From mboxrd@z Thu Jan  1 00:00:00 1970
From: thomas.petazzoni@free-electrons.com (Thomas Petazzoni)
Date: Thu, 27 Mar 2014 15:19:45 +0100
Subject: [PATCH 1/6] ARM: mvebu: introduce CPU reset code
In-Reply-To: <53342EF1.3020508@free-electrons.com>
References: <1395927485-11842-1-git-send-email-thomas.petazzoni@free-electrons.com>
 <1395927485-11842-2-git-send-email-thomas.petazzoni@free-electrons.com>
 <53342EF1.3020508@free-electrons.com>
Message-ID: <20140327151945.365db704@skate>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

Dear Gregory CLEMENT,

On Thu, 27 Mar 2014 15:00:17 +0100, Gregory CLEMENT wrote:

> > +static struct of_device_id of_cpu_reset_table[] = {
> > +	{.compatible = "marvell,armada-370-cpu-reset", .data = (void*) ARMADA_370_MAX_CPUS },
> What about removing the previous line. As explained in patch 5, the CPU
> reset driver is not really needed as Armada 370 is single core and the
> only use of the CPU reset driver is to boot secondary processors. So by
> removing this line we can keep the marvell,armada-370-cpu-reset node in
> the device tree without doing useless initialization.

I found it weird to have a compatible string marked as supported in the
DT binding document, but not actually supported by the kernel. I know
it's possible, but I found it odd, especially considering the fact that
mapping these registers, even if unused, isn't costing much.

I don't have a strong opinion on this, so if others voice in this way,
I'll change it.

Thomas
-- 
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com