diff --git a/HISTORY b/HISTORY
index fcd5c18..ce88fe2 100644
--- a/HISTORY
+++ b/HISTORY
@@ -1,5 +1,5 @@
 
-Version 1.44  2020-03-21
+Version 1.44  2020-03-22
   * add test file src/tests/test_pthread_lock.c
   * add uniq_skiplist.[hc]
   * add function split_string_ex
diff --git a/src/common_blocked_queue.c b/src/common_blocked_queue.c
index 89efa6d..83481b0 100644
--- a/src/common_blocked_queue.c
+++ b/src/common_blocked_queue.c
@@ -65,7 +65,8 @@ int common_blocked_queue_push(struct common_blocked_queue *queue, void *data)
 		return result;
 	}
 
-    node = (struct common_blocked_node *)fast_mblock_alloc_object(&queue->mblock);
+    node = (struct common_blocked_node *)fast_mblock_alloc_object(
+            &queue->mblock);
     if (node == NULL)
     {
         pthread_mutex_unlock(&(queue->lock));
@@ -202,9 +203,11 @@ void common_blocked_queue_free_all_nodes(struct common_blocked_queue *queue,
 {
     struct common_blocked_node *deleted;
 
+    pthread_mutex_lock(&(queue->lock));
     while (node != NULL) {
         deleted = node;
         node = node->next;
         fast_mblock_free_object(&queue->mblock, deleted);
     }
+    pthread_mutex_unlock(&(queue->lock));
 }
diff --git a/src/tests/Makefile b/src/tests/Makefile
index a2e621d..6e6adc3 100644
--- a/src/tests/Makefile
+++ b/src/tests/Makefile
@@ -8,7 +8,7 @@ ALL_PRGS = test_allocator test_skiplist test_multi_skiplist test_mblock test_blo
            test_id_generator test_ini_parser test_char_convert test_char_convert_loader \
            test_logger test_skiplist_set test_crc32 test_thourands_seperator test_sched_thread \
            test_json_parser test_pthread_lock test_uniq_skiplist test_split_string \
-           test_server_id_func test_pipe
+           test_server_id_func test_pipe test_atomic
 
 all: $(ALL_PRGS)
 .c:
diff --git a/src/tests/test_atomic.c b/src/tests/test_atomic.c
new file mode 100644
index 0000000..bf45d9c
--- /dev/null
+++ b/src/tests/test_atomic.c
@@ -0,0 +1,108 @@
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <math.h>
+#include <time.h>
+#include <inttypes.h>
+#include <sys/time.h>
+#include "fastcommon/logger.h"
+#include "fastcommon/shared_func.h"
+#include "fastcommon/sched_thread.h"
+#include "fastcommon/pthread_func.h"
+#include "fastcommon/ini_file_reader.h"
+#include "fastcommon/fast_allocator.h"
+
+#define LOOP_COUNT (100 * 1000 * 1000)
+#define barrier()  __asm__ __volatile__("" ::: "memory")
+
+static volatile int counter = 0;
+
+#define THREAD_COUNT 9
+pthread_t tids[THREAD_COUNT];
+
+void *thread2_func(void *args)
+{
+    int i;
+    for (i=0; i<LOOP_COUNT; i++) {
+        __sync_add_and_fetch(&counter, 1);
+    }
+    return NULL;
+}
+
+void *thread1_func(void *args)
+{
+    int i;
+    for (i=0; i<LOOP_COUNT; i++) {
+        __sync_sub_and_fetch(&counter, 1);
+    }
+    return NULL;
+}
+
+void *wait_thread_func(void *args)
+{
+    int i;
+    for (i=0; i<LOOP_COUNT; i++) {
+        __sync_add_and_fetch(&counter, 0);
+    }
+    return NULL;
+}
+
+
+int test()
+{
+    int result;
+    int i;
+    pthread_t *tid;
+
+    tid = tids;
+    for (i=0; i<THREAD_COUNT / 2; i++) {
+        if ((result=pthread_create(tid++, NULL, thread1_func, NULL)) != 0) {
+            logError("file: "__FILE__", line: %d, "
+                    "create thread fail, "
+                    "errno: %d, error info: %s",
+                    __LINE__, result, STRERROR(result));
+            return result;
+        }
+    }
+
+    for (i=0; i<THREAD_COUNT / 2; i++) {
+        if ((result=pthread_create(tid++, NULL, thread2_func, NULL)) != 0) {
+            logError("file: "__FILE__", line: %d, "
+                    "create thread fail, "
+                    "errno: %d, error info: %s",
+                    __LINE__, result, STRERROR(result));
+            return result;
+        }
+    }
+
+    if ((result=pthread_create(tid++, NULL, wait_thread_func, NULL)) != 0) {
+        logError("file: "__FILE__", line: %d, "
+                "create thread fail, "
+                "errno: %d, error info: %s",
+                __LINE__, result, STRERROR(result));
+        return result;
+    }
+
+    for (i=0; i<THREAD_COUNT; i++) {
+        pthread_join(tids[i], NULL);
+    }
+    return 0;
+}
+
+int main(int argc, char *argv[])
+{
+	int result;
+	int64_t start_time;
+
+	log_init();
+	srand(time(NULL));
+	g_log_context.log_level = LOG_DEBUG;
+	
+	start_time = get_current_time_ms();
+    result = test();
+	printf("counter: %d, time used: %"PRId64" ms\n",
+            __sync_add_and_fetch(&counter, 0),
+            get_current_time_ms() - start_time);
+
+	return result;
+}