netty内存规格
netty内存分配器类图如下
我们直接看
io.netty.buffer.PooledByteBufAllocator#newDirectBuffer
@Override
protected ByteBuf newDirectBuffer(int initialCapacity, int maxCapacity) {
PoolThreadCache cache = threadCache.get();
PoolArena directArena = cache.directArena;
final ByteBuf buf;
if (directArena != null) {
buf = directArena.allocate(cache, initialCapacity, maxCapacity);
} else {
buf = PlatformDependent.hasUnsafe() ?
UnsafeByteBufUtil.newUnsafeDirectByteBuf(this, initialCapacity, maxCapacity) :
new UnpooledDirectByteBuf(this, initialCapacity, maxCapacity);
}
return toLeakAwareBuffer(buf);
}
内存分配策略
1.先PoolThreadLocalCache获取欲线程绑定的缓存池PoolThreadCache
2.如果不存在,则获取Arean数组中的PoolArena,然后从中创建一个新的PoolThreadCache作为缓存池使用
具体分配代码如下
private void allocate(PoolThreadCache cache, PooledByteBuf buf, final int reqCapacity) {
final int normCapacity = normalizeCapacity(reqCapacity);
if (isTinyOrSmall(normCapacity)) { // capacity < pageSize
int tableIdx;
PoolSubpage[] table;
boolean tiny = isTiny(normCapacity);
if (tiny) { // < 512
if (cache.allocateTiny(this, buf, reqCapacity, normCapacity)) {
// was able to allocate out of the cache so move on
return;
}
tableIdx = tinyIdx(normCapacity);
table = tinySubpagePools;
} else {
if (cache.allocateSmall(this, buf, reqCapacity, normCapacity)) {
// was able to allocate out of the cache so move on
return;
}
tableIdx = smallIdx(normCapacity);
table = smallSubpagePools;
}
final PoolSubpage head = table[tableIdx];
/**
* Synchronize on the head. This is needed as {@link PoolChunk#allocateSubpage(int)} and
* {@link PoolChunk#free(long)} may modify the doubly linked list as well.
*/
synchronized (head) {
final PoolSubpage s = head.next;
if (s != head) {
assert s.doNotDestroy && s.elemSize == normCapacity;
long handle = s.allocate();
assert handle >= 0;
s.chunk.initBufWithSubpage(buf, handle, reqCapacity);
incTinySmallAllocation(tiny);
return;
}
}
synchronized (this) {
allocateNormal(buf, reqCapacity, normCapacity);
}
incTinySmallAllocation(tiny);
return;
}
if (normCapacity <= chunkSize) {
if (cache.allocateNormal(this, buf, reqCapacity, normCapacity)) {
// was able to allocate out of the cache so move on
return;
}
synchronized (this) {
allocateNormal(buf, reqCapacity, normCapacity);
++allocationsNormal;
}
} else {
// Huge allocations are never served via the cache so just call allocateHuge
allocateHuge(buf, reqCapacity);
}
}
小结如下
系统会根据需要申请的内存规格选择合适的MemoryRegionCache
集合上图我们可以知道MemoryRegionCache其实包含了PoolChunk我们最终也是通过调用PoolChunk的allocate()方法类进行正在的分配内存。
而PoolChunk则是通过二叉树来记录每个PoolSubpage的分配情况.
其中memoryMap存放的是PoolSubPage的分配信息,depthMap存放的是二叉树的深度。
depthMap初始化后就不会再变化,但memoryMap会随着PoolSubPage的变化而变化。
初始化时memoryMap与depthMap的取值相同。
节点分配情况有以下三种情况
(1)memroyMap[i]=depthMap[id]:表示当前节点可以分配内存
(2)memroyMap[i]>depthMap[id]:表示当前节点至少有一个子节点已经被分配,无法分配满足该深度的内存,但可以分配更小一些的内存(通过空闲的子节点)
(3)memroyMap[i]=最大深度+1:表示当前节点下的所有子节点都已经分配完了,没有课用内存
简要概述
分配内存,会从根节点往下找,找到合适的之后就会标记节点被占用.
标记完之后,依次向上更新父节点,直到根节点.将父节点的memoryMap[id]位置信息修改为两个子节点中的较小值.
获取到节点后,根据节点id计算并获取subpageIdx,通过subpageIdx从PoolSubpage
如果PoolSubpage为空,则新建,并将其加入到数组中,便于参与后续的内存分配.不为空,说明是使用完之后被放入到内存池了,重新初始化后将其加入到内存池的双向链表中,参与内存分配.
然后调用PoolSubpage的allocate方法，返回代表PoolSubpage块存储区域的占用情况(long类型,高32位表示subPage中分配的位置,低32位表示二叉树中分配的节点)
最终PoolAreana根据上述信息调用PoolChunk的initBuf方法完成PoolByteBuf的初始化.
查找到合适的MemoryReginCache后调用PoolChunk的allocate方法
,具体核心方法如下
(1)查找PoolChunk所在的PoolArena的PooSubpage头结点,因为在内存分配中会修改PoolSubpage链表,考虑到并发安全,对头结点进行了加锁
private long allocateSubpage(int normCapacity) {
// Obtain the head of the PoolSubPage pool that is owned by the PoolArena and synchronize on it.
// This is need as we may add it back and so alter the linked-list structure.
PoolSubpage head = arena.findSubpagePoolHead(normCapacity);
synchronized (head) {
}
(2)从二叉树中查找合适的节点,获取到节点ID
private int allocateNode(int d) {
int id = 1;
int initial = - (1 << d); // has last d bits = 0 and rest all = 1
byte val = value(id);
if (val > d) { // unusable
return -1;
}
while (val < d || (id & initial) == 0) { // id & initial == 1 << d for all ids at depth d, for < d it is 0
id <<= 1;
val = value(id);
if (val > d) {
id ^= 1;
val = value(id);
}
}
byte value = value(id);
assert value == d && (id & initial) == 1 << d : String.format("val = %d, id & initial = %d, d = %d",
value, id & initial, d);
setValue(id, unusable); // mark as unusable
updateParentsAlloc(id);
return id;
}
不断向下判断,当子节点满足分配条件时,深度加1,进入子节点,先判断左节点是否满足分配要求,如果不满足则寻找同层级的右节点,直到找到符合要求的节点,获取节点ID
(3)标记已分配的节点为不可用,并向上更新父节点的分配信息
private int allocateNode(int d) {
.....
updateParentsAlloc(id);
return id;
}
(4)根据nodeID获取对应的PoolSubPage
private long allocateSubpage(int normCapacity) {
....
int subpageIdx = subpageIdx(id);
PoolSubpage subpage = subpages[subpageIdx];
....
}
}
(5)判断PoolSubpage是新创建的还是被释放重用的,如果是新建的,创建完成之后加入到PoolSubpage
private long allocateSubpage(int normCapacity) {
.....
if (subpage == null) {
subpage = new PoolSubpage(head, this, id, runOffset(id), pageSize, normCapacity);
subpages[subpageIdx] = subpage;
} else {
subpage.init(head, normCapacity);
}
return subpage.allocate();
}
}
(6)调用PoolSubpage的allocate方法,返回PoolSubpage分配情况的位图索引
long allocate() {
if (elemSize == 0) {
return toHandle(0);
}
if (numAvail == 0 || !doNotDestroy) {
return -1;
}
final int bitmapIdx = getNextAvail();
int q = bitmapIdx >>> 6;
int r = bitmapIdx & 63;
assert (bitmap[q] >>> r & 1) == 0;
bitmap[q] |= 1L << r;
if (-- numAvail == 0) {
removeFromPool();
}
return toHandle(bitmapIdx);
}
(7)最后根据位图索引,需要申请的内存容量等参数交由PoolChunk进行内存分配
void initBuf(PooledByteBuf buf, long handle, int reqCapacity) {
int memoryMapIdx = memoryMapIdx(handle);
int bitmapIdx = bitmapIdx(handle);
if (bitmapIdx == 0) {
byte val = value(memoryMapIdx);
assert val == unusable : String.valueOf(val);
buf.init(this, handle, runOffset(memoryMapIdx) + offset, reqCapacity, runLength(memoryMapIdx),
arena.parent.threadCache());
} else {
initBufWithSubpage(buf, handle, bitmapIdx, reqCapacity);
}
}