sync.RWMutex: rewritten to add support for concurrent readers.

Also made sync.xadd public to help testing sync.RWMutex.

Also added unit tests for sync.RWMutex.

R=rsc
https://golang.org/cl/162044

src/pkg/sync/Makefile

@@ -7,6 +7,7 @@ include ../../Make.$(GOARCH)
 TARG=sync
 GOFILES=\
 	mutex.go\
+	rwmutex.go\
 
 OFILES=\
 	asm_$(GOARCH).$O\

src/pkg/sync/mutex.go

@@ -20,6 +20,9 @@ type Mutex struct {
 	sema	uint32;
 }
 
+// Add delta to *val, and return the new *val in a thread-safe way. If multiple
+// goroutines call xadd on the same val concurrently, the changes will be
+// serialized, and all the deltas will be added in an undefined order.
 func xadd(val *uint32, delta int32) (new uint32) {
 	for {
 		v := *val;
@@ -55,52 +58,3 @@ func (m *Mutex) Unlock() {
 	}
 	runtime.Semrelease(&m.sema);
 }
-
-// Stub implementation of r/w locks.
-// This satisfies the semantics but
-// is not terribly efficient.
-
-// The next comment goes in the BUGS section of the document,
-// in its own paragraph, without the (rsc) tag.
-
-// BUG(rsc): RWMutex does not (yet) allow multiple readers;
-// instead it behaves as if RLock and RUnlock were Lock and Unlock.
-
-// An RWMutex is a reader/writer mutual exclusion lock.
-// The lock can be held by an arbitrary number of readers
-// or a single writer.
-// RWMutexes can be created as part of other
-// structures; the zero value for a RWMutex is
-// an unlocked mutex.
-type RWMutex struct {
-	m Mutex;
-}
-
-// RLock locks rw for reading.
-// If the lock is already locked for writing or there is a writer already waiting
-// to acquire the lock, RLock blocks until the writer has released the lock.
-func (rw *RWMutex) RLock()	{ rw.m.Lock() }
-
-// RUnlock undoes a single RLock call;
-// it does not affect other simultaneous readers.
-// It is a run-time error if rw is not locked for reading
-// on entry to RUnlock.
-func (rw *RWMutex) RUnlock()	{ rw.m.Unlock() }
-
-// Lock locks rw for writing.
-// If the lock is already locked for reading or writing,
-// Lock blocks until the lock is available.
-// To ensure that the lock eventually becomes available,
-// a blocked Lock call excludes new readers from acquiring
-// the lock.
-func (rw *RWMutex) Lock()	{ rw.m.Lock() }
-
-// Unlock unlocks rw for writing.
-// It is a run-time error if rw is not locked for writing
-// on entry to Unlock.
-//
-// Like for Mutexes,
-// a locked RWMutex is not associated with a particular goroutine.
-// It is allowed for one goroutine to RLock (Lock) an RWMutex and then
-// arrange for another goroutine to RUnlock (Unlock) it.
-func (rw *RWMutex) Unlock()	{ rw.m.Unlock() }

src/pkg/sync/rwmutex.go

@@ -0,0 +1,75 @@
+// Copyright 2009 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package sync
+
+// An RWMutex is a reader/writer mutual exclusion lock.
+// The lock can be held by an arbitrary number of readers
+// or a single writer.
+// RWMutexes can be created as part of other
+// structures; the zero value for a RWMutex is
+// an unlocked mutex.
+//
+// Writers take priority over Readers: no new RLocks
+// are granted while a blocked Lock call is waiting.
+type RWMutex struct {
+	w		Mutex;	// held if there are pending readers or writers
+	r		Mutex;	// held if the w is being rd
+	readerCount	uint32;	// number of pending readers
+}
+
+// RLock locks rw for reading.
+// If the lock is already locked for writing or there is a writer already waiting
+// to r the lock, RLock blocks until the writer has released the lock.
+func (rw *RWMutex) RLock() {
+	// Use rw.r.Lock() to block granting the RLock if a goroutine
+	// is waiting for its Lock. This is the prevent starvation of W in
+	// this situation:
+	//   A: rw.RLock()  // granted
+	//   W: rw.Lock()   // waiting for rw.w().Lock()
+	//   B: rw.RLock()  // granted
+	//   C: rw.RLock()  // granted
+	//   B: rw.RUnlock()
+	//   ... (new readers come and go indefinitely, W is starving)
+	rw.r.Lock();
+	if xadd(&rw.readerCount, 1) == 1 {
+		// The first reader locks rw.w, so writers will be blocked
+		// while the readers have the RLock.
+		rw.w.Lock()
+	}
+	rw.r.Unlock();
+}
+
+// RUnlock undoes a single RLock call;
+// it does not affect other simultaneous readers.
+// It is a run-time error if rw is not locked for reading
+// on entry to RUnlock.
+func (rw *RWMutex) RUnlock() {
+	if xadd(&rw.readerCount, -1) == 0 {
+		// last reader finished, enable writers
+		rw.w.Unlock()
+	}
+}
+
+// Lock locks rw for writing.
+// If the lock is already locked for reading or writing,
+// Lock blocks until the lock is available.
+// To ensure that the lock eventually becomes available,
+// a blocked Lock call excludes new readers from acquiring
+// the lock.
+func (rw *RWMutex) Lock() {
+	rw.r.Lock();
+	rw.w.Lock();
+	rw.r.Unlock();
+}
+
+// Unlock unlocks rw for writing.
+// It is a run-time error if rw is not locked for writing
+// on entry to Unlock.
+//
+// Like for Mutexes,
+// a locked RWMutex is not associated with a particular goroutine.
+// It is allowed for one goroutine to RLock (Lock) an RWMutex and then
+// arrange for another goroutine to RUnlock (Unlock) it.
+func (rw *RWMutex) Unlock()	{ rw.w.Unlock() }

src/pkg/sync/rwmutex_test.go

@@ -0,0 +1,114 @@
+// Copyright 2009 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// GOMAXPROCS=10 gotest
+
+package sync_test
+
+import (
+	"fmt";
+	"runtime";
+	. "sync";
+	"testing";
+)
+
+func parallelReader(m *RWMutex, clocked, cunlock, cdone chan bool) {
+	m.RLock();
+	clocked <- true;
+	<-cunlock;
+	m.RUnlock();
+	cdone <- true;
+}
+
+func doTestParallelReaders(numReaders, gomaxprocs int) {
+	runtime.GOMAXPROCS(gomaxprocs);
+	var m RWMutex;
+	clocked := make(chan bool);
+	cunlock := make(chan bool);
+	cdone := make(chan bool);
+	for i := 0; i < numReaders; i++ {
+		go parallelReader(&m, clocked, cunlock, cdone)
+	}
+	// Wait for all parallel RLock()s to succeed.
+	for i := 0; i < numReaders; i++ {
+		<-clocked
+	}
+	for i := 0; i < numReaders; i++ {
+		cunlock <- true
+	}
+	// Wait for the goroutines to finish.
+	for i := 0; i < numReaders; i++ {
+		<-cdone
+	}
+}
+
+func TestParallelReaders(t *testing.T) {
+	doTestParallelReaders(1, 4);
+	doTestParallelReaders(3, 4);
+	doTestParallelReaders(4, 2);
+}
+
+func reader(rwm *RWMutex, num_iterations int, activity *uint32, cdone chan bool) {
+	for i := 0; i < num_iterations; i++ {
+		rwm.RLock();
+		n := Xadd(activity, 1);
+		if n < 1 || n >= 10000 {
+			panic(fmt.Sprintf("wlock(%d)\n", n))
+		}
+		for i := 0; i < 100; i++ {
+		}
+		Xadd(activity, -1);
+		rwm.RUnlock();
+	}
+	cdone <- true;
+}
+
+func writer(rwm *RWMutex, num_iterations int, activity *uint32, cdone chan bool) {
+	for i := 0; i < num_iterations; i++ {
+		rwm.Lock();
+		n := Xadd(activity, 10000);
+		if n != 10000 {
+			panic(fmt.Sprintf("wlock(%d)\n", n))
+		}
+		for i := 0; i < 100; i++ {
+		}
+		Xadd(activity, -10000);
+		rwm.Unlock();
+	}
+	cdone <- true;
+}
+
+func HammerRWMutex(gomaxprocs, numReaders, num_iterations int) {
+	runtime.GOMAXPROCS(gomaxprocs);
+	// Number of active readers + 10000 * number of active writers.
+	var activity uint32;
+	var rwm RWMutex;
+	cdone := make(chan bool);
+	go writer(&rwm, num_iterations, &activity, cdone);
+	var i int;
+	for i = 0; i < numReaders/2; i++ {
+		go reader(&rwm, num_iterations, &activity, cdone)
+	}
+	go writer(&rwm, num_iterations, &activity, cdone);
+	for ; i < numReaders; i++ {
+		go reader(&rwm, num_iterations, &activity, cdone)
+	}
+	// Wait for the 2 writers and all readers to finish.
+	for i := 0; i < 2+numReaders; i++ {
+		<-cdone
+	}
+}
+
+func TestRWMutex(t *testing.T) {
+	HammerRWMutex(1, 1, 1000);
+	HammerRWMutex(1, 3, 1000);
+	HammerRWMutex(1, 10, 1000);
+	HammerRWMutex(4, 1, 1000);
+	HammerRWMutex(4, 3, 1000);
+	HammerRWMutex(4, 10, 1000);
+	HammerRWMutex(10, 1, 1000);
+	HammerRWMutex(10, 3, 1000);
+	HammerRWMutex(10, 10, 1000);
+	HammerRWMutex(10, 5, 10000);
+}

src/pkg/sync/xadd_test.go

@@ -0,0 +1,9 @@
+// Copyright 2009 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package sync
+
+func Xadd(val *uint32, delta int32) (new uint32) {
+	return xadd(val, delta)
+}