IB/mlx5: Reset access mask when looping inside page fault handler

If page-fault handler spans multiple MRs then the access mask needs to
be reset before each MR handling or otherwise write access will be
granted to mapped pages instead of read-only.

Cc: <stable@vger.kernel.org> # 3.19
Fixes: 7bdf65d ("IB/mlx5: Handle page faults")
Reported-by: Jerome Glisse <jglisse@redhat.com>
Signed-off-by: Moni Shoua <monis@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>

drivers/infiniband/hw/mlx5/odp.c

@@ -585,7 +585,7 @@ static int pagefault_mr(struct mlx5_ib_dev *dev, struct mlx5_ib_mr *mr,
 	struct ib_umem_odp *odp_mr = to_ib_umem_odp(mr->umem);
 	bool downgrade = flags & MLX5_PF_FLAGS_DOWNGRADE;
 	bool prefetch = flags & MLX5_PF_FLAGS_PREFETCH;
-	u64 access_mask = ODP_READ_ALLOWED_BIT;
+	u64 access_mask;
 	u64 start_idx, page_mask;
 	struct ib_umem_odp *odp;
 	size_t size;
@@ -607,6 +607,7 @@ static int pagefault_mr(struct mlx5_ib_dev *dev, struct mlx5_ib_mr *mr,
 	page_shift = mr->umem->page_shift;
 	page_mask = ~(BIT(page_shift) - 1);
 	start_idx = (io_virt - (mr->mmkey.iova & page_mask)) >> page_shift;
+	access_mask = ODP_READ_ALLOWED_BIT;
 
 	if (prefetch && !downgrade && !mr->umem->writable) {
 		/* prefetch with write-access must