Hello All
I have just pushed the first cut of block summary code to
http://yaffs.net/gitweb?p=yaffs2/.git;a=shortlog;h=refs/heads/block-summary
The block summary feature writes all the tags in a block to the last chunk (or
in some cases chunks) of a block.
Instead of reading all the tags during a scan, the summary is read as a single
read and we have all the tags for the block. This can significantly reduce
the number of reads during a scan and should make scanning a lot faster. In
tests I've been doing this reduces the number of reads by a factor of 40 or
so.
If anyone would like to experiment and provide feedback I would appreciate
that.
-- CHarles