Dear Stefan Roese,
In message 200809251653.30832.sr@denx.de you wrote:
This will not work. Take a look at the lwmon5 configuration. It has a similar Intel FLASH which has two chips inside:
#define CFG_FLASH0 0xFC000000 #define CFG_FLASH1 0xF8000000 #define CFG_FLASH_BANKS_LIST { CFG_FLASH1, CFG_FLASH0 }
#define CFG_MAX_FLASH_BANKS 2
...
You need to change the addresses for your 32MB version though.
Please note that I really dislike such static configurations which will break as soon as the size of the flash banks changes.
Right from the begining U-Boot was always designed to allow for dynamic adjustment to the actual memory sizes found on the boards, and we should just use this capability.
Assuming a 32 bit configuration, above should rather be written as:
#define CFG_FLASH0 (0 - flash_info[0].size) #define CFG_FLASH1 (CFG_FLASH0 - flash_info[1].size) #define CFG_FLASH_BANKS_LIST { CFG_FLASH1, CFG_FLASH0 } ...
This may require a preliminary mapping of the flash banks that is then fixed after the CFI driver determined the real size; see for example how this was done for the TQM8xx* boards (of course this needs to be adapted for 4xx, but you get the idea).
Best regards,
Wolfgang Denk