From patchwork Sat Nov 16 03:15:03 2019 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Alex Shi X-Patchwork-Id: 11247439 Return-Path: Received: from mail.kernel.org (pdx-korg-mail-1.web.codeaurora.org [172.30.200.123]) by pdx-korg-patchwork-2.web.codeaurora.org (Postfix) with ESMTP id C8556930 for ; Sat, 16 Nov 2019 03:15:26 +0000 (UTC) Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.kernel.org (Postfix) with ESMTP id 8422C20700 for ; Sat, 16 Nov 2019 03:15:26 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 8422C20700 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=linux.alibaba.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=owner-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix) id E152A6B000A; Fri, 15 Nov 2019 22:15:18 -0500 (EST) Delivered-To: linux-mm-outgoing@kvack.org Received: by kanga.kvack.org (Postfix, from userid 40) id D9ED36B000C; Fri, 15 Nov 2019 22:15:18 -0500 (EST) X-Original-To: int-list-linux-mm@kvack.org X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id C3E9F6B000D; Fri, 15 Nov 2019 22:15:18 -0500 (EST) X-Original-To: linux-mm@kvack.org X-Delivered-To: linux-mm@kvack.org Received: from forelay.hostedemail.com (smtprelay0229.hostedemail.com [216.40.44.229]) by kanga.kvack.org (Postfix) with ESMTP id AD11C6B000A for ; Fri, 15 Nov 2019 22:15:18 -0500 (EST) Received: from smtpin15.hostedemail.com (10.5.19.251.rfc1918.com [10.5.19.251]) by forelay03.hostedemail.com (Postfix) with SMTP id 721DC82499A8 for ; Sat, 16 Nov 2019 03:15:18 +0000 (UTC) X-FDA: 76160674716.15.dress12_1109b33839804 X-Spam-Summary: 2,0,0,682e8a5b31896e6f,d41d8cd98f00b204,alex.shi@linux.alibaba.com,:alex.shi@linux.alibaba.com:cgroups@vger.kernel.org:linux-kernel@vger.kernel.org::akpm@linux-foundation.org:mgorman@techsingularity.net:tj@kernel.org:hughd@google.com:khlebnikov@yandex-team.ru:daniel.m.jordan@oracle.com:yang.shi@linux.alibaba.com:willy@infradead.org:hannes@cmpxchg.org:mhocko@kernel.org:vdavydov.dev@gmail.com:guro@fb.com:shakeelb@google.com:chris@chrisdown.name:tglx@linutronix.de:vbabka@suse.cz:aryabinin@virtuozzo.com:swkhack@gmail.com:stefan.potyra@elektrobit.com:jgg@ziepe.ca:mchehab+samsung@kernel.org:peng.fan@nxp.com:nborisov@suse.com:ira.weiny@intel.com:ktkhai@virtuozzo.com:laoar.shao@gmail.com,RULES_HIT:1:2:41:69:355:379:541:800:960:966:973:981:988:989:1260:1261:1345:1359:1431:1437:1605:1730:1747:1777:1792:2196:2198:2199:2200:2393:2553:2559:2562:2898:3138:3139:3140:3141:3142:3167:3865:3866:3867:3868:3870:3871:3874:4051:4250:4321:4385:4605:5007:6119:6261:6737:6742:7514:7903: 8957:901 X-HE-Tag: dress12_1109b33839804 X-Filterd-Recvd-Size: 10927 Received: from out30-133.freemail.mail.aliyun.com (out30-133.freemail.mail.aliyun.com [115.124.30.133]) by imf03.hostedemail.com (Postfix) with ESMTP for ; Sat, 16 Nov 2019 03:15:16 +0000 (UTC) X-Alimail-AntiSpam: AC=PASS;BC=-1|-1;BR=01201311R841e4;CH=green;DM=||false|;FP=0|-1|-1|-1|0|-1|-1|-1;HT=e01e04426;MF=alex.shi@linux.alibaba.com;NM=1;PH=DS;RN=30;SR=0;TI=SMTPD_---0TiC5Qf._1573874110; Received: from localhost(mailfrom:alex.shi@linux.alibaba.com fp:SMTPD_---0TiC5Qf._1573874110) by smtp.aliyun-inc.com(127.0.0.1); Sat, 16 Nov 2019 11:15:10 +0800 From: Alex Shi To: alex.shi@linux.alibaba.com, cgroups@vger.kernel.org, linux-kernel@vger.kernel.org, linux-mm@kvack.org, akpm@linux-foundation.org, mgorman@techsingularity.net, tj@kernel.org, hughd@google.com, khlebnikov@yandex-team.ru, daniel.m.jordan@oracle.com, yang.shi@linux.alibaba.com, willy@infradead.org Cc: Johannes Weiner , Michal Hocko , Vladimir Davydov , Roman Gushchin , Shakeel Butt , Chris Down , Thomas Gleixner , Vlastimil Babka , Andrey Ryabinin , swkhack , "Potyra, Stefan" , Jason Gunthorpe , Mauro Carvalho Chehab , Peng Fan , Nikolay Borisov , Ira Weiny , Kirill Tkhai , Yafang Shao Subject: [PATCH v3 4/7] mm/lru: only change the lru_lock iff page's lruvec is different Date: Sat, 16 Nov 2019 11:15:03 +0800 Message-Id: <1573874106-23802-5-git-send-email-alex.shi@linux.alibaba.com> X-Mailer: git-send-email 1.8.3.1 In-Reply-To: <1573874106-23802-1-git-send-email-alex.shi@linux.alibaba.com> References: <1573874106-23802-1-git-send-email-alex.shi@linux.alibaba.com> 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: During the pagevec locking, a new page's lruvec is may same as previous one. Thus we could save a re-locking, and only change lock iff lruvec is new. Function named relock_page_lruvec following Hugh Dickins patch. The first version of this patch used rcu_read_lock to guard lruvec assign and comparsion with locked_lruvev in relock_page_lruvec. But Rong Chen report a regression with PROVE_LOCKING config. The rcu_read locking causes qspinlock waiting to be locked for too long. Since we had hold a spinlock, rcu_read locking isn't necessary. [lkp@intel.com: Fix RCU-related regression reported by LKP robot] Signed-off-by: Alex Shi Cc: Johannes Weiner Cc: Michal Hocko Cc: Vladimir Davydov Cc: Andrew Morton Cc: Roman Gushchin Cc: Shakeel Butt Cc: Chris Down Cc: Thomas Gleixner Cc: Vlastimil Babka Cc: Andrey Ryabinin Cc: swkhack Cc: "Potyra, Stefan" Cc: Jason Gunthorpe Cc: Matthew Wilcox Cc: Mauro Carvalho Chehab Cc: Peng Fan Cc: Nikolay Borisov Cc: Ira Weiny Cc: Kirill Tkhai Cc: Yang Shi Cc: Yafang Shao Cc: Mel Gorman Cc: Konstantin Khlebnikov Cc: Hugh Dickins Cc: Tejun Heo Cc: linux-kernel@vger.kernel.org Cc: cgroups@vger.kernel.org Cc: linux-mm@kvack.org --- include/linux/memcontrol.h | 44 ++++++++++++++++++++++++++++++++++++++++++++ mm/mlock.c | 16 +++++++++------- mm/swap.c | 14 ++++++-------- mm/vmscan.c | 24 ++++++++++++------------ 4 files changed, 71 insertions(+), 27 deletions(-) diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h index 0b32eadd0eda..eaec01fb627f 100644 --- a/include/linux/memcontrol.h +++ b/include/linux/memcontrol.h @@ -1290,6 +1290,50 @@ static inline void dec_lruvec_page_state(struct page *page, mod_lruvec_page_state(page, idx, -1); } +/* Don't lock again iff page's lruvec locked */ +static inline struct lruvec *relock_page_lruvec_irq(struct page *page, + struct lruvec *locked_lruvec) +{ + struct pglist_data *pgdat = page_pgdat(page); + struct lruvec *lruvec; + + if (!locked_lruvec) + goto lock; + + lruvec = mem_cgroup_page_lruvec(page, pgdat); + + if (locked_lruvec == lruvec) + return lruvec; + + spin_unlock_irq(&locked_lruvec->lru_lock); + +lock: + lruvec = lock_page_lruvec_irq(page, pgdat); + return lruvec; +} + +/* Don't lock again iff page's lruvec locked */ +static inline struct lruvec *relock_page_lruvec_irqsave(struct page *page, + struct lruvec *locked_lruvec) +{ + struct pglist_data *pgdat = page_pgdat(page); + struct lruvec *lruvec; + + if (!locked_lruvec) + goto lock; + + lruvec = mem_cgroup_page_lruvec(page, pgdat); + + if (locked_lruvec == lruvec) + return lruvec; + + spin_unlock_irqrestore(&locked_lruvec->lru_lock, locked_lruvec->irqflags); + +lock: + lruvec = lock_page_lruvec_irqsave(page, pgdat); + return lruvec; +} + #ifdef CONFIG_CGROUP_WRITEBACK struct wb_domain *mem_cgroup_wb_domain(struct bdi_writeback *wb); diff --git a/mm/mlock.c b/mm/mlock.c index b509b80b8513..8b3a97b62c0a 100644 --- a/mm/mlock.c +++ b/mm/mlock.c @@ -290,6 +290,7 @@ static void __munlock_pagevec(struct pagevec *pvec, struct zone *zone) { int i; int nr = pagevec_count(pvec); + int delta_munlocked = -nr; struct pagevec pvec_putback; struct lruvec *lruvec = NULL; int pgrescued = 0; @@ -300,20 +301,19 @@ static void __munlock_pagevec(struct pagevec *pvec, struct zone *zone) for (i = 0; i < nr; i++) { struct page *page = pvec->pages[i]; - lruvec = lock_page_lruvec_irq(page, page_pgdat(page)); + lruvec = relock_page_lruvec_irq(page, lruvec); if (TestClearPageMlocked(page)) { /* * We already have pin from follow_page_mask() * so we can spare the get_page() here. */ - if (__munlock_isolate_lru_page(page, lruvec, false)) { - __mod_zone_page_state(zone, NR_MLOCK, -1); - spin_unlock_irq(&lruvec->lru_lock); + if (__munlock_isolate_lru_page(page, lruvec, false)) continue; - } else + else __munlock_isolation_failed(page); - } + } else + delta_munlocked++; /* * We won't be munlocking this page in the next phase @@ -323,8 +323,10 @@ static void __munlock_pagevec(struct pagevec *pvec, struct zone *zone) */ pagevec_add(&pvec_putback, pvec->pages[i]); pvec->pages[i] = NULL; - spin_unlock_irq(&lruvec->lru_lock); } + __mod_zone_page_state(zone, NR_MLOCK, delta_munlocked); + if (lruvec) + spin_unlock_irq(&lruvec->lru_lock); /* Now we can release pins of pages that we are not munlocking */ pagevec_release(&pvec_putback); diff --git a/mm/swap.c b/mm/swap.c index 60f04cb2b49e..a26e19738e96 100644 --- a/mm/swap.c +++ b/mm/swap.c @@ -195,11 +195,12 @@ static void pagevec_lru_move_fn(struct pagevec *pvec, for (i = 0; i < pagevec_count(pvec); i++) { struct page *page = pvec->pages[i]; - lruvec = lock_page_lruvec_irqsave(page, page_pgdat(page)); + lruvec = relock_page_lruvec_irqsave(page, lruvec); (*move_fn)(page, lruvec, arg); - spin_unlock_irqrestore(&lruvec->lru_lock, lruvec->irqflags); } + if (lruvec) + spin_unlock_irqrestore(&lruvec->lru_lock, lruvec->irqflags); release_pages(pvec->pages, pvec->nr); pagevec_reinit(pvec); @@ -821,15 +822,12 @@ void release_pages(struct page **pages, int nr) } if (PageLRU(page)) { - struct lruvec *new_lruvec = mem_cgroup_page_lruvec(page, page_pgdat(page)); + struct lruvec *pre_lruvec = lruvec; - if (new_lruvec != lruvec) { - if (lruvec) - spin_unlock_irqrestore(&lruvec->lru_lock, lruvec->irqflags); + lruvec = relock_page_lruvec_irqsave(page, lruvec); + if (pre_lruvec != lruvec) lock_batch = 0; - lruvec = lock_page_lruvec_irqsave(page, page_pgdat(page)); - } VM_BUG_ON_PAGE(!PageLRU(page), page); __ClearPageLRU(page); del_page_from_lru_list(page, lruvec, page_off_lru(page)); diff --git a/mm/vmscan.c b/mm/vmscan.c index 3cdf343e7a27..ba57c55a6a41 100644 --- a/mm/vmscan.c +++ b/mm/vmscan.c @@ -1825,22 +1825,25 @@ static int too_many_isolated(struct pglist_data *pgdat, int file, * Returns the number of pages moved to the given lruvec. */ -static unsigned noinline_for_stack move_pages_to_lru(struct lruvec *lruvec, +static unsigned noinline_for_stack move_pages_to_lru(struct lruvec *llvec, struct list_head *list) { int nr_pages, nr_moved = 0; LIST_HEAD(pages_to_free); struct page *page; enum lru_list lru; + struct lruvec *lruvec = llvec; while (!list_empty(list)) { page = lru_to_page(list); + lruvec = relock_page_lruvec_irq(page, lruvec); + VM_BUG_ON_PAGE(PageLRU(page), page); if (unlikely(!page_evictable(page))) { list_del(&page->lru); spin_unlock_irq(&lruvec->lru_lock); + lruvec = NULL; putback_lru_page(page); - spin_lock_irq(&lruvec->lru_lock); continue; } @@ -1858,8 +1861,8 @@ static unsigned noinline_for_stack move_pages_to_lru(struct lruvec *lruvec, if (unlikely(PageCompound(page))) { spin_unlock_irq(&lruvec->lru_lock); + lruvec = NULL; (*get_compound_page_dtor(page))(page); - spin_lock_irq(&lruvec->lru_lock); } else list_add(&page->lru, &pages_to_free); } else { @@ -1867,6 +1870,11 @@ static unsigned noinline_for_stack move_pages_to_lru(struct lruvec *lruvec, } } + if (lruvec != llvec) { + if (lruvec) + spin_unlock_irq(&lruvec->lru_lock); + spin_lock_irq(&llvec->lru_lock); + } /* * To save our caller's stack, now use input list for pages to free. */ @@ -4289,18 +4297,10 @@ void check_move_unevictable_pages(struct pagevec *pvec) for (i = 0; i < pvec->nr; i++) { struct page *page = pvec->pages[i]; - struct pglist_data *pgdat = page_pgdat(page); - struct lruvec *new_lruvec = mem_cgroup_page_lruvec(page, pgdat); - pgscanned++; - if (lruvec != new_lruvec) { - if (lruvec) - spin_unlock_irq(&lruvec->lru_lock); - lruvec = new_lruvec; - spin_lock_irq(&lruvec->lru_lock); - } + lruvec = relock_page_lruvec_irq(page, lruvec); if (!PageLRU(page) || !PageUnevictable(page)) continue;