Help on bad block table handling

Hi,

I have questions about the bad block table handling on nand flash with ext2. Do you update the bad block table when you have erase and program failures? The implementation that I am having creates the bad block table only when the flash is being formated. That is, there will be no updating of the bad block table until next formating of the flash.
Should I update the bad block table when erase and program failures are
detected? Does anyone know whether there is an implementation for that?
Please advise!! Thank you.

Best regards,
Eddy
eddy at emsoftltd.com