status_resync() will calculate 'curr_resync - recovery_active' to show
user a progress bar like following:
[============>........]  resync = 61.4%
'curr_resync' and 'recovery_active' is updated in md_do_sync(), and
status_resync() can read them concurrently, hence it's possible that
'curr_resync - recovery_active' can overflow to a huge number. In this
case status_resync() will be stuck in the loop to print a large amount
of '=', which will end up soft lockup.
Fix the problem by setting 'resync' to MD_RESYNC_ACTIVE in this case,
this way resync in progress will be reported to user.
Signed-off-by: Yu Kuai <yukuai3@huawei.com>
Signed-off-by: Song Liu <song@kernel.org>
Link: https://lore.kernel.org/r/20230310073855.1337560-3-yukuai1@huaweicloud.com
 
        } else if (resync > max_sectors) {
                resync = max_sectors;
        } else {
-               resync -= atomic_read(&mddev->recovery_active);
-               if (resync < MD_RESYNC_ACTIVE) {
-                       /*
-                        * Resync has started, but the subtraction has
-                        * yielded one of the special values. Force it
-                        * to active to ensure the status reports an
-                        * active resync.
-                        */
+               res = atomic_read(&mddev->recovery_active);
+               /*
+                * Resync has started, but the subtraction has overflowed or
+                * yielded one of the special values. Force it to active to
+                * ensure the status reports an active resync.
+                */
+               if (resync < res || resync - res < MD_RESYNC_ACTIVE)
                        resync = MD_RESYNC_ACTIVE;
-               }
+               else
+                       resync -= res;
        }
 
        if (resync == MD_RESYNC_NONE) {