3 ms·
Any idea why the keygen failed so badly? Debian's openssl patches generated predictable private keys for a while but they wouldn't be a single digit. It's Wired
by nailer 8y ago
Any idea why the keygen failed so badly? Debian's openssl patches generated predictable private keys for a while but they wouldn't be a single digit. It's Wired so I don't expect technical details but some info woukld be nice.
- tyingq 8y agoThe idea mentioned in the article of either unintentional or intentionally badly coded wallets seems likely.
- nailer 8y agoYeah but I'd like to know how the unintentional mistake was made. Eg - Debian OpenSSL was a Debian dev trying to prevent use of uninitialised memory (because Valgrind complained), without realising that uninitialised memory is used for randomness. - Years ago a bunch of PHP stuff (I forget what) was seeded with the string value of a randomness function, rather than it's output. What happened here?
- cesarb 8y ago> without realising that uninitialised memory is used for randomness It was more subtle than that: there were two identical calls, one of which added uninitialized memory, while the other added real entropy. The developer mistakenly removed both, thinking both were equally useless, instead of only removing the one with undefined behavior. To make things worse, another call added the current PID, so the results weren't identical every time. See more detail at https://research.swtch.com/openssl https://research.swtch.com/openssl
- tyingq 8y agoThere's lots of third party wallets, so the answer is probably all that and more. Broken KDFs, truncating things, etc.