Commit ca19f865f moved adding DI_BLACKLIST to the pathinfo flags out of
store_pathinfo(), but it didn't add it to all of the necessary callers.
Without this, store_pathinfo() callers can do unnecessary extra work,
including running the path checker on blacklisted devices. Also, running

multipathd add path <blacklisted_path_device>

will add a blacklisted path.

Fixes: ca19f865f "libmultipath: add 'cmd' as argument for get_refwwid()"
Cc: Hannes Reinecke <h...@suse.com>
Signed-off-by: Benjamin Marzinski <bmarz...@redhat.com>
---
 libmultipath/discovery.c  | 3 ++-
 multipathd/cli_handlers.c | 2 +-
 2 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/libmultipath/discovery.c b/libmultipath/discovery.c
index b3dc60b1..e9411dfc 100644
--- a/libmultipath/discovery.c
+++ b/libmultipath/discovery.c
@@ -134,7 +134,8 @@ path_discover (vector pathvec, struct config * conf,
                pp = find_path_by_devt(pathvec, devt);
                if (!pp)
                        return store_pathinfo(pathvec, conf,
-                                             udevice, flag, NULL);
+                                             udevice, flag | DI_BLACKLIST,
+                                             NULL);
        }
        return pathinfo(pp, conf, flag);
 }
diff --git a/multipathd/cli_handlers.c b/multipathd/cli_handlers.c
index 60e17d68..4c32d953 100644
--- a/multipathd/cli_handlers.c
+++ b/multipathd/cli_handlers.c
@@ -730,7 +730,7 @@ cli_add_path (void * v, char ** reply, int * len, void * 
data)
                conf = get_multipath_config();
                pthread_cleanup_push(put_multipath_config, conf);
                r = store_pathinfo(vecs->pathvec, conf,
-                                  udevice, DI_ALL, &pp);
+                                  udevice, DI_ALL | DI_BLACKLIST, &pp);
                pthread_cleanup_pop(1);
                udev_device_unref(udevice);
                if (!pp) {
-- 
2.17.2

--
dm-devel mailing list
dm-devel@redhat.com
https://www.redhat.com/mailman/listinfo/dm-devel

Reply via email to