|
Derived from: none
Declared in: be/support/Locker.h
Library: libbe.so
Summary: more...
The BLocker class provides a locking mechanism that can be used to protect a section of critical code. The code that you want to protect should be placed between BLocker's Lock() and Unlock() calls:
BLocker *aLock = new BLocker();
. . .
aLock->Lock();
/* Protected code goes here. */
aLock->Unlock();
This guarantees that only one thread at a time will pass through the lock. After a thread has locked the BLocker object, subsequent attempts to lock by other threads are blocked until the first thread calls Unlock().
BLocker keeps track of the locking thread—the thread that's currently between Lock() and Unlock() calls. It lets the thread make nested calls to Lock() without blocking. Because of this, you can wrap a BLocker's lock around a series of functions that might, themselves, lock the same BLocker object.
For example, let's say you have a class called BadDog that's declared thus:
class BadDog : public BArchivable
{
public:
void DoThis();
void DoThat();
void DoThisAndThat();
private:
BLocker lock;
};
Let's implement the member functions as follows:
void BadDog::DoThis()
{
lock.Lock();
/* Do this here. */
lock.Unlock();
}
void BadDog::DoThat()
{
lock.Lock();
/* Do that here. */
lock.Unlock();
}
void BadDog::DoThisAndThat()
{
lock.Lock();
DoThis();
DoThat();
lock.Unlock();
}
Notice that DoThisAndThat() wraps the lock around its calls to DoThis() and DoThat(), both of which contain locks as well. A thread that gets past the Lock() call in DoThisAndThat() won't block when it calls the nested Lock() calls that it runs into in DoThis() and DoThat().
Keep in mind that nested Lock() calls must be balanced by an equal number of Unlock() calls.
See also: the BAutolock class
BLocker() |
BLocker(void)
BLocker(const char *name)
BLocker(bool benaphore_style)
BLocker(const char *name, bool benaphore_style)Sets up the object, creating a semaphore to implement the locking mechanism. The optional name is for diagnostics and debugging.
By default, a BLocker is implemented as a benaphore –– a hybrid integer mutex and semaphore. This allows for faster performance when there is little contention for the lock. If you'd rather use a pure semaphore for locking, then pass false as the benaphore_style argument.
~BLocker() |
virtual ~BLocker() Destroys the lock, deleting the controlling semaphore. If there are any threads blocked waiting to lock the object, they're immediately unblocked.
CountLocks() see LockingThread()
|
CountLockRequests() see LockingThread()
|
|
IsLocked() see LockingThread() | |
Lock() , LockWithTimeout() , Unlock() |
bool Lock(void) status_t LockWithTimeout(bigtime_t timeout) void Unlock(void) Lock() tries to lock the BLocker. The function returns a) when the lock is acquired (return = true) or b) immediately if the BLocker's semaphore is deleted, most commonly as a result of deleting the object (return = false). A thread can nest Lock() calls, but each call must be balanced by a concomitant Unlock().
LockWithTimeout() lets you declare a time limit, specified in microseconds. If LockWithTimeout() can't acquire the lock before the time limit expires, it returns B_TIMED_OUT. If the timeout is 0, this function immediately returns B_OK (if it locked the BLooper) or B_ERROR (if it failed to obtain the lock). If the timeout is B_INFINITE_TIMEOUT, it blocks without limit, just as Lock() does. Note that if Lock() returns 0 (false), it has failed to lock the BLocker, but if LockWithTimeout() returns 0 (B_OK), it has succeeded.
Unlock() releases one level of nested locks and returns immediately. If there are threads blocked waiting for the lock when the lock is released, the thread that's been waiting the longest acquires the lock.
Any thread can call Unlock() on any BLocker—the thread needn't be the lock's current holder. Call IsLocked() before calling Unlock() if you want to make sure you own a lock before you unlock it.
LockingThread() , IsLocked() , CountLocks() , CountLockRequests() , Sem() |
thread_id LockingThread(void) const bool IsLocked(void) const int32 CountLocks(void) const int32 CountLockRequests(void) const sem_id Sem(void) const These functions provide information that may be useful for debugging purposes.
LockingThread() returns the thread that currently has the BLocker locked, or B_ERROR if the BLooper isn't locked.
IsLocked() returns true if the calling thread currently has the BLocker locked (if it's the locking thread) and false if not (if some other thread is the locking thread or the BLocker isn't locked).
CountLocks() returns the number of times the locking thread has locked the BLocker—the number of Lock() (or LockWithTimeout()) calls that have not yet been balanced by matching Unlock() calls.
CountLockRequests() returns the number of threads currently trying to lock the BLocker. The count includes the thread that currently holds the lock plus all threads currently waiting to acquire it.
Sem() returns the sem_id for the semaphore that the BLocker uses to implement the locking mechanism.
|
Copyright © 2000 Be, Inc. All rights reserved..