simd: enable core bpf programs

proposals/enable-core-bpf-programs.md

@@ -0,0 +1,171 @@
+---
+simd: 'XXXX'
+title: Enable Core BPF Programs
+authors:
+  - Joe Caulfield
+category: Standard
+type: Core
+status: Draft
+created: 2023-11-07
+feature: (fill in with feature tracking issues once accepted)
+supersedes: '0077'
+---
+
+## Summary
+
+### Roadmap
+
+This is SIMD 1/4 expected for **Multi-Client Feature Gates**. See
+<https://github.com/solana-foundation/solana-improvement-documents/issues/76>
+
+**Goals:**
+
+- An official process for upgrading core BPF programs and migrating native
+  programs to core BPF
+- Expanded control over pending feature activations
+- Automatic feature activation selection based on stake weight with supporting
+  software
+- Decentralized control over queuing new runtime features
+
+**Resulting Architecture:**
+
+1. 👉 **Core BPF Programs:** Enshrined BPF programs responsible for integral
+   cluster operations, including managing feature gates.
+2. **Feature Control:** Features can be created by anyone. Each is owned by the
+   core BPF Feature Gate program at 
+   `Feature111111111111111111111111111111111111`, which gives core contributors
+   added control over pending activations.
+3. **Feature Recognition & Activation:** Features qualify for activation based
+   on stake support of nodes who recognize the feature in their software
+   version. 
+4. **Feature Queuing:** A governance process nominates eligible features that
+   should be queued for activation.
+
+### Proposal
+
+This SIMD outlines the process by which engineers will manage upgrades to
+enshrined Core BPF programs, which are essential to network operations. This
+process will also be used to migrate existing native programs to Core BPF
+programs.
+
+## Motivation
+
+With the advent of Program Runtime v2 and Firedancer, native programs will
+eventually become BPF programs.
+
+Program Runtime v2 will introduce a brand-new ABI for on-chain programs, and the
+native programs will need to be revised to support it.
+
+BPF programs offer less complexity for other clients, such as Firedancer, since
+developers will no longer have to keep up with program changes in their runtime
+implementations. Instead, the program can just be modified once. 
+
+For these reasons, it makes sense to introduce the concept of Core BPF programs:
+BPF programs the network depends on that should be upgraded with a special
+process.
+
+With an explicit process in place for upgrading Core BPF programs, engineers can
+safely manage changes to essential programs through feature gates. The feature
+gate process requries incremental activations on each cluster: testnet, then
+devnet, then mainnet-beta. This is extremely valuable for ensuring a change has
+been safely integrated and will not have negative effects on mainnet-beta.
+
+This same explicit process can be used to migrate existing native programs to
+their new Core BPF versions.
+
+## New Terminology
+
+- `Core BPF Program`: A **non-upgradeable** BPF program relied on by any part of
+  the Solana stack, including (but not limited to) consensus, transaction
+  processing, voting, staking, and account creation.
+- `Target Program`: The core program intending to be changed by some
+  feature activation.
+- `Source Program`: The modified version of a core program intending to be moved
+  in place of the existing core program.
+
+## Detailed Design
+
+Core BPF programs shall be non-upgradeable BPF programs deployed with
+`BPFLoader2111111111111111111111111111111111`.
+
+Core BPF programs must only be modified through feature gates.
+
+**Upgrading a core BPF program** shall consist of deploying the modified program
+source code to a new arbitrary address as a **non-upgradeable** BPF program and
+creating a feature gate to, using the runtime, move the modified program in
+place of the existing program. 
+
+Note, since this deployed program will be deployed using the non-upgradeable BPF
+loader (`BPFLoader2111111111111111111111111111111111`), it will consist of only
+a program account, with no additional program _data_ account.
+
+This process must adhere to the following steps:
+
+1. Verifiably build the modified program.
+2. Generate a new keypair for the **source** program.
+3. Deploy the program to the **source** address.
+4. Generate a new keypair for the **feature gate**.
+5. Create a new feature gate for replacing the **target** program with the
+   **source** (modified) program.
+6. Follow the existing process for activating features.
+
+An additional optional field could be added to feature gate issues for the
+**target program** being upgraded.
+
+**Migrating a native program** to a core BPF program shall consist of following
+the exact steps outlined above with the modified BPF version of the program.
+This will effectively replace the program at the target address with its core
+BPF implementation.
+
+Some additional checks should be run when doing a migration instead of an
+upgrade, such as:
+
+- Validating the native program's owner is
+  `NativeLoader1111111111111111111111111111111`
+- Ensuring the core BPF program has the correct state
+
+## Impact
+
+This proposed change would result in a new process for upgrading core programs.
+Although this process will still require feature gates as it does now, the
+change is handled completely differently.
+
+The act of deploying the modified program and using a runtime feature gate to
+move it into the proper account will be an entirely new way of upgrading
+programs that were previously native.
+
+Core contributors will have to become comfortable with this process.
+
+## Security Considerations
+
+This proposal provides a secure means for upgrading core BPF programs - a
+process that will be increasingly valuable as new core BPF programs are created.
+
+However, this change also introduces a critical security consideration
+surrounding the replacement of a core program with the contents of another
+account. This is an extremely sensitive process that must be handled with
+maximum caution.
+
+If a core program is modified incorrectly, or somehow erased during migration,
+it could have immediate and fatal consequences for the network.
+
+## Backwards Compatibility
+
+This feature itself does not directly introduce any breaking changes.
+
+When this proposed process is used to migrate a native program to core BPF, that
+program will be completely backwards compatible.
+
+However, once a program has been migrated to core BPF, the process by which this
+program is upgraded will not be backward compatible. Core contributors must
+follow the new process.
+
+## Alternatives Considered
+
+An alternative approach to managing upgrades to essential core programs could be
+to make these core BPF programs _upgradeable_ and establish a multi-signature
+upgrade authority. This authority could be comprised of key-holders from each
+validator client: Solana Labs, Jito, Jump, and possibly Syndica (Sig).
+
+However, this approach adds additional layers of complexity around upgrading
+these programs behind a feature-gate.