From 0c3a095cf178bf7ba4125912709df61acc8609f7 Mon Sep 17 00:00:00 2001
From: Valentin Churavy <v.churavy@gmail.com>
Date: Tue, 26 Aug 2025 11:43:18 +0200
Subject: [PATCH 1/2] High-level disassembly API

---
 src/LLVM.jl      |  1 +
 src/disasm.jl    | 38 ++++++++++++++++++++++++++++++++++++++
 test/disasm.jl   | 10 ++++++++++
 test/runtests.jl |  1 +
 4 files changed, 50 insertions(+)
 create mode 100644 src/disasm.jl
 create mode 100644 test/disasm.jl

diff --git a/src/LLVM.jl b/src/LLVM.jl
index 5d59ba7e..d332ce4c 100644
--- a/src/LLVM.jl
+++ b/src/LLVM.jl
@@ -80,6 +80,7 @@ include("debuginfo.jl")
 include("utils.jl")
 include("orc.jl")
 include("newpm.jl")
+include("disasm.jl")
 
 # high-level functionality
 include("state.jl")
diff --git a/src/disasm.jl b/src/disasm.jl
new file mode 100644
index 00000000..cc77135e
--- /dev/null
+++ b/src/disasm.jl
@@ -0,0 +1,38 @@
+@checked struct DisasmContext
+    ref::API.LLVMDisasmContextRef
+end
+
+function create_disasm(triple)
+    ctx = DisasmContext(API.LLVMCreateDisasm(triple, C_NULL, 0, C_NULL, C_NULL))
+    return ctx
+end
+
+function dispose(ctx::DisasmContext)
+    API.LLVMDisasmDispose(ctx.ref)
+    return nothing
+end
+
+function disassemble_code(io, ctx::DisasmContext, native_code::Vector{UInt8}, code_addr::Csize_t = 0; bufsize=32)
+    pos = 1
+    buf = Vector{UInt8}(undef, bufsize)
+
+    while pos < length(native_code)
+        GC.@preserve native_code begin
+            bytes = pointer(native_code, pos)
+            pc = code_addr + pos - 1
+            nbytes = length(native_code) - pos + 1
+            nb = API.LLVMDisasmInstruction(ctx, bytes, nbytes, pc, buf, bufsize)
+        end
+        if nb == 0
+            # Disassembly failed
+            break
+        end
+        pos += nb
+        for byte in buf
+            byte == 0x00 && break
+            write(io, byte)
+        end
+        write(io, '\n')
+    end
+    return nothing
+end
\ No newline at end of file
diff --git a/test/disasm.jl b/test/disasm.jl
new file mode 100644
index 00000000..a2d0607e
--- /dev/null
+++ b/test/disasm.jl
@@ -0,0 +1,10 @@
+@testset "X86 Disassembly" begin
+    # Example from jitdump
+    native_code = UInt8[0x55, 0x48, 0x89, 0xe5, 0x49, 0x8b, 0x45, 0x10, 0x48, 0x8b, 0x40, 0x10, 0x48, 0x8b, 0x00, 0x48, 0x8b, 0x16, 0x48, 0x83, 0xc6, 0x08, 0x48, 0xb8, 0xe0, 0x7c, 0xf5, 0x81, 0xe4, 0x7f, 0x00, 0x00, 0xff, 0xd0, 0x5d, 0xc3]
+    code_addr = 0x00007fe48befcde0
+
+    ctx = create_disasm("x86_64-pc-linux-gnu")
+    disassembled = sprint(disassemble_code, ctx, native_code, code_addr)
+
+    @test disassembled == "\tpushq\t%rbp\n\tmovq\t%rsp, %rbp\n\tmovq\t16(%r13), %rax\n\tmovq\t16(%rax), %rax\n\tmovq\t(%rax), %rax\n\tmovq\t(%rsi), %rdx\n\taddq\t\$8, %rsi\n\tmovabsq\t\$140619409620192, %rax\n\tcallq\t*%rax\n\tpopq\t%rbp\n"
+end
\ No newline at end of file
diff --git a/test/runtests.jl b/test/runtests.jl
index d97a0f1b..81ead7d5 100644
--- a/test/runtests.jl
+++ b/test/runtests.jl
@@ -41,6 +41,7 @@ include("debuginfo.jl")
 include("util.jl")
 include("interop.jl")
 include("orc.jl")
+include("disasm.jl")
 if !Sys.iswindows()
     # XXX: hangs on Windows
     include("jljit.jl")

From 2fd8490a4c846df20ce1b463d713491ea5a03e04 Mon Sep 17 00:00:00 2001
From: Valentin Churavy <v.churavy@gmail.com>
Date: Tue, 26 Aug 2025 12:00:54 +0200
Subject: [PATCH 2/2] fixup! High-level disassembly API

---
 src/disasm.jl  | 6 ++++--
 test/disasm.jl | 3 +++
 2 files changed, 7 insertions(+), 2 deletions(-)

diff --git a/src/disasm.jl b/src/disasm.jl
index cc77135e..fc8f3e98 100644
--- a/src/disasm.jl
+++ b/src/disasm.jl
@@ -1,6 +1,7 @@
 @checked struct DisasmContext
     ref::API.LLVMDisasmContextRef
 end
+Base.unsafe_convert(::Type{API.LLVMDisasmContextRef}, dc::DisasmContext) = dc.ref
 
 function create_disasm(triple)
     ctx = DisasmContext(API.LLVMCreateDisasm(triple, C_NULL, 0, C_NULL, C_NULL))
@@ -17,11 +18,12 @@ function disassemble_code(io, ctx::DisasmContext, native_code::Vector{UInt8}, co
     buf = Vector{UInt8}(undef, bufsize)
 
     while pos < length(native_code)
-        GC.@preserve native_code begin
+        GC.@preserve native_code buf begin
             bytes = pointer(native_code, pos)
             pc = code_addr + pos - 1
             nbytes = length(native_code) - pos + 1
-            nb = API.LLVMDisasmInstruction(ctx, bytes, nbytes, pc, buf, bufsize)
+            # Need to use pointer(buf) here since the API is annotated as Cstring
+            nb = API.LLVMDisasmInstruction(ctx, bytes, nbytes, pc, pointer(buf), bufsize)
         end
         if nb == 0
             # Disassembly failed
diff --git a/test/disasm.jl b/test/disasm.jl
index a2d0607e..679358cc 100644
--- a/test/disasm.jl
+++ b/test/disasm.jl
@@ -1,3 +1,5 @@
+import LLVM: create_disasm, disassemble_code
+
 @testset "X86 Disassembly" begin
     # Example from jitdump
     native_code = UInt8[0x55, 0x48, 0x89, 0xe5, 0x49, 0x8b, 0x45, 0x10, 0x48, 0x8b, 0x40, 0x10, 0x48, 0x8b, 0x00, 0x48, 0x8b, 0x16, 0x48, 0x83, 0xc6, 0x08, 0x48, 0xb8, 0xe0, 0x7c, 0xf5, 0x81, 0xe4, 0x7f, 0x00, 0x00, 0xff, 0xd0, 0x5d, 0xc3]
@@ -5,6 +7,7 @@
 
     ctx = create_disasm("x86_64-pc-linux-gnu")
     disassembled = sprint(disassemble_code, ctx, native_code, code_addr)
+    LLVM.dispose(ctx)
 
     @test disassembled == "\tpushq\t%rbp\n\tmovq\t%rsp, %rbp\n\tmovq\t16(%r13), %rax\n\tmovq\t16(%rax), %rax\n\tmovq\t(%rax), %rax\n\tmovq\t(%rsi), %rdx\n\taddq\t\$8, %rsi\n\tmovabsq\t\$140619409620192, %rax\n\tcallq\t*%rax\n\tpopq\t%rbp\n"
 end
\ No newline at end of file