From 5470e808511a8aa0e8b40e55363a9add4e32fbf1 Mon Sep 17 00:00:00 2001
From: Philip Craig <philipjcraig@gmail.com>
Date: Wed, 15 Feb 2023 15:40:27 +1000
Subject: [PATCH] DWARF addresses are relative to the start of the code section

This partially reverts the previous commit.

Also fix the logic in `DebugLocReader::get_loc`.
---
 src/frontend.rs | 25 ++++++++++++++++---------
 src/ir/debug.rs |  8 +++++++-
 2 files changed, 23 insertions(+), 10 deletions(-)

diff --git a/src/frontend.rs b/src/frontend.rs
index 21f61bb..ae98a79 100644
--- a/src/frontend.rs
+++ b/src/frontend.rs
@@ -36,7 +36,7 @@ pub fn wasm_to_ir(bytes: &[u8]) -> Result<Module<'_>> {
         gimli::LocationLists::new(extra_sections.debug_loc, extra_sections.debug_loclists);
     dwarf.ranges =
         gimli::RangeLists::new(extra_sections.debug_ranges, extra_sections.debug_rnglists);
-    let debug_map = DebugMap::from_dwarf(dwarf, &mut module.debug)?;
+    let debug_map = DebugMap::from_dwarf(dwarf, &mut module.debug, extra_sections.code_offset)?;
     module.debug_map = debug_map;
 
     Ok(module)
@@ -74,6 +74,7 @@ struct ExtraSections<'a> {
     debug_loclists: gimli::DebugLocLists<gimli::EndianSlice<'a, gimli::LittleEndian>>,
     debug_ranges: gimli::DebugRanges<gimli::EndianSlice<'a, gimli::LittleEndian>>,
     debug_rnglists: gimli::DebugRngLists<gimli::EndianSlice<'a, gimli::LittleEndian>>,
+    code_offset: u32,
 }
 
 fn handle_payload<'a>(
@@ -170,6 +171,9 @@ fn handle_payload<'a>(
                 ));
             }
         }
+        Payload::CodeSectionStart { range, .. } => {
+            extra_sections.code_offset = range.start as u32;
+        }
         Payload::CodeSectionEntry(body) => {
             let func_idx = Func::new(*next_func);
             *next_func += 1;
@@ -285,7 +289,6 @@ fn handle_payload<'a>(
                 gimli::DebugRngLists::new(reader.data(), gimli::LittleEndian);
         }
         Payload::CustomSection(_) => {}
-        Payload::CodeSectionStart { .. } => {}
         Payload::Version { .. } => {}
         Payload::ElementSection(reader) => {
             for element in reader {
@@ -358,19 +361,22 @@ fn handle_payload<'a>(
 }
 
 struct DebugLocReader<'a> {
+    code_offset: u32,
     locs: &'a [(u32, u32, SourceLoc)],
 }
 
 impl<'a> DebugLocReader<'a> {
     fn new(module: &'a Module, func_offset_in_file: u32) -> Self {
+        let code_offset = module.debug_map.code_offset;
+        let func_address = func_offset_in_file - code_offset;
         let start = match module
             .debug_map
             .tuples
             .binary_search_by(|&(start, len, _)| {
                 use std::cmp::Ordering::*;
-                if start > func_offset_in_file {
+                if start > func_address {
                     Greater
-                } else if (start + len) <= func_offset_in_file {
+                } else if (start + len) <= func_address {
                     Less
                 } else {
                     Equal
@@ -380,21 +386,22 @@ impl<'a> DebugLocReader<'a> {
             Err(idx) => idx,
         };
         DebugLocReader {
+            code_offset,
             locs: &module.debug_map.tuples[start..],
         }
     }
 
     fn get_loc(&mut self, offset: usize) -> SourceLoc {
-        let offset = u32::try_from(offset).unwrap();
+        let address = u32::try_from(offset).unwrap() - self.code_offset;
         while self.locs.len() > 0 {
             let (start, len, loc) = self.locs[0];
-            if offset < start {
+            if address < start {
                 break;
             }
-            if offset > (start + len) {
-                self.locs = &self.locs[1..];
+            if address < start + len {
+                return loc;
             }
-            return loc;
+            self.locs = &self.locs[1..];
         }
         SourceLoc::invalid()
     }
diff --git a/src/ir/debug.rs b/src/ir/debug.rs
index 85f189d..b19b484 100644
--- a/src/ir/debug.rs
+++ b/src/ir/debug.rs
@@ -48,6 +48,8 @@ impl Debug {
 
 #[derive(Clone, Debug, Default)]
 pub struct DebugMap {
+    /// Offset of code section relative to the Wasm file start.
+    pub code_offset: u32,
     /// Each tuple is `(start, len, loc)`. The `start` offset is
     /// relative to the code section.
     pub tuples: Vec<(u32, u32, SourceLoc)>,
@@ -57,6 +59,7 @@ impl DebugMap {
     pub(crate) fn from_dwarf<R: gimli::Reader>(
         dwarf: gimli::Dwarf<R>,
         debug: &mut Debug,
+        code_offset: u32,
     ) -> anyhow::Result<DebugMap> {
         let ctx = addr2line::Context::from_dwarf(dwarf)?;
         let mut tuples = vec![];
@@ -79,6 +82,9 @@ impl DebugMap {
             retain
         });
 
-        Ok(DebugMap { tuples })
+        Ok(DebugMap {
+            code_offset,
+            tuples,
+        })
     }
 }