Increase conn. timeout in the timeout recovery tests

Due to increased number of the NM unit tests and, thus, increased load
on the system timeout recovery tests can sometimes fail, in particular
on FreeBSD. This commit fixes that. Besides, it seems that use of
T_SOFT here was unintentional to begin with.

tests/isc/doh_test.c

@@ -110,6 +110,7 @@ static isc_nm_t **nm = NULL;
 
 /* Timeout for soft-timeout tests (0.05 seconds) */
 #define T_SOFT	  50
+#define T_CONNECT 30 * 1000
 
 #define NSENDS	100
 #define NWRITES 10
@@ -697,7 +698,8 @@ doh_timeout_recovery(void *arg ISC_ATTR_UNUSED) {
 			ISC_NM_HTTP_DEFAULT_PATH);
 	isc_nm_httpconnect(connect_nm, NULL, &tcp_listen_addr, req_url,
 			   atomic_load(&POST), timeout_request_cb, NULL, ctx,
-			   client_sess_cache, T_SOFT, get_proxy_type(), NULL);
+			   client_sess_cache, T_CONNECT, get_proxy_type(),
+			   NULL);
 }
 
 static int
@@ -771,7 +773,7 @@ doh_connect_thread(void *arg) {
 	 */
 	int_fast64_t active = atomic_fetch_add(&active_cconnects, 1);
 	if (active > workers) {
-		goto next;
+		return;
 	}
 	connect_send_request(connect_nm, req_url, atomic_load(&POST),
 			     &(isc_region_t){ .base = (uint8_t *)send_msg.base,
@@ -782,8 +784,6 @@ doh_connect_thread(void *arg) {
 	if (sends <= 0) {
 		isc_loopmgr_shutdown(loopmgr);
 	}
-
-next: {}
 }
 
 static void

tests/isc/netmgr_common.c

@@ -978,7 +978,7 @@ stream_timeout_recovery(void **state ISC_ATTR_UNUSED) {
 	 */
 	isc_nm_settimeouts(connect_nm, T_SOFT, T_SOFT, T_SOFT, T_SOFT);
 	connect_readcb = timeout_retry_cb;
-	stream_connect(connect_connect_cb, NULL, T_SOFT);
+	stream_connect(connect_connect_cb, NULL, T_CONNECT);
 }
 
 int