Hey, I’m not really into programming, so I can't help much with changing the code. But if you wanna brute-force those 24 BIP39 words, check out BTCRecover. It's getting updates and supports GPUs.
Modifying Code for 24-Word Mnemonics
3 replies 447 views
To handle longer mnemonics, you gotta bump up the block size in HMAC from 128 bits to 256 bits. A 24-word mnemonic is over 128 bytes, so the current setup won't cut it. You’ll need to change a lot in the code, especially in those int_to_address.cl and just_seed.cl files since they’re set for 12 words.
I saw what mcdouglasx said and took it as a challenge. So, check this out: I made the ipad_key and opad_key arrays 128 bytes for HMAC. When the mnemonic is longer than 128 bytes (like with 24 words), the extra key part isn’t handled right. We need a key normalization process for HMAC. If the key is bigger than 128 bytes for SHA-512, it should be hashed down, then padded to 128 bytes if it’s short.
+1 to the key normalization idea. Sounds smart. But this seems like a huge code overhaul. Is it even worth it? Just curious, are you sure about the hashing bit? Seems like a lot of work for just a mnemonic.
Related topics
- Understanding the Differences Between Traditional and Simplified Chinese Mnemonics 6
- Issues with ripemd160 on Ubuntu 22 9
- New Bitcoin Improvement Proposal with $100 Reward 9
- Clipboard Vulnerabilities in Cryptocurrency Transactions 8
- Running Bitcoin Core on a Laptop with Limited Storage 20
- Exploring Blockstream's Satellite Tech and Its Potential 22