[RFC] BUG Power loss recovery issues in JFFS2

[Jörn Engel]

On Wed, 14 November 2007 15:17:36 -0500, David Woodhouse wrote:
> 
> That shouldn't happen. When one node obsoletes an earlier node covering
> the same range, we _should_ be checking the data CRC. I reimplemented
> this part of the code only recently.

Doesn't that fail for newly created files or for previously sparse
blocks of files?  Alexey's testcase creates a new file, so you should be
able to create such images yourself.  This testcase should be a bit more
reliable to hit the problem:

for i in `seq -w 1000000`; do echo 12345 > $i; done &
# wait half a sec
# hit reset button
cat `ls | sort | tail -1`

Completely untested, as always.

Jörn

-- 
All models are wrong. Some models are useful.
-- George Box
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://lists.infradead.org/pipermail/linux-mtd/attachments/20071115/df51c541/attachment.bin