网站视频弹窗广告代码,北京网站制作济南,关于加强门户网站建设的通知,wordpress xmlrpc攻击17.4 Java堆空间内存分配
分配Java堆内存前#xff0c;我们先通过两图来了解下C堆、Java堆、内核空间、native本地空间的关系。
1、从图17-1来看#xff0c;Java堆的分配其实就是从Java进程运行时堆中选中一块内存区域来映射
2、从图17-2#xff0c;可以看中各内存空间的…17.4 Java堆空间内存分配
分配Java堆内存前我们先通过两图来了解下C堆、Java堆、内核空间、native本地空间的关系。
1、从图17-1来看Java堆的分配其实就是从Java进程运行时堆中选中一块内存区域来映射
2、从图17-2可以看中各内存空间的关系当然实际的内存区域比这个复杂的多这里只是概括说明
图17-1
图17-2
17.4.1 genCollectedHeap.cpp
17.4.1.1 GenCollectedHeap::initialize
jint GenCollectedHeap::initialize() {// 这一步只是对c2编译器开通使用时做一些参数赋值操作这里就不展开讲CollectedHeap::pre_initialize();// 这里获取分代数_n_gens就是2int i;_n_gens gen_policy()-number_of_generations();// 保证2个值相等wordSize和HeapWordSize分别是在操作系统和Java堆中代表一个字word占用内存的大小这两个值必然相同否则出错guarantee(HeapWordSize wordSize, HeapWordSize must equal wordSize);// Java堆的对齐值这个在章节17.2.1.1中有介绍size_t gen_alignment Generation::GenGrain;// 获取分代对象数组这个在章节17.2.1.1中有介绍数组元素就2个索引0元素表示年轻代索引1元素表示老年代_gen_specs gen_policy()-generations();// 分别遍历新生代和老年代并设置各自分代的空间大小初始值和最大值同时确保内存对齐for (i 0; i _n_gens; i) {_gen_specs[i]-align(gen_alignment);}// 下面才是给Java堆分配空间char* heap_address;size_t total_reserved 0;int n_covered_regions 0;ReservedSpace heap_rs;// 这是最外层Java堆的内存对齐值size_t heap_alignment collector_policy()-heap_alignment();// 分配java堆内存看章节17.4.1.2heap_address allocate(heap_alignment, total_reserved,n_covered_regions, heap_rs);if (!heap_rs.is_reserved()) {vm_shutdown_during_initialization(Could not reserve enough space for object heap);return JNI_ENOMEM;}// 将分配的Java堆内存用 MemRegion 内存区域对象管理起来_reserved MemRegion((HeapWord*)heap_rs.base(),(HeapWord*)(heap_rs.base() heap_rs.size()));// 参数赋值_reserved.set_word_size(0);_reserved.set_start((HeapWord*)heap_rs.base()); // Java堆内存的首地址size_t actual_heap_size heap_rs.size(); // Java堆内存大小// Java堆内存的限制地址也就是不能超过这条线_reserved.set_end((HeapWord*)(heap_rs.base() actual_heap_size)); // 接下来就是创建记忆集、卡表的过程卡表和记忆集都是为了解决跨代引用的实现方案后续讲GC时会有涉及_rem_set collector_policy()-create_rem_set(_reserved, n_covered_regions);set_barrier_set(rem_set()-bs());_gch this;for (i 0; i _n_gens; i) {ReservedSpace this_rs heap_rs.first_part(_gen_specs[i]-max_size(), false, false);_gens[i] _gen_specs[i]-init(this_rs, i, rem_set());heap_rs heap_rs.last_part(_gen_specs[i]-max_size());}clear_incremental_collection_failed();#if INCLUDE_ALL_GCS// If we are running CMS, create the collector responsible// for collecting the CMS generations.if (collector_policy()-is_concurrent_mark_sweep_policy()) {bool success create_cms_collector();if (!success) return JNI_ENOMEM;}
#endif // INCLUDE_ALL_GCSreturn JNI_OK;
}17.4.1.2 GenCollectedHeap::allocate
char* GenCollectedHeap::allocate(size_t alignment,size_t* _total_reserved,int* _n_covered_regions,ReservedSpace* heap_rs){const char overflow_msg[] The size of the object heap VM data exceeds the maximum representable size;// Now figure out the total size.size_t total_reserved 0;int n_covered_regions 0;const size_t pageSize UseLargePages ?os::large_page_size() : os::vm_page_size();assert(alignment % pageSize 0, Must be);// 遍历_gen_specs求得新生代和老年代的分配大小for (int i 0; i _n_gens; i) {total_reserved _gen_specs[i]-max_size();if (total_reserved _gen_specs[i]-max_size()) {vm_exit_during_initialization(overflow_msg);}n_covered_regions _gen_specs[i]-n_covered_regions(); // 最终为2}assert(total_reserved % alignment 0,err_msg(Gen size; total_reserved SIZE_FORMAT , alignmentSIZE_FORMAT, total_reserved, alignment));// Needed until the cardtable is fixed to have the right number// of covered regions.n_covered_regions 2; // 再加2就是4也就是把堆最终分成4个区新生代、S1、S2、老年代*_total_reserved total_reserved;*_n_covered_regions n_covered_regions;// 分配内存实现细节看章节17.4.2*heap_rs Universe::reserve_heap(total_reserved, alignment);return heap_rs-base();
}17.4.2 universe.cpp
17.4.2.1 Universe::reserve_heap
ReservedSpace Universe::reserve_heap(size_t heap_size, size_t alignment) {assert(alignment Arguments::conservative_max_heap_alignment(),err_msg(actual alignment SIZE_FORMAT must be within maximum heap alignment SIZE_FORMAT,alignment, Arguments::conservative_max_heap_alignment()));// 通过内存对齐得到要分配的空间大小size_t total_reserved align_size_up(heap_size, alignment);assert(!UseCompressedOops || (total_reserved (OopEncodingHeapMax - os::vm_page_size())),heap size is too big for compressed oops);// 大页时考虑本系列文章中不考虑大而情况忽略bool use_large_pages UseLargePages is_size_aligned(alignment, os::large_page_size());assert(!UseLargePages|| UseParallelGC|| use_large_pages, Wrong alignment to use large pages);// 取出Java堆的基址base的值32位机器时就是0实现细节看章节17.4.2.2char* addr Universe::preferred_heap_base(total_reserved, alignment, Universe::UnscaledNarrowOop);// 创建一个ReservedHeapSpace对象该对象就是用来保留连续内存地址范围空间的数据结构实现细节看章节17.4.3ReservedHeapSpace total_rs(total_reserved, alignment, use_large_pages, addr);if (UseCompressedOops) {if (addr ! NULL !total_rs.is_reserved()) {// Failed to reserve at specified address - the requested memory// region is taken already, for example, by java launcher.// Try again to reserver heap higher.addr Universe::preferred_heap_base(total_reserved, alignment, Universe::ZeroBasedNarrowOop);ReservedHeapSpace total_rs0(total_reserved, alignment,use_large_pages, addr);if (addr ! NULL !total_rs0.is_reserved()) {// Failed to reserve at specified address again - give up.addr Universe::preferred_heap_base(total_reserved, alignment, Universe::HeapBasedNarrowOop);assert(addr NULL, );ReservedHeapSpace total_rs1(total_reserved, alignment,use_large_pages, addr);total_rs total_rs1;} else {total_rs total_rs0;}}}if (!total_rs.is_reserved()) {vm_exit_during_initialization(err_msg(Could not reserve enough space for SIZE_FORMAT KB object heap, total_reserved/K));return total_rs;}if (UseCompressedOops) {// Universe::initialize_heap() will reset this to NULL if unscaled// or zero-based narrow oops are actually used.address base (address)(total_rs.base() - os::vm_page_size());Universe::set_narrow_oop_base(base);}// 返回total_rsreturn total_rs;
}17.4.2.2 Universe::preferred_heap_base
char* Universe::preferred_heap_base(size_t heap_size, size_t alignment, NARROW_OOP_MODE mode) {assert(is_size_aligned((size_t)OopEncodingHeapMax, alignment), Must be);assert(is_size_aligned((size_t)UnscaledOopHeapMax, alignment), Must be);assert(is_size_aligned(heap_size, alignment), Must be);// HeapBaseMinAddress 是操作系统明确设定的堆内存的最低地址限制默认设置的是2*G这里按alignment对齐把HeapBaseMinAddress的值按alignment对齐后作为堆内存的最低地址uintx heap_base_min_address_aligned align_size_up(HeapBaseMinAddress, alignment);size_t base 0;
#ifdef _LP64 // 下面是对64位机器及使用压缩指针时的实现我们只讲32位的这块逻辑略过if (UseCompressedOops) {assert(mode UnscaledNarrowOop ||mode ZeroBasedNarrowOop ||mode HeapBasedNarrowOop, mode is invalid);const size_t total_size heap_size heap_base_min_address_aligned;// Return specified base for the first request.if (!FLAG_IS_DEFAULT(HeapBaseMinAddress) (mode UnscaledNarrowOop)) {base heap_base_min_address_aligned;// If the total size is small enough to allow UnscaledNarrowOop then// just use UnscaledNarrowOop.} else if ((total_size OopEncodingHeapMax) (mode ! HeapBasedNarrowOop)) {if ((total_size UnscaledOopHeapMax) (mode UnscaledNarrowOop) (Universe::narrow_oop_shift() 0)) {// Use 32-bits oops without encoding and// place heaps top on the 4Gb boundarybase (UnscaledOopHeapMax - heap_size);} else {// Cant reserve with NarrowOopShift 0Universe::set_narrow_oop_shift(LogMinObjAlignmentInBytes);if (mode UnscaledNarrowOop ||mode ZeroBasedNarrowOop total_size UnscaledOopHeapMax) {// Use zero based compressed oops with encoding and// place heaps top on the 32Gb boundary in case// total_size 4Gb or failed to reserve below 4Gb.uint64_t heap_top OopEncodingHeapMax;// For small heaps, save some space for compressed class pointer// space so it can be decoded with no base.if (UseCompressedClassPointers !UseSharedSpaces OopEncodingHeapMax 32*G) {uint64_t class_space align_size_up(CompressedClassSpaceSize, alignment);assert(is_size_aligned((size_t)OopEncodingHeapMax-class_space,alignment), difference must be aligned too);uint64_t new_top OopEncodingHeapMax-class_space;if (total_size new_top) {heap_top new_top;}}// Align base to the adjusted top of the heapbase heap_top - heap_size;}}} else {// UnscaledNarrowOop encoding didnt work, and no base was found for ZeroBasedOops or// HeapBasedNarrowOop encoding was requested. So, cant reserve below 32Gb.Universe::set_narrow_oop_shift(LogMinObjAlignmentInBytes);}// Set narrow_oop_base and narrow_oop_use_implicit_null_checks// used in ReservedHeapSpace() constructors.// The final values will be set in initialize_heap() below.if ((base ! 0) ((base heap_size) OopEncodingHeapMax)) {// Use zero based compressed oopsUniverse::set_narrow_oop_base(NULL);// Dont need guard page for implicit checks in indexed// addressing mode with zero based Compressed Oops.Universe::set_narrow_oop_use_implicit_null_checks(true);} else {// Set to a non-NULL value so the ReservedSpace ctor computes// the correct no-access prefix.// The final value will be set in initialize_heap() below.Universe::set_narrow_oop_base((address)UnscaledOopHeapMax);
#if defined(_WIN64) || defined(AIX)if (UseLargePages) {// Cannot allocate guard pages for implicit checks in indexed// addressing mode when large pages are specified on windows.Universe::set_narrow_oop_use_implicit_null_checks(false);}
#endif // _WIN64}}
#endifassert(is_ptr_aligned((char*)base, alignment), Must be);// 最终返回base,在32位机器时虚拟机就是返回0return (char*)base; // also return NULL (dont care) for 32-bit VM
}17.4.3 virtualspace.cpp
17.4.3.1 ReservedHeapSpace::ReservedHeapSpace
ReservedHeapSpace::ReservedHeapSpace(size_t size, size_t alignment,bool large, char* requested_address) :/* 先调用父类构造函数*/ReservedSpace(size, alignment, large,requested_address,(UseCompressedOops (Universe::narrow_oop_base() ! NULL) Universe::narrow_oop_use_implicit_null_checks()) ?lcm(os::vm_page_size(), alignment) : 0) {if (base() ! NULL) {MemTracker::record_virtual_memory_type((address)base(), mtJavaHeap);}// Only reserved space for the java heap should have a noaccess_prefix// if using compressed oops.protect_noaccess_prefix(size);
}17.4.3.2 ReservedSpace::ReservedSpace
ReservedSpace::ReservedSpace(size_t size, size_t alignment,bool large,char* requested_address,const size_t noaccess_prefix) {initialize(sizenoaccess_prefix, alignment, large, requested_address,noaccess_prefix, false);
}17.4.3.3 ReservedSpace::initialize
入口函数 ReservedHeapSpace total_rs(total_reserved, alignment, use_large_pages, addr);
参数
total_reserved 对应 size空间大小
alignment 对应 alignment内存对齐值
use_large_pages 对应 large这里不考虑大页就设置为false
addr 对应 requested_address32位时addr为0
noaccess_prefix 为 0
executable 为 false
void ReservedSpace::initialize(size_t size, size_t alignment, bool large,char* requested_address,const size_t noaccess_prefix,bool executable) {// 看源码得知这里就是取page size页大小没什么逻辑const size_t granularity os::vm_allocation_granularity();// 断言检验assert((size (granularity - 1)) 0,size not aligned to os::vm_allocation_granularity());assert((alignment (granularity - 1)) 0,alignment not aligned to os::vm_allocation_granularity());assert(alignment 0 || is_power_of_2((intptr_t)alignment),not a power of 2);// 取二者最大值对齐alignment MAX2(alignment, (size_t)os::vm_page_size());// Assert that if noaccess_prefix is used, it is the same as alignment.assert(noaccess_prefix 0 ||noaccess_prefix alignment, noaccess prefix wrong);_base NULL;_size 0;_special false;_executable executable;_alignment 0;_noaccess_prefix 0;if (size 0) {return;}// 不存在大页special 为 falsebool special large !os::can_commit_large_page_memory();char* base NULL;// 32位机器时 requested_address 0这条线也不会走if (requested_address ! 0) {requested_address - noaccess_prefix; // adjust requested addressassert(requested_address ! NULL, huge noaccess prefix?);}// special为false,这个if不会走if (special) {base os::reserve_memory_special(size, alignment, requested_address, executable);if (base ! NULL) {if (failed_to_reserve_as_requested(base, requested_address, size, true)) {// OS ignored requested address. Try different address.return;}// Check alignment constraints.assert((uintptr_t) base % alignment 0,err_msg(Large pages returned a non-aligned address, base: PTR_FORMAT alignment: PTR_FORMAT,base, (void*)(uintptr_t)alignment));_special true;} else {// failed; try to reserve regular memory belowif (UseLargePages (!FLAG_IS_DEFAULT(UseLargePages) ||!FLAG_IS_DEFAULT(LargePageSizeInBytes))) {if (PrintCompressedOopsMode) {tty-cr();tty-print_cr(Reserve regular memory without large pages.);}}}}if (base NULL) {if (requested_address ! 0) {base os::attempt_reserve_memory_at(size, requested_address);if (failed_to_reserve_as_requested(base, requested_address, size, false)) {// OS ignored requested address. Try different address.base NULL;}} else {// 这一步就是通过系统调用mmap映射一块size大小的内存Java堆内存就是mmap映射出来的base os::reserve_memory(size, NULL, alignment);}// 映射失败直接退出函数分配Java堆内存失败if (base NULL) return;// 验证对齐为啥要验证呢因为base是mmap映射后返回的内存首地址这个地址是os自己的规则选取的一个地址不一定能按照alignment对齐所以这一定要验证if ((((size_t)base noaccess_prefix) (alignment - 1)) ! 0) {// base没有对齐只能释放刚才mmap映射的内存然后重试if (!os::release_memory(base, size)) fatal(os::release_memory failed);// 确保对齐size align_size_up(size, alignment);// 再次mmap映射内存返回的base同样有上面一样的不对齐问题所以这个函数中包含了手动对齐操作细节看章节17.4.3.4base os::reserve_memory_aligned(size, alignment);if (requested_address ! 0 failed_to_reserve_as_requested(base, requested_address, size, false)) {// As a result of the alignment constraints, the allocated base differs// from the requested address. Return back to the caller who can// take remedial action (like try again without a requested address).assert(_base NULL, should be);return;}}}// Done_base base; // 最终拿到了Java堆的首地址_size size; // 最终拿到了Java堆的大小_alignment alignment; // 对齐值_noaccess_prefix noaccess_prefix; // 0// 断言判断assert(noaccess_prefix 0 ||noaccess_prefix _alignment, noaccess prefix wrong);assert(markOopDesc::encode_pointer_as_mark(_base)-decode_pointer() _base,area must be distinguisable from marks for mark-sweep);assert(markOopDesc::encode_pointer_as_mark(_base[size])-decode_pointer() _base[size],area must be distinguisable from marks for mark-sweep);
}17.4.3.4 os_posix.cpp-os::reserve_memory_aligned
char* os::reserve_memory_aligned(size_t size, size_t alignment) {assert((alignment (os::vm_allocation_granularity() - 1)) 0,Alignment must be a multiple of allocation granularity (page size));assert((size (alignment -1)) 0, size must be alignment aligned);size_t extra_size size alignment;assert(extra_size size, overflow, size is too large to allow alignment);// mmap映射一块内存区域返回首地址char* extra_base os::reserve_memory(extra_size, NULL, alignment);if (extra_base NULL) {return NULL;}// 手动对齐char* aligned_base (char*) align_size_up((uintptr_t) extra_base, alignment);// [ | | ]// ^ extra_base// ^ extra_base begin_offset aligned_base// extra_base begin_offset size ^// extra_base extra_size ^// || begin_offset// end_offset ||// 用对齐后的地址-mmap的首地址得出与首地址的偏移值size_t begin_offset aligned_base - extra_base;// 结束地址对齐后的偏移size_t end_offset (extra_base extra_size) - (aligned_base size);// begin_offset 0表示确实有偏移那就把extra_base到偏移的这部分释放掉因为有新的首地址了if (begin_offset 0) {os::release_memory(extra_base, begin_offset);}// end_offset 0表示确实有偏移那就把end_offset偏移的这部分释放掉因为有新的限制地址了if (end_offset 0) {os::release_memory(extra_base begin_offset size, end_offset);}// 返回首地址return aligned_base;
}
