]> www.infradead.org Git - nvme.git/commitdiff
md: Make sure md_do_sync() will set MD_RECOVERY_DONE
authorYu Kuai <yukuai3@huawei.com>
Thu, 1 Feb 2024 09:25:48 +0000 (17:25 +0800)
committerSong Liu <song@kernel.org>
Thu, 15 Feb 2024 22:17:18 +0000 (14:17 -0800)
stop_sync_thread() will interrupt md_do_sync(), and md_do_sync() must
set MD_RECOVERY_DONE, so that follow up md_check_recovery() will
unregister sync_thread, clear MD_RECOVERY_RUNNING and wake up
stop_sync_thread().

If MD_RECOVERY_WAIT is set or the array is read-only, md_do_sync() will
return without setting MD_RECOVERY_DONE, and after commit f52f5c71f3d4
("md: fix stopping sync thread"), dm-raid switch from
md_reap_sync_thread() to stop_sync_thread() to unregister sync_thread
from md_stop() and md_stop_writes(), causing the test
shell/lvconvert-raid-reshape.sh hang.

We shouldn't switch back to md_reap_sync_thread() because it's
problematic in the first place. Fix the problem by making sure
md_do_sync() will set MD_RECOVERY_DONE.

Reported-by: Mikulas Patocka <mpatocka@redhat.com>
Closes: https://lore.kernel.org/all/ece2b06f-d647-6613-a534-ff4c9bec1142@redhat.com/
Fixes: d5d885fd514f ("md: introduce new personality funciton start()")
Fixes: 5fd6c1dce06e ("[PATCH] md: allow checkpoint of recovery with version-1 superblock")
Fixes: f52f5c71f3d4 ("md: fix stopping sync thread")
Cc: stable@vger.kernel.org # v6.7+
Signed-off-by: Yu Kuai <yukuai3@huawei.com>
Signed-off-by: Song Liu <song@kernel.org>
Link: https://lore.kernel.org/r/20240201092559.910982-4-yukuai1@huaweicloud.com
drivers/md/md.c

index 2a1c19f87d01272abcaf0e0e18d81cc2caf6020d..fbf04160d0098d8cb7c6cad1383e40b652552279 100644 (file)
@@ -8792,12 +8792,16 @@ void md_do_sync(struct md_thread *thread)
        int ret;
 
        /* just incase thread restarts... */
-       if (test_bit(MD_RECOVERY_DONE, &mddev->recovery) ||
-           test_bit(MD_RECOVERY_WAIT, &mddev->recovery))
+       if (test_bit(MD_RECOVERY_DONE, &mddev->recovery))
                return;
-       if (!md_is_rdwr(mddev)) {/* never try to sync a read-only array */
+
+       if (test_bit(MD_RECOVERY_INTR, &mddev->recovery))
+               goto skip;
+
+       if (test_bit(MD_RECOVERY_WAIT, &mddev->recovery) ||
+           !md_is_rdwr(mddev)) {/* never try to sync a read-only array */
                set_bit(MD_RECOVERY_INTR, &mddev->recovery);
-               return;
+               goto skip;
        }
 
        if (mddev_is_clustered(mddev)) {