Document performance of QMutex

Change-Id: I4de2fd5ba717975b5de98ffe7ca6348afbed1b2f
Reviewed-by: Marc Mutz <marc.mutz@kdab.com>

1 files changed, 7 insertions, 0 deletions

diff --git a/src/corelib/thread/qmutex.cpp b/src/corelib/thread/qmutex.cpp
index 742a572bef..fa3bb080ae 100644
--- a/src/corelib/thread/qmutex.cpp
+++ b/src/corelib/thread/qmutex.cpp
@@ -139,6 +139,11 @@ public:
     lock calls unlock(). A non-blocking alternative to lock() is
     tryLock().
 
+    QMutex is optimized to be fast in the non-contended case. A non-recursive
+    QMutex will not allocate memory if there is no contention on that mutex.
+    It is constructed and destroyed with almost no overhead,
+    which means it is fine to have many mutexes as part of other classes.
+
     \sa QMutexLocker, QReadWriteLock, QSemaphore, QWaitCondition
 */
 
@@ -165,6 +170,8 @@ public:
     a thread may only lock a mutex once. The default is
     QMutex::NonRecursive.
 
+    Recursive mutexes are slower and take more memory than non-recursive ones.
+
     \sa lock(), unlock()
 */
 QMutex::QMutex(RecursionMode mode)