RDMA/mlx5: Simplify how the MR cache bucket is located

There are many bad APIs here that are accepting a cache bucket index
instead of a bucket pointer. Many of the callers already have a bucket
pointer, so this results in a lot of confusing uses of order2idx().

Pass the struct mlx5_cache_ent into add_keys(), remove_keys(), and
alloc_cached_mr().

Once the MR is in the cache, store the cache bucket pointer directly in
the MR, replacing the 'bool allocated_from cache'.

In the end there is only one place that needs to form index from order,
alloc_mr_from_cache(). Increase the safety of this function by disallowing
it from accessing cache entries in the ODP special area.

Link: https://lore.kernel.org/r/20200310082238.239865-7-leon@kernel.org
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Jason Gunthorpe <jgg@mellanox.com>

drivers/infiniband/hw/mlx5/mlx5_ib.h

@@ -617,8 +617,8 @@ struct mlx5_ib_mr {
 	struct ib_umem	       *umem;
 	struct mlx5_shared_mr_info	*smr_info;
 	struct list_head	list;
-	int			order;
-	bool			allocated_from_cache;
+	unsigned int		order;
+	struct mlx5_cache_ent  *cache_ent;
 	int			npages;
 	struct mlx5_ib_dev     *dev;
 	u32 out[MLX5_ST_SZ_DW(create_mkey_out)];
@@ -1274,7 +1274,8 @@ int mlx5_ib_get_cqe_size(struct ib_cq *ibcq);
 int mlx5_mr_cache_init(struct mlx5_ib_dev *dev);
 int mlx5_mr_cache_cleanup(struct mlx5_ib_dev *dev);
 
-struct mlx5_ib_mr *mlx5_mr_cache_alloc(struct mlx5_ib_dev *dev, int entry);
+struct mlx5_ib_mr *mlx5_mr_cache_alloc(struct mlx5_ib_dev *dev,
+				       unsigned int entry);
 void mlx5_mr_cache_free(struct mlx5_ib_dev *dev, struct mlx5_ib_mr *mr);
 int mlx5_mr_cache_invalidate(struct mlx5_ib_mr *mr);