From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
stable@vger.kernel.org, Tejun Heo <tj@kernel.org>,
Jiufei Xue <xuejiufei@gmail.com>, Jan Kara <jack@suse.cz>,
Jens Axboe <axboe@kernel.dk>, Sasha Levin <sashal@kernel.org>
Subject: [PATCH 4.20 45/88] writeback: synchronize sync(2) against cgroup writeback membership switches
Date: Mon, 4 Mar 2019 09:22:28 +0100 [thread overview]
Message-ID: <20190304081632.345744157@linuxfoundation.org> (raw)
In-Reply-To: <20190304081630.610632175@linuxfoundation.org>
4.20-stable review patch. If anyone has any objections, please let me know.
------------------
[ Upstream commit 7fc5854f8c6efae9e7624970ab49a1eac2faefb1 ]
sync_inodes_sb() can race against cgwb (cgroup writeback) membership
switches and fail to writeback some inodes. For example, if an inode
switches to another wb while sync_inodes_sb() is in progress, the new
wb might not be visible to bdi_split_work_to_wbs() at all or the inode
might jump from a wb which hasn't issued writebacks yet to one which
already has.
This patch adds backing_dev_info->wb_switch_rwsem to synchronize cgwb
switch path against sync_inodes_sb() so that sync_inodes_sb() is
guaranteed to see all the target wbs and inodes can't jump wbs to
escape syncing.
v2: Fixed misplaced rwsem init. Spotted by Jiufei.
Signed-off-by: Tejun Heo <tj@kernel.org>
Reported-by: Jiufei Xue <xuejiufei@gmail.com>
Link: http://lkml.kernel.org/r/dc694ae2-f07f-61e1-7097-7c8411cee12d@gmail.com
Acked-by: Jan Kara <jack@suse.cz>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
fs/fs-writeback.c | 40 ++++++++++++++++++++++++++++++--
include/linux/backing-dev-defs.h | 1 +
mm/backing-dev.c | 1 +
3 files changed, 40 insertions(+), 2 deletions(-)
diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c
index b40168fcc94a6..36855c1f8dafd 100644
--- a/fs/fs-writeback.c
+++ b/fs/fs-writeback.c
@@ -331,11 +331,22 @@ struct inode_switch_wbs_context {
struct work_struct work;
};
+static void bdi_down_write_wb_switch_rwsem(struct backing_dev_info *bdi)
+{
+ down_write(&bdi->wb_switch_rwsem);
+}
+
+static void bdi_up_write_wb_switch_rwsem(struct backing_dev_info *bdi)
+{
+ up_write(&bdi->wb_switch_rwsem);
+}
+
static void inode_switch_wbs_work_fn(struct work_struct *work)
{
struct inode_switch_wbs_context *isw =
container_of(work, struct inode_switch_wbs_context, work);
struct inode *inode = isw->inode;
+ struct backing_dev_info *bdi = inode_to_bdi(inode);
struct address_space *mapping = inode->i_mapping;
struct bdi_writeback *old_wb = inode->i_wb;
struct bdi_writeback *new_wb = isw->new_wb;
@@ -343,6 +354,12 @@ static void inode_switch_wbs_work_fn(struct work_struct *work)
struct page *page;
bool switched = false;
+ /*
+ * If @inode switches cgwb membership while sync_inodes_sb() is
+ * being issued, sync_inodes_sb() might miss it. Synchronize.
+ */
+ down_read(&bdi->wb_switch_rwsem);
+
/*
* By the time control reaches here, RCU grace period has passed
* since I_WB_SWITCH assertion and all wb stat update transactions
@@ -428,6 +445,8 @@ static void inode_switch_wbs_work_fn(struct work_struct *work)
spin_unlock(&new_wb->list_lock);
spin_unlock(&old_wb->list_lock);
+ up_read(&bdi->wb_switch_rwsem);
+
if (switched) {
wb_wakeup(new_wb);
wb_put(old_wb);
@@ -468,9 +487,18 @@ static void inode_switch_wbs(struct inode *inode, int new_wb_id)
if (inode->i_state & I_WB_SWITCH)
return;
+ /*
+ * Avoid starting new switches while sync_inodes_sb() is in
+ * progress. Otherwise, if the down_write protected issue path
+ * blocks heavily, we might end up starting a large number of
+ * switches which will block on the rwsem.
+ */
+ if (!down_read_trylock(&bdi->wb_switch_rwsem))
+ return;
+
isw = kzalloc(sizeof(*isw), GFP_ATOMIC);
if (!isw)
- return;
+ goto out_unlock;
/* find and pin the new wb */
rcu_read_lock();
@@ -504,12 +532,14 @@ static void inode_switch_wbs(struct inode *inode, int new_wb_id)
* Let's continue after I_WB_SWITCH is guaranteed to be visible.
*/
call_rcu(&isw->rcu_head, inode_switch_wbs_rcu_fn);
- return;
+ goto out_unlock;
out_free:
if (isw->new_wb)
wb_put(isw->new_wb);
kfree(isw);
+out_unlock:
+ up_read(&bdi->wb_switch_rwsem);
}
/**
@@ -887,6 +917,9 @@ fs_initcall(cgroup_writeback_init);
#else /* CONFIG_CGROUP_WRITEBACK */
+static void bdi_down_write_wb_switch_rwsem(struct backing_dev_info *bdi) { }
+static void bdi_up_write_wb_switch_rwsem(struct backing_dev_info *bdi) { }
+
static struct bdi_writeback *
locked_inode_to_wb_and_lock_list(struct inode *inode)
__releases(&inode->i_lock)
@@ -2413,8 +2446,11 @@ void sync_inodes_sb(struct super_block *sb)
return;
WARN_ON(!rwsem_is_locked(&sb->s_umount));
+ /* protect against inode wb switch, see inode_switch_wbs_work_fn() */
+ bdi_down_write_wb_switch_rwsem(bdi);
bdi_split_work_to_wbs(bdi, &work, false);
wb_wait_for_completion(bdi, &done);
+ bdi_up_write_wb_switch_rwsem(bdi);
wait_sb_inodes(sb);
}
diff --git a/include/linux/backing-dev-defs.h b/include/linux/backing-dev-defs.h
index c311571355981..07e02d6df5ad9 100644
--- a/include/linux/backing-dev-defs.h
+++ b/include/linux/backing-dev-defs.h
@@ -190,6 +190,7 @@ struct backing_dev_info {
struct radix_tree_root cgwb_tree; /* radix tree of active cgroup wbs */
struct rb_root cgwb_congested_tree; /* their congested states */
struct mutex cgwb_release_mutex; /* protect shutdown of wb structs */
+ struct rw_semaphore wb_switch_rwsem; /* no cgwb switch while syncing */
#else
struct bdi_writeback_congested *wb_congested;
#endif
diff --git a/mm/backing-dev.c b/mm/backing-dev.c
index 8a8bb8796c6c4..72e6d0c55cfad 100644
--- a/mm/backing-dev.c
+++ b/mm/backing-dev.c
@@ -689,6 +689,7 @@ static int cgwb_bdi_init(struct backing_dev_info *bdi)
INIT_RADIX_TREE(&bdi->cgwb_tree, GFP_ATOMIC);
bdi->cgwb_congested_tree = RB_ROOT;
mutex_init(&bdi->cgwb_release_mutex);
+ init_rwsem(&bdi->wb_switch_rwsem);
ret = wb_init(&bdi->wb, bdi, 1, GFP_KERNEL);
if (!ret) {
--
2.19.1
next prev parent reply other threads:[~2019-03-04 8:34 UTC|newest]
Thread overview: 101+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-03-04 8:21 [PATCH 4.20 00/88] 4.20.14-stable review Greg Kroah-Hartman
2019-03-04 8:21 ` [PATCH 4.20 01/88] genirq/matrix: Improve target CPU selection for managed interrupts Greg Kroah-Hartman
2019-03-04 8:21 ` [PATCH 4.20 02/88] scsi: libsas: Fix rphy phy_identifier for PHYs with end devices attached Greg Kroah-Hartman
2019-03-04 8:21 ` [PATCH 4.20 03/88] drm/msm: Unblock writer if reader closes file Greg Kroah-Hartman
2019-03-04 8:21 ` [PATCH 4.20 04/88] ASoC: Intel: Haswell/Broadwell: fix setting for .dynamic field Greg Kroah-Hartman
2019-03-04 8:21 ` [PATCH 4.20 05/88] ALSA: compress: prevent potential divide by zero bugs Greg Kroah-Hartman
2019-03-04 8:21 ` [PATCH 4.20 06/88] ASoC: rt5682: Fix recording no sound issue Greg Kroah-Hartman
2019-03-04 8:21 ` [PATCH 4.20 07/88] ASoC: Variable "val" in function rt274_i2c_probe() could be uninitialized Greg Kroah-Hartman
2019-03-04 8:21 ` [PATCH 4.20 08/88] clk: tegra: dfll: Fix a potential Oop in remove() Greg Kroah-Hartman
2019-03-04 8:21 ` [PATCH 4.20 09/88] clk: sysfs: fix invalid JSON in clk_dump Greg Kroah-Hartman
2019-03-04 8:21 ` [PATCH 4.20 10/88] clk: vc5: Abort clock configuration without upstream clock Greg Kroah-Hartman
2019-03-04 8:21 ` [PATCH 4.20 11/88] thermal: int340x_thermal: Fix a NULL vs IS_ERR() check Greg Kroah-Hartman
2019-03-04 8:21 ` [PATCH 4.20 12/88] usb: dwc3: gadget: synchronize_irq dwc irq in suspend Greg Kroah-Hartman
2019-03-04 8:46 ` He, Bo
2019-03-04 8:53 ` Greg Kroah-Hartman
2019-03-04 9:03 ` Marek Szyprowski
2019-03-04 9:44 ` Greg Kroah-Hartman
2019-03-04 8:21 ` [PATCH 4.20 13/88] usb: dwc3: gadget: Fix the uninitialized link_state when udc starts Greg Kroah-Hartman
2019-03-04 8:21 ` [PATCH 4.20 14/88] usb: gadget: Potential NULL dereference on allocation error Greg Kroah-Hartman
2019-03-04 8:21 ` [PATCH 4.20 15/88] HID: i2c-hid: Disable runtime PM on Goodix touchpad Greg Kroah-Hartman
2019-03-04 8:21 ` [PATCH 4.20 16/88] ASoC: core: Make snd_soc_find_component() more robust Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 17/88] selftests: rtc: rtctest: fix alarm tests Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 18/88] selftests: rtc: rtctest: add alarm test on minute boundary Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 19/88] genirq: Make sure the initial affinity is not empty Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 20/88] x86/mm/mem_encrypt: Fix erroneous sizeof() Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 21/88] ASoC: rt5682: Fix PLL source register definitions Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 22/88] ASoC: dapm: change snprintf to scnprintf for possible overflow Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 23/88] ASoC: imx-audmux: " Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 24/88] selftests/vm/gup_benchmark.c: match gup struct to kernel Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 25/88] phy: ath79-usb: Fix the power on error path Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 26/88] phy: ath79-usb: Fix the main reset name to match the DT binding Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 27/88] selftests: seccomp: use LDLIBS instead of LDFLAGS Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 28/88] selftests: gpio-mockup-chardev: Check asprintf() for error Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 29/88] irqchip/gic-v3-mbi: Fix uninitialized mbi_lock Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 30/88] ARC: fix __ffs return value to avoid build warnings Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 31/88] ARC: show_regs: lockdep: avoid page allocator Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 32/88] drivers: thermal: int340x_thermal: Fix sysfs race condition Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 33/88] staging: rtl8723bs: Fix build error with Clang when inlining is disabled Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 34/88] mac80211: fix miscounting of ttl-dropped frames Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 35/88] sched/wait: Fix rcuwait_wake_up() ordering Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 36/88] sched/wake_q: Fix wakeup ordering for wake_q Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 37/88] futex: Fix (possible) missed wakeup Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 38/88] locking/rwsem: " Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 39/88] drm/amd/powerplay: OD setting fix on Vega10 Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 40/88] tty: serial: qcom_geni_serial: Allow mctrl when flow control is disabled Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 41/88] serial: fsl_lpuart: fix maximum acceptable baud rate with over-sampling Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 42/88] drm/sun4i: hdmi: Fix usage of TMDS clock Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 43/88] staging: android: ion: Support cpu access during dma_buf_detach Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 44/88] direct-io: allow direct writes to empty inodes Greg Kroah-Hartman
2019-03-04 8:22 ` Greg Kroah-Hartman [this message]
2019-03-04 8:22 ` [PATCH 4.20 46/88] scsi: lpfc: nvme: avoid hang / use-after-free when destroying localport Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 47/88] scsi: lpfc: nvmet: avoid hang / use-after-free when destroying targetport Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 48/88] scsi: csiostor: fix NULL pointer dereference in csio_vport_set_state() Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 49/88] net: altera_tse: fix connect_local_phy error path Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 50/88] hv_netvsc: Fix ethtool change hash key error Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 51/88] hv_netvsc: Refactor assignments of struct netvsc_device_info Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 52/88] hv_netvsc: Fix hash key value reset after other ops Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 53/88] nvme-rdma: fix timeout handler Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 54/88] nvme-multipath: drop optimization for static ANA group IDs Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 55/88] cifs: fix memory leak of an allocated cifs_ntsd structure Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 56/88] drm/msm: Fix A6XX support for opp-level Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 57/88] drm/msm: avoid unused function warning Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 58/88] net: usb: asix: ax88772_bind return error when hw_reset fail Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 59/88] net: dev_is_mac_header_xmit() true for ARPHRD_RAWIP Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 60/88] ibmveth: Do not process frames after calling napi_reschedule Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 61/88] mac80211: dont initiate TDLS connection if station is not associated to AP Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 62/88] mac80211: Add attribute aligned(2) to struct action Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 63/88] cfg80211: extend range deviation for DMG Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 64/88] svm: Fix AVIC incomplete IPI emulation Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 65/88] KVM: nSVM: clear events pending from svm_complete_interrupts() when exiting to L1 Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 66/88] kvm: selftests: Fix region overlap check in kvm_util Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 67/88] KVM: selftests: check returned evmcs version range Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 68/88] mmc: spi: Fix card detection during probe Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 69/88] mmc: tmio_mmc_core: dont claim spurious interrupts Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 70/88] mmc: tmio: fix access width of Block Count Register Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 71/88] mmc: core: Fix NULL ptr crash from mmc_should_fail_request Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 72/88] mmc: cqhci: fix space allocated for transfer descriptor Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 73/88] mmc: cqhci: Fix a tiny potential memory leak on error condition Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 74/88] mmc: sdhci-esdhc-imx: correct the fix of ERR004536 Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 75/88] mm: enforce min addr even if capable() in expand_downwards() Greg Kroah-Hartman
2019-03-04 8:22 ` [PATCH 4.20 76/88] drm: Block fb changes for async plane updates Greg Kroah-Hartman
2019-03-04 8:23 ` [PATCH 4.20 77/88] hugetlbfs: fix races and page leaks during migration Greg Kroah-Hartman
2019-03-04 8:23 ` [PATCH 4.20 78/88] crypto: ccree - add missing inline qualifier Greg Kroah-Hartman
2019-03-04 8:23 ` [PATCH 4.20 79/88] MIPS: fix truncation in __cmpxchg_small for short values Greg Kroah-Hartman
2019-03-04 8:23 ` [PATCH 4.20 80/88] MIPS: BCM63XX: provide DMA masks for ethernet devices Greg Kroah-Hartman
2019-03-04 8:23 ` [PATCH 4.20 81/88] MIPS: fix memory setup for platforms with PHYS_OFFSET != 0 Greg Kroah-Hartman
2019-03-04 8:23 ` [PATCH 4.20 82/88] scsi: 3w-sas: fix calls to dma_set_mask_and_coherent() Greg Kroah-Hartman
2019-03-04 8:23 ` [PATCH 4.20 83/88] scsi: csiostor: " Greg Kroah-Hartman
2019-03-04 8:23 ` [PATCH 4.20 84/88] scsi: 3w-9xxx: " Greg Kroah-Hartman
2019-03-04 8:23 ` [PATCH 4.20 85/88] scsi: aic94xx: " Greg Kroah-Hartman
2019-03-04 8:23 ` [PATCH 4.20 86/88] arm64: dts: qcom: msm8998: Extend TZ reserved memory area Greg Kroah-Hartman
2019-03-04 8:23 ` [PATCH 4.20 87/88] MIPS: eBPF: Fix icache flush end address Greg Kroah-Hartman
2019-03-04 8:23 ` [PATCH 4.20 88/88] x86/uaccess: Dont leak the AC flag into __put_user() value evaluation Greg Kroah-Hartman
2019-03-04 20:36 ` [PATCH 4.20 00/88] 4.20.14-stable review Naresh Kamboju
2019-03-05 7:58 ` Greg Kroah-Hartman
2019-03-05 3:39 ` Guenter Roeck
2019-03-05 7:58 ` Greg Kroah-Hartman
2019-03-05 14:07 ` Jon Hunter
2019-03-05 14:55 ` Greg Kroah-Hartman
2019-03-05 16:20 ` shuah
2019-03-05 16:51 ` Greg Kroah-Hartman
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20190304081632.345744157@linuxfoundation.org \
--to=gregkh@linuxfoundation.org \
--cc=axboe@kernel.dk \
--cc=jack@suse.cz \
--cc=linux-kernel@vger.kernel.org \
--cc=sashal@kernel.org \
--cc=stable@vger.kernel.org \
--cc=tj@kernel.org \
--cc=xuejiufei@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).