On 08/02/2012 06:45 PM, Christoph Lameter wrote:
> On Thu, 2 Aug 2012, Glauber Costa wrote:
>
>> It also works okay both before the patches are applied, and with slab.
>
> Ok. I am seeing the same problem when using the following patch. That is
> pretty early during boot and so there may be issues with sysfs that the
> patchset caused. Looking into it.
>
> Index: linux-2.6/mm/slub.c
> ===================================================================
> --- linux-2.6.orig/mm/slub.c 2012-08-02 09:36:04.855637689 -0500
> +++ linux-2.6/mm/slub.c 2012-08-02 09:42:04.358089667 -0500
> @@ -3768,6 +3768,16 @@
> caches, cache_line_size(),
> slub_min_order, slub_max_order, slub_min_objects,
> nr_cpu_ids, nr_node_ids);
> +
> + { struct kmem_cache *qq;
> +
> + qq = create_kmalloc_cache("qq", 800, 0);
> + kmem_cache_destroy(qq);
> +
> + qq = create_kmalloc_cache("qq", 800, 0);
> + kmem_cache_destroy(qq);
> + }
> +
> }
>
> void __init kmem_cache_init_late(void)
>
Mine is similar, except:
1) I don't create a kmalloc cache (shouldn't matter)
2) I do it after SLAB_FULL.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>