Add DWARF 2+ line info support

Resolves #37

Cargo.lock

@@ -1483,6 +1483,12 @@ dependencies = [
  "libc",
 ]
 
+[[package]]
+name = "fallible-iterator"
+version = "0.3.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "2acce4a10f12dc2fb14a218589d4f1f62ef011b2d0cc4b3cb1bba8e94da14649"
+
 [[package]]
 name = "fastrand"
 version = "1.9.0"
@@ -1820,6 +1826,10 @@ name = "gimli"
 version = "0.28.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "4271d37baee1b8c7e4b708028c57d816cf9d2434acb33a549475f78c181f6253"
+dependencies = [
+ "fallible-iterator",
+ "stable_deref_trait",
+]
 
 [[package]]
 name = "gio-sys"
@@ -2867,6 +2877,7 @@ dependencies = [
  "flagset",
  "float-ord",
  "font-kit",
+ "gimli",
  "globset",
  "log",
  "memmap2",
@@ -3900,6 +3911,12 @@ dependencies = [
  "num-traits",
 ]
 
+[[package]]
+name = "stable_deref_trait"
+version = "1.2.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a8f112729512f8e442d81f95a8a7ddf2b7c6b8a1a6f509a95864142b30cab2d3"
+
 [[package]]
 name = "static_assertions"
 version = "1.1.0"

Cargo.toml

@@ -38,6 +38,7 @@ filetime = "0.2.23"
 flagset = "0.4.4"
 float-ord = "0.3.2"
 font-kit = "0.12.0"
+gimli = { version = "0.28.1", default-features = false, features = ["read-all"] }
 globset = { version = "0.4.14", features = ["serde1"] }
 log = "0.4.20"
 memmap2 = "0.9.3"

src/diff/code.rs

@@ -23,7 +23,7 @@ pub fn no_diff_code(
     data: &[u8],
     symbol: &mut ObjSymbol,
     relocs: &[ObjReloc],
-    line_info: &Option<BTreeMap<u32, u32>>,
+    line_info: &Option<BTreeMap<u64, u64>>,
 ) -> Result<()> {
     let code =
         &data[symbol.section_address as usize..(symbol.section_address + symbol.size) as usize];
@@ -57,8 +57,8 @@ pub fn diff_code(
     right_symbol: &mut ObjSymbol,
     left_relocs: &[ObjReloc],
     right_relocs: &[ObjReloc],
-    left_line_info: &Option<BTreeMap<u32, u32>>,
+    left_line_info: &Option<BTreeMap<u64, u64>>,
-    right_line_info: &Option<BTreeMap<u32, u32>>,
+    right_line_info: &Option<BTreeMap<u64, u64>>,
 ) -> Result<()> {
     let left_code = &left_data[left_symbol.section_address as usize
         ..(left_symbol.section_address + left_symbol.size) as usize];

src/obj/elf.rs

@@ -1,4 +1,4 @@
-use std::{collections::BTreeMap, fs, io::Cursor, path::Path};
+use std::{borrow::Cow, collections::BTreeMap, fs, io::Cursor, path::Path};
 
 use anyhow::{anyhow, bail, Context, Result};
 use byteorder::{BigEndian, ReadBytesExt};
@@ -6,8 +6,8 @@ use cwdemangle::demangle;
 use filetime::FileTime;
 use flagset::Flags;
 use object::{
-    elf, Architecture, File, Object, ObjectSection, ObjectSymbol, RelocationKind, RelocationTarget,
+    elf, Architecture, Endianness, File, Object, ObjectSection, ObjectSymbol, RelocationKind,
-    SectionIndex, SectionKind, Symbol, SymbolKind, SymbolScope, SymbolSection,
+    RelocationTarget, SectionIndex, SectionKind, Symbol, SymbolKind, SymbolScope, SymbolSection,
 };
 
 use crate::obj::{
@@ -278,7 +278,9 @@ fn relocations_by_section(
     Ok(relocations)
 }
 
-fn line_info(obj_file: &File<'_>) -> Result<Option<BTreeMap<u32, u32>>> {
+fn line_info(obj_file: &File<'_>) -> Result<Option<BTreeMap<u64, u64>>> {
+    // DWARF 1.1
+    let mut map = BTreeMap::new();
     if let Some(section) = obj_file.section_by_name(".line") {
         if section.size() == 0 {
             return Ok(None);
@@ -286,22 +288,49 @@ fn line_info(obj_file: &File<'_>) -> Result<Option<BTreeMap<u32, u32>>> {
         let data = section.uncompressed_data()?;
         let mut reader = Cursor::new(data.as_ref());
 
-        let mut map = BTreeMap::new();
         let size = reader.read_u32::<BigEndian>()?;
-        let base_address = reader.read_u32::<BigEndian>()?;
+        let base_address = reader.read_u32::<BigEndian>()? as u64;
         while reader.position() < size as u64 {
-            let line_number = reader.read_u32::<BigEndian>()?;
+            let line_number = reader.read_u32::<BigEndian>()? as u64;
             let statement_pos = reader.read_u16::<BigEndian>()?;
             if statement_pos != 0xFFFF {
                 log::warn!("Unhandled statement pos {}", statement_pos);
             }
-            let address_delta = reader.read_u32::<BigEndian>()?;
+            let address_delta = reader.read_u32::<BigEndian>()? as u64;
             map.insert(base_address + address_delta, line_number);
         }
-        // log::debug!("Line info: {map:#X?}");
-        return Ok(Some(map));
     }
-    Ok(None)
+
+    // DWARF 2+
+    let dwarf_cow = gimli::Dwarf::load(|id| {
+        Ok::<_, gimli::Error>(
+            obj_file
+                .section_by_name(id.name())
+                .and_then(|section| section.uncompressed_data().ok())
+                .unwrap_or(Cow::Borrowed(&[][..])),
+        )
+    })?;
+    let endian = match obj_file.endianness() {
+        Endianness::Little => gimli::RunTimeEndian::Little,
+        Endianness::Big => gimli::RunTimeEndian::Big,
+    };
+    let dwarf = dwarf_cow.borrow(|section| gimli::EndianSlice::new(section, endian));
+    let mut iter = dwarf.units();
+    while let Some(header) = iter.next()? {
+        let unit = dwarf.unit(header)?;
+        if let Some(program) = unit.line_program.clone() {
+            let mut rows = program.rows();
+            while let Some((_header, row)) = rows.next_row()? {
+                if let Some(line) = row.line() {
+                    map.insert(row.address(), line.get());
+                }
+            }
+        }
+    }
+    if map.is_empty() {
+        return Ok(None);
+    }
+    Ok(Some(map))
 }
 
 pub fn read(obj_path: &Path) -> Result<ObjInfo> {

src/obj/mips.rs

@@ -19,7 +19,7 @@ pub fn process_code(
     start_address: u64,
     end_address: u64,
     relocs: &[ObjReloc],
-    line_info: &Option<BTreeMap<u32, u32>>,
+    line_info: &Option<BTreeMap<u64, u64>>,
 ) -> Result<ProcessCodeResult> {
     configure_rabbitizer();
 
@@ -83,8 +83,9 @@ pub fn process_code(
                 }
             }
         }
-        let line =
+        let line = line_info
-            line_info.as_ref().and_then(|map| map.range(..=cur_addr).last().map(|(_, &b)| b));
+            .as_ref()
+            .and_then(|map| map.range(..=cur_addr as u64).last().map(|(_, &b)| b));
         insts.push(ObjIns {
             address: cur_addr,
             code,

src/obj/mod.rs

@@ -113,7 +113,7 @@ pub struct ObjIns {
     pub reloc: Option<ObjReloc>,
     pub branch_dest: Option<u32>,
     /// Line info
-    pub line: Option<u32>,
+    pub line: Option<u64>,
     /// Original (unsimplified) instruction
     pub orig: Option<String>,
 }
@@ -172,7 +172,7 @@ pub struct ObjInfo {
     pub timestamp: FileTime,
     pub sections: Vec<ObjSection>,
     pub common: Vec<ObjSymbol>,
-    pub line_info: Option<BTreeMap<u32, u32>>,
+    pub line_info: Option<BTreeMap<u64, u64>>,
 }
 #[derive(Debug, Eq, PartialEq, Copy, Clone)]
 pub enum ObjRelocKind {

src/obj/ppc.rs

@@ -24,7 +24,7 @@ pub fn process_code(
     data: &[u8],
     address: u64,
     relocs: &[ObjReloc],
-    line_info: &Option<BTreeMap<u32, u32>>,
+    line_info: &Option<BTreeMap<u64, u64>>,
 ) -> Result<ProcessCodeResult> {
     let ins_count = data.len() / 4;
     let mut ops = Vec::<u8>::with_capacity(ins_count);
@@ -82,7 +82,7 @@ pub fn process_code(
         ops.push(simplified.ins.op as u8);
         let line = line_info
             .as_ref()
-            .and_then(|map| map.range(..=simplified.ins.addr).last().map(|(_, &b)| b));
+            .and_then(|map| map.range(..=simplified.ins.addr as u64).last().map(|(_, &b)| b));
         insts.push(ObjIns {
             address: simplified.ins.addr,
             code: simplified.ins.code,