Use chacha in `get_secure_random_bytes()` #1974

danielgranhao · 2023-01-20T11:38:16Z

Addresses #1958

lightning/src/chain/keysinterface.rs

tnull

Github won't let me get there, but note that rand_bytes_unique_start is still initialized but now unused:

https://github.com/danielgranhao/rust-lightning/blob/a512d3dda038aec465122aeae2de4844c4cacfca/lightning/src/chain/keysinterface.rs#L1020-L1023

danielgranhao · 2023-01-20T16:07:44Z

@tnull Thank you for your comments. Somehow I didn't notice the unused value warnings. In d3f14e7 I also removed rand_bytes_unique_start.

I'm still missing fixing the failing test ln::functional_tests::test_duplicate_payment_hash_one_failure_one_success(). I will try to look a bit more into it, but if anybody has any intuition about the problem, suggestions on what to look for would be highly appreciated!

TheBlueMatt · 2023-01-25T03:11:19Z

#1984 should fix it :)

jkczyz · 2023-01-25T15:31:30Z

lightning/src/chain/keysinterface.rs

+				let mut nonce = Vec::new();
+				nonce.append(&mut starting_time_secs.to_be_bytes().to_vec());
+				nonce.append(&mut starting_time_nanos.to_be_bytes().to_vec());
+				let chacha = Mutex::new(ChaCha20::new(seed, &nonce));


I think we can avoid three vec allocations by having a fixed array and copying the bytes into it.

Thanks! I've changed to that approach in fac7b4e.

TheBlueMatt · 2023-01-26T00:34:28Z

lightning/src/chain/keysinterface.rs

-				rand_bytes_unique_start.input(&starting_time_nanos.to_be_bytes());
-				rand_bytes_unique_start.input(seed);
+				let mut rand_bytes_unique_start = [0u8; 32];
+				rand_bytes_unique_start[..8].copy_from_slice(&starting_time_secs.to_be_bytes());


We need to hash seed in here somehow. IMO we should keep it as a hash, basically the way it was and just complete the hash rather than leaving it as an unfinished engine.

Yeah, was thinking we may just want to restore and use the original rand_bytes_unique_start as the ChaCha seed?

Yea, restore the old code but complete the hash.

Ok, if I understood you correctly, this is what you propose -> 6472af6

TheBlueMatt · 2023-01-26T00:35:44Z

lightning/src/chain/keysinterface.rs

+
+		for _ in 1..5 {
+			let keys_manager_clone = Arc::clone(&keys_manager);
+			thread::spawn(move || {


You have to join these threads at the end, no?

yep, ty! Applied in 34fa5af

TheBlueMatt · 2023-01-26T00:39:21Z

lightning/src/chain/keysinterface.rs

-	rand_bytes_child_index: AtomicUsize,
-	rand_bytes_unique_start: Sha256State,
+	rand_bytes_unique_start: [u8; 32],
+	rand_bytes_index: AtomicUsize,


I think this should be a util::atomic_counter::AtomicCounter.

Applied in 6472af6

danielgranhao · 2023-01-26T17:58:06Z

I've rebased to take in the changes from #1984. I've added d08d4f8 just to check that fixing test_bump_penalty_txn_on_revoked_htlcs will fix the remaining CI checks. I suppose it needs an improvement similar to the one on test_duplicate_payment_hash_one_failure_one_success.

TheBlueMatt · 2023-01-26T18:37:38Z

LGTM, will let tnull take a look.

tnull

Alright, repeated the benchmarks locally, which drew a pretty clear picture:

| Threads | Method         | Time (ns) |
|---------+----------------+-----------|
| 1       | Hashing        | 1,805,139 |
| 1       | ChaCha/Mutex   | 6,308     |
| 1       | ChaCha/Counter | 10,725    |
| 3       | Hashing        | 1,875,495 |
| 3       | ChaCha/Mutex   | 144,687   |
| 3       | ChaCha/Counter | 18,778    |
| 5       | Hashing        | 1,898,584 |
| 5       | ChaCha/Mutex   | 92,116    |
| 5       | ChaCha/Counter | 40,465    |
|---------+----------------+-----------|

I'd say if we're confident in the counter method, it is probably the way to go.

That said, I'd also be happy to with the Mutex variant, as it's more straight forward, the performance difference is really not that big and realistically the kind of lock contention simulated in the benchmark is really unlikely to happen in production. Note however that in this case we may want to throw in a Sha256 for good measure, instead of using the seed directly.

tnull · 2023-01-26T17:18:58Z

lightning/src/chain/keysinterface.rs

-				nonce[..8].copy_from_slice(&starting_time_secs.to_be_bytes());
-				nonce[8..12].copy_from_slice(&starting_time_nanos.to_be_bytes());
-				let chacha = Mutex::new(ChaCha20::new(seed, &nonce));
+				let mut rand_bytes_unique_start = Sha256::engine();


nit: Rather than shadowing the variable, can we maybe rename this rand_engine_unique_start?

Sure, applied in 1456770.

TheBlueMatt · 2023-01-26T18:53:15Z

I'm more comfortable using the counter method - the ChaCha in-crate rolls doesn't handle chacha counter rollover, meaning it breaks at u32::MAX * 64 bytes, which, okay, we're not gonna hit, but still, the version with the counter doesn't.

TheBlueMatt · 2023-01-26T18:53:43Z

Please also rebase/squash the fixup commit at the end into the appropriate commit.

tnull

Looks good, I think I'm ACK after the tests pass.

lightning/src/chain/keysinterface.rs

danielgranhao force-pushed the speed-up-secure-random-byte-gen branch from c896c0b to a512d3d Compare January 20, 2023 12:13

danielgranhao mentioned this pull request Jan 20, 2023

Consier speeding up get_secure_random_bytes #1958

Closed

tnull reviewed Jan 20, 2023

View reviewed changes

lightning/src/chain/keysinterface.rs Outdated Show resolved Hide resolved

lightning/src/chain/keysinterface.rs Outdated Show resolved Hide resolved

tnull reviewed Jan 20, 2023

View reviewed changes

danielgranhao force-pushed the speed-up-secure-random-byte-gen branch from a512d3d to d3f14e7 Compare January 20, 2023 15:58

jkczyz reviewed Jan 25, 2023

View reviewed changes

danielgranhao force-pushed the speed-up-secure-random-byte-gen branch 3 times, most recently from d943b28 to 17860ac Compare January 25, 2023 23:03

TheBlueMatt reviewed Jan 26, 2023

View reviewed changes

danielgranhao force-pushed the speed-up-secure-random-byte-gen branch from 17860ac to 6472af6 Compare January 26, 2023 11:10

Add get_secure_random_bytes() benchmark

eedaf90

danielgranhao force-pushed the speed-up-secure-random-byte-gen branch from 6472af6 to d08d4f8 Compare January 26, 2023 17:53

tnull reviewed Jan 26, 2023

View reviewed changes

danielgranhao force-pushed the speed-up-secure-random-byte-gen branch from d08d4f8 to 1456770 Compare January 26, 2023 19:04

danielgranhao marked this pull request as ready for review January 26, 2023 19:06

tnull reviewed Jan 26, 2023

View reviewed changes

TheBlueMatt previously approved these changes Jan 26, 2023

View reviewed changes

lightning/src/chain/keysinterface.rs Show resolved Hide resolved

Use Chacha20 in get_secure_random_bytes()

f19821d

danielgranhao dismissed TheBlueMatt’s stale review via f19821d January 26, 2023 20:58

danielgranhao force-pushed the speed-up-secure-random-byte-gen branch from 1456770 to f19821d Compare January 26, 2023 20:58

TheBlueMatt approved these changes Jan 26, 2023

View reviewed changes

tnull approved these changes Jan 26, 2023

View reviewed changes

TheBlueMatt merged commit d4de913 into lightningdevkit:main Jan 26, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Use chacha in `get_secure_random_bytes()` #1974

Use chacha in `get_secure_random_bytes()` #1974

danielgranhao commented Jan 20, 2023 •

edited

tnull left a comment

danielgranhao commented Jan 20, 2023

TheBlueMatt commented Jan 25, 2023

jkczyz Jan 25, 2023

danielgranhao Jan 25, 2023

TheBlueMatt Jan 26, 2023

tnull Jan 26, 2023

TheBlueMatt Jan 26, 2023

danielgranhao Jan 26, 2023

TheBlueMatt Jan 26, 2023

danielgranhao Jan 26, 2023

TheBlueMatt Jan 26, 2023

danielgranhao Jan 26, 2023

danielgranhao commented Jan 26, 2023

TheBlueMatt commented Jan 26, 2023

tnull left a comment

tnull Jan 26, 2023

danielgranhao Jan 26, 2023

TheBlueMatt commented Jan 26, 2023

TheBlueMatt commented Jan 26, 2023

tnull left a comment

Use chacha in get_secure_random_bytes() #1974

Use chacha in get_secure_random_bytes() #1974

Conversation

danielgranhao commented Jan 20, 2023 • edited

tnull left a comment

Choose a reason for hiding this comment

danielgranhao commented Jan 20, 2023

TheBlueMatt commented Jan 25, 2023

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

danielgranhao commented Jan 26, 2023

TheBlueMatt commented Jan 26, 2023

tnull left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

TheBlueMatt commented Jan 26, 2023

TheBlueMatt commented Jan 26, 2023

tnull left a comment

Choose a reason for hiding this comment

Use chacha in `get_secure_random_bytes()` #1974

Use chacha in `get_secure_random_bytes()` #1974

danielgranhao commented Jan 20, 2023 •

edited