Bitcoin private keys & crypto puzzles

08 Jul 2024 03:17

ok ok.... i have found some realization

Bitcoin private keys & crypto puzzles

08 Jul 2024 03:00

Anybody knows the answer? Cmon guys (girls)....

Bitcoin private keys & crypto puzzles

08 Jul 2024 02:45

Use that PhD and help find that openssl flaw in the version Satoshi used back in 2012

Bitcoin private keys & crypto puzzles

08 Jul 2024 02:22

and nowhere to simply read this. Jeez. Why they dont simply written this in.. i dont know... pseudo code or whatever.....

Bitcoin private keys & crypto puzzles

08 Jul 2024 02:21

It feels soooo simple having that code above

Bitcoin private keys & crypto puzzles

08 Jul 2024 02:14

what you need to do which that?

Bitcoin private keys & crypto puzzles

08 Jul 2024 02:04

but nope - dont work

Bitcoin private keys & crypto puzzles

08 Jul 2024 02:04

I thought 192 == 128 + 64 and so the new lengths would be 256 + 64

Bitcoin private keys & crypto puzzles

08 Jul 2024 02:03

i almost sure, the "adaptation" simply requires to extend the ipad \ opad to 3 128 bytes chunks, fill the ipad and opad with mnemonic bytes and fill the left with "36" "5C" and place the "salt" in the third 128 bytes chunk.

Bitcoin private keys & crypto puzzles

08 Jul 2024 01:30

better than nothing but not really helpful

Bitcoin private keys & crypto puzzles

07 Jul 2024 21:54

I’m gonna honestly stop my laziness and write something in cuda

Bitcoin private keys & crypto puzzles

07 Jul 2024 21:53

compiller try itself to optimize the code

Bitcoin private keys & crypto puzzles

07 Jul 2024 21:52

Gave up and started accumulating hate for cuda

Bitcoin private keys & crypto puzzles

07 Jul 2024 21:51

But each core will be doing the whole series of operations sequentially again and again

Bitcoin private keys & crypto puzzles

07 Jul 2024 21:50

but even if i will understand this - would be actualy hard to manage all those bitwise hell into nice pipes... even sha256 looks like a hell to me, hmac-sha512 is some imbossible meat-sloter

Bitcoin private keys & crypto puzzles

08 Jul 2024 03:05

Why this is not the answer ??? WHYYYYYYYYYYYYY!!!!!!!!!!!!!!!!!!


uint8_t mnemonic[220] = { 0 };
uint32_t ipad[(3*128) / 4];
uint32_t opad[(3*128) / 4];
uint32_t seed[64 / 4] = { 0 };

for (int x = 0; x < (2*120) / 8; x++)
*(uint64_t*)((uint64_t*)ipad + x) = 0x3636363636363636ULL ^ SWAP512(*(uint64_t*)((uint64_t*)mnemonic + x));

for (int x = 0; x < (2*120) / 8; x++)
*(uint64_t*)((uint64_t*)opad + x) = 0x5C5C5C5C5C5C5C5CULL ^ SWAP512(*(uint64_t*)((uint64_t*)mnemonic + x));

for (int x = (2*120) / 4; x < (2*128) / 4; x++)
ipad[x] = 0x36363636;

for (int x = (2*120) / 4; x < (2*128) / 4; x++)
opad[x] = 0x5C5C5C5C;

for (int x = 0; x < 16 / 4; x++)
ipad[x + (2*128) / 4] = *(uint32_t*)((uint32_t*)&salt_swap + x);

sha512_swap((uint64_t*)ipad, (2*128+16-4), (uint64_t*)&opad[(2*128) / 4]);
sha512_swap((uint64_t*)opad, (2*128 + 64), (uint64_t*)&ipad[(2*128) / 4]);

for (int x = 0; x < 64 / 4; x++)
seed[x] = ipad[(2*128) / 4 + x];

for (int x = 1; x < 2048; x++){
sha512_swap((uint64_t*)ipad, (2*128 + 64), (uint64_t*)&opad[(2*128) / 4]);
sha512_swap((uint64_t*)opad, (2*128 + 64), (uint64_t*)&ipad[(2*128) / 4]);

for (int x = 0; x < 64 / 4; x++)
seed[x] = seed[x] ^ ipad[(2*128) / 4 + x];
}

Bitcoin private keys & crypto puzzles

08 Jul 2024 02:53

god damn..... how to get the seed for 24 words

Bitcoin private keys & crypto puzzles

08 Jul 2024 02:22

all wallets\clients simply uses openssl

Bitcoin private keys & crypto puzzles

08 Jul 2024 02:22

but same time .... nothing works right

Bitcoin private keys & crypto puzzles

08 Jul 2024 02:21

yeah i see, very interesting question but I never tested things with mnemonics and keys

Bitcoin private keys & crypto puzzles

08 Jul 2024 02:06

(this is real working code that suits me - but it is written for 12 words mnemonic and nope - it wont work for 24 words, the 24 words occupy much more than 120 bytes - actually up to 215 bytes)

Bitcoin private keys & crypto puzzles

08 Jul 2024 02:04

while 140.... well..... 128 + 16 - 4 and the adapted would be 256 + 16 -4

Bitcoin private keys & crypto puzzles

08 Jul 2024 02:03

but i am not sure what to do with those sha_512 swaps further

Bitcoin private keys & crypto puzzles

08 Jul 2024 02:02

uint8_t mnemonic[220] = { 0 };
uint32_t ipad[256 / 4];
uint32_t opad[256 / 4];
uint32_t seed[64 / 4] = { 0 };

for (int x = 0; x < 120 / 8; x++)
*(uint64_t*)((uint64_t*)ipad + x) = 0x3636363636363636ULL ^ SWAP512(*(uint64_t*)((uint64_t*)mnemonic + x));

for (int x = 0; x < 120 / 8; x++)
*(uint64_t*)((uint64_t*)opad + x) = 0x5C5C5C5C5C5C5C5CULL ^ SWAP512(*(uint64_t*)((uint64_t*)mnemonic + x));

for (int x = 120 / 4; x < 128 / 4; x++)
ipad[x] = 0x36363636;

for (int x = 120 / 4; x < 128 / 4; x++)
opad[x] = 0x5C5C5C5C;

for (int x = 0; x < 16 / 4; x++)
ipad[x + 128 / 4] = *(uint32_t*)((uint32_t*)&salt_swap + x);

sha512_swap((uint64_t*)ipad, 140, (uint64_t*)&opad[128 / 4]);
sha512_swap((uint64_t*)opad, 192, (uint64_t*)&ipad[128 / 4]);

for (int x = 0; x < 64 / 4; x++)
seed[x] = ipad[128 / 4 + x];

for (int x = 1; x < 2048; x++){
sha512_swap((uint64_t*)ipad, 192, (uint64_t*)&opad[128 / 4]);
sha512_swap((uint64_t*)opad, 192, (uint64_t*)&ipad[128 / 4]);

for (int x = 0; x < 64 / 4; x++)
seed[x] = seed[x] ^ ipad[128 / 4 + x];
}

This is the code for HMAC-SHA512 (PBKDF2) that takes 12 words mnemonic and turns it to the "seed" that i used further as master key in BIP39 keys derivation.

The SIMPLE question: "How to adapt it to at least 24 words? (better adapt to any mnemonic length)

Damn. I tried to find some real implementations and have not found anything. The only real thing i have found was the openssl implementation, but it is not straightforward for me to easily grab it.

Bitcoin private keys & crypto puzzles

08 Jul 2024 01:25

Hey.... are you pretty?

(can you help me understand the hmac-sha512 ?)

Bitcoin private keys & crypto puzzles

07 Jul 2024 21:54

so some of this thing could have been already made by compiller

Bitcoin private keys & crypto puzzles

07 Jul 2024 21:52

aaand yet its possible that something like that is already done by the compiller

Bitcoin private keys & crypto puzzles

07 Jul 2024 21:52

I tried with ripemd160 for few months

Bitcoin private keys & crypto puzzles

07 Jul 2024 21:51

I think same cores could run a set series of operations

Bitcoin private keys & crypto puzzles

07 Jul 2024 21:49

but what if same cores could run {operationA(thread1 | thread2 | thread3) operationB(thread4 | thread5 | thread6) operationC(thread7|thread8|thread9)) etc....