From: Guoqing Jiang <guoqing.jiang@linux.dev>
To: song@kernel.org
Cc: buczek@molgen.mpg.de, linux-raid@vger.kernel.org,
Guoqing Jiang <guoqing.jiang@cloud.ionos.com>,
Guoqing Jiang <guoqing.jiang@linux.dev>
Subject: [PATCH V3 1/2] md: don't unregister sync_thread with reconfig_mutex held
Date: Thu, 5 May 2022 16:16:40 +0800 [thread overview]
Message-ID: <20220505081641.21500-2-guoqing.jiang@linux.dev> (raw)
In-Reply-To: <20220505081641.21500-1-guoqing.jiang@linux.dev>
From: Guoqing Jiang <guoqing.jiang@cloud.ionos.com>
Unregister sync_thread doesn't need to hold reconfig_mutex since it
doesn't reconfigure array.
And it could cause deadlock problem for raid5 as follows:
1. process A tried to reap sync thread with reconfig_mutex held after echo
idle to sync_action.
2. raid5 sync thread was blocked if there were too many active stripes.
3. SB_CHANGE_PENDING was set (because of write IO comes from upper layer)
which causes the number of active stripes can't be decreased.
4. SB_CHANGE_PENDING can't be cleared since md_check_recovery was not able
to hold reconfig_mutex.
More details in the link:
https://lore.kernel.org/linux-raid/5ed54ffc-ce82-bf66-4eff-390cb23bc1ac@molgen.mpg.de/T/#t
Let's call md_unregister_thread between mddev_unlock and mddev_lock if
reconfig_mutex is held, and mddev_is_locked is introduced accordingly.
Reported-by: Donald Buczek <buczek@molgen.mpg.de>
Signed-off-by: Guoqing Jiang <guoqing.jiang@linux.dev>
---
v2: https://lore.kernel.org/all/1613177399-22024-1-git-send-email-guoqing.jiang@cloud.ionos.com/#t
v1: https://lore.kernel.org/all/1612923676-18294-1-git-send-email-guoqing.jiang@cloud.ionos.com/
drivers/md/md.c | 9 ++++++++-
drivers/md/md.h | 5 +++++
2 files changed, 13 insertions(+), 1 deletion(-)
diff --git a/drivers/md/md.c b/drivers/md/md.c
index 309b3af906ad..a70e7f0f9268 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -9432,10 +9432,17 @@ void md_reap_sync_thread(struct mddev *mddev)
{
struct md_rdev *rdev;
sector_t old_dev_sectors = mddev->dev_sectors;
- bool is_reshaped = false;
+ bool is_reshaped = false, is_locked = false;
+ if (mddev_is_locked(mddev)) {
+ is_locked = true;
+ mddev_unlock(mddev);
+ }
/* resync has finished, collect result */
md_unregister_thread(&mddev->sync_thread);
+ if (is_locked)
+ mddev_lock(mddev);
+
if (!test_bit(MD_RECOVERY_INTR, &mddev->recovery) &&
!test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery) &&
mddev->degraded != mddev->raid_disks) {
diff --git a/drivers/md/md.h b/drivers/md/md.h
index 6ac283864533..af6f3978b62b 100644
--- a/drivers/md/md.h
+++ b/drivers/md/md.h
@@ -549,6 +549,11 @@ static inline int mddev_trylock(struct mddev *mddev)
}
extern void mddev_unlock(struct mddev *mddev);
+static inline int mddev_is_locked(struct mddev *mddev)
+{
+ return mutex_is_locked(&mddev->reconfig_mutex);
+}
+
static inline void md_sync_acct(struct block_device *bdev, unsigned long nr_sectors)
{
atomic_add(nr_sectors, &bdev->bd_disk->sync_io);
--
2.31.1
next prev parent reply other threads:[~2022-05-05 8:19 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2022-05-05 8:16 [PATCH 0/2] two fixes for md Guoqing Jiang
2022-05-05 8:16 ` Guoqing Jiang [this message]
2022-05-05 14:02 ` [PATCH V3 1/2] md: don't unregister sync_thread with reconfig_mutex held kernel test robot
2022-05-05 18:04 ` kernel test robot
2022-05-06 2:34 ` Guoqing Jiang
2022-05-05 8:16 ` [PATCH 2/2] md: protect md_unregister_thread from reentrancy Guoqing Jiang
2022-05-09 6:39 ` Song Liu
2022-05-09 8:12 ` Guoqing Jiang
2022-05-06 11:36 ` [Update PATCH V3] md: don't unregister sync_thread with reconfig_mutex held Guoqing Jiang
2022-05-09 6:37 ` Song Liu
2022-05-09 8:09 ` Guoqing Jiang
2022-05-09 9:32 ` Wols Lists
2022-05-09 10:37 ` Guoqing Jiang
2022-05-09 11:19 ` Wols Lists
2022-05-09 11:26 ` Guoqing Jiang
2022-05-10 6:44 ` Song Liu
2022-05-10 12:01 ` Donald Buczek
2022-05-10 12:09 ` Guoqing Jiang
2022-05-10 12:35 ` Donald Buczek
2022-05-10 18:02 ` Song Liu
2022-05-11 8:10 ` Guoqing Jiang
2022-05-11 21:45 ` Song Liu
2022-05-20 18:27 ` Logan Gunthorpe
2022-05-21 18:23 ` Donald Buczek
2022-05-23 1:08 ` Guoqing Jiang
2022-05-23 5:41 ` Donald Buczek
2022-05-23 9:51 ` Guoqing Jiang
2022-05-24 16:13 ` Logan Gunthorpe
2022-05-25 9:04 ` Guoqing Jiang
2022-05-25 18:22 ` Logan Gunthorpe
2022-05-26 9:46 ` Jan Kara
2022-05-26 11:53 ` Jan Kara
2022-05-31 6:11 ` Christoph Hellwig
2022-05-31 7:43 ` Jan Kara
2022-05-30 9:55 ` Guoqing Jiang
2022-05-30 16:35 ` Logan Gunthorpe
2022-05-31 8:13 ` Guoqing Jiang
2022-05-24 15:58 ` Logan Gunthorpe
2022-05-24 18:16 ` Song Liu
2022-05-25 9:17 ` Guoqing Jiang
2022-05-24 15:51 ` Logan Gunthorpe
2022-06-02 8:12 ` Xiao Ni
2022-05-09 8:18 ` Donald Buczek
2022-05-09 8:48 ` Guoqing Jiang
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=20220505081641.21500-2-guoqing.jiang@linux.dev \
--to=guoqing.jiang@linux.dev \
--cc=buczek@molgen.mpg.de \
--cc=guoqing.jiang@cloud.ionos.com \
--cc=linux-raid@vger.kernel.org \
--cc=song@kernel.org \
/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).