Support R_MIPS_LITERAL, R_MIPS15_S3 relocations

Resolves #92
Resolves #95

.gitignore

@@ -22,3 +22,4 @@ android.keystore
 *.frag
 *.vert
 *.metal
+.vscode/launch.json

Cargo.lock

@@ -110,9 +110,9 @@ dependencies = [
 
 [[package]]
 name = "anyhow"
-version = "1.0.82"
+version = "1.0.86"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f538837af36e6f6a9be0faa67f9a314f8119e4e4b5867c6ab40ed60360142519"
+checksum = "b3d1d046238990b9cf5bcde22a3fb3584ee5cf65fb2765f454ed428c7a0063da"
 
 [[package]]
 name = "arboard"
@@ -524,9 +524,9 @@ checksum = "1fd0f2584146f6f2ef48085050886acf353beff7305ebd1ae69500e27c67f64b"
 
 [[package]]
 name = "bytes"
-version = "1.6.0"
+version = "1.6.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "514de17de45fdb8dc022b1a7975556c53c86f9f0aa5f534b98977b171857c2c9"
+checksum = "a12916984aab3fa6e39d655a33e09c0071eb36d6ab3aea5c2d78551f1df6d952"
 
 [[package]]
 name = "calloop"
@@ -970,6 +970,15 @@ version = "1.0.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "c2e06f9bce634a3c898eb1e5cb949ff63133cbb218af93cc9b38b31d6f3ea285"
 
+[[package]]
+name = "cwextab"
+version = "0.2.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "a92e840be31d11ead5f357b8fc503133d3657845d0ccf539afcc7d1212fad226"
+dependencies = [
+ "anyhow",
+]
+
 [[package]]
 name = "d3d12"
 version = "0.19.0"
@@ -1466,6 +1475,12 @@ dependencies = [
  "windows-sys 0.52.0",
 ]
 
+[[package]]
+name = "fixedbitset"
+version = "0.4.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "0ce7134b9999ecaf8bcd65542e436736ef32ddca1b3e06094cb6ec5755203b80"
+
 [[package]]
 name = "flagset"
 version = "0.4.5"
@@ -1922,6 +1937,12 @@ version = "0.4.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "95505c38b4572b2d910cecb0281560f54b440a19336cbbcb27bf6ce6adc6f5a8"
 
+[[package]]
+name = "heck"
+version = "0.5.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "2304e00983f87ffb38b55b444b5e3b60a884b5d30c0fca7d82fe33449bbe55ea"
+
 [[package]]
 name = "hermit-abi"
 version = "0.3.9"
@@ -2183,6 +2204,15 @@ dependencies = [
  "either",
 ]
 
+[[package]]
+name = "itertools"
+version = "0.13.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "413ee7dfc52ee1a4949ceeb7dbc8a33f2d6c088194d9f922fb8318faf1f01186"
+dependencies = [
+ "either",
+]
+
 [[package]]
 name = "itoa"
 version = "1.0.11"
@@ -2274,9 +2304,9 @@ checksum = "e2abad23fbc42b3700f2f279844dc832adb2b2eb069b2df918f455c4e18cc646"
 
 [[package]]
 name = "libc"
-version = "0.2.154"
+version = "0.2.155"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "ae743338b92ff9146ce83992f766a31066a91a8c84a45e0e9f21e7cf6de6d346"
+checksum = "97b3888a4aecf77e811145cadf6eef5901f4782c53886191b2f693f24761847c"
 
 [[package]]
 name = "libloading"
@@ -2470,6 +2500,12 @@ dependencies = [
  "bitflags 2.5.0",
 ]
 
+[[package]]
+name = "multimap"
+version = "0.10.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "defc4c55412d89136f966bbb339008b474350e5e6e78d2714439c386b3137a03"
+
 [[package]]
 name = "naga"
 version = "0.19.2"
@@ -2789,13 +2825,18 @@ dependencies = [
 
 [[package]]
 name = "objdiff-cli"
-version = "2.0.0-alpha.5"
+version = "2.0.0-beta.4"
 dependencies = [
  "anyhow",
  "argp",
  "crossterm",
  "enable-ansi-support",
+ "memmap2",
  "objdiff-core",
+ "pbjson",
+ "pbjson-build",
+ "prost",
+ "prost-build",
  "ratatui",
  "rayon",
  "serde",
@@ -2808,13 +2849,14 @@ dependencies = [
 
 [[package]]
 name = "objdiff-core"
-version = "2.0.0-alpha.5"
+version = "2.0.0-beta.4"
 dependencies = [
  "anyhow",
  "arm-attr",
  "byteorder",
  "cpp_demangle",
  "cwdemangle",
+ "cwextab",
  "filetime",
  "flagset",
  "gimli 0.29.0",
@@ -2838,7 +2880,7 @@ dependencies = [
 
 [[package]]
 name = "objdiff-gui"
-version = "2.0.0-alpha.5"
+version = "2.0.0-beta.4"
 dependencies = [
  "anyhow",
  "bytes",
@@ -2846,6 +2888,7 @@ dependencies = [
  "console_error_panic_hook",
  "const_format",
  "cwdemangle",
+ "cwextab",
  "dirs",
  "eframe",
  "egui",
@@ -2861,8 +2904,10 @@ dependencies = [
  "path-slash",
  "png",
  "pollster",
+ "regex",
  "reqwest",
  "rfd",
+ "rlwinmdec",
  "ron",
  "self_update",
  "serde",
@@ -2905,9 +2950,9 @@ checksum = "3fdb12b2476b595f9358c5161aa467c2438859caa136dec86c26fdd2efe17b92"
 
 [[package]]
 name = "openssl"
-version = "0.10.64"
+version = "0.10.66"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "95a0481286a310808298130d22dd1fef0fa571e05a8f44ec801801e84b216b1f"
+checksum = "9529f4786b70a3e8c61e11179af17ab6188ad8d0ded78c5529441ed39d4bd9c1"
 dependencies = [
  "bitflags 2.5.0",
  "cfg-if",
@@ -2937,9 +2982,9 @@ checksum = "ff011a302c396a5197692431fc1948019154afc178baf7d8e37367442a4601cf"
 
 [[package]]
 name = "openssl-sys"
-version = "0.9.102"
+version = "0.9.103"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c597637d56fbc83893a35eb0dd04b2b8e7a50c91e64e9493e398b5df4fb45fa2"
+checksum = "7f9e8deee91df40a943c71b917e5874b951d32a802526c85721ce3b776c929d6"
 dependencies = [
  "cc",
  "libc",
@@ -3047,12 +3092,44 @@ dependencies = [
  "rustc_version",
 ]
 
+[[package]]
+name = "pbjson"
+version = "0.7.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "c7e6349fa080353f4a597daffd05cb81572a9c031a6d4fff7e504947496fcc68"
+dependencies = [
+ "base64 0.21.7",
+ "serde",
+]
+
+[[package]]
+name = "pbjson-build"
+version = "0.7.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "6eea3058763d6e656105d1403cb04e0a41b7bbac6362d413e7c33be0c32279c9"
+dependencies = [
+ "heck 0.5.0",
+ "itertools 0.13.0",
+ "prost",
+ "prost-types",
+]
+
 [[package]]
 name = "percent-encoding"
 version = "2.3.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "e3148f5046208a5d56bcfc03053e3ca6334e51da8dfb19b6cdc8b306fae3283e"
 
+[[package]]
+name = "petgraph"
+version = "0.6.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b4c5cc86750666a3ed20bdaf5ca2a0344f9c67674cae0515bec2da16fbaa47db"
+dependencies = [
+ "fixedbitset",
+ "indexmap",
+]
+
 [[package]]
 name = "pin-project"
 version = "1.1.5"
@@ -3165,6 +3242,16 @@ version = "0.3.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "e8cf8e6a8aa66ce33f63993ffc4ea4271eb5b0530a9002db8455ea6050c77bfa"
 
+[[package]]
+name = "prettyplease"
+version = "0.2.20"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5f12335488a2f3b0a83b14edad48dca9879ce89b2edd10e80237e4e852dd645e"
+dependencies = [
+ "proc-macro2",
+ "syn 2.0.60",
+]
+
 [[package]]
 name = "proc-macro-crate"
 version = "1.3.1"
@@ -3199,6 +3286,59 @@ version = "1.0.15"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "43d84d1d7a6ac92673717f9f6d1518374ef257669c24ebc5ac25d5033828be58"
 
+[[package]]
+name = "prost"
+version = "0.13.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e13db3d3fde688c61e2446b4d843bc27a7e8af269a69440c0308021dc92333cc"
+dependencies = [
+ "bytes",
+ "prost-derive",
+]
+
+[[package]]
+name = "prost-build"
+version = "0.13.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5bb182580f71dd070f88d01ce3de9f4da5021db7115d2e1c3605a754153b77c1"
+dependencies = [
+ "bytes",
+ "heck 0.5.0",
+ "itertools 0.13.0",
+ "log",
+ "multimap",
+ "once_cell",
+ "petgraph",
+ "prettyplease",
+ "prost",
+ "prost-types",
+ "regex",
+ "syn 2.0.60",
+ "tempfile",
+]
+
+[[package]]
+name = "prost-derive"
+version = "0.13.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "18bec9b0adc4eba778b33684b7ba3e7137789434769ee3ce3930463ef904cfca"
+dependencies = [
+ "anyhow",
+ "itertools 0.13.0",
+ "proc-macro2",
+ "quote",
+ "syn 2.0.60",
+]
+
+[[package]]
+name = "prost-types"
+version = "0.13.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "cee5168b05f49d4b0ca581206eb14a7b22fafd963efe729ac48eb03266e25cc2"
+dependencies = [
+ "prost",
+]
+
 [[package]]
 name = "pulldown-cmark"
 version = "0.9.6"
@@ -3296,7 +3436,7 @@ dependencies = [
  "compact_str",
  "crossterm",
  "indoc",
- "itertools",
+ "itertools 0.12.1",
  "lru",
  "paste",
  "stability",
@@ -3377,9 +3517,9 @@ dependencies = [
 
 [[package]]
 name = "regex"
-version = "1.10.4"
+version = "1.10.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c117dbdfde9c8308975b6a18d71f3f385c89461f7b3fb054288ecf2a2058ba4c"
+checksum = "b91213439dad192326a0d7c6ee3955910425f441d7038e0d6933b0aec5c4517f"
 dependencies = [
  "aho-corasick",
  "memchr",
@@ -3509,6 +3649,15 @@ dependencies = [
  "windows-sys 0.52.0",
 ]
 
+[[package]]
+name = "rlwinmdec"
+version = "1.0.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "2076dbc187938f3db71c03c85d143febf01026631189dc8ca85f8c886d90ea12"
+dependencies = [
+ "anyhow",
+]
+
 [[package]]
 name = "ron"
 version = "0.8.1"
@@ -3966,7 +4115,7 @@ version = "0.26.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "c6cf59daf282c0a494ba14fd21610a0325f9f90ec9d1231dea26bcb1d696c946"
 dependencies = [
- "heck",
+ "heck 0.4.1",
  "proc-macro2",
  "quote",
  "rustversion",
@@ -4330,9 +4479,9 @@ dependencies = [
 
 [[package]]
 name = "unarm"
-version = "1.3.0"
+version = "1.4.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "c6bff109f0171a299559d82a7236e056093fc0dcd2a7da86aa745f82281e2d31"
+checksum = "379762d9433a2e6e498cde97801fb238318b024a513d0843eeac98b9056b9f3c"
 
 [[package]]
 name = "unicase"
@@ -4743,9 +4892,9 @@ dependencies = [
 
 [[package]]
 name = "wgpu-hal"
-version = "0.19.4"
+version = "0.19.5"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "fc1a4924366df7ab41a5d8546d6534f1f33231aa5b3f72b9930e300f254e39c3"
+checksum = "bfabcfc55fd86611a855816326b2d54c3b2fd7972c27ce414291562650552703"
 dependencies = [
  "android_system_properties",
  "arrayvec",

objdiff-cli/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "objdiff-cli"
-version = "2.0.0-alpha.5"
+version = "2.0.0-beta.4"
 edition = "2021"
 rust-version = "1.70"
 authors = ["Luke Street <luke@street.dev>"]
@@ -18,7 +18,10 @@ anyhow = "1.0.82"
 argp = "0.3.0"
 crossterm = "0.27.0"
 enable-ansi-support = "0.2.1"
+memmap2 = "0.9.4"
 objdiff-core = { path = "../objdiff-core", features = ["all"] }
+pbjson = "0.7.0"
+prost = "0.13.1"
 ratatui = "0.26.2"
 rayon = "1.10.0"
 serde = { version = "1", features = ["derive"] }
@@ -27,3 +30,8 @@ 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"] }
+
+[build-dependencies]
+prost-build = "0.13.1"
+pbjson-build = "0.7.0"
+

objdiff-cli/build.rs

@@ -1,3 +1,5 @@
+use std::path::{Path, PathBuf};
+
 fn main() {
     let output = std::process::Command::new("git")
         .args(["rev-parse", "HEAD"])
@@ -6,4 +8,55 @@ fn main() {
     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");
+
+    let root = PathBuf::from(env!("CARGO_MANIFEST_DIR")).join("protos");
+    let descriptor_path = root.join("proto_descriptor.bin");
+    println!("cargo:rerun-if-changed={}", descriptor_path.display());
+    let descriptor_mtime = std::fs::metadata(&descriptor_path)
+        .map(|m| m.modified().unwrap())
+        .unwrap_or(std::time::SystemTime::UNIX_EPOCH);
+    let mut run_protoc = false;
+    let proto_files = vec![root.join("report.proto")];
+    for proto_file in &proto_files {
+        println!("cargo:rerun-if-changed={}", proto_file.display());
+        let mtime = match std::fs::metadata(proto_file) {
+            Ok(m) => m.modified().unwrap(),
+            Err(e) => panic!("Failed to stat proto file {}: {:?}", proto_file.display(), e),
+        };
+        if mtime > descriptor_mtime {
+            run_protoc = true;
+        }
+    }
+
+    fn prost_config(descriptor_path: &Path, run_protoc: bool) -> prost_build::Config {
+        let mut config = prost_build::Config::new();
+        config.file_descriptor_set_path(descriptor_path);
+        // If our cached descriptor is up-to-date, we don't need to run protoc.
+        // This is helpful so that users don't need to have protoc installed
+        // unless they're updating the protos.
+        if !run_protoc {
+            config.skip_protoc_run();
+        }
+        config
+    }
+    if let Err(e) =
+        prost_config(&descriptor_path, run_protoc).compile_protos(&proto_files, &[root.as_path()])
+    {
+        if e.kind() == std::io::ErrorKind::NotFound && e.to_string().contains("protoc") {
+            eprintln!("protoc not found, skipping protobuf compilation");
+            prost_config(&descriptor_path, false)
+                .compile_protos(&proto_files, &[root.as_path()])
+                .expect("Failed to compile protos");
+        } else {
+            panic!("Failed to compile protos: {e:?}");
+        }
+    }
+
+    let descriptor_set = std::fs::read(descriptor_path).expect("Failed to read descriptor set");
+    pbjson_build::Builder::new()
+        .register_descriptors(&descriptor_set)
+        .expect("Failed to register descriptors")
+        .preserve_proto_field_names()
+        .build(&[".objdiff"])
+        .expect("Failed to build pbjson");
 }

objdiff-cli/protos/report.proto

@@ -0,0 +1,135 @@
+syntax = "proto3";
+
+package objdiff.report;
+
+// Progress info for a report or unit
+message Measures {
+  // Overall match percent, including partially matched functions and data
+  float fuzzy_match_percent = 1;
+  // Total size of code in bytes
+  uint64 total_code = 2;
+  // Fully matched code size in bytes
+  uint64 matched_code = 3;
+  // Fully matched code percent
+  float matched_code_percent = 4;
+  // Total size of data in bytes
+  uint64 total_data = 5;
+  // Fully matched data size in bytes
+  uint64 matched_data = 6;
+  // Fully matched data percent
+  float matched_data_percent = 7;
+  // Total number of functions
+  uint32 total_functions = 8;
+  // Fully matched functions
+  uint32 matched_functions = 9;
+  // Fully matched functions percent
+  float matched_functions_percent = 10;
+}
+
+// Project progress report
+message Report {
+  // Overall progress info
+  Measures measures = 1;
+  // Units within this report
+  repeated ReportUnit units = 2;
+}
+
+// A unit of the report (usually a translation unit)
+message ReportUnit {
+  // The name of the unit
+  string name = 1;
+  // Progress info for this unit
+  Measures measures = 2;
+  // Sections within this unit
+  repeated ReportItem sections = 3;
+  // Functions within this unit
+  repeated ReportItem functions = 4;
+  // Extra metadata for this unit
+  optional ReportUnitMetadata metadata = 5;
+}
+
+// Extra metadata for a unit
+message ReportUnitMetadata {
+  // Whether this unit is marked as complete (or "linked")
+  optional bool complete = 1;
+  // The name of the module this unit belongs to
+  optional string module_name = 2;
+  // The ID of the module this unit belongs to
+  optional uint32 module_id = 3;
+  // The path to the source file of this unit
+  optional string source_path = 4;
+}
+
+// A section or function within a unit
+message ReportItem {
+  // The name of the item
+  string name = 1;
+  // The size of the item in bytes
+  uint64 size = 2;
+  // The overall match percent for this item
+  float fuzzy_match_percent = 3;
+  // Extra metadata for this item
+  optional ReportItemMetadata metadata = 4;
+}
+
+// Extra metadata for an item
+message ReportItemMetadata {
+  // The demangled name of the function
+  optional string demangled_name = 1;
+  // The virtual address of the function or section
+  optional uint64 virtual_address = 2;
+}
+
+// A pair of reports to compare and generate changes
+message ChangesInput {
+  // The previous report
+  Report from = 1;
+  // The current report
+  Report to = 2;
+}
+
+// Changes between two reports
+message Changes {
+  // The progress info for the previous report
+  Measures from = 1;
+  // The progress info for the current report
+  Measures to = 2;
+  // Units that changed
+  repeated ChangeUnit units = 3;
+}
+
+// A changed unit
+message ChangeUnit {
+  // The name of the unit
+  string name = 1;
+  // The previous progress info (omitted if new)
+  optional Measures from = 2;
+  // The current progress info (omitted if removed)
+  optional Measures to = 3;
+  // Sections that changed
+  repeated ChangeItem sections = 4;
+  // Functions that changed
+  repeated ChangeItem functions = 5;
+  // Extra metadata for this unit
+  optional ReportUnitMetadata metadata = 6;
+}
+
+// A changed section or function
+message ChangeItem {
+  // The name of the item
+  string name = 1;
+  // The previous progress info (omitted if new)
+  optional ChangeItemInfo from = 2;
+  // The current progress info (omitted if removed)
+  optional ChangeItemInfo to = 3;
+  // Extra metadata for this item
+  optional ReportItemMetadata metadata = 4;
+}
+
+// Progress info for a section or function
+message ChangeItemInfo {
+  // The overall match percent for this item
+  float fuzzy_match_percent = 1;
+  // The size of the item in bytes
+  uint64 size = 2;
+}

objdiff-cli/src/cmd/diff.rs

@@ -750,8 +750,11 @@ impl FunctionDiffUi {
                             base_color = COLOR_ROTATION[diff.idx % COLOR_ROTATION.len()]
                         }
                     }
-                    DiffText::BranchDest(addr) => {
+                    DiffText::BranchDest(addr, diff) => {
                         label_text = format!("{addr:x}");
+                        if let Some(diff) = diff {
+                            base_color = COLOR_ROTATION[diff.idx % COLOR_ROTATION.len()]
+                        }
                     }
                     DiffText::Symbol(sym) => {
                         let name = sym.demangled_name.as_ref().unwrap_or(&sym.name);
@@ -817,6 +820,12 @@ impl FunctionDiffUi {
             mips_abi: Default::default(),            // TODO
             mips_instr_category: Default::default(), // TODO
             arm_arch_version: Default::default(),    // TODO
+            arm_unified_syntax: true,                // TODO
+            arm_av_registers: false,                 // TODO
+            arm_r9_usage: Default::default(),        // TODO
+            arm_sl_usage: false,                     // TODO
+            arm_fp_usage: false,                     // TODO
+            arm_ip_usage: false,                     // TODO
         };
         let target = self
             .target_path

objdiff-cli/src/main.rs

@@ -96,7 +96,7 @@ fn main() {
         // 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");
+        unsafe { env::set_var("IGNORE_IS_TERMINAL", "1") };
         supports_color::on(Stream::Stdout).is_some_and(|c| c.has_basic)
     };
 

objdiff-cli/src/util/mod.rs

@@ -1 +1,2 @@
+pub mod report;
 pub mod term;

objdiff-cli/src/util/report.rs

@@ -0,0 +1,226 @@
+use anyhow::{bail, Result};
+use prost::Message;
+use serde_json::error::Category;
+
+// Protobuf report types
+include!(concat!(env!("OUT_DIR"), "/objdiff.report.rs"));
+include!(concat!(env!("OUT_DIR"), "/objdiff.report.serde.rs"));
+
+impl Report {
+    pub fn parse(data: &[u8]) -> Result<Self> {
+        if data.is_empty() {
+            bail!(std::io::Error::from(std::io::ErrorKind::UnexpectedEof));
+        }
+        if data[0] == b'{' {
+            // Load as JSON
+            Self::from_json(data).map_err(anyhow::Error::new)
+        } else {
+            // Load as binary protobuf
+            Self::decode(data).map_err(anyhow::Error::new)
+        }
+    }
+
+    fn from_json(bytes: &[u8]) -> Result<Self, serde_json::Error> {
+        match serde_json::from_slice::<Self>(bytes) {
+            Ok(report) => Ok(report),
+            Err(e) => {
+                match e.classify() {
+                    Category::Io | Category::Eof | Category::Syntax => Err(e),
+                    Category::Data => {
+                        // Try to load as legacy report
+                        match serde_json::from_slice::<LegacyReport>(bytes) {
+                            Ok(legacy_report) => Ok(Report::from(legacy_report)),
+                            Err(_) => Err(e),
+                        }
+                    }
+                }
+            }
+        }
+    }
+}
+
+impl Measures {
+    /// Average the fuzzy match percentage over total code bytes.
+    pub fn calc_fuzzy_match_percent(&mut self) {
+        if self.total_code == 0 {
+            self.fuzzy_match_percent = 100.0;
+        } else {
+            self.fuzzy_match_percent /= self.total_code as f32;
+        }
+    }
+
+    /// Calculate the percentage of matched code, data, and functions.
+    pub fn calc_matched_percent(&mut self) {
+        self.matched_code_percent = if self.total_code == 0 {
+            100.0
+        } else {
+            self.matched_code as f32 / self.total_code as f32 * 100.0
+        };
+        self.matched_data_percent = if self.total_data == 0 {
+            100.0
+        } else {
+            self.matched_data as f32 / self.total_data as f32 * 100.0
+        };
+        self.matched_functions_percent = if self.total_functions == 0 {
+            100.0
+        } else {
+            self.matched_functions as f32 / self.total_functions as f32 * 100.0
+        };
+    }
+}
+
+/// Allows [collect](Iterator::collect) to be used on an iterator of [Measures].
+impl FromIterator<Measures> for Measures {
+    fn from_iter<T>(iter: T) -> Self
+    where T: IntoIterator<Item = Measures> {
+        let mut measures = Measures::default();
+        for other in iter {
+            measures.fuzzy_match_percent += other.fuzzy_match_percent * other.total_code as f32;
+            measures.total_code += other.total_code;
+            measures.matched_code += other.matched_code;
+            measures.total_data += other.total_data;
+            measures.matched_data += other.matched_data;
+            measures.total_functions += other.total_functions;
+            measures.matched_functions += other.matched_functions;
+        }
+        measures.calc_fuzzy_match_percent();
+        measures.calc_matched_percent();
+        measures
+    }
+}
+
+// Older JSON report types
+#[derive(Debug, Clone, Default, serde::Serialize, serde::Deserialize)]
+struct LegacyReport {
+    fuzzy_match_percent: f32,
+    total_code: u64,
+    matched_code: u64,
+    matched_code_percent: f32,
+    total_data: u64,
+    matched_data: u64,
+    matched_data_percent: f32,
+    total_functions: u32,
+    matched_functions: u32,
+    matched_functions_percent: f32,
+    units: Vec<LegacyReportUnit>,
+}
+
+impl From<LegacyReport> for Report {
+    fn from(value: LegacyReport) -> Self {
+        Self {
+            measures: Some(Measures {
+                fuzzy_match_percent: value.fuzzy_match_percent,
+                total_code: value.total_code,
+                matched_code: value.matched_code,
+                matched_code_percent: value.matched_code_percent,
+                total_data: value.total_data,
+                matched_data: value.matched_data,
+                matched_data_percent: value.matched_data_percent,
+                total_functions: value.total_functions,
+                matched_functions: value.matched_functions,
+                matched_functions_percent: value.matched_functions_percent,
+            }),
+            units: value.units.into_iter().map(ReportUnit::from).collect(),
+        }
+    }
+}
+
+#[derive(Debug, Clone, Default, serde::Serialize, serde::Deserialize)]
+struct LegacyReportUnit {
+    name: String,
+    fuzzy_match_percent: f32,
+    total_code: u64,
+    matched_code: u64,
+    total_data: u64,
+    matched_data: u64,
+    total_functions: u32,
+    matched_functions: u32,
+    #[serde(skip_serializing_if = "Option::is_none")]
+    complete: Option<bool>,
+    #[serde(skip_serializing_if = "Option::is_none")]
+    module_name: Option<String>,
+    #[serde(skip_serializing_if = "Option::is_none")]
+    module_id: Option<u32>,
+    sections: Vec<LegacyReportItem>,
+    functions: Vec<LegacyReportItem>,
+}
+
+impl From<LegacyReportUnit> for ReportUnit {
+    fn from(value: LegacyReportUnit) -> Self {
+        let mut measures = Measures {
+            fuzzy_match_percent: value.fuzzy_match_percent,
+            total_code: value.total_code,
+            matched_code: value.matched_code,
+            total_data: value.total_data,
+            matched_data: value.matched_data,
+            total_functions: value.total_functions,
+            matched_functions: value.matched_functions,
+            ..Default::default()
+        };
+        measures.calc_matched_percent();
+        Self {
+            name: value.name.clone(),
+            measures: Some(measures),
+            sections: value.sections.into_iter().map(ReportItem::from).collect(),
+            functions: value.functions.into_iter().map(ReportItem::from).collect(),
+            metadata: Some(ReportUnitMetadata {
+                complete: value.complete,
+                module_name: value.module_name.clone(),
+                module_id: value.module_id,
+                ..Default::default()
+            }),
+        }
+    }
+}
+
+#[derive(Debug, Clone, Default, serde::Serialize, serde::Deserialize)]
+struct LegacyReportItem {
+    name: String,
+    #[serde(skip_serializing_if = "Option::is_none")]
+    demangled_name: Option<String>,
+    #[serde(
+        default,
+        skip_serializing_if = "Option::is_none",
+        serialize_with = "serialize_hex",
+        deserialize_with = "deserialize_hex"
+    )]
+    address: Option<u64>,
+    size: u64,
+    fuzzy_match_percent: f32,
+}
+
+impl From<LegacyReportItem> for ReportItem {
+    fn from(value: LegacyReportItem) -> Self {
+        Self {
+            name: value.name,
+            size: value.size,
+            fuzzy_match_percent: value.fuzzy_match_percent,
+            metadata: Some(ReportItemMetadata {
+                demangled_name: value.demangled_name,
+                virtual_address: value.address,
+            }),
+        }
+    }
+}
+
+fn serialize_hex<S>(x: &Option<u64>, s: S) -> Result<S::Ok, S::Error>
+where S: serde::Serializer {
+    if let Some(x) = x {
+        s.serialize_str(&format!("{:#x}", x))
+    } else {
+        s.serialize_none()
+    }
+}
+
+fn deserialize_hex<'de, D>(d: D) -> Result<Option<u64>, D::Error>
+where D: serde::Deserializer<'de> {
+    use serde::Deserialize;
+    let s = String::deserialize(d)?;
+    if s.is_empty() {
+        Ok(None)
+    } else if !s.starts_with("0x") {
+        Err(serde::de::Error::custom("expected hex string"))
+    } else {
+        u64::from_str_radix(&s[2..], 16).map(Some).map_err(serde::de::Error::custom)
+    }
+}

objdiff-core/Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "objdiff-core"
-version = "2.0.0-alpha.5"
+version = "2.0.0-beta.4"
 edition = "2021"
 rust-version = "1.70"
 authors = ["Luke Street <luke@street.dev>"]
@@ -17,7 +17,7 @@ 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"]
+ppc = ["any-arch", "cwdemangle", "cwextab", "ppc750cl"]
 x86 = ["any-arch", "cpp_demangle", "iced-x86", "msvc-demangler"]
 arm = ["any-arch", "cpp_demangle", "unarm", "arm-attr"]
 
@@ -45,6 +45,7 @@ gimli = { version = "0.29.0", default-features = false, features = ["read-all"],
 
 # ppc
 cwdemangle = { version = "1.0.0", optional = true }
+cwextab = { version = "0.2.3", optional = true }
 ppc750cl = { git = "https://github.com/encounter/ppc750cl", rev = "6cbd7d888c7082c2c860f66cbb9848d633f753ed", optional = true }
 
 # mips
@@ -56,5 +57,5 @@ iced-x86 = { version = "1.21.0", default-features = false, features = ["std", "d
 msvc-demangler = { version = "0.10.0", optional = true }
 
 # arm
-unarm = { version = "1.3.0", optional = true }
+unarm = { version = "1.4.0", optional = true }
 arm-attr = { version = "0.1.1", optional = true }

objdiff-core/src/arch/arm.rs

@@ -13,12 +13,12 @@ use object::{
 use unarm::{
     args::{Argument, OffsetImm, OffsetReg, Register},
     parse::{ArmVersion, ParseMode, Parser},
-    ParsedIns,
+    DisplayOptions, ParseFlags, ParsedIns, RegNames,
 };
 
 use crate::{
     arch::{ObjArch, ProcessCodeResult},
-    diff::{ArmArchVersion, DiffObjConfig},
+    diff::{ArmArchVersion, ArmR9Usage, DiffObjConfig},
     obj::{ObjIns, ObjInsArg, ObjInsArgValue, ObjReloc, ObjSection},
 };
 
@@ -149,7 +149,24 @@ impl ObjArch for ObjArchArm {
             object::Endianness::Little => unarm::Endian::Little,
             object::Endianness::Big => unarm::Endian::Big,
         };
-        let mut parser = Parser::new(version, first_mapping, start_addr, endian, code);
+
+        let parse_flags = ParseFlags { ual: config.arm_unified_syntax };
+
+        let mut parser = Parser::new(version, first_mapping, start_addr, endian, parse_flags, code);
+
+        let display_options = DisplayOptions {
+            reg_names: RegNames {
+                av_registers: config.arm_av_registers,
+                r9_use: match config.arm_r9_usage {
+                    ArmR9Usage::GeneralPurpose => unarm::R9Use::GeneralPurpose,
+                    ArmR9Usage::Sb => unarm::R9Use::Pid,
+                    ArmR9Usage::Tr => unarm::R9Use::Tls,
+                },
+                explicit_stack_limit: config.arm_sl_usage,
+                frame_pointer: config.arm_fp_usage,
+                ip: config.arm_ip_usage,
+            },
+        };
 
         while let Some((address, op, ins)) = parser.next() {
             if let Some(next) = next_mapping {
@@ -187,7 +204,7 @@ impl ObjArch for ObjArchArm {
             let (args, branch_dest) = if reloc.is_some() && parser.mode == ParseMode::Data {
                 (vec![ObjInsArg::Reloc], None)
             } else {
-                push_args(&ins, config, reloc_arg, address)?
+                push_args(&ins, config, reloc_arg, address, display_options)?
             };
 
             ops.push(op.id());
@@ -200,7 +217,7 @@ impl ObjArch for ObjArchArm {
                 reloc,
                 branch_dest,
                 line,
-                formatted: ins.to_string(),
+                formatted: ins.display(display_options).to_string(),
                 orig: None,
             });
         }
@@ -210,6 +227,7 @@ impl ObjArch for ObjArchArm {
 
     fn implcit_addend(
         &self,
+        _file: &File<'_>,
         section: &ObjSection,
         address: u64,
         reloc: &Relocation,
@@ -281,6 +299,7 @@ fn push_args(
     config: &DiffObjConfig,
     reloc_arg: Option<usize>,
     cur_addr: u32,
+    display_options: DisplayOptions,
 ) -> Result<(Vec<ObjInsArg>, Option<u64>)> {
     let mut args = vec![];
     let mut branch_dest = None;
@@ -318,7 +337,9 @@ fn push_args(
                         deref = true;
                         args.push(ObjInsArg::PlainText("[".into()));
                     }
-                    args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(reg.reg.to_string().into())));
+                    args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(
+                        reg.reg.display(display_options.reg_names).to_string().into(),
+                    )));
                     if reg.writeback {
                         if reg.deref {
                             writeback = true;
@@ -336,7 +357,10 @@ fn push_args(
                                 args.push(ObjInsArg::PlainText(config.separator().into()));
                             }
                             args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(
-                                Register::parse(i).to_string().into(),
+                                Register::parse(i)
+                                    .display(display_options.reg_names)
+                                    .to_string()
+                                    .into(),
                             )));
                             first = false;
                         }
@@ -376,14 +400,16 @@ fn push_args(
                 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())));
+                    args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(
+                        shift.reg.display(display_options.reg_names).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(),
+                        offset.reg.display(display_options.reg_names).to_string().into(),
                     )));
                 }
                 Argument::CpsrMode(mode) => {
@@ -398,9 +424,9 @@ fn push_args(
                 | Argument::StatusMask(_)
                 | Argument::Shift(_)
                 | Argument::CpsrFlags(_)
-                | Argument::Endian(_) => {
-                    args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(arg.to_string().into())))
-                }
+                | Argument::Endian(_) => args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(
+                    arg.display(display_options).to_string().into(),
+                ))),
             }
         }
     }

objdiff-core/src/arch/mips.rs

@@ -1,7 +1,10 @@
 use std::{borrow::Cow, collections::BTreeMap, sync::Mutex};
 
 use anyhow::{anyhow, bail, Result};
-use object::{elf, Endian, Endianness, File, FileFlags, Object, Relocation, RelocationFlags};
+use object::{
+    elf, Endian, Endianness, File, FileFlags, Object, ObjectSection, ObjectSymbol, Relocation,
+    RelocationFlags, RelocationTarget,
+};
 use rabbitizer::{config, Abi, InstrCategory, Instruction, OperandType};
 
 use crate::{
@@ -22,6 +25,7 @@ pub struct ObjArchMips {
     pub endianness: Endianness,
     pub abi: Abi,
     pub instr_category: InstrCategory,
+    pub ri_gp_value: i32,
 }
 
 const EF_MIPS_ABI: u32 = 0x0000F000;
@@ -30,6 +34,8 @@ const EF_MIPS_MACH: u32 = 0x00FF0000;
 const EF_MIPS_MACH_ALLEGREX: u32 = 0x00840000;
 const EF_MIPS_MACH_5900: u32 = 0x00920000;
 
+const R_MIPS15_S3: u32 = 119;
+
 impl ObjArchMips {
     pub fn new(object: &File) -> Result<Self> {
         let mut abi = Abi::NUMERIC;
@@ -56,7 +62,19 @@ impl ObjArchMips {
             }
             _ => bail!("Unsupported MIPS file flags"),
         }
-        Ok(Self { endianness: object.endianness(), abi, instr_category })
+
+        // Parse the ri_gp_value stored in .reginfo to be able to correctly
+        // calculate R_MIPS_GPREL16 relocations later. The value is stored
+        // 0x14 bytes into .reginfo (on 32 bit platforms)
+        let ri_gp_value = object
+            .section_by_name(".reginfo")
+            .and_then(|section| section.data().ok())
+            .and_then(|data| data.get(0x14..0x18))
+            .and_then(|s| s.try_into().ok())
+            .map(|bytes| object.endianness().read_i32_bytes(bytes))
+            .unwrap_or(0);
+
+        Ok(Self { endianness: object.endianness(), abi, instr_category, ri_gp_value })
     }
 }
 
@@ -104,7 +122,7 @@ impl ObjArch for ObjArchMips {
             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 {
+            let mut branch_dest = if is_branch {
                 cur_addr.checked_add_signed(branch_offset).map(|a| a as u64)
             } else {
                 None
@@ -121,9 +139,7 @@ impl ObjArch for ObjArchMips {
                     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 let Some(reloc) = reloc {
                             if matches!(&reloc.target_section, Some(s) if s == ".text")
                                 && reloc.target.address > start_address
                                 && reloc.target.address < end_address
@@ -131,7 +147,10 @@ impl ObjArch for ObjArchMips {
                                 args.push(ObjInsArg::BranchDest(reloc.target.address));
                             } else {
                                 push_reloc(&mut args, reloc)?;
+                                branch_dest = None;
                             }
+                        } else if let Some(branch_dest) = branch_dest {
+                            args.push(ObjInsArg::BranchDest(branch_dest));
                         } else {
                             args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(
                                 op.disassemble(&instruction, None).into(),
@@ -152,6 +171,18 @@ impl ObjArch for ObjArchMips {
                         )));
                         args.push(ObjInsArg::PlainText(")".into()));
                     }
+                    // OperandType::r5900_immediate15 => match reloc {
+                    //     Some(reloc)
+                    //         if reloc.flags == RelocationFlags::Elf { r_type: R_MIPS15_S3 } =>
+                    //     {
+                    //         push_reloc(&mut args, reloc)?;
+                    //     }
+                    //     _ => {
+                    //         args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(
+                    //             op.disassemble(&instruction, None).into(),
+                    //         )));
+                    //     }
+                    // },
                     _ => {
                         args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(
                             op.disassemble(&instruction, None).into(),
@@ -179,6 +210,7 @@ impl ObjArch for ObjArchMips {
 
     fn implcit_addend(
         &self,
+        file: &File<'_>,
         section: &ObjSection,
         address: u64,
         reloc: &Relocation,
@@ -187,14 +219,29 @@ impl ObjArch for ObjArchMips {
         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_26 } => ((addend & 0x03FFFFFF) << 2) 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,
+                r_type: elf::R_MIPS_LO16 | elf::R_MIPS_GOT16 | elf::R_MIPS_CALL16,
             } => (addend & 0x0000FFFF) as i16 as i64,
-            RelocationFlags::Elf { r_type: elf::R_MIPS_26 } => ((addend & 0x03FFFFFF) << 2) as i64,
+            RelocationFlags::Elf { r_type: elf::R_MIPS_GPREL16 | elf::R_MIPS_LITERAL } => {
+                let RelocationTarget::Symbol(idx) = reloc.target() else {
+                    bail!("Unsupported R_MIPS_GPREL16 relocation against a non-symbol");
+                };
+                let sym = file.symbol_by_index(idx)?;
+
+                // if the symbol we are relocating against is in a local section we need to add
+                // the ri_gp_value from .reginfo to the addend.
+                if sym.section().index().is_some() {
+                    ((addend & 0x0000FFFF) as i16 as i64) + self.ri_gp_value as i64
+                } else {
+                    (addend & 0x0000FFFF) as i16 as i64
+                }
+            }
+            RelocationFlags::Elf { r_type: elf::R_MIPS_PC16 } => 0, // PC-relative relocation
+            RelocationFlags::Elf { r_type: R_MIPS15_S3 } => ((addend & 0x001FFFC0) >> 3) as i64,
             flags => bail!("Unsupported MIPS implicit relocation {flags:?}"),
         })
     }
@@ -202,13 +249,16 @@ impl ObjArch for ObjArchMips {
     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"),
+                elf::R_MIPS_HI16 => Cow::Borrowed("R_MIPS_HI16"),
+                elf::R_MIPS_LO16 => Cow::Borrowed("R_MIPS_LO16"),
+                elf::R_MIPS_GPREL16 => Cow::Borrowed("R_MIPS_GPREL16"),
+                elf::R_MIPS_LITERAL => Cow::Borrowed("R_MIPS_LITERAL"),
+                elf::R_MIPS_GOT16 => Cow::Borrowed("R_MIPS_GOT16"),
+                elf::R_MIPS_PC16 => Cow::Borrowed("R_MIPS_PC16"),
+                elf::R_MIPS_CALL16 => Cow::Borrowed("R_MIPS_CALL16"),
+                R_MIPS15_S3 => Cow::Borrowed("R_MIPS15_S3"),
                 _ => Cow::Owned(format!("<{flags:?}>")),
             },
             _ => Cow::Owned(format!("<{flags:?}>")),
@@ -244,7 +294,11 @@ fn push_reloc(args: &mut Vec<ObjInsArg>, reloc: &ObjReloc) -> Result<()> {
                 args.push(ObjInsArg::Reloc);
                 args.push(ObjInsArg::PlainText(")".into()));
             }
-            elf::R_MIPS_32 | elf::R_MIPS_26 => {
+            elf::R_MIPS_32
+            | elf::R_MIPS_26
+            | elf::R_MIPS_LITERAL
+            | elf::R_MIPS_PC16
+            | R_MIPS15_S3 => {
                 args.push(ObjInsArg::Reloc);
             }
             _ => bail!("Unsupported ELF MIPS relocation type {r_type}"),

objdiff-core/src/arch/mod.rs

@@ -1,7 +1,7 @@
 use std::{borrow::Cow, collections::BTreeMap};
 
 use anyhow::{bail, Result};
-use object::{Architecture, Object, ObjectSymbol, Relocation, RelocationFlags, Symbol};
+use object::{Architecture, File, Object, ObjectSymbol, Relocation, RelocationFlags, Symbol};
 
 use crate::{
     diff::DiffObjConfig,
@@ -28,8 +28,13 @@ pub trait ObjArch: Send + Sync {
         config: &DiffObjConfig,
     ) -> Result<ProcessCodeResult>;
 
-    fn implcit_addend(&self, section: &ObjSection, address: u64, reloc: &Relocation)
-        -> Result<i64>;
+    fn implcit_addend(
+        &self,
+        file: &File<'_>,
+        section: &ObjSection,
+        address: u64,
+        reloc: &Relocation,
+    ) -> Result<i64>;
 
     fn demangle(&self, _name: &str) -> Option<String> { None }
 

objdiff-core/src/arch/ppc.rs

@@ -150,6 +150,7 @@ impl ObjArch for ObjArchPpc {
 
     fn implcit_addend(
         &self,
+        _file: &File<'_>,
         _section: &ObjSection,
         address: u64,
         reloc: &Relocation,

objdiff-core/src/arch/x86.rs

@@ -128,6 +128,7 @@ impl ObjArch for ObjArchX86 {
 
     fn implcit_addend(
         &self,
+        _file: &File<'_>,
         section: &ObjSection,
         address: u64,
         reloc: &Relocation,

objdiff-core/src/diff/data.rs

@@ -11,27 +11,6 @@ use crate::{
     obj::{ObjInfo, ObjSection, SymbolRef},
 };
 
-/// Compare the addresses and sizes of each symbol in the BSS sections.
-pub fn diff_bss_section(
-    left: &ObjSection,
-    right: &ObjSection,
-) -> Result<(ObjSectionDiff, ObjSectionDiff)> {
-    let deadline = Instant::now() + Duration::from_secs(5);
-    let left_sizes = left.symbols.iter().map(|s| (s.section_address, s.size)).collect::<Vec<_>>();
-    let right_sizes = right.symbols.iter().map(|s| (s.section_address, s.size)).collect::<Vec<_>>();
-    let ops = capture_diff_slices_deadline(
-        Algorithm::Patience,
-        &left_sizes,
-        &right_sizes,
-        Some(deadline),
-    );
-    let match_percent = get_diff_ratio(&ops, left_sizes.len(), right_sizes.len()) * 100.0;
-    Ok((
-        ObjSectionDiff { symbols: vec![], data_diff: vec![], match_percent: Some(match_percent) },
-        ObjSectionDiff { symbols: vec![], data_diff: vec![], match_percent: Some(match_percent) },
-    ))
-}
-
 pub fn diff_bss_symbol(
     left_obj: &ObjInfo,
     right_obj: &ObjInfo,
@@ -65,10 +44,14 @@ pub fn no_diff_symbol(_obj: &ObjInfo, symbol_ref: SymbolRef) -> ObjSymbolDiff {
 pub fn diff_data_section(
     left: &ObjSection,
     right: &ObjSection,
+    left_section_diff: &ObjSectionDiff,
+    right_section_diff: &ObjSectionDiff,
 ) -> Result<(ObjSectionDiff, ObjSectionDiff)> {
     let deadline = Instant::now() + Duration::from_secs(5);
-    let left_max = left.symbols.iter().map(|s| s.section_address + s.size).max().unwrap_or(0);
-    let right_max = right.symbols.iter().map(|s| s.section_address + s.size).max().unwrap_or(0);
+    let left_max =
+        left.symbols.iter().map(|s| s.section_address + s.size).max().unwrap_or(0).min(left.size);
+    let right_max =
+        right.symbols.iter().map(|s| s.section_address + s.size).max().unwrap_or(0).min(right.size);
     let left_data = &left.data[..left_max as usize];
     let right_data = &right.data[..right_max as usize];
     let ops =
@@ -143,18 +126,18 @@ pub fn diff_data_section(
         }
     }
 
-    Ok((
-        ObjSectionDiff {
-            symbols: vec![],
-            data_diff: left_diff,
-            match_percent: Some(match_percent),
-        },
-        ObjSectionDiff {
-            symbols: vec![],
-            data_diff: right_diff,
-            match_percent: Some(match_percent),
-        },
-    ))
+    let (mut left_section_diff, mut right_section_diff) =
+        diff_generic_section(left, right, left_section_diff, right_section_diff)?;
+    left_section_diff.data_diff = left_diff;
+    right_section_diff.data_diff = right_diff;
+    // Use the highest match percent between two options:
+    // - Left symbols matching right symbols by name
+    // - Diff of the data itself
+    if left_section_diff.match_percent.unwrap_or(-1.0) < match_percent {
+        left_section_diff.match_percent = Some(match_percent);
+        right_section_diff.match_percent = Some(match_percent);
+    }
+    Ok((left_section_diff, right_section_diff))
 }
 
 pub fn diff_data_symbol(
@@ -195,21 +178,56 @@ pub fn diff_data_symbol(
     ))
 }
 
-/// Compare the text sections of two object files.
-/// This essentially adds up the match percentage of each symbol in the text section.
-pub fn diff_text_section(
+/// Compares a section of two object files.
+/// This essentially adds up the match percentage of each symbol in the section.
+pub fn diff_generic_section(
     left: &ObjSection,
     _right: &ObjSection,
     left_diff: &ObjSectionDiff,
     _right_diff: &ObjSectionDiff,
 ) -> Result<(ObjSectionDiff, ObjSectionDiff)> {
-    let match_percent = left
-        .symbols
-        .iter()
-        .zip(left_diff.symbols.iter())
-        .map(|(s, d)| d.match_percent.unwrap_or(0.0) * s.size as f32)
-        .sum::<f32>()
-        / left.size as f32;
+    let match_percent = if left_diff.symbols.iter().all(|d| d.match_percent == Some(100.0)) {
+        100.0 // Avoid fp precision issues
+    } else {
+        left.symbols
+            .iter()
+            .zip(left_diff.symbols.iter())
+            .map(|(s, d)| d.match_percent.unwrap_or(0.0) * s.size as f32)
+            .sum::<f32>()
+            / left.size as f32
+    };
+    Ok((
+        ObjSectionDiff { symbols: vec![], data_diff: vec![], match_percent: Some(match_percent) },
+        ObjSectionDiff { symbols: vec![], data_diff: vec![], match_percent: Some(match_percent) },
+    ))
+}
+
+/// Compare the addresses and sizes of each symbol in the BSS sections.
+pub fn diff_bss_section(
+    left: &ObjSection,
+    right: &ObjSection,
+    left_diff: &ObjSectionDiff,
+    right_diff: &ObjSectionDiff,
+) -> Result<(ObjSectionDiff, ObjSectionDiff)> {
+    let deadline = Instant::now() + Duration::from_secs(5);
+    let left_sizes = left.symbols.iter().map(|s| (s.section_address, s.size)).collect::<Vec<_>>();
+    let right_sizes = right.symbols.iter().map(|s| (s.section_address, s.size)).collect::<Vec<_>>();
+    let ops = capture_diff_slices_deadline(
+        Algorithm::Patience,
+        &left_sizes,
+        &right_sizes,
+        Some(deadline),
+    );
+    let mut match_percent = get_diff_ratio(&ops, left_sizes.len(), right_sizes.len()) * 100.0;
+
+    // Use the highest match percent between two options:
+    // - Left symbols matching right symbols by name
+    // - Diff of the addresses and sizes of each symbol
+    let (generic_diff, _) = diff_generic_section(left, right, left_diff, right_diff)?;
+    if generic_diff.match_percent.unwrap_or(-1.0) > match_percent {
+        match_percent = generic_diff.match_percent.unwrap();
+    }
+
     Ok((
         ObjSectionDiff { symbols: vec![], data_diff: vec![], match_percent: Some(match_percent) },
         ObjSectionDiff { symbols: vec![], data_diff: vec![], match_percent: Some(match_percent) },

objdiff-core/src/diff/display.rs

@@ -20,7 +20,7 @@ pub enum DiffText<'a> {
     /// Instruction argument
     Argument(&'a ObjInsArgValue, Option<&'a ObjInsArgDiff>),
     /// Branch destination
-    BranchDest(u64),
+    BranchDest(u64, Option<&'a ObjInsArgDiff>),
     /// Symbol name
     Symbol(&'a ObjSymbol),
     /// Number of spaces
@@ -62,12 +62,12 @@ pub fn display_diff<E>(
         if i == 0 {
             cb(DiffText::Spacing(1))?;
         }
+        let diff = ins_diff.arg_diff.get(i).and_then(|o| o.as_ref());
         match arg {
             ObjInsArg::PlainText(s) => {
                 cb(DiffText::Basic(s))?;
             }
             ObjInsArg::Arg(v) => {
-                let diff = ins_diff.arg_diff.get(i).and_then(|o| o.as_ref());
                 cb(DiffText::Argument(v, diff))?;
             }
             ObjInsArg::Reloc => {
@@ -75,7 +75,7 @@ pub fn display_diff<E>(
             }
             ObjInsArg::BranchDest(dest) => {
                 if let Some(dest) = dest.checked_sub(base_addr) {
-                    cb(DiffText::BranchDest(dest))?;
+                    cb(DiffText::BranchDest(dest, diff))?;
                 } else {
                     cb(DiffText::Basic("<unknown>"))?;
                 }
@@ -107,7 +107,9 @@ impl PartialEq<DiffText<'_>> for HighlightKind {
             (HighlightKind::Opcode(a), DiffText::Opcode(_, b)) => a == b,
             (HighlightKind::Arg(a), DiffText::Argument(b, _)) => a.loose_eq(b),
             (HighlightKind::Symbol(a), DiffText::Symbol(b)) => a == &b.name,
-            (HighlightKind::Address(a), DiffText::Address(b) | DiffText::BranchDest(b)) => a == b,
+            (HighlightKind::Address(a), DiffText::Address(b) | DiffText::BranchDest(b, _)) => {
+                a == b
+            }
             _ => false,
         }
     }
@@ -123,7 +125,7 @@ impl From<DiffText<'_>> for HighlightKind {
             DiffText::Opcode(_, op) => HighlightKind::Opcode(op),
             DiffText::Argument(arg, _) => HighlightKind::Arg(arg.clone()),
             DiffText::Symbol(sym) => HighlightKind::Symbol(sym.name.to_string()),
-            DiffText::Address(addr) | DiffText::BranchDest(addr) => HighlightKind::Address(addr),
+            DiffText::Address(addr) | DiffText::BranchDest(addr, _) => HighlightKind::Address(addr),
             _ => HighlightKind::None,
         }
     }

objdiff-core/src/diff/mod.rs

@@ -7,7 +7,7 @@ use crate::{
         code::{diff_code, no_diff_code, process_code_symbol},
         data::{
             diff_bss_section, diff_bss_symbol, diff_data_section, diff_data_symbol,
-            diff_text_section, no_diff_symbol,
+            diff_generic_section, no_diff_symbol,
         },
     },
     obj::{ObjInfo, ObjIns, ObjSection, ObjSectionKind, ObjSymbol, SymbolRef},
@@ -117,6 +117,34 @@ pub enum ArmArchVersion {
     V6K,
 }
 
+#[derive(
+    Debug,
+    Copy,
+    Clone,
+    Default,
+    Eq,
+    PartialEq,
+    serde::Deserialize,
+    serde::Serialize,
+    strum::VariantArray,
+    strum::EnumMessage,
+)]
+pub enum ArmR9Usage {
+    #[default]
+    #[strum(
+        message = "R9 or V6 (default)",
+        detailed_message = "Use R9 as a general-purpose register."
+    )]
+    GeneralPurpose,
+    #[strum(
+        message = "SB (static base)",
+        detailed_message = "Used for position-independent data (PID)."
+    )]
+    Sb,
+    #[strum(message = "TR (TLS register)", detailed_message = "Used for thread-local storage.")]
+    Tr,
+}
+
 #[inline]
 const fn default_true() -> bool { true }
 
@@ -134,6 +162,12 @@ pub struct DiffObjConfig {
     pub mips_instr_category: MipsInstrCategory,
     // ARM
     pub arm_arch_version: ArmArchVersion,
+    pub arm_unified_syntax: bool,
+    pub arm_av_registers: bool,
+    pub arm_r9_usage: ArmR9Usage,
+    pub arm_sl_usage: bool,
+    pub arm_fp_usage: bool,
+    pub arm_ip_usage: bool,
 }
 
 impl Default for DiffObjConfig {
@@ -146,6 +180,12 @@ impl Default for DiffObjConfig {
             mips_abi: Default::default(),
             mips_instr_category: Default::default(),
             arm_arch_version: Default::default(),
+            arm_unified_syntax: true,
+            arm_av_registers: false,
+            arm_r9_usage: Default::default(),
+            arm_sl_usage: false,
+            arm_fp_usage: false,
+            arm_ip_usage: false,
         }
     }
 }
@@ -446,7 +486,7 @@ pub fn diff_objs(
                 ObjSectionKind::Code => {
                     let left_section_diff = left_out.section_diff(left_section_idx);
                     let right_section_diff = right_out.section_diff(right_section_idx);
-                    let (left_diff, right_diff) = diff_text_section(
+                    let (left_diff, right_diff) = diff_generic_section(
                         left_section,
                         right_section,
                         left_section_diff,
@@ -456,12 +496,26 @@ pub fn diff_objs(
                     right_out.section_diff_mut(right_section_idx).merge(right_diff);
                 }
                 ObjSectionKind::Data => {
-                    let (left_diff, right_diff) = diff_data_section(left_section, right_section)?;
+                    let left_section_diff = left_out.section_diff(left_section_idx);
+                    let right_section_diff = right_out.section_diff(right_section_idx);
+                    let (left_diff, right_diff) = diff_data_section(
+                        left_section,
+                        right_section,
+                        left_section_diff,
+                        right_section_diff,
+                    )?;
                     left_out.section_diff_mut(left_section_idx).merge(left_diff);
                     right_out.section_diff_mut(right_section_idx).merge(right_diff);
                 }
                 ObjSectionKind::Bss => {
-                    let (left_diff, right_diff) = diff_bss_section(left_section, right_section)?;
+                    let left_section_diff = left_out.section_diff(left_section_idx);
+                    let right_section_diff = right_out.section_diff(right_section_idx);
+                    let (left_diff, right_diff) = diff_bss_section(
+                        left_section,
+                        right_section,
+                        left_section_diff,
+                        right_section_diff,
+                    )?;
                     left_out.section_diff_mut(left_section_idx).merge(left_diff);
                     right_out.section_diff_mut(right_section_idx).merge(right_diff);
                 }
@@ -504,8 +558,8 @@ fn matching_symbols(
             for (symbol_idx, symbol) in section.symbols.iter().enumerate() {
                 let symbol_match = SymbolMatch {
                     left: Some(SymbolRef { section_idx, symbol_idx }),
-                    right: find_symbol(right, symbol, section),
-                    prev: find_symbol(prev, symbol, section),
+                    right: find_symbol(right, symbol, section, Some(&right_used)),
+                    prev: find_symbol(prev, symbol, section, None),
                     section_kind: section.kind,
                 };
                 matches.push(symbol_match);
@@ -537,7 +591,7 @@ fn matching_symbols(
                 matches.push(SymbolMatch {
                     left: None,
                     right: Some(symbol_ref),
-                    prev: find_symbol(prev, symbol, section),
+                    prev: find_symbol(prev, symbol, section, None),
                     section_kind: section.kind,
                 });
             }
@@ -558,10 +612,25 @@ fn matching_symbols(
     Ok(matches)
 }
 
+fn unmatched_symbols<'section, 'used>(
+    section: &'section ObjSection,
+    section_idx: usize,
+    used: Option<&'used HashSet<SymbolRef>>,
+) -> impl Iterator<Item = (usize, &'section ObjSymbol)> + 'used
+where
+    'section: 'used,
+{
+    section.symbols.iter().enumerate().filter(move |&(symbol_idx, _)| {
+        // Skip symbols that have already been matched
+        !used.map(|u| u.contains(&SymbolRef { section_idx, symbol_idx })).unwrap_or(false)
+    })
+}
+
 fn find_symbol(
     obj: Option<&ObjInfo>,
     in_symbol: &ObjSymbol,
     in_section: &ObjSection,
+    used: Option<&HashSet<SymbolRef>>,
 ) -> Option<SymbolRef> {
     let obj = obj?;
     // Try to find an exact name match
@@ -569,8 +638,8 @@ fn find_symbol(
         if section.kind != in_section.kind {
             continue;
         }
-        if let Some(symbol_idx) =
-            section.symbols.iter().position(|symbol| symbol.name == in_symbol.name)
+        if let Some((symbol_idx, _)) = unmatched_symbols(section, section_idx, used)
+            .find(|(_, symbol)| symbol.name == in_symbol.name)
         {
             return Some(SymbolRef { section_idx, symbol_idx });
         }
@@ -583,9 +652,33 @@ fn find_symbol(
         if let Some((section_idx, section)) =
             obj.sections.iter().enumerate().find(|(_, s)| s.name == in_section.name)
         {
-            if let Some(symbol_idx) = section.symbols.iter().position(|symbol| {
-                symbol.address == in_symbol.address && symbol.name.starts_with('@')
-            }) {
+            if let Some((symbol_idx, _)) =
+                unmatched_symbols(section, section_idx, used).find(|(_, symbol)| {
+                    symbol.address == in_symbol.address && symbol.name.starts_with('@')
+                })
+            {
+                return Some(SymbolRef { section_idx, symbol_idx });
+            }
+        }
+    }
+    // Match Metrowerks symbol$1234 against symbol$2345
+    if let Some((prefix, suffix)) = in_symbol.name.split_once('$') {
+        if !suffix.chars().all(char::is_numeric) {
+            return None;
+        }
+        for (section_idx, section) in obj.sections.iter().enumerate() {
+            if section.kind != in_section.kind {
+                continue;
+            }
+            if let Some((symbol_idx, _)) =
+                unmatched_symbols(section, section_idx, used).find(|&(_, symbol)| {
+                    if let Some((p, s)) = symbol.name.split_once('$') {
+                        prefix == p && s.chars().all(char::is_numeric)
+                    } else {
+                        false
+                    }
+                })
+            {
                 return Some(SymbolRef { section_idx, symbol_idx });
             }
         }