[RESEND PATCH 2.6.26-git 1/2] MTD DataFlash: bugfix, binary page sizes now handled (v3)

[Andrew Morton]

On Fri, 25 Jul 2008 15:49:12 -0700
David Brownell <david-b at pacbell.net> wrote:

> This fixes a bug in handling certain newer DataFlash chips (at45db321d
> and at45db642d have current Linux users) when they were configured to
> use binary page sizes (possibly at the factory).
> 
>  - Use JEDEC probing (borrowed from m25p80 driver) to tell if the chip
>    supports binary page sizes ... if so, then query the chip to tell
>    whether it's currently using them.
> 
>  - Print that pagesize during probe; erasesize is still listed in 
>    the /proc/mtd file, useful for making filesystems.
> 
> Also fix a pre-existing bug with 2 MBit parts using the wrong pagesize;
> these are a bit smaller than Linux would normally use (even shared with
> a small FPGA's configuration bitstream).

I'm not sure what kernel this is against, but it gets a reject storm
against 2.6.27-rc1.  As does [2/2].