chacha20: returning rand_core feature #333
Conversation
|
Alright, I think I've added just about all I can think of. The There are also some functions in |
|
I'd suggest less unsafe code and more of a KISS implementation in an initial PR, and then circling back on some of it in a followup so it can be considered separately. |
|
FYI, #338 bumps the crate to v0.10.0-pre, but also includes some other breaking API changes. Rebasing on it shouldn't be too hard, I hope! |
…rom<[u32; 3]> for StreamId and From<[u8; 4]> for BlockPos
…he WordPosInput struct and adjusting the documentation; added checks in test_set_and_get_equivalence
|
I've made Some questions I have:
|
|
I've now got a working version where the backends seem to be safely zeroizing when the |
|
@nstilt1 can you rebase? This is looking mostly good and I'd like to get it merged. |
|
Sure thing. I apologize for the delay. I uh... I was mis-using ChaCha in several ways. I should have it rebased by tomorrow. |
…sting (and stopped working), fixed some spelling
…os to Rng instead of RngCore; updated docs to reflect changes
|
*block_pos. I moved set/get_block_pos to the RNG rather than its core, where it wasn't able to change the RNG's index after trying to change the block pos |
| // Tested for N=32; could be bugs in the loop bounds for other N | ||
| // returns bytes written, like fwrite: N means no error, 0 means error in all fwrites | ||
| size_t LongNumPrint( uint8_t *num, size_t N) | ||
| { | ||
| // caller can print a name if it wants | ||
|
|
||
| const int revbufsize = 8192; // 8kiB on the stack should be fine | ||
| alignas(32) char revbuf[revbufsize]; | ||
|
|
||
| if (N<32) { | ||
| // TODO: maybe use a smaller revbuf for this case to avoid touching new stack pages | ||
| ASCIIrev32B(revbuf, num); // the data we want is at the *end* of a 32-byte reverse | ||
| return fwrite(revbuf+32-N, 1, N, stdout); | ||
| } | ||
|
|
||
| size_t bytes_written = 0; | ||
| const uint8_t *inp = num+N; // start with last 32 bytes of num[] | ||
| do { | ||
| size_t chunksize = (inp - num >= revbufsize) ? revbufsize : inp - num; | ||
|
|
||
| const uint8_t *inp_stop = inp - chunksize + 32; // leave one full vector for the end | ||
| uint8_t *outp = revbuf; | ||
| while (inp > inp_stop) { // may run 0 times | ||
| inp -= 32; | ||
| ASCIIrev32B(outp, inp); | ||
| outp += 32; | ||
| } | ||
| // reverse first (lowest address) 32 bytes of this chunk of num | ||
| // into last 32 bytes of this chunk of revbuf | ||
| // if chunksize%32 != 0 this will overlap, which is fine. | ||
| ASCIIrev32B(revbuf + chunksize - 32, inp_stop - 32); | ||
| bytes_written += fwrite(revbuf, 1, chunksize, stdout); | ||
| inp = inp_stop - 32; | ||
| } while ( inp > num ); | ||
|
|
||
| return bytes_written; | ||
| // caller can putchar('\n') if it wants | ||
| } |
|
This looks very close now, though I'm not sure what's up with |
|
In the interest of moving this along since it's important, been taking awhile, and rather invasive, I will go ahead and merge and we can address the remaining issues as followups |
tarcieri
changed the title
|
I have no idea how that code ended up there, but it is from this post: https://stackoverflow.com/questions/61165307/ I think I was trying to explore palindromic cubes, looking for the fastest way to check if a large number is a palindrome... but I don't know how that code ended up there |
I borrowed some code from version
0.8.1ofchacha20, as well asrand_chachaand was able to get it to compile and pass tests. The main issues in my code that I identified were:[u8; 12]instead ofu128for set/get_stream(). It used to have au64when it had a 64 bit counterset_stream()takes the lower 96 bits of a u128 input. It now has support for using a[u8; 12]withset/get_stream_bytes()ParBlock<LesserBlock>as a temporary output buffer prior to copying it intoresultsinfn generate(). It will probably leave a copy of itself in memory after use. It might be possible to change theBlockRngResultstype toParBlock<LesserBlock>AsRef<[u32]> for BlockRngResultsandAsMut<[u32]> for BlockRngResultsset_word_pos()with an input that wasn't divisible by 16,get_word_pos()would round the input down to the nearest multiple of 16.rand_chacha.I've also added
ZeroizeOnDropfor theBlockRngResults.rust-random/rand#934