How to handle NFS patches for fscache I/O partial rewrite?

[David Howells <dhowells@redhat.com>]

Hi Trond, Anna,

As you may have seen, I've managed to cut down the fscache overhaul to add a
glue layer to handle reads and switch to using async DIO through kiocbs rather
than using sync write and async reads with page wait list snooping.  All the
rest of the API is unchanged (for the moment).

The glue layer (netfs helper library) handles the new VM readahead stuff and
reading transparent huge pages on behalf of the netfs, plus read shaping,
splitting and retrying.

Dave Wysochanski created some NFS patches for it, which I posted as part of my
branch.  I'd like to get them into linux-next with an eye to having them
pulled by Linus in the next merge window, along with the rest of my branch.

Should I keep them in my branch and thence into linux-next, or do they need to
go through one of your branches?

Thanks,
David