3 ms·
But RDRAND is a stupid backdoor RDRAND is a very good backdoor for code written by stupid people. There's a lot of application code which detects the presence
by cperciva 13y ago
But RDRAND is a stupid backdoor
RDRAND is a very good backdoor for code written by stupid people. There's a lot of application code which detects the presence of RDRAND and uses it directly instead of asking the OS for entropy.
(FWIW, I doubt RDRAND is backdoored in the traditional sense. I'm absolutely certain, however, that the NSA will take full advantage of the documented not-enough-entropy failure mode which returns all zeroes -- which most RDRAND-using software does not correctly handle.)
- tptacek 13y agoWhat application code are you thinking about here? Some examples would be great.
- pbsd 13y agoI've tried to diagnose when the RDRAND failures happen. From my experiments on Haswell, it only seems to happen when you have > N threads, where N is the number of physical cores, simultaneously reading continuously from RDRAND. But when that happens, nearly every read from every thread will fail. This means a user process could easily deplete RDRAND's contribution to the kernel's entropy pool. So encouraging all kinds of application code to use RDRAND is really bad advice. It should at least be a privileged instruction.