Skip to content
Complete x86/x64 JIT and AOT Assembler for C++
C++ C JavaScript Other
Branch: master
Clone or download

README.md

AsmJit

Complete x86/x64 JIT and AOT Assembler for C++.

Introduction

AsmJit is a complete JIT and AOT assembler for C++ language. It can generate native code for x86 and x64 architectures and supports the whole x86/x64 instruction set - from legacy MMX to the newest AVX512. It has a type-safe API that allows C++ compiler to do semantic checks at compile-time even before the assembled code is generated and/or executed.

AsmJit, as the name implies, started as a project that provided JIT code-generation and execution. However, AsmJit evolved and it now contains features that are far beyond the scope of a simple JIT compilation. To keep the library small and lightweight the functionality not strictly related to JIT is provided by a sister project called asmtk.

Minimal Example

#include <asmjit/asmjit.h>
#include <stdio.h>

using namespace asmjit;

// Signature of the generated function.
typedef int (*Func)(void);

int main(int argc, char* argv[]) {
  JitRuntime rt;                          // Runtime specialized for JIT code execution.

  CodeHolder code;                        // Holds code and relocation information.
  code.init(rt.codeInfo());               // Initialize to the same arch as JIT runtime.

  x86::Assembler a(&code);                // Create and attach x86::Assembler to `code`.
  a.mov(x86::eax, 1);                     // Move one to 'eax' register.
  a.ret();                                // Return from function.
  // ----> x86::Assembler is no longer needed from here and can be destroyed <----

  Func fn;
  Error err = rt.add(&fn, &code);         // Add the generated code to the runtime.
  if (err) return 1;                      // Handle a possible error returned by AsmJit.
  // ----> CodeHolder is no longer needed from here and can be destroyed <----

  int result = fn();                      // Execute the generated code.
  printf("%d\n", result);                 // Print the resulting "1".

  // All classes use RAII, all resources will be released before `main()` returns,
  // the generated function can be, however, released explicitly if you intend to
  // reuse or keep the runtime alive, which you should in a production-ready code.
  rt.release(fn);

  return 0;
}

AsmJit Summary

  • Complete x86/x64 instruction set - MMX, SSE+, BMI+, ADX, TBM, XOP, AVX+, FMA+, and AVX512+.
  • Different emitters providing various abstraction levels (Assembler, Builder, Compiler).
  • Support for sections for separating code and data.
  • Built-in CPU vendor and features detection.
  • Advanced logging, formatting, and error handling.
  • JIT memory allocator - interface similar to malloc/free for JIT code-generation and execution.
  • Lightweight and easily embeddable - ~300kB compiled with all built-in features.
  • Modular design - unneeded features can be disabled at compile-time to make the library smaller.
  • Zero dependencies - no external libraries, no STL/RTTI - easy to embed and/or link statically.
  • Doesn't use exceptions internally, but allows to attach a "throwable" error handler of your choice.

Advanced Features

  • AsmJit contains a highly compressed instruction database:
    • Instruction names - allows to convert instruction id to its name and vice versa.
    • Instruction metadata - access (read|write) of all operand combinations of all instructions.
    • Instruction signatures - allows to strictly validate if an instruction (with all its operands) is valid.
  • AsmJit allows to precisely control how instructions are encoded if there are multiple variations.
  • AsmJit is highly dynamic, constructing operands at runtime is a common practice.
  • Multiple emitters with the same interface - emit machine code directly or to a representation that can be post-processed.

Important

Breaking the official API is sometimes inevitable, what to do?

TODO

  • Add support for user external buffers in CodeHolder.

Supported Environments

C++ Compilers:

  • Requirements:
    • AsmJit won't build without C++11 enabled. If you use older GCC or Clang you would have to enable at least c++11 through compiler flags.
  • Tested:
    • Clang - tested by Travis-CI - Clang 3.9+ (with C++11 enabled) is officially supported (older Clang versions having C++11 support are probably fine, but are not regularly tested).
    • GNU - tested by Travis-CI - GCC 4.8+ (with C++11 enabled) is officially supported.
    • MINGW - tested by Travis-CI - Use the latest version, if possible.
    • MSVC - tested by Travis-CI - MSVC2017+ only! - there is a severe bug in MSVC2015's constexpr implementation that makes that compiler unusable.
  • Untested:
    • Intel - no maintainers and no CI environment to regularly test this compiler.
    • Other c++ compilers would require basic support in core/build.h.

Operating Systems:

  • Tested:
    • Linux - tested by Travis-CI - any distribution is generally supported.
    • OSX - tested by Travis-CI - any version is supported.
    • Windows - tested by Travis-CI - Windows 7+ is officially supported.
  • Untested:
    • BSDs - no maintainers, no CI environment to regularly test these OSes.
    • Haiku - not regularly tested, but reported to work.
    • Other operating systems would require some testing and support in core/build.h, core/osutils.cpp, and core/virtmem.cpp.

Backends:

  • X86 - tested by both Travis-CI - both 32-bit and 64-bit backends are fully functional.
  • ARM - work-in-progress (not public at the moment).

Project Organization

  • / - Project root.
    • src - Source code.
      • asmjit - Source code and headers (always point include path in here).
        • core - Core API, backend independent except relocations.
        • arm - ARM specific API, used only by ARM and AArch64 backends.
        • x86 - X86 specific API, used only by X86 and X64 backends.
    • test - Unit and integration tests (don't embed in your project).
    • tools - Tools used for configuring, documenting and generating data files.

Configuring & Feature Selection

AsmJit is designed to be easy embeddable in any project. However, it depends on some compile-time macros that can be used to build a specific version of AsmJit that includes or excludes certain features. A typical way of building AsmJit is to use cmake, but it's also possible to just include AsmJit source code in your project and just build it. The easiest way to include AsmJit in your project is to just include src directory in your project and to define ASMJIT_STATIC. AsmJit can be just updated from time to time without any changes to this integration process. Do not embed AsmJit's /test files in such case as these are used for testing.

Build Type:

  • ASMJIT_BUILD_DEBUG - Define to always turn debugging on (regardless of compile-time options detected).
  • ASMJIT_BUILD_RELEASE - Define to always turn debugging off (regardless of compile-time options detected).

By default none of these is defined, AsmJit detects build-type based on compile-time macros and supports most IDE and compiler settings out of box. By default AsmJit switches to release mode when NDEBUG is defined.

Build Mode:

  • ASMJIT_STATIC - Define to build AsmJit statically - either as a static library or as a part of another project. No symbols are exported in such case.

By default AsmJit build is configured to be built as a shared library, this means ASMJIT_STATIC must be explicitly enabled if you want to compile AsmJit statically.

Build Backends:

  • ASMJIT_BUILD_ARM - Build ARM backends (not ready, work-in-progress).
  • ASMJIT_BUILD_X86 - Build X86 backends (X86 and X86_64).
  • ASMJIT_BUILD_HOST - Build only the host backend (default).

If none of ASMJIT_BUILD_... is defined AsmJit bails to ASMJIT_BUILD_HOST, which will detect the target architecture at compile-time. Each backend automatically supports 32-bit and 64-bit targets, so for example AsmJit with X86 support can generate both 32-bit and 64-bit code.

Disabling Features:

  • ASMJIT_NO_BUILDER - Disables both Builder and Compiler emitters (only Assembler will be available). Ideal for users that don't use Builder concept and want to have AsmJit a bit smaller.
  • ASMJIT_NO_COMPILER - Disables Compiler emitter. For users that use Builder, but not Compiler.
  • ASMJIT_NO_JIT - Disables JIT execution engine, which includes JitUtils, JitAllocator, and JitRuntime.
  • ASMJIT_NO_LOGGING - Disables logging (Logger and all classes that inherit it) and instruction formatting.
  • ASMJIT_NO_TEXT - Disables everything that uses text-representation and that causes certain strings to be stored in the resulting binary. For example when this flag is set all instruction and error names (and related APIs) will not be available. This flag has to be disabled together with ASMJIT_NO_LOGGING. This option is suitable for deployment builds or builds that don't want to reveal the use of AsmJit.
  • ASMJIT_NO_INST_API - Disables instruction query features, strict validation, read/write information, and all additional data and APIs that can output information about instructions.

NOTE: Please don't disable any features if you plan to build AsmJit as a shared library that will be used by multiple projects that you don't control (for example asmjit in a Linux distribution). The possibility to disable certain features exists mainly for customized builds of AsmJit.

Using AsmJit

AsmJit library uses one global namespace called asmjit that provides the whole functionality. Architecture specific code is prefixed by the architecture name and architecture specific registers and operand builders have their own namespace. For example API targeting both X86 and X64 architectures is prefixed with X86 and registers & operand builders are accessible through x86 namespace. This design is very different from the initial version of AsmJit and it seems now as the most convenient one.

CodeHolder & Emitters

AsmJit provides two classes that are used together for code generation:

  • CodeHolder - Provides functionality to hold generated code and stores all necessary information about code sections, labels, symbols, and possible relocations.
  • BaseEmitter - Provides functionality to emit code into CodeHolder. BaseEmitter is abstract and provides just basic building blocks that are then implemented by BaseAssembler, BaseBuilder, BaseCompiler, and their architecture-specific implementations like x86::Assembler, x86::Builder, and x86::Compiler.

Code emitters:

  • [Base]Assembler - Emitter designed to emit machine code directly into a CodeBuffer held by CodeHolder.
  • [Base]Builder - Emitter designed to emit code into a representation that can be processed afterwards. It stores the whole code in a double linked list consisting of nodes (BaseNode and all derived classes). There are nodes that represent instructions (InstNode), labels (LabelNode), and other building blocks (AlignNode, DataNode, ...). Some nodes are used as markers (SentinelNode and comments (CommentNode).
  • [Base]Compiler - High-level code emitter that uses virtual registers and contains high-level function building features. Compiler extends [Base]Builder functionality and introduces new nodes like FuncNode, FuncRetNode, and FuncCallNode. Compiler is the simplest way to start with AsmJit as it abstracts lots of details required to generate a function that can be called from a C/C++ language.

Targets and JitRuntime

AsmJit's Target class is an interface that provides basic target abstraction. At the moment only one implementation called JitRuntime is provided, which as the name suggests provides JIT code target and execution runtime. JitRuntime provides all the necessary functionality to implement a simple JIT functionality with basic memory management. It only provides add() and release() functions that are used to either add code to the runtime or release it. The JitRuntime doesn't do any decisions on when the code should be released. Once you add new code into it you must decide when that code is no longer needed and should be released.

Instructions & Operands

Instructions specify operations performed by the CPU, and operands specify the operation's input(s) and output(s). Each AsmJit's instruction has it's own unique id (Inst::Id for example) and platform specific code emitters always provide a type safe intrinsic (or multiple overloads) to emit such instruction. There are two ways of emitting an instruction:

  • Using BaseEmitter::inst(operands...) - A type-safe way provided by platform specific emitters - for example x86::Assembler provides x86::Assembler::mov(x86::Gp, x86::Gp).
  • Using BaseEmitter::emit(instId, operands...) - Allows to emit an instruction in a dynamic way - you just need to know instruction's id and provide its operands.

AsmJit's operands all inherit from a base class called Operand and then specialize its type to:

  • None (not used or uninitialized operand).
  • Register (BaseReg) - Describes either physical or virtual register. Physical registers have id that matches the target's machine id directly whereas virtual registers must be allocated into physical registers by a register allocator pass. Register operand provides:
    • Register Type - Unique id that describes each possible register provided by the target architecture - for example X86 backend provides x86::Reg::RegType, which defines all variations of general purpose registers (GPB-LO, GPB-HI, GPW, GPD, and GPQ) and all types of other registers like K, MM, BND, XMM, YMM, and ZMM.
    • Register Group - Groups multiple register types under a single group - for example all general-purpose registers (of all sizes) on X86 are x86::Reg::kGroupGp, all SIMD registers (XMM, YMM, ZMM) are x86::Reg::kGroupVec, etc.
    • Register Size - Contains the size of the register in bytes. If the size depends on the mode (32-bit vs 64-bit) then generally the higher size is used (for example RIP register has size 8 by default).
    • Register ID - Contains physical or virtual id of the register.
    • Each architecture provides its own register that adds a architecture-specific API to BaseReg.
  • Memory Address (BaseMem) - Used to reference a memory location. Memory operand provides:
    • Base Register - A base register type and id (physical or virtual).
    • Index Register - An index register type and id (physical or virtual).
    • Offset - Displacement or absolute address to be referenced (32-bit if base register is used and 64-bit if base register is not used).
    • Flags that can describe various architecture dependent information (like scale and segment-override on X86).
    • Each architecture provides its own register that adds a architecture-specific API to BaseMem.
  • Immediate Value (Imm) - Immediate values are usually part of instructions (encoded within the instruction itself) or data.
  • Label - used to reference a location in code or data. Labels must be created by the BaseEmitter or by CodeHolder. Each label has its unique id per CodeHolder instance.

AsmJit allows to construct operands dynamically, to store them, and to query a complete information about them at run-time. Operands are small (always 16 bytes per Operand) and should be always copied (by value) if you intend to store them (don't create operands by using new keyword, it's not recommended). Operands are safe to be memcpy()ed and memset()ed if you need to work with arrays of operands.

Small example of manipulating and using operands:

#include <asmjit/asmjit.h>

using namespace asmjit;

x86::Gp dstRegByValue() { return x86::ecx; }

void usingOperandsExample(x86::Assembler& a) {
  // Create some operands.
  x86::Gp dst = dstRegByValue();          // Get `ecx` register returned by a function.
  x86::Gp src = x86::rax;                 // Get `rax` register directly from the provided `x86` namespace.
  x86::Gp idx = x86::gpq(10);             // Construct `r10` dynamically.
  x86::Mem m = x86::ptr(src, idx);        // Construct [src + idx] memory address - referencing [rax + r10].

  // Examine `m`:
  m.indexType();                          // Returns `x86::Reg::kTypeGpq`.
  m.indexId();                            // Returns 10 (`r10`).

  // Reconstruct `idx` stored in mem:
  x86::Gp idx_2 = x86::Gp::fromTypeAndId(m.indexType(), m.indexId());
  idx == idx_2;                           // True, `idx` and idx_2` are identical.

  Operand op = m;                         // Possible.
  op.isMem();                             // True (can be casted to BaseMem or architecture-specific Mem).

  m == op;                                // True, `op` is just a copy of `m`.
  static_cast<BaseMem&>(op).addOffset(1); // Static cast is fine and valid here.
  op.as<BaseMem>().addOffset(1);          // However, using `as<T>()` to cast to a derived type is preferred.
  m == op;                                // False, `op` now points to [rax + r10 + 1], which is not [rax + r10].

  // Emitting 'mov'
  a.mov(dst, m);                          // Type-safe way.
  a.mov(dst, op);                         // Not possible, `mov` doesn't provide `mov(x86::Gp, Operand)` overload.

  a.emit(x86::Inst::kIdMov, dst, m);      // Type-unsafe, but possible.
  a.emit(x86::Inst::kIdMov, dst, op);     // Also possible, `emit()` is typeless and can be used with raw `Operand`s.
}

Some operands have to be created explicitly by BaseEmitter. For example labels must be created by newLabel() before they are used.

Assembler Example

x86::Assembler is a code emitter that emits machine code into a CodeBuffer directly. It's capable of targeting both 32-bit and 64-bit instruction sets and it's possible to target both instruction sets within the same code-base. The following example shows how to generate a function that works in both 32-bit and 64-bit modes, and how to use JitRuntime, CodeHolder, and x86::Assembler together.

The example handles 3 calling conventions manually just to show how it could be done, however, AsmJit contains utilities that can be used to create function prologs and epilogs automatically, but these concepts will be explained later.

#include <asmjit/asmjit.h>
#include <stdio.h>

using namespace asmjit;

// Signature of the generated function.
typedef int (*SumFunc)(const int* arr, size_t count);

int main(int argc, char* argv[]) {
  JitRuntime jit;                         // Create a runtime specialized for JIT.
  CodeHolder code;                        // Create a CodeHolder.

  code.init(jit.codeInfo());              // Initialize it to be compatible with `jit`.
  x86::Assembler a(&code);                // Create and attach x86::Assembler to `code`.

  // Decide between 32-bit CDECL, WIN64, and SysV64 calling conventions:
  //   32-BIT - passed all arguments by stack.
  //   WIN64  - passes first 4 arguments by RCX, RDX, R8, and R9.
  //   UNIX64 - passes first 6 arguments by RDI, RSI, RCX, RDX, R8, and R9.
  x86::Gp arr, cnt;
  x86::Gp sum = x86::eax;                 // Use EAX as 'sum' as it's a return register.

  if (ASMJIT_ARCH_BITS == 64) {
  #if defined(_WIN32)
    arr = x86::rcx;                       // First argument (array ptr).
    cnt = x86::rdx;                       // Second argument (number of elements)
  #else
    arr = x86::rdi;                       // First argument (array ptr).
    cnt = x86::rsi;                       // Second argument (number of elements)
  #endif
  }
  else {
    arr = x86::edx;                       // Use EDX to hold the array pointer.
    cnt = x86::ecx;                       // Use ECX to hold the counter.
    a.mov(arr, x86::ptr(x86::esp, 4));    // Fetch first argument from [ESP + 4].
    a.mov(cnt, x86::ptr(x86::esp, 8));    // Fetch second argument from [ESP + 8].
  }

  Label Loop = a.newLabel();              // To construct the loop, we need some labels.
  Label Exit = a.newLabel();

  a.xor_(sum, sum);                       // Clear 'sum' register (shorter than 'mov').
  a.test(cnt, cnt);                       // Border case:
  a.jz(Exit);                             //   If 'cnt' is zero jump to 'Exit' now.

  a.bind(Loop);                           // Start of a loop iteration.
  a.add(sum, x86::dword_ptr(arr));        // Add int at [arr] to 'sum'.
  a.add(arr, 4);                          // Increment 'arr' pointer.
  a.dec(cnt);                             // Decrease 'cnt'.
  a.jnz(Loop);                            // If not zero jump to 'Loop'.

  a.bind(Exit);                           // Exit to handle the border case.
  a.ret();                                // Return from function ('sum' == 'eax').
  // ----> x86::Assembler is no longer needed from here and can be destroyed <----

  SumFunc fn;
  Error err = jit.add(&fn, &code);        // Add the generated code to the runtime.

  if (err) return 1;                      // Handle a possible error returned by AsmJit.
  // ----> CodeHolder is no longer needed from here and can be destroyed <----

  static const int array[6] = { 4, 8, 15, 16, 23, 42 };

  int result = fn(array, 6);              // Execute the generated code.
  printf("%d\n", result);                 // Print sum of array (108).

  jit.release(fn);                        // Remove the function from the runtime.
  return 0;
}

The example should be self-explanatory. It shows how to work with labels, how to use operands, and how to emit instructions that can use different registers based on runtime selection. It implements 32-bit CDECL, WIN64, and SysV64 caling conventions and will work on most X86 environments.

More About Memory Addresses

X86 provides a complex memory addressing model that allows to encode addresses having a BASE register, INDEX register with a possible scale (left shift), and displacement (called offset in AsmJit). Memory address can also specify memory segment (segment-override in X86 terminology) and some instructions (gather / scatter) require INDEX to be a VECTOR register instead of a general-purpose register. AsmJit allows to encode and work with all forms of addresses mentioned and implemented by X86. It also allows to construct a 64-bit memory address, which is only allowed in one form of 'mov' instruction.

#include <asmjit/asmjit.h>

// Memory operand construction is provided by x86 namespace.
using namespace asmjit;
using namespace asmjit::x86;              // Easier to access x86 regs.

// BASE + OFFSET.
x86::Mem a = ptr(rax);                    // a = [rax]
x86::Mem b = ptr(rax, 15)                 // b = [rax + 15]

// BASE + INDEX << SHIFT - Shift is in BITS as used by X86!
x86::Mem c = ptr(rax, rbx)                // c = [rax + rbx]
x86::Mem d = ptr(rax, rbx, 2)             // d = [rax + rbx << 2]
x86::Mem e = ptr(rax, rbx, 2, 15)         // e = [rax + rbx << 2 + 15]

// BASE + VM (Vector Index) (encoded as MOD+VSIB).
x86::Mem f = ptr(rax, xmm1)               // f = [rax + xmm1]
x86::Mem g = ptr(rax, xmm1, 2)            // g = [rax + xmm1 << 2]
x86::Mem h = ptr(rax, xmm1, 2, 15)        // h = [rax + xmm1 << 2 + 15]

// WITHOUT BASE:
uint64_t ADDR = (uint64_t)0x1234;
x86::Mem i = ptr(ADDR);                   // i = [0x1234]
x86::Mem j = ptr(ADDR, rbx);              // j = [0x1234 + rbx]
x86::Mem k = ptr(ADDR, rbx, 2);           // k = [0x1234 + rbx << 2]

// LABEL - Will be encoded as RIP (64-bit) or absolute address (32-bit).
Label L = ...;
x86::Mem m = ptr(L);                      // m = [L]
x86::Mem n = ptr(L, rbx);                 // n = [L + rbx]
x86::Mem o = ptr(L, rbx, 2);              // o = [L + rbx << 2]
x86::Mem p = ptr(L, rbx, 2, 15);          // p = [L + rbx << 2 + 15]

// RIP - 64-bit only (RIP can't use INDEX).
x86::Mem q = ptr(rip, 24);                // q = [rip + 24]

Memory operands can optionally contain memory size. This is required by instructions where the memory size cannot be deduced from other operands, like inc and dec:

x86::Mem a = x86::dword_ptr(rax, rbx);       // dword ptr [rax + rbx].
x86::Mem b = x86::qword_ptr(rdx, rsi, 0, 1); // qword ptr [rdx + rsi << 0 + 1].

Memory operands provide API that can be used to work with them:

x86::Mem mem = x86::dword_ptr(rax, 12);   // dword ptr [rax + 12].

mem.hasBase();                            // true.
mem.hasIndex();                           // false.
mem.size();                               // 4.
mem.offset();                             // 12.

mem.setSize(0);                           // Sets the size to 0 (makes it sizeless).
mem.addOffset(-1);                        // Adds -1 to the offset and makes it 11.
mem.setOffset(0);                         // Sets the offset to 0.
mem.setBase(rcx);                         // Changes BASE to RCX.
mem.setIndex(rax);                        // Changes INDEX to RAX.
mem.hasIndex();                           // true.

// ...

Making changes to memory operand is very comfortable when emitting loads and stores:

#include <asmjit/asmjit.h>

using namespace asmjit;

x86::Assembler a(...);                    // Your initialized x86::Assembler.
x86::Mem m = x86::ptr(eax);               // Construct [eax] memory operand.

// One way of emitting bunch of loads is to use `mem.adjusted()`. It returns
// a new memory operand and keeps the source operand unchanged.
a.movaps(x86::xmm0, m);                   // No adjustment needed to load [eax].
a.movaps(x86::xmm1, m.adjusted(16));      // Loads from [eax + 16].
a.movaps(x86::xmm2, m.adjusted(32));      // Loads from [eax + 32].
a.movaps(x86::xmm3, m.adjusted(48));      // Loads from [eax + 48].

// ... do something with xmm0-3 ...

// Another way of adjusting memory is to change the operand in-place. If you
// want to keep the original operand you can simply clone it.
x86::Mem mx = m.clone();
a.movaps(mx, x86::xmm0); mx.addOffset(16);// Stores to [eax]      (and adds 16 to mx).
a.movaps(mx, x86::xmm1); mx.addOffset(16);// Stores to [eax + 16] (and adds 16 to mx).
a.movaps(mx, x86::xmm2); mx.addOffset(16);// Stores to [eax + 32] (and adds 16 to mx).
a.movaps(mx, x86::xmm3);                  // Stores to [eax + 48].

You can explore the possibilities by taking a look at:

More About CodeInfo

In the first complete example the CodeInfo is retrieved from JitRuntime. It's logical as JitRuntime will always return a CodeInfo that is compatible with the runtime environment. For example if your application runs in 64-bit mode the CodeInfo will use ArchInfo::kIdX64 architecture in contrast to ArchInfo::kIdX86, which will be used in 32-bit mode. AsmJit also allows to setup CodeInfo manually, and to select a different architecture when needed. So let's do something else this time, let's always generate a 32-bit code and print it's binary representation. To do that, we create our own CodeInfo and initialize it to ArchInfo::kIdX86 architecture. CodeInfo will populate all basic fields just based on the architecture we provide, so it's super-easy:

#include <asmjit/asmjit.h>
#include <stdio.h>

using namespace asmjit;

int main(int argc, char* argv[]) {
  using namespace asmjit::x86;            // Easier access to x86/x64 registers.

  CodeHolder code;                        // Create a CodeHolder.
  code.init(CodeInfo(ArchInfo::kIdX86));// Initialize it for a 32-bit X86 target.

  // Generate a 32-bit function that sums 4 floats and looks like:
  //   void func(float* dst, const float* a, const float* b)
  x86::Assembler a(&code);                // Create and attach x86::Assembler to `code`.

  a.mov(eax, dword_ptr(esp, 4));          // Load the destination pointer.
  a.mov(ecx, dword_ptr(esp, 8));          // Load the first source pointer.
  a.mov(edx, dword_ptr(esp, 12));         // Load the second source pointer.

  a.movups(xmm0, ptr(ecx));               // Load 4 floats from [ecx] to XMM0.
  a.movups(xmm1, ptr(edx));               // Load 4 floats from [edx] to XMM1.
  a.addps(xmm0, xmm1);                    // Add 4 floats in XMM1 to XMM0.
  a.movups(ptr(eax), xmm0);               // Store the result to [eax].
  a.ret();                                // Return from function.

  // We have no Runtime this time, it's on us what we do with the code.
  // CodeHolder stores code in `Section`, which provides some basic properties
  // and CodeBuffer structure. We are interested in section's CodeBuffer only.
  //
  // NOTE: The first section is always '.text', so it's safe to just use 0 index.
  // Get it by using either `code.sectionById(0)` or `code.textSection()`.
  CodeBuffer& buffer = code.sectionById(0)->buffer();

  // Print the machine-code generated or do something more interesting with it?
  //   8B4424048B4C24048B5424040F28010F58010F2900C3
  for (size_t i = 0; i < buffer.length; i++)
    printf("%02X", buffer.data[i]);

  return 0;
}

Explicit Code Relocation

CodeInfo contains much more information than just the target architecture. It can be configured to specify a base-address (or a virtual base-address in a linker terminology), which could be static (useful when you know the location of the target's machine code) or dynamic. AsmJit assumes dynamic base-address by default and relocates the code held by CodeHolder to a user-provided address on-demand. To be able to relocate to a user-provided address it needs to store some information about relocations, which is represented by RelocEntry. Relocation entries are only required if you call external functions from the generated code that cannot be encoded by using a 32-bit displacement (X64 architecture doesn't provide an encodable 64-bit displacement).

There is also a concept called LabelLink - label links are lightweight structs that don't have any identifier and are stored per label in a single-linked list. Label links represent either unbound yet used labels (that are valid in cases in which label was not bound but was already referenced by an instruction) and links that cross-sections (only relevant to code that uses multiple sections). Since crossing sections is something that cannot be resolved immediately these links persist until offsets of these sections are assigned and CodeHolder::resolveUnresolvedLinks() is called. It's an error if you end up with code that has unresolved label links after flattening. You can verify it by calling CodeHolder::hasUnresolvedLinks() and CodeHolder::unresolvedLinkCount().

AsmJit can flatten code that uses multiple sections by assigning each section an incrementing offset that respects its alignment. Use CodeHolder::flatten() to do that. After the sections are flattened their offsets and virtual-sizes were adjusted to respect section's buffer size and alignment. You must call CodeHolder::resolveUnresolvedLinks() before relocating the code held by it. You can also flatten your code manually by iterating over all sections and calculating their offsets (relative to base) by your own algorithm. In that case you don't have to call CodeHolder::flatten(), but you must still call CodeHolder::resolveUnresolvedLinks().

Next example shows how to use a built-in virtual memory allocator JitAllocator instead of using JitRuntime (just in case you want to use your own memory management) and how to relocate the generated code into your own memory block - you can use your own virtual memory allocator if you prefer that, but that's OS specific and it's already provided by AsmJit, so we will use what AsmJit offers instead of going deep into OS specific APIs.

The following code is similar to the previous one, but implements a function working in both 32-bit and 64-bit environments:

#include <asmjit/asmjit.h>
#include <stdio.h>

using namespace asmjit;

typedef void (*SumIntsFunc)(int* dst, const int* a, const int* b);

int main(int argc, char* argv[]) {
  CodeHolder code;                        // Create a CodeHolder.
  code.init(CodeInfo(ArchInfo::kIdHost)); // Initialize it for the host architecture.

  x86::Assembler a(&code);                // Create and attach x86::Assembler to `code`.

  // Generate a function runnable in both 32-bit and 64-bit architectures:
  bool isX86 = ASMJIT_ARCH_X86 == 32;

  // Signature: 'void func(int* dst, const int* a, const int* b)'.
  x86::Gp dst;
  x86::Gp src_a;
  x86::Gp src_b;

  // Handle the difference between 32-bit and 64-bit calling convention.
  // (arguments passed through stack vs. arguments passed by registers).
  if (isX86) {
    dst   = x86::eax;
    src_a = x86::ecx;
    src_b = x86::edx;
    a.mov(dst  , x86::dword_ptr(x86::esp, 4));  // Load the destination pointer.
    a.mov(src_a, x86::dword_ptr(x86::esp, 8));  // Load the first source pointer.
    a.mov(src_b, x86::dword_ptr(x86::esp, 12)); // Load the second source pointer.
  }
  else {
  #if defined(_WIN32)
    dst   = x86::rcx;                     // First argument  (destination pointer).
    src_a = x86::rdx;                     // Second argument (source 'a' pointer).
    src_b = x86::r8;                      // Third argument  (source 'b' pointer).
  #else
    dst   = x86::rdi;                     // First argument  (destination pointer).
    src_a = x86::rsi;                     // Second argument (source 'a' pointer).
    src_b = x86::rdx;                     // Third argument  (source 'b' pointer).
  #endif
  }

  a.movdqu(x86::xmm0, x86::ptr(src_a));   // Load 4 ints from [src_a] to XMM0.
  a.movdqu(x86::xmm1, x86::ptr(src_b));   // Load 4 ints from [src_b] to XMM1.
  a.paddd(x86::xmm0, x86::xmm1);          // Add 4 ints in XMM1 to XMM0.
  a.movdqu(x86::ptr(dst), x86::xmm0);     // Store the result to [dst].
  a.ret();                                // Return from function.

  // Even when we didn't use multiple sections AsmJit could insert one section
  // called '.addrtab' (address table section), which would be filled by data
  // required by relocations (absolute jumps and calls). You can omit this code
  // if you are 100% sure your code doesn't contain multiple sections and
  // such relocations. You can use `CodeHolder::hasAddressTable()` to verify
  // whether the address table section does exist.
  code.flatten();
  code.resolveUnresolvedLinks();

  // After the code was generated it can be relocated manually to any memory
  // location, however, we need to know it's size before we perform memory
  // allocation. `CodeHolder::codeSize()` returns the worst estimated code
  // size in case that relocations are not possible without trampolines (in
  // that case some extra code at the end of the current code buffer is
  // generated during relocation).
  size_t estimatedSize = code.codeSize();

  // Instead of rolling up our own memory allocator we can use the one AsmJit
  // provides. It's decoupled so you don't need to use `JitRuntime` for that.
  JitAllocator allocator;

  // Allocate an executable virtual memory and handle a possible failure.
  void* p = allocator.alloc(estimatedSize);
  if (!p) return 0;

  // Now relocate the code to the address provided by the memory allocator.
  // Please note that this DOESN'T COPY anything to `p`. This function will
  // store the address in CodeInfo and use relocation entries to patch the
  // existing code in all sections to respect the base address provided.
  code.relocateToBase((uint64_t)p);

  // This is purely optional. There are cases in which the relocation can
  // omit unneeded data, which would shrink the size of address table. If
  // that happened the `codeSize` returned after `relocateToBase()` would
  // be smaller than the originally `estimatedSize`.
  size_t codeSize = code.codeSize();

  // This will copy code from all sections to `p`. Iterating over all
  // sections and calling `memcpy()` would work as well, however, this
  // function supports additional options that can be used to also zero
  // pad sections' virtual size, etc.
  //
  // With some additional features, copyFlattenData() does roughly this:
  //   for (Section* section : code.sections())
  //     memcpy((uint8_t*)p + section->offset(),
  //            section->data(),
  //            section->bufferSize());
  code.copyFlattenedData(p, codeSize, CodeHolder::kCopyWithPadding);

  // Execute the generated function.
  int inA[4] = { 4, 3, 2, 1 };
  int inB[4] = { 1, 5, 2, 8 };
  int out[4];

  // This code uses AsmJit's ptr_as_func<> to cast between void* and SumIntsFunc.
  ptr_as_func<SumIntsFunc>(p)(out, inA, inB);

  // Prints {5 8 4 9}
  printf("{%d %d %d %d}\n", out[0], out[1], out[2], out[3]);

  // Release 'p' is it's no longer needed. It will be destroyed with 'vm'
  // instance anyway, but it's a good practice to release it explicitly
  // when you know that the function will not be needed anymore.
  allocator.release(p);

  return 0;
}

If you know your base-address in advance (before code generation) you can use CodeInfo::setBaseAddress() to setup its initial value. In that case Assembler will know the absolute position of each instruction and would be able to use it during instruction encoding and prevent relocations in case the instruction is encodable. The following example shows how to configure the base address:

// Configure CodeInfo with base address.
CodeInfo ci(...);
ci.setBaseAddress(uint64_t(0x1234));

// Then initialize CodeHolder with it.
CodeHolder code;
code.init(ci);

Using Native Registers - zax, zbx, zcx, ...

AsmJit's X86 code emitters always provide functions to construct machine-size registers depending on the target. This feature is for people that want to write code targeting both 32-bit and 64-bit at the same time. In AsmJit terminology these registers are named zax, zcx, zdx, zbx, zsp, zbp, zsi, and zdi (they are defined in this exact order by X86). They are accessible through x86::Assembler, x86::Builder, and x86::Compiler. The following example illustrates how to use this feature:

#include <asmjit/asmjit.h>
#include <stdio.h>

using namespace asmjit;

typedef int (*Func)(void);

int main(int argc, char* argv[]) {
  JitRuntime jit;                         // Create a runtime specialized for JIT.
  CodeHolder code;                        // Create a CodeHolder.

  code.init(jit.codeInfo());              // Initialize it to be compatible with `jit`.
  x86::Assembler a(&code);                // Create and attach x86::Assembler to `code`.

  // Let's get these registers from x86::Assembler.
  x86::Gp zbp = a.zbp();
  x86::Gp zsp = a.zsp();

  int stackSize = 32;

  // Function prolog.
  a.push(zbp);
  a.mov(zbp, zsp);
  a.sub(zsp, stackSize);

  // ... emit some code (this just sets return value to zero) ...
  a.xor_(x86::eax, x86::eax);

  // Function epilog and return.
  a.mov(zsp, zbp);
  a.pop(zbp);
  a.ret();

  // To make the example complete let's call it.
  Func fn;
  Error err = jit.add(&fn, &code);        // Add the generated code to the runtime.
  if (err) return 1;                      // Handle a possible error returned by AsmJit.

  int result = fn();                      // Execute the generated code.
  printf("%d\n", result);                 // Print the resulting "0".

  jit.release(fn);                        // Remove the function from the runtime.
  return 0;
}

The example just returns 0, but the function generated contains a standard prolog and epilog sequence and the function itself reserves 32 bytes of local stack. The advantage is clear - a single code-base can handle multiple targets easily. If you want to create a register of native size dynamically by specifying its id it's also possible:

void example(x86::Assembler& a) {
  x86::Gp zax = a.gpz(x86::Gp::kIdAx);
  x86::Gp zbx = a.gpz(x86::Gp::kIdBx);
  x86::Gp zcx = a.gpz(x86::Gp::kIdCx);
  x86::Gp zdx = a.gpz(x86::Gp::kIdDx);

  // You can also change register's id easily.
  x86::Gp zsp = zax;
  zsp.setId(4); // or x86::Gp::kIdSp.
}

Cloning existing registers and chaning their IDs is fine in AsmJit; and this technique is used internally in many places.

Using Assembler as Code-Patcher

This is an advanced topic that is sometimes unavoidable. AsmJit by default appends machine-code it generates into a CodeBuffer, however, it also allows to set the offset in CodeBuffer explicitly and to overwrite its content. This technique is extremely dangerous for asm beginners as X86 instructions have variable length (see below), so you should in general only patch code to change instruction's offset or some basic other details you didn't know about the first time you emitted it. A typical scenario that requires code-patching is when you start emitting function and you don't know how much stack you want to reserve for it.

Before we go further it's important to introduce instruction options, because they can help with code-patching (and not only patching, but that will be explained in AVX-512 section):

  • Many general-purpose instructions (especially arithmetic ones) on X86 have multiple encodings - in AsmJit this is usually called 'short form' and 'long form'.
  • AsmJit always tries to use 'short form' as it makes the resulting machine-code smaller, which is always good - this decision is used by majority of assemblers out there.
  • AsmJit allows to override the default decision by using short_() and long_() instruction options to force short or long form, respectively. The most useful is long_() as it basically forces AsmJit to always emit the long form. The short_() is not that useful as it's automatic (except jumps to non-bound labels). Note the underscore after each function name as it avoids collision with built-in C++ types.

To illustrate what short form and long form means in binary let's assume we want to emit add esp, 16 instruction, which has two possible binary encodings:

  • 83C410 - This is a short form aka short add esp, 16 - You can see opcode byte (0x8C), MOD/RM byte (0xC4) and an 8-bit immediate value representing 16.
  • 81C410000000 - This is a long form aka long add esp, 16 - You can see a different opcode byte (0x81), the same Mod/RM byte (0xC4) and a 32-bit immediate in little-endian representing 16.

If you generate an instruction in a short form and then patch it in a long form or vice-versa then something really bad will happen when you try to execute such code. The following example illustrates how to patch the code properly (it just extends the previous example):

#include <asmjit/asmjit.h>
#include <stdio.h>

using namespace asmjit;

typedef int (*Func)(void);

int main(int argc, char* argv[]) {
  JitRuntime jit;                         // Create a runtime specialized for JIT.
  CodeHolder code;                        // Create a CodeHolder.

  code.init(jit.codeInfo());              // Initialize it to be compatible with `jit`.
  x86::Assembler a(&code);                // Create and attach x86::Assembler to `code`.

  // Let's get these registers from x86::Assembler.
  x86::Gp zbp = a.zbp();
  x86::Gp zsp = a.zsp();

  // Function prolog.
  a.push(zbp);
  a.mov(zbp, zsp);

  // This is where we are gonna patch the code later, so let's get the offset
  // (the current location) from the beginning of the code-buffer.
  size_t patchOffset = a.offset();
  // Let's just emit 'sub zsp, 0' for now, but don't forget to use LONG form.
  a.long_().sub(zsp, 0);

  // ... emit some code (this just sets return value to zero) ...
  a.xor_(x86::eax, x86::eax);

  // Function epilog and return.
  a.mov(zsp, zbp);
  a.pop(zbp);
  a.ret();

  // Now we know how much stack size we want to reserve. I have chosen 128
  // bytes on purpose as it's encodable only in long form that we have used.

  int stackSize = 128;                    // Number of bytes to reserve on the stack.
  a.setOffset(patchOffset);               // Move the current cursor to `patchOffset`.
  a.long_().sub(zsp, stackSize);          // Patch the code; don't forget to use LONG form.

  // Now the code is ready to be called
  Func fn;
  Error err = jit.add(&fn, &code);        // Add the generated code to the runtime.
  if (err) return 1;                      // Handle a possible error returned by AsmJit.

  int result = fn();                      // Execute the generated code.
  printf("%d\n", result);                 // Print the resulting "0".

  jit.release(fn);                        // Remove the function from the runtime.
  return 0;
}

If you run the example it would just work. As an experiment you can try removing long_() form to see what happens when wrong code is generated.

Code Patching and REX Prefix

In 64-bit mode there is one more thing to worry about when patching code - REX prefix. It's a single byte prefix designed to address registers with ids from 9 to 15 and to override the default width of operation from 32 to 64 bits. AsmJit, like other assemblers, only emits REX prefix when it's necessary. If the patched code only changes the immediate value as shown in the previous example then there is nothing to worry about as it doesn't change the logic behind emitting REX prefix, however, if the patched code changes register id or overrides the operation width then it's important to take care of REX prefix as well.

AsmJit contains another instruction option that controls (forces) REX prefix - rex(). If you use it the instruction emitted will always use REX prefix even when it's encodable without it. The following list contains some instructions and their binary representations to illustrate when it's emitted:

  • __83C410 - add esp, 16 - 32-bit operation in 64-bit mode doesn't require REX prefix.
  • 4083C410 - rex add esp, 16 - 32-bit operation in 64-bit mode with forced REX prefix (0x40).
  • 4883C410 - add rsp, 16 - 64-bit operation in 64-bit mode requires REX prefix (0x48).
  • 4183C410 - add r12d, 16 - 32-bit operation in 64-bit mode using R12D requires REX prefix (0x41).
  • 4983C410 - add r12, 16 - 64-bit operation in 64-bit mode using R12 requires REX prefix (0x49).

Generic Function API

So far all examples shown above handled creating function prologs and epilogs manually. While it's possible to do it that way it's much better to automate such process as function calling conventions vary across architectures and also across operating systems.

AsmJit contains a functionality that can be used to define function signatures and to calculate automatically optimal function frame that can be used directly by a prolog and epilog inserter. This feature was exclusive to AsmJit's Compiler for a very long time, but was abstracted out and is now available for all users regardless of BaseEmitter they use. The design of handling functions prologs and epilogs allows generally two use cases:

  • Calculate function frame before the function is generated - this is the only way if you use pure Assembler emitter and shown in the next example.
  • Calculate function frame after the function is generated - this way is generally used by Builder and Compiler emitters(will be described together with x86::Compiler).

The following concepts are used to describe and create functions in AsmJit:

  • Type - Type is an 8-bit value that describes a platform independent type as we know from C/C++. It provides abstractions for most common types like int8_t, uint32_t, uintptr_t, float, double, and all possible vector types to match ISAs up to AVX512. Type::Id was introduced originally to be used with the Compiler infrastucture, but is now used by FuncSignature as well.

  • CallConv - Describes a calling convention - this class contains instructions to assign registers and stack addresses to function arguments and return value(s), but doesn't specify any function signature. Calling conventions are architecture and OS dependent.

  • FuncSignature - Describes a function signature, for example int func(int, int). FuncSignature contains a function calling convention id, return value type, and function arguments. The signature itself is platform independent and uses Type::Id to describe types of function arguments and its return value(s).

  • FuncDetail - Architecture and ABI dependent information that describes CallConv and expanded FuncSignature. Each function argument and return value is represented as FuncValue that contains the original Type::Id enriched by additional information that specifies if the value is passed/returned by register (and which register) or by stack. Each value also contains some other metadata that provide additional information required to handle it properly (for example if a vector value is passed indirectly by a pointer as required by WIN64 calling convention, etc...).

  • FuncFrame - Contains information about the function frame that can be used by prolog/epilog inserter (PEI). Holds call stack size size and alignment, local stack size and alignment, and various attributes that describe how prolog and epilog should be constructed. FuncFrame doesn't know anything about function's arguments or return values, it hold only information necessary to create a valid and ABI conforming function prologs and epilogs.

  • FuncArgsAssignment - A helper class that can be used to reassign function arguments into user specified registers. It's architecture and ABI dependent mapping from function arguments described by CallConv and FuncDetail into registers specified by the user.

It's a lot of concepts where each represents one step in the function frame calculation. In addition, the whole machinery can also be used to create function calls, instead of function prologs and epilogs. The next example shows how AsmJit can be used to create functions for both 32-bit and 64-bit targets and various calling conventions:

#include <asmjit/asmjit.h>
#include <stdio.h>

using namespace asmjit;

typedef void (*SumIntsFunc)(int* dst, const int* a, const int* b);

int main(int argc, char* argv[]) {
  JitRuntime jit;                         // Create JIT Runtime.
  CodeHolder code;                        // Create a CodeHolder.

  code.init(jit.codeInfo());              // Initialize it to match `jit`.
  x86::Assembler a(&code);                // Create and attach x86::Assembler to `code`.

  // Decide which registers will be mapped to function arguments. Try changing
  // registers of `dst`, `src_a`, and `src_b` and see what happens in function's
  // prolog and epilog.
  x86::Gp dst   = a.zax();
  x86::Gp src_a = a.zcx();
  x86::Gp src_b = a.zdx();

  X86::Xmm vec0 = x86::xmm0;
  X86::Xmm vec1 = x86::xmm1;

  // Create and initialize `FuncDetail` and `FuncFrame`.
  FuncDetail func;
  func.init(FuncSignatureT<void, int*, const int*, const int*>(CallConv::kIdHost));

  FuncFrame frame;
  frame.init(func);

  // Make XMM0 and XMM1 dirty; `kGroupVec` describes XMM|YMM|ZMM registers.
  frame.setDirtyRegs(x86::Reg::kGroupVec, IntUtils::mask(0, 1));

  // Alternatively, if you don't want to use register masks you can pass `BaseReg`
  // to `addDirtyRegs()`. The following code would add both `xmm0` and `xmm1`.
  frame.addDirtyRegs(x86::xmm0, x86::xmm1);

  FuncArgsAssignment args(&func);         // Create arguments assignment context.
  args.assignAll(dst, src_a, src_b);      // Assign our registers to arguments.
  args.updateFrameInfo(frame);            // Reflect our args in FuncFrame.
  frame.finalize();                       // Finalize the FuncFrame (updates it).

  a.emitProlog(frame);                    // Emit function prolog.
  a.emitArgsAssignment(frame, args);      // Assign arguments to registers.
  a.movdqu(vec0, x86::ptr(src_a));        // Load 4 ints from [src_a] to XMM0.
  a.movdqu(vec1, x86::ptr(src_b));        // Load 4 ints from [src_b] to XMM1.
  a.paddd(vec0, vec1);                    // Add 4 ints in XMM1 to XMM0.
  a.movdqu(x86::ptr(dst), vec0);          // Store the result to [dst].
  a.emitEpilog(frame);                    // Emit function epilog and return.

  SumIntsFunc fn;
  Error err = jit.add(&fn, &code);        // Add the generated code to the runtime.
  if (err) return 1;                      // Handle a possible error case.

  // Execute the generated function.
  int inA[4] = { 4, 3, 2, 1 };
  int inB[4] = { 1, 5, 2, 8 };
  int out[4];
  fn(out, inA, inB);

  // Prints {5 8 4 9}
  printf("{%d %d %d %d}\n", out[0], out[1], out[2], out[3]);

  jit.release(fn);                        // Remove the function from the runtime.
  return 0;
}

Builder Interface

Both Builder and Compiler are emitters that emit everything to a representation that allows further processing. The code stored in such representation is completely safe to be patched, simplified, reordered, obfuscated, removed, injected, analyzed, and 'think-of-anything-else'. Each instruction, label, directive, etc... is stored in BaseNode (or derived class like InstNode or LabelNode) and contains all the information required to pass it later to the Assembler.

There is a huge difference between Builder and Compiler:

  • Builder (low-level):

    • Maximum compatibility with Assembler, easy to switch from Assembler to Builder and vice versa.
    • Doesn't generate machine code directly, allows to serialize to Assembler when the whole code is ready to be encoded.
  • Compiler (high-level):

    • Virtual registers - allows to use unlimited number of virtual registers which are allocated into physical registers by a built-in register allocator.
    • Function nodes - allows to create functions by specifying their signatures and assigning virtual registers to function arguments and return value(s).
    • Function calls - allows to call other functions within the generated code by using the same interface that is used to create functions.

There are multiple node types used by both Builder and Compiler:

  • Basic nodes:

    • BaseNode - Base class for all nodes.
    • InstNode - Instruction node.
    • AlignNode - Alignment directive (.align).
    • LabelNode - Label (location where to bound it).
  • Data nodes:

    • DataNode - Data embedded into the code.
    • ConstPoolNode - Constant pool data.
    • LabelDataNode - Label address embedded as data.
  • Informative nodes:

    • CommentNode - Contains a comment string, doesn't affect code generation.
    • SentinelNode - A marker that can be used to remember certain position, doesn't affect code generation.
  • Compiler-only nodes:

    • FuncNode - Start of a function.
    • FuncRetNode - Return from a function.
    • FuncCallNode - Function call.

Using Builder

The Builder interface was designed to be used as an Assembler replacement in case that post-processing of the generated code is required. The code can be modified during or after code generation. The post processing can be done manually or through Pass (Code-Builder Pass) object. Builder stores the emitted code as a double-linked list, which allows O(1) insertion and removal.

The code representation used by Builder is compatible with everything AsmJit provides. Each instruction is stored as InstNode, which contains instruction id, options, and operands. Each instruction emitted will create a new InstNode instance and add it to the current cursor in the double-linked list of nodes. Since the instruction stream used by Builder can be manipulated, we can rewrite the SumInts example into the following:

#include <asmjit/asmjit.h>
#include <stdio.h>

using namespace asmjit;

typedef void (*SumIntsFunc)(int* dst, const int* a, const int* b);

// Small helper function to print the current content of `cb`.
static void dumpCode(BaseBuilder& cb, const char* phase) {
  StringBuilder sb;
  cb.dump(sb);
  printf("%s:\n%s\n", phase, sb.data());
}

int main(int argc, char* argv[]) {
  JitRuntime jit;                         // Create JIT Runtime.
  CodeHolder code;                        // Create a CodeHolder.

  code.init(jit.codeInfo());              // Initialize it to match `jit`.
  x86::Builder cb(&code);                 // Create and attach x86::Builder to `code`.

  // Decide which registers will be mapped to function arguments. Try changing
  // registers of `dst`, `src_a`, and `src_b` and see what happens in function's
  // prolog and epilog.
  x86::Gp dst   = cb.zax();
  x86::Gp src_a = cb.zcx();
  x86::Gp src_b = cb.zdx();

  X86::Xmm vec0 = x86::xmm0;
  X86::Xmm vec1 = x86::xmm1;

  // Create and initialize `FuncDetail`.
  FuncDetail func;
  func.init(FuncSignatureT<void, int*, const int*, const int*>(CallConv::kIdHost));

  // Remember prolog insertion point.
  BaseNode* prologInsertionPoint = cb.cursor();

  // Emit function body:
  cb.movdqu(vec0, x86::ptr(src_a));       // Load 4 ints from [src_a] to XMM0.
  cb.movdqu(vec1, x86::ptr(src_b));       // Load 4 ints from [src_b] to XMM1.
  cb.paddd(vec0, vec1);                   // Add 4 ints in XMM1 to XMM0.
  cb.movdqu(x86::ptr(dst), vec0);         // Store the result to [dst].

  // Remember epilog insertion point.
  BaseNode* epilogInsertionPoint = cb.cursor();

  // Let's see what we have now.
  dumpCode(cb, "Raw Function");

  // Now, after we emitted the function body, we can insert the prolog, arguments
  // allocation, and epilog. This is not possible with using pure x86::Assembler.
  FuncFrame frame;
  frame.init(func);

  // Make XMM0 and XMM1 dirty; `kGroupVec` describes XMM|YMM|ZMM registers.
  frame.setDirtyRegs(x86::Reg::kGroupVec, IntUtils::mask(0, 1));

  FuncArgsAssignment args(&func);         // Create arguments assignment context.
  args.assignAll(dst, src_a, src_b);      // Assign our registers to arguments.
  args.updateFrame(frame);                // Reflect our args in FuncFrame.
  frame.finalize();                       // Finalize the FuncFrame (updates it).

  // Insert function prolog and allocate arguments to registers.
  cb.setCursor(prologInsertionPoint);
  cb.emitProlog(frame);
  cb.emitArgsAssignment(frame, args);

  // Insert function epilog.
  cb.setCursor(epilogInsertionPoint);
  cb.emitEpilog(frame);

  // Let's see how the function's prolog and epilog looks.
  dumpCode(cb, "Prolog & Epilog");

  // IMPORTANT: Builder requires `finalize()` to be called to serialize the code
  // to the Assembler (it automatically creates one if not attached).
  cb.finalize();

  SumIntsFunc fn;
  Error err = jit.add(&fn, &code);        // Add the generated code to the runtime.
  if (err) return 1;                      // Handle a possible error case.

  // Execute the generated function.
  int inA[4] = { 4, 3, 2, 1 };
  int inB[4] = { 1, 5, 2, 8 };
  int out[4];
  fn(out, inA, inB);

  // Prints {5 8 4 9}
  printf("{%d %d %d %d}\n", out[0], out[1], out[2], out[3]);

  jit.release(fn);                        // Remove the function from the runtime.
  return 0;
}

When the example is executed it should output the following (this one using AMD64-SystemV ABI):

Raw Function:
movdqu xmm0, [rcx]
movdqu xmm1, [rdx]
paddd xmm0, xmm1
movdqu [rax], xmm0

Prolog & Epilog:
mov rax, rdi
mov rcx, rsi
movdqu xmm0, [rcx]
movdqu xmm1, [rdx]
paddd xmm0, xmm1
movdqu [rax], xmm0
ret

{5 8 4 9}

The number of use-cases of x86::Builder is not limited and highly depends on your creativity and experience. The previous example can be easily improved to collect all dirty registers inside the function programmatically and to pass them to frame.setDirtyRegs():

#include <asmjit/asmjit.h>

using namespace asmjit;

// NOTE: This function doesn't cover all possible constructs. It ignores
// instructions that write to implicit registers that are not part of the
// operand list. It also counts read-only registers. Real implementation
// would be a bit more complicated, but still relatively easy to implement.
static void collectDirtyRegs(const BaseNode* first, const BaseNode* last, uint32_t regMask[BaseReg::kGroupVirt]) {
  const BaseNode* node = first;
  while (node) {
    if (node->actsAsInst()) {
      const InstNode* inst = node->as<InstNode>();
      const Operand* opArray = inst->operands();

      for (uint32_t i = 0, opCount = inst->opCount(); i < opCount; i++) {
        const Operand& op = opArray[i];
        if (op.isReg()) {
          const x86::Reg& reg = op.as<x86::Reg>();
          if (reg.group() < BaseReg::kGroupVirt)
            regMask[reg.group()] |= 1u << reg.id();
        }
      }
    }

    if (node == last) break;
    node = node->next();
  n}

static void setDirtyRegsOfFuncFrame(const x86::Builder& cb, FuncFrame& frame) {
  uint32_t regMask[BaseReg::kGroupVirt] = { 0 };
  collectDirtyRegs(cb.firstNode(), cb.lastNode(), regMask);

  // X86/X64 ABIs only require to save GP/XMM registers:
  frame.setDirtyRegs(x86::Reg::kGroupGp , regMask[x86::Reg::kGroupGp ]);
  frame.setDirtyRegs(x86::Reg::kGroupVec, regMask[x86::Reg::kGroupVec]);
}

Using x86::Assembler or x86::Builder through X86::Emitter

Even when Assembler and Builder provide the same interface as defined by BaseEmitter their platform dependent variants (x86::Assembler and x86::Builder, respective) cannot be interchanged or casted to each other by using C++'s static_cast<>. The main reason is the inheritance graph of these classes is different and cast-incompatible, as illustrated in the following graph:

                                            +--------------+      +=========================+
                   +----------------------->| x86::Emitter |<--+--# x86::EmitterImplicitT<> #<--+
                   |                        +--------------+   |  +=========================+   |
                   |                           (abstract)      |           (mixin)              |
                   |   +--------------+     +~~~~~~~~~~~~~~+   |                                |
                   +-->| BaseAssembler|---->|x86::Assembler|<--+                                |
                   |   +--------------+     +~~~~~~~~~~~~~~+   |                                |
                   |      (abstract)            (final)        |                                |
+===============+  |   +--------------+     +~~~~~~~~~~~~~~+   |                                |
#  BaseEmitter  #--+-->|  BaseBuilder |--+->| x86::Builder |<--+                                |
+===============+      +--------------+  |  +~~~~~~~~~~~~~~+                                    |
   (abstract)             (abstract)     |      (final)                                         |
                   +---------------------+                                                      |
                   |                                                                            |
                   |   +--------------+     +~~~~~~~~~~~~~~+      +=========================+   |
                   +-->| BaseCompiler |---->| x86::Compiler|<-----# x86::EmitterExplicitT<> #---+
                       +--------------+     +~~~~~~~~~~~~~~+      +=========================+
                          (abstract)            (final)                   (mixin)

The graph basically shows that it's not possible to cast x86::Assembler to x86::Builder and vice versa. However, since both x86::Assembler and x86::Builder share the same interface defined by both BaseEmitter and x86::EmmiterImplicitT a class called x86::Emitter was introduced to make it possible to write a function that can emit to both x86::Assembler and x86::Builder. Note that x86::Emitter cannot be created, it's abstract and has private constructors and destructors; it was only designed to be casted to and used as an interface.

Each X86 emitter implements a member function called as<x86::Emitter>(), which casts the instance to the x86::Emitter, as illustrated on the next example:

#include <asmjit/asmjit.h>

using namespace asmjit;

static void emitSomething(x86::Emitter* e) {
  e->mov(x86::eax, x86::ebx);
}

static void assemble(CodeHolder& code, bool useAsm) {
  if (useAsm) {
    x86::Assembler a(&code);
    emitSomething(a.as<x86::Emitter>());
  }
  else {
    x86::Builder cb(&code);
    emitSomething(cb.as<x86::Emitter>());

    // IMPORTANT: Builder requires `finalize()` to be called to serialize the
    // code to the Assembler (it automatically creates one if not attached).
    cb.finalize();
  }
}

The example above shows how to create a function that can emit code to either x86::Assembler or x86::Builder through x86::Emitter, which provides emitter-neutral functionality. x86::Emitter, however, doesn't provide any emitter x86::Assembler or x86::Builder specific functionality like setCursor().

Compiler Interface

Compiler is a high-level code emitter that provides virtual registers and automatically handles function calling conventions. It's still architecture dependent, but makes the code generation much easier by offering a built-in register allocator and function builder. Functions are essential; the first-step to generate some code is to define the signature of the function you want to generate (before generating the function body). Function arguments and return value(s) are handled by assigning virtual registers to them. Similarly, function calls are handled the same way.

Compiler also makes the use of passes (introduced by Builder) and automatically adds an architecture-dependent register allocator pass to the list of passes when attached to CodeHolder.

Compiler Basics

The first Compiler example shows how to generate a function that simply returns an integer value. It's an analogy to the very first example:

#include <asmjit/asmjit.h>
#include <stdio.h>

using namespace asmjit;

// Signature of the generated function.
typedef int (*Func)(void);

int main(int argc, char* argv[]) {
  JitRuntime jit;                         // Runtime specialized for JIT code execution.
  CodeHolder code;                        // Holds code and relocation information.

  code.init(jit.codeInfo());              // Initialize to the same arch as JIT runtime.
  x86::Compiler cc(&code);                // Create and attach x86::Compiler to `code`.

  cc.addFunc(FuncSignatureT<int>());      // Begin a function of `int fn(void)` signature.

  x86::Gp vReg = cc.newGpd();             // Create a 32-bit general purpose register.
  cc.mov(vReg, 1);                        // Move one to our virtual register `vReg`.
  cc.ret(vReg);                           // Return `vReg` from the function.

  cc.endFunc();                           // End of the function body.
  cc.finalize();                          // Translate and assemble the whole `cc` content.
  // ----> x86::Compiler is no longer needed from here and can be destroyed <----

  Func fn;
  Error err = jit.add(&fn, &code);        // Add the generated code to the runtime.
  if (err) return 1;                      // Handle a possible error returned by AsmJit.
  // ----> CodeHolder is no longer needed from here and can be destroyed <----

  int result = fn();                      // Execute the generated code.
  printf("%d\n", result);                 // Print the resulting "1".

  jit.release(fn);                        // RAII, but let's make it explicit.
  return 0;
}

The addFunc() and endFunc() methods define the body of the function. Both functions must be called per function, but the body doesn't have to be generated in sequence. An example of generating two functions will be shown later. The next example shows more complicated code that contain a loop and generates a memcpy32() function:

#include <asmjit/asmjit.h>
#include <stdio.h>

using namespace asmjit;

// Signature of the generated function.
typedef void (*MemCpy32)(uint32_t* dst, const uint32_t* src, size_t count);

int main(int argc, char* argv[]) {
  JitRuntime jit;                         // Runtime specialized for JIT code execution.
  CodeHolder code;                        // Holds code and relocation information.

  code.init(jit.codeInfo());              // Initialize to the same arch as JIT runtime.
  x86::Compiler cc(&code);                // Create and attach x86::Compiler to `code`.

  cc.addFunc(                             // Begin the function of the following signature:
    FuncSignatureT<void,                  //   Return value - void      (no return value).
      uint32_t*,                          //   1st argument - uint32_t* (machine reg-size).
      const uint32_t*,                    //   2nd argument - uint32_t* (machine reg-size).
      size_t>());                         //   3rd argument - size_t    (machine reg-size).

  Label L_Loop = cc.newLabel();           // Start of the loop.
  Label L_Exit = cc.newLabel();           // Used to exit early.

  x86::Gp dst = cc.newIntPtr("dst");      // Create `dst` register (destination pointer).
  x86::Gp src = cc.newIntPtr("src");      // Create `src` register (source pointer).
  x86::Gp cnt = cc.newUIntPtr("cnt");     // Create `cnt` register (loop counter).

  cc.setArg(0, dst);                      // Assign `dst` argument.
  cc.setArg(1, src);                      // Assign `src` argument.
  cc.setArg(2, cnt);                      // Assign `cnt` argument.

  cc.test(cnt, cnt);                      // Early exit if length is zero.
  cc.jz(L_Exit);

  cc.bind(L_Loop);                        // Bind the beginning of the loop here.

  x86::Gp tmp = cc.newInt32("tmp");       // Copy a single dword (4 bytes).
  cc.mov(tmp, x86::dword_ptr(src));       // Load DWORD from [src] address.
  cc.mov(x86::dword_ptr(dst), tmp);       // Store DWORD to [dst] address.

  cc.add(src, 4);                         // Increment `src`.
  cc.add(dst, 4);                         // Increment `dst`.

  cc.dec(cnt);                            // Loop until `cnt` is non-zero.
  cc.jnz(L_Loop);

  cc.bind(L_Exit);                        // Label used by early exit.
  cc.endFunc();                           // End of the function body.

  cc.finalize();                          // Translate and assemble the whole `cc` content.
  // ----> x86::Compiler is no longer needed from here and can be destroyed <----

  MemCpy32 memcpy32;
  Error err = jit.add(&memcpy32, &code);  // Add the generated code to the runtime.
  if (err) return 1;                      // Handle a possible error returned by AsmJit.
  // ----> CodeHolder is no longer needed from here and can be destroyed <----

  // Test the generated code.
  uint32_t input[6] = { 1, 2, 3, 5, 8, 13 };
  uint32_t output[6];
  memcpy32(output, input, 6);

  for (uint32_t i = 0; i < 6; i++)
    printf("%d\n", output[i]);

  jit.release(memcpy32);                  // RAII, but let's make it explicit.
  return 0;
}

Recursive Functions

It's possible to create more functions by using the same x86::Compiler instance and make links between them. In such case it's important to keep the pointer to the FuncNode node. The first example creates a simple Fibonacci function that calls itself recursively:

#include <asmjit/asmjit.h>
#include <stdio.h>

using namespace asmjit;

// Signature of the generated function.
typedef uint32_t (*Fibonacci)(uint32_t x);

int main(int argc, char* argv[]) {
  JitRuntime jit;                         // Runtime specialized for JIT code execution.
  CodeHolder code;                        // Holds code and relocation information.

  code.init(jit.codeInfo());              // Initialize to the same arch as JIT runtime.
  x86::Compiler cc(&code);                // Create and attach x86::Compiler to `code`.

  FuncNode* func = cc.addFunc(            // Begin of the Fibonacci function, `addFunc()`
    FuncSignatureT<int, int>());          // Returns a pointer to the `FuncNode` node.

  Label L_Exit = cc.newLabel()            // Exit label.
  x86::Gp x = cc.newU32();                // Function `x` argument.
  x86::Gp y = cc.newU32();                // Temporary.

  cc.setArg(0, x);

  cc.cmp(x, 3);                           // Return `x` if less than 3.
  cc.jb(L_Exit);

  cc.mov(y, x);                           // Make copy of the original `x`.
  cc.dec(x);                              // Decrease `x`.

  FuncCallNode* call = cc.call(           // Function call:
    func->label(),                        //   Function address or Label.
    FuncSignatureT<int, int>());          //   Function signature.

  call->setArg(0, x);                     // Assign `x` as the first argument and
  call->setRet(0, x);                     // assign `x` as a return value as well.

  cc.add(x, y);                           // Combine the return value with `y`.

  cc.bind(L_Exit);
  cc.ret(x);                              // Return `x`.
  cc.endFunc();                           // End of the function body.

  cc.finalize();                          // Translate and assemble the whole `cc` content.
  // ----> x86::Compiler is no longer needed from here and can be destroyed <----

  Fibonacci fib;
  Error err = jit.add(&fib, &code);       // Add the generated code to the runtime.
  if (err) return 1;                      // Handle a possible error returned by AsmJit.
  // ----> CodeHolder is no longer needed from here and can be destroyed <----

  printf("Fib(%u) -> %u\n", 8, fib(8));   // Test the generated code.

  jit.release(fib);                       // RAII, but let's make it explicit.
  return 0;
}

Stack Management

Compiler manages function's stack-frame, which is used by the register allocator to spill virtual registers. It also provides an interface to allocate user-defined block of the stack, which can be used as a temporary storage by the generated function. In the following example a stack of 256 bytes size is allocated, filled by bytes starting from 0 to 255 and then iterated again to sum all the values.

#include <asmjit/asmjit.h>
#include <stdio.h>

using namespace asmjit;

// Signature of the generated function.
typedef int (*Func)(void);

int main(int argc, char* argv[]) {
  JitRuntime jit;                         // Runtime specialized for JIT code execution.
  CodeHolder code;                        // Holds code and relocation information.

  code.init(jit.codeInfo());              // Initialize to the same arch as JIT runtime.
  x86::Compiler cc(&code);                // Create and attach x86::Compiler to `code`.

  cc.addFunc(FuncSignatureT<int>());      // Create a function that returns 'int'.

  x86::Gp p = cc.newIntPtr("p");
  x86::Gp i = cc.newIntPtr("i");

  x86::Mem stack = cc.newStack(256, 4);   // Allocate 256 bytes on the stack aligned to 4 bytes.
  x86::Mem stackIdx(stack);               // Copy of `stack` with `i` added.
  stackIdx.setIndex(i);                   // stackIdx <- stack[i].
  stackIdx.setSize(1);                    // stackIdx <- byte ptr stack[i].

  // Load a stack address to `p`. This step is purely optional and shows
  // that `lea` is useful to load a memory operands address (even absolute)
  // to a general purpose register.
  cc.lea(p, stack);

  // Clear `i` (`xor` as it's C++ keyword, hence `xor_` is used instead).
  cc.xor_(i, i);

  Label L1 = cc.newLabel();
  Label L2 = cc.newLabel();

  cc.bind(L1);                            // First loop, fill the stack.
  cc.mov(stackIdx, i.r8());               // stack[i] = uint8_t(i).

  cc.inc(i);                              // i++;
  cc.cmp(i, 256);                         // if (i < 256)
  cc.jb(L1);                              //   goto L1;

  // Second loop, sum all bytes stored in `stack`.
  x86::Gp sum = cc.newI32("sum");
  x86::Gp val = cc.newI32("val");

  cc.xor_(i, i);
  cc.xor_(sum, sum);

  cc.bind(L2);

  cc.movzx(val, stackIdx);                // val = uint32_t(stack[i]);
  cc.add(sum, val);                       // sum += val;

  cc.inc(i);                              // i++;
  cc.cmp(i, 256);                         // if (i < 256)
  cc.jb(L2);                              //   goto L2;

  cc.ret(sum);                            // Return the `sum` of all values.
  cc.endFunc();                           // End of the function body.

  cc.finalize();                          // Translate and assemble the whole `cc` content.
  // ----> x86::Compiler is no longer needed from here and can be destroyed <----

  Func func;
  Error err = jit.add(&func, &code);      // Add the generated code to the runtime.
  if (err) return 1;                      // Handle a possible error returned by AsmJit.
  // ----> CodeHolder is no longer needed from here and can be destroyed <----

  printf("Func() -> %d\n", func());       // Test the generated code.

  jit.release(func);                      // RAII, but let's make it explicit.
  return 0;
}

Constant Pool

Compiler provides two constant pools for a general purpose code generation - local and global. Local constant pool is related to a single FuncNode node and is generally flushed after the function body, and global constant pool is flushed at the end of the generated code by Compiler::finalize().

#include <asmjit/asmjit.h>

using namespace asmjit;

static void exampleUseOfConstPool(x86::Compiler& cc) {
  cc.addFunc(FuncSignatureT<int>());

  x86::Gp v0 = cc.newGpd("v0");
  x86::Gp v1 = cc.newGpd("v1");

  x86::Mem c0 = cc.newInt32Const(ConstPool::kScopeLocal, 200);
  x86::Mem c1 = cc.newInt32Const(ConstPool::kScopeLocal, 33);

  cc.mov(v0, c0);
  cc.mov(v1, c1);
  cc.add(v0, v1);

  cc.ret(v0);
  cc.endFunc();
}

Advanced Features

Logging

The initial phase of any project that generates machine code is not always smooth. Failure cases are common especially at the beginning of the project and AsmJit provides a logging functionality to address this issue. AsmJit does already a good job with function overloading to prevent from emitting semantically incorrect instructions, but it can't prevent from emitting machine code that is semantically correct, but doesn't work when it's executed. Logging has always been an important part of AsmJit's infrastructure and looking at logs can sometimes reveal code generation issues quickly.

AsmJit provides API for logging and formatting:

  • Logger - A logger that you can pass to CodeHolder and all emitters that inherit BaseEmitter.
  • FormatOptions - Formatting options that can change how instructions and operands are formatted.

AsmJit's Logger serves the following purposes:

  • Provides a basic foundation for logging.
  • Abstract class leaving the implementation (destination) on users. Two backends are built-in for simplicity:
    • FileLogger implements logging into a standard std::FILE stream.
    • StringLogger stores the logged text in StringBuilder instance.

AsmJit's FormatOptions provides the following to customize the formatting of instructions and operands:

  • Flags:
    • FormatOptions::kFlagMachineCode - Show a machine code of each encoded instruction.
    • FormatOptions::kFlagExplainConsts - Show a text explanation of some immediate values that are used as predicates.
    • FormatOptions::kFlagHexImms - Use hexadecimal notation to output immediates.
    • FormatOptions::kFlagHexOffsets - Use hexadecimal notation to output offsets.
    • FormatOptions::kFlagRegCasts - Show casts between various register types (compiler).
    • FormatOptions::kFlagPositions - Show positions associated with nodes (compiler).
  • Indentation:
    • FormatOptions::kIndentationCode - Indentation of instructions and directives.
    • FormatOptions::kIndentationLabel - Indentation of labels.
    • FormatOptions::kIndentationComment - Indentation of whole-line comments.

Logger is typically attached to CodeHolder and all attached code emitters automatically use it:

#include <asmjit/asmjit.h>
#include <stdio.h>

using namespace asmjit;

int main(int argc, char* argv[]) {
  JitRuntime jit;                         // Runtime specialized for JIT code execution.
  FileLogger logger(stdout);              // Logger should always survive the CodeHolder.

  CodeHolder code;                        // Holds code and relocation information.
  code.init(jit.codeInfo());              // Initialize to the same arch as JIT runtime.
  code.setLogger(&logger);                // Attach the `logger` to `code` holder.

  // ... code as usual, everything you emit will be logged to `stdout` ...

  return 0;
}

Error Handling

AsmJit uses error codes to represent and return errors. Every function where error can occur returns Error. Exceptions are never thrown by AsmJit even in extreme conditions like out-of-memory. Errors should never be ignored, however, checking errors after each asmjit API call would simply overcomplicate the whole code generation experience. To make life simpler AsmJit provides ErrorHandler, which provides handleError() function:

`virtual bool handleError(Error err, const char* message, BaseEmitter* origin) = 0;`

That can be overridden by AsmJit users and do the following:

    1. Record the error and continue (the way how the error is user-implemented).
    1. Throw an exception. AsmJit doesn't use exceptions and is completely exception-safe, but it's perfectly legal to throw an exception from the error handler.
    1. Use plain old C's setjmp() and longjmp(). Asmjit always puts Assembler and Compiler to a consistent state before calling the handleError() so longjmp() can be used without issues to cancel the code-generation if an error occurred. This method can be used if exception handling in your project is turned off and you still want some comfort. In most cases it should be safe as AsmJit uses Zone memory and the ownership of memory it allocates always ends with the instance that allocated it. If using this approach please never jump outside the life-time of CodeHolder and BaseEmitter.

ErrorHandler can be attached to CodeHolder and/or BaseEmitter (which has a priority). The first example uses error handler that just prints the error, but lets AsmJit continue:

// Error handling #1:
#include <asmjit/asmjit.h>

#include <stdio.h>

// Error handler that just prints the error and lets AsmJit ignore it.
class SimpleErrorHandler : public asmjit::ErrorHandler {
public:
  inline SimpleErrorHandler() : lastError(kErrorOk) {}

  void handleError(asmjit::Error err, const char* message, asmjit::BaseEmitter* origin) override {
    this->err = err;
    fprintf(stderr, "ERROR: %s\n", message);
  }

  Error err;
};

int main(int argc, char* argv[]) {
  using namespace asmjit;

  JitRuntime jit;
  SimpleErrorHandler eh;

  CodeHolder code;
  code.init(jit.codeInfo());
  code.setErrorHandler(&eh);

  // Try to emit instruction that doesn't exist.
  x86::Assembler a(&code);
  a.emit(x86::Inst::kIdMov, x86::xmm0, x86::xmm1);

  if (eh.err) {
    // Assembler failed!
  }

  return 0;
}

If error happens during instruction emitting / encoding the assembler behaves transactionally - the output buffer won't advance if encoding failed, thus either a fully encoded instruction or nothing is emitted. The error handling shown above is useful, but it's still not the best way of dealing with errors in AsmJit. The following example shows how to use exception handling to handle errors in a more C++ way:

// Error handling #2:
#include <asmjit/asmjit.h>

#include <exception>
#include <string>
#include <stdio.h>

// Error handler that throws a user-defined `AsmJitException`.
class AsmJitException : public std::exception {
public:
  AsmJitException(asmjit::Error err, const char* message) noexcept
    : err(err),
      message(message) {}

  const char* what() const noexcept override { return message.c_str(); }

  asmjit::Error err;
  std::string message;
};

class ThrowableErrorHandler : public asmjit::ErrorHandler {
public:
  // Throw is possible, functions that use ErrorHandler are never 'noexcept'.
  void handleError(asmjit::Error err, const char* message, asmjit::BaseEmitter* origin) override {
    throw AsmJitException(err, message);
  }
};

int main(int argc, char* argv[]) {
  using namespace asmjit;

  JitRuntime jit;
  ThrowableErrorHandler eh;

  CodeHolder code;
  code.init(jit.codeInfo());
  code.setErrorHandler(&eh);

  x86::Assembler a(&code);

  // Try to emit instruction that doesn't exist.
  try {
    a.emit(x86::Inst::kIdMov, x86::xmm0, x86::xmm1);
  }
  catch (const AsmJitException& ex) {
    printf("EXCEPTION THROWN: %s\n", ex.what());
  }

  return 0;
}

If C++ exceptions are not what you like or your project turns off them completely there is still a way of reducing the error handling to a minimum by using a standard setjmp/longjmp approach. AsmJit is exception-safe and cleans up everything before calling the ErrorHandler, so any approach is safe. You can simply jump from the error handler without causing any side-effects or memory leaks. The following example demonstrates how it could be done:

// Error handling #3:
#include <asmjit/asmjit.h>

#include <setjmp.h>
#include <stdio.h>

class LongJmpErrorHandler : public asmjit::ErrorHandler {
public:
  inline LongJmpErrorHandler() : err(asmjit::kErrorOk) {}

  void handleError(asmjit::Error err, const char* message, asmjit::BaseEmitter* origin) override {
    this->err = err;
    longjmp(state, 1);
  }

  jmp_buf state;
  asmjit::Error err;
};

int main(int argc, char* argv[]) {
  using namespace asmjit;

  JitRuntime jit;
  LongJmpErrorHandler eh;

  CodeHolder code;
  code.init(jit.codeInfo());
  code.setErrorHandler(&eh);

  x86::Assembler a(&code);

  if (!setjmp(eh.state)) {
    // Try to emit instruction that doesn't exist.
    a.emit(x86::Inst::kIdMov, x86::xmm0, x86::xmm1);
  }
  else {
    Error err = eh.err;
    printf("ASMJIT ERROR: 0x%08X [%s]\n", err, DebugUtils::errorAsString(err));
  }

  return 0;
}

Code Injection

Both Builder and Compiler emitters store their nodes in a double-linked list, which makes it easy to manipulate that list during the code generation or after. Each node is always emitted next to the current cursor and the cursor is changed to that newly emitted node. The cursor can be explicitly retrieved and changed by cursor() and setCursor(), respectively.

The following example shows how to inject code at the beginning of the function by implementing an XmmConstInjector helper class.

TODO

...More documentation...

Other Topics

This section provides quick answers to some recurring questions and topics.

Instruction Validation

AsmJit by default prefers performance when it comes to instruction encoding. The Assembler implementation would only validate operands that must be validated to select a proper encoding of the instruction. This means that by default it would accept instructions that do not really exist like mov rax, ebx. This is great in release mode as it makes the assembler faster, however, it's not that great for development as it allows to silently pass even when the instruction's operands are incorrect. To fix this Asmjit contains a feature called Strict Validation, which allows to validate each instruction before the Assembler tries to encode it. This feature can also be used without an Assembler instance through BaseInst::validate() API.

Emitter options are configured through CodeHolder:

CodeHolder code;

// Enables strict instruction validation for all emitters attached to `code`.
code.addEmitterOptions(BaseEmitter::kOptionStrictValidation);

// Use either ErrorHandler attached to CodeHolder or Error code returned by
// the Assembler.
x86::Assembler a(&code);
Error err = a.emit(x86::Inst::kIdMov, x86::eax, x86::al);
if (err) { /* failed */ }

Label Offsets and Links

When you use a label that is not yet bound the Assembler would create a LabelLink, which is then added to CodeHolder's LabelEntry. These links are also created for labels that are bound but reference some location in a different section. Firstly, here are some functions that can be used to check some basics:

CodeHolder code = ...;
Label L = ...;

// Returns whether the Label `L` is bound.
bool bound = code.isLabelBound(L or L.id());

// Returns true if the code contains either referenced, but unbound labels,
// or cross-section label links that are not resolved yet.
bool value = code.hasUnresolvedLinks();     // Boolean answer.
size_t count = code.unresolvedLinkCount();  // Count of links.

Please note that there is not API to return a count of unbound labels as this is completely unimportant from CodeHolder's perspective. If a label is not used then it doesn't matter whether it's bound or not, only used labels matter. After a Label is bound you can query it's offset relative to the start of the section where it was bound:

CodeHolder code = ...;
Label L = ...;

// After you are done you can check the offset. The offset provided
// is relative to the start of the section, see below for alternative.
// If the given label is not bound then the offset returned will be zero.
uint64_t offset = code.labelOffset(L or L.id());

// If you use multiple sections and want the offset relative to the base.
// NOTE: This function expects that the section has already an offset and
// the label-link was resolved (if this is not true you will still get an
// offset relative to the start of the section).
uint64_t offset = code.labelOffsetFromBase(L or L.id());

Sections

Sections is a relatively new feature that allows to create multiple sections. It's supported by Assembler, Builder, and Compiler. Please note that using multiple sections is advanced and requires more understanding about how AsmJit works. There is a test-case asmjit_test_x86_sections.cpp that shows how sections can be used.

CodeHolder code = ...;

// Text section is always provided as the first section.
Section* text = code.textSection(); // or code.sectionById(0);

// To create another section use `code.newSection()`.
Section* data;
Error err = code.newSection(&data,
  ".data",  // Section name
  SIZE_MAX, // Name length if the name is not null terminated (or SIZE_MAX).
  0,        // Section flags, see Section::Flags.
  8);       // Section alignment, must be power of 2.

// When you switch sections in Assembler, Builder, or Compiler the cursor
// will always move to the end of that section. When you create an Assembler
// the cursor would be placed at the end of the first (.text) section, which
// is initially empty.
x86::Assembler a(&code);
Label L_Data = a.newLabel();

a.mov(x86::eax, x86::ebx); // Emits in .text section.

a.section(data);           // Switches to the end of .data section.
a.bind(L_Data);            // Binds label in this .data section
a.db(0x01);                // Emits byte in .data section.

a.section(text);           // Switches to the end of .text section.
a.add(x86::ebx, x86::eax); // Emits in .text section.

// References a label bound in .data section in .text section. This
// would create a LabelLink even when the L_Data is already bound,
// because the reference crosses sections. See below...
a.lea(x86::rsi, x86::ptr(L_Data));

The last line in the example above shows that a LabelLink would be created even for bound labels that cross sections. In this case a referenced label was bound in another section, which means that the link couldn't be resolved at that moment. If your code uses sections, but you wish AsmJit to flatten these sections (you don't plan to flatten them manually) then there is an API for that.

// ... (continuing the previous example) ...
CodeHolder code = ...;

// Suppose we have some code that contains multiple sections and
// we would like to flatten them by using AsmJit's built-in API:
Error err = code.flatten();
if (err) { /* Error handling is necessary. */ }

// After flattening all sections would contain assigned offsets
// relative to base. Offsets are 64-bit unsigned integers so we
// cast them to `size_t` for simplicity. On 32-bit targets it's
// guaranteed that the offset cannot be greater than `2^32 - 1`.
printf("Data section offset %zu", size_t(data->offset()));

// The flattening doesn't resolve unresolved label links, this
// has to be done manually as flattening can be done separately.
err = code.resolveUnresolvedLinks();
if (err) { /* Error handling is necessary. */ }

if (code.hasUnresolvedLinks()) {
  // This would mean either unbound label or some other issue.
  printf("FAILED: UnresoledLinkCount=%zu\n", code.unresovedLinkCount());
}

Using AsmJit Data Structures

AsmJit stores its data in data structures allocated by ZoneAllocator. It's a fast allocator that allows AsmJit to allocate a lot of small data structures fast and without malloc() overhead. The most common data structure that you will probably inspect is ZoneVector<T>. It's like C++'s std::vector. but the implementation doesn't use exceptions and uses the mentioned ZoneAllocator for performance reasons. You don't have to worry about allocations as you should not need to add items to data structures that are managed by CodeHolder or advanced emitters like Builder/Compiler.

APIs that return ZoneVector:

CodeHolder code = ...;

// Contains all emitters attached to CodeHolder.
const ZoneVector<BaseEmitter*>& emitters = code.emitters();

// Contains all sections managed by CodeHolder.
const ZoneVector<Section*>& sections = code.sections();

// Contains all LabelEntry records associated with created Labels.
const ZoneVector<LabelEntry*>& labelEntries = code.labelEntries();

// Contains all RelocEntry records that describe relocations.
const ZoneVector<RelocEntry*>& relocEntries = code.relocEntries();

AsmJit's ZoneVector<T> has overloaded array access operator to make it possible accessing its elements through operator[]. Some standard functions like empty(), size(), and data() are provided as well. Vectors are also iterable through range-based for loop:

CodeHolder code = ...;

for (LabelEntry* le : code.labelEntries()) {
  printf("Label #%u {Bound=%s Offset=%llu}",
    le->id(),
    le->isBound() ? "true" : "false",
    (unsigned long long)le->offset());
}

Support

AsmJit is an open-source library released under a permissive ZLIB license, which makes it possible to use it freely in any open-source or commercial product. Free support is available through issues and gitter channel, which is very active. Commercial support is currently individual and can be negotiated on demand. It includes consultation, priority bug fixing, review of code that uses AsmJit, porting code to the latest AsmJit, and implementation of new AsmJit features.

If you use AsmJit in a non-commercial project and would like to appreciate the library in the form of a donation you are welcome to support us. Donations are anonymous unless the donor lets us know otherwise. The order and format of listed donors is not guaranteed and may change in the future. Additionally, donations should be considered as an appreciation of past work and not used to gain special privileges in terms of future development. AsmJit authors reserve the right to remove a donor from the list in extreme cases of disruptive behavior against other community members. Diversity of opinions and constructive criticism will always be welcome in the AsmJit community.

Donation Addresses:

  • BTC: 14dEp5h8jYSxgXB9vcjE8eh78uweD76o7W
  • ETH: 0xd4f0b9424cF31DF5a5359D029CF3A65c500a581E
  • Please contact us if you would like to donate through a different channel or to use a different crypto-currency. Wire transfers and SEPA payments are both possible.

Donors:

Authors & Maintainers

You can’t perform that action at this time.