path: root/erts/emulator/beam/erl_process.c

/*
 * %CopyrightBegin%
 *
 * Copyright Ericsson AB 1996-2014. All Rights Reserved.
 *
 * The contents of this file are subject to the Erlang Public License,
 * Version 1.1, (the "License"); you may not use this file except in
 * compliance with the License. You should have received a copy of the
 * Erlang Public License along with this software. If not, it can be
 * retrieved online at http://www.erlang.org/.
 *
 * Software distributed under the License is distributed on an "AS IS"
 * basis, WITHOUT WARRANTY OF ANY KIND, either express or implied. See
 * the License for the specific language governing rights and limitations
 * under the License.
 *
 * %CopyrightEnd%
 */

#define ERL_PROCESS_C__

#ifdef HAVE_CONFIG_H
#  include "config.h"
#endif

#include <stddef.h> /* offsetof() */
#include "sys.h"
#include "erl_vm.h"
#include "global.h"
#include "erl_process.h"
#include "error.h"
#include "bif.h"
#include "erl_db.h"
#include "dist.h"
#include "beam_catches.h"
#include "erl_instrument.h"
#include "erl_threads.h"
#include "erl_binary.h"
#include "beam_bp.h"
#include "erl_cpu_topology.h"
#include "erl_thr_progress.h"
#include "erl_thr_queue.h"
#include "erl_async.h"
#include "dtrace-wrapper.h"
#include "erl_ptab.h"


#define ERTS_DELAYED_WAKEUP_INFINITY (~(Uint64) 0)
#define ERTS_DELAYED_WAKEUP_REDUCTIONS ((Uint64) CONTEXT_REDS/2)

#define ERTS_RUNQ_CHECK_BALANCE_REDS_PER_SCHED (2000*CONTEXT_REDS)
#define ERTS_RUNQ_CALL_CHECK_BALANCE_REDS \
  (ERTS_RUNQ_CHECK_BALANCE_REDS_PER_SCHED/2)

#define ERTS_PROC_MIN_CONTEXT_SWITCH_REDS_COST (CONTEXT_REDS/10)

#ifndef ERTS_SCHED_MIN_SPIN
#define ERTS_SCHED_SPIN_UNTIL_YIELD 100
#else
#define ERTS_SCHED_SPIN_UNTIL_YIELD 1
#endif

#define ERTS_SCHED_SYS_SLEEP_SPINCOUNT_VERY_LONG 40
#define ERTS_SCHED_AUX_WORK_SLEEP_SPINCOUNT_FACT_VERY_LONG 1000
#define ERTS_SCHED_SYS_SLEEP_SPINCOUNT_LONG 20
#define ERTS_SCHED_AUX_WORK_SLEEP_SPINCOUNT_FACT_LONG 1000
#define ERTS_SCHED_SYS_SLEEP_SPINCOUNT_MEDIUM 10
#define ERTS_SCHED_AUX_WORK_SLEEP_SPINCOUNT_FACT_MEDIUM 1000
#define ERTS_SCHED_SYS_SLEEP_SPINCOUNT_SHORT 10
#define ERTS_SCHED_AUX_WORK_SLEEP_SPINCOUNT_FACT_SHORT 0
#define ERTS_SCHED_SYS_SLEEP_SPINCOUNT_VERY_SHORT 5
#define ERTS_SCHED_AUX_WORK_SLEEP_SPINCOUNT_FACT_VERY_SHORT 0
#define ERTS_SCHED_SYS_SLEEP_SPINCOUNT_NONE 0
#define ERTS_SCHED_AUX_WORK_SLEEP_SPINCOUNT_FACT_NONE 0

#define ERTS_SCHED_TSE_SLEEP_SPINCOUNT_FACT 1000
#define ERTS_SCHED_SUSPEND_SLEEP_SPINCOUNT 0

#if 0 || defined(DEBUG)
#define ERTS_FAKE_SCHED_BIND_PRINT_SORTED_CPU_DATA
#endif

#if defined(DEBUG) && 0
#define HARDDEBUG
#else
#undef HARDDEBUG
#endif

#ifdef HARDDEBUG
#define HARDDEBUG_RUNQS
#endif

#ifdef HIPE
#include "hipe_mode_switch.h"	/* for hipe_init_process() */
#include "hipe_signal.h"	/* for hipe_thread_signal_init() */
#endif

#ifdef ERTS_ENABLE_LOCK_COUNT
#include "erl_lock_count.h"
#endif

#define MAX_BIT       (1 << PRIORITY_MAX)
#define HIGH_BIT      (1 << PRIORITY_HIGH)
#define NORMAL_BIT    (1 << PRIORITY_NORMAL)
#define LOW_BIT       (1 << PRIORITY_LOW)
#define PORT_BIT      (1 << ERTS_PORT_PRIO_LEVEL)

#define ERTS_EMPTY_RUNQ(RQ)					\
    ((ERTS_RUNQ_FLGS_GET_NOB((RQ)) & ERTS_RUNQ_FLGS_QMASK) == 0	\
     && (RQ)->misc.start == NULL)

#undef RUNQ_READ_RQ
#undef RUNQ_SET_RQ
#define RUNQ_READ_RQ(X) ((ErtsRunQueue *) erts_smp_atomic_read_nob((X)))
#define RUNQ_SET_RQ(X, RQ) erts_smp_atomic_set_nob((X), (erts_aint_t) (RQ))

#ifdef DEBUG
#  if defined(ARCH_64) && !HALFWORD_HEAP
#    define ERTS_DBG_SET_INVALID_RUNQP(RQP, N) \
    (RUNQ_SET_RQ((RQP), (0xdeadbeefdead0003LL | ((N) << 4)))
#  define ERTS_DBG_VERIFY_VALID_RUNQP(RQP) \
do { \
    ASSERT((RQP) != NULL); \
    ASSERT(((((Uint) (RQP)) & ((Uint) 0x3))) == ((Uint) 0)); \
    ASSERT((((Uint) (RQP)) & ~((Uint) 0xffff)) != ((Uint) 0xdeadbeefdead0000LL));\
} while (0)
#  else
#    define ERTS_DBG_SET_INVALID_RUNQP(RQP, N) \
    (RUNQ_SET_RQ((RQP), (0xdead0003 | ((N) << 4))))
#  define ERTS_DBG_VERIFY_VALID_RUNQP(RQP) \
do { \
    ASSERT((RQP) != NULL); \
    ASSERT(((((UWord) (RQP)) & ((UWord) 1))) == ((UWord) 0)); \
    ASSERT((((UWord) (RQP)) & ~((UWord) 0xffff)) != ((UWord) 0xdead0000)); \
} while (0)
#  endif
#else
#  define ERTS_DBG_SET_INVALID_RUNQP(RQP, N)
#  define ERTS_DBG_VERIFY_VALID_RUNQP(RQP)
#endif

#define ERTS_EMPTY_RUNQ_PORTS(RQ) \
    (RUNQ_READ_LEN(&(RQ)->ports.info.len) == 0 && (RQ)->misc.start == NULL)

const Process erts_invalid_process = {{ERTS_INVALID_PID}};

extern BeamInstr beam_apply[];
extern BeamInstr beam_exit[];
extern BeamInstr beam_continue_exit[];

int erts_sched_compact_load;
int erts_sched_balance_util = 0;
Uint erts_no_schedulers;
#ifdef ERTS_DIRTY_SCHEDULERS
Uint erts_no_dirty_cpu_schedulers;
Uint erts_no_dirty_io_schedulers;
#endif

#define ERTS_THR_PRGR_LATER_CLEANUP_OP_THRESHOLD_VERY_LAZY		(4*1024*1024)
#define ERTS_THR_PRGR_LATER_CLEANUP_OP_THRESHOLD_LAZY			(512*1024)
#define ERTS_THR_PRGR_LATER_CLEANUP_OP_THRESHOLD_MEDIUM			(64*1024)
#define ERTS_THR_PRGR_LATER_CLEANUP_OP_THRESHOLD_EAGER			(16*1024)
#define ERTS_THR_PRGR_LATER_CLEANUP_OP_THRESHOLD_VERY_EAGER		(1024)

static UWord thr_prgr_later_cleanup_op_threshold = ERTS_THR_PRGR_LATER_CLEANUP_OP_THRESHOLD_MEDIUM;

ErtsPTab erts_proc erts_align_attribute(ERTS_CACHE_LINE_SIZE);

int erts_sched_thread_suggested_stack_size = -1;

#ifdef ERTS_ENABLE_LOCK_CHECK
ErtsLcPSDLocks erts_psd_required_locks[ERTS_PSD_SIZE];
#endif

static struct {
    int aux_work;
    int tse;
    int sys_schedule;
} sched_busy_wait;

#ifdef ERTS_SMP
int erts_disable_proc_not_running_opt;

static ErtsAuxWorkData *aux_thread_aux_work_data;

#define ERTS_SCHDLR_SSPND_CHNG_WAITER		(((erts_aint32_t) 1) << 0)
#define ERTS_SCHDLR_SSPND_CHNG_MSB		(((erts_aint32_t) 1) << 1)
#define ERTS_SCHDLR_SSPND_CHNG_ONLN		(((erts_aint32_t) 1) << 2)

#ifndef DEBUG

#define ERTS_SCHDLR_SSPND_CHNG_SET(VAL, OLD_VAL) \
  erts_smp_atomic32_set_nob(&schdlr_sspnd.changing, (VAL))

#ifdef ERTS_DIRTY_SCHEDULERS
#define ERTS_SCHDLR_SSPND_DIRTY_CPU_CHNG_SET(VAL, OLD_VAL) \
  erts_smp_atomic32_set_nob(&schdlr_sspnd.dirty_cpu_changing, (VAL))
#define ERTS_SCHDLR_SSPND_DIRTY_IO_CHNG_SET(VAL, OLD_VAL) \
  erts_smp_atomic32_set_nob(&schdlr_sspnd.dirty_io_changing, (VAL))
#endif

#else

#define ERTS_SCHDLR_SSPND_CHNG_SET(VAL, OLD_VAL)			\
do {									\
    erts_aint32_t old_val__;						\
    old_val__ = erts_smp_atomic32_xchg_nob(&schdlr_sspnd.changing,     	\
					   (VAL));			\
    ASSERT(old_val__ == (OLD_VAL));					\
} while (0)

#ifdef ERTS_DIRTY_SCHEDULERS
#define ERTS_SCHDLR_SSPND_DIRTY_CPU_CHNG_SET(VAL, OLD_VAL)		\
do {									\
    erts_aint32_t old_val__;						\
    old_val__ = erts_smp_atomic32_xchg_nob(&schdlr_sspnd.dirty_cpu_changing, \
					   (VAL));			\
    ASSERT(old_val__ == (OLD_VAL));					\
} while (0)
#define ERTS_SCHDLR_SSPND_DIRTY_IO_CHNG_SET(VAL, OLD_VAL)		\
do {									\
    erts_aint32_t old_val__;						\
    old_val__ = erts_smp_atomic32_xchg_nob(&schdlr_sspnd.dirty_io_changing, \
					   (VAL));			\
    ASSERT(old_val__ == (OLD_VAL));					\
} while (0)
#endif

#endif


static struct {
    erts_smp_mtx_t mtx;
    erts_smp_cnd_t cnd;
    int online;
    int curr_online;
    int wait_curr_online;
#ifdef ERTS_DIRTY_SCHEDULERS
    int dirty_cpu_online;
    int dirty_cpu_curr_online;
    int dirty_cpu_wait_curr_online;
    int dirty_io_online;
    int dirty_io_curr_online;
    int dirty_io_wait_curr_online;
#endif
    erts_smp_atomic32_t changing;
    erts_smp_atomic32_t active;
#ifdef ERTS_DIRTY_SCHEDULERS
    erts_smp_atomic32_t dirty_cpu_changing;
    erts_smp_atomic32_t dirty_cpu_active;
    erts_smp_atomic32_t dirty_io_changing;
    erts_smp_atomic32_t dirty_io_active;
#endif
    struct {
	int ongoing;
	long wait_active;
#ifdef ERTS_DIRTY_SCHEDULERS
	long dirty_cpu_wait_active;
	long dirty_io_wait_active;
#endif
	ErtsProcList *procs;
    } msb; /* Multi Scheduling Block */
} schdlr_sspnd;

static struct {
    erts_smp_mtx_t update_mtx;
    erts_smp_atomic32_t no_runqs;
    int last_active_runqs;
    int forced_check_balance;
    erts_smp_atomic32_t checking_balance;
    int halftime;
    int full_reds_history_index;
    struct {
	int active_runqs;
	int reds;
	erts_aint32_t max_len;
    } prev_rise;
    Uint n;
} balance_info;

#define ERTS_BLNCE_SAVE_RISE(ACTIVE, MAX_LEN, REDS)	\
do {							\
    balance_info.prev_rise.active_runqs = (ACTIVE);	\
    balance_info.prev_rise.max_len = (MAX_LEN);		\
    balance_info.prev_rise.reds = (REDS);		\
} while (0)

#endif

erts_sched_stat_t erts_sched_stat;

#ifdef USE_THREADS
static erts_tsd_key_t sched_data_key;
#endif

static erts_smp_atomic32_t function_calls;

#ifdef ERTS_SMP
static erts_smp_atomic32_t doing_sys_schedule;
static erts_smp_atomic32_t no_empty_run_queues;
long erts_runq_supervision_interval = 0;
static ethr_event runq_supervision_event;
static erts_tid_t runq_supervisor_tid;
static erts_atomic_t runq_supervisor_sleeping;
#else /* !ERTS_SMP */
ErtsSchedulerData *erts_scheduler_data;
#endif

ErtsAlignedRunQueue *erts_aligned_run_queues;
Uint erts_no_run_queues;

ErtsAlignedSchedulerData *erts_aligned_scheduler_data;
#ifdef ERTS_DIRTY_SCHEDULERS
ErtsAlignedSchedulerData *erts_aligned_dirty_cpu_scheduler_data;
ErtsAlignedSchedulerData *erts_aligned_dirty_io_scheduler_data;
#endif

typedef union {
    ErtsSchedulerSleepInfo ssi;
    char align[ERTS_ALC_CACHE_LINE_ALIGN_SIZE(sizeof(ErtsSchedulerSleepInfo))];
} ErtsAlignedSchedulerSleepInfo;

static ErtsAlignedSchedulerSleepInfo *aligned_sched_sleep_info;
#ifdef ERTS_DIRTY_SCHEDULERS
#ifdef ERTS_SMP
static ErtsAlignedSchedulerSleepInfo *aligned_dirty_cpu_sched_sleep_info;
static ErtsAlignedSchedulerSleepInfo *aligned_dirty_io_sched_sleep_info;
#endif
#endif

static Uint last_reductions;
static Uint last_exact_reductions;
Uint erts_default_process_flags;
Eterm erts_system_monitor;
Eterm erts_system_monitor_long_gc;
Uint erts_system_monitor_long_schedule;
Eterm erts_system_monitor_large_heap;
struct erts_system_monitor_flags_t erts_system_monitor_flags;

/* system performance monitor */
Eterm erts_system_profile;
struct erts_system_profile_flags_t erts_system_profile_flags;

#if ERTS_MAX_PROCESSES > 0x7fffffff
#error "Need to store process_count in another type"
#endif

typedef enum {
    ERTS_PSTT_GC,	/* Garbage Collect */
    ERTS_PSTT_CPC	/* Check Process Code */
} ErtsProcSysTaskType;

#define ERTS_MAX_PROC_SYS_TASK_ARGS 2

struct ErtsProcSysTask_ {
    ErtsProcSysTask *next;
    ErtsProcSysTask *prev;
    ErtsProcSysTaskType type;
    Eterm requester;
    Eterm reply_tag;
    Eterm req_id;
    Uint req_id_sz;
    Eterm arg[ERTS_MAX_PROC_SYS_TASK_ARGS];
    ErlOffHeap off_heap;
    Eterm heap[1];
};

#define ERTS_PROC_SYS_TASK_SIZE(HSz) \
    (sizeof(ErtsProcSysTask) - sizeof(Eterm) + sizeof(Eterm)*(HSz))

struct ErtsProcSysTaskQs_ {
    int qmask;
    int ncount;
    ErtsProcSysTask *q[ERTS_NO_PROC_PRIO_LEVELS];
};

ERTS_SCHED_PREF_QUICK_ALLOC_IMPL(proc_sys_task_queues,
				 ErtsProcSysTaskQs,
				 50,
				 ERTS_ALC_T_PROC_SYS_TSK_QS)

ERTS_SCHED_PREF_QUICK_ALLOC_IMPL(misc_op_list,
				 ErtsMiscOpList,
				 10,
				 ERTS_ALC_T_MISC_OP_LIST)

ERTS_SCHED_PREF_QUICK_ALLOC_IMPL(proclist,
				 ErtsProcList,
				 200,
				 ERTS_ALC_T_PROC_LIST)

#define ERTS_SCHED_SLEEP_INFO_IX(IX)					\
    (ASSERT(-1 <= ((int) (IX))					        \
		 && ((int) (IX)) < ((int) erts_no_schedulers)),		\
     &aligned_sched_sleep_info[(IX)].ssi)
#ifdef ERTS_DIRTY_SCHEDULERS
#define ERTS_DIRTY_CPU_SCHED_SLEEP_INFO_IX(IX)				\
    (ASSERT(0 <= ((int) (IX))					        \
	    && ((int) (IX)) < ((int) erts_no_dirty_cpu_schedulers)),	\
     &aligned_dirty_cpu_sched_sleep_info[(IX)].ssi)
#define ERTS_DIRTY_IO_SCHED_SLEEP_INFO_IX(IX)				\
    (ASSERT(0 <= ((int) (IX))					        \
	    && ((int) (IX)) < ((int) erts_no_dirty_io_schedulers)),	\
     &aligned_dirty_io_sched_sleep_info[(IX)].ssi)
#endif

#define ERTS_FOREACH_RUNQ(RQVAR, DO)					\
do {									\
    ErtsRunQueue *RQVAR;						\
    int ix__;								\
    for (ix__ = 0; ix__ < erts_no_run_queues; ix__++) {			\
	RQVAR = ERTS_RUNQ_IX(ix__);					\
	erts_smp_runq_lock(RQVAR);					\
	{ DO; }								\
	erts_smp_runq_unlock(RQVAR);					\
    }									\
} while (0)

#define ERTS_FOREACH_OP_RUNQ(RQVAR, DO)					\
do {									\
    ErtsRunQueue *RQVAR;						\
    int ix__;								\
    ERTS_SMP_LC_ASSERT(erts_smp_lc_mtx_is_locked(&schdlr_sspnd.mtx));	\
    for (ix__ = 0; ix__ < schdlr_sspnd.online; ix__++) {		\
	RQVAR = ERTS_RUNQ_IX(ix__);					\
	erts_smp_runq_lock(RQVAR);					\
	{ DO; }								\
	erts_smp_runq_unlock(RQVAR);					\
    }									\
} while (0)

#define ERTS_ATOMIC_FOREACH_RUNQ_X(RQVAR, DO, DOX)			\
do {									\
    ErtsRunQueue *RQVAR;						\
    int ix__;								\
    for (ix__ = 0; ix__ < erts_no_run_queues; ix__++) {			\
	RQVAR = ERTS_RUNQ_IX(ix__);					\
	erts_smp_runq_lock(RQVAR);					\
	{ DO; }								\
    }									\
    { DOX; }								\
    for (ix__ = 0; ix__ < erts_no_run_queues; ix__++)			\
	erts_smp_runq_unlock(ERTS_RUNQ_IX(ix__));			\
} while (0)

#define ERTS_ATOMIC_FOREACH_RUNQ(RQVAR, DO) \
  ERTS_ATOMIC_FOREACH_RUNQ_X(RQVAR, DO, )
/*
 * Local functions.
 */

static void exec_misc_ops(ErtsRunQueue *);
static void print_function_from_pc(int to, void *to_arg, BeamInstr* x);
static int stack_element_dump(int to, void *to_arg, Process* p, Eterm* sp,
			      int yreg);

static void aux_work_timeout(void *unused);
static void aux_work_timeout_early_init(int no_schedulers);
static void aux_work_timeout_late_init(void);
static void setup_aux_work_timer(void);

static int execute_sys_tasks(Process *c_p,
			     erts_aint32_t *statep,
			     int in_reds);
static int cleanup_sys_tasks(Process *c_p,
			     erts_aint32_t in_state,
			     int in_reds);


#if defined(DEBUG) || 0
#define ERTS_DBG_CHK_AUX_WORK_VAL(V) dbg_chk_aux_work_val((V))
static void
dbg_chk_aux_work_val(erts_aint32_t value)
{
    erts_aint32_t valid = 0;

    valid |= ERTS_SSI_AUX_WORK_SET_TMO;
    valid |= ERTS_SSI_AUX_WORK_MISC;
    valid |= ERTS_SSI_AUX_WORK_FIX_ALLOC_LOWER_LIM;
    valid |= ERTS_SSI_AUX_WORK_FIX_ALLOC_DEALLOC;
#if ERTS_USE_ASYNC_READY_Q
    valid |= ERTS_SSI_AUX_WORK_ASYNC_READY;
    valid |= ERTS_SSI_AUX_WORK_ASYNC_READY_CLEAN;
#endif
#ifdef ERTS_SMP
    valid |= ERTS_SSI_AUX_WORK_DELAYED_AW_WAKEUP;
    valid |= ERTS_SSI_AUX_WORK_MISC_THR_PRGR;
    valid |= ERTS_SSI_AUX_WORK_DD;
    valid |= ERTS_SSI_AUX_WORK_DD_THR_PRGR;
    valid |= ERTS_SSI_AUX_WORK_THR_PRGR_LATER_OP;
#endif
#if HAVE_ERTS_MSEG
    valid |= ERTS_SSI_AUX_WORK_MSEG_CACHE_CHECK;
#endif
#ifdef ERTS_SMP_SCHEDULERS_NEED_TO_CHECK_CHILDREN
    valid |= ERTS_SSI_AUX_WORK_CHECK_CHILDREN;
#endif
#ifdef ERTS_SSI_AUX_WORK_REAP_PORTS
    valid |= ERTS_SSI_AUX_WORK_REAP_PORTS;
#endif

    if (~valid & value)
	erl_exit(ERTS_ABORT_EXIT,
		 "Invalid aux_work value found: 0x%x\n",
		 ~valid & value);
}
#define ERTS_DBG_CHK_SSI_AUX_WORK(SSI) \
  ERTS_DBG_CHK_AUX_WORK_VAL(erts_atomic32_read_nob(&(SSI)->aux_work))
#else
#define ERTS_DBG_CHK_AUX_WORK_VAL(V)
#define ERTS_DBG_CHK_SSI_AUX_WORK(SSI)
#endif

#ifdef ERTS_SMP
static void handle_pending_exiters(ErtsProcList *);
static void wake_scheduler(ErtsRunQueue *rq);
#endif

#if defined(ERTS_SMP) && defined(ERTS_ENABLE_LOCK_CHECK)
int
erts_smp_lc_runq_is_locked(ErtsRunQueue *runq)
{
    return erts_smp_lc_mtx_is_locked(&runq->mtx);
}
#endif


static ERTS_INLINE Uint64
ensure_later_proc_interval(Uint64 interval)
{
    return erts_smp_ensure_later_interval_nob(erts_ptab_interval(&erts_proc), interval);
}

Uint64
erts_get_proc_interval(void)
{
    return erts_smp_current_interval_nob(erts_ptab_interval(&erts_proc));
}

Uint64
erts_ensure_later_proc_interval(Uint64 interval)
{
    return ensure_later_proc_interval(interval);
}

Uint64
erts_step_proc_interval(void)
{
    return erts_smp_step_interval_nob(erts_ptab_interval(&erts_proc));
}

void
erts_pre_init_process(void)
{
#ifdef USE_THREADS
    erts_tsd_key_create(&sched_data_key, "erts_sched_data_key");
#endif

#ifdef ERTS_ENABLE_LOCK_CHECK
 {
     int ix;

     erts_psd_required_locks[ERTS_PSD_ERROR_HANDLER].get_locks
	 = ERTS_PSD_ERROR_HANDLER_BUF_GET_LOCKS;
     erts_psd_required_locks[ERTS_PSD_ERROR_HANDLER].set_locks
	 = ERTS_PSD_ERROR_HANDLER_BUF_SET_LOCKS;

     erts_psd_required_locks[ERTS_PSD_SAVED_CALLS_BUF].get_locks
	 = ERTS_PSD_SAVED_CALLS_BUF_GET_LOCKS;
     erts_psd_required_locks[ERTS_PSD_SAVED_CALLS_BUF].set_locks
	 = ERTS_PSD_SAVED_CALLS_BUF_SET_LOCKS;

     erts_psd_required_locks[ERTS_PSD_SCHED_ID].get_locks
	 = ERTS_PSD_SCHED_ID_GET_LOCKS;
     erts_psd_required_locks[ERTS_PSD_SCHED_ID].set_locks
	 = ERTS_PSD_SCHED_ID_SET_LOCKS;

     erts_psd_required_locks[ERTS_PSD_DIST_ENTRY].get_locks
	 = ERTS_PSD_DIST_ENTRY_GET_LOCKS;
     erts_psd_required_locks[ERTS_PSD_DIST_ENTRY].set_locks
	 = ERTS_PSD_DIST_ENTRY_SET_LOCKS;

     erts_psd_required_locks[ERTS_PSD_CALL_TIME_BP].get_locks
	 = ERTS_PSD_CALL_TIME_BP_GET_LOCKS;
     erts_psd_required_locks[ERTS_PSD_CALL_TIME_BP].set_locks
	 = ERTS_PSD_CALL_TIME_BP_SET_LOCKS;

     erts_psd_required_locks[ERTS_PSD_DELAYED_GC_TASK_QS].get_locks
	 = ERTS_PSD_DELAYED_GC_TASK_QS_GET_LOCKS;
     erts_psd_required_locks[ERTS_PSD_DELAYED_GC_TASK_QS].set_locks
	 = ERTS_PSD_DELAYED_GC_TASK_QS_SET_LOCKS;

     erts_psd_required_locks[ERTS_PSD_NIF_TRAP_EXPORT].get_locks
	 = ERTS_PSD_NIF_TRAP_EXPORT_GET_LOCKS;
     erts_psd_required_locks[ERTS_PSD_NIF_TRAP_EXPORT].set_locks
	 = ERTS_PSD_NIF_TRAP_EXPORT_SET_LOCKS;

     /* Check that we have locks for all entries */
     for (ix = 0; ix < ERTS_PSD_SIZE; ix++) {
	 ERTS_SMP_LC_ASSERT(erts_psd_required_locks[ix].get_locks);
	 ERTS_SMP_LC_ASSERT(erts_psd_required_locks[ix].set_locks);
     }
 }
#endif
}

#ifdef ERTS_SMP
static void
release_process(void *vproc)
{
    erts_smp_proc_dec_refc((Process *) vproc);
}
#endif

/* initialize the scheduler */
void
erts_init_process(int ncpu, int proc_tab_size, int legacy_proc_tab)
{

#ifdef ERTS_SMP
    erts_disable_proc_not_running_opt = 0;
    erts_init_proc_lock(ncpu);
#endif

    init_proclist_alloc();

    erts_ptab_init_table(&erts_proc,
			 ERTS_ALC_T_PROC_TABLE,
#ifdef ERTS_SMP
			 release_process,
#else
			 NULL,
#endif
			 (ErtsPTabElementCommon *) &erts_invalid_process.common,
			 proc_tab_size,
			 sizeof(Process),
			 "process_table",
			 legacy_proc_tab);

    last_reductions = 0;
    last_exact_reductions = 0;
    erts_default_process_flags = 0;
}

void
erts_late_init_process(void)
{
    int ix;

    erts_smp_spinlock_init(&erts_sched_stat.lock, "sched_stat");
    for (ix = 0; ix < ERTS_NO_PRIO_LEVELS; ix++) {
	Eterm atom;
	char *atom_str;
	switch (ix) {
	case PRIORITY_MAX:
	    atom_str = "process_max";
	    break;
	case PRIORITY_HIGH:
	    atom_str = "process_high";
	    break;
	case PRIORITY_NORMAL:
	    atom_str = "process_normal";
	    break;
	case PRIORITY_LOW:
	    atom_str = "process_low";
	    break;
	case ERTS_PORT_PRIO_LEVEL:
	    atom_str = "port";
	    break;
	default:
	    atom_str = "bad_prio";
	    ASSERT(!"bad prio");
	    break;
	}
	atom = am_atom_put(atom_str, sys_strlen(atom_str));
	erts_sched_stat.prio[ix].name = atom;
	erts_sched_stat.prio[ix].total_executed = 0;
	erts_sched_stat.prio[ix].executed = 0;
	erts_sched_stat.prio[ix].total_migrated = 0;
	erts_sched_stat.prio[ix].migrated = 0;
    }

}

static void
init_sched_wall_time(ErtsSchedWallTime *swtp)
{
    swtp->need = erts_sched_balance_util;
    swtp->enabled = 0;
    swtp->start = 0;
    swtp->working.total = 0;
    swtp->working.start = 0;
    swtp->working.currently = 0;
}

static ERTS_INLINE Uint64
sched_wall_time_ts(void)
{
#ifdef HAVE_GETHRTIME
    return (Uint64) sys_gethrtime();
#else
    Uint64 res;
    SysTimeval tv;
    sys_gettimeofday(&tv);
    res = (Uint64) tv.tv_sec*1000000;
    res += (Uint64) tv.tv_usec;
    return res;
#endif
}

#if ERTS_HAVE_SCHED_UTIL_BALANCING_SUPPORT

#ifdef ARCH_64

static ERTS_INLINE Uint64
aschedtime_read(ErtsAtomicSchedTime *var)
{
    return (Uint64) erts_atomic_read_nob((erts_atomic_t *) var);
}

static ERTS_INLINE void
aschedtime_set(ErtsAtomicSchedTime *var, Uint64 val)
{
    erts_atomic_set_nob((erts_atomic_t *) var, (erts_aint_t) val);
}

static ERTS_INLINE void
aschedtime_init(ErtsAtomicSchedTime *var)
{
    erts_atomic_init_nob((erts_atomic_t *) var, (erts_aint_t) 0);
}

#elif defined(ARCH_32)

static ERTS_INLINE Uint64
aschedtime_read(ErtsAtomicSchedTime *var)
{
    erts_dw_aint_t dw;
    erts_dw_atomic_read_nob((erts_dw_atomic_t *) var, &dw);
#ifdef ETHR_SU_DW_NAINT_T__
    return (Uint64) dw.dw_sint;
#else
    {
	Uint64 res;
	res = (Uint64) ((Uint32) dw.sint[ERTS_DW_AINT_HIGH_WORD]);
	res <<= 32;
	res |= (Uint64) ((Uint32) dw.sint[ERTS_DW_AINT_LOW_WORD]);
	return res;
    }
#endif    
}

static ERTS_INLINE void
aschedtime_set(ErtsAtomicSchedTime *var, Uint64 val)
{
    erts_dw_aint_t dw;
#ifdef ETHR_SU_DW_NAINT_T__
    dw.dw_sint = (ETHR_SU_DW_NAINT_T__) val;
#else
    dw.sint[ERTS_DW_AINT_LOW_WORD] = (erts_aint_t) (val & 0xffffffff);
    dw.sint[ERTS_DW_AINT_HIGH_WORD] = (erts_aint_t) ((val >> 32) & 0xffffffff);
#endif
    erts_dw_atomic_set_nob((erts_dw_atomic_t *) var, &dw);
}

static ERTS_INLINE void
aschedtime_init(ErtsAtomicSchedTime *var)
{
    erts_dw_aint_t dw;
    dw.sint[ERTS_DW_AINT_LOW_WORD] = (erts_aint_t) 0;
    dw.sint[ERTS_DW_AINT_HIGH_WORD] = (erts_aint_t) 0;
    erts_dw_atomic_init_nob((erts_dw_atomic_t *) var, &dw);
}

#else
#  error :-/
#endif

#define ERTS_GET_AVG_MAX_UNLOCKED_TRY 50
#define ERTS_SCHED_AVG_UTIL_WRITE_MARKER (~((Uint64) 0))

/* Intervals in nanoseconds */
#define ERTS_SCHED_UTIL_SHORT_INTERVAL ((Uint64) 1*1000*1000*1000)
#define ERTS_SCHED_UTIL_LONG_INTERVAL ((Uint64) 10*1000*1000*1000)


#define ERTS_SCHED_UTIL_IGNORE_IMBALANCE_DIFF 5000 /* ppm */

static ERTS_INLINE Uint64
calc_sched_worktime(int is_working, Uint64 now, Uint64 last,
		    Uint64 interval, Uint64 old_worktime)
{
    Uint64 worktime;
    Uint64 new;

    if (now <= last)
	return old_worktime;

    new = now - last;

    if (new >= interval)
	return is_working ? interval : (Uint64) 0;


    /*
     * Division by 1000 in order to avoid
     * overflow. If changed update assertions
     * in init_runq_sched_util().
     */
    worktime = old_worktime;
    worktime *= (interval - new)/1000;
    worktime /= (interval/1000);
    if (is_working)
	worktime += new;

    ASSERT(0 <= worktime && worktime <= interval);

    return worktime;
}

static ERTS_INLINE void
update_avg_sched_util(ErtsSchedulerData *esdp, Uint64 now, int is_working)
{
    ErtsRunQueue *rq;
    int worked;
    Uint64 swt, lwt, last;

    rq = esdp->run_queue;
    last = aschedtime_read(&rq->sched_util.last);

    if (now <= last) {
	ASSERT(last == ERTS_SCHED_AVG_UTIL_WRITE_MARKER);
	return;
    }

    ASSERT(now >= last);

    worked = rq->sched_util.is_working;

    swt = calc_sched_worktime(worked, now, last, ERTS_SCHED_UTIL_SHORT_INTERVAL,
			      rq->sched_util.worktime.short_interval);
    lwt = calc_sched_worktime(worked, now, last, ERTS_SCHED_UTIL_LONG_INTERVAL,
			      rq->sched_util.worktime.long_interval);

    aschedtime_set(&rq->sched_util.last, ERTS_SCHED_AVG_UTIL_WRITE_MARKER);
    ERTS_THR_WRITE_MEMORY_BARRIER;
    rq->sched_util.is_working = is_working;
    rq->sched_util.worktime.short_interval = swt;
    rq->sched_util.worktime.long_interval = lwt;
    ERTS_THR_WRITE_MEMORY_BARRIER;
    aschedtime_set(&rq->sched_util.last, now);
}

int
erts_get_sched_util(ErtsRunQueue *rq, int initially_locked, int short_interval)
{
    /* Average scheduler utilization in ppm */
    int util, is_working, try = 0, locked = initially_locked;
    Uint64 worktime, old_worktime, now, last, interval, *old_worktimep;

    if (short_interval) {
	old_worktimep = &rq->sched_util.worktime.short_interval;
	interval = ERTS_SCHED_UTIL_SHORT_INTERVAL;
    }
    else {
	old_worktimep = &rq->sched_util.worktime.long_interval;
	interval = ERTS_SCHED_UTIL_LONG_INTERVAL;
    }

    while (1) {
	Uint64 chk_last;
	last = aschedtime_read(&rq->sched_util.last);
	ERTS_THR_READ_MEMORY_BARRIER;
	is_working = rq->sched_util.is_working;
	old_worktime = *old_worktimep;
	ERTS_THR_READ_MEMORY_BARRIER;
	chk_last = aschedtime_read(&rq->sched_util.last);
	if (chk_last == last)
	    break;
	if (!locked) {
	    if (++try >= ERTS_GET_AVG_MAX_UNLOCKED_TRY) {
		/* Writer will eventually block on runq-lock */
		erts_smp_runq_lock(rq);
		locked = 1;
	    }
	}
    }

    if (!initially_locked && locked)
	erts_smp_runq_unlock(rq);

    now = sched_wall_time_ts();
    worktime = calc_sched_worktime(is_working, now, last, interval, old_worktime);

    util = (int) ((worktime * 1000000)/interval);

    ASSERT(0 <= util && util <= 1000000);

    return util;
}

static void
init_runq_sched_util(ErtsRunQueueSchedUtil *rqsu, int enabled)
{
    aschedtime_init(&rqsu->last);
    if (!enabled)
	aschedtime_set(&rqsu->last, ERTS_SCHED_AVG_UTIL_WRITE_MARKER);
    rqsu->is_working = 0;
    rqsu->worktime.short_interval = (Uint64) 0;
    rqsu->worktime.long_interval = (Uint64) 0;

#ifdef DEBUG
    {
	Uint64 intrvl;
	/*
	 * If one of these asserts fail we may have
	 * overflow in calc_sched_worktime(). Which
	 * have to be fixed either by shrinking
	 * interval size, or fix calculation of
	 * worktime in calc_sched_worktime().
	 */
	intrvl = ERTS_SCHED_UTIL_SHORT_INTERVAL;
	ASSERT(intrvl*(intrvl/1000) > intrvl);
	intrvl = ERTS_SCHED_UTIL_LONG_INTERVAL;
	ASSERT(intrvl*(intrvl/1000) > intrvl);
    }
#endif
}

#endif /* ERTS_HAVE_SCHED_UTIL_BALANCING_SUPPORT */

static ERTS_INLINE void
sched_wall_time_change(ErtsSchedulerData *esdp, int working)
{
    if (esdp->sched_wall_time.need) {
	Uint64 ts = sched_wall_time_ts();
#if ERTS_HAVE_SCHED_UTIL_BALANCING_SUPPORT
	update_avg_sched_util(esdp, ts, working);
#endif
	if (esdp->sched_wall_time.enabled) {
	    if (working) {
#ifdef DEBUG
		ASSERT(!esdp->sched_wall_time.working.currently);
		esdp->sched_wall_time.working.currently = 1;
#endif
		ts -= esdp->sched_wall_time.start;
		esdp->sched_wall_time.working.start = ts;
	    }
	    else {
#ifdef DEBUG
		ASSERT(esdp->sched_wall_time.working.currently);
		esdp->sched_wall_time.working.currently = 0;
#endif
		ts -= esdp->sched_wall_time.start;
		ts -= esdp->sched_wall_time.working.start;
		esdp->sched_wall_time.working.total += ts;
	    }
	}
    }
}

typedef struct {
    int set;
    int enable;
    Process *proc;
    Eterm ref;
    Eterm ref_heap[REF_THING_SIZE];
    Uint req_sched;
    erts_smp_atomic32_t refc;
} ErtsSchedWallTimeReq;

#if !HALFWORD_HEAP
ERTS_SCHED_PREF_QUICK_ALLOC_IMPL(swtreq,
				 ErtsSchedWallTimeReq,
				 5,
				 ERTS_ALC_T_SCHED_WTIME_REQ)
#else
static ERTS_INLINE ErtsSchedWallTimeReq *
swtreq_alloc(void)
{
    return erts_alloc(ERTS_ALC_T_SCHED_WTIME_REQ,
		      sizeof(ErtsSchedWallTimeReq));
}

static ERTS_INLINE void
swtreq_free(ErtsSchedWallTimeReq *ptr)
{
    erts_free(ERTS_ALC_T_SCHED_WTIME_REQ, ptr);
}
#endif

static void
reply_sched_wall_time(void *vswtrp)
{
    Uint64 working = 0, total = 0;
    ErtsSchedulerData *esdp = erts_get_scheduler_data();
    ErtsSchedWallTimeReq *swtrp = (ErtsSchedWallTimeReq *) vswtrp;
    ErtsProcLocks rp_locks = (swtrp->req_sched == esdp->no
			      ? ERTS_PROC_LOCK_MAIN
			      : 0);
    Process *rp = swtrp->proc;
    Eterm ref_copy = NIL, msg;
    Eterm *hp = NULL;
    Eterm **hpp;
    Uint sz, *szp;
    ErlOffHeap *ohp = NULL;
    ErlHeapFragment *bp = NULL;

    ASSERT(esdp);
#ifdef ERTS_DIRTY_SCHEDULERS
    ASSERT(!ERTS_SCHEDULER_IS_DIRTY(esdp));
#endif
    if (swtrp->set) {
	if (!swtrp->enable && esdp->sched_wall_time.enabled) {
	    esdp->sched_wall_time.need = erts_sched_balance_util;
	    esdp->sched_wall_time.enabled = 0;
	}
	else if (swtrp->enable && !esdp->sched_wall_time.enabled) {
	    Uint64 ts = sched_wall_time_ts();
	    esdp->sched_wall_time.need = 1;
	    esdp->sched_wall_time.enabled = 1;
	    esdp->sched_wall_time.start = ts;
	    esdp->sched_wall_time.working.total = 0;
	    esdp->sched_wall_time.working.start = 0;
	    esdp->sched_wall_time.working.currently = 1;
	}
    }

    if (esdp->sched_wall_time.enabled) {
	Uint64 ts = sched_wall_time_ts();
	ASSERT(esdp->sched_wall_time.working.currently);
	ts -= esdp->sched_wall_time.start;
	total = ts;
	ts -= esdp->sched_wall_time.working.start;
	working = esdp->sched_wall_time.working.total + ts;
    }

    sz = 0;
    hpp = NULL;
    szp = &sz;

    while (1) {
	if (hpp)
	    ref_copy = STORE_NC(hpp, ohp, swtrp->ref);
	else
	    *szp += REF_THING_SIZE;

	if (swtrp->set)
	    msg = ref_copy;
	else {
	    msg = (!esdp->sched_wall_time.enabled
		   ? am_notsup
		   : erts_bld_tuple(hpp, szp, 3,
				    make_small(esdp->no),
				    erts_bld_uint64(hpp, szp, working),
				    erts_bld_uint64(hpp, szp, total)));

	    msg = erts_bld_tuple(hpp, szp, 2, ref_copy, msg);
	}
	if (hpp)
	    break;

	hp = erts_alloc_message_heap(sz, &bp, &ohp, rp, &rp_locks);
	szp = NULL;
	hpp = &hp;
    }

    erts_queue_message(rp, &rp_locks, bp, msg, NIL
#ifdef USE_VM_PROBES
			   , NIL
#endif
		       );

    if (swtrp->req_sched == esdp->no)
	rp_locks &= ~ERTS_PROC_LOCK_MAIN;
 
    if (rp_locks)
	erts_smp_proc_unlock(rp, rp_locks);

    erts_smp_proc_dec_refc(rp);

    if (erts_smp_atomic32_dec_read_nob(&swtrp->refc) == 0)
	swtreq_free(vswtrp);
}

Eterm
erts_sched_wall_time_request(Process *c_p, int set, int enable)
{
    ErtsSchedulerData *esdp = ERTS_PROC_GET_SCHDATA(c_p);
    Eterm ref;
    ErtsSchedWallTimeReq *swtrp;
    Eterm *hp;

    if (!set && !esdp->sched_wall_time.enabled)
	return THE_NON_VALUE;
#ifdef ERTS_DIRTY_SCHEDULERS
    ASSERT(!ERTS_SCHEDULER_IS_DIRTY(esdp));
#endif

    swtrp = swtreq_alloc();
    ref = erts_make_ref(c_p);
    hp = &swtrp->ref_heap[0];

    swtrp->set = set;
    swtrp->enable = enable;
    swtrp->proc = c_p;
    swtrp->ref = STORE_NC(&hp, NULL, ref);
    swtrp->req_sched = esdp->no;
    erts_smp_atomic32_init_nob(&swtrp->refc,
			       (erts_aint32_t) erts_no_schedulers);

    erts_smp_proc_add_refc(c_p, (Sint32) erts_no_schedulers);

#ifdef ERTS_SMP
    if (erts_no_schedulers > 1)
	erts_schedule_multi_misc_aux_work(1,
					  erts_no_schedulers,
					  reply_sched_wall_time,
					  (void *) swtrp);
#endif

    reply_sched_wall_time((void *) swtrp);

    return ref;
}

static ERTS_INLINE ErtsProcList *
proclist_create(Process *p)
{
    ErtsProcList *plp = proclist_alloc();
    ensure_later_proc_interval(p->common.u.alive.started_interval);
    plp->pid = p->common.id;
    plp->started_interval = p->common.u.alive.started_interval;
    return plp;
}

static ERTS_INLINE void
proclist_destroy(ErtsProcList *plp)
{
    proclist_free(plp);
}

ErtsProcList *
erts_proclist_create(Process *p)
{
    return proclist_create(p);
}

void
erts_proclist_destroy(ErtsProcList *plp)
{
    proclist_destroy(plp);
}

void *
erts_psd_set_init(Process *p, ErtsProcLocks plocks, int ix, void *data)
{
    void *old;
    ErtsProcLocks xplocks;
    int refc = 0;
    ErtsPSD *psd = erts_alloc(ERTS_ALC_T_PSD, sizeof(ErtsPSD));
    int i;
    for (i = 0; i < ERTS_PSD_SIZE; i++)
	psd->data[i] = NULL;

    ERTS_SMP_LC_ASSERT(plocks);
    ERTS_SMP_LC_ASSERT(plocks == erts_proc_lc_my_proc_locks(p));

    xplocks = ERTS_PROC_LOCKS_ALL;
    xplocks &= ~plocks;
    if (xplocks && erts_smp_proc_trylock(p, xplocks) == EBUSY) {
	if (xplocks & ERTS_PROC_LOCK_MAIN) {
	    erts_smp_proc_inc_refc(p);
	    erts_smp_proc_unlock(p, plocks);
	    erts_smp_proc_lock(p, ERTS_PROC_LOCKS_ALL);
	    refc = 1;
	}
	else {
	    if (plocks & ERTS_PROC_LOCKS_ALL_MINOR)
		erts_smp_proc_unlock(p, plocks & ERTS_PROC_LOCKS_ALL_MINOR);
	    erts_smp_proc_lock(p, ERTS_PROC_LOCKS_ALL_MINOR);
	}
    }
    if (!p->psd)
	p->psd = psd;
    if (xplocks)
	erts_smp_proc_unlock(p, xplocks);
    if (refc)
	erts_smp_proc_dec_refc(p);
    ASSERT(p->psd);
    if (p->psd != psd)
	erts_free(ERTS_ALC_T_PSD, psd);
    old = p->psd->data[ix];
    p->psd->data[ix] = data;
    ERTS_SMP_LC_ASSERT(plocks == erts_proc_lc_my_proc_locks(p));
    return old;
}

#ifdef ERTS_SMP

void
erts_sched_finish_poke(ErtsSchedulerSleepInfo *ssi, erts_aint32_t flags)
{
    switch (flags & ERTS_SSI_FLGS_SLEEP_TYPE) {
    case ERTS_SSI_FLG_POLL_SLEEPING:
	erts_sys_schedule_interrupt(1);
	break;
    case ERTS_SSI_FLG_POLL_SLEEPING|ERTS_SSI_FLG_TSE_SLEEPING:
	/*
	 * Thread progress blocking while poll sleeping; need
	 * to signal on both...
	 */
	erts_sys_schedule_interrupt(1);
	/* fall through */
    case ERTS_SSI_FLG_TSE_SLEEPING:
	erts_tse_set(ssi->event);
	break;
    case 0:
	break;
    default:
	erl_exit(ERTS_ABORT_EXIT, "%s:%d: Internal error\n",
		 __FILE__, __LINE__);
	break;
    }
}

#endif

static ERTS_INLINE void
set_aux_work_flags_wakeup_nob(ErtsSchedulerSleepInfo *ssi,
			      erts_aint32_t flgs)
{
    erts_aint32_t old_flgs;

    ERTS_DBG_CHK_SSI_AUX_WORK(ssi);

    old_flgs = erts_atomic32_read_nob(&ssi->aux_work);
    if ((old_flgs & flgs) == 0) {

	old_flgs = erts_atomic32_read_bor_nob(&ssi->aux_work, flgs);

	if ((old_flgs & flgs) == 0) {
#ifdef ERTS_SMP
	    erts_sched_poke(ssi);
#else
	    erts_sys_schedule_interrupt(1);
#endif
	}
    }
}

static ERTS_INLINE void
set_aux_work_flags_wakeup_relb(ErtsSchedulerSleepInfo *ssi,
			       erts_aint32_t flgs)
{
    erts_aint32_t old_flgs;

    ERTS_DBG_CHK_SSI_AUX_WORK(ssi);

    old_flgs = erts_atomic32_read_bor_relb(&ssi->aux_work, flgs);

    if ((old_flgs & flgs) == 0) {
#ifdef ERTS_SMP
	erts_sched_poke(ssi);
#else
	erts_sys_schedule_interrupt(1);
#endif
    }
}

static ERTS_INLINE erts_aint32_t
set_aux_work_flags(ErtsSchedulerSleepInfo *ssi, erts_aint32_t flgs)
{
    return erts_atomic32_read_bor_nob(&ssi->aux_work, flgs);
}

static ERTS_INLINE erts_aint32_t
unset_aux_work_flags(ErtsSchedulerSleepInfo *ssi, erts_aint32_t flgs)
{
    return erts_atomic32_read_band_nob(&ssi->aux_work, ~flgs);
}

#ifdef ERTS_SMP

static ERTS_INLINE void
haw_chk_later_cleanup_op_wakeup(ErtsAuxWorkData *awdp, ErtsThrPrgrVal val)
{
    if (awdp->later_op.first
	&& erts_thr_progress_cmp(val, awdp->later_op.thr_prgr) >= 0) {
	awdp->later_op.size = thr_prgr_later_cleanup_op_threshold;
    }
}

static ERTS_INLINE void
haw_thr_prgr_wakeup(ErtsAuxWorkData *awdp, ErtsThrPrgrVal val)
{
    int cmp = erts_thr_progress_cmp(val, awdp->latest_wakeup);
    if (cmp != 0) {
	if (cmp > 0) {
	    awdp->latest_wakeup = val;
	    haw_chk_later_cleanup_op_wakeup(awdp, val);
	}
	erts_thr_progress_wakeup(awdp->esdp, val);
    }
}

static ERTS_INLINE void
haw_thr_prgr_soft_wakeup(ErtsAuxWorkData *awdp, ErtsThrPrgrVal val)
{
    if (erts_thr_progress_cmp(val, awdp->latest_wakeup) > 0) {
	awdp->latest_wakeup = val;
	haw_chk_later_cleanup_op_wakeup(awdp, val);
	erts_thr_progress_wakeup(awdp->esdp, val);
    }
}

static ERTS_INLINE void
haw_thr_prgr_later_cleanup_op_wakeup(ErtsAuxWorkData *awdp, ErtsThrPrgrVal val, UWord size)
{
    if (erts_thr_progress_cmp(val, awdp->latest_wakeup) > 0) {
	awdp->later_op.thr_prgr = val;
	if (awdp->later_op.size > size)
	    awdp->later_op.size -= size;
	else {
	    awdp->latest_wakeup = val;
	    awdp->later_op.size = thr_prgr_later_cleanup_op_threshold;
	    erts_thr_progress_wakeup(awdp->esdp, val);
	}
    }
}

static ERTS_INLINE void
haw_thr_prgr_current_reset(ErtsAuxWorkData *awdp)
{
    awdp->current_thr_prgr = ERTS_THR_PRGR_INVALID;
}

static ERTS_INLINE ErtsThrPrgrVal
haw_thr_prgr_current(ErtsAuxWorkData *awdp)
{
    ErtsThrPrgrVal current = awdp->current_thr_prgr;
    if (current == ERTS_THR_PRGR_INVALID) {
	current = erts_thr_progress_current();
	awdp->current_thr_prgr = current;
    }
    return current;
}

static ERTS_INLINE void
haw_thr_prgr_current_check_progress(ErtsAuxWorkData *awdp)
{
    ErtsThrPrgrVal current = awdp->current_thr_prgr;
#ifdef ERTS_DIRTY_SCHEDULERS
    ASSERT(!awdp->esdp || !ERTS_SCHEDULER_IS_DIRTY(awdp->esdp));
#endif
    if (current != ERTS_THR_PRGR_INVALID
	&& !erts_thr_progress_equal(current, erts_thr_progress_current())) {
	/*
	 * We have used a previouly read current value that isn't the
	 * latest; need to poke ourselfs in order to guarantee no loss
	 * of wakeups.
	 */
	erts_sched_poke(awdp->ssi);
    }
}

static ERTS_INLINE erts_aint32_t
handle_delayed_aux_work_wakeup(ErtsAuxWorkData *awdp, erts_aint32_t aux_work, int waiting)
{
    int jix, max_jix;

#ifdef ERTS_DIRTY_SCHEDULERS
    ASSERT(!awdp->esdp || !ERTS_SCHEDULER_IS_DIRTY(awdp->esdp));
#endif

    ASSERT(awdp->delayed_wakeup.next != ERTS_DELAYED_WAKEUP_INFINITY);

    if (!waiting && awdp->delayed_wakeup.next > awdp->esdp->reductions)
	return aux_work;

    unset_aux_work_flags(awdp->ssi, ERTS_SSI_AUX_WORK_DELAYED_AW_WAKEUP);

    ERTS_THR_MEMORY_BARRIER;

    max_jix = awdp->delayed_wakeup.jix;
    awdp->delayed_wakeup.jix = -1;
    for (jix = 0; jix <= max_jix; jix++) {
	int sched = awdp->delayed_wakeup.job[jix].sched;
	erts_aint32_t aux_work = awdp->delayed_wakeup.job[jix].aux_work;

	ASSERT(awdp->delayed_wakeup.sched2jix[sched] == jix);
	awdp->delayed_wakeup.sched2jix[sched] = -1;
	set_aux_work_flags_wakeup_nob(ERTS_SCHED_SLEEP_INFO_IX(sched-1),
				      aux_work);
    }
    awdp->delayed_wakeup.next = ERTS_DELAYED_WAKEUP_INFINITY;
    return aux_work & ~ERTS_SSI_AUX_WORK_DELAYED_AW_WAKEUP;
}

static ERTS_INLINE void
schedule_aux_work_wakeup(ErtsAuxWorkData *awdp,
			 int sched,
			 erts_aint32_t aux_work)
{
    int jix = awdp->delayed_wakeup.sched2jix[sched];
    if (jix >= 0) {
	ASSERT(awdp->delayed_wakeup.job[jix].sched == sched);
	awdp->delayed_wakeup.job[jix].aux_work |= aux_work;
    }
    else {
	jix = ++awdp->delayed_wakeup.jix;
	awdp->delayed_wakeup.sched2jix[sched] = jix;
	awdp->delayed_wakeup.job[jix].sched = sched;
	awdp->delayed_wakeup.job[jix].aux_work = aux_work;
    }

    if (awdp->delayed_wakeup.next != ERTS_DELAYED_WAKEUP_INFINITY) {
	ASSERT(erts_atomic32_read_nob(&awdp->ssi->aux_work)
	       & ERTS_SSI_AUX_WORK_DELAYED_AW_WAKEUP);
    }
    else {
	awdp->delayed_wakeup.next = (awdp->esdp->reductions
				     + ERTS_DELAYED_WAKEUP_REDUCTIONS);

	ASSERT(!(erts_atomic32_read_nob(&awdp->ssi->aux_work)
		 & ERTS_SSI_AUX_WORK_DELAYED_AW_WAKEUP));
	set_aux_work_flags_wakeup_nob(awdp->ssi,
				      ERTS_SSI_AUX_WORK_DELAYED_AW_WAKEUP);
    }
}

#endif

typedef struct erts_misc_aux_work_t_ erts_misc_aux_work_t;
struct erts_misc_aux_work_t_ {
    void (*func)(void *);
    void *arg;
};

ERTS_SCHED_PREF_QUICK_ALLOC_IMPL(misc_aux_work,
				 erts_misc_aux_work_t,
				 200,
				 ERTS_ALC_T_MISC_AUX_WORK)

typedef union {
    ErtsThrQ_t q;
    char align[ERTS_ALC_CACHE_LINE_ALIGN_SIZE(sizeof(ErtsThrQ_t))];
} erts_algnd_misc_aux_work_q_t;

static erts_algnd_misc_aux_work_q_t *misc_aux_work_queues;

static void
notify_aux_work(void *vssi)
{
    set_aux_work_flags_wakeup_nob((ErtsSchedulerSleepInfo *) vssi,
				  ERTS_SSI_AUX_WORK_MISC);
}

static void
init_misc_aux_work(void)
{
    int ix;
    ErtsThrQInit_t qinit = ERTS_THR_Q_INIT_DEFAULT;
    qinit.notify = notify_aux_work;

    init_misc_aux_work_alloc();

    misc_aux_work_queues = 
	erts_alloc_permanent_cache_aligned(ERTS_ALC_T_MISC_AUX_WORK_Q,
					   sizeof(erts_algnd_misc_aux_work_q_t)
					   * (erts_no_schedulers+1));

#ifdef ERTS_SMP
    ix = 0; /* aux_thread + schedulers */
#else
    ix = 1; /* scheduler only */
#endif

    for (; ix <= erts_no_schedulers; ix++) {
	qinit.arg = (void *) ERTS_SCHED_SLEEP_INFO_IX(ix-1);
	erts_thr_q_initialize(&misc_aux_work_queues[ix].q, &qinit);
    }
}

static erts_aint32_t
misc_aux_work_clean(ErtsThrQ_t *q,
		    ErtsAuxWorkData *awdp,
		    erts_aint32_t aux_work)
{
    switch (erts_thr_q_clean(q)) {
    case ERTS_THR_Q_DIRTY:
	set_aux_work_flags(awdp->ssi, ERTS_SSI_AUX_WORK_MISC);
	return aux_work | ERTS_SSI_AUX_WORK_MISC;
    case ERTS_THR_Q_NEED_THR_PRGR:
#ifdef ERTS_SMP
	set_aux_work_flags(awdp->ssi, ERTS_SSI_AUX_WORK_MISC_THR_PRGR);
	haw_thr_prgr_soft_wakeup(awdp, erts_thr_q_need_thr_progress(q));
#endif
    case ERTS_THR_Q_CLEAN:
	break;
    }
    return aux_work;
}

static ERTS_INLINE erts_aint32_t
handle_misc_aux_work(ErtsAuxWorkData *awdp,
		     erts_aint32_t aux_work,
		     int waiting)
{
    ErtsThrQ_t *q = &misc_aux_work_queues[awdp->sched_id].q;

    unset_aux_work_flags(awdp->ssi, ERTS_SSI_AUX_WORK_MISC);
    while (1) {
	erts_misc_aux_work_t *mawp = erts_thr_q_dequeue(q);
	if (!mawp)
	    break;
	mawp->func(mawp->arg);
	misc_aux_work_free(mawp);
    }

    return misc_aux_work_clean(q, awdp, aux_work & ~ERTS_SSI_AUX_WORK_MISC);
}

#ifdef ERTS_SMP

static ERTS_INLINE erts_aint32_t
handle_misc_aux_work_thr_prgr(ErtsAuxWorkData *awdp,
			      erts_aint32_t aux_work,
			      int waiting)
{
#ifdef ERTS_DIRTY_SCHEDULERS
    ASSERT(!awdp->esdp || !ERTS_SCHEDULER_IS_DIRTY(awdp->esdp));
#endif
    if (!erts_thr_progress_has_reached_this(haw_thr_prgr_current(awdp),
					    awdp->misc.thr_prgr))
	return aux_work & ~ERTS_SSI_AUX_WORK_MISC_THR_PRGR;

    unset_aux_work_flags(awdp->ssi, ERTS_SSI_AUX_WORK_MISC_THR_PRGR);

    return misc_aux_work_clean(&misc_aux_work_queues[awdp->sched_id].q,
			       awdp,
			       aux_work & ~ERTS_SSI_AUX_WORK_MISC_THR_PRGR);
}

#endif

static ERTS_INLINE void
schedule_misc_aux_work(int sched_id,
		       void (*func)(void *),
		       void *arg)
{
    ErtsThrQ_t *q;
    erts_misc_aux_work_t *mawp;

#ifdef ERTS_SMP
    ASSERT(0 <= sched_id && sched_id <= erts_no_schedulers);
#else
    ASSERT(sched_id == 1);
#endif

    q = &misc_aux_work_queues[sched_id].q;
    mawp = misc_aux_work_alloc();
    mawp->func = func;
    mawp->arg = arg;
    erts_thr_q_enqueue(q, mawp);
}

void
erts_schedule_misc_aux_work(int sched_id,
			    void (*func)(void *),
			    void *arg)
{
    schedule_misc_aux_work(sched_id, func, arg);
}

void
erts_schedule_multi_misc_aux_work(int ignore_self,
				  int max_sched,
				  void (*func)(void *),
				  void *arg)
{
    int id, self = 0;

    if (ignore_self) {
	ErtsSchedulerData *esdp = erts_get_scheduler_data();
#ifdef ERTS_DIRTY_SCHEDULERS
	ASSERT(!ERTS_SCHEDULER_IS_DIRTY(esdp));
#endif
	if (esdp)
	    self = (int) esdp->no;
    }

    ASSERT(0 < max_sched && max_sched <= erts_no_schedulers);

    for (id = 1; id <= max_sched; id++) {
	if (id == self)
	    continue;
	schedule_misc_aux_work(id, func, arg);
   }
}

#if ERTS_USE_ASYNC_READY_Q

void
erts_notify_check_async_ready_queue(void *vno)
{
    int ix = ((int) (SWord) vno) -1;
    set_aux_work_flags_wakeup_nob(ERTS_SCHED_SLEEP_INFO_IX(ix),
				  ERTS_SSI_AUX_WORK_ASYNC_READY);
}

static ERTS_INLINE erts_aint32_t
handle_async_ready(ErtsAuxWorkData *awdp,
		   erts_aint32_t aux_work,
		   int waiting)
{
    ErtsSchedulerSleepInfo *ssi = awdp->ssi;
#ifdef ERTS_DIRTY_SCHEDULERS
    ASSERT(!awdp->esdp || !ERTS_SCHEDULER_IS_DIRTY(awdp->esdp));
#endif
    unset_aux_work_flags(ssi, ERTS_SSI_AUX_WORK_ASYNC_READY);
    if (erts_check_async_ready(awdp->async_ready.queue)) {
	if (set_aux_work_flags(ssi, ERTS_SSI_AUX_WORK_ASYNC_READY)
	    & ERTS_SSI_AUX_WORK_ASYNC_READY_CLEAN) {
	    unset_aux_work_flags(ssi, ERTS_SSI_AUX_WORK_ASYNC_READY_CLEAN);
	    aux_work &= ~ERTS_SSI_AUX_WORK_ASYNC_READY_CLEAN;
	}
	return aux_work;
    }
#ifdef ERTS_SMP
    awdp->async_ready.need_thr_prgr = 0;
#endif
    set_aux_work_flags(ssi, ERTS_SSI_AUX_WORK_ASYNC_READY_CLEAN);
    return ((aux_work & ~ERTS_SSI_AUX_WORK_ASYNC_READY)
	    | ERTS_SSI_AUX_WORK_ASYNC_READY_CLEAN);
}

static ERTS_INLINE erts_aint32_t
handle_async_ready_clean(ErtsAuxWorkData *awdp,
			 erts_aint32_t aux_work,
			 int waiting)
{
    void *thr_prgr_p;

#ifdef ERTS_DIRTY_SCHEDULERS
    ASSERT(!awdp->esdp || !ERTS_SCHEDULER_IS_DIRTY(awdp->esdp));
#endif
#ifdef ERTS_SMP
    if (awdp->async_ready.need_thr_prgr
	&& !erts_thr_progress_has_reached_this(haw_thr_prgr_current(awdp),
					       awdp->async_ready.thr_prgr)) {
	return aux_work & ~ERTS_SSI_AUX_WORK_ASYNC_READY_CLEAN;
    }

    awdp->async_ready.need_thr_prgr = 0;
    thr_prgr_p = (void *) &awdp->async_ready.thr_prgr;
#else
    thr_prgr_p = NULL;
#endif

    switch (erts_async_ready_clean(awdp->async_ready.queue, thr_prgr_p)) {
    case ERTS_ASYNC_READY_CLEAN:
	unset_aux_work_flags(awdp->ssi, ERTS_SSI_AUX_WORK_ASYNC_READY_CLEAN);
	return aux_work & ~ERTS_SSI_AUX_WORK_ASYNC_READY_CLEAN;
#ifdef ERTS_SMP
    case ERTS_ASYNC_READY_NEED_THR_PRGR:
	haw_thr_prgr_soft_wakeup(awdp, awdp->async_ready.thr_prgr);
	awdp->async_ready.need_thr_prgr = 1;
	return aux_work & ~ERTS_SSI_AUX_WORK_ASYNC_READY_CLEAN;
#endif
    default:
	return aux_work;
    }
}

#endif /* ERTS_USE_ASYNC_READY_Q */


static ERTS_INLINE erts_aint32_t
handle_fix_alloc(ErtsAuxWorkData *awdp, erts_aint32_t aux_work, int waiting)
{
    ErtsSchedulerSleepInfo *ssi = awdp->ssi;
    erts_aint32_t res;

#ifdef ERTS_DIRTY_SCHEDULERS
    ASSERT(!awdp->esdp || !ERTS_SCHEDULER_IS_DIRTY(awdp->esdp));
#endif
    unset_aux_work_flags(ssi, (ERTS_SSI_AUX_WORK_FIX_ALLOC_LOWER_LIM
			       | ERTS_SSI_AUX_WORK_FIX_ALLOC_DEALLOC));
    aux_work &= ~(ERTS_SSI_AUX_WORK_FIX_ALLOC_LOWER_LIM
		  | ERTS_SSI_AUX_WORK_FIX_ALLOC_DEALLOC);
    res = erts_alloc_fix_alloc_shrink(awdp->sched_id, aux_work);
    if (res) {
	set_aux_work_flags(ssi, res);
	aux_work |= res;
    }

    return aux_work;
}

#ifdef ERTS_SMP

void
erts_alloc_notify_delayed_dealloc(int ix)
{
    ErtsSchedulerData *esdp = erts_get_scheduler_data();
    if (esdp && !ERTS_SCHEDULER_IS_DIRTY(esdp))
	schedule_aux_work_wakeup(&esdp->aux_work_data,
				 ix,
				 ERTS_SSI_AUX_WORK_DD);
    else
	set_aux_work_flags_wakeup_relb(ERTS_SCHED_SLEEP_INFO_IX(ix-1),
				       ERTS_SSI_AUX_WORK_DD);
}

void
erts_alloc_ensure_handle_delayed_dealloc_call(int ix)
{
#ifdef DEBUG
    ErtsSchedulerData *esdp = erts_get_scheduler_data();
    ASSERT(!esdp || (ERTS_SCHEDULER_IS_DIRTY(esdp) || ix == (int) esdp->no));
#endif
    set_aux_work_flags_wakeup_nob(ERTS_SCHED_SLEEP_INFO_IX(ix-1),
				  ERTS_SSI_AUX_WORK_DD);
}

static ERTS_INLINE erts_aint32_t
handle_delayed_dealloc(ErtsAuxWorkData *awdp, erts_aint32_t aux_work, int waiting)
{
    ErtsSchedulerSleepInfo *ssi = awdp->ssi;
    int need_thr_progress = 0;
    ErtsThrPrgrVal wakeup = ERTS_THR_PRGR_INVALID;
    int more_work = 0;

#ifdef ERTS_DIRTY_SCHEDULERS
    ASSERT(!awdp->esdp || !ERTS_SCHEDULER_IS_DIRTY(awdp->esdp));
#endif
    unset_aux_work_flags(ssi, ERTS_SSI_AUX_WORK_DD);
    erts_alloc_scheduler_handle_delayed_dealloc((void *) awdp->esdp,
						&need_thr_progress,
						&wakeup,
						&more_work);
    if (more_work) {
	if (set_aux_work_flags(ssi, ERTS_SSI_AUX_WORK_DD)
	    & ERTS_SSI_AUX_WORK_DD_THR_PRGR) {
	    unset_aux_work_flags(ssi, ERTS_SSI_AUX_WORK_DD_THR_PRGR);
	    aux_work &= ~ERTS_SSI_AUX_WORK_DD_THR_PRGR;
	}
	return aux_work;
    }

    if (need_thr_progress) {
	if (wakeup == ERTS_THR_PRGR_INVALID)
	    wakeup = erts_thr_progress_later(awdp->esdp);
	awdp->dd.thr_prgr = wakeup;
	set_aux_work_flags(ssi, ERTS_SSI_AUX_WORK_DD_THR_PRGR);
	awdp->dd.thr_prgr = wakeup;
	haw_thr_prgr_soft_wakeup(awdp, wakeup);
    }
    else if (awdp->dd.completed_callback) {
	awdp->dd.completed_callback(awdp->dd.completed_arg);
	awdp->dd.completed_callback = NULL;
	awdp->dd.completed_arg = NULL;
    }
    return aux_work & ~ERTS_SSI_AUX_WORK_DD;
}

static ERTS_INLINE erts_aint32_t
handle_delayed_dealloc_thr_prgr(ErtsAuxWorkData *awdp, erts_aint32_t aux_work, int waiting)
{
    ErtsSchedulerSleepInfo *ssi;
    int need_thr_progress;
    int more_work;
    ErtsThrPrgrVal wakeup = ERTS_THR_PRGR_INVALID;
    ErtsThrPrgrVal current = haw_thr_prgr_current(awdp);

#ifdef ERTS_DIRTY_SCHEDULERS
    ASSERT(!awdp->esdp || !ERTS_SCHEDULER_IS_DIRTY(awdp->esdp));
#endif
    if (!erts_thr_progress_has_reached_this(current, awdp->dd.thr_prgr))
	return aux_work & ~ERTS_SSI_AUX_WORK_DD_THR_PRGR;

    ssi = awdp->ssi;
    need_thr_progress = 0;
    more_work = 0;

    erts_alloc_scheduler_handle_delayed_dealloc((void *) awdp->esdp,
						&need_thr_progress,
						&wakeup,
						&more_work);
    if (more_work) {
	set_aux_work_flags(ssi, ERTS_SSI_AUX_WORK_DD);
	unset_aux_work_flags(ssi, ERTS_SSI_AUX_WORK_DD_THR_PRGR);
	return ((aux_work & ~ERTS_SSI_AUX_WORK_DD_THR_PRGR)
		| ERTS_SSI_AUX_WORK_DD);
    }

    if (need_thr_progress) {
	if (wakeup == ERTS_THR_PRGR_INVALID)
	    wakeup = erts_thr_progress_later(awdp->esdp);
	awdp->dd.thr_prgr = wakeup;
	haw_thr_prgr_soft_wakeup(awdp, wakeup);
    }
    else {
	unset_aux_work_flags(ssi, ERTS_SSI_AUX_WORK_DD_THR_PRGR);
	if (awdp->dd.completed_callback) {
	    awdp->dd.completed_callback(awdp->dd.completed_arg);
	    awdp->dd.completed_callback = NULL;
	    awdp->dd.completed_arg = NULL;
	}
    }

    return aux_work & ~ERTS_SSI_AUX_WORK_DD_THR_PRGR;
}

/*
 * Handle scheduled thread progress later operations.
 */
#define ERTS_MAX_THR_PRGR_LATER_OPS 50

static ERTS_INLINE erts_aint32_t
handle_thr_prgr_later_op(ErtsAuxWorkData *awdp, erts_aint32_t aux_work, int waiting)
{
    int lops;
    ErtsThrPrgrVal current = haw_thr_prgr_current(awdp);

#ifdef ERTS_DIRTY_SCHEDULERS
    ASSERT(!awdp->esdp || !ERTS_SCHEDULER_IS_DIRTY(awdp->esdp));
#endif
    for (lops = 0; lops < ERTS_MAX_THR_PRGR_LATER_OPS; lops++) {
	ErtsThrPrgrLaterOp *lop = awdp->later_op.first;
	if (!erts_thr_progress_has_reached_this(current, lop->later))
	    return aux_work & ~ERTS_SSI_AUX_WORK_THR_PRGR_LATER_OP;
	awdp->later_op.first = lop->next;
	if (!awdp->later_op.first) {
	    awdp->later_op.last = NULL;
	}
	lop->func(lop->data);
	if (!awdp->later_op.first) {
	    awdp->later_op.size = thr_prgr_later_cleanup_op_threshold;
	    awdp->later_op.last = NULL;
	    unset_aux_work_flags(awdp->ssi,
				 ERTS_SSI_AUX_WORK_THR_PRGR_LATER_OP);
	    return aux_work & ~ERTS_SSI_AUX_WORK_THR_PRGR_LATER_OP;
	}
    }

    return aux_work;
}

static ERTS_INLINE ErtsThrPrgrVal
enqueue_later_op(ErtsSchedulerData *esdp,
		 void (*later_func)(void *),
		 void *later_data,
		 ErtsThrPrgrLaterOp *lop)
{
    ErtsThrPrgrVal later = erts_thr_progress_later(esdp);
    ASSERT(esdp);

    lop->func = later_func;
    lop->data = later_data;
    lop->later = later;
    lop->next = NULL;
    if (!esdp->aux_work_data.later_op.last)
	esdp->aux_work_data.later_op.first = lop;
    else
	esdp->aux_work_data.later_op.last->next = lop;
    esdp->aux_work_data.later_op.last = lop;
    set_aux_work_flags_wakeup_nob(esdp->ssi,
				  ERTS_SSI_AUX_WORK_THR_PRGR_LATER_OP);
    return later;
}

#endif /* ERTS_SMP */

void
erts_schedule_thr_prgr_later_op(void (*later_func)(void *),
				void *later_data,
				ErtsThrPrgrLaterOp *lop)
{
#ifndef ERTS_SMP
    later_func(later_data);
#else
    ErtsSchedulerData *esdp = erts_get_scheduler_data();
    ErtsThrPrgrVal later = enqueue_later_op(esdp, later_func, later_data, lop);
    haw_thr_prgr_wakeup(&esdp->aux_work_data, later);
#endif
}

void
erts_schedule_thr_prgr_later_cleanup_op(void (*later_func)(void *),
					void *later_data,
					ErtsThrPrgrLaterOp *lop,
					UWord size)
{
#ifndef ERTS_SMP
    later_func(later_data);
#else
    ErtsSchedulerData *esdp = erts_get_scheduler_data();
    ErtsThrPrgrVal later = enqueue_later_op(esdp, later_func, later_data, lop);
    haw_thr_prgr_later_cleanup_op_wakeup(&esdp->aux_work_data, later, size);
#endif
}

#ifdef ERTS_SMP

static erts_atomic32_t completed_dealloc_count;

static void
completed_dealloc(void *vproc)
{
    if (erts_atomic32_dec_read_mb(&completed_dealloc_count) == 0) {
	erts_resume((Process *) vproc, (ErtsProcLocks) 0);
	erts_smp_proc_dec_refc((Process *) vproc);
    }
}

static void
setup_completed_dealloc(void *vproc)
{
    ErtsSchedulerData *esdp = erts_get_scheduler_data();
    ErtsAuxWorkData *awdp = (esdp
			     ? &esdp->aux_work_data
			     : aux_thread_aux_work_data);
    erts_alloc_fix_alloc_shrink(awdp->sched_id, 0);
    set_aux_work_flags_wakeup_nob(awdp->ssi, ERTS_SSI_AUX_WORK_DD);
    awdp->dd.completed_callback = completed_dealloc;
    awdp->dd.completed_arg = vproc;
}

static void
prep_setup_completed_dealloc(void *vproc)
{
    erts_aint32_t count = (erts_aint32_t) (erts_no_schedulers+1);
    if (erts_atomic32_dec_read_mb(&completed_dealloc_count) == count) {
	/* scheduler threads */
	erts_schedule_multi_misc_aux_work(0,
					  erts_no_schedulers,
					  setup_completed_dealloc,
					  vproc);
	/* aux_thread */
	erts_schedule_misc_aux_work(0,
				    setup_completed_dealloc,
				    vproc);
    }
}

#endif /* ERTS_SMP */

int
erts_debug_wait_deallocations(Process *c_p)
{
#ifndef ERTS_SMP
    erts_alloc_fix_alloc_shrink(1, 0);
    return 1;
#else
    /* Only one process at a time can do this */
    erts_aint32_t count = (erts_aint32_t) (2*(erts_no_schedulers+1));
    if (0 == erts_atomic32_cmpxchg_mb(&completed_dealloc_count,
				      count,
				      0)) {
	erts_suspend(c_p, ERTS_PROC_LOCK_MAIN, NULL);
	erts_smp_proc_inc_refc(c_p);
	/* scheduler threads */
	erts_schedule_multi_misc_aux_work(0,
					  erts_no_schedulers,
					  prep_setup_completed_dealloc,
					  (void *) c_p);
	/* aux_thread */
	erts_schedule_misc_aux_work(0,
				    prep_setup_completed_dealloc,
				    (void *) c_p);
	return 1;
    }
    return 0;
#endif
}


#ifdef ERTS_SMP_SCHEDULERS_NEED_TO_CHECK_CHILDREN
void
erts_smp_notify_check_children_needed(void)
{
    int i;
    for (i = 0; i < erts_no_schedulers; i++)
	set_aux_work_flags_wakeup_nob(ERTS_SCHED_SLEEP_INFO_IX(i),
				      ERTS_SSI_AUX_WORK_CHECK_CHILDREN);
#ifdef ERTS_DIRTY_SCHEDULERS
    for (i = 0; i < erts_no_dirty_cpu_schedulers; i++)
	set_aux_work_flags_wakeup_nob(ERTS_DIRTY_CPU_SCHED_SLEEP_INFO_IX(i),
				      ERTS_SSI_AUX_WORK_CHECK_CHILDREN);
    for (i = 0; i < erts_no_dirty_io_schedulers; i++)
	set_aux_work_flags_wakeup_nob(ERTS_DIRTY_IO_SCHED_SLEEP_INFO_IX(i),
				      ERTS_SSI_AUX_WORK_CHECK_CHILDREN);
#endif
}

static ERTS_INLINE erts_aint32_t
handle_check_children(ErtsAuxWorkData *awdp, erts_aint32_t aux_work, int waiting)
{
    unset_aux_work_flags(awdp->ssi, ERTS_SSI_AUX_WORK_CHECK_CHILDREN);
    erts_check_children();
    return aux_work & ~ERTS_SSI_AUX_WORK_CHECK_CHILDREN;
}

#endif

static void
notify_reap_ports_relb(void)
{
    int i;
    for (i = 0; i < erts_no_schedulers; i++) {
	set_aux_work_flags_wakeup_relb(ERTS_SCHED_SLEEP_INFO_IX(i),
				       ERTS_SSI_AUX_WORK_REAP_PORTS);
    }
}

erts_smp_atomic32_t erts_halt_progress;
int erts_halt_code;

static ERTS_INLINE erts_aint32_t
handle_reap_ports(ErtsAuxWorkData *awdp, erts_aint32_t aux_work, int waiting)
{
    unset_aux_work_flags(awdp->ssi, ERTS_SSI_AUX_WORK_REAP_PORTS);
    awdp->esdp->run_queue->halt_in_progress = 1;
    if (erts_smp_atomic32_dec_read_acqb(&erts_halt_progress) == 0) {
	int i, max = erts_ptab_max(&erts_port);
	erts_smp_atomic32_set_nob(&erts_halt_progress, 1);
	for (i = 0; i < max; i++) {
	    erts_aint32_t state;
	    Port *prt = erts_pix2port(i);
	    if (!prt)
		continue;
	    state = erts_atomic32_read_acqb(&prt->state);
	    if (state & (ERTS_PORT_SFLGS_INVALID_DRIVER_LOOKUP
			 | ERTS_PORT_SFLG_HALT))
		continue;

	    /* We need to set the halt flag - get the port lock */

	    erts_smp_port_lock(prt);

	    state = erts_atomic32_read_nob(&prt->state);
	    if (!(state & (ERTS_PORT_SFLGS_INVALID_DRIVER_LOOKUP
			   | ERTS_PORT_SFLG_HALT))) {
		state = erts_atomic32_read_bor_relb(&prt->state,
						    ERTS_PORT_SFLG_HALT);
		erts_smp_atomic32_inc_nob(&erts_halt_progress);
		if (!(state & (ERTS_PORT_SFLG_EXITING|ERTS_PORT_SFLG_CLOSING)))
		    erts_deliver_port_exit(prt, prt->common.id, am_killed, 0);
	    }

	    erts_port_release(prt);
	}
	if (erts_smp_atomic32_dec_read_nob(&erts_halt_progress) == 0) {
	    erl_exit_flush_async(erts_halt_code, "");
	}
    }
    return aux_work & ~ERTS_SSI_AUX_WORK_REAP_PORTS;
}

#if HAVE_ERTS_MSEG

static ERTS_INLINE erts_aint32_t
handle_mseg_cache_check(ErtsAuxWorkData *awdp, erts_aint32_t aux_work, int waiting)
{
    unset_aux_work_flags(awdp->ssi, ERTS_SSI_AUX_WORK_MSEG_CACHE_CHECK);
    erts_mseg_cache_check();
    return aux_work & ~ERTS_SSI_AUX_WORK_MSEG_CACHE_CHECK;
}

#endif

static ERTS_INLINE erts_aint32_t
handle_setup_aux_work_timer(ErtsAuxWorkData *awdp, erts_aint32_t aux_work, int waiting)
{
    unset_aux_work_flags(awdp->ssi, ERTS_SSI_AUX_WORK_SET_TMO);
    setup_aux_work_timer();
    return aux_work & ~ERTS_SSI_AUX_WORK_SET_TMO;
}

static erts_aint32_t
handle_aux_work(ErtsAuxWorkData *awdp, erts_aint32_t orig_aux_work, int waiting)
{
#undef HANDLE_AUX_WORK
#define HANDLE_AUX_WORK(FLG, HNDLR) \
    ignore |= FLG; \
    if (aux_work & FLG) { \
	aux_work = HNDLR(awdp, aux_work, waiting);   \
	ERTS_DBG_CHK_AUX_WORK_VAL(aux_work); \
	if (!(aux_work & ~ignore)) { \
	    ERTS_DBG_CHK_AUX_WORK_VAL(aux_work); \
	    return aux_work; \
	} \
    }

    erts_aint32_t aux_work = orig_aux_work;
    erts_aint32_t ignore = 0;

#ifdef ERTS_SMP
    haw_thr_prgr_current_reset(awdp);
#endif

    ERTS_DBG_CHK_AUX_WORK_VAL(aux_work);
    ASSERT(aux_work);

    /*
     * Handlers are *only* allowed to modify flags in return value
     * and ssi flags that are explicity handled by the handler.
     * Handlers are, e.g., not allowed to read the ssi flag field and
     * then unconditionally return that value.
     *
     * Flag field returned should only contain flags for work that
     * can continue immediately.
     */

    /*
     * Keep ERTS_SSI_AUX_WORK flags in expected frequency order relative
     * eachother. Most frequent first.
     */
#ifdef ERTS_SMP
    HANDLE_AUX_WORK(ERTS_SSI_AUX_WORK_DELAYED_AW_WAKEUP,
		    handle_delayed_aux_work_wakeup);
    HANDLE_AUX_WORK(ERTS_SSI_AUX_WORK_DD,
		    handle_delayed_dealloc);
    /* DD must be before DD_THR_PRGR */
    HANDLE_AUX_WORK(ERTS_SSI_AUX_WORK_DD_THR_PRGR,
		    handle_delayed_dealloc_thr_prgr);
#endif

    HANDLE_AUX_WORK((ERTS_SSI_AUX_WORK_FIX_ALLOC_LOWER_LIM
		     | ERTS_SSI_AUX_WORK_FIX_ALLOC_DEALLOC),
		    handle_fix_alloc);

#ifdef ERTS_SMP
    HANDLE_AUX_WORK(ERTS_SSI_AUX_WORK_THR_PRGR_LATER_OP,
		    handle_thr_prgr_later_op);
#endif

#if ERTS_USE_ASYNC_READY_Q
    HANDLE_AUX_WORK(ERTS_SSI_AUX_WORK_ASYNC_READY,
		    handle_async_ready);
    /* ASYNC_READY must be before ASYNC_READY_CLEAN */
    HANDLE_AUX_WORK(ERTS_SSI_AUX_WORK_ASYNC_READY_CLEAN,
		    handle_async_ready_clean);
#endif

#ifdef ERTS_SMP
    HANDLE_AUX_WORK(ERTS_SSI_AUX_WORK_MISC_THR_PRGR,
		    handle_misc_aux_work_thr_prgr);
#endif
    /* MISC_THR_PRGR must be before MISC */
    HANDLE_AUX_WORK(ERTS_SSI_AUX_WORK_MISC,
		    handle_misc_aux_work);

#ifdef ERTS_SMP_SCHEDULERS_NEED_TO_CHECK_CHILDREN
    HANDLE_AUX_WORK(ERTS_SSI_AUX_WORK_CHECK_CHILDREN,
		    handle_check_children);
#endif

    HANDLE_AUX_WORK(ERTS_SSI_AUX_WORK_SET_TMO,
		    handle_setup_aux_work_timer);

#if HAVE_ERTS_MSEG
    HANDLE_AUX_WORK(ERTS_SSI_AUX_WORK_MSEG_CACHE_CHECK,
		    handle_mseg_cache_check);
#endif

    HANDLE_AUX_WORK(ERTS_SSI_AUX_WORK_REAP_PORTS,
		    handle_reap_ports);

    ERTS_DBG_CHK_AUX_WORK_VAL(aux_work);

#ifdef ERTS_SMP
    if (waiting && !aux_work)
	haw_thr_prgr_current_check_progress(awdp);
#endif

    return aux_work;

#undef HANDLE_AUX_WORK

}

typedef struct {
    union {
	ErlTimer data;
	char align__[ERTS_ALC_CACHE_LINE_ALIGN_SIZE(sizeof(ErlTimer))];
    } timer;

    int initialized;
    erts_atomic32_t refc;
    erts_atomic32_t type[1];
} ErtsAuxWorkTmo;

static ErtsAuxWorkTmo *aux_work_tmo;

static void
aux_work_timeout_early_init(int no_schedulers)
{
    int i;
    UWord p;

    /*
     * This is done really early. Our own allocators have
     * not been started yet.
     */

    p = (UWord) malloc((sizeof(ErtsAuxWorkTmo)
			+ sizeof(erts_atomic32_t)*(no_schedulers+1))
		       + ERTS_CACHE_LINE_SIZE-1);
    if (!p) {
        ERTS_INTERNAL_ERROR("malloc failed to allocate memory!");
    }
    if (p & ERTS_CACHE_LINE_MASK)
	p = (p & ~ERTS_CACHE_LINE_MASK) + ERTS_CACHE_LINE_SIZE;
    ASSERT((p & ERTS_CACHE_LINE_MASK) == 0);

    aux_work_tmo = (ErtsAuxWorkTmo *) p;
    aux_work_tmo->initialized = 0;
    erts_atomic32_init_nob(&aux_work_tmo->refc, 0);
    for (i = 0; i <= no_schedulers; i++)
	erts_atomic32_init_nob(&aux_work_tmo->type[i], 0);
}

void
aux_work_timeout_late_init(void)
{
    aux_work_tmo->initialized = 1;
    if (erts_atomic32_read_nob(&aux_work_tmo->refc)) {
	aux_work_tmo->timer.data.active = 0;
	erts_set_timer(&aux_work_tmo->timer.data,
		       aux_work_timeout,
		       NULL,
		       NULL,
		       1000);
    }
}

static void
aux_work_timeout(void *unused)
{
    erts_aint32_t refc;
    int i;
#ifdef ERTS_SMP
    i = 0;
#else
    i = 1;
#endif

    for (; i <= erts_no_schedulers; i++) {
	erts_aint32_t type;
	type = erts_atomic32_read_acqb(&aux_work_tmo->type[i]);
	if (type)
	    set_aux_work_flags_wakeup_nob(ERTS_SCHED_SLEEP_INFO_IX(i-1),
					  type);
    }

    refc = erts_atomic32_read_nob(&aux_work_tmo->refc);
    ASSERT(refc >= 1);
    if (refc != 1
	|| 1 != erts_atomic32_cmpxchg_relb(&aux_work_tmo->refc, 0, 1)) {
	/* Setup next timeout... */
	aux_work_tmo->timer.data.active = 0;
	erts_set_timer(&aux_work_tmo->timer.data,
		       aux_work_timeout,
		       NULL,
		       NULL,
		       1000);
    }
}

static void
setup_aux_work_timer(void)
{
#ifndef ERTS_SMP
    if (!erts_get_scheduler_data())
	set_aux_work_flags_wakeup_nob(ERTS_SCHED_SLEEP_INFO_IX(0),
				      ERTS_SSI_AUX_WORK_SET_TMO);
    else
#endif
    {
	aux_work_tmo->timer.data.active = 0;
	erts_set_timer(&aux_work_tmo->timer.data,
		       aux_work_timeout,
		       NULL,
		       NULL,
		       1000);
    }
}

erts_aint32_t
erts_set_aux_work_timeout(int ix, erts_aint32_t type, int enable)
{
    erts_aint32_t old, refc;

#ifndef ERTS_SMP
    ix = 1;
#endif

    ERTS_DBG_CHK_AUX_WORK_VAL(type);
    ERTS_DBG_CHK_AUX_WORK_VAL(erts_atomic32_read_nob(&aux_work_tmo->type[ix]));
    /* erts_fprintf(stderr, "t(%d, 0x%x, %d)\n", ix, type, enable); */

    if (!enable) {
	old = erts_atomic32_read_band_mb(&aux_work_tmo->type[ix], ~type);
	ERTS_DBG_CHK_AUX_WORK_VAL(erts_atomic32_read_nob(&aux_work_tmo->type[ix]));
	if (old != 0 && (old & ~type) == 0)
	    erts_atomic32_dec_relb(&aux_work_tmo->refc);
	return old;
    }

    old = erts_atomic32_read_bor_mb(&aux_work_tmo->type[ix], type);
    ERTS_DBG_CHK_AUX_WORK_VAL(erts_atomic32_read_nob(&aux_work_tmo->type[ix]));
    if (old == 0 && type != 0) {
	refc = erts_atomic32_inc_read_acqb(&aux_work_tmo->refc);
	if (refc == 1) {
	    erts_atomic32_inc_acqb(&aux_work_tmo->refc);
	    if (aux_work_tmo->initialized) 
		setup_aux_work_timer();
	}
    }
    return old;
}



static ERTS_INLINE void
sched_waiting_sys(Uint no, ErtsRunQueue *rq)
{
    ERTS_SMP_LC_ASSERT(erts_smp_lc_runq_is_locked(rq));
    ASSERT(rq->waiting >= 0);
    (void) ERTS_RUNQ_FLGS_SET(rq, (ERTS_RUNQ_FLG_OUT_OF_WORK
				   | ERTS_RUNQ_FLG_HALFTIME_OUT_OF_WORK));
    rq->waiting++;
    rq->waiting *= -1;
    rq->woken = 0;
    if (erts_system_profile_flags.scheduler)
	profile_scheduler(make_small(no), am_inactive);
}

static ERTS_INLINE void
sched_active_sys(Uint no, ErtsRunQueue *rq)
{
    ERTS_SMP_LC_ASSERT(erts_smp_lc_runq_is_locked(rq));
#ifdef ERTS_DIRTY_SCHEDULERS
    ASSERT(!ERTS_RUNQ_IX_IS_DIRTY(rq->ix));
#endif
    ASSERT(rq->waiting < 0);
    rq->waiting *= -1;
    rq->waiting--;
    if (erts_system_profile_flags.scheduler)
	profile_scheduler(make_small(no), am_active);
}

Uint
erts_active_schedulers(void)
{
    Uint as = erts_no_schedulers;

    ERTS_ATOMIC_FOREACH_RUNQ(rq, as -= abs(rq->waiting));

    ASSERT(as >= 0);
    return as;
}

#ifdef ERTS_SMP

static ERTS_INLINE void
clear_sys_scheduling(void)
{
    erts_smp_atomic32_set_mb(&doing_sys_schedule, 0);
}

static ERTS_INLINE int
try_set_sys_scheduling(void)
{
    return 0 == erts_smp_atomic32_cmpxchg_acqb(&doing_sys_schedule, 1, 0);
}

#endif

static ERTS_INLINE int
prepare_for_sys_schedule(ErtsSchedulerData *esdp)
{
#ifdef ERTS_SMP
    while (!erts_port_task_have_outstanding_io_tasks()
	   && try_set_sys_scheduling()) {
#ifdef ERTS_SCHED_ONLY_POLL_SCHED_1
      if (esdp->no != 1) {
	/* If we are not scheduler 1 and ERTS_SCHED_ONLY_POLL_SCHED_1 is used
	   then we make sure to wake scheduler 1 */
	ErtsRunQueue *rq = ERTS_RUNQ_IX(0);
	clear_sys_scheduling();
	wake_scheduler(rq);
	return 0;
      }
#endif
      if (!erts_port_task_have_outstanding_io_tasks())
	return 1;
      clear_sys_scheduling();
    }
    return 0;
#else
    return !erts_port_task_have_outstanding_io_tasks();
#endif
}

#ifdef ERTS_SMP

static ERTS_INLINE void
sched_change_waiting_sys_to_waiting(Uint no, ErtsRunQueue *rq)
{
    ERTS_SMP_LC_ASSERT(erts_smp_lc_runq_is_locked(rq));
#ifdef ERTS_DIRTY_SCHEDULERS
    ASSERT(!ERTS_RUNQ_IX_IS_DIRTY(rq->ix));
#endif
    ASSERT(rq->waiting < 0);
    rq->waiting *= -1;
}

static ERTS_INLINE void
sched_waiting(Uint no, ErtsRunQueue *rq)
{
    ERTS_SMP_LC_ASSERT(erts_smp_lc_runq_is_locked(rq));
    (void) ERTS_RUNQ_FLGS_SET(rq, (ERTS_RUNQ_FLG_OUT_OF_WORK
				   | ERTS_RUNQ_FLG_HALFTIME_OUT_OF_WORK));
    if (rq->waiting < 0)
	rq->waiting--;
    else
	rq->waiting++;
    rq->woken = 0;
    if (!ERTS_RUNQ_IX_IS_DIRTY(rq->ix) && erts_system_profile_flags.scheduler)
	profile_scheduler(make_small(no), am_inactive);
}

static ERTS_INLINE void
sched_active(Uint no, ErtsRunQueue *rq)
{
    ERTS_SMP_LC_ASSERT(erts_smp_lc_runq_is_locked(rq));
    if (rq->waiting < 0)
	rq->waiting++;
    else
	rq->waiting--;
    if (!ERTS_RUNQ_IX_IS_DIRTY(rq->ix) && erts_system_profile_flags.scheduler)
	profile_scheduler(make_small(no), am_active);
}

static int ERTS_INLINE
ongoing_multi_scheduling_block(void)
{
    ERTS_SMP_LC_ASSERT(erts_lc_mtx_is_locked(&schdlr_sspnd.mtx));
    return schdlr_sspnd.msb.ongoing;
}

static ERTS_INLINE void
empty_runq_aux(ErtsRunQueue *rq, Uint32 old_flags)
{
    if (!ERTS_RUNQ_IX_IS_DIRTY(rq->ix) && old_flags & ERTS_RUNQ_FLG_NONEMPTY) {
#ifdef DEBUG
	erts_aint32_t empty = erts_smp_atomic32_read_nob(&no_empty_run_queues);
	/*
	 * For a short period of time no_empty_run_queues may have
	 * been increased twice for a specific run queue.
	 */
	ASSERT(0 <= empty && empty < 2*erts_no_run_queues);
#endif
	if (!erts_runq_supervision_interval)
	    erts_smp_atomic32_inc_relb(&no_empty_run_queues);
	else {
	    erts_smp_atomic32_inc_mb(&no_empty_run_queues);
	    if (erts_atomic_read_nob(&runq_supervisor_sleeping))
		ethr_event_set(&runq_supervision_event);
	}
    }
}

static ERTS_INLINE void
empty_runq(ErtsRunQueue *rq)
{
    Uint32 old_flags = ERTS_RUNQ_FLGS_UNSET(rq, ERTS_RUNQ_FLG_NONEMPTY|ERTS_RUNQ_FLG_PROTECTED);
    empty_runq_aux(rq, old_flags);
}

static ERTS_INLINE Uint32
empty_protected_runq(ErtsRunQueue *rq)
{
    Uint32 old_flags = ERTS_RUNQ_FLGS_BSET(rq,
					   ERTS_RUNQ_FLG_NONEMPTY|ERTS_RUNQ_FLG_PROTECTED,
					   ERTS_RUNQ_FLG_PROTECTED);
    empty_runq_aux(rq, old_flags);
    return old_flags;
}

static ERTS_INLINE void
non_empty_runq(ErtsRunQueue *rq)
{
    Uint32 old_flags = ERTS_RUNQ_FLGS_SET(rq, ERTS_RUNQ_FLG_NONEMPTY);
    if (!ERTS_RUNQ_IX_IS_DIRTY(rq->ix) && (!(old_flags & ERTS_RUNQ_FLG_NONEMPTY))) {
#ifdef DEBUG
	erts_aint32_t empty = erts_smp_atomic32_read_nob(&no_empty_run_queues);
	/*
	 * For a short period of time no_empty_run_queues may have
	 * been increased twice for a specific run queue.
	 */
	ASSERT(0 < empty && empty <= 2*erts_no_run_queues);
#endif
	if (!erts_runq_supervision_interval)
	    erts_smp_atomic32_dec_relb(&no_empty_run_queues);
	else {
	    erts_aint32_t no;
	    no = erts_smp_atomic32_dec_read_mb(&no_empty_run_queues);
	    if (no > 0 && erts_atomic_read_nob(&runq_supervisor_sleeping))
		ethr_event_set(&runq_supervision_event);
	}
    }
}

void
erts_empty_runq(ErtsRunQueue *rq)
{
    empty_runq(rq);
}

void
erts_non_empty_runq(ErtsRunQueue *rq)
{
    non_empty_runq(rq);
}

static erts_aint32_t
sched_prep_spin_wait(ErtsSchedulerSleepInfo *ssi)
{
    erts_aint32_t oflgs;
    erts_aint32_t nflgs = (ERTS_SSI_FLG_SLEEPING
			   | ERTS_SSI_FLG_WAITING);
    erts_aint32_t xflgs = 0;

    do {
	oflgs = erts_smp_atomic32_cmpxchg_acqb(&ssi->flags, nflgs, xflgs);
	if (oflgs == xflgs)
	    return nflgs;
	xflgs = oflgs;
    } while (!(oflgs & ERTS_SSI_FLG_SUSPENDED));
    return oflgs;
}

static erts_aint32_t
sched_prep_cont_spin_wait(ErtsSchedulerSleepInfo *ssi)
{
    erts_aint32_t oflgs;
    erts_aint32_t nflgs = (ERTS_SSI_FLG_SLEEPING
			   | ERTS_SSI_FLG_WAITING);
    erts_aint32_t xflgs = ERTS_SSI_FLG_WAITING;

    do {
	oflgs = erts_smp_atomic32_cmpxchg_acqb(&ssi->flags, nflgs, xflgs);
	if (oflgs == xflgs)
	    return nflgs;
	xflgs = oflgs;
	nflgs |= oflgs & ERTS_SSI_FLG_SUSPENDED;
    } while (oflgs & ERTS_SSI_FLG_WAITING);
    return oflgs;
}

static erts_aint32_t
sched_spin_wait(ErtsSchedulerSleepInfo *ssi, int spincount)
{
    int until_yield = ERTS_SCHED_SPIN_UNTIL_YIELD;
    int sc = spincount;
    erts_aint32_t flgs;

    do {
	flgs = erts_smp_atomic32_read_acqb(&ssi->flags);
	if ((flgs & (ERTS_SSI_FLG_SLEEPING|ERTS_SSI_FLG_WAITING))
	    != (ERTS_SSI_FLG_SLEEPING|ERTS_SSI_FLG_WAITING)) {
	    break;
	}
	ERTS_SPIN_BODY;
	if (--until_yield == 0) {
	    until_yield = ERTS_SCHED_SPIN_UNTIL_YIELD;
	    erts_thr_yield();
	}
    } while (--sc > 0);
    return flgs;
}

static erts_aint32_t
sched_set_sleeptype(ErtsSchedulerSleepInfo *ssi, erts_aint32_t sleep_type)
{
    erts_aint32_t oflgs;
    erts_aint32_t nflgs = ERTS_SSI_FLG_SLEEPING|ERTS_SSI_FLG_WAITING|sleep_type;
    erts_aint32_t xflgs = ERTS_SSI_FLG_SLEEPING|ERTS_SSI_FLG_WAITING;

    if (sleep_type == ERTS_SSI_FLG_TSE_SLEEPING)
	erts_tse_reset(ssi->event);
    else {
	ASSERT(sleep_type == ERTS_SSI_FLG_POLL_SLEEPING);
	erts_sys_schedule_interrupt(0);
    }

    while (1) {
	oflgs = erts_smp_atomic32_cmpxchg_acqb(&ssi->flags, nflgs, xflgs);
	if (oflgs == xflgs)
	    return nflgs;
	if ((oflgs & (ERTS_SSI_FLG_SLEEPING|ERTS_SSI_FLG_WAITING))
	    != (ERTS_SSI_FLG_SLEEPING|ERTS_SSI_FLG_WAITING)) {
	    return oflgs;
	}
	xflgs = oflgs;
	nflgs |= oflgs & ERTS_SSI_FLG_SUSPENDED;
    }
}

#define ERTS_SCHED_WAIT_WOKEN(FLGS)				\
  (((FLGS) & (ERTS_SSI_FLG_WAITING|ERTS_SSI_FLG_SUSPENDED))	\
   != ERTS_SSI_FLG_WAITING)


static void
thr_prgr_wakeup(void *vssi)
{
    erts_sched_poke((ErtsSchedulerSleepInfo *) vssi);
}

static void
thr_prgr_prep_wait(void *vssi)
{
    ErtsSchedulerSleepInfo *ssi = (ErtsSchedulerSleepInfo *) vssi;
    erts_smp_atomic32_read_bor_acqb(&ssi->flags,
				    ERTS_SSI_FLG_SLEEPING);
}

static void
thr_prgr_wait(void *vssi)
{
    ErtsSchedulerSleepInfo *ssi = (ErtsSchedulerSleepInfo *) vssi;
    erts_aint32_t xflgs = ERTS_SSI_FLG_SLEEPING;

    erts_tse_reset(ssi->event);

    while (1) {
	erts_aint32_t aflgs, nflgs;
	nflgs = xflgs | ERTS_SSI_FLG_TSE_SLEEPING;
	aflgs = erts_smp_atomic32_cmpxchg_acqb(&ssi->flags, nflgs, xflgs);
	if (aflgs == xflgs) {
	    erts_tse_wait(ssi->event);
	    break;
	}
	if ((aflgs & ERTS_SSI_FLG_SLEEPING) == 0)
	    break;
	xflgs = aflgs;
    }
}

static void
thr_prgr_fin_wait(void *vssi)
{
    ErtsSchedulerSleepInfo *ssi = (ErtsSchedulerSleepInfo *) vssi;
    erts_smp_atomic32_read_band_nob(&ssi->flags,
				    ~(ERTS_SSI_FLG_SLEEPING
				      | ERTS_SSI_FLG_TSE_SLEEPING));
}

static void init_aux_work_data(ErtsAuxWorkData *awdp, ErtsSchedulerData *esdp, char *dawwp);

static void *
aux_thread(void *unused)
{
    ErtsAuxWorkData *awdp = aux_thread_aux_work_data;
    ErtsSchedulerSleepInfo *ssi = ERTS_SCHED_SLEEP_INFO_IX(-1);
    erts_aint32_t aux_work;
    ErtsThrPrgrCallbacks callbacks;
    int thr_prgr_active = 1;

#ifdef ERTS_ENABLE_LOCK_CHECK
    {
	char buf[] = "aux_thread";
	erts_lc_set_thread_name(buf);
    }
#endif

    ssi->event = erts_tse_fetch();

    callbacks.arg = (void *) ssi;
    callbacks.wakeup = thr_prgr_wakeup;
    callbacks.prepare_wait = thr_prgr_prep_wait;
    callbacks.wait = thr_prgr_wait;
    callbacks.finalize_wait = thr_prgr_fin_wait;

    erts_thr_progress_register_managed_thread(NULL, &callbacks, 1);
    init_aux_work_data(awdp, NULL, NULL);
    awdp->ssi = ssi;

    sched_prep_spin_wait(ssi);

    while (1) {
	erts_aint32_t flgs;

	aux_work = erts_atomic32_read_acqb(&ssi->aux_work);
	if (aux_work) {
	    if (!thr_prgr_active)
		erts_thr_progress_active(NULL, thr_prgr_active = 1);
	    aux_work = handle_aux_work(awdp, aux_work, 1);
	    if (aux_work && erts_thr_progress_update(NULL))
		erts_thr_progress_leader_update(NULL);
	}

	if (!aux_work) {
	    if (thr_prgr_active)
		erts_thr_progress_active(NULL, thr_prgr_active = 0);
	    erts_thr_progress_prepare_wait(NULL);

	    ERTS_SCHED_FAIR_YIELD();

	    flgs = sched_spin_wait(ssi, 0);

	    if (flgs & ERTS_SSI_FLG_SLEEPING) {
		ASSERT(flgs & ERTS_SSI_FLG_WAITING);
		flgs = sched_set_sleeptype(ssi, ERTS_SSI_FLG_TSE_SLEEPING);
		if (flgs & ERTS_SSI_FLG_SLEEPING) {
		    int res;
		    ASSERT(flgs & ERTS_SSI_FLG_TSE_SLEEPING);
		    ASSERT(flgs & ERTS_SSI_FLG_WAITING);
		    do {
			res = erts_tse_wait(ssi->event);
		    } while (res == EINTR);
		}
	    }
	    erts_thr_progress_finalize_wait(NULL);
	}

	flgs = sched_prep_spin_wait(ssi);
    }
    return NULL;
}

#endif /* ERTS_SMP */

static void
scheduler_wait(int *fcalls, ErtsSchedulerData *esdp, ErtsRunQueue *rq)
{
    int working = 1;
    ErtsSchedulerSleepInfo *ssi = esdp->ssi;
    int spincount;
    erts_aint32_t aux_work = 0;
#ifdef ERTS_SMP
    int thr_prgr_active = 1;
    erts_aint32_t flgs;

    ERTS_SMP_LC_ASSERT(erts_smp_lc_runq_is_locked(rq));

#ifdef ERTS_DIRTY_SCHEDULERS
    if (ERTS_RUNQ_IX_IS_DIRTY(rq->ix))
	erts_smp_spin_lock(&rq->sleepers.lock);
#endif
    flgs = sched_prep_spin_wait(ssi);
    if (flgs & ERTS_SSI_FLG_SUSPENDED) {
	/* Go suspend instead... */
#ifdef ERTS_DIRTY_SCHEDULERS
	if (ERTS_RUNQ_IX_IS_DIRTY(rq->ix))
	    erts_smp_spin_unlock(&rq->sleepers.lock);
#endif
	return;
    }

#ifdef ERTS_DIRTY_SCHEDULERS
    if (ERTS_RUNQ_IX_IS_DIRTY(rq->ix)) {
	ssi->prev = NULL;
	ssi->next = rq->sleepers.list;
	if (rq->sleepers.list)
	    rq->sleepers.list->prev = ssi;
	rq->sleepers.list = ssi;
	erts_smp_spin_unlock(&rq->sleepers.lock);
    }
#endif

    /*
     * If all schedulers are waiting, one of them *should*
     * be waiting in erl_sys_schedule()
     */

    if (ERTS_SCHEDULER_IS_DIRTY(esdp) || !prepare_for_sys_schedule(esdp)) {

	sched_waiting(esdp->no, rq);

	erts_smp_runq_unlock(rq);

	spincount = sched_busy_wait.tse;

    tse_wait:

	if (!ERTS_SCHEDULER_IS_DIRTY(esdp) && thr_prgr_active != working)
	    sched_wall_time_change(esdp, thr_prgr_active);

	while (1) {

	    aux_work = erts_atomic32_read_acqb(&ssi->aux_work);
	    if (aux_work) {
		if (!ERTS_SCHEDULER_IS_DIRTY(esdp) && !thr_prgr_active) {
		    erts_thr_progress_active(esdp, thr_prgr_active = 1);
		    sched_wall_time_change(esdp, 1);
		}
		aux_work = handle_aux_work(&esdp->aux_work_data, aux_work, 1);
		if (aux_work && !ERTS_SCHEDULER_IS_DIRTY(esdp)
		    && erts_thr_progress_update(esdp))
		    erts_thr_progress_leader_update(esdp);
	    }

	    if (aux_work)
		flgs = erts_smp_atomic32_read_acqb(&ssi->flags);
	    else {
		if (!ERTS_SCHEDULER_IS_DIRTY(esdp)) {
		    if (thr_prgr_active) {
			erts_thr_progress_active(esdp, thr_prgr_active = 0);
			sched_wall_time_change(esdp, 0);
		    }
		    erts_thr_progress_prepare_wait(esdp);
		}

		ERTS_SCHED_FAIR_YIELD();

		flgs = sched_spin_wait(ssi, spincount);
		if (flgs & ERTS_SSI_FLG_SLEEPING) {
		    ASSERT(flgs & ERTS_SSI_FLG_WAITING);
		    flgs = sched_set_sleeptype(ssi, ERTS_SSI_FLG_TSE_SLEEPING);
		    if (flgs & ERTS_SSI_FLG_SLEEPING) {
			int res;
			ASSERT(flgs & ERTS_SSI_FLG_TSE_SLEEPING);
			ASSERT(flgs & ERTS_SSI_FLG_WAITING);
			do {
			    res = erts_tse_wait(ssi->event);
			} while (res == EINTR);
		    }
		}
		if (!ERTS_SCHEDULER_IS_DIRTY(esdp))
		    erts_thr_progress_finalize_wait(esdp);
	    }

	    if (!(flgs & ERTS_SSI_FLG_WAITING)) {
		ASSERT(!(flgs & ERTS_SSI_FLG_SLEEPING));
		break;
	    }

	    flgs = sched_prep_cont_spin_wait(ssi);
	    spincount = sched_busy_wait.aux_work;

	    if (!(flgs & ERTS_SSI_FLG_WAITING)) {
		ASSERT(!(flgs & ERTS_SSI_FLG_SLEEPING));
		break;
	    }

	}

	if (flgs & ~ERTS_SSI_FLG_SUSPENDED)
	    erts_smp_atomic32_read_band_nob(&ssi->flags, ERTS_SSI_FLG_SUSPENDED);

	if (!ERTS_SCHEDULER_IS_DIRTY(esdp) && !thr_prgr_active) {
	    erts_thr_progress_active(esdp, thr_prgr_active = 1);
	    sched_wall_time_change(esdp, 1);
	}

	erts_smp_runq_lock(rq);
	sched_active(esdp->no, rq);

    }
    else
#endif
    {
	erts_aint_t dt;

	erts_smp_atomic32_set_relb(&function_calls, 0);
	*fcalls = 0;

#ifdef ERTS_DIRTY_SCHEDULERS
	ASSERT(!ERTS_SCHEDULER_IS_DIRTY(esdp));
#endif

#ifdef ERTS_SCHED_ONLY_POLL_SCHED_1
	ASSERT(esdp->no == 1);
#endif
	sched_waiting_sys(esdp->no, rq);


	erts_smp_runq_unlock(rq);

	ASSERT(working);
	sched_wall_time_change(esdp, working = 0);

	spincount = sched_busy_wait.sys_schedule;
	if (spincount == 0)
	    goto sys_aux_work;

	while (spincount-- > 0) {

	sys_poll_aux_work:

	    if (working)
		sched_wall_time_change(esdp, working = 0);

	    ASSERT(!erts_port_task_have_outstanding_io_tasks());
	    erl_sys_schedule(1); /* Might give us something to do */

	    dt = erts_do_time_read_and_reset();
	    if (dt) erts_bump_timer(dt);

	sys_aux_work:
#ifndef ERTS_SMP
	    erts_sys_schedule_interrupt(0);
#endif

	    aux_work = erts_atomic32_read_acqb(&ssi->aux_work);
	    if (aux_work) {
		if (!working)
		    sched_wall_time_change(esdp, working = 1);
#ifdef ERTS_SMP
		if (!thr_prgr_active)
		    erts_thr_progress_active(esdp, thr_prgr_active = 1);
#endif
		aux_work = handle_aux_work(&esdp->aux_work_data, aux_work, 1);
#ifdef ERTS_SMP
		if (aux_work && erts_thr_progress_update(esdp))
		    erts_thr_progress_leader_update(esdp);
#endif
	    }

#ifndef ERTS_SMP
	    if (rq->len != 0 || rq->misc.start)
		goto sys_woken;
#else
	    flgs = erts_smp_atomic32_read_acqb(&ssi->flags);
	    if (!(flgs & ERTS_SSI_FLG_WAITING)) {
		ASSERT(!(flgs & ERTS_SSI_FLG_SLEEPING));
		goto sys_woken;
	    }

	    /*
	     * If we got new I/O tasks we aren't allowed to
	     * call erl_sys_schedule() until it is handled.
	     */
	    if (erts_port_task_have_outstanding_io_tasks()) {
		clear_sys_scheduling();
		/*
		 * Got to check that we still got I/O tasks; otherwise
		 * we have to continue checking for I/O...
		 */
		if (!prepare_for_sys_schedule(esdp)) {
		    spincount *= ERTS_SCHED_TSE_SLEEP_SPINCOUNT_FACT;
		    goto tse_wait;
		}
	    }
#endif
	}

	erts_smp_runq_lock(rq);

#ifdef ERTS_SMP
	/*
	 * If we got new I/O tasks we aren't allowed to
	 * sleep in erl_sys_schedule().
	 */
	if (erts_port_task_have_outstanding_io_tasks()) {
	    clear_sys_scheduling();

	    /*
	     * Got to check that we still got I/O tasks; otherwise
	     * we have to wait in erl_sys_schedule() after all...
	     */
	    if (!prepare_for_sys_schedule(esdp)) {
		/*
		 * Not allowed to wait in erl_sys_schedule;
		 * do tse wait instead...
		 */
		sched_change_waiting_sys_to_waiting(esdp->no, rq);
		erts_smp_runq_unlock(rq);
		spincount = 0;
		goto tse_wait;
	    }
	}
#endif
	if (aux_work) {
	    erts_smp_runq_unlock(rq);
	    goto sys_poll_aux_work;
	}
#ifdef ERTS_SMP
	flgs = sched_set_sleeptype(ssi, ERTS_SSI_FLG_POLL_SLEEPING);
	if (!(flgs & ERTS_SSI_FLG_SLEEPING)) {
	    if (!(flgs & ERTS_SSI_FLG_WAITING)) {
		ASSERT(!(flgs & ERTS_SSI_FLG_SLEEPING));
		goto sys_locked_woken;
	    }
	    erts_smp_runq_unlock(rq);
	    flgs = sched_prep_cont_spin_wait(ssi);
	    if (!(flgs & ERTS_SSI_FLG_WAITING)) {
		ASSERT(!(flgs & ERTS_SSI_FLG_SLEEPING));
		goto sys_woken;
	    }
	    ASSERT(!erts_port_task_have_outstanding_io_tasks());
	    goto sys_poll_aux_work;
	}

	ASSERT(flgs & ERTS_SSI_FLG_POLL_SLEEPING);
	ASSERT(flgs & ERTS_SSI_FLG_WAITING);
#endif

	erts_smp_runq_unlock(rq);

	if (working)
	    sched_wall_time_change(esdp, working = 0);

#ifdef ERTS_SMP
	if (thr_prgr_active)
	    erts_thr_progress_active(esdp, thr_prgr_active = 0);
#endif

	ASSERT(!erts_port_task_have_outstanding_io_tasks());

	erl_sys_schedule(0);

	dt = erts_do_time_read_and_reset();
	if (dt) erts_bump_timer(dt);

#ifndef ERTS_SMP
	if (rq->len == 0 && !rq->misc.start)
	    goto sys_aux_work;
    sys_woken:
#else
	flgs = sched_prep_cont_spin_wait(ssi);
	if (flgs & ERTS_SSI_FLG_WAITING)
	    goto sys_aux_work;

    sys_woken:
	if (!thr_prgr_active)
	    erts_thr_progress_active(esdp, thr_prgr_active = 1);
	erts_smp_runq_lock(rq);
    sys_locked_woken:
	if (!thr_prgr_active) {
	    erts_smp_runq_unlock(rq);
	    erts_thr_progress_active(esdp, thr_prgr_active = 1);
	    erts_smp_runq_lock(rq);
	}
	clear_sys_scheduling();
	if (flgs & ~ERTS_SSI_FLG_SUSPENDED)
	    erts_smp_atomic32_read_band_nob(&ssi->flags, ERTS_SSI_FLG_SUSPENDED);
#endif
	if (!working)
	    sched_wall_time_change(esdp, working = 1);
	sched_active_sys(esdp->no, rq);
    }

    ERTS_SMP_LC_ASSERT(erts_smp_lc_runq_is_locked(rq));
}

#ifdef ERTS_SMP

static ERTS_INLINE erts_aint32_t
ssi_flags_set_wake(ErtsSchedulerSleepInfo *ssi)
{
    /* reset all flags but suspended */
    erts_aint32_t oflgs;
    erts_aint32_t nflgs = 0;
    erts_aint32_t xflgs = ERTS_SSI_FLG_SLEEPING|ERTS_SSI_FLG_WAITING;
    while (1) {
	oflgs = erts_smp_atomic32_cmpxchg_relb(&ssi->flags, nflgs, xflgs);
	if (oflgs == xflgs)
	    return oflgs;
	nflgs = oflgs & ERTS_SSI_FLG_SUSPENDED;
	xflgs = oflgs;
    }
}

static void
wake_scheduler(ErtsRunQueue *rq)
{
    ErtsSchedulerSleepInfo *ssi;
    erts_aint32_t flgs;

    /*
     * The unlocked run queue is not strictly necessary
     * from a thread safety or deadlock prevention
     * perspective. It will, however, cost us performance
     * if it is locked during wakup of another scheduler,
     * so all code *should* handle this without having
     * the lock on the run queue.
     */
    ERTS_SMP_LC_ASSERT(!erts_smp_lc_runq_is_locked(rq));

    ssi = rq->scheduler->ssi;

    flgs = ssi_flags_set_wake(ssi);
    erts_sched_finish_poke(ssi, flgs);
}

#ifdef ERTS_DIRTY_SCHEDULERS
static void
wake_dirty_schedulers(ErtsRunQueue *rq, int one)
{
    ErtsSchedulerSleepInfo *ssi;
    ErtsSchedulerSleepList *sl;

    ASSERT(ERTS_RUNQ_IX_IS_DIRTY(rq->ix));

    sl = &rq->sleepers;
    erts_smp_spin_lock(&sl->lock);
    ssi = sl->list;
    if (!ssi) {
	erts_smp_spin_unlock(&sl->lock);
	if (one)
	    wake_scheduler(rq);
    } else if (one) {
	erts_aint32_t flgs;
	if (ssi->prev)
	    ssi->prev->next = ssi->next;
	else {
	    ASSERT(sl->list == ssi);
	    sl->list = ssi->next;
	}
	if (ssi->next)
	    ssi->next->prev = ssi->prev;

	erts_smp_spin_unlock(&sl->lock);

	ERTS_THR_MEMORY_BARRIER;
	flgs = ssi_flags_set_wake(ssi);
	erts_sched_finish_poke(ssi, flgs);
    } else {
	sl->list = NULL;
	erts_smp_spin_unlock(&sl->lock);

	ERTS_THR_MEMORY_BARRIER;
	do {
	    ErtsSchedulerSleepInfo *wake_ssi = ssi;
	    ssi = ssi->next;
	    erts_sched_finish_poke(wake_ssi, ssi_flags_set_wake(wake_ssi));
	} while (ssi);
    }
}
#endif

#define ERTS_NO_USED_RUNQS_SHIFT 16
#define ERTS_NO_RUNQS_MASK 0xffff

#if ERTS_MAX_NO_OF_SCHEDULERS > ERTS_NO_RUNQS_MASK
#  error "Too large amount of schedulers allowed"
#endif

static ERTS_INLINE void
init_no_runqs(int active, int used)
{
    erts_aint32_t no_runqs = (erts_aint32_t) (active & ERTS_NO_RUNQS_MASK);
    no_runqs |= (erts_aint32_t) ((used & ERTS_NO_RUNQS_MASK) << ERTS_NO_USED_RUNQS_SHIFT);
    erts_smp_atomic32_init_nob(&balance_info.no_runqs, no_runqs);
}

static ERTS_INLINE void
get_no_runqs(int *active, int *used)
{
    erts_aint32_t no_runqs = erts_smp_atomic32_read_nob(&balance_info.no_runqs);
    if (active)
	*active = (int) (no_runqs & ERTS_NO_RUNQS_MASK);
    if (used)
	*used = (int) ((no_runqs >> ERTS_NO_USED_RUNQS_SHIFT) & ERTS_NO_RUNQS_MASK);
}

static ERTS_INLINE void
set_no_used_runqs(int used)
{
    erts_aint32_t exp = erts_smp_atomic32_read_nob(&balance_info.no_runqs);
    while (1) {
	erts_aint32_t act, new;
	new = (used & ERTS_NO_RUNQS_MASK) << ERTS_NO_USED_RUNQS_SHIFT;
	new |= exp & ERTS_NO_RUNQS_MASK;
	act = erts_smp_atomic32_cmpxchg_nob(&balance_info.no_runqs, new, exp);
	if (act == exp)
	    break;
	exp = act;
    }
}

static ERTS_INLINE void
set_no_active_runqs(int active)
{
    erts_aint32_t exp = erts_smp_atomic32_read_nob(&balance_info.no_runqs);
    while (1) {
	erts_aint32_t act, new;
	if ((exp & ERTS_NO_RUNQS_MASK) == active)
	    break;
	new = exp & (ERTS_NO_RUNQS_MASK << ERTS_NO_USED_RUNQS_SHIFT);
	new |= active & ERTS_NO_RUNQS_MASK;
	act = erts_smp_atomic32_cmpxchg_nob(&balance_info.no_runqs, new, exp);
	if (act == exp)
	    break;
	exp = act;
    }
}

static ERTS_INLINE int
try_inc_no_active_runqs(int active)
{
    erts_aint32_t exp = erts_smp_atomic32_read_nob(&balance_info.no_runqs);
    if (((exp >> ERTS_NO_USED_RUNQS_SHIFT) & ERTS_NO_RUNQS_MASK) < active)
	return 0;
    if ((exp & ERTS_NO_RUNQS_MASK) + 1 == active) {
	erts_aint32_t new, act;
	new = exp & (ERTS_NO_RUNQS_MASK << ERTS_NO_USED_RUNQS_SHIFT);
	new |= active & ERTS_NO_RUNQS_MASK;
	act = erts_smp_atomic32_cmpxchg_nob(&balance_info.no_runqs, new, exp);
	if (act == exp)
	    return 1;
    }
    return 0;
}

static ERTS_INLINE int
chk_wake_sched(ErtsRunQueue *crq, int ix, int activate)
{
    Uint32 flags;
    ErtsRunQueue *wrq;
    if (crq->ix == ix)
	return 0;
    wrq = ERTS_RUNQ_IX(ix);
    flags = ERTS_RUNQ_FLGS_GET(wrq);
    if (!(flags & (ERTS_RUNQ_FLG_SUSPENDED|ERTS_RUNQ_FLG_NONEMPTY))) {
	if (activate) {
	    if (try_inc_no_active_runqs(ix+1))
		(void) ERTS_RUNQ_FLGS_UNSET(wrq, ERTS_RUNQ_FLG_INACTIVE);
	}
	wake_scheduler(wrq);
	return 1;
    }
    return 0;
}

static void
wake_scheduler_on_empty_runq(ErtsRunQueue *crq)
{
    int ix = crq->ix;
    int stop_ix = ix;
    int active_ix, balance_ix;

    get_no_runqs(&active_ix, &balance_ix);

    if (active_ix > balance_ix)
	active_ix = balance_ix;

    if (ix >= active_ix)
	stop_ix = ix = active_ix;

    /* Try to wake a scheduler on an active run queue */
    while (1) {
	ix--;
	if (ix < 0) {
	    if (active_ix == stop_ix)
		break;
	    ix = active_ix - 1;
	}
	if (ix == stop_ix)
	    break;
	if (chk_wake_sched(crq, ix, 0))
	    return;
    }

    if (active_ix < balance_ix) {
	/* Try to activate a new run queue and wake its scheduler */
	(void) chk_wake_sched(crq, active_ix, 1);
    }
}

#endif /* ERTS_SMP */

static ERTS_INLINE void
smp_notify_inc_runq(ErtsRunQueue *runq)
{
#ifdef ERTS_SMP
    if (runq) {
#ifdef ERTS_DIRTY_SCHEDULERS
	if (ERTS_RUNQ_IX_IS_DIRTY(runq->ix))
	    wake_dirty_schedulers(runq, 1);
	else
#endif
	    wake_scheduler(runq);
    }
#endif
}

void
erts_smp_notify_inc_runq(ErtsRunQueue *runq)
{
    smp_notify_inc_runq(runq);
}

void
erts_sched_notify_check_cpu_bind(void)
{
#ifdef ERTS_SMP
    int ix;
    for (ix = 0; ix < erts_no_run_queues; ix++) {
	ErtsRunQueue *rq = ERTS_RUNQ_IX(ix);
	(void) ERTS_RUNQ_FLGS_SET(rq, ERTS_RUNQ_FLG_CHK_CPU_BIND);
	wake_scheduler(rq);
    }
#else
    erts_sched_check_cpu_bind(erts_get_scheduler_data());
#endif
}


static ERTS_INLINE void
enqueue_process(ErtsRunQueue *runq, int prio, Process *p)
{
    ErtsRunPrioQueue *rpq;

    ERTS_SMP_LC_ASSERT(erts_smp_lc_runq_is_locked(runq));

    erts_smp_inc_runq_len(runq, &runq->procs.prio_info[prio], prio);

    if (prio == PRIORITY_LOW) {
	p->schedule_count = RESCHEDULE_LOW;
	rpq = &runq->procs.prio[PRIORITY_NORMAL];
    }
    else {
	p->schedule_count = 1;
	rpq = &runq->procs.prio[prio];
    }

    p->next = NULL;
    if (rpq->last)
	rpq->last->next = p;
    else
	rpq->first = p;
    rpq->last = p;
}


static ERTS_INLINE void
unqueue_process(ErtsRunQueue *runq,
		ErtsRunPrioQueue *rpq,
		ErtsRunQueueInfo *rqi,
		int prio,
		Process *prev_proc,
		Process *proc)
{
    ERTS_SMP_LC_ASSERT(erts_smp_lc_runq_is_locked(runq));

    if (prev_proc)
	prev_proc->next = proc->next;
    else
	rpq->first = proc->next;
    if (!proc->next)
	rpq->last = prev_proc;

    if (!rpq->first)
	rpq->last = NULL;

    erts_smp_dec_runq_len(runq, rqi, prio);
}


static ERTS_INLINE Process *
dequeue_process(ErtsRunQueue *runq, int prio_q, erts_aint32_t *statep)
{
    erts_aint32_t state;
    int prio;
    ErtsRunPrioQueue *rpq;
    ErtsRunQueueInfo *rqi;
    Process *p;

    ERTS_SMP_LC_ASSERT(erts_smp_lc_runq_is_locked(runq));

    ASSERT(PRIORITY_NORMAL == prio_q
	   || PRIORITY_HIGH == prio_q
	   || PRIORITY_MAX == prio_q);

    rpq = &runq->procs.prio[prio_q];
    p = rpq->first;
    if (!p)
	return NULL;

    ERTS_SMP_DATA_DEPENDENCY_READ_MEMORY_BARRIER;

    state = erts_smp_atomic32_read_nob(&p->state);
    if (statep)
	*statep = state;

    prio = (int) ERTS_PSFLGS_GET_PRQ_PRIO(state);

    rqi = &runq->procs.prio_info[prio];

    if (p)
	unqueue_process(runq, rpq, rqi, prio, NULL, p);

    return p;
}

static ERTS_INLINE int
check_requeue_process(ErtsRunQueue *rq, int prio_q)
{
    ErtsRunPrioQueue *rpq = &rq->procs.prio[prio_q];
    Process *p = rpq->first;
    if (--p->schedule_count > 0 && p != rpq->last) {
	/* reschedule */
	rpq->first = p->next;
	rpq->last->next = p;
	rpq->last = p;
	p->next = NULL;
	return 1;
    }
    return 0;
}

#ifdef ERTS_SMP

static ErtsRunQueue *
check_immigration_need(ErtsRunQueue *c_rq, ErtsMigrationPath *mp, int prio)
{
    int len;
    Uint32 f_flags, f_rq_flags;
    ErtsRunQueue *f_rq;

    f_flags = mp->prio[prio].flags;

    ASSERT(ERTS_CHK_RUNQ_FLG_IMMIGRATE(mp->flags, prio));

    f_rq = mp->prio[prio].runq;
    if (!f_rq)
	return NULL;

#if ERTS_HAVE_SCHED_UTIL_BALANCING_SUPPORT
    if (mp->sched_util)
	return NULL;
#endif

    f_rq_flags = ERTS_RUNQ_FLGS_GET(f_rq);
    if (f_rq_flags & ERTS_RUNQ_FLG_PROTECTED)
	return NULL;

    if (ERTS_CHK_RUNQ_FLG_EVACUATE(f_flags, prio))
	return f_rq;

    if (f_rq_flags & ERTS_RUNQ_FLG_INACTIVE)
	return f_rq;

    if (prio == ERTS_PORT_PRIO_LEVEL)
	len = RUNQ_READ_LEN(&c_rq->ports.info.len);
    else
	len = RUNQ_READ_LEN(&c_rq->procs.prio_info[prio].len);

    if (len < mp->prio[prio].limit.this) {
	if (prio == ERTS_PORT_PRIO_LEVEL)
	    len = RUNQ_READ_LEN(&f_rq->ports.info.len);
	else
	    len = RUNQ_READ_LEN(&f_rq->procs.prio_info[prio].len);

	if (len > mp->prio[prio].limit.other)
	    return f_rq;
    }
    return NULL;
}

static void
immigrate(ErtsRunQueue *c_rq, ErtsMigrationPath *mp)
{
    Uint32 iflags, iflag;
    erts_smp_runq_unlock(c_rq);

    ASSERT(erts_thr_progress_is_managed_thread());

    iflags = mp->flags & ERTS_RUNQ_FLGS_IMMIGRATE_QMASK;

    iflag = iflags & -iflags;

    while (iflag) {
	ErtsRunQueue *rq;
	int prio;

	switch (iflag) {
	case (MAX_BIT << ERTS_RUNQ_FLGS_IMMIGRATE_SHFT):
	    prio = PRIORITY_MAX;
	    break;
	case (HIGH_BIT << ERTS_RUNQ_FLGS_IMMIGRATE_SHFT):
	    prio = PRIORITY_HIGH;
	    break;
	case (NORMAL_BIT << ERTS_RUNQ_FLGS_IMMIGRATE_SHFT):
	    prio = PRIORITY_NORMAL;
	    break;
	case (LOW_BIT << ERTS_RUNQ_FLGS_IMMIGRATE_SHFT):
	    prio = PRIORITY_LOW;
	    break;
	case (PORT_BIT << ERTS_RUNQ_FLGS_IMMIGRATE_SHFT):
	    prio = ERTS_PORT_PRIO_LEVEL;
	    break;
	default:
	    erl_exit(ERTS_ABORT_EXIT,
		     "%s:%d:%s(): Invalid immigrate queue mask",
		     __FILE__, __LINE__, __func__);
	    prio = 0;
	    break;
	}

	iflags &= ~iflag;
	iflag = iflags & -iflags;

	rq = check_immigration_need(c_rq, mp, prio);
	if (rq) {
	    erts_smp_runq_lock(rq);
	    if (prio == ERTS_PORT_PRIO_LEVEL) {
		Port *prt;
		prt = erts_dequeue_port(rq);
		if (prt)
		    RUNQ_SET_RQ(&prt->run_queue, c_rq);
		erts_smp_runq_unlock(rq);
		if (prt) {
		    /* port might terminate while we have no lock... */
		    rq = erts_port_runq(prt);
		    if (rq) {
			if (rq != c_rq)
			    erl_exit(ERTS_ABORT_EXIT,
				     "%s:%d:%s(): Internal error",
				     __FILE__, __LINE__, __func__);
			erts_enqueue_port(c_rq, prt);
			if (!iflag)
			    return; /* done */
			erts_smp_runq_unlock(c_rq);
		    }
		}
	    }
	    else {
		ErtsRunPrioQueue *rpq = &rq->procs.prio[prio == PRIORITY_LOW
							? PRIORITY_NORMAL
							: prio];
		Process *prev_proc = NULL;
		Process *proc = rpq->first;
		int rq_locked = 1;

		while (proc) {
		    erts_aint32_t state;
		    state = erts_smp_atomic32_read_acqb(&proc->state);
		    if (!(ERTS_PSFLG_BOUND & state)
			&& (prio == (int) ERTS_PSFLGS_GET_PRQ_PRIO(state))) {
			ErtsRunQueueInfo *rqi = &rq->procs.prio_info[prio];
			unqueue_process(rq, rpq, rqi, prio, prev_proc, proc);
			erts_smp_runq_unlock(rq);
			RUNQ_SET_RQ(&proc->run_queue, c_rq);
			rq_locked = 0;

			erts_smp_runq_lock(c_rq);
			enqueue_process(c_rq, prio, proc);
			if (!iflag)
			    return; /* done */
			erts_smp_runq_unlock(c_rq);
			break;
		    }
		    prev_proc = proc;
		    proc = proc->next;
		}
		if (rq_locked)
		    erts_smp_runq_unlock(rq);
	    }
	}
    }

    erts_smp_runq_lock(c_rq);
}

static ERTS_INLINE void
suspend_run_queue(ErtsRunQueue *rq)
{
    erts_smp_atomic32_read_bor_nob(&rq->scheduler->ssi->flags,
				   ERTS_SSI_FLG_SUSPENDED);
    (void) ERTS_RUNQ_FLGS_SET(rq, ERTS_RUNQ_FLG_SUSPENDED);

    wake_scheduler(rq);
}

static void scheduler_ix_resume_wake(Uint ix);
static void scheduler_ssi_resume_wake(ErtsSchedulerSleepInfo *ssi);

static ERTS_INLINE void
resume_run_queue(ErtsRunQueue *rq)
{
    int pix;

    erts_smp_runq_lock(rq);

    (void) ERTS_RUNQ_FLGS_READ_BSET(rq,
				    (ERTS_RUNQ_FLG_OUT_OF_WORK
				     | ERTS_RUNQ_FLG_HALFTIME_OUT_OF_WORK
				     | ERTS_RUNQ_FLG_SUSPENDED),
				    (ERTS_RUNQ_FLG_OUT_OF_WORK
				     | ERTS_RUNQ_FLG_HALFTIME_OUT_OF_WORK));

    rq->check_balance_reds = ERTS_RUNQ_CALL_CHECK_BALANCE_REDS;
    for (pix = 0; pix < ERTS_NO_PROC_PRIO_LEVELS; pix++) {
	rq->procs.prio_info[pix].max_len = 0;
	rq->procs.prio_info[pix].reds = 0;
    }
    rq->ports.info.max_len = 0;
    rq->ports.info.reds = 0;
    rq->max_len = 0;

    erts_smp_runq_unlock(rq);

#ifdef ERTS_DIRTY_SCHEDULERS
    if (!ERTS_RUNQ_IX_IS_DIRTY(rq->ix))
#endif
	scheduler_ix_resume_wake(rq->ix);
}

typedef struct {
    Process *first;
    Process *last;
} ErtsStuckBoundProcesses;

static void
schedule_bound_processes(ErtsRunQueue *rq,
			 ErtsStuckBoundProcesses *sbpp)
{
    Process *proc, *next;
    ERTS_SMP_LC_ASSERT(erts_smp_lc_runq_is_locked(rq));

    proc = sbpp->first;
    while (proc) {
	erts_aint32_t state = erts_smp_atomic32_read_acqb(&proc->state);
	next = proc->next;
	enqueue_process(rq, (int) ERTS_PSFLGS_GET_PRQ_PRIO(state), proc);
	proc = next;
    }
}

static void
evacuate_run_queue(ErtsRunQueue *rq,
		   ErtsStuckBoundProcesses *sbpp)
{
    int prio_q;
    ErtsRunQueue *to_rq;
    ErtsMigrationPaths *mps;
    ErtsMigrationPath *mp = NULL;

    ERTS_SMP_LC_ASSERT(erts_smp_lc_runq_is_locked(rq));

    (void) ERTS_RUNQ_FLGS_UNSET(rq, ERTS_RUNQ_FLG_PROTECTED);

#ifdef ERTS_DIRTY_SCHEDULERS
    if (!ERTS_RUNQ_IX_IS_DIRTY(rq->ix))
#endif
    {
	mps = erts_get_migration_paths_managed();
	mp = &mps->mpath[rq->ix];
    }

    /* Evacuate scheduled misc ops */

    if (rq->misc.start) {
	ErtsMiscOpList *start, *end;

#ifdef ERTS_DIRTY_SCHEDULERS
	ASSERT(!ERTS_RUNQ_IX_IS_DIRTY(rq->ix));
#endif
	to_rq = mp->misc_evac_runq;
	if (!to_rq)
	    return;

	start = rq->misc.start;
	end = rq->misc.end;
	rq->misc.start = NULL;
	rq->misc.end = NULL;
	erts_smp_runq_unlock(rq);

	erts_smp_runq_lock(to_rq);
	if (to_rq->misc.end)
	    to_rq->misc.end->next = start;
	else
	    to_rq->misc.start = start;

	to_rq->misc.end = end;

	non_empty_runq(to_rq);

	erts_smp_runq_unlock(to_rq);
	smp_notify_inc_runq(to_rq);
	erts_smp_runq_lock(to_rq);
    }

    if (rq->ports.start) {
	Port *prt;

#ifdef ERTS_DIRTY_SCHEDULERS
	ASSERT(!ERTS_RUNQ_IX_IS_DIRTY(rq->ix));
#endif
	to_rq = mp->prio[ERTS_PORT_PRIO_LEVEL].runq;
	if (!to_rq)
	    return;

	/* Evacuate scheduled ports */
	prt = rq->ports.start;
	while (prt) {
	    ErtsRunQueue *prt_rq;
	    prt = erts_dequeue_port(rq);
	    RUNQ_SET_RQ(&prt->run_queue, to_rq);
	    erts_smp_runq_unlock(rq);
	    /*
	     * The port might terminate while
	     * we have no lock on it...
	     */
	    prt_rq = erts_port_runq(prt);
	    if (prt_rq) {
		if (prt_rq != to_rq)
		    erl_exit(ERTS_ABORT_EXIT,
			     "%s:%d:%s() internal error\n",
			     __FILE__, __LINE__, __func__);
		erts_enqueue_port(to_rq, prt);
		erts_smp_runq_unlock(to_rq);
	    }
	    erts_smp_runq_lock(rq);
	    prt = rq->ports.start;
	}
	smp_notify_inc_runq(to_rq);
    }

    /* Evacuate scheduled processes */
    for (prio_q = 0; prio_q < ERTS_NO_PROC_PRIO_QUEUES; prio_q++) {
	erts_aint32_t state;
	Process *proc;
	int notify = 0;
#ifdef ERTS_DIRTY_SCHEDULERS
	int requeue;
#endif
	to_rq = NULL;

#ifdef ERTS_DIRTY_SCHEDULERS
	if (!ERTS_RUNQ_IX_IS_DIRTY(rq->ix))
#endif
	{
	    if (!mp->prio[prio_q].runq)
		return;
	    if (prio_q == PRIORITY_NORMAL && !mp->prio[PRIORITY_LOW].runq)
		return;
	}

	proc = dequeue_process(rq, prio_q, &state);
	while (proc) {
#ifdef ERTS_DIRTY_SCHEDULERS
	    requeue = 1;
#endif
	    if (ERTS_PSFLG_BOUND & state) {
		/* Bound processes get stuck here... */
		proc->next = NULL;
		if (sbpp->last)
		    sbpp->last->next = proc;
		else
		    sbpp->first = proc;
		sbpp->last = proc;
#ifdef ERTS_DIRTY_SCHEDULERS
		requeue = 0;
#endif
	    }
#ifdef ERTS_DIRTY_SCHEDULERS
	    else if (state & ERTS_PSFLG_DIRTY_CPU_PROC_IN_Q) {
#ifdef DEBUG
		erts_aint32_t old =
#else
		(void)
#endif
		    erts_smp_atomic32_read_band_nob(&proc->state,
						    ~(ERTS_PSFLG_DIRTY_CPU_PROC
						      | ERTS_PSFLG_DIRTY_CPU_PROC_IN_Q));
		/* assert that no other dirty flags are set */
		ASSERT(!(old & (ERTS_PSFLG_DIRTY_IO_PROC|ERTS_PSFLG_DIRTY_IO_PROC_IN_Q)));
	    } else if (state & ERTS_PSFLG_DIRTY_IO_PROC_IN_Q) {
#ifdef DEBUG
		erts_aint32_t old =
#else
		(void)
#endif
		    erts_smp_atomic32_read_band_nob(&proc->state,
						    ~(ERTS_PSFLG_DIRTY_IO_PROC
						      | ERTS_PSFLG_DIRTY_IO_PROC_IN_Q));
		/* assert that no other dirty flags are set */
		ASSERT(!(old & (ERTS_PSFLG_DIRTY_CPU_PROC|ERTS_PSFLG_DIRTY_CPU_PROC_IN_Q)));
	    }
	    if (requeue) {
#else
	    else {
#endif
		int prio = (int) ERTS_PSFLGS_GET_PRQ_PRIO(state);
		erts_smp_runq_unlock(rq);

#ifdef ERTS_DIRTY_SCHEDULERS
		if (ERTS_RUNQ_IX_IS_DIRTY(rq->ix))
		    /*
		     * dirty run queues evacuate only to run
		     * queue 0 during multi-scheduling blocking
		     */
		    to_rq = ERTS_RUNQ_IX(0);
		else
#endif
		    to_rq = mp->prio[prio].runq;
		RUNQ_SET_RQ(&proc->run_queue, to_rq);

		erts_smp_runq_lock(to_rq);
		enqueue_process(to_rq, prio, proc);
		erts_smp_runq_unlock(to_rq);
		notify = 1;

		erts_smp_runq_lock(rq);
	    }
	    proc = dequeue_process(rq, prio_q, &state);
	}
	if (notify)
	    smp_notify_inc_runq(to_rq);
    }
}

static int
try_steal_task_from_victim(ErtsRunQueue *rq, int *rq_lockedp, ErtsRunQueue *vrq, Uint32 flags)
{
    Uint32 procs_qmask = flags & ERTS_RUNQ_FLGS_PROCS_QMASK;
    int max_prio_bit;
    ErtsRunPrioQueue *rpq;

    if (*rq_lockedp) {
	erts_smp_runq_unlock(rq);
	*rq_lockedp = 0;
    }

    ERTS_SMP_LC_ASSERT(!erts_smp_lc_runq_is_locked(rq));

    erts_smp_runq_lock(vrq);

    if (rq->halt_in_progress)
	goto no_procs;

    /*
     * Check for a runnable process to steal...
     */

    while (procs_qmask) {
	Process *prev_proc;
	Process *proc;

	max_prio_bit = procs_qmask & -procs_qmask;
	switch (max_prio_bit) {
	case MAX_BIT:
	    rpq = &vrq->procs.prio[PRIORITY_MAX];
	    break;
	case HIGH_BIT:
	    rpq = &vrq->procs.prio[PRIORITY_HIGH];
	    break;
	case NORMAL_BIT:
	case LOW_BIT:
	    rpq = &vrq->procs.prio[PRIORITY_NORMAL];
	    break;
	case 0:
	    goto no_procs;
	default:
	    ASSERT(!"Invalid queue mask");
	    goto no_procs;
	}

	prev_proc = NULL;
	proc = rpq->first;

	while (proc) {
	    erts_aint32_t state = erts_smp_atomic32_read_acqb(&proc->state);
	    if (!(ERTS_PSFLG_BOUND & state)) {
		/* Steal process */
		int prio = (int) ERTS_PSFLGS_GET_PRQ_PRIO(state);
		ErtsRunQueueInfo *rqi = &vrq->procs.prio_info[prio];
		unqueue_process(vrq, rpq, rqi, prio, prev_proc, proc);
		erts_smp_runq_unlock(vrq);
		RUNQ_SET_RQ(&proc->run_queue, rq);

		erts_smp_runq_lock(rq);
		*rq_lockedp = 1;
		enqueue_process(rq, prio, proc);
		return !0;
	    }
	    prev_proc = proc;
	    proc = proc->next;
	}

	procs_qmask &= ~max_prio_bit;
    }

no_procs:

    ERTS_SMP_LC_ASSERT(erts_smp_lc_runq_is_locked(vrq));

    /*
     * Check for a runnable port to steal...
     */

    if (vrq->ports.start) {
	ErtsRunQueue *prt_rq;
	Port *prt = erts_dequeue_port(vrq);
	RUNQ_SET_RQ(&prt->run_queue, rq);
	erts_smp_runq_unlock(vrq);

	/*
	 * The port might terminate while
	 * we have no lock on it...
	 */

	prt_rq = erts_port_runq(prt);
	if (!prt_rq)
	    return 0;
	else {
	    if (prt_rq != rq)
		erl_exit(ERTS_ABORT_EXIT,
			 "%s:%d:%s() internal error\n",
			 __FILE__, __LINE__, __func__);
	    *rq_lockedp = 1;
	    erts_enqueue_port(rq, prt);
	    return !0;
	}
    }

    erts_smp_runq_unlock(vrq);

    return 0;
}


static ERTS_INLINE int
check_possible_steal_victim(ErtsRunQueue *rq, int *rq_lockedp, int vix)
{
    ErtsRunQueue *vrq = ERTS_RUNQ_IX(vix);
    Uint32 flags = ERTS_RUNQ_FLGS_GET(vrq);
    if ((flags & (ERTS_RUNQ_FLG_NONEMPTY
		  | ERTS_RUNQ_FLG_PROTECTED)) == ERTS_RUNQ_FLG_NONEMPTY)
	return try_steal_task_from_victim(rq, rq_lockedp, vrq, flags);
    else
	return 0;
}


static int
try_steal_task(ErtsRunQueue *rq)
{
    int res, rq_locked, vix, active_rqs, blnc_rqs;
    Uint32 flags;

    /* Protect jobs we steal from getting stolen from us... */
    flags = empty_protected_runq(rq);
    if (flags & ERTS_RUNQ_FLG_SUSPENDED)
	return 0; /* go suspend instead... */

    res = 0;
    rq_locked = 1;

    ERTS_SMP_LC_CHK_RUNQ_LOCK(rq, rq_locked);

    get_no_runqs(&active_rqs, &blnc_rqs);

    if (active_rqs > blnc_rqs)
	active_rqs = blnc_rqs;

    if (rq->ix < active_rqs) {

	/* First try to steal from an inactive run queue... */
	if (active_rqs < blnc_rqs) {
	    int no = blnc_rqs - active_rqs;
	    int stop_ix = vix = active_rqs + rq->ix % no;
	    while (erts_smp_atomic32_read_acqb(&no_empty_run_queues) < blnc_rqs) {
		res = check_possible_steal_victim(rq, &rq_locked, vix);
		if (res)
		    goto done;
		vix++;
		if (vix >= blnc_rqs)
		    vix = active_rqs;
		if (vix == stop_ix)
		    break;
	    }
	}

	vix = rq->ix;

	/* ... then try to steal a job from another active queue... */
	while (erts_smp_atomic32_read_acqb(&no_empty_run_queues) < blnc_rqs) {
	    vix++;
	    if (vix >= active_rqs)
		vix = 0;
	    if (vix == rq->ix)
		break;

	    res = check_possible_steal_victim(rq, &rq_locked, vix);
	    if (res)
		goto done;
	}

    }

 done:

    if (!rq_locked)
	erts_smp_runq_lock(rq);

    if (!res)
	res = rq->halt_in_progress ?
	    !ERTS_EMPTY_RUNQ_PORTS(rq) : !ERTS_EMPTY_RUNQ(rq);

    return res;
}

/* Run queue balancing */

typedef struct {
    Uint32 flags;
    struct {
	int max_len;
	int avail;
	int reds;
	int migration_limit;
	int emigrate_to;
	int immigrate_from;
    } prio[ERTS_NO_PRIO_LEVELS];
    int reds;
    int full_reds;
    int full_reds_history_sum;
    int full_reds_history_change;
    int oowc;
    int max_len;
#if ERTS_HAVE_SCHED_UTIL_BALANCING_SUPPORT
    int sched_util;
#endif
} ErtsRunQueueBalance;
static ErtsRunQueueBalance *run_queue_info;

typedef struct {
    int qix;
    int len;
} ErtsRunQueueCompare;
static ErtsRunQueueCompare *run_queue_compare;

static int
rqc_len_cmp(const void *x, const void *y)
{
    return ((ErtsRunQueueCompare *) x)->len - ((ErtsRunQueueCompare *) y)->len;
}

#define ERTS_PERCENT(X, Y) \
  ((Y) == 0 \
   ? ((X) == 0 ? 100 : INT_MAX) \
   : ((100*(X))/(Y)))

#define ERTS_UPDATE_FULL_REDS(QIX, LAST_REDS)				\
do {									\
    run_queue_info[(QIX)].full_reds					\
	= run_queue_info[(QIX)].full_reds_history_sum;			\
    run_queue_info[(QIX)].full_reds += (LAST_REDS);			\
    run_queue_info[(QIX)].full_reds					\
	>>= ERTS_FULL_REDS_HISTORY_AVG_SHFT;				\
    run_queue_info[(QIX)].full_reds_history_sum				\
	-= run_queue_info[(QIX)].full_reds_history_change;		\
    run_queue_info[(QIX)].full_reds_history_sum += (LAST_REDS);		\
    run_queue_info[(QIX)].full_reds_history_change = (LAST_REDS);	\
} while (0)

#define ERTS_DBG_CHK_FULL_REDS_HISTORY(RQ)				\
do {									\
    int sum__ = 0;							\
    int rix__;								\
    for (rix__ = 0; rix__ < ERTS_FULL_REDS_HISTORY_SIZE; rix__++)	\
	sum__ += (RQ)->full_reds_history[rix__];			\
    ASSERT(sum__ == (RQ)->full_reds_history_sum);			\
} while (0);

#define ERTS_PRE_ALLOCED_MPATHS 8

erts_atomic_t erts_migration_paths;

static struct {
    size_t size;
    ErtsMigrationPaths *freelist;
    struct {
	ErtsMigrationPaths *first;
	ErtsMigrationPaths *last;
    } retired;
} mpaths;

static void
init_migration_paths(void)
{
    int qix, i;
    char *p;
    ErtsMigrationPaths *mps;

    mpaths.size = sizeof(ErtsMigrationPaths);
    mpaths.size += sizeof(ErtsMigrationPath)*(erts_no_schedulers-1);
    mpaths.size = ERTS_ALC_CACHE_LINE_ALIGN_SIZE(mpaths.size);

    p = erts_alloc_permanent_cache_aligned(ERTS_ALC_T_LL_MPATHS,
					   (mpaths.size
					    * ERTS_PRE_ALLOCED_MPATHS));
    mpaths.freelist = NULL;
    for (i = 0; i < ERTS_PRE_ALLOCED_MPATHS-1; i++) {
	mps = (ErtsMigrationPaths *) p;
	mps->next = mpaths.freelist;
	mpaths.freelist = mps;
	p += mpaths.size;
    }

    mps = (ErtsMigrationPaths *) p;
    mps->block = NULL;
    for (qix = 0; qix < erts_no_run_queues; qix++) {
	int pix;
	mps->mpath[qix].flags = 0;
	mps->mpath[qix].misc_evac_runq = NULL;
	for (pix = 0; pix < ERTS_NO_PRIO_LEVELS; pix++) {
	    mps->mpath[qix].prio[pix].limit.this = -1;
	    mps->mpath[qix].prio[pix].limit.other = -1;
	    mps->mpath[qix].prio[pix].runq = NULL;
	    mps->mpath[qix].prio[pix].flags = 0;
	}
    }
    erts_atomic_init_wb(&erts_migration_paths, (erts_aint_t) mps);
}

static ERTS_INLINE ErtsMigrationPaths *
alloc_mpaths(void)
{
    void *block;
    ErtsMigrationPaths *res;
    ERTS_SMP_LC_ASSERT(erts_smp_lc_mtx_is_locked(&balance_info.update_mtx));

    res = mpaths.freelist;
    if (res) {
	mpaths.freelist = res->next;
	res->block = NULL;
	return res;
    }
    res = erts_alloc(ERTS_ALC_T_SL_MPATHS,
		     mpaths.size+ERTS_CACHE_LINE_SIZE);
    block = (void *) res;
    if (((UWord) res) & ERTS_CACHE_LINE_MASK)
	res = (ErtsMigrationPaths *) ((((UWord) res) & ~ERTS_CACHE_LINE_MASK)
				      + ERTS_CACHE_LINE_SIZE);
    res->block = block;
    return res;
}

static ERTS_INLINE void
retire_mpaths(ErtsMigrationPaths *mps)
{
    ErtsThrPrgrVal current;

    ERTS_SMP_LC_ASSERT(erts_smp_lc_mtx_is_locked(&balance_info.update_mtx));

    current = erts_thr_progress_current();

    while (mpaths.retired.first) {
	ErtsMigrationPaths *tmp = mpaths.retired.first;
	if (!erts_thr_progress_has_reached_this(current, tmp->thr_prgr))
	    break;
	mpaths.retired.first = tmp->next;
	if (tmp->block) {
	    erts_free(ERTS_ALC_T_SL_MPATHS, tmp->block);
	}
	else {
	    tmp->next = mpaths.freelist;
	    mpaths.freelist = tmp;
	}
    }

    if (!mpaths.retired.first)
	mpaths.retired.last = NULL;

    mps->thr_prgr = erts_thr_progress_later(NULL);
    mps->next = NULL;

    if (mpaths.retired.last)
	mpaths.retired.last->next = mps;
    else
	mpaths.retired.first = mps;
    mpaths.retired.last = mps;
}

static void
check_balance(ErtsRunQueue *c_rq)
{
#if ERTS_MAX_PROCESSES >= (1 << 27)
#  error check_balance() assumes ERTS_MAX_PROCESS < (1 << 27)
#endif
    ErtsMigrationPaths *new_mpaths, *old_mpaths;
    ErtsRunQueueBalance avg = {0};
    Sint64 scheds_reds, full_scheds_reds;
    int forced, active, current_active, oowc, half_full_scheds, full_scheds,
	mmax_len, blnc_no_rqs, qix, pix, freds_hist_ix;
#if ERTS_HAVE_SCHED_UTIL_BALANCING_SUPPORT
    int sched_util_balancing;
#endif

    if (erts_smp_atomic32_xchg_nob(&balance_info.checking_balance, 1)) {
	c_rq->check_balance_reds = INT_MAX;
	return;
    }

    get_no_runqs(NULL, &blnc_no_rqs);
    if (blnc_no_rqs == 1) {
	c_rq->check_balance_reds = INT_MAX;
	erts_smp_atomic32_set_nob(&balance_info.checking_balance, 0);
	return;
    }

    erts_smp_runq_unlock(c_rq);

    if (balance_info.halftime) {	
	balance_info.halftime = 0;
	erts_smp_atomic32_set_nob(&balance_info.checking_balance, 0);
	ERTS_FOREACH_RUNQ(rq,
	{
	    if (rq->waiting)
		(void) ERTS_RUNQ_FLGS_SET(rq, ERTS_RUNQ_FLG_HALFTIME_OUT_OF_WORK);
	    else
		(void) ERTS_RUNQ_FLGS_UNSET(rq, ERTS_RUNQ_FLG_HALFTIME_OUT_OF_WORK);
	    rq->check_balance_reds = ERTS_RUNQ_CALL_CHECK_BALANCE_REDS;
	});

	erts_smp_runq_lock(c_rq);
	return;
    }

    /*
     * check_balance() is never called in more threads
     * than one at a time, i.e., we will normally never
     * get any conflicts on the balance_info.update_mtx.
     * However, when blocking multi scheduling (which performance
     * critical applications do *not* do) migration information
     * is manipulated. Such updates of the migration information
     * might clash with balancing.
     */
    erts_smp_mtx_lock(&balance_info.update_mtx);

    forced = balance_info.forced_check_balance;
    balance_info.forced_check_balance = 0;

    get_no_runqs(&current_active, &blnc_no_rqs);

    if (blnc_no_rqs == 1) {
	erts_smp_mtx_unlock(&balance_info.update_mtx);
	erts_smp_runq_lock(c_rq);
	c_rq->check_balance_reds = INT_MAX;
	erts_smp_atomic32_set_nob(&balance_info.checking_balance, 0);
	return;
    }

#if ERTS_HAVE_SCHED_UTIL_BALANCING_SUPPORT
    sched_util_balancing = 0;
#endif

    freds_hist_ix = balance_info.full_reds_history_index;
    balance_info.full_reds_history_index++;
    if (balance_info.full_reds_history_index >= ERTS_FULL_REDS_HISTORY_SIZE)
	balance_info.full_reds_history_index = 0;

    /* Read balance information for all run queues */
    for (qix = 0; qix < blnc_no_rqs; qix++) {
	ErtsRunQueue *rq = ERTS_RUNQ_IX(qix);
	erts_smp_runq_lock(rq);

	run_queue_info[qix].flags = ERTS_RUNQ_FLGS_GET_NOB(rq);
	for (pix = 0; pix < ERTS_NO_PROC_PRIO_LEVELS; pix++) {
	    run_queue_info[qix].prio[pix].max_len
		= rq->procs.prio_info[pix].max_len;
	    run_queue_info[qix].prio[pix].reds
		= rq->procs.prio_info[pix].reds;
	}
	run_queue_info[qix].prio[ERTS_PORT_PRIO_LEVEL].max_len
	    = rq->ports.info.max_len;
	run_queue_info[qix].prio[ERTS_PORT_PRIO_LEVEL].reds
	    = rq->ports.info.reds;

	run_queue_info[qix].full_reds_history_sum
	    = rq->full_reds_history_sum;
	run_queue_info[qix].full_reds_history_change
	    = rq->full_reds_history[freds_hist_ix];

	run_queue_info[qix].oowc = rq->out_of_work_count;
	run_queue_info[qix].max_len = rq->max_len;
	rq->check_balance_reds = INT_MAX;

#if ERTS_HAVE_SCHED_UTIL_BALANCING_SUPPORT
	if (erts_sched_balance_util)
	    run_queue_info[qix].sched_util = erts_get_sched_util(rq, 1, 0);
#endif

	erts_smp_runq_unlock(rq);
    }

    full_scheds = 0;
    half_full_scheds = 0;
    full_scheds_reds = 0;
    scheds_reds = 0;
    oowc = 0;
    mmax_len = 0;

    /* Calculate availability for each priority in each run queues */
    for (qix = 0; qix < blnc_no_rqs; qix++) {
	int treds = 0;

	if (run_queue_info[qix].flags & ERTS_RUNQ_FLG_OUT_OF_WORK) {
	    for (pix = 0; pix < ERTS_NO_PRIO_LEVELS; pix++) {
		run_queue_info[qix].prio[pix].avail = 100;
		treds += run_queue_info[qix].prio[pix].reds;
	    }
	    if (!(run_queue_info[qix].flags & ERTS_RUNQ_FLG_HALFTIME_OUT_OF_WORK))
		half_full_scheds++;
	    ERTS_UPDATE_FULL_REDS(qix, ERTS_RUNQ_CHECK_BALANCE_REDS_PER_SCHED);
	}
	else {
	    ASSERT(!(run_queue_info[qix].flags & ERTS_RUNQ_FLG_HALFTIME_OUT_OF_WORK));
	    for (pix = 0; pix < ERTS_NO_PRIO_LEVELS; pix++)
		treds += run_queue_info[qix].prio[pix].reds;
	    if (treds == 0) {
		for (pix = 0; pix < ERTS_NO_PRIO_LEVELS; pix++)
		    run_queue_info[qix].prio[pix].avail = 0;
	    }
	    else {
		Sint64 xreds = 0;
		Sint64 procreds = treds;
		procreds -= 
		    ((Sint64)
		     run_queue_info[qix].prio[ERTS_PORT_PRIO_LEVEL].reds);

		for (pix = 0; pix < ERTS_NO_PROC_PRIO_LEVELS; pix++) {
		    Sint64 av;

		    if (xreds == 0)
			av = 100;
		    else if (procreds == xreds)
			av = 0;
		    else {
			av = (100*(procreds - xreds)) / procreds;
			if (av == 0)
			    av = 1;
		    }
		    run_queue_info[qix].prio[pix].avail = (int) av;
		    ASSERT(run_queue_info[qix].prio[pix].avail >= 0);
		    if (pix < PRIORITY_NORMAL) /* ie., max or high */
			xreds += (Sint64) run_queue_info[qix].prio[pix].reds;
		}
		run_queue_info[qix].prio[ERTS_PORT_PRIO_LEVEL].avail = 100;
	    }
	    ERTS_UPDATE_FULL_REDS(qix, treds);
	    full_scheds_reds += run_queue_info[qix].full_reds;
	    full_scheds++;
	    half_full_scheds++;
	}
	run_queue_info[qix].reds = treds;
	scheds_reds += treds;
	oowc += run_queue_info[qix].oowc;
	if (mmax_len < run_queue_info[qix].max_len)
	    mmax_len = run_queue_info[qix].max_len;
    }

    if (!erts_sched_compact_load) {
#if ERTS_HAVE_SCHED_UTIL_BALANCING_SUPPORT
	if (erts_sched_balance_util && full_scheds < blnc_no_rqs) {
	    int avg_util = 0;

	    for (qix = 0; qix < blnc_no_rqs; qix++)
		avg_util += run_queue_info[qix].sched_util;

	    avg_util /= blnc_no_rqs; /* in ppm */

	    sched_util_balancing = 1;
	    /*
	     * In order to avoid renaming a large amount of fields
	     * we write utilization values instead of lenght values
	     * in the 'max_len' and 'migration_limit' fields...
	     */
	    for (qix = 0; qix < blnc_no_rqs; qix++) {
		run_queue_info[qix].flags = 0; /* Reset for later use... */
		for (pix = 0; pix < ERTS_NO_PRIO_LEVELS; pix++) {
		    run_queue_info[qix].prio[pix].emigrate_to = -1;
		    run_queue_info[qix].prio[pix].immigrate_from = -1;
		    run_queue_info[qix].prio[pix].avail = 100;
		    run_queue_info[qix].prio[pix].max_len = run_queue_info[qix].sched_util;
		    run_queue_info[qix].prio[pix].migration_limit = avg_util;
		}
	    }
	    active = blnc_no_rqs;
	    goto setup_migration_paths;
	}
#endif
	goto all_active;
    }

    if (!forced && half_full_scheds != blnc_no_rqs) {
	int min = 1;
	if (min < half_full_scheds)
	    min = half_full_scheds;
	if (full_scheds) {
	    active = (scheds_reds - 1)/ERTS_RUNQ_CHECK_BALANCE_REDS_PER_SCHED+1;
	}
	else {
	    active = balance_info.last_active_runqs - 1;
	}

	if (balance_info.last_active_runqs < current_active) {
	    ERTS_BLNCE_SAVE_RISE(current_active, mmax_len, scheds_reds);
	    active = current_active;
	}
	else if (active < balance_info.prev_rise.active_runqs) {
	    if (ERTS_PERCENT(mmax_len,
			     balance_info.prev_rise.max_len) >= 90
		&& ERTS_PERCENT(scheds_reds,
				balance_info.prev_rise.reds) >= 90) {
		active = balance_info.prev_rise.active_runqs;
	    }
	}

	if (active < min)
	    active = min;
	else if (active > blnc_no_rqs)
	    active = blnc_no_rqs;

	if (active == blnc_no_rqs)
	    goto all_active;

	for (qix = 0; qix < active; qix++) {
	    run_queue_info[qix].flags = 0;
	    for (pix = 0; pix < ERTS_NO_PRIO_LEVELS; pix++) {
		run_queue_info[qix].prio[pix].emigrate_to = -1;
		run_queue_info[qix].prio[pix].immigrate_from = -1;
		run_queue_info[qix].prio[pix].migration_limit = 0;
	    }
	}
	for (qix = active; qix < blnc_no_rqs; qix++) {
	    run_queue_info[qix].flags = ERTS_RUNQ_FLG_INACTIVE;
	    for (pix = 0; pix < ERTS_NO_PRIO_LEVELS; pix++) {
		int tix = qix % active;
		ERTS_SET_RUNQ_FLG_EMIGRATE(run_queue_info[qix].flags, pix);
		run_queue_info[qix].prio[pix].emigrate_to = tix;
		run_queue_info[qix].prio[pix].immigrate_from = -1;
		run_queue_info[qix].prio[pix].migration_limit = 0;
	    }
	}
    }
    else {
	if (balance_info.last_active_runqs < current_active)
	    ERTS_BLNCE_SAVE_RISE(current_active, mmax_len, scheds_reds);
    all_active:

	active = blnc_no_rqs;

	for (qix = 0; qix < blnc_no_rqs; qix++) {

	    if (full_scheds_reds > 0) {
		/* Calculate availability compared to other schedulers */
		if (!(run_queue_info[qix].flags & ERTS_RUNQ_FLG_OUT_OF_WORK)) {
		    Sint64 tmp = ((Sint64) run_queue_info[qix].full_reds
				  * (Sint64) full_scheds);
		    for (pix = 0; pix < ERTS_NO_PRIO_LEVELS; pix++) {
			Sint64 avail = run_queue_info[qix].prio[pix].avail;
			avail = (avail*tmp)/full_scheds_reds;
			ASSERT(avail >= 0);
			run_queue_info[qix].prio[pix].avail = (int) avail;
		    }
		}
	    }

	    /* Calculate average max length */
	    for (pix = 0; pix < ERTS_NO_PRIO_LEVELS; pix++) {
		run_queue_info[qix].prio[pix].emigrate_to = -1;
		run_queue_info[qix].prio[pix].immigrate_from = -1;
		avg.prio[pix].max_len += run_queue_info[qix].prio[pix].max_len;
		avg.prio[pix].avail += run_queue_info[qix].prio[pix].avail;
	    }

	}

	for (pix = 0; pix < ERTS_NO_PRIO_LEVELS; pix++) {
	    int max_len = avg.prio[pix].max_len;
	    if (max_len != 0) {
		int avail = avg.prio[pix].avail;
		if (avail != 0) {
		    max_len = (int) ((100*((Sint64) max_len) - 1)
				     / ((Sint64) avail)) + 1;
		    avg.prio[pix].max_len = max_len;
		    ASSERT(max_len >= 0);
		}
	    }
	}

	/* Calculate migration limits for all priority queues in all
	   run queues */
	for (qix = 0; qix < blnc_no_rqs; qix++) {
	    run_queue_info[qix].flags = 0; /* Reset for later use... */
	    for (pix = 0; pix < ERTS_NO_PRIO_LEVELS; pix++) {
		int limit;
		if (avg.prio[pix].max_len == 0
		    || run_queue_info[qix].prio[pix].avail == 0)
		    limit = 0;
		else
		    limit = (int) (((((Sint64) avg.prio[pix].max_len)
				     * ((Sint64) run_queue_info[qix].prio[pix].avail))
				    - 1)
				   / 100 + 1);
		run_queue_info[qix].prio[pix].migration_limit = limit;
	    }
	}

#if ERTS_HAVE_SCHED_UTIL_BALANCING_SUPPORT
    setup_migration_paths:
#endif

	/* Setup migration paths for all priorities */
	for (pix = 0; pix < ERTS_NO_PRIO_LEVELS; pix++) {
	    int low = 0, high = 0;
	    for (qix = 0; qix < blnc_no_rqs; qix++) {
		int len_diff = run_queue_info[qix].prio[pix].max_len;
		len_diff -= run_queue_info[qix].prio[pix].migration_limit;

#ifdef DBG_PRINT
if (pix == 2) erts_fprintf(stderr, "%d ", len_diff);
#endif

#if ERTS_HAVE_SCHED_UTIL_BALANCING_SUPPORT
		if (sched_util_balancing
		    && -ERTS_SCHED_UTIL_IGNORE_IMBALANCE_DIFF <= len_diff
		    && len_diff <= ERTS_SCHED_UTIL_IGNORE_IMBALANCE_DIFF) {
		    /* ignore minor imbalance */
		    len_diff = 0;
		}
#endif
		    
		run_queue_compare[qix].qix = qix;
		run_queue_compare[qix].len = len_diff;
		if (len_diff != 0) {
		    if (len_diff < 0)
			low++;
		    else
			high++;
		}
	    }
#ifdef DBG_PRINT
if (pix == 2) erts_fprintf(stderr, "\n");
#endif
	    if (low && high) {
		int from_qix;
		int to_qix;
		int eof = 0;
		int eot = 0;
		int tix = 0;
		int fix = blnc_no_rqs-1;
		qsort(run_queue_compare,
		      blnc_no_rqs,
		      sizeof(ErtsRunQueueCompare),
		      rqc_len_cmp);

		while (1) {
		    if (run_queue_compare[fix].len <= 0)
			eof = 1;
		    if (run_queue_compare[tix].len >= 0)
			eot = 1;
		    if (eof || eot)
			break;
		    from_qix = run_queue_compare[fix].qix;
		    to_qix = run_queue_compare[tix].qix;
		    if (run_queue_info[from_qix].prio[pix].avail == 0) {
			ERTS_SET_RUNQ_FLG_EVACUATE(run_queue_info[from_qix].flags,
						   pix);
			ERTS_SET_RUNQ_FLG_EVACUATE(run_queue_info[to_qix].flags,
						   pix);
		    }
		    ERTS_SET_RUNQ_FLG_EMIGRATE(run_queue_info[from_qix].flags, pix);
		    ERTS_SET_RUNQ_FLG_IMMIGRATE(run_queue_info[to_qix].flags, pix);
		    run_queue_info[from_qix].prio[pix].emigrate_to = to_qix;
		    run_queue_info[to_qix].prio[pix].immigrate_from = from_qix;
		    tix++;
		    fix--;

#ifdef DBG_PRINT
if (pix == 2) erts_fprintf(stderr, "%d >--> %d\n", from_qix, to_qix);
#endif
		}

		if (!eot && eof) {
		    if (fix < blnc_no_rqs-1)
			fix++;

		    if (run_queue_compare[fix].len > 0) {
			int fix2 = -1;
			while (tix < fix) {
			    if (run_queue_compare[tix].len >= 0)
				break;
			    if (fix2 < fix)
				fix2 = blnc_no_rqs-1;
			    from_qix = run_queue_compare[fix2].qix;
			    to_qix = run_queue_compare[tix].qix;
			    ASSERT(to_qix != from_qix);
			    if (run_queue_info[from_qix].prio[pix].avail == 0)
				ERTS_SET_RUNQ_FLG_EVACUATE(run_queue_info[to_qix].flags,
							   pix);
			    ERTS_SET_RUNQ_FLG_IMMIGRATE(run_queue_info[to_qix].flags, pix);
			    run_queue_info[to_qix].prio[pix].immigrate_from = from_qix;
			    tix++;
			    fix2--;
#ifdef DBG_PRINT
if (pix == 2) erts_fprintf(stderr, "%d  --> %d\n", from_qix, to_qix);
#endif
			}
		    }
		}
		else if (!eof && eot) {
		    if (tix > 0)
			tix--;
		    if (run_queue_compare[tix].len < 0) {
			int tix2 = 0;
			while (tix < fix) {
			    if (run_queue_compare[fix].len <= 0)
				break;
			    if (tix2 > tix)
				tix2 = 0;
			    from_qix = run_queue_compare[fix].qix;
			    to_qix = run_queue_compare[tix2].qix;
			    ASSERT(to_qix != from_qix);
			    if (run_queue_info[from_qix].prio[pix].avail == 0)
				ERTS_SET_RUNQ_FLG_EVACUATE(run_queue_info[from_qix].flags,
							   pix);
			    ERTS_SET_RUNQ_FLG_EMIGRATE(run_queue_info[from_qix].flags, pix);
			    run_queue_info[from_qix].prio[pix].emigrate_to = to_qix;
			    fix--;
			    tix2++;
#ifdef DBG_PRINT
if (pix == 2) erts_fprintf(stderr, "%d >--  %d\n", from_qix, to_qix);
#endif

			}
		    }
		}
	    }
	}

#ifdef DBG_PRINT
erts_fprintf(stderr, "--------------------------------\n");
#endif
    }

    balance_info.last_active_runqs = active;
    set_no_active_runqs(active);

    balance_info.halftime = 1;
    new_mpaths = alloc_mpaths();

    /* Write migration paths */

    for (qix = 0; qix < blnc_no_rqs; qix++) {
	int mqix;
	Uint32 flags = run_queue_info[qix].flags;
	ErtsMigrationPath *mp = &new_mpaths->mpath[qix];

#if ERTS_HAVE_SCHED_UTIL_BALANCING_SUPPORT
	mp->sched_util = sched_util_balancing;
#endif
	mp->flags = flags;
	mp->misc_evac_runq = NULL;

	for (pix = 0; pix < ERTS_NO_PRIO_LEVELS; pix++) {
	    if (!(ERTS_CHK_RUNQ_FLG_EMIGRATE(flags, pix)
		  | ERTS_CHK_RUNQ_FLG_IMMIGRATE(flags, pix))) {
		ASSERT(run_queue_info[qix].prio[pix].immigrate_from < 0);
		ASSERT(run_queue_info[qix].prio[pix].emigrate_to < 0);
		mp->prio[pix].limit.this = -1;
		mp->prio[pix].limit.other = -1;
		mp->prio[pix].runq = NULL;
		mp->prio[pix].flags = 0;
	    }
	    else if (ERTS_CHK_RUNQ_FLG_EMIGRATE(flags, pix)) {
		ASSERT(!ERTS_CHK_RUNQ_FLG_IMMIGRATE(flags, pix));
		ASSERT(run_queue_info[qix].prio[pix].immigrate_from < 0);
		ASSERT(run_queue_info[qix].prio[pix].emigrate_to >= 0);

		mqix = run_queue_info[qix].prio[pix].emigrate_to;
		mp->prio[pix].limit.this
		    = run_queue_info[qix].prio[pix].migration_limit;
		mp->prio[pix].limit.other
		    = run_queue_info[mqix].prio[pix].migration_limit;
		mp->prio[pix].runq = ERTS_RUNQ_IX(mqix);
		mp->prio[pix].flags = run_queue_info[mqix].flags;
	    }
	    else {
		ASSERT(ERTS_CHK_RUNQ_FLG_IMMIGRATE(flags, pix));
		ASSERT(run_queue_info[qix].prio[pix].emigrate_to < 0);
		ASSERT(run_queue_info[qix].prio[pix].immigrate_from >= 0);

		mqix = run_queue_info[qix].prio[pix].immigrate_from;
		mp->prio[pix].limit.this
		    = run_queue_info[qix].prio[pix].migration_limit;
		mp->prio[pix].limit.other
		    = run_queue_info[mqix].prio[pix].migration_limit;
		mp->prio[pix].runq = ERTS_RUNQ_IX(mqix);
		mp->prio[pix].flags = run_queue_info[mqix].flags;
	    }
	}
    }

    old_mpaths = erts_get_migration_paths_managed();

    /* Keep offline run-queues as is */
    for (qix = blnc_no_rqs; qix < erts_no_schedulers; qix++) {
	ErtsMigrationPath *nmp = &new_mpaths->mpath[qix];
	ErtsMigrationPath *omp = &old_mpaths->mpath[qix];

	nmp->flags = omp->flags;
	nmp->misc_evac_runq = omp->misc_evac_runq;

	for (pix = 0; pix < ERTS_NO_PRIO_LEVELS; pix++) {
	    nmp->prio[pix].limit.this = omp->prio[pix].limit.this;
	    nmp->prio[pix].limit.other = omp->prio[pix].limit.other;
	    nmp->prio[pix].runq = omp->prio[pix].runq;
	    nmp->prio[pix].flags = omp->prio[pix].flags;
	}
    }


    /* Publish new migration paths... */
    erts_atomic_set_wb(&erts_migration_paths, (erts_aint_t) new_mpaths);

    /* Reset balance statistics in all online queues */
    for (qix = 0; qix < blnc_no_rqs; qix++) {
	Uint32 flags = run_queue_info[qix].flags;
	ErtsRunQueue *rq = ERTS_RUNQ_IX(qix);

	erts_smp_runq_lock(rq);
	ASSERT(!(flags & ERTS_RUNQ_FLG_OUT_OF_WORK));
	if (rq->waiting)
	    flags |= ERTS_RUNQ_FLG_OUT_OF_WORK;

	rq->full_reds_history_sum
	    = run_queue_info[qix].full_reds_history_sum;
	rq->full_reds_history[freds_hist_ix]
	    = run_queue_info[qix].full_reds_history_change;

	ERTS_DBG_CHK_FULL_REDS_HISTORY(rq);

	rq->out_of_work_count = 0;
	(void) ERTS_RUNQ_FLGS_READ_BSET(rq, ERTS_RUNQ_FLGS_MIGRATION_INFO, flags);

	rq->max_len = rq->len;
	for (pix = 0; pix < ERTS_NO_PRIO_LEVELS; pix++) {
	    ErtsRunQueueInfo *rqi;
	    rqi = (pix == ERTS_PORT_PRIO_LEVEL
		   ? &rq->ports.info
		   : &rq->procs.prio_info[pix]);
	    erts_smp_reset_max_len(rq, rqi);
	    rqi->reds = 0;
	}

	rq->check_balance_reds = ERTS_RUNQ_CALL_CHECK_BALANCE_REDS;
	erts_smp_runq_unlock(rq);
    }

    erts_smp_atomic32_set_nob(&balance_info.checking_balance, 0);

    balance_info.n++;
    retire_mpaths(old_mpaths);
    erts_smp_mtx_unlock(&balance_info.update_mtx);

    erts_smp_runq_lock(c_rq);
}

static void
change_no_used_runqs(int used)
{
    ErtsMigrationPaths *new_mpaths, *old_mpaths;
    int qix;
    erts_smp_mtx_lock(&balance_info.update_mtx);
    set_no_used_runqs(used);

    old_mpaths = erts_get_migration_paths_managed();
    new_mpaths = alloc_mpaths();

    /* Write migration paths... */

    for (qix = 0; qix < used; qix++) {
	int pix;
	ErtsMigrationPath *omp = &old_mpaths->mpath[qix];
	ErtsMigrationPath *nmp = &new_mpaths->mpath[qix];

	nmp->flags = omp->flags & ~ERTS_RUNQ_FLGS_MIGRATION_QMASKS;
	nmp->misc_evac_runq = NULL;

	for (pix = 0; pix < ERTS_NO_PRIO_LEVELS; pix++) {
	    nmp->prio[pix].limit.this = -1;
	    nmp->prio[pix].limit.other = -1;
	    nmp->prio[pix].runq = NULL;
	    nmp->prio[pix].flags = 0;
	}
    }
    for (qix = used; qix < erts_no_run_queues; qix++) {
	int pix;
	ErtsRunQueue *to_rq = ERTS_RUNQ_IX(qix % used);
	ErtsMigrationPath *nmp = &new_mpaths->mpath[qix];

	nmp->flags = (ERTS_RUNQ_FLGS_EMIGRATE_QMASK
		      | ERTS_RUNQ_FLGS_EVACUATE_QMASK);
	nmp->misc_evac_runq = to_rq;
	for (pix = 0; pix < ERTS_NO_PRIO_LEVELS; pix++) {
	    nmp->prio[pix].limit.this = -1;
	    nmp->prio[pix].limit.other = -1;
	    nmp->prio[pix].runq = to_rq;
	    nmp->prio[pix].flags = 0;
	}
    }

    /* ... and publish them. */
    erts_atomic_set_wb(&erts_migration_paths, (erts_aint_t) new_mpaths);

    retire_mpaths(old_mpaths);

    /* Make sure that we balance soon... */
    balance_info.forced_check_balance = 1;

    erts_smp_mtx_unlock(&balance_info.update_mtx);

    erts_smp_runq_lock(ERTS_RUNQ_IX(0));
    ERTS_RUNQ_IX(0)->check_balance_reds = 0;
    erts_smp_runq_unlock(ERTS_RUNQ_IX(0));
}


#endif /* #ifdef ERTS_SMP */

Uint
erts_debug_nbalance(void)
{
#ifdef ERTS_SMP
    Uint n;
    erts_smp_mtx_lock(&balance_info.update_mtx);
    n = balance_info.n;
    erts_smp_mtx_unlock(&balance_info.update_mtx);
    return n;
#else
    return 0;
#endif
}

/* Wakeup other schedulers */

typedef enum {
    ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_VERY_HIGH,
    ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_HIGH,
    ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_MEDIUM,
    ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_LOW,
    ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_VERY_LOW
} ErtsSchedWakeupOtherThreshold;

typedef enum {
    ERTS_SCHED_WAKEUP_OTHER_TYPE_DEFAULT,
    ERTS_SCHED_WAKEUP_OTHER_TYPE_LEGACY
} ErtsSchedWakeupOtherType;

/* Default */

#define ERTS_WAKEUP_OTHER_LIMIT_VERY_HIGH (200*CONTEXT_REDS)
#define ERTS_WAKEUP_OTHER_LIMIT_HIGH (50*CONTEXT_REDS)
#define ERTS_WAKEUP_OTHER_LIMIT_MEDIUM (10*CONTEXT_REDS)
#define ERTS_WAKEUP_OTHER_LIMIT_LOW (CONTEXT_REDS)
#define ERTS_WAKEUP_OTHER_LIMIT_VERY_LOW (CONTEXT_REDS/10)

#define ERTS_WAKEUP_OTHER_DEC_SHIFT_VERY_HIGH 3
#define ERTS_WAKEUP_OTHER_DEC_SHIFT_HIGH 1
#define ERTS_WAKEUP_OTHER_DEC_SHIFT_MEDIUM 0
#define ERTS_WAKEUP_OTHER_DEC_SHIFT_LOW -2
#define ERTS_WAKEUP_OTHER_DEC_SHIFT_VERY_LOW -5 

#define ERTS_WAKEUP_OTHER_DEC_SHIFT 2
#define ERTS_WAKEUP_OTHER_FIXED_INC (CONTEXT_REDS/10)

/* Legacy */

#define ERTS_WAKEUP_OTHER_LIMIT_VERY_HIGH_LEGACY (200*CONTEXT_REDS)
#define ERTS_WAKEUP_OTHER_LIMIT_HIGH_LEGACY (50*CONTEXT_REDS)
#define ERTS_WAKEUP_OTHER_LIMIT_MEDIUM_LEGACY (10*CONTEXT_REDS)
#define ERTS_WAKEUP_OTHER_LIMIT_LOW_LEGACY (CONTEXT_REDS)
#define ERTS_WAKEUP_OTHER_LIMIT_VERY_LOW_LEGACY (CONTEXT_REDS/10)

#define ERTS_WAKEUP_OTHER_DEC_LEGACY 10
#define ERTS_WAKEUP_OTHER_FIXED_INC_LEGACY (CONTEXT_REDS/10)

#ifdef ERTS_SMP

static struct {
    ErtsSchedWakeupOtherThreshold threshold;
    ErtsSchedWakeupOtherType type;
    int limit;
    int dec_shift;
    int dec_mask;
    void (*check)(ErtsRunQueue *rq, Uint32 flags);
} wakeup_other;

static void
wakeup_other_check(ErtsRunQueue *rq, Uint32 flags)
{
    int wo_reds = rq->wakeup_other_reds;
    if (wo_reds) {
	int left_len = rq->len - 1;
	if (left_len < 1) {
	    int wo_reduce = wo_reds << wakeup_other.dec_shift;
	    wo_reduce &= wakeup_other.dec_mask;
	    rq->wakeup_other -= wo_reduce;
	    if (rq->wakeup_other < 0)
		rq->wakeup_other = 0;
	}
	else {
	    rq->wakeup_other += (left_len*wo_reds
				 + ERTS_WAKEUP_OTHER_FIXED_INC);
	    if (rq->wakeup_other > wakeup_other.limit) {
#ifdef ERTS_DIRTY_SCHEDULERS
		if (ERTS_RUNQ_IX_IS_DIRTY(rq->ix) && rq->waiting)
		    wake_dirty_schedulers(rq, 1);
		else
#endif
		{
		    int empty_rqs =
			erts_smp_atomic32_read_acqb(&no_empty_run_queues);
		    if (flags & ERTS_RUNQ_FLG_PROTECTED)
			(void) ERTS_RUNQ_FLGS_UNSET(rq, ERTS_RUNQ_FLG_PROTECTED);
		    if (empty_rqs != 0)
			wake_scheduler_on_empty_runq(rq);
		    rq->wakeup_other = 0;
		}
	    }
	}
	rq->wakeup_other_reds = 0;
    }
}

static void
wakeup_other_set_limit(void)
{
    switch (wakeup_other.threshold) {
    case ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_VERY_HIGH:
	wakeup_other.limit = ERTS_WAKEUP_OTHER_LIMIT_VERY_HIGH;
	wakeup_other.dec_shift = ERTS_WAKEUP_OTHER_DEC_SHIFT_VERY_HIGH;
	break;
    case ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_HIGH:
	wakeup_other.limit = ERTS_WAKEUP_OTHER_LIMIT_HIGH;
	wakeup_other.dec_shift = ERTS_WAKEUP_OTHER_DEC_SHIFT_HIGH;
	break;
    case ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_MEDIUM:
	wakeup_other.limit = ERTS_WAKEUP_OTHER_LIMIT_MEDIUM;
	wakeup_other.dec_shift = ERTS_WAKEUP_OTHER_DEC_SHIFT_MEDIUM;
	break;
    case ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_LOW:
	wakeup_other.limit = ERTS_WAKEUP_OTHER_LIMIT_LOW;
	wakeup_other.dec_shift = ERTS_WAKEUP_OTHER_DEC_SHIFT_LOW;
	break;
    case ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_VERY_LOW:
	wakeup_other.limit = ERTS_WAKEUP_OTHER_LIMIT_VERY_LOW;
	wakeup_other.dec_shift = ERTS_WAKEUP_OTHER_DEC_SHIFT_VERY_LOW;
	break;
    }
    if (wakeup_other.dec_shift < 0)
	wakeup_other.dec_mask = (1 << (sizeof(wakeup_other.dec_mask)*8
				       + wakeup_other.dec_shift)) - 1;
    else {
	wakeup_other.dec_mask = 0;
	wakeup_other.dec_mask = ~wakeup_other.dec_mask;
    }
}

static void
wakeup_other_check_legacy(ErtsRunQueue *rq, Uint32 flags)
{
    int wo_reds = rq->wakeup_other_reds;
    if (wo_reds) {
	erts_aint32_t len = rq->len;
	if (len < 2) {
	    rq->wakeup_other -= ERTS_WAKEUP_OTHER_DEC_LEGACY*wo_reds;
	    if (rq->wakeup_other < 0)
		rq->wakeup_other = 0;
	}
	else if (rq->wakeup_other < wakeup_other.limit)
	    rq->wakeup_other += len*wo_reds + ERTS_WAKEUP_OTHER_FIXED_INC_LEGACY;
	else {
	    if (flags & ERTS_RUNQ_FLG_PROTECTED)
		(void) ERTS_RUNQ_FLGS_UNSET(rq, ERTS_RUNQ_FLG_PROTECTED);
	    if (erts_smp_atomic32_read_acqb(&no_empty_run_queues) != 0) {
		wake_scheduler_on_empty_runq(rq);
		rq->wakeup_other = 0;
	    }
	    rq->wakeup_other = 0;
	}
    }
    rq->wakeup_other_reds = 0;
}

static void
wakeup_other_set_limit_legacy(void)
{
    switch (wakeup_other.threshold) {
    case ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_VERY_HIGH:
	wakeup_other.limit = ERTS_WAKEUP_OTHER_LIMIT_VERY_HIGH_LEGACY;
	break;
    case ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_HIGH:
	wakeup_other.limit = ERTS_WAKEUP_OTHER_LIMIT_HIGH_LEGACY;
	break;
    case ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_MEDIUM:
	wakeup_other.limit = ERTS_WAKEUP_OTHER_LIMIT_MEDIUM_LEGACY;
	break;
    case ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_LOW:
	wakeup_other.limit = ERTS_WAKEUP_OTHER_LIMIT_LOW_LEGACY;
	break;
    case ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_VERY_LOW:
	wakeup_other.limit = ERTS_WAKEUP_OTHER_LIMIT_VERY_LOW_LEGACY;
	break;
    }
}

static void
set_wakeup_other_data(void)
{
    switch (wakeup_other.type) {
    case ERTS_SCHED_WAKEUP_OTHER_TYPE_DEFAULT:
	wakeup_other.check = wakeup_other_check;
	wakeup_other_set_limit();
	break;
    case ERTS_SCHED_WAKEUP_OTHER_TYPE_LEGACY:
	wakeup_other.check = wakeup_other_check_legacy;
	wakeup_other_set_limit_legacy();
	break;
    }
}

static int
no_runqs_to_supervise(void)
{
    int used;
    erts_aint32_t nerq = erts_smp_atomic32_read_acqb(&no_empty_run_queues);
    if (nerq <= 0)
	return 0;
    get_no_runqs(NULL, &used);
    if (nerq >= used)
	return 0;
    return used;
}

static void *
runq_supervisor(void *unused)
{
    while (1) {
	int ix, no_rqs;

	erts_milli_sleep(erts_runq_supervision_interval);
	no_rqs = no_runqs_to_supervise();
	if (!no_rqs) {
	    erts_atomic_set_nob(&runq_supervisor_sleeping, 1);
	    while (1) {
		ethr_event_reset(&runq_supervision_event);
		no_rqs = no_runqs_to_supervise();
		if (no_rqs) {
		    erts_atomic_set_nob(&runq_supervisor_sleeping, 0);
		    break;
		}
		ethr_event_wait(&runq_supervision_event);
	    }
	}

	for (ix = 0; ix < no_rqs; ix++) {
	    ErtsRunQueue *rq = ERTS_RUNQ_IX(ix);
	    if (ERTS_RUNQ_FLGS_GET(rq) & ERTS_RUNQ_FLG_NONEMPTY) {
		erts_smp_runq_lock(rq);
		if (rq->len != 0)
		    wake_scheduler_on_empty_runq(rq); /* forced wakeup... */
		erts_smp_runq_unlock(rq);
	    }
	}
    }
    return NULL;
}

#endif

void
erts_early_init_scheduling(int no_schedulers)
{
    aux_work_timeout_early_init(no_schedulers);
#ifdef ERTS_SMP
    wakeup_other.threshold = ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_MEDIUM;
    wakeup_other.type = ERTS_SCHED_WAKEUP_OTHER_TYPE_DEFAULT;
#endif
#ifndef ERTS_SCHED_MIN_SPIN
    sched_busy_wait.sys_schedule = ERTS_SCHED_SYS_SLEEP_SPINCOUNT_MEDIUM;
    sched_busy_wait.tse = (ERTS_SCHED_SYS_SLEEP_SPINCOUNT_MEDIUM
			   * ERTS_SCHED_TSE_SLEEP_SPINCOUNT_FACT);
    sched_busy_wait.aux_work = (ERTS_SCHED_SYS_SLEEP_SPINCOUNT_MEDIUM
				* ERTS_SCHED_AUX_WORK_SLEEP_SPINCOUNT_FACT_MEDIUM);
#else
    sched_busy_wait.sys_schedule = ERTS_SCHED_SYS_SLEEP_SPINCOUNT_NONE;
    sched_busy_wait.tse = ERTS_SCHED_SYS_SLEEP_SPINCOUNT_NONE;
    sched_busy_wait.aux_work = ERTS_SCHED_SYS_SLEEP_SPINCOUNT_NONE;
#endif
}

int
erts_sched_set_wakeup_other_thresold(char *str)
{
#ifdef ERTS_SMP
    ErtsSchedWakeupOtherThreshold threshold;
    if (sys_strcmp(str, "very_high") == 0)
	threshold = ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_VERY_HIGH;
    else if (sys_strcmp(str, "high") == 0)
	threshold = ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_HIGH;
    else if (sys_strcmp(str, "medium") == 0)
	threshold = ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_MEDIUM;
    else if (sys_strcmp(str, "low") == 0)
	threshold = ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_LOW;
    else if (sys_strcmp(str, "very_low") == 0)
	threshold = ERTS_SCHED_WAKEUP_OTHER_THRESHOLD_VERY_LOW;
    else
	return EINVAL;
    wakeup_other.threshold = threshold;
    set_wakeup_other_data();
    return 0;
#else
    if (sys_strcmp(str, "very_high") == 0 || sys_strcmp(str, "high") == 0 ||
	sys_strcmp(str, "medium") == 0 || sys_strcmp(str, "low") == 0 ||
	sys_strcmp(str, "very_low") == 0) {
	return 0;
    } 
    return EINVAL;
#endif
}

int
erts_sched_set_wakeup_other_type(char *str)
{
#ifdef ERTS_SMP
    ErtsSchedWakeupOtherType type;
    if (sys_strcmp(str, "default") == 0)
	type = ERTS_SCHED_WAKEUP_OTHER_TYPE_DEFAULT;
    else if (sys_strcmp(str, "legacy") == 0)
	type = ERTS_SCHED_WAKEUP_OTHER_TYPE_LEGACY;
    else
	return EINVAL;
    wakeup_other.type = type;
    return 0;
#else
    if (sys_strcmp(str, "default") == 0 || sys_strcmp(str, "legacy") == 0) {
	return 0;
    } 
    return EINVAL;
#endif
}

int
erts_sched_set_busy_wait_threshold(char *str)
{
    int sys_sched;
    int aux_work_fact;

    if (sys_strcmp(str, "very_long") == 0) {
	sys_sched = ERTS_SCHED_SYS_SLEEP_SPINCOUNT_VERY_LONG;
	aux_work_fact = ERTS_SCHED_AUX_WORK_SLEEP_SPINCOUNT_FACT_VERY_LONG;
    }
    else if (sys_strcmp(str, "long") == 0) {
	sys_sched = ERTS_SCHED_SYS_SLEEP_SPINCOUNT_LONG;
	aux_work_fact = ERTS_SCHED_AUX_WORK_SLEEP_SPINCOUNT_FACT_LONG;
    }
    else if (sys_strcmp(str, "medium") == 0) {
	sys_sched = ERTS_SCHED_SYS_SLEEP_SPINCOUNT_MEDIUM;
	aux_work_fact = ERTS_SCHED_AUX_WORK_SLEEP_SPINCOUNT_FACT_MEDIUM;
    }
    else if (sys_strcmp(str, "short") == 0) {
	sys_sched = ERTS_SCHED_SYS_SLEEP_SPINCOUNT_SHORT;
	aux_work_fact = ERTS_SCHED_AUX_WORK_SLEEP_SPINCOUNT_FACT_SHORT;
    }
    else if (sys_strcmp(str, "very_short") == 0) {
	sys_sched = ERTS_SCHED_SYS_SLEEP_SPINCOUNT_VERY_SHORT;
	aux_work_fact = ERTS_SCHED_AUX_WORK_SLEEP_SPINCOUNT_FACT_VERY_SHORT;
    }
    else if (sys_strcmp(str, "none") == 0) {
	sys_sched = ERTS_SCHED_SYS_SLEEP_SPINCOUNT_NONE;
	aux_work_fact = ERTS_SCHED_AUX_WORK_SLEEP_SPINCOUNT_FACT_NONE;
    }
    else {
	return EINVAL;
    }

    sched_busy_wait.sys_schedule = sys_sched;
    sched_busy_wait.tse = sys_sched*ERTS_SCHED_TSE_SLEEP_SPINCOUNT_FACT;
    sched_busy_wait.aux_work = sys_sched*aux_work_fact;

    return 0;
}

int
erts_sched_set_wake_cleanup_threshold(char *str)
{
    if (sys_strcmp(str, "very_lazy") == 0)
	thr_prgr_later_cleanup_op_threshold = ERTS_THR_PRGR_LATER_CLEANUP_OP_THRESHOLD_VERY_LAZY;
    else if (sys_strcmp(str, "lazy") == 0)
	thr_prgr_later_cleanup_op_threshold = ERTS_THR_PRGR_LATER_CLEANUP_OP_THRESHOLD_LAZY;
    else if (sys_strcmp(str, "medium") == 0)
	thr_prgr_later_cleanup_op_threshold = ERTS_THR_PRGR_LATER_CLEANUP_OP_THRESHOLD_MEDIUM;
    else if (sys_strcmp(str, "eager") == 0)
	thr_prgr_later_cleanup_op_threshold = ERTS_THR_PRGR_LATER_CLEANUP_OP_THRESHOLD_EAGER;
    else if (sys_strcmp(str, "very_eager") == 0)
	thr_prgr_later_cleanup_op_threshold = ERTS_THR_PRGR_LATER_CLEANUP_OP_THRESHOLD_VERY_EAGER;
    else
	return EINVAL;
    return 0;
}

static void
init_aux_work_data(ErtsAuxWorkData *awdp, ErtsSchedulerData *esdp, char *dawwp)
{
    if (!esdp)
	awdp->sched_id = 0;
#ifdef ERTS_DIRTY_SCHEDULERS
    else if (ERTS_SCHEDULER_IS_DIRTY(esdp))
	awdp->sched_id = (int) ERTS_DIRTY_SCHEDULER_NO(esdp);
#endif
    else
	awdp->sched_id = (int) esdp->no;
    awdp->esdp = esdp;
    awdp->ssi = esdp ? esdp->ssi : NULL;
#ifdef ERTS_SMP
    awdp->latest_wakeup = ERTS_THR_PRGR_VAL_FIRST;
    awdp->misc.thr_prgr = ERTS_THR_PRGR_VAL_WAITING;
    awdp->dd.thr_prgr = ERTS_THR_PRGR_VAL_WAITING;
    awdp->dd.completed_callback = NULL;
    awdp->dd.completed_arg = NULL;
    awdp->later_op.thr_prgr = ERTS_THR_PRGR_VAL_FIRST;
    awdp->later_op.size = 0;
    awdp->later_op.first = NULL;
    awdp->later_op.last = NULL;
#endif
#ifdef ERTS_USE_ASYNC_READY_Q
#ifdef ERTS_SMP
    awdp->async_ready.need_thr_prgr = 0;
    awdp->async_ready.thr_prgr = ERTS_THR_PRGR_VAL_WAITING;
#endif
    awdp->async_ready.queue = NULL;
#endif
#ifdef ERTS_SMP
    awdp->delayed_wakeup.next = ERTS_DELAYED_WAKEUP_INFINITY;
    if (!dawwp) {
	awdp->delayed_wakeup.job = NULL;
	awdp->delayed_wakeup.sched2jix = NULL;
	awdp->delayed_wakeup.jix = -1;
    }
    else {
	int i;
	awdp->delayed_wakeup.job = (ErtsDelayedAuxWorkWakeupJob *) dawwp;
	dawwp += sizeof(ErtsDelayedAuxWorkWakeupJob)*(erts_no_schedulers+1);
	awdp->delayed_wakeup.sched2jix = (int *) dawwp;
	awdp->delayed_wakeup.jix = -1;
	for (i = 0; i <= erts_no_schedulers; i++)
	    awdp->delayed_wakeup.sched2jix[i] = -1;
    }
#endif
}

static void
init_scheduler_data(ErtsSchedulerData* esdp, int num,
		    ErtsSchedulerSleepInfo* ssi,
		    ErtsRunQueue* runq,
		    char** daww_ptr, size_t daww_sz)
{
#ifdef ERTS_SMP
    erts_bits_init_state(&esdp->erl_bits_state);
    esdp->match_pseudo_process = NULL;
    esdp->free_process = NULL;
#endif
    esdp->x_reg_array =
	erts_alloc_permanent_cache_aligned(ERTS_ALC_T_BEAM_REGISTER,
					   ERTS_X_REGS_ALLOCATED *
					   sizeof(Eterm));
    esdp->f_reg_array =
	erts_alloc_permanent_cache_aligned(ERTS_ALC_T_BEAM_REGISTER,
					   MAX_REG * sizeof(FloatDef));
#if !HEAP_ON_C_STACK
    esdp->num_tmp_heap_used = 0;
#endif
#ifdef ERTS_DIRTY_SCHEDULERS
    if (ERTS_RUNQ_IX_IS_DIRTY(runq->ix)) {
	esdp->no = 0;
	ERTS_DIRTY_SCHEDULER_NO(esdp) = (Uint) num;
    }
    else {
	esdp->no = (Uint) num;
	ERTS_DIRTY_SCHEDULER_NO(esdp) = 0;
    }
#else
    esdp->no = (Uint) num;
#endif
    esdp->ssi = ssi;
    esdp->current_process = NULL;
    esdp->current_port = NULL;

    esdp->virtual_reds = 0;
    esdp->cpu_id = -1;

    erts_init_atom_cache_map(&esdp->atom_cache_map);

    esdp->run_queue = runq;
    esdp->run_queue->scheduler = esdp;

    if (daww_ptr) {
	init_aux_work_data(&esdp->aux_work_data, esdp, *daww_ptr);
#ifdef ERTS_SMP
	*daww_ptr += daww_sz;
#endif
    }

    esdp->reductions = 0;

    init_sched_wall_time(&esdp->sched_wall_time);
    erts_port_task_handle_init(&esdp->nosuspend_port_task_handle);
}

void
erts_init_scheduling(int no_schedulers, int no_schedulers_online
#ifdef ERTS_DIRTY_SCHEDULERS
		     , int no_dirty_cpu_schedulers, int no_dirty_cpu_schedulers_online,
		     int no_dirty_io_schedulers
#endif
		     )
{
    int ix, n, no_ssi;
    char *daww_ptr;
    size_t daww_sz;
    size_t size_runqs;

    init_misc_op_list_alloc();
    init_proc_sys_task_queues_alloc();

#ifdef ERTS_SMP
    set_wakeup_other_data();
#endif

#if ERTS_HAVE_SCHED_UTIL_BALANCING_SUPPORT
    if (erts_sched_balance_util)
	erts_sched_compact_load = 0;
#endif

    ASSERT(no_schedulers_online <= no_schedulers);
    ASSERT(no_schedulers_online >= 1);
    ASSERT(no_schedulers >= 1);
#ifdef ERTS_DIRTY_SCHEDULERS
    ASSERT(no_dirty_cpu_schedulers <= no_schedulers);
    ASSERT(no_dirty_cpu_schedulers >= 1);
    ASSERT(no_dirty_cpu_schedulers_online <= no_schedulers_online);
    ASSERT(no_dirty_cpu_schedulers_online >= 1);
#endif

    /* Create and initialize run queues */

    n = no_schedulers;
    size_runqs = sizeof(ErtsAlignedRunQueue) * (n + ERTS_NUM_DIRTY_RUNQS);
    erts_aligned_run_queues =
	erts_alloc_permanent_cache_aligned(ERTS_ALC_T_RUNQS, size_runqs);
#ifdef ERTS_SMP
#ifdef ERTS_DIRTY_SCHEDULERS
    erts_aligned_run_queues += ERTS_NUM_DIRTY_RUNQS;
#endif
    erts_smp_atomic32_init_nob(&no_empty_run_queues, 0);
#endif

    erts_no_run_queues = n;

    for (ix = -(ERTS_NUM_DIRTY_RUNQS); ix < n; ix++) {
	int pix, rix;
#ifdef ERTS_DIRTY_SCHEDULERS
	ErtsRunQueue *rq = ERTS_RUNQ_IX_IS_DIRTY(ix) ?
	    ERTS_DIRTY_RUNQ_IX(ix) : ERTS_RUNQ_IX(ix);
#else
	ErtsRunQueue *rq = ERTS_RUNQ_IX(ix);
#endif

	rq->ix = ix;

	/* make sure that the "extra" id correponds to the schedulers
	 * id if the esdp->no <-> ix+1 mapping change.
	 */

	erts_smp_mtx_init_x(&rq->mtx, "run_queue", make_small(ix + 1));
	erts_smp_cnd_init(&rq->cnd);

#ifdef ERTS_DIRTY_SCHEDULERS
#ifdef ERTS_SMP
	if (ERTS_RUNQ_IX_IS_DIRTY(ix))
	    erts_smp_spinlock_init(&rq->sleepers.lock, "dirty_run_queue_sleep_list");
	rq->sleepers.list = NULL;
#endif
#endif

	rq->waiting = 0;
	rq->woken = 0;
	ERTS_RUNQ_FLGS_INIT(rq, ERTS_RUNQ_FLG_NONEMPTY);
	rq->check_balance_reds = ERTS_RUNQ_CALL_CHECK_BALANCE_REDS;
	rq->full_reds_history_sum = 0;
	for (rix = 0; rix < ERTS_FULL_REDS_HISTORY_SIZE; rix++) {
	    rq->full_reds_history_sum += ERTS_RUNQ_CHECK_BALANCE_REDS_PER_SCHED;
	    rq->full_reds_history[rix] = ERTS_RUNQ_CHECK_BALANCE_REDS_PER_SCHED;
	}
	rq->out_of_work_count = 0;
	rq->max_len = 0;
	rq->len = 0;
	rq->wakeup_other = 0;
	rq->wakeup_other_reds = 0;
	rq->halt_in_progress = 0;

	rq->procs.pending_exiters = NULL;
	rq->procs.context_switches = 0;
	rq->procs.reductions = 0;

	for (pix = 0; pix < ERTS_NO_PROC_PRIO_LEVELS; pix++) {
	    erts_smp_atomic32_init_nob(&rq->procs.prio_info[pix].len, 0);
	    rq->procs.prio_info[pix].max_len = 0;
	    rq->procs.prio_info[pix].reds = 0;
	    if (pix < ERTS_NO_PROC_PRIO_LEVELS - 1) {
		rq->procs.prio[pix].first = NULL;
		rq->procs.prio[pix].last = NULL;
	    }
	}

	rq->misc.start = NULL;
	rq->misc.end = NULL;

	erts_smp_atomic32_init_nob(&rq->ports.info.len, 0);
	rq->ports.info.max_len = 0;
	rq->ports.info.reds = 0;
	rq->ports.start = NULL;
	rq->ports.end = NULL;

#if ERTS_HAVE_SCHED_UTIL_BALANCING_SUPPORT
	init_runq_sched_util(&rq->sched_util, erts_sched_balance_util);
#endif

    }

#ifdef ERTS_SMP

    if (erts_no_run_queues != 1) {
	run_queue_info = erts_alloc(ERTS_ALC_T_RUNQ_BLNS,
				    (sizeof(ErtsRunQueueBalance)
				    * erts_no_run_queues));
	run_queue_compare = erts_alloc(ERTS_ALC_T_RUNQ_BLNS,
				       (sizeof(ErtsRunQueueCompare)
					* erts_no_run_queues));
    }

#endif

    n = (int) no_schedulers;
    erts_no_schedulers = n;
#ifdef ERTS_DIRTY_SCHEDULERS
    erts_no_dirty_cpu_schedulers = no_dirty_cpu_schedulers;
    erts_no_dirty_io_schedulers = no_dirty_io_schedulers;
#endif

    /* Create and initialize scheduler sleep info */
#ifdef ERTS_SMP
    no_ssi = n+1;
#else
    no_ssi = 1;
#endif
    aligned_sched_sleep_info =
	erts_alloc_permanent_cache_aligned(
	    ERTS_ALC_T_SCHDLR_SLP_INFO,
	    no_ssi*sizeof(ErtsAlignedSchedulerSleepInfo));
    for (ix = 0; ix < no_ssi; ix++) {
	ErtsSchedulerSleepInfo *ssi = &aligned_sched_sleep_info[ix].ssi;
#ifdef ERTS_SMP
#if 0 /* no need to initialize these... */
	ssi->next = NULL;
	ssi->prev = NULL;
#endif
	erts_smp_atomic32_init_nob(&ssi->flags, 0);
	ssi->event = NULL; /* initialized in sched_thread_func */
#endif
	erts_atomic32_init_nob(&ssi->aux_work, 0);
    }

#ifdef ERTS_SMP
    aligned_sched_sleep_info++;

#ifdef ERTS_DIRTY_SCHEDULERS
    aligned_dirty_cpu_sched_sleep_info =
	erts_alloc_permanent_cache_aligned(
	    ERTS_ALC_T_SCHDLR_SLP_INFO,
	    no_dirty_cpu_schedulers*sizeof(ErtsAlignedSchedulerSleepInfo));
    for (ix = 0; ix < no_dirty_cpu_schedulers; ix++) {
	ErtsSchedulerSleepInfo *ssi = &aligned_dirty_cpu_sched_sleep_info[ix].ssi;
	erts_smp_atomic32_init_nob(&ssi->flags, 0);
	ssi->event = NULL; /* initialized in sched_dirty_cpu_thread_func */
	erts_atomic32_init_nob(&ssi->aux_work, 0);
    }
    aligned_dirty_io_sched_sleep_info =
	erts_alloc_permanent_cache_aligned(
	    ERTS_ALC_T_SCHDLR_SLP_INFO,
	    no_dirty_io_schedulers*sizeof(ErtsAlignedSchedulerSleepInfo));
    for (ix = 0; ix < no_dirty_io_schedulers; ix++) {
	ErtsSchedulerSleepInfo *ssi = &aligned_dirty_io_sched_sleep_info[ix].ssi;
	erts_smp_atomic32_init_nob(&ssi->flags, 0);
	ssi->event = NULL; /* initialized in sched_dirty_io_thread_func */
	erts_atomic32_init_nob(&ssi->aux_work, 0);
    }
#endif
#endif

    /* Create and initialize scheduler specific data */

#ifdef ERTS_SMP
    daww_sz = ERTS_ALC_CACHE_LINE_ALIGN_SIZE((sizeof(ErtsDelayedAuxWorkWakeupJob)
					      + sizeof(int))*(n+1));
    daww_ptr = erts_alloc_permanent_cache_aligned(ERTS_ALC_T_SCHDLR_DATA,
						  daww_sz*n);
#else
    daww_sz = 0;
    daww_ptr = NULL;
#endif

    erts_aligned_scheduler_data = 
	erts_alloc_permanent_cache_aligned(ERTS_ALC_T_SCHDLR_DATA,
					   n*sizeof(ErtsAlignedSchedulerData));					   

    for (ix = 0; ix < n; ix++) {
	ErtsSchedulerData *esdp = ERTS_SCHEDULER_IX(ix);
	init_scheduler_data(esdp, ix+1, ERTS_SCHED_SLEEP_INFO_IX(ix),
			    ERTS_RUNQ_IX(ix), &daww_ptr, daww_sz);
    }

#ifdef ERTS_DIRTY_SCHEDULERS
#ifdef ERTS_SMP
    erts_aligned_dirty_cpu_scheduler_data =
	erts_alloc_permanent_cache_aligned(
					   ERTS_ALC_T_SCHDLR_DATA,
					   no_dirty_cpu_schedulers*sizeof(ErtsAlignedSchedulerData));
    for (ix = 0; ix < no_dirty_cpu_schedulers; ix++) {
	ErtsSchedulerData *esdp = ERTS_DIRTY_CPU_SCHEDULER_IX(ix);
	init_scheduler_data(esdp, ix+1, ERTS_DIRTY_CPU_SCHED_SLEEP_INFO_IX(ix),
			    ERTS_DIRTY_CPU_RUNQ, NULL, 0);
    }
    erts_aligned_dirty_io_scheduler_data =
	erts_alloc_permanent_cache_aligned(
					   ERTS_ALC_T_SCHDLR_DATA,
					   no_dirty_io_schedulers*sizeof(ErtsAlignedSchedulerData));
    for (ix = 0; ix < no_dirty_io_schedulers; ix++) {
	ErtsSchedulerData *esdp = ERTS_DIRTY_IO_SCHEDULER_IX(ix);
	init_scheduler_data(esdp, ix+1, ERTS_DIRTY_IO_SCHED_SLEEP_INFO_IX(ix),
			    ERTS_DIRTY_IO_RUNQ, NULL, 0);
    }
#endif
#endif

    init_misc_aux_work();
#if !HALFWORD_HEAP
    init_swtreq_alloc();
#endif


#ifdef ERTS_SMP

    erts_atomic32_init_nob(&completed_dealloc_count, 0); /* debug only */

    aux_thread_aux_work_data =
	erts_alloc_permanent_cache_aligned(ERTS_ALC_T_SCHDLR_DATA,
					   sizeof(ErtsAuxWorkData));

    erts_smp_mtx_init(&schdlr_sspnd.mtx, "schdlr_sspnd");
    erts_smp_cnd_init(&schdlr_sspnd.cnd);

    erts_smp_atomic32_init_nob(&schdlr_sspnd.changing, 0);
    schdlr_sspnd.online = no_schedulers_online;
    schdlr_sspnd.curr_online = no_schedulers;
    schdlr_sspnd.msb.ongoing = 0;
    erts_smp_atomic32_init_nob(&schdlr_sspnd.active, no_schedulers);
#ifdef ERTS_DIRTY_SCHEDULERS
    erts_smp_atomic32_init_nob(&schdlr_sspnd.dirty_cpu_changing, 0);
    schdlr_sspnd.dirty_cpu_online = no_dirty_cpu_schedulers_online;
    schdlr_sspnd.dirty_cpu_curr_online = no_dirty_cpu_schedulers;
    erts_smp_atomic32_init_nob(&schdlr_sspnd.dirty_cpu_active, no_dirty_cpu_schedulers);
    erts_smp_atomic32_init_nob(&schdlr_sspnd.dirty_io_changing, 0);
    schdlr_sspnd.dirty_io_online = no_dirty_io_schedulers;
    schdlr_sspnd.dirty_io_curr_online = no_dirty_io_schedulers;
    erts_smp_atomic32_init_nob(&schdlr_sspnd.dirty_io_active, no_dirty_io_schedulers);
#endif
    schdlr_sspnd.msb.procs = NULL;
    init_no_runqs(no_schedulers_online, no_schedulers_online);
    balance_info.last_active_runqs = no_schedulers;
    erts_smp_mtx_init(&balance_info.update_mtx, "migration_info_update");
    balance_info.forced_check_balance = 0;
    balance_info.halftime = 1;
    balance_info.full_reds_history_index = 0;
    erts_smp_atomic32_init_nob(&balance_info.checking_balance, 0);
    balance_info.prev_rise.active_runqs = 0;
    balance_info.prev_rise.max_len = 0;
    balance_info.prev_rise.reds = 0;
    balance_info.n = 0;

    init_migration_paths();

    if (no_schedulers_online < no_schedulers) {
	change_no_used_runqs(no_schedulers_online);
	for (ix = no_schedulers_online; ix < erts_no_run_queues; ix++)
	    suspend_run_queue(ERTS_RUNQ_IX(ix));
    }

    schdlr_sspnd.wait_curr_online = no_schedulers_online;
    schdlr_sspnd.curr_online *= 2; /* Boot strapping... */
    ERTS_SCHDLR_SSPND_CHNG_SET((ERTS_SCHDLR_SSPND_CHNG_ONLN
				| ERTS_SCHDLR_SSPND_CHNG_WAITER), 0);
#ifdef ERTS_DIRTY_SCHEDULERS
    schdlr_sspnd.dirty_cpu_wait_curr_online = no_dirty_cpu_schedulers_online;
    schdlr_sspnd.dirty_cpu_curr_online *= 2;
    ERTS_SCHDLR_SSPND_DIRTY_CPU_CHNG_SET((ERTS_SCHDLR_SSPND_CHNG_ONLN
					  | ERTS_SCHDLR_SSPND_CHNG_WAITER), 0);
    for (ix = no_dirty_cpu_schedulers_online; ix < no_dirty_cpu_schedulers; ix++) {
	ErtsSchedulerData* esdp = ERTS_DIRTY_CPU_SCHEDULER_IX(ix);
	erts_smp_atomic32_read_bor_nob(&esdp->ssi->flags, ERTS_SSI_FLG_SUSPENDED);
    }

    schdlr_sspnd.dirty_io_wait_curr_online = no_dirty_io_schedulers;
    schdlr_sspnd.dirty_io_curr_online *= 2;
    ERTS_SCHDLR_SSPND_DIRTY_IO_CHNG_SET((ERTS_SCHDLR_SSPND_CHNG_ONLN
					 | ERTS_SCHDLR_SSPND_CHNG_WAITER), 0);
#endif

    erts_smp_atomic32_init_nob(&doing_sys_schedule, 0);

    init_misc_aux_work();

#else /* !ERTS_SMP */
    {
	ErtsSchedulerData *esdp;
	esdp = ERTS_SCHEDULER_IX(0);
	erts_scheduler_data = esdp;
#ifdef USE_THREADS
	erts_tsd_set(sched_data_key, (void *) esdp);
#endif
    }
    erts_no_schedulers = 1;
#ifdef ERTS_DIRTY_SCHEDULERS
    erts_no_dirty_cpu_schedulers = 0;
    erts_no_dirty_io_schedulers = 0;
#endif
#endif

    erts_smp_atomic32_init_nob(&function_calls, 0);

    /* init port tasks */
    erts_port_task_init();

    aux_work_timeout_late_init();

#ifndef ERTS_SMP
#ifdef ERTS_DO_VERIFY_UNUSED_TEMP_ALLOC
    erts_scheduler_data->verify_unused_temp_alloc
	= erts_alloc_get_verify_unused_temp_alloc(
	    &erts_scheduler_data->verify_unused_temp_alloc_data);
    ERTS_VERIFY_UNUSED_TEMP_ALLOC(NULL);
#endif
#endif

    erts_smp_atomic32_init_relb(&erts_halt_progress, -1);
    erts_halt_code = 0;

#if !defined(ERTS_SMP) && defined(ERTS_ENABLE_LOCK_CHECK)
    erts_lc_set_thread_name("scheduler 1");
#endif

}

ErtsRunQueue *
erts_schedid2runq(Uint id)
{
    int ix;
    ix = (int) id - 1;
    ASSERT(0 <= ix && ix < erts_no_run_queues);
    return ERTS_RUNQ_IX(ix);
}

#ifdef USE_THREADS

ErtsSchedulerData *
erts_get_scheduler_data(void)
{
    return (ErtsSchedulerData *) erts_tsd_get(sched_data_key);
}

#endif

static Process *
make_proxy_proc(Process *prev_proxy, Process *proc, erts_aint32_t prio)
{
    erts_aint32_t state;
    Process *proxy;
#ifdef ERTS_SMP
    ErtsRunQueue *rq = RUNQ_READ_RQ(&proc->run_queue);
#endif

    state = (ERTS_PSFLG_PROXY
	     | ERTS_PSFLG_IN_RUNQ
	     | (((erts_aint32_t) 1) << (prio + ERTS_PSFLGS_IN_PRQ_MASK_OFFSET))
	     | (prio << ERTS_PSFLGS_PRQ_PRIO_OFFSET)
	     | (prio << ERTS_PSFLGS_USR_PRIO_OFFSET)
	     | (prio << ERTS_PSFLGS_ACT_PRIO_OFFSET));

    if (prev_proxy) {
	proxy = prev_proxy;
	ASSERT(erts_smp_atomic32_read_nob(&proxy->state) & ERTS_PSFLG_PROXY);
	erts_smp_atomic32_set_nob(&proxy->state, state);
#ifdef ERTS_SMP
	RUNQ_SET_RQ(&proc->run_queue, rq);
#endif
    }
    else {
	proxy = erts_alloc(ERTS_ALC_T_PROC, sizeof(Process));
#ifdef DEBUG
	{
	    int i;
	    Uint32 *ui32 = (Uint32 *) (char *) proxy;
	    for (i = 0; i < sizeof(Process)/sizeof(Uint32); i++)
		ui32[i] = (Uint32) 0xdeadbeef;
	}
#endif
	erts_smp_atomic32_init_nob(&proxy->state, state);
#ifdef ERTS_SMP
	erts_smp_atomic_init_nob(&proxy->run_queue,
				 erts_smp_atomic_read_nob(&proc->run_queue));
#endif
    }

    proxy->common.id = proc->common.id;

    return proxy;
}

static ERTS_INLINE void
free_proxy_proc(Process *proxy)
{
    ASSERT(erts_smp_atomic32_read_nob(&proxy->state) & ERTS_PSFLG_PROXY);
    erts_free(ERTS_ALC_T_PROC, proxy);
}

#define ERTS_ENQUEUE_NOT 0
#define ERTS_ENQUEUE_NORMAL_QUEUE 1
#ifdef ERTS_DIRTY_SCHEDULERS
#define ERTS_ENQUEUE_DIRTY_CPU_QUEUE 2
#define ERTS_ENQUEUE_DIRTY_IO_QUEUE 3
#endif

static ERTS_INLINE int
check_enqueue_in_prio_queue(Process *c_p,
			    erts_aint32_t *prq_prio_p,
			    erts_aint32_t *newp,
			    erts_aint32_t actual)
{
    erts_aint32_t aprio, qbit, max_qbit;

    aprio = (actual >> ERTS_PSFLGS_ACT_PRIO_OFFSET) & ERTS_PSFLGS_PRIO_MASK;
    qbit = 1 << aprio;

    *prq_prio_p = aprio;

#ifdef ERTS_DIRTY_SCHEDULERS
    if (actual & (ERTS_PSFLG_DIRTY_CPU_PROC|ERTS_PSFLG_DIRTY_IO_PROC)) {
        /*
         * If we have system tasks of a priority higher
         * or equal to the user priority, we enqueue
         * on ordinary run-queue and take care of
         * those system tasks first.
         */
        if (actual & ERTS_PSFLG_ACTIVE_SYS) {
            erts_aint32_t uprio, stprio, qmask;
            uprio = (actual >> ERTS_PSFLGS_USR_PRIO_OFFSET) & ERTS_PSFLGS_PRIO_MASK;
            if (aprio < uprio)
                goto enqueue_normal_runq; /* system tasks with higher prio */
            erts_smp_proc_lock(c_p, ERTS_PROC_LOCK_STATUS);
            qmask = c_p->sys_task_qs->qmask;
            erts_smp_proc_unlock(c_p, ERTS_PROC_LOCK_STATUS);
            switch (qmask & -qmask) {
            case MAX_BIT:
                stprio = PRIORITY_MAX;
                break;
            case HIGH_BIT:
                stprio = PRIORITY_HIGH;
                break;
            case NORMAL_BIT:
                stprio = PRIORITY_NORMAL;
                break;
            case LOW_BIT:
                stprio = PRIORITY_LOW;
                break;
            default:
                stprio = PRIORITY_LOW+1;
                break;
            }
            if (stprio <= uprio)
                goto enqueue_normal_runq; /* system tasks with higher prio */
        }

        /* Enqueue in dirty run queue if not already enqueued */
	if (actual & (ERTS_PSFLG_DIRTY_CPU_PROC_IN_Q|ERTS_PSFLG_DIRTY_IO_PROC_IN_Q))
	    return ERTS_ENQUEUE_NOT; /* already in queue */
        if (actual & ERTS_PSFLG_DIRTY_CPU_PROC) {
	    *newp |= ERTS_PSFLG_DIRTY_CPU_PROC_IN_Q;
            if (actual & ERTS_PSFLG_IN_RUNQ)
                return -ERTS_ENQUEUE_DIRTY_CPU_QUEUE; /* use proxy */
	    *newp |= ERTS_PSFLG_IN_RUNQ;
            return ERTS_ENQUEUE_DIRTY_CPU_QUEUE;
        }
	*newp |= ERTS_PSFLG_DIRTY_IO_PROC_IN_Q;
        if (actual & ERTS_PSFLG_IN_RUNQ)
            return -ERTS_ENQUEUE_DIRTY_IO_QUEUE; /* use proxy */
	*newp |= ERTS_PSFLG_IN_RUNQ;
        return ERTS_ENQUEUE_DIRTY_IO_QUEUE;
    }

 enqueue_normal_runq:
#endif
    max_qbit = (actual >> ERTS_PSFLGS_IN_PRQ_MASK_OFFSET) & ERTS_PSFLGS_QMASK;
    max_qbit |= 1 << ERTS_PSFLGS_QMASK_BITS;
    max_qbit &= -max_qbit;
    /*
     * max_qbit now either contain bit set for highest prio queue or a bit
     * out of range (which will have a value larger than valid range).
     */

    if (qbit >= max_qbit)
	return ERTS_ENQUEUE_NOT; /* Already queued in higher or equal prio */

    /* Need to enqueue (if already enqueued, it is in lower prio) */
    *newp |= qbit << ERTS_PSFLGS_IN_PRQ_MASK_OFFSET;

    if ((actual & (ERTS_PSFLG_IN_RUNQ|ERTS_PSFLGS_USR_PRIO_MASK))
	!= (aprio << ERTS_PSFLGS_USR_PRIO_OFFSET)) {
	/*
	 * Process struct already enqueued, or actual prio not
	 * equal to user prio, i.e., enqueue using proxy.
	 */
	return -ERTS_ENQUEUE_NORMAL_QUEUE;
    }

    /*
     * Enqueue using process struct.
     */
    *newp &= ~ERTS_PSFLGS_PRQ_PRIO_MASK;
    *newp |= ERTS_PSFLG_IN_RUNQ | (aprio << ERTS_PSFLGS_PRQ_PRIO_OFFSET);
    return ERTS_ENQUEUE_NORMAL_QUEUE;
}

/*
 * schedule_out_process() return with c_rq locked.
 */
static ERTS_INLINE int
schedule_out_process(ErtsRunQueue *c_rq, erts_aint32_t state, Process *p, Process *proxy)
{
    erts_aint32_t a, e, n, enq_prio = -1;
    int enqueue; /* < 0 -> use proxy */
    Process* sched_p;
    ErtsRunQueue* runq;
#ifdef ERTS_SMP
    int check_emigration_need;
#endif

    a = state;

    while (1) {
	n = e = a;

	ASSERT(a & (ERTS_PSFLG_RUNNING|ERTS_PSFLG_RUNNING_SYS));

	enqueue = ERTS_ENQUEUE_NOT;

	n &= ~(ERTS_PSFLG_RUNNING|ERTS_PSFLG_RUNNING_SYS);
	if (a & ERTS_PSFLG_ACTIVE_SYS
	    || (a & (ERTS_PSFLG_ACTIVE|ERTS_PSFLG_SUSPENDED)) == ERTS_PSFLG_ACTIVE) {
	    enqueue = check_enqueue_in_prio_queue(p, &enq_prio, &n, a);
	}
	a = erts_smp_atomic32_cmpxchg_mb(&p->state, n, e);
	if (a == e)
	    break;
    }

    switch (enqueue) {
    case ERTS_ENQUEUE_NOT:
	if (erts_system_profile_flags.runnable_procs) {

	    if (!(a & ERTS_PSFLG_ACTIVE_SYS)
		&& (!(a & ERTS_PSFLG_ACTIVE)
		    || (a & ERTS_PSFLG_SUSPENDED))) {
		/* Process inactive */
		profile_runnable_proc(p, am_inactive);
	    }
	}

	if (proxy)
	    free_proxy_proc(proxy);

	erts_smp_runq_lock(c_rq);
	return 0;

#ifdef ERTS_DIRTY_SCHEDULERS
#ifdef ERTS_SMP
    case ERTS_ENQUEUE_DIRTY_CPU_QUEUE:
    case -ERTS_ENQUEUE_DIRTY_CPU_QUEUE:
	runq = ERTS_DIRTY_CPU_RUNQ;
	ASSERT(ERTS_SCHEDULER_IS_DIRTY_CPU(runq->scheduler));
#ifdef ERTS_SMP
	check_emigration_need = 0;
#endif
	break;

    case ERTS_ENQUEUE_DIRTY_IO_QUEUE:
    case -ERTS_ENQUEUE_DIRTY_IO_QUEUE:
	runq = ERTS_DIRTY_IO_RUNQ;
	ASSERT(ERTS_SCHEDULER_IS_DIRTY_IO(runq->scheduler));
#ifdef ERTS_SMP
	check_emigration_need = 0;
#endif
	break;
#endif
#endif

    default:
	ASSERT(enqueue == ERTS_ENQUEUE_NORMAL_QUEUE
	       || enqueue == -ERTS_ENQUEUE_NORMAL_QUEUE);

	runq = erts_get_runq_proc(p);
#ifdef ERTS_SMP
	check_emigration_need = !(ERTS_PSFLG_BOUND & n);
#endif
	break;
    }

    ASSERT(!(n & ERTS_PSFLG_SUSPENDED) || (n & ERTS_PSFLG_ACTIVE_SYS));

    if (enqueue < 0)
	sched_p = make_proxy_proc(proxy, p, enq_prio);
    else {
	sched_p = p;
	if (proxy)
	    free_proxy_proc(proxy);
    }

#ifdef ERTS_SMP
    if (check_emigration_need) {
	ErtsRunQueue *new_runq = erts_check_emigration_need(runq, enq_prio);
	if (new_runq) {
	    RUNQ_SET_RQ(&sched_p->run_queue, new_runq);
	    runq = new_runq;
	}
    }
#endif

    ASSERT(runq);

    erts_smp_runq_lock(runq);

    /* Enqueue the process */
    enqueue_process(runq, (int) enq_prio, sched_p);

    if (runq == c_rq)
	return 1;
    erts_smp_runq_unlock(runq);
    smp_notify_inc_runq(runq);
    erts_smp_runq_lock(c_rq);
    return 1;
}

static ERTS_INLINE void
add2runq(Process *p, erts_aint32_t state, erts_aint32_t prio)
{
    ErtsRunQueue *runq = erts_get_runq_proc(p);

#ifdef ERTS_SMP
    if (!(ERTS_PSFLG_BOUND & state)) {
	ErtsRunQueue *new_runq = erts_check_emigration_need(runq, (int) prio);
	if (new_runq) {
	    RUNQ_SET_RQ(&p->run_queue, new_runq);
	    runq = new_runq;
	}
    }
#endif
    ASSERT(runq);

    erts_smp_runq_lock(runq);

    /* Enqueue the process */
    enqueue_process(runq, (int) prio, p);

    erts_smp_runq_unlock(runq);
    smp_notify_inc_runq(runq);

}

static ERTS_INLINE int
change_proc_schedule_state(Process *p,
			   erts_aint32_t clear_state_flags,
			   erts_aint32_t set_state_flags,
			   erts_aint32_t *statep,
			   erts_aint32_t *enq_prio_p)
{
    /*
     * NOTE: ERTS_PSFLG_RUNNING, ERTS_PSFLG_RUNNING_SYS and
     *       ERTS_PSFLG_ACTIVE_SYS are not allowed to be
     *       altered by this function!
     */
    erts_aint32_t a = *statep, n;
    int enqueue; /* < 0 -> use proxy */

    ASSERT(!(a & ERTS_PSFLG_PROXY));
    ASSERT((clear_state_flags & (ERTS_PSFLG_RUNNING
				 | ERTS_PSFLG_RUNNING_SYS
				 | ERTS_PSFLG_ACTIVE_SYS)) == 0);
    ASSERT((set_state_flags & (ERTS_PSFLG_RUNNING
			       | ERTS_PSFLG_RUNNING_SYS
			       | ERTS_PSFLG_ACTIVE_SYS)) == 0);

    while (1) {
	erts_aint32_t e;
	n = e = a;

	enqueue = ERTS_ENQUEUE_NOT;

	if (a & ERTS_PSFLG_FREE)
	    break; /* We don't want to schedule free processes... */

	if (clear_state_flags)
	    n &= ~clear_state_flags;

	if (set_state_flags)
	    n |= set_state_flags;

	if ((n & (ERTS_PSFLG_SUSPENDED
		  | ERTS_PSFLG_RUNNING
		  | ERTS_PSFLG_RUNNING_SYS
		  | ERTS_PSFLG_IN_RUNQ
		  | ERTS_PSFLG_ACTIVE)) == ERTS_PSFLG_ACTIVE) {
	    /*
	     * Active and seemingly need to be enqueued, but
	     * process may be in a run queue via proxy, need
	     * further inspection...
	     */
	    enqueue = check_enqueue_in_prio_queue(p, enq_prio_p, &n, a);
	}

	a = erts_smp_atomic32_cmpxchg_mb(&p->state, n, e);
	if (a == e)
	    break;
	if (enqueue == ERTS_ENQUEUE_NOT && n == a)
	    break;
    }

    if (erts_system_profile_flags.runnable_procs) {

	if (((n & (ERTS_PSFLG_SUSPENDED
		   | ERTS_PSFLG_ACTIVE)) == ERTS_PSFLG_ACTIVE)
	    && (!(a & (ERTS_PSFLG_ACTIVE_SYS
		       | ERTS_PSFLG_RUNNING
		       | ERTS_PSFLG_RUNNING_SYS)
		  && (!(a & ERTS_PSFLG_ACTIVE)
		      || (a & ERTS_PSFLG_SUSPENDED))))) {
	    /* We activated a prevously inactive process */
	    profile_runnable_proc(p, am_active);
	}

    }

    *statep = a;

    return enqueue;
}

static ERTS_INLINE void
schedule_process(Process *p, erts_aint32_t in_state)
{
    erts_aint32_t enq_prio  = -1;
    erts_aint32_t state = in_state;
    int enqueue = change_proc_schedule_state(p,
					     0,
					     ERTS_PSFLG_ACTIVE,
					     &state,
					     &enq_prio);
    if (enqueue != ERTS_ENQUEUE_NOT)
	add2runq(enqueue > 0 ? p : make_proxy_proc(NULL, p, enq_prio),
		 state,
		 enq_prio);
}

void
erts_schedule_process(Process *p, erts_aint32_t state)
{
    schedule_process(p, state);
}

static void
schedule_process_sys_task(Process *p, erts_aint32_t state, Process *proxy)
{
    erts_aint32_t a = state, n, enq_prio = -1;
    int enqueue; /* < 0 -> use proxy */

    ASSERT(!(state & ERTS_PSFLG_PROXY));

    while (1) {
	erts_aint32_t e;
	n = e = a;

	if (a & ERTS_PSFLG_FREE)
	    return; /* We don't want to schedule free processes... */

	enqueue = ERTS_ENQUEUE_NOT;
	n |= ERTS_PSFLG_ACTIVE_SYS;
	if (!(a & (ERTS_PSFLG_RUNNING|ERTS_PSFLG_RUNNING_SYS)))
	    enqueue = check_enqueue_in_prio_queue(p, &enq_prio, &n, a);
	a = erts_smp_atomic32_cmpxchg_mb(&p->state, n, e);
	if (a == e)
	    break;
	if (a == n && enqueue == ERTS_ENQUEUE_NOT)
	    goto cleanup;
    }

    if (erts_system_profile_flags.runnable_procs) {

	if (!(a & (ERTS_PSFLG_ACTIVE_SYS
		   | ERTS_PSFLG_RUNNING
		   | ERTS_PSFLG_RUNNING_SYS))
	    && (!(a & ERTS_PSFLG_ACTIVE) || (a & ERTS_PSFLG_SUSPENDED))) {
	    /* We activated a prevously inactive process */
	    profile_runnable_proc(p, am_active);
	}

    }

    if (enqueue != ERTS_ENQUEUE_NOT) {
	Process *sched_p;
	if (enqueue > 0)
	    sched_p = p;
	else {
	    sched_p = make_proxy_proc(proxy, p, enq_prio);
	    proxy = NULL;
	}
	add2runq(sched_p, n, enq_prio);
    }

cleanup:
    if (proxy)
	free_proxy_proc(proxy);
}

static ERTS_INLINE int
suspend_process(Process *c_p, Process *p)
{
    erts_aint32_t state;
    int suspended = 0;
    ERTS_SMP_LC_ASSERT(ERTS_PROC_LOCK_STATUS & erts_proc_lc_my_proc_locks(p));

    state = erts_smp_atomic32_read_acqb(&p->state);

    if ((state & ERTS_PSFLG_SUSPENDED))
	suspended = -1;
    else {
	if (c_p == p) {
	    state = erts_smp_atomic32_read_bor_relb(&p->state,
						    ERTS_PSFLG_SUSPENDED);
	    ASSERT(state & (ERTS_PSFLG_RUNNING|ERTS_PSFLG_RUNNING_SYS));
	    suspended = (state & ERTS_PSFLG_SUSPENDED) ? -1: 1;
	}
	else {
	    while (!(state & (ERTS_PSFLG_RUNNING|ERTS_PSFLG_EXITING))) {
		erts_aint32_t n, e;

		n = e = state;
		n |= ERTS_PSFLG_SUSPENDED;
		state = erts_smp_atomic32_cmpxchg_relb(&p->state, n, e);
		if (state == e) {
		    suspended = 1;
		    break;
		}
		if (state & ERTS_PSFLG_SUSPENDED) {
		    suspended = -1;
		    break;
		}
	    }
	}
    }

    if (suspended) {

	ASSERT(!(ERTS_PSFLG_RUNNING & state)
	       || p == erts_get_current_process());

	if (suspended > 0 && erts_system_profile_flags.runnable_procs) {

	    /* 'state' is before our change... */

	    if ((state & (ERTS_PSFLG_ACTIVE
			  | ERTS_PSFLG_ACTIVE_SYS
			  | ERTS_PSFLG_RUNNING
			  | ERTS_PSFLG_RUNNING_SYS
			  | ERTS_PSFLG_SUSPENDED)) == ERTS_PSFLG_ACTIVE) {
		/* We made process inactive */
		profile_runnable_proc(p, am_inactive);
	    }

	}

	p->rcount++;  /* count number of suspend */
    }

    return suspended;
}

static ERTS_INLINE void
resume_process(Process *p)
{
    erts_aint32_t state, enq_prio = -1;
    int enqueue;

    ERTS_SMP_LC_ASSERT(ERTS_PROC_LOCK_STATUS & erts_proc_lc_my_proc_locks(p));

    ASSERT(p->rcount > 0);

    if (--p->rcount > 0)  /* multiple suspend */
	return;

    state = erts_smp_atomic32_read_nob(&p->state);
    enqueue = change_proc_schedule_state(p,
					 ERTS_PSFLG_SUSPENDED,
					 0,
					 &state,
					 &enq_prio);
    if (enqueue)
	add2runq(enqueue > 0 ? p : make_proxy_proc(NULL, p, enq_prio),
		 state,
		 enq_prio);
}

int
erts_get_max_no_executing_schedulers(void)
{
#ifdef ERTS_SMP
    if (erts_smp_atomic32_read_nob(&schdlr_sspnd.changing))
	return (int) erts_no_schedulers;
    ERTS_THR_MEMORY_BARRIER;
    return (int) erts_smp_atomic32_read_nob(&schdlr_sspnd.active);
#else
    return 1;
#endif
}

#ifdef ERTS_SMP

static void
scheduler_ix_resume_wake(Uint ix)
{
    ErtsSchedulerSleepInfo *ssi = ERTS_SCHED_SLEEP_INFO_IX(ix);
    scheduler_ssi_resume_wake(ssi);
}

static void
scheduler_ssi_resume_wake(ErtsSchedulerSleepInfo *ssi)
{
    erts_aint32_t xflgs = (ERTS_SSI_FLG_SLEEPING
			   | ERTS_SSI_FLG_TSE_SLEEPING
			   | ERTS_SSI_FLG_WAITING
			   | ERTS_SSI_FLG_SUSPENDED);
    erts_aint32_t oflgs;
    do {
	oflgs = erts_smp_atomic32_cmpxchg_relb(&ssi->flags, 0, xflgs);
	if (oflgs == xflgs) {
	    erts_sched_finish_poke(ssi, oflgs);
	    break;
	}
	xflgs = oflgs;
    } while (oflgs & ERTS_SSI_FLG_SUSPENDED);
}

static erts_aint32_t
sched_prep_spin_suspended(ErtsSchedulerSleepInfo *ssi, erts_aint32_t xpct)
{
    erts_aint32_t oflgs;
    erts_aint32_t nflgs = (ERTS_SSI_FLG_SLEEPING
			   | ERTS_SSI_FLG_WAITING
			   | ERTS_SSI_FLG_SUSPENDED);
    erts_aint32_t xflgs = xpct;

    do {
	oflgs = erts_smp_atomic32_cmpxchg_acqb(&ssi->flags, nflgs, xflgs);
	if (oflgs == xflgs)
	    return nflgs;
	xflgs = oflgs;
    } while (oflgs & ERTS_SSI_FLG_SUSPENDED);

    return oflgs;
}

static erts_aint32_t
sched_spin_suspended(ErtsSchedulerSleepInfo *ssi, int spincount)
{
    int until_yield = ERTS_SCHED_SPIN_UNTIL_YIELD;
    int sc = spincount;
    erts_aint32_t flgs;

    do {
	flgs = erts_smp_atomic32_read_acqb(&ssi->flags);
	if ((flgs & (ERTS_SSI_FLG_SLEEPING
		     | ERTS_SSI_FLG_WAITING
		     | ERTS_SSI_FLG_SUSPENDED))
	    != (ERTS_SSI_FLG_SLEEPING
		| ERTS_SSI_FLG_WAITING
		| ERTS_SSI_FLG_SUSPENDED)) {
	    break;
	}
	ERTS_SPIN_BODY;
	if (--until_yield == 0) {
	    until_yield = ERTS_SCHED_SPIN_UNTIL_YIELD;
	    erts_thr_yield();
	}
    } while (--sc > 0);
    return flgs;
}

static erts_aint32_t
sched_set_suspended_sleeptype(ErtsSchedulerSleepInfo *ssi)
{
    erts_aint32_t oflgs;
    erts_aint32_t nflgs = (ERTS_SSI_FLG_SLEEPING
			   | ERTS_SSI_FLG_TSE_SLEEPING
			   | ERTS_SSI_FLG_WAITING
			   | ERTS_SSI_FLG_SUSPENDED);
    erts_aint32_t xflgs = (ERTS_SSI_FLG_SLEEPING
			   | ERTS_SSI_FLG_WAITING
			   | ERTS_SSI_FLG_SUSPENDED);

    erts_tse_reset(ssi->event);

    while (1) {
	oflgs = erts_smp_atomic32_cmpxchg_acqb(&ssi->flags, nflgs, xflgs);
	if (oflgs == xflgs)
	    return nflgs;
	if ((oflgs & (ERTS_SSI_FLG_SLEEPING
		      | ERTS_SSI_FLG_WAITING
		      | ERTS_SSI_FLG_SUSPENDED))
	    != (ERTS_SSI_FLG_SLEEPING
		| ERTS_SSI_FLG_WAITING
		| ERTS_SSI_FLG_SUSPENDED)) {
	    return oflgs;
	}
	xflgs = oflgs;
    }
}

#ifdef ERTS_DIRTY_SCHEDULERS

static void
suspend_scheduler(ErtsSchedulerData *esdp)
{
    erts_aint32_t flgs;
    erts_aint32_t changing;
#ifdef ERTS_DIRTY_SCHEDULERS
    long no = (long) (ERTS_SCHEDULER_IS_DIRTY(esdp)
		      ? ERTS_DIRTY_SCHEDULER_NO(esdp)
		      : esdp->no);
#else
    long no = (long) esdp->no;
#endif
    ErtsSchedulerSleepInfo *ssi = esdp->ssi;
    long active_schedulers;
    int curr_online = 1;
    int wake = 0;
    erts_aint32_t aux_work;
    int thr_prgr_active = 1;
    ErtsStuckBoundProcesses sbp = {NULL, NULL};
    int* ss_onlinep;
    int* ss_curr_onlinep;
    int* ss_wait_curr_onlinep;
    long* ss_wait_activep;
    long ss_wait_active_target;
    erts_smp_atomic32_t* ss_changingp;
    erts_smp_atomic32_t* ss_activep;

    /*
     * Schedulers may be suspended in two different ways:
     * - A scheduler may be suspended since it is not online.
     *   All schedulers with scheduler ids greater than
     *   schdlr_sspnd.online are suspended; same for dirty
     *   schedulers and schdlr_sspnd.dirty_cpu_online and
     *   schdlr_sspnd.dirty_io_online.
     * - Multi scheduling is blocked. All schedulers except the
     *   scheduler with scheduler id 1 are suspended, and all
     *   dirty CPU and dirty I/O schedulers are suspended.
     *
     * Regardless of why a scheduler is suspended, it ends up here.
     */

    ASSERT(ERTS_SCHEDULER_IS_DIRTY(esdp) || no != 1);

#ifdef ERTS_DIRTY_SCHEDULERS
    if (ERTS_SCHEDULER_IS_DIRTY(esdp)) {
	if (erts_smp_mtx_trylock(&schdlr_sspnd.mtx) == EBUSY) {
	    erts_smp_runq_unlock(esdp->run_queue);
	    erts_smp_mtx_lock(&schdlr_sspnd.mtx);
	    erts_smp_runq_lock(esdp->run_queue);
	}
	if (ongoing_multi_scheduling_block())
	    evacuate_run_queue(esdp->run_queue, &sbp);
    } else
#endif
	evacuate_run_queue(esdp->run_queue, &sbp);

    erts_smp_runq_unlock(esdp->run_queue);

#ifdef ERTS_DIRTY_SCHEDULERS
    if (!ERTS_SCHEDULER_IS_DIRTY(esdp))
#endif
    {
	erts_sched_check_cpu_bind_prep_suspend(esdp);

	if (erts_system_profile_flags.scheduler)
	    profile_scheduler(make_small(esdp->no), am_inactive);

	sched_wall_time_change(esdp, 0);

	erts_smp_mtx_lock(&schdlr_sspnd.mtx);
    }

    flgs = sched_prep_spin_suspended(ssi, ERTS_SSI_FLG_SUSPENDED);
    if (flgs & ERTS_SSI_FLG_SUSPENDED) {

#ifdef ERTS_DIRTY_SCHEDULERS
	if (ERTS_SCHEDULER_IS_DIRTY(esdp)) {
	    if (ERTS_RUNQ_IS_DIRTY_CPU_RUNQ(esdp->run_queue)) {
		active_schedulers = erts_smp_atomic32_dec_read_nob(&schdlr_sspnd.dirty_cpu_active);
		ASSERT(active_schedulers >= 0);
		changing = erts_smp_atomic32_read_nob(&schdlr_sspnd.dirty_cpu_changing);
		ss_onlinep = &schdlr_sspnd.dirty_cpu_online;
		ss_curr_onlinep = &schdlr_sspnd.dirty_cpu_curr_online;
		ss_wait_curr_onlinep = &schdlr_sspnd.dirty_cpu_wait_curr_online;
		ss_changingp = &schdlr_sspnd.dirty_cpu_changing;
		ss_wait_activep = &schdlr_sspnd.msb.dirty_cpu_wait_active;
		ss_activep = &schdlr_sspnd.dirty_cpu_active;
	    } else {
		active_schedulers = erts_smp_atomic32_dec_read_nob(&schdlr_sspnd.dirty_io_active);
		ASSERT(active_schedulers >= 0);
		changing = erts_smp_atomic32_read_nob(&schdlr_sspnd.dirty_io_changing);
		ss_onlinep = &schdlr_sspnd.dirty_io_online;
		ss_curr_onlinep = &schdlr_sspnd.dirty_io_curr_online;
		ss_wait_curr_onlinep = &schdlr_sspnd.dirty_io_wait_curr_online;
		ss_changingp = &schdlr_sspnd.dirty_io_changing;
		ss_wait_activep = &schdlr_sspnd.msb.dirty_io_wait_active;
		ss_activep = &schdlr_sspnd.dirty_io_active;
	    }
	    ss_wait_active_target = 0;
	}
	else
#endif
	{
	    active_schedulers = erts_smp_atomic32_dec_read_nob(&schdlr_sspnd.active);
	    ASSERT(active_schedulers >= 1);
	    changing = erts_smp_atomic32_read_nob(&schdlr_sspnd.changing);
	    ss_onlinep = &schdlr_sspnd.online;
	    ss_curr_onlinep = &schdlr_sspnd.curr_online;
	    ss_wait_curr_onlinep = &schdlr_sspnd.wait_curr_online;
	    ss_changingp = &schdlr_sspnd.changing;
	    ss_wait_activep = &schdlr_sspnd.msb.wait_active;
	    ss_activep = &schdlr_sspnd.active;
	    ss_wait_active_target = 1;
	}
	if (changing & ERTS_SCHDLR_SSPND_CHNG_MSB) {
	    if (active_schedulers == *ss_wait_activep)
		wake = 1;
	    if (active_schedulers == ss_wait_active_target) {
		changing = erts_smp_atomic32_read_band_nob(ss_changingp,
							   ~ERTS_SCHDLR_SSPND_CHNG_MSB);
		changing &= ~ERTS_SCHDLR_SSPND_CHNG_MSB;
	    }
	}

	while (1) {
	    if (changing & ERTS_SCHDLR_SSPND_CHNG_ONLN) {
		int changed = 0;
		if (no > *ss_onlinep && curr_online) {
		    (*ss_curr_onlinep)--;
		    curr_online = 0;
		    changed = 1;
		}
		else if (no <= *ss_onlinep && !curr_online) {
		    (*ss_curr_onlinep)++;
		    curr_online = 1;
		    changed = 1;
		}
		if (changed
		    && *ss_curr_onlinep == *ss_wait_curr_onlinep)
		    wake = 1;
		if (*ss_onlinep == *ss_curr_onlinep) {
		    changing = erts_smp_atomic32_read_band_nob(ss_changingp,
							       ~ERTS_SCHDLR_SSPND_CHNG_ONLN);
		    changing &= ~ERTS_SCHDLR_SSPND_CHNG_ONLN;
		}
	    }

	    if (wake) {
		erts_smp_cnd_signal(&schdlr_sspnd.cnd);
		wake = 0;
	    }

	    if (curr_online && !ongoing_multi_scheduling_block()) {
		flgs = erts_smp_atomic32_read_acqb(&ssi->flags);
		if (!(flgs & ERTS_SSI_FLG_SUSPENDED))
		    break;
	    }
	    erts_smp_mtx_unlock(&schdlr_sspnd.mtx);

	    while (1) {
		erts_aint32_t qmask;
		erts_aint32_t flgs;

		qmask = (ERTS_RUNQ_FLGS_GET(esdp->run_queue)
			 & ERTS_RUNQ_FLGS_QMASK);
		aux_work = erts_atomic32_read_acqb(&ssi->aux_work);
		if (aux_work|qmask) {
		    if (!ERTS_SCHEDULER_IS_DIRTY(esdp) && !thr_prgr_active) {
			erts_thr_progress_active(esdp, thr_prgr_active = 1);
			sched_wall_time_change(esdp, 1);
		    }
		    if (aux_work)
			aux_work = handle_aux_work(&esdp->aux_work_data,
						   aux_work,
						   1);

		    if (!ERTS_SCHEDULER_IS_DIRTY(esdp) &&
			(aux_work && erts_thr_progress_update(esdp)))
			erts_thr_progress_leader_update(esdp);
		    if (qmask) {
#ifdef ERTS_DIRTY_SCHEDULERS
			if (ERTS_SCHEDULER_IS_DIRTY(esdp)) {
			    erts_smp_mtx_lock(&schdlr_sspnd.mtx);
			    erts_smp_runq_lock(esdp->run_queue);
			    if (ongoing_multi_scheduling_block())
				evacuate_run_queue(esdp->run_queue, &sbp);
			    erts_smp_runq_unlock(esdp->run_queue);
			    erts_smp_mtx_unlock(&schdlr_sspnd.mtx);
			} else
#endif
			{
			    erts_smp_runq_lock(esdp->run_queue);
			    evacuate_run_queue(esdp->run_queue, &sbp);
			    erts_smp_runq_unlock(esdp->run_queue);
			}
		    }
		}

		if (!aux_work) {
#ifdef ERTS_DIRTY_SCHEDULERS
		    if (!ERTS_SCHEDULER_IS_DIRTY(esdp))
#endif
		    {
			if (thr_prgr_active) {
			    erts_thr_progress_active(esdp, thr_prgr_active = 0);
			    sched_wall_time_change(esdp, 0);
			}
			erts_thr_progress_prepare_wait(esdp);
		    }
		    flgs = sched_spin_suspended(ssi,
						ERTS_SCHED_SUSPEND_SLEEP_SPINCOUNT);
		    if (flgs == (ERTS_SSI_FLG_SLEEPING
				 | ERTS_SSI_FLG_WAITING
				 | ERTS_SSI_FLG_SUSPENDED)) {
			flgs = sched_set_suspended_sleeptype(ssi);
			if (flgs == (ERTS_SSI_FLG_SLEEPING
				     | ERTS_SSI_FLG_TSE_SLEEPING
				     | ERTS_SSI_FLG_WAITING
				     | ERTS_SSI_FLG_SUSPENDED)) {
			    int res;

			    do {
				res = erts_tse_wait(ssi->event);
			    } while (res == EINTR);
			}
		    }
#ifdef ERTS_DIRTY_SCHEDULERS
		    if (!ERTS_SCHEDULER_IS_DIRTY(esdp))
#endif
			erts_thr_progress_finalize_wait(esdp);
		}

		flgs = sched_prep_spin_suspended(ssi, (ERTS_SSI_FLG_WAITING
						       | ERTS_SSI_FLG_SUSPENDED));
		if (!(flgs & ERTS_SSI_FLG_SUSPENDED))
		    break;
		changing = erts_smp_atomic32_read_nob(ss_changingp);
		if (changing & ~ERTS_SCHDLR_SSPND_CHNG_WAITER)
		    break;
	    }

	    erts_smp_mtx_lock(&schdlr_sspnd.mtx);
	    changing = erts_smp_atomic32_read_nob(ss_changingp);
	}

	active_schedulers = erts_smp_atomic32_inc_read_nob(ss_activep);
	changing = erts_smp_atomic32_read_nob(ss_changingp);
	if ((changing & ERTS_SCHDLR_SSPND_CHNG_MSB)
	    && *ss_onlinep == active_schedulers) {
	    erts_smp_atomic32_read_band_nob(ss_changingp,
					    ~ERTS_SCHDLR_SSPND_CHNG_MSB);
	}

#ifdef ERTS_DIRTY_SCHEDULERS
	if (!ERTS_SCHEDULER_IS_DIRTY(esdp))
#endif
	    ASSERT(no <= *ss_onlinep);
	ASSERT(!ongoing_multi_scheduling_block());

    }

    erts_smp_mtx_unlock(&schdlr_sspnd.mtx);

    ASSERT(curr_online);

#ifdef ERTS_DIRTY_SCHEDULERS
    if (!ERTS_SCHEDULER_IS_DIRTY(esdp))
#endif
    {
	if (erts_system_profile_flags.scheduler)
	    profile_scheduler(make_small(esdp->no), am_active);

	if (!thr_prgr_active) {
	    erts_thr_progress_active(esdp, thr_prgr_active = 1);
	    sched_wall_time_change(esdp, 1);
	}
    }

    erts_smp_runq_lock(esdp->run_queue);
    non_empty_runq(esdp->run_queue);

#ifdef ERTS_DIRTY_SCHEDULERS
    if (!ERTS_SCHEDULER_IS_DIRTY(esdp))
#endif
    {
	schedule_bound_processes(esdp->run_queue, &sbp);

	erts_sched_check_cpu_bind_post_suspend(esdp);
    }
}

#else /* !ERTS_DIRTY_SCHEDULERS */

static void
suspend_scheduler(ErtsSchedulerData *esdp)
{
    erts_aint32_t flgs;
    erts_aint32_t changing;
    long no = (long) esdp->no;
    ErtsSchedulerSleepInfo *ssi = esdp->ssi;
    long active_schedulers;
    int curr_online = 1;
    int wake = 0;
    erts_aint32_t aux_work;
    int thr_prgr_active = 1;
    ErtsStuckBoundProcesses sbp = {NULL, NULL};

    /*
     * Schedulers may be suspended in two different ways:
     * - A scheduler may be suspended since it is not online.
     *   All schedulers with scheduler ids greater than
     *   schdlr_sspnd.online are suspended.
     * - Multi scheduling is blocked. All schedulers except the
     *   scheduler with scheduler id 1 are suspended.
     *
     * Regardless of why a scheduler is suspended, it ends up here.
     */

    ASSERT(no != 1);

    evacuate_run_queue(esdp->run_queue, &sbp);

    erts_smp_runq_unlock(esdp->run_queue);

    erts_sched_check_cpu_bind_prep_suspend(esdp);

    if (erts_system_profile_flags.scheduler)
    	profile_scheduler(make_small(esdp->no), am_inactive);

    sched_wall_time_change(esdp, 0);

    erts_smp_mtx_lock(&schdlr_sspnd.mtx);

    flgs = sched_prep_spin_suspended(ssi, ERTS_SSI_FLG_SUSPENDED);
    if (flgs & ERTS_SSI_FLG_SUSPENDED) {

	active_schedulers = erts_smp_atomic32_dec_read_nob(&schdlr_sspnd.active);
	ASSERT(active_schedulers >= 1);
	changing = erts_smp_atomic32_read_nob(&schdlr_sspnd.changing);
	if (changing & ERTS_SCHDLR_SSPND_CHNG_MSB) {
	    if (active_schedulers == schdlr_sspnd.msb.wait_active)
		wake = 1;
	    if (active_schedulers == 1) {
		changing = erts_smp_atomic32_read_band_nob(&schdlr_sspnd.changing,
							   ~ERTS_SCHDLR_SSPND_CHNG_MSB);
		changing &= ~ERTS_SCHDLR_SSPND_CHNG_MSB;
	    }
	}

	while (1) {
	    if (changing & ERTS_SCHDLR_SSPND_CHNG_ONLN) {
		int changed = 0;
		if (no > schdlr_sspnd.online && curr_online) {
		    schdlr_sspnd.curr_online--;
		    curr_online = 0;
		    changed = 1;
		}
		else if (no <= schdlr_sspnd.online && !curr_online) {
		    schdlr_sspnd.curr_online++;
		    curr_online = 1;
		    changed = 1;
		}
		if (changed
		    && schdlr_sspnd.curr_online == schdlr_sspnd.wait_curr_online)
		    wake = 1;
		if (schdlr_sspnd.online == schdlr_sspnd.curr_online) {
		    changing = erts_smp_atomic32_read_band_nob(&schdlr_sspnd.changing,
							       ~ERTS_SCHDLR_SSPND_CHNG_ONLN);
		    changing &= ~ERTS_SCHDLR_SSPND_CHNG_ONLN;
		}
	    }

	    if (wake) {
		erts_smp_cnd_signal(&schdlr_sspnd.cnd);
		wake = 0;
	    }

	    if (curr_online && !ongoing_multi_scheduling_block()) {
		flgs = erts_smp_atomic32_read_acqb(&ssi->flags);
		if (!(flgs & ERTS_SSI_FLG_SUSPENDED))
		    break;
	    }
	    erts_smp_mtx_unlock(&schdlr_sspnd.mtx);

	    while (1) {
		erts_aint32_t qmask;
		erts_aint32_t flgs;

		qmask = (ERTS_RUNQ_FLGS_GET(esdp->run_queue)
			 & ERTS_RUNQ_FLGS_QMASK);
		aux_work = erts_atomic32_read_acqb(&ssi->aux_work);
		if (aux_work|qmask) {
		    if (!thr_prgr_active) {
			erts_thr_progress_active(esdp, thr_prgr_active = 1);
			sched_wall_time_change(esdp, 1);
		    }
		    if (aux_work)
			aux_work = handle_aux_work(&esdp->aux_work_data,
						   aux_work,
						   1);
		    if (aux_work && erts_thr_progress_update(esdp))
			erts_thr_progress_leader_update(esdp);
		    if (qmask) {
			erts_smp_runq_lock(esdp->run_queue);
			evacuate_run_queue(esdp->run_queue, &sbp);
			erts_smp_runq_unlock(esdp->run_queue);
		    }
		}

		if (!aux_work) {
		    if (thr_prgr_active) {
			erts_thr_progress_active(esdp, thr_prgr_active = 0);
			sched_wall_time_change(esdp, 0);
		    }
		    erts_thr_progress_prepare_wait(esdp);
		    flgs = sched_spin_suspended(ssi,
						ERTS_SCHED_SUSPEND_SLEEP_SPINCOUNT);
		    if (flgs == (ERTS_SSI_FLG_SLEEPING
				 | ERTS_SSI_FLG_WAITING
				 | ERTS_SSI_FLG_SUSPENDED)) {
			flgs = sched_set_suspended_sleeptype(ssi);
			if (flgs == (ERTS_SSI_FLG_SLEEPING
				     | ERTS_SSI_FLG_TSE_SLEEPING
				     | ERTS_SSI_FLG_WAITING
				     | ERTS_SSI_FLG_SUSPENDED)) {
			    int res;

			    do {
				res = erts_tse_wait(ssi->event);
			    } while (res == EINTR);
			}
		    }
		    erts_thr_progress_finalize_wait(esdp);
		}

		flgs = sched_prep_spin_suspended(ssi, (ERTS_SSI_FLG_WAITING
						       | ERTS_SSI_FLG_SUSPENDED));
		if (!(flgs & ERTS_SSI_FLG_SUSPENDED))
		    break;
		changing = erts_smp_atomic32_read_nob(&schdlr_sspnd.changing);
		if (changing & ~ERTS_SCHDLR_SSPND_CHNG_WAITER)
		    break;
	    }

	    erts_smp_mtx_lock(&schdlr_sspnd.mtx);
	    changing = erts_smp_atomic32_read_nob(&schdlr_sspnd.changing);
	}

	active_schedulers = erts_smp_atomic32_inc_read_nob(&schdlr_sspnd.active);
	changing = erts_smp_atomic32_read_nob(&schdlr_sspnd.changing);
	if ((changing & ERTS_SCHDLR_SSPND_CHNG_MSB)
	    && schdlr_sspnd.online == active_schedulers) {
	    erts_smp_atomic32_read_band_nob(&schdlr_sspnd.changing,
					    ~ERTS_SCHDLR_SSPND_CHNG_MSB);
	}

	ASSERT(no <= schdlr_sspnd.online);
	ASSERT(!ongoing_multi_scheduling_block());

    }

    erts_smp_mtx_unlock(&schdlr_sspnd.mtx);

    ASSERT(curr_online);

    if (erts_system_profile_flags.scheduler)
    	profile_scheduler(make_small(esdp->no), am_active);

    if (!thr_prgr_active) {
	erts_thr_progress_active(esdp, thr_prgr_active = 1);
	sched_wall_time_change(esdp, 1);
    }

    erts_smp_runq_lock(esdp->run_queue);
    non_empty_runq(esdp->run_queue);

    schedule_bound_processes(esdp->run_queue, &sbp);

    erts_sched_check_cpu_bind_post_suspend(esdp);
}

#endif

ErtsSchedSuspendResult
erts_schedulers_state(Uint *total,
		      Uint *online,
		      Uint *active,
		      Uint *dirty_cpu,
		      Uint *dirty_cpu_online,
		      Uint *dirty_io,
		      int yield_allowed)
{
    int res = ERTS_SCHDLR_SSPND_EINVAL;
    erts_aint32_t changing;
    erts_smp_mtx_lock(&schdlr_sspnd.mtx);
    changing = erts_smp_atomic32_read_nob(&schdlr_sspnd.changing);
#ifdef ERTS_DIRTY_SCHEDULERS
    changing |= (erts_smp_atomic32_read_nob(&schdlr_sspnd.dirty_cpu_changing)
		 | erts_smp_atomic32_read_nob(&schdlr_sspnd.dirty_io_changing));
#endif
    if (yield_allowed && (changing & ~ERTS_SCHDLR_SSPND_CHNG_WAITER))
	res = ERTS_SCHDLR_SSPND_YIELD_RESTART;
    else {
	if (active)
	    *active = schdlr_sspnd.online;
	if (online)
	    *online = schdlr_sspnd.online;
	if (ongoing_multi_scheduling_block() && active)
	    *active = 1;
#ifdef ERTS_DIRTY_SCHEDULERS
	if (dirty_cpu_online)
	    *dirty_cpu_online = schdlr_sspnd.dirty_cpu_online;
#endif
	res = ERTS_SCHDLR_SSPND_DONE;
    }
    erts_smp_mtx_unlock(&schdlr_sspnd.mtx);
    if (total)
	*total = erts_no_schedulers;
#ifdef ERTS_DIRTY_SCHEDULERS
    if (dirty_cpu)
	*dirty_cpu = erts_no_dirty_cpu_schedulers;
    if (dirty_io)
	*dirty_io = erts_no_dirty_io_schedulers;
#endif
    return res;
}

#ifdef ERTS_DIRTY_SCHEDULERS

ErtsSchedSuspendResult
erts_set_schedulers_online(Process *p,
			   ErtsProcLocks plocks,
			   Sint new_no,
			   Sint *old_no
#ifdef ERTS_DIRTY_SCHEDULERS
			   , int dirty_only
#endif
			   )
{
    ErtsSchedulerData *esdp;
    int ix, res = -1, no, have_unlocked_plocks, end_wait;
    erts_aint32_t changing = 0;
#ifdef ERTS_DIRTY_SCHEDULERS
    ErtsSchedulerSleepInfo* ssi;
    int dirty_no, change_dirty;
#endif

    if (new_no < 1)
	return ERTS_SCHDLR_SSPND_EINVAL;
#ifdef ERTS_DIRTY_SCHEDULERS
    else if (dirty_only && erts_no_dirty_cpu_schedulers < new_no)
	return ERTS_SCHDLR_SSPND_EINVAL;
#endif
    else if (erts_no_schedulers < new_no)
	return ERTS_SCHDLR_SSPND_EINVAL;

    esdp = ERTS_PROC_GET_SCHDATA(p);
    end_wait = 0;

    erts_smp_mtx_lock(&schdlr_sspnd.mtx);

    have_unlocked_plocks = 0;
    no = (int) new_no;

#ifdef ERTS_DIRTY_SCHEDULERS
    ASSERT(schdlr_sspnd.dirty_cpu_online <= erts_no_dirty_cpu_schedulers);
    if (dirty_only) {
	if (no > schdlr_sspnd.online) {
	    erts_smp_mtx_unlock(&schdlr_sspnd.mtx);
	    return ERTS_SCHDLR_SSPND_EINVAL;
	}
	dirty_no = no;
    } else {
	/*
	 * Adjust the number of dirty CPU schedulers online relative to the
	 * adjustment made to the number of normal schedulers online.
	 */
	int total_pct = erts_no_dirty_cpu_schedulers*100/erts_no_schedulers;
	int onln_pct = no*total_pct/schdlr_sspnd.online;
	dirty_no = schdlr_sspnd.dirty_cpu_online*onln_pct/100;
	if (dirty_no == 0)
	    dirty_no = 1;
	ASSERT(dirty_no <= erts_no_dirty_cpu_schedulers);
    }
#endif
    changing = erts_smp_atomic32_read_nob(&schdlr_sspnd.changing);
#ifdef ERTS_DIRTY_SCHEDULERS
    changing |= erts_smp_atomic32_read_nob(&schdlr_sspnd.dirty_cpu_changing);
#endif
    if (changing) {
	res = ERTS_SCHDLR_SSPND_YIELD_RESTART;
    }
    else {
	int online = *old_no = schdlr_sspnd.online;
#ifdef ERTS_DIRTY_SCHEDULERS
	int dirty_online = schdlr_sspnd.dirty_cpu_online;

	if (dirty_only) {
	    *old_no = schdlr_sspnd.dirty_cpu_online;
	    if (dirty_no == schdlr_sspnd.dirty_cpu_online) {
		res = ERTS_SCHDLR_SSPND_DONE;
	    }
	    change_dirty = 1;
	} else {
#endif
	    if (no == schdlr_sspnd.online) {
#ifdef ERTS_DIRTY_SCHEDULERS
		dirty_only = 1;
		if (dirty_no == schdlr_sspnd.dirty_cpu_online)
#endif
		    res = ERTS_SCHDLR_SSPND_DONE;
#ifdef ERTS_DIRTY_SCHEDULERS
		else
		    change_dirty = 1;
#endif
	    }
#ifdef ERTS_DIRTY_SCHEDULERS
	    else
		change_dirty = (dirty_no != schdlr_sspnd.dirty_cpu_online);
	}
#endif
	if (res == -1)
	{
	    int increase = (no > online);
#ifdef ERTS_DIRTY_SCHEDULERS
	    if (!dirty_only) {
#endif
		ERTS_SCHDLR_SSPND_CHNG_SET((ERTS_SCHDLR_SSPND_CHNG_ONLN
					    | ERTS_SCHDLR_SSPND_CHNG_WAITER), 0);
		schdlr_sspnd.online = no;
#ifdef ERTS_DIRTY_SCHEDULERS
	    } else
		increase = (dirty_no > dirty_online);
	    if (change_dirty) {
		ERTS_SCHDLR_SSPND_DIRTY_CPU_CHNG_SET((ERTS_SCHDLR_SSPND_CHNG_ONLN
						      | ERTS_SCHDLR_SSPND_CHNG_WAITER), 0);
		schdlr_sspnd.dirty_cpu_online = dirty_no;
	    }
#endif
	    if (increase) {
		int ix;
#ifdef ERTS_DIRTY_SCHEDULERS
		if (!dirty_only) {
#endif
		    schdlr_sspnd.wait_curr_online = no;
		    if (ongoing_multi_scheduling_block()) {
			for (ix = online; ix < no; ix++)
			    erts_sched_poke(ERTS_SCHED_SLEEP_INFO_IX(ix));
		    }
		    else {
			if (plocks) {
			    have_unlocked_plocks = 1;
			    erts_smp_proc_unlock(p, plocks);
			}
			change_no_used_runqs(no);

			for (ix = online; ix < no; ix++)
			    resume_run_queue(ERTS_RUNQ_IX(ix));

			for (ix = no; ix < erts_no_run_queues; ix++)
			    suspend_run_queue(ERTS_RUNQ_IX(ix));
		    }
#ifdef ERTS_DIRTY_SCHEDULERS
		}
		if (change_dirty) {
		    schdlr_sspnd.dirty_cpu_wait_curr_online = dirty_no;
		    ASSERT(schdlr_sspnd.dirty_cpu_curr_online !=
			   schdlr_sspnd.dirty_cpu_wait_curr_online);
		    if (ongoing_multi_scheduling_block()) {
			for (ix = dirty_online; ix < dirty_no; ix++) {
			    ssi = ERTS_DIRTY_CPU_SCHED_SLEEP_INFO_IX(ix);
			    erts_sched_poke(ssi);
			}
		    } else {
			for (ix = dirty_online; ix < dirty_no; ix++) {
			    ssi = ERTS_DIRTY_CPU_SCHED_SLEEP_INFO_IX(ix);
			    scheduler_ssi_resume_wake(ssi);
			    erts_smp_atomic32_read_band_nob(&ssi->flags,
							    ~ERTS_SSI_FLG_SUSPENDED);
			}
			wake_dirty_schedulers(ERTS_DIRTY_CPU_RUNQ, 0);
		    }
		}
#endif
		res = ERTS_SCHDLR_SSPND_DONE;
	    }
	    else /* if (no < online) */ {
#ifdef ERTS_DIRTY_SCHEDULERS
		if (change_dirty) {
		    schdlr_sspnd.dirty_cpu_wait_curr_online = dirty_no;
		    ASSERT(schdlr_sspnd.dirty_cpu_curr_online !=
			   schdlr_sspnd.dirty_cpu_wait_curr_online);
		    if (ongoing_multi_scheduling_block()) {
			for (ix = dirty_no; ix < dirty_online; ix++) {
			    ssi = ERTS_DIRTY_CPU_SCHED_SLEEP_INFO_IX(ix);
			    erts_sched_poke(ssi);
			}
		    } else {
			for (ix = dirty_no; ix < dirty_online; ix++) {
			    ssi = ERTS_DIRTY_CPU_SCHED_SLEEP_INFO_IX(ix);
			    erts_smp_atomic32_read_bor_nob(&ssi->flags,
							   ERTS_SSI_FLG_SUSPENDED);
			}
			wake_dirty_schedulers(ERTS_DIRTY_CPU_RUNQ, 0);
		    }
		}
		if (dirty_only) {
		    res = ERTS_SCHDLR_SSPND_DONE;
		}
		else
#endif
		{
		    if (p->scheduler_data->no <= no) {
			res = ERTS_SCHDLR_SSPND_DONE;
			schdlr_sspnd.wait_curr_online = no;
		    }
		    else {
			/*
			 * Yield! Current process needs to migrate
			 * before bif returns.
			 */
			res = ERTS_SCHDLR_SSPND_YIELD_DONE;
			schdlr_sspnd.wait_curr_online = no+1;
		    }

		    if (ongoing_multi_scheduling_block()) {
			for (ix = no; ix < online; ix++)
			    erts_sched_poke(ERTS_SCHED_SLEEP_INFO_IX(ix));
		    }
		    else {
			if (plocks) {
			    have_unlocked_plocks = 1;
			    erts_smp_proc_unlock(p, plocks);
			}

			change_no_used_runqs(no);
			for (ix = no; ix < erts_no_run_queues; ix++)
			    suspend_run_queue(ERTS_RUNQ_IX(ix));

			for (ix = no; ix < online; ix++) {
			    ErtsRunQueue *rq = ERTS_RUNQ_IX(ix);
			    wake_scheduler(rq);
			}
		    }
		}
	    }

#ifdef ERTS_DIRTY_SCHEDULERS
	    if (change_dirty) {
		while (schdlr_sspnd.dirty_cpu_curr_online != schdlr_sspnd.dirty_cpu_wait_curr_online)
		    erts_smp_cnd_wait(&schdlr_sspnd.cnd, &schdlr_sspnd.mtx);
		ERTS_SCHDLR_SSPND_DIRTY_CPU_CHNG_SET(0, ERTS_SCHDLR_SSPND_CHNG_WAITER);
		erts_smp_atomic32_read_band_nob(&schdlr_sspnd.dirty_cpu_changing,
						~ERTS_SCHDLR_SSPND_CHNG_WAITER);
	    }
	    if (!dirty_only)
#endif
	    {
		if (schdlr_sspnd.curr_online != schdlr_sspnd.wait_curr_online) {
		    erts_smp_mtx_unlock(&schdlr_sspnd.mtx);
		    if (plocks && !have_unlocked_plocks) {
			have_unlocked_plocks = 1;
			erts_smp_proc_unlock(p, plocks);
		    }
		    erts_thr_progress_active(esdp, 0);
		    erts_thr_progress_prepare_wait(esdp);
		    end_wait = 1;
		    erts_smp_mtx_lock(&schdlr_sspnd.mtx);
		}

		while (schdlr_sspnd.curr_online != schdlr_sspnd.wait_curr_online)
		    erts_smp_cnd_wait(&schdlr_sspnd.cnd, &schdlr_sspnd.mtx);

		ASSERT(res != ERTS_SCHDLR_SSPND_DONE
		       ? (ERTS_SCHDLR_SSPND_CHNG_WAITER
			  & erts_smp_atomic32_read_nob(&schdlr_sspnd.changing))
		       : (ERTS_SCHDLR_SSPND_CHNG_WAITER
			  == erts_smp_atomic32_read_nob(&schdlr_sspnd.changing)));
		erts_smp_atomic32_read_band_nob(&schdlr_sspnd.changing,
						~ERTS_SCHDLR_SSPND_CHNG_WAITER);
	    }
	}
    }

    erts_smp_mtx_unlock(&schdlr_sspnd.mtx);
#ifdef ERTS_DIRTY_SCHEDULERS
    ASSERT(schdlr_sspnd.dirty_cpu_online <= schdlr_sspnd.online);
    if (!dirty_only)
#endif
    {
	if (end_wait) {
	    erts_thr_progress_finalize_wait(esdp);
	    erts_thr_progress_active(esdp, 1);
	}
	if (have_unlocked_plocks)
	    erts_smp_proc_lock(p, plocks);
    }

    return res;
}

#else /* !ERTS_DIRTY_SCHEDULERS */

ErtsSchedSuspendResult
erts_set_schedulers_online(Process *p,
			   ErtsProcLocks plocks,
			   Sint new_no,
			   Sint *old_no)
{
    ErtsSchedulerData *esdp;
    int ix, res, no, have_unlocked_plocks, end_wait;
    erts_aint32_t changing;

    if (new_no < 1 || erts_no_schedulers < new_no)
	return ERTS_SCHDLR_SSPND_EINVAL;

    esdp = ERTS_PROC_GET_SCHDATA(p);
    end_wait = 0;

    erts_smp_mtx_lock(&schdlr_sspnd.mtx);

    have_unlocked_plocks = 0;
    no = (int) new_no;

    changing = erts_smp_atomic32_read_nob(&schdlr_sspnd.changing);
    if (changing) {
	res = ERTS_SCHDLR_SSPND_YIELD_RESTART;
    }
    else {
	int online = *old_no = schdlr_sspnd.online;
	if (no == schdlr_sspnd.online) {
	    res = ERTS_SCHDLR_SSPND_DONE;
	}
	else {
	    ERTS_SCHDLR_SSPND_CHNG_SET((ERTS_SCHDLR_SSPND_CHNG_ONLN
					| ERTS_SCHDLR_SSPND_CHNG_WAITER), 0);
	    schdlr_sspnd.online = no;
	    if (no > online) {
		int ix;
		schdlr_sspnd.wait_curr_online = no;
		if (ongoing_multi_scheduling_block()) {
		    for (ix = online; ix < no; ix++)
			erts_sched_poke(ERTS_SCHED_SLEEP_INFO_IX(ix));
		}
		else {
		    if (plocks) {
			have_unlocked_plocks = 1;
			erts_smp_proc_unlock(p, plocks);
		    }
		    change_no_used_runqs(no);

		    for (ix = online; ix < no; ix++)
			resume_run_queue(ERTS_RUNQ_IX(ix));

		    for (ix = no; ix < erts_no_run_queues; ix++)
			suspend_run_queue(ERTS_RUNQ_IX(ix));
		}
		res = ERTS_SCHDLR_SSPND_DONE;
	    }
	    else /* if (no < online) */ {
		if (p->scheduler_data->no <= no) {
		    res = ERTS_SCHDLR_SSPND_DONE;
		    schdlr_sspnd.wait_curr_online = no;
		}
		else {
		    /*
		     * Yield! Current process needs to migrate
		     * before bif returns.
		     */
		    res = ERTS_SCHDLR_SSPND_YIELD_DONE;
		    schdlr_sspnd.wait_curr_online = no+1;
		}

		if (ongoing_multi_scheduling_block()) {
		    for (ix = no; ix < online; ix++)
			erts_sched_poke(ERTS_SCHED_SLEEP_INFO_IX(ix));
		}
		else {
		    if (plocks) {
			have_unlocked_plocks = 1;
			erts_smp_proc_unlock(p, plocks);
		    }

		    change_no_used_runqs(no);
		    for (ix = no; ix < erts_no_run_queues; ix++)
			suspend_run_queue(ERTS_RUNQ_IX(ix));

		    for (ix = no; ix < online; ix++) {
			ErtsRunQueue *rq = ERTS_RUNQ_IX(ix);
			wake_scheduler(rq);
		    }
		}
	    }

	    if (schdlr_sspnd.curr_online != schdlr_sspnd.wait_curr_online) {
		erts_smp_mtx_unlock(&schdlr_sspnd.mtx);
		if (plocks && !have_unlocked_plocks) {
		    have_unlocked_plocks = 1;
		    erts_smp_proc_unlock(p, plocks);
		}
		erts_thr_progress_active(esdp, 0);
		erts_thr_progress_prepare_wait(esdp);
		end_wait = 1;
		erts_smp_mtx_lock(&schdlr_sspnd.mtx);
	    }

	    while (schdlr_sspnd.curr_online != schdlr_sspnd.wait_curr_online)
		erts_smp_cnd_wait(&schdlr_sspnd.cnd, &schdlr_sspnd.mtx);

	    ASSERT(res != ERTS_SCHDLR_SSPND_DONE
		   ? (ERTS_SCHDLR_SSPND_CHNG_WAITER
		      & erts_smp_atomic32_read_nob(&schdlr_sspnd.changing))
		   : (ERTS_SCHDLR_SSPND_CHNG_WAITER
		      == erts_smp_atomic32_read_nob(&schdlr_sspnd.changing)));
	    erts_smp_atomic32_read_band_nob(&schdlr_sspnd.changing,
					    ~ERTS_SCHDLR_SSPND_CHNG_WAITER);

	}
    }

    erts_smp_mtx_unlock(&schdlr_sspnd.mtx);
    if (end_wait) {
	erts_thr_progress_finalize_wait(esdp);
	erts_thr_progress_active(esdp, 1);
    }
    if (have_unlocked_plocks)
	erts_smp_proc_lock(p, plocks);

    return res;
}

#endif

ErtsSchedSuspendResult
erts_block_multi_scheduling(Process *p, ErtsProcLocks plocks, int on, int all)
{
    int ix, res, have_unlocked_plocks = 0, online;
    erts_aint32_t changing;
    ErtsProcList *plp;
#ifdef ERTS_DIRTY_SCHEDULERS
    ErtsSchedulerSleepInfo* ssi;
#endif

    erts_smp_mtx_lock(&schdlr_sspnd.mtx);
    changing = erts_smp_atomic32_read_nob(&schdlr_sspnd.changing);
#ifdef ERTS_DIRTY_SCHEDULERS
    changing |= (erts_smp_atomic32_read_nob(&schdlr_sspnd.dirty_cpu_changing)
		 | erts_smp_atomic32_read_nob(&schdlr_sspnd.dirty_io_changing));
#endif
    if (changing) {
	res = ERTS_SCHDLR_SSPND_YIELD_RESTART; /* Yield */
    }
    else if (on) { /* ------ BLOCK ------ */
	if (schdlr_sspnd.msb.procs) {
	    plp = proclist_create(p);
	    erts_proclist_store_last(&schdlr_sspnd.msb.procs, plp);
	    p->flags |= F_HAVE_BLCKD_MSCHED;
	    ASSERT(erts_smp_atomic32_read_nob(&schdlr_sspnd.active) == 1);
#ifdef ERTS_DIRTY_SCHEDULERS
	    ASSERT(erts_smp_atomic32_read_nob(&schdlr_sspnd.dirty_cpu_active) == 0);
	    ASSERT(erts_smp_atomic32_read_nob(&schdlr_sspnd.dirty_io_active) == 0);
#endif
	    ASSERT(p->scheduler_data->no == 1);
	    res = ERTS_SCHDLR_SSPND_DONE_MSCHED_BLOCKED;
	} else {
	    int online = schdlr_sspnd.online;
	    p->flags |= F_HAVE_BLCKD_MSCHED;
	    if (plocks) {
		have_unlocked_plocks = 1;
		erts_smp_proc_unlock(p, plocks);
	    }
	    ASSERT(!ongoing_multi_scheduling_block());
	    schdlr_sspnd.msb.ongoing = 1;
	    if (online == 1) {
		res = ERTS_SCHDLR_SSPND_DONE_MSCHED_BLOCKED;
		ASSERT(erts_smp_atomic32_read_nob(&schdlr_sspnd.active) == 1);
#ifdef ERTS_DIRTY_SCHEDULERS
		ASSERT(erts_smp_atomic32_read_nob(&schdlr_sspnd.dirty_cpu_active) == 1);
		ASSERT(!(erts_smp_atomic32_read_nob(&ERTS_DIRTY_CPU_SCHED_SLEEP_INFO_IX(0)->flags)
			 & ERTS_SSI_FLG_SUSPENDED));
		schdlr_sspnd.msb.dirty_cpu_wait_active = 0;
		ERTS_SCHDLR_SSPND_DIRTY_CPU_CHNG_SET((ERTS_SCHDLR_SSPND_CHNG_MSB
						      | ERTS_SCHDLR_SSPND_CHNG_WAITER), 0);
		ssi = ERTS_DIRTY_CPU_SCHED_SLEEP_INFO_IX(0);
		erts_smp_atomic32_read_bor_nob(&ssi->flags, ERTS_SSI_FLG_SUSPENDED);
		wake_dirty_schedulers(ERTS_DIRTY_CPU_RUNQ, 0);
		while (erts_smp_atomic32_read_nob(&schdlr_sspnd.dirty_cpu_active)
		       != schdlr_sspnd.msb.dirty_cpu_wait_active)
		    erts_smp_cnd_wait(&schdlr_sspnd.cnd, &schdlr_sspnd.mtx);
		ERTS_SCHDLR_SSPND_DIRTY_CPU_CHNG_SET(0, ERTS_SCHDLR_SSPND_CHNG_WAITER);

		schdlr_sspnd.msb.dirty_io_wait_active = 0;
		ERTS_SCHDLR_SSPND_DIRTY_IO_CHNG_SET((ERTS_SCHDLR_SSPND_CHNG_MSB
						     | ERTS_SCHDLR_SSPND_CHNG_WAITER), 0);
		for (ix = 0; ix < erts_no_dirty_io_schedulers; ix++) {
		    ssi = ERTS_DIRTY_IO_SCHED_SLEEP_INFO_IX(ix);
		    erts_smp_atomic32_read_bor_nob(&ssi->flags,
						   ERTS_SSI_FLG_SUSPENDED);
		}
		wake_dirty_schedulers(ERTS_DIRTY_IO_RUNQ, 0);
		while (erts_smp_atomic32_read_nob(&schdlr_sspnd.dirty_io_active)
		       != schdlr_sspnd.msb.dirty_io_wait_active)
		    erts_smp_cnd_wait(&schdlr_sspnd.cnd, &schdlr_sspnd.mtx);
		ERTS_SCHDLR_SSPND_DIRTY_IO_CHNG_SET(0, ERTS_SCHDLR_SSPND_CHNG_WAITER);
#endif
		ASSERT(p->scheduler_data->no == 1);
	    }
	    else {
		ERTS_SCHDLR_SSPND_CHNG_SET((ERTS_SCHDLR_SSPND_CHNG_MSB
					    | ERTS_SCHDLR_SSPND_CHNG_WAITER), 0);
		if (p->scheduler_data->no == 1) {
		    res = ERTS_SCHDLR_SSPND_DONE_MSCHED_BLOCKED;
		    schdlr_sspnd.msb.wait_active = 1;
		}
		else {
		    /*
		     * Yield! Current process needs to migrate
		     * before bif returns.
		     */
		    res = ERTS_SCHDLR_SSPND_YIELD_DONE_MSCHED_BLOCKED;
		    schdlr_sspnd.msb.wait_active = 2;
		}

#ifdef ERTS_DIRTY_SCHEDULERS
		schdlr_sspnd.msb.dirty_cpu_wait_active = 0;
		ERTS_SCHDLR_SSPND_DIRTY_CPU_CHNG_SET((ERTS_SCHDLR_SSPND_CHNG_MSB
						      | ERTS_SCHDLR_SSPND_CHNG_WAITER), 0);
		for (ix = 0; ix < erts_no_dirty_cpu_schedulers; ix++) {
		    ssi = ERTS_DIRTY_CPU_SCHED_SLEEP_INFO_IX(ix);
		    erts_smp_atomic32_read_bor_nob(&ssi->flags,
						   ERTS_SSI_FLG_SUSPENDED);
		}
		wake_dirty_schedulers(ERTS_DIRTY_CPU_RUNQ, 0);
		while (erts_smp_atomic32_read_nob(&schdlr_sspnd.dirty_cpu_active)
		       != schdlr_sspnd.msb.dirty_cpu_wait_active)
		    erts_smp_cnd_wait(&schdlr_sspnd.cnd, &schdlr_sspnd.mtx);
		ERTS_SCHDLR_SSPND_DIRTY_CPU_CHNG_SET(0, ERTS_SCHDLR_SSPND_CHNG_WAITER);
		ASSERT(schdlr_sspnd.dirty_cpu_curr_online == schdlr_sspnd.dirty_cpu_online);

		schdlr_sspnd.msb.dirty_io_wait_active = 0;
		ERTS_SCHDLR_SSPND_DIRTY_IO_CHNG_SET((ERTS_SCHDLR_SSPND_CHNG_MSB
						     | ERTS_SCHDLR_SSPND_CHNG_WAITER), 0);
		for (ix = 0; ix < erts_no_dirty_io_schedulers; ix++) {
		    ssi = ERTS_DIRTY_IO_SCHED_SLEEP_INFO_IX(ix);
		    erts_smp_atomic32_read_bor_nob(&ssi->flags,
						   ERTS_SSI_FLG_SUSPENDED);
		}
		wake_dirty_schedulers(ERTS_DIRTY_IO_RUNQ, 0);
		while (erts_smp_atomic32_read_nob(&schdlr_sspnd.dirty_io_active)
		       != schdlr_sspnd.msb.dirty_io_wait_active)
		    erts_smp_cnd_wait(&schdlr_sspnd.cnd, &schdlr_sspnd.mtx);
		ERTS_SCHDLR_SSPND_DIRTY_IO_CHNG_SET(0, ERTS_SCHDLR_SSPND_CHNG_WAITER);
		ASSERT(schdlr_sspnd.dirty_io_curr_online == schdlr_sspnd.dirty_io_online);
#endif
		change_no_used_runqs(1);
		for (ix = 1; ix < erts_no_run_queues; ix++)
		    suspend_run_queue(ERTS_RUNQ_IX(ix));

		for (ix = 1; ix < online; ix++) {
		    ErtsRunQueue *rq = ERTS_RUNQ_IX(ix);
		    wake_scheduler(rq);
		}

		if (erts_smp_atomic32_read_nob(&schdlr_sspnd.active)
		    != schdlr_sspnd.msb.wait_active) {
		    ErtsSchedulerData *esdp;

		    erts_smp_mtx_unlock(&schdlr_sspnd.mtx);

		    if (plocks && !have_unlocked_plocks) {
			have_unlocked_plocks = 1;
			erts_smp_proc_unlock(p, plocks);
		    }

		    esdp = ERTS_PROC_GET_SCHDATA(p);

		    erts_thr_progress_active(esdp, 0);
		    erts_thr_progress_prepare_wait(esdp);

		    erts_smp_mtx_lock(&schdlr_sspnd.mtx);

		    while (erts_smp_atomic32_read_nob(&schdlr_sspnd.active)
			   != schdlr_sspnd.msb.wait_active)
			erts_smp_cnd_wait(&schdlr_sspnd.cnd,
					  &schdlr_sspnd.mtx);
		    
		    erts_smp_mtx_unlock(&schdlr_sspnd.mtx);
		    
		    erts_thr_progress_active(esdp, 1);
		    erts_thr_progress_finalize_wait(esdp);

		    erts_smp_mtx_lock(&schdlr_sspnd.mtx);

		}

		ASSERT(res != ERTS_SCHDLR_SSPND_DONE_MSCHED_BLOCKED
		       ? (ERTS_SCHDLR_SSPND_CHNG_WAITER
			  & erts_smp_atomic32_read_nob(&schdlr_sspnd.changing))
		       : (ERTS_SCHDLR_SSPND_CHNG_WAITER
			  == erts_smp_atomic32_read_nob(&schdlr_sspnd.changing)));
		erts_smp_atomic32_read_band_nob(&schdlr_sspnd.changing,
						~ERTS_SCHDLR_SSPND_CHNG_WAITER);
	    }
	    plp = proclist_create(p);
	    erts_proclist_store_last(&schdlr_sspnd.msb.procs, plp);
	    ASSERT(p->scheduler_data);
	}
    }
    else if (!ongoing_multi_scheduling_block()) {
	/* unblock not ongoing */
	ASSERT(!schdlr_sspnd.msb.procs);
	res = ERTS_SCHDLR_SSPND_DONE;
    }
    else {  /* ------ UNBLOCK ------ */
	if (p->flags & F_HAVE_BLCKD_MSCHED) {
	    ErtsProcList *plp = erts_proclist_peek_first(schdlr_sspnd.msb.procs);

	    while (plp) {
		ErtsProcList *tmp_plp = plp;
		plp = erts_proclist_peek_next(schdlr_sspnd.msb.procs, plp);
		if (erts_proclist_same(tmp_plp, p)) {
		    erts_proclist_remove(&schdlr_sspnd.msb.procs, tmp_plp);
		    proclist_destroy(tmp_plp);
		    if (!all)
			break;
		}
	    }
	}
	if (schdlr_sspnd.msb.procs)
	    res = ERTS_SCHDLR_SSPND_DONE_MSCHED_BLOCKED;
	else {
	    ERTS_SCHDLR_SSPND_CHNG_SET(ERTS_SCHDLR_SSPND_CHNG_MSB, 0);
	    p->flags &= ~F_HAVE_BLCKD_MSCHED;
	    schdlr_sspnd.msb.ongoing = 0;
	    if (schdlr_sspnd.online == 1) {
		/* No normal schedulers to resume */
		ASSERT(erts_smp_atomic32_read_nob(&schdlr_sspnd.active) == 1);
		ERTS_SCHDLR_SSPND_CHNG_SET(0, ERTS_SCHDLR_SSPND_CHNG_MSB);
	    }
	    else {
		online = schdlr_sspnd.online;
		if (plocks) {
		    have_unlocked_plocks = 1;
		    erts_smp_proc_unlock(p, plocks);
		}

		change_no_used_runqs(online);

		/* Resume all online run queues */
		for (ix = 1; ix < online; ix++)
		    resume_run_queue(ERTS_RUNQ_IX(ix));

		for (ix = online; ix < erts_no_run_queues; ix++)
		    suspend_run_queue(ERTS_RUNQ_IX(ix));
	    }
#ifdef ERTS_DIRTY_SCHEDULERS
	    ERTS_SCHDLR_SSPND_DIRTY_CPU_CHNG_SET(ERTS_SCHDLR_SSPND_CHNG_MSB, 0);
	    schdlr_sspnd.msb.dirty_cpu_wait_active = schdlr_sspnd.dirty_cpu_online;
	    for (ix = 0; ix < schdlr_sspnd.dirty_cpu_online; ix++) {
		ssi = ERTS_DIRTY_CPU_SCHED_SLEEP_INFO_IX(ix);
		scheduler_ssi_resume_wake(ssi);
		erts_smp_atomic32_read_band_nob(&ssi->flags,
						~ERTS_SSI_FLG_SUSPENDED);
	    }
	    wake_dirty_schedulers(ERTS_DIRTY_CPU_RUNQ, 0);

	    ERTS_SCHDLR_SSPND_DIRTY_IO_CHNG_SET(ERTS_SCHDLR_SSPND_CHNG_MSB, 0);
	    schdlr_sspnd.msb.dirty_io_wait_active = erts_no_dirty_io_schedulers;
	    for (ix = 0; ix < erts_no_dirty_io_schedulers; ix++) {
		ssi = ERTS_DIRTY_IO_SCHED_SLEEP_INFO_IX(ix);
		scheduler_ssi_resume_wake(ssi);
		erts_smp_atomic32_read_band_nob(&ssi->flags,
						~ERTS_SSI_FLG_SUSPENDED);
	    }
	    wake_dirty_schedulers(ERTS_DIRTY_IO_RUNQ, 0);
#endif
	    res = ERTS_SCHDLR_SSPND_DONE;
	}
    }

    erts_smp_mtx_unlock(&schdlr_sspnd.mtx);
    if (have_unlocked_plocks)
	erts_smp_proc_lock(p, plocks);
    return res;
}

#ifdef DEBUG
void
erts_dbg_multi_scheduling_return_trap(Process *p, Eterm return_value)
{
    if (return_value == am_blocked) {
	erts_aint32_t active = erts_smp_atomic32_read_nob(&schdlr_sspnd.active);
	ASSERT(1 <= active && active <= 2);
	ASSERT(ERTS_PROC_GET_SCHDATA(p)->no == 1);
    }
}
#endif

int
erts_is_multi_scheduling_blocked(void)
{
    int res;
    erts_smp_mtx_lock(&schdlr_sspnd.mtx);
    res = schdlr_sspnd.msb.procs != NULL;
    erts_smp_mtx_unlock(&schdlr_sspnd.mtx);
    return res;
}

Eterm
erts_multi_scheduling_blockers(Process *p)
{
    Eterm res = NIL;

    erts_smp_mtx_lock(&schdlr_sspnd.mtx);
    if (!erts_proclist_is_empty(schdlr_sspnd.msb.procs)) {
	Eterm *hp, *hp_end;
	ErtsProcList *plp1, *plp2;
	Uint max_size = 0;

	for (plp1 = erts_proclist_peek_first(schdlr_sspnd.msb.procs);
	     plp1;
	     plp1 = erts_proclist_peek_next(schdlr_sspnd.msb.procs, plp1)) {
	    max_size += 2;
	}
	ASSERT(max_size);
	hp = HAlloc(p, max_size);
	hp_end = hp + max_size;
	for (plp1 = erts_proclist_peek_first(schdlr_sspnd.msb.procs);
	     plp1;
	     plp1 = erts_proclist_peek_next(schdlr_sspnd.msb.procs, plp1)) {
	    for (plp2 = erts_proclist_peek_first(schdlr_sspnd.msb.procs);
		 plp2->pid != plp1->pid;
		 plp2 = erts_proclist_peek_next(schdlr_sspnd.msb.procs, plp2));
	    if (plp2 == plp1) {
		res = CONS(hp, plp1->pid, res);
		hp += 2;
	    }
	    /* else: already in result list */
	}
	HRelease(p, hp_end, hp);
    }
    erts_smp_mtx_unlock(&schdlr_sspnd.mtx);
    return res;
}

static void *
sched_thread_func(void *vesdp)
{
    ErtsThrPrgrCallbacks callbacks;
    ErtsSchedulerData *esdp = vesdp;
    Uint no = esdp->no;
#ifdef ERTS_SMP
    ERTS_SCHED_SLEEP_INFO_IX(no - 1)->event = erts_tse_fetch();
    callbacks.arg = (void *) esdp->ssi;
    callbacks.wakeup = thr_prgr_wakeup;
    callbacks.prepare_wait = thr_prgr_prep_wait;
    callbacks.wait = thr_prgr_wait;
    callbacks.finalize_wait = thr_prgr_fin_wait;

    erts_thr_progress_register_managed_thread(esdp, &callbacks, 0);
    erts_alloc_register_scheduler(vesdp);
#endif
#ifdef ERTS_ENABLE_LOCK_CHECK
    {
	char buf[31];
	erts_snprintf(&buf[0], 31, "scheduler %beu", no);
	erts_lc_set_thread_name(&buf[0]);
    }
#endif
    erts_tsd_set(sched_data_key, vesdp);
#ifdef ERTS_SMP
#if HAVE_ERTS_MSEG
    erts_mseg_late_init();
#endif
#if ERTS_USE_ASYNC_READY_Q
    esdp->aux_work_data.async_ready.queue = erts_get_async_ready_queue(no);
#endif

    erts_sched_init_check_cpu_bind(esdp);

    erts_proc_lock_prepare_proc_lock_waiter();
#endif

#ifdef HIPE
    hipe_thread_signal_init();
#endif
    erts_thread_init_float();

    if (no == 1) {
	erts_thr_progress_active(esdp, 0);
	erts_thr_progress_prepare_wait(esdp);
    }

    erts_smp_mtx_lock(&schdlr_sspnd.mtx);

    ASSERT(erts_smp_atomic32_read_nob(&schdlr_sspnd.changing)
	   & ERTS_SCHDLR_SSPND_CHNG_ONLN);

    if (--schdlr_sspnd.curr_online == schdlr_sspnd.wait_curr_online) {
	erts_smp_atomic32_read_band_nob(&schdlr_sspnd.changing,
					~ERTS_SCHDLR_SSPND_CHNG_ONLN);
	if (no != 1)
#ifdef ERTS_DIRTY_SCHEDULERS
	    erts_smp_cnd_broadcast(&schdlr_sspnd.cnd);
#else
	    erts_smp_cnd_signal(&schdlr_sspnd.cnd);
#endif
    }

    if (no == 1) {
	while (schdlr_sspnd.curr_online != schdlr_sspnd.wait_curr_online)
	    erts_smp_cnd_wait(&schdlr_sspnd.cnd, &schdlr_sspnd.mtx);
	ERTS_SCHDLR_SSPND_CHNG_SET(0, ERTS_SCHDLR_SSPND_CHNG_WAITER);
    }
    erts_smp_mtx_unlock(&schdlr_sspnd.mtx);

    if (no == 1) {
	erts_thr_progress_finalize_wait(esdp);
	erts_thr_progress_active(esdp, 1);
    }

#ifdef ERTS_DO_VERIFY_UNUSED_TEMP_ALLOC
    esdp->verify_unused_temp_alloc
	= erts_alloc_get_verify_unused_temp_alloc(
	    &esdp->verify_unused_temp_alloc_data);
    ERTS_VERIFY_UNUSED_TEMP_ALLOC(NULL);
#endif

    process_main();
    /* No schedulers should *ever* terminate */
    erl_exit(ERTS_ABORT_EXIT,
	     "Scheduler thread number %beu terminated\n",
	     no);
    return NULL;
}

#ifdef ERTS_DIRTY_SCHEDULERS
#ifdef ERTS_SMP
static void*
sched_dirty_cpu_thread_func(void *vesdp)
{
    ErtsThrPrgrCallbacks callbacks;
    ErtsSchedulerData *esdp = vesdp;
    Uint no = ERTS_DIRTY_SCHEDULER_NO(esdp);
    ERTS_DIRTY_SCHEDULER_TYPE(esdp) = ERTS_DIRTY_CPU_SCHEDULER;
    ASSERT(no != 0);
    ERTS_DIRTY_CPU_SCHED_SLEEP_INFO_IX(no-1)->event = erts_tse_fetch();
    callbacks.arg = (void *) esdp->ssi;
    callbacks.wakeup = thr_prgr_wakeup;
    callbacks.prepare_wait = NULL;
    callbacks.wait = NULL;
    callbacks.finalize_wait = NULL;

    erts_thr_progress_register_unmanaged_thread(&callbacks);
#ifdef ERTS_ENABLE_LOCK_CHECK
    {
	char buf[31];
	erts_snprintf(&buf[0], 31, "dirty cpu scheduler %beu", no);
	erts_lc_set_thread_name(&buf[0]);
    }
#endif
    erts_tsd_set(sched_data_key, vesdp);
#if ERTS_USE_ASYNC_READY_Q
    esdp->aux_work_data.async_ready.queue = NULL;
#endif

    erts_proc_lock_prepare_proc_lock_waiter();

#ifdef HIPE
    hipe_thread_signal_init();
#endif
    erts_thread_init_float();

    erts_smp_mtx_lock(&schdlr_sspnd.mtx);
    ASSERT(erts_smp_atomic32_read_nob(&schdlr_sspnd.dirty_cpu_changing)
	   & ERTS_SCHDLR_SSPND_CHNG_ONLN);

    if (--schdlr_sspnd.dirty_cpu_curr_online == schdlr_sspnd.dirty_cpu_wait_curr_online) {
	erts_smp_atomic32_read_band_nob(&schdlr_sspnd.dirty_cpu_changing,
					~ERTS_SCHDLR_SSPND_CHNG_ONLN);
	if (no != 1)
	    erts_smp_cnd_broadcast(&schdlr_sspnd.cnd);
    }

    if (no == 1) {
	while (schdlr_sspnd.dirty_cpu_curr_online != schdlr_sspnd.dirty_cpu_wait_curr_online)
	    erts_smp_cnd_wait(&schdlr_sspnd.cnd, &schdlr_sspnd.mtx);
	ERTS_SCHDLR_SSPND_DIRTY_CPU_CHNG_SET(0, ERTS_SCHDLR_SSPND_CHNG_WAITER);
    }
    erts_smp_mtx_unlock(&schdlr_sspnd.mtx);

    process_main();
    /* No schedulers should *ever* terminate */
    erl_exit(ERTS_ABORT_EXIT,
	     "Dirty CPU scheduler thread number %beu terminated\n",
	     no);
    return NULL;
}

static void*
sched_dirty_io_thread_func(void *vesdp)
{
    ErtsThrPrgrCallbacks callbacks;
    ErtsSchedulerData *esdp = vesdp;
    Uint no = ERTS_DIRTY_SCHEDULER_NO(esdp);
    ERTS_DIRTY_SCHEDULER_TYPE(esdp) = ERTS_DIRTY_IO_SCHEDULER;
    ASSERT(no != 0);
    ERTS_DIRTY_IO_SCHED_SLEEP_INFO_IX(no-1)->event = erts_tse_fetch();
    callbacks.arg = (void *) esdp->ssi;
    callbacks.wakeup = thr_prgr_wakeup;
    callbacks.prepare_wait = NULL;
    callbacks.wait = NULL;
    callbacks.finalize_wait = NULL;

    erts_thr_progress_register_unmanaged_thread(&callbacks);
#ifdef ERTS_ENABLE_LOCK_CHECK
    {
	char buf[31];
	erts_snprintf(&buf[0], 31, "dirty io scheduler %beu", no);
	erts_lc_set_thread_name(&buf[0]);
    }
#endif
    erts_tsd_set(sched_data_key, vesdp);
#if ERTS_USE_ASYNC_READY_Q
    esdp->aux_work_data.async_ready.queue = NULL;
#endif

    erts_proc_lock_prepare_proc_lock_waiter();

#ifdef HIPE
    hipe_thread_signal_init();
#endif
    erts_thread_init_float();

    erts_smp_mtx_lock(&schdlr_sspnd.mtx);
    ASSERT(erts_smp_atomic32_read_nob(&schdlr_sspnd.dirty_io_changing)
	   & ERTS_SCHDLR_SSPND_CHNG_ONLN);

    if (--schdlr_sspnd.dirty_io_curr_online == schdlr_sspnd.dirty_io_wait_curr_online) {
	erts_smp_atomic32_read_band_nob(&schdlr_sspnd.dirty_io_changing,
					~ERTS_SCHDLR_SSPND_CHNG_ONLN);
	if (no != 1)
	    erts_smp_cnd_broadcast(&schdlr_sspnd.cnd);
    }

    if (no == 1) {
	while (schdlr_sspnd.dirty_io_curr_online != schdlr_sspnd.dirty_io_wait_curr_online)
	    erts_smp_cnd_wait(&schdlr_sspnd.cnd, &schdlr_sspnd.mtx);
	ERTS_SCHDLR_SSPND_DIRTY_IO_CHNG_SET(0, ERTS_SCHDLR_SSPND_CHNG_WAITER);
    }
    erts_smp_mtx_unlock(&schdlr_sspnd.mtx);

    process_main();
    /* No schedulers should *ever* terminate */
    erl_exit(ERTS_ABORT_EXIT,
	     "Dirty I/O scheduler thread number %beu terminated\n",
	     no);
    return NULL;
}
#endif
#endif

static ethr_tid aux_tid;

void
erts_start_schedulers(void)
{
    int res = 0;
    Uint actual;
    Uint wanted = erts_no_schedulers;
    Uint wanted_no_schedulers = erts_no_schedulers;
    ethr_thr_opts opts = ETHR_THR_OPTS_DEFAULT_INITER;

    opts.detached = 1;

#ifdef ETHR_HAVE_THREAD_NAMES
    opts.name = malloc(80);
    if (!opts.name) {
        ERTS_INTERNAL_ERROR("malloc failed to allocate memory!");
    }
#endif

#ifdef ERTS_SMP
    if (erts_runq_supervision_interval) {
	opts.suggested_stack_size = 16;
#ifdef ETHR_HAVE_THREAD_NAMES
        sprintf(opts.name, "runq_supervisor");
#endif
	erts_atomic_init_nob(&runq_supervisor_sleeping, 0);
	if (0 != ethr_event_init(&runq_supervision_event))
	    erl_exit(1, "Failed to create run-queue supervision event\n");
	if (0 != ethr_thr_create(&runq_supervisor_tid,
				 runq_supervisor,
				 NULL,
				 &opts))
	    erl_exit(1, "Failed to create run-queue supervision thread\n");

    }
#endif

    opts.suggested_stack_size = erts_sched_thread_suggested_stack_size;

    if (wanted < 1)
	wanted = 1;
    if (wanted > ERTS_MAX_NO_OF_SCHEDULERS) {
	wanted = ERTS_MAX_NO_OF_SCHEDULERS;
	res = ENOTSUP;
    }

    for (actual = 0; actual < wanted; actual++) {
	ErtsSchedulerData *esdp = ERTS_SCHEDULER_IX(actual);

	ASSERT(actual == esdp->no - 1);

#ifdef ETHR_HAVE_THREAD_NAMES
	sprintf(opts.name, "scheduler_%d", actual + 1);
#endif

#ifdef __OSE__
        /* This should be done in the bind strategy */
	opts.coreNo = (actual+1) % ose_num_cpus();
#endif

	res = ethr_thr_create(&esdp->tid, sched_thread_func, (void*)esdp, &opts);

	if (res != 0) {
           break;
	}
    }

    erts_no_schedulers = actual;

#ifdef ERTS_DIRTY_SCHEDULERS
#ifdef ERTS_SMP
    {
	int ix;
	for (ix = 0; ix < erts_no_dirty_cpu_schedulers; ix++) {
	    ErtsSchedulerData *esdp = ERTS_DIRTY_CPU_SCHEDULER_IX(ix);
#ifdef ETHR_HAVE_THREAD_NAMES
            sprintf(opts.name,"dirty_cpu_scheduler_%d", ix + 1);
#endif
	    res = ethr_thr_create(&esdp->tid,sched_dirty_cpu_thread_func,(void*)esdp,&opts);
	    if (res != 0)
		erl_exit(1, "Failed to create dirty cpu scheduler thread %d\n", ix);
	}
	for (ix = 0; ix < erts_no_dirty_io_schedulers; ix++) {
	    ErtsSchedulerData *esdp = ERTS_DIRTY_IO_SCHEDULER_IX(ix);
#ifdef ETHR_HAVE_THREAD_NAMES
            sprintf(opts.name,"dirty_io_scheduler_%d", ix + 1);
#endif
	    res = ethr_thr_create(&esdp->tid,sched_dirty_io_thread_func,(void*)esdp,&opts);
	    if (res != 0)
		erl_exit(1, "Failed to create dirty io scheduler thread %d\n", ix);
	}
    }
#endif
#endif

    ERTS_THR_MEMORY_BARRIER;

#ifdef ETHR_HAVE_THREAD_NAMES
    sprintf(opts.name, "aux");
#endif

#ifdef __OSE__
    opts.coreNo = 0;
#endif /* __OSE__ */

    res = ethr_thr_create(&aux_tid, aux_thread, NULL, &opts);
    if (res != 0)
	erl_exit(1, "Failed to create aux thread\n");

    if (actual < 1)
	erl_exit(1,
		 "Failed to create any scheduler-threads: %s (%d)\n",
		 erl_errno_id(res),
		 res);
    if (res != 0) {
	erts_dsprintf_buf_t *dsbufp = erts_create_logger_dsbuf();
	ASSERT(actual != wanted_no_schedulers);
	erts_dsprintf(dsbufp,
		      "Failed to create %beu scheduler-threads (%s:%d); "
		      "only %beu scheduler-thread%s created.\n",
		      wanted_no_schedulers, erl_errno_id(res), res,
		      actual, actual == 1 ? " was" : "s were");
	erts_send_error_to_logger_nogl(dsbufp);
    }

#ifdef ETHR_HAVE_THREAD_NAMES
    free(opts.name);
#endif
}

#endif /* ERTS_SMP */

#ifdef ERTS_SMP

static void
add_pend_suspend(Process *suspendee,
		 Eterm originator_pid,
		 void (*handle_func)(Process *,
				     ErtsProcLocks,
				     int,
				     Eterm))
{
    ErtsPendingSuspend *psp = erts_alloc(ERTS_ALC_T_PEND_SUSPEND,
					 sizeof(ErtsPendingSuspend));
    psp->next = NULL;
#ifdef DEBUG
#if defined(ARCH_64) && !HALFWORD_HEAP
    psp->end = (ErtsPendingSuspend *) 0xdeaddeaddeaddead;
#else
    psp->end = (ErtsPendingSuspend *) 0xdeaddead;
#endif
#endif
    psp->pid = originator_pid;
    psp->handle_func = handle_func;

    if (suspendee->pending_suspenders)
	suspendee->pending_suspenders->end->next = psp;
    else
	suspendee->pending_suspenders = psp;
    suspendee->pending_suspenders->end = psp;
}

static void
handle_pending_suspend(Process *p, ErtsProcLocks p_locks)
{
    ErtsPendingSuspend *psp;
    int is_alive = !ERTS_PROC_IS_EXITING(p);

    ERTS_SMP_LC_ASSERT(p_locks & ERTS_PROC_LOCK_STATUS);

    /*
     * New pending suspenders might appear while we are processing
     * (since we may release the status lock on p while processing).
     */
    while (p->pending_suspenders) {
	psp = p->pending_suspenders;
	p->pending_suspenders = NULL;
	while (psp) {
	    ErtsPendingSuspend *free_psp;
	    (*psp->handle_func)(p, p_locks, is_alive, psp->pid);
	    free_psp = psp;
	    psp = psp->next;
	    erts_free(ERTS_ALC_T_PEND_SUSPEND, (void *) free_psp);
	}
    }
    
}

static ERTS_INLINE void
cancel_suspend_of_suspendee(Process *p, ErtsProcLocks p_locks)
{
    if (is_not_nil(p->suspendee)) {
	Process *rp;
	if (!(p_locks & ERTS_PROC_LOCK_STATUS))
	    erts_smp_proc_lock(p, ERTS_PROC_LOCK_STATUS);
	rp = erts_pid2proc(p, p_locks|ERTS_PROC_LOCK_STATUS,
			   p->suspendee, ERTS_PROC_LOCK_STATUS);
	if (rp) {
	    erts_resume(rp, ERTS_PROC_LOCK_STATUS);
	    erts_smp_proc_unlock(rp, ERTS_PROC_LOCK_STATUS);
	}
	if (!(p_locks & ERTS_PROC_LOCK_STATUS))
	    erts_smp_proc_unlock(p, ERTS_PROC_LOCK_STATUS);
	p->suspendee = NIL;
    }
}

static void
handle_pend_sync_suspend(Process *suspendee,
			 ErtsProcLocks suspendee_locks,
			 int suspendee_alive,
			 Eterm suspender_pid)
{
    Process *suspender;

    ERTS_SMP_LC_ASSERT(suspendee_locks & ERTS_PROC_LOCK_STATUS);

    suspender = erts_pid2proc(suspendee,
			      suspendee_locks,
			      suspender_pid,
			      ERTS_PROC_LOCK_STATUS);
    if (suspender) {
	ASSERT(is_nil(suspender->suspendee));
	if (suspendee_alive) {
	    erts_suspend(suspendee, suspendee_locks, NULL);
	    suspender->suspendee = suspendee->common.id;
	}
	/* suspender is suspended waiting for suspendee to suspend;
	   resume suspender */
	resume_process(suspender);
	erts_smp_proc_unlock(suspender, ERTS_PROC_LOCK_STATUS);
    }
}

static Process *
pid2proc_not_running(Process *c_p, ErtsProcLocks c_p_locks,
		     Eterm pid, ErtsProcLocks pid_locks, int suspend)
{
    Process *rp;
    int unlock_c_p_status;

    ERTS_SMP_LC_ASSERT(c_p_locks == erts_proc_lc_my_proc_locks(c_p));

    ERTS_SMP_LC_ASSERT(c_p_locks & ERTS_PROC_LOCK_MAIN);
    ERTS_SMP_LC_ASSERT(pid_locks & (ERTS_PROC_LOCK_MAIN|ERTS_PROC_LOCK_STATUS));

    if (c_p->common.id == pid)
	return erts_pid2proc(c_p, c_p_locks, pid, pid_locks);

    if (c_p_locks & ERTS_PROC_LOCK_STATUS)
	unlock_c_p_status = 0;
    else {
	unlock_c_p_status = 1;
	erts_smp_proc_lock(c_p, ERTS_PROC_LOCK_STATUS);
    }

    if (c_p->suspendee == pid) {
	/* Process previously suspended by c_p (below)... */
	ErtsProcLocks rp_locks = pid_locks|ERTS_PROC_LOCK_STATUS;
	rp = erts_pid2proc(c_p, c_p_locks|ERTS_PROC_LOCK_STATUS, pid, rp_locks);
	c_p->suspendee = NIL;
	ASSERT(c_p->flags & F_P2PNR_RESCHED);
	c_p->flags &= ~F_P2PNR_RESCHED;
	if (!suspend && rp)
	    resume_process(rp);
    }
    else {

	rp = erts_pid2proc(c_p, c_p_locks|ERTS_PROC_LOCK_STATUS,
			   pid, pid_locks|ERTS_PROC_LOCK_STATUS);

	if (!rp) {
	    c_p->flags &= ~F_P2PNR_RESCHED;
	    goto done;
	}

	ASSERT(!(c_p->flags & F_P2PNR_RESCHED));

	if (suspend) {
	    if (suspend_process(c_p, rp))
		goto done;
	}
	else {
	    if (!((ERTS_PSFLG_RUNNING|ERTS_PSFLG_RUNNING_SYS)
		  & erts_smp_atomic32_read_acqb(&rp->state)))
		goto done;

	}

	/* Other process running */

	/*
	 * If we got pending suspenders and suspend ourselves waiting
	 * to suspend another process we might deadlock.
	 * In this case we have to yield, be suspended by
	 * someone else and then do it all over again.
	 */
	if (!c_p->pending_suspenders) {
	    /* Mark rp pending for suspend by c_p */
	    add_pend_suspend(rp, c_p->common.id, handle_pend_sync_suspend);
	    ASSERT(is_nil(c_p->suspendee));

	    /* Suspend c_p; when rp is suspended c_p will be resumed. */
	    suspend_process(c_p, c_p);
	    c_p->flags |= F_P2PNR_RESCHED;
	}
	/* Yield (caller is assumed to yield immediately in bif). */
	erts_smp_proc_unlock(rp, pid_locks|ERTS_PROC_LOCK_STATUS);
	rp = ERTS_PROC_LOCK_BUSY;
    }

 done:
    if (rp && rp != ERTS_PROC_LOCK_BUSY && !(pid_locks & ERTS_PROC_LOCK_STATUS))
	erts_smp_proc_unlock(rp, ERTS_PROC_LOCK_STATUS);
    if (unlock_c_p_status)
	erts_smp_proc_unlock(c_p, ERTS_PROC_LOCK_STATUS);
    return rp;
}


/*
 * Like erts_pid2proc() but:
 *
 * * At least ERTS_PROC_LOCK_MAIN have to be held on c_p.
 * * At least ERTS_PROC_LOCK_MAIN have to be taken on pid.
 * * It also waits for proc to be in a state != running and garbing.
 * * If ERTS_PROC_LOCK_BUSY is returned, the calling process has to
 *   yield (ERTS_BIF_YIELD[0-3]()). c_p might in this case have been
 *   suspended.
 */
Process *
erts_pid2proc_not_running(Process *c_p, ErtsProcLocks c_p_locks,
			  Eterm pid, ErtsProcLocks pid_locks)
{
    return pid2proc_not_running(c_p, c_p_locks, pid, pid_locks, 0);
}

/*
 * Like erts_pid2proc_not_running(), but hands over the process
 * in a suspended state unless (c_p is looked up).
 */
Process *
erts_pid2proc_suspend(Process *c_p, ErtsProcLocks c_p_locks,
		      Eterm pid, ErtsProcLocks pid_locks)
{
    return pid2proc_not_running(c_p, c_p_locks, pid, pid_locks, 1);
}

/*
 * erts_pid2proc_nropt() is normally the same as
 * erts_pid2proc_not_running(). However it is only
 * to be used when 'not running' is a pure optimization,
 * not a requirement.
 */

Process *
erts_pid2proc_nropt(Process *c_p, ErtsProcLocks c_p_locks,
		    Eterm pid, ErtsProcLocks pid_locks)
{
    if (erts_disable_proc_not_running_opt)
	return erts_pid2proc(c_p, c_p_locks, pid, pid_locks);
    else
	return erts_pid2proc_not_running(c_p, c_p_locks, pid, pid_locks);
}

static ERTS_INLINE int
do_bif_suspend_process(Process *c_p,
		       ErtsSuspendMonitor *smon,
		       Process *suspendee)
{
    ASSERT(suspendee);
    ASSERT(!ERTS_PROC_IS_EXITING(suspendee));
    ERTS_SMP_LC_ASSERT(ERTS_PROC_LOCK_STATUS
		       & erts_proc_lc_my_proc_locks(suspendee));
    if (smon) {
	if (!smon->active) {
	    if (!suspend_process(c_p, suspendee))
		return 0;
	}
	smon->active += smon->pending;
	ASSERT(smon->active);
	smon->pending = 0;
	return 1;
    }
    return 0;
}

static void
handle_pend_bif_sync_suspend(Process *suspendee,
			     ErtsProcLocks suspendee_locks,
			     int suspendee_alive,
			     Eterm suspender_pid)
{
    Process *suspender;

    ERTS_SMP_LC_ASSERT(suspendee_locks & ERTS_PROC_LOCK_STATUS);

    suspender = erts_pid2proc(suspendee,
			      suspendee_locks,
			      suspender_pid,
			      ERTS_PROC_LOCK_LINK|ERTS_PROC_LOCK_STATUS);
    if (suspender) {
	ASSERT(is_nil(suspender->suspendee));
	if (!suspendee_alive)
	    erts_delete_suspend_monitor(&suspender->suspend_monitors,
					suspendee->common.id);
	else {
#ifdef DEBUG
	    int res;
#endif
	    ErtsSuspendMonitor *smon;
	    smon = erts_lookup_suspend_monitor(suspender->suspend_monitors,
					       suspendee->common.id);
#ifdef DEBUG
	    res =
#endif
		do_bif_suspend_process(suspendee, smon, suspendee);
	    ASSERT(!smon || res != 0);
	    suspender->suspendee = suspendee->common.id;
	}
	/* suspender is suspended waiting for suspendee to suspend;
	   resume suspender */
	resume_process(suspender);
	erts_smp_proc_unlock(suspender,
			     ERTS_PROC_LOCK_LINK|ERTS_PROC_LOCK_STATUS);
    }
}

static void
handle_pend_bif_async_suspend(Process *suspendee,
			      ErtsProcLocks suspendee_locks,
			      int suspendee_alive,
			      Eterm suspender_pid)
{

    Process *suspender;

    ERTS_SMP_LC_ASSERT(suspendee_locks & ERTS_PROC_LOCK_STATUS);

    suspender = erts_pid2proc(suspendee,
			      suspendee_locks,
			      suspender_pid,
			      ERTS_PROC_LOCK_LINK);
    if (suspender) {
	ASSERT(is_nil(suspender->suspendee));
	if (!suspendee_alive)
	    erts_delete_suspend_monitor(&suspender->suspend_monitors,
					suspendee->common.id);
	else {
#ifdef DEBUG
	    int res;
#endif
	    ErtsSuspendMonitor *smon;
	    smon = erts_lookup_suspend_monitor(suspender->suspend_monitors,
					       suspendee->common.id);
#ifdef DEBUG
	    res =
#endif
	    do_bif_suspend_process(suspendee, smon, suspendee);
	    ASSERT(!smon || res != 0);
	}
	erts_smp_proc_unlock(suspender, ERTS_PROC_LOCK_LINK);
    }
}

#else

/*
 * Non-smp version of erts_pid2proc_suspend().
 */
Process *
erts_pid2proc_suspend(Process *c_p, ErtsProcLocks c_p_locks,
		      Eterm pid, ErtsProcLocks pid_locks)
{
    Process *rp = erts_pid2proc(c_p, c_p_locks, pid, pid_locks);
    if (rp)
	erts_suspend(rp, pid_locks, NULL);
    return rp;
}

#endif /* ERTS_SMP */

/*
 * The erlang:suspend_process/2 BIF
 */

BIF_RETTYPE
suspend_process_2(BIF_ALIST_2)
{
    Eterm res;
    Process* suspendee = NULL;
    ErtsSuspendMonitor *smon;
    ErtsProcLocks xlocks = (ErtsProcLocks) 0;

    /* Options and default values: */
    int asynchronous = 0;
    int unless_suspending = 0;


    if (BIF_P->common.id == BIF_ARG_1)
	goto badarg; /* We are not allowed to suspend ourselves */

    if (is_not_nil(BIF_ARG_2)) {
	/* Parse option list */
	Eterm arg = BIF_ARG_2;

	while (is_list(arg)) {
	    Eterm *lp = list_val(arg);
	    arg = CAR(lp);
	    switch (arg) {
	    case am_unless_suspending:
		unless_suspending = 1;
		break;
	    case am_asynchronous:
		asynchronous = 1;
		break;
	    default:
		goto badarg;
	    }
	    arg = CDR(lp);
	}
	if (is_not_nil(arg))
	    goto badarg;
    }

    xlocks = ERTS_PROC_LOCK_LINK | (asynchronous
				    ? (ErtsProcLocks) 0
				    : ERTS_PROC_LOCK_STATUS);

    erts_smp_proc_lock(BIF_P, xlocks);

    suspendee = erts_pid2proc(BIF_P,
			      ERTS_PROC_LOCK_MAIN|xlocks,
			      BIF_ARG_1,
			      ERTS_PROC_LOCK_STATUS);
    if (!suspendee)
	goto no_suspendee;

    smon = erts_add_or_lookup_suspend_monitor(&BIF_P->suspend_monitors,
					      BIF_ARG_1);
#ifndef ERTS_SMP /* no ERTS_SMP */

    /* This is really a piece of cake without SMP support... */
    if (!smon->active) {
	erts_smp_atomic32_read_bor_nob(&suspendee->state, ERTS_PSFLG_SUSPENDED);
	suspend_process(BIF_P, suspendee);
	smon->active++;
	res = am_true;
    }
    else if (unless_suspending)
	res = am_false;
    else if (smon->active == INT_MAX)
	goto system_limit;
    else {
	smon->active++;
	res = am_true;
    }

#else /* ERTS_SMP */

    /* ... but a little trickier with SMP support ... */

    if (asynchronous) {
	/* --- Asynchronous suspend begin ---------------------------------- */

	ERTS_SMP_LC_ASSERT(ERTS_PROC_LOCK_LINK
			   & erts_proc_lc_my_proc_locks(BIF_P));
	ERTS_SMP_LC_ASSERT(ERTS_PROC_LOCK_STATUS
			   == erts_proc_lc_my_proc_locks(suspendee));

	if (smon->active) {
	    smon->active += smon->pending;
	    smon->pending = 0;
	    if (unless_suspending)
		res = am_false;
	    else if (smon->active == INT_MAX)
		goto system_limit;
	    else {
		smon->active++;
		res = am_true;
	    }
	    /* done */
	}
	else {
	    /* We havn't got any active suspends on the suspendee */
	    if (smon->pending && unless_suspending)
		res = am_false;
	    else {
		if (smon->pending == INT_MAX)
		    goto system_limit;

		smon->pending++;

		if (!do_bif_suspend_process(BIF_P, smon, suspendee))
		    add_pend_suspend(suspendee,
				     BIF_P->common.id,
				     handle_pend_bif_async_suspend);

		res = am_true;
	    }
	    /* done */
	}
	/* --- Asynchronous suspend end ------------------------------------ */
    }
    else /* if (!asynchronous) */ {
	/* --- Synchronous suspend begin ----------------------------------- */

	ERTS_SMP_LC_ASSERT(((ERTS_PROC_LOCK_LINK|ERTS_PROC_LOCK_STATUS)
			    & erts_proc_lc_my_proc_locks(BIF_P))
			   == (ERTS_PROC_LOCK_LINK|ERTS_PROC_LOCK_STATUS));
	ERTS_SMP_LC_ASSERT(ERTS_PROC_LOCK_STATUS
			   == erts_proc_lc_my_proc_locks(suspendee));

	if (BIF_P->suspendee == BIF_ARG_1) {
	    /* We are back after a yield and the suspendee
	       has been suspended on behalf of us. */
	    ASSERT(smon->active >= 1);
	    BIF_P->suspendee = NIL;
	    res = (!unless_suspending || smon->active == 1
		   ? am_true
		   : am_false);
	    /* done */
	}
	else if (smon->active) {
	    if (unless_suspending)
		res = am_false;
	    else {
		smon->active++;
		res = am_true;
	    }
	    /* done */
	}
	else {
	    /* We haven't got any active suspends on the suspendee */

	    /*
	     * If we have pending suspenders and suspend ourselves waiting
	     * to suspend another process, or suspend another process
	     * we might deadlock. In this case we have to yield,
	     * be suspended by someone else, and then do it all over again.
	     */
	    if (BIF_P->pending_suspenders)
		goto yield;

	    if (!unless_suspending && smon->pending == INT_MAX)
		goto system_limit;
	    if (!unless_suspending || smon->pending == 0)
		smon->pending++;

	    if (do_bif_suspend_process(BIF_P, smon, suspendee)) {
		res = (!unless_suspending || smon->active == 1
		       ? am_true
		       : am_false);
		/* done */
	    }
	    else {
		/* Mark suspendee pending for suspend by BIF_P */
		add_pend_suspend(suspendee,
				 BIF_P->common.id,
				 handle_pend_bif_sync_suspend);

		ASSERT(is_nil(BIF_P->suspendee));

		/*
		 * Suspend BIF_P; when suspendee is suspended, BIF_P
		 * will be resumed and this BIF will be called again.
		 * This time with BIF_P->suspendee == BIF_ARG_1 (see
		 * above).
		 */
		suspend_process(BIF_P, BIF_P);
		goto yield;
	    }
	}
	/* --- Synchronous suspend end ------------------------------------- */
    }

#endif /* ERTS_SMP */
#ifdef DEBUG
    {
	erts_aint32_t state = erts_smp_atomic32_read_acqb(&suspendee->state);
	ASSERT((state & ERTS_PSFLG_SUSPENDED)
	       || (asynchronous && smon->pending));
	ASSERT((state & ERTS_PSFLG_SUSPENDED)
	       || !smon->active);
    }
#endif

    erts_smp_proc_unlock(suspendee, ERTS_PROC_LOCK_STATUS);
    erts_smp_proc_unlock(BIF_P, xlocks);
    BIF_RET(res);

 system_limit:
    ERTS_BIF_PREP_ERROR(res, BIF_P, SYSTEM_LIMIT);
    goto do_return;

 no_suspendee:
#ifdef ERTS_SMP
    BIF_P->suspendee = NIL;
#endif
    erts_delete_suspend_monitor(&BIF_P->suspend_monitors, BIF_ARG_1);

 badarg:
    ERTS_BIF_PREP_ERROR(res, BIF_P, BADARG);
#ifdef ERTS_SMP
    goto do_return;

 yield:
    ERTS_BIF_PREP_YIELD2(res, bif_export[BIF_suspend_process_2],
			 BIF_P, BIF_ARG_1, BIF_ARG_2);
#endif

 do_return:
    if (suspendee)
	erts_smp_proc_unlock(suspendee, ERTS_PROC_LOCK_STATUS);
    if (xlocks)
	erts_smp_proc_unlock(BIF_P, xlocks);
    return res;

}


/*
 * The erlang:resume_process/1 BIF
 */

BIF_RETTYPE
resume_process_1(BIF_ALIST_1)
{
    ErtsSuspendMonitor *smon;
    Process *suspendee;
    int is_active;
 
    if (BIF_P->common.id == BIF_ARG_1)
	BIF_ERROR(BIF_P, BADARG);

    erts_smp_proc_lock(BIF_P, ERTS_PROC_LOCK_LINK);
    smon = erts_lookup_suspend_monitor(BIF_P->suspend_monitors, BIF_ARG_1);

    if (!smon) {
	/* No previous suspend or dead suspendee */
	goto error;
    }
    else if (smon->pending) {
	smon->pending--;
	ASSERT(smon->pending >= 0);
	if (smon->active) {
	    smon->active += smon->pending;
	    smon->pending = 0;
	}
	is_active = smon->active;
    }
    else if (smon->active) {
	smon->active--;
	ASSERT(smon->pending >= 0);
	is_active = 1;
    }
    else {
	/* No previous suspend or dead suspendee */
	goto error;
    }

    if (smon->active || smon->pending || !is_active) {
	/* Leave the suspendee as it is; just verify that it is still alive */
	suspendee = erts_pid2proc(BIF_P,
				  ERTS_PROC_LOCK_MAIN|ERTS_PROC_LOCK_LINK,
				  BIF_ARG_1,
				  0);
	if (!suspendee)
	    goto no_suspendee;

    }
    else {
	/* Resume */
	suspendee = erts_pid2proc(BIF_P,
				  ERTS_PROC_LOCK_MAIN|ERTS_PROC_LOCK_LINK,
				  BIF_ARG_1,
				  ERTS_PROC_LOCK_STATUS);
	if (!suspendee)
	    goto no_suspendee;

	ASSERT(ERTS_PSFLG_SUSPENDED
	       & erts_smp_atomic32_read_nob(&suspendee->state));
	resume_process(suspendee);

	erts_smp_proc_unlock(suspendee, ERTS_PROC_LOCK_STATUS);
    }

    if (!smon->active && !smon->pending)
	erts_delete_suspend_monitor(&BIF_P->suspend_monitors, BIF_ARG_1);

    erts_smp_proc_unlock(BIF_P, ERTS_PROC_LOCK_LINK);

    BIF_RET(am_true);

 no_suspendee:
    /* cleanup */
    erts_delete_suspend_monitor(&BIF_P->suspend_monitors, BIF_ARG_1);

 error:
    erts_smp_proc_unlock(BIF_P, ERTS_PROC_LOCK_LINK);
    BIF_ERROR(BIF_P, BADARG);
}

Uint
erts_run_queues_len(Uint *qlen)
{
    int i = 0;
    Uint len = 0;
    ERTS_ATOMIC_FOREACH_RUNQ(rq,
    {
	Sint pqlen = 0;
	int pix;
	for (pix = 0; pix < ERTS_NO_PROC_PRIO_LEVELS; pix++)
	    pqlen += RUNQ_READ_LEN(&rq->procs.prio_info[pix].len);

	if (pqlen < 0)
	    pqlen = 0;
 	if (qlen)
	    qlen[i++] = pqlen;
	len += pqlen;
    }
	);
    return len;
}

Eterm
erts_process_status(Process *c_p, ErtsProcLocks c_p_locks,
		    Process *rp, Eterm rpid)
{
    Eterm res = am_undefined;
    Process *p = rp ? rp : erts_proc_lookup_raw(rpid);

    if (p) {
	erts_aint32_t state = erts_smp_atomic32_read_acqb(&p->state);
	if (state & ERTS_PSFLG_FREE)
	    res = am_free;
	else if (state & ERTS_PSFLG_EXITING)
	    res = am_exiting;
	else if (state & ERTS_PSFLG_GC)
	    res = am_garbage_collecting;
	else if (state & ERTS_PSFLG_SUSPENDED)
	    res = am_suspended;
	else if (state & ERTS_PSFLG_RUNNING)
	    res = am_running;
	else if (state & ERTS_PSFLG_ACTIVE)
	    res = am_runnable;
	else
	    res = am_waiting;
    }
#ifdef ERTS_SMP
    else {
	int i;
	ErtsSchedulerData *esdp;

	for (i = 0; i < erts_no_schedulers; i++) {
	    esdp = ERTS_SCHEDULER_IX(i);
	    erts_smp_runq_lock(esdp->run_queue);
	    if (esdp->free_process
		&& esdp->free_process->common.id == rpid) {
		res = am_free;
		erts_smp_runq_unlock(esdp->run_queue);
		break;
	    }
	    erts_smp_runq_unlock(esdp->run_queue);
	}
    }
#endif
    return res;
}

/*
** Suspend a currently executing process 
** If we are to suspend on a port the busy_port is the thing
** otherwise busy_port is NIL
*/

void
erts_suspend(Process* c_p, ErtsProcLocks c_p_locks, Port *busy_port)
{
    int suspend;

    ASSERT(c_p == erts_get_current_process());
    ERTS_SMP_LC_ASSERT(c_p_locks == erts_proc_lc_my_proc_locks(c_p));
    if (!(c_p_locks & ERTS_PROC_LOCK_STATUS))
	erts_smp_proc_lock(c_p, ERTS_PROC_LOCK_STATUS);

    if (busy_port)
	suspend = erts_save_suspend_process_on_port(busy_port, c_p);
    else
	suspend = 1;

    if (suspend) {
#ifdef DEBUG
	int res =
#endif
	    suspend_process(c_p, c_p);
	ASSERT(res);
    }

    if (!(c_p_locks & ERTS_PROC_LOCK_STATUS))
	erts_smp_proc_unlock(c_p, ERTS_PROC_LOCK_STATUS);

    if (suspend && busy_port && erts_system_monitor_flags.busy_port)
	monitor_generic(c_p, am_busy_port, busy_port->common.id);
}

void
erts_resume(Process* process, ErtsProcLocks process_locks)
{
    ERTS_SMP_LC_ASSERT(process_locks == erts_proc_lc_my_proc_locks(process));
    if (!(process_locks & ERTS_PROC_LOCK_STATUS))
	erts_smp_proc_lock(process, ERTS_PROC_LOCK_STATUS);
    resume_process(process);
    if (!(process_locks & ERTS_PROC_LOCK_STATUS))
	erts_smp_proc_unlock(process, ERTS_PROC_LOCK_STATUS);
}

int
erts_resume_processes(ErtsProcList *list)
{
    /* 'list' is expected to have been fetched (i.e. not a ring anymore) */
    int nresumed = 0;
    ErtsProcList *plp = list;

    while (plp) {
	Process *proc;
	ErtsProcList *fplp;
	ASSERT(is_internal_pid(plp->pid));
	proc = erts_pid2proc(NULL, 0, plp->pid, ERTS_PROC_LOCK_STATUS);
	if (proc) {
	    if (erts_proclist_same(plp, proc)) {
		resume_process(proc);
		nresumed++;
	    }
	    erts_smp_proc_unlock(proc, ERTS_PROC_LOCK_STATUS);
	}
	fplp = plp;
	plp = plp->next;
	proclist_destroy(fplp);
    }
    return nresumed;
}

Eterm
erts_get_process_priority(Process *p)
{
    erts_aint32_t state = erts_smp_atomic32_read_nob(&p->state);
    switch (ERTS_PSFLGS_GET_USR_PRIO(state)) {
    case PRIORITY_MAX:		return am_max;
    case PRIORITY_HIGH:		return am_high;
    case PRIORITY_NORMAL:	return am_normal;
    case PRIORITY_LOW:		return am_low;
    default: ASSERT(0);		return am_undefined;
    }
}

Eterm
erts_set_process_priority(Process *p, Eterm value)
{
    erts_aint32_t a, oprio, nprio;

    switch (value) {
    case am_max:	nprio = (erts_aint32_t) PRIORITY_MAX;		break;
    case am_high:	nprio = (erts_aint32_t) PRIORITY_HIGH;		break;
    case am_normal:	nprio = (erts_aint32_t) PRIORITY_NORMAL;	break;
    case am_low:	nprio = (erts_aint32_t) PRIORITY_LOW;		break;
    default:		return THE_NON_VALUE;				break;
    }

    a = erts_smp_atomic32_read_nob(&p->state);
    if (nprio == ERTS_PSFLGS_GET_USR_PRIO(a))
	oprio = nprio;
    else {
	int slocked = 0;
	erts_aint32_t e, n, aprio;

	if (a & ERTS_PSFLG_ACTIVE_SYS) {
	    erts_smp_proc_lock(p, ERTS_PROC_LOCK_STATUS);
	    slocked = 1;
	}

	do {
	    oprio = ERTS_PSFLGS_GET_USR_PRIO(a);
	    n = e = a;

	    if (!(a & (ERTS_PSFLG_ACTIVE_SYS|ERTS_PSFLG_DELAYED_SYS)))
		aprio = nprio;
	    else {
		int max_qbit;

		if (!slocked) {
		    erts_smp_proc_lock(p, ERTS_PROC_LOCK_STATUS);
		    slocked = 1;
		}

		max_qbit = 0;
		if (a & ERTS_PSFLG_ACTIVE_SYS)
		    max_qbit |= p->sys_task_qs->qmask;
		if (a & ERTS_PSFLG_DELAYED_SYS) {
		    ErtsProcSysTaskQs *qs;
		    qs = ERTS_PROC_GET_DELAYED_GC_TASK_QS(p);
		    ASSERT(qs);
		    max_qbit |= qs->qmask;
		}
		max_qbit &= -max_qbit;
		switch (max_qbit) {
		case MAX_BIT:
		    aprio = PRIORITY_MAX;
		    break;
		case HIGH_BIT:
		    aprio = PRIORITY_HIGH;
		    break;
		case NORMAL_BIT:
		    aprio = PRIORITY_NORMAL;
		    break;
		case LOW_BIT:
		    aprio = PRIORITY_LOW;
		    break;
		default:
		    ERTS_INTERNAL_ERROR("Invalid qmask");
		    aprio = -1;
		}

		if (aprio > nprio) /* low value -> high prio */
		    aprio = nprio;
	    }

	    n &= ~(ERTS_PSFLGS_USR_PRIO_MASK
		   | ERTS_PSFLGS_ACT_PRIO_MASK);
	    n |= ((nprio << ERTS_PSFLGS_USR_PRIO_OFFSET)
		  | (aprio << ERTS_PSFLGS_ACT_PRIO_OFFSET));

	    a = erts_smp_atomic32_cmpxchg_mb(&p->state, n, e);
	} while (a != e);
    }

    switch (oprio) {
    case PRIORITY_MAX:		return am_max;
    case PRIORITY_HIGH:		return am_high;
    case PRIORITY_NORMAL:	return am_normal;
    case PRIORITY_LOW:		return am_low;
    default: ASSERT(0);		return am_undefined;
    }
}

/*
 * schedule() is called from BEAM (process_main()) or HiPE
 * (hipe_mode_switch()) when the current process is to be
 * replaced by a new process. 'calls' is the number of reduction
 * steps the current process consumed.
 * schedule() returns the new process, and the new process'
 * ->fcalls field is initialised with its allowable number of
 * reduction steps.
 *
 * When no process is runnable, or when sufficiently many reduction
 * steps have been made, schedule() calls erl_sys_schedule() to
 * schedule system-level activities.
 *
 * We use the same queue for normal and low prio processes.
 * We reschedule low prio processes a certain number of times 
 * so that normal processes get to run more frequently. 
 */

Process *schedule(Process *p, int calls)
{
    Process *proxy_p = NULL;
    ErtsRunQueue *rq;
    erts_aint_t dt;
    ErtsSchedulerData *esdp;
    int context_reds;
    int fcalls;
    int input_reductions;
    int actual_reds;
    int reds;
    Uint32 flags;
    erts_aint32_t state = 0; /* Supress warning... */

#ifdef USE_VM_PROBES
    if (p != NULL && DTRACE_ENABLED(process_unscheduled)) {
        DTRACE_CHARBUF(process_buf, DTRACE_TERM_BUF_SIZE);

        dtrace_proc_str(p, process_buf);
        DTRACE1(process_unscheduled, process_buf);
    }
#endif

    if (ERTS_USE_MODIFIED_TIMING()) {
	context_reds = ERTS_MODIFIED_TIMING_CONTEXT_REDS;
	input_reductions = ERTS_MODIFIED_TIMING_INPUT_REDS;
    }
    else {
	context_reds = CONTEXT_REDS;
	input_reductions = INPUT_REDUCTIONS;
    }

    ERTS_SMP_LC_ASSERT(ERTS_SCHEDULER_IS_DIRTY(erts_get_scheduler_data())
		       || !erts_thr_progress_is_blocking());

    /*
     * Clean up after the process being scheduled out.
     */
    if (!p) {	/* NULL in the very first schedule() call */
	esdp = erts_get_scheduler_data();
	rq = erts_get_runq_current(esdp);
	ASSERT(esdp);
	fcalls = (int) erts_smp_atomic32_read_acqb(&function_calls);
	actual_reds = reds = 0;
	erts_smp_runq_lock(rq);
    } else {
    sched_out_proc:

#ifdef ERTS_SMP
	ERTS_SMP_CHK_HAVE_ONLY_MAIN_PROC_LOCK(p);
	esdp = p->scheduler_data;
	ASSERT(esdp->current_process == p
	       || esdp->free_process == p);
#else
	esdp = erts_scheduler_data;
	ASSERT(esdp->current_process == p);
#endif
	reds = actual_reds = calls - esdp->virtual_reds;
	if (reds < ERTS_PROC_MIN_CONTEXT_SWITCH_REDS_COST)
	    reds = ERTS_PROC_MIN_CONTEXT_SWITCH_REDS_COST;
	esdp->virtual_reds = 0;

	fcalls = (int) erts_smp_atomic32_add_read_acqb(&function_calls, reds);
	ASSERT(esdp && esdp == erts_get_scheduler_data());

	rq = erts_get_runq_current(esdp);

	p->reds += actual_reds;

	erts_smp_proc_lock(p, ERTS_PROC_LOCK_STATUS);

	state = erts_smp_atomic32_read_acqb(&p->state);

	if (IS_TRACED(p)) {
	    if (IS_TRACED_FL(p, F_TRACE_CALLS) && !(state & ERTS_PSFLG_FREE))
		erts_schedule_time_break(p, ERTS_BP_CALL_TIME_SCHEDULE_OUT);
	    if (state & (ERTS_PSFLG_FREE|ERTS_PSFLG_EXITING)) {
		if (ARE_TRACE_FLAGS_ON(p, F_TRACE_SCHED_EXIT))
		    trace_sched(p, ((state & ERTS_PSFLG_FREE)
				    ? am_out_exited
				    : am_out_exiting));
	    }
	    else {
		if (ARE_TRACE_FLAGS_ON(p, F_TRACE_SCHED))
		    trace_sched(p, am_out);
		else if (ARE_TRACE_FLAGS_ON(p, F_TRACE_SCHED_PROCS))
		    trace_virtual_sched(p, am_out);
	    }
	}

#ifdef ERTS_SMP
	if (state & ERTS_PSFLG_PENDING_EXIT)
	    erts_handle_pending_exit(p, (ERTS_PROC_LOCK_MAIN
					 | ERTS_PROC_LOCK_STATUS));
	if (p->pending_suspenders) 
	    handle_pending_suspend(p, (ERTS_PROC_LOCK_MAIN
				       | ERTS_PROC_LOCK_STATUS));
#endif

	esdp->reductions += reds;

	schedule_out_process(rq, state, p, proxy_p); /* Returns with rq locked! */
	proxy_p = NULL;

	ERTS_PROC_REDUCTIONS_EXECUTED(rq,
				      (int) ERTS_PSFLGS_GET_USR_PRIO(state),
				      reds,
				      actual_reds);

	esdp->current_process = NULL;
#ifdef ERTS_SMP
	p->scheduler_data = NULL;
#endif

	erts_smp_proc_unlock(p, ERTS_PROC_LOCK_MAIN|ERTS_PROC_LOCK_STATUS);

	if (state & ERTS_PSFLG_FREE) {
#ifdef ERTS_SMP
	    ASSERT(esdp->free_process == p);
	    esdp->free_process = NULL;
#else
	    state = erts_smp_atomic32_read_nob(&p->state);
	    if (!(state & ERTS_PSFLG_IN_RUNQ))
		erts_free_proc(p);
#endif
	}

#ifdef ERTS_SMP
	ASSERT(!esdp->free_process);
#endif
	ASSERT(!esdp->current_process);

	ERTS_SMP_CHK_NO_PROC_LOCKS;

	dt = erts_do_time_read_and_reset();
	if (dt) {
	    erts_smp_runq_unlock(rq);
	    erts_bump_timer(dt);
	    erts_smp_runq_lock(rq);
	}
	BM_STOP_TIMER(system);

    }

    ERTS_SMP_LC_ASSERT(ERTS_SCHEDULER_IS_DIRTY(esdp)
		       || !erts_thr_progress_is_blocking());

 check_activities_to_run: {
#ifdef ERTS_SMP
	ErtsMigrationPaths *mps;
	ErtsMigrationPath *mp;
	ErtsProcList *pnd_xtrs = rq->procs.pending_exiters;
	if (erts_proclist_fetch(&pnd_xtrs, NULL)) {
	    rq->procs.pending_exiters = NULL;
	    erts_smp_runq_unlock(rq);
	    handle_pending_exiters(pnd_xtrs);
	    erts_smp_runq_lock(rq);
	}

	if (!ERTS_SCHEDULER_IS_DIRTY(esdp)) {
	    if (rq->check_balance_reds <= 0)
		check_balance(rq);

	    ERTS_SMP_LC_ASSERT(!erts_thr_progress_is_blocking());

	    mps = erts_get_migration_paths_managed();
	    mp = &mps->mpath[rq->ix];

	    if (mp->flags & ERTS_RUNQ_FLGS_IMMIGRATE_QMASK)
		immigrate(rq, mp);
	}

	ERTS_SMP_LC_ASSERT(erts_smp_lc_runq_is_locked(rq));
    continue_check_activities_to_run:
	flags = ERTS_RUNQ_FLGS_GET_NOB(rq);
    continue_check_activities_to_run_known_flags:
	ASSERT(ERTS_SCHEDULER_IS_DIRTY(esdp)
	       || flags & ERTS_RUNQ_FLG_NONEMPTY);

	if (flags & (ERTS_RUNQ_FLG_CHK_CPU_BIND|ERTS_RUNQ_FLG_SUSPENDED)) {
	    if (flags & ERTS_RUNQ_FLG_SUSPENDED) {
		suspend_scheduler(esdp);
		flags = ERTS_RUNQ_FLGS_GET_NOB(rq);
	    }
	    if (flags & ERTS_RUNQ_FLG_CHK_CPU_BIND) {
		flags = ERTS_RUNQ_FLGS_UNSET(rq, ERTS_RUNQ_FLG_CHK_CPU_BIND);
		flags &= ~ ERTS_RUNQ_FLG_CHK_CPU_BIND;
		erts_sched_check_cpu_bind(esdp);
	    }
	}
#ifdef ERTS_DIRTY_SCHEDULERS
	else if (ERTS_SCHEDULER_IS_DIRTY(esdp)
		 && (erts_smp_atomic32_read_acqb(&esdp->ssi->flags)
		     & ERTS_SSI_FLG_SUSPENDED))
	    suspend_scheduler(esdp);
#endif

	{
	    erts_aint32_t aux_work;
	    int leader_update = ERTS_SCHEDULER_IS_DIRTY(esdp) ? 0
		: erts_thr_progress_update(esdp);
	    aux_work = erts_atomic32_read_acqb(&esdp->ssi->aux_work);
	    if (aux_work | leader_update | ERTS_SCHED_FAIR) {
		erts_smp_runq_unlock(rq);
		if (leader_update)
		    erts_thr_progress_leader_update(esdp);
		else if (ERTS_SCHED_FAIR)
		  ERTS_SCHED_FAIR_YIELD();
		if (aux_work)
		    handle_aux_work(&esdp->aux_work_data, aux_work, 0);
		erts_smp_runq_lock(rq);
	    }

	    ERTS_SMP_LC_ASSERT(ERTS_SCHEDULER_IS_DIRTY(esdp)
			       || !erts_thr_progress_is_blocking());
	}
	ERTS_SMP_LC_ASSERT(erts_smp_lc_runq_is_locked(rq));

#else /* ERTS_SMP */
	{
	    erts_aint32_t aux_work;
	    aux_work = erts_atomic32_read_acqb(&esdp->ssi->aux_work);
	    if (aux_work)
		handle_aux_work(&esdp->aux_work_data, aux_work, 0);
	}
#endif /* ERTS_SMP */

	flags = ERTS_RUNQ_FLGS_GET_NOB(rq);

#ifdef ERTS_DIRTY_SCHEDULERS
	if (ERTS_RUNQ_IX_IS_DIRTY(rq->ix) && rq->halt_in_progress) {
	    /*
	     * TODO: if halt in progress, need to put the dirty scheduler
	     * to sleep somewhere around here to prevent it from picking up
	     * new work
	     */
	}
	else
#endif

	if ((!(flags & ERTS_RUNQ_FLGS_QMASK) && !rq->misc.start)
	    || (rq->halt_in_progress && ERTS_EMPTY_RUNQ_PORTS(rq))) {
	    /* Prepare for scheduler wait */
#ifdef ERTS_SMP
	    ERTS_SMP_LC_ASSERT(erts_smp_lc_runq_is_locked(rq));

	    rq->wakeup_other = 0;
	    rq->wakeup_other_reds = 0;

	    flags = ERTS_RUNQ_FLGS_GET_NOB(rq);
	    if (flags & ERTS_RUNQ_FLG_SUSPENDED)
		goto continue_check_activities_to_run_known_flags;
	    if (flags & ERTS_RUNQ_FLG_INACTIVE)
		empty_runq(rq);
	    else {
		if (!ERTS_RUNQ_IX_IS_DIRTY(rq->ix) && try_steal_task(rq))
		    goto continue_check_activities_to_run;

		empty_runq(rq);

		/*
		 * Check for ERTS_RUNQ_FLG_SUSPENDED has to be done
		 * after trying to steal a task.
		 */
		flags = ERTS_RUNQ_FLGS_GET_NOB(rq);
		if (flags & ERTS_RUNQ_FLG_SUSPENDED) {
		    non_empty_runq(rq);
		    flags |= ERTS_RUNQ_FLG_NONEMPTY;
		    goto continue_check_activities_to_run_known_flags;
		}
	    }
#endif

	    scheduler_wait(&fcalls, esdp, rq);

#ifdef ERTS_SMP
	    non_empty_runq(rq);
#endif

	    goto check_activities_to_run;
	}
	else if (!ERTS_SCHEDULER_IS_DIRTY(esdp) &&
		 (fcalls > input_reductions &&
		  prepare_for_sys_schedule(esdp))) {
	    /*
	     * Schedule system-level activities.
	     */

	    erts_smp_atomic32_set_relb(&function_calls, 0);
	    fcalls = 0;

	    ASSERT(!erts_port_task_have_outstanding_io_tasks());

#if 0 /* Not needed since we wont wait in sys schedule */
	    erts_sys_schedule_interrupt(0);
#endif
	    erts_smp_runq_unlock(rq);
	    erl_sys_schedule(1);
	    dt = erts_do_time_read_and_reset();
	    if (dt) erts_bump_timer(dt);

#ifdef ERTS_SMP
	    erts_smp_runq_lock(rq);
	    clear_sys_scheduling();
	    goto continue_check_activities_to_run;
#else
	    goto check_activities_to_run;
#endif
	}

	if (rq->misc.start)
	    exec_misc_ops(rq);

#ifdef ERTS_SMP
	wakeup_other.check(rq, flags);
#endif

	/*
	 * Find a new port to run.
	 */

	if (RUNQ_READ_LEN(&rq->ports.info.len)) {
	    int have_outstanding_io;
	    have_outstanding_io = erts_port_task_execute(rq, &esdp->current_port);
	    if ((have_outstanding_io && fcalls > 2*input_reductions)
		|| rq->halt_in_progress) {
		/*
		 * If we have performed more than 2*INPUT_REDUCTIONS since
		 * last call to erl_sys_schedule() and we still haven't
		 * handled all I/O tasks we stop running processes and
		 * focus completely on ports.
		 *
		 * One could argue that this is a strange behavior. The
		 * reason for doing it this way is that it is similar
		 * to the behavior before port tasks were introduced.
		 * We don't want to change the behavior too much, at
		 * least not at the time of writing. This behavior
		 * might change in the future.
		 *
		 * /rickard
		 */
		goto check_activities_to_run;
	    }
	}

	/*
	 * Find a new process to run.
	 */
    pick_next_process: {
	    erts_aint32_t psflg_band_mask;
	    int prio_q;
	    int qmask;

	    flags = ERTS_RUNQ_FLGS_GET_NOB(rq);
	    qmask = (int) (flags & ERTS_RUNQ_FLGS_PROCS_QMASK);
	    switch (qmask & -qmask) {
	    case MAX_BIT:
		prio_q = PRIORITY_MAX;
		break;
	    case HIGH_BIT:
		prio_q = PRIORITY_HIGH;
		break;
	    case NORMAL_BIT:
	    case LOW_BIT:
		prio_q = PRIORITY_NORMAL;
		if (check_requeue_process(rq, PRIORITY_NORMAL))
		    goto pick_next_process;
		break;
	    case 0:			/* No process at all */
	    default:
		ASSERT(qmask == 0);
		goto check_activities_to_run;
	    }

	    BM_START_TIMER(system);

	    /*
	     * Take the chosen process out of the queue.
	     */
	    p = dequeue_process(rq, prio_q, &state);

	    ASSERT(p); /* Wrong qmask in rq->flags? */

	    psflg_band_mask = ~(((erts_aint32_t) 1) << (ERTS_PSFLGS_GET_PRQ_PRIO(state)
							+ ERTS_PSFLGS_IN_PRQ_MASK_OFFSET));

#ifdef ERTS_DIRTY_SCHEDULERS
	    ASSERT((state & (ERTS_PSFLG_DIRTY_CPU_PROC|ERTS_PSFLG_DIRTY_IO_PROC)) !=
		   (ERTS_PSFLG_DIRTY_CPU_PROC|ERTS_PSFLG_DIRTY_IO_PROC));
	    if (state & (ERTS_PSFLG_DIRTY_CPU_PROC|ERTS_PSFLG_DIRTY_IO_PROC)) {
		ASSERT((ERTS_SCHEDULER_IS_DIRTY_CPU(esdp) && (state & ERTS_PSFLG_DIRTY_CPU_PROC)) ||
		       (ERTS_SCHEDULER_IS_DIRTY_IO(esdp) && (state & ERTS_PSFLG_DIRTY_IO_PROC)));
		if (!ERTS_SCHEDULER_IS_DIRTY(esdp) && !(state & ERTS_PSFLG_ACTIVE_SYS))
		    goto pick_next_process;
		state &= ~(ERTS_PSFLG_DIRTY_CPU_PROC_IN_Q|ERTS_PSFLG_DIRTY_IO_PROC_IN_Q);
	    }
#endif

	    if (!(state & ERTS_PSFLG_PROXY))
		psflg_band_mask &= ~ERTS_PSFLG_IN_RUNQ;
	    else {
		proxy_p = p;
		p = erts_proc_lookup_raw(proxy_p->common.id);
		if (!p) {
		    free_proxy_proc(proxy_p);
		    proxy_p = NULL;
		    goto pick_next_process;
		}
		state = erts_smp_atomic32_read_nob(&p->state);
	    }

	    while (1) {
		erts_aint32_t exp, new, tmp;
		tmp = new = exp = state;
		new &= psflg_band_mask;
		if (!(state & (ERTS_PSFLG_RUNNING
			       | ERTS_PSFLG_RUNNING_SYS))) {
		    tmp = state & (ERTS_PSFLG_SUSPENDED
				   | ERTS_PSFLG_PENDING_EXIT
				   | ERTS_PSFLG_ACTIVE_SYS);
		    if (tmp != ERTS_PSFLG_SUSPENDED) {
			if (state & ERTS_PSFLG_ACTIVE_SYS)
			    new |= ERTS_PSFLG_RUNNING_SYS;
			else
			    new |= ERTS_PSFLG_RUNNING;
		    }
		}
		state = erts_smp_atomic32_cmpxchg_relb(&p->state, new, exp);
		if (state == exp) {
		    if ((state & (ERTS_PSFLG_RUNNING
				  | ERTS_PSFLG_RUNNING_SYS
				  | ERTS_PSFLG_FREE))
			|| ((state & (ERTS_PSFLG_SUSPENDED
				      | ERTS_PSFLG_PENDING_EXIT
				      | ERTS_PSFLG_ACTIVE_SYS))
			    == ERTS_PSFLG_SUSPENDED)) {
			if (state & ERTS_PSFLG_FREE) {
#ifdef ERTS_SMP
			    erts_smp_proc_dec_refc(p);
#else
			    erts_free_proc(p);
#endif
			}
			if (proxy_p) {
			    free_proxy_proc(proxy_p);
			    proxy_p = NULL;
			}
			goto pick_next_process;
		    }
		    state = new;
		    break;
		}
	    }

	    rq->procs.context_switches++;

	    esdp->current_process = p;

	}

#ifdef ERTS_SMP
	erts_smp_runq_unlock(rq);

	if (flags & ERTS_RUNQ_FLG_PROTECTED)
	    (void) ERTS_RUNQ_FLGS_UNSET(rq, ERTS_RUNQ_FLG_PROTECTED);

	ERTS_SMP_CHK_NO_PROC_LOCKS;

	erts_smp_proc_lock(p, ERTS_PROC_LOCK_MAIN|ERTS_PROC_LOCK_STATUS);

	if (erts_sched_stat.enabled) {
	    int prio;
	    UWord old = ERTS_PROC_SCHED_ID(p,
					  (ERTS_PROC_LOCK_MAIN
					   | ERTS_PROC_LOCK_STATUS),
					  (UWord) esdp->no);
	    int migrated = old && old != esdp->no;

#ifdef ERTS_DIRTY_SCHEDULERS
	    ASSERT(!ERTS_SCHEDULER_IS_DIRTY(esdp));
#endif

	    prio = (int) ERTS_PSFLGS_GET_USR_PRIO(state);

	    erts_smp_spin_lock(&erts_sched_stat.lock);
	    erts_sched_stat.prio[prio].total_executed++;
	    erts_sched_stat.prio[prio].executed++;
	    if (migrated) {
		erts_sched_stat.prio[prio].total_migrated++;
		erts_sched_stat.prio[prio].migrated++;
	    }
	    erts_smp_spin_unlock(&erts_sched_stat.lock);
	}

	if (ERTS_PROC_PENDING_EXIT(p)) {
	    erts_handle_pending_exit(p,
				     ERTS_PROC_LOCK_MAIN|ERTS_PROC_LOCK_STATUS);
	    state = erts_smp_atomic32_read_nob(&p->state);
	}
	ASSERT(!p->scheduler_data);
	p->scheduler_data = esdp;
#endif
	reds = context_reds;

	if (IS_TRACED(p)) {
	    if (state & ERTS_PSFLG_EXITING) {
		if (ARE_TRACE_FLAGS_ON(p, F_TRACE_SCHED_EXIT))
		    trace_sched(p, am_in_exiting);
	    }
	    else {
		if (ARE_TRACE_FLAGS_ON(p, F_TRACE_SCHED))
		    trace_sched(p, am_in);
		else if (ARE_TRACE_FLAGS_ON(p, F_TRACE_SCHED_PROCS))
		    trace_virtual_sched(p, am_in);
	    }
	    if (IS_TRACED_FL(p, F_TRACE_CALLS)) {
		erts_schedule_time_break(p, ERTS_BP_CALL_TIME_SCHEDULE_IN);
	    }
	}

	erts_smp_proc_unlock(p, ERTS_PROC_LOCK_STATUS);

#ifdef ERTS_SMP
	if (is_not_nil(ERTS_TRACER_PROC(p)))
	    erts_check_my_tracer_proc(p);
#endif

	if (state & ERTS_PSFLG_RUNNING_SYS) {
	    reds -= execute_sys_tasks(p, &state, reds);
	    if (reds <= 0
#ifdef ERTS_DIRTY_SCHEDULERS
		|| (state & (ERTS_PSFLG_DIRTY_CPU_PROC|ERTS_PSFLG_DIRTY_IO_PROC))
#endif
		) {
		p->fcalls = reds;
		goto sched_out_proc;
	    }

	    ASSERT(state & ERTS_PSFLG_RUNNING_SYS);
	    ASSERT(!(state & ERTS_PSFLG_RUNNING));

	    while (1) {
		erts_aint32_t n, e;

		if (((state & (ERTS_PSFLG_SUSPENDED
			       | ERTS_PSFLG_ACTIVE)) != ERTS_PSFLG_ACTIVE)
		    && !(state & ERTS_PSFLG_EXITING))
		    goto sched_out_proc;

		n = e = state;
		n &= ~ERTS_PSFLG_RUNNING_SYS;
		n |= ERTS_PSFLG_RUNNING;

		state = erts_smp_atomic32_cmpxchg_mb(&p->state, n, e);
		if (state == e) {
		    state = n;
		    break;
		}

		ASSERT(state & ERTS_PSFLG_RUNNING_SYS);
		ASSERT(!(state & ERTS_PSFLG_RUNNING));
	    }
	}

	if (!(state & ERTS_PSFLG_EXITING)
	    && ((FLAGS(p) & F_FORCE_GC)
		|| (MSO(p).overhead > BIN_VHEAP_SZ(p)))) {
	    reds -= erts_garbage_collect(p, 0, p->arg_reg, p->arity);
	    if (reds <= 0) {
		p->fcalls = reds;
		goto sched_out_proc;
	    }
	}

	if (proxy_p) {
	    free_proxy_proc(proxy_p);
	    proxy_p = NULL;
	}
	    
	p->fcalls = reds;
	ERTS_SMP_CHK_HAVE_ONLY_MAIN_PROC_LOCK(p);

	/* Never run a suspended process */
	ASSERT(!(ERTS_PSFLG_SUSPENDED & erts_smp_atomic32_read_nob(&p->state)));

	return p;
    }
}

static int
notify_sys_task_executed(Process *c_p, ErtsProcSysTask *st, Eterm st_result)
{
    Process *rp = erts_proc_lookup(st->requester);
    if (rp) {
	ErtsProcLocks rp_locks;
	ErlOffHeap *ohp;
	ErlHeapFragment* bp;
	Eterm *hp, msg, req_id, result;
	Uint st_result_sz, hsz;
#ifdef DEBUG
	Eterm *hp_start;
#endif

	rp_locks = (c_p == rp) ? ERTS_PROC_LOCK_MAIN : 0;

	st_result_sz = is_immed(st_result) ? 0 : size_object(st_result);
	hsz = st->req_id_sz + st_result_sz + 4 /* 3-tuple */;

	hp = erts_alloc_message_heap(hsz,
				     &bp,
				     &ohp,
				     rp,
				     &rp_locks);

#ifdef DEBUG
	hp_start = hp;
#endif

	req_id = st->req_id_sz == 0 ? st->req_id : copy_struct(st->req_id,
							       st->req_id_sz,
							       &hp,
							       ohp);

	result = st_result_sz == 0 ? st_result :  copy_struct(st_result,
							      st_result_sz,
							      &hp,
							      ohp);

	ASSERT(is_immed(st->reply_tag));

	msg = TUPLE3(hp, st->reply_tag, req_id, result);

#ifdef DEBUG
	hp += 4;
	ASSERT(hp_start + hsz == hp);
#endif

	erts_queue_message(rp,
			   &rp_locks,
			   bp,
			   msg,
			   NIL
#ifdef USE_VM_PROBES
			   , NIL
#endif
	    );

	if (c_p == rp)
	    rp_locks &= ~ERTS_PROC_LOCK_MAIN;

	if (rp_locks)
	    erts_smp_proc_unlock(rp, rp_locks);
    }

    erts_cleanup_offheap(&st->off_heap);

    erts_free(ERTS_ALC_T_PROC_SYS_TSK, st);

    return rp ? 1 : 0;
}

static ERTS_INLINE ErtsProcSysTask *
fetch_sys_task(Process *c_p, erts_aint32_t state, int *qmaskp, int *priop)
{
    ErtsProcSysTaskQs *unused_qs = NULL;
    int qbit, qmask;
    ErtsProcSysTask *st, **qp;

    *priop = -1; /* Shut up annoying erroneous warning */

    erts_smp_proc_lock(c_p, ERTS_PROC_LOCK_STATUS);

    if (!c_p->sys_task_qs) {
	qmask = 0;
	st = NULL;
	goto update_state;
    }

    qmask = c_p->sys_task_qs->qmask;

    if ((state & (ERTS_PSFLG_ACTIVE
		  | ERTS_PSFLG_EXITING
		  | ERTS_PSFLG_SUSPENDED)) == ERTS_PSFLG_ACTIVE) {
	/* No sys tasks if we got exclusively higher prio user work to do */
	st = NULL;
	switch (ERTS_PSFLGS_GET_USR_PRIO(state)) {
	case PRIORITY_MAX:
	    if (!(qmask & MAX_BIT))
		goto done;
	    break;
	case PRIORITY_HIGH:
	    if (!(qmask & (MAX_BIT|HIGH_BIT)))
		goto done;
	    break;
	default:
	    break;
	}
    }

    qbit = qmask & -qmask;
    switch (qbit) {
    case MAX_BIT:
	qp = &c_p->sys_task_qs->q[PRIORITY_MAX];
	*priop = PRIORITY_MAX;
	break;
    case HIGH_BIT:
	qp = &c_p->sys_task_qs->q[PRIORITY_HIGH];
	*priop = PRIORITY_HIGH;
	break;
    case NORMAL_BIT:
	if (!(qmask & PRIORITY_LOW)
	    || ++c_p->sys_task_qs->ncount <= RESCHEDULE_LOW) {
	    qp = &c_p->sys_task_qs->q[PRIORITY_NORMAL];
	    *priop = PRIORITY_NORMAL;
	    break;
	}
	c_p->sys_task_qs->ncount = 0;
	/* Fall through */
    case LOW_BIT:
	qp = &c_p->sys_task_qs->q[PRIORITY_LOW];
	*priop = PRIORITY_LOW;
	break;
    default:
	ERTS_INTERNAL_ERROR("Invalid qmask");
    }

    st = *qp;
    ASSERT(st);
    if (st->next != st) {
	*qp = st->next;
	st->next->prev = st->prev;
	st->prev->next = st->next;
    }
    else {
	erts_aint32_t a, e, n, st_prio, qmask2;

	*qp = NULL;
	qmask &= ~qbit;
	c_p->sys_task_qs->qmask = qmask;

    update_state:

	qmask2 = qmask;

	if (state & ERTS_PSFLG_DELAYED_SYS) {
	    ErtsProcSysTaskQs *qs = ERTS_PROC_GET_DELAYED_GC_TASK_QS(c_p);
	    ASSERT(qs);
	    qmask2 |= qs->qmask;
	}

	switch (qmask2 & -qmask2) {
	case MAX_BIT:
	    st_prio = PRIORITY_MAX;
	    break;
	case HIGH_BIT:
	    st_prio = PRIORITY_HIGH;
	    break;
	case NORMAL_BIT:
	    st_prio = PRIORITY_NORMAL;
	    break;
	case LOW_BIT:
	case 0:
	    st_prio = PRIORITY_LOW;
	    break;
	default:
	    ERTS_INTERNAL_ERROR("Invalid qmask");
	}

	if (!qmask) {
	    unused_qs = c_p->sys_task_qs;
	    c_p->sys_task_qs = NULL;
	}

	a = state;
	do {
	    erts_aint32_t prio = ERTS_PSFLGS_GET_USR_PRIO(a);

	    if (prio > st_prio)
		prio = st_prio;

	    n = e = a;

	    n &= ~ERTS_PSFLGS_ACT_PRIO_MASK;
	    n |= (prio << ERTS_PSFLGS_ACT_PRIO_OFFSET);

	    if (!qmask)
		n &= ~ERTS_PSFLG_ACTIVE_SYS;

	    if (a == n)
		break;
	    a = erts_smp_atomic32_cmpxchg_nob(&c_p->state, n, e);
	} while (a != e);
    }

done:

    erts_smp_proc_unlock(c_p, ERTS_PROC_LOCK_STATUS);

    if (unused_qs)
	proc_sys_task_queues_free(unused_qs);

    *qmaskp = qmask;

    return st;
}

static void save_gc_task(Process *c_p, ErtsProcSysTask *st, int prio);

static int
execute_sys_tasks(Process *c_p, erts_aint32_t *statep, int in_reds)
{
    int garbage_collected = 0;
    erts_aint32_t state = *statep;
    int max_reds = in_reds;
    int reds = 0;
    int qmask = 0;

    ERTS_SMP_LC_ASSERT(erts_proc_lc_my_proc_locks(c_p) == ERTS_PROC_LOCK_MAIN);

    do {
	ErtsProcSysTask *st;
	int st_prio;
	Eterm st_res;

	if (state & (ERTS_PSFLG_EXITING|ERTS_PSFLG_PENDING_EXIT)) {
#ifdef ERTS_SMP
	    if (state & ERTS_PSFLG_PENDING_EXIT)
		erts_handle_pending_exit(c_p, ERTS_PROC_LOCK_MAIN);
#endif
	    ASSERT(ERTS_PROC_IS_EXITING(c_p));
	    break;
	}

	st = fetch_sys_task(c_p, state, &qmask, &st_prio);
	if (!st)
	    break;

	switch (st->type) {
	case ERTS_PSTT_GC:
	    if (c_p->flags & F_DISABLE_GC) {
		save_gc_task(c_p, st, st_prio);
		st = NULL;
		reds++;
	    }
	    else {
		if (!garbage_collected) {
		    FLAGS(c_p) |= F_NEED_FULLSWEEP;
		    reds += erts_garbage_collect(c_p,
						 0,
						 c_p->arg_reg,
						 c_p->arity);
		    garbage_collected = 1;
		}
		st_res = am_true;
	    }
	    break;
	case ERTS_PSTT_CPC:
	    st_res = erts_check_process_code(c_p,
					     st->arg[0],
					     st->arg[1] == am_true,
					     &reds);
	    if (is_non_value(st_res)) {
		/* Needed gc, but gc was disabled */
		save_gc_task(c_p, st, st_prio);
		st = NULL;
	    }
	    break;
	default:
	    ERTS_INTERNAL_ERROR("Invalid process sys task type");
	    st_res = am_false;
	}

	if (st)
	    reds += notify_sys_task_executed(c_p, st, st_res);

	state = erts_smp_atomic32_read_acqb(&c_p->state);
    } while (qmask && reds < max_reds);

    *statep = state;

    return reds;
}

static int
cleanup_sys_tasks(Process *c_p, erts_aint32_t in_state, int in_reds)
{
    erts_aint32_t state = in_state;
    int max_reds = in_reds;
    int reds = 0;
    int qmask = 0;

    ERTS_SMP_LC_ASSERT(erts_proc_lc_my_proc_locks(c_p) == ERTS_PROC_LOCK_MAIN);

    do {
	ErtsProcSysTask *st;
	Eterm st_res;
	int st_prio;

	st = fetch_sys_task(c_p, state, &qmask, &st_prio);
	if (!st)
	    break;

	switch (st->type) {
	case ERTS_PSTT_GC:
	    st_res = am_false;
	    break;
	case ERTS_PSTT_CPC:
	    st_res = am_false;
	    break;
	default:
	    ERTS_INTERNAL_ERROR("Invalid process sys task type");
	    st_res = am_false;
	    break;
	}

	reds += notify_sys_task_executed(c_p, st, st_res);

	state = erts_smp_atomic32_read_acqb(&c_p->state);
    } while (qmask && reds < max_reds);

    return reds;
}

BIF_RETTYPE
erts_internal_request_system_task_3(BIF_ALIST_3)
{
    Process *rp = erts_proc_lookup(BIF_ARG_1);
    ErtsProcSysTaskQs *stqs, *free_stqs = NULL;
    ErtsProcSysTask *st = NULL;
    erts_aint32_t prio, rp_state;
    int rp_locked;
    Eterm noproc_res, req_type;

    if (!rp && !is_internal_pid(BIF_ARG_1)) {
	if (!is_external_pid(BIF_ARG_1))
	    goto badarg;
	if (external_pid_dist_entry(BIF_ARG_1) != erts_this_dist_entry)
	    goto badarg;
    }

    switch (BIF_ARG_2) {
    case am_max:	prio = PRIORITY_MAX;	break;
    case am_high:	prio = PRIORITY_HIGH;	break;
    case am_normal:	prio = PRIORITY_NORMAL;	break;
    case am_low:	prio = PRIORITY_LOW;	break;
    default: goto badarg;
    }

    if (is_not_tuple(BIF_ARG_3))
	goto badarg;
    else {
	int i;
	Eterm *tp = tuple_val(BIF_ARG_3);
	Uint arity = arityval(*tp);
	Eterm req_id;
	Uint req_id_sz;
	Eterm arg[ERTS_MAX_PROC_SYS_TASK_ARGS];
	Uint arg_sz[ERTS_MAX_PROC_SYS_TASK_ARGS];
	Uint tot_sz;
	Eterm *hp;

	if (arity < 2)
	    goto badarg;
	if (arity > 2 + ERTS_MAX_PROC_SYS_TASK_ARGS)
	    goto badarg;
	req_type = tp[1];
	req_id = tp[2];
	req_id_sz = is_immed(req_id) ? req_id : size_object(req_id);
	tot_sz = req_id_sz;
	for (i = 0; i < ERTS_MAX_PROC_SYS_TASK_ARGS; i++) {
	    int tix = 3 + i;
	    if (tix > arity) {
		arg[i] = THE_NON_VALUE;
		arg_sz[i] = 0;
	    }
	    else {
		arg[i] = tp[tix];
		if (is_immed(arg[i]))
		    arg_sz[i] = 0;
		else {
		    arg_sz[i] = size_object(arg[i]);
		    tot_sz += arg_sz[i];
		}
	    }
	}
	st = erts_alloc(ERTS_ALC_T_PROC_SYS_TSK,
			ERTS_PROC_SYS_TASK_SIZE(tot_sz));
	st->next = st->prev = st; /* Prep for empty prio queue */
	ERTS_INIT_OFF_HEAP(&st->off_heap);
	hp = &st->heap[0];

	st->requester = BIF_P->common.id;
	st->reply_tag = req_type;
	st->req_id_sz = req_id_sz;
	st->req_id = req_id_sz == 0 ? req_id : copy_struct(req_id,
							   req_id_sz,
							   &hp,
							   &st->off_heap);

	for (i = 0; i < ERTS_MAX_PROC_SYS_TASK_ARGS; i++)
	    st->arg[i] = arg_sz[i] == 0 ? arg[i] : copy_struct(arg[i],
							       arg_sz[i],
							       &hp,
							       &st->off_heap);
	ASSERT(&st->heap[0] + tot_sz == hp);
    }

    switch (req_type) {

    case am_garbage_collect:
	st->type = ERTS_PSTT_GC;
	noproc_res = am_false;
	if (!rp)
	    goto noproc;
	break;

    case am_check_process_code:
	if (is_not_atom(st->arg[0]))
	    goto badarg;
	if (st->arg[1] != am_true && st->arg[1] != am_false)
	    goto badarg;
	noproc_res = am_false;
	st->type = ERTS_PSTT_CPC;
	if (!rp)
	    goto noproc;
	break;

    default:
	goto badarg;
    }

    rp_state = erts_smp_atomic32_read_nob(&rp->state);

    rp_locked = 0;

    free_stqs = NULL;
    if (rp_state & ERTS_PSFLG_ACTIVE_SYS)
	stqs = NULL;
    else {
    alloc_qs:
	stqs = proc_sys_task_queues_alloc();
	stqs->qmask = 1 << prio;
	stqs->ncount = 0;
	stqs->q[PRIORITY_MAX] = NULL;
	stqs->q[PRIORITY_HIGH] = NULL;
	stqs->q[PRIORITY_NORMAL] = NULL;
	stqs->q[PRIORITY_LOW] = NULL;
	stqs->q[prio] = st;
    }

    if (!rp_locked) {
	rp_locked = 1;
	erts_smp_proc_lock(rp, ERTS_PROC_LOCK_STATUS);

	rp_state = erts_smp_atomic32_read_nob(&rp->state);
	if (rp_state & ERTS_PSFLG_EXITING) {
	    erts_smp_proc_unlock(rp, ERTS_PROC_LOCK_STATUS);
	    rp = NULL;
	    free_stqs = stqs;
	    goto noproc;
	}
    }

    if (!rp->sys_task_qs) {
	if (stqs)
	    rp->sys_task_qs = stqs;
	else
	    goto alloc_qs;
    }
    else {
	if (stqs)
	    free_stqs = stqs;
	stqs = rp->sys_task_qs;
	if (!stqs->q[prio]) {
	    stqs->q[prio] = st;
	    stqs->qmask |= 1 << prio;
	}
	else {
	    st->next = stqs->q[prio];
	    st->prev = stqs->q[prio]->prev;
	    st->next->prev = st;
	    st->prev->next = st;
	    ASSERT(stqs->qmask & (1 << prio));
	}
    }

    if (ERTS_PSFLGS_GET_ACT_PRIO(rp_state) > prio) {
	erts_aint32_t n, a, e;
	/* Need to elevate actual prio */

	a = rp_state;
	do {
	    if (ERTS_PSFLGS_GET_ACT_PRIO(a) <= prio) {
		n = a;
		break;
	    }
	    n = e = a;
	    n &= ~ERTS_PSFLGS_ACT_PRIO_MASK;
	    n |= (prio << ERTS_PSFLGS_ACT_PRIO_OFFSET);
	    a = erts_smp_atomic32_cmpxchg_nob(&rp->state, n, e);    
	} while (a != e);
	rp_state = n;
    }

    erts_smp_proc_unlock(rp, ERTS_PROC_LOCK_STATUS);

    schedule_process_sys_task(rp, rp_state, NULL);

    if (free_stqs)
	proc_sys_task_queues_free(free_stqs);

    BIF_RET(am_ok);

noproc:

    notify_sys_task_executed(BIF_P, st, noproc_res);
    if (free_stqs)
	proc_sys_task_queues_free(free_stqs);
    BIF_RET(am_ok);

badarg:

    if (st) {
	erts_cleanup_offheap(&st->off_heap);
	erts_free(ERTS_ALC_T_PROC_SYS_TSK, st);
    }
    if (free_stqs)
	proc_sys_task_queues_free(free_stqs);
    BIF_ERROR(BIF_P, BADARG);
}

static void
save_gc_task(Process *c_p, ErtsProcSysTask *st, int prio)
{
    erts_aint32_t state;
    ErtsProcSysTaskQs *qs;

    ERTS_SMP_LC_ASSERT(ERTS_PROC_LOCK_MAIN == erts_proc_lc_my_proc_locks(c_p));

    qs = ERTS_PROC_GET_DELAYED_GC_TASK_QS(c_p);
    if (!qs) {
	st->next = st->prev = st;
	qs = proc_sys_task_queues_alloc();
	qs->qmask = 1 << prio;
	qs->ncount = 0;
	qs->q[PRIORITY_MAX] = NULL;
	qs->q[PRIORITY_HIGH] = NULL;
	qs->q[PRIORITY_NORMAL] = NULL;
	qs->q[PRIORITY_LOW] = NULL;
	qs->q[prio] = st;
	(void) ERTS_PROC_SET_DELAYED_GC_TASK_QS(c_p, ERTS_PROC_LOCK_MAIN, qs);
    }
    else {
	if (!qs->q[prio]) {
	    st->next = st->prev = st;
	    qs->q[prio] = st;
	    qs->qmask |= 1 << prio;
	}
	else {
	    st->next = qs->q[prio];
	    st->prev = qs->q[prio]->prev;
	    st->next->prev = st;
	    st->prev->next = st;
	    ASSERT(qs->qmask & (1 << prio));
	}
    }

    state = erts_smp_atomic32_read_nob(&c_p->state);
    ASSERT((ERTS_PSFLG_RUNNING|ERTS_PSFLG_RUNNING_SYS) & state);

    while (!(state & ERTS_PSFLG_DELAYED_SYS)
	   || prio < ERTS_PSFLGS_GET_ACT_PRIO(state)) {
	erts_aint32_t n, e;

	n = e = state;
	n |= ERTS_PSFLG_DELAYED_SYS;
	if (prio < ERTS_PSFLGS_GET_ACT_PRIO(state)) {
	    n &= ~ERTS_PSFLGS_ACT_PRIO_MASK;
	    n |= prio << ERTS_PSFLGS_ACT_PRIO_OFFSET;
	}
	state = erts_smp_atomic32_cmpxchg_relb(&c_p->state, n, e);
	if (state == e)
	    break;
    }
}

int
erts_set_gc_state(Process *c_p, int enable)
{
    ErtsProcSysTaskQs *dgc_tsk_qs;
    ASSERT(c_p == erts_get_current_process());
    ASSERT((ERTS_PSFLG_RUNNING|ERTS_PSFLG_RUNNING_SYS)
	   & erts_smp_atomic32_read_nob(&c_p->state));
    ERTS_SMP_LC_ASSERT(ERTS_PROC_LOCK_MAIN == erts_proc_lc_my_proc_locks(c_p));

    if (!enable) {
	c_p->flags |= F_DISABLE_GC;
	return 0;
    }

    c_p->flags &= ~F_DISABLE_GC;

    dgc_tsk_qs = ERTS_PROC_GET_DELAYED_GC_TASK_QS(c_p);
    if (!dgc_tsk_qs)
	return 0;

    /* Move delayed gc tasks into sys tasks queues. */

    erts_smp_proc_lock(c_p, ERTS_PROC_LOCK_STATUS);

    if (!c_p->sys_task_qs) {
	c_p->sys_task_qs = dgc_tsk_qs;
	dgc_tsk_qs = NULL;
    }
    else {
	ErtsProcSysTaskQs *stsk_qs;
	int prio;

	/*
	 * We push delayed tasks to the front of the queue
	 * since they have already made it to the front
	 * once and then been delayed after that.
	 */

	stsk_qs = c_p->sys_task_qs;

	while (dgc_tsk_qs->qmask) {
	    int qbit = dgc_tsk_qs->qmask & -dgc_tsk_qs->qmask;
	    dgc_tsk_qs->qmask &= ~qbit;
	    switch (qbit) {
	    case MAX_BIT:
		prio = PRIORITY_MAX;
		break;
	    case HIGH_BIT:
		prio = PRIORITY_HIGH;
		break;
	    case NORMAL_BIT:
		prio = PRIORITY_NORMAL;
		break;
	    case LOW_BIT:
		prio = PRIORITY_LOW;
		break;
	    default:
		ERTS_INTERNAL_ERROR("Invalid qmask");
		prio = -1;
		break;
	    }

	    ASSERT(dgc_tsk_qs->q[prio]);

	    if (!stsk_qs->q[prio]) {
		stsk_qs->q[prio] = dgc_tsk_qs->q[prio];
		stsk_qs->qmask |= 1 << prio;
	    }
	    else {
		ErtsProcSysTask *first1, *last1, *first2, *last2;

		ASSERT(stsk_qs->qmask & (1 << prio));
		first1 = dgc_tsk_qs->q[prio];
		last1 = first1->prev;
		first2 = stsk_qs->q[prio];
		last2 = first1->prev;

		last1->next = first2;
		first2->prev = last1;

		first1->prev = last2;
		last2->next = first1;

		stsk_qs->q[prio] = first1;
	    }

	}
    }

#ifdef DEBUG
    {
	int qmask;
	erts_aint32_t aprio, state =
#endif

	    erts_smp_atomic32_read_bset_nob(&c_p->state,
					    (ERTS_PSFLG_DELAYED_SYS
					     | ERTS_PSFLG_ACTIVE_SYS),
					    ERTS_PSFLG_ACTIVE_SYS);

#ifdef DEBUG
	ASSERT(state & ERTS_PSFLG_DELAYED_SYS);
	qmask = c_p->sys_task_qs->qmask;
	aprio = ERTS_PSFLGS_GET_ACT_PRIO(state);
	ASSERT(ERTS_PSFLGS_GET_USR_PRIO(state) >= aprio);
	ASSERT((qmask & -qmask) >= (1 << aprio));
    }
#endif

    erts_smp_proc_unlock(c_p, ERTS_PROC_LOCK_STATUS);

    (void) ERTS_PROC_SET_DELAYED_GC_TASK_QS(c_p, ERTS_PROC_LOCK_MAIN, NULL);

    if (dgc_tsk_qs)
	proc_sys_task_queues_free(dgc_tsk_qs);

    return 1;
}

void
erts_sched_stat_modify(int what)
{
    int ix;
    switch (what) {
    case ERTS_SCHED_STAT_MODIFY_ENABLE:
	erts_smp_thr_progress_block();
	erts_sched_stat.enabled = 1;
	erts_smp_thr_progress_unblock();
	break;
    case ERTS_SCHED_STAT_MODIFY_DISABLE:
	erts_smp_thr_progress_block();
	erts_sched_stat.enabled = 0;
	erts_smp_thr_progress_unblock();
	break;
    case ERTS_SCHED_STAT_MODIFY_CLEAR:
	erts_smp_spin_lock(&erts_sched_stat.lock);
	for (ix = 0; ix < ERTS_NO_PRIO_LEVELS; ix++) {
	    erts_sched_stat.prio[ix].total_executed = 0;
	    erts_sched_stat.prio[ix].executed = 0;
	    erts_sched_stat.prio[ix].total_migrated = 0;
	    erts_sched_stat.prio[ix].migrated = 0;
	}
	erts_smp_spin_unlock(&erts_sched_stat.lock);
	break;
    }
}

Eterm
erts_sched_stat_term(Process *p, int total)
{
    Uint sz;
    Uint *hp;
    Eterm prio[ERTS_NO_PRIO_LEVELS];
    Uint executed[ERTS_NO_PRIO_LEVELS];
    Uint migrated[ERTS_NO_PRIO_LEVELS];

    erts_smp_spin_lock(&erts_sched_stat.lock);
    if (total) {
	int i;
	for (i = 0; i < ERTS_NO_PRIO_LEVELS; i++) {
	    prio[i] = erts_sched_stat.prio[i].name;
	    executed[i] = erts_sched_stat.prio[i].total_executed;
	    migrated[i] = erts_sched_stat.prio[i].total_migrated;
	}
    }
    else {
	int i;
	for (i = 0; i < ERTS_NO_PRIO_LEVELS; i++) {
	    prio[i] = erts_sched_stat.prio[i].name;
	    executed[i] = erts_sched_stat.prio[i].executed;
	    erts_sched_stat.prio[i].executed = 0;
	    migrated[i] = erts_sched_stat.prio[i].migrated;
	    erts_sched_stat.prio[i].migrated = 0;
	}
    }
    erts_smp_spin_unlock(&erts_sched_stat.lock);

    sz = 0;
    (void) erts_bld_atom_2uint_3tup_list(NULL, &sz, ERTS_NO_PRIO_LEVELS,
					 prio, executed, migrated);
    hp = HAlloc(p, sz);
    return erts_bld_atom_2uint_3tup_list(&hp, NULL, ERTS_NO_PRIO_LEVELS,
					 prio, executed, migrated);
}

/*
 * Scheduling of misc stuff
 */

void
erts_schedule_misc_op(void (*func)(void *), void *arg)
{
    ErtsSchedulerData *esdp = erts_get_scheduler_data();
    ErtsRunQueue *rq = esdp ? esdp->run_queue : ERTS_RUNQ_IX(0);
    ErtsMiscOpList *molp = misc_op_list_alloc();
#ifdef ERTS_SMP
    ErtsMigrationPaths *mpaths = erts_get_migration_paths();

    if (!mpaths)
	rq = ERTS_RUNQ_IX(0);
    else {
	ErtsRunQueue *erq = mpaths->mpath[rq->ix].misc_evac_runq;
	if (erq)
	    rq = erq;
    }
#endif

    erts_smp_runq_lock(rq);

    molp->next = NULL;
    molp->func = func;
    molp->arg = arg;
    if (rq->misc.end)
	rq->misc.end->next = molp;
    else
	rq->misc.start = molp;
    rq->misc.end = molp;

#ifdef ERTS_SMP
    non_empty_runq(rq);
#endif

    erts_smp_runq_unlock(rq);

    smp_notify_inc_runq(rq);
}

static void
exec_misc_ops(ErtsRunQueue *rq)
{
    int i;
    ErtsMiscOpList *molp = rq->misc.start;
    ErtsMiscOpList *tmp_molp = molp;

    for (i = 0; i < ERTS_MAX_MISC_OPS-1; i++) {
	if (!tmp_molp) 
	    goto mtq;
	tmp_molp = tmp_molp->next;
    }
    
    if (!tmp_molp) {
    mtq:
	rq->misc.start = NULL;
	rq->misc.end = NULL;
    }
    else {
	rq->misc.start = tmp_molp->next;
	tmp_molp->next = NULL;
	if (!rq->misc.start)
	    rq->misc.end = NULL;
    }

    erts_smp_runq_unlock(rq);

    while (molp) {
	tmp_molp = molp;
	(*molp->func)(molp->arg);
	molp = molp->next;
	misc_op_list_free(tmp_molp);
    }

    erts_smp_runq_lock(rq);
}

Uint
erts_get_total_context_switches(void)
{
    Uint res = 0;
    ERTS_ATOMIC_FOREACH_RUNQ(rq, res += rq->procs.context_switches);
    return res;
}

void
erts_get_total_reductions(Uint *redsp, Uint *diffp)
{
    Uint reds = 0;
    ERTS_ATOMIC_FOREACH_RUNQ_X(rq,

			       reds += rq->procs.reductions,

			       if (redsp) *redsp = reds;
			       if (diffp) *diffp = reds - last_reductions;
			       last_reductions = reds);
}

void
erts_get_exact_total_reductions(Process *c_p, Uint *redsp, Uint *diffp)
{
    Uint reds = erts_current_reductions(c_p, c_p);
    int ix;
    erts_smp_proc_unlock(c_p, ERTS_PROC_LOCK_MAIN);
    /*
     * Wait for other schedulers to schedule out their processes
     * and update 'reductions'.
     */
    erts_smp_thr_progress_block();
    for (reds = 0, ix = 0; ix < erts_no_run_queues; ix++)
	reds += ERTS_RUNQ_IX(ix)->procs.reductions;
    if (redsp)
	*redsp = reds;
    if (diffp)
	*diffp = reds - last_exact_reductions;
    last_exact_reductions = reds;
    erts_smp_thr_progress_unblock();
    erts_smp_proc_lock(c_p, ERTS_PROC_LOCK_MAIN);
}

void
erts_free_proc(Process *p)
{
#ifdef ERTS_SMP
    erts_proc_lock_fin(p);
#endif
    erts_free(ERTS_ALC_T_PROC, (void *) p);
}

typedef struct {
    Process *proc;
    erts_aint32_t state;
    ErtsRunQueue *run_queue;
} ErtsEarlyProcInit;

static void early_init_process_struct(void *varg, Eterm data)
{
    ErtsEarlyProcInit *arg = (ErtsEarlyProcInit *) varg;
    Process *proc = arg->proc;

    proc->common.id = make_internal_pid(data);
    erts_smp_atomic32_init_relb(&proc->state, arg->state);

#ifdef ERTS_SMP
    RUNQ_SET_RQ(&proc->run_queue, arg->run_queue);

    erts_proc_lock_init(proc); /* All locks locked */
#endif

}

/*
** Allocate process and find out where to place next process.
*/
static Process*
alloc_process(ErtsRunQueue *rq, erts_aint32_t state)
{
    ErtsEarlyProcInit init_arg;
    Process *p;

    p = erts_alloc_fnf(ERTS_ALC_T_PROC, sizeof(Process));
    if (!p)
	return NULL;

    init_arg.proc = (Process *) p;
    init_arg.run_queue = rq;
    init_arg.state = state;

    ASSERT(((char *) p) == ((char *) &p->common));

    if (!erts_ptab_new_element(&erts_proc,
			       &p->common,
			       (void *) &init_arg,
			       early_init_process_struct)) {
	erts_free(ERTS_ALC_T_PROC, p);
	return NULL;
    }

    ASSERT(internal_pid_serial(p->common.id) <= ERTS_MAX_PID_SERIAL);
    
    p->approx_started = erts_get_approx_time();
    p->rcount = 0;
    p->heap = NULL;


    ASSERT(p == (Process *) (erts_ptab_pix2intptr_nob(
				 &erts_proc,
				 internal_pid_index(p->common.id))));

    return p;
}

Eterm
erl_create_process(Process* parent, /* Parent of process (default group leader). */
		   Eterm mod,	/* Tagged atom for module. */
		   Eterm func,	/* Tagged atom for function. */
		   Eterm args,	/* Arguments for function (must be well-formed list). */
		   ErlSpawnOpts* so) /* Options for spawn. */
{
    ErtsRunQueue *rq = NULL;
    Process *p;
    Sint arity;			/* Number of arguments. */
    Uint arg_size;		/* Size of arguments. */
    Uint sz;			/* Needed words on heap. */
    Uint heap_need;		/* Size needed on heap. */
    Eterm res = THE_NON_VALUE;
    erts_aint32_t state = 0;
    erts_aint32_t prio = (erts_aint32_t) PRIORITY_NORMAL;

#ifdef ERTS_SMP
    erts_smp_proc_lock(parent, ERTS_PROC_LOCKS_ALL_MINOR);
#endif

    /*
     * Check for errors.
     */

    if (is_not_atom(mod) || is_not_atom(func) || ((arity = erts_list_length(args)) < 0)) {
	so->error_code = BADARG;
	goto error;
    }

    if (so->flags & SPO_USE_ARGS) {
	if (so->scheduler) {
	    int ix = so->scheduler-1;
	    ASSERT(0 <= ix && ix < erts_no_run_queues);
	    rq = ERTS_RUNQ_IX(ix);
	    state |= ERTS_PSFLG_BOUND;
	}
	prio = (erts_aint32_t) so->priority;
    }

    state |= (((prio & ERTS_PSFLGS_PRIO_MASK) << ERTS_PSFLGS_ACT_PRIO_OFFSET)
	      | ((prio & ERTS_PSFLGS_PRIO_MASK) << ERTS_PSFLGS_USR_PRIO_OFFSET));

    if (!rq)
	rq = erts_get_runq_proc(parent);

    p = alloc_process(rq, state); /* All proc locks are locked by this thread
				     on success */
    if (!p) {
	erts_send_error_to_logger_str(parent->group_leader,
				      "Too many processes\n");
	so->error_code = SYSTEM_LIMIT;
	goto error;
    }

#ifdef BM_COUNTERS
    processes_busy++;
#endif
    BM_COUNT(processes_spawned);

    BM_SWAP_TIMER(system,size);
    arg_size = size_object(args);
    BM_SWAP_TIMER(size,system);
    heap_need = arg_size;

    p->flags = erts_default_process_flags;

    if (so->flags & SPO_USE_ARGS) {
	p->min_heap_size  = so->min_heap_size;
	p->min_vheap_size = so->min_vheap_size;
	p->max_gen_gcs    = so->max_gen_gcs;
    } else {
	p->min_heap_size  = H_MIN_SIZE;
	p->min_vheap_size = BIN_VH_MIN_SIZE;
	p->max_gen_gcs    = (Uint16) erts_smp_atomic32_read_nob(&erts_max_gen_gcs);
    }
    p->schedule_count = 0;
    ASSERT(p->min_heap_size == erts_next_heap_size(p->min_heap_size, 0));
    
    p->initial[INITIAL_MOD] = mod;
    p->initial[INITIAL_FUN] = func;
    p->initial[INITIAL_ARI] = (Uint) arity;

    /*
     * Must initialize binary lists here before copying binaries to process.
     */
    p->off_heap.first = NULL;
    p->off_heap.overhead = 0;

    heap_need +=
	IS_CONST(parent->group_leader) ? 0 : NC_HEAP_SIZE(parent->group_leader);

    if (heap_need < p->min_heap_size) {
	sz = heap_need = p->min_heap_size;
    } else {
	sz = erts_next_heap_size(heap_need, 0);
    }

#ifdef HIPE
    hipe_init_process(&p->hipe);
#ifdef ERTS_SMP
    hipe_init_process_smp(&p->hipe_smp);
#endif
#endif
    p->heap = (Eterm *) ERTS_HEAP_ALLOC(ERTS_ALC_T_HEAP, sizeof(Eterm)*sz);
    p->old_hend = p->old_htop = p->old_heap = NULL;
    p->high_water = p->heap;
    p->gen_gcs = 0;
    p->stop = p->hend = p->heap + sz;
    p->htop = p->heap;
    p->heap_sz = sz;
    p->catches = 0;

    p->bin_vheap_sz     = p->min_vheap_size;
    p->bin_old_vheap_sz = p->min_vheap_size;
    p->bin_old_vheap    = 0;
    p->bin_vheap_mature = 0;

    p->sys_task_qs = NULL;

    /* No need to initialize p->fcalls. */

    p->current = p->initial+INITIAL_MOD;

    p->i = (BeamInstr *) beam_apply;
    p->cp = (BeamInstr *) beam_apply+1;

    p->arg_reg = p->def_arg_reg;
    p->max_arg_reg = sizeof(p->def_arg_reg)/sizeof(p->def_arg_reg[0]);
    p->arg_reg[0] = mod;
    p->arg_reg[1] = func;
    BM_STOP_TIMER(system);
    BM_MESSAGE(args,p,parent);
    BM_START_TIMER(system);
    BM_SWAP_TIMER(system,copy);
    p->arg_reg[2] = copy_struct(args, arg_size, &p->htop, &p->off_heap);
    BM_MESSAGE_COPIED(arg_size);
    BM_SWAP_TIMER(copy,system);
    p->arity = 3;

    p->fvalue = NIL;
    p->freason = EXC_NULL;
    p->ftrace = NIL;
    p->reds = 0;

#ifdef ERTS_SMP
    p->common.u.alive.ptimer = NULL;
#else
    sys_memset(&p->common.u.alive.tm, 0, sizeof(ErlTimer));
#endif

    p->common.u.alive.reg = NULL;
    ERTS_P_LINKS(p) = NULL;
    ERTS_P_MONITORS(p) = NULL;
    p->nodes_monitors = NULL;
    p->suspend_monitors = NULL;

    ASSERT(is_pid(parent->group_leader));

    if (parent->group_leader == ERTS_INVALID_PID)
	p->group_leader = p->common.id;
    else {
	/* Needs to be done after the heap has been set up */
	p->group_leader =
	    IS_CONST(parent->group_leader)
	    ? parent->group_leader
	    : STORE_NC(&p->htop, &p->off_heap, parent->group_leader);
    }

    erts_get_default_tracing(&ERTS_TRACE_FLAGS(p), &ERTS_TRACER_PROC(p));

    p->msg.first = NULL;
    p->msg.last = &p->msg.first;
    p->msg.save = &p->msg.first;
    p->msg.len = 0;
#ifdef ERTS_SMP
    p->msg_inq.first = NULL;
    p->msg_inq.last = &p->msg_inq.first;
    p->msg_inq.len = 0;
#endif
    p->u.bif_timers = NULL;
    p->mbuf = NULL;
    p->mbuf_sz = 0;
    p->psd = NULL;
    p->dictionary = NULL;
    p->seq_trace_lastcnt = 0;
    p->seq_trace_clock = 0;
    SEQ_TRACE_TOKEN(p) = NIL;
#ifdef USE_VM_PROBES
    DT_UTAG(p) = NIL;
    DT_UTAG_FLAGS(p) = 0;
#endif
    p->parent = (parent->common.id == ERTS_INVALID_PID
		 ? NIL
		 : parent->common.id);

    INIT_HOLE_CHECK(p);
#ifdef DEBUG
    p->last_old_htop = NULL;
#endif

    if (IS_TRACED(parent)) {
	if (ERTS_TRACE_FLAGS(parent) & F_TRACE_SOS) {
	    ERTS_TRACE_FLAGS(p) |= (ERTS_TRACE_FLAGS(parent) & TRACEE_FLAGS);
	    ERTS_TRACER_PROC(p) = ERTS_TRACER_PROC(parent);
	}
	if (ARE_TRACE_FLAGS_ON(parent, F_TRACE_PROCS)) {
	    trace_proc_spawn(parent, p->common.id, mod, func, args);
	}
	if (ERTS_TRACE_FLAGS(parent) & F_TRACE_SOS1) {
	    /* Overrides TRACE_CHILDREN */
	    ERTS_TRACE_FLAGS(p) |= (ERTS_TRACE_FLAGS(parent) & TRACEE_FLAGS);
	    ERTS_TRACER_PROC(p) = ERTS_TRACER_PROC(parent);
	    ERTS_TRACE_FLAGS(p) &= ~(F_TRACE_SOS1 | F_TRACE_SOS);
	    ERTS_TRACE_FLAGS(parent) &= ~(F_TRACE_SOS1 | F_TRACE_SOS);
	}
    }

    /*
     * Check if this process should be initially linked to its parent.
     */

    if (so->flags & SPO_LINK) {
#ifdef DEBUG
	int ret;
#endif
	if (IS_TRACED_FL(parent, F_TRACE_PROCS)) {
	    trace_proc(parent, parent, am_link, p->common.id);
	}

#ifdef DEBUG
	ret = erts_add_link(&ERTS_P_LINKS(parent),  LINK_PID, p->common.id);
	ASSERT(ret == 0);
	ret = erts_add_link(&ERTS_P_LINKS(p), LINK_PID, parent->common.id);
	ASSERT(ret == 0);
#else	
	erts_add_link(&ERTS_P_LINKS(parent), LINK_PID, p->common.id);
	erts_add_link(&ERTS_P_LINKS(p), LINK_PID, parent->common.id);
#endif

	if (IS_TRACED(parent)) {
	    if (ERTS_TRACE_FLAGS(parent) & (F_TRACE_SOL|F_TRACE_SOL1)) {
		ERTS_TRACE_FLAGS(p) |= (ERTS_TRACE_FLAGS(parent)&TRACEE_FLAGS);
		ERTS_TRACER_PROC(p) = ERTS_TRACER_PROC(parent); /*maybe steal*/

		if (ERTS_TRACE_FLAGS(parent) & F_TRACE_SOL1) {/*maybe override*/
		    ERTS_TRACE_FLAGS(p) &= ~(F_TRACE_SOL1 | F_TRACE_SOL);
		    ERTS_TRACE_FLAGS(parent) &= ~(F_TRACE_SOL1 | F_TRACE_SOL);
		}
	    }
	}
    }

    /*
     * Test whether this process should be initially monitored by its parent.
     */
    if (so->flags & SPO_MONITOR) {
	Eterm mref;

	mref = erts_make_ref(parent);
	erts_add_monitor(&ERTS_P_MONITORS(parent), MON_ORIGIN, mref, p->common.id, NIL);
	erts_add_monitor(&ERTS_P_MONITORS(p), MON_TARGET, mref, parent->common.id, NIL);
	so->mref = mref;
    }

#ifdef ERTS_SMP
    p->scheduler_data = NULL;
    p->suspendee = NIL;
    p->pending_suspenders = NULL;
    p->pending_exit.reason = THE_NON_VALUE;
    p->pending_exit.bp = NULL;
#endif

#if !defined(NO_FPE_SIGNALS) || defined(HIPE)
    p->fp_exception = 0;
#endif

    erts_smp_proc_unlock(p, ERTS_PROC_LOCKS_ALL);

    res = p->common.id;

    /*
     * Schedule process for execution.
     */

    schedule_process(p, state);

    VERBOSE(DEBUG_PROCESSES, ("Created a new process: %T\n",p->common.id));

#ifdef USE_VM_PROBES
    if (DTRACE_ENABLED(process_spawn)) {
        DTRACE_CHARBUF(process_name, DTRACE_TERM_BUF_SIZE);
        DTRACE_CHARBUF(mfa, DTRACE_TERM_BUF_SIZE);

        dtrace_fun_decode(p, mod, func, arity, process_name, mfa);
        DTRACE2(process_spawn, process_name, mfa);
    }
#endif

 error:

    erts_smp_proc_unlock(parent, ERTS_PROC_LOCKS_ALL_MINOR);

    return res;
}

/*
 * Initiates a pseudo process that can be used
 * for arithmetic BIFs.
 */

void erts_init_empty_process(Process *p)
{
    p->htop = NULL;
    p->stop = NULL;
    p->hend = NULL;
    p->heap = NULL;
    p->gen_gcs = 0;
    p->max_gen_gcs = 0;
    p->min_heap_size = 0;
    p->min_vheap_size = 0;
    p->rcount = 0;
    p->common.id = ERTS_INVALID_PID;
    p->reds = 0;
    ERTS_TRACER_PROC(p) = NIL;
    ERTS_TRACE_FLAGS(p) = F_INITIAL_TRACE_FLAGS;
    p->group_leader = ERTS_INVALID_PID;
    p->flags = 0;
    p->fvalue = NIL;
    p->freason = EXC_NULL;
    p->ftrace = NIL;
    p->fcalls = 0;

    p->bin_vheap_sz = BIN_VH_MIN_SIZE;
    p->bin_old_vheap_sz = BIN_VH_MIN_SIZE;
    p->bin_old_vheap = 0;
    p->sys_task_qs = NULL;
    p->bin_vheap_mature = 0;
#ifdef ERTS_SMP
    p->common.u.alive.ptimer = NULL;
#else
    memset(&(p->common.u.alive.tm), 0, sizeof(ErlTimer));
#endif
    p->next = NULL;
    p->off_heap.first = NULL;
    p->off_heap.overhead = 0;
    p->common.u.alive.reg = NULL;
    p->heap_sz = 0;
    p->high_water = NULL;
    p->old_hend = NULL;
    p->old_htop = NULL;
    p->old_heap = NULL;
    p->mbuf = NULL;
    p->mbuf_sz = 0;
    p->psd = NULL;
    ERTS_P_MONITORS(p) = NULL;
    ERTS_P_LINKS(p) = NULL;         /* List of links */
    p->nodes_monitors = NULL;
    p->suspend_monitors = NULL;
    p->msg.first = NULL;
    p->msg.last = &p->msg.first;
    p->msg.save = &p->msg.first;
    p->msg.len = 0;
    p->u.bif_timers = NULL;
    p->dictionary = NULL;
    p->seq_trace_clock = 0;
    p->seq_trace_lastcnt = 0;
    p->seq_trace_token = NIL;
    p->initial[0] = 0;
    p->initial[1] = 0;
    p->initial[2] = 0;
    p->catches = 0;
    p->cp = NULL;
    p->i = NULL;
    p->current = NULL;

    /*
     * Saved x registers.
     */
    p->arity = 0;
    p->arg_reg = NULL;
    p->max_arg_reg = 0;
    p->def_arg_reg[0] = 0;
    p->def_arg_reg[1] = 0;
    p->def_arg_reg[2] = 0;
    p->def_arg_reg[3] = 0;
    p->def_arg_reg[4] = 0;
    p->def_arg_reg[5] = 0;

    p->parent = NIL;
    p->approx_started = 0;
    p->common.u.alive.started_interval = 0;

#ifdef HIPE
    hipe_init_process(&p->hipe);
#ifdef ERTS_SMP
    hipe_init_process_smp(&p->hipe_smp);
#endif
#endif

    INIT_HOLE_CHECK(p);
#ifdef DEBUG
    p->last_old_htop = NULL;
#endif

    erts_smp_atomic32_init_nob(&p->state, (erts_aint32_t) PRIORITY_NORMAL);

#ifdef ERTS_SMP
    p->scheduler_data = NULL;
    p->msg_inq.first = NULL;
    p->msg_inq.last = &p->msg_inq.first;
    p->msg_inq.len = 0;
    p->suspendee = NIL;
    p->pending_suspenders = NULL;
    p->pending_exit.reason = THE_NON_VALUE;
    p->pending_exit.bp = NULL;
    erts_proc_lock_init(p);
    erts_smp_proc_unlock(p, ERTS_PROC_LOCKS_ALL);
    RUNQ_SET_RQ(&p->run_queue, ERTS_RUNQ_IX(0));
#endif

#if !defined(NO_FPE_SIGNALS) || defined(HIPE)
    p->fp_exception = 0;
#endif

}    

#ifdef DEBUG

void
erts_debug_verify_clean_empty_process(Process* p)
{
    /* Things that erts_cleanup_empty_process() will *not* cleanup... */
    ASSERT(p->htop == NULL);
    ASSERT(p->stop == NULL);
    ASSERT(p->hend == NULL);
    ASSERT(p->heap == NULL);
    ASSERT(p->common.id == ERTS_INVALID_PID);
    ASSERT(ERTS_TRACER_PROC(p) == NIL);
    ASSERT(ERTS_TRACE_FLAGS(p) == F_INITIAL_TRACE_FLAGS);
    ASSERT(p->group_leader == ERTS_INVALID_PID);
    ASSERT(p->next == NULL);
    ASSERT(p->common.u.alive.reg == NULL);
    ASSERT(p->heap_sz == 0);
    ASSERT(p->high_water == NULL);
    ASSERT(p->old_hend == NULL);
    ASSERT(p->old_htop == NULL);
    ASSERT(p->old_heap == NULL);

    ASSERT(ERTS_P_MONITORS(p) == NULL);
    ASSERT(ERTS_P_LINKS(p) == NULL);
    ASSERT(p->nodes_monitors == NULL);
    ASSERT(p->suspend_monitors == NULL);
    ASSERT(p->msg.first == NULL);
    ASSERT(p->msg.len == 0);
    ASSERT(p->u.bif_timers == NULL);
    ASSERT(p->dictionary == NULL);
    ASSERT(p->catches == 0);
    ASSERT(p->cp == NULL);
    ASSERT(p->i == NULL);
    ASSERT(p->current == NULL);

    ASSERT(p->parent == NIL);

#ifdef ERTS_SMP
    ASSERT(p->msg_inq.first == NULL);
    ASSERT(p->msg_inq.len == 0);
    ASSERT(p->suspendee == NIL);
    ASSERT(p->pending_suspenders == NULL);
    ASSERT(p->pending_exit.reason == THE_NON_VALUE);
    ASSERT(p->pending_exit.bp == NULL);
#endif

    /* Thing that erts_cleanup_empty_process() cleans up */

    ASSERT(p->off_heap.first == NULL);
    ASSERT(p->off_heap.overhead == 0);

    ASSERT(p->mbuf == NULL);
}

#endif

void
erts_cleanup_empty_process(Process* p)
{
    /* We only check fields that are known to be used... */

    erts_cleanup_offheap(&p->off_heap);
    p->off_heap.first = NULL;
    p->off_heap.overhead = 0;

    if (p->mbuf != NULL) {
	free_message_buffer(p->mbuf);
	p->mbuf = NULL;
    }
#ifdef ERTS_SMP
    erts_proc_lock_fin(p);
#endif
#ifdef DEBUG
    erts_debug_verify_clean_empty_process(p);
#endif
}

/*
 * p must be the currently executing process.
 */
static void
delete_process(Process* p)
{
    ErlMessage* mp;

    VERBOSE(DEBUG_PROCESSES, ("Removing process: %T\n",p->common.id));

    /* Cleanup psd */

    if (p->psd)
	erts_free(ERTS_ALC_T_PSD, p->psd);

    /* Clean binaries and funs */
    erts_cleanup_offheap(&p->off_heap);

    /*
     * The mso list should not be used anymore, but if it is, make sure that
     * we'll notice.
     */
    p->off_heap.first = (void *) 0x8DEFFACD;

    if (p->arg_reg != p->def_arg_reg) {
	erts_free(ERTS_ALC_T_ARG_REG, p->arg_reg);
    }

    /*
     * Release heaps. Clobber contents in DEBUG build.
     */


#ifdef DEBUG
    sys_memset(p->heap, DEBUG_BAD_BYTE, p->heap_sz*sizeof(Eterm));
#endif

#ifdef HIPE
    hipe_delete_process(&p->hipe);
#endif

    ERTS_HEAP_FREE(ERTS_ALC_T_HEAP, (void*) p->heap, p->heap_sz*sizeof(Eterm));
    if (p->old_heap != NULL) {

#ifdef DEBUG
	sys_memset(p->old_heap, DEBUG_BAD_BYTE,
                   (p->old_hend-p->old_heap)*sizeof(Eterm));
#endif
	ERTS_HEAP_FREE(ERTS_ALC_T_OLD_HEAP,
		       p->old_heap,
		       (p->old_hend-p->old_heap)*sizeof(Eterm));
    }

    /*
     * Free all pending message buffers.
     */
    if (p->mbuf != NULL) {	
	free_message_buffer(p->mbuf);
    }

    erts_erase_dicts(p);

    /* free all pending messages */
    mp = p->msg.first;
    while(mp != NULL) {
	ErlMessage* next_mp = mp->next;
	if (mp->data.attached) {
	    if (is_value(mp->m[0]))
		free_message_buffer(mp->data.heap_frag);
	    else {
		if (is_not_nil(mp->m[1])) {
		    ErlHeapFragment *heap_frag;
		    heap_frag = (ErlHeapFragment *) mp->data.dist_ext->ext_endp;
		    erts_cleanup_offheap(&heap_frag->off_heap);
		}
		erts_free_dist_ext_copy(mp->data.dist_ext);
	    }
	}
	free_message(mp);
	mp = next_mp;
    }

    ASSERT(!p->nodes_monitors);
    ASSERT(!p->suspend_monitors);

    p->fvalue = NIL;
}

static ERTS_INLINE void
set_proc_exiting(Process *p,
		 erts_aint32_t in_state,
		 Eterm reason,
		 ErlHeapFragment *bp)
{
    erts_aint32_t state = in_state, enq_prio = -1;
    int enqueue;
    ERTS_SMP_LC_ASSERT(erts_proc_lc_my_proc_locks(p) == ERTS_PROC_LOCKS_ALL);

    enqueue = change_proc_schedule_state(p,
					 ERTS_PSFLG_SUSPENDED|ERTS_PSFLG_PENDING_EXIT,
					 ERTS_PSFLG_EXITING|ERTS_PSFLG_ACTIVE,
					 &state,
					 &enq_prio);

    p->fvalue = reason;
    if (bp)
	erts_link_mbuf_to_proc(p, bp);
    /*
     * We used to set freason to EXC_EXIT here, but there is no need to
     * save the stack trace since this process irreversibly is going to
     * exit.
     */
    p->freason = EXTAG_EXIT;
    KILL_CATCHES(p);
    cancel_timer(p);
    p->i = (BeamInstr *) beam_exit;

    if (enqueue)
	add2runq(enqueue > 0 ? p : make_proxy_proc(NULL, p, enq_prio),
		 state,
		 enq_prio);
}

static ERTS_INLINE erts_aint32_t
set_proc_self_exiting(Process *c_p)
{
#ifdef DEBUG
    int enqueue;
#endif
    erts_aint32_t state, enq_prio = -1;

    ERTS_SMP_LC_ASSERT(erts_proc_lc_my_proc_locks(c_p) == ERTS_PROC_LOCKS_ALL);

    state = erts_smp_atomic32_read_nob(&c_p->state);
    ASSERT(state & (ERTS_PSFLG_RUNNING|ERTS_PSFLG_RUNNING_SYS));

#ifdef DEBUG
    enqueue =
#endif
	change_proc_schedule_state(c_p,
				   ERTS_PSFLG_SUSPENDED|ERTS_PSFLG_PENDING_EXIT,
				   ERTS_PSFLG_EXITING|ERTS_PSFLG_ACTIVE,
				   &state,
				   &enq_prio);

    ASSERT(!enqueue);
    return state;
}

#ifdef ERTS_SMP

void
erts_handle_pending_exit(Process *c_p, ErtsProcLocks locks)
{
    ErtsProcLocks xlocks;
    ASSERT(is_value(c_p->pending_exit.reason));
    ERTS_SMP_LC_ASSERT(erts_proc_lc_my_proc_locks(c_p) == locks);
    ERTS_SMP_LC_ASSERT(locks & ERTS_PROC_LOCK_MAIN);
    ERTS_SMP_LC_ASSERT(!((ERTS_PSFLG_EXITING|ERTS_PSFLG_FREE)
			 & erts_smp_atomic32_read_nob(&c_p->state)));

    /* Ensure that all locks on c_p are locked before proceeding... */
    if (locks == ERTS_PROC_LOCKS_ALL)
	xlocks = 0;
    else {
	xlocks = ~locks & ERTS_PROC_LOCKS_ALL;
	if (erts_smp_proc_trylock(c_p, xlocks) == EBUSY) {
	    erts_smp_proc_unlock(c_p, locks & ~ERTS_PROC_LOCK_MAIN);
	    erts_smp_proc_lock(c_p, ERTS_PROC_LOCKS_ALL_MINOR);
	}
    }

    set_proc_exiting(c_p,
		     erts_smp_atomic32_read_acqb(&c_p->state),
		     c_p->pending_exit.reason,
		     c_p->pending_exit.bp);
    c_p->pending_exit.reason = THE_NON_VALUE;
    c_p->pending_exit.bp = NULL;

    if (xlocks)
	erts_smp_proc_unlock(c_p, xlocks);
}

static void
handle_pending_exiters(ErtsProcList *pnd_xtrs)
{
    /* 'list' is expected to have been fetched (i.e. not a ring anymore) */
    ErtsProcList *plp = pnd_xtrs;

    while (plp) {
	ErtsProcList *free_plp;
	Process *p = erts_pid2proc(NULL, 0, plp->pid, ERTS_PROC_LOCKS_ALL);
	if (p) {
	    if (erts_proclist_same(plp, p)) {
		erts_aint32_t state = erts_smp_atomic32_read_acqb(&p->state);
		if (!(state & (ERTS_PSFLG_RUNNING|ERTS_PSFLG_RUNNING_SYS))) {
		    ASSERT(state & ERTS_PSFLG_PENDING_EXIT);
		    erts_handle_pending_exit(p, ERTS_PROC_LOCKS_ALL);
		}
	    }
	    erts_smp_proc_unlock(p, ERTS_PROC_LOCKS_ALL);
	}
	free_plp = plp;
	plp = plp->next;
	proclist_destroy(free_plp);
    }
}

static void
save_pending_exiter(Process *p)
{
    ErtsProcList *plp;
    ErtsRunQueue *rq;

    ERTS_SMP_LC_ASSERT(ERTS_PROC_LOCK_STATUS & erts_proc_lc_my_proc_locks(p));

    rq = erts_get_runq_current(NULL);

    plp = proclist_create(p);

    erts_smp_runq_lock(rq);

    erts_proclist_store_last(&rq->procs.pending_exiters, plp);

    non_empty_runq(rq);

    erts_smp_runq_unlock(rq);
#ifdef ERTS_DIRTY_SCHEDULERS
    if (ERTS_RUNQ_IX_IS_DIRTY(rq->ix))
	wake_dirty_schedulers(rq, 0);
    else
#endif
	wake_scheduler(rq);
}

#endif

/*
 * This function delivers an EXIT message to a process
 * which is trapping EXITs.
 */

static ERTS_INLINE void
send_exit_message(Process *to, ErtsProcLocks *to_locksp,
		  Eterm exit_term, Uint term_size, Eterm token)
{
    if (token == NIL 
#ifdef USE_VM_PROBES
	|| token == am_have_dt_utag
#endif
	) {
	Eterm* hp;
	Eterm mess;
	ErlHeapFragment* bp;
	ErlOffHeap *ohp;

	hp = erts_alloc_message_heap(term_size, &bp, &ohp, to, to_locksp);
	mess = copy_struct(exit_term, term_size, &hp, ohp);
	erts_queue_message(to, to_locksp, bp, mess, NIL
#ifdef USE_VM_PROBES
			   , NIL
#endif
			   );
    } else {
	ErlHeapFragment* bp;
	Eterm* hp;
	Eterm mess;
	Eterm temp_token;
	Uint sz_token;

	ASSERT(is_tuple(token));
	sz_token = size_object(token);
	bp = new_message_buffer(term_size+sz_token);
	hp = bp->mem;
	mess = copy_struct(exit_term, term_size, &hp, &bp->off_heap);
	/* the trace token must in this case be updated by the caller */
	seq_trace_output(token, mess, SEQ_TRACE_SEND, to->common.id, NULL);
	temp_token = copy_struct(token, sz_token, &hp, &bp->off_heap);
	erts_queue_message(to, to_locksp, bp, mess, temp_token
#ifdef USE_VM_PROBES
			   , NIL
#endif
			   );
    }
}

/*
 *
 * *** Exit signal behavior ***
 *
 * Exit signals are asynchronous (truly asynchronous in the
 * SMP emulator). When the signal is received the receiver receives an
 * 'EXIT' message if it is trapping exits; otherwise, it will either
 * ignore the signal if the exit reason is normal, or go into an
 * exiting state (ERTS_PSFLG_EXITING). When a process has gone into the
 * exiting state it will not execute any more Erlang code, but it might
 * take a while before it actually exits. The exit signal is being
 * received when the 'EXIT' message is put in the message queue, the
 * signal is dropped, or when it changes state into exiting. The time it
 * is in the exiting state before actually exiting is undefined (it
 * might take a really long time under certain conditions). The
 * receiver of the exit signal does not break links or trigger monitors
 * until it actually exits.
 *
 * Exit signals and other signals, e.g. messages, have to be received
 * by a receiver in the same order as sent by a sender.
 *
 *
 *
 * Exit signal implementation in the SMP emulator:
 *
 * If the receiver is trapping exits, the signal is transformed
 * into an 'EXIT' message and sent as a normal message, if the
 * reason is normal the signal is dropped; otherwise, the process
 * is determined to be exited. The interesting case is when the
 * process is to be exited and this is what is described below.
 *
 * If it is possible, the receiver is set in the exiting state straight
 * away and we are done; otherwise, the sender places the exit reason
 * in the pending_exit field of the process struct and if necessary
 * adds the receiver to the run queue. It is typically not possible
 * to set a scheduled process or a process which we cannot get all locks
 * on without releasing locks on it in an exiting state straight away.
 *
 * The receiver will poll the pending_exit field when it reach certain
 * places during it's execution. When it discovers the pending exit
 * it will change state into the exiting state. If the receiver wasn't
 * scheduled when the pending exit was set, the first scheduler that
 * schedules a new process will set the receiving process in the exiting
 * state just before it schedules next process.
 * 
 * When the exit signal is placed in the pending_exit field, the signal
 * is considered as being in transit on the Erlang level. The signal is
 * actually in some kind of semi transit state, since we have already
 * determined how it should be received. It will exit the process no
 * matter what if it is received (the process may exit by itself before
 * reception of the exit signal). The signal is received when it is
 * discovered in the pending_exit field by the receiver.
 *
 * The receiver have to poll the pending_exit field at least before:
 * - moving messages from the message in queue to the private message
 *   queue. This in order to preserve signal order.
 * - unlink. Otherwise the process might get exited on a link that
 *   have been removed.
 * - changing the trap_exit flag to true. This in order to simplify the
 *   implementation; otherwise, we would have to transform the signal
 *   into an 'EXIT' message when setting the trap_exit flag to true. We
 *   would also have to maintain a queue of exit signals in transit.
 * - being scheduled in or out.
 */

static ERTS_INLINE int
send_exit_signal(Process *c_p,		/* current process if and only
					   if reason is stored on it */
		 Eterm from,		/* Id of sender of signal */
		 Process *rp,		/* receiving process */
		 ErtsProcLocks *rp_locks,/* current locks on receiver */
		 Eterm reason,		/* exit reason */
		 Eterm exit_tuple,	/* Prebuild exit tuple
					   or THE_NON_VALUE */
		 Uint exit_tuple_sz,	/* Size of prebuilt exit tuple
					   (if exit_tuple != THE_NON_VALUE) */
		 Eterm token,		/* token */
		 Process *token_update, /* token updater */
		 Uint32 flags		/* flags */
    )		
{
    erts_aint32_t state = erts_smp_atomic32_read_nob(&rp->state);
    Eterm rsn = reason == am_kill ? am_killed : reason;

    ERTS_SMP_LC_ASSERT(*rp_locks == erts_proc_lc_my_proc_locks(rp));
    ERTS_SMP_LC_ASSERT((*rp_locks & ERTS_PROC_LOCKS_XSIG_SEND)
		       == ERTS_PROC_LOCKS_XSIG_SEND);

    ASSERT(reason != THE_NON_VALUE);

#ifdef USE_VM_PROBES
    if(DTRACE_ENABLED(process_exit_signal) && is_pid(from)) {
        DTRACE_CHARBUF(sender_str, DTRACE_TERM_BUF_SIZE);
        DTRACE_CHARBUF(receiver_str, DTRACE_TERM_BUF_SIZE);
        DTRACE_CHARBUF(reason_buf, DTRACE_TERM_BUF_SIZE);

        dtrace_pid_str(from, sender_str);
        dtrace_proc_str(rp, receiver_str);
        erts_snprintf(reason_buf, sizeof(DTRACE_CHARBUF_NAME(reason_buf)) - 1, "%T", reason);
        DTRACE3(process_exit_signal, sender_str, receiver_str, reason_buf);
    }
#endif

    if ((state & ERTS_PSFLG_TRAP_EXIT)
	&& (reason != am_kill || (flags & ERTS_XSIG_FLG_IGN_KILL))) {
	if (is_not_nil(token) 
#ifdef USE_VM_PROBES
	    && token != am_have_dt_utag
#endif
	    && token_update)
	    seq_trace_update_send(token_update);
	if (is_value(exit_tuple))
	    send_exit_message(rp, rp_locks, exit_tuple, exit_tuple_sz, token);
	else
	    erts_deliver_exit_message(from, rp, rp_locks, rsn, token);
	return 1; /* Receiver will get a message */
    }
    else if (reason != am_normal || (flags & ERTS_XSIG_FLG_NO_IGN_NORMAL)) {
#ifdef ERTS_SMP
	if (!(state & (ERTS_PSFLG_EXITING|ERTS_PSFLG_PENDING_EXIT))) {
	    ASSERT(!rp->pending_exit.bp);

	    if (rp == c_p && (*rp_locks & ERTS_PROC_LOCK_MAIN)) {
		/* Ensure that all locks on c_p are locked before
		   proceeding... */
		if (*rp_locks != ERTS_PROC_LOCKS_ALL) {
		    ErtsProcLocks need_locks = (~(*rp_locks)
						& ERTS_PROC_LOCKS_ALL);
		    if (erts_smp_proc_trylock(c_p, need_locks) == EBUSY) {
			erts_smp_proc_unlock(c_p,
					     *rp_locks & ~ERTS_PROC_LOCK_MAIN);
			erts_smp_proc_lock(c_p, ERTS_PROC_LOCKS_ALL_MINOR);
		    }
		    *rp_locks = ERTS_PROC_LOCKS_ALL;
		}
		set_proc_exiting(c_p, state, rsn, NULL);
	    }
	    else if (!(state & (ERTS_PSFLG_RUNNING|ERTS_PSFLG_RUNNING_SYS))) {
		/* Process not running ... */
		ErtsProcLocks need_locks = ~(*rp_locks) & ERTS_PROC_LOCKS_ALL;
		if (need_locks
		    && erts_smp_proc_trylock(rp, need_locks) == EBUSY) {
		    /* ... but we havn't got all locks on it ... */
		    save_pending_exiter(rp);
		    /*
		     * The pending exit will be discovered when next
		     * process is scheduled in
		     */
		    goto set_pending_exit;
		}
		else {
		    /* ...and we have all locks on it... */
		    *rp_locks = ERTS_PROC_LOCKS_ALL;
		    set_proc_exiting(rp,
				     state,
				     (is_immed(rsn)
				      ? rsn
				      : copy_object(rsn, rp)),
				     NULL);
		}
	    }
	    else { /* Process running... */

		/*
		 * The pending exit will be discovered when the process
		 * is scheduled out if not discovered earlier.
		 */

	    set_pending_exit:
		if (is_immed(rsn)) {
		    rp->pending_exit.reason = rsn;
		}
		else {
		    Eterm *hp;
		    Uint sz = size_object(rsn);
		    ErlHeapFragment *bp = new_message_buffer(sz);

		    hp = &bp->mem[0];
		    rp->pending_exit.reason = copy_struct(rsn,
							  sz,
							  &hp,
							  &bp->off_heap);
		    rp->pending_exit.bp = bp;
		}
		erts_smp_atomic32_read_bor_relb(&rp->state,
						ERTS_PSFLG_PENDING_EXIT);
	    }
	}
	/* else:
	 *
	 *    The receiver already has a pending exit (or is exiting)
	 *    so we drop this signal.
	 *
	 *    NOTE: dropping this exit signal is based on the assumption
	 *          that the receiver *will* exit; either on the pending
	 *          exit or by itself before seeing the pending exit.
	 */
#else /* !ERTS_SMP */
	erts_aint32_t state = erts_smp_atomic32_read_nob(&rp->state);
	if (!(state & ERTS_PSFLG_EXITING)) {
	    set_proc_exiting(rp,
			     state,
			     (is_immed(rsn) || c_p == rp
			      ? rsn
			      : copy_object(rsn, rp)),
			     NULL);
	}
#endif
	return -1; /* Receiver will exit */
    }

    return 0; /* Receiver unaffected */
}


int
erts_send_exit_signal(Process *c_p,
		      Eterm from,
		      Process *rp,
		      ErtsProcLocks *rp_locks,
		      Eterm reason,
		      Eterm token,
		      Process *token_update,
		      Uint32 flags)
{
    return send_exit_signal(c_p,
			    from,
			    rp,
			    rp_locks,
			    reason,
			    THE_NON_VALUE,
			    0,
			    token,
			    token_update,
			    flags);
}

typedef struct {
    Eterm reason;
    Process *p;
} ExitMonitorContext;

static void doit_exit_monitor(ErtsMonitor *mon, void *vpcontext)
{
    ExitMonitorContext *pcontext = vpcontext;
    DistEntry *dep;
    ErtsMonitor *rmon;
    Process *rp;

    if (mon->type == MON_ORIGIN) {
	/* We are monitoring someone else, we need to demonitor that one.. */
	if (is_atom(mon->pid)) { /* remote by name */
	    ASSERT(is_node_name_atom(mon->pid));
	    dep = erts_sysname_to_connected_dist_entry(mon->pid);
	    if (dep) {
		erts_smp_de_links_lock(dep);
		rmon = erts_remove_monitor(&(dep->monitors), mon->ref);
		erts_smp_de_links_unlock(dep);
		if (rmon) {
		    ErtsDSigData dsd;
		    int code = erts_dsig_prepare(&dsd, dep, NULL,
						 ERTS_DSP_NO_LOCK, 0);
		    if (code == ERTS_DSIG_PREP_CONNECTED) {
			code = erts_dsig_send_demonitor(&dsd,
							rmon->pid,
							mon->name,
							mon->ref,
							1);
			ASSERT(code == ERTS_DSIG_SEND_OK);
		    }
		    erts_destroy_monitor(rmon);
		}
		erts_deref_dist_entry(dep);
	    }
	} else {
	    ASSERT(is_pid(mon->pid));
	    if (is_internal_pid(mon->pid)) { /* local by pid or name */
		rp = erts_pid2proc(NULL, 0, mon->pid, ERTS_PROC_LOCK_LINK);
		if (!rp) {
		    goto done;
		}
		rmon = erts_remove_monitor(&ERTS_P_MONITORS(rp), mon->ref);
		erts_smp_proc_unlock(rp, ERTS_PROC_LOCK_LINK);
		if (rmon == NULL) {
		    goto done;
		}
		erts_destroy_monitor(rmon);
	    } else { /* remote by pid */
		ASSERT(is_external_pid(mon->pid));
		dep = external_pid_dist_entry(mon->pid);
		ASSERT(dep != NULL);
		if (dep) {
		    erts_smp_de_links_lock(dep);
		    rmon = erts_remove_monitor(&(dep->monitors), mon->ref);
		    erts_smp_de_links_unlock(dep);
		    if (rmon) {
			ErtsDSigData dsd;
			int code = erts_dsig_prepare(&dsd, dep, NULL,
						     ERTS_DSP_NO_LOCK, 0);
			if (code == ERTS_DSIG_PREP_CONNECTED) {
			    code = erts_dsig_send_demonitor(&dsd,
							    rmon->pid,
							    mon->pid,
							    mon->ref,
							    1);
			    ASSERT(code == ERTS_DSIG_SEND_OK);
			}
			erts_destroy_monitor(rmon);
		    }
		}
	    }
	}
    } else { /* type == MON_TARGET */
	ASSERT(mon->type == MON_TARGET);
	ASSERT(is_pid(mon->pid) || is_internal_port(mon->pid));
	if (is_internal_port(mon->pid)) {
	    Port *prt = erts_id2port(mon->pid);
	    if (prt == NULL) {
		goto done;
	    }
	    erts_fire_port_monitor(prt, mon->ref);
	    erts_port_release(prt); 
	} else if (is_internal_pid(mon->pid)) {/* local by name or pid */
	    Eterm watched;
	    DeclareTmpHeapNoproc(lhp,3);
	    ErtsProcLocks rp_locks = (ERTS_PROC_LOCK_LINK
				      | ERTS_PROC_LOCKS_MSG_SEND);
	    rp = erts_pid2proc(NULL, 0, mon->pid, rp_locks);
	    if (rp == NULL) {
		goto done;
	    }
	    UseTmpHeapNoproc(3);
	    rmon = erts_remove_monitor(&ERTS_P_MONITORS(rp), mon->ref);
	    if (rmon) {
		erts_destroy_monitor(rmon);
		watched = (is_atom(mon->name)
			   ? TUPLE2(lhp, mon->name, 
				    erts_this_dist_entry->sysname)
			   : pcontext->p->common.id);
		erts_queue_monitor_message(rp, &rp_locks, mon->ref, am_process, 
					   watched, pcontext->reason);
	    }
	    UnUseTmpHeapNoproc(3);
	    /* else: demonitor while we exited, i.e. do nothing... */
	    erts_smp_proc_unlock(rp, rp_locks);
	} else { /* external by pid or name */
	    ASSERT(is_external_pid(mon->pid));    
	    dep = external_pid_dist_entry(mon->pid);
	    ASSERT(dep != NULL);
	    if (dep) {
		erts_smp_de_links_lock(dep);
		rmon = erts_remove_monitor(&(dep->monitors), mon->ref);
		erts_smp_de_links_unlock(dep);
		if (rmon) {
		    ErtsDSigData dsd;
		    int code = erts_dsig_prepare(&dsd, dep, NULL,
						 ERTS_DSP_NO_LOCK, 0);
		    if (code == ERTS_DSIG_PREP_CONNECTED) {
			code = erts_dsig_send_m_exit(&dsd,
						     mon->pid,
						     (rmon->name != NIL
						      ? rmon->name
						      : rmon->pid),
						     mon->ref,
						     pcontext->reason);
			ASSERT(code == ERTS_DSIG_SEND_OK);
		    }
		    erts_destroy_monitor(rmon);
		}
	    }
	}
    }
 done:
    /* As the monitors are previously removed from the process, 
       distribution operations will not cause monitors to disappear,
       we can safely delete it. */
       
    erts_destroy_monitor(mon);
}

typedef struct {
    Process *p;
    Eterm reason;
    Eterm exit_tuple;
    Uint exit_tuple_sz;
} ExitLinkContext;

static void doit_exit_link(ErtsLink *lnk, void *vpcontext)
{
    ExitLinkContext *pcontext = vpcontext;
    /* Unpack context, it's readonly */
    Process *p = pcontext->p;
    Eterm reason = pcontext->reason;
    Eterm exit_tuple = pcontext->exit_tuple;
    Uint exit_tuple_sz = pcontext->exit_tuple_sz;
    Eterm item = lnk->pid;
    ErtsLink *rlnk;
    DistEntry *dep;
    Process *rp;

    switch(lnk->type) {
    case LINK_PID:
	if(is_internal_port(item)) {
	    Port *prt = erts_port_lookup(item, ERTS_PORT_SFLGS_INVALID_LOOKUP);
	    if (prt)
		erts_port_exit(NULL,
			       (ERTS_PORT_SIG_FLG_FORCE_SCHED
				| ERTS_PORT_SIG_FLG_BROKEN_LINK),
			       prt,
			       p->common.id,
			       reason,
			       NULL);
	}
	else if(is_external_port(item)) {
	    erts_dsprintf_buf_t *dsbufp = erts_create_logger_dsbuf();
	    erts_dsprintf(dsbufp,
			  "Erroneous link between %T and external port %T "
			  "found\n",
			  p->common.id,
			  item);
	    erts_send_error_to_logger_nogl(dsbufp);
	    ASSERT(0); /* It isn't possible to setup such a link... */
	}
	else if (is_internal_pid(item)) {
	    ErtsProcLocks rp_locks = (ERTS_PROC_LOCK_LINK
				      | ERTS_PROC_LOCKS_XSIG_SEND);
	    rp = erts_pid2proc(NULL, 0, item, rp_locks);
	    if (rp) {
		rlnk = erts_remove_link(&ERTS_P_LINKS(rp), p->common.id);
		/* If rlnk == NULL, we got unlinked while exiting,
		   i.e., do nothing... */
		if (rlnk) {
		    int xres;
		    erts_destroy_link(rlnk);
		    xres = send_exit_signal(NULL,
					    p->common.id,
					    rp,
					    &rp_locks, 
					    reason,
					    exit_tuple,
					    exit_tuple_sz,
					    SEQ_TRACE_TOKEN(p),
					    p,
					    ERTS_XSIG_FLG_IGN_KILL);
		    if (xres >= 0 && IS_TRACED_FL(rp, F_TRACE_PROCS)) {
			/* We didn't exit the process and it is traced */
			if (IS_TRACED_FL(rp, F_TRACE_PROCS)) {
			    trace_proc(p, rp, am_getting_unlinked, p->common.id);
			}
		    }
		}
		ASSERT(rp != p);
		erts_smp_proc_unlock(rp, rp_locks);
	    }
	}
	else if (is_external_pid(item)) {
	    dep = external_pid_dist_entry(item);
	    if(dep != erts_this_dist_entry) {
		ErtsDSigData dsd;
		int code;
		ErtsDistLinkData dld;
		erts_remove_dist_link(&dld, p->common.id, item, dep);
		erts_smp_proc_lock(p, ERTS_PROC_LOCK_MAIN);
		code = erts_dsig_prepare(&dsd, dep, p, ERTS_DSP_NO_LOCK, 0);
		if (code == ERTS_DSIG_PREP_CONNECTED) {
		    code = erts_dsig_send_exit_tt(&dsd, p->common.id, item,
						  reason, SEQ_TRACE_TOKEN(p));
		    ASSERT(code == ERTS_DSIG_SEND_OK);
		}
		erts_smp_proc_unlock(p, ERTS_PROC_LOCK_MAIN);
		erts_destroy_dist_link(&dld);
	    }
	}
	break;
    case LINK_NODE:
	ASSERT(is_node_name_atom(item));
	dep = erts_sysname_to_connected_dist_entry(item);
	if(dep) {
	    /* dist entries have node links in a separate structure to 
	       avoid confusion */
	    erts_smp_de_links_lock(dep);
	    rlnk = erts_remove_link(&(dep->node_links), p->common.id);
	    erts_smp_de_links_unlock(dep);
	    if (rlnk)
		erts_destroy_link(rlnk);
	    erts_deref_dist_entry(dep);
	}
	break;
	
    default:
	erl_exit(1, "bad type in link list\n");
	break;
    }
    erts_destroy_link(lnk);
}

static void
resume_suspend_monitor(ErtsSuspendMonitor *smon, void *vc_p)
{
    Process *suspendee = erts_pid2proc((Process *) vc_p, ERTS_PROC_LOCK_MAIN,
				       smon->pid, ERTS_PROC_LOCK_STATUS);
    if (suspendee) {
	if (smon->active)
	    resume_process(suspendee);
	erts_smp_proc_unlock(suspendee, ERTS_PROC_LOCK_STATUS);
    }
    erts_destroy_suspend_monitor(smon);
}

/* this function fishishes a process and propagates exit messages - called
   by process_main when a process dies */
void 
erts_do_exit_process(Process* p, Eterm reason)
{
    p->arity = 0;		/* No live registers */
    p->fvalue = reason;

#ifdef USE_VM_PROBES
    if (DTRACE_ENABLED(process_exit)) {
        DTRACE_CHARBUF(process_buf, DTRACE_TERM_BUF_SIZE);
        DTRACE_CHARBUF(reason_buf, DTRACE_TERM_BUF_SIZE);

        dtrace_proc_str(p, process_buf);
        erts_snprintf(reason_buf, DTRACE_TERM_BUF_SIZE - 1, "%T", reason);
        DTRACE2(process_exit, process_buf, reason_buf);
    }
#endif

#ifdef ERTS_SMP
    ERTS_SMP_CHK_HAVE_ONLY_MAIN_PROC_LOCK(p);
    /* By locking all locks (main lock is already locked) when going
       to exiting state (ERTS_PSFLG_EXITING), it is enough to take any lock when
       looking up a process (erts_pid2proc()) to prevent the looked up
       process from exiting until the lock has been released. */
    erts_smp_proc_lock(p, ERTS_PROC_LOCKS_ALL_MINOR);
#endif

#ifndef ERTS_SMP
    set_proc_self_exiting(p);
#else
    if (ERTS_PSFLG_PENDING_EXIT & set_proc_self_exiting(p)) {
	/* Process exited before pending exit was received... */
	p->pending_exit.reason = THE_NON_VALUE;
	if (p->pending_exit.bp) {
	    free_message_buffer(p->pending_exit.bp);
	    p->pending_exit.bp = NULL;
	}
    }

    cancel_suspend_of_suspendee(p, ERTS_PROC_LOCKS_ALL); 

    ERTS_SMP_MSGQ_MV_INQ2PRIVQ(p);
#endif

    if (IS_TRACED(p)) {
	if (IS_TRACED_FL(p, F_TRACE_CALLS))
	    erts_schedule_time_break(p, ERTS_BP_CALL_TIME_SCHEDULE_EXITING);

	if (IS_TRACED_FL(p,F_TRACE_PROCS))
	    trace_proc(p, p, am_exit, reason);
    }

    erts_trace_check_exiting(p->common.id);

    ASSERT((ERTS_TRACE_FLAGS(p) & F_INITIAL_TRACE_FLAGS)
	   == F_INITIAL_TRACE_FLAGS);

    cancel_timer(p);		/* Always cancel timer just in case */

    if (p->u.bif_timers)
	erts_cancel_bif_timers(p, ERTS_PROC_LOCKS_ALL);

    erts_smp_proc_unlock(p, ERTS_PROC_LOCKS_ALL_MINOR);

    /*
     * The p->u.bif_timers of this process can *not* be used anymore;
     * will be overwritten by misc termination data.
     */
    p->u.terminate = NULL;


    erts_continue_exit_process(p);
}

void
erts_continue_exit_process(Process *p)
{
    ErtsLink* lnk;
    ErtsMonitor *mon;
    ErtsProcLocks curr_locks = ERTS_PROC_LOCK_MAIN;
    Eterm reason = p->fvalue;
    DistEntry *dep;
    struct saved_calls *scb;
    process_breakpoint_time_t *pbt;
    erts_aint32_t state;
    void *nif_export;

#ifdef DEBUG
    int yield_allowed = 1;
#endif

    ERTS_SMP_LC_ASSERT(ERTS_PROC_LOCK_MAIN == erts_proc_lc_my_proc_locks(p));

    ASSERT(ERTS_PROC_IS_EXITING(p));

#ifdef ERTS_SMP
    if (p->flags & F_HAVE_BLCKD_MSCHED) {
	ErtsSchedSuspendResult ssr;
	ssr = erts_block_multi_scheduling(p, ERTS_PROC_LOCK_MAIN, 0, 1);
	switch (ssr) {
	case ERTS_SCHDLR_SSPND_YIELD_RESTART:
	    goto yield;
	case ERTS_SCHDLR_SSPND_DONE_MSCHED_BLOCKED:
	case ERTS_SCHDLR_SSPND_YIELD_DONE_MSCHED_BLOCKED:
	case ERTS_SCHDLR_SSPND_DONE:
	case ERTS_SCHDLR_SSPND_YIELD_DONE:
	    p->flags &= ~F_HAVE_BLCKD_MSCHED;
	    break;
	case ERTS_SCHDLR_SSPND_EINVAL:
	default:
	    erl_exit(ERTS_ABORT_EXIT, "%s:%d: Internal error: %d\n",
		     __FILE__, __LINE__, (int) ssr);
	}
    }
#endif

    if (p->flags & F_USING_DB) {
	if (erts_db_process_exiting(p, ERTS_PROC_LOCK_MAIN))
	    goto yield;
	p->flags &= ~F_USING_DB;
    }

    erts_set_gc_state(p, 1);
    state = erts_smp_atomic32_read_acqb(&p->state);
    if (state & ERTS_PSFLG_ACTIVE_SYS) {
	if (cleanup_sys_tasks(p, state, CONTEXT_REDS) >= CONTEXT_REDS/2)
	    goto yield;
    }

    if (p->flags & F_USING_DDLL) {
	erts_ddll_proc_dead(p, ERTS_PROC_LOCK_MAIN);
	p->flags &= ~F_USING_DDLL;
    }

    if (p->nodes_monitors) {
	erts_delete_nodes_monitors(p, ERTS_PROC_LOCK_MAIN);
	p->nodes_monitors = NULL;
    }
	

    if (p->suspend_monitors) {
	erts_sweep_suspend_monitors(p->suspend_monitors,
				    resume_suspend_monitor,
				    p);
	p->suspend_monitors = NULL;
    }

    /*
     * The registered name *should* be the last "erlang resource" to
     * cleanup.
     */
    if (p->common.u.alive.reg) {
	(void) erts_unregister_name(p, ERTS_PROC_LOCK_MAIN, NULL, THE_NON_VALUE);
	ASSERT(!p->common.u.alive.reg);
    }

    erts_smp_proc_lock(p, ERTS_PROC_LOCKS_ALL_MINOR);
    curr_locks = ERTS_PROC_LOCKS_ALL;

    /*
     * From this point on we are no longer allowed to yield
     * this process.
     */
#ifdef DEBUG
    yield_allowed = 0;
#endif

    /*
     * Note! The monitor and link fields will be overwritten 
     * by erts_ptab_delete_element() below.
     */
    mon = ERTS_P_MONITORS(p);
    lnk = ERTS_P_LINKS(p);

    {
	/* Do *not* use erts_get_runq_proc() */
	ErtsRunQueue *rq;
	rq = erts_get_runq_current(ERTS_GET_SCHEDULER_DATA_FROM_PROC(p));

	erts_smp_runq_lock(rq);

#ifdef ERTS_SMP
	ASSERT(p->scheduler_data);
	ASSERT(p->scheduler_data->current_process == p);
	ASSERT(p->scheduler_data->free_process == NULL);

	p->scheduler_data->current_process = NULL;
	p->scheduler_data->free_process = p;
#endif

	/* Time of death! */
	erts_ptab_delete_element(&erts_proc, &p->common);

	erts_smp_runq_unlock(rq);
    }

    /*
     * All "erlang resources" have to be deallocated before this point,
     * e.g. registered name, so monitoring and linked processes can
     * be sure that all interesting resources have been deallocated
     * when the monitors and/or links hit.
     */

    {
	/* Inactivate and notify free */
	erts_aint32_t n, e, a = erts_smp_atomic32_read_nob(&p->state);
#ifdef ERTS_SMP
	int refc_inced = 0;
#endif
	while (1) {
	    n = e = a;
	    ASSERT(a & ERTS_PSFLG_EXITING);
	    n |= ERTS_PSFLG_FREE;
	    n &= ~ERTS_PSFLG_ACTIVE;
#ifdef ERTS_SMP
	    if ((n & ERTS_PSFLG_IN_RUNQ) && !refc_inced) {
		erts_smp_proc_inc_refc(p);
		refc_inced = 1;
	    }
#endif
	    a = erts_smp_atomic32_cmpxchg_mb(&p->state, n, e);
	    if (a == e)
		break;
	}

#ifdef ERTS_SMP
	if (refc_inced && !(n & ERTS_PSFLG_IN_RUNQ))
	    erts_smp_proc_dec_refc(p);
#endif
    }
    
    dep = ((p->flags & F_DISTRIBUTION)
	   ? ERTS_PROC_SET_DIST_ENTRY(p, ERTS_PROC_LOCKS_ALL, NULL)
	   : NULL);
    scb = ERTS_PROC_SET_SAVED_CALLS_BUF(p, ERTS_PROC_LOCKS_ALL, NULL);
    pbt = ERTS_PROC_SET_CALL_TIME(p, ERTS_PROC_LOCKS_ALL, NULL);
    nif_export = ERTS_PROC_SET_NIF_TRAP_EXPORT(p, ERTS_PROC_LOCKS_ALL, NULL);

    erts_smp_proc_unlock(p, ERTS_PROC_LOCKS_ALL);
#ifdef BM_COUNTERS
    processes_busy--;
#endif

    if (dep) {
	erts_do_net_exits(dep, reason);
	if(dep)
	    erts_deref_dist_entry(dep);
    }

    /*
     * Pre-build the EXIT tuple if there are any links.
     */
    if (lnk) {
	DeclareTmpHeap(tmp_heap,4,p);
	Eterm exit_tuple;
	Uint exit_tuple_sz;
	Eterm* hp;

	UseTmpHeap(4,p);
	hp = &tmp_heap[0];

	exit_tuple = TUPLE3(hp, am_EXIT, p->common.id, reason);

	exit_tuple_sz = size_object(exit_tuple);

	{
	    ExitLinkContext context = {p, reason, exit_tuple, exit_tuple_sz};
	    erts_sweep_links(lnk, &doit_exit_link, &context);
	}
	UnUseTmpHeap(4,p);
    }

    {
	ExitMonitorContext context = {reason, p};
	erts_sweep_monitors(mon,&doit_exit_monitor,&context); /* Allocates TmpHeap, but we
								 have none here */
    }

    if (scb)
        erts_free(ERTS_ALC_T_CALLS_BUF, (void *) scb);

    if (pbt)
        erts_free(ERTS_ALC_T_BPD, (void *) pbt);

    if (nif_export)
	erts_destroy_nif_export(nif_export);

    delete_process(p);

#ifdef ERTS_SMP
    erts_smp_proc_lock(p, ERTS_PROC_LOCK_MAIN);
    ERTS_SMP_CHK_HAVE_ONLY_MAIN_PROC_LOCK(p);
#endif

    return;

 yield:

#ifdef DEBUG
    ASSERT(yield_allowed);
#endif

    ERTS_SMP_LC_ASSERT(curr_locks == erts_proc_lc_my_proc_locks(p));
    ERTS_SMP_LC_ASSERT(ERTS_PROC_LOCK_MAIN & curr_locks);

    p->i = (BeamInstr *) beam_continue_exit;

    if (!(curr_locks & ERTS_PROC_LOCK_STATUS)) {
	erts_smp_proc_lock(p, ERTS_PROC_LOCK_STATUS);
	curr_locks |= ERTS_PROC_LOCK_STATUS;
    }

    if (curr_locks != ERTS_PROC_LOCK_MAIN)
	erts_smp_proc_unlock(p, ~ERTS_PROC_LOCK_MAIN & curr_locks);

    ERTS_SMP_LC_ASSERT(ERTS_PROC_LOCK_MAIN == erts_proc_lc_my_proc_locks(p));

}

/* Callback for process timeout */
static void
timeout_proc(Process* p)
{
    erts_aint32_t state;
    BeamInstr** pi = (BeamInstr **) p->def_arg_reg;
    p->i = *pi;
    p->flags |= F_TIMO;
    p->flags &= ~F_INSLPQUEUE;

    state = erts_smp_atomic32_read_acqb(&p->state);
    if (!(state & ERTS_PSFLG_ACTIVE))
	schedule_process(p, state);
}


void
cancel_timer(Process* p)
{
    ERTS_SMP_LC_ASSERT(ERTS_PROC_LOCK_MAIN & erts_proc_lc_my_proc_locks(p));
    p->flags &= ~(F_INSLPQUEUE|F_TIMO);
#ifdef ERTS_SMP
    erts_cancel_smp_ptimer(p->common.u.alive.ptimer);
#else
    erts_cancel_timer(&p->common.u.alive.tm);
#endif
}

/*
 * Insert a process into the time queue, with a timeout 'timeout' in ms.
 */
void
set_timer(Process* p, Uint timeout)
{
    ERTS_SMP_LC_ASSERT(ERTS_PROC_LOCK_MAIN & erts_proc_lc_my_proc_locks(p));

    /* check for special case timeout=0 DONT ADD TO time queue */
    if (timeout == 0) {
	p->flags |= F_TIMO;
	return;
    }
    p->flags |= F_INSLPQUEUE;
    p->flags &= ~F_TIMO;

#ifdef ERTS_SMP
    erts_create_smp_ptimer(&p->common.u.alive.ptimer,
			   p->common.id,
			   (ErlTimeoutProc) timeout_proc,
			   timeout);
#else
    erts_set_timer(&p->common.u.alive.tm,
		  (ErlTimeoutProc) timeout_proc,
		  NULL,
		  (void*) p,
		  timeout);
#endif
}

/*
 * Stack dump functions follow.
 */

void
erts_stack_dump(int to, void *to_arg, Process *p)
{
    Eterm* sp;
    int yreg = -1;

    if (ERTS_TRACE_FLAGS(p) & F_SENSITIVE) {
	return;
    }
    erts_program_counter_info(to, to_arg, p);
    for (sp = p->stop; sp < STACK_START(p); sp++) {
        yreg = stack_element_dump(to, to_arg, p, sp, yreg);
    }
}

void
erts_program_counter_info(int to, void *to_arg, Process *p)
{
    erts_aint32_t state;
    int i;

    erts_print(to, to_arg, "Program counter: %p (", p->i);
    print_function_from_pc(to, to_arg, p->i);
    erts_print(to, to_arg, ")\n");
    erts_print(to, to_arg, "CP: %p (", p->cp);
    print_function_from_pc(to, to_arg, p->cp);
    erts_print(to, to_arg, ")\n");
    state = erts_smp_atomic32_read_acqb(&p->state);
    if (!(state & (ERTS_PSFLG_RUNNING
		   | ERTS_PSFLG_RUNNING_SYS
		   | ERTS_PSFLG_GC))) {
        erts_print(to, to_arg, "arity = %d\n",p->arity);
	if (!ERTS_IS_CRASH_DUMPING) {
	    /*
	     * Only print the arguments if we are not writing a
	     * crash dump file. The arguments cannot be interpreted
	     * by the crashdump_viewer application and will therefore
	     * only cause problems.
	     */
	    for (i = 0; i < p->arity; i++)
		erts_print(to, to_arg, "   %.*T\n", INT_MAX, p->arg_reg[i]);
	}
    }
}

static void
print_function_from_pc(int to, void *to_arg, BeamInstr* x)
{
    BeamInstr* addr = find_function_from_pc(x);
    if (addr == NULL) {
        if (x == beam_exit) {
            erts_print(to, to_arg, "<terminate process>");
        } else if (x == beam_continue_exit) {
            erts_print(to, to_arg, "<continue terminate process>");
        } else if (x == beam_apply+1) {
            erts_print(to, to_arg, "<terminate process normally>");
	} else if (x == 0) {
            erts_print(to, to_arg, "invalid");
        } else {
            erts_print(to, to_arg, "unknown function");
        }
    } else {
	erts_print(to, to_arg, "%T:%T/%d + %d",
		   addr[0], addr[1], addr[2], ((x-addr)-2) * sizeof(Eterm));
    }
}

static int
stack_element_dump(int to, void *to_arg, Process* p, Eterm* sp, int yreg)
{
    Eterm x = *sp;

    if (yreg < 0 || is_CP(x)) {
        erts_print(to, to_arg, "\n%p ", sp);
    } else {
        char sbuf[16];
        erts_snprintf(sbuf, sizeof(sbuf), "y(%d)", yreg);
        erts_print(to, to_arg, "%-8s ", sbuf);
        yreg++;
    }

    if (is_CP(x)) {
        erts_print(to, to_arg, "Return addr %p (", (Eterm *) EXPAND_POINTER(x));
        print_function_from_pc(to, to_arg, cp_val(x));
        erts_print(to, to_arg, ")\n");
        yreg = 0;
    } else if is_catch(x) {
        erts_print(to, to_arg, "Catch %p (", catch_pc(x));
        print_function_from_pc(to, to_arg, catch_pc(x));
        erts_print(to, to_arg, ")\n");
    } else {
	erts_print(to, to_arg, "%T\n", x);
    }
    return yreg;
}

/*
 * A nice system halt closing all open port goes as follows:
 * 1) This function schedules the aux work ERTS_SSI_AUX_WORK_REAP_PORTS
 *    on all schedulers, then schedules itself out.
 * 2) All shedulers detect this and set the flag halt_in_progress
 *    on their run queue. The last scheduler sets all non-closed ports
 *    ERTS_PORT_SFLG_HALT. Global atomic erts_halt_progress is used
 *    as refcount to determine which is last.
 * 3) While the run ques has flag halt_in_progress no processes
 *    will be scheduled, only ports.
 * 4) When the last port closes that scheduler calls erlang:halt/1.
 *    The same global atomic is used as refcount.
 *
 * A BIF that calls this should make sure to schedule out to never come back:
 *    erl_halt((int)(- code));
 *    ERTS_BIF_YIELD1(bif_export[BIF_erlang_halt_1], BIF_P, NIL);
 */
void erl_halt(int code)
{
    if (-1 == erts_smp_atomic32_cmpxchg_acqb(&erts_halt_progress,
					     erts_no_schedulers,
					     -1)) {
#ifdef ERTS_DIRTY_SCHEDULERS
	ERTS_DIRTY_CPU_RUNQ->halt_in_progress = 1;
	ERTS_DIRTY_IO_RUNQ->halt_in_progress = 1;
#endif
	erts_halt_code = code;
	notify_reap_ports_relb();
    }
}

#if defined(ERTS_SMP) && defined(ERTS_ENABLE_LOCK_CHECK)
int
erts_dbg_check_halloc_lock(Process *p)
{
    if (ERTS_PROC_LOCK_MAIN & erts_proc_lc_my_proc_locks(p))
	return 1;
    if (p->common.id == ERTS_INVALID_PID)
	return 1;
    if (p->scheduler_data && p == p->scheduler_data->match_pseudo_process)
	return 1;
    if (erts_thr_progress_is_blocking())
	return 1;
    return 0;
}
#endif