From: Dave Chinner <david@fromorbit.com>
To: Nikolay Borisov <nborisov@suse.com>
Cc: linux-xfs@vger.kernel.org, linux-mm@kvack.org,
linux-fsdevel@vger.kernel.org
Subject: Re: [PATCH 04/24] shrinker: defer work only to kswapd
Date: Mon, 5 Aug 2019 07:37:32 +1000 [thread overview]
Message-ID: <20190804213732.GU7777@dread.disaster.area> (raw)
In-Reply-To: <625f5e1e-b362-7a76-be01-7f1057646588@suse.com>
On Sun, Aug 04, 2019 at 07:48:01PM +0300, Nikolay Borisov wrote:
>
>
> On 1.08.19 г. 5:17 ч., Dave Chinner wrote:
> > From: Dave Chinner <dchinner@redhat.com>
> >
> > Right now deferred work is picked up by whatever GFP_KERNEL context
> > reclaimer that wins the race to empty the node's deferred work
> > counter. However, if there are lots of direct reclaimers, that
> > work might be continually picked up by contexts taht can't do any
> > work and so the opportunities to do the work are missed by contexts
> > that could do them.
> >
> > A further problem with the current code is that the deferred work
> > can be picked up by a random direct reclaimer, resulting in that
> > specific process having to do all the deferred reclaim work and
> > hence can take extremely long latencies if the reclaim work blocks
> > regularly. This is not good for direct reclaim fairness or for
> > minimising long tail latency events.
> >
> > To avoid these problems, simply limit deferred work to kswapd
> > contexts. We know kswapd is a context that can always do reclaim
> > work, and hence deferring work to kswapd allows the deferred work to
> > be done in the background and not adversely affect any specific
> > process context doing direct reclaim.
> >
> > The advantage of this is that amount of work to be done in direct
> > reclaim is now bound and predictable - it is entirely based on
> > the cache's freeable objects and the reclaim priority. hence all
> > direct reclaimers running at the same time should be doing
> > relatively equal amounts of work, thereby reducing the incidence of
> > long tail latencies due to uneven reclaim workloads.
> >
> > Signed-off-by: Dave Chinner <dchinner@redhat.com>
> > ---
> > mm/vmscan.c | 93 ++++++++++++++++++++++++++++-------------------------
> > 1 file changed, 50 insertions(+), 43 deletions(-)
> >
> > diff --git a/mm/vmscan.c b/mm/vmscan.c
> > index b7472953b0e6..c583b4efb9bf 100644
> > --- a/mm/vmscan.c
> > +++ b/mm/vmscan.c
> > @@ -500,15 +500,15 @@ static unsigned long do_shrink_slab(struct shrink_control *shrinkctl,
> > struct shrinker *shrinker, int priority)
> > {
> > unsigned long freed = 0;
> > - long total_scan;
> > int64_t freeable_objects = 0;
> > int64_t scan_count;
> > - long nr;
> > + int64_t scanned_objects = 0;
> > + int64_t next_deferred = 0;
> > + int64_t deferred_count = 0;
> > long new_nr;
> > int nid = shrinkctl->nid;
> > long batch_size = shrinker->batch ? shrinker->batch
> > : SHRINK_BATCH;
> > - long scanned = 0, next_deferred;
> >
> > if (!(shrinker->flags & SHRINKER_NUMA_AWARE))
> > nid = 0;
> > @@ -519,47 +519,53 @@ static unsigned long do_shrink_slab(struct shrink_control *shrinkctl,
> > return scan_count;
> >
> > /*
> > - * copy the current shrinker scan count into a local variable
> > - * and zero it so that other concurrent shrinker invocations
> > - * don't also do this scanning work.
> > + * If kswapd, we take all the deferred work and do it here. We don't let
> > + * direct reclaim do this, because then it means some poor sod is going
> > + * to have to do somebody else's GFP_NOFS reclaim, and it hides the real
> > + * amount of reclaim work from concurrent kswapd operations. Hence we do
> > + * the work in the wrong place, at the wrong time, and it's largely
> > + * unpredictable.
> > + *
> > + * By doing the deferred work only in kswapd, we can schedule the work
> > + * according the the reclaim priority - low priority reclaim will do
> > + * less deferred work, hence we'll do more of the deferred work the more
> > + * desperate we become for free memory. This avoids the need for needing
> > + * to specifically avoid deferred work windup as low amount os memory
> > + * pressure won't excessive trim caches anymore.
> > */
> > - nr = atomic_long_xchg(&shrinker->nr_deferred[nid], 0);
> > + if (current_is_kswapd()) {
> > + int64_t deferred_scan;
> >
> > - total_scan = nr + scan_count;
> > - if (total_scan < 0) {
> > - pr_err("shrink_slab: %pS negative objects to delete nr=%ld\n",
> > - shrinker->scan_objects, total_scan);
> > - total_scan = scan_count;
> > - next_deferred = nr;
> > - } else
> > - next_deferred = total_scan;
> > + deferred_count = atomic64_xchg(&shrinker->nr_deferred[nid], 0);
> >
> > - /*
> > - * We need to avoid excessive windup on filesystem shrinkers
> > - * due to large numbers of GFP_NOFS allocations causing the
> > - * shrinkers to return -1 all the time. This results in a large
> > - * nr being built up so when a shrink that can do some work
> > - * comes along it empties the entire cache due to nr >>>
> > - * freeable. This is bad for sustaining a working set in
> > - * memory.
> > - *
> > - * Hence only allow the shrinker to scan the entire cache when
> > - * a large delta change is calculated directly.
> > - */
> > - if (scan_count < freeable_objects / 4)
> > - total_scan = min_t(long, total_scan, freeable_objects / 2);
> > + /* we want to scan 5-10% of the deferred work here at minimum */
> > + deferred_scan = deferred_count;
> > + if (priority)
> > + do_div(deferred_scan, priority);
> > + scan_count += deferred_scan;
> > +
> > + /*
> > + * If there is more deferred work than the number of freeable
> > + * items in the cache, limit the amount of work we will carry
> > + * over to the next kswapd run on this cache. This prevents
> > + * deferred work windup.
> > + */
> > + if (deferred_count > freeable_objects * 2)
> > + deferred_count = freeable_objects * 2;
>
> nit : deferred_count = min(deferred_count, freeable_objects * 2).
*nod*
> How can we have more deferred objects than are currently on the LRU?
deferred work is aggregated. Put enough direct reclaimers in action
in GFP_NOFS context (e.g. fsmark create workload) and it will wind
up the deferred count much faster than kswapd can drain it.
> Aren't deferred objects always some part of freeable objects.
For a single scan, yes. In aggregate, no.
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
next prev parent reply other threads:[~2019-08-04 21:38 UTC|newest]
Thread overview: 87+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-08-01 2:17 [RFC] [PATCH 00/24] mm, xfs: non-blocking inode reclaim Dave Chinner
2019-08-01 2:17 ` [PATCH 01/24] mm: directed shrinker work deferral Dave Chinner
2019-08-02 15:27 ` Brian Foster
2019-08-04 1:49 ` Dave Chinner
2019-08-05 17:42 ` Brian Foster
2019-08-05 23:43 ` Dave Chinner
2019-08-06 12:27 ` Brian Foster
2019-08-06 22:22 ` Dave Chinner
2019-08-07 11:13 ` Brian Foster
2019-08-01 2:17 ` [PATCH 02/24] shrinkers: use will_defer for GFP_NOFS sensitive shrinkers Dave Chinner
2019-08-02 15:27 ` Brian Foster
2019-08-04 1:50 ` Dave Chinner
2019-08-01 2:17 ` [PATCH 03/24] mm: factor shrinker work calculations Dave Chinner
2019-08-02 15:08 ` Nikolay Borisov
2019-08-04 2:05 ` Dave Chinner
2019-08-02 15:31 ` Brian Foster
2019-08-01 2:17 ` [PATCH 04/24] shrinker: defer work only to kswapd Dave Chinner
2019-08-02 15:34 ` Brian Foster
2019-08-04 16:48 ` Nikolay Borisov
2019-08-04 21:37 ` Dave Chinner [this message]
2019-08-07 16:12 ` kbuild test robot
2019-08-07 18:00 ` kbuild test robot
2019-08-01 2:17 ` [PATCH 05/24] shrinker: clean up variable types and tracepoints Dave Chinner
2019-08-01 2:17 ` [PATCH 06/24] mm: reclaim_state records pages reclaimed, not slabs Dave Chinner
2019-08-01 2:17 ` [PATCH 07/24] mm: back off direct reclaim on excessive shrinker deferral Dave Chinner
2019-08-01 2:17 ` [PATCH 08/24] mm: kswapd backoff for shrinkers Dave Chinner
2019-08-01 2:17 ` [PATCH 09/24] xfs: don't allow log IO to be throttled Dave Chinner
2019-08-01 13:39 ` Chris Mason
2019-08-01 23:58 ` Dave Chinner
2019-08-02 8:12 ` Christoph Hellwig
2019-08-02 14:11 ` Chris Mason
2019-08-02 18:34 ` Matthew Wilcox
2019-08-02 23:28 ` Dave Chinner
2019-08-05 18:32 ` Chris Mason
2019-08-05 23:09 ` Dave Chinner
2019-08-01 2:17 ` [PATCH 10/24] xfs: fix missed wakeup on l_flush_wait Dave Chinner
2019-08-01 2:17 ` [PATCH 11/24] xfs:: account for memory freed from metadata buffers Dave Chinner
2019-08-01 8:16 ` Christoph Hellwig
2019-08-01 9:21 ` Dave Chinner
2019-08-06 5:51 ` Christoph Hellwig
2019-08-01 2:17 ` [PATCH 12/24] xfs: correctly acount for reclaimable slabs Dave Chinner
2019-08-06 5:52 ` Christoph Hellwig
2019-08-06 21:05 ` Dave Chinner
2019-08-01 2:17 ` [PATCH 13/24] xfs: synchronous AIL pushing Dave Chinner
2019-08-05 17:51 ` Brian Foster
2019-08-05 23:21 ` Dave Chinner
2019-08-06 12:29 ` Brian Foster
2019-08-01 2:17 ` [PATCH 14/24] xfs: tail updates only need to occur when LSN changes Dave Chinner
2019-08-05 17:53 ` Brian Foster
2019-08-05 23:28 ` Dave Chinner
2019-08-06 5:33 ` Dave Chinner
2019-08-06 12:53 ` Brian Foster
2019-08-06 21:11 ` Dave Chinner
2019-08-01 2:17 ` [PATCH 15/24] xfs: eagerly free shadow buffers to reduce CIL footprint Dave Chinner
2019-08-05 18:03 ` Brian Foster
2019-08-05 23:33 ` Dave Chinner
2019-08-06 12:57 ` Brian Foster
2019-08-06 21:21 ` Dave Chinner
2019-08-01 2:17 ` [PATCH 16/24] xfs: Lower CIL flush limit for large logs Dave Chinner
2019-08-04 17:12 ` Nikolay Borisov
2019-08-01 2:17 ` [PATCH 17/24] xfs: don't block kswapd in inode reclaim Dave Chinner
2019-08-06 18:21 ` Brian Foster
2019-08-06 21:27 ` Dave Chinner
2019-08-07 11:14 ` Brian Foster
2019-08-01 2:17 ` [PATCH 18/24] xfs: reduce kswapd blocking on inode locking Dave Chinner
2019-08-06 18:22 ` Brian Foster
2019-08-06 21:33 ` Dave Chinner
2019-08-07 11:30 ` Brian Foster
2019-08-07 23:16 ` Dave Chinner
2019-08-01 2:17 ` [PATCH 19/24] xfs: kill background reclaim work Dave Chinner
2019-08-01 2:17 ` [PATCH 20/24] xfs: use AIL pushing for inode reclaim IO Dave Chinner
2019-08-07 18:09 ` Brian Foster
2019-08-07 23:10 ` Dave Chinner
2019-08-08 16:20 ` Brian Foster
2019-08-01 2:17 ` [PATCH 21/24] xfs: remove mode from xfs_reclaim_inodes() Dave Chinner
2019-08-01 2:17 ` [PATCH 22/24] xfs: track reclaimable inodes using a LRU list Dave Chinner
2019-08-08 16:36 ` Brian Foster
2019-08-09 0:10 ` Dave Chinner
2019-08-01 2:17 ` [PATCH 23/24] xfs: reclaim inodes from the LRU Dave Chinner
2019-08-08 16:39 ` Brian Foster
2019-08-09 1:20 ` Dave Chinner
2019-08-09 12:36 ` Brian Foster
2019-08-11 2:17 ` Dave Chinner
2019-08-11 12:46 ` Brian Foster
2019-08-01 2:17 ` [PATCH 24/24] xfs: remove unusued old inode reclaim code Dave Chinner
2019-08-06 5:57 ` [RFC] [PATCH 00/24] mm, xfs: non-blocking inode reclaim Christoph Hellwig
2019-08-06 21:37 ` Dave Chinner
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20190804213732.GU7777@dread.disaster.area \
--to=david@fromorbit.com \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-xfs@vger.kernel.org \
--cc=nborisov@suse.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).