summaryrefslogtreecommitdiffstats
path: root/toolkit/components/url-classifier/HashStore.h
blob: ccb665efa7e3a69948db7f010296af57b2be9faf (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
/* This Source Code Form is subject to the terms of the Mozilla Public
 * License, v. 2.0. If a copy of the MPL was not distributed with this
 * file, You can obtain one at http://mozilla.org/MPL/2.0/. */

#ifndef HashStore_h__
#define HashStore_h__

#include "Entries.h"
#include "ChunkSet.h"

#include "nsString.h"
#include "nsTArray.h"
#include "nsIFile.h"
#include "nsISupports.h"
#include "nsCOMPtr.h"
#include <string>

namespace mozilla {
namespace safebrowsing {

// The abstract class of TableUpdateV2 and TableUpdateV4. This
// is convenient for passing the TableUpdate* around associated
// with v2 and v4 instance.
class TableUpdate {
 public:
  TableUpdate(const nsACString& aTable) : mTable(aTable) {}

  NS_INLINE_DECL_THREADSAFE_REFCOUNTING(TableUpdate);

  // To be overriden.
  virtual bool Empty() const = 0;

  // Common interfaces.
  const nsCString& TableName() const { return mTable; }

  template <typename T>
  static T* Cast(TableUpdate* aThat) {
    return (T::TAG == aThat->Tag() ? reinterpret_cast<T*>(aThat) : nullptr);
  }
  template <typename T>
  static const T* Cast(const TableUpdate* aThat) {
    return (T::TAG == aThat->Tag() ? reinterpret_cast<const T*>(aThat)
                                   : nullptr);
  }

 protected:
  virtual ~TableUpdate() = default;

 private:
  virtual int Tag() const = 0;

  const nsCString mTable;
};

typedef nsTArray<RefPtr<TableUpdate>> TableUpdateArray;
typedef nsTArray<RefPtr<const TableUpdate>> ConstTableUpdateArray;

// A table update is built from a single update chunk from the server. As the
// protocol parser processes each chunk, it constructs a table update with the
// new hashes.
class TableUpdateV2 : public TableUpdate {
 public:
  explicit TableUpdateV2(const nsACString& aTable) : TableUpdate(aTable) {}

  bool Empty() const override {
    return mAddChunks.Length() == 0 && mSubChunks.Length() == 0 &&
           mAddExpirations.Length() == 0 && mSubExpirations.Length() == 0 &&
           mAddPrefixes.Length() == 0 && mSubPrefixes.Length() == 0 &&
           mAddCompletes.Length() == 0 && mSubCompletes.Length() == 0 &&
           mMissPrefixes.Length() == 0;
  }

  // Throughout, uint32_t aChunk refers only to the chunk number. Chunk data is
  // stored in the Prefix structures.
  [[nodiscard]] nsresult NewAddChunk(uint32_t aChunk) {
    return mAddChunks.Set(aChunk);
  };
  [[nodiscard]] nsresult NewSubChunk(uint32_t aChunk) {
    return mSubChunks.Set(aChunk);
  };
  [[nodiscard]] nsresult NewAddExpiration(uint32_t aChunk) {
    return mAddExpirations.Set(aChunk);
  };
  [[nodiscard]] nsresult NewSubExpiration(uint32_t aChunk) {
    return mSubExpirations.Set(aChunk);
  };
  [[nodiscard]] nsresult NewAddPrefix(uint32_t aAddChunk,
                                      const Prefix& aPrefix);
  [[nodiscard]] nsresult NewSubPrefix(uint32_t aAddChunk, const Prefix& aPrefix,
                                      uint32_t aSubChunk);
  [[nodiscard]] nsresult NewAddComplete(uint32_t aChunk,
                                        const Completion& aCompletion);
  [[nodiscard]] nsresult NewSubComplete(uint32_t aAddChunk,
                                        const Completion& aCompletion,
                                        uint32_t aSubChunk);
  [[nodiscard]] nsresult NewMissPrefix(const Prefix& aPrefix);

  const ChunkSet& AddChunks() const { return mAddChunks; }
  const ChunkSet& SubChunks() const { return mSubChunks; }

  // Expirations for chunks.
  const ChunkSet& AddExpirations() const { return mAddExpirations; }
  const ChunkSet& SubExpirations() const { return mSubExpirations; }

  // Hashes associated with this chunk.
  AddPrefixArray& AddPrefixes() { return mAddPrefixes; }
  SubPrefixArray& SubPrefixes() { return mSubPrefixes; }
  const AddCompleteArray& AddCompletes() const { return mAddCompletes; }
  AddCompleteArray& AddCompletes() { return mAddCompletes; }
  SubCompleteArray& SubCompletes() { return mSubCompletes; }

  // Entries that cannot be completed.
  const MissPrefixArray& MissPrefixes() const { return mMissPrefixes; }

  // For downcasting.
  static const int TAG = 2;

 private:
  // The list of chunk numbers that we have for each of the type of chunks.
  ChunkSet mAddChunks;
  ChunkSet mSubChunks;
  ChunkSet mAddExpirations;
  ChunkSet mSubExpirations;

  // 4-byte sha256 prefixes.
  AddPrefixArray mAddPrefixes;
  SubPrefixArray mSubPrefixes;

  // This is only used by gethash so don't add this to Header.
  MissPrefixArray mMissPrefixes;

  // 32-byte hashes.
  AddCompleteArray mAddCompletes;
  SubCompleteArray mSubCompletes;

  virtual int Tag() const override { return TAG; }
};

// Structure for DBService/HashStore/Classifiers to update.
// It would contain the prefixes (both fixed and variable length)
// for addition and indices to removal. See Bug 1283009.
class TableUpdateV4 : public TableUpdate {
 public:
  typedef nsTArray<int32_t> RemovalIndiceArray;

 public:
  explicit TableUpdateV4(const nsACString& aTable)
      : TableUpdate(aTable), mFullUpdate(false) {}

  bool Empty() const override {
    return mPrefixesMap.IsEmpty() && mRemovalIndiceArray.IsEmpty() &&
           mFullHashResponseMap.IsEmpty();
  }

  bool IsFullUpdate() const { return mFullUpdate; }
  const PrefixStringMap& Prefixes() const { return mPrefixesMap; }
  const RemovalIndiceArray& RemovalIndices() const {
    return mRemovalIndiceArray;
  }
  const nsACString& ClientState() const { return mClientState; }
  const nsACString& SHA256() const { return mSHA256; }
  const FullHashResponseMap& FullHashResponse() const {
    return mFullHashResponseMap;
  }

  // For downcasting.
  static const int TAG = 4;

  void SetFullUpdate(bool aIsFullUpdate) { mFullUpdate = aIsFullUpdate; }
  void NewPrefixes(int32_t aSize, const nsACString& aPrefixes);
  void SetNewClientState(const nsACString& aState) { mClientState = aState; }
  void SetSHA256(const std::string& aSHA256);

  nsresult NewRemovalIndices(const uint32_t* aIndices, size_t aNumOfIndices);
  nsresult NewFullHashResponse(const Prefix& aPrefix,
                               const CachedFullHashResponse& aResponse);

  // Clear Prefixes & Removal indice.
  void Clear();

 private:
  virtual int Tag() const override { return TAG; }

  bool mFullUpdate;
  PrefixStringMap mPrefixesMap;
  RemovalIndiceArray mRemovalIndiceArray;
  nsCString mClientState;
  nsCString mSHA256;

  // This is used to store response from fullHashes.find.
  FullHashResponseMap mFullHashResponseMap;
};

// There is one hash store per table.
class HashStore {
 public:
  HashStore(const nsACString& aTableName, const nsACString& aProvider,
            nsIFile* aRootStoreFile);
  ~HashStore();

  const nsCString& TableName() const { return mTableName; }

  // Version is set to 0 by default, it is only used when we want to open
  // a specific version of HashStore. Note that the intention of aVersion
  // is only to pass SanityCheck, reading data from older version should
  // be handled additionally.
  nsresult Open(uint32_t aVersion = 0);

  // Add Prefixes/Completes are stored partly in the PrefixSet (contains the
  // Prefix data organized for fast lookup/low RAM usage) and partly in the
  // HashStore (Add Chunk numbers - only used for updates, slow retrieval).
  // AugmentAdds function joins the separate datasets into one complete
  // prefixes+chunknumbers dataset.
  nsresult AugmentAdds(const nsTArray<uint32_t>& aPrefixes,
                       const nsTArray<nsCString>& aCompletes);

  ChunkSet& AddChunks();
  ChunkSet& SubChunks();
  AddPrefixArray& AddPrefixes() { return mAddPrefixes; }
  SubPrefixArray& SubPrefixes() { return mSubPrefixes; }
  AddCompleteArray& AddCompletes() { return mAddCompletes; }
  SubCompleteArray& SubCompletes() { return mSubCompletes; }

  // =======
  // Updates
  // =======
  // Begin the update process.  Reads the store into memory.
  nsresult BeginUpdate();

  // Imports the data from a TableUpdate.
  nsresult ApplyUpdate(RefPtr<TableUpdateV2> aUpdate);

  // Process expired chunks
  nsresult Expire();

  // Rebuild the store, Incorporating all the applied updates.
  nsresult Rebuild();

  // Write the current state of the store to disk.
  // If you call between ApplyUpdate() and Rebuild(), you'll
  // have a mess on your hands.
  nsresult WriteFile();

  nsresult ReadCompletionsLegacyV3(AddCompleteArray& aCompletes);

  nsresult Reset();

 private:
  nsresult ReadHeader();
  nsresult SanityCheck(uint32_t aVersion = 0) const;
  nsresult CalculateChecksum(nsAutoCString& aChecksum, uint32_t aFileSize,
                             bool aChecksumPresent);
  nsresult CheckChecksum(uint32_t aFileSize);
  void UpdateHeader();

  nsresult ReadCompletions();
  nsresult ReadChunkNumbers();
  nsresult ReadHashes();

  nsresult ReadAddPrefixes();
  nsresult ReadSubPrefixes();
  nsresult ReadAddCompletes();

  nsresult WriteAddPrefixChunks(nsIOutputStream* aOut);
  nsresult WriteSubPrefixes(nsIOutputStream* aOut);
  nsresult WriteAddCompleteChunks(nsIOutputStream* aOut);

  nsresult ProcessSubs();

  nsresult PrepareForUpdate();

  // This is used for checking that the database is correct and for figuring out
  // the number of chunks, etc. to read from disk on restart.
  struct Header {
    uint32_t magic;
    uint32_t version;
    uint32_t numAddChunks;
    uint32_t numSubChunks;
    uint32_t numAddPrefixes;
    uint32_t numSubPrefixes;
    uint32_t numAddCompletes;
    uint32_t numSubCompletes;
  };

  Header mHeader;

  // The name of the table (must end in -shavar or -digest256, or evidently
  // -simple for unittesting.
  const nsCString mTableName;
  nsCOMPtr<nsIFile> mStoreDirectory;

  bool mInUpdate;

  nsCOMPtr<nsIInputStream> mInputStream;

  // Chunk numbers, stored as uint32_t arrays.
  ChunkSet mAddChunks;
  ChunkSet mSubChunks;

  ChunkSet mAddExpirations;
  ChunkSet mSubExpirations;

  // Chunk data for shavar tables. See Entries.h for format.
  AddPrefixArray mAddPrefixes;
  SubPrefixArray mSubPrefixes;

  // See bug 806422 for background. We must be able to distinguish between
  // updates from the completion server and updates from the regular server.
  AddCompleteArray mAddCompletes;
  SubCompleteArray mSubCompletes;

  uint32_t mFileSize;

  // For gtest to inspect private members.
  friend class PerProviderDirectoryTestUtils;
};

}  // namespace safebrowsing
}  // namespace mozilla

#endif