pnfs: support for non-rpc layout drivers

Non-rpc layout driver such as for objects and blocks
implement their own I/O path and error handling logic.
Therefore bypass NFS-based error handling for these layout drivers.

[fix lseg ref-count bugs, and null de-refs]
[Fall out from: non-rpc layout drivers]
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
[get rid of PNFS_USE_RPC_CODE]
[get rid of __nfs4_write_done_cb]
[revert useless change in nfs4_write_done_cb]
Signed-off-by: Benny Halevy <bhalevy@panasas.com>

fs/nfs/internal.h

@@ -310,6 +310,7 @@ extern int nfs_migrate_page(struct address_space *,
 #endif
 
 /* nfs4proc.c */
+extern void __nfs4_read_done_cb(struct nfs_read_data *);
 extern void nfs4_reset_read(struct rpc_task *task, struct nfs_read_data *data);
 extern int nfs4_init_client(struct nfs_client *clp,
 			    const struct rpc_timeout *timeparms,

fs/nfs/nfs4proc.c

@@ -3175,6 +3175,11 @@ static int nfs4_proc_pathconf(struct nfs_server *server, struct nfs_fh *fhandle,
 	return err;
 }
 
+void __nfs4_read_done_cb(struct nfs_read_data *data)
+{
+	nfs_invalidate_atime(data->inode);
+}
+
 static int nfs4_read_done_cb(struct rpc_task *task, struct nfs_read_data *data)
 {
 	struct nfs_server *server = NFS_SERVER(data->inode);
@@ -3184,7 +3189,7 @@ static int nfs4_read_done_cb(struct rpc_task *task, struct nfs_read_data *data)
 		return -EAGAIN;
 	}
 
-	nfs_invalidate_atime(data->inode);
+	__nfs4_read_done_cb(data);
 	if (task->tk_status > 0)
 		renew_lease(server, data->timestamp);
 	return 0;
@@ -3198,7 +3203,8 @@ static int nfs4_read_done(struct rpc_task *task, struct nfs_read_data *data)
 	if (!nfs4_sequence_done(task, &data->res.seq_res))
 		return -EAGAIN;
 
-	return data->read_done_cb(task, data);
+	return data->read_done_cb ? data->read_done_cb(task, data) :
+				    nfs4_read_done_cb(task, data);
 }
 
 static void nfs4_proc_read_setup(struct nfs_read_data *data, struct rpc_message *msg)
@@ -3243,7 +3249,8 @@ static int nfs4_write_done(struct rpc_task *task, struct nfs_write_data *data)
 {
 	if (!nfs4_sequence_done(task, &data->res.seq_res))
 		return -EAGAIN;
-	return data->write_done_cb(task, data);
+	return data->write_done_cb ? data->write_done_cb(task, data) :
+		nfs4_write_done_cb(task, data);
 }
 
 /* Reset the the nfs_write_data to send the write to the MDS. */

fs/nfs/pnfs.c

@@ -243,7 +243,7 @@ put_lseg_common(struct pnfs_layout_segment *lseg)
 {
 	struct inode *inode = lseg->pls_layout->plh_inode;
 
-	BUG_ON(test_bit(NFS_LSEG_VALID, &lseg->pls_flags));
+	WARN_ON(test_bit(NFS_LSEG_VALID, &lseg->pls_flags));
 	list_del_init(&lseg->pls_list);
 	if (list_empty(&lseg->pls_layout->plh_segs)) {
 		set_bit(NFS_LAYOUT_DESTROYED, &lseg->pls_layout->plh_flags);
@@ -1054,6 +1054,29 @@ pnfs_pageio_init_write(struct nfs_pageio_descriptor *pgio, struct inode *inode)
 	pgio->pg_test = (ld && ld->pg_test) ? pnfs_write_pg_test : NULL;
 }
 
+/*
+ * Called by non rpc-based layout drivers
+ */
+int
+pnfs_ld_write_done(struct nfs_write_data *data)
+{
+	int status;
+
+	if (!data->pnfs_error) {
+		pnfs_set_layoutcommit(data);
+		data->mds_ops->rpc_call_done(&data->task, data);
+		data->mds_ops->rpc_release(data);
+		return 0;
+	}
+
+	dprintk("%s: pnfs_error=%d, retry via MDS\n", __func__,
+		data->pnfs_error);
+	status = nfs_initiate_write(data, NFS_CLIENT(data->inode),
+				    data->mds_ops, NFS_FILE_SYNC);
+	return status ? : -EAGAIN;
+}
+EXPORT_SYMBOL_GPL(pnfs_ld_write_done);
+
 enum pnfs_try_status
 pnfs_try_to_write_data(struct nfs_write_data *wdata,
 			const struct rpc_call_ops *call_ops, int how)
@@ -1078,6 +1101,29 @@ pnfs_try_to_write_data(struct nfs_write_data *wdata,
 	return trypnfs;
 }
 
+/*
+ * Called by non rpc-based layout drivers
+ */
+int
+pnfs_ld_read_done(struct nfs_read_data *data)
+{
+	int status;
+
+	if (!data->pnfs_error) {
+		__nfs4_read_done_cb(data);
+		data->mds_ops->rpc_call_done(&data->task, data);
+		data->mds_ops->rpc_release(data);
+		return 0;
+	}
+
+	dprintk("%s: pnfs_error=%d, retry via MDS\n", __func__,
+		data->pnfs_error);
+	status = nfs_initiate_read(data, NFS_CLIENT(data->inode),
+				   data->mds_ops);
+	return status ? : -EAGAIN;
+}
+EXPORT_SYMBOL_GPL(pnfs_ld_read_done);
+
 /*
  * Call the appropriate parallel I/O subsystem read function.
  */

fs/nfs/pnfs.h

@@ -165,6 +165,8 @@ void pnfs_roc_set_barrier(struct inode *ino, u32 barrier);
 bool pnfs_roc_drain(struct inode *ino, u32 *barrier);
 void pnfs_set_layoutcommit(struct nfs_write_data *wdata);
 int pnfs_layoutcommit_inode(struct inode *inode, bool sync);
+int pnfs_ld_write_done(struct nfs_write_data *);
+int pnfs_ld_read_done(struct nfs_read_data *);
 
 /* pnfs_dev.c */
 struct nfs4_deviceid_node {

include/linux/nfs_xdr.h

@@ -1087,6 +1087,7 @@ struct nfs_read_data {
 	const struct rpc_call_ops *mds_ops;
 	int (*read_done_cb) (struct rpc_task *task, struct nfs_read_data *data);
 	__u64			mds_offset;
+	int			pnfs_error;
 	struct page		*page_array[NFS_PAGEVEC_SIZE];
 };
 
@@ -1112,6 +1113,7 @@ struct nfs_write_data {
 	unsigned long		timestamp;	/* For lease renewal */
 #endif
 	__u64			mds_offset;	/* Filelayout dense stripe */
+	int			pnfs_error;
 	struct page		*page_array[NFS_PAGEVEC_SIZE];
 };