"This all comes after CNN cited unnamed sources in reporting that while testing a supposedly fixed version of the MCAS software in a simulator, pilots found that a microprocessor in the anti-stall system would lock up. As a result, the plane would pitch down and, for several seconds, the pilots would struggle to regain control. Presumably the new software was run on actual MCAS hardware within the simulator, triggering the hardware freeze.
To us, it sounds as though code in the MCAS update either forces the processor into a locked state, such as a tight unbreakable and uninterruptable infinite loop, or triggers an exception that can’t be handled and the CPU halts. It is remotely possible the code encounters a design flaw in the unidentified microprocessor that causes the circuitry to freeze."
My take is that the simulator runs the same software on the same hardware as the actual plane, which would make sense. And the software fix probably fixed the problem it was supposed to fix, but somehow borked something else.1)
This still leaves the systems that provide the simulated input for the simulator as a potential source for problems, though.
Sounds like debug hell.
1) There is a German expression for that, verschlimmbessern.
