Merge pull request #5071 from liukangcc/master

[update] support armclang and c++11.

27 files changed, 2594 insertions, 11 deletions

diff --git a/components/cplusplus/Kconfig b/components/cplusplus/Kconfig
index f2b817823d..dfd6786004 100644
--- a/components/cplusplus/Kconfig
+++ b/components/cplusplus/Kconfig
@@ -4,4 +4,16 @@ config RT_USING_CPLUSPLUS
     bool "Support C++ features"
     default n
 
+if RT_USING_CPLUSPLUS
+
+    config RT_USING_CPLUSPLUS11
+        bool "Enable c++11 feature support"
+        default n
+        select RT_USING_LIBC
+        select RT_USING_DFS
+        select RT_USING_PTHREADS
+        select RT_USING_RTC
+
+endif
+
 endmenu
diff --git a/components/cplusplus/SConscript b/components/cplusplus/SConscript
index 4f7fab9b9d..325a27f65f 100644
--- a/components/cplusplus/SConscript
+++ b/components/cplusplus/SConscript
@@ -1,12 +1,23 @@
 # RT-Thread building script for component
 
 from building import *
+Import('rtconfig')
 
 cwd = GetCurrentDir()
 src = Glob('*.cpp') + Glob('*.c')
 
+if GetDepend('RT_USING_CPLUSPLUS11'):
+    src += Glob('cpp11/*.cpp') + Glob('cpp11/*.c')
+
 CPPPATH = [cwd]
 
+if rtconfig.PLATFORM == 'armclang' and GetDepend('RT_USING_CPLUSPLUS11'):
+    src += Glob('cpp11/armclang/*.cpp') + Glob('cpp11/armclang/*.c')
+    CPPPATH += [cwd + '/cpp11/armclang']
+elif rtconfig.PLATFORM == 'gcc' and GetDepend('RT_USING_CPLUSPLUS11'):
+    src += Glob('cpp11/gcc/*.cpp') + Glob('cpp11/gcc/*.c')
+    CPPPATH += [cwd + '/cpp11/gcc']
+
 group = DefineGroup('CPlusPlus', src, depend = ['RT_USING_CPLUSPLUS'], CPPPATH = CPPPATH)
 
-Return('group')
+Return('group')
\ No newline at end of file
diff --git a/components/cplusplus/cpp11/README.md b/components/cplusplus/cpp11/README.md
new file mode 100644
index 0000000000..a292051608
--- /dev/null
+++ b/components/cplusplus/cpp11/README.md
@@ -0,0 +1,49 @@
+# C++ 11 support for RT-Thread
+
+## Features
+
+Here are some features about rt-thread c++11 threading futures.
+
+- Atomic.
+- Conditional variables.
+- Clocks.
+- Future.
+- Mutexes.
+- Threads.
+- TLS.
+
+## How To Use
+
+Note that using C++ 11 in rt-thread requires modifying some of the files in the toolchain. Before modifying the tool, back up the tool chain.  
+
+1. Enable c++11 support
+
+   ![](figures/Snipaste_2021-09-02_16-00-09.png)
+
+2. Download toolchain GCC 10.2.1:
+
+   ```shell
+   gcc version 10.2.1 20201103 (release) (GNU Arm Embedded Toolchain 10-2020-q4-major)
+   ```
+
+3. Delete the following files:
+
+   ```shell
+   rm -f toolchain/arm-none-eabi/include/c++/10.2.1/thread
+   rm -f toolchain/arm-none-eabi/include/c++/10.2.1/mutex
+   rm -f toolchain/arm-none-eabi/include/c++/10.2.1/condition_variable
+   rm -f toolchain/arm-none-eabi/include/c++/10.2.1/future
+   rm -f toolchain/arm-none-eabi/include/pthread.h
+   ```
+
+4. Clear the contents of the following files and keep them to prevent compilation failures:
+
+   ```shell
+   toolchain/arm-none-eabi/include/sys/_pthreadtypes.h
+   ```
+   
+5. Update `rtconfig.py` file. add compilation parameters:
+
+   ```shell
+   CXXFLAGS = CFLAGS  + ' -std=c++11 -fabi-version=0 -MMD -MP -MF'
+   ```
diff --git a/components/cplusplus/cpp11/README_ZH.md b/components/cplusplus/cpp11/README_ZH.md
new file mode 100644
index 0000000000..e46bcb90c4
--- /dev/null
+++ b/components/cplusplus/cpp11/README_ZH.md
@@ -0,0 +1,48 @@
+# cpp 11 support for rt-thread
+
+## 特性
+
+下面是 RT-Thread 支持的 C++ 11 线程特性。
+
+- Atomic.
+- Conditional variables.
+- Clocks.
+- Future.
+- Mutexes.
+- Threads.
+- TLS.
+
+## 如何使用
+
+请注意，在 RT-Thread 中使用 C++ 11，需要修改工具链中的部分文件。请在修改之前，备份好工具链。
+
+1. 使能 c++11
+   ![](figures/Snipaste_2021-09-02_16-00-09.png)
+
+2. 下载 GCC 工具链
+
+   ```shell
+   gcc version 10.2.1 20201103 (release) (GNU Arm Embedded Toolchain 10-2020-q4-major)
+   ```
+
+3. 删除下面的文件
+
+   ```shell
+   rm -f toolchain/arm-none-eabi/include/c++/10.2.1/thread
+   rm -f toolchain/arm-none-eabi/include/c++/10.2.1/mutex
+   rm -f toolchain/arm-none-eabi/include/c++/10.2.1/condition_variable
+   rm -f toolchain/arm-none-eabi/include/c++/10.2.1/future
+   rm -f toolchain/arm-none-eabi/include/pthread.h
+   ```
+
+4. 请清除下面文件的内容，保留文件避免编译失败
+
+   ```shell
+   toolchain/arm-none-eabi/include/sys/_pthreadtypes.h
+   ```
+
+5. 更新 `rtconfig.py` 文件，添加 c++ 编译参数:
+
+   ```shell
+   CXXFLAGS = CFLAGS  + ' -std=c++11 -fabi-version=0 -MMD -MP -MF'
+   ```
diff --git a/components/cplusplus/cpp11/armclang/clock.cpp b/components/cplusplus/cpp11/armclang/clock.cpp
new file mode 100644
index 0000000000..137af393b0
--- /dev/null
+++ b/components/cplusplus/cpp11/armclang/clock.cpp
@@ -0,0 +1,29 @@
+/*
+ * Copyright (c) 2006-2021, RT-Thread Development Team
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Change Logs:
+ * Date           Author       Notes
+ * 2021-04-27     flybreak     the first version.
+ */
+
+#include <arm-tpl.h>
+#include <sys/time.h>
+#include <rtthread.h>
+
+extern "C" int __ARM_TPL_clock_realtime(__ARM_TPL_timespec_t* __ts)
+{
+    unsigned int t = std::time(nullptr);
+    __ts->tv_sec = t;
+    __ts->tv_nsec = 0;
+    return 0;
+}
+
+extern "C" int __ARM_TPL_clock_monotonic(__ARM_TPL_timespec_t* __ts)
+{
+    unsigned int t = rt_tick_get();
+    __ts->tv_sec = t / RT_TICK_PER_SECOND;
+    __ts->tv_nsec = (t %RT_TICK_PER_SECOND) * NANOSECOND_PER_TICK  ;
+    return 0;
+}
diff --git a/components/cplusplus/cpp11/armclang/condvar.cpp b/components/cplusplus/cpp11/armclang/condvar.cpp
new file mode 100644
index 0000000000..7943ccee60
--- /dev/null
+++ b/components/cplusplus/cpp11/armclang/condvar.cpp
@@ -0,0 +1,178 @@
+/*
+ * Copyright (c) 2006-2021, RT-Thread Development Team
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Change Logs:
+ * Date           Author       Notes
+ * 2021-04-27     flybreak     the first version.
+ */
+
+#include <arm-tpl.h>
+#include "tpl.h"
+#include <new>
+#include <cstdint>
+#include <stdatomic.h>
+
+arm_tpl_cv::arm_tpl_cv()
+{
+    s = rt_sem_create("semxs", 0, RT_IPC_FLAG_PRIO);
+    if (s == nullptr)
+        RT_ASSERT(0);
+    h = rt_sem_create("semxh", 0, RT_IPC_FLAG_PRIO);
+    if (h == nullptr)
+    {
+        rt_sem_delete(s);
+        RT_ASSERT(0);
+    }
+    x = rt_mutex_create("mutx", RT_IPC_FLAG_PRIO);
+    if (x == nullptr)
+    {
+        rt_sem_delete(s);
+        rt_sem_delete(h);
+        RT_ASSERT(0);
+    }
+}
+
+arm_tpl_cv::~arm_tpl_cv()
+{
+    rt_mutex_delete(x);
+    rt_sem_delete(h);
+    rt_sem_delete(s);
+}
+
+void arm_tpl_cv::wait(rt_mutex_t lock, bool recursive)
+{
+    while (rt_mutex_take(x, ARM_TPL_MAX_DELAY) != 0);
+    rt_sem_release(s);
+    rt_mutex_release(x);
+    if (recursive)
+        rt_mutex_release(lock);
+    else
+        rt_mutex_release(lock);
+    while (rt_sem_take(h, ARM_TPL_MAX_DELAY) != 0);
+    if (recursive)
+        while (rt_mutex_take(lock, ARM_TPL_MAX_DELAY) != 0);
+    else
+        while (rt_mutex_take(lock, ARM_TPL_MAX_DELAY) != 0);
+}
+
+int arm_tpl_cv::timedwait(rt_mutex_t lock, bool recursive, unsigned int timeout_ms)
+{
+    int result = 0;
+    while (rt_mutex_take(x, ARM_TPL_MAX_DELAY) != 0);
+    rt_sem_release(s);
+    rt_mutex_release(x);
+    if (recursive)
+        rt_mutex_release(lock);
+    else
+        rt_mutex_release(lock);
+    if (rt_sem_take(h, rt_tick_from_millisecond(timeout_ms)) != 0)
+    {
+        while (rt_mutex_take(x, ARM_TPL_MAX_DELAY) != 0);
+        if (rt_sem_take(h, 0) != 0)
+        {
+            if (rt_sem_take(s, 0) != 0)
+                result = -1;
+            else
+                result = 1;
+        }
+        rt_mutex_release(x);
+    }
+    if (recursive)
+        while (rt_mutex_take(lock, ARM_TPL_MAX_DELAY) != 0);
+    else
+        while (rt_mutex_take(lock, ARM_TPL_MAX_DELAY) != 0);
+    return result;
+}
+
+void arm_tpl_cv::signal()
+{
+    while (rt_mutex_take(x, ARM_TPL_MAX_DELAY) != 0);
+    if (rt_sem_take(s, 0) == 0)
+        rt_sem_release(h);
+    rt_mutex_release(x);
+}
+
+void arm_tpl_cv::broadcast()
+{
+    while (rt_mutex_take(x, ARM_TPL_MAX_DELAY) != 0);
+    auto count = s->value;
+    for (auto i = 0; i < count; i++)
+    {
+        while (rt_sem_take(s, ARM_TPL_MAX_DELAY) != 0);
+        rt_sem_release(h);
+    }
+    rt_mutex_release(x);
+}
+
+static int check_create(volatile __ARM_TPL_condvar_t *__vcv)
+{
+    if (__vcv->data == 0)
+    {
+        uintptr_t cv_new;
+        cv_new = reinterpret_cast<uintptr_t>(new arm_tpl_cv());
+        if (cv_new == 0)
+        {
+            return -1;
+        }
+        uintptr_t cv_null = 0;
+        if (!atomic_compare_exchange_strong(&__vcv->data, &cv_null, cv_new))
+            delete reinterpret_cast<arm_tpl_cv *>(cv_new);
+    }
+    return 0;
+}
+
+extern "C" int __ARM_TPL_condvar_wait(__ARM_TPL_condvar_t *__cv, __ARM_TPL_mutex_t *__m)
+{
+    volatile __ARM_TPL_condvar_t *__vcv = __cv;
+    if (check_create(__vcv) != 0)
+        return -1;
+    struct arm_tpl_mutex_struct *tmutex = (struct arm_tpl_mutex_struct *)(__m->data);
+    ((arm_tpl_cv *) __vcv->data)->wait(tmutex->mutex, tmutex->type == RECURSIVE);
+    return 0;
+}
+
+extern "C" int __ARM_TPL_condvar_timedwait(__ARM_TPL_condvar_t *__cv,
+                                __ARM_TPL_mutex_t *__m,
+                                __ARM_TPL_timespec_t *__ts)
+{
+    volatile __ARM_TPL_condvar_t *__vcv = __cv;
+    if (check_create(__vcv) != 0)
+        return -1;
+    __ARM_TPL_timespec_t now;
+    if (__ARM_TPL_clock_realtime(&now) != 0)
+        return -1;
+    struct arm_tpl_mutex_struct *tmutex = (struct arm_tpl_mutex_struct *)(__m->data);
+    unsigned int timeout_ms = (__ts->tv_sec - now.tv_sec) * 1000 + (__ts->tv_nsec - now.tv_nsec) / 1000000;
+    if (((arm_tpl_cv *) __vcv->data)->timedwait(tmutex->mutex, tmutex->type == RECURSIVE, timeout_ms) < 0)
+        return -1;
+    return 0;
+}
+
+extern "C" int __ARM_TPL_condvar_signal(__ARM_TPL_condvar_t *__cv)
+{
+    volatile __ARM_TPL_condvar_t *__vcv = __cv;
+    if (__vcv->data != 0)
+        ((arm_tpl_cv *) __vcv->data)->signal();
+    return 0;
+}
+
+extern "C" int __ARM_TPL_condvar_broadcast(__ARM_TPL_condvar_t *__cv)
+{
+    volatile __ARM_TPL_condvar_t *__vcv = __cv;
+    if (__vcv->data != 0)
+        ((arm_tpl_cv *) __vcv->data)->broadcast();
+    return 0;
+}
+
+extern "C" int __ARM_TPL_condvar_destroy(__ARM_TPL_condvar_t *__cv)
+{
+    volatile __ARM_TPL_condvar_t *__vcv = __cv;
+    if (__vcv->data != 0)
+    {
+        delete (arm_tpl_cv *) __vcv->data;
+        __vcv->data = 0;
+    }
+    return 0;
+}
diff --git a/components/cplusplus/cpp11/armclang/miscellaneous.cpp b/components/cplusplus/cpp11/armclang/miscellaneous.cpp
new file mode 100644
index 0000000000..a084eda8c8
--- /dev/null
+++ b/components/cplusplus/cpp11/armclang/miscellaneous.cpp
@@ -0,0 +1,22 @@
+/*
+ * Copyright (c) 2006-2021, RT-Thread Development Team
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Change Logs:
+ * Date           Author       Notes
+ * 2021-04-27     flybreak     the first version.
+ */
+
+#include <arm-tpl.h>
+
+extern "C" int __ARM_TPL_execute_once(__ARM_TPL_exec_once_flag *__flag,
+                                      void (*__init_routine)(void))
+{
+    if (*__flag == 0)
+    {
+        __init_routine();
+        *__flag = 1;
+    }
+    return 0;
+}
diff --git a/components/cplusplus/cpp11/armclang/mutex.cpp b/components/cplusplus/cpp11/armclang/mutex.cpp
new file mode 100644
index 0000000000..44a70efcf7
--- /dev/null
+++ b/components/cplusplus/cpp11/armclang/mutex.cpp
@@ -0,0 +1,108 @@
+/*
+ * Copyright (c) 2006-2021, RT-Thread Development Team
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Change Logs:
+ * Date           Author       Notes
+ * 2021-04-27     flybreak     the first version.
+ */
+
+#include <arm-tpl.h>
+#include <cstdint>
+#include <stdatomic.h>
+#include "tpl.h"
+
+static int check_create(volatile __ARM_TPL_mutex_t *__vm, bool recursive = false)
+{
+    if (__vm->data == 0)
+    {
+        uintptr_t mut_null = 0;
+        arm_tpl_mutex_struct *mutex_p = (arm_tpl_mutex_struct *)rt_malloc(sizeof(arm_tpl_mutex_struct));
+        if (mutex_p == nullptr) return -1;
+
+        if (recursive)
+            mutex_p->mutex = rt_mutex_create("mutexx", RT_IPC_FLAG_PRIO);
+        else
+            mutex_p->mutex = rt_mutex_create("mutexx", RT_IPC_FLAG_PRIO);
+
+        if (mutex_p->mutex == nullptr)
+        {
+            rt_free(mutex_p);
+            return -1;
+        }
+        mutex_p->type = recursive ? RECURSIVE : NORMAL;
+        uintptr_t mut_new = reinterpret_cast<uintptr_t>(mutex_p);
+        if (!atomic_compare_exchange_strong(&__vm->data, &mut_null, mut_new))
+        {
+            rt_mutex_delete(mutex_p->mutex);
+            rt_free(mutex_p);
+        }
+    }
+    return 0;
+}
+
+static int mutexLock(arm_tpl_mutex_struct *mutex_p, rt_tick_t timeOut)
+{
+    if (mutex_p->type == RECURSIVE)
+    {
+        if (rt_mutex_take(mutex_p->mutex, timeOut) == 0)
+            return 0;
+    }
+    else
+    {
+        if (rt_mutex_take(mutex_p->mutex, timeOut) == 0)
+            return 0;
+    }
+    return -1;
+}
+
+static int mutexUnlock(arm_tpl_mutex_struct *mutex_p)
+{
+    if (mutex_p->type == RECURSIVE)
+        rt_mutex_release(mutex_p->mutex);
+    else
+        rt_mutex_release(mutex_p->mutex);
+    return 0;
+}
+
+extern "C" int __ARM_TPL_recursive_mutex_init(__ARM_TPL_mutex_t *__m)
+{
+    volatile __ARM_TPL_mutex_t *__vm = __m;
+    return check_create(__vm, true);
+}
+
+extern "C" int __ARM_TPL_mutex_lock(__ARM_TPL_mutex_t *__m)
+{
+    volatile __ARM_TPL_mutex_t *__vm = __m;
+    if (check_create(__vm))
+        return -1;
+    while (mutexLock((arm_tpl_mutex_struct *)(__vm->data), ARM_TPL_MAX_DELAY) != 0);
+    return 0;
+}
+
+extern "C" int __ARM_TPL_mutex_trylock(__ARM_TPL_mutex_t *__m)
+{
+    volatile __ARM_TPL_mutex_t *__vm = __m;
+    if (check_create(__vm))
+        return -1;
+    return mutexLock((arm_tpl_mutex_struct *)(__vm->data), 0);
+}
+
+extern "C" int __ARM_TPL_mutex_unlock(__ARM_TPL_mutex_t *__m)
+{
+    volatile __ARM_TPL_mutex_t *__vm = __m;
+    return mutexUnlock((arm_tpl_mutex_struct *)(__vm->data));
+}
+
+extern "C" int __ARM_TPL_mutex_destroy(__ARM_TPL_mutex_t *__m)
+{
+    volatile __ARM_TPL_mutex_t *__vm = __m;
+    if (__vm->data != 0)
+    {
+        rt_mutex_delete(((arm_tpl_mutex_struct *)(__vm->data))->mutex);
+        rt_free((void *)(__vm->data));
+        __vm->data = 0;
+    }
+    return 0;
+}
diff --git a/components/cplusplus/cpp11/armclang/thread.cpp b/components/cplusplus/cpp11/armclang/thread.cpp
new file mode 100644
index 0000000000..f7f0cb4077
--- /dev/null
+++ b/components/cplusplus/cpp11/armclang/thread.cpp
@@ -0,0 +1,109 @@
+/*
+ * Copyright (c) 2006-2021, RT-Thread Development Team
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Change Logs:
+ * Date           Author       Notes
+ * 2021-04-27     flybreak     the first version.
+ */
+
+#include <arm-tpl.h>
+#include "tpl.h"
+#include <cstdio>
+#include <pthread.h>
+
+extern "C" int __ARM_TPL_thread_create(__ARM_TPL_thread_t *__t,
+                                       void *(*__func)(void *),
+                                       void *__arg)
+{
+    int ret = 0;
+    /* TODO memory leek */
+    pthread_t *pid = (pthread_t *)rt_malloc(sizeof(pthread_t));
+    if (pid == nullptr)
+        return -1;
+    ret = pthread_create(pid, RT_NULL, __func, __arg);
+    if (ret == 0)
+    {
+        __t->data = (std::uintptr_t)pid;
+        return 0;
+    }
+    return -1;
+}
+
+extern "C" int __ARM_TPL_thread_id_compare(__ARM_TPL_thread_id __tid1,
+        __ARM_TPL_thread_id __tid2)
+{
+    if (__tid1 > __tid2)
+        return 1;
+    else if (__tid1 < __tid2)
+        return -1;
+    else
+        return 0;
+}
+
+extern "C" __ARM_TPL_thread_id __ARM_TPL_thread_get_current_id()
+{
+    return (__ARM_TPL_thread_id)pthread_self();
+}
+
+extern "C" __ARM_TPL_thread_id __ARM_TPL_thread_get_id(
+    const __ARM_TPL_thread_t *__t)
+{
+    return (__ARM_TPL_thread_id)((*(pthread_t *)__t->data));
+}
+
+extern "C" int __ARM_TPL_thread_join(__ARM_TPL_thread_t *__t)
+{
+    pthread_join((*(pthread_t *)__t->data), RT_NULL);
+    return 0;
+}
+
+extern "C" int __ARM_TPL_thread_detach(__ARM_TPL_thread_t *__t)
+{
+    pthread_detach((*(pthread_t *)__t->data));
+    return 0;
+}
+
+extern "C" void __ARM_TPL_thread_yield()
+{
+    rt_thread_yield();
+}
+
+extern "C" int __ARM_TPL_thread_nanosleep(const __ARM_TPL_timespec_t *__req,
+        __ARM_TPL_timespec_t *__rem)
+{
+    return nanosleep(__req, rem);
+}
+
+extern "C" unsigned __ARM_TPL_thread_hw_concurrency()
+{
+    return 1;
+}
+
+extern "C" int __ARM_TPL_tls_create(__ARM_TPL_tls_key *__key,
+                                    void (*__at_exit)(void *))
+{
+    pthread_key_t key; 
+    
+    if (pthread_key_create(&key, __at_exit) == 0)
+    {
+        *__key = key;
+        return 0;
+    }
+    return -1;
+}
+
+extern "C" void *__ARM_TPL_tls_get(__ARM_TPL_tls_key __key)
+{
+    return pthread_getspecific(__key);
+}
+
+extern "C" int __ARM_TPL_tls_set(__ARM_TPL_tls_key __key, void *__p)
+{
+    if (pthread_setspecific(__key, (void*)__p) != 0)
+    {
+        return -1;
+    }
+    return 0;
+}
diff --git a/components/cplusplus/cpp11/armclang/tpl.h b/components/cplusplus/cpp11/armclang/tpl.h
new file mode 100644
index 0000000000..70c9503cec
--- /dev/null
+++ b/components/cplusplus/cpp11/armclang/tpl.h
@@ -0,0 +1,56 @@
+/*
+ * Copyright (c) 2006-2021, RT-Thread Development Team
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Change Logs:
+ * Date           Author       Notes
+ * 2021-04-27     flybreak     the first version.
+ */
+
+#pragma once
+#ifndef __cplusplus
+void ARMTPLInit();
+#else
+#include <rtthread.h>
+
+#define ARM_TPL_MAX_DELAY 1000
+#define ARM_TPL_THREAD_STACK_SIZE 4096
+
+enum arm_tpl_mutex_type
+{
+    NORMAL,
+    RECURSIVE,
+};
+
+struct arm_tpl_mutex_struct
+{
+    rt_mutex_t mutex;
+    arm_tpl_mutex_type type;
+};
+
+struct arm_tpl_thread_struct
+{
+    rt_thread_t task;
+    void *(*func)(void *);
+    void *arg;
+    rt_sem_t join_sem;
+    rt_sem_t detach_sem;
+};
+
+class arm_tpl_cv
+{
+public:
+    arm_tpl_cv();
+    ~arm_tpl_cv();
+    void wait(rt_mutex_t lock, bool recursive);
+    int timedwait(rt_mutex_t lock, bool recursive, unsigned int timeout_ms);
+    void signal();
+    void broadcast();
+private:
+    rt_sem_t s;
+    rt_sem_t h;
+    rt_mutex_t x;
+};
+
+#endif
diff --git a/components/cplusplus/cpp11/atomic_8.c b/components/cplusplus/cpp11/atomic_8.c
new file mode 100644
index 0000000000..64d3311b4a
--- /dev/null
+++ b/components/cplusplus/cpp11/atomic_8.c
@@ -0,0 +1,88 @@
+/*
+ * Copyright (c) 2006-2021, RT-Thread Development Team
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Change Logs:
+ * Date           Author       Notes
+ * 2021-04-27     peterfan     Add copyright header.
+ */
+
+#include <rthw.h>
+#include <stdint.h>
+#include <stdbool.h>
+
+/*
+* override gcc builtin atomic function for std::atomic<int64_t>, std::atomic<uint64_t>
+* @see https://gcc.gnu.org/onlinedocs/gcc/_005f_005fatomic-Builtins.html
+*/
+uint64_t __atomic_load_8(volatile void *ptr, int memorder)
+{
+    volatile uint64_t *val_ptr = (volatile uint64_t *)ptr;
+    register rt_base_t level;
+    uint64_t tmp;
+    level = rt_hw_interrupt_disable();
+    tmp = *val_ptr;
+    rt_hw_interrupt_enable(level);
+    return tmp;
+}
+
+void __atomic_store_8(volatile void *ptr, uint64_t val, int memorder)
+{
+    volatile uint64_t *val_ptr = (volatile uint64_t *)ptr;
+    register rt_base_t level;
+    level = rt_hw_interrupt_disable();
+    *val_ptr = val;
+    rt_hw_interrupt_enable(level);
+}
+
+uint64_t __atomic_exchange_8(volatile void *ptr, uint64_t val, int memorder)
+{
+    volatile uint64_t *val_ptr = (volatile uint64_t *)ptr;
+    register rt_base_t level;
+    uint64_t tmp;
+    level = rt_hw_interrupt_disable();
+    tmp = *val_ptr;
+    *val_ptr = val;
+    rt_hw_interrupt_enable(level);
+    return tmp;
+}
+
+bool __atomic_compare_exchange_8(volatile void *ptr, volatile void *expected, uint64_t desired, bool weak, int success_memorder, int failure_memorder)
+{
+    volatile uint64_t *val_ptr = (volatile uint64_t *)ptr;
+    volatile uint64_t *expected_ptr = (volatile uint64_t *)expected;
+    register rt_base_t level;
+    bool exchanged;
+    level = rt_hw_interrupt_disable();
+    if (*val_ptr == *expected_ptr)
+    {
+        *val_ptr = desired;
+        exchanged = true;
+    }
+    else
+    {
+        *expected_ptr = *val_ptr;
+        exchanged = false;
+    }
+    rt_hw_interrupt_enable(level);
+    return exchanged;
+}
+
+#define __atomic_fetch_op_8(OPNAME, OP) \
+uint64_t __atomic_fetch_##OPNAME##_8(volatile void *ptr, uint64_t val, int memorder) {\
+    volatile uint64_t* val_ptr = (volatile uint64_t*)ptr;\
+    register rt_base_t level;\
+    uint64_t tmp;\
+    level = rt_hw_interrupt_disable();\
+    tmp = *val_ptr;\
+    *val_ptr OP##= val;\
+    rt_hw_interrupt_enable(level);\
+    return tmp;\
+}
+
+__atomic_fetch_op_8(add, +)
+__atomic_fetch_op_8(sub, -)
+__atomic_fetch_op_8( and, &)
+__atomic_fetch_op_8( or, |)
+__atomic_fetch_op_8(xor, ^)
diff --git a/components/cplusplus/cpp11/emutls.c b/components/cplusplus/cpp11/emutls.c
new file mode 100644
index 0000000000..576888f16f
--- /dev/null
+++ b/components/cplusplus/cpp11/emutls.c
@@ -0,0 +1,216 @@
+/*
+ * Copyright (c) 2006-2021, RT-Thread Development Team
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Change Logs:
+ * Date           Author       Notes
+ * 2021-04-27     peterfan     Add copyright header.
+ */
+
+/* ===---------- emutls.c - Implements __emutls_get_address ---------------===
+ *
+ *                     The LLVM Compiler Infrastructure
+ *
+ * This file is dual licensed under the MIT and the University of Illinois Open
+ * Source Licenses. See LICENSE.TXT for details.
+ *
+ * ===----------------------------------------------------------------------===
+ */
+
+#include <pthread.h>
+#include <stdint.h>
+#include <stdlib.h>
+#include <string.h>
+
+#define COMPILE_TIME_ASSERT(x)
+
+extern int pthread_key_create(pthread_key_t *key, void (*destructor)(void *));
+extern int pthread_key_delete(pthread_key_t key);
+extern void *pthread_getspecific(pthread_key_t key);
+extern int pthread_setspecific(pthread_key_t key, const void *value);
+
+/* Default is not to use posix_memalign, so systems like Android
+ * can use thread local data without heavier POSIX memory allocators.
+ */
+#ifndef EMUTLS_USE_POSIX_MEMALIGN
+#define EMUTLS_USE_POSIX_MEMALIGN 0
+#endif
+
+/* For every TLS variable xyz,
+ * there is one __emutls_control variable named __emutls_v.xyz.
+ * If xyz has non-zero initial value, __emutls_v.xyz's "value"
+ * will point to __emutls_t.xyz, which has the initial value.
+ */
+typedef struct __emutls_control
+{
+    size_t size;  /* size of the object in bytes */
+    size_t align; /* alignment of the object in bytes */
+    union
+    {
+        uintptr_t index; /* data[index-1] is the object address */
+        void *address;   /* object address, when in single thread env */
+    } object;
+    void *value; /* null or non-zero initial value for the object */
+} __emutls_control;
+
+static __inline void *emutls_memalign_alloc(size_t align, size_t size)
+{
+    void *base;
+#if EMUTLS_USE_POSIX_MEMALIGN
+    if (posix_memalign(&base, align, size) != 0)
+        abort();
+#else
+#define EXTRA_ALIGN_PTR_BYTES (align - 1 + sizeof(void *))
+    char *object;
+    if ((object = malloc(EXTRA_ALIGN_PTR_BYTES + size)) == NULL)
+        abort();
+    base = (void *)(((uintptr_t)(object + EXTRA_ALIGN_PTR_BYTES)) & ~(uintptr_t)(align - 1));
+
+    ((void **)base)[-1] = object;
+#endif
+    return base;
+}
+
+static __inline void emutls_memalign_free(void *base)
+{
+#if EMUTLS_USE_POSIX_MEMALIGN
+    free(base);
+#else
+    /* The mallocated address is in ((void**)base)[-1] */
+    free(((void **)base)[-1]);
+#endif
+}
+
+/* Emulated TLS objects are always allocated at run-time. */
+static __inline void *emutls_allocate_object(__emutls_control *control)
+{
+    /* Use standard C types, check with gcc's emutls.o. */
+    typedef unsigned int gcc_word __attribute__((mode(word)));
+    typedef unsigned int gcc_pointer __attribute__((mode(pointer)));
+    COMPILE_TIME_ASSERT(sizeof(size_t) == sizeof(gcc_word));
+    COMPILE_TIME_ASSERT(sizeof(uintptr_t) == sizeof(gcc_pointer));
+    COMPILE_TIME_ASSERT(sizeof(uintptr_t) == sizeof(void *));
+
+    size_t size = control->size;
+    size_t align = control->align;
+    if (align < sizeof(void *))
+        align = sizeof(void *);
+    /* Make sure that align is power of 2. */
+    if ((align & (align - 1)) != 0)
+        abort();
+
+    void *base = emutls_memalign_alloc(align, size);
+    if (control->value)
+        memcpy(base, control->value, size);
+    else
+        memset(base, 0, size);
+    return base;
+}
+
+static pthread_mutex_t emutls_mutex = PTHREAD_MUTEX_INITIALIZER;
+
+static size_t emutls_num_object = 0; /* number of allocated TLS objects */
+
+typedef struct emutls_address_array
+{
+    uintptr_t size; /* number of elements in the 'data' array */
+    void *data[];
+} emutls_address_array;
+
+static pthread_key_t emutls_pthread_key;
+
+static void emutls_key_destructor(void *ptr)
+{
+    emutls_address_array *array = (emutls_address_array *)ptr;
+    uintptr_t i;
+    for (i = 0; i < array->size; ++i)
+    {
+        if (array->data[i])
+            emutls_memalign_free(array->data[i]);
+    }
+    free(ptr);
+}
+
+static void emutls_init(void)
+{
+    if (pthread_key_create(&emutls_pthread_key, emutls_key_destructor) != 0)
+        abort();
+}
+
+/* Returns control->object.index; set index if not allocated yet. */
+static __inline uintptr_t emutls_get_index(__emutls_control *control)
+{
+    uintptr_t index = __atomic_load_n(&control->object.index, __ATOMIC_ACQUIRE);
+    if (!index)
+    {
+        static pthread_once_t once = PTHREAD_ONCE_INIT;
+        pthread_once(&once, emutls_init);
+        pthread_mutex_lock(&emutls_mutex);
+        index = control->object.index;
+        if (!index)
+        {
+            index = ++emutls_num_object;
+            __atomic_store_n(&control->object.index, index, __ATOMIC_RELEASE);
+        }
+        pthread_mutex_unlock(&emutls_mutex);
+    }
+    return index;
+}
+
+/* Updates newly allocated thread local emutls_address_array. */
+static __inline void emutls_check_array_set_size(emutls_address_array *array,
+                                                 uintptr_t size)
+{
+    if (array == NULL)
+        abort();
+    array->size = size;
+    pthread_setspecific(emutls_pthread_key, (void *)array);
+}
+
+/* Returns the new 'data' array size, number of elements,
+ * which must be no smaller than the given index.
+ */
+static __inline uintptr_t emutls_new_data_array_size(uintptr_t index)
+{
+    /* Need to allocate emutls_address_array with one extra slot
+     * to store the data array size.
+     * Round up the emutls_address_array size to multiple of 16.
+     */
+    return ((index + 1 + 15) & ~((uintptr_t)15)) - 1;
+}
+
+/* Returns the thread local emutls_address_array.
+ * Extends its size if necessary to hold address at index.
+ */
+static __inline emutls_address_array *
+emutls_get_address_array(uintptr_t index)
+{
+    emutls_address_array *array = pthread_getspecific(emutls_pthread_key);
+    if (array == NULL)
+    {
+        uintptr_t new_size = emutls_new_data_array_size(index);
+        array = calloc(new_size + 1, sizeof(void *));
+        emutls_check_array_set_size(array, new_size);
+    }
+    else if (index > array->size)
+    {
+        uintptr_t orig_size = array->size;
+        uintptr_t new_size = emutls_new_data_array_size(index);
+        array = realloc(array, (new_size + 1) * sizeof(void *));
+        if (array)
+            memset(array->data + orig_size, 0,
+                   (new_size - orig_size) * sizeof(void *));
+        emutls_check_array_set_size(array, new_size);
+    }
+    return array;
+}
+
+void *__emutls_get_address(void *control)
+{
+    uintptr_t index = emutls_get_index((__emutls_control *)control);
+    emutls_address_array *array = emutls_get_address_array(index);
+    if (array->data[index - 1] == NULL)
+        array->data[index - 1] = emutls_allocate_object((__emutls_control *)control);
+    return array->data[index - 1];
+}
diff --git a/components/cplusplus/cpp11/figures/Snipaste_2021-09-02_16-00-09.png b/components/cplusplus/cpp11/figures/Snipaste_2021-09-02_16-00-09.png
new file mode 100644
index 0000000000..0e2381b5d9
--- /dev/null
+++ b/components/cplusplus/cpp11/figures/Snipaste_2021-09-02_16-00-09.png
diff --git a/components/cplusplus/cpp11/gcc/__utils.h b/components/cplusplus/cpp11/gcc/__utils.h
new file mode 100644
index 0000000000..f9a3c93796
--- /dev/null
+++ b/components/cplusplus/cpp11/gcc/__utils.h
@@ -0,0 +1,59 @@
+/*
+ * Copyright (c) 2006-2021, RT-Thread Development Team
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Change Logs:
+ * Date           Author       Notes
+ * 2021-04-27     flybreak     the first version.
+ */
+
+#pragma once
+
+#include <cstdlib>
+#include <system_error>
+#include <chrono>
+#include <ratio>
+
+#include <rtthread.h>
+
+#define RT_USING_CPP_EXCEPTION
+
+inline void throw_system_error(int err, const char *what_msg)
+{
+#ifdef RT_USING_CPP_EXCEPTION
+    throw std::system_error(std::error_code(err, std::system_category()), what_msg);
+#else
+    (void)err;
+    (void)what_msg;
+    ::abort();
+#endif
+}
+
+class tick_clock
+{
+public:
+    typedef clock_t rep;
+    typedef std::ratio<1, RT_TICK_PER_SECOND> period;
+
+    typedef std::chrono::duration<tick_clock::rep, tick_clock::period> duration;
+    typedef std::chrono::time_point<tick_clock> time_point;
+
+    constexpr static bool is_ready = true;
+
+    static time_point now();
+};
+
+class real_time_clock
+{
+public:
+    typedef std::chrono::nanoseconds duration;
+    typedef duration::rep rep;
+    typedef duration::period period;
+    typedef std::chrono::time_point<real_time_clock, duration> time_point;
+
+    static constexpr bool is_steady = true;
+
+    static time_point
+    now() noexcept;
+};
diff --git a/components/cplusplus/cpp11/gcc/condition_variable b/components/cplusplus/cpp11/gcc/condition_variable
new file mode 100644
index 0000000000..f2d48dd5b5
--- /dev/null
+++ b/components/cplusplus/cpp11/gcc/condition_variable
@@ -0,0 +1,222 @@
+#pragma once
+
+#if __cplusplus < 201103L
+#error "C++ version lower than C++11"
+#endif
+
+#include <pthread.h>
+
+#include <system_error>
+#include <chrono>
+#include <utility>
+#include <functional>
+#include <memory>
+
+#include "__utils.h"
+#include "mutex"
+
+#define rt_cpp_cond_var pthread_cond_t
+
+namespace std
+{
+
+    enum class cv_status
+    {
+        no_timeout,
+        timeout
+    };
+
+    class condition_variable
+    {
+    public:
+        typedef rt_cpp_cond_var *native_handle_type;
+
+        condition_variable(const condition_variable &) = delete;
+        condition_variable &operator=(const condition_variable &) = delete;
+
+        condition_variable() = default;
+
+        ~condition_variable()
+        {
+            pthread_cond_destroy(&_m_cond);
+        }
+
+        void wait(unique_lock<mutex> &lock);
+
+        void notify_one() noexcept
+        {
+            pthread_cond_signal(&_m_cond);
+        }
+
+        void notify_all() noexcept
+        {
+            pthread_cond_broadcast(&_m_cond);
+        }
+
+        template <class Predicate>
+        void wait(unique_lock<mutex> &lock, Predicate pred)
+        {
+            while (!pred())
+                wait(lock);
+        }
+
+        template <class Clock, class Duration>
+        cv_status wait_until(unique_lock<mutex> &lock,
+                             const chrono::time_point<Clock, Duration> &abs_time)
+        {
+            if (!lock.owns_lock())
+                throw_system_error((int)errc::operation_not_permitted,
+                                   "condition_variable::wailt_until: waiting on unlocked lock");
+            auto secs = chrono::time_point_cast<chrono::seconds>(abs_time);
+            auto nano_secs = chrono::duration_cast<chrono::nanoseconds>(abs_time - secs);
+
+            struct timespec c_abs_time = {static_cast<time_t>(secs.time_since_epoch().count()),
+                                          static_cast<long>(nano_secs.count())};
+
+            pthread_cond_timedwait(&_m_cond, lock.mutex()->native_handle(), &c_abs_time);
+
+            return (Clock::now() < abs_time) ? cv_status::no_timeout : cv_status::timeout;
+        }
+
+        template <class Clock, class Duration, class Predicate>
+        bool wait_until(unique_lock<mutex> &lock,
+                        const chrono::time_point<Clock, Duration> &abs_time,
+                        Predicate pred)
+        {
+            while (!pred())
+                if (wait_until(lock, abs_time) == cv_status::timeout)
+                    return pred();
+            return true;
+        }
+
+        template <class Rep, class Period>
+        cv_status wait_for(unique_lock<mutex> &lock,
+                           const chrono::duration<Rep, Period> &rel_time)
+        {
+            return wait_until(lock, real_time_clock::now() + rel_time);
+        }
+
+        template <class Rep, class Period, class Predicate>
+        bool wait_for(unique_lock<mutex> &lock,
+                      const chrono::duration<Rep, Period> &rel_time,
+                      Predicate pred)
+        {
+            return wait_until(lock, real_time_clock::now() + rel_time, std::move(pred));
+        }
+
+        native_handle_type native_handle()
+        {
+            return &_m_cond;
+        }
+
+    private:
+        rt_cpp_cond_var _m_cond = PTHREAD_COND_INITIALIZER;
+    };
+
+    // Lockable is only required to have `lock()` and `unlock()`
+    class condition_variable_any
+    {
+    private:
+        condition_variable _m_cond;
+        shared_ptr<mutex> _m_mtx;
+
+        // so that Lockable automatically unlocks when waiting and locks after waiting
+        template <class Lockable>
+        struct unlocker
+        {
+            Lockable &_m_lock;
+
+            explicit unlocker(Lockable &lk)
+                : _m_lock(lk)
+            {
+                _m_lock.unlock();
+            }
+
+            ~unlocker()
+            {
+                _m_lock.lock();
+            }
+
+            unlocker(const unlocker &) = delete;
+            unlocker &operator=(const unlocker &) = delete;
+        };
+
+    public:
+        condition_variable_any() : _m_mtx(std::make_shared<mutex>()) {}
+        ~condition_variable_any() = default;
+
+        condition_variable_any(const condition_variable_any &) = delete;
+        condition_variable_any &operator=(const condition_variable_any &) = delete;
+
+        void notify_one() noexcept
+        {
+            lock_guard<mutex> lk(*_m_mtx);
+            _m_cond.notify_one();
+        }
+
+        void notify_all() noexcept
+        {
+            lock_guard<mutex> lk(*_m_mtx);
+            _m_cond.notify_all();
+        }
+
+        template <class Lock>
+        void wait(Lock &lock)
+        {
+            shared_ptr<mutex> mut = _m_mtx;
+            unique_lock<mutex> lk(*mut);
+            unlocker<Lock> auto_lk(lock); // unlock here
+
+            unique_lock<mutex> lk2(std::move(lk));
+            _m_cond.wait(lk2);
+        } // mut.unlock(); lock.lock();
+
+        template <class Lock, class Predicate>
+        void wait(Lock &lock, Predicate pred)
+        {
+            while (!pred())
+                wait(lock);
+        }
+
+        template <class Lock, class Clock, class Duration>
+        cv_status wait_until(Lock &lock,
+                             const chrono::time_point<Clock, Duration> &abs_time)
+        {
+            shared_ptr<mutex> mut = _m_mtx;
+            unique_lock<mutex> lk(*mut);
+            unlocker<Lock> auto_lk(lock); // unlock here
+
+            unique_lock<mutex> lk2(std::move(lk));
+            return _m_cond.wait_until(lk2, abs_time);
+        }
+
+        template <class Lock, class Clock, class Duration, class Predicate>
+        bool wait_until(Lock &lock,
+                        const chrono::time_point<Clock, Duration> &abs_time,
+                        Predicate pred)
+        {
+            while (!pred())
+                if (wait_until(lock, abs_time) == cv_status::timeout)
+                    return pred();
+            return true;
+        }
+
+        template <class Lock, class Rep, class Period>
+        cv_status wait_for(Lock &lock,
+                           const chrono::duration<Rep, Period> &rel_time)
+        {
+            return wait_until(lock, real_time_clock::now() + rel_time);
+        }
+
+        template <class Lock, class Rep, class Period, class Predicate>
+        bool wait_for(Lock &lock,
+                      const chrono::duration<Rep, Period> &rel_time,
+                      Predicate pred)
+        {
+            return wait_until(lock, real_time_clock::now() + rel_time, std::move(pred));
+        }
+    };
+
+    void notify_all_at_thread_exit(condition_variable &cond, unique_lock<mutex> lk);
+
+} // namespace std
diff --git a/components/cplusplus/cpp11/gcc/condition_variable.cpp b/components/cplusplus/cpp11/gcc/condition_variable.cpp
new file mode 100644
index 0000000000..2860dd9161
--- /dev/null
+++ b/components/cplusplus/cpp11/gcc/condition_variable.cpp
@@ -0,0 +1,34 @@
+/*
+ * Copyright (c) 2006-2021, RT-Thread Development Team
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Change Logs:
+ * Date           Author       Notes
+ * 2021-04-27     flybreak     the first version.
+ */
+
+#include "condition_variable"
+
+namespace std
+{
+    void condition_variable::wait(unique_lock<mutex>& lock)
+    {
+        int err = pthread_cond_wait(&_m_cond, lock.mutex()->native_handle());
+
+        if (err)
+        {
+            throw_system_error(err, "condition_variable::wait: failed to wait on a condition");
+        }
+    }
+
+    void notify_all_at_thread_exit(condition_variable& cond, unique_lock<mutex> lk)
+    {
+        // TLS currently not available
+        mutex* mut = lk.release();
+        mut->unlock();
+        cond.notify_all();
+    }
+
+
+} // namespace std
diff --git a/components/cplusplus/cpp11/gcc/future b/components/cplusplus/cpp11/gcc/future
new file mode 100644
index 0000000000..6a686063fb
--- /dev/null
+++ b/components/cplusplus/cpp11/gcc/future
@@ -0,0 +1,336 @@
+#pragma once
+
+#if __cplusplus < 201103L
+#error "C++ version lower than C++11"
+#endif
+
+#include <mutex>
+#include <condition_variable>
+#include <memory>
+#include <chrono>
+#include <cassert>
+
+namespace std {
+
+enum class future_status {
+    ready,
+    timeout,
+    deferred
+};
+
+namespace detail {
+
+class shared_state_base {
+protected:
+    typedef void (*deleter_fn)(void *v);
+
+    using scoped_lock = std::lock_guard<std::mutex>;
+    using unique_lock = std::unique_lock<std::mutex>;
+public:
+    explicit shared_state_base(deleter_fn d) : v_(nullptr), d_(d), valid_(true) {}
+
+    ~shared_state_base() { d_(v_); }
+
+    shared_state_base(shared_state_base &&other) = delete;
+
+    shared_state_base(const shared_state_base &other) = delete;
+
+    shared_state_base &operator=(shared_state_base &&other) = delete;
+
+    shared_state_base &operator=(const shared_state_base &other) = delete;
+
+    void wait() {
+        unique_lock lock(m_);
+        c_.wait(lock, [this] { return has_value(); });
+    }
+
+    template <class Rep, class Period>
+    std::future_status
+    wait_for(const std::chrono::duration<Rep, Period> &rel_time) {
+        unique_lock lock(m_);
+        if (c_.wait_for(lock, rel_time, [this] { return has_value(); })) {
+            return std::future_status::ready;
+        }
+        return std::future_status::timeout;
+    }
+
+    template <class Clock, class Duration>
+    std::future_status
+    wait_until(const std::chrono::time_point<Clock, Duration> &abs_time) {
+        unique_lock lock(m_);
+        if (c_.wait_until(lock, abs_time, [this] { return has_value(); })) {
+            return std::future_status::ready;
+        }
+        return std::future_status::timeout;
+    }
+
+protected:
+    bool has_value() { return v_ != nullptr; }
+
+protected:
+    std::mutex m_;
+    std::condition_variable c_;
+    void *v_;
+    deleter_fn d_;
+    bool valid_;
+};
+
+
+template <typename R>
+class shared_state: public shared_state_base {
+public:
+    shared_state() :shared_state_base(default_deleter_) {}
+
+    ~shared_state() {}
+
+    R &get() {
+        wait();
+        scoped_lock lock(m_);
+        assert(valid_);
+        valid_ = false;
+        return *(static_cast<R *>(v_));
+    }
+
+    void set(const R &v) {
+        scoped_lock lock(m_);
+        assert(!has_value());
+        v_ = new R(v);
+        valid_ = true;
+        c_.notify_one();
+    }
+
+    void set(R &&v) {
+        scoped_lock lock(m_);
+        assert(!has_value());
+        v_ = new R(std::move(v));
+        valid_ = true;
+        c_.notify_one();
+    }
+
+    bool valid() {
+        scoped_lock lock(m_);
+        return valid_;
+    }
+
+private:
+    static void default_deleter_(void *v) { delete static_cast<R *>(v); }
+};
+
+} // namespace detail
+
+template <typename R>
+class shared_future {
+};
+
+
+template <typename R>
+class future {
+    using state_type = std::shared_ptr<detail::shared_state<R>>;
+public:
+    future() {}
+
+    explicit future(const state_type &state) : state_(state) {}
+
+    future(future &&other) noexcept: state_(std::move(other.state_)) {
+        other.state_.reset();
+    }
+
+    future(const future &other) = delete;
+
+    ~future() {}
+
+    future &operator=(future &&other) noexcept {
+        if (&other != this) {
+            state_ = std::move(other.state_);
+            other.state_.reset();
+        }
+        return *this;
+    }
+
+    future &operator=(const future &other) = delete;
+
+    void swap(future &other) noexcept {
+        std::swap(state_, other.state_);
+    }
+
+    std::shared_future<R> share() noexcept { return std::shared_future<R>(); }
+
+    R get() { return state_->get(); }
+
+    bool valid() const noexcept { return state_->valid(); }
+
+    void wait() const { state_->wait(); }
+
+    template <class Rep, class Period>
+    std::future_status
+    wait_for(const std::chrono::duration<Rep, Period> &rel_time) const {
+        return state_->wait_for(rel_time);
+    }
+
+    template <class Clock, class Duration>
+    std::future_status
+    wait_until(const std::chrono::time_point<Clock, Duration> &abs_time) const {
+        return state_->wait_until(abs_time);
+    }
+
+private:
+    state_type state_;
+};
+
+
+template <>
+class future<void> {
+    using state_type = std::shared_ptr<detail::shared_state<int>>;
+public:
+    future() {}
+
+    explicit future(const state_type &state) : state_(state) {}
+
+    future(future &&other) noexcept: state_(std::move(other.state_)) {
+        other.state_.reset();
+    }
+
+    future(const future &other) = delete;
+
+    ~future() {}
+
+    future &operator=(future &&other) noexcept {
+        if (&other != this) {
+            state_ = std::move(other.state_);
+            other.state_.reset();
+        }
+        return *this;
+    }
+
+    future &operator=(const future &other) = delete;
+
+    void swap(future &other) noexcept {
+        std::swap(state_, other.state_);
+    }
+
+    std::shared_future<void> share() noexcept { return std::shared_future<void>(); }
+
+    void get() { state_->get(); }
+
+    bool valid() const noexcept { return state_->valid(); }
+
+    void wait() const { state_->wait(); }
+
+    template <class Rep, class Period>
+    std::future_status
+    wait_for(const std::chrono::duration<Rep, Period> &rel_time) const {
+        return state_->wait_for(rel_time);
+    }
+
+    template <class Clock, class Duration>
+    std::future_status
+    wait_until(const std::chrono::time_point<Clock, Duration> &abs_time) const {
+        return state_->wait_until(abs_time);
+    }
+
+private:
+    state_type state_;
+};
+
+
+template <typename R>
+class promise {
+    using state_type = std::shared_ptr<detail::shared_state<R>>;
+public:
+    promise() : state_(new detail::shared_state<R>()) {}
+
+    promise(promise &&other) noexcept: state_(std::move(other.state_)) {
+        other.state_.reset();
+    }
+
+    promise(const promise &other) = delete;
+
+    ~promise() {}
+
+    promise &operator=(promise &&other) noexcept {
+        if (&other != this) {
+            state_ = std::move(other.state_);
+            other.state_.reset();
+        }
+        return *this;
+    }
+
+    promise &operator=(const promise &other) = delete;
+
+    void swap(promise &other) noexcept {
+        std::swap(state_, other.state_);
+    }
+
+    std::future<R> get_future() { return std::future<R>(state_); }
+
+    void set_value(const R &value) { state_->set(value); }
+
+    void set_value(R &&value) { state_->set(std::move(value)); }
+
+    void set_value_at_thread_exit(const R &value);
+
+    void set_value_at_thread_exit(R &&value);
+
+    void set_exception(std::exception_ptr p);
+
+    void set_exception_at_thread_exit(std::exception_ptr p);
+
+private:
+    state_type state_;
+};
+
+
+template <>
+class promise<void> {
+    using state_type = std::shared_ptr<detail::shared_state<int>>;
+public:
+    promise() : state_(new detail::shared_state<int>()) {}
+
+    promise(promise &&other) noexcept: state_(std::move(other.state_)) {
+        other.state_.reset();
+    }
+
+    promise(const promise &other) = delete;
+
+    ~promise() {}
+
+    promise &operator=(promise &&other) noexcept {
+        if (&other != this) {
+            state_ = std::move(other.state_);
+            other.state_.reset();
+        }
+        return *this;
+    }
+
+    promise &operator=(const promise &other) = delete;
+
+    void swap(promise &other) noexcept {
+        std::swap(state_, other.state_);
+    }
+
+    std::future<void> get_future() { return std::future<void>(state_); }
+
+    void set_value() { state_->set(0); }
+
+    void set_value_at_thread_exit();
+
+    void set_exception(std::exception_ptr p);
+
+    void set_exception_at_thread_exit(std::exception_ptr p);
+
+private:
+    state_type state_;
+};
+
+
+template <class R>
+void swap(std::future<R> &lhs, std::future<R> &rhs) noexcept {
+    lhs.swap(rhs);
+}
+
+template <class R>
+void swap(std::promise<R> &lhs, std::promise<R> &rhs) noexcept {
+    lhs.swap(rhs);
+}
+
+} // namespace std
diff --git a/components/cplusplus/cpp11/gcc/mutex b/components/cplusplus/cpp11/gcc/mutex
new file mode 100644
index 0000000000..cc593cc139
--- /dev/null
+++ b/components/cplusplus/cpp11/gcc/mutex
@@ -0,0 +1,512 @@
+#pragma once
+
+#if __cplusplus < 201103L
+#error "C++ version lower than C++11"
+#endif
+
+//#if defined(RT_USING_LIBC) && defined(RT_USING_PTHREADS)
+
+#include <pthread.h>
+
+#include <system_error>
+#include <chrono>
+#include <utility>
+#include <functional>
+
+#include "__utils.h"
+
+#define rt_cpp_mutex_t  pthread_mutex_t
+
+namespace std 
+{
+    // Base class on which to build std::mutex and std::timed_mutex
+    class __mutex_base
+    {
+        protected:
+            typedef rt_cpp_mutex_t  __native_type;
+
+            __native_type _m_mutex = PTHREAD_MUTEX_INITIALIZER;
+
+            constexpr __mutex_base() noexcept = default;
+            __mutex_base(const __mutex_base&) = delete;
+            __mutex_base& operator=(const __mutex_base&) = delete;
+    };
+
+    
+    class mutex : private __mutex_base
+    {
+        public:
+            constexpr mutex() = default;
+            ~mutex() = default;
+
+            mutex(const mutex&) = delete;
+            mutex& operator=(const mutex&) = delete;
+
+            void lock()
+            {
+                int err  = pthread_mutex_lock(&_m_mutex);
+
+                if (err)
+                {
+                    throw_system_error(err, "mutex:lock failed.");
+                }
+            }
+
+            bool try_lock() noexcept
+            {
+                return !pthread_mutex_trylock(&_m_mutex);
+            }
+
+            void unlock() noexcept
+            {
+                pthread_mutex_unlock(&_m_mutex);
+            }
+
+            typedef __native_type* native_handle_type;
+
+            native_handle_type native_handle() 
+            {
+                return &_m_mutex;
+            };
+
+    };
+
+    inline int __rt_cpp_recursive_mutex_init(rt_cpp_mutex_t* m)
+    {
+        pthread_mutexattr_t attr;
+        int res;
+
+        res = pthread_mutexattr_init(&attr);
+        if (res)
+            return res;
+        res = pthread_mutexattr_settype(&attr, PTHREAD_MUTEX_RECURSIVE);
+        if (res)
+            goto attr_cleanup;
+        res = pthread_mutex_init(m, &attr);
+
+        attr_cleanup:
+            int err = pthread_mutexattr_destroy(&attr);
+            return res ? res : err;
+    }
+
+    class __recursive_mutex_base
+    {
+        protected:
+            typedef rt_cpp_mutex_t __native_type;
+
+            __native_type _m_recursive_mutex;
+
+            __recursive_mutex_base(const __recursive_mutex_base&) = delete;
+            __recursive_mutex_base& operator=(const __recursive_mutex_base&) = delete;
+
+            __recursive_mutex_base()
+            {
+                int err = __rt_cpp_recursive_mutex_init(&_m_recursive_mutex);
+                if (err)
+                    throw_system_error(err, "Recursive mutex failed to construct");
+            }
+
+            ~__recursive_mutex_base()
+            {
+                pthread_mutex_destroy(&_m_recursive_mutex);
+            }
+    };
+
+    class recursive_mutex : private __recursive_mutex_base
+    {
+        public:
+            typedef __native_type* native_handle_type; 
+            recursive_mutex() = default;
+            ~recursive_mutex() = default;
+
+            recursive_mutex(const recursive_mutex&) = delete;
+            recursive_mutex& operator=(const recursive_mutex&) = delete;
+            void lock()
+            {
+                int err = pthread_mutex_lock(&_m_recursive_mutex);
+
+                if (err)
+                    throw_system_error(err, "recursive_mutex::lock failed");
+            }
+
+            bool try_lock() noexcept
+            {
+                return !pthread_mutex_trylock(&_m_recursive_mutex);
+            }
+
+            void unlock() noexcept
+            {
+                pthread_mutex_unlock(&_m_recursive_mutex);
+            }
+
+            native_handle_type native_handle()
+            { return &_m_recursive_mutex; }
+    };
+
+#ifdef RT_PTHREAD_TIMED_MUTEX
+
+    class timed_mutex;
+
+    class recursive_timed_mutex;
+
+#endif // RT_PTHREAD_TIMED_MUTEX
+
+    
+    struct defer_lock_t {};
+    struct try_to_lock_t {};
+    struct adopt_lock_t {}; // take ownership of a locked mtuex
+
+    constexpr defer_lock_t defer_lock { };
+    constexpr try_to_lock_t try_to_lock { };
+    constexpr adopt_lock_t adopt_lock { };
+
+    template <class Mutex> 
+    class lock_guard
+    {
+        public:
+            typedef Mutex mutex_type;
+
+            explicit lock_guard(mutex_type& m) : pm(m) { pm.lock(); }
+            lock_guard(mutex_type& m, adopt_lock_t) noexcept : pm(m)
+            { }
+            ~lock_guard() 
+            { pm.unlock(); }
+
+            lock_guard(lock_guard const&) = delete;
+            lock_guard& operator=(lock_guard const&) = delete;
+
+        private:
+            mutex_type& pm;
+
+    };
+
+    template <class Mutex>
+    class unique_lock
+    {
+        public:
+            typedef Mutex mutex_type;
+
+            unique_lock() noexcept : pm(nullptr), owns(false) { }
+            
+            explicit unique_lock(mutex_type& m) 
+                : pm(std::addressof(m)), owns(false)
+            {
+                lock();
+                owns = true;
+            }
+
+            unique_lock(mutex_type& m, defer_lock_t) noexcept 
+                : pm(std::addressof(m)), owns(false)
+            { }
+
+            unique_lock(mutex_type& m, try_to_lock_t) noexcept
+                : pm(std::addressof(m)), owns(pm->try_lock())
+            { }
+
+            unique_lock(mutex_type& m, adopt_lock_t) noexcept
+                : pm(std::addressof(m)), owns(true)
+            { }
+
+            // any lock-involving timed mutex API is currently only for custom implementations
+            // the standard ones are not available
+            template <class Clock, class Duration>
+            unique_lock(mutex_type& m, const chrono::time_point<Clock, Duration>& abs_time) noexcept
+                : pm(std::addressof(m)), owns(pm->try_lock_until(abs_time))
+            { }
+
+            template <class Rep, class Period>
+            unique_lock(mutex_type& m, const chrono::duration<Rep, Period>& rel_time) noexcept
+                : pm(std::addressof(m)), owns(pm->try_lock_for(rel_time))
+            { }
+            
+            ~unique_lock()
+            {
+                if (owns)
+                    unlock();
+            }
+            
+            unique_lock(unique_lock const&) = delete;
+            unique_lock& operator=(unique_lock const&) = delete;
+
+            unique_lock(unique_lock&& u) noexcept
+                : pm(u.pm), owns(u.owns)
+            {
+                u.pm = nullptr;
+                u.owns = false;
+            }
+
+            unique_lock& operator=(unique_lock&& u) noexcept                
+            {
+                if (owns)
+                    unlock();
+                
+                unique_lock(std::move(u)).swap(*this);
+
+                u.pm = nullptr;
+                u.owns = false;
+
+                return *this;
+            }
+
+            void lock()
+            {
+                if (!pm)
+                    throw_system_error(int(errc::operation_not_permitted), 
+                                        "unique_lock::lock: references null mutex");
+                else if (owns)
+                    throw_system_error(int(errc::resource_deadlock_would_occur),
+                                        "unique_lock::lock: already locked" );
+                else {
+                    pm->lock();
+                    owns = true;
+                }
+            }
+
+            bool try_lock()
+            {
+                if (!pm)
+                    throw_system_error(int(errc::operation_not_permitted), 
+                                        "unique_lock::try_lock: references null mutex");
+                else if (owns)
+                    throw_system_error(int(errc::resource_deadlock_would_occur),
+                                        "unique_lock::try_lock: already locked" );
+                else {
+                    owns = pm->try_lock();
+                }
+                return owns;
+            }
+
+            template <class Rep, class Period>
+            bool try_lock_for(const chrono::duration<Rep, Period>& rel_time)
+            {
+                if (!pm)
+                    throw_system_error(int(errc::operation_not_permitted),
+                                        "unique_lock::try_lock_for: references null mutex");
+                else if (owns)
+                    throw_system_error(int(errc::resource_deadlock_would_occur),
+                                        "unique_lock::try_lock_for: already locked");
+                else {
+                    owns = pm->try_lock_for(rel_time);
+                }
+                return owns;
+            }
+            
+            template <class Clock, class Duration>
+            bool try_lock_until(const chrono::time_point<Clock, Duration>& abs_time)
+            {
+                if (!pm)
+                    throw_system_error(int(errc::operation_not_permitted),
+                                        "unique_lock::try_lock_until: references null mutex");
+                else if (owns)
+                    throw_system_error(int(errc::resource_deadlock_would_occur),
+                                        "unique_lock::try_lock_until: already locked");
+                else {
+                    owns = pm->try_lock_until(abs_time);
+                }
+                return owns;
+            }
+            
+            void unlock()
+            {
+                if (!owns)
+                    throw_system_error(int(errc::operation_not_permitted), 
+                                        "unique_lock::unlock: not locked");
+                else {
+                    pm->unlock();
+                    owns = false;
+                }
+            }
+
+            void swap(unique_lock& u) noexcept
+            {
+                std::swap(pm, u.pm);
+                std::swap(owns, u.owns);
+            }
+
+            mutex_type *release() noexcept
+            {
+                mutex_type* ret_mutex = pm;
+                pm = nullptr;
+                owns = false;
+                
+                return ret_mutex;
+            }
+
+            bool owns_lock() const noexcept
+            { return owns; }
+
+            explicit operator bool() const noexcept
+            { return owns_lock(); }
+
+            mutex_type* mutex() const noexcept
+            { return pm; }
+            
+            
+        private:
+            mutex_type *pm; 
+            bool owns;
+    };
+
+    template <class Mutex>
+    void swap(unique_lock<Mutex>& x, unique_lock<Mutex>& y)
+    {
+        x.swap(y);
+    }
+
+    template <class L0, class L1>
+    int try_lock(L0& l0, L1& l1)
+    {
+        unique_lock<L0> u0(l0, try_to_lock); // try to lock the first Lockable
+        // using unique_lock since we don't want to unlock l0 manually if l1 fails to lock
+        if (u0.owns_lock())
+        {
+            if (l1.try_lock()) // lock the second one
+            {
+                u0.release(); // do not let RAII of a unique_lock unlock l0
+                return -1;
+            }
+            else
+                return 1; 
+        }
+        return 0;
+    } 
+
+    
+    template <class L0, class L1, class L2, class... L3>
+    int try_lock(L0& l0, L1& l1, L2& l2, L3&... l3)
+    {
+        int r = 0;
+        unique_lock<L0> u0(l0, try_to_lock);
+        // automatically unlock is done through RAII of unique_lock
+        if (u0.owns_lock())
+        {
+            r = try_lock(l1, l2, l3...);
+            if (r == -1)
+                u0.release();
+            else
+                ++r;
+        }
+        return r;
+    }
+
+    template <class L0, class L1, class L2, class ...L3>
+    void
+    __lock_first(int i, L0& l0, L1& l1, L2& l2, L3&... l3)
+    {
+        while (true)
+        {
+            // we first lock the one that is the most difficult to lock 
+            switch (i) 
+            {
+            case 0:
+                {
+                    unique_lock<L0> u0(l0);
+                    i = try_lock(l1, l2, l3...);
+                    if (i == -1)
+                    {
+                        u0.release();
+                        return;
+                    }
+                }
+                ++i;
+                sched_yield();
+                break;
+            case 1:
+                {
+                    unique_lock<L1> u1(l1);
+                    i = try_lock(l2, l3..., l0);
+                    if (i == -1)
+                    {
+                        u1.release();
+                        return;
+                    }
+                }
+                if (i == sizeof...(L3) + 1) // all except l0 are locked
+                    i = 0;
+                else
+                    i += 2; // since i was two-based above
+                sched_yield();
+                break;
+            default:
+                __lock_first(i - 2, l2, l3..., l0, l1);
+                return;
+            }
+        }
+    }
+
+
+    template <class L0, class L1>
+    void lock(L0& l0, L1& l1)
+    {
+        while (true)
+        {
+            {
+                unique_lock<L0> u0(l0);
+                if (l1.try_lock())
+                {
+                    u0.release();
+                    break;
+                }
+            }
+            sched_yield();
+            // wait and try the other way
+            {
+                unique_lock<L1> u1(l1);
+                if (l0.try_lock())
+                {
+                    u1.release();
+                    break;
+                }
+            }
+            sched_yield();
+        }
+    }
+
+    template <class L0, class L1, class... L2>
+    void lock(L0& l0, L1& l1, L2&... l2)
+    {
+        __lock_first(0, l0, l1, l2...);
+    }
+
+    struct once_flag 
+    {
+        constexpr once_flag() noexcept = default;
+
+        once_flag(const once_flag&) = delete;
+        once_flag& operator=(const once_flag&) = delete;
+
+        template <class Callable, class... Args>
+        friend void call_once(once_flag& flag, Callable&& func, Args&&... args);
+
+        private:
+            pthread_once_t _m_once = PTHREAD_ONCE_INIT;
+    };
+
+    mutex& get_once_mutex();
+    extern function<void()> once_functor;
+    extern void set_once_functor_lock_ptr(unique_lock<mutex>*);
+
+    extern "C" void once_proxy(); // passed into pthread_once
+
+    template <class Callable, class... Args>
+    void call_once(once_flag& flag, Callable&& func, Args&&... args)
+    {
+        // use a lock to ensure the call to the functor
+        // is exclusive to only the first calling thread
+        unique_lock<mutex> functor_lock(get_once_mutex()); 
+
+        auto call_wrapper = std::bind(std::forward<Callable>(func), std::forward<Args>(args)...);
+        once_functor = [&]() { call_wrapper(); };
+
+        set_once_functor_lock_ptr(&functor_lock); // so as to unlock when actually calling
+
+        int err = pthread_once(&flag._m_once, &once_proxy);
+
+        if (functor_lock)
+            set_once_functor_lock_ptr(nullptr);
+        if (err)
+            throw_system_error(err, "call_once failed");
+    }
+}
+
+//#endif // (RT_USING_LIBC) && (RT_USING_PTHREADS)
\ No newline at end of file
diff --git a/components/cplusplus/cpp11/gcc/mutex.cpp b/components/cplusplus/cpp11/gcc/mutex.cpp
new file mode 100644
index 0000000000..014460ffde
--- /dev/null
+++ b/components/cplusplus/cpp11/gcc/mutex.cpp
@@ -0,0 +1,52 @@
+/*
+ * Copyright (c) 2006-2021, RT-Thread Development Team
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Change Logs:
+ * Date           Author       Notes
+ * 2021-04-27     flybreak     the first version.
+ */
+
+#include "mutex"
+
+namespace std
+{
+    // use a set of global and static objects
+    // a proxy function to pthread_once
+
+    function<void()> once_functor;
+
+    mutex& get_once_mutex()
+    {
+        static mutex once_mutex;
+        return once_mutex;
+    }
+
+    inline unique_lock<mutex>*& get_once_functor_lock_ptr()
+    {
+        static unique_lock<mutex>* once_functor_mutex_ptr = nullptr;
+        return once_functor_mutex_ptr;
+    }
+
+    void set_once_functor_lock_ptr(unique_lock<mutex>* m_ptr)
+    {
+        get_once_functor_lock_ptr() = m_ptr;
+    }
+
+    extern "C"  
+    {
+        void once_proxy()
+        {
+            // need to first transfer the functor's ownership so as to call it
+            function<void()> once_call = std::move(once_functor);
+
+            // no need to hold the lock anymore
+            unique_lock<mutex>* lock_ptr = get_once_functor_lock_ptr();            
+            get_once_functor_lock_ptr() = nullptr;
+            lock_ptr->unlock();
+
+            once_call();
+        }
+    }
+}
diff --git a/components/cplusplus/cpp11/gcc/thread b/components/cplusplus/cpp11/gcc/thread
new file mode 100644
index 0000000000..5b64decd6a
--- /dev/null
+++ b/components/cplusplus/cpp11/gcc/thread
@@ -0,0 +1,239 @@
+#pragma once
+
+#if __cplusplus < 201103L
+#error "C++ version lower than C++11"
+#endif
+
+//#if defined(RT_USING_LIBC) && defined(RT_USING_PTHREADS)
+
+#include <unistd.h>
+#include <pthread.h>
+#include <sched.h>
+#include <rtthread.h>
+
+#include <cstddef>
+#include <cerrno>
+#include <ostream>
+#include <functional>
+#include <utility>
+#include <chrono>
+#include <memory>
+
+#define rt_cpp_thread_t pthread_t
+#ifndef PTHREAD_NUM_MAX
+#define PTHREAD_NUM_MAX 32
+#endif
+#define CPP_UNJOINABLE_THREAD PTHREAD_NUM_MAX
+
+namespace std
+{
+    #define __STDCPP_THREADS__ __cplusplus
+
+    
+
+    class thread
+    {
+    public:
+        typedef rt_cpp_thread_t   native_handle_type;
+        
+        struct invoker_base;
+        typedef shared_ptr<invoker_base> invoker_base_ptr;
+
+        class id
+        {
+            // basically a wrapper around native_handle_type
+            native_handle_type __cpp_thread_t;
+
+        public:
+            id() noexcept : __cpp_thread_t(CPP_UNJOINABLE_THREAD) {} 
+
+            explicit id(native_handle_type hid) 
+                : __cpp_thread_t(hid) {}
+        private:
+            friend class thread;
+            friend class hash<thread::id>;
+
+            friend bool operator==(thread::id x, thread::id y) noexcept;
+
+            friend bool operator<(thread::id x, thread::id y) noexcept;
+
+            template <class charT, class traits>
+            friend basic_ostream<charT, traits>& 
+            operator<<(basic_ostream<charT, traits>& out, thread::id id);
+        };
+
+        thread() noexcept = default;
+        thread(const thread&) = delete;
+        thread& operator=(const thread&) = delete;
+        ~thread();
+        
+        template <class F, class ...Args>
+        explicit thread(F&& f, Args&&... args)
+        {
+            start_thread(make_invoker_ptr(std::bind(
+                std::forward<F>(f),
+                std::forward<Args>(args)...
+            )));
+        }
+
+        thread(thread&& t) noexcept
+        {
+            swap(t);
+        }
+
+        thread& operator=(thread&& t) noexcept
+        {
+            if (joinable())
+                terminate();
+            swap(t);
+            return *this;
+        }
+
+        // member functions
+        void swap(thread& t) noexcept
+        {
+            std::swap(_m_thr, t._m_thr);
+        }
+
+        bool joinable() const noexcept
+        {
+            return (_m_thr.__cpp_thread_t < PTHREAD_NUM_MAX);
+        }
+
+        void join();
+
+        void detach();
+
+        id get_id() const noexcept { return _m_thr; }
+
+        native_handle_type native_handle() { return _m_thr.__cpp_thread_t; }
+
+        // static members
+        static unsigned hardware_concurrency() noexcept;
+
+    private:
+        id _m_thr;
+
+        void start_thread(invoker_base_ptr b);
+    public:
+        struct invoker_base
+        {
+            invoker_base_ptr this_ptr;
+            
+            virtual ~invoker_base() = default;
+
+            virtual void invoke() = 0;
+        };
+
+
+        template<typename Callable>
+        struct invoker : public invoker_base
+        {
+            Callable func;
+
+            invoker(Callable&& F) : func(std::forward<Callable>(F)) { }
+
+            void invoke() { func(); }
+        };
+
+        template <typename Callable>
+        shared_ptr<invoker<Callable>> make_invoker_ptr(Callable&& F)
+        {
+            return std::make_shared<invoker<Callable>>(std::forward<Callable>(F));
+        }
+
+        
+    };
+
+    inline void swap(thread& x, thread& y) noexcept
+    {
+        x.swap(y);
+    }
+
+
+    inline bool operator==(thread::id x, thread::id y) noexcept
+    {
+        // From POSIX for pthread_equal: 
+        //"If either t1 or t2 are not valid thread IDs,  the behavior is undefined."
+        return x.__cpp_thread_t == y.__cpp_thread_t;
+    }
+
+    inline bool operator!=(thread::id x, thread::id y) noexcept
+    {
+        return !(x == y);
+    }
+
+    inline bool operator<(thread::id x, thread::id y) noexcept
+    {
+        return x.__cpp_thread_t < y.__cpp_thread_t;
+    }
+
+    inline bool operator<=(thread::id x, thread::id y) noexcept
+    {
+        return !(y < x);
+    }
+
+    inline bool operator>(thread::id x, thread::id y) noexcept
+    {
+        return !(x <= y);
+    }
+
+    inline bool operator>=(thread::id x, thread::id y) noexcept
+    {
+        return !(x < y);
+    }
+
+    template <class charT, class traits>
+    inline basic_ostream<charT, traits>& 
+    operator<<(basic_ostream<charT, traits>& out, thread::id id)
+    {
+        if (id == thread::id()) // id is invalid, representing no pthread
+            out << "thread::id of a non-executing thread";
+        else
+            out << id.__cpp_thread_t;
+        return out;
+    }
+
+    template <> 
+    struct hash<thread::id> 
+    {
+        typedef size_t result_type;
+        typedef thread::id argument_type;
+        size_t operator()(const thread::id& id) const noexcept
+        {
+            return hash<rt_cpp_thread_t>()(id.__cpp_thread_t);
+        }
+    };
+
+    namespace this_thread 
+    {
+        inline thread::id get_id() noexcept
+        {
+            return thread::id(pthread_self());
+        }
+
+        inline void yield() noexcept
+        {
+            sched_yield();
+        } 
+
+        template <class Rep, class Period>
+        inline void sleep_for(const chrono::duration<Rep, Period>& rel_time)
+        {
+            if (rel_time <= rel_time.zero()) // less than zero, no need to sleep
+                return;
+            auto milli_secs = chrono::duration_cast<chrono::milliseconds>(rel_time);
+            // the precision is limited by rt-thread thread API
+            rt_thread_mdelay(milli_secs.count());
+        }
+
+        template <class Clock, class Duration>
+        inline void sleep_until(const chrono::time_point<Clock, Duration>& abs_time)
+        {
+            auto now = Clock::now();
+            if (abs_time > now)
+                sleep_for(abs_time - now);
+        }
+
+    }
+}
diff --git a/components/cplusplus/cpp11/gcc/thread.cpp b/components/cplusplus/cpp11/gcc/thread.cpp
new file mode 100644
index 0000000000..947170da25
--- /dev/null
+++ b/components/cplusplus/cpp11/gcc/thread.cpp
@@ -0,0 +1,94 @@
+/*
+ * Copyright (c) 2006-2021, RT-Thread Development Team
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Change Logs:
+ * Date           Author       Notes
+ * 2021-04-27     flybreak     the first version.
+ */
+
+#include "thread"
+#include "__utils.h"
+
+
+#define _RT_NPROCS 0
+
+namespace std
+{
+
+    extern "C"
+    {
+        static void* execute_native_thread_routine(void *p)
+        {
+            thread::invoker_base* t = static_cast<thread::invoker_base*>(p);
+            thread::invoker_base_ptr local;
+            local.swap(t->this_ptr); // tranfer the ownership of the invoker into the thread entry
+
+            local->invoke();
+
+            return NULL;
+        }
+    }
+
+    void thread::start_thread(invoker_base_ptr b)
+    {
+        auto raw_ptr = b.get();
+        // transfer the ownership of the invoker to the new thread
+        raw_ptr->this_ptr = std::move(b); 
+        int err = pthread_create(&_m_thr.__cpp_thread_t, NULL,
+                &execute_native_thread_routine, raw_ptr);
+
+        if (err) 
+        {
+            raw_ptr->this_ptr.reset();
+            throw_system_error(err, "Failed to create a thread");
+        }
+        
+    }    
+
+    thread::~thread()
+    {
+        if (joinable()) // when either not joined or not detached
+            terminate();
+    }
+
+    void thread::join()
+    {
+        int err = EINVAL;
+
+        if (joinable())
+            err = pthread_join(native_handle(), NULL);
+        
+        if (err) 
+        {
+            throw_system_error(err, "thread::join failed");
+        }
+    
+        _m_thr = id();
+    }
+
+    void thread::detach()
+    {
+        int err = EINVAL;
+
+        if (joinable())
+            err = pthread_detach(native_handle());
+        if (err)
+        {
+            throw_system_error(err, "thread::detach failed");
+        }
+    
+        _m_thr = id();
+    }
+
+    // TODO: not yet actually implemented.
+    // The standard states that the returned value should only be considered a hint.
+    unsigned thread::hardware_concurrency() noexcept 
+    {   
+        int __n = _RT_NPROCS;
+        if (__n < 0)
+            __n = 0;
+        return __n;
+    }
+}
diff --git a/components/cplusplus/cpp11/gcc/utils.cpp b/components/cplusplus/cpp11/gcc/utils.cpp
new file mode 100644
index 0000000000..1ec3b70cad
--- /dev/null
+++ b/components/cplusplus/cpp11/gcc/utils.cpp
@@ -0,0 +1,29 @@
+/*
+ * Copyright (c) 2006-2021, RT-Thread Development Team
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Change Logs:
+ * Date           Author       Notes
+ * 2021-04-27     flybreak     the first version.
+ */
+
+#include "__utils.h"
+#include <sys/time.h>
+
+tick_clock::time_point tick_clock::now()
+{
+    tick_clock::rep cur_tk = clock();
+    tick_clock::duration cur_time(cur_tk);
+
+    return tick_clock::time_point(cur_time);
+}
+
+real_time_clock::time_point real_time_clock::now() noexcept
+{
+    timespec tp;
+    clock_gettime(CLOCK_REALTIME, &tp);
+
+    return time_point(duration(std::chrono::seconds(tp.tv_sec))
+                        + std::chrono::nanoseconds(tp.tv_nsec));
+}
diff --git a/components/cplusplus/cpp11/thread_local_impl.cpp b/components/cplusplus/cpp11/thread_local_impl.cpp
new file mode 100644
index 0000000000..43761c1e62
--- /dev/null
+++ b/components/cplusplus/cpp11/thread_local_impl.cpp
@@ -0,0 +1,34 @@
+/*
+ * Copyright (c) 2006-2021, RT-Thread Development Team
+ *
+ * SPDX-License-Identifier: Apache-2.0
+ *
+ * Change Logs:
+ * Date           Author       Notes
+ * 2021-04-27     flybreak     the first version.
+ */
+
+#include <pthread.h>
+#include <cstdlib>
+
+typedef void (*destructor) (void *);
+
+extern "C"
+int __cxa_thread_atexit_impl(destructor dtor, void* obj, void* dso_symbol)
+{
+    pthread_key_t key_tmp;
+    if (pthread_key_create(&key_tmp, dtor) != 0)
+        abort();
+    pthread_setspecific(key_tmp, obj);
+    return 0;
+}
+
+#if defined(__GNUC__) && !defined(__ARMCC_VERSION)/*GCC*/
+#include <cxxabi.h>
+
+extern"C"
+int __cxxabiv1::__cxa_thread_atexit(destructor dtor, void *obj, void *dso_handle)
+{
+    return __cxa_thread_atexit_impl(dtor, obj, dso_handle);
+}
+#endif
diff --git a/components/libc/compilers/common/nogcc/sys/types.h b/components/libc/compilers/common/nogcc/sys/types.h
index 2c4c12055f..5672f54a4d 100644
--- a/components/libc/compilers/common/nogcc/sys/types.h
+++ b/components/libc/compilers/common/nogcc/sys/types.h
@@ -27,6 +27,7 @@ typedef signed int       ssize_t;       /* Used for a count of bytes or an error
 #else
 typedef long signed int  ssize_t;       /* Used for a count of bytes or an error indication. */
 #endif
+typedef long             suseconds_t;   /* microseconds. */
 typedef unsigned long    useconds_t;    /* microseconds (unsigned) */
 
 typedef unsigned long    dev_t;
diff --git a/components/libc/compilers/common/sys/time.h b/components/libc/compilers/common/sys/time.h
index 9d4422e17a..ecf6e85ac6 100644
--- a/components/libc/compilers/common/sys/time.h
+++ b/components/libc/compilers/common/sys/time.h
@@ -48,9 +48,10 @@ rt_int8_t rt_tz_is_dst(void);
 #ifndef _TIMEVAL_DEFINED
 #define _TIMEVAL_DEFINED
 #if !(defined(_WIN32))
-struct timeval {
-    long    tv_sec;     /* seconds */
-    long    tv_usec;    /* and microseconds */
+struct timeval
+{
+    time_t      tv_sec;     /* seconds */
+    suseconds_t tv_usec;    /* and microseconds */
 };
 #endif
 #endif /* _TIMEVAL_DEFINED */
diff --git a/components/libc/pthreads/pthread.c b/components/libc/pthreads/pthread.c
index 79104d1fd5..b0bf39dff6 100644
--- a/components/libc/pthreads/pthread.c
+++ b/components/libc/pthreads/pthread.c
@@ -90,9 +90,30 @@ void _pthread_data_destroy(pthread_t pth)
 {
     RT_DECLARE_SPINLOCK(pth_lock);
 
+    extern _pthread_key_data_t _thread_keys[PTHREAD_KEY_MAX];
     _pthread_data_t *ptd = _pthread_get_data(pth);
     if (ptd)
     {
+        /* destruct thread local key */
+        if (ptd->tls != RT_NULL)
+        {
+            void *data;
+            rt_uint32_t index;
+            for (index = 0; index < PTHREAD_KEY_MAX; index ++)
+            {
+                if (_thread_keys[index].is_used)
+                {
+                    data = ptd->tls[index];
+                    if (data && _thread_keys[index].destructor)
+                        _thread_keys[index].destructor(data);
+                }
+            }
+
+            /* release tls area */
+            rt_free(ptd->tls);
+            ptd->tls = RT_NULL;
+        }
+
         /* remove from pthread table */
         rt_hw_spin_lock(&pth_lock);
         pth_table[pth] = NULL;
@@ -103,13 +124,17 @@ void _pthread_data_destroy(pthread_t pth)
             rt_sem_delete(ptd->joinable_sem);
 
         /* release thread resource */
-        if (ptd->attr.stackaddr == RT_NULL && ptd->tid->stack_addr != RT_NULL)
+        if (ptd->attr.stackaddr == RT_NULL)
         {
             /* release thread allocated stack */
-            rt_free(ptd->tid->stack_addr);
+            if (ptd->tid)
+            {
+                rt_free(ptd->tid->stack_addr);
+            }
         }
         /* clean stack addr pointer */
-        ptd->tid->stack_addr = RT_NULL;
+        if (ptd->tid)
+            ptd->tid->stack_addr = RT_NULL;
 
         /*
         * if this thread create the local thread data,
@@ -221,6 +246,12 @@ int pthread_create(pthread_t            *pid,
         pthread_attr_init(&ptd->attr);
     }
 
+    if (ptd->attr.stacksize == 0)
+    {
+        ret = EINVAL;
+        goto __exit;
+    }
+
     rt_snprintf(name, sizeof(name), "pth%02d", pthread_number ++);
 
     /* pthread is a static thread object */
@@ -269,7 +300,7 @@ int pthread_create(pthread_t            *pid,
     /* initial this pthread to system */
     if (rt_thread_init(ptd->tid, name, pthread_entry_stub, ptd,
                        stack, ptd->attr.stacksize,
-                       ptd->attr.schedparam.sched_priority, 5) != RT_EOK)
+                       ptd->attr.schedparam.sched_priority, 20) != RT_EOK)
     {
         ret = EINVAL;
         goto __exit;
@@ -460,7 +491,7 @@ void pthread_exit(void *value)
             if (_thread_keys[index].is_used)
             {
                 data = ptd->tls[index];
-                if (data)
+                if (data && _thread_keys[index].destructor)
                     _thread_keys[index].destructor(data);
             }
         }
@@ -507,14 +538,21 @@ int pthread_kill(pthread_t thread, int sig)
 {
 #ifdef RT_USING_SIGNALS
     _pthread_data_t *ptd;
+    int ret;
 
     ptd = _pthread_get_data(thread);
     if (ptd)
     {
-        return rt_thread_kill(ptd->tid, sig);
+        ret = rt_thread_kill(ptd->tid, sig);
+        if (ret == -RT_EINVAL)
+        {
+            return EINVAL;
+        }
+
+        return ret;
     }
 
-    return EINVAL;
+    return ESRCH;
 #else
     return ENOSYS;
 #endif
@@ -707,3 +745,4 @@ int pthread_cancel(pthread_t thread)
     return 0;
 }
 RTM_EXPORT(pthread_cancel);
+
diff --git a/components/libc/pthreads/pthread.h b/components/libc/pthreads/pthread.h
index 8639c1604b..177b2539c5 100644
--- a/components/libc/pthreads/pthread.h
+++ b/components/libc/pthreads/pthread.h
@@ -265,6 +265,11 @@ int pthread_barrier_init(pthread_barrier_t           *barrier,
 
 int pthread_barrier_wait(pthread_barrier_t *barrier);
 
+int pthread_setspecific(pthread_key_t key, const void *value);
+void *pthread_getspecific(pthread_key_t key);
+int pthread_key_create(pthread_key_t *key, void (*destructor)(void *));
+int pthread_key_delete(pthread_key_t key);
+
 #ifdef __cplusplus
 }
 #endif