代码拉取完成,页面将自动刷新
同步操作将从 src-openEuler/openjdk-1.8.0 强制同步,此操作会覆盖自 Fork 仓库以来所做的任何修改,且无法恢复!!!
确定后同步将在后台操作,完成时将刷新页面,请耐心等待。
From 20a2cb586843089c328041057057187f37ffe56f Mon Sep 17 00:00:00 2001
Date: Fri, 22 Jan 2021 14:31:49 +0800
Subject: Backport of JDK-8166197.
Summary: assert(RelaxAssert || w != Thread::current()->_MutexEvent) failed: invariant
LLT: hotspot/test/stress/gc/TestStressRSetCoarsening.java
Bug url: https://bugs.openjdk.java.net/browse/JDK-8166197
---
hotspot/src/share/vm/runtime/mutex.cpp | 31 +++++++++++++++-----------
1 file changed, 18 insertions(+), 13 deletions(-)
diff --git a/hotspot/src/share/vm/runtime/mutex.cpp b/hotspot/src/share/vm/runtime/mutex.cpp
index 84207eae0..646f7f152 100644
--- a/hotspot/src/share/vm/runtime/mutex.cpp
+++ b/hotspot/src/share/vm/runtime/mutex.cpp
@@ -465,7 +465,7 @@ void Monitor::ILock (Thread * Self) {
ParkEvent * const ESelf = Self->_MutexEvent ;
assert (_OnDeck != ESelf, "invariant") ;
- // As an optimization, spinners could conditionally try to set ONDECK to _LBIT
+ // As an optimization, spinners could conditionally try to set _OnDeck to _LBIT
// Synchronizer.cpp uses a similar optimization.
if (TrySpin (Self)) goto Exeunt ;
@@ -476,7 +476,7 @@ void Monitor::ILock (Thread * Self) {
OrderAccess::fence() ;
// Optional optimization ... try barging on the inner lock
- if ((NativeMonitorFlags & 32) && CASPTR (&_OnDeck, NULL, UNS(Self)) == 0) {
+ if ((NativeMonitorFlags & 32) && CASPTR (&_OnDeck, NULL, UNS(ESelf)) == 0) {
goto OnDeck_LOOP ;
}
@@ -484,14 +484,14 @@ void Monitor::ILock (Thread * Self) {
// At any given time there is at most one ondeck thread.
// ondeck implies not resident on cxq and not resident on EntryList
- // Only the OnDeck thread can try to acquire -- contended for -- the lock.
+ // Only the OnDeck thread can try to acquire -- contend for -- the lock.
// CONSIDER: use Self->OnDeck instead of m->OnDeck.
// Deschedule Self so that others may run.
- while (_OnDeck != ESelf) {
+ while (OrderAccess::load_ptr_acquire(&_OnDeck) != ESelf) {
ParkCommon (ESelf, 0) ;
}
- // Self is now in the ONDECK position and will remain so until it
+ // Self is now in the OnDeck position and will remain so until it
// manages to acquire the lock.
OnDeck_LOOP:
for (;;) {
@@ -515,8 +515,8 @@ void Monitor::ILock (Thread * Self) {
// A. Shift or defer dropping the inner lock until the subsequent IUnlock() operation.
// This might avoid potential reacquisition of the inner lock in IUlock().
// B. While still holding the inner lock, attempt to opportunistically select
- // and unlink the next ONDECK thread from the EntryList.
- // If successful, set ONDECK to refer to that thread, otherwise clear ONDECK.
+ // and unlink the next OnDeck thread from the EntryList.
+ // If successful, set OnDeck to refer to that thread, otherwise clear OnDeck.
// It's critical that the select-and-unlink operation run in constant-time as
// it executes when holding the outer lock and may artificially increase the
// effective length of the critical section.
@@ -543,7 +543,7 @@ void Monitor::IUnlock (bool RelaxAssert) {
OrderAccess::release_store(&_LockWord.Bytes[_LSBINDEX], 0); // drop outer lock
OrderAccess::storeload ();
- ParkEvent * const w = _OnDeck ;
+ ParkEvent * const w = _OnDeck ; // raw load as we will just return if non-NULL
assert (RelaxAssert || w != Thread::current()->_MutexEvent, "invariant") ;
if (w != NULL) {
// Either we have a valid ondeck thread or ondeck is transiently "locked"
@@ -551,7 +551,7 @@ void Monitor::IUnlock (bool RelaxAssert) {
// OnDeck allows us to discriminate two cases. If the latter, the
// responsibility for progress and succession lies with that other thread.
// For good performance, we also depend on the fact that redundant unpark()
- // operations are cheap. That is, repeated Unpark()ing of the ONDECK thread
+ // operations are cheap. That is, repeated Unpark()ing of the OnDeck thread
// is inexpensive. This approach provides implicit futile wakeup throttling.
// Note that the referent "w" might be stale with respect to the lock.
// In that case the following unpark() is harmless and the worst that'll happen
@@ -600,8 +600,13 @@ void Monitor::IUnlock (bool RelaxAssert) {
_EntryList = w->ListNext ;
// as a diagnostic measure consider setting w->_ListNext = BAD
assert (UNS(_OnDeck) == _LBIT, "invariant") ;
- _OnDeck = w ; // pass OnDeck to w.
- // w will clear OnDeck once it acquires the outer lock
+
+ // Pass OnDeck role to w, ensuring that _EntryList has been set first.
+ // w will clear _OnDeck once it acquires the outer lock.
+ // Note that once we set _OnDeck that thread can acquire the mutex, proceed
+ // with its critical section and then enter this code to unlock the mutex. So
+ // you can have multiple threads active in IUnlock at the same time.
+ OrderAccess::release_store_ptr(&_OnDeck, w);
// Another optional optimization ...
// For heavily contended locks it's not uncommon that some other
@@ -849,7 +854,7 @@ int Monitor::IWait (Thread * Self, jlong timo) {
// ESelf is now on the cxq, EntryList or at the OnDeck position.
// The following fragment is extracted from Monitor::ILock()
for (;;) {
- if (_OnDeck == ESelf && TrySpin(Self)) break ;
+ if (OrderAccess::load_ptr_acquire(&_OnDeck) == ESelf && TrySpin(Self)) break ;
ParkCommon (ESelf, 0) ;
}
assert (_OnDeck == ESelf, "invariant") ;
@@ -1060,7 +1065,7 @@ void Monitor::jvm_raw_lock() {
// Only the OnDeck thread can try to acquire -- contended for -- the lock.
// CONSIDER: use Self->OnDeck instead of m->OnDeck.
for (;;) {
- if (_OnDeck == ESelf && TrySpin(NULL)) break ;
+ if (OrderAccess::load_ptr_acquire(&_OnDeck) == ESelf && TrySpin(NULL)) break ;
ParkCommon (ESelf, 0) ;
}
--
2.19.0
此处可能存在不合适展示的内容,页面不予展示。您可通过相关编辑功能自查并修改。
如您确认内容无涉及 不当用语 / 纯广告导流 / 暴力 / 低俗色情 / 侵权 / 盗版 / 虚假 / 无价值内容或违法国家有关法律法规的内容,可点击提交进行申诉,我们将尽快为您处理。