[platform/reboot] Fix the reboot stuck issue

[wangxin]

Description of PR

Summary:
Fixes # (issue)
The test_reboot.py script may stuck like for ever while rebooting the SONiC switch.

Reason:
Some switch may need more time to go down after the reboot
command is issued. If switch does not go down in time, terminating the
multiprocessing.Process could stuck. The terminate() call sends SIGTERM
which is not able to terminate the asynchronous process.

The fix:

1. Increase the timeout to wait for switch to go down.
2. Replace multiprocessing.Process with multiprocessing.pool.ThreadPool.
3. Try to kill any reboot process on the switch if switch does not go down
   in time.
4. On switch that reboot fast, the wait_for module may fail to detect
   that the switch was down for rebooting. Ignore the error of waiting
   for switch to go down.
5. Use uptime to verify whether switch has rebooted.

Type of change

• Bug fix
• [] Testbed and Framework(new/improvement)
• [] Test case(new/improvement)

Approach

How did you do it?

1. Increase the timeout to wait for switch to go down.
2. Replace multiprocessing.Process with multiprocessing.pool.ThreadPool.
3. Try to kill any reboot process on the switch if switch does not go down
   in time.
4. On switch that reboot fast, the wait_for module may fail to detect
   that the switch was down for rebooting. Ignore the error of waiting
   for switch to go down.
5. Use uptime to verify whether switch has rebooted.

How did you verify/test it?

Tested on Mellanox platform.

Any platform specific information?

No.

Supported testbed topology if it's a new test case?

NA

Documentation

[wangxin]

This PR may conflict with #1079. Close it for now. I will create a new one based on PR#1079 after it is merged.