X-Git-Url: https://git.saurik.com/apple/xnu.git/blobdiff_plain/cf7d32b81c573a0536dc4da4157f9c26f8d0bed3..8a3053a07cee346dca737a5670e546fd26a7c9d6:/osfmk/kern/task.c?ds=sidebyside diff --git a/osfmk/kern/task.c b/osfmk/kern/task.c index b7dd90765..aa5d2b360 100644 --- a/osfmk/kern/task.c +++ b/osfmk/kern/task.c @@ -1,5 +1,5 @@ /* - * Copyright (c) 2000-2008 Apple Inc. All rights reserved. + * Copyright (c) 2000-2010 Apple Inc. All rights reserved. * * @APPLE_OSREFERENCE_LICENSE_HEADER_START@ * @@ -86,7 +86,6 @@ * Copyright (c) 2005 SPARTA, Inc. */ -#include #include #include @@ -102,6 +101,7 @@ #include #include #include +#include #include #include @@ -113,29 +113,25 @@ #include #include /* for thread_wakeup */ #include -#include #include #include #include #include #include #include +#include +#if CONFIG_TELEMETRY +#include +#endif #include #include #include /* for kernel_map, ipc_kernel_map */ #include #include +#include -#if MACH_KDB -#include -#endif /* MACH_KDB */ - -#ifdef __ppc__ -#include -#include -#endif - +#include /* * Exported interfaces */ @@ -152,15 +148,86 @@ #include #endif -task_t kernel_task; -zone_t task_zone; +#if CONFIG_COUNTERS +#include +#endif /* CONFIG_COUNTERS */ + +#include +#include + +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; + +/* Flag set by core audio when audio is playing. Used to stifle EXC_RESOURCE generation when active. */ +int audio_active = 0; + +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_t dead_task_statistics_lock; + +static ledger_template_t task_ledger_template = NULL; +struct _task_ledger_indices task_ledgers __attribute__((used)) = {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1}; +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 __attribute__((noinline)) THIS_PROCESS_IS_CAUSING_TOO_MANY_WAKEUPS__SENDING_EXC_RESOURCE(void); +void __attribute__((noinline)) THIS_PROCESS_CROSSED_HIGH_WATERMARK__SENDING_EXC_RESOURCE(int max_footprint_mb); +int coredump(void *core_proc, int reserve_mb, int ignore_ulimit); + +kern_return_t task_suspend_internal(task_t); +kern_return_t task_resume_internal(task_t); + +void proc_init_cpumon_params(void); + +// 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 */ + +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. */ + +int max_task_footprint = 0; /* Per-task limit on physical memory consumption */ +int task_max = CONFIG_TASK_MAX; /* Max number of tasks */ + +int hwm_user_cores = 0; /* high watermark violations generate user core files */ + +#ifdef MACH_BSD +extern void proc_getexecutableuuid(void *, unsigned char *, unsigned long); +extern int proc_pid(struct proc *p); +extern int proc_selfpid(void); +extern char *proc_name_address(struct proc *p); +#if CONFIG_JETSAM +extern void memorystatus_on_ledger_footprint_exceeded(int warning, const int max_footprint_mb); +#endif +#endif /* Forwards */ void task_hold_locked( task_t task); void task_wait_locked( - task_t task); + task_t task, + boolean_t until_not_runnable); void task_release_locked( task_t task); void task_free( @@ -168,10 +235,8 @@ void task_free( void task_synchronizer_destroy_all( task_t task); -kern_return_t task_set_ledger( - task_t task, - ledger_t wired, - ledger_t paged); +int check_for_tasksuspend( + task_t task); void task_backing_store_privileged( @@ -189,47 +254,19 @@ task_set_64bit( task_t task, boolean_t is64bit) { -#ifdef __i386__ +#if defined(__i386__) || defined(__x86_64__) thread_t thread; -#endif /* __i386__ */ - int vm_flags = 0; +#endif /* defined(__i386__) || defined(__x86_64__) */ + + task_lock(task); if (is64bit) { if (task_has_64BitAddr(task)) - return; - + goto out; task_set_64BitAddr(task); } else { if ( !task_has_64BitAddr(task)) - return; - - /* - * 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__ - /* LP64todo - make this clean */ - /* - * 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); + goto out; task_clear_64BitAddr(task); } /* FIXME: On x86, the thread save state flavor can diverge from the @@ -238,30 +275,179 @@ task_set_64bit( * certain routines may observe the thread as being in an inconsistent * state with respect to its task's 64-bitness. */ -#ifdef __i386__ + +#if defined(__i386__) || defined(__x86_64__) queue_iterate(&task->threads, thread, thread_t, task_threads) { + thread_mtx_lock(thread); machine_thread_switch_addrmode(thread); + thread_mtx_unlock(thread); } -#endif /* __i386__ */ +#endif /* defined(__i386__) || defined(__x86_64__) */ + +out: + task_unlock(task); +} + + +void +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; + task->all_image_info_size = size; + task_unlock(task); +} + +#if TASK_REFERENCE_LEAK_DEBUG +#include + +decl_simple_lock_data(static,task_ref_lock); +static btlog_t *task_ref_btlog; +#define TASK_REF_OP_INCR 0x1 +#define TASK_REF_OP_DECR 0x2 + +#define TASK_REF_BTDEPTH 7 + +static void +task_ref_lock_lock(void *context) +{ + simple_lock((simple_lock_t)context); +} +static void +task_ref_lock_unlock(void *context) +{ + simple_unlock((simple_lock_t)context); +} + +void +task_reference_internal(task_t task) +{ + void * bt[TASK_REF_BTDEPTH]; + int numsaved = 0; + + numsaved = OSBacktrace(bt, TASK_REF_BTDEPTH); + + (void)hw_atomic_add(&(task)->ref_count, 1); + btlog_add_entry(task_ref_btlog, task, TASK_REF_OP_INCR, + bt, numsaved); } +uint32_t +task_deallocate_internal(task_t task) +{ + 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 hw_atomic_sub(&(task)->ref_count, 1); +} + +#endif /* TASK_REFERENCE_LEAK_DEBUG */ + void task_init(void) { + + 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_max * sizeof(struct task), TASK_CHUNK * sizeof(struct task), "tasks"); + zone_change(task_zone, Z_NOENCRYPT, TRUE); + + /* + * Configure per-task memory limit. The boot arg takes precedence over the + * device tree. + */ + if (!PE_parse_boot_argn("max_task_pmem", &max_task_footprint, + sizeof (max_task_footprint))) { + max_task_footprint = 0; + } + + if (max_task_footprint == 0) { + /* + * 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, + sizeof(max_task_footprint))) { + max_task_footprint = 0; + } + } + + if (max_task_footprint != 0) { +#if CONFIG_JETSAM + if (max_task_footprint < 50) { + printf("Warning: max_task_pmem %d below minimum.\n", + max_task_footprint); + max_task_footprint = 50; + } + printf("Limiting task physical memory footprint to %d MB\n", + max_task_footprint); + max_task_footprint *= 1024 * 1024; // Convert MB to bytes +#else + printf("Warning: max_task_footprint specified, but jetsam not configured; ignoring.\n"); +#endif + } + + if (!PE_parse_boot_argn("hwm_user_cores", &hwm_user_cores, + sizeof (hwm_user_cores))) { + hwm_user_cores = 0; + } + + 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; + } + + init_task_ledgers(); + +#if TASK_REFERENCE_LEAK_DEBUG + simple_lock_init(&task_ref_lock, 0); + task_ref_btlog = btlog_create(100000, + TASK_REF_BTDEPTH, + task_ref_lock_lock, + task_ref_lock_unlock, + &task_ref_lock); + assert(task_ref_btlog); +#endif + /* * 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; + lck_spin_init(&dead_task_statistics_lock, &task_lck_grp, &task_lck_attr); } /* @@ -320,6 +506,76 @@ host_security_create_task_token( return(KERN_FAILURE); } +/* + * Task ledgers + * ------------ + * + * phys_footprint + * Physical footprint: This is the sum of: + * + phys_mem [task's resident memory] + * + phys_compressed + * + iokit_mem + * + * iokit_mem + * IOKit mappings: The total size of all IOKit mappings in this task [regardless of clean/dirty state]. + * + * phys_compressed + * Physical compressed: Amount of this task's resident 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. + */ +void +init_task_ledgers(void) +{ + ledger_template_t t; + + assert(task_ledger_template == NULL); + assert(kernel_task == TASK_NULL); + + 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.iokit_mem = ledger_entry_add(t, "iokit_mem", "mappings", + "bytes"); + task_ledgers.phys_footprint = ledger_entry_add(t, "phys_footprint", "physmem", + "bytes"); + task_ledgers.phys_compressed = ledger_entry_add(t, "phys_compressed", "physmem", + "bytes"); + 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 ((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.iokit_mem < 0) || + (task_ledgers.phys_footprint < 0) || (task_ledgers.phys_compressed < 0) || + (task_ledgers.platform_idle_wakeups < 0) || (task_ledgers.interrupt_wakeups < 0)) { + panic("couldn't create entries for task ledger template"); + } + + ledger_track_maximum(t, task_ledgers.phys_footprint, 60); + +#if CONFIG_JETSAM + ledger_set_callback(t, task_ledgers.phys_footprint, task_footprint_exceeded, NULL, NULL); +#endif + + ledger_set_callback(t, task_ledgers.interrupt_wakeups, + task_wakeups_rate_exceeded, NULL, NULL); + + task_ledger_template = t; +} + kern_return_t task_create_internal( task_t parent_task, @@ -329,6 +585,7 @@ task_create_internal( { task_t new_task; vm_shared_region_t shared_region; + ledger_t ledger = NULL; new_task = (task_t) zalloc(task_zone); @@ -338,26 +595,37 @@ task_create_internal( /* one ref for just being alive; one for our caller */ new_task->ref_count = 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); + } + + new_task->ledger = ledger; + + /* if inherit_memory is true, parent_task MUST not be NULL */ if (inherit_memory) - new_task->map = vm_map_fork(parent_task->map); + new_task->map = vm_map_fork(ledger, 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); + new_task->map = vm_map_create(pmap_create(ledger, 0, is_64bit), + (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, parent_task->map->user_wire_limit); + vm_map_set_user_wire_limit(new_task->map, (vm_size_t)parent_task->map->user_wire_limit); - mutex_init(&new_task->lock, 0); + 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->pset_hint = PROCESSOR_SET_NULL; - new_task->role = TASK_UNSPECIFIED; + new_task->legacy_stop_count = 0; new_task->active = TRUE; + new_task->halting = FALSE; new_task->user_data = NULL; new_task->faults = 0; new_task->cow_faults = 0; @@ -368,28 +636,37 @@ task_create_internal( 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 */ + new_task->t_flags = 0; + new_task->importance = 0; + + zinfo_task_init(new_task); #ifdef MACH_BSD new_task->bsd_info = NULL; #endif /* MACH_BSD */ -#ifdef __i386__ - new_task->i386_ldt = 0; +#if CONFIG_JETSAM + if (max_task_footprint != 0) { + ledger_set_limit(ledger, task_ledgers.phys_footprint, max_task_footprint, PHYS_FOOTPRINT_WARNING_LEVEL); + } #endif -#ifdef __ppc__ - if(BootProcInfo.pf.Available & pf64Bit) new_task->taskFeatures[0] |= tf64BitData; /* If 64-bit machine, show we have 64-bit registers at least */ + 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 defined(__i386__) || defined(__x86_64__) + new_task->i386_ldt = 0; #endif + new_task->task_debug = NULL; + queue_init(&new_task->semaphore_list); - queue_init(&new_task->lock_set_list); new_task->semaphores_owned = 0; - new_task->lock_sets_owned = 0; #if CONFIG_MACF_MACH - /*mutex_init(&new_task->labellock, ETAP_NO_TRACE);*/ new_task->label = labelh_new(1); mac_task_label_init (&new_task->maclabel); #endif @@ -405,6 +682,44 @@ task_create_internal( new_task->affinity_space = NULL; +#if CONFIG_COUNTERS + new_task->t_chud = 0U; +#endif + + 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 + + + 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->mem_notify_reserved = 0; +#if IMPORTANCE_INHERITANCE + new_task->imp_receiver = 0; + new_task->imp_donor = 0; + new_task->imp_reserved = 0; + new_task->task_imp_assertcnt = 0; + new_task->task_imp_externcnt = 0; +#endif /* IMPORTANCE_INHERITANCE */ + +#if defined(__x86_64__) + new_task->uexc_range_start = new_task->uexc_range_size = new_task->uexc_handler = 0; +#endif + + new_task->requested_policy = default_task_requested_policy; + new_task->effective_policy = default_task_effective_policy; + new_task->pended_policy = default_task_pended_policy; + if (parent_task != TASK_NULL) { new_task->sec_token = parent_task->sec_token; new_task->audit_token = parent_task->audit_token; @@ -413,44 +728,74 @@ task_create_internal( shared_region = vm_shared_region_get(parent_task); vm_shared_region_set(new_task, shared_region); - 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; -#ifdef __i386__ +#if defined(__i386__) || defined(__x86_64__) if (inherit_memory && parent_task->i386_ldt) new_task->i386_ldt = user_ldt_copy(parent_task->i386_ldt); #endif if (inherit_memory && parent_task->affinity_space) task_affinity_create(parent_task, new_task); - } - else { + + new_task->pset_hint = parent_task->pset_hint = task_choose_pset(parent_task); + +#if IMPORTANCE_INHERITANCE + new_task->imp_donor = parent_task->imp_donor; + /* Embedded doesn't want this to inherit */ + new_task->imp_receiver = parent_task->imp_receiver; +#endif /* IMPORTANCE_INHERITANCE */ + + new_task->requested_policy.t_apptype = parent_task->requested_policy.t_apptype; + + new_task->requested_policy.int_darwinbg = parent_task->requested_policy.int_darwinbg; + new_task->requested_policy.ext_darwinbg = parent_task->requested_policy.ext_darwinbg; + new_task->requested_policy.int_iotier = parent_task->requested_policy.int_iotier; + new_task->requested_policy.ext_iotier = parent_task->requested_policy.ext_iotier; + new_task->requested_policy.int_iopassive = parent_task->requested_policy.int_iopassive; + new_task->requested_policy.ext_iopassive = parent_task->requested_policy.ext_iopassive; + new_task->requested_policy.bg_iotier = parent_task->requested_policy.bg_iotier; + new_task->requested_policy.terminated = parent_task->requested_policy.terminated; + + task_policy_create(new_task, parent_task->requested_policy.t_boosted); + } 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 + 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 { + } else if (proc_get_effective_task_policy(new_task, TASK_POLICY_LOWPRI_CPU)) { + new_task->priority = MAXPRI_THROTTLE; + new_task->max_priority = MAXPRI_THROTTLE; + } else { new_task->priority = BASEPRI_DEFAULT; new_task->max_priority = MAXPRI_USER; } - - mutex_lock(&tasks_threads_lock); + + bzero(&new_task->extmod_statistics, sizeof(new_task->extmod_statistics)); + new_task->task_timer_wakeups_bin_1 = new_task->task_timer_wakeups_bin_2 = 0; + lck_mtx_lock(&tasks_threads_lock); queue_enter(&tasks, new_task, task_t, tasks); tasks_count++; - mutex_unlock(&tasks_threads_lock); + lck_mtx_unlock(&tasks_threads_lock); if (vm_backing_store_low && parent_task != NULL) new_task->priv_flags |= (parent_task->priv_flags&VM_BACKING_STORE_PRIV); + new_task->task_volatile_objects = 0; + ipc_task_enable(new_task); *child_task = new_task; @@ -466,12 +811,26 @@ void task_deallocate( task_t task) { + ledger_amount_t credit, debit, interrupt_wakeups, platform_idle_wakeups; + if (task == TASK_NULL) return; if (task_deallocate_internal(task) > 0) return; + lck_mtx_lock(&tasks_threads_lock); + queue_remove(&terminated_tasks, task, task_t, tasks); + terminated_tasks_count--; + lck_mtx_unlock(&tasks_threads_lock); + + /* + * Give the machine dependent code a chance + * to perform cleanup before ripping apart + * the task. + */ + machine_task_terminate(task); + ipc_task_terminate(task); if (task->affinity_space) @@ -480,9 +839,56 @@ task_deallocate( vm_map_deallocate(task->map); is_release(task->itk_space); + 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); + + /* 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; + + lck_spin_unlock(&dead_task_statistics_lock); + lck_mtx_destroy(&task->lock, &task_lck_grp); + #if CONFIG_MACF_MACH labelh_release(task->label); #endif + + 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); + zinfo_task_free(task); + +#if TASK_REFERENCE_LEAK_DEBUG + btlog_remove_entries_for_element(task_ref_btlog, task); +#endif + + if (task->task_volatile_objects) { + /* + * This task still "owns" some volatile VM objects. + * Disown them now to avoid leaving them pointing back at + * an invalid task. + */ + vm_purgeable_disown(task); + assert(task->task_volatile_objects == 0); + } + zfree(task_zone, task); } @@ -498,6 +904,17 @@ task_name_deallocate( return(task_deallocate((task_t)task_name)); } +/* + * 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_terminate: @@ -548,9 +965,9 @@ task_terminate_internal( task_lock(task); } - if (!task->active || !self->active) { + if (!task->active) { /* - * Task or current act is already being terminated. + * 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. @@ -562,6 +979,14 @@ task_terminate_internal( return (KERN_FAILURE); } +#if MACH_ASSERT + if (task->suspends_outstanding != 0) { + printf("WARNING: %s (%d) exiting with %d outstanding suspensions\n", + proc_name_address(task->bsd_info), proc_pid(task->bsd_info), + task->suspends_outstanding); + } +#endif + if (self_task != task) task_unlock(self_task); @@ -583,6 +1008,13 @@ task_terminate_internal( 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. */ @@ -590,16 +1022,9 @@ task_terminate_internal( thread_terminate_internal(thread); } - /* - * Give the machine dependent code a chance - * to perform cleanup before ripping apart - * the task. - */ - if (self_task == task) - machine_thread_terminate_self(); - task_unlock(task); + /* * Destroy all synchronizers owned by the task. */ @@ -608,15 +1033,7 @@ task_terminate_internal( /* * Destroy the IPC space, leaving just a reference for it. */ - ipc_space_destroy(task->itk_space); - -#ifdef __ppc__ - /* LP64todo - make this clean */ - /* - * PPC51: ppc64 is limited to 51-bit addresses. - */ - pmap_unmap_sharedpage(task->map->pmap); /* Unmap commpage */ -#endif /* __ppc__ */ + ipc_space_terminate(task->itk_space); if (vm_map_has_4GB_pagezero(task->map)) vm_map_clear_4GB_pagezero(task->map); @@ -637,10 +1054,12 @@ task_terminate_internal( /* release our shared region */ vm_shared_region_set(task, NULL); - mutex_lock(&tasks_threads_lock); + lck_mtx_lock(&tasks_threads_lock); queue_remove(&tasks, task, task_t, tasks); + queue_enter(&terminated_tasks, task, task_t, tasks); tasks_count--; - mutex_unlock(&tasks_threads_lock); + terminated_tasks_count++; + lck_mtx_unlock(&tasks_threads_lock); /* * We no longer need to guard against being aborted, so restore @@ -648,10 +1067,6 @@ task_terminate_internal( */ thread_interrupt_level(interrupt_save); -#if __ppc__ - perfmon_release_facility(task); // notify the perfmon facility -#endif - /* * Get rid of the task active reference on itself. */ @@ -661,16 +1076,15 @@ task_terminate_internal( } /* - * task_halt: + * 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, terminate all other threads in the task and - * wait for them to terminate, clean up the portspace, and when - * all done, let the current thread go. + * We hold the task and mark all other threads in the task for + * termination. */ kern_return_t -task_halt( +task_start_halt( task_t task) { thread_t thread, self; @@ -684,7 +1098,7 @@ task_halt( task_lock(task); - if (!task->active || !self->active) { + 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 @@ -696,7 +1110,10 @@ task_halt( return (KERN_FAILURE); } + task->halting = TRUE; + if (task->thread_count > 1) { + /* * Mark all the threads to keep them from starting any more * user-level execution. The thread_terminate_internal code @@ -715,15 +1132,45 @@ task_halt( task_release_locked(task); } + task_unlock(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. + */ +void +task_complete_halt(task_t task) +{ + task_lock(task); + assert(task->halting); + assert(task == current_task()); /* - * Give the machine dependent code a chance - * to perform cleanup before ripping apart - * the task. + * Wait for the other threads to get shut down. + * When the last other thread is reaped, we'll be + * woken up. */ - machine_thread_terminate_self(); + 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); + } - 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. @@ -743,7 +1190,7 @@ task_halt( vm_map_remove(task->map, task->map->min_offset, task->map->max_offset, VM_MAP_NO_FLAGS); - return (KERN_SUCCESS); + task->halting = FALSE; } /* @@ -808,6 +1255,28 @@ task_hold( 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: * @@ -818,7 +1287,8 @@ task_hold( */ void task_wait_locked( - register task_t task) + register task_t task, + boolean_t until_not_runnable) { register thread_t thread, self; @@ -834,7 +1304,7 @@ task_wait_locked( */ queue_iterate(&task->threads, thread, thread_t, task_threads) { if (thread != self) - thread_wait(thread); + thread_wait(thread, until_not_runnable); } } @@ -1002,36 +1472,37 @@ task_threads( return (KERN_SUCCESS); } -/* - * task_suspend: - * - * Implement a user-level suspension on a task. - * - * Conditions: - * The caller holds a reference to the task - */ -kern_return_t -task_suspend( - register task_t task) -{ - if (task == TASK_NULL || task == kernel_task) - return (KERN_INVALID_ARGUMENT); - - task_lock(task); +#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 ( + register task_t task, + int mode) +{ if (!task->active) { - task_unlock(task); - return (KERN_FAILURE); } + KERNEL_DEBUG_CONSTANT_IST(KDEBUG_TRACE, + MACHDBG_CODE(DBG_MACH_IPC,MACH_TASK_SUSPEND) | DBG_FUNC_NONE, + proc_pid(task->bsd_info), ((thread_t)queue_first(&task->threads))->thread_id, + task->user_stop_count, task->user_stop_count + 1, 0); + +#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. */ - task_unlock(task); - return (KERN_SUCCESS); } @@ -1042,16 +1513,159 @@ task_suspend( * to stop executing user code. */ task_hold_locked(task); - task_wait_locked(task); + task_wait_locked(task, FALSE); + + return (KERN_SUCCESS); +} + +static kern_return_t +release_task_hold ( + register task_t task, + int mode) +{ + register boolean_t release = FALSE; + + if (!task->active) { + return (KERN_FAILURE); + } + + 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, + proc_pid(task->bsd_info), ((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); +} + + +/* + * 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( + register task_t task) +{ + kern_return_t kr; + mach_port_t port, send, old_notify; + mach_port_name_t name; + + if (task == TASK_NULL || task == kernel_task) + return (KERN_INVALID_ARGUMENT); + + task_lock(task); + + /* + * Claim a send right on the task resume port, and request a no-senders + * notification on that port (if none outstanding). + */ + if (task->itk_resume == IP_NULL) { + task->itk_resume = ipc_port_alloc_kernel(); + if (!IP_VALID(task->itk_resume)) + panic("failed to create resume port"); + ipc_kobject_set(task->itk_resume, (ipc_kobject_t)task, IKOT_TASK_RESUME); + } + + port = task->itk_resume; + ip_lock(port); + assert(ip_active(port)); + + send = ipc_port_make_send_locked(port); + assert(IP_VALID(send)); + + if (port->ip_nsrequest == IP_NULL) { + 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); + } + + /* + * place a legacy hold on the task. + */ + kr = place_task_hold(task, TASK_HOLD_LEGACY); + if (kr != KERN_SUCCESS) { + task_unlock(task); + ipc_port_release_send(send); + return kr; + } task_unlock(task); - return (KERN_SUCCESS); + /* + * 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 ((kr = ipc_kmsg_copyout_object(current_task()->itk_space, (ipc_object_t)send, + MACH_MSG_TYPE_MOVE_SEND, &name)) != KERN_SUCCESS) { + printf("warning: %s(%d) failed to copyout suspension token for task %s(%d) with error: %d\n", + proc_name_address(current_task()->bsd_info), proc_pid(current_task()->bsd_info), + proc_name_address(task->bsd_info), proc_pid(task->bsd_info), kr); + return (kr); + } + + return (kr); } /* * task_resume: - * Release a kernel hold on a task. + * Release a user hold on a task. * * Conditions: * The caller holds a reference to the task @@ -1060,40 +1674,413 @@ kern_return_t task_resume( register task_t task) { - register boolean_t release = FALSE; + 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, (ipc_object_t)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 %s(%d)\n", + proc_name_address(current_task()->bsd_info), proc_pid(current_task()->bsd_info), + proc_name_address(task->bsd_info), proc_pid(task->bsd_info)); + } + + 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); +} - if (!task->active) { +/* + * Suspend the target task, and return a suspension token. The token + * represents a reference on the suspended task. + */ +kern_return_t +task_suspend2( + register 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( + register 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( + register 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 = (ipc_port_t) 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; +} + +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( + register 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); + + return (kr); +} + +/* If enabled, we bring all the frozen pages back in prior to resumption; otherwise, they're faulted back in on demand */ +#define THAW_ON_RESUME 1 + +/* + * task_pidresume: + * Resumes a previously suspended task. + * + * Conditions: + * The caller holds a reference to the task + */ +kern_return_t +task_pidresume( + register task_t task) +{ + kern_return_t kr; + + if (task == TASK_NULL || task == kernel_task) + return (KERN_INVALID_ARGUMENT); + + task_lock(task); + +#if (CONFIG_FREEZE && THAW_ON_RESUME) + + 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 (CONFIG_FREEZE && THAW_ON_RESUME) + if ((kr == KERN_SUCCESS) && (task->frozen == TRUE)) { + + if (COMPRESSED_PAGER_IS_ACTIVE || DEFAULT_FREEZER_COMPRESSED_PAGER_IS_ACTIVE) { + + kr = KERN_SUCCESS; + } else { + + kr = vm_map_thaw(task->map); + } + } + 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); +} + +#if CONFIG_FREEZE + +/* + * task_freeze: + * + * Freeze a task. + * + * Conditions: + * The caller holds a reference to the task + */ +kern_return_t +task_freeze( + register task_t task, + uint32_t *purgeable_count, + uint32_t *wired_count, + uint32_t *clean_count, + uint32_t *dirty_count, + uint32_t dirty_budget, + boolean_t *shared, + boolean_t walk_only) +{ + kern_return_t kr; + + 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; + + task_unlock(task); - if (task->user_stop_count > 0) { - if (--task->user_stop_count == 0) - release = TRUE; + if (walk_only) { + kr = vm_map_freeze_walk(task->map, purgeable_count, wired_count, clean_count, dirty_count, dirty_budget, shared); + } else { + kr = vm_map_freeze(task->map, purgeable_count, wired_count, clean_count, dirty_count, dirty_budget, shared); } - else { + + task_lock(task); + + if (walk_only == FALSE && kr == KERN_SUCCESS) + task->frozen = TRUE; + task->changing_freeze_state = FALSE; + thread_wakeup(&task->changing_freeze_state); + + task_unlock(task); + + return (kr); +} + +/* + * task_thaw: + * + * Thaw a currently frozen task. + * + * Conditions: + * The caller holds a reference to the task + */ +extern void +vm_consider_waking_compactor_swapper(void); + +kern_return_t +task_thaw( + register task_t task) +{ + kern_return_t kr; + + 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; - /* - * Release the task if necessary. - */ - if (release) - task_release_locked(task); + if (DEFAULT_PAGER_IS_ACTIVE || DEFAULT_FREEZER_IS_ACTIVE) { + task_unlock(task); + kr = vm_map_thaw(task->map); + + task_lock(task); + + if (kr == KERN_SUCCESS) + task->frozen = FALSE; + } else { + task->frozen = FALSE; + kr = KERN_SUCCESS; + } + + task->changing_freeze_state = FALSE; + thread_wakeup(&task->changing_freeze_state); + task_unlock(task); - return (KERN_SUCCESS); + if (COMPRESSED_PAGER_IS_ACTIVE || DEFAULT_FREEZER_COMPRESSED_PAGER_IS_ACTIVE) { + vm_consider_waking_compactor_swapper(); + } + + return (kr); } +#endif /* CONFIG_FREEZE */ + kern_return_t host_security_set_task_token( host_security_t host_security, @@ -1114,7 +2101,8 @@ host_security_set_task_token( task_lock(task); task->sec_token = sec_token; task->audit_token = audit_token; - task_unlock(task); + + task_unlock(task); if (host_priv != HOST_PRIV_NULL) { kr = host_get_host_priv_port(host_priv, &host_port); @@ -1126,32 +2114,6 @@ host_security_set_task_token( return(kr); } -/* - * Utility routine to set a ledger - */ -kern_return_t -task_set_ledger( - task_t task, - ledger_t wired, - ledger_t paged) -{ - if (task == TASK_NULL) - return(KERN_INVALID_ARGUMENT); - - 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); - - return(KERN_SUCCESS); -} - /* * This routine was added, pretty much exclusively, for registering the * RPC glue vector for in-kernel short circuited tasks. Rather than @@ -1179,29 +2141,42 @@ task_set_info( kern_return_t task_info( - task_t task, - task_flavor_t flavor, - task_info_t task_info_out, + 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; + if (task == TASK_NULL) return (KERN_INVALID_ARGUMENT); + 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: { task_basic_info_32_t basic_info; - vm_map_t map; + vm_map_t map; + clock_sec_t secs; + clock_usec_t usecs; - if (*task_info_count < TASK_BASIC_INFO_32_COUNT) - return (KERN_INVALID_ARGUMENT); + 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 = CAST_DOWN(vm_offset_t,map->size); + basic_info->virtual_size = (typeof(basic_info->virtual_size))map->size; if (flavor == TASK_BASIC2_INFO_32) { /* * The "BASIC2" flavor gets the maximum resident @@ -1213,18 +2188,19 @@ task_info( } basic_info->resident_size *= PAGE_SIZE; - task_lock(task); 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, - (unsigned *)&basic_info->user_time.seconds, - (unsigned *)&basic_info->user_time.microseconds); - absolutetime_to_microtime(task->total_system_time, - (unsigned *)&basic_info->system_time.seconds, - (unsigned *)&basic_info->system_time.microseconds); - task_unlock(task); + 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; @@ -1233,10 +2209,14 @@ task_info( case TASK_BASIC_INFO_64: { task_basic_info_64_t basic_info; - vm_map_t map; + vm_map_t map; + clock_sec_t secs; + clock_usec_t usecs; - if (*task_info_count < TASK_BASIC_INFO_64_COUNT) - return (KERN_INVALID_ARGUMENT); + if (*task_info_count < TASK_BASIC_INFO_64_COUNT) { + error = KERN_INVALID_ARGUMENT; + break; + } basic_info = (task_basic_info_64_t)task_info_out; @@ -1246,30 +2226,78 @@ task_info( (mach_vm_size_t)(pmap_resident_count(map->pmap)) * PAGE_SIZE_64; - task_lock(task); 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, - (unsigned *)&basic_info->user_time.seconds, - (unsigned *)&basic_info->user_time.microseconds); - absolutetime_to_microtime(task->total_system_time, - (unsigned *)&basic_info->system_time.seconds, - (unsigned *)&basic_info->system_time.microseconds); - task_unlock(task); + 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; } + 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 = map->size; + + 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: { register task_thread_times_info_t times_info; register thread_t thread; - if (*task_info_count < TASK_THREAD_TIMES_INFO_COUNT) - return (KERN_INVALID_ARGUMENT); + 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; @@ -1277,18 +2305,18 @@ task_info( times_info->system_time.seconds = 0; times_info->system_time.microseconds = 0; - task_lock(task); queue_iterate(&task->threads, thread, thread_t, task_threads) { - time_value_t user_time, system_time; + time_value_t user_time, system_time; - thread_read_times(thread, &user_time, &system_time); + if (thread->options & TH_OPT_IDLE_THREAD) + continue; - time_value_add(×_info->user_time, &user_time); - time_value_add(×_info->system_time, &system_time); - } + thread_read_times(thread, &user_time, &system_time); - task_unlock(task); + time_value_add(×_info->user_time, &user_time); + time_value_add(×_info->system_time, &system_time); + } *task_info_count = TASK_THREAD_TIMES_INFO_COUNT; break; @@ -1299,65 +2327,213 @@ task_info( task_absolutetime_info_t info; register thread_t thread; - if (*task_info_count < TASK_ABSOLUTETIME_INFO_COUNT) - return (KERN_INVALID_ARGUMENT); + 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; - task_lock(task); 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); - info->threads_system += tval; - info->total_system += tval; + 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_unlock(task); *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_64BitAddr(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) - return (KERN_INVALID_ARGUMENT); + if (*task_info_count < POLICY_FIFO_BASE_COUNT) { + error = KERN_INVALID_ARGUMENT; + break; + } - return (KERN_INVALID_POLICY); + error = KERN_INVALID_POLICY; + break; } /* OBSOLETE */ case TASK_SCHED_RR_INFO: { register policy_rr_base_t rr_base; + uint32_t quantum_time; + uint64_t quantum_ns; - if (*task_info_count < POLICY_RR_BASE_COUNT) - return (KERN_INVALID_ARGUMENT); + if (*task_info_count < POLICY_RR_BASE_COUNT) { + error = KERN_INVALID_ARGUMENT; + break; + } rr_base = (policy_rr_base_t) task_info_out; - task_lock(task); if (task != kernel_task) { - task_unlock(task); - return (KERN_INVALID_POLICY); + error = KERN_INVALID_POLICY; + break; } rr_base->base_priority = task->priority; - task_unlock(task); - rr_base->quantum = std_quantum_us / 1000; + 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; @@ -1368,19 +2544,19 @@ task_info( { register policy_timeshare_base_t ts_base; - if (*task_info_count < POLICY_TIMESHARE_BASE_COUNT) - return (KERN_INVALID_ARGUMENT); + if (*task_info_count < POLICY_TIMESHARE_BASE_COUNT) { + error = KERN_INVALID_ARGUMENT; + break; + } ts_base = (policy_timeshare_base_t) task_info_out; - task_lock(task); if (task == kernel_task) { - task_unlock(task); - return (KERN_INVALID_POLICY); + error = KERN_INVALID_POLICY; + break; } ts_base->base_priority = task->priority; - task_unlock(task); *task_info_count = POLICY_TIMESHARE_BASE_COUNT; break; @@ -1390,14 +2566,14 @@ task_info( { register security_token_t *sec_token_p; - if (*task_info_count < TASK_SECURITY_TOKEN_COUNT) - return (KERN_INVALID_ARGUMENT); + if (*task_info_count < TASK_SECURITY_TOKEN_COUNT) { + error = KERN_INVALID_ARGUMENT; + break; + } sec_token_p = (security_token_t *) task_info_out; - task_lock(task); *sec_token_p = task->sec_token; - task_unlock(task); *task_info_count = TASK_SECURITY_TOKEN_COUNT; break; @@ -1407,66 +2583,248 @@ task_info( { register audit_token_t *audit_token_p; - if (*task_info_count < TASK_AUDIT_TOKEN_COUNT) - return (KERN_INVALID_ARGUMENT); + if (*task_info_count < TASK_AUDIT_TOKEN_COUNT) { + error = KERN_INVALID_ARGUMENT; + break; + } audit_token_p = (audit_token_t *) task_info_out; - task_lock(task); *audit_token_p = task->audit_token; - task_unlock(task); *task_info_count = TASK_AUDIT_TOKEN_COUNT; break; } - - case TASK_SCHED_INFO: - return (KERN_INVALID_ARGUMENT); + + case TASK_SCHED_INFO: + error = KERN_INVALID_ARGUMENT; + break; + + case TASK_EVENTS_INFO: + { + register task_events_info_t events_info; + register 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 = 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; + + 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); + break; + } + + case TASK_VM_INFO: + case TASK_VM_INFO_PURGEABLE: + { + task_vm_info_t vm_info; + vm_map_t map; + + if (*task_info_count < TASK_VM_INFO_COUNT) { + error = KERN_INVALID_ARGUMENT; + break; + } + + vm_info = (task_vm_info_t)task_info_out; + + if (task == kernel_task) { + map = kernel_map; + /* no lock */ + } else { + map = task->map; + vm_map_lock_read(map); + } + + vm_info->virtual_size = (typeof(vm_info->virtual_size))map->size; + 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 (task == 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_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_pmap_size); + if (kr == KERN_SUCCESS) { + vm_info->purgeable_volatile_pmap = + volatile_pmap_size; + vm_info->purgeable_volatile_resident = + volatile_resident_size; + vm_info->purgeable_volatile_virtual = + volatile_virtual_size; + } + } + vm_map_unlock_read(map); + } + + *task_info_count = TASK_VM_INFO_COUNT; + break; + } + + default: + error = KERN_INVALID_ARGUMENT; + } - case TASK_EVENTS_INFO: - { - register task_events_info_t events_info; - register thread_t thread; + task_unlock(task); + return (error); +} - if (*task_info_count < TASK_EVENTS_INFO_COUNT) - return (KERN_INVALID_ARGUMENT); +/* + * 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) +{ + thread_t thread; + ledger_amount_t tmp; - events_info = (task_events_info_t) task_info_out; + task_lock_assert_owned(task); - task_lock(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); - 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; + info->task_timer_wakeups_bin_1 = task->task_timer_wakeups_bin_1; + info->task_timer_wakeups_bin_2 = task->task_timer_wakeups_bin_2; - events_info->csw = task->c_switch; + info->total_user = task->total_user_time; + info->total_system = task->total_system_time; - queue_iterate(&task->threads, thread, thread_t, task_threads) { - events_info->csw += thread->c_switch; - } + queue_iterate(&task->threads, thread, thread_t, task_threads) { + uint64_t tval; + spl_t x; - task_unlock(task); + if (thread->options & TH_OPT_IDLE_THREAD) + continue; - *task_info_count = TASK_EVENTS_INFO_COUNT; - break; - } - case TASK_AFFINITY_TAG_INFO: - { - if (*task_info_count < TASK_AFFINITY_TAG_INFO_COUNT) - return (KERN_INVALID_ARGUMENT); + x = splsched(); + thread_lock(thread); - return task_affinity_info(task, task_info_out, task_info_count); - } + info->task_timer_wakeups_bin_1 += thread->thread_timer_wakeups_bin_1; + info->task_timer_wakeups_bin_2 += thread->thread_timer_wakeups_bin_2; - default: - return (KERN_INVALID_ARGUMENT); + tval = timer_grab(&thread->user_timer); + info->total_user += tval; + + 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; + } + + thread_unlock(thread); + splx(x); } +} - return (KERN_SUCCESS); +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 @@ -1475,6 +2833,7 @@ task_vtimer_set( integer_t which) { thread_t thread; + spl_t x; /* assert(task == current_task()); */ /* bogus assert 4803227 4807483 */ @@ -1486,21 +2845,36 @@ task_vtimer_set( case TASK_VTIMER_USER: queue_iterate(&task->threads, thread, thread_t, task_threads) { - thread->vtimer_user_save = timer_grab(&thread->user_timer); + 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; } @@ -1530,38 +2904,48 @@ __unused uint32_t *microsecs) { thread_t thread = current_thread(); - uint32_t tdelt, secs; + uint32_t tdelt; + clock_sec_t secs; uint64_t tsum; assert(task == current_task()); assert(task->vtimers & which); - tdelt = secs = 0; + secs = tdelt = 0; switch (which) { case TASK_VTIMER_USER: - tdelt = timer_delta(&thread->user_timer, + 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 = tsum - thread->vtimer_prof_save; - thread->vtimer_prof_save = tsum; + 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 = tsum - thread->vtimer_rlim_save; + tdelt = (uint32_t)(tsum - thread->vtimer_rlim_save); thread->vtimer_rlim_save = tsum; + absolutetime_to_microtime(tdelt, &secs, microsecs); break; } - absolutetime_to_microtime(tdelt, &secs, microsecs); } /* @@ -1685,7 +3069,6 @@ void task_synchronizer_destroy_all(task_t task) { semaphore_t semaphore; - lock_set_t lock_set; /* * Destroy owned semaphores @@ -1695,16 +3078,252 @@ task_synchronizer_destroy_all(task_t task) semaphore = (semaphore_t) queue_first(&task->semaphore_list); (void) semaphore_destroy(task, semaphore); } +} + +/* + * 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); + } + + task_lock(task); + + if (!task->active) { + task_unlock(task); + return (KERN_FAILURE); + } + + ret = machine_task_set_state(task, flavor, state, state_count); + + 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) +{ + kern_return_t ret; + + if (task == TASK_NULL) { + return (KERN_INVALID_ARGUMENT); + } + + 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; +} + +#if CONFIG_JETSAM +#define HWM_USERCORE_MINSPACE 250 // free space (in MB) required *after* core file creation + +void __attribute__((noinline)) +THIS_PROCESS_CROSSED_HIGH_WATERMARK__SENDING_EXC_RESOURCE(int max_footprint_mb) +{ + task_t task = current_task(); + int pid = 0; + char *procname = (char *) "unknown"; + mach_exception_data_type_t code[EXCEPTION_CODE_MAX]; + +#ifdef MACH_BSD + pid = proc_selfpid(); + if (task->bsd_info != NULL) + procname = proc_name_address(current_task()->bsd_info); +#endif + + if (hwm_user_cores) { + int error; + uint64_t starttime, end; + clock_sec_t secs = 0; + uint32_t microsecs = 0; + + starttime = mach_absolute_time(); + /* + * 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. + */ + if ((error = coredump(current_task()->bsd_info, HWM_USERCORE_MINSPACE, 1)) != 0) { + printf("couldn't take coredump of %s[%d]: %d\n", procname, pid, error); + } + /* + * coredump() leaves the task suspended. + */ + task_resume_internal(current_task()); + + end = mach_absolute_time(); + absolutetime_to_microtime(end - starttime, &secs, µsecs); + printf("coredump of %s[%d] taken in %d secs %d microsecs\n", + proc_name_address(current_task()->bsd_info), pid, (int)secs, microsecs); + } + + 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; + } + + printf("process %s[%d] crossed memory high watermark (%d MB); sending " + "EXC_RESOURCE.\n", procname, pid, max_footprint_mb); + + 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); + exception_triage(EXC_RESOURCE, code, EXCEPTION_CODE_MAX); +} + +/* + * Callback invoked when a task exceeds its physical footprint limit. + */ +void +task_footprint_exceeded(int warning, __unused const void *param0, __unused const void *param1) +{ + ledger_amount_t max_footprint_mb; + + if (warning == LEDGER_WARNING_DIPPED_BELOW) { + /* + * Task memory limits only provide a warning on the way up. + */ + return; + } + + ledger_get_limit(current_task()->ledger, task_ledgers.phys_footprint, &max_footprint_mb); + max_footprint_mb >>= 20; /* - * Destroy owned lock sets + * If this an actual violation (not a warning), + * generate a non-fatal high watermark EXC_RESOURCE. */ + if ((warning == 0) && (current_task()->rusage_cpu_flags & TASK_RUSECPU_FLAGS_PHYS_FOOTPRINT_EXCEPTION)) { + THIS_PROCESS_CROSSED_HIGH_WATERMARK__SENDING_EXC_RESOURCE((int)max_footprint_mb); + } + + memorystatus_on_ledger_footprint_exceeded((warning == LEDGER_WARNING_ROSE_ABOVE) ? TRUE : FALSE, + (int)max_footprint_mb); +} + +extern int proc_check_footprint_priv(void); + +kern_return_t +task_set_phys_footprint_limit( + task_t task, + int new_limit_mb, + int *old_limit_mb) +{ + kern_return_t error; + + if ((error = proc_check_footprint_priv())) { + return (KERN_NO_ACCESS); + } + + return task_set_phys_footprint_limit_internal(task, new_limit_mb, old_limit_mb, FALSE); +} + +kern_return_t +task_set_phys_footprint_limit_internal( + task_t task, + int new_limit_mb, + int *old_limit_mb, + boolean_t trigger_exception) +{ + ledger_amount_t old; + + ledger_get_limit(task->ledger, task_ledgers.phys_footprint, &old); + + if (old_limit_mb) { + *old_limit_mb = old >> 20; + } + + 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 ? PHYS_FOOTPRINT_WARNING_LEVEL : 0); + return (KERN_SUCCESS); + } + +#ifdef CONFIG_NOMONITORS + return (KERN_SUCCESS); +#endif /* CONFIG_NOMONITORS */ - 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 (trigger_exception) { + task->rusage_cpu_flags |= TASK_RUSECPU_FLAGS_PHYS_FOOTPRINT_EXCEPTION; + } else { + task->rusage_cpu_flags &= ~TASK_RUSECPU_FLAGS_PHYS_FOOTPRINT_EXCEPTION; } + + ledger_set_limit(task->ledger, task_ledgers.phys_footprint, + (ledger_amount_t)new_limit_mb << 20, PHYS_FOOTPRINT_WARNING_LEVEL); + + task_unlock(task); + + return (KERN_SUCCESS); +} + +kern_return_t +task_get_phys_footprint_limit( + task_t task, + int *limit_mb) +{ + ledger_amount_t limit; + + ledger_get_limit(task->ledger, task_ledgers.phys_footprint, &limit); + *limit_mb = limit >> 20; + + return (KERN_SUCCESS); +} +#else /* CONFIG_JETSAM */ +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); +} + +kern_return_t +task_get_phys_footprint_limit( + __unused task_t task, + __unused int *limit_mb) +{ + return (KERN_FAILURE); } +#endif /* CONFIG_JETSAM */ /* * We need to export some functions to other components that @@ -1719,6 +3338,16 @@ boolean_t is_kerneltask(task_t t) return (FALSE); } +int +check_for_tasksuspend(task_t 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) @@ -1736,6 +3365,24 @@ task_reference( task_reference_internal(task); } +/* + * This routine is called always with task lock held. + * And it returns a thread handle without reference as the caller + * operates on it under the task lock held. + */ +thread_t +task_findtid(task_t task, uint64_t tid) +{ + thread_t thread= THREAD_NULL; + + queue_iterate(&task->threads, thread, thread_t, task_threads) { + if (thread->thread_id == tid) + return(thread); + } + return(THREAD_NULL); +} + + #if CONFIG_MACF_MACH /* * Protect 2 task labels against modification by adding a reference on @@ -1796,3 +3443,184 @@ mac_task_get_label(struct task *task) return (&task->maclabel); } #endif + +/* + * Control the CPU usage monitor for a task. + */ +kern_return_t +task_cpu_usage_monitor_ctl(task_t task, uint32_t *flags) +{ + int error = KERN_SUCCESS; + + if (*flags & CPUMON_MAKE_FATAL) { + task->rusage_cpu_flags |= TASK_RUSECPU_FLAGS_FATAL_CPUMON; + } else { + error = KERN_INVALID_ARGUMENT; + } + + return error; +} + +/* + * Control the wakeups monitor for a task. + */ +kern_return_t +task_wakeups_monitor_ctl(task_t task, uint32_t *flags, int32_t *rate_hz) +{ + ledger_t ledger = task->ledger; + + task_lock(task); + if (*flags & WAKEMON_GET_PARAMS) { + ledger_amount_t limit; + uint64_t period; + + ledger_get_limit(ledger, task_ledgers.interrupt_wakeups, &limit); + ledger_get_period(ledger, task_ledgers.interrupt_wakeups, &period); + + 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 WAKEMON_GET_PARAMS is present in flags, all other flags are ignored. + */ + task_unlock(task); + return KERN_SUCCESS; + } + + if (*flags & WAKEMON_ENABLE) { + if (*flags & WAKEMON_SET_DEFAULTS) { + *rate_hz = task_wakeups_monitor_rate; + } + +#ifndef CONFIG_NOMONITORS + if (*flags & WAKEMON_MAKE_FATAL) { + task->rusage_cpu_flags |= TASK_RUSECPU_FLAGS_FATAL_WAKEUPSMON; + } +#endif /* CONFIG_NOMONITORS */ + + if (*rate_hz < 0) { + task_unlock(task); + return KERN_INVALID_ARGUMENT; + } + +#ifndef CONFIG_NOMONITORS + ledger_set_limit(ledger, task_ledgers.interrupt_wakeups, *rate_hz * task_wakeups_monitor_interval, + 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(current_task(), TF_WAKEMON_WARNING, 0); +#endif + ledger_disable_refill(ledger, task_ledgers.interrupt_wakeups); + ledger_disable_callback(ledger, task_ledgers.interrupt_wakeups); + } + + task_unlock(task); + return KERN_SUCCESS; +} + +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; + } + +#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 + + if (warning == 0) { + THIS_PROCESS_IS_CAUSING_TOO_MANY_WAKEUPS__SENDING_EXC_RESOURCE(); + } +} + +void __attribute__((noinline)) +THIS_PROCESS_IS_CAUSING_TOO_MANY_WAKEUPS__SENDING_EXC_RESOURCE(void) +{ + task_t task = current_task(); + int pid = 0; + char *procname = (char *) "unknown"; + uint64_t observed_wakeups_rate; + uint64_t permitted_wakeups_rate; + uint64_t observation_interval; + mach_exception_data_type_t code[EXCEPTION_CODE_MAX]; + struct ledger_entry_info lei; + +#ifdef MACH_BSD + pid = proc_selfpid(); + if (task->bsd_info != NULL) + procname = proc_name_address(current_task()->bsd_info); +#endif + + ledger_get_entry_info(task->ledger, task_ledgers.interrupt_wakeups, &lei); + + /* + * Disable the exception notification so we don't overwhelm + * the listener with an endless stream of redundant exceptions. + */ + uint32_t flags = WAKEMON_DISABLE; + task_wakeups_monitor_ctl(task, &flags, NULL); + + observed_wakeups_rate = (lei.lei_balance * (int64_t)NSEC_PER_SEC) / lei.lei_last_refill; + permitted_wakeups_rate = lei.lei_limit / task_wakeups_monitor_interval; + observation_interval = lei.lei_refill_period / NSEC_PER_SEC; + + 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) { + printf("process %s[%d] caught causing excessive wakeups. EXC_RESOURCE " + "supressed due to audio playback\n", procname, pid); + return; + } + printf("process %s[%d] caught causing excessive wakeups. Observed wakeups rate " + "(per sec): %lld; Maximum permitted wakeups rate (per sec): %lld; Observation " + "period: %lld seconds; Task lifetime number of wakeups: %lld\n", + procname, pid, observed_wakeups_rate, permitted_wakeups_rate, + observation_interval, lei.lei_credit); + + 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], task_wakeups_monitor_rate); + EXC_RESOURCE_CPUMONITOR_ENCODE_OBSERVATION_INTERVAL(code[0], observation_interval); + EXC_RESOURCE_CPUMONITOR_ENCODE_WAKEUPS_OBSERVED(code[1], lei.lei_balance * (int64_t)NSEC_PER_SEC / lei.lei_last_refill); + exception_triage(EXC_RESOURCE, code, EXCEPTION_CODE_MAX); + + if (task->rusage_cpu_flags & TASK_RUSECPU_FLAGS_FATAL_WAKEUPSMON) { + task_terminate_internal(task); + } +}