Worker Pool 防止饿死 by yousongyang · Pull Request #51 · owent/libatapp

yousongyang · 2026-05-19T03:23:43Z

No description provided.

owent

AICR problem for src/atframe/modules/worker_pool_module.cpp:282

owent · 2026-05-19T03:26:50Z

  worker_tick_action_container_type tick_handles_;
  std::recursive_mutex tick_handle_lock_;
  std::atomic<std::chrono::microseconds::rep> current_tick_interval_us_;
+  std::chrono::microseconds::rep current_tick_second_busy_us_;


MEDIUM
·
correctness
新增成员变量 current_tick_second_busy_us_ 和 current_tick_second_waited_us_ 使用普通标量类型，而同类中类似用途的变量 cpu_time_busy_us_ 和 cpu_time_sleep_us_ 使用 std::atomic 并通过 fetch_add 更新（见第473、491行）。如果这些新变量预期会被其他线程读取（例如用于诊断或监控），则存在数据竞争风险。建议：若需跨线程访问，应声明为 std::atomic 类型；若仅在工作线程内部访问，建议添加注释说明设计意图。

Location: src/atframe/modules/worker_pool_module.cpp:282-283

Referenced code (src/atframe/modules/worker_pool_module.cpp:282-283):

std::recursive_mutex tick_handle_lock_; std::atomic<std::chrono::microseconds::rep> current_tick_interval_us_; std::chrono::microseconds::rep current_tick_second_busy_us_; std::chrono::microseconds::rep current_tick_second_waited_us_; std::atomic<std::chrono::microseconds::rep> cpu_time_busy_us_;

owent · 2026-05-19T03:26:52Z

AI Code Review

Updated: 2026-05-19 03:33:57 UTC | Commit: 27ae5bc

AI Code Review Summary

worker_pool_module.cpp 时序追踪问题

Target: Worker Pool 防止饿死
Author: @owent
Branch: main
Reviewers: @owent

审查范围

审查了 src/atframe/modules/worker_pool_module.cpp 中新增的每秒时间追踪和 preserve sleep 逻辑。

发现问题

中等严重度 (正确性): current_tick_second_waited_us_ 累加的是预期等待时间而非实际等待时间，可能导致 preserve sleep 过早触发。

背景

新增的 current_tick_second_busy_us_ 和 current_tick_second_waited_us_ 用于追踪每秒的忙碌和等待时间，当累计达到1秒时触发 preserve sleep。但 wait 计数器在 wait_for 执行前就已累加预期值，而非实际等待时长。

Problems (1)

#	Severity	Category	Location	Message
0	MEDIUM	correctness	`src/atframe/modules/worker_pool_module.cpp:463`	`current_tick_second_waited_us_` 在第463行累加的是预期等待时间而非实际等待时间。当条件变量在第466行被提前唤醒（spurious wakeup 或通知）时，计数器会高估实际经过的墙钟时间，导致第477行的检查条件过早触发 preserve sleep。例如：预期等待500ms但实际只等待了100ms时，计数器仍增加500ms，与 busy 时间求和后可能远超1秒墙钟时间。

Code reference: `src/atframe/modules/worker_pool_module.cpp:463`

      if (busy_end_time - start_time < tick_interval) {
        std::unique_lock<std::mutex> lk_cv(self->waker_lock_);
        self->current_tick_second_waited_us_ += std::chrono::duration_cast<std::chrono::microseconds>(tick_interval - (busy_end_time - start_time)).count();

        self->status_.store(static_cast<uint8_t>(worker_status::kSleeping), std::memory_order_release);

Powered by AICodeReviewer*

Open Issues (1)

[MEDIUM] correctness — src/atframe/modules/worker_pool_module.cpp:463 (new in 27ae5bc)

Resolved (1)

The following previously reported issues are no longer present:

~~862f8171808563e6~~ ✅ Resolved

Copilot

Pull request overview

This PR adds worker-pool throttling intended to prevent starvation by tracking per-worker busy/wait time and inserting a preserved sleep window once accumulated tick time reaches one second.

Changes:

Adds per-worker counters for current-second busy and waited microseconds.
Adds a worker-set preserve interval defaulting to 8000µs.
Inserts an extra condition-variable wait after each accumulated second of worker activity.

Comments suppressed due to low confidence (1)

src/atframe/modules/worker_pool_module.cpp:481

This added line exceeds the repository's .clang-format ColumnLimit: 120; please wrap it or run clang-format to keep formatting consistent.

        auto wait_preserve_us = self->get_owner().configure_tick_preserve_microseconds_in_second.load(std::memory_order_acquire);

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

+      self->current_tick_second_busy_us_ += std::chrono::duration_cast<std::chrono::microseconds>(busy_end_time - start_time).count();
      if (busy_end_time - start_time < tick_interval) {
        std::unique_lock<std::mutex> lk_cv(self->waker_lock_);
+        self->current_tick_second_waited_us_ += std::chrono::duration_cast<std::chrono::microseconds>(tick_interval - (busy_end_time - start_time)).count();


+        auto wait_preserve_us = self->get_owner().configure_tick_preserve_microseconds_in_second.load(std::memory_order_acquire);
+        if (wait_preserve_us <= 0) {
+          wait_preserve_us = 8000;
+        }
+
+        std::unique_lock<std::mutex> lk_cv(self->waker_lock_);
+        self->status_.store(static_cast<uint8_t>(worker_status::kSleeping), std::memory_order_release);
+        self->waker_cv_.wait_for(lk_cv, std::chrono::microseconds(wait_preserve_us));
+        self->status_.store(static_cast<uint8_t>(worker_status::kRunning), std::memory_order_release);
+
+        self->cpu_time_sleep_us_.fetch_add(wait_preserve_us, std::memory_order_release);


+      if (self->current_tick_second_busy_us_ + self->current_tick_second_waited_us_ >= 1000000) {
+        self->current_tick_second_busy_us_ = 0;
+        self->current_tick_second_waited_us_ = 0;
+
+        auto wait_preserve_us = self->get_owner().configure_tick_preserve_microseconds_in_second.load(std::memory_order_acquire);


+      self->current_tick_second_busy_us_ += std::chrono::duration_cast<std::chrono::microseconds>(busy_end_time - start_time).count();
      if (busy_end_time - start_time < tick_interval) {
        std::unique_lock<std::mutex> lk_cv(self->waker_lock_);
+        self->current_tick_second_waited_us_ += std::chrono::duration_cast<std::chrono::microseconds>(tick_interval - (busy_end_time - start_time)).count();


owent

AICR problem for src/atframe/modules/worker_pool_module.cpp:491

owent · 2026-05-19T03:30:05Z

+        self->waker_cv_.wait_for(lk_cv, std::chrono::microseconds(wait_preserve_us));
+        self->status_.store(static_cast<uint8_t>(worker_status::kRunning), std::memory_order_release);
+
+        self->cpu_time_sleep_us_.fetch_add(wait_preserve_us, std::memory_order_release);


MEDIUM
·
correctness
cpu_time_sleep_us_ 的统计方式不一致：普通 tick 等待使用实际测量的睡眠时间（第 469-473 行），但新增的 preserve sleep 直接使用计划时间 wait_preserve_us（第 491 行）。如果 preserve sleep 因虚假唤醒或 waker_cv_ 被通知而提前结束，cpu_time_sleep_us_ 会高估实际睡眠时间。建议在第 489 行后添加类似第 469-473 行的实际时间测量逻辑。

Location: src/atframe/modules/worker_pool_module.cpp:491

Referenced code (src/atframe/modules/worker_pool_module.cpp:491):

self->status_.store(static_cast<uint8_t>(worker_status::kRunning), std::memory_order_release); self->cpu_time_sleep_us_.fetch_add(wait_preserve_us, std::memory_order_release); } }

owent

AICR problem for src/atframe/modules/worker_pool_module.cpp:463

owent · 2026-05-19T03:33:56Z

+      self->current_tick_second_busy_us_ += std::chrono::duration_cast<std::chrono::microseconds>(busy_end_time - start_time).count();
      if (busy_end_time - start_time < tick_interval) {
        std::unique_lock<std::mutex> lk_cv(self->waker_lock_);
+        self->current_tick_second_waited_us_ += std::chrono::duration_cast<std::chrono::microseconds>(tick_interval - (busy_end_time - start_time)).count();


MEDIUM
·
correctness
current_tick_second_waited_us_ 在第463行累加的是预期等待时间而非实际等待时间。当条件变量在第466行被提前唤醒（spurious wakeup 或通知）时，计数器会高估实际经过的墙钟时间，导致第477行的检查条件过早触发 preserve sleep。例如：预期等待500ms但实际只等待了100ms时，计数器仍增加500ms，与 busy 时间求和后可能远超1秒墙钟时间。

Location: src/atframe/modules/worker_pool_module.cpp:463

Referenced code (src/atframe/modules/worker_pool_module.cpp:463):

if (busy_end_time - start_time < tick_interval) { std::unique_lock<std::mutex> lk_cv(self->waker_lock_); self->current_tick_second_waited_us_ += std::chrono::duration_cast<std::chrono::microseconds>(tick_interval - (busy_end_time - start_time)).count(); self->status_.store(static_cast<uint8_t>(worker_status::kSleeping), std::memory_order_release);

Suggested fix:
考虑在 wait_for 返回后测量实际等待时间并累加，或使用墙钟时间戳来判断是否已过一秒。

Worker Pool 防止饿死

27ae5bc

Copilot AI review requested due to automatic review settings May 19, 2026 03:23

Copilot started reviewing on behalf of yousongyang May 19, 2026 03:24 View session

owent reviewed May 19, 2026

View reviewed changes

owent added aicr aicr:problem:medium aicr:reviewed labels May 19, 2026

Copilot AI reviewed May 19, 2026

View reviewed changes

owent reviewed May 19, 2026

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Worker Pool 防止饿死#51

Worker Pool 防止饿死#51
yousongyang wants to merge 1 commit into
owent:mainfrom
yousongyang:main

yousongyang commented May 19, 2026

Uh oh!

owent left a comment

Uh oh!

owent May 19, 2026

Uh oh!

owent commented May 19, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

owent left a comment

Uh oh!

owent May 19, 2026

Uh oh!

owent left a comment

Uh oh!

owent May 19, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

yousongyang commented May 19, 2026

Uh oh!

owent left a comment

Choose a reason for hiding this comment

Uh oh!

owent May 19, 2026

Choose a reason for hiding this comment

Uh oh!

owent commented May 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

AI Code Review

AI Code Review Summary

worker_pool_module.cpp 时序追踪问题

审查范围

发现问题

背景

Problems (1)

Code reference: src/atframe/modules/worker_pool_module.cpp:463

Open Issues (1)

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

owent left a comment

Choose a reason for hiding this comment

Uh oh!

owent May 19, 2026

Choose a reason for hiding this comment

Uh oh!

owent left a comment

Choose a reason for hiding this comment

Uh oh!

owent May 19, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

owent commented May 19, 2026 •

edited

Loading

Code reference: `src/atframe/modules/worker_pool_module.cpp:463`