Re: Expected getdents behaviour

[Anton Altaparmakov <aia21@cam.ac.uk>]

On Thu, 15 Sep 2005, Jan Blunck wrote:
> On Thu, Sep 15, Anton Altaparmakov wrote:
> > That is completely untrue.  It is up to the user.  The file position in
> > a directory is completely up to the caller and in particular when a
> > directory is modified the file position in that directory is not changed
> > by the filesystem.  Doing this would in fact be impossible since the
> > creation/deletion in the kernel has no access to all the "struct file"s
> > with which a directory has been opened.
> 
> No, thats not true. The file position in a regular (flat) file is completely
> up to the user. "Thanks" to the short-comings of POSIX on not defining the
> functions on directories, the file position in a directory is completely up to
> the file system. Due to that fact, modern file systems don't need to provide
> a "flat" view (in terms of sequential access to a directory file) on

We are talking past each other and actually agreeing with each other.  Our 
definitions of "up to the user" and "up to the fs" are at opposites.

> them. Although, the file system must guarantee, that readdir's must not fail
> because of changes directory contents. AFAIR ext2 is achieving this through
> not shrinking directories while the filesystem is online.

ntfs uses B trees on disk for directories.  The tree must be balanced at 
all times and each insert/delete will be a balancing tree operation.  
This can completely change the tree shape and hence the on disk data.

Best regards,

	Anton
-- 
Anton Altaparmakov <aia21 at cam.ac.uk> (replace at with @)
Unix Support, Computing Service, University of Cambridge, CB2 3QH, UK
Linux NTFS maintainer / IRC: #ntfs on irc.freenode.net
WWW: http://linux-ntfs.sf.net/ & http://www-stu.christs.cam.ac.uk/~aia21/