Re: [RFC PATCH v2] SLUB: enhance slub to handle memory nodes without normal memory

[Jiang Liu <liuj97@gmail.com>]

On 07/24/2012 10:45 PM, Christoph Lameter wrote:
> On Tue, 24 Jul 2012, Jiang Liu wrote:
> 
>>
>> diff --git a/mm/slub.c b/mm/slub.c
>> index 8c691fa..3976745 100644
>> --- a/mm/slub.c
>> +++ b/mm/slub.c
>> @@ -2803,6 +2803,17 @@ static inline int alloc_kmem_cache_cpus(struct kmem_cache *s)
>>
>>  static struct kmem_cache *kmem_cache_node;
>>
>> +static bool node_has_normal_memory(int node)
>> +{
>> +	int i;
>> +
>> +	for (i = ZONE_NORMAL; i >= 0; i--)
>> +		if (populated_zone(&NODE_DATA(node)->node_zones[i]))
>> +			return true;
>> +
>> +	return false;
>> +}
> 
> There is already a N_NORMAL_MEMORY node map that contains a list of node
> that have *normal* memory usable by slab allocators etc. I think the
> cleanest solution would be to clear the corresponding node bits for your
> special movable only zones. Then you wont be needing to modify other
> subsystems anymore.
> 
Hi Chris,
	Thanks for your comments! I have thought about the solution mentioned,
but seems it doesn't work. We have node masks for both N_NORMAL_MEMORY and
N_HIGH_MEMORY to distinguish between normal and highmem on platforms such as x86.
But we still don't have such a mechanism to distinguish between "normal" and "movable"
memory. So for memory nodes with only movable zones, we still set N_NORMAL_MEMORY for
them. One possible solution is to add a node mask for "N_NORMAL_OR_MOVABLE_MEMORY",
but haven't tried that yet. Will have a try for that.
	Thanks!
	Gerry

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>