Gregory's Server@SwiftOnSecurity here's an interesting one. We are working on a factory test for #tinytapeout . Many of the stateful designs don't have a reset (unlike all industry chips). I was finding that post reset or power cycle the test results were changing - the flops inside were retaining state. It took 10 minutes for the flops to lose state!
@matthewvenn @SwiftOnSecurity Plenty of industry chips don’t simply reset every flop when you assert the reset line. It’s commonplace that SRAMs and I/Dcaches don’t bother to wipe their contents across reset, and even common they don’t reset the “valid” bits, which introduces some interesting vulnerabilities. And similarly weird bugs to diagnose when someone assumed they reset.