public inbox for linux-mtd@lists.infradead.org
 help / color / mirror / Atom feed
* Cramfs on NAND
       [not found] <E1FErAE-00004u-2M@canuck.infradead.org>
@ 2006-03-03  1:37 ` Terence Soh
  2006-03-10 13:06   ` Josh Boyer
  0 siblings, 1 reply; 5+ messages in thread
From: Terence Soh @ 2006-03-03  1:37 UTC (permalink / raw)
  To: linux-mtd

Hi again,

I hear that cramfs on NAND flash is not a good idea because of the bad 
blocks. Is this still true? There was talks some time  ago about making 
cramfs work on top on NAND. 
<http://lists.infradead.org/pipermail/linux-mtd/2004-May/009784.html>

If this is true, are there any recommendations of what other fs to use. 
I already have one partition of jffs2 to store read-write files.

Thanks again,
Terence.

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: Cramfs on NAND
  2006-03-03  1:37 ` Cramfs on NAND Terence Soh
@ 2006-03-10 13:06   ` Josh Boyer
  2006-03-10 17:36     ` Russ Dill
  0 siblings, 1 reply; 5+ messages in thread
From: Josh Boyer @ 2006-03-10 13:06 UTC (permalink / raw)
  To: Terence Soh; +Cc: linux-mtd

On 3/2/06, Terence Soh <gopher@singnet.com.sg> wrote:
> Hi again,
>
> I hear that cramfs on NAND flash is not a good idea because of the bad
> blocks. Is this still true? There was talks some time  ago about making
> cramfs work on top on NAND.
> <http://lists.infradead.org/pipermail/linux-mtd/2004-May/009784.html>

Yes, it's still true.  CRAMFS expects it's data to be contiguous so
unless you have something similar to a flash translation layer that
hides the bad blocks, insanity will ensue.

> If this is true, are there any recommendations of what other fs to use.
> I already have one partition of jffs2 to store read-write files.

Could you use a read-only JFFS2 partition?

josh

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: Cramfs on NAND
  2006-03-10 13:06   ` Josh Boyer
@ 2006-03-10 17:36     ` Russ Dill
  2006-03-13  0:34       ` Terence Soh
  2006-03-18  2:43       ` Sean Kelley
  0 siblings, 2 replies; 5+ messages in thread
From: Russ Dill @ 2006-03-10 17:36 UTC (permalink / raw)
  To: Josh Boyer; +Cc: linux-mtd, Terence Soh

> > I hear that cramfs on NAND flash is not a good idea because of the bad
> > blocks. Is this still true? There was talks some time  ago about making
> > cramfs work on top on NAND.
> > <http://lists.infradead.org/pipermail/linux-mtd/2004-May/009784.html>
>
> Yes, it's still true.  CRAMFS expects it's data to be contiguous so
> unless you have something similar to a flash translation layer that
> hides the bad blocks, insanity will ensue.
>
> > If this is true, are there any recommendations of what other fs to use.
> > I already have one partition of jffs2 to store read-write files.
>
> Could you use a read-only JFFS2 partition?

The other solution is to use an ftl ontop of the flash, and then put a
cramfs on that. Of course, that depends on you living somewhere where
algorithms are not restricted to those with the license to use them.
(iirc)

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: Cramfs on NAND
  2006-03-10 17:36     ` Russ Dill
@ 2006-03-13  0:34       ` Terence Soh
  2006-03-18  2:43       ` Sean Kelley
  1 sibling, 0 replies; 5+ messages in thread
From: Terence Soh @ 2006-03-13  0:34 UTC (permalink / raw)
  To: Russ Dill, Josh Boyer; +Cc: linux-mtd

Hi,

>>Yes, it's still true.  CRAMFS expects it's data to be contiguous so
>>unless you have something similar to a flash translation layer that
>>hides the bad blocks, insanity will ensue.
FTL will not work for me because of the licensing issues :(.

>>Could you use a read-only JFFS2 partition?
Besides the read-only characteristic, I'm also looking at fast mount 
time. I guess I should take a look at YAFFS.

Question: Does a read-only JFFS2/YAFFS partition means that there will 
not be any wear-levelling performed?

> 
> 
> The other solution is to use an ftl ontop of the flash, and then put a
> cramfs on that. Of course, that depends on you living somewhere where
> algorithms are not restricted to those with the license to use them.
> (iirc)
> 

Bye and thanks for your time,
Terence.

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: Cramfs on NAND
  2006-03-10 17:36     ` Russ Dill
  2006-03-13  0:34       ` Terence Soh
@ 2006-03-18  2:43       ` Sean Kelley
  1 sibling, 0 replies; 5+ messages in thread
From: Sean Kelley @ 2006-03-18  2:43 UTC (permalink / raw)
  To: linux-mtd

So is a NAND based FTL for linux encumbered with patents/licensing by M-Systems?

Sean

On 3/10/06, Russ Dill <russ.dill@gmail.com> wrote:
> > > I hear that cramfs on NAND flash is not a good idea because of the bad
> > > blocks. Is this still true? There was talks some time  ago about making
> > > cramfs work on top on NAND.
> > > <http://lists.infradead.org/pipermail/linux-mtd/2004-May/009784.html>
> >
> > Yes, it's still true.  CRAMFS expects it's data to be contiguous so
> > unless you have something similar to a flash translation layer that
> > hides the bad blocks, insanity will ensue.
> >
> > > If this is true, are there any recommendations of what other fs to use.
> > > I already have one partition of jffs2 to store read-write files.
> >
> > Could you use a read-only JFFS2 partition?
>
> The other solution is to use an ftl ontop of the flash, and then put a
> cramfs on that. Of course, that depends on you living somewhere where
> algorithms are not restricted to those with the license to use them.
> (iirc)
>
> ______________________________________________________
> Linux MTD discussion mailing list
> http://lists.infradead.org/mailman/listinfo/linux-mtd/
>

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2006-03-18  2:43 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
     [not found] <E1FErAE-00004u-2M@canuck.infradead.org>
2006-03-03  1:37 ` Cramfs on NAND Terence Soh
2006-03-10 13:06   ` Josh Boyer
2006-03-10 17:36     ` Russ Dill
2006-03-13  0:34       ` Terence Soh
2006-03-18  2:43       ` Sean Kelley

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