process core dumps when they exist, after test runs

[prakashsurya]

It's possible and sometimes expected for a system to panic while running ztest or the zfstest suite. When this occurs, we should automatically process any core dumps when gathering the test results, after the test has finished. This would save us the trouble of having to perform this analysis manually, and also potentially losing this state if the VM containing the core dump is destroyed prior to being inspected.

As an example, see this comment: openzfs/openzfs#219 (comment)