Discussion:
creating jffs2 filesystem on ext3
Tomasz Chmielewski
2005-04-05 13:08:14 UTC
Permalink
Hi,

As jffs2 is the only stable filesystem for Linux that is capable of r/w
compression, I wanted to test it a bit.

I have a 23 GB ext3 /mnt/test partition on which I would like to create
a ~23 GB jffs2 image.

My / partiton is about 9 GB.

How can I create that big jffs2 image?

man doesn't say much (at least for me); it seems that it's only possible
to create images from an existing directory.

Does it mean I won't be able to create an empty 23 GB jffs2 image?

Or perhaps I didn't see some option?


Tomek


To unsubscribe from this list: send the line "unsubscribe jffs-dev" in
the body of a message to ***@axis.com
David Woodhouse
2005-04-05 13:45:16 UTC
Permalink
Post by Tomasz Chmielewski
Does it mean I won't be able to create an empty 23 GB jffs2 image?
Or perhaps I didn't see some option?
JFFS2 won't handle images larger than 4GiB, and as a fully log-
structured file system its RAM usage will probably become prohibitive
before you get that big. We're working on tricks to make it scale
better, but until that's done there's little point in dealing with the
4GiB limit. That one isn't a JFFS2 problem per se; it's just that the
MTD core doesn't handle devices larger than that.
--
dwmw2


To unsubscribe from this list: send the line "unsubscribe jffs-dev" in
the body of a message to ***@axis.com
Jörn Engel
2005-04-05 14:34:01 UTC
Permalink
Post by Tomasz Chmielewski
As jffs2 is the only stable filesystem for Linux that is capable of r/w
compression, I wanted to test it a bit.
I have a 23 GB ext3 /mnt/test partition on which I would like to create
a ~23 GB jffs2 image.
My / partiton is about 9 GB.
How can I create that big jffs2 image?
You cannot. The mtd interface is limited to 4GiB. Not sure if anyone
has tested a 4GiB filesystem yet, so the effective limit might be
lower, due to bugs. 2GiB minus a bit should be safe.

Easiest would be to fill the device with 0xff, use an mtd translation
and mount it as jffs2.

Also note that block2mtd, as of 2.6.11[.*] is broken, so either update
to mtd cvs or use blkmtd as translation.

Jörn
--
When in doubt, punt. When somebody actually complains, go back and fix it...
The 90% solution is a good thing.
-- Rob Landley

To unsubscribe from this list: send the line "unsubscribe jffs-dev" in
the body of a message to ***@axis.com
David Woodhouse
2005-04-05 15:07:06 UTC
Permalink
Post by Jörn Engel
Also note that block2mtd, as of 2.6.11[.*] is broken, so either update
to mtd cvs or use blkmtd as translation.
Is the fix not already in the BK tree and hence in Andrew's 2.6.x-mm
tree.
--
dwmw2


To unsubscribe from this list: send the line "unsubscribe jffs-dev" in
the body of a message to ***@axis.com
Jörn Engel
2005-04-05 15:17:45 UTC
Permalink
Post by David Woodhouse
Post by Jörn Engel
Also note that block2mtd, as of 2.6.11[.*] is broken, so either update
to mtd cvs or use blkmtd as translation.
Is the fix not already in the BK tree and hence in Andrew's 2.6.x-mm
tree.
Didn't see it yet, checking...
Kernel.org is slow, it may take a byte (just to ridicule Artem ;)

Jörn
--
tglx1 thinks that joern should get a (TM) for "Thinking Is Hard"
-- Thomas Gleixner

To unsubscribe from this list: send the line "unsubscribe jffs-dev" in
the body of a message to ***@axis.com
Jörn Engel
2005-04-05 16:00:39 UTC
Permalink
Post by David Woodhouse
Post by Jörn Engel
Also note that block2mtd, as of 2.6.11[.*] is broken, so either update
to mtd cvs or use blkmtd as translation.
Is the fix not already in the BK tree and hence in Andrew's 2.6.x-mm
tree.
2.6.12-rc2-mm1 has the fix.

Jörn
--
Optimizations always bust things, because all optimizations are, in
the long haul, a form of cheating, and cheaters eventually get caught.
-- Larry Wall

To unsubscribe from this list: send the line "unsubscribe jffs-dev" in
the body of a message to ***@axis.com
Loading...