Mask error in nandbase.c: nand_scan(): oob_size

Thomas Gleixner tglx at linutronix.de
Thu Aug 4 12:21:09 EDT 2005


On Thu, 2005-08-04 at 09:54 -0600, Nathan H. wrote:
> >Please look into
> >
> >http://www.samsung.com/Products/Semiconductor/Flash/NAND/16Gbit/K9WAG08U1M/ds_k9wag08u1m_rev02.pdf
> >
> >to see why the bits are reserved are why they are guaranteed to be zero.
> 
> I understand that currently the extra bit is going to be zero, but that bit 
> is part of a completely different field. Someday one of those reserved 
> values may be used at which point things will break. The fact is that the 
> field is 1-bit wide and thus the mask should be 0x1 not 0x3.

Sorry. The field is 2 (two) bits wide for _ANY_ FLASH which supports the
extended ID. The code looks at bit 0 and bit 1 of 4th ID data. Why do
you insist that it is a 1-bit field ?

>From the datasheet you referred to

        I/O1  I/O0
1K       0      0
2k       0      1
reserved 1      0
reserved 1      1

>From the datasheet I referred to

        I/O1  I/O0
1K       0      0
2k       0      1
4k       1      0
8k       1      1

> >No error to see here :)
> 
> See no evil, hear no evil, eh? ;)

I still do not see one :)

tglx






More information about the linux-mtd mailing list