emit YUL builtins debug line info and fix debug info source file (#358)

This PR fixes and enhances debug info generation:
1. Adds line information for each lowered YUL builtin and the `if`
statement.
2. Fixes the debug info source path to match the YUL file of the
contract dumped to the `--debug-output-dir`.

This improves inspection of the generated code a lot. Excerpt from
`llvm-objdump -Sl /tmp/dbg/contracts_EndpointV2.sol.EndpointV2.so`:

```
; /tmp/dbg/contracts_EndpointV2.sol.EndpointV2.yul:203
;                 let _1 := memoryguard(0x80)
   13c3e: 3aa5b023      sd      a0, 0x3a0(a1)
   13c42: 38a5bc23      sd      a0, 0x398(a1)
   13c46: 38a5b823      sd      a0, 0x390(a1)
   13c4a: 08000513      li      a0, 0x80
   13c4e: 38a5b423      sd      a0, 0x388(a1)
; /tmp/dbg/contracts_EndpointV2.sol.EndpointV2.yul:204
;                 mstore(64, _1)
   13c52: 3885b503      ld      a0, 0x388(a1)
   13c56: 3905b603      ld      a2, 0x390(a1)
   13c5a: 3985b683      ld      a3, 0x398(a1)
   13c5e: 3a05b703      ld      a4, 0x3a0(a1)
   13c62: 38e5b023      sd      a4, 0x380(a1)
   13c66: 36d5bc23      sd      a3, 0x378(a1)
   13c6a: 36c5b823      sd      a2, 0x370(a1)
   13c6e: 36a5b423      sd      a0, 0x368(a1)
   13c72: 04000513      li      a0, 0x40
   13c76: 65d9          lui     a1, 0x16
   13c78: a605859b      addiw   a1, a1, -0x5a0
   13c7c: 95a6          add     a1, a1, s1
   13c7e: 00000097      auipc   ra, 0x0
   13c82: 000080e7      jalr    ra <__runtime+0x1de>

0000000000013c86 <.Lpcrel_hi27>:
; /tmp/dbg/contracts_EndpointV2.sol.EndpointV2.yul:205
;                 if iszero(lt(calldatasize(), 4))
   13c86: 00000517      auipc   a0, 0x0
   13c8a: 00053503      ld      a0, 0x0(a0)
   13c8e: 4108          lw      a0, 0x0(a0)
   13c90: 4591          li      a1, 0x4
   13c92: 06b56263      bltu    a0, a1, 0x13cf6 <.Lpcrel_hi27+0x70>
   13c96: a009          j       0x13c98 <.Lpcrel_hi27+0x12>
```

---------

Signed-off-by: Cyrill Leutwiler <bigcyrill@hotmail.com>

CHANGELOG.md

@@ -6,6 +6,18 @@ This is a development pre-release.
 
 Supported `polkadot-sdk` rev: `2503.0.1`
 
+## v0.4.0
+
+This is a development pre-release.
+
+Supported `polkadot-sdk` rev: `2503.0.1`
+
+### Added
+- Line debug information per YUL builtin and for `if` statements.
+
+### Fixed
+- The debug info source file matches the YUL path in `--debug-output-dir`, allowing tools to display the source line. 
+
 ## v0.3.0
 
 This is a development pre-release.

crates/llvm-context/src/debug_config/mod.rs

@@ -2,6 +2,7 @@
 
 pub mod ir_type;
 
+use std::path::Path;
 use std::path::PathBuf;
 
 use serde::Deserialize;
@@ -16,6 +17,14 @@ pub struct DebugConfig {
     pub output_directory: Option<PathBuf>,
     /// Whether debug info should be emitted.
     pub emit_debug_info: bool,
+    /// The YUL debug output file path.
+    ///
+    /// Is expected to be configured when running in YUL mode.
+    pub contract_path: Option<PathBuf>,
+    /// The YUL input file path.
+    ///
+    /// Is expected to be configured when not running in YUL mode.
+    pub yul_path: Option<PathBuf>,
 }
 
 impl DebugConfig {
@@ -24,15 +33,41 @@ impl DebugConfig {
         Self {
             output_directory,
             emit_debug_info,
+            contract_path: None,
+            yul_path: None,
         }
     }
 
+    /// Set the current YUL path.
+    pub fn set_yul_path(&mut self, yul_path: &Path) {
+        self.yul_path = yul_path.to_path_buf().into();
+    }
+
+    /// Set the current contract path.
+    pub fn set_contract_path(&mut self, contract_path: &str) {
+        self.contract_path = self.yul_source_path(contract_path);
+    }
+
+    /// Returns with the following precedence:
+    /// 1. The YUL source path if it was configured.
+    /// 2. The source YUL path from the debug output dir if it was configured.
+    /// 3. `None` if there is no debug output directory.
+    pub fn yul_source_path(&self, contract_path: &str) -> Option<PathBuf> {
+        if let Some(path) = self.yul_path.as_ref() {
+            return Some(path.clone());
+        }
+
+        self.output_directory.as_ref().map(|output_directory| {
+            let mut file_path = output_directory.to_owned();
+            let full_file_name = Self::full_file_name(contract_path, None, IRType::Yul);
+            file_path.push(full_file_name);
+            file_path
+        })
+    }
+
     /// Dumps the Yul IR.
     pub fn dump_yul(&self, contract_path: &str, code: &str) -> anyhow::Result<()> {
-        if let Some(output_directory) = self.output_directory.as_ref() {
-            let mut file_path = output_directory.to_owned();
-            let full_file_name = Self::full_file_name(contract_path, None, IRType::Yul);
-            file_path.push(full_file_name);
+        if let Some(file_path) = self.yul_source_path(contract_path) {
             std::fs::write(file_path, code)?;
         }
 

crates/llvm-context/src/polkavm/context/debug_info.rs

@@ -51,11 +51,20 @@ pub struct DebugInfo<'ctx> {
 
 impl<'ctx> DebugInfo<'ctx> {
     /// A shortcut constructor.
-    pub fn new(module: &inkwell::module::Module<'ctx>) -> Self {
+    pub fn new(
+        module: &inkwell::module::Module<'ctx>,
+        debug_config: &crate::debug_config::DebugConfig,
+    ) -> Self {
+        let module_name = module.get_name().to_string_lossy();
+        let yul_name = debug_config
+            .contract_path
+            .as_ref()
+            .map(|path| path.display().to_string());
+
         let (builder, compile_unit) = module.create_debug_info_builder(
             true,
             inkwell::debug_info::DWARFSourceLanguage::C,
-            module.get_name().to_string_lossy().as_ref(),
+            yul_name.as_deref().unwrap_or_else(|| module_name.as_ref()),
             "",
             "",
             false,

crates/llvm-context/src/polkavm/context/mod.rs

@@ -247,7 +247,7 @@ where
         let intrinsics = Intrinsics::new(llvm, &module);
         let llvm_runtime = LLVMRuntime::new(llvm, &module, &optimizer);
         let debug_info = debug_config.emit_debug_info.then(|| {
-            let debug_info = DebugInfo::new(&module);
+            let debug_info = DebugInfo::new(&module, &debug_config);
             debug_info.initialize_module(llvm, &module);
             debug_info
         });

crates/resolc/src/lib.rs

@@ -54,7 +54,7 @@ pub fn yul<T: Compiler>(
     solc: &mut T,
     optimizer_settings: revive_llvm_context::OptimizerSettings,
     include_metadata_hash: bool,
-    debug_config: revive_llvm_context::DebugConfig,
+    mut debug_config: revive_llvm_context::DebugConfig,
     llvm_arguments: &[String],
     memory_config: SolcStandardJsonInputSettingsPolkaVMMemory,
 ) -> anyhow::Result<Build> {
@@ -77,6 +77,7 @@ pub fn yul<T: Compiler>(
     let solc_validator = Some(&*solc);
     let project = Project::try_from_yul_path(path, solc_validator)?;
 
+    debug_config.set_yul_path(path);
     let build = project.compile(
         optimizer_settings,
         include_metadata_hash,

crates/resolc/src/project/contract/mod.rs

@@ -77,7 +77,7 @@ impl Contract {
         project: Project,
         optimizer_settings: revive_llvm_context::OptimizerSettings,
         include_metadata_hash: bool,
-        debug_config: revive_llvm_context::DebugConfig,
+        mut debug_config: revive_llvm_context::DebugConfig,
         llvm_arguments: &[String],
         memory_config: SolcStandardJsonInputSettingsPolkaVMMemory,
     ) -> anyhow::Result<ContractBuild> {
@@ -117,6 +117,7 @@ impl Contract {
             _ => llvm.create_module(self.path.as_str()),
         };
 
+        debug_config.set_contract_path(&self.path);
         let mut context = revive_llvm_context::PolkaVMContext::new(
             &llvm,
             module,

crates/yul/src/parser/statement/expression/function_call/mod.rs

@@ -123,6 +123,7 @@ impl FunctionCall {
         D: revive_llvm_context::PolkaVMDependency + Clone,
     {
         let location = self.location;
+        context.set_debug_location(location.line, 0, None)?;
 
         match self.name {
             Name::UserDefined(name) => {

crates/yul/src/parser/statement/if_conditional.rs

@@ -53,6 +53,7 @@ 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, 0, None)?;
         let condition = self
             .condition
             .into_llvm(context)?