add fix timer race code #412

M-iceberg · 2024-11-14T03:15:38Z

No description provided.

M-iceberg · 2024-11-14T03:15:49Z

/azp run #Resolved

azure-pipelines · 2024-11-14T03:16:03Z

Azure Pipelines successfully started running 1 pipeline(s). #Resolved

M-iceberg · 2024-11-19T00:07:57Z

/azp run #Resolved

azure-pipelines · 2024-11-19T00:08:11Z

Azure Pipelines successfully started running 1 pipeline(s). #Resolved

darobs · 2024-11-19T00:51:59Z

interfaces/inc/c_pal/threadpool.h

@@ -21,10 +21,13 @@ extern "C" {
 #endif

 typedef struct THREADPOOL_TAG THREADPOOL;
-typedef struct TIMER_INSTANCE_TAG* TIMER_INSTANCE_HANDLE;
+typedef struct TIMER_INSTANCE_TAG TIMER_INSTANCE;
+typedef struct TIMER_INSTANCE_TAG* TIMER_INSTANCE_HANDLE; //todo: if this one is needed


should probably remove TIMER_INSTANCE_HANDLE #Resolved

Not removed, I can still see it.

darobs · 2024-11-19T00:54:05Z

linux/devdoc/threadpool_linux_requirements.md


-MOCKABLE_FUNCTION(, int, threadpool_timer_restart, TIMER_INSTANCE_HANDLE, timer, uint32_t, start_delay_ms, uint32_t, timer_period_ms);
+MOCKABLE_FUNCTION(, int, threadpool_timer_restart, THANDLE(TIMER_INSTANCE)*, timer, uint32_t, start_delay_ms, uint32_t, timer_period_ms);


)*

timer is just THANDLE(TIMER_INSTANCE), not THANDLE(TIMER_INSTANCE)* Same for line 92. #Resolved

darobs · 2024-11-19T01:28:47Z

linux/tests/threadpool_linux_ut/threadpool_linux_ut.c

@@ -1140,9 +1141,10 @@ TEST_FUNCTION(threadpool_timer_start_with_NULL_work_function_context_succeeds)
    // arrange
    THANDLE(THREADPOOL) threadpool = test_create_and_open_threadpool();

-    TIMER_INSTANCE_HANDLE timer_instance;
+    THANDLE(TIMER_INSTANCE) timer_instance = NULL;


timer_instance = NULL

Do we have to assign this to be NULL for threadpool_timer_start to work? I can see this might be an annoyance when the timer is part of a struct, like it generally is. #Resolved

will getting segmentation fault if not set this THANDLE to NULL

The common pattern we have in the code that uses this is

typedef struct SOME_STRUCT_TAG { // ... THANDLE(TIMER_INSTANCE) timer; // ... } SOME_STRUCT; // then in the "_create" function: SOME_STRUCT_HANDLE result = (SOME_STRUCT_HANDLE)malloc(sizeof(SOME_STRUCT); threadpool_timer_start(threadpool, start_delay, period, test_work_function, context, &result->timer);

The timer is being passed in uninitialized, and so requiring it to be initialized in the test implies that the above pattern will need to change. BTW, this comment goes along with using THANDLE_INITIALIZE_MOVE instead of THANDLE_MOVE

modified to initialize_move

darobs · 2024-11-19T01:35:59Z

win32/src/threadpool_win32.c

+                        threadpool_internal_set_timer(tp_timer, start_delay_ms, timer_period_ms);
+
+                        /* Codes_SRS_THREADPOOL_WIN32_42_009: [ threadpool_timer_start shall return the allocated handle in timer_handle. ]*/
+                        THANDLE_MOVE(TIMER_INSTANCE)(timer_handle, &timer_temp);


THANDLE_MOVE(TIMER_INSTANCE)(timer_handle, &timer_temp);

Would THANDLE_INITIALIZE_MOVE make more sense here? #Resolved

Same for Linux?

darobs · 2024-11-19T02:22:47Z

linux/linux_reals/real_threadpool.h

@@ -18,8 +18,7 @@
        threadpool_schedule_work, \
        threadpool_timer_start, \
        threadpool_timer_restart, \
-        threadpool_timer_cancel, \
-        threadpool_timer_destroy \
+        threadpool_timer_cancel
    ) \


build failed, missing \ at the end of this line #Resolved

azure-pipelines · 2024-11-19T21:12:23Z

No pipelines are associated with this pull request. #Resolved

azure-pipelines · 2024-11-19T21:12:42Z

No pipelines are associated with this pull request. #Resolved

M-iceberg · 2024-11-19T22:25:12Z

/azp run #Resolved

azure-pipelines · 2024-11-19T22:25:25Z

Azure Pipelines successfully started running 1 pipeline(s). #Resolved

M-iceberg · 2024-11-19T23:37:53Z

/azp run #Resolved

azure-pipelines · 2024-11-19T23:38:03Z

Azure Pipelines successfully started running 1 pipeline(s). #Resolved

dcristoloveanu · 2024-11-20T00:34:37Z

interfaces/inc/c_pal/threadpool.h

@@ -21,10 +21,13 @@ extern "C" {
 #endif

 typedef struct THREADPOOL_TAG THREADPOOL;
+typedef struct TIMER_INSTANCE_TAG TIMER_INSTANCE;
 typedef struct TIMER_INSTANCE_TAG* TIMER_INSTANCE_HANDLE;


TIMER_INSTANCE_HANDLE

Who uses the TIMER_INSTANCE_HANDLE? I am imagining that after this change there should be only one, right? #Resolved

Or are there more PRs planned? If so, we should add a comment explaining that, otherwise it is sort of confusing.

removed the timer_instance_handle, not using anymore

dcristoloveanu · 2024-11-20T00:59:48Z

win32/tests/threadpool_win32_int/threadpool_win32_int.c

    for (uint32_t i = 0; i < N_TIMERS; i++)
    {
-        ASSERT_ARE_EQUAL(int, 0, threadpool_timer_start(threadpool, 100, 500, work_function, (void*)&call_count, &timers[i]));
+        THANDLE_INITIALIZE(TIMER_INSTANCE)((void*) & timers[i], NULL);
+        ASSERT_ARE_EQUAL(int, 0, threadpool_timer_start(threadpool, 100, 500, work_function, (void*)&call_count, (void*) & timers[i]));


(void*)

Needed? #Resolved

dcristoloveanu · 2024-11-20T01:00:17Z

win32/tests/threadpool_win32_int/threadpool_win32_int.c

+            (void)InterlockedDecrement(&chaos_test_data->timers_starting);
+            WakeByAddressSingle((PVOID)&chaos_test_data->timers_starting);
+        }
+        break;


break;

break should go within statement scope (before line 962). #Resolved

dcristoloveanu · 2024-11-20T01:00:33Z

win32/tests/threadpool_win32_int/threadpool_win32_int.c

+            (void)InterlockedDecrement(&chaos_test_data->timers_starting);
+            WakeByAddressSingle((PVOID)&chaos_test_data->timers_starting);
+        }
+        case TEST_ACTION_SCHEDULE_WORK_ITEM:


Missing break; #Resolved

dcristoloveanu · 2024-11-20T01:01:42Z

win32/tests/threadpool_win32_int/threadpool_win32_int.c

+                if (threadpool_schedule_work_item(chaos_test_data->threadpool, chaos_test_data->work_item_context) == 0)
+                {
+                    (void)InterlockedIncrement64(&chaos_test_data->expected_call_count);
+                }


}

Consider adding the else branch empty (just with a comment saying that we don't care if it failed. #Resolved

dcristoloveanu · 2024-11-20T01:02:09Z

    break;

Break should go one line higher #Resolved

Refers to: win32/tests/threadpool_win32_int/threadpool_win32_int.c:1064 in 491e0ff. [](commit_id = 491e0ff, deletion_comment = False)

dcristoloveanu · 2024-11-20T01:02:17Z

    break;

Here and below.

In reply to: 2487086153

Refers to: win32/tests/threadpool_win32_int/threadpool_win32_int.c:1064 in 491e0ff. [](commit_id = 491e0ff, deletion_comment = False)

dcristoloveanu · 2024-11-20T01:02:25Z

win32/tests/threadpool_win32_int/threadpool_win32_int.c

@@ -1008,12 +1099,9 @@ static DWORD WINAPI chaos_thread_with_timers_no_lock_func(LPVOID lpThreadParamet
        }


}

Missing break; #Resolved

dcristoloveanu · 2024-11-20T01:03:22Z

win32/tests/threadpool_win32_int/threadpool_win32_int.c

+    // assert that all scheduled items were executed
+    ASSERT_ARE_EQUAL(int64_t, (int64_t)InterlockedAdd64(&chaos_test_data.expected_call_count, 0), (int64_t)InterlockedAdd64(&chaos_test_data.executed_work_functions, 0));
+
+    LogInfo("Chaos test executed %" PRIu64 " work items, %" PRIu64 " timers",


PRIu64

interlocked APIs work with signed values, this should be PRId64. Please check the other places too. #Resolved

dcristoloveanu · 2024-11-20T01:06:28Z

TEST_FUNCTION(schedule_after_close_works_v2)

I think the same chaos tests should be run on Linux also.
As a matter of fact, why are we even having 2 separate test suites for Linux and Windows for the integration part?

Unit tests should be different, but at integration level we should be running exactly the same test code on both platforms and it should pass ....

Refers to: linux/tests/threadpool_linux_int/threadpool_linux_int.c:869 in 491e0ff. [](commit_id = 491e0ff, deletion_comment = False)

dcristoloveanu

🕐

azure-pipelines · 2024-11-20T23:54:45Z

Pull request contains merge conflicts. #Resolved

azure-pipelines · 2024-11-20T23:54:55Z

Pull request contains merge conflicts. #Resolved

M-iceberg · 2024-11-22T23:41:28Z

    break;

done

In reply to: 2487086295

Refers to: win32/tests/threadpool_win32_int/threadpool_win32_int.c:1064 in 491e0ff. [](commit_id = 491e0ff, deletion_comment = False)

M-iceberg · 2024-11-26T00:46:01Z

/azp run

azure-pipelines · 2024-11-26T00:46:07Z

Pull request contains merge conflicts.

M-iceberg · 2024-11-26T00:52:22Z

/azp run

azure-pipelines · 2024-11-26T00:52:34Z

Azure Pipelines successfully started running 1 pipeline(s).

M-iceberg · 2024-12-02T18:12:51Z

/azp run

azure-pipelines · 2024-12-02T18:13:04Z

Azure Pipelines successfully started running 1 pipeline(s).

M-iceberg · 2024-12-02T19:28:53Z

/azp run

azure-pipelines · 2024-12-02T19:29:09Z

Azure Pipelines successfully started running 1 pipeline(s).

M-iceberg · 2024-12-02T19:59:42Z

/azp run

azure-pipelines · 2024-12-02T19:59:53Z

Azure Pipelines successfully started running 1 pipeline(s).

add fix timer race code

81f5069

microsoft-github-policy-service bot requested review from anporumb, avranju, mattdurak, dcristoloveanu, jebrando, darobs, ipg2013 and parth21999 November 14, 2024 03:16

upload linux code change

9284da2

darobs reviewed Nov 19, 2024

View reviewed changes

M-iceberg added 2 commits November 19, 2024 14:22

address comments

70d23f1

small fix

1d74206

microsoft-github-policy-service bot requested a review from darobs November 19, 2024 22:25

fix linux build

491e0ff

dcristoloveanu reviewed Nov 20, 2024

View reviewed changes

dcristoloveanu requested changes Nov 20, 2024

View reviewed changes

address comments

7352c74

microsoft-github-policy-service bot requested a review from dcristoloveanu November 26, 2024 00:45

resolve merge conflict

9b64051

M-iceberg added 2 commits November 25, 2024 17:18

manually resolve conflict

b03551c

resolve merge conflict

64f7d74

resolve merge conflict error

e57a31b

fix test failure due to merge conflict

01d3122


		MOCKABLE_FUNCTION(, int, threadpool_timer_restart, TIMER_INSTANCE_HANDLE, timer, uint32_t, start_delay_ms, uint32_t, timer_period_ms);
		MOCKABLE_FUNCTION(, int, threadpool_timer_restart, THANDLE(TIMER_INSTANCE)*, timer, uint32_t, start_delay_ms, uint32_t, timer_period_ms);

		@@ -1008,12 +1099,9 @@ static DWORD WINAPI chaos_thread_with_timers_no_lock_func(LPVOID lpThreadParamet
		}

add fix timer race code #412

Are you sure you want to change the base?

add fix timer race code #412

Conversation

M-iceberg commented Nov 14, 2024

M-iceberg commented Nov 14, 2024 • edited Loading

azure-pipelines bot commented Nov 14, 2024 • edited by M-iceberg Loading

M-iceberg commented Nov 19, 2024 • edited Loading

azure-pipelines bot commented Nov 19, 2024 • edited by M-iceberg Loading

darobs Nov 19, 2024 • edited by M-iceberg Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

darobs Nov 19, 2024 • edited by M-iceberg Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

darobs Nov 19, 2024 • edited by M-iceberg Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

darobs Nov 19, 2024 • edited by M-iceberg Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

darobs Nov 19, 2024 • edited by M-iceberg Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

azure-pipelines bot commented Nov 19, 2024 • edited by M-iceberg Loading

azure-pipelines bot commented Nov 19, 2024 • edited by M-iceberg Loading

M-iceberg commented Nov 19, 2024 • edited Loading

azure-pipelines bot commented Nov 19, 2024 • edited by M-iceberg Loading

M-iceberg commented Nov 19, 2024 • edited Loading

azure-pipelines bot commented Nov 19, 2024 • edited by M-iceberg Loading

dcristoloveanu Nov 20, 2024 • edited by M-iceberg Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

dcristoloveanu Nov 20, 2024 • edited by M-iceberg Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

dcristoloveanu Nov 20, 2024 • edited by M-iceberg Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

dcristoloveanu Nov 20, 2024 • edited by M-iceberg Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

dcristoloveanu Nov 20, 2024 • edited by M-iceberg Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

dcristoloveanu commented Nov 20, 2024 • edited by M-iceberg Loading

dcristoloveanu commented Nov 20, 2024

dcristoloveanu Nov 20, 2024 • edited by M-iceberg Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

dcristoloveanu Nov 20, 2024 • edited by M-iceberg Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

dcristoloveanu commented Nov 20, 2024

dcristoloveanu left a comment

Choose a reason for hiding this comment

azure-pipelines bot commented Nov 20, 2024 • edited by M-iceberg Loading

azure-pipelines bot commented Nov 20, 2024 • edited by M-iceberg Loading

M-iceberg commented Nov 22, 2024

M-iceberg commented Nov 26, 2024

azure-pipelines bot commented Nov 26, 2024

M-iceberg commented Nov 26, 2024

azure-pipelines bot commented Nov 26, 2024

M-iceberg commented Dec 2, 2024

azure-pipelines bot commented Dec 2, 2024

M-iceberg commented Dec 2, 2024

azure-pipelines bot commented Dec 2, 2024

M-iceberg commented Dec 2, 2024

azure-pipelines bot commented Dec 2, 2024

M-iceberg commented Nov 14, 2024 •

edited

Loading

azure-pipelines bot commented Nov 14, 2024 •

edited by M-iceberg

Loading

M-iceberg commented Nov 19, 2024 •

edited

Loading

azure-pipelines bot commented Nov 19, 2024 •

edited by M-iceberg

Loading

darobs Nov 19, 2024 •

edited by M-iceberg

Loading

darobs Nov 19, 2024 •

edited by M-iceberg

Loading

darobs Nov 19, 2024 •

edited by M-iceberg

Loading

darobs Nov 19, 2024 •

edited by M-iceberg

Loading

darobs Nov 19, 2024 •

edited by M-iceberg

Loading

azure-pipelines bot commented Nov 19, 2024 •

edited by M-iceberg

Loading

azure-pipelines bot commented Nov 19, 2024 •

edited by M-iceberg

Loading

M-iceberg commented Nov 19, 2024 •

edited

Loading

azure-pipelines bot commented Nov 19, 2024 •

edited by M-iceberg

Loading

M-iceberg commented Nov 19, 2024 •

edited

Loading

azure-pipelines bot commented Nov 19, 2024 •

edited by M-iceberg

Loading

dcristoloveanu Nov 20, 2024 •

edited by M-iceberg

Loading

dcristoloveanu Nov 20, 2024 •

edited by M-iceberg

Loading

dcristoloveanu Nov 20, 2024 •

edited by M-iceberg

Loading

dcristoloveanu Nov 20, 2024 •

edited by M-iceberg

Loading

dcristoloveanu Nov 20, 2024 •

edited by M-iceberg

Loading

dcristoloveanu commented Nov 20, 2024 •

edited by M-iceberg

Loading

dcristoloveanu Nov 20, 2024 •

edited by M-iceberg

Loading

dcristoloveanu Nov 20, 2024 •

edited by M-iceberg

Loading

azure-pipelines bot commented Nov 20, 2024 •

edited by M-iceberg

Loading

azure-pipelines bot commented Nov 20, 2024 •

edited by M-iceberg

Loading