Re: semaphore: lockless fastpath using atomic_{inc,dec}_return

[Arjan van de Ven]

On Wed, 09 Jul 2008 14:13:29 +0100
Bruno Santos <bsantos@xxxxxxxx> wrote:

> So far the machine I'm testing this (Core2 Duo) it's been up to
> almost 24H.
> 
> It seems the patch got screwed by the mailer, so I'm posting it again.
> 
> 
>  From 343d08a5d172d103e49c77e5580a45f02fab2b5e Mon Sep 17 00:00:00
> 2001 From: Bruno Santos <bsantos@xxxxxxxx>
> Date: Tue, 8 Jul 2008 23:40:53 +0100
> Subject: [PATCH] semaphore lockless fastpath
> 
> 
> Signed-off-by: Bruno Santos <bsantos@xxxxxxxx>
> ---
>  include/linux/semaphore.h |    4 +-
>  kernel/semaphore.c        |  131 
> +++++++++++++++++++++++++--------------------
>  2 files changed, 75 insertions(+), 60 deletions(-)
> 
> diff --git a/include/linux/semaphore.h b/include/linux/semaphore.h
> index 9cae64b..d7850f0 100644
> --- a/include/linux/semaphore.h
> +++ b/include/linux/semaphore.h
> @@ -14,15 +14,15 @@
>  
>  /* Please don't access any members of this structure directly */
>  struct semaphore {
> +    atomic_t        count;
>      spinlock_t        lock;
> -    unsigned int        count;
>      struct list_head    wait_list;
>  };
> 



hi,

not to ruin the party but... how is this lockless? An atomic variable
is every bit a "lock" as a spinlock is... and very much equally
expensive as well for most cases ;-(
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/