The aptly-named [chip-red-pill] workforce is featuring you a likelihood to go down the Intel rabbit gap. If you figured out how to build CPUs back in the 1970s, you would study that your instruction decoder would, for example, observe a sign up to sign up shift and then light-weight up one sign up to compose to a typical bus and a further sign-up to examine from the popular bus. These times, it is not that simple. In addition to compiling to an underlying instruction set, processors almost never encode guidance in components any more. Rather, every instruction has microcode that causes the proper things to happen at the appropriate time. But Intel encrypts their microcode. Of system, what can be encrypted can also be decrypted.
Employing vulnerabilities, you can activate an undocumented debugging method named red unlock. This lets a microcode dump and the decryption keys are inside. The staff did a paper for OffensiveCon22 on this strategy and you can see a movie about it, down below.
So far, the keys for Gemini Lake and Apollo Lake processors are obtainable. That covers rather a number of processors. Of training course, there are a lot of much more processors out there if you want to test your hand at a very similar exploit.
This exact workforce has carried out other exploits, these as executing arbitrary microcode inside of an Atom CPU. If you want to play along, you may well discover this beneficial. You do know that your CPU has guidance it is retaining from you, really do not you?