feat(l1): add support for ZisK backend #45

xqft · 2025-11-20T15:12:32Z

Motivation

a ZisK backend was added for ethrex-prover in lambdaclass/ethrex#5392. This PR adds supports to use it with ethrex-replay

…_decoded_nodes

Copilot

Pull Request Overview

This PR adds support for the ZisK backend to ethrex-replay by updating dependencies and adding the necessary feature flag and backend support.

Key Changes

Added ZisK feature flag and backend variant support
Updated all ethrex dependencies from tag v7.0.0 to branch add_zisk_zkvm_backend
Added Default::default() parameter to execution_witness_from_rpc_chain_config calls
Updated error messages for backend features

Reviewed Changes

Copilot reviewed 6 out of 10 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
Cargo.toml	Added `zisk` feature flag for ZisK backend support
src/cli.rs	Added ZisK backend handling in `backend()` function with error messages
src/run.rs	Added default parameter to witness conversion function calls
src/rpc/mod.rs	Updated trie node handling to decode nodes and use mutable reference
src/rpc/db.rs	Removed duplicate node insertion in trie construction
.DS_Store	macOS system file that should not be committed
.github/.DS_Store	macOS system file that should not be committed
src/.DS_Store	macOS system file that should not be committed
diff	Temporary diff artifact that should not be committed
Cargo.lock	Auto-generated dependency lock file updates

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

src/cli.rs

diff

src/cli.rs

JereSalo · 2025-11-26T13:22:06Z

src/cli.rs

+    let first_block_number = cache.get_first_block_number()?;
+    let Some(parent_block_header) = cache.blocks.iter().find_map(|b| {
+        if b.header.number == first_block_number - 1 {
+            Some(b.header.clone())
+        } else {
+            None
+        }
+    }) else {
+        eyre::bail!("No parent block header");
+    };


I think this doesn't make much sense.
The first_block_number is the first number that appears in cache.blocks, so the parent block header is never going to be in cache.blocks because you are looking for the parent of the first block, which is not going to be there. It is actually going to be in cache.witness.headers.
In these PRs I made changes that I think would work though:
#48
lambdaclass/ethrex#5416
I was planning on merging those after this is merged. The simplest thing to do for now IMO is to comment replay_no_zkvm (or make it work but it's going to change later)

This approach should be taken instead, as it's done in run.rs

let initial_state_root = db .headers .iter() .map(|h| { BlockHeader::decode(h).map_err(|_| eyre::Error::msg("Failed to decode block header")) }) .collect::<Result<Vec<_>, _>>()? .into_iter() .find(|h| h.number == first_block_number - 1) .map(|h| h.state_root) .ok_or_else(|| eyre::eyre!("Initial state root not found"))?;

good catch! fixed, thank you!!

JereSalo · 2025-11-26T13:35:35Z

src/cli.rs

+                    rpc_url: Some(rpc_url.clone()),
+                    cached: false,
+                    network: None,
+                    cache_dir: PathBuf::default(),


We should use the default cache dir that we always use, otherwise it's not going to look into the specific folder we designed for that. Maybe we should have a constant for DEFAULT_CACHE_DIR or something like that.
cache_dir: PathBuf::from("./replay_cache"),

JereSalo · 2025-11-26T13:47:30Z

src/rpc/db.rs

-        let hash = H256::from_slice(&Keccak256::digest(root));
-        state_nodes.insert(hash, root.clone());
-        hash
+        H256::from_slice(&Keccak256::digest(root))


Here we aren't inserting the root in the state_nodes as before. I don't know if it's critical but above we are doing proof.iter().skip(1) so the root should be inserted here, right?
I think the person who wrote this code tried not to hash the root twice I guess but I believe that if we remove this insert we are not going to have the root in the state_nodes.

Maybe we can change the proof.iter().skip(1) for proof.iter() and leave the rest of the code as is, hashing the root twice but we couldn't care less, right?

I can't tell if the reason is that but I recently tried to execute a block gathering the data with eth_getProof and I got Inconsisntent Trie. So that could be a possible cause.

Edit: Oh, It failed when trying to apply account updates it seems because of the known bug of missing nodes with eth_getProof, now I'm not sure if what I commented is a problem or not haha

I'm not sure what led me to changed this in the first place, solved!

JereSalo

Left comments, some things should be fixed. The Zisk integration works fine though

Additional Comment:

I know it's pretty basic but it would be good if we had brief docs on how to generate input for running a block and then feeding that into a zkvm. Saying that we should get the elf from replay releases I guess, right?

xqft · 2025-11-26T14:26:33Z

@JereSalo yes that's a good idea. added docs!

JereSalo · 2025-11-26T14:41:16Z

src/run.rs

        execution_witness,
        chain_config,
        block.header.number,
+        Default::default(),


This one is just for compiling I believe but we shouldn't use the default state root here

JereSalo

😃

…hrex-replay into add_support_for_zisk_zkvm

ilitteri and others added 7 commits November 4, 2025 11:16

Minimal changes

7481fea

Add proof type

d21c6ba

minimal changes

07edd3d

Merge branch 'main' into add_support_for_zisk_zkvm

56b7f6b

Merge branch 'main' of github.com:lambdaclass/ethrex-replay into pass…

75f5b73

…_decoded_nodes

Merge branch 'pass_decoded_nodes' into add_support_for_zisk_zkvm

db8643c

update feature

3d67455

github-actions bot assigned xqft Nov 20, 2025

fixes

4df31cd

xqft changed the title ~~Add support for ZisK backend~~ feat: add support for ZisK backend Nov 20, 2025

xqft marked this pull request as ready for review November 20, 2025 16:51

Copilot AI review requested due to automatic review settings November 20, 2025 16:51

Copilot started reviewing on behalf of xqft November 20, 2025 16:51 View session

Copilot finished reviewing on behalf of xqft November 20, 2025 16:52

Copilot AI reviewed Nov 20, 2025

View reviewed changes

src/cli.rs Show resolved Hide resolved

diff Outdated Show resolved Hide resolved

src/cli.rs Show resolved Hide resolved

ilitteri added 3 commits November 20, 2025 15:35

Delete .github/.DS_Store

af3d22b

Delete src/.DS_Store

f78b4ee

Delete .DS_Store

5bea82f

ilitteri changed the title ~~feat: add support for ZisK backend~~ feat(l1): add support for ZisK backend Nov 20, 2025

github-actions bot added the L1 label Nov 20, 2025

ilitteri and others added 10 commits November 21, 2025 01:30

update lock

a1345f1

Add binary inputs generation command

e5b1907

remove file

1d7da93

fix and add zisk job

af6951c

Update Cargo.lock

00be59f

fix locks

be72e14

fix zisk toml

1d22f34

update ethrewx

aea5c39

fmt

9a7bdb0

clippy

028d29f

gianbelinche approved these changes Nov 25, 2025

View reviewed changes

xqft added 3 commits November 25, 2025 18:18

cargo update

45f4c5a

added ci to lint

6982de5

add ci flag

36b4751

JereSalo reviewed Nov 26, 2025

View reviewed changes

JereSalo requested changes Nov 26, 2025

View reviewed changes

xqft added 5 commits November 26, 2025 11:06

update branch

f6607fb

fix parent block header

79b2810

fix cache path

8c23cde

add state root to nodes hashmap

31e4181

add generate-input docs

876de05

JereSalo reviewed Nov 26, 2025

View reviewed changes

xqft and others added 5 commits November 26, 2025 11:47

fix generate-input

c21d52e

fix run-tx

642a81d

fix all_nodes problem

443428c

remove imports

e156e61

fix import

5672aef

JereSalo approved these changes Nov 26, 2025

View reviewed changes

JereSalo added 2 commits November 26, 2025 13:08

add profile release with debug

68f45a1

Merge branch 'add_support_for_zisk_zkvm' of github.com:lambdaclass/et…

6f2a01c

…hrex-replay into add_support_for_zisk_zkvm

ilitteri enabled auto-merge November 26, 2025 16:13

fix debug compilation

9f5c939

ilitteri added this pull request to the merge queue Nov 26, 2025

Merged via the queue into main with commit 6ea9480 Nov 26, 2025
17 checks passed

ilitteri deleted the add_support_for_zisk_zkvm branch November 26, 2025 16:50

JereSalo mentioned this pull request Nov 26, 2025

fix(l1, l2): add release with debug compilation and also fix compilation issues with recent ethrex #44

Closed

feat(l1): add support for ZisK backend #45

feat(l1): add support for ZisK backend #45

Uh oh!

Conversation

xqft commented Nov 20, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Key Changes

Reviewed Changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

JereSalo Nov 26, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

JereSalo left a comment

Choose a reason for hiding this comment

Uh oh!

xqft commented Nov 26, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

JereSalo left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

xqft commented Nov 20, 2025 •

edited

Loading

JereSalo Nov 26, 2025 •

edited

Loading