Commit | Line | Data |
---|---|---|
14bfc3f5 ILT |
1 | // stringpool.h -- a string pool for gold -*- C++ -*- |
2 | ||
6cb15b7f ILT |
3 | // Copyright 2006, 2007 Free Software Foundation, Inc. |
4 | // Written by Ian Lance Taylor <iant@google.com>. | |
5 | ||
6 | // This file is part of gold. | |
7 | ||
8 | // This program is free software; you can redistribute it and/or modify | |
9 | // it under the terms of the GNU General Public License as published by | |
10 | // the Free Software Foundation; either version 3 of the License, or | |
11 | // (at your option) any later version. | |
12 | ||
13 | // This program is distributed in the hope that it will be useful, | |
14 | // but WITHOUT ANY WARRANTY; without even the implied warranty of | |
15 | // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
16 | // GNU General Public License for more details. | |
17 | ||
18 | // You should have received a copy of the GNU General Public License | |
19 | // along with this program; if not, write to the Free Software | |
20 | // Foundation, Inc., 51 Franklin Street - Fifth Floor, Boston, | |
21 | // MA 02110-1301, USA. | |
22 | ||
14bfc3f5 ILT |
23 | #include <string> |
24 | #include <list> | |
25 | ||
14bfc3f5 ILT |
26 | #ifndef GOLD_STRINGPOOL_H |
27 | #define GOLD_STRINGPOOL_H | |
28 | ||
29 | namespace gold | |
30 | { | |
31 | ||
61ba1cf9 ILT |
32 | class Output_file; |
33 | ||
835965e6 ILT |
34 | // A Stringpool is a pool of unique strings. It provides the |
35 | // following features: | |
36 | ||
37 | // Every string in the pool is unique. Thus, if you have two strings | |
38 | // in the Stringpool, you can compare them for equality by using | |
39 | // pointer comparison rather than string comparison. | |
40 | ||
41 | // There is a key associated with every string in the pool. If you | |
42 | // add strings to the Stringpool in the same order, then the key for | |
43 | // each string will always be the same for any run of the linker. | |
44 | // This is not true of the string pointers themselves, as they may | |
45 | // change due to address space randomization. Some parts of the | |
46 | // linker (e.g., the symbol table) use the key value instead of the | |
47 | // string pointer so that repeated runs of the linker will generate | |
48 | // precisely the same output. | |
49 | ||
cfd73a4e ILT |
50 | // When you add a string to a Stringpool, Stringpool will optionally |
51 | // make a copy of it. Thus there is no requirement to keep a copy | |
52 | // elsewhere. | |
bc56b3fb | 53 | |
835965e6 ILT |
54 | // A Stringpool can be turned into a string table, a sequential series |
55 | // of null terminated strings. The first string may optionally be a | |
56 | // single zero byte, as required for SHT_STRTAB sections. This | |
57 | // conversion is only permitted after all strings have been added to | |
58 | // the Stringpool. After doing this conversion, you can ask for the | |
59 | // offset of any string in the stringpool in the string table, and you | |
60 | // can write the resulting string table to an output file. | |
61 | ||
62 | // When a Stringpool is turned into a string table, then as an | |
63 | // optimization it will reuse string suffixes to avoid duplicating | |
64 | // strings. That is, given the strings "abc" and "bc", only the | |
65 | // string "abc" will be stored, and "bc" will be represented by an | |
66 | // offset into the middle of the string "abc". | |
67 | ||
68 | // Stringpools are implemented in terms of Stringpool_template, which | |
69 | // is generalized on the type of character used for the strings. Most | |
70 | // uses will want the Stringpool type which uses char. Other cases | |
71 | // are used for merging wide string constants. | |
72 | ||
b8e6aad9 ILT |
73 | template<typename Stringpool_char> |
74 | class Stringpool_template | |
14bfc3f5 ILT |
75 | { |
76 | public: | |
835965e6 ILT |
77 | // The type of a key into the stringpool. As described above, a key |
78 | // value will always be the same during any run of the linker. Zero | |
79 | // is never a valid key value. | |
f0641a0b ILT |
80 | typedef size_t Key; |
81 | ||
a8b2552e ILT |
82 | // Create a Stringpool. |
83 | Stringpool_template(); | |
14bfc3f5 | 84 | |
b8e6aad9 | 85 | ~Stringpool_template(); |
14bfc3f5 | 86 | |
9a0910c3 ILT |
87 | // Clear all the data from the stringpool. |
88 | void | |
89 | clear(); | |
90 | ||
a8b2552e ILT |
91 | // Indicate that we should not reserve offset 0 to hold the empty |
92 | // string when converting the stringpool to a string table. This | |
93 | // should not be called for a proper ELF SHT_STRTAB section. | |
94 | void | |
95 | set_no_zero_null() | |
96 | { this->zero_null_ = false; } | |
97 | ||
835965e6 | 98 | // Add the string S to the pool. This returns a canonical permanent |
cfd73a4e ILT |
99 | // pointer to the string in the pool. If COPY is true, the string |
100 | // is copied into permanent storage. If PKEY is not NULL, this sets | |
101 | // *PKEY to the key for the string. | |
b8e6aad9 | 102 | const Stringpool_char* |
cfd73a4e | 103 | add(const Stringpool_char* s, bool copy, Key* pkey); |
14bfc3f5 | 104 | |
835965e6 | 105 | // Add the prefix of length LEN of string S to the pool. |
b8e6aad9 | 106 | const Stringpool_char* |
cfd73a4e | 107 | add_prefix(const Stringpool_char* s, size_t len, Key* pkey); |
61ba1cf9 | 108 | |
835965e6 ILT |
109 | // If the string S is present in the pool, return the canonical |
110 | // string pointer. Otherwise, return NULL. If PKEY is not NULL, | |
111 | // set *PKEY to the key. | |
b8e6aad9 | 112 | const Stringpool_char* |
835965e6 | 113 | find(const Stringpool_char* s, Key* pkey) const; |
61ba1cf9 | 114 | |
835965e6 ILT |
115 | // Turn the stringpool into a string table: determine the offsets of |
116 | // all the strings. After this is called, no more strings may be | |
117 | // added to the stringpool. | |
61ba1cf9 ILT |
118 | void |
119 | set_string_offsets(); | |
120 | ||
835965e6 ILT |
121 | // Get the offset of the string S in the string table. This returns |
122 | // the offset in bytes, not in units of Stringpool_char. This may | |
123 | // only be called after set_string_offsets has been called. | |
61ba1cf9 | 124 | off_t |
835965e6 | 125 | get_offset(const Stringpool_char* s) const; |
61ba1cf9 | 126 | |
835965e6 | 127 | // Get the offset of the string S in the string table. |
61ba1cf9 | 128 | off_t |
b8e6aad9 | 129 | get_offset(const std::basic_string<Stringpool_char>& s) const |
61ba1cf9 ILT |
130 | { return this->get_offset(s.c_str()); } |
131 | ||
835965e6 ILT |
132 | // Get the size of the string table. This returns the number of |
133 | // bytes, not in units of Stringpool_char. | |
61ba1cf9 ILT |
134 | off_t |
135 | get_strtab_size() const | |
a3ad94ed ILT |
136 | { |
137 | gold_assert(this->strtab_size_ != 0); | |
138 | return this->strtab_size_; | |
139 | } | |
61ba1cf9 | 140 | |
835965e6 ILT |
141 | // Write the string table into the output file at the specified |
142 | // offset. | |
61ba1cf9 ILT |
143 | void |
144 | write(Output_file*, off_t offset); | |
14bfc3f5 | 145 | |
9a0910c3 ILT |
146 | // Write the string table into the specified buffer, of the |
147 | // specified size. buffer_size should be at least | |
148 | // get_strtab_size(). | |
149 | void | |
96803768 | 150 | write_to_buffer(unsigned char* buffer, size_t buffer_size); |
9a0910c3 | 151 | |
ad8f37d1 ILT |
152 | // Dump statistical information to stderr. |
153 | void | |
154 | print_stats(const char*) const; | |
155 | ||
14bfc3f5 | 156 | private: |
b8e6aad9 ILT |
157 | Stringpool_template(const Stringpool_template&); |
158 | Stringpool_template& operator=(const Stringpool_template&); | |
159 | ||
835965e6 | 160 | // Return the length of a string in units of Stringpool_char. |
b8e6aad9 ILT |
161 | static size_t |
162 | string_length(const Stringpool_char*); | |
14bfc3f5 | 163 | |
987cc251 ILT |
164 | // Return whether two strings are equal. |
165 | static bool | |
166 | string_equal(const Stringpool_char*, const Stringpool_char*); | |
167 | ||
168 | // Compute a hash code for a string. LENGTH is the length of the | |
169 | // string in characters. | |
170 | static size_t | |
171 | string_hash(const Stringpool_char*, size_t length); | |
172 | ||
61ba1cf9 ILT |
173 | // We store the actual data in a list of these buffers. |
174 | struct Stringdata | |
14bfc3f5 ILT |
175 | { |
176 | // Length of data in buffer. | |
177 | size_t len; | |
178 | // Allocated size of buffer. | |
179 | size_t alc; | |
f0641a0b ILT |
180 | // Buffer index. |
181 | unsigned int index; | |
14bfc3f5 ILT |
182 | // Buffer. |
183 | char data[1]; | |
184 | }; | |
185 | ||
61ba1cf9 | 186 | // Copy a string into the buffers, returning a canonical string. |
b8e6aad9 | 187 | const Stringpool_char* |
987cc251 ILT |
188 | add_string(const Stringpool_char*, size_t, Key*); |
189 | ||
190 | // Return whether s1 is a suffix of s2. | |
191 | static bool | |
192 | is_suffix(const Stringpool_char* s1, size_t len1, | |
193 | const Stringpool_char* s2, size_t len2); | |
194 | ||
195 | // The hash table key includes the string, the length of the string, | |
196 | // and the hash code for the string. We put the hash code | |
197 | // explicitly into the key so that we can do a find()/insert() | |
198 | // sequence without having to recompute the hash. Computing the | |
199 | // hash code is a significant user of CPU time in the linker. | |
200 | struct Hashkey | |
201 | { | |
202 | const Stringpool_char* string; | |
203 | // Length is in characters, not bytes. | |
204 | size_t length; | |
205 | size_t hash_code; | |
206 | ||
207 | // This goes in an STL container, so we need a default | |
208 | // constructor. | |
209 | Hashkey() | |
210 | : string(NULL), length(0), hash_code(0) | |
211 | { } | |
212 | ||
213 | // Note that these constructors are relatively expensive, because | |
214 | // they compute the hash code. | |
215 | Hashkey(const Stringpool_char* s) | |
216 | : string(s), length(string_length(s)), hash_code(string_hash(s, length)) | |
217 | { } | |
218 | ||
219 | Hashkey(const Stringpool_char* s, size_t len) | |
220 | : string(s), length(len), hash_code(string_hash(s, len)) | |
221 | { } | |
222 | }; | |
14bfc3f5 | 223 | |
987cc251 ILT |
224 | // Hash function. This is trivial, since we have already computed |
225 | // the hash. | |
14bfc3f5 ILT |
226 | struct Stringpool_hash |
227 | { | |
228 | size_t | |
987cc251 ILT |
229 | operator()(const Hashkey& hk) const |
230 | { return hk.hash_code; } | |
14bfc3f5 ILT |
231 | }; |
232 | ||
835965e6 | 233 | // Equality comparison function for hash table. |
14bfc3f5 ILT |
234 | struct Stringpool_eq |
235 | { | |
236 | bool | |
987cc251 | 237 | operator()(const Hashkey&, const Hashkey&) const; |
14bfc3f5 ILT |
238 | }; |
239 | ||
987cc251 ILT |
240 | // The hash table is a map from strings to a pair of Key and string |
241 | // table offsets. We only use the offsets if we turn this into an | |
242 | // string table section. | |
f0641a0b | 243 | |
987cc251 | 244 | typedef std::pair<Key, off_t> Hashval; |
61ba1cf9 | 245 | |
987cc251 | 246 | typedef Unordered_map<Hashkey, Hashval, Stringpool_hash, |
61ba1cf9 | 247 | Stringpool_eq> String_set_type; |
274e99f9 | 248 | |
987cc251 ILT |
249 | // Comparison routine used when sorting into a string table. |
250 | ||
251 | typedef typename String_set_type::iterator Stringpool_sort_info; | |
61ba1cf9 ILT |
252 | |
253 | struct Stringpool_sort_comparison | |
254 | { | |
255 | bool | |
614f30a2 | 256 | operator()(const Stringpool_sort_info&, const Stringpool_sort_info&) const; |
61ba1cf9 ILT |
257 | }; |
258 | ||
f0641a0b ILT |
259 | // List of Stringdata structures. |
260 | typedef std::list<Stringdata*> Stringdata_list; | |
261 | ||
262 | // Mapping from const char* to namepool entry. | |
14bfc3f5 | 263 | String_set_type string_set_; |
f0641a0b ILT |
264 | // List of buffers. |
265 | Stringdata_list strings_; | |
835965e6 | 266 | // Size of string table. |
61ba1cf9 | 267 | off_t strtab_size_; |
f0641a0b ILT |
268 | // Next Stringdata index. |
269 | unsigned int next_index_; | |
cfd73a4e ILT |
270 | // Next key value for a string we don't copy. |
271 | int next_uncopied_key_; | |
b8e6aad9 ILT |
272 | // Whether to reserve offset 0 to hold the null string. |
273 | bool zero_null_; | |
14bfc3f5 ILT |
274 | }; |
275 | ||
b8e6aad9 ILT |
276 | // The most common type of Stringpool. |
277 | typedef Stringpool_template<char> Stringpool; | |
278 | ||
14bfc3f5 ILT |
279 | } // End namespace gold. |
280 | ||
281 | #endif // !defined(GOLD_STRINGPOOL_H) |