[PATCH] read-write semaphore downgrade and trylock

Here's a patch from Christoph Hellwig and myself to supply write->read
semaphore downgrade, and also from Brian Watson to supply trylock for rwsems.

include/asm-i386/rwsem.h

@@ -46,6 +46,7 @@ struct rwsem_waiter;
 extern struct rw_semaphore *FASTCALL(rwsem_down_read_failed(struct rw_semaphore *sem));
 extern struct rw_semaphore *FASTCALL(rwsem_down_write_failed(struct rw_semaphore *sem));
 extern struct rw_semaphore *FASTCALL(rwsem_wake(struct rw_semaphore *));
+extern struct rw_semaphore *FASTCALL(rwsem_downgrade_write(struct rw_semaphore *sem));
 
 /*
  * the semaphore definition
@@ -195,6 +196,31 @@ LOCK_PREFIX	"  xaddl     %%edx,(%%eax)\n\t" /* tries to transition 0xffff0001 ->
 		: "memory", "cc", "edx");
 }
 
+/*
+ * downgrade write lock to read lock
+ */
+static inline void __downgrade_write(struct rw_semaphore *sem)
+{
+	__asm__ __volatile__(
+		"# beginning __downgrade_write\n\t"
+LOCK_PREFIX	"  addl      %2,(%%eax)\n\t" /* transitions 0xZZZZ0001 -> 0xYYYY0001 */
+		"  js        2f\n\t" /* jump if the lock is being waited upon */
+		"1:\n\t"
+		LOCK_SECTION_START("")
+		"2:\n\t"
+		"  pushl     %%ecx\n\t"
+		"  pushl     %%edx\n\t"
+		"  call      rwsem_downgrade_wake\n\t"
+		"  popl      %%edx\n\t"
+		"  popl      %%ecx\n\t"
+		"  jmp       1b\n"
+		LOCK_SECTION_END
+		"# ending __downgrade_write\n"
+		: "=m"(sem->count)
+		: "a"(sem), "i"(-RWSEM_WAITING_BIAS), "m"(sem->count)
+		: "memory", "cc");
+}
+
 /*
  * implement atomic add functionality
  */

include/linux/rwsem.h

@@ -75,6 +75,16 @@ static inline void up_write(struct rw_semaphore *sem)
 	rwsemtrace(sem,"Leaving up_write");
 }
 
+/*
+ * downgrade write lock to read lock
+ */
+static inline void downgrade_write(struct rw_semaphore *sem)
+{
+	rwsemtrace(sem,"Entering downgrade_write");
+	__downgrade_write(sem);
+	rwsemtrace(sem,"Leaving downgrade_write");
+}
+
 
 #endif /* __KERNEL__ */
 #endif /* _LINUX_RWSEM_H */

lib/rwsem.c

@@ -34,8 +34,9 @@ void rwsemtrace(struct rw_semaphore *sem, const char *str)
  *   - there must be someone on the queue
  * - the spinlock must be held by the caller
  * - woken process blocks are discarded from the list after having flags zeroised
+ * - writers are only woken if wakewrite is non-zero
  */
-static inline struct rw_semaphore *__rwsem_do_wake(struct rw_semaphore *sem)
+static inline struct rw_semaphore *__rwsem_do_wake(struct rw_semaphore *sem, int wakewrite)
 {
 	struct rwsem_waiter *waiter;
 	struct list_head *next;
@@ -44,6 +45,9 @@ static inline struct rw_semaphore *__rwsem_do_wake(struct rw_semaphore *sem)
 
 	rwsemtrace(sem,"Entering __rwsem_do_wake");
 
+	if (!wakewrite)
+		goto dont_wake_writers;
+
 	/* only wake someone up if we can transition the active part of the count from 0 -> 1 */
  try_again:
 	oldcount = rwsem_atomic_update(RWSEM_ACTIVE_BIAS,sem) - RWSEM_ACTIVE_BIAS;
@@ -64,6 +68,12 @@ static inline struct rw_semaphore *__rwsem_do_wake(struct rw_semaphore *sem)
 	wake_up_process(waiter->task);
 	goto out;
 
+	/* don't want to wake any writers */
+ dont_wake_writers:
+	waiter = list_entry(sem->wait_list.next,struct rwsem_waiter,list);
+	if (waiter->flags & RWSEM_WAITING_FOR_WRITE)
+		goto out;
+
 	/* grant an infinite number of read locks to the readers at the front of the queue
 	 * - note we increment the 'active part' of the count by the number of readers (less one
 	 *   for the activity decrement we've already done) before waking any processes up
@@ -132,7 +142,7 @@ static inline struct rw_semaphore *rwsem_down_failed_common(struct rw_semaphore
 	 * - it might even be this process, since the waker takes a more active part
 	 */
 	if (!(count & RWSEM_ACTIVE_MASK))
-		sem = __rwsem_do_wake(sem);
+		sem = __rwsem_do_wake(sem,1);
 
 	spin_unlock(&sem->wait_lock);
 
@@ -193,7 +203,7 @@ struct rw_semaphore *rwsem_wake(struct rw_semaphore *sem)
 
 	/* do nothing if list empty */
 	if (!list_empty(&sem->wait_list))
-		sem = __rwsem_do_wake(sem);
+		sem = __rwsem_do_wake(sem,1);
 
 	spin_unlock(&sem->wait_lock);
 
@@ -202,6 +212,27 @@ struct rw_semaphore *rwsem_wake(struct rw_semaphore *sem)
 	return sem;
 }
 
+/*
+ * downgrade a write lock into a read lock
+ * - caller incremented waiting part of count, and discovered it to be still negative
+ * - just wake up any readers at the front of the queue
+ */
+struct rw_semaphore *rwsem_downgrade_wake(struct rw_semaphore *sem)
+{
+	rwsemtrace(sem,"Entering rwsem_downgrade_wake");
+
+	spin_lock(&sem->wait_lock);
+
+	/* do nothing if list empty */
+	if (!list_empty(&sem->wait_list))
+		sem = __rwsem_do_wake(sem,0);
+
+	spin_unlock(&sem->wait_lock);
+
+	rwsemtrace(sem,"Leaving rwsem_downgrade_wake");
+	return sem;
+}
+
 EXPORT_SYMBOL_NOVERS(rwsem_down_read_failed);
 EXPORT_SYMBOL_NOVERS(rwsem_down_write_failed);
 EXPORT_SYMBOL_NOVERS(rwsem_wake);