1. That is an age old question but the answer is pretty simple.... As part of your test planning you need to have figured out what your test completion criteria are such as "all Functional System Tests executed and passed with no errors" or "all Functional System Tests executed with no showstopper or high priority bugs in the system".
The criteria must be measurable, acheivable and realistic above all else, they must also be agreed with the project manager etc. so they can agree on what to expect in advance.
Once you have criteria to base your actual testing against, it gives you the ability to prioritise what you will do first and assist in tracking your progress to meet the criteria whilst you execute your tests.
Just keep testing until you meet the completion criteria or you are told to stop due to time constraints
2. If you've got decent tests, you should know what to expect from them in terms of actual results. Once a test passes, note it down and update your metrics to compare against progress with your completion criteria.