Re: [Yaffs] Problem with Yaffs on Routerboard RB112 and kern…

Top Page
Attachments:
Message as email
+ (text/plain)
Delete this message
Reply to this message
Author: ian@brightstareng.com
Date:  
To: yaffs
CC: yaffs, Andrea Conti
Subject: Re: [Yaffs] Problem with Yaffs on Routerboard RB112 and kernel 2.6.19
On Wednesday 09 May 2007 14:30, Andrea Conti wrote:
> This is the dump of a working mikrotik boot partition, created
> with "nanddump -ni" under 2.6.19.2
> It should contain a single file named "kernel"
>
> http://www.alyf.net/rb/boot-dump.bz2


$ hexdump -C boot-dump
00000000 01 00 00 00 01 00 00 00 ff ff 6b 65 72 6e 65 6c |..........kernel|
00000010 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 |................|
*
00000100 00 00 00 00 00 00 00 00 00 00 ff ff a4 81 00 00 |................|
00000110 00 00 00 00 00 00 00 00 ad 43 6d 38 ad 43 6d 38 |.........Cm8.Cm8|
00000120 ad 43 6d 38 00 00 00 00 ff ff ff ff ff ff ff ff |.Cm8............|
00000130 ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff |................|
*
000001c0 ff ff ff ff ff ff ff ff ff ff ff ff 00 00 00 00 |................|
000001d0 ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff |................|
*
000001f0 ff ff ff ff ff ff ff ff 00 00 00 00 00 00 00 00 |................|
00000200 00 00 10 00 00 ff 7f 01 ff f0 0f 70 81 aa 5a 97 |...........p..Z.|
00000210 7f 45 4c 46 01 01 01 00 00 00 00 00 00 00 00 00 |.ELF............|
00000220 02 00 08 00 01 00 00 00 40 00 28 80 34 00 00 00 |........@.(.4...|
00000230 44 fb 1f 00 01 10 00 50 34 00 20 00 03 00 28 00 |D......P4. ...(.|
00000240 14 00 13 00 00 00 00 70 70 73 19 00 70 93 29 80 |.......pps..p.).|
00000250 70 93 29 80 18 00 00 00 18 00 00 00 04 00 00 00 |p.).............|

The line starting at 00000200 is the oob/spare for the first page.
This contains 6 bytes of ECC, 1 byte of block-status (6th byte),
and 8 bytes of Yaffs PackedTags1 -- the question is, how is it ordered?
The ordering has to be consistent between readers and writers.

The autoplacement ordering in MTD is defined by the following in
nand_base.c:

static struct nand_oobinfo nand_oob_16 = {
        .useecc = MTD_NANDECC_AUTOPLACE,
        .eccbytes = 6,
        .eccpos = {0, 1, 2, 3, 6, 7},
        .oobfree = { {8, 8} }
};


Whether this ordering is used or not depend on the (oobinfo) argument passed
to MTD via the ioctl (from userspace) or the MTD driver read/write functions.
The current version of Yaffs(1) has the following:

#if (LINUX_VERSION_CODE < KERNEL_VERSION(2,6,18))
static struct nand_oobinfo yaffs_oobinfo = {
        .useecc = 1,        /* MTD_NANDECC_PLACE */
        .eccbytes = 6,
        .eccpos = {8, 9, 10, 13, 14, 15}
};
#endif


So that's a different ordering. And the mikrotik boot may have another?!

Now Charles is really good and decoding tags and guessing the right answer.
I think is looks like the Tags start at the beginning and MTD's ECC is at
the end -- so so perhaps the ordering from Yaffs' yaffs_oobinfo.

You could try using the yaffs_oobinfo as your default in MTD. Make sure
you do this experiment only on data you can afford to loose! There are
also some other orderings out there, some with the ECC bytes in a different
sequence.

One could take the page data (512 bytes) and compute the 6 ECC bytes,
then look for them in this sample -- this could reveal the answer.

-Ian