On 06.03.2023 17:29, Juergen Gross wrote: > tools/tests/vhpet tests don't build since ages (at least since 4.10) > and they can't be activated from outside of tools/tests/vhpet.
This isn't exactly true - see the run-tests-% goal in the top level Makefile. > Remove them as they seem to be irrelevant. > > Signed-off-by: Juergen Gross <[email protected]> > --- > Andrew seems to remember that Roger wanted to keep those tests, but > this information might be stale or based on false assumptions. > So this patch should only go in with Roger's Ack. Having tried a little I can see that it's going to be quite a bit of work to get the thing to at least build again. In fact it looks as if it, having been introduced in 4.5, may already not have built successfully anymore by the time 4.5 was released: hpet_init() takes a struct domain *, but is passed &vcpu0 (matching what I can see in 4.4); I've noticed this while trying to figure out from where it got a declaration of the function in the first place. Nevertheless the test once served a purpose, so it may be worth making it work again. So yes, to decide which direction to move we will want Roger's opinion. Whether just building it by default is useful I'm not entirely certain. What I'd consider more useful is if tests like this were _run_ in the course of routine automated testing. Jan
