llvm: expose deterministic archive writing

archive.go

@@ -0,0 +1,109 @@
+//===- archive.go - Bindings for archive writing --------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+//
+// This file defines bindings for llvm::writeArchive.
+//
+//===----------------------------------------------------------------------===//
+
+package llvm
+
+/*
+#include "llvm-c/Core.h"
+#include "backports.h"
+#include <stdlib.h>
+*/
+import "C"
+
+import (
+	"errors"
+	"fmt"
+	"unsafe"
+)
+
+// ArchiveMember is an object or bitcode file to add to an archive.
+//
+// A memory-buffer member borrows its buffer. The caller must keep the buffer
+// alive until WriteArchive returns.
+type ArchiveMember struct {
+	name   string
+	path   string
+	buffer MemoryBuffer
+}
+
+// NewArchiveMemberFromFile creates a path-backed archive member. The archive
+// member name is the base name of path.
+func NewArchiveMemberFromFile(path string) ArchiveMember {
+	return ArchiveMember{path: path}
+}
+
+// NewArchiveMemberFromMemoryBuffer creates a memory-backed archive member.
+// name is the name stored in the archive.
+func NewArchiveMemberFromMemoryBuffer(name string, buffer MemoryBuffer) ArchiveMember {
+	return ArchiveMember{name: name, buffer: buffer}
+}
+
+// WriteArchive creates a deterministic, non-thin archive with a symbol table.
+// targetTriple determines the platform archive format.
+func WriteArchive(path, targetTriple string, members []ArchiveMember) error {
+	if path == "" {
+		return errors.New("archive path is empty")
+	}
+	if len(members) == 0 {
+		return errors.New("archive has no members")
+	}
+
+	cTriple := C.CString(targetTriple)
+	defer C.free(unsafe.Pointer(cTriple))
+	writer := C.LLVMGoCreateArchiveWriter(cTriple)
+	if writer == nil {
+		return errors.New("failed to create archive writer")
+	}
+	defer C.LLVMGoDisposeArchiveWriter(writer)
+
+	for i, member := range members {
+		var cErr *C.char
+		switch {
+		case member.path != "":
+			cPath := C.CString(member.path)
+			var cName *C.char
+			if member.name != "" {
+				cName = C.CString(member.name)
+			}
+			cErr = C.LLVMGoArchiveWriterAddFile(writer, cPath, cName)
+			C.free(unsafe.Pointer(cPath))
+			C.free(unsafe.Pointer(cName))
+		case !member.buffer.IsNil():
+			if member.name == "" {
+				return fmt.Errorf("archive member %d has an empty name", i)
+			}
+			cName := C.CString(member.name)
+			cErr = C.LLVMGoArchiveWriterAddMemoryBuffer(writer, member.buffer.C, cName)
+			C.free(unsafe.Pointer(cName))
+		default:
+			return fmt.Errorf("archive member %d has no file or memory buffer", i)
+		}
+		if cErr != nil {
+			err := errors.New(C.GoString(cErr))
+			C.LLVMDisposeMessage(cErr)
+			label := member.name
+			if member.path != "" {
+				label = member.path
+			}
+			return fmt.Errorf("add archive member %d (%q): %w", i, label, err)
+		}
+	}
+
+	cPath := C.CString(path)
+	defer C.free(unsafe.Pointer(cPath))
+	if cErr := C.LLVMGoArchiveWriterWrite(writer, cPath); cErr != nil {
+		err := errors.New(C.GoString(cErr))
+		C.LLVMDisposeMessage(cErr)
+		return err
+	}
+	return nil
+}

archive_test.go

@@ -0,0 +1,87 @@
+//===- archive_test.go - Tests for archive writing ------------------------===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+
+package llvm
+
+import (
+	"bytes"
+	"os"
+	"path/filepath"
+	"strings"
+	"testing"
+)
+
+func TestWriteArchiveMemoryAndFileMembers(t *testing.T) {
+	ctx := NewContext()
+	defer ctx.Dispose()
+
+	mod := ctx.NewModule("archive")
+	defer mod.Dispose()
+	mod.SetTarget("x86_64-unknown-linux-gnu")
+	AddFunction(mod, "answer", FunctionType(ctx.Int32Type(), nil, false))
+
+	memoryBuf := WriteBitcodeToMemoryBuffer(mod)
+	defer memoryBuf.Dispose()
+
+	fileBuf := WriteBitcodeToMemoryBuffer(mod)
+	defer fileBuf.Dispose()
+	filePath := filepath.Join(t.TempDir(), "from-file.bc")
+	if err := os.WriteFile(filePath, fileBuf.Bytes(), 0o644); err != nil {
+		t.Fatal(err)
+	}
+
+	archivePath := filepath.Join(t.TempDir(), "libarchive.a")
+	err := WriteArchive(archivePath, mod.Target(), []ArchiveMember{
+		NewArchiveMemberFromMemoryBuffer("from-memory.bc", memoryBuf),
+		NewArchiveMemberFromFile(filePath),
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+
+	data, err := os.ReadFile(archivePath)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if !bytes.HasPrefix(data, []byte("!<arch>\n")) {
+		end := len(data)
+		if end > 8 {
+			end = 8
+		}
+		t.Fatalf("archive magic = %q", data[:end])
+	}
+	for _, name := range []string{"from-memory.bc", "from-file.bc"} {
+		if !bytes.Contains(data, []byte(name)) {
+			t.Errorf("archive does not contain member name %q", name)
+		}
+	}
+}
+
+func TestWriteArchiveRejectsInvalidMembers(t *testing.T) {
+	path := filepath.Join(t.TempDir(), "invalid.a")
+	missing := filepath.Join(t.TempDir(), "missing.o")
+	tests := []struct {
+		name    string
+		path    string
+		members []ArchiveMember
+		want    string
+	}{
+		{name: "empty path", members: []ArchiveMember{{}}, want: "archive path is empty"},
+		{name: "no members", path: path, want: "archive has no members"},
+		{name: "empty member", path: path, members: []ArchiveMember{{}}, want: "has no file or memory buffer"},
+		{name: "missing file", path: path, members: []ArchiveMember{NewArchiveMemberFromFile(missing)}, want: missing},
+	}
+	for _, tt := range tests {
+		t.Run(tt.name, func(t *testing.T) {
+			err := WriteArchive(tt.path, "x86_64-unknown-linux-gnu", tt.members)
+			if err == nil || !strings.Contains(err.Error(), tt.want) {
+				t.Fatalf("WriteArchive error = %v, want %q", err, tt.want)
+			}
+		})
+	}
+}

backports.cpp

@@ -15,10 +15,111 @@
 #include "llvm/Transforms/IPO/PassManagerBuilder.h"
 #endif
 #include "llvm/IR/Module.h"
+#include "llvm/Object/Archive.h"
+#include "llvm/Object/ArchiveWriter.h"
 #include "llvm/Pass.h"
+#include "llvm/Support/Error.h"
+#if LLVM_VERSION_MAJOR >= 18
+#include "llvm/TargetParser/Host.h"
+#else
+#include "llvm/Support/Host.h"
+#endif
 #include "llvm/Support/MemoryBuffer.h"
+#include "llvm/Support/Path.h"
+#if LLVM_VERSION_MAJOR >= 16
+#include "llvm/TargetParser/Triple.h"
+#else
+#include "llvm/ADT/Triple.h"
+#endif
 #include "llvm/Transforms/IPO.h"
+#include "llvm-c/Core.h"
 #include "llvm-c/DebugInfo.h"
+#include <deque>
+
+struct LLVMGoArchiveWriterOpaque {
+  llvm::object::Archive::Kind Kind;
+  std::vector<llvm::NewArchiveMember> Members;
+  std::deque<std::string> Names;
+};
+
+static char *LLVMGoArchiveError(llvm::Error Err) {
+  if (!Err)
+    return nullptr;
+  return LLVMCreateMessage(llvm::toString(std::move(Err)).c_str());
+}
+
+LLVMGoArchiveWriterRef LLVMGoCreateArchiveWriter(const char *TargetTriple) {
+  auto TripleText = TargetTriple && TargetTriple[0]
+                        ? llvm::Triple::normalize(TargetTriple)
+                        : llvm::sys::getDefaultTargetTriple();
+  llvm::Triple Triple(TripleText);
+  auto Kind = llvm::object::Archive::K_GNU;
+  if (Triple.isOSDarwin())
+    Kind = llvm::object::Archive::K_DARWIN;
+  else if (Triple.isOSAIX())
+    Kind = llvm::object::Archive::K_AIXBIG;
+  else if (Triple.isOSWindows())
+    Kind = llvm::object::Archive::K_COFF;
+  return new LLVMGoArchiveWriterOpaque{Kind, {}, {}};
+}
+
+char *LLVMGoArchiveWriterAddFile(LLVMGoArchiveWriterRef Writer,
+                                 const char *Path, const char *Name) {
+  if (!Writer)
+    return LLVMCreateMessage("archive writer is nil");
+  if (!Path || !Path[0])
+    return LLVMCreateMessage("archive member path is empty");
+
+  auto MemberOrErr = llvm::NewArchiveMember::getFile(Path, true);
+  if (!MemberOrErr)
+    return LLVMGoArchiveError(MemberOrErr.takeError());
+
+  Writer->Names.emplace_back(
+      Name && Name[0] ? Name
+                      : llvm::sys::path::filename(Path).str());
+  MemberOrErr->MemberName = Writer->Names.back();
+  Writer->Members.push_back(std::move(*MemberOrErr));
+  return nullptr;
+}
+
+char *LLVMGoArchiveWriterAddMemoryBuffer(LLVMGoArchiveWriterRef Writer,
+                                         LLVMMemoryBufferRef Buffer,
+                                         const char *Name) {
+  if (!Writer)
+    return LLVMCreateMessage("archive writer is nil");
+  if (!Buffer)
+    return LLVMCreateMessage("archive member buffer is nil");
+  if (!Name || !Name[0])
+    return LLVMCreateMessage("archive member name is empty");
+
+  Writer->Names.emplace_back(Name);
+  Writer->Members.emplace_back(llvm::unwrap(Buffer)->getMemBufferRef());
+  Writer->Members.back().MemberName = Writer->Names.back();
+  return nullptr;
+}
+
+char *LLVMGoArchiveWriterWrite(LLVMGoArchiveWriterRef Writer,
+                               const char *ArchivePath) {
+  if (!Writer)
+    return LLVMCreateMessage("archive writer is nil");
+  if (!ArchivePath || !ArchivePath[0])
+    return LLVMCreateMessage("archive path is empty");
+  if (Writer->Members.empty())
+    return LLVMCreateMessage("archive has no members");
+
+#if LLVM_VERSION_MAJOR >= 18
+  auto WriteSymtab = llvm::SymtabWritingMode::NormalSymtab;
+#else
+  auto WriteSymtab = true;
+#endif
+  return LLVMGoArchiveError(llvm::writeArchive(
+      ArchivePath, Writer->Members, WriteSymtab, Writer->Kind,
+      /*Deterministic=*/true, /*Thin=*/false));
+}
+
+void LLVMGoDisposeArchiveWriter(LLVMGoArchiveWriterRef Writer) {
+  delete Writer;
+}
 
 void LLVMGlobalObjectAddMetadata(LLVMValueRef Global, unsigned KindID, LLVMMetadataRef MD) {
   llvm::MDNode *N = MD ? llvm::unwrap<llvm::MDNode>(MD) : nullptr;

backports.h

@@ -13,6 +13,22 @@ LLVMMemoryBufferRef LLVMGoWriteThinLTOBitcodeToMemoryBuffer(LLVMModuleRef M);
 LLVMMemoryBufferRef LLVMGoWriteFullLTOBitcodeToMemoryBuffer(
     LLVMModuleRef M, LLVMBool EnableSplitLTOUnit);
 
+typedef struct LLVMGoArchiveWriterOpaque *LLVMGoArchiveWriterRef;
+
+LLVMGoArchiveWriterRef LLVMGoCreateArchiveWriter(const char *TargetTriple);
+
+char *LLVMGoArchiveWriterAddFile(LLVMGoArchiveWriterRef Writer,
+                                 const char *Path, const char *Name);
+
+char *LLVMGoArchiveWriterAddMemoryBuffer(LLVMGoArchiveWriterRef Writer,
+                                         LLVMMemoryBufferRef Buffer,
+                                         const char *Name);
+
+char *LLVMGoArchiveWriterWrite(LLVMGoArchiveWriterRef Writer,
+                               const char *ArchivePath);
+
+void LLVMGoDisposeArchiveWriter(LLVMGoArchiveWriterRef Writer);
+
 void LLVMGoDIBuilderInsertDbgValueRecordAtEnd(
     LLVMDIBuilderRef Builder, LLVMValueRef Val, LLVMMetadataRef VarInfo,
     LLVMMetadataRef Expr, LLVMMetadataRef DebugLoc, LLVMBasicBlockRef Block);