Re: [PATCH 00/26] generic btree implementation, version 3

[Christoph Hellwig <hch@xxxxxx>]

On Mon, Aug 04, 2008 at 11:54:25AM +1000, Dave Chinner wrote:
> > The alloc and inobt trees were in fact using the almost exact sequence
> > built from ->set_root and ->free_block but the odd way to pass the
> > block to be freed made this non-obvious.  By changing the calling
> > convention to the normal one this one could be unified.  I've left
> > in some nasty asserts to ensure the assumptions about these blocks
> > wasn't wrong.
> 
> Ok. I'll comment on it when the patches come through....

The killroot unification turned out to be buggy, I'll fix it in the
next round of patches.