From nobody Tue Dec 16 19:57:19 2025 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by smtp.lore.kernel.org (Postfix) with ESMTP id 5A0E6CDB482 for ; Sat, 14 Oct 2023 12:55:23 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S233193AbjJNMzW (ORCPT ); Sat, 14 Oct 2023 08:55:22 -0400 Received: from lindbergh.monkeyblade.net ([23.128.96.19]:51450 "EHLO lindbergh.monkeyblade.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S232957AbjJNMzV (ORCPT ); Sat, 14 Oct 2023 08:55:21 -0400 Received: from out30-100.freemail.mail.aliyun.com (out30-100.freemail.mail.aliyun.com [115.124.30.100]) by lindbergh.monkeyblade.net (Postfix) with ESMTPS id 21F8EB7; Sat, 14 Oct 2023 05:55:17 -0700 (PDT) X-Alimail-AntiSpam: AC=PASS;BC=-1|-1;BR=01201311R671e4;CH=green;DM=||false|;DS=||;FP=0|-1|-1|-1|0|-1|-1|-1;HT=ay29a033018045192;MF=jefflexu@linux.alibaba.com;NM=1;PH=DS;RN=12;SR=0;TI=SMTPD_---0Vu5Ns.c_1697288111; Received: from localhost(mailfrom:jefflexu@linux.alibaba.com fp:SMTPD_---0Vu5Ns.c_1697288111) by smtp.aliyun-inc.com; Sat, 14 Oct 2023 20:55:12 +0800 From: Jingbo Xu To: tj@kernel.org, guro@fb.com, jack@suse.cz Cc: lizefan.x@bytedance.com, hannes@cmpxchg.org, cgroups@vger.kernel.org, linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, viro@zeniv.linux.org.uk, brauner@kernel.org, willy@infradead.org, joseph.qi@linux.alibaba.com Subject: [PATCH v3] writeback, cgroup: switch inodes with dirty timestamps to release dying cgwbs Date: Sat, 14 Oct 2023 20:55:11 +0800 Message-Id: <20231014125511.102978-1-jefflexu@linux.alibaba.com> X-Mailer: git-send-email 2.19.1.6.gb485710b MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Precedence: bulk List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Type: text/plain; charset="utf-8" The cgwb cleanup routine will try to release the dying cgwb by switching the attached inodes. It fetches the attached inodes from wb->b_attached list, omitting the fact that inodes only with dirty timestamps reside in wb->b_dirty_time list, which is the case when lazytime is enabled. This causes enormous zombie memory cgroup when lazytime is enabled, as inodes with dirty timestamps can not be switched to a live cgwb for a long time. It is reasonable not to switch cgwb for inodes with dirty data, as otherwise it may break the bandwidth restrictions. However since the writeback of inode metadata is not accounted for, let's also switch inodes with dirty timestamps to avoid zombie memory and block cgroups when laztytime is enabled. Fixes: c22d70a162d3 ("writeback, cgroup: release dying cgwbs by switching a= ttached inodes") Reviewed-by: Jan Kara Signed-off-by: Jingbo Xu Acked-by: Tejun Heo --- v3: fix spelling of "Fixes"; add "Reviewed-by" tag from Jan Kara (Thanks!) v1: https://lore.kernel.org/all/20231011084228.77615-1-jefflexu@linux.aliba= ba.com/ v2: https://lore.kernel.org/all/20231013055208.15457-1-jefflexu@linux.aliba= ba.com/ --- fs/fs-writeback.c | 41 +++++++++++++++++++++++++++++------------ 1 file changed, 29 insertions(+), 12 deletions(-) diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index c1af01b2c42d..1767493dffda 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -613,6 +613,24 @@ static void inode_switch_wbs(struct inode *inode, int = new_wb_id) kfree(isw); } =20 +static bool isw_prepare_wbs_switch(struct inode_switch_wbs_context *isw, + struct list_head *list, int *nr) +{ + struct inode *inode; + + list_for_each_entry(inode, list, i_io_list) { + if (!inode_prepare_wbs_switch(inode, isw->new_wb)) + continue; + + isw->inodes[*nr] =3D inode; + (*nr)++; + + if (*nr >=3D WB_MAX_INODES_PER_ISW - 1) + return true; + } + return false; +} + /** * cleanup_offline_cgwb - detach associated inodes * @wb: target wb @@ -625,7 +643,6 @@ bool cleanup_offline_cgwb(struct bdi_writeback *wb) { struct cgroup_subsys_state *memcg_css; struct inode_switch_wbs_context *isw; - struct inode *inode; int nr; bool restart =3D false; =20 @@ -647,17 +664,17 @@ bool cleanup_offline_cgwb(struct bdi_writeback *wb) =20 nr =3D 0; spin_lock(&wb->list_lock); - list_for_each_entry(inode, &wb->b_attached, i_io_list) { - if (!inode_prepare_wbs_switch(inode, isw->new_wb)) - continue; - - isw->inodes[nr++] =3D inode; - - if (nr >=3D WB_MAX_INODES_PER_ISW - 1) { - restart =3D true; - break; - } - } + /* + * In addition to the inodes that have completed writeback, also switch + * cgwbs for those inodes only with dirty timestamps. Otherwise, those + * inodes won't be written back for a long time when lazytime is + * enabled, and thus pinning the dying cgwbs. It won't break the + * bandwidth restrictions, as writeback of inode metadata is not + * accounted for. + */ + restart =3D isw_prepare_wbs_switch(isw, &wb->b_attached, &nr); + if (!restart) + restart =3D isw_prepare_wbs_switch(isw, &wb->b_dirty_time, &nr); spin_unlock(&wb->list_lock); =20 /* no attached inodes? bail out */ --=20 2.19.1.6.gb485710b