1 |
jsr166 |
1.1 |
/* |
2 |
|
|
* Written by Doug Lea with assistance from members of JCP JSR-166 |
3 |
|
|
* Expert Group and released to the public domain, as explained at |
4 |
|
|
* http://creativecommons.org/publicdomain/zero/1.0/ |
5 |
|
|
*/ |
6 |
|
|
|
7 |
|
|
package java.util.concurrent.locks; |
8 |
|
|
|
9 |
|
|
/** |
10 |
|
|
* A {@code ReadWriteLock} maintains a pair of associated {@link |
11 |
|
|
* Lock locks}, one for read-only operations and one for writing. |
12 |
|
|
* The {@linkplain #readLock read lock} may be held simultaneously |
13 |
|
|
* by multiple reader threads, so long as there are no writers. |
14 |
|
|
* The {@linkplain #writeLock write lock} is exclusive. |
15 |
|
|
* |
16 |
|
|
* <p>All {@code ReadWriteLock} implementations must guarantee that |
17 |
|
|
* the memory synchronization effects of {@code writeLock} operations |
18 |
|
|
* (as specified in the {@link Lock} interface) also hold with respect |
19 |
|
|
* to the associated {@code readLock}. That is, a thread successfully |
20 |
|
|
* acquiring the read lock will see all updates made upon previous |
21 |
|
|
* release of the write lock. |
22 |
|
|
* |
23 |
|
|
* <p>A read-write lock allows for a greater level of concurrency in |
24 |
|
|
* accessing shared data than that permitted by a mutual exclusion lock. |
25 |
|
|
* It exploits the fact that while only a single thread at a time (a |
26 |
|
|
* <em>writer</em> thread) can modify the shared data, in many cases any |
27 |
|
|
* number of threads can concurrently read the data (hence <em>reader</em> |
28 |
|
|
* threads). |
29 |
|
|
* In theory, the increase in concurrency permitted by the use of a read-write |
30 |
|
|
* lock will lead to performance improvements over the use of a mutual |
31 |
|
|
* exclusion lock. In practice this increase in concurrency will only be fully |
32 |
|
|
* realized on a multi-processor, and then only if the access patterns for |
33 |
|
|
* the shared data are suitable. |
34 |
|
|
* |
35 |
|
|
* <p>Whether or not a read-write lock will improve performance over the use |
36 |
|
|
* of a mutual exclusion lock depends on the frequency that the data is |
37 |
|
|
* read compared to being modified, the duration of the read and write |
38 |
|
|
* operations, and the contention for the data - that is, the number of |
39 |
|
|
* threads that will try to read or write the data at the same time. |
40 |
|
|
* For example, a collection that is initially populated with data and |
41 |
|
|
* thereafter infrequently modified, while being frequently searched |
42 |
|
|
* (such as a directory of some kind) is an ideal candidate for the use of |
43 |
|
|
* a read-write lock. However, if updates become frequent then the data |
44 |
|
|
* spends most of its time being exclusively locked and there is little, if any |
45 |
|
|
* increase in concurrency. Further, if the read operations are too short |
46 |
|
|
* the overhead of the read-write lock implementation (which is inherently |
47 |
|
|
* more complex than a mutual exclusion lock) can dominate the execution |
48 |
|
|
* cost, particularly as many read-write lock implementations still serialize |
49 |
|
|
* all threads through a small section of code. Ultimately, only profiling |
50 |
|
|
* and measurement will establish whether the use of a read-write lock is |
51 |
|
|
* suitable for your application. |
52 |
|
|
* |
53 |
|
|
* <p>Although the basic operation of a read-write lock is straight-forward, |
54 |
|
|
* there are many policy decisions that an implementation must make, which |
55 |
|
|
* may affect the effectiveness of the read-write lock in a given application. |
56 |
|
|
* Examples of these policies include: |
57 |
|
|
* <ul> |
58 |
|
|
* <li>Determining whether to grant the read lock or the write lock, when |
59 |
|
|
* both readers and writers are waiting, at the time that a writer releases |
60 |
|
|
* the write lock. Writer preference is common, as writes are expected to be |
61 |
|
|
* short and infrequent. Reader preference is less common as it can lead to |
62 |
|
|
* lengthy delays for a write if the readers are frequent and long-lived as |
63 |
|
|
* expected. Fair, or "in-order" implementations are also possible. |
64 |
|
|
* |
65 |
|
|
* <li>Determining whether readers that request the read lock while a |
66 |
|
|
* reader is active and a writer is waiting, are granted the read lock. |
67 |
|
|
* Preference to the reader can delay the writer indefinitely, while |
68 |
|
|
* preference to the writer can reduce the potential for concurrency. |
69 |
|
|
* |
70 |
|
|
* <li>Determining whether the locks are reentrant: can a thread with the |
71 |
|
|
* write lock reacquire it? Can it acquire a read lock while holding the |
72 |
|
|
* write lock? Is the read lock itself reentrant? |
73 |
|
|
* |
74 |
|
|
* <li>Can the write lock be downgraded to a read lock without allowing |
75 |
|
|
* an intervening writer? Can a read lock be upgraded to a write lock, |
76 |
|
|
* in preference to other waiting readers or writers? |
77 |
|
|
* |
78 |
|
|
* </ul> |
79 |
|
|
* You should consider all of these things when evaluating the suitability |
80 |
|
|
* of a given implementation for your application. |
81 |
|
|
* |
82 |
|
|
* @see ReentrantReadWriteLock |
83 |
|
|
* @see Lock |
84 |
|
|
* @see ReentrantLock |
85 |
|
|
* |
86 |
|
|
* @since 1.5 |
87 |
|
|
* @author Doug Lea |
88 |
|
|
*/ |
89 |
|
|
public interface ReadWriteLock { |
90 |
|
|
/** |
91 |
|
|
* Returns the lock used for reading. |
92 |
|
|
* |
93 |
|
|
* @return the lock used for reading |
94 |
|
|
*/ |
95 |
|
|
Lock readLock(); |
96 |
|
|
|
97 |
|
|
/** |
98 |
|
|
* Returns the lock used for writing. |
99 |
|
|
* |
100 |
|
|
* @return the lock used for writing |
101 |
|
|
*/ |
102 |
|
|
Lock writeLock(); |
103 |
|
|
} |