Version v2.0.0-alpha.5

* Initial ARM support

* Disassemble const pool reloc

* Disasm ARM/Thumb/data based on mapping symbols

* Fallback to mapping symbol `$a`

* Support multiple DWARF sequences

* Update line info

* Rework DWARF line info parsing

- Properly handles multiple sections
  in DWARF 1
- line_info moved into ObjSection
- DWARF 2 parser no longer errors with
  no .text section
- Both parsers properly skip empty
  sections

* Simplify line_info (no Option)

* Get line info from section; output formatted ins string

* Unwrap code section in `arm.rs`

* Handle reloc `R_ARM_SBREL32`

* Update ARM disassembler

* Update README.md

* Format

* Revert "Update README.md"

This reverts commit 8bbfcc6f45.

* Update README.md

* Detect ARM version; support ARMv4T and v6K

* Combobox to force ARM version

* Clear LSB in ARM symbol addresses

* Support big-endian ARM ELF files

* Bump `unarm`, `arm-attr`

* Handle ARM implicit addends

* Update README.md

* Explicitly handle all ARM argument types

* Format

* Display more ARM relocs

* Mask LSB on ARM code symbols only

* Read ARM implicit addends

* Format

---------

Co-authored-by: Luke Street <luke.street@encounterpc.com>

.github/workflows/build.yaml

@@ -9,7 +9,7 @@ on:
   workflow_dispatch:
 
 env:
-  CARGO_BIN_NAME: objdiff
+  BUILD_PROFILE: release-lto
   CARGO_TARGET_DIR: target
 
 jobs:
@@ -24,15 +24,39 @@ jobs:
           sudo apt-get update
           sudo apt-get -y install libgtk-3-dev
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
       - name: Setup Rust toolchain
         uses: dtolnay/rust-toolchain@stable
         with:
-          components: rustfmt, clippy
+          components: clippy
+      - name: Setup sccache
+        uses: mozilla-actions/sccache-action@v0.0.4
       - name: Cargo check
-        run: cargo check --all-features
+        env:
+          RUSTC_WRAPPER: sccache
+          SCCACHE_GHA_ENABLED: "true"
+        run: cargo check
       - name: Cargo clippy
-        run: cargo clippy --all-features
+        env:
+          RUSTC_WRAPPER: sccache
+          SCCACHE_GHA_ENABLED: "true"
+        run: cargo clippy
+
+  fmt:
+    name: Format
+    runs-on: ubuntu-latest
+    env:
+      RUSTFLAGS: -D warnings
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+      - name: Setup Rust toolchain
+        # We use nightly options in rustfmt.toml
+        uses: dtolnay/rust-toolchain@nightly
+        with:
+          components: rustfmt
+      - name: Cargo fmt
+        run: cargo fmt --all --check
 
   deny:
     name: Deny
@@ -45,13 +69,14 @@ jobs:
     # Prevent new advisories from failing CI
     continue-on-error: ${{ matrix.checks == 'advisories' }}
     steps:
-      - uses: actions/checkout@v3
+      - uses: actions/checkout@v4
       - uses: EmbarkStudios/cargo-deny-action@v1
         with:
           command: check ${{ matrix.checks }}
 
   test:
     name: Test
+    if: 'false' # No tests yet
     strategy:
       matrix:
         platform: [ ubuntu-latest, windows-latest, macos-latest ]
@@ -64,11 +89,16 @@ jobs:
           sudo apt-get update
           sudo apt-get -y install libgtk-3-dev
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
       - name: Setup Rust toolchain
         uses: dtolnay/rust-toolchain@stable
+      - name: Setup sccache
+        uses: mozilla-actions/sccache-action@v0.0.4
       - name: Cargo test
-        run: cargo test --release --all-features
+        env:
+          RUSTC_WRAPPER: sccache
+          SCCACHE_GHA_ENABLED: "true"
+        run: cargo test --release
 
   build:
     name: Build
@@ -79,15 +109,19 @@ jobs:
             target: x86_64-unknown-linux-gnu
             name: linux-x86_64
             packages: libgtk-3-dev
+            features: default
           - platform: windows-latest
             target: x86_64-pc-windows-msvc
             name: windows-x86_64
+            features: default
           - platform: macos-latest
             target: x86_64-apple-darwin
             name: macos-x86_64
+            features: default
           - platform: macos-latest
             target: aarch64-apple-darwin
             name: macos-arm64
+            features: default
       fail-fast: false
     runs-on: ${{ matrix.platform }}
     steps:
@@ -97,22 +131,29 @@ jobs:
           sudo apt-get update
           sudo apt-get -y install ${{ matrix.packages }}
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
       - name: Setup Rust toolchain
         uses: dtolnay/rust-toolchain@stable
         with:
           targets: ${{ matrix.target }}
+      - name: Setup sccache
+        uses: mozilla-actions/sccache-action@v0.0.4
       - name: Cargo build
-        run: cargo build --release --all-features --target ${{ matrix.target }} --bin ${{ env.CARGO_BIN_NAME }}
+        env:
+          RUSTC_WRAPPER: sccache
+          SCCACHE_GHA_ENABLED: "true"
+        run: >
+          cargo build --profile ${{ env.BUILD_PROFILE }} --target ${{ matrix.target }} 
+          --bin objdiff-cli --bin objdiff --features ${{ matrix.features }}
       - name: Upload artifacts
-        uses: actions/upload-artifact@v3
+        uses: actions/upload-artifact@v4
         with:
           name: ${{ matrix.name }}
           path: |
-            ${{ env.CARGO_TARGET_DIR }}/release/${{ env.CARGO_BIN_NAME }}
-            ${{ env.CARGO_TARGET_DIR }}/release/${{ env.CARGO_BIN_NAME }}.exe
-            ${{ env.CARGO_TARGET_DIR }}/${{ matrix.target }}/release/${{ env.CARGO_BIN_NAME }}
-            ${{ env.CARGO_TARGET_DIR }}/${{ matrix.target }}/release/${{ env.CARGO_BIN_NAME }}.exe
+            ${{ env.CARGO_TARGET_DIR }}/${{ matrix.target }}/${{ env.BUILD_PROFILE }}/objdiff-cli
+            ${{ env.CARGO_TARGET_DIR }}/${{ matrix.target }}/${{ env.BUILD_PROFILE }}/objdiff-cli.exe
+            ${{ env.CARGO_TARGET_DIR }}/${{ matrix.target }}/${{ env.BUILD_PROFILE }}/objdiff
+            ${{ env.CARGO_TARGET_DIR }}/${{ matrix.target }}/${{ env.BUILD_PROFILE }}/objdiff.exe
           if-no-files-found: error
 
   release:
@@ -120,17 +161,31 @@ jobs:
     if: startsWith(github.ref, 'refs/tags/')
     runs-on: ubuntu-latest
     needs: [ build ]
+    permissions:
+      contents: write
     steps:
       - name: Download artifacts
-        uses: actions/download-artifact@v3
+        uses: actions/download-artifact@v4
         with:
           path: artifacts
       - name: Rename artifacts
         working-directory: artifacts
         run: |
+          set -euo pipefail
           mkdir ../out
-          for i in */*/release/$CARGO_BIN_NAME*; do
-            mv "$i" "../out/$(sed -E "s/([^/]+)\/[^/]+\/release\/($CARGO_BIN_NAME)/\2-\1/" <<< "$i")"
+          for dir in */; do
+            for file in "$dir"*; do
+              base=$(basename "$file")
+              name="${base%.*}"
+              ext="${base##*.}"
+              if [ "$ext" = "$base" ]; then
+                  ext=""
+              else
+                  ext=".$ext"
+              fi
+              dst="../out/${name}-${dir%/}${ext}"
+              mv "$file" "$dst"
+            done
           done
           ls -R ../out
       - name: Release

Cargo.toml

@@ -1,87 +1,12 @@
-[package]
-name = "objdiff"
-version = "0.5.0"
-edition = "2021"
-rust-version = "1.70"
-authors = ["Luke Street <luke@street.dev>"]
-license = "MIT OR Apache-2.0"
-repository = "https://github.com/encounter/objdiff"
-readme = "README.md"
-description = """
-A local diffing tool for decompilation projects.
-"""
-publish = false
-build = "build.rs"
+[workspace]
+members = [
+    "objdiff-cli",
+    "objdiff-core",
+    "objdiff-gui",
+]
+resolver = "2"
 
-[profile.release]
+[profile.release-lto]
+inherits = "release"
 lto = "thin"
 strip = "debuginfo"
-
-[features]
-default = []
-wgpu = ["eframe/wgpu"]
-
-[dependencies]
-anyhow = "1.0.75"
-byteorder = "1.5.0"
-bytes = "1.5.0"
-cfg-if = "1.0.0"
-const_format = "0.2.31"
-cwdemangle = "0.1.6"
-dirs = "5.0.1"
-eframe = { version = "0.23.0", features = ["persistence"] }
-egui = "0.23.0"
-egui_extras = "0.23.0"
-flagset = "0.4.4"
-globset = { version = "0.4.13", features = ["serde1"] }
-log = "0.4.20"
-memmap2 = "0.9.0"
-notify = "6.1.1"
-object = { version = "0.32.1", features = ["read_core", "std", "elf"], default-features = false }
-png = "0.17.10"
-ppc750cl = { git = "https://github.com/terorie/ppc750cl", rev = "9ae36eef34aa6d74e00972c7671f547a2acfd0aa" }
-rabbitizer = "1.7.10"
-rfd = { version = "0.12.0" } #, default-features = false, features = ['xdg-portal']
-ron = "0.8.1"
-semver = "1.0.19"
-serde = { version = "1", features = ["derive"] }
-serde_json = "1.0.107"
-serde_yaml = "0.9.25"
-tempfile = "3.8.0"
-thiserror = "1.0.49"
-time = { version = "0.3.29", features = ["formatting", "local-offset"] }
-toml = "0.8.2"
-twox-hash = "1.6.3"
-
-# For Linux static binaries, use rustls
-[target.'cfg(target_os = "linux")'.dependencies]
-reqwest = { version = "0.11.22", default-features = false, features = ["blocking", "json", "rustls"] }
-self_update = { version = "0.38.0", default-features = false, features = ["rustls"] }
-
-# For all other platforms, use native TLS
-[target.'cfg(not(target_os = "linux"))'.dependencies]
-reqwest = "0.11.22"
-self_update = "0.38.0"
-
-[target.'cfg(windows)'.dependencies]
-path-slash = "0.2.1"
-winapi = "0.3.9"
-
-[target.'cfg(windows)'.build-dependencies]
-winres = "0.1.12"
-
-[target.'cfg(unix)'.dependencies]
-exec = "0.3.1"
-
-# native:
-[target.'cfg(not(target_arch = "wasm32"))'.dependencies]
-tracing-subscriber = "0.3"
-
-# web:
-[target.'cfg(target_arch = "wasm32")'.dependencies]
-console_error_panic_hook = "0.1.7"
-tracing-wasm = "0.2"
-
-[build-dependencies]
-anyhow = "1.0.75"
-vergen = { version = "8.2.5", features = ["build", "cargo", "git", "gitcl"] }

README.md

@@ -3,11 +3,21 @@
 [Build Status]: https://github.com/encounter/objdiff/actions/workflows/build.yaml/badge.svg
 [actions]: https://github.com/encounter/objdiff/actions
 
-A local diffing tool for decompilation projects.
+A local diffing tool for decompilation projects. Inspired by [decomp.me](https://decomp.me) and [asm-differ](https://github.com/simonlindholm/asm-differ).
+
+Features:
+- Compare entire object files: functions and data.
+- Built-in symbol demangling for C++. (CodeWarrior, Itanium & MSVC)
+- Automatic rebuild on source file changes.
+- Project integration via [configuration file](#configuration).
+- Search and filter all of a project's objects and quickly switch.
+- Click to highlight all instances of values and registers.
 
 Supports:
-- PowerPC 750CL (GameCube & Wii)
-- MIPS (Nintendo 64)
+- PowerPC 750CL (GameCube, Wii)
+- MIPS (N64, PS1, PS2, PSP)
+- x86 (COFF only at the moment)
+- ARM (GBA, DS, 3DS)
 
 See [Usage](#usage) for more information.
 
@@ -50,6 +60,10 @@ file as well. You can then add `objdiff.json` to your `.gitignore` to prevent it
 // objdiff.json
 {
   "custom_make": "ninja",
+  "custom_args": [
+    "-d",
+    "keeprsp"
+  ],
 
   // Only required if objects use "path" instead of "target_path" and "base_path".
   "target_dir": "build/asm",
@@ -83,7 +97,10 @@ file as well. You can then add `objdiff.json` to your `.gitignore` to prevent it
 ```
 
 `custom_make` _(optional)_: By default, objdiff will use `make` to build the project.  
-If the project uses a different build system (e.g. `ninja`), specify it here.
+If the project uses a different build system (e.g. `ninja`), specify it here.  
+The build command will be `[custom_make] [custom_args] path/to/object.o`.
+
+`custom_args` _(optional)_: Additional arguments to pass to the build command prior to the object path.
 
 `target_dir` _(optional)_: Relative from the root of the project, this where the "target" or "expected" objects are located.  
 These are the **intended result** of the match.
@@ -118,7 +135,17 @@ If not specified, objdiff will use the default patterns listed above.
 > `reverse_fn_order` _(optional)_: Displays function symbols in reversed order.  
 Used to support MWCC's `-inline deferred` option, which reverses the order of functions in the object file.
 
+## Building
 
+Install Rust via [rustup](https://rustup.rs).
+
+```shell
+$ git clone https://github.com/encounter/objdiff.git
+$ cd objdiff
+$ cargo run --release
+# or, for wgpu backend (recommended on macOS)
+$ cargo run --release --features wgpu
+```
 
 ## License
 

deny.toml

@@ -47,13 +47,7 @@ yanked = "warn"
 notice = "warn"
 # A list of advisory IDs to ignore. Note that ignored advisories will still
 # output a note when they are encountered.
-ignore = [
-    "RUSTSEC-2023-0022",
-    "RUSTSEC-2023-0023",
-    "RUSTSEC-2023-0024",
-    "RUSTSEC-2023-0034",
-    "RUSTSEC-2023-0044",
-]
+ignore = []
 # Threshold for security vulnerabilities, any vulnerability with a CVSS score
 # lower than the range specified will be ignored. Note that ignored advisories
 # will still output a note when they are encountered.
@@ -76,6 +70,7 @@ unlicensed = "deny"
 allow = [
     "MIT",
     "Apache-2.0",
+    "Apache-2.0 WITH LLVM-exception",
     "ISC",
     "BSD-2-Clause",
     "BSD-3-Clause",
@@ -159,7 +154,7 @@ registries = [
 # https://embarkstudios.github.io/cargo-deny/checks/bans/cfg.html
 [bans]
 # Lint level for when multiple versions of the same crate are detected
-multiple-versions = "warn"
+multiple-versions = "allow"
 # Lint level for when a crate version requirement is `*`
 wildcards = "allow"
 # The graph highlighting used when creating dotgraphs for crates

objdiff-cli/Cargo.toml

@@ -0,0 +1,29 @@
+[package]
+name = "objdiff-cli"
+version = "2.0.0-alpha.5"
+edition = "2021"
+rust-version = "1.70"
+authors = ["Luke Street <luke@street.dev>"]
+license = "MIT OR Apache-2.0"
+repository = "https://github.com/encounter/objdiff"
+readme = "../README.md"
+description = """
+A local diffing tool for decompilation projects.
+"""
+publish = false
+build = "build.rs"
+
+[dependencies]
+anyhow = "1.0.82"
+argp = "0.3.0"
+crossterm = "0.27.0"
+enable-ansi-support = "0.2.1"
+objdiff-core = { path = "../objdiff-core", features = ["all"] }
+ratatui = "0.26.2"
+rayon = "1.10.0"
+serde = { version = "1", features = ["derive"] }
+serde_json = "1.0.116"
+supports-color = "3.0.0"
+time = { version = "0.3.36", features = ["formatting", "local-offset"] }
+tracing = "0.1.40"
+tracing-subscriber = { version = "0.3.18", features = ["env-filter"] }

objdiff-cli/build.rs

@@ -0,0 +1,9 @@
+fn main() {
+    let output = std::process::Command::new("git")
+        .args(["rev-parse", "HEAD"])
+        .output()
+        .expect("Failed to execute git");
+    let rev = String::from_utf8(output.stdout).expect("Failed to parse git output");
+    println!("cargo:rustc-env=GIT_COMMIT_SHA={rev}");
+    println!("cargo:rustc-rerun-if-changed=.git/HEAD");
+}

objdiff-cli/src/argp_version.rs

@@ -0,0 +1,64 @@
+// Originally from https://gist.github.com/suluke/e0c672492126be0a4f3b4f0e1115d77c
+//! Extend `argp` to be better integrated with the `cargo` ecosystem
+//!
+//! For now, this only adds a --version/-V option which causes early-exit.
+use std::ffi::OsStr;
+
+use argp::{parser::ParseGlobalOptions, EarlyExit, FromArgs, TopLevelCommand};
+
+struct ArgsOrVersion<T>(T)
+where T: FromArgs;
+
+impl<T> TopLevelCommand for ArgsOrVersion<T> where T: FromArgs {}
+
+impl<T> FromArgs for ArgsOrVersion<T>
+where T: FromArgs
+{
+    fn _from_args(
+        command_name: &[&str],
+        args: &[&OsStr],
+        parent: Option<&mut dyn ParseGlobalOptions>,
+    ) -> Result<Self, EarlyExit> {
+        /// Also use argp for catching `--version`-only invocations
+        #[derive(FromArgs)]
+        struct Version {
+            /// Print version information and exit.
+            #[argp(switch, short = 'V')]
+            pub version: bool,
+        }
+
+        match Version::from_args(command_name, args) {
+            Ok(v) => {
+                if v.version {
+                    println!(
+                        "{} {} {}",
+                        command_name.first().unwrap_or(&""),
+                        env!("CARGO_PKG_VERSION"),
+                        env!("GIT_COMMIT_SHA"),
+                    );
+                    std::process::exit(0);
+                } else {
+                    // Pass through empty arguments
+                    T::_from_args(command_name, args, parent).map(Self)
+                }
+            }
+            Err(exit) => match exit {
+                EarlyExit::Help(_help) => {
+                    // TODO: Chain help info from Version
+                    // For now, we just put the switch on T as well
+                    T::from_args(command_name, &["--help"]).map(Self)
+                }
+                EarlyExit::Err(_) => T::_from_args(command_name, args, parent).map(Self),
+            },
+        }
+    }
+}
+
+/// Create a `FromArgs` type from the current process’s `env::args`.
+///
+/// This function will exit early from the current process if argument parsing was unsuccessful or if information like `--help` was requested.
+/// Error messages will be printed to stderr, and `--help` output to stdout.
+pub fn from_env<T>() -> T
+where T: TopLevelCommand {
+    argp::parse_args_or_exit::<ArgsOrVersion<T>>(argp::DEFAULT).0
+}

objdiff-cli/src/cmd/mod.rs

@@ -0,0 +1,2 @@
+pub mod diff;
+pub mod report;

objdiff-cli/src/main.rs

@@ -0,0 +1,141 @@
+mod argp_version;
+mod cmd;
+mod util;
+
+use std::{env, ffi::OsStr, fmt::Display, path::PathBuf, str::FromStr};
+
+use anyhow::{Error, Result};
+use argp::{FromArgValue, FromArgs};
+use enable_ansi_support::enable_ansi_support;
+use supports_color::Stream;
+use tracing_subscriber::{filter::LevelFilter, EnvFilter};
+
+#[derive(Debug, Eq, PartialEq, Copy, Clone)]
+enum LogLevel {
+    Error,
+    Warn,
+    Info,
+    Debug,
+    Trace,
+}
+
+impl FromStr for LogLevel {
+    type Err = ();
+
+    fn from_str(s: &str) -> Result<Self, Self::Err> {
+        Ok(match s {
+            "error" => Self::Error,
+            "warn" => Self::Warn,
+            "info" => Self::Info,
+            "debug" => Self::Debug,
+            "trace" => Self::Trace,
+            _ => return Err(()),
+        })
+    }
+}
+
+impl Display for LogLevel {
+    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+        f.write_str(match self {
+            LogLevel::Error => "error",
+            LogLevel::Warn => "warn",
+            LogLevel::Info => "info",
+            LogLevel::Debug => "debug",
+            LogLevel::Trace => "trace",
+        })
+    }
+}
+
+impl FromArgValue for LogLevel {
+    fn from_arg_value(value: &OsStr) -> Result<Self, String> {
+        String::from_arg_value(value)
+            .and_then(|s| Self::from_str(&s).map_err(|_| "Invalid log level".to_string()))
+    }
+}
+
+#[derive(FromArgs, PartialEq, Debug)]
+/// Yet another GameCube/Wii decompilation toolkit.
+struct TopLevel {
+    #[argp(subcommand)]
+    command: SubCommand,
+    #[argp(option, short = 'C')]
+    /// Change working directory.
+    chdir: Option<PathBuf>,
+    #[argp(option, short = 'L')]
+    /// Minimum logging level. (Default: info)
+    /// Possible values: error, warn, info, debug, trace
+    log_level: Option<LogLevel>,
+    /// Print version information and exit.
+    #[argp(switch, short = 'V')]
+    version: bool,
+    /// Disable color output. (env: NO_COLOR)
+    #[argp(switch)]
+    no_color: bool,
+}
+
+#[derive(FromArgs, PartialEq, Debug)]
+#[argp(subcommand)]
+enum SubCommand {
+    Diff(cmd::diff::Args),
+    Report(cmd::report::Args),
+}
+
+// Duplicated from supports-color so we can check early.
+fn env_no_color() -> bool {
+    match env::var("NO_COLOR").as_deref() {
+        Ok("") | Ok("0") | Err(_) => false,
+        Ok(_) => true,
+    }
+}
+
+fn main() {
+    let args: TopLevel = argp_version::from_env();
+    let use_colors = if args.no_color || env_no_color() {
+        false
+    } else {
+        // Try to enable ANSI support on Windows.
+        let _ = enable_ansi_support();
+        // Disable isatty check for supports-color. (e.g. when used with ninja)
+        env::set_var("IGNORE_IS_TERMINAL", "1");
+        supports_color::on(Stream::Stdout).is_some_and(|c| c.has_basic)
+    };
+
+    let format =
+        tracing_subscriber::fmt::format().with_ansi(use_colors).with_target(false).without_time();
+    let builder = tracing_subscriber::fmt().event_format(format).with_writer(std::io::stderr);
+    if let Some(level) = args.log_level {
+        builder
+            .with_max_level(match level {
+                LogLevel::Error => LevelFilter::ERROR,
+                LogLevel::Warn => LevelFilter::WARN,
+                LogLevel::Info => LevelFilter::INFO,
+                LogLevel::Debug => LevelFilter::DEBUG,
+                LogLevel::Trace => LevelFilter::TRACE,
+            })
+            .init();
+    } else {
+        builder
+            .with_env_filter(
+                EnvFilter::builder()
+                    .with_default_directive(LevelFilter::INFO.into())
+                    .from_env_lossy(),
+            )
+            .init();
+    }
+
+    let mut result = Ok(());
+    if let Some(dir) = &args.chdir {
+        result = env::set_current_dir(dir).map_err(|e| {
+            Error::new(e)
+                .context(format!("Failed to change working directory to '{}'", dir.display()))
+        });
+    }
+    result = result.and_then(|_| match args.command {
+        SubCommand::Diff(c_args) => cmd::diff::run(c_args),
+        SubCommand::Report(c_args) => cmd::report::run(c_args),
+    });
+    if let Err(e) = result {
+        eprintln!("Failed: {e:?}");
+        std::process::exit(1);
+    }
+}

objdiff-cli/src/util/mod.rs

@@ -0,0 +1 @@
+pub mod term;

objdiff-cli/src/util/term.rs

@@ -0,0 +1,16 @@
+use std::{io::stdout, panic};
+
+use crossterm::{
+    cursor::Show,
+    event::DisableMouseCapture,
+    terminal::{disable_raw_mode, LeaveAlternateScreen},
+};
+
+pub fn crossterm_panic_handler() {
+    let original_hook = panic::take_hook();
+    panic::set_hook(Box::new(move |panic_info| {
+        let _ = crossterm::execute!(stdout(), LeaveAlternateScreen, DisableMouseCapture, Show);
+        let _ = disable_raw_mode();
+        original_hook(panic_info);
+    }));
+}

objdiff-core/Cargo.toml

@@ -0,0 +1,60 @@
+[package]
+name = "objdiff-core"
+version = "2.0.0-alpha.5"
+edition = "2021"
+rust-version = "1.70"
+authors = ["Luke Street <luke@street.dev>"]
+license = "MIT OR Apache-2.0"
+repository = "https://github.com/encounter/objdiff"
+readme = "../README.md"
+description = """
+A local diffing tool for decompilation projects.
+"""
+
+[features]
+all = ["config", "dwarf", "mips", "ppc", "x86", "arm"]
+any-arch = [] # Implicit, used to check if any arch is enabled
+config = ["globset", "semver", "serde_json", "serde_yaml"]
+dwarf = ["gimli"]
+mips = ["any-arch", "rabbitizer"]
+ppc = ["any-arch", "cwdemangle", "ppc750cl"]
+x86 = ["any-arch", "cpp_demangle", "iced-x86", "msvc-demangler"]
+arm = ["any-arch", "cpp_demangle", "unarm", "arm-attr"]
+
+[dependencies]
+anyhow = "1.0.82"
+byteorder = "1.5.0"
+filetime = "0.2.23"
+flagset = "0.4.5"
+log = "0.4.21"
+memmap2 = "0.9.4"
+num-traits = "0.2.18"
+object = { version = "0.35.0", features = ["read_core", "std", "elf", "pe"], default-features = false }
+serde = { version = "1", features = ["derive"] }
+similar = { version = "2.5.0", default-features = false }
+strum = { version = "0.26.2", features = ["derive"] }
+
+# config
+globset = { version = "0.4.14", features = ["serde1"], optional = true }
+semver = { version = "1.0.22", optional = true }
+serde_json = { version = "1.0.116", optional = true }
+serde_yaml = { version = "0.9.34", optional = true }
+
+# dwarf
+gimli = { version = "0.29.0", default-features = false, features = ["read-all"], optional = true }
+
+# ppc
+cwdemangle = { version = "1.0.0", optional = true }
+ppc750cl = { git = "https://github.com/encounter/ppc750cl", rev = "6cbd7d888c7082c2c860f66cbb9848d633f753ed", optional = true }
+
+# mips
+rabbitizer = { version = "1.11.0", optional = true }
+
+# x86
+cpp_demangle = { version = "0.4.3", optional = true }
+iced-x86 = { version = "1.21.0", default-features = false, features = ["std", "decoder", "intel", "gas", "masm", "nasm", "exhaustive_enums"], optional = true }
+msvc-demangler = { version = "0.10.0", optional = true }
+
+# arm
+unarm = { version = "1.3.0", optional = true }
+arm-attr = { version = "0.1.1", optional = true }

objdiff-core/src/arch/arm.rs

@@ -0,0 +1,414 @@
+use std::{
+    borrow::Cow,
+    collections::{BTreeMap, HashMap},
+};
+
+use anyhow::{bail, Result};
+use arm_attr::{enums::CpuArch, tag::Tag, BuildAttrs};
+use object::{
+    elf::{self, SHT_ARM_ATTRIBUTES},
+    Endian, File, Object, ObjectSection, ObjectSymbol, Relocation, RelocationFlags, SectionIndex,
+    SectionKind, Symbol, SymbolKind,
+};
+use unarm::{
+    args::{Argument, OffsetImm, OffsetReg, Register},
+    parse::{ArmVersion, ParseMode, Parser},
+    ParsedIns,
+};
+
+use crate::{
+    arch::{ObjArch, ProcessCodeResult},
+    diff::{ArmArchVersion, DiffObjConfig},
+    obj::{ObjIns, ObjInsArg, ObjInsArgValue, ObjReloc, ObjSection},
+};
+
+pub struct ObjArchArm {
+    /// Maps section index, to list of disasm modes (arm, thumb or data) sorted by address
+    disasm_modes: HashMap<SectionIndex, Vec<DisasmMode>>,
+    detected_version: Option<ArmVersion>,
+    endianness: object::Endianness,
+}
+
+impl ObjArchArm {
+    pub fn new(file: &File) -> Result<Self> {
+        let endianness = file.endianness();
+        match file {
+            File::Elf32(_) => {
+                let disasm_modes = Self::elf_get_mapping_symbols(file);
+                let detected_version = Self::elf_detect_arm_version(file)?;
+                Ok(Self { disasm_modes, detected_version, endianness })
+            }
+            _ => bail!("Unsupported file format {:?}", file.format()),
+        }
+    }
+
+    fn elf_detect_arm_version(file: &File) -> Result<Option<ArmVersion>> {
+        // Check ARM attributes
+        if let Some(arm_attrs) = file.sections().find(|s| {
+            s.kind() == SectionKind::Elf(SHT_ARM_ATTRIBUTES) && s.name() == Ok(".ARM.attributes")
+        }) {
+            let attr_data = arm_attrs.uncompressed_data()?;
+            let build_attrs = BuildAttrs::new(&attr_data, match file.endianness() {
+                object::Endianness::Little => arm_attr::Endian::Little,
+                object::Endianness::Big => arm_attr::Endian::Big,
+            })?;
+            for subsection in build_attrs.subsections() {
+                let subsection = subsection?;
+                if !subsection.is_aeabi() {
+                    continue;
+                }
+                // Only checking first CpuArch tag. Others may exist, but that's very unlikely.
+                let cpu_arch = subsection.into_public_tag_iter()?.find_map(|(_, tag)| {
+                    if let Tag::CpuArch(cpu_arch) = tag {
+                        Some(cpu_arch)
+                    } else {
+                        None
+                    }
+                });
+                match cpu_arch {
+                    Some(CpuArch::V4T) => return Ok(Some(ArmVersion::V4T)),
+                    Some(CpuArch::V5TE) => return Ok(Some(ArmVersion::V5Te)),
+                    Some(CpuArch::V6K) => return Ok(Some(ArmVersion::V6K)),
+                    Some(arch) => bail!("ARM arch {} not supported", arch),
+                    None => {}
+                };
+            }
+        }
+
+        Ok(None)
+    }
+
+    fn elf_get_mapping_symbols(file: &File) -> HashMap<SectionIndex, Vec<DisasmMode>> {
+        file.sections()
+            .filter(|s| s.kind() == SectionKind::Text)
+            .map(|s| {
+                let index = s.index();
+                let mut mapping_symbols: Vec<_> = file
+                    .symbols()
+                    .filter(|s| s.section_index().map(|i| i == index).unwrap_or(false))
+                    .filter_map(|s| DisasmMode::from_symbol(&s))
+                    .collect();
+                mapping_symbols.sort_unstable_by_key(|x| x.address);
+                (s.index(), mapping_symbols)
+            })
+            .collect()
+    }
+}
+
+impl ObjArch for ObjArchArm {
+    fn symbol_address(&self, symbol: &Symbol) -> u64 {
+        let address = symbol.address();
+        if symbol.kind() == SymbolKind::Text {
+            address & !1
+        } else {
+            address
+        }
+    }
+
+    fn process_code(
+        &self,
+        address: u64,
+        code: &[u8],
+        section_index: usize,
+        relocations: &[ObjReloc],
+        line_info: &BTreeMap<u64, u64>,
+        config: &DiffObjConfig,
+    ) -> Result<ProcessCodeResult> {
+        let start_addr = address as u32;
+        let end_addr = start_addr + code.len() as u32;
+
+        // Mapping symbols decide what kind of data comes after it. $a for ARM code, $t for Thumb code and $d for data.
+        let fallback_mappings = [DisasmMode { address: start_addr, mapping: ParseMode::Arm }];
+        let mapping_symbols = self
+            .disasm_modes
+            .get(&SectionIndex(section_index))
+            .map(|x| x.as_slice())
+            .unwrap_or(&fallback_mappings);
+        let first_mapping_idx =
+            match mapping_symbols.binary_search_by_key(&start_addr, |x| x.address) {
+                Ok(idx) => idx,
+                Err(idx) => idx - 1,
+            };
+        let first_mapping = mapping_symbols[first_mapping_idx].mapping;
+
+        let mut mappings_iter =
+            mapping_symbols.iter().skip(first_mapping_idx + 1).take_while(|x| x.address < end_addr);
+        let mut next_mapping = mappings_iter.next();
+
+        let ins_count = code.len() / first_mapping.instruction_size(start_addr);
+        let mut ops = Vec::<u16>::with_capacity(ins_count);
+        let mut insts = Vec::<ObjIns>::with_capacity(ins_count);
+
+        let version = match config.arm_arch_version {
+            ArmArchVersion::Auto => self.detected_version.unwrap_or(ArmVersion::V5Te),
+            ArmArchVersion::V4T => ArmVersion::V4T,
+            ArmArchVersion::V5TE => ArmVersion::V5Te,
+            ArmArchVersion::V6K => ArmVersion::V6K,
+        };
+        let endian = match self.endianness {
+            object::Endianness::Little => unarm::Endian::Little,
+            object::Endianness::Big => unarm::Endian::Big,
+        };
+        let mut parser = Parser::new(version, first_mapping, start_addr, endian, code);
+
+        while let Some((address, op, ins)) = parser.next() {
+            if let Some(next) = next_mapping {
+                let next_address = parser.address;
+                if next_address >= next.address {
+                    // Change mapping
+                    parser.mode = next.mapping;
+                    next_mapping = mappings_iter.next();
+                }
+            }
+            let line = line_info.range(..=address as u64).last().map(|(_, &b)| b);
+
+            let reloc = relocations.iter().find(|r| (r.address as u32 & !1) == address).cloned();
+
+            let mut reloc_arg = None;
+            if let Some(reloc) = &reloc {
+                match reloc.flags {
+                    // Calls
+                    RelocationFlags::Elf { r_type: elf::R_ARM_THM_XPC22 }
+                    | RelocationFlags::Elf { r_type: elf::R_ARM_THM_PC22 }
+                    | RelocationFlags::Elf { r_type: elf::R_ARM_PC24 }
+                    | RelocationFlags::Elf { r_type: elf::R_ARM_XPC25 }
+                    | RelocationFlags::Elf { r_type: elf::R_ARM_CALL } => {
+                        reloc_arg =
+                            ins.args.iter().rposition(|a| matches!(a, Argument::BranchDest(_)));
+                    }
+                    // Data
+                    RelocationFlags::Elf { r_type: elf::R_ARM_ABS32 } => {
+                        reloc_arg = ins.args.iter().rposition(|a| matches!(a, Argument::UImm(_)));
+                    }
+                    _ => (),
+                }
+            };
+
+            let (args, branch_dest) = if reloc.is_some() && parser.mode == ParseMode::Data {
+                (vec![ObjInsArg::Reloc], None)
+            } else {
+                push_args(&ins, config, reloc_arg, address)?
+            };
+
+            ops.push(op.id());
+            insts.push(ObjIns {
+                address: address as u64,
+                size: (parser.address - address) as u8,
+                op: op.id(),
+                mnemonic: ins.mnemonic.to_string(),
+                args,
+                reloc,
+                branch_dest,
+                line,
+                formatted: ins.to_string(),
+                orig: None,
+            });
+        }
+
+        Ok(ProcessCodeResult { ops, insts })
+    }
+
+    fn implcit_addend(
+        &self,
+        section: &ObjSection,
+        address: u64,
+        reloc: &Relocation,
+    ) -> anyhow::Result<i64> {
+        let address = address as usize;
+        Ok(match reloc.flags() {
+            // ARM calls
+            RelocationFlags::Elf { r_type: elf::R_ARM_PC24 }
+            | RelocationFlags::Elf { r_type: elf::R_ARM_XPC25 }
+            | RelocationFlags::Elf { r_type: elf::R_ARM_CALL } => {
+                let data = section.data[address..address + 4].try_into()?;
+                let addend = self.endianness.read_i32_bytes(data);
+                let imm24 = addend & 0xffffff;
+                (imm24 << 2) << 8 >> 8
+            }
+
+            // Thumb calls
+            RelocationFlags::Elf { r_type: elf::R_ARM_THM_PC22 }
+            | RelocationFlags::Elf { r_type: elf::R_ARM_THM_XPC22 } => {
+                let data = section.data[address..address + 2].try_into()?;
+                let high = self.endianness.read_i16_bytes(data) as i32;
+                let data = section.data[address + 2..address + 4].try_into()?;
+                let low = self.endianness.read_i16_bytes(data) as i32;
+
+                let imm22 = ((high & 0x7ff) << 11) | (low & 0x7ff);
+                (imm22 << 1) << 9 >> 9
+            }
+
+            // Data
+            RelocationFlags::Elf { r_type: elf::R_ARM_ABS32 } => {
+                let data = section.data[address..address + 4].try_into()?;
+                self.endianness.read_i32_bytes(data)
+            }
+
+            flags => bail!("Unsupported ARM implicit relocation {flags:?}"),
+        } as i64)
+    }
+
+    fn demangle(&self, name: &str) -> Option<String> {
+        cpp_demangle::Symbol::new(name)
+            .ok()
+            .and_then(|s| s.demangle(&cpp_demangle::DemangleOptions::default()).ok())
+    }
+
+    fn display_reloc(&self, flags: RelocationFlags) -> Cow<'static, str> {
+        Cow::Owned(format!("<{flags:?}>"))
+    }
+}
+
+#[derive(Clone, Copy, Debug)]
+struct DisasmMode {
+    address: u32,
+    mapping: ParseMode,
+}
+
+impl DisasmMode {
+    fn from_symbol<'a>(sym: &Symbol<'a, '_, &'a [u8]>) -> Option<Self> {
+        if let Ok(name) = sym.name() {
+            ParseMode::from_mapping_symbol(name)
+                .map(|mapping| DisasmMode { address: sym.address() as u32, mapping })
+        } else {
+            None
+        }
+    }
+}
+
+fn push_args(
+    parsed_ins: &ParsedIns,
+    config: &DiffObjConfig,
+    reloc_arg: Option<usize>,
+    cur_addr: u32,
+) -> Result<(Vec<ObjInsArg>, Option<u64>)> {
+    let mut args = vec![];
+    let mut branch_dest = None;
+    let mut writeback = false;
+    let mut deref = false;
+    for (i, arg) in parsed_ins.args_iter().enumerate() {
+        // Emit punctuation before separator
+        if deref {
+            match arg {
+                Argument::OffsetImm(OffsetImm { post_indexed: true, value: _ })
+                | Argument::OffsetReg(OffsetReg { add: _, post_indexed: true, reg: _ })
+                | Argument::CoOption(_) => {
+                    deref = false;
+                    args.push(ObjInsArg::PlainText("]".into()));
+                    if writeback {
+                        writeback = false;
+                        args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque("!".into())));
+                    }
+                }
+                _ => {}
+            }
+        }
+
+        if i > 0 {
+            args.push(ObjInsArg::PlainText(config.separator().into()));
+        }
+
+        if reloc_arg == Some(i) {
+            args.push(ObjInsArg::Reloc);
+        } else {
+            match arg {
+                Argument::None => {}
+                Argument::Reg(reg) => {
+                    if reg.deref {
+                        deref = true;
+                        args.push(ObjInsArg::PlainText("[".into()));
+                    }
+                    args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(reg.reg.to_string().into())));
+                    if reg.writeback {
+                        if reg.deref {
+                            writeback = true;
+                        } else {
+                            args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque("!".into())));
+                        }
+                    }
+                }
+                Argument::RegList(reg_list) => {
+                    args.push(ObjInsArg::PlainText("{".into()));
+                    let mut first = true;
+                    for i in 0..16 {
+                        if (reg_list.regs & (1 << i)) != 0 {
+                            if !first {
+                                args.push(ObjInsArg::PlainText(config.separator().into()));
+                            }
+                            args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(
+                                Register::parse(i).to_string().into(),
+                            )));
+                            first = false;
+                        }
+                    }
+                    args.push(ObjInsArg::PlainText("}".into()));
+                    if reg_list.user_mode {
+                        args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque("^".to_string().into())));
+                    }
+                }
+                Argument::UImm(value) | Argument::CoOpcode(value) | Argument::SatImm(value) => {
+                    args.push(ObjInsArg::PlainText("#".into()));
+                    args.push(ObjInsArg::Arg(ObjInsArgValue::Unsigned(*value as u64)));
+                }
+                Argument::SImm(value)
+                | Argument::OffsetImm(OffsetImm { post_indexed: _, value }) => {
+                    args.push(ObjInsArg::PlainText("#".into()));
+                    args.push(ObjInsArg::Arg(ObjInsArgValue::Signed(*value as i64)));
+                }
+                Argument::BranchDest(value) => {
+                    let dest = cur_addr.wrapping_add_signed(*value) as u64;
+                    args.push(ObjInsArg::BranchDest(dest));
+                    branch_dest = Some(dest);
+                }
+                Argument::CoOption(value) => {
+                    args.push(ObjInsArg::PlainText("{".into()));
+                    args.push(ObjInsArg::Arg(ObjInsArgValue::Unsigned(*value as u64)));
+                    args.push(ObjInsArg::PlainText("}".into()));
+                }
+                Argument::CoprocNum(value) => {
+                    args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(format!("p{}", value).into())));
+                }
+                Argument::ShiftImm(shift) => {
+                    args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(shift.op.to_string().into())));
+                    args.push(ObjInsArg::PlainText(" #".into()));
+                    args.push(ObjInsArg::Arg(ObjInsArgValue::Unsigned(shift.imm as u64)));
+                }
+                Argument::ShiftReg(shift) => {
+                    args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(shift.op.to_string().into())));
+                    args.push(ObjInsArg::PlainText(" ".into()));
+                    args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(shift.reg.to_string().into())));
+                }
+                Argument::OffsetReg(offset) => {
+                    if !offset.add {
+                        args.push(ObjInsArg::PlainText("-".into()));
+                    }
+                    args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(
+                        offset.reg.to_string().into(),
+                    )));
+                }
+                Argument::CpsrMode(mode) => {
+                    args.push(ObjInsArg::PlainText("#".into()));
+                    args.push(ObjInsArg::Arg(ObjInsArgValue::Unsigned(mode.mode as u64)));
+                    if mode.writeback {
+                        args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque("!".into())));
+                    }
+                }
+                Argument::CoReg(_)
+                | Argument::StatusReg(_)
+                | Argument::StatusMask(_)
+                | Argument::Shift(_)
+                | Argument::CpsrFlags(_)
+                | Argument::Endian(_) => {
+                    args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(arg.to_string().into())))
+                }
+            }
+        }
+    }
+    if deref {
+        args.push(ObjInsArg::PlainText("]".into()));
+        if writeback {
+            args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque("!".into())));
+        }
+    }
+    Ok((args, branch_dest))
+}

objdiff-core/src/arch/mips.rs

@@ -0,0 +1,255 @@
+use std::{borrow::Cow, collections::BTreeMap, sync::Mutex};
+
+use anyhow::{anyhow, bail, Result};
+use object::{elf, Endian, Endianness, File, FileFlags, Object, Relocation, RelocationFlags};
+use rabbitizer::{config, Abi, InstrCategory, Instruction, OperandType};
+
+use crate::{
+    arch::{ObjArch, ProcessCodeResult},
+    diff::{DiffObjConfig, MipsAbi, MipsInstrCategory},
+    obj::{ObjIns, ObjInsArg, ObjInsArgValue, ObjReloc, ObjSection},
+};
+
+static RABBITIZER_MUTEX: Mutex<()> = Mutex::new(());
+
+fn configure_rabbitizer(abi: Abi) {
+    unsafe {
+        config::RabbitizerConfig_Cfg.reg_names.fpr_abi_names = abi;
+    }
+}
+
+pub struct ObjArchMips {
+    pub endianness: Endianness,
+    pub abi: Abi,
+    pub instr_category: InstrCategory,
+}
+
+const EF_MIPS_ABI: u32 = 0x0000F000;
+const EF_MIPS_MACH: u32 = 0x00FF0000;
+
+const EF_MIPS_MACH_ALLEGREX: u32 = 0x00840000;
+const EF_MIPS_MACH_5900: u32 = 0x00920000;
+
+impl ObjArchMips {
+    pub fn new(object: &File) -> Result<Self> {
+        let mut abi = Abi::NUMERIC;
+        let mut instr_category = InstrCategory::CPU;
+        match object.flags() {
+            FileFlags::None => {}
+            FileFlags::Elf { e_flags, .. } => {
+                abi = match e_flags & EF_MIPS_ABI {
+                    elf::EF_MIPS_ABI_O32 | elf::EF_MIPS_ABI_O64 => Abi::O32,
+                    elf::EF_MIPS_ABI_EABI32 | elf::EF_MIPS_ABI_EABI64 => Abi::N32,
+                    _ => {
+                        if e_flags & elf::EF_MIPS_ABI2 != 0 {
+                            Abi::N32
+                        } else {
+                            Abi::NUMERIC
+                        }
+                    }
+                };
+                instr_category = match e_flags & EF_MIPS_MACH {
+                    EF_MIPS_MACH_ALLEGREX => InstrCategory::R4000ALLEGREX,
+                    EF_MIPS_MACH_5900 => InstrCategory::R5900,
+                    _ => InstrCategory::CPU,
+                };
+            }
+            _ => bail!("Unsupported MIPS file flags"),
+        }
+        Ok(Self { endianness: object.endianness(), abi, instr_category })
+    }
+}
+
+impl ObjArch for ObjArchMips {
+    fn process_code(
+        &self,
+        address: u64,
+        code: &[u8],
+        _section_index: usize,
+        relocations: &[ObjReloc],
+        line_info: &BTreeMap<u64, u64>,
+        config: &DiffObjConfig,
+    ) -> Result<ProcessCodeResult> {
+        let _guard = RABBITIZER_MUTEX.lock().map_err(|e| anyhow!("Failed to lock mutex: {e}"))?;
+        configure_rabbitizer(match config.mips_abi {
+            MipsAbi::Auto => self.abi,
+            MipsAbi::O32 => Abi::O32,
+            MipsAbi::N32 => Abi::N32,
+            MipsAbi::N64 => Abi::N64,
+        });
+        let instr_category = match config.mips_instr_category {
+            MipsInstrCategory::Auto => self.instr_category,
+            MipsInstrCategory::Cpu => InstrCategory::CPU,
+            MipsInstrCategory::Rsp => InstrCategory::RSP,
+            MipsInstrCategory::R3000Gte => InstrCategory::R3000GTE,
+            MipsInstrCategory::R4000Allegrex => InstrCategory::R4000ALLEGREX,
+            MipsInstrCategory::R5900 => InstrCategory::R5900,
+        };
+
+        let start_address = address;
+        let end_address = address + code.len() as u64;
+        let ins_count = code.len() / 4;
+        let mut ops = Vec::<u16>::with_capacity(ins_count);
+        let mut insts = Vec::<ObjIns>::with_capacity(ins_count);
+        let mut cur_addr = start_address as u32;
+        for chunk in code.chunks_exact(4) {
+            let reloc = relocations.iter().find(|r| (r.address as u32 & !3) == cur_addr);
+            let code = self.endianness.read_u32_bytes(chunk.try_into()?);
+            let instruction = Instruction::new(code, cur_addr, instr_category);
+
+            let formatted = instruction.disassemble(None, 0);
+            let op = instruction.unique_id as u16;
+            ops.push(op);
+
+            let mnemonic = instruction.opcode_name().to_string();
+            let is_branch = instruction.is_branch();
+            let branch_offset = instruction.branch_offset();
+            let branch_dest = if is_branch {
+                cur_addr.checked_add_signed(branch_offset).map(|a| a as u64)
+            } else {
+                None
+            };
+
+            let operands = instruction.get_operands_slice();
+            let mut args = Vec::with_capacity(operands.len() + 1);
+            for (idx, op) in operands.iter().enumerate() {
+                if idx > 0 {
+                    args.push(ObjInsArg::PlainText(config.separator().into()));
+                }
+
+                match op {
+                    OperandType::cpu_immediate
+                    | OperandType::cpu_label
+                    | OperandType::cpu_branch_target_label => {
+                        if let Some(branch_dest) = branch_dest {
+                            args.push(ObjInsArg::BranchDest(branch_dest));
+                        } else if let Some(reloc) = reloc {
+                            if matches!(&reloc.target_section, Some(s) if s == ".text")
+                                && reloc.target.address > start_address
+                                && reloc.target.address < end_address
+                            {
+                                args.push(ObjInsArg::BranchDest(reloc.target.address));
+                            } else {
+                                push_reloc(&mut args, reloc)?;
+                            }
+                        } else {
+                            args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(
+                                op.disassemble(&instruction, None).into(),
+                            )));
+                        }
+                    }
+                    OperandType::cpu_immediate_base => {
+                        if let Some(reloc) = reloc {
+                            push_reloc(&mut args, reloc)?;
+                        } else {
+                            args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(
+                                OperandType::cpu_immediate.disassemble(&instruction, None).into(),
+                            )));
+                        }
+                        args.push(ObjInsArg::PlainText("(".into()));
+                        args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(
+                            OperandType::cpu_rs.disassemble(&instruction, None).into(),
+                        )));
+                        args.push(ObjInsArg::PlainText(")".into()));
+                    }
+                    _ => {
+                        args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(
+                            op.disassemble(&instruction, None).into(),
+                        )));
+                    }
+                }
+            }
+            let line = line_info.range(..=cur_addr as u64).last().map(|(_, &b)| b);
+            insts.push(ObjIns {
+                address: cur_addr as u64,
+                size: 4,
+                op,
+                mnemonic,
+                args,
+                reloc: reloc.cloned(),
+                branch_dest,
+                line,
+                formatted,
+                orig: None,
+            });
+            cur_addr += 4;
+        }
+        Ok(ProcessCodeResult { ops, insts })
+    }
+
+    fn implcit_addend(
+        &self,
+        section: &ObjSection,
+        address: u64,
+        reloc: &Relocation,
+    ) -> Result<i64> {
+        let data = section.data[address as usize..address as usize + 4].try_into()?;
+        let addend = self.endianness.read_u32_bytes(data);
+        Ok(match reloc.flags() {
+            RelocationFlags::Elf { r_type: elf::R_MIPS_32 } => addend as i64,
+            RelocationFlags::Elf { r_type: elf::R_MIPS_HI16 } => {
+                ((addend & 0x0000FFFF) << 16) as i32 as i64
+            }
+            RelocationFlags::Elf {
+                r_type:
+                    elf::R_MIPS_LO16 | elf::R_MIPS_GOT16 | elf::R_MIPS_CALL16 | elf::R_MIPS_GPREL16,
+            } => (addend & 0x0000FFFF) as i16 as i64,
+            RelocationFlags::Elf { r_type: elf::R_MIPS_26 } => ((addend & 0x03FFFFFF) << 2) as i64,
+            flags => bail!("Unsupported MIPS implicit relocation {flags:?}"),
+        })
+    }
+
+    fn display_reloc(&self, flags: RelocationFlags) -> Cow<'static, str> {
+        match flags {
+            RelocationFlags::Elf { r_type } => match r_type {
+                elf::R_MIPS_HI16 => Cow::Borrowed("R_MIPS_HI16"),
+                elf::R_MIPS_LO16 => Cow::Borrowed("R_MIPS_LO16"),
+                elf::R_MIPS_GOT16 => Cow::Borrowed("R_MIPS_GOT16"),
+                elf::R_MIPS_CALL16 => Cow::Borrowed("R_MIPS_CALL16"),
+                elf::R_MIPS_GPREL16 => Cow::Borrowed("R_MIPS_GPREL16"),
+                elf::R_MIPS_32 => Cow::Borrowed("R_MIPS_32"),
+                elf::R_MIPS_26 => Cow::Borrowed("R_MIPS_26"),
+                _ => Cow::Owned(format!("<{flags:?}>")),
+            },
+            _ => Cow::Owned(format!("<{flags:?}>")),
+        }
+    }
+}
+
+fn push_reloc(args: &mut Vec<ObjInsArg>, reloc: &ObjReloc) -> Result<()> {
+    match reloc.flags {
+        RelocationFlags::Elf { r_type } => match r_type {
+            elf::R_MIPS_HI16 => {
+                args.push(ObjInsArg::PlainText("%hi(".into()));
+                args.push(ObjInsArg::Reloc);
+                args.push(ObjInsArg::PlainText(")".into()));
+            }
+            elf::R_MIPS_LO16 => {
+                args.push(ObjInsArg::PlainText("%lo(".into()));
+                args.push(ObjInsArg::Reloc);
+                args.push(ObjInsArg::PlainText(")".into()));
+            }
+            elf::R_MIPS_GOT16 => {
+                args.push(ObjInsArg::PlainText("%got(".into()));
+                args.push(ObjInsArg::Reloc);
+                args.push(ObjInsArg::PlainText(")".into()));
+            }
+            elf::R_MIPS_CALL16 => {
+                args.push(ObjInsArg::PlainText("%call16(".into()));
+                args.push(ObjInsArg::Reloc);
+                args.push(ObjInsArg::PlainText(")".into()));
+            }
+            elf::R_MIPS_GPREL16 => {
+                args.push(ObjInsArg::PlainText("%gp_rel(".into()));
+                args.push(ObjInsArg::Reloc);
+                args.push(ObjInsArg::PlainText(")".into()));
+            }
+            elf::R_MIPS_32 | elf::R_MIPS_26 => {
+                args.push(ObjInsArg::Reloc);
+            }
+            _ => bail!("Unsupported ELF MIPS relocation type {r_type}"),
+        },
+        flags => panic!("Unsupported MIPS relocation flags {flags:?}"),
+    }
+    Ok(())
+}

objdiff-core/src/arch/mod.rs

@@ -0,0 +1,58 @@
+use std::{borrow::Cow, collections::BTreeMap};
+
+use anyhow::{bail, Result};
+use object::{Architecture, Object, ObjectSymbol, Relocation, RelocationFlags, Symbol};
+
+use crate::{
+    diff::DiffObjConfig,
+    obj::{ObjIns, ObjReloc, ObjSection},
+};
+
+#[cfg(feature = "arm")]
+mod arm;
+#[cfg(feature = "mips")]
+pub mod mips;
+#[cfg(feature = "ppc")]
+pub mod ppc;
+#[cfg(feature = "x86")]
+pub mod x86;
+
+pub trait ObjArch: Send + Sync {
+    fn process_code(
+        &self,
+        address: u64,
+        code: &[u8],
+        section_index: usize,
+        relocations: &[ObjReloc],
+        line_info: &BTreeMap<u64, u64>,
+        config: &DiffObjConfig,
+    ) -> Result<ProcessCodeResult>;
+
+    fn implcit_addend(&self, section: &ObjSection, address: u64, reloc: &Relocation)
+        -> Result<i64>;
+
+    fn demangle(&self, _name: &str) -> Option<String> { None }
+
+    fn display_reloc(&self, flags: RelocationFlags) -> Cow<'static, str>;
+
+    fn symbol_address(&self, symbol: &Symbol) -> u64 { symbol.address() }
+}
+
+pub struct ProcessCodeResult {
+    pub ops: Vec<u16>,
+    pub insts: Vec<ObjIns>,
+}
+
+pub fn new_arch(object: &object::File) -> Result<Box<dyn ObjArch>> {
+    Ok(match object.architecture() {
+        #[cfg(feature = "ppc")]
+        Architecture::PowerPc => Box::new(ppc::ObjArchPpc::new(object)?),
+        #[cfg(feature = "mips")]
+        Architecture::Mips => Box::new(mips::ObjArchMips::new(object)?),
+        #[cfg(feature = "x86")]
+        Architecture::I386 | Architecture::X86_64 => Box::new(x86::ObjArchX86::new(object)?),
+        #[cfg(feature = "arm")]
+        Architecture::Arm => Box::new(arm::ObjArchArm::new(object)?),
+        arch => bail!("Unsupported architecture: {arch:?}"),
+    })
+}

objdiff-core/src/arch/ppc.rs

@@ -0,0 +1,209 @@
+use std::{borrow::Cow, collections::BTreeMap};
+
+use anyhow::{bail, Result};
+use object::{elf, File, Relocation, RelocationFlags};
+use ppc750cl::{Argument, InsIter, GPR};
+
+use crate::{
+    arch::{ObjArch, ProcessCodeResult},
+    diff::DiffObjConfig,
+    obj::{ObjIns, ObjInsArg, ObjInsArgValue, ObjReloc, ObjSection},
+};
+
+// Relative relocation, can be Simm, Offset or BranchDest
+fn is_relative_arg(arg: &Argument) -> bool {
+    matches!(arg, Argument::Simm(_) | Argument::Offset(_) | Argument::BranchDest(_))
+}
+
+// Relative or absolute relocation, can be Uimm, Simm or Offset
+fn is_rel_abs_arg(arg: &Argument) -> bool {
+    matches!(arg, Argument::Uimm(_) | Argument::Simm(_) | Argument::Offset(_))
+}
+
+fn is_offset_arg(arg: &Argument) -> bool { matches!(arg, Argument::Offset(_)) }
+
+pub struct ObjArchPpc {}
+
+impl ObjArchPpc {
+    pub fn new(_file: &File) -> Result<Self> { Ok(Self {}) }
+}
+
+impl ObjArch for ObjArchPpc {
+    fn process_code(
+        &self,
+        address: u64,
+        code: &[u8],
+        _section_index: usize,
+        relocations: &[ObjReloc],
+        line_info: &BTreeMap<u64, u64>,
+        config: &DiffObjConfig,
+    ) -> Result<ProcessCodeResult> {
+        let ins_count = code.len() / 4;
+        let mut ops = Vec::<u16>::with_capacity(ins_count);
+        let mut insts = Vec::<ObjIns>::with_capacity(ins_count);
+        for (cur_addr, mut ins) in InsIter::new(code, address as u32) {
+            let reloc = relocations.iter().find(|r| (r.address as u32 & !3) == cur_addr);
+            if let Some(reloc) = reloc {
+                // Zero out relocations
+                ins.code = match reloc.flags {
+                    RelocationFlags::Elf { r_type: elf::R_PPC_EMB_SDA21 } => ins.code & !0x1FFFFF,
+                    RelocationFlags::Elf { r_type: elf::R_PPC_REL24 } => ins.code & !0x3FFFFFC,
+                    RelocationFlags::Elf { r_type: elf::R_PPC_REL14 } => ins.code & !0xFFFC,
+                    RelocationFlags::Elf {
+                        r_type: elf::R_PPC_ADDR16_HI | elf::R_PPC_ADDR16_HA | elf::R_PPC_ADDR16_LO,
+                    } => ins.code & !0xFFFF,
+                    _ => ins.code,
+                };
+            }
+
+            let orig = ins.basic().to_string();
+            let simplified = ins.simplified();
+            let formatted = simplified.to_string();
+
+            let mut reloc_arg = None;
+            if let Some(reloc) = reloc {
+                match reloc.flags {
+                    RelocationFlags::Elf { r_type: elf::R_PPC_EMB_SDA21 } => {
+                        reloc_arg = Some(1);
+                    }
+                    RelocationFlags::Elf { r_type: elf::R_PPC_REL24 | elf::R_PPC_REL14 } => {
+                        reloc_arg = simplified.args.iter().rposition(is_relative_arg);
+                    }
+                    RelocationFlags::Elf {
+                        r_type: elf::R_PPC_ADDR16_HI | elf::R_PPC_ADDR16_HA | elf::R_PPC_ADDR16_LO,
+                    } => {
+                        reloc_arg = simplified.args.iter().rposition(is_rel_abs_arg);
+                    }
+                    _ => {}
+                }
+            }
+
+            let mut args = vec![];
+            let mut branch_dest = None;
+            let mut writing_offset = false;
+            for (idx, arg) in simplified.args_iter().enumerate() {
+                if idx > 0 && !writing_offset {
+                    args.push(ObjInsArg::PlainText(config.separator().into()));
+                }
+
+                if reloc_arg == Some(idx) {
+                    let reloc = reloc.unwrap();
+                    push_reloc(&mut args, reloc)?;
+                    // For @sda21, we can omit the register argument
+                    if matches!(reloc.flags, RelocationFlags::Elf { r_type: elf::R_PPC_EMB_SDA21 })
+                        // Sanity check: the next argument should be r0
+                        && matches!(simplified.args.get(idx + 1), Some(Argument::GPR(GPR(0))))
+                    {
+                        break;
+                    }
+                } else {
+                    match arg {
+                        Argument::Simm(simm) => {
+                            args.push(ObjInsArg::Arg(ObjInsArgValue::Signed(simm.0 as i64)));
+                        }
+                        Argument::Uimm(uimm) => {
+                            args.push(ObjInsArg::Arg(ObjInsArgValue::Unsigned(uimm.0 as u64)));
+                        }
+                        Argument::Offset(offset) => {
+                            args.push(ObjInsArg::Arg(ObjInsArgValue::Signed(offset.0 as i64)));
+                        }
+                        Argument::BranchDest(dest) => {
+                            let dest = cur_addr.wrapping_add_signed(dest.0) as u64;
+                            args.push(ObjInsArg::BranchDest(dest));
+                            branch_dest = Some(dest);
+                        }
+                        _ => {
+                            args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(
+                                arg.to_string().into(),
+                            )));
+                        }
+                    };
+                }
+
+                if writing_offset {
+                    args.push(ObjInsArg::PlainText(")".into()));
+                    writing_offset = false;
+                }
+                if is_offset_arg(arg) {
+                    args.push(ObjInsArg::PlainText("(".into()));
+                    writing_offset = true;
+                }
+            }
+
+            ops.push(ins.op as u16);
+            let line = line_info.range(..=cur_addr as u64).last().map(|(_, &b)| b);
+            insts.push(ObjIns {
+                address: cur_addr as u64,
+                size: 4,
+                mnemonic: simplified.mnemonic.to_string(),
+                args,
+                reloc: reloc.cloned(),
+                op: ins.op as u16,
+                branch_dest,
+                line,
+                formatted,
+                orig: Some(orig),
+            });
+        }
+        Ok(ProcessCodeResult { ops, insts })
+    }
+
+    fn implcit_addend(
+        &self,
+        _section: &ObjSection,
+        address: u64,
+        reloc: &Relocation,
+    ) -> Result<i64> {
+        bail!("Unsupported PPC implicit relocation {:#x}:{:?}", address, reloc.flags())
+    }
+
+    fn demangle(&self, name: &str) -> Option<String> {
+        cwdemangle::demangle(name, &cwdemangle::DemangleOptions::default())
+    }
+
+    fn display_reloc(&self, flags: RelocationFlags) -> Cow<'static, str> {
+        match flags {
+            RelocationFlags::Elf { r_type } => match r_type {
+                elf::R_PPC_ADDR16_LO => Cow::Borrowed("R_PPC_ADDR16_LO"),
+                elf::R_PPC_ADDR16_HI => Cow::Borrowed("R_PPC_ADDR16_HI"),
+                elf::R_PPC_ADDR16_HA => Cow::Borrowed("R_PPC_ADDR16_HA"),
+                elf::R_PPC_EMB_SDA21 => Cow::Borrowed("R_PPC_EMB_SDA21"),
+                elf::R_PPC_ADDR32 => Cow::Borrowed("R_PPC_ADDR32"),
+                elf::R_PPC_UADDR32 => Cow::Borrowed("R_PPC_UADDR32"),
+                elf::R_PPC_REL24 => Cow::Borrowed("R_PPC_REL24"),
+                elf::R_PPC_REL14 => Cow::Borrowed("R_PPC_REL14"),
+                _ => Cow::Owned(format!("<{flags:?}>")),
+            },
+            _ => Cow::Owned(format!("<{flags:?}>")),
+        }
+    }
+}
+
+fn push_reloc(args: &mut Vec<ObjInsArg>, reloc: &ObjReloc) -> Result<()> {
+    match reloc.flags {
+        RelocationFlags::Elf { r_type } => match r_type {
+            elf::R_PPC_ADDR16_LO => {
+                args.push(ObjInsArg::Reloc);
+                args.push(ObjInsArg::PlainText("@l".into()));
+            }
+            elf::R_PPC_ADDR16_HI => {
+                args.push(ObjInsArg::Reloc);
+                args.push(ObjInsArg::PlainText("@h".into()));
+            }
+            elf::R_PPC_ADDR16_HA => {
+                args.push(ObjInsArg::Reloc);
+                args.push(ObjInsArg::PlainText("@ha".into()));
+            }
+            elf::R_PPC_EMB_SDA21 => {
+                args.push(ObjInsArg::Reloc);
+                args.push(ObjInsArg::PlainText("@sda21".into()));
+            }
+            elf::R_PPC_ADDR32 | elf::R_PPC_UADDR32 | elf::R_PPC_REL24 | elf::R_PPC_REL14 => {
+                args.push(ObjInsArg::Reloc);
+            }
+            _ => bail!("Unsupported ELF PPC relocation type {r_type}"),
+        },
+        flags => bail!("Unsupported PPC relocation kind: {flags:?}"),
+    };
+    Ok(())
+}

objdiff-core/src/arch/x86.rs

@@ -0,0 +1,342 @@
+use std::{borrow::Cow, collections::BTreeMap};
+
+use anyhow::{anyhow, bail, ensure, Result};
+use iced_x86::{
+    Decoder, DecoderOptions, DecoratorKind, Formatter, FormatterOutput, FormatterTextKind,
+    GasFormatter, Instruction, IntelFormatter, MasmFormatter, NasmFormatter, NumberKind, OpKind,
+    PrefixKind, Register,
+};
+use object::{pe, Endian, Endianness, File, Object, Relocation, RelocationFlags};
+
+use crate::{
+    arch::{ObjArch, ProcessCodeResult},
+    diff::{DiffObjConfig, X86Formatter},
+    obj::{ObjIns, ObjInsArg, ObjInsArgValue, ObjReloc, ObjSection},
+};
+
+pub struct ObjArchX86 {
+    bits: u32,
+    endianness: Endianness,
+}
+
+impl ObjArchX86 {
+    pub fn new(object: &File) -> Result<Self> {
+        Ok(Self { bits: if object.is_64() { 64 } else { 32 }, endianness: object.endianness() })
+    }
+}
+
+impl ObjArch for ObjArchX86 {
+    fn process_code(
+        &self,
+        address: u64,
+        code: &[u8],
+        _section_index: usize,
+        relocations: &[ObjReloc],
+        line_info: &BTreeMap<u64, u64>,
+        config: &DiffObjConfig,
+    ) -> Result<ProcessCodeResult> {
+        let mut result = ProcessCodeResult { ops: Vec::new(), insts: Vec::new() };
+        let mut decoder = Decoder::with_ip(self.bits, code, address, DecoderOptions::NONE);
+        let mut formatter: Box<dyn Formatter> = match config.x86_formatter {
+            X86Formatter::Intel => Box::new(IntelFormatter::new()),
+            X86Formatter::Gas => Box::new(GasFormatter::new()),
+            X86Formatter::Nasm => Box::new(NasmFormatter::new()),
+            X86Formatter::Masm => Box::new(MasmFormatter::new()),
+        };
+        formatter.options_mut().set_space_after_operand_separator(config.space_between_args);
+
+        let mut output = InstructionFormatterOutput {
+            formatted: String::new(),
+            ins: ObjIns {
+                address: 0,
+                size: 0,
+                op: 0,
+                mnemonic: String::new(),
+                args: vec![],
+                reloc: None,
+                branch_dest: None,
+                line: None,
+                formatted: String::new(),
+                orig: None,
+            },
+            error: None,
+            ins_operands: vec![],
+        };
+        let mut instruction = Instruction::default();
+        while decoder.can_decode() {
+            decoder.decode_out(&mut instruction);
+
+            let address = instruction.ip();
+            let op = instruction.mnemonic() as u16;
+            let reloc = relocations
+                .iter()
+                .find(|r| r.address >= address && r.address < address + instruction.len() as u64);
+            let line = line_info.range(..=address).last().map(|(_, &b)| b);
+            output.ins = ObjIns {
+                address,
+                size: instruction.len() as u8,
+                op,
+                mnemonic: String::new(),
+                args: vec![],
+                reloc: reloc.cloned(),
+                branch_dest: None,
+                line,
+                formatted: String::new(),
+                orig: None,
+            };
+            // Run the formatter, which will populate output.ins
+            formatter.format(&instruction, &mut output);
+            if let Some(error) = output.error.take() {
+                return Err(error);
+            }
+            ensure!(output.ins_operands.len() == output.ins.args.len());
+            output.ins.formatted.clone_from(&output.formatted);
+
+            // Make sure we've put the relocation somewhere in the instruction
+            if reloc.is_some() && !output.ins.args.iter().any(|a| matches!(a, ObjInsArg::Reloc)) {
+                let mut found = replace_arg(
+                    OpKind::Memory,
+                    ObjInsArg::Reloc,
+                    &mut output.ins.args,
+                    &instruction,
+                    &output.ins_operands,
+                )?;
+                if !found {
+                    found = replace_arg(
+                        OpKind::Immediate32,
+                        ObjInsArg::Reloc,
+                        &mut output.ins.args,
+                        &instruction,
+                        &output.ins_operands,
+                    )?;
+                }
+                ensure!(found, "x86: Failed to find operand for Absolute relocation");
+            }
+            if reloc.is_some() && !output.ins.args.iter().any(|a| matches!(a, ObjInsArg::Reloc)) {
+                bail!("Failed to find relocation in instruction");
+            }
+
+            result.ops.push(op);
+            result.insts.push(output.ins.clone());
+
+            // Clear for next iteration
+            output.formatted.clear();
+            output.ins_operands.clear();
+        }
+        Ok(result)
+    }
+
+    fn implcit_addend(
+        &self,
+        section: &ObjSection,
+        address: u64,
+        reloc: &Relocation,
+    ) -> Result<i64> {
+        match reloc.flags() {
+            RelocationFlags::Coff { typ: pe::IMAGE_REL_I386_DIR32 | pe::IMAGE_REL_I386_REL32 } => {
+                let data = section.data[address as usize..address as usize + 4].try_into()?;
+                Ok(self.endianness.read_i32_bytes(data) as i64)
+            }
+            flags => bail!("Unsupported x86 implicit relocation {flags:?}"),
+        }
+    }
+
+    fn demangle(&self, name: &str) -> Option<String> {
+        if name.starts_with('?') {
+            msvc_demangler::demangle(name, msvc_demangler::DemangleFlags::llvm()).ok()
+        } else {
+            cpp_demangle::Symbol::new(name)
+                .ok()
+                .and_then(|s| s.demangle(&cpp_demangle::DemangleOptions::default()).ok())
+        }
+    }
+
+    fn display_reloc(&self, flags: RelocationFlags) -> Cow<'static, str> {
+        match flags {
+            RelocationFlags::Coff { typ } => match typ {
+                pe::IMAGE_REL_I386_DIR32 => Cow::Borrowed("IMAGE_REL_I386_DIR32"),
+                pe::IMAGE_REL_I386_REL32 => Cow::Borrowed("IMAGE_REL_I386_REL32"),
+                _ => Cow::Owned(format!("<{flags:?}>")),
+            },
+            _ => Cow::Owned(format!("<{flags:?}>")),
+        }
+    }
+}
+
+fn replace_arg(
+    from: OpKind,
+    to: ObjInsArg,
+    args: &mut [ObjInsArg],
+    instruction: &Instruction,
+    ins_operands: &[Option<u32>],
+) -> Result<bool> {
+    let mut replace = None;
+    for i in 0..instruction.op_count() {
+        let op_kind = instruction.op_kind(i);
+        if op_kind == from {
+            replace = Some(i);
+            break;
+        }
+    }
+    if let Some(i) = replace {
+        for (j, arg) in args.iter_mut().enumerate() {
+            if ins_operands[j] == Some(i) {
+                *arg = to;
+                return Ok(true);
+            }
+        }
+    }
+    Ok(false)
+}
+
+struct InstructionFormatterOutput {
+    formatted: String,
+    ins: ObjIns,
+    error: Option<anyhow::Error>,
+    ins_operands: Vec<Option<u32>>,
+}
+
+impl InstructionFormatterOutput {
+    fn push_signed(&mut self, value: i64) {
+        // The formatter writes the '-' operator and then gives us a negative value,
+        // so convert it to a positive value to avoid double negatives
+        if value < 0
+            && matches!(self.ins.args.last(), Some(ObjInsArg::Arg(ObjInsArgValue::Opaque(v))) if v == "-")
+        {
+            self.ins.args.push(ObjInsArg::Arg(ObjInsArgValue::Signed(value.wrapping_abs())));
+        } else {
+            self.ins.args.push(ObjInsArg::Arg(ObjInsArgValue::Signed(value)));
+        }
+    }
+}
+
+impl FormatterOutput for InstructionFormatterOutput {
+    fn write(&mut self, text: &str, kind: FormatterTextKind) {
+        self.formatted.push_str(text);
+        // Skip whitespace after the mnemonic
+        if self.ins.args.is_empty() && kind == FormatterTextKind::Text {
+            return;
+        }
+        self.ins_operands.push(None);
+        match kind {
+            FormatterTextKind::Text | FormatterTextKind::Punctuation => {
+                self.ins.args.push(ObjInsArg::PlainText(text.to_string().into()));
+            }
+            FormatterTextKind::Keyword | FormatterTextKind::Operator => {
+                self.ins.args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(text.to_string().into())));
+            }
+            _ => {
+                if self.error.is_none() {
+                    self.error = Some(anyhow!("x86: Unsupported FormatterTextKind {:?}", kind));
+                }
+            }
+        }
+    }
+
+    fn write_prefix(&mut self, _instruction: &Instruction, text: &str, _prefix: PrefixKind) {
+        self.formatted.push_str(text);
+        self.ins_operands.push(None);
+        self.ins.args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(text.to_string().into())));
+    }
+
+    fn write_mnemonic(&mut self, _instruction: &Instruction, text: &str) {
+        self.formatted.push_str(text);
+        self.ins.mnemonic = text.to_string();
+    }
+
+    fn write_number(
+        &mut self,
+        _instruction: &Instruction,
+        _operand: u32,
+        instruction_operand: Option<u32>,
+        text: &str,
+        value: u64,
+        number_kind: NumberKind,
+        kind: FormatterTextKind,
+    ) {
+        self.formatted.push_str(text);
+        self.ins_operands.push(instruction_operand);
+
+        // Handle relocations
+        match kind {
+            FormatterTextKind::LabelAddress => {
+                if let Some(reloc) = self.ins.reloc.as_ref() {
+                    if matches!(reloc.flags, RelocationFlags::Coff {
+                        typ: pe::IMAGE_REL_I386_DIR32 | pe::IMAGE_REL_I386_REL32
+                    }) {
+                        self.ins.args.push(ObjInsArg::Reloc);
+                        return;
+                    } else if self.error.is_none() {
+                        self.error = Some(anyhow!(
+                            "x86: Unsupported LabelAddress relocation flags {:?}",
+                            reloc.flags
+                        ));
+                    }
+                }
+                self.ins.args.push(ObjInsArg::BranchDest(value));
+                self.ins.branch_dest = Some(value);
+                return;
+            }
+            FormatterTextKind::FunctionAddress => {
+                if let Some(reloc) = self.ins.reloc.as_ref() {
+                    if matches!(reloc.flags, RelocationFlags::Coff {
+                        typ: pe::IMAGE_REL_I386_REL32
+                    }) {
+                        self.ins.args.push(ObjInsArg::Reloc);
+                        return;
+                    } else if self.error.is_none() {
+                        self.error = Some(anyhow!(
+                            "x86: Unsupported FunctionAddress relocation flags {:?}",
+                            reloc.flags
+                        ));
+                    }
+                }
+            }
+            _ => {}
+        }
+
+        match number_kind {
+            NumberKind::Int8 => {
+                self.push_signed(value as i8 as i64);
+            }
+            NumberKind::Int16 => {
+                self.push_signed(value as i16 as i64);
+            }
+            NumberKind::Int32 => {
+                self.push_signed(value as i32 as i64);
+            }
+            NumberKind::Int64 => {
+                self.push_signed(value as i64);
+            }
+            NumberKind::UInt8 | NumberKind::UInt16 | NumberKind::UInt32 | NumberKind::UInt64 => {
+                self.ins.args.push(ObjInsArg::Arg(ObjInsArgValue::Unsigned(value)));
+            }
+        }
+    }
+
+    fn write_decorator(
+        &mut self,
+        _instruction: &Instruction,
+        _operand: u32,
+        instruction_operand: Option<u32>,
+        text: &str,
+        _decorator: DecoratorKind,
+    ) {
+        self.formatted.push_str(text);
+        self.ins_operands.push(instruction_operand);
+        self.ins.args.push(ObjInsArg::PlainText(text.to_string().into()));
+    }
+
+    fn write_register(
+        &mut self,
+        _instruction: &Instruction,
+        _operand: u32,
+        instruction_operand: Option<u32>,
+        text: &str,
+        _register: Register,
+    ) {
+        self.formatted.push_str(text);
+        self.ins_operands.push(instruction_operand);
+        self.ins.args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(text.to_string().into())));
+    }
+}