On 31/10/2018 15:30, Michael Walle wrote: >> >> Just one thing, I think an atomic test-and-clear is a better API. I >> understand that what you have works, because >> get_irq_latched+clear_irq_latch don't have any intervening access to the >> qtest socket. > > Ok thanks, like get_and_clear_irq_latched()? Do you think of we still > need the get_irq_latched() to get the state without actually clearing > it? I mean my simple test case won't need it, but there might be some > cases where you may want peek at the state.
Maybe - it's easy enough to add it, so if your test doesn't need it you can certainly leave it out. Paolo