I'm pretty familiar with this code since I hacked on it pretty extensively
last month and it does what it tries to do pretty well. However it relies
heavily on the assumption that directory blocks can be grouped into page-sized
units. This assumption doesn't hold in my directory indexing code, and I
don't see any clean way to extend the approach you use in this patch to my
index design.
On the other hand, there is an alternative approach, suggested to me by
Stephen Tweedie, that will run just as fast as this code and have a lot less
cruft in it, namely - perform buffer operations on the underlying buffers in
the page cache. Now perhaps we should discuss that idea and see if it goes
anywhere.
Aside from the part that's tied to the page cache, your patch is generally a
whole lot nicer to read that the original, and I have already incorporated
parts of it in my directory index patch.
-- Daniel - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/