[U-Boot] [PATCH 01/13] mxc nand: Merge mtd and spl register definitions

[Scott Wood]

On 08/14/2012 03:37 AM, Stefano Babic wrote:
> On 13/08/2012 22:48, Benoît Thébaudeau wrote:
>> This patches fixes the TODO to use same register definitions in mtd mxc_nand and
>> nand_spl fsl nfc drivers.
>>
>> Signed-off-by: Benoît Thébaudeau <benoit.thebaudeau at advansee.com>
>> Cc: Scott Wood <scottwood at freescale.com>
>> Cc: Stefano Babic <sbabic at denx.de>
>> ---
> 
> Hi Benoît,
> 
> before going deeper in the review there is a general issue.
> 
> nand_spl was used in the past and it was the first implementation to
> have a first loader able to copy u-boot into RAM when the SOC boots from
> NAND. It is replaced by the general SPL framework. All new
> implementations must use the new SPL framework, What you see under
> nand_spl is referred to already supported boards.
> 
> With general SPL, we do not have two drivers for NAND, and we are able
> to boot from other devices, too. So it makes no sense to try to uniform
> the driver under nand_spl with the main driver, we want to use only the
> main driver.

It is not true that new-SPL will always have one uniform NAND driver,
nor is it true that nand_spl always has separate NAND drivers (it always
has a separate NAND subsystem, but the basic nand_spl/nand_boot.c does
use the standard device-specific driver).

It depends on space constraints of the target, and how big the standard
device-specific driver is.  With Wolfgang's mandate that no new boards
be added to nand_spl, we're going to have to support boards with a small
buffer for NAND booting in new SPL (even if it turns out this isn't one
of them).  It's not just 4xx; Freescale PPC chips are like this too,
including some that have not yet had NAND boot support added to mainline
U-Boot and thus can't use nand_spl.

-Scott