1 /*
2 * Copyright (c) 1997, 2026, Oracle and/or its affiliates. All rights reserved.
3 * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER.
4 *
5 * This code is free software; you can redistribute it and/or modify it
6 * under the terms of the GNU General Public License version 2 only, as
7 * published by the Free Software Foundation.
8 *
9 * This code is distributed in the hope that it will be useful, but WITHOUT
10 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
11 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
12 * version 2 for more details (a copy is included in the LICENSE file that
13 * accompanied this code).
14 *
15 * You should have received a copy of the GNU General Public License version
16 * 2 along with this work; if not, write to the Free Software Foundation,
17 * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
18 *
19 * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA
20 * or visit www.oracle.com if you need additional information or have any
21 * questions.
22 *
23 */
24
25 #ifndef SHARE_CLASSFILE_COMPACTHASHTABLE_HPP
26 #define SHARE_CLASSFILE_COMPACTHASHTABLE_HPP
27
28 #include "cds/aotCompressedPointers.hpp"
29 #include "cds/cds_globals.hpp"
30 #include "oops/array.hpp"
31 #include "oops/symbol.hpp"
32 #include "runtime/globals.hpp"
33 #include "utilities/growableArray.hpp"
34
35
36 template <
37 typename K,
38 typename V,
39 V (*DECODE)(address base_address, u4 encoded_value),
40 bool (*EQUALS)(V value, K key, int len)
41 >
42 class CompactHashtable;
43 class NumberSeq;
44 class SimpleCompactHashtable;
45 class SerializeClosure;
46
47 // Stats for symbol tables in the CDS archive
48 class CompactHashtableStats {
49 public:
50 int hashentry_count;
51 int hashentry_bytes;
52 int bucket_count;
53 int bucket_bytes;
54
55 CompactHashtableStats() :
56 hashentry_count(0), hashentry_bytes(0),
57 bucket_count(0), bucket_bytes(0) {}
58 };
59
60 #if INCLUDE_CDS
61 /////////////////////////////////////////////////////////////////////////
62 //
63 // The compact hash table writer. Used at dump time for writing out
64 // the compact table to the shared archive.
65 //
66 // At dump time, the CompactHashtableWriter obtains all entries from
67 // a table (the table could be in any form of a collection of <hash, encoded_value> pair)
68 // and adds them to a new temporary hash table (_buckets). The hash
69 // table size (number of buckets) is calculated using
70 // '(num_entries + bucket_size - 1) / bucket_size'. The default bucket
71 // size is 4 and can be changed by -XX:SharedSymbolTableBucketSize option.
72 // 4 is chosen because it produces smaller sized bucket on average for
73 // faster lookup. It also has relatively small number of empty buckets and
74 // good distribution of the entries.
75 //
76 // We use a simple hash function (hash % num_bucket) for the table.
77 // The new table is compacted when written out. Please see comments
78 // above the CompactHashtable class for the table layout detail. The bucket
79 // offsets are written to the archive as part of the compact table. The
80 // bucket offset is encoded in the low 30-bit (0-29) and the bucket type
81 // (regular or value_only) are encoded in bit[31, 30]. For buckets with more
82 // than one entry, both hash and encoded_value are written to the
83 // table. For buckets with only one entry, only the encoded_value is written
84 // to the table and the buckets are tagged as value_only in their type bits.
85 // Buckets without entry are skipped from the table. Their offsets are
86 // still written out for faster lookup.
87 //
88 class CompactHashtableWriter: public StackObj {
89 public:
90 class Entry {
91 unsigned int _hash;
92 u4 _encoded_value;
93
94 public:
95 Entry() {}
96 Entry(unsigned int hash, u4 encoded_value) : _hash(hash), _encoded_value(encoded_value) {}
97
98 u4 encoded_value() {
99 return _encoded_value;
100 }
101 unsigned int hash() {
102 return _hash;
103 }
104
105 bool operator==(const CompactHashtableWriter::Entry& other) {
106 return (_encoded_value == other._encoded_value && _hash == other._hash);
107 }
108 }; // class CompactHashtableWriter::Entry
109
110 private:
111 int _num_entries_written;
112 int _num_buckets;
113 int _num_empty_buckets;
114 int _num_value_only_buckets;
115 int _num_other_buckets;
116 GrowableArray<Entry>** _buckets;
117 CompactHashtableStats* _stats;
118 Array<u4>* _compact_buckets;
119 Array<u4>* _compact_entries;
120
121 public:
122 // This is called at dump-time only
123 CompactHashtableWriter(int num_entries, CompactHashtableStats* stats);
124 ~CompactHashtableWriter();
125
126 void add(unsigned int hash, u4 encoded_value);
127 void add(unsigned int hash, AOTCompressedPointers::narrowPtr encoded_value) {
128 add(hash, cast_to_u4(encoded_value));
129 }
130 void dump(SimpleCompactHashtable *cht, const char* table_name);
131
132 private:
133 void allocate_table();
134 void dump_table(NumberSeq* summary);
135 static int calculate_num_buckets(int num_entries) {
136 int num_buckets = num_entries / SharedSymbolTableBucketSize;
137 // calculation of num_buckets can result in zero buckets, we need at least one
138 return (num_buckets < 1) ? 1 : num_buckets;
139 }
140 };
141 #endif // INCLUDE_CDS
142
143 #define REGULAR_BUCKET_TYPE 0
144 #define VALUE_ONLY_BUCKET_TYPE 1
145 #define TABLEEND_BUCKET_TYPE 3
146 #define BUCKET_OFFSET_MASK 0x3FFFFFFF
147 #define BUCKET_OFFSET(info) ((info) & BUCKET_OFFSET_MASK)
148 #define BUCKET_TYPE_SHIFT 30
149 #define BUCKET_TYPE(info) (((info) & ~BUCKET_OFFSET_MASK) >> BUCKET_TYPE_SHIFT)
150 #define BUCKET_INFO(offset, type) (((type) << BUCKET_TYPE_SHIFT) | ((offset) & BUCKET_OFFSET_MASK))
151
152 /////////////////////////////////////////////////////////////////////////////
153 //
154 // CompactHashtable is used to store the CDS archive's tables.
155 // A table could be in any form of a collection of <hash, encoded_value> pair.
156 //
157 // Because these tables are read-only (no entries can be added/deleted) at run-time
158 // and tend to have large number of entries, we try to minimize the footprint
159 // cost per entry.
160 //
161 // The CompactHashtable is split into two arrays
162 //
163 // u4 buckets[num_buckets+1]; // bit[31,30]: type; bit[29-0]: offset
164 // u4 entries[<variable size>]
165 //
166 // The size of buckets[] is 'num_buckets + 1'. Each entry of
167 // buckets[] is a 32-bit encoding of the bucket type and bucket offset,
168 // with the type in the left-most 2-bit and offset in the remaining 30-bit.
169 //
170 // There are three types of buckets: regular, value_only, and table_end.
171 // . The regular buckets have '00' in their highest 2-bit.
172 // . The value_only buckets have '01' in their highest 2-bit.
173 // . There is only a single table_end bucket that marks the end of buckets[].
174 // It has '11' in its highest 2-bit.
175 //
176 // For regular buckets, each entry is 8 bytes in the entries[]:
177 // u4 hash; // entry hash
178 // u4 encoded_value; // A 32-bit encoding of the template type V. The template parameter DECODE
179 // // converts this to type V. Many CompactHashtables encode a pointer as a 32-bit offset, where
180 // // V entry = (V)(base_address + offset)
181 // // see StringTable, SymbolTable and AdapterHandlerLibrary for examples
182 //
183 // For value_only buckets, each entry has only the 4-byte 'encoded_value' in the entries[].
184 //
185 // The single table_end bucket has no corresponding entry.
186 //
187 // The number of entries in bucket <i> can be calculated like this:
188 // my_offset = _buckets[i] & 0x3fffffff; // mask off top 2-bit
189 // next_offset = _buckets[i+1] & 0x3fffffff
190 // For REGULAR_BUCKET_TYPE
191 // num_entries = (next_offset - my_offset) / 8;
192 // For VALUE_ONLY_BUCKET_TYPE
193 // num_entries = (next_offset - my_offset) / 4;
194 //
195 // If bucket <i> is empty, we have my_offset == next_offset. Empty buckets are
196 // always encoded as regular buckets.
197 //
198 // In the following example:
199 // - Bucket #0 is a REGULAR_BUCKET_TYPE with two entries
200 // - Bucket #1 is a VALUE_ONLY_BUCKET_TYPE with one entry.
201 // - Bucket #2 is a REGULAR_BUCKET_TYPE with zero entries.
202 //
203 // buckets[0, 4, 5(empty), 5, ...., N(table_end)]
204 // | | | | |
205 // | | +---+-----+ |
206 // | | | |
207 // | +----+ + |
208 // v v v v
209 // entries[H,O,H,O,O,H,O,H,O........]
210 //
211 // See CompactHashtable::lookup() for how the table is searched at runtime.
212 // See CompactHashtableWriter::dump() for how the table is written at CDS
213 // dump time.
214 //
215 class SimpleCompactHashtable {
216 protected:
217 address _base_address;
218 u4 _bucket_count;
219 u4 _entry_count;
220 u4* _buckets;
221 u4* _entries;
222
223 public:
224 SimpleCompactHashtable() :
225 _base_address(nullptr),
226 _bucket_count(0),
227 _entry_count(0),
228 _buckets(nullptr),
229 _entries(nullptr)
230 {}
231
232 void reset() {
233 _base_address = nullptr;
234 _bucket_count = 0;
235 _entry_count = 0;
236 _buckets = nullptr;
237 _entries = nullptr;
238 }
239
240 void init(address base_address, u4 entry_count, u4 bucket_count, u4* buckets, u4* entries);
241
242 // Read/Write the table's header from/to the CDS archive
243 void serialize_header(SerializeClosure* soc) NOT_CDS_RETURN;
244
245 inline bool empty() const {
246 return (_entry_count == 0);
247 }
248
249 inline size_t entry_count() const {
250 return _entry_count;
251 }
252 };
253
254 template <
255 typename K,
256 typename V,
257 V (*DECODE)(address base_address, u4 encoded_value),
258 bool (*EQUALS)(V value, K key, int len)
259 >
260 class CompactHashtable : public SimpleCompactHashtable {
261
262 V decode(u4 encoded_value) const {
263 return DECODE(_base_address, encoded_value);
264 }
265
266 public:
267 // Lookup a value V from the compact table using key K
268 inline V lookup(K key, unsigned int hash, int len) const {
269 if (_entry_count > 0) {
270 int index = hash % _bucket_count;
271 u4 bucket_info = _buckets[index];
272 u4 bucket_offset = BUCKET_OFFSET(bucket_info);
273 int bucket_type = BUCKET_TYPE(bucket_info);
274 u4* entry = _entries + bucket_offset;
275
276 if (bucket_type == VALUE_ONLY_BUCKET_TYPE) {
277 V value = decode(entry[0]);
278 if (EQUALS(value, key, len)) {
279 return value;
280 }
281 } else {
282 // This is a regular bucket, which has more than one
283 // entries. Each entry is a (hash, value) pair.
284 // Seek until the end of the bucket.
285 u4* entry_max = _entries + BUCKET_OFFSET(_buckets[index + 1]);
286 while (entry < entry_max) {
287 unsigned int h = (unsigned int)(entry[0]);
288 if (h == hash) {
289 V value = decode(entry[1]);
290 if (EQUALS(value, key, len)) {
291 return value;
292 }
293 }
294 entry += 2;
295 }
296 }
297 }
298 return nullptr;
299 }
300
301 // Iterate through the values in the table, stopping when do_value() returns false.
302 template <class ITER>
303 inline void iterate(ITER* iter) const { iterate([&](V v) { iter->do_value(v); }); }
304
305 template<typename Function>
306 inline void iterate(const Function& function) const { // lambda enabled API
307 iterate(const_cast<Function&>(function));
308 }
309
310 // Iterate through the values in the table, stopping when the lambda returns false.
311 template<typename Function>
312 inline void iterate(Function& function) const { // lambda enabled API
313 for (u4 i = 0; i < _bucket_count; i++) {
314 u4 bucket_info = _buckets[i];
315 u4 bucket_offset = BUCKET_OFFSET(bucket_info);
316 int bucket_type = BUCKET_TYPE(bucket_info);
317 u4* entry = _entries + bucket_offset;
318
319 if (bucket_type == VALUE_ONLY_BUCKET_TYPE) {
320 if (!function(decode(entry[0]))) {
321 return;
322 }
323 } else {
324 u4* entry_max = _entries + BUCKET_OFFSET(_buckets[i + 1]);
325 while (entry < entry_max) {
326 if (!function(decode(entry[1]))) {
327 return;
328 }
329 entry += 2;
330 }
331 }
332 }
333 }
334
335 // Unconditionally iterate through all the values in the table
336 template <class ITER>
337 inline void iterate_all(ITER* iter) const { iterate_all([&](V v) { iter->do_value(v); }); }
338
339 // Unconditionally iterate through all the values in the table using lambda
340 template<typename Function>
341 void iterate_all(Function function) const { // lambda enabled API
342 auto wrapper = [&] (V v) {
343 function(v);
344 return true;
345 };
346 iterate(wrapper);
347 }
348
349 void print_table_statistics(outputStream* st, const char* name) {
350 st->print_cr("%s statistics:", name);
351 int total_entries = 0;
352 int max_bucket = 0;
353 for (u4 i = 0; i < _bucket_count; i++) {
354 u4 bucket_info = _buckets[i];
355 int bucket_type = BUCKET_TYPE(bucket_info);
356 int bucket_size;
357
358 if (bucket_type == VALUE_ONLY_BUCKET_TYPE) {
359 bucket_size = 1;
360 } else {
361 bucket_size = (BUCKET_OFFSET(_buckets[i + 1]) - BUCKET_OFFSET(bucket_info)) / 2;
362 }
363 total_entries += bucket_size;
364 if (max_bucket < bucket_size) {
365 max_bucket = bucket_size;
366 }
367 }
368 st->print_cr("Number of buckets : %9d", _bucket_count);
369 st->print_cr("Number of entries : %9d", total_entries);
370 st->print_cr("Maximum bucket size : %9d", max_bucket);
371 }
372 };
373
374 ////////////////////////////////////////////////////////////////////////
375 //
376 // OffsetCompactHashtable -- This is used to store many types of objects
377 // in the CDS archive. On 64-bit platforms, we save space by using a 32-bit
378 // narrowPtr from the CDS base address.
379
380 template <typename V>
381 inline V read_value_from_compact_hashtable(address base_address, u4 narrowp) {
382 return AOTCompressedPointers::decode_not_null<V>(cast_from_u4(narrowp), base_address);
383 }
384
385 template <
386 typename K,
387 typename V,
388 bool (*EQUALS)(V value, K key, int len)
389 >
390 class OffsetCompactHashtable : public CompactHashtable<
391 K, V, read_value_from_compact_hashtable<V>, EQUALS> {
392 };
393
394
395 ////////////////////////////////////////////////////////////////////////
396 //
397 // Read/Write the contents of a hashtable textual dump (created by
398 // SymbolTable::dump and StringTable::dump).
399 // Because the dump file may be big (hundred of MB in extreme cases),
400 // we use mmap for fast access when reading it.
401 //
402 class HashtableTextDump {
403 int _fd;
404 const char* _base;
405 const char* _p;
406 const char* _end;
407 const char* _filename;
408 size_t _size;
409 int _prefix_type;
410 int _line_no;
411 public:
412 HashtableTextDump(const char* filename);
413 ~HashtableTextDump();
414
415 enum {
416 SymbolPrefix = 1 << 0,
417 StringPrefix = 1 << 1,
418 Unknown = 1 << 2
419 };
420
421 void quit(const char* err, const char* msg);
422
423 inline int remain() {
424 return (int)(_end - _p);
425 }
426 int last_line_no() {
427 return _line_no - 1;
428 }
429
430 void corrupted(const char *p, const char *msg);
431
432 inline void corrupted_if(bool cond, const char *msg) {
433 if (cond) {
434 corrupted(_p, msg);
435 }
436 }
437
438 bool skip_newline();
439 int skip(char must_be_char);
440 void skip_past(char c);
441 void check_version(const char* ver);
442
443 inline void get_num(char delim, int *num) {
444 const char* p = _p;
445 const char* end = _end;
446 u8 n = 0;
447
448 while (p < end) {
449 char c = *p++;
450 if ('0' <= c && c <= '9') {
451 n = n * 10 + (c - '0');
452 if (n > (u8)INT_MAX) {
453 corrupted(_p, "Num overflow");
454 }
455 } else if (c == delim) {
456 _p = p;
457 *num = (int)n;
458 return;
459 } else {
460 // Not [0-9], not 'delim'
461 corrupted(_p, "Unrecognized format");;
462 }
463 }
464
465 corrupted(_end, "Incorrect format");
466 ShouldNotReachHere();
467 }
468
469 void scan_prefix_type();
470 int scan_prefix(int* utf8_length);
471 int scan_string_prefix();
472 int scan_symbol_prefix();
473
474 int unescape(const char* from, const char* end, int count);
475 void get_utf8(char* utf8_buffer, int utf8_length);
476 static void put_utf8(outputStream* st, const char* utf8_string, size_t utf8_length);
477 };
478
479 #endif // SHARE_CLASSFILE_COMPACTHASHTABLE_HPP