Btrfs: avoid allocation clusters that are too spread out
authorChris Mason <chris.mason@oracle.com>
Tue, 9 Jun 2009 22:35:15 +0000 (18:35 -0400)
committerChris Mason <chris.mason@oracle.com>
Wed, 10 Jun 2009 15:29:51 +0000 (11:29 -0400)
In SSD mode for data, and all the time for metadata the allocator
will try to find a cluster of nearby blocks for allocations.  This
commit adds extra checks to make sure that each free block in the
cluster is close to the last one.

Signed-off-by: Chris Mason <chris.mason@oracle.com>

No differences found