Thanks Charles.

I would like to explain my problem in detail!

I want to use YAFFS2 as flash file system.

I need to store two stage boot loader in NAND. Stage - 1 bootloader stored (First  4k code can be loaded in Internal RAM )  in block 0, which is responsible to copy the stage -2  boot loader from other portion of the NAND to SDRAM and pass control to it for execution. Stage - 2 boot loader is responsible initializing peripherals and booting up kernel image. 

Design Approach - 1 : Have  separate partitions for Stage -2 boot loader , kernel image and root file system.

Advantage:-  kernel images can be loaded using TFTP / FTP support available in boot loader.

Disadvantage:- If only 'n' number of blocks are allocated for loading kernel image and all the allocated blocks goes bad after some program / erase cycles then you can not load kernel image again!

Design Approach 2: have two partition boot and root. Bootloader will mount root partition and boot  the kernel image!

Advantage:- Kernel image program / erase is not restricted to some partition space!

Disadvantage: - As root file system can be mounted only as read only from boot loader. I can not reload kernel image from boot loader. I need to flash entire root file system again!

What efforts are required to mount root file system as read / write from boot loader?

How mkyaffs utility works? How it marks bad block when first time it is used to flash the images
on NAND flash?

--Rakhi

On 9/29/05, Charles Manning <manningc2@actrix.gen.nz> wrote:
On Thursday 29 September 2005 01:11, Rakhi Rathi wrote:
> Hi All,
>
> Is YAFFS boot loader code available under cvs directory 'direct' supports
> mounting of root file system in read / write mode?
This only allows read mode.

It is intended to support reading a boot image file, that is all.

-- Charles