ocfs2: Negotiate locking protocol versions.
Currently, when ocfs2 nodes connect via TCP, they advertise their
compatibility level. If the versions do not match, two nodes cannot speak
to each other and they disconnect. As a result, this provides no forward or
backwards compatibility.
This patch implements a simple protocol negotiation at the dlm level by
introducing a major/minor version number scheme for entities that
communicate. Specifically, o2dlm has a major/minor version for interaction
with o2dlm on other nodes, and ocfs2 itself has a major/minor version for
interacting with the filesystem on other nodes.
This will allow rolling upgrades of ocfs2 clusters when changes to the
locking or network protocols can be done in a backwards compatible manner.
In those cases, only the minor number is changed and the negotatied protocol
minor is returned from dlm join. In the far less likely event that a
required protocol change makes backwards compatibility impossible, we simply
bump the major number.
Signed-off-by: Joel Becker <joel.becker@oracle.com>
Signed-off-by: Mark Fasheh <mark.fasheh@oracle.com>
diff --git a/fs/ocfs2/dlmglue.c b/fs/ocfs2/dlmglue.c
index 3867244..351130c 100644
--- a/fs/ocfs2/dlmglue.c
+++ b/fs/ocfs2/dlmglue.c
@@ -43,6 +43,7 @@
#include <cluster/masklog.h>
#include "ocfs2.h"
+#include "ocfs2_lockingver.h"
#include "alloc.h"
#include "dcache.h"
@@ -258,6 +259,31 @@
.flags = 0,
};
+/*
+ * This is the filesystem locking protocol version.
+ *
+ * Whenever the filesystem does new things with locks (adds or removes a
+ * lock, orders them differently, does different things underneath a lock),
+ * the version must be changed. The protocol is negotiated when joining
+ * the dlm domain. A node may join the domain if its major version is
+ * identical to all other nodes and its minor version is greater than
+ * or equal to all other nodes. When its minor version is greater than
+ * the other nodes, it will run at the minor version specified by the
+ * other nodes.
+ *
+ * If a locking change is made that will not be compatible with older
+ * versions, the major number must be increased and the minor version set
+ * to zero. If a change merely adds a behavior that can be disabled when
+ * speaking to older versions, the minor version must be increased. If a
+ * change adds a fully backwards compatible change (eg, LVB changes that
+ * are just ignored by older versions), the version does not need to be
+ * updated.
+ */
+const struct dlm_protocol_version ocfs2_locking_protocol = {
+ .pv_major = OCFS2_LOCKING_PROTOCOL_MAJOR,
+ .pv_minor = OCFS2_LOCKING_PROTOCOL_MINOR,
+};
+
static inline int ocfs2_is_inode_lock(struct ocfs2_lock_res *lockres)
{
return lockres->l_type == OCFS2_LOCK_TYPE_META ||
@@ -2506,7 +2532,8 @@
dlm_key = crc32_le(0, osb->uuid_str, strlen(osb->uuid_str));
/* for now, uuid == domain */
- dlm = dlm_register_domain(osb->uuid_str, dlm_key);
+ dlm = dlm_register_domain(osb->uuid_str, dlm_key,
+ &osb->osb_locking_proto);
if (IS_ERR(dlm)) {
status = PTR_ERR(dlm);
mlog_errno(status);