From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-10.3 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_PATCH,MAILING_LIST_MULTI,NICE_REPLY_A, SPF_HELO_NONE,SPF_PASS,USER_AGENT_SANE_1 autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id BE655C56202 for ; Thu, 26 Nov 2020 15:44:09 +0000 (UTC) Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.kernel.org (Postfix) with ESMTP id 7BAC221D1A for ; Thu, 26 Nov 2020 15:44:08 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 7BAC221D1A Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=suse.cz Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=owner-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix) id C0F116B005D; Thu, 26 Nov 2020 10:44:07 -0500 (EST) Received: by kanga.kvack.org (Postfix, from userid 40) id BBFEE6B0071; Thu, 26 Nov 2020 10:44:07 -0500 (EST) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id AD4816B0074; Thu, 26 Nov 2020 10:44:07 -0500 (EST) X-Delivered-To: linux-mm@kvack.org Received: from forelay.hostedemail.com (smtprelay0230.hostedemail.com [216.40.44.230]) by kanga.kvack.org (Postfix) with ESMTP id 946C36B005D for ; Thu, 26 Nov 2020 10:44:07 -0500 (EST) Received: from smtpin08.hostedemail.com (10.5.19.251.rfc1918.com [10.5.19.251]) by forelay04.hostedemail.com (Postfix) with ESMTP id 5905D1EE6 for ; Thu, 26 Nov 2020 15:44:07 +0000 (UTC) X-FDA: 77526990534.08.air73_151331227380 Received: from filter.hostedemail.com (10.5.16.251.rfc1918.com [10.5.16.251]) by smtpin08.hostedemail.com (Postfix) with ESMTP id F391B1819E793 for ; Thu, 26 Nov 2020 15:44:06 +0000 (UTC) X-HE-Tag: air73_151331227380 X-Filterd-Recvd-Size: 4472 Received: from mx2.suse.de (mx2.suse.de [195.135.220.15]) by imf45.hostedemail.com (Postfix) with ESMTP for ; Thu, 26 Nov 2020 15:44:06 +0000 (UTC) X-Virus-Scanned: by amavisd-new at test-mx.suse.de Received: from relay2.suse.de (unknown [195.135.221.27]) by mx2.suse.de (Postfix) with ESMTP id 2A8F6ACC4; Thu, 26 Nov 2020 15:44:05 +0000 (UTC) Subject: Re: [PATCH next] mm/swap.c: reduce lock contention in lru_cache_add From: Vlastimil Babka To: Yu Zhao , Alex Shi , Matthew Wilcox Cc: Konstantin Khlebnikov , Andrew Morton , Hugh Dickins , Michal Hocko , linux-mm@kvack.org, linux-kernel@vger.kernel.org References: <1605860847-47445-1-git-send-email-alex.shi@linux.alibaba.com> <20201126045234.GA1014081@google.com> <20201126072402.GA1047005@google.com> <464fa387-9dfd-a8c7-3d86-040f26fd4115@suse.cz> Message-ID: Date: Thu, 26 Nov 2020 16:44:04 +0100 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:78.0) Gecko/20100101 Thunderbird/78.5.0 MIME-Version: 1.0 In-Reply-To: <464fa387-9dfd-a8c7-3d86-040f26fd4115@suse.cz> Content-Type: text/plain; charset=utf-8; format=flowed Content-Language: en-US Content-Transfer-Encoding: quoted-printable X-Bogosity: Ham, tests=bogofilter, spamicity=0.000000, version=1.2.4 Sender: owner-linux-mm@kvack.org Precedence: bulk X-Loop: owner-majordomo@kvack.org List-ID: On 11/26/20 12:22 PM, Vlastimil Babka wrote: > On 11/26/20 8:24 AM, Yu Zhao wrote: >> On Thu, Nov 26, 2020 at 02:39:03PM +0800, Alex Shi wrote: >>>=20 >>>=20 >>> =E5=9C=A8 2020/11/26 =E4=B8=8B=E5=8D=8812:52, Yu Zhao =E5=86=99=E9=81= =93: >>> >> */ >>> >> void __pagevec_lru_add(struct pagevec *pvec) >>> >> { >>> >> - int i; >>> >> - struct lruvec *lruvec =3D NULL; >>> >> + int i, nr_lruvec; >>> >> unsigned long flags =3D 0; >>> >> + struct page *page; >>> >> + struct lruvecs lruvecs; >>> >> =20 >>> >> - for (i =3D 0; i < pagevec_count(pvec); i++) { >>> >> - struct page *page =3D pvec->pages[i]; >>> >> + nr_lruvec =3D sort_page_lruvec(&lruvecs, pvec); >>> > Simply looping pvec multiple times (15 at most) for different lruve= cs >>> > would be better because 1) it requires no extra data structures and >>> > therefore has better cache locality (theoretically faster) 2) it on= ly >>> > loops once when !CONFIG_MEMCG and !CONFIG_NUMA and therefore has no >>> > impact on Android and Chrome OS. >>> >=20 >>>=20 >>> With multiple memcgs, it do help a lot, I had gotten 30% grain on rea= dtwice >>> case. but yes, w/o MEMCG and NUMA, it's good to keep old behavior. So= =20 >>> would you like has a proposal for this? >>=20 >> Oh, no, I'm not against your idea. I was saying it doesn't seem >> necessary to sort -- a nested loop would just do the job given >> pagevec is small. >=20 > Yeah that could work. The worst case doesn't look nice (O(n^2)) but it = should be > rather rare to have pages from really multiple memcgs/nodes? However, Matthew wanted to increase pagevec size [1] and once 15^2 become= s 63^2,=20 it starts to be somewhat more worrying. [1] https://lore.kernel.org/linux-mm/20201105172651.2455-1-willy@infradea= d.org/ > Maybe with the following change? Avoids going through the nulls if we g= ot lucky > (or have !MEMCG !NUMA). >=20 >> diff --git a/mm/swap.c b/mm/swap.c >> index cb3794e13b48..1d238edc2907 100644 >> --- a/mm/swap.c >> +++ b/mm/swap.c >> @@ -996,15 +996,26 @@ static void __pagevec_lru_add_fn(struct page *pa= ge, struct lruvec *lruvec) >> */ >> void __pagevec_lru_add(struct pagevec *pvec) >> { >> - int i; >> + int i, j; >=20 > int i, j, n; >=20 >> struct lruvec *lruvec =3D NULL; >> unsigned long flags =3D 0; >> =20 >=20 > n =3D pagevec_count(pvec); >> for (i =3D 0; i < pagevec_count(pvec); i++) { >=20 > for (i =3D 0; n; i++) { >=20 >> struct page *page =3D pvec->pages[i]; >> =20 >> + if (!page) >> + continue; >> + >> lruvec =3D relock_page_lruvec_irqsave(page, lruvec, &flags); >> - __pagevec_lru_add_fn(page, lruvec); >=20 > n--; >=20 >> + >> + for (j =3D i; j < pagevec_count(pvec); j++) { >> + if (page_to_nid(pvec->pages[j]) !=3D page_to_nid(page) || >> + page_memcg(pvec->pages[j]) !=3D page_memcg(page)) >> + continue; >> + >> + __pagevec_lru_add_fn(pvec->pages[j], lruvec); >> + pvec->pages[j] =3D NULL; >=20 > n--; >> + } >> } >> if (lruvec) >> unlock_page_lruvec_irqrestore(lruvec, flags); >>=20 >=20