For the checkpoint, how do you manage those obsoleted "checkpoint". And when mount the flash, how do you know which is the newest version?
Thank you.
 
> From: manningc2@actrix.gen.nz
> To: yaffs@lists.aleph1.co.uk
> Subject: Re: [Yaffs] Yaffs2 Question
> Date: Mon, 28 Mar 2011 09:48:13 +1300
> CC: by_pacific@hotmail.com
> 
> On Saturday 26 March 2011 17:23:28 HuangAlex wrote:
> > Hi all,
> >
> > I have some questions.
> > 1. For yaffs2, I know there is a checkpoint strategy. The yaffs will
> > write the current status of the system into block before the next mount. So
> > I wonder do the yaffs rewrite that old chunks stored the information every
> > time or it will use new chunks for every ¡°checkpoint¡±? 
> 
> The yaffs allocation policy tends to partially randomise block use. THat means 
> the checkpoint will not always end up on the same blocks.
> 
> > 2. For the 
> > update file, can I said that any change in a file will lead to write a new
> > object header in the end? If not, what situation will lead to update a new
> > object header?
> 
> New object headers are written whenever:
> * file metadata is changed (file attributes, name etc).
> * modified file is closed/flushed.
> * directory contents change.
> 
> Garbage collection will also cause object headers to be rewritten.
> 
> 
>