From mboxrd@z Thu Jan  1 00:00:00 1970
From: shawn.guo@freescale.com (Shawn Guo)
Date: Mon, 20 Dec 2010 22:25:17 +0800
Subject: [PATCH v7 05/15] ARM: mxs: Add low-level debug UART support
In-Reply-To: <20101220135535.GB8885@pengutronix.de>
References: <1292679575-2635-1-git-send-email-shawn.guo@freescale.com>
	<1292679575-2635-4-git-send-email-shawn.guo@freescale.com>
	<20101220115059.GA8885@pengutronix.de>
	<20101220132147.GB16497@freescale.com>
	<20101220135535.GB8885@pengutronix.de>
Message-ID: <20101220142516.GC16497@freescale.com>
To: linux-arm-kernel@lists.infradead.org
List-Id: linux-arm-kernel.lists.infradead.org

Hi Wolfram,

On Mon, Dec 20, 2010 at 02:55:35PM +0100, Wolfram Sang wrote:
> 
> > > We should use asm/hardware/debug-pl01x.S, like the other users:
> > 
> > Just tested it, and it works.  Will do in v8.  Thanks for the comment.
> 
> Cool.
> 
> > > 
> > > >  arch/arm/mach-mxs/include/mach/uncompress.h  |   76 ++++++++++++++++++++++++++
> > > 
> > > Auto-detection could be used here, too?
> > > 
> > I'm afraid not.  How do you expect the detection is done, by reading
> > some register like we are doing in system.c?
> 
> I am not sure which register you mean in system.c?
> 
> > Which register specifically?
> 
> The signature of the amba device (base + size - 0x20). You can check
> drivers/amba/bus.c how it checks for the signature.
> 
> > If the register gets different address between
> > mx23 and mx28, we need to utilize the current mechanism to
> > distinguish the address anyway.
> 
> We could simply test both locations and see where it is present?
> 
Is there any problem of simply using arch_id here? I do not see
significant benefit of using the detection you suggest here
over the current way.

-- 
Regards,
Shawn