1 Star 0 Fork 2.1K

freeman_zsk/rt-thread

forked from RT-Thread/rt-thread 
加入 Gitee
与超过 1200万 开发者一起发现、参与优秀开源项目,私有仓库也完全免费 :)
免费加入
文件
克隆/下载
scheduler_comm.c 7.96 KB
一键复制 编辑 原始数据 按行查看 历史
Shell 提交于 2024-08-28 20:54 . feat: kernel/libcpu: fit into ilp32d
/*
* Copyright (c) 2006-2024 RT-Thread Development Team
*
* SPDX-License-Identifier: Apache-2.0
*
* (scheduler_comm.c) Common API of scheduling routines.
*
* Change Logs:
* Date Author Notes
* 2024-01-18 Shell Separate scheduling related codes from thread.c, scheduler_.*
*/
#define DBG_TAG "kernel.sched"
#define DBG_LVL DBG_INFO
#include <rtdbg.h>
#include <rtthread.h>
void rt_sched_thread_init_ctx(struct rt_thread *thread, rt_uint32_t tick, rt_uint8_t priority)
{
/* setup thread status */
RT_SCHED_CTX(thread).stat = RT_THREAD_INIT;
#ifdef RT_USING_SMP
/* not bind on any cpu */
RT_SCHED_CTX(thread).bind_cpu = RT_CPUS_NR;
RT_SCHED_CTX(thread).oncpu = RT_CPU_DETACHED;
#endif /* RT_USING_SMP */
rt_sched_thread_init_priv(thread, tick, priority);
}
rt_err_t rt_sched_thread_timer_start(struct rt_thread *thread)
{
RT_SCHED_DEBUG_IS_LOCKED;
RT_SCHED_CTX(thread).sched_flag_ttmr_set = 1;
return RT_EOK;
}
rt_err_t rt_sched_thread_timer_stop(struct rt_thread *thread)
{
rt_err_t error;
RT_SCHED_DEBUG_IS_LOCKED;
if (RT_SCHED_CTX(thread).sched_flag_ttmr_set)
{
error = rt_timer_stop(&thread->thread_timer);
/* mask out timer flag no matter stop success or not */
RT_SCHED_CTX(thread).sched_flag_ttmr_set = 0;
}
else
{
error = RT_EOK;
}
return error;
}
rt_uint8_t rt_sched_thread_get_stat(struct rt_thread *thread)
{
RT_SCHED_DEBUG_IS_LOCKED;
return RT_SCHED_CTX(thread).stat & RT_THREAD_STAT_MASK;
}
rt_uint8_t rt_sched_thread_get_curr_prio(struct rt_thread *thread)
{
RT_SCHED_DEBUG_IS_LOCKED;
return RT_SCHED_PRIV(thread).current_priority;
}
rt_uint8_t rt_sched_thread_get_init_prio(struct rt_thread *thread)
{
/* read only fields, so lock is unecessary */
return RT_SCHED_PRIV(thread).init_priority;
}
/**
* @note Caller must hold the scheduler lock
*/
rt_uint8_t rt_sched_thread_is_suspended(struct rt_thread *thread)
{
RT_SCHED_DEBUG_IS_LOCKED;
return (RT_SCHED_CTX(thread).stat & RT_THREAD_SUSPEND_MASK) == RT_THREAD_SUSPEND_MASK;
}
rt_err_t rt_sched_thread_close(struct rt_thread *thread)
{
RT_SCHED_DEBUG_IS_LOCKED;
RT_SCHED_CTX(thread).stat = RT_THREAD_CLOSE;
return RT_EOK;
}
rt_err_t rt_sched_thread_yield(struct rt_thread *thread)
{
RT_SCHED_DEBUG_IS_LOCKED;
RT_SCHED_PRIV(thread).remaining_tick = RT_SCHED_PRIV(thread).init_tick;
RT_SCHED_CTX(thread).stat |= RT_THREAD_STAT_YIELD;
return RT_EOK;
}
rt_err_t rt_sched_thread_ready(struct rt_thread *thread)
{
rt_err_t error;
RT_SCHED_DEBUG_IS_LOCKED;
if (!rt_sched_thread_is_suspended(thread))
{
/* failed to proceed, and that's possibly due to a racing condition */
error = -RT_EINVAL;
}
else
{
if (RT_SCHED_CTX(thread).sched_flag_ttmr_set)
{
/**
* Quiet timeout timer first if set. and don't continue if we
* failed, because it probably means that a timeout ISR racing to
* resume thread before us.
*/
error = rt_sched_thread_timer_stop(thread);
}
else
{
error = RT_EOK;
}
if (!error)
{
/* remove from suspend list */
rt_list_remove(&RT_THREAD_LIST_NODE(thread));
#ifdef RT_USING_SMART
thread->wakeup_handle.func = RT_NULL;
#endif
/* insert to schedule ready list and remove from susp list */
rt_sched_insert_thread(thread);
}
}
return error;
}
rt_err_t rt_sched_tick_increase(rt_tick_t tick)
{
struct rt_thread *thread;
rt_sched_lock_level_t slvl;
thread = rt_thread_self();
rt_sched_lock(&slvl);
if(RT_SCHED_PRIV(thread).remaining_tick > tick)
{
RT_SCHED_PRIV(thread).remaining_tick -= tick;
}
else
{
RT_SCHED_PRIV(thread).remaining_tick = 0;
}
if (RT_SCHED_PRIV(thread).remaining_tick)
{
rt_sched_unlock(slvl);
}
else
{
rt_sched_thread_yield(thread);
/* request a rescheduling even though we are probably in an ISR */
rt_sched_unlock_n_resched(slvl);
}
return RT_EOK;
}
/**
* @brief Update priority of the target thread
*/
rt_err_t rt_sched_thread_change_priority(struct rt_thread *thread, rt_uint8_t priority)
{
RT_ASSERT(priority < RT_THREAD_PRIORITY_MAX);
RT_SCHED_DEBUG_IS_LOCKED;
/* for ready thread, change queue; otherwise simply update the priority */
if ((RT_SCHED_CTX(thread).stat & RT_THREAD_STAT_MASK) == RT_THREAD_READY)
{
/* remove thread from schedule queue first */
rt_sched_remove_thread(thread);
/* change thread priority */
RT_SCHED_PRIV(thread).current_priority = priority;
/* recalculate priority attribute */
#if RT_THREAD_PRIORITY_MAX > 32
RT_SCHED_PRIV(thread).number = RT_SCHED_PRIV(thread).current_priority >> 3; /* 5bit */
RT_SCHED_PRIV(thread).number_mask = 1 << RT_SCHED_PRIV(thread).number;
RT_SCHED_PRIV(thread).high_mask = 1 << (RT_SCHED_PRIV(thread).current_priority & 0x07); /* 3bit */
#else
RT_SCHED_PRIV(thread).number_mask = 1 << RT_SCHED_PRIV(thread).current_priority;
#endif /* RT_THREAD_PRIORITY_MAX > 32 */
RT_SCHED_CTX(thread).stat = RT_THREAD_INIT;
/* insert thread to schedule queue again */
rt_sched_insert_thread(thread);
}
else
{
RT_SCHED_PRIV(thread).current_priority = priority;
/* recalculate priority attribute */
#if RT_THREAD_PRIORITY_MAX > 32
RT_SCHED_PRIV(thread).number = RT_SCHED_PRIV(thread).current_priority >> 3; /* 5bit */
RT_SCHED_PRIV(thread).number_mask = 1 << RT_SCHED_PRIV(thread).number;
RT_SCHED_PRIV(thread).high_mask = 1 << (RT_SCHED_PRIV(thread).current_priority & 0x07); /* 3bit */
#else
RT_SCHED_PRIV(thread).number_mask = 1 << RT_SCHED_PRIV(thread).current_priority;
#endif /* RT_THREAD_PRIORITY_MAX > 32 */
}
return RT_EOK;
}
#ifdef RT_USING_OVERFLOW_CHECK
void rt_scheduler_stack_check(struct rt_thread *thread)
{
RT_ASSERT(thread != RT_NULL);
#ifdef RT_USING_SMART
#ifndef ARCH_MM_MMU
struct rt_lwp *lwp = thread ? (struct rt_lwp *)thread->lwp : 0;
/* if stack pointer locate in user data section skip stack check. */
if (lwp && ((rt_uint32_t)thread->sp > (rt_uint32_t)lwp->data_entry &&
(rt_uint32_t)thread->sp <= (rt_uint32_t)lwp->data_entry + (rt_uint32_t)lwp->data_size))
{
return;
}
#endif /* not defined ARCH_MM_MMU */
#endif /* RT_USING_SMART */
#ifndef RT_USING_HW_STACK_GUARD
#ifdef ARCH_CPU_STACK_GROWS_UPWARD
if (*((rt_uint8_t *)((rt_uintptr_t)thread->stack_addr + thread->stack_size - 1)) != '#' ||
#else
if (*((rt_uint8_t *)thread->stack_addr) != '#' ||
#endif /* ARCH_CPU_STACK_GROWS_UPWARD */
(rt_uintptr_t)thread->sp <= (rt_uintptr_t)thread->stack_addr ||
(rt_uintptr_t)thread->sp >
(rt_uintptr_t)thread->stack_addr + (rt_uintptr_t)thread->stack_size)
{
rt_base_t dummy = 1;
LOG_E("thread:%s stack overflow\n", thread->parent.name);
while (dummy);
}
#endif /* RT_USING_HW_STACK_GUARD */
#ifdef ARCH_CPU_STACK_GROWS_UPWARD
#ifndef RT_USING_HW_STACK_GUARD
else if ((rt_uintptr_t)thread->sp > ((rt_uintptr_t)thread->stack_addr + thread->stack_size))
#else
if ((rt_uintptr_t)thread->sp > ((rt_uintptr_t)thread->stack_addr + thread->stack_size))
#endif
{
LOG_W("warning: %s stack is close to the top of stack address.\n",
thread->parent.name);
}
#else
#ifndef RT_USING_HW_STACK_GUARD
else if ((rt_uintptr_t)thread->sp <= ((rt_uintptr_t)thread->stack_addr + 32))
#else
if ((rt_uintptr_t)thread->sp <= ((rt_uintptr_t)thread->stack_addr + 32))
#endif
{
LOG_W("warning: %s stack is close to end of stack address.\n",
thread->parent.name);
}
#endif /* ARCH_CPU_STACK_GROWS_UPWARD */
}
#endif /* RT_USING_OVERFLOW_CHECK */
Loading...
马建仓 AI 助手
尝试更多
代码解读
代码找茬
代码优化
C
1
https://gitee.com/freeman_mayun_2019/rt-thread.git
git@gitee.com:freeman_mayun_2019/rt-thread.git
freeman_mayun_2019
rt-thread
rt-thread
master

搜索帮助