We know that the Grover algorithm outputs a marked item. Now we want to know the locations of all items. I can't find any paper to solve this problem.
Grover's search returns a uniform superposition of all marked items. So, yes, in your last step, you measure it and find a random sample out of that set. If you want others, just repeat and you'll get another random sample.
If you want to be a bit more directed, you can explicitly exclude any items you've previously found by unmarking them in your oracle step.