test/libcephfs: put inodes after lookup

Otherwise, the client umount will hang due to inability to trim the
inodes looked up using the low-level interface. This results in slow-op
warnings and an eviction:

	2021-09-11T17:23:31.097+0000 7f99c3522700  0 log_channel(cluster) log [WRN] : evicting unresponsive client smithi176 (9756), after 303.924 seconds
	2021-09-11T17:23:31.097+0000 7f99c3522700 10 mds.0.server autoclosing stale session client.9756 172.21.15.176:0/3891214934 last renewed caps 303.924s ago

From: /ceph/teuthology-archive/yuriw-2021-09-11_16:21:09-smoke-pacific-distro-basic-smithi/6385038/remote/smithi175/log/ceph-mds.b.log.gz

Fixes: https://tracker.ceph.com/issues/52572
Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit c0252063b94d811dc7863058999856ac5614d1eb)

 Conflicts:
	src/test/libcephfs/test.cc

diff --git a/src/test/libcephfs/test.cc b/src/test/libcephfs/test.cc
index 1c28bb051b2e59ab6b0532e4ebc02ac16bb62296..950b9aa54e954bf616c3fc69d79fc3b596f3b9c0 100644 (file)
--- a/src/test/libcephfs/test.cc
+++ b/src/test/libcephfs/test.cc
@@ -3524,11 +3524,16 @@ TEST(LibCephFS, LookupMdsPrivateInos) {
       ASSERT_EQ(-ESTALE, ceph_ll_lookup_inode(cmount, ino, &inode));
     } else if (ino == CEPH_INO_ROOT || ino == CEPH_INO_GLOBAL_SNAPREALM) {
       ASSERT_EQ(0, ceph_ll_lookup_inode(cmount, ino, &inode));
+      ceph_ll_put(cmount, inode);
     } else if (ino == CEPH_INO_LOST_AND_FOUND) {
       // the ino 3 will only exists after the recovery tool ran, so
       // it may return -ESTALE with a fresh fs cluster
       int r = ceph_ll_lookup_inode(cmount, ino, &inode);
-      ASSERT_TRUE(r == -ESTALE || r == 0);
+      if (r == 0) {
+        ceph_ll_put(cmount, inode);
+      } else {
+        ASSERT_TRUE(r == -ESTALE);
+      }
     } else {
       // currently the ino 0 and 4~99 is not useded yet.
       ASSERT_EQ(-ESTALE, ceph_ll_lookup_inode(cmount, ino, &inode));