/* * Copyright (c) 1997, 2015, Oracle and/or its affiliates. All rights reserved. * DO NOT ALTER OR REMOVE COPYRIGHT NOTICES OR THIS FILE HEADER. * * This code is free software; you can redistribute it and/or modify it * under the terms of the GNU General Public License version 2 only, as * published by the Free Software Foundation. * * This code is distributed in the hope that it will be useful, but WITHOUT * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License * version 2 for more details (a copy is included in the LICENSE file that * accompanied this code). * * You should have received a copy of the GNU General Public License version * 2 along with this work; if not, write to the Free Software Foundation, * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA. * * Please contact Oracle, 500 Oracle Parkway, Redwood Shores, CA 94065 USA * or visit www.oracle.com if you need additional information or have any * questions. * */ #include "precompiled.hpp" #include "classfile/altHashing.hpp" #include "classfile/compactHashtable.hpp" #include "classfile/javaClasses.hpp" #include "classfile/stringTable.hpp" #include "classfile/systemDictionary.hpp" #include "gc/shared/collectedHeap.inline.hpp" #include "gc/shared/gcLocker.inline.hpp" #include "memory/allocation.inline.hpp" #include "memory/filemap.hpp" #include "oops/oop.inline.hpp" #include "runtime/atomic.inline.hpp" #include "runtime/mutexLocker.hpp" #include "utilities/hashtable.inline.hpp" #include "utilities/macros.hpp" #if INCLUDE_ALL_GCS #include "gc/g1/g1CollectedHeap.hpp" #include "gc/g1/g1SATBCardTableModRefBS.hpp" #include "gc/g1/g1StringDedup.hpp" #endif // the number of buckets a thread claims const int ClaimChunkSize = 32; #ifdef ASSERT class StableMemoryChecker : public StackObj { enum { _bufsize = wordSize*4 }; address _region; jint _size; u1 _save_buf[_bufsize]; int sample(u1* save_buf) { if (_size <= _bufsize) { memcpy(save_buf, _region, _size); return _size; } else { // copy head and tail memcpy(&save_buf[0], _region, _bufsize/2); memcpy(&save_buf[_bufsize/2], _region + _size - _bufsize/2, _bufsize/2); return (_bufsize/2)*2; } } public: StableMemoryChecker(const void* region, jint size) { _region = (address) region; _size = size; sample(_save_buf); } bool verify() { u1 check_buf[sizeof(_save_buf)]; int check_size = sample(check_buf); return (0 == memcmp(_save_buf, check_buf, check_size)); } void set_region(const void* region) { _region = (address) region; } }; #endif // -------------------------------------------------------------------------- StringTable* StringTable::_the_table = NULL; bool StringTable::_ignore_shared_strings = false; bool StringTable::_needs_rehashing = false; volatile int StringTable::_parallel_claimed_idx = 0; CompactHashtable StringTable::_shared_table; // Pick hashing algorithm template unsigned int StringTable::hash_string(const T* s, int len) { return use_alternate_hashcode() ? AltHashing::murmur3_32(seed(), s, len) : java_lang_String::hash_code(s, len); } // Explicit instantiation for all supported types. template unsigned int StringTable::hash_string(const jchar* s, int len); template unsigned int StringTable::hash_string(const jbyte* s, int len); oop StringTable::lookup_shared(jchar* name, int len) { // java_lang_String::hash_code() was used to compute hash values in the shared table. Don't // use the hash value from StringTable::hash_string() as it might use alternate hashcode. return _shared_table.lookup((const char*)name, java_lang_String::hash_code(name, len), len); } oop StringTable::lookup_in_main_table(int index, jchar* name, int len, unsigned int hash) { int count = 0; for (HashtableEntry* l = bucket(index); l != NULL; l = l->next()) { count++; if (l->hash() == hash) { if (java_lang_String::equals(l->literal(), name, len)) { return l->literal(); } } } // If the bucket size is too deep check if this hash code is insufficient. if (count >= rehash_count && !needs_rehashing()) { _needs_rehashing = check_rehash_table(count); } return NULL; } oop StringTable::basic_add(int index_arg, Handle string, jchar* name, int len, unsigned int hashValue_arg, TRAPS) { assert(java_lang_String::equals(string(), name, len), "string must be properly initialized"); // Cannot hit a safepoint in this function because the "this" pointer can move. No_Safepoint_Verifier nsv; // Check if the symbol table has been rehashed, if so, need to recalculate // the hash value and index before second lookup. unsigned int hashValue; int index; if (use_alternate_hashcode()) { hashValue = hash_string(name, len); index = hash_to_index(hashValue); } else { hashValue = hashValue_arg; index = index_arg; } // Since look-up was done lock-free, we need to check if another // thread beat us in the race to insert the symbol. // No need to lookup the shared table from here since the caller (intern()) already did oop test = lookup_in_main_table(index, name, len, hashValue); // calls lookup(u1*, int) if (test != NULL) { // Entry already added return test; } HashtableEntry* entry = new_entry(hashValue, string()); add_entry(index, entry); return string(); } oop StringTable::lookup(Symbol* symbol) { ResourceMark rm; int length; jchar* chars = symbol->as_unicode(length); return lookup(chars, length); } // Tell the GC that this string was looked up in the StringTable. static void ensure_string_alive(oop string) { // A lookup in the StringTable could return an object that was previously // considered dead. The SATB part of G1 needs to get notified about this // potential resurrection, otherwise the marking might not find the object. #if INCLUDE_ALL_GCS if (UseG1GC && string != NULL) { G1SATBCardTableModRefBS::enqueue(string); } #endif } oop StringTable::lookup(jchar* name, int len) { oop string = lookup_shared(name, len); if (string != NULL) { return string; } unsigned int hash = hash_string(name, len); int index = the_table()->hash_to_index(hash); string = the_table()->lookup_in_main_table(index, name, len, hash); ensure_string_alive(string); return string; } oop StringTable::intern(Handle string_or_null, jchar* name, int len, TRAPS) { oop found_string = lookup_shared(name, len); if (found_string != NULL) { return found_string; } unsigned int hashValue = hash_string(name, len); int index = the_table()->hash_to_index(hashValue); found_string = the_table()->lookup_in_main_table(index, name, len, hashValue); // Found if (found_string != NULL) { ensure_string_alive(found_string); return found_string; } debug_only(StableMemoryChecker smc(name, len * sizeof(name[0]))); assert(!Universe::heap()->is_in_reserved(name), "proposed name of symbol must be stable"); Handle string; // try to reuse the string if possible if (!string_or_null.is_null()) { string = string_or_null; } else { string = java_lang_String::create_from_unicode(name, len, CHECK_NULL); } #if INCLUDE_ALL_GCS if (G1StringDedup::is_enabled()) { // Deduplicate the string before it is interned. Note that we should never // deduplicate a string after it has been interned. Doing so will counteract // compiler optimizations done on e.g. interned string literals. G1StringDedup::deduplicate(string()); } #endif // Grab the StringTable_lock before getting the_table() because it could // change at safepoint. oop added_or_found; { MutexLocker ml(StringTable_lock, THREAD); // Otherwise, add to symbol to table added_or_found = the_table()->basic_add(index, string, name, len, hashValue, CHECK_NULL); } ensure_string_alive(added_or_found); return added_or_found; } oop StringTable::intern(Symbol* symbol, TRAPS) { if (symbol == NULL) return NULL; ResourceMark rm(THREAD); int length; jchar* chars = symbol->as_unicode(length); Handle string; oop result = intern(string, chars, length, CHECK_NULL); return result; } oop StringTable::intern(oop string, TRAPS) { if (string == NULL) return NULL; ResourceMark rm(THREAD); int length; Handle h_string (THREAD, string); jchar* chars = java_lang_String::as_unicode_string(string, length, CHECK_NULL); oop result = intern(h_string, chars, length, CHECK_NULL); return result; } oop StringTable::intern(const char* utf8_string, TRAPS) { if (utf8_string == NULL) return NULL; ResourceMark rm(THREAD); int length = UTF8::unicode_length(utf8_string); jchar* chars = NEW_RESOURCE_ARRAY(jchar, length); UTF8::convert_to_unicode(utf8_string, chars, length); Handle string; oop result = intern(string, chars, length, CHECK_NULL); return result; } void StringTable::unlink_or_oops_do(BoolObjectClosure* is_alive, OopClosure* f, int* processed, int* removed) { buckets_unlink_or_oops_do(is_alive, f, 0, the_table()->table_size(), processed, removed); } void StringTable::possibly_parallel_unlink_or_oops_do(BoolObjectClosure* is_alive, OopClosure* f, int* processed, int* removed) { // Readers of the table are unlocked, so we should only be removing // entries at a safepoint. assert(SafepointSynchronize::is_at_safepoint(), "must be at safepoint"); const int limit = the_table()->table_size(); for (;;) { // Grab next set of buckets to scan int start_idx = Atomic::add(ClaimChunkSize, &_parallel_claimed_idx) - ClaimChunkSize; if (start_idx >= limit) { // End of table break; } int end_idx = MIN2(limit, start_idx + ClaimChunkSize); buckets_unlink_or_oops_do(is_alive, f, start_idx, end_idx, processed, removed); } } void StringTable::buckets_oops_do(OopClosure* f, int start_idx, int end_idx) { const int limit = the_table()->table_size(); assert(0 <= start_idx && start_idx <= limit, "start_idx (%d) is out of bounds", start_idx); assert(0 <= end_idx && end_idx <= limit, "end_idx (%d) is out of bounds", end_idx); assert(start_idx <= end_idx, "Index ordering: start_idx=%d, end_idx=%d", start_idx, end_idx); for (int i = start_idx; i < end_idx; i += 1) { HashtableEntry* entry = the_table()->bucket(i); while (entry != NULL) { assert(!entry->is_shared(), "CDS not used for the StringTable"); f->do_oop((oop*)entry->literal_addr()); entry = entry->next(); } } } void StringTable::buckets_unlink_or_oops_do(BoolObjectClosure* is_alive, OopClosure* f, int start_idx, int end_idx, int* processed, int* removed) { const int limit = the_table()->table_size(); assert(0 <= start_idx && start_idx <= limit, "start_idx (%d) is out of bounds", start_idx); assert(0 <= end_idx && end_idx <= limit, "end_idx (%d) is out of bounds", end_idx); assert(start_idx <= end_idx, "Index ordering: start_idx=%d, end_idx=%d", start_idx, end_idx); for (int i = start_idx; i < end_idx; ++i) { HashtableEntry** p = the_table()->bucket_addr(i); HashtableEntry* entry = the_table()->bucket(i); while (entry != NULL) { assert(!entry->is_shared(), "CDS not used for the StringTable"); if (is_alive->do_object_b(entry->literal())) { if (f != NULL) { f->do_oop((oop*)entry->literal_addr()); } p = entry->next_addr(); } else { *p = entry->next(); the_table()->free_entry(entry); (*removed)++; } (*processed)++; entry = *p; } } } void StringTable::oops_do(OopClosure* f) { buckets_oops_do(f, 0, the_table()->table_size()); } void StringTable::possibly_parallel_oops_do(OopClosure* f) { const int limit = the_table()->table_size(); for (;;) { // Grab next set of buckets to scan int start_idx = Atomic::add(ClaimChunkSize, &_parallel_claimed_idx) - ClaimChunkSize; if (start_idx >= limit) { // End of table break; } int end_idx = MIN2(limit, start_idx + ClaimChunkSize); buckets_oops_do(f, start_idx, end_idx); } } // This verification is part of Universe::verify() and needs to be quick. // See StringTable::verify_and_compare() below for exhaustive verification. void StringTable::verify() { for (int i = 0; i < the_table()->table_size(); ++i) { HashtableEntry* p = the_table()->bucket(i); for ( ; p != NULL; p = p->next()) { oop s = p->literal(); guarantee(s != NULL, "interned string is NULL"); unsigned int h = java_lang_String::hash_string(s); guarantee(p->hash() == h, "broken hash in string table entry"); guarantee(the_table()->hash_to_index(h) == i, "wrong index in string table"); } } } void StringTable::dump(outputStream* st, bool verbose) { if (!verbose) { the_table()->dump_table(st, "StringTable"); } else { Thread* THREAD = Thread::current(); st->print_cr("VERSION: 1.1"); for (int i = 0; i < the_table()->table_size(); ++i) { HashtableEntry* p = the_table()->bucket(i); for ( ; p != NULL; p = p->next()) { oop s = p->literal(); typeArrayOop value = java_lang_String::value(s); int length = java_lang_String::length(s); bool is_latin1 = java_lang_String::is_latin1(s); if (length <= 0) { st->print("%d: ", length); } else { ResourceMark rm(THREAD); int utf8_length; char* utf8_string; if (!is_latin1) { jchar* chars = value->char_at_addr(0); utf8_length = UNICODE::utf8_length(chars, length); utf8_string = UNICODE::as_utf8(chars, length); } else { jbyte* bytes = value->byte_at_addr(0); utf8_length = UNICODE::utf8_length(bytes, length); utf8_string = UNICODE::as_utf8(bytes, length); } st->print("%d: ", utf8_length); HashtableTextDump::put_utf8(st, utf8_string, utf8_length); } st->cr(); } } } } StringTable::VerifyRetTypes StringTable::compare_entries( int bkt1, int e_cnt1, HashtableEntry* e_ptr1, int bkt2, int e_cnt2, HashtableEntry* e_ptr2) { // These entries are sanity checked by verify_and_compare_entries() // before this function is called. oop str1 = e_ptr1->literal(); oop str2 = e_ptr2->literal(); if (str1 == str2) { tty->print_cr("ERROR: identical oop values (0x" PTR_FORMAT ") " "in entry @ bucket[%d][%d] and entry @ bucket[%d][%d]", p2i(str1), bkt1, e_cnt1, bkt2, e_cnt2); return _verify_fail_continue; } if (java_lang_String::equals(str1, str2)) { tty->print_cr("ERROR: identical String values in entry @ " "bucket[%d][%d] and entry @ bucket[%d][%d]", bkt1, e_cnt1, bkt2, e_cnt2); return _verify_fail_continue; } return _verify_pass; } StringTable::VerifyRetTypes StringTable::verify_entry(int bkt, int e_cnt, HashtableEntry* e_ptr, StringTable::VerifyMesgModes mesg_mode) { VerifyRetTypes ret = _verify_pass; // be optimistic oop str = e_ptr->literal(); if (str == NULL) { if (mesg_mode == _verify_with_mesgs) { tty->print_cr("ERROR: NULL oop value in entry @ bucket[%d][%d]", bkt, e_cnt); } // NULL oop means no more verifications are possible return _verify_fail_done; } if (str->klass() != SystemDictionary::String_klass()) { if (mesg_mode == _verify_with_mesgs) { tty->print_cr("ERROR: oop is not a String in entry @ bucket[%d][%d]", bkt, e_cnt); } // not a String means no more verifications are possible return _verify_fail_done; } unsigned int h = java_lang_String::hash_string(str); if (e_ptr->hash() != h) { if (mesg_mode == _verify_with_mesgs) { tty->print_cr("ERROR: broken hash value in entry @ bucket[%d][%d], " "bkt_hash=%d, str_hash=%d", bkt, e_cnt, e_ptr->hash(), h); } ret = _verify_fail_continue; } if (the_table()->hash_to_index(h) != bkt) { if (mesg_mode == _verify_with_mesgs) { tty->print_cr("ERROR: wrong index value for entry @ bucket[%d][%d], " "str_hash=%d, hash_to_index=%d", bkt, e_cnt, h, the_table()->hash_to_index(h)); } ret = _verify_fail_continue; } return ret; } // See StringTable::verify() above for the quick verification that is // part of Universe::verify(). This verification is exhaustive and // reports on every issue that is found. StringTable::verify() only // reports on the first issue that is found. // // StringTable::verify_entry() checks: // - oop value != NULL (same as verify()) // - oop value is a String // - hash(String) == hash in entry (same as verify()) // - index for hash == index of entry (same as verify()) // // StringTable::compare_entries() checks: // - oops are unique across all entries // - String values are unique across all entries // int StringTable::verify_and_compare_entries() { assert(StringTable_lock->is_locked(), "sanity check"); int fail_cnt = 0; // first, verify all the entries individually: for (int bkt = 0; bkt < the_table()->table_size(); bkt++) { HashtableEntry* e_ptr = the_table()->bucket(bkt); for (int e_cnt = 0; e_ptr != NULL; e_ptr = e_ptr->next(), e_cnt++) { VerifyRetTypes ret = verify_entry(bkt, e_cnt, e_ptr, _verify_with_mesgs); if (ret != _verify_pass) { fail_cnt++; } } } // Optimization: if the above check did not find any failures, then // the comparison loop below does not need to call verify_entry() // before calling compare_entries(). If there were failures, then we // have to call verify_entry() to see if the entry can be passed to // compare_entries() safely. When we call verify_entry() in the loop // below, we do so quietly to void duplicate messages and we don't // increment fail_cnt because the failures have already been counted. bool need_entry_verify = (fail_cnt != 0); // second, verify all entries relative to each other: for (int bkt1 = 0; bkt1 < the_table()->table_size(); bkt1++) { HashtableEntry* e_ptr1 = the_table()->bucket(bkt1); for (int e_cnt1 = 0; e_ptr1 != NULL; e_ptr1 = e_ptr1->next(), e_cnt1++) { if (need_entry_verify) { VerifyRetTypes ret = verify_entry(bkt1, e_cnt1, e_ptr1, _verify_quietly); if (ret == _verify_fail_done) { // cannot use the current entry to compare against other entries continue; } } for (int bkt2 = bkt1; bkt2 < the_table()->table_size(); bkt2++) { HashtableEntry* e_ptr2 = the_table()->bucket(bkt2); int e_cnt2; for (e_cnt2 = 0; e_ptr2 != NULL; e_ptr2 = e_ptr2->next(), e_cnt2++) { if (bkt1 == bkt2 && e_cnt2 <= e_cnt1) { // skip the entries up to and including the one that // we're comparing against continue; } if (need_entry_verify) { VerifyRetTypes ret = verify_entry(bkt2, e_cnt2, e_ptr2, _verify_quietly); if (ret == _verify_fail_done) { // cannot compare against this entry continue; } } // compare two entries, report and count any failures: if (compare_entries(bkt1, e_cnt1, e_ptr1, bkt2, e_cnt2, e_ptr2) != _verify_pass) { fail_cnt++; } } } } } return fail_cnt; } // Create a new table and using alternate hash code, populate the new table // with the existing strings. Set flag to use the alternate hash code afterwards. void StringTable::rehash_table() { assert(SafepointSynchronize::is_at_safepoint(), "must be at safepoint"); // This should never happen with -Xshare:dump but it might in testing mode. if (DumpSharedSpaces) return; StringTable* new_table = new StringTable(); // Rehash the table the_table()->move_to(new_table); // Delete the table and buckets (entries are reused in new table). delete _the_table; // Don't check if we need rehashing until the table gets unbalanced again. // Then rehash with a new global seed. _needs_rehashing = false; _the_table = new_table; } // Utility for dumping strings StringtableDCmd::StringtableDCmd(outputStream* output, bool heap) : DCmdWithParser(output, heap), _verbose("-verbose", "Dump the content of each string in the table", "BOOLEAN", false, "false") { _dcmdparser.add_dcmd_option(&_verbose); } void StringtableDCmd::execute(DCmdSource source, TRAPS) { VM_DumpHashtable dumper(output(), VM_DumpHashtable::DumpStrings, _verbose.value()); VMThread::execute(&dumper); } int StringtableDCmd::num_arguments() { ResourceMark rm; StringtableDCmd* dcmd = new StringtableDCmd(NULL, false); if (dcmd != NULL) { DCmdMark mark(dcmd); return dcmd->_dcmdparser.num_arguments(); } else { return 0; } } // Sharing bool StringTable::copy_shared_string(GrowableArray *string_space, CompactHashtableWriter* ch_table) { #if INCLUDE_CDS && INCLUDE_ALL_GCS && defined(_LP64) && !defined(_WINDOWS) assert(UseG1GC, "Only support G1 GC"); assert(UseCompressedOops && UseCompressedClassPointers, "Only support UseCompressedOops and UseCompressedClassPointers enabled"); Thread* THREAD = Thread::current(); G1CollectedHeap::heap()->begin_archive_alloc_range(); for (int i = 0; i < the_table()->table_size(); ++i) { HashtableEntry* bucket = the_table()->bucket(i); for ( ; bucket != NULL; bucket = bucket->next()) { oop s = bucket->literal(); unsigned int hash = java_lang_String::hash_code(s); if (hash == 0) { continue; } // allocate the new 'value' array first typeArrayOop v = java_lang_String::value(s); int v_len = v->size(); typeArrayOop new_v; if (G1CollectedHeap::heap()->is_archive_alloc_too_large(v_len)) { continue; // skip the current String. The 'value' array is too large to handle } else { new_v = (typeArrayOop)G1CollectedHeap::heap()->archive_mem_allocate(v_len); if (new_v == NULL) { return false; // allocation failed } } // now allocate the new String object int s_len = s->size(); oop new_s = (oop)G1CollectedHeap::heap()->archive_mem_allocate(s_len); if (new_s == NULL) { return false; } s->identity_hash(); v->identity_hash(); // copy the objects' data Copy::aligned_disjoint_words((HeapWord*)s, (HeapWord*)new_s, s_len); Copy::aligned_disjoint_words((HeapWord*)v, (HeapWord*)new_v, v_len); // adjust the pointer to the 'value' field in the new String oop. Also pre-compute and set the // 'hash' field. That avoids "write" to the shared strings at runtime by the deduplication process. java_lang_String::set_value_raw(new_s, new_v); if (java_lang_String::hash(new_s) == 0) { java_lang_String::set_hash(new_s, hash); } // add to the compact table ch_table->add(hash, new_s); } } G1CollectedHeap::heap()->end_archive_alloc_range(string_space, os::vm_allocation_granularity()); assert(string_space->length() <= 2, "sanity"); #endif return true; } bool StringTable::copy_compact_table(char** top, char *end, GrowableArray *string_space, size_t* space_size) { #if INCLUDE_CDS && defined(_LP64) && !defined(_WINDOWS) if (!(UseG1GC && UseCompressedOops && UseCompressedClassPointers)) { if (PrintSharedSpaces) { tty->print_cr("Shared strings are excluded from the archive as UseG1GC, " "UseCompressedOops and UseCompressedClassPointers are required."); } return true; } CompactHashtableWriter ch_table(CompactHashtable::_string_table, the_table()->number_of_entries(), &MetaspaceShared::stats()->string); // Copy the interned strings into the "string space" within the java heap if (!copy_shared_string(string_space, &ch_table)) { return false; } for (int i = 0; i < string_space->length(); i++) { *space_size += string_space->at(i).byte_size(); } // Now dump the compact table if (*top + ch_table.get_required_bytes() > end) { // not enough space left return false; } ch_table.dump(top, end); *top = (char*)align_pointer_up(*top, sizeof(void*)); #endif return true; } void StringTable::shared_oops_do(OopClosure* f) { #if INCLUDE_CDS && defined(_LP64) && !defined(_WINDOWS) _shared_table.oops_do(f); #endif } const char* StringTable::init_shared_table(FileMapInfo *mapinfo, char *buffer) { #if INCLUDE_CDS && defined(_LP64) && !defined(_WINDOWS) if (mapinfo->space_capacity(MetaspaceShared::first_string) == 0) { // no shared string data return buffer; } // initialize the shared table juint *p = (juint*)buffer; const char* end = _shared_table.init( CompactHashtable::_string_table, (char*)p); const char* aligned_end = (const char*)align_pointer_up(end, sizeof(void*)); if (_ignore_shared_strings) { _shared_table.reset(); } return aligned_end; #endif return buffer; }