Re: [PATCH 1/4] numa,x86: store maximum numa node distance

[Rik van Riel <riel@redhat.com>]

On 05/09/2014 05:45 AM, Peter Zijlstra wrote:
> On Thu, May 08, 2014 at 01:23:28PM -0400, riel@redhat.com wrote:
>> From: Rik van Riel <riel@redhat.com>
>>
>> Store the maximum node distance, so the numa placement code can do
>> better placement on systems with complex numa topology.
>>
>> The function max_node_distance will return LOCAL_DISTANCE if the
>> system has simple NUMA topology, with only a single level of
>> remote distance.
>>
>> Signed-off-by: Rik van Riel <riel@redhat.com>
>> Tested-by: Chegu Vinod <chegu_vinod@hp.com>
>> ---
>>   arch/x86/include/asm/topology.h |  3 +++
>>   arch/x86/mm/numa.c              | 25 +++++++++++++++++++++++++
>>   include/linux/topology.h        |  3 +++
>>   3 files changed, 31 insertions(+)
>>
>
> Why are you doing this in arch code? I would've expected some extra code
> to sched_init_numa() which is generic code that analyses the distance
> table and reconstructs the actual topology from it.

The only reason it is in the arch code is that node_distance()
is in arch code today.

If there is no good reason for node_distance and the node
distances array to be in arch code, I could take a stab at
making it generic...