-
Notifications
You must be signed in to change notification settings - Fork 98
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Merge branch 'ethereum-optimism:main' into plasma-tx-version
- Loading branch information
Showing
17 changed files
with
399 additions
and
191 deletions.
There are no files selected for viewing
Validating CODEOWNERS rules …
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -1,22 +1,41 @@ | ||
| set positional-arguments | ||
|
|
||
| # default recipe to display help information | ||
| default: | ||
| @just --list | ||
|
|
||
| # Install required dependencies | ||
| deps: | ||
| pnpm i --frozen-lockfile | ||
|
|
||
| # Lint the workspace for all available targets | ||
| lint: lint-specs-md-check lint-specs-toc-check lint-links | ||
|
|
||
| # Updates all files to fix linting issues | ||
| lint-fix: lint-specs-md-fix lint-specs-toc | ||
|
|
||
| # Validates markdown file formatting | ||
| lint-specs-md-check: | ||
| npx markdownlint-cli2 "./specs/**/*.md" | ||
|
|
||
| # Updates markdown files formatting to satisfy lints | ||
| lint-specs-md-fix: | ||
| npx markdownlint-cli2-fix "./specs/**/*.md" | ||
| npx markdownlint-cli2 --fix "./specs/**/*.md" | ||
|
|
||
| # Validates Table of Content Sections with doctoc | ||
| lint-specs-toc-check: | ||
| npx doctoc '--title=**Table of Contents**' ./specs && git diff --exit-code . | ||
| npx doctoc '--title=**Table of Contents**' ./specs && git diff --exit-code ./specs | ||
|
|
||
| # Updates Table of Content Sections with doctoc | ||
| lint-specs-toc: | ||
| npx doctoc '--title=**Table of Contents**' ./specs | ||
|
|
||
| # Validates all hyperlinks respond with status 200 | ||
| lint-links: | ||
| docker run --init -it -v `pwd`:/input lycheeverse/lychee --verbose --no-progress --exclude-loopback \ | ||
| --exclude twitter.com --exclude explorer.optimism.io --exclude linux-mips.org --exclude vitalik.ca \ | ||
| --exclude-mail /input/README.md "/input/specs/**/*.md" | ||
|
|
||
| serve: | ||
| mdbook serve | ||
| # Serves the mdbook locally | ||
| serve *args='': | ||
| mdbook serve $@ |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,88 @@ | ||
| # Diff Based L1Attributes Transaction | ||
|
|
||
| <!-- START doctoc generated TOC please keep comment here to allow auto update --> | ||
| <!-- DON'T EDIT THIS SECTION, INSTEAD RE-RUN doctoc TO UPDATE --> | ||
| **Table of Contents** | ||
|
|
||
| - [Implementation](#implementation) | ||
| - [Ecotone L1Attributes Example](#ecotone-l1attributes-example) | ||
| - [Optimization](#optimization) | ||
| - [Node Implementation](#node-implementation) | ||
|
|
||
| <!-- END doctoc generated TOC please keep comment here to allow auto update --> | ||
|
|
||
| The `L1Attributes` transaction is the first transaction included in every L2 block. | ||
| It can be thought of as an extension to the L2 block header that is 100% backwards | ||
| compatible with L1 Ethereum by not changing the structure of the block header itself. | ||
| A major problem with the `L1Attributes` transaction is that it constantly submits the | ||
| same information over and over again. This adds chain history that must be preserved | ||
| forever. Ideally, the `L1Attributes` transaction only includes diffs to reduce the | ||
| growth of the chain history. | ||
|
|
||
| ## Implementation | ||
|
|
||
| The `L1Attributes` transaction already commits to a schema that is ABI encoded. | ||
| There is no concept of an `Optional` type in ABI encoding, so instead we can | ||
| utilize a custom data structure to indicate the presence of a field in the schema. | ||
| A leading bitfield can be used to use presence. Each bit indicates a specific item | ||
| in the schema where `1` means it is present and `0` means it is not present. Each | ||
| item MUST have at least a fixed size portion so that the decoder knows how far | ||
| forward to seek after a read when the item is present. | ||
|
|
||
| ### Ecotone L1Attributes Example | ||
|
|
||
| The following fields are required as part of the `L1Attributes` transaction after | ||
| the Ecotone network upgrade. | ||
|
|
||
| | Field | Size (bytes) | Bitfield Representation | | ||
| | ----------------- | ------------ | ------------------------ | | ||
| | baseFeeScalar | 4 | 0b1000\_0000\_0000\_0000 | | ||
| | blobBaseFeeScalar | 4 | 0b0100\_0000\_0000\_0000 | | ||
| | sequenceNumber | 8 | 0b0010\_0000\_0000\_0000 | | ||
| | l1BlockTimestamp | 8 | 0b0001\_0000\_0000\_0000 | | ||
| | l1BlockNumber | 8 | 0b0000\_1000\_0000\_0000 | | ||
| | basefee | 32 | 0b0000\_0100\_0000\_0000 | | ||
| | blobBaseFee | 32 | 0b0000\_0010\_0000\_0000 | | ||
| | l1BlockHash | 32 | 0b0000\_0001\_0000\_0000 | | ||
| | batcherHash | 32 | 0b0000\_0000\_1000\_0000 | | ||
|
|
||
| This is a total of 9 fields, so 9 bits are required to represent all of them. | ||
| It cannot fit into a single byte, so 2 bytes should be used to represent the bitfield. | ||
| Following the bitfield is the data tightly packed together. | ||
| When the value is set to `1` in the location that corresponds to the field, then | ||
| it indicates that the value is present and the the `size` number of bytes should | ||
| be read from the `calldata` and the offset of the next read from the calldata is | ||
| incremented by the `size`. If the value in the bitfield is a `0`, then the data | ||
| is not present and any reading/seeking is skipped. | ||
|
|
||
| ### Optimization | ||
|
|
||
| The above example can be optimized by grouping the attributes that always change together | ||
| into a single bit. This would be the `l1BlockTimestamp`, the `l1BlockNumber` and `l1BlockHash`. | ||
| In the case of a missed L1 slot or L1 reorg, it is possible to have the `l1BlockNumber` not change but | ||
| the others may change. This case doesn't happen often and would result in some repeat data being | ||
| posted. | ||
|
|
||
| It is not safe to omit the `sequenceNumber` due to the fact that it always increases because we need | ||
| to guarantee that L1 Attributes Transactions have unique transaction hashes. In this way it acts as a nonce. | ||
|
|
||
| ## Node Implementation | ||
|
|
||
| Previously, the node could read the calldata from the latest `L1Attributes` transaction | ||
| to be able to populate it's internal view of the `SystemConfig`. This will not be possible | ||
| with a diff based `L1Attributes` transaction implementation, but the problem can be solved | ||
| easily with a batch RPC call that reads each value directly from state using `eth_call`. | ||
|
|
||
| The RPC receipts are hydrated with the L1 fee parameters such that it is possible to | ||
| compute the L1 portion of the fee based on observing the receipt. These fields are not | ||
| part of the consensus encoding of the receipt, simply to provide better UX. These fields | ||
| are populated by deserializing the L1Attributes transaction's calldata. If the L1Attributes | ||
| transaction no longer always contains the latest values, then this approach is no longer feasible. | ||
| This means that the values would need to be instead read directly from state. | ||
| To prevent the need of an archive node to read the values from state for historical receipts, | ||
| an event can be emitted that contains the values required to hydrate this data. | ||
| Ideally this event does not need to be emitted for every L1Attributes transaction, but a simple | ||
| implementation may do so. It could be emitted everytime that the `sequenceNumber` is `0` and | ||
| the rehydration method could observe the L1Attributes transactions from the last L1 origin | ||
| update to the block including the receipt being hydrated and apply the diffs against the value | ||
| that was emitted from the event. |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.