[llvm] [CAS] Add OnDiskTrieRawHashMap (PR #114100)

[Steven Wu via llvm-commits]

================
@@ -0,0 +1,153 @@
+//===----------------------------------------------------------------------===//
+//
+// 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
+//
+//===----------------------------------------------------------------------===//
+//
+/// \file
+/// This file declares the common interface for a DatabaseFile that is used to
+/// implement OnDiskCAS.
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_LIB_CAS_DATABASEFILE_H
+#define LLVM_LIB_CAS_DATABASEFILE_H
+
+#include "llvm/ADT/StringRef.h"
+#include "llvm/CAS/MappedFileRegionArena.h"
+#include "llvm/Support/Error.h"
+
+namespace llvm::cas::ondisk {
+
+using MappedFileRegion = MappedFileRegionArena::RegionT;
+
+/// Generic handle for a table.
+///
+/// Generic table header layout:
+/// - 2-bytes: TableKind
+/// - 2-bytes: TableNameSize
+/// - 4-bytes: TableNameRelOffset (relative to header)
+class TableHandle {
+public:
+  enum class TableKind : uint16_t {
+    TrieRawHashMap = 1,
+    DataAllocator = 2,
+  };
+  struct Header {
+    TableKind Kind;
+    uint16_t NameSize;
+    int32_t NameRelOffset; ///< Relative to Header.
+  };
+
+  explicit operator bool() const { return H; }
+  const Header &getHeader() const { return *H; }
+  MappedFileRegion &getRegion() const { return *Region; }
+
+  template <class T> static void check() {
+    static_assert(
+        std::is_same<decltype(T::Header::GenericHeader), Header>::value,
+        "T::GenericHeader should be of type TableHandle::Header");
+    static_assert(offsetof(typename T::Header, GenericHeader) == 0,
+                  "T::GenericHeader must be the head of T::Header");
+  }
+  template <class T> bool is() const { return T::Kind == H->Kind; }
+  template <class T> T dyn_cast() const {
+    check<T>();
+    if (is<T>())
+      return T(*Region, *reinterpret_cast<typename T::Header *>(H));
+    return T();
+  }
+  template <class T> T cast() const {
+    assert(is<T>());
+    return dyn_cast<T>();
+  }
+
+  StringRef getName() const {
+    auto *Begin = reinterpret_cast<const char *>(H) + H->NameRelOffset;
+    return StringRef(Begin, H->NameSize);
+  }
+
+  TableHandle() = default;
+  TableHandle(MappedFileRegion &Region, Header &H) : Region(&Region), H(&H) {}
+  TableHandle(MappedFileRegion &Region, intptr_t HeaderOffset)
+      : TableHandle(Region,
+                    *reinterpret_cast<Header *>(Region.data() + HeaderOffset)) {
+  }
+
+private:
+  MappedFileRegion *Region = nullptr;
+  Header *H = nullptr;
+};
+
+/// Encapsulate a database file, which:
+/// - Sets/checks magic.
+/// - Sets/checks version.
+/// - Points at an arbitrary root table.
+/// - Sets up a MappedFileRegionArena for allocation.
+///
+/// Top-level layout:
+/// - 8-bytes: Magic
+/// - 8-bytes: Version
+/// - 8-bytes: RootTable (16-bits: Kind; 48-bits: Offset)
+/// - 8-bytes: BumpPtr
+class DatabaseFile {
+public:
+  static constexpr uint64_t getMagic() { return 0x00FFDA7ABA53FF00ULL; }
+  static constexpr uint64_t getVersion() { return 1ULL; }
+  struct Header {
+    uint64_t Magic;
+    uint64_t Version;
----------------
cachemeifyoucan wrote:

The magic is picked in the first ever commit and no one ever gives it a second thought, maybe because this header only appears once per file so the space it takes is not a big deal.

The common magic is 32 bits, so I guess it is possible we can just use that. Note assume 64 bits in needed for Offset and we want to make the atomic boundary of TableOffset, we have 64 bits before that anyway. A 32 bit magic + 32 bit for version is probably enough.

https://github.com/llvm/llvm-project/pull/114100