[GSOC] HAMMER2 compression feature week9 report

Daniel Flores daniel5555 at gmail.com
Sun Aug 18 10:47:09 PDT 2013

Hello everyone,
This week I was working very closely with my mentor on implementing write
path using threads. The write path is, essentially, the same, but now it is
inside a supporting thread. This change was quite challenging and we faced
a lot of bugs, mostly related to racing conditions. But, right now, most of
those bugs are fixed and the whole feature seems to be pretty stable. I
can’t say it’s ready, because there are still things to improve and I still
have to do real stress tests, but it’s closer now to a really stable state.

Aside from this, I don’t really have much to say for now. There were some
minor changes to hammer2 utility code and I also started to make some
changes to LZ4 compression files, because they give us many warnings during
compilation and we don’t need many of the features they offer. I’ll
continue to do this during next days and then I’ll start to exhaustively
test the file system. If everything will be fine, I’m also planning to add
an alternative compression algorithm. This will be the next goal for my
project, if there won’t be any important bugs that will be very
time-consuming to fix.

As usually, you can check my code from my repository [1]. I’ll be grateful
for any advices, feedback, and criticism.


[1] git://leaf.dragonflybsd.org/~iostream/dragonfly.git<http://leaf.dragonflybsd.org/%7Eiostream/dragonfly.git>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.dragonflybsd.org/pipermail/kernel/attachments/20130818/d71fcbfd/attachment.html>

More information about the Kernel mailing list