@q I checked their other intermediates; they are 115, 155, 56, 66, 156, 237, and 230.
I think they used a real RNG but only requested one byte
@q I checked their other intermediates; they are 115, 155, 56, 66, 156, 237, and 230.
I think they used a real RNG but only requested one byte
When all their PKs are < 256 I wonder if this could be a bug in some off-the-shelf software. Sure, maybe they rolled their own crypto, but also, maybe they used buggy software. Maybe this affects other intermediates or certs from other CAs?
I mean who looks at their PK material to check that it's not filled with 0s?
In the mozilla report they claim it came out of an HSM. It turns out doing the thing in magical space woo woo hardware is no proof that the implementation is good.