1 /*
2 * Copyright (c) 1997, 2024, 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/cds_globals.hpp"
29 #include "oops/array.hpp"
30 #include "oops/symbol.hpp"
31 #include "runtime/globals.hpp"
32 #include "utilities/growableArray.hpp"
33
34
35 template <
36 typename K,
37 typename V,
38 V (*DECODE)(address base_address, u4 offset),
39 bool (*EQUALS)(V value, K key, int len)
40 >
41 class CompactHashtable;
42 class NumberSeq;
43 class SimpleCompactHashtable;
44 class SerializeClosure;
45
46 // Stats for symbol tables in the CDS archive
47 class CompactHashtableStats {
48 public:
49 int hashentry_count;
50 int hashentry_bytes;
51 int bucket_count;
52 int bucket_bytes;
53
54 CompactHashtableStats() :
55 hashentry_count(0), hashentry_bytes(0),
56 bucket_count(0), bucket_bytes(0) {}
57 };
58
59 #if INCLUDE_CDS
60 /////////////////////////////////////////////////////////////////////////
61 //
62 // The compact hash table writer. Used at dump time for writing out
63 // the compact table to the shared archive.
64 //
65 // At dump time, the CompactHashtableWriter obtains all entries from the
66 // symbol/string table and adds them to a new temporary hash table. The hash
67 // table size (number of buckets) is calculated using
68 // '(num_entries + bucket_size - 1) / bucket_size'. The default bucket
69 // size is 4 and can be changed by -XX:SharedSymbolTableBucketSize option.
70 // 4 is chosen because it produces smaller sized bucket on average for
71 // faster lookup. It also has relatively small number of empty buckets and
72 // good distribution of the entries.
73 //
74 // We use a simple hash function (hash % num_bucket) for the table.
75 // The new table is compacted when written out. Please see comments
76 // above the CompactHashtable class for the table layout detail. The bucket
77 // offsets are written to the archive as part of the compact table. The
78 // bucket offset is encoded in the low 30-bit (0-29) and the bucket type
79 // (regular or compact) are encoded in bit[31, 30]. For buckets with more
80 // than one entry, both hash and entry offset are written to the
81 // table. For buckets with only one entry, only the entry offset is written
82 // to the table and the buckets are tagged as compact in their type bits.
83 // Buckets without entry are skipped from the table. Their offsets are
84 // still written out for faster lookup.
85 //
86 class CompactHashtableWriter: public StackObj {
87 public:
88 class Entry {
89 unsigned int _hash;
90 u4 _value;
91
92 public:
93 Entry() {}
94 Entry(unsigned int hash, u4 val) : _hash(hash), _value(val) {}
95
96 u4 value() {
97 return _value;
98 }
99 unsigned int hash() {
100 return _hash;
101 }
102
103 bool operator==(const CompactHashtableWriter::Entry& other) {
104 return (_value == other._value && _hash == other._hash);
105 }
106 }; // class CompactHashtableWriter::Entry
107
108 private:
109 int _num_entries_written;
110 int _num_buckets;
111 int _num_empty_buckets;
112 int _num_value_only_buckets;
113 int _num_other_buckets;
114 GrowableArray<Entry>** _buckets;
115 CompactHashtableStats* _stats;
116 Array<u4>* _compact_buckets;
117 Array<u4>* _compact_entries;
118
119 public:
120 // This is called at dump-time only
121 CompactHashtableWriter(int num_entries, CompactHashtableStats* stats);
122 ~CompactHashtableWriter();
123
124 void add(unsigned int hash, u4 value);
125
126 private:
127 void allocate_table();
128 void dump_table(NumberSeq* summary);
129 static int calculate_num_buckets(int num_entries) {
130 int num_buckets = num_entries / SharedSymbolTableBucketSize;
131 // calculation of num_buckets can result in zero buckets, we need at least one
132 return (num_buckets < 1) ? 1 : num_buckets;
133 }
134
135 public:
136 void dump(SimpleCompactHashtable *cht, const char* table_name);
137
138 static size_t estimate_size(int num_entries);
139 };
140 #endif // INCLUDE_CDS
141
142 #define REGULAR_BUCKET_TYPE 0
143 #define VALUE_ONLY_BUCKET_TYPE 1
144 #define TABLEEND_BUCKET_TYPE 3
145 #define BUCKET_OFFSET_MASK 0x3FFFFFFF
146 #define BUCKET_OFFSET(info) ((info) & BUCKET_OFFSET_MASK)
147 #define BUCKET_TYPE_SHIFT 30
148 #define BUCKET_TYPE(info) (((info) & ~BUCKET_OFFSET_MASK) >> BUCKET_TYPE_SHIFT)
149 #define BUCKET_INFO(offset, type) (((type) << BUCKET_TYPE_SHIFT) | ((offset) & BUCKET_OFFSET_MASK))
150
151 /////////////////////////////////////////////////////////////////////////////
152 //
153 // CompactHashtable is used to store the CDS archive's symbol/string tables.
154 //
155 // Because these tables are read-only (no entries can be added/deleted) at run-time
156 // and tend to have large number of entries, we try to minimize the footprint
157 // cost per entry.
158 //
159 // The CompactHashtable is split into two arrays
160 //
161 // u4 buckets[num_buckets+1]; // bit[31,30]: type; bit[29-0]: offset
162 // u4 entries[<variable size>]
163 //
164 // The size of buckets[] is 'num_buckets + 1'. Each entry of
165 // buckets[] is a 32-bit encoding of the bucket type and bucket offset,
166 // with the type in the left-most 2-bit and offset in the remaining 30-bit.
167 // The last entry is a special type. It contains the end of the last
168 // bucket.
169 //
170 // There are two types of buckets, regular buckets and value_only buckets. The
171 // value_only buckets have '01' in their highest 2-bit, and regular buckets have
172 // '00' in their highest 2-bit.
173 //
174 // For normal buckets, each entry is 8 bytes in the entries[]:
175 // u4 hash; /* symbol/string hash */
176 // union {
177 // u4 offset; /* Symbol* sym = (Symbol*)(base_address + offset) */
178 // narrowOop str; /* String narrowOop encoding */
179 // }
180 //
181 //
182 // For value_only buckets, each entry has only the 4-byte 'offset' in the entries[].
183 //
184 // Example -- note that the second bucket is a VALUE_ONLY_BUCKET_TYPE so the hash code
185 // is skipped.
186 // buckets[0, 4, 5, ....]
187 // | | |
188 // | | +---+
189 // | | |
190 // | +----+ |
191 // v v v
192 // entries[H,O,H,O,O,H,O,H,O.....]
193 //
194 // See CompactHashtable::lookup() for how the table is searched at runtime.
195 // See CompactHashtableWriter::dump() for how the table is written at CDS
196 // dump time.
197 //
198 class SimpleCompactHashtable {
199 protected:
200 address _base_address;
201 u4 _bucket_count;
202 u4 _entry_count;
203 u4* _buckets;
204 u4* _entries;
205
206 public:
207 SimpleCompactHashtable() :
208 _base_address(nullptr),
209 _bucket_count(0),
210 _entry_count(0),
211 _buckets(nullptr),
212 _entries(nullptr)
213 {}
214
215 void reset() {
216 _base_address = nullptr;
217 _bucket_count = 0;
218 _entry_count = 0;
219 _buckets = nullptr;
220 _entries = nullptr;
221 }
222
223 void init(address base_address, u4 entry_count, u4 bucket_count, u4* buckets, u4* entries);
224
225 // Read/Write the table's header from/to the CDS archive
226 void serialize_header(SerializeClosure* soc) NOT_CDS_RETURN;
227
228 inline bool empty() const {
229 return (_entry_count == 0);
230 }
231
232 inline size_t entry_count() const {
233 return _entry_count;
234 }
235
236 static size_t calculate_header_size();
237 };
238
239 template <
240 typename K,
241 typename V,
242 V (*DECODE)(address base_address, u4 offset),
243 bool (*EQUALS)(V value, K key, int len)
244 >
245 class CompactHashtable : public SimpleCompactHashtable {
246 friend class VMStructs;
247
248 V decode(u4 offset) const {
249 return DECODE(_base_address, offset);
250 }
251
252 public:
253 // Lookup a value V from the compact table using key K
254 inline V lookup(K key, unsigned int hash, int len) const {
255 if (_entry_count > 0) {
256 int index = hash % _bucket_count;
257 u4 bucket_info = _buckets[index];
258 u4 bucket_offset = BUCKET_OFFSET(bucket_info);
259 int bucket_type = BUCKET_TYPE(bucket_info);
260 u4* entry = _entries + bucket_offset;
261
262 if (bucket_type == VALUE_ONLY_BUCKET_TYPE) {
263 V value = decode(entry[0]);
264 if (EQUALS(value, key, len)) {
265 return value;
266 }
267 } else {
268 // This is a regular bucket, which has more than one
269 // entries. Each entry is a pair of entry (hash, offset).
270 // Seek until the end of the bucket.
271 u4* entry_max = _entries + BUCKET_OFFSET(_buckets[index + 1]);
272 while (entry < entry_max) {
273 unsigned int h = (unsigned int)(entry[0]);
274 if (h == hash) {
275 V value = decode(entry[1]);
276 if (EQUALS(value, key, len)) {
277 return value;
278 }
279 }
280 entry += 2;
281 }
282 }
283 }
284 return nullptr;
285 }
286
287 template <class ITER>
288 inline void iterate(ITER* iter) const {
289 for (u4 i = 0; i < _bucket_count; i++) {
290 u4 bucket_info = _buckets[i];
291 u4 bucket_offset = BUCKET_OFFSET(bucket_info);
292 int bucket_type = BUCKET_TYPE(bucket_info);
293 u4* entry = _entries + bucket_offset;
294
295 if (bucket_type == VALUE_ONLY_BUCKET_TYPE) {
296 iter->do_value(decode(entry[0]));
297 } else {
298 u4*entry_max = _entries + BUCKET_OFFSET(_buckets[i + 1]);
299 while (entry < entry_max) {
300 iter->do_value(decode(entry[1]));
301 entry += 2;
302 }
303 }
304 }
305 }
306
307 template<typename Function>
308 inline void iterate(Function function) const { // lambda enabled API
309 for (u4 i = 0; i < _bucket_count; i++) {
310 u4 bucket_info = _buckets[i];
311 u4 bucket_offset = BUCKET_OFFSET(bucket_info);
312 int bucket_type = BUCKET_TYPE(bucket_info);
313 u4* entry = _entries + bucket_offset;
314
315 if (bucket_type == VALUE_ONLY_BUCKET_TYPE) {
316 function(decode(entry[0]));
317 } else {
318 u4* entry_max = _entries + BUCKET_OFFSET(_buckets[i + 1]);
319 while (entry < entry_max) {
320 function(decode(entry[1]));
321 entry += 2;
322 }
323 }
324 }
325 }
326
327 void print_table_statistics(outputStream* st, const char* name) {
328 st->print_cr("%s statistics:", name);
329 int total_entries = 0;
330 int max_bucket = 0;
331 for (u4 i = 0; i < _bucket_count; i++) {
332 u4 bucket_info = _buckets[i];
333 int bucket_type = BUCKET_TYPE(bucket_info);
334 int bucket_size;
335
336 if (bucket_type == VALUE_ONLY_BUCKET_TYPE) {
337 bucket_size = 1;
338 } else {
339 bucket_size = (BUCKET_OFFSET(_buckets[i + 1]) - BUCKET_OFFSET(bucket_info)) / 2;
340 }
341 total_entries += bucket_size;
342 if (max_bucket < bucket_size) {
343 max_bucket = bucket_size;
344 }
345 }
346 st->print_cr("Number of buckets : %9d", _bucket_count);
347 st->print_cr("Number of entries : %9d", total_entries);
348 st->print_cr("Maximum bucket size : %9d", max_bucket);
349 }
350 };
351
352 ////////////////////////////////////////////////////////////////////////
353 //
354 // OffsetCompactHashtable -- This is used to store many types of objects
355 // in the CDS archive. On 64-bit platforms, we save space by using a 32-bit
356 // offset from the CDS base address.
357
358 template <typename V>
359 inline V read_value_from_compact_hashtable(address base_address, u4 offset) {
360 return (V)(base_address + offset);
361 }
362
363 template <
364 typename K,
365 typename V,
366 bool (*EQUALS)(V value, K key, int len)
367 >
368 class OffsetCompactHashtable : public CompactHashtable<
369 K, V, read_value_from_compact_hashtable<V>, EQUALS> {
370 };
371
372
373 ////////////////////////////////////////////////////////////////////////
374 //
375 // Read/Write the contents of a hashtable textual dump (created by
376 // SymbolTable::dump and StringTable::dump).
377 // Because the dump file may be big (hundred of MB in extreme cases),
378 // we use mmap for fast access when reading it.
379 //
380 class HashtableTextDump {
381 int _fd;
382 const char* _base;
383 const char* _p;
384 const char* _end;
385 const char* _filename;
386 size_t _size;
387 int _prefix_type;
388 int _line_no;
389 public:
390 HashtableTextDump(const char* filename);
391 ~HashtableTextDump();
392
393 enum {
394 SymbolPrefix = 1 << 0,
395 StringPrefix = 1 << 1,
396 Unknown = 1 << 2
397 };
398
399 void quit(const char* err, const char* msg);
400
401 inline int remain() {
402 return (int)(_end - _p);
403 }
404 int last_line_no() {
405 return _line_no - 1;
406 }
407
408 void corrupted(const char *p, const char *msg);
409
410 inline void corrupted_if(bool cond, const char *msg) {
411 if (cond) {
412 corrupted(_p, msg);
413 }
414 }
415
416 bool skip_newline();
417 int skip(char must_be_char);
418 void skip_past(char c);
419 void check_version(const char* ver);
420
421 inline void get_num(char delim, int *num) {
422 const char* p = _p;
423 const char* end = _end;
424 u8 n = 0;
425
426 while (p < end) {
427 char c = *p++;
428 if ('0' <= c && c <= '9') {
429 n = n * 10 + (c - '0');
430 if (n > (u8)INT_MAX) {
431 corrupted(_p, "Num overflow");
432 }
433 } else if (c == delim) {
434 _p = p;
435 *num = (int)n;
436 return;
437 } else {
438 // Not [0-9], not 'delim'
439 corrupted(_p, "Unrecognized format");;
440 }
441 }
442
443 corrupted(_end, "Incorrect format");
444 ShouldNotReachHere();
445 }
446
447 void scan_prefix_type();
448 int scan_prefix(int* utf8_length);
449 int scan_string_prefix();
450 int scan_symbol_prefix();
451
452 int unescape(const char* from, const char* end, int count);
453 void get_utf8(char* utf8_buffer, int utf8_length);
454 static void put_utf8(outputStream* st, const char* utf8_string, size_t utf8_length);
455 };
456
457 #endif // SHARE_CLASSFILE_COMPACTHASHTABLE_HPP