dm raid: fix raid "check" regression due to improper cleanup in raid_message()
authorMike Snitzer <snitzer@redhat.com>
Tue, 28 Feb 2017 16:55:16 +0000 (11:55 -0500)
committerMike Snitzer <snitzer@redhat.com>
Tue, 28 Feb 2017 21:47:50 +0000 (16:47 -0500)
While cleaning up awkward branching in raid_message() a raid set "check"
regression was introduced because "check" needs both MD_RECOVERY_SYNC
and MD_RECOVERY_REQUESTED flags set.

Fix this regression by explicitly setting both flags for the "check"
case (like is also done for the "repair" case, but redundant set_bit()s
are perfectly fine because it adds clarity to what is needed in response
to both messages -- in addition this isn't fast path code).

Fixes: 105db59912 ("dm raid: cleanup awkward branching in raid_message() option processing")
Reported-by: Heinz Mauelshagen <heinzm@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
drivers/md/dm-raid.c

index 5c9e95d66f3b64d14355a91abd936eb5f3c46440..0460cf84fd0e74ac5141b1e5ae001c51156d373f 100644 (file)
@@ -3462,9 +3462,11 @@ static int raid_message(struct dm_target *ti, unsigned int argc, char **argv)
        else if (!strcasecmp(argv[0], "recover"))
                set_bit(MD_RECOVERY_RECOVER, &mddev->recovery);
        else {
-               if (!strcasecmp(argv[0], "check"))
+               if (!strcasecmp(argv[0], "check")) {
                        set_bit(MD_RECOVERY_CHECK, &mddev->recovery);
-               else if (!strcasecmp(argv[0], "repair")) {
+                       set_bit(MD_RECOVERY_REQUESTED, &mddev->recovery);
+                       set_bit(MD_RECOVERY_SYNC, &mddev->recovery);
+               } else if (!strcasecmp(argv[0], "repair")) {
                        set_bit(MD_RECOVERY_REQUESTED, &mddev->recovery);
                        set_bit(MD_RECOVERY_SYNC, &mddev->recovery);
                } else