[U-Boot-Users] [PATCH 1/1] Add support for ATMELAT91SAM9G20EKboard

[Jean-Christophe PLAGNIOL-VILLARD]

> > 
> 
> As I see it, CONFIG_AT91 would mean that you have a 
> a certain class of peripherals which is developed for the 
> AT91 range of processors (and is used by the AVR32 as well)
> The name is probably slightly misleading, but convenient.
> 
> Not having this definition, will soon mean that you have statements
> like the following:
> 
> #if  defined(CONFIG_AT91RM9200) || \
> defined(CONFIG_AT91SAM9260) || \
> defined(CONFIG_AT91SAM9261) || \
> defined(CONFIG_AT91SAM9263) || \
> defined(CONFIG_AT91SAM9G10) || \
> defined(CONFIG_AT91SAM9G15) || \
> defined(CONFIG_AT91SAM9G20) || \
> defined(CONFIG_AT91SAM9G41) || \
> defined(CONFIG_AT91SAM9M10) || \
> defined(CONFIG_AT91SAM9M11) || \
> defined(CONFIG_AT91CAP9) || \
> defined(CONFIG_AT572D940)
> 
> This statement will have to be updated every time a new MCU is released.
> 
> Having the statment 
> #if  defined(CONFIG_AT91)
> 
> should definitely reduce the maintenance of drivers.
I fully agree with Ulf, we need to have a better CONFIG, but I'm not
sure that CONFIG_AT91 is the right one

We have actually in the tree the following CPU
CONFIG_AT91CAP9
CONFIG_AT91RM9200
CONFIG_AT91SAM9260
CONFIG_AT91SAM9261
CONFIG_AT91SAM9263
CONFIG_AT91SAM9RL

and as mention Ulf are going to have more and more

maybe we can use config related to the functionnality or the sub-class
CPU
ex :
CONFIG_MACB_INCLK

Best Regards,
J.