From mboxrd@z Thu Jan  1 00:00:00 1970
From: Russell King - ARM Linux <linux@arm.linux.org.uk>
Subject: Re: [PATCH] phylib: add mdio-gpio bus driver (v2)
Date: Tue, 28 Oct 2008 15:17:55 +0000
Message-ID: <20081028151755.GA1034@flint.arm.linux.org.uk>
References: <20081027105318.21923.24436.stgit@Programuotojas.82-135-208-232.ip.zebra.lt> <20081027143734.GA13565@secretlab.ca> <4906C338.6010300@teltonika.lt> <fa686aa40810280608q1c58e07cq7c5f2cc9feeb8ef6@mail.gmail.com>
Mime-Version: 1.0
Return-path: <linux-embedded-owner@vger.kernel.org>
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed;
	d=arm.linux.org.uk; s=caramon; h=Date:From:To:Cc:Subject:
	Message-ID:References:Mime-Version:Content-Type:In-Reply-To:
	Sender; bh=7EqBiCTk0TrGa9qYQFXYiH/q/Y9ITr3aaxniMqrCUPs=; b=Idf5b
	1WkPtsaAPGEU7QE5Erje7kYzT9lPjwEWFdkATDd/nxQwiVYYzZX+AG5HMbfgedmU
	2EyEVbCOxKiv0a331H+XgVR9jwrC10U7dW0spm508MV3c4GX2VOxUstKdnNn0B90
	I+NG+2YclsAUVGG5b6Abongn+HWPbZK5d02Wgk=
Content-Disposition: inline
In-Reply-To: <fa686aa40810280608q1c58e07cq7c5f2cc9feeb8ef6@mail.gmail.com>
Sender: linux-embedded-owner@vger.kernel.org
List-ID: <linux-embedded.vger.kernel.org>
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: Grant Likely <grant.likely@secretlab.ca>
Cc: Paulius Zaleckas <paulius.zaleckas@teltonika.lt>, netdev@vger.kernel.org, linux-arm-kernel@lists.arm.linux.org.uk, linux-embedded@vger.kernel.org

On Tue, Oct 28, 2008 at 07:08:06AM -0600, Grant Likely wrote:
> On Tue, Oct 28, 2008 at 1:46 AM, Paulius Zaleckas
> <paulius.zaleckas@teltonika.lt> wrote:
> > Grant Likely wrote:
> >> The IRQ array is fixed size.  You can add it to the mdio_gpio_info
> >> structure and then just set the pointer here so that only one kzalloc
> >> is needed.
> >
> > It can be put in mdio_gpio_info, but please note that mdio_gpio_info is
> > allocated with kzalloc() and irq with kmalloc(), because there is no need
> > to fill this array with zeros(see below).
> 
> Adding an additional 32 words to be zeroed in the mdio_gpio_info
> kzalloc is considerably cheaper than doing an additional kmalloc.
> Plus, once the array is zeroed it is then in the cache and so the
> filling it with -1 also becomes cheaper.

Actually no, it doesn't become cheaper.  You're making the assumption
that cache lines are allocated when memory is written to.  This isn't
the case with the vast majority of ARM CPUs.

That means zeroing allocated memory and then filling it with -1 makes
it twice as expensive as just filling it with -1.