Integrate built-in disassemblers, drop capstone dependency

Replace capstone with pure-Rust disassemblers for x86_64 and aarch64.
The disasm feature flag is kept to control compilation in release builds
but no longer pulls in any external dependency.

Author: tekknolagi

zjit/Cargo.toml

@@ -6,9 +6,6 @@ rust-version = "1.85.0" # Minimally supported rust version
 publish = false         # Don't publish to crates.io
 
 [dependencies]
-# No required dependencies to simplify build process.
-# Optional For development and testing purposes.
-capstone = { version = "0.14.0", optional = true }
 jit = { version = "0.1.0", path = "../jit" }
 
 [dev-dependencies]
@@ -18,7 +15,7 @@ rand = "0.9"
 # NOTE: Development builds select a set of these via configure.ac
 # For debugging, `make V=1` shows exact cargo invocation.
 [features]
-# Support --yjit-dump-disasm and RubyVM::YJIT.disasm using libcapstone.
-disasm = ["capstone"]
+# Support --zjit-dump-disasm using built-in disassembler.
+disasm = []
 runtime_checks = []
 stats_allocator = []

zjit/src/disasm.rs

@@ -23,49 +23,52 @@ pub fn dump_disasm_addr_range(cb: &CodeBlock, start_addr: CodePtr, end_addr: Cod
 pub fn disasm_addr_range(cb: &CodeBlock, start_addr: usize, end_addr: usize) -> String {
     use std::fmt::Write;
 
-    let mut out = String::from("");
+    let mut out = String::new();
 
-    // Initialize capstone
-    use capstone::prelude::*;
-
-    #[cfg(target_arch = "x86_64")]
-    let mut cs = Capstone::new()
-        .x86()
-        .mode(arch::x86::ArchMode::Mode64)
-        .syntax(arch::x86::ArchSyntax::Intel)
-        .build()
-        .unwrap();
-    #[cfg(target_arch = "aarch64")]
-    let mut cs = Capstone::new()
-        .arm64()
-        .mode(arch::arm64::ArchMode::Arm)
-        .detail(true)
-        .build()
-        .unwrap();
-
-    cs.set_skipdata(true).unwrap();
-
-    // Disassemble the instructions
     let code_size = end_addr - start_addr;
-    let code_slice = unsafe { std::slice::from_raw_parts(start_addr as _, code_size) };
+    let code_slice = unsafe { std::slice::from_raw_parts(start_addr as *const u8, code_size) };
     // Stabilize output for cargo test
     #[cfg(test)]
     let start_addr = 0;
-    let insns = cs.disasm_all(code_slice, start_addr as u64).unwrap();
 
     let colors = crate::ttycolors::get_colors();
     let bold_begin = colors.bold_begin;
     let bold_end = colors.bold_end;
 
-    // For each instruction in this block
-    for insn in insns.as_ref() {
-        // Comments for this block
-        if let Some(comment_list) = cb.comments_at(insn.address() as usize) {
+    let mut offset = 0;
+    while offset < code_size {
+        let addr = start_addr + offset;
+
+        // Comments for this address
+        if let Some(comment_list) = cb.comments_at(addr) {
             for comment in comment_list {
                 writeln!(&mut out, "  {bold_begin}# {comment}{bold_end}").unwrap();
             }
         }
-        writeln!(&mut out, "  {}", format!("{insn}").trim()).unwrap();
+
+        // Decode one instruction
+        #[cfg(target_arch = "x86_64")]
+        let (text, len) = crate::disasm_x86_64::disassemble_one(&code_slice[offset..], addr);
+        #[cfg(target_arch = "aarch64")]
+        let (text, len) = {
+            if code_slice.len() - offset < 4 {
+                ("(truncated)".to_string(), code_slice.len() - offset)
+            } else {
+                let word = u32::from_le_bytes([
+                    code_slice[offset],
+                    code_slice[offset + 1],
+                    code_slice[offset + 2],
+                    code_slice[offset + 3],
+                ]);
+                (crate::disasm_aarch64::disassemble_instruction(word), 4)
+            }
+        };
+
+        writeln!(&mut out, "  0x{addr:x}: {text}").unwrap();
+        if len == 0 {
+            break;
+        }
+        offset += len;
     }
 
     out

zjit/src/lib.rs

@@ -23,6 +23,10 @@ mod virtualmem;
 mod asm;
 mod backend;
 #[cfg(feature = "disasm")]
+mod disasm_x86_64;
+#[cfg(feature = "disasm")]
+mod disasm_aarch64;
+#[cfg(feature = "disasm")]
 mod disasm;
 mod options;
 mod profile;