Compare commits

...

4 Commits

Author SHA1 Message Date
Omar Abdulla 425e970eb1 Make the all field of Selection public 2025-07-25 14:54:16 +03:00
dependabot[bot] 56aadce0a9 Bump @eslint/plugin-kit from 0.3.1 to 0.3.3 (#363)
Bumps
[@eslint/plugin-kit](https://github.com/eslint/rewrite/tree/HEAD/packages/plugin-kit)
from 0.3.1 to 0.3.3.
<details>
<summary>Release notes</summary>
<p><em>Sourced from <a
href="https://github.com/eslint/rewrite/releases"><code>@​eslint/plugin-kit</code>'s
releases</a>.</em></p>
<blockquote>
<h2>plugin-kit: v0.3.3</h2>
<h2><a
href="https://github.com/eslint/rewrite/compare/plugin-kit-v0.3.2...plugin-kit-v0.3.3">0.3.3</a>
(2025-06-25)</h2>
<h3>Dependencies</h3>
<ul>
<li>The following workspace dependencies were updated
<ul>
<li>dependencies
<ul>
<li><code>@​eslint/core</code> bumped from ^0.15.0 to ^0.15.1</li>
</ul>
</li>
</ul>
</li>
</ul>
<h2>plugin-kit: v0.3.2</h2>
<h2><a
href="https://github.com/eslint/rewrite/compare/plugin-kit-v0.3.1...plugin-kit-v0.3.2">0.3.2</a>
(2025-06-09)</h2>
<h3>Dependencies</h3>
<ul>
<li>The following workspace dependencies were updated
<ul>
<li>dependencies
<ul>
<li><code>@​eslint/core</code> bumped from ^0.14.0 to ^0.15.0</li>
</ul>
</li>
</ul>
</li>
</ul>
</blockquote>
</details>
<details>
<summary>Changelog</summary>
<p><em>Sourced from <a
href="https://github.com/eslint/rewrite/blob/main/packages/plugin-kit/CHANGELOG.md"><code>@​eslint/plugin-kit</code>'s
changelog</a>.</em></p>
<blockquote>
<h2><a
href="https://github.com/eslint/rewrite/compare/plugin-kit-v0.3.2...plugin-kit-v0.3.3">0.3.3</a>
(2025-06-25)</h2>
<h3>Dependencies</h3>
<ul>
<li>The following workspace dependencies were updated
<ul>
<li>dependencies
<ul>
<li><code>@​eslint/core</code> bumped from ^0.15.0 to ^0.15.1</li>
</ul>
</li>
</ul>
</li>
</ul>
<h2><a
href="https://github.com/eslint/rewrite/compare/plugin-kit-v0.3.1...plugin-kit-v0.3.2">0.3.2</a>
(2025-06-09)</h2>
<h3>Dependencies</h3>
<ul>
<li>The following workspace dependencies were updated
<ul>
<li>dependencies
<ul>
<li><code>@​eslint/core</code> bumped from ^0.14.0 to ^0.15.0</li>
</ul>
</li>
</ul>
</li>
</ul>
</blockquote>
</details>
<details>
<summary>Commits</summary>
<ul>
<li><a
href="https://github.com/eslint/rewrite/commit/0496201974aad87fdcf3aa2a63ec74e91b54825e"><code>0496201</code></a>
chore: release main (<a
href="https://github.com/eslint/rewrite/tree/HEAD/packages/plugin-kit/issues/229">#229</a>)</li>
<li><a
href="https://github.com/eslint/rewrite/commit/f5e6d683ee00b24b98777291c0a9a83719fe3402"><code>f5e6d68</code></a>
chore: hoist cli tools to root level (<a
href="https://github.com/eslint/rewrite/tree/HEAD/packages/plugin-kit/issues/224">#224</a>)</li>
<li><a
href="https://github.com/eslint/rewrite/commit/48b1f849476582257e1b6a110c4af55adbbec2e8"><code>48b1f84</code></a>
chore: release main (<a
href="https://github.com/eslint/rewrite/tree/HEAD/packages/plugin-kit/issues/216">#216</a>)</li>
<li><a
href="https://github.com/eslint/rewrite/commit/147afec6be22b6ed4151c1e0a8fc40c061d626d6"><code>147afec</code></a>
chore: update <code>package.json</code> to follow template (<a
href="https://github.com/eslint/rewrite/tree/HEAD/packages/plugin-kit/issues/225">#225</a>)</li>
<li><a
href="https://github.com/eslint/rewrite/commit/0a6aad0a70cc7261a303df614a4466e0b1f525b8"><code>0a6aad0</code></a>
docs: fix bun command and update documentation (<a
href="https://github.com/eslint/rewrite/tree/HEAD/packages/plugin-kit/issues/222">#222</a>)</li>
<li><a
href="https://github.com/eslint/rewrite/commit/2a8913cdae0aa57dfea993256dbe6a04029909da"><code>2a8913c</code></a>
docs: Update README sponsors</li>
<li><a
href="https://github.com/eslint/rewrite/commit/cb858ffb8b77ea76187a857546c7a838a1fb4881"><code>cb858ff</code></a>
refactor: update and fix error types in JSDoc comments (<a
href="https://github.com/eslint/rewrite/tree/HEAD/packages/plugin-kit/issues/213">#213</a>)</li>
<li><a
href="https://github.com/eslint/rewrite/commit/4ec089e5b60d64f695d09a973dbc4eb72026f112"><code>4ec089e</code></a>
docs: Update README sponsors</li>
<li><a
href="https://github.com/eslint/rewrite/commit/1ddb2ed129e2999c2164f933282d8df1884bbd6c"><code>1ddb2ed</code></a>
docs: Update README sponsors</li>
<li><a
href="https://github.com/eslint/rewrite/commit/aa6a48b789a5a48eb405c03502a71e5879905876"><code>aa6a48b</code></a>
docs: Update README sponsors</li>
<li>See full diff in <a
href="https://github.com/eslint/rewrite/commits/plugin-kit-v0.3.3/packages/plugin-kit">compare
view</a></li>
</ul>
</details>
<br />


[![Dependabot compatibility
score](https://dependabot-badges.githubapp.com/badges/compatibility_score?dependency-name=@eslint/plugin-kit&package-manager=npm_and_yarn&previous-version=0.3.1&new-version=0.3.3)](https://docs.github.com/en/github/managing-security-vulnerabilities/about-dependabot-security-updates#about-compatibility-scores)

Dependabot will resolve any conflicts with this PR as long as you don't
alter it yourself. You can also trigger a rebase manually by commenting
`@dependabot rebase`.

[//]: # (dependabot-automerge-start)
[//]: # (dependabot-automerge-end)

---

<details>
<summary>Dependabot commands and options</summary>
<br />

You can trigger Dependabot actions by commenting on this PR:
- `@dependabot rebase` will rebase this PR
- `@dependabot recreate` will recreate this PR, overwriting any edits
that have been made to it
- `@dependabot merge` will merge this PR after your CI passes on it
- `@dependabot squash and merge` will squash and merge this PR after
your CI passes on it
- `@dependabot cancel merge` will cancel a previously requested merge
and block automerging
- `@dependabot reopen` will reopen this PR if it is closed
- `@dependabot close` will close this PR and stop Dependabot recreating
it. You can achieve the same result by closing it manually
- `@dependabot show <dependency name> ignore conditions` will show all
of the ignore conditions of the specified dependency
- `@dependabot ignore this major version` will close this PR and stop
Dependabot creating any more for this major version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this minor version` will close this PR and stop
Dependabot creating any more for this minor version (unless you reopen
the PR or upgrade to it yourself)
- `@dependabot ignore this dependency` will close this PR and stop
Dependabot creating any more for this dependency (unless you reopen the
PR or upgrade to it yourself)
You can disable automated security fix PRs for this repo from the
[Security Alerts
page](https://github.com/paritytech/revive/network/alerts).

</details>

Signed-off-by: dependabot[bot] <support@github.com>
Co-authored-by: dependabot[bot] <49699333+dependabot[bot]@users.noreply.github.com>
Co-authored-by: xermicus <cyrill@parity.io>
2025-07-22 18:45:52 +02:00
xermicus 9751481f6b the revive-explorer utility (#364)
A maintainable and more precise version of what was a hacky but useful
script, exploring the compilers YUL lowering unit.

It analyzes a given shared objects from the debug dump and outputs:
- The count of each YUL statement translated.
- A per YUL statement break-down of bytecode size contributed per.
- Estimated `yul-phaser` cost parameters.

Signed-off-by: Cyrill Leutwiler <bigcyrill@hotmail.com>
2025-07-22 09:17:55 +02:00
xermicus c285a6ec3d add columns to debug information (#362)
- Add column numbers to debug information.
- Do not build allocas at entry for now.

---------

Signed-off-by: Cyrill Leutwiler <bigcyrill@hotmail.com>
2025-07-17 09:53:54 +02:00
31 changed files with 801 additions and 40 deletions
+2
View File
@@ -14,7 +14,9 @@ Supported `polkadot-sdk` rev: `2503.0.1`
### Added
- Line debug information per YUL builtin and for `if` statements.
- Column numbers in debug information.
- Support for the YUL optimizer details in the standard json input definition.
- The `revive-explorer` compiler utility.
### Fixed
- The debug info source file matches the YUL path in `--debug-output-dir`, allowing tools to display the source line.
Generated
+10
View File
@@ -8615,6 +8615,16 @@ dependencies = [
"tempfile",
]
[[package]]
name = "revive-explorer"
version = "0.1.0"
dependencies = [
"anyhow",
"clap",
"num_cpus",
"revive-yul",
]
[[package]]
name = "revive-integration"
version = "0.1.1"
+1
View File
@@ -19,6 +19,7 @@ revive-benchmarks = { version = "0.1.0", path = "crates/benchmarks" }
revive-builtins = { version = "0.1.0", path = "crates/builtins" }
revive-common = { version = "0.1.0", path = "crates/common" }
revive-differential = { version = "0.1.0", path = "crates/differential" }
revive-explorer = { version = "0.1.0", path = "crates/explore" }
revive-integration = { version = "0.1.1", path = "crates/integration" }
revive-linker = { version = "0.1.0", path = "crates/linker" }
lld-sys = { version = "0.1.0", path = "crates/lld-sys" }
+5 -1
View File
@@ -6,6 +6,7 @@
install-llvm-builder \
install-llvm \
install-revive-runner \
install-revive-explorer \
format \
clippy \
machete \
@@ -43,6 +44,9 @@ install-llvm: install-llvm-builder
install-revive-runner:
cargo install --locked --force --path crates/runner --no-default-features
install-revive-explorer:
cargo install --locked --force --path crates/explorer --no-default-features
format:
cargo fmt --all --check
@@ -53,7 +57,7 @@ machete:
cargo install cargo-machete
cargo machete
test: format clippy machete test-cli test-workspace install-revive-runner
test: format clippy machete test-cli test-workspace install-revive-runner install-revive-explorer
test-integration: install-bin
cargo test --package revive-integration
+20
View File
@@ -0,0 +1,20 @@
[package]
name = "revive-explorer"
version.workspace = true
license.workspace = true
edition.workspace = true
repository.workspace = true
authors.workspace = true
description = "Helper utility to inspect debug builds"
[[bin]]
name = "revive-explorer"
path = "src/main.rs"
[dependencies]
anyhow = { workspace = true }
clap = { workspace = true, features = ["help", "std", "derive"] }
num_cpus = { workspace = true }
revive-yul = { workspace = true }
+49
View File
@@ -0,0 +1,49 @@
# revive-explorer
The `revive-explorer` is a helper utility for exploring the compilers YUL lowering unit.
It analyzes a given shared objects from the debug dump and outputs:
- The count of each YUL statement translated.
- A per YUL statement break-down of bytecode size contributed per.
- Estimated `yul-phaser` cost parameters.
Example:
```
statements count:
block 532
Caller 20
Not 73
Gas 24
Shr 2
...
Shl 259
SetImmutable 2
CodeSize 1
CallDataLoad 87
Return 56
bytes per statement:
Or 756
CodeCopy 158
Log3 620
Return 1562
MStore 36128
...
ReturnDataCopy 2854
DataOffset 28
assignment 1194
Number 540
CallValue 4258
yul-phaser parameters:
--break-cost 1
--variable-declaration-cost 3
--function-call-cost 8
--if-cost 4
--expression-statement-cost 6
--function-definition-cost 11
--switch-cost 3
--block-cost 1
--leave-cost 1
--assignment-cost 1
```
+59
View File
@@ -0,0 +1,59 @@
//! The `llvm-dwarfdump` utility helper library.
use std::{
path::{Path, PathBuf},
process::{Command, Stdio},
};
pub static EXECUTABLE: &str = "llvm-dwarfdump";
pub static DEBUG_LINES_ARGUMENTS: [&str; 1] = ["--debug-line"];
pub static SOURCE_FILE_ARGUMENTS: [&str; 1] = ["--show-sources"];
/// Calls the `llvm-dwarfdump` tool to extract debug line information
/// from the shared object at `path`. Returns the output.
///
/// Provide `Some(dwarfdump_exectuable)` to override the default executable.
pub fn debug_lines(
shared_object: &Path,
dwarfdump_executable: &Option<PathBuf>,
) -> anyhow::Result<String> {
dwarfdump(shared_object, dwarfdump_executable, &DEBUG_LINES_ARGUMENTS)
}
/// Calls the `llvm-dwarfdump` tool to extract the source file name.
/// Returns the source file path.
///
/// Provide `Some(dwarfdump_exectuable)` to override the default executable.
pub fn source_file(
shared_object: &Path,
dwarfdump_executable: &Option<PathBuf>,
) -> anyhow::Result<PathBuf> {
let output = dwarfdump(shared_object, dwarfdump_executable, &SOURCE_FILE_ARGUMENTS)?;
Ok(output.trim().into())
}
/// The internal `llvm-dwarfdump` helper function.
fn dwarfdump(
shared_object: &Path,
dwarfdump_executable: &Option<PathBuf>,
arguments: &[&str],
) -> anyhow::Result<String> {
let executable = dwarfdump_executable
.to_owned()
.unwrap_or_else(|| PathBuf::from(EXECUTABLE));
let output = Command::new(executable)
.args(arguments)
.arg(shared_object)
.stdin(Stdio::null())
.stdout(Stdio::piped())
.stderr(Stdio::piped())
.spawn()?
.wait_with_output()?;
if !output.status.success() {
anyhow::bail!(String::from_utf8_lossy(&output.stderr).to_string());
}
Ok(String::from_utf8_lossy(&output.stdout).to_string())
}
+250
View File
@@ -0,0 +1,250 @@
//! The core dwarf dump analyzer library.
use std::{
collections::HashMap,
path::{Path, PathBuf},
};
use revive_yul::lexer::token::location::Location;
use crate::location_mapper::{self, map_locations, LocationMap};
/// Unknwon code.
pub const OTHER: &str = "other";
/// Compiler internal code.
pub const INTERNAL: &str = "internal";
/// YUL block code.
pub const BLOCK: &str = "block";
/// YUL function call code.
pub const FUNCTION_CALL: &str = "function_call";
/// YUL conditional code.
pub const IF: &str = "if";
/// YUL loop code.
pub const FOR: &str = "for";
/// YUL loop continue code.
pub const CONTINUE: &str = "continue";
/// YUL loop break code.
pub const BREAK: &str = "break";
/// YUL switch code.
pub const SWITCH: &str = "switch";
/// YUL variable declaration code.
pub const DECLARATION: &str = "let";
/// YUL variable assignment code.
pub const ASSIGNMENT: &str = "assignment";
/// YUL function definition code.
pub const FUNCTION_DEFINITION: &str = "function_definition";
/// YUL function leave code.
pub const LEAVE: &str = "leave";
/// The dwarf dump analyzer.
///
/// Loads debug information from `llvm-dwarfdump` and calculates statistics
/// about the compiled YUL statements:
/// - Statements count
/// - Per-statement
#[derive(Debug, Default)]
pub struct DwarfdumpAnalyzer {
/// The YUL source file path.
source: PathBuf,
/// The YUL location to statements map.
location_map: LocationMap,
/// The `llvm-dwarfdump --debug-lines` output.
debug_lines: String,
/// The observed statements.
statements_count: HashMap<String, usize>,
/// The observed statement to instructions size.
statements_size: HashMap<String, u64>,
}
impl DwarfdumpAnalyzer {
/// The debug info analyzer constructor.
///
/// `source` is the path to the YUL source file.
/// `debug_lines` is the `llvm-dwarfdump --debug-lines` output.
pub fn new(source: &Path, debug_lines: String) -> Self {
Self {
source: source.to_path_buf(),
debug_lines,
..Default::default()
}
}
/// Run the analysis.
pub fn analyze(&mut self) -> anyhow::Result<()> {
self.map_locations()?;
self.analyze_statements()?;
Ok(())
}
/// Populate the maps so that we can always unwrap later.
fn map_locations(&mut self) -> anyhow::Result<()> {
self.location_map = map_locations(&self.source)?;
self.statements_count = HashMap::with_capacity(self.location_map.len());
self.statements_size = HashMap::with_capacity(self.location_map.len());
for statement in self.location_map.values() {
if !self.statements_size.contains_key(statement) {
self.statements_size.insert(statement.clone(), 0);
}
*self.statements_count.entry(statement.clone()).or_insert(0) += 1;
}
Ok(())
}
/// Analyze how much bytes of insturctions each statement contributes.
fn analyze_statements(&mut self) -> anyhow::Result<()> {
let mut previous_offset = 0;
let mut previous_location = Location::new(0, 0);
for line in self
.debug_lines
.lines()
.skip_while(|line| !line.starts_with("Address"))
.skip(2)
{
let mut parts = line.split_whitespace();
let (Some(offset), Some(line), Some(column)) =
(parts.next(), parts.next(), parts.next())
else {
continue;
};
let current_offset = u64::from_str_radix(offset.trim_start_matches("0x"), 16)?;
let mut current_location = Location::new(line.parse()?, column.parse()?);
// TODO: A bug? Needs further investigation.
if current_location.line == 0 && current_location.column != 0 {
current_location.line = previous_location.line;
}
if let Some(statement) = self.location_map.get(&previous_location) {
let contribution = current_offset - previous_offset;
*self.statements_size.get_mut(statement).unwrap() += contribution;
}
previous_offset = current_offset;
previous_location = current_location;
}
Ok(())
}
/// Print the per-statement count break-down.
pub fn display_statement_count(&self) {
println!("statements count:");
for (statement, count) in self.statements_count.iter() {
println!("\t{statement} {count}");
}
}
/// Print the per-statement byte size contribution break-down.
pub fn display_statement_size(&self) {
println!("bytes per statement:");
for (statement, size) in self.statements_size.iter() {
println!("\t{statement} {size}");
}
}
/// Print the estimated `yul-phaser` cost parameters.
pub fn display_phaser_costs(&self, yul_phaser_scale: u64) {
println!("yul-phaser parameters:");
for (parameter, cost) in self.phaser_costs(yul_phaser_scale) {
println!("\t{parameter} {cost}");
}
}
/// Estimate the `yul-phaser` costs using the simplified weight function:
/// `Total size / toal count = cost`
pub fn phaser_costs(&self, yul_phaser_scale: u64) -> Vec<(String, u64)> {
let mut costs: HashMap<String, (usize, u64)> = HashMap::with_capacity(16);
for (statement, count) in self
.statements_count
.iter()
.filter(|(_, count)| **count > 0)
{
let size = self.statements_size.get(statement).unwrap();
let cost = match statement.as_str() {
location_mapper::FOR => "--for-loop-cost",
location_mapper::OTHER => continue,
location_mapper::INTERNAL => continue,
location_mapper::BLOCK => "--block-cost",
location_mapper::FUNCTION_CALL => "--function-call-cost",
location_mapper::IF => "--if-cost",
location_mapper::CONTINUE => "--continue-cost",
location_mapper::BREAK => "--break-cost",
location_mapper::LEAVE => "--leave-cost",
location_mapper::SWITCH => "--switch-cost",
location_mapper::DECLARATION => "--variable-declaration-cost",
location_mapper::ASSIGNMENT => "--assignment-cost",
location_mapper::FUNCTION_DEFINITION => "--function-definition-cost",
_ => "--expression-statement-cost",
};
let entry = costs.entry(cost.to_string()).or_default();
entry.0 += count;
entry.1 += size;
}
let costs = costs
.iter()
.map(|(cost, (count, size))| {
let ratio = *size / *count as u64;
(cost.to_string(), ratio.min(100))
})
.collect::<Vec<_>>();
let scaled_costs = scale_to(
costs
.iter()
.map(|(_, ratio)| *ratio)
.collect::<Vec<_>>()
.as_slice(),
yul_phaser_scale,
);
costs
.iter()
.zip(scaled_costs)
.map(|((cost, _), scaled_ratio)| (cost.to_string(), scaled_ratio))
.collect()
}
}
/// Given a slice of u64 values, returns a Vec<u64> where each element
/// is linearly scaled into the closed interval [1, 10].
fn scale_to(data: &[u64], scale_max: u64) -> Vec<u64> {
if data.is_empty() {
return Vec::new();
}
let mut min = data[0];
let mut max = data[0];
for &x in &data[1..] {
if x < min {
min = x;
}
if x > max {
max = x;
}
}
if max < scale_max {
return data.to_vec();
}
let range = max - min;
data.iter()
.map(|&x| {
if range == 0 {
1
} else {
1 + (x - min) * scale_max / range
}
})
.collect()
}
+6
View File
@@ -0,0 +1,6 @@
//! The revive explorer leverages debug info to get insights into emitted code.
pub mod dwarfdump;
pub mod dwarfdump_analyzer;
pub mod location_mapper;
pub mod yul_phaser;
+158
View File
@@ -0,0 +1,158 @@
//! The location mapper utility maps YUL source locations to AST statements.
//!
//! TODO: Refactor when the AST visitor is implemented.
use std::{collections::HashMap, path::Path};
use revive_yul::{
lexer::{token::location::Location, Lexer},
parser::statement::{
block::Block,
expression::{function_call::name::Name, Expression},
object::Object,
Statement,
},
};
/// Code attributed to an unknown location.
pub const OTHER: &str = "other";
/// Code attributed to a compiler internal location.
pub const INTERNAL: &str = "internal";
/// Code attributed to a
pub const BLOCK: &str = "block";
pub const FUNCTION_CALL: &str = "function_call";
pub const FOR: &str = "for";
pub const IF: &str = "if";
pub const CONTINUE: &str = "continue";
pub const BREAK: &str = "break";
pub const LEAVE: &str = "leave";
pub const SWITCH: &str = "switch";
pub const DECLARATION: &str = "let";
pub const ASSIGNMENT: &str = "assignment";
pub const FUNCTION_DEFINITION: &str = "function_definition";
/// The location to statements map type alias.
pub type LocationMap = HashMap<Location, String>;
/// Construct a [LocationMap] from the given YUL `source` file.
pub fn map_locations(source: &Path) -> anyhow::Result<LocationMap> {
let mut lexer = Lexer::new(std::fs::read_to_string(source)?);
let ast = Object::parse(&mut lexer, None).map_err(|error| {
anyhow::anyhow!("Contract `{}` parsing error: {:?}", source.display(), error)
})?;
let mut location_map = HashMap::with_capacity(1024);
crate::location_mapper::object_mapper(&mut location_map, &ast);
location_map.insert(Location::new(0, 0), OTHER.to_string());
location_map.insert(Location::new(1, 0), INTERNAL.to_string());
Ok(location_map)
}
/// Map the [Block].
fn block_mapper(map: &mut LocationMap, block: &Block) {
map.insert(block.location, BLOCK.to_string());
for statement in &block.statements {
statement_mapper(map, statement);
}
}
/// Map the [Expression].
fn expression_mapper(map: &mut LocationMap, expression: &Expression) {
if let Expression::FunctionCall(call) = expression {
let id = match call.name {
Name::UserDefined(_) => FUNCTION_CALL.to_string(),
_ => format!("{:?}", call.name),
};
map.insert(expression.location(), id);
for expression in &call.arguments {
expression_mapper(map, expression);
}
}
}
/// Map the [Statement].
fn statement_mapper(map: &mut LocationMap, statement: &Statement) {
match statement {
Statement::Object(object) => object_mapper(map, object),
Statement::Code(code) => block_mapper(map, &code.block),
Statement::Block(block) => block_mapper(map, block),
Statement::ForLoop(for_loop) => {
map.insert(for_loop.location, FOR.to_string());
expression_mapper(map, &for_loop.condition);
block_mapper(map, &for_loop.body);
block_mapper(map, &for_loop.initializer);
block_mapper(map, &for_loop.finalizer);
}
Statement::IfConditional(if_conditional) => {
map.insert(if_conditional.location, IF.to_string());
expression_mapper(map, &if_conditional.condition);
block_mapper(map, &if_conditional.block);
}
Statement::Expression(expression) => expression_mapper(map, expression),
Statement::Continue(location) => {
map.insert(*location, CONTINUE.to_string());
}
Statement::Leave(location) => {
map.insert(*location, LEAVE.to_string());
}
Statement::Break(location) => {
map.insert(*location, BREAK.to_string());
}
Statement::Switch(switch) => {
map.insert(switch.expression.location(), SWITCH.to_string());
expression_mapper(map, &switch.expression);
for case in &switch.cases {
block_mapper(map, &case.block);
}
if let Some(block) = switch.default.as_ref() {
block_mapper(map, block);
}
}
Statement::Assignment(assignment) => {
map.insert(assignment.location, ASSIGNMENT.to_string());
expression_mapper(map, &assignment.initializer);
}
Statement::VariableDeclaration(declaration) => {
map.insert(declaration.location, DECLARATION.to_string());
if let Some(expression) = declaration.expression.as_ref() {
expression_mapper(map, expression);
}
}
Statement::FunctionDefinition(definition) => {
map.insert(definition.location, FUNCTION_DEFINITION.to_string());
block_mapper(map, &definition.body);
}
}
}
/// Map the [Object].
fn object_mapper(map: &mut LocationMap, object: &Object) {
map.insert(object.location, object.identifier.clone());
block_mapper(map, &object.code.block);
if let Some(object) = object.inner_object.as_ref() {
object_mapper(map, object);
}
}
+56
View File
@@ -0,0 +1,56 @@
use std::path::PathBuf;
use clap::Parser;
use revive_explorer::{dwarfdump, dwarfdump_analyzer::DwarfdumpAnalyzer, yul_phaser};
/// The `revive-explorer` is a helper utility for exploring the compilers YUL lowering unit.
///
/// It analyzes a given shared objects from the debug dump and outputs:
/// - The count of each YUL statement translated.
/// - A per YUL statement break-down of bytecode size contributed per.
/// - Estimated `yul-phaser` cost parameters.
#[derive(Parser, Debug)]
#[command(version, about, long_about = None)]
struct Args {
/// Path of the dwarfdump executable.
#[arg(short, long)]
dwarfdump: Option<PathBuf>,
/// The YUL phaser cost scale maximum value.
#[arg(short, long, default_value_t = 10)]
cost_scale: u64,
/// Run the provided yul-phaser executable using the estimated costs.
#[arg(short, long)]
yul_phaser: Option<PathBuf>,
/// Path of the shared object to analyze.
file: PathBuf,
}
fn main() -> anyhow::Result<()> {
let args = Args::parse();
let source_file = dwarfdump::source_file(&args.file, &args.dwarfdump)?;
let debug_lines = dwarfdump::debug_lines(&args.file, &args.dwarfdump)?;
let mut analyzer = DwarfdumpAnalyzer::new(source_file.as_path(), debug_lines);
analyzer.analyze()?;
if let Some(path) = args.yul_phaser.as_ref() {
yul_phaser::run(
path,
source_file.as_path(),
analyzer.phaser_costs(args.cost_scale).as_slice(),
num_cpus::get() / 2, // TODO: should be configurable.
)?;
return Ok(());
}
analyzer.display_statement_count();
analyzer.display_statement_size();
analyzer.display_phaser_costs(args.cost_scale);
Ok(())
}
+79
View File
@@ -0,0 +1,79 @@
//! The revive explorer YUL phaser utility library.
//!
//! This can be used to invoke the `yul-phaser` utility,
//! used to find better YUL optimizer sequences.
use std::{
path::{Path, PathBuf},
process::{Command, Stdio},
thread,
time::{SystemTime, UNIX_EPOCH},
};
/// The `yul-phaser` sane default arguments:
/// - Less verbose output.
/// - Sufficient rounds.
/// - Sufficient random population start.
const ARGUMENTS: [&str; 6] = [
"--hide-round",
"--rounds",
"1000",
"--random-population",
"100",
"--show-only-top-chromosome",
];
/// Run multiple YUL phaser executables in parallel.
pub fn run(
executable: &Path,
source: &Path,
costs: &[(String, u64)],
n_threads: usize,
) -> anyhow::Result<()> {
let mut handles = Vec::with_capacity(n_threads);
for n in 0..n_threads {
let executable = executable.to_path_buf();
let source = source.to_path_buf();
let costs = costs.to_vec();
handles.push(thread::spawn(move || {
spawn_process(executable, source, costs, n)
}));
}
for handle in handles {
let _ = handle.join();
}
Ok(())
}
/// The `yul-phaser` process spawning helper function.
fn spawn_process(
executable: PathBuf,
source: PathBuf,
costs: Vec<(String, u64)>,
seed: usize,
) -> anyhow::Result<()> {
let cost_parameters = costs
.iter()
.flat_map(|(parameter, cost)| vec![parameter.clone(), cost.to_string()]);
let secs = SystemTime::now()
.duration_since(UNIX_EPOCH)
.expect("Time went backwards")
.as_secs();
Command::new(executable)
.args(cost_parameters)
.args(ARGUMENTS)
.arg("--seed")
.arg((seed + secs as usize).to_string())
.arg(source)
.stdin(Stdio::null())
.spawn()?
.wait()?;
Ok(())
}
+8 -8
View File
@@ -1,10 +1,10 @@
{
"Baseline": 939,
"Computation": 2282,
"DivisionArithmetics": 8849,
"ERC20": 18308,
"Events": 1640,
"FibonacciIterative": 1497,
"Flipper": 2099,
"SHA1": 8243
"Baseline": 960,
"Computation": 2367,
"DivisionArithmetics": 9108,
"ERC20": 17655,
"Events": 1680,
"FibonacciIterative": 1536,
"Flipper": 2137,
"SHA1": 8299
}
@@ -48,6 +48,7 @@ where
function_type,
0,
Some(inkwell::module::Linkage::External),
None,
)?;
self.inner.declare(context)
@@ -74,7 +75,6 @@ where
}
context.set_basic_block(context.current_function().borrow().return_block());
context.set_debug_location(0, 0, None)?;
context.build_return(None);
context.pop_debug_scope();
@@ -145,6 +145,7 @@ where
entry_function_type,
0,
Some(inkwell::module::Linkage::External),
None,
)?;
context.declare_global(
@@ -48,6 +48,7 @@ where
function_type,
0,
Some(inkwell::module::Linkage::External),
None,
)?;
self.inner.declare(context)
+25 -12
View File
@@ -463,6 +463,7 @@ where
r#type: inkwell::types::FunctionType<'ctx>,
return_values_length: usize,
linkage: Option<inkwell::module::Linkage>,
location: Option<(u32, u32)>,
) -> anyhow::Result<Rc<RefCell<Function<'ctx>>>> {
let value = self.module().add_function(name, r#type, linkage);
@@ -478,7 +479,8 @@ where
Some(scp) => scp,
};
self.push_debug_scope(func_scope.as_debug_info_scope());
self.set_debug_location(0, 0, Some(func_scope.as_debug_info_scope()))?;
let (line, column) = location.unwrap_or((0, 0));
self.set_debug_location(line, column, Some(func_scope.as_debug_info_scope()))?;
}
let entry_block = self.llvm.append_basic_block(value, "entry");
@@ -533,7 +535,11 @@ where
/// Sets the current active function. If debug-info generation is enabled,
/// constructs a debug-scope and pushes in on the scope-stack.
pub fn set_current_function(&mut self, name: &str, line: Option<u32>) -> anyhow::Result<()> {
pub fn set_current_function(
&mut self,
name: &str,
location: Option<(u32, u32)>,
) -> anyhow::Result<()> {
let function = self.functions.get(name).cloned().ok_or_else(|| {
anyhow::anyhow!("Failed to activate an undeclared function `{}`", name)
})?;
@@ -542,7 +548,8 @@ where
if let Some(scope) = self.current_function().borrow().get_debug_scope() {
self.push_debug_scope(scope);
}
self.set_debug_location(line.unwrap_or_default(), 0, None)?;
let (line, column) = location.unwrap_or_default();
self.set_debug_location(line, column, None)?;
Ok(())
}
@@ -723,17 +730,23 @@ where
r#type: T,
name: &str,
) -> Pointer<'ctx> {
let current_block = self.basic_block();
let entry_block = self.current_function().borrow().entry_block();
// TODO: Revisit. While at entry should be preferred in theory:
// - It has negligible code size impact on real word contracts.
// - Sometimes has negative impact on code size.
// - Messes up debug information used to analyze code size issues.
self.build_alloca(r#type, name)
match entry_block.get_first_instruction() {
Some(instruction) => self.builder().position_before(&instruction),
None => self.builder().position_at_end(entry_block),
}
// let current_block = self.basic_block();
// let entry_block = self.current_function().borrow().entry_block();
let pointer = self.build_alloca(r#type, name);
self.set_basic_block(current_block);
pointer
// match entry_block.get_first_instruction() {
// Some(instruction) => self.builder().position_before(&instruction),
// None => self.builder().position_at_end(entry_block),
// }
// let pointer = self.build_alloca(r#type, name);
// self.set_basic_block(current_block);
// pointer
}
/// Builds an aligned stack allocation at the current position.
@@ -35,6 +35,7 @@ where
Self::r#type(context),
0,
Some(inkwell::module::Linkage::External),
None,
)?;
let mut attributes = Self::ATTRIBUTES.to_vec();
@@ -40,6 +40,7 @@ pub fn check_attribute_null_pointer_is_invalid() {
.fn_type(&[context.word_type().into()], false),
1,
Some(inkwell::module::Linkage::External),
None,
)
.expect("Failed to add function");
assert!(!function
@@ -63,6 +64,7 @@ pub fn check_attribute_optimize_for_size_mode_3() {
.fn_type(&[context.word_type().into()], false),
1,
Some(inkwell::module::Linkage::External),
None,
)
.expect("Failed to add function");
assert!(!function
@@ -86,6 +88,7 @@ pub fn check_attribute_optimize_for_size_mode_z() {
.fn_type(&[context.word_type().into()], false),
1,
Some(inkwell::module::Linkage::External),
None,
)
.expect("Failed to add function");
assert!(function
@@ -109,6 +112,7 @@ pub fn check_attribute_min_size_mode_3() {
.fn_type(&[context.word_type().into()], false),
1,
Some(inkwell::module::Linkage::External),
None,
)
.expect("Failed to add function");
assert!(!function
@@ -132,6 +136,7 @@ pub fn check_attribute_min_size_mode_z() {
.fn_type(&[context.word_type().into()], false),
1,
Some(inkwell::module::Linkage::External),
None,
)
.expect("Failed to add function");
assert!(function
@@ -14,7 +14,7 @@ use self::file::File as FileSelection;
pub struct Selection {
/// Only the 'all' wildcard is available for robustness reasons.
#[serde(rename = "*", skip_serializing_if = "Option::is_none")]
all: Option<FileSelection>,
pub all: Option<FileSelection>,
#[serde(skip_serializing_if = "BTreeMap::is_empty", flatten)]
pub files: BTreeMap<String, FileSelection>,
+10
View File
@@ -1,5 +1,8 @@
//! The lexical token location.
use std::hash::Hash;
use std::hash::Hasher;
use serde::Deserialize;
use serde::Serialize;
@@ -48,6 +51,13 @@ impl PartialEq for Location {
}
}
impl Hash for Location {
fn hash<H: Hasher>(&self, state: &mut H) {
self.line.hash(state);
self.column.hash(state);
}
}
impl std::fmt::Display for Location {
fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
write!(f, "{}:{}", self.line, self.column)
@@ -119,7 +119,7 @@ where
mut self,
context: &mut revive_llvm_context::PolkaVMContext<D>,
) -> anyhow::Result<()> {
context.set_debug_location(self.location.line, 0, None)?;
context.set_debug_location(self.location.line, self.location.column, None)?;
let value = match self.initializer.into_llvm(context)? {
Some(value) => value,
@@ -149,7 +149,7 @@ where
context.build_store(tuple_pointer, value)?;
for (index, binding) in self.bindings.into_iter().enumerate() {
context.set_debug_location(self.location.line, 0, None)?;
context.set_debug_location(self.location.line, self.location.column, None)?;
let field_pointer = context.build_gep(
tuple_pointer,
+10 -3
View File
@@ -155,7 +155,10 @@ where
function.into_llvm(context)?;
}
context.set_current_function(current_function.as_str(), Some(self.location.line))?;
context.set_current_function(
current_function.as_str(),
Some((self.location.line, self.location.column)),
)?;
if let Some(debug_info) = context.debug_info() {
let di_builder = debug_info.builder();
@@ -169,12 +172,16 @@ where
)
.as_debug_info_scope();
context.push_debug_scope(di_block_scope);
context.set_debug_location(self.location.line, 0, None)?;
context.set_debug_location(self.location.line, self.location.column, None)?;
}
context.set_basic_block(current_block);
for statement in local_statements.into_iter() {
context.set_debug_location(statement.location().line, 0, None)?;
context.set_debug_location(
statement.location().line,
statement.location().column,
None,
)?;
if context.basic_block().get_terminator().is_some() {
break;
}
@@ -123,7 +123,7 @@ impl FunctionCall {
D: revive_llvm_context::PolkaVMDependency + Clone,
{
let location = self.location;
context.set_debug_location(location.line, 0, None)?;
context.set_debug_location(location.line, location.column, None)?;
match self.name {
Name::UserDefined(name) => {
@@ -997,6 +997,8 @@ impl FunctionCall {
}
arguments.reverse();
context.set_debug_location(self.location.line, self.location.column, None)?;
Ok(arguments.try_into().expect("Always successful"))
}
@@ -1014,6 +1016,8 @@ impl FunctionCall {
}
arguments.reverse();
context.set_debug_location(self.location.line, self.location.column, None)?;
Ok(arguments.try_into().expect("Always successful"))
}
}
@@ -72,6 +72,7 @@ where
let increment_block = context.append_basic_block("for_increment");
let join_block = context.append_basic_block("for_join");
context.set_debug_location(self.location.line, self.location.column, None)?;
context.build_unconditional_branch(condition_block);
context.set_basic_block(condition_block);
let condition = self
@@ -80,6 +81,7 @@ where
.expect("Always exists")
.access(context)?
.into_int_value();
context.set_debug_location(self.location.line, self.location.column, None)?;
let condition = context.builder().build_int_z_extend_or_bit_cast(
condition,
context.word_type(),
@@ -98,10 +100,12 @@ where
context.set_basic_block(body_block);
self.body.into_llvm(context)?;
context.build_unconditional_branch(increment_block);
context.set_debug_location(self.location.line, self.location.column, None)?;
context.set_basic_block(increment_block);
self.finalizer.into_llvm(context)?;
context.build_unconditional_branch(condition_block);
context.set_debug_location(self.location.line, self.location.column, None)?;
context.pop_loop();
context.set_basic_block(join_block);
@@ -196,6 +196,7 @@ where
&mut self,
context: &mut revive_llvm_context::PolkaVMContext<D>,
) -> anyhow::Result<()> {
context.set_debug_location(self.location.line, self.location.column, None)?;
let argument_types: Vec<_> = self
.arguments
.iter()
@@ -212,6 +213,7 @@ where
function_type,
self.result.len(),
Some(inkwell::module::Linkage::External),
Some((self.location.line, self.location.column)),
)?;
revive_llvm_context::PolkaVMFunction::set_attributes(
context.llvm(),
@@ -230,7 +232,10 @@ where
mut self,
context: &mut revive_llvm_context::PolkaVMContext<D>,
) -> anyhow::Result<()> {
context.set_current_function(self.identifier.as_str(), Some(self.location.line))?;
context.set_current_function(
self.identifier.as_str(),
Some((self.location.line, self.location.column)),
)?;
context.set_basic_block(context.current_function().borrow().entry_block());
let r#return = context.current_function().borrow().r#return();
@@ -290,7 +295,7 @@ where
}
self.body.into_llvm(context)?;
context.set_debug_location(self.location.line, 0, None)?;
context.set_debug_location(self.location.line, self.location.column, None)?;
match context
.basic_block()
@@ -53,13 +53,13 @@ where
D: revive_llvm_context::PolkaVMDependency + Clone,
{
fn into_llvm(self, context: &mut revive_llvm_context::PolkaVMContext<D>) -> anyhow::Result<()> {
context.set_debug_location(self.location.line, 0, None)?;
let condition = self
.condition
.into_llvm(context)?
.expect("Always exists")
.access(context)?
.into_int_value();
context.set_debug_location(self.location.line, self.location.column, None)?;
let condition = context.builder().build_int_z_extend_or_bit_cast(
condition,
context.word_type(),
+2 -2
View File
@@ -279,17 +279,17 @@ where
context.push_debug_scope(object_scope.as_debug_info_scope());
}
context.set_debug_location(self.location.line, self.location.column, None)?;
if self.identifier.ends_with("_deployed") {
revive_llvm_context::PolkaVMRuntimeCodeFunction::new(self.code).into_llvm(context)?;
} else {
revive_llvm_context::PolkaVMDeployCodeFunction::new(self.code).into_llvm(context)?;
}
context.set_debug_location(self.location.line, 0, None)?;
if let Some(object) = self.inner_object {
object.into_llvm(context)?;
}
context.set_debug_location(self.location.line, 0, None)?;
context.set_debug_location(self.location.line, self.location.column, None)?;
context.pop_debug_scope();
@@ -123,6 +123,7 @@ where
D: revive_llvm_context::PolkaVMDependency + Clone,
{
fn into_llvm(self, context: &mut revive_llvm_context::PolkaVMContext<D>) -> anyhow::Result<()> {
context.set_debug_location(self.location.line, self.location.column, None)?;
let scrutinee = self.expression.into_llvm(context)?;
if self.cases.is_empty() {
@@ -143,6 +144,7 @@ where
.append_basic_block(format!("switch_case_branch_{}_block", index + 1).as_str());
context.set_basic_block(expression_block);
case.block.into_llvm(context)?;
context.set_debug_location(self.location.line, self.location.column, None)?;
context.build_unconditional_branch(join_block);
branches.push((constant.into_int_value(), expression_block));
@@ -159,6 +161,7 @@ where
None => join_block,
};
context.set_debug_location(self.location.line, self.location.column, None)?;
context.set_basic_block(current_block);
context.builder().build_switch(
scrutinee
@@ -169,6 +172,7 @@ where
branches.as_slice(),
)?;
context.set_debug_location(self.location.line, self.location.column, None)?;
context.set_basic_block(join_block);
Ok(())
@@ -101,7 +101,7 @@ where
) -> anyhow::Result<()> {
if self.bindings.len() == 1 {
let identifier = self.bindings.remove(0);
context.set_debug_location(self.location.line, 0, None)?;
context.set_debug_location(self.location.line, self.location.column, None)?;
let identifier_type = identifier.r#type.clone().unwrap_or_default();
let r#type = identifier_type.into_llvm(context);
let pointer = context.build_alloca(r#type, identifier.inner.as_str());
@@ -133,7 +133,7 @@ where
}
for (index, binding) in self.bindings.iter().enumerate() {
context.set_debug_location(self.location.line, 0, None)?;
context.set_debug_location(self.location.line, self.location.column, None)?;
let yul_type = binding
.r#type
+16 -4
View File
@@ -743,18 +743,30 @@
}
},
"node_modules/@eslint/plugin-kit": {
"version": "0.3.1",
"resolved": "https://registry.npmjs.org/@eslint/plugin-kit/-/plugin-kit-0.3.1.tgz",
"integrity": "sha512-0J+zgWxHN+xXONWIyPWKFMgVuJoZuGiIFu8yxk7RJjxkzpGmyja5wRFqZIVtjDVOQpV+Rw0iOAjYPE2eQyjr0w==",
"version": "0.3.3",
"resolved": "https://registry.npmjs.org/@eslint/plugin-kit/-/plugin-kit-0.3.3.tgz",
"integrity": "sha512-1+WqvgNMhmlAambTvT3KPtCl/Ibr68VldY2XY40SL1CE0ZXiakFR/cbTspaF5HsnpDMvcYYoJHfl4980NBjGag==",
"license": "Apache-2.0",
"dependencies": {
"@eslint/core": "^0.14.0",
"@eslint/core": "^0.15.1",
"levn": "^0.4.1"
},
"engines": {
"node": "^18.18.0 || ^20.9.0 || >=21.1.0"
}
},
"node_modules/@eslint/plugin-kit/node_modules/@eslint/core": {
"version": "0.15.1",
"resolved": "https://registry.npmjs.org/@eslint/core/-/core-0.15.1.tgz",
"integrity": "sha512-bkOp+iumZCCbt1K1CmWf0R9pM5yKpDv+ZXtvSyQpudrI9kuFLp+bM2WOPXImuD/ceQuaa8f5pj93Y7zyECIGNA==",
"license": "Apache-2.0",
"dependencies": {
"@types/json-schema": "^7.0.15"
},
"engines": {
"node": "^18.18.0 || ^20.9.0 || >=21.1.0"
}
},
"node_modules/@eth-optimism/contracts": {
"version": "0.5.40",
"resolved": "https://registry.npmjs.org/@eth-optimism/contracts/-/contracts-0.5.40.tgz",