Add ESP32 Wi-Fi driver for Renesas GR-LYCHEE

[toyowata]

[x] I confirm this contribution is my own and I agree to license it with Apache 2.0.
[x] I confirm the moderators may change the PR before merging it in.

Summary of changes

I added the ESP32 Wi-Fi driver which accidentally missed in the PR #96

[iot-pdmc]

Thanks for your changes.
@ARMmbed/pdmc-integrators please review.

[teetak01]

Please provide new E2E test-results using the WIFI_ESP32. To me it seems that the target is not using this driver (considering that it compiles even without the driver).

Should the esp32.provide-default be set to true for this target to enable the driver?

[JanneKiiskila]

Should the esp32.provide-default be set to true for this target to enable the driver?

Yes it should. GR_LYCHEE does not seem to have Ethernet at all, it only has ESP32.

[toyowata]

@teetak01 @JanneKiiskila
The GR-LYCHEE board has on-board ESP32 module and the esp32.provide-default is defined here: https://github.com/ARMmbed/esp32-driver/blob/d6bd83f8710b4c475b3cce101d8b2c77e17feaf6/mbed_lib.json#L49

E2E test-results (same as here)

============================== 4 passed in 39.15s ==============================
2020-01-15 10:58:02,240:tests.conftest:MainThread:INFO: -----  TEST RESULTS SUMMARY  -----
2020-01-15 10:58:02,240:tests.conftest:MainThread:INFO: [ check the complete fail reasons and code locations from this log or html report ]
2020-01-15 10:58:02,241:tests.conftest:MainThread:INFO: [passed] - dev-client-tests.py::test_01_get_device_id - (0.772s)
2020-01-15 10:58:02,241:tests.conftest:MainThread:INFO: [passed] - dev-client-tests.py::test_02_put_resource - (4.831s)
2020-01-15 10:58:02,242:tests.conftest:MainThread:INFO: [passed] - dev-client-tests.py::test_03_get_resource - (1.839s)
2020-01-15 10:58:02,242:tests.conftest:MainThread:INFO: [passed] - dev-client-tests.py::test_04_subscribe_resource - (3.470s)

[teetak01]

I really do not like that practice. ISM43362 is also silently enabling it for some boards. It would be better that application would always have to enable the driver if it actually wants to use it.

Well, this is one reason we use the COMPONENT formalism as these would cause suprise-issues if some board is used with different network-stack than some driver is expecting.

[teetak01]

build-only

[iot-pdmc]

Test run: SUCCESS

---

Build artifacts

posted-by-tag: https://jenkins-client.isgtesting.com/job/mbed-os-example-pelion/