Re: [PATCH 20/35] Use a pre-calculated value fornum_online_nodes()

[Christoph Lameter]

On Thu, 19 Mar 2009, Mel Gorman wrote:

> This patch actually alters the API. node_set_online() called when
> MAX_NUMNODES == 1 will now fail to compile. That situation wouldn't make
> any sense anyway but is it intentional?

Yes MAX_NUMNODES means that this is not a NUMA configuration. Setting an
ode online would make no sense. Node 0 is always online.

> For reference here is the patch I had for a similar goal which kept the
> API as it was. I'll drop it if you prefer your own version.

Lets look through it and get the best pieces from both.

>  static inline void node_set_state(int node, enum node_states state)
>  {
>  	__node_set(node, &node_states[state]);
> +	if (state == N_ONLINE)
> +		nr_online_nodes = num_node_state(N_ONLINE);
>  }

That assumes uses of node_set_state N_ONLINE. Are there such users or are
all using node_set_online()?

> @@ -449,7 +457,8 @@ static inline int num_node_state(enum node_states state)
>  	node;					\
>  })
>
> -#define num_online_nodes()	num_node_state(N_ONLINE)
> +
> +#define num_online_nodes()	(nr_online_nodes)
>  #define num_possible_nodes()	num_node_state(N_POSSIBLE)
>  #define node_online(node)	node_state((node), N_ONLINE)
>  #define node_possible(node)	node_state((node), N_POSSIBLE)

Hmmmm... Yes we could get rid of those.

I'd also like to see nr_possible_nodes(). nr_possible_nodes is important
if you want to check if the system could ever bring up a second node
(which would make the current optimization not viable) whereas
nr_online_nodes is the check for how many nodes are currently online.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/