• Home
  • Line#
  • Scopes#
  • Navigate#
  • Raw
  • Download
1 // Copyright 2007, 2008 Google Inc.
2 // Authors: Jeff Dean, Sanjay Ghemawat, Lincoln Smith
3 //
4 // Licensed under the Apache License, Version 2.0 (the "License");
5 // you may not use this file except in compliance with the License.
6 // You may obtain a copy of the License at
7 //
8 //      http://www.apache.org/licenses/LICENSE-2.0
9 //
10 // Unless required by applicable law or agreed to in writing, software
11 // distributed under the License is distributed on an "AS IS" BASIS,
12 // WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13 // See the License for the specific language governing permissions and
14 // limitations under the License.
15 
16 #ifndef OPEN_VCDIFF_ROLLING_HASH_H_
17 #define OPEN_VCDIFF_ROLLING_HASH_H_
18 
19 #include <config.h>
20 #include <stdint.h>  // uint32_t
21 #include "logging.h"
22 
23 namespace open_vcdiff {
24 
25 // Rabin-Karp hasher module -- this is a faster version with different
26 // constants, so it's not quite Rabin-Karp fingerprinting, but its behavior is
27 // close enough for most applications.
28 
29 // Definitions common to all hash window sizes.
30 class RollingHashUtil {
31  public:
32   // Multiplier for incremental hashing.  The compiler should be smart enough to
33   // convert (val * kMult) into ((val << 8) + val).
34   static const uint32_t kMult = 257;
35 
36   // All hashes are returned modulo "kBase".  Current implementation requires
37   // kBase <= 2^32/kMult to avoid overflow.  Also, kBase must be a power of two
38   // so that we can compute modulus efficiently.
39   static const uint32_t kBase = (1 << 23);
40 
41   // Returns operand % kBase, assuming that kBase is a power of two.
ModBase(uint32_t operand)42   static inline uint32_t ModBase(uint32_t operand) {
43     return operand & (kBase - 1);
44   }
45 
46   // Given an unsigned integer "operand", returns an unsigned integer "result"
47   // such that
48   //     result < kBase
49   // and
50   //     ModBase(operand + result) == 0
FindModBaseInverse(uint32_t operand)51   static inline uint32_t FindModBaseInverse(uint32_t operand) {
52     // The subtraction (0 - operand) produces an unsigned underflow for any
53     // operand except 0.  The underflow results in a (very large) unsigned
54     // number.  Binary subtraction is used instead of unary negation because
55     // some compilers (e.g. Visual Studio 7+) produce a warning if an unsigned
56     // value is negated.
57     //
58     // The C++ mod operation (operand % kBase) may produce different results for
59     // different compilers if operand is negative.  That is not a problem in
60     // this case, since all numbers used are unsigned, and ModBase does its work
61     // using bitwise arithmetic rather than the % operator.
62     return ModBase(uint32_t(0) - operand);
63   }
64 
65   // Here's the heart of the hash algorithm.  Start with a partial_hash value of
66   // 0, and run this HashStep once against each byte in the data window to be
67   // hashed.  The result will be the hash value for the entire data window.  The
68   // Hash() function, below, does exactly this, albeit with some refinements.
HashStep(uint32_t partial_hash,unsigned char next_byte)69   static inline uint32_t HashStep(uint32_t partial_hash,
70                                   unsigned char next_byte) {
71     return ModBase((partial_hash * kMult) + next_byte);
72   }
73 
74   // Use this function to start computing a new hash value based on the first
75   // two bytes in the window.  It is equivalent to calling
76   //     HashStep(HashStep(0, ptr[0]), ptr[1])
77   // but takes advantage of the fact that the maximum value of
78   // (ptr[0] * kMult) + ptr[1] is not large enough to exceed kBase, thus
79   // avoiding an unnecessary ModBase operation.
HashFirstTwoBytes(const char * ptr)80   static inline uint32_t HashFirstTwoBytes(const char* ptr) {
81     return (static_cast<unsigned char>(ptr[0]) * kMult)
82         + static_cast<unsigned char>(ptr[1]);
83   }
84  private:
85   // Making these private avoids copy constructor and assignment operator.
86   // No objects of this type should be constructed.
87   RollingHashUtil();
88   RollingHashUtil(const RollingHashUtil&);  // NOLINT
89   void operator=(const RollingHashUtil&);
90 };
91 
92 // window_size must be >= 2.
93 template<int window_size>
94 class RollingHash {
95  public:
96   // Perform global initialization that is required in order to instantiate a
97   // RollingHash.  This function *must* be called (preferably on startup) by any
98   // program that uses a RollingHash.  It is harmless to call this function more
99   // than once.  It is not thread-safe, but calling it from two different
100   // threads at the same time can only cause a memory leak, not incorrect
101   // behavior.  Make sure to call it before spawning any threads that could use
102   // RollingHash.  The function returns true if initialization succeeds, or
103   // false if initialization fails, in which case the caller should not proceed
104   // to construct any objects of type RollingHash.
105   static bool Init();
106 
107   // Initialize hasher to maintain a window of the specified size.  You need an
108   // instance of this type to use UpdateHash(), but Hash() does not depend on
109   // remove_table_, so it is static.
RollingHash()110   RollingHash() {
111     if (!remove_table_) {
112       LOG(DFATAL) << "RollingHash object instantiated"
113                      " before calling RollingHash::Init()" << LOG_ENDL;
114     }
115   }
116 
117   // Compute a hash of the window "ptr[0, window_size - 1]".
Hash(const char * ptr)118   static uint32_t Hash(const char* ptr) {
119     uint32_t h = RollingHashUtil::HashFirstTwoBytes(ptr);
120     for (int i = 2; i < window_size; ++i) {
121       h = RollingHashUtil::HashStep(h, ptr[i]);
122     }
123     return h;
124   }
125 
126   // Update a hash by removing the oldest byte and adding a new byte.
127   //
128   // UpdateHash takes the hash value of buffer[0] ... buffer[window_size -1]
129   // along with the value of buffer[0] (the "old_first_byte" argument)
130   // and the value of buffer[window_size] (the "new_last_byte" argument).
131   // It quickly computes the hash value of buffer[1] ... buffer[window_size]
132   // without having to run Hash() on the entire window.
133   //
134   // The larger the window, the more advantage comes from using UpdateHash()
135   // (which runs in time independent of window_size) instead of Hash().
136   // Each time window_size doubles, the time to execute Hash() also doubles,
137   // while the time to execute UpdateHash() remains constant.  Empirical tests
138   // have borne out this statement.
UpdateHash(uint32_t old_hash,const char old_first_byte,const char new_last_byte)139   uint32_t UpdateHash(uint32_t old_hash,
140                       const char old_first_byte,
141                       const char new_last_byte) const {
142     uint32_t partial_hash = RemoveFirstByteFromHash(old_hash, old_first_byte);
143     return RollingHashUtil::HashStep(partial_hash, new_last_byte);
144   }
145 
146  protected:
147   // Given a full hash value for buffer[0] ... buffer[window_size -1], plus the
148   // value of the first byte buffer[0], this function returns a *partial* hash
149   // value for buffer[1] ... buffer[window_size -1].  See the comments in
150   // Init(), below, for a description of how the contents of remove_table_ are
151   // computed.
RemoveFirstByteFromHash(uint32_t full_hash,unsigned char first_byte)152   static uint32_t RemoveFirstByteFromHash(uint32_t full_hash,
153                                           unsigned char first_byte) {
154     return RollingHashUtil::ModBase(full_hash + remove_table_[first_byte]);
155   }
156 
157  private:
158   // We keep a table that maps from any byte "b" to
159   //    (- b * pow(kMult, window_size - 1)) % kBase
160   static const uint32_t* remove_table_;
161 };
162 
163 // For each window_size, fill a 256-entry table such that
164 //        the hash value of buffer[0] ... buffer[window_size - 1]
165 //      + remove_table_[buffer[0]]
166 //     == the hash value of buffer[1] ... buffer[window_size - 1]
167 // See the comments in Init(), below, for a description of how the contents of
168 // remove_table_ are computed.
169 template<int window_size>
170 const uint32_t* RollingHash<window_size>::remove_table_ = NULL;
171 
172 // Init() checks to make sure that the static object remove_table_ has been
173 // initialized; if not, it does the considerable work of populating it.  Once
174 // it's ready, the table can be used for any number of RollingHash objects of
175 // the same window_size.
176 //
177 template<int window_size>
Init()178 bool RollingHash<window_size>::Init() {
179   if (window_size < 2) {
180     LOG(ERROR) << "RollingHash window size " << window_size
181                << " is too small" << LOG_ENDL;
182     return false;
183   }
184   if (remove_table_ == NULL) {
185     // The new object is placed into a local pointer instead of directly into
186     // remove_table_, for two reasons:
187     //   1. remove_table_ is a pointer to const.  The table is populated using
188     //      the non-const local pointer and then assigned to the global const
189     //      pointer once it's ready.
190     //   2. No other thread will ever see remove_table_ pointing to a
191     //      partially-initialized table.  If two threads happen to call Init()
192     //      at the same time, two tables with the same contents may be created
193     //      (causing a memory leak), but the results will be consistent
194     //      no matter which of the two tables is used.
195     uint32_t* new_remove_table = new uint32_t[256];
196     // Compute multiplier.  Concisely, it is:
197     //     pow(kMult, (window_size - 1)) % kBase,
198     // but we compute the power in integer form.
199     uint32_t multiplier = 1;
200     for (int i = 0; i < window_size - 1; ++i) {
201       multiplier =
202           RollingHashUtil::ModBase(multiplier * RollingHashUtil::kMult);
203     }
204     // For each character removed_byte, compute
205     //     remove_table_[removed_byte] ==
206     //         (- (removed_byte * pow(kMult, (window_size - 1)))) % kBase
207     // where the power operator "pow" is taken in integer form.
208     //
209     // If you take a hash value fp representing the hash of
210     //     buffer[0] ... buffer[window_size - 1]
211     // and add the value of remove_table_[buffer[0]] to it, the result will be
212     // a partial hash value for
213     //     buffer[1] ... buffer[window_size - 1]
214     // that is to say, it no longer includes buffer[0].
215     //
216     // The following byte at buffer[window_size] can then be merged with this
217     // partial hash value to arrive quickly at the hash value for a window that
218     // has advanced by one byte, to
219     //     buffer[1] ... buffer[window_size]
220     // In fact, that is precisely what happens in UpdateHash, above.
221     uint32_t byte_times_multiplier = 0;
222     for (int removed_byte = 0; removed_byte < 256; ++removed_byte) {
223       new_remove_table[removed_byte] =
224           RollingHashUtil::FindModBaseInverse(byte_times_multiplier);
225       // Iteratively adding the multiplier in this loop is equivalent to
226       // computing (removed_byte * multiplier), and is faster
227       byte_times_multiplier =
228           RollingHashUtil::ModBase(byte_times_multiplier + multiplier);
229     }
230     remove_table_ = new_remove_table;
231   }
232   return true;
233 }
234 
235 }  // namespace open_vcdiff
236 
237 #endif  // OPEN_VCDIFF_ROLLING_HASH_H_
238