locks: provide a file lease method enabling cluster-coherent leases
Currently leases are only kept locally, so there's no way for a distributed
filesystem to enforce them against multiple clients. We're particularly
interested in the case of nfsd exporting a cluster filesystem, in which
case nfsd needs cluster-coherent leases in order to implement delegations
correctly.
Also add some documentation.
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
diff --git a/fs/locks.c b/fs/locks.c
index a65d85c..94f5d80 100644
--- a/fs/locks.c
+++ b/fs/locks.c
@@ -1429,7 +1429,24 @@
* @lease: file_lock to use
*
* Call this to establish a lease on the file.
- * The fl_lmops fl_break function is required by break_lease
+ * The (*lease)->fl_lmops->fl_break operation must be set; if not,
+ * break_lease will oops!
+ *
+ * This will call the filesystem's setlease file method, if
+ * defined. Note that there is no getlease method; instead, the
+ * filesystem setlease method should call back to setlease() to
+ * add a lease to the inode's lease list, where fcntl_getlease() can
+ * find it. Since fcntl_getlease() only reports whether the current
+ * task holds a lease, a cluster filesystem need only do this for
+ * leases held by processes on this node.
+ *
+ * There is also no break_lease method; filesystems that
+ * handle their own leases shoud break leases themselves from the
+ * filesystem's open, create, and (on truncate) setattr methods.
+ *
+ * Warning: the only current setlease methods exist only to disable
+ * leases in certain cases. More vfs changes may be required to
+ * allow a full filesystem lease implementation.
*/
int vfs_setlease(struct file *filp, long arg, struct file_lock **lease)
@@ -1437,7 +1454,10 @@
int error;
lock_kernel();
- error = setlease(filp, arg, lease);
+ if (filp->f_op && filp->f_op->setlease)
+ error = filp->f_op->setlease(filp, arg, lease);
+ else
+ error = setlease(filp, arg, lease);
unlock_kernel();
return error;