From: Brian Foster <bfoster@redhat.com>
To: Dave Chinner <david@fromorbit.com>
Cc: linux-xfs@vger.kernel.org, linux-fsdevel@vger.kernel.org,
linux-mm@kvack.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH 12/28] shrinker: defer work only to kswapd
Date: Mon, 4 Nov 2019 10:29:54 -0500 [thread overview]
Message-ID: <20191104152954.GC10665@bfoster> (raw)
In-Reply-To: <20191031234618.15403-13-david@fromorbit.com>
On Fri, Nov 01, 2019 at 10:46:02AM +1100, 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.
>
> Note that we use signed integers for everything except the freed
> count as the returns from the shrinker callouts cannot be guaranteed
> untainted. Indeed, the shrinkers can return scan counts larger that
> were fed in, so we need scan counts to underflow in a detectable
> manner to terminate loops. This is necessary to avoid a misbehaving
> shrinker from triggering endless scanning loops.
>
> Signed-off-by: Dave Chinner <dchinner@redhat.com>
> ---
> include/linux/shrinker.h | 2 +-
> mm/vmscan.c | 100 ++++++++++++++++++++-------------------
> 2 files changed, 53 insertions(+), 49 deletions(-)
>
> diff --git a/include/linux/shrinker.h b/include/linux/shrinker.h
> index 3405c39ab92c..30c10f42109f 100644
> --- a/include/linux/shrinker.h
> +++ b/include/linux/shrinker.h
> @@ -81,7 +81,7 @@ struct shrinker {
> int id;
> #endif
> /* objs pending delete, per node */
> - atomic_long_t *nr_deferred;
> + atomic64_t *nr_deferred;
> };
> #define DEFAULT_SEEKS 2 /* A good number if you don't know better. */
>
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index 2d39ec37c04d..c0e2bf656e3f 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -517,16 +517,16 @@ static int64_t shrink_scan_count(struct shrink_control *shrinkctl,
> static unsigned long do_shrink_slab(struct shrink_control *shrinkctl,
> struct shrinker *shrinker, int priority)
> {
> - unsigned long freed = 0;
> - long total_scan;
> + uint64_t freed = 0;
> int64_t freeable_objects = 0;
> int64_t scan_count;
> - long nr;
> - long new_nr;
> + int64_t scanned_objects = 0;
> + int64_t next_deferred = 0;
> + int64_t deferred_count = 0;
> + int64_t 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;
> @@ -537,47 +537,51 @@ 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.
> - */
> - nr = atomic_long_xchg(&shrinker->nr_deferred[nid], 0);
> -
> - 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;
> -
> - /*
> - * 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.
> + * 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.
> *
> - * Hence only allow the shrinker to scan the entire cache when
> - * a large delta change is calculated directly.
> + * 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.
That last sentence is hard to read. ;)
> */
> - if (scan_count < freeable_objects / 4)
> - total_scan = min_t(long, total_scan, freeable_objects / 2);
> + if (current_is_kswapd()) {
> + int64_t deferred_scan;
> +
> + deferred_count = atomic64_xchg(&shrinker->nr_deferred[nid], 0);
> +
> + /* 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.
> + */
> + deferred_count = min(deferred_count, freeable_objects * 2);
> +
Extra whitespace above.
> + }
>
> /*
> * Avoid risking looping forever due to too large nr value:
> * never try to free more than twice the estimate number of
> * freeable entries.
> */
The comment refers to a variable that no longer exists.
I also wonder if it's a little cleaner to move the deferred_count =
min(...); statement above down here and condense the two comments.
> - if (total_scan > freeable_objects * 2)
> - total_scan = freeable_objects * 2;
> + scan_count = min(scan_count, freeable_objects * 2);
>
> - trace_mm_shrink_slab_start(shrinker, shrinkctl, nr,
> + trace_mm_shrink_slab_start(shrinker, shrinkctl, deferred_count,
> freeable_objects, scan_count,
> - total_scan, priority);
> + scan_count, priority);
>
> /*
> * If the shrinker can't run (e.g. due to gfp_mask constraints), then
> @@ -601,10 +605,10 @@ static unsigned long do_shrink_slab(struct shrink_control *shrinkctl,
> * scanning at high prio and therefore should try to reclaim as much as
> * possible.
> */
> - while (total_scan >= batch_size ||
> - total_scan >= freeable_objects) {
> + while (scan_count >= batch_size ||
> + scan_count >= freeable_objects) {
> unsigned long ret;
> - unsigned long nr_to_scan = min(batch_size, total_scan);
> + unsigned long nr_to_scan = min_t(long, batch_size, scan_count);
>
> shrinkctl->nr_to_scan = nr_to_scan;
> shrinkctl->nr_scanned = nr_to_scan;
> @@ -614,29 +618,29 @@ static unsigned long do_shrink_slab(struct shrink_control *shrinkctl,
> freed += ret;
>
> count_vm_events(SLABS_SCANNED, shrinkctl->nr_scanned);
> - total_scan -= shrinkctl->nr_scanned;
> - scanned += shrinkctl->nr_scanned;
> + scan_count -= shrinkctl->nr_scanned;
> + scanned_objects += shrinkctl->nr_scanned;
>
> cond_resched();
> }
> -
> done:
> - if (next_deferred >= scanned)
> - next_deferred -= scanned;
> + if (deferred_count)
> + next_deferred = deferred_count - scanned_objects;
> else
> - next_deferred = 0;
> + next_deferred = scan_count;
Hmm.. so if there was no deferred count on this cycle, we set
next_deferred to whatever is left from scan_count and add that back into
the shrinker struct below. If there was a pending deferred count on this
cycle, we subtract what we scanned from that and add that value back.
But what happens to the remaining scan_count in the latter case? Is it
lost, or am I missing something?
For example, suppose we start this cycle with a large scan_count and
->scan_objects() returned SHRINK_STOP before doing much work. In that
scenario, it looks like whether ->nr_deferred is 0 or not is the only
thing that determines whether we defer the entire remaining scan_count
or just what is left from the previous ->nr_deferred. The existing code
appears to consistently factor in what is left from the current scan
with the previous deferred count. Hm?
> /*
> * move the unused scan count back into the shrinker in a
> * manner that handles concurrent updates. If we exhausted the
> * scan, there is no need to do an update.
> */
> if (next_deferred > 0)
> - new_nr = atomic_long_add_return(next_deferred,
> + new_nr = atomic64_add_return(next_deferred,
> &shrinker->nr_deferred[nid]);
> else
> - new_nr = atomic_long_read(&shrinker->nr_deferred[nid]);
> + new_nr = atomic64_read(&shrinker->nr_deferred[nid]);
It looks like we could kill new_nr and just reuse next_deferred here
too.
Brian
>
> - trace_mm_shrink_slab_end(shrinker, nid, freed, nr, new_nr, total_scan);
> + trace_mm_shrink_slab_end(shrinker, nid, freed, deferred_count, new_nr,
> + scan_count);
> return freed;
> }
>
> --
> 2.24.0.rc0
>
next prev parent reply other threads:[~2019-11-04 15:30 UTC|newest]
Thread overview: 72+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-10-31 23:45 [PATCH 00/28] mm, xfs: non-blocking inode reclaim Dave Chinner
2019-10-31 23:45 ` [PATCH 01/28] xfs: Lower CIL flush limit for large logs Dave Chinner
2019-10-31 23:45 ` [PATCH 02/28] xfs: Throttle commits on delayed background CIL push Dave Chinner
2019-11-01 12:04 ` Brian Foster
2019-11-01 21:40 ` Dave Chinner
2019-11-04 22:48 ` Darrick J. Wong
2019-10-31 23:45 ` [PATCH 03/28] xfs: don't allow log IO to be throttled Dave Chinner
2019-10-31 23:45 ` [PATCH 04/28] xfs: Improve metadata buffer reclaim accountability Dave Chinner
2019-11-01 12:05 ` Brian Foster
2019-11-04 23:21 ` Darrick J. Wong
2019-10-31 23:45 ` [PATCH 05/28] xfs: correctly acount for reclaimable slabs Dave Chinner
2019-10-31 23:45 ` [PATCH 06/28] xfs: factor common AIL item deletion code Dave Chinner
2019-11-04 23:16 ` Darrick J. Wong
2019-10-31 23:45 ` [PATCH 07/28] xfs: tail updates only need to occur when LSN changes Dave Chinner
2019-11-04 23:18 ` Darrick J. Wong
2019-10-31 23:45 ` [PATCH 08/28] xfs: factor inode lookup from xfs_ifree_cluster Dave Chinner
2019-11-01 12:05 ` Brian Foster
2019-11-04 23:20 ` Darrick J. Wong
2019-10-31 23:45 ` [PATCH 09/28] mm: directed shrinker work deferral Dave Chinner
2019-11-04 15:25 ` Brian Foster
2019-11-14 20:49 ` Dave Chinner
2019-11-15 17:21 ` Brian Foster
2019-11-18 0:49 ` Dave Chinner
2019-11-19 15:12 ` Brian Foster
2019-10-31 23:46 ` [PATCH 10/28] shrinkers: use defer_work for GFP_NOFS sensitive shrinkers Dave Chinner
2019-10-31 23:46 ` [PATCH 11/28] mm: factor shrinker work calculations Dave Chinner
2019-11-02 10:55 ` kbuild test robot
2019-11-04 15:29 ` Brian Foster
2019-11-14 20:59 ` Dave Chinner
2019-10-31 23:46 ` [PATCH 12/28] shrinker: defer work only to kswapd Dave Chinner
2019-11-04 15:29 ` Brian Foster [this message]
2019-11-14 21:11 ` Dave Chinner
2019-11-15 17:23 ` Brian Foster
2019-10-31 23:46 ` [PATCH 13/28] shrinker: clean up variable types and tracepoints Dave Chinner
2019-11-04 15:30 ` Brian Foster
2019-10-31 23:46 ` [PATCH 14/28] mm: reclaim_state records pages reclaimed, not slabs Dave Chinner
2019-11-04 19:58 ` Brian Foster
2019-10-31 23:46 ` [PATCH 15/28] mm: back off direct reclaim on excessive shrinker deferral Dave Chinner
2019-11-04 19:58 ` Brian Foster
2019-11-14 21:28 ` Dave Chinner
2019-10-31 23:46 ` [PATCH 16/28] mm: kswapd backoff for shrinkers Dave Chinner
2019-11-04 19:58 ` Brian Foster
2019-11-14 21:41 ` Dave Chinner
2019-10-31 23:46 ` [PATCH 17/28] xfs: synchronous AIL pushing Dave Chinner
2019-11-05 17:05 ` Brian Foster
2019-10-31 23:46 ` [PATCH 18/28] xfs: don't block kswapd in inode reclaim Dave Chinner
2019-10-31 23:46 ` [PATCH 19/28] xfs: reduce kswapd blocking on inode locking Dave Chinner
2019-11-05 17:05 ` Brian Foster
2019-10-31 23:46 ` [PATCH 20/28] xfs: kill background reclaim work Dave Chinner
2019-11-05 17:05 ` Brian Foster
2019-10-31 23:46 ` [PATCH 21/28] xfs: use AIL pushing for inode reclaim IO Dave Chinner
2019-11-05 17:06 ` Brian Foster
2019-10-31 23:46 ` [PATCH 22/28] xfs: remove mode from xfs_reclaim_inodes() Dave Chinner
2019-10-31 23:46 ` [PATCH 23/28] xfs: track reclaimable inodes using a LRU list Dave Chinner
2019-10-31 23:46 ` [PATCH 24/28] xfs: reclaim inodes from the LRU Dave Chinner
2019-11-06 17:21 ` Brian Foster
2019-11-14 21:51 ` Dave Chinner
2019-10-31 23:46 ` [PATCH 25/28] xfs: remove unusued old inode reclaim code Dave Chinner
2019-11-06 17:21 ` Brian Foster
2019-10-31 23:46 ` [PATCH 26/28] xfs: use xfs_ail_push_all in xfs_reclaim_inodes Dave Chinner
2019-11-06 17:22 ` Brian Foster
2019-11-14 21:53 ` Dave Chinner
2019-10-31 23:46 ` [PATCH 27/28] rwsem: introduce down/up_write_non_owner Dave Chinner
2019-10-31 23:46 ` [PATCH 28/28] xfs: rework unreferenced inode lookups Dave Chinner
2019-11-06 22:18 ` Brian Foster
2019-11-14 22:16 ` Dave Chinner
2019-11-15 13:13 ` Christoph Hellwig
2019-11-15 17:26 ` Brian Foster
2019-11-18 1:00 ` Dave Chinner
2019-11-19 15:13 ` Brian Foster
2019-11-19 21:18 ` Dave Chinner
2019-11-20 12:42 ` Brian Foster
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=20191104152954.GC10665@bfoster \
--to=bfoster@redhat.com \
--cc=david@fromorbit.com \
--cc=linux-fsdevel@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=linux-xfs@vger.kernel.org \
/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).