ARMv4T (GBA) and ARMv6K (3DS) support (#75)

* 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>

Cargo.lock

@@ -150,6 +150,15 @@ dependencies = [
  "syn 1.0.109",
 ]
 
+[[package]]
+name = "arm-attr"
+version = "0.1.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "4d0cabd3a7d2dfa96ab3faa7b532a83c5e090061bf6d83197ca2bc91f5afac6c"
+dependencies = [
+ "thiserror",
+]
+
 [[package]]
 name = "arrayvec"
 version = "0.7.4"
@@ -968,7 +977,7 @@ source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "3e3d747f100290a1ca24b752186f61f6637e1deffe3bf6320de6fcb29510a307"
 dependencies = [
  "bitflags 2.5.0",
- "libloading 0.7.4",
+ "libloading 0.8.3",
  "winapi",
 ]
 
@@ -2802,6 +2811,7 @@ name = "objdiff-core"
 version = "2.0.0-alpha.4"
 dependencies = [
  "anyhow",
+ "arm-attr",
  "byteorder",
  "cpp_demangle",
  "cwdemangle",
@@ -4029,18 +4039,18 @@ dependencies = [
 
 [[package]]
 name = "thiserror"
-version = "1.0.59"
+version = "1.0.61"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "f0126ad08bff79f29fc3ae6a55cc72352056dfff61e3ff8bb7129476d44b23aa"
+checksum = "c546c80d6be4bc6a00c0f01730c08df82eaa7a7a61f11d656526506112cc1709"
 dependencies = [
  "thiserror-impl",
 ]
 
 [[package]]
 name = "thiserror-impl"
-version = "1.0.59"
+version = "1.0.61"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "d1cd413b5d558b4c5bf3680e324a6fa5014e7b7c067a51e69dbdf47eb7148b66"
+checksum = "46c3384250002a6d5af4d114f2845d37b57521033f30d5c3f46c4d70e1197533"
 dependencies = [
  "proc-macro2",
  "quote",
@@ -4320,9 +4330,9 @@ dependencies = [
 
 [[package]]
 name = "unarm"
-version = "1.0.0"
+version = "1.3.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "1c13fc9a9c95348bf7565e5c30688fc288239962958cac0ccdc7cd009141d850"
+checksum = "c6bff109f0171a299559d82a7236e056093fc0dcd2a7da86aa745f82281e2d31"
 
 [[package]]
 name = "unicase"

README.md

@@ -17,7 +17,7 @@ Supports:
 - PowerPC 750CL (GameCube, Wii)
 - MIPS (N64, PS1, PS2, PSP)
 - x86 (COFF only at the moment)
-- ARMv5 (DS)
+- ARM (GBA, DS, 3DS)
 
 See [Usage](#usage) for more information.
 

objdiff-cli/src/cmd/diff.rs

@@ -816,6 +816,7 @@ impl FunctionDiffUi {
             x86_formatter: Default::default(),       // TODO
             mips_abi: Default::default(),            // TODO
             mips_instr_category: Default::default(), // TODO
+            arm_arch_version: Default::default(),    // TODO
         };
         let target = self
             .target_path

objdiff-core/Cargo.toml

@@ -19,7 +19,7 @@ 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 = ["any-arch", "cpp_demangle", "unarm", "arm-attr"]
 
 [dependencies]
 anyhow = "1.0.82"
@@ -56,4 +56,5 @@ iced-x86 = { version = "1.21.0", default-features = false, features = ["std", "d
 msvc-demangler = { version = "0.10.0", optional = true }
 
 # arm
-unarm = { version = "1.0.0", optional = true }
+unarm = { version = "1.3.0", optional = true }
+arm-attr = { version = "0.1.1", optional = true }

objdiff-core/src/arch/arm.rs

@@ -4,9 +4,11 @@ use std::{
 };
 
 use anyhow::{bail, Result};
+use arm_attr::{enums::CpuArch, tag::Tag, BuildAttrs};
 use object::{
-    elf, File, Object, ObjectSection, ObjectSymbol, Relocation, RelocationFlags, SectionIndex,
+    elf::{self, SHT_ARM_ATTRIBUTES},
-    SectionKind, Symbol,
+    Endian, File, Object, ObjectSection, ObjectSymbol, Relocation, RelocationFlags, SectionIndex,
+    SectionKind, Symbol, SymbolKind,
 };
 use unarm::{
     args::{Argument, OffsetImm, OffsetReg, Register},
@@ -16,41 +18,93 @@ use unarm::{
 
 use crate::{
     arch::{ObjArch, ProcessCodeResult},
-    diff::DiffObjConfig,
+    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: HashMap<_, _> = file
+                let disasm_modes = Self::elf_get_mapping_symbols(file);
-                    .sections()
+                let detected_version = Self::elf_detect_arm_version(file)?;
-                    .filter(|s| s.kind() == SectionKind::Text)
+                Ok(Self { disasm_modes, detected_version, endianness })
-                    .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();
-                Ok(Self { disasm_modes })
             }
             _ => 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,
@@ -81,10 +135,21 @@ impl ObjArch for ObjArchArm {
             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();
+        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 mut parser = Parser::new(ArmVersion::V5Te, first_mapping, start_addr, code);
+
+        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 {
@@ -95,7 +160,6 @@ impl ObjArch for ObjArchArm {
                     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();
@@ -103,11 +167,19 @@ impl ObjArch for ObjArchArm {
             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_PC24 } => {
+                    | 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(_)));
+                    }
                     _ => (),
                 }
             };
@@ -138,11 +210,42 @@ impl ObjArch for ObjArchArm {
 
     fn implcit_addend(
         &self,
-        _section: &ObjSection,
+        section: &ObjSection,
         address: u64,
         reloc: &Relocation,
     ) -> anyhow::Result<i64> {
-        bail!("Unsupported ARM implicit relocation {:#x}{:?}", address, reloc.flags())
+        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> {
@@ -209,6 +312,7 @@ fn push_args(
             args.push(ObjInsArg::Reloc);
         } else {
             match arg {
+                Argument::None => {}
                 Argument::Reg(reg) => {
                     if reg.deref {
                         deref = true;
@@ -242,7 +346,7 @@ fn push_args(
                         args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque("^".to_string().into())));
                     }
                 }
-                Argument::UImm(value) | Argument::CoOpcode(value) => {
+                Argument::UImm(value) | Argument::CoOpcode(value) | Argument::SatImm(value) => {
                     args.push(ObjInsArg::PlainText("#".into()));
                     args.push(ObjInsArg::Arg(ObjInsArgValue::Unsigned(*value as u64)));
                 }
@@ -282,7 +386,21 @@ fn push_args(
                         offset.reg.to_string().into(),
                     )));
                 }
-                _ => args.push(ObjInsArg::Arg(ObjInsArgValue::Opaque(arg.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())))
+                }
             }
         }
     }

objdiff-core/src/arch/mod.rs

@@ -1,7 +1,7 @@
 use std::{borrow::Cow, collections::BTreeMap};
 
 use anyhow::{bail, Result};
-use object::{Architecture, Object, Relocation, RelocationFlags};
+use object::{Architecture, Object, ObjectSymbol, Relocation, RelocationFlags, Symbol};
 
 use crate::{
     diff::DiffObjConfig,
@@ -34,6 +34,8 @@ pub trait ObjArch: Send + Sync {
     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 {

objdiff-core/src/diff/mod.rs

@@ -93,6 +93,30 @@ pub enum MipsInstrCategory {
     R5900,
 }
 
+#[derive(
+    Debug,
+    Copy,
+    Clone,
+    Default,
+    Eq,
+    PartialEq,
+    serde::Deserialize,
+    serde::Serialize,
+    strum::VariantArray,
+    strum::EnumMessage,
+)]
+pub enum ArmArchVersion {
+    #[default]
+    #[strum(message = "Auto (default)")]
+    Auto,
+    #[strum(message = "ARMv4T (GBA)")]
+    V4T,
+    #[strum(message = "ARMv5TE (DS)")]
+    V5TE,
+    #[strum(message = "ARMv6K (3DS)")]
+    V6K,
+}
+
 #[inline]
 const fn default_true() -> bool { true }
 
@@ -108,6 +132,8 @@ pub struct DiffObjConfig {
     // MIPS
     pub mips_abi: MipsAbi,
     pub mips_instr_category: MipsInstrCategory,
+    // ARM
+    pub arm_arch_version: ArmArchVersion,
 }
 
 impl Default for DiffObjConfig {
@@ -119,6 +145,7 @@ impl Default for DiffObjConfig {
             x86_formatter: Default::default(),
             mips_abi: Default::default(),
             mips_instr_category: Default::default(),
+            arm_arch_version: Default::default(),
         }
     }
 }

objdiff-core/src/obj/read.rs

@@ -55,12 +55,13 @@ fn to_obj_symbol(
     if obj_file.format() == BinaryFormat::Elf && symbol.scope() == SymbolScope::Linkage {
         flags = ObjSymbolFlagSet(flags.0 | ObjSymbolFlags::Hidden);
     }
+    let address = arch.symbol_address(symbol);
     let section_address = if let Some(section) =
         symbol.section_index().and_then(|idx| obj_file.section_by_index(idx).ok())
     {
-        symbol.address() - section.address()
+        address - section.address()
     } else {
-        symbol.address()
+        address
     };
     let demangled_name = arch.demangle(name);
     // Find the virtual address for the symbol if available
@@ -70,7 +71,7 @@ fn to_obj_symbol(
     Ok(ObjSymbol {
         name: name.to_string(),
         demangled_name,
-        address: symbol.address(),
+        address,
         section_address,
         size: symbol.size(),
         size_known: symbol.size() != 0,

objdiff-gui/src/views/config.rs

@@ -16,7 +16,7 @@ use egui::{
 use globset::Glob;
 use objdiff_core::{
     config::{ProjectObject, DEFAULT_WATCH_PATTERNS},
-    diff::{MipsAbi, MipsInstrCategory, X86Formatter},
+    diff::{ArmArchVersion, MipsAbi, MipsInstrCategory, X86Formatter},
 };
 use self_update::cargo_crate_version;
 use strum::{EnumMessage, VariantArray};
@@ -907,4 +907,22 @@ fn arch_config_ui(ui: &mut egui::Ui, config: &mut AppConfig, _appearance: &Appea
                 }
             }
         });
+    ui.separator();
+    ui.heading("ARM");
+    egui::ComboBox::new("arm_arch_version", "Architecture Version")
+        .selected_text(config.diff_obj_config.arm_arch_version.get_message().unwrap())
+        .show_ui(ui, |ui| {
+            for &version in ArmArchVersion::VARIANTS {
+                if ui
+                    .selectable_label(
+                        config.diff_obj_config.arm_arch_version == version,
+                        version.get_message().unwrap(),
+                    )
+                    .clicked()
+                {
+                    config.diff_obj_config.arm_arch_version = version;
+                    config.queue_reload = true;
+                }
+            }
+        });
 }