persistence of files using jffs

persistence of files using jffs

[Juan Gonzo]

Me again.  More questions.  Oy!

Our system is using a ramdisk (with a prebuilt image)
as its root device.  Before I boot up the system I burn
a small image into flash (this image was created with
mkfs.jffs and contains a couple of files and one directory).

So, the system boots up and I do the following:
   (1) mknod /dev/mtd3 b 31 3   (mtd3 is where the fs 'partition' is,
                                 using examples from nora.c)
   (2) mount -t jffs /dev/mtd3 /mnt

The mount is successful and after this I can see the files and
directory I burned into flash.  I can copy another file from the
ramdisk to this mount.  And I can 'umount /mnt' successfully as
well.

However, when I cycle power on the system,
any *new* files I create or copy into this space are lost.

Everytime I restart the system I have to 'mknod /dev/mtd3', does
this have something to do with that?  What am I missing to
get the new files to be present when I remount JFFS?

Thanks,
   -- kelly



_______________________________________________________
Do You Yahoo!?
Get your free @yahoo.ca address at http://mail.yahoo.ca


To unsubscribe, send "unsubscribe mtd" to majordomo@infradead.org

Re: persistence of files using jffs

[David Woodhouse]

kayemmess@yahoo.ca said:
>    (1) mknod /dev/mtd3 b 31 3   (mtd3 is where the fs 'partition' is,
>                                  using examples from nora.c)

The block devices should be /dev/mtdblock<n>. /dev/mtd3 should be a 
character device. But the names don't really matter.


kayemmess@yahoo.ca said:
>  And I can 'umount /mnt' successfully as well.
> However, when I cycle power on the system, any *new* files I create or
> copy into this space are lost. 

Hmmm. Are they there when you unmount and remount? Can you make sure you're 
using the latest version from CVS, and enable CONFIG_JFFS_FS_VERBOSE?



--
dwmw2




To unsubscribe, send "unsubscribe mtd" to majordomo@infradead.org

Re: persistence of files using jffs

[Juan Gonzo]

When umount and then remount, any new files I copied are
not there.

Our system has 4 Intel Strataflash chips that are coupled pairs
to give us 32MB at 32bits data width.  In the structure 'map_info',
I set the 'buswidth' field to 4.

The startup code uses the write32 functions, but when I mount 
I noticed that write16 functions are being used.  As well, in 
the file cfi_cmdset_0001.c, only write16 functions are called.  
Is this causing a problem for our system?

Thanks,
   -- kelly


--- David Woodhouse <dwmw2@infradead.org> wrote:
> 
> kayemmess@yahoo.ca said:
> >    (1) mknod /dev/mtd3 b 31 3   (mtd3 is where the fs 'partition'
> is,
> >                                  using examples from nora.c)
> 
> The block devices should be /dev/mtdblock<n>. /dev/mtd3 should be a 
> character device. But the names don't really matter.
> 
> 
> kayemmess@yahoo.ca said:
> >  And I can 'umount /mnt' successfully as well.
> > However, when I cycle power on the system, any *new* files I create
> or
> > copy into this space are lost. 
> 
> Hmmm. Are they there when you unmount and remount? Can you make sure
> you're 
> using the latest version from CVS, and enable CONFIG_JFFS_FS_VERBOSE?
> 
> 
> 
> --
> dwmw2
> 
> 
> 
> 
> To unsubscribe, send "unsubscribe mtd" to majordomo@infradead.org


_______________________________________________________
Do You Yahoo!?
Get your free @yahoo.ca address at http://mail.yahoo.ca


To unsubscribe, send "unsubscribe mtd" to majordomo@infradead.org

Re: persistence of files using jffs

[David Woodhouse]

On Fri, 14 Jul 2000, Juan Gonzo wrote:

> Our system has 4 Intel Strataflash chips that are coupled pairs
> to give us 32MB at 32bits data width.  In the structure 'map_info',
> I set the 'buswidth' field to 4.
> 
> The startup code uses the write32 functions, but when I mount 
> I noticed that write16 functions are being used.  As well, in 
> the file cfi_cmdset_0001.c, only write16 functions are called.  
> Is this causing a problem for our system?

Almost certainly. That geometry isn't supported yet, for the simple reason
that I haven't actually met a board which does it yet.

Copy the cfi_intelext_*_1_by_16 routines and make 2_by_16 versions. It
shouldn't be difficult. We also need to check the geometry in
cfi_intelext_setup() and make the MTD device's methods point to the right
set of functions.

-- 
dwmw2




To unsubscribe, send "unsubscribe mtd" to majordomo@infradead.org