[TE1]: Provide test interface in chip-shell to test CRMP.

[yufengwangca]

Problem

This is requested by up-coming TE 1, after discussing with TSG, we decided use chip-shell/chip-echo-responder as the first setup to test CRMP protocol. But chip-shell does not have command to test CRMP right now.

Summary of Changes

Add ping command to chip-shell for secure channel tests

Usage:
yufengw@yufengw-intel:~/Workarea/connectedhomeip/out/debug$ ./chip-shell
Init CHIP Stack
Init Thread stack
CHIP:IN: TransportMgr initialized
CHIP:IN: TransportMgr initialized

ping -h
Usage: ping [options]

Options:
-h print help information
-u use UDP (default)
-t use TCP
-i ping interval time in seconds
-c stop after replies
-r <1|0> enalbe/disable CRMP
Done

Fixes #5120

[todo]

temprary create a SecureSessionHandle from node id to unblock end-to-end test. Complete solution is tracked in PR:4451

connectedhomeip/examples/shell/shell_common/cmd_ping.cpp

Lines 267 to 277 in f006d05

	// TODO: temprary create a SecureSessionHandle from node id to unblock end-to-end test. Complete solution is tracked in PR:4451
	err = gEchoClient.Init(&gExchangeManager, { kTestDeviceNodeId, 0, gAdminId });
	SuccessOrExit(err);
	// Arrange to get a callback whenever an Echo Response is received.
	gEchoClient.SetEchoResponseReceived(HandleEchoResponseReceived);
	// Connection has been established. Now send the EchoRequests.
	for (unsigned int i = 0; i < gMaxEchoCount; i++)
	{
	err = SendEchoRequest(stream);

---

This comment was generated by todo based on a TODO comment in f006d05 in #5228. cc @yufengwangca.

[andy31415]

Do we have a separation of what each utility is intended to do?

We have: chip-tool, mobile apps (iOS and Android), chip-shell and python (chip-device-ctrl and chip-repl).

My understanding is that python is intended to superseed the C++ chip-tool (so that it is scriptable). What is the intent with chip-shell? Can we standardise on python instead?

[yufengwangca]

Do we have a separation of what each utility is intended to do?

We have: chip-tool, mobile apps (iOS and Android), chip-shell and python (chip-device-ctrl and chip-repl).

My understanding is that python is intended to superseed the C++ chip-tool (so that it is scriptable). What is the intent with chip-shell? Can we standardise on python instead?

I have a lot of discussion with TSG regarding the test setup for TE1.

On the client side, we have chip-tool, chip-shell, python-controller, chip-tool is designed only for testing cluster commands, we need to adjust its architecture a lot for CRMP test, and we also have decided to phase it out with python controller, so this tool is not an option.

python-controller is our target, but this tools is not stable right now and it is also not integrated with messaging layer because of #4451. So TSG decided to use python-controller in TE2 after we solve all the remaining issues.

On the server side, we decided to use lighting-app as the accessory for CRMP test. #5236. TSG could use chip-echo-responder temporarily before lighting-app echo support is ready.