On Mon, Jan 15, 2018 at 12:43:44PM -0500, Mike Snitzer wrote:
> On Mon, Jan 15 2018 at 11:58am -0500,
> Ming Lei <ming.lei@xxxxxxxxxx> wrote:
>
> > Hi Guys,
> >
> > The 3 paches changes the blk-mq part of blk_insert_cloned_request(),
> > in which we switch to blk_mq_try_issue_directly(), so that both dm-rq
> > and blk-mq can get the dispatch result of underlying queue, and with
> > this information, blk-mq can handle IO merge much better, then
> > sequential I/O performance is improved much.
> >
> > In my dm-mpath over virtio-scsi test, this whole patchset improves
> > sequential IO by 3X ~ 5X.
> >
> > V4:
> > - remove dm patches which are in DM tree already
> > - cleanup __blk_mq_issue_req as suggested by Jens
> >
>
> Ming,
>
> You dropped the header cleanups that I did in v3 ("blk-mq: issue request
> directly for blk_insert_cloned_request") being the one header I care
> about being updated).
>
> I also worked in parallel on my own v4 to address Jens' dislike for v3's
> 3 returns. But I skinned the cat a different way, by dropping your
> first patch that introduces the __blk_mq_issue_req helper, please see
> these 2 commits:
>
> https://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm.git/commit/?h=block-4.16&id=40f8947784128bb83dc5f7a6aed7ed230222f675
> https://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm.git/commit/?h=block-4.16&id=f641015f42f41df75220313ee62e8241f2feeeed
>
> I think it makes the changes more obvious (not spread across 2 methods
> and doesn't require use of BLK_STS_AGAIN).
>
> Happy to yield to Jens to decide which he prefers.
>
> Jens, if you'd like to pick my variant of v4 up they are here, thanks!
Hi Mike,
Looks we were working on it at the same time, I am fine with your V4.
Jens, please let us know if Mike's V4 is OK, if not, we can make V5/.., :-)
Thanks,
Ming
