base/sched/api.c

Go to the documentation of this file.

/** 
 * @ingroup lxrt
 * @file
 * Common scheduling function 
 * @author Paolo Mantegazza
 *
 * This file is part of the RTAI project.
 *
 * @note Copyright &copy; 1999-2003 Paolo Mantegazza <mantegazza@aero.polimi.it>
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public License as
 * published by the Free Software Foundation; either version 2 of the
 * License, or (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
 */

#include <rtai_schedcore.h>
#include <rtai_registry.h>
#include <linux/module.h>

/* ++++++++++++++++++++++++ COMMON FUNCTIONALITIES ++++++++++++++++++++++++++ */

/* +++++++++++++++++++++++++ PRIORITY MANAGEMENT ++++++++++++++++++++++++++++ */

void rt_set_sched_policy(RT_TASK *task, int policy, int rr_quantum_ns)
{
    if ((task->policy = policy ? 1 : 0)) {
        task->rr_quantum = nano2count_cpuid(rr_quantum_ns, task->runnable_on_cpus);
        if ((task->rr_quantum & 0xF0000000) || !task->rr_quantum) {
#ifdef CONFIG_SMP
            task->rr_quantum = rt_smp_times[task->runnable_on_cpus].linux_tick;
#else
            task->rr_quantum = rt_times.linux_tick;
#endif
        }
        task->rr_remaining = task->rr_quantum;
        task->yield_time = 0;
    }
}


/**
 * @anchor rt_get_prio
 * @brief Check a task priority.
 * 
 * rt_get_prio returns the base priority of task @e task.
 *
 * Recall that a task has a base native priority, assigned at its
 * birth or by @ref rt_change_prio(), and an actual, inherited,
 * priority. They can be different because of priority inheritance.
 *
 * @param task is the affected task.
 *
 * @return rt_get_prio returns the priority of task @e task.
 *
 * @note To be used only with RTAI24.x.xx.
 */
int rt_get_prio(RT_TASK *task)
{
    if (task->magic != RT_TASK_MAGIC) {
        return -EINVAL;
    }
    return task->base_priority;
}


/**
 * @anchor rt_get_inher_prio
 * @brief Check a task priority.
 * 
 * rt_get_prio returns the base priority task @e task has inherited
 * from other tasks, either blocked on resources owned by or waiting
 * to pass a message to task @e task.
 *
 * Recall that a task has a base native priority, assigned at its
 * birth or by @ref rt_change_prio(), and an actual, inherited,
 * priority. They can be different because of priority inheritance.
 *
 * @param task is the affected task.
 *
 * @return rt_get_inher_prio returns the priority of task @e task.
 *
 * @note To be used only with RTAI24.x.xx.
 */
int rt_get_inher_prio(RT_TASK *task)
{
    if (task->magic != RT_TASK_MAGIC) {
        return -EINVAL;
    }
    return task->base_priority;
}


/**
 * @anchor rt_change_prio
 * @brief Change a task priority.
 * 
 * rt_change_prio changes the base priority of task @e task to @e
 * prio. 
 *
 * Recall that a task has a base native priority, assigned at its
 * birth or by @ref rt_change_prio(), and an actual, inherited,
 * priority. They can be different because of priority inheritance.
 *
 * @param task is the affected task.
 *
 * @param priority is the new priority, it can range within 0 < prio < 
 * RT_SCHED_LOWEST_PRIORITY. 
 *
 * @return rt_change_prio returns the base priority task @e task had
 * before the change.
 *
 * @note To be used only with RTAI24.x.xx (FIXME).
 */
int rt_change_prio(RT_TASK *task, int priority)
{
    unsigned long flags;
    int prio;

    if (task->magic != RT_TASK_MAGIC || priority < 0) {
        return -EINVAL;
    }

    prio = task->base_priority;
    flags = rt_global_save_flags_and_cli();
    if ((task->base_priority = priority) < task->priority) {
        unsigned long schedmap;
        QUEUE *q;
        schedmap = 0;
        do {
            task->priority = priority;
            if (task->state == RT_SCHED_READY) {
                (task->rprev)->rnext = task->rnext;
                (task->rnext)->rprev = task->rprev;
                enq_ready_task(task);
#ifdef CONFIG_SMP
                set_bit(task->runnable_on_cpus & 0x1F, &schedmap);
#else
                schedmap = 1;
#endif
            } else if ((q = task->blocked_on) && !((task->state & RT_SCHED_SEMAPHORE) && ((SEM *)q)->qtype)) {
                (task->queue.prev)->next = task->queue.next;
                (task->queue.next)->prev = task->queue.prev;
                while ((q = q->next) != task->blocked_on && (q->task)->priority <= priority);
                q->prev = (task->queue.prev = q->prev)->next  = &(task->queue);
                task->queue.next = q;
#ifdef CONFIG_SMP
                set_bit(task->runnable_on_cpus & 0x1F, &schedmap);
#else
                schedmap = 1;
#endif
            }
        } while ((task = task->prio_passed_to) && task->priority > priority);
        if (schedmap) {
#ifdef CONFIG_SMP
            if (test_and_clear_bit(rtai_cpuid(), &schedmap)) {
                RT_SCHEDULE_MAP_BOTH(schedmap);
            } else {
                RT_SCHEDULE_MAP(schedmap);
            }
#else
            rt_schedule();
#endif
        }
    }
    rt_global_restore_flags(flags);
    return prio;
}

/* +++++++++++++++++++++ TASK RELATED SCHEDULER SERVICES ++++++++++++++++++++ */


/**
 * @anchor rt_whoami
 * @brief Get the task pointer of the current task.
 *
 * Calling rt_whoami from a task can get a pointer to its own task
 * structure.
 * 
 * @return The pointer to the current task.
 */
RT_TASK *rt_whoami(void)
{
    return _rt_whoami();
}



/**
 * @anchor rt_task_yield
 * Yield the current task.
 *
 * @ref rt_task_yield() stops the current task and takes it at the end
 * of the list of ready tasks having its same priority. The scheduler
 * makes the next ready task of the same priority active.
 *
 * Recall that RTAI schedulers allow only higher priority tasks to
 * preempt the execution of lower priority ones. So equal priority
 * tasks cannot preempt each other and @ref rt_task_yield() should be
 * used if a user needs a cooperative time slicing among equal
 * priority tasks. The implementation of the related policy is wholly
 * in the hand of the user. It is believed that time slicing is too
 * much an overhead for the most demanding real time applications, so
 * it is left up to you.
 */
void rt_task_yield(void)
{
    RT_TASK *rt_current, *task;
    unsigned long flags;

    flags = rt_global_save_flags_and_cli();
    task = (rt_current = RT_CURRENT)->rnext;
    while (rt_current->priority == task->priority) {
        task = task->rnext;
    }
    if (task != rt_current->rnext) {
        (rt_current->rprev)->rnext = rt_current->rnext;
        (rt_current->rnext)->rprev = rt_current->rprev;
        task->rprev = (rt_current->rprev = task->rprev)->rnext = rt_current;
        rt_current->rnext = task;
        rt_schedule();
    }
    rt_global_restore_flags(flags);
}



/**
 * @anchor rt_task_suspend
 * rt_task_suspend suspends execution of the task task.
 *
 * It will not be executed until a call to @ref rt_task_resume() or
 * @ref rt_task_make_periodic() is made. No account is made for
 * multiple suspends, i.e. a multiply suspended task is made ready as
 * soon as it is rt_task_resumed, thus immediately resuming its
 * execution if it is the highest in priority.
 *
 * @param task pointer to a task structure.
 *
 * @return 0 on success. A negative value on failure as described below:
 * - @b EINVAL: task does not refer to a valid task.
 *
 * @note the new RTAI 24.1.xx (FIXME) development releases take into
 * account multiple suspend and require as many @ref rt_task_resume()
 * as the rt_task_suspends placed on a task.
 */
int rt_task_suspend(RT_TASK *task)
{
    unsigned long flags;

    if (!task) {
        task = RT_CURRENT;
    } else if (task->magic != RT_TASK_MAGIC) {
        return -EINVAL;
    }

    flags = rt_global_save_flags_and_cli();
    if (!task->owndres) {
        if (!task->suspdepth++) {
            rem_ready_task(task);
            rem_timed_task(task);
            task->state |= RT_SCHED_SUSPENDED;
            if (task == RT_CURRENT) {
                rt_schedule();
            }
        }
    } else if (task->suspdepth < 0) {
        task->suspdepth++;
    }
    rt_global_restore_flags(flags);
    return task->suspdepth;
}


int rt_task_suspend_if(RT_TASK *task)
{
    unsigned long flags;

    if (!task) {
        task = RT_CURRENT;
    } else if (task->magic != RT_TASK_MAGIC) {
        return -EINVAL;
    }

    flags = rt_global_save_flags_and_cli();
    if (!task->owndres && task->suspdepth < 0) {
        task->suspdepth++;
    }
    rt_global_restore_flags(flags);
    return task->suspdepth;
}


int rt_task_suspend_until(RT_TASK *task, RTIME time)
{
    unsigned long flags;

    if (!task) {
        task = RT_CURRENT;
    } else if (task->magic != RT_TASK_MAGIC) {
        return -EINVAL;
    }

    flags = rt_global_save_flags_and_cli();
    if (!task->owndres) {
        if (!task->suspdepth) {
#ifdef CONFIG_SMP
            int cpuid = rtai_cpuid();
#endif
            if ((task->resume_time = time) > rt_time_h) {
                task->suspdepth++;
                rem_ready_task(task);
                enq_timed_task(task);
                task->state |= (RT_SCHED_SUSPENDED | RT_SCHED_DELAYED);
                task->blocked_on = (void *)task;
                if (task == RT_CURRENT) {
                    rt_schedule();
                    if (task->blocked_on) {
                        task->suspdepth--;
                        rt_global_restore_flags(flags);
                        return SEM_TIMOUT;
                    }
                }
            }
        } else {
            task->suspdepth++;
        }
    } else if (task->suspdepth < 0) {
        task->suspdepth++;
    }
    rt_global_restore_flags(flags);
    return task->suspdepth;
}


int rt_task_suspend_timed(RT_TASK *task, RTIME delay)
{
    return rt_task_suspend_until(task, get_time() + delay);
}


/**
 * @anchor rt_task_resume
 * Resume a task.
 *
 * rt_task_resume resumes execution of the task @e task previously
 * suspended by @ref rt_task_suspend(), or makes a newly created task
 * ready to run, if it makes the task ready. Since no account is made
 * for multiple suspend rt_task_resume unconditionally resumes any
 * task it makes ready.
 *
 * @param task pointer to a task structure.
 *
 * @return 0 on success. A negative value on failure as described below:
 * - @b EINVAL: task does not refer to a valid task.
 *
 * @note the new RTAI 24.1.xx (FIXME) development releases take into
 *       account multiple suspend and require as many rt_task_resumes
 *   as the rt_task_suspends placed on a task.
 */
int rt_task_resume(RT_TASK *task)
{
    unsigned long flags;

    if (task->magic != RT_TASK_MAGIC) {
        return -EINVAL;
    }

    flags = rt_global_save_flags_and_cli();
    if (!(--task->suspdepth)) {
        rem_timed_task(task);
        if ((task->state &= ~(RT_SCHED_SUSPENDED | RT_SCHED_DELAYED)) == RT_SCHED_READY) {
            task->blocked_on = NOTHING;
            enq_ready_task(task);
            RT_SCHEDULE(task, rtai_cpuid());
        }
    }
    rt_global_restore_flags(flags);
    return 0;
}


/**
 * @anchor rt_get_task_state
 * Query task state
 *
 * rt_get_task_state returns the state of a real time task.
 *
 * @param task is a pointer to the task structure.
 *
 * Task state is formed by the bitwise OR of one or more of the
 * following flags:
 *
 * @retval READY Task @e task is ready to run (i.e. unblocked).
 * Note that on a UniProcessor machine the currently running task is
 * just in READY state, while on MultiProcessors can be (READY |
 * RUNNING), see below. 
 * @retval SUSPENDED Task @e task blocked waiting for a resume.
 * @retval DELAYED Task @e task blocked waiting for its next running
 * period or expiration of a timeout.
 * @retval SEMAPHORE Task @e task blocked on a semaphore, waiting for
 * the semaphore to be signaled.
 * @retval SEND Task @e task blocked on sending a message, receiver
 * was not in RECEIVE state.
 * @retval RECEIVE Task @e task blocked waiting for incoming messages,
 * sends or rpcs. 
 * @retval RPC Task @e task blocked on a Remote Procedure Call,
 * receiver was not in RECEIVE state.
 * @retval RETURN Task @e task blocked waiting for a return from a
 * Remote Procedure Call, receiver got the RPC but has not replied
 * yet. 
 * @retval RUNNING Task @e task is running, used only for SMP
 * schedulers. 
 *
 * The returned task state is just an approximate information. Timer
 * and other hardware interrupts may cause a change in the state of
 * the queried task before the caller could evaluate the returned
 * value. Caller should disable interrupts if it wants reliable info
 * about an other task.  rt_get_task_state does not perform any check
 * on pointer task.
 */
int rt_get_task_state(RT_TASK *task)
{
    return task->state;
}


/**
 * @anchor rt_linux_use_fpu
 * @brief Set indication of FPU usage.
 *
 * rt_linux_use_fpu informs the scheduler that floating point
 * arithmetic operations will be used also by foreground Linux
 * processes, i.e. the Linux kernel itself (unlikely) and any of its
 * processes. 
 *
 * @param use_fpu_flag If this parameter has a nonzero value, the
 * Floating Point Unit (FPU) context is also switched when @e task or
 * the kernel becomes active.
 * This makes task switching slower, negligibly, on all 32 bits CPUs
 * but 386s and the oldest 486s. 
 * This flag can be set also by rt_task_init when the real time task
 * is created. With UP and MUP schedulers care is taken to avoid
 * useless saves/ restores of the FPU environment. 
 * Under SMP tasks can be moved from CPU to CPU so saves/restores for
 * tasks using the FPU are always carried out. 
 * Note that by default Linux has this flag cleared. Beside by using
 * rt_linux_use_fpu you can change the Linux FPU flag when you insmod
 * any RTAI scheduler module by setting the LinuxFpu command line
 * parameter of the rtai_sched module itself.
 *
 * @return 0 on success. A negative value on failure as described below:
 * - @b EINVAL: task does not refer to a valid task.
 *
 * See also: @ref rt_linux_use_fpu().
 */
void rt_linux_use_fpu(int use_fpu_flag)
{
    int cpuid;
    for (cpuid = 0; cpuid < num_online_cpus(); cpuid++) {
        rt_linux_task.uses_fpu = use_fpu_flag ? 1 : 0;
    }
}


/**
 * @anchor rt_task_use_fpu
 * @brief 
 *
 * rt_task_use_fpu informs the scheduler that floating point
 * arithmetic operations will be used by the real time task @e task.
 *
 * @param task is a pointer to the real time task.
 * 
 * @param use_fpu_flag If this parameter has a nonzero value, the
 * Floating Point Unit (FPU) context is also switched when @e task or
 * the kernel becomes active.
 * This makes task switching slower, negligibly, on all 32 bits CPUs
 * but 386s and the oldest 486s.
 * This flag can be set also by @ref rt_task_init() when the real time
 * task is created. With UP and MUP schedulers care is taken to avoid
 * useless saves/restores of the FPU environment.
 * Under SMP tasks can be moved from CPU to CPU so saves/restores for
 * tasks using the FPU are always carried out. 
 *
 * @return 0 on success. A negative value on failure as described below:
 * - @b EINVAL: task does not refer to a valid task.
 *
 * See also: @ref rt_linux_use_fpu().
 */
int rt_task_use_fpu(RT_TASK *task, int use_fpu_flag)
{
    if (task->magic != RT_TASK_MAGIC) {
        return -EINVAL;
    }
    task->uses_fpu = use_fpu_flag ? 1 : 0;
    return 0;
}


/**
 * @anchor rt_task_signal_handler
 * @brief Set the signal handler of a task.
 *
 * rt_task_signal_handler installs, or changes, the signal function
 * of a real time task.
 *
 * @param task is a pointer to the real time task.
 *
 * @param handler is the entry point of the signal function.
 *
 * A signal handler function can be set also when the task is newly
 * created with @ref rt_task_init(). The signal handler is a function
 * called within the task environment and with interrupts disabled,
 * when the task becomes the current running task after a context
 * switch, except at its very first scheduling. It allows you to
 * implement whatever signal management policy you think useful, and
 * many other things as well (FIXME).
 *
 * @return 0 on success.A negative value on failure as described below:
 * - @b EINVAL: task does not refer to a valid task.
 */
int rt_task_signal_handler(RT_TASK *task, void (*handler)(void))
{
    if (task->magic != RT_TASK_MAGIC) {
        return -EINVAL;
    }
    task->signal = handler;
    return 0;
}

/* ++++++++++++++++++++++++++++ MEASURING TIME ++++++++++++++++++++++++++++++ */

void rt_gettimeorig(RTIME time_orig[])
{
    unsigned long flags;
    struct timeval tv;
    rtai_save_flags_and_cli(flags);
    do_gettimeofday(&tv);
    time_orig[0] = rtai_rdtsc();
    rtai_restore_flags(flags);
    time_orig[0] = tv.tv_sec*(long long)tuned.cpu_freq + llimd(tv.tv_usec, tuned.cpu_freq, 1000000) - time_orig[0];
    time_orig[1] = llimd(time_orig[0], 1000000000, tuned.cpu_freq);
}

/* +++++++++++++++++++++++++++ CONTROLLING TIME ++++++++++++++++++++++++++++++ */

/**
 * @anchor rt_task_make_periodic_relative_ns
 * Make a task run periodically.
 *
 * rt_task_make_periodic_relative_ns mark the task @e task, previously
 * created with @ref rt_task_init(), as suitable for a periodic
 * execution, with period @e period, when @ref rt_task_wait_period()
 * is called.
 *
 * The time of first execution is defined through @e start_time or @e
 * start_delay. @e start_time is an absolute value measured in clock
 * ticks. @e start_delay is relative to the current time and measured
 * in nanoseconds. 
 *
 * @param task is a pointer to the task you want to make periodic.
 *
 * @param start_delay is the time, to wait before the task start
 *    running, in nanoseconds.
 *
 * @param period corresponds to the period of the task, in nanoseconds.
 *
 * @retval 0 on success. A negative value on failure as described below:
 * - @b EINVAL: task does not refer to a valid task.
 *
 * Recall that the term clock ticks depends on the mode in which the hard
 * timer runs. So if the hard timer was set as periodic a clock tick will
 * last as the period set in start_rt_timer, while if oneshot mode is used
 * a clock tick will last as the inverse of the running frequency of the
 * hard timer in use and irrespective of any period used in the call to
 * start_rt_timer.
 */
int rt_task_make_periodic_relative_ns(RT_TASK *task, RTIME start_delay, RTIME period)
{
    long flags;

    if (!task) {
        task = RT_CURRENT;
    } else if (task->magic != RT_TASK_MAGIC) {
        return -EINVAL;
    }
    start_delay = nano2count_cpuid(start_delay, task->runnable_on_cpus);
    period = nano2count_cpuid(period, task->runnable_on_cpus);
    flags = rt_global_save_flags_and_cli();
    task->resume_time = rt_get_time_cpuid(task->runnable_on_cpus) + start_delay;
    task->period = period;
    task->suspdepth = 0;
        if (!(task->state & RT_SCHED_DELAYED)) {
        rem_ready_task(task);
        task->state = (task->state & ~RT_SCHED_SUSPENDED) | RT_SCHED_DELAYED;
        enq_timed_task(task);
}
    RT_SCHEDULE(task, rtai_cpuid());
    rt_global_restore_flags(flags);
    return 0;
}


/**
 * @anchor rt_task_make_periodic
 * Make a task run periodically
 *
 * rt_task_make_periodic mark the task @e task, previously created
 * with @ref rt_task_init(), as suitable for a periodic execution, with
 * period @e period, when @ref rt_task_wait_period() is called.
 *
 * The time of first execution is defined through @e start_time or @e
 * start_delay. @e start_time is an absolute value measured in clock
 * ticks.  @e start_delay is relative to the current time and measured
 * in nanoseconds.
 *
 * @param task is a pointer to the task you want to make periodic.
 *
 * @param start_time is the absolute time to wait before the task start
 *    running, in clock ticks.
 *
 * @param period corresponds to the period of the task, in clock ticks.
 *
 * @retval 0 on success. A negative value on failure as described
 * below: 
 * - @b EINVAL: task does not refer to a valid task.
 *
 * See also: @ref rt_task_make_periodic_relative_ns().
 * Recall that the term clock ticks depends on the mode in which the hard
 * timer runs. So if the hard timer was set as periodic a clock tick will
 * last as the period set in start_rt_timer, while if oneshot mode is used
 * a clock tick will last as the inverse of the running frequency of the
 * hard timer in use and irrespective of any period used in the call to
 * start_rt_timer.
 *
 */
int rt_task_make_periodic(RT_TASK *task, RTIME start_time, RTIME period)
{
    long flags;

    if (!task) {
        task = RT_CURRENT;
    } else if (task->magic != RT_TASK_MAGIC) {
        return -EINVAL;
    }
    flags = rt_global_save_flags_and_cli();
    task->resume_time = start_time;
    task->period = period;
    task->suspdepth = 0;
        if (!(task->state & RT_SCHED_DELAYED)) {
        rem_ready_task(task);
        task->state = (task->state & ~RT_SCHED_SUSPENDED) | RT_SCHED_DELAYED;
        enq_timed_task(task);
    }
    RT_SCHEDULE(task, rtai_cpuid());
    rt_global_restore_flags(flags);
    return 0;
}


/**
 * @anchor rt_task_wait_period
 * Wait till next period.
 *
 * rt_task_wait_period suspends the execution of the currently running
 * real time task until the next period is reached.
 * The task must have
 * been previously marked for a periodic execution by calling
 * @ref rt_task_make_periodic() or @ref rt_task_make_periodic_relative_ns().
 *
 * @note The task is suspended only temporarily, i.e. it simply gives
 * up control until the next time period.
 */
int rt_task_wait_period(void)
{
    DECLARE_RT_CURRENT;
    long flags;

    flags = rt_global_save_flags_and_cli();
    ASSIGN_RT_CURRENT;
    if (rt_current->resync_frame) { // Request from watchdog
            rt_current->resync_frame = 0;
#ifdef CONFIG_SMP
        rt_current->resume_time = oneshot_timer ? rtai_rdtsc() : rt_smp_times[cpuid].tick_time;
#else
        rt_current->resume_time = oneshot_timer ? rtai_rdtsc() : rt_times.tick_time;
#endif
    } else if ((rt_current->resume_time += rt_current->period) > rt_time_h) {
        rt_current->state |= RT_SCHED_DELAYED;
        rem_ready_current(rt_current);
        enq_timed_task(rt_current);
        rt_schedule();
        rt_global_restore_flags(flags);
        return 0;
    }
    rt_global_restore_flags(flags);
    return 1;
}

void rt_task_set_resume_end_times(RTIME resume, RTIME end)
{
    RT_TASK *rt_current;
    long flags;

    flags = rt_global_save_flags_and_cli();
    rt_current = RT_CURRENT;
    rt_current->policy   = -1;
    rt_current->priority =  0;
    if (resume > 0) {
        rt_current->resume_time = resume;
    } else {
        rt_current->resume_time -= resume;
    }
    if (end > 0) {
        rt_current->period = end;
    } else {
        rt_current->period = rt_current->resume_time - end;
    }
    rt_current->state |= RT_SCHED_DELAYED;
    rem_ready_current(rt_current);
    enq_timed_task(rt_current);
    rt_schedule();
    rt_global_restore_flags(flags);
}

int rt_set_resume_time(RT_TASK *task, RTIME new_resume_time)
{
    long flags;

    if (task->magic != RT_TASK_MAGIC) {
        return -EINVAL;
    }

    flags = rt_global_save_flags_and_cli();
    if (task->state & RT_SCHED_DELAYED) {
        if (((task->resume_time = new_resume_time) - (task->tnext)->resume_time) > 0) {
            rem_timed_task(task);
            enq_timed_task(task);
            rt_global_restore_flags(flags);
            return 0;
            }
        }
    rt_global_restore_flags(flags);
    return -ETIME;
}

int rt_set_period(RT_TASK *task, RTIME new_period)
{
    long flags;

    if (task->magic != RT_TASK_MAGIC) {
        return -EINVAL;
    }
    rtai_save_flags_and_cli(flags);
    task->period = new_period;
    rtai_restore_flags(flags);
    return 0;
}

/**
 * @anchor next_period
 * @brief Get the time a periodic task will be resumed after calling
 *  rt_task_wait_period.
 *
 * this function returns the time when the caller task will run
 * next. Combined with the appropriate @ref rt_get_time function() it
 * can be used for checking the fraction of period used or any period
 * overrun.
 *
 * @return Next period time in internal count units.
 */
RTIME next_period(void)
{
    RT_TASK *rt_current;
    unsigned long flags;
    flags = rt_global_save_flags_and_cli();
    rt_current = RT_CURRENT;
    rt_global_restore_flags(flags);
    return rt_current->resume_time + rt_current->period;
}

/**
 * @anchor rt_busy_sleep
 * @brief Delay/suspend execution for a while.
 *
 * rt_busy_sleep delays the execution of the caller task without
 * giving back the control to the scheduler. This function burns away
 * CPU cycles in a busy wait loop so it should be used only for very
 * short synchronization delays. On machine not having a TSC clock it
 * can lead to many microseconds uncertain busy sleeps because of the
 * need of reading the 8254 timer.
 *
 * @param ns is the number of nanoseconds to wait.
 * 
 * See also: @ref rt_sleep(), @ref rt_sleep_until().
 *
 * @note A higher priority task or interrupt handler can run before
 *   the task goes to sleep, so the actual time spent in these
 *   functions may be longer than that specified.
 */
void rt_busy_sleep(int ns)
{
    RTIME end_time;
    end_time = rtai_rdtsc() + llimd(ns, tuned.cpu_freq, 1000000000);
    while (rtai_rdtsc() < end_time);
}

/**
 * @anchor rt_sleep
 * @brief Delay/suspend execution for a while.
 *
 * rt_sleep suspends execution of the caller task for a time of delay
 * internal count units. During this time the CPU is used by other
 * tasks.
 * 
 * @param delay Corresponds to the time the task is going to be suspended.
 *
 * See also: @ref rt_busy_sleep(), @ref rt_sleep_until().
 *
 * @note A higher priority task or interrupt handler can run before
 *   the task goes to sleep, so the actual time spent in these
 *   functions may be longer than the the one specified.
 */
int rt_sleep(RTIME delay)
{
    DECLARE_RT_CURRENT;
    unsigned long flags;
    flags = rt_global_save_flags_and_cli();
    ASSIGN_RT_CURRENT;
    if ((rt_current->resume_time = get_time() + delay) > rt_time_h) {
        rt_current->state |= RT_SCHED_DELAYED;
        rem_ready_current(rt_current);
        enq_timed_task(rt_current);
        rt_schedule();
        rt_global_restore_flags(flags);
        return 0;
    }
    rt_global_restore_flags(flags);
    return 1;
}

/**
 * @anchor rt_sleep_until
 * @brief Delay/suspend execution for a while.
 *
 * rt_sleep_until is similar to @ref rt_sleep() but the parameter time
 * is the absolute time till the task have to be suspended. If the
 * given time is already passed this call has no effect.
 * 
 * @param time Absolute time till the task have to be suspended
 *
 * See also: @ref rt_busy_sleep(), @ref rt_sleep_until().
 *
 * @note A higher priority task or interrupt handler can run before
 *   the task goes to sleep, so the actual time spent in these
 *   functions may be longer than the the one specified.
 */
int rt_sleep_until(RTIME time)
{
    DECLARE_RT_CURRENT;
    unsigned long flags;
    flags = rt_global_save_flags_and_cli();
    ASSIGN_RT_CURRENT;
    if ((rt_current->resume_time = time) > rt_time_h) {
        rt_current->state |= RT_SCHED_DELAYED;
        rem_ready_current(rt_current);
        enq_timed_task(rt_current);
        rt_schedule();
        rt_global_restore_flags(flags);
        return 0;
    }
    rt_global_restore_flags(flags);
    return 1;
}

int rt_task_masked_unblock(RT_TASK *task, unsigned long mask)
{
    unsigned long flags;

    if (task->magic != RT_TASK_MAGIC) {
        return -EINVAL;
    }

    if (task->state && task->state != RT_SCHED_READY) {
        flags = rt_global_save_flags_and_cli();
        if (mask & RT_SCHED_DELAYED) {
            rem_timed_task(task);
        }
        if (task->blocked_on && (mask & (RT_SCHED_SEMAPHORE | RT_SCHED_SEND | RT_SCHED_RPC | RT_SCHED_RETURN))) {
            (task->queue.prev)->next = task->queue.next;
            (task->queue.next)->prev = task->queue.prev;
            if (task->state & RT_SCHED_SEMAPHORE) {
                SEM *sem = (SEM *)task->blocked_on;
                if (++sem->count > 1 && sem->type) {
                    sem->count = 1;
                }
            }
        }
        if (task->state != RT_SCHED_READY && (task->state &= ~mask) == RT_SCHED_READY) {
            enq_ready_task(task);
            RT_SCHEDULE(task, rtai_cpuid());
        }
        rt_global_restore_flags(flags);
        return task->unblocked = 1;
    }
    return 0;
}

int rt_nanosleep(struct timespec *rqtp, struct timespec *rmtp)
{
    RTIME expire;

    if (rqtp->tv_nsec >= 1000000000L || rqtp->tv_nsec < 0 || rqtp->tv_sec < 0) {
        return -EINVAL;
    }
    rt_sleep_until(expire = rt_get_time() + timespec2count(rqtp));
    if ((expire -= rt_get_time()) > 0) {
        if (rmtp) {
            count2timespec(expire, rmtp);
        }
        return -EINTR;
    }
    return 0;
}

/* +++++++++++++++++++ READY AND TIMED QUEUE MANIPULATION +++++++++++++++++++ */

void rt_enq_ready_edf_task(RT_TASK *ready_task)
{
    enq_ready_edf_task(ready_task);
}

void rt_enq_ready_task(RT_TASK *ready_task)
{
    enq_ready_task(ready_task);
}

int rt_renq_ready_task(RT_TASK *ready_task, int priority)
{
    return renq_ready_task(ready_task, priority);
}

void rt_rem_ready_task(RT_TASK *task)
{
    rem_ready_task(task);
}

void rt_rem_ready_current(RT_TASK *rt_current)
{
    rem_ready_current(rt_current);
}

void rt_enq_timed_task(RT_TASK *timed_task)
{
    enq_timed_task(timed_task);
}

void rt_wake_up_timed_tasks(int cpuid)
{
#ifdef CONFIG_SMP
    wake_up_timed_tasks(cpuid);
#else
        wake_up_timed_tasks(0);
#endif
}

void rt_rem_timed_task(RT_TASK *task)
{
    rem_timed_task(task);
}

void rt_enqueue_blocked(RT_TASK *task, QUEUE *queue, int qtype)
{
    enqueue_blocked(task, queue, qtype);
}

void rt_dequeue_blocked(RT_TASK *task)
{
    dequeue_blocked(task);
}

int rt_renq_current(RT_TASK *rt_current, int priority)
{
    return renq_current(rt_current, priority);
}

/* ++++++++++++++++++++++++ NAMED TASK INIT/DELETE ++++++++++++++++++++++++++ */

RT_TASK *rt_named_task_init(const char *task_name, void (*thread)(long), long data, int stack_size, int prio, int uses_fpu, void(*signal)(void))
{
    RT_TASK *task;
    unsigned long name;

    if ((task = rt_get_adr(name = nam2num(task_name)))) {
        return task;
    }
        if ((task = rt_malloc(sizeof(RT_TASK))) && !rt_task_init(task, thread, data, stack_size, prio, uses_fpu, signal)) {
        if (rt_register(name, task, IS_TASK, 0)) {
            return task;
        }
        rt_task_delete(task);
    }
    rt_free(task);
    return (RT_TASK *)0;
}

RT_TASK *rt_named_task_init_cpuid(const char *task_name, void (*thread)(long), long data, int stack_size, int prio, int uses_fpu, void(*signal)(void), unsigned int run_on_cpu)
{
    RT_TASK *task;
    unsigned long name;

    if ((task = rt_get_adr(name = nam2num(task_name)))) {
        return task;
    }
        if ((task = rt_malloc(sizeof(RT_TASK))) && !rt_task_init_cpuid(task, thread, data, stack_size, prio, uses_fpu, signal, run_on_cpu)) {
        if (rt_register(name, task, IS_TASK, 0)) {
            return task;
        }
        rt_task_delete(task);
    }
    rt_free(task);
    return (RT_TASK *)0;
}

int rt_named_task_delete(RT_TASK *task)
{
    if (!rt_task_delete(task)) {
        rt_free(task);
    }
    return rt_drg_on_adr(task);
}

/* +++++++++++++++++++++++++++++++ REGISTRY +++++++++++++++++++++++++++++++++ */

#define HASHED_REGISTRY

#ifdef HASHED_REGISTRY

int max_slots;
static struct rt_registry_entry *lxrt_list;
static spinlock_t list_lock = SPIN_LOCK_UNLOCKED;

#define COLLISION_COUNT() do { col++; } while(0)
static unsigned long long col;
#ifndef COLLISION_COUNT
#define COLLISION_COUNT()
#endif

#define NONAME  (1UL)
#define NOADR   ((void *)1)

#define PRIMES_TAB_GRANULARITY  100

static unsigned short primes[ ] = { 1, 103, 211, 307, 401, 503, 601, 701, 809, 907, 1009, 1103, 1201, 1301, 1409, 1511, 1601, 1709, 1801, 1901, 2003, 2111, 2203, 2309, 2411, 2503, 2609, 2707, 2801, 2903, 3001, 3109, 3203, 3301, 3407, 3511,
3607, 3701, 3803, 3907, 4001, 4111, 4201, 4327, 4409, 4507, 4603, 4703, 4801, 4903, 5003, 5101, 5209, 5303, 5407, 5501, 5623, 5701, 5801, 5903, 6007, 6101, 6203, 6301, 6421, 6521, 6607, 6703, 6803, 6907, 7001, 7103, 7207, 7307, 7411, 7507,
7603, 7703, 7817, 7901, 8009, 8101, 8209, 8311, 8419, 8501, 8609, 8707, 8803, 8923, 9001, 9103, 9203, 9311, 9403, 9511, 9601, 9719, 9803, 9901, 10007, 10103, 10211, 10301, 10427, 10501, 10601, 10709, 10831, 10903, 11003, 11113, 11213, 11311, 11411, 11503, 11597, 11617, 11701, 11801, 11903, 12007, 12101, 12203, 12301, 12401, 12503, 12601, 12703, 12809, 12907, 13001, 13103, 13217, 13309, 13411, 13513, 13613, 13709, 13807, 13901, 14009, 14107, 14207, 14303, 14401, 14503, 14621,
14713, 14813, 14923, 15013, 15101, 15217, 15307, 15401, 15511, 15601, 15727, 15803, 15901, 16001, 16103, 16217, 16301, 16411, 16519, 16603, 16703, 16811, 16901, 17011, 17107, 17203, 17317, 17401, 17509, 17609, 17707, 17807, 17903, 18013, 18119, 18211, 18301, 18401, 18503, 18617, 18701, 18803, 18911, 19001, 19121, 19207, 19301, 19403, 19501, 19603, 19709, 19801, 19913, 20011, 20101 };

#define hash_fun(m, n) ((m)%(n) + 1)

static int hash_ins_adr(void *adr, struct rt_registry_entry *list, int lstlen, int nlink)
{
    int i, k;
    unsigned long flags;

    i = hash_fun((unsigned long)adr, lstlen);
    while (1) {
        k = i;
        while (list[k].adr > NOADR && list[k].adr != adr) {
COLLISION_COUNT();
            if (++k > lstlen) {
                k = 1;
            }
            if (k == i) {
                return 0;
            }
        }
        flags = rt_spin_lock_irqsave(&list_lock);
        if (list[k].adr == adr) {
            rt_spin_unlock_irqrestore(flags, &list_lock);
            return -k;
        } else if (list[k].adr <= NOADR) {
            list[k].adr       = adr;
            list[k].nlink     = nlink;
            list[nlink].alink = k;
            rt_spin_unlock_irqrestore(flags, &list_lock);
            return k;
        }
    }
}

static int hash_ins_name(unsigned long name, void *adr, int type, struct task_struct *lnxtsk, struct rt_registry_entry *list, int lstlen, int inc)
{
    int i, k;
    unsigned long flags;

    i = hash_fun(name, lstlen);
    while (1) {
        k = i;
        while (list[k].name > NONAME && list[k].name != name) {
COLLISION_COUNT();
            if (++k > lstlen) {
                k = 1;
            }
            if (k == i) {
                return 0;
            }
        }
        flags = rt_spin_lock_irqsave(&list_lock);
        if (list[k].name == name) {
            if (inc) {
                list[k].count++;
            }
            rt_spin_unlock_irqrestore(flags, &list_lock);
            return -k;
        } else if (list[k].name <= NONAME) {
            list[k].name  = name;
            list[k].type  = type;
            list[k].tsk   = lnxtsk;
            list[k].count = 1;
            list[k].alink = 0;
            rt_spin_unlock_irqrestore(flags, &list_lock);
                    if (hash_ins_adr(adr, list, lstlen, k) <= 0) {
                rt_spin_unlock_irqrestore(flags, &list_lock);
                            return 0;
                    }
            return k;
        }
    }
}

static void *hash_find_name(unsigned long name, struct rt_registry_entry *list, long lstlen, int inc, int *slot)
{
    int i, k;
    unsigned long flags;

    i = hash_fun(name, lstlen);
    while (1) {
        k = i;
        while (list[k].name > NONAME && list[k].name != name) {
COLLISION_COUNT();
            if (++k > lstlen) {
                k = 1;
            }
            if (k == i) {
                return NULL;
            }
        }
        flags = rt_spin_lock_irqsave(&list_lock);
        if (list[k].name == name) {
            if (inc) {
                list[k].count++;
            }
            rt_spin_unlock_irqrestore(flags, &list_lock);
            if (slot) {
                *slot = k;
            }
            return list[list[k].alink].adr;
        } else if (list[k].name <= NONAME) {
            rt_spin_unlock_irqrestore(flags, &list_lock);
            return NULL;
        }
    }
}

static unsigned long hash_find_adr(void *adr, struct rt_registry_entry *list, long lstlen, int inc)
{
    int i, k;
    unsigned long flags;

    i = hash_fun((unsigned long)adr, lstlen);
    while (1) {
        k = i;
        while (list[k].adr > NOADR && list[k].adr != adr) {
COLLISION_COUNT();
            if (++k > lstlen) {
                k = 1;
            }
            if (k == i) {
                return 0;
            }
        }
        flags = rt_spin_lock_irqsave(&list_lock);
        if (list[k].adr == adr) {
            if (inc) {
                list[list[k].nlink].count++;
            }
            rt_spin_unlock_irqrestore(flags, &list_lock);
            return list[list[k].nlink].name;
        } else if (list[k].adr <= NOADR) {
            rt_spin_unlock_irqrestore(flags, &list_lock);
            return 0;
        }
    }
}

static int hash_rem_name(unsigned long name, struct rt_registry_entry *list, long lstlen, int dec)
{
    int i, k;
    unsigned long flags;

    k = i = hash_fun(name, lstlen);
    while (list[k].name && list[k].name != name) {
COLLISION_COUNT();
        if (++k > lstlen) {
            k = 1;
        }
        if (k == i) {
            return 0;
        }
    }
    flags = rt_spin_lock_irqsave(&list_lock);
    if (list[k].name == name) {
        if (!dec || (list[k].count && !--list[k].count)) {
            int j;
            if ((i = k + 1) > lstlen) {
                i = 1;
            }
            list[k].name = !list[i].name ? 0UL : NONAME;
            if ((j = list[k].alink)) {
                if ((i = j + 1) > lstlen) {
                    i = 1;
                }
                list[j].adr = !list[i].adr ? NULL : NOADR;
            }
        }
        if (dec) {
            k = list[k].count;
        }
        rt_spin_unlock_irqrestore(flags, &list_lock);
        return k;
    }
    rt_spin_unlock_irqrestore(flags, &list_lock);
    return dec;
}

static int hash_rem_adr(void *adr, struct rt_registry_entry *list, long lstlen, int dec)
{
    int i, k;
    unsigned long flags;

    k = i = hash_fun((unsigned long)adr, lstlen);
    while (list[k].adr && list[k].adr != adr) {
COLLISION_COUNT();
        if (++k > lstlen) {
            k = 1;
        }
        if (k == i) {
            return 0;
        }
    }
    flags = rt_spin_lock_irqsave(&list_lock);
    if (list[k].adr == adr) {
        if (!dec || (list[list[k].nlink].count && !--list[list[k].nlink].count)) {
            int j;
            if ((i = k + 1) > lstlen) {
                i = 1;
            }
            list[k].adr = !list[i].adr ? NULL : NOADR;
            j = list[k].nlink;
            if ((i = j + 1) > lstlen) {
                i = 1;
            }
            list[j].name = !list[i].name ? 0UL : NONAME;
        }
        if (dec) {
            k = list[list[k].nlink].count;
        }
        rt_spin_unlock_irqrestore(flags, &list_lock);
        return k;
    }
    rt_spin_unlock_irqrestore(flags, &list_lock);
    return dec;
}

static inline int registr(unsigned long name, void *adr, int type, struct task_struct *lnxtsk)
{
    return abs(hash_ins_name(name, adr, type, lnxtsk, lxrt_list, max_slots, 1));
}

static inline int drg_on_name(unsigned long name)
{
    return hash_rem_name(name, lxrt_list, max_slots, 0);
} 

static inline int drg_on_name_cnt(unsigned long name)
{
    return hash_rem_name(name, lxrt_list, max_slots, -EFAULT);
} 

static inline int drg_on_adr(void *adr)
{
    return hash_rem_adr(adr, lxrt_list, max_slots, 0);
} 

static inline int drg_on_adr_cnt(void *adr)
{
    return hash_rem_adr(adr, lxrt_list, max_slots, -EFAULT);
} 

static inline unsigned long get_name(void *adr)
{
    static unsigned long nameseed = 3518743764UL;
    if (!adr) {
        unsigned long flags;
        unsigned long name;
        flags = rt_spin_lock_irqsave(&list_lock);
        if ((name = ++nameseed) == 0xFFFFFFFFUL) {
            nameseed = 3518743764UL;
        }
        rt_spin_unlock_irqrestore(flags, &list_lock);
        return name;
    } else {
        return hash_find_adr(adr, lxrt_list, max_slots, 0);
    }
    return 0;
} 

static inline void *get_adr(unsigned long name)
{
    return hash_find_name(name, lxrt_list, max_slots, 0, NULL);
} 

static inline void *get_adr_cnt(unsigned long name)
{
    return hash_find_name(name, lxrt_list, max_slots, 1, NULL);
} 

static inline int get_type(unsigned long name)
{
    int slot;

    if (hash_find_name(name, lxrt_list, max_slots, 0, &slot)) {
        return lxrt_list[slot].type;
    }
        return -EINVAL;
}

unsigned long is_process_registered(struct task_struct *lnxtsk)
{
    void *adr = lnxtsk->rtai_tskext(TSKEXT0);
    return adr ? hash_find_adr(adr, lxrt_list, max_slots, 0) : 0;
}

int rt_get_registry_slot(int slot, struct rt_registry_entry *entry)
{
    unsigned long flags;
        flags = rt_spin_lock_irqsave(&list_lock);
    if (lxrt_list[slot].name > NONAME) {
        *entry = lxrt_list[slot];
        entry->adr = lxrt_list[entry->alink].adr;
        rt_spin_unlock_irqrestore(flags, &list_lock);
        return slot;
        }
        rt_spin_unlock_irqrestore(flags, &list_lock);
        return 0;
}

int rt_registry_alloc(void)
{
    if ((max_slots = (MAX_SLOTS + PRIMES_TAB_GRANULARITY - 1)/(PRIMES_TAB_GRANULARITY)) >= sizeof(primes)/sizeof(primes[0])) {
        printk("REGISTRY TABLE TOO LARGE FOR AVAILABLE PRIMES\n");
                return -ENOMEM;
        }
    max_slots = primes[max_slots];
    if (!(lxrt_list = vmalloc((max_slots + 1)*sizeof(struct rt_registry_entry)))) {
        printk("NO MEMORY FOR REGISTRY TABLE\n");
                return -ENOMEM;
    }
    memset(lxrt_list, 0, (max_slots + 1)*sizeof(struct rt_registry_entry));
    return 0;
}

void rt_registry_free(void)
{
    if (lxrt_list) {
        vfree(lxrt_list);
    }
}
#else
volatile int max_slots;
static struct rt_registry_entry *lxrt_list;
static spinlock_t list_lock = SPIN_LOCK_UNLOCKED;

int rt_registry_alloc(void)
{
    if (!(lxrt_list = vmalloc((MAX_SLOTS + 1)*sizeof(struct rt_registry_entry)))) {
                printk("NO MEMORY FOR REGISTRY TABLE\n");
                return -ENOMEM;
        }
        memset(lxrt_list, 0, (MAX_SLOTS + 1)*sizeof(struct rt_registry_entry));
        return 0;
}

void rt_registry_free(void)
{
    if (lxrt_list) {
        vfree(lxrt_list);
    }
}

static inline int registr(unsigned long name, void *adr, int type, struct task_struct *tsk)
{
        unsigned long flags;
        int i, slot;
/*
 * Register a resource. This allows other programs (RTAI and/or user space)
 * to use the same resource because they can find the address from the name.
*/
        // index 0 is reserved for the null slot.
    while ((slot = max_slots) < MAX_SLOTS) {
            for (i = 1; i <= max_slots; i++) {
                    if (lxrt_list[i].name == name) {
                return 0;
            }
        }
            flags = rt_spin_lock_irqsave(&list_lock);
                if (slot == max_slots && max_slots < MAX_SLOTS) {
            slot = ++max_slots;
                        lxrt_list[slot].name  = name;
                        lxrt_list[slot].adr   = adr;
                        lxrt_list[slot].tsk   = tsk;
                        lxrt_list[slot].type  = type;
                        lxrt_list[slot].count = 1;
                        rt_spin_unlock_irqrestore(flags, &list_lock);
                        return slot;
                }
            rt_spin_unlock_irqrestore(flags, &list_lock);
        }
        return 0;
}

static inline int drg_on_name(unsigned long name)
{
    unsigned long flags;
    int slot;
    for (slot = 1; slot <= max_slots; slot++) {
        flags = rt_spin_lock_irqsave(&list_lock);
        if (lxrt_list[slot].name == name) {
            if (slot < max_slots) {
                lxrt_list[slot] = lxrt_list[max_slots];
            }
            if (max_slots > 0) {
                max_slots--;
            }
            rt_spin_unlock_irqrestore(flags, &list_lock);
            return slot;
        }
        rt_spin_unlock_irqrestore(flags, &list_lock);
    }
    return 0;
} 

static inline int drg_on_name_cnt(unsigned long name)
{
    unsigned long flags;
    int slot, count;
    for (slot = 1; slot <= max_slots; slot++) {
        flags = rt_spin_lock_irqsave(&list_lock);
        if (lxrt_list[slot].name == name && lxrt_list[slot].count > 0 && !(count = --lxrt_list[slot].count)) {
            if (slot < max_slots) {
                lxrt_list[slot] = lxrt_list[max_slots];
            }
            if (max_slots > 0) {
                max_slots--;
            }
            rt_spin_unlock_irqrestore(flags, &list_lock);
            return count;
        }
        rt_spin_unlock_irqrestore(flags, &list_lock);
    }
    return -EFAULT;
} 

static inline int drg_on_adr(void *adr)
{
    unsigned long flags;
    int slot;
    for (slot = 1; slot <= max_slots; slot++) {
        flags = rt_spin_lock_irqsave(&list_lock);
        if (lxrt_list[slot].adr == adr) {
            if (slot < max_slots) {
                lxrt_list[slot] = lxrt_list[max_slots];
            }
            if (max_slots > 0) {
                max_slots--;
            }
            rt_spin_unlock_irqrestore(flags, &list_lock);
            return slot;
        }
        rt_spin_unlock_irqrestore(flags, &list_lock);
    }
    return 0;
} 

static inline int drg_on_adr_cnt(void *adr)
{
    unsigned long flags;
    int slot, count;
    for (slot = 1; slot <= max_slots; slot++) {
        flags = rt_spin_lock_irqsave(&list_lock);
        if (lxrt_list[slot].adr == adr && lxrt_list[slot].count > 0 && !(count = --lxrt_list[slot].count)) {
            if (slot < max_slots) {
                lxrt_list[slot] = lxrt_list[max_slots];
            }
            if (max_slots > 0) {
                max_slots--;
            }
            rt_spin_unlock_irqrestore(flags, &list_lock);
            return count;
        }
        rt_spin_unlock_irqrestore(flags, &list_lock);
    }
    return -EFAULT;
} 

static inline unsigned long get_name(void *adr)
{
    static unsigned long nameseed = 3518743764UL;
    int slot;
        if (!adr) {
        unsigned long flags;
        unsigned long name;
        flags = rt_spin_lock_irqsave(&list_lock);
        if ((name = ++nameseed) == 0xFFFFFFFFUL) {
            nameseed = 3518743764UL;
        }
        rt_spin_unlock_irqrestore(flags, &list_lock);
        return name;
        }
    for (slot = 1; slot <= max_slots; slot++) {
        if (lxrt_list[slot].adr == adr) {
            return lxrt_list[slot].name;
        }
    }
    return 0;
} 

static inline void *get_adr(unsigned long name)
{
    int slot;
    for (slot = 1; slot <= max_slots; slot++) {
        if (lxrt_list[slot].name == name) {
            return lxrt_list[slot].adr;
        }
    }
    return 0;
} 

static inline void *get_adr_cnt(unsigned long name)
{
    unsigned long flags;
    int slot;
    for (slot = 1; slot <= max_slots; slot++) {
        flags = rt_spin_lock_irqsave(&list_lock);
        if (lxrt_list[slot].name == name) {
            ++lxrt_list[slot].count;
            rt_spin_unlock_irqrestore(flags, &list_lock);
            return lxrt_list[slot].adr;
        }
        rt_spin_unlock_irqrestore(flags, &list_lock);
    }
    return 0;
} 

static inline int get_type(unsigned long name)
{
        int slot;
        for (slot = 1; slot <= max_slots; slot++) {
                if (lxrt_list[slot].name == name) {
                        return lxrt_list[slot].type;
                }
        }
        return -EINVAL;
}

unsigned long is_process_registered(struct task_struct *tsk)
{
        void *adr;

        if ((adr = tsk->rtai_tskext(TSKEXT0))) {
        int slot;
        for (slot = 1; slot <= max_slots; slot++) {
            if (lxrt_list[slot].adr == adr) {
                return lxrt_list[slot].name;
            }
                }
        }
        return 0;
}

int rt_get_registry_slot(int slot, struct rt_registry_entry *entry)
{
    unsigned long flags;

    if(entry == 0) {
        return 0;
    }
    flags = rt_spin_lock_irqsave(&list_lock);
    if (slot > 0 && slot <= max_slots ) {
        if (lxrt_list[slot].name != 0) {
            *entry = lxrt_list[slot];
            rt_spin_unlock_irqrestore(flags, &list_lock);
            return slot;
        }
    }
    rt_spin_unlock_irqrestore(flags, &list_lock);

    return 0;
}
#endif

/**
 * @ingroup lxrt
 * Register an object.
 *
 * rt_register registers the object to be identified with @a name, which is
 * pointed by @a adr.
 *
 * @return a positive number on success, 0 on failure.
 */
int rt_register(unsigned long name, void *adr, int type, struct task_struct *t)
{
/*
 * Register a resource. This function provides the service to all RTAI tasks.
*/
    return get_adr(name) ? 0 : registr(name, adr, type, t );
}


/**
 * @ingroup lxrt
 * Deregister an object by its name.
 *
 * rt_drg_on_name deregisters the object identified by its @a name.
 *
 * @return a positive number on success, 0 on failure.
 */
int rt_drg_on_name(unsigned long name)
{
    return drg_on_name(name);
} 

/**
 * @ingroup lxrt
 * Deregister an object by its address.
 *
 * rt_drg_on_adr deregisters the object identified by its @a adr.
 *
 * @return a positive number on success, 0 on failure.
 */
int rt_drg_on_adr(void *adr)
{
    return drg_on_adr(adr);
} 

unsigned long rt_get_name(void *adr)
{
    return get_name(adr);
} 

void *rt_get_adr(unsigned long name)
{
    return get_adr(name);
}

int rt_get_type(unsigned long name)
{
    return get_type(name);
}

int rt_drg_on_name_cnt(unsigned long name)
{
    return drg_on_name_cnt(name);
}

int rt_drg_on_adr_cnt(void *adr)
{
    return drg_on_adr_cnt(adr);
}

void *rt_get_adr_cnt(unsigned long name)
{
    return get_adr_cnt(name);
}

#include <rtai_lxrt.h>

extern struct rt_fun_entry rt_fun_lxrt[];

void krtai_objects_release(void)
{
    int slot;
        struct rt_registry_entry entry;
    char name[8], *type;

    for (slot = 1; slot <= max_slots; slot++) {
                if (rt_get_registry_slot(slot, &entry)) {
            switch (entry.type) {
                            case IS_TASK:
                    type = "TASK";
                    rt_named_task_delete(entry.adr);
                    break;
                case IS_SEM:
                    type = "SEM ";
                    ((void (*)(void *))rt_fun_lxrt[NAMED_SEM_DELETE].fun)(entry.adr);
                    break;
                case IS_RWL:
                    type = "RWL ";
                    ((void (*)(void *))rt_fun_lxrt[NAMED_RWL_DELETE].fun)(entry.adr);
                    break;
                case IS_SPL:
                    type = "SPL ";
                    ((void (*)(void *))rt_fun_lxrt[NAMED_SPL_DELETE].fun)(entry.adr);
                    break;
                case IS_MBX:
                    type = "MBX ";
                    ((void (*)(void *))rt_fun_lxrt[NAMED_MBX_DELETE].fun)(entry.adr);
                                break;  
                case IS_PRX:
                    type = "PRX ";
                    ((void (*)(void *))rt_fun_lxrt[PROXY_DETACH].fun)(entry.adr);
                    rt_drg_on_adr(entry.adr); 
                    break;
                            default:
                    type = "ALIEN";
                    break;
            }
            num2nam(entry.name, name);
            rt_printk("SCHED releases registered named %s %s\n", type, name);
        }
    }
}

/* +++++++++++++++++++++++++ SUPPORT FOR IRQ TASKS ++++++++++++++++++++++++++ */

#include <rtai_tasklets.h>

extern struct {
    int (*handler)(unsigned irq, void *cookie);
    void *cookie;
    int retmode;
    int cpumask;
} rtai_realtime_irq[];

int rt_irq_wait(unsigned irq)
{   
    int retval;
    retval = rt_task_suspend(0);
    return rtai_realtime_irq[irq].handler ? -retval : RT_IRQ_TASK_ERR;
}

int rt_irq_wait_if(unsigned irq)
{
    int retval;
    retval = rt_task_suspend_if(0);
    return rtai_realtime_irq[irq].handler ? -retval : RT_IRQ_TASK_ERR;
}

int rt_irq_wait_until(unsigned irq, RTIME time)
{
    int retval;
    retval = rt_task_suspend_until(0, time);
    return rtai_realtime_irq[irq].handler ? -retval : RT_IRQ_TASK_ERR;
}

int rt_irq_wait_timed(unsigned irq, RTIME delay)
{
    return rt_irq_wait_until(irq, get_time() + delay);
}

void rt_irq_signal(unsigned irq)
{
    if (rtai_realtime_irq[irq].handler) {
        rt_task_resume((void *)rtai_realtime_irq[irq].cookie);
    }
}

static int rt_irq_task_handler(unsigned irq, RT_TASK *irq_task)
{
    rt_task_resume(irq_task);
    return 0;
}

int rt_request_irq_task (unsigned irq, void *handler, int type, int affine2task)
{
    RT_TASK *task;
    if (!handler) {
        task = _rt_whoami();
    } else {
        task = type == RT_IRQ_TASKLET ? ((struct rt_tasklet_struct *)handler)->task : handler;
    }
    if (affine2task) {
        rt_assign_irq_to_cpu(irq, (1 << task->runnable_on_cpus));
    }
    return rt_request_irq(irq, (void *)rt_irq_task_handler, task, 0);
}

int rt_release_irq_task (unsigned irq)
{
    int retval;
    RT_TASK *task;
    task = (void *)rtai_realtime_irq[irq].cookie;
    if (!(retval = rt_release_irq(irq))) {
        rt_task_resume(task);
        rt_reset_irq_to_sym_mode(irq);
    }
    return retval;
}

extern void usp_request_rtc(int, void *);
void usp_request_rtc(int rtc_freq, void *handler)
{
    rt_request_rtc(rtc_freq, !handler || (handler && handler == (void *)1) ? handler : rt_irq_signal);
        
}

/* +++++++++++++++++ SUPPORT FOR THE LINUX SYSCALL SERVER +++++++++++++++++++ */

RT_TASK *rt_exec_linux_syscall(RT_TASK *rt_current, RT_TASK *task, struct pt_regs *regs)
{
    unsigned long flags;

    flags = rt_global_save_flags_and_cli();
    if (task->state & RT_SCHED_RECEIVE) {
        rt_current->msg = task->msg = (unsigned long)regs;
        task->msg_queue.task = rt_current;
        task->ret_queue.task = NOTHING;
        task->state = RT_SCHED_READY;
        enq_ready_task(task);
        enqueue_blocked(rt_current, &task->ret_queue, 1);
        rt_current->state |= RT_SCHED_RETURN;
    } else {
        rt_current->msg = (unsigned long)regs;
                enqueue_blocked(rt_current, &task->msg_queue, 1);
        rt_current->state |= RT_SCHED_RPC;
    }
    task->priority = rt_current->priority;
    rem_ready_current(rt_current);
    rt_current->msg_queue.task = task;
    rt_schedule();
    rt_global_restore_flags(flags);
    return rt_current->msg_queue.task != rt_current ? NULL : task;
}

#include <asm/uaccess.h>
RT_TASK *rt_receive_linux_syscall(RT_TASK *task, struct pt_regs *regs)
{
    unsigned long flags;
    RT_TASK *rt_current;

    flags = rt_global_save_flags_and_cli();
    rt_current = rt_smp_current[rtai_cpuid()];
    if ((task->state & RT_SCHED_RPC) && task->msg_queue.task == rt_current) {
        dequeue_blocked(task);
        *regs = *((struct pt_regs *)task->msg);
        rt_current->msg_queue.task = task;
        enqueue_blocked(task, &rt_current->ret_queue, 1);
        task->state = (task->state & ~RT_SCHED_RPC) | RT_SCHED_RETURN;
    } else {
        rt_current->ret_queue.task = SOMETHING;
        rt_current->state |= RT_SCHED_RECEIVE;
        rem_ready_current(rt_current);
        rt_current->msg_queue.task = task != rt_current ? task : NULL;
        rt_schedule();
        *regs = *((struct pt_regs *)rt_current->msg);
    }
    rt_current->msg_queue.task = rt_current;
    rt_global_restore_flags(flags);
    return rt_current->ret_queue.task ? NULL : task;
}

void rt_return_linux_syscall(RT_TASK *task, unsigned long retval)
{
    unsigned long flags;

    ((struct pt_regs *)task->msg)->LINUX_SYSCALL_RETREG = retval;
    flags = rt_global_save_flags_and_cli();
    dequeue_blocked(task);
    task->msg = 0;
    task->msg_queue.task = task;
    if ((task->state &= ~RT_SCHED_RETURN) == RT_SCHED_READY) {
        enq_ready_task(task);
    }
    rt_global_restore_flags(flags);
}

/* ++++++++++++++++++++ END OF COMMON FUNCTIONALITIES +++++++++++++++++++++++ */

#ifdef CONFIG_PROC_FS
#include <linux/stat.h>
#include <linux/proc_fs.h>
#include <rtai_proc_fs.h>
#include <rtai_nam2num.h>

extern struct proc_dir_entry *rtai_proc_root;

/* ----------------------< proc filesystem section >----------------------*/

static int rtai_read_lxrt(char *page, char **start, off_t off, int count, int *eof, void *data)
{
    PROC_PRINT_VARS;
    struct rt_registry_entry entry;
    char *type_name[] = { "TASK", "SEM", "RWL", "SPL", "MBX", "PRX", "BITS", "TBX", "HPCK" };
    unsigned int i = 1;
    char name[8];

    PROC_PRINT("\nRTAI LXRT Information.\n\n");
    PROC_PRINT("    MAX_SLOTS = %d\n\n", MAX_SLOTS);

//                  1234 123456 0x12345678 ALIEN  0x12345678 0x12345678   1234567      1234567

    PROC_PRINT("                                         Linux_Owner         Parent PID\n");
    PROC_PRINT("Slot Name   ID         Type   RT_Handle    Pointer   Tsk_PID   MEM_Sz   USG Cnt\n");
    PROC_PRINT("-------------------------------------------------------------------------------\n");
    for (i = 1; i <= max_slots; i++) {
        if (rt_get_registry_slot(i, &entry)) {
            num2nam(entry.name, name);
            PROC_PRINT("%4d %-6.6s 0x%08lx %-6.6s 0x%p 0x%p  %7d   %8d %7d\n",
            i,              // the slot number
            name,               // the name in 6 char asci
            entry.name,         // the name as unsigned long hex
            entry.type >= PAGE_SIZE ? "SHMEM" : 
            entry.type > sizeof(type_name)/sizeof(char *) ? 
            "ALIEN" : 
            type_name[entry.type],  // the Type
            entry.adr,      // The RT Handle
            entry.tsk,          // The Owner task pointer
            entry.tsk ? entry.tsk->pid : 0, // The Owner PID
            entry.type == IS_TASK && ((RT_TASK *)entry.adr)->lnxtsk ? (((RT_TASK *)entry.adr)->lnxtsk)->pid : entry.type >= PAGE_SIZE ? entry.type : 0, entry.count);
         }
    }
        PROC_PRINT_DONE;
}  /* End function - rtai_read_lxrt */

int rtai_proc_lxrt_register(void)
{
    struct proc_dir_entry *proc_lxrt_ent;


    proc_lxrt_ent = create_proc_entry("names", S_IFREG|S_IRUGO|S_IWUSR, rtai_proc_root);
    if (!proc_lxrt_ent) {
        printk("Unable to initialize /proc/rtai/lxrt\n");
        return(-1);
    }
    proc_lxrt_ent->read_proc = rtai_read_lxrt;
    return(0);
}  /* End function - rtai_proc_lxrt_register */


void rtai_proc_lxrt_unregister(void)
{
    remove_proc_entry("names", rtai_proc_root);
}  /* End function - rtai_proc_lxrt_unregister */

/* ------------------< end of proc filesystem section >------------------*/
#endif /* CONFIG_PROC_FS */

#ifdef CONFIG_KBUILD

EXPORT_SYMBOL(rt_set_sched_policy);
EXPORT_SYMBOL(rt_get_prio);
EXPORT_SYMBOL(rt_get_inher_prio);
EXPORT_SYMBOL(rt_change_prio);
EXPORT_SYMBOL(rt_whoami);
EXPORT_SYMBOL(rt_task_yield);
EXPORT_SYMBOL(rt_task_suspend);
EXPORT_SYMBOL(rt_task_suspend_if);
EXPORT_SYMBOL(rt_task_suspend_until);
EXPORT_SYMBOL(rt_task_suspend_timed);
EXPORT_SYMBOL(rt_task_resume);
EXPORT_SYMBOL(rt_get_task_state);
EXPORT_SYMBOL(rt_linux_use_fpu);
EXPORT_SYMBOL(rt_task_use_fpu);
EXPORT_SYMBOL(rt_task_signal_handler);
EXPORT_SYMBOL(rt_gettimeorig);
EXPORT_SYMBOL(rt_task_make_periodic_relative_ns);
EXPORT_SYMBOL(rt_task_make_periodic);
EXPORT_SYMBOL(rt_task_wait_period);
EXPORT_SYMBOL(rt_task_set_resume_end_times);
EXPORT_SYMBOL(rt_set_resume_time);
EXPORT_SYMBOL(rt_set_period);
EXPORT_SYMBOL(next_period);
EXPORT_SYMBOL(rt_busy_sleep);
EXPORT_SYMBOL(rt_sleep);
EXPORT_SYMBOL(rt_sleep_until);
EXPORT_SYMBOL(rt_task_masked_unblock);
EXPORT_SYMBOL(rt_nanosleep);
EXPORT_SYMBOL(rt_enq_ready_edf_task);
EXPORT_SYMBOL(rt_enq_ready_task);
EXPORT_SYMBOL(rt_renq_ready_task);
EXPORT_SYMBOL(rt_rem_ready_task);
EXPORT_SYMBOL(rt_rem_ready_current);
EXPORT_SYMBOL(rt_enq_timed_task);
EXPORT_SYMBOL(rt_wake_up_timed_tasks);
EXPORT_SYMBOL(rt_rem_timed_task);
EXPORT_SYMBOL(rt_enqueue_blocked);
EXPORT_SYMBOL(rt_dequeue_blocked);
EXPORT_SYMBOL(rt_renq_current);
EXPORT_SYMBOL(rt_named_task_init);
EXPORT_SYMBOL(rt_named_task_init_cpuid);
EXPORT_SYMBOL(rt_named_task_delete);
EXPORT_SYMBOL(is_process_registered);
EXPORT_SYMBOL(rt_register);
EXPORT_SYMBOL(rt_drg_on_name);
EXPORT_SYMBOL(rt_drg_on_adr);
EXPORT_SYMBOL(rt_get_name);
EXPORT_SYMBOL(rt_get_adr);
EXPORT_SYMBOL(rt_get_type);
EXPORT_SYMBOL(rt_drg_on_name_cnt);
EXPORT_SYMBOL(rt_drg_on_adr_cnt);
EXPORT_SYMBOL(rt_get_adr_cnt);
EXPORT_SYMBOL(rt_get_registry_slot);

EXPORT_SYMBOL(rt_task_init);
EXPORT_SYMBOL(rt_task_init_cpuid);
EXPORT_SYMBOL(rt_set_runnable_on_cpus);
EXPORT_SYMBOL(rt_set_runnable_on_cpuid);
EXPORT_SYMBOL(rt_check_current_stack);
EXPORT_SYMBOL(rt_schedule);
EXPORT_SYMBOL(rt_spv_RMS);
EXPORT_SYMBOL(rt_sched_lock);
EXPORT_SYMBOL(rt_sched_unlock);
EXPORT_SYMBOL(rt_task_delete);
EXPORT_SYMBOL(rt_is_hard_timer_running);
EXPORT_SYMBOL(rt_set_periodic_mode);
EXPORT_SYMBOL(rt_set_oneshot_mode);
EXPORT_SYMBOL(rt_get_timer_cpu);
EXPORT_SYMBOL(start_rt_timer);
EXPORT_SYMBOL(stop_rt_timer);
EXPORT_SYMBOL(start_rt_apic_timers);
EXPORT_SYMBOL(rt_sched_type);
EXPORT_SYMBOL(rt_hard_timer_tick_count);
EXPORT_SYMBOL(rt_hard_timer_tick_count_cpuid);
EXPORT_SYMBOL(rt_set_task_trap_handler);
EXPORT_SYMBOL(rt_get_time);
EXPORT_SYMBOL(rt_get_time_cpuid);
EXPORT_SYMBOL(rt_get_time_ns);
EXPORT_SYMBOL(rt_get_time_ns_cpuid);
EXPORT_SYMBOL(rt_get_cpu_time_ns);
EXPORT_SYMBOL(rt_get_base_linux_task);
EXPORT_SYMBOL(rt_alloc_dynamic_task);
EXPORT_SYMBOL(rt_register_watchdog);
EXPORT_SYMBOL(rt_deregister_watchdog);
EXPORT_SYMBOL(count2nano);
EXPORT_SYMBOL(nano2count);
EXPORT_SYMBOL(count2nano_cpuid);
EXPORT_SYMBOL(nano2count_cpuid);

EXPORT_SYMBOL(rt_kthread_init);
EXPORT_SYMBOL(rt_smp_linux_task);
EXPORT_SYMBOL(rt_smp_current);
EXPORT_SYMBOL(rt_smp_time_h);
EXPORT_SYMBOL(rt_smp_oneshot_timer);
EXPORT_SYMBOL(wake_up_srq);
EXPORT_SYMBOL(set_rt_fun_entries);
EXPORT_SYMBOL(reset_rt_fun_entries);
EXPORT_SYMBOL(set_rt_fun_ext_index);
EXPORT_SYMBOL(reset_rt_fun_ext_index);
EXPORT_SYMBOL(max_slots);

#ifdef CONFIG_SMP
#endif /* CONFIG_SMP */

#endif /* CONFIG_KBUILD */

---