Hi,
On Tue, Apr 18, 2000 at 09:12:47AM -0400, Alexander Viro wrote:
>
> Ouch. b-tree => potentially unbound amount of blocks to be written
> during the operation.
Yes, but if you are talking about storing extents rather than individual
block mappings, then the amount of data in the tree will be a _tiny_
fraction of the amount needed in the current scheme.
The only need for btrees is if you want to deal with holes cleanly.
If you are willing to lose that, then storing extent maps in a static
tree is even easier, but you probably have to be willing to convert
the mapping into a straight indirection array if you ever write into
a hole if you don't have a btree structure.
The goal is the minimisation of the mapping datastructures. btree
just makes it easier to achieve with holes: the _important_ part is
not the btree part, but the use of extent mappings rather than
discrete block mappings.
--Stephen
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Sun Apr 23 2000 - 21:00:13 EST