From: Patrick Donnelly <pdonnell@redhat.com>
Date: Fri, 5 Jan 2024 20:06:20 +0000 (-0500)
Subject: mds: add comment on locking order rules
X-Git-Tag: testing/wip-root-testing-20240411.174241~89^2~56
X-Git-Url: http://git.apps.os.sepia.ceph.com/?a=commitdiff_plain;h=48be885ae1e56cf3eb034773b8e1bc7b441d5213;p=ceph-ci.git

mds: add comment on locking order rules

From a past commit.

Signed-off-by: Patrick Donnelly <pdonnell@redhat.com>
(cherry picked from commit 32a544f9a71aa68547db5bb0fdd517cafbf7d928)
---

diff --git a/src/mds/MDCache.cc b/src/mds/MDCache.cc
index f79e806494b..5022b23dd0f 100644
--- a/src/mds/MDCache.cc
+++ b/src/mds/MDCache.cc
@@ -8240,6 +8240,24 @@ void MDCache::dispatch(const cref_t<Message> &m)
   }
 }
 
+/**
+ * In 246f647566095c173e5e0e54661696cea230f96e, an updated rule for locking order
+ * was established (differing from past strategies):
+ *
+ *  [The helper function is for requests that operate on two paths. It
+ *  ensures that the two paths get locks in proper order.] The rule is:
+ *
+ *   1. Lock directory inodes or dentries according to which trees they
+ *      are under. Lock objects under fs root before objects under mdsdir.
+ *   2. Lock directory inodes or dentries according to their depth, in
+ *      ascending order.
+ *   3. Lock directory inodes or dentries according to inode numbers or
+ *      dentries' parent inode numbers, in ascending order.
+ *   4. Lock dentries in the same directory in order of their keys.
+ *   5. Lock non-directory inodes according to inode numbers, in ascending
+ *      order.
+ */
+
 int MDCache::path_traverse(const MDRequestRef& mdr, MDSContextFactory& cf,
                            const filepath& path, int flags,
                            vector<CDentry*> *pdnvec, CInode **pin)