Re: [PATCH] x86, numa: seperate alloc_numa_distance from numa_reset_distance.

[Cyrill Gorcunov <gorcunov@gmail.com>]

On 02/17/2011 01:29 AM, Yinghai Lu wrote:
> On 02/16/2011 01:44 PM, Cyrill Gorcunov wrote:
>> On 02/16/2011 11:57 PM, Yinghai Lu wrote:
>>>
>>> alloc code is much bigger the setting self.
>>>
>> ...
>>> +
>>> +    numa_distance = __va(phys);
>>> +    numa_distance_cnt = cnt;
>>> +
>>> +    /* fill with the default distances */
>>> +    for (i = 0; i<   cnt; i++)
>>> +        for (j = 0; j<   cnt; j++)
>>> +            numa_distance[i * cnt + j] = i == j ?
>>> +                LOCAL_DISTANCE : REMOTE_DISTANCE;
>>> +    printk(KERN_DEBUG "NUMA: Initialized distance table, cnt=%d\n",
>>> cnt);
>>> +}
>>> +
>>
>>    Hi Yinghai, btw would it be convenient to use node_distance() helper here
>> (since this snippet is touched anyway and there is a big merging of numa
>>   code), ie
>>
>>      /* fill with the default distances */
>>      for (i = 0; i<  cnt; i++)
>>          for (j = 0; j<   cnt; j++)
>>              numa_distance[i * cnt + j] = node_distance(i, j);
>>
>> though it should be patch on top of yours to not mix with plain code move
>
> do you mean
>
> int __node_distance(int from, int to)
> {
>          if (from>= numa_distance_cnt || to>= numa_distance_cnt)
>                  return from == to ? LOCAL_DISTANCE : REMOTE_DISTANCE;
>          return numa_distance[from * numa_distance_cnt + to];
> }
> EXPORT_SYMBOL(__node_distance);
>
> or node_distance()?
>
> then you will have
>
>   numa_distance[i * cnt + j] =  numa_distance[i * cnt + j];
>
> ...

Ah, yeah, I mentally swapped >= to <= which is not true of course,
drop this idea, sorry.

-- 
     Cyrill