Re: [PATCH v5 0/2] ext4: Improve parallel I/O performance on NVDIMM

From: Christoph Hellwig
Date: Thu May 05 2016 - 10:19:39 EST

On Thu, May 05, 2016 at 11:57:30AM +1000, Dave Chinner wrote:
> We did this intentionally. DAX IO needs to have the same parallel
> write semantics of direct IO, because otherwise a single writer
> prevents any IO concurrency and that's a bigger problem for DAX that
> traditional storage due to the access speed and bandwidth available.

But this is secondard - even buffered I/O would benefit from not
having the global lock. The difference is that Posix requires exclusive
writers, and we can't just break those semantics because we're sitting
on whizz bang fancy backend storage.