A while ago we got the RB532 NAND support going, but as I recall while it
worked on my old 532 there was a problem with the YAFFS support and some newer
cards did not work properly.

I am currently working on the RB112 NAND support, and I may be having a 
similar problem, but not having followed the resolution to the original 
problem very closely I am unsure.

The symptom that I am getting is that the NAND support finds the chip,
recognises it for what it is, and defines the two partitions.  /proc/mtd
and the like are all set up correctly.  When the NAND module loads it does
a bad block scan, and reports no bad blocks.

Using the YAFFS from generic-2.6 I then try to mount the partitions, and
they complain that every single YAFFS block is bad, resulting in zero free
space on the disk.  I mount the partitions using:-

mount -t yaffs /dev/mtdblock? /mnt

Looking in the YAFFS code, this happens in yaffs_guts.c at line 5300 and is
because it things that state == YAFFS_BLOCK_STATE_DEAD.  Quite how it got here
I do not know.

What I am trying to find out is if this is a repeat (or minor variation) or
the original problem, or whether this is a new problem.

Any ideas?

David
_______________________________________________
openwrt-devel mailing list
[email protected]
http://lists.openwrt.org/cgi-bin/mailman/listinfo/openwrt-devel

Reply via email to