[RESUBMIT] [PATCH] [MTD] NAND nand_ecc.c: rewrite for improved performance

[Troy Kisky]

David Woodhouse wrote:
> On Fri, 2008-08-15 at 12:04 +0200, Frans Meulenbroeks wrote:
>> 2008/8/15, David Woodhouse <dwmw2 at infradead.org>:
>>> On Fri, 2008-08-15 at 11:23 +0200, Frans Meulenbroeks wrote:
>>>  > 2008/8/15, David Woodhouse <dwmw2 at infradead.org>:
>>>
>>> No need -- if you've thought about it and believe it should work, that's
>>>  probably enough. I just saw some 'unsigned long' data types, which are
>>>  going to have a different size between 32-bit and 64-bit systems, and
>>>  wondered if that would introduce differences.
>> I was unaware of that. For me unsigned long is more or less a synonym
>> for 32 bit. Maybe I'm just getting too old :-(
>> Anyway, if you have a suggestion for a better type, I'll happily change things.
>> Would uint32_t be better?
> 
> If it needs to be 32-bit, then yes -- uint32_t is the correct type to
> use.
> 
> If 64-bit is OK, then 'unsigned long' is likely to be more efficient on
> some platforms.
> 

You might also test it on a big endian system to be safe.

Troy