features/locks: Clean up logging of cleanup in DISCONNECT codepath

        Backport of http://review.gluster.org/7981

Now, gfid is printed as opposed to path in cleanup messages.

Also, refkeeper update is eliminated in inodelk and entrylk.
Instead, the patch ensures inode and pl_inode are kept alive as
long as there is atleast one lock (granted/blocked) on an inode.

Also, every inode is unref'd appropriately on a DISCONNECT from the
lock-owning client.

Change-Id: I234db688ad0d314f4936a16cc5af70a3bd071970
BUG: 1104915
Signed-off-by: Krutika Dhananjay <kdhananj@redhat.com>
Reviewed-on: http://review.gluster.org/8042
Tested-by: Gluster Build System <jenkins@build.gluster.com>
Reviewed-by: Pranith Kumar Karampuri <pkarampu@redhat.com>
Reviewed-by: Niels de Vos <ndevos@redhat.com>

1 files changed, 4 insertions, 0 deletions

diff --git a/xlators/features/locks/src/locks.h b/xlators/features/locks/src/locks.h
index 8c2a6f867ee..f761b3d4a00 100644
--- a/xlators/features/locks/src/locks.h
+++ b/xlators/features/locks/src/locks.h
@@ -146,6 +146,10 @@ struct __pl_inode {
 
         inode_t          *refkeeper;     /* hold refs on an inode while locks are
                                             held to prevent pruning */
+        uuid_t            gfid;          /* placeholder for gfid of the inode */
+        inode_t          *inode;         /* pointer to be used for ref and unref
+                                            of inode_t as long as there are
+                                            locks on it */
 };
 typedef struct __pl_inode pl_inode_t;