1 // stringpool.cc -- a string pool for gold
3 // Copyright 2006, 2007, 2008 Free Software Foundation, Inc.
4 // Written by Ian Lance Taylor <iant@google.com>.
6 // This file is part of gold.
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.
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.
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.
30 #include "parameters.h"
31 #include "stringpool.h"
36 template<typename Stringpool_char
>
37 Stringpool_template
<Stringpool_char
>::Stringpool_template()
38 : string_set_(), key_to_offset_(), strings_(), strtab_size_(0),
39 zero_null_(true), optimize_(false), offset_(0)
41 if (parameters
->options_valid() && parameters
->options().optimize() >= 2)
42 this->optimize_
= true;
45 template<typename Stringpool_char
>
47 Stringpool_template
<Stringpool_char
>::clear()
49 for (typename
std::list
<Stringdata
*>::iterator p
= this->strings_
.begin();
50 p
!= this->strings_
.end();
52 delete[] reinterpret_cast<char*>(*p
);
53 this->strings_
.clear();
54 this->key_to_offset_
.clear();
55 this->string_set_
.clear();
58 template<typename Stringpool_char
>
59 Stringpool_template
<Stringpool_char
>::~Stringpool_template()
64 // Resize the internal hashtable with the expectation we'll get n new
65 // elements. Note that the hashtable constructor takes a "number of
66 // buckets you'd like," rather than "number of elements you'd like,"
67 // but that's the best we can do.
69 template<typename Stringpool_char
>
71 Stringpool_template
<Stringpool_char
>::reserve(unsigned int n
)
73 this->key_to_offset_
.reserve(n
);
75 #if defined(HAVE_TR1_UNORDERED_MAP)
76 // rehash() implementation is broken in gcc 4.0.3's stl
77 //this->string_set_.rehash(this->string_set_.size() + n);
79 #elif defined(HAVE_EXT_HASH_MAP)
80 this->string_set_
.resize(this->string_set_
.size() + n
);
84 // This is the generic "reserve" code, if no #ifdef above triggers.
85 String_set_type
new_string_set(this->string_set_
.size() + n
);
86 new_string_set
.insert(this->string_set_
.begin(), this->string_set_
.end());
87 this->string_set_
.swap(new_string_set
);
90 // Return the length of a string of arbitrary character type.
92 template<typename Stringpool_char
>
94 Stringpool_template
<Stringpool_char
>::string_length(const Stringpool_char
* p
)
102 // Specialize string_length for char. Maybe we could just use
103 // std::char_traits<>::length?
107 Stringpool_template
<char>::string_length(const char* p
)
112 // Compare two strings of arbitrary character type for equality.
114 template<typename Stringpool_char
>
116 Stringpool_template
<Stringpool_char
>::string_equal(const Stringpool_char
* s1
,
117 const Stringpool_char
* s2
)
125 // Specialize string_equal for char.
129 Stringpool_template
<char>::string_equal(const char* s1
, const char* s2
)
131 return strcmp(s1
, s2
) == 0;
134 // Equality comparison function for the hash table.
136 template<typename Stringpool_char
>
138 Stringpool_template
<Stringpool_char
>::Stringpool_eq::operator()(
140 const Hashkey
& h2
) const
142 return (h1
.hash_code
== h2
.hash_code
143 && h1
.length
== h2
.length
144 && (h1
.string
== h2
.string
145 || memcmp(h1
.string
, h2
.string
,
146 h1
.length
* sizeof(Stringpool_char
)) == 0));
149 // Hash function. The length is in characters, not bytes.
151 template<typename Stringpool_char
>
153 Stringpool_template
<Stringpool_char
>::string_hash(const Stringpool_char
* s
,
156 return gold::string_hash
<Stringpool_char
>(s
, length
);
159 // Add the string S to the list of canonical strings. Return a
160 // pointer to the canonical string. If PKEY is not NULL, set *PKEY to
161 // the key. LENGTH is the length of S in characters. Note that S may
162 // not be NUL terminated.
164 template<typename Stringpool_char
>
165 const Stringpool_char
*
166 Stringpool_template
<Stringpool_char
>::add_string(const Stringpool_char
* s
,
169 // We are in trouble if we've already computed the string offsets.
170 gold_assert(this->strtab_size_
== 0);
172 // The size we allocate for a new Stringdata.
173 const size_t buffer_size
= 1000;
174 // The amount we multiply the Stringdata index when calculating the
176 const size_t key_mult
= 1024;
177 gold_assert(key_mult
>= buffer_size
);
179 // Convert len to the number of bytes we need to allocate, including
180 // the null character.
181 len
= (len
+ 1) * sizeof(Stringpool_char
);
185 if (len
> buffer_size
)
187 alc
= sizeof(Stringdata
) + len
;
190 else if (this->strings_
.empty())
191 alc
= sizeof(Stringdata
) + buffer_size
;
194 Stringdata
*psd
= this->strings_
.front();
195 if (len
> psd
->alc
- psd
->len
)
196 alc
= sizeof(Stringdata
) + buffer_size
;
199 char* ret
= psd
->data
+ psd
->len
;
200 memcpy(ret
, s
, len
- sizeof(Stringpool_char
));
201 memset(ret
+ len
- sizeof(Stringpool_char
), 0,
202 sizeof(Stringpool_char
));
206 return reinterpret_cast<const Stringpool_char
*>(ret
);
210 Stringdata
*psd
= reinterpret_cast<Stringdata
*>(new char[alc
]);
211 psd
->alc
= alc
- sizeof(Stringdata
);
212 memcpy(psd
->data
, s
, len
- sizeof(Stringpool_char
));
213 memset(psd
->data
+ len
- sizeof(Stringpool_char
), 0,
214 sizeof(Stringpool_char
));
218 this->strings_
.push_front(psd
);
220 this->strings_
.push_back(psd
);
222 return reinterpret_cast<const Stringpool_char
*>(psd
->data
);
225 // Add a string to a string pool.
227 template<typename Stringpool_char
>
228 const Stringpool_char
*
229 Stringpool_template
<Stringpool_char
>::add(const Stringpool_char
* s
, bool copy
,
232 return this->add_with_length(s
, string_length(s
), copy
, pkey
);
235 // Add a new key offset entry.
237 template<typename Stringpool_char
>
239 Stringpool_template
<Stringpool_char
>::new_key_offset(size_t length
)
241 if (this->key_to_offset_
.size() == 0)
242 this->offset_
= this->zero_null_
? sizeof(Stringpool_char
) : 0;
244 section_offset_type offset
;
245 if (this->zero_null_
&& length
== 0)
249 offset
= this->offset_
;
250 this->offset_
+= (length
+ 1) * sizeof(Stringpool_char
);
252 this->key_to_offset_
.push_back(offset
);
255 template<typename Stringpool_char
>
256 const Stringpool_char
*
257 Stringpool_template
<Stringpool_char
>::add_with_length(const Stringpool_char
* s
,
262 typedef std::pair
<typename
String_set_type::iterator
, bool> Insert_type
;
264 // We add 1 so that 0 is always invalid.
265 const Key k
= this->key_to_offset_
.size() + 1;
269 // When we don't need to copy the string, we can call insert
272 std::pair
<Hashkey
, Hashval
> element(Hashkey(s
, length
), k
);
274 Insert_type ins
= this->string_set_
.insert(element
);
276 typename
String_set_type::const_iterator p
= ins
.first
;
280 // We just added the string. The key value has now been
282 this->new_key_offset(length
);
286 gold_assert(k
!= p
->second
);
291 return p
->first
.string
;
294 // When we have to copy the string, we look it up twice in the hash
295 // table. The problem is that we can't insert S before we
296 // canonicalize it by copying it into the canonical list. The hash
297 // code will only be computed once.
299 Hashkey
hk(s
, length
);
300 typename
String_set_type::const_iterator p
= this->string_set_
.find(hk
);
301 if (p
!= this->string_set_
.end())
305 return p
->first
.string
;
308 this->new_key_offset(length
);
310 hk
.string
= this->add_string(s
, length
);
311 // The contents of the string stay the same, so we don't need to
312 // adjust hk.hash_code or hk.length.
314 std::pair
<Hashkey
, Hashval
> element(hk
, k
);
316 Insert_type ins
= this->string_set_
.insert(element
);
317 gold_assert(ins
.second
);
324 template<typename Stringpool_char
>
325 const Stringpool_char
*
326 Stringpool_template
<Stringpool_char
>::find(const Stringpool_char
* s
,
330 typename
String_set_type::const_iterator p
= this->string_set_
.find(hk
);
331 if (p
== this->string_set_
.end())
337 return p
->first
.string
;
340 // Comparison routine used when sorting into an ELF strtab. We want
341 // to sort this so that when one string is a suffix of another, we
342 // always see the shorter string immediately after the longer string.
343 // For example, we want to see these strings in this order:
347 // When strings are not suffixes, we don't care what order they are
348 // in, but we need to ensure that suffixes wind up next to each other.
349 // So we do a reversed lexicographic sort on the reversed string.
351 template<typename Stringpool_char
>
353 Stringpool_template
<Stringpool_char
>::Stringpool_sort_comparison::operator()(
354 const Stringpool_sort_info
& sort_info1
,
355 const Stringpool_sort_info
& sort_info2
) const
357 const Hashkey
& h1(sort_info1
->first
);
358 const Hashkey
& h2(sort_info2
->first
);
359 const Stringpool_char
* s1
= h1
.string
;
360 const Stringpool_char
* s2
= h2
.string
;
361 const size_t len1
= h1
.length
;
362 const size_t len2
= h2
.length
;
363 const size_t minlen
= len1
< len2
? len1
: len2
;
364 const Stringpool_char
* p1
= s1
+ len1
- 1;
365 const Stringpool_char
* p2
= s2
+ len2
- 1;
366 for (size_t i
= minlen
; i
> 0; --i
, --p1
, --p2
)
374 // Return whether s1 is a suffix of s2.
376 template<typename Stringpool_char
>
378 Stringpool_template
<Stringpool_char
>::is_suffix(const Stringpool_char
* s1
,
380 const Stringpool_char
* s2
,
385 return memcmp(s1
, s2
+ len2
- len1
, len1
* sizeof(Stringpool_char
)) == 0;
388 // Turn the stringpool into an ELF strtab: determine the offsets of
389 // each string in the table.
391 template<typename Stringpool_char
>
393 Stringpool_template
<Stringpool_char
>::set_string_offsets()
395 if (this->strtab_size_
!= 0)
397 // We've already computed the offsets.
401 const size_t charsize
= sizeof(Stringpool_char
);
403 // Offset 0 may be reserved for the empty string.
404 section_offset_type offset
= this->zero_null_
? charsize
: 0;
406 // Sorting to find suffixes can take over 25% of the total CPU time
407 // used by the linker. Since it's merely an optimization to reduce
408 // the strtab size, and gives a relatively small benefit (it's
409 // typically rare for a symbol to be a suffix of another), we only
410 // take the time to sort when the user asks for heavy optimization.
411 if (!this->optimize_
)
413 // If we are not optimizing, the offsets are already assigned.
414 offset
= this->offset_
;
418 size_t count
= this->string_set_
.size();
420 std::vector
<Stringpool_sort_info
> v
;
423 for (typename
String_set_type::iterator p
= this->string_set_
.begin();
424 p
!= this->string_set_
.end();
426 v
.push_back(Stringpool_sort_info(p
));
428 std::sort(v
.begin(), v
.end(), Stringpool_sort_comparison());
430 section_offset_type last_offset
= -1;
431 for (typename
std::vector
<Stringpool_sort_info
>::iterator last
= v
.end(),
436 section_offset_type this_offset
;
437 if (this->zero_null_
&& (*curr
)->first
.string
[0] == 0)
439 else if (last
!= v
.end()
440 && is_suffix((*curr
)->first
.string
,
441 (*curr
)->first
.length
,
442 (*last
)->first
.string
,
443 (*last
)->first
.length
))
444 this_offset
= (last_offset
445 + (((*last
)->first
.length
- (*curr
)->first
.length
)
449 this_offset
= offset
;
450 offset
+= ((*curr
)->first
.length
+ 1) * charsize
;
452 this->key_to_offset_
[(*curr
)->second
- 1] = this_offset
;
453 last_offset
= this_offset
;
457 this->strtab_size_
= offset
;
460 // Get the offset of a string in the ELF strtab. The string must
463 template<typename Stringpool_char
>
465 Stringpool_template
<Stringpool_char
>::get_offset(const Stringpool_char
* s
)
468 return this->get_offset_with_length(s
, string_length(s
));
471 template<typename Stringpool_char
>
473 Stringpool_template
<Stringpool_char
>::get_offset_with_length(
474 const Stringpool_char
* s
,
477 gold_assert(this->strtab_size_
!= 0);
478 Hashkey
hk(s
, length
);
479 typename
String_set_type::const_iterator p
= this->string_set_
.find(hk
);
480 if (p
!= this->string_set_
.end())
481 return this->key_to_offset_
[p
->second
- 1];
485 // Write the ELF strtab into the buffer.
487 template<typename Stringpool_char
>
489 Stringpool_template
<Stringpool_char
>::write_to_buffer(
490 unsigned char* buffer
,
491 section_size_type bufsize
)
493 gold_assert(this->strtab_size_
!= 0);
494 gold_assert(bufsize
>= this->strtab_size_
);
495 if (this->zero_null_
)
497 for (typename
String_set_type::const_iterator p
= this->string_set_
.begin();
498 p
!= this->string_set_
.end();
501 const int len
= (p
->first
.length
+ 1) * sizeof(Stringpool_char
);
502 const section_offset_type offset
= this->key_to_offset_
[p
->second
- 1];
503 gold_assert(static_cast<section_size_type
>(offset
) + len
504 <= this->strtab_size_
);
505 memcpy(buffer
+ offset
, p
->first
.string
, len
);
509 // Write the ELF strtab into the output file at the specified offset.
511 template<typename Stringpool_char
>
513 Stringpool_template
<Stringpool_char
>::write(Output_file
* of
, off_t offset
)
515 gold_assert(this->strtab_size_
!= 0);
516 unsigned char* view
= of
->get_output_view(offset
, this->strtab_size_
);
517 this->write_to_buffer(view
, this->strtab_size_
);
518 of
->write_output_view(offset
, this->strtab_size_
, view
);
521 // Print statistical information to stderr. This is used for --stats.
523 template<typename Stringpool_char
>
525 Stringpool_template
<Stringpool_char
>::print_stats(const char* name
) const
527 #if defined(HAVE_TR1_UNORDERED_MAP) || defined(HAVE_EXT_HASH_MAP)
528 fprintf(stderr
, _("%s: %s entries: %zu; buckets: %zu\n"),
529 program_name
, name
, this->string_set_
.size(),
530 this->string_set_
.bucket_count());
532 fprintf(stderr
, _("%s: %s entries: %zu\n"),
533 program_name
, name
, this->table_
.size());
535 fprintf(stderr
, _("%s: %s Stringdata structures: %zu\n"),
536 program_name
, name
, this->strings_
.size());
539 // Instantiate the templates we need.
542 class Stringpool_template
<char>;
545 class Stringpool_template
<uint16_t>;
548 class Stringpool_template
<uint32_t>;
550 } // End namespace gold.