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=-3.8 required=3.0 tests=DKIMWL_WL_HIGH,DKIM_SIGNED, DKIM_VALID,HEADER_FROM_DIFFERENT_DOMAINS,MAILING_LIST_MULTI,SIGNED_OFF_BY, SPF_HELO_NONE,SPF_PASS autolearn=no 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 90616C432C0 for ; Sun, 1 Dec 2019 01:56:02 +0000 (UTC) Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.kernel.org (Postfix) with ESMTP id 388532084D for ; Sun, 1 Dec 2019 01:56:02 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (1024-bit key) header.d=kernel.org header.i=@kernel.org header.b="pUAS8l/E" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 388532084D Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=linux-foundation.org Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=owner-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix) id DBEFC6B033E; Sat, 30 Nov 2019 20:56:01 -0500 (EST) Received: by kanga.kvack.org (Postfix, from userid 40) id D47866B0340; Sat, 30 Nov 2019 20:56:01 -0500 (EST) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id C0EC86B0341; Sat, 30 Nov 2019 20:56:01 -0500 (EST) X-Delivered-To: linux-mm@kvack.org Received: from forelay.hostedemail.com (smtprelay0070.hostedemail.com [216.40.44.70]) by kanga.kvack.org (Postfix) with ESMTP id A17A26B033E for ; Sat, 30 Nov 2019 20:56:01 -0500 (EST) Received: from smtpin17.hostedemail.com (10.5.19.251.rfc1918.com [10.5.19.251]) by forelay04.hostedemail.com (Postfix) with SMTP id 54F232C0C for ; Sun, 1 Dec 2019 01:56:01 +0000 (UTC) X-FDA: 76214906922.17.sleep71_34847e8376f15 X-HE-Tag: sleep71_34847e8376f15 X-Filterd-Recvd-Size: 12952 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by imf50.hostedemail.com (Postfix) with ESMTP for ; Sun, 1 Dec 2019 01:56:00 +0000 (UTC) Received: from localhost.localdomain (c-73-231-172-41.hsd1.ca.comcast.net [73.231.172.41]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPSA id CBD9121774; Sun, 1 Dec 2019 01:55:59 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=default; t=1575165360; bh=nzlJCFn2Dx0dYQLiKaWl6qdKPfprNDO4KCbOY+W+xbg=; h=Date:From:To:Subject:From; b=pUAS8l/E73+z0oIVF83qm+BfH2tl1mbJRPjA51gkMR4Hs+QYnlheO2pGaw7oEMLV4 9/FdN5tz/aKLr/G805K02+/3xln7ddpc00n9c+Hek8QU5Pu/OKo8ZQmBN0VKuLtGg4 xNMKroxEqTcuxUgwPQevO8eFySqRrkdHbFMuA0/M= Date: Sat, 30 Nov 2019 17:55:59 -0800 From: akpm@linux-foundation.org To: akpm@linux-foundation.org, aryabinin@virtuozzo.com, hannes@cmpxchg.org, linux-mm@kvack.org, mhocko@suse.com, mm-commits@vger.kernel.org, riel@surriel.com, shakeelb@google.com, surenb@google.com, torvalds@linux-foundation.org Subject: [patch 113/158] mm: vmscan: detect file thrashing at the reclaim root Message-ID: <20191201015559.rp-zSepME%akpm@linux-foundation.org> User-Agent: s-nail v14.8.16 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: From: Johannes Weiner Subject: mm: vmscan: detect file thrashing at the reclaim root We use refault information to determine whether the cache workingset is stable or transitioning, and dynamically adjust the inactive:active file LRU ratio so as to maximize protection from one-off cache during stable periods, and minimize IO during transitions. With cgroups and their nested LRU lists, we currently don't do this correctly. While recursive cgroup reclaim establishes a relative LRU order among the pages of all involved cgroups, refaults only affect the local LRU order in the cgroup in which they are occuring. As a result, cache transitions can take longer in a cgrouped system as the active pages of sibling cgroups aren't challenged when they should be. [ Right now, this is somewhat theoretical, because the siblings, under continued regular reclaim pressure, should eventually run out of inactive pages - and since inactive:active *size* balancing is also done on a cgroup-local level, we will challenge the active pages eventually in most cases. But the next patch will move that relative size enforcement to the reclaim root as well, and then this patch here will be necessary to propagate refault pressure to siblings. ] This patch moves refault detection to the root of reclaim. Instead of remembering the cgroup owner of an evicted page, remember the cgroup that caused the reclaim to happen. When refaults later occur, they'll correctly influence the cross-cgroup LRU order that reclaim follows. I.e. if global reclaim kicked out pages in some subgroup A/B/C, the refault of those pages will challenge the global LRU order, and not just the local order down inside C. [hannes@cmpxchg.org: use page_memcg() instead of another lookup] Link: http://lkml.kernel.org/r/20191115160722.GA309754@cmpxchg.org Link: http://lkml.kernel.org/r/20191107205334.158354-3-hannes@cmpxchg.org Signed-off-by: Johannes Weiner Reviewed-by: Suren Baghdasaryan Cc: Andrey Ryabinin Cc: Michal Hocko Cc: Rik van Riel Cc: Shakeel Butt Signed-off-by: Andrew Morton --- include/linux/memcontrol.h | 5 ++ include/linux/swap.h | 2 - mm/vmscan.c | 32 +++++++--------- mm/workingset.c | 67 +++++++++++++++++++++++++++-------- 4 files changed, 73 insertions(+), 33 deletions(-) --- a/include/linux/memcontrol.h~mm-vmscan-detect-file-thrashing-at-the-reclaim-root +++ a/include/linux/memcontrol.h @@ -901,6 +901,11 @@ static inline struct lruvec *mem_cgroup_ return &pgdat->__lruvec; } +static inline struct mem_cgroup *parent_mem_cgroup(struct mem_cgroup *memcg) +{ + return NULL; +} + static inline bool mm_match_cgroup(struct mm_struct *mm, struct mem_cgroup *memcg) { --- a/include/linux/swap.h~mm-vmscan-detect-file-thrashing-at-the-reclaim-root +++ a/include/linux/swap.h @@ -307,7 +307,7 @@ struct vma_swap_readahead { }; /* linux/mm/workingset.c */ -void *workingset_eviction(struct page *page); +void *workingset_eviction(struct page *page, struct mem_cgroup *target_memcg); void workingset_refault(struct page *page, void *shadow); void workingset_activation(struct page *page); --- a/mm/vmscan.c~mm-vmscan-detect-file-thrashing-at-the-reclaim-root +++ a/mm/vmscan.c @@ -853,7 +853,7 @@ static pageout_t pageout(struct page *pa * gets returned with a refcount of 0. */ static int __remove_mapping(struct address_space *mapping, struct page *page, - bool reclaimed) + bool reclaimed, struct mem_cgroup *target_memcg) { unsigned long flags; int refcount; @@ -925,7 +925,7 @@ static int __remove_mapping(struct addre */ if (reclaimed && page_is_file_cache(page) && !mapping_exiting(mapping) && !dax_mapping(mapping)) - shadow = workingset_eviction(page); + shadow = workingset_eviction(page, target_memcg); __delete_from_page_cache(page, shadow); xa_unlock_irqrestore(&mapping->i_pages, flags); @@ -948,7 +948,7 @@ cannot_free: */ int remove_mapping(struct address_space *mapping, struct page *page) { - if (__remove_mapping(mapping, page, false)) { + if (__remove_mapping(mapping, page, false, NULL)) { /* * Unfreezing the refcount with 1 rather than 2 effectively * drops the pagecache ref for us without requiring another @@ -1426,7 +1426,8 @@ static unsigned long shrink_page_list(st count_vm_event(PGLAZYFREED); count_memcg_page_event(page, PGLAZYFREED); - } else if (!mapping || !__remove_mapping(mapping, page, true)) + } else if (!mapping || !__remove_mapping(mapping, page, true, + sc->target_mem_cgroup)) goto keep_locked; unlock_page(page); @@ -2189,6 +2190,7 @@ static bool inactive_list_is_low(struct enum lru_list inactive_lru = file * LRU_FILE; unsigned long inactive, active; unsigned long inactive_ratio; + struct lruvec *target_lruvec; unsigned long refaults; unsigned long gb; @@ -2200,8 +2202,9 @@ static bool inactive_list_is_low(struct * is being established. Disable active list protection to get * rid of the stale workingset quickly. */ - refaults = lruvec_page_state_local(lruvec, WORKINGSET_ACTIVATE); - if (file && lruvec->refaults != refaults) { + target_lruvec = mem_cgroup_lruvec(sc->target_mem_cgroup, pgdat); + refaults = lruvec_page_state(target_lruvec, WORKINGSET_ACTIVATE); + if (file && target_lruvec->refaults != refaults) { inactive_ratio = 0; } else { gb = (inactive + active) >> (30 - PAGE_SHIFT); @@ -2973,19 +2976,14 @@ static void shrink_zones(struct zonelist sc->gfp_mask = orig_mask; } -static void snapshot_refaults(struct mem_cgroup *root_memcg, pg_data_t *pgdat) +static void snapshot_refaults(struct mem_cgroup *target_memcg, pg_data_t *pgdat) { - struct mem_cgroup *memcg; - - memcg = mem_cgroup_iter(root_memcg, NULL, NULL); - do { - unsigned long refaults; - struct lruvec *lruvec; + struct lruvec *target_lruvec; + unsigned long refaults; - lruvec = mem_cgroup_lruvec(memcg, pgdat); - refaults = lruvec_page_state_local(lruvec, WORKINGSET_ACTIVATE); - lruvec->refaults = refaults; - } while ((memcg = mem_cgroup_iter(root_memcg, memcg, NULL))); + target_lruvec = mem_cgroup_lruvec(target_memcg, pgdat); + refaults = lruvec_page_state(target_lruvec, WORKINGSET_ACTIVATE); + target_lruvec->refaults = refaults; } /* --- a/mm/workingset.c~mm-vmscan-detect-file-thrashing-at-the-reclaim-root +++ a/mm/workingset.c @@ -213,28 +213,53 @@ static void unpack_shadow(void *shadow, *workingsetp = workingset; } +static void advance_inactive_age(struct mem_cgroup *memcg, pg_data_t *pgdat) +{ + /* + * Reclaiming a cgroup means reclaiming all its children in a + * round-robin fashion. That means that each cgroup has an LRU + * order that is composed of the LRU orders of its child + * cgroups; and every page has an LRU position not just in the + * cgroup that owns it, but in all of that group's ancestors. + * + * So when the physical inactive list of a leaf cgroup ages, + * the virtual inactive lists of all its parents, including + * the root cgroup's, age as well. + */ + do { + struct lruvec *lruvec; + + lruvec = mem_cgroup_lruvec(memcg, pgdat); + atomic_long_inc(&lruvec->inactive_age); + } while (memcg && (memcg = parent_mem_cgroup(memcg))); +} + /** * workingset_eviction - note the eviction of a page from memory + * @target_memcg: the cgroup that is causing the reclaim * @page: the page being evicted * * Returns a shadow entry to be stored in @page->mapping->i_pages in place * of the evicted @page so that a later refault can be detected. */ -void *workingset_eviction(struct page *page) +void *workingset_eviction(struct page *page, struct mem_cgroup *target_memcg) { struct pglist_data *pgdat = page_pgdat(page); - struct mem_cgroup *memcg = page_memcg(page); - int memcgid = mem_cgroup_id(memcg); unsigned long eviction; struct lruvec *lruvec; + int memcgid; /* Page is fully exclusive and pins page->mem_cgroup */ VM_BUG_ON_PAGE(PageLRU(page), page); VM_BUG_ON_PAGE(page_count(page), page); VM_BUG_ON_PAGE(!PageLocked(page), page); - lruvec = mem_cgroup_lruvec(memcg, pgdat); - eviction = atomic_long_inc_return(&lruvec->inactive_age); + advance_inactive_age(page_memcg(page), pgdat); + + lruvec = mem_cgroup_lruvec(target_memcg, pgdat); + /* XXX: target_memcg can be NULL, go through lruvec */ + memcgid = mem_cgroup_id(lruvec_memcg(lruvec)); + eviction = atomic_long_read(&lruvec->inactive_age); return pack_shadow(memcgid, pgdat, eviction, PageWorkingset(page)); } @@ -244,10 +269,13 @@ void *workingset_eviction(struct page *p * @shadow: shadow entry of the evicted page * * Calculates and evaluates the refault distance of the previously - * evicted page in the context of the node it was allocated in. + * evicted page in the context of the node and the memcg whose memory + * pressure caused the eviction. */ void workingset_refault(struct page *page, void *shadow) { + struct mem_cgroup *eviction_memcg; + struct lruvec *eviction_lruvec; unsigned long refault_distance; struct pglist_data *pgdat; unsigned long active_file; @@ -277,12 +305,12 @@ void workingset_refault(struct page *pag * would be better if the root_mem_cgroup existed in all * configurations instead. */ - memcg = mem_cgroup_from_id(memcgid); - if (!mem_cgroup_disabled() && !memcg) + eviction_memcg = mem_cgroup_from_id(memcgid); + if (!mem_cgroup_disabled() && !eviction_memcg) goto out; - lruvec = mem_cgroup_lruvec(memcg, pgdat); - refault = atomic_long_read(&lruvec->inactive_age); - active_file = lruvec_lru_size(lruvec, LRU_ACTIVE_FILE, MAX_NR_ZONES); + eviction_lruvec = mem_cgroup_lruvec(eviction_memcg, pgdat); + refault = atomic_long_read(&eviction_lruvec->inactive_age); + active_file = lruvec_page_state(eviction_lruvec, NR_ACTIVE_FILE); /* * Calculate the refault distance @@ -302,6 +330,17 @@ void workingset_refault(struct page *pag */ refault_distance = (refault - eviction) & EVICTION_MASK; + /* + * The activation decision for this page is made at the level + * where the eviction occurred, as that is where the LRU order + * during page reclaim is being determined. + * + * However, the cgroup that will own the page is the one that + * is actually experiencing the refault event. + */ + memcg = page_memcg(page); + lruvec = mem_cgroup_lruvec(memcg, pgdat); + inc_lruvec_state(lruvec, WORKINGSET_REFAULT); /* @@ -313,7 +352,7 @@ void workingset_refault(struct page *pag goto out; SetPageActive(page); - atomic_long_inc(&lruvec->inactive_age); + advance_inactive_age(memcg, pgdat); inc_lruvec_state(lruvec, WORKINGSET_ACTIVATE); /* Page was active prior to eviction */ @@ -332,7 +371,6 @@ out: void workingset_activation(struct page *page) { struct mem_cgroup *memcg; - struct lruvec *lruvec; rcu_read_lock(); /* @@ -345,8 +383,7 @@ void workingset_activation(struct page * memcg = page_memcg_rcu(page); if (!mem_cgroup_disabled() && !memcg) goto out; - lruvec = mem_cgroup_lruvec(memcg, page_pgdat(page)); - atomic_long_inc(&lruvec->inactive_age); + advance_inactive_age(memcg, page_pgdat(page)); out: rcu_read_unlock(); } _