The patch below does not apply to the 5.4-stable tree.
If someone wants it applied there, or to any other stable or longterm
tree, then please email the backport, including the original git commit
id to <[email protected]>.

Thanks,
Sasha

------------------ original commit in Linus's tree ------------------

>From cd32b27a66db8776d8b8e82ec7d7dde97a8693b0 Mon Sep 17 00:00:00 2001
From: Yu Kuai <[email protected]>
Date: Tue, 5 Mar 2024 15:23:03 +0800
Subject: [PATCH] md/dm-raid: don't call md_reap_sync_thread() directly

Currently md_reap_sync_thread() is called from raid_message() directly
without holding 'reconfig_mutex', this is definitely unsafe because
md_reap_sync_thread() can change many fields that is protected by
'reconfig_mutex'.

However, hold 'reconfig_mutex' here is still problematic because this
will cause deadlock, for example, commit 130443d60b1b ("md: refactor
idle/frozen_sync_thread() to fix deadlock").

Fix this problem by using stop_sync_thread() to unregister sync_thread,
like md/raid did.

Fixes: be83651f0050 ("DM RAID: Add message/status support for changing sync 
action")
Cc: [email protected] # v6.7+
Signed-off-by: Yu Kuai <[email protected]>
Signed-off-by: Xiao Ni <[email protected]>
Acked-by: Mike Snitzer <[email protected]>
Signed-off-by: Song Liu <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
---
 drivers/md/dm-raid.c | 28 ++++++++++++++++++----------
 1 file changed, 18 insertions(+), 10 deletions(-)

diff --git a/drivers/md/dm-raid.c b/drivers/md/dm-raid.c
index e2d7a73c0f874..47c4b1b6e532a 100644
--- a/drivers/md/dm-raid.c
+++ b/drivers/md/dm-raid.c
@@ -3719,6 +3719,7 @@ static int raid_message(struct dm_target *ti, unsigned 
int argc, char **argv,
 {
        struct raid_set *rs = ti->private;
        struct mddev *mddev = &rs->md;
+       int ret = 0;
 
        if (!mddev->pers || !mddev->pers->sync_request)
                return -EINVAL;
@@ -3726,17 +3727,24 @@ static int raid_message(struct dm_target *ti, unsigned 
int argc, char **argv,
        if (test_bit(RT_FLAG_RS_SUSPENDED, &rs->runtime_flags))
                return -EBUSY;
 
-       if (!strcasecmp(argv[0], "frozen"))
-               set_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
-       else
-               clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
+       if (!strcasecmp(argv[0], "frozen")) {
+               ret = mddev_lock(mddev);
+               if (ret)
+                       return ret;
 
-       if (!strcasecmp(argv[0], "idle") || !strcasecmp(argv[0], "frozen")) {
-               if (mddev->sync_thread) {
-                       set_bit(MD_RECOVERY_INTR, &mddev->recovery);
-                       md_reap_sync_thread(mddev);
-               }
-       } else if (decipher_sync_action(mddev, mddev->recovery) != st_idle)
+               md_frozen_sync_thread(mddev);
+               mddev_unlock(mddev);
+       } else if (!strcasecmp(argv[0], "idle")) {
+               ret = mddev_lock(mddev);
+               if (ret)
+                       return ret;
+
+               md_idle_sync_thread(mddev);
+               mddev_unlock(mddev);
+       }
+
+       clear_bit(MD_RECOVERY_FROZEN, &mddev->recovery);
+       if (decipher_sync_action(mddev, mddev->recovery) != st_idle)
                return -EBUSY;
        else if (!strcasecmp(argv[0], "resync"))
                ; /* MD_RECOVERY_NEEDED set below */
-- 
2.43.0





Reply via email to