zig

fork of https://codeberg.org/ziglang/zig
Log | Files | Refs | README | LICENSE

commit f2feb4e47aa7d74f26f5bda1f8383ccd0f54026a (tree)
parent 64f4ef75566ef34289e9e6a455b0173e4e58df47
Author: Andrew Kelley <andrew@ziglang.org>
Date:   Fri, 15 May 2020 21:44:33 -0400

move Module to its own file

Diffstat:
Asrc-self-hosted/Module.zig | 2018+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Msrc-self-hosted/codegen.zig | 18++++++++++--------
Msrc-self-hosted/ir.zig | 2018+------------------------------------------------------------------------------
Dsrc-self-hosted/ir/text.zig | 1476-------------------------------------------------------------------------------
Msrc-self-hosted/link.zig | 21+++++++++++----------
Msrc-self-hosted/main.zig | 25+++++++++++++------------
Msrc-self-hosted/value.zig | 6+++---
Asrc-self-hosted/zir.zig | 1477+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
8 files changed, 3534 insertions(+), 3525 deletions(-)

diff --git a/src-self-hosted/Module.zig b/src-self-hosted/Module.zig @@ -0,0 +1,2018 @@ +const std = @import("std"); +const mem = std.mem; +const Allocator = std.mem.Allocator; +const ArrayListUnmanaged = std.ArrayListUnmanaged; +const Value = @import("value.zig").Value; +const Type = @import("type.zig").Type; +const TypedValue = @import("TypedValue.zig"); +const assert = std.debug.assert; +const BigIntConst = std.math.big.int.Const; +const BigIntMutable = std.math.big.int.Mutable; +const Target = std.Target; +const Package = @import("Package.zig"); +const link = @import("link.zig"); +const ir = @import("ir.zig"); +const zir = @import("zir.zig"); +const Module = @This(); +const Inst = ir.Inst; + +/// General-purpose allocator. +allocator: *Allocator, +/// Module owns this resource. +root_pkg: *Package, +/// Module owns this resource. +root_scope: *Scope.ZIRModule, +/// Pointer to externally managed resource. +bin_file: *link.ElfFile, +/// It's rare for a decl to be exported, so we save memory by having a sparse map of +/// Decl pointers to details about them being exported. +/// The Export memory is owned by the `export_owners` table; the slice itself is owned by this table. +decl_exports: std.AutoHashMap(*Decl, []*Export), +/// This models the Decls that perform exports, so that `decl_exports` can be updated when a Decl +/// is modified. Note that the key of this table is not the Decl being exported, but the Decl that +/// is performing the export of another Decl. +/// This table owns the Export memory. +export_owners: std.AutoHashMap(*Decl, []*Export), +/// Maps fully qualified namespaced names to the Decl struct for them. +decl_table: std.AutoHashMap(Decl.Hash, *Decl), + +optimize_mode: std.builtin.Mode, +link_error_flags: link.ElfFile.ErrorFlags = link.ElfFile.ErrorFlags{}, + +work_queue: std.fifo.LinearFifo(WorkItem, .Dynamic), + +/// We optimize memory usage for a compilation with no compile errors by storing the +/// error messages and mapping outside of `Decl`. +/// The ErrorMsg memory is owned by the decl, using Module's allocator. +/// Note that a Decl can succeed but the Fn it represents can fail. In this case, +/// a Decl can have a failed_decls entry but have analysis status of success. +failed_decls: std.AutoHashMap(*Decl, *ErrorMsg), +/// Using a map here for consistency with the other fields here. +/// The ErrorMsg memory is owned by the `Scope.ZIRModule`, using Module's allocator. +failed_files: std.AutoHashMap(*Scope.ZIRModule, *ErrorMsg), +/// Using a map here for consistency with the other fields here. +/// The ErrorMsg memory is owned by the `Export`, using Module's allocator. +failed_exports: std.AutoHashMap(*Export, *ErrorMsg), + +pub const WorkItem = union(enum) { + /// Write the machine code for a Decl to the output file. + codegen_decl: *Decl, +}; + +pub const Export = struct { + options: std.builtin.ExportOptions, + /// Byte offset into the file that contains the export directive. + src: usize, + /// Represents the position of the export, if any, in the output file. + link: link.ElfFile.Export, + /// The Decl that performs the export. Note that this is *not* the Decl being exported. + owner_decl: *Decl, + status: enum { + in_progress, + failed, + /// Indicates that the failure was due to a temporary issue, such as an I/O error + /// when writing to the output file. Retrying the export may succeed. + failed_retryable, + complete, + }, +}; + +pub const Decl = struct { + /// This name is relative to the containing namespace of the decl. It uses a null-termination + /// to save bytes, since there can be a lot of decls in a compilation. The null byte is not allowed + /// in symbol names, because executable file formats use null-terminated strings for symbol names. + /// All Decls have names, even values that are not bound to a zig namespace. This is necessary for + /// mapping them to an address in the output file. + /// Memory owned by this decl, using Module's allocator. + name: [*:0]const u8, + /// The direct parent container of the Decl. This field will need to get more fleshed out when + /// self-hosted supports proper struct types and Zig AST => ZIR. + /// Reference to externally owned memory. + scope: *Scope.ZIRModule, + /// Byte offset into the source file that contains this declaration. + /// This is the base offset that src offsets within this Decl are relative to. + src: usize, + /// The most recent value of the Decl after a successful semantic analysis. + /// The tag for this union is determined by the tag value of the analysis field. + typed_value: union { + never_succeeded: void, + most_recent: TypedValue.Managed, + }, + /// Represents the "shallow" analysis status. For example, for decls that are functions, + /// the function type is analyzed with this set to `in_progress`, however, the semantic + /// analysis of the function body is performed with this value set to `success`. Functions + /// have their own analysis status field. + analysis: enum { + initial_in_progress, + /// This Decl might be OK but it depends on another one which did not successfully complete + /// semantic analysis. This Decl never had a value computed. + initial_dependency_failure, + /// Semantic analysis failure. This Decl never had a value computed. + /// There will be a corresponding ErrorMsg in Module.failed_decls. + initial_sema_failure, + /// In this case the `typed_value.most_recent` can still be accessed. + /// There will be a corresponding ErrorMsg in Module.failed_decls. + codegen_failure, + /// In this case the `typed_value.most_recent` can still be accessed. + /// There will be a corresponding ErrorMsg in Module.failed_decls. + /// This indicates the failure was something like running out of disk space, + /// and attempting codegen again may succeed. + codegen_failure_retryable, + /// This Decl might be OK but it depends on another one which did not successfully complete + /// semantic analysis. There is a most recent value available. + repeat_dependency_failure, + /// Semantic anlaysis failure, but the `typed_value.most_recent` can be accessed. + /// There will be a corresponding ErrorMsg in Module.failed_decls. + repeat_sema_failure, + /// Completed successfully before; the `typed_value.most_recent` can be accessed, and + /// new semantic analysis is in progress. + repeat_in_progress, + /// Everything is done and updated. + complete, + }, + + /// Represents the position of the code in the output file. + /// This is populated regardless of semantic analysis and code generation. + link: link.ElfFile.Decl = link.ElfFile.Decl.empty, + + /// The shallow set of other decls whose typed_value could possibly change if this Decl's + /// typed_value is modified. + /// TODO look into using a lightweight map/set data structure rather than a linear array. + dependants: ArrayListUnmanaged(*Decl) = ArrayListUnmanaged(*Decl){}, + + contents_hash: Hash, + + pub fn destroy(self: *Decl, allocator: *Allocator) void { + allocator.free(mem.spanZ(self.name)); + if (self.typedValueManaged()) |tvm| { + tvm.deinit(allocator); + } + allocator.destroy(self); + } + + pub const Hash = [16]u8; + + /// If the name is small enough, it is used directly as the hash. + /// If it is long, blake3 hash is computed. + pub fn hashSimpleName(name: []const u8) Hash { + var out: Hash = undefined; + if (name.len <= Hash.len) { + mem.copy(u8, &out, name); + mem.set(u8, out[name.len..], 0); + } else { + std.crypto.Blake3.hash(name, &out); + } + return out; + } + + /// Must generate unique bytes with no collisions with other decls. + /// The point of hashing here is only to limit the number of bytes of + /// the unique identifier to a fixed size (16 bytes). + pub fn fullyQualifiedNameHash(self: Decl) Hash { + // Right now we only have ZIRModule as the source. So this is simply the + // relative name of the decl. + return hashSimpleName(mem.spanZ(u8, self.name)); + } + + pub fn typedValue(self: *Decl) error{AnalysisFail}!TypedValue { + const tvm = self.typedValueManaged() orelse return error.AnalysisFail; + return tvm.typed_value; + } + + pub fn value(self: *Decl) error{AnalysisFail}!Value { + return (try self.typedValue()).val; + } + + pub fn dump(self: *Decl) void { + const loc = std.zig.findLineColumn(self.scope.source.bytes, self.src); + std.debug.warn("{}:{}:{} name={} status={}", .{ + self.scope.sub_file_path, + loc.line + 1, + loc.column + 1, + mem.spanZ(self.name), + @tagName(self.analysis), + }); + if (self.typedValueManaged()) |tvm| { + std.debug.warn(" ty={} val={}", .{ tvm.typed_value.ty, tvm.typed_value.val }); + } + std.debug.warn("\n", .{}); + } + + fn typedValueManaged(self: *Decl) ?*TypedValue.Managed { + switch (self.analysis) { + .initial_in_progress, + .initial_dependency_failure, + .initial_sema_failure, + => return null, + .codegen_failure, + .codegen_failure_retryable, + .repeat_dependency_failure, + .repeat_sema_failure, + .repeat_in_progress, + .complete, + => return &self.typed_value.most_recent, + } + } +}; + +/// Fn struct memory is owned by the Decl's TypedValue.Managed arena allocator. +pub const Fn = struct { + /// This memory owned by the Decl's TypedValue.Managed arena allocator. + fn_type: Type, + analysis: union(enum) { + /// The value is the source instruction. + queued: *zir.Inst.Fn, + in_progress: *Analysis, + /// There will be a corresponding ErrorMsg in Module.failed_decls + sema_failure, + /// This Fn might be OK but it depends on another Decl which did not successfully complete + /// semantic analysis. + dependency_failure, + success: Body, + }, + + /// This memory is temporary and points to stack memory for the duration + /// of Fn analysis. + pub const Analysis = struct { + inner_block: Scope.Block, + /// TODO Performance optimization idea: instead of this inst_table, + /// use a field in the zir.Inst instead to track corresponding instructions + inst_table: std.AutoHashMap(*zir.Inst, *Inst), + needed_inst_capacity: usize, + }; +}; + +pub const Scope = struct { + tag: Tag, + + pub fn cast(base: *Scope, comptime T: type) ?*T { + if (base.tag != T.base_tag) + return null; + + return @fieldParentPtr(T, "base", base); + } + + /// Asserts the scope has a parent which is a DeclAnalysis and + /// returns the arena Allocator. + pub fn arena(self: *Scope) *Allocator { + switch (self.tag) { + .block => return self.cast(Block).?.arena, + .decl => return &self.cast(DeclAnalysis).?.arena.allocator, + .zir_module => return &self.cast(ZIRModule).?.contents.module.arena.allocator, + } + } + + /// Asserts the scope has a parent which is a DeclAnalysis and + /// returns the Decl. + pub fn decl(self: *Scope) *Decl { + switch (self.tag) { + .block => return self.cast(Block).?.decl, + .decl => return self.cast(DeclAnalysis).?.decl, + .zir_module => unreachable, + } + } + + /// Asserts the scope has a parent which is a ZIRModule and + /// returns it. + pub fn namespace(self: *Scope) *ZIRModule { + switch (self.tag) { + .block => return self.cast(Block).?.decl.scope, + .decl => return self.cast(DeclAnalysis).?.decl.scope, + .zir_module => return self.cast(ZIRModule).?, + } + } + + pub fn dumpInst(self: *Scope, inst: *Inst) void { + const zir_module = self.namespace(); + const loc = std.zig.findLineColumn(zir_module.source.bytes, inst.src); + std.debug.warn("{}:{}:{}: {}: ty={}\n", .{ + zir_module.sub_file_path, + loc.line + 1, + loc.column + 1, + @tagName(inst.tag), + inst.ty, + }); + } + + pub const Tag = enum { + zir_module, + block, + decl, + }; + + pub const ZIRModule = struct { + pub const base_tag: Tag = .zir_module; + base: Scope = Scope{ .tag = base_tag }, + /// Relative to the owning package's root_src_dir. + /// Reference to external memory, not owned by ZIRModule. + sub_file_path: []const u8, + source: union { + unloaded: void, + bytes: [:0]const u8, + }, + contents: union { + not_available: void, + module: *zir.Module, + }, + status: enum { + never_loaded, + unloaded_success, + unloaded_parse_failure, + unloaded_sema_failure, + loaded_parse_failure, + loaded_sema_failure, + loaded_success, + }, + + pub fn unload(self: *ZIRModule, allocator: *Allocator) void { + switch (self.status) { + .never_loaded, + .unloaded_parse_failure, + .unloaded_sema_failure, + .unloaded_success, + => {}, + + .loaded_success => { + allocator.free(self.source.bytes); + self.contents.module.deinit(allocator); + allocator.destroy(self.contents.module); + self.status = .unloaded_success; + }, + .loaded_sema_failure => { + allocator.free(self.source.bytes); + self.contents.module.deinit(allocator); + allocator.destroy(self.contents.module); + self.status = .unloaded_sema_failure; + }, + .loaded_parse_failure => { + allocator.free(self.source.bytes); + self.status = .unloaded_parse_failure; + }, + } + } + + pub fn deinit(self: *ZIRModule, allocator: *Allocator) void { + self.unload(allocator); + self.* = undefined; + } + + pub fn dumpSrc(self: *ZIRModule, src: usize) void { + const loc = std.zig.findLineColumn(self.source.bytes, src); + std.debug.warn("{}:{}:{}\n", .{ self.sub_file_path, loc.line + 1, loc.column + 1 }); + } + }; + + /// This is a temporary structure, references to it are valid only + /// during semantic analysis of the block. + pub const Block = struct { + pub const base_tag: Tag = .block; + base: Scope = Scope{ .tag = base_tag }, + func: *Fn, + decl: *Decl, + instructions: ArrayListUnmanaged(*Inst), + /// Points to the arena allocator of DeclAnalysis + arena: *Allocator, + }; + + /// This is a temporary structure, references to it are valid only + /// during semantic analysis of the decl. + pub const DeclAnalysis = struct { + pub const base_tag: Tag = .decl; + base: Scope = Scope{ .tag = base_tag }, + decl: *Decl, + arena: std.heap.ArenaAllocator, + }; +}; + +pub const Body = struct { + instructions: []*Inst, +}; + +pub const AllErrors = struct { + arena: std.heap.ArenaAllocator.State, + list: []const Message, + + pub const Message = struct { + src_path: []const u8, + line: usize, + column: usize, + byte_offset: usize, + msg: []const u8, + }; + + pub fn deinit(self: *AllErrors, allocator: *Allocator) void { + self.arena.promote(allocator).deinit(); + } + + fn add( + arena: *std.heap.ArenaAllocator, + errors: *std.ArrayList(Message), + sub_file_path: []const u8, + source: []const u8, + simple_err_msg: ErrorMsg, + ) !void { + const loc = std.zig.findLineColumn(source, simple_err_msg.byte_offset); + try errors.append(.{ + .src_path = try arena.allocator.dupe(u8, sub_file_path), + .msg = try arena.allocator.dupe(u8, simple_err_msg.msg), + .byte_offset = simple_err_msg.byte_offset, + .line = loc.line, + .column = loc.column, + }); + } +}; + +pub fn deinit(self: *Module) void { + const allocator = self.allocator; + self.work_queue.deinit(); + { + var it = self.decl_table.iterator(); + while (it.next()) |kv| { + kv.value.destroy(allocator); + } + self.decl_table.deinit(); + } + { + var it = self.failed_decls.iterator(); + while (it.next()) |kv| { + kv.value.destroy(allocator); + } + self.failed_decls.deinit(); + } + { + var it = self.failed_files.iterator(); + while (it.next()) |kv| { + kv.value.destroy(allocator); + } + self.failed_files.deinit(); + } + { + var it = self.failed_exports.iterator(); + while (it.next()) |kv| { + kv.value.destroy(allocator); + } + self.failed_exports.deinit(); + } + { + var it = self.decl_exports.iterator(); + while (it.next()) |kv| { + const export_list = kv.value; + allocator.free(export_list); + } + self.decl_exports.deinit(); + } + { + var it = self.export_owners.iterator(); + while (it.next()) |kv| { + const export_list = kv.value; + for (export_list) |exp| { + allocator.destroy(exp); + } + allocator.free(export_list); + } + self.export_owners.deinit(); + } + self.root_pkg.destroy(); + { + self.root_scope.deinit(allocator); + allocator.destroy(self.root_scope); + } + self.* = undefined; +} + +pub fn target(self: Module) std.Target { + return self.bin_file.options.target; +} + +/// Detect changes to source files, perform semantic analysis, and update the output files. +pub fn update(self: *Module) !void { + // TODO Use the cache hash file system to detect which source files changed. + // Here we simulate a full cache miss. + // Analyze the root source file now. + self.analyzeRoot(self.root_scope) catch |err| switch (err) { + error.AnalysisFail => { + assert(self.totalErrorCount() != 0); + }, + else => |e| return e, + }; + + try self.performAllTheWork(); + + // Unload all the source files from memory. + self.root_scope.unload(self.allocator); + + try self.bin_file.flush(); + self.link_error_flags = self.bin_file.error_flags; +} + +pub fn totalErrorCount(self: *Module) usize { + return self.failed_decls.size + + self.failed_files.size + + self.failed_exports.size + + @boolToInt(self.link_error_flags.no_entry_point_found); +} + +pub fn getAllErrorsAlloc(self: *Module) !AllErrors { + var arena = std.heap.ArenaAllocator.init(self.allocator); + errdefer arena.deinit(); + + var errors = std.ArrayList(AllErrors.Message).init(self.allocator); + defer errors.deinit(); + + { + var it = self.failed_files.iterator(); + while (it.next()) |kv| { + const scope = kv.key; + const err_msg = kv.value; + const source = scope.source.bytes; + try AllErrors.add(&arena, &errors, scope.sub_file_path, source, err_msg.*); + } + } + { + var it = self.failed_decls.iterator(); + while (it.next()) |kv| { + const decl = kv.key; + const err_msg = kv.value; + const source = decl.scope.source.bytes; + try AllErrors.add(&arena, &errors, decl.scope.sub_file_path, source, err_msg.*); + } + } + { + var it = self.failed_exports.iterator(); + while (it.next()) |kv| { + const decl = kv.key.owner_decl; + const err_msg = kv.value; + const source = decl.scope.source.bytes; + try AllErrors.add(&arena, &errors, decl.scope.sub_file_path, source, err_msg.*); + } + } + + if (self.link_error_flags.no_entry_point_found) { + try errors.append(.{ + .src_path = self.root_pkg.root_src_path, + .line = 0, + .column = 0, + .byte_offset = 0, + .msg = try std.fmt.allocPrint(&arena.allocator, "no entry point found", .{}), + }); + } + + assert(errors.items.len == self.totalErrorCount()); + + return AllErrors{ + .arena = arena.state, + .list = try arena.allocator.dupe(AllErrors.Message, errors.items), + }; +} + +const InnerError = error{ OutOfMemory, AnalysisFail }; + +pub fn performAllTheWork(self: *Module) error{OutOfMemory}!void { + while (self.work_queue.readItem()) |work_item| switch (work_item) { + .codegen_decl => |decl| switch (decl.analysis) { + .initial_in_progress, + .repeat_in_progress, + => unreachable, + + .initial_sema_failure, + .repeat_sema_failure, + .codegen_failure, + .initial_dependency_failure, + .repeat_dependency_failure, + => continue, + + .complete, .codegen_failure_retryable => { + if (decl.typed_value.most_recent.typed_value.val.cast(Value.Payload.Function)) |payload| { + switch (payload.func.analysis) { + .queued => self.analyzeFnBody(decl, payload.func) catch |err| switch (err) { + error.AnalysisFail => { + if (payload.func.analysis == .queued) { + payload.func.analysis = .dependency_failure; + } + continue; + }, + else => |e| return e, + }, + .in_progress => unreachable, + .sema_failure, .dependency_failure => continue, + .success => {}, + } + } + + assert(decl.typed_value.most_recent.typed_value.ty.hasCodeGenBits()); + + self.bin_file.updateDecl(self, decl) catch |err| switch (err) { + error.OutOfMemory => return error.OutOfMemory, + error.AnalysisFail => { + decl.analysis = .repeat_dependency_failure; + }, + else => { + try self.failed_decls.ensureCapacity(self.failed_decls.size + 1); + self.failed_decls.putAssumeCapacityNoClobber(decl, try ErrorMsg.create( + self.allocator, + decl.src, + "unable to codegen: {}", + .{@errorName(err)}, + )); + decl.analysis = .codegen_failure_retryable; + }, + }; + }, + }, + }; +} + +fn getSrcModule(self: *Module, root_scope: *Scope.ZIRModule) !*zir.Module { + switch (root_scope.status) { + .never_loaded, .unloaded_success => { + try self.failed_files.ensureCapacity(self.failed_files.size + 1); + + var keep_source = false; + const source = try self.root_pkg.root_src_dir.readFileAllocOptions( + self.allocator, + self.root_pkg.root_src_path, + std.math.maxInt(u32), + 1, + 0, + ); + defer if (!keep_source) self.allocator.free(source); + + var keep_zir_module = false; + const zir_module = try self.allocator.create(zir.Module); + defer if (!keep_zir_module) self.allocator.destroy(zir_module); + + zir_module.* = try zir.parse(self.allocator, source); + defer if (!keep_zir_module) zir_module.deinit(self.allocator); + + if (zir_module.error_msg) |src_err_msg| { + self.failed_files.putAssumeCapacityNoClobber( + root_scope, + try ErrorMsg.create(self.allocator, src_err_msg.byte_offset, "{}", .{src_err_msg.msg}), + ); + root_scope.status = .loaded_parse_failure; + root_scope.source = .{ .bytes = source }; + keep_source = true; + return error.AnalysisFail; + } + + root_scope.status = .loaded_success; + root_scope.source = .{ .bytes = source }; + keep_source = true; + root_scope.contents = .{ .module = zir_module }; + keep_zir_module = true; + + return zir_module; + }, + + .unloaded_parse_failure, + .unloaded_sema_failure, + .loaded_parse_failure, + .loaded_sema_failure, + => return error.AnalysisFail, + .loaded_success => return root_scope.contents.module, + } +} + +fn analyzeRoot(self: *Module, root_scope: *Scope.ZIRModule) !void { + // TODO use the cache to identify, from the modified source files, the decls which have + // changed based on the span of memory that represents the decl in the re-parsed source file. + // Use the cached dependency graph to recursively determine the set of decls which need + // regeneration. + // Here we simulate adding a source file which was previously not part of the compilation, + // which means scanning the decls looking for exports. + // TODO also identify decls that need to be deleted. + switch (root_scope.status) { + .never_loaded => { + const src_module = try self.getSrcModule(root_scope); + + // Here we ensure enough queue capacity to store all the decls, so that later we can use + // appendAssumeCapacity. + try self.work_queue.ensureUnusedCapacity(src_module.decls.len); + + for (src_module.decls) |decl| { + if (decl.cast(zir.Inst.Export)) |export_inst| { + _ = try self.resolveDecl(&root_scope.base, &export_inst.base, link.ElfFile.Decl.empty); + } + } + }, + + .unloaded_parse_failure, + .unloaded_sema_failure, + .loaded_parse_failure, + .loaded_sema_failure, + .loaded_success, + .unloaded_success, + => { + const src_module = try self.getSrcModule(root_scope); + + // Look for changed decls. + for (src_module.decls) |src_decl| { + const name_hash = Decl.hashSimpleName(src_decl.name); + if (self.decl_table.get(name_hash)) |kv| { + const decl = kv.value; + const new_contents_hash = Decl.hashSimpleName(src_decl.contents); + if (!mem.eql(u8, &new_contents_hash, &decl.contents_hash)) { + // TODO recursive dependency management + std.debug.warn("noticed that '{}' changed\n", .{src_decl.name}); + self.decl_table.removeAssertDiscard(name_hash); + const saved_link = decl.link; + decl.destroy(self.allocator); + if (self.export_owners.getValue(decl)) |exports| { + @panic("TODO handle updating a decl that does an export"); + } + const new_decl = self.resolveDecl( + &root_scope.base, + src_decl, + saved_link, + ) catch |err| switch (err) { + error.OutOfMemory => return error.OutOfMemory, + error.AnalysisFail => continue, + }; + if (self.decl_exports.remove(decl)) |entry| { + self.decl_exports.putAssumeCapacityNoClobber(new_decl, entry.value); + } + } + } else if (src_decl.cast(zir.Inst.Export)) |export_inst| { + _ = try self.resolveDecl(&root_scope.base, &export_inst.base, link.ElfFile.Decl.empty); + } + } + }, + } +} + +fn analyzeFnBody(self: *Module, decl: *Decl, func: *Fn) !void { + // Use the Decl's arena for function memory. + var arena = decl.typed_value.most_recent.arena.?.promote(self.allocator); + defer decl.typed_value.most_recent.arena.?.* = arena.state; + var analysis: Fn.Analysis = .{ + .inner_block = .{ + .func = func, + .decl = decl, + .instructions = .{}, + .arena = &arena.allocator, + }, + .needed_inst_capacity = 0, + .inst_table = std.AutoHashMap(*zir.Inst, *Inst).init(self.allocator), + }; + defer analysis.inner_block.instructions.deinit(self.allocator); + defer analysis.inst_table.deinit(); + + const fn_inst = func.analysis.queued; + func.analysis = .{ .in_progress = &analysis }; + + try self.analyzeBody(&analysis.inner_block.base, fn_inst.positionals.body); + + func.analysis = .{ + .success = .{ + .instructions = try arena.allocator.dupe(*Inst, analysis.inner_block.instructions.items), + }, + }; +} + +fn resolveDecl( + self: *Module, + scope: *Scope, + old_inst: *zir.Inst, + bin_file_link: link.ElfFile.Decl, +) InnerError!*Decl { + const hash = Decl.hashSimpleName(old_inst.name); + if (self.decl_table.get(hash)) |kv| { + return kv.value; + } else { + const new_decl = blk: { + try self.decl_table.ensureCapacity(self.decl_table.size + 1); + const new_decl = try self.allocator.create(Decl); + errdefer self.allocator.destroy(new_decl); + const name = try mem.dupeZ(self.allocator, u8, old_inst.name); + errdefer self.allocator.free(name); + new_decl.* = .{ + .name = name, + .scope = scope.namespace(), + .src = old_inst.src, + .typed_value = .{ .never_succeeded = {} }, + .analysis = .initial_in_progress, + .contents_hash = Decl.hashSimpleName(old_inst.contents), + .link = bin_file_link, + }; + self.decl_table.putAssumeCapacityNoClobber(hash, new_decl); + break :blk new_decl; + }; + + var decl_scope: Scope.DeclAnalysis = .{ + .decl = new_decl, + .arena = std.heap.ArenaAllocator.init(self.allocator), + }; + errdefer decl_scope.arena.deinit(); + + const typed_value = self.analyzeInstConst(&decl_scope.base, old_inst) catch |err| switch (err) { + error.OutOfMemory => return error.OutOfMemory, + error.AnalysisFail => { + switch (new_decl.analysis) { + .initial_in_progress => new_decl.analysis = .initial_dependency_failure, + .repeat_in_progress => new_decl.analysis = .repeat_dependency_failure, + else => {}, + } + return error.AnalysisFail; + }, + }; + const arena_state = try decl_scope.arena.allocator.create(std.heap.ArenaAllocator.State); + + const has_codegen_bits = typed_value.ty.hasCodeGenBits(); + if (has_codegen_bits) { + // We don't fully codegen the decl until later, but we do need to reserve a global + // offset table index for it. This allows us to codegen decls out of dependency order, + // increasing how many computations can be done in parallel. + try self.bin_file.allocateDeclIndexes(new_decl); + } + + arena_state.* = decl_scope.arena.state; + + new_decl.typed_value = .{ + .most_recent = .{ + .typed_value = typed_value, + .arena = arena_state, + }, + }; + new_decl.analysis = .complete; + if (has_codegen_bits) { + // We ensureCapacity when scanning for decls. + self.work_queue.writeItemAssumeCapacity(.{ .codegen_decl = new_decl }); + } + return new_decl; + } +} + +fn resolveCompleteDecl(self: *Module, scope: *Scope, old_inst: *zir.Inst) InnerError!*Decl { + const decl = try self.resolveDecl(scope, old_inst, link.ElfFile.Decl.empty); + switch (decl.analysis) { + .initial_in_progress => unreachable, + .repeat_in_progress => unreachable, + .initial_dependency_failure, + .repeat_dependency_failure, + .initial_sema_failure, + .repeat_sema_failure, + .codegen_failure, + .codegen_failure_retryable, + => return error.AnalysisFail, + + .complete => return decl, + } +} + +fn resolveInst(self: *Module, scope: *Scope, old_inst: *zir.Inst) InnerError!*Inst { + if (scope.cast(Scope.Block)) |block| { + if (block.func.analysis.in_progress.inst_table.get(old_inst)) |kv| { + return kv.value; + } + } + + const decl = try self.resolveCompleteDecl(scope, old_inst); + const decl_ref = try self.analyzeDeclRef(scope, old_inst.src, decl); + return self.analyzeDeref(scope, old_inst.src, decl_ref, old_inst.src); +} + +fn requireRuntimeBlock(self: *Module, scope: *Scope, src: usize) !*Scope.Block { + return scope.cast(Scope.Block) orelse + return self.fail(scope, src, "instruction illegal outside function body", .{}); +} + +fn resolveInstConst(self: *Module, scope: *Scope, old_inst: *zir.Inst) InnerError!TypedValue { + const new_inst = try self.resolveInst(scope, old_inst); + const val = try self.resolveConstValue(scope, new_inst); + return TypedValue{ + .ty = new_inst.ty, + .val = val, + }; +} + +fn resolveConstValue(self: *Module, scope: *Scope, base: *Inst) !Value { + return (try self.resolveDefinedValue(scope, base)) orelse + return self.fail(scope, base.src, "unable to resolve comptime value", .{}); +} + +fn resolveDefinedValue(self: *Module, scope: *Scope, base: *Inst) !?Value { + if (base.value()) |val| { + if (val.isUndef()) { + return self.fail(scope, base.src, "use of undefined value here causes undefined behavior", .{}); + } + return val; + } + return null; +} + +fn resolveConstString(self: *Module, scope: *Scope, old_inst: *zir.Inst) ![]u8 { + const new_inst = try self.resolveInst(scope, old_inst); + const wanted_type = Type.initTag(.const_slice_u8); + const coerced_inst = try self.coerce(scope, wanted_type, new_inst); + const val = try self.resolveConstValue(scope, coerced_inst); + return val.toAllocatedBytes(scope.arena()); +} + +fn resolveType(self: *Module, scope: *Scope, old_inst: *zir.Inst) !Type { + const new_inst = try self.resolveInst(scope, old_inst); + const wanted_type = Type.initTag(.@"type"); + const coerced_inst = try self.coerce(scope, wanted_type, new_inst); + const val = try self.resolveConstValue(scope, coerced_inst); + return val.toType(); +} + +fn analyzeExport(self: *Module, scope: *Scope, export_inst: *zir.Inst.Export) InnerError!void { + try self.decl_exports.ensureCapacity(self.decl_exports.size + 1); + try self.export_owners.ensureCapacity(self.export_owners.size + 1); + const symbol_name = try self.resolveConstString(scope, export_inst.positionals.symbol_name); + const exported_decl = try self.resolveCompleteDecl(scope, export_inst.positionals.value); + const typed_value = exported_decl.typed_value.most_recent.typed_value; + switch (typed_value.ty.zigTypeTag()) { + .Fn => {}, + else => return self.fail( + scope, + export_inst.positionals.value.src, + "unable to export type '{}'", + .{typed_value.ty}, + ), + } + const new_export = try self.allocator.create(Export); + errdefer self.allocator.destroy(new_export); + + const owner_decl = scope.decl(); + + new_export.* = .{ + .options = .{ .name = symbol_name }, + .src = export_inst.base.src, + .link = .{}, + .owner_decl = owner_decl, + .status = .in_progress, + }; + + // Add to export_owners table. + const eo_gop = self.export_owners.getOrPut(owner_decl) catch unreachable; + if (!eo_gop.found_existing) { + eo_gop.kv.value = &[0]*Export{}; + } + eo_gop.kv.value = try self.allocator.realloc(eo_gop.kv.value, eo_gop.kv.value.len + 1); + eo_gop.kv.value[eo_gop.kv.value.len - 1] = new_export; + errdefer eo_gop.kv.value = self.allocator.shrink(eo_gop.kv.value, eo_gop.kv.value.len - 1); + + // Add to exported_decl table. + const de_gop = self.decl_exports.getOrPut(exported_decl) catch unreachable; + if (!de_gop.found_existing) { + de_gop.kv.value = &[0]*Export{}; + } + de_gop.kv.value = try self.allocator.realloc(de_gop.kv.value, de_gop.kv.value.len + 1); + de_gop.kv.value[de_gop.kv.value.len - 1] = new_export; + errdefer de_gop.kv.value = self.allocator.shrink(de_gop.kv.value, de_gop.kv.value.len - 1); + + self.bin_file.updateDeclExports(self, exported_decl, de_gop.kv.value) catch |err| switch (err) { + error.OutOfMemory => return error.OutOfMemory, + else => { + try self.failed_exports.ensureCapacity(self.failed_exports.size + 1); + self.failed_exports.putAssumeCapacityNoClobber(new_export, try ErrorMsg.create( + self.allocator, + export_inst.base.src, + "unable to export: {}", + .{@errorName(err)}, + )); + new_export.status = .failed_retryable; + }, + }; +} + +/// TODO should not need the cast on the last parameter at the callsites +fn addNewInstArgs( + self: *Module, + block: *Scope.Block, + src: usize, + ty: Type, + comptime T: type, + args: Inst.Args(T), +) !*Inst { + const inst = try self.addNewInst(block, src, ty, T); + inst.args = args; + return &inst.base; +} + +fn addNewInst(self: *Module, block: *Scope.Block, src: usize, ty: Type, comptime T: type) !*T { + const inst = try block.arena.create(T); + inst.* = .{ + .base = .{ + .tag = T.base_tag, + .ty = ty, + .src = src, + }, + .args = undefined, + }; + try block.instructions.append(self.allocator, &inst.base); + return inst; +} + +fn constInst(self: *Module, scope: *Scope, src: usize, typed_value: TypedValue) !*Inst { + const const_inst = try scope.arena().create(Inst.Constant); + const_inst.* = .{ + .base = .{ + .tag = Inst.Constant.base_tag, + .ty = typed_value.ty, + .src = src, + }, + .val = typed_value.val, + }; + return &const_inst.base; +} + +fn constStr(self: *Module, scope: *Scope, src: usize, str: []const u8) !*Inst { + const ty_payload = try scope.arena().create(Type.Payload.Array_u8_Sentinel0); + ty_payload.* = .{ .len = str.len }; + + const bytes_payload = try scope.arena().create(Value.Payload.Bytes); + bytes_payload.* = .{ .data = str }; + + return self.constInst(scope, src, .{ + .ty = Type.initPayload(&ty_payload.base), + .val = Value.initPayload(&bytes_payload.base), + }); +} + +fn constType(self: *Module, scope: *Scope, src: usize, ty: Type) !*Inst { + return self.constInst(scope, src, .{ + .ty = Type.initTag(.type), + .val = try ty.toValue(scope.arena()), + }); +} + +fn constVoid(self: *Module, scope: *Scope, src: usize) !*Inst { + return self.constInst(scope, src, .{ + .ty = Type.initTag(.void), + .val = Value.initTag(.the_one_possible_value), + }); +} + +fn constUndef(self: *Module, scope: *Scope, src: usize, ty: Type) !*Inst { + return self.constInst(scope, src, .{ + .ty = ty, + .val = Value.initTag(.undef), + }); +} + +fn constBool(self: *Module, scope: *Scope, src: usize, v: bool) !*Inst { + return self.constInst(scope, src, .{ + .ty = Type.initTag(.bool), + .val = ([2]Value{ Value.initTag(.bool_false), Value.initTag(.bool_true) })[@boolToInt(v)], + }); +} + +fn constIntUnsigned(self: *Module, scope: *Scope, src: usize, ty: Type, int: u64) !*Inst { + const int_payload = try scope.arena().create(Value.Payload.Int_u64); + int_payload.* = .{ .int = int }; + + return self.constInst(scope, src, .{ + .ty = ty, + .val = Value.initPayload(&int_payload.base), + }); +} + +fn constIntSigned(self: *Module, scope: *Scope, src: usize, ty: Type, int: i64) !*Inst { + const int_payload = try scope.arena().create(Value.Payload.Int_i64); + int_payload.* = .{ .int = int }; + + return self.constInst(scope, src, .{ + .ty = ty, + .val = Value.initPayload(&int_payload.base), + }); +} + +fn constIntBig(self: *Module, scope: *Scope, src: usize, ty: Type, big_int: BigIntConst) !*Inst { + const val_payload = if (big_int.positive) blk: { + if (big_int.to(u64)) |x| { + return self.constIntUnsigned(scope, src, ty, x); + } else |err| switch (err) { + error.NegativeIntoUnsigned => unreachable, + error.TargetTooSmall => {}, // handled below + } + const big_int_payload = try scope.arena().create(Value.Payload.IntBigPositive); + big_int_payload.* = .{ .limbs = big_int.limbs }; + break :blk &big_int_payload.base; + } else blk: { + if (big_int.to(i64)) |x| { + return self.constIntSigned(scope, src, ty, x); + } else |err| switch (err) { + error.NegativeIntoUnsigned => unreachable, + error.TargetTooSmall => {}, // handled below + } + const big_int_payload = try scope.arena().create(Value.Payload.IntBigNegative); + big_int_payload.* = .{ .limbs = big_int.limbs }; + break :blk &big_int_payload.base; + }; + + return self.constInst(scope, src, .{ + .ty = ty, + .val = Value.initPayload(val_payload), + }); +} + +fn analyzeInstConst(self: *Module, scope: *Scope, old_inst: *zir.Inst) InnerError!TypedValue { + const new_inst = try self.analyzeInst(scope, old_inst); + return TypedValue{ + .ty = new_inst.ty, + .val = try self.resolveConstValue(scope, new_inst), + }; +} + +fn analyzeInst(self: *Module, scope: *Scope, old_inst: *zir.Inst) InnerError!*Inst { + switch (old_inst.tag) { + .breakpoint => return self.analyzeInstBreakpoint(scope, old_inst.cast(zir.Inst.Breakpoint).?), + .call => return self.analyzeInstCall(scope, old_inst.cast(zir.Inst.Call).?), + .declref => return self.analyzeInstDeclRef(scope, old_inst.cast(zir.Inst.DeclRef).?), + .str => { + const bytes = old_inst.cast(zir.Inst.Str).?.positionals.bytes; + // The bytes references memory inside the ZIR module, which can get deallocated + // after semantic analysis is complete. We need the memory to be in the Decl's arena. + const arena_bytes = try scope.arena().dupe(u8, bytes); + return self.constStr(scope, old_inst.src, arena_bytes); + }, + .int => { + const big_int = old_inst.cast(zir.Inst.Int).?.positionals.int; + return self.constIntBig(scope, old_inst.src, Type.initTag(.comptime_int), big_int); + }, + .ptrtoint => return self.analyzeInstPtrToInt(scope, old_inst.cast(zir.Inst.PtrToInt).?), + .fieldptr => return self.analyzeInstFieldPtr(scope, old_inst.cast(zir.Inst.FieldPtr).?), + .deref => return self.analyzeInstDeref(scope, old_inst.cast(zir.Inst.Deref).?), + .as => return self.analyzeInstAs(scope, old_inst.cast(zir.Inst.As).?), + .@"asm" => return self.analyzeInstAsm(scope, old_inst.cast(zir.Inst.Asm).?), + .@"unreachable" => return self.analyzeInstUnreachable(scope, old_inst.cast(zir.Inst.Unreachable).?), + .@"return" => return self.analyzeInstRet(scope, old_inst.cast(zir.Inst.Return).?), + .@"fn" => return self.analyzeInstFn(scope, old_inst.cast(zir.Inst.Fn).?), + .@"export" => { + try self.analyzeExport(scope, old_inst.cast(zir.Inst.Export).?); + return self.constVoid(scope, old_inst.src); + }, + .primitive => return self.analyzeInstPrimitive(scope, old_inst.cast(zir.Inst.Primitive).?), + .ref => return self.analyzeInstRef(scope, old_inst.cast(zir.Inst.Ref).?), + .fntype => return self.analyzeInstFnType(scope, old_inst.cast(zir.Inst.FnType).?), + .intcast => return self.analyzeInstIntCast(scope, old_inst.cast(zir.Inst.IntCast).?), + .bitcast => return self.analyzeInstBitCast(scope, old_inst.cast(zir.Inst.BitCast).?), + .elemptr => return self.analyzeInstElemPtr(scope, old_inst.cast(zir.Inst.ElemPtr).?), + .add => return self.analyzeInstAdd(scope, old_inst.cast(zir.Inst.Add).?), + .cmp => return self.analyzeInstCmp(scope, old_inst.cast(zir.Inst.Cmp).?), + .condbr => return self.analyzeInstCondBr(scope, old_inst.cast(zir.Inst.CondBr).?), + .isnull => return self.analyzeInstIsNull(scope, old_inst.cast(zir.Inst.IsNull).?), + .isnonnull => return self.analyzeInstIsNonNull(scope, old_inst.cast(zir.Inst.IsNonNull).?), + } +} + +fn analyzeInstBreakpoint(self: *Module, scope: *Scope, inst: *zir.Inst.Breakpoint) InnerError!*Inst { + const b = try self.requireRuntimeBlock(scope, inst.base.src); + return self.addNewInstArgs(b, inst.base.src, Type.initTag(.void), Inst.Breakpoint, Inst.Args(Inst.Breakpoint){}); +} + +fn analyzeInstRef(self: *Module, scope: *Scope, inst: *zir.Inst.Ref) InnerError!*Inst { + const decl = try self.resolveCompleteDecl(scope, inst.positionals.operand); + return self.analyzeDeclRef(scope, inst.base.src, decl); +} + +fn analyzeInstDeclRef(self: *Module, scope: *Scope, inst: *zir.Inst.DeclRef) InnerError!*Inst { + const decl_name = try self.resolveConstString(scope, inst.positionals.name); + // This will need to get more fleshed out when there are proper structs & namespaces. + const zir_module = scope.namespace(); + for (zir_module.contents.module.decls) |src_decl| { + if (mem.eql(u8, src_decl.name, decl_name)) { + const decl = try self.resolveCompleteDecl(scope, src_decl); + return self.analyzeDeclRef(scope, inst.base.src, decl); + } + } + return self.fail(scope, inst.positionals.name.src, "use of undeclared identifier '{}'", .{decl_name}); +} + +fn analyzeDeclRef(self: *Module, scope: *Scope, src: usize, decl: *Decl) InnerError!*Inst { + const decl_tv = try decl.typedValue(); + const ty_payload = try scope.arena().create(Type.Payload.SingleConstPointer); + ty_payload.* = .{ .pointee_type = decl_tv.ty }; + const val_payload = try scope.arena().create(Value.Payload.DeclRef); + val_payload.* = .{ .decl = decl }; + return self.constInst(scope, src, .{ + .ty = Type.initPayload(&ty_payload.base), + .val = Value.initPayload(&val_payload.base), + }); +} + +fn analyzeInstCall(self: *Module, scope: *Scope, inst: *zir.Inst.Call) InnerError!*Inst { + const func = try self.resolveInst(scope, inst.positionals.func); + if (func.ty.zigTypeTag() != .Fn) + return self.fail(scope, inst.positionals.func.src, "type '{}' not a function", .{func.ty}); + + const cc = func.ty.fnCallingConvention(); + if (cc == .Naked) { + // TODO add error note: declared here + return self.fail( + scope, + inst.positionals.func.src, + "unable to call function with naked calling convention", + .{}, + ); + } + const call_params_len = inst.positionals.args.len; + const fn_params_len = func.ty.fnParamLen(); + if (func.ty.fnIsVarArgs()) { + if (call_params_len < fn_params_len) { + // TODO add error note: declared here + return self.fail( + scope, + inst.positionals.func.src, + "expected at least {} arguments, found {}", + .{ fn_params_len, call_params_len }, + ); + } + return self.fail(scope, inst.base.src, "TODO implement support for calling var args functions", .{}); + } else if (fn_params_len != call_params_len) { + // TODO add error note: declared here + return self.fail( + scope, + inst.positionals.func.src, + "expected {} arguments, found {}", + .{ fn_params_len, call_params_len }, + ); + } + + if (inst.kw_args.modifier == .compile_time) { + return self.fail(scope, inst.base.src, "TODO implement comptime function calls", .{}); + } + if (inst.kw_args.modifier != .auto) { + return self.fail(scope, inst.base.src, "TODO implement call with modifier {}", .{inst.kw_args.modifier}); + } + + // TODO handle function calls of generic functions + + const fn_param_types = try self.allocator.alloc(Type, fn_params_len); + defer self.allocator.free(fn_param_types); + func.ty.fnParamTypes(fn_param_types); + + const casted_args = try scope.arena().alloc(*Inst, fn_params_len); + for (inst.positionals.args) |src_arg, i| { + const uncasted_arg = try self.resolveInst(scope, src_arg); + casted_args[i] = try self.coerce(scope, fn_param_types[i], uncasted_arg); + } + + const b = try self.requireRuntimeBlock(scope, inst.base.src); + return self.addNewInstArgs(b, inst.base.src, Type.initTag(.void), Inst.Call, Inst.Args(Inst.Call){ + .func = func, + .args = casted_args, + }); +} + +fn analyzeInstFn(self: *Module, scope: *Scope, fn_inst: *zir.Inst.Fn) InnerError!*Inst { + const fn_type = try self.resolveType(scope, fn_inst.positionals.fn_type); + const new_func = try scope.arena().create(Fn); + new_func.* = .{ + .fn_type = fn_type, + .analysis = .{ .queued = fn_inst }, + }; + const fn_payload = try scope.arena().create(Value.Payload.Function); + fn_payload.* = .{ .func = new_func }; + return self.constInst(scope, fn_inst.base.src, .{ + .ty = fn_type, + .val = Value.initPayload(&fn_payload.base), + }); +} + +fn analyzeInstFnType(self: *Module, scope: *Scope, fntype: *zir.Inst.FnType) InnerError!*Inst { + const return_type = try self.resolveType(scope, fntype.positionals.return_type); + + if (return_type.zigTypeTag() == .NoReturn and + fntype.positionals.param_types.len == 0 and + fntype.kw_args.cc == .Unspecified) + { + return self.constType(scope, fntype.base.src, Type.initTag(.fn_noreturn_no_args)); + } + + if (return_type.zigTypeTag() == .NoReturn and + fntype.positionals.param_types.len == 0 and + fntype.kw_args.cc == .Naked) + { + return self.constType(scope, fntype.base.src, Type.initTag(.fn_naked_noreturn_no_args)); + } + + if (return_type.zigTypeTag() == .Void and + fntype.positionals.param_types.len == 0 and + fntype.kw_args.cc == .C) + { + return self.constType(scope, fntype.base.src, Type.initTag(.fn_ccc_void_no_args)); + } + + return self.fail(scope, fntype.base.src, "TODO implement fntype instruction more", .{}); +} + +fn analyzeInstPrimitive(self: *Module, scope: *Scope, primitive: *zir.Inst.Primitive) InnerError!*Inst { + return self.constType(scope, primitive.base.src, primitive.positionals.tag.toType()); +} + +fn analyzeInstAs(self: *Module, scope: *Scope, as: *zir.Inst.As) InnerError!*Inst { + const dest_type = try self.resolveType(scope, as.positionals.dest_type); + const new_inst = try self.resolveInst(scope, as.positionals.value); + return self.coerce(scope, dest_type, new_inst); +} + +fn analyzeInstPtrToInt(self: *Module, scope: *Scope, ptrtoint: *zir.Inst.PtrToInt) InnerError!*Inst { + const ptr = try self.resolveInst(scope, ptrtoint.positionals.ptr); + if (ptr.ty.zigTypeTag() != .Pointer) { + return self.fail(scope, ptrtoint.positionals.ptr.src, "expected pointer, found '{}'", .{ptr.ty}); + } + // TODO handle known-pointer-address + const b = try self.requireRuntimeBlock(scope, ptrtoint.base.src); + const ty = Type.initTag(.usize); + return self.addNewInstArgs(b, ptrtoint.base.src, ty, Inst.PtrToInt, Inst.Args(Inst.PtrToInt){ .ptr = ptr }); +} + +fn analyzeInstFieldPtr(self: *Module, scope: *Scope, fieldptr: *zir.Inst.FieldPtr) InnerError!*Inst { + const object_ptr = try self.resolveInst(scope, fieldptr.positionals.object_ptr); + const field_name = try self.resolveConstString(scope, fieldptr.positionals.field_name); + + const elem_ty = switch (object_ptr.ty.zigTypeTag()) { + .Pointer => object_ptr.ty.elemType(), + else => return self.fail(scope, fieldptr.positionals.object_ptr.src, "expected pointer, found '{}'", .{object_ptr.ty}), + }; + switch (elem_ty.zigTypeTag()) { + .Array => { + if (mem.eql(u8, field_name, "len")) { + const len_payload = try scope.arena().create(Value.Payload.Int_u64); + len_payload.* = .{ .int = elem_ty.arrayLen() }; + + const ref_payload = try scope.arena().create(Value.Payload.RefVal); + ref_payload.* = .{ .val = Value.initPayload(&len_payload.base) }; + + return self.constInst(scope, fieldptr.base.src, .{ + .ty = Type.initTag(.single_const_pointer_to_comptime_int), + .val = Value.initPayload(&ref_payload.base), + }); + } else { + return self.fail( + scope, + fieldptr.positionals.field_name.src, + "no member named '{}' in '{}'", + .{ field_name, elem_ty }, + ); + } + }, + else => return self.fail(scope, fieldptr.base.src, "type '{}' does not support field access", .{elem_ty}), + } +} + +fn analyzeInstIntCast(self: *Module, scope: *Scope, intcast: *zir.Inst.IntCast) InnerError!*Inst { + const dest_type = try self.resolveType(scope, intcast.positionals.dest_type); + const new_inst = try self.resolveInst(scope, intcast.positionals.value); + + const dest_is_comptime_int = switch (dest_type.zigTypeTag()) { + .ComptimeInt => true, + .Int => false, + else => return self.fail( + scope, + intcast.positionals.dest_type.src, + "expected integer type, found '{}'", + .{ + dest_type, + }, + ), + }; + + switch (new_inst.ty.zigTypeTag()) { + .ComptimeInt, .Int => {}, + else => return self.fail( + scope, + intcast.positionals.value.src, + "expected integer type, found '{}'", + .{new_inst.ty}, + ), + } + + if (dest_is_comptime_int or new_inst.value() != null) { + return self.coerce(scope, dest_type, new_inst); + } + + return self.fail(scope, intcast.base.src, "TODO implement analyze widen or shorten int", .{}); +} + +fn analyzeInstBitCast(self: *Module, scope: *Scope, inst: *zir.Inst.BitCast) InnerError!*Inst { + const dest_type = try self.resolveType(scope, inst.positionals.dest_type); + const operand = try self.resolveInst(scope, inst.positionals.operand); + return self.bitcast(scope, dest_type, operand); +} + +fn analyzeInstElemPtr(self: *Module, scope: *Scope, inst: *zir.Inst.ElemPtr) InnerError!*Inst { + const array_ptr = try self.resolveInst(scope, inst.positionals.array_ptr); + const uncasted_index = try self.resolveInst(scope, inst.positionals.index); + const elem_index = try self.coerce(scope, Type.initTag(.usize), uncasted_index); + + if (array_ptr.ty.isSinglePointer() and array_ptr.ty.elemType().zigTypeTag() == .Array) { + if (array_ptr.value()) |array_ptr_val| { + if (elem_index.value()) |index_val| { + // Both array pointer and index are compile-time known. + const index_u64 = index_val.toUnsignedInt(); + // @intCast here because it would have been impossible to construct a value that + // required a larger index. + const elem_ptr = try array_ptr_val.elemPtr(scope.arena(), @intCast(usize, index_u64)); + + const type_payload = try scope.arena().create(Type.Payload.SingleConstPointer); + type_payload.* = .{ .pointee_type = array_ptr.ty.elemType().elemType() }; + + return self.constInst(scope, inst.base.src, .{ + .ty = Type.initPayload(&type_payload.base), + .val = elem_ptr, + }); + } + } + } + + return self.fail(scope, inst.base.src, "TODO implement more analyze elemptr", .{}); +} + +fn analyzeInstAdd(self: *Module, scope: *Scope, inst: *zir.Inst.Add) InnerError!*Inst { + const lhs = try self.resolveInst(scope, inst.positionals.lhs); + const rhs = try self.resolveInst(scope, inst.positionals.rhs); + + if (lhs.ty.zigTypeTag() == .Int and rhs.ty.zigTypeTag() == .Int) { + if (lhs.value()) |lhs_val| { + if (rhs.value()) |rhs_val| { + // TODO is this a performance issue? maybe we should try the operation without + // resorting to BigInt first. + var lhs_space: Value.BigIntSpace = undefined; + var rhs_space: Value.BigIntSpace = undefined; + const lhs_bigint = lhs_val.toBigInt(&lhs_space); + const rhs_bigint = rhs_val.toBigInt(&rhs_space); + const limbs = try scope.arena().alloc( + std.math.big.Limb, + std.math.max(lhs_bigint.limbs.len, rhs_bigint.limbs.len) + 1, + ); + var result_bigint = BigIntMutable{ .limbs = limbs, .positive = undefined, .len = undefined }; + result_bigint.add(lhs_bigint, rhs_bigint); + const result_limbs = result_bigint.limbs[0..result_bigint.len]; + + if (!lhs.ty.eql(rhs.ty)) { + return self.fail(scope, inst.base.src, "TODO implement peer type resolution", .{}); + } + + const val_payload = if (result_bigint.positive) blk: { + const val_payload = try scope.arena().create(Value.Payload.IntBigPositive); + val_payload.* = .{ .limbs = result_limbs }; + break :blk &val_payload.base; + } else blk: { + const val_payload = try scope.arena().create(Value.Payload.IntBigNegative); + val_payload.* = .{ .limbs = result_limbs }; + break :blk &val_payload.base; + }; + + return self.constInst(scope, inst.base.src, .{ + .ty = lhs.ty, + .val = Value.initPayload(val_payload), + }); + } + } + } + + return self.fail(scope, inst.base.src, "TODO implement more analyze add", .{}); +} + +fn analyzeInstDeref(self: *Module, scope: *Scope, deref: *zir.Inst.Deref) InnerError!*Inst { + const ptr = try self.resolveInst(scope, deref.positionals.ptr); + return self.analyzeDeref(scope, deref.base.src, ptr, deref.positionals.ptr.src); +} + +fn analyzeDeref(self: *Module, scope: *Scope, src: usize, ptr: *Inst, ptr_src: usize) InnerError!*Inst { + const elem_ty = switch (ptr.ty.zigTypeTag()) { + .Pointer => ptr.ty.elemType(), + else => return self.fail(scope, ptr_src, "expected pointer, found '{}'", .{ptr.ty}), + }; + if (ptr.value()) |val| { + return self.constInst(scope, src, .{ + .ty = elem_ty, + .val = try val.pointerDeref(scope.arena()), + }); + } + + return self.fail(scope, src, "TODO implement runtime deref", .{}); +} + +fn analyzeInstAsm(self: *Module, scope: *Scope, assembly: *zir.Inst.Asm) InnerError!*Inst { + const return_type = try self.resolveType(scope, assembly.positionals.return_type); + const asm_source = try self.resolveConstString(scope, assembly.positionals.asm_source); + const output = if (assembly.kw_args.output) |o| try self.resolveConstString(scope, o) else null; + + const inputs = try scope.arena().alloc([]const u8, assembly.kw_args.inputs.len); + const clobbers = try scope.arena().alloc([]const u8, assembly.kw_args.clobbers.len); + const args = try scope.arena().alloc(*Inst, assembly.kw_args.args.len); + + for (inputs) |*elem, i| { + elem.* = try self.resolveConstString(scope, assembly.kw_args.inputs[i]); + } + for (clobbers) |*elem, i| { + elem.* = try self.resolveConstString(scope, assembly.kw_args.clobbers[i]); + } + for (args) |*elem, i| { + const arg = try self.resolveInst(scope, assembly.kw_args.args[i]); + elem.* = try self.coerce(scope, Type.initTag(.usize), arg); + } + + const b = try self.requireRuntimeBlock(scope, assembly.base.src); + return self.addNewInstArgs(b, assembly.base.src, return_type, Inst.Assembly, Inst.Args(Inst.Assembly){ + .asm_source = asm_source, + .is_volatile = assembly.kw_args.@"volatile", + .output = output, + .inputs = inputs, + .clobbers = clobbers, + .args = args, + }); +} + +fn analyzeInstCmp(self: *Module, scope: *Scope, inst: *zir.Inst.Cmp) InnerError!*Inst { + const lhs = try self.resolveInst(scope, inst.positionals.lhs); + const rhs = try self.resolveInst(scope, inst.positionals.rhs); + const op = inst.positionals.op; + + const is_equality_cmp = switch (op) { + .eq, .neq => true, + else => false, + }; + const lhs_ty_tag = lhs.ty.zigTypeTag(); + const rhs_ty_tag = rhs.ty.zigTypeTag(); + if (is_equality_cmp and lhs_ty_tag == .Null and rhs_ty_tag == .Null) { + // null == null, null != null + return self.constBool(scope, inst.base.src, op == .eq); + } else if (is_equality_cmp and + ((lhs_ty_tag == .Null and rhs_ty_tag == .Optional) or + rhs_ty_tag == .Null and lhs_ty_tag == .Optional)) + { + // comparing null with optionals + const opt_operand = if (lhs_ty_tag == .Optional) lhs else rhs; + if (opt_operand.value()) |opt_val| { + const is_null = opt_val.isNull(); + return self.constBool(scope, inst.base.src, if (op == .eq) is_null else !is_null); + } + const b = try self.requireRuntimeBlock(scope, inst.base.src); + switch (op) { + .eq => return self.addNewInstArgs( + b, + inst.base.src, + Type.initTag(.bool), + Inst.IsNull, + Inst.Args(Inst.IsNull){ .operand = opt_operand }, + ), + .neq => return self.addNewInstArgs( + b, + inst.base.src, + Type.initTag(.bool), + Inst.IsNonNull, + Inst.Args(Inst.IsNonNull){ .operand = opt_operand }, + ), + else => unreachable, + } + } else if (is_equality_cmp and + ((lhs_ty_tag == .Null and rhs.ty.isCPtr()) or (rhs_ty_tag == .Null and lhs.ty.isCPtr()))) + { + return self.fail(scope, inst.base.src, "TODO implement C pointer cmp", .{}); + } else if (lhs_ty_tag == .Null or rhs_ty_tag == .Null) { + const non_null_type = if (lhs_ty_tag == .Null) rhs.ty else lhs.ty; + return self.fail(scope, inst.base.src, "comparison of '{}' with null", .{non_null_type}); + } else if (is_equality_cmp and + ((lhs_ty_tag == .EnumLiteral and rhs_ty_tag == .Union) or + (rhs_ty_tag == .EnumLiteral and lhs_ty_tag == .Union))) + { + return self.fail(scope, inst.base.src, "TODO implement equality comparison between a union's tag value and an enum literal", .{}); + } else if (lhs_ty_tag == .ErrorSet and rhs_ty_tag == .ErrorSet) { + if (!is_equality_cmp) { + return self.fail(scope, inst.base.src, "{} operator not allowed for errors", .{@tagName(op)}); + } + return self.fail(scope, inst.base.src, "TODO implement equality comparison between errors", .{}); + } else if (lhs.ty.isNumeric() and rhs.ty.isNumeric()) { + // This operation allows any combination of integer and float types, regardless of the + // signed-ness, comptime-ness, and bit-width. So peer type resolution is incorrect for + // numeric types. + return self.cmpNumeric(scope, inst.base.src, lhs, rhs, op); + } + return self.fail(scope, inst.base.src, "TODO implement more cmp analysis", .{}); +} + +fn analyzeInstIsNull(self: *Module, scope: *Scope, inst: *zir.Inst.IsNull) InnerError!*Inst { + const operand = try self.resolveInst(scope, inst.positionals.operand); + return self.analyzeIsNull(scope, inst.base.src, operand, true); +} + +fn analyzeInstIsNonNull(self: *Module, scope: *Scope, inst: *zir.Inst.IsNonNull) InnerError!*Inst { + const operand = try self.resolveInst(scope, inst.positionals.operand); + return self.analyzeIsNull(scope, inst.base.src, operand, false); +} + +fn analyzeInstCondBr(self: *Module, scope: *Scope, inst: *zir.Inst.CondBr) InnerError!*Inst { + const uncasted_cond = try self.resolveInst(scope, inst.positionals.condition); + const cond = try self.coerce(scope, Type.initTag(.bool), uncasted_cond); + + if (try self.resolveDefinedValue(scope, cond)) |cond_val| { + const body = if (cond_val.toBool()) &inst.positionals.true_body else &inst.positionals.false_body; + try self.analyzeBody(scope, body.*); + return self.constVoid(scope, inst.base.src); + } + + const parent_block = try self.requireRuntimeBlock(scope, inst.base.src); + + var true_block: Scope.Block = .{ + .func = parent_block.func, + .decl = parent_block.decl, + .instructions = .{}, + .arena = parent_block.arena, + }; + defer true_block.instructions.deinit(self.allocator); + try self.analyzeBody(&true_block.base, inst.positionals.true_body); + + var false_block: Scope.Block = .{ + .func = parent_block.func, + .decl = parent_block.decl, + .instructions = .{}, + .arena = parent_block.arena, + }; + defer false_block.instructions.deinit(self.allocator); + try self.analyzeBody(&false_block.base, inst.positionals.false_body); + + return self.addNewInstArgs(parent_block, inst.base.src, Type.initTag(.void), Inst.CondBr, Inst.Args(Inst.CondBr){ + .condition = cond, + .true_body = .{ .instructions = try scope.arena().dupe(*Inst, true_block.instructions.items) }, + .false_body = .{ .instructions = try scope.arena().dupe(*Inst, false_block.instructions.items) }, + }); +} + +fn wantSafety(self: *Module, scope: *Scope) bool { + return switch (self.optimize_mode) { + .Debug => true, + .ReleaseSafe => true, + .ReleaseFast => false, + .ReleaseSmall => false, + }; +} + +fn analyzeInstUnreachable(self: *Module, scope: *Scope, unreach: *zir.Inst.Unreachable) InnerError!*Inst { + const b = try self.requireRuntimeBlock(scope, unreach.base.src); + if (self.wantSafety(scope)) { + // TODO Once we have a panic function to call, call it here instead of this. + _ = try self.addNewInstArgs(b, unreach.base.src, Type.initTag(.void), Inst.Breakpoint, {}); + } + return self.addNewInstArgs(b, unreach.base.src, Type.initTag(.noreturn), Inst.Unreach, {}); +} + +fn analyzeInstRet(self: *Module, scope: *Scope, inst: *zir.Inst.Return) InnerError!*Inst { + const b = try self.requireRuntimeBlock(scope, inst.base.src); + return self.addNewInstArgs(b, inst.base.src, Type.initTag(.noreturn), Inst.Ret, {}); +} + +fn analyzeBody(self: *Module, scope: *Scope, body: zir.Module.Body) !void { + if (scope.cast(Scope.Block)) |b| { + const analysis = b.func.analysis.in_progress; + analysis.needed_inst_capacity += body.instructions.len; + try analysis.inst_table.ensureCapacity(analysis.needed_inst_capacity); + for (body.instructions) |src_inst| { + const new_inst = try self.analyzeInst(scope, src_inst); + analysis.inst_table.putAssumeCapacityNoClobber(src_inst, new_inst); + } + } else { + for (body.instructions) |src_inst| { + _ = try self.analyzeInst(scope, src_inst); + } + } +} + +fn analyzeIsNull( + self: *Module, + scope: *Scope, + src: usize, + operand: *Inst, + invert_logic: bool, +) InnerError!*Inst { + return self.fail(scope, src, "TODO implement analysis of isnull and isnotnull", .{}); +} + +/// Asserts that lhs and rhs types are both numeric. +fn cmpNumeric( + self: *Module, + scope: *Scope, + src: usize, + lhs: *Inst, + rhs: *Inst, + op: std.math.CompareOperator, +) !*Inst { + assert(lhs.ty.isNumeric()); + assert(rhs.ty.isNumeric()); + + const lhs_ty_tag = lhs.ty.zigTypeTag(); + const rhs_ty_tag = rhs.ty.zigTypeTag(); + + if (lhs_ty_tag == .Vector and rhs_ty_tag == .Vector) { + if (lhs.ty.arrayLen() != rhs.ty.arrayLen()) { + return self.fail(scope, src, "vector length mismatch: {} and {}", .{ + lhs.ty.arrayLen(), + rhs.ty.arrayLen(), + }); + } + return self.fail(scope, src, "TODO implement support for vectors in cmpNumeric", .{}); + } else if (lhs_ty_tag == .Vector or rhs_ty_tag == .Vector) { + return self.fail(scope, src, "mixed scalar and vector operands to comparison operator: '{}' and '{}'", .{ + lhs.ty, + rhs.ty, + }); + } + + if (lhs.value()) |lhs_val| { + if (rhs.value()) |rhs_val| { + return self.constBool(scope, src, Value.compare(lhs_val, op, rhs_val)); + } + } + + // TODO handle comparisons against lazy zero values + // Some values can be compared against zero without being runtime known or without forcing + // a full resolution of their value, for example `@sizeOf(@Frame(function))` is known to + // always be nonzero, and we benefit from not forcing the full evaluation and stack frame layout + // of this function if we don't need to. + + // It must be a runtime comparison. + const b = try self.requireRuntimeBlock(scope, src); + // For floats, emit a float comparison instruction. + const lhs_is_float = switch (lhs_ty_tag) { + .Float, .ComptimeFloat => true, + else => false, + }; + const rhs_is_float = switch (rhs_ty_tag) { + .Float, .ComptimeFloat => true, + else => false, + }; + if (lhs_is_float and rhs_is_float) { + // Implicit cast the smaller one to the larger one. + const dest_type = x: { + if (lhs_ty_tag == .ComptimeFloat) { + break :x rhs.ty; + } else if (rhs_ty_tag == .ComptimeFloat) { + break :x lhs.ty; + } + if (lhs.ty.floatBits(self.target()) >= rhs.ty.floatBits(self.target())) { + break :x lhs.ty; + } else { + break :x rhs.ty; + } + }; + const casted_lhs = try self.coerce(scope, dest_type, lhs); + const casted_rhs = try self.coerce(scope, dest_type, rhs); + return self.addNewInstArgs(b, src, dest_type, Inst.Cmp, Inst.Args(Inst.Cmp){ + .lhs = casted_lhs, + .rhs = casted_rhs, + .op = op, + }); + } + // For mixed unsigned integer sizes, implicit cast both operands to the larger integer. + // For mixed signed and unsigned integers, implicit cast both operands to a signed + // integer with + 1 bit. + // For mixed floats and integers, extract the integer part from the float, cast that to + // a signed integer with mantissa bits + 1, and if there was any non-integral part of the float, + // add/subtract 1. + const lhs_is_signed = if (lhs.value()) |lhs_val| + lhs_val.compareWithZero(.lt) + else + (lhs.ty.isFloat() or lhs.ty.isSignedInt()); + const rhs_is_signed = if (rhs.value()) |rhs_val| + rhs_val.compareWithZero(.lt) + else + (rhs.ty.isFloat() or rhs.ty.isSignedInt()); + const dest_int_is_signed = lhs_is_signed or rhs_is_signed; + + var dest_float_type: ?Type = null; + + var lhs_bits: usize = undefined; + if (lhs.value()) |lhs_val| { + if (lhs_val.isUndef()) + return self.constUndef(scope, src, Type.initTag(.bool)); + const is_unsigned = if (lhs_is_float) x: { + var bigint_space: Value.BigIntSpace = undefined; + var bigint = try lhs_val.toBigInt(&bigint_space).toManaged(self.allocator); + defer bigint.deinit(); + const zcmp = lhs_val.orderAgainstZero(); + if (lhs_val.floatHasFraction()) { + switch (op) { + .eq => return self.constBool(scope, src, false), + .neq => return self.constBool(scope, src, true), + else => {}, + } + if (zcmp == .lt) { + try bigint.addScalar(bigint.toConst(), -1); + } else { + try bigint.addScalar(bigint.toConst(), 1); + } + } + lhs_bits = bigint.toConst().bitCountTwosComp(); + break :x (zcmp != .lt); + } else x: { + lhs_bits = lhs_val.intBitCountTwosComp(); + break :x (lhs_val.orderAgainstZero() != .lt); + }; + lhs_bits += @boolToInt(is_unsigned and dest_int_is_signed); + } else if (lhs_is_float) { + dest_float_type = lhs.ty; + } else { + const int_info = lhs.ty.intInfo(self.target()); + lhs_bits = int_info.bits + @boolToInt(!int_info.signed and dest_int_is_signed); + } + + var rhs_bits: usize = undefined; + if (rhs.value()) |rhs_val| { + if (rhs_val.isUndef()) + return self.constUndef(scope, src, Type.initTag(.bool)); + const is_unsigned = if (rhs_is_float) x: { + var bigint_space: Value.BigIntSpace = undefined; + var bigint = try rhs_val.toBigInt(&bigint_space).toManaged(self.allocator); + defer bigint.deinit(); + const zcmp = rhs_val.orderAgainstZero(); + if (rhs_val.floatHasFraction()) { + switch (op) { + .eq => return self.constBool(scope, src, false), + .neq => return self.constBool(scope, src, true), + else => {}, + } + if (zcmp == .lt) { + try bigint.addScalar(bigint.toConst(), -1); + } else { + try bigint.addScalar(bigint.toConst(), 1); + } + } + rhs_bits = bigint.toConst().bitCountTwosComp(); + break :x (zcmp != .lt); + } else x: { + rhs_bits = rhs_val.intBitCountTwosComp(); + break :x (rhs_val.orderAgainstZero() != .lt); + }; + rhs_bits += @boolToInt(is_unsigned and dest_int_is_signed); + } else if (rhs_is_float) { + dest_float_type = rhs.ty; + } else { + const int_info = rhs.ty.intInfo(self.target()); + rhs_bits = int_info.bits + @boolToInt(!int_info.signed and dest_int_is_signed); + } + + const dest_type = if (dest_float_type) |ft| ft else blk: { + const max_bits = std.math.max(lhs_bits, rhs_bits); + const casted_bits = std.math.cast(u16, max_bits) catch |err| switch (err) { + error.Overflow => return self.fail(scope, src, "{} exceeds maximum integer bit count", .{max_bits}), + }; + break :blk try self.makeIntType(scope, dest_int_is_signed, casted_bits); + }; + const casted_lhs = try self.coerce(scope, dest_type, lhs); + const casted_rhs = try self.coerce(scope, dest_type, lhs); + + return self.addNewInstArgs(b, src, dest_type, Inst.Cmp, Inst.Args(Inst.Cmp){ + .lhs = casted_lhs, + .rhs = casted_rhs, + .op = op, + }); +} + +fn makeIntType(self: *Module, scope: *Scope, signed: bool, bits: u16) !Type { + if (signed) { + const int_payload = try scope.arena().create(Type.Payload.IntSigned); + int_payload.* = .{ .bits = bits }; + return Type.initPayload(&int_payload.base); + } else { + const int_payload = try scope.arena().create(Type.Payload.IntUnsigned); + int_payload.* = .{ .bits = bits }; + return Type.initPayload(&int_payload.base); + } +} + +fn coerce(self: *Module, scope: *Scope, dest_type: Type, inst: *Inst) !*Inst { + // If the types are the same, we can return the operand. + if (dest_type.eql(inst.ty)) + return inst; + + const in_memory_result = coerceInMemoryAllowed(dest_type, inst.ty); + if (in_memory_result == .ok) { + return self.bitcast(scope, dest_type, inst); + } + + // *[N]T to []T + if (inst.ty.isSinglePointer() and dest_type.isSlice() and + (!inst.ty.pointerIsConst() or dest_type.pointerIsConst())) + { + const array_type = inst.ty.elemType(); + const dst_elem_type = dest_type.elemType(); + if (array_type.zigTypeTag() == .Array and + coerceInMemoryAllowed(dst_elem_type, array_type.elemType()) == .ok) + { + return self.coerceArrayPtrToSlice(scope, dest_type, inst); + } + } + + // comptime_int to fixed-width integer + if (inst.ty.zigTypeTag() == .ComptimeInt and dest_type.zigTypeTag() == .Int) { + // The representation is already correct; we only need to make sure it fits in the destination type. + const val = inst.value().?; // comptime_int always has comptime known value + if (!val.intFitsInType(dest_type, self.target())) { + return self.fail(scope, inst.src, "type {} cannot represent integer value {}", .{ inst.ty, val }); + } + return self.constInst(scope, inst.src, .{ .ty = dest_type, .val = val }); + } + + // integer widening + if (inst.ty.zigTypeTag() == .Int and dest_type.zigTypeTag() == .Int) { + const src_info = inst.ty.intInfo(self.target()); + const dst_info = dest_type.intInfo(self.target()); + if (src_info.signed == dst_info.signed and dst_info.bits >= src_info.bits) { + if (inst.value()) |val| { + return self.constInst(scope, inst.src, .{ .ty = dest_type, .val = val }); + } else { + return self.fail(scope, inst.src, "TODO implement runtime integer widening", .{}); + } + } else { + return self.fail(scope, inst.src, "TODO implement more int widening {} to {}", .{ inst.ty, dest_type }); + } + } + + return self.fail(scope, inst.src, "TODO implement type coercion from {} to {}", .{ inst.ty, dest_type }); +} + +fn bitcast(self: *Module, scope: *Scope, dest_type: Type, inst: *Inst) !*Inst { + if (inst.value()) |val| { + // Keep the comptime Value representation; take the new type. + return self.constInst(scope, inst.src, .{ .ty = dest_type, .val = val }); + } + // TODO validate the type size and other compile errors + const b = try self.requireRuntimeBlock(scope, inst.src); + return self.addNewInstArgs(b, inst.src, dest_type, Inst.BitCast, Inst.Args(Inst.BitCast){ .operand = inst }); +} + +fn coerceArrayPtrToSlice(self: *Module, scope: *Scope, dest_type: Type, inst: *Inst) !*Inst { + if (inst.value()) |val| { + // The comptime Value representation is compatible with both types. + return self.constInst(scope, inst.src, .{ .ty = dest_type, .val = val }); + } + return self.fail(scope, inst.src, "TODO implement coerceArrayPtrToSlice runtime instruction", .{}); +} + +fn fail(self: *Module, scope: *Scope, src: usize, comptime format: []const u8, args: var) InnerError { + @setCold(true); + try self.failed_decls.ensureCapacity(self.failed_decls.size + 1); + try self.failed_files.ensureCapacity(self.failed_files.size + 1); + const err_msg = try ErrorMsg.create(self.allocator, src, format, args); + switch (scope.tag) { + .decl => { + const decl = scope.cast(Scope.DeclAnalysis).?.decl; + switch (decl.analysis) { + .initial_in_progress => decl.analysis = .initial_sema_failure, + .repeat_in_progress => decl.analysis = .repeat_sema_failure, + else => unreachable, + } + self.failed_decls.putAssumeCapacityNoClobber(decl, err_msg); + }, + .block => { + const block = scope.cast(Scope.Block).?; + block.func.analysis = .sema_failure; + self.failed_decls.putAssumeCapacityNoClobber(block.decl, err_msg); + }, + .zir_module => { + const zir_module = scope.cast(Scope.ZIRModule).?; + zir_module.status = .loaded_sema_failure; + self.failed_files.putAssumeCapacityNoClobber(zir_module, err_msg); + }, + } + return error.AnalysisFail; +} + +const InMemoryCoercionResult = enum { + ok, + no_match, +}; + +fn coerceInMemoryAllowed(dest_type: Type, src_type: Type) InMemoryCoercionResult { + if (dest_type.eql(src_type)) + return .ok; + + // TODO: implement more of this function + + return .no_match; +} + +pub const ErrorMsg = struct { + byte_offset: usize, + msg: []const u8, + + pub fn create(allocator: *Allocator, byte_offset: usize, comptime format: []const u8, args: var) !*ErrorMsg { + const self = try allocator.create(ErrorMsg); + errdefer allocator.destroy(self); + self.* = try init(allocator, byte_offset, format, args); + return self; + } + + /// Assumes the ErrorMsg struct and msg were both allocated with allocator. + pub fn destroy(self: *ErrorMsg, allocator: *Allocator) void { + self.deinit(allocator); + allocator.destroy(self); + } + + pub fn init(allocator: *Allocator, byte_offset: usize, comptime format: []const u8, args: var) !ErrorMsg { + return ErrorMsg{ + .byte_offset = byte_offset, + .msg = try std.fmt.allocPrint(allocator, format, args), + }; + } + + pub fn deinit(self: *ErrorMsg, allocator: *Allocator) void { + allocator.free(self.msg); + self.* = undefined; + } +}; diff --git a/src-self-hosted/codegen.zig b/src-self-hosted/codegen.zig @@ -6,6 +6,8 @@ const Type = @import("type.zig").Type; const Value = @import("value.zig").Value; const TypedValue = @import("TypedValue.zig"); const link = @import("link.zig"); +const Module = @import("Module.zig"); +const ErrorMsg = Module.ErrorMsg; const Target = std.Target; const Allocator = mem.Allocator; @@ -14,7 +16,7 @@ pub const Result = union(enum) { appended: void, /// The value is available externally, `code` is unused. externally_managed: []const u8, - fail: *ir.ErrorMsg, + fail: *Module.ErrorMsg, }; pub fn generateSymbol( @@ -77,7 +79,7 @@ pub fn generateSymbol( } } return Result{ - .fail = try ir.ErrorMsg.create( + .fail = try ErrorMsg.create( bin_file.allocator, src, "TODO implement generateSymbol for more kinds of arrays", @@ -107,7 +109,7 @@ pub fn generateSymbol( return Result{ .appended = {} }; } return Result{ - .fail = try ir.ErrorMsg.create( + .fail = try ErrorMsg.create( bin_file.allocator, src, "TODO implement generateSymbol for pointer {}", @@ -123,7 +125,7 @@ pub fn generateSymbol( return Result{ .appended = {} }; } return Result{ - .fail = try ir.ErrorMsg.create( + .fail = try ErrorMsg.create( bin_file.allocator, src, "TODO implement generateSymbol for int type '{}'", @@ -133,7 +135,7 @@ pub fn generateSymbol( }, else => |t| { return Result{ - .fail = try ir.ErrorMsg.create( + .fail = try ErrorMsg.create( bin_file.allocator, src, "TODO implement generateSymbol for type '{}'", @@ -147,10 +149,10 @@ pub fn generateSymbol( const Function = struct { bin_file: *link.ElfFile, target: *const std.Target, - mod_fn: *const ir.Module.Fn, + mod_fn: *const Module.Fn, code: *std.ArrayList(u8), inst_table: std.AutoHashMap(*ir.Inst, MCValue), - err_msg: ?*ir.ErrorMsg, + err_msg: ?*ErrorMsg, const MCValue = union(enum) { none, @@ -570,7 +572,7 @@ const Function = struct { fn fail(self: *Function, src: usize, comptime format: []const u8, args: var) error{ CodegenFail, OutOfMemory } { @setCold(true); assert(self.err_msg == null); - self.err_msg = try ir.ErrorMsg.create(self.code.allocator, src, format, args); + self.err_msg = try ErrorMsg.create(self.code.allocator, src, format, args); return error.CodegenFail; } }; diff --git a/src-self-hosted/ir.zig b/src-self-hosted/ir.zig @@ -1,20 +1,9 @@ const std = @import("std"); -const mem = std.mem; -const Allocator = std.mem.Allocator; -const ArrayListUnmanaged = std.ArrayListUnmanaged; const Value = @import("value.zig").Value; const Type = @import("type.zig").Type; -const TypedValue = @import("TypedValue.zig"); -const assert = std.debug.assert; -const BigIntConst = std.math.big.int.Const; -const BigIntMutable = std.math.big.int.Mutable; -const Target = std.Target; -const Package = @import("Package.zig"); -const link = @import("link.zig"); +const Module = @import("Module.zig"); -pub const text = @import("ir/text.zig"); - -/// These are in-memory, analyzed instructions. See `text.Inst` for the representation +/// These are in-memory, analyzed instructions. See `zir.Inst` for the representation /// of instructions that correspond to the ZIR text format. /// This struct owns the `Value` and `Type` memory. When the struct is deallocated, /// so are the `Value` and `Type`. The value of a constant must be copied into @@ -166,2006 +155,3 @@ pub const Inst = struct { args: void, }; }; - -pub const Module = struct { - /// General-purpose allocator. - allocator: *Allocator, - /// Module owns this resource. - root_pkg: *Package, - /// Module owns this resource. - root_scope: *Scope.ZIRModule, - /// Pointer to externally managed resource. - bin_file: *link.ElfFile, - /// It's rare for a decl to be exported, so we save memory by having a sparse map of - /// Decl pointers to details about them being exported. - /// The Export memory is owned by the `export_owners` table; the slice itself is owned by this table. - decl_exports: std.AutoHashMap(*Decl, []*Export), - /// This models the Decls that perform exports, so that `decl_exports` can be updated when a Decl - /// is modified. Note that the key of this table is not the Decl being exported, but the Decl that - /// is performing the export of another Decl. - /// This table owns the Export memory. - export_owners: std.AutoHashMap(*Decl, []*Export), - /// Maps fully qualified namespaced names to the Decl struct for them. - decl_table: std.AutoHashMap(Decl.Hash, *Decl), - - optimize_mode: std.builtin.Mode, - link_error_flags: link.ElfFile.ErrorFlags = link.ElfFile.ErrorFlags{}, - - work_queue: std.fifo.LinearFifo(WorkItem, .Dynamic), - - /// We optimize memory usage for a compilation with no compile errors by storing the - /// error messages and mapping outside of `Decl`. - /// The ErrorMsg memory is owned by the decl, using Module's allocator. - /// Note that a Decl can succeed but the Fn it represents can fail. In this case, - /// a Decl can have a failed_decls entry but have analysis status of success. - failed_decls: std.AutoHashMap(*Decl, *ErrorMsg), - /// Using a map here for consistency with the other fields here. - /// The ErrorMsg memory is owned by the `Scope.ZIRModule`, using Module's allocator. - failed_files: std.AutoHashMap(*Scope.ZIRModule, *ErrorMsg), - /// Using a map here for consistency with the other fields here. - /// The ErrorMsg memory is owned by the `Export`, using Module's allocator. - failed_exports: std.AutoHashMap(*Export, *ErrorMsg), - - pub const WorkItem = union(enum) { - /// Write the machine code for a Decl to the output file. - codegen_decl: *Decl, - }; - - pub const Export = struct { - options: std.builtin.ExportOptions, - /// Byte offset into the file that contains the export directive. - src: usize, - /// Represents the position of the export, if any, in the output file. - link: link.ElfFile.Export, - /// The Decl that performs the export. Note that this is *not* the Decl being exported. - owner_decl: *Decl, - status: enum { - in_progress, - failed, - /// Indicates that the failure was due to a temporary issue, such as an I/O error - /// when writing to the output file. Retrying the export may succeed. - failed_retryable, - complete, - }, - }; - - pub const Decl = struct { - /// This name is relative to the containing namespace of the decl. It uses a null-termination - /// to save bytes, since there can be a lot of decls in a compilation. The null byte is not allowed - /// in symbol names, because executable file formats use null-terminated strings for symbol names. - /// All Decls have names, even values that are not bound to a zig namespace. This is necessary for - /// mapping them to an address in the output file. - /// Memory owned by this decl, using Module's allocator. - name: [*:0]const u8, - /// The direct parent container of the Decl. This field will need to get more fleshed out when - /// self-hosted supports proper struct types and Zig AST => ZIR. - /// Reference to externally owned memory. - scope: *Scope.ZIRModule, - /// Byte offset into the source file that contains this declaration. - /// This is the base offset that src offsets within this Decl are relative to. - src: usize, - /// The most recent value of the Decl after a successful semantic analysis. - /// The tag for this union is determined by the tag value of the analysis field. - typed_value: union { - never_succeeded: void, - most_recent: TypedValue.Managed, - }, - /// Represents the "shallow" analysis status. For example, for decls that are functions, - /// the function type is analyzed with this set to `in_progress`, however, the semantic - /// analysis of the function body is performed with this value set to `success`. Functions - /// have their own analysis status field. - analysis: enum { - initial_in_progress, - /// This Decl might be OK but it depends on another one which did not successfully complete - /// semantic analysis. This Decl never had a value computed. - initial_dependency_failure, - /// Semantic analysis failure. This Decl never had a value computed. - /// There will be a corresponding ErrorMsg in Module.failed_decls. - initial_sema_failure, - /// In this case the `typed_value.most_recent` can still be accessed. - /// There will be a corresponding ErrorMsg in Module.failed_decls. - codegen_failure, - /// In this case the `typed_value.most_recent` can still be accessed. - /// There will be a corresponding ErrorMsg in Module.failed_decls. - /// This indicates the failure was something like running out of disk space, - /// and attempting codegen again may succeed. - codegen_failure_retryable, - /// This Decl might be OK but it depends on another one which did not successfully complete - /// semantic analysis. There is a most recent value available. - repeat_dependency_failure, - /// Semantic anlaysis failure, but the `typed_value.most_recent` can be accessed. - /// There will be a corresponding ErrorMsg in Module.failed_decls. - repeat_sema_failure, - /// Completed successfully before; the `typed_value.most_recent` can be accessed, and - /// new semantic analysis is in progress. - repeat_in_progress, - /// Everything is done and updated. - complete, - }, - - /// Represents the position of the code in the output file. - /// This is populated regardless of semantic analysis and code generation. - link: link.ElfFile.Decl = link.ElfFile.Decl.empty, - - /// The shallow set of other decls whose typed_value could possibly change if this Decl's - /// typed_value is modified. - /// TODO look into using a lightweight map/set data structure rather than a linear array. - dependants: ArrayListUnmanaged(*Decl) = ArrayListUnmanaged(*Decl){}, - - contents_hash: Hash, - - pub fn destroy(self: *Decl, allocator: *Allocator) void { - allocator.free(mem.spanZ(self.name)); - if (self.typedValueManaged()) |tvm| { - tvm.deinit(allocator); - } - allocator.destroy(self); - } - - pub const Hash = [16]u8; - - /// If the name is small enough, it is used directly as the hash. - /// If it is long, blake3 hash is computed. - pub fn hashSimpleName(name: []const u8) Hash { - var out: Hash = undefined; - if (name.len <= Hash.len) { - mem.copy(u8, &out, name); - mem.set(u8, out[name.len..], 0); - } else { - std.crypto.Blake3.hash(name, &out); - } - return out; - } - - /// Must generate unique bytes with no collisions with other decls. - /// The point of hashing here is only to limit the number of bytes of - /// the unique identifier to a fixed size (16 bytes). - pub fn fullyQualifiedNameHash(self: Decl) Hash { - // Right now we only have ZIRModule as the source. So this is simply the - // relative name of the decl. - return hashSimpleName(mem.spanZ(u8, self.name)); - } - - pub fn typedValue(self: *Decl) error{AnalysisFail}!TypedValue { - const tvm = self.typedValueManaged() orelse return error.AnalysisFail; - return tvm.typed_value; - } - - pub fn value(self: *Decl) error{AnalysisFail}!Value { - return (try self.typedValue()).val; - } - - pub fn dump(self: *Decl) void { - const loc = std.zig.findLineColumn(self.scope.source.bytes, self.src); - std.debug.warn("{}:{}:{} name={} status={}", .{ - self.scope.sub_file_path, - loc.line + 1, - loc.column + 1, - mem.spanZ(self.name), - @tagName(self.analysis), - }); - if (self.typedValueManaged()) |tvm| { - std.debug.warn(" ty={} val={}", .{ tvm.typed_value.ty, tvm.typed_value.val }); - } - std.debug.warn("\n", .{}); - } - - fn typedValueManaged(self: *Decl) ?*TypedValue.Managed { - switch (self.analysis) { - .initial_in_progress, - .initial_dependency_failure, - .initial_sema_failure, - => return null, - .codegen_failure, - .codegen_failure_retryable, - .repeat_dependency_failure, - .repeat_sema_failure, - .repeat_in_progress, - .complete, - => return &self.typed_value.most_recent, - } - } - }; - - /// Fn struct memory is owned by the Decl's TypedValue.Managed arena allocator. - pub const Fn = struct { - /// This memory owned by the Decl's TypedValue.Managed arena allocator. - fn_type: Type, - analysis: union(enum) { - /// The value is the source instruction. - queued: *text.Inst.Fn, - in_progress: *Analysis, - /// There will be a corresponding ErrorMsg in Module.failed_decls - sema_failure, - /// This Fn might be OK but it depends on another Decl which did not successfully complete - /// semantic analysis. - dependency_failure, - success: Body, - }, - - /// This memory is temporary and points to stack memory for the duration - /// of Fn analysis. - pub const Analysis = struct { - inner_block: Scope.Block, - /// TODO Performance optimization idea: instead of this inst_table, - /// use a field in the text.Inst instead to track corresponding instructions - inst_table: std.AutoHashMap(*text.Inst, *Inst), - needed_inst_capacity: usize, - }; - }; - - pub const Scope = struct { - tag: Tag, - - pub fn cast(base: *Scope, comptime T: type) ?*T { - if (base.tag != T.base_tag) - return null; - - return @fieldParentPtr(T, "base", base); - } - - /// Asserts the scope has a parent which is a DeclAnalysis and - /// returns the arena Allocator. - pub fn arena(self: *Scope) *Allocator { - switch (self.tag) { - .block => return self.cast(Block).?.arena, - .decl => return &self.cast(DeclAnalysis).?.arena.allocator, - .zir_module => return &self.cast(ZIRModule).?.contents.module.arena.allocator, - } - } - - /// Asserts the scope has a parent which is a DeclAnalysis and - /// returns the Decl. - pub fn decl(self: *Scope) *Decl { - switch (self.tag) { - .block => return self.cast(Block).?.decl, - .decl => return self.cast(DeclAnalysis).?.decl, - .zir_module => unreachable, - } - } - - /// Asserts the scope has a parent which is a ZIRModule and - /// returns it. - pub fn namespace(self: *Scope) *ZIRModule { - switch (self.tag) { - .block => return self.cast(Block).?.decl.scope, - .decl => return self.cast(DeclAnalysis).?.decl.scope, - .zir_module => return self.cast(ZIRModule).?, - } - } - - pub fn dumpInst(self: *Scope, inst: *Inst) void { - const zir_module = self.namespace(); - const loc = std.zig.findLineColumn(zir_module.source.bytes, inst.src); - std.debug.warn("{}:{}:{}: {}: ty={}\n", .{ - zir_module.sub_file_path, - loc.line + 1, - loc.column + 1, - @tagName(inst.tag), - inst.ty, - }); - } - - pub const Tag = enum { - zir_module, - block, - decl, - }; - - pub const ZIRModule = struct { - pub const base_tag: Tag = .zir_module; - base: Scope = Scope{ .tag = base_tag }, - /// Relative to the owning package's root_src_dir. - /// Reference to external memory, not owned by ZIRModule. - sub_file_path: []const u8, - source: union { - unloaded: void, - bytes: [:0]const u8, - }, - contents: union { - not_available: void, - module: *text.Module, - }, - status: enum { - never_loaded, - unloaded_success, - unloaded_parse_failure, - unloaded_sema_failure, - loaded_parse_failure, - loaded_sema_failure, - loaded_success, - }, - - pub fn unload(self: *ZIRModule, allocator: *Allocator) void { - switch (self.status) { - .never_loaded, - .unloaded_parse_failure, - .unloaded_sema_failure, - .unloaded_success, - => {}, - - .loaded_success => { - allocator.free(self.source.bytes); - self.contents.module.deinit(allocator); - allocator.destroy(self.contents.module); - self.status = .unloaded_success; - }, - .loaded_sema_failure => { - allocator.free(self.source.bytes); - self.contents.module.deinit(allocator); - allocator.destroy(self.contents.module); - self.status = .unloaded_sema_failure; - }, - .loaded_parse_failure => { - allocator.free(self.source.bytes); - self.status = .unloaded_parse_failure; - }, - } - } - - pub fn deinit(self: *ZIRModule, allocator: *Allocator) void { - self.unload(allocator); - self.* = undefined; - } - - pub fn dumpSrc(self: *ZIRModule, src: usize) void { - const loc = std.zig.findLineColumn(self.source.bytes, src); - std.debug.warn("{}:{}:{}\n", .{ self.sub_file_path, loc.line + 1, loc.column + 1 }); - } - }; - - /// This is a temporary structure, references to it are valid only - /// during semantic analysis of the block. - pub const Block = struct { - pub const base_tag: Tag = .block; - base: Scope = Scope{ .tag = base_tag }, - func: *Fn, - decl: *Decl, - instructions: ArrayListUnmanaged(*Inst), - /// Points to the arena allocator of DeclAnalysis - arena: *Allocator, - }; - - /// This is a temporary structure, references to it are valid only - /// during semantic analysis of the decl. - pub const DeclAnalysis = struct { - pub const base_tag: Tag = .decl; - base: Scope = Scope{ .tag = base_tag }, - decl: *Decl, - arena: std.heap.ArenaAllocator, - }; - }; - - pub const Body = struct { - instructions: []*Inst, - }; - - pub const AllErrors = struct { - arena: std.heap.ArenaAllocator.State, - list: []const Message, - - pub const Message = struct { - src_path: []const u8, - line: usize, - column: usize, - byte_offset: usize, - msg: []const u8, - }; - - pub fn deinit(self: *AllErrors, allocator: *Allocator) void { - self.arena.promote(allocator).deinit(); - } - - fn add( - arena: *std.heap.ArenaAllocator, - errors: *std.ArrayList(Message), - sub_file_path: []const u8, - source: []const u8, - simple_err_msg: ErrorMsg, - ) !void { - const loc = std.zig.findLineColumn(source, simple_err_msg.byte_offset); - try errors.append(.{ - .src_path = try arena.allocator.dupe(u8, sub_file_path), - .msg = try arena.allocator.dupe(u8, simple_err_msg.msg), - .byte_offset = simple_err_msg.byte_offset, - .line = loc.line, - .column = loc.column, - }); - } - }; - - pub fn deinit(self: *Module) void { - const allocator = self.allocator; - self.work_queue.deinit(); - { - var it = self.decl_table.iterator(); - while (it.next()) |kv| { - kv.value.destroy(allocator); - } - self.decl_table.deinit(); - } - { - var it = self.failed_decls.iterator(); - while (it.next()) |kv| { - kv.value.destroy(allocator); - } - self.failed_decls.deinit(); - } - { - var it = self.failed_files.iterator(); - while (it.next()) |kv| { - kv.value.destroy(allocator); - } - self.failed_files.deinit(); - } - { - var it = self.failed_exports.iterator(); - while (it.next()) |kv| { - kv.value.destroy(allocator); - } - self.failed_exports.deinit(); - } - { - var it = self.decl_exports.iterator(); - while (it.next()) |kv| { - const export_list = kv.value; - allocator.free(export_list); - } - self.decl_exports.deinit(); - } - { - var it = self.export_owners.iterator(); - while (it.next()) |kv| { - const export_list = kv.value; - for (export_list) |exp| { - allocator.destroy(exp); - } - allocator.free(export_list); - } - self.export_owners.deinit(); - } - self.root_pkg.destroy(); - { - self.root_scope.deinit(allocator); - allocator.destroy(self.root_scope); - } - self.* = undefined; - } - - pub fn target(self: Module) std.Target { - return self.bin_file.options.target; - } - - /// Detect changes to source files, perform semantic analysis, and update the output files. - pub fn update(self: *Module) !void { - // TODO Use the cache hash file system to detect which source files changed. - // Here we simulate a full cache miss. - // Analyze the root source file now. - self.analyzeRoot(self.root_scope) catch |err| switch (err) { - error.AnalysisFail => { - assert(self.totalErrorCount() != 0); - }, - else => |e| return e, - }; - - try self.performAllTheWork(); - - // Unload all the source files from memory. - self.root_scope.unload(self.allocator); - - try self.bin_file.flush(); - self.link_error_flags = self.bin_file.error_flags; - } - - pub fn totalErrorCount(self: *Module) usize { - return self.failed_decls.size + - self.failed_files.size + - self.failed_exports.size + - @boolToInt(self.link_error_flags.no_entry_point_found); - } - - pub fn getAllErrorsAlloc(self: *Module) !AllErrors { - var arena = std.heap.ArenaAllocator.init(self.allocator); - errdefer arena.deinit(); - - var errors = std.ArrayList(AllErrors.Message).init(self.allocator); - defer errors.deinit(); - - { - var it = self.failed_files.iterator(); - while (it.next()) |kv| { - const scope = kv.key; - const err_msg = kv.value; - const source = scope.source.bytes; - try AllErrors.add(&arena, &errors, scope.sub_file_path, source, err_msg.*); - } - } - { - var it = self.failed_decls.iterator(); - while (it.next()) |kv| { - const decl = kv.key; - const err_msg = kv.value; - const source = decl.scope.source.bytes; - try AllErrors.add(&arena, &errors, decl.scope.sub_file_path, source, err_msg.*); - } - } - { - var it = self.failed_exports.iterator(); - while (it.next()) |kv| { - const decl = kv.key.owner_decl; - const err_msg = kv.value; - const source = decl.scope.source.bytes; - try AllErrors.add(&arena, &errors, decl.scope.sub_file_path, source, err_msg.*); - } - } - - if (self.link_error_flags.no_entry_point_found) { - try errors.append(.{ - .src_path = self.root_pkg.root_src_path, - .line = 0, - .column = 0, - .byte_offset = 0, - .msg = try std.fmt.allocPrint(&arena.allocator, "no entry point found", .{}), - }); - } - - assert(errors.items.len == self.totalErrorCount()); - - return AllErrors{ - .arena = arena.state, - .list = try arena.allocator.dupe(AllErrors.Message, errors.items), - }; - } - - const InnerError = error{ OutOfMemory, AnalysisFail }; - - pub fn performAllTheWork(self: *Module) error{OutOfMemory}!void { - while (self.work_queue.readItem()) |work_item| switch (work_item) { - .codegen_decl => |decl| switch (decl.analysis) { - .initial_in_progress, - .repeat_in_progress, - => unreachable, - - .initial_sema_failure, - .repeat_sema_failure, - .codegen_failure, - .initial_dependency_failure, - .repeat_dependency_failure, - => continue, - - .complete, .codegen_failure_retryable => { - if (decl.typed_value.most_recent.typed_value.val.cast(Value.Payload.Function)) |payload| { - switch (payload.func.analysis) { - .queued => self.analyzeFnBody(decl, payload.func) catch |err| switch (err) { - error.AnalysisFail => { - if (payload.func.analysis == .queued) { - payload.func.analysis = .dependency_failure; - } - continue; - }, - else => |e| return e, - }, - .in_progress => unreachable, - .sema_failure, .dependency_failure => continue, - .success => {}, - } - } - - assert(decl.typed_value.most_recent.typed_value.ty.hasCodeGenBits()); - - self.bin_file.updateDecl(self, decl) catch |err| switch (err) { - error.OutOfMemory => return error.OutOfMemory, - error.AnalysisFail => { - decl.analysis = .repeat_dependency_failure; - }, - else => { - try self.failed_decls.ensureCapacity(self.failed_decls.size + 1); - self.failed_decls.putAssumeCapacityNoClobber(decl, try ErrorMsg.create( - self.allocator, - decl.src, - "unable to codegen: {}", - .{@errorName(err)}, - )); - decl.analysis = .codegen_failure_retryable; - }, - }; - }, - }, - }; - } - - fn getTextModule(self: *Module, root_scope: *Scope.ZIRModule) !*text.Module { - switch (root_scope.status) { - .never_loaded, .unloaded_success => { - try self.failed_files.ensureCapacity(self.failed_files.size + 1); - - var keep_source = false; - const source = try self.root_pkg.root_src_dir.readFileAllocOptions( - self.allocator, - self.root_pkg.root_src_path, - std.math.maxInt(u32), - 1, - 0, - ); - defer if (!keep_source) self.allocator.free(source); - - var keep_zir_module = false; - const zir_module = try self.allocator.create(text.Module); - defer if (!keep_zir_module) self.allocator.destroy(zir_module); - - zir_module.* = try text.parse(self.allocator, source); - defer if (!keep_zir_module) zir_module.deinit(self.allocator); - - if (zir_module.error_msg) |src_err_msg| { - self.failed_files.putAssumeCapacityNoClobber( - root_scope, - try ErrorMsg.create(self.allocator, src_err_msg.byte_offset, "{}", .{src_err_msg.msg}), - ); - root_scope.status = .loaded_parse_failure; - root_scope.source = .{ .bytes = source }; - keep_source = true; - return error.AnalysisFail; - } - - root_scope.status = .loaded_success; - root_scope.source = .{ .bytes = source }; - keep_source = true; - root_scope.contents = .{ .module = zir_module }; - keep_zir_module = true; - - return zir_module; - }, - - .unloaded_parse_failure, - .unloaded_sema_failure, - .loaded_parse_failure, - .loaded_sema_failure, - => return error.AnalysisFail, - .loaded_success => return root_scope.contents.module, - } - } - - fn analyzeRoot(self: *Module, root_scope: *Scope.ZIRModule) !void { - // TODO use the cache to identify, from the modified source files, the decls which have - // changed based on the span of memory that represents the decl in the re-parsed source file. - // Use the cached dependency graph to recursively determine the set of decls which need - // regeneration. - // Here we simulate adding a source file which was previously not part of the compilation, - // which means scanning the decls looking for exports. - // TODO also identify decls that need to be deleted. - switch (root_scope.status) { - .never_loaded => { - const src_module = try self.getTextModule(root_scope); - - // Here we ensure enough queue capacity to store all the decls, so that later we can use - // appendAssumeCapacity. - try self.work_queue.ensureUnusedCapacity(src_module.decls.len); - - for (src_module.decls) |decl| { - if (decl.cast(text.Inst.Export)) |export_inst| { - _ = try self.resolveDecl(&root_scope.base, &export_inst.base, link.ElfFile.Decl.empty); - } - } - }, - - .unloaded_parse_failure, - .unloaded_sema_failure, - .loaded_parse_failure, - .loaded_sema_failure, - .loaded_success, - .unloaded_success, - => { - const src_module = try self.getTextModule(root_scope); - - // Look for changed decls. - for (src_module.decls) |src_decl| { - const name_hash = Decl.hashSimpleName(src_decl.name); - if (self.decl_table.get(name_hash)) |kv| { - const decl = kv.value; - const new_contents_hash = Decl.hashSimpleName(src_decl.contents); - if (!mem.eql(u8, &new_contents_hash, &decl.contents_hash)) { - // TODO recursive dependency management - std.debug.warn("noticed that '{}' changed\n", .{src_decl.name}); - self.decl_table.removeAssertDiscard(name_hash); - const saved_link = decl.link; - decl.destroy(self.allocator); - if (self.export_owners.getValue(decl)) |exports| { - @panic("TODO handle updating a decl that does an export"); - } - const new_decl = self.resolveDecl( - &root_scope.base, - src_decl, - saved_link, - ) catch |err| switch (err) { - error.OutOfMemory => return error.OutOfMemory, - error.AnalysisFail => continue, - }; - if (self.decl_exports.remove(decl)) |entry| { - self.decl_exports.putAssumeCapacityNoClobber(new_decl, entry.value); - } - } - } else if (src_decl.cast(text.Inst.Export)) |export_inst| { - _ = try self.resolveDecl(&root_scope.base, &export_inst.base, link.ElfFile.Decl.empty); - } - } - }, - } - } - - fn analyzeFnBody(self: *Module, decl: *Decl, func: *Fn) !void { - // Use the Decl's arena for function memory. - var arena = decl.typed_value.most_recent.arena.?.promote(self.allocator); - defer decl.typed_value.most_recent.arena.?.* = arena.state; - var analysis: Fn.Analysis = .{ - .inner_block = .{ - .func = func, - .decl = decl, - .instructions = .{}, - .arena = &arena.allocator, - }, - .needed_inst_capacity = 0, - .inst_table = std.AutoHashMap(*text.Inst, *Inst).init(self.allocator), - }; - defer analysis.inner_block.instructions.deinit(self.allocator); - defer analysis.inst_table.deinit(); - - const fn_inst = func.analysis.queued; - func.analysis = .{ .in_progress = &analysis }; - - try self.analyzeBody(&analysis.inner_block.base, fn_inst.positionals.body); - - func.analysis = .{ - .success = .{ - .instructions = try arena.allocator.dupe(*Inst, analysis.inner_block.instructions.items), - }, - }; - } - - fn resolveDecl( - self: *Module, - scope: *Scope, - old_inst: *text.Inst, - bin_file_link: link.ElfFile.Decl, - ) InnerError!*Decl { - const hash = Decl.hashSimpleName(old_inst.name); - if (self.decl_table.get(hash)) |kv| { - return kv.value; - } else { - const new_decl = blk: { - try self.decl_table.ensureCapacity(self.decl_table.size + 1); - const new_decl = try self.allocator.create(Decl); - errdefer self.allocator.destroy(new_decl); - const name = try mem.dupeZ(self.allocator, u8, old_inst.name); - errdefer self.allocator.free(name); - new_decl.* = .{ - .name = name, - .scope = scope.namespace(), - .src = old_inst.src, - .typed_value = .{ .never_succeeded = {} }, - .analysis = .initial_in_progress, - .contents_hash = Decl.hashSimpleName(old_inst.contents), - .link = bin_file_link, - }; - self.decl_table.putAssumeCapacityNoClobber(hash, new_decl); - break :blk new_decl; - }; - - var decl_scope: Scope.DeclAnalysis = .{ - .decl = new_decl, - .arena = std.heap.ArenaAllocator.init(self.allocator), - }; - errdefer decl_scope.arena.deinit(); - - const typed_value = self.analyzeInstConst(&decl_scope.base, old_inst) catch |err| switch (err) { - error.OutOfMemory => return error.OutOfMemory, - error.AnalysisFail => { - switch (new_decl.analysis) { - .initial_in_progress => new_decl.analysis = .initial_dependency_failure, - .repeat_in_progress => new_decl.analysis = .repeat_dependency_failure, - else => {}, - } - return error.AnalysisFail; - }, - }; - const arena_state = try decl_scope.arena.allocator.create(std.heap.ArenaAllocator.State); - - const has_codegen_bits = typed_value.ty.hasCodeGenBits(); - if (has_codegen_bits) { - // We don't fully codegen the decl until later, but we do need to reserve a global - // offset table index for it. This allows us to codegen decls out of dependency order, - // increasing how many computations can be done in parallel. - try self.bin_file.allocateDeclIndexes(new_decl); - } - - arena_state.* = decl_scope.arena.state; - - new_decl.typed_value = .{ - .most_recent = .{ - .typed_value = typed_value, - .arena = arena_state, - }, - }; - new_decl.analysis = .complete; - if (has_codegen_bits) { - // We ensureCapacity when scanning for decls. - self.work_queue.writeItemAssumeCapacity(.{ .codegen_decl = new_decl }); - } - return new_decl; - } - } - - fn resolveCompleteDecl(self: *Module, scope: *Scope, old_inst: *text.Inst) InnerError!*Decl { - const decl = try self.resolveDecl(scope, old_inst, link.ElfFile.Decl.empty); - switch (decl.analysis) { - .initial_in_progress => unreachable, - .repeat_in_progress => unreachable, - .initial_dependency_failure, - .repeat_dependency_failure, - .initial_sema_failure, - .repeat_sema_failure, - .codegen_failure, - .codegen_failure_retryable, - => return error.AnalysisFail, - - .complete => return decl, - } - } - - fn resolveInst(self: *Module, scope: *Scope, old_inst: *text.Inst) InnerError!*Inst { - if (scope.cast(Scope.Block)) |block| { - if (block.func.analysis.in_progress.inst_table.get(old_inst)) |kv| { - return kv.value; - } - } - - const decl = try self.resolveCompleteDecl(scope, old_inst); - const decl_ref = try self.analyzeDeclRef(scope, old_inst.src, decl); - return self.analyzeDeref(scope, old_inst.src, decl_ref, old_inst.src); - } - - fn requireRuntimeBlock(self: *Module, scope: *Scope, src: usize) !*Scope.Block { - return scope.cast(Scope.Block) orelse - return self.fail(scope, src, "instruction illegal outside function body", .{}); - } - - fn resolveInstConst(self: *Module, scope: *Scope, old_inst: *text.Inst) InnerError!TypedValue { - const new_inst = try self.resolveInst(scope, old_inst); - const val = try self.resolveConstValue(scope, new_inst); - return TypedValue{ - .ty = new_inst.ty, - .val = val, - }; - } - - fn resolveConstValue(self: *Module, scope: *Scope, base: *Inst) !Value { - return (try self.resolveDefinedValue(scope, base)) orelse - return self.fail(scope, base.src, "unable to resolve comptime value", .{}); - } - - fn resolveDefinedValue(self: *Module, scope: *Scope, base: *Inst) !?Value { - if (base.value()) |val| { - if (val.isUndef()) { - return self.fail(scope, base.src, "use of undefined value here causes undefined behavior", .{}); - } - return val; - } - return null; - } - - fn resolveConstString(self: *Module, scope: *Scope, old_inst: *text.Inst) ![]u8 { - const new_inst = try self.resolveInst(scope, old_inst); - const wanted_type = Type.initTag(.const_slice_u8); - const coerced_inst = try self.coerce(scope, wanted_type, new_inst); - const val = try self.resolveConstValue(scope, coerced_inst); - return val.toAllocatedBytes(scope.arena()); - } - - fn resolveType(self: *Module, scope: *Scope, old_inst: *text.Inst) !Type { - const new_inst = try self.resolveInst(scope, old_inst); - const wanted_type = Type.initTag(.@"type"); - const coerced_inst = try self.coerce(scope, wanted_type, new_inst); - const val = try self.resolveConstValue(scope, coerced_inst); - return val.toType(); - } - - fn analyzeExport(self: *Module, scope: *Scope, export_inst: *text.Inst.Export) InnerError!void { - try self.decl_exports.ensureCapacity(self.decl_exports.size + 1); - try self.export_owners.ensureCapacity(self.export_owners.size + 1); - const symbol_name = try self.resolveConstString(scope, export_inst.positionals.symbol_name); - const exported_decl = try self.resolveCompleteDecl(scope, export_inst.positionals.value); - const typed_value = exported_decl.typed_value.most_recent.typed_value; - switch (typed_value.ty.zigTypeTag()) { - .Fn => {}, - else => return self.fail( - scope, - export_inst.positionals.value.src, - "unable to export type '{}'", - .{typed_value.ty}, - ), - } - const new_export = try self.allocator.create(Export); - errdefer self.allocator.destroy(new_export); - - const owner_decl = scope.decl(); - - new_export.* = .{ - .options = .{ .name = symbol_name }, - .src = export_inst.base.src, - .link = .{}, - .owner_decl = owner_decl, - .status = .in_progress, - }; - - // Add to export_owners table. - const eo_gop = self.export_owners.getOrPut(owner_decl) catch unreachable; - if (!eo_gop.found_existing) { - eo_gop.kv.value = &[0]*Export{}; - } - eo_gop.kv.value = try self.allocator.realloc(eo_gop.kv.value, eo_gop.kv.value.len + 1); - eo_gop.kv.value[eo_gop.kv.value.len - 1] = new_export; - errdefer eo_gop.kv.value = self.allocator.shrink(eo_gop.kv.value, eo_gop.kv.value.len - 1); - - // Add to exported_decl table. - const de_gop = self.decl_exports.getOrPut(exported_decl) catch unreachable; - if (!de_gop.found_existing) { - de_gop.kv.value = &[0]*Export{}; - } - de_gop.kv.value = try self.allocator.realloc(de_gop.kv.value, de_gop.kv.value.len + 1); - de_gop.kv.value[de_gop.kv.value.len - 1] = new_export; - errdefer de_gop.kv.value = self.allocator.shrink(de_gop.kv.value, de_gop.kv.value.len - 1); - - self.bin_file.updateDeclExports(self, exported_decl, de_gop.kv.value) catch |err| switch (err) { - error.OutOfMemory => return error.OutOfMemory, - else => { - try self.failed_exports.ensureCapacity(self.failed_exports.size + 1); - self.failed_exports.putAssumeCapacityNoClobber(new_export, try ErrorMsg.create( - self.allocator, - export_inst.base.src, - "unable to export: {}", - .{@errorName(err)}, - )); - new_export.status = .failed_retryable; - }, - }; - } - - /// TODO should not need the cast on the last parameter at the callsites - fn addNewInstArgs( - self: *Module, - block: *Scope.Block, - src: usize, - ty: Type, - comptime T: type, - args: Inst.Args(T), - ) !*Inst { - const inst = try self.addNewInst(block, src, ty, T); - inst.args = args; - return &inst.base; - } - - fn addNewInst(self: *Module, block: *Scope.Block, src: usize, ty: Type, comptime T: type) !*T { - const inst = try block.arena.create(T); - inst.* = .{ - .base = .{ - .tag = T.base_tag, - .ty = ty, - .src = src, - }, - .args = undefined, - }; - try block.instructions.append(self.allocator, &inst.base); - return inst; - } - - fn constInst(self: *Module, scope: *Scope, src: usize, typed_value: TypedValue) !*Inst { - const const_inst = try scope.arena().create(Inst.Constant); - const_inst.* = .{ - .base = .{ - .tag = Inst.Constant.base_tag, - .ty = typed_value.ty, - .src = src, - }, - .val = typed_value.val, - }; - return &const_inst.base; - } - - fn constStr(self: *Module, scope: *Scope, src: usize, str: []const u8) !*Inst { - const ty_payload = try scope.arena().create(Type.Payload.Array_u8_Sentinel0); - ty_payload.* = .{ .len = str.len }; - - const bytes_payload = try scope.arena().create(Value.Payload.Bytes); - bytes_payload.* = .{ .data = str }; - - return self.constInst(scope, src, .{ - .ty = Type.initPayload(&ty_payload.base), - .val = Value.initPayload(&bytes_payload.base), - }); - } - - fn constType(self: *Module, scope: *Scope, src: usize, ty: Type) !*Inst { - return self.constInst(scope, src, .{ - .ty = Type.initTag(.type), - .val = try ty.toValue(scope.arena()), - }); - } - - fn constVoid(self: *Module, scope: *Scope, src: usize) !*Inst { - return self.constInst(scope, src, .{ - .ty = Type.initTag(.void), - .val = Value.initTag(.the_one_possible_value), - }); - } - - fn constUndef(self: *Module, scope: *Scope, src: usize, ty: Type) !*Inst { - return self.constInst(scope, src, .{ - .ty = ty, - .val = Value.initTag(.undef), - }); - } - - fn constBool(self: *Module, scope: *Scope, src: usize, v: bool) !*Inst { - return self.constInst(scope, src, .{ - .ty = Type.initTag(.bool), - .val = ([2]Value{ Value.initTag(.bool_false), Value.initTag(.bool_true) })[@boolToInt(v)], - }); - } - - fn constIntUnsigned(self: *Module, scope: *Scope, src: usize, ty: Type, int: u64) !*Inst { - const int_payload = try scope.arena().create(Value.Payload.Int_u64); - int_payload.* = .{ .int = int }; - - return self.constInst(scope, src, .{ - .ty = ty, - .val = Value.initPayload(&int_payload.base), - }); - } - - fn constIntSigned(self: *Module, scope: *Scope, src: usize, ty: Type, int: i64) !*Inst { - const int_payload = try scope.arena().create(Value.Payload.Int_i64); - int_payload.* = .{ .int = int }; - - return self.constInst(scope, src, .{ - .ty = ty, - .val = Value.initPayload(&int_payload.base), - }); - } - - fn constIntBig(self: *Module, scope: *Scope, src: usize, ty: Type, big_int: BigIntConst) !*Inst { - const val_payload = if (big_int.positive) blk: { - if (big_int.to(u64)) |x| { - return self.constIntUnsigned(scope, src, ty, x); - } else |err| switch (err) { - error.NegativeIntoUnsigned => unreachable, - error.TargetTooSmall => {}, // handled below - } - const big_int_payload = try scope.arena().create(Value.Payload.IntBigPositive); - big_int_payload.* = .{ .limbs = big_int.limbs }; - break :blk &big_int_payload.base; - } else blk: { - if (big_int.to(i64)) |x| { - return self.constIntSigned(scope, src, ty, x); - } else |err| switch (err) { - error.NegativeIntoUnsigned => unreachable, - error.TargetTooSmall => {}, // handled below - } - const big_int_payload = try scope.arena().create(Value.Payload.IntBigNegative); - big_int_payload.* = .{ .limbs = big_int.limbs }; - break :blk &big_int_payload.base; - }; - - return self.constInst(scope, src, .{ - .ty = ty, - .val = Value.initPayload(val_payload), - }); - } - - fn analyzeInstConst(self: *Module, scope: *Scope, old_inst: *text.Inst) InnerError!TypedValue { - const new_inst = try self.analyzeInst(scope, old_inst); - return TypedValue{ - .ty = new_inst.ty, - .val = try self.resolveConstValue(scope, new_inst), - }; - } - - fn analyzeInst(self: *Module, scope: *Scope, old_inst: *text.Inst) InnerError!*Inst { - switch (old_inst.tag) { - .breakpoint => return self.analyzeInstBreakpoint(scope, old_inst.cast(text.Inst.Breakpoint).?), - .call => return self.analyzeInstCall(scope, old_inst.cast(text.Inst.Call).?), - .declref => return self.analyzeInstDeclRef(scope, old_inst.cast(text.Inst.DeclRef).?), - .str => { - const bytes = old_inst.cast(text.Inst.Str).?.positionals.bytes; - // The bytes references memory inside the ZIR text module, which can get deallocated - // after semantic analysis is complete. We need the memory to be in the Decl's arena. - const arena_bytes = try scope.arena().dupe(u8, bytes); - return self.constStr(scope, old_inst.src, arena_bytes); - }, - .int => { - const big_int = old_inst.cast(text.Inst.Int).?.positionals.int; - return self.constIntBig(scope, old_inst.src, Type.initTag(.comptime_int), big_int); - }, - .ptrtoint => return self.analyzeInstPtrToInt(scope, old_inst.cast(text.Inst.PtrToInt).?), - .fieldptr => return self.analyzeInstFieldPtr(scope, old_inst.cast(text.Inst.FieldPtr).?), - .deref => return self.analyzeInstDeref(scope, old_inst.cast(text.Inst.Deref).?), - .as => return self.analyzeInstAs(scope, old_inst.cast(text.Inst.As).?), - .@"asm" => return self.analyzeInstAsm(scope, old_inst.cast(text.Inst.Asm).?), - .@"unreachable" => return self.analyzeInstUnreachable(scope, old_inst.cast(text.Inst.Unreachable).?), - .@"return" => return self.analyzeInstRet(scope, old_inst.cast(text.Inst.Return).?), - .@"fn" => return self.analyzeInstFn(scope, old_inst.cast(text.Inst.Fn).?), - .@"export" => { - try self.analyzeExport(scope, old_inst.cast(text.Inst.Export).?); - return self.constVoid(scope, old_inst.src); - }, - .primitive => return self.analyzeInstPrimitive(scope, old_inst.cast(text.Inst.Primitive).?), - .ref => return self.analyzeInstRef(scope, old_inst.cast(text.Inst.Ref).?), - .fntype => return self.analyzeInstFnType(scope, old_inst.cast(text.Inst.FnType).?), - .intcast => return self.analyzeInstIntCast(scope, old_inst.cast(text.Inst.IntCast).?), - .bitcast => return self.analyzeInstBitCast(scope, old_inst.cast(text.Inst.BitCast).?), - .elemptr => return self.analyzeInstElemPtr(scope, old_inst.cast(text.Inst.ElemPtr).?), - .add => return self.analyzeInstAdd(scope, old_inst.cast(text.Inst.Add).?), - .cmp => return self.analyzeInstCmp(scope, old_inst.cast(text.Inst.Cmp).?), - .condbr => return self.analyzeInstCondBr(scope, old_inst.cast(text.Inst.CondBr).?), - .isnull => return self.analyzeInstIsNull(scope, old_inst.cast(text.Inst.IsNull).?), - .isnonnull => return self.analyzeInstIsNonNull(scope, old_inst.cast(text.Inst.IsNonNull).?), - } - } - - fn analyzeInstBreakpoint(self: *Module, scope: *Scope, inst: *text.Inst.Breakpoint) InnerError!*Inst { - const b = try self.requireRuntimeBlock(scope, inst.base.src); - return self.addNewInstArgs(b, inst.base.src, Type.initTag(.void), Inst.Breakpoint, Inst.Args(Inst.Breakpoint){}); - } - - fn analyzeInstRef(self: *Module, scope: *Scope, inst: *text.Inst.Ref) InnerError!*Inst { - const decl = try self.resolveCompleteDecl(scope, inst.positionals.operand); - return self.analyzeDeclRef(scope, inst.base.src, decl); - } - - fn analyzeInstDeclRef(self: *Module, scope: *Scope, inst: *text.Inst.DeclRef) InnerError!*Inst { - const decl_name = try self.resolveConstString(scope, inst.positionals.name); - // This will need to get more fleshed out when there are proper structs & namespaces. - const zir_module = scope.namespace(); - for (zir_module.contents.module.decls) |src_decl| { - if (mem.eql(u8, src_decl.name, decl_name)) { - const decl = try self.resolveCompleteDecl(scope, src_decl); - return self.analyzeDeclRef(scope, inst.base.src, decl); - } - } - return self.fail(scope, inst.positionals.name.src, "use of undeclared identifier '{}'", .{decl_name}); - } - - fn analyzeDeclRef(self: *Module, scope: *Scope, src: usize, decl: *Decl) InnerError!*Inst { - const decl_tv = try decl.typedValue(); - const ty_payload = try scope.arena().create(Type.Payload.SingleConstPointer); - ty_payload.* = .{ .pointee_type = decl_tv.ty }; - const val_payload = try scope.arena().create(Value.Payload.DeclRef); - val_payload.* = .{ .decl = decl }; - return self.constInst(scope, src, .{ - .ty = Type.initPayload(&ty_payload.base), - .val = Value.initPayload(&val_payload.base), - }); - } - - fn analyzeInstCall(self: *Module, scope: *Scope, inst: *text.Inst.Call) InnerError!*Inst { - const func = try self.resolveInst(scope, inst.positionals.func); - if (func.ty.zigTypeTag() != .Fn) - return self.fail(scope, inst.positionals.func.src, "type '{}' not a function", .{func.ty}); - - const cc = func.ty.fnCallingConvention(); - if (cc == .Naked) { - // TODO add error note: declared here - return self.fail( - scope, - inst.positionals.func.src, - "unable to call function with naked calling convention", - .{}, - ); - } - const call_params_len = inst.positionals.args.len; - const fn_params_len = func.ty.fnParamLen(); - if (func.ty.fnIsVarArgs()) { - if (call_params_len < fn_params_len) { - // TODO add error note: declared here - return self.fail( - scope, - inst.positionals.func.src, - "expected at least {} arguments, found {}", - .{ fn_params_len, call_params_len }, - ); - } - return self.fail(scope, inst.base.src, "TODO implement support for calling var args functions", .{}); - } else if (fn_params_len != call_params_len) { - // TODO add error note: declared here - return self.fail( - scope, - inst.positionals.func.src, - "expected {} arguments, found {}", - .{ fn_params_len, call_params_len }, - ); - } - - if (inst.kw_args.modifier == .compile_time) { - return self.fail(scope, inst.base.src, "TODO implement comptime function calls", .{}); - } - if (inst.kw_args.modifier != .auto) { - return self.fail(scope, inst.base.src, "TODO implement call with modifier {}", .{inst.kw_args.modifier}); - } - - // TODO handle function calls of generic functions - - const fn_param_types = try self.allocator.alloc(Type, fn_params_len); - defer self.allocator.free(fn_param_types); - func.ty.fnParamTypes(fn_param_types); - - const casted_args = try scope.arena().alloc(*Inst, fn_params_len); - for (inst.positionals.args) |src_arg, i| { - const uncasted_arg = try self.resolveInst(scope, src_arg); - casted_args[i] = try self.coerce(scope, fn_param_types[i], uncasted_arg); - } - - const b = try self.requireRuntimeBlock(scope, inst.base.src); - return self.addNewInstArgs(b, inst.base.src, Type.initTag(.void), Inst.Call, Inst.Args(Inst.Call){ - .func = func, - .args = casted_args, - }); - } - - fn analyzeInstFn(self: *Module, scope: *Scope, fn_inst: *text.Inst.Fn) InnerError!*Inst { - const fn_type = try self.resolveType(scope, fn_inst.positionals.fn_type); - const new_func = try scope.arena().create(Fn); - new_func.* = .{ - .fn_type = fn_type, - .analysis = .{ .queued = fn_inst }, - }; - const fn_payload = try scope.arena().create(Value.Payload.Function); - fn_payload.* = .{ .func = new_func }; - return self.constInst(scope, fn_inst.base.src, .{ - .ty = fn_type, - .val = Value.initPayload(&fn_payload.base), - }); - } - - fn analyzeInstFnType(self: *Module, scope: *Scope, fntype: *text.Inst.FnType) InnerError!*Inst { - const return_type = try self.resolveType(scope, fntype.positionals.return_type); - - if (return_type.zigTypeTag() == .NoReturn and - fntype.positionals.param_types.len == 0 and - fntype.kw_args.cc == .Unspecified) - { - return self.constType(scope, fntype.base.src, Type.initTag(.fn_noreturn_no_args)); - } - - if (return_type.zigTypeTag() == .NoReturn and - fntype.positionals.param_types.len == 0 and - fntype.kw_args.cc == .Naked) - { - return self.constType(scope, fntype.base.src, Type.initTag(.fn_naked_noreturn_no_args)); - } - - if (return_type.zigTypeTag() == .Void and - fntype.positionals.param_types.len == 0 and - fntype.kw_args.cc == .C) - { - return self.constType(scope, fntype.base.src, Type.initTag(.fn_ccc_void_no_args)); - } - - return self.fail(scope, fntype.base.src, "TODO implement fntype instruction more", .{}); - } - - fn analyzeInstPrimitive(self: *Module, scope: *Scope, primitive: *text.Inst.Primitive) InnerError!*Inst { - return self.constType(scope, primitive.base.src, primitive.positionals.tag.toType()); - } - - fn analyzeInstAs(self: *Module, scope: *Scope, as: *text.Inst.As) InnerError!*Inst { - const dest_type = try self.resolveType(scope, as.positionals.dest_type); - const new_inst = try self.resolveInst(scope, as.positionals.value); - return self.coerce(scope, dest_type, new_inst); - } - - fn analyzeInstPtrToInt(self: *Module, scope: *Scope, ptrtoint: *text.Inst.PtrToInt) InnerError!*Inst { - const ptr = try self.resolveInst(scope, ptrtoint.positionals.ptr); - if (ptr.ty.zigTypeTag() != .Pointer) { - return self.fail(scope, ptrtoint.positionals.ptr.src, "expected pointer, found '{}'", .{ptr.ty}); - } - // TODO handle known-pointer-address - const b = try self.requireRuntimeBlock(scope, ptrtoint.base.src); - const ty = Type.initTag(.usize); - return self.addNewInstArgs(b, ptrtoint.base.src, ty, Inst.PtrToInt, Inst.Args(Inst.PtrToInt){ .ptr = ptr }); - } - - fn analyzeInstFieldPtr(self: *Module, scope: *Scope, fieldptr: *text.Inst.FieldPtr) InnerError!*Inst { - const object_ptr = try self.resolveInst(scope, fieldptr.positionals.object_ptr); - const field_name = try self.resolveConstString(scope, fieldptr.positionals.field_name); - - const elem_ty = switch (object_ptr.ty.zigTypeTag()) { - .Pointer => object_ptr.ty.elemType(), - else => return self.fail(scope, fieldptr.positionals.object_ptr.src, "expected pointer, found '{}'", .{object_ptr.ty}), - }; - switch (elem_ty.zigTypeTag()) { - .Array => { - if (mem.eql(u8, field_name, "len")) { - const len_payload = try scope.arena().create(Value.Payload.Int_u64); - len_payload.* = .{ .int = elem_ty.arrayLen() }; - - const ref_payload = try scope.arena().create(Value.Payload.RefVal); - ref_payload.* = .{ .val = Value.initPayload(&len_payload.base) }; - - return self.constInst(scope, fieldptr.base.src, .{ - .ty = Type.initTag(.single_const_pointer_to_comptime_int), - .val = Value.initPayload(&ref_payload.base), - }); - } else { - return self.fail( - scope, - fieldptr.positionals.field_name.src, - "no member named '{}' in '{}'", - .{ field_name, elem_ty }, - ); - } - }, - else => return self.fail(scope, fieldptr.base.src, "type '{}' does not support field access", .{elem_ty}), - } - } - - fn analyzeInstIntCast(self: *Module, scope: *Scope, intcast: *text.Inst.IntCast) InnerError!*Inst { - const dest_type = try self.resolveType(scope, intcast.positionals.dest_type); - const new_inst = try self.resolveInst(scope, intcast.positionals.value); - - const dest_is_comptime_int = switch (dest_type.zigTypeTag()) { - .ComptimeInt => true, - .Int => false, - else => return self.fail( - scope, - intcast.positionals.dest_type.src, - "expected integer type, found '{}'", - .{ - dest_type, - }, - ), - }; - - switch (new_inst.ty.zigTypeTag()) { - .ComptimeInt, .Int => {}, - else => return self.fail( - scope, - intcast.positionals.value.src, - "expected integer type, found '{}'", - .{new_inst.ty}, - ), - } - - if (dest_is_comptime_int or new_inst.value() != null) { - return self.coerce(scope, dest_type, new_inst); - } - - return self.fail(scope, intcast.base.src, "TODO implement analyze widen or shorten int", .{}); - } - - fn analyzeInstBitCast(self: *Module, scope: *Scope, inst: *text.Inst.BitCast) InnerError!*Inst { - const dest_type = try self.resolveType(scope, inst.positionals.dest_type); - const operand = try self.resolveInst(scope, inst.positionals.operand); - return self.bitcast(scope, dest_type, operand); - } - - fn analyzeInstElemPtr(self: *Module, scope: *Scope, inst: *text.Inst.ElemPtr) InnerError!*Inst { - const array_ptr = try self.resolveInst(scope, inst.positionals.array_ptr); - const uncasted_index = try self.resolveInst(scope, inst.positionals.index); - const elem_index = try self.coerce(scope, Type.initTag(.usize), uncasted_index); - - if (array_ptr.ty.isSinglePointer() and array_ptr.ty.elemType().zigTypeTag() == .Array) { - if (array_ptr.value()) |array_ptr_val| { - if (elem_index.value()) |index_val| { - // Both array pointer and index are compile-time known. - const index_u64 = index_val.toUnsignedInt(); - // @intCast here because it would have been impossible to construct a value that - // required a larger index. - const elem_ptr = try array_ptr_val.elemPtr(scope.arena(), @intCast(usize, index_u64)); - - const type_payload = try scope.arena().create(Type.Payload.SingleConstPointer); - type_payload.* = .{ .pointee_type = array_ptr.ty.elemType().elemType() }; - - return self.constInst(scope, inst.base.src, .{ - .ty = Type.initPayload(&type_payload.base), - .val = elem_ptr, - }); - } - } - } - - return self.fail(scope, inst.base.src, "TODO implement more analyze elemptr", .{}); - } - - fn analyzeInstAdd(self: *Module, scope: *Scope, inst: *text.Inst.Add) InnerError!*Inst { - const lhs = try self.resolveInst(scope, inst.positionals.lhs); - const rhs = try self.resolveInst(scope, inst.positionals.rhs); - - if (lhs.ty.zigTypeTag() == .Int and rhs.ty.zigTypeTag() == .Int) { - if (lhs.value()) |lhs_val| { - if (rhs.value()) |rhs_val| { - // TODO is this a performance issue? maybe we should try the operation without - // resorting to BigInt first. - var lhs_space: Value.BigIntSpace = undefined; - var rhs_space: Value.BigIntSpace = undefined; - const lhs_bigint = lhs_val.toBigInt(&lhs_space); - const rhs_bigint = rhs_val.toBigInt(&rhs_space); - const limbs = try scope.arena().alloc( - std.math.big.Limb, - std.math.max(lhs_bigint.limbs.len, rhs_bigint.limbs.len) + 1, - ); - var result_bigint = BigIntMutable{ .limbs = limbs, .positive = undefined, .len = undefined }; - result_bigint.add(lhs_bigint, rhs_bigint); - const result_limbs = result_bigint.limbs[0..result_bigint.len]; - - if (!lhs.ty.eql(rhs.ty)) { - return self.fail(scope, inst.base.src, "TODO implement peer type resolution", .{}); - } - - const val_payload = if (result_bigint.positive) blk: { - const val_payload = try scope.arena().create(Value.Payload.IntBigPositive); - val_payload.* = .{ .limbs = result_limbs }; - break :blk &val_payload.base; - } else blk: { - const val_payload = try scope.arena().create(Value.Payload.IntBigNegative); - val_payload.* = .{ .limbs = result_limbs }; - break :blk &val_payload.base; - }; - - return self.constInst(scope, inst.base.src, .{ - .ty = lhs.ty, - .val = Value.initPayload(val_payload), - }); - } - } - } - - return self.fail(scope, inst.base.src, "TODO implement more analyze add", .{}); - } - - fn analyzeInstDeref(self: *Module, scope: *Scope, deref: *text.Inst.Deref) InnerError!*Inst { - const ptr = try self.resolveInst(scope, deref.positionals.ptr); - return self.analyzeDeref(scope, deref.base.src, ptr, deref.positionals.ptr.src); - } - - fn analyzeDeref(self: *Module, scope: *Scope, src: usize, ptr: *Inst, ptr_src: usize) InnerError!*Inst { - const elem_ty = switch (ptr.ty.zigTypeTag()) { - .Pointer => ptr.ty.elemType(), - else => return self.fail(scope, ptr_src, "expected pointer, found '{}'", .{ptr.ty}), - }; - if (ptr.value()) |val| { - return self.constInst(scope, src, .{ - .ty = elem_ty, - .val = try val.pointerDeref(scope.arena()), - }); - } - - return self.fail(scope, src, "TODO implement runtime deref", .{}); - } - - fn analyzeInstAsm(self: *Module, scope: *Scope, assembly: *text.Inst.Asm) InnerError!*Inst { - const return_type = try self.resolveType(scope, assembly.positionals.return_type); - const asm_source = try self.resolveConstString(scope, assembly.positionals.asm_source); - const output = if (assembly.kw_args.output) |o| try self.resolveConstString(scope, o) else null; - - const inputs = try scope.arena().alloc([]const u8, assembly.kw_args.inputs.len); - const clobbers = try scope.arena().alloc([]const u8, assembly.kw_args.clobbers.len); - const args = try scope.arena().alloc(*Inst, assembly.kw_args.args.len); - - for (inputs) |*elem, i| { - elem.* = try self.resolveConstString(scope, assembly.kw_args.inputs[i]); - } - for (clobbers) |*elem, i| { - elem.* = try self.resolveConstString(scope, assembly.kw_args.clobbers[i]); - } - for (args) |*elem, i| { - const arg = try self.resolveInst(scope, assembly.kw_args.args[i]); - elem.* = try self.coerce(scope, Type.initTag(.usize), arg); - } - - const b = try self.requireRuntimeBlock(scope, assembly.base.src); - return self.addNewInstArgs(b, assembly.base.src, return_type, Inst.Assembly, Inst.Args(Inst.Assembly){ - .asm_source = asm_source, - .is_volatile = assembly.kw_args.@"volatile", - .output = output, - .inputs = inputs, - .clobbers = clobbers, - .args = args, - }); - } - - fn analyzeInstCmp(self: *Module, scope: *Scope, inst: *text.Inst.Cmp) InnerError!*Inst { - const lhs = try self.resolveInst(scope, inst.positionals.lhs); - const rhs = try self.resolveInst(scope, inst.positionals.rhs); - const op = inst.positionals.op; - - const is_equality_cmp = switch (op) { - .eq, .neq => true, - else => false, - }; - const lhs_ty_tag = lhs.ty.zigTypeTag(); - const rhs_ty_tag = rhs.ty.zigTypeTag(); - if (is_equality_cmp and lhs_ty_tag == .Null and rhs_ty_tag == .Null) { - // null == null, null != null - return self.constBool(scope, inst.base.src, op == .eq); - } else if (is_equality_cmp and - ((lhs_ty_tag == .Null and rhs_ty_tag == .Optional) or - rhs_ty_tag == .Null and lhs_ty_tag == .Optional)) - { - // comparing null with optionals - const opt_operand = if (lhs_ty_tag == .Optional) lhs else rhs; - if (opt_operand.value()) |opt_val| { - const is_null = opt_val.isNull(); - return self.constBool(scope, inst.base.src, if (op == .eq) is_null else !is_null); - } - const b = try self.requireRuntimeBlock(scope, inst.base.src); - switch (op) { - .eq => return self.addNewInstArgs( - b, - inst.base.src, - Type.initTag(.bool), - Inst.IsNull, - Inst.Args(Inst.IsNull){ .operand = opt_operand }, - ), - .neq => return self.addNewInstArgs( - b, - inst.base.src, - Type.initTag(.bool), - Inst.IsNonNull, - Inst.Args(Inst.IsNonNull){ .operand = opt_operand }, - ), - else => unreachable, - } - } else if (is_equality_cmp and - ((lhs_ty_tag == .Null and rhs.ty.isCPtr()) or (rhs_ty_tag == .Null and lhs.ty.isCPtr()))) - { - return self.fail(scope, inst.base.src, "TODO implement C pointer cmp", .{}); - } else if (lhs_ty_tag == .Null or rhs_ty_tag == .Null) { - const non_null_type = if (lhs_ty_tag == .Null) rhs.ty else lhs.ty; - return self.fail(scope, inst.base.src, "comparison of '{}' with null", .{non_null_type}); - } else if (is_equality_cmp and - ((lhs_ty_tag == .EnumLiteral and rhs_ty_tag == .Union) or - (rhs_ty_tag == .EnumLiteral and lhs_ty_tag == .Union))) - { - return self.fail(scope, inst.base.src, "TODO implement equality comparison between a union's tag value and an enum literal", .{}); - } else if (lhs_ty_tag == .ErrorSet and rhs_ty_tag == .ErrorSet) { - if (!is_equality_cmp) { - return self.fail(scope, inst.base.src, "{} operator not allowed for errors", .{@tagName(op)}); - } - return self.fail(scope, inst.base.src, "TODO implement equality comparison between errors", .{}); - } else if (lhs.ty.isNumeric() and rhs.ty.isNumeric()) { - // This operation allows any combination of integer and float types, regardless of the - // signed-ness, comptime-ness, and bit-width. So peer type resolution is incorrect for - // numeric types. - return self.cmpNumeric(scope, inst.base.src, lhs, rhs, op); - } - return self.fail(scope, inst.base.src, "TODO implement more cmp analysis", .{}); - } - - fn analyzeInstIsNull(self: *Module, scope: *Scope, inst: *text.Inst.IsNull) InnerError!*Inst { - const operand = try self.resolveInst(scope, inst.positionals.operand); - return self.analyzeIsNull(scope, inst.base.src, operand, true); - } - - fn analyzeInstIsNonNull(self: *Module, scope: *Scope, inst: *text.Inst.IsNonNull) InnerError!*Inst { - const operand = try self.resolveInst(scope, inst.positionals.operand); - return self.analyzeIsNull(scope, inst.base.src, operand, false); - } - - fn analyzeInstCondBr(self: *Module, scope: *Scope, inst: *text.Inst.CondBr) InnerError!*Inst { - const uncasted_cond = try self.resolveInst(scope, inst.positionals.condition); - const cond = try self.coerce(scope, Type.initTag(.bool), uncasted_cond); - - if (try self.resolveDefinedValue(scope, cond)) |cond_val| { - const body = if (cond_val.toBool()) &inst.positionals.true_body else &inst.positionals.false_body; - try self.analyzeBody(scope, body.*); - return self.constVoid(scope, inst.base.src); - } - - const parent_block = try self.requireRuntimeBlock(scope, inst.base.src); - - var true_block: Scope.Block = .{ - .func = parent_block.func, - .decl = parent_block.decl, - .instructions = .{}, - .arena = parent_block.arena, - }; - defer true_block.instructions.deinit(self.allocator); - try self.analyzeBody(&true_block.base, inst.positionals.true_body); - - var false_block: Scope.Block = .{ - .func = parent_block.func, - .decl = parent_block.decl, - .instructions = .{}, - .arena = parent_block.arena, - }; - defer false_block.instructions.deinit(self.allocator); - try self.analyzeBody(&false_block.base, inst.positionals.false_body); - - return self.addNewInstArgs(parent_block, inst.base.src, Type.initTag(.void), Inst.CondBr, Inst.Args(Inst.CondBr){ - .condition = cond, - .true_body = .{ .instructions = try scope.arena().dupe(*Inst, true_block.instructions.items) }, - .false_body = .{ .instructions = try scope.arena().dupe(*Inst, false_block.instructions.items) }, - }); - } - - fn wantSafety(self: *Module, scope: *Scope) bool { - return switch (self.optimize_mode) { - .Debug => true, - .ReleaseSafe => true, - .ReleaseFast => false, - .ReleaseSmall => false, - }; - } - - fn analyzeInstUnreachable(self: *Module, scope: *Scope, unreach: *text.Inst.Unreachable) InnerError!*Inst { - const b = try self.requireRuntimeBlock(scope, unreach.base.src); - if (self.wantSafety(scope)) { - // TODO Once we have a panic function to call, call it here instead of this. - _ = try self.addNewInstArgs(b, unreach.base.src, Type.initTag(.void), Inst.Breakpoint, {}); - } - return self.addNewInstArgs(b, unreach.base.src, Type.initTag(.noreturn), Inst.Unreach, {}); - } - - fn analyzeInstRet(self: *Module, scope: *Scope, inst: *text.Inst.Return) InnerError!*Inst { - const b = try self.requireRuntimeBlock(scope, inst.base.src); - return self.addNewInstArgs(b, inst.base.src, Type.initTag(.noreturn), Inst.Ret, {}); - } - - fn analyzeBody(self: *Module, scope: *Scope, body: text.Module.Body) !void { - if (scope.cast(Scope.Block)) |b| { - const analysis = b.func.analysis.in_progress; - analysis.needed_inst_capacity += body.instructions.len; - try analysis.inst_table.ensureCapacity(analysis.needed_inst_capacity); - for (body.instructions) |src_inst| { - const new_inst = try self.analyzeInst(scope, src_inst); - analysis.inst_table.putAssumeCapacityNoClobber(src_inst, new_inst); - } - } else { - for (body.instructions) |src_inst| { - _ = try self.analyzeInst(scope, src_inst); - } - } - } - - fn analyzeIsNull( - self: *Module, - scope: *Scope, - src: usize, - operand: *Inst, - invert_logic: bool, - ) InnerError!*Inst { - return self.fail(scope, src, "TODO implement analysis of isnull and isnotnull", .{}); - } - - /// Asserts that lhs and rhs types are both numeric. - fn cmpNumeric( - self: *Module, - scope: *Scope, - src: usize, - lhs: *Inst, - rhs: *Inst, - op: std.math.CompareOperator, - ) !*Inst { - assert(lhs.ty.isNumeric()); - assert(rhs.ty.isNumeric()); - - const lhs_ty_tag = lhs.ty.zigTypeTag(); - const rhs_ty_tag = rhs.ty.zigTypeTag(); - - if (lhs_ty_tag == .Vector and rhs_ty_tag == .Vector) { - if (lhs.ty.arrayLen() != rhs.ty.arrayLen()) { - return self.fail(scope, src, "vector length mismatch: {} and {}", .{ - lhs.ty.arrayLen(), - rhs.ty.arrayLen(), - }); - } - return self.fail(scope, src, "TODO implement support for vectors in cmpNumeric", .{}); - } else if (lhs_ty_tag == .Vector or rhs_ty_tag == .Vector) { - return self.fail(scope, src, "mixed scalar and vector operands to comparison operator: '{}' and '{}'", .{ - lhs.ty, - rhs.ty, - }); - } - - if (lhs.value()) |lhs_val| { - if (rhs.value()) |rhs_val| { - return self.constBool(scope, src, Value.compare(lhs_val, op, rhs_val)); - } - } - - // TODO handle comparisons against lazy zero values - // Some values can be compared against zero without being runtime known or without forcing - // a full resolution of their value, for example `@sizeOf(@Frame(function))` is known to - // always be nonzero, and we benefit from not forcing the full evaluation and stack frame layout - // of this function if we don't need to. - - // It must be a runtime comparison. - const b = try self.requireRuntimeBlock(scope, src); - // For floats, emit a float comparison instruction. - const lhs_is_float = switch (lhs_ty_tag) { - .Float, .ComptimeFloat => true, - else => false, - }; - const rhs_is_float = switch (rhs_ty_tag) { - .Float, .ComptimeFloat => true, - else => false, - }; - if (lhs_is_float and rhs_is_float) { - // Implicit cast the smaller one to the larger one. - const dest_type = x: { - if (lhs_ty_tag == .ComptimeFloat) { - break :x rhs.ty; - } else if (rhs_ty_tag == .ComptimeFloat) { - break :x lhs.ty; - } - if (lhs.ty.floatBits(self.target()) >= rhs.ty.floatBits(self.target())) { - break :x lhs.ty; - } else { - break :x rhs.ty; - } - }; - const casted_lhs = try self.coerce(scope, dest_type, lhs); - const casted_rhs = try self.coerce(scope, dest_type, rhs); - return self.addNewInstArgs(b, src, dest_type, Inst.Cmp, Inst.Args(Inst.Cmp){ - .lhs = casted_lhs, - .rhs = casted_rhs, - .op = op, - }); - } - // For mixed unsigned integer sizes, implicit cast both operands to the larger integer. - // For mixed signed and unsigned integers, implicit cast both operands to a signed - // integer with + 1 bit. - // For mixed floats and integers, extract the integer part from the float, cast that to - // a signed integer with mantissa bits + 1, and if there was any non-integral part of the float, - // add/subtract 1. - const lhs_is_signed = if (lhs.value()) |lhs_val| - lhs_val.compareWithZero(.lt) - else - (lhs.ty.isFloat() or lhs.ty.isSignedInt()); - const rhs_is_signed = if (rhs.value()) |rhs_val| - rhs_val.compareWithZero(.lt) - else - (rhs.ty.isFloat() or rhs.ty.isSignedInt()); - const dest_int_is_signed = lhs_is_signed or rhs_is_signed; - - var dest_float_type: ?Type = null; - - var lhs_bits: usize = undefined; - if (lhs.value()) |lhs_val| { - if (lhs_val.isUndef()) - return self.constUndef(scope, src, Type.initTag(.bool)); - const is_unsigned = if (lhs_is_float) x: { - var bigint_space: Value.BigIntSpace = undefined; - var bigint = try lhs_val.toBigInt(&bigint_space).toManaged(self.allocator); - defer bigint.deinit(); - const zcmp = lhs_val.orderAgainstZero(); - if (lhs_val.floatHasFraction()) { - switch (op) { - .eq => return self.constBool(scope, src, false), - .neq => return self.constBool(scope, src, true), - else => {}, - } - if (zcmp == .lt) { - try bigint.addScalar(bigint.toConst(), -1); - } else { - try bigint.addScalar(bigint.toConst(), 1); - } - } - lhs_bits = bigint.toConst().bitCountTwosComp(); - break :x (zcmp != .lt); - } else x: { - lhs_bits = lhs_val.intBitCountTwosComp(); - break :x (lhs_val.orderAgainstZero() != .lt); - }; - lhs_bits += @boolToInt(is_unsigned and dest_int_is_signed); - } else if (lhs_is_float) { - dest_float_type = lhs.ty; - } else { - const int_info = lhs.ty.intInfo(self.target()); - lhs_bits = int_info.bits + @boolToInt(!int_info.signed and dest_int_is_signed); - } - - var rhs_bits: usize = undefined; - if (rhs.value()) |rhs_val| { - if (rhs_val.isUndef()) - return self.constUndef(scope, src, Type.initTag(.bool)); - const is_unsigned = if (rhs_is_float) x: { - var bigint_space: Value.BigIntSpace = undefined; - var bigint = try rhs_val.toBigInt(&bigint_space).toManaged(self.allocator); - defer bigint.deinit(); - const zcmp = rhs_val.orderAgainstZero(); - if (rhs_val.floatHasFraction()) { - switch (op) { - .eq => return self.constBool(scope, src, false), - .neq => return self.constBool(scope, src, true), - else => {}, - } - if (zcmp == .lt) { - try bigint.addScalar(bigint.toConst(), -1); - } else { - try bigint.addScalar(bigint.toConst(), 1); - } - } - rhs_bits = bigint.toConst().bitCountTwosComp(); - break :x (zcmp != .lt); - } else x: { - rhs_bits = rhs_val.intBitCountTwosComp(); - break :x (rhs_val.orderAgainstZero() != .lt); - }; - rhs_bits += @boolToInt(is_unsigned and dest_int_is_signed); - } else if (rhs_is_float) { - dest_float_type = rhs.ty; - } else { - const int_info = rhs.ty.intInfo(self.target()); - rhs_bits = int_info.bits + @boolToInt(!int_info.signed and dest_int_is_signed); - } - - const dest_type = if (dest_float_type) |ft| ft else blk: { - const max_bits = std.math.max(lhs_bits, rhs_bits); - const casted_bits = std.math.cast(u16, max_bits) catch |err| switch (err) { - error.Overflow => return self.fail(scope, src, "{} exceeds maximum integer bit count", .{max_bits}), - }; - break :blk try self.makeIntType(scope, dest_int_is_signed, casted_bits); - }; - const casted_lhs = try self.coerce(scope, dest_type, lhs); - const casted_rhs = try self.coerce(scope, dest_type, lhs); - - return self.addNewInstArgs(b, src, dest_type, Inst.Cmp, Inst.Args(Inst.Cmp){ - .lhs = casted_lhs, - .rhs = casted_rhs, - .op = op, - }); - } - - fn makeIntType(self: *Module, scope: *Scope, signed: bool, bits: u16) !Type { - if (signed) { - const int_payload = try scope.arena().create(Type.Payload.IntSigned); - int_payload.* = .{ .bits = bits }; - return Type.initPayload(&int_payload.base); - } else { - const int_payload = try scope.arena().create(Type.Payload.IntUnsigned); - int_payload.* = .{ .bits = bits }; - return Type.initPayload(&int_payload.base); - } - } - - fn coerce(self: *Module, scope: *Scope, dest_type: Type, inst: *Inst) !*Inst { - // If the types are the same, we can return the operand. - if (dest_type.eql(inst.ty)) - return inst; - - const in_memory_result = coerceInMemoryAllowed(dest_type, inst.ty); - if (in_memory_result == .ok) { - return self.bitcast(scope, dest_type, inst); - } - - // *[N]T to []T - if (inst.ty.isSinglePointer() and dest_type.isSlice() and - (!inst.ty.pointerIsConst() or dest_type.pointerIsConst())) - { - const array_type = inst.ty.elemType(); - const dst_elem_type = dest_type.elemType(); - if (array_type.zigTypeTag() == .Array and - coerceInMemoryAllowed(dst_elem_type, array_type.elemType()) == .ok) - { - return self.coerceArrayPtrToSlice(scope, dest_type, inst); - } - } - - // comptime_int to fixed-width integer - if (inst.ty.zigTypeTag() == .ComptimeInt and dest_type.zigTypeTag() == .Int) { - // The representation is already correct; we only need to make sure it fits in the destination type. - const val = inst.value().?; // comptime_int always has comptime known value - if (!val.intFitsInType(dest_type, self.target())) { - return self.fail(scope, inst.src, "type {} cannot represent integer value {}", .{ inst.ty, val }); - } - return self.constInst(scope, inst.src, .{ .ty = dest_type, .val = val }); - } - - // integer widening - if (inst.ty.zigTypeTag() == .Int and dest_type.zigTypeTag() == .Int) { - const src_info = inst.ty.intInfo(self.target()); - const dst_info = dest_type.intInfo(self.target()); - if (src_info.signed == dst_info.signed and dst_info.bits >= src_info.bits) { - if (inst.value()) |val| { - return self.constInst(scope, inst.src, .{ .ty = dest_type, .val = val }); - } else { - return self.fail(scope, inst.src, "TODO implement runtime integer widening", .{}); - } - } else { - return self.fail(scope, inst.src, "TODO implement more int widening {} to {}", .{ inst.ty, dest_type }); - } - } - - return self.fail(scope, inst.src, "TODO implement type coercion from {} to {}", .{ inst.ty, dest_type }); - } - - fn bitcast(self: *Module, scope: *Scope, dest_type: Type, inst: *Inst) !*Inst { - if (inst.value()) |val| { - // Keep the comptime Value representation; take the new type. - return self.constInst(scope, inst.src, .{ .ty = dest_type, .val = val }); - } - // TODO validate the type size and other compile errors - const b = try self.requireRuntimeBlock(scope, inst.src); - return self.addNewInstArgs(b, inst.src, dest_type, Inst.BitCast, Inst.Args(Inst.BitCast){ .operand = inst }); - } - - fn coerceArrayPtrToSlice(self: *Module, scope: *Scope, dest_type: Type, inst: *Inst) !*Inst { - if (inst.value()) |val| { - // The comptime Value representation is compatible with both types. - return self.constInst(scope, inst.src, .{ .ty = dest_type, .val = val }); - } - return self.fail(scope, inst.src, "TODO implement coerceArrayPtrToSlice runtime instruction", .{}); - } - - fn fail(self: *Module, scope: *Scope, src: usize, comptime format: []const u8, args: var) InnerError { - @setCold(true); - try self.failed_decls.ensureCapacity(self.failed_decls.size + 1); - try self.failed_files.ensureCapacity(self.failed_files.size + 1); - const err_msg = try ErrorMsg.create(self.allocator, src, format, args); - switch (scope.tag) { - .decl => { - const decl = scope.cast(Scope.DeclAnalysis).?.decl; - switch (decl.analysis) { - .initial_in_progress => decl.analysis = .initial_sema_failure, - .repeat_in_progress => decl.analysis = .repeat_sema_failure, - else => unreachable, - } - self.failed_decls.putAssumeCapacityNoClobber(decl, err_msg); - }, - .block => { - const block = scope.cast(Scope.Block).?; - block.func.analysis = .sema_failure; - self.failed_decls.putAssumeCapacityNoClobber(block.decl, err_msg); - }, - .zir_module => { - const zir_module = scope.cast(Scope.ZIRModule).?; - zir_module.status = .loaded_sema_failure; - self.failed_files.putAssumeCapacityNoClobber(zir_module, err_msg); - }, - } - return error.AnalysisFail; - } - - const InMemoryCoercionResult = enum { - ok, - no_match, - }; - - fn coerceInMemoryAllowed(dest_type: Type, src_type: Type) InMemoryCoercionResult { - if (dest_type.eql(src_type)) - return .ok; - - // TODO: implement more of this function - - return .no_match; - } -}; - -pub const ErrorMsg = struct { - byte_offset: usize, - msg: []const u8, - - pub fn create(allocator: *Allocator, byte_offset: usize, comptime format: []const u8, args: var) !*ErrorMsg { - const self = try allocator.create(ErrorMsg); - errdefer allocator.destroy(self); - self.* = try init(allocator, byte_offset, format, args); - return self; - } - - /// Assumes the ErrorMsg struct and msg were both allocated with allocator. - pub fn destroy(self: *ErrorMsg, allocator: *Allocator) void { - self.deinit(allocator); - allocator.destroy(self); - } - - pub fn init(allocator: *Allocator, byte_offset: usize, comptime format: []const u8, args: var) !ErrorMsg { - return ErrorMsg{ - .byte_offset = byte_offset, - .msg = try std.fmt.allocPrint(allocator, format, args), - }; - } - - pub fn deinit(self: *ErrorMsg, allocator: *Allocator) void { - allocator.free(self.msg); - self.* = undefined; - } -}; diff --git a/src-self-hosted/ir/text.zig b/src-self-hosted/ir/text.zig @@ -1,1476 +0,0 @@ -//! This file has to do with parsing and rendering the ZIR text format. - -const std = @import("std"); -const mem = std.mem; -const Allocator = std.mem.Allocator; -const assert = std.debug.assert; -const BigIntConst = std.math.big.int.Const; -const BigIntMutable = std.math.big.int.Mutable; -const Type = @import("../type.zig").Type; -const Value = @import("../value.zig").Value; -const TypedValue = @import("../TypedValue.zig"); -const ir = @import("../ir.zig"); - -/// These are instructions that correspond to the ZIR text format. See `ir.Inst` for -/// in-memory, analyzed instructions with types and values. -pub const Inst = struct { - tag: Tag, - /// Byte offset into the source. - src: usize, - name: []const u8, - - /// Slice into the source of the part after the = and before the next instruction. - contents: []const u8 = &[0]u8{}, - - /// These names are used directly as the instruction names in the text format. - pub const Tag = enum { - breakpoint, - call, - /// Represents a reference to a global decl by name. - /// The syntax `@foo` is equivalent to `declref("foo")`. - declref, - str, - int, - ptrtoint, - fieldptr, - deref, - as, - @"asm", - @"unreachable", - @"return", - @"fn", - @"export", - primitive, - ref, - fntype, - intcast, - bitcast, - elemptr, - add, - cmp, - condbr, - isnull, - isnonnull, - }; - - pub fn TagToType(tag: Tag) type { - return switch (tag) { - .breakpoint => Breakpoint, - .call => Call, - .declref => DeclRef, - .str => Str, - .int => Int, - .ptrtoint => PtrToInt, - .fieldptr => FieldPtr, - .deref => Deref, - .as => As, - .@"asm" => Asm, - .@"unreachable" => Unreachable, - .@"return" => Return, - .@"fn" => Fn, - .@"export" => Export, - .primitive => Primitive, - .ref => Ref, - .fntype => FnType, - .intcast => IntCast, - .bitcast => BitCast, - .elemptr => ElemPtr, - .add => Add, - .cmp => Cmp, - .condbr => CondBr, - .isnull => IsNull, - .isnonnull => IsNonNull, - }; - } - - pub fn cast(base: *Inst, comptime T: type) ?*T { - if (base.tag != T.base_tag) - return null; - - return @fieldParentPtr(T, "base", base); - } - - pub const Breakpoint = struct { - pub const base_tag = Tag.breakpoint; - base: Inst, - - positionals: struct {}, - kw_args: struct {}, - }; - - pub const Call = struct { - pub const base_tag = Tag.call; - base: Inst, - - positionals: struct { - func: *Inst, - args: []*Inst, - }, - kw_args: struct { - modifier: std.builtin.CallOptions.Modifier = .auto, - }, - }; - - pub const DeclRef = struct { - pub const base_tag = Tag.declref; - base: Inst, - - positionals: struct { - name: *Inst, - }, - kw_args: struct {}, - }; - - pub const Str = struct { - pub const base_tag = Tag.str; - base: Inst, - - positionals: struct { - bytes: []const u8, - }, - kw_args: struct {}, - }; - - pub const Int = struct { - pub const base_tag = Tag.int; - base: Inst, - - positionals: struct { - int: BigIntConst, - }, - kw_args: struct {}, - }; - - pub const PtrToInt = struct { - pub const base_tag = Tag.ptrtoint; - base: Inst, - - positionals: struct { - ptr: *Inst, - }, - kw_args: struct {}, - }; - - pub const FieldPtr = struct { - pub const base_tag = Tag.fieldptr; - base: Inst, - - positionals: struct { - object_ptr: *Inst, - field_name: *Inst, - }, - kw_args: struct {}, - }; - - pub const Deref = struct { - pub const base_tag = Tag.deref; - base: Inst, - - positionals: struct { - ptr: *Inst, - }, - kw_args: struct {}, - }; - - pub const As = struct { - pub const base_tag = Tag.as; - base: Inst, - - positionals: struct { - dest_type: *Inst, - value: *Inst, - }, - kw_args: struct {}, - }; - - pub const Asm = struct { - pub const base_tag = Tag.@"asm"; - base: Inst, - - positionals: struct { - asm_source: *Inst, - return_type: *Inst, - }, - kw_args: struct { - @"volatile": bool = false, - output: ?*Inst = null, - inputs: []*Inst = &[0]*Inst{}, - clobbers: []*Inst = &[0]*Inst{}, - args: []*Inst = &[0]*Inst{}, - }, - }; - - pub const Unreachable = struct { - pub const base_tag = Tag.@"unreachable"; - base: Inst, - - positionals: struct {}, - kw_args: struct {}, - }; - - pub const Return = struct { - pub const base_tag = Tag.@"return"; - base: Inst, - - positionals: struct {}, - kw_args: struct {}, - }; - - pub const Fn = struct { - pub const base_tag = Tag.@"fn"; - base: Inst, - - positionals: struct { - fn_type: *Inst, - body: Module.Body, - }, - kw_args: struct {}, - }; - - pub const Export = struct { - pub const base_tag = Tag.@"export"; - base: Inst, - - positionals: struct { - symbol_name: *Inst, - value: *Inst, - }, - kw_args: struct {}, - }; - - pub const Ref = struct { - pub const base_tag = Tag.ref; - base: Inst, - - positionals: struct { - operand: *Inst, - }, - kw_args: struct {}, - }; - - pub const Primitive = struct { - pub const base_tag = Tag.primitive; - base: Inst, - - positionals: struct { - tag: BuiltinType, - }, - kw_args: struct {}, - - pub const BuiltinType = enum { - isize, - usize, - c_short, - c_ushort, - c_int, - c_uint, - c_long, - c_ulong, - c_longlong, - c_ulonglong, - c_longdouble, - c_void, - f16, - f32, - f64, - f128, - bool, - void, - noreturn, - type, - anyerror, - comptime_int, - comptime_float, - - fn toType(self: BuiltinType) Type { - return switch (self) { - .isize => Type.initTag(.isize), - .usize => Type.initTag(.usize), - .c_short => Type.initTag(.c_short), - .c_ushort => Type.initTag(.c_ushort), - .c_int => Type.initTag(.c_int), - .c_uint => Type.initTag(.c_uint), - .c_long => Type.initTag(.c_long), - .c_ulong => Type.initTag(.c_ulong), - .c_longlong => Type.initTag(.c_longlong), - .c_ulonglong => Type.initTag(.c_ulonglong), - .c_longdouble => Type.initTag(.c_longdouble), - .c_void => Type.initTag(.c_void), - .f16 => Type.initTag(.f16), - .f32 => Type.initTag(.f32), - .f64 => Type.initTag(.f64), - .f128 => Type.initTag(.f128), - .bool => Type.initTag(.bool), - .void => Type.initTag(.void), - .noreturn => Type.initTag(.noreturn), - .type => Type.initTag(.type), - .anyerror => Type.initTag(.anyerror), - .comptime_int => Type.initTag(.comptime_int), - .comptime_float => Type.initTag(.comptime_float), - }; - } - }; - }; - - pub const FnType = struct { - pub const base_tag = Tag.fntype; - base: Inst, - - positionals: struct { - param_types: []*Inst, - return_type: *Inst, - }, - kw_args: struct { - cc: std.builtin.CallingConvention = .Unspecified, - }, - }; - - pub const IntCast = struct { - pub const base_tag = Tag.intcast; - base: Inst, - - positionals: struct { - dest_type: *Inst, - value: *Inst, - }, - kw_args: struct {}, - }; - - pub const BitCast = struct { - pub const base_tag = Tag.bitcast; - base: Inst, - - positionals: struct { - dest_type: *Inst, - operand: *Inst, - }, - kw_args: struct {}, - }; - - pub const ElemPtr = struct { - pub const base_tag = Tag.elemptr; - base: Inst, - - positionals: struct { - array_ptr: *Inst, - index: *Inst, - }, - kw_args: struct {}, - }; - - pub const Add = struct { - pub const base_tag = Tag.add; - base: Inst, - - positionals: struct { - lhs: *Inst, - rhs: *Inst, - }, - kw_args: struct {}, - }; - - pub const Cmp = struct { - pub const base_tag = Tag.cmp; - base: Inst, - - positionals: struct { - lhs: *Inst, - op: std.math.CompareOperator, - rhs: *Inst, - }, - kw_args: struct {}, - }; - - pub const CondBr = struct { - pub const base_tag = Tag.condbr; - base: Inst, - - positionals: struct { - condition: *Inst, - true_body: Module.Body, - false_body: Module.Body, - }, - kw_args: struct {}, - }; - - pub const IsNull = struct { - pub const base_tag = Tag.isnull; - base: Inst, - - positionals: struct { - operand: *Inst, - }, - kw_args: struct {}, - }; - - pub const IsNonNull = struct { - pub const base_tag = Tag.isnonnull; - base: Inst, - - positionals: struct { - operand: *Inst, - }, - kw_args: struct {}, - }; -}; - -pub const ErrorMsg = struct { - byte_offset: usize, - msg: []const u8, -}; - -pub const Module = struct { - decls: []*Inst, - arena: std.heap.ArenaAllocator, - error_msg: ?ErrorMsg = null, - - pub const Body = struct { - instructions: []*Inst, - }; - - pub fn deinit(self: *Module, allocator: *Allocator) void { - allocator.free(self.decls); - self.arena.deinit(); - self.* = undefined; - } - - /// This is a debugging utility for rendering the tree to stderr. - pub fn dump(self: Module) void { - self.writeToStream(std.heap.page_allocator, std.io.getStdErr().outStream()) catch {}; - } - - const InstPtrTable = std.AutoHashMap(*Inst, struct { index: usize, fn_body: ?*Module.Body }); - - /// The allocator is used for temporary storage, but this function always returns - /// with no resources allocated. - pub fn writeToStream(self: Module, allocator: *Allocator, stream: var) !void { - // First, build a map of *Inst to @ or % indexes - var inst_table = InstPtrTable.init(allocator); - defer inst_table.deinit(); - - try inst_table.ensureCapacity(self.decls.len); - - for (self.decls) |decl, decl_i| { - try inst_table.putNoClobber(decl, .{ .index = decl_i, .fn_body = null }); - - if (decl.cast(Inst.Fn)) |fn_inst| { - for (fn_inst.positionals.body.instructions) |inst, inst_i| { - try inst_table.putNoClobber(inst, .{ .index = inst_i, .fn_body = &fn_inst.positionals.body }); - } - } - } - - for (self.decls) |decl, i| { - try stream.print("@{} ", .{i}); - try self.writeInstToStream(stream, decl, &inst_table); - try stream.writeByte('\n'); - } - } - - fn writeInstToStream( - self: Module, - stream: var, - decl: *Inst, - inst_table: *const InstPtrTable, - ) @TypeOf(stream).Error!void { - // TODO I tried implementing this with an inline for loop and hit a compiler bug - switch (decl.tag) { - .breakpoint => return self.writeInstToStreamGeneric(stream, .breakpoint, decl, inst_table), - .call => return self.writeInstToStreamGeneric(stream, .call, decl, inst_table), - .declref => return self.writeInstToStreamGeneric(stream, .declref, decl, inst_table), - .str => return self.writeInstToStreamGeneric(stream, .str, decl, inst_table), - .int => return self.writeInstToStreamGeneric(stream, .int, decl, inst_table), - .ptrtoint => return self.writeInstToStreamGeneric(stream, .ptrtoint, decl, inst_table), - .fieldptr => return self.writeInstToStreamGeneric(stream, .fieldptr, decl, inst_table), - .deref => return self.writeInstToStreamGeneric(stream, .deref, decl, inst_table), - .as => return self.writeInstToStreamGeneric(stream, .as, decl, inst_table), - .@"asm" => return self.writeInstToStreamGeneric(stream, .@"asm", decl, inst_table), - .@"unreachable" => return self.writeInstToStreamGeneric(stream, .@"unreachable", decl, inst_table), - .@"return" => return self.writeInstToStreamGeneric(stream, .@"return", decl, inst_table), - .@"fn" => return self.writeInstToStreamGeneric(stream, .@"fn", decl, inst_table), - .@"export" => return self.writeInstToStreamGeneric(stream, .@"export", decl, inst_table), - .ref => return self.writeInstToStreamGeneric(stream, .ref, decl, inst_table), - .primitive => return self.writeInstToStreamGeneric(stream, .primitive, decl, inst_table), - .fntype => return self.writeInstToStreamGeneric(stream, .fntype, decl, inst_table), - .intcast => return self.writeInstToStreamGeneric(stream, .intcast, decl, inst_table), - .bitcast => return self.writeInstToStreamGeneric(stream, .bitcast, decl, inst_table), - .elemptr => return self.writeInstToStreamGeneric(stream, .elemptr, decl, inst_table), - .add => return self.writeInstToStreamGeneric(stream, .add, decl, inst_table), - .cmp => return self.writeInstToStreamGeneric(stream, .cmp, decl, inst_table), - .condbr => return self.writeInstToStreamGeneric(stream, .condbr, decl, inst_table), - .isnull => return self.writeInstToStreamGeneric(stream, .isnull, decl, inst_table), - .isnonnull => return self.writeInstToStreamGeneric(stream, .isnonnull, decl, inst_table), - } - } - - fn writeInstToStreamGeneric( - self: Module, - stream: var, - comptime inst_tag: Inst.Tag, - base: *Inst, - inst_table: *const InstPtrTable, - ) !void { - const SpecificInst = Inst.TagToType(inst_tag); - const inst = @fieldParentPtr(SpecificInst, "base", base); - const Positionals = @TypeOf(inst.positionals); - try stream.writeAll("= " ++ @tagName(inst_tag) ++ "("); - const pos_fields = @typeInfo(Positionals).Struct.fields; - inline for (pos_fields) |arg_field, i| { - if (i != 0) { - try stream.writeAll(", "); - } - try self.writeParamToStream(stream, @field(inst.positionals, arg_field.name), inst_table); - } - - comptime var need_comma = pos_fields.len != 0; - const KW_Args = @TypeOf(inst.kw_args); - inline for (@typeInfo(KW_Args).Struct.fields) |arg_field, i| { - if (@typeInfo(arg_field.field_type) == .Optional) { - if (@field(inst.kw_args, arg_field.name)) |non_optional| { - if (need_comma) try stream.writeAll(", "); - try stream.print("{}=", .{arg_field.name}); - try self.writeParamToStream(stream, non_optional, inst_table); - need_comma = true; - } - } else { - if (need_comma) try stream.writeAll(", "); - try stream.print("{}=", .{arg_field.name}); - try self.writeParamToStream(stream, @field(inst.kw_args, arg_field.name), inst_table); - need_comma = true; - } - } - - try stream.writeByte(')'); - } - - fn writeParamToStream(self: Module, stream: var, param: var, inst_table: *const InstPtrTable) !void { - if (@typeInfo(@TypeOf(param)) == .Enum) { - return stream.writeAll(@tagName(param)); - } - switch (@TypeOf(param)) { - *Inst => return self.writeInstParamToStream(stream, param, inst_table), - []*Inst => { - try stream.writeByte('['); - for (param) |inst, i| { - if (i != 0) { - try stream.writeAll(", "); - } - try self.writeInstParamToStream(stream, inst, inst_table); - } - try stream.writeByte(']'); - }, - Module.Body => { - try stream.writeAll("{\n"); - for (param.instructions) |inst, i| { - try stream.print(" %{} ", .{i}); - try self.writeInstToStream(stream, inst, inst_table); - try stream.writeByte('\n'); - } - try stream.writeByte('}'); - }, - bool => return stream.writeByte("01"[@boolToInt(param)]), - []u8, []const u8 => return std.zig.renderStringLiteral(param, stream), - BigIntConst => return stream.print("{}", .{param}), - else => |T| @compileError("unimplemented: rendering parameter of type " ++ @typeName(T)), - } - } - - fn writeInstParamToStream(self: Module, stream: var, inst: *Inst, inst_table: *const InstPtrTable) !void { - const info = inst_table.getValue(inst).?; - const prefix = if (info.fn_body == null) "@" else "%"; - try stream.print("{}{}", .{ prefix, info.index }); - } -}; - -pub fn parse(allocator: *Allocator, source: [:0]const u8) Allocator.Error!Module { - var global_name_map = std.StringHashMap(usize).init(allocator); - defer global_name_map.deinit(); - - var parser: Parser = .{ - .allocator = allocator, - .arena = std.heap.ArenaAllocator.init(allocator), - .i = 0, - .source = source, - .global_name_map = &global_name_map, - .decls = .{}, - .unnamed_index = 0, - }; - errdefer parser.arena.deinit(); - - parser.parseRoot() catch |err| switch (err) { - error.ParseFailure => { - assert(parser.error_msg != null); - }, - else => |e| return e, - }; - - return Module{ - .decls = parser.decls.toOwnedSlice(allocator), - .arena = parser.arena, - .error_msg = parser.error_msg, - }; -} - -const Parser = struct { - allocator: *Allocator, - arena: std.heap.ArenaAllocator, - i: usize, - source: [:0]const u8, - decls: std.ArrayListUnmanaged(*Inst), - global_name_map: *std.StringHashMap(usize), - error_msg: ?ErrorMsg = null, - unnamed_index: usize, - - const Body = struct { - instructions: std.ArrayList(*Inst), - name_map: std.StringHashMap(usize), - }; - - fn parseBody(self: *Parser) !Module.Body { - var body_context = Body{ - .instructions = std.ArrayList(*Inst).init(self.allocator), - .name_map = std.StringHashMap(usize).init(self.allocator), - }; - defer body_context.instructions.deinit(); - defer body_context.name_map.deinit(); - - try requireEatBytes(self, "{"); - skipSpace(self); - - while (true) : (self.i += 1) switch (self.source[self.i]) { - ';' => _ = try skipToAndOver(self, '\n'), - '%' => { - self.i += 1; - const ident = try skipToAndOver(self, ' '); - skipSpace(self); - try requireEatBytes(self, "="); - skipSpace(self); - const inst = try parseInstruction(self, &body_context, ident); - const ident_index = body_context.instructions.items.len; - if (try body_context.name_map.put(ident, ident_index)) |_| { - return self.fail("redefinition of identifier '{}'", .{ident}); - } - try body_context.instructions.append(inst); - continue; - }, - ' ', '\n' => continue, - '}' => { - self.i += 1; - break; - }, - else => |byte| return self.failByte(byte), - }; - - // Move the instructions to the arena - const instrs = try self.arena.allocator.alloc(*Inst, body_context.instructions.items.len); - mem.copy(*Inst, instrs, body_context.instructions.items); - return Module.Body{ .instructions = instrs }; - } - - fn parseStringLiteral(self: *Parser) ![]u8 { - const start = self.i; - try self.requireEatBytes("\""); - - while (true) : (self.i += 1) switch (self.source[self.i]) { - '"' => { - self.i += 1; - const span = self.source[start..self.i]; - var bad_index: usize = undefined; - const parsed = std.zig.parseStringLiteral(&self.arena.allocator, span, &bad_index) catch |err| switch (err) { - error.InvalidCharacter => { - self.i = start + bad_index; - const bad_byte = self.source[self.i]; - return self.fail("invalid string literal character: '{c}'\n", .{bad_byte}); - }, - else => |e| return e, - }; - return parsed; - }, - '\\' => { - self.i += 1; - continue; - }, - 0 => return self.failByte(0), - else => continue, - }; - } - - fn parseIntegerLiteral(self: *Parser) !BigIntConst { - const start = self.i; - if (self.source[self.i] == '-') self.i += 1; - while (true) : (self.i += 1) switch (self.source[self.i]) { - '0'...'9' => continue, - else => break, - }; - const number_text = self.source[start..self.i]; - const base = 10; - // TODO reuse the same array list for this - const limbs_buffer_len = std.math.big.int.calcSetStringLimbsBufferLen(base, number_text.len); - const limbs_buffer = try self.allocator.alloc(std.math.big.Limb, limbs_buffer_len); - defer self.allocator.free(limbs_buffer); - const limb_len = std.math.big.int.calcSetStringLimbCount(base, number_text.len); - const limbs = try self.arena.allocator.alloc(std.math.big.Limb, limb_len); - var result = BigIntMutable{ .limbs = limbs, .positive = undefined, .len = undefined }; - result.setString(base, number_text, limbs_buffer, self.allocator) catch |err| switch (err) { - error.InvalidCharacter => { - self.i = start; - return self.fail("invalid digit in integer literal", .{}); - }, - }; - return result.toConst(); - } - - fn parseRoot(self: *Parser) !void { - // The IR format is designed so that it can be tokenized and parsed at the same time. - while (true) { - switch (self.source[self.i]) { - ';' => _ = try skipToAndOver(self, '\n'), - '@' => { - self.i += 1; - const ident = try skipToAndOver(self, ' '); - skipSpace(self); - try requireEatBytes(self, "="); - skipSpace(self); - const inst = try parseInstruction(self, null, ident); - const ident_index = self.decls.items.len; - if (try self.global_name_map.put(ident, ident_index)) |_| { - return self.fail("redefinition of identifier '{}'", .{ident}); - } - try self.decls.append(self.allocator, inst); - }, - ' ', '\n' => self.i += 1, - 0 => break, - else => |byte| return self.fail("unexpected byte: '{c}'", .{byte}), - } - } - } - - fn eatByte(self: *Parser, byte: u8) bool { - if (self.source[self.i] != byte) return false; - self.i += 1; - return true; - } - - fn skipSpace(self: *Parser) void { - while (self.source[self.i] == ' ' or self.source[self.i] == '\n') { - self.i += 1; - } - } - - fn requireEatBytes(self: *Parser, bytes: []const u8) !void { - const start = self.i; - for (bytes) |byte| { - if (self.source[self.i] != byte) { - self.i = start; - return self.fail("expected '{}'", .{bytes}); - } - self.i += 1; - } - } - - fn skipToAndOver(self: *Parser, byte: u8) ![]const u8 { - const start_i = self.i; - while (self.source[self.i] != 0) : (self.i += 1) { - if (self.source[self.i] == byte) { - const result = self.source[start_i..self.i]; - self.i += 1; - return result; - } - } - return self.fail("unexpected EOF", .{}); - } - - /// ParseFailure is an internal error code; handled in `parse`. - const InnerError = error{ ParseFailure, OutOfMemory }; - - fn failByte(self: *Parser, byte: u8) InnerError { - if (byte == 0) { - return self.fail("unexpected EOF", .{}); - } else { - return self.fail("unexpected byte: '{c}'", .{byte}); - } - } - - fn fail(self: *Parser, comptime format: []const u8, args: var) InnerError { - @setCold(true); - self.error_msg = ErrorMsg{ - .byte_offset = self.i, - .msg = try std.fmt.allocPrint(&self.arena.allocator, format, args), - }; - return error.ParseFailure; - } - - fn parseInstruction(self: *Parser, body_ctx: ?*Body, name: []const u8) InnerError!*Inst { - const contents_start = self.i; - const fn_name = try skipToAndOver(self, '('); - inline for (@typeInfo(Inst.Tag).Enum.fields) |field| { - if (mem.eql(u8, field.name, fn_name)) { - const tag = @field(Inst.Tag, field.name); - return parseInstructionGeneric(self, field.name, Inst.TagToType(tag), body_ctx, name, contents_start); - } - } - return self.fail("unknown instruction '{}'", .{fn_name}); - } - - fn parseInstructionGeneric( - self: *Parser, - comptime fn_name: []const u8, - comptime InstType: type, - body_ctx: ?*Body, - inst_name: []const u8, - contents_start: usize, - ) InnerError!*Inst { - const inst_specific = try self.arena.allocator.create(InstType); - inst_specific.base = .{ - .name = inst_name, - .src = self.i, - .tag = InstType.base_tag, - }; - - if (@hasField(InstType, "ty")) { - inst_specific.ty = opt_type orelse { - return self.fail("instruction '" ++ fn_name ++ "' requires type", .{}); - }; - } - - const Positionals = @TypeOf(inst_specific.positionals); - inline for (@typeInfo(Positionals).Struct.fields) |arg_field| { - if (self.source[self.i] == ',') { - self.i += 1; - skipSpace(self); - } else if (self.source[self.i] == ')') { - return self.fail("expected positional parameter '{}'", .{arg_field.name}); - } - @field(inst_specific.positionals, arg_field.name) = try parseParameterGeneric( - self, - arg_field.field_type, - body_ctx, - ); - skipSpace(self); - } - - const KW_Args = @TypeOf(inst_specific.kw_args); - inst_specific.kw_args = .{}; // assign defaults - skipSpace(self); - while (eatByte(self, ',')) { - skipSpace(self); - const name = try skipToAndOver(self, '='); - inline for (@typeInfo(KW_Args).Struct.fields) |arg_field| { - const field_name = arg_field.name; - if (mem.eql(u8, name, field_name)) { - const NonOptional = switch (@typeInfo(arg_field.field_type)) { - .Optional => |info| info.child, - else => arg_field.field_type, - }; - @field(inst_specific.kw_args, field_name) = try parseParameterGeneric(self, NonOptional, body_ctx); - break; - } - } else { - return self.fail("unrecognized keyword parameter: '{}'", .{name}); - } - skipSpace(self); - } - try requireEatBytes(self, ")"); - - inst_specific.base.contents = self.source[contents_start..self.i]; - - return &inst_specific.base; - } - - fn parseParameterGeneric(self: *Parser, comptime T: type, body_ctx: ?*Body) !T { - if (@typeInfo(T) == .Enum) { - const start = self.i; - while (true) : (self.i += 1) switch (self.source[self.i]) { - ' ', '\n', ',', ')' => { - const enum_name = self.source[start..self.i]; - return std.meta.stringToEnum(T, enum_name) orelse { - return self.fail("tag '{}' not a member of enum '{}'", .{ enum_name, @typeName(T) }); - }; - }, - 0 => return self.failByte(0), - else => continue, - }; - } - switch (T) { - Module.Body => return parseBody(self), - bool => { - const bool_value = switch (self.source[self.i]) { - '0' => false, - '1' => true, - else => |byte| return self.fail("expected '0' or '1' for boolean value, found {c}", .{byte}), - }; - self.i += 1; - return bool_value; - }, - []*Inst => { - try requireEatBytes(self, "["); - skipSpace(self); - if (eatByte(self, ']')) return &[0]*Inst{}; - - var instructions = std.ArrayList(*Inst).init(&self.arena.allocator); - while (true) { - skipSpace(self); - try instructions.append(try parseParameterInst(self, body_ctx)); - skipSpace(self); - if (!eatByte(self, ',')) break; - } - try requireEatBytes(self, "]"); - return instructions.toOwnedSlice(); - }, - *Inst => return parseParameterInst(self, body_ctx), - []u8, []const u8 => return self.parseStringLiteral(), - BigIntConst => return self.parseIntegerLiteral(), - else => @compileError("Unimplemented: ir parseParameterGeneric for type " ++ @typeName(T)), - } - return self.fail("TODO parse parameter {}", .{@typeName(T)}); - } - - fn parseParameterInst(self: *Parser, body_ctx: ?*Body) !*Inst { - const local_ref = switch (self.source[self.i]) { - '@' => false, - '%' => true, - else => |byte| return self.fail("unexpected byte: '{c}'", .{byte}), - }; - const map = if (local_ref) - if (body_ctx) |bc| - &bc.name_map - else - return self.fail("referencing a % instruction in global scope", .{}) - else - self.global_name_map; - - self.i += 1; - const name_start = self.i; - while (true) : (self.i += 1) switch (self.source[self.i]) { - 0, ' ', '\n', ',', ')', ']' => break, - else => continue, - }; - const ident = self.source[name_start..self.i]; - const kv = map.get(ident) orelse { - const bad_name = self.source[name_start - 1 .. self.i]; - const src = name_start - 1; - if (local_ref) { - self.i = src; - return self.fail("unrecognized identifier: {}", .{bad_name}); - } else { - const name = try self.arena.allocator.create(Inst.Str); - name.* = .{ - .base = .{ - .name = try self.generateName(), - .src = src, - .tag = Inst.Str.base_tag, - }, - .positionals = .{ .bytes = ident }, - .kw_args = .{}, - }; - const declref = try self.arena.allocator.create(Inst.DeclRef); - declref.* = .{ - .base = .{ - .name = try self.generateName(), - .src = src, - .tag = Inst.DeclRef.base_tag, - }, - .positionals = .{ .name = &name.base }, - .kw_args = .{}, - }; - return &declref.base; - } - }; - if (local_ref) { - return body_ctx.?.instructions.items[kv.value]; - } else { - return self.decls.items[kv.value]; - } - } - - fn generateName(self: *Parser) ![]u8 { - const result = try std.fmt.allocPrint(&self.arena.allocator, "unnamed${}", .{self.unnamed_index}); - self.unnamed_index += 1; - return result; - } -}; - -pub fn emit_zir(allocator: *Allocator, old_module: ir.Module) !Module { - var ctx: EmitZIR = .{ - .allocator = allocator, - .decls = .{}, - .decl_table = std.AutoHashMap(*ir.Inst, *Inst).init(allocator), - .arena = std.heap.ArenaAllocator.init(allocator), - .old_module = &old_module, - }; - defer ctx.decls.deinit(allocator); - defer ctx.decl_table.deinit(); - errdefer ctx.arena.deinit(); - - try ctx.emit(); - - return Module{ - .decls = ctx.decls.toOwnedSlice(allocator), - .arena = ctx.arena, - }; -} - -const EmitZIR = struct { - allocator: *Allocator, - arena: std.heap.ArenaAllocator, - old_module: *const ir.Module, - decls: std.ArrayListUnmanaged(*Inst), - decl_table: std.AutoHashMap(*ir.Inst, *Inst), - - fn emit(self: *EmitZIR) !void { - var it = self.old_module.decl_exports.iterator(); - while (it.next()) |kv| { - const decl = kv.key; - const exports = kv.value; - const export_value = try self.emitTypedValue(decl.src, decl.typed_value.most_recent.typed_value); - for (exports) |module_export| { - const symbol_name = try self.emitStringLiteral(module_export.src, module_export.options.name); - const export_inst = try self.arena.allocator.create(Inst.Export); - export_inst.* = .{ - .base = .{ - .name = try self.autoName(), - .src = module_export.src, - .tag = Inst.Export.base_tag, - }, - .positionals = .{ - .symbol_name = symbol_name, - .value = export_value, - }, - .kw_args = .{}, - }; - try self.decls.append(self.allocator, &export_inst.base); - } - } - } - - fn resolveInst(self: *EmitZIR, inst_table: *const std.AutoHashMap(*ir.Inst, *Inst), inst: *ir.Inst) !*Inst { - if (inst.cast(ir.Inst.Constant)) |const_inst| { - if (self.decl_table.getValue(inst)) |decl| { - return decl; - } - const new_decl = try self.emitTypedValue(inst.src, .{ .ty = inst.ty, .val = const_inst.val }); - try self.decl_table.putNoClobber(inst, new_decl); - return new_decl; - } else { - return inst_table.getValue(inst).?; - } - } - - fn emitComptimeIntVal(self: *EmitZIR, src: usize, val: Value) !*Inst { - const big_int_space = try self.arena.allocator.create(Value.BigIntSpace); - const int_inst = try self.arena.allocator.create(Inst.Int); - int_inst.* = .{ - .base = .{ - .name = try self.autoName(), - .src = src, - .tag = Inst.Int.base_tag, - }, - .positionals = .{ - .int = val.toBigInt(big_int_space), - }, - .kw_args = .{}, - }; - try self.decls.append(self.allocator, &int_inst.base); - return &int_inst.base; - } - - fn emitTypedValue(self: *EmitZIR, src: usize, typed_value: TypedValue) Allocator.Error!*Inst { - const allocator = &self.arena.allocator; - switch (typed_value.ty.zigTypeTag()) { - .Pointer => { - const ptr_elem_type = typed_value.ty.elemType(); - switch (ptr_elem_type.zigTypeTag()) { - .Array => { - // TODO more checks to make sure this can be emitted as a string literal - //const array_elem_type = ptr_elem_type.elemType(); - //if (array_elem_type.eql(Type.initTag(.u8)) and - // ptr_elem_type.hasSentinel(Value.initTag(.zero))) - //{ - //} - const bytes = typed_value.val.toAllocatedBytes(allocator) catch |err| switch (err) { - error.AnalysisFail => unreachable, - else => |e| return e, - }; - return self.emitStringLiteral(src, bytes); - }, - else => |t| std.debug.panic("TODO implement emitTypedValue for pointer to {}", .{@tagName(t)}), - } - }, - .ComptimeInt => return self.emitComptimeIntVal(src, typed_value.val), - .Int => { - const as_inst = try self.arena.allocator.create(Inst.As); - as_inst.* = .{ - .base = .{ - .name = try self.autoName(), - .src = src, - .tag = Inst.As.base_tag, - }, - .positionals = .{ - .dest_type = try self.emitType(src, typed_value.ty), - .value = try self.emitComptimeIntVal(src, typed_value.val), - }, - .kw_args = .{}, - }; - try self.decls.append(self.allocator, &as_inst.base); - - return &as_inst.base; - }, - .Type => { - const ty = typed_value.val.toType(); - return self.emitType(src, ty); - }, - .Fn => { - const module_fn = typed_value.val.cast(Value.Payload.Function).?.func; - - var inst_table = std.AutoHashMap(*ir.Inst, *Inst).init(self.allocator); - defer inst_table.deinit(); - - var instructions = std.ArrayList(*Inst).init(self.allocator); - defer instructions.deinit(); - - try self.emitBody(module_fn.analysis.success, &inst_table, &instructions); - - const fn_type = try self.emitType(src, module_fn.fn_type); - - const arena_instrs = try self.arena.allocator.alloc(*Inst, instructions.items.len); - mem.copy(*Inst, arena_instrs, instructions.items); - - const fn_inst = try self.arena.allocator.create(Inst.Fn); - fn_inst.* = .{ - .base = .{ - .name = try self.autoName(), - .src = src, - .tag = Inst.Fn.base_tag, - }, - .positionals = .{ - .fn_type = fn_type, - .body = .{ .instructions = arena_instrs }, - }, - .kw_args = .{}, - }; - try self.decls.append(self.allocator, &fn_inst.base); - return &fn_inst.base; - }, - else => |t| std.debug.panic("TODO implement emitTypedValue for {}", .{@tagName(t)}), - } - } - - fn emitTrivial(self: *EmitZIR, src: usize, comptime T: type) Allocator.Error!*Inst { - const new_inst = try self.arena.allocator.create(T); - new_inst.* = .{ - .base = .{ - .name = try self.autoName(), - .src = src, - .tag = T.base_tag, - }, - .positionals = .{}, - .kw_args = .{}, - }; - return &new_inst.base; - } - - fn emitBody( - self: *EmitZIR, - body: ir.Module.Body, - inst_table: *std.AutoHashMap(*ir.Inst, *Inst), - instructions: *std.ArrayList(*Inst), - ) Allocator.Error!void { - for (body.instructions) |inst| { - const new_inst = switch (inst.tag) { - .breakpoint => try self.emitTrivial(inst.src, Inst.Breakpoint), - .call => blk: { - const old_inst = inst.cast(ir.Inst.Call).?; - const new_inst = try self.arena.allocator.create(Inst.Call); - - const args = try self.arena.allocator.alloc(*Inst, old_inst.args.args.len); - for (args) |*elem, i| { - elem.* = try self.resolveInst(inst_table, old_inst.args.args[i]); - } - new_inst.* = .{ - .base = .{ - .name = try self.autoName(), - .src = inst.src, - .tag = Inst.Call.base_tag, - }, - .positionals = .{ - .func = try self.resolveInst(inst_table, old_inst.args.func), - .args = args, - }, - .kw_args = .{}, - }; - break :blk &new_inst.base; - }, - .unreach => try self.emitTrivial(inst.src, Inst.Unreachable), - .ret => try self.emitTrivial(inst.src, Inst.Return), - .constant => unreachable, // excluded from function bodies - .assembly => blk: { - const old_inst = inst.cast(ir.Inst.Assembly).?; - const new_inst = try self.arena.allocator.create(Inst.Asm); - - const inputs = try self.arena.allocator.alloc(*Inst, old_inst.args.inputs.len); - for (inputs) |*elem, i| { - elem.* = try self.emitStringLiteral(inst.src, old_inst.args.inputs[i]); - } - - const clobbers = try self.arena.allocator.alloc(*Inst, old_inst.args.clobbers.len); - for (clobbers) |*elem, i| { - elem.* = try self.emitStringLiteral(inst.src, old_inst.args.clobbers[i]); - } - - const args = try self.arena.allocator.alloc(*Inst, old_inst.args.args.len); - for (args) |*elem, i| { - elem.* = try self.resolveInst(inst_table, old_inst.args.args[i]); - } - - new_inst.* = .{ - .base = .{ - .name = try self.autoName(), - .src = inst.src, - .tag = Inst.Asm.base_tag, - }, - .positionals = .{ - .asm_source = try self.emitStringLiteral(inst.src, old_inst.args.asm_source), - .return_type = try self.emitType(inst.src, inst.ty), - }, - .kw_args = .{ - .@"volatile" = old_inst.args.is_volatile, - .output = if (old_inst.args.output) |o| - try self.emitStringLiteral(inst.src, o) - else - null, - .inputs = inputs, - .clobbers = clobbers, - .args = args, - }, - }; - break :blk &new_inst.base; - }, - .ptrtoint => blk: { - const old_inst = inst.cast(ir.Inst.PtrToInt).?; - const new_inst = try self.arena.allocator.create(Inst.PtrToInt); - new_inst.* = .{ - .base = .{ - .name = try self.autoName(), - .src = inst.src, - .tag = Inst.PtrToInt.base_tag, - }, - .positionals = .{ - .ptr = try self.resolveInst(inst_table, old_inst.args.ptr), - }, - .kw_args = .{}, - }; - break :blk &new_inst.base; - }, - .bitcast => blk: { - const old_inst = inst.cast(ir.Inst.BitCast).?; - const new_inst = try self.arena.allocator.create(Inst.BitCast); - new_inst.* = .{ - .base = .{ - .name = try self.autoName(), - .src = inst.src, - .tag = Inst.BitCast.base_tag, - }, - .positionals = .{ - .dest_type = try self.emitType(inst.src, inst.ty), - .operand = try self.resolveInst(inst_table, old_inst.args.operand), - }, - .kw_args = .{}, - }; - break :blk &new_inst.base; - }, - .cmp => blk: { - const old_inst = inst.cast(ir.Inst.Cmp).?; - const new_inst = try self.arena.allocator.create(Inst.Cmp); - new_inst.* = .{ - .base = .{ - .name = try self.autoName(), - .src = inst.src, - .tag = Inst.Cmp.base_tag, - }, - .positionals = .{ - .lhs = try self.resolveInst(inst_table, old_inst.args.lhs), - .rhs = try self.resolveInst(inst_table, old_inst.args.rhs), - .op = old_inst.args.op, - }, - .kw_args = .{}, - }; - break :blk &new_inst.base; - }, - .condbr => blk: { - const old_inst = inst.cast(ir.Inst.CondBr).?; - - var true_body = std.ArrayList(*Inst).init(self.allocator); - var false_body = std.ArrayList(*Inst).init(self.allocator); - - defer true_body.deinit(); - defer false_body.deinit(); - - try self.emitBody(old_inst.args.true_body, inst_table, &true_body); - try self.emitBody(old_inst.args.false_body, inst_table, &false_body); - - const new_inst = try self.arena.allocator.create(Inst.CondBr); - new_inst.* = .{ - .base = .{ - .name = try self.autoName(), - .src = inst.src, - .tag = Inst.CondBr.base_tag, - }, - .positionals = .{ - .condition = try self.resolveInst(inst_table, old_inst.args.condition), - .true_body = .{ .instructions = true_body.toOwnedSlice() }, - .false_body = .{ .instructions = false_body.toOwnedSlice() }, - }, - .kw_args = .{}, - }; - break :blk &new_inst.base; - }, - .isnull => blk: { - const old_inst = inst.cast(ir.Inst.IsNull).?; - const new_inst = try self.arena.allocator.create(Inst.IsNull); - new_inst.* = .{ - .base = .{ - .name = try self.autoName(), - .src = inst.src, - .tag = Inst.IsNull.base_tag, - }, - .positionals = .{ - .operand = try self.resolveInst(inst_table, old_inst.args.operand), - }, - .kw_args = .{}, - }; - break :blk &new_inst.base; - }, - .isnonnull => blk: { - const old_inst = inst.cast(ir.Inst.IsNonNull).?; - const new_inst = try self.arena.allocator.create(Inst.IsNonNull); - new_inst.* = .{ - .base = .{ - .name = try self.autoName(), - .src = inst.src, - .tag = Inst.IsNonNull.base_tag, - }, - .positionals = .{ - .operand = try self.resolveInst(inst_table, old_inst.args.operand), - }, - .kw_args = .{}, - }; - break :blk &new_inst.base; - }, - }; - try instructions.append(new_inst); - try inst_table.putNoClobber(inst, new_inst); - } - } - - fn emitType(self: *EmitZIR, src: usize, ty: Type) Allocator.Error!*Inst { - switch (ty.tag()) { - .isize => return self.emitPrimitiveType(src, .isize), - .usize => return self.emitPrimitiveType(src, .usize), - .c_short => return self.emitPrimitiveType(src, .c_short), - .c_ushort => return self.emitPrimitiveType(src, .c_ushort), - .c_int => return self.emitPrimitiveType(src, .c_int), - .c_uint => return self.emitPrimitiveType(src, .c_uint), - .c_long => return self.emitPrimitiveType(src, .c_long), - .c_ulong => return self.emitPrimitiveType(src, .c_ulong), - .c_longlong => return self.emitPrimitiveType(src, .c_longlong), - .c_ulonglong => return self.emitPrimitiveType(src, .c_ulonglong), - .c_longdouble => return self.emitPrimitiveType(src, .c_longdouble), - .c_void => return self.emitPrimitiveType(src, .c_void), - .f16 => return self.emitPrimitiveType(src, .f16), - .f32 => return self.emitPrimitiveType(src, .f32), - .f64 => return self.emitPrimitiveType(src, .f64), - .f128 => return self.emitPrimitiveType(src, .f128), - .anyerror => return self.emitPrimitiveType(src, .anyerror), - else => switch (ty.zigTypeTag()) { - .Bool => return self.emitPrimitiveType(src, .bool), - .Void => return self.emitPrimitiveType(src, .void), - .NoReturn => return self.emitPrimitiveType(src, .noreturn), - .Type => return self.emitPrimitiveType(src, .type), - .ComptimeInt => return self.emitPrimitiveType(src, .comptime_int), - .ComptimeFloat => return self.emitPrimitiveType(src, .comptime_float), - .Fn => { - const param_types = try self.allocator.alloc(Type, ty.fnParamLen()); - defer self.allocator.free(param_types); - - ty.fnParamTypes(param_types); - const emitted_params = try self.arena.allocator.alloc(*Inst, param_types.len); - for (param_types) |param_type, i| { - emitted_params[i] = try self.emitType(src, param_type); - } - - const fntype_inst = try self.arena.allocator.create(Inst.FnType); - fntype_inst.* = .{ - .base = .{ - .name = try self.autoName(), - .src = src, - .tag = Inst.FnType.base_tag, - }, - .positionals = .{ - .param_types = emitted_params, - .return_type = try self.emitType(src, ty.fnReturnType()), - }, - .kw_args = .{ - .cc = ty.fnCallingConvention(), - }, - }; - try self.decls.append(self.allocator, &fntype_inst.base); - return &fntype_inst.base; - }, - else => std.debug.panic("TODO implement emitType for {}", .{ty}), - }, - } - } - - fn autoName(self: *EmitZIR) ![]u8 { - return std.fmt.allocPrint(&self.arena.allocator, "{}", .{self.decls.items.len}); - } - - fn emitPrimitiveType(self: *EmitZIR, src: usize, tag: Inst.Primitive.BuiltinType) !*Inst { - const primitive_inst = try self.arena.allocator.create(Inst.Primitive); - primitive_inst.* = .{ - .base = .{ - .name = try self.autoName(), - .src = src, - .tag = Inst.Primitive.base_tag, - }, - .positionals = .{ - .tag = tag, - }, - .kw_args = .{}, - }; - try self.decls.append(self.allocator, &primitive_inst.base); - return &primitive_inst.base; - } - - fn emitStringLiteral(self: *EmitZIR, src: usize, str: []const u8) !*Inst { - const str_inst = try self.arena.allocator.create(Inst.Str); - str_inst.* = .{ - .base = .{ - .name = try self.autoName(), - .src = src, - .tag = Inst.Str.base_tag, - }, - .positionals = .{ - .bytes = str, - }, - .kw_args = .{}, - }; - try self.decls.append(self.allocator, &str_inst.base); - - const ref_inst = try self.arena.allocator.create(Inst.Ref); - ref_inst.* = .{ - .base = .{ - .name = try self.autoName(), - .src = src, - .tag = Inst.Ref.base_tag, - }, - .positionals = .{ - .operand = &str_inst.base, - }, - .kw_args = .{}, - }; - try self.decls.append(self.allocator, &ref_inst.base); - - return &ref_inst.base; - } -}; diff --git a/src-self-hosted/link.zig b/src-self-hosted/link.zig @@ -3,6 +3,7 @@ const mem = std.mem; const assert = std.debug.assert; const Allocator = std.mem.Allocator; const ir = @import("ir.zig"); +const Module = @import("Module.zig"); const fs = std.fs; const elf = std.elf; const codegen = @import("codegen.zig"); @@ -45,8 +46,8 @@ pub fn writeFilePath( allocator: *Allocator, dir: fs.Dir, sub_path: []const u8, - module: ir.Module, - errors: *std.ArrayList(ir.ErrorMsg), + module: Module, + errors: *std.ArrayList(Module.ErrorMsg), ) !void { const options: Options = .{ .target = module.target, @@ -755,7 +756,7 @@ pub const ElfFile = struct { }; } - pub fn allocateDeclIndexes(self: *ElfFile, decl: *ir.Module.Decl) !void { + pub fn allocateDeclIndexes(self: *ElfFile, decl: *Module.Decl) !void { if (decl.link.local_sym_index != 0) return; try self.local_symbols.ensureCapacity(self.allocator, self.local_symbols.items.len + 1); @@ -784,7 +785,7 @@ pub const ElfFile = struct { }; } - pub fn updateDecl(self: *ElfFile, module: *ir.Module, decl: *ir.Module.Decl) !void { + pub fn updateDecl(self: *ElfFile, module: *Module, decl: *Module.Decl) !void { var code_buffer = std.ArrayList(u8).init(self.allocator); defer code_buffer.deinit(); @@ -878,16 +879,16 @@ pub const ElfFile = struct { try self.file.pwriteAll(code, file_offset); // Since we updated the vaddr and the size, each corresponding export symbol also needs to be updated. - const decl_exports = module.decl_exports.getValue(decl) orelse &[0]*ir.Module.Export{}; + const decl_exports = module.decl_exports.getValue(decl) orelse &[0]*Module.Export{}; return self.updateDeclExports(module, decl, decl_exports); } /// Must be called only after a successful call to `updateDecl`. pub fn updateDeclExports( self: *ElfFile, - module: *ir.Module, - decl: *const ir.Module.Decl, - exports: []const *ir.Module.Export, + module: *Module, + decl: *const Module.Decl, + exports: []const *Module.Export, ) !void { try self.global_symbols.ensureCapacity(self.allocator, self.global_symbols.items.len + exports.len); const typed_value = decl.typed_value.most_recent.typed_value; @@ -900,7 +901,7 @@ pub const ElfFile = struct { try module.failed_exports.ensureCapacity(module.failed_exports.size + 1); module.failed_exports.putAssumeCapacityNoClobber( exp, - try ir.ErrorMsg.create(self.allocator, 0, "Unimplemented: ExportOptions.section", .{}), + try Module.ErrorMsg.create(self.allocator, 0, "Unimplemented: ExportOptions.section", .{}), ); continue; } @@ -918,7 +919,7 @@ pub const ElfFile = struct { try module.failed_exports.ensureCapacity(module.failed_exports.size + 1); module.failed_exports.putAssumeCapacityNoClobber( exp, - try ir.ErrorMsg.create(self.allocator, 0, "Unimplemented: GlobalLinkage.LinkOnce", .{}), + try Module.ErrorMsg.create(self.allocator, 0, "Unimplemented: GlobalLinkage.LinkOnce", .{}), ); continue; }, diff --git a/src-self-hosted/main.zig b/src-self-hosted/main.zig @@ -6,9 +6,10 @@ const process = std.process; const Allocator = mem.Allocator; const ArrayList = std.ArrayList; const ast = std.zig.ast; -const ir = @import("ir.zig"); +const Module = @import("Module.zig"); const link = @import("link.zig"); const Package = @import("Package.zig"); +const zir = @import("zir.zig"); const LibCInstallation = @import("libc_installation.zig").LibCInstallation; @@ -438,7 +439,7 @@ fn buildOutputType( const root_pkg = try Package.create(gpa, fs.cwd(), ".", src_path); errdefer root_pkg.destroy(); - const root_scope = try gpa.create(ir.Module.Scope.ZIRModule); + const root_scope = try gpa.create(Module.Scope.ZIRModule); errdefer gpa.destroy(root_scope); root_scope.* = .{ .sub_file_path = root_pkg.root_src_path, @@ -447,19 +448,19 @@ fn buildOutputType( .status = .never_loaded, }; - break :blk ir.Module{ + break :blk Module{ .allocator = gpa, .root_pkg = root_pkg, .root_scope = root_scope, .bin_file = &bin_file, .optimize_mode = .Debug, - .decl_table = std.AutoHashMap(ir.Module.Decl.Hash, *ir.Module.Decl).init(gpa), - .decl_exports = std.AutoHashMap(*ir.Module.Decl, []*ir.Module.Export).init(gpa), - .export_owners = std.AutoHashMap(*ir.Module.Decl, []*ir.Module.Export).init(gpa), - .failed_decls = std.AutoHashMap(*ir.Module.Decl, *ir.ErrorMsg).init(gpa), - .failed_files = std.AutoHashMap(*ir.Module.Scope.ZIRModule, *ir.ErrorMsg).init(gpa), - .failed_exports = std.AutoHashMap(*ir.Module.Export, *ir.ErrorMsg).init(gpa), - .work_queue = std.fifo.LinearFifo(ir.Module.WorkItem, .Dynamic).init(gpa), + .decl_table = std.AutoHashMap(Module.Decl.Hash, *Module.Decl).init(gpa), + .decl_exports = std.AutoHashMap(*Module.Decl, []*Module.Export).init(gpa), + .export_owners = std.AutoHashMap(*Module.Decl, []*Module.Export).init(gpa), + .failed_decls = std.AutoHashMap(*Module.Decl, *Module.ErrorMsg).init(gpa), + .failed_files = std.AutoHashMap(*Module.Scope.ZIRModule, *Module.ErrorMsg).init(gpa), + .failed_exports = std.AutoHashMap(*Module.Export, *Module.ErrorMsg).init(gpa), + .work_queue = std.fifo.LinearFifo(Module.WorkItem, .Dynamic).init(gpa), }; }; defer module.deinit(); @@ -491,7 +492,7 @@ fn buildOutputType( } } -fn updateModule(gpa: *Allocator, module: *ir.Module, zir_out_path: ?[]const u8) !void { +fn updateModule(gpa: *Allocator, module: *Module, zir_out_path: ?[]const u8) !void { try module.update(); var errors = try module.getAllErrorsAlloc(); @@ -509,7 +510,7 @@ fn updateModule(gpa: *Allocator, module: *ir.Module, zir_out_path: ?[]const u8) } if (zir_out_path) |zop| { - var new_zir_module = try ir.text.emit_zir(gpa, module.*); + var new_zir_module = try zir.emit(gpa, module.*); defer new_zir_module.deinit(gpa); const baf = try io.BufferedAtomicFile.create(gpa, fs.cwd(), zop, .{}); diff --git a/src-self-hosted/value.zig b/src-self-hosted/value.zig @@ -6,7 +6,7 @@ const BigIntConst = std.math.big.int.Const; const BigIntMutable = std.math.big.int.Mutable; const Target = std.Target; const Allocator = std.mem.Allocator; -const ir = @import("ir.zig"); +const Module = @import("Module.zig"); /// This is the raw data, with no bookkeeping, no memory awareness, /// no de-duplication, and no type system awareness. @@ -904,7 +904,7 @@ pub const Value = extern union { pub const Function = struct { base: Payload = Payload{ .tag = .function }, - func: *ir.Module.Fn, + func: *Module.Fn, }; pub const ArraySentinel0_u8_Type = struct { @@ -926,7 +926,7 @@ pub const Value = extern union { /// Represents a pointer to a decl, not the value of the decl. pub const DeclRef = struct { base: Payload = Payload{ .tag = .decl_ref }, - decl: *ir.Module.Decl, + decl: *Module.Decl, }; pub const ElemPtr = struct { diff --git a/src-self-hosted/zir.zig b/src-self-hosted/zir.zig @@ -0,0 +1,1477 @@ +//! This file has to do with parsing and rendering the ZIR text format. + +const std = @import("std"); +const mem = std.mem; +const Allocator = std.mem.Allocator; +const assert = std.debug.assert; +const BigIntConst = std.math.big.int.Const; +const BigIntMutable = std.math.big.int.Mutable; +const Type = @import("type.zig").Type; +const Value = @import("value.zig").Value; +const TypedValue = @import("TypedValue.zig"); +const ir = @import("ir.zig"); +const IrModule = @import("Module.zig"); + +/// These are instructions that correspond to the ZIR text format. See `ir.Inst` for +/// in-memory, analyzed instructions with types and values. +pub const Inst = struct { + tag: Tag, + /// Byte offset into the source. + src: usize, + name: []const u8, + + /// Slice into the source of the part after the = and before the next instruction. + contents: []const u8 = &[0]u8{}, + + /// These names are used directly as the instruction names in the text format. + pub const Tag = enum { + breakpoint, + call, + /// Represents a reference to a global decl by name. + /// The syntax `@foo` is equivalent to `declref("foo")`. + declref, + str, + int, + ptrtoint, + fieldptr, + deref, + as, + @"asm", + @"unreachable", + @"return", + @"fn", + @"export", + primitive, + ref, + fntype, + intcast, + bitcast, + elemptr, + add, + cmp, + condbr, + isnull, + isnonnull, + }; + + pub fn TagToType(tag: Tag) type { + return switch (tag) { + .breakpoint => Breakpoint, + .call => Call, + .declref => DeclRef, + .str => Str, + .int => Int, + .ptrtoint => PtrToInt, + .fieldptr => FieldPtr, + .deref => Deref, + .as => As, + .@"asm" => Asm, + .@"unreachable" => Unreachable, + .@"return" => Return, + .@"fn" => Fn, + .@"export" => Export, + .primitive => Primitive, + .ref => Ref, + .fntype => FnType, + .intcast => IntCast, + .bitcast => BitCast, + .elemptr => ElemPtr, + .add => Add, + .cmp => Cmp, + .condbr => CondBr, + .isnull => IsNull, + .isnonnull => IsNonNull, + }; + } + + pub fn cast(base: *Inst, comptime T: type) ?*T { + if (base.tag != T.base_tag) + return null; + + return @fieldParentPtr(T, "base", base); + } + + pub const Breakpoint = struct { + pub const base_tag = Tag.breakpoint; + base: Inst, + + positionals: struct {}, + kw_args: struct {}, + }; + + pub const Call = struct { + pub const base_tag = Tag.call; + base: Inst, + + positionals: struct { + func: *Inst, + args: []*Inst, + }, + kw_args: struct { + modifier: std.builtin.CallOptions.Modifier = .auto, + }, + }; + + pub const DeclRef = struct { + pub const base_tag = Tag.declref; + base: Inst, + + positionals: struct { + name: *Inst, + }, + kw_args: struct {}, + }; + + pub const Str = struct { + pub const base_tag = Tag.str; + base: Inst, + + positionals: struct { + bytes: []const u8, + }, + kw_args: struct {}, + }; + + pub const Int = struct { + pub const base_tag = Tag.int; + base: Inst, + + positionals: struct { + int: BigIntConst, + }, + kw_args: struct {}, + }; + + pub const PtrToInt = struct { + pub const base_tag = Tag.ptrtoint; + base: Inst, + + positionals: struct { + ptr: *Inst, + }, + kw_args: struct {}, + }; + + pub const FieldPtr = struct { + pub const base_tag = Tag.fieldptr; + base: Inst, + + positionals: struct { + object_ptr: *Inst, + field_name: *Inst, + }, + kw_args: struct {}, + }; + + pub const Deref = struct { + pub const base_tag = Tag.deref; + base: Inst, + + positionals: struct { + ptr: *Inst, + }, + kw_args: struct {}, + }; + + pub const As = struct { + pub const base_tag = Tag.as; + base: Inst, + + positionals: struct { + dest_type: *Inst, + value: *Inst, + }, + kw_args: struct {}, + }; + + pub const Asm = struct { + pub const base_tag = Tag.@"asm"; + base: Inst, + + positionals: struct { + asm_source: *Inst, + return_type: *Inst, + }, + kw_args: struct { + @"volatile": bool = false, + output: ?*Inst = null, + inputs: []*Inst = &[0]*Inst{}, + clobbers: []*Inst = &[0]*Inst{}, + args: []*Inst = &[0]*Inst{}, + }, + }; + + pub const Unreachable = struct { + pub const base_tag = Tag.@"unreachable"; + base: Inst, + + positionals: struct {}, + kw_args: struct {}, + }; + + pub const Return = struct { + pub const base_tag = Tag.@"return"; + base: Inst, + + positionals: struct {}, + kw_args: struct {}, + }; + + pub const Fn = struct { + pub const base_tag = Tag.@"fn"; + base: Inst, + + positionals: struct { + fn_type: *Inst, + body: Module.Body, + }, + kw_args: struct {}, + }; + + pub const Export = struct { + pub const base_tag = Tag.@"export"; + base: Inst, + + positionals: struct { + symbol_name: *Inst, + value: *Inst, + }, + kw_args: struct {}, + }; + + pub const Ref = struct { + pub const base_tag = Tag.ref; + base: Inst, + + positionals: struct { + operand: *Inst, + }, + kw_args: struct {}, + }; + + pub const Primitive = struct { + pub const base_tag = Tag.primitive; + base: Inst, + + positionals: struct { + tag: BuiltinType, + }, + kw_args: struct {}, + + pub const BuiltinType = enum { + isize, + usize, + c_short, + c_ushort, + c_int, + c_uint, + c_long, + c_ulong, + c_longlong, + c_ulonglong, + c_longdouble, + c_void, + f16, + f32, + f64, + f128, + bool, + void, + noreturn, + type, + anyerror, + comptime_int, + comptime_float, + + fn toType(self: BuiltinType) Type { + return switch (self) { + .isize => Type.initTag(.isize), + .usize => Type.initTag(.usize), + .c_short => Type.initTag(.c_short), + .c_ushort => Type.initTag(.c_ushort), + .c_int => Type.initTag(.c_int), + .c_uint => Type.initTag(.c_uint), + .c_long => Type.initTag(.c_long), + .c_ulong => Type.initTag(.c_ulong), + .c_longlong => Type.initTag(.c_longlong), + .c_ulonglong => Type.initTag(.c_ulonglong), + .c_longdouble => Type.initTag(.c_longdouble), + .c_void => Type.initTag(.c_void), + .f16 => Type.initTag(.f16), + .f32 => Type.initTag(.f32), + .f64 => Type.initTag(.f64), + .f128 => Type.initTag(.f128), + .bool => Type.initTag(.bool), + .void => Type.initTag(.void), + .noreturn => Type.initTag(.noreturn), + .type => Type.initTag(.type), + .anyerror => Type.initTag(.anyerror), + .comptime_int => Type.initTag(.comptime_int), + .comptime_float => Type.initTag(.comptime_float), + }; + } + }; + }; + + pub const FnType = struct { + pub const base_tag = Tag.fntype; + base: Inst, + + positionals: struct { + param_types: []*Inst, + return_type: *Inst, + }, + kw_args: struct { + cc: std.builtin.CallingConvention = .Unspecified, + }, + }; + + pub const IntCast = struct { + pub const base_tag = Tag.intcast; + base: Inst, + + positionals: struct { + dest_type: *Inst, + value: *Inst, + }, + kw_args: struct {}, + }; + + pub const BitCast = struct { + pub const base_tag = Tag.bitcast; + base: Inst, + + positionals: struct { + dest_type: *Inst, + operand: *Inst, + }, + kw_args: struct {}, + }; + + pub const ElemPtr = struct { + pub const base_tag = Tag.elemptr; + base: Inst, + + positionals: struct { + array_ptr: *Inst, + index: *Inst, + }, + kw_args: struct {}, + }; + + pub const Add = struct { + pub const base_tag = Tag.add; + base: Inst, + + positionals: struct { + lhs: *Inst, + rhs: *Inst, + }, + kw_args: struct {}, + }; + + pub const Cmp = struct { + pub const base_tag = Tag.cmp; + base: Inst, + + positionals: struct { + lhs: *Inst, + op: std.math.CompareOperator, + rhs: *Inst, + }, + kw_args: struct {}, + }; + + pub const CondBr = struct { + pub const base_tag = Tag.condbr; + base: Inst, + + positionals: struct { + condition: *Inst, + true_body: Module.Body, + false_body: Module.Body, + }, + kw_args: struct {}, + }; + + pub const IsNull = struct { + pub const base_tag = Tag.isnull; + base: Inst, + + positionals: struct { + operand: *Inst, + }, + kw_args: struct {}, + }; + + pub const IsNonNull = struct { + pub const base_tag = Tag.isnonnull; + base: Inst, + + positionals: struct { + operand: *Inst, + }, + kw_args: struct {}, + }; +}; + +pub const ErrorMsg = struct { + byte_offset: usize, + msg: []const u8, +}; + +pub const Module = struct { + decls: []*Inst, + arena: std.heap.ArenaAllocator, + error_msg: ?ErrorMsg = null, + + pub const Body = struct { + instructions: []*Inst, + }; + + pub fn deinit(self: *Module, allocator: *Allocator) void { + allocator.free(self.decls); + self.arena.deinit(); + self.* = undefined; + } + + /// This is a debugging utility for rendering the tree to stderr. + pub fn dump(self: Module) void { + self.writeToStream(std.heap.page_allocator, std.io.getStdErr().outStream()) catch {}; + } + + const InstPtrTable = std.AutoHashMap(*Inst, struct { index: usize, fn_body: ?*Module.Body }); + + /// The allocator is used for temporary storage, but this function always returns + /// with no resources allocated. + pub fn writeToStream(self: Module, allocator: *Allocator, stream: var) !void { + // First, build a map of *Inst to @ or % indexes + var inst_table = InstPtrTable.init(allocator); + defer inst_table.deinit(); + + try inst_table.ensureCapacity(self.decls.len); + + for (self.decls) |decl, decl_i| { + try inst_table.putNoClobber(decl, .{ .index = decl_i, .fn_body = null }); + + if (decl.cast(Inst.Fn)) |fn_inst| { + for (fn_inst.positionals.body.instructions) |inst, inst_i| { + try inst_table.putNoClobber(inst, .{ .index = inst_i, .fn_body = &fn_inst.positionals.body }); + } + } + } + + for (self.decls) |decl, i| { + try stream.print("@{} ", .{i}); + try self.writeInstToStream(stream, decl, &inst_table); + try stream.writeByte('\n'); + } + } + + fn writeInstToStream( + self: Module, + stream: var, + decl: *Inst, + inst_table: *const InstPtrTable, + ) @TypeOf(stream).Error!void { + // TODO I tried implementing this with an inline for loop and hit a compiler bug + switch (decl.tag) { + .breakpoint => return self.writeInstToStreamGeneric(stream, .breakpoint, decl, inst_table), + .call => return self.writeInstToStreamGeneric(stream, .call, decl, inst_table), + .declref => return self.writeInstToStreamGeneric(stream, .declref, decl, inst_table), + .str => return self.writeInstToStreamGeneric(stream, .str, decl, inst_table), + .int => return self.writeInstToStreamGeneric(stream, .int, decl, inst_table), + .ptrtoint => return self.writeInstToStreamGeneric(stream, .ptrtoint, decl, inst_table), + .fieldptr => return self.writeInstToStreamGeneric(stream, .fieldptr, decl, inst_table), + .deref => return self.writeInstToStreamGeneric(stream, .deref, decl, inst_table), + .as => return self.writeInstToStreamGeneric(stream, .as, decl, inst_table), + .@"asm" => return self.writeInstToStreamGeneric(stream, .@"asm", decl, inst_table), + .@"unreachable" => return self.writeInstToStreamGeneric(stream, .@"unreachable", decl, inst_table), + .@"return" => return self.writeInstToStreamGeneric(stream, .@"return", decl, inst_table), + .@"fn" => return self.writeInstToStreamGeneric(stream, .@"fn", decl, inst_table), + .@"export" => return self.writeInstToStreamGeneric(stream, .@"export", decl, inst_table), + .ref => return self.writeInstToStreamGeneric(stream, .ref, decl, inst_table), + .primitive => return self.writeInstToStreamGeneric(stream, .primitive, decl, inst_table), + .fntype => return self.writeInstToStreamGeneric(stream, .fntype, decl, inst_table), + .intcast => return self.writeInstToStreamGeneric(stream, .intcast, decl, inst_table), + .bitcast => return self.writeInstToStreamGeneric(stream, .bitcast, decl, inst_table), + .elemptr => return self.writeInstToStreamGeneric(stream, .elemptr, decl, inst_table), + .add => return self.writeInstToStreamGeneric(stream, .add, decl, inst_table), + .cmp => return self.writeInstToStreamGeneric(stream, .cmp, decl, inst_table), + .condbr => return self.writeInstToStreamGeneric(stream, .condbr, decl, inst_table), + .isnull => return self.writeInstToStreamGeneric(stream, .isnull, decl, inst_table), + .isnonnull => return self.writeInstToStreamGeneric(stream, .isnonnull, decl, inst_table), + } + } + + fn writeInstToStreamGeneric( + self: Module, + stream: var, + comptime inst_tag: Inst.Tag, + base: *Inst, + inst_table: *const InstPtrTable, + ) !void { + const SpecificInst = Inst.TagToType(inst_tag); + const inst = @fieldParentPtr(SpecificInst, "base", base); + const Positionals = @TypeOf(inst.positionals); + try stream.writeAll("= " ++ @tagName(inst_tag) ++ "("); + const pos_fields = @typeInfo(Positionals).Struct.fields; + inline for (pos_fields) |arg_field, i| { + if (i != 0) { + try stream.writeAll(", "); + } + try self.writeParamToStream(stream, @field(inst.positionals, arg_field.name), inst_table); + } + + comptime var need_comma = pos_fields.len != 0; + const KW_Args = @TypeOf(inst.kw_args); + inline for (@typeInfo(KW_Args).Struct.fields) |arg_field, i| { + if (@typeInfo(arg_field.field_type) == .Optional) { + if (@field(inst.kw_args, arg_field.name)) |non_optional| { + if (need_comma) try stream.writeAll(", "); + try stream.print("{}=", .{arg_field.name}); + try self.writeParamToStream(stream, non_optional, inst_table); + need_comma = true; + } + } else { + if (need_comma) try stream.writeAll(", "); + try stream.print("{}=", .{arg_field.name}); + try self.writeParamToStream(stream, @field(inst.kw_args, arg_field.name), inst_table); + need_comma = true; + } + } + + try stream.writeByte(')'); + } + + fn writeParamToStream(self: Module, stream: var, param: var, inst_table: *const InstPtrTable) !void { + if (@typeInfo(@TypeOf(param)) == .Enum) { + return stream.writeAll(@tagName(param)); + } + switch (@TypeOf(param)) { + *Inst => return self.writeInstParamToStream(stream, param, inst_table), + []*Inst => { + try stream.writeByte('['); + for (param) |inst, i| { + if (i != 0) { + try stream.writeAll(", "); + } + try self.writeInstParamToStream(stream, inst, inst_table); + } + try stream.writeByte(']'); + }, + Module.Body => { + try stream.writeAll("{\n"); + for (param.instructions) |inst, i| { + try stream.print(" %{} ", .{i}); + try self.writeInstToStream(stream, inst, inst_table); + try stream.writeByte('\n'); + } + try stream.writeByte('}'); + }, + bool => return stream.writeByte("01"[@boolToInt(param)]), + []u8, []const u8 => return std.zig.renderStringLiteral(param, stream), + BigIntConst => return stream.print("{}", .{param}), + else => |T| @compileError("unimplemented: rendering parameter of type " ++ @typeName(T)), + } + } + + fn writeInstParamToStream(self: Module, stream: var, inst: *Inst, inst_table: *const InstPtrTable) !void { + const info = inst_table.getValue(inst).?; + const prefix = if (info.fn_body == null) "@" else "%"; + try stream.print("{}{}", .{ prefix, info.index }); + } +}; + +pub fn parse(allocator: *Allocator, source: [:0]const u8) Allocator.Error!Module { + var global_name_map = std.StringHashMap(usize).init(allocator); + defer global_name_map.deinit(); + + var parser: Parser = .{ + .allocator = allocator, + .arena = std.heap.ArenaAllocator.init(allocator), + .i = 0, + .source = source, + .global_name_map = &global_name_map, + .decls = .{}, + .unnamed_index = 0, + }; + errdefer parser.arena.deinit(); + + parser.parseRoot() catch |err| switch (err) { + error.ParseFailure => { + assert(parser.error_msg != null); + }, + else => |e| return e, + }; + + return Module{ + .decls = parser.decls.toOwnedSlice(allocator), + .arena = parser.arena, + .error_msg = parser.error_msg, + }; +} + +const Parser = struct { + allocator: *Allocator, + arena: std.heap.ArenaAllocator, + i: usize, + source: [:0]const u8, + decls: std.ArrayListUnmanaged(*Inst), + global_name_map: *std.StringHashMap(usize), + error_msg: ?ErrorMsg = null, + unnamed_index: usize, + + const Body = struct { + instructions: std.ArrayList(*Inst), + name_map: std.StringHashMap(usize), + }; + + fn parseBody(self: *Parser) !Module.Body { + var body_context = Body{ + .instructions = std.ArrayList(*Inst).init(self.allocator), + .name_map = std.StringHashMap(usize).init(self.allocator), + }; + defer body_context.instructions.deinit(); + defer body_context.name_map.deinit(); + + try requireEatBytes(self, "{"); + skipSpace(self); + + while (true) : (self.i += 1) switch (self.source[self.i]) { + ';' => _ = try skipToAndOver(self, '\n'), + '%' => { + self.i += 1; + const ident = try skipToAndOver(self, ' '); + skipSpace(self); + try requireEatBytes(self, "="); + skipSpace(self); + const inst = try parseInstruction(self, &body_context, ident); + const ident_index = body_context.instructions.items.len; + if (try body_context.name_map.put(ident, ident_index)) |_| { + return self.fail("redefinition of identifier '{}'", .{ident}); + } + try body_context.instructions.append(inst); + continue; + }, + ' ', '\n' => continue, + '}' => { + self.i += 1; + break; + }, + else => |byte| return self.failByte(byte), + }; + + // Move the instructions to the arena + const instrs = try self.arena.allocator.alloc(*Inst, body_context.instructions.items.len); + mem.copy(*Inst, instrs, body_context.instructions.items); + return Module.Body{ .instructions = instrs }; + } + + fn parseStringLiteral(self: *Parser) ![]u8 { + const start = self.i; + try self.requireEatBytes("\""); + + while (true) : (self.i += 1) switch (self.source[self.i]) { + '"' => { + self.i += 1; + const span = self.source[start..self.i]; + var bad_index: usize = undefined; + const parsed = std.zig.parseStringLiteral(&self.arena.allocator, span, &bad_index) catch |err| switch (err) { + error.InvalidCharacter => { + self.i = start + bad_index; + const bad_byte = self.source[self.i]; + return self.fail("invalid string literal character: '{c}'\n", .{bad_byte}); + }, + else => |e| return e, + }; + return parsed; + }, + '\\' => { + self.i += 1; + continue; + }, + 0 => return self.failByte(0), + else => continue, + }; + } + + fn parseIntegerLiteral(self: *Parser) !BigIntConst { + const start = self.i; + if (self.source[self.i] == '-') self.i += 1; + while (true) : (self.i += 1) switch (self.source[self.i]) { + '0'...'9' => continue, + else => break, + }; + const number_text = self.source[start..self.i]; + const base = 10; + // TODO reuse the same array list for this + const limbs_buffer_len = std.math.big.int.calcSetStringLimbsBufferLen(base, number_text.len); + const limbs_buffer = try self.allocator.alloc(std.math.big.Limb, limbs_buffer_len); + defer self.allocator.free(limbs_buffer); + const limb_len = std.math.big.int.calcSetStringLimbCount(base, number_text.len); + const limbs = try self.arena.allocator.alloc(std.math.big.Limb, limb_len); + var result = BigIntMutable{ .limbs = limbs, .positive = undefined, .len = undefined }; + result.setString(base, number_text, limbs_buffer, self.allocator) catch |err| switch (err) { + error.InvalidCharacter => { + self.i = start; + return self.fail("invalid digit in integer literal", .{}); + }, + }; + return result.toConst(); + } + + fn parseRoot(self: *Parser) !void { + // The IR format is designed so that it can be tokenized and parsed at the same time. + while (true) { + switch (self.source[self.i]) { + ';' => _ = try skipToAndOver(self, '\n'), + '@' => { + self.i += 1; + const ident = try skipToAndOver(self, ' '); + skipSpace(self); + try requireEatBytes(self, "="); + skipSpace(self); + const inst = try parseInstruction(self, null, ident); + const ident_index = self.decls.items.len; + if (try self.global_name_map.put(ident, ident_index)) |_| { + return self.fail("redefinition of identifier '{}'", .{ident}); + } + try self.decls.append(self.allocator, inst); + }, + ' ', '\n' => self.i += 1, + 0 => break, + else => |byte| return self.fail("unexpected byte: '{c}'", .{byte}), + } + } + } + + fn eatByte(self: *Parser, byte: u8) bool { + if (self.source[self.i] != byte) return false; + self.i += 1; + return true; + } + + fn skipSpace(self: *Parser) void { + while (self.source[self.i] == ' ' or self.source[self.i] == '\n') { + self.i += 1; + } + } + + fn requireEatBytes(self: *Parser, bytes: []const u8) !void { + const start = self.i; + for (bytes) |byte| { + if (self.source[self.i] != byte) { + self.i = start; + return self.fail("expected '{}'", .{bytes}); + } + self.i += 1; + } + } + + fn skipToAndOver(self: *Parser, byte: u8) ![]const u8 { + const start_i = self.i; + while (self.source[self.i] != 0) : (self.i += 1) { + if (self.source[self.i] == byte) { + const result = self.source[start_i..self.i]; + self.i += 1; + return result; + } + } + return self.fail("unexpected EOF", .{}); + } + + /// ParseFailure is an internal error code; handled in `parse`. + const InnerError = error{ ParseFailure, OutOfMemory }; + + fn failByte(self: *Parser, byte: u8) InnerError { + if (byte == 0) { + return self.fail("unexpected EOF", .{}); + } else { + return self.fail("unexpected byte: '{c}'", .{byte}); + } + } + + fn fail(self: *Parser, comptime format: []const u8, args: var) InnerError { + @setCold(true); + self.error_msg = ErrorMsg{ + .byte_offset = self.i, + .msg = try std.fmt.allocPrint(&self.arena.allocator, format, args), + }; + return error.ParseFailure; + } + + fn parseInstruction(self: *Parser, body_ctx: ?*Body, name: []const u8) InnerError!*Inst { + const contents_start = self.i; + const fn_name = try skipToAndOver(self, '('); + inline for (@typeInfo(Inst.Tag).Enum.fields) |field| { + if (mem.eql(u8, field.name, fn_name)) { + const tag = @field(Inst.Tag, field.name); + return parseInstructionGeneric(self, field.name, Inst.TagToType(tag), body_ctx, name, contents_start); + } + } + return self.fail("unknown instruction '{}'", .{fn_name}); + } + + fn parseInstructionGeneric( + self: *Parser, + comptime fn_name: []const u8, + comptime InstType: type, + body_ctx: ?*Body, + inst_name: []const u8, + contents_start: usize, + ) InnerError!*Inst { + const inst_specific = try self.arena.allocator.create(InstType); + inst_specific.base = .{ + .name = inst_name, + .src = self.i, + .tag = InstType.base_tag, + }; + + if (@hasField(InstType, "ty")) { + inst_specific.ty = opt_type orelse { + return self.fail("instruction '" ++ fn_name ++ "' requires type", .{}); + }; + } + + const Positionals = @TypeOf(inst_specific.positionals); + inline for (@typeInfo(Positionals).Struct.fields) |arg_field| { + if (self.source[self.i] == ',') { + self.i += 1; + skipSpace(self); + } else if (self.source[self.i] == ')') { + return self.fail("expected positional parameter '{}'", .{arg_field.name}); + } + @field(inst_specific.positionals, arg_field.name) = try parseParameterGeneric( + self, + arg_field.field_type, + body_ctx, + ); + skipSpace(self); + } + + const KW_Args = @TypeOf(inst_specific.kw_args); + inst_specific.kw_args = .{}; // assign defaults + skipSpace(self); + while (eatByte(self, ',')) { + skipSpace(self); + const name = try skipToAndOver(self, '='); + inline for (@typeInfo(KW_Args).Struct.fields) |arg_field| { + const field_name = arg_field.name; + if (mem.eql(u8, name, field_name)) { + const NonOptional = switch (@typeInfo(arg_field.field_type)) { + .Optional => |info| info.child, + else => arg_field.field_type, + }; + @field(inst_specific.kw_args, field_name) = try parseParameterGeneric(self, NonOptional, body_ctx); + break; + } + } else { + return self.fail("unrecognized keyword parameter: '{}'", .{name}); + } + skipSpace(self); + } + try requireEatBytes(self, ")"); + + inst_specific.base.contents = self.source[contents_start..self.i]; + + return &inst_specific.base; + } + + fn parseParameterGeneric(self: *Parser, comptime T: type, body_ctx: ?*Body) !T { + if (@typeInfo(T) == .Enum) { + const start = self.i; + while (true) : (self.i += 1) switch (self.source[self.i]) { + ' ', '\n', ',', ')' => { + const enum_name = self.source[start..self.i]; + return std.meta.stringToEnum(T, enum_name) orelse { + return self.fail("tag '{}' not a member of enum '{}'", .{ enum_name, @typeName(T) }); + }; + }, + 0 => return self.failByte(0), + else => continue, + }; + } + switch (T) { + Module.Body => return parseBody(self), + bool => { + const bool_value = switch (self.source[self.i]) { + '0' => false, + '1' => true, + else => |byte| return self.fail("expected '0' or '1' for boolean value, found {c}", .{byte}), + }; + self.i += 1; + return bool_value; + }, + []*Inst => { + try requireEatBytes(self, "["); + skipSpace(self); + if (eatByte(self, ']')) return &[0]*Inst{}; + + var instructions = std.ArrayList(*Inst).init(&self.arena.allocator); + while (true) { + skipSpace(self); + try instructions.append(try parseParameterInst(self, body_ctx)); + skipSpace(self); + if (!eatByte(self, ',')) break; + } + try requireEatBytes(self, "]"); + return instructions.toOwnedSlice(); + }, + *Inst => return parseParameterInst(self, body_ctx), + []u8, []const u8 => return self.parseStringLiteral(), + BigIntConst => return self.parseIntegerLiteral(), + else => @compileError("Unimplemented: ir parseParameterGeneric for type " ++ @typeName(T)), + } + return self.fail("TODO parse parameter {}", .{@typeName(T)}); + } + + fn parseParameterInst(self: *Parser, body_ctx: ?*Body) !*Inst { + const local_ref = switch (self.source[self.i]) { + '@' => false, + '%' => true, + else => |byte| return self.fail("unexpected byte: '{c}'", .{byte}), + }; + const map = if (local_ref) + if (body_ctx) |bc| + &bc.name_map + else + return self.fail("referencing a % instruction in global scope", .{}) + else + self.global_name_map; + + self.i += 1; + const name_start = self.i; + while (true) : (self.i += 1) switch (self.source[self.i]) { + 0, ' ', '\n', ',', ')', ']' => break, + else => continue, + }; + const ident = self.source[name_start..self.i]; + const kv = map.get(ident) orelse { + const bad_name = self.source[name_start - 1 .. self.i]; + const src = name_start - 1; + if (local_ref) { + self.i = src; + return self.fail("unrecognized identifier: {}", .{bad_name}); + } else { + const name = try self.arena.allocator.create(Inst.Str); + name.* = .{ + .base = .{ + .name = try self.generateName(), + .src = src, + .tag = Inst.Str.base_tag, + }, + .positionals = .{ .bytes = ident }, + .kw_args = .{}, + }; + const declref = try self.arena.allocator.create(Inst.DeclRef); + declref.* = .{ + .base = .{ + .name = try self.generateName(), + .src = src, + .tag = Inst.DeclRef.base_tag, + }, + .positionals = .{ .name = &name.base }, + .kw_args = .{}, + }; + return &declref.base; + } + }; + if (local_ref) { + return body_ctx.?.instructions.items[kv.value]; + } else { + return self.decls.items[kv.value]; + } + } + + fn generateName(self: *Parser) ![]u8 { + const result = try std.fmt.allocPrint(&self.arena.allocator, "unnamed${}", .{self.unnamed_index}); + self.unnamed_index += 1; + return result; + } +}; + +pub fn emit(allocator: *Allocator, old_module: IrModule) !Module { + var ctx: EmitZIR = .{ + .allocator = allocator, + .decls = .{}, + .decl_table = std.AutoHashMap(*ir.Inst, *Inst).init(allocator), + .arena = std.heap.ArenaAllocator.init(allocator), + .old_module = &old_module, + }; + defer ctx.decls.deinit(allocator); + defer ctx.decl_table.deinit(); + errdefer ctx.arena.deinit(); + + try ctx.emit(); + + return Module{ + .decls = ctx.decls.toOwnedSlice(allocator), + .arena = ctx.arena, + }; +} + +const EmitZIR = struct { + allocator: *Allocator, + arena: std.heap.ArenaAllocator, + old_module: *const IrModule, + decls: std.ArrayListUnmanaged(*Inst), + decl_table: std.AutoHashMap(*ir.Inst, *Inst), + + fn emit(self: *EmitZIR) !void { + var it = self.old_module.decl_exports.iterator(); + while (it.next()) |kv| { + const decl = kv.key; + const exports = kv.value; + const export_value = try self.emitTypedValue(decl.src, decl.typed_value.most_recent.typed_value); + for (exports) |module_export| { + const symbol_name = try self.emitStringLiteral(module_export.src, module_export.options.name); + const export_inst = try self.arena.allocator.create(Inst.Export); + export_inst.* = .{ + .base = .{ + .name = try self.autoName(), + .src = module_export.src, + .tag = Inst.Export.base_tag, + }, + .positionals = .{ + .symbol_name = symbol_name, + .value = export_value, + }, + .kw_args = .{}, + }; + try self.decls.append(self.allocator, &export_inst.base); + } + } + } + + fn resolveInst(self: *EmitZIR, inst_table: *const std.AutoHashMap(*ir.Inst, *Inst), inst: *ir.Inst) !*Inst { + if (inst.cast(ir.Inst.Constant)) |const_inst| { + if (self.decl_table.getValue(inst)) |decl| { + return decl; + } + const new_decl = try self.emitTypedValue(inst.src, .{ .ty = inst.ty, .val = const_inst.val }); + try self.decl_table.putNoClobber(inst, new_decl); + return new_decl; + } else { + return inst_table.getValue(inst).?; + } + } + + fn emitComptimeIntVal(self: *EmitZIR, src: usize, val: Value) !*Inst { + const big_int_space = try self.arena.allocator.create(Value.BigIntSpace); + const int_inst = try self.arena.allocator.create(Inst.Int); + int_inst.* = .{ + .base = .{ + .name = try self.autoName(), + .src = src, + .tag = Inst.Int.base_tag, + }, + .positionals = .{ + .int = val.toBigInt(big_int_space), + }, + .kw_args = .{}, + }; + try self.decls.append(self.allocator, &int_inst.base); + return &int_inst.base; + } + + fn emitTypedValue(self: *EmitZIR, src: usize, typed_value: TypedValue) Allocator.Error!*Inst { + const allocator = &self.arena.allocator; + switch (typed_value.ty.zigTypeTag()) { + .Pointer => { + const ptr_elem_type = typed_value.ty.elemType(); + switch (ptr_elem_type.zigTypeTag()) { + .Array => { + // TODO more checks to make sure this can be emitted as a string literal + //const array_elem_type = ptr_elem_type.elemType(); + //if (array_elem_type.eql(Type.initTag(.u8)) and + // ptr_elem_type.hasSentinel(Value.initTag(.zero))) + //{ + //} + const bytes = typed_value.val.toAllocatedBytes(allocator) catch |err| switch (err) { + error.AnalysisFail => unreachable, + else => |e| return e, + }; + return self.emitStringLiteral(src, bytes); + }, + else => |t| std.debug.panic("TODO implement emitTypedValue for pointer to {}", .{@tagName(t)}), + } + }, + .ComptimeInt => return self.emitComptimeIntVal(src, typed_value.val), + .Int => { + const as_inst = try self.arena.allocator.create(Inst.As); + as_inst.* = .{ + .base = .{ + .name = try self.autoName(), + .src = src, + .tag = Inst.As.base_tag, + }, + .positionals = .{ + .dest_type = try self.emitType(src, typed_value.ty), + .value = try self.emitComptimeIntVal(src, typed_value.val), + }, + .kw_args = .{}, + }; + try self.decls.append(self.allocator, &as_inst.base); + + return &as_inst.base; + }, + .Type => { + const ty = typed_value.val.toType(); + return self.emitType(src, ty); + }, + .Fn => { + const module_fn = typed_value.val.cast(Value.Payload.Function).?.func; + + var inst_table = std.AutoHashMap(*ir.Inst, *Inst).init(self.allocator); + defer inst_table.deinit(); + + var instructions = std.ArrayList(*Inst).init(self.allocator); + defer instructions.deinit(); + + try self.emitBody(module_fn.analysis.success, &inst_table, &instructions); + + const fn_type = try self.emitType(src, module_fn.fn_type); + + const arena_instrs = try self.arena.allocator.alloc(*Inst, instructions.items.len); + mem.copy(*Inst, arena_instrs, instructions.items); + + const fn_inst = try self.arena.allocator.create(Inst.Fn); + fn_inst.* = .{ + .base = .{ + .name = try self.autoName(), + .src = src, + .tag = Inst.Fn.base_tag, + }, + .positionals = .{ + .fn_type = fn_type, + .body = .{ .instructions = arena_instrs }, + }, + .kw_args = .{}, + }; + try self.decls.append(self.allocator, &fn_inst.base); + return &fn_inst.base; + }, + else => |t| std.debug.panic("TODO implement emitTypedValue for {}", .{@tagName(t)}), + } + } + + fn emitTrivial(self: *EmitZIR, src: usize, comptime T: type) Allocator.Error!*Inst { + const new_inst = try self.arena.allocator.create(T); + new_inst.* = .{ + .base = .{ + .name = try self.autoName(), + .src = src, + .tag = T.base_tag, + }, + .positionals = .{}, + .kw_args = .{}, + }; + return &new_inst.base; + } + + fn emitBody( + self: *EmitZIR, + body: IrModule.Body, + inst_table: *std.AutoHashMap(*ir.Inst, *Inst), + instructions: *std.ArrayList(*Inst), + ) Allocator.Error!void { + for (body.instructions) |inst| { + const new_inst = switch (inst.tag) { + .breakpoint => try self.emitTrivial(inst.src, Inst.Breakpoint), + .call => blk: { + const old_inst = inst.cast(ir.Inst.Call).?; + const new_inst = try self.arena.allocator.create(Inst.Call); + + const args = try self.arena.allocator.alloc(*Inst, old_inst.args.args.len); + for (args) |*elem, i| { + elem.* = try self.resolveInst(inst_table, old_inst.args.args[i]); + } + new_inst.* = .{ + .base = .{ + .name = try self.autoName(), + .src = inst.src, + .tag = Inst.Call.base_tag, + }, + .positionals = .{ + .func = try self.resolveInst(inst_table, old_inst.args.func), + .args = args, + }, + .kw_args = .{}, + }; + break :blk &new_inst.base; + }, + .unreach => try self.emitTrivial(inst.src, Inst.Unreachable), + .ret => try self.emitTrivial(inst.src, Inst.Return), + .constant => unreachable, // excluded from function bodies + .assembly => blk: { + const old_inst = inst.cast(ir.Inst.Assembly).?; + const new_inst = try self.arena.allocator.create(Inst.Asm); + + const inputs = try self.arena.allocator.alloc(*Inst, old_inst.args.inputs.len); + for (inputs) |*elem, i| { + elem.* = try self.emitStringLiteral(inst.src, old_inst.args.inputs[i]); + } + + const clobbers = try self.arena.allocator.alloc(*Inst, old_inst.args.clobbers.len); + for (clobbers) |*elem, i| { + elem.* = try self.emitStringLiteral(inst.src, old_inst.args.clobbers[i]); + } + + const args = try self.arena.allocator.alloc(*Inst, old_inst.args.args.len); + for (args) |*elem, i| { + elem.* = try self.resolveInst(inst_table, old_inst.args.args[i]); + } + + new_inst.* = .{ + .base = .{ + .name = try self.autoName(), + .src = inst.src, + .tag = Inst.Asm.base_tag, + }, + .positionals = .{ + .asm_source = try self.emitStringLiteral(inst.src, old_inst.args.asm_source), + .return_type = try self.emitType(inst.src, inst.ty), + }, + .kw_args = .{ + .@"volatile" = old_inst.args.is_volatile, + .output = if (old_inst.args.output) |o| + try self.emitStringLiteral(inst.src, o) + else + null, + .inputs = inputs, + .clobbers = clobbers, + .args = args, + }, + }; + break :blk &new_inst.base; + }, + .ptrtoint => blk: { + const old_inst = inst.cast(ir.Inst.PtrToInt).?; + const new_inst = try self.arena.allocator.create(Inst.PtrToInt); + new_inst.* = .{ + .base = .{ + .name = try self.autoName(), + .src = inst.src, + .tag = Inst.PtrToInt.base_tag, + }, + .positionals = .{ + .ptr = try self.resolveInst(inst_table, old_inst.args.ptr), + }, + .kw_args = .{}, + }; + break :blk &new_inst.base; + }, + .bitcast => blk: { + const old_inst = inst.cast(ir.Inst.BitCast).?; + const new_inst = try self.arena.allocator.create(Inst.BitCast); + new_inst.* = .{ + .base = .{ + .name = try self.autoName(), + .src = inst.src, + .tag = Inst.BitCast.base_tag, + }, + .positionals = .{ + .dest_type = try self.emitType(inst.src, inst.ty), + .operand = try self.resolveInst(inst_table, old_inst.args.operand), + }, + .kw_args = .{}, + }; + break :blk &new_inst.base; + }, + .cmp => blk: { + const old_inst = inst.cast(ir.Inst.Cmp).?; + const new_inst = try self.arena.allocator.create(Inst.Cmp); + new_inst.* = .{ + .base = .{ + .name = try self.autoName(), + .src = inst.src, + .tag = Inst.Cmp.base_tag, + }, + .positionals = .{ + .lhs = try self.resolveInst(inst_table, old_inst.args.lhs), + .rhs = try self.resolveInst(inst_table, old_inst.args.rhs), + .op = old_inst.args.op, + }, + .kw_args = .{}, + }; + break :blk &new_inst.base; + }, + .condbr => blk: { + const old_inst = inst.cast(ir.Inst.CondBr).?; + + var true_body = std.ArrayList(*Inst).init(self.allocator); + var false_body = std.ArrayList(*Inst).init(self.allocator); + + defer true_body.deinit(); + defer false_body.deinit(); + + try self.emitBody(old_inst.args.true_body, inst_table, &true_body); + try self.emitBody(old_inst.args.false_body, inst_table, &false_body); + + const new_inst = try self.arena.allocator.create(Inst.CondBr); + new_inst.* = .{ + .base = .{ + .name = try self.autoName(), + .src = inst.src, + .tag = Inst.CondBr.base_tag, + }, + .positionals = .{ + .condition = try self.resolveInst(inst_table, old_inst.args.condition), + .true_body = .{ .instructions = true_body.toOwnedSlice() }, + .false_body = .{ .instructions = false_body.toOwnedSlice() }, + }, + .kw_args = .{}, + }; + break :blk &new_inst.base; + }, + .isnull => blk: { + const old_inst = inst.cast(ir.Inst.IsNull).?; + const new_inst = try self.arena.allocator.create(Inst.IsNull); + new_inst.* = .{ + .base = .{ + .name = try self.autoName(), + .src = inst.src, + .tag = Inst.IsNull.base_tag, + }, + .positionals = .{ + .operand = try self.resolveInst(inst_table, old_inst.args.operand), + }, + .kw_args = .{}, + }; + break :blk &new_inst.base; + }, + .isnonnull => blk: { + const old_inst = inst.cast(ir.Inst.IsNonNull).?; + const new_inst = try self.arena.allocator.create(Inst.IsNonNull); + new_inst.* = .{ + .base = .{ + .name = try self.autoName(), + .src = inst.src, + .tag = Inst.IsNonNull.base_tag, + }, + .positionals = .{ + .operand = try self.resolveInst(inst_table, old_inst.args.operand), + }, + .kw_args = .{}, + }; + break :blk &new_inst.base; + }, + }; + try instructions.append(new_inst); + try inst_table.putNoClobber(inst, new_inst); + } + } + + fn emitType(self: *EmitZIR, src: usize, ty: Type) Allocator.Error!*Inst { + switch (ty.tag()) { + .isize => return self.emitPrimitiveType(src, .isize), + .usize => return self.emitPrimitiveType(src, .usize), + .c_short => return self.emitPrimitiveType(src, .c_short), + .c_ushort => return self.emitPrimitiveType(src, .c_ushort), + .c_int => return self.emitPrimitiveType(src, .c_int), + .c_uint => return self.emitPrimitiveType(src, .c_uint), + .c_long => return self.emitPrimitiveType(src, .c_long), + .c_ulong => return self.emitPrimitiveType(src, .c_ulong), + .c_longlong => return self.emitPrimitiveType(src, .c_longlong), + .c_ulonglong => return self.emitPrimitiveType(src, .c_ulonglong), + .c_longdouble => return self.emitPrimitiveType(src, .c_longdouble), + .c_void => return self.emitPrimitiveType(src, .c_void), + .f16 => return self.emitPrimitiveType(src, .f16), + .f32 => return self.emitPrimitiveType(src, .f32), + .f64 => return self.emitPrimitiveType(src, .f64), + .f128 => return self.emitPrimitiveType(src, .f128), + .anyerror => return self.emitPrimitiveType(src, .anyerror), + else => switch (ty.zigTypeTag()) { + .Bool => return self.emitPrimitiveType(src, .bool), + .Void => return self.emitPrimitiveType(src, .void), + .NoReturn => return self.emitPrimitiveType(src, .noreturn), + .Type => return self.emitPrimitiveType(src, .type), + .ComptimeInt => return self.emitPrimitiveType(src, .comptime_int), + .ComptimeFloat => return self.emitPrimitiveType(src, .comptime_float), + .Fn => { + const param_types = try self.allocator.alloc(Type, ty.fnParamLen()); + defer self.allocator.free(param_types); + + ty.fnParamTypes(param_types); + const emitted_params = try self.arena.allocator.alloc(*Inst, param_types.len); + for (param_types) |param_type, i| { + emitted_params[i] = try self.emitType(src, param_type); + } + + const fntype_inst = try self.arena.allocator.create(Inst.FnType); + fntype_inst.* = .{ + .base = .{ + .name = try self.autoName(), + .src = src, + .tag = Inst.FnType.base_tag, + }, + .positionals = .{ + .param_types = emitted_params, + .return_type = try self.emitType(src, ty.fnReturnType()), + }, + .kw_args = .{ + .cc = ty.fnCallingConvention(), + }, + }; + try self.decls.append(self.allocator, &fntype_inst.base); + return &fntype_inst.base; + }, + else => std.debug.panic("TODO implement emitType for {}", .{ty}), + }, + } + } + + fn autoName(self: *EmitZIR) ![]u8 { + return std.fmt.allocPrint(&self.arena.allocator, "{}", .{self.decls.items.len}); + } + + fn emitPrimitiveType(self: *EmitZIR, src: usize, tag: Inst.Primitive.BuiltinType) !*Inst { + const primitive_inst = try self.arena.allocator.create(Inst.Primitive); + primitive_inst.* = .{ + .base = .{ + .name = try self.autoName(), + .src = src, + .tag = Inst.Primitive.base_tag, + }, + .positionals = .{ + .tag = tag, + }, + .kw_args = .{}, + }; + try self.decls.append(self.allocator, &primitive_inst.base); + return &primitive_inst.base; + } + + fn emitStringLiteral(self: *EmitZIR, src: usize, str: []const u8) !*Inst { + const str_inst = try self.arena.allocator.create(Inst.Str); + str_inst.* = .{ + .base = .{ + .name = try self.autoName(), + .src = src, + .tag = Inst.Str.base_tag, + }, + .positionals = .{ + .bytes = str, + }, + .kw_args = .{}, + }; + try self.decls.append(self.allocator, &str_inst.base); + + const ref_inst = try self.arena.allocator.create(Inst.Ref); + ref_inst.* = .{ + .base = .{ + .name = try self.autoName(), + .src = src, + .tag = Inst.Ref.base_tag, + }, + .positionals = .{ + .operand = &str_inst.base, + }, + .kw_args = .{}, + }; + try self.decls.append(self.allocator, &ref_inst.base); + + return &ref_inst.base; + } +};