public inbox for linux-mtd@lists.infradead.org
 help / color / mirror / Atom feed
* compare JFFS2 vs YAFFS
@ 2003-04-19  6:03 Paul Wong
  2003-04-21 20:28 ` Charles Manning
  0 siblings, 1 reply; 3+ messages in thread
From: Paul Wong @ 2003-04-19  6:03 UTC (permalink / raw)
  To: linux-mtd

Hi All,
    could u tell me the JFFS2 and YAFFS comparsion? speed, save, reliable,
and future

Thanks

Paul

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: compare JFFS2 vs YAFFS
  2003-04-19  6:03 Paul Wong
@ 2003-04-21 20:28 ` Charles Manning
  0 siblings, 0 replies; 3+ messages in thread
From: Charles Manning @ 2003-04-21 20:28 UTC (permalink / raw)
  To: Paul Wong, linux-mtd

On Saturday 19 April 2003 18:03, Paul Wong wrote:
> Hi All,
>     could u tell me the JFFS2 and YAFFS comparsion? speed, save, reliable,
> and future
>
> Thanks
>
> Paul

Paul

This is a tricky one. Both YAFFS and JFFS2 have their pros and cons.

I have not used JFFS2 myself, but I have studied it a bit. Herewith my 
impressions.
* YAFFS is likely faster than JFFS2 in many situations. YAFFS does not do 
compression (which takes time). YAFFS has a simpler garbage collection 
mechanism that is likely to be faster.
* YAFFS uses smaller management structures (and therefore uses less RAM).
* JFFS2 provides compression. Nice if you have a small NAND. Some people 
achieve almost the same thing (for read) by loop mounting compressed file 
systems (eg. cramfs)  on a YAFFS file.
* You might already have JFFS2 for NOR, therefore you get an image footprint 
saving if you use it for NAND too.
* YAFFS is very reliable. I think JFFS2 is also.
* Both file systems are widely used and will continue to be supported. 

As a hunch, I'd say it is better to use JFFS2 on smaller NAND arrays where 
you gain from the compression and YAFFS on larger sizes.

My suggestion to you is that you try both and go with the one that is the 
best fit for your needs.

-- Charles

^ permalink raw reply	[flat|nested] 3+ messages in thread

* RE: compare JFFS2 vs YAFFS
       [not found] <20030423202700.0DEF715788@desire.actrix.co.nz>
@ 2003-04-24  9:30 ` Stephan Linke
  0 siblings, 0 replies; 3+ messages in thread
From: Stephan Linke @ 2003-04-24  9:30 UTC (permalink / raw)
  To: Linux-Mtd, manningc2

Hi Charles,

Indeed I meant the over head that comes from reserved areas for garbage collection etc.
You say: "Since even the smallest NAND device holds many hundred blocks this is generally not an issue."
But if you are going to create a verry small NAND partition of a few hundred kilobyte this may be become an issue.

Thanks for the info,

Stephan

> -----Original Message-----
> From: Charles Manning [mailto:manningc2@actrix.gen.nz]
> Sent: Mittwoch, 23. April 2003 22:28
> To: Stephan Linke
> Subject: Re: compare JFFS2 vs YAFFS
> 
> 
> 
> >
> > How about the "out-take" that JAFFS2 requires for garbage collection? I
> > guess it is X times mtd->erasesize (X=2..5)? What's the "out-take" of
> > YAFFS?
> > I think this is the most importnat factor if you are going to use YAFFS or
> > JFFS2 on a small NAND partition.
> 
> 
> Can you explain what you mean by out-take a bit better? Do you mean 
> "overhead"?
> 
> If so, there are two types of overhead:
> 
> * NAND space. This is run-time configurable. YAFFS normally uses a reserve 
> space of 5 blocks (ie. 5x16kB), but should work fine with 2 blocks. 5 blocks 
> just gives extra comfort for blocks going bad at the same time as garbage 
> collection. Since even the smallest NAND device holds many hundred blocks 
> this is generally not an issue.
> 
> * Time: YAFFS does not stop for a long time while it does garbage collection. 
> The worst case is just the time to erase and rewrite a block (ie approx 
> 7milliseconds).
> 
> Another area where YAFFS is good is boot time. Systems with 512Mbytes of NAND 
> usually boot within 1 minute. There are plans to reduce this to a few seconds.
> 
> 
> -- Charles
> 
> 
> 
> 

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2003-04-24  9:30 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
     [not found] <20030423202700.0DEF715788@desire.actrix.co.nz>
2003-04-24  9:30 ` compare JFFS2 vs YAFFS Stephan Linke
2003-04-19  6:03 Paul Wong
2003-04-21 20:28 ` Charles Manning

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox