Re: [rfc] Near-constant time directory index for Ext2

[Linus Torvalds]

In article <[e-mail]>,
Daniel Phillips  <[e-mail]> wrote:
>
>I mentioned this earlier but it's worth repeating: the desire to use a
>small block size is purely an artifact of the fact that ext2 has no
>handling for tail block fragmentation.  That's a temporary situation -
>once we've dealt with it your 2,000,000 file directory will be happier
>with 4K filesystem blocks.

I'd rather see a whole new filesystem than have ext2 do tail-block
fragmentation. 

Once you do tail fragments, you might as well do the whole filesystem
over and have it do fancier stuff than just handling sub-blocking. 

Another way of saying this: if you go to the complexity of no longer
being a purely block-based filesystem, please go the whole way. Make the
thing be extent-based, and get away from the notion that you have to
allocate blocks one at a time. Make the blocksize something nice and
big, not just 4kB or 8kB or something.

And don't call it ext2. 

		Linus