- changed name convention in cobd, thus all MD related methods have _md_...

- changed name convention in cobd, thus all MD related methods have _md_ prefix like cobd_md_getstatus(),
all data layer related have _dt_prefix and common one like cobd_connect() have no prefix at all.

- added hack to mdc_intent_lock() and cobd_md_intent_lock() to distinguish cobd case in mdc_intent_lock() and do not raise -ESTALE if obtained lock is not coherent wiht store id due to generation after switchin to cache cobd. This is dirty hack of course, but it should be for a while to pass cmobd/cobd related tests and Peter agreed it ok for now.

lustre/include/linux/lustre_compat25.h

@@ -95,6 +95,14 @@ static inline int cleanup_group_info(void)
 }
 #endif /* end of groups_info stuff */
 
+/*
+ * this define is from the same namespace as other lookup flags in linux, will
+ * be gone when COBD medium switching will be done more correct manner.
+ * Currently this is dirty hack and this flag is needed to let MDC layer know
+ * that it does not have to check if requested id is the same as receviced from
+ * MDS in mdc_intent_lock() --umka
+ */
+#define LOOKUP_COBD 4096
 
 #if LINUX_VERSION_CODE >= KERNEL_VERSION(2,5,0)
 

lustre/mdc/mdc_locks.c

@@ -602,7 +602,7 @@ int mdc_intent_lock(struct obd_export *exp, struct lustre_id *pid,
                 
                 /* we have to compare all the fields but type, because MDS can
                  * return fid/mds/ino/gen if inode lives on another MDS -bzzz */
-                if (!id_equal(cid, &mds_body->id1))
+                if ((lookup_flags & LOOKUP_COBD) && !id_equal(cid, &mds_body->id1))
                         RETURN(-ESTALE);
         }