From nobody Mon May 25 06:40:17 2026 Received: from out30-131.freemail.mail.aliyun.com (out30-131.freemail.mail.aliyun.com [115.124.30.131]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id CBFF83BE17F; Sun, 17 May 2026 14:22:24 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=115.124.30.131 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1779027757; cv=none; b=Q6BxTQFA64xeynhNdoIOclwY4ocKYgBM9wTZAKjK2s15SFDuflLxZJyxI1HJ9oAse5RuPakGZSVQEbPX1ir2QMEG46SEfJHYUdOpPJHGqyi7XBAnINhpI5Iq8GADQt5tVWy8e6s1ecuxzk6NZc7qDKyTsBMQK6s4aM3A1dQHG4w= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1779027757; c=relaxed/simple; bh=iVqgbsmoUShKzc8sehabHkz/7kiOXcuiLF9YU7dEtZU=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=s7+dVxnsKXtUmxCOlszY3zmtTMc2E7IwbTAJ/VjB9zOebZPc3Pmj7LwdBPAtFfWf1j7O1MZ7Z04kuU0yfFGl+RV1nQAXyluQDRSr3TsK79PPg1zb+dgZRVotsD54V8DeKhNsUK0v8xx8mUmrT87NeBtWJd9ZENNMsY9i5fX/+ZM= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.alibaba.com; spf=pass smtp.mailfrom=linux.alibaba.com; dkim=pass (1024-bit key) header.d=linux.alibaba.com header.i=@linux.alibaba.com header.b=HoR/2PVr; arc=none smtp.client-ip=115.124.30.131 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.alibaba.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=linux.alibaba.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linux.alibaba.com header.i=@linux.alibaba.com header.b="HoR/2PVr" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linux.alibaba.com; s=default; t=1779027732; h=From:To:Subject:Date:Message-ID:MIME-Version; bh=pEeH7zHkW8IEVUYlNw8sKefpVLWTLyEFqi5D7xTTfZ0=; b=HoR/2PVrdJE7Ku1lN6WOTGMpKWhHTKWqhVAT378NGf/d4VB/K/4p4N+7pJp4Ax+S4wNfZHqR4ovIc0yo0+vidrFMo/o1iDpcAFaYlpCDHToLRyAnIWG55zF8D882rVTa47lhUVgXfPcluOSfb5Y8I8haRstd6tJZxurj12nKP2U= X-Alimail-AntiSpam: AC=PASS;BC=-1|-1;BR=01201311R111e4;CH=green;DM=||false|;DS=||;FP=0|-1|-1|-1|0|-1|-1|-1;HT=maildocker-contentspam033045133197;MF=libaokun@linux.alibaba.com;NM=1;PH=DS;RN=8;SR=0;TI=SMTPD_---0X31pQEL_1779027731; Received: from x31h02109.sqa.na131.tbsite.net(mailfrom:libaokun@linux.alibaba.com fp:SMTPD_---0X31pQEL_1779027731 cluster:ay36) by smtp.aliyun-inc.com; Sun, 17 May 2026 22:22:12 +0800 From: Baokun Li To: linux-fsdevel@vger.kernel.org Cc: viro@zeniv.linux.org.uk, brauner@kernel.org, jack@suse.cz, tj@kernel.org, linux-kernel@vger.kernel.org Subject: [PATCH v2 1/3] writeback: fix race between cgroup_writeback_umount() and inode_switch_wbs() Date: Sun, 17 May 2026 22:21:30 +0800 Message-ID: <20260517142147.3354909-2-libaokun@linux.alibaba.com> X-Mailer: git-send-email 2.43.7 In-Reply-To: <20260517142147.3354909-1-libaokun@linux.alibaba.com> References: <20260517142147.3354909-1-libaokun@linux.alibaba.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" When a container exits, the following BUG_ON() is occasionally triggered: =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D VFS: Busy inodes after unmount of sdb (ext4) ------------[ cut here ]------------ kernel BUG at fs/super.c:695! CPU: 3 PID: 6 Comm: containerd-shim Tainted: G OE K 6.6 #1 pstate: 63400009 (nZCv daif +PAN -UAO +TCO +DIT -SSBS BTYPE=3D--) pc : generic_shutdown_super+0xf0/0x100 lr : generic_shutdown_super+0xf0/0x100 Call trace: generic_shutdown_super+0xf0/0x100 kill_block_super+0x20/0x48 ext4_kill_sb+0x28/0x60 deactivate_locked_super+0x54/0x130 deactivate_super+0x84/0xa0 cleanup_mnt+0xa4/0x140 __cleanup_mnt+0x18/0x28 task_work_run+0x78/0xe0 do_notify_resume+0x204/0x240 =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D The root cause is a race between cgroup_writeback_umount() and inode_switch_wbs()/cleanup_offline_cgwb(). There is a window between inode_prepare_wbs_switch() returning true and the subsequent wb_queue_isw() call. Following is the process that triggers the issue: CPU A (umount) | CPU B (writeback) ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ inode_switch_wbs/cleanup_offline_cgwb atomic_inc(&isw_nr_in_flight) inode_prepare_wbs_switch -> passes SB_ACTIVE check __iget(inode) generic_shutdown_super sb->s_flags &=3D ~SB_ACTIVE cgroup_writeback_umount(sb) smp_mb() atomic_read(&isw_nr_in_flight) rcu_barrier() -> no pending RCU callbacks flush_workqueue(isw_wq) -> nothing queued, returns evict_inodes(sb) -> Inode skipped as isw still holds a ref. sop->put_super(sb) /* destroys percpu counters */ -> VFS: Busy inodes after unmount! wb_queue_isw() queue_work(isw_wq, ...) /* later in work function */ inode_switch_wbs_work_fn process_inode_switch_wbs iput() -> evict percpu_counter_dec() // UAF! Fix this by extending the RCU read-side critical section in inode_switch_wbs() and cleanup_offline_cgwb() to cover from inode_prepare_wbs_switch() through wb_queue_isw(). Since there is no sleep in this window, rcu_read_lock() can be used. Then add a synchronize_rcu() in cgroup_writeback_umount() before the existing rcu_barrier(), so that all in-flight switchers that have passed the SB_ACTIVE check have completed queue_work() before flush_workqueue() is called. Fixes: a1a0e23e4903 ("writeback: flush inode cgroup wb switches instead of = pinning super_block") Suggested-by: Jan Kara Link: https://lore.kernel.org/all/mxnjq2l6guusfchvauxr3v7c4bwjasybxlleqbbh4= efloeqspz@iqylk76ohufz Signed-off-by: Baokun Li Reviewed-by: Jan Kara --- fs/fs-writeback.c | 27 ++++++++++++++++++++++++++- 1 file changed, 26 insertions(+), 1 deletion(-) diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index a65694cbfe68..1f95ddcee363 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -665,7 +665,6 @@ static void inode_switch_wbs(struct inode *inode, int n= ew_wb_id) memcg_css =3D css_from_id(new_wb_id, &memory_cgrp_subsys); if (memcg_css && !css_tryget(memcg_css)) memcg_css =3D NULL; - rcu_read_unlock(); if (!memcg_css) goto out_free; =20 @@ -680,10 +679,18 @@ static void inode_switch_wbs(struct inode *inode, int= new_wb_id) isw->inodes[0] =3D inode; =20 trace_inode_switch_wbs_queue(inode->i_wb, new_wb, 1); + /* + * Paired with synchronize_rcu() in cgroup_writeback_umount(). + * Holding rcu_read_lock across wb_queue_isw() ensures + * synchronize_rcu() cannot return until the work is queued, so + * the subsequent flush_workqueue() will wait for the switch. + */ wb_queue_isw(new_wb, isw); + rcu_read_unlock(); return; =20 out_free: + rcu_read_unlock(); atomic_dec(&isw_nr_in_flight); if (new_wb) wb_put(new_wb); @@ -741,6 +748,14 @@ bool cleanup_offline_cgwb(struct bdi_writeback *wb) new_wb =3D &wb->bdi->wb; /* wb_get() is noop for bdi's wb */ =20 nr =3D 0; + /* + * Paired with synchronize_rcu() in cgroup_writeback_umount(). + * Holding rcu_read_lock across the SB_ACTIVE check, the inode grab + * and wb_queue_isw() ensures synchronize_rcu() cannot return until + * the work is queued, so the subsequent flush_workqueue() will wait + * for the switch. + */ + rcu_read_lock(); spin_lock(&wb->list_lock); /* * In addition to the inodes that have completed writeback, also switch @@ -758,6 +773,7 @@ bool cleanup_offline_cgwb(struct bdi_writeback *wb) =20 /* no attached inodes? bail out */ if (nr =3D=3D 0) { + rcu_read_unlock(); atomic_dec(&isw_nr_in_flight); wb_put(new_wb); kfree(isw); @@ -766,6 +782,7 @@ bool cleanup_offline_cgwb(struct bdi_writeback *wb) =20 trace_inode_switch_wbs_queue(wb, new_wb, nr); wb_queue_isw(new_wb, isw); + rcu_read_unlock(); =20 return restart; } @@ -1221,6 +1238,14 @@ void cgroup_writeback_umount(struct super_block *sb) smp_mb(); =20 if (atomic_read(&isw_nr_in_flight)) { + /* + * Paired with rcu_read_lock() in inode_switch_wbs() and + * cleanup_offline_cgwb(). synchronize_rcu() waits for any + * in-flight switcher that already passed the SB_ACTIVE check + * to finish queueing its work, so flush_workqueue() below + * will then drain it. + */ + synchronize_rcu(); /* * Use rcu_barrier() to wait for all pending callbacks to * ensure that all in-flight wb switches are in the workqueue. --=20 2.43.7 From nobody Mon May 25 06:40:17 2026 Received: from out30-98.freemail.mail.aliyun.com (out30-98.freemail.mail.aliyun.com [115.124.30.98]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 80F703B6340; Sun, 17 May 2026 14:22:22 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=115.124.30.98 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1779027755; cv=none; b=sIg2Nc5EIojHacddBhlIYIBNm9bxlnxb6VEGJqiJhu4bnHXAmNwKfWyOxD8iYAA1dZu1AfeYT5QGRp4m2lit9LEeXVoFUe6QNcTrlDxSMG4I+wHYrHZqef1u5osMO02imEe8X5RMJkR4ZLOcF+zsqwFS0BLAKA1jt7KX7ZRNtLc= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1779027755; c=relaxed/simple; bh=5hY3F0Ab1d0mPnYcm8X5rkr2dRXRed9/0i2JO6zD8OI=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=q7nL8pUcplKLNZKf+7OClQcgN46KKpJdFdLgQ7IWYJRiWd20pihRpoYF/zmwBVr5LAPgOpdtaaAFYYaPIYC+oqWfWV45gFUiKRQ+lIIE05qsjvuBppAcR8xyQC3GQ1oQlCv8SzMktMmNzGGiVLEJ1eF9mn3U5GcvQOwLEf5P2xE= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.alibaba.com; spf=pass smtp.mailfrom=linux.alibaba.com; dkim=pass (1024-bit key) header.d=linux.alibaba.com header.i=@linux.alibaba.com header.b=AXILGgDN; arc=none smtp.client-ip=115.124.30.98 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.alibaba.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=linux.alibaba.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linux.alibaba.com header.i=@linux.alibaba.com header.b="AXILGgDN" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linux.alibaba.com; s=default; t=1779027733; h=From:To:Subject:Date:Message-ID:MIME-Version; bh=CfBA/2IgnyZniAFp9jsEvw4nY5aItReLADys2BWbIxs=; b=AXILGgDNA1nzKP5TqoNZ2dgJA0Q6PJwMe6OsTJwvXyLSoZ5XrRuY/O/59KzHubZULOdtMsLahD4Y9/k3smOWxEeZ7uzBii3RrJUxjG0IFyNxshabnPKtYBC7xNggE+YyW4t9+GbQt/slw8dMLew8lOZojfJq01ldWKYkKLVed14= X-Alimail-AntiSpam: AC=PASS;BC=-1|-1;BR=01201311R131e4;CH=green;DM=||false|;DS=||;FP=0|-1|-1|-1|0|-1|-1|-1;HT=maildocker-contentspam033037009110;MF=libaokun@linux.alibaba.com;NM=1;PH=DS;RN=8;SR=0;TI=SMTPD_---0X31pQEj_1779027732; Received: from x31h02109.sqa.na131.tbsite.net(mailfrom:libaokun@linux.alibaba.com fp:SMTPD_---0X31pQEj_1779027732 cluster:ay36) by smtp.aliyun-inc.com; Sun, 17 May 2026 22:22:12 +0800 From: Baokun Li To: linux-fsdevel@vger.kernel.org Cc: viro@zeniv.linux.org.uk, brauner@kernel.org, jack@suse.cz, tj@kernel.org, linux-kernel@vger.kernel.org Subject: [PATCH v2 2/3] writeback: drop now-unnecessary rcu_barrier() in cgroup_writeback_umount() Date: Sun, 17 May 2026 22:21:31 +0800 Message-ID: <20260517142147.3354909-3-libaokun@linux.alibaba.com> X-Mailer: git-send-email 2.43.7 In-Reply-To: <20260517142147.3354909-1-libaokun@linux.alibaba.com> References: <20260517142147.3354909-1-libaokun@linux.alibaba.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Commit e1b849cfa6b6 ("writeback: Avoid contention on wb->list_lock when switching inodes") replaced the queue_rcu_work() based scheduling of inode wb switches with a plain queue_work(). Since then no switcher goes through call_rcu(), so rcu_barrier() in cgroup_writeback_umount() has no work to wait for and is effectively a no-op. Fixes: e1b849cfa6b6 ("writeback: Avoid contention on wb->list_lock when swi= tching inodes") Signed-off-by: Baokun Li Reviewed-by: Jan Kara --- fs/fs-writeback.c | 5 ----- 1 file changed, 5 deletions(-) diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index 1f95ddcee363..9ae290547eb2 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -1246,11 +1246,6 @@ void cgroup_writeback_umount(struct super_block *sb) * will then drain it. */ synchronize_rcu(); - /* - * Use rcu_barrier() to wait for all pending callbacks to - * ensure that all in-flight wb switches are in the workqueue. - */ - rcu_barrier(); flush_workqueue(isw_wq); } } --=20 2.43.7 From nobody Mon May 25 06:40:17 2026 Received: from out30-99.freemail.mail.aliyun.com (out30-99.freemail.mail.aliyun.com [115.124.30.99]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 293F130BBB8; Sun, 17 May 2026 14:22:22 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=115.124.30.99 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1779027751; cv=none; b=GB7DRs3P/LX9l7UsmOH8SZlct4VUY9GJUBCzMwikzgUlnjSRQXQ8AAuEqHrudBfXWR/foObCu+7QjKm/xhGDWiLfGZp0LmDZ4vukfZlddQTvgGO86JywN5QLZUgWgNG/FEo/voM96/PtmwQItDPy02uTDrNsP48B2O07/LUvXN4= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1779027751; c=relaxed/simple; bh=lurzRAxIpmYepF6PXpBqusEmZv/wOtLAIdE2S3Wq1bY=; h=From:To:Cc:Subject:Date:Message-ID:In-Reply-To:References: MIME-Version; b=RIuyD37iNPMPkaghhO41B2yUWA9l4GBUUmT+MFopiMEBa4FMguDsOnUEmFLzxYIFWnnPSPadF/WVCRJO1b00s8x20wSUGeB8B+SEvImI44QtwqEkOKTqCuXY+IkWLzfV2JBsM/WqX4QWV2x/uRN+jBYCoFIWnN0XNBTmcdQqhu8= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.alibaba.com; spf=pass smtp.mailfrom=linux.alibaba.com; dkim=pass (1024-bit key) header.d=linux.alibaba.com header.i=@linux.alibaba.com header.b=nMsikzCa; arc=none smtp.client-ip=115.124.30.99 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=linux.alibaba.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=linux.alibaba.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key) header.d=linux.alibaba.com header.i=@linux.alibaba.com header.b="nMsikzCa" DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linux.alibaba.com; s=default; t=1779027734; h=From:To:Subject:Date:Message-ID:MIME-Version; bh=RAGy0yHiPLGaug7QeMFsI9pLhb7Hn7COALaw60KsCfE=; b=nMsikzCaom/JCxuS2cDErVzlGNgs1JJJtDjKSfI++34Rkn/H6YW1rmjecT1TdLl8S+frDjXjM434zB0WxcrtGKsR4zX12GVptlNdM0VUwqHXXUqZqu8s8zktJLDhso8sg+fAm48JfUbMjvFvu43VAeEs5u4hiPINI+ok5XzrW4I= X-Alimail-AntiSpam: AC=PASS;BC=-1|-1;BR=01201311R161e4;CH=green;DM=||false|;DS=||;FP=0|-1|-1|-1|0|-1|-1|-1;HT=maildocker-contentspam033045133197;MF=libaokun@linux.alibaba.com;NM=1;PH=DS;RN=8;SR=0;TI=SMTPD_---0X31pQF-_1779027733; Received: from x31h02109.sqa.na131.tbsite.net(mailfrom:libaokun@linux.alibaba.com fp:SMTPD_---0X31pQF-_1779027733 cluster:ay36) by smtp.aliyun-inc.com; Sun, 17 May 2026 22:22:13 +0800 From: Baokun Li To: linux-fsdevel@vger.kernel.org Cc: viro@zeniv.linux.org.uk, brauner@kernel.org, jack@suse.cz, tj@kernel.org, linux-kernel@vger.kernel.org Subject: [PATCH v2 3/3] writeback: use a per-sb counter to drain inode wb switches at umount Date: Sun, 17 May 2026 22:21:32 +0800 Message-ID: <20260517142147.3354909-4-libaokun@linux.alibaba.com> X-Mailer: git-send-email 2.43.7 In-Reply-To: <20260517142147.3354909-1-libaokun@linux.alibaba.com> References: <20260517142147.3354909-1-libaokun@linux.alibaba.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable Content-Type: text/plain; charset="utf-8" Tracking in-flight inode wb switches with a single global counter (isw_nr_in_flight) plus a synchronize_rcu() based wait in cgroup_writeback_umount() forces every umount to take a global hit whenever any other superblock on the system has wb switches in flight, even if the superblock being unmounted has none of its own. Replace the global synchronize_rcu()/flush_workqueue() pair with a per-sb counter, s_isw_nr_in_flight: - inode_prepare_wbs_switch() increments before checking SB_ACTIVE and grabbing the inode; failure paths decrement before returning. - process_inode_switch_wbs() decrements after the matching iput(). - cgroup_writeback_umount() waits for the per-sb counter to reach zero. The smp_mb() pair between inode_prepare_wbs_switch() and cgroup_writeback_umount() keeps the SB_ACTIVE / counter ordering: either the umounter sees a non-zero counter and waits, or the switcher sees SB_ACTIVE cleared and aborts before grabbing the inode. The existing global isw_nr_in_flight is left in place, since it is still used to throttle in-flight switches via WB_FRN_MAX_IN_FLIGHT. The rcu_read_lock() extension in inode_switch_wbs() and cleanup_offline_cgwb() introduced by the race fix is no longer needed and is reverted. The synchronize_rcu() in cgroup_writeback_umount() is dropped as well. inode_prepare_wbs_switch() also gains a lockless SB_ACTIVE check at the top to skip the atomic_inc/smp_mb cost for inodes whose sb is already being torn down. The following numbers were measured with 4 background superblocks each churning "create memcg -> write 1 MiB -> rmdir memcg" to keep the global isw_nr_in_flight non-zero. A separate idle target sb is mounted and umounted in a loop (N=3D100); only its umount latency is measured. Idle target umount latency under cross-sb cgwb-switch pressure: p50 p95 p99 max global synchronize_rcu() 64.4 ms 95.8 ms 101.4 ms 110.5 ms per-sb counter (this) 5.3 ms 6.9 ms 7.4 ms 7.7 ms no-pressure baseline 5.2 ms 5.9 ms 6.0 ms 6.1 ms 8 concurrent umounts of idle sbs under the same pressure: p50 p95 max global synchronize_rcu() 57.9 ms 82.1 ms 90.0 ms per-sb counter (this) 7.5 ms 7.8 ms 8.0 ms In-kernel cgroup_writeback_umount() time over 286 calls (bpftrace): global synchronize_rcu() 8717 ms total (~30 ms / call) per-sb counter (this) 1.16 ms total (~4 us / call) When s_isw_nr_in_flight is zero the loop body is never entered. Signed-off-by: Baokun Li --- fs/fs-writeback.c | 74 ++++++++++++++-------------------- include/linux/fs/super_types.h | 8 ++++ 2 files changed, 38 insertions(+), 44 deletions(-) diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index 9ae290547eb2..4282fcfe027b 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -554,8 +554,12 @@ static void process_inode_switch_wbs(struct bdi_writeb= ack *new_wb, wb_put_many(old_wb, nr_switched); } =20 - for (inodep =3D isw->inodes; *inodep; inodep++) + for (inodep =3D isw->inodes; *inodep; inodep++) { + struct super_block *sb =3D (*inodep)->i_sb; + iput(*inodep); + atomic_dec(&sb->s_isw_nr_in_flight); + } wb_put(new_wb); kfree(isw); atomic_dec(&isw_nr_in_flight); @@ -598,16 +602,19 @@ void inode_switch_wbs_work_fn(struct work_struct *wor= k) static bool inode_prepare_wbs_switch(struct inode *inode, struct bdi_writeback *new_wb) { + /* Avoid the atomic_inc/smp_mb dance once SB_ACTIVE is gone. */ + if (!(inode->i_sb->s_flags & SB_ACTIVE)) + return false; + /* - * Paired with smp_mb() in cgroup_writeback_umount(). - * isw_nr_in_flight must be increased before checking SB_ACTIVE and - * grabbing an inode, otherwise isw_nr_in_flight can be observed as 0 - * in cgroup_writeback_umount() and the isw_wq will be not flushed. + * Pairs with smp_mb() in cgroup_writeback_umount(): the umounter either + * sees a non-zero counter and waits, or we see SB_ACTIVE clear below. */ + atomic_inc(&inode->i_sb->s_isw_nr_in_flight); smp_mb(); =20 if (IS_DAX(inode)) - return false; + goto out_dec; =20 /* while holding I_WB_SWITCH, no one else can update the association */ spin_lock(&inode->i_lock); @@ -615,13 +622,17 @@ static bool inode_prepare_wbs_switch(struct inode *in= ode, inode_state_read(inode) & (I_WB_SWITCH | I_FREEING | I_WILL_FREE) || inode_to_wb(inode) =3D=3D new_wb) { spin_unlock(&inode->i_lock); - return false; + goto out_dec; } inode_state_set(inode, I_WB_SWITCH); __iget(inode); spin_unlock(&inode->i_lock); =20 return true; + +out_dec: + atomic_dec(&inode->i_sb->s_isw_nr_in_flight); + return false; } =20 static void wb_queue_isw(struct bdi_writeback *wb, @@ -665,6 +676,7 @@ static void inode_switch_wbs(struct inode *inode, int n= ew_wb_id) memcg_css =3D css_from_id(new_wb_id, &memory_cgrp_subsys); if (memcg_css && !css_tryget(memcg_css)) memcg_css =3D NULL; + rcu_read_unlock(); if (!memcg_css) goto out_free; =20 @@ -679,18 +691,10 @@ static void inode_switch_wbs(struct inode *inode, int= new_wb_id) isw->inodes[0] =3D inode; =20 trace_inode_switch_wbs_queue(inode->i_wb, new_wb, 1); - /* - * Paired with synchronize_rcu() in cgroup_writeback_umount(). - * Holding rcu_read_lock across wb_queue_isw() ensures - * synchronize_rcu() cannot return until the work is queued, so - * the subsequent flush_workqueue() will wait for the switch. - */ wb_queue_isw(new_wb, isw); - rcu_read_unlock(); return; =20 out_free: - rcu_read_unlock(); atomic_dec(&isw_nr_in_flight); if (new_wb) wb_put(new_wb); @@ -748,14 +752,6 @@ bool cleanup_offline_cgwb(struct bdi_writeback *wb) new_wb =3D &wb->bdi->wb; /* wb_get() is noop for bdi's wb */ =20 nr =3D 0; - /* - * Paired with synchronize_rcu() in cgroup_writeback_umount(). - * Holding rcu_read_lock across the SB_ACTIVE check, the inode grab - * and wb_queue_isw() ensures synchronize_rcu() cannot return until - * the work is queued, so the subsequent flush_workqueue() will wait - * for the switch. - */ - rcu_read_lock(); spin_lock(&wb->list_lock); /* * In addition to the inodes that have completed writeback, also switch @@ -773,7 +769,6 @@ bool cleanup_offline_cgwb(struct bdi_writeback *wb) =20 /* no attached inodes? bail out */ if (nr =3D=3D 0) { - rcu_read_unlock(); atomic_dec(&isw_nr_in_flight); wb_put(new_wb); kfree(isw); @@ -782,7 +777,6 @@ bool cleanup_offline_cgwb(struct bdi_writeback *wb) =20 trace_inode_switch_wbs_queue(wb, new_wb, nr); wb_queue_isw(new_wb, isw); - rcu_read_unlock(); =20 return restart; } @@ -1215,38 +1209,30 @@ int cgroup_writeback_by_id(u64 bdi_id, int memcg_id, } =20 /** - * cgroup_writeback_umount - flush inode wb switches for umount + * cgroup_writeback_umount - wait for in-flight inode wb switches on @sb * @sb: target super_block * - * This function is called when a super_block is about to be destroyed and - * flushes in-flight inode wb switches. An inode wb switch goes through - * RCU and then workqueue, so the two need to be flushed in order to ensure - * that all previously scheduled switches are finished. As wb switches are - * rare occurrences and synchronize_rcu() can take a while, perform - * flushing iff wb switches are in flight. + * Wait until every inode wb switch that already passed the SB_ACTIVE + * check on this superblock has been completed by the worker. Since + * SB_ACTIVE is cleared before this is called, no new switches can start + * for @sb, so s_isw_nr_in_flight will monotonically drop to zero. */ void cgroup_writeback_umount(struct super_block *sb) { - if (!(sb->s_bdi->capabilities & BDI_CAP_WRITEBACK)) return; =20 /* - * SB_ACTIVE should be reliably cleared before checking - * isw_nr_in_flight, see generic_shutdown_super(). + * Pairs with smp_mb() in inode_prepare_wbs_switch(): we either observe + * a non-zero counter and wait, or the switcher sees SB_ACTIVE clear + * (cleared by generic_shutdown_super()) and bails before grabbing the + * inode. */ smp_mb(); =20 - if (atomic_read(&isw_nr_in_flight)) { - /* - * Paired with rcu_read_lock() in inode_switch_wbs() and - * cleanup_offline_cgwb(). synchronize_rcu() waits for any - * in-flight switcher that already passed the SB_ACTIVE check - * to finish queueing its work, so flush_workqueue() below - * will then drain it. - */ - synchronize_rcu(); + while (atomic_read(&sb->s_isw_nr_in_flight)) { flush_workqueue(isw_wq); + cond_resched(); } } =20 diff --git a/include/linux/fs/super_types.h b/include/linux/fs/super_types.h index 383050e7fdf5..1ab4e2265129 100644 --- a/include/linux/fs/super_types.h +++ b/include/linux/fs/super_types.h @@ -274,6 +274,14 @@ struct super_block { =20 /* number of fserrors that are being sent to fsnotify/filesystems */ refcount_t s_pending_errors; + +#ifdef CONFIG_CGROUP_WRITEBACK + /* + * Number of in-flight inode wb switches for this sb. Drained by + * cgroup_writeback_umount() before tear-down. + */ + atomic_t s_isw_nr_in_flight; +#endif } __randomize_layout; =20 /* --=20 2.43.7