On Tue, Jul 28, 2009 at 10:42 PM, Jean-Christophe PLAGNIOL-VILLARDplagnioj@jcrosoft.com wrote:
On 07:58 Tue 28 Jul , Kyungmin Park wrote:
On Tue, Jul 28, 2009 at 5:09 AM, Jean-Christophe PLAGNIOL-VILLARDplagnioj@jcrosoft.com wrote:
On 14:28 Fri 24 Jul , Kyungmin Park wrote:
On Fri, Jul 24, 2009 at 7:42 AM, Jean-Christophe PLAGNIOL-VILLARDplagnioj@jcrosoft.com wrote:
On 09:55 Tue 21 Jul , Kyungmin Park wrote:
S5PC100 has own OneNAND controller and has different interface. OneNAND IPL use it to S5PC100 board.
Signed-off-by: Kyungmin Park kyungmin.park@samsung.com
is there any better than put soc specific code in generic implementatioN
I hope so. only s3c64xx series and s5pc100 use own OneNAND controller. I also don't understand why need these controller. The OneNAND has its own controller at chips already.
I known so what do you propose?
just commit the patch I sent. It's difficult to remove the ifdef since size limitation.
I understand your problem of size but is there any otherway to do ti without the onenand soc controler ?
No, no way we should use soc controller.
I'll prefer we find a way to not put soc specific code in the generic code maybe we can crete a header which will be soc specific or a generic that we will include depending on the soc
something like weak alias?
The new problem is now I want to support S5PC100 & S5PC110 Samsung SoC simultaneously. But these has different interface. c100 has CMD_MAP method, c110 has similar generic but different. At runtime it detects the cpu id and determine which read function is used. At this case, s5pc100/c110 has no size limitation because of internal ROM code. At boot it loads 16KiB and 8KiB data to internal RAM respectively.
Okay I will try to make it generic and resend it after finishing the port s5pc110 works
Thank you, Kyungmin Park
Best Regards, J.