Re: RE : ONFI patch

[Matthieu CASTET <matthieu.castet@parrot.com>]

David Mosberger-Tang a écrit :
> Matthieu,
> 
> On Sat, Mar 16, 2013 at 12:21 PM, Matthieu Castet
> <matthieu.castet@parrot.com <mailto:matthieu.castet@parrot.com>> wrote:

>  
> 
>     > I see that the linux-mtd tree has a NAND_BUSWIDTH_AUTO features, but
>     > I don't see how that's supposed to help for, say,
>     Did you look at the omap example ?
> 
> 
> I don't see any users of NAND_BUSWIDTH_AUTO in the linux-mtd GIT tree.
>  Am I missing something?


Yes the patch was pending. You can find it here :
http://article.gmane.org/gmane.linux.ports.arm.omap/93502/match=
>  
> 
>     > nand_onfi-set_features() or nand_onfi_get_features() as
>     > nand_set_defaults() will eventually set the 16-bit transfer routines
>     > for 16-bit devices.
>     If that's a problem of nand_onfi_get/set_features why don't you fix
>     them to use only 8 bits transfert ?
> 
> 
> That's what my patch does.
>  
> 
>     Isn't there a patch already submitted for doing that ?
> 
> 
> I don't know.  I joined the mailing list only recently.  Can you point
> me to such a patch?
> 
http://thread.gmane.org/gmane.linux.drivers.mtd/45578/focus=45610

But it only fix the nand_onfi_get_features and I don't think it fix the column
shift.

BTW you can't use nand_write_buf/nand_read_buf in common code. Some controller
overide chip->read_buf/chip->write_buf with their version and don't expect
nand_write_buf/nand_read_buf to be called.


Matthieu