On Tue, Nov 11, 2014 at 11:02:00AM -0500, Milosz Tanski wrote:

On Tue, Nov 11, 2014 at 1:44 AM, Dave Chinner [off-list ref] wrote:

quoted

On Mon, Nov 10, 2014 at 11:40:23AM -0500, Milosz Tanski wrote:

quoted

This patcheset introduces an ability to perform a non-blocking read from
regular files in buffered IO mode. This works by only for those filesystems
that have data in the page cache.

It does this by introducing new syscalls new syscalls preadv2/pwritev2. These
new syscalls behave like the network sendmsg, recvmsg syscalls that accept an
extra flag argument (RWF_NONBLOCK).

It's a very common patern today (samba, libuv, etc..) use a large threadpool to
perform buffered IO operations. They submit the work form another thread
that performs network IO and epoll or other threads that perform CPU work. This
leads to increased latency for processing, esp. in the case of data that's
already cached in the page cache.

With the new interface the applications will now be able to fetch the data in
their network / cpu bound thread(s) and only defer to a threadpool if it's not
there. In our own application (VLDB) we've observed a decrease in latency for
"fast" request by avoiding unnecessary queuing and having to swap out current
tasks in IO bound work threads.

Can you write a test (or set of) for fstests that exercises this new
functionality? I'm not worried about performance, just
correctness....

Sure thing. Can you point me at the fstests repo? A quick google
search reveals lots of projects named fstests, most of them abandoned.

git://git.kernel.org/pub/scm/fs/xfs/xfstests-dev.git

Cheers,

Dave.
-- 
Dave Chinner
david@fromorbit.com

--
To unsubscribe, send a message with 'unsubscribe linux-aio' in
the body to majordomo@kvack.org.  For more info on Linux AIO,
see: http://www.kvack.org/aio/
Don't email: <a href=mailto:"aart@kvack.org">aart@kvack.org</a>