If dm_get_opencount() fails, open_count will be negative, which means
mpath_in_use() will never fail. However, dm_flush_map__() will
eventually fail. There are multiple callers of dm_get_opencount() that
run when trying to remove a device. All the others treat a failed call
as if the device was in use, including one that will later be called
directly on this device. So save ourselves the work, and exit early
Signed-off-by: Benjamin Marzinski <bmarz...@redhat.com>
---
libmultipath/devmapper.c | 5 +++++
1 file changed, 5 insertions(+)
diff --git a/libmultipath/devmapper.c b/libmultipath/devmapper.c
index 751b45d8..3667c51b 100644
--- a/libmultipath/devmapper.c
+++ b/libmultipath/devmapper.c
@@ -961,6 +961,11 @@ int mpath_in_use(const char *name)
{
int open_count = dm_get_opencount(name);
+ if (open_count < 0) {
+ condlog(0, "%s: %s: failed to get open count, assuming in use",
+ __func__, name);
+ return 1;
+ }
if (open_count) {
int part_count = 0;
--
2.46.2
