Re: If not readdir() then what?

["Bob Copeland" <me@bobcopeland.com>]

On 4/10/07, Neil Brown <neilb@suse.de> wrote:
> 2/ Some data structure using an ordered search key that is based on
>   the filename (e.g. a B-tree with a search key that is a hash of the
>   filename).
>
> In the first case, you just use a fixed opaque cookie for location in
> a directory.
> In the second you use the filename.  If the file has been deleted,
> that shouldn't stop you finding the place where it would have been in
> the overall sort order.

I can think of one (admittedly insane) FS that is between those two:

3/ an unsorted hash table, where each directory entry has an indirect
pointer to its neighbor in case of hash collisions.

   a
   b -> d -> c -> e
   g
   f -> x

Given 'c' as the "last" thing returned, you can hash c to find out
that you are in the bucket with 'b', but if 'c' was deleted, the best
you can do is return b twice or skip the chain entirely.  I maintain
an out-of-tree driver for such an fs (I promise I did not invent it);
the best I could come up with is to encode the hash chain index in the
top byte of f_pos.   Needless to say, readdir is very not performant
with all the seeking this hash scheme entails.

-Bob