Add option to select KZG library for PeerDAS

[jimmygchen]

Description

We've recently switched from the c-kzg library to the Rust peerdas-kzg library on the das branch (#5941). The purpose for the switch was to:

1. Try out a new alternative to c-kzg;
2. Allow us to merge das branch to unstable without having to update c-kzg to a development version

However we've now lost the option to run the das branch with ckzg, which has been quite stable for a while. We've recently run into a few issues with peerdas-kzg, including slow startup time and stack overflow during block production (#6105). Therefore I think it would be great to have the option to use ckzg library, especially for public denvets, where a small number of reliable supernodes is required.

Ideally we want to be able to:

1. maintain the ckzg library version for the mainnet code path
2. optionally include the ckzg peerdas library when it's enabled via a feature flag, so both ckzg library will co-exist when this ckzg-peerdas feature is enabled, however this library will only be referenced by the PeerDAS portion of the code.
3. allow selecting the ckzg-peerdas library via a hidden runtime flag

[jimmygchen]

We currently use c-kzg for Deneb, and rust-eth-kzg for PeerDAS. To fully support this we'd need:

• Re-integrate PeerDAS with c-kzg, so we can have the option to load c-kzg only
• Use rust-eth-kzg to perform deneb KZG operations, so we can have the option to load rust-eth-kzg only
• If we complete both changes, then we can have option to choose which KZG library to use, or feature gate them (Add option to select KZG library for PeerDAS #6107)

rust-eth-kzg seems to perform better in PeerDAS - I'm not sure if this has changed.

Alternatively, an option is to maintain the status quo for now, and gradually phase out c-kzg over time? We can't completely switch over to rust KZG yet, AFAIK it hasn't been audited, but I'd assume there are plans to do this.

[jimmygchen]

@pawanjay176 also expressed his preference to maintain the status quo now and eventually get rid of c-kzg. It's unclear whether both libraries will be maintained in the longer term.