[verification] Ensure each test cleans up derived contexts

[jhand2]

In the verification tests, if a test derives contexts that don't get destroyed, the target DPE will run out of contexts if it doesn't get reset. This is especially a problem for real hardware targets since they take too long to reset in between tests.

I would recommend some common setup code that derives a random context for each test to use that can be destroyed at the end of each test.

For example, if a target has an auto-initialized default context:

• DeriveChild(retain-parent = true). This makes sure the test can hold onto the initial context to restore it at the end. So now we have two handles, initialParent and rootChild.
• The test can derive all necessary contexts from rootChild
• And the end of the test, rootChild can be destroyed
• Can use RotateContextHandle make make initialParent the default context again.

[hpya93]

@jhand2, I have a similar logic in TagTCI testcase, can I move the logic to a new function in initializeContext.go?