FOUND IT + how to fix, Was: Using MTD + cfi_probe with RPXLite ?

[David Woodhouse]

AKC at pel.dk said:
> cfi_probe couldn't detect the flash on the RPXLite because it only has
> x32 access, which means A30,A31 isn't connected to the flash. So to
> detect the flash and get information from it all address' needs *4,
> eg.   0xAA becomes 0x2a8.

> After the flash is detected in cfi_probe_new_chip, all the address'
> needs *2  

> There must be a better way than touching up the code by hand ;-)

Try the rpxlite map driver which I've added to the CVS tree. If it doesn't 
work, let me have a patch for it.

There's also a 'physmap' driver for normal physical mappings of flash 
chips, which require ioremap() to get a virtual mapping - unlike nora.c 
which has a virtual mapping set up for it by the kernel's VM init code.

--
dwmw2




To unsubscribe, send "unsubscribe mtd" to majordomo at infradead.org