IB/mlx4: Add support for memory management extensions and local DMA L_Key

Add support for the following operations to mlx4 when device firmware
supports them:

 - Send with invalidate and local invalidate send queue work requests;
 - Allocate/free fast register MRs;
 - Allocate/free fast register MR page lists;
 - Fast register MR send queue work requests;
 - Local DMA L_Key.

Signed-off-by: Roland Dreier <rolandd@cisco.com>
diff --git a/include/linux/mlx4/qp.h b/include/linux/mlx4/qp.h
index f02e9ed..e27082c 100644
--- a/include/linux/mlx4/qp.h
+++ b/include/linux/mlx4/qp.h
@@ -233,6 +233,14 @@
 	__be64			length;
 };
 
+enum {
+	MLX4_WQE_FMR_PERM_LOCAL_READ	= 1 << 27,
+	MLX4_WQE_FMR_PERM_LOCAL_WRITE	= 1 << 28,
+	MLX4_WQE_FMR_PERM_REMOTE_READ	= 1 << 29,
+	MLX4_WQE_FMR_PERM_REMOTE_WRITE	= 1 << 30,
+	MLX4_WQE_FMR_PERM_ATOMIC	= 1 << 31
+};
+
 struct mlx4_wqe_fmr_seg {
 	__be32			flags;
 	__be32			mem_key;
@@ -255,11 +263,11 @@
 };
 
 struct mlx4_wqe_local_inval_seg {
-	u8			flags;
-	u8			reserved1[3];
+	__be32			flags;
+	u32			reserved1;
 	__be32			mem_key;
-	u8			reserved2[3];
-	u8			guest_id;
+	u32			reserved2[2];
+	__be32			guest_id;
 	__be64			pa;
 };