Re: Patch: linux-2.5.29 __downgrade_write() for CONFIG_RWSEM_GENERIC_SPINLOCK

[Roman Zippel <zippel@linux-m68k.org>]

Hi,

On Sun, 28 Jul 2002, Christoph Hellwig wrote:

> -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 *__rws
>
>  	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) {

You don't really need that extra argument, testing sem->activity should do
the same job.
If you exchange the wakewrite (or sem->activity) test and the
waiter->flags you can fold it into the next test (this means all the extra
work would only be done, if we have a writer waiting at the top).

bye, Roman