I think the cleanest way to solve this problem is to use the go-pointers library and pass in the opaque pointer from the application through to the go callback.
Here's a blog post I found detailing how this is done: https://eli.thegreenplace.net/2019/passing-callbacks-and-pointers-to-cgo/ On Tuesday, October 18, 2022 at 9:50:21 PM UTC-4 David Stainton wrote: > > Okay here I've tried a similar solution to the go-pointers library... but > i had to change it a bit to fit how the C library uses the context pointer: > > https://git.xx.network/elixxir/ctidh_cgo/-/commit/8f02611b943c269eab5b6e888958e4d41d72cd51 > > It builds and runs... and then panics on the "rng is nil"... not sure why. > So I'm going to try another solution where I change the C api slightly. > > I'm convinced there are many possible solutions to this and picking one is > a matter of deciding among a few tradeoffs. > > > On Tuesday, October 18, 2022 at 6:33:11 PM UTC-4 David Stainton wrote: > >> Greetings cgo experts, perhaps this post might also be of some interest >> to cryptography people as well. >> >> What's the best way for a C cryptography library to receive entropy from >> golang? >> >> There exists a C cryptography library (CTIDH, it's a PQ NIKE)... I am >> collaborating with C programmers who are adding features to it. I maintain >> cgo bindings for this library. Thus far the cgo is straight forward and it >> works. However the C programmers recently added a constructor for private >> keys which takes a function pointer as an argument: >> >> https://codeberg.org/io/highctidh/src/branch/main/csidh.h#L50 >> >> /* >> * generate a new private key using rng_callback and write the result to >> (priv). >> * (priv) is passed as (context) to the rng_callback. >> */ >> void csidh_private_withrng(private_key *priv, ctidh_fillrandom >> rng_callback); >> >> I did get this to work with an ugly hack where the golang constructor >> takes an io.Reader interface object as an argument and sets an unexported >> module scoped global variable to the rng/io.Reader object, guarded by a >> mutex, gross: >> >> >> https://git.xx.network/elixxir/ctidh_cgo/-/commit/2e7fb078722decb07ab8da7fb59c8ec46fb32fd1 >> >> But that's a really ugly hack and if called by multiple threads would >> result in the rng getting overwritten... which in most cases is probably >> fine for that to happen but seems like a bad design. >> >> >> >> >> -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/a8b5fa3e-3ff0-45f1-8864-c169d6cafa23n%40googlegroups.com.
