[U-Boot] [PATCH][v2] driver/nand: Update SRAM initialize logic for IFC.

[Prabhakar Kushwaha]

Hi Scott,


On 6/14/2014 2:10 AM, Scott Wood wrote:
> On Thu, 2014-06-12 at 12:14 +0530, Prabhakar Kushwaha wrote:
>> IFC controller v1.1.0 requires internal SRAM initialize by reading
>> NAND flash. Higher controller versions have provided "SRAM init" bit in
>> NCFGR register space.
>>
>> update SRAM initialize logic to reflect the same.
>>
>> Also print error message in case of Page read error.
>>
>> Signed-off-by: Prabhakar Kushwaha <prabhakar at freescale.com>
>> ---
>> Changes for v2:
>> 	- Updated error handling
>>
>>   drivers/mtd/nand/fsl_ifc_nand.c |   35 +++++++++++++++++++++++++++++++----
>>   include/fsl_ifc.h               |    2 ++
>>   2 files changed, 33 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/mtd/nand/fsl_ifc_nand.c b/drivers/mtd/nand/fsl_ifc_nand.c
>> index 27f5177..280e14e 100644
>> --- a/drivers/mtd/nand/fsl_ifc_nand.c
>> +++ b/drivers/mtd/nand/fsl_ifc_nand.c
>> @@ -806,12 +806,30 @@ static void fsl_ifc_select_chip(struct mtd_info *mtd, int chip)
>>   {
>>   }
>>   
>> -static void fsl_ifc_sram_init(void)
>> +static int fsl_ifc_sram_init(uint32_t ver)
>>   {
>>   	struct fsl_ifc *ifc = ifc_ctrl->regs;
>>   	uint32_t cs = 0, csor = 0, csor_8k = 0, csor_ext = 0;
>> +	uint32_t ncfgr = 0;
>>   	long long end_tick;
>>   
>> +	if (ver > FSL_IFC_V1_1_0) {
> It would be better to check that ver >= the first version that supports
> this, rather than > the last version that doesn't.

here only 2 type of code is present i.e.  (ver == FSL_IFC_V1_1_0) and 
(ver > FSL_IFC_V1_1_0).
if i put  (ver == FSL_IFC_V1_1_0) earlier a lots of code go in "if ". I 
am trying to avoid it to make code more cleaner.

> How much benefit is there in doing this versus continuing with the
> current code?

Existing code is perfect for 8KB internal SRAM as we are reading 8KB Page.
As IFC 2.0 has 16KB SRAM, either I have to read 16KB page or I have to 
read 2 8KB page to fill complete SRAM.

in future, there may be possibility of > 16KB internal SRAM. So I 
decided to use SRAM init bit.

> Should we determine the IFC version at compile time instead, so that we
> don't need to carry around both versions of the code in the binary?
>

Compile time option can be used. But in Linux driver it may not be 
possible.
I am trying to have same code base in u-boot and Linux for maintainability.

--prabhakar