FW: "du" returns wrong values on yaffs

Stephan Linke Stephan.Linke at epygi.de
Fri Aug 29 08:50:08 EDT 2003


Hi,

in MTD list there are more filesystem guys so I copy this one from YAFFS list...

Stephan

> -----Original Message-----
> From: Stephan Linke [mailto:Stephan.Linke at epygi.de]
> Sent: Freitag, 29. August 2003 14:48
> To: manningc2 at actrix.gen.nz; Yaffs List
> Subject: RE: "du" returns wrong values on yaffs
> 
> 
> Hi,
> 
> Now there's something strange:
> The "du" I use (busybox) still returnes strange values. The reason is that it assumes a block size of 512 bytes and 
> doesn't care for the file block size returned by lstat. "du" simply divides by 2 to get the number of 1k blocks (already 
> a wrong calculation I guess).
> 
> Filesystems like cramfs and jffs show a similiar stupid behavior. They sai that their block size is PAGE_CACHE_SIZE or 
> PAGE_SIZE (4k in our case) but use a hard coded 512 to calculate the block count.
> It looks to me like there's something wrong in most of the filesystem inodes...
> 
> Thanx, Stephan
> 
> 
> > -----Original Message-----
> > From: Charles Manning [mailto:manningc2 at actrix.gen.nz]
> > Sent: Mittwoch, 20. August 2003 06:03
> > To: Stephan Linke; Yaffs List
> > Subject: Re: "du" returns wrong values on yaffs
> > 
> > 
> > On Tuesday 19 August 2003 21:39, Stephan Linke wrote:
> > > Hi,
> > > we are using "du" to detect the amount of space occupied by some of our
> > > files to keep us from running out of space. Unfortunately YAFFS returns a
> > > block count of zero.
> > >
> > Stepahn I found and ??corrected?? a problem, though I don't know it is 100% 
> > yet. THis is in cvs. 
> > 
> > At best du only provides an estimate and does not account for various other 
> > overheads (eg. bad blocks, reserved blocks). statfs() , as used by 'df', 
> > gives a better indicator of free space.
> > 
> > Stephan: please try the new yaffs_fs and see if this helps. 
> > 
> > -- Charles
> > 



More information about the linux-mtd mailing list