From: Alex Shi <alex.shi@linux.alibaba.com>
To: Vlastimil Babka <vbabka@suse.cz>
Cc: Konstantin Khlebnikov <koct9i@gmail.com>,
Andrew Morton <akpm@linux-foundation.org>,
Hugh Dickins <hughd@google.com>, Yu Zhao <yuzhao@google.com>,
Michal Hocko <mhocko@suse.com>,
linux-mm@kvack.org, linux-kernel@vger.kernel.org
Subject: Re: [PATCH next] mm/swap.c: reduce lock contention in lru_cache_add
Date: Thu, 26 Nov 2020 11:12:30 +0800 [thread overview]
Message-ID: <df62191b-e694-1928-2ffc-c4cfb4dea14d@linux.alibaba.com> (raw)
In-Reply-To: <b92a1a50-b7cd-0b33-de2e-52d74c91925c@suse.cz>
在 2020/11/25 下午11:38, Vlastimil Babka 写道:
> On 11/20/20 9:27 AM, Alex Shi wrote:
>> The current relock logical will change lru_lock when found a new
>> lruvec, so if 2 memcgs are reading file or alloc page at same time,
>> they could hold the lru_lock alternately, and wait for each other for
>> fairness attribute of ticket spin lock.
>>
>> This patch will sort that all lru_locks and only hold them once in
>> above scenario. That could reduce fairness waiting for lock reget.
>> Than, vm-scalability/case-lru-file-readtwice could get ~5% performance
>> gain on my 2P*20core*HT machine.
>
> Hm, once you sort the pages like this, it's a shame not to splice them instead of more list_del() + list_add() iterations. update_lru_size() could be also called once?
Yes, looks it's a good idea to use splice instead of list_del/add, but pages
may on different lru list in a same lruvec, and also may come from different
zones. That could involve 5 cycles for different lists, and more for zones...
I give up the try.
next prev parent reply other threads:[~2020-11-26 3:12 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-11-20 8:27 [PATCH next] mm/swap.c: reduce lock contention in lru_cache_add Alex Shi
2020-11-20 23:19 ` Andrew Morton
2020-11-23 4:46 ` Alex Shi
2020-11-25 15:38 ` Vlastimil Babka
2020-11-26 3:12 ` Alex Shi [this message]
2020-11-26 11:05 ` Vlastimil Babka
2020-11-26 4:52 ` Yu Zhao
2020-11-26 6:39 ` Alex Shi
2020-11-26 7:24 ` Yu Zhao
2020-11-26 8:09 ` Alex Shi
2020-11-26 11:22 ` Vlastimil Babka
2020-11-26 15:44 ` Vlastimil Babka
2020-11-26 15:55 ` Matthew Wilcox
2020-11-27 3:14 ` Alex Shi
2020-12-01 8:02 ` [PATCH 1/3] mm/swap.c: pre-sort pages in pagevec for pagevec_lru_move_fn Alex Shi
2020-12-01 8:02 ` [PATCH 2/3] mm/swap.c: bail out early for no memcg and no numa Alex Shi
2020-12-01 8:02 ` [PATCH 3/3] mm/swap.c: extend the usage to pagevec_lru_add Alex Shi
2020-12-01 8:10 ` [PATCH 1/3] mm/swap.c: pre-sort pages in pagevec for pagevec_lru_move_fn Michal Hocko
2020-12-01 8:20 ` Alex Shi
2020-12-25 9:59 ` [RFC PATCH 0/4] pre sort pages on lruvec in pagevec Alex Shi
2020-12-25 9:59 ` [RFC PATCH 1/4] mm/swap.c: pre-sort pages in pagevec for pagevec_lru_move_fn Alex Shi
2020-12-25 9:59 ` [RFC PATCH 2/4] mm/swap.c: bail out early for no memcg and no numa Alex Shi
2020-12-25 9:59 ` [RFC PATCH 3/4] mm/swap.c: extend the usage to pagevec_lru_add 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=df62191b-e694-1928-2ffc-c4cfb4dea14d@linux.alibaba.com \
--to=alex.shi@linux.alibaba.com \
--cc=akpm@linux-foundation.org \
--cc=hughd@google.com \
--cc=koct9i@gmail.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-mm@kvack.org \
--cc=mhocko@suse.com \
--cc=vbabka@suse.cz \
--cc=yuzhao@google.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).