Josh Boyer | 12 Aug 03:54 2003
Picon

[patch] 2.6.0-test3 fs/jffs/inode-v23.c

This patch fixes a compiler warning due to an incorrect structure type
for a function parameter.  Not sure if jffs is supported anymore, but I
thought I would send this anyway.

jb

--- linux-2.6.0-test3.orig/fs/jffs/inode-v23.c	2003-07-27
20:46:52.000000000 -0500
+++ linux-2.6.0-test3/fs/jffs/inode-v23.c	2003-08-11 20:45:38.382051048
-0500
 <at>  <at>  -383,7 +383,7  <at>  <at> 

 /* Get statistics of the file system.  */
 int
-jffs_statfs(struct super_block *sb, struct statfs *buf)
+jffs_statfs(struct super_block *sb, struct kstatfs *buf)
 {
 	struct jffs_control *c = (struct jffs_control *) sb->s_fs_info;
 	struct jffs_fmcontrol *fmc;

jffs | 12 Aug 04:38 2003

Can I use whole flash chip in one mtd partition for jffs2 filesystem?

I have two flash chips, I want use second chip to store my application data, 
make it jffs2 filesytem. The chip has 2M bytes and the sectors is:
 16K(1) 8K(2) 32K(1) 64K (31)
I made mtd partion: 

static struct mtd_partition my_partitions[] = {
       {
               name: "user (2048K)",
               size: 0x200000,
               offset: 0x0
       }
}; 

When i erase /dev/mtd0, found :
the numEraseRegions is 0 

=================================
the debug information: 

MTD_open minor=0
MTD_open devnum=0
MTD_open numeraseregions=0
Erase Total 1 Units
MTD_ioctl cmd=80044d07
numeraseregions=0
MTD_ioctl cmd=801e4d01
PerformTing Flash Erase of length 65536 at offset 0x0
done
MTD_close
=================================== 
(Continue reading)

Alex Shnitman | 12 Aug 18:08 2003
Picon

Using JFFS on a block device?

Hi,

I have a board which has a CompactFlash slot, and it emulates a hard
disk (both the BIOS and Linux see it that way). I'd like to use JFFS on
it, mostly because it offers wear-leveling -- without that I'm afraid
that the flash will die pretty quickly. It it possible at all to do
something like that?

Or perhaps you have a better idea how to handle this situation -- using
the flash as a hard disk, and having some kind of wear-leveling
protection at the same time?

Thanks!

--

-- 
Alex Shnitman <alexsh <at> hectic.net>
http://www.hectic.net/   UIN 188956
PGP 0xEC5D619D / E1 F2 7B 6C A0 31 80 28  63 B8 02 BA 65 C7 8B BA

To unsubscribe from this list: send the line "unsubscribe jffs-dev" in
the body of a message to majordomo <at> axis.com

Pieter Grimmerink | 12 Aug 21:32 2003
Picon

RE: Using JFFS on a block device?


> -----Original Message-----
> From: owner-jffs-dev <at> axis.com [mailto:owner-jffs-dev <at> axis.com]On Behalf
> Of Alex Shnitman
> Sent: dinsdag 12 augustus 2003 18:08
> To: jffs-dev <at> axis.com
> Subject: Using JFFS on a block device?

> I have a board which has a CompactFlash slot, and it emulates a hard
> disk (both the BIOS and Linux see it that way). I'd like to use JFFS on
> it, mostly because it offers wear-leveling -- without that I'm afraid
> that the flash will die pretty quickly. It it possible at all to do
> something like that?

No, JFFS won't provide wear-levelling on a device that behaves like a
harddrive.

JFFS(2) is for 'raw' flash only.

> Or perhaps you have a better idea how to handle this situation -- using
> the flash as a hard disk, and having some kind of wear-leveling
> protection at the same time?

CompactFlashes usually have their own internal wear-levelling.
SanDisk seems to work quite good, if you have to use CompactFlash.

Regards,

Pieter

(Continue reading)

Jörn Engel | 18 Aug 11:17 2003
Picon

Re: Can I use whole flash chip in one mtd partition for jffs2 filesystem?

Yes.

Your output looks good, congratulations!

Jörn

--

-- 
Anything that can go wrong, will.
-- Finagle's Law

To unsubscribe from this list: send the line "unsubscribe jffs-dev" in
the body of a message to majordomo <at> axis.com

Lokesh Kumar | 20 Aug 16:31 2003

jffs2 flash life calculation


We are developing a product, which requires a flash life span of 10 years.
Specs are - 350KB data / 10 minutes, at any given time 2MB data in flash,
NOR flash which has 100K erase cycles/ block, with 128k sector size. This is
Intel startaflash, with JFFS2 over MTD.

How do I calculate the required flash size, taking wear-leveling of JFFS2
into account. Where can I find this information?

Lokesh Kumar
Wireless Matrix USA Inc.
703-262-4032

To unsubscribe from this list: send the line "unsubscribe jffs-dev" in
the body of a message to majordomo <at> axis.com

Charles Manning | 21 Aug 03:18 2003
Picon

Re: jffs2 flash life calculation

On Thursday 21 August 2003 02:31, Lokesh Kumar wrote:
> We are developing a product, which requires a flash life span of 10 years.
> Specs are - 350KB data / 10 minutes, at any given time 2MB data in flash,
> NOR flash which has 100K erase cycles/ block, with 128k sector size. This
> is Intel startaflash, with JFFS2 over MTD.
>
> How do I calculate the required flash size, taking wear-leveling of JFFS2
> into account. Where can I find this information?
>

Answering questions like this is not easy. The impact will depend on various 
factors such as what your writes do and how they impact the generation of 
discarded data and thus garbage collection etc.

For example, if you write 350kB as a single write() call you will generate 
far less garbage than is you do 10240 writes of 35 bytes, or worse still 
overlapping writes.

-- Charles

To unsubscribe from this list: send the line "unsubscribe jffs-dev" in
the body of a message to majordomo <at> axis.com

Lokesh Kumar | 21 Aug 14:51 2003

RE: jffs2 flash life calculation

Lets say I do rougly 1K writes, 35 times in 10 minutes.

Lokesh

-----Original Message-----
From: Charles Manning [mailto:manningc2 <at> actrix.gen.nz]
Sent: Wednesday, August 20, 2003 9:19 PM
To: Lokesh Kumar; 'jffs-dev <at> axis.com'
Subject: Re: jffs2 flash life calculation

On Thursday 21 August 2003 02:31, Lokesh Kumar wrote:
> We are developing a product, which requires a flash life span of 10 years.
> Specs are - 350KB data / 10 minutes, at any given time 2MB data in flash,
> NOR flash which has 100K erase cycles/ block, with 128k sector size. This
> is Intel startaflash, with JFFS2 over MTD.
>
> How do I calculate the required flash size, taking wear-leveling of JFFS2
> into account. Where can I find this information?
>

Answering questions like this is not easy. The impact will depend on various

factors such as what your writes do and how they impact the generation of 
discarded data and thus garbage collection etc.

For example, if you write 350kB as a single write() call you will generate 
far less garbage than is you do 10240 writes of 35 bytes, or worse still 
overlapping writes.

-- Charles
(Continue reading)

Lokesh Kumar | 21 Aug 16:37 2003

RE: jffs2 flash life calculation

Roughly 350K of data will get processed every 10 minutes or so, and get
removed from the flash. So, there will be multiple files, each of 350K, may
be 6 of them in the file system at any given time, and then the first 350K
file will be processed and removed.

Thanks for your help

Lokesh

-----Original Message-----
From: David Woodhouse [mailto:dwmw2 <at> infradead.org]
Sent: Thursday, August 21, 2003 10:33 AM
To: Lokesh Kumar
Cc: 'manningc2 <at> actrix.gen.nz'; 'jffs-dev <at> axis.com'
Subject: RE: jffs2 flash life calculation

On Thu, 2003-08-21 at 08:51 -0400, Lokesh Kumar wrote:
> Lets say I do rougly 1K writes, 35 times in 10 minutes.

And how well does it compress? How are you 'expiring' old data?
Truncating the file every time it gets too big, or something else?

How big is your flash?

--

-- 
dwmw2

To unsubscribe from this list: send the line "unsubscribe jffs-dev" in
the body of a message to majordomo <at> axis.com

(Continue reading)

David Woodhouse | 21 Aug 16:32 2003

RE: jffs2 flash life calculation

On Thu, 2003-08-21 at 08:51 -0400, Lokesh Kumar wrote:
> Lets say I do rougly 1K writes, 35 times in 10 minutes.

And how well does it compress? How are you 'expiring' old data?
Truncating the file every time it gets too big, or something else?

How big is your flash?

--

-- 
dwmw2

To unsubscribe from this list: send the line "unsubscribe jffs-dev" in
the body of a message to majordomo <at> axis.com


Gmane