[U-Boot-Users] RE: Nand OOB layout, u-boot and the kernel sou rces do not agree.. ??

[Woodruff, Richard]

The u-boot from CVS 4.4 has CONFIG_MTD_NAND_ECC_JFFS commented out by
default.  This results in NAND_NOOB_ECCPOS5 being defined for ECC.  The
header file sets defines this as position 5....right below the badblock pos
is set position 5.  An ECC value will be written to the bad block spot when
calculated...turn on JFFS2 and this problem goes away...however, something
is fishy in the bad_block and position 5 resolution code for some reason.
The codes addressing doesn't look like whats in the kernel tree.

At some point I'll submit a patch, but if someone else does it first that's
just fine.

Richard W.  

> -----Original Message-----
> From: Dave Ellis [mailto:dge at sixnetio.com] 
> Sent: Thursday, July 24, 2003 2:00 PM
> To: Woodruff, Richard
> Cc: u-boot-users at lists.sourceforge.net
> Subject: RE: [U-Boot-Users] RE: Nand OOB layout, u-boot and 
> the kernel sources do not agree.. ??
> 
> 
> Richard Woodruff wrote: 
> > Sent: Thursday, July 24, 2003 1:57 PM
> > > -----Original Message-----
> > > From: Woodruff, Richard
> > > Sent: Thursday, July 24, 2003 12:45 PM
> > > While trying to resolve what the OOB data layout should be I 
> > > see that the kernel headers as of 8-10-2002 have changed such 
> > > that both the NAND_JFFS2 and NAND_NOOB use position 5 for bad 
> > > block data.  The u-boot headers do not reflect this 
> > > change...doesn't this mean u-boot will be incompatible with 
> > > more recent kernels?  Should u-boot's headers be updated here?
> 
> Position 5 is where the chip makers mark bad sectors, so we 
> do not get a choice. The NAND_NOOB values in U-BOOT are wrong 
> and should be changed to match the new ones in the Linux 
> kernel. I think the original cmd_nand.c was based on a very 
> old version of MTD.
> 
> > ... Having more up to date definitions would seem better....as
> > raw nand doesn't seem to be well supported except with jffs2 &
> > possibly yaffs I don't suppose the NAND_NOOB is such a concern.
> 
> I don't know what software (if any) uses NAND_NOOB, but I 
> think the definitions still should be fixed (or removed). In 
> my patch to cmd_nand.c I hard coded the bad block position at 
> 5, so they can't be used as they are.
> 
> Dave
> 
> Dave Ellis
> ~~~~~~~~~~~~~~~~~~~~~~~~~~
> SIXNET - "Leading the Industrial Ethernet Revolution"
> 331 Ushers Road,   P.O. Box 767, Clifton Park, NY 12065 USA
> Tel +1 (518) 877-5173   Fax +1 (518) 877-8346
> Email me at: dge at sixnetio.com 
> Detailed product info: www.sixnetio.com 
> ~~~~~~~~~~~~~~~~~~~~~~~~~~
> 
>  
>