No space left on almost empty JFFS2 filesystem

[Thomas Gleixner]

On Wednesday 25 June 2003 20:24, Paul van Gool wrote:
> Creating am image with '-e 0x10000' and running with it, results in:
>
> # mount -t jffs2 /dev/mtdblock4 /mnt/jffs
> mtdblock_open
> ok
> jffs2_scan_empty(): Empty block at 0x000017ac ends at 0x00010000 (with
> 0x200319y CLEANMARKER node found at 0x00010000, not first node in block
> (0x00000000) jffs2_scan_empty(): Empty block at 0x0002000c ends at
> 0x00030000 (with 0x200319y CLEANMARKER node found at 0x00030000, not first
> node in block (0x00020000) jffs2_scan_empty(): Empty block at 0x0004000c
> ends at 0x00050000 (with 0x200319y CLEANMARKER node found at 0x00050000,
> not first node in block (0x00040000) jffs2_scan_empty(): Empty block at
> 0x0006000c ends at 0x00070000 (with 0x200319y CLEANMARKER node found at
> 0x00070000, not first node in block (0x00060000) jffs2: No clean _or_ dirty
> blocks to GC from! Where are they all?
> jffs2: Couldn't find erase block to garbage collect!
> jffs2: No clean _or_ dirty blocks to GC from! Where are they all?
> jffs2: Couldn't find erase block to garbage collect!
> jffs2: No clean _or_ dirty blocks to GC from! Where are they all?
> jffs2: Couldn't find erase block to garbage collect!
Don't try to betray the pyhsiccal facts. Even if you tell mkfs.jffs2 that your 
eraseblock size is < 128K,  the physical eraseblock size will remain at 128K, 
as this is given by the chip. 

You can only erase eraseblocks and not a quarter of them, as you can't buy a 
1/4 of a beerbottle. You can fill the bottle with 1/4 of the maximum content, 
but the bottle still remains the same size. :)

-- 
Thomas
________________________________________________________________________
linutronix - competence in embedded & realtime linux
http://www.linutronix.de
mail: tglx at linutronix.de