From: Vlastimil Babka <vbabka@suse.cz>
To: Mel Gorman <mgorman@techsingularity.net>,
Andrew Morton <akpm@linux-foundation.org>
Cc: Nicolas Saenz Julienne <nsaenzju@redhat.com>,
Marcelo Tosatti <mtosatti@redhat.com>,
Michal Hocko <mhocko@kernel.org>, Hugh Dickins <hughd@google.com>,
LKML <linux-kernel@vger.kernel.org>,
Linux-MM <linux-mm@kvack.org>
Subject: Re: [PATCH 6/7] mm/page_alloc: Remotely drain per-cpu lists
Date: Thu, 16 Jun 2022 18:41:44 +0200 [thread overview]
Message-ID: <4a5d90b3-afe4-7c13-b989-3ece4a797b1f@suse.cz> (raw)
In-Reply-To: <20220613125622.18628-7-mgorman@techsingularity.net>
On 6/13/22 14:56, Mel Gorman wrote:
> From: Nicolas Saenz Julienne <nsaenzju@redhat.com>
>
> Some setups, notably NOHZ_FULL CPUs, are too busy to handle the per-cpu
> drain work queued by __drain_all_pages(). So introduce a new mechanism to
> remotely drain the per-cpu lists. It is made possible by remotely locking
> 'struct per_cpu_pages' new per-cpu spinlocks. A benefit of this new
> scheme is that drain operations are now migration safe.
>
> There was no observed performance degradation vs. the previous scheme.
> Both netperf and hackbench were run in parallel to triggering the
> __drain_all_pages(NULL, true) code path around ~100 times per second. The
> new scheme performs a bit better (~5%), although the important point here
> is there are no performance regressions vs. the previous mechanism.
> Per-cpu lists draining happens only in slow paths.
>
> Minchan Kim tested this independently and reported;
>
> My workload is not NOHZ CPUs but run apps under heavy memory
> pressure so they goes to direct reclaim and be stuck on
> drain_all_pages until work on workqueue run.
>
> unit: nanosecond
> max(dur) avg(dur) count(dur)
> 166713013 487511.77786438033 1283
>
> From traces, system encountered the drain_all_pages 1283 times and
> worst case was 166ms and avg was 487us.
>
> The other problem was alloc_contig_range in CMA. The PCP draining
> takes several hundred millisecond sometimes though there is no
> memory pressure or a few of pages to be migrated out but CPU were
> fully booked.
>
> Your patch perfectly removed those wasted time.
>
> Signed-off-by: Nicolas Saenz Julienne <nsaenzju@redhat.com>
> Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
> Tested-by: Minchan Kim <minchan@kernel.org>
> Acked-by: Minchan Kim <minchan@kernel.org>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
next prev parent reply other threads:[~2022-06-16 16:41 UTC|newest]
Thread overview: 40+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-06-13 12:56 [PATCH v4 00/7] Drain remote per-cpu directly Mel Gorman
2022-06-13 12:56 ` [PATCH 1/7] mm/page_alloc: Add page->buddy_list and page->pcp_list Mel Gorman
2022-06-13 12:56 ` [PATCH 2/7] mm/page_alloc: Use only one PCP list for THP-sized allocations Mel Gorman
2022-06-13 12:56 ` [PATCH 3/7] mm/page_alloc: Split out buddy removal code from rmqueue into separate helper Mel Gorman
2022-06-13 12:56 ` [PATCH 4/7] mm/page_alloc: Remove mistaken page == NULL check in rmqueue Mel Gorman
2022-06-13 12:56 ` [PATCH 5/7] mm/page_alloc: Protect PCP lists with a spinlock Mel Gorman
2022-06-16 15:59 ` Vlastimil Babka
2022-06-13 12:56 ` [PATCH 6/7] mm/page_alloc: Remotely drain per-cpu lists Mel Gorman
2022-06-16 16:41 ` Vlastimil Babka [this message]
2022-06-13 12:56 ` [PATCH 7/7] mm/page_alloc: Replace local_lock with normal spinlock Mel Gorman
2022-06-15 22:43 ` Yu Zhao
[not found] ` <CGME20220615224855eucas1p1ea6d90c23ec9423dfe04b267f6dddd2a@eucas1p1.samsung.com>
2022-06-15 22:48 ` Marek Szyprowski
2022-06-15 23:04 ` Andrew Morton
2022-06-16 3:05 ` Yu Zhao
2022-06-17 7:55 ` Vlastimil Babka
2022-06-17 6:47 ` Marek Szyprowski
2022-06-21 9:21 ` Mel Gorman
2022-06-16 17:01 ` Vlastimil Babka
2022-06-16 21:07 ` Yu Zhao
2022-06-17 7:57 ` Vlastimil Babka
2022-06-21 9:27 ` Mel Gorman
2022-06-21 9:26 ` Mel Gorman
2022-06-17 9:39 ` Nicolas Saenz Julienne
2022-06-21 9:29 ` Mel Gorman
2022-06-21 9:31 ` Nicolas Saenz Julienne
2022-07-03 9:44 ` [mm/page_alloc] 2bd8eec68f: BUG:sleeping_function_called_from_invalid_context_at_mm/gup.c kernel test robot
2022-07-03 20:22 ` Andrew Morton
2022-07-05 13:51 ` Oliver Sang
2022-07-06 9:55 ` Mel Gorman
2022-07-06 11:53 ` Mel Gorman
2022-07-06 14:21 ` Oliver Sang
2022-07-06 14:52 ` Mel Gorman
2022-07-07 8:22 ` Oliver Sang
2022-07-06 14:25 ` Oliver Sang
2022-07-06 14:53 ` Mel Gorman
2022-07-07 21:55 ` Vlastimil Babka
2022-07-08 10:56 ` Mel Gorman
2022-07-12 5:04 ` Oliver Sang
2022-06-24 12:54 [PATCH v5 00/7] Drain remote per-cpu directly Mel Gorman
2022-06-24 12:54 ` [PATCH 6/7] mm/page_alloc: Remotely drain per-cpu lists Mel Gorman
2022-07-04 14:28 ` Vlastimil Babka
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=4a5d90b3-afe4-7c13-b989-3ece4a797b1f@suse.cz \
--to=vbabka@suse.cz \
--cc=akpm@linux-foundation.org \
--cc=hughd@google.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mgorman@techsingularity.net \
--cc=mhocko@kernel.org \
--cc=mtosatti@redhat.com \
--cc=nsaenzju@redhat.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).