[event_log] Stop leaking semaphore resources

With the global state and fork, we are left with uncleaned resources. Isolate mulitprocessing.Value in a function so we stop the leak. Bug: 353656374 Change-Id: If50bb544bda12b72f00c02bc1d2c0d19de000b88 Reviewed-on: https://gerrit-review.googlesource.com/c/git-repo/+/440261 Commit-Queue: Josip Sokcevic <[email protected]> Reviewed-by: Gavin Mak <[email protected]> Tested-by: Josip Sokcevic <[email protected]>
GerritCodeReview · Oct 24, 2024 · ae384f8 · ae384f8
1 parent 70a4e64
commit ae384f8
Showing 1 changed file with 10 additions and 2 deletions.
diff --git a/event_log.py b/event_log.py
@@ -168,8 +168,10 @@ def Write(self, filename):
                 f.write("\n")
 
 
-# An integer id that is unique across this invocation of the program.
-_EVENT_ID = multiprocessing.Value("i", 1)
+# An integer id that is unique across this invocation of the program, to be set
+# by the first Add event. We can't set it here since it results in leaked
+# resources (see: https://issues.gerritcodereview.com/353656374).
+_EVENT_ID = None
 
 
 def _NextEventId():
@@ -178,6 +180,12 @@ def _NextEventId():
     Returns:
         A unique, to this invocation of the program, integer id.
     """
+    global _EVENT_ID
+    if _EVENT_ID is None:
+        # There is a small chance of race condition - two parallel processes
+        # setting up _EVENT_ID. However, we expect TASK_COMMAND to happen before
+        # mp kicks in.
+        _EVENT_ID = multiprocessing.Value("i", 1)
     with _EVENT_ID.get_lock():
         val = _EVENT_ID.value
         _EVENT_ID.value += 1