Dear York Sun,
In message 1365195056-20188-1-git-send-email-yorksun@freescale.com you wrote:
When we copy code/data to the main memory, we may need to flush the cache if required by architecture. It uses the existing function flush_cache. Syntax is
flush <addr> <size>
Plain "flush" is way too generic a name. I think we should make it clear from the command invocation that we are dealing with caches here.
Actually I think we should not even use a new command for this - we already have the "dcahe" and "icache" commands for this purpose.
What do you think about implementiung this as a subcommand to "dcache"? Something like:
So far:
dcache on dcache off
adding new:
dcache flush => flush all dcache flush <addr> <size> => flush range
I think this makes more sense. Comments?
Best regards,
Wolfgang Denk