t kr = KERN_SUCCESS;
+
+       if (task == TASK_NULL || task == kernel_task) {
+               return KERN_INVALID_ARGUMENT;
+       }
+
+       task_lock(task);
+
+       while (task->changing_freeze_state) {
+               assert_wait((event_t)&task->changing_freeze_state, THREAD_UNINT);
+               task_unlock(task);
+               thread_block(THREAD_CONTINUE_NULL);
+
+               task_lock(task);
+       }
+       if (task->frozen) {
+               task_unlock(task);
+               return KERN_FAILURE;
+       }
+       task->changing_freeze_state = TRUE;
+
+       freezer_context_global.freezer_ctx_task = task;
+
+       task_unlock(task);
+
+       kr = vm_map_freeze(task,
+           purgeable_count,
+           wired_count,
+           clean_count,
+           dirty_count,
+           dirty_budget,
+           shared_count,
+           freezer_error_code,
+           eval_only);
+
+       task_lock(task);
+
+       if ((kr == KERN_SUCCESS) && (eval_only == FALSE)) {
+               task->frozen = TRUE;
+
+               freezer_context_global.freezer_ctx_task = NULL;
+               freezer_context_global.freezer_ctx_uncompressed_pages = 0;
+
+               if (VM_CONFIG_FREEZER_SWAP_IS_ACTIVE) {
+                       /*
+                        * reset the counter tracking the # of swapped compressed pages
+                        * because we are now done with this freeze session and task.
+                        */
+
+                       *dirty_count = (uint32_t) (freezer_context_global.freezer_ctx_swapped_bytes / PAGE_SIZE_64); /*used to track pageouts*/
+               }
+
+               freezer_context_global.freezer_ctx_swapped_bytes = 0;
+       }
+
+       task->changing_freeze_state = FALSE;
+       thread_wakeup(&task->changing_freeze_state);
+
+       task_unlock(task);
+
+       if (VM_CONFIG_COMPRESSOR_IS_PRESENT &&
+           (kr == KERN_SUCCESS) &&
+           (eval_only == FALSE)) {
+               vm_wake_compactor_swapper();
+               /*
+                * We do an explicit wakeup of the swapout thread here
+                * because the compact_and_swap routines don't have
+                * knowledge about these kind of "per-task packed c_segs"
+                * and so will not be evaluating whether we need to do
+                * a wakeup there.
+                */
+               thread_wakeup((event_t)&c_swapout_list_head);
+       }
+
+       return kr;
+}
+
+/*
+ *     task_thaw:
+ *
+ *     Thaw a currently frozen task.
+ *
+ * Conditions:
+ *      The caller holds a reference to the task
+ */
+kern_return_t
+task_thaw(
+       task_t          task)
+{
+       if (task == TASK_NULL || task == kernel_task) {
+               return KERN_INVALID_ARGUMENT;
+       }
+
+       task_lock(task);
+
+       while (task->changing_freeze_state) {
+               assert_wait((event_t)&task->changing_freeze_state, THREAD_UNINT);
+               task_unlock(task);
+               thread_block(THREAD_CONTINUE_NULL);
+
+               task_lock(task);
+       }
+       if (!task->frozen) {
+               task_unlock(task);
+               return KERN_FAILURE;
+       }
+       task->frozen = FALSE;
+
+       task_unlock(task);
+
+       return KERN_SUCCESS;
+}
+
+void
+task_update_frozen_to_swap_acct(task_t task, int64_t amount, freezer_acct_op_t op)
+{
+       /*
+        * We don't assert that the task lock is held because we call this
+        * routine from the decompression path and we won't be holding the
+        * task lock. However, since we are in the context of the task we are
+        * safe.
+        * In the case of the task_freeze path, we call it from behind the task
+        * lock but we don't need to because we have a reference on the proc
+        * being frozen.
+        */
+
+       assert(task);
+       if (amount == 0) {
+               return;
+       }
+
+       if (op == CREDIT_TO_SWAP) {
+               ledger_credit_nocheck(task->ledger, task_ledgers.frozen_to_swap, amount);
+       } else if (op == DEBIT_FROM_SWAP) {
+               ledger_debit_nocheck(task->ledger, task_ledgers.frozen_to_swap, amount);
+       } else {
+               panic("task_update_frozen_to_swap_acct: Invalid ledger op\n");
+       }
+}
+#endif /* CONFIG_FREEZE */
+
+kern_return_t
+host_security_set_task_token(
+       host_security_t  host_security,
+       task_t           task,
+       security_token_t sec_token,
+       audit_token_t    audit_token,
+       host_priv_t      host_priv)
+{
+       ipc_port_t       host_port;
+       kern_return_t    kr;
+
+       if (task == TASK_NULL) {
+               return KERN_INVALID_ARGUMENT;
+       }
+
+       if (host_security == HOST_NULL) {
+               return KERN_INVALID_SECURITY;
+       }
+
+       task_lock(task);
+       task->sec_token = sec_token;
+       task->audit_token = audit_token;
+       task_unlock(task);
+
+       if (host_priv != HOST_PRIV_NULL) {
+               kr = host_get_host_priv_port(host_priv, &host_port);
+       } else {
+               kr = host_get_host_port(host_priv_self(), &host_port);
+       }
+       assert(kr == KERN_SUCCESS);
+
+       kr = task_set_special_port_internal(task, TASK_HOST_PORT, host_port);
+       return kr;
+}
+
+kern_return_t
+task_send_trace_memory(
+       __unused task_t   target_task,
+       __unused uint32_t pid,
+       __unused uint64_t uniqueid)
+{
+       return KERN_INVALID_ARGUMENT;
+}
+
+/*
+ * This routine was added, pretty much exclusively, for registering the
+ * RPC glue vector for in-kernel short circuited tasks.  Rather than
+ * removing it completely, I have only disabled that feature (which was
+ * the only feature at the time).  It just appears that we are going to
+ * want to add some user data to tasks in the future (i.e. bsd info,
+ * task names, etc...), so I left it in the formal task interface.
+ */
+kern_return_t
+task_set_info(
+       task_t          task,
+       task_flavor_t   flavor,
+       __unused task_info_t    task_info_in,           /* pointer to IN array */
+       __unused mach_msg_type_number_t task_info_count)
+{
+       if (task == TASK_NULL) {
+               return KERN_INVALID_ARGUMENT;
+       }
+       switch (flavor) {
+#if CONFIG_ATM
+       case TASK_TRACE_MEMORY_INFO:
+               return KERN_NOT_SUPPORTED;
+#endif // CONFIG_ATM
+       default:
+               return KERN_INVALID_ARGUMENT;
+       }
+}
+
+int radar_20146450 = 1;
+kern_return_t
+task_info(
+       task_t                  task,
+       task_flavor_t           flavor,
+       task_info_t             task_info_out,
+       mach_msg_type_number_t  *task_info_count)
+{
+       kern_return_t error = KERN_SUCCESS;
+       mach_msg_type_number_t  original_task_info_count;
+       bool is_kernel_task = (task == kernel_task);
+
+       if (task == TASK_NULL) {
+               return KERN_INVALID_ARGUMENT;
+       }
+
+       original_task_info_count = *task_info_count;
+       task_lock(task);
+
+       if ((task != current_task()) && (!task->active)) {
+               task_unlock(task);
+               return KERN_INVALID_ARGUMENT;
+       }
+
+
+       switch (flavor) {
+       case TASK_BASIC_INFO_32:
+       case TASK_BASIC2_INFO_32:
+#if defined(__arm__) || defined(__arm64__)
+       case TASK_BASIC_INFO_64:
+#endif
+               {
+                       task_basic_info_32_t    basic_info;
+                       vm_map_t                                map;
+                       clock_sec_t                             secs;
+                       clock_usec_t                    usecs;
+
+                       if (*task_info_count < TASK_BASIC_INFO_32_COUNT) {
+                               error = KERN_INVALID_ARGUMENT;
+                               break;
+                       }
+
+                       basic_info = (task_basic_info_32_t)task_info_out;
+
+                       map = (task == kernel_task)? kernel_map: task->map;
+                       basic_info->virtual_size = (typeof(basic_info->virtual_size))vm_map_adjusted_size(map);
+                       if (flavor == TASK_BASIC2_INFO_32) {
+                               /*
+                                * The "BASIC2" flavor gets the maximum resident
+                                * size instead of the current resident size...
+                                */
+                               basic_info->resident_size = pmap_resident_max(map->pmap);
+                       } else {
+                               basic_info->resident_size = pmap_resident_count(map->pmap);
+                       }
+                       basic_info->resident_size *= PAGE_SIZE;
+
+                       basic_info->policy = ((task != kernel_task)?
+                           POLICY_TIMESHARE: POLICY_RR);
+                       basic_info->suspend_count = task->user_stop_count;
+
+                       absolutetime_to_microtime(task->total_user_time, &secs, &usecs);
+                       basic_info->user_time.seconds =
+                           (typeof(basic_info->user_time.seconds))secs;
+                       basic_info->user_time.microseconds = usecs;
+
+                       absolutetime_to_microtime(task->total_system_time, &secs, &usecs);
+                       basic_info->system_time.seconds =
+                           (typeof(basic_info->system_time.seconds))secs;
+                       basic_info->system_time.microseconds = usecs;
+
+                       *task_info_count = TASK_BASIC_INFO_32_COUNT;
+                       break;
+               }
+
+#if defined(__arm__) || defined(__arm64__)
+       case TASK_BASIC_INFO_64_2:
+       {
+               task_basic_info_64_2_t  basic_info;
+               vm_map_t                                map;
+               clock_sec_t                             secs;
+               clock_usec_t                    usecs;
+
+               if (*task_info_count < TASK_BASIC_INFO_64_2_COUNT) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               basic_info = (task_basic_info_64_2_t)task_info_out;
+
+               map = (task == kernel_task)? kernel_map: task->map;
+               basic_info->virtual_size  = vm_map_adjusted_size(map);
+               basic_info->resident_size =
+                   (mach_vm_size_t)(pmap_resident_count(map->pmap))
+                   * PAGE_SIZE_64;
+
+               basic_info->policy = ((task != kernel_task)?
+                   POLICY_TIMESHARE: POLICY_RR);
+               basic_info->suspend_count = task->user_stop_count;
+
+               absolutetime_to_microtime(task->total_user_time, &secs, &usecs);
+               basic_info->user_time.seconds =
+                   (typeof(basic_info->user_time.seconds))secs;
+               basic_info->user_time.microseconds = usecs;
+
+               absolutetime_to_microtime(task->total_system_time, &secs, &usecs);
+               basic_info->system_time.seconds =
+                   (typeof(basic_info->system_time.seconds))secs;
+               basic_info->system_time.microseconds = usecs;
+
+               *task_info_count = TASK_BASIC_INFO_64_2_COUNT;
+               break;
+       }
+
+#else /* defined(__arm__) || defined(__arm64__) */
+       case TASK_BASIC_INFO_64:
+       {
+               task_basic_info_64_t    basic_info;
+               vm_map_t                                map;
+               clock_sec_t                             secs;
+               clock_usec_t                    usecs;
+
+               if (*task_info_count < TASK_BASIC_INFO_64_COUNT) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               basic_info = (task_basic_info_64_t)task_info_out;
+
+               map = (task == kernel_task)? kernel_map: task->map;
+               basic_info->virtual_size  = vm_map_adjusted_size(map);
+               basic_info->resident_size =
+                   (mach_vm_size_t)(pmap_resident_count(map->pmap))
+                   * PAGE_SIZE_64;
+
+               basic_info->policy = ((task != kernel_task)?
+                   POLICY_TIMESHARE: POLICY_RR);
+               basic_info->suspend_count = task->user_stop_count;
+
+               absolutetime_to_microtime(task->total_user_time, &secs, &usecs);
+               basic_info->user_time.seconds =
+                   (typeof(basic_info->user_time.seconds))secs;
+               basic_info->user_time.microseconds = usecs;
+
+               absolutetime_to_microtime(task->total_system_time, &secs, &usecs);
+               basic_info->system_time.seconds =
+                   (typeof(basic_info->system_time.seconds))secs;
+               basic_info->system_time.microseconds = usecs;
+
+               *task_info_count = TASK_BASIC_INFO_64_COUNT;
+               break;
+       }
+#endif /* defined(__arm__) || defined(__arm64__) */
+
+       case MACH_TASK_BASIC_INFO:
+       {
+               mach_task_basic_info_t  basic_info;
+               vm_map_t                map;
+               clock_sec_t             secs;
+               clock_usec_t            usecs;
+
+               if (*task_info_count < MACH_TASK_BASIC_INFO_COUNT) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               basic_info = (mach_task_basic_info_t)task_info_out;
+
+               map = (task == kernel_task) ? kernel_map : task->map;
+
+               basic_info->virtual_size  = vm_map_adjusted_size(map);
+
+               basic_info->resident_size =
+                   (mach_vm_size_t)(pmap_resident_count(map->pmap));
+               basic_info->resident_size *= PAGE_SIZE_64;
+
+               basic_info->resident_size_max =
+                   (mach_vm_size_t)(pmap_resident_max(map->pmap));
+               basic_info->resident_size_max *= PAGE_SIZE_64;
+
+               basic_info->policy = ((task != kernel_task) ?
+                   POLICY_TIMESHARE : POLICY_RR);
+
+               basic_info->suspend_count = task->user_stop_count;
+
+               absolutetime_to_microtime(task->total_user_time, &secs, &usecs);
+               basic_info->user_time.seconds =
+                   (typeof(basic_info->user_time.seconds))secs;
+               basic_info->user_time.microseconds = usecs;
+
+               absolutetime_to_microtime(task->total_system_time, &secs, &usecs);
+               basic_info->system_time.seconds =
+                   (typeof(basic_info->system_time.seconds))secs;
+               basic_info->system_time.microseconds = usecs;
+
+               *task_info_count = MACH_TASK_BASIC_INFO_COUNT;
+               break;
+       }
+
+       case TASK_THREAD_TIMES_INFO:
+       {
+               task_thread_times_info_t        times_info;
+               thread_t                                        thread;
+
+               if (*task_info_count < TASK_THREAD_TIMES_INFO_COUNT) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               times_info = (task_thread_times_info_t) task_info_out;
+               times_info->user_time.seconds = 0;
+               times_info->user_time.microseconds = 0;
+               times_info->system_time.seconds = 0;
+               times_info->system_time.microseconds = 0;
+
+
+               queue_iterate(&task->threads, thread, thread_t, task_threads) {
+                       time_value_t    user_time, system_time;
+
+                       if (thread->options & TH_OPT_IDLE_THREAD) {
+                               continue;
+                       }
+
+                       thread_read_times(thread, &user_time, &system_time, NULL);
+
+                       time_value_add(&times_info->user_time, &user_time);
+                       time_value_add(&times_info->system_time, &system_time);
+               }
+
+               *task_info_count = TASK_THREAD_TIMES_INFO_COUNT;
+               break;
+       }
+
+       case TASK_ABSOLUTETIME_INFO:
+       {
+               task_absolutetime_info_t        info;
+               thread_t                        thread;
+
+               if (*task_info_count < TASK_ABSOLUTETIME_INFO_COUNT) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               info = (task_absolutetime_info_t)task_info_out;
+               info->threads_user = info->threads_system = 0;
+
+
+               info->total_user = task->total_user_time;
+               info->total_system = task->total_system_time;
+
+               queue_iterate(&task->threads, thread, thread_t, task_threads) {
+                       uint64_t        tval;
+                       spl_t           x;
+
+                       if (thread->options & TH_OPT_IDLE_THREAD) {
+                               continue;
+                       }
+
+                       x = splsched();
+                       thread_lock(thread);
+
+                       tval = timer_grab(&thread->user_timer);
+                       info->threads_user += tval;
+                       info->total_user += tval;
+
+                       tval = timer_grab(&thread->system_timer);
+                       if (thread->precise_user_kernel_time) {
+                               info->threads_system += tval;
+                               info->total_system += tval;
+                       } else {
+                               /* system_timer may represent either sys or user */
+                               info->threads_user += tval;
+                               info->total_user += tval;
+                       }
+
+                       thread_unlock(thread);
+                       splx(x);
+               }
+
+
+               *task_info_count = TASK_ABSOLUTETIME_INFO_COUNT;
+               break;
+       }
+
+       case TASK_DYLD_INFO:
+       {
+               task_dyld_info_t info;
+
+               /*
+                * We added the format field to TASK_DYLD_INFO output.  For
+                * temporary backward compatibility, accept the fact that
+                * clients may ask for the old version - distinquished by the
+                * size of the expected result structure.
+                */
+#define TASK_LEGACY_DYLD_INFO_COUNT \
+               offsetof(struct task_dyld_info, all_image_info_format)/sizeof(natural_t)
+
+               if (*task_info_count < TASK_LEGACY_DYLD_INFO_COUNT) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               info = (task_dyld_info_t)task_info_out;
+               info->all_image_info_addr = task->all_image_info_addr;
+               info->all_image_info_size = task->all_image_info_size;
+
+               /* only set format on output for those expecting it */
+               if (*task_info_count >= TASK_DYLD_INFO_COUNT) {
+                       info->all_image_info_format = task_has_64Bit_addr(task) ?
+                           TASK_DYLD_ALL_IMAGE_INFO_64 :
+                           TASK_DYLD_ALL_IMAGE_INFO_32;
+                       *task_info_count = TASK_DYLD_INFO_COUNT;
+               } else {
+                       *task_info_count = TASK_LEGACY_DYLD_INFO_COUNT;
+               }
+               break;
+       }
+
+       case TASK_EXTMOD_INFO:
+       {
+               task_extmod_info_t info;
+               void *p;
+
+               if (*task_info_count < TASK_EXTMOD_INFO_COUNT) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               info = (task_extmod_info_t)task_info_out;
+
+               p = get_bsdtask_info(task);
+               if (p) {
+                       proc_getexecutableuuid(p, info->task_uuid, sizeof(info->task_uuid));
+               } else {
+                       bzero(info->task_uuid, sizeof(info->task_uuid));
+               }
+               info->extmod_statistics = task->extmod_statistics;
+               *task_info_count = TASK_EXTMOD_INFO_COUNT;
+
+               break;
+       }
+
+       case TASK_KERNELMEMORY_INFO:
+       {
+               task_kernelmemory_info_t        tkm_info;
+               ledger_amount_t                 credit, debit;
+
+               if (*task_info_count < TASK_KERNELMEMORY_INFO_COUNT) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               tkm_info = (task_kernelmemory_info_t) task_info_out;
+               tkm_info->total_palloc = 0;
+               tkm_info->total_pfree = 0;
+               tkm_info->total_salloc = 0;
+               tkm_info->total_sfree = 0;
+
+               if (task == kernel_task) {
+                       /*
+                        * All shared allocs/frees from other tasks count against
+                        * the kernel private memory usage.  If we are looking up
+                        * info for the kernel task, gather from everywhere.
+                        */
+                       task_unlock(task);
+
+                       /* start by accounting for all the terminated tasks against the kernel */
+                       tkm_info->total_palloc = tasks_tkm_private.alloc + tasks_tkm_shared.alloc;
+                       tkm_info->total_pfree = tasks_tkm_private.free + tasks_tkm_shared.free;
+
+                       /* count all other task/thread shared alloc/free against the kernel */
+                       lck_mtx_lock(&tasks_threads_lock);
+
+                       /* XXX this really shouldn't be using the function parameter 'task' as a local var! */
+                       queue_iterate(&tasks, task, task_t, tasks) {
+                               if (task == kernel_task) {
+                                       if (ledger_get_entries(task->ledger,
+                                           task_ledgers.tkm_private, &credit,
+                                           &debit) == KERN_SUCCESS) {
+                                               tkm_info->total_palloc += credit;
+                                               tkm_info->total_pfree += debit;
+                                       }
+                               }
+                               if (!ledger_get_entries(task->ledger,
+                                   task_ledgers.tkm_shared, &credit, &debit)) {
+                                       tkm_info->total_palloc += credit;
+                                       tkm_info->total_pfree += debit;
+                               }
+                       }
+                       lck_mtx_unlock(&tasks_threads_lock);
+               } else {
+                       if (!ledger_get_entries(task->ledger,
+                           task_ledgers.tkm_private, &credit, &debit)) {
+                               tkm_info->total_palloc = credit;
+                               tkm_info->total_pfree = debit;
+                       }
+                       if (!ledger_get_entries(task->ledger,
+                           task_ledgers.tkm_shared, &credit, &debit)) {
+                               tkm_info->total_salloc = credit;
+                               tkm_info->total_sfree = debit;
+                       }
+                       task_unlock(task);
+               }
+
+               *task_info_count = TASK_KERNELMEMORY_INFO_COUNT;
+               return KERN_SUCCESS;
+       }
+
+       /* OBSOLETE */
+       case TASK_SCHED_FIFO_INFO:
+       {
+               if (*task_info_count < POLICY_FIFO_BASE_COUNT) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               error = KERN_INVALID_POLICY;
+               break;
+       }
+
+       /* OBSOLETE */
+       case TASK_SCHED_RR_INFO:
+       {
+               policy_rr_base_t        rr_base;
+               uint32_t quantum_time;
+               uint64_t quantum_ns;
+
+               if (*task_info_count < POLICY_RR_BASE_COUNT) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               rr_base = (policy_rr_base_t) task_info_out;
+
+               if (task != kernel_task) {
+                       error = KERN_INVALID_POLICY;
+                       break;
+               }
+
+               rr_base->base_priority = task->priority;
+
+               quantum_time = SCHED(initial_quantum_size)(THREAD_NULL);
+               absolutetime_to_nanoseconds(quantum_time, &quantum_ns);
+
+               rr_base->quantum = (uint32_t)(quantum_ns / 1000 / 1000);
+
+               *task_info_count = POLICY_RR_BASE_COUNT;
+               break;
+       }
+
+       /* OBSOLETE */
+       case TASK_SCHED_TIMESHARE_INFO:
+       {
+               policy_timeshare_base_t ts_base;
+
+               if (*task_info_count < POLICY_TIMESHARE_BASE_COUNT) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               ts_base = (policy_timeshare_base_t) task_info_out;
+
+               if (task == kernel_task) {
+                       error = KERN_INVALID_POLICY;
+                       break;
+               }
+
+               ts_base->base_priority = task->priority;
+
+               *task_info_count = POLICY_TIMESHARE_BASE_COUNT;
+               break;
+       }
+
+       case TASK_SECURITY_TOKEN:
+       {
+               security_token_t        *sec_token_p;
+
+               if (*task_info_count < TASK_SECURITY_TOKEN_COUNT) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               sec_token_p = (security_token_t *) task_info_out;
+
+               *sec_token_p = task->sec_token;
+
+               *task_info_count = TASK_SECURITY_TOKEN_COUNT;
+               break;
+       }
+
+       case TASK_AUDIT_TOKEN:
+       {
+               audit_token_t   *audit_token_p;
+
+               if (*task_info_count < TASK_AUDIT_TOKEN_COUNT) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               audit_token_p = (audit_token_t *) task_info_out;
+
+               *audit_token_p = task->audit_token;
+
+               *task_info_count = TASK_AUDIT_TOKEN_COUNT;
+               break;
+       }
+
+       case TASK_SCHED_INFO:
+               error = KERN_INVALID_ARGUMENT;
+               break;
+
+       case TASK_EVENTS_INFO:
+       {
+               task_events_info_t      events_info;
+               thread_t                        thread;
+
+               if (*task_info_count < TASK_EVENTS_INFO_COUNT) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               events_info = (task_events_info_t) task_info_out;
+
+
+               events_info->faults = (int32_t) MIN(counter_load(&task->faults), INT32_MAX);
+               events_info->pageins = task->pageins;
+               events_info->cow_faults = task->cow_faults;
+               events_info->messages_sent = task->messages_sent;
+               events_info->messages_received = task->messages_received;
+               events_info->syscalls_mach = task->syscalls_mach;
+               events_info->syscalls_unix = task->syscalls_unix;
+
+               events_info->csw = task->c_switch;
+
+               queue_iterate(&task->threads, thread, thread_t, task_threads) {
+                       events_info->csw           += thread->c_switch;
+                       events_info->syscalls_mach += thread->syscalls_mach;
+                       events_info->syscalls_unix += thread->syscalls_unix;
+               }
+
+
+               *task_info_count = TASK_EVENTS_INFO_COUNT;
+               break;
+       }
+       case TASK_AFFINITY_TAG_INFO:
+       {
+               if (*task_info_count < TASK_AFFINITY_TAG_INFO_COUNT) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               error = task_affinity_info(task, task_info_out, task_info_count);
+               break;
+       }
+       case TASK_POWER_INFO:
+       {
+               if (*task_info_count < TASK_POWER_INFO_COUNT) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               task_power_info_locked(task, (task_power_info_t)task_info_out, NULL, NULL, NULL);
+               break;
+       }
+
+       case TASK_POWER_INFO_V2:
+       {
+               if (*task_info_count < TASK_POWER_INFO_V2_COUNT_OLD) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+               task_power_info_v2_t tpiv2 = (task_power_info_v2_t) task_info_out;
+               task_power_info_locked(task, &tpiv2->cpu_energy, &tpiv2->gpu_energy, tpiv2, NULL);
+               break;
+       }
+
+       case TASK_VM_INFO:
+       case TASK_VM_INFO_PURGEABLE:
+       {
+               task_vm_info_t          vm_info;
+               vm_map_t                map;
+
+#if __arm64__
+               struct proc *p;
+               uint32_t platform, sdk;
+               p = current_proc();
+               platform = proc_platform(p);
+               sdk = proc_min_sdk(p);
+               if (original_task_info_count > TASK_VM_INFO_REV2_COUNT &&
+                   platform == PLATFORM_IOS &&
+                   sdk != 0 &&
+                   (sdk >> 16) <= 12) {
+                       /*
+                        * Some iOS apps pass an incorrect value for
+                        * task_info_count, expressed in number of bytes
+                        * instead of number of "natural_t" elements.
+                        * For the sake of backwards binary compatibility
+                        * for apps built with an iOS12 or older SDK and using
+                        * the "rev2" data structure, let's fix task_info_count
+                        * for them, to avoid stomping past the actual end
+                        * of their buffer.
+                        */
+#if DEVELOPMENT || DEBUG
+                       printf("%s:%d %d[%s] rdar://49484582 task_info_count %d -> %d platform %d sdk %d.%d.%d\n", __FUNCTION__, __LINE__, proc_pid(p), proc_name_address(p), original_task_info_count, TASK_VM_INFO_REV2_COUNT, platform, (sdk >> 16), ((sdk >> 8) & 0xff), (sdk & 0xff));
+#endif /* DEVELOPMENT || DEBUG */
+                       DTRACE_VM4(workaround_task_vm_info_count,
+                           mach_msg_type_number_t, original_task_info_count,
+                           mach_msg_type_number_t, TASK_VM_INFO_REV2_COUNT,
+                           uint32_t, platform,
+                           uint32_t, sdk);
+                       original_task_info_count = TASK_VM_INFO_REV2_COUNT;
+                       *task_info_count = original_task_info_count;
+               }
+#endif /* __arm64__ */
+
+               if (*task_info_count < TASK_VM_INFO_REV0_COUNT) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               vm_info = (task_vm_info_t)task_info_out;
+
+               /*
+                * Do not hold both the task and map locks,
+                * so convert the task lock into a map reference,
+                * drop the task lock, then lock the map.
+                */
+               if (is_kernel_task) {
+                       map = kernel_map;
+                       task_unlock(task);
+                       /* no lock, no reference */
+               } else {
+                       map = task->map;
+                       vm_map_reference(map);
+                       task_unlock(task);
+                       vm_map_lock_read(map);
+               }
+
+               vm_info->virtual_size = (typeof(vm_info->virtual_size))vm_map_adjusted_size(map);
+               vm_info->region_count = map->hdr.nentries;
+               vm_info->page_size = vm_map_page_size(map);
+
+               vm_info->resident_size = pmap_resident_count(map->pmap);
+               vm_info->resident_size *= PAGE_SIZE;
+               vm_info->resident_size_peak = pmap_resident_max(map->pmap);
+               vm_info->resident_size_peak *= PAGE_SIZE;
+
+#define _VM_INFO(_name) \
+       vm_info->_name = ((mach_vm_size_t) map->pmap->stats._name) * PAGE_SIZE
+
+               _VM_INFO(device);
+               _VM_INFO(device_peak);
+               _VM_INFO(external);
+               _VM_INFO(external_peak);
+               _VM_INFO(internal);
+               _VM_INFO(internal_peak);
+               _VM_INFO(reusable);
+               _VM_INFO(reusable_peak);
+               _VM_INFO(compressed);
+               _VM_INFO(compressed_peak);
+               _VM_INFO(compressed_lifetime);
+
+               vm_info->purgeable_volatile_pmap = 0;
+               vm_info->purgeable_volatile_resident = 0;
+               vm_info->purgeable_volatile_virtual = 0;
+               if (is_kernel_task) {
+                       /*
+                        * We do not maintain the detailed stats for the
+                        * kernel_pmap, so just count everything as
+                        * "internal"...
+                        */
+                       vm_info->internal = vm_info->resident_size;
+                       /*
+                        * ... but since the memory held by the VM compressor
+                        * in the kernel address space ought to be attributed
+                        * to user-space tasks, we subtract it from "internal"
+                        * to give memory reporting tools a more accurate idea
+                        * of what the kernel itself is actually using, instead
+                        * of making it look like the kernel is leaking memory
+                        * when the system is under memory pressure.
+                        */
+                       vm_info->internal -= (VM_PAGE_COMPRESSOR_COUNT *
+                           PAGE_SIZE);
+               } else {
+                       mach_vm_size_t  volatile_virtual_size;
+                       mach_vm_size_t  volatile_resident_size;
+                       mach_vm_size_t  volatile_compressed_size;
+                       mach_vm_size_t  volatile_pmap_size;
+                       mach_vm_size_t  volatile_compressed_pmap_size;
+                       kern_return_t   kr;
+
+                       if (flavor == TASK_VM_INFO_PURGEABLE) {
+                               kr = vm_map_query_volatile(
+                                       map,
+                                       &volatile_virtual_size,
+                                       &volatile_resident_size,
+                                       &volatile_compressed_size,
+                                       &volatile_pmap_size,
+                                       &volatile_compressed_pmap_size);
+                               if (kr == KERN_SUCCESS) {
+                                       vm_info->purgeable_volatile_pmap =
+                                           volatile_pmap_size;
+                                       if (radar_20146450) {
+                                               vm_info->compressed -=
+                                                   volatile_compressed_pmap_size;
+                                       }
+                                       vm_info->purgeable_volatile_resident =
+                                           volatile_resident_size;
+                                       vm_info->purgeable_volatile_virtual =
+                                           volatile_virtual_size;
+                               }
+                       }
+               }
+               *task_info_count = TASK_VM_INFO_REV0_COUNT;
+
+               if (original_task_info_count >= TASK_VM_INFO_REV2_COUNT) {
+                       /* must be captured while we still have the map lock */
+                       vm_info->min_address = map->min_offset;
+                       vm_info->max_address = map->max_offset;
+               }
+
+               /*
+                * Done with vm map things, can drop the map lock and reference,
+                * and take the task lock back.
+                *
+                * Re-validate that the task didn't die on us.
+                */
+               if (!is_kernel_task) {
+                       vm_map_unlock_read(map);
+                       vm_map_deallocate(map);
+               }
+               map = VM_MAP_NULL;
+
+               task_lock(task);
+
+               if ((task != current_task()) && (!task->active)) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               if (original_task_info_count >= TASK_VM_INFO_REV1_COUNT) {
+                       vm_info->phys_footprint =
+                           (mach_vm_size_t) get_task_phys_footprint(task);
+                       *task_info_count = TASK_VM_INFO_REV1_COUNT;
+               }
+               if (original_task_info_count >= TASK_VM_INFO_REV2_COUNT) {
+                       /* data was captured above */
+                       *task_info_count = TASK_VM_INFO_REV2_COUNT;
+               }
+
+               if (original_task_info_count >= TASK_VM_INFO_REV3_COUNT) {
+                       ledger_get_lifetime_max(task->ledger,
+                           task_ledgers.phys_footprint,
+                           &vm_info->ledger_phys_footprint_peak);
+                       ledger_get_balance(task->ledger,
+                           task_ledgers.purgeable_nonvolatile,
+                           &vm_info->ledger_purgeable_nonvolatile);
+                       ledger_get_balance(task->ledger,
+                           task_ledgers.purgeable_nonvolatile_compressed,
+                           &vm_info->ledger_purgeable_novolatile_compressed);
+                       ledger_get_balance(task->ledger,
+                           task_ledgers.purgeable_volatile,
+                           &vm_info->ledger_purgeable_volatile);
+                       ledger_get_balance(task->ledger,
+                           task_ledgers.purgeable_volatile_compressed,
+                           &vm_info->ledger_purgeable_volatile_compressed);
+                       ledger_get_balance(task->ledger,
+                           task_ledgers.network_nonvolatile,
+                           &vm_info->ledger_tag_network_nonvolatile);
+                       ledger_get_balance(task->ledger,
+                           task_ledgers.network_nonvolatile_compressed,
+                           &vm_info->ledger_tag_network_nonvolatile_compressed);
+                       ledger_get_balance(task->ledger,
+                           task_ledgers.network_volatile,
+                           &vm_info->ledger_tag_network_volatile);
+                       ledger_get_balance(task->ledger,
+                           task_ledgers.network_volatile_compressed,
+                           &vm_info->ledger_tag_network_volatile_compressed);
+                       ledger_get_balance(task->ledger,
+                           task_ledgers.media_footprint,
+                           &vm_info->ledger_tag_media_footprint);
+                       ledger_get_balance(task->ledger,
+                           task_ledgers.media_footprint_compressed,
+                           &vm_info->ledger_tag_media_footprint_compressed);
+                       ledger_get_balance(task->ledger,
+                           task_ledgers.media_nofootprint,
+                           &vm_info->ledger_tag_media_nofootprint);
+                       ledger_get_balance(task->ledger,
+                           task_ledgers.media_nofootprint_compressed,
+                           &vm_info->ledger_tag_media_nofootprint_compressed);
+                       ledger_get_balance(task->ledger,
+                           task_ledgers.graphics_footprint,
+                           &vm_info->ledger_tag_graphics_footprint);
+                       ledger_get_balance(task->ledger,
+                           task_ledgers.graphics_footprint_compressed,
+                           &vm_info->ledger_tag_graphics_footprint_compressed);
+                       ledger_get_balance(task->ledger,
+                           task_ledgers.graphics_nofootprint,
+                           &vm_info->ledger_tag_graphics_nofootprint);
+                       ledger_get_balance(task->ledger,
+                           task_ledgers.graphics_nofootprint_compressed,
+                           &vm_info->ledger_tag_graphics_nofootprint_compressed);
+                       ledger_get_balance(task->ledger,
+                           task_ledgers.neural_footprint,
+                           &vm_info->ledger_tag_neural_footprint);
+                       ledger_get_balance(task->ledger,
+                           task_ledgers.neural_footprint_compressed,
+                           &vm_info->ledger_tag_neural_footprint_compressed);
+                       ledger_get_balance(task->ledger,
+                           task_ledgers.neural_nofootprint,
+                           &vm_info->ledger_tag_neural_nofootprint);
+                       ledger_get_balance(task->ledger,
+                           task_ledgers.neural_nofootprint_compressed,
+                           &vm_info->ledger_tag_neural_nofootprint_compressed);
+                       *task_info_count = TASK_VM_INFO_REV3_COUNT;
+               }
+               if (original_task_info_count >= TASK_VM_INFO_REV4_COUNT) {
+                       if (task->bsd_info) {
+                               vm_info->limit_bytes_remaining =
+                                   memorystatus_available_memory_internal(task->bsd_info);
+                       } else {
+                               vm_info->limit_bytes_remaining = 0;
+                       }
+                       *task_info_count = TASK_VM_INFO_REV4_COUNT;
+               }
+               if (original_task_info_count >= TASK_VM_INFO_REV5_COUNT) {
+                       thread_t thread;
+                       integer_t total = task->decompressions;
+                       queue_iterate(&task->threads, thread, thread_t, task_threads) {
+                               total += thread->decompressions;
+                       }
+                       vm_info->decompressions = total;
+                       *task_info_count = TASK_VM_INFO_REV5_COUNT;
+               }
+
+               break;
+       }
+
+       case TASK_WAIT_STATE_INFO:
+       {
+               /*
+                * Deprecated flavor. Currently allowing some results until all users
+                * stop calling it. The results may not be accurate.
+                */
+               task_wait_state_info_t  wait_state_info;
+               uint64_t total_sfi_ledger_val = 0;
+
+               if (*task_info_count < TASK_WAIT_STATE_INFO_COUNT) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               wait_state_info = (task_wait_state_info_t) task_info_out;
+
+               wait_state_info->total_wait_state_time = 0;
+               bzero(wait_state_info->_reserved, sizeof(wait_state_info->_reserved));
+
+#if CONFIG_SCHED_SFI
+               int i, prev_lentry = -1;
+               int64_t  val_credit, val_debit;
+
+               for (i = 0; i < MAX_SFI_CLASS_ID; i++) {
+                       val_credit = 0;
+                       /*
+                        * checking with prev_lentry != entry ensures adjacent classes
+                        * which share the same ledger do not add wait times twice.
+                        * Note: Use ledger() call to get data for each individual sfi class.
+                        */
+                       if (prev_lentry != task_ledgers.sfi_wait_times[i] &&
+                           KERN_SUCCESS == ledger_get_entries(task->ledger,
+                           task_ledgers.sfi_wait_times[i], &val_credit, &val_debit)) {
+                               total_sfi_ledger_val += val_credit;
+                       }
+                       prev_lentry = task_ledgers.sfi_wait_times[i];
+               }
+
+#endif /* CONFIG_SCHED_SFI */
+               wait_state_info->total_wait_sfi_state_time = total_sfi_ledger_val;
+               *task_info_count = TASK_WAIT_STATE_INFO_COUNT;
+
+               break;
+       }
+       case TASK_VM_INFO_PURGEABLE_ACCOUNT:
+       {
+#if DEVELOPMENT || DEBUG
+               pvm_account_info_t      acnt_info;
+
+               if (*task_info_count < PVM_ACCOUNT_INFO_COUNT) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               if (task_info_out == NULL) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               acnt_info = (pvm_account_info_t) task_info_out;
+
+               error = vm_purgeable_account(task, acnt_info);
+
+               *task_info_count = PVM_ACCOUNT_INFO_COUNT;
+
+               break;
+#else /* DEVELOPMENT || DEBUG */
+               error = KERN_NOT_SUPPORTED;
+               break;
+#endif /* DEVELOPMENT || DEBUG */
+       }
+       case TASK_FLAGS_INFO:
+       {
+               task_flags_info_t               flags_info;
+
+               if (*task_info_count < TASK_FLAGS_INFO_COUNT) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+
+               flags_info = (task_flags_info_t)task_info_out;
+
+               /* only publish the 64-bit flag of the task */
+               flags_info->flags = task->t_flags & (TF_64B_ADDR | TF_64B_DATA);
+
+               *task_info_count = TASK_FLAGS_INFO_COUNT;
+               break;
+       }
+
+       case TASK_DEBUG_INFO_INTERNAL:
+       {
+#if DEVELOPMENT || DEBUG
+               task_debug_info_internal_t dbg_info;
+               ipc_space_t space = task->itk_space;
+               if (*task_info_count < TASK_DEBUG_INFO_INTERNAL_COUNT) {
+                       error = KERN_NOT_SUPPORTED;
+                       break;
+               }
+
+               if (task_info_out == NULL) {
+                       error = KERN_INVALID_ARGUMENT;
+                       break;
+               }
+               dbg_info = (task_debug_info_internal_t) task_info_out;
+               dbg_info->ipc_space_size = 0;
+
+               if (space) {
+                       is_read_lock(space);
+                       dbg_info->ipc_space_size = space->is_table_size;
+                       is_read_unlock(space);
+               }
+
+               dbg_info->suspend_count = task->suspend_count;
+
+               error = KERN_SUCCESS;
+               *task_info_count = TASK_DEBUG_INFO_INTERNAL_COUNT;
+               break;
+#else /* DEVELOPMENT || DEBUG */
+               error = KERN_NOT_SUPPORTED;
+               break;
+#endif /* DEVELOPMENT || DEBUG */
+       }
+       default:
+               error = KERN_INVALID_ARGUMENT;
+       }
+
+       task_unlock(task);
+       return error;
+}
+
+/*
+ * task_info_from_user
+ *
+ * When calling task_info from user space,
+ * this function will be executed as mig server side
+ * instead of calling directly into task_info.
+ * This gives the possibility to perform more security
+ * checks on task_port.
+ *
+ * In the case of TASK_DYLD_INFO, we require the more
+ * privileged task_read_port not the less-privileged task_name_port.
+ *
+ */
+kern_return_t
+task_info_from_user(
+       mach_port_t             task_port,
+       task_flavor_t           flavor,
+       task_info_t             task_info_out,
+       mach_msg_type_number_t  *task_info_count)
+{
+       task_t task;
+       kern_return_t ret;
+
+       if (flavor == TASK_DYLD_INFO) {
+               task = convert_port_to_task_read(task_port);
+       } else {
+               task = convert_port_to_task_name(task_port);
+       }
+
+       ret = task_info(task, flavor, task_info_out, task_info_count);
+
+       task_deallocate(task);
+
+       return ret;
+}
+
+/*
+ * Routine: task_dyld_process_info_update_helper
+ *
+ * Release send rights in release_ports.
+ *
+ * If no active ports found in task's dyld notifier array, unset the magic value
+ * in user space to indicate so.
+ *
+ * Condition:
+ *      task's itk_lock is locked, and is unlocked upon return.
+ *      Global g_dyldinfo_mtx is locked, and is unlocked upon return.
+ */
+void
+task_dyld_process_info_update_helper(
+       task_t                  task,
+       size_t                  active_count,
+       vm_map_address_t        magic_addr,    /* a userspace address */
+       ipc_port_t             *release_ports,
+       size_t                  release_count)
+{
+       void *notifiers_ptr = NULL;
+
+       assert(release_count <= DYLD_MAX_PROCESS_INFO_NOTIFY_COUNT);
+
+       if (active_count == 0) {
+               assert(task->itk_dyld_notify != NULL);
+               notifiers_ptr = task->itk_dyld_notify;
+               task->itk_dyld_notify = NULL;
+               itk_unlock(task);
+
+               kfree(notifiers_ptr, (vm_size_t)sizeof(ipc_port_t) * DYLD_MAX_PROCESS_INFO_NOTIFY_COUNT);
+               (void)copyoutmap_atomic32(task->map, MACH_PORT_NULL, magic_addr); /* unset magic */
+       } else {
+               itk_unlock(task);
+               (void)copyoutmap_atomic32(task->map, (mach_port_name_t)DYLD_PROCESS_INFO_NOTIFY_MAGIC,
+                   magic_addr);     /* reset magic */
+       }
+
+       lck_mtx_unlock(&g_dyldinfo_mtx);
+
+       for (size_t i = 0; i < release_count; i++) {
+               ipc_port_release_send(release_ports[i]);
+       }
+}
+
+/*
+ * Routine: task_dyld_process_info_notify_register
+ *
+ * Insert a send right to target task's itk_dyld_notify array. Allocate kernel
+ * memory for the array if it's the first port to be registered. Also cleanup
+ * any dead rights found in the array.
+ *
+ * Consumes sright if returns KERN_SUCCESS, otherwise MIG will destroy it.
+ *
+ * Args:
+ *     task:   Target task for the registration.
+ *     sright: A send right.
+ *
+ * Returns:
+ *     KERN_SUCCESS: Registration succeeded.
+ *     KERN_INVALID_TASK: task is invalid.
+ *     KERN_INVALID_RIGHT: sright is invalid.
+ *     KERN_DENIED: Security policy denied this call.
+ *     KERN_RESOURCE_SHORTAGE: Kernel memory allocation failed.
+ *     KERN_NO_SPACE: No available notifier port slot left for this task.
+ *     KERN_RIGHT_EXISTS: The notifier port is already registered and active.
+ *
+ *     Other error code see task_info().
+ *
+ * See Also:
+ *     task_dyld_process_info_notify_get_trap() in mach_kernelrpc.c
+ */
+kern_return_t
+task_dyld_process_info_notify_register(
+       task_t                  task,
+       ipc_port_t              sright)
+{
+       struct task_dyld_info dyld_info;
+       mach_msg_type_number_t info_count = TASK_DYLD_INFO_COUNT;
+       ipc_port_t release_ports[DYLD_MAX_PROCESS_INFO_NOTIFY_COUNT];
+       uint32_t release_count = 0, active_count = 0;
+       mach_vm_address_t ports_addr; /* a user space address */
+       kern_return_t kr;
+       boolean_t right_exists = false;
+       ipc_port_t *notifiers_ptr = NULL;
+       ipc_port_t *portp;
+
+       if (task == TASK_NULL || task == kernel_task) {
+               return KERN_INVALID_TASK;
+       }
+
+       if (!IP_VALID(sright)) {
+               return KERN_INVALID_RIGHT;
+       }
+
+#if CONFIG_MACF
+       if (mac_task_check_dyld_process_info_notify_register()) {
+               return KERN_DENIED;
+       }
+#endif
+
+       kr = task_info(task, TASK_DYLD_INFO, (task_info_t)&dyld_info, &info_count);
+       if (kr) {
+               return kr;
+       }
+
+       if (dyld_info.all_image_info_format == TASK_DYLD_ALL_IMAGE_INFO_32) {
+               ports_addr = (mach_vm_address_t)(dyld_info.all_image_info_addr +
+                   offsetof(struct user32_dyld_all_image_infos, notifyMachPorts));
+       } else {
+               ports_addr = (mach_vm_address_t)(dyld_info.all_image_info_addr +
+                   offsetof(struct user64_dyld_all_image_infos, notifyMachPorts));
+       }
+
+       if (task->itk_dyld_notify == NULL) {
+               notifiers_ptr = (ipc_port_t *)
+                   kalloc_flags(sizeof(ipc_port_t) * DYLD_MAX_PROCESS_INFO_NOTIFY_COUNT, Z_ZERO);
+               if (!notifiers_ptr) {
+                       return KERN_RESOURCE_SHORTAGE;
+               }
+       }
+
+       lck_mtx_lock(&g_dyldinfo_mtx);
+       itk_lock(task);
+
+       if (task->itk_dyld_notify == NULL) {
+               task->itk_dyld_notify = notifiers_ptr;
+               notifiers_ptr = NULL;
+       }
+
+       assert(task->itk_dyld_notify != NULL);
+       /* First pass: clear dead names and check for duplicate registration */
+       for (int slot = 0; slot < DYLD_MAX_PROCESS_INFO_NOTIFY_COUNT; slot++) {
+               portp = &task->itk_dyld_notify[slot];
+               if (*portp != IPC_PORT_NULL && !ip_active(*portp)) {
+                       release_ports[release_count++] = *portp;
+                       *portp = IPC_PORT_NULL;
+               } else if (*portp == sright) {
+                       /* the port is already registered and is active */
+                       right_exists = true;
+               }
+
+               if (*portp != IPC_PORT_NULL) {
+                       active_count++;
+               }
+       }
+
+       if (right_exists) {
+               /* skip second pass */
+               kr = KERN_RIGHT_EXISTS;
+               goto out;
+       }
+
+       /* Second pass: register the port */
+       kr = KERN_NO_SPACE;
+       for (int slot = 0; slot < DYLD_MAX_PROCESS_INFO_NOTIFY_COUNT; slot++) {
+               portp = &task->itk_dyld_notify[slot];
+               if (*portp == IPC_PORT_NULL) {
+                       *portp = sright;
+                       active_count++;
+                       kr = KERN_SUCCESS;
+                       break;
+               }
+       }
+
+out:
+       assert(active_count > 0);
+
+       task_dyld_process_info_update_helper(task, active_count,
+           (vm_map_address_t)ports_addr, release_ports, release_count);
+       /* itk_lock, g_dyldinfo_mtx are unlocked upon return */
+
+       if (notifiers_ptr) {
+               kfree(notifiers_ptr, sizeof(ipc_port_t) * DYLD_MAX_PROCESS_INFO_NOTIFY_COUNT);
+       }
+
+       return kr;
+}
+
+/*
+ * Routine: task_dyld_process_info_notify_deregister
+ *
+ * Remove a send right in target task's itk_dyld_notify array matching the receive
+ * right name passed in. Deallocate kernel memory for the array if it's the last port to
+ * be deregistered, or all ports have died. Also cleanup any dead rights found in the array.
+ *
+ * Does not consume any reference.
+ *
+ * Args:
+ *     task: Target task for the deregistration.
+ *     rcv_name: The name denoting the receive right in caller's space.
+ *
+ * Returns:
+ *     KERN_SUCCESS: A matching entry found and degistration succeeded.
+ *     KERN_INVALID_TASK: task is invalid.
+ *     KERN_INVALID_NAME: name is invalid.
+ *     KERN_DENIED: Security policy denied this call.
+ *     KERN_FAILURE: A matching entry is not found.
+ *     KERN_INVALID_RIGHT: The name passed in does not represent a valid rcv right.
+ *
+ *     Other error code see task_info().
+ *
+ * See Also:
+ *     task_dyld_process_info_notify_get_trap() in mach_kernelrpc.c
+ */
+kern_return_t
+task_dyld_process_info_notify_deregister(
+       task_t                  task,
+       mach_port_name_t        rcv_name)
+{
+       struct task_dyld_info dyld_info;
+       mach_msg_type_number_t info_count = TASK_DYLD_INFO_COUNT;
+       ipc_port_t release_ports[DYLD_MAX_PROCESS_INFO_NOTIFY_COUNT];
+       uint32_t release_count = 0, active_count = 0;
+       boolean_t port_found = false;
+       mach_vm_address_t ports_addr; /* a user space address */
+       ipc_port_t sright;
+       kern_return_t kr;
+       ipc_port_t *portp;
+
+       if (task == TASK_NULL || task == kernel_task) {
+               return KERN_INVALID_TASK;
+       }
+
+       if (!MACH_PORT_VALID(rcv_name)) {
+               return KERN_INVALID_NAME;
+       }
+
+#if CONFIG_MACF
+       if (mac_task_check_dyld_process_info_notify_register()) {
+               return KERN_DENIED;
+       }
+#endif
+
+       kr = task_info(task, TASK_DYLD_INFO, (task_info_t)&dyld_info, &info_count);
+       if (kr) {
+               return kr;
+       }
+
+       if (dyld_info.all_image_info_format == TASK_DYLD_ALL_IMAGE_INFO_32) {
+               ports_addr = (mach_vm_address_t)(dyld_info.all_image_info_addr +
+                   offsetof(struct user32_dyld_all_image_infos, notifyMachPorts));
+       } else {
+               ports_addr = (mach_vm_address_t)(dyld_info.all_image_info_addr +
+                   offsetof(struct user64_dyld_all_image_infos, notifyMachPorts));
+       }
+
+       kr = ipc_port_translate_receive(current_space(), rcv_name, &sright); /* does not produce port ref */
+       if (kr) {
+               return KERN_INVALID_RIGHT;
+       }
+
+       ip_reference(sright);
+       ip_unlock(sright);
+
+       assert(sright != IPC_PORT_NULL);
+
+       lck_mtx_lock(&g_dyldinfo_mtx);
+       itk_lock(task);
+
+       if (task->itk_dyld_notify == NULL) {
+               itk_unlock(task);
+               lck_mtx_unlock(&g_dyldinfo_mtx);
+               ip_release(sright);
+               return KERN_FAILURE;
+       }
+
+       for (int slot = 0; slot < DYLD_MAX_PROCESS_INFO_NOTIFY_COUNT; slot++) {
+               portp = &task->itk_dyld_notify[slot];
+               if (*portp == sright) {
+                       release_ports[release_count++] = *portp;
+                       *portp = IPC_PORT_NULL;
+                       port_found = true;
+               } else if ((*portp != IPC_PORT_NULL && !ip_active(*portp))) {
+                       release_ports[release_count++] = *portp;
+                       *portp = IPC_PORT_NULL;
+               }
+
+               if (*portp != IPC_PORT_NULL) {
+                       active_count++;
+               }
+       }
+
+       task_dyld_process_info_update_helper(task, active_count,
+           (vm_map_address_t)ports_addr, release_ports, release_count);
+       /* itk_lock, g_dyldinfo_mtx are unlocked upon return */
+
+       ip_release(sright);
+
+       return port_found ? KERN_SUCCESS : KERN_FAILURE;
+}
+
+/*
+ *     task_power_info
+ *
+ *     Returns power stats for the task.
+ *     Note: Called with task locked.
+ */
+void
+task_power_info_locked(
+       task_t                  task,
+       task_power_info_t       info,
+       gpu_energy_data_t       ginfo,
+       task_power_info_v2_t    infov2,
+       uint64_t                *runnable_time)
+{
+       thread_t                thread;
+       ledger_amount_t         tmp;
+
+       uint64_t                runnable_time_sum = 0;
+
+       task_lock_assert_owned(task);
+
+       ledger_get_entries(task->ledger, task_ledgers.interrupt_wakeups,
+           (ledger_amount_t *)&info->task_interrupt_wakeups, &tmp);
+       ledger_get_entries(task->ledger, task_ledgers.platform_idle_wakeups,
+           (ledger_amount_t *)&info->task_platform_idle_wakeups, &tmp);
+
+       info->task_timer_wakeups_bin_1 = task->task_timer_wakeups_bin_1;
+       info->task_timer_wakeups_bin_2 = task->task_timer_wakeups_bin_2;
+
+       info->total_user = task->total_user_time;
+       info->total_system = task->total_system_time;
+       runnable_time_sum = task->total_runnable_time;
+
+#if defined(__arm__) || defined(__arm64__)
+       if (infov2) {
+               infov2->task_energy = task->task_energy;
+       }
+#endif /* defined(__arm__) || defined(__arm64__) */
+
+       if (ginfo) {
+               ginfo->task_gpu_utilisation = task->task_gpu_ns;
+       }
+
+       if (infov2) {
+               infov2->task_ptime = task->total_ptime;
+               infov2->task_pset_switches = task->ps_switch;
+       }
+
+       queue_iterate(&task->threads, thread, thread_t, task_threads) {
+               uint64_t        tval;
+               spl_t           x;
+
+               if (thread->options & TH_OPT_IDLE_THREAD) {
+                       continue;
+               }
+
+               x = splsched();
+               thread_lock(thread);
+
+               info->task_timer_wakeups_bin_1 += thread->thread_timer_wakeups_bin_1;
+               info->task_timer_wakeups_bin_2 += thread->thread_timer_wakeups_bin_2;
+
+#if defined(__arm__) || defined(__arm64__)
+               if (infov2) {
+                       infov2->task_energy += ml_energy_stat(thread);
+               }
+#endif /* defined(__arm__) || defined(__arm64__) */
+
+               tval = timer_grab(&thread->user_timer);
+               info->total_user += tval;
+
+               if (infov2) {
+                       tval = timer_grab(&thread->ptime);
+                       infov2->task_ptime += tval;
+                       infov2->task_pset_switches += thread->ps_switch;
+               }
+
+               tval = timer_grab(&thread->system_timer);
+               if (thread->precise_user_kernel_time) {
+                       info->total_system += tval;
+               } else {
+                       /* system_timer may represent either sys or user */
+                       info->total_user += tval;
+               }
+
+               tval = timer_grab(&thread->runnable_timer);
+
+               runnable_time_sum += tval;
+
+               if (ginfo) {
+                       ginfo->task_gpu_utilisation += ml_gpu_stat(thread);
+               }
+               thread_unlock(thread);
+               splx(x);
+       }
+
+       if (runnable_time) {
+               *runnable_time = runnable_time_sum;
+       }
+}
+
+/*
+ *     task_gpu_utilisation
+ *
+ *     Returns the total gpu time used by the all the threads of the task
+ *  (both dead and alive)
+ */
+uint64_t
+task_gpu_utilisation(
+       task_t  task)
+{
+       uint64_t gpu_time = 0;
+#if defined(__x86_64__)
+       thread_t thread;
+
+       task_lock(task);
+       gpu_time += task->task_gpu_ns;
+
+       queue_iterate(&task->threads, thread, thread_t, task_threads) {
+               spl_t x;
+               x = splsched();
+               thread_lock(thread);
+               gpu_time += ml_gpu_stat(thread);
+               thread_unlock(thread);
+               splx(x);
+       }
+
+       task_unlock(task);
+#else /* defined(__x86_64__) */
+       /* silence compiler warning */
+       (void)task;
+#endif /* defined(__x86_64__) */
+       return gpu_time;
+}
+
+/*
+ *     task_energy
+ *
+ *     Returns the total energy used by the all the threads of the task
+ *  (both dead and alive)
+ */
+uint64_t
+task_energy(
+       task_t  task)
+{
+       uint64_t energy = 0;
+       thread_t thread;
+
+       task_lock(task);
+       energy += task->task_energy;
+
+       queue_iterate(&task->threads, thread, thread_t, task_threads) {
+               spl_t x;
+               x = splsched();
+               thread_lock(thread);
+               energy += ml_energy_stat(thread);
+               thread_unlock(thread);
+               splx(x);
+       }
+
+       task_unlock(task);
+       return energy;
+}
+
+#if __AMP__
+
+uint64_t
+task_cpu_ptime(
+       task_t  task)
+{
+       uint64_t cpu_ptime = 0;
+       thread_t thread;
+
+       task_lock(task);
+       cpu_ptime += task->total_ptime;
+
+       queue_iterate(&task->threads, thread, thread_t, task_threads) {
+               cpu_ptime += timer_grab(&thread->ptime);
+       }
+
+       task_unlock(task);
+       return cpu_ptime;
+}
+
+#else /* __AMP__ */
+
+uint64_t
+task_cpu_ptime(
+       __unused task_t  task)
+{
+       return 0;
+}
+
+#endif /* __AMP__ */
+
+/* This function updates the cpu time in the arrays for each
+ * effective and requested QoS class
+ */
+void
+task_update_cpu_time_qos_stats(
+       task_t  task,
+       uint64_t *eqos_stats,
+       uint64_t *rqos_stats)
+{
+       if (!eqos_stats && !rqos_stats) {
+               return;
+       }
+
+       task_lock(task);
+       thread_t thread;
+       queue_iterate(&task->threads, thread, thread_t, task_threads) {
+               if (thread->options & TH_OPT_IDLE_THREAD) {
+                       continue;
+               }
+
+               thread_update_qos_cpu_time(thread);
+       }
+
+       if (eqos_stats) {
+               eqos_stats[THREAD_QOS_DEFAULT] += task->cpu_time_eqos_stats.cpu_time_qos_default;
+               eqos_stats[THREAD_QOS_MAINTENANCE] += task->cpu_time_eqos_stats.cpu_time_qos_maintenance;
+               eqos_stats[THREAD_QOS_BACKGROUND] += task->cpu_time_eqos_stats.cpu_time_qos_background;
+               eqos_stats[THREAD_QOS_UTILITY] += task->cpu_time_eqos_stats.cpu_time_qos_utility;
+               eqos_stats[THREAD_QOS_LEGACY] += task->cpu_time_eqos_stats.cpu_time_qos_legacy;
+               eqos_stats[THREAD_QOS_USER_INITIATED] += task->cpu_time_eqos_stats.cpu_time_qos_user_initiated;
+               eqos_stats[THREAD_QOS_USER_INTERACTIVE] += task->cpu_time_eqos_stats.cpu_time_qos_user_interactive;
+       }
+
+       if (rqos_stats) {
+               rqos_stats[THREAD_QOS_DEFAULT] += task->cpu_time_rqos_stats.cpu_time_qos_default;
+               rqos_stats[THREAD_QOS_MAINTENANCE] += task->cpu_time_rqos_stats.cpu_time_qos_maintenance;
+               rqos_stats[THREAD_QOS_BACKGROUND] += task->cpu_time_rqos_stats.cpu_time_qos_background;
+               rqos_stats[THREAD_QOS_UTILITY] += task->cpu_time_rqos_stats.cpu_time_qos_utility;
+               rqos_stats[THREAD_QOS_LEGACY] += task->cpu_time_rqos_stats.cpu_time_qos_legacy;
+               rqos_stats[THREAD_QOS_USER_INITIATED] += task->cpu_time_rqos_stats.cpu_time_qos_user_initiated;
+               rqos_stats[THREAD_QOS_USER_INTERACTIVE] += task->cpu_time_rqos_stats.cpu_time_qos_user_interactive;
+       }
+
+       task_unlock(task);
+}
+
+kern_return_t
+task_purgable_info(
+       task_t                  task,
+       task_purgable_info_t    *stats)
+{
+       if (task == TASK_NULL || stats == NULL) {
+               return KERN_INVALID_ARGUMENT;
+       }
+       /* Take task reference */
+       task_reference(task);
+       vm_purgeable_stats((vm_purgeable_info_t)stats, task);
+       /* Drop task reference */
+       task_deallocate(task);
+       return KERN_SUCCESS;
+}
+
+void
+task_vtimer_set(
+       task_t          task,
+       integer_t       which)
+{
+       thread_t        thread;
+       spl_t           x;
+
+       task_lock(task);
+
+       task->vtimers |= which;
+
+       switch (which) {
+       case TASK_VTIMER_USER:
+               queue_iterate(&task->threads, thread, thread_t, task_threads) {
+                       x = splsched();
+                       thread_lock(thread);
+                       if (thread->precise_user_kernel_time) {
+                               thread->vtimer_user_save = timer_grab(&thread->user_timer);
+                       } else {
+                               thread->vtimer_user_save = timer_grab(&thread->system_timer);
+                       }
+                       thread_unlock(thread);
+                       splx(x);
+               }
+               break;
+
+       case TASK_VTIMER_PROF:
+               queue_iterate(&task->threads, thread, thread_t, task_threads) {
+                       x = splsched();
+                       thread_lock(thread);
+                       thread->vtimer_prof_save = timer_grab(&thread->user_timer);
+                       thread->vtimer_prof_save += timer_grab(&thread->system_timer);
+                       thread_unlock(thread);
+                       splx(x);
+               }
+               break;
+
+       case TASK_VTIMER_RLIM:
+               queue_iterate(&task->threads, thread, thread_t, task_threads) {
+                       x = splsched();
+                       thread_lock(thread);
+                       thread->vtimer_rlim_save = timer_grab(&thread->user_timer);
+                       thread->vtimer_rlim_save += timer_grab(&thread->system_timer);
+                       thread_unlock(thread);
+                       splx(x);
+               }
+               break;
+       }
+
+       task_unlock(task);
+}
+
+void
+task_vtimer_clear(
+       task_t          task,
+       integer_t       which)
+{
+       assert(task == current_task());
+
+       task_lock(task);
+
+       task->vtimers &= ~which;
+
+       task_unlock(task);
+}
+
+void
+task_vtimer_update(
+       __unused
+       task_t          task,
+       integer_t       which,
+       uint32_t        *microsecs)
+{
+       thread_t        thread = current_thread();
+       uint32_t        tdelt = 0;
+       clock_sec_t     secs = 0;
+       uint64_t        tsum;
+
+       assert(task == current_task());
+
+       spl_t s = splsched();
+       thread_lock(thread);
+
+       if ((task->vtimers & which) != (uint32_t)which) {
+               thread_unlock(thread);
+               splx(s);
+               return;
+       }
+
+       switch (which) {
+       case TASK_VTIMER_USER:
+               if (thread->precise_user_kernel_time) {
+                       tdelt = (uint32_t)timer_delta(&thread->user_timer,
+                           &thread->vtimer_user_save);
+               } else {
+                       tdelt = (uint32_t)timer_delta(&thread->system_timer,
+                           &thread->vtimer_user_save);
+               }
+               absolutetime_to_microtime(tdelt, &secs, microsecs);
+               break;
+
+       case TASK_VTIMER_PROF:
+               tsum = timer_grab(&thread->user_timer);
+               tsum += timer_grab(&thread->system_timer);
+               tdelt = (uint32_t)(tsum - thread->vtimer_prof_save);
+               absolutetime_to_microtime(tdelt, &secs, microsecs);
+               /* if the time delta is smaller than a usec, ignore */
+               if (*microsecs != 0) {
+                       thread->vtimer_prof_save = tsum;
+               }
+               break;
+
+       case TASK_VTIMER_RLIM:
+               tsum = timer_grab(&thread->user_timer);
+               tsum += timer_grab(&thread->system_timer);
+               tdelt = (uint32_t)(tsum - thread->vtimer_rlim_save);
+               thread->vtimer_rlim_save = tsum;
+               absolutetime_to_microtime(tdelt, &secs, microsecs);
+               break;
+       }
+
+       thread_unlock(thread);
+       splx(s);
+}
 
 /*
- *     task_terminate:
+ *     task_assign:
  *
- *     Terminate the specified task.  See comments on thread_terminate
- *     (kern/thread.c) about problems with terminating the "current task."
+ *     Change the assigned processor set for the task
+ */
+kern_return_t
+task_assign(
+       __unused task_t         task,
+       __unused processor_set_t        new_pset,
+       __unused boolean_t      assign_threads)
+{
+       return KERN_FAILURE;
+}
+
+/*
+ *     task_assign_default:
+ *
+ *     Version of task_assign to assign to default processor set.
  */
+kern_return_t
+task_assign_default(
+       task_t          task,
+       boolean_t       assign_threads)
+{
+       return task_assign(task, &pset0, assign_threads);
+}
 
+/*
+ *     task_get_assignment
+ *
+ *     Return name of processor set that task is assigned to.
+ */
 kern_return_t
-task_terminate(
-       task_t          task)
+task_get_assignment(
+       task_t          task,
+       processor_set_t *pset)
 {
-       if (task == TASK_NULL)
-               return (KERN_INVALID_ARGUMENT);
+       if (!task || !task->active) {
+               return KERN_FAILURE;
+       }
+
+       *pset = &pset0;
 
-       if (task->bsd_info)
-               return (KERN_FAILURE);
+       return KERN_SUCCESS;
+}
 
-       return (task_terminate_internal(task));
+uint64_t
+get_task_dispatchqueue_offset(
+       task_t          task)
+{
+       return task->dispatchqueue_offset;
 }
 
+/*
+ *      task_policy
+ *
+ *     Set scheduling policy and parameters, both base and limit, for
+ *     the given task. Policy must be a policy which is enabled for the
+ *     processor set. Change contained threads if requested.
+ */
 kern_return_t
-task_terminate_internal(
-       task_t                  task)
+task_policy(
+       __unused task_t                 task,
+       __unused policy_t                       policy_id,
+       __unused policy_base_t          base,
+       __unused mach_msg_type_number_t count,
+       __unused boolean_t                      set_limit,
+       __unused boolean_t                      change)
 {
-       thread_t                        thread, self;
-       task_t                          self_task;
-       boolean_t                       interrupt_save;
+       return KERN_FAILURE;
+}
 
-       assert(task != kernel_task);
+/*
+ *     task_set_policy
+ *
+ *     Set scheduling policy and parameters, both base and limit, for
+ *     the given task. Policy can be any policy implemented by the
+ *     processor set, whether enabled or not. Change contained threads
+ *     if requested.
+ */
+kern_return_t
+task_set_policy(
+       __unused task_t                 task,
+       __unused processor_set_t                pset,
+       __unused policy_t                       policy_id,
+       __unused policy_base_t          base,
+       __unused mach_msg_type_number_t base_count,
+       __unused policy_limit_t         limit,
+       __unused mach_msg_type_number_t limit_count,
+       __unused boolean_t                      change)
+{
+       return KERN_FAILURE;
+}
 
-       self = current_thread();
-       self_task = self->task;
+kern_return_t
+task_set_ras_pc(
+       __unused task_t task,
+       __unused vm_offset_t    pc,
+       __unused vm_offset_t    endpc)
+{
+       return KERN_FAILURE;
+}
 
+void
+task_synchronizer_destroy_all(task_t task)
+{
        /*
-        *      Get the task locked and make sure that we are not racing
-        *      with someone else trying to terminate us.
+        *  Destroy owned semaphores
         */
-       if (task == self_task)
-               task_lock(task);
-       else
-       if (task < self_task) {
-               task_lock(task);
-               task_lock(self_task);
-       }
-       else {
-               task_lock(self_task);
-               task_lock(task);
+       semaphore_destroy_all(task);
+}
+
+/*
+ * Install default (machine-dependent) initial thread state
+ * on the task.  Subsequent thread creation will have this initial
+ * state set on the thread by machine_thread_inherit_taskwide().
+ * Flavors and structures are exactly the same as those to thread_set_state()
+ */
+kern_return_t
+task_set_state(
+       task_t task,
+       int flavor,
+       thread_state_t state,
+       mach_msg_type_number_t state_count)
+{
+       kern_return_t ret;
+
+       if (task == TASK_NULL) {
+               return KERN_INVALID_ARGUMENT;
        }
 
-       if (!task->active || !self->active) {
-               /*
-                *      Task or current act is already being terminated.
-                *      Just return an error. If we are dying, this will
-                *      just get us to our AST special handler and that
-                *      will get us to finalize the termination of ourselves.
-                */
-               task_unlock(task);
-               if (self_task != task)
-                       task_unlock(self_task);
+       task_lock(task);
 
-               return (KERN_FAILURE);
+       if (!task->active) {
+               task_unlock(task);
+               return KERN_FAILURE;
        }
 
-       if (self_task != task)
-               task_unlock(self_task);
+       ret = machine_task_set_state(task, flavor, state, state_count);
 
-       /*
-        * Make sure the current thread does not get aborted out of
-        * the waits inside these operations.
-        */
-       interrupt_save = thread_interrupt_level(THREAD_UNINT);
+       task_unlock(task);
+       return ret;
+}
 
-       /*
-        *      Indicate that we want all the threads to stop executing
-        *      at user space by holding the task (we would have held
-        *      each thread independently in thread_terminate_internal -
-        *      but this way we may be more likely to already find it
-        *      held there).  Mark the task inactive, and prevent
-        *      further task operations via the task port.
-        */
-       task_hold_locked(task);
-       task->active = FALSE;
-       ipc_task_disable(task);
+/*
+ * Examine the default (machine-dependent) initial thread state
+ * on the task, as set by task_set_state().  Flavors and structures
+ * are exactly the same as those passed to thread_get_state().
+ */
+kern_return_t
+task_get_state(
+       task_t  task,
+       int     flavor,
+       thread_state_t state,
+       mach_msg_type_number_t *state_count)
+{
+       kern_return_t ret;
 
-       /*
-        *      Terminate each thread in the task.
-        */
-       queue_iterate(&task->threads, thread, thread_t, task_threads) {
-                       thread_terminate_internal(thread);
+       if (task == TASK_NULL) {
+               return KERN_INVALID_ARGUMENT;
        }
 
-       /*
-        *      Give the machine dependent code a chance
-        *      to perform cleanup before ripping apart
-        *      the task.
-        */
-       if (self_task == task)
-               machine_thread_terminate_self();
+       task_lock(task);
+
+       if (!task->active) {
+               task_unlock(task);
+               return KERN_FAILURE;
+       }
+
+       ret = machine_task_get_state(task, flavor, state, state_count);
 
        task_unlock(task);
+       return ret;
+}
 
-       /*
-        *      Destroy all synchronizers owned by the task.
-        */
-       task_synchronizer_destroy_all(task);
 
-       /*
-        *      Destroy the IPC space, leaving just a reference for it.
-        */
-       ipc_space_destroy(task->itk_space);
+static kern_return_t __attribute__((noinline, not_tail_called))
+PROC_VIOLATED_GUARD__SEND_EXC_GUARD_AND_SUSPEND(
+       mach_exception_code_t code,
+       mach_exception_subcode_t subcode,
+       void *reason)
+{
+#ifdef MACH_BSD
+       if (1 == proc_selfpid()) {
+               return KERN_NOT_SUPPORTED;              // initproc is immune
+       }
+#endif
+       mach_exception_data_type_t codes[EXCEPTION_CODE_MAX] = {
+               [0] = code,
+               [1] = subcode,
+       };
+       task_t task = current_task();
+       kern_return_t kr;
+
+       /* (See jetsam-related comments below) */
+
+       proc_memstat_terminated(task->bsd_info, TRUE);
+       kr = task_enqueue_exception_with_corpse(task, EXC_GUARD, codes, 2, reason);
+       proc_memstat_terminated(task->bsd_info, FALSE);
+       return kr;
+}
 
-#ifdef __ppc__
-       /*
-        * PPC51: ppc64 is limited to 51-bit addresses.
-        */
-       pmap_unmap_sharedpage(task->map->pmap);         /* Unmap commpage */
-#endif /* __ppc__ */
+kern_return_t
+task_violated_guard(
+       mach_exception_code_t code,
+       mach_exception_subcode_t subcode,
+       void *reason)
+{
+       return PROC_VIOLATED_GUARD__SEND_EXC_GUARD_AND_SUSPEND(code, subcode, reason);
+}
 
-       if (vm_map_has_4GB_pagezero(task->map))
-               vm_map_clear_4GB_pagezero(task->map);
 
-       /*
-        * If the current thread is a member of the task
-        * being terminated, then the last reference to
-        * the task will not be dropped until the thread
-        * is finally reaped.  To avoid incurring the
-        * expense of removing the address space regions
-        * at reap time, we do it explictly here.
-        */
-       vm_map_remove(task->map,
-                     task->map->min_offset,
-                     task->map->max_offset,
-                     VM_MAP_NO_FLAGS);
+#if CONFIG_MEMORYSTATUS
 
-       /* release our shared region */
-       vm_shared_region_set(task, NULL);
+boolean_t
+task_get_memlimit_is_active(task_t task)
+{
+       assert(task != NULL);
 
-       lck_mtx_lock(&tasks_threads_lock);
-       queue_remove(&tasks, task, task_t, tasks);
-       tasks_count--;
-       lck_mtx_unlock(&tasks_threads_lock);
+       if (task->memlimit_is_active == 1) {
+               return TRUE;
+       } else {
+               return FALSE;
+       }
+}
+
+void
+task_set_memlimit_is_active(task_t task, boolean_t memlimit_is_active)
+{
+       assert(task != NULL);
+
+       if (memlimit_is_active) {
+               task->memlimit_is_active = 1;
+       } else {
+               task->memlimit_is_active = 0;
+       }
+}
+
+boolean_t
+task_get_memlimit_is_fatal(task_t task)
+{
+       assert(task != NULL);
+
+       if (task->memlimit_is_fatal == 1) {
+               return TRUE;
+       } else {
+               return FALSE;
+       }
+}
+
+void
+task_set_memlimit_is_fatal(task_t task, boolean_t memlimit_is_fatal)
+{
+       assert(task != NULL);
+
+       if (memlimit_is_fatal) {
+               task->memlimit_is_fatal = 1;
+       } else {
+               task->memlimit_is_fatal = 0;
+       }
+}
 
-       /*
-        * We no longer need to guard against being aborted, so restore
-        * the previous interruptible state.
-        */
-       thread_interrupt_level(interrupt_save);
+boolean_t
+task_has_triggered_exc_resource(task_t task, boolean_t memlimit_is_active)
+{
+       boolean_t triggered = FALSE;
 
-#if __ppc__
-    perfmon_release_facility(task); // notify the perfmon facility
-#endif
+       assert(task == current_task());
 
        /*
-        * Get rid of the task active reference on itself.
+        * Returns true, if task has already triggered an exc_resource exception.
         */
-       task_deallocate(task);
 
-       return (KERN_SUCCESS);
+       if (memlimit_is_active) {
+               triggered = (task->memlimit_active_exc_resource ? TRUE : FALSE);
+       } else {
+               triggered = (task->memlimit_inactive_exc_resource ? TRUE : FALSE);
+       }
+
+       return triggered;
 }
 
-/*
- * task_start_halt:
- *
- *     Shut the current task down (except for the current thread) in
- *     preparation for dramatic changes to the task (probably exec).
- *     We hold the task and mark all other threads in the task for
- *     termination.
- */
-kern_return_t
-task_start_halt(
-       task_t          task)
+void
+task_mark_has_triggered_exc_resource(task_t task, boolean_t memlimit_is_active)
 {
-       thread_t        thread, self;
+       assert(task == current_task());
 
-       assert(task != kernel_task);
+       /*
+        * We allow one exc_resource per process per active/inactive limit.
+        * The limit's fatal attribute does not come into play.
+        */
 
-       self = current_thread();
+       if (memlimit_is_active) {
+               task->memlimit_active_exc_resource = 1;
+       } else {
+               task->memlimit_inactive_exc_resource = 1;
+       }
+}
 
-       if (task != self->task)
-               return (KERN_INVALID_ARGUMENT);
+#define HWM_USERCORE_MINSPACE 250 // free space (in MB) required *after* core file creation
 
-       task_lock(task);
+void __attribute__((noinline))
+PROC_CROSSED_HIGH_WATERMARK__SEND_EXC_RESOURCE_AND_SUSPEND(int max_footprint_mb, boolean_t is_fatal)
+{
+       task_t                                          task            = current_task();
+       int                                                     pid         = 0;
+       const char                                      *procname       = "unknown";
+       mach_exception_data_type_t      code[EXCEPTION_CODE_MAX];
+       boolean_t send_sync_exc_resource = FALSE;
 
-       if (task->halting || !task->active || !self->active) {
+#ifdef MACH_BSD
+       pid = proc_selfpid();
+
+       if (pid == 1) {
                /*
-                *      Task or current thread is already being terminated.
-                *      Hurry up and return out of the current kernel context
-                *      so that we run our AST special handler to terminate
-                *      ourselves.
+                * Cannot have ReportCrash analyzing
+                * a suspended initproc.
                 */
-               task_unlock(task);
-
-               return (KERN_FAILURE);
+               return;
        }
 
-       task->halting = TRUE;
-
-       if (task->thread_count > 1) {
-
+       if (task->bsd_info != NULL) {
+               procname = proc_name_address(current_task()->bsd_info);
+               send_sync_exc_resource = proc_send_synchronous_EXC_RESOURCE(current_task()->bsd_info);
+       }
+#endif
+#if CONFIG_COREDUMP
+       if (hwm_user_cores) {
+               int                             error;
+               uint64_t                starttime, end;
+               clock_sec_t             secs = 0;
+               uint32_t                microsecs = 0;
+
+               starttime = mach_absolute_time();
                /*
-                * Mark all the threads to keep them from starting any more
-                * user-level execution.  The thread_terminate_internal code
-                * would do this on a thread by thread basis anyway, but this
-                * gives us a better chance of not having to wait there.
+                * Trigger a coredump of this process. Don't proceed unless we know we won't
+                * be filling up the disk; and ignore the core size resource limit for this
+                * core file.
                 */
-               task_hold_locked(task);
-
+               if ((error = coredump(current_task()->bsd_info, HWM_USERCORE_MINSPACE, COREDUMP_IGNORE_ULIMIT)) != 0) {
+                       printf("couldn't take coredump of %s[%d]: %d\n", procname, pid, error);
+               }
                /*
-                *      Terminate all the other threads in the task.
+                * coredump() leaves the task suspended.
                 */
-               queue_iterate(&task->threads, thread, thread_t, task_threads) {
-                       if (thread != self)
-                               thread_terminate_internal(thread);
-               }
+               task_resume_internal(current_task());
 
-               task_release_locked(task);
+               end = mach_absolute_time();
+               absolutetime_to_microtime(end - starttime, &secs, &microsecs);
+               printf("coredump of %s[%d] taken in %d secs %d microsecs\n",
+                   proc_name_address(current_task()->bsd_info), pid, (int)secs, microsecs);
        }
-       task_unlock(task);
-       return KERN_SUCCESS;
-}
+#endif /* CONFIG_COREDUMP */
 
-
-/*
- * task_complete_halt:
- *
- *     Complete task halt by waiting for threads to terminate, then clean
- *     up task resources (VM, port namespace, etc...) and then let the
- *     current thread go in the (practically empty) task context.
- */
-void
-task_complete_halt(task_t task)
-{
-       task_lock(task);
-       assert(task->halting);
-       assert(task == current_task());
+       if (disable_exc_resource) {
+               printf("process %s[%d] crossed memory high watermark (%d MB); EXC_RESOURCE "
+                   "supressed by a boot-arg.\n", procname, pid, max_footprint_mb);
+               return;
+       }
 
        /*
-        *      Give the machine dependent code a chance
-        *      to perform cleanup of task-level resources
-        *      associated with the current thread before
-        *      ripping apart the task.
-        *
-        *      This must be done with the task locked.
+        * A task that has triggered an EXC_RESOURCE, should not be
+        * jetsammed when the device is under memory pressure.  Here
+        * we set the P_MEMSTAT_TERMINATED flag so that the process
+        * will be skipped if the memorystatus_thread wakes up.
         */
-       machine_thread_terminate_self();
+       proc_memstat_terminated(current_task()->bsd_info, TRUE);
+
+       code[0] = code[1] = 0;
+       EXC_RESOURCE_ENCODE_TYPE(code[0], RESOURCE_TYPE_MEMORY);
+       EXC_RESOURCE_ENCODE_FLAVOR(code[0], FLAVOR_HIGH_WATERMARK);
+       EXC_RESOURCE_HWM_ENCODE_LIMIT(code[0], max_footprint_mb);
 
        /*
-        *      Wait for the other threads to get shut down.
-        *      When the last other thread is reaped, we'll be
-        *      worken up.
+        * Do not generate a corpse fork if the violation is a fatal one
+        * or the process wants synchronous EXC_RESOURCE exceptions.
         */
-       if (task->thread_count > 1) {
-               assert_wait((event_t)&task->halting, THREAD_UNINT);
-               task_unlock(task);
-               thread_block(THREAD_CONTINUE_NULL);
+       if (is_fatal || send_sync_exc_resource || exc_via_corpse_forking == 0) {
+               /* Do not send a EXC_RESOURCE if corpse_for_fatal_memkill is set */
+               if (send_sync_exc_resource || corpse_for_fatal_memkill == 0) {
+                       /*
+                        * Use the _internal_ variant so that no user-space
+                        * process can resume our task from under us.
+                        */
+                       task_suspend_internal(task);
+                       exception_triage(EXC_RESOURCE, code, EXCEPTION_CODE_MAX);
+                       task_resume_internal(task);
+               }
        } else {
-               task_unlock(task);
+               if (audio_active) {
+                       printf("process %s[%d] crossed memory high watermark (%d MB); EXC_RESOURCE "
+                           "supressed due to audio playback.\n", procname, pid, max_footprint_mb);
+               } else {
+                       task_enqueue_exception_with_corpse(task, EXC_RESOURCE,
+                           code, EXCEPTION_CODE_MAX, NULL);
+               }
        }
 
        /*
-        *      Destroy all synchronizers owned by the task.
-        */
-       task_synchronizer_destroy_all(task);
-
-       /*
-        *      Destroy the contents of the IPC space, leaving just
-        *      a reference for it.
-        */
-       ipc_space_clean(task->itk_space);
-
-       /*
-        * Clean out the address space, as we are going to be
-        * getting a new one.
+        * After the EXC_RESOURCE has been handled, we must clear the
+        * P_MEMSTAT_TERMINATED flag so that the process can again be
+        * considered for jetsam if the memorystatus_thread wakes up.
         */
-       vm_map_remove(task->map, task->map->min_offset,
-                     task->map->max_offset, VM_MAP_NO_FLAGS);
-
-       task->halting = FALSE;
+       proc_memstat_terminated(current_task()->bsd_info, FALSE);  /* clear the flag */
 }
 
 /*
- *     task_hold_locked:
- *
- *     Suspend execution of the specified task.
- *     This is a recursive-style suspension of the task, a count of
- *     suspends is maintained.
- *
- *     CONDITIONS: the task is locked and active.
+ * Callback invoked when a task exceeds its physical footprint limit.
  */
 void
-task_hold_locked(
-       register task_t         task)
+task_footprint_exceeded(int warning, __unused const void *param0, __unused const void *param1)
 {
-       register thread_t       thread;
+       ledger_amount_t max_footprint, max_footprint_mb;
+       task_t task;
+       boolean_t is_warning;
+       boolean_t memlimit_is_active;
+       boolean_t memlimit_is_fatal;
 
-       assert(task->active);
-
-       if (task->suspend_count++ > 0)
+       if (warning == LEDGER_WARNING_DIPPED_BELOW) {
+               /*
+                * Task memory limits only provide a warning on the way up.
+                */
                return;
+       } else if (warning == LEDGER_WARNING_ROSE_ABOVE) {
+               /*
+                * This task is in danger of violating a memory limit,
+                * It has exceeded a percentage level of the limit.
+                */
+               is_warning = TRUE;
+       } else {
+               /*
+                * The task has exceeded the physical footprint limit.
+                * This is not a warning but a true limit violation.
+                */
+               is_warning = FALSE;
+       }
+
+       task = current_task();
+
+       ledger_get_limit(task->ledger, task_ledgers.phys_footprint, &max_footprint);
+       max_footprint_mb = max_footprint >> 20;
+
+       memlimit_is_active = task_get_memlimit_is_active(task);
+       memlimit_is_fatal = task_get_memlimit_is_fatal(task);
 
        /*
-        *      Iterate through all the threads and hold them.
+        * If this is an actual violation (not a warning), then generate EXC_RESOURCE exception.
+        * We only generate the exception once per process per memlimit (active/inactive limit).
+        * To enforce this, we monitor state based on the  memlimit's active/inactive attribute
+        * and we disable it by marking that memlimit as exception triggered.
         */
-       queue_iterate(&task->threads, thread, thread_t, task_threads) {
-               thread_mtx_lock(thread);
-               thread_hold(thread);
-               thread_mtx_unlock(thread);
+       if ((is_warning == FALSE) && (!task_has_triggered_exc_resource(task, memlimit_is_active))) {
+               PROC_CROSSED_HIGH_WATERMARK__SEND_EXC_RESOURCE_AND_SUSPEND((int)max_footprint_mb, memlimit_is_fatal);
+               memorystatus_log_exception((int)max_footprint_mb, memlimit_is_active, memlimit_is_fatal);
+               task_mark_has_triggered_exc_resource(task, memlimit_is_active);
        }
+
+       memorystatus_on_ledger_footprint_exceeded(is_warning, memlimit_is_active, memlimit_is_fatal);
 }
 
-/*
- *     task_hold:
- *
- *     Same as the internal routine above, except that is must lock
- *     and verify that the task is active.  This differs from task_suspend
- *     in that it places a kernel hold on the task rather than just a 
- *     user-level hold.  This keeps users from over resuming and setting
- *     it running out from under the kernel.
- *
- *     CONDITIONS: the caller holds a reference on the task
- */
+extern int proc_check_footprint_priv(void);
+
 kern_return_t
-task_hold(
-       register task_t         task)
+task_set_phys_footprint_limit(
+       task_t task,
+       int new_limit_mb,
+       int *old_limit_mb)
 {
-       if (task == TASK_NULL)
-               return (KERN_INVALID_ARGUMENT);
-
-       task_lock(task);
+       kern_return_t error;
 
-       if (!task->active) {
-               task_unlock(task);
+       boolean_t memlimit_is_active;
+       boolean_t memlimit_is_fatal;
 
-               return (KERN_FAILURE);
+       if ((error = proc_check_footprint_priv())) {
+               return KERN_NO_ACCESS;
        }
 
-       task_hold_locked(task);
-       task_unlock(task);
+       /*
+        * This call should probably be obsoleted.
+        * But for now, we default to current state.
+        */
+       memlimit_is_active = task_get_memlimit_is_active(task);
+       memlimit_is_fatal = task_get_memlimit_is_fatal(task);
 
-       return (KERN_SUCCESS);
+       return task_set_phys_footprint_limit_internal(task, new_limit_mb, old_limit_mb, memlimit_is_active, memlimit_is_fatal);
 }
 
-/*
- *     task_wait_locked:
- *
- *     Wait for all threads in task to stop.
- *
- * Conditions:
- *     Called with task locked, active, and held.
- */
-void
-task_wait_locked(
-       register task_t         task)
+kern_return_t
+task_convert_phys_footprint_limit(
+       int limit_mb,
+       int *converted_limit_mb)
+{
+       if (limit_mb == -1) {
+               /*
+                * No limit
+                */
+               if (max_task_footprint != 0) {
+                       *converted_limit_mb = (int)(max_task_footprint / 1024 / 1024);   /* bytes to MB */
+               } else {
+                       *converted_limit_mb = (int)(LEDGER_LIMIT_INFINITY >> 20);
+               }
+       } else {
+               /* nothing to convert */
+               *converted_limit_mb = limit_mb;
+       }
+       return KERN_SUCCESS;
+}
+
+
+kern_return_t
+task_set_phys_footprint_limit_internal(
+       task_t task,
+       int new_limit_mb,
+       int *old_limit_mb,
+       boolean_t memlimit_is_active,
+       boolean_t memlimit_is_fatal)
 {
-       register thread_t       thread, self;
+       ledger_amount_t old;
+       kern_return_t ret;
 
-       assert(task->active);
-       assert(task->suspend_count > 0);
+       ret = ledger_get_limit(task->ledger, task_ledgers.phys_footprint, &old);
 
-       self = current_thread();
+       if (ret != KERN_SUCCESS) {
+               return ret;
+       }
 
        /*
-        *      Iterate through all the threads and wait for them to
-        *      stop.  Do not wait for the current thread if it is within
-        *      the task.
+        * Check that limit >> 20 will not give an "unexpected" 32-bit
+        * result. There are, however, implicit assumptions that -1 mb limit
+        * equates to LEDGER_LIMIT_INFINITY.
         */
-       queue_iterate(&task->threads, thread, thread_t, task_threads) {
-               if (thread != self)
-                       thread_wait(thread);
-       }
-}
+       assert(((old & 0xFFF0000000000000LL) == 0) || (old == LEDGER_LIMIT_INFINITY));
 
-/*
- *     task_release_locked:
- *
- *     Release a kernel hold on a task.
- *
- *     CONDITIONS: the task is locked and active
- */
-void
-task_release_locked(
-       register task_t         task)
-{
-       register thread_t       thread;
+       if (old_limit_mb) {
+               *old_limit_mb = (int)(old >> 20);
+       }
 
-       assert(task->active);
-       assert(task->suspend_count > 0);
+       if (new_limit_mb == -1) {
+               /*
+                * Caller wishes to remove the limit.
+                */
+               ledger_set_limit(task->ledger, task_ledgers.phys_footprint,
+                   max_task_footprint ? max_task_footprint : LEDGER_LIMIT_INFINITY,
+                   max_task_footprint ? (uint8_t)max_task_footprint_warning_level : 0);
 
-       if (--task->suspend_count > 0)
-               return;
+               task_lock(task);
+               task_set_memlimit_is_active(task, memlimit_is_active);
+               task_set_memlimit_is_fatal(task, memlimit_is_fatal);
+               task_unlock(task);
 
-       queue_iterate(&task->threads, thread, thread_t, task_threads) {
-               thread_mtx_lock(thread);
-               thread_release(thread);
-               thread_mtx_unlock(thread);
+               return KERN_SUCCESS;
        }
-}
 
-/*
- *     task_release:
- *
- *     Same as the internal routine above, except that it must lock
- *     and verify that the task is active.
- *
- *     CONDITIONS: The caller holds a reference to the task
- */
-kern_return_t
-task_release(
-       task_t          task)
-{
-       if (task == TASK_NULL)
-               return (KERN_INVALID_ARGUMENT);
+#ifdef CONFIG_NOMONITORS
+       return KERN_SUCCESS;
+#endif /* CONFIG_NOMONITORS */
 
        task_lock(task);
 
-       if (!task->active) {
+       if ((memlimit_is_active == task_get_memlimit_is_active(task)) &&
+           (memlimit_is_fatal == task_get_memlimit_is_fatal(task)) &&
+           (((ledger_amount_t)new_limit_mb << 20) == old)) {
+               /*
+                * memlimit state is not changing
+                */
                task_unlock(task);
+               return KERN_SUCCESS;
+       }
+
+       task_set_memlimit_is_active(task, memlimit_is_active);
+       task_set_memlimit_is_fatal(task, memlimit_is_fatal);
 
-               return (KERN_FAILURE);
+       ledger_set_limit(task->ledger, task_ledgers.phys_footprint,
+           (ledger_amount_t)new_limit_mb << 20, PHYS_FOOTPRINT_WARNING_LEVEL);
+
+       if (task == current_task()) {
+               ledger_check_new_balance(current_thread(), task->ledger,
+                   task_ledgers.phys_footprint);
        }
 
-       task_release_locked(task);
        task_unlock(task);
 
-       return (KERN_SUCCESS);
+       return KERN_SUCCESS;
 }
 
 kern_return_t
-task_threads(
-       task_t                                  task,
-       thread_act_array_t              *threads_out,
-       mach_msg_type_number_t  *count)
+task_get_phys_footprint_limit(
+       task_t task,
+       int *limit_mb)
 {
-       mach_msg_type_number_t  actual;
-       thread_t                                *thread_list;
-       thread_t                                thread;
-       vm_size_t                               size, size_needed;
-       void                                    *addr;
-       unsigned int                    i, j;
-
-       if (task == TASK_NULL)
-               return (KERN_INVALID_ARGUMENT);
-
-       size = 0; addr = NULL;
-
-       for (;;) {
-               task_lock(task);
-               if (!task->active) {
-                       task_unlock(task);
-
-                       if (size != 0)
-                               kfree(addr, size);
-
-                       return (KERN_FAILURE);
-               }
-
-               actual = task->thread_count;
-
-               /* do we have the memory we need? */
-               size_needed = actual * sizeof (mach_port_t);
-               if (size_needed <= size)
-                       break;
-
-               /* unlock the task and allocate more memory */
-               task_unlock(task);
-
-               if (size != 0)
-                       kfree(addr, size);
-
-               assert(size_needed > 0);
-               size = size_needed;
+       ledger_amount_t limit;
+       kern_return_t ret;
 
-               addr = kalloc(size);
-               if (addr == 0)
-                       return (KERN_RESOURCE_SHORTAGE);
+       ret = ledger_get_limit(task->ledger, task_ledgers.phys_footprint, &limit);
+       if (ret != KERN_SUCCESS) {
+               return ret;
        }
 
-       /* OK, have memory and the task is locked & active */
-       thread_list = (thread_t *)addr;
+       /*
+        * Check that limit >> 20 will not give an "unexpected" signed, 32-bit
+        * result. There are, however, implicit assumptions that -1 mb limit
+        * equates to LEDGER_LIMIT_INFINITY.
+        */
+       assert(((limit & 0xFFF0000000000000LL) == 0) || (limit == LEDGER_LIMIT_INFINITY));
+       *limit_mb = (int)(limit >> 20);
+
+       return KERN_SUCCESS;
+}
+#else /* CONFIG_MEMORYSTATUS */
+kern_return_t
+task_set_phys_footprint_limit(
+       __unused task_t task,
+       __unused int new_limit_mb,
+       __unused int *old_limit_mb)
+{
+       return KERN_FAILURE;
+}
 
-       i = j = 0;
+kern_return_t
+task_get_phys_footprint_limit(
+       __unused task_t task,
+       __unused int *limit_mb)
+{
+       return KERN_FAILURE;
+}
+#endif /* CONFIG_MEMORYSTATUS */
 
-       for (thread = (thread_t)queue_first(&task->threads); i < actual;
-                               ++i, thread = (thread_t)queue_next(&thread->task_threads)) {
-               thread_reference_internal(thread);
-               thread_list[j++] = thread;
+void
+task_set_thread_limit(task_t task, uint16_t thread_limit)
+{
+       assert(task != kernel_task);
+       if (thread_limit <= TASK_MAX_THREAD_LIMIT) {
+               task_lock(task);
+               task->task_thread_limit = thread_limit;
+               task_unlock(task);
        }
+}
 
-       assert(queue_end(&task->threads, (queue_entry_t)thread));
-
-       actual = j;
-       size_needed = actual * sizeof (mach_port_t);
+#if XNU_TARGET_OS_OSX
+boolean_t
+task_has_system_version_compat_enabled(task_t task)
+{
+       boolean_t enabled = FALSE;
 
-       /* can unlock task now that we've got the thread refs */
+       task_lock(task);
+       enabled = (task->t_flags & TF_SYS_VERSION_COMPAT);
        task_unlock(task);
 
-       if (actual == 0) {
-               /* no threads, so return null pointer and deallocate memory */
+       return enabled;
+}
 
-               *threads_out = NULL;
-               *count = 0;
+void
+task_set_system_version_compat_enabled(task_t task, boolean_t enable_system_version_compat)
+{
+       assert(task == current_task());
+       assert(task != kernel_task);
 
-               if (size != 0)
-                       kfree(addr, size);
+       task_lock(task);
+       if (enable_system_version_compat) {
+               task->t_flags |= TF_SYS_VERSION_COMPAT;
+       } else {
+               task->t_flags &= ~TF_SYS_VERSION_COMPAT;
        }
-       else {
-               /* if we allocated too much, must copy */
-
-               if (size_needed < size) {
-                       void *newaddr;
+       task_unlock(task);
+}
+#endif /* XNU_TARGET_OS_OSX */
 
-                       newaddr = kalloc(size_needed);
-                       if (newaddr == 0) {
-                               for (i = 0; i < actual; ++i)
-                                       thread_deallocate(thread_list[i]);
-                               kfree(addr, size);
-                               return (KERN_RESOURCE_SHORTAGE);
-                       }
+/*
+ * We need to export some functions to other components that
+ * are currently implemented in macros within the osfmk
+ * component.  Just export them as functions of the same name.
+ */
+boolean_t
+is_kerneltask(task_t t)
+{
+       if (t == kernel_task) {
+               return TRUE;
+       }
 
-                       bcopy(addr, newaddr, size_needed);
-                       kfree(addr, size);
-                       thread_list = (thread_t *)newaddr;
-               }
+       return FALSE;
+}
 
-               *threads_out = thread_list;
-               *count = actual;
+boolean_t
+is_corpsetask(task_t t)
+{
+       return task_is_a_corpse(t);
+}
 
-               /* do the conversion that Mig should handle */
+#undef current_task
+task_t current_task(void);
+task_t
+current_task(void)
+{
+       return current_task_fast();
+}
 
-               for (i = 0; i < actual; ++i)
-                       ((ipc_port_t *) thread_list)[i] = convert_thread_to_port(thread_list[i]);
+#undef task_reference
+void task_reference(task_t task);
+void
+task_reference(
+       task_t          task)
+{
+       if (task != TASK_NULL) {
+               task_reference_internal(task);
        }
+}
+
+/* defined in bsd/kern/kern_prot.c */
+extern int get_audit_token_pid(audit_token_t *audit_token);
 
-       return (KERN_SUCCESS);
+int
+task_pid(task_t task)
+{
+       if (task) {
+               return get_audit_token_pid(&task->audit_token);
+       }
+       return -1;
 }
 
+#if __has_feature(ptrauth_calls)
 /*
- *     task_suspend:
- *
- *     Implement a user-level suspension on a task.
- *
- * Conditions:
- *     The caller holds a reference to the task
+ * Get the shared region id and jop signing key for the task.
+ * The function will allocate a kalloc buffer and return
+ * it to caller, the caller needs to free it. This is used
+ * for getting the information via task port.
  */
-kern_return_t
-task_suspend(
-       register task_t         task)
+char *
+task_get_vm_shared_region_id_and_jop_pid(task_t task, uint64_t *jop_pid)
 {
-       if (task == TASK_NULL || task == kernel_task)
-               return (KERN_INVALID_ARGUMENT);
+       size_t len;
+       char *shared_region_id = NULL;
 
        task_lock(task);
-
-       if (!task->active) {
+       if (task->shared_region_id == NULL) {
                task_unlock(task);
-
-               return (KERN_FAILURE);
+               return NULL;
        }
+       len = strlen(task->shared_region_id) + 1;
 
-       if (task->user_stop_count++ > 0) {
-               /*
-                *      If the stop count was positive, the task is
-                *      already stopped and we can exit.
-                */
+       /* don't hold task lock while allocating */
+       task_unlock(task);
+       shared_region_id = kheap_alloc(KHEAP_DATA_BUFFERS, len, Z_WAITOK);
+       task_lock(task);
+
+       if (task->shared_region_id == NULL) {
                task_unlock(task);
+               kheap_free(KHEAP_DATA_BUFFERS, shared_region_id, len);
+               return NULL;
+       }
+       assert(len == strlen(task->shared_region_id) + 1);      /* should never change */
+       strlcpy(shared_region_id, task->shared_region_id, len);
+       task_unlock(task);
 
-               return (KERN_SUCCESS);
+       /* find key from its auth pager */
+       if (jop_pid != NULL) {
+               *jop_pid = shared_region_find_key(shared_region_id);
        }
 
-       /*
-        * Put a kernel-level hold on the threads in the task (all
-        * user-level task suspensions added together represent a
-        * single kernel-level hold).  We then wait for the threads
-        * to stop executing user code.
-        */
-       task_hold_locked(task);
-       task_wait_locked(task);
+       return shared_region_id;
+}
+
+/*
+ * set the shared region id for a task
+ */
+void
+task_set_shared_region_id(task_t task, char *id)
+{
+       char *old_id;
 
+       task_lock(task);
+       old_id = task->shared_region_id;
+       task->shared_region_id = id;
+       task->shared_region_auth_remapped = FALSE;
        task_unlock(task);
 
-       return (KERN_SUCCESS);
+       /* free any pre-existing shared region id */
+       if (old_id != NULL) {
+               shared_region_key_dealloc(old_id);
+               kheap_free(KHEAP_DATA_BUFFERS, old_id, strlen(old_id) + 1);
+       }
 }
+#endif /* __has_feature(ptrauth_calls) */
 
 /*
- *     task_resume:
- *             Release a kernel hold on a task.
- *             
- * Conditions:
- *             The caller holds a reference to the task
+ * This routine finds a thread in a task by its unique id
+ * Returns a referenced thread or THREAD_NULL if the thread was not found
+ *
+ * TODO: This is super inefficient - it's an O(threads in task) list walk!
+ *       We should make a tid hash, or transition all tid clients to thread ports
+ *
+ * Precondition: No locks held (will take task lock)
  */
-kern_return_t 
-task_resume(
-       register task_t task)
+thread_t
+task_findtid(task_t task, uint64_t tid)
 {
-       register boolean_t      release = FALSE;
-
-       if (task == TASK_NULL || task == kernel_task)
-               return (KERN_INVALID_ARGUMENT);
+       thread_t self           = current_thread();
+       thread_t found_thread   = THREAD_NULL;
+       thread_t iter_thread    = THREAD_NULL;
 
-       task_lock(task);
+       /* Short-circuit the lookup if we're looking up ourselves */
+       if (tid == self->thread_id || tid == TID_NULL) {
+               assert(self->task == task);
 
-       if (!task->active) {
-               task_unlock(task);
+               thread_reference(self);
 
-               return (KERN_FAILURE);
+               return self;
        }
 
-       if (task->user_stop_count > 0) {
-               if (--task->user_stop_count == 0)
-                       release = TRUE;
-       }
-       else {
-               task_unlock(task);
+       task_lock(task);
 
-               return (KERN_FAILURE);
+       queue_iterate(&task->threads, iter_thread, thread_t, task_threads) {
+               if (iter_thread->thread_id == tid) {
+                       found_thread = iter_thread;
+                       thread_reference(found_thread);
+                       break;
+               }
        }
 
-       /*
-        *      Release the task if necessary.
-        */
-       if (release)
-               task_release_locked(task);
-
        task_unlock(task);
 
-       return (KERN_SUCCESS);
+       return found_thread;
 }
 
-kern_return_t
-host_security_set_task_token(
-        host_security_t  host_security,
-        task_t          task,
-        security_token_t sec_token,
-       audit_token_t    audit_token,
-       host_priv_t      host_priv)
+int
+pid_from_task(task_t task)
 {
-       ipc_port_t       host_port;
-       kern_return_t    kr;
+       int pid = -1;
 
-       if (task == TASK_NULL)
-               return(KERN_INVALID_ARGUMENT);
-
-       if (host_security == HOST_NULL)
-               return(KERN_INVALID_SECURITY);
-
-        task_lock(task);
-        task->sec_token = sec_token;
-       task->audit_token = audit_token;
-        task_unlock(task);
-
-       if (host_priv != HOST_PRIV_NULL) {
-               kr = host_get_host_priv_port(host_priv, &host_port);
+       if (task->bsd_info) {
+               pid = proc_pid(task->bsd_info);
        } else {
-               kr = host_get_host_port(host_priv_self(), &host_port);
+               pid = task_pid(task);
        }
-       assert(kr == KERN_SUCCESS);
-       kr = task_set_special_port(task, TASK_HOST_PORT, host_port);
-        return(kr);
+
+       return pid;
 }
 
 /*
- * Utility routine to set a ledger
+ * Control the CPU usage monitor for a task.
  */
 kern_return_t
-task_set_ledger(
-        task_t         task,
-        ledger_t       wired,
-        ledger_t       paged)
+task_cpu_usage_monitor_ctl(task_t task, uint32_t *flags)
 {
-       if (task == TASK_NULL)
-               return(KERN_INVALID_ARGUMENT);
+       int error = KERN_SUCCESS;
 
-        task_lock(task);
-        if (wired) {
-                ipc_port_release_send(task->wired_ledger_port);
-                task->wired_ledger_port = ledger_copy(wired);
-        }                
-        if (paged) {
-                ipc_port_release_send(task->paged_ledger_port);
-                task->paged_ledger_port = ledger_copy(paged);
-        }                
-        task_unlock(task);
+       if (*flags & CPUMON_MAKE_FATAL) {
+               task->rusage_cpu_flags |= TASK_RUSECPU_FLAGS_FATAL_CPUMON;
+       } else {
+               error = KERN_INVALID_ARGUMENT;
+       }
 
-        return(KERN_SUCCESS);
+       return error;
 }
 
 /*
- * This routine was added, pretty much exclusively, for registering the
- * RPC glue vector for in-kernel short circuited tasks.  Rather than
- * removing it completely, I have only disabled that feature (which was
- * the only feature at the time).  It just appears that we are going to
- * want to add some user data to tasks in the future (i.e. bsd info,
- * task names, etc...), so I left it in the formal task interface.
+ * Control the wakeups monitor for a task.
  */
 kern_return_t
-task_set_info(
-       task_t          task,
-       task_flavor_t   flavor,
-       __unused task_info_t    task_info_in,           /* pointer to IN array */
-       __unused mach_msg_type_number_t task_info_count)
+task_wakeups_monitor_ctl(task_t task, uint32_t *flags, int32_t *rate_hz)
 {
-       if (task == TASK_NULL)
-               return(KERN_INVALID_ARGUMENT);
-
-       switch (flavor) {
-           default:
-               return (KERN_INVALID_ARGUMENT);
-       }
-       return (KERN_SUCCESS);
-}
+       ledger_t ledger = task->ledger;
 
-kern_return_t
-task_info(
-       task_t                                  task,
-       task_flavor_t                   flavor,
-       task_info_t                             task_info_out,
-       mach_msg_type_number_t  *task_info_count)
-{
-       kern_return_t error = KERN_SUCCESS;
+       task_lock(task);
+       if (*flags & WAKEMON_GET_PARAMS) {
+               ledger_amount_t limit;
+               uint64_t                period;
 
-       if (task == TASK_NULL)
-               return (KERN_INVALID_ARGUMENT);
+               ledger_get_limit(ledger, task_ledgers.interrupt_wakeups, &limit);
+               ledger_get_period(ledger, task_ledgers.interrupt_wakeups, &period);
 
-       task_lock(task);
+               if (limit != LEDGER_LIMIT_INFINITY) {
+                       /*
+                        * An active limit means the wakeups monitor is enabled.
+                        */
+                       *rate_hz = (int32_t)(limit / (int64_t)(period / NSEC_PER_SEC));
+                       *flags = WAKEMON_ENABLE;
+                       if (task->rusage_cpu_flags & TASK_RUSECPU_FLAGS_FATAL_WAKEUPSMON) {
+                               *flags |= WAKEMON_MAKE_FATAL;
+                       }
+               } else {
+                       *flags = WAKEMON_DISABLE;
+                       *rate_hz = -1;
+               }
 
-       if ((task != current_task()) && (!task->active)) {
+               /*
+                * If WAKEMON_GET_PARAMS is present in flags, all other flags are ignored.
+                */
                task_unlock(task);
-               return (KERN_INVALID_ARGUMENT);
+               return KERN_SUCCESS;
        }
 
-       switch (flavor) {
-
-       case TASK_BASIC_INFO_32:
-       case TASK_BASIC2_INFO_32:
-       {
-               task_basic_info_32_t    basic_info;
-               vm_map_t                                map;
-               clock_sec_t                             secs;
-               clock_usec_t                    usecs;
-
-               if (*task_info_count < TASK_BASIC_INFO_32_COUNT) {
-                   error = KERN_INVALID_ARGUMENT;
-                   break;
+       if (*flags & WAKEMON_ENABLE) {
+               if (*flags & WAKEMON_SET_DEFAULTS) {
+                       *rate_hz = task_wakeups_monitor_rate;
                }
 
-               basic_info = (task_basic_info_32_t)task_info_out;
-
-               map = (task == kernel_task)? kernel_map: task->map;
-               basic_info->virtual_size = (typeof(basic_info->virtual_size))map->size;
-               if (flavor == TASK_BASIC2_INFO_32) {
-                       /*
-                        * The "BASIC2" flavor gets the maximum resident
-                        * size instead of the current resident size...
-                        */
-                       basic_info->resident_size = pmap_resident_max(map->pmap);
-               } else {
-                       basic_info->resident_size = pmap_resident_count(map->pmap);
+#ifndef CONFIG_NOMONITORS
+               if (*flags & WAKEMON_MAKE_FATAL) {
+                       task->rusage_cpu_flags |= TASK_RUSECPU_FLAGS_FATAL_WAKEUPSMON;
                }
-               basic_info->resident_size *= PAGE_SIZE;
+#endif /* CONFIG_NOMONITORS */
 
-               basic_info->policy = ((task != kernel_task)?
-                                                                                 POLICY_TIMESHARE: POLICY_RR);
-               basic_info->suspend_count = task->user_stop_count;
+               if (*rate_hz <= 0) {
+                       task_unlock(task);
+                       return KERN_INVALID_ARGUMENT;
+               }
 
-               absolutetime_to_microtime(task->total_user_time, &secs, &usecs);
-               basic_info->user_time.seconds = 
-                       (typeof(basic_info->user_time.seconds))secs;
-               basic_info->user_time.microseconds = usecs;
+#ifndef CONFIG_NOMONITORS
+               ledger_set_limit(ledger, task_ledgers.interrupt_wakeups, *rate_hz * task_wakeups_monitor_interval,
+                   (uint8_t)task_wakeups_monitor_ustackshots_trigger_pct);
+               ledger_set_period(ledger, task_ledgers.interrupt_wakeups, task_wakeups_monitor_interval * NSEC_PER_SEC);
+               ledger_enable_callback(ledger, task_ledgers.interrupt_wakeups);
+#endif /* CONFIG_NOMONITORS */
+       } else if (*flags & WAKEMON_DISABLE) {
+               /*
+                * Caller wishes to disable wakeups monitor on the task.
+                *
+                * Disable telemetry if it was triggered by the wakeups monitor, and
+                * remove the limit & callback on the wakeups ledger entry.
+                */
+#if CONFIG_TELEMETRY
+               telemetry_task_ctl_locked(task, TF_WAKEMON_WARNING, 0);
+#endif
+               ledger_disable_refill(ledger, task_ledgers.interrupt_wakeups);
+               ledger_disable_callback(ledger, task_ledgers.interrupt_wakeups);
+       }
 
-               absolutetime_to_microtime(task->total_system_time, &secs, &usecs);
-               basic_info->system_time.seconds = 
-                       (typeof(basic_info->system_time.seconds))secs;
-               basic_info->system_time.microseconds = usecs;
+       task_unlock(task);
+       return KERN_SUCCESS;
+}
 
-               *task_info_count = TASK_BASIC_INFO_32_COUNT;
-               break;
+void
+task_wakeups_rate_exceeded(int warning, __unused const void *param0, __unused const void *param1)
+{
+       if (warning == LEDGER_WARNING_ROSE_ABOVE) {
+#if CONFIG_TELEMETRY
+               /*
+                * This task is in danger of violating the wakeups monitor. Enable telemetry on this task
+                * so there are micro-stackshots available if and when EXC_RESOURCE is triggered.
+                */
+               telemetry_task_ctl(current_task(), TF_WAKEMON_WARNING, 1);
+#endif
+               return;
        }
 
-       case TASK_BASIC_INFO_64:
-       {
-               task_basic_info_64_t    basic_info;
-               vm_map_t                                map;
-               clock_sec_t                             secs;
-               clock_usec_t                    usecs;
-
-               if (*task_info_count < TASK_BASIC_INFO_64_COUNT) {
-                   error = KERN_INVALID_ARGUMENT;
-                   break;
-               }
+#if CONFIG_TELEMETRY
+       /*
+        * If the balance has dipped below the warning level (LEDGER_WARNING_DIPPED_BELOW) or
+        * exceeded the limit, turn telemetry off for the task.
+        */
+       telemetry_task_ctl(current_task(), TF_WAKEMON_WARNING, 0);
+#endif
 
-               basic_info = (task_basic_info_64_t)task_info_out;
+       if (warning == 0) {
+               SENDING_NOTIFICATION__THIS_PROCESS_IS_CAUSING_TOO_MANY_WAKEUPS();
+       }
+}
 
-               map = (task == kernel_task)? kernel_map: task->map;
-               basic_info->virtual_size  = map->size;
-               basic_info->resident_size =
-                       (mach_vm_size_t)(pmap_resident_count(map->pmap))
-                       * PAGE_SIZE_64;
+void __attribute__((noinline))
+SENDING_NOTIFICATION__THIS_PROCESS_IS_CAUSING_TOO_MANY_WAKEUPS(void)
+{
+       task_t                      task        = current_task();
+       int                         pid         = 0;
+       const char                  *procname   = "unknown";
+       boolean_t                   fatal;
+       kern_return_t               kr;
+#ifdef EXC_RESOURCE_MONITORS
+       mach_exception_data_type_t  code[EXCEPTION_CODE_MAX];
+#endif /* EXC_RESOURCE_MONITORS */
+       struct ledger_entry_info    lei;
 
-               basic_info->policy = ((task != kernel_task)?
-                                                                                 POLICY_TIMESHARE: POLICY_RR);
-               basic_info->suspend_count = task->user_stop_count;
+#ifdef MACH_BSD
+       pid = proc_selfpid();
+       if (task->bsd_info != NULL) {
+               procname = proc_name_address(current_task()->bsd_info);
+       }
+#endif
 
-               absolutetime_to_microtime(task->total_user_time, &secs, &usecs);
-               basic_info->user_time.seconds = 
-                       (typeof(basic_info->user_time.seconds))secs;
-               basic_info->user_time.microseconds = usecs;
+       ledger_get_entry_info(task->ledger, task_ledgers.interrupt_wakeups, &lei);
 
-               absolutetime_to_microtime(task->total_system_time, &secs, &usecs);
-               basic_info->system_time.seconds =
-                       (typeof(basic_info->system_time.seconds))secs;
-               basic_info->system_time.microseconds = usecs;
+       /*
+        * Disable the exception notification so we don't overwhelm
+        * the listener with an endless stream of redundant exceptions.
+        * TODO: detect whether another thread is already reporting the violation.
+        */
+       uint32_t flags = WAKEMON_DISABLE;
+       task_wakeups_monitor_ctl(task, &flags, NULL);
+
+       fatal = task->rusage_cpu_flags & TASK_RUSECPU_FLAGS_FATAL_WAKEUPSMON;
+       trace_resource_violation(RMON_CPUWAKES_VIOLATED, &lei);
+       os_log(OS_LOG_DEFAULT, "process %s[%d] caught waking the CPU %llu times "
+           "over ~%llu seconds, averaging %llu wakes / second and "
+           "violating a %slimit of %llu wakes over %llu seconds.\n",
+           procname, pid,
+           lei.lei_balance, lei.lei_last_refill / NSEC_PER_SEC,
+           lei.lei_last_refill == 0 ? 0 :
+           (NSEC_PER_SEC * lei.lei_balance / lei.lei_last_refill),
+           fatal ? "FATAL " : "",
+           lei.lei_limit, lei.lei_refill_period / NSEC_PER_SEC);
+
+       kr = send_resource_violation(send_cpu_wakes_violation, task, &lei,
+           fatal ? kRNFatalLimitFlag : 0);
+       if (kr) {
+               printf("send_resource_violation(CPU wakes, ...): error %#x\n", kr);
+       }
 
-               *task_info_count = TASK_BASIC_INFO_64_COUNT;
-               break;
+#ifdef EXC_RESOURCE_MONITORS
+       if (disable_exc_resource) {
+               printf("process %s[%d] caught causing excessive wakeups. EXC_RESOURCE "
+                   "supressed by a boot-arg\n", procname, pid);
+               return;
+       }
+       if (audio_active) {
+               os_log(OS_LOG_DEFAULT, "process %s[%d] caught causing excessive wakeups. EXC_RESOURCE "
+                   "supressed due to audio playback\n", procname, pid);
+               return;
+       }
+       if (lei.lei_last_refill == 0) {
+               os_log(OS_LOG_DEFAULT, "process %s[%d] caught causing excessive wakeups. EXC_RESOURCE "
+                   "supressed due to lei.lei_last_refill = 0 \n", procname, pid);
        }
 
-       case TASK_THREAD_TIMES_INFO:
-       {
-               register task_thread_times_info_t       times_info;
-               register thread_t                                       thread;
+       code[0] = code[1] = 0;
+       EXC_RESOURCE_ENCODE_TYPE(code[0], RESOURCE_TYPE_WAKEUPS);
+       EXC_RESOURCE_ENCODE_FLAVOR(code[0], FLAVOR_WAKEUPS_MONITOR);
+       EXC_RESOURCE_CPUMONITOR_ENCODE_WAKEUPS_PERMITTED(code[0],
+           NSEC_PER_SEC * lei.lei_limit / lei.lei_refill_period);
+       EXC_RESOURCE_CPUMONITOR_ENCODE_OBSERVATION_INTERVAL(code[0],
+           lei.lei_last_refill);
+       EXC_RESOURCE_CPUMONITOR_ENCODE_WAKEUPS_OBSERVED(code[1],
+           NSEC_PER_SEC * lei.lei_balance / lei.lei_last_refill);
+       exception_triage(EXC_RESOURCE, code, EXCEPTION_CODE_MAX);
+#endif /* EXC_RESOURCE_MONITORS */
+
+       if (fatal) {
+               task_terminate_internal(task);
+       }
+}
 
-               if (*task_info_count < TASK_THREAD_TIMES_INFO_COUNT) {
-                   error = KERN_INVALID_ARGUMENT;
-                   break;
+static boolean_t
+global_update_logical_writes(int64_t io_delta, int64_t *global_write_count)
+{
+       int64_t old_count, new_count;
+       boolean_t needs_telemetry;
+
+       do {
+               new_count = old_count = *global_write_count;
+               new_count += io_delta;
+               if (new_count >= io_telemetry_limit) {
+                       new_count = 0;
+                       needs_telemetry = TRUE;
+               } else {
+                       needs_telemetry = FALSE;
                }
+       } while (!OSCompareAndSwap64(old_count, new_count, global_write_count));
+       return needs_telemetry;
+}
 
-               times_info = (task_thread_times_info_t) task_info_out;
-               times_info->user_time.seconds = 0;
-               times_info->user_time.microseconds = 0;
-               times_info->system_time.seconds = 0;
-               times_info->system_time.microseconds = 0;
-
+void
+task_update_physical_writes(__unused task_t task, __unused task_physical_write_flavor_t flavor, __unused uint64_t io_size, __unused task_balance_flags_t flags)
+{
+#if CONFIG_PHYS_WRITE_ACCT
+       if (!io_size) {
+               return;
+       }
 
-               queue_iterate(&task->threads, thread, thread_t, task_threads) {
-                   time_value_t        user_time, system_time;
+       /*
+        * task == NULL means that we have to update kernel_task ledgers
+        */
+       if (!task) {
+               task = kernel_task;
+       }
 
-                   thread_read_times(thread, &user_time, &system_time);
+       KERNEL_DEBUG_CONSTANT((MACHDBG_CODE(DBG_MACH_VM, VM_PHYS_WRITE_ACCT)) | DBG_FUNC_NONE,
+           task_pid(task), flavor, io_size, flags, 0);
+       DTRACE_IO4(physical_writes, struct task *, task, task_physical_write_flavor_t, flavor, uint64_t, io_size, task_balance_flags_t, flags);
 
-                   time_value_add(&times_info->user_time, &user_time);
-                   time_value_add(&times_info->system_time, &system_time);
+       if (flags & TASK_BALANCE_CREDIT) {
+               if (flavor == TASK_PHYSICAL_WRITE_METADATA) {
+                       OSAddAtomic64(io_size, (SInt64 *)&(task->task_fs_metadata_writes));
+                       ledger_credit_nocheck(task->ledger, task_ledgers.fs_metadata_writes, io_size);
                }
+       } else if (flags & TASK_BALANCE_DEBIT) {
+               if (flavor == TASK_PHYSICAL_WRITE_METADATA) {
+                       OSAddAtomic64(-1 * io_size, (SInt64 *)&(task->task_fs_metadata_writes));
+                       ledger_debit_nocheck(task->ledger, task_ledgers.fs_metadata_writes, io_size);
+               }
+       }
+#endif /* CONFIG_PHYS_WRITE_ACCT */
+}
 
-
-               *task_info_count = TASK_THREAD_TIMES_INFO_COUNT;
-               break;
+void
+task_update_logical_writes(task_t task, uint32_t io_size, int flags, void *vp)
+{
+       int64_t io_delta = 0;
+       int64_t * global_counter_to_update;
+       boolean_t needs_telemetry = FALSE;
+       boolean_t is_external_device = FALSE;
+       int ledger_to_update = 0;
+       struct task_writes_counters * writes_counters_to_update;
+
+       if ((!task) || (!io_size) || (!vp)) {
+               return;
        }
 
-       case TASK_ABSOLUTETIME_INFO:
-       {
-               task_absolutetime_info_t        info;
-               register thread_t                       thread;
+       KERNEL_DEBUG_CONSTANT((MACHDBG_CODE(DBG_MACH_VM, VM_DATA_WRITE)) | DBG_FUNC_NONE,
+           task_pid(task), io_size, flags, (uintptr_t)VM_KERNEL_ADDRPERM(vp), 0);
+       DTRACE_IO4(logical_writes, struct task *, task, uint32_t, io_size, int, flags, vnode *, vp);
 
-               if (*task_info_count < TASK_ABSOLUTETIME_INFO_COUNT) {
-                       error = KERN_INVALID_ARGUMENT;
-                       break;
-               }
+       // Is the drive backing this vnode internal or external to the system?
+       if (vnode_isonexternalstorage(vp) == false) {
+               global_counter_to_update = &global_logical_writes_count;
+               ledger_to_update = task_ledgers.logical_writes;
+               writes_counters_to_update = &task->task_writes_counters_internal;
+               is_external_device = FALSE;
+       } else {
+               global_counter_to_update = &global_logical_writes_to_external_count;
+               ledger_to_update = task_ledgers.logical_writes_to_external;
+               writes_counters_to_update = &task->task_writes_counters_external;
+               is_external_device = TRUE;
+       }
 
-               info = (task_absolutetime_info_t)task_info_out;
-               info->threads_user = info->threads_system = 0;
+       switch (flags) {
+       case TASK_WRITE_IMMEDIATE:
+               OSAddAtomic64(io_size, (SInt64 *)&(writes_counters_to_update->task_immediate_writes));
+               ledger_credit(task->ledger, ledger_to_update, io_size);
+               if (!is_external_device) {
+                       coalition_io_ledger_update(task, FLAVOR_IO_LOGICAL_WRITES, TRUE, io_size);
+               }
+               break;
+       case TASK_WRITE_DEFERRED:
+               OSAddAtomic64(io_size, (SInt64 *)&(writes_counters_to_update->task_deferred_writes));
+               ledger_credit(task->ledger, ledger_to_update, io_size);
+               if (!is_external_device) {
+                       coalition_io_ledger_update(task, FLAVOR_IO_LOGICAL_WRITES, TRUE, io_size);
+               }
+               break;
+       case TASK_WRITE_INVALIDATED:
+               OSAddAtomic64(io_size, (SInt64 *)&(writes_counters_to_update->task_invalidated_writes));
+               ledger_debit(task->ledger, ledger_to_update, io_size);
+               if (!is_external_device) {
+                       coalition_io_ledger_update(task, FLAVOR_IO_LOGICAL_WRITES, FALSE, io_size);
+               }
+               break;
+       case TASK_WRITE_METADATA:
+               OSAddAtomic64(io_size, (SInt64 *)&(writes_counters_to_update->task_metadata_writes));
+               ledger_credit(task->ledger, ledger_to_update, io_size);
+               if (!is_external_device) {
+                       coalition_io_ledger_update(task, FLAVOR_IO_LOGICAL_WRITES, TRUE, io_size);
+               }
+               break;
+       }
 
+       io_delta = (flags == TASK_WRITE_INVALIDATED) ? ((int64_t)io_size * -1ll) : ((int64_t)io_size);
+       if (io_telemetry_limit != 0) {
+               /* If io_telemetry_limit is 0, disable global updates and I/O telemetry */
+               needs_telemetry = global_update_logical_writes(io_delta, global_counter_to_update);
+               if (needs_telemetry && !is_external_device) {
+                       act_set_io_telemetry_ast(current_thread());
+               }
+       }
+}
 
-               info->total_user = task->total_user_time;
-               info->total_system = task->total_system_time;
+/*
+ * Control the I/O monitor for a task.
+ */
+kern_return_t
+task_io_monitor_ctl(task_t task, uint32_t *flags)
+{
+       ledger_t ledger = task->ledger;
 
-               queue_iterate(&task->threads, thread, thread_t, task_threads) {
-                       uint64_t        tval;
+       task_lock(task);
+       if (*flags & IOMON_ENABLE) {
+               /* Configure the physical I/O ledger */
+               ledger_set_limit(ledger, task_ledgers.physical_writes, (task_iomon_limit_mb * 1024 * 1024), 0);
+               ledger_set_period(ledger, task_ledgers.physical_writes, (task_iomon_interval_secs * NSEC_PER_SEC));
+       } else if (*flags & IOMON_DISABLE) {
+               /*
+                * Caller wishes to disable I/O monitor on the task.
+                */
+               ledger_disable_refill(ledger, task_ledgers.physical_writes);
+               ledger_disable_callback(ledger, task_ledgers.physical_writes);
+       }
 
-                       tval = timer_grab(&thread->user_timer);
-                       info->threads_user += tval;
-                       info->total_user += tval;
+       task_unlock(task);
+       return KERN_SUCCESS;
+}
 
-                       tval = timer_grab(&thread->system_timer);
-                       info->threads_system += tval;
-                       info->total_system += tval;
-               }
+void
+task_io_rate_exceeded(int warning, const void *param0, __unused const void *param1)
+{
+       if (warning == 0) {
+               SENDING_NOTIFICATION__THIS_PROCESS_IS_CAUSING_TOO_MUCH_IO((int)param0);
+       }
+}
 
+void __attribute__((noinline))
+SENDING_NOTIFICATION__THIS_PROCESS_IS_CAUSING_TOO_MUCH_IO(int flavor)
+{
+       int                             pid = 0;
+       task_t                          task = current_task();
+#ifdef EXC_RESOURCE_MONITORS
+       mach_exception_data_type_t      code[EXCEPTION_CODE_MAX];
+#endif /* EXC_RESOURCE_MONITORS */
+       struct ledger_entry_info        lei;
+       kern_return_t                   kr;
 
-               *task_info_count = TASK_ABSOLUTETIME_INFO_COUNT;
+#ifdef MACH_BSD
+       pid = proc_selfpid();
+#endif
+       /*
+        * Get the ledger entry info. We need to do this before disabling the exception
+        * to get correct values for all fields.
+        */
+       switch (flavor) {
+       case FLAVOR_IO_PHYSICAL_WRITES:
+               ledger_get_entry_info(task->ledger, task_ledgers.physical_writes, &lei);
                break;
        }
 
-       case TASK_DYLD_INFO:
-       {
-               task_dyld_info_t info;
 
-               if (*task_info_count < TASK_DYLD_INFO_COUNT) {
-                       error = KERN_INVALID_ARGUMENT;
-                       break;
-               }
-               info = (task_dyld_info_t)task_info_out;
-               info->all_image_info_addr = task->all_image_info_addr;
-               info->all_image_info_size = task->all_image_info_size;
-               *task_info_count = TASK_DYLD_INFO_COUNT;
-               break;
+       /*
+        * Disable the exception notification so we don't overwhelm
+        * the listener with an endless stream of redundant exceptions.
+        * TODO: detect whether another thread is already reporting the violation.
+        */
+       uint32_t flags = IOMON_DISABLE;
+       task_io_monitor_ctl(task, &flags);
+
+       if (flavor == FLAVOR_IO_LOGICAL_WRITES) {
+               trace_resource_violation(RMON_LOGWRITES_VIOLATED, &lei);
        }
+       os_log(OS_LOG_DEFAULT, "process [%d] caught causing excessive I/O (flavor: %d). Task I/O: %lld MB. [Limit : %lld MB per %lld secs]\n",
+           pid, flavor, (lei.lei_balance / (1024 * 1024)), (lei.lei_limit / (1024 * 1024)), (lei.lei_refill_period / NSEC_PER_SEC));
 
-       /* OBSOLETE */
-       case TASK_SCHED_FIFO_INFO:
-       {
+       kr = send_resource_violation(send_disk_writes_violation, task, &lei, kRNFlagsNone);
+       if (kr) {
+               printf("send_resource_violation(disk_writes, ...): error %#x\n", kr);
+       }
 
-               if (*task_info_count < POLICY_FIFO_BASE_COUNT) {
-                       error = KERN_INVALID_ARGUMENT;
-                       break;
-               }
+#ifdef EXC_RESOURCE_MONITORS
+       code[0] = code[1] = 0;
+       EXC_RESOURCE_ENCODE_TYPE(code[0], RESOURCE_TYPE_IO);
+       EXC_RESOURCE_ENCODE_FLAVOR(code[0], flavor);
+       EXC_RESOURCE_IO_ENCODE_INTERVAL(code[0], (lei.lei_refill_period / NSEC_PER_SEC));
+       EXC_RESOURCE_IO_ENCODE_LIMIT(code[0], (lei.lei_limit / (1024 * 1024)));
+       EXC_RESOURCE_IO_ENCODE_OBSERVED(code[1], (lei.lei_balance / (1024 * 1024)));
+       exception_triage(EXC_RESOURCE, code, EXCEPTION_CODE_MAX);
+#endif /* EXC_RESOURCE_MONITORS */
+}
 
-               error = KERN_INVALID_POLICY;
+/* Placeholders for the task set/get voucher interfaces */
+kern_return_t
+task_get_mach_voucher(
+       task_t                  task,
+       mach_voucher_selector_t __unused which,
+       ipc_voucher_t           *voucher)
+{
+       if (TASK_NULL == task) {
+               return KERN_INVALID_TASK;
        }
 
-       /* OBSOLETE */
-       case TASK_SCHED_RR_INFO:
-       {
-               register policy_rr_base_t       rr_base;
-
-               if (*task_info_count < POLICY_RR_BASE_COUNT) {
-                       error = KERN_INVALID_ARGUMENT;
-                       break;
-               }
+       *voucher = NULL;
+       return KERN_SUCCESS;
+}
 
-               rr_base = (policy_rr_base_t) task_info_out;
+kern_return_t
+task_set_mach_voucher(
+       task_t                  task,
+       ipc_voucher_t           __unused voucher)
+{
+       if (TASK_NULL == task) {
+               return KERN_INVALID_TASK;
+       }
 
-               if (task != kernel_task) {
-                       error = KERN_INVALID_POLICY;
-                       break;
-               }
+       return KERN_SUCCESS;
+}
 
-               rr_base->base_priority = task->priority;
+kern_return_t
+task_swap_mach_voucher(
+       __unused task_t         task,
+       __unused ipc_voucher_t  new_voucher,
+       ipc_voucher_t          *in_out_old_voucher)
+{
+       /*
+        * Currently this function is only called from a MIG generated
+        * routine which doesn't release the reference on the voucher
+        * addressed by in_out_old_voucher. To avoid leaking this reference,
+        * a call to release it has been added here.
+        */
+       ipc_voucher_release(*in_out_old_voucher);
+       return KERN_NOT_SUPPORTED;
+}
 
-               rr_base->quantum = std_quantum_us / 1000;
+void
+task_set_gpu_denied(task_t task, boolean_t denied)
+{
+       task_lock(task);
 
-               *task_info_count = POLICY_RR_BASE_COUNT;
-               break;
+       if (denied) {
+               task->t_flags |= TF_GPU_DENIED;
+       } else {
+               task->t_flags &= ~TF_GPU_DENIED;
        }
 
-       /* OBSOLETE */
-       case TASK_SCHED_TIMESHARE_INFO:
-       {
-               register policy_timeshare_base_t        ts_base;
+       task_unlock(task);
+}
 
-               if (*task_info_count < POLICY_TIMESHARE_BASE_COUNT) {
-                       error = KERN_INVALID_ARGUMENT;
-                       break;
-               }
+boolean_t
+task_is_gpu_denied(task_t task)
+{
+       /* We don't need the lock to read this flag */
+       return (task->t_flags & TF_GPU_DENIED) ? TRUE : FALSE;
+}
 
-               ts_base = (policy_timeshare_base_t) task_info_out;
 
-               if (task == kernel_task) {
-                       error = KERN_INVALID_POLICY;
-                       break;
-               }
+uint64_t
+get_task_memory_region_count(task_t task)
+{
+       vm_map_t map;
+       map = (task == kernel_task) ? kernel_map: task->map;
+       return (uint64_t)get_map_nentries(map);
+}
 
-               ts_base->base_priority = task->priority;
+static void
+kdebug_trace_dyld_internal(uint32_t base_code,
+    struct dyld_kernel_image_info *info)
+{
+       static_assert(sizeof(info->uuid) >= 16);
+
+#if defined(__LP64__)
+       uint64_t *uuid = (uint64_t *)&(info->uuid);
+
+       KERNEL_DEBUG_CONSTANT_IST(KDEBUG_TRACE,
+           KDBG_EVENTID(DBG_DYLD, DBG_DYLD_UUID, base_code), uuid[0],
+           uuid[1], info->load_addr,
+           (uint64_t)info->fsid.val[0] | ((uint64_t)info->fsid.val[1] << 32),
+           0);
+       KERNEL_DEBUG_CONSTANT_IST(KDEBUG_TRACE,
+           KDBG_EVENTID(DBG_DYLD, DBG_DYLD_UUID, base_code + 1),
+           (uint64_t)info->fsobjid.fid_objno |
+           ((uint64_t)info->fsobjid.fid_generation << 32),
+           0, 0, 0, 0);
+#else /* defined(__LP64__) */
+       uint32_t *uuid = (uint32_t *)&(info->uuid);
+
+       KERNEL_DEBUG_CONSTANT_IST(KDEBUG_TRACE,
+           KDBG_EVENTID(DBG_DYLD, DBG_DYLD_UUID, base_code + 2), uuid[0],
+           uuid[1], uuid[2], uuid[3], 0);
+       KERNEL_DEBUG_CONSTANT_IST(KDEBUG_TRACE,
+           KDBG_EVENTID(DBG_DYLD, DBG_DYLD_UUID, base_code + 3),
+           (uint32_t)info->load_addr, info->fsid.val[0], info->fsid.val[1],
+           info->fsobjid.fid_objno, 0);
+       KERNEL_DEBUG_CONSTANT_IST(KDEBUG_TRACE,
+           KDBG_EVENTID(DBG_DYLD, DBG_DYLD_UUID, base_code + 4),
+           info->fsobjid.fid_generation, 0, 0, 0, 0);
+#endif /* !defined(__LP64__) */
+}
 
-               *task_info_count = POLICY_TIMESHARE_BASE_COUNT;
-               break;
+static kern_return_t
+kdebug_trace_dyld(task_t task, uint32_t base_code,
+    vm_map_copy_t infos_copy, mach_msg_type_number_t infos_len)
+{
+       kern_return_t kr;
+       dyld_kernel_image_info_array_t infos;
+       vm_map_offset_t map_data;
+       vm_offset_t data;
+
+       if (!infos_copy) {
+               return KERN_INVALID_ADDRESS;
        }
 
-       case TASK_SECURITY_TOKEN:
-       {
-               register security_token_t       *sec_token_p;
+       if (!kdebug_enable ||
+           !kdebug_debugid_enabled(KDBG_EVENTID(DBG_DYLD, DBG_DYLD_UUID, 0))) {
+               vm_map_copy_discard(infos_copy);
+               return KERN_SUCCESS;
+       }
 
-               if (*task_info_count < TASK_SECURITY_TOKEN_COUNT) {
-                   error = KERN_INVALID_ARGUMENT;
-                   break;
-               }
+       if (task == NULL || task != current_task()) {
+               return KERN_INVALID_TASK;
+       }
 
-               sec_token_p = (security_token_t *) task_info_out;
+       kr = vm_map_copyout(ipc_kernel_map, &map_data, (vm_map_copy_t)infos_copy);
+       if (kr != KERN_SUCCESS) {
+               return kr;
+       }
 
-               *sec_token_p = task->sec_token;
+       infos = CAST_DOWN(dyld_kernel_image_info_array_t, map_data);
 
-               *task_info_count = TASK_SECURITY_TOKEN_COUNT;
-               break;
+       for (mach_msg_type_number_t i = 0; i < infos_len; i++) {
+               kdebug_trace_dyld_internal(base_code, &(infos[i]));
        }
-            
-       case TASK_AUDIT_TOKEN:
-       {
-               register audit_token_t  *audit_token_p;
 
-               if (*task_info_count < TASK_AUDIT_TOKEN_COUNT) {
-                   error = KERN_INVALID_ARGUMENT;
-                   break;
-               }
+       data = CAST_DOWN(vm_offset_t, map_data);
+       mach_vm_deallocate(ipc_kernel_map, data, infos_len * sizeof(infos[0]));
+       return KERN_SUCCESS;
+}
 
-               audit_token_p = (audit_token_t *) task_info_out;
+kern_return_t
+task_register_dyld_image_infos(task_t task,
+    dyld_kernel_image_info_array_t infos_copy,
+    mach_msg_type_number_t infos_len)
+{
+       return kdebug_trace_dyld(task, DBG_DYLD_UUID_MAP_A,
+                  (vm_map_copy_t)infos_copy, infos_len);
+}
 
-               *audit_token_p = task->audit_token;
+kern_return_t
+task_unregister_dyld_image_infos(task_t task,
+    dyld_kernel_image_info_array_t infos_copy,
+    mach_msg_type_number_t infos_len)
+{
+       return kdebug_trace_dyld(task, DBG_DYLD_UUID_UNMAP_A,
+                  (vm_map_copy_t)infos_copy, infos_len);
+}
 
-               *task_info_count = TASK_AUDIT_TOKEN_COUNT;
-               break;
-       }
-            
-       case TASK_SCHED_INFO:
-               error = KERN_INVALID_ARGUMENT;
+kern_return_t
+task_get_dyld_image_infos(__unused task_t task,
+    __unused dyld_kernel_image_info_array_t * dyld_images,
+    __unused mach_msg_type_number_t * dyld_imagesCnt)
+{
+       return KERN_NOT_SUPPORTED;
+}
 
-       case TASK_EVENTS_INFO:
-       {
-               register task_events_info_t     events_info;
-               register thread_t                       thread;
+kern_return_t
+task_register_dyld_shared_cache_image_info(task_t task,
+    dyld_kernel_image_info_t cache_img,
+    __unused boolean_t no_cache,
+    __unused boolean_t private_cache)
+{
+       if (task == NULL || task != current_task()) {
+               return KERN_INVALID_TASK;
+       }
 
-               if (*task_info_count < TASK_EVENTS_INFO_COUNT) {
-                  error = KERN_INVALID_ARGUMENT;
-                  break;
-               }
+       kdebug_trace_dyld_internal(DBG_DYLD_UUID_SHARED_CACHE_A, &cache_img);
+       return KERN_SUCCESS;
+}
 
-               events_info = (task_events_info_t) task_info_out;
+kern_return_t
+task_register_dyld_set_dyld_state(__unused task_t task,
+    __unused uint8_t dyld_state)
+{
+       return KERN_NOT_SUPPORTED;
+}
 
+kern_return_t
+task_register_dyld_get_process_state(__unused task_t task,
+    __unused dyld_kernel_process_info_t * dyld_process_state)
+{
+       return KERN_NOT_SUPPORTED;
+}
 
-               events_info->faults = task->faults;
-               events_info->pageins = task->pageins;
-               events_info->cow_faults = task->cow_faults;
-               events_info->messages_sent = task->messages_sent;
-               events_info->messages_received = task->messages_received;
-               events_info->syscalls_mach = task->syscalls_mach;
-               events_info->syscalls_unix = task->syscalls_unix;
+kern_return_t
+task_inspect(task_inspect_t task_insp, task_inspect_flavor_t flavor,
+    task_inspect_info_t info_out, mach_msg_type_number_t *size_in_out)
+{
+#if MONOTONIC
+       task_t task = (task_t)task_insp;
+       kern_return_t kr = KERN_SUCCESS;
+       mach_msg_type_number_t size;
 
-               events_info->csw = task->c_switch;
+       if (task == TASK_NULL) {
+               return KERN_INVALID_ARGUMENT;
+       }
 
-               queue_iterate(&task->threads, thread, thread_t, task_threads) {
-                       events_info->csw += thread->c_switch;
-               }
+       size = *size_in_out;
 
+       switch (flavor) {
+       case TASK_INSPECT_BASIC_COUNTS: {
+               struct task_inspect_basic_counts *bc;
+               uint64_t task_counts[MT_CORE_NFIXED] = { 0 };
 
-               *task_info_count = TASK_EVENTS_INFO_COUNT;
-               break;
-       }
-       case TASK_AFFINITY_TAG_INFO:
-       {
-               if (*task_info_count < TASK_AFFINITY_TAG_INFO_COUNT) {
-                   error = KERN_INVALID_ARGUMENT;
-                   break;
+               if (size < TASK_INSPECT_BASIC_COUNTS_COUNT) {
+                       kr = KERN_INVALID_ARGUMENT;
+                       break;
                }
 
-               error = task_affinity_info(task, task_info_out, task_info_count);
+               mt_fixed_task_counts(task, task_counts);
+               bc = (struct task_inspect_basic_counts *)info_out;
+#ifdef MT_CORE_INSTRS
+               bc->instructions = task_counts[MT_CORE_INSTRS];
+#else /* defined(MT_CORE_INSTRS) */
+               bc->instructions = 0;
+#endif /* !defined(MT_CORE_INSTRS) */
+               bc->cycles = task_counts[MT_CORE_CYCLES];
+               size = TASK_INSPECT_BASIC_COUNTS_COUNT;
+               break;
        }
-
        default:
-               error = KERN_INVALID_ARGUMENT;
+               kr = KERN_INVALID_ARGUMENT;
+               break;
        }
 
-       task_unlock(task);
-       return (error);
+       if (kr == KERN_SUCCESS) {
+               *size_in_out = size;
+       }
+       return kr;
+#else /* MONOTONIC */
+#pragma unused(task_insp, flavor, info_out, size_in_out)
+       return KERN_NOT_SUPPORTED;
+#endif /* !MONOTONIC */
 }
 
+#if CONFIG_SECLUDED_MEMORY
+int num_tasks_can_use_secluded_mem = 0;
+
 void
-task_vtimer_set(
-       task_t          task,
-       integer_t       which)
+task_set_can_use_secluded_mem(
+       task_t          task,
+       boolean_t       can_use_secluded_mem)
 {
-       thread_t        thread;
-
-       /* assert(task == current_task()); */ /* bogus assert 4803227 4807483 */
-
+       if (!task->task_could_use_secluded_mem) {
+               return;
+       }
        task_lock(task);
+       task_set_can_use_secluded_mem_locked(task, can_use_secluded_mem);
+       task_unlock(task);
+}
 
-       task->vtimers |= which;
-
-       switch (which) {
+void
+task_set_can_use_secluded_mem_locked(
+       task_t          task,
+       boolean_t       can_use_secluded_mem)
+{
+       assert(task->task_could_use_secluded_mem);
+       if (can_use_secluded_mem &&
+           secluded_for_apps && /* global boot-arg */
+           !task->task_can_use_secluded_mem) {
+               assert(num_tasks_can_use_secluded_mem >= 0);
+               OSAddAtomic(+1,
+                   (volatile SInt32 *)&num_tasks_can_use_secluded_mem);
+               task->task_can_use_secluded_mem = TRUE;
+       } else if (!can_use_secluded_mem &&
+           task->task_can_use_secluded_mem) {
+               assert(num_tasks_can_use_secluded_mem > 0);
+               OSAddAtomic(-1,
+                   (volatile SInt32 *)&num_tasks_can_use_secluded_mem);
+               task->task_can_use_secluded_mem = FALSE;
+       }
+}
 
-       case TASK_VTIMER_USER:
-               queue_iterate(&task->threads, thread, thread_t, task_threads) {
-                       thread->vtimer_user_save = timer_grab(&thread->user_timer);
-               }
-               break;
+void
+task_set_could_use_secluded_mem(
+       task_t          task,
+       boolean_t       could_use_secluded_mem)
+{
+       task->task_could_use_secluded_mem = !!could_use_secluded_mem;
+}
 
-       case TASK_VTIMER_PROF:
-               queue_iterate(&task->threads, thread, thread_t, task_threads) {
-                       thread->vtimer_prof_save = timer_grab(&thread->user_timer);
-                       thread->vtimer_prof_save += timer_grab(&thread->system_timer);
-               }
-               break;
+void
+task_set_could_also_use_secluded_mem(
+       task_t          task,
+       boolean_t       could_also_use_secluded_mem)
+{
+       task->task_could_also_use_secluded_mem = !!could_also_use_secluded_mem;
+}
 
-       case TASK_VTIMER_RLIM:
-               queue_iterate(&task->threads, thread, thread_t, task_threads) {
-                       thread->vtimer_rlim_save = timer_grab(&thread->user_timer);
-                       thread->vtimer_rlim_save += timer_grab(&thread->system_timer);
+boolean_t
+task_can_use_secluded_mem(
+       task_t          task,
+       boolean_t       is_alloc)
+{
+       if (task->task_can_use_secluded_mem) {
+               assert(task->task_could_use_secluded_mem);
+               assert(num_tasks_can_use_secluded_mem > 0);
+               return TRUE;
+       }
+       if (task->task_could_also_use_secluded_mem &&
+           num_tasks_can_use_secluded_mem > 0) {
+               assert(num_tasks_can_use_secluded_mem > 0);
+               return TRUE;
+       }
+
+       /*
+        * If a single task is using more than some large amount of
+        * memory (i.e. secluded_shutoff_trigger) and is approaching
+        * its task limit, allow it to dip into secluded and begin
+        * suppression of rebuilding secluded memory until that task exits.
+        */
+       if (is_alloc && secluded_shutoff_trigger != 0) {
+               uint64_t phys_used = get_task_phys_footprint(task);
+               uint64_t limit = get_task_phys_footprint_limit(task);
+               if (phys_used > secluded_shutoff_trigger &&
+                   limit > secluded_shutoff_trigger &&
+                   phys_used > limit - secluded_shutoff_headroom) {
+                       start_secluded_suppression(task);
+                       return TRUE;
                }
-               break;
        }
 
-       task_unlock(task);
+       return FALSE;
 }
 
-void
-task_vtimer_clear(
-       task_t          task,
-       integer_t       which)
+boolean_t
+task_could_use_secluded_mem(
+       task_t  task)
 {
-       assert(task == current_task());
+       return task->task_could_use_secluded_mem;
+}
 
-       task_lock(task);
+boolean_t
+task_could_also_use_secluded_mem(
+       task_t  task)
+{
+       return task->task_could_also_use_secluded_mem;
+}
+#endif /* CONFIG_SECLUDED_MEMORY */
 
-       task->vtimers &= ~which;
+queue_head_t *
+task_io_user_clients(task_t task)
+{
+       return &task->io_user_clients;
+}
 
-       task_unlock(task);
+void
+task_set_message_app_suspended(task_t task, boolean_t enable)
+{
+       task->message_app_suspended = enable;
 }
 
 void
-task_vtimer_update(
-__unused
-       task_t          task,
-       integer_t       which,
-       uint32_t        *microsecs)
+task_copy_fields_for_exec(task_t dst_task, task_t src_task)
 {
-       thread_t        thread = current_thread();
-       uint32_t        tdelt;
-       clock_sec_t     secs;
-       uint64_t        tsum;
+       dst_task->vtimers = src_task->vtimers;
+}
 
-       assert(task == current_task());
+#if DEVELOPMENT || DEBUG
+int vm_region_footprint = 0;
+#endif /* DEVELOPMENT || DEBUG */
 
-       assert(task->vtimers & which);
+boolean_t
+task_self_region_footprint(void)
+{
+#if DEVELOPMENT || DEBUG
+       if (vm_region_footprint) {
+               /* system-wide override */
+               return TRUE;
+       }
+#endif /* DEVELOPMENT || DEBUG */
+       return current_task()->task_region_footprint;
+}
 
-       secs = tdelt = 0;
+void
+task_self_region_footprint_set(
+       boolean_t newval)
+{
+       task_t  curtask;
 
-       switch (which) {
+       curtask = current_task();
+       task_lock(curtask);
+       if (newval) {
+               curtask->task_region_footprint = TRUE;
+       } else {
+               curtask->task_region_footprint = FALSE;
+       }
+       task_unlock(curtask);
+}
 
-       case TASK_VTIMER_USER:
-               tdelt = (uint32_t)timer_delta(&thread->user_timer,
-                                                               &thread->vtimer_user_save);
-               absolutetime_to_microtime(tdelt, &secs, microsecs);
-               break;
+void
+task_set_darkwake_mode(task_t task, boolean_t set_mode)
+{
+       assert(task);
 
-       case TASK_VTIMER_PROF:
-               tsum = timer_grab(&thread->user_timer);
-               tsum += timer_grab(&thread->system_timer);
-               tdelt = (uint32_t)(tsum - thread->vtimer_prof_save);
-               absolutetime_to_microtime(tdelt, &secs, microsecs);
-               /* if the time delta is smaller than a usec, ignore */
-               if (*microsecs != 0)
-                       thread->vtimer_prof_save = tsum;
-               break;
+       task_lock(task);
 
-       case TASK_VTIMER_RLIM:
-               tsum = timer_grab(&thread->user_timer);
-               tsum += timer_grab(&thread->system_timer);
-               tdelt = (uint32_t)(tsum - thread->vtimer_rlim_save);
-               thread->vtimer_rlim_save = tsum;
-               absolutetime_to_microtime(tdelt, &secs, microsecs);
-               break;
+       if (set_mode) {
+               task->t_flags |= TF_DARKWAKE_MODE;
+       } else {
+               task->t_flags &= ~(TF_DARKWAKE_MODE);
        }
 
+       task_unlock(task);
 }
 
-/*
- *     task_assign:
- *
- *     Change the assigned processor set for the task
- */
-kern_return_t
-task_assign(
-       __unused task_t         task,
-       __unused processor_set_t        new_pset,
-       __unused boolean_t      assign_threads)
+boolean_t
+task_get_darkwake_mode(task_t task)
 {
-       return(KERN_FAILURE);
+       assert(task);
+       return (task->t_flags & TF_DARKWAKE_MODE) != 0;
 }
 
-/*
- *     task_assign_default:
- *
- *     Version of task_assign to assign to default processor set.
- */
 kern_return_t
-task_assign_default(
-       task_t          task,
-       boolean_t       assign_threads)
+task_get_exc_guard_behavior(
+       task_t task,
+       task_exc_guard_behavior_t *behaviorp)
 {
-    return (task_assign(task, &pset0, assign_threads));
+       if (task == TASK_NULL) {
+               return KERN_INVALID_TASK;
+       }
+       *behaviorp = task->task_exc_guard;
+       return KERN_SUCCESS;
 }
 
-/*
- *     task_get_assignment
- *
- *     Return name of processor set that task is assigned to.
- */
+#ifndef TASK_EXC_GUARD_ALL
+/* Temporary define until two branches are merged */
+#define TASK_EXC_GUARD_ALL (TASK_EXC_GUARD_VM_ALL | 0xf0)
+#endif
+
 kern_return_t
-task_get_assignment(
-       task_t          task,
-       processor_set_t *pset)
+task_set_exc_guard_behavior(
+       task_t task,
+       task_exc_guard_behavior_t behavior)
 {
-       if (!task->active)
-               return(KERN_FAILURE);
-
-       *pset = &pset0;
-
-       return (KERN_SUCCESS);
+       if (task == TASK_NULL) {
+               return KERN_INVALID_TASK;
+       }
+       if (behavior & ~TASK_EXC_GUARD_ALL) {
+               return KERN_INVALID_VALUE;
+       }
+       task->task_exc_guard = behavior;
+       return KERN_SUCCESS;
 }
 
+#if __arm64__
+extern int legacy_footprint_entitlement_mode;
+extern void memorystatus_act_on_legacy_footprint_entitlement(struct proc *, boolean_t);
+extern void memorystatus_act_on_ios13extended_footprint_entitlement(struct proc *);
 
-/*
- *     task_policy
- *
- *     Set scheduling policy and parameters, both base and limit, for
- *     the given task. Policy must be a policy which is enabled for the
- *     processor set. Change contained threads if requested. 
- */
-kern_return_t
-task_policy(
-       __unused task_t                 task,
-       __unused policy_t                       policy_id,
-       __unused policy_base_t          base,
-       __unused mach_msg_type_number_t count,
-       __unused boolean_t                      set_limit,
-       __unused boolean_t                      change)
-{
-       return(KERN_FAILURE);
-}
 
-/*
- *     task_set_policy
- *
- *     Set scheduling policy and parameters, both base and limit, for 
- *     the given task. Policy can be any policy implemented by the
- *     processor set, whether enabled or not. Change contained threads
- *     if requested.
- */
-kern_return_t
-task_set_policy(
-       __unused task_t                 task,
-       __unused processor_set_t                pset,
-       __unused policy_t                       policy_id,
-       __unused policy_base_t          base,
-       __unused mach_msg_type_number_t base_count,
-       __unused policy_limit_t         limit,
-       __unused mach_msg_type_number_t limit_count,
-       __unused boolean_t                      change)
+void
+task_set_legacy_footprint(
+       task_t task)
 {
-       return(KERN_FAILURE);
+       task_lock(task);
+       task->task_legacy_footprint = TRUE;
+       task_unlock(task);
 }
 
-#if    FAST_TAS
-kern_return_t
-task_set_ras_pc(
-       task_t          task,
-       vm_offset_t     pc,
-       vm_offset_t     endpc)
+void
+task_set_extra_footprint_limit(
+       task_t task)
 {
-       extern int fast_tas_debug;
-       if (fast_tas_debug) {
-               printf("task 0x%x: setting fast_tas to [0x%x, 0x%x]\n",
-                      task, pc, endpc);
+       if (task->task_extra_footprint_limit) {
+               return;
        }
        task_lock(task);
-       task->fast_tas_base = pc;
-       task->fast_tas_end =  endpc;
+       if (task->task_extra_footprint_limit) {
+               task_unlock(task);
+               return;
+       }
+       task->task_extra_footprint_limit = TRUE;
        task_unlock(task);
-       return KERN_SUCCESS;
-} 
-#else  /* FAST_TAS */
-kern_return_t
-task_set_ras_pc(
-       __unused task_t task,
-       __unused vm_offset_t    pc,
-       __unused vm_offset_t    endpc)
-{
-       return KERN_FAILURE;
+       memorystatus_act_on_legacy_footprint_entitlement(task->bsd_info, TRUE);
 }
-#endif /* FAST_TAS */
 
 void
-task_synchronizer_destroy_all(task_t task)
+task_set_ios13extended_footprint_limit(
+       task_t task)
 {
-       semaphore_t     semaphore;
-       lock_set_t      lock_set;
-
-       /*
-        *  Destroy owned semaphores
-        */
-
-       while (!queue_empty(&task->semaphore_list)) {
-               semaphore = (semaphore_t) queue_first(&task->semaphore_list);
-               (void) semaphore_destroy(task, semaphore);
+       if (task->task_ios13extended_footprint_limit) {
+               return;
        }
-
-       /*
-        *  Destroy owned lock sets
-        */
-
-       while (!queue_empty(&task->lock_set_list)) {
-               lock_set = (lock_set_t) queue_first(&task->lock_set_list);
-               (void) lock_set_destroy(task, lock_set);
+       task_lock(task);
+       if (task->task_ios13extended_footprint_limit) {
+               task_unlock(task);
+               return;
        }
+       task->task_ios13extended_footprint_limit = TRUE;
+       task_unlock(task);
+       memorystatus_act_on_ios13extended_footprint_entitlement(task->bsd_info);
+}
+#endif /* __arm64__ */
+
+static inline ledger_amount_t
+task_ledger_get_balance(
+       ledger_t        ledger,
+       int             ledger_idx)
+{
+       ledger_amount_t amount;
+       amount = 0;
+       ledger_get_balance(ledger, ledger_idx, &amount);
+       return amount;
 }
 
 /*
- * Install default (machine-dependent) initial thread state 
- * on the task.  Subsequent thread creation will have this initial
- * state set on the thread by machine_thread_inherit_taskwide().
- * Flavors and structures are exactly the same as those to thread_set_state()
+ * Gather the amount of memory counted in a task's footprint due to
+ * being in a specific set of ledgers.
  */
-kern_return_t 
-task_set_state(
-       task_t task, 
-       int flavor, 
-       thread_state_t state, 
-       mach_msg_type_number_t state_count)
+void
+task_ledgers_footprint(
+       ledger_t        ledger,
+       ledger_amount_t *ledger_resident,
+       ledger_amount_t *ledger_compressed)
 {
-       kern_return_t ret;
+       *ledger_resident = 0;
+       *ledger_compressed = 0;
 
-       if (task == TASK_NULL) {
-               return (KERN_INVALID_ARGUMENT);
-       }
+       /* purgeable non-volatile memory */
+       *ledger_resident += task_ledger_get_balance(ledger, task_ledgers.purgeable_nonvolatile);
+       *ledger_compressed += task_ledger_get_balance(ledger, task_ledgers.purgeable_nonvolatile_compressed);
 
-       task_lock(task);
+       /* "default" tagged memory */
+       *ledger_resident += task_ledger_get_balance(ledger, task_ledgers.tagged_footprint);
+       *ledger_compressed += task_ledger_get_balance(ledger, task_ledgers.tagged_footprint_compressed);
 
-       if (!task->active) {
-               task_unlock(task);
-               return (KERN_FAILURE);
-       }
+       /* "network" currently never counts in the footprint... */
 
-       ret = machine_task_set_state(task, flavor, state, state_count);
+       /* "media" tagged memory */
+       *ledger_resident += task_ledger_get_balance(ledger, task_ledgers.media_footprint);
+       *ledger_compressed += task_ledger_get_balance(ledger, task_ledgers.media_footprint_compressed);
+
+       /* "graphics" tagged memory */
+       *ledger_resident += task_ledger_get_balance(ledger, task_ledgers.graphics_footprint);
+       *ledger_compressed += task_ledger_get_balance(ledger, task_ledgers.graphics_footprint_compressed);
+
+       /* "neural" tagged memory */
+       *ledger_resident += task_ledger_get_balance(ledger, task_ledgers.neural_footprint);
+       *ledger_compressed += task_ledger_get_balance(ledger, task_ledgers.neural_footprint_compressed);
+}
 
+void
+task_set_memory_ownership_transfer(
+       task_t    task,
+       boolean_t value)
+{
+       task_lock(task);
+       task->task_can_transfer_memory_ownership = !!value;
        task_unlock(task);
-       return ret;
 }
 
-/*
- * Examine the default (machine-dependent) initial thread state 
- * on the task, as set by task_set_state().  Flavors and structures
- * are exactly the same as those passed to thread_get_state().
- */
-kern_return_t 
-task_get_state(
-       task_t  task, 
-       int     flavor,
-       thread_state_t state,
-       mach_msg_type_number_t *state_count)
+void
+task_copy_vmobjects(task_t task, vm_object_query_t query, size_t len, size_t *num)
 {
-       kern_return_t ret;
+       vm_object_t find_vmo;
+       size_t size = 0;
+
+       task_objq_lock(task);
+       if (query != NULL) {
+               queue_iterate(&task->task_objq, find_vmo, vm_object_t, task_objq)
+               {
+                       vm_object_query_t p = &query[size++];
+
+                       /* make sure to not overrun */
+                       if (size * sizeof(vm_object_query_data_t) > len) {
+                               --size;
+                               break;
+                       }
 
-       if (task == TASK_NULL) {
-               return (KERN_INVALID_ARGUMENT);
+                       bzero(p, sizeof(*p));
+                       p->object_id = (vm_object_id_t) VM_KERNEL_ADDRPERM(find_vmo);
+                       p->virtual_size = find_vmo->internal ? find_vmo->vo_size : 0;
+                       p->resident_size = find_vmo->resident_page_count * PAGE_SIZE;
+                       p->wired_size = find_vmo->wired_page_count * PAGE_SIZE;
+                       p->reusable_size = find_vmo->reusable_page_count * PAGE_SIZE;
+                       p->vo_no_footprint = find_vmo->vo_no_footprint;
+                       p->vo_ledger_tag = find_vmo->vo_ledger_tag;
+                       p->purgable = find_vmo->purgable;
+
+                       if (find_vmo->internal && find_vmo->pager_created && find_vmo->pager != NULL) {
+                               p->compressed_size = vm_compressor_pager_get_count(find_vmo->pager) * PAGE_SIZE;
+                       } else {
+                               p->compressed_size = 0;
+                       }
+               }
+       } else {
+               size = (size_t)task->task_owned_objects;
        }
+       task_objq_unlock(task);
 
-       task_lock(task);
-
-       if (!task->active) {
-               task_unlock(task);
-               return (KERN_FAILURE);
-       }
+       *num = size;
+}
 
-       ret = machine_task_get_state(task, flavor, state, state_count);
+void
+task_set_filter_msg_flag(
+       task_t task,
+       boolean_t flag)
+{
+       assert(task != TASK_NULL);
 
+       task_lock(task);
+       if (flag) {
+               task->t_flags |= TF_FILTER_MSG;
+       } else {
+               task->t_flags &= ~TF_FILTER_MSG;
+       }
        task_unlock(task);
-       return ret;
 }
 
-
-/*
- * We need to export some functions to other components that
- * are currently implemented in macros within the osfmk
- * component.  Just export them as functions of the same name.
- */
-boolean_t is_kerneltask(task_t t)
+boolean_t
+task_get_filter_msg_flag(
+       task_t task)
 {
-       if (t == kernel_task)
-               return (TRUE);
+       uint32_t flags = 0;
+
+       if (!task) {
+               return false;
+       }
 
-       return (FALSE);
+       flags = os_atomic_load(&task->t_flags, relaxed);
+       return (flags & TF_FILTER_MSG) ? TRUE : FALSE;
+}
+bool
+task_is_exotic(
+       task_t task)
+{
+       if (task == TASK_NULL) {
+               return false;
+       }
+       return vm_map_is_exotic(get_task_map(task));
 }
 
-int
-check_for_tasksuspend(task_t task)
+bool
+task_is_alien(
+       task_t task)
 {
+       if (task == TASK_NULL) {
+               return false;
+       }
+       return vm_map_is_alien(get_task_map(task));
+}
 
-       if (task == TASK_NULL)
-               return (0);
 
-       return (task->suspend_count > 0);
-}
 
-#undef current_task
-task_t current_task(void);
-task_t current_task(void)
+#if CONFIG_MACF
+/* Set the filter mask for Mach traps. */
+void
+mac_task_set_mach_filter_mask(task_t task, uint8_t *maskptr)
 {
-       return (current_task_fast());
+       assert(task);
+
+       task->mach_trap_filter_mask = maskptr;
 }
 
-#undef task_reference
-void task_reference(task_t task);
+/* Set the filter mask for kobject msgs. */
 void
-task_reference(
-       task_t          task)
+mac_task_set_kobj_filter_mask(task_t task, uint8_t *maskptr)
 {
-       if (task != TASK_NULL)
-               task_reference_internal(task);
+       assert(task);
+
+       task->mach_kobj_filter_mask = maskptr;
 }
 
-#if CONFIG_MACF_MACH
-/*
- * Protect 2 task labels against modification by adding a reference on
- * both label handles. The locks do not actually have to be held while
- * using the labels as only labels with one reference can be modified
- * in place.
- */
+/* Hook for mach trap/sc filter evaluation policy. */
+mac_task_mach_filter_cbfunc_t mac_task_mach_trap_evaluate = NULL;
 
-void
-tasklabel_lock2(
-       task_t a,
-       task_t b)
+/* Hook for kobj message filter evaluation policy. */
+mac_task_kobj_filter_cbfunc_t mac_task_kobj_msg_evaluate = NULL;
+
+/* Set the callback hooks for the filtering policy. */
+int
+mac_task_register_filter_callbacks(
+       const mac_task_mach_filter_cbfunc_t mach_cbfunc,
+       const mac_task_kobj_filter_cbfunc_t kobj_cbfunc)
 {
-       labelh_reference(a->label);
-       labelh_reference(b->label);
+       if (mach_cbfunc != NULL) {
+               if (mac_task_mach_trap_evaluate != NULL) {
+                       return KERN_FAILURE;
+               }
+               mac_task_mach_trap_evaluate = mach_cbfunc;
+       }
+       if (kobj_cbfunc != NULL) {
+               if (mac_task_kobj_msg_evaluate != NULL) {
+                       return KERN_FAILURE;
+               }
+               mac_task_kobj_msg_evaluate = kobj_cbfunc;
+       }
+
+       return KERN_SUCCESS;
 }
+#endif /* CONFIG_MACF */
 
 void
-tasklabel_unlock2(
-       task_t a,
-       task_t b)
+task_transfer_mach_filter_bits(
+       task_t new_task,
+       task_t old_task)
 {
-       labelh_release(a->label);
-       labelh_release(b->label);
+#ifdef CONFIG_MACF
+       /* Copy mach trap and kernel object mask pointers to new task. */
+       new_task->mach_trap_filter_mask = old_task->mach_trap_filter_mask;
+       new_task->mach_kobj_filter_mask = old_task->mach_kobj_filter_mask;
+#endif
+       /* If filter message flag is set then set it in the new task. */
+       if (task_get_filter_msg_flag(old_task)) {
+               new_task->t_flags |= TF_FILTER_MSG;
+       }
 }
 
+
+#if __has_feature(ptrauth_calls)
+
+#define PAC_EXCEPTION_ENTITLEMENT "com.apple.private.pac.exception"
+
 void
-mac_task_label_update_internal(
-       struct label    *pl,
-       struct task     *task)
+task_set_pac_exception_fatal_flag(
+       task_t task)
 {
+       assert(task != TASK_NULL);
+
+       if (!IOTaskHasEntitlement(task, PAC_EXCEPTION_ENTITLEMENT)) {
+               return;
+       }
 
-       tasklabel_lock(task);
-       task->label = labelh_modify(task->label);
-       mac_task_label_update(pl, &task->maclabel);
-       tasklabel_unlock(task);
-       ip_lock(task->itk_self);
-       mac_port_label_update_cred(pl, &task->itk_self->ip_label);
-       ip_unlock(task->itk_self);
+       task_lock(task);
+       task->t_flags |= TF_PAC_EXC_FATAL;
+       task_unlock(task);
 }
 
-void
-mac_task_label_modify(
-       struct task     *task,
-       void            *arg,
-       void (*f)       (struct label *l, void *arg))
+bool
+task_is_pac_exception_fatal(
+       task_t task)
 {
+       uint32_t flags = 0;
+
+       assert(task != TASK_NULL);
 
-       tasklabel_lock(task);
-       task->label = labelh_modify(task->label);
-       (*f)(&task->maclabel, arg);
-       tasklabel_unlock(task);
+       flags = os_atomic_load(&task->t_flags, relaxed);
+       return (bool)(flags & TF_PAC_EXC_FATAL);
 }
+#endif /* __has_feature(ptrauth_calls) */
 
-struct label *
-mac_task_get_label(struct task *task)
+void
+task_set_tecs(task_t task)
 {
-       return (&task->maclabel);
+       if (task == TASK_NULL) {
+               task = current_task();
+       }
+
+       if (!machine_csv(CPUVN_CI)) {
+               return;
+       }
+
+       LCK_MTX_ASSERT(&task->lock, LCK_MTX_ASSERT_NOTOWNED);
+
+       task_lock(task);
+
+       task->t_flags |= TF_TECS;
+
+       thread_t thread;
+       queue_iterate(&task->threads, thread, thread_t, task_threads) {
+               machine_tecs(thread);
+       }
+       task_unlock(task);
 }
-#endif