Merge bk://bkbits.ras.ucalgary.ca/rgooch-2.5

into home.transmeta.com:/home/torvalds/v2.5/linux

arch/i386/kernel/process.c

@@ -674,6 +674,14 @@ void __switch_to(struct task_struct *prev_p, struct task_struct *next_p)
 	 */
 	tss->esp0 = next->esp0;
 
+	/*
+	 * Load the per-thread Thread-Local Storage descriptor.
+	 *
+	 * NOTE: it's faster to do the two stores unconditionally
+	 * than to branch away.
+	 */
+	load_TLS_desc(next, cpu);
+
 	/*
 	 * Save away %fs and %gs. No need to save %es and %ds, as
 	 * those are always kernel segments while inside the kernel.
@@ -689,14 +697,6 @@ void __switch_to(struct task_struct *prev_p, struct task_struct *next_p)
 		loadsegment(gs, next->gs);
 	}
 
-	/*
-	 * Load the per-thread Thread-Local Storage descriptor.
-	 *
-	 * NOTE: it's faster to do the two stores unconditionally
-	 * than to branch away.
-	 */
-	load_TLS_desc(next, cpu);
-
 	/*
 	 * Now maybe reload the debug registers
 	 */

include/asm-i386/rwsem.h

@@ -117,6 +117,29 @@ LOCK_PREFIX	"  incl      (%%eax)\n\t" /* adds 0x00000001, returns the old value
 		: "memory", "cc");
 }
 
+/*
+ * trylock for reading -- returns 1 if successful, 0 if contention
+ */
+static inline int __down_read_trylock(struct rw_semaphore *sem)
+{
+	__s32 result, tmp;
+	__asm__ __volatile__(
+		"# beginning __down_read_trylock\n\t"
+		"  movl      %0,%1\n\t"
+		"1:\n\t"
+		"  movl	     %1,%2\n\t"
+		"  addl      %3,%2\n\t"
+		"  jle	     2f\n\t"
+LOCK_PREFIX	"  cmpxchgl  %2,%0\n\t"
+		"  jnz	     1b\n\t"
+		"2:\n\t"
+		"# ending __down_read_trylock\n\t"
+		: "+m"(sem->count), "=&a"(result), "=&r"(tmp)
+		: "i"(RWSEM_ACTIVE_READ_BIAS)
+		: "memory", "cc");
+	return result>=0 ? 1 : 0;
+}
+
 /*
  * lock for writing
  */
@@ -144,6 +167,19 @@ LOCK_PREFIX	"  xadd      %%edx,(%%eax)\n\t" /* subtract 0x0000ffff, returns the
 		: "memory", "cc");
 }
 
+/*
+ * trylock for writing -- returns 1 if successful, 0 if contention
+ */
+static inline int __down_write_trylock(struct rw_semaphore *sem)
+{
+	signed long ret = cmpxchg(&sem->count,
+				  RWSEM_UNLOCKED_VALUE, 
+				  RWSEM_ACTIVE_WRITE_BIAS);
+	if (ret == RWSEM_UNLOCKED_VALUE)
+		return 1;
+	return 0;
+}
+
 /*
  * unlock after reading
  */

include/linux/rwsem-spinlock.h

@@ -54,9 +54,12 @@ struct rw_semaphore {
 
 extern void FASTCALL(init_rwsem(struct rw_semaphore *sem));
 extern void FASTCALL(__down_read(struct rw_semaphore *sem));
+extern int FASTCALL(__down_read_trylock(struct rw_semaphore *sem));
 extern void FASTCALL(__down_write(struct rw_semaphore *sem));
+extern int FASTCALL(__down_write_trylock(struct rw_semaphore *sem));
 extern void FASTCALL(__up_read(struct rw_semaphore *sem));
 extern void FASTCALL(__up_write(struct rw_semaphore *sem));
+extern void FASTCALL(__downgrade_write(struct rw_semaphore *sem));
 
 #endif /* __KERNEL__ */
 #endif /* _LINUX_RWSEM_SPINLOCK_H */

include/linux/rwsem.h

@@ -45,6 +45,18 @@ static inline void down_read(struct rw_semaphore *sem)
 	rwsemtrace(sem,"Leaving down_read");
 }
 
+/*
+ * trylock for reading -- returns 1 if successful, 0 if contention
+ */
+static inline int down_read_trylock(struct rw_semaphore *sem)
+{
+	int ret;
+	rwsemtrace(sem,"Entering down_read_trylock");
+	ret = __down_read_trylock(sem);
+	rwsemtrace(sem,"Leaving down_read_trylock");
+	return ret;
+}
+
 /*
  * lock for writing
  */
@@ -55,6 +67,18 @@ static inline void down_write(struct rw_semaphore *sem)
 	rwsemtrace(sem,"Leaving down_write");
 }
 
+/*
+ * trylock for writing -- returns 1 if successful, 0 if contention
+ */
+static inline int down_write_trylock(struct rw_semaphore *sem)
+{
+	int ret;
+	rwsemtrace(sem,"Entering down_write_trylock");
+	ret = __down_write_trylock(sem);
+	rwsemtrace(sem,"Leaving down_write_trylock");
+	return ret;
+}
+
 /*
  * release a read lock
  */
@@ -85,6 +109,5 @@ static inline void downgrade_write(struct rw_semaphore *sem)
 	rwsemtrace(sem,"Leaving downgrade_write");
 }
 
-
 #endif /* __KERNEL__ */
 #endif /* _LINUX_RWSEM_H */

lib/rwsem-spinlock.c

@@ -46,8 +46,9 @@ void init_rwsem(struct rw_semaphore *sem)
  *   - the 'waiting count' is non-zero
  * - 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;
 	int woken;
@@ -56,7 +57,14 @@ static inline struct rw_semaphore *__rwsem_do_wake(struct rw_semaphore *sem)
 
 	waiter = list_entry(sem->wait_list.next,struct rwsem_waiter,list);
 
-	/* try to grant a single write lock if there's a writer at the front of the queue
+	if (!wakewrite) {
+		if (waiter->flags & RWSEM_WAITING_FOR_WRITE)
+			goto out;
+		goto dont_wake_writers;
+	}
+
+	/* if we are allowed to wake writers try to grant a single write lock if there's a
+	 * writer at the front of the queue
 	 * - we leave the 'waiting count' incremented to signify potential contention
 	 */
 	if (waiter->flags & RWSEM_WAITING_FOR_WRITE) {
@@ -68,16 +76,19 @@ static inline struct rw_semaphore *__rwsem_do_wake(struct rw_semaphore *sem)
 	}
 
 	/* grant an infinite number of read locks to the readers at the front of the queue */
+ dont_wake_writers:
 	woken = 0;
-	do {
+	while (waiter->flags&RWSEM_WAITING_FOR_READ) {
+		struct list_head *next = waiter->list.next;
+
 		list_del(&waiter->list);
 		waiter->flags = 0;
 		wake_up_process(waiter->task);
 		woken++;
 		if (list_empty(&sem->wait_list))
 			break;
-		waiter = list_entry(sem->wait_list.next,struct rwsem_waiter,list);
-	} while (waiter->flags&RWSEM_WAITING_FOR_READ);
+		waiter = list_entry(next,struct rwsem_waiter,list);
+	}
 
 	sem->activity += woken;
 
@@ -148,6 +159,28 @@ void __down_read(struct rw_semaphore *sem)
 	rwsemtrace(sem,"Leaving __down_read");
 }
 
+/*
+ * trylock for reading -- returns 1 if successful, 0 if contention
+ */
+int __down_read_trylock(struct rw_semaphore *sem)
+{
+	int ret = 0;
+	rwsemtrace(sem,"Entering __down_read_trylock");
+
+	spin_lock(&sem->wait_lock);
+
+	if (sem->activity>=0 && list_empty(&sem->wait_list)) {
+		/* granted */
+		sem->activity++;
+		ret = 1;
+	}
+
+	spin_unlock(&sem->wait_lock);
+
+	rwsemtrace(sem,"Leaving __down_read_trylock");
+	return ret;
+}
+
 /*
  * get a write lock on the semaphore
  * - note that we increment the waiting count anyway to indicate an exclusive lock
@@ -194,6 +227,28 @@ void __down_write(struct rw_semaphore *sem)
 	rwsemtrace(sem,"Leaving __down_write");
 }
 
+/*
+ * trylock for writing -- returns 1 if successful, 0 if contention
+ */
+int __down_write_trylock(struct rw_semaphore *sem)
+{
+	int ret = 0;
+	rwsemtrace(sem,"Entering __down_write_trylock");
+
+	spin_lock(&sem->wait_lock);
+
+	if (sem->activity==0 && list_empty(&sem->wait_list)) {
+		/* granted */
+		sem->activity = -1;
+		ret = 1;
+	}
+
+	spin_unlock(&sem->wait_lock);
+
+	rwsemtrace(sem,"Leaving __down_write_trylock");
+	return ret;
+}
+
 /*
  * release a read lock on the semaphore
  */
@@ -222,18 +277,40 @@ void __up_write(struct rw_semaphore *sem)
 
 	sem->activity = 0;
 	if (!list_empty(&sem->wait_list))
-		sem = __rwsem_do_wake(sem);
+		sem = __rwsem_do_wake(sem, 1);
 
 	spin_unlock(&sem->wait_lock);
 
 	rwsemtrace(sem,"Leaving __up_write");
 }
 
+/*
+ * downgrade a write lock into a read lock
+ * - just wake up any readers at the front of the queue
+ */
+void __downgrade_write(struct rw_semaphore *sem)
+{
+	rwsemtrace(sem,"Entering __rwsem_downgrade");
+
+	spin_lock(&sem->wait_lock);
+
+	sem->activity = 1;
+	if (!list_empty(&sem->wait_list))
+		sem = __rwsem_do_wake(sem,0);
+
+	spin_unlock(&sem->wait_lock);
+
+	rwsemtrace(sem,"Leaving __rwsem_downgrade");
+}
+
 EXPORT_SYMBOL(init_rwsem);
 EXPORT_SYMBOL(__down_read);
+EXPORT_SYMBOL(__down_read_trylock);
 EXPORT_SYMBOL(__down_write);
+EXPORT_SYMBOL(__down_write_trylock);
 EXPORT_SYMBOL(__up_read);
 EXPORT_SYMBOL(__up_write);
+EXPORT_SYMBOL(__downgrade_write);
 #if RWSEM_DEBUG
 EXPORT_SYMBOL(rwsemtrace);
 #endif