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=-12.8 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_CR_TRAILER,INCLUDES_PATCH, MAILING_LIST_MULTI,PDS_BAD_THREAD_QP_64,SPF_HELO_NONE,SPF_PASS,URIBL_BLOCKED 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 20353C433DB for ; Tue, 2 Mar 2021 07:05:43 +0000 (UTC) Received: from kanga.kvack.org (kanga.kvack.org [205.233.56.17]) by mail.kernel.org (Postfix) with ESMTP id 13A2964DE4 for ; Tue, 2 Mar 2021 07:05:42 +0000 (UTC) DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 13A2964DE4 Authentication-Results: mail.kernel.org; dmarc=none (p=none dis=none) header.from=huawei.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=owner-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix) id 28F028D00E3; Tue, 2 Mar 2021 02:05:41 -0500 (EST) Received: by kanga.kvack.org (Postfix, from userid 40) id 23EB48D0063; Tue, 2 Mar 2021 02:05:41 -0500 (EST) X-Delivered-To: int-list-linux-mm@kvack.org Received: by kanga.kvack.org (Postfix, from userid 63042) id 12E378D00E3; Tue, 2 Mar 2021 02:05:41 -0500 (EST) X-Delivered-To: linux-mm@kvack.org Received: from forelay.hostedemail.com (smtprelay0208.hostedemail.com [216.40.44.208]) by kanga.kvack.org (Postfix) with ESMTP id F1F218D0063 for ; Tue, 2 Mar 2021 02:05:40 -0500 (EST) Received: from smtpin23.hostedemail.com (10.5.19.251.rfc1918.com [10.5.19.251]) by forelay02.hostedemail.com (Postfix) with ESMTP id B6F4A75B3 for ; Tue, 2 Mar 2021 07:05:40 +0000 (UTC) X-FDA: 77874048840.23.26398D9 Received: from szxga01-in.huawei.com (szxga01-in.huawei.com [45.249.212.187]) by imf03.hostedemail.com (Postfix) with ESMTP id AAF11C0007C0 for ; Tue, 2 Mar 2021 07:05:38 +0000 (UTC) Received: from DGGEMM401-HUB.china.huawei.com (unknown [172.30.72.53]) by szxga01-in.huawei.com (SkyGuard) with ESMTP id 4DqSm42pyGzYD8V; Tue, 2 Mar 2021 15:04:04 +0800 (CST) Received: from dggema774-chm.china.huawei.com (10.1.198.216) by DGGEMM401-HUB.china.huawei.com (10.3.20.209) with Microsoft SMTP Server (TLS) id 14.3.498.0; Tue, 2 Mar 2021 15:05:35 +0800 Received: from dggeme755-chm.china.huawei.com (10.3.19.101) by dggema774-chm.china.huawei.com (10.1.198.216) with Microsoft SMTP Server (version=TLS1_2, cipher=TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA256_P256) id 15.1.2106.2; Tue, 2 Mar 2021 15:05:35 +0800 Received: from dggeme755-chm.china.huawei.com ([10.7.64.71]) by dggeme755-chm.china.huawei.com ([10.7.64.71]) with mapi id 15.01.2106.006; Tue, 2 Mar 2021 15:05:35 +0800 From: "Zhouguanghui (OS Kernel)" To: Zi Yan CC: "linux-kernel@vger.kernel.org" , "linux-mm@kvack.org" , "akpm@linux-foundation.org" , "npiggin@suse.de" , "Wangkefeng (OS Kernel Lab)" , "Guohanjun (Hanjun Guo)" , Dingtianhong , Chenweilong , "Xiangrui (Euler)" , Johannes Weiner , Michal Hocko , Vladimir Davydov Subject: Re: [PATCH] mm/memcg: set memcg when split pages Thread-Topic: [PATCH] mm/memcg: set memcg when split pages Thread-Index: AQHXDwe4PfGiAC+zfUG9IaYiC6n5KA== Date: Tue, 2 Mar 2021 07:05:35 +0000 Message-ID: References: <20210302013451.118701-1-zhouguanghui1@huawei.com> <733264DE-1845-4615-8573-481E63895397@nvidia.com> Accept-Language: en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [10.174.178.106] Content-Type: text/plain; charset="iso-2022-jp" Content-Transfer-Encoding: quoted-printable MIME-Version: 1.0 X-CFilter-Loop: Reflected X-Rspamd-Server: rspam04 X-Rspamd-Queue-Id: AAF11C0007C0 X-Stat-Signature: y5ncj5qjjdiqf4fzcaira1ao1c6wrdmo Received-SPF: none (huawei.com>: No applicable sender policy available) receiver=imf03; identity=mailfrom; envelope-from=""; helo=szxga01-in.huawei.com; client-ip=45.249.212.187 X-HE-DKIM-Result: none/none X-HE-Tag: 1614668738-297830 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: =1B$B:_=1B(B 2021/3/2 10:00, Zi Yan =1B$B On 1 Mar 2021, at 20:34, Zhou Guanghui wrote:=0A= > =0A= >> When split page, the memory cgroup info recorded in first page is=0A= >> not copied to tail pages. In this case, when the tail pages are=0A= >> freed, the uncharge operation is not performed. As a result, the=0A= >> usage of this memcg keeps increasing, and the OOM may occur.=0A= >>=0A= >> So, the copying of first page's memory cgroup info to tail pages=0A= >> is needed when split page.=0A= >>=0A= >> Signed-off-by: Zhou Guanghui =0A= >> ---=0A= >> include/linux/memcontrol.h | 10 ++++++++++=0A= >> mm/page_alloc.c | 4 +++-=0A= >> 2 files changed, 13 insertions(+), 1 deletion(-)=0A= >>=0A= >> diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h=0A= >> index e6dc793d587d..c7e2b4421dc1 100644=0A= >> --- a/include/linux/memcontrol.h=0A= >> +++ b/include/linux/memcontrol.h=0A= >> @@ -867,6 +867,12 @@ void mem_cgroup_print_oom_group(struct mem_cgroup *= memcg);=0A= >> extern bool cgroup_memory_noswap;=0A= >> #endif=0A= >>=0A= >> +static inline void copy_page_memcg(struct page *dst, struct page *src)= =0A= >> +{=0A= >> + if (src->memcg_data)=0A= >> + dst->memcg_data =3D src->memcg_data;=0A= >> +}=0A= >> +=0A= >> struct mem_cgroup *lock_page_memcg(struct page *page);=0A= >> void __unlock_page_memcg(struct mem_cgroup *memcg);=0A= >> void unlock_page_memcg(struct page *page);=0A= >> @@ -1291,6 +1297,10 @@ mem_cgroup_print_oom_meminfo(struct mem_cgroup *m= emcg)=0A= >> {=0A= >> }=0A= >>=0A= >> +static inline void copy_page_memcg(struct page *dst, struct page *src)= =0A= >> +{=0A= >> +}=0A= >> +=0A= >> static inline struct mem_cgroup *lock_page_memcg(struct page *page)=0A= >> {=0A= >> return NULL;=0A= >> diff --git a/mm/page_alloc.c b/mm/page_alloc.c=0A= >> index 3e4b29ee2b1e..ee0a63dc1c9b 100644=0A= >> --- a/mm/page_alloc.c=0A= >> +++ b/mm/page_alloc.c=0A= >> @@ -3307,8 +3307,10 @@ void split_page(struct page *page, unsigned int o= rder)=0A= >> VM_BUG_ON_PAGE(PageCompound(page), page);=0A= >> VM_BUG_ON_PAGE(!page_count(page), page);=0A= >>=0A= >> - for (i =3D 1; i < (1 << order); i++)=0A= >> + for (i =3D 1; i < (1 << order); i++) {=0A= >> set_page_refcounted(page + i);=0A= >> + copy_page_memcg(page + i, page);=0A= >> + }=0A= >> split_page_owner(page, 1 << order);=0A= >> }=0A= >> EXPORT_SYMBOL_GPL(split_page);=0A= >> -- =0A= >> 2.25.0=0A= > =0A= > +memcg maintainers=0A= > =0A= > split_page() is used for non-compound higher-order pages. I am not sure= =0A= > if there is any such pages monitored by memcg. Please let me know=0A= > if I miss anything.=0A= =0A= Thank you for taking time for this.=0A= =0A= This should be put in kmemcg, and I'll modify it.=0A= =0A= When the kmemcg is enabled and _GFP_ACCOUNT is set, the charged and =0A= uncharged sizes do not match when alloc/free_pages_exact method is used =0A= to apply for or free memory with exact size. This is because memcg data =0A= of the tail page is not set during the split page.=0A= =0A=