Enhancements on wasm function execution time statistic (#2985)

Enhance the statistic of wasm function execution time, or the performance
profiling feature:
- Add os_time_thread_cputime_us() to get the cputime of a thread,
  and use it to calculate the execution time of a wasm function
- Support the statistic of the children execution time of a function,
  and dump it in wasm_runtime_dump_perf_profiling
- Expose two APIs:
  wasm_runtime_sum_wasm_exec_time
  wasm_runtime_get_wasm_func_exec_time

And rename os_time_get_boot_microsecond to os_time_get_boot_us.

core/shared/platform/alios/alios_time.c

@@ -6,7 +6,14 @@
 #include "platform_api_vmcore.h"
 
 uint64
-os_time_get_boot_microsecond()
+os_time_get_boot_us()
 {
     return (uint64)aos_now_ms() * 1000;
 }
+
+uint64
+os_time_thread_cputime_us(void)
+{
+    /* FIXME if u know the right api */
+    return os_time_get_boot_us();
+}

core/shared/platform/common/freertos/freertos_time.c

@@ -6,8 +6,15 @@
 #include "platform_api_vmcore.h"
 
 uint64
-os_time_get_boot_microsecond()
+os_time_get_boot_us()
 {
     TickType_t ticks = xTaskGetTickCount();
     return (uint64)1000 * 1000 / configTICK_RATE_HZ * ticks;
 }
+
+uint64
+os_time_thread_cputime_us(void)
+{
+    /* FIXME if u know the right api */
+    return os_time_get_boot_us();
+}

core/shared/platform/common/posix/posix_time.c

@@ -6,7 +6,7 @@
 #include "platform_api_vmcore.h"
 
 uint64
-os_time_get_boot_microsecond()
+os_time_get_boot_us()
 {
     struct timespec ts;
     if (clock_gettime(CLOCK_MONOTONIC, &ts) != 0) {
@@ -15,3 +15,14 @@ os_time_get_boot_microsecond()
 
     return ((uint64)ts.tv_sec) * 1000 * 1000 + ((uint64)ts.tv_nsec) / 1000;
 }
+
+uint64
+os_time_thread_cputime_us()
+{
+    struct timespec ts;
+    if (clock_gettime(CLOCK_THREAD_CPUTIME_ID, &ts) != 0) {
+        return 0;
+    }
+
+    return ((uint64)ts.tv_sec) * 1000 * 1000 + ((uint64)ts.tv_nsec) / 1000;
+}

core/shared/platform/esp-idf/espidf_platform.c

@@ -36,11 +36,18 @@ os_vprintf(const char *format, va_list ap)
 }
 
 uint64
-os_time_get_boot_microsecond(void)
+os_time_get_boot_us(void)
 {
     return (uint64)esp_timer_get_time();
 }
 
+uint64
+os_time_thread_cputime_us(void)
+{
+    /* FIXME if u know the right api */
+    return os_time_get_boot_us();
+}
+
 uint8 *
 os_thread_get_stack_boundary(void)
 {

core/shared/platform/include/platform_api_vmcore.h

@@ -64,7 +64,13 @@ os_vprintf(const char *format, va_list ap);
  * Get microseconds after boot.
  */
 uint64
-os_time_get_boot_microsecond(void);
+os_time_get_boot_us(void);
+
+/**
+ * Get thread-specific CPU-time clock in microseconds
+ */
+uint64
+os_time_thread_cputime_us(void);
 
 /**
  * Get current thread id.

core/shared/platform/linux-sgx/sgx_time.c

@@ -26,7 +26,7 @@ ocall_clock_nanosleep(int *p_ret, unsigned clock_id, int flags,
                       const void *rem_buf, unsigned int rem_buf_size);
 
 uint64
-os_time_get_boot_microsecond()
+os_time_get_boot_us()
 {
 #ifndef SGX_DISABLE_WASI
     struct timespec ts;
@@ -40,6 +40,21 @@ os_time_get_boot_microsecond()
 #endif
 }
 
+uint64
+os_time_thread_cputime_us(void)
+{
+#ifndef SGX_DISABLE_WASI
+    struct timespec ts;
+    if (clock_gettime(CLOCK_THREAD_CPUTIME_ID, &ts) != 0) {
+        return 0;
+    }
+
+    return ((uint64)ts.tv_sec) * 1000 * 1000 + ((uint64)ts.tv_nsec) / 1000;
+#else
+    return 0;
+#endif
+}
+
 #ifndef SGX_DISABLE_WASI
 
 int

core/shared/platform/riot/riot_time.c

@@ -10,25 +10,32 @@
 
 #if IS_USED(MODULE_ZTIMER64_USEC)
 uint64
-os_time_get_boot_microsecond()
+os_time_get_boot_us()
 {
     return ztimer64_now(ZTIMER64_USEC);
 }
 #elif IS_USED(MODULE_ZTIMER64_MSEC)
 uint64
-os_time_get_boot_microsecond()
+os_time_get_boot_us()
 {
     return ztimer64_now(ZTIMER64_MSEC) * 1000;
 }
 #else
 #ifdef __GNUC__
 __attribute__((weak)) uint64
-os_time_get_boot_microsecond();
+os_time_get_boot_us();
 #endif
 uint64
-os_time_get_boot_microsecond()
+os_time_get_boot_us()
 {
     static uint64_t times;
     return ++times;
 }
 #endif
+
+uint64
+os_time_thread_cputime_us(void)
+{
+    /* FIXME if u know the right api */
+    return os_time_get_boot_us();
+}

core/shared/platform/rt-thread/rtt_platform.c

@@ -120,13 +120,20 @@ os_vprintf(const char *format, va_list ap)
 }
 
 uint64
-os_time_get_boot_microsecond(void)
+os_time_get_boot_us(void)
 {
     uint64 ret = rt_tick_get() * 1000;
     ret /= RT_TICK_PER_SECOND;
     return ret;
 }
 
+uint64
+os_time_thread_cputime_us(void)
+{
+    /* FIXME if u know the right api */
+    return os_time_get_boot_us();
+}
+
 korp_tid
 os_self_thread(void)
 {

core/shared/platform/windows/win_time.c

@@ -6,7 +6,7 @@
 #include "platform_api_vmcore.h"
 
 uint64
-os_time_get_boot_microsecond()
+os_time_get_boot_us()
 {
     struct timespec ts;
 #if defined(__MINGW32__)
@@ -18,3 +18,10 @@ os_time_get_boot_microsecond()
 
     return ((uint64)ts.tv_sec) * 1000 * 1000 + ((uint64)ts.tv_nsec) / 1000;
 }
+
+uint64
+os_time_thread_cputime_us(void)
+{
+    /* FIXME if u know the right api */
+    return os_time_get_boot_us();
+}

core/shared/platform/zephyr/zephyr_time.c

@@ -6,7 +6,14 @@
 #include "platform_api_vmcore.h"
 
 uint64
-os_time_get_boot_microsecond()
+os_time_get_boot_us()
 {
     return k_uptime_get() * 1000;
 }
+
+uint64
+os_time_thread_cputime_us(void)
+{
+    /* FIXME if u know the right api */
+    return os_time_get_boot_us();
+}

core/shared/utils/bh_log.c

@@ -31,7 +31,7 @@ bh_log(LogLevel log_level, const char *file, int line, const char *fmt, ...)
 
     self = os_self_thread();
 
-    usec = os_time_get_boot_microsecond();
+    usec = os_time_get_boot_us();
     t = (uint32)(usec / 1000000) % (24 * 60 * 60);
     h = t / (60 * 60);
     t = t % (60 * 60);

core/shared/utils/runtime_timer.c

@@ -38,7 +38,7 @@ struct _timer_ctx {
 uint64
 bh_get_tick_ms()
 {
-    return os_time_get_boot_microsecond() / 1000;
+    return os_time_get_boot_us() / 1000;
 }
 
 uint32