chm的put涉及的一些逻辑代码注释解析。
final V putVal(K key, V value, boolean onlyIfAbsent) {
if (key == null || value == null) throw new NullPointerException();
//(h ^ (h >>> 16)) & HASH_BITS,HASH_BITS=0x7fffffff,即chm对于hash的计算除了与hashmap类似的hashcode抑或hashcode无符号高16位之外,
//&hash_bits结果就是计算出的hash值是非负数
int hash = spread(key.hashCode());
int binCount = 0;
//自旋重试
for (Node<K,V>[] tab = table;;) {
Node<K,V> f; int n, i, fh;
//在数组未初始化时,承担初始化工作
if (tab == null || (n = tab.length) == 0)
tab = initTable();
//如果定位到的槽上没有节点,则尝试通过cas将该key value的节点赋值到槽上
else if ((f = tabAt(tab, i = (n - 1) & hash)) == null) {
//确保只有一个线程cas成功
if (casTabAt(tab, i, null,
new Node<K,V>(hash, key, value, null)))
break; // no lock when adding to empty bin
}
else if ((fh = f.hash) == MOVED)//当前chm正处于扩容阶段,注意在这对fh赋值了tab数组第i槽的首节点f的hash值
//进入协助扩容方法
tab = helpTransfer(tab, f);
else {
V oldVal = null;
//通过锁,锁定槽的第一个元素
synchronized (f) {
if (tabAt(tab, i) == f) {
//一般情况下,如果是链表状态下节点的hash值都是非负数,
if (fh >= 0) {
//在链表状态下,作为槽上节点的计数变量
binCount = 1;
for (Node<K,V> e = f;; ++binCount) {
K ek;
//最简单的情况,便是这种key相同的情况,只需要覆盖原来的value即可
if (e.hash == hash &&
((ek = e.key) == key ||
(ek != null && key.equals(ek)))) {
oldVal = e.val;
if (!onlyIfAbsent)
e.val = value;
break;
}
//走到这说明可能不是值的覆盖,那就考虑如果到了链表的尾节点,则插入到链表尾部
Node<K,V> pred = e;
if ((e = e.next) == null) {
pred.next = new Node<K,V>(hash, key,
value, null);
break;
}
}
}
//如果是红黑树类型
else if (f instanceof TreeBin) {
Node<K,V> p;
binCount = 2;
//调用红黑树方法插入节点
if ((p = ((TreeBin<K,V>)f).putTreeVal(hash, key,
value)) != null) {
oldVal = p.val;
if (!onlyIfAbsent)
p.val = value;
}
}
}
}
//bincount!=0说明是进行了操作,
if (binCount != 0) {
//判断是否需要将链表改为红黑树结构
if (binCount >= TREEIFY_THRESHOLD)
//要注意的一点是除了单个槽上的链表节点数量大于等于8之外,还需要tab数组的长度不小于64才能变成红黑树,否则会先尝试扩容
treeifyBin(tab, i);
if (oldVal != null)
return oldVal;
break;
}
}
}
//chm节点数量增加
addCount(1L, binCount);
return null;
}
helpTransfer是在put操作时,遇到了处于moved状态节点,则感知到正在进行扩容时的线程操作,即进入协助库容逻辑
final Node<K,V>[] helpTransfer(Node<K,V>[] tab, Node<K,V> f) {
Node<K,V>[] nextTab; int sc;
//此时,chm处于扩容期间,尝试获取一段扩容任务,协助扩容
if (tab != null && (f instanceof ForwardingNode) &&
(nextTab = ((ForwardingNode<K,V>)f).nextTable) != null) {
int rs = resizeStamp(tab.length);
while (nextTab == nextTable && table == tab &&
(sc = sizeCtl) < 0) {
if ((sc >>> RESIZE_STAMP_SHIFT) != rs || sc == rs + 1 ||
sc == rs + MAX_RESIZERS || transferIndex <= 0)
break;
if (U.compareAndSwapInt(this, SIZECTL, sc, sc + 1)) {
transfer(tab, nextTab);
break;
}
}
return nextTab;
}
return table;
}
addCount便是对于chm元素数量的统计 和扩容逻辑的 处理函数
private final void addCount(long x, int check) {
CounterCell[] as; long b, s;
//初始情况下,counterCells未被初始化,即as为null,则尝试通过cas 修改basecount的值为原来的值+1
//这种情况下,也可能会在线程竞争激烈时出现失败的情况,则进入if内部
if ((as = counterCells) != null ||
!U.compareAndSwapLong(this, BASECOUNT, b = baseCount, s = b + x)) {
CounterCell a; long v; int m;
boolean uncontended = true;
//初始时as为null,则直接调用fulladdcount
if (as == null || (m = as.length - 1) < 0 ||
(a = as[ThreadLocalRandom.getProbe() & m]) == null ||
!(uncontended =
U.compareAndSwapLong(a, CELLVALUE, v = a.value, v + x))) {
//如果是countcell数组为初始化或者随机定位的countcell数组上槽对应元素未初始化或cas countcell值失败了,都走fulladdcount方法
//uncontended是是否进行过cas countcell数组元素值的操作标志(只有cas过且失败,才会走到这,否则都是true)
fullAddCount(x, uncontended);
return;
}
if (check <= 1)
return;
//统计元素个数
s = sumCount();
}
//只有check为非负数的情况下,才会检查是否需要扩容
if (check >= 0) {
Node<K,V>[] tab, nt; int n, sc;
while (s >= (long)(sc = sizeCtl) && (tab = table) != null &&
(n = tab.length) < MAXIMUM_CAPACITY) {
//扩容戳
int rs = resizeStamp(n);
//其他非首个进入扩容的线程,则是sc<0,
if (sc < 0) {
if ((sc >>> RESIZE_STAMP_SHIFT) != rs || sc == rs + 1 ||
sc == rs + MAX_RESIZERS || (nt = nextTable) == null ||
transferIndex <= 0)
break;
//其他线程都是+1,再进入transfer,并带入新的tab,即nt
//此时可以将SIZECTL理解为一个既保存了是多少长度的tab扩容,又记录了参与扩容的线程数的一个值
//以便于在之后的线程处理完扩容后,最后一个线程进行收尾工作,
if (U.compareAndSwapInt(this, SIZECTL, sc, sc + 1))
transfer(tab, nt);
}
//rs=Integer.numberOfLeadingZeros(n) | (1 << (RESIZE_STAMP_BITS - 1))就是说rs的低16位 第一位是1
//那么rs << RESIZE_STAMP_SHIFT,RESIZE_STAMP_SHIFT=16,也就是rs<<16 那么rs变成了负数,
//这个负数的高16位是rs的低16位,然后cas +2,
else if (U.compareAndSwapInt(this, SIZECTL, sc,
(rs << RESIZE_STAMP_SHIFT) + 2))
//首次扩容线程
transfer(tab, null);
s = sumCount();
}
}
}
fullAddCount函数,便是在并发竞争大时的最终元素数量累加处理函数,该函数会在
- CounterCell数组已初始化,
a.根据随机数定位到的CounterCell未初始化时初始化,并设置累加值
b. 直接cas已经初始化的CounterCell
c.CounterCell数组扩容 - CounterCell数组未初始化,初始化数组,并根据随机数定位到的CounterCell初始化,并设置值
- 进行basecount的cas累加
函数大致是这种逻辑思路,细节上有出入
private final void fullAddCount(long x, boolean wasUncontended) {
int h;
if ((h = ThreadLocalRandom.getProbe()) == 0) {
ThreadLocalRandom.localInit(); // force initialization
h = ThreadLocalRandom.getProbe();
wasUncontended = true;
}
boolean collide = false; // True if last slot nonempty
for (;;) {
CounterCell[] as; CounterCell a; int n; long v;
//如果countcells已经初始化了,
if ((as = counterCells) != null && (n = as.length) > 0) {
//如果随机数h取模得到的countcells数组位置上的对象还未初始化,
if ((a = as[(n - 1) & h]) == null) {
//探测cellsbusy==0即countcells数组未被占用
if (cellsBusy == 0) { // Try to attach new Cell
CounterCell r = new CounterCell(x); // Optimistic create
//尝试通过cellsbusy的值,获取到对countcells数组的操作权
if (cellsBusy == 0 &&
U.compareAndSwapInt(this, CELLSBUSY, 0, 1)) {
boolean created = false;
try { // Recheck under lock
CounterCell[] rs; int m, j;
//拿到操作权后,也只有在数组已经初始化,且h取模定位到的元素为空的情况下赋值
if ((rs = counterCells) != null &&
(m = rs.length) > 0 &&
rs[j = (m - 1) & h] == null) {
rs[j] = r;
//只有成功覆盖了原为null的countcell,才会设置created=true
created = true;
}
} finally {
cellsBusy = 0;
}
//只有created=true才说明上面覆盖成功,否则可能失败,继续自旋重试
if (created)
break;
continue; // Slot is now non-empty
}
}
collide = false;
}
//wasUncontended是 在调用fullAddCount函数时,是否进行了cas操作失败的标志,为false说明尝试过cas但失败了
else if (!wasUncontended) // CAS already known to fail
wasUncontended = true; // Continue after rehash
//尝试cas cellval的值,累加x
else if (U.compareAndSwapLong(a, CELLVALUE, v = a.value, v + x))
break;
//如果counterCells != as说明已经被其他线程修改过了,可以通过自旋重试
//n>=ncpu,即数组长度已经大于等于cpu个数了,这时也没必必要再进行下面的扩容,因而选择重试
else if (counterCells != as || n >= NCPU)
collide = false; // At max size or stale
//如果第二次走到这,则选择走下面的扩容
else if (!collide)
collide = true;
//首先获取操作countercells的权限,即cellsbusy=0且cas成1成功
else if (cellsBusy == 0 &&
U.compareAndSwapInt(this, CELLSBUSY, 0, 1)) {
try {
//说明未被其他线程修改
if (counterCells == as) {// Expand table unless stale
//扩容一倍
CounterCell[] rs = new CounterCell[n << 1];
//原有对象复制
for (int i = 0; i < n; ++i)
rs[i] = as[i];
counterCells = rs;
}
} finally {
cellsBusy = 0;
}
collide = false;
continue; // Retry with expanded table
}
h = ThreadLocalRandom.advanceProbe(h);
}
//初始时,是走这,通过cellbusy和cas cellsbusy的值获得扩容权限,
else if (cellsBusy == 0 && counterCells == as &&
U.compareAndSwapInt(this, CELLSBUSY, 0, 1)) {
boolean init = false;
try { // Initialize table
if (counterCells == as) {
//初始化countercell数组
CounterCell[] rs = new CounterCell[2];
//将随机数h取模定位到的统计节点初始化对象,并赋值初始值x
rs[h & 1] = new CounterCell(x);
counterCells = rs;
init = true;
}
} finally {
cellsBusy = 0;
}
if (init)
break;
}
//尝试cas basecount的值,如果成功,则返回
else if (U.compareAndSwapLong(this, BASECOUNT, v = baseCount, v + x))
break; // Fall back on using base
}
}
chm将对于元素记录的个数记录,首先时基于基本的一个int类型basecount,在并发竞争大时,会引入CounterCell数组协助记录个数,最终的总数是两者合并的结果
//计算chm的元素个数为basecount+countcell数组上所有非空元素记录值的和
final long sumCount() {
CounterCell[] as = counterCells; CounterCell a;
long sum = baseCount;
if (as != null) {
for (int i = 0; i < as.length; ++i) {
if ((a = as[i]) != null)
sum += a.value;
}
}
return sum;
}
chm的扩容,将扩容拆分为一个个的扩容任务,每个任务最小跨度16,线程进入先获取任务,再进行扩容, 此时如果有线程进入put时,发现正在进行扩容,也会进行协助扩容,最终都是走下面的transfer方法
private final void transfer(Node<K,V>[] tab, Node<K,V>[] nextTab) {
int n = tab.length, stride;
//stride每次线程扩容的任务步长,单核状态下为n,多核状态下为n/8/cpu核数,
//计算结果与16取较大值,即步长最低为16
if ((stride = (NCPU > 1) ? (n >>> 3) / NCPU : n) < MIN_TRANSFER_STRIDE)
stride = MIN_TRANSFER_STRIDE; // subdivide range
//只有开启扩容的线程,进入方法时,nextab才是null,也就是需要新建新容器的标志
if (nextTab == null) { // initiating
try {
@SuppressWarnings("unchecked")
Node<K,V>[] nt = (Node<K,V>[])new Node<?,?>[n << 1];
nextTab = nt;
} catch (Throwable ex) { // try to cope with OOME
sizeCtl = Integer.MAX_VALUE;
return;
}
//chm中扩容时,新table的记录位置
nextTable = nextTab;
//记录要迁移的数组大小
transferIndex = n;
}
int nextn = nextTab.length;
//在扩容时,作为占位在无数据存放的槽的节点上
ForwardingNode<K,V> fwd = new ForwardingNode<K,V>(nextTab);
boolean advance = true;
boolean finishing = false; // to ensure sweep before committing nextTab
for (int i = 0, bound = 0;;) {
Node<K,V> f; int fh;
//自旋获取本线程可以扩容的任务区间
//i为扩容的原数组索引任务的索引最大值,
//例如要扩容32位长度数组,该线程是第一个进入扩容的,
//则拿到的i为32-1=31,nextbound=32-16=16
//第二个拿到的任务为i=15,nextbound=0
while (advance) {
int nextIndex, nextBound;
//--i首先是可以作为i的递减使用
//--i>=bound可以作为边界的判断,finishing作为扩容结束的标志,既然扩容结束了便不需要进行扩容任务的抢占了
if (--i >= bound || finishing)
advance = false;
//nextIndex的赋值,
else if ((nextIndex = transferIndex) <= 0) {
i = -1;
advance = false;
}
//进行cas抢占任务,chm将扩容任务拆分为段,每一段任务最小跨度16位长度,线程进入抢占任务后处理完毕,可以继续参与抢占任务,
else if (U.compareAndSwapInt
(this, TRANSFERINDEX, nextIndex,
nextBound = (nextIndex > stride ?
nextIndex - stride : 0))) {
bound = nextBound;
i = nextIndex - 1;
advance = false;
}
}
if (i < 0 || i >= n || i + n >= nextn) {
int sc;
//下面的cas会在最后一个离开的线程中执行finishing =true
//然后在该线程的下次自旋中进行table的赋值和sizeCtl 的赋值,完成最终扩容
if (finishing) {
nextTable = null;
table = nextTab;
sizeCtl = (n << 1) - (n >>> 1);
return;
}
//相较于上一个 if (finishing) 判断,这个才是需要先执行的,
//因为addCount函数中第一个进入扩容的线程是+2,其他每一个进入的线程都是+1
//这里每个扩容完毕自己分配的任务后-1,
//所以最后一个扩容离开的线程肯定是sc-2 ==(resizeStamp(n) << RESIZE_STAMP_SHIFT)
if (U.compareAndSwapInt(this, SIZECTL, sc = sizeCtl, sc - 1)) {
if ((sc - 2) != resizeStamp(n) << RESIZE_STAMP_SHIFT)
return;
//最后一个扩容离开的线程才需要赋值finishing =true
finishing = advance = true;
i = n; // recheck before commit
}
}
//在槽为空时放入fwd占位,如果此时有线程进行put操作,会感知,并先进入扩容流程
else if ((f = tabAt(tab, i)) == null)
advance = casTabAt(tab, i, null, fwd);
//这里便是在最终扩容完毕前进行的检查工作,扩容完毕前会再次检查原tab上的每个元素,如果都是moved则扩容完毕,
//因为有可能存在某个修改了tab[i]上第一个节点的情况,毕竟此时tab也还在使用并且tab[i]还未加锁锁定
else if ((fh = f.hash) == MOVED)
advance = true; // already processed
else {
//拿到锁
synchronized (f) {
//有可能之前拿到的tab[i]的第一个节点f,已经被其他线程修改了,不是第一个线程了,
//所以再次进行自旋,这时才体现出最后一个线程recheck的必要性,因为自选后i变了
if (tabAt(tab, i) == f) {
Node<K,V> ln, hn;
//通过节点hash值判断是否为链表结构
if (fh >= 0) {
int runBit = fh & n;
Node<K,V> lastRun = f;
//找到最后一段连续的高位或低位节点
for (Node<K,V> p = f.next; p != null; p = p.next) {
//p.hash & n如果为1,那么就是高位,需要改变位置到i+n的位置
//例如在数组长度16中,节点hashcode二进制为10110==22,那么原本的位置是i=22&15==6
//扩容后是22&31 是取低5位。这时影响他位置的就是倒数第五位上是不是1,
//如果是1,那么hash&(1<<4)!=0,而1<<4==16也就是这里的n
int b = p.hash & n;
if (b != runBit) {
runBit = b;
lastRun = p;
}
}
//如果节点为高位或低位,则先分别赋值
if (runBit == 0) {
ln = lastRun;
hn = null;
}
else {
hn = lastRun;
ln = null;
}
//枚举节点,直到p == lastRun为止,因为最后这一段已经添加完毕
for (Node<K,V> p = f; p != lastRun; p = p.next) {
int ph = p.hash; K pk = p.key; V pv = p.val;
if ((ph & n) == 0)
ln = new Node<K,V>(ph, pk, pv, ln);
else
hn = new Node<K,V>(ph, pk, pv, hn);
}
//将高位和低位分别赋值到新的数组槽上
setTabAt(nextTab, i, ln);
setTabAt(nextTab, i + n, hn);
//扩容完成的标志,将fwd cas到原数组槽上
setTabAt(tab, i, fwd);
advance = true;
}
else if (f instanceof TreeBin) {
TreeBin<K,V> t = (TreeBin<K,V>)f;
TreeNode<K,V> lo = null, loTail = null;
TreeNode<K,V> hi = null, hiTail = null;
int lc = 0, hc = 0;
for (Node<K,V> e = t.first; e != null; e = e.next) {
int h = e.hash;
TreeNode<K,V> p = new TreeNode<K,V>
(h, e.key, e.val, null, null);
if ((h & n) == 0) {
if ((p.prev = loTail) == null)
lo = p;
else
loTail.next = p;
loTail = p;
++lc;
}
else {
if ((p.prev = hiTail) == null)
hi = p;
else
hiTail.next = p;
hiTail = p;
++hc;
}
}
ln = (lc <= UNTREEIFY_THRESHOLD) ? untreeify(lo) :
(hc != 0) ? new TreeBin<K,V>(lo) : t;
hn = (hc <= UNTREEIFY_THRESHOLD) ? untreeify(hi) :
(lc != 0) ? new TreeBin<K,V>(hi) : t;
setTabAt(nextTab, i, ln);
setTabAt(nextTab, i + n, hn);
setTabAt(tab, i, fwd);
advance = true;
}
}
}
}
}
}
以上为个人对于源码的理解和注释,如有错误或疑问,欢迎一起讨论,谢谢!