Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

FIP-TBD: Export sector termination method from miner actor #1035

Open
wants to merge 4 commits into
base: master
Choose a base branch
from
Open
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
115 changes: 115 additions & 0 deletions FIPS/fip-tbd.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,115 @@
---
fip: "tbd"
jennijuju marked this conversation as resolved.
Show resolved Hide resolved
title: Export sector termination method from miner actor
author: @jennijuju, @stebalien, @anorth, @Schwartz10,
discussions-to: https://github.com/filecoin-project/FIPs/discussions/1034
status: Draft
type: Technical
category (*only required for Standard Track): <Core>
created: 2024-07-12
---


# FIP-TBD: Export sector termination method from miner actor

## Simple Summary

Adds a new `TerminateSectors2` (method 37) that allows callers to submit sectors for termination and define the maximum amount of sectors to be terminated immediately.

Expose `TerminateSectors2` method for invocation by user actors, so to enable more decentralized DeFi protocol for Storage Provider (SP) services.

## Abstract

Following FEVM launch, [FIP-0050](https://github.com/filecoin-project/FIPs/blob/master/FIPS/fip-0050.md) introduced a list of builtin actor methods that are exposed as callables from user actors. Since then, we have seen a number of protocols use miner actor collateral and/or rewards to build lending and leasing applications.

It is a common practice in DeFi applications to manage risk for liquidity providers (i.e., those lending money) by "liquidating" collateral supplied by borrowers. In the Filecoin ecosystem, 'liquidation' is achieved when Storage Providers terminate a miner's active sectors, pay the required penalty, recover their pledged collateral, and repay liquidity providers. This process is a critical part of FEVM economic security, as is understanding the precise value of existing collateral.


This FIP aims to introduce an exported sector termination method in miner actor to support on-chain liquidation procedures.

## Change Motivation

As DeFi protocols on Filecoin gain traction and lock larger amounts of value, they will be targets for more attacks. As an ecosystem, we should create the necessary primitives to build secure economic applications that use miner actors as collateral.

There are a few motivations to this change:

- Trustless DeFi - it’s important that users of a DeFi protocol can trust code and not humans to protect their assets. When humans are in the loop, there is necessarily a “judgment call zone”, which violates all aspects of DeFi, it also creates personal safety issues for individuals who are known to process liquidations with no plausible deniability.
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I am not sure how this statement supports the changes proposed by this FIP. Is the correct interpretation that exporting these actors simply allows for more/better DeFi operations to be executed via code?

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
- Trustless DeFi - it’s important that users of a DeFi protocol can trust code and not humans to protect their assets. When humans are in the loop, there is necessarily a “judgment call zone”, which violates all aspects of DeFi, it also creates personal safety issues for individuals who are known to process liquidations with no plausible deniability.
- Trustless DeFi - it’s important that users of a DeFi protocol can trust code and not humans to protect their assets. When humans are in the loop, there is necessarily a “judgment call zone”, which violates all aspects of DeFi. It also creates personal safety issues for individuals who are known to process liquidations with no plausible deniability.

- Liability concern - from the DeFi protocol creator’s perspective, being an external third party that has to process liquidations off chain (rather than creating a keeper network of liquidators) introduces liability and regulatory concern to the protocol, making it harder to operate in certain jurisdictions
- Incentive design - by designing their own rulesets for liquidations, application developers can incentivize keepers to protect their protocol in a decentralized manner

This FIPs propose a builtin actors APIs that can trigger sector termination in smart contracts that staking protocols often need to build offchain logics/oracles for currently.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
This FIPs propose a builtin actors APIs that can trigger sector termination in smart contracts that staking protocols often need to build offchain logics/oracles for currently.
This FIP proposes a builtin actors APIs that can trigger sector termination in smart contracts that staking protocols often need to build offchain logics/oracles for currently.


## Specification

Expose the following APIs, following the FIP-0050 specification for exported builtin actor methods.

**TerminateSectors2** In Miner Actor

To allow caller (storage providers, stake pools and etc) to have more precise control over the amount of the sectors to be terminated based on their operational needs, this FIP introduces a new `TerminateSector2` method. This method follows most of the behaviour of the existing `TerminateSectors` method (method 9), notably:
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
To allow caller (storage providers, stake pools and etc) to have more precise control over the amount of the sectors to be terminated based on their operational needs, this FIP introduces a new `TerminateSector2` method. This method follows most of the behaviour of the existing `TerminateSectors` method (method 9), notably:
To allow callers (e.g., storage providers, stake pools) to have more precise control over the amount of the sectors to be terminated based on their operational needs, this FIP introduces a new `TerminateSector2` method. This method follows most of the behaviour of the existing `TerminateSectors` method (method 9), notably:

- Only a control address of the miner actor can call this method.
- Attempt to terminate sectors that are in the current or the next proving deadline will fail with `USR_ILLEGAL_ARGUMENT`.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
- Attempt to terminate sectors that are in the current or the next proving deadline will fail with `USR_ILLEGAL_ARGUMENT`.
- Attempts to terminate sectors that are in the current or the next proving deadline will fail with `USR_ILLEGAL_ARGUMENT`.

- The method will always process the sector(s) that are already in the `early_termination` queue before processing newly submitted sectors.

This FIP proposes the following behaviour changes:
- Currently, `TerminateSectors` allows callers to submit early termination for at most 3000 sectors in one message. `TerminateSectors2` will remove this hardcoded upper bound. Instead, `TerminateSectors2` takes a `max_termination` parameter for caller to specify the maximum number of sectors to be terminated.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
- Currently, `TerminateSectors` allows callers to submit early termination for at most 3000 sectors in one message. `TerminateSectors2` will remove this hardcoded upper bound. Instead, `TerminateSectors2` takes a `max_termination` parameter for caller to specify the maximum number of sectors to be terminated.
- Currently, `TerminateSectors` allows callers to submit early termination for at most 3000 sectors in one message. `TerminateSectors2` will remove this hardcoded upper bound. Instead, `TerminateSectors2` takes a `max_termination` parameter for callers to specify the maximum number of sectors to be terminated.

- `TerminateSectors2` will always first terminate the sectors that are already in the `early_termination` queue (either added through cron or through `TerminateSectors(2)` messages), then process the new sectors submitted in the message. This prevents the same sectors to be added to the termination queue in the cron and consume block space maliciously.
- A newly submitted sector will be skipped if it was already in the `early_termination` queue and got executed first, so that the sector will not be terminated twice.
- `TerminateSectors2` will only terminate sectors up to the number caller specified in the `max_termination` parameter. If there are more sectors in `early_termination` queue and the new sectors submitted in the message than the `max_termination` amount, the message will fail with `USR_ILLEGAL_ARGUMENT`.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
- `TerminateSectors2` will only terminate sectors up to the number caller specified in the `max_termination` parameter. If there are more sectors in `early_termination` queue and the new sectors submitted in the message than the `max_termination` amount, the message will fail with `USR_ILLEGAL_ARGUMENT`.
- `TerminateSectors2` will only terminate sectors up to the number specified in the `max_termination` parameter. If there are more sectors in `early_termination` queue and the new sectors submitted in the message than the `max_termination` amount, the message will fail with `USR_ILLEGAL_ARGUMENT`.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I know we discussed this a lot but I have forgotten the various reasoning. Why does this fail if there were no new sectors submitted for termination but the queue is longer than max_terminations? I think we should allow manually terminating part of the queue even if it's too long to ever do in a single message.

Failing if it can't terminate the sectors that were submitted extra (because the queue is too long) is ok

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think we should allow manually terminating part of the queue even if it's too long to ever do in a single message.

this makes sense.

- Caller can query the `early_termination` queue size ahead of submitting the message to determine the maximum number of sectors and/or new sectors it can terminate in one message.
- That said, `TerminateSectors2` will not add sectors to the termination queue.
- `TerminateSectors2` will fail with `SYS_OUT_OF_GAS` and abort the all execution if the caller doesn't have enough gas to execute the termination and/or the message itself run out of gas.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
- `TerminateSectors2` will fail with `SYS_OUT_OF_GAS` and abort the all execution if the caller doesn't have enough gas to execute the termination and/or the message itself run out of gas.
- `TerminateSectors2` will fail with `SYS_OUT_OF_GAS` and abort all execution if the caller doesn't have enough gas to execute the termination and/or the message itself runs out of gas.


The `TerminateSectors2Params` follows the existing `TerminateSectorsParams` structure, with an additional `max_termination` parameter.

`Vec<TerminationDeclaration>` can be nullable. When it is null, flush/terminate the sectors that are queued up in the cron for termination.
jennijuju marked this conversation as resolved.
Show resolved Hide resolved

```rust
pub struct TerminateSectors2Params {
pub terminations: Option<Vec<TerminationDeclaration>>,
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Do we need both optional and empty vector? How does the behaviour of an option with Some([]) differ from submitting None? Would just a vector suffice?

pub max_termination: u64,
}

pub struct TerminateSectors2Return {
// Set to true if all early termination work and new sectors termination work have been completed. Set to false otherwise.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This needs more explanation of how this return value could be false and when the entire method call fails instead. Again, I know we discussed this a while ago, but I have forgotten the outcomes.

pub done: bool,
}
```

jennijuju marked this conversation as resolved.
Show resolved Hide resolved
**TerminateSectors2Exported** In Miner Actor

Add `TerminateSectors2Exported` method to Miner builtin actor. The method ID is `2777360141`.


## Backwards Compatibility

This FIP introduces new builtin actor methods, therefore needs a new actors version shipped in a network upgrade. No breaking changes to existing methods, however, existing methods may be deprecated in future network upgrades.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
This FIP introduces new builtin actor methods, therefore needs a new actors version shipped in a network upgrade. No breaking changes to existing methods, however, existing methods may be deprecated in future network upgrades.
This FIP introduces new builtin actor methods, therefore needs a new actors version shipped in a network upgrade. There are no breaking changes to existing methods, however, existing methods may be deprecated in future network upgrades.


This FIP does not require a state migration.

## Test Cases
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Should we have a TODO to enumerate this out more? I assume we ultimately want a lit of asserts to make in the test cases to make sure everyone is aligned.


- TODO


## Security Considerations

No new method is being introduced that can be used to attack the network, and bugs in the implementation will only lead to incorrect behavior being observed by user actors.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Are we sure? Isn't having a smart contract being able to invoke terminatesector on an SP's behalf a big deal.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

added to the secuirty consideration

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This line doesn't seem right to me at all. The method could be used to attack the network if it has a problem, and bugs could lead to all kinds of unexpected behaviour. But nevertheless, that's not the kind of thing we've bothered writing in any FIP that introduces a new method in the past. This isn't any safer or more dangerous than other FIPs. I don't think this sentence belongs at all.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Will remove!



To prevents the same sectors to be added to the termination queue in the cron and consume block space maliciously with programmatic smart contracts, `TerminateSectors2` will not add any sectors to the cron queue and it will always terminate the sectors that are already in the `early_termination` queue first, then process the new sectors submitted in the message and skip the sectors that are already in the `early_termination` queue.

While the newly introduced `TerminateSectors2` method allows a smart contract to terminate a sector on SPs behalf, it still requires the SP to add the smart contract as a control address first. When SPs signs and send the message to add smart contract as a control address, this should be considered as a high trust action, as the SP is authorizing the control right to the smart contract.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
While the newly introduced `TerminateSectors2` method allows a smart contract to terminate a sector on SPs behalf, it still requires the SP to add the smart contract as a control address first. When SPs signs and send the message to add smart contract as a control address, this should be considered as a high trust action, as the SP is authorizing the control right to the smart contract.
While the newly introduced `TerminateSectors2` method allows a smart contract to terminate a sector on an SP's behalf, it still requires the SP to add the smart contract as a control address first. When a SP signs and sends the message to add smart contract as a control address, this should be considered as a high trust action, as the SP is authorizing the control right to the smart contract.


## Product and Incentive Considerations

This FIP enables more trustless and decentralized staking and DeFi protocol for token holders to participate in. Staking and Defi protocol can drive Filecoin onchain activities, improve FIL utilities and enabling more (1) storage providers to get liquidity to join the network and provide storage services & secure the network consensus (2) token holders to put their FIL into use and improve Filecoin GDP.


## Implementation

TODO

## Copyright
Copyright and related rights waived via [CC0](https://creativecommons.org/publicdomain/zero/1.0/).


Loading