Re: [PATCH x86/mm UPDATED] x86-64, NUMA: Fix distance table handling

[Yinghai Lu <yinghai@kernel.org>]

On 03/02/2011 08:37 AM, Tejun Heo wrote:
> Hey, Yinghai.
> 
> On Wed, Mar 02, 2011 at 08:16:18AM -0800, Yinghai Lu wrote:
>> my original part:
>>
>> @@ -393,7 +393,7 @@ void __init numa_reset_distance(void)
>>         size_t size;
>>  
>>         if (numa_distance_cnt) {
>> -               size = numa_distance_cnt * sizeof(numa_distance[0]);
>> +               size = numa_distance_cnt * numa_distance_cnt * sizeof(numa_distance[0]);
>>                 memblock_x86_free_range(__pa(numa_distance),
>>                                         __pa(numa_distance) + size);
>>                 numa_distance_cnt = 0;
>>
>> So can you tell me why you need to make those change?
>> 	move out assigning or numa_distance_cnt and size of the the IF
> 
> Please read the patch description.  I actually wrote that down.  :-)

well you said:
> while at it, take numa_distance_cnt resetting in
> numa_reset_distance() out of the if block to simplify the code a bit.

what are you talking about? what do you mean "simplify the code a bit" ?

> 
>> the change include:
>> 1. you only need to go over new_nr*new_nr instead huge MAX_NUMNODES * MAX_NUMNODES
>> 2. you do NOT need to go over it if you don't have phys_dist assigned before.
>>    numa_alloc_distance already have that default set.
>> 3. do need to check if phys_dist is assigned before referring phys_dist.
> 
> * If you wanted to make that change, split it into a separate patch.
>   Don't mix it with changes which actually fix the bug.
> 
> * I don't think it's gonna matter all that much.  It's one time and
>   only used if emulation is enabled, but then again yeap MAX_NUMNODES
>   * MAX_NUMNODES can get quite high, but it looks way too complicated
>   for what it achieves.  Just looping over enabled nodes should
>   achieve about the same thing in much simpler way, right?

what kind of excuse to put inefficiency code there!

Yinghai