diff --git a/kernel/aster-nix/src/sched/priority_scheduler.rs b/kernel/aster-nix/src/sched/priority_scheduler.rs
index c4928b61..c769876d 100644
--- a/kernel/aster-nix/src/sched/priority_scheduler.rs
+++ b/kernel/aster-nix/src/sched/priority_scheduler.rs
@@ -1,56 +1,234 @@
 // SPDX-License-Identifier: MPL-2.0
 
-use intrusive_collections::LinkedList;
-use ostd::task::{set_scheduler, Scheduler, Task, TaskAdapter};
+use ostd::{
+    cpu::{num_cpus, this_cpu},
+    task::{
+        scheduler::{inject_scheduler, EnqueueFlags, LocalRunQueue, Scheduler, UpdateFlags},
+        AtomicCpuId, Priority, Task,
+    },
+};
 
 use crate::prelude::*;
 
 pub fn init() {
-    let preempt_scheduler = Box::new(PreemptScheduler::new());
-    let scheduler = Box::<PreemptScheduler>::leak(preempt_scheduler);
-    set_scheduler(scheduler);
+    let preempt_scheduler = Box::new(PreemptScheduler::default());
+    let scheduler = Box::<PreemptScheduler<Task>>::leak(preempt_scheduler);
+    inject_scheduler(scheduler);
 }
 
-/// The preempt scheduler
+/// The preempt scheduler.
 ///
-/// Real-time tasks are placed in the `real_time_tasks` queue and
+/// Real-time tasks are placed in the `real_time_entities` queue and
 /// are always prioritized during scheduling.
-/// Normal tasks are placed in the `normal_tasks` queue and are only
+/// Normal tasks are placed in the `normal_entities` queue and are only
 /// scheduled for execution when there are no real-time tasks.
-struct PreemptScheduler {
-    /// Tasks with a priority of less than 100 are regarded as real-time tasks.
-    real_time_tasks: SpinLock<LinkedList<TaskAdapter>>,
-    /// Tasks with a priority greater than or equal to 100 are regarded as normal tasks.
-    normal_tasks: SpinLock<LinkedList<TaskAdapter>>,
+struct PreemptScheduler<T: PreemptSchedInfo> {
+    rq: Vec<SpinLock<PreemptRunQueue<T>>>,
 }
 
-impl PreemptScheduler {
+impl<T: PreemptSchedInfo> PreemptScheduler<T> {
+    fn new(nr_cpus: u32) -> Self {
+        let mut rq = Vec::with_capacity(nr_cpus as usize);
+        for _ in 0..nr_cpus {
+            rq.push(SpinLock::new(PreemptRunQueue::new()));
+        }
+        Self { rq }
+    }
+
+    /// Selects a cpu for task to run on.
+    fn select_cpu(&self, _runnable: &Arc<T>) -> u32 {
+        // FIXME: adopt more reasonable policy once we fully enable SMP.
+        0
+    }
+}
+
+impl<T: Sync + Send + PreemptSchedInfo> Scheduler<T> for PreemptScheduler<T> {
+    fn enqueue(&self, runnable: Arc<T>, flags: EnqueueFlags) -> Option<u32> {
+        let mut still_in_rq = false;
+        let target_cpu = {
+            let mut cpu_id = self.select_cpu(&runnable);
+            if let Err(task_cpu_id) = runnable.cpu().set_if_is_none(cpu_id) {
+                debug_assert!(flags != EnqueueFlags::Spawn);
+                still_in_rq = true;
+                cpu_id = task_cpu_id;
+            }
+
+            cpu_id
+        };
+
+        let mut rq = self.rq[target_cpu as usize].lock_irq_disabled();
+        if still_in_rq && let Err(_) = runnable.cpu().set_if_is_none(target_cpu) {
+            return None;
+        }
+        let entity = PreemptSchedEntity::new(runnable);
+        if entity.is_real_time() {
+            rq.real_time_entities.push_back(entity);
+        } else {
+            rq.normal_entities.push_back(entity);
+        }
+
+        Some(target_cpu)
+    }
+
+    fn local_rq_with(&self, f: &mut dyn FnMut(&dyn LocalRunQueue<T>)) {
+        let local_rq: &PreemptRunQueue<T> = &self.rq[this_cpu() as usize].lock_irq_disabled();
+        f(local_rq);
+    }
+
+    fn local_mut_rq_with(&self, f: &mut dyn FnMut(&mut dyn LocalRunQueue<T>)) {
+        let local_rq: &mut PreemptRunQueue<T> =
+            &mut self.rq[this_cpu() as usize].lock_irq_disabled();
+        f(local_rq);
+    }
+}
+
+impl Default for PreemptScheduler<Task> {
+    fn default() -> Self {
+        Self::new(num_cpus())
+    }
+}
+
+struct PreemptRunQueue<T: PreemptSchedInfo> {
+    current: Option<PreemptSchedEntity<T>>,
+    real_time_entities: VecDeque<PreemptSchedEntity<T>>,
+    normal_entities: VecDeque<PreemptSchedEntity<T>>,
+}
+
+impl<T: PreemptSchedInfo> PreemptRunQueue<T> {
     pub fn new() -> Self {
         Self {
-            real_time_tasks: SpinLock::new(LinkedList::new(TaskAdapter::new())),
-            normal_tasks: SpinLock::new(LinkedList::new(TaskAdapter::new())),
+            current: None,
+            real_time_entities: VecDeque::new(),
+            normal_entities: VecDeque::new(),
         }
     }
 }
 
-impl Scheduler for PreemptScheduler {
-    fn enqueue(&self, task: Arc<Task>) {
-        if task.is_real_time() {
-            self.real_time_tasks.lock_irq_disabled().push_back(task);
-        } else {
-            self.normal_tasks.lock_irq_disabled().push_back(task);
+impl<T: Sync + Send + PreemptSchedInfo> LocalRunQueue<T> for PreemptRunQueue<T> {
+    fn current(&self) -> Option<&Arc<T>> {
+        self.current.as_ref().map(|entity| &entity.runnable)
+    }
+
+    fn update_current(&mut self, flags: UpdateFlags) -> bool {
+        match flags {
+            UpdateFlags::Tick => {
+                let Some(ref mut current_entity) = self.current else {
+                    return false;
+                };
+                current_entity.tick()
+                    || (!current_entity.is_real_time() && !self.real_time_entities.is_empty())
+            }
+            _ => true,
         }
     }
 
-    fn dequeue(&self) -> Option<Arc<Task>> {
-        if !self.real_time_tasks.lock_irq_disabled().is_empty() {
-            self.real_time_tasks.lock_irq_disabled().pop_front()
+    fn pick_next_current(&mut self) -> Option<&Arc<T>> {
+        let next_entity = if !self.real_time_entities.is_empty() {
+            self.real_time_entities.pop_front()
         } else {
-            self.normal_tasks.lock_irq_disabled().pop_front()
+            self.normal_entities.pop_front()
+        }?;
+        if let Some(prev_entity) = self.current.replace(next_entity) {
+            if prev_entity.is_real_time() {
+                self.real_time_entities.push_back(prev_entity);
+            } else {
+                self.normal_entities.push_back(prev_entity);
+            }
         }
+
+        Some(&self.current.as_ref().unwrap().runnable)
     }
 
-    fn should_preempt(&self, task: &Arc<Task>) -> bool {
-        !task.is_real_time() && !self.real_time_tasks.lock_irq_disabled().is_empty()
+    fn dequeue_current(&mut self) -> Option<Arc<T>> {
+        self.current.take().map(|entity| {
+            let runnable = entity.runnable;
+            runnable.cpu().set_to_none();
+
+            runnable
+        })
+    }
+}
+
+struct PreemptSchedEntity<T: PreemptSchedInfo> {
+    runnable: Arc<T>,
+    time_slice: TimeSlice,
+}
+
+impl<T: PreemptSchedInfo> PreemptSchedEntity<T> {
+    fn new(runnable: Arc<T>) -> Self {
+        Self {
+            runnable,
+            time_slice: TimeSlice::default(),
+        }
+    }
+
+    fn is_real_time(&self) -> bool {
+        self.runnable.is_real_time()
+    }
+
+    fn tick(&mut self) -> bool {
+        self.time_slice.elapse()
+    }
+}
+
+impl<T: PreemptSchedInfo> Clone for PreemptSchedEntity<T> {
+    fn clone(&self) -> Self {
+        Self {
+            runnable: self.runnable.clone(),
+            time_slice: self.time_slice,
+        }
+    }
+}
+
+#[derive(Clone, Copy)]
+pub struct TimeSlice {
+    elapsed_ticks: u32,
+}
+
+impl TimeSlice {
+    const DEFAULT_TIME_SLICE: u32 = 100;
+
+    pub const fn new() -> Self {
+        TimeSlice { elapsed_ticks: 0 }
+    }
+
+    pub fn elapse(&mut self) -> bool {
+        self.elapsed_ticks = (self.elapsed_ticks + 1) % Self::DEFAULT_TIME_SLICE;
+
+        self.elapsed_ticks == 0
+    }
+}
+
+impl Default for TimeSlice {
+    fn default() -> Self {
+        Self::new()
+    }
+}
+
+impl PreemptSchedInfo for Task {
+    type PRIORITY = Priority;
+
+    const REAL_TIME_TASK_PRIORITY: Self::PRIORITY = Priority::new(100);
+
+    fn priority(&self) -> Self::PRIORITY {
+        self.priority()
+    }
+
+    fn cpu(&self) -> &AtomicCpuId {
+        self.cpu()
+    }
+}
+
+trait PreemptSchedInfo {
+    type PRIORITY: Ord + PartialOrd + Eq + PartialEq;
+
+    const REAL_TIME_TASK_PRIORITY: Self::PRIORITY;
+
+    fn priority(&self) -> Self::PRIORITY;
+
+    fn cpu(&self) -> &AtomicCpuId;
+
+    fn is_real_time(&self) -> bool {
+        self.priority() < Self::REAL_TIME_TASK_PRIORITY
     }
 }
diff --git a/kernel/aster-nix/src/thread/task.rs b/kernel/aster-nix/src/thread/task.rs
index 63692339..d7b6ed3c 100644
--- a/kernel/aster-nix/src/thread/task.rs
+++ b/kernel/aster-nix/src/thread/task.rs
@@ -1,7 +1,7 @@
 // SPDX-License-Identifier: MPL-2.0
 
 use ostd::{
-    task::{preempt, Task, TaskOptions},
+    task::{Task, TaskOptions},
     user::{ReturnReason, UserContextApi, UserMode, UserSpace},
 };
 
@@ -84,8 +84,6 @@ pub fn create_new_user_task(user_space: Arc<UserSpace>, thread_ref: Weak<Thread>
                 debug!("exit due to signal");
                 break;
             }
-            // a preemption point after handling user event.
-            preempt(current_task);
         }
         debug!("exit user loop");
     }
diff --git a/kernel/aster-nix/src/thread/work_queue/worker_pool.rs b/kernel/aster-nix/src/thread/work_queue/worker_pool.rs
index 77ad2f5a..4dfe9aa2 100644
--- a/kernel/aster-nix/src/thread/work_queue/worker_pool.rs
+++ b/kernel/aster-nix/src/thread/work_queue/worker_pool.rs
@@ -7,11 +7,7 @@ use core::{
     time::Duration,
 };
 
-use ostd::{
-    cpu::CpuSet,
-    sync::WaitQueue,
-    task::{add_task, Priority},
-};
+use ostd::{cpu::CpuSet, sync::WaitQueue, task::Priority};
 
 use super::{simple_scheduler::SimpleScheduler, worker::Worker, WorkItem, WorkPriority, WorkQueue};
 use crate::{
@@ -81,7 +77,7 @@ impl LocalWorkerPool {
     fn add_worker(&self) {
         let worker = Worker::new(self.parent.clone(), self.cpu_id);
         self.workers.lock_irq_disabled().push_back(worker.clone());
-        add_task(worker.bound_thread().task().clone());
+        worker.bound_thread().run();
     }
 
     fn remove_worker(&self) {
diff --git a/ostd/src/arch/x86/cpu/mod.rs b/ostd/src/arch/x86/cpu/mod.rs
index b0c2f1ab..20cacf6c 100644
--- a/ostd/src/arch/x86/cpu/mod.rs
+++ b/ostd/src/arch/x86/cpu/mod.rs
@@ -17,6 +17,7 @@ use trapframe::UserContext as RawUserContext;
 use x86_64::registers::rflags::RFlags;
 
 use crate::{
+    task::scheduler,
     trap::call_irq_callback_functions,
     user::{ReturnReason, UserContextApi, UserContextApiInternal},
 };
@@ -50,32 +51,6 @@ pub struct CpuExceptionInfo {
     pub page_fault_addr: usize,
 }
 
-/// User Preemption.
-pub struct UserPreemption {
-    count: u32,
-}
-
-impl UserPreemption {
-    const PREEMPTION_INTERVAL: u32 = 100;
-
-    /// Creates a new instance of `UserPreemption`.
-    #[allow(clippy::new_without_default)]
-    pub const fn new() -> Self {
-        UserPreemption { count: 0 }
-    }
-
-    /// Checks if preemption might occur and takes necessary actions.
-    pub fn might_preempt(&mut self) {
-        self.count = (self.count + 1) % Self::PREEMPTION_INTERVAL;
-
-        if self.count == 0 {
-            crate::arch::irq::enable_local();
-            crate::task::schedule();
-            crate::arch::irq::disable_local();
-        }
-    }
-}
-
 impl UserContext {
     /// Returns a reference to the general registers.
     pub fn general_regs(&self) -> &RawGeneralRegs {
@@ -115,9 +90,9 @@ impl UserContextApiInternal for UserContext {
         let return_reason: ReturnReason;
         const SYSCALL_TRAPNUM: u16 = 0x100;
 
-        let mut user_preemption = UserPreemption::new();
         // return when it is syscall or cpu exception type is Fault or Trap.
         loop {
+            scheduler::might_preempt();
             self.user_context.run();
             match CpuException::to_cpu_exception(self.user_context.trap_num as u16) {
                 Some(exception) => {
@@ -146,8 +121,6 @@ impl UserContextApiInternal for UserContext {
                 return_reason = ReturnReason::KernelEvent;
                 break;
             }
-
-            user_preemption.might_preempt();
         }
 
         crate::arch::irq::enable_local();
diff --git a/ostd/src/boot/mod.rs b/ostd/src/boot/mod.rs
index 941f92ea..56f2ede7 100644
--- a/ostd/src/boot/mod.rs
+++ b/ostd/src/boot/mod.rs
@@ -129,7 +129,7 @@ pub fn call_ostd_main() -> ! {
     unsafe {
         use alloc::boxed::Box;
 
-        use crate::task::{set_scheduler, FifoScheduler, Scheduler, TaskOptions};
+        use crate::task::TaskOptions;
 
         crate::init();
         // The whitelists that will be generated by OSDK runner as static consts.
@@ -137,10 +137,6 @@ pub fn call_ostd_main() -> ! {
             static KTEST_TEST_WHITELIST: Option<&'static [&'static str]>;
             static KTEST_CRATE_WHITELIST: Option<&'static [&'static str]>;
         }
-        // Set the global scheduler a FIFO scheduler.
-        let simple_scheduler = Box::new(FifoScheduler::new());
-        let static_scheduler: &'static dyn Scheduler = Box::leak(simple_scheduler);
-        set_scheduler(static_scheduler);
 
         let test_task = move || {
             run_ktests(KTEST_TEST_WHITELIST, KTEST_CRATE_WHITELIST);
diff --git a/ostd/src/cpu/local/mod.rs b/ostd/src/cpu/local/mod.rs
index 6e9c7b16..04b981ad 100644
--- a/ostd/src/cpu/local/mod.rs
+++ b/ostd/src/cpu/local/mod.rs
@@ -60,19 +60,20 @@ extern "C" {
 }
 
 cpu_local_cell! {
-    /// The count of the preempt lock.
+    /// A 4-byte preemption information consisting of a should_preempt flag at
+    /// the highest bit and a preemption counter in the lower 31 bits.
     ///
-    /// We need to access the preemption count before we can copy the section
-    /// for application processors. So, the preemption count is not copied from
+    /// We need to access the preemption info before we can copy the section
+    /// for application processors. So, the preemption info is not copied from
     /// bootstrap processor's section as the initialization. Instead it is
     /// initialized to zero for application processors.
-    pub(crate) static PREEMPT_LOCK_COUNT: u32 = 0;
+    pub(crate) static PREEMPT_INFO: u32 = 0;
 }
 
 /// Sets the base address of the CPU-local storage for the bootstrap processor.
 ///
 /// It should be called early to let [`crate::task::disable_preempt`] work,
-/// which needs to update a CPU-local preempt lock count. Otherwise it may
+/// which needs to update a CPU-local preemption info. Otherwise it may
 /// panic when calling [`crate::task::disable_preempt`].
 ///
 /// # Safety
@@ -133,16 +134,16 @@ pub unsafe fn init_on_bsp() {
             (ap_pages_ptr as *mut u32).write(cpu_i);
         }
 
-        // SAFETY: the `PREEMPT_LOCK_COUNT` may be dirty on the BSP, so we need
+        // SAFETY: the `PREEMPT_INFO` may be dirty on the BSP, so we need
         // to ensure that it is initialized to zero for APs. The safety
         // requirements are met since the static is defined in the `.cpu_local`
         // section and the pointer to that static is the offset in the CPU-
         // local area. It is a `usize` so it is safe to be overwritten.
         unsafe {
-            let preempt_count_ptr = &PREEMPT_LOCK_COUNT as *const _ as usize;
-            let preempt_count_offset = preempt_count_ptr - __cpu_local_start as usize;
-            let ap_preempt_count_ptr = ap_pages_ptr.add(preempt_count_offset) as *mut usize;
-            ap_preempt_count_ptr.write(0);
+            let preempt_info_ptr = &PREEMPT_INFO as *const _ as usize;
+            let preempt_info_offset = preempt_info_ptr - __cpu_local_start as usize;
+            let ap_preempt_info_ptr = ap_pages_ptr.add(preempt_info_offset) as *mut usize;
+            ap_preempt_info_ptr.write(0);
         }
 
         // SAFETY: bytes `8:16` are reserved for storing the pointer to the
diff --git a/ostd/src/sync/wait.rs b/ostd/src/sync/wait.rs
index 81cd950e..34b4a1c2 100644
--- a/ostd/src/sync/wait.rs
+++ b/ostd/src/sync/wait.rs
@@ -4,7 +4,7 @@ use alloc::{collections::VecDeque, sync::Arc};
 use core::sync::atomic::{AtomicBool, AtomicU32, Ordering};
 
 use super::SpinLock;
-use crate::task::{add_task, schedule, Task, TaskStatus};
+use crate::task::{scheduler, Task};
 
 // # Explanation on the memory orders
 //
@@ -255,36 +255,15 @@ impl Waker {
         if self.has_woken.swap(true, Ordering::Release) {
             return false;
         }
-
-        let mut task = self.task.inner_exclusive_access();
-        match task.task_status {
-            TaskStatus::Sleepy => {
-                task.task_status = TaskStatus::Runnable;
-            }
-            TaskStatus::Sleeping => {
-                task.task_status = TaskStatus::Runnable;
-
-                // Avoid holding the lock when doing `add_task`
-                drop(task);
-                add_task(self.task.clone());
-            }
-            _ => (),
-        }
+        scheduler::unpark_target(self.task.clone());
 
         true
     }
 
     fn do_wait(&self) {
-        while !self.has_woken.swap(false, Ordering::Acquire) {
-            let mut task = self.task.inner_exclusive_access();
-            // After holding the lock, check again to avoid races
-            if self.has_woken.swap(false, Ordering::Acquire) {
-                break;
-            }
-            task.task_status = TaskStatus::Sleepy;
-            drop(task);
-
-            schedule();
+        let has_woken = &self.has_woken;
+        while !has_woken.swap(false, Ordering::Acquire) {
+            scheduler::park_current(has_woken);
         }
     }
 
diff --git a/ostd/src/task/mod.rs b/ostd/src/task/mod.rs
index 61b76cd9..5b97c6a7 100644
--- a/ostd/src/task/mod.rs
+++ b/ostd/src/task/mod.rs
@@ -2,15 +2,12 @@
 
 //! Tasks are the unit of code execution.
 
-mod priority;
 mod processor;
-mod scheduler;
+pub mod scheduler;
 #[allow(clippy::module_inception)]
 mod task;
 
 pub use self::{
-    priority::Priority,
-    processor::{disable_preempt, preempt, schedule, DisablePreemptGuard},
-    scheduler::{add_task, set_scheduler, FifoScheduler, Scheduler},
-    task::{Task, TaskAdapter, TaskContextApi, TaskOptions, TaskStatus},
+    processor::{disable_preempt, DisablePreemptGuard},
+    task::{AtomicCpuId, Priority, Task, TaskAdapter, TaskContextApi, TaskOptions},
 };
diff --git a/ostd/src/task/processor.rs b/ostd/src/task/processor.rs
index 48471908..9412a295 100644
--- a/ostd/src/task/processor.rs
+++ b/ostd/src/task/processor.rs
@@ -2,12 +2,8 @@
 
 use alloc::sync::Arc;
 
-use super::{
-    scheduler::{fetch_task, GLOBAL_SCHEDULER},
-    task::{context_switch, TaskContext},
-    Task, TaskStatus,
-};
-use crate::{cpu::local::PREEMPT_LOCK_COUNT, cpu_local_cell};
+use super::task::{context_switch, Task, TaskContext};
+use crate::{cpu::local::PREEMPT_INFO, cpu_local_cell};
 
 cpu_local_cell! {
     /// The `Arc<Task>` (casted by [`Arc::into_raw`]) that is the current task.
@@ -37,52 +33,19 @@ pub(super) fn current_task() -> Option<Arc<Task>> {
     Some(restored)
 }
 
-/// Calls this function to switch to other task by using GLOBAL_SCHEDULER
-pub fn schedule() {
-    if let Some(task) = fetch_task() {
-        switch_to_task(task);
-    }
-}
-
-/// Preempts the `task`.
-///
-/// TODO: This interface of this method is error prone.
-/// The method takes an argument for the current task to optimize its efficiency,
-/// but the argument provided by the caller may not be the current task, really.
-/// Thus, this method should be removed or reworked in the future.
-pub fn preempt(task: &Arc<Task>) {
-    // TODO: Refactor `preempt` and `schedule`
-    // after the Atomic mode and `might_break` is enabled.
-    let mut scheduler = GLOBAL_SCHEDULER.lock_irq_disabled();
-    if !scheduler.should_preempt(task) {
-        return;
-    }
-    let Some(next_task) = scheduler.dequeue() else {
-        return;
-    };
-    drop(scheduler);
-    switch_to_task(next_task);
-}
-
 /// Calls this function to switch to other task
 ///
 /// If current task is none, then it will use the default task context and it
 /// will not return to this function again.
 ///
-/// If the current task's status not [`TaskStatus::Runnable`], it will not be
-/// added to the scheduler.
-///
 /// # Panics
 ///
 /// This function will panic if called while holding preemption locks or with
 /// local IRQ disabled.
-fn switch_to_task(next_task: Arc<Task>) {
-    let preemt_lock_count = PREEMPT_LOCK_COUNT.load();
-    if preemt_lock_count != 0 {
-        panic!(
-            "Calling schedule() while holding {} locks",
-            preemt_lock_count
-        );
+pub(super) fn switch_to_task(next_task: Arc<Task>) {
+    let preemt_count = PREEMPT_COUNT.get();
+    if preemt_count != 0 {
+        panic!("Switching task while holding {} locks", preemt_count);
     }
 
     assert!(
@@ -93,7 +56,6 @@ fn switch_to_task(next_task: Arc<Task>) {
     let irq_guard = crate::trap::disable_local();
 
     let current_task_ptr = CURRENT_TASK_PTR.load();
-
     let current_task_ctx_ptr = if current_task_ptr.is_null() {
         // SAFETY: Interrupts are disabled, so the pointer is safe to be fetched.
         unsafe { BOOTSTRAP_CONTEXT.as_ptr_mut() }
@@ -104,24 +66,12 @@ fn switch_to_task(next_task: Arc<Task>) {
             let _ = core::mem::ManuallyDrop::new(restored.clone());
             restored
         };
-
         let ctx_ptr = cur_task_arc.ctx().get();
 
-        let mut task_inner = cur_task_arc.inner_exclusive_access();
-
-        debug_assert_ne!(task_inner.task_status, TaskStatus::Sleeping);
-        if task_inner.task_status == TaskStatus::Runnable {
-            drop(task_inner);
-            GLOBAL_SCHEDULER.lock().enqueue(cur_task_arc);
-        } else if task_inner.task_status == TaskStatus::Sleepy {
-            task_inner.task_status = TaskStatus::Sleeping;
-        }
-
         ctx_ptr
     };
 
     let next_task_ctx_ptr = next_task.ctx().get().cast_const();
-
     if let Some(next_user_space) = next_task.user_space() {
         next_user_space.vm_space().activate();
     }
@@ -144,7 +94,7 @@ fn switch_to_task(next_task: Arc<Task>) {
     drop(irq_guard);
 
     // SAFETY:
-    // 1. `ctx` is only used in `schedule()`. We have exclusive access to both the current task
+    // 1. `ctx` is only used in `reschedule()`. We have exclusive access to both the current task
     //    context and the next task context.
     // 2. The next task context is a valid task context.
     unsafe {
@@ -159,6 +109,34 @@ fn switch_to_task(next_task: Arc<Task>) {
     // to the next task switching.
 }
 
+static PREEMPT_COUNT: PreemptCount = PreemptCount::new();
+
+struct PreemptCount {}
+
+impl PreemptCount {
+    const SHIFT: u8 = 0;
+
+    const BITS: u8 = 31;
+
+    const MASK: u32 = ((1 << Self::BITS) - 1) << Self::SHIFT;
+
+    const fn new() -> Self {
+        Self {}
+    }
+
+    fn inc(&self) {
+        PREEMPT_INFO.add_assign(1 << Self::SHIFT);
+    }
+
+    fn dec(&self) {
+        PREEMPT_INFO.sub_assign(1 << Self::SHIFT);
+    }
+
+    fn get(&self) -> u32 {
+        PREEMPT_INFO.load() & Self::MASK
+    }
+}
+
 /// A guard for disable preempt.
 #[clippy::has_significant_drop]
 #[must_use]
@@ -171,7 +149,7 @@ impl !Send for DisablePreemptGuard {}
 
 impl DisablePreemptGuard {
     fn new() -> Self {
-        PREEMPT_LOCK_COUNT.add_assign(1);
+        PREEMPT_COUNT.inc();
         Self { _private: () }
     }
 
@@ -184,7 +162,7 @@ impl DisablePreemptGuard {
 
 impl Drop for DisablePreemptGuard {
     fn drop(&mut self) {
-        PREEMPT_LOCK_COUNT.sub_assign(1);
+        PREEMPT_COUNT.dec();
     }
 }
 
diff --git a/ostd/src/task/scheduler.rs b/ostd/src/task/scheduler.rs
deleted file mode 100644
index aada4154..00000000
--- a/ostd/src/task/scheduler.rs
+++ /dev/null
@@ -1,107 +0,0 @@
-// SPDX-License-Identifier: MPL-2.0
-
-#![allow(dead_code)]
-
-use alloc::collections::VecDeque;
-
-use crate::{prelude::*, sync::SpinLock, task::Task};
-
-static DEFAULT_SCHEDULER: FifoScheduler = FifoScheduler::new();
-pub(crate) static GLOBAL_SCHEDULER: SpinLock<GlobalScheduler> = SpinLock::new(GlobalScheduler {
-    scheduler: &DEFAULT_SCHEDULER,
-});
-
-/// A scheduler for tasks.
-///
-/// An implementation of scheduler can attach scheduler-related information
-/// with the `TypeMap` returned from `task.data()`.
-pub trait Scheduler: Sync + Send {
-    /// Enqueues a task to the scheduler.
-    fn enqueue(&self, task: Arc<Task>);
-
-    /// Dequeues a task from the scheduler.
-    fn dequeue(&self) -> Option<Arc<Task>>;
-
-    /// Tells whether the given task should be preempted by other tasks in the queue.
-    fn should_preempt(&self, task: &Arc<Task>) -> bool;
-}
-
-pub struct GlobalScheduler {
-    scheduler: &'static dyn Scheduler,
-}
-
-impl GlobalScheduler {
-    pub const fn new(scheduler: &'static dyn Scheduler) -> Self {
-        Self { scheduler }
-    }
-
-    /// dequeue a task using scheduler
-    /// require the scheduler is not none
-    pub fn dequeue(&mut self) -> Option<Arc<Task>> {
-        self.scheduler.dequeue()
-    }
-    /// enqueue a task using scheduler
-    /// require the scheduler is not none
-    pub fn enqueue(&mut self, task: Arc<Task>) {
-        self.scheduler.enqueue(task)
-    }
-
-    pub fn should_preempt(&self, task: &Arc<Task>) -> bool {
-        self.scheduler.should_preempt(task)
-    }
-}
-/// Sets the global task scheduler.
-///
-/// This must be called before invoking `Task::spawn`.
-pub fn set_scheduler(scheduler: &'static dyn Scheduler) {
-    let mut global_scheduler = GLOBAL_SCHEDULER.lock_irq_disabled();
-    // When setting a new scheduler, the old scheduler should be empty
-    assert!(global_scheduler.dequeue().is_none());
-    global_scheduler.scheduler = scheduler;
-}
-
-pub fn fetch_task() -> Option<Arc<Task>> {
-    GLOBAL_SCHEDULER.lock_irq_disabled().dequeue()
-}
-
-/// Adds a task to the global scheduler.
-pub fn add_task(task: Arc<Task>) {
-    GLOBAL_SCHEDULER.lock_irq_disabled().enqueue(task);
-}
-
-/// A simple FIFO (First-In-First-Out) task scheduler.
-pub struct FifoScheduler {
-    /// A thread-safe queue to hold tasks waiting to be executed.
-    task_queue: SpinLock<VecDeque<Arc<Task>>>,
-}
-
-impl FifoScheduler {
-    /// Creates a new instance of `FifoScheduler`.
-    pub const fn new() -> Self {
-        FifoScheduler {
-            task_queue: SpinLock::new(VecDeque::new()),
-        }
-    }
-}
-
-impl Default for FifoScheduler {
-    fn default() -> Self {
-        Self::new()
-    }
-}
-
-impl Scheduler for FifoScheduler {
-    /// Enqueues a task to the end of the queue.
-    fn enqueue(&self, task: Arc<Task>) {
-        self.task_queue.lock_irq_disabled().push_back(task);
-    }
-    /// Dequeues a task from the front of the queue, if any.
-    fn dequeue(&self) -> Option<Arc<Task>> {
-        self.task_queue.lock_irq_disabled().pop_front()
-    }
-    /// In this simple implementation, task preemption is not supported.
-    /// Once a task starts running, it will continue to run until completion.
-    fn should_preempt(&self, _task: &Arc<Task>) -> bool {
-        false
-    }
-}
diff --git a/ostd/src/task/scheduler/fifo_scheduler.rs b/ostd/src/task/scheduler/fifo_scheduler.rs
new file mode 100644
index 00000000..a153e9b5
--- /dev/null
+++ b/ostd/src/task/scheduler/fifo_scheduler.rs
@@ -0,0 +1,125 @@
+// SPDX-License-Identifier: MPL-2.0
+
+use alloc::{boxed::Box, collections::VecDeque, sync::Arc, vec::Vec};
+
+use super::{inject_scheduler, EnqueueFlags, LocalRunQueue, Scheduler, UpdateFlags};
+use crate::{
+    cpu::{num_cpus, this_cpu},
+    sync::SpinLock,
+    task::{AtomicCpuId, Task},
+};
+
+pub fn init() {
+    let fifo_scheduler = Box::new(FifoScheduler::default());
+    let scheduler = Box::<FifoScheduler<Task>>::leak(fifo_scheduler);
+    inject_scheduler(scheduler);
+}
+
+/// A simple FIFO (First-In-First-Out) task scheduler.
+struct FifoScheduler<T: FifoSchedInfo> {
+    /// A thread-safe queue to hold tasks waiting to be executed.
+    rq: Vec<SpinLock<FifoRunQueue<T>>>,
+}
+
+impl<T: FifoSchedInfo> FifoScheduler<T> {
+    /// Creates a new instance of `FifoScheduler`.
+    fn new(nr_cpus: u32) -> Self {
+        let mut rq = Vec::new();
+        for _ in 0..nr_cpus {
+            rq.push(SpinLock::new(FifoRunQueue::new()));
+        }
+        Self { rq }
+    }
+
+    fn select_cpu(&self) -> u32 {
+        // FIXME: adopt more reasonable policy once we fully enable SMP.
+        0
+    }
+}
+
+impl<T: FifoSchedInfo + Send + Sync> Scheduler<T> for FifoScheduler<T> {
+    fn enqueue(&self, runnable: Arc<T>, flags: EnqueueFlags) -> Option<u32> {
+        let mut still_in_rq = false;
+        let target_cpu = {
+            let mut cpu_id = self.select_cpu();
+            if let Err(task_cpu_id) = runnable.cpu().set_if_is_none(cpu_id) {
+                debug_assert!(flags != EnqueueFlags::Spawn);
+                still_in_rq = true;
+                cpu_id = task_cpu_id;
+            }
+
+            cpu_id
+        };
+
+        let mut rq = self.rq[target_cpu as usize].lock_irq_disabled();
+        if still_in_rq && let Err(_) = runnable.cpu().set_if_is_none(target_cpu) {
+            return None;
+        }
+        rq.queue.push_back(runnable);
+
+        Some(target_cpu)
+    }
+
+    fn local_rq_with(&self, f: &mut dyn FnMut(&dyn LocalRunQueue<T>)) {
+        let local_rq: &FifoRunQueue<T> = &self.rq[this_cpu() as usize].lock_irq_disabled();
+        f(local_rq);
+    }
+
+    fn local_mut_rq_with(&self, f: &mut dyn FnMut(&mut dyn LocalRunQueue<T>)) {
+        let local_rq: &mut FifoRunQueue<T> = &mut self.rq[this_cpu() as usize].lock_irq_disabled();
+        f(local_rq);
+    }
+}
+
+struct FifoRunQueue<T: FifoSchedInfo> {
+    current: Option<Arc<T>>,
+    queue: VecDeque<Arc<T>>,
+}
+
+impl<T: FifoSchedInfo> FifoRunQueue<T> {
+    pub const fn new() -> Self {
+        Self {
+            current: None,
+            queue: VecDeque::new(),
+        }
+    }
+}
+
+impl<T: FifoSchedInfo> LocalRunQueue<T> for FifoRunQueue<T> {
+    fn current(&self) -> Option<&Arc<T>> {
+        self.current.as_ref()
+    }
+
+    fn update_current(&mut self, flags: super::UpdateFlags) -> bool {
+        !matches!(flags, UpdateFlags::Tick)
+    }
+
+    fn pick_next_current(&mut self) -> Option<&Arc<T>> {
+        let next_task = self.queue.pop_front()?;
+        if let Some(prev_task) = self.current.replace(next_task) {
+            self.queue.push_back(prev_task);
+        }
+
+        self.current.as_ref()
+    }
+
+    fn dequeue_current(&mut self) -> Option<Arc<T>> {
+        self.current.take().inspect(|task| task.cpu().set_to_none())
+    }
+}
+
+impl Default for FifoScheduler<Task> {
+    fn default() -> Self {
+        Self::new(num_cpus())
+    }
+}
+
+impl FifoSchedInfo for Task {
+    fn cpu(&self) -> &AtomicCpuId {
+        self.cpu()
+    }
+}
+
+trait FifoSchedInfo {
+    fn cpu(&self) -> &AtomicCpuId;
+}
diff --git a/ostd/src/task/scheduler/mod.rs b/ostd/src/task/scheduler/mod.rs
new file mode 100644
index 00000000..15f81770
--- /dev/null
+++ b/ostd/src/task/scheduler/mod.rs
@@ -0,0 +1,271 @@
+// SPDX-License-Identifier: MPL-2.0
+
+//! Scheduling subsystem (in-OSTD part).
+//!
+//! This module defines what OSTD expects from a scheduling implementation
+//! and provides useful functions for controlling the execution flow.
+
+mod fifo_scheduler;
+
+use core::sync::atomic::{AtomicBool, Ordering};
+
+use spin::Once;
+
+use super::{processor, task::Task};
+use crate::{
+    arch::timer,
+    cpu::{local::PREEMPT_INFO, this_cpu},
+    prelude::*,
+};
+
+/// Injects a scheduler implementation into framework.
+///
+/// This function can only be called once and must be called during the initialization of kernel.
+pub fn inject_scheduler(scheduler: &'static dyn Scheduler<Task>) {
+    SCHEDULER.call_once(|| scheduler);
+
+    timer::register_callback(|| {
+        SCHEDULER.get().unwrap().local_mut_rq_with(&mut |local_rq| {
+            if local_rq.update_current(UpdateFlags::Tick) {
+                SHOULD_PREEMPT.set();
+            }
+        })
+    });
+}
+
+static SCHEDULER: Once<&'static dyn Scheduler<Task>> = Once::new();
+
+/// A per-CPU task scheduler.
+pub trait Scheduler<T = Task>: Sync + Send {
+    /// Enqueues a runnable task.
+    ///
+    /// Scheduler developers can perform load-balancing or some accounting work here.
+    ///
+    /// If the `current` of a CPU should be preempted, this method returns the id of
+    /// that CPU.
+    fn enqueue(&self, runnable: Arc<T>, flags: EnqueueFlags) -> Option<u32>;
+
+    /// Gets an immutable access to the local runqueue of the current CPU core.
+    fn local_rq_with(&self, f: &mut dyn FnMut(&dyn LocalRunQueue<T>));
+
+    /// Gets a mutable access to the local runqueue of the current CPU core.
+    fn local_mut_rq_with(&self, f: &mut dyn FnMut(&mut dyn LocalRunQueue<T>));
+}
+
+/// The _local_ view of a per-CPU runqueue.
+///
+/// This local view provides the interface for the runqueue of a CPU core
+/// to be inspected and manipulated by the code running on this particular CPU core.
+///
+/// Conceptually, a local runqueue consists of two parts:
+/// (1) a priority queue of runnable tasks;
+/// (2) the current running task.
+/// The exact definition of "priority" is left for the concrete implementation to decide.
+pub trait LocalRunQueue<T = Task> {
+    /// Gets the current runnable task.
+    fn current(&self) -> Option<&Arc<T>>;
+
+    /// Updates the current runnable task's scheduling statistics and potentially its
+    /// position in the queue.
+    ///
+    /// If the current runnable task should be preempted, the method returns `true`.
+    fn update_current(&mut self, flags: UpdateFlags) -> bool;
+
+    /// Picks the next current runnable task.
+    ///
+    /// This method returns the chosen next current runnable task. If there is no
+    /// candidate for next current runnable task, this method returns `None`.
+    fn pick_next_current(&mut self) -> Option<&Arc<T>>;
+
+    /// Removes the current runnable task from runqueue.
+    ///
+    /// This method returns the current runnable task. If there is no current runnable
+    /// task, this method returns `None`.
+    fn dequeue_current(&mut self) -> Option<Arc<T>>;
+}
+
+/// Possible triggers of an `enqueue` action.
+#[derive(PartialEq, Copy, Clone)]
+pub enum EnqueueFlags {
+    /// Spawn a new task.
+    Spawn,
+    /// Wake a sleeping task.
+    Wake,
+}
+
+/// Possible triggers of an `update_current` action.
+#[derive(PartialEq, Copy, Clone)]
+pub enum UpdateFlags {
+    /// Timer interrupt.
+    Tick,
+    /// Task waiting.
+    Wait,
+    /// Task yielding.
+    Yield,
+}
+
+/// Preempts the current task.
+pub(crate) fn might_preempt() {
+    fn preempt_check() -> bool {
+        PREEMPT_INFO.load() == 0
+    }
+
+    if !preempt_check() {
+        return;
+    }
+    yield_now();
+}
+
+/// Blocks the current task unless `has_woken` is `true`.
+pub(crate) fn park_current(has_woken: &AtomicBool) {
+    let mut current = None;
+    let mut is_first_try = true;
+    reschedule(&mut |local_rq: &mut dyn LocalRunQueue| {
+        if is_first_try {
+            if has_woken.load(Ordering::Acquire) {
+                return ReschedAction::DoNothing;
+            }
+            current = local_rq.dequeue_current();
+            local_rq.update_current(UpdateFlags::Wait);
+        }
+        if let Some(next_task) = local_rq.pick_next_current() {
+            if Arc::ptr_eq(current.as_ref().unwrap(), next_task) {
+                return ReschedAction::DoNothing;
+            }
+            ReschedAction::SwitchTo(next_task.clone())
+        } else {
+            is_first_try = false;
+            ReschedAction::Retry
+        }
+    });
+}
+
+/// Unblocks a target task.
+pub(crate) fn unpark_target(runnable: Arc<Task>) {
+    let should_preempt_info = SCHEDULER
+        .get()
+        .unwrap()
+        .enqueue(runnable, EnqueueFlags::Wake);
+    if should_preempt_info.is_some() {
+        let cpu_id = should_preempt_info.unwrap();
+        // FIXME: send IPI to set remote CPU's `SHOULD_PREEMPT` if needed.
+        if cpu_id == this_cpu() {
+            SHOULD_PREEMPT.set();
+        }
+    }
+}
+
+/// Enqueues a newly built task.
+///
+/// Note that the new task is not guranteed to run at once.
+pub(super) fn run_new_task(runnable: Arc<Task>) {
+    // FIXME: remove this check for `SCHEDULER`.
+    // Currently OSTD cannot know whether its user has injected a scheduler.
+    if !SCHEDULER.is_completed() {
+        fifo_scheduler::init();
+    }
+
+    let should_preempt_info = SCHEDULER
+        .get()
+        .unwrap()
+        .enqueue(runnable, EnqueueFlags::Spawn);
+    if should_preempt_info.is_some() {
+        let cpu_id = should_preempt_info.unwrap();
+        // FIXME: send IPI to set remote CPU's `SHOULD_PREEMPT` if needed.
+        if cpu_id == this_cpu() {
+            SHOULD_PREEMPT.set();
+        }
+    }
+
+    might_preempt();
+}
+
+/// Dequeues the current task from its runqueue.
+///
+/// This should only be called if the current is to exit.
+pub(super) fn exit_current() {
+    reschedule(&mut |local_rq: &mut dyn LocalRunQueue| {
+        let _ = local_rq.dequeue_current();
+        if let Some(next_task) = local_rq.pick_next_current() {
+            ReschedAction::SwitchTo(next_task.clone())
+        } else {
+            ReschedAction::Retry
+        }
+    })
+}
+
+/// Yields execution.
+pub(super) fn yield_now() {
+    reschedule(&mut |local_rq| {
+        local_rq.update_current(UpdateFlags::Yield);
+
+        if let Some(next_task) = local_rq.pick_next_current() {
+            ReschedAction::SwitchTo(next_task.clone())
+        } else {
+            ReschedAction::DoNothing
+        }
+    })
+}
+
+/// Do rescheduling by acting on the scheduling decision (`ReschedAction`) made by a
+/// user-given closure.
+///
+/// The closure makes the scheduling decision by taking the local runqueue has its input.
+fn reschedule<F>(f: &mut F)
+where
+    F: FnMut(&mut dyn LocalRunQueue) -> ReschedAction,
+{
+    let next_task = loop {
+        let mut action = ReschedAction::DoNothing;
+        SCHEDULER.get().unwrap().local_mut_rq_with(&mut |rq| {
+            action = f(rq);
+        });
+
+        match action {
+            ReschedAction::DoNothing => {
+                return;
+            }
+            ReschedAction::Retry => {
+                continue;
+            }
+            ReschedAction::SwitchTo(next_task) => {
+                break next_task;
+            }
+        };
+    };
+
+    SHOULD_PREEMPT.clear();
+    processor::switch_to_task(next_task);
+}
+
+/// Possible actions of a rescheduling.
+enum ReschedAction {
+    /// Keep running current task and do nothing.
+    DoNothing,
+    /// Loop until finding a task to swap out the current.
+    Retry,
+    /// Switch to target task.
+    SwitchTo(Arc<Task>),
+}
+
+static SHOULD_PREEMPT: ShouldPreemptFlag = ShouldPreemptFlag::new();
+
+struct ShouldPreemptFlag {}
+
+impl ShouldPreemptFlag {
+    const SHIFT: u8 = 31;
+
+    const MASK: u32 = 1 << Self::SHIFT;
+
+    const fn new() -> Self {
+        Self {}
+    }
+
+    fn set(&self) {
+        PREEMPT_INFO.bitand_assign(!Self::MASK);
+    }
+
+    fn clear(&self) {
+        PREEMPT_INFO.bitor_assign(Self::MASK);
+    }
+}
diff --git a/ostd/src/task/task.rs b/ostd/src/task/task/mod.rs
similarity index 85%
rename from ostd/src/task/task.rs
rename to ostd/src/task/task/mod.rs
index e875678f..4d5a2390 100644
--- a/ostd/src/task/task.rs
+++ b/ostd/src/task/task/mod.rs
@@ -4,22 +4,23 @@
 // So we temporary allow missing_docs for this module.
 #![allow(missing_docs)]
 
-use alloc::{boxed::Box, sync::Arc};
-use core::{any::Any, cell::UnsafeCell};
+mod priority;
+
+use core::{
+    any::Any,
+    cell::UnsafeCell,
+    sync::atomic::{AtomicU32, Ordering},
+};
 
 use intrusive_collections::{intrusive_adapter, LinkedListAtomicLink};
+pub use priority::Priority;
 
-use super::{
-    add_task,
-    priority::Priority,
-    processor::{current_task, schedule},
-};
+use super::{processor::current_task, scheduler};
 pub(crate) use crate::arch::task::{context_switch, TaskContext};
 use crate::{
     cpu::CpuSet,
     mm::{kspace::KERNEL_PAGE_TABLE, FrameAllocOptions, Paddr, PageFlags, Segment, PAGE_SIZE},
     prelude::*,
-    sync::{SpinLock, SpinLockGuard},
     user::UserSpace,
 };
 
@@ -103,6 +104,41 @@ impl Drop for KernelStack {
     }
 }
 
+/// An atomic CPUID container.
+pub struct AtomicCpuId(AtomicU32);
+
+impl AtomicCpuId {
+    /// The null value of CPUID.
+    ///
+    /// An `AtomicCpuId` with `AtomicCpuId::NONE` as its inner value is empty.
+    const NONE: u32 = u32::MAX;
+
+    fn new(cpu_id: u32) -> Self {
+        Self(AtomicU32::new(cpu_id))
+    }
+
+    /// Sets the inner value of an `AtomicCpuId` if it's empty.
+    ///
+    /// The return value is a result indicating whether the new value was written
+    /// and containing the previous value.
+    pub fn set_if_is_none(&self, cpu_id: u32) -> core::result::Result<u32, u32> {
+        self.0
+            .compare_exchange(Self::NONE, cpu_id, Ordering::Relaxed, Ordering::Relaxed)
+    }
+
+    /// Sets the inner value of an `AtomicCpuId` to `AtomicCpuId::NONE`, i.e. makes
+    /// an `AtomicCpuId` empty.
+    pub fn set_to_none(&self) {
+        self.0.store(Self::NONE, Ordering::Relaxed);
+    }
+}
+
+impl Default for AtomicCpuId {
+    fn default() -> Self {
+        Self::new(Self::NONE)
+    }
+}
+
 /// A task that executes a function to the end.
 ///
 /// Each task is associated with per-task data and an optional user space.
@@ -112,11 +148,11 @@ pub struct Task {
     func: Box<dyn Fn() + Send + Sync>,
     data: Box<dyn Any + Send + Sync>,
     user_space: Option<Arc<UserSpace>>,
-    task_inner: SpinLock<TaskInner>,
     ctx: UnsafeCell<TaskContext>,
     /// kernel stack, note that the top is SyscallFrame/TrapFrame
     kstack: KernelStack,
     link: LinkedListAtomicLink,
+    cpu: AtomicCpuId,
     priority: Priority,
     // TODO: add multiprocessor support
     #[allow(dead_code)]
@@ -130,11 +166,6 @@ intrusive_adapter!(pub TaskAdapter = Arc<Task>: Task { link: LinkedListAtomicLin
 // we have exclusive access to the field.
 unsafe impl Sync for Task {}
 
-#[derive(Debug)]
-pub(crate) struct TaskInner {
-    pub task_status: TaskStatus,
-}
-
 impl Task {
     /// Gets the current task.
     ///
@@ -143,11 +174,6 @@ impl Task {
         current_task()
     }
 
-    /// Gets inner
-    pub(crate) fn inner_exclusive_access(&self) -> SpinLockGuard<TaskInner> {
-        self.task_inner.lock_irq_disabled()
-    }
-
     pub(super) fn ctx(&self) -> &UnsafeCell<TaskContext> {
         &self.ctx
     }
@@ -157,18 +183,14 @@ impl Task {
     /// Note that this method cannot be simply named "yield" as the name is
     /// a Rust keyword.
     pub fn yield_now() {
-        schedule();
+        scheduler::yield_now()
     }
 
     /// Runs the task.
+    ///
+    /// BUG: This method highly depends on the current scheduling policy.
     pub fn run(self: &Arc<Self>) {
-        add_task(self.clone());
-        schedule();
-    }
-
-    /// Returns the task status.
-    pub fn status(&self) -> TaskStatus {
-        self.task_inner.lock_irq_disabled().task_status
+        scheduler::run_new_task(self.clone());
     }
 
     /// Returns the task data.
@@ -185,6 +207,16 @@ impl Task {
         }
     }
 
+    // Returns the cpu of this task.
+    pub fn cpu(&self) -> &AtomicCpuId {
+        &self.cpu
+    }
+
+    /// Returns the priority.
+    pub fn priority(&self) -> Priority {
+        self.priority
+    }
+
     /// Exits the current task.
     ///
     /// The task `self` must be the task that is currently running.
@@ -192,13 +224,10 @@ impl Task {
     /// **NOTE:** If there is anything left on the stack, it will be forgotten. This behavior may
     /// lead to resource leakage.
     fn exit(self: Arc<Self>) -> ! {
-        self.inner_exclusive_access().task_status = TaskStatus::Exited;
-
         // `current_task()` still holds a strong reference, so nothing is destroyed at this point,
         // neither is the kernel stack.
         drop(self);
-
-        schedule();
+        scheduler::exit_current();
         unreachable!()
     }
 
@@ -208,19 +237,6 @@ impl Task {
     }
 }
 
-#[derive(Clone, Copy, PartialEq, Eq, PartialOrd, Ord, Debug)]
-/// The status of a task.
-pub enum TaskStatus {
-    /// The task is runnable.
-    Runnable,
-    /// The task is running in the foreground but will sleep when it goes to the background.
-    Sleepy,
-    /// The task is sleeping in the background.
-    Sleeping,
-    /// The task has exited.
-    Exited,
-}
-
 /// Options to create or spawn a new task.
 pub struct TaskOptions {
     func: Option<Box<dyn Fn() + Send + Sync>>,
@@ -236,13 +252,12 @@ impl TaskOptions {
     where
         F: Fn() + Send + Sync + 'static,
     {
-        let cpu_affinity = CpuSet::new_full();
         Self {
             func: Some(Box::new(func)),
             data: None,
             user_space: None,
             priority: Priority::normal(),
-            cpu_affinity,
+            cpu_affinity: CpuSet::new_full(),
         }
     }
 
@@ -300,11 +315,9 @@ impl TaskOptions {
             func: self.func.unwrap(),
             data: self.data.unwrap(),
             user_space: self.user_space,
-            task_inner: SpinLock::new(TaskInner {
-                task_status: TaskStatus::Runnable,
-            }),
             ctx: UnsafeCell::new(TaskContext::default()),
             kstack: KernelStack::new_with_guard_page()?,
+            cpu: AtomicCpuId::default(),
             link: LinkedListAtomicLink::new(),
             priority: self.priority,
             cpu_affinity: self.cpu_affinity,
diff --git a/ostd/src/task/priority.rs b/ostd/src/task/task/priority.rs
similarity index 96%
rename from ostd/src/task/priority.rs
rename to ostd/src/task/task/priority.rs
index 4bbad8c5..fd5b3ebe 100644
--- a/ostd/src/task/priority.rs
+++ b/ostd/src/task/task/priority.rs
@@ -7,7 +7,7 @@ pub const REAL_TIME_TASK_PRIORITY: u16 = 100;
 /// Similar to Linux, a larger value represents a lower priority,
 /// with a range of 0 to 139. Priorities ranging from 0 to 99 are considered real-time,
 /// while those ranging from 100 to 139 are considered normal.
-#[derive(Copy, Clone, Debug)]
+#[derive(Copy, Clone, Eq, Ord, PartialEq, PartialOrd)]
 pub struct Priority(u16);
 
 impl Priority {