Bug in nand_select_chip?

[llandre]

> > I fixed the driver as shown in the NAND Flash document:
> > - nCE pin low when hwcontrol is invoked with NAND_CTL_SETNCE
> > - nCE pin high when hwcontrol is invoked with NAND_CTL_CLRNCE
> > and now the driver works with the new MTD code.
>
>That's the reason why documents are written. :)
>
>But the driver should work with the old code too, as this never changed since
>we started NAND support in JFFS2.

In my understanding the driver does not work with the old code because
the old nand_select_chip() implementation called this->hwcontrol(mtd, 
NAND_CTL_CLRNCE),
with chip==0, setting the nCE pin high.

By the way, as described in the mentioned document, I send you the driver 
source code in order to
include it in the public source tree. It supports the chip K9F5608U0B 
(32MB) by Samsung, mounted on
the PPChameleonEVB board. To build it, it is necessary to add the following 
line in the Makefile:
obj-$(CONFIG_MTD_NAND_PPCHAMELEONEVB)   += ppchameleonevb.o

and the following lines in the Config.in:

if [ "$CONFIG_PPCHAMELEONEVB" = "y" ]; then
    dep_tristate '  NAND Flash device on PPChameleonEVB board' 
CONFIG_MTD_NAND_PPCHAMELEONEVB $CONFIG_MTD_NAND
fi

We tested it with the kernel 2.4.20 included the linuxppc_2_4_devel tree 
hosted by Denx Software
Engineering (www.denx.de).



Best regards,

llandre 
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: ppchameleonevb.c
Url: http://lists.infradead.org/pipermail/linux-mtd/attachments/20040318/ab959061/attachment.c