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=-15.7 required=3.0 tests=BAYES_00,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU,HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER, INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED autolearn=unavailable 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 16EBFC433E0 for ; Wed, 24 Feb 2021 20:05:45 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id C747D64F19 for ; Wed, 24 Feb 2021 20:05:44 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S234738AbhBXUFm (ORCPT ); Wed, 24 Feb 2021 15:05:42 -0500 Received: from mail.kernel.org ([198.145.29.99]:55458 "EHLO mail.kernel.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S235286AbhBXUEW (ORCPT ); Wed, 24 Feb 2021 15:04:22 -0500 Received: by mail.kernel.org (Postfix) with ESMTPSA id 3053264F26; Wed, 24 Feb 2021 20:03:36 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linux-foundation.org; s=korg; t=1614197017; bh=AFxRM4kCInF5En0Aet2ikFxdWfUXoU3bNgDO7Rr+GAc=; h=Date:From:To:Subject:In-Reply-To:From; b=QrAgp6k5EiRQgL2JrGZm1gueidEhTexULU2KUGXc++oQ5DxbNkxXGaLiQyrY/R28I /C9Wz6Dw/cUAHX5aXuW9lSIWtp43t/XwjoBfI4mRuwYL0nFll+SALxMgs1pVR0NtnW 9+WPIcZZCjg/4WId8Bf4WzU1WE0Pxszu+JbOlcI4= Date: Wed, 24 Feb 2021 12:03:35 -0800 From: Andrew Morton To: adobriyan@gmail.com, akpm@linux-foundation.org, feng.tang@intel.com, gregkh@linuxfoundation.org, guro@fb.com, hannes@cmpxchg.org, hughd@google.com, iamjoonsoo.kim@lge.com, linux-mm@kvack.org, mhocko@suse.com, mm-commits@vger.kernel.org, neilb@suse.de, pankaj.gupta@cloud.ionos.com, rafael@kernel.org, rdunlap@infradead.org, samitolvanen@google.com, shakeelb@google.com, songmuchun@bytedance.com, torvalds@linux-foundation.org, vdavydov.dev@gmail.com Subject: [patch 061/173] mm: memcontrol: convert NR_SHMEM_PMDMAPPED account to pages Message-ID: <20210224200335.i3uyijHmX%akpm@linux-foundation.org> In-Reply-To: <20210224115824.1e289a6895087f10c41dd8d6@linux-foundation.org> User-Agent: s-nail v14.8.16 Precedence: bulk Reply-To: linux-kernel@vger.kernel.org List-ID: X-Mailing-List: mm-commits@vger.kernel.org From: Muchun Song Subject: mm: memcontrol: convert NR_SHMEM_PMDMAPPED account to pages Currently we use struct per_cpu_nodestat to cache the vmstat counters, which leads to inaccurate statistics especially THP vmstat counters. In the systems with hundreds of processors it can be GBs of memory. For example, for a 96 CPUs system, the threshold is the maximum number of 125. And the per cpu counters can cache 23.4375 GB in total. The THP page is already a form of batched addition (it will add 512 worth of memory in one go) so skipping the batching seems like sensible. Although every THP stats update overflows the per-cpu counter, resorting to atomic global updates. But it can make the statistics more accuracy for the THP vmstat counters. So we convert the NR_SHMEM_PMDMAPPED account to pages. This patch is consistent with 8f182270dfec ("mm/swap.c: flush lru pvecs on compound page arrival"). Doing this also can make the unit of vmstat counters more unified. Finally, the unit of the vmstat counters are pages, kB and bytes. The B/KB suffix can tell us that the unit is bytes or kB. The rest which is without suffix are pages. Link: https://lkml.kernel.org/r/20201228164110.2838-6-songmuchun@bytedance.com Signed-off-by: Muchun Song Cc: Alexey Dobriyan Cc: Feng Tang Cc: Greg Kroah-Hartman Cc: Hugh Dickins Cc: Johannes Weiner Cc: Joonsoo Kim Cc: Michal Hocko Cc: NeilBrown Cc: Pankaj Gupta Cc: Rafael. J. Wysocki Cc: Randy Dunlap Cc: Roman Gushchin Cc: Sami Tolvanen Cc: Shakeel Butt Cc: Vladimir Davydov Signed-off-by: Andrew Morton --- drivers/base/node.c | 3 +-- fs/proc/meminfo.c | 2 +- include/linux/mmzone.h | 3 ++- mm/page_alloc.c | 3 +-- mm/rmap.c | 14 ++++++++++---- 5 files changed, 15 insertions(+), 10 deletions(-) --- a/drivers/base/node.c~mm-memcontrol-convert-nr_shmem_pmdmapped-account-to-pages +++ a/drivers/base/node.c @@ -463,8 +463,7 @@ static ssize_t node_read_meminfo(struct , nid, K(node_page_state(pgdat, NR_ANON_THPS)), nid, K(node_page_state(pgdat, NR_SHMEM_THPS)), - nid, K(node_page_state(pgdat, NR_SHMEM_PMDMAPPED) * - HPAGE_PMD_NR), + nid, K(node_page_state(pgdat, NR_SHMEM_PMDMAPPED)), nid, K(node_page_state(pgdat, NR_FILE_THPS)), nid, K(node_page_state(pgdat, NR_FILE_PMDMAPPED) * HPAGE_PMD_NR) --- a/fs/proc/meminfo.c~mm-memcontrol-convert-nr_shmem_pmdmapped-account-to-pages +++ a/fs/proc/meminfo.c @@ -133,7 +133,7 @@ static int meminfo_proc_show(struct seq_ show_val_kb(m, "ShmemHugePages: ", global_node_page_state(NR_SHMEM_THPS)); show_val_kb(m, "ShmemPmdMapped: ", - global_node_page_state(NR_SHMEM_PMDMAPPED) * HPAGE_PMD_NR); + global_node_page_state(NR_SHMEM_PMDMAPPED)); show_val_kb(m, "FileHugePages: ", global_node_page_state(NR_FILE_THPS)); show_val_kb(m, "FilePmdMapped: ", --- a/include/linux/mmzone.h~mm-memcontrol-convert-nr_shmem_pmdmapped-account-to-pages +++ a/include/linux/mmzone.h @@ -221,7 +221,8 @@ static __always_inline bool vmstat_item_ return item == NR_ANON_THPS || item == NR_FILE_THPS || - item == NR_SHMEM_THPS; + item == NR_SHMEM_THPS || + item == NR_SHMEM_PMDMAPPED; } /* --- a/mm/page_alloc.c~mm-memcontrol-convert-nr_shmem_pmdmapped-account-to-pages +++ a/mm/page_alloc.c @@ -5585,8 +5585,7 @@ void show_free_areas(unsigned int filter K(node_page_state(pgdat, NR_SHMEM)), #ifdef CONFIG_TRANSPARENT_HUGEPAGE K(node_page_state(pgdat, NR_SHMEM_THPS)), - K(node_page_state(pgdat, NR_SHMEM_PMDMAPPED) - * HPAGE_PMD_NR), + K(node_page_state(pgdat, NR_SHMEM_PMDMAPPED)), K(node_page_state(pgdat, NR_ANON_THPS)), #endif K(node_page_state(pgdat, NR_WRITEBACK_TEMP)), --- a/mm/rmap.c~mm-memcontrol-convert-nr_shmem_pmdmapped-account-to-pages +++ a/mm/rmap.c @@ -1211,14 +1211,17 @@ void page_add_file_rmap(struct page *pag VM_BUG_ON_PAGE(compound && !PageTransHuge(page), page); lock_page_memcg(page); if (compound && PageTransHuge(page)) { - for (i = 0, nr = 0; i < thp_nr_pages(page); i++) { + int nr_pages = thp_nr_pages(page); + + for (i = 0, nr = 0; i < nr_pages; i++) { if (atomic_inc_and_test(&page[i]._mapcount)) nr++; } if (!atomic_inc_and_test(compound_mapcount_ptr(page))) goto out; if (PageSwapBacked(page)) - __inc_node_page_state(page, NR_SHMEM_PMDMAPPED); + __mod_lruvec_page_state(page, NR_SHMEM_PMDMAPPED, + nr_pages); else __inc_node_page_state(page, NR_FILE_PMDMAPPED); } else { @@ -1252,14 +1255,17 @@ static void page_remove_file_rmap(struct /* page still mapped by someone else? */ if (compound && PageTransHuge(page)) { - for (i = 0, nr = 0; i < thp_nr_pages(page); i++) { + int nr_pages = thp_nr_pages(page); + + for (i = 0, nr = 0; i < nr_pages; i++) { if (atomic_add_negative(-1, &page[i]._mapcount)) nr++; } if (!atomic_add_negative(-1, compound_mapcount_ptr(page))) return; if (PageSwapBacked(page)) - __dec_node_page_state(page, NR_SHMEM_PMDMAPPED); + __mod_lruvec_page_state(page, NR_SHMEM_PMDMAPPED, + -nr_pages); else __dec_node_page_state(page, NR_FILE_PMDMAPPED); } else { _