From mboxrd@z Thu Jan  1 00:00:00 1970
From: sebastian.hesselbarth@gmail.com (Sebastian Hesselbarth)
Date: Thu, 13 Mar 2014 10:17:04 +0000
Subject: [PATCH 1/2] ARM: dts: berlin2q: add the Marvell Armada 1500 pro
 (BG2Q) device tree
In-Reply-To: <20140313100838.GA4418@piout.net>
References: <1394622364-6848-1-git-send-email-antoine.tenart@free-electrons.com>
 <1394622364-6848-2-git-send-email-antoine.tenart@free-electrons.com>
 <20140312122210.GA3040@piout.net> <532180E0.9060108@gmail.com>
 <20140313100838.GA4418@piout.net>
Message-ID: <532185A0.3020007@gmail.com>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

On 03/13/2014 10:08 AM, Alexandre Belloni wrote:
> On 13/03/2014 at 09:56:48 +0000, Sebastian Hesselbarth wrote :
>> On 03/12/2014 12:22 PM, Alexandre Belloni wrote:
>>>
>>> The 25MHz crystal is on the board, please move it to the board dts.
>>
>> If you can confirm that sysmgr clock comes directly from this crystal,
>> I agree. If it is fed into a pll and can possibly be modified or gated,
>> make it depend on a 25MHz board crystal. If 25MHz is the only option
>> for this clock input, I could also live with it being part of the SoC
>> description.
>>
>
> Yeah, after more discussion, the 25 MHz crystl is the only option and it
> is fed to an oscillator that generate the reference clock for the PLLs.
> It is probably fine to let it in the dtsi.
>

Fine, I am planing to do more work on Berlin and especially clocks next
cycle. So most of the fixed-clocks will vanish anyway.

BTW, back at ELCE/ARM summit I successfully enabled SMP for berlin2.
I'll dig that patch out for you to re-test on Quad-core bg2q.

Sebastian