[Drbd-dev] [PATCH 00/13] convert block layer to bioset_init()/mempool_init()

[Mike Snitzer]

On Mon, May 21 2018 at 10:52am -0400,
Jens Axboe <axboe at kernel.dk> wrote:

> On 5/21/18 8:47 AM, Mike Snitzer wrote:
> > On Mon, May 21 2018 at 10:36am -0400,
> > Jens Axboe <axboe at kernel.dk> wrote:
> > 
> >> On 5/21/18 8:31 AM, Mike Snitzer wrote:
> >>> On Mon, May 21 2018 at 10:19am -0400,
> >>> Jens Axboe <axboe at kernel.dk> wrote:
> >>>
> >>>> On 5/21/18 8:03 AM, Mike Snitzer wrote:
> >>>>> On Sun, May 20 2018 at  6:25pm -0400,
> >>>>> Kent Overstreet <kent.overstreet at gmail.com> wrote:
> >>>>>
> >>>>>> Jens - this series does the rest of the conversions that Christoph wanted, and
> >>>>>> drops bioset_create().
> >>>>>>
> >>>>>> Only lightly tested, but the changes are pretty mechanical. Based on your
> >>>>>> for-next tree.
> >>>>>
> >>>>> By switching 'mempool_t *' to 'mempool_t' and 'bio_set *' to 'bio_set'
> >>>>> you've altered the alignment of members in data structures.  So I'll
> >>>>> need to audit all the data structures you've modified in DM.
> >>>>>
> >>>>> Could we get the backstory on _why_ you're making this change?
> >>>>> Would go a long way to helping me appreciate why this is a good use of
> >>>>> anyone's time.
> >>>>
> >>>> Yeah, it's in the first series, it gets rid of a pointer indirection.
> >>>
> >>> "Allows mempools to be embedded in other structs, getting rid of a
> >>> pointer indirection from allocation fastpaths."
> >>>
> >>> So this is about using contiguous memory or avoiding partial allocation
> >>> failure?  Or both?
> >>>
> >>> Or more to it?  Just trying to fully appreciate the theory behind the
> >>> perceived associated benefit.
> >>
> >> It's about avoiding a pointer indirection. Instead of having to
> >> follow a pointer to get to that struct, it's simple offset math off
> >> your main structure.
> >>
> >>> I do think the increased risk of these embedded bio_set and mempool_t
> >>> themselves crossing cachelines, or struct members that follow them doing
> >>> so, really detracts from these types of changes.
> >>
> >> Definitely something to look out for, though most of them should be
> >> per-dev structures and not in-flight structures. That makes it a bit
> >> less sensitive. But can't hurt to audit the layouts and adjust if
> >> necessary. This is why it's posted for review :-)
> > 
> > This isn't something that is easily caught upfront.  Yes we can all be
> > busy little beavers with pahole to audit alignment.  But chances are
> > most people won't do it.
> > 
> > Reality is there is potential for a regression due to false sharing to
> > creep in if a hot struct member suddenly starts straddling a cacheline.
> > That type of NUMA performance killer is pretty insidious and somewhat
> > tedious to hunt down even when looking for it with specialized tools:
> > https://joemario.github.io/blog/2016/09/01/c2c-blog/
> 
> IMHO you're making a big deal out of something that should not be.

I raised an issue that had seemingly not been considered at all.  Not
making a big deal.  Raising it for others' benefit.

> If the dm bits are that sensitive and cache line honed to perfection
> already due to previous regressions in that area, then it might
> not be a bad idea to have some compile checks for false cacheline
> sharing between sensitive members, or spilling of a sub-struct
> into multiple cachelines.
> 
> It's not like this was pushed behind your back. It's posted for
> review. It's quite possible the net change is a win for dm. Let's
> focus on getting it reviewed, rather than pontificate on what
> could potentially go all wrong with this.

Why are you making this personal?  Or purely about DM?  I'm merely
pointing out this change isn't something that can be given a quick
blanket "looks good".

Mike