From mboxrd@z Thu Jan  1 00:00:00 1970
From: Hong Xu <hong.xu@atmel.com>
Date: Mon, 01 Aug 2011 16:13:59 +0800
Subject: [U-Boot] [PATCH 1/2 V3] AT91: Makes AT91SAM9RL SoC build
 correctly against u-boot-atmel/master
In-Reply-To: <1312185615.3635.105.camel@ubuntu>
References: <1308107471-22295-1-git-send-email-hong.xu@atmel.com>
	<1312185615.3635.105.camel@ubuntu>
Message-ID: <4E366047.3050908@atmel.com>
List-Id: <u-boot.lists.denx.de>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: u-boot@lists.denx.de

Hi Reinhard,

On 08/01/2011 04:00 PM, Reinhard Meyer wrote:
> Dear Hong Xu,
>  > Rework for AT91SAM9RL SoC, makes it build again.
>  > arch/arm/include/asm/arch-at91/at91sam9rl_matrix.h | 121 +++++---------
>
> When I compare at91sam9rl_matrix.h to at91sam9260_matrix.h there is only
> one subtle difference:
>
> at91sam9260_matrix.h:
>
> struct at91_matrix {
> u32 mcfg[16]; /* Master Configuration Registers */
> u32 scfg[16]; /* Slave Configuration Registers */
> u32 pras[16][2]; /* Priority Assignment Slave Registers */
> u32 mrcr; /* Master Remap Control Register */
> u32 filler[0x06];
> u32 ebicsa; /* EBI Chip Select Assignment Register */
>
> at91sam9rl_matrix.h:
>
> struct at91_matrix {
> u32 mcfg[16]; /* Master Configuration Registers */
> u32 scfg[16]; /* Slave Configuration Registers */
> u32 pras[16][2]; /* Priority Assignment Slave Registers */
> u32 mrcr; /* Master Remap Control Register */
> u32 filler[7];
> u32 ebicsa; /* EBI Chip Select Assignment Register */
> };
>
> 1. Is it really filler[7] for the 9rl - can you verify with the data
> sheet?

Confirmed. ebicsa is at 0x120 for SAM9RL, but 0x11c for SAM9260.

> 2. It seems there is a possibility to unify all those _matrix files into
> one in the future (we are not going to do this now!) ?

There are many differents among chips. For example, on SAM9260/SAM9G10

struct at91_matrix {
         u32     mcfg;   /* Master Configuration Registers */
         u32     scfg[5];        /* Slave Configuration Registers */
         u32     filler[6];
         u32     ebicsa;         /* EBI Chip Select Assignment Register */
};

And for newer chips, some new fields are added. Can we keep current 
scheme and then have a look to see whether we have an elegant way to 
handle this? BTW, we're going to add some new chips soon.

Thanks.

BR,
Eric