Re: [PATCH] [JFFS2] load_xattr_datum need to return a positive number in case of unrecoverable error

[Jean-Christophe DUBOIS <jcd@tribudubois.net>]

On 01/05/2012 14:15, Artem Bityutskiy wrote:
> On Mon, 2012-04-30 at 22:54 +0200, Jean-Christophe DUBOIS wrote:
>> On 29/04/2012 17:44, Artem Bityutskiy wrote:
>>> On Thu, 2012-04-26 at 23:09 +0200, Jean-Christophe DUBOIS wrote:
>>>> Is this convincing enough?
>>> Sure, thanks.
>>>
>>>> You might think that the attibute corruption I am showing here is very
>>>> artificial but I can assure you that it does happen in the wild when you
>>>> get a power cut on your equipment while the GC was moving an attribute
>>>> node for example.
>>> This issue is for sure worth fixing.
>> So, will you accept this patch?
> No, I meant that I agree that there is an issue, but I think that it
> should be solved for all node types the same way. So I suggest to look
> at what happens when data nodes are corrupted, how JFFS2 picks the older
> node (if it does) and do it similarly for xattrs.

I just want to restate that this patch is just re-enabling a feature 
that was (wrongly I think) disabled a bit more than a year ago.

I will have a look at the other way you suggested but this might mean in 
the end a lot more restructuring of the JFFS2 code (at least the xattr 
part). As you seem to be very cautious about JFFS2 code change this will 
certainly mean a slow merge process. In the mean time the broken JFFS2 
xattr behavior will still be present.

JC