[llvm] [CGData] llvm-cgdata (PR #89884)

[Ellis Hoag via llvm-commits]

================
@@ -0,0 +1,174 @@
+//===- CodeGenDataReader.cpp ----------------------------------------------===//
+//
+// 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 contains support for reading codegen data.
+//
+//===----------------------------------------------------------------------===//
+
+#include "llvm/CodeGenData/CodeGenDataReader.h"
+#include "llvm/CodeGenData/OutlinedHashTreeRecord.h"
+#include "llvm/Support/MemoryBuffer.h"
+
+#define DEBUG_TYPE "cg-data-reader"
+
+using namespace llvm;
+
+namespace llvm {
+
+static Expected<std::unique_ptr<MemoryBuffer>>
+setupMemoryBuffer(const Twine &Filename, vfs::FileSystem &FS) {
+  auto BufferOrErr = Filename.str() == "-" ? MemoryBuffer::getSTDIN()
+                                           : FS.getBufferForFile(Filename);
+  if (std::error_code EC = BufferOrErr.getError())
+    return errorCodeToError(EC);
+  return std::move(BufferOrErr.get());
+}
+
+Error CodeGenDataReader::mergeFromObjectFile(
+    const object::ObjectFile *Obj,
+    OutlinedHashTreeRecord &GlobalOutlineRecord) {
+  Triple TT = Obj->makeTriple();
+  auto CGOutLineName =
+      getCodeGenDataSectionName(CG_outline, TT.getObjectFormat(), false);
+
+  for (auto &Section : Obj->sections()) {
+    Expected<StringRef> NameOrErr = Section.getName();
+    if (!NameOrErr)
+      return NameOrErr.takeError();
+    Expected<StringRef> ContentsOrErr = Section.getContents();
+    if (!ContentsOrErr)
+      return ContentsOrErr.takeError();
+    auto *Data = reinterpret_cast<const unsigned char *>(ContentsOrErr->data());
+    auto *EndData = Data + ContentsOrErr->size();
+
+    if (*NameOrErr == CGOutLineName) {
+      // In case dealing with an executable that has concatenaed cgdata,
+      // we want to merge them into a single cgdata.
+      // Although it's not a typical workflow, we support this scenario.
+      while (Data != EndData) {
+        OutlinedHashTreeRecord LocalOutlineRecord;
+        LocalOutlineRecord.deserialize(Data);
+        GlobalOutlineRecord.merge(LocalOutlineRecord);
+      }
+    }
+    // TODO: Add support for other cgdata sections.
+  }
+
+  return Error::success();
+}
+
+Error IndexedCodeGenDataReader::read() {
+  using namespace support;
+
+  // The smallest header with the version 1 is 24 bytes
+  const unsigned MinHeaderSize = 24;
+  if (DataBuffer->getBufferSize() < MinHeaderSize)
+    return error(cgdata_error::bad_header);
+
+  auto *Start =
+      reinterpret_cast<const unsigned char *>(DataBuffer->getBufferStart());
+  auto *End =
+      reinterpret_cast<const unsigned char *>(DataBuffer->getBufferEnd());
+  auto HeaderOr = IndexedCGData::Header::readFromBuffer(Start);
+  if (!HeaderOr)
+    return HeaderOr.takeError();
+  Header = HeaderOr.get();
+
+  if (hasOutlinedHashTree()) {
+    const unsigned char *Ptr = Start + Header.OutlinedHashTreeOffset;
+    if (Ptr >= End)
+      return error(cgdata_error::eof);
+    HashTreeRecord.deserialize(Ptr);
+  }
+
+  return success();
+}
+
+Expected<std::unique_ptr<CodeGenDataReader>>
+CodeGenDataReader::create(const Twine &Path, vfs::FileSystem &FS) {
+  // Set up the buffer to read.
+  auto BufferOrError = setupMemoryBuffer(Path, FS);
+  if (Error E = BufferOrError.takeError())
+    return std::move(E);
+  return CodeGenDataReader::create(std::move(BufferOrError.get()));
+}
+
+Expected<std::unique_ptr<CodeGenDataReader>>
+CodeGenDataReader::create(std::unique_ptr<MemoryBuffer> Buffer) {
+  if (Buffer->getBufferSize() == 0)
+    return make_error<CGDataError>(cgdata_error::empty_cgdata);
+
+  std::unique_ptr<CodeGenDataReader> Reader;
+  // Create the reader.
+  if (IndexedCodeGenDataReader::hasFormat(*Buffer))
+    Reader.reset(new IndexedCodeGenDataReader(std::move(Buffer)));
+  else if (TextCodeGenDataReader::hasFormat(*Buffer))
+    Reader.reset(new TextCodeGenDataReader(std::move(Buffer)));
+  else
+    return make_error<CGDataError>(cgdata_error::malformed);
+
+  // Initialize the reader and return the result.
+  if (Error E = Reader->read())
+    return std::move(E);
+
+  return std::move(Reader);
+}
+
+bool IndexedCodeGenDataReader::hasFormat(const MemoryBuffer &DataBuffer) {
+  using namespace support;
+  if (DataBuffer.getBufferSize() < 8)
+    return false;
+
+  uint64_t Magic = endian::read<uint64_t, llvm::endianness::little, aligned>(
+      DataBuffer.getBufferStart());
+  // Verify that it's magical.
+  return Magic == IndexedCGData::Magic;
+}
+
+bool TextCodeGenDataReader::hasFormat(const MemoryBuffer &Buffer) {
+  // Verify that this really looks like plain ASCII text by checking a
+  // 'reasonable' number of characters (up to profile magic size).
+  size_t count = std::min(Buffer.getBufferSize(), sizeof(uint64_t));
+  StringRef buffer = Buffer.getBufferStart();
+  return count == 0 ||
+         std::all_of(buffer.begin(), buffer.begin() + count,
+                     [](char c) { return isPrint(c) || isSpace(c); });
+}
+Error TextCodeGenDataReader::read() {
+  using namespace support;
+
+  // Parse the custom header line by line.
+  while (Line->starts_with(":")) {
+    StringRef Str = Line->substr(1);
----------------
ellishg wrote:

If we want to handle whitespace.
```suggestion
    StringRef Str = Line->drop_front().rtrim();
```

https://github.com/llvm/llvm-project/pull/89884