The most awesome part:

> This bug was independently discovered by multiple research teams within Google, including the silifuzz team and Google Information Security Engineering.

It may be worth this submission being merged into https://news.ycombinator.com/item?id=38266773

Yup. Done. Thanks!

Thanks, this is a way more detailed and informative explanation!

    ...our validation pipeline produced an interesting assertion...

What is a validation pipeline?

It's described one paragraph earlier.

> I’ve written previously about a processor validation technique called Oracle Serialization that we’ve been using. The idea is to generate two forms of the same randomly generated program and verify their final state is identical.

Sounds like the real story should be that Google solved the halting problem. :-P

You're free to solve the halting problem for restricted sets of programs, that doesn't break any rules of the universe.

They also could be just discarding any where it runs for longer than X time, or a bunch of other possibilities.

They might be generating programs that they know will halt. Like: applications with finite loops and such. There are not enough details.

The blog has a link to https://lock.cmpxchg8b.com/zenbleed.html#discovery which presents the concept.

have you never heard of continuous integration?

Rather, it looks to be:

https://en.wikipedia.org/wiki/Instruction_pipelining

From the Zenbleed link in sibling:

> I found this bug by fuzzing, big surprise [..] In fact, vendors fuzz their own products extensively - the industry term for it is Post-Silicon Validation.

That is, "a pipeline to validate the pipeline".

This was a lot more fun than the Google puff piece.