block-ciphers VS ring

If found this set of crates for other algorithms : https://github.com/RustCrypto/hashes And also found this set of crates that seem to include a lot of block cyphers : https://github.com/RustCrypto/block-ciphers Even if "des" is listed as a crate in this last link, it doesn't seem to provide the DES algorithm entirely.

If you are trying to use it as a library in your own crate then I would suggest looking at the [patch.crates-io] section of your Cargo.toml. It should allow you to override the dependency and point it to something else. Under that section set aes = { git = 'https://github.com/RustCrypto/block-ciphers', rev = 'e59142b26edcaa5e287c7e5067be8a501b42f9cb' }, changing the rev key to whichever commit has the right version of the crate when it was published. Then do the same for block-cipher and any others that it cannot find the version for but with the correct repository and commit.

aes (GitHub: RustCrypto / block-ciphers / aes) good: still maintained as of now - last commit on GitHub is from October 2021 good: examples look easy to use good: has received an audit by NCC Group bad: seems a bit too low level - the example provided only shows usage with data that is exactly block sized - seems there is no padding handling for real world use cases

I have no idea which libraries have the best code for Raspberry Pi 4. I think it doesn't have hardware AES, so an implementation of AES that doesn't leak secret bits through side channels and is fast would be complicated. The code I would trust is this: https://github.com/RustCrypto/block-ciphers but I have no idea whether it has python bindings. I would also sorry about correctly reusing buffers or else the memory allocation would be the bottleneck.

You should look at this this: https://github.com/rust-cc/awesome-cryptography-rust and you probably need this: https://github.com/RustCrypto/block-ciphers

For example, here is a bitsliced AES S-box written in single assignment form. Granted that's not the entire cipher, but the entire cipher can be implemented that way if you so desire.

Again, this is just a temporary situation, and a matter of burning down a list of small tasks. Not that the OpenSSL license issue is a big deal for most anyway. Feel free to help; see this issue filed by Josh Triplett: https://github.com/briansmith/ring/issues/1318#issuecomment-...

Note also that rustls depends on ring, which has architecture-dependent code in it that is not as widely compatible as eg. OpenSSL/GnuTLS/Mbed-TLS. For example, MIPS is not supported by ring.

The AWS Rust library we were using as a dependency depended on a cryptography library called ring. This library leverages C and assembly code to implement its cryptographic primitives. Unfortunately, cross compiling when C is involved can add complexity to the build process. While it might've been possible to overcome these issues I decided that it wasn't worth digging into more.

That'd be great. Thanks Brian. Re: making ring portable to all platforms: IBM have been graciously maintaining a up to date patchset for Ring for years now and there's an outstanding PR here you may not have seen since they filed it in 2020... https://github.com/briansmith/ring/pull/1057

Beyond the simple matter of Rust being much newer than OpenSSL, one concern for some cryptographic primitives is the timing side-channel.

https://en.wikipedia.org/wiki/Timing_attack

In high level languages like Rust, the compiler does not prioritise trying to emit machine code which executes in constant time for all inputs. OpenSSL has implementations for some primitives which are known to be constant time, which can be important.

One option if you're working with Rust anyway would be use something like Ring:

https://github.com/briansmith/ring

Ring's primitives are just taken from BoringSSL which is Google's fork of OpenSSL, they're a mix of C and assembly language, it's possible (though fraught) to write some constant time algorithms in C if you know which compiler will be used, and of course it's possible (if you read the performance manuals carefully) to write constant time assembly in many cases.

In the C / assembly language code of course you do not have any safety benefits.

It can certainly make sense to do this very tricky primitive stuff in dangerous C or assembly, but then write all the higher level stuff in Rust, and that's the sort of thing Ring is intended for. BoringSSL for example includes code to do X.509 parsing and signature validation in C, but those things aren't sensitive, a timing attack on my X.509 parsing tells you nothing of value, and it's complicated to do correctly so Rust could make sense.

machine learning, neural networks, image processing, cryptography (though it is getting better), font shaping/rendering (though it is getting better), CPU/software rendering (though it is getting better)