From: Konstantin Khlebnikov <koct9i@gmail.com>
To: Alex Shi <alex.shi@linux.alibaba.com>
Cc: "Andrew Morton" <akpm@linux-foundation.org>,
"Mel Gorman" <mgorman@techsingularity.net>,
"Tejun Heo" <tj@kernel.org>, "Hugh Dickins" <hughd@google.com>,
"Константин Хлебников" <khlebnikov@yandex-team.ru>,
daniel.m.jordan@oracle.com, yang.shi@linux.alibaba.com,
"Matthew Wilcox" <willy@infradead.org>,
"Johannes Weiner" <hannes@cmpxchg.org>,
lkp@intel.com, linux-mm@kvack.org,
"Linux Kernel Mailing List" <linux-kernel@vger.kernel.org>,
Cgroups <cgroups@vger.kernel.org>,
shakeelb@google.com, "Joonsoo Kim" <iamjoonsoo.kim@lge.com>,
richard.weiyang@gmail.com
Subject: Re: [PATCH v14 15/20] mm/swap: serialize memcg changes during pagevec_lru_move_fn
Date: Fri, 3 Jul 2020 12:13:14 +0300 [thread overview]
Message-ID: <CALYGNiOkA_ZsycF_hqm3XLk55Ek1Mo9w1gO=6EeE35fUtA0i_w@mail.gmail.com> (raw)
In-Reply-To: <1593752873-4493-16-git-send-email-alex.shi@linux.alibaba.com>
On Fri, Jul 3, 2020 at 8:09 AM Alex Shi <alex.shi@linux.alibaba.com> wrote:
>
> Hugh Dickins' found a memcg change bug on original version:
> If we want to change the pgdat->lru_lock to memcg's lruvec lock, we have
> to serialize mem_cgroup_move_account during pagevec_lru_move_fn. The
> possible bad scenario would like:
>
> cpu 0 cpu 1
> lruvec = mem_cgroup_page_lruvec()
> if (!isolate_lru_page())
> mem_cgroup_move_account
>
> spin_lock_irqsave(&lruvec->lru_lock <== wrong lock.
>
> So we need the ClearPageLRU to block isolate_lru_page(), then serialize
> the memcg change here.
>
> Reported-by: Hugh Dickins <hughd@google.com>
> Signed-off-by: Alex Shi <alex.shi@linux.alibaba.com>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Cc: linux-mm@kvack.org
> Cc: linux-kernel@vger.kernel.org
> ---
> mm/swap.c | 31 +++++++++++++++++++------------
> 1 file changed, 19 insertions(+), 12 deletions(-)
>
> diff --git a/mm/swap.c b/mm/swap.c
> index b24d5f69b93a..55eb2c2eed03 100644
> --- a/mm/swap.c
> +++ b/mm/swap.c
> @@ -203,7 +203,7 @@ int get_kernel_page(unsigned long start, int write, struct page **pages)
> EXPORT_SYMBOL_GPL(get_kernel_page);
>
> static void pagevec_lru_move_fn(struct pagevec *pvec,
> - void (*move_fn)(struct page *page, struct lruvec *lruvec))
> + void (*move_fn)(struct page *page, struct lruvec *lruvec), bool add)
> {
> int i;
> struct pglist_data *pgdat = NULL;
> @@ -221,8 +221,15 @@ static void pagevec_lru_move_fn(struct pagevec *pvec,
> spin_lock_irqsave(&pgdat->lru_lock, flags);
> }
>
> + /* new page add to lru or page moving between lru */
> + if (!add && !TestClearPageLRU(page))
> + continue;
> +
> lruvec = mem_cgroup_page_lruvec(page, pgdat);
> (*move_fn)(page, lruvec);
> +
> + if (!add)
> + SetPageLRU(page);
> }
> if (pgdat)
> spin_unlock_irqrestore(&pgdat->lru_lock, flags);
> @@ -259,7 +266,7 @@ void rotate_reclaimable_page(struct page *page)
> local_lock_irqsave(&lru_rotate.lock, flags);
> pvec = this_cpu_ptr(&lru_rotate.pvec);
> if (!pagevec_add(pvec, page) || PageCompound(page))
> - pagevec_lru_move_fn(pvec, pagevec_move_tail_fn);
> + pagevec_lru_move_fn(pvec, pagevec_move_tail_fn, false);
> local_unlock_irqrestore(&lru_rotate.lock, flags);
> }
> }
> @@ -328,7 +335,7 @@ static void activate_page_drain(int cpu)
> struct pagevec *pvec = &per_cpu(lru_pvecs.activate_page, cpu);
>
> if (pagevec_count(pvec))
> - pagevec_lru_move_fn(pvec, __activate_page);
> + pagevec_lru_move_fn(pvec, __activate_page, false);
> }
>
> static bool need_activate_page_drain(int cpu)
> @@ -346,7 +353,7 @@ void activate_page(struct page *page)
> pvec = this_cpu_ptr(&lru_pvecs.activate_page);
> get_page(page);
> if (!pagevec_add(pvec, page) || PageCompound(page))
> - pagevec_lru_move_fn(pvec, __activate_page);
> + pagevec_lru_move_fn(pvec, __activate_page, false);
> local_unlock(&lru_pvecs.lock);
> }
> }
> @@ -621,21 +628,21 @@ void lru_add_drain_cpu(int cpu)
>
> /* No harm done if a racing interrupt already did this */
> local_lock_irqsave(&lru_rotate.lock, flags);
> - pagevec_lru_move_fn(pvec, pagevec_move_tail_fn);
> + pagevec_lru_move_fn(pvec, pagevec_move_tail_fn, false);
> local_unlock_irqrestore(&lru_rotate.lock, flags);
> }
>
> pvec = &per_cpu(lru_pvecs.lru_deactivate_file, cpu);
> if (pagevec_count(pvec))
> - pagevec_lru_move_fn(pvec, lru_deactivate_file_fn);
> + pagevec_lru_move_fn(pvec, lru_deactivate_file_fn, false);
>
> pvec = &per_cpu(lru_pvecs.lru_deactivate, cpu);
> if (pagevec_count(pvec))
> - pagevec_lru_move_fn(pvec, lru_deactivate_fn);
> + pagevec_lru_move_fn(pvec, lru_deactivate_fn, false);
>
> pvec = &per_cpu(lru_pvecs.lru_lazyfree, cpu);
> if (pagevec_count(pvec))
> - pagevec_lru_move_fn(pvec, lru_lazyfree_fn);
> + pagevec_lru_move_fn(pvec, lru_lazyfree_fn, false);
>
> activate_page_drain(cpu);
> }
> @@ -664,7 +671,7 @@ void deactivate_file_page(struct page *page)
> pvec = this_cpu_ptr(&lru_pvecs.lru_deactivate_file);
>
> if (!pagevec_add(pvec, page) || PageCompound(page))
> - pagevec_lru_move_fn(pvec, lru_deactivate_file_fn);
> + pagevec_lru_move_fn(pvec, lru_deactivate_file_fn, false);
> local_unlock(&lru_pvecs.lock);
> }
> }
> @@ -686,7 +693,7 @@ void deactivate_page(struct page *page)
> pvec = this_cpu_ptr(&lru_pvecs.lru_deactivate);
> get_page(page);
> if (!pagevec_add(pvec, page) || PageCompound(page))
> - pagevec_lru_move_fn(pvec, lru_deactivate_fn);
> + pagevec_lru_move_fn(pvec, lru_deactivate_fn, false);
> local_unlock(&lru_pvecs.lock);
> }
> }
> @@ -708,7 +715,7 @@ void mark_page_lazyfree(struct page *page)
> pvec = this_cpu_ptr(&lru_pvecs.lru_lazyfree);
> get_page(page);
> if (!pagevec_add(pvec, page) || PageCompound(page))
> - pagevec_lru_move_fn(pvec, lru_lazyfree_fn);
> + pagevec_lru_move_fn(pvec, lru_lazyfree_fn, false);
> local_unlock(&lru_pvecs.lock);
> }
> }
> @@ -976,7 +983,7 @@ static void __pagevec_lru_add_fn(struct page *page, struct lruvec *lruvec)
> */
> void __pagevec_lru_add(struct pagevec *pvec)
> {
> - pagevec_lru_move_fn(pvec, __pagevec_lru_add_fn);
> + pagevec_lru_move_fn(pvec, __pagevec_lru_add_fn, true);
> }
It seems better to open code version in lru_add than adding a bool
argument which is true just for one user.
Also with this new lru protection logic lru_add could be optimized:
It could prepare a list of pages and under lru_lock do only list
splice and bumping counter.
Since PageLRU isn't set yet nobody could touch these pages in lru.
After that lru_add could iterate pages from first to last without
lru_lock to set PageLRU and drop reference.
So, lru_add will do O(1) operations under lru_lock regardless of the
count of pages it added.
Actually per-cpu vector for adding could be replaced with per-cpu
lists and\or per-lruvec atomic slist.
Thus incommig pages will be already in list structure rather than page vector.
This allows to accumulate more pages and offload adding to kswapd or
direct reclaim.
>
> /**
> --
> 1.8.3.1
>
>
next prev parent reply other threads:[~2020-07-03 9:13 UTC|newest]
Thread overview: 35+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-07-03 5:07 [PATCH v14 00/20] per memcg lru lock Alex Shi
2020-07-03 5:07 ` [PATCH v14 01/20] mm/vmscan: remove unnecessary lruvec adding Alex Shi
2020-07-03 5:07 ` [PATCH v14 02/20] mm/page_idle: no unlikely double check for idle page counting Alex Shi
2020-07-03 5:07 ` [PATCH v14 03/20] mm/compaction: correct the comments of compact_defer_shift Alex Shi
2020-07-03 5:07 ` [PATCH v14 04/20] mm/compaction: rename compact_deferred as compact_should_defer Alex Shi
2020-07-03 5:07 ` [PATCH v14 05/20] mm/thp: move lru_add_page_tail func to huge_memory.c Alex Shi
2020-07-03 5:07 ` [PATCH v14 06/20] mm/thp: clean up lru_add_page_tail Alex Shi
2020-07-03 5:07 ` [PATCH v14 07/20] mm/thp: narrow lru locking Alex Shi
2020-07-06 9:15 ` Alex Shi
2020-07-06 11:35 ` Matthew Wilcox
2020-07-07 4:52 ` Hugh Dickins
2020-07-09 14:02 ` Alex Shi
2020-07-09 15:48 ` Kirill A. Shutemov
2020-07-10 8:23 ` Alex Shi
[not found] ` <20200710112831.jrv4hzjzjqtxtc7u@box>
2020-07-10 14:09 ` Alex Shi
2020-07-07 10:51 ` Alex Shi
2020-07-03 5:07 ` [PATCH v14 08/20] mm/memcg: add debug checking in lock_page_memcg Alex Shi
2020-07-03 5:07 ` [PATCH v14 09/20] mm/swap: fold vm event PGROTATED into pagevec_move_tail_fn Alex Shi
2020-07-03 5:07 ` [PATCH v14 10/20] mm/lru: move lru_lock holding in func lru_note_cost_page Alex Shi
2020-07-03 5:07 ` [PATCH v14 11/20] mm/lru: move lock into lru_note_cost Alex Shi
2020-07-03 5:07 ` [PATCH v14 12/20] mm/lru: introduce TestClearPageLRU Alex Shi
2020-07-03 5:07 ` [PATCH v14 13/20] mm/compaction: do page isolation first in compaction Alex Shi
2020-07-03 5:07 ` [PATCH v14 14/20] mm/mlock: reorder isolation sequence during munlock Alex Shi
2020-07-03 5:07 ` [PATCH v14 15/20] mm/swap: serialize memcg changes during pagevec_lru_move_fn Alex Shi
2020-07-03 9:13 ` Konstantin Khlebnikov [this message]
2020-07-04 11:34 ` Alex Shi
2020-07-04 11:39 ` Matthew Wilcox
2020-07-04 13:12 ` Alex Shi
2020-07-04 13:33 ` Matthew Wilcox
2020-07-04 15:47 ` Alex Shi
2020-07-03 5:07 ` [PATCH v14 16/20] mm/lru: replace pgdat lru_lock with lruvec lock Alex Shi
2020-07-03 5:07 ` [PATCH v14 17/20] mm/lru: introduce the relock_page_lruvec function Alex Shi
2020-07-03 5:07 ` [PATCH v14 18/20] mm/vmscan: use relock for move_pages_to_lru Alex Shi
2020-07-03 5:07 ` [PATCH v14 19/20] mm/pgdat: remove pgdat lru_lock Alex Shi
2020-07-03 5:07 ` [PATCH v14 20/20] mm/lru: revise the comments of lru_lock Alex Shi
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='CALYGNiOkA_ZsycF_hqm3XLk55Ek1Mo9w1gO=6EeE35fUtA0i_w@mail.gmail.com' \
--to=koct9i@gmail.com \
--cc=akpm@linux-foundation.org \
--cc=alex.shi@linux.alibaba.com \
--cc=cgroups@vger.kernel.org \
--cc=daniel.m.jordan@oracle.com \
--cc=hannes@cmpxchg.org \
--cc=hughd@google.com \
--cc=iamjoonsoo.kim@lge.com \
--cc=khlebnikov@yandex-team.ru \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=lkp@intel.com \
--cc=mgorman@techsingularity.net \
--cc=richard.weiyang@gmail.com \
--cc=shakeelb@google.com \
--cc=tj@kernel.org \
--cc=willy@infradead.org \
--cc=yang.shi@linux.alibaba.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).