Support Secp256r1 signature verification

[pluscubed]

Following up on a "Bonus Point" in #751, it'd be great to have Secp256r1 signature verification.

This would allow iOS/Android hardware-backed signatures, e.g. for smart contract wallets. (Similar: cosmos/cosmos-sdk#7718)

[webmaster128]

This would be doable. What do you need it for? We are curious about the use case and priority.

[webmaster128]

@reuvenpo do you have any opinion on this? You asked for Secp256r1 recently in a different context.

[reuvenpo]

I was asking about it because I thought Cosmos were adding support for it on the client side, so I thought that I'd need to support it on the Secret Network's enclave side. But since that's not the case, I do not have any use for secp256r1 in any context.

[assafmo]

For the record, I don't think it's a good design choice to put these functions in the final VM API. I think it should be in cw-plus and I'm sure there is a way to eliminate all the bloat that comes with it by default in order to minimize binary size.

[maurolacy]

For the record, I don't think it's a good design choice to put these functions in the final VM API. I think it should be in cw-plus and I'm sure there is a way to eliminate all the bloat that comes with it by default in order to minimize binary size.

@webmaster128 can probably answer this better than me, but I think it's not so much an issue of module size (though that's relevant, too), but performance (gas usage).

By putting the crypto stuff in the VM, it compiles to native code instead of wasm, with a significant performance improvement.

Given that cryptographic primitives are usually computationally intensive, this makes particular sense here.

[webmaster128]

This request is resonable but out of scope for 1.0.0. A draft implementation is developed in #1083. It can be added to a running chain later on via a new feature flag in a non-breaking way.

[webmaster128]

This seems to become more important recently for account abstraction. I put this on the agenda for after CosmWasm 2.0.

[webmaster128]

See follow-up ticket #2001 for everything you need to know about secp256r1 support.