Re: Expected getdents behaviour

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

On Thu, 2005-09-15 at 15:07 +0100, Anton Altaparmakov wrote:
> On Thu, 2005-09-15 at 09:57 -0400, Akshat Aranya wrote:
> > I noticed that bonnie++, in its directory tests, does the following on
> > a large directory
> > 
> > open(dir);
> > while (getdents() != 0)
> > {
> >    unlink all the returned entries from getdents
> > }
> > close(dir);
> > 
> > My question is whether the filesystem's readdir is expected to
> > consider the offset value in the second readdir to still be valid,
> > given that entries from the directory were deleted after the first
> > readdir.
> 
> This would fail horribly on ntfs for example (once it supports file
> deletion anyway) since the offset value of all entries changes when the
> directory is modified.

Oops.  I forgot to answer your question.  Yes, the filesystem needs to
consider the offset value in the second readdir to still be valid.  You
cannot keep rewinding back to zero every time you make a modification or
you would keep returning entries you have already returned and never
make any progress if e.g. some user does this in a loop at the same
time:

while 1; do
	touch blah
	rm blah
done

It would be a very effective DOS that any user can trigger.

Bonnie++'s code is just complete crap...  It is the author's fault that
it will not work on filesystems where the directory entries are not in
fixed locations...

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/