const std = @import("std"); const Allocator = std.mem.Allocator; const assert = std.debug.assert; const Compilation = @import("../Compilation.zig"); const write = Compilation.write; // const log = Compilation.log; // const logln = Compilation.logln; const Module = Compilation.Module; const data_structures = @import("../library.zig"); const MyHashMap = data_structures.MyHashMap; const UnpinnedArray = data_structures.UnpinnedArray; pub const bindings = @import("llvm_bindings.zig"); pub const Logger = enum { print_module, function, pub var bitset = std.EnumSet(Logger).initMany(&.{ .print_module, .function, }); }; pub const LLVM = struct { context: *LLVM.Context, module: *LLVM.Module, builder: *LLVM.Builder, debug_info_builder: *LLVM.DebugInfo.Builder, debug_info_file_map: MyHashMap(Compilation.Debug.File.Index, *LLVM.DebugInfo.File) = .{}, debug_type_map: MyHashMap(Compilation.Type.Index, *LLVM.DebugInfo.Type) = .{}, type_name_map: MyHashMap(Compilation.Type.Index, []const u8) = .{}, type_map: MyHashMap(Compilation.Type.Index, *LLVM.Type) = .{}, function_declaration_map: MyHashMap(*Compilation.Debug.Declaration.Global, *LLVM.Value.Constant.Function) = .{}, function_definition_map: MyHashMap(*Compilation.Debug.Declaration.Global, *LLVM.Value.Constant.Function) = .{}, llvm_instruction_map: MyHashMap(Compilation.Instruction.Index, *LLVM.Value) = .{}, llvm_value_map: MyHashMap(Compilation.V, *LLVM.Value) = .{}, llvm_block_map: MyHashMap(Compilation.BasicBlock.Index, *LLVM.Value.BasicBlock) = .{}, llvm_external_functions: MyHashMap(*Compilation.Debug.Declaration.Global, *LLVM.Value.Constant.Function) = .{}, global_variable_map: MyHashMap(*Compilation.Debug.Declaration.Global, *LLVM.Value.Constant.GlobalVariable) = .{}, scope_map: MyHashMap(*Compilation.Debug.Scope, *LLVM.DebugInfo.Scope) = .{}, pointer_type: ?*LLVM.Type.Pointer = null, function: *LLVM.Value.Constant.Function = undefined, exit_block: *LLVM.Value.BasicBlock = undefined, sema_function: *Compilation.Debug.Declaration.Global = undefined, argument_allocas: MyHashMap(Compilation.Instruction.Index, *LLVM.Value) = .{}, return_phi_node: ?*LLVM.Value.Instruction.PhiNode = null, scope: *LLVM.DebugInfo.Scope = undefined, file: *LLVM.DebugInfo.File = undefined, attributes: Attributes, // subprogram: *LLVM.DebugInfo.Subprogram = undefined, arg_index: u32 = 0, tag_count: c_uint = 0, inside_branch: bool = false, pub const x86_64 = struct { pub const initializeTarget = bindings.LLVMInitializeX86Target; pub const initializeTargetInfo = bindings.LLVMInitializeX86TargetInfo; pub const initializeTargetMC = bindings.LLVMInitializeX86TargetMC; pub const initializeAsmPrinter = bindings.LLVMInitializeX86AsmPrinter; pub const initializeAsmParser = bindings.LLVMInitializeX86AsmParser; }; pub const aarch64 = struct { pub const initializeTarget = bindings.LLVMInitializeAArch64Target; pub const initializeTargetInfo = bindings.LLVMInitializeAArch64TargetInfo; pub const initializeTargetMC = bindings.LLVMInitializeAArch64TargetMC; pub const initializeAsmPrinter = bindings.LLVMInitializeAArch64AsmPrinter; pub const initializeAsmParser = bindings.LLVMInitializeAArch64AsmParser; }; pub const Attributes = struct { noreturn: *Attribute, naked: *Attribute, nounwind: *Attribute, inreg: *Attribute, @"noalias": *Attribute, }; pub const Linkage = enum(c_uint) { @"extern" = 0, available_external = 1, link_once_any = 2, link_once_odr = 3, weak_any = 4, weak_odr = 5, appending = 6, internal = 7, private = 8, external_weak = 9, common = 10, }; pub const ThreadLocalMode = enum(c_uint) { not_thread_local = 0, }; pub const Context = opaque { const create = bindings.NativityLLVMCreateContext; const createBasicBlock = bindings.NativityLLVMCreateBasicBlock; const getConstantInt = bindings.NativityLLVMContextGetConstantInt; const getConstString = bindings.NativityLLVMContextGetConstString; const getVoidType = bindings.NativityLLVMGetVoidType; const getIntegerType = bindings.NativityLLVMGetIntegerType; const getFunctionType = bindings.NativityLLVMGetFunctionType; const getPointerType = bindings.NativityLLVMGetPointerType; const getStructType = bindings.NativityLLVMGetStructType; const getIntrinsicType = bindings.NativityLLVMContextGetIntrinsicType; const getAttributeFromEnum = bindings.NativityLLVMContextGetAttributeFromEnum; const getAttributeFromString = bindings.NativityLLVMContextGetAttributeFromString; const getAttributeFromType = bindings.NativityLLVMContextGetAttributeFromType; const getAttributeSet = bindings.NativityLLVMContextGetAttributeSet; }; pub const Module = opaque { const addGlobalVariable = bindings.NativityLLVMModuleAddGlobalVariable; const create = bindings.NativityLLVMCreateModule; const getFunction = bindings.NativityLLVMModuleGetFunction; const createFunction = bindings.NativityLLVModuleCreateFunction; const verify = bindings.NativityLLVMVerifyModule; const toString = bindings.NativityLLVMModuleToString; const getIntrinsicDeclaration = bindings.NativityLLVMModuleGetIntrinsicDeclaration; const createDebugInfoBuilder = bindings.NativityLLVMModuleCreateDebugInfoBuilder; const setTargetMachineDataLayout = bindings.NativityLLVMModuleSetTargetMachineDataLayout; const setTargetTriple = bindings.NativityLLVMModuleSetTargetTriple; const runOptimizationPipeline = bindings.NativityLLVMRunOptimizationPipeline; const addPassesToEmitFile = bindings.NativityLLVMModuleAddPassesToEmitFile; }; pub const Builder = opaque { const create = bindings.NativityLLVMCreateBuilder; const setInsertPoint = bindings.NativityLLVMBuilderSetInsertPoint; const createAdd = bindings.NativityLLVMBuilderCreateAdd; const createAlloca = bindings.NativityLLVMBuilderCreateAlloca; const createAnd = bindings.NativityLLVMBuilderCreateAnd; const createOr = bindings.NativityLLVMBuilderCreateOr; const createCall = bindings.NativityLLVMBuilderCreateCall; const createCast = bindings.NativityLLVMBuilderCreateCast; const createBranch = bindings.NativityLLVMBuilderCreateBranch; const createConditionalBranch = bindings.NativityLLVMBuilderCreateConditionalBranch; const createSwitch = bindings.NativityLLVMBuilderCreateSwitch; const createGEP = bindings.NativityLLVMBuilderCreateGEP; const createStructGEP = bindings.NativityLLVMBuilderCreateStructGEP; const createICmp = bindings.NativityLLVMBuilderCreateICmp; const createLoad = bindings.NativityLLVMBuilderCreateLoad; const createMultiply = bindings.NativityLLVMBuilderCreateMultiply; const createRet = bindings.NativityLLVMBuilderCreateRet; const createShiftLeft = bindings.NativityLLVMBuilderCreateShiftLeft; const createArithmeticShiftRight = bindings.NativityLLVMBuilderCreateArithmeticShiftRight; const createLogicalShiftRight = bindings.NativityLLVMBuilderCreateLogicalShiftRight; const createStore = bindings.NativityLLVMBuilderCreateStore; const createSub = bindings.NativityLLVMBuilderCreateSub; const createUnreachable = bindings.NativityLLVMBuilderCreateUnreachable; const createXor = bindings.NativityLLVMBuilderCreateXor; const createUDiv = bindings.NativityLLVMBuilderCreateUDiv; const createSDiv = bindings.NativityLLVMBuilderCreateSDiv; const createURem = bindings.NativityLLVMBuilderCreateURem; const createSRem = bindings.NativityLLVMBuilderCreateSRem; const createExtractValue = bindings.NativityLLVMBuilderCreateExtractValue; const createInsertValue = bindings.NativityLLVMBuilderCreateInsertValue; const createGlobalString = bindings.NativityLLVMBuilderCreateGlobalString; const createGlobalStringPointer = bindings.NativityLLVMBuilderCreateGlobalStringPointer; const createPhi = bindings.NativityLLVMBuilderCreatePhi; const createMemcpy = bindings.NativityLLVMBuilderCreateMemcpy; const getInsertBlock = bindings.NativityLLVMBuilderGetInsertBlock; const isCurrentBlockTerminated = bindings.NativityLLVMBuilderIsCurrentBlockTerminated; const setCurrentDebugLocation = bindings.NativityLLVMBuilderSetCurrentDebugLocation; }; pub const DebugInfo = struct { pub const AttributeType = enum(c_uint) { address = 0x01, boolean = 0x02, complex_float = 0x03, float = 0x04, signed = 0x05, signed_char = 0x06, unsigned = 0x07, unsigned_char = 0x08, imaginary_float = 0x09, packed_decimal = 0x0a, numeric_string = 0x0b, edited = 0x0c, signed_fixed = 0x0d, unsigned_fixed = 0x0e, decimal_float = 0x0f, UTF = 0x10, UCS = 0x11, ASCII = 0x12, }; pub const CallingConvention = enum(c_uint) { none = 0, normal = 0x01, program = 0x02, nocall = 0x03, pass_by_reference = 0x04, pass_by_value = 0x05, // Vendor extensions GNU_renesas_sh = 0x40, GNU_borland_fastcall_i386 = 0x41, BORLAND_safecall = 0xb0, BORLAND_stdcall = 0xb1, BORLAND_pascal = 0xb2, BORLAND_msfastcall = 0xb3, BORLAND_msreturn = 0xb4, BORLAND_thiscall = 0xb5, BORLAND_fastcall = 0xb6, LLVM_vectorcall = 0xc0, LLVM_Win64 = 0xc1, LLVM_X86_64SysV = 0xc2, LLVM_AAPCS = 0xc3, LLVM_AAPCS_VFP = 0xc4, LLVM_IntelOclBicc = 0xc5, LLVM_SpirFunction = 0xc6, LLVM_OpenCLKernel = 0xc7, LLVM_Swift = 0xc8, LLVM_PreserveMost = 0xc9, LLVM_PreserveAll = 0xca, LLVM_X86RegCall = 0xcb, GDB_IBM_OpenCL = 0xff, }; pub const Builder = opaque { const createCompileUnit = bindings.NativityLLVMDebugInfoBuilderCreateCompileUnit; const createFile = bindings.NativityLLVMDebugInfoBuilderCreateFile; const createFunction = bindings.NativityLLVMDebugInfoBuilderCreateFunction; const createSubroutineType = bindings.NativityLLVMDebugInfoBuilderCreateSubroutineType; const createLexicalBlock = bindings.NativityLLVMDebugInfoBuilderCreateLexicalBlock; const createParameterVariable = bindings.NativityLLVMDebugInfoBuilderCreateParameterVariable; const createAutoVariable = bindings.NativityLLVMDebugInfoBuilderCreateAutoVariable; const createGlobalVariableExpression = bindings.NativityLLVMDebugInfoBuilderCreateGlobalVariableExpression; const createExpression = bindings.NativityLLVMDebugInfoBuilderCreateExpression; const createBasicType = bindings.NativityLLVMDebugInfoBuilderCreateBasicType; const createPointerType = bindings.NativityLLVMDebugInfoBuilderCreatePointerType; const createStructType = bindings.NativityLLVMDebugInfoBuilderCreateStructType; const createArrayType = bindings.NativityLLVMDebugInfoBuilderCreateArrayType; const createEnumerationType = bindings.NativityLLVMDebugInfoBuilderCreateEnumerationType; const createEnumerator = bindings.NativityLLVMDebugInfoBuilderCreateEnumerator; const createReplaceableCompositeType = bindings.NativityLLVMDebugInfoBuilderCreateReplaceableCompositeType; const createMemberType = bindings.NativityLLVMDebugInfoBuilderCreateMemberType; const insertDeclare = bindings.NativityLLVMDebugInfoBuilderInsertDeclare; const finalizeSubprogram = bindings.NativityLLVMDebugInfoBuilderFinalizeSubprogram; const finalize = bindings.NativityLLVMDebugInfoBuilderFinalize; const replaceCompositeTypes = bindings.NativityLLVMDebugInfoBuilderCompositeTypeReplaceTypes; }; pub const CompileUnit = opaque { fn toScope(this: *@This()) *Scope { return @ptrCast(this); } pub const EmissionKind = enum(c_uint) { no_debug = 0, full_debug = 1, line_tables_only = 2, debug_directives_only = 3, }; pub const NameTableKind = enum(c_uint) { default = 0, gnu = 1, none = 2, }; }; pub const Expression = opaque {}; pub const GlobalVariableExpression = opaque {}; pub const LocalVariable = opaque {}; pub const LexicalBlock = opaque { fn toScope(this: *@This()) *Scope { return @ptrCast(this); } }; pub const Node = opaque { pub const Flags = packed struct(u32) { visibility: Visibility, forward_declaration: bool, apple_block: bool, block_by_ref_struct: bool, virtual: bool, artificial: bool, explicit: bool, prototyped: bool, objective_c_class_complete: bool, object_pointer: bool, vector: bool, static_member: bool, lvalue_reference: bool, rvalue_reference: bool, reserved: bool = false, inheritance: Inheritance, introduced_virtual: bool, bit_field: bool, no_return: bool, type_pass_by_value: bool, type_pass_by_reference: bool, enum_class: bool, thunk: bool, non_trivial: bool, big_endian: bool, little_endian: bool, all_calls_described: bool, _: u3 = 0, const Visibility = enum(u2) { none = 0, private = 1, protected = 2, public = 3, }; const Inheritance = enum(u2) { none = 0, single = 1, multiple = 2, virtual = 3, }; }; }; pub const File = opaque { fn toScope(this: *@This()) *Scope { return @ptrCast(this); } }; pub const Language = enum(c_uint) { c = 0x02, }; pub const Scope = opaque { const toSubprogram = bindings.NativityLLVMDebugInfoScopeToSubprogram; }; pub const LocalScope = opaque { fn toScope(this: *@This()) *Scope { return @ptrCast(this); } }; pub const Subprogram = opaque { const getFile = bindings.NativityLLVMDebugInfoSubprogramGetFile; const getArgumentType = bindings.NativityLLVMDebugInfoSubprogramGetArgumentType; fn toLocalScope(this: *@This()) *LocalScope { return @ptrCast(this); } pub const Flags = packed struct(u32) { virtuality: Virtuality, local_to_unit: bool, definition: bool, optimized: bool, pure: bool, elemental: bool, recursive: bool, main_subprogram: bool, deleted: bool, reserved: bool = false, object_c_direct: bool, _: u20 = 0, const Virtuality = enum(u2) { none = 0, virtual = 1, pure_virtual = 2, }; }; }; pub const Type = opaque { const isResolved = bindings.NativityLLLVMDITypeIsResolved; fn toScope(this: *@This()) *Scope { return @ptrCast(this); } pub const Derived = opaque { fn toType(this: *@This()) *LLVM.DebugInfo.Type { return @ptrCast(this); } }; pub const Composite = opaque { fn toType(this: *@This()) *LLVM.DebugInfo.Type { return @ptrCast(this); } }; pub const Enumerator = opaque {}; pub const Subroutine = opaque { fn toType(this: *@This()) *LLVM.DebugInfo.Type { return @ptrCast(this); } }; }; }; pub const FloatAbi = enum(c_uint) { default = 0, soft = 1, hard = 2, }; pub const FloatOperationFusionMode = enum(c_uint) { fast = 0, standard = 1, strict = 2, }; pub const JumpTableType = enum(c_uint) { single = 0, arity = 1, simplified = 2, full = 3, }; pub const ThreadModel = enum(c_uint) { posix = 0, single = 1, }; pub const BasicBlockSection = enum(c_uint) { all = 0, list = 1, labels = 2, preset = 3, none = 4, }; pub const EAbi = enum(c_uint) { unknown = 0, default = 1, eabi4 = 2, eabi5 = 3, gnu = 4, }; pub const DebuggerKind = enum(c_uint) { default = 0, gdb = 1, lldb = 2, sce = 3, dbx = 4, }; pub const GlobalISelAbortMode = enum(c_uint) { disable = 0, enable = 1, disable_with_diagnostic = 2, }; pub const DebugCompressionType = enum(c_uint) { none = 0, zlib = 1, zstd = 2, }; pub const RelocationModel = enum(c_uint) { static = 0, pic = 1, dynamic_no_pic = 2, ropi = 3, rwpi = 4, ropi_rwpi = 5, }; pub const CodeModel = enum(c_uint) { tiny = 0, small = 1, kernel = 2, medium = 3, large = 4, }; pub const PicLevel = enum(c_uint) { not_pic = 0, small_pic = 1, big_pic = 2, }; pub const PieLevel = enum(c_uint) { default = 0, small = 1, large = 2, }; pub const TlsModel = enum(c_uint) { general_dynamic = 0, local_dynamic = 1, initial_exec = 2, local_exec = 3, }; pub const CodegenOptimizationLevel = enum(c_int) { none = 0, less = 1, default = 2, aggressive = 3, }; pub const OptimizationLevel = extern struct { speed_level: c_uint, size_level: c_uint, }; pub const FramePointerKind = enum(c_uint) { none = 0, non_leaf = 1, all = 2, }; pub const CodeGenFileType = enum(c_uint) { assembly = 0, object = 1, null = 2, }; pub const Target = opaque { const createTargetMachine = bindings.NativityLLVMTargetCreateTargetMachine; pub const Machine = opaque {}; // This is a non-LLVM struct const Options = extern struct { bin_utils_version: struct { i32, i32 }, fp_math: extern struct { unsafe: bool, no_infs: bool, no_nans: bool, no_traping: bool, no_signed_zeroes: bool, approx_func: bool, enable_aix_extended_altivec_abi: bool, honor_sign_dependent_rounding: bool, }, no_zeroes_in_bss: bool, guaranteed_tail_call_optimization: bool, stack_symbol_ordering: bool, enable_fast_isel: bool, enable_global_isel: bool, global_isel_abort_mode: GlobalISelAbortMode, use_init_array: bool, disable_integrated_assembler: bool, debug_compression_type: DebugCompressionType, relax_elf_relocations: bool, function_sections: bool, data_sections: bool, ignore_xcoff_visibility: bool, xcoff_traceback_table: bool, unique_section_names: bool, unique_basic_block_section_names: bool, trap_unreachable: bool, no_trap_after_noreturn: bool, tls_size: u8, emulated_tls: bool, enable_ipra: bool, emit_stack_size_section: bool, enable_machine_outliner: bool, enable_machine_function_splitter: bool, support_default_outlining: bool, emit_address_significance_table: bool, bb_sections: BasicBlockSection, emit_call_site_info: bool, support_debug_entry_values: bool, enable_debug_entry_values: bool, value_tracking_variable_locations: bool, force_dwarf_frame_section: bool, xray_function_index: bool, debug_strict_dwarf: bool, hotpatch: bool, ppc_gen_scalar_mass_entries: bool, jmc_instrument: bool, cfi_fixup: bool, loop_alignment: u32 = 0, float_abi_type: FloatAbi, fp_operation_fusion: FloatOperationFusionMode, thread_model: ThreadModel, eabi_version: EAbi, debugger_tuning: DebuggerKind, }; }; const lookupIntrinsic = bindings.NativityLLVMLookupIntrinsic; const newPhiNode = bindings.NativityLLVMCreatePhiNode; pub const Metadata = opaque { pub const Node = opaque {}; pub const Tuple = opaque {}; }; pub const Attribute = opaque { pub const Set = opaque {}; pub const Id = enum(u32) { AllocAlign = 1, AllocatedPointer = 2, AlwaysInline = 3, Builtin = 4, Cold = 5, Convergent = 6, DisableSanitizerInstrumentation = 7, FnRetThunkExtern = 8, Hot = 9, ImmArg = 10, InReg = 11, InlineHint = 12, JumpTable = 13, MinSize = 14, MustProgress = 15, Naked = 16, Nest = 17, NoAlias = 18, NoBuiltin = 19, NoCallback = 20, NoCapture = 21, NoCfCheck = 22, NoDuplicate = 23, NoFree = 24, NoImplicitFloat = 25, NoInline = 26, NoMerge = 27, NoProfile = 28, NoRecurse = 29, NoRedZone = 30, NoReturn = 31, NoSanitizeBounds = 32, NoSanitizeCoverage = 33, NoSync = 34, NoUndef = 35, NoUnwind = 36, NonLazyBind = 37, NonNull = 38, NullPointerIsValid = 39, OptForFuzzing = 40, OptimizeForSize = 41, OptimizeNone = 42, PresplitCoroutine = 43, ReadNone = 44, ReadOnly = 45, Returned = 46, ReturnsTwice = 47, SExt = 48, SafeStack = 49, SanitizeAddress = 50, SanitizeHWAddress = 51, SanitizeMemTag = 52, SanitizeMemory = 53, SanitizeThread = 54, ShadowCallStack = 55, SkipProfile = 56, Speculatable = 57, SpeculativeLoadHardening = 58, StackProtect = 59, StackProtectReq = 60, StackProtectStrong = 61, StrictFP = 62, SwiftAsync = 63, SwiftError = 64, SwiftSelf = 65, WillReturn = 66, WriteOnly = 67, ZExt = 68, ByRef = 69, ByVal = 70, ElementType = 71, InAlloca = 72, Preallocated = 73, StructRet = 74, Alignment = 75, AllocKind = 76, AllocSize = 77, Dereferenceable = 78, DereferenceableOrNull = 79, Memory = 80, StackAlignment = 81, UWTable = 82, VScaleRange = 83, }; }; pub const Type = opaque { const compare = bindings.NativityLLVMCompareTypes; const toStruct = bindings.NativityLLVMTypeToStruct; const toFunction = bindings.NativityLLVMTypeToFunction; const toArray = bindings.NativityLLVMTypeToArray; const toPointer = bindings.NativityLLVMTypeToPointer; const isPointer = bindings.NativityLLVMTypeIsPointer; const isInteger = bindings.NativityLLVMTypeIsInteger; const isVoid = bindings.NativityLLVMTypeIsVoid; const assertEqual = bindings.NativityLLVMTypeAssertEqual; pub const Array = opaque { fn toType(integer: *@This()) *Type { return @ptrCast(integer); } const get = bindings.NativityLLVMGetArrayType; const getConstant = bindings.NativityLLVMGetConstantArray; const getElementType = bindings.NativityLLVMArrayTypeGetElementType; }; pub const Integer = opaque { fn toType(integer: *@This()) *Type { return @ptrCast(integer); } }; pub const Function = opaque { fn toType(integer: *@This()) *Type { return @ptrCast(integer); } const getArgumentType = bindings.NativityLLVMFunctionTypeGetArgumentType; const getReturnType = bindings.NativityLLVMFunctionTypeGetReturnType; }; pub const Pointer = opaque { fn toType(integer: *@This()) *Type { return @ptrCast(integer); } const getNull = bindings.NativityLLVMPointerTypeGetNull; }; pub const Struct = opaque { const getConstant = bindings.NativityLLVMGetConstantStruct; fn toType(integer: *@This()) *Type { return @ptrCast(integer); } }; pub const Error = error{ void, function, integer, pointer, @"struct", intrinsic, array, }; const getPoison = bindings.NativityLLVMGetPoisonValue; }; pub const Value = opaque { const setName = bindings.NativityLLVMValueSetName; const getType = bindings.NativityLLVMValueGetType; const toConstant = bindings.NativityLLVMValueToConstant; const toFunction = bindings.NativityLLVMValueToFunction; const toAlloca = bindings.NativityLLVMValueToAlloca; const toString = bindings.NativityLLVMValueToString; pub const IntrinsicID = enum(u32) { none = 0, _, }; pub const BasicBlock = opaque { const remove = bindings.NativityLLVMBasicBlockRemoveFromParent; fn toValue(this: *@This()) *Value { return @ptrCast(this); } }; pub const Argument = opaque { const getIndex = bindings.NativityLLVMArgumentGetIndex; fn toValue(this: *@This()) *Value { return @ptrCast(this); } }; pub const Instruction = opaque { fn toValue(this: *@This()) *Value { return @ptrCast(this); } pub const Alloca = opaque { fn toValue(this: *@This()) *Value { return @ptrCast(this); } const getAllocatedType = bindings.NativityLLVMAllocatGetAllocatedType; }; pub const Branch = opaque { fn toValue(this: *@This()) *Value { return @ptrCast(this); } }; pub const Call = opaque { const setCallingConvention = bindings.NativityLLVMCallSetCallingConvention; const setAttributes = bindings.NativityLLVMCallSetAttributes; fn toValue(this: *@This()) *Value { return @ptrCast(this); } }; pub const Cast = opaque { pub const Type = enum(c_uint) { truncate = 38, zero_extend = 39, sign_extend = 40, float_to_unsigned_integer = 41, float_to_signed_integer = 42, unsigned_integer_to_float = 43, signed_integer_to_float = 44, float_truncate = 45, float_extend = 46, pointer_to_int = 47, int_to_pointer = 48, bitcast = 49, address_space_cast = 50, }; fn toValue(this: *@This()) *Value { return @ptrCast(this); } }; pub const ICmp = opaque { fn toValue(this: *@This()) *Value { return @ptrCast(this); } pub const Kind = enum(c_uint) { eq = 32, // equal ne = 33, // not equal ugt = 34, // unsigned greater than uge = 35, // unsigned greater or equal ult = 36, // unsigned less than ule = 37, // unsigned less or equal sgt = 38, // signed greater than sge = 39, // signed greater or equal slt = 40, // signed less than sle = 41, // signed less or equal }; }; pub const Load = opaque { fn toValue(this: *@This()) *Value { return @ptrCast(this); } }; pub const PhiNode = opaque { pub const addIncoming = bindings.NativityLLVMPhiAddIncoming; fn toValue(this: *@This()) *Value { return @ptrCast(this); } }; pub const Store = opaque { fn toValue(this: *@This()) *Value { return @ptrCast(this); } }; pub const Switch = opaque { fn toValue(this: *@This()) *Value { return @ptrCast(this); } }; pub const Ret = opaque { fn toValue(this: *@This()) *Value { return @ptrCast(this); } }; pub const Unreachable = opaque { fn toValue(this: *@This()) *Value { return @ptrCast(this); } }; pub const Error = error{ add, alloca, @"and", arithmetic_shift_right, call, cast, conditional_branch, extract_value, gep, icmp, insert_value, load, logical_shift_right, multiply, @"or", ret, sdiv, shift_left, store, udiv, @"unreachable", xor, }; }; pub const Constant = opaque { pub const Function = opaque { const getArgument = bindings.NativityLLVMFunctionGetArgument; const getArguments = bindings.NativityLLVMFunctionGetArguments; const getType = bindings.NativityLLVMFunctionGetType; // const addAttributeKey = bindings.NativityLLVMFunctionAddAttributeKey; const verify = bindings.NativityLLVMVerifyFunction; const toString = bindings.NativityLLVMFunctionToString; const setCallingConvention = bindings.NativityLLVMFunctionSetCallingConvention; const getCallingConvention = bindings.NativityLLVMFunctionGetCallingConvention; const setSubprogram = bindings.NativityLLVMFunctionSetSubprogram; const getSubprogram = bindings.NativityLLVMFunctionGetSubprogram; const setAttributes = bindings.NativityLLVMFunctionSetAttributes; fn toValue(this: *@This()) *Value { return @ptrCast(this); } fn toConstant(this: *@This()) *Constant { return @ptrCast(this); } pub const CallingConvention = enum(c_uint) { /// The default llvm calling convention, compatible with C. This convention /// is the only one that supports varargs calls. As with typical C calling /// conventions, the callee/caller have to tolerate certain amounts of /// prototype mismatch. C = 0, // Generic LLVM calling conventions. None of these support varargs calls, // and all assume that the caller and callee prototype exactly match. /// Attempts to make calls as fast as possible (e.g. by passing things in /// registers). Fast = 8, /// Attempts to make code in the caller as efficient as possible under the /// assumption that the call is not commonly executed. As such, these calls /// often preserve all registers so that the call does not break any live /// ranges in the caller side. Cold = 9, /// Used by the Glasgow Haskell Compiler (GHC). GHC = 10, /// Used by the High-Performance Erlang Compiler (HiPE). HiPE = 11, /// Used for stack based JavaScript calls WebKit_JS = 12, /// Used for dynamic register based calls (e.g. stackmap and patchpoint /// intrinsics). AnyReg = 13, /// Used for runtime calls that preserves most registers. PreserveMost = 14, /// Used for runtime calls that preserves (almost) all registers. PreserveAll = 15, /// Calling convention for Swift. Swift = 16, /// Used for access functions. CXX_FAST_TLS = 17, /// Attemps to make calls as fast as possible while guaranteeing that tail /// call optimization can always be performed. Tail = 18, /// Special calling convention on Windows for calling the Control Guard /// Check ICall funtion. The function takes exactly one argument (address of /// the target function) passed in the first argument register, and has no /// return value. All register values are preserved. CFGuard_Check = 19, /// This follows the Swift calling convention in how arguments are passed /// but guarantees tail calls will be made by making the callee clean up /// their stack. SwiftTail = 20, /// This is the start of the target-specific calling conventions, e.g. /// fastcall and thiscall on X86. // FirstTargetCC = 64, /// stdcall is mostly used by the Win32 API. It is basically the same as the /// C convention with the difference in that the callee is responsible for /// popping the arguments from the stack. X86_StdCall = 64, /// 'fast' analog of X86_StdCall. Passes first two arguments in ECX:EDX /// registers, others - via stack. Callee is responsible for stack cleaning. X86_FastCall = 65, /// ARM Procedure Calling Standard (obsolete, but still used on some /// targets). ARM_APCS = 66, /// ARM Architecture Procedure Calling Standard calling convention (aka /// EABI). Soft float variant. ARM_AAPCS = 67, /// Same as ARM_AAPCS, but uses hard floating point ABI. ARM_AAPCS_VFP = 68, /// Used for MSP430 interrupt routines. MSP430_INTR = 69, /// Similar to X86_StdCall. Passes first argument in ECX, others via stack. /// Callee is responsible for stack cleaning. MSVC uses this by default for /// methods in its ABI. X86_ThisCall = 70, /// Call to a PTX kernel. Passes all arguments in parameter space. PTX_Kernel = 71, /// Call to a PTX device function. Passes all arguments in register or /// parameter space. PTX_Device = 72, /// Used for SPIR non-kernel device functions. No lowering or expansion of /// arguments. Structures are passed as a pointer to a struct with the /// byval attribute. Functions can only call SPIR_FUNC and SPIR_KERNEL /// functions. Functions can only have zero or one return values. Variable /// arguments are not allowed, except for printf. How arguments/return /// values are lowered are not specified. Functions are only visible to the /// devices. SPIR_FUNC = 75, /// Used for SPIR kernel functions. Inherits the restrictions of SPIR_FUNC, /// except it cannot have non-void return values, it cannot have variable /// arguments, it can also be called by the host or it is externally /// visible. SPIR_KERNEL = 76, /// Used for Intel OpenCL built-ins. Intel_OCL_BI = 77, /// The C convention as specified in the x86-64 supplement to the System V /// ABI, used on most non-Windows systems. X86_64_SysV = 78, /// The C convention as implemented on Windows/x86-64 and AArch64. It /// differs from the more common \c X86_64_SysV convention in a number of /// ways, most notably in that XMM registers used to pass arguments are /// shadowed by GPRs, and vice versa. On AArch64, this is identical to the /// normal C (AAPCS) calling convention for normal functions, but floats are /// passed in integer registers to variadic functions. Win64 = 79, /// MSVC calling convention that passes vectors and vector aggregates in SSE /// registers. X86_VectorCall = 80, /// Used by HipHop Virtual Machine (HHVM) to perform calls to and from /// translation cache, and for calling PHP functions. HHVM calling /// convention supports tail/sibling call elimination. HHVM = 81, /// HHVM calling convention for invoking C/C++ helpers. HHVM_C = 82, /// x86 hardware interrupt context. Callee may take one or two parameters, /// where the 1st represents a pointer to hardware context frame and the 2nd /// represents hardware error code, the presence of the later depends on the /// interrupt vector taken. Valid for both 32- and 64-bit subtargets. X86_INTR = 83, /// Used for AVR interrupt routines. AVR_INTR = 84, /// Used for AVR signal routines. AVR_SIGNAL = 85, /// Used for special AVR rtlib functions which have an "optimized" /// convention to preserve registers. AVR_BUILTIN = 86, /// Used for Mesa vertex shaders, or AMDPAL last shader stage before /// rasterization (vertex shader if tessellation and geometry are not in /// use, or otherwise copy shader if one is needed). AMDGPU_VS = 87, /// Used for Mesa/AMDPAL geometry shaders. AMDGPU_GS = 88, /// Used for Mesa/AMDPAL pixel shaders. AMDGPU_PS = 89, /// Used for Mesa/AMDPAL compute shaders. AMDGPU_CS = 90, /// Used for AMDGPU code object kernels. AMDGPU_KERNEL = 91, /// Register calling convention used for parameters transfer optimization X86_RegCall = 92, /// Used for Mesa/AMDPAL hull shaders (= tessellation control shaders). AMDGPU_HS = 93, /// Used for special MSP430 rtlib functions which have an "optimized" /// convention using additional registers. MSP430_BUILTIN = 94, /// Used for AMDPAL vertex shader if tessellation is in use. AMDGPU_LS = 95, /// Used for AMDPAL shader stage before geometry shader if geometry is in /// use. So either the domain (= tessellation evaluation) shader if /// tessellation is in use, or otherwise the vertex shader. AMDGPU_ES = 96, /// Used between AArch64 Advanced SIMD functions AArch64_VectorCall = 97, /// Used between AArch64 SVE functions AArch64_SVE_VectorCall = 98, /// For emscripten __invoke_* functions. The first argument is required to /// be the function ptr being indirectly called. The remainder matches the /// regular calling convention. WASM_EmscriptenInvoke = 99, /// Used for AMD graphics targets. AMDGPU_Gfx = 100, /// Used for M68k interrupt routines. M68k_INTR = 101, /// Preserve X0-X13, X19-X29, SP, Z0-Z31, P0-P15. AArch64_SME_ABI_Support_Routines_PreserveMost_From_X0 = 102, /// Preserve X2-X15, X19-X29, SP, Z0-Z31, P0-P15. AArch64_SME_ABI_Support_Routines_PreserveMost_From_X2 = 103, /// The highest possible ID. Must be some 2^k - 1. MaxID = 1023, }; }; pub const Int = opaque { fn toValue(this: *@This()) *Value { return @ptrCast(this); } fn toConstant(this: *@This()) *Constant { return @ptrCast(this); } }; pub const GlobalVariable = opaque { pub const setInitializer = bindings.NativityLLVMGlobalVariableSetInitializer; fn toValue(this: *@This()) *Value { return @ptrCast(this); } fn toConstant(this: *@This()) *Constant { return @ptrCast(this); } }; pub const PointerNull = opaque { fn toValue(this: *@This()) *Value { return @ptrCast(this); } fn toConstant(this: *@This()) *Constant { return @ptrCast(this); } }; pub const Undefined = opaque { fn toConstant(this: *@This()) *Constant { return @ptrCast(this); } fn toValue(this: *@This()) *Value { return @ptrCast(this); } }; pub const Poison = opaque { fn toConstant(this: *@This()) *Constant { return @ptrCast(this); } fn toValue(this: *@This()) *Value { return @ptrCast(this); } }; fn toValue(this: *@This()) *Value { return @ptrCast(this); } const toInt = bindings.NativityLLVMConstantToInt; }; pub const InlineAssembly = opaque { pub const Dialect = enum(c_uint) { @"at&t", intel, }; const get = bindings.NativityLLVMGetInlineAssembly; fn toValue(this: *@This()) *Value { return @ptrCast(this); } }; pub const Error = error{ constant_struct, constant_int, constant_array, inline_assembly, global_variable, intrinsic, }; }; fn getType(llvm: *LLVM, unit: *Compilation.Unit, context: *const Compilation.Context, type_index: Compilation.Type.Index) !*LLVM.Type { if (llvm.type_map.get(type_index)) |llvm_type| { return llvm_type; } else { const sema_type = unit.types.get(type_index); const llvm_type: *LLVM.Type = switch (sema_type.*) { .function => |function_prototype_index| blk: { const sema_function_prototype = unit.function_prototypes.get(function_prototype_index); const llvm_return_type = try llvm.getType(unit, context, sema_function_prototype.abi.return_type); var parameter_types = try UnpinnedArray(*LLVM.Type).initialize_with_capacity(context.my_allocator, @intCast(sema_function_prototype.abi.parameter_types.len)); for (sema_function_prototype.abi.parameter_types) |argument_type_index| { parameter_types.append_with_capacity(try llvm.getType(unit, context, argument_type_index)); } const is_var_args = false; const llvm_function_type = LLVM.Context.getFunctionType(llvm_return_type, parameter_types.pointer, parameter_types.length, is_var_args) orelse return Type.Error.function; break :blk llvm_function_type.toType(); }, .integer => |integer| switch (integer.kind) { .comptime_int => unreachable, else => blk: { const llvm_integer_type = llvm.context.getIntegerType(integer.bit_count) orelse return Type.Error.integer; break :blk llvm_integer_type.toType(); }, }, .pointer => { if (llvm.pointer_type) |pointer_type| { return pointer_type.toType(); } else { const pointer_type = llvm.context.getPointerType(address_space) orelse return Type.Error.pointer; llvm.pointer_type = pointer_type; return pointer_type.toType(); } }, .noreturn, .void, => blk: { const void_type = llvm.context.getVoidType() orelse return Type.Error.void; break :blk void_type; }, .slice => |slice| blk: { const llvm_pointer_type = try llvm.getType(unit, context, slice.child_pointer_type); const llvm_usize_type = try llvm.getType(unit, context, Compilation.Type.usize); const slice_types = [_]*Type{ llvm_pointer_type, llvm_usize_type }; const is_packed = false; const struct_type = llvm.context.getStructType(&slice_types, slice_types.len, is_packed) orelse return Type.Error.@"struct"; break :blk struct_type.toType(); }, .@"struct" => |struct_type_index| switch (unit.structs.get(struct_type_index).kind) { .error_union => |error_union| try llvm.getType(unit, context, error_union.abi), .raw_error_union => |error_union_type_index| blk: { const error_union_type = try llvm.getType(unit, context, error_union_type_index); const boolean_type = try llvm.getType(unit, context, .bool); const types = [2]*LLVM.Type{ error_union_type, boolean_type }; const is_packed = false; const struct_type = llvm.context.getStructType(&types, types.len, is_packed) orelse return Type.Error.@"struct"; break :blk struct_type.toType(); }, .abi_compatible_error_union => |error_union| blk: { const error_union_type = try llvm.getType(unit, context, error_union.type); const boolean_type = try llvm.getType(unit, context, .bool); var three_types: [3]*LLVM.Type = undefined; var two_types: [2]*LLVM.Type = undefined; const types: []const *LLVM.Type = if (error_union.padding == .null) b: { two_types = .{ error_union_type, boolean_type }; break :b &two_types; } else b: { const padding_type = try llvm.getType(unit, context, error_union.padding); three_types = .{ error_union_type, padding_type, boolean_type }; break :b &three_types; }; const is_packed = false; const struct_type = llvm.context.getStructType(types.ptr, types.len, is_packed) orelse return Type.Error.@"struct"; break :blk struct_type.toType(); }, .@"struct" => |*sema_struct_type| blk: { var field_type_list = try UnpinnedArray(*LLVM.Type).initialize_with_capacity(context.my_allocator, sema_struct_type.fields.length); for (sema_struct_type.fields.slice()) |sema_field_index| { const sema_field = unit.struct_fields.get(sema_field_index); const llvm_type = try llvm.getType(unit, context, sema_field.type); field_type_list.append_with_capacity(llvm_type); } // TODO: const is_packed = false; const struct_type = llvm.context.getStructType(field_type_list.pointer, field_type_list.length, is_packed) orelse return Type.Error.@"struct"; break :blk struct_type.toType(); }, .two_struct => |pair| blk: { const types = [2]*LLVM.Type{ try llvm.getType(unit, context, pair[0]), try llvm.getType(unit, context, pair[1]), }; const is_packed = false; const struct_type = llvm.context.getStructType(&types, types.len, is_packed) orelse return Type.Error.@"struct"; break :blk struct_type.toType(); }, else => |t| @panic(@tagName(t)), }, .array => |array| blk: { const element_type = try llvm.getType(unit, context, array.type); const extra_element = switch (array.termination) { .none => false, else => true, }; const array_type = LLVM.Type.Array.get(element_type, array.count + @intFromBool(extra_element)) orelse return Type.Error.array; break :blk array_type.toType(); }, .any => (llvm.pointer_type orelse unreachable).toType(), else => |t| @panic(@tagName(t)), }; try llvm.type_map.put_no_clobber(context.my_allocator, type_index, llvm_type); return llvm_type; } } fn getDebugInfoFile(llvm: *LLVM, unit: *Compilation.Unit, context: *const Compilation.Context, sema_file_index: Compilation.Debug.File.Index) !*DebugInfo.File { if (llvm.debug_info_file_map.get(sema_file_index)) |file| { return file; } else { // if (@intFromEnum(sema_file_index) == 4) @breakpoint(); const sema_file = unit.files.get(sema_file_index); const full_path = try Compilation.joinPath(context, sema_file.package.directory.path, sema_file.relative_path); const filename = std.fs.path.basename(full_path); const directory = full_path[0 .. full_path.len - filename.len]; const debug_file = llvm.debug_info_builder.createFile(filename.ptr, filename.len, directory.ptr, directory.len) orelse unreachable; try llvm.debug_info_file_map.put_no_clobber(context.my_allocator, sema_file_index, debug_file); return debug_file; } } fn renderTypeName(llvm: *LLVM, unit: *Compilation.Unit, context: *const Compilation.Context, sema_type_index: Compilation.Type.Index) ![]const u8 { if (llvm.type_name_map.get(sema_type_index)) |result| { return result; } else { if (unit.type_declarations.get(sema_type_index)) |global_declaration| { const result = unit.getIdentifier(global_declaration.declaration.name); try llvm.type_name_map.put_no_clobber(context.my_allocator, sema_type_index, result); return result; } else { const sema_type = unit.types.get(sema_type_index); const result: []const u8 = switch (sema_type.*) { .integer => |integer| switch (integer.kind) { .materialized_int => b: { var buffer: [65]u8 = undefined; const format = data_structures.format_int(&buffer, integer.bit_count, 10, false); const slice_ptr = format.ptr - 1; const slice = slice_ptr[0 .. format.len + 1]; slice[0] = switch (integer.signedness) { .signed => 's', .unsigned => 'u', }; break :b try context.my_allocator.duplicate_bytes(slice); }, .bool => "bool", else => |t| @panic(@tagName(t)), }, // .bool => "bool", .pointer => |pointer| b: { var name = UnpinnedArray(u8){}; try name.append(context.my_allocator, '&'); if (pointer.mutability == .@"const") { try name.append_slice(context.my_allocator, "const"); } try name.append(context.my_allocator, ' '); const element_type_name = try llvm.renderTypeName(unit, context, pointer.type); try name.append_slice(context.my_allocator, element_type_name); break :b name.slice(); }, .@"struct" => |struct_index| switch (unit.structs.get(struct_index).kind) { .@"struct" => "anon_struct", // TODO: else => |t| @panic(@tagName(t)), }, // TODO: termination .slice => |slice| b: { var name = UnpinnedArray(u8){}; try name.append_slice(context.my_allocator, "[] "); if (slice.mutability == .@"const") { try name.append_slice(context.my_allocator, "const "); } const element_type_name = try llvm.renderTypeName(unit, context, slice.child_type); try name.append_slice(context.my_allocator, element_type_name); break :b name.slice(); }, .array => |array| b: { var name = UnpinnedArray(u8){}; try name.append(context.my_allocator, '['); var buffer: [65]u8 = undefined; const array_count = data_structures.format_int(&buffer, array.count, 10, false); try name.append_slice(context.my_allocator, array_count); try name.append(context.my_allocator, ']'); const element_type_name = try llvm.renderTypeName(unit, context, array.type); try name.append_slice(context.my_allocator, element_type_name); break :b name.slice(); }, // TODO .function => "fn_type", .any => "any", else => |t| @panic(@tagName(t)), }; try llvm.type_name_map.put(context.my_allocator, sema_type_index, result); return result; } } } fn createDebugStructType(llvm: *LLVM, arguments: struct { scope: ?*LLVM.DebugInfo.Scope, name: []const u8, file: ?*LLVM.DebugInfo.File, line: u32, bitsize: u64, alignment: u32, field_types: []const *LLVM.DebugInfo.Type, forward_declaration: ?*LLVM.DebugInfo.Type.Composite, }) *LLVM.DebugInfo.Type.Composite { const flags = LLVM.DebugInfo.Node.Flags{ .visibility = .none, .forward_declaration = false, .apple_block = false, .block_by_ref_struct = false, .virtual = false, .artificial = false, .explicit = false, .prototyped = false, .objective_c_class_complete = false, .object_pointer = false, .vector = false, .static_member = false, .lvalue_reference = false, .rvalue_reference = false, .reserved = false, .inheritance = .none, .introduced_virtual = false, .bit_field = false, .no_return = false, .type_pass_by_value = false, .type_pass_by_reference = false, .enum_class = false, .thunk = false, .non_trivial = false, .big_endian = false, .little_endian = false, .all_calls_described = false, }; const struct_type = llvm.debug_info_builder.createStructType(arguments.scope, arguments.name.ptr, arguments.name.len, arguments.file, arguments.line, arguments.bitsize, arguments.alignment, flags, null, arguments.field_types.ptr, arguments.field_types.len, arguments.forward_declaration) orelse unreachable; return struct_type; } fn getDebugStructType(llvm: *LLVM, unit: *Compilation.Unit, context: *const Compilation.Context, sema_type_index: Compilation.Type.Index, scope: *const Compilation.Debug.Scope, fields: []const Compilation.Struct.Field.Index, name: []const u8) !*LLVM.DebugInfo.Type { const file = try llvm.getDebugInfoFile(unit, context, scope.file); const line = 0; const flags = LLVM.DebugInfo.Node.Flags{ .visibility = .none, .forward_declaration = false, .apple_block = false, .block_by_ref_struct = false, .virtual = false, .artificial = false, .explicit = false, .prototyped = false, .objective_c_class_complete = false, .object_pointer = false, .vector = false, .static_member = false, .lvalue_reference = false, .rvalue_reference = false, .reserved = false, .inheritance = .none, .introduced_virtual = false, .bit_field = false, .no_return = false, .type_pass_by_value = false, .type_pass_by_reference = false, .enum_class = false, .thunk = false, .non_trivial = false, .big_endian = false, .little_endian = false, .all_calls_described = false, }; var bit_size: u32 = 0; for (fields) |struct_field_index| { const struct_field = unit.struct_fields.get(struct_field_index); const struct_field_type = unit.types.get(struct_field.type); const struct_field_bit_size = struct_field_type.getBitSize(unit); bit_size += struct_field_bit_size; } const struct_type = llvm.createDebugStructType(.{ .scope = null, .name = name, .file = file, .line = line, .bitsize = bit_size, .alignment = 0, .field_types = &.{}, .forward_declaration = null, }); try llvm.debug_type_map.put_no_clobber(context.my_allocator, sema_type_index, struct_type.toType()); var field_types = try UnpinnedArray(*LLVM.DebugInfo.Type).initialize_with_capacity(context.my_allocator, @intCast(fields.len)); bit_size = 0; for (fields) |struct_field_index| { const struct_field = unit.struct_fields.get(struct_field_index); const struct_field_type = unit.types.get(struct_field.type); const struct_field_bit_size = struct_field_type.getBitSize(unit); const field_type = try llvm.getDebugType(unit, context, struct_field.type); const field_name = unit.getIdentifier(struct_field.name); const alignment = struct_field_bit_size; const member_type = llvm.debug_info_builder.createMemberType(null, field_name.ptr, field_name.len, file, 0, struct_field_bit_size, alignment, bit_size, flags, field_type).toType(); field_types.append_with_capacity(member_type); bit_size += struct_field_bit_size; } llvm.debug_info_builder.replaceCompositeTypes(struct_type, field_types.pointer, field_types.length); return struct_type.toType(); } fn getDebugType(llvm: *LLVM, unit: *Compilation.Unit, context: *const Compilation.Context, sema_type_index: Compilation.Type.Index) anyerror!*LLVM.DebugInfo.Type { if (llvm.debug_type_map.get(sema_type_index)) |result| { return result; } else { const name = try llvm.renderTypeName(unit, context, sema_type_index); const sema_type = unit.types.get(sema_type_index); const result = switch (sema_type.*) { .integer => |*integer| switch (integer.kind) { .bitfield => |*bitfield| try llvm.getDebugStructType(unit, context, sema_type_index, &bitfield.scope.scope, bitfield.fields.slice(), name), .materialized_int => b: { const dwarf_encoding: LLVM.DebugInfo.AttributeType = switch (integer.signedness) { .unsigned => .unsigned, .signed => .signed, }; const flags = LLVM.DebugInfo.Node.Flags{ .visibility = .none, .forward_declaration = false, .apple_block = false, .block_by_ref_struct = false, .virtual = false, .artificial = false, .explicit = false, .prototyped = false, .objective_c_class_complete = false, .object_pointer = false, .vector = false, .static_member = false, .lvalue_reference = false, .rvalue_reference = false, .reserved = false, .inheritance = .none, .introduced_virtual = false, .bit_field = false, .no_return = false, .type_pass_by_value = false, .type_pass_by_reference = false, .enum_class = false, .thunk = false, .non_trivial = false, .big_endian = false, .little_endian = false, .all_calls_described = false, }; const integer_type = llvm.debug_info_builder.createBasicType(name.ptr, name.len, integer.bit_count, dwarf_encoding, flags) orelse unreachable; break :b integer_type; }, .bool => b: { const flags = LLVM.DebugInfo.Node.Flags{ .visibility = .none, .forward_declaration = false, .apple_block = false, .block_by_ref_struct = false, .virtual = false, .artificial = false, .explicit = false, .prototyped = false, .objective_c_class_complete = false, .object_pointer = false, .vector = false, .static_member = false, .lvalue_reference = false, .rvalue_reference = false, .reserved = false, .inheritance = .none, .introduced_virtual = false, .bit_field = false, .no_return = false, .type_pass_by_value = false, .type_pass_by_reference = false, .enum_class = false, .thunk = false, .non_trivial = false, .big_endian = false, .little_endian = false, .all_calls_described = false, }; const boolean_type = llvm.debug_info_builder.createBasicType("bool", "bool".len, 1, .boolean, flags) orelse unreachable; break :b boolean_type; }, .@"enum" => |*enum_type| b: { var enumerators = try UnpinnedArray(*LLVM.DebugInfo.Type.Enumerator).initialize_with_capacity(context.my_allocator, enum_type.fields.length); for (enum_type.fields.slice()) |enum_field_index| { const enum_field = unit.enum_fields.get(enum_field_index); const enum_field_name = unit.getIdentifier(enum_field.name); const is_unsigned = true; const enumerator = llvm.debug_info_builder.createEnumerator(enum_field_name.ptr, enum_field_name.len, enum_field.value, is_unsigned) orelse unreachable; enumerators.append_with_capacity(enumerator); } const type_declaration = unit.type_declarations.get(sema_type_index).?; const file = try llvm.getDebugInfoFile(unit, context, type_declaration.declaration.scope.file); const bit_size = integer.bit_count; const sema_backing_type = try unit.getIntegerType(context, .{ .kind = .materialized_int, .bit_count = integer.bit_count, .signedness = integer.signedness, }); const backing_type = try llvm.getDebugType(unit, context, sema_backing_type); const alignment = 0; const line = type_declaration.declaration.line + 1; const scope = try llvm.getScope(unit, context, enum_type.scope.scope.parent.?); const enumeration_type = llvm.debug_info_builder.createEnumerationType(scope, name.ptr, name.len, file, line, bit_size, alignment, enumerators.pointer, enumerators.length, backing_type) orelse unreachable; break :b enumeration_type.toType(); }, .@"error" => |*error_type| b: { var enumerators = try UnpinnedArray(*LLVM.DebugInfo.Type.Enumerator).initialize_with_capacity(context.my_allocator, error_type.fields.length); for (error_type.fields.slice()) |error_field_index| { const error_field = unit.error_fields.get(error_field_index); const error_field_name = unit.getIdentifier(error_field.name); const is_unsigned = true; const enumerator = llvm.debug_info_builder.createEnumerator(error_field_name.ptr, error_field_name.len, error_field.value, is_unsigned) orelse unreachable; enumerators.append_with_capacity(enumerator); } const type_declaration = unit.type_declarations.get(sema_type_index).?; const file = try llvm.getDebugInfoFile(unit, context, type_declaration.declaration.scope.file); const bit_size = integer.bit_count; const sema_backing_type = try unit.getIntegerType(context, .{ .kind = .materialized_int, .bit_count = integer.bit_count, .signedness = integer.signedness, }); const backing_type = try llvm.getDebugType(unit, context, sema_backing_type); const alignment = 0; const line = type_declaration.declaration.line + 1; const scope = try llvm.getScope(unit, context, error_type.scope.scope.parent.?); const enumeration_type = llvm.debug_info_builder.createEnumerationType(scope, name.ptr, name.len, file, line, bit_size, alignment, enumerators.pointer, enumerators.length, backing_type) orelse unreachable; break :b enumeration_type.toType(); }, else => |t| @panic(@tagName(t)), }, .@"struct" => |struct_index| switch (unit.structs.get(struct_index).kind) { .@"struct" => |*sema_struct_type| try llvm.getDebugStructType(unit, context, sema_type_index, &sema_struct_type.scope.scope, sema_struct_type.fields.slice(), name), else => |t| @panic(@tagName(t)), }, .pointer => |pointer| b: { if (pointer.type == .any) { const flags = LLVM.DebugInfo.Node.Flags{ .visibility = .none, .forward_declaration = false, .apple_block = false, .block_by_ref_struct = false, .virtual = false, .artificial = false, .explicit = false, .prototyped = false, .objective_c_class_complete = false, .object_pointer = false, .vector = false, .static_member = false, .lvalue_reference = false, .rvalue_reference = false, .reserved = false, .inheritance = .none, .introduced_virtual = false, .bit_field = false, .no_return = false, .type_pass_by_value = false, .type_pass_by_reference = false, .enum_class = false, .thunk = false, .non_trivial = false, .big_endian = false, .little_endian = false, .all_calls_described = false, }; const pointer_type = llvm.debug_info_builder.createBasicType(name.ptr, name.len, 64, .address, flags) orelse unreachable; break :b pointer_type; } else { const element_type = try llvm.getDebugType(unit, context, pointer.type); const pointer_width = @bitSizeOf(usize); const alignment = 3; const pointer_type = llvm.debug_info_builder.createPointerType(element_type, pointer_width, alignment, name.ptr, name.len) orelse unreachable; break :b pointer_type.toType(); } }, .slice => |slice| b: { const pointer_type = try llvm.getDebugType(unit, context, slice.child_pointer_type); const len_type = try llvm.getDebugType(unit, context, Compilation.Type.usize); const scope = null; const file = null; const line = 1; const flags = LLVM.DebugInfo.Node.Flags{ .visibility = .none, .forward_declaration = false, .apple_block = false, .block_by_ref_struct = false, .virtual = false, .artificial = false, .explicit = false, .prototyped = false, .objective_c_class_complete = false, .object_pointer = false, .vector = false, .static_member = false, .lvalue_reference = false, .rvalue_reference = false, .reserved = false, .inheritance = .none, .introduced_virtual = false, .bit_field = false, .no_return = false, .type_pass_by_value = false, .type_pass_by_reference = false, .enum_class = false, .thunk = false, .non_trivial = false, .big_endian = false, .little_endian = false, .all_calls_described = false, }; const types = [2]*LLVM.DebugInfo.Type{ pointer_type, len_type }; const member_types = [2]*LLVM.DebugInfo.Type{ llvm.debug_info_builder.createMemberType(null, "pointer", "pointer".len, null, 0, 64, 3, 0, flags, types[0]).toType(), llvm.debug_info_builder.createMemberType(null, "length", "length".len, null, 0, 64, 3, 64, flags, types[1]).toType(), }; const struct_type = llvm.createDebugStructType(.{ .scope = scope, .name = name, .file = file, .line = line, .bitsize = 2 * @bitSizeOf(usize), .alignment = @alignOf(usize), .field_types = &member_types, .forward_declaration = null, }); break :b struct_type.toType(); }, .array => |array| b: { // TODO: compute const byte_size = 1; // array.count * unit.types.get(array.element_type).getSize(); const bit_size = byte_size * 8; const element_type = try llvm.getDebugType(unit, context, array.type); assert(array.count != 0); const array_type = llvm.debug_info_builder.createArrayType(bit_size, 1, element_type, array.count) orelse unreachable; break :b array_type.toType(); }, .function => |function_prototype_index| b: { const function_prototype = unit.function_prototypes.get(function_prototype_index); var parameter_types = try UnpinnedArray(*LLVM.DebugInfo.Type).initialize_with_capacity(context.my_allocator, @intCast(function_prototype.argument_types.len)); for (function_prototype.argument_types) |argument_type_index| { const argument_type = try llvm.getDebugType(unit, context, argument_type_index); parameter_types.append_with_capacity(argument_type); } const subroutine_type_flags = LLVM.DebugInfo.Node.Flags{ .visibility = .none, .forward_declaration = false, .apple_block = false, .block_by_ref_struct = false, .virtual = false, .artificial = false, .explicit = false, .prototyped = false, .objective_c_class_complete = false, .object_pointer = false, .vector = false, .static_member = false, .lvalue_reference = false, .rvalue_reference = false, .reserved = false, .inheritance = .none, .introduced_virtual = false, .bit_field = false, .no_return = false, .type_pass_by_value = false, .type_pass_by_reference = false, .enum_class = false, .thunk = false, .non_trivial = false, .big_endian = false, .little_endian = false, .all_calls_described = false, }; const subroutine_type_calling_convention = LLVM.DebugInfo.CallingConvention.none; const subroutine_type = llvm.debug_info_builder.createSubroutineType(parameter_types.pointer, parameter_types.length, subroutine_type_flags, subroutine_type_calling_convention) orelse unreachable; break :b subroutine_type.toType(); }, else => |t| @panic(@tagName(t)), }; try llvm.debug_type_map.put(context.my_allocator, sema_type_index, result); assert(@intFromPtr(result) != 0xaaaa_aaaa_aaaa_aaaa); return result; } } fn createGEP(llvm: *LLVM, unit: *Compilation.Unit, context: *const Compilation.Context, instruction_index: Compilation.Instruction.Index) !*LLVM.Value { const instruction = unit.instructions.get(instruction_index); const gep = instruction.get_element_pointer; const pointer = llvm.llvm_instruction_map.get(gep.pointer).?; const index = try llvm.emitRightValue(unit, context, gep.index); const struct_index = llvm.context.getConstantInt(@bitSizeOf(u32), 0, false) orelse unreachable; const index_buffer = [2]*LLVM.Value{ struct_index.toValue(), index }; const indices = index_buffer[@intFromBool(!gep.is_struct)..]; if (gep.is_struct) assert(indices.len == 2) else assert(indices.len == 1); const base_type = try llvm.getType(unit, context, gep.base_type); const in_bounds = true; if (gep.is_struct and gep.index.type != .u32) unreachable; const gep_name = unit.getIdentifier(gep.name); const get_element_pointer = llvm.builder.createGEP(base_type, pointer, indices.ptr, indices.len, gep_name.ptr, gep_name.len, in_bounds) orelse unreachable; try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, get_element_pointer); return get_element_pointer; } fn emitLeftValue(llvm: *LLVM, unit: *Compilation.Unit, context: *const Compilation.Context, v: Compilation.V) !*LLVM.Value { switch (v.value) { .runtime => |instruction_index| { if (llvm.llvm_instruction_map.get(instruction_index)) |value| { return value; } else { const instruction = unit.instructions.get(instruction_index); switch (instruction.*) { .get_element_pointer => { return try llvm.createGEP(unit, context, instruction_index); }, else => |t| @panic(@tagName(t)), } } }, .@"comptime" => |ct| switch (ct) { .global => |global| switch (global.initial_value) { .function_definition => return llvm.function_definition_map.get(global).?.toValue(), else => return llvm.global_variable_map.get(global).?.toValue(), }, else => |t| @panic(@tagName(t)), }, else => |t| @panic(@tagName(t)), } } fn emitComptimeRightValue(llvm: *LLVM, unit: *Compilation.Unit, context: *const Compilation.Context, ct: Compilation.V.Comptime, type_index: Compilation.Type.Index) anyerror!*LLVM.Value.Constant { switch (ct) { .constant_int => |integer| { const integer_type = unit.types.get(type_index); switch (integer_type.*) { .integer => |integer_t| switch (integer_t.kind) { .materialized_int, .bitfield => { const signed = switch (integer_t.signedness) { .signed => true, .unsigned => false, }; const constant_int = llvm.context.getConstantInt(integer_t.bit_count, integer.value, signed) orelse unreachable; return constant_int.toConstant(); }, else => |t| @panic(@tagName(t)), }, else => |t| @panic(@tagName(t)), } }, .comptime_int => |integer| { const integer_type = unit.types.get(type_index); switch (integer_type.*) { .integer => |integer_t| switch (integer_t.kind) { else => |t| @panic(@tagName(t)), .materialized_int => { const signed = switch (integer_t.signedness) { .signed => true, .unsigned => false, }; const constant_int = llvm.context.getConstantInt(integer_t.bit_count, integer.value, signed) orelse unreachable; return constant_int.toConstant(); }, }, else => |t| @panic(@tagName(t)), } }, .enum_value => |enum_field_index| { const enum_field = unit.enum_fields.get(enum_field_index); const integer = unit.types.get(enum_field.parent).integer; const signed = switch (integer.signedness) { .signed => true, .unsigned => false, }; const constant_int = llvm.context.getConstantInt(integer.bit_count, enum_field.value, signed) orelse unreachable; return constant_int.toConstant(); }, .constant_bitfield => |value| { const integer = unit.types.get(type_index).integer; const signed = switch (integer.signedness) { .signed => true, .unsigned => false, }; const constant_int = llvm.context.getConstantInt(integer.bit_count, value, signed) orelse unreachable; return constant_int.toConstant(); }, .constant_struct => |constant_struct_index| return try llvm.getConstantStruct(unit, context, constant_struct_index), .undefined => { const undefined_type = try llvm.getType(unit, context, type_index); const poison = undefined_type.getPoison() orelse unreachable; return poison.toConstant(); }, .bool => |boolean| { const bit_count = 1; const signed = false; const constant_bool = llvm.context.getConstantInt(bit_count, @intFromBool(boolean), signed) orelse unreachable; return constant_bool.toConstant(); }, .constant_slice => |constant_slice_index| { const constant_slice = try llvm.getConstantSlice(unit, context, constant_slice_index); return constant_slice; }, .constant_array => |constant_array_index| { const constant_array = try llvm.getConstantArray(unit, context, constant_array_index); return constant_array; }, .global => |global| { const constant = switch (global.initial_value) { .function_definition => llvm.function_definition_map.get(global).?.toConstant(), .function_declaration => llvm.llvm_external_functions.get(global).?.toConstant(), else => llvm.global_variable_map.get(global).?.toConstant(), }; return constant; }, .null_pointer => { const value_type = try llvm.getType(unit, context, type_index); const pointer_type = value_type.toPointer() orelse unreachable; const constant_null_pointer = pointer_type.getNull(); return constant_null_pointer.toConstant(); }, .error_value => |error_field_index| { const error_field = unit.error_fields.get(error_field_index); const error_type_index = error_field.type; const integer = unit.types.get(error_type_index).integer; const bit_count = integer.bit_count; const signed = switch (integer.signedness) { .unsigned => false, .signed => true, }; const constant_int = llvm.context.getConstantInt(bit_count, error_field.value, signed) orelse unreachable; return constant_int.toConstant(); }, else => |t| @panic(@tagName(t)), } } fn emitRightValue(llvm: *LLVM, unit: *Compilation.Unit, context: *const Compilation.Context, v: Compilation.V) !*LLVM.Value { switch (v.value) { .@"comptime" => |ct| { const constant_value = try llvm.emitComptimeRightValue(unit, context, ct, v.type); return constant_value.toValue(); }, .runtime => |instruction_index| { if (llvm.llvm_instruction_map.get(instruction_index)) |instruction| { return instruction; } else { unreachable; } }, else => |t| @panic(@tagName(t)), } } fn getScope(llvm: *LLVM, unit: *Compilation.Unit, context: *const Compilation.Context, sema_scope: *Compilation.Debug.Scope) anyerror!*LLVM.DebugInfo.Scope { switch (sema_scope.kind) { .function, .block, .compilation_unit, => { return llvm.scope_map.get(sema_scope).?; }, .file => { unreachable; }, .file_container, .struct_type => { if (llvm.scope_map.get(sema_scope)) |scope| { if (true) unreachable; return scope; } else { const global_scope: *Compilation.Debug.Scope.Global = @fieldParentPtr("scope", sema_scope); const struct_type = try llvm.getDebugType(unit, context, global_scope.type); return struct_type.toScope(); } }, else => |t| @panic(@tagName(t)), } unreachable; } fn createBasicBlock(llvm: *LLVM, context: *const Compilation.Context, basic_block_index: Compilation.BasicBlock.Index, name: []const u8) !*BasicBlockList.Node { const basic_block = llvm.context.createBasicBlock(name.ptr, name.len, llvm.function, null) orelse return Error.basic_block; const basic_block_node = try context.allocator.create(BasicBlockList.Node); basic_block_node.* = .{ .data = basic_block_index, }; try llvm.llvm_block_map.put_no_clobber(context.my_allocator, basic_block_index, basic_block); return basic_block_node; } fn getConstantSlice(llvm: *LLVM, unit: *Compilation.Unit, context: *const Compilation.Context, constant_slice_index: Compilation.V.Comptime.ConstantSlice.Index) !*LLVM.Value.Constant { const const_slice = unit.constant_slices.get(constant_slice_index); const const_slice_type = try llvm.getType(unit, context, const_slice.type); const slice_struct_type = const_slice_type.toStruct() orelse unreachable; const slice_fields: [2]*LLVM.Value.Constant = if (const_slice.array) |array| b: { const ptr = llvm.global_variable_map.get(array).?; const signed = false; assert(const_slice.start == 0); const len = llvm.context.getConstantInt(@bitSizeOf(usize), const_slice.end, signed) orelse unreachable; const slice_fields = [2]*LLVM.Value.Constant{ ptr.toConstant(), len.toConstant(), }; break :b slice_fields; } else b: { const ptr = llvm.pointer_type.?.getNull(); const len = llvm.context.getConstantInt(64, 0, false) orelse unreachable; break :b .{ ptr.toConstant(), len.toConstant() }; }; const constant_slice = slice_struct_type.getConstant(&slice_fields, slice_fields.len) orelse unreachable; return constant_slice; } fn getConstantArray(llvm: *LLVM, unit: *Compilation.Unit, context: *const Compilation.Context, constant_array_index: Compilation.V.Comptime.ConstantArray.Index) !*LLVM.Value.Constant { const constant_array = unit.constant_arrays.get(constant_array_index); const sema_array_type = unit.types.get(constant_array.type).array; const constant_type = try llvm.getType(unit, context, constant_array.type); const array_type = constant_type.toArray() orelse unreachable; var list = try UnpinnedArray(*LLVM.Value.Constant).initialize_with_capacity(context.my_allocator, @intCast(constant_array.values.len)); for (constant_array.values) |sema_value| { const value = switch (sema_value) { .constant_int => |const_int| b: { const integer_type = unit.types.get(sema_array_type.type).integer; const signed = switch (integer_type.signedness) { .signed => true, .unsigned => false, }; assert(!signed); const constant_int = llvm.context.getConstantInt(integer_type.bit_count, const_int.value, signed) orelse unreachable; break :b constant_int.toConstant(); }, .constant_slice => |constant_slice_index| try llvm.getConstantSlice(unit, context, constant_slice_index), .constant_struct => |constant_struct_index| try llvm.getConstantStruct(unit, context, constant_struct_index), .enum_value => |enum_field_index| b: { const integer_type = unit.types.get(sema_array_type.type).integer; const signed = switch (integer_type.signedness) { .signed => true, .unsigned => false, }; assert(!signed); const constant_int = llvm.context.getConstantInt(integer_type.bit_count, unit.enum_fields.get(enum_field_index).value, signed) orelse unreachable; break :b constant_int.toConstant(); }, else => |t| @panic(@tagName(t)), }; list.append_with_capacity(value); } const result = array_type.getConstant(list.pointer, list.length) orelse unreachable; return result; } fn getConstantStruct(llvm: *LLVM, unit: *Compilation.Unit, context: *const Compilation.Context, constant_struct_index: Compilation.V.Comptime.ConstantStruct.Index) !*LLVM.Value.Constant { const constant_struct = unit.constant_structs.get(constant_struct_index); var field_values = try UnpinnedArray(*LLVM.Value.Constant).initialize_with_capacity(context.my_allocator, @intCast(constant_struct.fields.len)); const sema_struct_index = unit.types.get(constant_struct.type).@"struct"; const sema_struct = unit.structs.get(sema_struct_index); const llvm_type = try llvm.getType(unit, context, constant_struct.type); const struct_type = llvm_type.toStruct() orelse unreachable; switch (sema_struct.kind) { .@"struct" => |*sema_struct_type| { for (constant_struct.fields, sema_struct_type.fields.slice()) |field_value, field_index| { const field = unit.struct_fields.get(field_index); const constant = try llvm.emitComptimeRightValue(unit, context, field_value, field.type); field_values.append_with_capacity(constant); } }, .error_union => |error_union| { const abi_ty = unit.types.get(error_union.abi); switch (abi_ty.*) { .@"struct" => |struct_index| switch (unit.structs.get(struct_index).kind) { .raw_error_union => |err_union_base_type| { const field_types = [2]Compilation.Type.Index{ err_union_base_type, .bool }; for (field_types, constant_struct.fields) |field_type_index, field_value| { const constant = try llvm.emitComptimeRightValue(unit, context, field_value, field_type_index); field_values.append_with_capacity(constant); } }, else => |t| @panic(@tagName(t)), }, else => |t| @panic(@tagName(t)), } }, else => |t| @panic(@tagName(t)), } const const_struct = struct_type.getConstant(field_values.pointer, field_values.length) orelse unreachable; return const_struct; } fn callIntrinsic(llvm: *LLVM, intrinsic_name: []const u8, intrinsic_parameter_types: []const *LLVM.Type, intrinsic_arguments: []const *LLVM.Value) !*LLVM.Value { const intrinsic_id = LLVM.lookupIntrinsic(intrinsic_name.ptr, intrinsic_name.len); assert(intrinsic_id != .none); const intrinsic_function = llvm.module.getIntrinsicDeclaration(intrinsic_id, intrinsic_parameter_types.ptr, intrinsic_parameter_types.len) orelse return LLVM.Value.Error.intrinsic; const intrinsic_type = intrinsic_function.getType(); const void_name: []const u8 = ""; const name = switch (intrinsic_type.getReturnType().isVoid()) { true => void_name, false => intrinsic_name, }; const call = llvm.builder.createCall(intrinsic_type, intrinsic_function.toValue(), intrinsic_arguments.ptr, intrinsic_arguments.len, name.ptr, name.len, null) orelse return LLVM.Value.Instruction.Error.call; return call.toValue(); } fn emitParameterAttributes(llvm: *LLVM, unit: *Compilation.Unit, context: *const Compilation.Context, abi: Compilation.Function.AbiInfo, is_return: bool) !*const LLVM.Attribute.Set { var attributes = UnpinnedArray(*LLVM.Attribute){}; if (abi.attributes.by_reg) { try attributes.append(context.my_allocator, llvm.attributes.inreg); } switch (abi.kind) { .ignore => { assert(is_return); }, .direct, .direct_pair, .direct_coerce => {}, .indirect => |indirect| { const indirect_type = try llvm.getType(unit, context, indirect.type); if (is_return) { const sret = llvm.context.getAttributeFromType(.StructRet, indirect_type); try attributes.append(context.my_allocator, sret); try attributes.append(context.my_allocator, llvm.attributes.@"noalias"); // TODO: alignment } else { if (abi.attributes.by_value) { const byval = llvm.context.getAttributeFromType(.ByVal, indirect_type); try attributes.append(context.my_allocator, byval); } //TODO: alignment } }, else => |t| @panic(@tagName(t)), } const attribute_set = llvm.context.getAttributeSet(attributes.pointer, attributes.length); return attribute_set; } fn getFunctionAttributes(llvm: *LLVM, unit: *Compilation.Unit, context: *const Compilation.Context, function_prototype: *Compilation.Function.Prototype) !*const LLVM.Attribute.Set { var function_attributes = UnpinnedArray(*LLVM.Attribute){}; try function_attributes.append(context.my_allocator, llvm.attributes.nounwind); switch (unit.types.get(function_prototype.return_type).*) { .noreturn => { try function_attributes.append(context.my_allocator, llvm.attributes.noreturn); }, else => {}, } if (function_prototype.attributes.naked) { try function_attributes.append(context.my_allocator, llvm.attributes.naked); } const function_attribute_set = llvm.context.getAttributeSet(function_attributes.pointer, function_attributes.length); return function_attribute_set; } const CallOrFunction = union(enum) { call: *LLVM.Value.Instruction.Call, function: *LLVM.Value.Constant.Function, }; fn setCallOrFunctionAttributes(llvm: *LLVM, unit: *Compilation.Unit, context: *const Compilation.Context, function_prototype: *Compilation.Function.Prototype, call_or_function: CallOrFunction) !void { const function_attribute_set = try llvm.getFunctionAttributes(unit, context, function_prototype); var parameter_attribute_sets = try UnpinnedArray(*const LLVM.Attribute.Set).initialize_with_capacity(context.my_allocator, @intCast(function_prototype.abi.parameter_types_abi.len + @intFromBool(function_prototype.abi.return_type_abi.kind == .indirect))); const return_attribute_set = blk: { const attribute_set = try llvm.emitParameterAttributes(unit, context, function_prototype.abi.return_type_abi, true); break :blk switch (function_prototype.abi.return_type_abi.kind) { .indirect => b: { parameter_attribute_sets.append_with_capacity(attribute_set); break :b llvm.context.getAttributeSet(null, 0); }, else => attribute_set, }; }; for (function_prototype.abi.parameter_types_abi) |parameter_type_abi| { const parameter_attribute_set = try llvm.emitParameterAttributes(unit, context, parameter_type_abi, false); parameter_attribute_sets.append_with_capacity(parameter_attribute_set); } const calling_convention = getCallingConvention(function_prototype.calling_convention); switch (call_or_function) { .call => |call| { call.setAttributes(llvm.context, function_attribute_set, return_attribute_set, parameter_attribute_sets.pointer, parameter_attribute_sets.length); call.setCallingConvention(calling_convention); }, .function => |function| { function.setAttributes(llvm.context, function_attribute_set, return_attribute_set, parameter_attribute_sets.pointer, parameter_attribute_sets.length); function.setCallingConvention(calling_convention); }, } } fn emitFunctionDeclaration(llvm: *LLVM, unit: *Compilation.Unit, context: *const Compilation.Context, declaration: *Compilation.Debug.Declaration.Global) !void { const name = unit.getIdentifier(declaration.declaration.name); const function_type = try llvm.getType(unit, context, declaration.declaration.type); const is_export = declaration.attributes.contains(.@"export"); const is_extern = declaration.attributes.contains(.@"extern"); const export_or_extern = is_export or is_extern; const linkage: LLVM.Linkage = switch (export_or_extern) { true => .@"extern", false => .internal, }; // TODO: Check name collision const mangle_name = !export_or_extern; _ = mangle_name; // autofix const function = llvm.module.createFunction(function_type.toFunction() orelse unreachable, linkage, address_space, name.ptr, name.len) orelse return Error.function; const function_prototype = unit.function_prototypes.get(unit.types.get(declaration.declaration.type).function); try llvm.setCallOrFunctionAttributes(unit, context, function_prototype, .{ .function = function, }); switch (declaration.initial_value) { .function_declaration => try llvm.function_declaration_map.put_no_clobber(context.my_allocator, declaration, function), .function_definition => try llvm.function_definition_map.put_no_clobber(context.my_allocator, declaration, function), else => unreachable, } const generate_debug_information = unit.descriptor.generate_debug_information and switch (declaration.initial_value) { .function_declaration => true, .function_definition => |function_definition_index| unit.function_definitions.get(function_definition_index).has_debug_info, else => unreachable, }; if (generate_debug_information) { // if (data_structures.byte_equal(name, "nat_split_struct_ints")) @breakpoint(); const debug_file = try llvm.getDebugInfoFile(unit, context, declaration.declaration.scope.file); var parameter_types = try UnpinnedArray(*LLVM.DebugInfo.Type).initialize_with_capacity(context.my_allocator, @intCast(function_prototype.argument_types.len)); for (function_prototype.argument_types) |argument_type_index| { const argument_type = try llvm.getDebugType(unit, context, argument_type_index); parameter_types.append_with_capacity(argument_type); } const subroutine_type_flags = LLVM.DebugInfo.Node.Flags{ .visibility = .none, .forward_declaration = is_extern, .apple_block = false, .block_by_ref_struct = false, .virtual = false, .artificial = false, .explicit = false, .prototyped = false, .objective_c_class_complete = false, .object_pointer = false, .vector = false, .static_member = false, .lvalue_reference = false, .rvalue_reference = false, .reserved = false, .inheritance = .none, .introduced_virtual = false, .bit_field = false, .no_return = false, .type_pass_by_value = false, .type_pass_by_reference = false, .enum_class = false, .thunk = false, .non_trivial = false, .big_endian = false, .little_endian = false, .all_calls_described = false, }; const subroutine_type_calling_convention = LLVM.DebugInfo.CallingConvention.none; const subroutine_type = llvm.debug_info_builder.createSubroutineType(parameter_types.pointer, parameter_types.length, subroutine_type_flags, subroutine_type_calling_convention) orelse unreachable; const subprogram_flags = LLVM.DebugInfo.Subprogram.Flags{ .virtuality = .none, .local_to_unit = !export_or_extern, .definition = !is_extern, .optimized = false, .pure = false, .elemental = false, .recursive = false, .main_subprogram = false, .deleted = false, .object_c_direct = false, }; const subprogram_declaration = null; const function_name = unit.getIdentifier(declaration.declaration.name); const subprogram = llvm.debug_info_builder.createFunction(debug_file.toScope(), function_name.ptr, function_name.len, function_name.ptr, function_name.len, debug_file, declaration.declaration.line + 1, subroutine_type, declaration.declaration.line + 1, subroutine_type_flags, subprogram_flags, subprogram_declaration) orelse unreachable; function.setSubprogram(subprogram); switch (declaration.initial_value) { .function_definition => |function_definition_index| { const function_definition = unit.function_definitions.get(function_definition_index); const scope = subprogram.toLocalScope().toScope(); try llvm.scope_map.put_no_clobber(context.my_allocator, &function_definition.scope.scope, scope); }, .function_declaration => {}, else => |t| @panic(@tagName(t)), } } switch (declaration.initial_value) { .function_declaration => try llvm.llvm_external_functions.put_no_clobber(context.my_allocator, declaration, function), .function_definition => {}, else => |t| @panic(@tagName(t)), } } }; fn getCallingConvention(calling_convention: Compilation.Function.CallingConvention) LLVM.Value.Constant.Function.CallingConvention { return switch (calling_convention) { .auto => .Fast, .c => .C, }; } const BasicBlockList = std.DoublyLinkedList(Compilation.BasicBlock.Index); const Error = error{ context, module, builder, function, basic_block, debug_info_builder, }; const address_space = 0; pub const Format = enum(c_uint) { elf = 0, macho = 1, coff = 2, }; pub fn codegen(unit: *Compilation.Unit, context: *const Compilation.Context) !void { const llvm_context = LLVM.Context.create() orelse return Error.context; const module = LLVM.Module.create(@ptrCast(unit.descriptor.name.ptr), unit.descriptor.name.len, llvm_context) orelse return Error.module; // TODO: const builder = LLVM.Builder.create(llvm_context) orelse return Error.builder; var llvm = LLVM{ .context = llvm_context, .module = module, .builder = builder, .debug_info_builder = if (unit.descriptor.generate_debug_information) module.createDebugInfoBuilder() orelse return Error.debug_info_builder else undefined, .attributes = .{ .naked = llvm_context.getAttributeFromEnum(.Naked, 0), .noreturn = llvm_context.getAttributeFromEnum(.NoReturn, 0), .nounwind = llvm_context.getAttributeFromEnum(.NoUnwind, 0), .inreg = llvm_context.getAttributeFromEnum(.InReg, 0), .@"noalias" = llvm_context.getAttributeFromEnum(.NoAlias, 0), }, }; if (unit.descriptor.generate_debug_information) { const full_path = try std.fs.cwd().realpathAlloc(context.allocator, unit.descriptor.main_package_path); const filename = std.fs.path.basename(full_path); const directory = full_path[0 .. full_path.len - filename.len]; const debug_info_file = llvm.debug_info_builder.createFile(filename.ptr, filename.len, directory.ptr, directory.len) orelse unreachable; const producer = "nativity"; const is_optimized = false; const flags = ""; const runtime_version = 0; const splitname = ""; const DWOId = 0; const debug_info_kind = LLVM.DebugInfo.CompileUnit.EmissionKind.full_debug; const split_debug_inlining = true; const debug_info_for_profiling = false; const name_table_kind = LLVM.DebugInfo.CompileUnit.NameTableKind.default; const ranges_base_address = false; const sysroot = ""; const sdk = ""; const compile_unit = llvm.debug_info_builder.createCompileUnit(LLVM.DebugInfo.Language.c, debug_info_file, producer, producer.len, is_optimized, flags, flags.len, runtime_version, splitname, splitname.len, debug_info_kind, DWOId, split_debug_inlining, debug_info_for_profiling, name_table_kind, ranges_base_address, sysroot, sysroot.len, sdk, sdk.len) orelse unreachable; llvm.scope = compile_unit.toScope(); try llvm.scope_map.put_no_clobber(context.my_allocator, &unit.scope.scope, llvm.scope); } for (unit.external_functions.values()) |external_function_declaration| { try llvm.emitFunctionDeclaration(unit, context, external_function_declaration); } const functions = unit.code_to_emit.values(); { var function_i: usize = functions.len; // Emit it in reverse order so the code goes the right order, from entry point to leaves while (function_i > 0) { function_i -= 1; const function_declaration = functions[function_i]; try llvm.emitFunctionDeclaration(unit, context, function_declaration); } } // First, cache all the global variables for (unit.data_to_emit.slice()) |global_declaration| { const name = unit.getIdentifier(global_declaration.declaration.name); switch (global_declaration.initial_value) { .string_literal => |hash| { const string_literal = unit.string_literal_values.get(hash).?; const global_variable = llvm.builder.createGlobalString(string_literal.ptr, string_literal.len, name.ptr, name.len, address_space, llvm.module) orelse unreachable; try llvm.global_variable_map.put_no_clobber(context.my_allocator, global_declaration, global_variable); }, else => { const global_type = try llvm.getType(unit, context, global_declaration.declaration.type); const linkage: LLVM.Linkage = switch (global_declaration.attributes.contains(.@"export")) { true => .@"extern", false => .internal, }; const constant = switch (global_declaration.declaration.mutability) { .@"var" => false, .@"const" => true, }; const initializer: ?*LLVM.Value.Constant = null; const thread_local_mode = LLVM.ThreadLocalMode.not_thread_local; const externally_initialized = false; const global_variable = llvm.module.addGlobalVariable(global_type, constant, linkage, initializer, name.ptr, name.len, null, thread_local_mode, address_space, externally_initialized) orelse return LLVM.Value.Error.constant_int; try llvm.global_variable_map.put_no_clobber(context.my_allocator, global_declaration, global_variable); }, } if (unit.descriptor.generate_debug_information) { // Don't emit debug information for strings if (@intFromEnum(global_declaration.declaration.scope.kind) < @intFromEnum(Compilation.Debug.Scope.Kind.function)) { const scope = try llvm.getScope(unit, context, global_declaration.declaration.scope); const file = try llvm.getDebugInfoFile(unit, context, global_declaration.declaration.scope.file); const debug_type = try llvm.getDebugType(unit, context, global_declaration.declaration.type); const is_local_to_unit = !global_declaration.attributes.contains(.@"export"); const is_defined = true; const expression = null; const declaration = null; const template_parameters = null; const alignment = 0; const debug_global_variable = llvm.debug_info_builder.createGlobalVariableExpression(scope, name.ptr, name.len, name.ptr, name.len, file, global_declaration.declaration.line, debug_type, is_local_to_unit, is_defined, expression, declaration, template_parameters, alignment) orelse unreachable; _ = debug_global_variable; // autofix } } } for (llvm.global_variable_map.keys(), llvm.global_variable_map.values()) |global_declaration, global_variable| { if (global_declaration.initial_value == .string_literal) continue; const constant_initializer = try llvm.emitComptimeRightValue(unit, context, global_declaration.initial_value, global_declaration.declaration.type); global_variable.setInitializer(constant_initializer); } for (llvm.function_definition_map.keys(), llvm.function_definition_map.values()) |function_declaration, function| { const function_definition_index = function_declaration.getFunctionDefinitionIndex(); const function_definition = unit.function_definitions.get(function_definition_index); llvm.function = function; llvm.sema_function = function_declaration; llvm.inside_branch = false; const generate_debug_information = unit.descriptor.generate_debug_information and function_definition.has_debug_info; if (generate_debug_information) { const subprogram = llvm.function.getSubprogram() orelse unreachable; llvm.file = subprogram.getFile() orelse unreachable; llvm.scope = subprogram.toLocalScope().toScope(); } var alloca_map = MyHashMap(Compilation.Instruction.Index, *LLVM.Value){}; var block_command_list = BasicBlockList{}; const entry_block_node = try llvm.createBasicBlock(context, function_definition.basic_blocks.pointer[0], "fn_entry"); block_command_list.append(entry_block_node); var phis = MyHashMap(Compilation.Instruction.Index, *LLVM.Value.Instruction.PhiNode){}; while (block_command_list.len != 0) { const block_node = block_command_list.first orelse unreachable; const basic_block_index = block_node.data; const sema_basic_block = unit.basic_blocks.get(basic_block_index); const basic_block = llvm.llvm_block_map.get(basic_block_index).?; llvm.builder.setInsertPoint(basic_block); var last_block = block_node; for (sema_basic_block.instructions.slice()) |instruction_index| { //if (@intFromEnum(instruction_index) == 474) @breakpoint(); const sema_instruction = unit.instructions.get(instruction_index); switch (sema_instruction.*) { .push_scope => |push_scope| { if (generate_debug_information) { const old_scope = try llvm.getScope(unit, context, push_scope.old); assert(@intFromEnum(push_scope.old.kind) >= @intFromEnum(Compilation.Debug.Scope.Kind.function)); const lexical_block = llvm.debug_info_builder.createLexicalBlock(old_scope, llvm.file, push_scope.new.line + 1, push_scope.new.column + 1) orelse unreachable; try llvm.scope_map.put_no_clobber(context.my_allocator, push_scope.new, lexical_block.toScope()); llvm.scope = lexical_block.toScope(); } }, .pop_scope => |pop_scope| { if (generate_debug_information) { const new = try llvm.getScope(unit, context, pop_scope.new); if (pop_scope.new.kind == .function) { assert(new.toSubprogram() orelse unreachable == llvm.function.getSubprogram() orelse unreachable); } llvm.scope = new; var scope = pop_scope.old; while (scope.kind != .function) { scope = scope.parent.?; } const subprogram_scope = try llvm.getScope(unit, context, scope); assert(llvm.function.getSubprogram() orelse unreachable == subprogram_scope.toSubprogram() orelse unreachable); } }, .debug_checkpoint => |debug_checkpoint| { if (generate_debug_information) { const scope = try llvm.getScope(unit, context, debug_checkpoint.scope); llvm.builder.setCurrentDebugLocation(llvm.context, debug_checkpoint.line + 1, debug_checkpoint.column + 1, scope, llvm.function); } }, .inline_assembly => |inline_assembly_index| { const assembly_block = unit.inline_assembly.get(inline_assembly_index); var assembly_statements = UnpinnedArray(u8){}; var constraints = UnpinnedArray(u8){}; var operand_values = UnpinnedArray(*LLVM.Value){}; var operand_types = UnpinnedArray(*LLVM.Type){}; switch (unit.descriptor.arch) { .x86_64 => { for (assembly_block.instructions) |assembly_instruction_index| { const instruction = unit.assembly_instructions.get(assembly_instruction_index); const instruction_id: Compilation.InlineAssembly.x86_64.Instruction = @enumFromInt(instruction.id); try assembly_statements.append_slice(context.my_allocator, switch (instruction_id) { .xor => "xorl", .mov => "movq", .@"and" => "andq", .call => "callq", }); try assembly_statements.append(context.my_allocator, ' '); if (instruction.operands.len > 0) { var reverse_operand_iterator = std.mem.reverseIterator(instruction.operands); while (reverse_operand_iterator.next()) |operand| { switch (operand) { .register => |register_value| { const register: Compilation.InlineAssembly.x86_64.Register = @enumFromInt(register_value); try assembly_statements.append(context.my_allocator, '%'); try assembly_statements.append_slice(context.my_allocator, @tagName(register)); }, .number_literal => |literal| { var buffer: [65]u8 = undefined; const number_literal = data_structures.format_int(&buffer, literal, 16, false); const slice_ptr = number_literal.ptr - 4; const literal_slice = slice_ptr[0 .. number_literal.len + 4]; literal_slice[0] = '$'; literal_slice[1] = '$'; literal_slice[2] = '0'; literal_slice[3] = 'x'; try assembly_statements.append_slice(context.my_allocator, try context.my_allocator.duplicate_bytes(literal_slice)); }, .value => |sema_value| { if (llvm.llvm_value_map.get(sema_value)) |v| { _ = v; // autofix unreachable; } else { const value = try llvm.emitLeftValue(unit, context, sema_value); var buffer: [65]u8 = undefined; const operand_number = data_structures.format_int(&buffer, operand_values.length, 16, false); const slice_ptr = operand_number.ptr - 2; const operand_slice = slice_ptr[0 .. operand_number.len + 2]; operand_slice[0] = '$'; operand_slice[1] = '{'; var new_buffer: [65]u8 = undefined; @memcpy(new_buffer[0..operand_slice.len], operand_slice); new_buffer[operand_slice.len] = ':'; new_buffer[operand_slice.len + 1] = 'P'; new_buffer[operand_slice.len + 2] = '}'; const new_slice = try context.my_allocator.duplicate_bytes(new_buffer[0 .. operand_slice.len + 3]); try assembly_statements.append_slice(context.my_allocator, new_slice); try operand_values.append(context.my_allocator, value); const value_type = value.getType(); try operand_types.append(context.my_allocator, value_type); try constraints.append(context.my_allocator, 'X'); } }, } try assembly_statements.append_slice(context.my_allocator, ", "); } _ = assembly_statements.pop(); _ = assembly_statements.pop(); } try assembly_statements.append_slice(context.my_allocator, "\n\t"); } // try constraints.append_slice(context.allocator, ",~{dirflag},~{fpsr},~{flags}"); }, else => |t| @panic(@tagName(t)), } const is_var_args = false; const function_type = LLVM.Context.getFunctionType(try llvm.getType(unit, context, Compilation.Type.Index.void), operand_types.pointer, operand_types.length, is_var_args) orelse unreachable; const has_side_effects = true; const is_align_stack = true; const dialect = LLVM.Value.InlineAssembly.Dialect.@"at&t"; const can_throw = false; const inline_assembly = LLVM.Value.InlineAssembly.get(function_type, assembly_statements.pointer, assembly_statements.length, constraints.pointer, constraints.length, has_side_effects, is_align_stack, dialect, can_throw) orelse return LLVM.Value.Error.inline_assembly; const call = llvm.builder.createCall(function_type, inline_assembly.toValue(), operand_values.pointer, operand_values.length, "", "".len, null) orelse return LLVM.Value.Instruction.Error.call; try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, call.toValue()); }, .stack_slot => |stack_slot| { // const stack_slot_type = unit.types.get(stack_slot.type); // const stack_slot_alignment = stack_slot_type.getAbiAlignment(unit); const declaration_type = try llvm.getType(unit, context, stack_slot.type); const type_alignment = unit.types.get(stack_slot.type).getAbiAlignment(unit); const alloca_array_size = null; const declaration_alloca = llvm.builder.createAlloca(declaration_type, address_space, alloca_array_size, "", "".len, type_alignment) orelse return LLVM.Value.Instruction.Error.alloca; try alloca_map.put_no_clobber(context.my_allocator, instruction_index, declaration_alloca.toValue()); try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, declaration_alloca.toValue()); }, .store => |store| { const right = try llvm.emitRightValue(unit, context, store.source); const source_type = unit.types.get(store.source.type); const alignment = source_type.getAbiAlignment(unit); const is_volatile = false; const left = try llvm.emitLeftValue(unit, context, store.destination); const store_instruction = llvm.builder.createStore(right, left, is_volatile, alignment) orelse return LLVM.Value.Instruction.Error.store; _ = store_instruction; // autofix }, .cast => |cast| { const value = try llvm.emitRightValue(unit, context, cast.value); const dest_type = try llvm.getType(unit, context, cast.type); switch (cast.id) { .int_to_pointer => { const cast_type = LLVM.Value.Instruction.Cast.Type.int_to_pointer; const cast_name = @tagName(cast_type); const cast_instruction = llvm.builder.createCast(cast_type, value, value.getType(), cast_name.ptr, cast_name.len) orelse return LLVM.Value.Instruction.Error.cast; try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, cast_instruction); }, // TODO: Poke metadata .pointer_var_to_const, .slice_var_to_const, .enum_to_int, .slice_to_nullable, .slice_to_not_null, .slice_coerce_to_zero_termination, .slice_zero_to_no_termination, .pointer_to_nullable, .pointer_to_not_nullable, .pointer_const_to_var, .pointer_to_array_to_pointer_to_many, .pointer_source_type_to_destination_type, .pointer_none_terminated_to_zero, => { try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, value); }, .sign_extend => { const sign_extend = llvm.builder.createCast(.sign_extend, value, dest_type, "sign_extend", "sign_extend".len) orelse return LLVM.Value.Instruction.Error.cast; try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, sign_extend); }, .zero_extend => { const zero_extend = llvm.builder.createCast(.zero_extend, value, dest_type, "zero_extend", "zero_extend".len) orelse return LLVM.Value.Instruction.Error.cast; try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, zero_extend); }, .bitcast => { const bitcast = llvm.builder.createCast(.bitcast, value, dest_type, "bitcast", "bitcast".len) orelse return LLVM.Value.Instruction.Error.cast; try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, bitcast); }, .pointer_to_int => { const pointer_to_int = llvm.builder.createCast(.pointer_to_int, value, dest_type, "pointer_to_int", "pointer_to_int".len) orelse return LLVM.Value.Instruction.Error.cast; try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, pointer_to_int); }, .truncate => { const truncate = llvm.builder.createCast(.truncate, value, dest_type, "truncate", "truncate".len) orelse return LLVM.Value.Instruction.Error.cast; try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, truncate); }, .error_union_type_int_to_pointer, .error_union_type_upcast, .error_union_type_downcast => unreachable, } }, .load => |load| { const value = if (llvm.llvm_value_map.get(load.value)) |v| v else blk: { const value = switch (load.value.value) { .runtime => |instr_index| llvm.llvm_instruction_map.get(instr_index) orelse switch (unit.instructions.get(instr_index).*) { else => |t| @panic(@tagName(t)), }, .@"comptime" => |ct| switch (ct) { .global => |global| llvm.global_variable_map.get(global).?.toValue(), else => |t| @panic(@tagName(t)), }, else => |t| @panic(@tagName(t)), }; try llvm.llvm_value_map.put_no_clobber(context.my_allocator, load.value, value); break :blk value; }; const load_type = unit.types.get(load.type); const alignment = if (load.alignment) |alignment| alignment else load_type.getAbiAlignment(unit); const value_type = try llvm.getType(unit, context, load.type); const is_volatile = false; const load_i = llvm.builder.createLoad(value_type, value, is_volatile, "", "".len, alignment) orelse return LLVM.Value.Instruction.Error.load; try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, load_i.toValue()); }, .integer_binary_operation => |binary_operation| { assert(binary_operation.left.type == binary_operation.right.type); const left = try llvm.emitRightValue(unit, context, binary_operation.left); const right = try llvm.emitRightValue(unit, context, binary_operation.right); assert(left.getType() == right.getType()); const no_signed_wrapping = binary_operation.signedness == .signed or switch (binary_operation.id) { .wrapping_add, .wrapping_sub, .wrapping_mul => false, else => true, }; const no_unsigned_wrapping = binary_operation.signedness == .unsigned or switch (binary_operation.id) { .wrapping_add, .wrapping_sub, .wrapping_mul => false, else => true, }; const name = @tagName(binary_operation.id); const is_exact = false; const instruction = switch (binary_operation.id) { .add, .wrapping_add => llvm.builder.createAdd(left, right, name.ptr, name.len, no_unsigned_wrapping, no_signed_wrapping) orelse return LLVM.Value.Instruction.Error.add, .sub, .wrapping_sub => llvm.builder.createSub(left, right, name.ptr, name.len, no_unsigned_wrapping, no_signed_wrapping) orelse return LLVM.Value.Instruction.Error.add, .mul, .wrapping_mul => llvm.builder.createMultiply(left, right, name.ptr, name.len, no_unsigned_wrapping, no_signed_wrapping) orelse return LLVM.Value.Instruction.Error.multiply, .div => switch (binary_operation.signedness) { .unsigned => llvm.builder.createUDiv(left, right, name.ptr, name.len, is_exact) orelse unreachable, .signed => llvm.builder.createSDiv(left, right, name.ptr, name.len, is_exact) orelse unreachable, }, .mod => switch (binary_operation.signedness) { .unsigned => llvm.builder.createURem(left, right, name.ptr, name.len) orelse unreachable, .signed => llvm.builder.createSRem(left, right, name.ptr, name.len) orelse unreachable, }, .bit_and => llvm.builder.createAnd(left, right, name.ptr, name.len) orelse unreachable, .bit_or => llvm.builder.createOr(left, right, name.ptr, name.len) orelse unreachable, .bit_xor => llvm.builder.createXor(left, right, name.ptr, name.len) orelse unreachable, .shift_left => llvm.builder.createShiftLeft(left, right, name.ptr, name.len, no_unsigned_wrapping, no_signed_wrapping) orelse unreachable, .shift_right => switch (binary_operation.signedness) { .unsigned => llvm.builder.createLogicalShiftRight(left, right, name.ptr, name.len, is_exact) orelse unreachable, .signed => llvm.builder.createArithmeticShiftRight(left, right, name.ptr, name.len, is_exact) orelse unreachable, }, else => unreachable, }; try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, instruction); }, .call => |sema_call| { var argument_buffer: [32]*LLVM.Value = undefined; const argument_count = sema_call.arguments.len; const arguments = argument_buffer[0..argument_count]; const call_function_type = unit.types.get(sema_call.function_type); const function_prototype = unit.function_prototypes.get(call_function_type.function); const call_type = try llvm.getType(unit, context, sema_call.function_type); const call = switch (sema_call.callable.value) { .@"comptime" => |ct| switch (ct) { .global => |call_function_declaration| b: { const callee = switch (call_function_declaration.initial_value) { .function_definition => llvm.function_definition_map.get(call_function_declaration).?, .function_declaration => llvm.function_declaration_map.get(call_function_declaration).?, else => |t| @panic(@tagName(t)), }; for (sema_call.arguments, arguments) |argument_value, *argument| { argument.* = try llvm.emitRightValue(unit, context, argument_value); } const name = ""; const function_name = unit.getIdentifier(call_function_declaration.declaration.name); _ = function_name; // autofix const function_type = call_type.toFunction() orelse unreachable; for (sema_call.arguments, arguments, function_prototype.abi.parameter_types, 0..) |sema_argument, argument, sema_argument_type, i| { assert(sema_argument.type == sema_argument_type); const argument_type = function_type.getArgumentType(@intCast(i)); argument_type.assertEqual(argument.getType()); } const call_instruction = llvm.builder.createCall(function_type, callee.toValue(), arguments.ptr, arguments.len, name.ptr, name.len, null) orelse return LLVM.Value.Instruction.Error.call; break :b call_instruction; }, else => |t| @panic(@tagName(t)), }, .runtime => |ii| b: { const callee = llvm.llvm_instruction_map.get(ii).?; for (sema_call.arguments, arguments) |argument_value, *argument| { argument.* = try llvm.emitRightValue(unit, context, argument_value); } const name = ""; const function_type = call_type.toFunction() orelse unreachable; const call_instruction = llvm.builder.createCall(function_type, callee, arguments.ptr, arguments.len, name.ptr, name.len, null) orelse return LLVM.Value.Instruction.Error.call; break :b call_instruction; }, else => |t| @panic(@tagName(t)), }; try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, call.toValue()); try llvm.setCallOrFunctionAttributes(unit, context, function_prototype, .{ .call = call, }); }, .ret => |return_value| { const value = switch (return_value.type) { .void => null, else => try llvm.emitRightValue(unit, context, return_value), }; const ret = llvm.builder.createRet(value) orelse return LLVM.Value.Instruction.Error.ret; _ = ret; // autofix }, .syscall => |syscall| { var syscall_argument_buffer: [7]*LLVM.Value = undefined; var syscall_argument_type_buffer: [7]*LLVM.Type = undefined; const sema_syscall_arguments = syscall.arguments; const syscall_argument_count: usize = sema_syscall_arguments.len; const syscall_arguments = syscall_argument_buffer[0..syscall_argument_count]; const syscall_argument_types = syscall_argument_type_buffer[0..syscall_argument_count]; for (sema_syscall_arguments, syscall_arguments, syscall_argument_types) |sema_syscall_argument_value_index, *syscall_argument, *syscall_argument_type| { syscall_argument.* = try llvm.emitRightValue(unit, context, sema_syscall_argument_value_index); syscall_argument_type.* = syscall_argument.*.getType(); } const return_type = try llvm.getType(unit, context, Compilation.Type.usize); const is_var_args = false; const function_type = LLVM.Context.getFunctionType(return_type, syscall_argument_types.ptr, syscall_argument_types.len, is_var_args) orelse unreachable; var constraints = UnpinnedArray(u8){}; const inline_asm = switch (unit.descriptor.arch) { .x86_64 => blk: { try constraints.append_slice(context.my_allocator, "={rax}"); const syscall_registers = [7][]const u8{ "rax", "rdi", "rsi", "rdx", "r10", "r8", "r9" }; for (syscall_registers[0..syscall_argument_count]) |syscall_register| { try constraints.append(context.my_allocator, ','); try constraints.append(context.my_allocator, '{'); try constraints.append_slice(context.my_allocator, syscall_register); try constraints.append(context.my_allocator, '}'); } try constraints.append_slice(context.my_allocator, ",~{rcx},~{r11},~{memory}"); const assembly = "syscall"; const has_side_effects = true; const is_align_stack = true; const can_throw = false; const inline_assembly = LLVM.Value.InlineAssembly.get(function_type, assembly, assembly.len, constraints.pointer, constraints.length, has_side_effects, is_align_stack, LLVM.Value.InlineAssembly.Dialect.@"at&t", can_throw) orelse return LLVM.Value.Error.inline_assembly; break :blk inline_assembly; }, else => |t| @panic(@tagName(t)), }; const call_to_asm = llvm.builder.createCall(function_type, inline_asm.toValue(), syscall_arguments.ptr, syscall_arguments.len, "syscall", "syscall".len, null) orelse return LLVM.Value.Instruction.Error.call; try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, call_to_asm.toValue()); }, .@"unreachable" => { _ = llvm.builder.createUnreachable() orelse return LLVM.Value.Instruction.Error.@"unreachable"; }, .abi_argument => |argument_index| { const argument = llvm.function.getArgument(argument_index) orelse unreachable; try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, argument.toValue()); }, .debug_declare_argument => |debug_declare| { if (generate_debug_information) { const argument_index: c_uint = debug_declare.argument.index; const declaration = &debug_declare.argument.declaration; const debug_declaration_type = try llvm.getDebugType(unit, context, declaration.type); const declaration_alloca = llvm.llvm_instruction_map.get(debug_declare.stack).?; const always_preserve = true; const flags = LLVM.DebugInfo.Node.Flags{ .visibility = .none, .forward_declaration = false, .apple_block = false, .block_by_ref_struct = false, .virtual = false, .artificial = false, .explicit = false, .prototyped = false, .objective_c_class_complete = false, .object_pointer = false, .vector = false, .static_member = false, .lvalue_reference = false, .rvalue_reference = false, .reserved = false, .inheritance = .none, .introduced_virtual = false, .bit_field = false, .no_return = false, .type_pass_by_value = false, .type_pass_by_reference = false, .enum_class = false, .thunk = false, .non_trivial = false, .big_endian = false, .little_endian = false, .all_calls_described = false, }; const declaration_name = unit.getIdentifier(declaration.name); const line = declaration.line; const column = declaration.column; const debug_parameter_variable = llvm.debug_info_builder.createParameterVariable(llvm.scope, declaration_name.ptr, declaration_name.len, argument_index + 1, llvm.file, line, debug_declaration_type, always_preserve, flags) orelse unreachable; const insert_declare = llvm.debug_info_builder.insertDeclare(declaration_alloca, debug_parameter_variable, llvm.context, line, column, (llvm.function.getSubprogram() orelse unreachable).toLocalScope().toScope(), llvm.builder.getInsertBlock() orelse unreachable); _ = insert_declare; } }, .debug_declare_local_variable => |declare_local_variable| { if (generate_debug_information) { const local_variable = declare_local_variable.variable; const debug_declaration_type = try llvm.getDebugType(unit, context, local_variable.declaration.type); const always_preserve = true; const flags = LLVM.DebugInfo.Node.Flags{ .visibility = .none, .forward_declaration = false, .apple_block = false, .block_by_ref_struct = false, .virtual = false, .artificial = false, .explicit = false, .prototyped = false, .objective_c_class_complete = false, .object_pointer = false, .vector = false, .static_member = false, .lvalue_reference = false, .rvalue_reference = false, .reserved = false, .inheritance = .none, .introduced_virtual = false, .bit_field = false, .no_return = false, .type_pass_by_value = false, .type_pass_by_reference = false, .enum_class = false, .thunk = false, .non_trivial = false, .big_endian = false, .little_endian = false, .all_calls_described = false, }; const alignment = 0; const declaration_name = unit.getIdentifier(local_variable.declaration.name); const line = local_variable.declaration.line; const column = local_variable.declaration.column; const debug_local_variable = llvm.debug_info_builder.createAutoVariable(llvm.scope, declaration_name.ptr, declaration_name.len, llvm.file, line, debug_declaration_type, always_preserve, flags, alignment) orelse unreachable; const local = alloca_map.get(declare_local_variable.stack).?; const insert_declare = llvm.debug_info_builder.insertDeclare(local, debug_local_variable, llvm.context, line, column, (llvm.function.getSubprogram() orelse unreachable).toLocalScope().toScope(), llvm.builder.getInsertBlock() orelse unreachable); _ = insert_declare; } }, .insert_value => |insert_value| { const aggregate = try llvm.emitRightValue(unit, context, insert_value.expression); const value = try llvm.emitRightValue(unit, context, insert_value.new_value); const indices = [1]c_uint{insert_value.index}; const instruction = llvm.builder.createInsertValue(aggregate, value, &indices, indices.len, "", "".len) orelse unreachable; try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, instruction); }, .extract_value => |extract_value| { switch (unit.types.get(extract_value.expression.type).*) { .pointer => unreachable, else => {}, } const aggregate = try llvm.emitRightValue(unit, context, extract_value.expression); const aggregate_type = try llvm.getType(unit, context, extract_value.expression.type); assert(aggregate_type == aggregate.getType()); assert(!aggregate.getType().isPointer()); const indices = [1]c_uint{extract_value.index}; const instruction = llvm.builder.createExtractValue(aggregate, &indices, indices.len, "", "".len) orelse unreachable; try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, instruction); }, .integer_compare => |integer_compare| { assert(integer_compare.left.type == integer_compare.right.type); const left = try llvm.emitRightValue(unit, context, integer_compare.left); const right = try llvm.emitRightValue(unit, context, integer_compare.right); assert(left.getType() == right.getType()); const comparison_id: LLVM.Value.Instruction.ICmp.Kind = switch (integer_compare.id) { .equal => .eq, .not_equal => .ne, .unsigned_less => .ult, .unsigned_less_equal => .ule, .unsigned_greater => .ugt, .unsigned_greater_equal => .uge, .signed_less => .slt, .signed_less_equal => .sle, .signed_greater => .sgt, .signed_greater_equal => .sge, }; const icmp = llvm.builder.createICmp(comparison_id, left, right, "", "".len) orelse unreachable; try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, icmp); }, .jump => |jump| { const target_block = if (llvm.llvm_block_map.get(jump.to)) |target_block| target_block else blk: { const jump_target_block_node = try llvm.createBasicBlock(context, jump.to, "jmp_target"); block_command_list.insertAfter(last_block, jump_target_block_node); last_block = jump_target_block_node; // TODO: make this efficient break :blk llvm.llvm_block_map.get(jump_target_block_node.data).?; }; const br = llvm.builder.createBranch(target_block) orelse unreachable; try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, br.toValue()); }, .branch => |branch| { const taken_node = try llvm.createBasicBlock(context, branch.taken, "taken_block"); const not_taken_node = try llvm.createBasicBlock(context, branch.not_taken, "not_taken_block"); block_command_list.insertAfter(last_block, taken_node); block_command_list.insertAfter(taken_node, not_taken_node); last_block = not_taken_node; // TODO: make this fast const taken_block = llvm.llvm_block_map.get(taken_node.data).?; const not_taken_block = llvm.llvm_block_map.get(not_taken_node.data).?; const condition = llvm.llvm_instruction_map.get(branch.condition).?; const branch_weights = null; const unpredictable = null; const br = llvm.builder.createConditionalBranch(condition, taken_block, not_taken_block, branch_weights, unpredictable) orelse unreachable; try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, br.toValue()); }, .phi => |phi| { const phi_type = try llvm.getType(unit, context, phi.type); const reserved_value_count: c_uint = @intCast(phi.values.length); const phi_name = "phi"; const phi_node = llvm.builder.createPhi(phi_type, reserved_value_count, phi_name, phi_name.len) orelse unreachable; try phis.put_no_clobber(context.my_allocator, instruction_index, phi_node); try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, phi_node.toValue()); }, .umin => |umin| { const intrinsic_type = try llvm.getType(unit, context, umin.type); const parameter_types = [_]*LLVM.Type{intrinsic_type}; const left = try llvm.emitRightValue(unit, context, umin.left); const right = try llvm.emitRightValue(unit, context, umin.right); const arguments = [_]*LLVM.Value{ left, right }; const intrinsic_call = try llvm.callIntrinsic("llvm.umin", ¶meter_types, &arguments); try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, intrinsic_call); }, .get_element_pointer => { _ = try llvm.createGEP(unit, context, instruction_index); }, .trap => { const parameter_types: []const *LLVM.Type = &.{}; const parameter_values: []const *LLVM.Value = &.{}; const intrinsic_call = try llvm.callIntrinsic("llvm.trap", parameter_types, parameter_values); try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, intrinsic_call); }, .add_overflow => |add_overflow| { const intrinsic_type = try llvm.getType(unit, context, add_overflow.type); const parameter_types = [_]*LLVM.Type{intrinsic_type}; const left = try llvm.emitRightValue(unit, context, add_overflow.left); const right = try llvm.emitRightValue(unit, context, add_overflow.right); const arguments = [_]*LLVM.Value{ left, right }; const intrinsic_call = try llvm.callIntrinsic("llvm.sadd.with.overflow", ¶meter_types, &arguments); try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, intrinsic_call); }, .trailing_zeroes => |v| { const intrinsic_type = try llvm.getType(unit, context, v.type); const parameter_types = [_]*LLVM.Type{intrinsic_type}; const value = try llvm.emitRightValue(unit, context, v); const is_poison = llvm.context.getConstantInt(1, 0, false) orelse unreachable; const arguments = [_]*LLVM.Value{ value, is_poison.toValue() }; const intrinsic_call = try llvm.callIntrinsic("llvm.cttz", ¶meter_types, &arguments); try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, intrinsic_call); }, .@"switch" => |switch_expression| { const condition = try llvm.emitRightValue(unit, context, switch_expression.condition); const else_block: ?*LLVM.Value.BasicBlock = if (switch_expression.else_block != .null) b: { const else_block_node = try llvm.createBasicBlock(context, switch_expression.else_block, "switch_else"); const bb = llvm.llvm_block_map.get(switch_expression.else_block).?; block_command_list.insertAfter(last_block, else_block_node); last_block = else_block_node; break :b bb; } else null; var basic_block_array = try UnpinnedArray(*LLVM.Value.BasicBlock).initialize_with_capacity(context.my_allocator, switch_expression.cases.length); var condition_array = try UnpinnedArray(*LLVM.Value.Constant.Int).initialize_with_capacity(context.my_allocator, switch_expression.cases.length); for (switch_expression.cases.pointer[0..switch_expression.cases.length]) |case| { const constant_value = try llvm.emitComptimeRightValue(unit, context, case.condition, switch_expression.condition.type); const constant_int = constant_value.toInt() orelse unreachable; const block = if (llvm.llvm_block_map.get(case.basic_block)) |bb| bb else b: { const switch_block_node = try llvm.createBasicBlock(context, case.basic_block, "case_block"); block_command_list.insertAfter(last_block, switch_block_node); last_block = switch_block_node; const block = llvm.llvm_block_map.get(case.basic_block).?; break :b block; }; condition_array.append_with_capacity(constant_int); basic_block_array.append_with_capacity(block); } const branch_weights = null; const unpredictable = null; const switch_instruction = llvm.builder.createSwitch(condition, else_block, condition_array.pointer, basic_block_array.pointer, condition_array.length, branch_weights, unpredictable); try llvm.llvm_instruction_map.put_no_clobber(context.my_allocator, instruction_index, switch_instruction.toValue()); }, .memcpy => |memcpy| { const destination = try llvm.emitLeftValue(unit, context, memcpy.destination); const source = try llvm.emitLeftValue(unit, context, memcpy.source); const destination_alignment = if (memcpy.destination_alignment) |alignment| alignment else b: { const ty = unit.types.get(memcpy.destination.type); const alignment = ty.getAbiAlignment(unit); break :b alignment; }; const source_alignment = if (memcpy.source_alignment) |alignment| alignment else b: { const ty = unit.types.get(memcpy.source.type); const alignment = ty.getAbiAlignment(unit); break :b alignment; }; _ = llvm.builder.createMemcpy(destination, destination_alignment, source, source_alignment, memcpy.size, memcpy.is_volatile); }, else => |t| @panic(@tagName(t)), } } _ = block_command_list.popFirst(); } for (phis.keys(), phis.values()) |instruction_index, phi| { const instruction = unit.instructions.get(instruction_index); const sema_phi = &instruction.phi; for (sema_phi.values.slice(), sema_phi.basic_blocks.slice()) |sema_value, sema_block| { assert(sema_value.type == sema_phi.type); const value_basic_block = llvm.llvm_block_map.get(sema_block).?; const value = llvm.llvm_value_map.get(sema_value) orelse try llvm.emitRightValue(unit, context, sema_value); phi.addIncoming(value, value_basic_block); } } if (!builder.isCurrentBlockTerminated()) { var message_len: usize = 0; const function_str = llvm.function.toString(&message_len); const function_dump = function_str[0..message_len]; try write(.panic, function_dump); @panic("Function block with no termination"); } if (generate_debug_information) { llvm.debug_info_builder.finalizeSubprogram(llvm.function.getSubprogram() orelse unreachable, llvm.function); } const verify_function = true; if (verify_function) { var function_len: usize = 0; const function_ptr = llvm.function.toString(&function_len); const function_ir = function_ptr[0..function_len]; var message_ptr: [*]const u8 = undefined; var message_len: usize = 0; const result = llvm.function.verify(&message_ptr, &message_len); if (!result) { // std.debug.print("PANIC: Failed to verify function:\n{s}\n", .{error_message}); var module_len: usize = 0; const module_ptr = llvm.module.toString(&module_len); const module_dump = module_ptr[0..module_len]; _ = module_dump; // autofix try write(.panic, function_ir); const error_message = message_ptr[0..message_len]; try write(.panic, error_message); // std.debug.print("\nLLVM verification for function inside module failed:\nFull module: {s}\n```\n{s}\n```\n{s}\n", .{ module_dump, function_ir, error_message }); @panic("LLVM function verification failed"); } } } llvm.debug_info_builder.finalize(); var module_len: usize = 0; const module_ptr = llvm.module.toString(&module_len); const module_string = module_ptr[0..module_len]; // logln(.llvm, .print_module, "{s}", .{module_string}); const verify_module = true; const print_module = true; if (verify_module) { var message_ptr: [*]const u8 = undefined; var message_len: usize = 0; const result = llvm.module.verify(&message_ptr, &message_len); if (!result) { try write(.llvm, "Module: \n"); try write(.llvm, module_string); try write(.llvm, "\n"); try write(.llvm, message_ptr[0..message_len]); @panic("\nLLVM module verification failed"); } } if (print_module) { try write(.llvm, "Module: \n"); try write(.llvm, module_string); try write(.llvm, "\n"); } switch (unit.descriptor.arch) { inline else => |a| { const arch = @field(LLVM, @tagName(a)); arch.initializeTarget(); arch.initializeTargetInfo(); arch.initializeTargetMC(); arch.initializeAsmPrinter(); arch.initializeAsmParser(); }, } // TODO: proper target selection const target_triple = switch (unit.descriptor.os) { .linux => "x86_64-linux-none", .macos => "aarch64-apple-macosx-none", .windows => "x86_64-windows-gnu", }; const cpu = "generic"; const temp_use_native_features = true; const features = if (temp_use_native_features) blk: { var buffer = UnpinnedArray(u8){}; for (@import("builtin").cpu.arch.allFeaturesList(), 0..) |feature, index_usize| { const index = @as(std.Target.Cpu.Feature.Set.Index, @intCast(index_usize)); const is_enabled = @import("builtin").cpu.features.isEnabled(index); if (feature.llvm_name) |llvm_name| { const plus_or_minus = "-+"[@intFromBool(is_enabled)]; try buffer.append(context.my_allocator, plus_or_minus); try buffer.append_slice(context.my_allocator, llvm_name); try buffer.append_slice(context.my_allocator, ","); } } if (buffer.length == 0) break :blk ""; assert(std.mem.endsWith(u8, buffer.slice(), ",")); buffer.slice()[buffer.length - 1] = 0; break :blk buffer.slice()[0 .. buffer.length - 1 :0]; } else ""; const target = blk: { var error_message: [*]const u8 = undefined; var error_message_len: usize = 0; const target = bindings.NativityLLVMGetTarget(target_triple.ptr, target_triple.len, &error_message, &error_message_len) orelse unreachable; break :blk target; }; const jit = false; const code_model: LLVM.CodeModel = undefined; const is_code_model_present = false; const codegen_optimization_level: LLVM.CodegenOptimizationLevel = switch (unit.descriptor.optimization) { .none => .none, .debug_prefer_fast, .debug_prefer_size => .none, .lightly_optimize_for_speed => .less, .optimize_for_speed, .optimize_for_size => .default, .aggressively_optimize_for_speed, .aggressively_optimize_for_size => .aggressive, }; const target_machine = target.createTargetMachine(target_triple.ptr, target_triple.len, cpu, cpu.len, features.ptr, features.len, LLVM.RelocationModel.static, code_model, is_code_model_present, codegen_optimization_level, jit) orelse unreachable; llvm.module.setTargetMachineDataLayout(target_machine); llvm.module.setTargetTriple(target_triple.ptr, target_triple.len); const file_path = unit.descriptor.executable_path; const object_file_path = blk: { const slice = try context.allocator.alloc(u8, file_path.len + 3); @memcpy(slice[0..file_path.len], file_path); slice[file_path.len] = '.'; slice[file_path.len + 1] = 'o'; slice[file_path.len + 2] = 0; const object_file_path = slice[0 .. slice.len - 1 :0]; break :blk object_file_path; }; if (unit.descriptor.optimization != .none) { const optimization_level: LLVM.OptimizationLevel = switch (unit.descriptor.optimization) { .none => unreachable, .debug_prefer_fast, .debug_prefer_size => .{ .speed_level = 0, .size_level = 0 }, // -O0 .lightly_optimize_for_speed => .{ .speed_level = 1, .size_level = 0 }, // -O1 .optimize_for_speed => .{ .speed_level = 2, .size_level = 0 }, // -O2 .optimize_for_size => .{ .speed_level = 2, .size_level = 1 }, // -Os .aggressively_optimize_for_speed => .{ .speed_level = 3, .size_level = 0 }, // -O3 .aggressively_optimize_for_size => .{ .speed_level = 2, .size_level = 2 }, // -Oz }; llvm.module.runOptimizationPipeline(target_machine, optimization_level); } const disable_verify = false; const result = llvm.module.addPassesToEmitFile(target_machine, object_file_path.ptr, object_file_path.len, LLVM.CodeGenFileType.object, disable_verify); if (!result) { @panic("can't generate machine code"); } }