Hey,
I know the JFFS2 filesystem underneath has support for bad blocks,
however do I have to turn it on or is enabled by default? The options flag
seems to indicate otherwise.
Also, here is what I mean by skipping bad blocks
I know nandwriter during its initial burning of uboot+ubl to nand, just
crashes when it encounters a bad block during erase..
The correct approach is to build a BBT table and update it with this
location of the bad block...a solution I was thinking of was a physical to
logical mapping table that would allow for address translation to handle
corrupt blocks, sorta like hard drives
That way, the table will be updated so that uboot and ubl will be written
seamlessly past the corrupt block...
Later during ubl bootup, it would read the BBT table when copying u-boot out
of NAND, and ensure the bad block is skipped..
And also, later on during kernel boot, it would take care of the BBT table
if it detects a bad block that the previous two methods didn't notice...this
could be handled by the JFFS2 system
Either way, nandwriter and ubl need to be updated so that they can handle
bad blocks and can copy the appropriately mapped data from NAND to DDR2
The kernel also needs to know that so that when you burn the jffs2+kernel
uImage to nand using ftl_format and nandwrite, it can use the same
convention to ensure bad blocks are skipped when writing...currently,
ftl_format just crashes on encountering bad block
I don't know about nandwrite, don't have source for it
If this doesn't convey my intended use, please clarify, I'll try to explain
better hehe
Jerry
_______________________________________________
Davinci-linux-open-source mailing list
[email protected]
http://linux.davincidsp.com/mailman/listinfo/davinci-linux-open-source
