From: Arnd Bergmann <arnd@arndb.de>
To: Hillf Danton <dhillf@gmail.com>
Subject: Re: [PATCH v0] add nano semaphore in kernel
Date: Mon, 27 Dec 2010 22:15:52 +0100
User-Agent: KMail/1.13.5 (Linux/2.6.35-16-generic; KDE/4.5.1; x86_64; ; )
Cc: linux-kernel@vger.kernel.org
References: <AANLkTim0Xmsnzz5yMLAx_+w-c7sSSReu0SKMAmjpbS+1@mail.gmail.com>
In-Reply-To: <AANLkTim0Xmsnzz5yMLAx_+w-c7sSSReu0SKMAmjpbS+1@mail.gmail.com>
MIME-Version: 1.0
Content-Type: Text/Plain;
  charset="utf-8"
Content-Transfer-Encoding: 7bit
Message-Id: <201012272215.52642.arnd@arndb.de>
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1332
Lines: 33

On Sunday 26 December 2010, Hillf Danton wrote:
> 
> Based upon high resolution timer and idea borrowed from semaphore,
> nano semaphore is created.
> 
> Nano semaphore provides finer time resolution depending on system
> configuration and capabilities.
> 
> Nano semaphore is not to replace semaphore, but used in application
> environments where nano seconds are required.
> 
> Three methods, nano_semaphore_try_down, nano_semaphore_down and
> nano_semaphore_up are implemented in a header file, and there is no
> corresponding C file since nano semaphore is not complex.
> 
> Signed-off-by: Hillf Danton <dhillf@gmail.com>

There are very few users of real semaphores today, and we're trying to
get rid of them. It's not clear what your requirements are, since you
have not posted any new users of this, but instead of adding more
locking primitives, I would recommend changing one of the existing
ones (mutex, semaphore, rwsem) to have nanosecond timeouts instead
of jiffies.

The easiest way would certainly be to change the three users of
down_timeout() to use nanoseconds.

	Arnd
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/