RE: [Yaffs] About phsyical NAND FLASH cloning

<-- Warning: charset 'ks_c_5601-1987' is not supported -->

The process Brad wrote will work fine, mostly. It does, however, have one or two problems:

The image will include garbage as well (ie. Old data that has been deleted and not garbage collected). This, in itself, is not really a problem but it can mean that if you try progam it inot a device with more bad blocks than the one used to generate the image then you might run out of space.

A possibly better approach is to copy the yaffs file system into a as a set of files and use mkyaffsimage to generate an image file from that. This image will have no garbage in it.

So how do you detect bad blocks in the image? Look for blocks where the bad block marker in the spare area is zero.
You should also ignore empty blocks too.

The mkyaffs utility can "format" a NAND for yaffs (in other words just erase all non-bad blocks) and load an image. Even if you don't use this, reading the code will give you a good idea of what is involved (ie. What the NAND programming machine needs to do).


There are a few people using NAND loading machines to load NAND with yaffs images for cellphones etc.

-- Charles


> -----Original Message-----
> From: yaffs-admin@stoneboat.aleph1.co.uk
> [mailto:yaffs-admin@stoneboat.aleph1.co.uk] On Behalf Of
> noshel@idis.co.kr
> Sent: Thursday, 3 February 2005 1:38 p.m.
> To: Brad Beveridge
> Cc: yaffs@stoneboat.aleph1.co.uk
> Subject: Re: [Yaffs] About phsyical NAND FLASH cloning
>
>
> Brad Beveridge wrote:
>
> >Edward J. Lee(??????) wrote:
> >
> >
> >
> >>Hello,
> >>I'm using yaffs on a SAMSUNG 16MB flash chip.
> >>
> >>I've just heard that there is a device that can 'mass produce'
> >>programmed NAND chips, and just needs a 'bin' file.
> >>
> >>The machine has abilities to
> >>- write OOB data (which is included in the bin file),
> >>- check and skip bad blocks
> >>
> >>I guess it will work just fine for the parts that are written
> >>sequentially in raw binary, but I'm not sure about the
> yaffs part. So
> >>my question is : 'can a binary image of yaffs be extracted,
> and is it
> >>able to be dumped on a different NAND chip (same model, of course),
> >>just by writing the bin contents
> >>sequentially (skipping bad blocks)?'
> >>
> >>Thanks in advance.
> >>
> >>
> >>Regards,
> >>Ed
> >>
> >>_______________________________________________
> >>
> >>
> >>
> >>
> >Yes. Yaffs doesn't care about block order. You could place
> blocks from
> >one nand chip running yaffs, to another chip (as long as the page &
> >block sizes are the same) in any order that you like, and Yaffs will
> >scan and run fine. The only exception to this is that block 0 should
> >not be used, because for each yaffs partition block zero is reserved.
> >
> >Brad
> >
> >
> >_______________________________________________
> >yaffs mailing list
> >yaffs@stoneboat.aleph1.co.uk
> >http://stoneboat.aleph1.co.uk/cgi-bin/mailman/listinfo/yaffs
> >
> >
> >
> Hello, Brad,
> Wow, great. thx a lot.
> then here comes another question :
> 'How can a binary image of yaffs be obtained?'
> Should I hexdump the whole partition and search for blocks
> that seem dirty? Dumping the device wouldn't be hard,
> but the deciding of 'clean or dirty': I'm not sure how.
>
>
> _______________________________________________
> yaffs mailing list
> yaffs@stoneboat.aleph1.co.uk
> http://stoneboat.aleph1.co.uk/cgi-bin/mailman/listinfo/yaffs
>