Same problem here: we use surefire + junit to run integration tests (maybe there is a better way, I don't know...) and our test suite takes hours to complete right now, yet most of the time we expect 100% correct execution and would like to know - as soon as possible - which test failed and why.
I half expect I'm using the wrong tool or using the tool wrong and one or more simple solutions are available. That said, I'm a bit puzzled that this issue remains open for three years: it seems to me enough to add a switch (e.g. -DfastFail=true, default=false) and wrap up test execution as soon as the first test fails, producing the usual test run output. Test methods which haven't been executed might have to be removed from internal data structures so that they do not appear in the reports.