NFS: Use nfs_refresh_inode() in ops that aren't expected to change the inode

nfs_post_op_update_inode() is really only meant to be used if we expect the
inode and its attributes to have changed in some way.

Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>

+3 -5
+1 -2
fs/nfs/nfs3proc.c
··· 777 777 { 778 778 if (nfs3_async_handle_jukebox(task, data->inode)) 779 779 return -EAGAIN; 780 - if (task->tk_status >= 0) 781 - nfs_post_op_update_inode(data->inode, data->res.fattr); 780 + nfs_refresh_inode(data->inode, data->res.fattr); 782 781 return 0; 783 782 } 784 783
+2 -3
fs/nfs/nfs4proc.c
··· 2475 2475 rpc_restart_call(task); 2476 2476 return -EAGAIN; 2477 2477 } 2478 - if (task->tk_status >= 0) 2479 - nfs_post_op_update_inode(inode, data->res.fattr); 2478 + nfs_refresh_inode(inode, data->res.fattr); 2480 2479 return 0; 2481 2480 } 2482 2481 ··· 3045 3046 if (status == 0) { 3046 3047 status = data->rpc_status; 3047 3048 if (status == 0) 3048 - nfs_post_op_update_inode(inode, &data->fattr); 3049 + nfs_refresh_inode(inode, &data->fattr); 3049 3050 } 3050 3051 rpc_put_task(task); 3051 3052 return status;