20:43:12stassatshow do we know that an interrupt's context is stored on the stack?
20:45:45stassatspinning the registers explicitly fixes the problem
20:48:56stassatsok, now is it specific to darwin, darwin rosetta? i think the assumption that the registers are on the stack is not guaranteed, yet every other OS works ok
21:16:05stassatsvmmap says Rosetta Thread Context, that's where the context seems to be
22:52:09stassatsit just allows me to look at x86-64 things without going to the linux machine
22:52:15stassatsnot really using it for anything serious
22:53:24stassatsand it breaks some implicit assumptions, like sigcontext not allocated on the stack, so that's interesting, even if not applicable to other oses at the moment (but maybe in the future?)
22:56:50stassatsit's always -31, where could that come from