tcp_transport: Rework test to split test code and use fixtures

This commit is contained in:
David Cermak 2021-05-19 12:36:36 +02:00
parent f249ddd9ae
commit 6cf077169e
7 changed files with 587 additions and 462 deletions

View File

@ -0,0 +1,41 @@
#ifndef _TCP_TRANSPORT_FIXTURES_H_
#define _TCP_TRANSPORT_FIXTURES_H_
/**
* @brief Structures and types for passing socket options
*/
enum expected_sock_option_types {
SOCK_OPT_TYPE_BOOL,
SOCK_OPT_TYPE_INT,
};
struct expected_sock_option {
int level;
int optname;
int optval;
enum expected_sock_option_types opttype;
};
/**
* @brief Helper test functions for timeout connection tests
*
* This case simulates connection timeout running tcp connect asynchronously with other socket connection
* consuming entire socket listener backlog.
* Important: Both tasks must run on the same core, with listener's prio higher to make sure that
* 1) first the localhost_listener() creates and connects all sockets until the last one blocks
* 2) before the tcp_connect_task() attempts to connect and thus fails with connection timeout
*/
void tcp_transport_test_connection_timeout(esp_transport_handle_t transport_under_test);
/**
* @brief Helper test function to check socket options configured separately by transports
*
* This sets up the connection test to start two tasks, but unlike tcp_transport_test_connection_timeout,
* here we just let the connection to happen or at least open on TCP layer so we get the internal socket
* descriptor. While the connection is in progress or connected, we can check the socket options configured
* by the tcp_transport API.
*/
void tcp_transport_test_socket_options(esp_transport_handle_t transport_under_test, bool async,
const struct expected_sock_option *expected_opts, size_t sock_options_len);
#endif //_TCP_TRANSPORT_FIXTURES_H_

View File

@ -1,455 +0,0 @@
#include "unity.h"
#include "esp_transport.h"
#include "esp_transport_tcp.h"
#include "esp_transport_ssl.h"
#include "esp_transport_ws.h"
#include "test_utils.h"
#include "esp_log.h"
#include "lwip/err.h"
#include "lwip/sockets.h"
#include "lwip/sys.h"
#include <lwip/netdb.h>
#include "freertos/event_groups.h"
#define TCP_CONNECT_DONE (1)
#define TCP_LISTENER_DONE (2)
#define TCP_ACCEPTOR_DONE (4)
#define TCP_LISTENER_ACCEPTED (8)
#define TCP_LISTENER_READY (16)
struct tcp_connect_task_params {
int timeout_ms;
int port;
EventGroupHandle_t tcp_connect_done;
int ret;
int listen_sock;
int accepted_sock;
int last_connect_sock;
bool tcp_listener_failed;
esp_transport_handle_t transport_under_test;
bool accept_connection;
bool consume_sock_backlog;
};
#define TEST_TRANSPORT_BIND_IFNAME() \
struct ifreq ifr; \
ifr.ifr_name[0] = 'l'; \
ifr.ifr_name[1] = 'o'; \
ifr.ifr_name[2] = '\0';
/**
* @brief Recursively connects with a new socket to loopback interface until the last one blocks.
* The last socket is closed upon test teardown, that initiates recursive cleanup (close) for all
* active/connected sockets.
*/
static void connect_once(struct tcp_connect_task_params *params)
{
struct sockaddr_in dest_addr_ip4 = { .sin_addr.s_addr = htonl(INADDR_LOOPBACK),
.sin_family = AF_INET,
.sin_port = htons(params->port) };
int connect_sock = socket(AF_INET, SOCK_STREAM, IPPROTO_IP);
if (connect_sock < 0) {
params->tcp_listener_failed = true;
return;
}
params->last_connect_sock = connect_sock;
int err = connect(connect_sock, (struct sockaddr *)&dest_addr_ip4, sizeof(dest_addr_ip4));
if (err != 0) {
// The last connection is expected to fail here, since the both sockets get closed on test cleanup
return;
}
connect_once(params);
close(connect_sock);
}
/**
* @brief creates a listener (and an acceptor if configured)
*
* if consume_sock_backlog set: connect as many times as possible to prepare an endpoint which
* would make the client block but not complete TCP handshake
*
* if accept_connection set: waiting normally for connection creating an acceptor to mimic tcp-transport endpoint
*/
static void localhost_listener(void *pvParameters)
{
const char* TAG = "tcp_transport_test";
struct tcp_connect_task_params *params = pvParameters;
struct sockaddr_in dest_addr_ip4 = { .sin_addr.s_addr = htonl(INADDR_ANY),
.sin_family = AF_INET,
.sin_port = htons(params->port) };
// Create listener socket and bind it to ANY address
params->listen_sock = socket(AF_INET, SOCK_STREAM, IPPROTO_IP);
int opt = 1;
setsockopt(params->listen_sock, SOL_SOCKET, SO_REUSEADDR, &opt, sizeof(opt));
if (params->listen_sock < 0) {
ESP_LOGE(TAG, "Unable to create socket");
params->tcp_listener_failed = true;
goto failed;
}
int err = bind(params->listen_sock, (struct sockaddr *)&dest_addr_ip4, sizeof(dest_addr_ip4));
if (err != 0) {
ESP_LOGE(TAG, "Socket unable to bind: errno %d", errno);
params->tcp_listener_failed = true;
goto failed;
}
// Listen with backlog set to a low number
err = listen(params->listen_sock, 4);
if (err != 0) {
ESP_LOGE(TAG, "Error occurred during listen: errno %d", errno);
params->tcp_listener_failed = true;
goto failed;
}
// Listener is ready at this point
xEventGroupSetBits(params->tcp_connect_done, TCP_LISTENER_READY);
if (params->consume_sock_backlog) {
// Ideally we would set backlog to 0, but since this is an implementation specific recommendation parameter,
// we recursively create sockets and try to connect to this listener in order to consume the backlog. After
// the backlog is consumed, the last connection blocks (waiting for accept), but at that point we are sure
// that any other connection would also block
connect_once(params);
} else if (params->accept_connection) {
struct sockaddr_storage source_addr;
socklen_t addr_len = sizeof(source_addr);
params->accepted_sock = accept(params->listen_sock, (struct sockaddr *)&source_addr, &addr_len);
if (params->accepted_sock < 0) {
ESP_LOGE(TAG, "Unable to accept connection: errno %d", errno);
goto failed;
}
xEventGroupSetBits(params->tcp_connect_done, TCP_LISTENER_ACCEPTED); // Mark the socket as accepted
// ...and wait for the "acceptor" tests to finish
xEventGroupWaitBits(params->tcp_connect_done, TCP_ACCEPTOR_DONE, true, true, params->timeout_ms * 10);
}
failed:
xEventGroupSetBits(params->tcp_connect_done, TCP_LISTENER_DONE);
vTaskSuspend(NULL);
}
static void tcp_connect_task(void *pvParameters)
{
struct tcp_connect_task_params *params = pvParameters;
params->ret = esp_transport_connect(params->transport_under_test, "localhost", params->port, params->timeout_ms);
if (params->accept_connection) {
// If we test the accepted connection, need to wait until the test completes
xEventGroupWaitBits(params->tcp_connect_done, TCP_ACCEPTOR_DONE, true, true, params->timeout_ms * 10);
}
xEventGroupSetBits(params->tcp_connect_done, TCP_CONNECT_DONE);
vTaskSuspend(NULL);
}
TEST_CASE("tcp_transport: init and deinit transport list", "[tcp_transport][leaks=0]")
{
esp_transport_list_handle_t transport_list = esp_transport_list_init();
esp_transport_handle_t tcp = esp_transport_tcp_init();
esp_transport_list_add(transport_list, tcp, "tcp");
TEST_ASSERT_EQUAL(ESP_OK, esp_transport_list_destroy(transport_list));
}
TEST_CASE("tcp_transport: using ssl transport separately", "[tcp_transport][leaks=0]")
{
esp_transport_handle_t h = esp_transport_ssl_init();
TEST_ASSERT_EQUAL(ESP_OK, esp_transport_destroy(h));
}
TEST_CASE("tcp_transport: using ws transport separately", "[tcp_transport][leaks=0]")
{
esp_transport_handle_t tcp = esp_transport_tcp_init();
esp_transport_handle_t ws = esp_transport_ws_init(tcp);
TEST_ASSERT_EQUAL(ESP_OK, esp_transport_destroy(ws));
TEST_ASSERT_EQUAL(ESP_OK, esp_transport_destroy(tcp));
}
static void transport_connection_timeout_test(esp_transport_handle_t transport_under_test)
{
// This case simulates connection timeout running tcp connect asynchronously with other socket connection
// consuming entire socket listener backlog.
// Important: Both tasks must run on the same core, with listener's prio higher to make sure that
// 1) first the localhost_listener() creates and connects all sockets until the last one blocks
// 2) before the tcp_connect_task() attempts to connect and thus fails with connection timeout
struct tcp_connect_task_params params = { .tcp_connect_done = xEventGroupCreate(),
.timeout_ms = 200,
.port = 80,
.consume_sock_backlog = true,
.transport_under_test = transport_under_test };
TickType_t max_wait = pdMS_TO_TICKS(params.timeout_ms * 10);
TaskHandle_t localhost_listener_task_handle = NULL;
TaskHandle_t tcp_connect_task_handle = NULL;
test_case_uses_tcpip();
// Create listener and connect it with as many sockets until the last one blocks
xTaskCreatePinnedToCore(localhost_listener, "localhost_listener", 4096, (void*)&params, 5, &localhost_listener_task_handle, 0);
// Perform tcp-connect in a separate task to check asynchronously for the timeout
xTaskCreatePinnedToCore(tcp_connect_task, "tcp_connect_task", 4096, (void*)&params, 4, &tcp_connect_task_handle, 0);
// Roughly measure tick-time spent while trying to connect
TickType_t start = xTaskGetTickCount();
EventBits_t bits = xEventGroupWaitBits(params.tcp_connect_done, TCP_CONNECT_DONE, true, true, max_wait);
TickType_t end = xTaskGetTickCount();
TEST_ASSERT_EQUAL(TCP_CONNECT_DONE, TCP_CONNECT_DONE & bits); // Connection has finished
TEST_ASSERT_EQUAL(-1, params.ret); // Connection failed with -1
// Test connection attempt took expected timeout value
TEST_ASSERT_INT_WITHIN(pdMS_TO_TICKS(params.timeout_ms/5), pdMS_TO_TICKS(params.timeout_ms), end-start);
// Closing both parties of the last "blocking" connection to unwind localhost_listener() and let other connected sockets closed
close(params.listen_sock);
close(params.last_connect_sock);
// Cleanup
xEventGroupWaitBits(params.tcp_connect_done, TCP_LISTENER_DONE, true, true, max_wait);
TEST_ASSERT_EQUAL(false, params.tcp_listener_failed);
vEventGroupDelete(params.tcp_connect_done);
test_utils_task_delete(localhost_listener_task_handle);
test_utils_task_delete(tcp_connect_task_handle);
}
TEST_CASE("tcp_transport: connect timeout", "[tcp_transport]")
{
// Init the transport under test
esp_transport_list_handle_t transport_list = esp_transport_list_init();
esp_transport_handle_t tcp = esp_transport_tcp_init();
esp_transport_list_add(transport_list, tcp, "tcp");
transport_connection_timeout_test(tcp);
esp_transport_close(tcp);
esp_transport_list_destroy(transport_list);
}
TEST_CASE("ssl_transport: connect timeout", "[tcp_transport]")
{
// Init the transport under test
esp_transport_list_handle_t transport_list = esp_transport_list_init();
esp_transport_handle_t tcp = esp_transport_tcp_init();
esp_transport_list_add(transport_list, tcp, "tcp");
esp_transport_handle_t ssl = esp_transport_ssl_init();
esp_transport_list_add(transport_list, ssl, "ssl");
transport_connection_timeout_test(ssl);
esp_transport_close(tcp);
esp_transport_close(ssl);
esp_transport_list_destroy(transport_list);
}
TEST_CASE("transport: init and deinit multiple transport items", "[tcp_transport][leaks=0]")
{
esp_transport_list_handle_t transport_list = esp_transport_list_init();
esp_transport_handle_t tcp = esp_transport_tcp_init();
esp_transport_list_add(transport_list, tcp, "tcp");
esp_transport_handle_t ssl = esp_transport_ssl_init();
esp_transport_list_add(transport_list, ssl, "ssl");
esp_transport_handle_t ws = esp_transport_ws_init(tcp);
esp_transport_list_add(transport_list, ws, "ws");
esp_transport_handle_t wss = esp_transport_ws_init(ssl);
esp_transport_list_add(transport_list, wss, "wss");
TEST_ASSERT_EQUAL(ESP_OK, esp_transport_list_destroy(transport_list));
}
// This is a private API of the tcp transport, but needed for socket operation tests
int esp_transport_get_socket(esp_transport_handle_t t);
// Structures and types for passing socket options
enum expected_sock_option_types {
SOCK_OPT_TYPE_BOOL,
SOCK_OPT_TYPE_INT,
};
struct expected_sock_option {
int level;
int optname;
int optval;
enum expected_sock_option_types opttype;
};
static void socket_operation_test(esp_transport_handle_t transport_under_test,
const struct expected_sock_option expected_opts[], size_t sock_options_len)
{
struct tcp_connect_task_params params = { .tcp_connect_done = xEventGroupCreate(),
.timeout_ms = 200,
.port = 80,
.accept_connection = true,
.transport_under_test = transport_under_test };
TickType_t max_wait = pdMS_TO_TICKS(params.timeout_ms * 10);
TaskHandle_t localhost_listener_task_handle = NULL;
TaskHandle_t tcp_connect_task_handle = NULL;
test_case_uses_tcpip();
// Create a listener and wait for it to be ready
xTaskCreatePinnedToCore(localhost_listener, "localhost_listener", 4096, (void*)&params, 5, &localhost_listener_task_handle, 0);
xEventGroupWaitBits(params.tcp_connect_done, TCP_LISTENER_READY, true, true, max_wait);
// Perform tcp-connect in a separate task
xTaskCreatePinnedToCore(tcp_connect_task, "tcp_connect_task", 4096, (void*)&params, 6, &tcp_connect_task_handle, 0);
// Wait till the connection gets accepted to get the client's socket
xEventGroupWaitBits(params.tcp_connect_done, TCP_LISTENER_ACCEPTED, true, true, max_wait);
int sock = esp_transport_get_socket(params.transport_under_test);
for (int i=0; i<sock_options_len; ++i) {
int value = -1;
socklen_t optlen = (socklen_t)sizeof(value);
TEST_ASSERT_EQUAL(getsockopt(sock, expected_opts[i].level, expected_opts[i].optname,
(void*)&value, &optlen), 0);
if (expected_opts[i].opttype == SOCK_OPT_TYPE_BOOL) {
TEST_ASSERT_EQUAL((bool)value, (bool) expected_opts[i].optval);
} else if (expected_opts[i].opttype == SOCK_OPT_TYPE_INT) {
TEST_ASSERT_EQUAL(value, expected_opts[i].optval);
} else {
TEST_FAIL_MESSAGE("Unsupported socket option type");
}
}
close(sock); // close the tcp_transport's socket so we don't have to wait for connection timeout
xEventGroupSetBits(params.tcp_connect_done, TCP_ACCEPTOR_DONE);
xEventGroupWaitBits(params.tcp_connect_done, TCP_CONNECT_DONE, true, true, max_wait);
// Closing the listener and acceptor sockets
close(params.listen_sock);
close(params.accepted_sock);
xEventGroupWaitBits(params.tcp_connect_done, TCP_LISTENER_DONE, true, true, max_wait);
// Cleanup
TEST_ASSERT_EQUAL(false, params.tcp_listener_failed);
vEventGroupDelete(params.tcp_connect_done);
test_utils_task_delete(localhost_listener_task_handle);
test_utils_task_delete(tcp_connect_task_handle);
}
static void tcp_transport_keepalive_test(esp_transport_handle_t transport_under_test, esp_transport_keep_alive_t *config)
{
static struct expected_sock_option expected_opts[4] = {
{ .level = SOL_SOCKET, .optname = SO_KEEPALIVE, .optval = 1, .opttype = SOCK_OPT_TYPE_BOOL },
{ .level = IPPROTO_TCP },
{ .level = IPPROTO_TCP },
{ .level = IPPROTO_TCP }
};
expected_opts[1].optname = TCP_KEEPIDLE;
expected_opts[1].optval = config->keep_alive_idle;
expected_opts[2].optname = TCP_KEEPINTVL;
expected_opts[2].optval = config->keep_alive_interval;
expected_opts[3].optname = TCP_KEEPCNT;
expected_opts[3].optval = config->keep_alive_count;
socket_operation_test(transport_under_test, expected_opts, sizeof(expected_opts)/sizeof(struct expected_sock_option));
}
TEST_CASE("tcp_transport: Keep alive test", "[tcp_transport]")
{
// Init the transport under test
esp_transport_list_handle_t transport_list = esp_transport_list_init();
esp_transport_handle_t tcp = esp_transport_tcp_init();
esp_transport_list_add(transport_list, tcp, "tcp");
// Perform the test
esp_transport_keep_alive_t keep_alive_cfg = {
.keep_alive_interval = 5,
.keep_alive_idle = 4,
.keep_alive_enable = true,
.keep_alive_count = 3 };
esp_transport_tcp_set_keep_alive(tcp, &keep_alive_cfg);
// Bind device interface to loopback
TEST_TRANSPORT_BIND_IFNAME();
esp_transport_tcp_set_interface_name(tcp, &ifr);
tcp_transport_keepalive_test(tcp, &keep_alive_cfg);
// Cleanup
esp_transport_close(tcp);
esp_transport_list_destroy(transport_list);
}
TEST_CASE("ssl_transport: Keep alive test", "[tcp_transport]")
{
// Init the transport under test
esp_transport_list_handle_t transport_list = esp_transport_list_init();
esp_transport_handle_t ssl = esp_transport_ssl_init();
esp_transport_list_add(transport_list, ssl, "ssl");
esp_tls_init_global_ca_store();
esp_transport_ssl_enable_global_ca_store(ssl);
// Perform the test
esp_transport_keep_alive_t keep_alive_cfg = {
.keep_alive_interval = 2,
.keep_alive_idle = 3,
.keep_alive_enable = true,
.keep_alive_count = 4 };
esp_transport_ssl_set_keep_alive(ssl, &keep_alive_cfg);
// Bind device interface to loopback
TEST_TRANSPORT_BIND_IFNAME();
esp_transport_ssl_set_interface_name(ssl, &ifr);
tcp_transport_keepalive_test(ssl, &keep_alive_cfg);
// Cleanup
esp_transport_close(ssl);
esp_transport_list_destroy(transport_list);
}
TEST_CASE("ws_transport: Keep alive test", "[tcp_transport]")
{
// Init the transport under test
esp_transport_list_handle_t transport_list = esp_transport_list_init();
esp_transport_handle_t ssl = esp_transport_ssl_init();
esp_transport_list_add(transport_list, ssl, "ssl");
esp_tls_init_global_ca_store();
esp_transport_ssl_enable_global_ca_store(ssl);
esp_transport_handle_t ws = esp_transport_ws_init(ssl);
esp_transport_list_add(transport_list, ws, "wss");
// Perform the test
esp_transport_keep_alive_t keep_alive_cfg = {
.keep_alive_interval = 1,
.keep_alive_idle = 2,
.keep_alive_enable = true,
.keep_alive_count = 3 };
esp_transport_tcp_set_keep_alive(ssl, &keep_alive_cfg);
// Bind device interface to loopback
TEST_TRANSPORT_BIND_IFNAME();
esp_transport_ssl_set_interface_name(ssl, &ifr);
tcp_transport_keepalive_test(ws, &keep_alive_cfg);
// Cleanup
esp_transport_close(ssl);
esp_transport_list_destroy(transport_list);
}
// Note: This functionality is tested and kept only for compatibility reasons with IDF <= 4.x
// It is strongly encouraged to use transport within lists only
TEST_CASE("ssl_transport: Check that parameters (keepalive) are set independently on the list", "[tcp_transport]")
{
// Init the transport under test
esp_transport_handle_t ssl = esp_transport_ssl_init();
esp_tls_init_global_ca_store();
esp_transport_ssl_enable_global_ca_store(ssl);
// Perform the test
esp_transport_keep_alive_t keep_alive_cfg = {
.keep_alive_interval = 2,
.keep_alive_idle = 4,
.keep_alive_enable = true,
.keep_alive_count = 3 };
esp_transport_ssl_set_keep_alive(ssl, &keep_alive_cfg);
// Bind device interface to loopback
TEST_TRANSPORT_BIND_IFNAME();
esp_transport_ssl_set_interface_name(ssl, &ifr);
tcp_transport_keepalive_test(ssl, &keep_alive_cfg);
// Cleanup
esp_transport_close(ssl);
esp_transport_destroy(ssl);
}

View File

@ -0,0 +1,44 @@
#include "unity.h"
#include "esp_transport.h"
#include "esp_transport_tcp.h"
#include "esp_transport_ssl.h"
#include "esp_transport_ws.h"
#include "esp_log.h"
TEST_CASE("tcp_transport: init and deinit transport list", "[tcp_transport][leaks=0]")
{
esp_transport_list_handle_t transport_list = esp_transport_list_init();
esp_transport_handle_t tcp = esp_transport_tcp_init();
esp_transport_list_add(transport_list, tcp, "tcp");
TEST_ASSERT_EQUAL(ESP_OK, esp_transport_list_destroy(transport_list));
}
TEST_CASE("tcp_transport: using ssl transport separately", "[tcp_transport][leaks=0]")
{
esp_transport_handle_t h = esp_transport_ssl_init();
TEST_ASSERT_EQUAL(ESP_OK, esp_transport_destroy(h));
}
TEST_CASE("tcp_transport: using ws transport separately", "[tcp_transport][leaks=0]")
{
esp_transport_handle_t tcp = esp_transport_tcp_init();
esp_transport_handle_t ws = esp_transport_ws_init(tcp);
TEST_ASSERT_EQUAL(ESP_OK, esp_transport_destroy(ws));
TEST_ASSERT_EQUAL(ESP_OK, esp_transport_destroy(tcp));
}
TEST_CASE("transport: init and deinit multiple transport items", "[tcp_transport][leaks=0]")
{
esp_transport_list_handle_t transport_list = esp_transport_list_init();
esp_transport_handle_t tcp = esp_transport_tcp_init();
esp_transport_list_add(transport_list, tcp, "tcp");
esp_transport_handle_t ssl = esp_transport_ssl_init();
esp_transport_list_add(transport_list, ssl, "ssl");
esp_transport_handle_t ws = esp_transport_ws_init(tcp);
esp_transport_list_add(transport_list, ws, "ws");
esp_transport_handle_t wss = esp_transport_ws_init(ssl);
esp_transport_list_add(transport_list, wss, "wss");
TEST_ASSERT_EQUAL(ESP_OK, esp_transport_list_destroy(transport_list));
}

View File

@ -0,0 +1,178 @@
#include "unity.h"
#include "esp_transport.h"
#include "esp_transport_tcp.h"
#include "esp_transport_ssl.h"
#include "esp_transport_ws.h"
#include "esp_log.h"
#include "lwip/sockets.h"
#include "tcp_transport_fixtures.h"
#define TEST_TRANSPORT_BIND_IFNAME() \
struct ifreq ifr; \
ifr.ifr_name[0] = 'l'; \
ifr.ifr_name[1] = 'o'; \
ifr.ifr_name[2] = '\0';
static void tcp_transport_keepalive_test(esp_transport_handle_t transport_under_test, bool async, esp_transport_keep_alive_t *config)
{
static struct expected_sock_option expected_opts[4] = {
{ .level = SOL_SOCKET, .optname = SO_KEEPALIVE, .optval = 1, .opttype = SOCK_OPT_TYPE_BOOL },
{ .level = IPPROTO_TCP },
{ .level = IPPROTO_TCP },
{ .level = IPPROTO_TCP }
};
expected_opts[1].optname = TCP_KEEPIDLE;
expected_opts[1].optval = config->keep_alive_idle;
expected_opts[2].optname = TCP_KEEPINTVL;
expected_opts[2].optval = config->keep_alive_interval;
expected_opts[3].optname = TCP_KEEPCNT;
expected_opts[3].optval = config->keep_alive_count;
tcp_transport_test_socket_options(transport_under_test, async, expected_opts,
sizeof(expected_opts) / sizeof(struct expected_sock_option));
}
TEST_CASE("tcp_transport: connect timeout", "[tcp_transport]")
{
// Init the transport under test
esp_transport_list_handle_t transport_list = esp_transport_list_init();
esp_transport_handle_t tcp = esp_transport_tcp_init();
esp_transport_list_add(transport_list, tcp, "tcp");
tcp_transport_test_connection_timeout(tcp);
esp_transport_close(tcp);
esp_transport_list_destroy(transport_list);
}
TEST_CASE("ssl_transport: connect timeout", "[tcp_transport]")
{
// Init the transport under test
esp_transport_list_handle_t transport_list = esp_transport_list_init();
esp_transport_handle_t tcp = esp_transport_tcp_init();
esp_transport_list_add(transport_list, tcp, "tcp");
esp_transport_handle_t ssl = esp_transport_ssl_init();
esp_transport_list_add(transport_list, ssl, "ssl");
tcp_transport_test_connection_timeout(ssl);
esp_transport_close(tcp);
esp_transport_close(ssl);
esp_transport_list_destroy(transport_list);
}
TEST_CASE("tcp_transport: Keep alive test", "[tcp_transport]")
{
// Init the transport under test
esp_transport_list_handle_t transport_list = esp_transport_list_init();
esp_transport_handle_t tcp = esp_transport_tcp_init();
esp_transport_list_add(transport_list, tcp, "tcp");
// Perform the test
esp_transport_keep_alive_t keep_alive_cfg = {
.keep_alive_interval = 5,
.keep_alive_idle = 4,
.keep_alive_enable = true,
.keep_alive_count = 3 };
esp_transport_tcp_set_keep_alive(tcp, &keep_alive_cfg);
// Bind device interface to loopback
TEST_TRANSPORT_BIND_IFNAME();
esp_transport_tcp_set_interface_name(tcp, &ifr);
// Run the test for both sync and async_connect
tcp_transport_keepalive_test(tcp, true, &keep_alive_cfg);
tcp_transport_keepalive_test(tcp, false, &keep_alive_cfg);
// Cleanup
esp_transport_close(tcp);
esp_transport_list_destroy(transport_list);
}
TEST_CASE("ssl_transport: Keep alive test", "[tcp_transport]")
{
// Init the transport under test
esp_transport_list_handle_t transport_list = esp_transport_list_init();
esp_transport_handle_t ssl = esp_transport_ssl_init();
esp_transport_list_add(transport_list, ssl, "ssl");
esp_tls_init_global_ca_store();
esp_transport_ssl_enable_global_ca_store(ssl);
// Perform the test
esp_transport_keep_alive_t keep_alive_cfg = {
.keep_alive_interval = 2,
.keep_alive_idle = 3,
.keep_alive_enable = true,
.keep_alive_count = 4 };
esp_transport_ssl_set_keep_alive(ssl, &keep_alive_cfg);
// Bind device interface to loopback
TEST_TRANSPORT_BIND_IFNAME();
esp_transport_ssl_set_interface_name(ssl, &ifr);
// Run the test for async_connect only
// - TLS connection would connect on socket level only, returning tls-handshake in progress
tcp_transport_keepalive_test(ssl, true, &keep_alive_cfg);
// Cleanup
esp_transport_close(ssl);
esp_transport_list_destroy(transport_list);
}
TEST_CASE("ws_transport: Keep alive test", "[tcp_transport]")
{
// Init the transport under test
esp_transport_list_handle_t transport_list = esp_transport_list_init();
esp_transport_handle_t tcp = esp_transport_tcp_init();
esp_transport_list_add(transport_list, tcp, "tcp");
esp_transport_handle_t ws = esp_transport_ws_init(tcp);
esp_transport_list_add(transport_list, ws, "ws");
// Perform the test
esp_transport_keep_alive_t keep_alive_cfg = {
.keep_alive_interval = 11,
.keep_alive_idle = 22,
.keep_alive_enable = true,
.keep_alive_count = 33 };
esp_transport_tcp_set_keep_alive(tcp, &keep_alive_cfg);
// Bind device interface to loopback
TEST_TRANSPORT_BIND_IFNAME();
esp_transport_ssl_set_interface_name(tcp, &ifr);
// Run the test for sync_connect only (ws doesn't support async)
tcp_transport_keepalive_test(ws, false, &keep_alive_cfg);
// Cleanup
esp_transport_close(tcp);
esp_transport_list_destroy(transport_list);
}
// Note: This functionality is tested and kept only for compatibility reasons with IDF <= 4.x
// It is strongly encouraged to use transport within lists only
TEST_CASE("ssl_transport: Check that parameters (keepalive) are set independently on the list", "[tcp_transport]")
{
// Init the transport under test
esp_transport_handle_t ssl = esp_transport_ssl_init();
esp_tls_init_global_ca_store();
esp_transport_ssl_enable_global_ca_store(ssl);
// Perform the test
esp_transport_keep_alive_t keep_alive_cfg = {
.keep_alive_interval = 2,
.keep_alive_idle = 4,
.keep_alive_enable = true,
.keep_alive_count = 3 };
esp_transport_ssl_set_keep_alive(ssl, &keep_alive_cfg);
// Bind device interface to loopback
TEST_TRANSPORT_BIND_IFNAME();
esp_transport_ssl_set_interface_name(ssl, &ifr);
tcp_transport_keepalive_test(ssl, true, &keep_alive_cfg);
// Cleanup
esp_transport_close(ssl);
esp_transport_destroy(ssl);
}

View File

@ -0,0 +1,311 @@
#include "unity.h"
#include "esp_transport.h"
#include "esp_transport_tcp.h"
#include "test_utils.h"
#include "esp_log.h"
#include "lwip/err.h"
#include "lwip/sockets.h"
#include "freertos/event_groups.h"
#include "tcp_transport_fixtures.h"
// This is a private API of the tcp transport, but needed for socket operation tests
int esp_transport_get_socket(esp_transport_handle_t t);
/**
* @brief Event flags for synchronization between the listener task, the connection task and the test task
*/
enum {
TCP_CONNECT_DONE = 1 << 0, /*!< Indicates that the connection task has finished, so the transport_connect() exited */
TCP_LISTENER_DONE = 1 << 1, /*!< Indicates that the listener task has finished either with success for failure */
TCP_TEST_DONE = 1 << 2, /*!< Indicates that the test case finished, test tear-down() called */
TCP_LISTENER_READY = 1 << 3, /*!< Indicates that the listener task is ready to accept connections */
TCP_LISTENER_ACCEPTED = 1 << 4, /*!< Indicates that the listener task has accepted a connection (from transport_connect()) */
};
/**
* @brief Connection test configuration parameters
*/
struct tcp_connect_test_config {
esp_transport_handle_t transport_under_test;
bool accept_connection;
bool consume_sock_backlog;
bool connect_async;
int timeout_ms;
int port;
bool listener_task_prio_higher;
};
/**
* @brief Test setup structure containing all the info needed for the connection tests
*/
struct tcp_connect_test_storage {
struct tcp_connect_test_config config;
TickType_t max_wait;
EventGroupHandle_t tcp_connect_done;
int connect_return_value;
int listen_sock;
int accepted_sock;
int last_connect_sock;
bool tcp_listener_failed;
TaskHandle_t listener_task;
TaskHandle_t tcp_connect_task;
};
typedef struct tcp_connect_test_storage *tcp_connect_test_t;
/**
* @brief Recursively connects with a new socket to loopback interface until the last one blocks.
* The last socket is closed upon test teardown, that initiates recursive cleanup (close) for all
* active/connected sockets.
*/
static void connect_once(struct tcp_connect_test_storage *storage)
{
struct sockaddr_in dest_addr_ip4 = { .sin_addr.s_addr = htonl(INADDR_LOOPBACK),
.sin_family = AF_INET,
.sin_port = htons(storage->config.port) };
int connect_sock = socket(AF_INET, SOCK_STREAM, IPPROTO_IP);
if (connect_sock < 0) {
storage->tcp_listener_failed = true;
return;
}
storage->last_connect_sock = connect_sock;
int err = connect(connect_sock, (struct sockaddr *)&dest_addr_ip4, sizeof(dest_addr_ip4));
if (err != 0) {
// The last connection is expected to fail here, since the both sockets get closed on test cleanup
return;
}
connect_once(storage);
close(connect_sock);
}
/**
* @brief creates a listener (and an acceptor if configured)
*
* if consume_sock_backlog set: connect as many times as possible to prepare an endpoint which
* would make the client block but not complete TCP handshake
*
* if accept_connection set: waiting normally for connection creating an acceptor to mimic tcp-transport endpoint
*/
static void localhost_listener(void *pvParameters)
{
const char* TAG = "tcp_transport_test";
struct tcp_connect_test_storage *storage = pvParameters;
struct sockaddr_in dest_addr_ip4 = { .sin_addr.s_addr = htonl(INADDR_ANY),
.sin_family = AF_INET,
.sin_port = htons(storage->config.port) };
// Create listener socket and bind it to ANY address
storage->listen_sock = socket(AF_INET, SOCK_STREAM, IPPROTO_IP);
int opt = 1;
setsockopt(storage->listen_sock, SOL_SOCKET, SO_REUSEADDR, &opt, sizeof(opt));
if (storage->listen_sock < 0) {
ESP_LOGE(TAG, "Unable to create socket");
storage->tcp_listener_failed = true;
goto failed;
}
int err = bind(storage->listen_sock, (struct sockaddr *)&dest_addr_ip4, sizeof(dest_addr_ip4));
if (err != 0) {
ESP_LOGE(TAG, "Socket unable to bind: errno %d", errno);
storage->tcp_listener_failed = true;
goto failed;
}
// Listen with backlog set to a low number
err = listen(storage->listen_sock, 4);
if (err != 0) {
ESP_LOGE(TAG, "Error occurred during listen: errno %d", errno);
storage->tcp_listener_failed = true;
goto failed;
}
// Listener is ready at this point
xEventGroupSetBits(storage->tcp_connect_done, TCP_LISTENER_READY);
if (storage->config.consume_sock_backlog) {
// Ideally we would set backlog to 0, but since this is an implementation specific recommendation parameter,
// we recursively create sockets and try to connect to this listener in order to consume the backlog. After
// the backlog is consumed, the last connection blocks (waiting for accept), but at that point we are sure
// that any other connection would also block
connect_once(storage);
} else if (storage->config.accept_connection) {
struct sockaddr_storage source_addr;
socklen_t addr_len = sizeof(source_addr);
storage->accepted_sock = accept(storage->listen_sock, (struct sockaddr *)&source_addr, &addr_len);
if (storage->accepted_sock < 0) {
ESP_LOGE(TAG, "Unable to accept connection: errno %d", errno);
goto failed;
}
xEventGroupSetBits(storage->tcp_connect_done, TCP_LISTENER_ACCEPTED); // Mark the socket as accepted
// ...and wait for the "acceptor" tests to finish
xEventGroupWaitBits(storage->tcp_connect_done, TCP_TEST_DONE, true, true, storage->config.timeout_ms * 10);
}
failed:
xEventGroupSetBits(storage->tcp_connect_done, TCP_LISTENER_DONE);
vTaskSuspend(NULL);
}
/**
* @brief This task simply tries to connect to localhost (server provided by listner's task) using tcp_transport
*/
static void tcp_connect_task(void *pvParameters)
{
struct tcp_connect_test_storage *storage = pvParameters;
int (*connect_fn)(esp_transport_handle_t, const char *, int, int) =
storage->config.connect_async ? esp_transport_connect_async : esp_transport_connect;
storage->connect_return_value = connect_fn(storage->config.transport_under_test, "localhost", storage->config.port, storage->config.timeout_ms);
if (storage->config.accept_connection) {
// If we test the accepted connection, need to wait until the test completes
xEventGroupWaitBits(storage->tcp_connect_done, TCP_TEST_DONE, true, true, storage->config.timeout_ms * 10);
}
xEventGroupSetBits(storage->tcp_connect_done, TCP_CONNECT_DONE);
vTaskSuspend(NULL);
}
static inline void close_if_valid(int *s)
{
if (*s >= 0) {
close(*s);
*s = -1;
}
}
/**
* @brief Connect test setup function
*
* Creates the Test storage, configures it accordingly and starts two tasks
* * localhost_listener -- to provide a simple server endpoint for the transport layers to connect to
* * tcp_connect_task -- to perform the connection
*/
static tcp_connect_test_t connect_test_setup(struct tcp_connect_test_config *config)
{
tcp_connect_test_t t = calloc(1, sizeof(struct tcp_connect_test_storage));
if (!t) {
return NULL;
}
memcpy(&t->config, config, sizeof(struct tcp_connect_test_config));
t->tcp_connect_done = xEventGroupCreate();
if (!t->tcp_connect_done) {
return NULL;
}
t->max_wait = pdMS_TO_TICKS(config->timeout_ms * 10);
t->listen_sock = t->last_connect_sock = t->accepted_sock = -1; // mark all sockets invalid
test_case_uses_tcpip();
// Create listener task
xTaskCreatePinnedToCore(localhost_listener, "localhost_listener", 4096, t, 5, &t->listener_task, 0);
xEventGroupWaitBits(t->tcp_connect_done, TCP_LISTENER_READY, true, true, t->max_wait);
// Perform tcp-connect in a separate task to check asynchronously for the timeout or to connect (depends on the test config)
xTaskCreatePinnedToCore(tcp_connect_task, "tcp_connect_task", 4096, t,
config->listener_task_prio_higher? 4 : 6, &t->tcp_connect_task, 0);
return t;
}
/**
* @brief Destroys and cleans out the test environment
*/
static void connect_test_teardown(tcp_connect_test_t t)
{
// Mark the test done and wait for the listener to check if finished with no issues
xEventGroupSetBits(t->tcp_connect_done, TCP_TEST_DONE);
xEventGroupWaitBits(t->tcp_connect_done, TCP_LISTENER_DONE, true, true, t->max_wait);
TEST_ASSERT_EQUAL(false, t->tcp_listener_failed);
// Closing both parties of the last "blocking" connection to unwind localhost_listener() and let other connected sockets closed
close_if_valid(&t->listen_sock);
close_if_valid(&t->last_connect_sock);
close_if_valid(&t->accepted_sock);
// Cleanup
vTaskSuspend(t->tcp_connect_task);
vTaskSuspend(t->listener_task);
vEventGroupDelete(t->tcp_connect_done);
test_utils_task_delete(t->tcp_connect_task);
test_utils_task_delete(t->listener_task);
free(t);
}
/**
* @brief Utility function for testing timeouts for different transports
*/
void tcp_transport_test_connection_timeout(esp_transport_handle_t transport_under_test)
{
struct tcp_connect_test_config params = {
.timeout_ms = 200,
.port = 80,
.consume_sock_backlog = true,
.connect_async = false,
.transport_under_test = transport_under_test,
.listener_task_prio_higher = true
};
tcp_connect_test_t test = connect_test_setup(&params);
TEST_ASSERT_NOT_NULL(test);
// Roughly measure tick-time spent while trying to connect
TickType_t start = xTaskGetTickCount();
EventBits_t bits = xEventGroupWaitBits(test->tcp_connect_done, TCP_CONNECT_DONE, true, true, test->max_wait);
TickType_t end = xTaskGetTickCount();
TEST_ASSERT_EQUAL(TCP_CONNECT_DONE, TCP_CONNECT_DONE & bits); // Connection has finished
TEST_ASSERT_EQUAL(-1, test->connect_return_value); // Connection failed with -1
// Test connection attempt took expected timeout value
TEST_ASSERT_INT_WITHIN(pdMS_TO_TICKS(params.timeout_ms/5), pdMS_TO_TICKS(params.timeout_ms), end-start);
// Close the last bound connection, to recursively unwind the consumed backlog
close_if_valid(&test->last_connect_sock);
connect_test_teardown(test);
}
/**
* @brief Utility function for testing timeouts for different transports, options and both sync and async connection
*/
void tcp_transport_test_socket_options(esp_transport_handle_t transport_under_test, bool async,
const struct expected_sock_option *expected_opts, size_t sock_options_len)
{
struct tcp_connect_test_config params = {
.timeout_ms = 200,
.port = 80,
.accept_connection = true,
.consume_sock_backlog = false,
.transport_under_test = transport_under_test,
.connect_async = async,
.listener_task_prio_higher = false
};
tcp_connect_test_t test = connect_test_setup(&params);
TEST_ASSERT_NOT_NULL(test);
// Wait till the connection gets accepted to get the client's socket
xEventGroupWaitBits(test->tcp_connect_done, TCP_LISTENER_ACCEPTED, true, true, test->max_wait);
int sock = esp_transport_get_socket(params.transport_under_test);
for (int i=0; i<sock_options_len; ++i) {
int value = -1;
socklen_t optlen = (socklen_t)sizeof(value);
TEST_ASSERT_EQUAL(getsockopt(sock, expected_opts[i].level, expected_opts[i].optname,
(void*)&value, &optlen), 0);
if (expected_opts[i].opttype == SOCK_OPT_TYPE_BOOL) {
TEST_ASSERT_EQUAL((bool)value, (bool) expected_opts[i].optval);
} else if (expected_opts[i].opttype == SOCK_OPT_TYPE_INT) {
TEST_ASSERT_EQUAL(value, expected_opts[i].optval);
} else {
TEST_FAIL_MESSAGE("Unsupported socket option type");
}
}
// close the tcp_transport's socket so we don't have to wait for connection timeout
close(sock);
connect_test_teardown(test);
}

View File

@ -79,9 +79,15 @@ static int esp_tls_connect_async(esp_transport_handle_t t, const char *host, int
return -1;
}
ssl->conn_state = TRANS_SSL_CONNECTING;
ssl->sockfd = -1;
}
if (ssl->conn_state == TRANS_SSL_CONNECTING) {
return esp_tls_conn_new_async(host, strlen(host), port, &ssl->cfg, ssl->tls);
int progress = esp_tls_conn_new_async(host, strlen(host), port, &ssl->cfg, ssl->tls);
if (progress >= 0) {
ssl->sockfd = ssl->tls->sockfd;
}
return progress;
}
return 0;
}
@ -161,7 +167,7 @@ static int ssl_poll_read(esp_transport_handle_t t, int timeout_ms)
uint32_t optlen = sizeof(sock_errno);
getsockopt(ssl->sockfd, SOL_SOCKET, SO_ERROR, &sock_errno, &optlen);
esp_transport_capture_errno(t, sock_errno);
ESP_LOGE(TAG, "poll_read select error %d, errno = %s, fd = %d", sock_errno, strerror(sock_errno), ssl->tls->sockfd);
ESP_LOGE(TAG, "poll_read select error %d, errno = %s, fd = %d", sock_errno, strerror(sock_errno), ssl->sockfd);
ret = -1;
}
return ret;
@ -184,7 +190,7 @@ static int ssl_poll_write(esp_transport_handle_t t, int timeout_ms)
uint32_t optlen = sizeof(sock_errno);
getsockopt(ssl->sockfd, SOL_SOCKET, SO_ERROR, &sock_errno, &optlen);
esp_transport_capture_errno(t, sock_errno);
ESP_LOGE(TAG, "poll_write select error %d, errno = %s, fd = %d", sock_errno, strerror(sock_errno), ssl->tls->sockfd);
ESP_LOGE(TAG, "poll_write select error %d, errno = %s, fd = %d", sock_errno, strerror(sock_errno), ssl->sockfd);
ret = -1;
}
return ret;
@ -196,7 +202,7 @@ static int ssl_write(esp_transport_handle_t t, const char *buffer, int len, int
transport_esp_tls_t *ssl = ssl_get_context_data(t);
if ((poll = esp_transport_poll_write(t, timeout_ms)) <= 0) {
ESP_LOGW(TAG, "Poll timeout or error, errno=%s, fd=%d, timeout_ms=%d", strerror(errno), ssl->tls->sockfd, timeout_ms);
ESP_LOGW(TAG, "Poll timeout or error, errno=%s, fd=%d, timeout_ms=%d", strerror(errno), ssl->sockfd, timeout_ms);
return poll;
}
int ret = esp_tls_conn_write(ssl->tls, (const unsigned char *) buffer, len);
@ -213,7 +219,7 @@ static int tcp_write(esp_transport_handle_t t, const char *buffer, int len, int
transport_esp_tls_t *ssl = ssl_get_context_data(t);
if ((poll = esp_transport_poll_write(t, timeout_ms)) <= 0) {
ESP_LOGW(TAG, "Poll timeout or error, errno=%s, fd=%d, timeout_ms=%d", strerror(errno), ssl->tls->sockfd, timeout_ms);
ESP_LOGW(TAG, "Poll timeout or error, errno=%s, fd=%d, timeout_ms=%d", strerror(errno), ssl->sockfd, timeout_ms);
return poll;
}
int ret = send(ssl->sockfd,(const unsigned char *) buffer, len, 0);
@ -392,7 +398,7 @@ static int ssl_get_socket(esp_transport_handle_t t)
{
transport_esp_tls_t *ssl = ssl_get_context_data(t);
if (ssl && ssl->tls) {
return ssl->tls->sockfd;
return ssl->sockfd;
}
return -1;
}

View File

@ -602,7 +602,7 @@ esp_transport_handle_t esp_transport_ws_init(esp_transport_handle_t parent_handl
});
esp_transport_set_func(t, ws_connect, ws_read, ws_write, ws_close, ws_poll_read, ws_poll_write, ws_destroy);
// webocket underlying transfer is the payload transfer handle
// websocket underlying transfer is the payload transfer handle
esp_transport_set_parent_transport_func(t, ws_get_payload_transport_handle);
esp_transport_set_context_data(t, ws);