test(kmod): add kunit test for receive_msg

[Koichi98]

Description

Added receive_msg kunit tests:

• test_case_no_topic_when_receive
  check error handling when topic_wrapper is not found.

• test_case_no_subscriber_when_receive
  check error handling when sub_info is not found.

• test_case_no_publish_no_receive
  check that receive_msg will return without any error and without getting any message

• test_case_receive_one
  simple test for receiving one message

• test_case_receive_msg_sub_qos_depth_smaller_than_publish_num_smaller_than_pub_qos_depth
  check for correct behavior when sub_qos_depth < publish_num < pub_qos_depth

• test_case_receive_msg_publish_num_smaller_than_sub_qos_depth_smaller_than_pub_qos_depth
  check for correct behavior when publish_num < sub_qos_depth < pub_qos_depth

• test_case_receive_msg_sub_qos_depth_smaller_than_pub_qos_depth_smaller_than_publish_num
  check for correct behavior when sub_qos_depth < pub_qos_depth = MAX_QOS_DEPTH < publish_num

• test_case_receive_msg_publish_num_and_sub_qos_depth_and_pub_qos_depth_are_all_max_qos_depth
  check for correct behavior when sub_qos_depth = pub_qos_depth = publish_num = MAX_QOS_DEPTH

• test_case_too_many_rc
  check error handling when the number of pub/sub entities referencing to the message exceeds MAX_REFERENCING_PUBSUB_NUM_PER_ENTRY

• test_case_one_new_pub
  check that receive_msg gets the information of the publisher correctly, when the new publisher appears

• test_case_pubsub_in_same_process
  check that receive_msg doesn't get the information of the publisher, when the new publisher appears is the same process as the subscriber

• test_case_2pub_in_same_process
  check that receive_msg get the information of the publisher only once, when the new publisher appears is the same process as the other publisher.

• test_case_2sub_in_same_process
  check that receive_msg doesn't get the information of the publisher, if the other subscriber in the same process has already mapped the memory of the publisher process.

• test_case_subscriber_too_many_mapping_processes
  add many subscribers to reach MAX_PROCESS_NUM_PER_MEMPOOL. This is a complex test and will be implemented later.

Related links

How was this PR tested?

• Autoware (required)
• bash scripts/e2e_test_1to1_with_ros2sub (required)
• bash scripts/e2e_test_2to2 (required)
• sample application

Notes for reviewers

[veqcc]

実装を見てパッとわかりにくい（何番目の引数が qos かわからん）ので、
subscriber_add の書き方を真似るのは微妙ですか？個人的にはわかりやすいと思っており

agnocast/agnocast_kmod/agnocast_kunit/agnocast_kunit_subscriber_add.c

Line 219 in da9ab69

// publisher_qos_depth > entries_num > subscriber_qos_depth

// publisher_qos_depth > entries_num > subscriber_qos_depth
void test_case_subscriber_add_normal_with_many_entries_and_transient_local_1(struct kunit * test) {
  ...
  const uint32_t publisher_qos_depth = 7;
  const uint32_t subscriber_qos_depth = 3;
  const uint32_t entries_num = 5;
  ...

[Koichi98]

関数が見た目的にでかくなることを避けたのですがさすがに提案してもらった方法の方がわかりやすそうですね。

[Koichi98]

修正します！

[Koichi98]

修正しました！